diff --git a/drivers/char/agp/amd64-agp.c b/drivers/char/agp/amd64-agp.c index 780498d..444f8b6 100644 --- a/drivers/char/agp/amd64-agp.c +++ b/drivers/char/agp/amd64-agp.c @@ -33,7 +33,7 @@ #define ULI_X86_64_ENU_SCR_REG 0x54 static struct resource *aperture_resource; -static int __initdata agp_try_unsupported = 1; +static bool __initdata agp_try_unsupported = 1; static int agp_bridges_found; static void amd64_tlbflush(struct agp_memory *temp) diff --git a/drivers/char/agp/backend.c b/drivers/char/agp/backend.c index 4b71647..317c28c 100644 --- a/drivers/char/agp/backend.c +++ b/drivers/char/agp/backend.c @@ -194,10 +194,10 @@ static int agp_backend_initialize(struct agp_bridge_data *bridge) err_out: if (bridge->driver->needs_scratch_page) { - void *va = page_address(bridge->scratch_page_page); + struct page *page = bridge->scratch_page_page; - bridge->driver->agp_destroy_page(va, AGP_PAGE_DESTROY_UNMAP); - bridge->driver->agp_destroy_page(va, AGP_PAGE_DESTROY_FREE); + bridge->driver->agp_destroy_page(page, AGP_PAGE_DESTROY_UNMAP); + bridge->driver->agp_destroy_page(page, AGP_PAGE_DESTROY_FREE); } if (got_gatt) bridge->driver->free_gatt_table(bridge); @@ -221,10 +221,10 @@ static void agp_backend_cleanup(struct agp_bridge_data *bridge) if (bridge->driver->agp_destroy_page && bridge->driver->needs_scratch_page) { - void *va = page_address(bridge->scratch_page_page); + struct page *page = bridge->scratch_page_page; - bridge->driver->agp_destroy_page(va, AGP_PAGE_DESTROY_UNMAP); - bridge->driver->agp_destroy_page(va, AGP_PAGE_DESTROY_FREE); + bridge->driver->agp_destroy_page(page, AGP_PAGE_DESTROY_UNMAP); + bridge->driver->agp_destroy_page(page, AGP_PAGE_DESTROY_FREE); } } diff --git a/drivers/char/agp/generic.c b/drivers/char/agp/generic.c index b072648..17e05d1 100644 --- a/drivers/char/agp/generic.c +++ b/drivers/char/agp/generic.c @@ -514,12 +514,12 @@ static void agp_v2_parse_one(u32 *requested_mode, u32 *bridge_agpstat, u32 *vga_ switch (*bridge_agpstat & 7) { case 4: *bridge_agpstat |= (AGPSTAT2_2X | AGPSTAT2_1X); - printk(KERN_INFO PFX "BIOS bug. AGP bridge claims to only support x4 rate" + printk(KERN_INFO PFX "BIOS bug. AGP bridge claims to only support x4 rate. " "Fixing up support for x2 & x1\n"); break; case 2: *bridge_agpstat |= AGPSTAT2_1X; - printk(KERN_INFO PFX "BIOS bug. AGP bridge claims to only support x2 rate" + printk(KERN_INFO PFX "BIOS bug. AGP bridge claims to only support x2 rate. " "Fixing up support for x1\n"); break; default: @@ -693,7 +693,7 @@ static void agp_v3_parse_one(u32 *requested_mode, u32 *bridge_agpstat, u32 *vga_ *bridge_agpstat &= ~(AGPSTAT3_4X | AGPSTAT3_RSVD); *vga_agpstat &= ~(AGPSTAT3_4X | AGPSTAT3_RSVD); } else { - printk(KERN_INFO PFX "Fell back to AGPx4 mode because"); + printk(KERN_INFO PFX "Fell back to AGPx4 mode because "); if (!(*bridge_agpstat & AGPSTAT3_8X)) { printk(KERN_INFO PFX "bridge couldn't do x8. bridge_agpstat:%x (orig=%x)\n", *bridge_agpstat, origbridge); @@ -956,7 +956,7 @@ int agp_generic_create_gatt_table(struct agp_bridge_data *bridge) bridge->driver->cache_flush(); #ifdef CONFIG_X86 if (set_memory_uc((unsigned long)table, 1 << page_order)) - printk(KERN_WARNING "Could not set GATT table memory to UC!"); + printk(KERN_WARNING "Could not set GATT table memory to UC!\n"); bridge->gatt_table = (void *)table; #else diff --git a/drivers/char/agp/intel-agp.c b/drivers/char/agp/intel-agp.c index 58b49d1..4293c48 100644 --- a/drivers/char/agp/intel-agp.c +++ b/drivers/char/agp/intel-agp.c @@ -850,6 +850,7 @@ static struct pci_device_id agp_intel_pci_table[] = { .subvendor = PCI_ANY_ID, \ .subdevice = PCI_ANY_ID, \ } + ID(PCI_DEVICE_ID_INTEL_82441), /* for HAS2 support */ ID(PCI_DEVICE_ID_INTEL_82443LX_0), ID(PCI_DEVICE_ID_INTEL_82443BX_0), ID(PCI_DEVICE_ID_INTEL_82443GX_0), diff --git a/drivers/char/agp/intel-gtt.c b/drivers/char/agp/intel-gtt.c index 43c4ec3..7f025fb 100644 --- a/drivers/char/agp/intel-gtt.c +++ b/drivers/char/agp/intel-gtt.c @@ -76,7 +76,6 @@ static struct _intel_private { struct resource ifp_resource; int resource_valid; struct page *scratch_page; - dma_addr_t scratch_page_dma; } intel_private; #define INTEL_GTT_GEN intel_private.driver->gen @@ -306,9 +305,9 @@ static int intel_gtt_setup_scratch_page(void) if (pci_dma_mapping_error(intel_private.pcidev, dma_addr)) return -EINVAL; - intel_private.scratch_page_dma = dma_addr; + intel_private.base.scratch_page_dma = dma_addr; } else - intel_private.scratch_page_dma = page_to_phys(page); + intel_private.base.scratch_page_dma = page_to_phys(page); intel_private.scratch_page = page; @@ -631,7 +630,7 @@ static unsigned int intel_gtt_mappable_entries(void) static void intel_gtt_teardown_scratch_page(void) { set_pages_wb(intel_private.scratch_page, 1); - pci_unmap_page(intel_private.pcidev, intel_private.scratch_page_dma, + pci_unmap_page(intel_private.pcidev, intel_private.base.scratch_page_dma, PAGE_SIZE, PCI_DMA_BIDIRECTIONAL); put_page(intel_private.scratch_page); __free_page(intel_private.scratch_page); @@ -681,6 +680,7 @@ static int intel_gtt_init(void) iounmap(intel_private.registers); return -ENOMEM; } + intel_private.base.gtt = intel_private.gtt; global_cache_flush(); /* FIXME: ? */ @@ -975,7 +975,7 @@ void intel_gtt_clear_range(unsigned int first_entry, unsigned int num_entries) unsigned int i; for (i = first_entry; i < (first_entry + num_entries); i++) { - intel_private.driver->write_entry(intel_private.scratch_page_dma, + intel_private.driver->write_entry(intel_private.base.scratch_page_dma, i, 0); } readl(intel_private.gtt+i-1); @@ -1190,7 +1190,6 @@ static inline int needs_idle_maps(void) { #ifdef CONFIG_INTEL_IOMMU const unsigned short gpu_devid = intel_private.pcidev->device; - extern int intel_iommu_gfx_mapped; /* Query intel_iommu to see if we need the workaround. Presumably that * was loaded first. diff --git a/drivers/char/agp/sis-agp.c b/drivers/char/agp/sis-agp.c index 29aacd8..08704ae 100644 --- a/drivers/char/agp/sis-agp.c +++ b/drivers/char/agp/sis-agp.c @@ -17,7 +17,7 @@ #define PCI_DEVICE_ID_SI_662 0x0662 #define PCI_DEVICE_ID_SI_671 0x0671 -static int __devinitdata agp_sis_force_delay = 0; +static bool __devinitdata agp_sis_force_delay = 0; static int __devinitdata agp_sis_agp_spec = -1; static int sis_fetch_size(void) diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig index 1368826..e354bc0 100644 --- a/drivers/gpu/drm/Kconfig +++ b/drivers/gpu/drm/Kconfig @@ -9,6 +9,7 @@ menuconfig DRM depends on (AGP || AGP=n) && !EMULATED_CMPXCHG && MMU select I2C select I2C_ALGOBIT + select DMA_SHARED_BUFFER help Kernel-level support for the Direct Rendering Infrastructure (DRI) introduced in XFree86 4.0. If you say Y here, you need to select @@ -18,6 +19,11 @@ menuconfig DRM details. You should also select and configure AGP (/dev/agpgart) support if it is available for your platform. +config DRM_USB + tristate + depends on DRM + select USB + config DRM_KMS_HELPER tristate depends on DRM @@ -27,6 +33,18 @@ config DRM_KMS_HELPER help FB and CRTC helpers for KMS drivers. +config DRM_LOAD_EDID_FIRMWARE + bool "Allow to specify an EDID data set instead of probing for it" + depends on DRM_KMS_HELPER + help + Say Y here, if you want to use EDID data to be loaded from the + /lib/firmware directory or one of the provided built-in + data sets. This may be necessary, if the graphics adapter or + monitor are unable to provide appropriate EDID data. Since this + feature is provided as a workaround for broken hardware, the + default case is N. Details and instructions how to build your own + EDID data are given in Documentation/EDID/HOWTO.txt. + config DRM_TTM tristate depends on DRM @@ -71,6 +89,8 @@ config DRM_RADEON source "drivers/gpu/drm/radeon/Kconfig" +source "drivers/gpu/drm/nouveau/Kconfig" + config DRM_I810 tristate "Intel I810" # !PREEMPT because of missing ioctl locking @@ -162,3 +182,7 @@ config DRM_SAVAGE source "drivers/gpu/drm/exynos/Kconfig" source "drivers/gpu/drm/vmwgfx/Kconfig" + +source "drivers/gpu/drm/gma500/Kconfig" + +source "drivers/gpu/drm/udl/Kconfig" diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile index c0496f6..c20da5b 100644 --- a/drivers/gpu/drm/Makefile +++ b/drivers/gpu/drm/Makefile @@ -9,20 +9,24 @@ drm-y := drm_auth.o drm_buffer.o drm_bufs.o drm_cache.o \ drm_drv.o drm_fops.o drm_gem.o drm_ioctl.o drm_irq.o \ drm_lock.o drm_memory.o drm_proc.o drm_stub.o drm_vm.o \ drm_agpsupport.o drm_scatter.o ati_pcigart.o drm_pci.o \ - drm_platform.o drm_sysfs.o drm_hashtab.o drm_sman.o drm_mm.o \ + drm_platform.o drm_sysfs.o drm_hashtab.o drm_mm.o \ drm_crtc.o drm_modes.o drm_edid.o \ drm_info.o drm_debugfs.o drm_encoder_slave.o \ - drm_trace_points.o drm_global.o drm_usb.o + drm_trace_points.o drm_global.o drm_prime.o drm-$(CONFIG_COMPAT) += drm_ioc32.o +drm-usb-y := drm_usb.o + drm_kms_helper-y := drm_fb_helper.o drm_crtc_helper.o drm_dp_i2c_helper.o +drm_kms_helper-$(CONFIG_DRM_LOAD_EDID_FIRMWARE) += drm_edid_load.o obj-$(CONFIG_DRM_KMS_HELPER) += drm_kms_helper.o CFLAGS_drm_trace_points.o := -I$(src) obj-$(CONFIG_DRM) += drm.o +obj-$(CONFIG_DRM_USB) += drm_usb.o obj-$(CONFIG_DRM_TTM) += ttm/ obj-$(CONFIG_DRM_TDFX) += tdfx/ obj-$(CONFIG_DRM_R128) += r128/ @@ -36,4 +40,6 @@ obj-$(CONFIG_DRM_VMWGFX)+= vmwgfx/ obj-$(CONFIG_DRM_VIA) +=via/ obj-$(CONFIG_DRM_NOUVEAU) +=nouveau/ obj-$(CONFIG_DRM_EXYNOS) +=exynos/ +obj-$(CONFIG_DRM_GMA500) += gma500/ +obj-$(CONFIG_DRM_UDL) += udl/ obj-y += i2c/ diff --git a/drivers/gpu/drm/drm_bufs.c b/drivers/gpu/drm/drm_bufs.c index 30372f7..348b367 100644 --- a/drivers/gpu/drm/drm_bufs.c +++ b/drivers/gpu/drm/drm_bufs.c @@ -1510,8 +1510,8 @@ int drm_freebufs(struct drm_device *dev, void *data, * \param arg pointer to a drm_buf_map structure. * \return zero on success or a negative number on failure. * - * Maps the AGP, SG or PCI buffer region with do_mmap(), and copies information - * about each buffer into user space. For PCI buffers, it calls do_mmap() with + * Maps the AGP, SG or PCI buffer region with vm_mmap(), and copies information + * about each buffer into user space. For PCI buffers, it calls vm_mmap() with * offset equal to 0, which drm_mmap() interpretes as PCI buffers and calls * drm_mmap_dma(). */ @@ -1553,18 +1553,14 @@ int drm_mapbufs(struct drm_device *dev, void *data, retcode = -EINVAL; goto done; } - down_write(¤t->mm->mmap_sem); - virtual = do_mmap(file_priv->filp, 0, map->size, + virtual = vm_mmap(file_priv->filp, 0, map->size, PROT_READ | PROT_WRITE, MAP_SHARED, token); - up_write(¤t->mm->mmap_sem); } else { - down_write(¤t->mm->mmap_sem); - virtual = do_mmap(file_priv->filp, 0, dma->byte_count, + virtual = vm_mmap(file_priv->filp, 0, dma->byte_count, PROT_READ | PROT_WRITE, MAP_SHARED, 0); - up_write(¤t->mm->mmap_sem); } if (virtual > -1024UL) { /* Real error */ diff --git a/drivers/gpu/drm/drm_cache.c b/drivers/gpu/drm/drm_cache.c index 5928653..4b8653b 100644 --- a/drivers/gpu/drm/drm_cache.c +++ b/drivers/gpu/drm/drm_cache.c @@ -41,10 +41,10 @@ drm_clflush_page(struct page *page) if (unlikely(page == NULL)) return; - page_virtual = kmap_atomic(page, KM_USER0); + page_virtual = kmap_atomic(page); for (i = 0; i < PAGE_SIZE; i += boot_cpu_data.x86_clflush_size) clflush(page_virtual + i); - kunmap_atomic(page_virtual, KM_USER0); + kunmap_atomic(page_virtual); } static void drm_cache_flush_clflush(struct page *pages[], @@ -87,10 +87,10 @@ drm_clflush_pages(struct page *pages[], unsigned long num_pages) if (unlikely(page == NULL)) continue; - page_virtual = kmap_atomic(page, KM_USER0); + page_virtual = kmap_atomic(page); flush_dcache_range((unsigned long)page_virtual, (unsigned long)page_virtual + PAGE_SIZE); - kunmap_atomic(page_virtual, KM_USER0); + kunmap_atomic(page_virtual); } #else printk(KERN_ERR "Architecture has no drm_cache.c support\n"); diff --git a/drivers/gpu/drm/drm_context.c b/drivers/gpu/drm/drm_context.c index 6d440fb..325365f 100644 --- a/drivers/gpu/drm/drm_context.c +++ b/drivers/gpu/drm/drm_context.c @@ -154,8 +154,6 @@ int drm_getsareactx(struct drm_device *dev, void *data, return -EINVAL; } - mutex_unlock(&dev->struct_mutex); - request->handle = NULL; list_for_each_entry(_entry, &dev->maplist, head) { if (_entry->map == map) { @@ -164,6 +162,9 @@ int drm_getsareactx(struct drm_device *dev, void *data, break; } } + + mutex_unlock(&dev->struct_mutex); + if (request->handle == NULL) return -EINVAL; diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c index 3f1799b..4fd363f 100644 --- a/drivers/gpu/drm/drm_crtc.c +++ b/drivers/gpu/drm/drm_crtc.c @@ -36,11 +36,7 @@ #include "drmP.h" #include "drm_crtc.h" #include "drm_edid.h" - -struct drm_prop_enum_list { - int type; - char *name; -}; +#include "drm_fourcc.h" /* Avoid boilerplate. I'm tired of typing. */ #define DRM_ENUM_NAME_FN(fnname, list) \ @@ -297,9 +293,8 @@ int drm_framebuffer_init(struct drm_device *dev, struct drm_framebuffer *fb, int ret; ret = drm_mode_object_get(dev, &fb->base, DRM_MODE_OBJECT_FB); - if (ret) { + if (ret) return ret; - } fb->dev = dev; fb->funcs = funcs; @@ -324,6 +319,7 @@ void drm_framebuffer_cleanup(struct drm_framebuffer *fb) { struct drm_device *dev = fb->dev; struct drm_crtc *crtc; + struct drm_plane *plane; struct drm_mode_set set; int ret; @@ -340,6 +336,18 @@ void drm_framebuffer_cleanup(struct drm_framebuffer *fb) } } + list_for_each_entry(plane, &dev->mode_config.plane_list, head) { + if (plane->fb == fb) { + /* should turn off the crtc */ + ret = plane->funcs->disable_plane(plane); + if (ret) + DRM_ERROR("failed to disable plane with busy fb\n"); + /* disconnect the plane from the fb and crtc: */ + plane->fb = NULL; + plane->crtc = NULL; + } + } + drm_mode_object_put(dev, &fb->base); list_del(&fb->head); dev->mode_config.num_fb--; @@ -356,19 +364,31 @@ EXPORT_SYMBOL(drm_framebuffer_cleanup); * Caller must hold mode config lock. * * Inits a new object created as base part of an driver crtc object. + * + * RETURNS: + * Zero on success, error code on failure. */ -void drm_crtc_init(struct drm_device *dev, struct drm_crtc *crtc, +int drm_crtc_init(struct drm_device *dev, struct drm_crtc *crtc, const struct drm_crtc_funcs *funcs) { + int ret; + crtc->dev = dev; crtc->funcs = funcs; mutex_lock(&dev->mode_config.mutex); - drm_mode_object_get(dev, &crtc->base, DRM_MODE_OBJECT_CRTC); + + ret = drm_mode_object_get(dev, &crtc->base, DRM_MODE_OBJECT_CRTC); + if (ret) + goto out; list_add_tail(&crtc->head, &dev->mode_config.crtc_list); dev->mode_config.num_crtc++; + + out: mutex_unlock(&dev->mode_config.mutex); + + return ret; } EXPORT_SYMBOL(drm_crtc_init); @@ -428,7 +448,7 @@ void drm_mode_remove(struct drm_connector *connector, struct drm_display_mode *mode) { list_del(&mode->head); - kfree(mode); + drm_mode_destroy(connector->dev, mode); } EXPORT_SYMBOL(drm_mode_remove); @@ -440,21 +460,29 @@ EXPORT_SYMBOL(drm_mode_remove); * @name: user visible name of the connector * * LOCKING: - * Caller must hold @dev's mode_config lock. + * Takes mode config lock. * * Initialises a preallocated connector. Connectors should be * subclassed as part of driver connector objects. + * + * RETURNS: + * Zero on success, error code on failure. */ -void drm_connector_init(struct drm_device *dev, - struct drm_connector *connector, - const struct drm_connector_funcs *funcs, - int connector_type) +int drm_connector_init(struct drm_device *dev, + struct drm_connector *connector, + const struct drm_connector_funcs *funcs, + int connector_type) { + int ret; + mutex_lock(&dev->mode_config.mutex); + ret = drm_mode_object_get(dev, &connector->base, DRM_MODE_OBJECT_CONNECTOR); + if (ret) + goto out; + connector->dev = dev; connector->funcs = funcs; - drm_mode_object_get(dev, &connector->base, DRM_MODE_OBJECT_CONNECTOR); connector->connector_type = connector_type; connector->connector_type_id = ++drm_connector_enum_list[connector_type].count; /* TODO */ @@ -474,7 +502,10 @@ void drm_connector_init(struct drm_device *dev, drm_connector_attach_property(connector, dev->mode_config.dpms_property, 0); + out: mutex_unlock(&dev->mode_config.mutex); + + return ret; } EXPORT_SYMBOL(drm_connector_init); @@ -483,7 +514,7 @@ EXPORT_SYMBOL(drm_connector_init); * @connector: connector to cleanup * * LOCKING: - * Caller must hold @dev's mode_config lock. + * Takes mode config lock. * * Cleans up the connector but doesn't free the object. */ @@ -509,23 +540,41 @@ void drm_connector_cleanup(struct drm_connector *connector) } EXPORT_SYMBOL(drm_connector_cleanup); -void drm_encoder_init(struct drm_device *dev, +void drm_connector_unplug_all(struct drm_device *dev) +{ + struct drm_connector *connector; + + /* taking the mode config mutex ends up in a clash with sysfs */ + list_for_each_entry(connector, &dev->mode_config.connector_list, head) + drm_sysfs_connector_remove(connector); + +} +EXPORT_SYMBOL(drm_connector_unplug_all); + +int drm_encoder_init(struct drm_device *dev, struct drm_encoder *encoder, const struct drm_encoder_funcs *funcs, int encoder_type) { + int ret; + mutex_lock(&dev->mode_config.mutex); - encoder->dev = dev; + ret = drm_mode_object_get(dev, &encoder->base, DRM_MODE_OBJECT_ENCODER); + if (ret) + goto out; - drm_mode_object_get(dev, &encoder->base, DRM_MODE_OBJECT_ENCODER); + encoder->dev = dev; encoder->encoder_type = encoder_type; encoder->funcs = funcs; list_add_tail(&encoder->head, &dev->mode_config.encoder_list); dev->mode_config.num_encoder++; + out: mutex_unlock(&dev->mode_config.mutex); + + return ret; } EXPORT_SYMBOL(drm_encoder_init); @@ -540,6 +589,69 @@ void drm_encoder_cleanup(struct drm_encoder *encoder) } EXPORT_SYMBOL(drm_encoder_cleanup); +int drm_plane_init(struct drm_device *dev, struct drm_plane *plane, + unsigned long possible_crtcs, + const struct drm_plane_funcs *funcs, + const uint32_t *formats, uint32_t format_count, + bool priv) +{ + int ret; + + mutex_lock(&dev->mode_config.mutex); + + ret = drm_mode_object_get(dev, &plane->base, DRM_MODE_OBJECT_PLANE); + if (ret) + goto out; + + plane->dev = dev; + plane->funcs = funcs; + plane->format_types = kmalloc(sizeof(uint32_t) * format_count, + GFP_KERNEL); + if (!plane->format_types) { + DRM_DEBUG_KMS("out of memory when allocating plane\n"); + drm_mode_object_put(dev, &plane->base); + ret = -ENOMEM; + goto out; + } + + memcpy(plane->format_types, formats, format_count * sizeof(uint32_t)); + plane->format_count = format_count; + plane->possible_crtcs = possible_crtcs; + + /* private planes are not exposed to userspace, but depending on + * display hardware, might be convenient to allow sharing programming + * for the scanout engine with the crtc implementation. + */ + if (!priv) { + list_add_tail(&plane->head, &dev->mode_config.plane_list); + dev->mode_config.num_plane++; + } else { + INIT_LIST_HEAD(&plane->head); + } + + out: + mutex_unlock(&dev->mode_config.mutex); + + return ret; +} +EXPORT_SYMBOL(drm_plane_init); + +void drm_plane_cleanup(struct drm_plane *plane) +{ + struct drm_device *dev = plane->dev; + + mutex_lock(&dev->mode_config.mutex); + kfree(plane->format_types); + drm_mode_object_put(dev, &plane->base); + /* if not added to a list, it must be a private plane */ + if (!list_empty(&plane->head)) { + list_del(&plane->head); + dev->mode_config.num_plane--; + } + mutex_unlock(&dev->mode_config.mutex); +} +EXPORT_SYMBOL(drm_plane_cleanup); + /** * drm_mode_create - create a new display mode * @dev: DRM device @@ -560,7 +672,11 @@ struct drm_display_mode *drm_mode_create(struct drm_device *dev) if (!nmode) return NULL; - drm_mode_object_get(dev, &nmode->base, DRM_MODE_OBJECT_MODE); + if (drm_mode_object_get(dev, &nmode->base, DRM_MODE_OBJECT_MODE)) { + kfree(nmode); + return NULL; + } + return nmode; } EXPORT_SYMBOL(drm_mode_create); @@ -577,6 +693,9 @@ EXPORT_SYMBOL(drm_mode_create); */ void drm_mode_destroy(struct drm_device *dev, struct drm_display_mode *mode) { + if (!mode) + return; + drm_mode_object_put(dev, &mode->base); kfree(mode); @@ -587,7 +706,6 @@ static int drm_mode_create_standard_connector_properties(struct drm_device *dev) { struct drm_property *edid; struct drm_property *dpms; - int i; /* * Standard properties (apply to all connectors) @@ -597,11 +715,9 @@ static int drm_mode_create_standard_connector_properties(struct drm_device *dev) "EDID", 0); dev->mode_config.edid_property = edid; - dpms = drm_property_create(dev, DRM_MODE_PROP_ENUM, - "DPMS", ARRAY_SIZE(drm_dpms_enum_list)); - for (i = 0; i < ARRAY_SIZE(drm_dpms_enum_list); i++) - drm_property_add_enum(dpms, i, drm_dpms_enum_list[i].type, - drm_dpms_enum_list[i].name); + dpms = drm_property_create_enum(dev, 0, + "DPMS", drm_dpms_enum_list, + ARRAY_SIZE(drm_dpms_enum_list)); dev->mode_config.dpms_property = dpms; return 0; @@ -617,30 +733,21 @@ int drm_mode_create_dvi_i_properties(struct drm_device *dev) { struct drm_property *dvi_i_selector; struct drm_property *dvi_i_subconnector; - int i; if (dev->mode_config.dvi_i_select_subconnector_property) return 0; dvi_i_selector = - drm_property_create(dev, DRM_MODE_PROP_ENUM, + drm_property_create_enum(dev, 0, "select subconnector", + drm_dvi_i_select_enum_list, ARRAY_SIZE(drm_dvi_i_select_enum_list)); - for (i = 0; i < ARRAY_SIZE(drm_dvi_i_select_enum_list); i++) - drm_property_add_enum(dvi_i_selector, i, - drm_dvi_i_select_enum_list[i].type, - drm_dvi_i_select_enum_list[i].name); dev->mode_config.dvi_i_select_subconnector_property = dvi_i_selector; - dvi_i_subconnector = - drm_property_create(dev, DRM_MODE_PROP_ENUM | - DRM_MODE_PROP_IMMUTABLE, + dvi_i_subconnector = drm_property_create_enum(dev, DRM_MODE_PROP_IMMUTABLE, "subconnector", + drm_dvi_i_subconnector_enum_list, ARRAY_SIZE(drm_dvi_i_subconnector_enum_list)); - for (i = 0; i < ARRAY_SIZE(drm_dvi_i_subconnector_enum_list); i++) - drm_property_add_enum(dvi_i_subconnector, i, - drm_dvi_i_subconnector_enum_list[i].type, - drm_dvi_i_subconnector_enum_list[i].name); dev->mode_config.dvi_i_subconnector_property = dvi_i_subconnector; return 0; @@ -671,51 +778,33 @@ int drm_mode_create_tv_properties(struct drm_device *dev, int num_modes, /* * Basic connector properties */ - tv_selector = drm_property_create(dev, DRM_MODE_PROP_ENUM, + tv_selector = drm_property_create_enum(dev, 0, "select subconnector", + drm_tv_select_enum_list, ARRAY_SIZE(drm_tv_select_enum_list)); - for (i = 0; i < ARRAY_SIZE(drm_tv_select_enum_list); i++) - drm_property_add_enum(tv_selector, i, - drm_tv_select_enum_list[i].type, - drm_tv_select_enum_list[i].name); dev->mode_config.tv_select_subconnector_property = tv_selector; tv_subconnector = - drm_property_create(dev, DRM_MODE_PROP_ENUM | - DRM_MODE_PROP_IMMUTABLE, "subconnector", + drm_property_create_enum(dev, DRM_MODE_PROP_IMMUTABLE, + "subconnector", + drm_tv_subconnector_enum_list, ARRAY_SIZE(drm_tv_subconnector_enum_list)); - for (i = 0; i < ARRAY_SIZE(drm_tv_subconnector_enum_list); i++) - drm_property_add_enum(tv_subconnector, i, - drm_tv_subconnector_enum_list[i].type, - drm_tv_subconnector_enum_list[i].name); dev->mode_config.tv_subconnector_property = tv_subconnector; /* * Other, TV specific properties: margins & TV modes. */ dev->mode_config.tv_left_margin_property = - drm_property_create(dev, DRM_MODE_PROP_RANGE, - "left margin", 2); - dev->mode_config.tv_left_margin_property->values[0] = 0; - dev->mode_config.tv_left_margin_property->values[1] = 100; + drm_property_create_range(dev, 0, "left margin", 0, 100); dev->mode_config.tv_right_margin_property = - drm_property_create(dev, DRM_MODE_PROP_RANGE, - "right margin", 2); - dev->mode_config.tv_right_margin_property->values[0] = 0; - dev->mode_config.tv_right_margin_property->values[1] = 100; + drm_property_create_range(dev, 0, "right margin", 0, 100); dev->mode_config.tv_top_margin_property = - drm_property_create(dev, DRM_MODE_PROP_RANGE, - "top margin", 2); - dev->mode_config.tv_top_margin_property->values[0] = 0; - dev->mode_config.tv_top_margin_property->values[1] = 100; + drm_property_create_range(dev, 0, "top margin", 0, 100); dev->mode_config.tv_bottom_margin_property = - drm_property_create(dev, DRM_MODE_PROP_RANGE, - "bottom margin", 2); - dev->mode_config.tv_bottom_margin_property->values[0] = 0; - dev->mode_config.tv_bottom_margin_property->values[1] = 100; + drm_property_create_range(dev, 0, "bottom margin", 0, 100); dev->mode_config.tv_mode_property = drm_property_create(dev, DRM_MODE_PROP_ENUM, @@ -725,40 +814,22 @@ int drm_mode_create_tv_properties(struct drm_device *dev, int num_modes, i, modes[i]); dev->mode_config.tv_brightness_property = - drm_property_create(dev, DRM_MODE_PROP_RANGE, - "brightness", 2); - dev->mode_config.tv_brightness_property->values[0] = 0; - dev->mode_config.tv_brightness_property->values[1] = 100; + drm_property_create_range(dev, 0, "brightness", 0, 100); dev->mode_config.tv_contrast_property = - drm_property_create(dev, DRM_MODE_PROP_RANGE, - "contrast", 2); - dev->mode_config.tv_contrast_property->values[0] = 0; - dev->mode_config.tv_contrast_property->values[1] = 100; + drm_property_create_range(dev, 0, "contrast", 0, 100); dev->mode_config.tv_flicker_reduction_property = - drm_property_create(dev, DRM_MODE_PROP_RANGE, - "flicker reduction", 2); - dev->mode_config.tv_flicker_reduction_property->values[0] = 0; - dev->mode_config.tv_flicker_reduction_property->values[1] = 100; + drm_property_create_range(dev, 0, "flicker reduction", 0, 100); dev->mode_config.tv_overscan_property = - drm_property_create(dev, DRM_MODE_PROP_RANGE, - "overscan", 2); - dev->mode_config.tv_overscan_property->values[0] = 0; - dev->mode_config.tv_overscan_property->values[1] = 100; + drm_property_create_range(dev, 0, "overscan", 0, 100); dev->mode_config.tv_saturation_property = - drm_property_create(dev, DRM_MODE_PROP_RANGE, - "saturation", 2); - dev->mode_config.tv_saturation_property->values[0] = 0; - dev->mode_config.tv_saturation_property->values[1] = 100; + drm_property_create_range(dev, 0, "saturation", 0, 100); dev->mode_config.tv_hue_property = - drm_property_create(dev, DRM_MODE_PROP_RANGE, - "hue", 2); - dev->mode_config.tv_hue_property->values[0] = 0; - dev->mode_config.tv_hue_property->values[1] = 100; + drm_property_create_range(dev, 0, "hue", 0, 100); return 0; } @@ -774,18 +845,14 @@ EXPORT_SYMBOL(drm_mode_create_tv_properties); int drm_mode_create_scaling_mode_property(struct drm_device *dev) { struct drm_property *scaling_mode; - int i; if (dev->mode_config.scaling_mode_property) return 0; scaling_mode = - drm_property_create(dev, DRM_MODE_PROP_ENUM, "scaling mode", + drm_property_create_enum(dev, 0, "scaling mode", + drm_scaling_mode_enum_list, ARRAY_SIZE(drm_scaling_mode_enum_list)); - for (i = 0; i < ARRAY_SIZE(drm_scaling_mode_enum_list); i++) - drm_property_add_enum(scaling_mode, i, - drm_scaling_mode_enum_list[i].type, - drm_scaling_mode_enum_list[i].name); dev->mode_config.scaling_mode_property = scaling_mode; @@ -803,18 +870,14 @@ EXPORT_SYMBOL(drm_mode_create_scaling_mode_property); int drm_mode_create_dithering_property(struct drm_device *dev) { struct drm_property *dithering_mode; - int i; if (dev->mode_config.dithering_mode_property) return 0; dithering_mode = - drm_property_create(dev, DRM_MODE_PROP_ENUM, "dithering", + drm_property_create_enum(dev, 0, "dithering", + drm_dithering_mode_enum_list, ARRAY_SIZE(drm_dithering_mode_enum_list)); - for (i = 0; i < ARRAY_SIZE(drm_dithering_mode_enum_list); i++) - drm_property_add_enum(dithering_mode, i, - drm_dithering_mode_enum_list[i].type, - drm_dithering_mode_enum_list[i].name); dev->mode_config.dithering_mode_property = dithering_mode; return 0; @@ -831,20 +894,15 @@ EXPORT_SYMBOL(drm_mode_create_dithering_property); int drm_mode_create_dirty_info_property(struct drm_device *dev) { struct drm_property *dirty_info; - int i; if (dev->mode_config.dirty_info_property) return 0; dirty_info = - drm_property_create(dev, DRM_MODE_PROP_ENUM | - DRM_MODE_PROP_IMMUTABLE, + drm_property_create_enum(dev, DRM_MODE_PROP_IMMUTABLE, "dirty", + drm_dirty_info_enum_list, ARRAY_SIZE(drm_dirty_info_enum_list)); - for (i = 0; i < ARRAY_SIZE(drm_dirty_info_enum_list); i++) - drm_property_add_enum(dirty_info, i, - drm_dirty_info_enum_list[i].type, - drm_dirty_info_enum_list[i].name); dev->mode_config.dirty_info_property = dirty_info; return 0; @@ -871,6 +929,7 @@ void drm_mode_config_init(struct drm_device *dev) INIT_LIST_HEAD(&dev->mode_config.encoder_list); INIT_LIST_HEAD(&dev->mode_config.property_list); INIT_LIST_HEAD(&dev->mode_config.property_blob_list); + INIT_LIST_HEAD(&dev->mode_config.plane_list); idr_init(&dev->mode_config.crtc_idr); mutex_lock(&dev->mode_config.mutex); @@ -927,6 +986,7 @@ int drm_mode_group_init_legacy_group(struct drm_device *dev, return 0; } +EXPORT_SYMBOL(drm_mode_group_init_legacy_group); /** * drm_mode_config_cleanup - free up DRM mode_config info @@ -947,6 +1007,7 @@ void drm_mode_config_cleanup(struct drm_device *dev) struct drm_encoder *encoder, *enct; struct drm_framebuffer *fb, *fbt; struct drm_property *property, *pt; + struct drm_plane *plane, *plt; list_for_each_entry_safe(encoder, enct, &dev->mode_config.encoder_list, head) { @@ -967,10 +1028,17 @@ void drm_mode_config_cleanup(struct drm_device *dev) fb->funcs->destroy(fb); } + list_for_each_entry_safe(plane, plt, &dev->mode_config.plane_list, + head) { + plane->funcs->destroy(plane); + } + list_for_each_entry_safe(crtc, ct, &dev->mode_config.crtc_list, head) { crtc->funcs->destroy(crtc); } + idr_remove_all(&dev->mode_config.crtc_idr); + idr_destroy(&dev->mode_config.crtc_idr); } EXPORT_SYMBOL(drm_mode_config_cleanup); @@ -985,9 +1053,16 @@ EXPORT_SYMBOL(drm_mode_config_cleanup); * Convert a drm_display_mode into a drm_mode_modeinfo structure to return to * the user. */ -void drm_crtc_convert_to_umode(struct drm_mode_modeinfo *out, - struct drm_display_mode *in) +static void drm_crtc_convert_to_umode(struct drm_mode_modeinfo *out, + const struct drm_display_mode *in) { + WARN(in->hdisplay > USHRT_MAX || in->hsync_start > USHRT_MAX || + in->hsync_end > USHRT_MAX || in->htotal > USHRT_MAX || + in->hskew > USHRT_MAX || in->vdisplay > USHRT_MAX || + in->vsync_start > USHRT_MAX || in->vsync_end > USHRT_MAX || + in->vtotal > USHRT_MAX || in->vscan > USHRT_MAX, + "timing values too large for mode info\n"); + out->clock = in->clock; out->hdisplay = in->hdisplay; out->hsync_start = in->hsync_start; @@ -1016,10 +1091,16 @@ void drm_crtc_convert_to_umode(struct drm_mode_modeinfo *out, * * Convert a drm_mode_modeinfo into a drm_display_mode structure to return to * the caller. + * + * RETURNS: + * Zero on success, errno on failure. */ -void drm_crtc_convert_umode(struct drm_display_mode *out, - struct drm_mode_modeinfo *in) +static int drm_crtc_convert_umode(struct drm_display_mode *out, + const struct drm_mode_modeinfo *in) { + if (in->clock > INT_MAX || in->vrefresh > INT_MAX) + return -ERANGE; + out->clock = in->clock; out->hdisplay = in->hdisplay; out->hsync_start = in->hsync_start; @@ -1036,6 +1117,8 @@ void drm_crtc_convert_umode(struct drm_display_mode *out, out->type = in->type; strncpy(out->name, in->name, DRM_DISPLAY_MODE_LEN); out->name[DRM_DISPLAY_MODE_LEN-1] = 0; + + return 0; } /** @@ -1234,7 +1317,7 @@ out: * @arg: arg from ioctl * * LOCKING: - * Caller? (FIXME) + * Takes mode config lock. * * Construct a CRTC configuration structure to return to the user. * @@ -1294,7 +1377,7 @@ out: * @arg: arg from ioctl * * LOCKING: - * Caller? (FIXME) + * Takes mode config lock. * * Construct a connector configuration structure to return to the user. * @@ -1379,7 +1462,7 @@ int drm_mode_getconnector(struct drm_device *dev, void *data, */ if ((out_resp->count_modes >= mode_count) && mode_count) { copied = 0; - mode_ptr = (struct drm_mode_modeinfo *)(unsigned long)out_resp->modes_ptr; + mode_ptr = (struct drm_mode_modeinfo __user *)(unsigned long)out_resp->modes_ptr; list_for_each_entry(mode, &connector->modes, head) { drm_crtc_convert_to_umode(&u_mode, mode); if (copy_to_user(mode_ptr + copied, @@ -1394,8 +1477,8 @@ int drm_mode_getconnector(struct drm_device *dev, void *data, if ((out_resp->count_props >= props_count) && props_count) { copied = 0; - prop_ptr = (uint32_t *)(unsigned long)(out_resp->props_ptr); - prop_values = (uint64_t *)(unsigned long)(out_resp->prop_values_ptr); + prop_ptr = (uint32_t __user *)(unsigned long)(out_resp->props_ptr); + prop_values = (uint64_t __user *)(unsigned long)(out_resp->prop_values_ptr); for (i = 0; i < DRM_CONNECTOR_MAX_PROPERTY; i++) { if (connector->property_ids[i] != 0) { if (put_user(connector->property_ids[i], @@ -1417,7 +1500,7 @@ int drm_mode_getconnector(struct drm_device *dev, void *data, if ((out_resp->count_encoders >= encoders_count) && encoders_count) { copied = 0; - encoder_ptr = (uint32_t *)(unsigned long)(out_resp->encoders_ptr); + encoder_ptr = (uint32_t __user *)(unsigned long)(out_resp->encoders_ptr); for (i = 0; i < DRM_CONNECTOR_MAX_ENCODER; i++) { if (connector->encoder_ids[i] != 0) { if (put_user(connector->encoder_ids[i], @@ -1471,6 +1554,254 @@ out: } /** + * drm_mode_getplane_res - get plane info + * @dev: DRM device + * @data: ioctl data + * @file_priv: DRM file info + * + * LOCKING: + * Takes mode config lock. + * + * Return an plane count and set of IDs. + */ +int drm_mode_getplane_res(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + struct drm_mode_get_plane_res *plane_resp = data; + struct drm_mode_config *config; + struct drm_plane *plane; + uint32_t __user *plane_ptr; + int copied = 0, ret = 0; + + if (!drm_core_check_feature(dev, DRIVER_MODESET)) + return -EINVAL; + + mutex_lock(&dev->mode_config.mutex); + config = &dev->mode_config; + + /* + * This ioctl is called twice, once to determine how much space is + * needed, and the 2nd time to fill it. + */ + if (config->num_plane && + (plane_resp->count_planes >= config->num_plane)) { + plane_ptr = (uint32_t __user *)(unsigned long)plane_resp->plane_id_ptr; + + list_for_each_entry(plane, &config->plane_list, head) { + if (put_user(plane->base.id, plane_ptr + copied)) { + ret = -EFAULT; + goto out; + } + copied++; + } + } + plane_resp->count_planes = config->num_plane; + +out: + mutex_unlock(&dev->mode_config.mutex); + return ret; +} + +/** + * drm_mode_getplane - get plane info + * @dev: DRM device + * @data: ioctl data + * @file_priv: DRM file info + * + * LOCKING: + * Takes mode config lock. + * + * Return plane info, including formats supported, gamma size, any + * current fb, etc. + */ +int drm_mode_getplane(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + struct drm_mode_get_plane *plane_resp = data; + struct drm_mode_object *obj; + struct drm_plane *plane; + uint32_t __user *format_ptr; + int ret = 0; + + if (!drm_core_check_feature(dev, DRIVER_MODESET)) + return -EINVAL; + + mutex_lock(&dev->mode_config.mutex); + obj = drm_mode_object_find(dev, plane_resp->plane_id, + DRM_MODE_OBJECT_PLANE); + if (!obj) { + ret = -ENOENT; + goto out; + } + plane = obj_to_plane(obj); + + if (plane->crtc) + plane_resp->crtc_id = plane->crtc->base.id; + else + plane_resp->crtc_id = 0; + + if (plane->fb) + plane_resp->fb_id = plane->fb->base.id; + else + plane_resp->fb_id = 0; + + plane_resp->plane_id = plane->base.id; + plane_resp->possible_crtcs = plane->possible_crtcs; + plane_resp->gamma_size = plane->gamma_size; + + /* + * This ioctl is called twice, once to determine how much space is + * needed, and the 2nd time to fill it. + */ + if (plane->format_count && + (plane_resp->count_format_types >= plane->format_count)) { + format_ptr = (uint32_t __user *)(unsigned long)plane_resp->format_type_ptr; + if (copy_to_user(format_ptr, + plane->format_types, + sizeof(uint32_t) * plane->format_count)) { + ret = -EFAULT; + goto out; + } + } + plane_resp->count_format_types = plane->format_count; + +out: + mutex_unlock(&dev->mode_config.mutex); + return ret; +} + +/** + * drm_mode_setplane - set up or tear down an plane + * @dev: DRM device + * @data: ioctl data* + * @file_prive: DRM file info + * + * LOCKING: + * Takes mode config lock. + * + * Set plane info, including placement, fb, scaling, and other factors. + * Or pass a NULL fb to disable. + */ +int drm_mode_setplane(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + struct drm_mode_set_plane *plane_req = data; + struct drm_mode_object *obj; + struct drm_plane *plane; + struct drm_crtc *crtc; + struct drm_framebuffer *fb; + int ret = 0; + unsigned int fb_width, fb_height; + int i; + + if (!drm_core_check_feature(dev, DRIVER_MODESET)) + return -EINVAL; + + mutex_lock(&dev->mode_config.mutex); + + /* + * First, find the plane, crtc, and fb objects. If not available, + * we don't bother to call the driver. + */ + obj = drm_mode_object_find(dev, plane_req->plane_id, + DRM_MODE_OBJECT_PLANE); + if (!obj) { + DRM_DEBUG_KMS("Unknown plane ID %d\n", + plane_req->plane_id); + ret = -ENOENT; + goto out; + } + plane = obj_to_plane(obj); + + /* No fb means shut it down */ + if (!plane_req->fb_id) { + plane->funcs->disable_plane(plane); + plane->crtc = NULL; + plane->fb = NULL; + goto out; + } + + obj = drm_mode_object_find(dev, plane_req->crtc_id, + DRM_MODE_OBJECT_CRTC); + if (!obj) { + DRM_DEBUG_KMS("Unknown crtc ID %d\n", + plane_req->crtc_id); + ret = -ENOENT; + goto out; + } + crtc = obj_to_crtc(obj); + + obj = drm_mode_object_find(dev, plane_req->fb_id, + DRM_MODE_OBJECT_FB); + if (!obj) { + DRM_DEBUG_KMS("Unknown framebuffer ID %d\n", + plane_req->fb_id); + ret = -ENOENT; + goto out; + } + fb = obj_to_fb(obj); + + /* Check whether this plane supports the fb pixel format. */ + for (i = 0; i < plane->format_count; i++) + if (fb->pixel_format == plane->format_types[i]) + break; + if (i == plane->format_count) { + DRM_DEBUG_KMS("Invalid pixel format 0x%08x\n", fb->pixel_format); + ret = -EINVAL; + goto out; + } + + fb_width = fb->width << 16; + fb_height = fb->height << 16; + + /* Make sure source coordinates are inside the fb. */ + if (plane_req->src_w > fb_width || + plane_req->src_x > fb_width - plane_req->src_w || + plane_req->src_h > fb_height || + plane_req->src_y > fb_height - plane_req->src_h) { + DRM_DEBUG_KMS("Invalid source coordinates " + "%u.%06ux%u.%06u+%u.%06u+%u.%06u\n", + plane_req->src_w >> 16, + ((plane_req->src_w & 0xffff) * 15625) >> 10, + plane_req->src_h >> 16, + ((plane_req->src_h & 0xffff) * 15625) >> 10, + plane_req->src_x >> 16, + ((plane_req->src_x & 0xffff) * 15625) >> 10, + plane_req->src_y >> 16, + ((plane_req->src_y & 0xffff) * 15625) >> 10); + ret = -ENOSPC; + goto out; + } + + /* Give drivers some help against integer overflows */ + if (plane_req->crtc_w > INT_MAX || + plane_req->crtc_x > INT_MAX - (int32_t) plane_req->crtc_w || + plane_req->crtc_h > INT_MAX || + plane_req->crtc_y > INT_MAX - (int32_t) plane_req->crtc_h) { + DRM_DEBUG_KMS("Invalid CRTC coordinates %ux%u+%d+%d\n", + plane_req->crtc_w, plane_req->crtc_h, + plane_req->crtc_x, plane_req->crtc_y); + ret = -ERANGE; + goto out; + } + + ret = plane->funcs->update_plane(plane, crtc, fb, + plane_req->crtc_x, plane_req->crtc_y, + plane_req->crtc_w, plane_req->crtc_h, + plane_req->src_x, plane_req->src_y, + plane_req->src_w, plane_req->src_h); + if (!ret) { + plane->crtc = crtc; + plane->fb = fb; + } + +out: + mutex_unlock(&dev->mode_config.mutex); + + return ret; +} + +/** * drm_mode_setcrtc - set CRTC configuration * @inode: inode from the ioctl * @filp: file * from the ioctl @@ -1478,7 +1809,7 @@ out: * @arg: arg from ioctl * * LOCKING: - * Caller? (FIXME) + * Takes mode config lock. * * Build a new CRTC configuration based on user request. * @@ -1493,7 +1824,7 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data, struct drm_mode_config *config = &dev->mode_config; struct drm_mode_crtc *crtc_req = data; struct drm_mode_object *obj; - struct drm_crtc *crtc, *crtcfb; + struct drm_crtc *crtc; struct drm_connector **connector_set = NULL, *connector; struct drm_framebuffer *fb = NULL; struct drm_display_mode *mode = NULL; @@ -1505,6 +1836,10 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data, if (!drm_core_check_feature(dev, DRIVER_MODESET)) return -EINVAL; + /* For some reason crtc x/y offsets are signed internally. */ + if (crtc_req->x > INT_MAX || crtc_req->y > INT_MAX) + return -ERANGE; + mutex_lock(&dev->mode_config.mutex); obj = drm_mode_object_find(dev, crtc_req->crtc_id, DRM_MODE_OBJECT_CRTC); @@ -1520,14 +1855,12 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data, /* If we have a mode we need a framebuffer. */ /* If we pass -1, set the mode with the currently bound fb */ if (crtc_req->fb_id == -1) { - list_for_each_entry(crtcfb, - &dev->mode_config.crtc_list, head) { - if (crtcfb == crtc) { - DRM_DEBUG_KMS("Using current fb for " - "setmode\n"); - fb = crtc->fb; - } + if (!crtc->fb) { + DRM_DEBUG_KMS("CRTC doesn't have current FB\n"); + ret = -EINVAL; + goto out; } + fb = crtc->fb; } else { obj = drm_mode_object_find(dev, crtc_req->fb_id, DRM_MODE_OBJECT_FB); @@ -1541,8 +1874,30 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data, } mode = drm_mode_create(dev); - drm_crtc_convert_umode(mode, &crtc_req->mode); + if (!mode) { + ret = -ENOMEM; + goto out; + } + + ret = drm_crtc_convert_umode(mode, &crtc_req->mode); + if (ret) { + DRM_DEBUG_KMS("Invalid mode\n"); + goto out; + } + drm_mode_set_crtcinfo(mode, CRTC_INTERLACE_HALVE_V); + + if (mode->hdisplay > fb->width || + mode->vdisplay > fb->height || + crtc_req->x > fb->width - mode->hdisplay || + crtc_req->y > fb->height - mode->vdisplay) { + DRM_DEBUG_KMS("Invalid CRTC viewport %ux%u+%u+%u for fb size %ux%u.\n", + mode->hdisplay, mode->vdisplay, + crtc_req->x, crtc_req->y, + fb->width, fb->height); + ret = -ENOSPC; + goto out; + } } if (crtc_req->count_connectors == 0 && mode) { @@ -1576,7 +1931,7 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data, } for (i = 0; i < crtc_req->count_connectors; i++) { - set_connectors_ptr = (uint32_t *)(unsigned long)crtc_req->set_connectors_ptr; + set_connectors_ptr = (uint32_t __user *)(unsigned long)crtc_req->set_connectors_ptr; if (get_user(out_id, &set_connectors_ptr[i])) { ret = -EFAULT; goto out; @@ -1610,6 +1965,7 @@ int drm_mode_setcrtc(struct drm_device *dev, void *data, out: kfree(connector_set); + drm_mode_destroy(dev, mode); mutex_unlock(&dev->mode_config.mutex); return ret; } @@ -1660,6 +2016,42 @@ out: return ret; } +/* Original addfb only supported RGB formats, so figure out which one */ +uint32_t drm_mode_legacy_fb_format(uint32_t bpp, uint32_t depth) +{ + uint32_t fmt; + + switch (bpp) { + case 8: + fmt = DRM_FORMAT_RGB332; + break; + case 16: + if (depth == 15) + fmt = DRM_FORMAT_XRGB1555; + else + fmt = DRM_FORMAT_RGB565; + break; + case 24: + fmt = DRM_FORMAT_RGB888; + break; + case 32: + if (depth == 24) + fmt = DRM_FORMAT_XRGB8888; + else if (depth == 30) + fmt = DRM_FORMAT_XRGB2101010; + else + fmt = DRM_FORMAT_ARGB8888; + break; + default: + DRM_ERROR("bad bpp, assuming x8r8g8b8 pixel format\n"); + fmt = DRM_FORMAT_XRGB8888; + break; + } + + return fmt; +} +EXPORT_SYMBOL(drm_mode_legacy_fb_format); + /** * drm_mode_addfb - add an FB to the graphics configuration * @inode: inode from the ioctl @@ -1680,18 +2072,27 @@ out: int drm_mode_addfb(struct drm_device *dev, void *data, struct drm_file *file_priv) { - struct drm_mode_fb_cmd *r = data; + struct drm_mode_fb_cmd *or = data; + struct drm_mode_fb_cmd2 r = {}; struct drm_mode_config *config = &dev->mode_config; struct drm_framebuffer *fb; int ret = 0; + /* Use new struct with format internally */ + r.fb_id = or->fb_id; + r.width = or->width; + r.height = or->height; + r.pitches[0] = or->pitch; + r.pixel_format = drm_mode_legacy_fb_format(or->bpp, or->depth); + r.handles[0] = or->handle; + if (!drm_core_check_feature(dev, DRIVER_MODESET)) return -EINVAL; - if ((config->min_width > r->width) || (r->width > config->max_width)) + if ((config->min_width > r.width) || (r.width > config->max_width)) return -EINVAL; - if ((config->min_height > r->height) || (r->height > config->max_height)) + if ((config->min_height > r.height) || (r.height > config->max_height)) return -EINVAL; mutex_lock(&dev->mode_config.mutex); @@ -1699,6 +2100,138 @@ int drm_mode_addfb(struct drm_device *dev, /* TODO check buffer is sufficiently large */ /* TODO setup destructor callback */ + fb = dev->mode_config.funcs->fb_create(dev, file_priv, &r); + if (IS_ERR(fb)) { + DRM_ERROR("could not create framebuffer\n"); + ret = PTR_ERR(fb); + goto out; + } + + or->fb_id = fb->base.id; + list_add(&fb->filp_head, &file_priv->fbs); + DRM_DEBUG_KMS("[FB:%d]\n", fb->base.id); + +out: + mutex_unlock(&dev->mode_config.mutex); + return ret; +} + +static int format_check(struct drm_mode_fb_cmd2 *r) +{ + uint32_t format = r->pixel_format & ~DRM_FORMAT_BIG_ENDIAN; + + switch (format) { + case DRM_FORMAT_C8: + case DRM_FORMAT_RGB332: + case DRM_FORMAT_BGR233: + case DRM_FORMAT_XRGB4444: + case DRM_FORMAT_XBGR4444: + case DRM_FORMAT_RGBX4444: + case DRM_FORMAT_BGRX4444: + case DRM_FORMAT_ARGB4444: + case DRM_FORMAT_ABGR4444: + case DRM_FORMAT_RGBA4444: + case DRM_FORMAT_BGRA4444: + case DRM_FORMAT_XRGB1555: + case DRM_FORMAT_XBGR1555: + case DRM_FORMAT_RGBX5551: + case DRM_FORMAT_BGRX5551: + case DRM_FORMAT_ARGB1555: + case DRM_FORMAT_ABGR1555: + case DRM_FORMAT_RGBA5551: + case DRM_FORMAT_BGRA5551: + case DRM_FORMAT_RGB565: + case DRM_FORMAT_BGR565: + case DRM_FORMAT_RGB888: + case DRM_FORMAT_BGR888: + case DRM_FORMAT_XRGB8888: + case DRM_FORMAT_XBGR8888: + case DRM_FORMAT_RGBX8888: + case DRM_FORMAT_BGRX8888: + case DRM_FORMAT_ARGB8888: + case DRM_FORMAT_ABGR8888: + case DRM_FORMAT_RGBA8888: + case DRM_FORMAT_BGRA8888: + case DRM_FORMAT_XRGB2101010: + case DRM_FORMAT_XBGR2101010: + case DRM_FORMAT_RGBX1010102: + case DRM_FORMAT_BGRX1010102: + case DRM_FORMAT_ARGB2101010: + case DRM_FORMAT_ABGR2101010: + case DRM_FORMAT_RGBA1010102: + case DRM_FORMAT_BGRA1010102: + case DRM_FORMAT_YUYV: + case DRM_FORMAT_YVYU: + case DRM_FORMAT_UYVY: + case DRM_FORMAT_VYUY: + case DRM_FORMAT_AYUV: + case DRM_FORMAT_NV12: + case DRM_FORMAT_NV21: + case DRM_FORMAT_NV16: + case DRM_FORMAT_NV61: + case DRM_FORMAT_YUV410: + case DRM_FORMAT_YVU410: + case DRM_FORMAT_YUV411: + case DRM_FORMAT_YVU411: + case DRM_FORMAT_YUV420: + case DRM_FORMAT_YVU420: + case DRM_FORMAT_YUV422: + case DRM_FORMAT_YVU422: + case DRM_FORMAT_YUV444: + case DRM_FORMAT_YVU444: + return 0; + default: + return -EINVAL; + } +} + +/** + * drm_mode_addfb2 - add an FB to the graphics configuration + * @inode: inode from the ioctl + * @filp: file * from the ioctl + * @cmd: cmd from ioctl + * @arg: arg from ioctl + * + * LOCKING: + * Takes mode config lock. + * + * Add a new FB to the specified CRTC, given a user request with format. + * + * Called by the user via ioctl. + * + * RETURNS: + * Zero on success, errno on failure. + */ +int drm_mode_addfb2(struct drm_device *dev, + void *data, struct drm_file *file_priv) +{ + struct drm_mode_fb_cmd2 *r = data; + struct drm_mode_config *config = &dev->mode_config; + struct drm_framebuffer *fb; + int ret = 0; + + if (!drm_core_check_feature(dev, DRIVER_MODESET)) + return -EINVAL; + + if ((config->min_width > r->width) || (r->width > config->max_width)) { + DRM_ERROR("bad framebuffer width %d, should be >= %d && <= %d\n", + r->width, config->min_width, config->max_width); + return -EINVAL; + } + if ((config->min_height > r->height) || (r->height > config->max_height)) { + DRM_ERROR("bad framebuffer height %d, should be >= %d && <= %d\n", + r->height, config->min_height, config->max_height); + return -EINVAL; + } + + ret = format_check(r); + if (ret) { + DRM_ERROR("bad framebuffer format 0x%08x\n", r->pixel_format); + return ret; + } + + mutex_lock(&dev->mode_config.mutex); + fb = dev->mode_config.funcs->fb_create(dev, file_priv, r); if (IS_ERR(fb)) { DRM_ERROR("could not create framebuffer\n"); @@ -1782,7 +2315,7 @@ out: * @arg: arg from ioctl * * LOCKING: - * Caller? (FIXME) + * Takes mode config lock. * * Lookup the FB given its ID and return info about it. * @@ -1814,7 +2347,7 @@ int drm_mode_getfb(struct drm_device *dev, r->width = fb->width; r->depth = fb->depth; r->bpp = fb->bits_per_pixel; - r->pitch = fb->pitch; + r->pitch = fb->pitches[0]; fb->funcs->create_handle(fb, file_priv, &r->handle); out: @@ -1846,7 +2379,7 @@ int drm_mode_dirtyfb_ioctl(struct drm_device *dev, fb = obj_to_fb(obj); num_clips = r->num_clips; - clips_ptr = (struct drm_clip_rect *)(unsigned long)r->clips_ptr; + clips_ptr = (struct drm_clip_rect __user *)(unsigned long)r->clips_ptr; if (!num_clips != !clips_ptr) { ret = -EINVAL; @@ -1931,38 +2464,48 @@ void drm_fb_release(struct drm_file *priv) * * Add @mode to @connector's user mode list. */ -static int drm_mode_attachmode(struct drm_device *dev, - struct drm_connector *connector, - struct drm_display_mode *mode) +static void drm_mode_attachmode(struct drm_device *dev, + struct drm_connector *connector, + struct drm_display_mode *mode) { - int ret = 0; - list_add_tail(&mode->head, &connector->user_modes); - return ret; } int drm_mode_attachmode_crtc(struct drm_device *dev, struct drm_crtc *crtc, - struct drm_display_mode *mode) + const struct drm_display_mode *mode) { struct drm_connector *connector; int ret = 0; - struct drm_display_mode *dup_mode; - int need_dup = 0; + struct drm_display_mode *dup_mode, *next; + LIST_HEAD(list); + list_for_each_entry(connector, &dev->mode_config.connector_list, head) { if (!connector->encoder) - break; + continue; if (connector->encoder->crtc == crtc) { - if (need_dup) - dup_mode = drm_mode_duplicate(dev, mode); - else - dup_mode = mode; - ret = drm_mode_attachmode(dev, connector, dup_mode); - if (ret) - return ret; - need_dup = 1; + dup_mode = drm_mode_duplicate(dev, mode); + if (!dup_mode) { + ret = -ENOMEM; + goto out; + } + list_add_tail(&dup_mode->head, &list); } } - return 0; + + list_for_each_entry(connector, &dev->mode_config.connector_list, head) { + if (!connector->encoder) + continue; + if (connector->encoder->crtc == crtc) + list_move_tail(list.next, &connector->user_modes); + } + + WARN_ON(!list_empty(&list)); + + out: + list_for_each_entry_safe(dup_mode, next, &list, head) + drm_mode_destroy(dev, dup_mode); + + return ret; } EXPORT_SYMBOL(drm_mode_attachmode_crtc); @@ -2041,9 +2584,14 @@ int drm_mode_attachmode_ioctl(struct drm_device *dev, goto out; } - drm_crtc_convert_umode(mode, umode); + ret = drm_crtc_convert_umode(mode, umode); + if (ret) { + DRM_DEBUG_KMS("Invalid mode\n"); + drm_mode_destroy(dev, mode); + goto out; + } - ret = drm_mode_attachmode(dev, connector, mode); + drm_mode_attachmode(dev, connector, mode); out: mutex_unlock(&dev->mode_config.mutex); return ret; @@ -2084,7 +2632,12 @@ int drm_mode_detachmode_ioctl(struct drm_device *dev, } connector = obj_to_connector(obj); - drm_crtc_convert_umode(&mode, umode); + ret = drm_crtc_convert_umode(&mode, umode); + if (ret) { + DRM_DEBUG_KMS("Invalid mode\n"); + goto out; + } + ret = drm_mode_detachmode(dev, connector, &mode); out: mutex_unlock(&dev->mode_config.mutex); @@ -2095,6 +2648,7 @@ struct drm_property *drm_property_create(struct drm_device *dev, int flags, const char *name, int num_values) { struct drm_property *property = NULL; + int ret; property = kzalloc(sizeof(struct drm_property), GFP_KERNEL); if (!property) @@ -2106,7 +2660,10 @@ struct drm_property *drm_property_create(struct drm_device *dev, int flags, goto fail; } - drm_mode_object_get(dev, &property->base, DRM_MODE_OBJECT_PROPERTY); + ret = drm_mode_object_get(dev, &property->base, DRM_MODE_OBJECT_PROPERTY); + if (ret) + goto fail; + property->flags = flags; property->num_values = num_values; INIT_LIST_HEAD(&property->enum_blob_list); @@ -2119,11 +2676,59 @@ struct drm_property *drm_property_create(struct drm_device *dev, int flags, list_add_tail(&property->head, &dev->mode_config.property_list); return property; fail: + kfree(property->values); kfree(property); return NULL; } EXPORT_SYMBOL(drm_property_create); +struct drm_property *drm_property_create_enum(struct drm_device *dev, int flags, + const char *name, + const struct drm_prop_enum_list *props, + int num_values) +{ + struct drm_property *property; + int i, ret; + + flags |= DRM_MODE_PROP_ENUM; + + property = drm_property_create(dev, flags, name, num_values); + if (!property) + return NULL; + + for (i = 0; i < num_values; i++) { + ret = drm_property_add_enum(property, i, + props[i].type, + props[i].name); + if (ret) { + drm_property_destroy(dev, property); + return NULL; + } + } + + return property; +} +EXPORT_SYMBOL(drm_property_create_enum); + +struct drm_property *drm_property_create_range(struct drm_device *dev, int flags, + const char *name, + uint64_t min, uint64_t max) +{ + struct drm_property *property; + + flags |= DRM_MODE_PROP_RANGE; + + property = drm_property_create(dev, flags, name, 2); + if (!property) + return NULL; + + property->values[0] = min; + property->values[1] = max; + + return property; +} +EXPORT_SYMBOL(drm_property_create_range); + int drm_property_add_enum(struct drm_property *property, int index, uint64_t value, const char *name) { @@ -2242,7 +2847,7 @@ int drm_mode_getproperty_ioctl(struct drm_device *dev, struct drm_property_enum *prop_enum; struct drm_mode_property_enum __user *enum_ptr; struct drm_property_blob *prop_blob; - uint32_t *blob_id_ptr; + uint32_t __user *blob_id_ptr; uint64_t __user *values_ptr; uint32_t __user *blob_length_ptr; @@ -2272,7 +2877,7 @@ int drm_mode_getproperty_ioctl(struct drm_device *dev, out_resp->flags = property->flags; if ((out_resp->count_values >= value_count) && value_count) { - values_ptr = (uint64_t *)(unsigned long)out_resp->values_ptr; + values_ptr = (uint64_t __user *)(unsigned long)out_resp->values_ptr; for (i = 0; i < value_count; i++) { if (copy_to_user(values_ptr + i, &property->values[i], sizeof(uint64_t))) { ret = -EFAULT; @@ -2285,7 +2890,7 @@ int drm_mode_getproperty_ioctl(struct drm_device *dev, if (property->flags & DRM_MODE_PROP_ENUM) { if ((out_resp->count_enum_blobs >= enum_count) && enum_count) { copied = 0; - enum_ptr = (struct drm_mode_property_enum *)(unsigned long)out_resp->enum_blob_ptr; + enum_ptr = (struct drm_mode_property_enum __user *)(unsigned long)out_resp->enum_blob_ptr; list_for_each_entry(prop_enum, &property->enum_blob_list, head) { if (copy_to_user(&enum_ptr[copied].value, &prop_enum->value, sizeof(uint64_t))) { @@ -2307,8 +2912,8 @@ int drm_mode_getproperty_ioctl(struct drm_device *dev, if (property->flags & DRM_MODE_PROP_BLOB) { if ((out_resp->count_enum_blobs >= blob_count) && blob_count) { copied = 0; - blob_id_ptr = (uint32_t *)(unsigned long)out_resp->enum_blob_ptr; - blob_length_ptr = (uint32_t *)(unsigned long)out_resp->values_ptr; + blob_id_ptr = (uint32_t __user *)(unsigned long)out_resp->enum_blob_ptr; + blob_length_ptr = (uint32_t __user *)(unsigned long)out_resp->values_ptr; list_for_each_entry(prop_blob, &property->enum_blob_list, head) { if (put_user(prop_blob->base.id, blob_id_ptr + copied)) { @@ -2335,6 +2940,7 @@ static struct drm_property_blob *drm_property_create_blob(struct drm_device *dev void *data) { struct drm_property_blob *blob; + int ret; if (!length || !data) return NULL; @@ -2343,13 +2949,16 @@ static struct drm_property_blob *drm_property_create_blob(struct drm_device *dev if (!blob) return NULL; - blob->data = (void *)((char *)blob + sizeof(struct drm_property_blob)); + ret = drm_mode_object_get(dev, &blob->base, DRM_MODE_OBJECT_BLOB); + if (ret) { + kfree(blob); + return NULL; + } + blob->length = length; memcpy(blob->data, data, length); - drm_mode_object_get(dev, &blob->base, DRM_MODE_OBJECT_BLOB); - list_add_tail(&blob->head, &dev->mode_config.property_blob_list); return blob; } @@ -2369,7 +2978,7 @@ int drm_mode_getblob_ioctl(struct drm_device *dev, struct drm_mode_get_blob *out_resp = data; struct drm_property_blob *blob; int ret = 0; - void *blob_ptr; + void __user *blob_ptr; if (!drm_core_check_feature(dev, DRIVER_MODESET)) return -EINVAL; @@ -2383,7 +2992,7 @@ int drm_mode_getblob_ioctl(struct drm_device *dev, blob = obj_to_blob(obj); if (out_resp->length == blob->length) { - blob_ptr = (void *)(unsigned long)out_resp->data; + blob_ptr = (void __user *)(unsigned long)out_resp->data; if (copy_to_user(blob_ptr, blob->data, blob->length)){ ret = -EFAULT; goto done; @@ -2528,7 +3137,7 @@ void drm_mode_connector_detach_encoder(struct drm_connector *connector, } EXPORT_SYMBOL(drm_mode_connector_detach_encoder); -bool drm_mode_crtc_set_gamma_size(struct drm_crtc *crtc, +int drm_mode_crtc_set_gamma_size(struct drm_crtc *crtc, int gamma_size) { crtc->gamma_size = gamma_size; @@ -2536,10 +3145,10 @@ bool drm_mode_crtc_set_gamma_size(struct drm_crtc *crtc, crtc->gamma_store = kzalloc(gamma_size * sizeof(uint16_t) * 3, GFP_KERNEL); if (!crtc->gamma_store) { crtc->gamma_size = 0; - return false; + return -ENOMEM; } - return true; + return 0; } EXPORT_SYMBOL(drm_mode_crtc_set_gamma_size); @@ -2685,6 +3294,18 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev, goto out; fb = obj_to_fb(obj); + if (crtc->mode.hdisplay > fb->width || + crtc->mode.vdisplay > fb->height || + crtc->x > fb->width - crtc->mode.hdisplay || + crtc->y > fb->height - crtc->mode.vdisplay) { + DRM_DEBUG_KMS("Invalid fb size %ux%u for CRTC viewport %ux%u+%d+%d.\n", + fb->width, fb->height, + crtc->mode.hdisplay, crtc->mode.vdisplay, + crtc->x, crtc->y); + ret = -ENOSPC; + goto out; + } + if (page_flip->flags & DRM_MODE_PAGE_FLIP_EVENT) { ret = -ENOMEM; spin_lock_irqsave(&dev->event_lock, flags); @@ -2714,10 +3335,12 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev, ret = crtc->funcs->page_flip(crtc, fb, e); if (ret) { - spin_lock_irqsave(&dev->event_lock, flags); - file_priv->event_space += sizeof e->event; - spin_unlock_irqrestore(&dev->event_lock, flags); - kfree(e); + if (page_flip->flags & DRM_MODE_PAGE_FLIP_EVENT) { + spin_lock_irqsave(&dev->event_lock, flags); + file_priv->event_space += sizeof e->event; + spin_unlock_irqrestore(&dev->event_lock, flags); + kfree(e); + } } out: @@ -2777,3 +3400,71 @@ int drm_mode_destroy_dumb_ioctl(struct drm_device *dev, return dev->driver->dumb_destroy(file_priv, dev, args->handle); } + +/* + * Just need to support RGB formats here for compat with code that doesn't + * use pixel formats directly yet. + */ +void drm_fb_get_bpp_depth(uint32_t format, unsigned int *depth, + int *bpp) +{ + switch (format) { + case DRM_FORMAT_RGB332: + case DRM_FORMAT_BGR233: + *depth = 8; + *bpp = 8; + break; + case DRM_FORMAT_XRGB1555: + case DRM_FORMAT_XBGR1555: + case DRM_FORMAT_RGBX5551: + case DRM_FORMAT_BGRX5551: + case DRM_FORMAT_ARGB1555: + case DRM_FORMAT_ABGR1555: + case DRM_FORMAT_RGBA5551: + case DRM_FORMAT_BGRA5551: + *depth = 15; + *bpp = 16; + break; + case DRM_FORMAT_RGB565: + case DRM_FORMAT_BGR565: + *depth = 16; + *bpp = 16; + break; + case DRM_FORMAT_RGB888: + case DRM_FORMAT_BGR888: + *depth = 24; + *bpp = 24; + break; + case DRM_FORMAT_XRGB8888: + case DRM_FORMAT_XBGR8888: + case DRM_FORMAT_RGBX8888: + case DRM_FORMAT_BGRX8888: + *depth = 24; + *bpp = 32; + break; + case DRM_FORMAT_XRGB2101010: + case DRM_FORMAT_XBGR2101010: + case DRM_FORMAT_RGBX1010102: + case DRM_FORMAT_BGRX1010102: + case DRM_FORMAT_ARGB2101010: + case DRM_FORMAT_ABGR2101010: + case DRM_FORMAT_RGBA1010102: + case DRM_FORMAT_BGRA1010102: + *depth = 30; + *bpp = 32; + break; + case DRM_FORMAT_ARGB8888: + case DRM_FORMAT_ABGR8888: + case DRM_FORMAT_RGBA8888: + case DRM_FORMAT_BGRA8888: + *depth = 32; + *bpp = 32; + break; + default: + DRM_DEBUG_KMS("unsupported pixel format\n"); + *depth = 0; + *bpp = 0; + break; + } +} +EXPORT_SYMBOL(drm_fb_get_bpp_depth); diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c index 11788f7..8111889 100644 --- a/drivers/gpu/drm/drm_crtc_helper.c +++ b/drivers/gpu/drm/drm_crtc_helper.c @@ -34,8 +34,10 @@ #include "drmP.h" #include "drm_crtc.h" +#include "drm_fourcc.h" #include "drm_crtc_helper.h" #include "drm_fb_helper.h" +#include "drm_edid.h" static bool drm_kms_helper_poll = true; module_param_named(poll, drm_kms_helper_poll, bool, 0600); @@ -43,12 +45,12 @@ module_param_named(poll, drm_kms_helper_poll, bool, 0600); static void drm_mode_validate_flag(struct drm_connector *connector, int flags) { - struct drm_display_mode *mode, *t; + struct drm_display_mode *mode; if (flags == (DRM_MODE_FLAG_DBLSCAN | DRM_MODE_FLAG_INTERLACE)) return; - list_for_each_entry_safe(mode, t, &connector->modes, head) { + list_for_each_entry(mode, &connector->modes, head) { if ((mode->flags & DRM_MODE_FLAG_INTERLACE) && !(flags & DRM_MODE_FLAG_INTERLACE)) mode->status = MODE_NO_INTERLACE; @@ -86,7 +88,7 @@ int drm_helper_probe_single_connector_modes(struct drm_connector *connector, uint32_t maxX, uint32_t maxY) { struct drm_device *dev = connector->dev; - struct drm_display_mode *mode, *t; + struct drm_display_mode *mode; struct drm_connector_helper_funcs *connector_funcs = connector->helper_private; int count = 0; @@ -95,7 +97,7 @@ int drm_helper_probe_single_connector_modes(struct drm_connector *connector, DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n", connector->base.id, drm_get_connector_name(connector)); /* set all modes to the unverified state */ - list_for_each_entry_safe(mode, t, &connector->modes, head) + list_for_each_entry(mode, &connector->modes, head) mode->status = MODE_UNVERIFIED; if (connector->force) { @@ -117,7 +119,12 @@ int drm_helper_probe_single_connector_modes(struct drm_connector *connector, goto prune; } - count = (*connector_funcs->get_modes)(connector); +#ifdef CONFIG_DRM_LOAD_EDID_FIRMWARE + count = drm_load_edid_firmware(connector); + if (count == 0) +#endif + count = (*connector_funcs->get_modes)(connector); + if (count == 0 && connector->status == connector_status_connected) count = drm_add_modes_noedid(connector, 1024, 768); if (count == 0) @@ -135,7 +142,7 @@ int drm_helper_probe_single_connector_modes(struct drm_connector *connector, mode_flags |= DRM_MODE_FLAG_DBLSCAN; drm_mode_validate_flag(connector, mode_flags); - list_for_each_entry_safe(mode, t, &connector->modes, head) { + list_for_each_entry(mode, &connector->modes, head) { if (mode->status == MODE_OK) mode->status = connector_funcs->mode_valid(connector, mode); @@ -151,7 +158,7 @@ prune: DRM_DEBUG_KMS("[CONNECTOR:%d:%s] probed modes :\n", connector->base.id, drm_get_connector_name(connector)); - list_for_each_entry_safe(mode, t, &connector->modes, head) { + list_for_each_entry(mode, &connector->modes, head) { mode->vrefresh = drm_mode_vrefresh(mode); drm_mode_set_crtcinfo(mode, CRTC_INTERLACE_HALVE_V); @@ -321,8 +328,8 @@ drm_crtc_prepare_encoders(struct drm_device *dev) * drm_crtc_set_mode - set a mode * @crtc: CRTC to program * @mode: mode to use - * @x: horizontal offset into the surface - * @y: vertical offset into the surface + * @x: width of mode + * @y: height of mode * * LOCKING: * Caller must hold mode config lock. @@ -351,6 +358,8 @@ bool drm_crtc_helper_set_mode(struct drm_crtc *crtc, return true; adjusted_mode = drm_mode_duplicate(dev, mode); + if (!adjusted_mode) + return false; saved_hwmode = crtc->hwmode; saved_mode = crtc->mode; @@ -710,7 +719,7 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set) for (i = 0; i < set->num_connectors; i++) { DRM_DEBUG_KMS("\t[CONNECTOR:%d:%s] set DPMS on\n", set->connectors[i]->base.id, drm_get_connector_name(set->connectors[i])); - set->connectors[i]->dpms = DRM_MODE_DPMS_ON; + set->connectors[i]->funcs->dpms(set->connectors[i], DRM_MODE_DPMS_ON); } } drm_helper_disable_unused_functions(dev); @@ -847,13 +856,19 @@ void drm_helper_connector_dpms(struct drm_connector *connector, int mode) EXPORT_SYMBOL(drm_helper_connector_dpms); int drm_helper_mode_fill_fb_struct(struct drm_framebuffer *fb, - struct drm_mode_fb_cmd *mode_cmd) + struct drm_mode_fb_cmd2 *mode_cmd) { + int i; + fb->width = mode_cmd->width; fb->height = mode_cmd->height; - fb->pitch = mode_cmd->pitch; - fb->bits_per_pixel = mode_cmd->bpp; - fb->depth = mode_cmd->depth; + for (i = 0; i < 4; i++) { + fb->pitches[i] = mode_cmd->pitches[i]; + fb->offsets[i] = mode_cmd->offsets[i]; + } + drm_fb_get_bpp_depth(mode_cmd->pixel_format, &fb->depth, + &fb->bits_per_pixel); + fb->pixel_format = mode_cmd->pixel_format; return 0; } @@ -1008,3 +1023,36 @@ void drm_helper_hpd_irq_event(struct drm_device *dev) queue_delayed_work(system_nrt_wq, &dev->mode_config.output_poll_work, 0); } EXPORT_SYMBOL(drm_helper_hpd_irq_event); + + +/** + * drm_format_num_planes - get the number of planes for format + * @format: pixel format (DRM_FORMAT_*) + * + * RETURNS: + * The number of planes used by the specified pixel format. + */ +int drm_format_num_planes(uint32_t format) +{ + switch (format) { + case DRM_FORMAT_YUV410: + case DRM_FORMAT_YVU410: + case DRM_FORMAT_YUV411: + case DRM_FORMAT_YVU411: + case DRM_FORMAT_YUV420: + case DRM_FORMAT_YVU420: + case DRM_FORMAT_YUV422: + case DRM_FORMAT_YVU422: + case DRM_FORMAT_YUV444: + case DRM_FORMAT_YVU444: + return 3; + case DRM_FORMAT_NV12: + case DRM_FORMAT_NV21: + case DRM_FORMAT_NV16: + case DRM_FORMAT_NV61: + return 2; + default: + return 1; + } +} +EXPORT_SYMBOL(drm_format_num_planes); diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c index 40c187c..6116e3b 100644 --- a/drivers/gpu/drm/drm_drv.c +++ b/drivers/gpu/drm/drm_drv.c @@ -61,14 +61,14 @@ static int drm_version(struct drm_device *dev, void *data, /** Ioctl table */ static struct drm_ioctl_desc drm_ioctls[] = { - DRM_IOCTL_DEF(DRM_IOCTL_VERSION, drm_version, 0), + DRM_IOCTL_DEF(DRM_IOCTL_VERSION, drm_version, DRM_UNLOCKED), DRM_IOCTL_DEF(DRM_IOCTL_GET_UNIQUE, drm_getunique, 0), DRM_IOCTL_DEF(DRM_IOCTL_GET_MAGIC, drm_getmagic, 0), DRM_IOCTL_DEF(DRM_IOCTL_IRQ_BUSID, drm_irq_by_busid, DRM_MASTER|DRM_ROOT_ONLY), - DRM_IOCTL_DEF(DRM_IOCTL_GET_MAP, drm_getmap, 0), - DRM_IOCTL_DEF(DRM_IOCTL_GET_CLIENT, drm_getclient, 0), - DRM_IOCTL_DEF(DRM_IOCTL_GET_STATS, drm_getstats, 0), - DRM_IOCTL_DEF(DRM_IOCTL_GET_CAP, drm_getcap, 0), + DRM_IOCTL_DEF(DRM_IOCTL_GET_MAP, drm_getmap, DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_GET_CLIENT, drm_getclient, DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_GET_STATS, drm_getstats, DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_GET_CAP, drm_getcap, DRM_UNLOCKED), DRM_IOCTL_DEF(DRM_IOCTL_SET_VERSION, drm_setversion, DRM_MASTER), DRM_IOCTL_DEF(DRM_IOCTL_SET_UNIQUE, drm_setunique, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY), @@ -135,21 +135,29 @@ static struct drm_ioctl_desc drm_ioctls[] = { DRM_IOCTL_DEF(DRM_IOCTL_GEM_FLINK, drm_gem_flink_ioctl, DRM_AUTH|DRM_UNLOCKED), DRM_IOCTL_DEF(DRM_IOCTL_GEM_OPEN, drm_gem_open_ioctl, DRM_AUTH|DRM_UNLOCKED), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETRESOURCES, drm_mode_getresources, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCRTC, drm_mode_getcrtc, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETRESOURCES, drm_mode_getresources, DRM_CONTROL_ALLOW|DRM_UNLOCKED), + + DRM_IOCTL_DEF(DRM_IOCTL_PRIME_HANDLE_TO_FD, drm_prime_handle_to_fd_ioctl, DRM_AUTH|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_PRIME_FD_TO_HANDLE, drm_prime_fd_to_handle_ioctl, DRM_AUTH|DRM_UNLOCKED), + + DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPLANERESOURCES, drm_mode_getplane_res, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCRTC, drm_mode_getcrtc, DRM_CONTROL_ALLOW|DRM_UNLOCKED), DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETCRTC, drm_mode_setcrtc, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPLANE, drm_mode_getplane, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETPLANE, drm_mode_setplane, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), DRM_IOCTL_DEF(DRM_IOCTL_MODE_CURSOR, drm_mode_cursor_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETGAMMA, drm_mode_gamma_get_ioctl, DRM_MASTER|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETGAMMA, drm_mode_gamma_get_ioctl, DRM_UNLOCKED), DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETGAMMA, drm_mode_gamma_set_ioctl, DRM_MASTER|DRM_UNLOCKED), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETENCODER, drm_mode_getencoder, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCONNECTOR, drm_mode_getconnector, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETENCODER, drm_mode_getencoder, DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETCONNECTOR, drm_mode_getconnector, DRM_CONTROL_ALLOW|DRM_UNLOCKED), DRM_IOCTL_DEF(DRM_IOCTL_MODE_ATTACHMODE, drm_mode_attachmode_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), DRM_IOCTL_DEF(DRM_IOCTL_MODE_DETACHMODE, drm_mode_detachmode_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPROPERTY, drm_mode_getproperty_ioctl, DRM_MASTER | DRM_CONTROL_ALLOW|DRM_UNLOCKED), DRM_IOCTL_DEF(DRM_IOCTL_MODE_SETPROPERTY, drm_mode_connector_property_set_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETPROPBLOB, drm_mode_getblob_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), - DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETFB, drm_mode_getfb, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_GETFB, drm_mode_getfb, DRM_CONTROL_ALLOW|DRM_UNLOCKED), DRM_IOCTL_DEF(DRM_IOCTL_MODE_ADDFB, drm_mode_addfb, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF(DRM_IOCTL_MODE_ADDFB2, drm_mode_addfb2, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), DRM_IOCTL_DEF(DRM_IOCTL_MODE_RMFB, drm_mode_rmfb, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), DRM_IOCTL_DEF(DRM_IOCTL_MODE_PAGE_FLIP, drm_mode_page_flip_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), DRM_IOCTL_DEF(DRM_IOCTL_MODE_DIRTYFB, drm_mode_dirtyfb_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), @@ -386,6 +394,10 @@ long drm_ioctl(struct file *filp, unsigned int usize, asize; dev = file_priv->minor->dev; + + if (drm_device_is_unplugged(dev)) + return -ENODEV; + atomic_inc(&dev->ioctl_count); atomic_inc(&dev->counts[_DRM_STAT_IOCTLS]); ++file_priv->ioctl_count; diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c index bb95d59..9d9835a 100644 --- a/drivers/gpu/drm/drm_edid.c +++ b/drivers/gpu/drm/drm_edid.c @@ -87,9 +87,6 @@ static struct edid_quirk { int product_id; u32 quirks; } edid_quirk_list[] = { - /* ASUS VW222S */ - { "ACI", 0x22a2, EDID_QUIRK_FORCE_REDUCED_BLANKING }, - /* Acer AL1706 */ { "ACR", 44358, EDID_QUIRK_PREFER_LARGE_60 }, /* Acer F51 */ @@ -157,8 +154,7 @@ EXPORT_SYMBOL(drm_edid_header_is_valid); * Sanity check the EDID block (base or extension). Return 0 if the block * doesn't check out, or 1 if it's valid. */ -static bool -drm_edid_block_valid(u8 *raw_edid) +bool drm_edid_block_valid(u8 *raw_edid) { int i; u8 csum = 0; @@ -211,6 +207,7 @@ bad: } return 0; } +EXPORT_SYMBOL(drm_edid_block_valid); /** * drm_edid_is_valid - sanity check EDID data @@ -234,7 +231,6 @@ bool drm_edid_is_valid(struct edid *edid) } EXPORT_SYMBOL(drm_edid_is_valid); -#define DDC_ADDR 0x50 #define DDC_SEGMENT_ADDR 0x30 /** * Get EDID information via I2C. @@ -521,25 +517,10 @@ static void cea_for_each_detailed_block(u8 *ext, detailed_cb *cb, void *closure) { int i, n = 0; - u8 rev = ext[0x01], d = ext[0x02]; + u8 d = ext[0x02]; u8 *det_base = ext + d; - switch (rev) { - case 0: - /* can't happen */ - return; - case 1: - /* have to infer how many blocks we have, check pixel clock */ - for (i = 0; i < 6; i++) - if (det_base[18*i] || det_base[18*i+1]) - n++; - break; - default: - /* explicit count */ - n = min(ext[0x03] & 0x0f, 6); - break; - } - + n = (127 - d) / 18; for (i = 0; i < n; i++) cb((struct detailed_timing *)(det_base + 18 * i), closure); } @@ -773,7 +754,7 @@ drm_mode_std(struct drm_connector *connector, struct edid *edid, */ mode = drm_gtf_mode(dev, hsize, vsize, vrefresh_rate, 0, 0); if (drm_mode_hsync(mode) > drm_gtf2_hbreak(edid)) { - kfree(mode); + drm_mode_destroy(dev, mode); mode = drm_gtf_mode_complex(dev, hsize, vsize, vrefresh_rate, 0, 0, drm_gtf2_m(edid), @@ -1341,6 +1322,7 @@ add_detailed_modes(struct drm_connector *connector, struct edid *edid, #define HDMI_IDENTIFIER 0x000C03 #define AUDIO_BLOCK 0x01 +#define VIDEO_BLOCK 0x02 #define VENDOR_BLOCK 0x03 #define SPEAKER_BLOCK 0x04 #define EDID_BASIC_AUDIO (1 << 6) @@ -1371,6 +1353,47 @@ u8 *drm_find_cea_extension(struct edid *edid) } EXPORT_SYMBOL(drm_find_cea_extension); +static int +do_cea_modes (struct drm_connector *connector, u8 *db, u8 len) +{ + struct drm_device *dev = connector->dev; + u8 * mode, cea_mode; + int modes = 0; + + for (mode = db; mode < db + len; mode++) { + cea_mode = (*mode & 127) - 1; /* CEA modes are numbered 1..127 */ + if (cea_mode < drm_num_cea_modes) { + struct drm_display_mode *newmode; + newmode = drm_mode_duplicate(dev, + &edid_cea_modes[cea_mode]); + if (newmode) { + drm_mode_probed_add(connector, newmode); + modes++; + } + } + } + + return modes; +} + +static int +add_cea_modes(struct drm_connector *connector, struct edid *edid) +{ + u8 * cea = drm_find_cea_extension(edid); + u8 * db, dbl; + int modes = 0; + + if (cea && cea[1] >= 3) { + for (db = cea + 4; db < cea + cea[2]; db += dbl + 1) { + dbl = db[0] & 0x1f; + if (((db[0] & 0xe0) >> 5) == VIDEO_BLOCK) + modes += do_cea_modes (connector, db+1, dbl); + } + } + + return modes; +} + static void parse_hdmi_vsdb(struct drm_connector *connector, uint8_t *db) { @@ -1454,26 +1477,29 @@ void drm_edid_to_eld(struct drm_connector *connector, struct edid *edid) eld[18] = edid->prod_code[0]; eld[19] = edid->prod_code[1]; - for (db = cea + 4; db < cea + cea[2]; db += dbl + 1) { - dbl = db[0] & 0x1f; - - switch ((db[0] & 0xe0) >> 5) { - case AUDIO_BLOCK: /* Audio Data Block, contains SADs */ - sad_count = dbl / 3; - memcpy(eld + 20 + mnl, &db[1], dbl); - break; - case SPEAKER_BLOCK: /* Speaker Allocation Data Block */ - eld[7] = db[1]; - break; - case VENDOR_BLOCK: - /* HDMI Vendor-Specific Data Block */ - if (db[1] == 0x03 && db[2] == 0x0c && db[3] == 0) - parse_hdmi_vsdb(connector, db); - break; - default: - break; + if (cea[1] >= 3) + for (db = cea + 4; db < cea + cea[2]; db += dbl + 1) { + dbl = db[0] & 0x1f; + + switch ((db[0] & 0xe0) >> 5) { + case AUDIO_BLOCK: + /* Audio Data Block, contains SADs */ + sad_count = dbl / 3; + memcpy(eld + 20 + mnl, &db[1], dbl); + break; + case SPEAKER_BLOCK: + /* Speaker Allocation Data Block */ + eld[7] = db[1]; + break; + case VENDOR_BLOCK: + /* HDMI Vendor-Specific Data Block */ + if (db[1] == 0x03 && db[2] == 0x0c && db[3] == 0) + parse_hdmi_vsdb(connector, db); + break; + default: + break; + } } - } eld[5] |= sad_count << 4; eld[2] = (20 + mnl + sad_count * 3 + 3) / 4; @@ -1744,6 +1770,7 @@ int drm_add_edid_modes(struct drm_connector *connector, struct edid *edid) num_modes += add_standard_modes(connector, edid); num_modes += add_established_modes(connector, edid); num_modes += add_inferred_modes(connector, edid); + num_modes += add_cea_modes(connector, edid); if (quirks & (EDID_QUIRK_PREFER_LARGE_60 | EDID_QUIRK_PREFER_LARGE_75)) edid_fixup_preferred(connector, quirks); diff --git a/drivers/gpu/drm/drm_edid_load.c b/drivers/gpu/drm/drm_edid_load.c new file mode 100644 index 0000000..da9acba --- /dev/null +++ b/drivers/gpu/drm/drm_edid_load.c @@ -0,0 +1,250 @@ +/* + drm_edid_load.c: use a built-in EDID data set or load it via the firmware + interface + + Copyright (C) 2012 Carsten Emde + + This program is free software; you can redistribute it and/or + modify it under the terms of the GNU General Public License + as published by the Free Software Foundation; either version 2 + of the License, or (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. +*/ + +#include +#include +#include "drmP.h" +#include "drm_crtc.h" +#include "drm_crtc_helper.h" +#include "drm_edid.h" + +static char edid_firmware[PATH_MAX]; +module_param_string(edid_firmware, edid_firmware, sizeof(edid_firmware), 0644); +MODULE_PARM_DESC(edid_firmware, "Do not probe monitor, use specified EDID blob " + "from built-in data or /lib/firmware instead. "); + +#define GENERIC_EDIDS 4 +static char *generic_edid_name[GENERIC_EDIDS] = { + "edid/1024x768.bin", + "edid/1280x1024.bin", + "edid/1680x1050.bin", + "edid/1920x1080.bin", +}; + +static u8 generic_edid[GENERIC_EDIDS][128] = { + { + 0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00, + 0x31, 0xd8, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x05, 0x16, 0x01, 0x03, 0x6d, 0x23, 0x1a, 0x78, + 0xea, 0x5e, 0xc0, 0xa4, 0x59, 0x4a, 0x98, 0x25, + 0x20, 0x50, 0x54, 0x00, 0x08, 0x00, 0x61, 0x40, + 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, + 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x64, 0x19, + 0x00, 0x40, 0x41, 0x00, 0x26, 0x30, 0x08, 0x90, + 0x36, 0x00, 0x63, 0x0a, 0x11, 0x00, 0x00, 0x18, + 0x00, 0x00, 0x00, 0xff, 0x00, 0x4c, 0x69, 0x6e, + 0x75, 0x78, 0x20, 0x23, 0x30, 0x0a, 0x20, 0x20, + 0x20, 0x20, 0x00, 0x00, 0x00, 0xfd, 0x00, 0x3b, + 0x3d, 0x2f, 0x31, 0x07, 0x00, 0x0a, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x00, 0x00, 0x00, 0xfc, + 0x00, 0x4c, 0x69, 0x6e, 0x75, 0x78, 0x20, 0x58, + 0x47, 0x41, 0x0a, 0x20, 0x20, 0x20, 0x00, 0x55, + }, + { + 0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00, + 0x31, 0xd8, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x05, 0x16, 0x01, 0x03, 0x6d, 0x2c, 0x23, 0x78, + 0xea, 0x5e, 0xc0, 0xa4, 0x59, 0x4a, 0x98, 0x25, + 0x20, 0x50, 0x54, 0x00, 0x00, 0x00, 0x81, 0x80, + 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, + 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x30, 0x2a, + 0x00, 0x98, 0x51, 0x00, 0x2a, 0x40, 0x30, 0x70, + 0x13, 0x00, 0xbc, 0x63, 0x11, 0x00, 0x00, 0x1e, + 0x00, 0x00, 0x00, 0xff, 0x00, 0x4c, 0x69, 0x6e, + 0x75, 0x78, 0x20, 0x23, 0x30, 0x0a, 0x20, 0x20, + 0x20, 0x20, 0x00, 0x00, 0x00, 0xfd, 0x00, 0x3b, + 0x3d, 0x3e, 0x40, 0x0b, 0x00, 0x0a, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x00, 0x00, 0x00, 0xfc, + 0x00, 0x4c, 0x69, 0x6e, 0x75, 0x78, 0x20, 0x53, + 0x58, 0x47, 0x41, 0x0a, 0x20, 0x20, 0x00, 0xa0, + }, + { + 0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00, + 0x31, 0xd8, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x05, 0x16, 0x01, 0x03, 0x6d, 0x2b, 0x1b, 0x78, + 0xea, 0x5e, 0xc0, 0xa4, 0x59, 0x4a, 0x98, 0x25, + 0x20, 0x50, 0x54, 0x00, 0x00, 0x00, 0xb3, 0x00, + 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, + 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x21, 0x39, + 0x90, 0x30, 0x62, 0x1a, 0x27, 0x40, 0x68, 0xb0, + 0x36, 0x00, 0xb5, 0x11, 0x11, 0x00, 0x00, 0x1e, + 0x00, 0x00, 0x00, 0xff, 0x00, 0x4c, 0x69, 0x6e, + 0x75, 0x78, 0x20, 0x23, 0x30, 0x0a, 0x20, 0x20, + 0x20, 0x20, 0x00, 0x00, 0x00, 0xfd, 0x00, 0x3b, + 0x3d, 0x40, 0x42, 0x0f, 0x00, 0x0a, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x00, 0x00, 0x00, 0xfc, + 0x00, 0x4c, 0x69, 0x6e, 0x75, 0x78, 0x20, 0x57, + 0x53, 0x58, 0x47, 0x41, 0x0a, 0x20, 0x00, 0x26, + }, + { + 0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00, + 0x31, 0xd8, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x05, 0x16, 0x01, 0x03, 0x6d, 0x32, 0x1c, 0x78, + 0xea, 0x5e, 0xc0, 0xa4, 0x59, 0x4a, 0x98, 0x25, + 0x20, 0x50, 0x54, 0x00, 0x00, 0x00, 0xd1, 0xc0, + 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, + 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x02, 0x3a, + 0x80, 0x18, 0x71, 0x38, 0x2d, 0x40, 0x58, 0x2c, + 0x45, 0x00, 0xf4, 0x19, 0x11, 0x00, 0x00, 0x1e, + 0x00, 0x00, 0x00, 0xff, 0x00, 0x4c, 0x69, 0x6e, + 0x75, 0x78, 0x20, 0x23, 0x30, 0x0a, 0x20, 0x20, + 0x20, 0x20, 0x00, 0x00, 0x00, 0xfd, 0x00, 0x3b, + 0x3d, 0x42, 0x44, 0x0f, 0x00, 0x0a, 0x20, 0x20, + 0x20, 0x20, 0x20, 0x20, 0x00, 0x00, 0x00, 0xfc, + 0x00, 0x4c, 0x69, 0x6e, 0x75, 0x78, 0x20, 0x46, + 0x48, 0x44, 0x0a, 0x20, 0x20, 0x20, 0x00, 0x05, + }, +}; + +static int edid_load(struct drm_connector *connector, char *name, + char *connector_name) +{ + const struct firmware *fw; + struct platform_device *pdev; + u8 *fwdata = NULL, *edid; + int fwsize, expected; + int builtin = 0, err = 0; + int i, valid_extensions = 0; + + pdev = platform_device_register_simple(connector_name, -1, NULL, 0); + if (IS_ERR(pdev)) { + DRM_ERROR("Failed to register EDID firmware platform device " + "for connector \"%s\"\n", connector_name); + err = -EINVAL; + goto out; + } + + err = request_firmware(&fw, name, &pdev->dev); + platform_device_unregister(pdev); + + if (err) { + i = 0; + while (i < GENERIC_EDIDS && strcmp(name, generic_edid_name[i])) + i++; + if (i < GENERIC_EDIDS) { + err = 0; + builtin = 1; + fwdata = generic_edid[i]; + fwsize = sizeof(generic_edid[i]); + } + } + + if (err) { + DRM_ERROR("Requesting EDID firmware \"%s\" failed (err=%d)\n", + name, err); + goto out; + } + + if (fwdata == NULL) { + fwdata = (u8 *) fw->data; + fwsize = fw->size; + } + + expected = (fwdata[0x7e] + 1) * EDID_LENGTH; + if (expected != fwsize) { + DRM_ERROR("Size of EDID firmware \"%s\" is invalid " + "(expected %d, got %d)\n", name, expected, (int) fwsize); + err = -EINVAL; + goto relfw_out; + } + + edid = kmalloc(fwsize, GFP_KERNEL); + if (edid == NULL) { + err = -ENOMEM; + goto relfw_out; + } + memcpy(edid, fwdata, fwsize); + + if (!drm_edid_block_valid(edid)) { + DRM_ERROR("Base block of EDID firmware \"%s\" is invalid ", + name); + kfree(edid); + err = -EINVAL; + goto relfw_out; + } + + for (i = 1; i <= edid[0x7e]; i++) { + if (i != valid_extensions + 1) + memcpy(edid + (valid_extensions + 1) * EDID_LENGTH, + edid + i * EDID_LENGTH, EDID_LENGTH); + if (drm_edid_block_valid(edid + i * EDID_LENGTH)) + valid_extensions++; + } + + if (valid_extensions != edid[0x7e]) { + edid[EDID_LENGTH-1] += edid[0x7e] - valid_extensions; + DRM_INFO("Found %d valid extensions instead of %d in EDID data " + "\"%s\" for connector \"%s\"\n", valid_extensions, + edid[0x7e], name, connector_name); + edid[0x7e] = valid_extensions; + edid = krealloc(edid, (valid_extensions + 1) * EDID_LENGTH, + GFP_KERNEL); + if (edid == NULL) { + err = -ENOMEM; + goto relfw_out; + } + } + + connector->display_info.raw_edid = edid; + DRM_INFO("Got %s EDID base block and %d extension%s from " + "\"%s\" for connector \"%s\"\n", builtin ? "built-in" : + "external", valid_extensions, valid_extensions == 1 ? "" : "s", + name, connector_name); + +relfw_out: + release_firmware(fw); + +out: + return err; +} + +int drm_load_edid_firmware(struct drm_connector *connector) +{ + char *connector_name = drm_get_connector_name(connector); + char *edidname = edid_firmware, *last, *colon; + int ret = 0; + + if (*edidname == '\0') + return ret; + + colon = strchr(edidname, ':'); + if (colon != NULL) { + if (strncmp(connector_name, edidname, colon - edidname)) + return ret; + edidname = colon + 1; + if (*edidname == '\0') + return ret; + } + + last = edidname + strlen(edidname) - 1; + if (*last == '\n') + *last = '\0'; + + ret = edid_load(connector, edidname, connector_name); + if (ret) + return 0; + + drm_mode_connector_update_edid_property(connector, + (struct edid *) connector->display_info.raw_edid); + + return drm_add_edid_modes(connector, (struct edid *) + connector->display_info.raw_edid); +} diff --git a/drivers/gpu/drm/drm_edid_modes.h b/drivers/gpu/drm/drm_edid_modes.h index 5f20644..a91ffb1 100644 --- a/drivers/gpu/drm/drm_edid_modes.h +++ b/drivers/gpu/drm/drm_edid_modes.h @@ -378,3 +378,287 @@ static const struct { { 1920, 1440, 75, 0 }, }; static const int num_est3_modes = sizeof(est3_modes) / sizeof(est3_modes[0]); + +/* + * Probably taken from CEA-861 spec. + * This table is converted from xorg's hw/xfree86/modes/xf86EdidModes.c. + */ +static const struct drm_display_mode edid_cea_modes[] = { + /* 640x480@60Hz */ + { DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 25175, 640, 656, + 752, 800, 0, 480, 490, 492, 525, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 720x480@60Hz */ + { DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 27000, 720, 736, + 798, 858, 0, 480, 489, 495, 525, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 720x480@60Hz */ + { DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 27000, 720, 736, + 798, 858, 0, 480, 489, 495, 525, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 1280x720@60Hz */ + { DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 74250, 1280, 1390, + 1430, 1650, 0, 720, 725, 730, 750, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + /* 1920x1080i@60Hz */ + { DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2008, + 2052, 2200, 0, 1080, 1084, 1094, 1125, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC | + DRM_MODE_FLAG_INTERLACE) }, + /* 1440x480i@60Hz */ + { DRM_MODE("1440x480", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478, + 1602, 1716, 0, 480, 488, 494, 525, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC | + DRM_MODE_FLAG_INTERLACE) }, + /* 1440x480i@60Hz */ + { DRM_MODE("1440x480", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478, + 1602, 1716, 0, 480, 488, 494, 525, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC | + DRM_MODE_FLAG_INTERLACE) }, + /* 1440x240@60Hz */ + { DRM_MODE("1440x240", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478, + 1602, 1716, 0, 240, 244, 247, 262, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 1440x240@60Hz */ + { DRM_MODE("1440x240", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1478, + 1602, 1716, 0, 240, 244, 247, 262, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 2880x480i@60Hz */ + { DRM_MODE("2880x480", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2956, + 3204, 3432, 0, 480, 488, 494, 525, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC | + DRM_MODE_FLAG_INTERLACE) }, + /* 2880x480i@60Hz */ + { DRM_MODE("2880x480", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2956, + 3204, 3432, 0, 480, 488, 494, 525, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC | + DRM_MODE_FLAG_INTERLACE) }, + /* 2880x240@60Hz */ + { DRM_MODE("2880x240", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2956, + 3204, 3432, 0, 240, 244, 247, 262, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 2880x240@60Hz */ + { DRM_MODE("2880x240", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2956, + 3204, 3432, 0, 240, 244, 247, 262, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 1440x480@60Hz */ + { DRM_MODE("1440x480", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1472, + 1596, 1716, 0, 480, 489, 495, 525, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 1440x480@60Hz */ + { DRM_MODE("1440x480", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1472, + 1596, 1716, 0, 480, 489, 495, 525, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 1920x1080@60Hz */ + { DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 148500, 1920, 2008, + 2052, 2200, 0, 1080, 1084, 1089, 1125, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + /* 720x576@50Hz */ + { DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 27000, 720, 732, + 796, 864, 0, 576, 581, 586, 625, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 720x576@50Hz */ + { DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 27000, 720, 732, + 796, 864, 0, 576, 581, 586, 625, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 1280x720@50Hz */ + { DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 74250, 1280, 1720, + 1760, 1980, 0, 720, 725, 730, 750, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + /* 1920x1080i@50Hz */ + { DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2448, + 2492, 2640, 0, 1080, 1084, 1094, 1125, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC | + DRM_MODE_FLAG_INTERLACE) }, + /* 1440x576i@50Hz */ + { DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464, + 1590, 1728, 0, 576, 580, 586, 625, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC | + DRM_MODE_FLAG_INTERLACE) }, + /* 1440x576i@50Hz */ + { DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464, + 1590, 1728, 0, 576, 580, 586, 625, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC | + DRM_MODE_FLAG_INTERLACE) }, + /* 1440x288@50Hz */ + { DRM_MODE("1440x288", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464, + 1590, 1728, 0, 288, 290, 293, 312, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 1440x288@50Hz */ + { DRM_MODE("1440x288", DRM_MODE_TYPE_DRIVER, 27000, 1440, 1464, + 1590, 1728, 0, 288, 290, 293, 312, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 2880x576i@50Hz */ + { DRM_MODE("2880x576", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2928, + 3180, 3456, 0, 576, 580, 586, 625, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC | + DRM_MODE_FLAG_INTERLACE) }, + /* 2880x576i@50Hz */ + { DRM_MODE("2880x576", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2928, + 3180, 3456, 0, 576, 580, 586, 625, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC | + DRM_MODE_FLAG_INTERLACE) }, + /* 2880x288@50Hz */ + { DRM_MODE("2880x288", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2928, + 3180, 3456, 0, 288, 290, 293, 312, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 2880x288@50Hz */ + { DRM_MODE("2880x288", DRM_MODE_TYPE_DRIVER, 54000, 2880, 2928, + 3180, 3456, 0, 288, 290, 293, 312, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 1440x576@50Hz */ + { DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464, + 1592, 1728, 0, 576, 581, 586, 625, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 1440x576@50Hz */ + { DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464, + 1592, 1728, 0, 576, 581, 586, 625, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 1920x1080@50Hz */ + { DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 148500, 1920, 2448, + 2492, 2640, 0, 1080, 1084, 1089, 1125, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + /* 1920x1080@24Hz */ + { DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2558, + 2602, 2750, 0, 1080, 1084, 1089, 1125, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + /* 1920x1080@25Hz */ + { DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2448, + 2492, 2640, 0, 1080, 1084, 1089, 1125, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + /* 1920x1080@30Hz */ + { DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 74250, 1920, 2008, + 2052, 2200, 0, 1080, 1084, 1089, 1125, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + /* 2880x480@60Hz */ + { DRM_MODE("2880x480", DRM_MODE_TYPE_DRIVER, 108000, 2880, 2944, + 3192, 3432, 0, 480, 489, 495, 525, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 2880x480@60Hz */ + { DRM_MODE("2880x480", DRM_MODE_TYPE_DRIVER, 108000, 2880, 2944, + 3192, 3432, 0, 480, 489, 495, 525, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 2880x576@50Hz */ + { DRM_MODE("2880x576", DRM_MODE_TYPE_DRIVER, 108000, 2880, 2928, + 3184, 3456, 0, 576, 581, 586, 625, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 2880x576@50Hz */ + { DRM_MODE("2880x576", DRM_MODE_TYPE_DRIVER, 108000, 2880, 2928, + 3184, 3456, 0, 576, 581, 586, 625, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 1920x1080i@50Hz */ + { DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 72000, 1920, 1952, + 2120, 2304, 0, 1080, 1126, 1136, 1250, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC | + DRM_MODE_FLAG_INTERLACE) }, + /* 1920x1080i@100Hz */ + { DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 148500, 1920, 2448, + 2492, 2640, 0, 1080, 1084, 1094, 1125, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC | + DRM_MODE_FLAG_INTERLACE) }, + /* 1280x720@100Hz */ + { DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 148500, 1280, 1720, + 1760, 1980, 0, 720, 725, 730, 750, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + /* 720x576@100Hz */ + { DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 54000, 720, 732, + 796, 864, 0, 576, 581, 586, 625, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 720x576@100Hz */ + { DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 54000, 720, 732, + 796, 864, 0, 576, 581, 586, 625, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 1440x576i@100Hz */ + { DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464, + 1590, 1728, 0, 576, 580, 586, 625, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 1440x576i@100Hz */ + { DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1464, + 1590, 1728, 0, 576, 580, 586, 625, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 1920x1080i@120Hz */ + { DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 148500, 1920, 2008, + 2052, 2200, 0, 1080, 1084, 1094, 1125, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC | + DRM_MODE_FLAG_INTERLACE) }, + /* 1280x720@120Hz */ + { DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 148500, 1280, 1390, + 1430, 1650, 0, 720, 725, 730, 750, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + /* 720x480@120Hz */ + { DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 54000, 720, 736, + 798, 858, 0, 480, 489, 495, 525, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 720x480@120Hz */ + { DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 54000, 720, 736, + 798, 858, 0, 480, 489, 495, 525, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 1440x480i@120Hz */ + { DRM_MODE("1440x480", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1478, + 1602, 1716, 0, 480, 488, 494, 525, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC | + DRM_MODE_FLAG_INTERLACE) }, + /* 1440x480i@120Hz */ + { DRM_MODE("1440x480", DRM_MODE_TYPE_DRIVER, 54000, 1440, 1478, + 1602, 1716, 0, 480, 488, 494, 525, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC | + DRM_MODE_FLAG_INTERLACE) }, + /* 720x576@200Hz */ + { DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 108000, 720, 732, + 796, 864, 0, 576, 581, 586, 625, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 720x576@200Hz */ + { DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 108000, 720, 732, + 796, 864, 0, 576, 581, 586, 625, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 1440x576i@200Hz */ + { DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1464, + 1590, 1728, 0, 576, 580, 586, 625, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC | + DRM_MODE_FLAG_INTERLACE) }, + /* 1440x576i@200Hz */ + { DRM_MODE("1440x576", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1464, + 1590, 1728, 0, 576, 580, 586, 625, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC | + DRM_MODE_FLAG_INTERLACE) }, + /* 720x480@240Hz */ + { DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 108000, 720, 736, + 798, 858, 0, 480, 489, 495, 525, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 720x480@240Hz */ + { DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 108000, 720, 736, + 798, 858, 0, 480, 489, 495, 525, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) }, + /* 1440x480i@240 */ + { DRM_MODE("1440x480", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1478, + 1602, 1716, 0, 480, 488, 494, 525, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC | + DRM_MODE_FLAG_INTERLACE) }, + /* 1440x480i@240 */ + { DRM_MODE("1440x480", DRM_MODE_TYPE_DRIVER, 108000, 1440, 1478, + 1602, 1716, 0, 480, 488, 494, 525, 0, + DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC | + DRM_MODE_FLAG_INTERLACE) }, + /* 1280x720@24Hz */ + { DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 59400, 1280, 3040, + 3080, 3300, 0, 720, 725, 730, 750, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + /* 1280x720@25Hz */ + { DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 74250, 1280, 3700, + 3740, 3960, 0, 720, 725, 730, 750, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + /* 1280x720@30Hz */ + { DRM_MODE("1280x720", DRM_MODE_TYPE_DRIVER, 74250, 1280, 3040, + 3080, 3300, 0, 720, 725, 730, 750, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + /* 1920x1080@120Hz */ + { DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 297000, 1920, 2008, + 2052, 2200, 0, 1080, 1084, 1089, 1125, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + /* 1920x1080@100Hz */ + { DRM_MODE("1920x1080", DRM_MODE_TYPE_DRIVER, 297000, 1920, 2448, + 2492, 2640, 0, 1080, 1084, 1094, 1125, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, +}; +static const int drm_num_cea_modes = + sizeof (edid_cea_modes) / sizeof (edid_cea_modes[0]); diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c index dd58373..a0d6e89 100644 --- a/drivers/gpu/drm/drm_fb_helper.c +++ b/drivers/gpu/drm/drm_fb_helper.c @@ -255,6 +255,13 @@ bool drm_fb_helper_force_kernel_mode(void) int drm_fb_helper_panic(struct notifier_block *n, unsigned long ununsed, void *panic_str) { + /* + * It's a waste of time and effort to switch back to text console + * if the kernel should reboot before panic messages can be seen. + */ + if (panic_timeout < 0) + return 0; + printk(KERN_ERR "panic occurred, switching back to text console\n"); return drm_fb_helper_force_kernel_mode(); } @@ -299,91 +306,31 @@ static struct sysrq_key_op sysrq_drm_fb_helper_restore_op = { static struct sysrq_key_op sysrq_drm_fb_helper_restore_op = { }; #endif -static void drm_fb_helper_on(struct fb_info *info) -{ - struct drm_fb_helper *fb_helper = info->par; - struct drm_device *dev = fb_helper->dev; - struct drm_crtc *crtc; - struct drm_crtc_helper_funcs *crtc_funcs; - struct drm_connector *connector; - struct drm_encoder *encoder; - int i, j; - - /* - * For each CRTC in this fb, turn the crtc on then, - * find all associated encoders and turn them on. - */ - mutex_lock(&dev->mode_config.mutex); - for (i = 0; i < fb_helper->crtc_count; i++) { - crtc = fb_helper->crtc_info[i].mode_set.crtc; - crtc_funcs = crtc->helper_private; - - if (!crtc->enabled) - continue; - - crtc_funcs->dpms(crtc, DRM_MODE_DPMS_ON); - - /* Walk the connectors & encoders on this fb turning them on */ - for (j = 0; j < fb_helper->connector_count; j++) { - connector = fb_helper->connector_info[j]->connector; - connector->dpms = DRM_MODE_DPMS_ON; - drm_connector_property_set_value(connector, - dev->mode_config.dpms_property, - DRM_MODE_DPMS_ON); - } - /* Found a CRTC on this fb, now find encoders */ - list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) { - if (encoder->crtc == crtc) { - struct drm_encoder_helper_funcs *encoder_funcs; - - encoder_funcs = encoder->helper_private; - encoder_funcs->dpms(encoder, DRM_MODE_DPMS_ON); - } - } - } - mutex_unlock(&dev->mode_config.mutex); -} - -static void drm_fb_helper_off(struct fb_info *info, int dpms_mode) +static void drm_fb_helper_dpms(struct fb_info *info, int dpms_mode) { struct drm_fb_helper *fb_helper = info->par; struct drm_device *dev = fb_helper->dev; struct drm_crtc *crtc; - struct drm_crtc_helper_funcs *crtc_funcs; struct drm_connector *connector; - struct drm_encoder *encoder; int i, j; /* - * For each CRTC in this fb, find all associated encoders - * and turn them off, then turn off the CRTC. + * For each CRTC in this fb, turn the connectors on/off. */ mutex_lock(&dev->mode_config.mutex); for (i = 0; i < fb_helper->crtc_count; i++) { crtc = fb_helper->crtc_info[i].mode_set.crtc; - crtc_funcs = crtc->helper_private; if (!crtc->enabled) continue; - /* Walk the connectors on this fb and mark them off */ + /* Walk the connectors & encoders on this fb turning them on/off */ for (j = 0; j < fb_helper->connector_count; j++) { connector = fb_helper->connector_info[j]->connector; - connector->dpms = dpms_mode; + drm_helper_connector_dpms(connector, dpms_mode); drm_connector_property_set_value(connector, - dev->mode_config.dpms_property, - dpms_mode); + dev->mode_config.dpms_property, dpms_mode); } - /* Found a CRTC on this fb, now find encoders */ - list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) { - if (encoder->crtc == crtc) { - struct drm_encoder_helper_funcs *encoder_funcs; - - encoder_funcs = encoder->helper_private; - encoder_funcs->dpms(encoder, dpms_mode); - } - } - crtc_funcs->dpms(crtc, DRM_MODE_DPMS_OFF); } mutex_unlock(&dev->mode_config.mutex); } @@ -393,23 +340,23 @@ int drm_fb_helper_blank(int blank, struct fb_info *info) switch (blank) { /* Display: On; HSync: On, VSync: On */ case FB_BLANK_UNBLANK: - drm_fb_helper_on(info); + drm_fb_helper_dpms(info, DRM_MODE_DPMS_ON); break; /* Display: Off; HSync: On, VSync: On */ case FB_BLANK_NORMAL: - drm_fb_helper_off(info, DRM_MODE_DPMS_STANDBY); + drm_fb_helper_dpms(info, DRM_MODE_DPMS_STANDBY); break; /* Display: Off; HSync: Off, VSync: On */ case FB_BLANK_HSYNC_SUSPEND: - drm_fb_helper_off(info, DRM_MODE_DPMS_STANDBY); + drm_fb_helper_dpms(info, DRM_MODE_DPMS_STANDBY); break; /* Display: Off; HSync: On, VSync: Off */ case FB_BLANK_VSYNC_SUSPEND: - drm_fb_helper_off(info, DRM_MODE_DPMS_SUSPEND); + drm_fb_helper_dpms(info, DRM_MODE_DPMS_SUSPEND); break; /* Display: Off; HSync: Off, VSync: Off */ case FB_BLANK_POWERDOWN: - drm_fb_helper_off(info, DRM_MODE_DPMS_OFF); + drm_fb_helper_dpms(info, DRM_MODE_DPMS_OFF); break; } return 0; @@ -423,8 +370,11 @@ static void drm_fb_helper_crtc_free(struct drm_fb_helper *helper) for (i = 0; i < helper->connector_count; i++) kfree(helper->connector_info[i]); kfree(helper->connector_info); - for (i = 0; i < helper->crtc_count; i++) + for (i = 0; i < helper->crtc_count; i++) { kfree(helper->crtc_info[i].mode_set.connectors); + if (helper->crtc_info[i].mode_set.mode) + drm_mode_destroy(helper->dev, helper->crtc_info[i].mode_set.mode); + } kfree(helper->crtc_info); } @@ -467,11 +417,10 @@ int drm_fb_helper_init(struct drm_device *dev, i = 0; list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { - fb_helper->crtc_info[i].crtc_id = crtc->base.id; fb_helper->crtc_info[i].mode_set.crtc = crtc; i++; } - fb_helper->conn_limit = max_conn_count; + return 0; out_free: drm_fb_helper_crtc_free(fb_helper); diff --git a/drivers/gpu/drm/drm_fops.c b/drivers/gpu/drm/drm_fops.c index 020b103..b90abff 100644 --- a/drivers/gpu/drm/drm_fops.c +++ b/drivers/gpu/drm/drm_fops.c @@ -133,6 +133,9 @@ int drm_open(struct inode *inode, struct file *filp) if (!(dev = minor->dev)) return -ENODEV; + if (drm_device_is_unplugged(dev)) + return -ENODEV; + retcode = drm_open_helper(inode, filp, dev); if (!retcode) { atomic_inc(&dev->counts[_DRM_STAT_OPENS]); @@ -184,8 +187,11 @@ int drm_stub_open(struct inode *inode, struct file *filp) if (!(dev = minor->dev)) goto out; + if (drm_device_is_unplugged(dev)) + goto out; + old_fops = filp->f_op; - filp->f_op = fops_get(&dev->driver->fops); + filp->f_op = fops_get(dev->driver->fops); if (filp->f_op == NULL) { filp->f_op = old_fops; goto out; @@ -268,6 +274,9 @@ static int drm_open_helper(struct inode *inode, struct file *filp, if (dev->driver->driver_features & DRIVER_GEM) drm_gem_open(dev, priv); + if (drm_core_check_feature(dev, DRIVER_PRIME)) + drm_prime_init_file_private(&priv->prime); + if (dev->driver->open) { ret = dev->driver->open(dev, priv); if (ret < 0) @@ -501,12 +510,12 @@ int drm_release(struct inode *inode, struct file *filp) drm_events_release(file_priv); - if (dev->driver->driver_features & DRIVER_GEM) - drm_gem_release(dev, file_priv); - if (dev->driver->driver_features & DRIVER_MODESET) drm_fb_release(file_priv); + if (dev->driver->driver_features & DRIVER_GEM) + drm_gem_release(dev, file_priv); + mutex_lock(&dev->ctxlist_mutex); if (!list_empty(&dev->ctxlist)) { struct drm_ctx_list *pos, *n; @@ -568,6 +577,10 @@ int drm_release(struct inode *inode, struct file *filp) if (dev->driver->postclose) dev->driver->postclose(dev, file_priv); + + if (drm_core_check_feature(dev, DRIVER_PRIME)) + drm_prime_destroy_file_private(&file_priv->prime); + kfree(file_priv); /* ======================================================== @@ -582,6 +595,8 @@ int drm_release(struct inode *inode, struct file *filp) retcode = -EBUSY; } else retcode = drm_lastclose(dev); + if (drm_device_is_unplugged(dev)) + drm_put_dev(dev); } mutex_unlock(&drm_global_mutex); diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c index 396e60c..83114b5 100644 --- a/drivers/gpu/drm/drm_gem.c +++ b/drivers/gpu/drm/drm_gem.c @@ -35,6 +35,7 @@ #include #include #include +#include #include "drmP.h" /** @file drm_gem.c @@ -140,7 +141,7 @@ int drm_gem_object_init(struct drm_device *dev, obj->dev = dev; obj->filp = shmem_file_setup("drm mm object", size, VM_NORESERVE); if (IS_ERR(obj->filp)) - return -ENOMEM; + return PTR_ERR(obj->filp); kref_init(&obj->refcount); atomic_set(&obj->handle_count, 0); @@ -232,6 +233,10 @@ drm_gem_handle_delete(struct drm_file *filp, u32 handle) idr_remove(&filp->object_idr, handle); spin_unlock(&filp->table_lock); + if (obj->import_attach) + drm_prime_remove_imported_buf_handle(&filp->prime, + obj->import_attach->dmabuf); + if (dev->driver->gem_close_object) dev->driver->gem_close_object(obj, filp); drm_gem_object_handle_unreference_unlocked(obj); @@ -527,6 +532,10 @@ drm_gem_object_release_handle(int id, void *ptr, void *data) struct drm_gem_object *obj = ptr; struct drm_device *dev = obj->dev; + if (obj->import_attach) + drm_prime_remove_imported_buf_handle(&file_priv->prime, + obj->import_attach->dmabuf); + if (dev->driver->gem_close_object) dev->driver->gem_close_object(obj, file_priv); @@ -661,6 +670,9 @@ int drm_gem_mmap(struct file *filp, struct vm_area_struct *vma) struct drm_hash_item *hash; int ret = 0; + if (drm_device_is_unplugged(dev)) + return -ENODEV; + mutex_lock(&dev->struct_mutex); if (drm_ht_find_item(&mm->offset_hash, vma->vm_pgoff, &hash)) { @@ -700,7 +712,6 @@ int drm_gem_mmap(struct file *filp, struct vm_area_struct *vma) */ drm_gem_object_reference(obj); - vma->vm_file = filp; /* Needed for drm_vm_open() */ drm_vm_open_locked(vma); out_unlock: diff --git a/drivers/gpu/drm/drm_ioc32.c b/drivers/gpu/drm/drm_ioc32.c index ddd70db..637fcc3 100644 --- a/drivers/gpu/drm/drm_ioc32.c +++ b/drivers/gpu/drm/drm_ioc32.c @@ -315,7 +315,8 @@ static int compat_drm_getclient(struct file *file, unsigned int cmd, if (err) return err; - if (__get_user(c32.auth, &client->auth) + if (__get_user(c32.idx, &client->idx) + || __get_user(c32.auth, &client->auth) || __get_user(c32.pid, &client->pid) || __get_user(c32.uid, &client->uid) || __get_user(c32.magic, &client->magic) diff --git a/drivers/gpu/drm/drm_ioctl.c b/drivers/gpu/drm/drm_ioctl.c index 904d7e9..cf85155 100644 --- a/drivers/gpu/drm/drm_ioctl.c +++ b/drivers/gpu/drm/drm_ioctl.c @@ -37,6 +37,7 @@ #include "drm_core.h" #include "linux/pci.h" +#include "linux/export.h" /** * Get the bus id. @@ -158,14 +159,11 @@ int drm_getmap(struct drm_device *dev, void *data, int i; idx = map->offset; - - mutex_lock(&dev->struct_mutex); - if (idx < 0) { - mutex_unlock(&dev->struct_mutex); + if (idx < 0) return -EINVAL; - } i = 0; + mutex_lock(&dev->struct_mutex); list_for_each(list, &dev->maplist) { if (i == idx) { r_list = list_entry(list, struct drm_map_list, head); @@ -211,9 +209,9 @@ int drm_getclient(struct drm_device *dev, void *data, int i; idx = client->idx; - mutex_lock(&dev->struct_mutex); - i = 0; + + mutex_lock(&dev->struct_mutex); list_for_each_entry(pt, &dev->filelist, lhead) { if (i++ >= idx) { client->auth = pt->authenticated; @@ -249,8 +247,6 @@ int drm_getstats(struct drm_device *dev, void *data, memset(stats, 0, sizeof(*stats)); - mutex_lock(&dev->struct_mutex); - for (i = 0; i < dev->counters; i++) { if (dev->types[i] == _DRM_STAT_LOCK) stats->data[i].value = @@ -262,8 +258,6 @@ int drm_getstats(struct drm_device *dev, void *data, stats->count = dev->counters; - mutex_unlock(&dev->struct_mutex); - return 0; } @@ -283,6 +277,12 @@ int drm_getcap(struct drm_device *dev, void *data, struct drm_file *file_priv) case DRM_CAP_VBLANK_HIGH_CRTC: req->value = 1; break; + case DRM_CAP_DUMB_PREFERRED_DEPTH: + req->value = dev->mode_config.preferred_depth; + break; + case DRM_CAP_DUMB_PREFER_SHADOW: + req->value = dev->mode_config.prefer_shadow; + break; default: return -EINVAL; } @@ -353,3 +353,4 @@ int drm_noop(struct drm_device *dev, void *data, DRM_DEBUG("\n"); return 0; } +EXPORT_SYMBOL(drm_noop); diff --git a/drivers/gpu/drm/drm_irq.c b/drivers/gpu/drm/drm_irq.c index 44a5d0a..c869436 100644 --- a/drivers/gpu/drm/drm_irq.c +++ b/drivers/gpu/drm/drm_irq.c @@ -305,7 +305,7 @@ static void drm_irq_vgaarb_nokms(void *cookie, bool state) * \param dev DRM device. * * Initializes the IRQ related data. Installs the handler, calling the driver - * \c drm_driver_irq_preinstall() and \c drm_driver_irq_postinstall() functions + * \c irq_preinstall() and \c irq_postinstall() functions * before and after the installation. */ int drm_irq_install(struct drm_device *dev) @@ -385,7 +385,7 @@ EXPORT_SYMBOL(drm_irq_install); * * \param dev DRM device. * - * Calls the driver's \c drm_driver_irq_uninstall() function, and stops the irq. + * Calls the driver's \c irq_uninstall() function, and stops the irq. */ int drm_irq_uninstall(struct drm_device *dev) { diff --git a/drivers/gpu/drm/drm_lock.c b/drivers/gpu/drm/drm_lock.c index 632ae24..c79c713 100644 --- a/drivers/gpu/drm/drm_lock.c +++ b/drivers/gpu/drm/drm_lock.c @@ -33,6 +33,7 @@ * OTHER DEALINGS IN THE SOFTWARE. */ +#include #include "drmP.h" static int drm_notifier(void *priv); @@ -345,6 +346,7 @@ void drm_idlelock_take(struct drm_lock_data *lock_data) } spin_unlock_bh(&lock_data->spinlock); } +EXPORT_SYMBOL(drm_idlelock_take); void drm_idlelock_release(struct drm_lock_data *lock_data) { @@ -364,6 +366,7 @@ void drm_idlelock_release(struct drm_lock_data *lock_data) } spin_unlock_bh(&lock_data->spinlock); } +EXPORT_SYMBOL(drm_idlelock_release); int drm_i_have_hw_lock(struct drm_device *dev, struct drm_file *file_priv) { diff --git a/drivers/gpu/drm/drm_memory.c b/drivers/gpu/drm/drm_memory.c index c8b6b66..c86a0f1 100644 --- a/drivers/gpu/drm/drm_memory.c +++ b/drivers/gpu/drm/drm_memory.c @@ -37,25 +37,6 @@ #include #include "drmP.h" -/** - * Called when "/proc/dri/%dev%/mem" is read. - * - * \param buf output buffer. - * \param start start of output data. - * \param offset requested start offset. - * \param len requested number of bytes. - * \param eof whether there is no more data to return. - * \param data private data. - * \return number of written bytes. - * - * No-op. - */ -int drm_mem_info(char *buf, char **start, off_t offset, - int len, int *eof, void *data) -{ - return 0; -} - #if __OS_HAS_AGP static void *agp_remap(unsigned long offset, unsigned long size, struct drm_device * dev) diff --git a/drivers/gpu/drm/drm_modes.c b/drivers/gpu/drm/drm_modes.c index fb8e46b..b7adb4a 100644 --- a/drivers/gpu/drm/drm_modes.c +++ b/drivers/gpu/drm/drm_modes.c @@ -686,8 +686,6 @@ void drm_mode_set_crtcinfo(struct drm_display_mode *p, int adjust_flags) p->crtc_vsync_end /= 2; p->crtc_vtotal /= 2; } - - p->crtc_vtotal |= 1; } if (p->flags & DRM_MODE_FLAG_DBLSCAN) { @@ -716,6 +714,27 @@ EXPORT_SYMBOL(drm_mode_set_crtcinfo); /** + * drm_mode_copy - copy the mode + * @dst: mode to overwrite + * @src: mode to copy + * + * LOCKING: + * None. + * + * Copy an existing mode into another mode, preserving the object id + * of the destination mode. + */ +void drm_mode_copy(struct drm_display_mode *dst, const struct drm_display_mode *src) +{ + int id = dst->base.id; + + *dst = *src; + dst->base.id = id; + INIT_LIST_HEAD(&dst->head); +} +EXPORT_SYMBOL(drm_mode_copy); + +/** * drm_mode_duplicate - allocate and duplicate an existing mode * @m: mode to duplicate * @@ -729,16 +748,13 @@ struct drm_display_mode *drm_mode_duplicate(struct drm_device *dev, const struct drm_display_mode *mode) { struct drm_display_mode *nmode; - int new_id; nmode = drm_mode_create(dev); if (!nmode) return NULL; - new_id = nmode->base.id; - *nmode = *mode; - nmode->base.id = new_id; - INIT_LIST_HEAD(&nmode->head); + drm_mode_copy(nmode, mode); + return nmode; } EXPORT_SYMBOL(drm_mode_duplicate); diff --git a/drivers/gpu/drm/drm_pci.c b/drivers/gpu/drm/drm_pci.c index d4d10b7..13f3d93 100644 --- a/drivers/gpu/drm/drm_pci.c +++ b/drivers/gpu/drm/drm_pci.c @@ -324,8 +324,6 @@ int drm_get_pci_dev(struct pci_dev *pdev, const struct pci_device_id *ent, if (ret) goto err_g1; - pci_set_master(pdev); - dev->pdev = pdev; dev->dev = &pdev->dev; diff --git a/drivers/gpu/drm/drm_platform.c b/drivers/gpu/drm/drm_platform.c index ae9db5e..82431dc 100644 --- a/drivers/gpu/drm/drm_platform.c +++ b/drivers/gpu/drm/drm_platform.c @@ -122,7 +122,7 @@ static const char *drm_platform_get_name(struct drm_device *dev) static int drm_platform_set_busid(struct drm_device *dev, struct drm_master *master) { - int len, ret; + int len, ret, id; master->unique_len = 13 + strlen(dev->platformdev->name); master->unique_size = master->unique_len; @@ -131,8 +131,16 @@ static int drm_platform_set_busid(struct drm_device *dev, struct drm_master *mas if (master->unique == NULL) return -ENOMEM; + id = dev->platformdev->id; + + /* if only a single instance of the platform device, id will be + * set to -1.. use 0 instead to avoid a funny looking bus-id: + */ + if (id == -1) + id = 0; + len = snprintf(master->unique, master->unique_len, - "platform:%s:%02d", dev->platformdev->name, dev->platformdev->id); + "platform:%s:%02d", dev->platformdev->name, id); if (len > master->unique_len) { DRM_ERROR("Unique buffer overflowed\n"); diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c new file mode 100644 index 0000000..1bdf2b5 --- /dev/null +++ b/drivers/gpu/drm/drm_prime.c @@ -0,0 +1,304 @@ +/* + * Copyright © 2012 Red Hat + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + * + * Authors: + * Dave Airlie + * Rob Clark + * + */ + +#include +#include +#include "drmP.h" + +/* + * DMA-BUF/GEM Object references and lifetime overview: + * + * On the export the dma_buf holds a reference to the exporting GEM + * object. It takes this reference in handle_to_fd_ioctl, when it + * first calls .prime_export and stores the exporting GEM object in + * the dma_buf priv. This reference is released when the dma_buf + * object goes away in the driver .release function. + * + * On the import the importing GEM object holds a reference to the + * dma_buf (which in turn holds a ref to the exporting GEM object). + * It takes that reference in the fd_to_handle ioctl. + * It calls dma_buf_get, creates an attachment to it and stores the + * attachment in the GEM object. When this attachment is destroyed + * when the imported object is destroyed, we remove the attachment + * and drop the reference to the dma_buf. + * + * Thus the chain of references always flows in one direction + * (avoiding loops): importing_gem -> dmabuf -> exporting_gem + * + * Self-importing: if userspace is using PRIME as a replacement for flink + * then it will get a fd->handle request for a GEM object that it created. + * Drivers should detect this situation and return back the gem object + * from the dma-buf private. + */ + +struct drm_prime_member { + struct list_head entry; + struct dma_buf *dma_buf; + uint32_t handle; +}; + +int drm_gem_prime_handle_to_fd(struct drm_device *dev, + struct drm_file *file_priv, uint32_t handle, uint32_t flags, + int *prime_fd) +{ + struct drm_gem_object *obj; + void *buf; + + obj = drm_gem_object_lookup(dev, file_priv, handle); + if (!obj) + return -ENOENT; + + mutex_lock(&file_priv->prime.lock); + /* re-export the original imported object */ + if (obj->import_attach) { + get_dma_buf(obj->import_attach->dmabuf); + *prime_fd = dma_buf_fd(obj->import_attach->dmabuf, flags); + drm_gem_object_unreference_unlocked(obj); + mutex_unlock(&file_priv->prime.lock); + return 0; + } + + if (obj->export_dma_buf) { + get_dma_buf(obj->export_dma_buf); + *prime_fd = dma_buf_fd(obj->export_dma_buf, flags); + drm_gem_object_unreference_unlocked(obj); + } else { + buf = dev->driver->gem_prime_export(dev, obj, flags); + if (IS_ERR(buf)) { + /* normally the created dma-buf takes ownership of the ref, + * but if that fails then drop the ref + */ + drm_gem_object_unreference_unlocked(obj); + mutex_unlock(&file_priv->prime.lock); + return PTR_ERR(buf); + } + obj->export_dma_buf = buf; + *prime_fd = dma_buf_fd(buf, flags); + } + mutex_unlock(&file_priv->prime.lock); + return 0; +} +EXPORT_SYMBOL(drm_gem_prime_handle_to_fd); + +int drm_gem_prime_fd_to_handle(struct drm_device *dev, + struct drm_file *file_priv, int prime_fd, uint32_t *handle) +{ + struct dma_buf *dma_buf; + struct drm_gem_object *obj; + int ret; + + dma_buf = dma_buf_get(prime_fd); + if (IS_ERR(dma_buf)) + return PTR_ERR(dma_buf); + + mutex_lock(&file_priv->prime.lock); + + ret = drm_prime_lookup_imported_buf_handle(&file_priv->prime, + dma_buf, handle); + if (!ret) { + ret = 0; + goto out_put; + } + + /* never seen this one, need to import */ + obj = dev->driver->gem_prime_import(dev, dma_buf); + if (IS_ERR(obj)) { + ret = PTR_ERR(obj); + goto out_put; + } + + ret = drm_gem_handle_create(file_priv, obj, handle); + drm_gem_object_unreference_unlocked(obj); + if (ret) + goto out_put; + + ret = drm_prime_add_imported_buf_handle(&file_priv->prime, + dma_buf, *handle); + if (ret) + goto fail; + + mutex_unlock(&file_priv->prime.lock); + return 0; + +fail: + /* hmm, if driver attached, we are relying on the free-object path + * to detach.. which seems ok.. + */ + drm_gem_object_handle_unreference_unlocked(obj); +out_put: + dma_buf_put(dma_buf); + mutex_unlock(&file_priv->prime.lock); + return ret; +} +EXPORT_SYMBOL(drm_gem_prime_fd_to_handle); + +int drm_prime_handle_to_fd_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + struct drm_prime_handle *args = data; + uint32_t flags; + + if (!drm_core_check_feature(dev, DRIVER_PRIME)) + return -EINVAL; + + if (!dev->driver->prime_handle_to_fd) + return -ENOSYS; + + /* check flags are valid */ + if (args->flags & ~DRM_CLOEXEC) + return -EINVAL; + + /* we only want to pass DRM_CLOEXEC which is == O_CLOEXEC */ + flags = args->flags & DRM_CLOEXEC; + + return dev->driver->prime_handle_to_fd(dev, file_priv, + args->handle, flags, &args->fd); +} + +int drm_prime_fd_to_handle_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + struct drm_prime_handle *args = data; + + if (!drm_core_check_feature(dev, DRIVER_PRIME)) + return -EINVAL; + + if (!dev->driver->prime_fd_to_handle) + return -ENOSYS; + + return dev->driver->prime_fd_to_handle(dev, file_priv, + args->fd, &args->handle); +} + +/* + * drm_prime_pages_to_sg + * + * this helper creates an sg table object from a set of pages + * the driver is responsible for mapping the pages into the + * importers address space + */ +struct sg_table *drm_prime_pages_to_sg(struct page **pages, int nr_pages) +{ + struct sg_table *sg = NULL; + struct scatterlist *iter; + int i; + int ret; + + sg = kmalloc(sizeof(struct sg_table), GFP_KERNEL); + if (!sg) + goto out; + + ret = sg_alloc_table(sg, nr_pages, GFP_KERNEL); + if (ret) + goto out; + + for_each_sg(sg->sgl, iter, nr_pages, i) + sg_set_page(iter, pages[i], PAGE_SIZE, 0); + + return sg; +out: + kfree(sg); + return NULL; +} +EXPORT_SYMBOL(drm_prime_pages_to_sg); + +/* helper function to cleanup a GEM/prime object */ +void drm_prime_gem_destroy(struct drm_gem_object *obj, struct sg_table *sg) +{ + struct dma_buf_attachment *attach; + struct dma_buf *dma_buf; + attach = obj->import_attach; + if (sg) + dma_buf_unmap_attachment(attach, sg, DMA_BIDIRECTIONAL); + dma_buf = attach->dmabuf; + dma_buf_detach(attach->dmabuf, attach); + /* remove the reference */ + dma_buf_put(dma_buf); +} +EXPORT_SYMBOL(drm_prime_gem_destroy); + +void drm_prime_init_file_private(struct drm_prime_file_private *prime_fpriv) +{ + INIT_LIST_HEAD(&prime_fpriv->head); + mutex_init(&prime_fpriv->lock); +} +EXPORT_SYMBOL(drm_prime_init_file_private); + +void drm_prime_destroy_file_private(struct drm_prime_file_private *prime_fpriv) +{ + struct drm_prime_member *member, *safe; + list_for_each_entry_safe(member, safe, &prime_fpriv->head, entry) { + list_del(&member->entry); + kfree(member); + } +} +EXPORT_SYMBOL(drm_prime_destroy_file_private); + +int drm_prime_add_imported_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf, uint32_t handle) +{ + struct drm_prime_member *member; + + member = kmalloc(sizeof(*member), GFP_KERNEL); + if (!member) + return -ENOMEM; + + member->dma_buf = dma_buf; + member->handle = handle; + list_add(&member->entry, &prime_fpriv->head); + return 0; +} +EXPORT_SYMBOL(drm_prime_add_imported_buf_handle); + +int drm_prime_lookup_imported_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf, uint32_t *handle) +{ + struct drm_prime_member *member; + + list_for_each_entry(member, &prime_fpriv->head, entry) { + if (member->dma_buf == dma_buf) { + *handle = member->handle; + return 0; + } + } + return -ENOENT; +} +EXPORT_SYMBOL(drm_prime_lookup_imported_buf_handle); + +void drm_prime_remove_imported_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf) +{ + struct drm_prime_member *member, *safe; + + mutex_lock(&prime_fpriv->lock); + list_for_each_entry_safe(member, safe, &prime_fpriv->head, entry) { + if (member->dma_buf == dma_buf) { + list_del(&member->entry); + kfree(member); + } + } + mutex_unlock(&prime_fpriv->lock); +} +EXPORT_SYMBOL(drm_prime_remove_imported_buf_handle); diff --git a/drivers/gpu/drm/drm_sman.c b/drivers/gpu/drm/drm_sman.c deleted file mode 100644 index cebce45..0000000 --- a/drivers/gpu/drm/drm_sman.c +++ /dev/null @@ -1,351 +0,0 @@ -/************************************************************************** - * - * Copyright 2006 Tungsten Graphics, Inc., Bismarck., ND., USA. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, - * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR - * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE - * USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * - **************************************************************************/ -/* - * Simple memory manager interface that keeps track on allocate regions on a - * per "owner" basis. All regions associated with an "owner" can be released - * with a simple call. Typically if the "owner" exists. The owner is any - * "unsigned long" identifier. Can typically be a pointer to a file private - * struct or a context identifier. - * - * Authors: - * Thomas Hellström - */ - -#include -#include "drm_sman.h" - -struct drm_owner_item { - struct drm_hash_item owner_hash; - struct list_head sman_list; - struct list_head mem_blocks; -}; - -void drm_sman_takedown(struct drm_sman * sman) -{ - drm_ht_remove(&sman->user_hash_tab); - drm_ht_remove(&sman->owner_hash_tab); - kfree(sman->mm); -} - -EXPORT_SYMBOL(drm_sman_takedown); - -int -drm_sman_init(struct drm_sman * sman, unsigned int num_managers, - unsigned int user_order, unsigned int owner_order) -{ - int ret = 0; - - sman->mm = kcalloc(num_managers, sizeof(*sman->mm), GFP_KERNEL); - if (!sman->mm) { - ret = -ENOMEM; - goto out; - } - sman->num_managers = num_managers; - INIT_LIST_HEAD(&sman->owner_items); - ret = drm_ht_create(&sman->owner_hash_tab, owner_order); - if (ret) - goto out1; - ret = drm_ht_create(&sman->user_hash_tab, user_order); - if (!ret) - goto out; - - drm_ht_remove(&sman->owner_hash_tab); -out1: - kfree(sman->mm); -out: - return ret; -} - -EXPORT_SYMBOL(drm_sman_init); - -static void *drm_sman_mm_allocate(void *private, unsigned long size, - unsigned alignment) -{ - struct drm_mm *mm = (struct drm_mm *) private; - struct drm_mm_node *tmp; - - tmp = drm_mm_search_free(mm, size, alignment, 1); - if (!tmp) { - return NULL; - } - tmp = drm_mm_get_block(tmp, size, alignment); - return tmp; -} - -static void drm_sman_mm_free(void *private, void *ref) -{ - struct drm_mm_node *node = (struct drm_mm_node *) ref; - - drm_mm_put_block(node); -} - -static void drm_sman_mm_destroy(void *private) -{ - struct drm_mm *mm = (struct drm_mm *) private; - drm_mm_takedown(mm); - kfree(mm); -} - -static unsigned long drm_sman_mm_offset(void *private, void *ref) -{ - struct drm_mm_node *node = (struct drm_mm_node *) ref; - return node->start; -} - -int -drm_sman_set_range(struct drm_sman * sman, unsigned int manager, - unsigned long start, unsigned long size) -{ - struct drm_sman_mm *sman_mm; - struct drm_mm *mm; - int ret; - - BUG_ON(manager >= sman->num_managers); - - sman_mm = &sman->mm[manager]; - mm = kzalloc(sizeof(*mm), GFP_KERNEL); - if (!mm) { - return -ENOMEM; - } - sman_mm->private = mm; - ret = drm_mm_init(mm, start, size); - - if (ret) { - kfree(mm); - return ret; - } - - sman_mm->allocate = drm_sman_mm_allocate; - sman_mm->free = drm_sman_mm_free; - sman_mm->destroy = drm_sman_mm_destroy; - sman_mm->offset = drm_sman_mm_offset; - - return 0; -} - -EXPORT_SYMBOL(drm_sman_set_range); - -int -drm_sman_set_manager(struct drm_sman * sman, unsigned int manager, - struct drm_sman_mm * allocator) -{ - BUG_ON(manager >= sman->num_managers); - sman->mm[manager] = *allocator; - - return 0; -} -EXPORT_SYMBOL(drm_sman_set_manager); - -static struct drm_owner_item *drm_sman_get_owner_item(struct drm_sman * sman, - unsigned long owner) -{ - int ret; - struct drm_hash_item *owner_hash_item; - struct drm_owner_item *owner_item; - - ret = drm_ht_find_item(&sman->owner_hash_tab, owner, &owner_hash_item); - if (!ret) { - return drm_hash_entry(owner_hash_item, struct drm_owner_item, - owner_hash); - } - - owner_item = kzalloc(sizeof(*owner_item), GFP_KERNEL); - if (!owner_item) - goto out; - - INIT_LIST_HEAD(&owner_item->mem_blocks); - owner_item->owner_hash.key = owner; - if (drm_ht_insert_item(&sman->owner_hash_tab, &owner_item->owner_hash)) - goto out1; - - list_add_tail(&owner_item->sman_list, &sman->owner_items); - return owner_item; - -out1: - kfree(owner_item); -out: - return NULL; -} - -struct drm_memblock_item *drm_sman_alloc(struct drm_sman *sman, unsigned int manager, - unsigned long size, unsigned alignment, - unsigned long owner) -{ - void *tmp; - struct drm_sman_mm *sman_mm; - struct drm_owner_item *owner_item; - struct drm_memblock_item *memblock; - - BUG_ON(manager >= sman->num_managers); - - sman_mm = &sman->mm[manager]; - tmp = sman_mm->allocate(sman_mm->private, size, alignment); - - if (!tmp) { - return NULL; - } - - memblock = kzalloc(sizeof(*memblock), GFP_KERNEL); - - if (!memblock) - goto out; - - memblock->mm_info = tmp; - memblock->mm = sman_mm; - memblock->sman = sman; - - if (drm_ht_just_insert_please - (&sman->user_hash_tab, &memblock->user_hash, - (unsigned long)memblock, 32, 0, 0)) - goto out1; - - owner_item = drm_sman_get_owner_item(sman, owner); - if (!owner_item) - goto out2; - - list_add_tail(&memblock->owner_list, &owner_item->mem_blocks); - - return memblock; - -out2: - drm_ht_remove_item(&sman->user_hash_tab, &memblock->user_hash); -out1: - kfree(memblock); -out: - sman_mm->free(sman_mm->private, tmp); - - return NULL; -} - -EXPORT_SYMBOL(drm_sman_alloc); - -static void drm_sman_free(struct drm_memblock_item *item) -{ - struct drm_sman *sman = item->sman; - - list_del(&item->owner_list); - drm_ht_remove_item(&sman->user_hash_tab, &item->user_hash); - item->mm->free(item->mm->private, item->mm_info); - kfree(item); -} - -int drm_sman_free_key(struct drm_sman *sman, unsigned int key) -{ - struct drm_hash_item *hash_item; - struct drm_memblock_item *memblock_item; - - if (drm_ht_find_item(&sman->user_hash_tab, key, &hash_item)) - return -EINVAL; - - memblock_item = drm_hash_entry(hash_item, struct drm_memblock_item, - user_hash); - drm_sman_free(memblock_item); - return 0; -} - -EXPORT_SYMBOL(drm_sman_free_key); - -static void drm_sman_remove_owner(struct drm_sman *sman, - struct drm_owner_item *owner_item) -{ - list_del(&owner_item->sman_list); - drm_ht_remove_item(&sman->owner_hash_tab, &owner_item->owner_hash); - kfree(owner_item); -} - -int drm_sman_owner_clean(struct drm_sman *sman, unsigned long owner) -{ - - struct drm_hash_item *hash_item; - struct drm_owner_item *owner_item; - - if (drm_ht_find_item(&sman->owner_hash_tab, owner, &hash_item)) { - return -1; - } - - owner_item = drm_hash_entry(hash_item, struct drm_owner_item, owner_hash); - if (owner_item->mem_blocks.next == &owner_item->mem_blocks) { - drm_sman_remove_owner(sman, owner_item); - return -1; - } - - return 0; -} - -EXPORT_SYMBOL(drm_sman_owner_clean); - -static void drm_sman_do_owner_cleanup(struct drm_sman *sman, - struct drm_owner_item *owner_item) -{ - struct drm_memblock_item *entry, *next; - - list_for_each_entry_safe(entry, next, &owner_item->mem_blocks, - owner_list) { - drm_sman_free(entry); - } - drm_sman_remove_owner(sman, owner_item); -} - -void drm_sman_owner_cleanup(struct drm_sman *sman, unsigned long owner) -{ - - struct drm_hash_item *hash_item; - struct drm_owner_item *owner_item; - - if (drm_ht_find_item(&sman->owner_hash_tab, owner, &hash_item)) { - - return; - } - - owner_item = drm_hash_entry(hash_item, struct drm_owner_item, owner_hash); - drm_sman_do_owner_cleanup(sman, owner_item); -} - -EXPORT_SYMBOL(drm_sman_owner_cleanup); - -void drm_sman_cleanup(struct drm_sman *sman) -{ - struct drm_owner_item *entry, *next; - unsigned int i; - struct drm_sman_mm *sman_mm; - - list_for_each_entry_safe(entry, next, &sman->owner_items, sman_list) { - drm_sman_do_owner_cleanup(sman, entry); - } - if (sman->mm) { - for (i = 0; i < sman->num_managers; ++i) { - sman_mm = &sman->mm[i]; - if (sman_mm->private) { - sman_mm->destroy(sman_mm->private); - sman_mm->private = NULL; - } - } - } -} - -EXPORT_SYMBOL(drm_sman_cleanup); diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c index 6d7b083..aa454f8 100644 --- a/drivers/gpu/drm/drm_stub.c +++ b/drivers/gpu/drm/drm_stub.c @@ -319,6 +319,7 @@ int drm_fill_in_dev(struct drm_device *dev, drm_lastclose(dev); return retcode; } +EXPORT_SYMBOL(drm_fill_in_dev); /** @@ -397,6 +398,7 @@ err_idr: *minor = NULL; return ret; } +EXPORT_SYMBOL(drm_get_minor); /** * Put a secondary minor number. @@ -428,6 +430,12 @@ int drm_put_minor(struct drm_minor **minor_p) *minor_p = NULL; return 0; } +EXPORT_SYMBOL(drm_put_minor); + +static void drm_unplug_minor(struct drm_minor *minor) +{ + drm_sysfs_device_remove(minor); +} /** * Called via drm_exit() at module unload time or when pci device is @@ -492,3 +500,21 @@ void drm_put_dev(struct drm_device *dev) kfree(dev); } EXPORT_SYMBOL(drm_put_dev); + +void drm_unplug_dev(struct drm_device *dev) +{ + /* for a USB device */ + if (drm_core_check_feature(dev, DRIVER_MODESET)) + drm_unplug_minor(dev->control); + drm_unplug_minor(dev->primary); + + mutex_lock(&drm_global_mutex); + + drm_device_set_unplugged(dev); + + if (dev->open_count == 0) { + drm_put_dev(dev); + } + mutex_unlock(&drm_global_mutex); +} +EXPORT_SYMBOL(drm_unplug_dev); diff --git a/drivers/gpu/drm/drm_sysfs.c b/drivers/gpu/drm/drm_sysfs.c index 0f9ef9b..5a7bd51 100644 --- a/drivers/gpu/drm/drm_sysfs.c +++ b/drivers/gpu/drm/drm_sysfs.c @@ -72,7 +72,7 @@ static int drm_class_resume(struct device *dev) return 0; } -static char *drm_devnode(struct device *dev, mode_t *mode) +static char *drm_devnode(struct device *dev, umode_t *mode) { return kasprintf(GFP_KERNEL, "dri/%s", dev_name(dev)); } @@ -454,6 +454,8 @@ void drm_sysfs_connector_remove(struct drm_connector *connector) { int i; + if (!connector->kdev.parent) + return; DRM_DEBUG("removing \"%s\" from sysfs\n", drm_get_connector_name(connector)); @@ -461,6 +463,7 @@ void drm_sysfs_connector_remove(struct drm_connector *connector) device_remove_file(&connector->kdev, &connector_attrs[i]); sysfs_remove_bin_file(&connector->kdev.kobj, &edid_attr); device_unregister(&connector->kdev); + connector->kdev.parent = NULL; } EXPORT_SYMBOL(drm_sysfs_connector_remove); @@ -533,7 +536,9 @@ err_out: */ void drm_sysfs_device_remove(struct drm_minor *minor) { - device_unregister(&minor->kdev); + if (minor->kdev.parent) + device_unregister(&minor->kdev); + minor->kdev.parent = NULL; } diff --git a/drivers/gpu/drm/drm_usb.c b/drivers/gpu/drm/drm_usb.c index 445003f..37c9a52 100644 --- a/drivers/gpu/drm/drm_usb.c +++ b/drivers/gpu/drm/drm_usb.c @@ -1,8 +1,7 @@ #include "drmP.h" #include -#include +#include -#ifdef CONFIG_USB int drm_get_usb_dev(struct usb_interface *interface, const struct usb_device_id *id, struct drm_driver *driver) @@ -115,4 +114,7 @@ void drm_usb_exit(struct drm_driver *driver, usb_deregister(udriver); } EXPORT_SYMBOL(drm_usb_exit); -#endif + +MODULE_AUTHOR("David Airlie"); +MODULE_DESCRIPTION("USB DRM support"); +MODULE_LICENSE("GPL and additional rights"); diff --git a/drivers/gpu/drm/drm_vm.c b/drivers/gpu/drm/drm_vm.c index 8c03eaf..1495618 100644 --- a/drivers/gpu/drm/drm_vm.c +++ b/drivers/gpu/drm/drm_vm.c @@ -519,7 +519,6 @@ static int drm_mmap_dma(struct file *filp, struct vm_area_struct *vma) vma->vm_flags |= VM_RESERVED; /* Don't swap */ vma->vm_flags |= VM_DONTEXPAND; - vma->vm_file = filp; /* Needed for drm_vm_open() */ drm_vm_open_locked(vma); return 0; } @@ -671,7 +670,6 @@ int drm_mmap_locked(struct file *filp, struct vm_area_struct *vma) vma->vm_flags |= VM_RESERVED; /* Don't swap */ vma->vm_flags |= VM_DONTEXPAND; - vma->vm_file = filp; /* Needed for drm_vm_open() */ drm_vm_open_locked(vma); return 0; } @@ -682,6 +680,9 @@ int drm_mmap(struct file *filp, struct vm_area_struct *vma) struct drm_device *dev = priv->minor->dev; int ret; + if (drm_device_is_unplugged(dev)) + return -ENODEV; + mutex_lock(&dev->struct_mutex); ret = drm_mmap_locked(filp, vma); mutex_unlock(&dev->struct_mutex); diff --git a/drivers/gpu/drm/exynos/Kconfig b/drivers/gpu/drm/exynos/Kconfig index 847466a..3343ac4 100644 --- a/drivers/gpu/drm/exynos/Kconfig +++ b/drivers/gpu/drm/exynos/Kconfig @@ -1,7 +1,6 @@ config DRM_EXYNOS tristate "DRM Support for Samsung SoC EXYNOS Series" depends on DRM && PLAT_SAMSUNG - default n select DRM_KMS_HELPER select FB_CFB_FILLRECT select FB_CFB_COPYAREA @@ -12,9 +11,19 @@ config DRM_EXYNOS If M is selected the module will be called exynosdrm. config DRM_EXYNOS_FIMD - tristate "Exynos DRM FIMD" - depends on DRM_EXYNOS - default n + bool "Exynos DRM FIMD" + depends on DRM_EXYNOS && !FB_S3C help Choose this option if you want to use Exynos FIMD for DRM. - If M is selected, the module will be called exynos_drm_fimd + +config DRM_EXYNOS_HDMI + bool "Exynos DRM HDMI" + depends on DRM_EXYNOS && !VIDEO_SAMSUNG_S5P_TV + help + Choose this option if you want to use Exynos HDMI for DRM. + +config DRM_EXYNOS_VIDI + bool "Exynos DRM Virtual Display" + depends on DRM_EXYNOS + help + Choose this option if you want to use Exynos VIDI for DRM. diff --git a/drivers/gpu/drm/exynos/Makefile b/drivers/gpu/drm/exynos/Makefile index 0496d3f..9e0bff8 100644 --- a/drivers/gpu/drm/exynos/Makefile +++ b/drivers/gpu/drm/exynos/Makefile @@ -5,7 +5,13 @@ ccflags-y := -Iinclude/drm -Idrivers/gpu/drm/exynos exynosdrm-y := exynos_drm_drv.o exynos_drm_encoder.o exynos_drm_connector.o \ exynos_drm_crtc.o exynos_drm_fbdev.o exynos_drm_fb.o \ - exynos_drm_buf.o exynos_drm_gem.o exynos_drm_core.o + exynos_drm_buf.o exynos_drm_gem.o exynos_drm_core.o \ + exynos_drm_plane.o -obj-$(CONFIG_DRM_EXYNOS) += exynosdrm.o -obj-$(CONFIG_DRM_EXYNOS_FIMD) += exynos_drm_fimd.o +exynosdrm-$(CONFIG_DRM_EXYNOS_FIMD) += exynos_drm_fimd.o +exynosdrm-$(CONFIG_DRM_EXYNOS_HDMI) += exynos_hdmi.o exynos_mixer.o \ + exynos_ddc.o exynos_hdmiphy.o \ + exynos_drm_hdmi.o +exynosdrm-$(CONFIG_DRM_EXYNOS_VIDI) += exynos_drm_vidi.o + +obj-$(CONFIG_DRM_EXYNOS) += exynosdrm.o diff --git a/drivers/gpu/drm/exynos/exynos_ddc.c b/drivers/gpu/drm/exynos/exynos_ddc.c new file mode 100644 index 0000000..7e1051d --- /dev/null +++ b/drivers/gpu/drm/exynos/exynos_ddc.c @@ -0,0 +1,57 @@ +/* + * Copyright (C) 2011 Samsung Electronics Co.Ltd + * Authors: + * Seung-Woo Kim + * Inki Dae + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or (at your + * option) any later version. + * + */ + +#include "drmP.h" + +#include +#include +#include + + +#include "exynos_drm_drv.h" +#include "exynos_hdmi.h" + +static int s5p_ddc_probe(struct i2c_client *client, + const struct i2c_device_id *dev_id) +{ + hdmi_attach_ddc_client(client); + + dev_info(&client->adapter->dev, "attached s5p_ddc " + "into i2c adapter successfully\n"); + + return 0; +} + +static int s5p_ddc_remove(struct i2c_client *client) +{ + dev_info(&client->adapter->dev, "detached s5p_ddc " + "from i2c adapter successfully\n"); + + return 0; +} + +static struct i2c_device_id ddc_idtable[] = { + {"s5p_ddc", 0}, + { }, +}; + +struct i2c_driver ddc_driver = { + .driver = { + .name = "s5p_ddc", + .owner = THIS_MODULE, + }, + .id_table = ddc_idtable, + .probe = s5p_ddc_probe, + .remove = __devexit_p(s5p_ddc_remove), + .command = NULL, +}; diff --git a/drivers/gpu/drm/exynos/exynos_drm_buf.c b/drivers/gpu/drm/exynos/exynos_drm_buf.c index 2bb07bc..de8d209 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_buf.c +++ b/drivers/gpu/drm/exynos/exynos_drm_buf.c @@ -25,45 +25,142 @@ #include "drmP.h" #include "drm.h" +#include "exynos_drm.h" #include "exynos_drm_drv.h" #include "exynos_drm_gem.h" #include "exynos_drm_buf.h" static int lowlevel_buffer_allocate(struct drm_device *dev, - struct exynos_drm_gem_buf *buffer) + unsigned int flags, struct exynos_drm_gem_buf *buf) { + dma_addr_t start_addr; + unsigned int npages, page_size, i = 0; + struct scatterlist *sgl; + int ret = 0; + DRM_DEBUG_KMS("%s\n", __FILE__); - buffer->kvaddr = dma_alloc_writecombine(dev->dev, buffer->size, - &buffer->dma_addr, GFP_KERNEL); - if (!buffer->kvaddr) { - DRM_ERROR("failed to allocate buffer.\n"); + if (IS_NONCONTIG_BUFFER(flags)) { + DRM_DEBUG_KMS("not support allocation type.\n"); + return -EINVAL; + } + + if (buf->dma_addr) { + DRM_DEBUG_KMS("already allocated.\n"); + return 0; + } + + if (buf->size >= SZ_1M) { + npages = buf->size >> SECTION_SHIFT; + page_size = SECTION_SIZE; + } else if (buf->size >= SZ_64K) { + npages = buf->size >> 16; + page_size = SZ_64K; + } else { + npages = buf->size >> PAGE_SHIFT; + page_size = PAGE_SIZE; + } + + buf->sgt = kzalloc(sizeof(struct sg_table), GFP_KERNEL); + if (!buf->sgt) { + DRM_ERROR("failed to allocate sg table.\n"); return -ENOMEM; } - DRM_DEBUG_KMS("vaddr(0x%lx), dma_addr(0x%lx), size(0x%lx)\n", - (unsigned long)buffer->kvaddr, - (unsigned long)buffer->dma_addr, - buffer->size); + ret = sg_alloc_table(buf->sgt, npages, GFP_KERNEL); + if (ret < 0) { + DRM_ERROR("failed to initialize sg table.\n"); + kfree(buf->sgt); + buf->sgt = NULL; + return -ENOMEM; + } - return 0; + buf->kvaddr = dma_alloc_writecombine(dev->dev, buf->size, + &buf->dma_addr, GFP_KERNEL); + if (!buf->kvaddr) { + DRM_ERROR("failed to allocate buffer.\n"); + ret = -ENOMEM; + goto err1; + } + + buf->pages = kzalloc(sizeof(struct page) * npages, GFP_KERNEL); + if (!buf->pages) { + DRM_ERROR("failed to allocate pages.\n"); + ret = -ENOMEM; + goto err2; + } + + sgl = buf->sgt->sgl; + start_addr = buf->dma_addr; + + while (i < npages) { + buf->pages[i] = phys_to_page(start_addr); + sg_set_page(sgl, buf->pages[i], page_size, 0); + sg_dma_address(sgl) = start_addr; + start_addr += page_size; + sgl = sg_next(sgl); + i++; + } + + DRM_DEBUG_KMS("vaddr(0x%lx), dma_addr(0x%lx), size(0x%lx)\n", + (unsigned long)buf->kvaddr, + (unsigned long)buf->dma_addr, + buf->size); + + return ret; +err2: + dma_free_writecombine(dev->dev, buf->size, buf->kvaddr, + (dma_addr_t)buf->dma_addr); + buf->dma_addr = (dma_addr_t)NULL; +err1: + sg_free_table(buf->sgt); + kfree(buf->sgt); + buf->sgt = NULL; + + return ret; } static void lowlevel_buffer_deallocate(struct drm_device *dev, - struct exynos_drm_gem_buf *buffer) + unsigned int flags, struct exynos_drm_gem_buf *buf) { DRM_DEBUG_KMS("%s.\n", __FILE__); - if (buffer->dma_addr && buffer->size) - dma_free_writecombine(dev->dev, buffer->size, buffer->kvaddr, - (dma_addr_t)buffer->dma_addr); - else - DRM_DEBUG_KMS("buffer data are invalid.\n"); + /* + * release only physically continuous memory and + * non-continuous memory would be released by exynos + * gem framework. + */ + if (IS_NONCONTIG_BUFFER(flags)) { + DRM_DEBUG_KMS("not support allocation type.\n"); + return; + } + + if (!buf->dma_addr) { + DRM_DEBUG_KMS("dma_addr is invalid.\n"); + return; + } + + DRM_DEBUG_KMS("vaddr(0x%lx), dma_addr(0x%lx), size(0x%lx)\n", + (unsigned long)buf->kvaddr, + (unsigned long)buf->dma_addr, + buf->size); + + sg_free_table(buf->sgt); + + kfree(buf->sgt); + buf->sgt = NULL; + + kfree(buf->pages); + buf->pages = NULL; + + dma_free_writecombine(dev->dev, buf->size, buf->kvaddr, + (dma_addr_t)buf->dma_addr); + buf->dma_addr = (dma_addr_t)NULL; } -struct exynos_drm_gem_buf *exynos_drm_buf_create(struct drm_device *dev, - unsigned int size) +struct exynos_drm_gem_buf *exynos_drm_init_buf(struct drm_device *dev, + unsigned int size) { struct exynos_drm_gem_buf *buffer; @@ -73,26 +170,15 @@ struct exynos_drm_gem_buf *exynos_drm_buf_create(struct drm_device *dev, buffer = kzalloc(sizeof(*buffer), GFP_KERNEL); if (!buffer) { DRM_ERROR("failed to allocate exynos_drm_gem_buf.\n"); - return ERR_PTR(-ENOMEM); + return NULL; } buffer->size = size; - - /* - * allocate memory region with size and set the memory information - * to vaddr and dma_addr of a buffer object. - */ - if (lowlevel_buffer_allocate(dev, buffer) < 0) { - kfree(buffer); - buffer = NULL; - return ERR_PTR(-ENOMEM); - } - return buffer; } -void exynos_drm_buf_destroy(struct drm_device *dev, - struct exynos_drm_gem_buf *buffer) +void exynos_drm_fini_buf(struct drm_device *dev, + struct exynos_drm_gem_buf *buffer) { DRM_DEBUG_KMS("%s.\n", __FILE__); @@ -101,12 +187,27 @@ void exynos_drm_buf_destroy(struct drm_device *dev, return; } - lowlevel_buffer_deallocate(dev, buffer); - kfree(buffer); buffer = NULL; } -MODULE_AUTHOR("Inki Dae "); -MODULE_DESCRIPTION("Samsung SoC DRM Buffer Management Module"); -MODULE_LICENSE("GPL"); +int exynos_drm_alloc_buf(struct drm_device *dev, + struct exynos_drm_gem_buf *buf, unsigned int flags) +{ + + /* + * allocate memory region and set the memory information + * to vaddr and dma_addr of a buffer object. + */ + if (lowlevel_buffer_allocate(dev, flags, buf) < 0) + return -ENOMEM; + + return 0; +} + +void exynos_drm_free_buf(struct drm_device *dev, + unsigned int flags, struct exynos_drm_gem_buf *buffer) +{ + + lowlevel_buffer_deallocate(dev, flags, buffer); +} diff --git a/drivers/gpu/drm/exynos/exynos_drm_buf.h b/drivers/gpu/drm/exynos/exynos_drm_buf.h index 6e91f9c..3388e4e 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_buf.h +++ b/drivers/gpu/drm/exynos/exynos_drm_buf.h @@ -26,15 +26,22 @@ #ifndef _EXYNOS_DRM_BUF_H_ #define _EXYNOS_DRM_BUF_H_ -/* allocate physical memory. */ -struct exynos_drm_gem_buf *exynos_drm_buf_create(struct drm_device *dev, - unsigned int size); +/* create and initialize buffer object. */ +struct exynos_drm_gem_buf *exynos_drm_init_buf(struct drm_device *dev, + unsigned int size); -/* get memory information of a drm framebuffer. */ -struct exynos_drm_gem_buf *exynos_drm_fb_get_buf(struct drm_framebuffer *fb); +/* destroy buffer object. */ +void exynos_drm_fini_buf(struct drm_device *dev, + struct exynos_drm_gem_buf *buffer); -/* remove allocated physical memory. */ -void exynos_drm_buf_destroy(struct drm_device *dev, - struct exynos_drm_gem_buf *buffer); +/* allocate physical memory region and setup sgt and pages. */ +int exynos_drm_alloc_buf(struct drm_device *dev, + struct exynos_drm_gem_buf *buf, + unsigned int flags); + +/* release physical memory region, sgt and pages. */ +void exynos_drm_free_buf(struct drm_device *dev, + unsigned int flags, + struct exynos_drm_gem_buf *buffer); #endif diff --git a/drivers/gpu/drm/exynos/exynos_drm_connector.c b/drivers/gpu/drm/exynos/exynos_drm_connector.c index d620b07..bf791fa 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_connector.c +++ b/drivers/gpu/drm/exynos/exynos_drm_connector.c @@ -28,6 +28,7 @@ #include "drmP.h" #include "drm_crtc_helper.h" +#include #include "exynos_drm_drv.h" #include "exynos_drm_encoder.h" @@ -44,22 +45,25 @@ struct exynos_drm_connector { /* convert exynos_video_timings to drm_display_mode */ static inline void convert_to_display_mode(struct drm_display_mode *mode, - struct fb_videomode *timing) + struct exynos_drm_panel_info *panel) { + struct fb_videomode *timing = &panel->timing; DRM_DEBUG_KMS("%s\n", __FILE__); mode->clock = timing->pixclock / 1000; mode->vrefresh = timing->refresh; mode->hdisplay = timing->xres; - mode->hsync_start = mode->hdisplay + timing->left_margin; + mode->hsync_start = mode->hdisplay + timing->right_margin; mode->hsync_end = mode->hsync_start + timing->hsync_len; - mode->htotal = mode->hsync_end + timing->right_margin; + mode->htotal = mode->hsync_end + timing->left_margin; mode->vdisplay = timing->yres; - mode->vsync_start = mode->vdisplay + timing->upper_margin; + mode->vsync_start = mode->vdisplay + timing->lower_margin; mode->vsync_end = mode->vsync_start + timing->vsync_len; - mode->vtotal = mode->vsync_end + timing->lower_margin; + mode->vtotal = mode->vsync_end + timing->upper_margin; + mode->width_mm = panel->width_mm; + mode->height_mm = panel->height_mm; if (timing->vmode & FB_VMODE_INTERLACED) mode->flags |= DRM_MODE_FLAG_INTERLACE; @@ -81,14 +85,14 @@ convert_to_video_timing(struct fb_videomode *timing, timing->refresh = drm_mode_vrefresh(mode); timing->xres = mode->hdisplay; - timing->left_margin = mode->hsync_start - mode->hdisplay; + timing->right_margin = mode->hsync_start - mode->hdisplay; timing->hsync_len = mode->hsync_end - mode->hsync_start; - timing->right_margin = mode->htotal - mode->hsync_end; + timing->left_margin = mode->htotal - mode->hsync_end; timing->yres = mode->vdisplay; - timing->upper_margin = mode->vsync_start - mode->vdisplay; + timing->lower_margin = mode->vsync_start - mode->vdisplay; timing->vsync_len = mode->vsync_end - mode->vsync_start; - timing->lower_margin = mode->vtotal - mode->vsync_end; + timing->upper_margin = mode->vtotal - mode->vsync_end; if (mode->flags & DRM_MODE_FLAG_INTERLACE) timing->vmode = FB_VMODE_INTERLACED; @@ -148,16 +152,18 @@ static int exynos_drm_connector_get_modes(struct drm_connector *connector) connector->display_info.raw_edid = edid; } else { struct drm_display_mode *mode = drm_mode_create(connector->dev); - struct fb_videomode *timing; + struct exynos_drm_panel_info *panel; - if (display_ops->get_timing) - timing = display_ops->get_timing(manager->dev); + if (display_ops->get_panel) + panel = display_ops->get_panel(manager->dev); else { drm_mode_destroy(connector->dev, mode); return 0; } - convert_to_display_mode(mode, timing); + convert_to_display_mode(mode, panel); + connector->display_info.width_mm = mode->width_mm; + connector->display_info.height_mm = mode->height_mm; mode->type = DRM_MODE_TYPE_DRIVER | DRM_MODE_TYPE_PREFERRED; drm_mode_set_name(mode); @@ -219,6 +225,29 @@ static struct drm_connector_helper_funcs exynos_connector_helper_funcs = { .best_encoder = exynos_drm_best_encoder, }; +static int exynos_drm_connector_fill_modes(struct drm_connector *connector, + unsigned int max_width, unsigned int max_height) +{ + struct exynos_drm_connector *exynos_connector = + to_exynos_connector(connector); + struct exynos_drm_manager *manager = exynos_connector->manager; + struct exynos_drm_manager_ops *ops = manager->ops; + unsigned int width, height; + + width = max_width; + height = max_height; + + /* + * if specific driver want to find desired_mode using maxmum + * resolution then get max width and height from that driver. + */ + if (ops && ops->get_max_resol) + ops->get_max_resol(manager->dev, &width, &height); + + return drm_helper_probe_single_connector_modes(connector, width, + height); +} + /* get detection status of display device. */ static enum drm_connector_status exynos_drm_connector_detect(struct drm_connector *connector, bool force) @@ -256,7 +285,7 @@ static void exynos_drm_connector_destroy(struct drm_connector *connector) static struct drm_connector_funcs exynos_connector_funcs = { .dpms = drm_helper_connector_dpms, - .fill_modes = drm_helper_probe_single_connector_modes, + .fill_modes = exynos_drm_connector_fill_modes, .detect = exynos_drm_connector_detect, .destroy = exynos_drm_connector_destroy, }; @@ -286,6 +315,10 @@ struct drm_connector *exynos_drm_connector_create(struct drm_device *dev, connector->interlace_allowed = true; connector->polled = DRM_CONNECTOR_POLL_HPD; break; + case EXYNOS_DISPLAY_TYPE_VIDI: + type = DRM_MODE_CONNECTOR_VIRTUAL; + connector->polled = DRM_CONNECTOR_POLL_HPD; + break; default: type = DRM_MODE_CONNECTOR_Unknown; break; @@ -319,9 +352,3 @@ err_connector: kfree(exynos_connector); return NULL; } - -MODULE_AUTHOR("Inki Dae "); -MODULE_AUTHOR("Joonyoung Shim "); -MODULE_AUTHOR("Seung-Woo Kim "); -MODULE_DESCRIPTION("Samsung SoC DRM Connector Driver"); -MODULE_LICENSE("GPL"); diff --git a/drivers/gpu/drm/exynos/exynos_drm_core.c b/drivers/gpu/drm/exynos/exynos_drm_core.c index 661a035..eaf630d 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_core.c +++ b/drivers/gpu/drm/exynos/exynos_drm_core.c @@ -32,7 +32,6 @@ #include "exynos_drm_connector.h" #include "exynos_drm_fbdev.h" -static DEFINE_MUTEX(exynos_drm_mutex); static LIST_HEAD(exynos_drm_subdrv_list); static struct drm_device *drm_dev; @@ -55,13 +54,18 @@ static int exynos_drm_subdrv_probe(struct drm_device *dev, * * P.S. note that this driver is considered for modularization. */ - ret = subdrv->probe(dev, subdrv->manager.dev); + ret = subdrv->probe(dev, subdrv->dev); if (ret) return ret; } + if (!subdrv->manager) + return 0; + + subdrv->manager->dev = subdrv->dev; + /* create and initialize a encoder for this sub driver. */ - encoder = exynos_drm_encoder_create(dev, &subdrv->manager, + encoder = exynos_drm_encoder_create(dev, subdrv->manager, (1 << MAX_CRTC) - 1); if (!encoder) { DRM_ERROR("failed to create encoder\n"); @@ -116,13 +120,10 @@ int exynos_drm_device_register(struct drm_device *dev) if (!dev) return -EINVAL; - if (drm_dev) { - DRM_ERROR("Already drm device were registered\n"); - return -EBUSY; - } + drm_dev = dev; - mutex_lock(&exynos_drm_mutex); list_for_each_entry_safe(subdrv, n, &exynos_drm_subdrv_list, list) { + subdrv->drm_dev = dev; err = exynos_drm_subdrv_probe(dev, subdrv); if (err) { DRM_DEBUG("exynos drm subdrv probe failed.\n"); @@ -130,9 +131,6 @@ int exynos_drm_device_register(struct drm_device *dev) } } - drm_dev = dev; - mutex_unlock(&exynos_drm_mutex); - return 0; } EXPORT_SYMBOL_GPL(exynos_drm_device_register); @@ -143,83 +141,28 @@ int exynos_drm_device_unregister(struct drm_device *dev) DRM_DEBUG_DRIVER("%s\n", __FILE__); - if (!dev || dev != drm_dev) { + if (!dev) { WARN(1, "Unexpected drm device unregister!\n"); return -EINVAL; } - mutex_lock(&exynos_drm_mutex); list_for_each_entry(subdrv, &exynos_drm_subdrv_list, list) exynos_drm_subdrv_remove(dev, subdrv); drm_dev = NULL; - mutex_unlock(&exynos_drm_mutex); return 0; } EXPORT_SYMBOL_GPL(exynos_drm_device_unregister); -static int exynos_drm_mode_group_reinit(struct drm_device *dev) -{ - struct drm_mode_group *group = &dev->primary->mode_group; - uint32_t *id_list = group->id_list; - int ret; - - DRM_DEBUG_DRIVER("%s\n", __FILE__); - - ret = drm_mode_group_init_legacy_group(dev, group); - if (ret < 0) - return ret; - - kfree(id_list); - return 0; -} - int exynos_drm_subdrv_register(struct exynos_drm_subdrv *subdrv) { - int err; - DRM_DEBUG_DRIVER("%s\n", __FILE__); if (!subdrv) return -EINVAL; - mutex_lock(&exynos_drm_mutex); - if (drm_dev) { - err = exynos_drm_subdrv_probe(drm_dev, subdrv); - if (err) { - DRM_ERROR("failed to probe exynos drm subdrv\n"); - mutex_unlock(&exynos_drm_mutex); - return err; - } - - /* - * if any specific driver such as fimd or hdmi driver called - * exynos_drm_subdrv_register() later than drm_load(), - * the fb helper should be re-initialized and re-configured. - */ - err = exynos_drm_fbdev_reinit(drm_dev); - if (err) { - DRM_ERROR("failed to reinitialize exynos drm fbdev\n"); - exynos_drm_subdrv_remove(drm_dev, subdrv); - mutex_unlock(&exynos_drm_mutex); - return err; - } - - err = exynos_drm_mode_group_reinit(drm_dev); - if (err) { - DRM_ERROR("failed to reinitialize mode group\n"); - exynos_drm_fbdev_fini(drm_dev); - exynos_drm_subdrv_remove(drm_dev, subdrv); - mutex_unlock(&exynos_drm_mutex); - return err; - } - } - - subdrv->drm_dev = drm_dev; - list_add_tail(&subdrv->list, &exynos_drm_subdrv_list); - mutex_unlock(&exynos_drm_mutex); return 0; } @@ -227,46 +170,48 @@ EXPORT_SYMBOL_GPL(exynos_drm_subdrv_register); int exynos_drm_subdrv_unregister(struct exynos_drm_subdrv *subdrv) { - int ret = -EFAULT; - DRM_DEBUG_DRIVER("%s\n", __FILE__); - if (!subdrv) { - DRM_DEBUG("Unexpected exynos drm subdrv unregister!\n"); - return ret; - } + if (!subdrv) + return -EINVAL; - mutex_lock(&exynos_drm_mutex); - if (drm_dev) { - exynos_drm_subdrv_remove(drm_dev, subdrv); - list_del(&subdrv->list); + list_del(&subdrv->list); - /* - * fb helper should be updated once a sub driver is released - * to re-configure crtc and connector and also to re-setup - * drm framebuffer. - */ - ret = exynos_drm_fbdev_reinit(drm_dev); - if (ret < 0) { - DRM_ERROR("failed fb helper reinit.\n"); - goto fail; - } + return 0; +} +EXPORT_SYMBOL_GPL(exynos_drm_subdrv_unregister); - ret = exynos_drm_mode_group_reinit(drm_dev); - if (ret < 0) { - DRM_ERROR("failed drm mode group reinit.\n"); - goto fail; +int exynos_drm_subdrv_open(struct drm_device *dev, struct drm_file *file) +{ + struct exynos_drm_subdrv *subdrv; + int ret; + + list_for_each_entry(subdrv, &exynos_drm_subdrv_list, list) { + if (subdrv->open) { + ret = subdrv->open(dev, subdrv->dev, file); + if (ret) + goto err; } } -fail: - mutex_unlock(&exynos_drm_mutex); + return 0; + +err: + list_for_each_entry_reverse(subdrv, &subdrv->list, list) { + if (subdrv->close) + subdrv->close(dev, subdrv->dev, file); + } return ret; } -EXPORT_SYMBOL_GPL(exynos_drm_subdrv_unregister); +EXPORT_SYMBOL_GPL(exynos_drm_subdrv_open); + +void exynos_drm_subdrv_close(struct drm_device *dev, struct drm_file *file) +{ + struct exynos_drm_subdrv *subdrv; -MODULE_AUTHOR("Inki Dae "); -MODULE_AUTHOR("Joonyoung Shim "); -MODULE_AUTHOR("Seung-Woo Kim "); -MODULE_DESCRIPTION("Samsung SoC DRM Core Driver"); -MODULE_LICENSE("GPL"); + list_for_each_entry(subdrv, &exynos_drm_subdrv_list, list) { + if (subdrv->close) + subdrv->close(dev, subdrv->dev, file); + } +} +EXPORT_SYMBOL_GPL(exynos_drm_subdrv_close); diff --git a/drivers/gpu/drm/exynos/exynos_drm_crtc.c b/drivers/gpu/drm/exynos/exynos_drm_crtc.c index ee43cc2..3486ffe 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_crtc.c +++ b/drivers/gpu/drm/exynos/exynos_drm_crtc.c @@ -34,7 +34,6 @@ #include "exynos_drm_fb.h" #include "exynos_drm_encoder.h" #include "exynos_drm_gem.h" -#include "exynos_drm_buf.h" #define to_exynos_crtc(x) container_of(x, struct exynos_drm_crtc,\ drm_crtc) @@ -52,11 +51,13 @@ * drm framework doesn't support multiple irq yet. * we can refer to the crtc to current hardware interrupt occured through * this pipe value. + * @dpms: store the crtc dpms value */ struct exynos_drm_crtc { struct drm_crtc drm_crtc; struct exynos_drm_overlay overlay; unsigned int pipe; + unsigned int dpms; }; static void exynos_drm_crtc_apply(struct drm_crtc *crtc) @@ -78,19 +79,23 @@ int exynos_drm_overlay_update(struct exynos_drm_overlay *overlay, struct exynos_drm_gem_buf *buffer; unsigned int actual_w; unsigned int actual_h; + int nr = exynos_drm_format_num_buffers(fb->pixel_format); + int i; + + for (i = 0; i < nr; i++) { + buffer = exynos_drm_fb_buffer(fb, i); + if (!buffer) { + DRM_LOG_KMS("buffer is null\n"); + return -EFAULT; + } - buffer = exynos_drm_fb_get_buf(fb); - if (!buffer) { - DRM_LOG_KMS("buffer is null.\n"); - return -EFAULT; - } - - overlay->dma_addr = buffer->dma_addr; - overlay->vaddr = buffer->kvaddr; + overlay->dma_addr[i] = buffer->dma_addr; + overlay->vaddr[i] = buffer->kvaddr; - DRM_DEBUG_KMS("vaddr = 0x%lx, dma_addr = 0x%lx\n", - (unsigned long)overlay->vaddr, - (unsigned long)overlay->dma_addr); + DRM_DEBUG_KMS("buffer: %d, vaddr = 0x%lx, dma_addr = 0x%lx\n", + i, (unsigned long)overlay->vaddr[i], + (unsigned long)overlay->dma_addr[i]); + } actual_w = min((mode->hdisplay - pos->crtc_x), pos->crtc_w); actual_h = min((mode->vdisplay - pos->crtc_y), pos->crtc_h); @@ -101,7 +106,8 @@ int exynos_drm_overlay_update(struct exynos_drm_overlay *overlay, overlay->fb_width = fb->width; overlay->fb_height = fb->height; overlay->bpp = fb->bits_per_pixel; - overlay->pitch = fb->pitch; + overlay->pitch = fb->pitches[0]; + overlay->pixel_format = fb->pixel_format; /* set overlay range to be displayed. */ overlay->crtc_x = pos->crtc_x; @@ -153,26 +159,37 @@ static int exynos_drm_crtc_update(struct drm_crtc *crtc) static void exynos_drm_crtc_dpms(struct drm_crtc *crtc, int mode) { + struct drm_device *dev = crtc->dev; struct exynos_drm_crtc *exynos_crtc = to_exynos_crtc(crtc); DRM_DEBUG_KMS("crtc[%d] mode[%d]\n", crtc->base.id, mode); + if (exynos_crtc->dpms == mode) { + DRM_DEBUG_KMS("desired dpms mode is same as previous one.\n"); + return; + } + + mutex_lock(&dev->struct_mutex); + switch (mode) { case DRM_MODE_DPMS_ON: - exynos_drm_fn_encoder(crtc, &exynos_crtc->pipe, - exynos_drm_encoder_crtc_commit); + exynos_drm_fn_encoder(crtc, &mode, + exynos_drm_encoder_crtc_dpms); + exynos_crtc->dpms = mode; break; case DRM_MODE_DPMS_STANDBY: case DRM_MODE_DPMS_SUSPEND: case DRM_MODE_DPMS_OFF: - /* TODO */ - exynos_drm_fn_encoder(crtc, NULL, - exynos_drm_encoder_crtc_disable); + exynos_drm_fn_encoder(crtc, &mode, + exynos_drm_encoder_crtc_dpms); + exynos_crtc->dpms = mode; break; default: - DRM_DEBUG_KMS("unspecified mode %d\n", mode); + DRM_ERROR("unspecified mode %d\n", mode); break; } + + mutex_unlock(&dev->struct_mutex); } static void exynos_drm_crtc_prepare(struct drm_crtc *crtc) @@ -188,6 +205,28 @@ static void exynos_drm_crtc_commit(struct drm_crtc *crtc) DRM_DEBUG_KMS("%s\n", __FILE__); + /* + * when set_crtc is requested from user or at booting time, + * crtc->commit would be called without dpms call so if dpms is + * no power on then crtc->dpms should be called + * with DRM_MODE_DPMS_ON for the hardware power to be on. + */ + if (exynos_crtc->dpms != DRM_MODE_DPMS_ON) { + int mode = DRM_MODE_DPMS_ON; + + /* + * enable hardware(power on) to all encoders hdmi connected + * to current crtc. + */ + exynos_drm_crtc_dpms(crtc, mode); + /* + * enable dma to all encoders connected to current crtc and + * lcd panel. + */ + exynos_drm_fn_encoder(crtc, &mode, + exynos_drm_encoder_dpms_from_crtc); + } + exynos_drm_fn_encoder(crtc, &exynos_crtc->pipe, exynos_drm_encoder_crtc_commit); } @@ -210,7 +249,11 @@ exynos_drm_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode, { DRM_DEBUG_KMS("%s\n", __FILE__); - mode = adjusted_mode; + /* + * copy the mode data adjusted by mode_fixup() into crtc->mode + * so that hardware can be seet to proper mode. + */ + memcpy(&crtc->mode, adjusted_mode, sizeof(*adjusted_mode)); return exynos_drm_crtc_update(crtc); } @@ -268,9 +311,6 @@ static int exynos_drm_crtc_page_flip(struct drm_crtc *crtc, */ event->pipe = exynos_crtc->pipe; - list_add_tail(&event->base.link, - &dev_priv->pageflip_event_list); - ret = drm_vblank_get(dev, exynos_crtc->pipe); if (ret) { DRM_DEBUG("failed to acquire vblank counter\n"); @@ -279,6 +319,9 @@ static int exynos_drm_crtc_page_flip(struct drm_crtc *crtc, goto out; } + list_add_tail(&event->base.link, + &dev_priv->pageflip_event_list); + crtc->fb = fb; ret = exynos_drm_crtc_update(crtc); if (ret) { @@ -344,6 +387,8 @@ int exynos_drm_crtc_create(struct drm_device *dev, unsigned int nr) } exynos_crtc->pipe = nr; + exynos_crtc->dpms = DRM_MODE_DPMS_OFF; + exynos_crtc->overlay.zpos = DEFAULT_ZPOS; crtc = &exynos_crtc->drm_crtc; private->crtc[nr] = crtc; @@ -357,9 +402,14 @@ int exynos_drm_crtc_create(struct drm_device *dev, unsigned int nr) int exynos_drm_crtc_enable_vblank(struct drm_device *dev, int crtc) { struct exynos_drm_private *private = dev->dev_private; + struct exynos_drm_crtc *exynos_crtc = + to_exynos_crtc(private->crtc[crtc]); DRM_DEBUG_KMS("%s\n", __FILE__); + if (exynos_crtc->dpms != DRM_MODE_DPMS_ON) + return -EPERM; + exynos_drm_fn_encoder(private->crtc[crtc], &crtc, exynos_drm_enable_vblank); @@ -369,15 +419,14 @@ int exynos_drm_crtc_enable_vblank(struct drm_device *dev, int crtc) void exynos_drm_crtc_disable_vblank(struct drm_device *dev, int crtc) { struct exynos_drm_private *private = dev->dev_private; + struct exynos_drm_crtc *exynos_crtc = + to_exynos_crtc(private->crtc[crtc]); DRM_DEBUG_KMS("%s\n", __FILE__); + if (exynos_crtc->dpms != DRM_MODE_DPMS_ON) + return; + exynos_drm_fn_encoder(private->crtc[crtc], &crtc, exynos_drm_disable_vblank); } - -MODULE_AUTHOR("Inki Dae "); -MODULE_AUTHOR("Joonyoung Shim "); -MODULE_AUTHOR("Seung-Woo Kim "); -MODULE_DESCRIPTION("Samsung SoC DRM CRTC Driver"); -MODULE_LICENSE("GPL"); diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.c b/drivers/gpu/drm/exynos/exynos_drm_drv.c index 53e2216..a6819b5 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_drv.c +++ b/drivers/gpu/drm/exynos/exynos_drm_drv.c @@ -33,16 +33,21 @@ #include "exynos_drm_drv.h" #include "exynos_drm_crtc.h" +#include "exynos_drm_encoder.h" #include "exynos_drm_fbdev.h" #include "exynos_drm_fb.h" #include "exynos_drm_gem.h" +#include "exynos_drm_plane.h" +#include "exynos_drm_vidi.h" -#define DRIVER_NAME "exynos-drm" +#define DRIVER_NAME "exynos" #define DRIVER_DESC "Samsung SoC DRM" #define DRIVER_DATE "20110530" #define DRIVER_MAJOR 1 #define DRIVER_MINOR 0 +#define VBLANK_OFF_DELAY 50000 + static int exynos_drm_load(struct drm_device *dev, unsigned long flags) { struct exynos_drm_private *private; @@ -77,6 +82,12 @@ static int exynos_drm_load(struct drm_device *dev, unsigned long flags) goto err_crtc; } + for (nr = 0; nr < MAX_PLANE; nr++) { + ret = exynos_plane_init(dev, nr); + if (ret) + goto err_crtc; + } + ret = drm_vblank_init(dev, MAX_CRTC); if (ret) goto err_crtc; @@ -90,6 +101,9 @@ static int exynos_drm_load(struct drm_device *dev, unsigned long flags) if (ret) goto err_vblank; + /* setup possible_clones. */ + exynos_drm_encoder_setup(dev); + /* * create and configure fb helper and also exynos specific * fbdev object. @@ -100,6 +114,8 @@ static int exynos_drm_load(struct drm_device *dev, unsigned long flags) goto err_drm_device; } + drm_vblank_offdelay = VBLANK_OFF_DELAY; + return 0; err_drm_device: @@ -129,17 +145,45 @@ static int exynos_drm_unload(struct drm_device *dev) return 0; } +static int exynos_drm_open(struct drm_device *dev, struct drm_file *file) +{ + DRM_DEBUG_DRIVER("%s\n", __FILE__); + + return exynos_drm_subdrv_open(dev, file); +} + static void exynos_drm_preclose(struct drm_device *dev, - struct drm_file *file_priv) + struct drm_file *file) { - struct exynos_drm_private *dev_priv = dev->dev_private; + struct exynos_drm_private *private = dev->dev_private; + struct drm_pending_vblank_event *e, *t; + unsigned long flags; - /* - * drm framework frees all events at release time, - * so private event list should be cleared. - */ - if (!list_empty(&dev_priv->pageflip_event_list)) - INIT_LIST_HEAD(&dev_priv->pageflip_event_list); + DRM_DEBUG_DRIVER("%s\n", __FILE__); + + /* release events of current file */ + spin_lock_irqsave(&dev->event_lock, flags); + list_for_each_entry_safe(e, t, &private->pageflip_event_list, + base.link) { + if (e->base.file_priv == file) { + list_del(&e->base.link); + e->base.destroy(&e->base); + } + } + spin_unlock_irqrestore(&dev->event_lock, flags); + + exynos_drm_subdrv_close(dev, file); +} + +static void exynos_drm_postclose(struct drm_device *dev, struct drm_file *file) +{ + DRM_DEBUG_DRIVER("%s\n", __FILE__); + + if (!file->driver_priv) + return; + + kfree(file->driver_priv); + file->driver_priv = NULL; } static void exynos_drm_lastclose(struct drm_device *dev) @@ -163,6 +207,20 @@ static struct drm_ioctl_desc exynos_ioctls[] = { DRM_AUTH), DRM_IOCTL_DEF_DRV(EXYNOS_GEM_MMAP, exynos_drm_gem_mmap_ioctl, DRM_UNLOCKED | DRM_AUTH), + DRM_IOCTL_DEF_DRV(EXYNOS_PLANE_SET_ZPOS, exynos_plane_set_zpos_ioctl, + DRM_UNLOCKED | DRM_AUTH), + DRM_IOCTL_DEF_DRV(EXYNOS_VIDI_CONNECTION, + vidi_connection_ioctl, DRM_UNLOCKED | DRM_AUTH), +}; + +static const struct file_operations exynos_drm_driver_fops = { + .owner = THIS_MODULE, + .open = drm_open, + .mmap = exynos_drm_gem_mmap, + .poll = drm_poll, + .read = drm_read, + .unlocked_ioctl = drm_ioctl, + .release = drm_release, }; static struct drm_driver exynos_drm_driver = { @@ -170,8 +228,10 @@ static struct drm_driver exynos_drm_driver = { DRIVER_MODESET | DRIVER_GEM, .load = exynos_drm_load, .unload = exynos_drm_unload, + .open = exynos_drm_open, .preclose = exynos_drm_preclose, .lastclose = exynos_drm_lastclose, + .postclose = exynos_drm_postclose, .get_vblank_counter = drm_vblank_count, .enable_vblank = exynos_drm_crtc_enable_vblank, .disable_vblank = exynos_drm_crtc_disable_vblank, @@ -182,15 +242,7 @@ static struct drm_driver exynos_drm_driver = { .dumb_map_offset = exynos_drm_gem_dumb_map_offset, .dumb_destroy = exynos_drm_gem_dumb_destroy, .ioctls = exynos_ioctls, - .fops = { - .owner = THIS_MODULE, - .open = drm_open, - .mmap = exynos_drm_gem_mmap, - .poll = drm_poll, - .read = drm_read, - .unlocked_ioctl = drm_ioctl, - .release = drm_release, - }, + .fops = &exynos_drm_driver_fops, .name = DRIVER_NAME, .desc = DRIVER_DESC, .date = DRIVER_DATE, @@ -221,15 +273,66 @@ static struct platform_driver exynos_drm_platform_driver = { .remove = __devexit_p(exynos_drm_platform_remove), .driver = { .owner = THIS_MODULE, - .name = DRIVER_NAME, + .name = "exynos-drm", }, }; static int __init exynos_drm_init(void) { + int ret; + DRM_DEBUG_DRIVER("%s\n", __FILE__); - return platform_driver_register(&exynos_drm_platform_driver); +#ifdef CONFIG_DRM_EXYNOS_FIMD + ret = platform_driver_register(&fimd_driver); + if (ret < 0) + goto out_fimd; +#endif + +#ifdef CONFIG_DRM_EXYNOS_HDMI + ret = platform_driver_register(&hdmi_driver); + if (ret < 0) + goto out_hdmi; + ret = platform_driver_register(&mixer_driver); + if (ret < 0) + goto out_mixer; + ret = platform_driver_register(&exynos_drm_common_hdmi_driver); + if (ret < 0) + goto out_common_hdmi; +#endif + +#ifdef CONFIG_DRM_EXYNOS_VIDI + ret = platform_driver_register(&vidi_driver); + if (ret < 0) + goto out_vidi; +#endif + + ret = platform_driver_register(&exynos_drm_platform_driver); + if (ret < 0) + goto out; + + return 0; + +out: +#ifdef CONFIG_DRM_EXYNOS_VIDI +out_vidi: + platform_driver_unregister(&vidi_driver); +#endif + +#ifdef CONFIG_DRM_EXYNOS_HDMI + platform_driver_unregister(&exynos_drm_common_hdmi_driver); +out_common_hdmi: + platform_driver_unregister(&mixer_driver); +out_mixer: + platform_driver_unregister(&hdmi_driver); +out_hdmi: +#endif + +#ifdef CONFIG_DRM_EXYNOS_FIMD + platform_driver_unregister(&fimd_driver); +out_fimd: +#endif + return ret; } static void __exit exynos_drm_exit(void) @@ -237,6 +340,20 @@ static void __exit exynos_drm_exit(void) DRM_DEBUG_DRIVER("%s\n", __FILE__); platform_driver_unregister(&exynos_drm_platform_driver); + +#ifdef CONFIG_DRM_EXYNOS_HDMI + platform_driver_unregister(&exynos_drm_common_hdmi_driver); + platform_driver_unregister(&mixer_driver); + platform_driver_unregister(&hdmi_driver); +#endif + +#ifdef CONFIG_DRM_EXYNOS_VIDI + platform_driver_unregister(&vidi_driver); +#endif + +#ifdef CONFIG_DRM_EXYNOS_FIMD + platform_driver_unregister(&fimd_driver); +#endif } module_init(exynos_drm_init); diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.h b/drivers/gpu/drm/exynos/exynos_drm_drv.h index 5e02e6e..1d81417 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_drv.h +++ b/drivers/gpu/drm/exynos/exynos_drm_drv.h @@ -32,12 +32,17 @@ #include #include "drm.h" -#define MAX_CRTC 2 +#define MAX_CRTC 3 +#define MAX_PLANE 5 +#define MAX_FB_BUFFER 4 +#define DEFAULT_ZPOS -1 struct drm_device; struct exynos_drm_overlay; struct drm_connector; +extern unsigned int drm_vblank_offdelay; + /* this enumerates display type. */ enum exynos_drm_output_type { EXYNOS_DISPLAY_TYPE_NONE, @@ -45,6 +50,8 @@ enum exynos_drm_output_type { EXYNOS_DISPLAY_TYPE_LCD, /* HDMI Interface. */ EXYNOS_DISPLAY_TYPE_HDMI, + /* Virtual Display Interface. */ + EXYNOS_DISPLAY_TYPE_VIDI, }; /* @@ -57,8 +64,8 @@ enum exynos_drm_output_type { struct exynos_drm_overlay_ops { void (*mode_set)(struct device *subdrv_dev, struct exynos_drm_overlay *overlay); - void (*commit)(struct device *subdrv_dev); - void (*disable)(struct device *subdrv_dev); + void (*commit)(struct device *subdrv_dev, int zpos); + void (*disable)(struct device *subdrv_dev, int zpos); }; /* @@ -80,9 +87,11 @@ struct exynos_drm_overlay_ops { * @scan_flag: interlace or progressive way. * (it could be DRM_MODE_FLAG_*) * @bpp: pixel size.(in bit) - * @dma_addr: bus(accessed by dma) address to the memory region allocated - * for a overlay. - * @vaddr: virtual memory addresss to this overlay. + * @pixel_format: fourcc pixel format of this overlay + * @dma_addr: array of bus(accessed by dma) address to the memory region + * allocated for a overlay. + * @vaddr: array of virtual memory addresss to this overlay. + * @zpos: order of overlay layer(z position). * @default_win: a window to be enabled. * @color_key: color key on or off. * @index_color: if using color key feature then this value would be used @@ -109,8 +118,10 @@ struct exynos_drm_overlay { unsigned int scan_flag; unsigned int bpp; unsigned int pitch; - dma_addr_t dma_addr; - void __iomem *vaddr; + uint32_t pixel_format; + dma_addr_t dma_addr[MAX_FB_BUFFER]; + void __iomem *vaddr[MAX_FB_BUFFER]; + int zpos; bool default_win; bool color_key; @@ -127,7 +138,7 @@ struct exynos_drm_overlay { * @type: one of EXYNOS_DISPLAY_TYPE_LCD and HDMI. * @is_connected: check for that display is connected or not. * @get_edid: get edid modes from display driver. - * @get_timing: get timing object from display driver. + * @get_panel: get panel object from display driver. * @check_timing: check if timing is valid or not. * @power_on: display device on or off. */ @@ -136,7 +147,7 @@ struct exynos_drm_display_ops { bool (*is_connected)(struct device *dev); int (*get_edid)(struct device *dev, struct drm_connector *connector, u8 *edid, int len); - void *(*get_timing)(struct device *dev); + void *(*get_panel)(struct device *dev); int (*check_timing)(struct device *dev, void *timing); int (*power_on)(struct device *dev, int mode); }; @@ -144,17 +155,27 @@ struct exynos_drm_display_ops { /* * Exynos drm manager ops * + * @dpms: control device power. + * @apply: set timing, vblank and overlay data to registers. + * @mode_fixup: fix mode data comparing to hw specific display mode. * @mode_set: convert drm_display_mode to hw specific display mode and * would be called by encoder->mode_set(). + * @get_max_resol: get maximum resolution to specific hardware. * @commit: set current hw specific display mode to hw. - * @disable: disable hardware specific display mode. * @enable_vblank: specific driver callback for enabling vblank interrupt. * @disable_vblank: specific driver callback for disabling vblank interrupt. */ struct exynos_drm_manager_ops { + void (*dpms)(struct device *subdrv_dev, int mode); + void (*apply)(struct device *subdrv_dev); + void (*mode_fixup)(struct device *subdrv_dev, + struct drm_connector *connector, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode); void (*mode_set)(struct device *subdrv_dev, void *mode); + void (*get_max_resol)(struct device *subdrv_dev, unsigned int *width, + unsigned int *height); void (*commit)(struct device *subdrv_dev); - void (*disable)(struct device *subdrv_dev); int (*enable_vblank)(struct device *subdrv_dev); void (*disable_vblank)(struct device *subdrv_dev); }; @@ -204,25 +225,33 @@ struct exynos_drm_private { * Exynos drm sub driver structure. * * @list: sub driver has its own list object to register to exynos drm driver. + * @dev: pointer to device object for subdrv device driver. * @drm_dev: pointer to drm_device and this pointer would be set * when sub driver calls exynos_drm_subdrv_register(). + * @manager: subdrv has its own manager to control a hardware appropriately + * and we can access a hardware drawing on this manager. * @probe: this callback would be called by exynos drm driver after * subdrv is registered to it. * @remove: this callback is used to release resources created * by probe callback. - * @manager: subdrv has its own manager to control a hardware appropriately - * and we can access a hardware drawing on this manager. + * @open: this would be called with drm device file open. + * @close: this would be called with drm device file close. * @encoder: encoder object owned by this sub driver. * @connector: connector object owned by this sub driver. */ struct exynos_drm_subdrv { struct list_head list; + struct device *dev; struct drm_device *drm_dev; + struct exynos_drm_manager *manager; int (*probe)(struct drm_device *drm_dev, struct device *dev); void (*remove)(struct drm_device *dev); + int (*open)(struct drm_device *drm_dev, struct device *dev, + struct drm_file *file); + void (*close)(struct drm_device *drm_dev, struct device *dev, + struct drm_file *file); - struct exynos_drm_manager manager; struct drm_encoder *encoder; struct drm_connector *connector; }; @@ -243,15 +272,19 @@ int exynos_drm_device_unregister(struct drm_device *dev); * this function would be called by sub drivers such as display controller * or hdmi driver to register this sub driver object to exynos drm driver * and when a sub driver is registered to exynos drm driver a probe callback - * of the sub driver is called and creates its own encoder and connector - * and then fb helper and drm mode group would be re-initialized. + * of the sub driver is called and creates its own encoder and connector. */ int exynos_drm_subdrv_register(struct exynos_drm_subdrv *drm_subdrv); -/* - * this function removes subdrv list from exynos drm driver and fb helper - * and drm mode group would be re-initialized. - */ +/* this function removes subdrv list from exynos drm driver */ int exynos_drm_subdrv_unregister(struct exynos_drm_subdrv *drm_subdrv); +int exynos_drm_subdrv_open(struct drm_device *dev, struct drm_file *file); +void exynos_drm_subdrv_close(struct drm_device *dev, struct drm_file *file); + +extern struct platform_driver fimd_driver; +extern struct platform_driver hdmi_driver; +extern struct platform_driver mixer_driver; +extern struct platform_driver exynos_drm_common_hdmi_driver; +extern struct platform_driver vidi_driver; #endif diff --git a/drivers/gpu/drm/exynos/exynos_drm_encoder.c b/drivers/gpu/drm/exynos/exynos_drm_encoder.c index 1530614..6e9ac7b 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_encoder.c +++ b/drivers/gpu/drm/exynos/exynos_drm_encoder.c @@ -42,49 +42,68 @@ * @drm_encoder: encoder object. * @manager: specific encoder has its own manager to control a hardware * appropriately and we can access a hardware drawing on this manager. + * @dpms: store the encoder dpms value. */ struct exynos_drm_encoder { struct drm_encoder drm_encoder; struct exynos_drm_manager *manager; + int dpms; }; -static void exynos_drm_encoder_dpms(struct drm_encoder *encoder, int mode) +static void exynos_drm_display_power(struct drm_encoder *encoder, int mode) { struct drm_device *dev = encoder->dev; struct drm_connector *connector; struct exynos_drm_manager *manager = exynos_drm_get_manager(encoder); + + list_for_each_entry(connector, &dev->mode_config.connector_list, head) { + if (connector->encoder == encoder) { + struct exynos_drm_display_ops *display_ops = + manager->display_ops; + + DRM_DEBUG_KMS("connector[%d] dpms[%d]\n", + connector->base.id, mode); + if (display_ops && display_ops->power_on) + display_ops->power_on(manager->dev, mode); + } + } +} + +static void exynos_drm_encoder_dpms(struct drm_encoder *encoder, int mode) +{ + struct drm_device *dev = encoder->dev; + struct exynos_drm_manager *manager = exynos_drm_get_manager(encoder); struct exynos_drm_manager_ops *manager_ops = manager->ops; + struct exynos_drm_encoder *exynos_encoder = to_exynos_encoder(encoder); DRM_DEBUG_KMS("%s, encoder dpms: %d\n", __FILE__, mode); + if (exynos_encoder->dpms == mode) { + DRM_DEBUG_KMS("desired dpms mode is same as previous one.\n"); + return; + } + + mutex_lock(&dev->struct_mutex); + switch (mode) { case DRM_MODE_DPMS_ON: - if (manager_ops && manager_ops->commit) - manager_ops->commit(manager->dev); + if (manager_ops && manager_ops->apply) + manager_ops->apply(manager->dev); + exynos_drm_display_power(encoder, mode); + exynos_encoder->dpms = mode; break; case DRM_MODE_DPMS_STANDBY: case DRM_MODE_DPMS_SUSPEND: case DRM_MODE_DPMS_OFF: - /* TODO */ - if (manager_ops && manager_ops->disable) - manager_ops->disable(manager->dev); + exynos_drm_display_power(encoder, mode); + exynos_encoder->dpms = mode; break; default: DRM_ERROR("unspecified mode %d\n", mode); break; } - list_for_each_entry(connector, &dev->mode_config.connector_list, head) { - if (connector->encoder == encoder) { - struct exynos_drm_display_ops *display_ops = - manager->display_ops; - - DRM_DEBUG_KMS("connector[%d] dpms[%d]\n", - connector->base.id, mode); - if (display_ops && display_ops->power_on) - display_ops->power_on(manager->dev, mode); - } - } + mutex_unlock(&dev->struct_mutex); } static bool @@ -92,9 +111,19 @@ exynos_drm_encoder_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) { + struct drm_device *dev = encoder->dev; + struct drm_connector *connector; + struct exynos_drm_manager *manager = exynos_drm_get_manager(encoder); + struct exynos_drm_manager_ops *manager_ops = manager->ops; + DRM_DEBUG_KMS("%s\n", __FILE__); - /* drm framework doesn't check NULL. */ + list_for_each_entry(connector, &dev->mode_config.connector_list, head) { + if (connector->encoder == encoder) + if (manager_ops && manager_ops->mode_fixup) + manager_ops->mode_fixup(manager->dev, connector, + mode, adjusted_mode); + } return true; } @@ -113,12 +142,11 @@ static void exynos_drm_encoder_mode_set(struct drm_encoder *encoder, DRM_DEBUG_KMS("%s\n", __FILE__); - mode = adjusted_mode; - list_for_each_entry(connector, &dev->mode_config.connector_list, head) { if (connector->encoder == encoder) { if (manager_ops && manager_ops->mode_set) - manager_ops->mode_set(manager->dev, mode); + manager_ops->mode_set(manager->dev, + adjusted_mode); if (overlay_ops && overlay_ops->mode_set) overlay_ops->mode_set(manager->dev, overlay); @@ -169,7 +197,6 @@ static void exynos_drm_encoder_destroy(struct drm_encoder *encoder) exynos_encoder->manager->pipe = -1; drm_encoder_cleanup(encoder); - encoder->dev->mode_config.num_encoder--; kfree(exynos_encoder); } @@ -177,6 +204,41 @@ static struct drm_encoder_funcs exynos_encoder_funcs = { .destroy = exynos_drm_encoder_destroy, }; +static unsigned int exynos_drm_encoder_clones(struct drm_encoder *encoder) +{ + struct drm_encoder *clone; + struct drm_device *dev = encoder->dev; + struct exynos_drm_encoder *exynos_encoder = to_exynos_encoder(encoder); + struct exynos_drm_display_ops *display_ops = + exynos_encoder->manager->display_ops; + unsigned int clone_mask = 0; + int cnt = 0; + + list_for_each_entry(clone, &dev->mode_config.encoder_list, head) { + switch (display_ops->type) { + case EXYNOS_DISPLAY_TYPE_LCD: + case EXYNOS_DISPLAY_TYPE_HDMI: + case EXYNOS_DISPLAY_TYPE_VIDI: + clone_mask |= (1 << (cnt++)); + break; + default: + continue; + } + } + + return clone_mask; +} + +void exynos_drm_encoder_setup(struct drm_device *dev) +{ + struct drm_encoder *encoder; + + DRM_DEBUG_KMS("%s\n", __FILE__); + + list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) + encoder->possible_clones = exynos_drm_encoder_clones(encoder); +} + struct drm_encoder * exynos_drm_encoder_create(struct drm_device *dev, struct exynos_drm_manager *manager, @@ -199,6 +261,7 @@ exynos_drm_encoder_create(struct drm_device *dev, return NULL; } + exynos_encoder->dpms = DRM_MODE_DPMS_OFF; exynos_encoder->manager = manager; encoder = &exynos_encoder->drm_encoder; encoder->possible_crtcs = possible_crtcs; @@ -275,12 +338,27 @@ void exynos_drm_disable_vblank(struct drm_encoder *encoder, void *data) manager_ops->disable_vblank(manager->dev); } -void exynos_drm_encoder_crtc_commit(struct drm_encoder *encoder, void *data) +void exynos_drm_encoder_crtc_plane_commit(struct drm_encoder *encoder, + void *data) { struct exynos_drm_manager *manager = to_exynos_encoder(encoder)->manager; struct exynos_drm_overlay_ops *overlay_ops = manager->overlay_ops; + int zpos = DEFAULT_ZPOS; + + if (data) + zpos = *(int *)data; + + if (overlay_ops && overlay_ops->commit) + overlay_ops->commit(manager->dev, zpos); +} + +void exynos_drm_encoder_crtc_commit(struct drm_encoder *encoder, void *data) +{ + struct exynos_drm_manager *manager = + to_exynos_encoder(encoder)->manager; int crtc = *(int *)data; + int zpos = DEFAULT_ZPOS; DRM_DEBUG_KMS("%s\n", __FILE__); @@ -290,8 +368,53 @@ void exynos_drm_encoder_crtc_commit(struct drm_encoder *encoder, void *data) */ manager->pipe = crtc; - if (overlay_ops && overlay_ops->commit) - overlay_ops->commit(manager->dev); + exynos_drm_encoder_crtc_plane_commit(encoder, &zpos); +} + +void exynos_drm_encoder_dpms_from_crtc(struct drm_encoder *encoder, void *data) +{ + struct exynos_drm_encoder *exynos_encoder = to_exynos_encoder(encoder); + int mode = *(int *)data; + + DRM_DEBUG_KMS("%s\n", __FILE__); + + exynos_drm_encoder_dpms(encoder, mode); + + exynos_encoder->dpms = mode; +} + +void exynos_drm_encoder_crtc_dpms(struct drm_encoder *encoder, void *data) +{ + struct drm_device *dev = encoder->dev; + struct exynos_drm_encoder *exynos_encoder = to_exynos_encoder(encoder); + struct exynos_drm_manager *manager = exynos_encoder->manager; + struct exynos_drm_manager_ops *manager_ops = manager->ops; + struct drm_connector *connector; + int mode = *(int *)data; + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (manager_ops && manager_ops->dpms) + manager_ops->dpms(manager->dev, mode); + + /* + * set current dpms mode to the connector connected to + * current encoder. connector->dpms would be checked + * at drm_helper_connector_dpms() + */ + list_for_each_entry(connector, &dev->mode_config.connector_list, head) + if (connector->encoder == encoder) + connector->dpms = mode; + + /* + * if this condition is ok then it means that the crtc is already + * detached from encoder and last function for detaching is properly + * done, so clear pipe from manager to prevent repeated call. + */ + if (mode > DRM_MODE_DPMS_ON) { + if (!encoder->crtc) + manager->pipe = -1; + } } void exynos_drm_encoder_crtc_mode_set(struct drm_encoder *encoder, void *data) @@ -310,23 +433,13 @@ void exynos_drm_encoder_crtc_disable(struct drm_encoder *encoder, void *data) struct exynos_drm_manager *manager = to_exynos_encoder(encoder)->manager; struct exynos_drm_overlay_ops *overlay_ops = manager->overlay_ops; + int zpos = DEFAULT_ZPOS; DRM_DEBUG_KMS("\n"); - if (overlay_ops && overlay_ops->disable) - overlay_ops->disable(manager->dev); + if (data) + zpos = *(int *)data; - /* - * crtc is already detached from encoder and last - * function for detaching is properly done, so - * clear pipe from manager to prevent repeated call - */ - if (!encoder->crtc) - manager->pipe = -1; + if (overlay_ops && overlay_ops->disable) + overlay_ops->disable(manager->dev, zpos); } - -MODULE_AUTHOR("Inki Dae "); -MODULE_AUTHOR("Joonyoung Shim "); -MODULE_AUTHOR("Seung-Woo Kim "); -MODULE_DESCRIPTION("Samsung SoC DRM Encoder Driver"); -MODULE_LICENSE("GPL"); diff --git a/drivers/gpu/drm/exynos/exynos_drm_encoder.h b/drivers/gpu/drm/exynos/exynos_drm_encoder.h index a22acfb..eb7d231 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_encoder.h +++ b/drivers/gpu/drm/exynos/exynos_drm_encoder.h @@ -30,6 +30,7 @@ struct exynos_drm_manager; +void exynos_drm_encoder_setup(struct drm_device *dev); struct drm_encoder *exynos_drm_encoder_create(struct drm_device *dev, struct exynos_drm_manager *mgr, unsigned int possible_crtcs); @@ -39,7 +40,12 @@ void exynos_drm_fn_encoder(struct drm_crtc *crtc, void *data, void (*fn)(struct drm_encoder *, void *)); void exynos_drm_enable_vblank(struct drm_encoder *encoder, void *data); void exynos_drm_disable_vblank(struct drm_encoder *encoder, void *data); +void exynos_drm_encoder_crtc_plane_commit(struct drm_encoder *encoder, + void *data); void exynos_drm_encoder_crtc_commit(struct drm_encoder *encoder, void *data); +void exynos_drm_encoder_dpms_from_crtc(struct drm_encoder *encoder, + void *data); +void exynos_drm_encoder_crtc_dpms(struct drm_encoder *encoder, void *data); void exynos_drm_encoder_crtc_mode_set(struct drm_encoder *encoder, void *data); void exynos_drm_encoder_crtc_disable(struct drm_encoder *encoder, void *data); diff --git a/drivers/gpu/drm/exynos/exynos_drm_fb.c b/drivers/gpu/drm/exynos/exynos_drm_fb.c index 5bf4a1a..c38c8f4 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_fb.c +++ b/drivers/gpu/drm/exynos/exynos_drm_fb.c @@ -33,7 +33,6 @@ #include "exynos_drm_drv.h" #include "exynos_drm_fb.h" -#include "exynos_drm_buf.h" #include "exynos_drm_gem.h" #define to_exynos_fb(x) container_of(x, struct exynos_drm_fb, fb) @@ -42,15 +41,11 @@ * exynos specific framebuffer structure. * * @fb: drm framebuffer obejct. - * @exynos_gem_obj: exynos specific gem object containing a gem object. - * @buffer: pointer to exynos_drm_gem_buffer object. - * - contain the memory information to memory region allocated - * at default framebuffer creation. + * @exynos_gem_obj: array of exynos specific gem object containing a gem object. */ struct exynos_drm_fb { struct drm_framebuffer fb; - struct exynos_drm_gem_obj *exynos_gem_obj; - struct exynos_drm_gem_buf *buffer; + struct exynos_drm_gem_obj *exynos_gem_obj[MAX_FB_BUFFER]; }; static void exynos_drm_fb_destroy(struct drm_framebuffer *fb) @@ -61,13 +56,6 @@ static void exynos_drm_fb_destroy(struct drm_framebuffer *fb) drm_framebuffer_cleanup(fb); - /* - * default framebuffer has no gem object so - * a buffer of the default framebuffer should be released at here. - */ - if (!exynos_fb->exynos_gem_obj && exynos_fb->buffer) - exynos_drm_buf_destroy(fb->dev, exynos_fb->buffer); - kfree(exynos_fb); exynos_fb = NULL; } @@ -81,7 +69,7 @@ static int exynos_drm_fb_create_handle(struct drm_framebuffer *fb, DRM_DEBUG_KMS("%s\n", __FILE__); return drm_gem_handle_create(file_priv, - &exynos_fb->exynos_gem_obj->base, handle); + &exynos_fb->exynos_gem_obj[0]->base, handle); } static int exynos_drm_fb_dirty(struct drm_framebuffer *fb, @@ -102,134 +90,88 @@ static struct drm_framebuffer_funcs exynos_drm_fb_funcs = { .dirty = exynos_drm_fb_dirty, }; -static struct drm_framebuffer * -exynos_drm_fb_init(struct drm_file *file_priv, struct drm_device *dev, - struct drm_mode_fb_cmd *mode_cmd) +struct drm_framebuffer * +exynos_drm_framebuffer_init(struct drm_device *dev, + struct drm_mode_fb_cmd2 *mode_cmd, + struct drm_gem_object *obj) { struct exynos_drm_fb *exynos_fb; - struct drm_framebuffer *fb; - struct exynos_drm_gem_obj *exynos_gem_obj = NULL; - struct drm_gem_object *obj; - unsigned int size; int ret; - DRM_DEBUG_KMS("%s\n", __FILE__); - - mode_cmd->pitch = max(mode_cmd->pitch, - mode_cmd->width * (mode_cmd->bpp >> 3)); - - DRM_LOG_KMS("drm fb create(%dx%d)\n", - mode_cmd->width, mode_cmd->height); - exynos_fb = kzalloc(sizeof(*exynos_fb), GFP_KERNEL); if (!exynos_fb) { - DRM_ERROR("failed to allocate exynos drm framebuffer.\n"); + DRM_ERROR("failed to allocate exynos drm framebuffer\n"); return ERR_PTR(-ENOMEM); } - fb = &exynos_fb->fb; - ret = drm_framebuffer_init(dev, fb, &exynos_drm_fb_funcs); + ret = drm_framebuffer_init(dev, &exynos_fb->fb, &exynos_drm_fb_funcs); if (ret) { - DRM_ERROR("failed to initialize framebuffer.\n"); - goto err_init; + DRM_ERROR("failed to initialize framebuffer\n"); + return ERR_PTR(ret); } - DRM_LOG_KMS("create: fb id: %d\n", fb->base.id); + drm_helper_mode_fill_fb_struct(&exynos_fb->fb, mode_cmd); + exynos_fb->exynos_gem_obj[0] = to_exynos_gem_obj(obj); - size = mode_cmd->pitch * mode_cmd->height; + return &exynos_fb->fb; +} - /* - * mode_cmd->handle could be NULL at booting time or - * with user request. if NULL, a new buffer or a gem object - * would be allocated. - */ - if (!mode_cmd->handle) { - if (!file_priv) { - struct exynos_drm_gem_buf *buffer; - - /* - * in case that file_priv is NULL, it allocates - * only buffer and this buffer would be used - * for default framebuffer. - */ - buffer = exynos_drm_buf_create(dev, size); - if (IS_ERR(buffer)) { - ret = PTR_ERR(buffer); - goto err_buffer; - } - - exynos_fb->buffer = buffer; - - DRM_LOG_KMS("default: dma_addr = 0x%lx, size = 0x%x\n", - (unsigned long)buffer->dma_addr, size); - - goto out; - } else { - exynos_gem_obj = exynos_drm_gem_create(dev, file_priv, - &mode_cmd->handle, - size); - if (IS_ERR(exynos_gem_obj)) { - ret = PTR_ERR(exynos_gem_obj); - goto err_buffer; - } - } - } else { - obj = drm_gem_object_lookup(dev, file_priv, mode_cmd->handle); - if (!obj) { - DRM_ERROR("failed to lookup gem object.\n"); - goto err_buffer; - } +static struct drm_framebuffer * +exynos_user_fb_create(struct drm_device *dev, struct drm_file *file_priv, + struct drm_mode_fb_cmd2 *mode_cmd) +{ + struct drm_gem_object *obj; + struct drm_framebuffer *fb; + struct exynos_drm_fb *exynos_fb; + int nr; + int i; - exynos_gem_obj = to_exynos_gem_obj(obj); + DRM_DEBUG_KMS("%s\n", __FILE__); - drm_gem_object_unreference_unlocked(obj); + obj = drm_gem_object_lookup(dev, file_priv, mode_cmd->handles[0]); + if (!obj) { + DRM_ERROR("failed to lookup gem object\n"); + return ERR_PTR(-ENOENT); } - /* - * if got a exynos_gem_obj from either a handle or - * a new creation then exynos_fb->exynos_gem_obj is NULL - * so that default framebuffer has no its own gem object, - * only its own buffer object. - */ - exynos_fb->buffer = exynos_gem_obj->buffer; - - DRM_LOG_KMS("dma_addr = 0x%lx, size = 0x%x, gem object = 0x%x\n", - (unsigned long)exynos_fb->buffer->dma_addr, size, - (unsigned int)&exynos_gem_obj->base); + drm_gem_object_unreference_unlocked(obj); -out: - exynos_fb->exynos_gem_obj = exynos_gem_obj; + fb = exynos_drm_framebuffer_init(dev, mode_cmd, obj); + if (IS_ERR(fb)) + return fb; - drm_helper_mode_fill_fb_struct(fb, mode_cmd); + exynos_fb = to_exynos_fb(fb); + nr = exynos_drm_format_num_buffers(fb->pixel_format); - return fb; - -err_buffer: - drm_framebuffer_cleanup(fb); - -err_init: - kfree(exynos_fb); + for (i = 1; i < nr; i++) { + obj = drm_gem_object_lookup(dev, file_priv, + mode_cmd->handles[i]); + if (!obj) { + DRM_ERROR("failed to lookup gem object\n"); + exynos_drm_fb_destroy(fb); + return ERR_PTR(-ENOENT); + } - return ERR_PTR(ret); -} + drm_gem_object_unreference_unlocked(obj); -struct drm_framebuffer *exynos_drm_fb_create(struct drm_device *dev, - struct drm_file *file_priv, - struct drm_mode_fb_cmd *mode_cmd) -{ - DRM_DEBUG_KMS("%s\n", __FILE__); + exynos_fb->exynos_gem_obj[i] = to_exynos_gem_obj(obj); + } - return exynos_drm_fb_init(file_priv, dev, mode_cmd); + return fb; } -struct exynos_drm_gem_buf *exynos_drm_fb_get_buf(struct drm_framebuffer *fb) +struct exynos_drm_gem_buf *exynos_drm_fb_buffer(struct drm_framebuffer *fb, + int index) { struct exynos_drm_fb *exynos_fb = to_exynos_fb(fb); struct exynos_drm_gem_buf *buffer; DRM_DEBUG_KMS("%s\n", __FILE__); - buffer = exynos_fb->buffer; + if (index >= MAX_FB_BUFFER) + return NULL; + + buffer = exynos_fb->exynos_gem_obj[index]->buffer; if (!buffer) return NULL; @@ -250,7 +192,7 @@ static void exynos_drm_output_poll_changed(struct drm_device *dev) } static struct drm_mode_config_funcs exynos_drm_mode_config_funcs = { - .fb_create = exynos_drm_fb_create, + .fb_create = exynos_user_fb_create, .output_poll_changed = exynos_drm_output_poll_changed, }; @@ -269,9 +211,3 @@ void exynos_drm_mode_config_init(struct drm_device *dev) dev->mode_config.funcs = &exynos_drm_mode_config_funcs; } - -MODULE_AUTHOR("Inki Dae "); -MODULE_AUTHOR("Joonyoung Shim "); -MODULE_AUTHOR("Seung-Woo Kim "); -MODULE_DESCRIPTION("Samsung SoC DRM FB Driver"); -MODULE_LICENSE("GPL"); diff --git a/drivers/gpu/drm/exynos/exynos_drm_fb.h b/drivers/gpu/drm/exynos/exynos_drm_fb.h index eb35931..3ecb30d 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_fb.h +++ b/drivers/gpu/drm/exynos/exynos_drm_fb.h @@ -28,9 +28,27 @@ #ifndef _EXYNOS_DRM_FB_H_ #define _EXYNOS_DRM_FB_H -struct drm_framebuffer *exynos_drm_fb_create(struct drm_device *dev, - struct drm_file *filp, - struct drm_mode_fb_cmd *mode_cmd); +static inline int exynos_drm_format_num_buffers(uint32_t format) +{ + switch (format) { + case DRM_FORMAT_NV12M: + case DRM_FORMAT_NV12MT: + return 2; + case DRM_FORMAT_YUV420M: + return 3; + default: + return 1; + } +} + +struct drm_framebuffer * +exynos_drm_framebuffer_init(struct drm_device *dev, + struct drm_mode_fb_cmd2 *mode_cmd, + struct drm_gem_object *obj); + +/* get memory information of a drm framebuffer */ +struct exynos_drm_gem_buf *exynos_drm_fb_buffer(struct drm_framebuffer *fb, + int index); void exynos_drm_mode_config_init(struct drm_device *dev); diff --git a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c index 836f410..d5586cc 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c +++ b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c @@ -34,7 +34,6 @@ #include "exynos_drm_drv.h" #include "exynos_drm_fb.h" #include "exynos_drm_gem.h" -#include "exynos_drm_buf.h" #define MAX_CONNECTOR 4 #define PREFERRED_BPP 32 @@ -43,43 +42,17 @@ drm_fb_helper) struct exynos_drm_fbdev { - struct drm_fb_helper drm_fb_helper; - struct drm_framebuffer *fb; + struct drm_fb_helper drm_fb_helper; + struct exynos_drm_gem_obj *exynos_gem_obj; }; -static int exynos_drm_fbdev_set_par(struct fb_info *info) -{ - struct fb_var_screeninfo *var = &info->var; - - switch (var->bits_per_pixel) { - case 32: - case 24: - case 18: - case 16: - case 12: - info->fix.visual = FB_VISUAL_TRUECOLOR; - break; - case 1: - info->fix.visual = FB_VISUAL_MONO01; - break; - default: - info->fix.visual = FB_VISUAL_PSEUDOCOLOR; - break; - } - - info->fix.line_length = (var->xres_virtual * var->bits_per_pixel) / 8; - - return drm_fb_helper_set_par(info); -} - - static struct fb_ops exynos_drm_fb_ops = { .owner = THIS_MODULE, .fb_fillrect = cfb_fillrect, .fb_copyarea = cfb_copyarea, .fb_imageblit = cfb_imageblit, .fb_check_var = drm_fb_helper_check_var, - .fb_set_par = exynos_drm_fbdev_set_par, + .fb_set_par = drm_fb_helper_set_par, .fb_blank = drm_fb_helper_blank, .fb_pan_display = drm_fb_helper_pan_display, .fb_setcmap = drm_fb_helper_setcmap, @@ -90,26 +63,24 @@ static int exynos_drm_fbdev_update(struct drm_fb_helper *helper, { struct fb_info *fbi = helper->fbdev; struct drm_device *dev = helper->dev; - struct exynos_drm_fbdev *exynos_fb = to_exynos_fbdev(helper); struct exynos_drm_gem_buf *buffer; unsigned int size = fb->width * fb->height * (fb->bits_per_pixel >> 3); unsigned long offset; DRM_DEBUG_KMS("%s\n", __FILE__); - exynos_fb->fb = fb; - - drm_fb_helper_fill_fix(fbi, fb->pitch, fb->depth); + drm_fb_helper_fill_fix(fbi, fb->pitches[0], fb->depth); drm_fb_helper_fill_var(fbi, helper, fb->width, fb->height); - buffer = exynos_drm_fb_get_buf(fb); + /* RGB formats use only one buffer */ + buffer = exynos_drm_fb_buffer(fb, 0); if (!buffer) { DRM_LOG_KMS("buffer is null.\n"); return -EFAULT; } offset = fbi->var.xoffset * (fb->bits_per_pixel >> 3); - offset += fbi->var.yoffset * fb->pitch; + offset += fbi->var.yoffset * fb->pitches[0]; dev->mode_config.fb_base = (resource_size_t)buffer->dma_addr; fbi->screen_base = buffer->kvaddr + offset; @@ -124,10 +95,12 @@ static int exynos_drm_fbdev_create(struct drm_fb_helper *helper, struct drm_fb_helper_surface_size *sizes) { struct exynos_drm_fbdev *exynos_fbdev = to_exynos_fbdev(helper); + struct exynos_drm_gem_obj *exynos_gem_obj; struct drm_device *dev = helper->dev; struct fb_info *fbi; - struct drm_mode_fb_cmd mode_cmd = { 0 }; + struct drm_mode_fb_cmd2 mode_cmd = { 0 }; struct platform_device *pdev = dev->platformdev; + unsigned long size; int ret; DRM_DEBUG_KMS("%s\n", __FILE__); @@ -138,8 +111,9 @@ static int exynos_drm_fbdev_create(struct drm_fb_helper *helper, mode_cmd.width = sizes->surface_width; mode_cmd.height = sizes->surface_height; - mode_cmd.bpp = sizes->surface_bpp; - mode_cmd.depth = sizes->surface_depth; + mode_cmd.pitches[0] = sizes->surface_width * (sizes->surface_bpp >> 3); + mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp, + sizes->surface_depth); mutex_lock(&dev->struct_mutex); @@ -150,14 +124,25 @@ static int exynos_drm_fbdev_create(struct drm_fb_helper *helper, goto out; } - exynos_fbdev->fb = exynos_drm_fb_create(dev, NULL, &mode_cmd); - if (IS_ERR_OR_NULL(exynos_fbdev->fb)) { + size = mode_cmd.pitches[0] * mode_cmd.height; + + /* 0 means to allocate physically continuous memory */ + exynos_gem_obj = exynos_drm_gem_create(dev, 0, size); + if (IS_ERR(exynos_gem_obj)) { + ret = PTR_ERR(exynos_gem_obj); + goto out; + } + + exynos_fbdev->exynos_gem_obj = exynos_gem_obj; + + helper->fb = exynos_drm_framebuffer_init(dev, &mode_cmd, + &exynos_gem_obj->base); + if (IS_ERR_OR_NULL(helper->fb)) { DRM_ERROR("failed to create drm framebuffer.\n"); - ret = PTR_ERR(exynos_fbdev->fb); + ret = PTR_ERR(helper->fb); goto out; } - helper->fb = exynos_fbdev->fb; helper->fbdev = fbi; fbi->par = helper; @@ -171,8 +156,10 @@ static int exynos_drm_fbdev_create(struct drm_fb_helper *helper, } ret = exynos_drm_fbdev_update(helper, helper->fb); - if (ret < 0) + if (ret < 0) { fb_dealloc_cmap(&fbi->cmap); + goto out; + } /* * if failed, all resources allocated above would be released by @@ -184,58 +171,6 @@ out: return ret; } -static bool -exynos_drm_fbdev_is_samefb(struct drm_framebuffer *fb, - struct drm_fb_helper_surface_size *sizes) -{ - if (fb->width != sizes->surface_width) - return false; - if (fb->height != sizes->surface_height) - return false; - if (fb->bits_per_pixel != sizes->surface_bpp) - return false; - if (fb->depth != sizes->surface_depth) - return false; - - return true; -} - -static int exynos_drm_fbdev_recreate(struct drm_fb_helper *helper, - struct drm_fb_helper_surface_size *sizes) -{ - struct drm_device *dev = helper->dev; - struct exynos_drm_fbdev *exynos_fbdev = to_exynos_fbdev(helper); - struct drm_framebuffer *fb = exynos_fbdev->fb; - struct drm_mode_fb_cmd mode_cmd = { 0 }; - - DRM_DEBUG_KMS("%s\n", __FILE__); - - if (helper->fb != fb) { - DRM_ERROR("drm framebuffer is different\n"); - return -EINVAL; - } - - if (exynos_drm_fbdev_is_samefb(fb, sizes)) - return 0; - - mode_cmd.width = sizes->surface_width; - mode_cmd.height = sizes->surface_height; - mode_cmd.bpp = sizes->surface_bpp; - mode_cmd.depth = sizes->surface_depth; - - if (fb->funcs->destroy) - fb->funcs->destroy(fb); - - exynos_fbdev->fb = exynos_drm_fb_create(dev, NULL, &mode_cmd); - if (IS_ERR(exynos_fbdev->fb)) { - DRM_ERROR("failed to allocate fb.\n"); - return PTR_ERR(exynos_fbdev->fb); - } - - helper->fb = exynos_fbdev->fb; - return exynos_drm_fbdev_update(helper, helper->fb); -} - static int exynos_drm_fbdev_probe(struct drm_fb_helper *helper, struct drm_fb_helper_surface_size *sizes) { @@ -243,6 +178,10 @@ static int exynos_drm_fbdev_probe(struct drm_fb_helper *helper, DRM_DEBUG_KMS("%s\n", __FILE__); + /* + * with !helper->fb, it means that this funcion is called first time + * and after that, the helper->fb would be used as clone mode. + */ if (!helper->fb) { ret = exynos_drm_fbdev_create(helper, sizes); if (ret < 0) { @@ -255,12 +194,6 @@ static int exynos_drm_fbdev_probe(struct drm_fb_helper *helper, * because register_framebuffer() should be called. */ ret = 1; - } else { - ret = exynos_drm_fbdev_recreate(helper, sizes); - if (ret < 0) { - DRM_ERROR("failed to reconfigure fbdev\n"); - return ret; - } } return ret; @@ -366,6 +299,9 @@ void exynos_drm_fbdev_fini(struct drm_device *dev) fbdev = to_exynos_fbdev(private->fb_helper); + if (fbdev->exynos_gem_obj) + exynos_drm_gem_destroy(fbdev->exynos_gem_obj); + exynos_drm_fbdev_destroy(dev, private->fb_helper); kfree(fbdev); private->fb_helper = NULL; @@ -380,89 +316,3 @@ void exynos_drm_fbdev_restore_mode(struct drm_device *dev) drm_fb_helper_restore_fbdev_mode(private->fb_helper); } - -int exynos_drm_fbdev_reinit(struct drm_device *dev) -{ - struct exynos_drm_private *private = dev->dev_private; - struct drm_fb_helper *fb_helper; - int ret; - - if (!private) - return -EINVAL; - - /* - * if all sub drivers were unloaded then num_connector is 0 - * so at this time, the framebuffers also should be destroyed. - */ - if (!dev->mode_config.num_connector) { - exynos_drm_fbdev_fini(dev); - return 0; - } - - fb_helper = private->fb_helper; - - if (fb_helper) { - struct list_head temp_list; - - INIT_LIST_HEAD(&temp_list); - - /* - * fb_helper is reintialized but kernel fb is reused - * so kernel_fb_list need to be backuped and restored - */ - if (!list_empty(&fb_helper->kernel_fb_list)) - list_replace_init(&fb_helper->kernel_fb_list, - &temp_list); - - drm_fb_helper_fini(fb_helper); - - ret = drm_fb_helper_init(dev, fb_helper, - dev->mode_config.num_crtc, MAX_CONNECTOR); - if (ret < 0) { - DRM_ERROR("failed to initialize drm fb helper\n"); - return ret; - } - - if (!list_empty(&temp_list)) - list_replace(&temp_list, &fb_helper->kernel_fb_list); - - ret = drm_fb_helper_single_add_all_connectors(fb_helper); - if (ret < 0) { - DRM_ERROR("failed to add fb helper to connectors\n"); - goto err; - } - - ret = drm_fb_helper_initial_config(fb_helper, PREFERRED_BPP); - if (ret < 0) { - DRM_ERROR("failed to set up hw configuration.\n"); - goto err; - } - } else { - /* - * if drm_load() failed whem drm load() was called prior - * to specific drivers, fb_helper must be NULL and so - * this fuction should be called again to re-initialize and - * re-configure the fb helper. it means that this function - * has been called by the specific drivers. - */ - ret = exynos_drm_fbdev_init(dev); - } - - return ret; - -err: - /* - * if drm_load() failed when drm load() was called prior - * to specific drivers, the fb_helper must be NULL and so check it. - */ - if (fb_helper) - drm_fb_helper_fini(fb_helper); - - return ret; -} - -MODULE_AUTHOR("Inki Dae "); -MODULE_AUTHOR("Joonyoung Shim "); -MODULE_AUTHOR("Seung-Woo Kim "); -MODULE_DESCRIPTION("Samsung SoC DRM FBDEV Driver"); -MODULE_LICENSE("GPL"); diff --git a/drivers/gpu/drm/exynos/exynos_drm_fimd.c b/drivers/gpu/drm/exynos/exynos_drm_fimd.c index db3b3d9..29fdbfe 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_fimd.c +++ b/drivers/gpu/drm/exynos/exynos_drm_fimd.c @@ -17,6 +17,7 @@ #include #include #include +#include #include #include @@ -68,6 +69,7 @@ struct fimd_win_data { void __iomem *vaddr; unsigned int buf_offsize; unsigned int line_size; /* bytes */ + bool enabled; }; struct fimd_context { @@ -84,8 +86,10 @@ struct fimd_context { unsigned long irq_flags; u32 vidcon0; u32 vidcon1; + bool suspended; + struct mutex lock; - struct fb_videomode *timing; + struct exynos_drm_panel_info *panel; }; static bool fimd_display_is_connected(struct device *dev) @@ -97,13 +101,13 @@ static bool fimd_display_is_connected(struct device *dev) return true; } -static void *fimd_get_timing(struct device *dev) +static void *fimd_get_panel(struct device *dev) { struct fimd_context *ctx = get_fimd_context(dev); DRM_DEBUG_KMS("%s\n", __FILE__); - return ctx->timing; + return ctx->panel; } static int fimd_check_timing(struct device *dev, void *timing) @@ -119,7 +123,7 @@ static int fimd_display_power_on(struct device *dev, int mode) { DRM_DEBUG_KMS("%s\n", __FILE__); - /* TODO. */ + /* TODO */ return 0; } @@ -127,17 +131,75 @@ static int fimd_display_power_on(struct device *dev, int mode) static struct exynos_drm_display_ops fimd_display_ops = { .type = EXYNOS_DISPLAY_TYPE_LCD, .is_connected = fimd_display_is_connected, - .get_timing = fimd_get_timing, + .get_panel = fimd_get_panel, .check_timing = fimd_check_timing, .power_on = fimd_display_power_on, }; +static void fimd_dpms(struct device *subdrv_dev, int mode) +{ + struct fimd_context *ctx = get_fimd_context(subdrv_dev); + + DRM_DEBUG_KMS("%s, %d\n", __FILE__, mode); + + mutex_lock(&ctx->lock); + + switch (mode) { + case DRM_MODE_DPMS_ON: + /* + * enable fimd hardware only if suspended status. + * + * P.S. fimd_dpms function would be called at booting time so + * clk_enable could be called double time. + */ + if (ctx->suspended) + pm_runtime_get_sync(subdrv_dev); + break; + case DRM_MODE_DPMS_STANDBY: + case DRM_MODE_DPMS_SUSPEND: + case DRM_MODE_DPMS_OFF: + if (!ctx->suspended) + pm_runtime_put_sync(subdrv_dev); + break; + default: + DRM_DEBUG_KMS("unspecified mode %d\n", mode); + break; + } + + mutex_unlock(&ctx->lock); +} + +static void fimd_apply(struct device *subdrv_dev) +{ + struct fimd_context *ctx = get_fimd_context(subdrv_dev); + struct exynos_drm_manager *mgr = ctx->subdrv.manager; + struct exynos_drm_manager_ops *mgr_ops = mgr->ops; + struct exynos_drm_overlay_ops *ovl_ops = mgr->overlay_ops; + struct fimd_win_data *win_data; + int i; + + DRM_DEBUG_KMS("%s\n", __FILE__); + + for (i = 0; i < WINDOWS_NR; i++) { + win_data = &ctx->win_data[i]; + if (win_data->enabled && (ovl_ops && ovl_ops->commit)) + ovl_ops->commit(subdrv_dev, i); + } + + if (mgr_ops && mgr_ops->commit) + mgr_ops->commit(subdrv_dev); +} + static void fimd_commit(struct device *dev) { struct fimd_context *ctx = get_fimd_context(dev); - struct fb_videomode *timing = ctx->timing; + struct exynos_drm_panel_info *panel = ctx->panel; + struct fb_videomode *timing = &panel->timing; u32 val; + if (ctx->suspended) + return; + DRM_DEBUG_KMS("%s\n", __FILE__); /* setup polarity values from machine code. */ @@ -177,40 +239,6 @@ static void fimd_commit(struct device *dev) writel(val, ctx->regs + VIDCON0); } -static void fimd_disable(struct device *dev) -{ - struct fimd_context *ctx = get_fimd_context(dev); - struct exynos_drm_subdrv *subdrv = &ctx->subdrv; - struct drm_device *drm_dev = subdrv->drm_dev; - struct exynos_drm_manager *manager = &subdrv->manager; - u32 val; - - DRM_DEBUG_KMS("%s\n", __FILE__); - - /* fimd dma off */ - val = readl(ctx->regs + VIDCON0); - val &= ~(VIDCON0_ENVID | VIDCON0_ENVID_F); - writel(val, ctx->regs + VIDCON0); - - /* - * if vblank is enabled status with dma off then - * it disables vsync interrupt. - */ - if (drm_dev->vblank_enabled[manager->pipe] && - atomic_read(&drm_dev->vblank_refcount[manager->pipe])) { - drm_vblank_put(drm_dev, manager->pipe); - - /* - * if vblank_disable_allowed is 0 then disable - * vsync interrupt right now else the vsync interrupt - * would be disabled by drm timer once a current process - * gives up ownershop of vblank event. - */ - if (!drm_dev->vblank_disable_allowed) - drm_vblank_off(drm_dev, manager->pipe); - } -} - static int fimd_enable_vblank(struct device *dev) { struct fimd_context *ctx = get_fimd_context(dev); @@ -218,6 +246,9 @@ static int fimd_enable_vblank(struct device *dev) DRM_DEBUG_KMS("%s\n", __FILE__); + if (ctx->suspended) + return -EPERM; + if (!test_and_set_bit(0, &ctx->irq_flags)) { val = readl(ctx->regs + VIDINTCON0); @@ -242,6 +273,9 @@ static void fimd_disable_vblank(struct device *dev) DRM_DEBUG_KMS("%s\n", __FILE__); + if (ctx->suspended) + return; + if (test_and_clear_bit(0, &ctx->irq_flags)) { val = readl(ctx->regs + VIDINTCON0); @@ -253,8 +287,9 @@ static void fimd_disable_vblank(struct device *dev) } static struct exynos_drm_manager_ops fimd_manager_ops = { + .dpms = fimd_dpms, + .apply = fimd_apply, .commit = fimd_commit, - .disable = fimd_disable, .enable_vblank = fimd_enable_vblank, .disable_vblank = fimd_disable_vblank, }; @@ -264,6 +299,7 @@ static void fimd_win_mode_set(struct device *dev, { struct fimd_context *ctx = get_fimd_context(dev); struct fimd_win_data *win_data; + int win; unsigned long offset; DRM_DEBUG_KMS("%s\n", __FILE__); @@ -273,12 +309,19 @@ static void fimd_win_mode_set(struct device *dev, return; } + win = overlay->zpos; + if (win == DEFAULT_ZPOS) + win = ctx->default_win; + + if (win < 0 || win > WINDOWS_NR) + return; + offset = overlay->fb_x * (overlay->bpp >> 3); offset += overlay->fb_y * overlay->pitch; DRM_DEBUG_KMS("offset = 0x%lx, pitch = %x\n", offset, overlay->pitch); - win_data = &ctx->win_data[ctx->default_win]; + win_data = &ctx->win_data[win]; win_data->offset_x = overlay->crtc_x; win_data->offset_y = overlay->crtc_y; @@ -286,8 +329,8 @@ static void fimd_win_mode_set(struct device *dev, win_data->ovl_height = overlay->crtc_height; win_data->fb_width = overlay->fb_width; win_data->fb_height = overlay->fb_height; - win_data->dma_addr = overlay->dma_addr + offset; - win_data->vaddr = overlay->vaddr + offset; + win_data->dma_addr = overlay->dma_addr[0] + offset; + win_data->vaddr = overlay->vaddr[0] + offset; win_data->bpp = overlay->bpp; win_data->buf_offsize = (overlay->fb_width - overlay->crtc_width) * (overlay->bpp >> 3); @@ -381,15 +424,21 @@ static void fimd_win_set_colkey(struct device *dev, unsigned int win) writel(keycon1, ctx->regs + WKEYCON1_BASE(win)); } -static void fimd_win_commit(struct device *dev) +static void fimd_win_commit(struct device *dev, int zpos) { struct fimd_context *ctx = get_fimd_context(dev); struct fimd_win_data *win_data; - int win = ctx->default_win; + int win = zpos; unsigned long val, alpha, size; DRM_DEBUG_KMS("%s\n", __FILE__); + if (ctx->suspended) + return; + + if (win == DEFAULT_ZPOS) + win = ctx->default_win; + if (win < 0 || win > WINDOWS_NR) return; @@ -472,24 +521,37 @@ static void fimd_win_commit(struct device *dev) if (win != 0) fimd_win_set_colkey(dev, win); + /* wincon */ + val = readl(ctx->regs + WINCON(win)); + val |= WINCONx_ENWIN; + writel(val, ctx->regs + WINCON(win)); + /* Enable DMA channel and unprotect windows */ val = readl(ctx->regs + SHADOWCON); val |= SHADOWCON_CHx_ENABLE(win); val &= ~SHADOWCON_WINx_PROTECT(win); writel(val, ctx->regs + SHADOWCON); + + win_data->enabled = true; } -static void fimd_win_disable(struct device *dev) +static void fimd_win_disable(struct device *dev, int zpos) { struct fimd_context *ctx = get_fimd_context(dev); - int win = ctx->default_win; + struct fimd_win_data *win_data; + int win = zpos; u32 val; DRM_DEBUG_KMS("%s\n", __FILE__); + if (win == DEFAULT_ZPOS) + win = ctx->default_win; + if (win < 0 || win > WINDOWS_NR) return; + win_data = &ctx->win_data[win]; + /* protect windows */ val = readl(ctx->regs + SHADOWCON); val |= SHADOWCON_WINx_PROTECT(win); @@ -505,6 +567,8 @@ static void fimd_win_disable(struct device *dev) val &= ~SHADOWCON_CHx_ENABLE(win); val &= ~SHADOWCON_WINx_PROTECT(win); writel(val, ctx->regs + SHADOWCON); + + win_data->enabled = false; } static struct exynos_drm_overlay_ops fimd_overlay_ops = { @@ -513,6 +577,13 @@ static struct exynos_drm_overlay_ops fimd_overlay_ops = { .disable = fimd_win_disable, }; +static struct exynos_drm_manager fimd_manager = { + .pipe = -1, + .ops = &fimd_manager_ops, + .overlay_ops = &fimd_overlay_ops, + .display_ops = &fimd_display_ops, +}; + static void fimd_finish_pageflip(struct drm_device *drm_dev, int crtc) { struct exynos_drm_private *dev_priv = drm_dev->dev_private; @@ -540,8 +611,21 @@ static void fimd_finish_pageflip(struct drm_device *drm_dev, int crtc) wake_up_interruptible(&e->base.file_priv->event_wait); } - if (is_checked) - drm_vblank_put(drm_dev, crtc); + if (is_checked) { + /* + * call drm_vblank_put only in case that drm_vblank_get was + * called. + */ + if (atomic_read(&drm_dev->vblank_refcount[crtc]) > 0) + drm_vblank_put(drm_dev, crtc); + + /* + * don't off vblank if vblank_disable_allowed is 1, + * because vblank would be off by timer handler. + */ + if (!drm_dev->vblank_disable_allowed) + drm_vblank_off(drm_dev, crtc); + } spin_unlock_irqrestore(&drm_dev->event_lock, flags); } @@ -551,7 +635,7 @@ static irqreturn_t fimd_irq_handler(int irq, void *dev_id) struct fimd_context *ctx = (struct fimd_context *)dev_id; struct exynos_drm_subdrv *subdrv = &ctx->subdrv; struct drm_device *drm_dev = subdrv->drm_dev; - struct exynos_drm_manager *manager = &subdrv->manager; + struct exynos_drm_manager *manager = subdrv->manager; u32 val; val = readl(ctx->regs + VIDINTCON1); @@ -560,19 +644,14 @@ static irqreturn_t fimd_irq_handler(int irq, void *dev_id) /* VSYNC interrupt */ writel(VIDINTCON1_INT_FRAME, ctx->regs + VIDINTCON1); - /* - * in case that vblank_disable_allowed is 1, it could induce - * the problem that manager->pipe could be -1 because with - * disable callback, vsync interrupt isn't disabled and at this moment, - * vsync interrupt could occur. the vsync interrupt would be disabled - * by timer handler later. - */ - if (manager->pipe == -1) - return IRQ_HANDLED; + /* check the crtc is detached already from encoder */ + if (manager->pipe < 0) + goto out; drm_handle_vblank(drm_dev, manager->pipe); fimd_finish_pageflip(drm_dev, manager->pipe); +out: return IRQ_HANDLED; } @@ -590,6 +669,13 @@ static int fimd_subdrv_probe(struct drm_device *drm_dev, struct device *dev) */ drm_dev->irq_enabled = 1; + /* + * with vblank_disable_allowed = 1, vblank interrupt will be disabled + * by drm timer once a current process gives up ownership of + * vblank event.(after drm_vblank_put function is called) + */ + drm_dev->vblank_disable_allowed = 1; + return 0; } @@ -662,13 +748,53 @@ static void fimd_clear_win(struct fimd_context *ctx, int win) writel(val, ctx->regs + SHADOWCON); } +static int fimd_power_on(struct fimd_context *ctx, bool enable) +{ + struct exynos_drm_subdrv *subdrv = &ctx->subdrv; + struct device *dev = subdrv->dev; + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (enable != false && enable != true) + return -EINVAL; + + if (enable) { + int ret; + + ret = clk_enable(ctx->bus_clk); + if (ret < 0) + return ret; + + ret = clk_enable(ctx->lcd_clk); + if (ret < 0) { + clk_disable(ctx->bus_clk); + return ret; + } + + ctx->suspended = false; + + /* if vblank was enabled status, enable it again. */ + if (test_and_clear_bit(0, &ctx->irq_flags)) + fimd_enable_vblank(dev); + + fimd_apply(dev); + } else { + clk_disable(ctx->lcd_clk); + clk_disable(ctx->bus_clk); + + ctx->suspended = true; + } + + return 0; +} + static int __devinit fimd_probe(struct platform_device *pdev) { struct device *dev = &pdev->dev; struct fimd_context *ctx; struct exynos_drm_subdrv *subdrv; struct exynos_drm_fimd_pdata *pdata; - struct fb_videomode *timing; + struct exynos_drm_panel_info *panel; struct resource *res; int win; int ret = -EINVAL; @@ -681,9 +807,9 @@ static int __devinit fimd_probe(struct platform_device *pdev) return -EINVAL; } - timing = &pdata->timing; - if (!timing) { - dev_err(dev, "timing is null.\n"); + panel = &pdata->panel; + if (!panel) { + dev_err(dev, "panel is null.\n"); return -EINVAL; } @@ -698,8 +824,6 @@ static int __devinit fimd_probe(struct platform_device *pdev) goto err_clk_get; } - clk_enable(ctx->bus_clk); - ctx->lcd_clk = clk_get(dev, "sclk_fimd"); if (IS_ERR(ctx->lcd_clk)) { dev_err(dev, "failed to get lcd clock\n"); @@ -707,8 +831,6 @@ static int __devinit fimd_probe(struct platform_device *pdev) goto err_bus_clk; } - clk_enable(ctx->lcd_clk); - res = platform_get_resource(pdev, IORESOURCE_MEM, 0); if (!res) { dev_err(dev, "failed to find registers\n"); @@ -739,37 +861,40 @@ static int __devinit fimd_probe(struct platform_device *pdev) ctx->irq = res->start; - for (win = 0; win < WINDOWS_NR; win++) - fimd_clear_win(ctx, win); - ret = request_irq(ctx->irq, fimd_irq_handler, 0, "drm_fimd", ctx); if (ret < 0) { dev_err(dev, "irq request failed.\n"); goto err_req_irq; } - ctx->clkdiv = fimd_calc_clkdiv(ctx, timing); ctx->vidcon0 = pdata->vidcon0; ctx->vidcon1 = pdata->vidcon1; ctx->default_win = pdata->default_win; - ctx->timing = timing; - - timing->pixclock = clk_get_rate(ctx->lcd_clk) / ctx->clkdiv; - - DRM_DEBUG_KMS("pixel clock = %d, clkdiv = %d\n", - timing->pixclock, ctx->clkdiv); + ctx->panel = panel; subdrv = &ctx->subdrv; + subdrv->dev = dev; + subdrv->manager = &fimd_manager; subdrv->probe = fimd_subdrv_probe; subdrv->remove = fimd_subdrv_remove; - subdrv->manager.pipe = -1; - subdrv->manager.ops = &fimd_manager_ops; - subdrv->manager.overlay_ops = &fimd_overlay_ops; - subdrv->manager.display_ops = &fimd_display_ops; - subdrv->manager.dev = dev; + + mutex_init(&ctx->lock); platform_set_drvdata(pdev, ctx); + + pm_runtime_enable(dev); + pm_runtime_get_sync(dev); + + ctx->clkdiv = fimd_calc_clkdiv(ctx, &panel->timing); + panel->timing.pixclock = clk_get_rate(ctx->lcd_clk) / ctx->clkdiv; + + DRM_DEBUG_KMS("pixel clock = %d, clkdiv = %d\n", + panel->timing.pixclock, ctx->clkdiv); + + for (win = 0; win < WINDOWS_NR; win++) + fimd_clear_win(ctx, win); + exynos_drm_subdrv_register(subdrv); return 0; @@ -797,14 +922,25 @@ err_clk_get: static int __devexit fimd_remove(struct platform_device *pdev) { + struct device *dev = &pdev->dev; struct fimd_context *ctx = platform_get_drvdata(pdev); DRM_DEBUG_KMS("%s\n", __FILE__); exynos_drm_subdrv_unregister(&ctx->subdrv); + if (ctx->suspended) + goto out; + clk_disable(ctx->lcd_clk); clk_disable(ctx->bus_clk); + + pm_runtime_set_suspended(dev); + pm_runtime_put_sync(dev); + +out: + pm_runtime_disable(dev); + clk_put(ctx->lcd_clk); clk_put(ctx->bus_clk); @@ -818,29 +954,69 @@ static int __devexit fimd_remove(struct platform_device *pdev) return 0; } -static struct platform_driver fimd_driver = { - .probe = fimd_probe, - .remove = __devexit_p(fimd_remove), - .driver = { - .name = "exynos4-fb", - .owner = THIS_MODULE, - }, -}; +#ifdef CONFIG_PM_SLEEP +static int fimd_suspend(struct device *dev) +{ + struct fimd_context *ctx = get_fimd_context(dev); + + if (pm_runtime_suspended(dev)) + return 0; + + /* + * do not use pm_runtime_suspend(). if pm_runtime_suspend() is + * called here, an error would be returned by that interface + * because the usage_count of pm runtime is more than 1. + */ + return fimd_power_on(ctx, false); +} + +static int fimd_resume(struct device *dev) +{ + struct fimd_context *ctx = get_fimd_context(dev); + + /* + * if entered to sleep when lcd panel was on, the usage_count + * of pm runtime would still be 1 so in this case, fimd driver + * should be on directly not drawing on pm runtime interface. + */ + if (!pm_runtime_suspended(dev)) + return fimd_power_on(ctx, true); -static int __init fimd_init(void) + return 0; +} +#endif + +#ifdef CONFIG_PM_RUNTIME +static int fimd_runtime_suspend(struct device *dev) { - return platform_driver_register(&fimd_driver); + struct fimd_context *ctx = get_fimd_context(dev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + return fimd_power_on(ctx, false); } -static void __exit fimd_exit(void) +static int fimd_runtime_resume(struct device *dev) { - platform_driver_unregister(&fimd_driver); + struct fimd_context *ctx = get_fimd_context(dev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + return fimd_power_on(ctx, true); } +#endif -module_init(fimd_init); -module_exit(fimd_exit); +static const struct dev_pm_ops fimd_pm_ops = { + SET_SYSTEM_SLEEP_PM_OPS(fimd_suspend, fimd_resume) + SET_RUNTIME_PM_OPS(fimd_runtime_suspend, fimd_runtime_resume, NULL) +}; -MODULE_AUTHOR("Joonyoung Shim "); -MODULE_AUTHOR("Inki Dae "); -MODULE_DESCRIPTION("Samsung DRM FIMD Driver"); -MODULE_LICENSE("GPL"); +struct platform_driver fimd_driver = { + .probe = fimd_probe, + .remove = __devexit_p(fimd_remove), + .driver = { + .name = "exynos4-fb", + .owner = THIS_MODULE, + .pm = &fimd_pm_ops, + }, +}; diff --git a/drivers/gpu/drm/exynos/exynos_drm_gem.c b/drivers/gpu/drm/exynos/exynos_drm_gem.c index aba0fe4..1dffa83 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_gem.c +++ b/drivers/gpu/drm/exynos/exynos_drm_gem.c @@ -26,6 +26,7 @@ #include "drmP.h" #include "drm.h" +#include #include #include "exynos_drm_drv.h" @@ -55,118 +56,413 @@ static unsigned int convert_to_vm_err_msg(int msg) return out_msg; } -static unsigned int get_gem_mmap_offset(struct drm_gem_object *obj) +static int check_gem_flags(unsigned int flags) { - DRM_DEBUG_KMS("%s\n", __FILE__); + if (flags & ~(EXYNOS_BO_MASK)) { + DRM_ERROR("invalid flags.\n"); + return -EINVAL; + } - return (unsigned int)obj->map_list.hash.key << PAGE_SHIFT; + return 0; } -static struct exynos_drm_gem_obj - *exynos_drm_gem_init(struct drm_device *drm_dev, - struct drm_file *file_priv, unsigned int *handle, - unsigned int size) +static unsigned long roundup_gem_size(unsigned long size, unsigned int flags) { - struct exynos_drm_gem_obj *exynos_gem_obj; - struct drm_gem_object *obj; - int ret; + if (!IS_NONCONTIG_BUFFER(flags)) { + if (size >= SZ_1M) + return roundup(size, SECTION_SIZE); + else if (size >= SZ_64K) + return roundup(size, SZ_64K); + else + goto out; + } +out: + return roundup(size, PAGE_SIZE); +} - exynos_gem_obj = kzalloc(sizeof(*exynos_gem_obj), GFP_KERNEL); - if (!exynos_gem_obj) { - DRM_ERROR("failed to allocate exynos gem object.\n"); +static struct page **exynos_gem_get_pages(struct drm_gem_object *obj, + gfp_t gfpmask) +{ + struct inode *inode; + struct address_space *mapping; + struct page *p, **pages; + int i, npages; + + /* This is the shared memory object that backs the GEM resource */ + inode = obj->filp->f_path.dentry->d_inode; + mapping = inode->i_mapping; + + npages = obj->size >> PAGE_SHIFT; + + pages = drm_malloc_ab(npages, sizeof(struct page *)); + if (pages == NULL) return ERR_PTR(-ENOMEM); + + gfpmask |= mapping_gfp_mask(mapping); + + for (i = 0; i < npages; i++) { + p = shmem_read_mapping_page_gfp(mapping, i, gfpmask); + if (IS_ERR(p)) + goto fail; + pages[i] = p; } - obj = &exynos_gem_obj->base; + return pages; - ret = drm_gem_object_init(drm_dev, obj, size); - if (ret < 0) { - DRM_ERROR("failed to initialize gem object.\n"); - ret = -EINVAL; - goto err_object_init; +fail: + while (i--) + page_cache_release(pages[i]); + + drm_free_large(pages); + return ERR_PTR(PTR_ERR(p)); +} + +static void exynos_gem_put_pages(struct drm_gem_object *obj, + struct page **pages, + bool dirty, bool accessed) +{ + int i, npages; + + npages = obj->size >> PAGE_SHIFT; + + for (i = 0; i < npages; i++) { + if (dirty) + set_page_dirty(pages[i]); + + if (accessed) + mark_page_accessed(pages[i]); + + /* Undo the reference we took when populating the table */ + page_cache_release(pages[i]); } - DRM_DEBUG_KMS("created file object = 0x%x\n", (unsigned int)obj->filp); + drm_free_large(pages); +} - ret = drm_gem_create_mmap_offset(obj); +static int exynos_drm_gem_map_pages(struct drm_gem_object *obj, + struct vm_area_struct *vma, + unsigned long f_vaddr, + pgoff_t page_offset) +{ + struct exynos_drm_gem_obj *exynos_gem_obj = to_exynos_gem_obj(obj); + struct exynos_drm_gem_buf *buf = exynos_gem_obj->buffer; + unsigned long pfn; + + if (exynos_gem_obj->flags & EXYNOS_BO_NONCONTIG) { + if (!buf->pages) + return -EINTR; + + pfn = page_to_pfn(buf->pages[page_offset++]); + } else + pfn = (buf->dma_addr >> PAGE_SHIFT) + page_offset; + + return vm_insert_mixed(vma, f_vaddr, pfn); +} + +static int exynos_drm_gem_get_pages(struct drm_gem_object *obj) +{ + struct exynos_drm_gem_obj *exynos_gem_obj = to_exynos_gem_obj(obj); + struct exynos_drm_gem_buf *buf = exynos_gem_obj->buffer; + struct scatterlist *sgl; + struct page **pages; + unsigned int npages, i = 0; + int ret; + + if (buf->pages) { + DRM_DEBUG_KMS("already allocated.\n"); + return -EINVAL; + } + + pages = exynos_gem_get_pages(obj, GFP_KERNEL); + if (IS_ERR(pages)) { + DRM_ERROR("failed to get pages.\n"); + return PTR_ERR(pages); + } + + npages = obj->size >> PAGE_SHIFT; + + buf->sgt = kzalloc(sizeof(struct sg_table), GFP_KERNEL); + if (!buf->sgt) { + DRM_ERROR("failed to allocate sg table.\n"); + ret = -ENOMEM; + goto err; + } + + ret = sg_alloc_table(buf->sgt, npages, GFP_KERNEL); if (ret < 0) { - DRM_ERROR("failed to allocate mmap offset.\n"); - goto err_create_mmap_offset; + DRM_ERROR("failed to initialize sg table.\n"); + ret = -EFAULT; + goto err1; + } + + sgl = buf->sgt->sgl; + + /* set all pages to sg list. */ + while (i < npages) { + sg_set_page(sgl, pages[i], PAGE_SIZE, 0); + sg_dma_address(sgl) = page_to_phys(pages[i]); + i++; + sgl = sg_next(sgl); } + /* add some codes for UNCACHED type here. TODO */ + + buf->pages = pages; + return ret; +err1: + kfree(buf->sgt); + buf->sgt = NULL; +err: + exynos_gem_put_pages(obj, pages, true, false); + return ret; + +} + +static void exynos_drm_gem_put_pages(struct drm_gem_object *obj) +{ + struct exynos_drm_gem_obj *exynos_gem_obj = to_exynos_gem_obj(obj); + struct exynos_drm_gem_buf *buf = exynos_gem_obj->buffer; + + /* + * if buffer typs is EXYNOS_BO_NONCONTIG then release all pages + * allocated at gem fault handler. + */ + sg_free_table(buf->sgt); + kfree(buf->sgt); + buf->sgt = NULL; + + exynos_gem_put_pages(obj, buf->pages, true, false); + buf->pages = NULL; + + /* add some codes for UNCACHED type here. TODO */ +} + +static int exynos_drm_gem_handle_create(struct drm_gem_object *obj, + struct drm_file *file_priv, + unsigned int *handle) +{ + int ret; + /* * allocate a id of idr table where the obj is registered * and handle has the id what user can see. */ ret = drm_gem_handle_create(file_priv, obj, handle); if (ret) - goto err_handle_create; + return ret; DRM_DEBUG_KMS("gem handle = 0x%x\n", *handle); /* drop reference from allocate - handle holds it now. */ drm_gem_object_unreference_unlocked(obj); - return exynos_gem_obj; + return 0; +} + +void exynos_drm_gem_destroy(struct exynos_drm_gem_obj *exynos_gem_obj) +{ + struct drm_gem_object *obj; + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (!exynos_gem_obj) + return; -err_handle_create: - drm_gem_free_mmap_offset(obj); + obj = &exynos_gem_obj->base; + + DRM_DEBUG_KMS("handle count = %d\n", atomic_read(&obj->handle_count)); + + if ((exynos_gem_obj->flags & EXYNOS_BO_NONCONTIG) && + exynos_gem_obj->buffer->pages) + exynos_drm_gem_put_pages(obj); + else + exynos_drm_free_buf(obj->dev, exynos_gem_obj->flags, + exynos_gem_obj->buffer); + + exynos_drm_fini_buf(obj->dev, exynos_gem_obj->buffer); + exynos_gem_obj->buffer = NULL; -err_create_mmap_offset: + if (obj->map_list.map) + drm_gem_free_mmap_offset(obj); + + /* release file pointer to gem object. */ drm_gem_object_release(obj); -err_object_init: kfree(exynos_gem_obj); + exynos_gem_obj = NULL; +} - return ERR_PTR(ret); +static struct exynos_drm_gem_obj *exynos_drm_gem_init(struct drm_device *dev, + unsigned long size) +{ + struct exynos_drm_gem_obj *exynos_gem_obj; + struct drm_gem_object *obj; + int ret; + + exynos_gem_obj = kzalloc(sizeof(*exynos_gem_obj), GFP_KERNEL); + if (!exynos_gem_obj) { + DRM_ERROR("failed to allocate exynos gem object\n"); + return NULL; + } + + exynos_gem_obj->size = size; + obj = &exynos_gem_obj->base; + + ret = drm_gem_object_init(dev, obj, size); + if (ret < 0) { + DRM_ERROR("failed to initialize gem object\n"); + kfree(exynos_gem_obj); + return NULL; + } + + DRM_DEBUG_KMS("created file object = 0x%x\n", (unsigned int)obj->filp); + + return exynos_gem_obj; } struct exynos_drm_gem_obj *exynos_drm_gem_create(struct drm_device *dev, - struct drm_file *file_priv, - unsigned int *handle, unsigned long size) + unsigned int flags, + unsigned long size) { + struct exynos_drm_gem_obj *exynos_gem_obj; + struct exynos_drm_gem_buf *buf; + int ret; - struct exynos_drm_gem_obj *exynos_gem_obj = NULL; - struct exynos_drm_gem_buf *buffer; + if (!size) { + DRM_ERROR("invalid size.\n"); + return ERR_PTR(-EINVAL); + } - size = roundup(size, PAGE_SIZE); + size = roundup_gem_size(size, flags); + DRM_DEBUG_KMS("%s\n", __FILE__); - DRM_DEBUG_KMS("%s: size = 0x%lx\n", __FILE__, size); + ret = check_gem_flags(flags); + if (ret) + return ERR_PTR(ret); - buffer = exynos_drm_buf_create(dev, size); - if (IS_ERR(buffer)) { - return ERR_CAST(buffer); - } + buf = exynos_drm_init_buf(dev, size); + if (!buf) + return ERR_PTR(-ENOMEM); - exynos_gem_obj = exynos_drm_gem_init(dev, file_priv, handle, size); - if (IS_ERR(exynos_gem_obj)) { - exynos_drm_buf_destroy(dev, buffer); - return exynos_gem_obj; + exynos_gem_obj = exynos_drm_gem_init(dev, size); + if (!exynos_gem_obj) { + ret = -ENOMEM; + goto err_fini_buf; } - exynos_gem_obj->buffer = buffer; + exynos_gem_obj->buffer = buf; + + /* set memory type and cache attribute from user side. */ + exynos_gem_obj->flags = flags; + + /* + * allocate all pages as desired size if user wants to allocate + * physically non-continuous memory. + */ + if (flags & EXYNOS_BO_NONCONTIG) { + ret = exynos_drm_gem_get_pages(&exynos_gem_obj->base); + if (ret < 0) { + drm_gem_object_release(&exynos_gem_obj->base); + goto err_fini_buf; + } + } else { + ret = exynos_drm_alloc_buf(dev, buf, flags); + if (ret < 0) { + drm_gem_object_release(&exynos_gem_obj->base); + goto err_fini_buf; + } + } return exynos_gem_obj; + +err_fini_buf: + exynos_drm_fini_buf(dev, buf); + return ERR_PTR(ret); } int exynos_drm_gem_create_ioctl(struct drm_device *dev, void *data, - struct drm_file *file_priv) + struct drm_file *file_priv) { struct drm_exynos_gem_create *args = data; - struct exynos_drm_gem_obj *exynos_gem_obj = NULL; + struct exynos_drm_gem_obj *exynos_gem_obj; + int ret; DRM_DEBUG_KMS("%s\n", __FILE__); - exynos_gem_obj = exynos_drm_gem_create(dev, file_priv, - &args->handle, args->size); + exynos_gem_obj = exynos_drm_gem_create(dev, args->flags, args->size); if (IS_ERR(exynos_gem_obj)) return PTR_ERR(exynos_gem_obj); + ret = exynos_drm_gem_handle_create(&exynos_gem_obj->base, file_priv, + &args->handle); + if (ret) { + exynos_drm_gem_destroy(exynos_gem_obj); + return ret; + } + return 0; } +void *exynos_drm_gem_get_dma_addr(struct drm_device *dev, + unsigned int gem_handle, + struct drm_file *file_priv) +{ + struct exynos_drm_gem_obj *exynos_gem_obj; + struct drm_gem_object *obj; + + obj = drm_gem_object_lookup(dev, file_priv, gem_handle); + if (!obj) { + DRM_ERROR("failed to lookup gem object.\n"); + return ERR_PTR(-EINVAL); + } + + exynos_gem_obj = to_exynos_gem_obj(obj); + + if (exynos_gem_obj->flags & EXYNOS_BO_NONCONTIG) { + DRM_DEBUG_KMS("not support NONCONTIG type.\n"); + drm_gem_object_unreference_unlocked(obj); + + /* TODO */ + return ERR_PTR(-EINVAL); + } + + return &exynos_gem_obj->buffer->dma_addr; +} + +void exynos_drm_gem_put_dma_addr(struct drm_device *dev, + unsigned int gem_handle, + struct drm_file *file_priv) +{ + struct exynos_drm_gem_obj *exynos_gem_obj; + struct drm_gem_object *obj; + + obj = drm_gem_object_lookup(dev, file_priv, gem_handle); + if (!obj) { + DRM_ERROR("failed to lookup gem object.\n"); + return; + } + + exynos_gem_obj = to_exynos_gem_obj(obj); + + if (exynos_gem_obj->flags & EXYNOS_BO_NONCONTIG) { + DRM_DEBUG_KMS("not support NONCONTIG type.\n"); + drm_gem_object_unreference_unlocked(obj); + + /* TODO */ + return; + } + + drm_gem_object_unreference_unlocked(obj); + + /* + * decrease obj->refcount one more time because we has already + * increased it at exynos_drm_gem_get_dma_addr(). + */ + drm_gem_object_unreference_unlocked(obj); +} + int exynos_drm_gem_map_offset_ioctl(struct drm_device *dev, void *data, - struct drm_file *file_priv) + struct drm_file *file_priv) { struct drm_exynos_gem_map_off *args = data; @@ -185,21 +481,23 @@ int exynos_drm_gem_map_offset_ioctl(struct drm_device *dev, void *data, } static int exynos_drm_gem_mmap_buffer(struct file *filp, - struct vm_area_struct *vma) + struct vm_area_struct *vma) { struct drm_gem_object *obj = filp->private_data; struct exynos_drm_gem_obj *exynos_gem_obj = to_exynos_gem_obj(obj); struct exynos_drm_gem_buf *buffer; - unsigned long pfn, vm_size; + unsigned long pfn, vm_size, usize, uaddr = vma->vm_start; + int ret; DRM_DEBUG_KMS("%s\n", __FILE__); vma->vm_flags |= (VM_IO | VM_RESERVED); + /* in case of direct mapping, always having non-cachable attribute */ vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); - vma->vm_file = filp; - vm_size = vma->vm_end - vma->vm_start; + vm_size = usize = vma->vm_end - vma->vm_start; + /* * a buffer contains information to physically continuous memory * allocated by user request or at framebuffer creation. @@ -210,18 +508,39 @@ static int exynos_drm_gem_mmap_buffer(struct file *filp, if (vm_size > buffer->size) return -EINVAL; - /* - * get page frame number to physical memory to be mapped - * to user space. - */ - pfn = ((unsigned long)exynos_gem_obj->buffer->dma_addr) >> PAGE_SHIFT; - - DRM_DEBUG_KMS("pfn = 0x%lx\n", pfn); - - if (remap_pfn_range(vma, vma->vm_start, pfn, vm_size, - vma->vm_page_prot)) { - DRM_ERROR("failed to remap pfn range.\n"); - return -EAGAIN; + if (exynos_gem_obj->flags & EXYNOS_BO_NONCONTIG) { + int i = 0; + + if (!buffer->pages) + return -EINVAL; + + vma->vm_flags |= VM_MIXEDMAP; + + do { + ret = vm_insert_page(vma, uaddr, buffer->pages[i++]); + if (ret) { + DRM_ERROR("failed to remap user space.\n"); + return ret; + } + + uaddr += PAGE_SIZE; + usize -= PAGE_SIZE; + } while (usize > 0); + } else { + /* + * get page frame number to physical memory to be mapped + * to user space. + */ + pfn = ((unsigned long)exynos_gem_obj->buffer->dma_addr) >> + PAGE_SHIFT; + + DRM_DEBUG_KMS("pfn = 0x%lx\n", pfn); + + if (remap_pfn_range(vma, vma->vm_start, pfn, vm_size, + vma->vm_page_prot)) { + DRM_ERROR("failed to remap pfn range.\n"); + return -EAGAIN; + } } return 0; @@ -232,7 +551,7 @@ static const struct file_operations exynos_drm_gem_fops = { }; int exynos_drm_gem_mmap_ioctl(struct drm_device *dev, void *data, - struct drm_file *file_priv) + struct drm_file *file_priv) { struct drm_exynos_gem_mmap *args = data; struct drm_gem_object *obj; @@ -254,10 +573,8 @@ int exynos_drm_gem_mmap_ioctl(struct drm_device *dev, void *data, obj->filp->f_op = &exynos_drm_gem_fops; obj->filp->private_data = obj; - down_write(¤t->mm->mmap_sem); - addr = do_mmap(obj->filp, 0, args->size, + addr = vm_mmap(obj->filp, 0, args->size, PROT_READ | PROT_WRITE, MAP_SHARED, 0); - up_write(¤t->mm->mmap_sem); drm_gem_object_unreference_unlocked(obj); @@ -278,32 +595,19 @@ int exynos_drm_gem_init_object(struct drm_gem_object *obj) return 0; } -void exynos_drm_gem_free_object(struct drm_gem_object *gem_obj) +void exynos_drm_gem_free_object(struct drm_gem_object *obj) { - struct exynos_drm_gem_obj *exynos_gem_obj; - DRM_DEBUG_KMS("%s\n", __FILE__); - DRM_DEBUG_KMS("handle count = %d\n", - atomic_read(&gem_obj->handle_count)); - - if (gem_obj->map_list.map) - drm_gem_free_mmap_offset(gem_obj); - - /* release file pointer to gem object. */ - drm_gem_object_release(gem_obj); - - exynos_gem_obj = to_exynos_gem_obj(gem_obj); - - exynos_drm_buf_destroy(gem_obj->dev, exynos_gem_obj->buffer); - - kfree(exynos_gem_obj); + exynos_drm_gem_destroy(to_exynos_gem_obj(obj)); } int exynos_drm_gem_dumb_create(struct drm_file *file_priv, - struct drm_device *dev, struct drm_mode_create_dumb *args) + struct drm_device *dev, + struct drm_mode_create_dumb *args) { struct exynos_drm_gem_obj *exynos_gem_obj; + int ret; DRM_DEBUG_KMS("%s\n", __FILE__); @@ -314,21 +618,29 @@ int exynos_drm_gem_dumb_create(struct drm_file *file_priv, */ args->pitch = args->width * args->bpp >> 3; - args->size = args->pitch * args->height; + args->size = PAGE_ALIGN(args->pitch * args->height); - exynos_gem_obj = exynos_drm_gem_create(dev, file_priv, &args->handle, - args->size); + exynos_gem_obj = exynos_drm_gem_create(dev, args->flags, args->size); if (IS_ERR(exynos_gem_obj)) return PTR_ERR(exynos_gem_obj); + ret = exynos_drm_gem_handle_create(&exynos_gem_obj->base, file_priv, + &args->handle); + if (ret) { + exynos_drm_gem_destroy(exynos_gem_obj); + return ret; + } + return 0; } int exynos_drm_gem_dumb_map_offset(struct drm_file *file_priv, - struct drm_device *dev, uint32_t handle, uint64_t *offset) + struct drm_device *dev, uint32_t handle, + uint64_t *offset) { struct exynos_drm_gem_obj *exynos_gem_obj; struct drm_gem_object *obj; + int ret = 0; DRM_DEBUG_KMS("%s\n", __FILE__); @@ -343,19 +655,46 @@ int exynos_drm_gem_dumb_map_offset(struct drm_file *file_priv, obj = drm_gem_object_lookup(dev, file_priv, handle); if (!obj) { DRM_ERROR("failed to lookup gem object.\n"); - mutex_unlock(&dev->struct_mutex); - return -EINVAL; + ret = -EINVAL; + goto unlock; } exynos_gem_obj = to_exynos_gem_obj(obj); - *offset = get_gem_mmap_offset(&exynos_gem_obj->base); - - drm_gem_object_unreference(obj); + if (!exynos_gem_obj->base.map_list.map) { + ret = drm_gem_create_mmap_offset(&exynos_gem_obj->base); + if (ret) + goto out; + } + *offset = (u64)exynos_gem_obj->base.map_list.hash.key << PAGE_SHIFT; DRM_DEBUG_KMS("offset = 0x%lx\n", (unsigned long)*offset); +out: + drm_gem_object_unreference(obj); +unlock: mutex_unlock(&dev->struct_mutex); + return ret; +} + +int exynos_drm_gem_dumb_destroy(struct drm_file *file_priv, + struct drm_device *dev, + unsigned int handle) +{ + int ret; + + DRM_DEBUG_KMS("%s\n", __FILE__); + + /* + * obj->refcount and obj->handle_count are decreased and + * if both them are 0 then exynos_drm_gem_free_object() + * would be called by callback to release resources. + */ + ret = drm_gem_handle_delete(file_priv, handle); + if (ret < 0) { + DRM_ERROR("failed to delete drm_gem_handle.\n"); + return ret; + } return 0; } @@ -363,21 +702,20 @@ int exynos_drm_gem_dumb_map_offset(struct drm_file *file_priv, int exynos_drm_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf) { struct drm_gem_object *obj = vma->vm_private_data; - struct exynos_drm_gem_obj *exynos_gem_obj = to_exynos_gem_obj(obj); struct drm_device *dev = obj->dev; - unsigned long pfn; + unsigned long f_vaddr; pgoff_t page_offset; int ret; page_offset = ((unsigned long)vmf->virtual_address - vma->vm_start) >> PAGE_SHIFT; + f_vaddr = (unsigned long)vmf->virtual_address; mutex_lock(&dev->struct_mutex); - pfn = (((unsigned long)exynos_gem_obj->buffer->dma_addr) >> - PAGE_SHIFT) + page_offset; - - ret = vm_insert_mixed(vma, (unsigned long)vmf->virtual_address, pfn); + ret = exynos_drm_gem_map_pages(obj, vma, f_vaddr, page_offset); + if (ret < 0) + DRM_ERROR("failed to map pages.\n"); mutex_unlock(&dev->struct_mutex); @@ -402,29 +740,3 @@ int exynos_drm_gem_mmap(struct file *filp, struct vm_area_struct *vma) return ret; } - - -int exynos_drm_gem_dumb_destroy(struct drm_file *file_priv, - struct drm_device *dev, unsigned int handle) -{ - int ret; - - DRM_DEBUG_KMS("%s\n", __FILE__); - - /* - * obj->refcount and obj->handle_count are decreased and - * if both them are 0 then exynos_drm_gem_free_object() - * would be called by callback to release resources. - */ - ret = drm_gem_handle_delete(file_priv, handle); - if (ret < 0) { - DRM_ERROR("failed to delete drm_gem_handle.\n"); - return ret; - } - - return 0; -} - -MODULE_AUTHOR("Inki Dae "); -MODULE_DESCRIPTION("Samsung SoC DRM GEM Module"); -MODULE_LICENSE("GPL"); diff --git a/drivers/gpu/drm/exynos/exynos_drm_gem.h b/drivers/gpu/drm/exynos/exynos_drm_gem.h index ef87973..4ed8420 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_gem.h +++ b/drivers/gpu/drm/exynos/exynos_drm_gem.h @@ -29,6 +29,8 @@ #define to_exynos_gem_obj(x) container_of(x,\ struct exynos_drm_gem_obj, base) +#define IS_NONCONTIG_BUFFER(f) (f & EXYNOS_BO_NONCONTIG) + /* * exynos drm gem buffer structure. * @@ -36,11 +38,15 @@ * @dma_addr: bus address(accessed by dma) to allocated memory region. * - this address could be physical address without IOMMU and * device address with IOMMU. + * @sgt: sg table to transfer page data. + * @pages: contain all pages to allocated memory region. * @size: size of allocated memory region. */ struct exynos_drm_gem_buf { void __iomem *kvaddr; dma_addr_t dma_addr; + struct sg_table *sgt; + struct page **pages; unsigned long size; }; @@ -55,19 +61,26 @@ struct exynos_drm_gem_buf { * by user request or at framebuffer creation. * continuous memory region allocated by user request * or at framebuffer creation. + * @size: total memory size to physically non-continuous memory region. + * @flags: indicate memory type to allocated buffer and cache attruibute. * * P.S. this object would be transfered to user as kms_bo.handle so * user can access the buffer through kms_bo.handle. */ struct exynos_drm_gem_obj { - struct drm_gem_object base; - struct exynos_drm_gem_buf *buffer; + struct drm_gem_object base; + struct exynos_drm_gem_buf *buffer; + unsigned long size; + unsigned int flags; }; -/* create a new buffer and get a new gem handle. */ +/* destroy a buffer with gem object */ +void exynos_drm_gem_destroy(struct exynos_drm_gem_obj *exynos_gem_obj); + +/* create a new buffer with gem object */ struct exynos_drm_gem_obj *exynos_drm_gem_create(struct drm_device *dev, - struct drm_file *file_priv, - unsigned int *handle, unsigned long size); + unsigned int flags, + unsigned long size); /* * request gem object creation and buffer allocation as the size @@ -75,15 +88,36 @@ struct exynos_drm_gem_obj *exynos_drm_gem_create(struct drm_device *dev, * height and bpp. */ int exynos_drm_gem_create_ioctl(struct drm_device *dev, void *data, - struct drm_file *file_priv); + struct drm_file *file_priv); + +/* + * get dma address from gem handle and this function could be used for + * other drivers such as 2d/3d acceleration drivers. + * with this function call, gem object reference count would be increased. + */ +void *exynos_drm_gem_get_dma_addr(struct drm_device *dev, + unsigned int gem_handle, + struct drm_file *file_priv); + +/* + * put dma address from gem handle and this function could be used for + * other drivers such as 2d/3d acceleration drivers. + * with this function call, gem object reference count would be decreased. + */ +void exynos_drm_gem_put_dma_addr(struct drm_device *dev, + unsigned int gem_handle, + struct drm_file *file_priv); /* get buffer offset to map to user space. */ int exynos_drm_gem_map_offset_ioctl(struct drm_device *dev, void *data, - struct drm_file *file_priv); + struct drm_file *file_priv); -/* unmap a buffer from user space. */ -int exynos_drm_gem_munmap_ioctl(struct drm_device *dev, void *data, - struct drm_file *file_priv); +/* + * mmap the physically continuous memory that a gem object contains + * to user space. + */ +int exynos_drm_gem_mmap_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv); /* initialize gem object. */ int exynos_drm_gem_init_object(struct drm_gem_object *obj); @@ -93,24 +127,13 @@ void exynos_drm_gem_free_object(struct drm_gem_object *gem_obj); /* create memory region for drm framebuffer. */ int exynos_drm_gem_dumb_create(struct drm_file *file_priv, - struct drm_device *dev, struct drm_mode_create_dumb *args); + struct drm_device *dev, + struct drm_mode_create_dumb *args); /* map memory region for drm framebuffer to user space. */ int exynos_drm_gem_dumb_map_offset(struct drm_file *file_priv, - struct drm_device *dev, uint32_t handle, uint64_t *offset); - -/* page fault handler and mmap fault address(virtual) to physical memory. */ -int exynos_drm_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf); - -/* - * mmap the physically continuous memory that a gem object contains - * to user space. - */ -int exynos_drm_gem_mmap_ioctl(struct drm_device *dev, void *data, - struct drm_file *file_priv); - -/* set vm_flags and we can change the vm attribute to other one at here. */ -int exynos_drm_gem_mmap(struct file *filp, struct vm_area_struct *vma); + struct drm_device *dev, uint32_t handle, + uint64_t *offset); /* * destroy memory region allocated. @@ -118,6 +141,13 @@ int exynos_drm_gem_mmap(struct file *filp, struct vm_area_struct *vma); * would be released by drm_gem_handle_delete(). */ int exynos_drm_gem_dumb_destroy(struct drm_file *file_priv, - struct drm_device *dev, unsigned int handle); + struct drm_device *dev, + unsigned int handle); + +/* page fault handler and mmap fault address(virtual) to physical memory. */ +int exynos_drm_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf); + +/* set vm_flags and we can change the vm attribute to other one at here. */ +int exynos_drm_gem_mmap(struct file *filp, struct vm_area_struct *vma); #endif diff --git a/drivers/gpu/drm/exynos/exynos_drm_hdmi.c b/drivers/gpu/drm/exynos/exynos_drm_hdmi.c new file mode 100644 index 0000000..3424463 --- /dev/null +++ b/drivers/gpu/drm/exynos/exynos_drm_hdmi.c @@ -0,0 +1,377 @@ +/* + * Copyright (C) 2011 Samsung Electronics Co.Ltd + * Authors: + * Inki Dae + * Seung-Woo Kim + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or (at your + * option) any later version. + * + */ + +#include "drmP.h" + +#include +#include +#include +#include +#include + +#include + +#include "exynos_drm_drv.h" +#include "exynos_drm_hdmi.h" + +#define to_context(dev) platform_get_drvdata(to_platform_device(dev)) +#define to_subdrv(dev) to_context(dev) +#define get_ctx_from_subdrv(subdrv) container_of(subdrv,\ + struct drm_hdmi_context, subdrv); + +/* these callback points shoud be set by specific drivers. */ +static struct exynos_hdmi_ops *hdmi_ops; +static struct exynos_mixer_ops *mixer_ops; + +struct drm_hdmi_context { + struct exynos_drm_subdrv subdrv; + struct exynos_drm_hdmi_context *hdmi_ctx; + struct exynos_drm_hdmi_context *mixer_ctx; +}; + +void exynos_hdmi_ops_register(struct exynos_hdmi_ops *ops) +{ + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (ops) + hdmi_ops = ops; +} + +void exynos_mixer_ops_register(struct exynos_mixer_ops *ops) +{ + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (ops) + mixer_ops = ops; +} + +static bool drm_hdmi_is_connected(struct device *dev) +{ + struct drm_hdmi_context *ctx = to_context(dev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (hdmi_ops && hdmi_ops->is_connected) + return hdmi_ops->is_connected(ctx->hdmi_ctx->ctx); + + return false; +} + +static int drm_hdmi_get_edid(struct device *dev, + struct drm_connector *connector, u8 *edid, int len) +{ + struct drm_hdmi_context *ctx = to_context(dev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (hdmi_ops && hdmi_ops->get_edid) + return hdmi_ops->get_edid(ctx->hdmi_ctx->ctx, connector, edid, + len); + + return 0; +} + +static int drm_hdmi_check_timing(struct device *dev, void *timing) +{ + struct drm_hdmi_context *ctx = to_context(dev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (hdmi_ops && hdmi_ops->check_timing) + return hdmi_ops->check_timing(ctx->hdmi_ctx->ctx, timing); + + return 0; +} + +static int drm_hdmi_power_on(struct device *dev, int mode) +{ + struct drm_hdmi_context *ctx = to_context(dev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (hdmi_ops && hdmi_ops->power_on) + return hdmi_ops->power_on(ctx->hdmi_ctx->ctx, mode); + + return 0; +} + +static struct exynos_drm_display_ops drm_hdmi_display_ops = { + .type = EXYNOS_DISPLAY_TYPE_HDMI, + .is_connected = drm_hdmi_is_connected, + .get_edid = drm_hdmi_get_edid, + .check_timing = drm_hdmi_check_timing, + .power_on = drm_hdmi_power_on, +}; + +static int drm_hdmi_enable_vblank(struct device *subdrv_dev) +{ + struct drm_hdmi_context *ctx = to_context(subdrv_dev); + struct exynos_drm_subdrv *subdrv = &ctx->subdrv; + struct exynos_drm_manager *manager = subdrv->manager; + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (mixer_ops && mixer_ops->enable_vblank) + return mixer_ops->enable_vblank(ctx->mixer_ctx->ctx, + manager->pipe); + + return 0; +} + +static void drm_hdmi_disable_vblank(struct device *subdrv_dev) +{ + struct drm_hdmi_context *ctx = to_context(subdrv_dev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (mixer_ops && mixer_ops->disable_vblank) + return mixer_ops->disable_vblank(ctx->mixer_ctx->ctx); +} + +static void drm_hdmi_mode_fixup(struct device *subdrv_dev, + struct drm_connector *connector, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + struct drm_hdmi_context *ctx = to_context(subdrv_dev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (hdmi_ops && hdmi_ops->mode_fixup) + hdmi_ops->mode_fixup(ctx->hdmi_ctx->ctx, connector, mode, + adjusted_mode); +} + +static void drm_hdmi_mode_set(struct device *subdrv_dev, void *mode) +{ + struct drm_hdmi_context *ctx = to_context(subdrv_dev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (hdmi_ops && hdmi_ops->mode_set) + hdmi_ops->mode_set(ctx->hdmi_ctx->ctx, mode); +} + +static void drm_hdmi_get_max_resol(struct device *subdrv_dev, + unsigned int *width, unsigned int *height) +{ + struct drm_hdmi_context *ctx = to_context(subdrv_dev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (hdmi_ops && hdmi_ops->get_max_resol) + hdmi_ops->get_max_resol(ctx->hdmi_ctx->ctx, width, height); +} + +static void drm_hdmi_commit(struct device *subdrv_dev) +{ + struct drm_hdmi_context *ctx = to_context(subdrv_dev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (hdmi_ops && hdmi_ops->commit) + hdmi_ops->commit(ctx->hdmi_ctx->ctx); +} + +static void drm_hdmi_dpms(struct device *subdrv_dev, int mode) +{ + struct drm_hdmi_context *ctx = to_context(subdrv_dev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + switch (mode) { + case DRM_MODE_DPMS_ON: + break; + case DRM_MODE_DPMS_STANDBY: + case DRM_MODE_DPMS_SUSPEND: + case DRM_MODE_DPMS_OFF: + if (hdmi_ops && hdmi_ops->disable) + hdmi_ops->disable(ctx->hdmi_ctx->ctx); + break; + default: + DRM_DEBUG_KMS("unkown dps mode: %d\n", mode); + break; + } +} + +static struct exynos_drm_manager_ops drm_hdmi_manager_ops = { + .dpms = drm_hdmi_dpms, + .enable_vblank = drm_hdmi_enable_vblank, + .disable_vblank = drm_hdmi_disable_vblank, + .mode_fixup = drm_hdmi_mode_fixup, + .mode_set = drm_hdmi_mode_set, + .get_max_resol = drm_hdmi_get_max_resol, + .commit = drm_hdmi_commit, +}; + +static void drm_mixer_mode_set(struct device *subdrv_dev, + struct exynos_drm_overlay *overlay) +{ + struct drm_hdmi_context *ctx = to_context(subdrv_dev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (mixer_ops && mixer_ops->win_mode_set) + mixer_ops->win_mode_set(ctx->mixer_ctx->ctx, overlay); +} + +static void drm_mixer_commit(struct device *subdrv_dev, int zpos) +{ + struct drm_hdmi_context *ctx = to_context(subdrv_dev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (mixer_ops && mixer_ops->win_commit) + mixer_ops->win_commit(ctx->mixer_ctx->ctx, zpos); +} + +static void drm_mixer_disable(struct device *subdrv_dev, int zpos) +{ + struct drm_hdmi_context *ctx = to_context(subdrv_dev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (mixer_ops && mixer_ops->win_disable) + mixer_ops->win_disable(ctx->mixer_ctx->ctx, zpos); +} + +static struct exynos_drm_overlay_ops drm_hdmi_overlay_ops = { + .mode_set = drm_mixer_mode_set, + .commit = drm_mixer_commit, + .disable = drm_mixer_disable, +}; + +static struct exynos_drm_manager hdmi_manager = { + .pipe = -1, + .ops = &drm_hdmi_manager_ops, + .overlay_ops = &drm_hdmi_overlay_ops, + .display_ops = &drm_hdmi_display_ops, +}; + +static int hdmi_subdrv_probe(struct drm_device *drm_dev, + struct device *dev) +{ + struct exynos_drm_subdrv *subdrv = to_subdrv(dev); + struct drm_hdmi_context *ctx; + struct platform_device *pdev = to_platform_device(dev); + struct exynos_drm_common_hdmi_pd *pd; + + DRM_DEBUG_KMS("%s\n", __FILE__); + + pd = pdev->dev.platform_data; + + if (!pd) { + DRM_DEBUG_KMS("platform data is null.\n"); + return -EFAULT; + } + + if (!pd->hdmi_dev) { + DRM_DEBUG_KMS("hdmi device is null.\n"); + return -EFAULT; + } + + if (!pd->mixer_dev) { + DRM_DEBUG_KMS("mixer device is null.\n"); + return -EFAULT; + } + + ctx = get_ctx_from_subdrv(subdrv); + + ctx->hdmi_ctx = (struct exynos_drm_hdmi_context *) + to_context(pd->hdmi_dev); + if (!ctx->hdmi_ctx) { + DRM_DEBUG_KMS("hdmi context is null.\n"); + return -EFAULT; + } + + ctx->hdmi_ctx->drm_dev = drm_dev; + + ctx->mixer_ctx = (struct exynos_drm_hdmi_context *) + to_context(pd->mixer_dev); + if (!ctx->mixer_ctx) { + DRM_DEBUG_KMS("mixer context is null.\n"); + return -EFAULT; + } + + ctx->mixer_ctx->drm_dev = drm_dev; + + return 0; +} + +static int __devinit exynos_drm_hdmi_probe(struct platform_device *pdev) +{ + struct device *dev = &pdev->dev; + struct exynos_drm_subdrv *subdrv; + struct drm_hdmi_context *ctx; + + DRM_DEBUG_KMS("%s\n", __FILE__); + + ctx = kzalloc(sizeof(*ctx), GFP_KERNEL); + if (!ctx) { + DRM_LOG_KMS("failed to alloc common hdmi context.\n"); + return -ENOMEM; + } + + subdrv = &ctx->subdrv; + + subdrv->dev = dev; + subdrv->manager = &hdmi_manager; + subdrv->probe = hdmi_subdrv_probe; + + platform_set_drvdata(pdev, subdrv); + + exynos_drm_subdrv_register(subdrv); + + return 0; +} + +static int hdmi_runtime_suspend(struct device *dev) +{ + DRM_DEBUG_KMS("%s\n", __FILE__); + + return 0; +} + +static int hdmi_runtime_resume(struct device *dev) +{ + DRM_DEBUG_KMS("%s\n", __FILE__); + + return 0; +} + +static const struct dev_pm_ops hdmi_pm_ops = { + .runtime_suspend = hdmi_runtime_suspend, + .runtime_resume = hdmi_runtime_resume, +}; + +static int __devexit exynos_drm_hdmi_remove(struct platform_device *pdev) +{ + struct drm_hdmi_context *ctx = platform_get_drvdata(pdev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + exynos_drm_subdrv_unregister(&ctx->subdrv); + kfree(ctx); + + return 0; +} + +struct platform_driver exynos_drm_common_hdmi_driver = { + .probe = exynos_drm_hdmi_probe, + .remove = __devexit_p(exynos_drm_hdmi_remove), + .driver = { + .name = "exynos-drm-hdmi", + .owner = THIS_MODULE, + .pm = &hdmi_pm_ops, + }, +}; diff --git a/drivers/gpu/drm/exynos/exynos_drm_hdmi.h b/drivers/gpu/drm/exynos/exynos_drm_hdmi.h new file mode 100644 index 0000000..f3ae192 --- /dev/null +++ b/drivers/gpu/drm/exynos/exynos_drm_hdmi.h @@ -0,0 +1,73 @@ +/* exynos_drm_hdmi.h + * + * Copyright (c) 2011 Samsung Electronics Co., Ltd. + * Authoer: Inki Dae + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * VA LINUX SYSTEMS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ + +#ifndef _EXYNOS_DRM_HDMI_H_ +#define _EXYNOS_DRM_HDMI_H_ + +/* + * exynos hdmi common context structure. + * + * @drm_dev: pointer to drm_device. + * @ctx: pointer to the context of specific device driver. + * this context should be hdmi_context or mixer_context. + */ +struct exynos_drm_hdmi_context { + struct drm_device *drm_dev; + void *ctx; +}; + +struct exynos_hdmi_ops { + /* display */ + bool (*is_connected)(void *ctx); + int (*get_edid)(void *ctx, struct drm_connector *connector, + u8 *edid, int len); + int (*check_timing)(void *ctx, void *timing); + int (*power_on)(void *ctx, int mode); + + /* manager */ + void (*mode_fixup)(void *ctx, struct drm_connector *connector, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode); + void (*mode_set)(void *ctx, void *mode); + void (*get_max_resol)(void *ctx, unsigned int *width, + unsigned int *height); + void (*commit)(void *ctx); + void (*disable)(void *ctx); +}; + +struct exynos_mixer_ops { + /* manager */ + int (*enable_vblank)(void *ctx, int pipe); + void (*disable_vblank)(void *ctx); + + /* overlay */ + void (*win_mode_set)(void *ctx, struct exynos_drm_overlay *overlay); + void (*win_commit)(void *ctx, int zpos); + void (*win_disable)(void *ctx, int zpos); +}; + +void exynos_hdmi_ops_register(struct exynos_hdmi_ops *ops); +void exynos_mixer_ops_register(struct exynos_mixer_ops *ops); +#endif diff --git a/drivers/gpu/drm/exynos/exynos_drm_plane.c b/drivers/gpu/drm/exynos/exynos_drm_plane.c new file mode 100644 index 0000000..f92fe4c --- /dev/null +++ b/drivers/gpu/drm/exynos/exynos_drm_plane.c @@ -0,0 +1,171 @@ +/* + * Copyright (C) 2011 Samsung Electronics Co.Ltd + * Authors: Joonyoung Shim + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or (at your + * option) any later version. + * + */ + +#include "drmP.h" + +#include "exynos_drm.h" +#include "exynos_drm_crtc.h" +#include "exynos_drm_drv.h" +#include "exynos_drm_encoder.h" + +struct exynos_plane { + struct drm_plane base; + struct exynos_drm_overlay overlay; + bool enabled; +}; + +static const uint32_t formats[] = { + DRM_FORMAT_XRGB8888, + DRM_FORMAT_ARGB8888, + DRM_FORMAT_NV12, + DRM_FORMAT_NV12M, + DRM_FORMAT_NV12MT, +}; + +static int +exynos_update_plane(struct drm_plane *plane, struct drm_crtc *crtc, + struct drm_framebuffer *fb, int crtc_x, int crtc_y, + unsigned int crtc_w, unsigned int crtc_h, + uint32_t src_x, uint32_t src_y, + uint32_t src_w, uint32_t src_h) +{ + struct exynos_plane *exynos_plane = + container_of(plane, struct exynos_plane, base); + struct exynos_drm_overlay *overlay = &exynos_plane->overlay; + struct exynos_drm_crtc_pos pos; + unsigned int x = src_x >> 16; + unsigned int y = src_y >> 16; + int ret; + + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + memset(&pos, 0, sizeof(struct exynos_drm_crtc_pos)); + pos.crtc_x = crtc_x; + pos.crtc_y = crtc_y; + pos.crtc_w = crtc_w; + pos.crtc_h = crtc_h; + + pos.fb_x = x; + pos.fb_y = y; + + /* TODO: scale feature */ + ret = exynos_drm_overlay_update(overlay, fb, &crtc->mode, &pos); + if (ret < 0) + return ret; + + exynos_drm_fn_encoder(crtc, overlay, + exynos_drm_encoder_crtc_mode_set); + exynos_drm_fn_encoder(crtc, &overlay->zpos, + exynos_drm_encoder_crtc_plane_commit); + + exynos_plane->enabled = true; + + return 0; +} + +static int exynos_disable_plane(struct drm_plane *plane) +{ + struct exynos_plane *exynos_plane = + container_of(plane, struct exynos_plane, base); + struct exynos_drm_overlay *overlay = &exynos_plane->overlay; + + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + if (!exynos_plane->enabled) + return 0; + + exynos_drm_fn_encoder(plane->crtc, &overlay->zpos, + exynos_drm_encoder_crtc_disable); + + exynos_plane->enabled = false; + exynos_plane->overlay.zpos = DEFAULT_ZPOS; + + return 0; +} + +static void exynos_plane_destroy(struct drm_plane *plane) +{ + struct exynos_plane *exynos_plane = + container_of(plane, struct exynos_plane, base); + + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + exynos_disable_plane(plane); + drm_plane_cleanup(plane); + kfree(exynos_plane); +} + +static struct drm_plane_funcs exynos_plane_funcs = { + .update_plane = exynos_update_plane, + .disable_plane = exynos_disable_plane, + .destroy = exynos_plane_destroy, +}; + +int exynos_plane_init(struct drm_device *dev, unsigned int nr) +{ + struct exynos_plane *exynos_plane; + uint32_t possible_crtcs; + + exynos_plane = kzalloc(sizeof(struct exynos_plane), GFP_KERNEL); + if (!exynos_plane) + return -ENOMEM; + + /* all CRTCs are available */ + possible_crtcs = (1 << MAX_CRTC) - 1; + + exynos_plane->overlay.zpos = DEFAULT_ZPOS; + + return drm_plane_init(dev, &exynos_plane->base, possible_crtcs, + &exynos_plane_funcs, formats, ARRAY_SIZE(formats), + false); +} + +int exynos_plane_set_zpos_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + struct drm_exynos_plane_set_zpos *zpos_req = data; + struct drm_mode_object *obj; + struct drm_plane *plane; + struct exynos_plane *exynos_plane; + int ret = 0; + + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + if (!drm_core_check_feature(dev, DRIVER_MODESET)) + return -EINVAL; + + if (zpos_req->zpos < 0 || zpos_req->zpos >= MAX_PLANE) { + if (zpos_req->zpos != DEFAULT_ZPOS) { + DRM_ERROR("zpos not within limits\n"); + return -EINVAL; + } + } + + mutex_lock(&dev->mode_config.mutex); + + obj = drm_mode_object_find(dev, zpos_req->plane_id, + DRM_MODE_OBJECT_PLANE); + if (!obj) { + DRM_DEBUG_KMS("Unknown plane ID %d\n", + zpos_req->plane_id); + ret = -EINVAL; + goto out; + } + + plane = obj_to_plane(obj); + exynos_plane = container_of(plane, struct exynos_plane, base); + + exynos_plane->overlay.zpos = zpos_req->zpos; + +out: + mutex_unlock(&dev->mode_config.mutex); + return ret; +} diff --git a/drivers/gpu/drm/exynos/exynos_drm_plane.h b/drivers/gpu/drm/exynos/exynos_drm_plane.h new file mode 100644 index 0000000..16b71f8 --- /dev/null +++ b/drivers/gpu/drm/exynos/exynos_drm_plane.h @@ -0,0 +1,14 @@ +/* + * Copyright (C) 2011 Samsung Electronics Co.Ltd + * Authors: Joonyoung Shim + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or (at your + * option) any later version. + * + */ + +int exynos_plane_init(struct drm_device *dev, unsigned int nr); +int exynos_plane_set_zpos_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv); diff --git a/drivers/gpu/drm/exynos/exynos_drm_vidi.c b/drivers/gpu/drm/exynos/exynos_drm_vidi.c new file mode 100644 index 0000000..7b9c153 --- /dev/null +++ b/drivers/gpu/drm/exynos/exynos_drm_vidi.c @@ -0,0 +1,680 @@ +/* exynos_drm_vidi.c + * + * Copyright (C) 2012 Samsung Electronics Co.Ltd + * Authors: + * Inki Dae + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or (at your + * option) any later version. + * + */ +#include "drmP.h" + +#include +#include +#include + +#include + +#include "drm_edid.h" +#include "drm_crtc_helper.h" + +#include "exynos_drm_drv.h" +#include "exynos_drm_crtc.h" +#include "exynos_drm_encoder.h" + +/* vidi has totally three virtual windows. */ +#define WINDOWS_NR 3 + +#define get_vidi_context(dev) platform_get_drvdata(to_platform_device(dev)) + +struct vidi_win_data { + unsigned int offset_x; + unsigned int offset_y; + unsigned int ovl_width; + unsigned int ovl_height; + unsigned int fb_width; + unsigned int fb_height; + unsigned int bpp; + dma_addr_t dma_addr; + void __iomem *vaddr; + unsigned int buf_offsize; + unsigned int line_size; /* bytes */ + bool enabled; +}; + +struct vidi_context { + struct exynos_drm_subdrv subdrv; + struct drm_crtc *crtc; + struct vidi_win_data win_data[WINDOWS_NR]; + struct edid *raw_edid; + unsigned int clkdiv; + unsigned int default_win; + unsigned long irq_flags; + unsigned int connected; + bool vblank_on; + bool suspended; + struct work_struct work; + struct mutex lock; +}; + +static const char fake_edid_info[] = { + 0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00, 0x4c, 0x2d, 0x05, 0x05, + 0x00, 0x00, 0x00, 0x00, 0x30, 0x12, 0x01, 0x03, 0x80, 0x10, 0x09, 0x78, + 0x0a, 0xee, 0x91, 0xa3, 0x54, 0x4c, 0x99, 0x26, 0x0f, 0x50, 0x54, 0xbd, + 0xee, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, + 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x66, 0x21, 0x50, 0xb0, 0x51, 0x00, + 0x1b, 0x30, 0x40, 0x70, 0x36, 0x00, 0xa0, 0x5a, 0x00, 0x00, 0x00, 0x1e, + 0x01, 0x1d, 0x00, 0x72, 0x51, 0xd0, 0x1e, 0x20, 0x6e, 0x28, 0x55, 0x00, + 0xa0, 0x5a, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, 0xfd, 0x00, 0x18, + 0x4b, 0x1a, 0x44, 0x17, 0x00, 0x0a, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, + 0x00, 0x00, 0x00, 0xfc, 0x00, 0x53, 0x41, 0x4d, 0x53, 0x55, 0x4e, 0x47, + 0x0a, 0x20, 0x20, 0x20, 0x20, 0x20, 0x01, 0xbc, 0x02, 0x03, 0x1e, 0xf1, + 0x46, 0x84, 0x05, 0x03, 0x10, 0x20, 0x22, 0x23, 0x09, 0x07, 0x07, 0x83, + 0x01, 0x00, 0x00, 0xe2, 0x00, 0x0f, 0x67, 0x03, 0x0c, 0x00, 0x10, 0x00, + 0xb8, 0x2d, 0x01, 0x1d, 0x80, 0x18, 0x71, 0x1c, 0x16, 0x20, 0x58, 0x2c, + 0x25, 0x00, 0xa0, 0x5a, 0x00, 0x00, 0x00, 0x9e, 0x8c, 0x0a, 0xd0, 0x8a, + 0x20, 0xe0, 0x2d, 0x10, 0x10, 0x3e, 0x96, 0x00, 0xa0, 0x5a, 0x00, 0x00, + 0x00, 0x18, 0x02, 0x3a, 0x80, 0x18, 0x71, 0x38, 0x2d, 0x40, 0x58, 0x2c, + 0x45, 0x00, 0xa0, 0x5a, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x06 +}; + +static void vidi_fake_vblank_handler(struct work_struct *work); + +static bool vidi_display_is_connected(struct device *dev) +{ + struct vidi_context *ctx = get_vidi_context(dev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + /* + * connection request would come from user side + * to do hotplug through specific ioctl. + */ + return ctx->connected ? true : false; +} + +static int vidi_get_edid(struct device *dev, struct drm_connector *connector, + u8 *edid, int len) +{ + struct vidi_context *ctx = get_vidi_context(dev); + struct edid *raw_edid; + + DRM_DEBUG_KMS("%s\n", __FILE__); + + /* + * the edid data comes from user side and it would be set + * to ctx->raw_edid through specific ioctl. + */ + if (!ctx->raw_edid) { + DRM_DEBUG_KMS("raw_edid is null.\n"); + return -EFAULT; + } + + raw_edid = kzalloc(len, GFP_KERNEL); + if (!raw_edid) { + DRM_DEBUG_KMS("failed to allocate raw_edid.\n"); + return -ENOMEM; + } + + memcpy(raw_edid, ctx->raw_edid, min((1 + ctx->raw_edid->extensions) + * EDID_LENGTH, len)); + + /* attach the edid data to connector. */ + connector->display_info.raw_edid = (char *)raw_edid; + + memcpy(edid, ctx->raw_edid, min((1 + ctx->raw_edid->extensions) + * EDID_LENGTH, len)); + + return 0; +} + +static void *vidi_get_panel(struct device *dev) +{ + DRM_DEBUG_KMS("%s\n", __FILE__); + + /* TODO. */ + + return NULL; +} + +static int vidi_check_timing(struct device *dev, void *timing) +{ + DRM_DEBUG_KMS("%s\n", __FILE__); + + /* TODO. */ + + return 0; +} + +static int vidi_display_power_on(struct device *dev, int mode) +{ + DRM_DEBUG_KMS("%s\n", __FILE__); + + /* TODO */ + + return 0; +} + +static struct exynos_drm_display_ops vidi_display_ops = { + .type = EXYNOS_DISPLAY_TYPE_VIDI, + .is_connected = vidi_display_is_connected, + .get_edid = vidi_get_edid, + .get_panel = vidi_get_panel, + .check_timing = vidi_check_timing, + .power_on = vidi_display_power_on, +}; + +static void vidi_dpms(struct device *subdrv_dev, int mode) +{ + struct vidi_context *ctx = get_vidi_context(subdrv_dev); + + DRM_DEBUG_KMS("%s, %d\n", __FILE__, mode); + + mutex_lock(&ctx->lock); + + switch (mode) { + case DRM_MODE_DPMS_ON: + /* TODO. */ + break; + case DRM_MODE_DPMS_STANDBY: + case DRM_MODE_DPMS_SUSPEND: + case DRM_MODE_DPMS_OFF: + /* TODO. */ + break; + default: + DRM_DEBUG_KMS("unspecified mode %d\n", mode); + break; + } + + mutex_unlock(&ctx->lock); +} + +static void vidi_apply(struct device *subdrv_dev) +{ + struct vidi_context *ctx = get_vidi_context(subdrv_dev); + struct exynos_drm_manager *mgr = ctx->subdrv.manager; + struct exynos_drm_manager_ops *mgr_ops = mgr->ops; + struct exynos_drm_overlay_ops *ovl_ops = mgr->overlay_ops; + struct vidi_win_data *win_data; + int i; + + DRM_DEBUG_KMS("%s\n", __FILE__); + + for (i = 0; i < WINDOWS_NR; i++) { + win_data = &ctx->win_data[i]; + if (win_data->enabled && (ovl_ops && ovl_ops->commit)) + ovl_ops->commit(subdrv_dev, i); + } + + if (mgr_ops && mgr_ops->commit) + mgr_ops->commit(subdrv_dev); +} + +static void vidi_commit(struct device *dev) +{ + struct vidi_context *ctx = get_vidi_context(dev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (ctx->suspended) + return; +} + +static int vidi_enable_vblank(struct device *dev) +{ + struct vidi_context *ctx = get_vidi_context(dev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (ctx->suspended) + return -EPERM; + + if (!test_and_set_bit(0, &ctx->irq_flags)) + ctx->vblank_on = true; + + return 0; +} + +static void vidi_disable_vblank(struct device *dev) +{ + struct vidi_context *ctx = get_vidi_context(dev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (ctx->suspended) + return; + + if (test_and_clear_bit(0, &ctx->irq_flags)) + ctx->vblank_on = false; +} + +static struct exynos_drm_manager_ops vidi_manager_ops = { + .dpms = vidi_dpms, + .apply = vidi_apply, + .commit = vidi_commit, + .enable_vblank = vidi_enable_vblank, + .disable_vblank = vidi_disable_vblank, +}; + +static void vidi_win_mode_set(struct device *dev, + struct exynos_drm_overlay *overlay) +{ + struct vidi_context *ctx = get_vidi_context(dev); + struct vidi_win_data *win_data; + int win; + unsigned long offset; + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (!overlay) { + dev_err(dev, "overlay is NULL\n"); + return; + } + + win = overlay->zpos; + if (win == DEFAULT_ZPOS) + win = ctx->default_win; + + if (win < 0 || win > WINDOWS_NR) + return; + + offset = overlay->fb_x * (overlay->bpp >> 3); + offset += overlay->fb_y * overlay->pitch; + + DRM_DEBUG_KMS("offset = 0x%lx, pitch = %x\n", offset, overlay->pitch); + + win_data = &ctx->win_data[win]; + + win_data->offset_x = overlay->crtc_x; + win_data->offset_y = overlay->crtc_y; + win_data->ovl_width = overlay->crtc_width; + win_data->ovl_height = overlay->crtc_height; + win_data->fb_width = overlay->fb_width; + win_data->fb_height = overlay->fb_height; + win_data->dma_addr = overlay->dma_addr[0] + offset; + win_data->vaddr = overlay->vaddr[0] + offset; + win_data->bpp = overlay->bpp; + win_data->buf_offsize = (overlay->fb_width - overlay->crtc_width) * + (overlay->bpp >> 3); + win_data->line_size = overlay->crtc_width * (overlay->bpp >> 3); + + /* + * some parts of win_data should be transferred to user side + * through specific ioctl. + */ + + DRM_DEBUG_KMS("offset_x = %d, offset_y = %d\n", + win_data->offset_x, win_data->offset_y); + DRM_DEBUG_KMS("ovl_width = %d, ovl_height = %d\n", + win_data->ovl_width, win_data->ovl_height); + DRM_DEBUG_KMS("paddr = 0x%lx, vaddr = 0x%lx\n", + (unsigned long)win_data->dma_addr, + (unsigned long)win_data->vaddr); + DRM_DEBUG_KMS("fb_width = %d, crtc_width = %d\n", + overlay->fb_width, overlay->crtc_width); +} + +static void vidi_win_commit(struct device *dev, int zpos) +{ + struct vidi_context *ctx = get_vidi_context(dev); + struct vidi_win_data *win_data; + int win = zpos; + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (ctx->suspended) + return; + + if (win == DEFAULT_ZPOS) + win = ctx->default_win; + + if (win < 0 || win > WINDOWS_NR) + return; + + win_data = &ctx->win_data[win]; + + win_data->enabled = true; + + DRM_DEBUG_KMS("dma_addr = 0x%x\n", win_data->dma_addr); + + if (ctx->vblank_on) + schedule_work(&ctx->work); +} + +static void vidi_win_disable(struct device *dev, int zpos) +{ + struct vidi_context *ctx = get_vidi_context(dev); + struct vidi_win_data *win_data; + int win = zpos; + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (win == DEFAULT_ZPOS) + win = ctx->default_win; + + if (win < 0 || win > WINDOWS_NR) + return; + + win_data = &ctx->win_data[win]; + win_data->enabled = false; + + /* TODO. */ +} + +static struct exynos_drm_overlay_ops vidi_overlay_ops = { + .mode_set = vidi_win_mode_set, + .commit = vidi_win_commit, + .disable = vidi_win_disable, +}; + +static struct exynos_drm_manager vidi_manager = { + .pipe = -1, + .ops = &vidi_manager_ops, + .overlay_ops = &vidi_overlay_ops, + .display_ops = &vidi_display_ops, +}; + +static void vidi_finish_pageflip(struct drm_device *drm_dev, int crtc) +{ + struct exynos_drm_private *dev_priv = drm_dev->dev_private; + struct drm_pending_vblank_event *e, *t; + struct timeval now; + unsigned long flags; + bool is_checked = false; + + spin_lock_irqsave(&drm_dev->event_lock, flags); + + list_for_each_entry_safe(e, t, &dev_priv->pageflip_event_list, + base.link) { + /* if event's pipe isn't same as crtc then ignore it. */ + if (crtc != e->pipe) + continue; + + is_checked = true; + + do_gettimeofday(&now); + e->event.sequence = 0; + e->event.tv_sec = now.tv_sec; + e->event.tv_usec = now.tv_usec; + + list_move_tail(&e->base.link, &e->base.file_priv->event_list); + wake_up_interruptible(&e->base.file_priv->event_wait); + } + + if (is_checked) { + /* + * call drm_vblank_put only in case that drm_vblank_get was + * called. + */ + if (atomic_read(&drm_dev->vblank_refcount[crtc]) > 0) + drm_vblank_put(drm_dev, crtc); + + /* + * don't off vblank if vblank_disable_allowed is 1, + * because vblank would be off by timer handler. + */ + if (!drm_dev->vblank_disable_allowed) + drm_vblank_off(drm_dev, crtc); + } + + spin_unlock_irqrestore(&drm_dev->event_lock, flags); +} + +static void vidi_fake_vblank_handler(struct work_struct *work) +{ + struct vidi_context *ctx = container_of(work, struct vidi_context, + work); + struct exynos_drm_subdrv *subdrv = &ctx->subdrv; + struct exynos_drm_manager *manager = subdrv->manager; + + if (manager->pipe < 0) + return; + + /* refresh rate is about 50Hz. */ + usleep_range(16000, 20000); + + drm_handle_vblank(subdrv->drm_dev, manager->pipe); + vidi_finish_pageflip(subdrv->drm_dev, manager->pipe); +} + +static int vidi_subdrv_probe(struct drm_device *drm_dev, struct device *dev) +{ + DRM_DEBUG_KMS("%s\n", __FILE__); + + /* + * enable drm irq mode. + * - with irq_enabled = 1, we can use the vblank feature. + * + * P.S. note that we wouldn't use drm irq handler but + * just specific driver own one instead because + * drm framework supports only one irq handler. + */ + drm_dev->irq_enabled = 1; + + /* + * with vblank_disable_allowed = 1, vblank interrupt will be disabled + * by drm timer once a current process gives up ownership of + * vblank event.(after drm_vblank_put function is called) + */ + drm_dev->vblank_disable_allowed = 1; + + return 0; +} + +static void vidi_subdrv_remove(struct drm_device *drm_dev) +{ + DRM_DEBUG_KMS("%s\n", __FILE__); + + /* TODO. */ +} + +static int vidi_power_on(struct vidi_context *ctx, bool enable) +{ + struct exynos_drm_subdrv *subdrv = &ctx->subdrv; + struct device *dev = subdrv->dev; + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (enable != false && enable != true) + return -EINVAL; + + if (enable) { + ctx->suspended = false; + + /* if vblank was enabled status, enable it again. */ + if (test_and_clear_bit(0, &ctx->irq_flags)) + vidi_enable_vblank(dev); + + vidi_apply(dev); + } else { + ctx->suspended = true; + } + + return 0; +} + +static int vidi_show_connection(struct device *dev, + struct device_attribute *attr, char *buf) +{ + int rc; + struct vidi_context *ctx = get_vidi_context(dev); + + mutex_lock(&ctx->lock); + + rc = sprintf(buf, "%d\n", ctx->connected); + + mutex_unlock(&ctx->lock); + + return rc; +} + +static int vidi_store_connection(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t len) +{ + struct vidi_context *ctx = get_vidi_context(dev); + int ret; + + DRM_DEBUG_KMS("%s\n", __FILE__); + + ret = kstrtoint(buf, 0, &ctx->connected); + if (ret) + return ret; + + if (ctx->connected > 1) + return -EINVAL; + + DRM_DEBUG_KMS("requested connection.\n"); + + drm_helper_hpd_irq_event(ctx->subdrv.drm_dev); + + return len; +} + +static DEVICE_ATTR(connection, 0644, vidi_show_connection, + vidi_store_connection); + +int vidi_connection_ioctl(struct drm_device *drm_dev, void *data, + struct drm_file *file_priv) +{ + struct vidi_context *ctx = NULL; + struct drm_encoder *encoder; + struct exynos_drm_manager *manager; + struct exynos_drm_display_ops *display_ops; + struct drm_exynos_vidi_connection *vidi = data; + + DRM_DEBUG_KMS("%s\n", __FILE__); + + if (!vidi) { + DRM_DEBUG_KMS("user data for vidi is null.\n"); + return -EINVAL; + } + + if (!vidi->edid) { + DRM_DEBUG_KMS("edid data is null.\n"); + return -EINVAL; + } + + if (vidi->connection > 1) { + DRM_DEBUG_KMS("connection should be 0 or 1.\n"); + return -EINVAL; + } + + list_for_each_entry(encoder, &drm_dev->mode_config.encoder_list, + head) { + manager = exynos_drm_get_manager(encoder); + display_ops = manager->display_ops; + + if (display_ops->type == EXYNOS_DISPLAY_TYPE_VIDI) { + ctx = get_vidi_context(manager->dev); + break; + } + } + + if (!ctx) { + DRM_DEBUG_KMS("not found virtual device type encoder.\n"); + return -EINVAL; + } + + if (ctx->connected == vidi->connection) { + DRM_DEBUG_KMS("same connection request.\n"); + return -EINVAL; + } + + if (vidi->connection) + ctx->raw_edid = (struct edid *)vidi->edid; + + ctx->connected = vidi->connection; + drm_helper_hpd_irq_event(ctx->subdrv.drm_dev); + + return 0; +} + +static int __devinit vidi_probe(struct platform_device *pdev) +{ + struct device *dev = &pdev->dev; + struct vidi_context *ctx; + struct exynos_drm_subdrv *subdrv; + int ret; + + DRM_DEBUG_KMS("%s\n", __FILE__); + + ctx = kzalloc(sizeof(*ctx), GFP_KERNEL); + if (!ctx) + return -ENOMEM; + + ctx->default_win = 0; + + INIT_WORK(&ctx->work, vidi_fake_vblank_handler); + + /* for test */ + ctx->raw_edid = (struct edid *)fake_edid_info; + + subdrv = &ctx->subdrv; + subdrv->dev = dev; + subdrv->manager = &vidi_manager; + subdrv->probe = vidi_subdrv_probe; + subdrv->remove = vidi_subdrv_remove; + + mutex_init(&ctx->lock); + + platform_set_drvdata(pdev, ctx); + + ret = device_create_file(&pdev->dev, &dev_attr_connection); + if (ret < 0) + DRM_INFO("failed to create connection sysfs.\n"); + + exynos_drm_subdrv_register(subdrv); + + return 0; +} + +static int __devexit vidi_remove(struct platform_device *pdev) +{ + struct vidi_context *ctx = platform_get_drvdata(pdev); + + DRM_DEBUG_KMS("%s\n", __FILE__); + + exynos_drm_subdrv_unregister(&ctx->subdrv); + + kfree(ctx); + + return 0; +} + +#ifdef CONFIG_PM_SLEEP +static int vidi_suspend(struct device *dev) +{ + struct vidi_context *ctx = get_vidi_context(dev); + + return vidi_power_on(ctx, false); +} + +static int vidi_resume(struct device *dev) +{ + struct vidi_context *ctx = get_vidi_context(dev); + + return vidi_power_on(ctx, true); +} +#endif + +static const struct dev_pm_ops vidi_pm_ops = { + SET_SYSTEM_SLEEP_PM_OPS(vidi_suspend, vidi_resume) +}; + +struct platform_driver vidi_driver = { + .probe = vidi_probe, + .remove = __devexit_p(vidi_remove), + .driver = { + .name = "exynos-drm-vidi", + .owner = THIS_MODULE, + .pm = &vidi_pm_ops, + }, +}; diff --git a/drivers/gpu/drm/exynos/exynos_drm_vidi.h b/drivers/gpu/drm/exynos/exynos_drm_vidi.h new file mode 100644 index 0000000..a4babe4 --- /dev/null +++ b/drivers/gpu/drm/exynos/exynos_drm_vidi.h @@ -0,0 +1,36 @@ +/* exynos_drm_vidi.h + * + * Copyright (c) 2012 Samsung Electronics Co., Ltd. + * Author: Inki Dae + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * VA LINUX SYSTEMS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ + +#ifndef _EXYNOS_DRM_VIDI_H_ +#define _EXYNOS_DRM_VIDI_H_ + +#ifdef CONFIG_DRM_EXYNOS_VIDI +int vidi_connection_ioctl(struct drm_device *drm_dev, void *data, + struct drm_file *file_priv); +#else +#define vidi_connection_ioctl NULL +#endif + +#endif diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.c b/drivers/gpu/drm/exynos/exynos_hdmi.c new file mode 100644 index 0000000..b003538 --- /dev/null +++ b/drivers/gpu/drm/exynos/exynos_hdmi.c @@ -0,0 +1,2389 @@ +/* + * Copyright (C) 2011 Samsung Electronics Co.Ltd + * Authors: + * Seung-Woo Kim + * Inki Dae + * Joonyoung Shim + * + * Based on drivers/media/video/s5p-tv/hdmi_drv.c + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or (at your + * option) any later version. + * + */ + +#include "drmP.h" +#include "drm_edid.h" +#include "drm_crtc_helper.h" + +#include "regs-hdmi.h" + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include + +#include "exynos_drm_drv.h" +#include "exynos_drm_hdmi.h" + +#include "exynos_hdmi.h" + +#define MAX_WIDTH 1920 +#define MAX_HEIGHT 1080 +#define get_hdmi_context(dev) platform_get_drvdata(to_platform_device(dev)) + +struct hdmi_resources { + struct clk *hdmi; + struct clk *sclk_hdmi; + struct clk *sclk_pixel; + struct clk *sclk_hdmiphy; + struct clk *hdmiphy; + struct regulator_bulk_data *regul_bulk; + int regul_count; +}; + +struct hdmi_context { + struct device *dev; + struct drm_device *drm_dev; + struct fb_videomode *default_timing; + unsigned int is_v13:1; + unsigned int default_win; + unsigned int default_bpp; + bool hpd_handle; + bool enabled; + + struct resource *regs_res; + void __iomem *regs; + unsigned int irq; + struct workqueue_struct *wq; + struct work_struct hotplug_work; + + struct i2c_client *ddc_port; + struct i2c_client *hdmiphy_port; + + /* current hdmiphy conf index */ + int cur_conf; + + struct hdmi_resources res; + void *parent_ctx; +}; + +/* HDMI Version 1.3 */ +static const u8 hdmiphy_v13_conf27[32] = { + 0x01, 0x05, 0x00, 0xD8, 0x10, 0x1C, 0x30, 0x40, + 0x6B, 0x10, 0x02, 0x51, 0xDF, 0xF2, 0x54, 0x87, + 0x84, 0x00, 0x30, 0x38, 0x00, 0x08, 0x10, 0xE0, + 0x22, 0x40, 0xE3, 0x26, 0x00, 0x00, 0x00, 0x00, +}; + +static const u8 hdmiphy_v13_conf27_027[32] = { + 0x01, 0x05, 0x00, 0xD4, 0x10, 0x9C, 0x09, 0x64, + 0x6B, 0x10, 0x02, 0x51, 0xDF, 0xF2, 0x54, 0x87, + 0x84, 0x00, 0x30, 0x38, 0x00, 0x08, 0x10, 0xE0, + 0x22, 0x40, 0xE3, 0x26, 0x00, 0x00, 0x00, 0x00, +}; + +static const u8 hdmiphy_v13_conf74_175[32] = { + 0x01, 0x05, 0x00, 0xD8, 0x10, 0x9C, 0xef, 0x5B, + 0x6D, 0x10, 0x01, 0x51, 0xef, 0xF3, 0x54, 0xb9, + 0x84, 0x00, 0x30, 0x38, 0x00, 0x08, 0x10, 0xE0, + 0x22, 0x40, 0xa5, 0x26, 0x01, 0x00, 0x00, 0x00, +}; + +static const u8 hdmiphy_v13_conf74_25[32] = { + 0x01, 0x05, 0x00, 0xd8, 0x10, 0x9c, 0xf8, 0x40, + 0x6a, 0x10, 0x01, 0x51, 0xff, 0xf1, 0x54, 0xba, + 0x84, 0x00, 0x10, 0x38, 0x00, 0x08, 0x10, 0xe0, + 0x22, 0x40, 0xa4, 0x26, 0x01, 0x00, 0x00, 0x00, +}; + +static const u8 hdmiphy_v13_conf148_5[32] = { + 0x01, 0x05, 0x00, 0xD8, 0x10, 0x9C, 0xf8, 0x40, + 0x6A, 0x18, 0x00, 0x51, 0xff, 0xF1, 0x54, 0xba, + 0x84, 0x00, 0x10, 0x38, 0x00, 0x08, 0x10, 0xE0, + 0x22, 0x40, 0xa4, 0x26, 0x02, 0x00, 0x00, 0x00, +}; + +struct hdmi_v13_tg_regs { + u8 cmd; + u8 h_fsz_l; + u8 h_fsz_h; + u8 hact_st_l; + u8 hact_st_h; + u8 hact_sz_l; + u8 hact_sz_h; + u8 v_fsz_l; + u8 v_fsz_h; + u8 vsync_l; + u8 vsync_h; + u8 vsync2_l; + u8 vsync2_h; + u8 vact_st_l; + u8 vact_st_h; + u8 vact_sz_l; + u8 vact_sz_h; + u8 field_chg_l; + u8 field_chg_h; + u8 vact_st2_l; + u8 vact_st2_h; + u8 vsync_top_hdmi_l; + u8 vsync_top_hdmi_h; + u8 vsync_bot_hdmi_l; + u8 vsync_bot_hdmi_h; + u8 field_top_hdmi_l; + u8 field_top_hdmi_h; + u8 field_bot_hdmi_l; + u8 field_bot_hdmi_h; +}; + +struct hdmi_v13_core_regs { + u8 h_blank[2]; + u8 v_blank[3]; + u8 h_v_line[3]; + u8 vsync_pol[1]; + u8 int_pro_mode[1]; + u8 v_blank_f[3]; + u8 h_sync_gen[3]; + u8 v_sync_gen1[3]; + u8 v_sync_gen2[3]; + u8 v_sync_gen3[3]; +}; + +struct hdmi_v13_preset_conf { + struct hdmi_v13_core_regs core; + struct hdmi_v13_tg_regs tg; +}; + +struct hdmi_v13_conf { + int width; + int height; + int vrefresh; + bool interlace; + const u8 *hdmiphy_data; + const struct hdmi_v13_preset_conf *conf; +}; + +static const struct hdmi_v13_preset_conf hdmi_v13_conf_480p = { + .core = { + .h_blank = {0x8a, 0x00}, + .v_blank = {0x0d, 0x6a, 0x01}, + .h_v_line = {0x0d, 0xa2, 0x35}, + .vsync_pol = {0x01}, + .int_pro_mode = {0x00}, + .v_blank_f = {0x00, 0x00, 0x00}, + .h_sync_gen = {0x0e, 0x30, 0x11}, + .v_sync_gen1 = {0x0f, 0x90, 0x00}, + /* other don't care */ + }, + .tg = { + 0x00, /* cmd */ + 0x5a, 0x03, /* h_fsz */ + 0x8a, 0x00, 0xd0, 0x02, /* hact */ + 0x0d, 0x02, /* v_fsz */ + 0x01, 0x00, 0x33, 0x02, /* vsync */ + 0x2d, 0x00, 0xe0, 0x01, /* vact */ + 0x33, 0x02, /* field_chg */ + 0x49, 0x02, /* vact_st2 */ + 0x01, 0x00, 0x33, 0x02, /* vsync top/bot */ + 0x01, 0x00, 0x33, 0x02, /* field top/bot */ + }, +}; + +static const struct hdmi_v13_preset_conf hdmi_v13_conf_720p60 = { + .core = { + .h_blank = {0x72, 0x01}, + .v_blank = {0xee, 0xf2, 0x00}, + .h_v_line = {0xee, 0x22, 0x67}, + .vsync_pol = {0x00}, + .int_pro_mode = {0x00}, + .v_blank_f = {0x00, 0x00, 0x00}, /* don't care */ + .h_sync_gen = {0x6c, 0x50, 0x02}, + .v_sync_gen1 = {0x0a, 0x50, 0x00}, + .v_sync_gen2 = {0x01, 0x10, 0x00}, + .v_sync_gen3 = {0x01, 0x10, 0x00}, + /* other don't care */ + }, + .tg = { + 0x00, /* cmd */ + 0x72, 0x06, /* h_fsz */ + 0x71, 0x01, 0x01, 0x05, /* hact */ + 0xee, 0x02, /* v_fsz */ + 0x01, 0x00, 0x33, 0x02, /* vsync */ + 0x1e, 0x00, 0xd0, 0x02, /* vact */ + 0x33, 0x02, /* field_chg */ + 0x49, 0x02, /* vact_st2 */ + 0x01, 0x00, 0x01, 0x00, /* vsync top/bot */ + 0x01, 0x00, 0x33, 0x02, /* field top/bot */ + }, +}; + +static const struct hdmi_v13_preset_conf hdmi_v13_conf_1080i50 = { + .core = { + .h_blank = {0xd0, 0x02}, + .v_blank = {0x32, 0xB2, 0x00}, + .h_v_line = {0x65, 0x04, 0xa5}, + .vsync_pol = {0x00}, + .int_pro_mode = {0x01}, + .v_blank_f = {0x49, 0x2A, 0x23}, + .h_sync_gen = {0x0E, 0xEA, 0x08}, + .v_sync_gen1 = {0x07, 0x20, 0x00}, + .v_sync_gen2 = {0x39, 0x42, 0x23}, + .v_sync_gen3 = {0x38, 0x87, 0x73}, + /* other don't care */ + }, + .tg = { + 0x00, /* cmd */ + 0x50, 0x0A, /* h_fsz */ + 0xCF, 0x02, 0x81, 0x07, /* hact */ + 0x65, 0x04, /* v_fsz */ + 0x01, 0x00, 0x33, 0x02, /* vsync */ + 0x16, 0x00, 0x1c, 0x02, /* vact */ + 0x33, 0x02, /* field_chg */ + 0x49, 0x02, /* vact_st2 */ + 0x01, 0x00, 0x33, 0x02, /* vsync top/bot */ + 0x01, 0x00, 0x33, 0x02, /* field top/bot */ + }, +}; + +static const struct hdmi_v13_preset_conf hdmi_v13_conf_1080p50 = { + .core = { + .h_blank = {0xd0, 0x02}, + .v_blank = {0x65, 0x6c, 0x01}, + .h_v_line = {0x65, 0x04, 0xa5}, + .vsync_pol = {0x00}, + .int_pro_mode = {0x00}, + .v_blank_f = {0x00, 0x00, 0x00}, /* don't care */ + .h_sync_gen = {0x0e, 0xea, 0x08}, + .v_sync_gen1 = {0x09, 0x40, 0x00}, + .v_sync_gen2 = {0x01, 0x10, 0x00}, + .v_sync_gen3 = {0x01, 0x10, 0x00}, + /* other don't care */ + }, + .tg = { + 0x00, /* cmd */ + 0x50, 0x0A, /* h_fsz */ + 0xCF, 0x02, 0x81, 0x07, /* hact */ + 0x65, 0x04, /* v_fsz */ + 0x01, 0x00, 0x33, 0x02, /* vsync */ + 0x2d, 0x00, 0x38, 0x04, /* vact */ + 0x33, 0x02, /* field_chg */ + 0x48, 0x02, /* vact_st2 */ + 0x01, 0x00, 0x01, 0x00, /* vsync top/bot */ + 0x01, 0x00, 0x33, 0x02, /* field top/bot */ + }, +}; + +static const struct hdmi_v13_preset_conf hdmi_v13_conf_1080i60 = { + .core = { + .h_blank = {0x18, 0x01}, + .v_blank = {0x32, 0xB2, 0x00}, + .h_v_line = {0x65, 0x84, 0x89}, + .vsync_pol = {0x00}, + .int_pro_mode = {0x01}, + .v_blank_f = {0x49, 0x2A, 0x23}, + .h_sync_gen = {0x56, 0x08, 0x02}, + .v_sync_gen1 = {0x07, 0x20, 0x00}, + .v_sync_gen2 = {0x39, 0x42, 0x23}, + .v_sync_gen3 = {0xa4, 0x44, 0x4a}, + /* other don't care */ + }, + .tg = { + 0x00, /* cmd */ + 0x98, 0x08, /* h_fsz */ + 0x17, 0x01, 0x81, 0x07, /* hact */ + 0x65, 0x04, /* v_fsz */ + 0x01, 0x00, 0x33, 0x02, /* vsync */ + 0x16, 0x00, 0x1c, 0x02, /* vact */ + 0x33, 0x02, /* field_chg */ + 0x49, 0x02, /* vact_st2 */ + 0x01, 0x00, 0x33, 0x02, /* vsync top/bot */ + 0x01, 0x00, 0x33, 0x02, /* field top/bot */ + }, +}; + +static const struct hdmi_v13_preset_conf hdmi_v13_conf_1080p60 = { + .core = { + .h_blank = {0x18, 0x01}, + .v_blank = {0x65, 0x6c, 0x01}, + .h_v_line = {0x65, 0x84, 0x89}, + .vsync_pol = {0x00}, + .int_pro_mode = {0x00}, + .v_blank_f = {0x00, 0x00, 0x00}, /* don't care */ + .h_sync_gen = {0x56, 0x08, 0x02}, + .v_sync_gen1 = {0x09, 0x40, 0x00}, + .v_sync_gen2 = {0x01, 0x10, 0x00}, + .v_sync_gen3 = {0x01, 0x10, 0x00}, + /* other don't care */ + }, + .tg = { + 0x00, /* cmd */ + 0x98, 0x08, /* h_fsz */ + 0x17, 0x01, 0x81, 0x07, /* hact */ + 0x65, 0x04, /* v_fsz */ + 0x01, 0x00, 0x33, 0x02, /* vsync */ + 0x2d, 0x00, 0x38, 0x04, /* vact */ + 0x33, 0x02, /* field_chg */ + 0x48, 0x02, /* vact_st2 */ + 0x01, 0x00, 0x01, 0x00, /* vsync top/bot */ + 0x01, 0x00, 0x33, 0x02, /* field top/bot */ + }, +}; + +static const struct hdmi_v13_conf hdmi_v13_confs[] = { + { 1280, 720, 60, false, hdmiphy_v13_conf74_25, &hdmi_v13_conf_720p60 }, + { 1280, 720, 50, false, hdmiphy_v13_conf74_25, &hdmi_v13_conf_720p60 }, + { 720, 480, 60, false, hdmiphy_v13_conf27_027, &hdmi_v13_conf_480p }, + { 1920, 1080, 50, true, hdmiphy_v13_conf74_25, &hdmi_v13_conf_1080i50 }, + { 1920, 1080, 50, false, hdmiphy_v13_conf148_5, + &hdmi_v13_conf_1080p50 }, + { 1920, 1080, 60, true, hdmiphy_v13_conf74_25, &hdmi_v13_conf_1080i60 }, + { 1920, 1080, 60, false, hdmiphy_v13_conf148_5, + &hdmi_v13_conf_1080p60 }, +}; + +/* HDMI Version 1.4 */ +static const u8 hdmiphy_conf27_027[32] = { + 0x01, 0xd1, 0x2d, 0x72, 0x40, 0x64, 0x12, 0x08, + 0x43, 0xa0, 0x0e, 0xd9, 0x45, 0xa0, 0xac, 0x80, + 0x08, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86, + 0x54, 0xe3, 0x24, 0x00, 0x00, 0x00, 0x01, 0x00, +}; + +static const u8 hdmiphy_conf74_25[32] = { + 0x01, 0xd1, 0x1f, 0x10, 0x40, 0x40, 0xf8, 0x08, + 0x81, 0xa0, 0xba, 0xd8, 0x45, 0xa0, 0xac, 0x80, + 0x3c, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86, + 0x54, 0xa5, 0x24, 0x01, 0x00, 0x00, 0x01, 0x00, +}; + +static const u8 hdmiphy_conf148_5[32] = { + 0x01, 0xd1, 0x1f, 0x00, 0x40, 0x40, 0xf8, 0x08, + 0x81, 0xa0, 0xba, 0xd8, 0x45, 0xa0, 0xac, 0x80, + 0x3c, 0x80, 0x11, 0x04, 0x02, 0x22, 0x44, 0x86, + 0x54, 0x4b, 0x25, 0x03, 0x00, 0x00, 0x01, 0x00, +}; + +struct hdmi_tg_regs { + u8 cmd; + u8 h_fsz_l; + u8 h_fsz_h; + u8 hact_st_l; + u8 hact_st_h; + u8 hact_sz_l; + u8 hact_sz_h; + u8 v_fsz_l; + u8 v_fsz_h; + u8 vsync_l; + u8 vsync_h; + u8 vsync2_l; + u8 vsync2_h; + u8 vact_st_l; + u8 vact_st_h; + u8 vact_sz_l; + u8 vact_sz_h; + u8 field_chg_l; + u8 field_chg_h; + u8 vact_st2_l; + u8 vact_st2_h; + u8 vact_st3_l; + u8 vact_st3_h; + u8 vact_st4_l; + u8 vact_st4_h; + u8 vsync_top_hdmi_l; + u8 vsync_top_hdmi_h; + u8 vsync_bot_hdmi_l; + u8 vsync_bot_hdmi_h; + u8 field_top_hdmi_l; + u8 field_top_hdmi_h; + u8 field_bot_hdmi_l; + u8 field_bot_hdmi_h; + u8 tg_3d; +}; + +struct hdmi_core_regs { + u8 h_blank[2]; + u8 v2_blank[2]; + u8 v1_blank[2]; + u8 v_line[2]; + u8 h_line[2]; + u8 hsync_pol[1]; + u8 vsync_pol[1]; + u8 int_pro_mode[1]; + u8 v_blank_f0[2]; + u8 v_blank_f1[2]; + u8 h_sync_start[2]; + u8 h_sync_end[2]; + u8 v_sync_line_bef_2[2]; + u8 v_sync_line_bef_1[2]; + u8 v_sync_line_aft_2[2]; + u8 v_sync_line_aft_1[2]; + u8 v_sync_line_aft_pxl_2[2]; + u8 v_sync_line_aft_pxl_1[2]; + u8 v_blank_f2[2]; /* for 3D mode */ + u8 v_blank_f3[2]; /* for 3D mode */ + u8 v_blank_f4[2]; /* for 3D mode */ + u8 v_blank_f5[2]; /* for 3D mode */ + u8 v_sync_line_aft_3[2]; + u8 v_sync_line_aft_4[2]; + u8 v_sync_line_aft_5[2]; + u8 v_sync_line_aft_6[2]; + u8 v_sync_line_aft_pxl_3[2]; + u8 v_sync_line_aft_pxl_4[2]; + u8 v_sync_line_aft_pxl_5[2]; + u8 v_sync_line_aft_pxl_6[2]; + u8 vact_space_1[2]; + u8 vact_space_2[2]; + u8 vact_space_3[2]; + u8 vact_space_4[2]; + u8 vact_space_5[2]; + u8 vact_space_6[2]; +}; + +struct hdmi_preset_conf { + struct hdmi_core_regs core; + struct hdmi_tg_regs tg; +}; + +struct hdmi_conf { + int width; + int height; + int vrefresh; + bool interlace; + const u8 *hdmiphy_data; + const struct hdmi_preset_conf *conf; +}; + +static const struct hdmi_preset_conf hdmi_conf_480p60 = { + .core = { + .h_blank = {0x8a, 0x00}, + .v2_blank = {0x0d, 0x02}, + .v1_blank = {0x2d, 0x00}, + .v_line = {0x0d, 0x02}, + .h_line = {0x5a, 0x03}, + .hsync_pol = {0x01}, + .vsync_pol = {0x01}, + .int_pro_mode = {0x00}, + .v_blank_f0 = {0xff, 0xff}, + .v_blank_f1 = {0xff, 0xff}, + .h_sync_start = {0x0e, 0x00}, + .h_sync_end = {0x4c, 0x00}, + .v_sync_line_bef_2 = {0x0f, 0x00}, + .v_sync_line_bef_1 = {0x09, 0x00}, + .v_sync_line_aft_2 = {0xff, 0xff}, + .v_sync_line_aft_1 = {0xff, 0xff}, + .v_sync_line_aft_pxl_2 = {0xff, 0xff}, + .v_sync_line_aft_pxl_1 = {0xff, 0xff}, + .v_blank_f2 = {0xff, 0xff}, + .v_blank_f3 = {0xff, 0xff}, + .v_blank_f4 = {0xff, 0xff}, + .v_blank_f5 = {0xff, 0xff}, + .v_sync_line_aft_3 = {0xff, 0xff}, + .v_sync_line_aft_4 = {0xff, 0xff}, + .v_sync_line_aft_5 = {0xff, 0xff}, + .v_sync_line_aft_6 = {0xff, 0xff}, + .v_sync_line_aft_pxl_3 = {0xff, 0xff}, + .v_sync_line_aft_pxl_4 = {0xff, 0xff}, + .v_sync_line_aft_pxl_5 = {0xff, 0xff}, + .v_sync_line_aft_pxl_6 = {0xff, 0xff}, + .vact_space_1 = {0xff, 0xff}, + .vact_space_2 = {0xff, 0xff}, + .vact_space_3 = {0xff, 0xff}, + .vact_space_4 = {0xff, 0xff}, + .vact_space_5 = {0xff, 0xff}, + .vact_space_6 = {0xff, 0xff}, + /* other don't care */ + }, + .tg = { + 0x00, /* cmd */ + 0x5a, 0x03, /* h_fsz */ + 0x8a, 0x00, 0xd0, 0x02, /* hact */ + 0x0d, 0x02, /* v_fsz */ + 0x01, 0x00, 0x33, 0x02, /* vsync */ + 0x2d, 0x00, 0xe0, 0x01, /* vact */ + 0x33, 0x02, /* field_chg */ + 0x48, 0x02, /* vact_st2 */ + 0x00, 0x00, /* vact_st3 */ + 0x00, 0x00, /* vact_st4 */ + 0x01, 0x00, 0x01, 0x00, /* vsync top/bot */ + 0x01, 0x00, 0x33, 0x02, /* field top/bot */ + 0x00, /* 3d FP */ + }, +}; + +static const struct hdmi_preset_conf hdmi_conf_720p50 = { + .core = { + .h_blank = {0xbc, 0x02}, + .v2_blank = {0xee, 0x02}, + .v1_blank = {0x1e, 0x00}, + .v_line = {0xee, 0x02}, + .h_line = {0xbc, 0x07}, + .hsync_pol = {0x00}, + .vsync_pol = {0x00}, + .int_pro_mode = {0x00}, + .v_blank_f0 = {0xff, 0xff}, + .v_blank_f1 = {0xff, 0xff}, + .h_sync_start = {0xb6, 0x01}, + .h_sync_end = {0xde, 0x01}, + .v_sync_line_bef_2 = {0x0a, 0x00}, + .v_sync_line_bef_1 = {0x05, 0x00}, + .v_sync_line_aft_2 = {0xff, 0xff}, + .v_sync_line_aft_1 = {0xff, 0xff}, + .v_sync_line_aft_pxl_2 = {0xff, 0xff}, + .v_sync_line_aft_pxl_1 = {0xff, 0xff}, + .v_blank_f2 = {0xff, 0xff}, + .v_blank_f3 = {0xff, 0xff}, + .v_blank_f4 = {0xff, 0xff}, + .v_blank_f5 = {0xff, 0xff}, + .v_sync_line_aft_3 = {0xff, 0xff}, + .v_sync_line_aft_4 = {0xff, 0xff}, + .v_sync_line_aft_5 = {0xff, 0xff}, + .v_sync_line_aft_6 = {0xff, 0xff}, + .v_sync_line_aft_pxl_3 = {0xff, 0xff}, + .v_sync_line_aft_pxl_4 = {0xff, 0xff}, + .v_sync_line_aft_pxl_5 = {0xff, 0xff}, + .v_sync_line_aft_pxl_6 = {0xff, 0xff}, + .vact_space_1 = {0xff, 0xff}, + .vact_space_2 = {0xff, 0xff}, + .vact_space_3 = {0xff, 0xff}, + .vact_space_4 = {0xff, 0xff}, + .vact_space_5 = {0xff, 0xff}, + .vact_space_6 = {0xff, 0xff}, + /* other don't care */ + }, + .tg = { + 0x00, /* cmd */ + 0xbc, 0x07, /* h_fsz */ + 0xbc, 0x02, 0x00, 0x05, /* hact */ + 0xee, 0x02, /* v_fsz */ + 0x01, 0x00, 0x33, 0x02, /* vsync */ + 0x1e, 0x00, 0xd0, 0x02, /* vact */ + 0x33, 0x02, /* field_chg */ + 0x48, 0x02, /* vact_st2 */ + 0x00, 0x00, /* vact_st3 */ + 0x00, 0x00, /* vact_st4 */ + 0x01, 0x00, 0x01, 0x00, /* vsync top/bot */ + 0x01, 0x00, 0x33, 0x02, /* field top/bot */ + 0x00, /* 3d FP */ + }, +}; + +static const struct hdmi_preset_conf hdmi_conf_720p60 = { + .core = { + .h_blank = {0x72, 0x01}, + .v2_blank = {0xee, 0x02}, + .v1_blank = {0x1e, 0x00}, + .v_line = {0xee, 0x02}, + .h_line = {0x72, 0x06}, + .hsync_pol = {0x00}, + .vsync_pol = {0x00}, + .int_pro_mode = {0x00}, + .v_blank_f0 = {0xff, 0xff}, + .v_blank_f1 = {0xff, 0xff}, + .h_sync_start = {0x6c, 0x00}, + .h_sync_end = {0x94, 0x00}, + .v_sync_line_bef_2 = {0x0a, 0x00}, + .v_sync_line_bef_1 = {0x05, 0x00}, + .v_sync_line_aft_2 = {0xff, 0xff}, + .v_sync_line_aft_1 = {0xff, 0xff}, + .v_sync_line_aft_pxl_2 = {0xff, 0xff}, + .v_sync_line_aft_pxl_1 = {0xff, 0xff}, + .v_blank_f2 = {0xff, 0xff}, + .v_blank_f3 = {0xff, 0xff}, + .v_blank_f4 = {0xff, 0xff}, + .v_blank_f5 = {0xff, 0xff}, + .v_sync_line_aft_3 = {0xff, 0xff}, + .v_sync_line_aft_4 = {0xff, 0xff}, + .v_sync_line_aft_5 = {0xff, 0xff}, + .v_sync_line_aft_6 = {0xff, 0xff}, + .v_sync_line_aft_pxl_3 = {0xff, 0xff}, + .v_sync_line_aft_pxl_4 = {0xff, 0xff}, + .v_sync_line_aft_pxl_5 = {0xff, 0xff}, + .v_sync_line_aft_pxl_6 = {0xff, 0xff}, + .vact_space_1 = {0xff, 0xff}, + .vact_space_2 = {0xff, 0xff}, + .vact_space_3 = {0xff, 0xff}, + .vact_space_4 = {0xff, 0xff}, + .vact_space_5 = {0xff, 0xff}, + .vact_space_6 = {0xff, 0xff}, + /* other don't care */ + }, + .tg = { + 0x00, /* cmd */ + 0x72, 0x06, /* h_fsz */ + 0x72, 0x01, 0x00, 0x05, /* hact */ + 0xee, 0x02, /* v_fsz */ + 0x01, 0x00, 0x33, 0x02, /* vsync */ + 0x1e, 0x00, 0xd0, 0x02, /* vact */ + 0x33, 0x02, /* field_chg */ + 0x48, 0x02, /* vact_st2 */ + 0x00, 0x00, /* vact_st3 */ + 0x00, 0x00, /* vact_st4 */ + 0x01, 0x00, 0x01, 0x00, /* vsync top/bot */ + 0x01, 0x00, 0x33, 0x02, /* field top/bot */ + 0x00, /* 3d FP */ + }, +}; + +static const struct hdmi_preset_conf hdmi_conf_1080i50 = { + .core = { + .h_blank = {0xd0, 0x02}, + .v2_blank = {0x32, 0x02}, + .v1_blank = {0x16, 0x00}, + .v_line = {0x65, 0x04}, + .h_line = {0x50, 0x0a}, + .hsync_pol = {0x00}, + .vsync_pol = {0x00}, + .int_pro_mode = {0x01}, + .v_blank_f0 = {0x49, 0x02}, + .v_blank_f1 = {0x65, 0x04}, + .h_sync_start = {0x0e, 0x02}, + .h_sync_end = {0x3a, 0x02}, + .v_sync_line_bef_2 = {0x07, 0x00}, + .v_sync_line_bef_1 = {0x02, 0x00}, + .v_sync_line_aft_2 = {0x39, 0x02}, + .v_sync_line_aft_1 = {0x34, 0x02}, + .v_sync_line_aft_pxl_2 = {0x38, 0x07}, + .v_sync_line_aft_pxl_1 = {0x38, 0x07}, + .v_blank_f2 = {0xff, 0xff}, + .v_blank_f3 = {0xff, 0xff}, + .v_blank_f4 = {0xff, 0xff}, + .v_blank_f5 = {0xff, 0xff}, + .v_sync_line_aft_3 = {0xff, 0xff}, + .v_sync_line_aft_4 = {0xff, 0xff}, + .v_sync_line_aft_5 = {0xff, 0xff}, + .v_sync_line_aft_6 = {0xff, 0xff}, + .v_sync_line_aft_pxl_3 = {0xff, 0xff}, + .v_sync_line_aft_pxl_4 = {0xff, 0xff}, + .v_sync_line_aft_pxl_5 = {0xff, 0xff}, + .v_sync_line_aft_pxl_6 = {0xff, 0xff}, + .vact_space_1 = {0xff, 0xff}, + .vact_space_2 = {0xff, 0xff}, + .vact_space_3 = {0xff, 0xff}, + .vact_space_4 = {0xff, 0xff}, + .vact_space_5 = {0xff, 0xff}, + .vact_space_6 = {0xff, 0xff}, + /* other don't care */ + }, + .tg = { + 0x00, /* cmd */ + 0x50, 0x0a, /* h_fsz */ + 0xd0, 0x02, 0x80, 0x07, /* hact */ + 0x65, 0x04, /* v_fsz */ + 0x01, 0x00, 0x33, 0x02, /* vsync */ + 0x16, 0x00, 0x1c, 0x02, /* vact */ + 0x33, 0x02, /* field_chg */ + 0x49, 0x02, /* vact_st2 */ + 0x00, 0x00, /* vact_st3 */ + 0x00, 0x00, /* vact_st4 */ + 0x01, 0x00, 0x33, 0x02, /* vsync top/bot */ + 0x01, 0x00, 0x33, 0x02, /* field top/bot */ + 0x00, /* 3d FP */ + }, +}; + +static const struct hdmi_preset_conf hdmi_conf_1080i60 = { + .core = { + .h_blank = {0x18, 0x01}, + .v2_blank = {0x32, 0x02}, + .v1_blank = {0x16, 0x00}, + .v_line = {0x65, 0x04}, + .h_line = {0x98, 0x08}, + .hsync_pol = {0x00}, + .vsync_pol = {0x00}, + .int_pro_mode = {0x01}, + .v_blank_f0 = {0x49, 0x02}, + .v_blank_f1 = {0x65, 0x04}, + .h_sync_start = {0x56, 0x00}, + .h_sync_end = {0x82, 0x00}, + .v_sync_line_bef_2 = {0x07, 0x00}, + .v_sync_line_bef_1 = {0x02, 0x00}, + .v_sync_line_aft_2 = {0x39, 0x02}, + .v_sync_line_aft_1 = {0x34, 0x02}, + .v_sync_line_aft_pxl_2 = {0xa4, 0x04}, + .v_sync_line_aft_pxl_1 = {0xa4, 0x04}, + .v_blank_f2 = {0xff, 0xff}, + .v_blank_f3 = {0xff, 0xff}, + .v_blank_f4 = {0xff, 0xff}, + .v_blank_f5 = {0xff, 0xff}, + .v_sync_line_aft_3 = {0xff, 0xff}, + .v_sync_line_aft_4 = {0xff, 0xff}, + .v_sync_line_aft_5 = {0xff, 0xff}, + .v_sync_line_aft_6 = {0xff, 0xff}, + .v_sync_line_aft_pxl_3 = {0xff, 0xff}, + .v_sync_line_aft_pxl_4 = {0xff, 0xff}, + .v_sync_line_aft_pxl_5 = {0xff, 0xff}, + .v_sync_line_aft_pxl_6 = {0xff, 0xff}, + .vact_space_1 = {0xff, 0xff}, + .vact_space_2 = {0xff, 0xff}, + .vact_space_3 = {0xff, 0xff}, + .vact_space_4 = {0xff, 0xff}, + .vact_space_5 = {0xff, 0xff}, + .vact_space_6 = {0xff, 0xff}, + /* other don't care */ + }, + .tg = { + 0x00, /* cmd */ + 0x98, 0x08, /* h_fsz */ + 0x18, 0x01, 0x80, 0x07, /* hact */ + 0x65, 0x04, /* v_fsz */ + 0x01, 0x00, 0x33, 0x02, /* vsync */ + 0x16, 0x00, 0x1c, 0x02, /* vact */ + 0x33, 0x02, /* field_chg */ + 0x49, 0x02, /* vact_st2 */ + 0x00, 0x00, /* vact_st3 */ + 0x00, 0x00, /* vact_st4 */ + 0x01, 0x00, 0x33, 0x02, /* vsync top/bot */ + 0x01, 0x00, 0x33, 0x02, /* field top/bot */ + 0x00, /* 3d FP */ + }, +}; + +static const struct hdmi_preset_conf hdmi_conf_1080p50 = { + .core = { + .h_blank = {0xd0, 0x02}, + .v2_blank = {0x65, 0x04}, + .v1_blank = {0x2d, 0x00}, + .v_line = {0x65, 0x04}, + .h_line = {0x50, 0x0a}, + .hsync_pol = {0x00}, + .vsync_pol = {0x00}, + .int_pro_mode = {0x00}, + .v_blank_f0 = {0xff, 0xff}, + .v_blank_f1 = {0xff, 0xff}, + .h_sync_start = {0x0e, 0x02}, + .h_sync_end = {0x3a, 0x02}, + .v_sync_line_bef_2 = {0x09, 0x00}, + .v_sync_line_bef_1 = {0x04, 0x00}, + .v_sync_line_aft_2 = {0xff, 0xff}, + .v_sync_line_aft_1 = {0xff, 0xff}, + .v_sync_line_aft_pxl_2 = {0xff, 0xff}, + .v_sync_line_aft_pxl_1 = {0xff, 0xff}, + .v_blank_f2 = {0xff, 0xff}, + .v_blank_f3 = {0xff, 0xff}, + .v_blank_f4 = {0xff, 0xff}, + .v_blank_f5 = {0xff, 0xff}, + .v_sync_line_aft_3 = {0xff, 0xff}, + .v_sync_line_aft_4 = {0xff, 0xff}, + .v_sync_line_aft_5 = {0xff, 0xff}, + .v_sync_line_aft_6 = {0xff, 0xff}, + .v_sync_line_aft_pxl_3 = {0xff, 0xff}, + .v_sync_line_aft_pxl_4 = {0xff, 0xff}, + .v_sync_line_aft_pxl_5 = {0xff, 0xff}, + .v_sync_line_aft_pxl_6 = {0xff, 0xff}, + .vact_space_1 = {0xff, 0xff}, + .vact_space_2 = {0xff, 0xff}, + .vact_space_3 = {0xff, 0xff}, + .vact_space_4 = {0xff, 0xff}, + .vact_space_5 = {0xff, 0xff}, + .vact_space_6 = {0xff, 0xff}, + /* other don't care */ + }, + .tg = { + 0x00, /* cmd */ + 0x50, 0x0a, /* h_fsz */ + 0xd0, 0x02, 0x80, 0x07, /* hact */ + 0x65, 0x04, /* v_fsz */ + 0x01, 0x00, 0x33, 0x02, /* vsync */ + 0x2d, 0x00, 0x38, 0x04, /* vact */ + 0x33, 0x02, /* field_chg */ + 0x48, 0x02, /* vact_st2 */ + 0x00, 0x00, /* vact_st3 */ + 0x00, 0x00, /* vact_st4 */ + 0x01, 0x00, 0x01, 0x00, /* vsync top/bot */ + 0x01, 0x00, 0x33, 0x02, /* field top/bot */ + 0x00, /* 3d FP */ + }, +}; + +static const struct hdmi_preset_conf hdmi_conf_1080p60 = { + .core = { + .h_blank = {0x18, 0x01}, + .v2_blank = {0x65, 0x04}, + .v1_blank = {0x2d, 0x00}, + .v_line = {0x65, 0x04}, + .h_line = {0x98, 0x08}, + .hsync_pol = {0x00}, + .vsync_pol = {0x00}, + .int_pro_mode = {0x00}, + .v_blank_f0 = {0xff, 0xff}, + .v_blank_f1 = {0xff, 0xff}, + .h_sync_start = {0x56, 0x00}, + .h_sync_end = {0x82, 0x00}, + .v_sync_line_bef_2 = {0x09, 0x00}, + .v_sync_line_bef_1 = {0x04, 0x00}, + .v_sync_line_aft_2 = {0xff, 0xff}, + .v_sync_line_aft_1 = {0xff, 0xff}, + .v_sync_line_aft_pxl_2 = {0xff, 0xff}, + .v_sync_line_aft_pxl_1 = {0xff, 0xff}, + .v_blank_f2 = {0xff, 0xff}, + .v_blank_f3 = {0xff, 0xff}, + .v_blank_f4 = {0xff, 0xff}, + .v_blank_f5 = {0xff, 0xff}, + .v_sync_line_aft_3 = {0xff, 0xff}, + .v_sync_line_aft_4 = {0xff, 0xff}, + .v_sync_line_aft_5 = {0xff, 0xff}, + .v_sync_line_aft_6 = {0xff, 0xff}, + .v_sync_line_aft_pxl_3 = {0xff, 0xff}, + .v_sync_line_aft_pxl_4 = {0xff, 0xff}, + .v_sync_line_aft_pxl_5 = {0xff, 0xff}, + .v_sync_line_aft_pxl_6 = {0xff, 0xff}, + /* other don't care */ + }, + .tg = { + 0x00, /* cmd */ + 0x98, 0x08, /* h_fsz */ + 0x18, 0x01, 0x80, 0x07, /* hact */ + 0x65, 0x04, /* v_fsz */ + 0x01, 0x00, 0x33, 0x02, /* vsync */ + 0x2d, 0x00, 0x38, 0x04, /* vact */ + 0x33, 0x02, /* field_chg */ + 0x48, 0x02, /* vact_st2 */ + 0x00, 0x00, /* vact_st3 */ + 0x00, 0x00, /* vact_st4 */ + 0x01, 0x00, 0x01, 0x00, /* vsync top/bot */ + 0x01, 0x00, 0x33, 0x02, /* field top/bot */ + 0x00, /* 3d FP */ + }, +}; + +static const struct hdmi_conf hdmi_confs[] = { + { 720, 480, 60, false, hdmiphy_conf27_027, &hdmi_conf_480p60 }, + { 1280, 720, 50, false, hdmiphy_conf74_25, &hdmi_conf_720p50 }, + { 1280, 720, 60, false, hdmiphy_conf74_25, &hdmi_conf_720p60 }, + { 1920, 1080, 50, true, hdmiphy_conf74_25, &hdmi_conf_1080i50 }, + { 1920, 1080, 60, true, hdmiphy_conf74_25, &hdmi_conf_1080i60 }, + { 1920, 1080, 50, false, hdmiphy_conf148_5, &hdmi_conf_1080p50 }, + { 1920, 1080, 60, false, hdmiphy_conf148_5, &hdmi_conf_1080p60 }, +}; + + +static inline u32 hdmi_reg_read(struct hdmi_context *hdata, u32 reg_id) +{ + return readl(hdata->regs + reg_id); +} + +static inline void hdmi_reg_writeb(struct hdmi_context *hdata, + u32 reg_id, u8 value) +{ + writeb(value, hdata->regs + reg_id); +} + +static inline void hdmi_reg_writemask(struct hdmi_context *hdata, + u32 reg_id, u32 value, u32 mask) +{ + u32 old = readl(hdata->regs + reg_id); + value = (value & mask) | (old & ~mask); + writel(value, hdata->regs + reg_id); +} + +static void hdmi_v13_regs_dump(struct hdmi_context *hdata, char *prefix) +{ +#define DUMPREG(reg_id) \ + DRM_DEBUG_KMS("%s:" #reg_id " = %08x\n", prefix, \ + readl(hdata->regs + reg_id)) + DRM_DEBUG_KMS("%s: ---- CONTROL REGISTERS ----\n", prefix); + DUMPREG(HDMI_INTC_FLAG); + DUMPREG(HDMI_INTC_CON); + DUMPREG(HDMI_HPD_STATUS); + DUMPREG(HDMI_V13_PHY_RSTOUT); + DUMPREG(HDMI_V13_PHY_VPLL); + DUMPREG(HDMI_V13_PHY_CMU); + DUMPREG(HDMI_V13_CORE_RSTOUT); + + DRM_DEBUG_KMS("%s: ---- CORE REGISTERS ----\n", prefix); + DUMPREG(HDMI_CON_0); + DUMPREG(HDMI_CON_1); + DUMPREG(HDMI_CON_2); + DUMPREG(HDMI_SYS_STATUS); + DUMPREG(HDMI_V13_PHY_STATUS); + DUMPREG(HDMI_STATUS_EN); + DUMPREG(HDMI_HPD); + DUMPREG(HDMI_MODE_SEL); + DUMPREG(HDMI_V13_HPD_GEN); + DUMPREG(HDMI_V13_DC_CONTROL); + DUMPREG(HDMI_V13_VIDEO_PATTERN_GEN); + + DRM_DEBUG_KMS("%s: ---- CORE SYNC REGISTERS ----\n", prefix); + DUMPREG(HDMI_H_BLANK_0); + DUMPREG(HDMI_H_BLANK_1); + DUMPREG(HDMI_V13_V_BLANK_0); + DUMPREG(HDMI_V13_V_BLANK_1); + DUMPREG(HDMI_V13_V_BLANK_2); + DUMPREG(HDMI_V13_H_V_LINE_0); + DUMPREG(HDMI_V13_H_V_LINE_1); + DUMPREG(HDMI_V13_H_V_LINE_2); + DUMPREG(HDMI_VSYNC_POL); + DUMPREG(HDMI_INT_PRO_MODE); + DUMPREG(HDMI_V13_V_BLANK_F_0); + DUMPREG(HDMI_V13_V_BLANK_F_1); + DUMPREG(HDMI_V13_V_BLANK_F_2); + DUMPREG(HDMI_V13_H_SYNC_GEN_0); + DUMPREG(HDMI_V13_H_SYNC_GEN_1); + DUMPREG(HDMI_V13_H_SYNC_GEN_2); + DUMPREG(HDMI_V13_V_SYNC_GEN_1_0); + DUMPREG(HDMI_V13_V_SYNC_GEN_1_1); + DUMPREG(HDMI_V13_V_SYNC_GEN_1_2); + DUMPREG(HDMI_V13_V_SYNC_GEN_2_0); + DUMPREG(HDMI_V13_V_SYNC_GEN_2_1); + DUMPREG(HDMI_V13_V_SYNC_GEN_2_2); + DUMPREG(HDMI_V13_V_SYNC_GEN_3_0); + DUMPREG(HDMI_V13_V_SYNC_GEN_3_1); + DUMPREG(HDMI_V13_V_SYNC_GEN_3_2); + + DRM_DEBUG_KMS("%s: ---- TG REGISTERS ----\n", prefix); + DUMPREG(HDMI_TG_CMD); + DUMPREG(HDMI_TG_H_FSZ_L); + DUMPREG(HDMI_TG_H_FSZ_H); + DUMPREG(HDMI_TG_HACT_ST_L); + DUMPREG(HDMI_TG_HACT_ST_H); + DUMPREG(HDMI_TG_HACT_SZ_L); + DUMPREG(HDMI_TG_HACT_SZ_H); + DUMPREG(HDMI_TG_V_FSZ_L); + DUMPREG(HDMI_TG_V_FSZ_H); + DUMPREG(HDMI_TG_VSYNC_L); + DUMPREG(HDMI_TG_VSYNC_H); + DUMPREG(HDMI_TG_VSYNC2_L); + DUMPREG(HDMI_TG_VSYNC2_H); + DUMPREG(HDMI_TG_VACT_ST_L); + DUMPREG(HDMI_TG_VACT_ST_H); + DUMPREG(HDMI_TG_VACT_SZ_L); + DUMPREG(HDMI_TG_VACT_SZ_H); + DUMPREG(HDMI_TG_FIELD_CHG_L); + DUMPREG(HDMI_TG_FIELD_CHG_H); + DUMPREG(HDMI_TG_VACT_ST2_L); + DUMPREG(HDMI_TG_VACT_ST2_H); + DUMPREG(HDMI_TG_VSYNC_TOP_HDMI_L); + DUMPREG(HDMI_TG_VSYNC_TOP_HDMI_H); + DUMPREG(HDMI_TG_VSYNC_BOT_HDMI_L); + DUMPREG(HDMI_TG_VSYNC_BOT_HDMI_H); + DUMPREG(HDMI_TG_FIELD_TOP_HDMI_L); + DUMPREG(HDMI_TG_FIELD_TOP_HDMI_H); + DUMPREG(HDMI_TG_FIELD_BOT_HDMI_L); + DUMPREG(HDMI_TG_FIELD_BOT_HDMI_H); +#undef DUMPREG +} + +static void hdmi_v14_regs_dump(struct hdmi_context *hdata, char *prefix) +{ + int i; + +#define DUMPREG(reg_id) \ + DRM_DEBUG_KMS("%s:" #reg_id " = %08x\n", prefix, \ + readl(hdata->regs + reg_id)) + + DRM_DEBUG_KMS("%s: ---- CONTROL REGISTERS ----\n", prefix); + DUMPREG(HDMI_INTC_CON); + DUMPREG(HDMI_INTC_FLAG); + DUMPREG(HDMI_HPD_STATUS); + DUMPREG(HDMI_INTC_CON_1); + DUMPREG(HDMI_INTC_FLAG_1); + DUMPREG(HDMI_PHY_STATUS_0); + DUMPREG(HDMI_PHY_STATUS_PLL); + DUMPREG(HDMI_PHY_CON_0); + DUMPREG(HDMI_PHY_RSTOUT); + DUMPREG(HDMI_PHY_VPLL); + DUMPREG(HDMI_PHY_CMU); + DUMPREG(HDMI_CORE_RSTOUT); + + DRM_DEBUG_KMS("%s: ---- CORE REGISTERS ----\n", prefix); + DUMPREG(HDMI_CON_0); + DUMPREG(HDMI_CON_1); + DUMPREG(HDMI_CON_2); + DUMPREG(HDMI_SYS_STATUS); + DUMPREG(HDMI_PHY_STATUS_0); + DUMPREG(HDMI_STATUS_EN); + DUMPREG(HDMI_HPD); + DUMPREG(HDMI_MODE_SEL); + DUMPREG(HDMI_ENC_EN); + DUMPREG(HDMI_DC_CONTROL); + DUMPREG(HDMI_VIDEO_PATTERN_GEN); + + DRM_DEBUG_KMS("%s: ---- CORE SYNC REGISTERS ----\n", prefix); + DUMPREG(HDMI_H_BLANK_0); + DUMPREG(HDMI_H_BLANK_1); + DUMPREG(HDMI_V2_BLANK_0); + DUMPREG(HDMI_V2_BLANK_1); + DUMPREG(HDMI_V1_BLANK_0); + DUMPREG(HDMI_V1_BLANK_1); + DUMPREG(HDMI_V_LINE_0); + DUMPREG(HDMI_V_LINE_1); + DUMPREG(HDMI_H_LINE_0); + DUMPREG(HDMI_H_LINE_1); + DUMPREG(HDMI_HSYNC_POL); + + DUMPREG(HDMI_VSYNC_POL); + DUMPREG(HDMI_INT_PRO_MODE); + DUMPREG(HDMI_V_BLANK_F0_0); + DUMPREG(HDMI_V_BLANK_F0_1); + DUMPREG(HDMI_V_BLANK_F1_0); + DUMPREG(HDMI_V_BLANK_F1_1); + + DUMPREG(HDMI_H_SYNC_START_0); + DUMPREG(HDMI_H_SYNC_START_1); + DUMPREG(HDMI_H_SYNC_END_0); + DUMPREG(HDMI_H_SYNC_END_1); + + DUMPREG(HDMI_V_SYNC_LINE_BEF_2_0); + DUMPREG(HDMI_V_SYNC_LINE_BEF_2_1); + DUMPREG(HDMI_V_SYNC_LINE_BEF_1_0); + DUMPREG(HDMI_V_SYNC_LINE_BEF_1_1); + + DUMPREG(HDMI_V_SYNC_LINE_AFT_2_0); + DUMPREG(HDMI_V_SYNC_LINE_AFT_2_1); + DUMPREG(HDMI_V_SYNC_LINE_AFT_1_0); + DUMPREG(HDMI_V_SYNC_LINE_AFT_1_1); + + DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_2_0); + DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_2_1); + DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_1_0); + DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_1_1); + + DUMPREG(HDMI_V_BLANK_F2_0); + DUMPREG(HDMI_V_BLANK_F2_1); + DUMPREG(HDMI_V_BLANK_F3_0); + DUMPREG(HDMI_V_BLANK_F3_1); + DUMPREG(HDMI_V_BLANK_F4_0); + DUMPREG(HDMI_V_BLANK_F4_1); + DUMPREG(HDMI_V_BLANK_F5_0); + DUMPREG(HDMI_V_BLANK_F5_1); + + DUMPREG(HDMI_V_SYNC_LINE_AFT_3_0); + DUMPREG(HDMI_V_SYNC_LINE_AFT_3_1); + DUMPREG(HDMI_V_SYNC_LINE_AFT_4_0); + DUMPREG(HDMI_V_SYNC_LINE_AFT_4_1); + DUMPREG(HDMI_V_SYNC_LINE_AFT_5_0); + DUMPREG(HDMI_V_SYNC_LINE_AFT_5_1); + DUMPREG(HDMI_V_SYNC_LINE_AFT_6_0); + DUMPREG(HDMI_V_SYNC_LINE_AFT_6_1); + + DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_3_0); + DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_3_1); + DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_4_0); + DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_4_1); + DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_5_0); + DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_5_1); + DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_6_0); + DUMPREG(HDMI_V_SYNC_LINE_AFT_PXL_6_1); + + DUMPREG(HDMI_VACT_SPACE_1_0); + DUMPREG(HDMI_VACT_SPACE_1_1); + DUMPREG(HDMI_VACT_SPACE_2_0); + DUMPREG(HDMI_VACT_SPACE_2_1); + DUMPREG(HDMI_VACT_SPACE_3_0); + DUMPREG(HDMI_VACT_SPACE_3_1); + DUMPREG(HDMI_VACT_SPACE_4_0); + DUMPREG(HDMI_VACT_SPACE_4_1); + DUMPREG(HDMI_VACT_SPACE_5_0); + DUMPREG(HDMI_VACT_SPACE_5_1); + DUMPREG(HDMI_VACT_SPACE_6_0); + DUMPREG(HDMI_VACT_SPACE_6_1); + + DRM_DEBUG_KMS("%s: ---- TG REGISTERS ----\n", prefix); + DUMPREG(HDMI_TG_CMD); + DUMPREG(HDMI_TG_H_FSZ_L); + DUMPREG(HDMI_TG_H_FSZ_H); + DUMPREG(HDMI_TG_HACT_ST_L); + DUMPREG(HDMI_TG_HACT_ST_H); + DUMPREG(HDMI_TG_HACT_SZ_L); + DUMPREG(HDMI_TG_HACT_SZ_H); + DUMPREG(HDMI_TG_V_FSZ_L); + DUMPREG(HDMI_TG_V_FSZ_H); + DUMPREG(HDMI_TG_VSYNC_L); + DUMPREG(HDMI_TG_VSYNC_H); + DUMPREG(HDMI_TG_VSYNC2_L); + DUMPREG(HDMI_TG_VSYNC2_H); + DUMPREG(HDMI_TG_VACT_ST_L); + DUMPREG(HDMI_TG_VACT_ST_H); + DUMPREG(HDMI_TG_VACT_SZ_L); + DUMPREG(HDMI_TG_VACT_SZ_H); + DUMPREG(HDMI_TG_FIELD_CHG_L); + DUMPREG(HDMI_TG_FIELD_CHG_H); + DUMPREG(HDMI_TG_VACT_ST2_L); + DUMPREG(HDMI_TG_VACT_ST2_H); + DUMPREG(HDMI_TG_VACT_ST3_L); + DUMPREG(HDMI_TG_VACT_ST3_H); + DUMPREG(HDMI_TG_VACT_ST4_L); + DUMPREG(HDMI_TG_VACT_ST4_H); + DUMPREG(HDMI_TG_VSYNC_TOP_HDMI_L); + DUMPREG(HDMI_TG_VSYNC_TOP_HDMI_H); + DUMPREG(HDMI_TG_VSYNC_BOT_HDMI_L); + DUMPREG(HDMI_TG_VSYNC_BOT_HDMI_H); + DUMPREG(HDMI_TG_FIELD_TOP_HDMI_L); + DUMPREG(HDMI_TG_FIELD_TOP_HDMI_H); + DUMPREG(HDMI_TG_FIELD_BOT_HDMI_L); + DUMPREG(HDMI_TG_FIELD_BOT_HDMI_H); + DUMPREG(HDMI_TG_3D); + + DRM_DEBUG_KMS("%s: ---- PACKET REGISTERS ----\n", prefix); + DUMPREG(HDMI_AVI_CON); + DUMPREG(HDMI_AVI_HEADER0); + DUMPREG(HDMI_AVI_HEADER1); + DUMPREG(HDMI_AVI_HEADER2); + DUMPREG(HDMI_AVI_CHECK_SUM); + DUMPREG(HDMI_VSI_CON); + DUMPREG(HDMI_VSI_HEADER0); + DUMPREG(HDMI_VSI_HEADER1); + DUMPREG(HDMI_VSI_HEADER2); + for (i = 0; i < 7; ++i) + DUMPREG(HDMI_VSI_DATA(i)); + +#undef DUMPREG +} + +static void hdmi_regs_dump(struct hdmi_context *hdata, char *prefix) +{ + if (hdata->is_v13) + hdmi_v13_regs_dump(hdata, prefix); + else + hdmi_v14_regs_dump(hdata, prefix); +} + +static int hdmi_v13_conf_index(struct drm_display_mode *mode) +{ + int i; + + for (i = 0; i < ARRAY_SIZE(hdmi_v13_confs); ++i) + if (hdmi_v13_confs[i].width == mode->hdisplay && + hdmi_v13_confs[i].height == mode->vdisplay && + hdmi_v13_confs[i].vrefresh == mode->vrefresh && + hdmi_v13_confs[i].interlace == + ((mode->flags & DRM_MODE_FLAG_INTERLACE) ? + true : false)) + return i; + + return -EINVAL; +} + +static int hdmi_v14_conf_index(struct drm_display_mode *mode) +{ + int i; + + for (i = 0; i < ARRAY_SIZE(hdmi_confs); ++i) + if (hdmi_confs[i].width == mode->hdisplay && + hdmi_confs[i].height == mode->vdisplay && + hdmi_confs[i].vrefresh == mode->vrefresh && + hdmi_confs[i].interlace == + ((mode->flags & DRM_MODE_FLAG_INTERLACE) ? + true : false)) + return i; + + return -EINVAL; +} + +static int hdmi_conf_index(struct hdmi_context *hdata, + struct drm_display_mode *mode) +{ + if (hdata->is_v13) + return hdmi_v13_conf_index(mode); + + return hdmi_v14_conf_index(mode); +} + +static bool hdmi_is_connected(void *ctx) +{ + struct hdmi_context *hdata = ctx; + u32 val = hdmi_reg_read(hdata, HDMI_HPD_STATUS); + + if (val) + return true; + + return false; +} + +static int hdmi_get_edid(void *ctx, struct drm_connector *connector, + u8 *edid, int len) +{ + struct edid *raw_edid; + struct hdmi_context *hdata = ctx; + + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + if (!hdata->ddc_port) + return -ENODEV; + + raw_edid = drm_get_edid(connector, hdata->ddc_port->adapter); + if (raw_edid) { + memcpy(edid, raw_edid, min((1 + raw_edid->extensions) + * EDID_LENGTH, len)); + DRM_DEBUG_KMS("width[%d] x height[%d]\n", + raw_edid->width_cm, raw_edid->height_cm); + } else { + return -ENODEV; + } + + return 0; +} + +static int hdmi_v13_check_timing(struct fb_videomode *check_timing) +{ + int i; + + DRM_DEBUG_KMS("valid mode : xres=%d, yres=%d, refresh=%d, intl=%d\n", + check_timing->xres, check_timing->yres, + check_timing->refresh, (check_timing->vmode & + FB_VMODE_INTERLACED) ? true : false); + + for (i = 0; i < ARRAY_SIZE(hdmi_v13_confs); ++i) + if (hdmi_v13_confs[i].width == check_timing->xres && + hdmi_v13_confs[i].height == check_timing->yres && + hdmi_v13_confs[i].vrefresh == check_timing->refresh && + hdmi_v13_confs[i].interlace == + ((check_timing->vmode & FB_VMODE_INTERLACED) ? + true : false)) + return 0; + + /* TODO */ + + return -EINVAL; +} + +static int hdmi_v14_check_timing(struct fb_videomode *check_timing) +{ + int i; + + DRM_DEBUG_KMS("valid mode : xres=%d, yres=%d, refresh=%d, intl=%d\n", + check_timing->xres, check_timing->yres, + check_timing->refresh, (check_timing->vmode & + FB_VMODE_INTERLACED) ? true : false); + + for (i = 0; i < ARRAY_SIZE(hdmi_confs); i++) + if (hdmi_confs[i].width == check_timing->xres && + hdmi_confs[i].height == check_timing->yres && + hdmi_confs[i].vrefresh == check_timing->refresh && + hdmi_confs[i].interlace == + ((check_timing->vmode & FB_VMODE_INTERLACED) ? + true : false)) + return 0; + + /* TODO */ + + return -EINVAL; +} + +static int hdmi_check_timing(void *ctx, void *timing) +{ + struct hdmi_context *hdata = ctx; + struct fb_videomode *check_timing = timing; + + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + DRM_DEBUG_KMS("[%d]x[%d] [%d]Hz [%x]\n", check_timing->xres, + check_timing->yres, check_timing->refresh, + check_timing->vmode); + + if (hdata->is_v13) + return hdmi_v13_check_timing(check_timing); + else + return hdmi_v14_check_timing(check_timing); +} + +static int hdmi_display_power_on(void *ctx, int mode) +{ + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + switch (mode) { + case DRM_MODE_DPMS_ON: + DRM_DEBUG_KMS("hdmi [on]\n"); + break; + case DRM_MODE_DPMS_STANDBY: + break; + case DRM_MODE_DPMS_SUSPEND: + break; + case DRM_MODE_DPMS_OFF: + DRM_DEBUG_KMS("hdmi [off]\n"); + break; + default: + break; + } + + return 0; +} + +static void hdmi_set_acr(u32 freq, u8 *acr) +{ + u32 n, cts; + + switch (freq) { + case 32000: + n = 4096; + cts = 27000; + break; + case 44100: + n = 6272; + cts = 30000; + break; + case 88200: + n = 12544; + cts = 30000; + break; + case 176400: + n = 25088; + cts = 30000; + break; + case 48000: + n = 6144; + cts = 27000; + break; + case 96000: + n = 12288; + cts = 27000; + break; + case 192000: + n = 24576; + cts = 27000; + break; + default: + n = 0; + cts = 0; + break; + } + + acr[1] = cts >> 16; + acr[2] = cts >> 8 & 0xff; + acr[3] = cts & 0xff; + + acr[4] = n >> 16; + acr[5] = n >> 8 & 0xff; + acr[6] = n & 0xff; +} + +static void hdmi_reg_acr(struct hdmi_context *hdata, u8 *acr) +{ + hdmi_reg_writeb(hdata, HDMI_ACR_N0, acr[6]); + hdmi_reg_writeb(hdata, HDMI_ACR_N1, acr[5]); + hdmi_reg_writeb(hdata, HDMI_ACR_N2, acr[4]); + hdmi_reg_writeb(hdata, HDMI_ACR_MCTS0, acr[3]); + hdmi_reg_writeb(hdata, HDMI_ACR_MCTS1, acr[2]); + hdmi_reg_writeb(hdata, HDMI_ACR_MCTS2, acr[1]); + hdmi_reg_writeb(hdata, HDMI_ACR_CTS0, acr[3]); + hdmi_reg_writeb(hdata, HDMI_ACR_CTS1, acr[2]); + hdmi_reg_writeb(hdata, HDMI_ACR_CTS2, acr[1]); + + if (hdata->is_v13) + hdmi_reg_writeb(hdata, HDMI_V13_ACR_CON, 4); + else + hdmi_reg_writeb(hdata, HDMI_ACR_CON, 4); +} + +static void hdmi_audio_init(struct hdmi_context *hdata) +{ + u32 sample_rate, bits_per_sample, frame_size_code; + u32 data_num, bit_ch, sample_frq; + u32 val; + u8 acr[7]; + + sample_rate = 44100; + bits_per_sample = 16; + frame_size_code = 0; + + switch (bits_per_sample) { + case 20: + data_num = 2; + bit_ch = 1; + break; + case 24: + data_num = 3; + bit_ch = 1; + break; + default: + data_num = 1; + bit_ch = 0; + break; + } + + hdmi_set_acr(sample_rate, acr); + hdmi_reg_acr(hdata, acr); + + hdmi_reg_writeb(hdata, HDMI_I2S_MUX_CON, HDMI_I2S_IN_DISABLE + | HDMI_I2S_AUD_I2S | HDMI_I2S_CUV_I2S_ENABLE + | HDMI_I2S_MUX_ENABLE); + + hdmi_reg_writeb(hdata, HDMI_I2S_MUX_CH, HDMI_I2S_CH0_EN + | HDMI_I2S_CH1_EN | HDMI_I2S_CH2_EN); + + hdmi_reg_writeb(hdata, HDMI_I2S_MUX_CUV, HDMI_I2S_CUV_RL_EN); + + sample_frq = (sample_rate == 44100) ? 0 : + (sample_rate == 48000) ? 2 : + (sample_rate == 32000) ? 3 : + (sample_rate == 96000) ? 0xa : 0x0; + + hdmi_reg_writeb(hdata, HDMI_I2S_CLK_CON, HDMI_I2S_CLK_DIS); + hdmi_reg_writeb(hdata, HDMI_I2S_CLK_CON, HDMI_I2S_CLK_EN); + + val = hdmi_reg_read(hdata, HDMI_I2S_DSD_CON) | 0x01; + hdmi_reg_writeb(hdata, HDMI_I2S_DSD_CON, val); + + /* Configuration I2S input ports. Configure I2S_PIN_SEL_0~4 */ + hdmi_reg_writeb(hdata, HDMI_I2S_PIN_SEL_0, HDMI_I2S_SEL_SCLK(5) + | HDMI_I2S_SEL_LRCK(6)); + hdmi_reg_writeb(hdata, HDMI_I2S_PIN_SEL_1, HDMI_I2S_SEL_SDATA1(1) + | HDMI_I2S_SEL_SDATA2(4)); + hdmi_reg_writeb(hdata, HDMI_I2S_PIN_SEL_2, HDMI_I2S_SEL_SDATA3(1) + | HDMI_I2S_SEL_SDATA2(2)); + hdmi_reg_writeb(hdata, HDMI_I2S_PIN_SEL_3, HDMI_I2S_SEL_DSD(0)); + + /* I2S_CON_1 & 2 */ + hdmi_reg_writeb(hdata, HDMI_I2S_CON_1, HDMI_I2S_SCLK_FALLING_EDGE + | HDMI_I2S_L_CH_LOW_POL); + hdmi_reg_writeb(hdata, HDMI_I2S_CON_2, HDMI_I2S_MSB_FIRST_MODE + | HDMI_I2S_SET_BIT_CH(bit_ch) + | HDMI_I2S_SET_SDATA_BIT(data_num) + | HDMI_I2S_BASIC_FORMAT); + + /* Configure register related to CUV information */ + hdmi_reg_writeb(hdata, HDMI_I2S_CH_ST_0, HDMI_I2S_CH_STATUS_MODE_0 + | HDMI_I2S_2AUD_CH_WITHOUT_PREEMPH + | HDMI_I2S_COPYRIGHT + | HDMI_I2S_LINEAR_PCM + | HDMI_I2S_CONSUMER_FORMAT); + hdmi_reg_writeb(hdata, HDMI_I2S_CH_ST_1, HDMI_I2S_CD_PLAYER); + hdmi_reg_writeb(hdata, HDMI_I2S_CH_ST_2, HDMI_I2S_SET_SOURCE_NUM(0)); + hdmi_reg_writeb(hdata, HDMI_I2S_CH_ST_3, HDMI_I2S_CLK_ACCUR_LEVEL_2 + | HDMI_I2S_SET_SMP_FREQ(sample_frq)); + hdmi_reg_writeb(hdata, HDMI_I2S_CH_ST_4, + HDMI_I2S_ORG_SMP_FREQ_44_1 + | HDMI_I2S_WORD_LEN_MAX24_24BITS + | HDMI_I2S_WORD_LEN_MAX_24BITS); + + hdmi_reg_writeb(hdata, HDMI_I2S_CH_ST_CON, HDMI_I2S_CH_STATUS_RELOAD); +} + +static void hdmi_audio_control(struct hdmi_context *hdata, bool onoff) +{ + u32 mod; + + mod = hdmi_reg_read(hdata, HDMI_MODE_SEL); + if (mod & HDMI_DVI_MODE_EN) + return; + + hdmi_reg_writeb(hdata, HDMI_AUI_CON, onoff ? 2 : 0); + hdmi_reg_writemask(hdata, HDMI_CON_0, onoff ? + HDMI_ASP_EN : HDMI_ASP_DIS, HDMI_ASP_MASK); +} + +static void hdmi_conf_reset(struct hdmi_context *hdata) +{ + u32 reg; + + /* disable hpd handle for drm */ + hdata->hpd_handle = false; + + if (hdata->is_v13) + reg = HDMI_V13_CORE_RSTOUT; + else + reg = HDMI_CORE_RSTOUT; + + /* resetting HDMI core */ + hdmi_reg_writemask(hdata, reg, 0, HDMI_CORE_SW_RSTOUT); + mdelay(10); + hdmi_reg_writemask(hdata, reg, ~0, HDMI_CORE_SW_RSTOUT); + mdelay(10); + + /* enable hpd handle for drm */ + hdata->hpd_handle = true; +} + +static void hdmi_conf_init(struct hdmi_context *hdata) +{ + /* disable hpd handle for drm */ + hdata->hpd_handle = false; + + /* enable HPD interrupts */ + hdmi_reg_writemask(hdata, HDMI_INTC_CON, 0, HDMI_INTC_EN_GLOBAL | + HDMI_INTC_EN_HPD_PLUG | HDMI_INTC_EN_HPD_UNPLUG); + mdelay(10); + hdmi_reg_writemask(hdata, HDMI_INTC_CON, ~0, HDMI_INTC_EN_GLOBAL | + HDMI_INTC_EN_HPD_PLUG | HDMI_INTC_EN_HPD_UNPLUG); + + /* choose HDMI mode */ + hdmi_reg_writemask(hdata, HDMI_MODE_SEL, + HDMI_MODE_HDMI_EN, HDMI_MODE_MASK); + /* disable bluescreen */ + hdmi_reg_writemask(hdata, HDMI_CON_0, 0, HDMI_BLUE_SCR_EN); + + if (hdata->is_v13) { + /* choose bluescreen (fecal) color */ + hdmi_reg_writeb(hdata, HDMI_V13_BLUE_SCREEN_0, 0x12); + hdmi_reg_writeb(hdata, HDMI_V13_BLUE_SCREEN_1, 0x34); + hdmi_reg_writeb(hdata, HDMI_V13_BLUE_SCREEN_2, 0x56); + + /* enable AVI packet every vsync, fixes purple line problem */ + hdmi_reg_writeb(hdata, HDMI_V13_AVI_CON, 0x02); + /* force RGB, look to CEA-861-D, table 7 for more detail */ + hdmi_reg_writeb(hdata, HDMI_V13_AVI_BYTE(0), 0 << 5); + hdmi_reg_writemask(hdata, HDMI_CON_1, 0x10 << 5, 0x11 << 5); + + hdmi_reg_writeb(hdata, HDMI_V13_SPD_CON, 0x02); + hdmi_reg_writeb(hdata, HDMI_V13_AUI_CON, 0x02); + hdmi_reg_writeb(hdata, HDMI_V13_ACR_CON, 0x04); + } else { + /* enable AVI packet every vsync, fixes purple line problem */ + hdmi_reg_writeb(hdata, HDMI_AVI_CON, 0x02); + hdmi_reg_writeb(hdata, HDMI_AVI_BYTE(1), 2 << 5); + hdmi_reg_writemask(hdata, HDMI_CON_1, 2, 3 << 5); + } + + /* enable hpd handle for drm */ + hdata->hpd_handle = true; +} + +static void hdmi_v13_timing_apply(struct hdmi_context *hdata) +{ + const struct hdmi_v13_preset_conf *conf = + hdmi_v13_confs[hdata->cur_conf].conf; + const struct hdmi_v13_core_regs *core = &conf->core; + const struct hdmi_v13_tg_regs *tg = &conf->tg; + int tries; + + /* setting core registers */ + hdmi_reg_writeb(hdata, HDMI_H_BLANK_0, core->h_blank[0]); + hdmi_reg_writeb(hdata, HDMI_H_BLANK_1, core->h_blank[1]); + hdmi_reg_writeb(hdata, HDMI_V13_V_BLANK_0, core->v_blank[0]); + hdmi_reg_writeb(hdata, HDMI_V13_V_BLANK_1, core->v_blank[1]); + hdmi_reg_writeb(hdata, HDMI_V13_V_BLANK_2, core->v_blank[2]); + hdmi_reg_writeb(hdata, HDMI_V13_H_V_LINE_0, core->h_v_line[0]); + hdmi_reg_writeb(hdata, HDMI_V13_H_V_LINE_1, core->h_v_line[1]); + hdmi_reg_writeb(hdata, HDMI_V13_H_V_LINE_2, core->h_v_line[2]); + hdmi_reg_writeb(hdata, HDMI_VSYNC_POL, core->vsync_pol[0]); + hdmi_reg_writeb(hdata, HDMI_INT_PRO_MODE, core->int_pro_mode[0]); + hdmi_reg_writeb(hdata, HDMI_V13_V_BLANK_F_0, core->v_blank_f[0]); + hdmi_reg_writeb(hdata, HDMI_V13_V_BLANK_F_1, core->v_blank_f[1]); + hdmi_reg_writeb(hdata, HDMI_V13_V_BLANK_F_2, core->v_blank_f[2]); + hdmi_reg_writeb(hdata, HDMI_V13_H_SYNC_GEN_0, core->h_sync_gen[0]); + hdmi_reg_writeb(hdata, HDMI_V13_H_SYNC_GEN_1, core->h_sync_gen[1]); + hdmi_reg_writeb(hdata, HDMI_V13_H_SYNC_GEN_2, core->h_sync_gen[2]); + hdmi_reg_writeb(hdata, HDMI_V13_V_SYNC_GEN_1_0, core->v_sync_gen1[0]); + hdmi_reg_writeb(hdata, HDMI_V13_V_SYNC_GEN_1_1, core->v_sync_gen1[1]); + hdmi_reg_writeb(hdata, HDMI_V13_V_SYNC_GEN_1_2, core->v_sync_gen1[2]); + hdmi_reg_writeb(hdata, HDMI_V13_V_SYNC_GEN_2_0, core->v_sync_gen2[0]); + hdmi_reg_writeb(hdata, HDMI_V13_V_SYNC_GEN_2_1, core->v_sync_gen2[1]); + hdmi_reg_writeb(hdata, HDMI_V13_V_SYNC_GEN_2_2, core->v_sync_gen2[2]); + hdmi_reg_writeb(hdata, HDMI_V13_V_SYNC_GEN_3_0, core->v_sync_gen3[0]); + hdmi_reg_writeb(hdata, HDMI_V13_V_SYNC_GEN_3_1, core->v_sync_gen3[1]); + hdmi_reg_writeb(hdata, HDMI_V13_V_SYNC_GEN_3_2, core->v_sync_gen3[2]); + /* Timing generator registers */ + hdmi_reg_writeb(hdata, HDMI_TG_H_FSZ_L, tg->h_fsz_l); + hdmi_reg_writeb(hdata, HDMI_TG_H_FSZ_H, tg->h_fsz_h); + hdmi_reg_writeb(hdata, HDMI_TG_HACT_ST_L, tg->hact_st_l); + hdmi_reg_writeb(hdata, HDMI_TG_HACT_ST_H, tg->hact_st_h); + hdmi_reg_writeb(hdata, HDMI_TG_HACT_SZ_L, tg->hact_sz_l); + hdmi_reg_writeb(hdata, HDMI_TG_HACT_SZ_H, tg->hact_sz_h); + hdmi_reg_writeb(hdata, HDMI_TG_V_FSZ_L, tg->v_fsz_l); + hdmi_reg_writeb(hdata, HDMI_TG_V_FSZ_H, tg->v_fsz_h); + hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_L, tg->vsync_l); + hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_H, tg->vsync_h); + hdmi_reg_writeb(hdata, HDMI_TG_VSYNC2_L, tg->vsync2_l); + hdmi_reg_writeb(hdata, HDMI_TG_VSYNC2_H, tg->vsync2_h); + hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST_L, tg->vact_st_l); + hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST_H, tg->vact_st_h); + hdmi_reg_writeb(hdata, HDMI_TG_VACT_SZ_L, tg->vact_sz_l); + hdmi_reg_writeb(hdata, HDMI_TG_VACT_SZ_H, tg->vact_sz_h); + hdmi_reg_writeb(hdata, HDMI_TG_FIELD_CHG_L, tg->field_chg_l); + hdmi_reg_writeb(hdata, HDMI_TG_FIELD_CHG_H, tg->field_chg_h); + hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST2_L, tg->vact_st2_l); + hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST2_H, tg->vact_st2_h); + hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_TOP_HDMI_L, tg->vsync_top_hdmi_l); + hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_TOP_HDMI_H, tg->vsync_top_hdmi_h); + hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_BOT_HDMI_L, tg->vsync_bot_hdmi_l); + hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_BOT_HDMI_H, tg->vsync_bot_hdmi_h); + hdmi_reg_writeb(hdata, HDMI_TG_FIELD_TOP_HDMI_L, tg->field_top_hdmi_l); + hdmi_reg_writeb(hdata, HDMI_TG_FIELD_TOP_HDMI_H, tg->field_top_hdmi_h); + hdmi_reg_writeb(hdata, HDMI_TG_FIELD_BOT_HDMI_L, tg->field_bot_hdmi_l); + hdmi_reg_writeb(hdata, HDMI_TG_FIELD_BOT_HDMI_H, tg->field_bot_hdmi_h); + + /* waiting for HDMIPHY's PLL to get to steady state */ + for (tries = 100; tries; --tries) { + u32 val = hdmi_reg_read(hdata, HDMI_V13_PHY_STATUS); + if (val & HDMI_PHY_STATUS_READY) + break; + mdelay(1); + } + /* steady state not achieved */ + if (tries == 0) { + DRM_ERROR("hdmiphy's pll could not reach steady state.\n"); + hdmi_regs_dump(hdata, "timing apply"); + } + + clk_disable(hdata->res.sclk_hdmi); + clk_set_parent(hdata->res.sclk_hdmi, hdata->res.sclk_hdmiphy); + clk_enable(hdata->res.sclk_hdmi); + + /* enable HDMI and timing generator */ + hdmi_reg_writemask(hdata, HDMI_CON_0, ~0, HDMI_EN); + if (core->int_pro_mode[0]) + hdmi_reg_writemask(hdata, HDMI_TG_CMD, ~0, HDMI_TG_EN | + HDMI_FIELD_EN); + else + hdmi_reg_writemask(hdata, HDMI_TG_CMD, ~0, HDMI_TG_EN); +} + +static void hdmi_v14_timing_apply(struct hdmi_context *hdata) +{ + const struct hdmi_preset_conf *conf = hdmi_confs[hdata->cur_conf].conf; + const struct hdmi_core_regs *core = &conf->core; + const struct hdmi_tg_regs *tg = &conf->tg; + int tries; + + /* setting core registers */ + hdmi_reg_writeb(hdata, HDMI_H_BLANK_0, core->h_blank[0]); + hdmi_reg_writeb(hdata, HDMI_H_BLANK_1, core->h_blank[1]); + hdmi_reg_writeb(hdata, HDMI_V2_BLANK_0, core->v2_blank[0]); + hdmi_reg_writeb(hdata, HDMI_V2_BLANK_1, core->v2_blank[1]); + hdmi_reg_writeb(hdata, HDMI_V1_BLANK_0, core->v1_blank[0]); + hdmi_reg_writeb(hdata, HDMI_V1_BLANK_1, core->v1_blank[1]); + hdmi_reg_writeb(hdata, HDMI_V_LINE_0, core->v_line[0]); + hdmi_reg_writeb(hdata, HDMI_V_LINE_1, core->v_line[1]); + hdmi_reg_writeb(hdata, HDMI_H_LINE_0, core->h_line[0]); + hdmi_reg_writeb(hdata, HDMI_H_LINE_1, core->h_line[1]); + hdmi_reg_writeb(hdata, HDMI_HSYNC_POL, core->hsync_pol[0]); + hdmi_reg_writeb(hdata, HDMI_VSYNC_POL, core->vsync_pol[0]); + hdmi_reg_writeb(hdata, HDMI_INT_PRO_MODE, core->int_pro_mode[0]); + hdmi_reg_writeb(hdata, HDMI_V_BLANK_F0_0, core->v_blank_f0[0]); + hdmi_reg_writeb(hdata, HDMI_V_BLANK_F0_1, core->v_blank_f0[1]); + hdmi_reg_writeb(hdata, HDMI_V_BLANK_F1_0, core->v_blank_f1[0]); + hdmi_reg_writeb(hdata, HDMI_V_BLANK_F1_1, core->v_blank_f1[1]); + hdmi_reg_writeb(hdata, HDMI_H_SYNC_START_0, core->h_sync_start[0]); + hdmi_reg_writeb(hdata, HDMI_H_SYNC_START_1, core->h_sync_start[1]); + hdmi_reg_writeb(hdata, HDMI_H_SYNC_END_0, core->h_sync_end[0]); + hdmi_reg_writeb(hdata, HDMI_H_SYNC_END_1, core->h_sync_end[1]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_BEF_2_0, + core->v_sync_line_bef_2[0]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_BEF_2_1, + core->v_sync_line_bef_2[1]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_BEF_1_0, + core->v_sync_line_bef_1[0]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_BEF_1_1, + core->v_sync_line_bef_1[1]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_2_0, + core->v_sync_line_aft_2[0]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_2_1, + core->v_sync_line_aft_2[1]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_1_0, + core->v_sync_line_aft_1[0]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_1_1, + core->v_sync_line_aft_1[1]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_2_0, + core->v_sync_line_aft_pxl_2[0]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_2_1, + core->v_sync_line_aft_pxl_2[1]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_1_0, + core->v_sync_line_aft_pxl_1[0]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_1_1, + core->v_sync_line_aft_pxl_1[1]); + hdmi_reg_writeb(hdata, HDMI_V_BLANK_F2_0, core->v_blank_f2[0]); + hdmi_reg_writeb(hdata, HDMI_V_BLANK_F2_1, core->v_blank_f2[1]); + hdmi_reg_writeb(hdata, HDMI_V_BLANK_F3_0, core->v_blank_f3[0]); + hdmi_reg_writeb(hdata, HDMI_V_BLANK_F3_1, core->v_blank_f3[1]); + hdmi_reg_writeb(hdata, HDMI_V_BLANK_F4_0, core->v_blank_f4[0]); + hdmi_reg_writeb(hdata, HDMI_V_BLANK_F4_1, core->v_blank_f4[1]); + hdmi_reg_writeb(hdata, HDMI_V_BLANK_F5_0, core->v_blank_f5[0]); + hdmi_reg_writeb(hdata, HDMI_V_BLANK_F5_1, core->v_blank_f5[1]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_3_0, + core->v_sync_line_aft_3[0]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_3_1, + core->v_sync_line_aft_3[1]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_4_0, + core->v_sync_line_aft_4[0]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_4_1, + core->v_sync_line_aft_4[1]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_5_0, + core->v_sync_line_aft_5[0]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_5_1, + core->v_sync_line_aft_5[1]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_6_0, + core->v_sync_line_aft_6[0]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_6_1, + core->v_sync_line_aft_6[1]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_3_0, + core->v_sync_line_aft_pxl_3[0]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_3_1, + core->v_sync_line_aft_pxl_3[1]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_4_0, + core->v_sync_line_aft_pxl_4[0]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_4_1, + core->v_sync_line_aft_pxl_4[1]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_5_0, + core->v_sync_line_aft_pxl_5[0]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_5_1, + core->v_sync_line_aft_pxl_5[1]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_6_0, + core->v_sync_line_aft_pxl_6[0]); + hdmi_reg_writeb(hdata, HDMI_V_SYNC_LINE_AFT_PXL_6_1, + core->v_sync_line_aft_pxl_6[1]); + hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_1_0, core->vact_space_1[0]); + hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_1_1, core->vact_space_1[1]); + hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_2_0, core->vact_space_2[0]); + hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_2_1, core->vact_space_2[1]); + hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_3_0, core->vact_space_3[0]); + hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_3_1, core->vact_space_3[1]); + hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_4_0, core->vact_space_4[0]); + hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_4_1, core->vact_space_4[1]); + hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_5_0, core->vact_space_5[0]); + hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_5_1, core->vact_space_5[1]); + hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_6_0, core->vact_space_6[0]); + hdmi_reg_writeb(hdata, HDMI_VACT_SPACE_6_1, core->vact_space_6[1]); + + /* Timing generator registers */ + hdmi_reg_writeb(hdata, HDMI_TG_H_FSZ_L, tg->h_fsz_l); + hdmi_reg_writeb(hdata, HDMI_TG_H_FSZ_H, tg->h_fsz_h); + hdmi_reg_writeb(hdata, HDMI_TG_HACT_ST_L, tg->hact_st_l); + hdmi_reg_writeb(hdata, HDMI_TG_HACT_ST_H, tg->hact_st_h); + hdmi_reg_writeb(hdata, HDMI_TG_HACT_SZ_L, tg->hact_sz_l); + hdmi_reg_writeb(hdata, HDMI_TG_HACT_SZ_H, tg->hact_sz_h); + hdmi_reg_writeb(hdata, HDMI_TG_V_FSZ_L, tg->v_fsz_l); + hdmi_reg_writeb(hdata, HDMI_TG_V_FSZ_H, tg->v_fsz_h); + hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_L, tg->vsync_l); + hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_H, tg->vsync_h); + hdmi_reg_writeb(hdata, HDMI_TG_VSYNC2_L, tg->vsync2_l); + hdmi_reg_writeb(hdata, HDMI_TG_VSYNC2_H, tg->vsync2_h); + hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST_L, tg->vact_st_l); + hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST_H, tg->vact_st_h); + hdmi_reg_writeb(hdata, HDMI_TG_VACT_SZ_L, tg->vact_sz_l); + hdmi_reg_writeb(hdata, HDMI_TG_VACT_SZ_H, tg->vact_sz_h); + hdmi_reg_writeb(hdata, HDMI_TG_FIELD_CHG_L, tg->field_chg_l); + hdmi_reg_writeb(hdata, HDMI_TG_FIELD_CHG_H, tg->field_chg_h); + hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST2_L, tg->vact_st2_l); + hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST2_H, tg->vact_st2_h); + hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST3_L, tg->vact_st3_l); + hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST3_H, tg->vact_st3_h); + hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST4_L, tg->vact_st4_l); + hdmi_reg_writeb(hdata, HDMI_TG_VACT_ST4_H, tg->vact_st4_h); + hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_TOP_HDMI_L, tg->vsync_top_hdmi_l); + hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_TOP_HDMI_H, tg->vsync_top_hdmi_h); + hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_BOT_HDMI_L, tg->vsync_bot_hdmi_l); + hdmi_reg_writeb(hdata, HDMI_TG_VSYNC_BOT_HDMI_H, tg->vsync_bot_hdmi_h); + hdmi_reg_writeb(hdata, HDMI_TG_FIELD_TOP_HDMI_L, tg->field_top_hdmi_l); + hdmi_reg_writeb(hdata, HDMI_TG_FIELD_TOP_HDMI_H, tg->field_top_hdmi_h); + hdmi_reg_writeb(hdata, HDMI_TG_FIELD_BOT_HDMI_L, tg->field_bot_hdmi_l); + hdmi_reg_writeb(hdata, HDMI_TG_FIELD_BOT_HDMI_H, tg->field_bot_hdmi_h); + hdmi_reg_writeb(hdata, HDMI_TG_3D, tg->tg_3d); + + /* waiting for HDMIPHY's PLL to get to steady state */ + for (tries = 100; tries; --tries) { + u32 val = hdmi_reg_read(hdata, HDMI_PHY_STATUS_0); + if (val & HDMI_PHY_STATUS_READY) + break; + mdelay(1); + } + /* steady state not achieved */ + if (tries == 0) { + DRM_ERROR("hdmiphy's pll could not reach steady state.\n"); + hdmi_regs_dump(hdata, "timing apply"); + } + + clk_disable(hdata->res.sclk_hdmi); + clk_set_parent(hdata->res.sclk_hdmi, hdata->res.sclk_hdmiphy); + clk_enable(hdata->res.sclk_hdmi); + + /* enable HDMI and timing generator */ + hdmi_reg_writemask(hdata, HDMI_CON_0, ~0, HDMI_EN); + if (core->int_pro_mode[0]) + hdmi_reg_writemask(hdata, HDMI_TG_CMD, ~0, HDMI_TG_EN | + HDMI_FIELD_EN); + else + hdmi_reg_writemask(hdata, HDMI_TG_CMD, ~0, HDMI_TG_EN); +} + +static void hdmi_timing_apply(struct hdmi_context *hdata) +{ + if (hdata->is_v13) + hdmi_v13_timing_apply(hdata); + else + hdmi_v14_timing_apply(hdata); +} + +static void hdmiphy_conf_reset(struct hdmi_context *hdata) +{ + u8 buffer[2]; + u32 reg; + + clk_disable(hdata->res.sclk_hdmi); + clk_set_parent(hdata->res.sclk_hdmi, hdata->res.sclk_pixel); + clk_enable(hdata->res.sclk_hdmi); + + /* operation mode */ + buffer[0] = 0x1f; + buffer[1] = 0x00; + + if (hdata->hdmiphy_port) + i2c_master_send(hdata->hdmiphy_port, buffer, 2); + + if (hdata->is_v13) + reg = HDMI_V13_PHY_RSTOUT; + else + reg = HDMI_PHY_RSTOUT; + + /* reset hdmiphy */ + hdmi_reg_writemask(hdata, reg, ~0, HDMI_PHY_SW_RSTOUT); + mdelay(10); + hdmi_reg_writemask(hdata, reg, 0, HDMI_PHY_SW_RSTOUT); + mdelay(10); +} + +static void hdmiphy_conf_apply(struct hdmi_context *hdata) +{ + const u8 *hdmiphy_data; + u8 buffer[32]; + u8 operation[2]; + u8 read_buffer[32] = {0, }; + int ret; + int i; + + if (!hdata->hdmiphy_port) { + DRM_ERROR("hdmiphy is not attached\n"); + return; + } + + /* pixel clock */ + if (hdata->is_v13) + hdmiphy_data = hdmi_v13_confs[hdata->cur_conf].hdmiphy_data; + else + hdmiphy_data = hdmi_confs[hdata->cur_conf].hdmiphy_data; + + memcpy(buffer, hdmiphy_data, 32); + ret = i2c_master_send(hdata->hdmiphy_port, buffer, 32); + if (ret != 32) { + DRM_ERROR("failed to configure HDMIPHY via I2C\n"); + return; + } + + mdelay(10); + + /* operation mode */ + operation[0] = 0x1f; + operation[1] = 0x80; + + ret = i2c_master_send(hdata->hdmiphy_port, operation, 2); + if (ret != 2) { + DRM_ERROR("failed to enable hdmiphy\n"); + return; + } + + ret = i2c_master_recv(hdata->hdmiphy_port, read_buffer, 32); + if (ret < 0) { + DRM_ERROR("failed to read hdmiphy config\n"); + return; + } + + for (i = 0; i < ret; i++) + DRM_DEBUG_KMS("hdmiphy[0x%02x] write[0x%02x] - " + "recv [0x%02x]\n", i, buffer[i], read_buffer[i]); +} + +static void hdmi_conf_apply(struct hdmi_context *hdata) +{ + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + hdmiphy_conf_reset(hdata); + hdmiphy_conf_apply(hdata); + + hdmi_conf_reset(hdata); + hdmi_conf_init(hdata); + hdmi_audio_init(hdata); + + /* setting core registers */ + hdmi_timing_apply(hdata); + hdmi_audio_control(hdata, true); + + hdmi_regs_dump(hdata, "start"); +} + +static void hdmi_mode_fixup(void *ctx, struct drm_connector *connector, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + struct drm_display_mode *m; + struct hdmi_context *hdata = ctx; + int index; + + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + drm_mode_set_crtcinfo(adjusted_mode, 0); + + if (hdata->is_v13) + index = hdmi_v13_conf_index(adjusted_mode); + else + index = hdmi_v14_conf_index(adjusted_mode); + + /* just return if user desired mode exists. */ + if (index >= 0) + return; + + /* + * otherwise, find the most suitable mode among modes and change it + * to adjusted_mode. + */ + list_for_each_entry(m, &connector->modes, head) { + if (hdata->is_v13) + index = hdmi_v13_conf_index(m); + else + index = hdmi_v14_conf_index(m); + + if (index >= 0) { + DRM_INFO("desired mode doesn't exist so\n"); + DRM_INFO("use the most suitable mode among modes.\n"); + memcpy(adjusted_mode, m, sizeof(*m)); + break; + } + } +} + +static void hdmi_mode_set(void *ctx, void *mode) +{ + struct hdmi_context *hdata = ctx; + int conf_idx; + + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + conf_idx = hdmi_conf_index(hdata, mode); + if (conf_idx >= 0) + hdata->cur_conf = conf_idx; + else + DRM_DEBUG_KMS("not supported mode\n"); +} + +static void hdmi_get_max_resol(void *ctx, unsigned int *width, + unsigned int *height) +{ + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + *width = MAX_WIDTH; + *height = MAX_HEIGHT; +} + +static void hdmi_commit(void *ctx) +{ + struct hdmi_context *hdata = ctx; + + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + hdmi_conf_apply(hdata); + + hdata->enabled = true; +} + +static void hdmi_disable(void *ctx) +{ + struct hdmi_context *hdata = ctx; + + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + if (hdata->enabled) { + hdmi_audio_control(hdata, false); + hdmiphy_conf_reset(hdata); + hdmi_conf_reset(hdata); + } +} + +static struct exynos_hdmi_ops hdmi_ops = { + /* display */ + .is_connected = hdmi_is_connected, + .get_edid = hdmi_get_edid, + .check_timing = hdmi_check_timing, + .power_on = hdmi_display_power_on, + + /* manager */ + .mode_fixup = hdmi_mode_fixup, + .mode_set = hdmi_mode_set, + .get_max_resol = hdmi_get_max_resol, + .commit = hdmi_commit, + .disable = hdmi_disable, +}; + +/* + * Handle hotplug events outside the interrupt handler proper. + */ +static void hdmi_hotplug_func(struct work_struct *work) +{ + struct hdmi_context *hdata = + container_of(work, struct hdmi_context, hotplug_work); + struct exynos_drm_hdmi_context *ctx = + (struct exynos_drm_hdmi_context *)hdata->parent_ctx; + + drm_helper_hpd_irq_event(ctx->drm_dev); +} + +static irqreturn_t hdmi_irq_handler(int irq, void *arg) +{ + struct exynos_drm_hdmi_context *ctx = arg; + struct hdmi_context *hdata = ctx->ctx; + u32 intc_flag; + + intc_flag = hdmi_reg_read(hdata, HDMI_INTC_FLAG); + /* clearing flags for HPD plug/unplug */ + if (intc_flag & HDMI_INTC_FLAG_HPD_UNPLUG) { + DRM_DEBUG_KMS("unplugged, handling:%d\n", hdata->hpd_handle); + hdmi_reg_writemask(hdata, HDMI_INTC_FLAG, ~0, + HDMI_INTC_FLAG_HPD_UNPLUG); + } + if (intc_flag & HDMI_INTC_FLAG_HPD_PLUG) { + DRM_DEBUG_KMS("plugged, handling:%d\n", hdata->hpd_handle); + hdmi_reg_writemask(hdata, HDMI_INTC_FLAG, ~0, + HDMI_INTC_FLAG_HPD_PLUG); + } + + if (ctx->drm_dev && hdata->hpd_handle) + queue_work(hdata->wq, &hdata->hotplug_work); + + return IRQ_HANDLED; +} + +static int __devinit hdmi_resources_init(struct hdmi_context *hdata) +{ + struct device *dev = hdata->dev; + struct hdmi_resources *res = &hdata->res; + static char *supply[] = { + "hdmi-en", + "vdd", + "vdd_osc", + "vdd_pll", + }; + int i, ret; + + DRM_DEBUG_KMS("HDMI resource init\n"); + + memset(res, 0, sizeof *res); + + /* get clocks, power */ + res->hdmi = clk_get(dev, "hdmi"); + if (IS_ERR_OR_NULL(res->hdmi)) { + DRM_ERROR("failed to get clock 'hdmi'\n"); + goto fail; + } + res->sclk_hdmi = clk_get(dev, "sclk_hdmi"); + if (IS_ERR_OR_NULL(res->sclk_hdmi)) { + DRM_ERROR("failed to get clock 'sclk_hdmi'\n"); + goto fail; + } + res->sclk_pixel = clk_get(dev, "sclk_pixel"); + if (IS_ERR_OR_NULL(res->sclk_pixel)) { + DRM_ERROR("failed to get clock 'sclk_pixel'\n"); + goto fail; + } + res->sclk_hdmiphy = clk_get(dev, "sclk_hdmiphy"); + if (IS_ERR_OR_NULL(res->sclk_hdmiphy)) { + DRM_ERROR("failed to get clock 'sclk_hdmiphy'\n"); + goto fail; + } + res->hdmiphy = clk_get(dev, "hdmiphy"); + if (IS_ERR_OR_NULL(res->hdmiphy)) { + DRM_ERROR("failed to get clock 'hdmiphy'\n"); + goto fail; + } + + clk_set_parent(res->sclk_hdmi, res->sclk_pixel); + + res->regul_bulk = kzalloc(ARRAY_SIZE(supply) * + sizeof res->regul_bulk[0], GFP_KERNEL); + if (!res->regul_bulk) { + DRM_ERROR("failed to get memory for regulators\n"); + goto fail; + } + for (i = 0; i < ARRAY_SIZE(supply); ++i) { + res->regul_bulk[i].supply = supply[i]; + res->regul_bulk[i].consumer = NULL; + } + ret = regulator_bulk_get(dev, ARRAY_SIZE(supply), res->regul_bulk); + if (ret) { + DRM_ERROR("failed to get regulators\n"); + goto fail; + } + res->regul_count = ARRAY_SIZE(supply); + + return 0; +fail: + DRM_ERROR("HDMI resource init - failed\n"); + return -ENODEV; +} + +static int hdmi_resources_cleanup(struct hdmi_context *hdata) +{ + struct hdmi_resources *res = &hdata->res; + + regulator_bulk_free(res->regul_count, res->regul_bulk); + /* kfree is NULL-safe */ + kfree(res->regul_bulk); + if (!IS_ERR_OR_NULL(res->hdmiphy)) + clk_put(res->hdmiphy); + if (!IS_ERR_OR_NULL(res->sclk_hdmiphy)) + clk_put(res->sclk_hdmiphy); + if (!IS_ERR_OR_NULL(res->sclk_pixel)) + clk_put(res->sclk_pixel); + if (!IS_ERR_OR_NULL(res->sclk_hdmi)) + clk_put(res->sclk_hdmi); + if (!IS_ERR_OR_NULL(res->hdmi)) + clk_put(res->hdmi); + memset(res, 0, sizeof *res); + + return 0; +} + +static void hdmi_resource_poweron(struct hdmi_context *hdata) +{ + struct hdmi_resources *res = &hdata->res; + + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + /* turn HDMI power on */ + regulator_bulk_enable(res->regul_count, res->regul_bulk); + /* power-on hdmi physical interface */ + clk_enable(res->hdmiphy); + /* turn clocks on */ + clk_enable(res->hdmi); + clk_enable(res->sclk_hdmi); + + hdmiphy_conf_reset(hdata); + hdmi_conf_reset(hdata); + hdmi_conf_init(hdata); + hdmi_audio_init(hdata); +} + +static void hdmi_resource_poweroff(struct hdmi_context *hdata) +{ + struct hdmi_resources *res = &hdata->res; + + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + /* turn clocks off */ + clk_disable(res->sclk_hdmi); + clk_disable(res->hdmi); + /* power-off hdmiphy */ + clk_disable(res->hdmiphy); + /* turn HDMI power off */ + regulator_bulk_disable(res->regul_count, res->regul_bulk); +} + +static int hdmi_runtime_suspend(struct device *dev) +{ + struct exynos_drm_hdmi_context *ctx = get_hdmi_context(dev); + + DRM_DEBUG_KMS("%s\n", __func__); + + hdmi_resource_poweroff(ctx->ctx); + + return 0; +} + +static int hdmi_runtime_resume(struct device *dev) +{ + struct exynos_drm_hdmi_context *ctx = get_hdmi_context(dev); + + DRM_DEBUG_KMS("%s\n", __func__); + + hdmi_resource_poweron(ctx->ctx); + + return 0; +} + +static const struct dev_pm_ops hdmi_pm_ops = { + .runtime_suspend = hdmi_runtime_suspend, + .runtime_resume = hdmi_runtime_resume, +}; + +static struct i2c_client *hdmi_ddc, *hdmi_hdmiphy; + +void hdmi_attach_ddc_client(struct i2c_client *ddc) +{ + if (ddc) + hdmi_ddc = ddc; +} + +void hdmi_attach_hdmiphy_client(struct i2c_client *hdmiphy) +{ + if (hdmiphy) + hdmi_hdmiphy = hdmiphy; +} + +static int __devinit hdmi_probe(struct platform_device *pdev) +{ + struct device *dev = &pdev->dev; + struct exynos_drm_hdmi_context *drm_hdmi_ctx; + struct hdmi_context *hdata; + struct exynos_drm_hdmi_pdata *pdata; + struct resource *res; + int ret; + + DRM_DEBUG_KMS("[%d]\n", __LINE__); + + pdata = pdev->dev.platform_data; + if (!pdata) { + DRM_ERROR("no platform data specified\n"); + return -EINVAL; + } + + drm_hdmi_ctx = kzalloc(sizeof(*drm_hdmi_ctx), GFP_KERNEL); + if (!drm_hdmi_ctx) { + DRM_ERROR("failed to allocate common hdmi context.\n"); + return -ENOMEM; + } + + hdata = kzalloc(sizeof(struct hdmi_context), GFP_KERNEL); + if (!hdata) { + DRM_ERROR("out of memory\n"); + kfree(drm_hdmi_ctx); + return -ENOMEM; + } + + drm_hdmi_ctx->ctx = (void *)hdata; + hdata->parent_ctx = (void *)drm_hdmi_ctx; + + platform_set_drvdata(pdev, drm_hdmi_ctx); + + hdata->is_v13 = pdata->is_v13; + hdata->default_win = pdata->default_win; + hdata->default_timing = &pdata->timing; + hdata->default_bpp = pdata->bpp; + hdata->dev = dev; + + ret = hdmi_resources_init(hdata); + if (ret) { + ret = -EINVAL; + goto err_data; + } + + res = platform_get_resource(pdev, IORESOURCE_MEM, 0); + if (!res) { + DRM_ERROR("failed to find registers\n"); + ret = -ENOENT; + goto err_resource; + } + + hdata->regs_res = request_mem_region(res->start, resource_size(res), + dev_name(dev)); + if (!hdata->regs_res) { + DRM_ERROR("failed to claim register region\n"); + ret = -ENOENT; + goto err_resource; + } + + hdata->regs = ioremap(res->start, resource_size(res)); + if (!hdata->regs) { + DRM_ERROR("failed to map registers\n"); + ret = -ENXIO; + goto err_req_region; + } + + /* DDC i2c driver */ + if (i2c_add_driver(&ddc_driver)) { + DRM_ERROR("failed to register ddc i2c driver\n"); + ret = -ENOENT; + goto err_iomap; + } + + hdata->ddc_port = hdmi_ddc; + + /* hdmiphy i2c driver */ + if (i2c_add_driver(&hdmiphy_driver)) { + DRM_ERROR("failed to register hdmiphy i2c driver\n"); + ret = -ENOENT; + goto err_ddc; + } + + hdata->hdmiphy_port = hdmi_hdmiphy; + + res = platform_get_resource(pdev, IORESOURCE_IRQ, 0); + if (res == NULL) { + DRM_ERROR("get interrupt resource failed.\n"); + ret = -ENXIO; + goto err_hdmiphy; + } + + /* create workqueue and hotplug work */ + hdata->wq = alloc_workqueue("exynos-drm-hdmi", + WQ_UNBOUND | WQ_NON_REENTRANT, 1); + if (hdata->wq == NULL) { + DRM_ERROR("Failed to create workqueue.\n"); + ret = -ENOMEM; + goto err_hdmiphy; + } + INIT_WORK(&hdata->hotplug_work, hdmi_hotplug_func); + + /* register hpd interrupt */ + ret = request_irq(res->start, hdmi_irq_handler, 0, "drm_hdmi", + drm_hdmi_ctx); + if (ret) { + DRM_ERROR("request interrupt failed.\n"); + goto err_workqueue; + } + hdata->irq = res->start; + + /* register specific callbacks to common hdmi. */ + exynos_hdmi_ops_register(&hdmi_ops); + + hdmi_resource_poweron(hdata); + + return 0; + +err_workqueue: + destroy_workqueue(hdata->wq); +err_hdmiphy: + i2c_del_driver(&hdmiphy_driver); +err_ddc: + i2c_del_driver(&ddc_driver); +err_iomap: + iounmap(hdata->regs); +err_req_region: + release_mem_region(hdata->regs_res->start, + resource_size(hdata->regs_res)); +err_resource: + hdmi_resources_cleanup(hdata); +err_data: + kfree(hdata); + kfree(drm_hdmi_ctx); + return ret; +} + +static int __devexit hdmi_remove(struct platform_device *pdev) +{ + struct exynos_drm_hdmi_context *ctx = platform_get_drvdata(pdev); + struct hdmi_context *hdata = ctx->ctx; + + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + hdmi_resource_poweroff(hdata); + + disable_irq(hdata->irq); + free_irq(hdata->irq, hdata); + + cancel_work_sync(&hdata->hotplug_work); + destroy_workqueue(hdata->wq); + + hdmi_resources_cleanup(hdata); + + iounmap(hdata->regs); + + release_mem_region(hdata->regs_res->start, + resource_size(hdata->regs_res)); + + /* hdmiphy i2c driver */ + i2c_del_driver(&hdmiphy_driver); + /* DDC i2c driver */ + i2c_del_driver(&ddc_driver); + + kfree(hdata); + + return 0; +} + +struct platform_driver hdmi_driver = { + .probe = hdmi_probe, + .remove = __devexit_p(hdmi_remove), + .driver = { + .name = "exynos4-hdmi", + .owner = THIS_MODULE, + .pm = &hdmi_pm_ops, + }, +}; diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.h b/drivers/gpu/drm/exynos/exynos_hdmi.h new file mode 100644 index 0000000..1c3b6d8 --- /dev/null +++ b/drivers/gpu/drm/exynos/exynos_hdmi.h @@ -0,0 +1,37 @@ +/* + * + * Copyright (c) 2011 Samsung Electronics Co., Ltd. + * Authors: + * Inki Dae + * Seung-Woo Kim + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * VA LINUX SYSTEMS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ + +#ifndef _EXYNOS_HDMI_H_ +#define _EXYNOS_HDMI_H_ + +void hdmi_attach_ddc_client(struct i2c_client *ddc); +void hdmi_attach_hdmiphy_client(struct i2c_client *hdmiphy); + +extern struct i2c_driver hdmiphy_driver; +extern struct i2c_driver ddc_driver; + +#endif diff --git a/drivers/gpu/drm/exynos/exynos_hdmiphy.c b/drivers/gpu/drm/exynos/exynos_hdmiphy.c new file mode 100644 index 0000000..9fe2995 --- /dev/null +++ b/drivers/gpu/drm/exynos/exynos_hdmiphy.c @@ -0,0 +1,58 @@ +/* + * Copyright (C) 2011 Samsung Electronics Co.Ltd + * Authors: + * Seung-Woo Kim + * Inki Dae + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or (at your + * option) any later version. + * + */ + +#include "drmP.h" + +#include +#include +#include + +#include "exynos_drm_drv.h" +#include "exynos_hdmi.h" + + +static int hdmiphy_probe(struct i2c_client *client, + const struct i2c_device_id *id) +{ + hdmi_attach_hdmiphy_client(client); + + dev_info(&client->adapter->dev, "attached s5p_hdmiphy " + "into i2c adapter successfully\n"); + + return 0; +} + +static int hdmiphy_remove(struct i2c_client *client) +{ + dev_info(&client->adapter->dev, "detached s5p_hdmiphy " + "from i2c adapter successfully\n"); + + return 0; +} + +static const struct i2c_device_id hdmiphy_id[] = { + { "s5p_hdmiphy", 0 }, + { }, +}; + +struct i2c_driver hdmiphy_driver = { + .driver = { + .name = "s5p-hdmiphy", + .owner = THIS_MODULE, + }, + .id_table = hdmiphy_id, + .probe = hdmiphy_probe, + .remove = __devexit_p(hdmiphy_remove), + .command = NULL, +}; +EXPORT_SYMBOL(hdmiphy_driver); diff --git a/drivers/gpu/drm/exynos/exynos_mixer.c b/drivers/gpu/drm/exynos/exynos_mixer.c new file mode 100644 index 0000000..e15438c --- /dev/null +++ b/drivers/gpu/drm/exynos/exynos_mixer.c @@ -0,0 +1,1112 @@ +/* + * Copyright (C) 2011 Samsung Electronics Co.Ltd + * Authors: + * Seung-Woo Kim + * Inki Dae + * Joonyoung Shim + * + * Based on drivers/media/video/s5p-tv/mixer_reg.c + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2 of the License, or (at your + * option) any later version. + * + */ + +#include "drmP.h" + +#include "regs-mixer.h" +#include "regs-vp.h" + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include + +#include "exynos_drm_drv.h" +#include "exynos_drm_hdmi.h" + +#define MIXER_WIN_NR 3 +#define MIXER_DEFAULT_WIN 0 + +#define get_mixer_context(dev) platform_get_drvdata(to_platform_device(dev)) + +struct hdmi_win_data { + dma_addr_t dma_addr; + void __iomem *vaddr; + dma_addr_t chroma_dma_addr; + void __iomem *chroma_vaddr; + uint32_t pixel_format; + unsigned int bpp; + unsigned int crtc_x; + unsigned int crtc_y; + unsigned int crtc_width; + unsigned int crtc_height; + unsigned int fb_x; + unsigned int fb_y; + unsigned int fb_width; + unsigned int fb_height; + unsigned int mode_width; + unsigned int mode_height; + unsigned int scan_flags; +}; + +struct mixer_resources { + struct device *dev; + int irq; + void __iomem *mixer_regs; + void __iomem *vp_regs; + spinlock_t reg_slock; + struct clk *mixer; + struct clk *vp; + struct clk *sclk_mixer; + struct clk *sclk_hdmi; + struct clk *sclk_dac; +}; + +struct mixer_context { + unsigned int irq; + int pipe; + bool interlace; + + struct mixer_resources mixer_res; + struct hdmi_win_data win_data[MIXER_WIN_NR]; +}; + +static const u8 filter_y_horiz_tap8[] = { + 0, -1, -1, -1, -1, -1, -1, -1, + -1, -1, -1, -1, -1, 0, 0, 0, + 0, 2, 4, 5, 6, 6, 6, 6, + 6, 5, 5, 4, 3, 2, 1, 1, + 0, -6, -12, -16, -18, -20, -21, -20, + -20, -18, -16, -13, -10, -8, -5, -2, + 127, 126, 125, 121, 114, 107, 99, 89, + 79, 68, 57, 46, 35, 25, 16, 8, +}; + +static const u8 filter_y_vert_tap4[] = { + 0, -3, -6, -8, -8, -8, -8, -7, + -6, -5, -4, -3, -2, -1, -1, 0, + 127, 126, 124, 118, 111, 102, 92, 81, + 70, 59, 48, 37, 27, 19, 11, 5, + 0, 5, 11, 19, 27, 37, 48, 59, + 70, 81, 92, 102, 111, 118, 124, 126, + 0, 0, -1, -1, -2, -3, -4, -5, + -6, -7, -8, -8, -8, -8, -6, -3, +}; + +static const u8 filter_cr_horiz_tap4[] = { + 0, -3, -6, -8, -8, -8, -8, -7, + -6, -5, -4, -3, -2, -1, -1, 0, + 127, 126, 124, 118, 111, 102, 92, 81, + 70, 59, 48, 37, 27, 19, 11, 5, +}; + +static inline u32 vp_reg_read(struct mixer_resources *res, u32 reg_id) +{ + return readl(res->vp_regs + reg_id); +} + +static inline void vp_reg_write(struct mixer_resources *res, u32 reg_id, + u32 val) +{ + writel(val, res->vp_regs + reg_id); +} + +static inline void vp_reg_writemask(struct mixer_resources *res, u32 reg_id, + u32 val, u32 mask) +{ + u32 old = vp_reg_read(res, reg_id); + + val = (val & mask) | (old & ~mask); + writel(val, res->vp_regs + reg_id); +} + +static inline u32 mixer_reg_read(struct mixer_resources *res, u32 reg_id) +{ + return readl(res->mixer_regs + reg_id); +} + +static inline void mixer_reg_write(struct mixer_resources *res, u32 reg_id, + u32 val) +{ + writel(val, res->mixer_regs + reg_id); +} + +static inline void mixer_reg_writemask(struct mixer_resources *res, + u32 reg_id, u32 val, u32 mask) +{ + u32 old = mixer_reg_read(res, reg_id); + + val = (val & mask) | (old & ~mask); + writel(val, res->mixer_regs + reg_id); +} + +static void mixer_regs_dump(struct mixer_context *ctx) +{ +#define DUMPREG(reg_id) \ +do { \ + DRM_DEBUG_KMS(#reg_id " = %08x\n", \ + (u32)readl(ctx->mixer_res.mixer_regs + reg_id)); \ +} while (0) + + DUMPREG(MXR_STATUS); + DUMPREG(MXR_CFG); + DUMPREG(MXR_INT_EN); + DUMPREG(MXR_INT_STATUS); + + DUMPREG(MXR_LAYER_CFG); + DUMPREG(MXR_VIDEO_CFG); + + DUMPREG(MXR_GRAPHIC0_CFG); + DUMPREG(MXR_GRAPHIC0_BASE); + DUMPREG(MXR_GRAPHIC0_SPAN); + DUMPREG(MXR_GRAPHIC0_WH); + DUMPREG(MXR_GRAPHIC0_SXY); + DUMPREG(MXR_GRAPHIC0_DXY); + + DUMPREG(MXR_GRAPHIC1_CFG); + DUMPREG(MXR_GRAPHIC1_BASE); + DUMPREG(MXR_GRAPHIC1_SPAN); + DUMPREG(MXR_GRAPHIC1_WH); + DUMPREG(MXR_GRAPHIC1_SXY); + DUMPREG(MXR_GRAPHIC1_DXY); +#undef DUMPREG +} + +static void vp_regs_dump(struct mixer_context *ctx) +{ +#define DUMPREG(reg_id) \ +do { \ + DRM_DEBUG_KMS(#reg_id " = %08x\n", \ + (u32) readl(ctx->mixer_res.vp_regs + reg_id)); \ +} while (0) + + DUMPREG(VP_ENABLE); + DUMPREG(VP_SRESET); + DUMPREG(VP_SHADOW_UPDATE); + DUMPREG(VP_FIELD_ID); + DUMPREG(VP_MODE); + DUMPREG(VP_IMG_SIZE_Y); + DUMPREG(VP_IMG_SIZE_C); + DUMPREG(VP_PER_RATE_CTRL); + DUMPREG(VP_TOP_Y_PTR); + DUMPREG(VP_BOT_Y_PTR); + DUMPREG(VP_TOP_C_PTR); + DUMPREG(VP_BOT_C_PTR); + DUMPREG(VP_ENDIAN_MODE); + DUMPREG(VP_SRC_H_POSITION); + DUMPREG(VP_SRC_V_POSITION); + DUMPREG(VP_SRC_WIDTH); + DUMPREG(VP_SRC_HEIGHT); + DUMPREG(VP_DST_H_POSITION); + DUMPREG(VP_DST_V_POSITION); + DUMPREG(VP_DST_WIDTH); + DUMPREG(VP_DST_HEIGHT); + DUMPREG(VP_H_RATIO); + DUMPREG(VP_V_RATIO); + +#undef DUMPREG +} + +static inline void vp_filter_set(struct mixer_resources *res, + int reg_id, const u8 *data, unsigned int size) +{ + /* assure 4-byte align */ + BUG_ON(size & 3); + for (; size; size -= 4, reg_id += 4, data += 4) { + u32 val = (data[0] << 24) | (data[1] << 16) | + (data[2] << 8) | data[3]; + vp_reg_write(res, reg_id, val); + } +} + +static void vp_default_filter(struct mixer_resources *res) +{ + vp_filter_set(res, VP_POLY8_Y0_LL, + filter_y_horiz_tap8, sizeof filter_y_horiz_tap8); + vp_filter_set(res, VP_POLY4_Y0_LL, + filter_y_vert_tap4, sizeof filter_y_vert_tap4); + vp_filter_set(res, VP_POLY4_C0_LL, + filter_cr_horiz_tap4, sizeof filter_cr_horiz_tap4); +} + +static void mixer_vsync_set_update(struct mixer_context *ctx, bool enable) +{ + struct mixer_resources *res = &ctx->mixer_res; + + /* block update on vsync */ + mixer_reg_writemask(res, MXR_STATUS, enable ? + MXR_STATUS_SYNC_ENABLE : 0, MXR_STATUS_SYNC_ENABLE); + + vp_reg_write(res, VP_SHADOW_UPDATE, enable ? + VP_SHADOW_UPDATE_ENABLE : 0); +} + +static void mixer_cfg_scan(struct mixer_context *ctx, unsigned int height) +{ + struct mixer_resources *res = &ctx->mixer_res; + u32 val; + + /* choosing between interlace and progressive mode */ + val = (ctx->interlace ? MXR_CFG_SCAN_INTERLACE : + MXR_CFG_SCAN_PROGRASSIVE); + + /* choosing between porper HD and SD mode */ + if (height == 480) + val |= MXR_CFG_SCAN_NTSC | MXR_CFG_SCAN_SD; + else if (height == 576) + val |= MXR_CFG_SCAN_PAL | MXR_CFG_SCAN_SD; + else if (height == 720) + val |= MXR_CFG_SCAN_HD_720 | MXR_CFG_SCAN_HD; + else if (height == 1080) + val |= MXR_CFG_SCAN_HD_1080 | MXR_CFG_SCAN_HD; + else + val |= MXR_CFG_SCAN_HD_720 | MXR_CFG_SCAN_HD; + + mixer_reg_writemask(res, MXR_CFG, val, MXR_CFG_SCAN_MASK); +} + +static void mixer_cfg_rgb_fmt(struct mixer_context *ctx, unsigned int height) +{ + struct mixer_resources *res = &ctx->mixer_res; + u32 val; + + if (height == 480) { + val = MXR_CFG_RGB601_0_255; + } else if (height == 576) { + val = MXR_CFG_RGB601_0_255; + } else if (height == 720) { + val = MXR_CFG_RGB709_16_235; + mixer_reg_write(res, MXR_CM_COEFF_Y, + (1 << 30) | (94 << 20) | (314 << 10) | + (32 << 0)); + mixer_reg_write(res, MXR_CM_COEFF_CB, + (972 << 20) | (851 << 10) | (225 << 0)); + mixer_reg_write(res, MXR_CM_COEFF_CR, + (225 << 20) | (820 << 10) | (1004 << 0)); + } else if (height == 1080) { + val = MXR_CFG_RGB709_16_235; + mixer_reg_write(res, MXR_CM_COEFF_Y, + (1 << 30) | (94 << 20) | (314 << 10) | + (32 << 0)); + mixer_reg_write(res, MXR_CM_COEFF_CB, + (972 << 20) | (851 << 10) | (225 << 0)); + mixer_reg_write(res, MXR_CM_COEFF_CR, + (225 << 20) | (820 << 10) | (1004 << 0)); + } else { + val = MXR_CFG_RGB709_16_235; + mixer_reg_write(res, MXR_CM_COEFF_Y, + (1 << 30) | (94 << 20) | (314 << 10) | + (32 << 0)); + mixer_reg_write(res, MXR_CM_COEFF_CB, + (972 << 20) | (851 << 10) | (225 << 0)); + mixer_reg_write(res, MXR_CM_COEFF_CR, + (225 << 20) | (820 << 10) | (1004 << 0)); + } + + mixer_reg_writemask(res, MXR_CFG, val, MXR_CFG_RGB_FMT_MASK); +} + +static void mixer_cfg_layer(struct mixer_context *ctx, int win, bool enable) +{ + struct mixer_resources *res = &ctx->mixer_res; + u32 val = enable ? ~0 : 0; + + switch (win) { + case 0: + mixer_reg_writemask(res, MXR_CFG, val, MXR_CFG_GRP0_ENABLE); + break; + case 1: + mixer_reg_writemask(res, MXR_CFG, val, MXR_CFG_GRP1_ENABLE); + break; + case 2: + vp_reg_writemask(res, VP_ENABLE, val, VP_ENABLE_ON); + mixer_reg_writemask(res, MXR_CFG, val, MXR_CFG_VP_ENABLE); + break; + } +} + +static void mixer_run(struct mixer_context *ctx) +{ + struct mixer_resources *res = &ctx->mixer_res; + + mixer_reg_writemask(res, MXR_STATUS, ~0, MXR_STATUS_REG_RUN); + + mixer_regs_dump(ctx); +} + +static void vp_video_buffer(struct mixer_context *ctx, int win) +{ + struct mixer_resources *res = &ctx->mixer_res; + unsigned long flags; + struct hdmi_win_data *win_data; + unsigned int full_width, full_height, width, height; + unsigned int x_ratio, y_ratio; + unsigned int src_x_offset, src_y_offset, dst_x_offset, dst_y_offset; + unsigned int mode_width, mode_height; + unsigned int buf_num; + dma_addr_t luma_addr[2], chroma_addr[2]; + bool tiled_mode = false; + bool crcb_mode = false; + u32 val; + + win_data = &ctx->win_data[win]; + + switch (win_data->pixel_format) { + case DRM_FORMAT_NV12MT: + tiled_mode = true; + case DRM_FORMAT_NV12M: + crcb_mode = false; + buf_num = 2; + break; + /* TODO: single buffer format NV12, NV21 */ + default: + /* ignore pixel format at disable time */ + if (!win_data->dma_addr) + break; + + DRM_ERROR("pixel format for vp is wrong [%d].\n", + win_data->pixel_format); + return; + } + + full_width = win_data->fb_width; + full_height = win_data->fb_height; + width = win_data->crtc_width; + height = win_data->crtc_height; + mode_width = win_data->mode_width; + mode_height = win_data->mode_height; + + /* scaling feature: (src << 16) / dst */ + x_ratio = (width << 16) / width; + y_ratio = (height << 16) / height; + + src_x_offset = win_data->fb_x; + src_y_offset = win_data->fb_y; + dst_x_offset = win_data->crtc_x; + dst_y_offset = win_data->crtc_y; + + if (buf_num == 2) { + luma_addr[0] = win_data->dma_addr; + chroma_addr[0] = win_data->chroma_dma_addr; + } else { + luma_addr[0] = win_data->dma_addr; + chroma_addr[0] = win_data->dma_addr + + (full_width * full_height); + } + + if (win_data->scan_flags & DRM_MODE_FLAG_INTERLACE) { + ctx->interlace = true; + if (tiled_mode) { + luma_addr[1] = luma_addr[0] + 0x40; + chroma_addr[1] = chroma_addr[0] + 0x40; + } else { + luma_addr[1] = luma_addr[0] + full_width; + chroma_addr[1] = chroma_addr[0] + full_width; + } + } else { + ctx->interlace = false; + luma_addr[1] = 0; + chroma_addr[1] = 0; + } + + spin_lock_irqsave(&res->reg_slock, flags); + mixer_vsync_set_update(ctx, false); + + /* interlace or progressive scan mode */ + val = (ctx->interlace ? ~0 : 0); + vp_reg_writemask(res, VP_MODE, val, VP_MODE_LINE_SKIP); + + /* setup format */ + val = (crcb_mode ? VP_MODE_NV21 : VP_MODE_NV12); + val |= (tiled_mode ? VP_MODE_MEM_TILED : VP_MODE_MEM_LINEAR); + vp_reg_writemask(res, VP_MODE, val, VP_MODE_FMT_MASK); + + /* setting size of input image */ + vp_reg_write(res, VP_IMG_SIZE_Y, VP_IMG_HSIZE(full_width) | + VP_IMG_VSIZE(full_height)); + /* chroma height has to reduced by 2 to avoid chroma distorions */ + vp_reg_write(res, VP_IMG_SIZE_C, VP_IMG_HSIZE(full_width) | + VP_IMG_VSIZE(full_height / 2)); + + vp_reg_write(res, VP_SRC_WIDTH, width); + vp_reg_write(res, VP_SRC_HEIGHT, height); + vp_reg_write(res, VP_SRC_H_POSITION, + VP_SRC_H_POSITION_VAL(src_x_offset)); + vp_reg_write(res, VP_SRC_V_POSITION, src_y_offset); + + vp_reg_write(res, VP_DST_WIDTH, width); + vp_reg_write(res, VP_DST_H_POSITION, dst_x_offset); + if (ctx->interlace) { + vp_reg_write(res, VP_DST_HEIGHT, height / 2); + vp_reg_write(res, VP_DST_V_POSITION, dst_y_offset / 2); + } else { + vp_reg_write(res, VP_DST_HEIGHT, height); + vp_reg_write(res, VP_DST_V_POSITION, dst_y_offset); + } + + vp_reg_write(res, VP_H_RATIO, x_ratio); + vp_reg_write(res, VP_V_RATIO, y_ratio); + + vp_reg_write(res, VP_ENDIAN_MODE, VP_ENDIAN_MODE_LITTLE); + + /* set buffer address to vp */ + vp_reg_write(res, VP_TOP_Y_PTR, luma_addr[0]); + vp_reg_write(res, VP_BOT_Y_PTR, luma_addr[1]); + vp_reg_write(res, VP_TOP_C_PTR, chroma_addr[0]); + vp_reg_write(res, VP_BOT_C_PTR, chroma_addr[1]); + + mixer_cfg_scan(ctx, mode_height); + mixer_cfg_rgb_fmt(ctx, mode_height); + mixer_cfg_layer(ctx, win, true); + mixer_run(ctx); + + mixer_vsync_set_update(ctx, true); + spin_unlock_irqrestore(&res->reg_slock, flags); + + vp_regs_dump(ctx); +} + +static void mixer_graph_buffer(struct mixer_context *ctx, int win) +{ + struct mixer_resources *res = &ctx->mixer_res; + unsigned long flags; + struct hdmi_win_data *win_data; + unsigned int full_width, width, height; + unsigned int x_ratio, y_ratio; + unsigned int src_x_offset, src_y_offset, dst_x_offset, dst_y_offset; + unsigned int mode_width, mode_height; + dma_addr_t dma_addr; + unsigned int fmt; + u32 val; + + win_data = &ctx->win_data[win]; + + #define RGB565 4 + #define ARGB1555 5 + #define ARGB4444 6 + #define ARGB8888 7 + + switch (win_data->bpp) { + case 16: + fmt = ARGB4444; + break; + case 32: + fmt = ARGB8888; + break; + default: + fmt = ARGB8888; + } + + dma_addr = win_data->dma_addr; + full_width = win_data->fb_width; + width = win_data->crtc_width; + height = win_data->crtc_height; + mode_width = win_data->mode_width; + mode_height = win_data->mode_height; + + /* 2x scaling feature */ + x_ratio = 0; + y_ratio = 0; + + src_x_offset = win_data->fb_x; + src_y_offset = win_data->fb_y; + dst_x_offset = win_data->crtc_x; + dst_y_offset = win_data->crtc_y; + + /* converting dma address base and source offset */ + dma_addr = dma_addr + + (src_x_offset * win_data->bpp >> 3) + + (src_y_offset * full_width * win_data->bpp >> 3); + src_x_offset = 0; + src_y_offset = 0; + + if (win_data->scan_flags & DRM_MODE_FLAG_INTERLACE) + ctx->interlace = true; + else + ctx->interlace = false; + + spin_lock_irqsave(&res->reg_slock, flags); + mixer_vsync_set_update(ctx, false); + + /* setup format */ + mixer_reg_writemask(res, MXR_GRAPHIC_CFG(win), + MXR_GRP_CFG_FORMAT_VAL(fmt), MXR_GRP_CFG_FORMAT_MASK); + + /* setup geometry */ + mixer_reg_write(res, MXR_GRAPHIC_SPAN(win), full_width); + + val = MXR_GRP_WH_WIDTH(width); + val |= MXR_GRP_WH_HEIGHT(height); + val |= MXR_GRP_WH_H_SCALE(x_ratio); + val |= MXR_GRP_WH_V_SCALE(y_ratio); + mixer_reg_write(res, MXR_GRAPHIC_WH(win), val); + + /* setup offsets in source image */ + val = MXR_GRP_SXY_SX(src_x_offset); + val |= MXR_GRP_SXY_SY(src_y_offset); + mixer_reg_write(res, MXR_GRAPHIC_SXY(win), val); + + /* setup offsets in display image */ + val = MXR_GRP_DXY_DX(dst_x_offset); + val |= MXR_GRP_DXY_DY(dst_y_offset); + mixer_reg_write(res, MXR_GRAPHIC_DXY(win), val); + + /* set buffer address to mixer */ + mixer_reg_write(res, MXR_GRAPHIC_BASE(win), dma_addr); + + mixer_cfg_scan(ctx, mode_height); + mixer_cfg_rgb_fmt(ctx, mode_height); + mixer_cfg_layer(ctx, win, true); + mixer_run(ctx); + + mixer_vsync_set_update(ctx, true); + spin_unlock_irqrestore(&res->reg_slock, flags); +} + +static void vp_win_reset(struct mixer_context *ctx) +{ + struct mixer_resources *res = &ctx->mixer_res; + int tries = 100; + + vp_reg_write(res, VP_SRESET, VP_SRESET_PROCESSING); + for (tries = 100; tries; --tries) { + /* waiting until VP_SRESET_PROCESSING is 0 */ + if (~vp_reg_read(res, VP_SRESET) & VP_SRESET_PROCESSING) + break; + mdelay(10); + } + WARN(tries == 0, "failed to reset Video Processor\n"); +} + +static int mixer_enable_vblank(void *ctx, int pipe) +{ + struct mixer_context *mixer_ctx = ctx; + struct mixer_resources *res = &mixer_ctx->mixer_res; + + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + mixer_ctx->pipe = pipe; + + /* enable vsync interrupt */ + mixer_reg_writemask(res, MXR_INT_EN, MXR_INT_EN_VSYNC, + MXR_INT_EN_VSYNC); + + return 0; +} + +static void mixer_disable_vblank(void *ctx) +{ + struct mixer_context *mixer_ctx = ctx; + struct mixer_resources *res = &mixer_ctx->mixer_res; + + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + /* disable vsync interrupt */ + mixer_reg_writemask(res, MXR_INT_EN, 0, MXR_INT_EN_VSYNC); +} + +static void mixer_win_mode_set(void *ctx, + struct exynos_drm_overlay *overlay) +{ + struct mixer_context *mixer_ctx = ctx; + struct hdmi_win_data *win_data; + int win; + + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + if (!overlay) { + DRM_ERROR("overlay is NULL\n"); + return; + } + + DRM_DEBUG_KMS("set [%d]x[%d] at (%d,%d) to [%d]x[%d] at (%d,%d)\n", + overlay->fb_width, overlay->fb_height, + overlay->fb_x, overlay->fb_y, + overlay->crtc_width, overlay->crtc_height, + overlay->crtc_x, overlay->crtc_y); + + win = overlay->zpos; + if (win == DEFAULT_ZPOS) + win = MIXER_DEFAULT_WIN; + + if (win < 0 || win > MIXER_WIN_NR) { + DRM_ERROR("overlay plane[%d] is wrong\n", win); + return; + } + + win_data = &mixer_ctx->win_data[win]; + + win_data->dma_addr = overlay->dma_addr[0]; + win_data->vaddr = overlay->vaddr[0]; + win_data->chroma_dma_addr = overlay->dma_addr[1]; + win_data->chroma_vaddr = overlay->vaddr[1]; + win_data->pixel_format = overlay->pixel_format; + win_data->bpp = overlay->bpp; + + win_data->crtc_x = overlay->crtc_x; + win_data->crtc_y = overlay->crtc_y; + win_data->crtc_width = overlay->crtc_width; + win_data->crtc_height = overlay->crtc_height; + + win_data->fb_x = overlay->fb_x; + win_data->fb_y = overlay->fb_y; + win_data->fb_width = overlay->fb_width; + win_data->fb_height = overlay->fb_height; + + win_data->mode_width = overlay->mode_width; + win_data->mode_height = overlay->mode_height; + + win_data->scan_flags = overlay->scan_flag; +} + +static void mixer_win_commit(void *ctx, int zpos) +{ + struct mixer_context *mixer_ctx = ctx; + int win = zpos; + + DRM_DEBUG_KMS("[%d] %s, win: %d\n", __LINE__, __func__, win); + + if (win == DEFAULT_ZPOS) + win = MIXER_DEFAULT_WIN; + + if (win < 0 || win > MIXER_WIN_NR) { + DRM_ERROR("overlay plane[%d] is wrong\n", win); + return; + } + + if (win > 1) + vp_video_buffer(mixer_ctx, win); + else + mixer_graph_buffer(mixer_ctx, win); +} + +static void mixer_win_disable(void *ctx, int zpos) +{ + struct mixer_context *mixer_ctx = ctx; + struct mixer_resources *res = &mixer_ctx->mixer_res; + unsigned long flags; + int win = zpos; + + DRM_DEBUG_KMS("[%d] %s, win: %d\n", __LINE__, __func__, win); + + if (win == DEFAULT_ZPOS) + win = MIXER_DEFAULT_WIN; + + if (win < 0 || win > MIXER_WIN_NR) { + DRM_ERROR("overlay plane[%d] is wrong\n", win); + return; + } + + spin_lock_irqsave(&res->reg_slock, flags); + mixer_vsync_set_update(mixer_ctx, false); + + mixer_cfg_layer(mixer_ctx, win, false); + + mixer_vsync_set_update(mixer_ctx, true); + spin_unlock_irqrestore(&res->reg_slock, flags); +} + +static struct exynos_mixer_ops mixer_ops = { + /* manager */ + .enable_vblank = mixer_enable_vblank, + .disable_vblank = mixer_disable_vblank, + + /* overlay */ + .win_mode_set = mixer_win_mode_set, + .win_commit = mixer_win_commit, + .win_disable = mixer_win_disable, +}; + +/* for pageflip event */ +static void mixer_finish_pageflip(struct drm_device *drm_dev, int crtc) +{ + struct exynos_drm_private *dev_priv = drm_dev->dev_private; + struct drm_pending_vblank_event *e, *t; + struct timeval now; + unsigned long flags; + bool is_checked = false; + + spin_lock_irqsave(&drm_dev->event_lock, flags); + + list_for_each_entry_safe(e, t, &dev_priv->pageflip_event_list, + base.link) { + /* if event's pipe isn't same as crtc then ignore it. */ + if (crtc != e->pipe) + continue; + + is_checked = true; + do_gettimeofday(&now); + e->event.sequence = 0; + e->event.tv_sec = now.tv_sec; + e->event.tv_usec = now.tv_usec; + + list_move_tail(&e->base.link, &e->base.file_priv->event_list); + wake_up_interruptible(&e->base.file_priv->event_wait); + } + + if (is_checked) + /* + * call drm_vblank_put only in case that drm_vblank_get was + * called. + */ + if (atomic_read(&drm_dev->vblank_refcount[crtc]) > 0) + drm_vblank_put(drm_dev, crtc); + + spin_unlock_irqrestore(&drm_dev->event_lock, flags); +} + +static irqreturn_t mixer_irq_handler(int irq, void *arg) +{ + struct exynos_drm_hdmi_context *drm_hdmi_ctx = arg; + struct mixer_context *ctx = drm_hdmi_ctx->ctx; + struct mixer_resources *res = &ctx->mixer_res; + u32 val, val_base; + + spin_lock(&res->reg_slock); + + /* read interrupt status for handling and clearing flags for VSYNC */ + val = mixer_reg_read(res, MXR_INT_STATUS); + + /* handling VSYNC */ + if (val & MXR_INT_STATUS_VSYNC) { + /* interlace scan need to check shadow register */ + if (ctx->interlace) { + val_base = mixer_reg_read(res, MXR_GRAPHIC_BASE_S(0)); + if (ctx->win_data[0].dma_addr != val_base) + goto out; + + val_base = mixer_reg_read(res, MXR_GRAPHIC_BASE_S(1)); + if (ctx->win_data[1].dma_addr != val_base) + goto out; + } + + drm_handle_vblank(drm_hdmi_ctx->drm_dev, ctx->pipe); + mixer_finish_pageflip(drm_hdmi_ctx->drm_dev, ctx->pipe); + } + +out: + /* clear interrupts */ + if (~val & MXR_INT_EN_VSYNC) { + /* vsync interrupt use different bit for read and clear */ + val &= ~MXR_INT_EN_VSYNC; + val |= MXR_INT_CLEAR_VSYNC; + } + mixer_reg_write(res, MXR_INT_STATUS, val); + + spin_unlock(&res->reg_slock); + + return IRQ_HANDLED; +} + +static void mixer_win_reset(struct mixer_context *ctx) +{ + struct mixer_resources *res = &ctx->mixer_res; + unsigned long flags; + u32 val; /* value stored to register */ + + spin_lock_irqsave(&res->reg_slock, flags); + mixer_vsync_set_update(ctx, false); + + mixer_reg_writemask(res, MXR_CFG, MXR_CFG_DST_HDMI, MXR_CFG_DST_MASK); + + /* set output in RGB888 mode */ + mixer_reg_writemask(res, MXR_CFG, MXR_CFG_OUT_RGB888, MXR_CFG_OUT_MASK); + + /* 16 beat burst in DMA */ + mixer_reg_writemask(res, MXR_STATUS, MXR_STATUS_16_BURST, + MXR_STATUS_BURST_MASK); + + /* setting default layer priority: layer1 > layer0 > video + * because typical usage scenario would be + * layer1 - OSD + * layer0 - framebuffer + * video - video overlay + */ + val = MXR_LAYER_CFG_GRP1_VAL(3); + val |= MXR_LAYER_CFG_GRP0_VAL(2); + val |= MXR_LAYER_CFG_VP_VAL(1); + mixer_reg_write(res, MXR_LAYER_CFG, val); + + /* setting background color */ + mixer_reg_write(res, MXR_BG_COLOR0, 0x008080); + mixer_reg_write(res, MXR_BG_COLOR1, 0x008080); + mixer_reg_write(res, MXR_BG_COLOR2, 0x008080); + + /* setting graphical layers */ + + val = MXR_GRP_CFG_COLOR_KEY_DISABLE; /* no blank key */ + val |= MXR_GRP_CFG_WIN_BLEND_EN; + val |= MXR_GRP_CFG_ALPHA_VAL(0xff); /* non-transparent alpha */ + + /* the same configuration for both layers */ + mixer_reg_write(res, MXR_GRAPHIC_CFG(0), val); + + val |= MXR_GRP_CFG_BLEND_PRE_MUL; + val |= MXR_GRP_CFG_PIXEL_BLEND_EN; + mixer_reg_write(res, MXR_GRAPHIC_CFG(1), val); + + /* configuration of Video Processor Registers */ + vp_win_reset(ctx); + vp_default_filter(res); + + /* disable all layers */ + mixer_reg_writemask(res, MXR_CFG, 0, MXR_CFG_GRP0_ENABLE); + mixer_reg_writemask(res, MXR_CFG, 0, MXR_CFG_GRP1_ENABLE); + mixer_reg_writemask(res, MXR_CFG, 0, MXR_CFG_VP_ENABLE); + + mixer_vsync_set_update(ctx, true); + spin_unlock_irqrestore(&res->reg_slock, flags); +} + +static void mixer_resource_poweron(struct mixer_context *ctx) +{ + struct mixer_resources *res = &ctx->mixer_res; + + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + clk_enable(res->mixer); + clk_enable(res->vp); + clk_enable(res->sclk_mixer); + + mixer_win_reset(ctx); +} + +static void mixer_resource_poweroff(struct mixer_context *ctx) +{ + struct mixer_resources *res = &ctx->mixer_res; + + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + + clk_disable(res->mixer); + clk_disable(res->vp); + clk_disable(res->sclk_mixer); +} + +static int mixer_runtime_resume(struct device *dev) +{ + struct exynos_drm_hdmi_context *ctx = get_mixer_context(dev); + + DRM_DEBUG_KMS("resume - start\n"); + + mixer_resource_poweron(ctx->ctx); + + return 0; +} + +static int mixer_runtime_suspend(struct device *dev) +{ + struct exynos_drm_hdmi_context *ctx = get_mixer_context(dev); + + DRM_DEBUG_KMS("suspend - start\n"); + + mixer_resource_poweroff(ctx->ctx); + + return 0; +} + +static const struct dev_pm_ops mixer_pm_ops = { + .runtime_suspend = mixer_runtime_suspend, + .runtime_resume = mixer_runtime_resume, +}; + +static int __devinit mixer_resources_init(struct exynos_drm_hdmi_context *ctx, + struct platform_device *pdev) +{ + struct mixer_context *mixer_ctx = ctx->ctx; + struct device *dev = &pdev->dev; + struct mixer_resources *mixer_res = &mixer_ctx->mixer_res; + struct resource *res; + int ret; + + mixer_res->dev = dev; + spin_lock_init(&mixer_res->reg_slock); + + mixer_res->mixer = clk_get(dev, "mixer"); + if (IS_ERR_OR_NULL(mixer_res->mixer)) { + dev_err(dev, "failed to get clock 'mixer'\n"); + ret = -ENODEV; + goto fail; + } + mixer_res->vp = clk_get(dev, "vp"); + if (IS_ERR_OR_NULL(mixer_res->vp)) { + dev_err(dev, "failed to get clock 'vp'\n"); + ret = -ENODEV; + goto fail; + } + mixer_res->sclk_mixer = clk_get(dev, "sclk_mixer"); + if (IS_ERR_OR_NULL(mixer_res->sclk_mixer)) { + dev_err(dev, "failed to get clock 'sclk_mixer'\n"); + ret = -ENODEV; + goto fail; + } + mixer_res->sclk_hdmi = clk_get(dev, "sclk_hdmi"); + if (IS_ERR_OR_NULL(mixer_res->sclk_hdmi)) { + dev_err(dev, "failed to get clock 'sclk_hdmi'\n"); + ret = -ENODEV; + goto fail; + } + mixer_res->sclk_dac = clk_get(dev, "sclk_dac"); + if (IS_ERR_OR_NULL(mixer_res->sclk_dac)) { + dev_err(dev, "failed to get clock 'sclk_dac'\n"); + ret = -ENODEV; + goto fail; + } + res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "mxr"); + if (res == NULL) { + dev_err(dev, "get memory resource failed.\n"); + ret = -ENXIO; + goto fail; + } + + clk_set_parent(mixer_res->sclk_mixer, mixer_res->sclk_hdmi); + + mixer_res->mixer_regs = ioremap(res->start, resource_size(res)); + if (mixer_res->mixer_regs == NULL) { + dev_err(dev, "register mapping failed.\n"); + ret = -ENXIO; + goto fail; + } + + res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "vp"); + if (res == NULL) { + dev_err(dev, "get memory resource failed.\n"); + ret = -ENXIO; + goto fail_mixer_regs; + } + + mixer_res->vp_regs = ioremap(res->start, resource_size(res)); + if (mixer_res->vp_regs == NULL) { + dev_err(dev, "register mapping failed.\n"); + ret = -ENXIO; + goto fail_mixer_regs; + } + + res = platform_get_resource_byname(pdev, IORESOURCE_IRQ, "irq"); + if (res == NULL) { + dev_err(dev, "get interrupt resource failed.\n"); + ret = -ENXIO; + goto fail_vp_regs; + } + + ret = request_irq(res->start, mixer_irq_handler, 0, "drm_mixer", ctx); + if (ret) { + dev_err(dev, "request interrupt failed.\n"); + goto fail_vp_regs; + } + mixer_res->irq = res->start; + + return 0; + +fail_vp_regs: + iounmap(mixer_res->vp_regs); + +fail_mixer_regs: + iounmap(mixer_res->mixer_regs); + +fail: + if (!IS_ERR_OR_NULL(mixer_res->sclk_dac)) + clk_put(mixer_res->sclk_dac); + if (!IS_ERR_OR_NULL(mixer_res->sclk_hdmi)) + clk_put(mixer_res->sclk_hdmi); + if (!IS_ERR_OR_NULL(mixer_res->sclk_mixer)) + clk_put(mixer_res->sclk_mixer); + if (!IS_ERR_OR_NULL(mixer_res->vp)) + clk_put(mixer_res->vp); + if (!IS_ERR_OR_NULL(mixer_res->mixer)) + clk_put(mixer_res->mixer); + mixer_res->dev = NULL; + return ret; +} + +static void mixer_resources_cleanup(struct mixer_context *ctx) +{ + struct mixer_resources *res = &ctx->mixer_res; + + disable_irq(res->irq); + free_irq(res->irq, ctx); + + iounmap(res->vp_regs); + iounmap(res->mixer_regs); +} + +static int __devinit mixer_probe(struct platform_device *pdev) +{ + struct device *dev = &pdev->dev; + struct exynos_drm_hdmi_context *drm_hdmi_ctx; + struct mixer_context *ctx; + int ret; + + dev_info(dev, "probe start\n"); + + drm_hdmi_ctx = kzalloc(sizeof(*drm_hdmi_ctx), GFP_KERNEL); + if (!drm_hdmi_ctx) { + DRM_ERROR("failed to allocate common hdmi context.\n"); + return -ENOMEM; + } + + ctx = kzalloc(sizeof(*ctx), GFP_KERNEL); + if (!ctx) { + DRM_ERROR("failed to alloc mixer context.\n"); + kfree(drm_hdmi_ctx); + return -ENOMEM; + } + + drm_hdmi_ctx->ctx = (void *)ctx; + + platform_set_drvdata(pdev, drm_hdmi_ctx); + + /* acquire resources: regs, irqs, clocks */ + ret = mixer_resources_init(drm_hdmi_ctx, pdev); + if (ret) + goto fail; + + /* register specific callback point to common hdmi. */ + exynos_mixer_ops_register(&mixer_ops); + + mixer_resource_poweron(ctx); + + return 0; + + +fail: + dev_info(dev, "probe failed\n"); + return ret; +} + +static int mixer_remove(struct platform_device *pdev) +{ + struct device *dev = &pdev->dev; + struct exynos_drm_hdmi_context *drm_hdmi_ctx = + platform_get_drvdata(pdev); + struct mixer_context *ctx = drm_hdmi_ctx->ctx; + + dev_info(dev, "remove successful\n"); + + mixer_resource_poweroff(ctx); + mixer_resources_cleanup(ctx); + + return 0; +} + +struct platform_driver mixer_driver = { + .driver = { + .name = "s5p-mixer", + .owner = THIS_MODULE, + .pm = &mixer_pm_ops, + }, + .probe = mixer_probe, + .remove = __devexit_p(mixer_remove), +}; diff --git a/drivers/gpu/drm/exynos/regs-hdmi.h b/drivers/gpu/drm/exynos/regs-hdmi.h new file mode 100644 index 0000000..3c04bea --- /dev/null +++ b/drivers/gpu/drm/exynos/regs-hdmi.h @@ -0,0 +1,561 @@ +/* + * + * Cloned from drivers/media/video/s5p-tv/regs-hdmi.h + * + * Copyright (c) 2010-2011 Samsung Electronics Co., Ltd. + * http://www.samsung.com/ + * + * HDMI register header file for Samsung TVOUT driver + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. +*/ + +#ifndef SAMSUNG_REGS_HDMI_H +#define SAMSUNG_REGS_HDMI_H + +/* + * Register part +*/ + +/* HDMI Version 1.3 & Common */ +#define HDMI_CTRL_BASE(x) ((x) + 0x00000000) +#define HDMI_CORE_BASE(x) ((x) + 0x00010000) +#define HDMI_I2S_BASE(x) ((x) + 0x00040000) +#define HDMI_TG_BASE(x) ((x) + 0x00050000) + +/* Control registers */ +#define HDMI_INTC_CON HDMI_CTRL_BASE(0x0000) +#define HDMI_INTC_FLAG HDMI_CTRL_BASE(0x0004) +#define HDMI_HPD_STATUS HDMI_CTRL_BASE(0x000C) +#define HDMI_V13_PHY_RSTOUT HDMI_CTRL_BASE(0x0014) +#define HDMI_V13_PHY_VPLL HDMI_CTRL_BASE(0x0018) +#define HDMI_V13_PHY_CMU HDMI_CTRL_BASE(0x001C) +#define HDMI_V13_CORE_RSTOUT HDMI_CTRL_BASE(0x0020) + +/* Core registers */ +#define HDMI_CON_0 HDMI_CORE_BASE(0x0000) +#define HDMI_CON_1 HDMI_CORE_BASE(0x0004) +#define HDMI_CON_2 HDMI_CORE_BASE(0x0008) +#define HDMI_SYS_STATUS HDMI_CORE_BASE(0x0010) +#define HDMI_V13_PHY_STATUS HDMI_CORE_BASE(0x0014) +#define HDMI_STATUS_EN HDMI_CORE_BASE(0x0020) +#define HDMI_HPD HDMI_CORE_BASE(0x0030) +#define HDMI_MODE_SEL HDMI_CORE_BASE(0x0040) +#define HDMI_ENC_EN HDMI_CORE_BASE(0x0044) +#define HDMI_V13_BLUE_SCREEN_0 HDMI_CORE_BASE(0x0050) +#define HDMI_V13_BLUE_SCREEN_1 HDMI_CORE_BASE(0x0054) +#define HDMI_V13_BLUE_SCREEN_2 HDMI_CORE_BASE(0x0058) +#define HDMI_H_BLANK_0 HDMI_CORE_BASE(0x00A0) +#define HDMI_H_BLANK_1 HDMI_CORE_BASE(0x00A4) +#define HDMI_V13_V_BLANK_0 HDMI_CORE_BASE(0x00B0) +#define HDMI_V13_V_BLANK_1 HDMI_CORE_BASE(0x00B4) +#define HDMI_V13_V_BLANK_2 HDMI_CORE_BASE(0x00B8) +#define HDMI_V13_H_V_LINE_0 HDMI_CORE_BASE(0x00C0) +#define HDMI_V13_H_V_LINE_1 HDMI_CORE_BASE(0x00C4) +#define HDMI_V13_H_V_LINE_2 HDMI_CORE_BASE(0x00C8) +#define HDMI_VSYNC_POL HDMI_CORE_BASE(0x00E4) +#define HDMI_INT_PRO_MODE HDMI_CORE_BASE(0x00E8) +#define HDMI_V13_V_BLANK_F_0 HDMI_CORE_BASE(0x0110) +#define HDMI_V13_V_BLANK_F_1 HDMI_CORE_BASE(0x0114) +#define HDMI_V13_V_BLANK_F_2 HDMI_CORE_BASE(0x0118) +#define HDMI_V13_H_SYNC_GEN_0 HDMI_CORE_BASE(0x0120) +#define HDMI_V13_H_SYNC_GEN_1 HDMI_CORE_BASE(0x0124) +#define HDMI_V13_H_SYNC_GEN_2 HDMI_CORE_BASE(0x0128) +#define HDMI_V13_V_SYNC_GEN_1_0 HDMI_CORE_BASE(0x0130) +#define HDMI_V13_V_SYNC_GEN_1_1 HDMI_CORE_BASE(0x0134) +#define HDMI_V13_V_SYNC_GEN_1_2 HDMI_CORE_BASE(0x0138) +#define HDMI_V13_V_SYNC_GEN_2_0 HDMI_CORE_BASE(0x0140) +#define HDMI_V13_V_SYNC_GEN_2_1 HDMI_CORE_BASE(0x0144) +#define HDMI_V13_V_SYNC_GEN_2_2 HDMI_CORE_BASE(0x0148) +#define HDMI_V13_V_SYNC_GEN_3_0 HDMI_CORE_BASE(0x0150) +#define HDMI_V13_V_SYNC_GEN_3_1 HDMI_CORE_BASE(0x0154) +#define HDMI_V13_V_SYNC_GEN_3_2 HDMI_CORE_BASE(0x0158) +#define HDMI_V13_ACR_CON HDMI_CORE_BASE(0x0180) +#define HDMI_V13_AVI_CON HDMI_CORE_BASE(0x0300) +#define HDMI_V13_AVI_BYTE(n) HDMI_CORE_BASE(0x0320 + 4 * (n)) +#define HDMI_V13_DC_CONTROL HDMI_CORE_BASE(0x05C0) +#define HDMI_V13_VIDEO_PATTERN_GEN HDMI_CORE_BASE(0x05C4) +#define HDMI_V13_HPD_GEN HDMI_CORE_BASE(0x05C8) +#define HDMI_V13_AUI_CON HDMI_CORE_BASE(0x0360) +#define HDMI_V13_SPD_CON HDMI_CORE_BASE(0x0400) + +/* Timing generator registers */ +#define HDMI_TG_CMD HDMI_TG_BASE(0x0000) +#define HDMI_TG_H_FSZ_L HDMI_TG_BASE(0x0018) +#define HDMI_TG_H_FSZ_H HDMI_TG_BASE(0x001C) +#define HDMI_TG_HACT_ST_L HDMI_TG_BASE(0x0020) +#define HDMI_TG_HACT_ST_H HDMI_TG_BASE(0x0024) +#define HDMI_TG_HACT_SZ_L HDMI_TG_BASE(0x0028) +#define HDMI_TG_HACT_SZ_H HDMI_TG_BASE(0x002C) +#define HDMI_TG_V_FSZ_L HDMI_TG_BASE(0x0030) +#define HDMI_TG_V_FSZ_H HDMI_TG_BASE(0x0034) +#define HDMI_TG_VSYNC_L HDMI_TG_BASE(0x0038) +#define HDMI_TG_VSYNC_H HDMI_TG_BASE(0x003C) +#define HDMI_TG_VSYNC2_L HDMI_TG_BASE(0x0040) +#define HDMI_TG_VSYNC2_H HDMI_TG_BASE(0x0044) +#define HDMI_TG_VACT_ST_L HDMI_TG_BASE(0x0048) +#define HDMI_TG_VACT_ST_H HDMI_TG_BASE(0x004C) +#define HDMI_TG_VACT_SZ_L HDMI_TG_BASE(0x0050) +#define HDMI_TG_VACT_SZ_H HDMI_TG_BASE(0x0054) +#define HDMI_TG_FIELD_CHG_L HDMI_TG_BASE(0x0058) +#define HDMI_TG_FIELD_CHG_H HDMI_TG_BASE(0x005C) +#define HDMI_TG_VACT_ST2_L HDMI_TG_BASE(0x0060) +#define HDMI_TG_VACT_ST2_H HDMI_TG_BASE(0x0064) +#define HDMI_TG_VSYNC_TOP_HDMI_L HDMI_TG_BASE(0x0078) +#define HDMI_TG_VSYNC_TOP_HDMI_H HDMI_TG_BASE(0x007C) +#define HDMI_TG_VSYNC_BOT_HDMI_L HDMI_TG_BASE(0x0080) +#define HDMI_TG_VSYNC_BOT_HDMI_H HDMI_TG_BASE(0x0084) +#define HDMI_TG_FIELD_TOP_HDMI_L HDMI_TG_BASE(0x0088) +#define HDMI_TG_FIELD_TOP_HDMI_H HDMI_TG_BASE(0x008C) +#define HDMI_TG_FIELD_BOT_HDMI_L HDMI_TG_BASE(0x0090) +#define HDMI_TG_FIELD_BOT_HDMI_H HDMI_TG_BASE(0x0094) + +/* + * Bit definition part + */ + +/* HDMI_INTC_CON */ +#define HDMI_INTC_EN_GLOBAL (1 << 6) +#define HDMI_INTC_EN_HPD_PLUG (1 << 3) +#define HDMI_INTC_EN_HPD_UNPLUG (1 << 2) + +/* HDMI_INTC_FLAG */ +#define HDMI_INTC_FLAG_HPD_PLUG (1 << 3) +#define HDMI_INTC_FLAG_HPD_UNPLUG (1 << 2) + +/* HDMI_PHY_RSTOUT */ +#define HDMI_PHY_SW_RSTOUT (1 << 0) + +/* HDMI_CORE_RSTOUT */ +#define HDMI_CORE_SW_RSTOUT (1 << 0) + +/* HDMI_CON_0 */ +#define HDMI_BLUE_SCR_EN (1 << 5) +#define HDMI_ASP_EN (1 << 2) +#define HDMI_ASP_DIS (0 << 2) +#define HDMI_ASP_MASK (1 << 2) +#define HDMI_EN (1 << 0) + +/* HDMI_PHY_STATUS */ +#define HDMI_PHY_STATUS_READY (1 << 0) + +/* HDMI_MODE_SEL */ +#define HDMI_MODE_HDMI_EN (1 << 1) +#define HDMI_MODE_DVI_EN (1 << 0) +#define HDMI_DVI_MODE_EN (1) +#define HDMI_DVI_MODE_DIS (0) +#define HDMI_MODE_MASK (3 << 0) + +/* HDMI_TG_CMD */ +#define HDMI_TG_EN (1 << 0) +#define HDMI_FIELD_EN (1 << 1) + + +/* HDMI Version 1.4 */ +/* Control registers */ +/* #define HDMI_INTC_CON HDMI_CTRL_BASE(0x0000) */ +/* #define HDMI_INTC_FLAG HDMI_CTRL_BASE(0x0004) */ +#define HDMI_HDCP_KEY_LOAD HDMI_CTRL_BASE(0x0008) +/* #define HDMI_HPD_STATUS HDMI_CTRL_BASE(0x000C) */ +#define HDMI_INTC_CON_1 HDMI_CTRL_BASE(0x0010) +#define HDMI_INTC_FLAG_1 HDMI_CTRL_BASE(0x0014) +#define HDMI_PHY_STATUS_0 HDMI_CTRL_BASE(0x0020) +#define HDMI_PHY_STATUS_CMU HDMI_CTRL_BASE(0x0024) +#define HDMI_PHY_STATUS_PLL HDMI_CTRL_BASE(0x0028) +#define HDMI_PHY_CON_0 HDMI_CTRL_BASE(0x0030) +#define HDMI_HPD_CTRL HDMI_CTRL_BASE(0x0040) +#define HDMI_HPD_ST HDMI_CTRL_BASE(0x0044) +#define HDMI_HPD_TH_X HDMI_CTRL_BASE(0x0050) +#define HDMI_AUDIO_CLKSEL HDMI_CTRL_BASE(0x0070) +#define HDMI_PHY_RSTOUT HDMI_CTRL_BASE(0x0074) +#define HDMI_PHY_VPLL HDMI_CTRL_BASE(0x0078) +#define HDMI_PHY_CMU HDMI_CTRL_BASE(0x007C) +#define HDMI_CORE_RSTOUT HDMI_CTRL_BASE(0x0080) + +/* Video related registers */ +#define HDMI_YMAX HDMI_CORE_BASE(0x0060) +#define HDMI_YMIN HDMI_CORE_BASE(0x0064) +#define HDMI_CMAX HDMI_CORE_BASE(0x0068) +#define HDMI_CMIN HDMI_CORE_BASE(0x006C) + +#define HDMI_V2_BLANK_0 HDMI_CORE_BASE(0x00B0) +#define HDMI_V2_BLANK_1 HDMI_CORE_BASE(0x00B4) +#define HDMI_V1_BLANK_0 HDMI_CORE_BASE(0x00B8) +#define HDMI_V1_BLANK_1 HDMI_CORE_BASE(0x00BC) + +#define HDMI_V_LINE_0 HDMI_CORE_BASE(0x00C0) +#define HDMI_V_LINE_1 HDMI_CORE_BASE(0x00C4) +#define HDMI_H_LINE_0 HDMI_CORE_BASE(0x00C8) +#define HDMI_H_LINE_1 HDMI_CORE_BASE(0x00CC) + +#define HDMI_HSYNC_POL HDMI_CORE_BASE(0x00E0) + +#define HDMI_V_BLANK_F0_0 HDMI_CORE_BASE(0x0110) +#define HDMI_V_BLANK_F0_1 HDMI_CORE_BASE(0x0114) +#define HDMI_V_BLANK_F1_0 HDMI_CORE_BASE(0x0118) +#define HDMI_V_BLANK_F1_1 HDMI_CORE_BASE(0x011C) + +#define HDMI_H_SYNC_START_0 HDMI_CORE_BASE(0x0120) +#define HDMI_H_SYNC_START_1 HDMI_CORE_BASE(0x0124) +#define HDMI_H_SYNC_END_0 HDMI_CORE_BASE(0x0128) +#define HDMI_H_SYNC_END_1 HDMI_CORE_BASE(0x012C) + +#define HDMI_V_SYNC_LINE_BEF_2_0 HDMI_CORE_BASE(0x0130) +#define HDMI_V_SYNC_LINE_BEF_2_1 HDMI_CORE_BASE(0x0134) +#define HDMI_V_SYNC_LINE_BEF_1_0 HDMI_CORE_BASE(0x0138) +#define HDMI_V_SYNC_LINE_BEF_1_1 HDMI_CORE_BASE(0x013C) + +#define HDMI_V_SYNC_LINE_AFT_2_0 HDMI_CORE_BASE(0x0140) +#define HDMI_V_SYNC_LINE_AFT_2_1 HDMI_CORE_BASE(0x0144) +#define HDMI_V_SYNC_LINE_AFT_1_0 HDMI_CORE_BASE(0x0148) +#define HDMI_V_SYNC_LINE_AFT_1_1 HDMI_CORE_BASE(0x014C) + +#define HDMI_V_SYNC_LINE_AFT_PXL_2_0 HDMI_CORE_BASE(0x0150) +#define HDMI_V_SYNC_LINE_AFT_PXL_2_1 HDMI_CORE_BASE(0x0154) +#define HDMI_V_SYNC_LINE_AFT_PXL_1_0 HDMI_CORE_BASE(0x0158) +#define HDMI_V_SYNC_LINE_AFT_PXL_1_1 HDMI_CORE_BASE(0x015C) + +#define HDMI_V_BLANK_F2_0 HDMI_CORE_BASE(0x0160) +#define HDMI_V_BLANK_F2_1 HDMI_CORE_BASE(0x0164) +#define HDMI_V_BLANK_F3_0 HDMI_CORE_BASE(0x0168) +#define HDMI_V_BLANK_F3_1 HDMI_CORE_BASE(0x016C) +#define HDMI_V_BLANK_F4_0 HDMI_CORE_BASE(0x0170) +#define HDMI_V_BLANK_F4_1 HDMI_CORE_BASE(0x0174) +#define HDMI_V_BLANK_F5_0 HDMI_CORE_BASE(0x0178) +#define HDMI_V_BLANK_F5_1 HDMI_CORE_BASE(0x017C) + +#define HDMI_V_SYNC_LINE_AFT_3_0 HDMI_CORE_BASE(0x0180) +#define HDMI_V_SYNC_LINE_AFT_3_1 HDMI_CORE_BASE(0x0184) +#define HDMI_V_SYNC_LINE_AFT_4_0 HDMI_CORE_BASE(0x0188) +#define HDMI_V_SYNC_LINE_AFT_4_1 HDMI_CORE_BASE(0x018C) +#define HDMI_V_SYNC_LINE_AFT_5_0 HDMI_CORE_BASE(0x0190) +#define HDMI_V_SYNC_LINE_AFT_5_1 HDMI_CORE_BASE(0x0194) +#define HDMI_V_SYNC_LINE_AFT_6_0 HDMI_CORE_BASE(0x0198) +#define HDMI_V_SYNC_LINE_AFT_6_1 HDMI_CORE_BASE(0x019C) + +#define HDMI_V_SYNC_LINE_AFT_PXL_3_0 HDMI_CORE_BASE(0x01A0) +#define HDMI_V_SYNC_LINE_AFT_PXL_3_1 HDMI_CORE_BASE(0x01A4) +#define HDMI_V_SYNC_LINE_AFT_PXL_4_0 HDMI_CORE_BASE(0x01A8) +#define HDMI_V_SYNC_LINE_AFT_PXL_4_1 HDMI_CORE_BASE(0x01AC) +#define HDMI_V_SYNC_LINE_AFT_PXL_5_0 HDMI_CORE_BASE(0x01B0) +#define HDMI_V_SYNC_LINE_AFT_PXL_5_1 HDMI_CORE_BASE(0x01B4) +#define HDMI_V_SYNC_LINE_AFT_PXL_6_0 HDMI_CORE_BASE(0x01B8) +#define HDMI_V_SYNC_LINE_AFT_PXL_6_1 HDMI_CORE_BASE(0x01BC) + +#define HDMI_VACT_SPACE_1_0 HDMI_CORE_BASE(0x01C0) +#define HDMI_VACT_SPACE_1_1 HDMI_CORE_BASE(0x01C4) +#define HDMI_VACT_SPACE_2_0 HDMI_CORE_BASE(0x01C8) +#define HDMI_VACT_SPACE_2_1 HDMI_CORE_BASE(0x01CC) +#define HDMI_VACT_SPACE_3_0 HDMI_CORE_BASE(0x01D0) +#define HDMI_VACT_SPACE_3_1 HDMI_CORE_BASE(0x01D4) +#define HDMI_VACT_SPACE_4_0 HDMI_CORE_BASE(0x01D8) +#define HDMI_VACT_SPACE_4_1 HDMI_CORE_BASE(0x01DC) +#define HDMI_VACT_SPACE_5_0 HDMI_CORE_BASE(0x01E0) +#define HDMI_VACT_SPACE_5_1 HDMI_CORE_BASE(0x01E4) +#define HDMI_VACT_SPACE_6_0 HDMI_CORE_BASE(0x01E8) +#define HDMI_VACT_SPACE_6_1 HDMI_CORE_BASE(0x01EC) + +#define HDMI_GCP_CON HDMI_CORE_BASE(0x0200) +#define HDMI_GCP_BYTE1 HDMI_CORE_BASE(0x0210) +#define HDMI_GCP_BYTE2 HDMI_CORE_BASE(0x0214) +#define HDMI_GCP_BYTE3 HDMI_CORE_BASE(0x0218) + +/* Audio related registers */ +#define HDMI_ASP_CON HDMI_CORE_BASE(0x0300) +#define HDMI_ASP_SP_FLAT HDMI_CORE_BASE(0x0304) +#define HDMI_ASP_CHCFG0 HDMI_CORE_BASE(0x0310) +#define HDMI_ASP_CHCFG1 HDMI_CORE_BASE(0x0314) +#define HDMI_ASP_CHCFG2 HDMI_CORE_BASE(0x0318) +#define HDMI_ASP_CHCFG3 HDMI_CORE_BASE(0x031C) + +#define HDMI_ACR_CON HDMI_CORE_BASE(0x0400) +#define HDMI_ACR_MCTS0 HDMI_CORE_BASE(0x0410) +#define HDMI_ACR_MCTS1 HDMI_CORE_BASE(0x0414) +#define HDMI_ACR_MCTS2 HDMI_CORE_BASE(0x0418) +#define HDMI_ACR_CTS0 HDMI_CORE_BASE(0x0420) +#define HDMI_ACR_CTS1 HDMI_CORE_BASE(0x0424) +#define HDMI_ACR_CTS2 HDMI_CORE_BASE(0x0428) +#define HDMI_ACR_N0 HDMI_CORE_BASE(0x0430) +#define HDMI_ACR_N1 HDMI_CORE_BASE(0x0434) +#define HDMI_ACR_N2 HDMI_CORE_BASE(0x0438) + +/* Packet related registers */ +#define HDMI_ACP_CON HDMI_CORE_BASE(0x0500) +#define HDMI_ACP_TYPE HDMI_CORE_BASE(0x0514) +#define HDMI_ACP_DATA(n) HDMI_CORE_BASE(0x0520 + 4 * (n)) + +#define HDMI_ISRC_CON HDMI_CORE_BASE(0x0600) +#define HDMI_ISRC1_HEADER1 HDMI_CORE_BASE(0x0614) +#define HDMI_ISRC1_DATA(n) HDMI_CORE_BASE(0x0620 + 4 * (n)) +#define HDMI_ISRC2_DATA(n) HDMI_CORE_BASE(0x06A0 + 4 * (n)) + +#define HDMI_AVI_CON HDMI_CORE_BASE(0x0700) +#define HDMI_AVI_HEADER0 HDMI_CORE_BASE(0x0710) +#define HDMI_AVI_HEADER1 HDMI_CORE_BASE(0x0714) +#define HDMI_AVI_HEADER2 HDMI_CORE_BASE(0x0718) +#define HDMI_AVI_CHECK_SUM HDMI_CORE_BASE(0x071C) +#define HDMI_AVI_BYTE(n) HDMI_CORE_BASE(0x0720 + 4 * (n)) + +#define HDMI_AUI_CON HDMI_CORE_BASE(0x0800) +#define HDMI_AUI_HEADER0 HDMI_CORE_BASE(0x0810) +#define HDMI_AUI_HEADER1 HDMI_CORE_BASE(0x0814) +#define HDMI_AUI_HEADER2 HDMI_CORE_BASE(0x0818) +#define HDMI_AUI_CHECK_SUM HDMI_CORE_BASE(0x081C) +#define HDMI_AUI_BYTE(n) HDMI_CORE_BASE(0x0820 + 4 * (n)) + +#define HDMI_MPG_CON HDMI_CORE_BASE(0x0900) +#define HDMI_MPG_CHECK_SUM HDMI_CORE_BASE(0x091C) +#define HDMI_MPG_DATA(n) HDMI_CORE_BASE(0x0920 + 4 * (n)) + +#define HDMI_SPD_CON HDMI_CORE_BASE(0x0A00) +#define HDMI_SPD_HEADER0 HDMI_CORE_BASE(0x0A10) +#define HDMI_SPD_HEADER1 HDMI_CORE_BASE(0x0A14) +#define HDMI_SPD_HEADER2 HDMI_CORE_BASE(0x0A18) +#define HDMI_SPD_DATA(n) HDMI_CORE_BASE(0x0A20 + 4 * (n)) + +#define HDMI_GAMUT_CON HDMI_CORE_BASE(0x0B00) +#define HDMI_GAMUT_HEADER0 HDMI_CORE_BASE(0x0B10) +#define HDMI_GAMUT_HEADER1 HDMI_CORE_BASE(0x0B14) +#define HDMI_GAMUT_HEADER2 HDMI_CORE_BASE(0x0B18) +#define HDMI_GAMUT_METADATA(n) HDMI_CORE_BASE(0x0B20 + 4 * (n)) + +#define HDMI_VSI_CON HDMI_CORE_BASE(0x0C00) +#define HDMI_VSI_HEADER0 HDMI_CORE_BASE(0x0C10) +#define HDMI_VSI_HEADER1 HDMI_CORE_BASE(0x0C14) +#define HDMI_VSI_HEADER2 HDMI_CORE_BASE(0x0C18) +#define HDMI_VSI_DATA(n) HDMI_CORE_BASE(0x0C20 + 4 * (n)) + +#define HDMI_DC_CONTROL HDMI_CORE_BASE(0x0D00) +#define HDMI_VIDEO_PATTERN_GEN HDMI_CORE_BASE(0x0D04) + +#define HDMI_AN_SEED_SEL HDMI_CORE_BASE(0x0E48) +#define HDMI_AN_SEED_0 HDMI_CORE_BASE(0x0E58) +#define HDMI_AN_SEED_1 HDMI_CORE_BASE(0x0E5C) +#define HDMI_AN_SEED_2 HDMI_CORE_BASE(0x0E60) +#define HDMI_AN_SEED_3 HDMI_CORE_BASE(0x0E64) + +/* HDCP related registers */ +#define HDMI_HDCP_SHA1(n) HDMI_CORE_BASE(0x7000 + 4 * (n)) +#define HDMI_HDCP_KSV_LIST(n) HDMI_CORE_BASE(0x7050 + 4 * (n)) + +#define HDMI_HDCP_KSV_LIST_CON HDMI_CORE_BASE(0x7064) +#define HDMI_HDCP_SHA_RESULT HDMI_CORE_BASE(0x7070) +#define HDMI_HDCP_CTRL1 HDMI_CORE_BASE(0x7080) +#define HDMI_HDCP_CTRL2 HDMI_CORE_BASE(0x7084) +#define HDMI_HDCP_CHECK_RESULT HDMI_CORE_BASE(0x7090) +#define HDMI_HDCP_BKSV(n) HDMI_CORE_BASE(0x70A0 + 4 * (n)) +#define HDMI_HDCP_AKSV(n) HDMI_CORE_BASE(0x70C0 + 4 * (n)) +#define HDMI_HDCP_AN(n) HDMI_CORE_BASE(0x70E0 + 4 * (n)) + +#define HDMI_HDCP_BCAPS HDMI_CORE_BASE(0x7100) +#define HDMI_HDCP_BSTATUS_0 HDMI_CORE_BASE(0x7110) +#define HDMI_HDCP_BSTATUS_1 HDMI_CORE_BASE(0x7114) +#define HDMI_HDCP_RI_0 HDMI_CORE_BASE(0x7140) +#define HDMI_HDCP_RI_1 HDMI_CORE_BASE(0x7144) +#define HDMI_HDCP_I2C_INT HDMI_CORE_BASE(0x7180) +#define HDMI_HDCP_AN_INT HDMI_CORE_BASE(0x7190) +#define HDMI_HDCP_WDT_INT HDMI_CORE_BASE(0x71A0) +#define HDMI_HDCP_RI_INT HDMI_CORE_BASE(0x71B0) +#define HDMI_HDCP_RI_COMPARE_0 HDMI_CORE_BASE(0x71D0) +#define HDMI_HDCP_RI_COMPARE_1 HDMI_CORE_BASE(0x71D4) +#define HDMI_HDCP_FRAME_COUNT HDMI_CORE_BASE(0x71E0) + +#define HDMI_RGB_ROUND_EN HDMI_CORE_BASE(0xD500) +#define HDMI_VACT_SPACE_R_0 HDMI_CORE_BASE(0xD504) +#define HDMI_VACT_SPACE_R_1 HDMI_CORE_BASE(0xD508) +#define HDMI_VACT_SPACE_G_0 HDMI_CORE_BASE(0xD50C) +#define HDMI_VACT_SPACE_G_1 HDMI_CORE_BASE(0xD510) +#define HDMI_VACT_SPACE_B_0 HDMI_CORE_BASE(0xD514) +#define HDMI_VACT_SPACE_B_1 HDMI_CORE_BASE(0xD518) + +#define HDMI_BLUE_SCREEN_B_0 HDMI_CORE_BASE(0xD520) +#define HDMI_BLUE_SCREEN_B_1 HDMI_CORE_BASE(0xD524) +#define HDMI_BLUE_SCREEN_G_0 HDMI_CORE_BASE(0xD528) +#define HDMI_BLUE_SCREEN_G_1 HDMI_CORE_BASE(0xD52C) +#define HDMI_BLUE_SCREEN_R_0 HDMI_CORE_BASE(0xD530) +#define HDMI_BLUE_SCREEN_R_1 HDMI_CORE_BASE(0xD534) + +/* HDMI I2S register */ +#define HDMI_I2S_CLK_CON HDMI_I2S_BASE(0x000) +#define HDMI_I2S_CON_1 HDMI_I2S_BASE(0x004) +#define HDMI_I2S_CON_2 HDMI_I2S_BASE(0x008) +#define HDMI_I2S_PIN_SEL_0 HDMI_I2S_BASE(0x00c) +#define HDMI_I2S_PIN_SEL_1 HDMI_I2S_BASE(0x010) +#define HDMI_I2S_PIN_SEL_2 HDMI_I2S_BASE(0x014) +#define HDMI_I2S_PIN_SEL_3 HDMI_I2S_BASE(0x018) +#define HDMI_I2S_DSD_CON HDMI_I2S_BASE(0x01c) +#define HDMI_I2S_MUX_CON HDMI_I2S_BASE(0x020) +#define HDMI_I2S_CH_ST_CON HDMI_I2S_BASE(0x024) +#define HDMI_I2S_CH_ST_0 HDMI_I2S_BASE(0x028) +#define HDMI_I2S_CH_ST_1 HDMI_I2S_BASE(0x02c) +#define HDMI_I2S_CH_ST_2 HDMI_I2S_BASE(0x030) +#define HDMI_I2S_CH_ST_3 HDMI_I2S_BASE(0x034) +#define HDMI_I2S_CH_ST_4 HDMI_I2S_BASE(0x038) +#define HDMI_I2S_CH_ST_SH_0 HDMI_I2S_BASE(0x03c) +#define HDMI_I2S_CH_ST_SH_1 HDMI_I2S_BASE(0x040) +#define HDMI_I2S_CH_ST_SH_2 HDMI_I2S_BASE(0x044) +#define HDMI_I2S_CH_ST_SH_3 HDMI_I2S_BASE(0x048) +#define HDMI_I2S_CH_ST_SH_4 HDMI_I2S_BASE(0x04c) +#define HDMI_I2S_MUX_CH HDMI_I2S_BASE(0x054) +#define HDMI_I2S_MUX_CUV HDMI_I2S_BASE(0x058) + +/* I2S bit definition */ + +/* I2S_CLK_CON */ +#define HDMI_I2S_CLK_DIS (0) +#define HDMI_I2S_CLK_EN (1) + +/* I2S_CON_1 */ +#define HDMI_I2S_SCLK_FALLING_EDGE (0 << 1) +#define HDMI_I2S_SCLK_RISING_EDGE (1 << 1) +#define HDMI_I2S_L_CH_LOW_POL (0) +#define HDMI_I2S_L_CH_HIGH_POL (1) + +/* I2S_CON_2 */ +#define HDMI_I2S_MSB_FIRST_MODE (0 << 6) +#define HDMI_I2S_LSB_FIRST_MODE (1 << 6) +#define HDMI_I2S_BIT_CH_32FS (0 << 4) +#define HDMI_I2S_BIT_CH_48FS (1 << 4) +#define HDMI_I2S_BIT_CH_RESERVED (2 << 4) +#define HDMI_I2S_SDATA_16BIT (1 << 2) +#define HDMI_I2S_SDATA_20BIT (2 << 2) +#define HDMI_I2S_SDATA_24BIT (3 << 2) +#define HDMI_I2S_BASIC_FORMAT (0) +#define HDMI_I2S_L_JUST_FORMAT (2) +#define HDMI_I2S_R_JUST_FORMAT (3) +#define HDMI_I2S_CON_2_CLR (~(0xFF)) +#define HDMI_I2S_SET_BIT_CH(x) (((x) & 0x7) << 4) +#define HDMI_I2S_SET_SDATA_BIT(x) (((x) & 0x7) << 2) + +/* I2S_PIN_SEL_0 */ +#define HDMI_I2S_SEL_SCLK(x) (((x) & 0x7) << 4) +#define HDMI_I2S_SEL_LRCK(x) ((x) & 0x7) + +/* I2S_PIN_SEL_1 */ +#define HDMI_I2S_SEL_SDATA1(x) (((x) & 0x7) << 4) +#define HDMI_I2S_SEL_SDATA2(x) ((x) & 0x7) + +/* I2S_PIN_SEL_2 */ +#define HDMI_I2S_SEL_SDATA3(x) (((x) & 0x7) << 4) +#define HDMI_I2S_SEL_SDATA2(x) ((x) & 0x7) + +/* I2S_PIN_SEL_3 */ +#define HDMI_I2S_SEL_DSD(x) ((x) & 0x7) + +/* I2S_DSD_CON */ +#define HDMI_I2S_DSD_CLK_RI_EDGE (1 << 1) +#define HDMI_I2S_DSD_CLK_FA_EDGE (0 << 1) +#define HDMI_I2S_DSD_ENABLE (1) +#define HDMI_I2S_DSD_DISABLE (0) + +/* I2S_MUX_CON */ +#define HDMI_I2S_NOISE_FILTER_ZERO (0 << 5) +#define HDMI_I2S_NOISE_FILTER_2_STAGE (1 << 5) +#define HDMI_I2S_NOISE_FILTER_3_STAGE (2 << 5) +#define HDMI_I2S_NOISE_FILTER_4_STAGE (3 << 5) +#define HDMI_I2S_NOISE_FILTER_5_STAGE (4 << 5) +#define HDMI_I2S_IN_DISABLE (1 << 4) +#define HDMI_I2S_IN_ENABLE (0 << 4) +#define HDMI_I2S_AUD_SPDIF (0 << 2) +#define HDMI_I2S_AUD_I2S (1 << 2) +#define HDMI_I2S_AUD_DSD (2 << 2) +#define HDMI_I2S_CUV_SPDIF_ENABLE (0 << 1) +#define HDMI_I2S_CUV_I2S_ENABLE (1 << 1) +#define HDMI_I2S_MUX_DISABLE (0) +#define HDMI_I2S_MUX_ENABLE (1) +#define HDMI_I2S_MUX_CON_CLR (~(0xFF)) + +/* I2S_CH_ST_CON */ +#define HDMI_I2S_CH_STATUS_RELOAD (1) +#define HDMI_I2S_CH_ST_CON_CLR (~(1)) + +/* I2S_CH_ST_0 / I2S_CH_ST_SH_0 */ +#define HDMI_I2S_CH_STATUS_MODE_0 (0 << 6) +#define HDMI_I2S_2AUD_CH_WITHOUT_PREEMPH (0 << 3) +#define HDMI_I2S_2AUD_CH_WITH_PREEMPH (1 << 3) +#define HDMI_I2S_DEFAULT_EMPHASIS (0 << 3) +#define HDMI_I2S_COPYRIGHT (0 << 2) +#define HDMI_I2S_NO_COPYRIGHT (1 << 2) +#define HDMI_I2S_LINEAR_PCM (0 << 1) +#define HDMI_I2S_NO_LINEAR_PCM (1 << 1) +#define HDMI_I2S_CONSUMER_FORMAT (0) +#define HDMI_I2S_PROF_FORMAT (1) +#define HDMI_I2S_CH_ST_0_CLR (~(0xFF)) + +/* I2S_CH_ST_1 / I2S_CH_ST_SH_1 */ +#define HDMI_I2S_CD_PLAYER (0x00) +#define HDMI_I2S_DAT_PLAYER (0x03) +#define HDMI_I2S_DCC_PLAYER (0x43) +#define HDMI_I2S_MINI_DISC_PLAYER (0x49) + +/* I2S_CH_ST_2 / I2S_CH_ST_SH_2 */ +#define HDMI_I2S_CHANNEL_NUM_MASK (0xF << 4) +#define HDMI_I2S_SOURCE_NUM_MASK (0xF) +#define HDMI_I2S_SET_CHANNEL_NUM(x) (((x) & (0xF)) << 4) +#define HDMI_I2S_SET_SOURCE_NUM(x) ((x) & (0xF)) + +/* I2S_CH_ST_3 / I2S_CH_ST_SH_3 */ +#define HDMI_I2S_CLK_ACCUR_LEVEL_1 (1 << 4) +#define HDMI_I2S_CLK_ACCUR_LEVEL_2 (0 << 4) +#define HDMI_I2S_CLK_ACCUR_LEVEL_3 (2 << 4) +#define HDMI_I2S_SMP_FREQ_44_1 (0x0) +#define HDMI_I2S_SMP_FREQ_48 (0x2) +#define HDMI_I2S_SMP_FREQ_32 (0x3) +#define HDMI_I2S_SMP_FREQ_96 (0xA) +#define HDMI_I2S_SET_SMP_FREQ(x) ((x) & (0xF)) + +/* I2S_CH_ST_4 / I2S_CH_ST_SH_4 */ +#define HDMI_I2S_ORG_SMP_FREQ_44_1 (0xF << 4) +#define HDMI_I2S_ORG_SMP_FREQ_88_2 (0x7 << 4) +#define HDMI_I2S_ORG_SMP_FREQ_22_05 (0xB << 4) +#define HDMI_I2S_ORG_SMP_FREQ_176_4 (0x3 << 4) +#define HDMI_I2S_WORD_LEN_NOT_DEFINE (0x0 << 1) +#define HDMI_I2S_WORD_LEN_MAX24_20BITS (0x1 << 1) +#define HDMI_I2S_WORD_LEN_MAX24_22BITS (0x2 << 1) +#define HDMI_I2S_WORD_LEN_MAX24_23BITS (0x4 << 1) +#define HDMI_I2S_WORD_LEN_MAX24_24BITS (0x5 << 1) +#define HDMI_I2S_WORD_LEN_MAX24_21BITS (0x6 << 1) +#define HDMI_I2S_WORD_LEN_MAX20_16BITS (0x1 << 1) +#define HDMI_I2S_WORD_LEN_MAX20_18BITS (0x2 << 1) +#define HDMI_I2S_WORD_LEN_MAX20_19BITS (0x4 << 1) +#define HDMI_I2S_WORD_LEN_MAX20_20BITS (0x5 << 1) +#define HDMI_I2S_WORD_LEN_MAX20_17BITS (0x6 << 1) +#define HDMI_I2S_WORD_LEN_MAX_24BITS (1) +#define HDMI_I2S_WORD_LEN_MAX_20BITS (0) + +/* I2S_MUX_CH */ +#define HDMI_I2S_CH3_R_EN (1 << 7) +#define HDMI_I2S_CH3_L_EN (1 << 6) +#define HDMI_I2S_CH3_EN (3 << 6) +#define HDMI_I2S_CH2_R_EN (1 << 5) +#define HDMI_I2S_CH2_L_EN (1 << 4) +#define HDMI_I2S_CH2_EN (3 << 4) +#define HDMI_I2S_CH1_R_EN (1 << 3) +#define HDMI_I2S_CH1_L_EN (1 << 2) +#define HDMI_I2S_CH1_EN (3 << 2) +#define HDMI_I2S_CH0_R_EN (1 << 1) +#define HDMI_I2S_CH0_L_EN (1) +#define HDMI_I2S_CH0_EN (3) +#define HDMI_I2S_CH_ALL_EN (0xFF) +#define HDMI_I2S_MUX_CH_CLR (~HDMI_I2S_CH_ALL_EN) + +/* I2S_MUX_CUV */ +#define HDMI_I2S_CUV_R_EN (1 << 1) +#define HDMI_I2S_CUV_L_EN (1) +#define HDMI_I2S_CUV_RL_EN (0x03) + +/* I2S_CUV_L_R */ +#define HDMI_I2S_CUV_R_DATA_MASK (0x7 << 4) +#define HDMI_I2S_CUV_L_DATA_MASK (0x7) + +/* Timing generator registers */ +/* TG configure/status registers */ +#define HDMI_TG_VACT_ST3_L HDMI_TG_BASE(0x0068) +#define HDMI_TG_VACT_ST3_H HDMI_TG_BASE(0x006c) +#define HDMI_TG_VACT_ST4_L HDMI_TG_BASE(0x0070) +#define HDMI_TG_VACT_ST4_H HDMI_TG_BASE(0x0074) +#define HDMI_TG_3D HDMI_TG_BASE(0x00F0) + +#endif /* SAMSUNG_REGS_HDMI_H */ diff --git a/drivers/gpu/drm/exynos/regs-mixer.h b/drivers/gpu/drm/exynos/regs-mixer.h new file mode 100644 index 0000000..fd2f4d1 --- /dev/null +++ b/drivers/gpu/drm/exynos/regs-mixer.h @@ -0,0 +1,141 @@ +/* + * + * Cloned from drivers/media/video/s5p-tv/regs-mixer.h + * + * Copyright (c) 2010-2011 Samsung Electronics Co., Ltd. + * http://www.samsung.com/ + * + * Mixer register header file for Samsung Mixer driver + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. +*/ +#ifndef SAMSUNG_REGS_MIXER_H +#define SAMSUNG_REGS_MIXER_H + +/* + * Register part + */ +#define MXR_STATUS 0x0000 +#define MXR_CFG 0x0004 +#define MXR_INT_EN 0x0008 +#define MXR_INT_STATUS 0x000C +#define MXR_LAYER_CFG 0x0010 +#define MXR_VIDEO_CFG 0x0014 +#define MXR_GRAPHIC0_CFG 0x0020 +#define MXR_GRAPHIC0_BASE 0x0024 +#define MXR_GRAPHIC0_SPAN 0x0028 +#define MXR_GRAPHIC0_SXY 0x002C +#define MXR_GRAPHIC0_WH 0x0030 +#define MXR_GRAPHIC0_DXY 0x0034 +#define MXR_GRAPHIC0_BLANK 0x0038 +#define MXR_GRAPHIC1_CFG 0x0040 +#define MXR_GRAPHIC1_BASE 0x0044 +#define MXR_GRAPHIC1_SPAN 0x0048 +#define MXR_GRAPHIC1_SXY 0x004C +#define MXR_GRAPHIC1_WH 0x0050 +#define MXR_GRAPHIC1_DXY 0x0054 +#define MXR_GRAPHIC1_BLANK 0x0058 +#define MXR_BG_CFG 0x0060 +#define MXR_BG_COLOR0 0x0064 +#define MXR_BG_COLOR1 0x0068 +#define MXR_BG_COLOR2 0x006C +#define MXR_CM_COEFF_Y 0x0080 +#define MXR_CM_COEFF_CB 0x0084 +#define MXR_CM_COEFF_CR 0x0088 +#define MXR_GRAPHIC0_BASE_S 0x2024 +#define MXR_GRAPHIC1_BASE_S 0x2044 + +/* for parametrized access to layer registers */ +#define MXR_GRAPHIC_CFG(i) (0x0020 + (i) * 0x20) +#define MXR_GRAPHIC_BASE(i) (0x0024 + (i) * 0x20) +#define MXR_GRAPHIC_SPAN(i) (0x0028 + (i) * 0x20) +#define MXR_GRAPHIC_SXY(i) (0x002C + (i) * 0x20) +#define MXR_GRAPHIC_WH(i) (0x0030 + (i) * 0x20) +#define MXR_GRAPHIC_DXY(i) (0x0034 + (i) * 0x20) +#define MXR_GRAPHIC_BLANK(i) (0x0038 + (i) * 0x20) +#define MXR_GRAPHIC_BASE_S(i) (0x2024 + (i) * 0x20) + +/* + * Bit definition part + */ + +/* generates mask for range of bits */ +#define MXR_MASK(high_bit, low_bit) \ + (((2 << ((high_bit) - (low_bit))) - 1) << (low_bit)) + +#define MXR_MASK_VAL(val, high_bit, low_bit) \ + (((val) << (low_bit)) & MXR_MASK(high_bit, low_bit)) + +/* bits for MXR_STATUS */ +#define MXR_STATUS_16_BURST (1 << 7) +#define MXR_STATUS_BURST_MASK (1 << 7) +#define MXR_STATUS_BIG_ENDIAN (1 << 3) +#define MXR_STATUS_ENDIAN_MASK (1 << 3) +#define MXR_STATUS_SYNC_ENABLE (1 << 2) +#define MXR_STATUS_REG_RUN (1 << 0) + +/* bits for MXR_CFG */ +#define MXR_CFG_RGB601_0_255 (0 << 9) +#define MXR_CFG_RGB601_16_235 (1 << 9) +#define MXR_CFG_RGB709_0_255 (2 << 9) +#define MXR_CFG_RGB709_16_235 (3 << 9) +#define MXR_CFG_RGB_FMT_MASK 0x600 +#define MXR_CFG_OUT_YUV444 (0 << 8) +#define MXR_CFG_OUT_RGB888 (1 << 8) +#define MXR_CFG_OUT_MASK (1 << 8) +#define MXR_CFG_DST_SDO (0 << 7) +#define MXR_CFG_DST_HDMI (1 << 7) +#define MXR_CFG_DST_MASK (1 << 7) +#define MXR_CFG_SCAN_HD_720 (0 << 6) +#define MXR_CFG_SCAN_HD_1080 (1 << 6) +#define MXR_CFG_GRP1_ENABLE (1 << 5) +#define MXR_CFG_GRP0_ENABLE (1 << 4) +#define MXR_CFG_VP_ENABLE (1 << 3) +#define MXR_CFG_SCAN_INTERLACE (0 << 2) +#define MXR_CFG_SCAN_PROGRASSIVE (1 << 2) +#define MXR_CFG_SCAN_NTSC (0 << 1) +#define MXR_CFG_SCAN_PAL (1 << 1) +#define MXR_CFG_SCAN_SD (0 << 0) +#define MXR_CFG_SCAN_HD (1 << 0) +#define MXR_CFG_SCAN_MASK 0x47 + +/* bits for MXR_GRAPHICn_CFG */ +#define MXR_GRP_CFG_COLOR_KEY_DISABLE (1 << 21) +#define MXR_GRP_CFG_BLEND_PRE_MUL (1 << 20) +#define MXR_GRP_CFG_WIN_BLEND_EN (1 << 17) +#define MXR_GRP_CFG_PIXEL_BLEND_EN (1 << 16) +#define MXR_GRP_CFG_FORMAT_VAL(x) MXR_MASK_VAL(x, 11, 8) +#define MXR_GRP_CFG_FORMAT_MASK MXR_GRP_CFG_FORMAT_VAL(~0) +#define MXR_GRP_CFG_ALPHA_VAL(x) MXR_MASK_VAL(x, 7, 0) + +/* bits for MXR_GRAPHICn_WH */ +#define MXR_GRP_WH_H_SCALE(x) MXR_MASK_VAL(x, 28, 28) +#define MXR_GRP_WH_V_SCALE(x) MXR_MASK_VAL(x, 12, 12) +#define MXR_GRP_WH_WIDTH(x) MXR_MASK_VAL(x, 26, 16) +#define MXR_GRP_WH_HEIGHT(x) MXR_MASK_VAL(x, 10, 0) + +/* bits for MXR_GRAPHICn_SXY */ +#define MXR_GRP_SXY_SX(x) MXR_MASK_VAL(x, 26, 16) +#define MXR_GRP_SXY_SY(x) MXR_MASK_VAL(x, 10, 0) + +/* bits for MXR_GRAPHICn_DXY */ +#define MXR_GRP_DXY_DX(x) MXR_MASK_VAL(x, 26, 16) +#define MXR_GRP_DXY_DY(x) MXR_MASK_VAL(x, 10, 0) + +/* bits for MXR_INT_EN */ +#define MXR_INT_EN_VSYNC (1 << 11) +#define MXR_INT_EN_ALL (0x0f << 8) + +/* bit for MXR_INT_STATUS */ +#define MXR_INT_CLEAR_VSYNC (1 << 11) +#define MXR_INT_STATUS_VSYNC (1 << 0) + +/* bit for MXR_LAYER_CFG */ +#define MXR_LAYER_CFG_GRP1_VAL(x) MXR_MASK_VAL(x, 11, 8) +#define MXR_LAYER_CFG_GRP0_VAL(x) MXR_MASK_VAL(x, 7, 4) +#define MXR_LAYER_CFG_VP_VAL(x) MXR_MASK_VAL(x, 3, 0) + +#endif /* SAMSUNG_REGS_MIXER_H */ + diff --git a/drivers/gpu/drm/exynos/regs-vp.h b/drivers/gpu/drm/exynos/regs-vp.h new file mode 100644 index 0000000..10b737a --- /dev/null +++ b/drivers/gpu/drm/exynos/regs-vp.h @@ -0,0 +1,91 @@ +/* + * + * Cloned from drivers/media/video/s5p-tv/regs-vp.h + * + * Copyright (c) 2010-2011 Samsung Electronics Co., Ltd. + * http://www.samsung.com/ + * + * Video processor register header file for Samsung Mixer driver + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 as + * published by the Free Software Foundation. + */ + +#ifndef SAMSUNG_REGS_VP_H +#define SAMSUNG_REGS_VP_H + +/* + * Register part + */ + +#define VP_ENABLE 0x0000 +#define VP_SRESET 0x0004 +#define VP_SHADOW_UPDATE 0x0008 +#define VP_FIELD_ID 0x000C +#define VP_MODE 0x0010 +#define VP_IMG_SIZE_Y 0x0014 +#define VP_IMG_SIZE_C 0x0018 +#define VP_PER_RATE_CTRL 0x001C +#define VP_TOP_Y_PTR 0x0028 +#define VP_BOT_Y_PTR 0x002C +#define VP_TOP_C_PTR 0x0030 +#define VP_BOT_C_PTR 0x0034 +#define VP_ENDIAN_MODE 0x03CC +#define VP_SRC_H_POSITION 0x0044 +#define VP_SRC_V_POSITION 0x0048 +#define VP_SRC_WIDTH 0x004C +#define VP_SRC_HEIGHT 0x0050 +#define VP_DST_H_POSITION 0x0054 +#define VP_DST_V_POSITION 0x0058 +#define VP_DST_WIDTH 0x005C +#define VP_DST_HEIGHT 0x0060 +#define VP_H_RATIO 0x0064 +#define VP_V_RATIO 0x0068 +#define VP_POLY8_Y0_LL 0x006C +#define VP_POLY4_Y0_LL 0x00EC +#define VP_POLY4_C0_LL 0x012C + +/* + * Bit definition part + */ + +/* generates mask for range of bits */ + +#define VP_MASK(high_bit, low_bit) \ + (((2 << ((high_bit) - (low_bit))) - 1) << (low_bit)) + +#define VP_MASK_VAL(val, high_bit, low_bit) \ + (((val) << (low_bit)) & VP_MASK(high_bit, low_bit)) + + /* VP_ENABLE */ +#define VP_ENABLE_ON (1 << 0) + +/* VP_SRESET */ +#define VP_SRESET_PROCESSING (1 << 0) + +/* VP_SHADOW_UPDATE */ +#define VP_SHADOW_UPDATE_ENABLE (1 << 0) + +/* VP_MODE */ +#define VP_MODE_NV12 (0 << 6) +#define VP_MODE_NV21 (1 << 6) +#define VP_MODE_LINE_SKIP (1 << 5) +#define VP_MODE_MEM_LINEAR (0 << 4) +#define VP_MODE_MEM_TILED (1 << 4) +#define VP_MODE_FMT_MASK (5 << 4) +#define VP_MODE_FIELD_ID_AUTO_TOGGLING (1 << 2) +#define VP_MODE_2D_IPC (1 << 1) + +/* VP_IMG_SIZE_Y */ +/* VP_IMG_SIZE_C */ +#define VP_IMG_HSIZE(x) VP_MASK_VAL(x, 29, 16) +#define VP_IMG_VSIZE(x) VP_MASK_VAL(x, 13, 0) + +/* VP_SRC_H_POSITION */ +#define VP_SRC_H_POSITION_VAL(x) VP_MASK_VAL(x, 14, 4) + +/* VP_ENDIAN_MODE */ +#define VP_ENDIAN_MODE_LITTLE (1 << 0) + +#endif /* SAMSUNG_REGS_VP_H */ diff --git a/drivers/gpu/drm/gma500/Kconfig b/drivers/gpu/drm/gma500/Kconfig new file mode 100644 index 0000000..42e665c --- /dev/null +++ b/drivers/gpu/drm/gma500/Kconfig @@ -0,0 +1,33 @@ +config DRM_GMA500 + tristate "Intel GMA5/600 KMS Framebuffer" + depends on DRM && PCI && X86 && EXPERIMENTAL + select FB_CFB_COPYAREA + select FB_CFB_FILLRECT + select FB_CFB_IMAGEBLIT + select DRM_KMS_HELPER + select DRM_TTM + help + Say yes for an experimental 2D KMS framebuffer driver for the + Intel GMA500 ('Poulsbo') and other Intel IMG based graphics + devices. + +config DRM_GMA600 + bool "Intel GMA600 support (Experimental)" + depends on DRM_GMA500 + help + Say yes to include support for GMA600 (Intel Moorestown/Oaktrail) + platforms with LVDS ports. MIPI is not currently supported. + +config DRM_GMA3600 + bool "Intel GMA3600/3650 support (Experimental)" + depends on DRM_GMA500 + help + Say yes to include basic support for Intel GMA3600/3650 (Intel + Cedar Trail) platforms. + +config DRM_MEDFIELD + bool "Intel Medfield support (Experimental)" + depends on DRM_GMA500 && X86_INTEL_MID + help + Say yes to include support for the Intel Medfield platform. + diff --git a/drivers/gpu/drm/gma500/Makefile b/drivers/gpu/drm/gma500/Makefile new file mode 100644 index 0000000..1583982 --- /dev/null +++ b/drivers/gpu/drm/gma500/Makefile @@ -0,0 +1,50 @@ +# +# KMS driver for the GMA500 +# +ccflags-y += -Iinclude/drm + +gma500_gfx-y += gem_glue.o \ + accel_2d.o \ + backlight.o \ + framebuffer.o \ + gem.o \ + gtt.o \ + intel_bios.o \ + intel_i2c.o \ + intel_gmbus.o \ + intel_opregion.o \ + mmu.o \ + power.o \ + psb_drv.o \ + psb_intel_display.o \ + psb_intel_lvds.o \ + psb_intel_modes.o \ + psb_intel_sdvo.o \ + psb_lid.o \ + psb_irq.o \ + psb_device.o \ + mid_bios.o + +gma500_gfx-$(CONFIG_DRM_GMA3600) += cdv_device.o \ + cdv_intel_crt.o \ + cdv_intel_display.o \ + cdv_intel_hdmi.o \ + cdv_intel_lvds.o + +gma500_gfx-$(CONFIG_DRM_GMA600) += oaktrail_device.o \ + oaktrail_crtc.o \ + oaktrail_lvds.o \ + oaktrail_hdmi.o \ + oaktrail_hdmi_i2c.o + +gma500_gfx-$(CONFIG_DRM_MEDFIELD) += mdfld_device.o \ + mdfld_output.o \ + mdfld_intel_display.o \ + mdfld_dsi_output.o \ + mdfld_dsi_dpi.o \ + mdfld_dsi_pkg_sender.o \ + mdfld_tpo_vid.o \ + mdfld_tmd_vid.o \ + tc35876x-dsi-lvds.o + +obj-$(CONFIG_DRM_GMA500) += gma500_gfx.o diff --git a/drivers/gpu/drm/gma500/accel_2d.c b/drivers/gpu/drm/gma500/accel_2d.c new file mode 100644 index 0000000..d5ef1a5 --- /dev/null +++ b/drivers/gpu/drm/gma500/accel_2d.c @@ -0,0 +1,364 @@ +/************************************************************************** + * Copyright (c) 2007-2011, Intel Corporation. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + * Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to + * develop this driver. + * + **************************************************************************/ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include + +#include "psb_drv.h" +#include "psb_reg.h" +#include "framebuffer.h" + +/** + * psb_spank - reset the 2D engine + * @dev_priv: our PSB DRM device + * + * Soft reset the graphics engine and then reload the necessary registers. + * We use this at initialisation time but it will become relevant for + * accelerated X later + */ +void psb_spank(struct drm_psb_private *dev_priv) +{ + PSB_WSGX32(_PSB_CS_RESET_BIF_RESET | _PSB_CS_RESET_DPM_RESET | + _PSB_CS_RESET_TA_RESET | _PSB_CS_RESET_USE_RESET | + _PSB_CS_RESET_ISP_RESET | _PSB_CS_RESET_TSP_RESET | + _PSB_CS_RESET_TWOD_RESET, PSB_CR_SOFT_RESET); + PSB_RSGX32(PSB_CR_SOFT_RESET); + + msleep(1); + + PSB_WSGX32(0, PSB_CR_SOFT_RESET); + wmb(); + PSB_WSGX32(PSB_RSGX32(PSB_CR_BIF_CTRL) | _PSB_CB_CTRL_CLEAR_FAULT, + PSB_CR_BIF_CTRL); + wmb(); + (void) PSB_RSGX32(PSB_CR_BIF_CTRL); + + msleep(1); + PSB_WSGX32(PSB_RSGX32(PSB_CR_BIF_CTRL) & ~_PSB_CB_CTRL_CLEAR_FAULT, + PSB_CR_BIF_CTRL); + (void) PSB_RSGX32(PSB_CR_BIF_CTRL); + PSB_WSGX32(dev_priv->gtt.gatt_start, PSB_CR_BIF_TWOD_REQ_BASE); +} + +/** + * psb2_2d_wait_available - wait for FIFO room + * @dev_priv: our DRM device + * @size: size (in dwords) of the command we want to issue + * + * Wait until there is room to load the FIFO with our data. If the + * device is not responding then reset it + */ +static int psb_2d_wait_available(struct drm_psb_private *dev_priv, + unsigned size) +{ + uint32_t avail = PSB_RSGX32(PSB_CR_2D_SOCIF); + unsigned long t = jiffies + HZ; + + while (avail < size) { + avail = PSB_RSGX32(PSB_CR_2D_SOCIF); + if (time_after(jiffies, t)) { + psb_spank(dev_priv); + return -EIO; + } + } + return 0; +} + +/** + * psb_2d_submit - submit a 2D command + * @dev_priv: our DRM device + * @cmdbuf: command to issue + * @size: length (in dwords) + * + * Issue one or more 2D commands to the accelerator. This needs to be + * serialized later when we add the GEM interfaces for acceleration + */ +static int psbfb_2d_submit(struct drm_psb_private *dev_priv, uint32_t *cmdbuf, + unsigned size) +{ + int ret = 0; + int i; + unsigned submit_size; + unsigned long flags; + + spin_lock_irqsave(&dev_priv->lock_2d, flags); + while (size > 0) { + submit_size = (size < 0x60) ? size : 0x60; + size -= submit_size; + ret = psb_2d_wait_available(dev_priv, submit_size); + if (ret) + break; + + submit_size <<= 2; + + for (i = 0; i < submit_size; i += 4) + PSB_WSGX32(*cmdbuf++, PSB_SGX_2D_SLAVE_PORT + i); + + (void)PSB_RSGX32(PSB_SGX_2D_SLAVE_PORT + i - 4); + } + spin_unlock_irqrestore(&dev_priv->lock_2d, flags); + return ret; +} + + +/** + * psb_accel_2d_copy_direction - compute blit order + * @xdir: X direction of move + * @ydir: Y direction of move + * + * Compute the correct order setings to ensure that an overlapping blit + * correctly copies all the pixels. + */ +static u32 psb_accel_2d_copy_direction(int xdir, int ydir) +{ + if (xdir < 0) + return (ydir < 0) ? PSB_2D_COPYORDER_BR2TL : + PSB_2D_COPYORDER_TR2BL; + else + return (ydir < 0) ? PSB_2D_COPYORDER_BL2TR : + PSB_2D_COPYORDER_TL2BR; +} + +/** + * psb_accel_2d_copy - accelerated 2D copy + * @dev_priv: our DRM device + * @src_offset in bytes + * @src_stride in bytes + * @src_format psb 2D format defines + * @dst_offset in bytes + * @dst_stride in bytes + * @dst_format psb 2D format defines + * @src_x offset in pixels + * @src_y offset in pixels + * @dst_x offset in pixels + * @dst_y offset in pixels + * @size_x of the copied area + * @size_y of the copied area + * + * Format and issue a 2D accelerated copy command. + */ +static int psb_accel_2d_copy(struct drm_psb_private *dev_priv, + uint32_t src_offset, uint32_t src_stride, + uint32_t src_format, uint32_t dst_offset, + uint32_t dst_stride, uint32_t dst_format, + uint16_t src_x, uint16_t src_y, + uint16_t dst_x, uint16_t dst_y, + uint16_t size_x, uint16_t size_y) +{ + uint32_t blit_cmd; + uint32_t buffer[10]; + uint32_t *buf; + uint32_t direction; + + buf = buffer; + + direction = + psb_accel_2d_copy_direction(src_x - dst_x, src_y - dst_y); + + if (direction == PSB_2D_COPYORDER_BR2TL || + direction == PSB_2D_COPYORDER_TR2BL) { + src_x += size_x - 1; + dst_x += size_x - 1; + } + if (direction == PSB_2D_COPYORDER_BR2TL || + direction == PSB_2D_COPYORDER_BL2TR) { + src_y += size_y - 1; + dst_y += size_y - 1; + } + + blit_cmd = + PSB_2D_BLIT_BH | + PSB_2D_ROT_NONE | + PSB_2D_DSTCK_DISABLE | + PSB_2D_SRCCK_DISABLE | + PSB_2D_USE_PAT | PSB_2D_ROP3_SRCCOPY | direction; + + *buf++ = PSB_2D_FENCE_BH; + *buf++ = + PSB_2D_DST_SURF_BH | dst_format | (dst_stride << + PSB_2D_DST_STRIDE_SHIFT); + *buf++ = dst_offset; + *buf++ = + PSB_2D_SRC_SURF_BH | src_format | (src_stride << + PSB_2D_SRC_STRIDE_SHIFT); + *buf++ = src_offset; + *buf++ = + PSB_2D_SRC_OFF_BH | (src_x << PSB_2D_SRCOFF_XSTART_SHIFT) | + (src_y << PSB_2D_SRCOFF_YSTART_SHIFT); + *buf++ = blit_cmd; + *buf++ = + (dst_x << PSB_2D_DST_XSTART_SHIFT) | (dst_y << + PSB_2D_DST_YSTART_SHIFT); + *buf++ = + (size_x << PSB_2D_DST_XSIZE_SHIFT) | (size_y << + PSB_2D_DST_YSIZE_SHIFT); + *buf++ = PSB_2D_FLUSH_BH; + + return psbfb_2d_submit(dev_priv, buffer, buf - buffer); +} + +/** + * psbfb_copyarea_accel - copyarea acceleration for /dev/fb + * @info: our framebuffer + * @a: copyarea parameters from the framebuffer core + * + * Perform a 2D copy via the accelerator + */ +static void psbfb_copyarea_accel(struct fb_info *info, + const struct fb_copyarea *a) +{ + struct psb_fbdev *fbdev = info->par; + struct psb_framebuffer *psbfb = &fbdev->pfb; + struct drm_device *dev = psbfb->base.dev; + struct drm_framebuffer *fb = fbdev->psb_fb_helper.fb; + struct drm_psb_private *dev_priv = dev->dev_private; + uint32_t offset; + uint32_t stride; + uint32_t src_format; + uint32_t dst_format; + + if (!fb) + return; + + offset = psbfb->gtt->offset; + stride = fb->pitches[0]; + + switch (fb->depth) { + case 8: + src_format = PSB_2D_SRC_332RGB; + dst_format = PSB_2D_DST_332RGB; + break; + case 15: + src_format = PSB_2D_SRC_555RGB; + dst_format = PSB_2D_DST_555RGB; + break; + case 16: + src_format = PSB_2D_SRC_565RGB; + dst_format = PSB_2D_DST_565RGB; + break; + case 24: + case 32: + /* this is wrong but since we don't do blending its okay */ + src_format = PSB_2D_SRC_8888ARGB; + dst_format = PSB_2D_DST_8888ARGB; + break; + default: + /* software fallback */ + cfb_copyarea(info, a); + return; + } + + if (!gma_power_begin(dev, false)) { + cfb_copyarea(info, a); + return; + } + psb_accel_2d_copy(dev_priv, + offset, stride, src_format, + offset, stride, dst_format, + a->sx, a->sy, a->dx, a->dy, a->width, a->height); + gma_power_end(dev); +} + +/** + * psbfb_copyarea - 2D copy interface + * @info: our framebuffer + * @region: region to copy + * + * Copy an area of the framebuffer console either by the accelerator + * or directly using the cfb helpers according to the request + */ +void psbfb_copyarea(struct fb_info *info, + const struct fb_copyarea *region) +{ + if (unlikely(info->state != FBINFO_STATE_RUNNING)) + return; + + /* Avoid the 8 pixel erratum */ + if (region->width == 8 || region->height == 8 || + (info->flags & FBINFO_HWACCEL_DISABLED)) + return cfb_copyarea(info, region); + + psbfb_copyarea_accel(info, region); +} + +/** + * psbfb_sync - synchronize 2D + * @info: our framebuffer + * + * Wait for the 2D engine to quiesce so that we can do CPU + * access to the framebuffer again + */ +int psbfb_sync(struct fb_info *info) +{ + struct psb_fbdev *fbdev = info->par; + struct psb_framebuffer *psbfb = &fbdev->pfb; + struct drm_device *dev = psbfb->base.dev; + struct drm_psb_private *dev_priv = dev->dev_private; + unsigned long _end = jiffies + DRM_HZ; + int busy = 0; + unsigned long flags; + + spin_lock_irqsave(&dev_priv->lock_2d, flags); + /* + * First idle the 2D engine. + */ + + if ((PSB_RSGX32(PSB_CR_2D_SOCIF) == _PSB_C2_SOCIF_EMPTY) && + ((PSB_RSGX32(PSB_CR_2D_BLIT_STATUS) & _PSB_C2B_STATUS_BUSY) == 0)) + goto out; + + do { + busy = (PSB_RSGX32(PSB_CR_2D_SOCIF) != _PSB_C2_SOCIF_EMPTY); + cpu_relax(); + } while (busy && !time_after_eq(jiffies, _end)); + + if (busy) + busy = (PSB_RSGX32(PSB_CR_2D_SOCIF) != _PSB_C2_SOCIF_EMPTY); + if (busy) + goto out; + + do { + busy = ((PSB_RSGX32(PSB_CR_2D_BLIT_STATUS) & + _PSB_C2B_STATUS_BUSY) != 0); + cpu_relax(); + } while (busy && !time_after_eq(jiffies, _end)); + if (busy) + busy = ((PSB_RSGX32(PSB_CR_2D_BLIT_STATUS) & + _PSB_C2B_STATUS_BUSY) != 0); + +out: + spin_unlock_irqrestore(&dev_priv->lock_2d, flags); + return (busy) ? -EBUSY : 0; +} diff --git a/drivers/gpu/drm/gma500/backlight.c b/drivers/gpu/drm/gma500/backlight.c new file mode 100644 index 0000000..2079395 --- /dev/null +++ b/drivers/gpu/drm/gma500/backlight.c @@ -0,0 +1,49 @@ +/* + * GMA500 Backlight Interface + * + * Copyright (c) 2009-2011, Intel Corporation. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + * Authors: Eric Knopp + * + */ + +#include "psb_drv.h" +#include "psb_intel_reg.h" +#include "psb_intel_drv.h" +#include "intel_bios.h" +#include "power.h" + +int gma_backlight_init(struct drm_device *dev) +{ +#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE + struct drm_psb_private *dev_priv = dev->dev_private; + return dev_priv->ops->backlight_init(dev); +#else + return 0; +#endif +} + +void gma_backlight_exit(struct drm_device *dev) +{ +#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE + struct drm_psb_private *dev_priv = dev->dev_private; + if (dev_priv->backlight_device) { + dev_priv->backlight_device->props.brightness = 0; + backlight_update_status(dev_priv->backlight_device); + backlight_device_unregister(dev_priv->backlight_device); + } +#endif +} diff --git a/drivers/gpu/drm/gma500/cdv_device.c b/drivers/gpu/drm/gma500/cdv_device.c new file mode 100644 index 0000000..a54cc73 --- /dev/null +++ b/drivers/gpu/drm/gma500/cdv_device.c @@ -0,0 +1,484 @@ +/************************************************************************** + * Copyright (c) 2011, Intel Corporation. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + **************************************************************************/ + +#include +#include +#include +#include "gma_drm.h" +#include "psb_drv.h" +#include "psb_reg.h" +#include "psb_intel_reg.h" +#include "intel_bios.h" +#include "cdv_device.h" + +#define VGA_SR_INDEX 0x3c4 +#define VGA_SR_DATA 0x3c5 + +static void cdv_disable_vga(struct drm_device *dev) +{ + u8 sr1; + u32 vga_reg; + + vga_reg = VGACNTRL; + + outb(1, VGA_SR_INDEX); + sr1 = inb(VGA_SR_DATA); + outb(sr1 | 1<<5, VGA_SR_DATA); + udelay(300); + + REG_WRITE(vga_reg, VGA_DISP_DISABLE); + REG_READ(vga_reg); +} + +static int cdv_output_init(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + cdv_disable_vga(dev); + + cdv_intel_crt_init(dev, &dev_priv->mode_dev); + cdv_intel_lvds_init(dev, &dev_priv->mode_dev); + + /* These bits indicate HDMI not SDVO on CDV, but we don't yet support + the HDMI interface */ + if (REG_READ(SDVOB) & SDVO_DETECTED) + cdv_hdmi_init(dev, &dev_priv->mode_dev, SDVOB); + if (REG_READ(SDVOC) & SDVO_DETECTED) + cdv_hdmi_init(dev, &dev_priv->mode_dev, SDVOC); + return 0; +} + +#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE + +/* + * Poulsbo Backlight Interfaces + */ + +#define BLC_PWM_PRECISION_FACTOR 100 /* 10000000 */ +#define BLC_PWM_FREQ_CALC_CONSTANT 32 +#define MHz 1000000 + +#define PSB_BLC_PWM_PRECISION_FACTOR 10 +#define PSB_BLC_MAX_PWM_REG_FREQ 0xFFFE +#define PSB_BLC_MIN_PWM_REG_FREQ 0x2 + +#define PSB_BACKLIGHT_PWM_POLARITY_BIT_CLEAR (0xFFFE) +#define PSB_BACKLIGHT_PWM_CTL_SHIFT (16) + +static int cdv_brightness; +static struct backlight_device *cdv_backlight_device; + +static int cdv_get_brightness(struct backlight_device *bd) +{ + /* return locally cached var instead of HW read (due to DPST etc.) */ + /* FIXME: ideally return actual value in case firmware fiddled with + it */ + return cdv_brightness; +} + + +static int cdv_backlight_setup(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + unsigned long core_clock; + /* u32 bl_max_freq; */ + /* unsigned long value; */ + u16 bl_max_freq; + uint32_t value; + uint32_t blc_pwm_precision_factor; + + /* get bl_max_freq and pol from dev_priv*/ + if (!dev_priv->lvds_bl) { + dev_err(dev->dev, "Has no valid LVDS backlight info\n"); + return -ENOENT; + } + bl_max_freq = dev_priv->lvds_bl->freq; + blc_pwm_precision_factor = PSB_BLC_PWM_PRECISION_FACTOR; + + core_clock = dev_priv->core_freq; + + value = (core_clock * MHz) / BLC_PWM_FREQ_CALC_CONSTANT; + value *= blc_pwm_precision_factor; + value /= bl_max_freq; + value /= blc_pwm_precision_factor; + + if (value > (unsigned long long)PSB_BLC_MAX_PWM_REG_FREQ || + value < (unsigned long long)PSB_BLC_MIN_PWM_REG_FREQ) + return -ERANGE; + else { + /* FIXME */ + } + return 0; +} + +static int cdv_set_brightness(struct backlight_device *bd) +{ + int level = bd->props.brightness; + + /* Percentage 1-100% being valid */ + if (level < 1) + level = 1; + + /*cdv_intel_lvds_set_brightness(dev, level); FIXME */ + cdv_brightness = level; + return 0; +} + +static const struct backlight_ops cdv_ops = { + .get_brightness = cdv_get_brightness, + .update_status = cdv_set_brightness, +}; + +static int cdv_backlight_init(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + int ret; + struct backlight_properties props; + + memset(&props, 0, sizeof(struct backlight_properties)); + props.max_brightness = 100; + props.type = BACKLIGHT_PLATFORM; + + cdv_backlight_device = backlight_device_register("psb-bl", + NULL, (void *)dev, &cdv_ops, &props); + if (IS_ERR(cdv_backlight_device)) + return PTR_ERR(cdv_backlight_device); + + ret = cdv_backlight_setup(dev); + if (ret < 0) { + backlight_device_unregister(cdv_backlight_device); + cdv_backlight_device = NULL; + return ret; + } + cdv_backlight_device->props.brightness = 100; + cdv_backlight_device->props.max_brightness = 100; + backlight_update_status(cdv_backlight_device); + dev_priv->backlight_device = cdv_backlight_device; + return 0; +} + +#endif + +/* + * Provide the Cedarview specific chip logic and low level methods + * for power management + * + * FIXME: we need to implement the apm/ospm base management bits + * for this and the MID devices. + */ + +static inline u32 CDV_MSG_READ32(uint port, uint offset) +{ + int mcr = (0x10<<24) | (port << 16) | (offset << 8); + uint32_t ret_val = 0; + struct pci_dev *pci_root = pci_get_bus_and_slot(0, 0); + pci_write_config_dword(pci_root, 0xD0, mcr); + pci_read_config_dword(pci_root, 0xD4, &ret_val); + pci_dev_put(pci_root); + return ret_val; +} + +static inline void CDV_MSG_WRITE32(uint port, uint offset, u32 value) +{ + int mcr = (0x11<<24) | (port << 16) | (offset << 8) | 0xF0; + struct pci_dev *pci_root = pci_get_bus_and_slot(0, 0); + pci_write_config_dword(pci_root, 0xD4, value); + pci_write_config_dword(pci_root, 0xD0, mcr); + pci_dev_put(pci_root); +} + +#define PSB_PM_SSC 0x20 +#define PSB_PM_SSS 0x30 +#define PSB_PWRGT_GFX_ON 0x02 +#define PSB_PWRGT_GFX_OFF 0x01 +#define PSB_PWRGT_GFX_D0 0x00 +#define PSB_PWRGT_GFX_D3 0x03 + +static void cdv_init_pm(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + u32 pwr_cnt; + int i; + + dev_priv->apm_base = CDV_MSG_READ32(PSB_PUNIT_PORT, + PSB_APMBA) & 0xFFFF; + dev_priv->ospm_base = CDV_MSG_READ32(PSB_PUNIT_PORT, + PSB_OSPMBA) & 0xFFFF; + + /* Power status */ + pwr_cnt = inl(dev_priv->apm_base + PSB_APM_CMD); + + /* Enable the GPU */ + pwr_cnt &= ~PSB_PWRGT_GFX_MASK; + pwr_cnt |= PSB_PWRGT_GFX_ON; + outl(pwr_cnt, dev_priv->apm_base + PSB_APM_CMD); + + /* Wait for the GPU power */ + for (i = 0; i < 5; i++) { + u32 pwr_sts = inl(dev_priv->apm_base + PSB_APM_STS); + if ((pwr_sts & PSB_PWRGT_GFX_MASK) == 0) + return; + udelay(10); + } + dev_err(dev->dev, "GPU: power management timed out.\n"); +} + +/** + * cdv_save_display_registers - save registers lost on suspend + * @dev: our DRM device + * + * Save the state we need in order to be able to restore the interface + * upon resume from suspend + */ +static int cdv_save_display_registers(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_save_area *regs = &dev_priv->regs; + struct drm_connector *connector; + + dev_info(dev->dev, "Saving GPU registers.\n"); + + pci_read_config_byte(dev->pdev, 0xF4, ®s->cdv.saveLBB); + + regs->cdv.saveDSPCLK_GATE_D = REG_READ(DSPCLK_GATE_D); + regs->cdv.saveRAMCLK_GATE_D = REG_READ(RAMCLK_GATE_D); + + regs->cdv.saveDSPARB = REG_READ(DSPARB); + regs->cdv.saveDSPFW[0] = REG_READ(DSPFW1); + regs->cdv.saveDSPFW[1] = REG_READ(DSPFW2); + regs->cdv.saveDSPFW[2] = REG_READ(DSPFW3); + regs->cdv.saveDSPFW[3] = REG_READ(DSPFW4); + regs->cdv.saveDSPFW[4] = REG_READ(DSPFW5); + regs->cdv.saveDSPFW[5] = REG_READ(DSPFW6); + + regs->cdv.saveADPA = REG_READ(ADPA); + + regs->cdv.savePP_CONTROL = REG_READ(PP_CONTROL); + regs->cdv.savePFIT_PGM_RATIOS = REG_READ(PFIT_PGM_RATIOS); + regs->saveBLC_PWM_CTL = REG_READ(BLC_PWM_CTL); + regs->saveBLC_PWM_CTL2 = REG_READ(BLC_PWM_CTL2); + regs->cdv.saveLVDS = REG_READ(LVDS); + + regs->cdv.savePFIT_CONTROL = REG_READ(PFIT_CONTROL); + + regs->cdv.savePP_ON_DELAYS = REG_READ(PP_ON_DELAYS); + regs->cdv.savePP_OFF_DELAYS = REG_READ(PP_OFF_DELAYS); + regs->cdv.savePP_CYCLE = REG_READ(PP_CYCLE); + + regs->cdv.saveVGACNTRL = REG_READ(VGACNTRL); + + regs->cdv.saveIER = REG_READ(PSB_INT_ENABLE_R); + regs->cdv.saveIMR = REG_READ(PSB_INT_MASK_R); + + list_for_each_entry(connector, &dev->mode_config.connector_list, head) + connector->funcs->dpms(connector, DRM_MODE_DPMS_OFF); + + return 0; +} + +/** + * cdv_restore_display_registers - restore lost register state + * @dev: our DRM device + * + * Restore register state that was lost during suspend and resume. + * + * FIXME: review + */ +static int cdv_restore_display_registers(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_save_area *regs = &dev_priv->regs; + struct drm_connector *connector; + u32 temp; + + pci_write_config_byte(dev->pdev, 0xF4, regs->cdv.saveLBB); + + REG_WRITE(DSPCLK_GATE_D, regs->cdv.saveDSPCLK_GATE_D); + REG_WRITE(RAMCLK_GATE_D, regs->cdv.saveRAMCLK_GATE_D); + + /* BIOS does below anyway */ + REG_WRITE(DPIO_CFG, 0); + REG_WRITE(DPIO_CFG, DPIO_MODE_SELECT_0 | DPIO_CMN_RESET_N); + + temp = REG_READ(DPLL_A); + if ((temp & DPLL_SYNCLOCK_ENABLE) == 0) { + REG_WRITE(DPLL_A, temp | DPLL_SYNCLOCK_ENABLE); + REG_READ(DPLL_A); + } + + temp = REG_READ(DPLL_B); + if ((temp & DPLL_SYNCLOCK_ENABLE) == 0) { + REG_WRITE(DPLL_B, temp | DPLL_SYNCLOCK_ENABLE); + REG_READ(DPLL_B); + } + + udelay(500); + + REG_WRITE(DSPFW1, regs->cdv.saveDSPFW[0]); + REG_WRITE(DSPFW2, regs->cdv.saveDSPFW[1]); + REG_WRITE(DSPFW3, regs->cdv.saveDSPFW[2]); + REG_WRITE(DSPFW4, regs->cdv.saveDSPFW[3]); + REG_WRITE(DSPFW5, regs->cdv.saveDSPFW[4]); + REG_WRITE(DSPFW6, regs->cdv.saveDSPFW[5]); + + REG_WRITE(DSPARB, regs->cdv.saveDSPARB); + REG_WRITE(ADPA, regs->cdv.saveADPA); + + REG_WRITE(BLC_PWM_CTL2, regs->saveBLC_PWM_CTL2); + REG_WRITE(LVDS, regs->cdv.saveLVDS); + REG_WRITE(PFIT_CONTROL, regs->cdv.savePFIT_CONTROL); + REG_WRITE(PFIT_PGM_RATIOS, regs->cdv.savePFIT_PGM_RATIOS); + REG_WRITE(BLC_PWM_CTL, regs->saveBLC_PWM_CTL); + REG_WRITE(PP_ON_DELAYS, regs->cdv.savePP_ON_DELAYS); + REG_WRITE(PP_OFF_DELAYS, regs->cdv.savePP_OFF_DELAYS); + REG_WRITE(PP_CYCLE, regs->cdv.savePP_CYCLE); + REG_WRITE(PP_CONTROL, regs->cdv.savePP_CONTROL); + + REG_WRITE(VGACNTRL, regs->cdv.saveVGACNTRL); + + REG_WRITE(PSB_INT_ENABLE_R, regs->cdv.saveIER); + REG_WRITE(PSB_INT_MASK_R, regs->cdv.saveIMR); + + /* Fix arbitration bug */ + CDV_MSG_WRITE32(3, 0x30, 0x08027108); + + drm_mode_config_reset(dev); + + list_for_each_entry(connector, &dev->mode_config.connector_list, head) + connector->funcs->dpms(connector, DRM_MODE_DPMS_ON); + + /* Resume the modeset for every activated CRTC */ + drm_helper_resume_force_mode(dev); + return 0; +} + +static int cdv_power_down(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + u32 pwr_cnt, pwr_mask, pwr_sts; + int tries = 5; + + pwr_cnt = inl(dev_priv->apm_base + PSB_APM_CMD); + pwr_cnt &= ~PSB_PWRGT_GFX_MASK; + pwr_cnt |= PSB_PWRGT_GFX_OFF; + pwr_mask = PSB_PWRGT_GFX_MASK; + + outl(pwr_cnt, dev_priv->apm_base + PSB_APM_CMD); + + while (tries--) { + pwr_sts = inl(dev_priv->apm_base + PSB_APM_STS); + if ((pwr_sts & pwr_mask) == PSB_PWRGT_GFX_D3) + return 0; + udelay(10); + } + return 0; +} + +static int cdv_power_up(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + u32 pwr_cnt, pwr_mask, pwr_sts; + int tries = 5; + + pwr_cnt = inl(dev_priv->apm_base + PSB_APM_CMD); + pwr_cnt &= ~PSB_PWRGT_GFX_MASK; + pwr_cnt |= PSB_PWRGT_GFX_ON; + pwr_mask = PSB_PWRGT_GFX_MASK; + + outl(pwr_cnt, dev_priv->apm_base + PSB_APM_CMD); + + while (tries--) { + pwr_sts = inl(dev_priv->apm_base + PSB_APM_STS); + if ((pwr_sts & pwr_mask) == PSB_PWRGT_GFX_D0) + return 0; + udelay(10); + } + return 0; +} + +/* FIXME ? - shared with Poulsbo */ +static void cdv_get_core_freq(struct drm_device *dev) +{ + uint32_t clock; + struct pci_dev *pci_root = pci_get_bus_and_slot(0, 0); + struct drm_psb_private *dev_priv = dev->dev_private; + + pci_write_config_dword(pci_root, 0xD0, 0xD0050300); + pci_read_config_dword(pci_root, 0xD4, &clock); + pci_dev_put(pci_root); + + switch (clock & 0x07) { + case 0: + dev_priv->core_freq = 100; + break; + case 1: + dev_priv->core_freq = 133; + break; + case 2: + dev_priv->core_freq = 150; + break; + case 3: + dev_priv->core_freq = 178; + break; + case 4: + dev_priv->core_freq = 200; + break; + case 5: + case 6: + case 7: + dev_priv->core_freq = 266; + default: + dev_priv->core_freq = 0; + } +} + +static int cdv_chip_setup(struct drm_device *dev) +{ + cdv_get_core_freq(dev); + gma_intel_opregion_init(dev); + psb_intel_init_bios(dev); + REG_WRITE(PORT_HOTPLUG_EN, 0); + REG_WRITE(PORT_HOTPLUG_STAT, REG_READ(PORT_HOTPLUG_STAT)); + return 0; +} + +/* CDV is much like Poulsbo but has MID like SGX offsets and PM */ + +const struct psb_ops cdv_chip_ops = { + .name = "GMA3600/3650", + .accel_2d = 0, + .pipes = 2, + .crtcs = 2, + .sgx_offset = MRST_SGX_OFFSET, + .chip_setup = cdv_chip_setup, + + .crtc_helper = &cdv_intel_helper_funcs, + .crtc_funcs = &cdv_intel_crtc_funcs, + + .output_init = cdv_output_init, + +#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE + .backlight_init = cdv_backlight_init, +#endif + + .init_pm = cdv_init_pm, + .save_regs = cdv_save_display_registers, + .restore_regs = cdv_restore_display_registers, + .power_down = cdv_power_down, + .power_up = cdv_power_up, +}; diff --git a/drivers/gpu/drm/gma500/cdv_device.h b/drivers/gpu/drm/gma500/cdv_device.h new file mode 100644 index 0000000..9561e17 --- /dev/null +++ b/drivers/gpu/drm/gma500/cdv_device.h @@ -0,0 +1,36 @@ +/* + * Copyright © 2011 Intel Corporation + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + */ + +extern const struct drm_crtc_helper_funcs cdv_intel_helper_funcs; +extern const struct drm_crtc_funcs cdv_intel_crtc_funcs; +extern void cdv_intel_crt_init(struct drm_device *dev, + struct psb_intel_mode_device *mode_dev); +extern void cdv_intel_lvds_init(struct drm_device *dev, + struct psb_intel_mode_device *mode_dev); +extern void cdv_hdmi_init(struct drm_device *dev, struct psb_intel_mode_device *mode_dev, + int reg); +extern struct drm_display_mode *cdv_intel_crtc_mode_get(struct drm_device *dev, + struct drm_crtc *crtc); + +static inline void cdv_intel_wait_for_vblank(struct drm_device *dev) +{ + /* Wait for 20ms, i.e. one cycle at 50hz. */ + /* FIXME: msleep ?? */ + mdelay(20); +} + + diff --git a/drivers/gpu/drm/gma500/cdv_intel_crt.c b/drivers/gpu/drm/gma500/cdv_intel_crt.c new file mode 100644 index 0000000..a71a6cd --- /dev/null +++ b/drivers/gpu/drm/gma500/cdv_intel_crt.c @@ -0,0 +1,340 @@ +/* + * Copyright © 2006-2007 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * Eric Anholt + */ + +#include +#include + +#include "intel_bios.h" +#include "psb_drv.h" +#include "psb_intel_drv.h" +#include "psb_intel_reg.h" +#include "power.h" +#include "cdv_device.h" +#include + + +static void cdv_intel_crt_dpms(struct drm_encoder *encoder, int mode) +{ + struct drm_device *dev = encoder->dev; + u32 temp, reg; + reg = ADPA; + + temp = REG_READ(reg); + temp &= ~(ADPA_HSYNC_CNTL_DISABLE | ADPA_VSYNC_CNTL_DISABLE); + temp &= ~ADPA_DAC_ENABLE; + + switch (mode) { + case DRM_MODE_DPMS_ON: + temp |= ADPA_DAC_ENABLE; + break; + case DRM_MODE_DPMS_STANDBY: + temp |= ADPA_DAC_ENABLE | ADPA_HSYNC_CNTL_DISABLE; + break; + case DRM_MODE_DPMS_SUSPEND: + temp |= ADPA_DAC_ENABLE | ADPA_VSYNC_CNTL_DISABLE; + break; + case DRM_MODE_DPMS_OFF: + temp |= ADPA_HSYNC_CNTL_DISABLE | ADPA_VSYNC_CNTL_DISABLE; + break; + } + + REG_WRITE(reg, temp); +} + +static int cdv_intel_crt_mode_valid(struct drm_connector *connector, + struct drm_display_mode *mode) +{ + struct drm_psb_private *dev_priv = connector->dev->dev_private; + int max_clock = 0; + if (mode->flags & DRM_MODE_FLAG_DBLSCAN) + return MODE_NO_DBLESCAN; + + /* The lowest clock for CDV is 20000KHz */ + if (mode->clock < 20000) + return MODE_CLOCK_LOW; + + /* The max clock for CDV is 355 instead of 400 */ + max_clock = 355000; + if (mode->clock > max_clock) + return MODE_CLOCK_HIGH; + + if (mode->hdisplay > 1680 || mode->vdisplay > 1050) + return MODE_PANEL; + + /* We assume worst case scenario of 32 bpp here, since we don't know */ + if ((ALIGN(mode->hdisplay * 4, 64) * mode->vdisplay) > + dev_priv->vram_stolen_size) + return MODE_MEM; + + return MODE_OK; +} + +static bool cdv_intel_crt_mode_fixup(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + return true; +} + +static void cdv_intel_crt_mode_set(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + + struct drm_device *dev = encoder->dev; + struct drm_crtc *crtc = encoder->crtc; + struct psb_intel_crtc *psb_intel_crtc = + to_psb_intel_crtc(crtc); + int dpll_md_reg; + u32 adpa, dpll_md; + u32 adpa_reg; + + if (psb_intel_crtc->pipe == 0) + dpll_md_reg = DPLL_A_MD; + else + dpll_md_reg = DPLL_B_MD; + + adpa_reg = ADPA; + + /* + * Disable separate mode multiplier used when cloning SDVO to CRT + * XXX this needs to be adjusted when we really are cloning + */ + { + dpll_md = REG_READ(dpll_md_reg); + REG_WRITE(dpll_md_reg, + dpll_md & ~DPLL_MD_UDI_MULTIPLIER_MASK); + } + + adpa = 0; + if (adjusted_mode->flags & DRM_MODE_FLAG_PHSYNC) + adpa |= ADPA_HSYNC_ACTIVE_HIGH; + if (adjusted_mode->flags & DRM_MODE_FLAG_PVSYNC) + adpa |= ADPA_VSYNC_ACTIVE_HIGH; + + if (psb_intel_crtc->pipe == 0) + adpa |= ADPA_PIPE_A_SELECT; + else + adpa |= ADPA_PIPE_B_SELECT; + + REG_WRITE(adpa_reg, adpa); +} + + +/** + * Uses CRT_HOTPLUG_EN and CRT_HOTPLUG_STAT to detect CRT presence. + * + * \return true if CRT is connected. + * \return false if CRT is disconnected. + */ +static bool cdv_intel_crt_detect_hotplug(struct drm_connector *connector, + bool force) +{ + struct drm_device *dev = connector->dev; + u32 hotplug_en; + int i, tries = 0, ret = false; + u32 adpa_orig; + + /* disable the DAC when doing the hotplug detection */ + + adpa_orig = REG_READ(ADPA); + + REG_WRITE(ADPA, adpa_orig & ~(ADPA_DAC_ENABLE)); + + /* + * On a CDV thep, CRT detect sequence need to be done twice + * to get a reliable result. + */ + tries = 2; + + hotplug_en = REG_READ(PORT_HOTPLUG_EN); + hotplug_en &= ~(CRT_HOTPLUG_DETECT_MASK); + hotplug_en |= CRT_HOTPLUG_FORCE_DETECT; + + hotplug_en |= CRT_HOTPLUG_ACTIVATION_PERIOD_64; + hotplug_en |= CRT_HOTPLUG_VOLTAGE_COMPARE_50; + + for (i = 0; i < tries ; i++) { + unsigned long timeout; + /* turn on the FORCE_DETECT */ + REG_WRITE(PORT_HOTPLUG_EN, hotplug_en); + timeout = jiffies + msecs_to_jiffies(1000); + /* wait for FORCE_DETECT to go off */ + do { + if (!(REG_READ(PORT_HOTPLUG_EN) & + CRT_HOTPLUG_FORCE_DETECT)) + break; + msleep(1); + } while (time_after(timeout, jiffies)); + } + + if ((REG_READ(PORT_HOTPLUG_STAT) & CRT_HOTPLUG_MONITOR_MASK) != + CRT_HOTPLUG_MONITOR_NONE) + ret = true; + + /* Restore the saved ADPA */ + REG_WRITE(ADPA, adpa_orig); + return ret; +} + +static enum drm_connector_status cdv_intel_crt_detect( + struct drm_connector *connector, bool force) +{ + if (cdv_intel_crt_detect_hotplug(connector, force)) + return connector_status_connected; + else + return connector_status_disconnected; +} + +static void cdv_intel_crt_destroy(struct drm_connector *connector) +{ + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(connector); + + psb_intel_i2c_destroy(psb_intel_encoder->ddc_bus); + drm_sysfs_connector_remove(connector); + drm_connector_cleanup(connector); + kfree(connector); +} + +static int cdv_intel_crt_get_modes(struct drm_connector *connector) +{ + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(connector); + return psb_intel_ddc_get_modes(connector, &psb_intel_encoder->ddc_bus->adapter); +} + +static int cdv_intel_crt_set_property(struct drm_connector *connector, + struct drm_property *property, + uint64_t value) +{ + return 0; +} + +/* + * Routines for controlling stuff on the analog port + */ + +static const struct drm_encoder_helper_funcs cdv_intel_crt_helper_funcs = { + .dpms = cdv_intel_crt_dpms, + .mode_fixup = cdv_intel_crt_mode_fixup, + .prepare = psb_intel_encoder_prepare, + .commit = psb_intel_encoder_commit, + .mode_set = cdv_intel_crt_mode_set, +}; + +static const struct drm_connector_funcs cdv_intel_crt_connector_funcs = { + .dpms = drm_helper_connector_dpms, + .detect = cdv_intel_crt_detect, + .fill_modes = drm_helper_probe_single_connector_modes, + .destroy = cdv_intel_crt_destroy, + .set_property = cdv_intel_crt_set_property, +}; + +static const struct drm_connector_helper_funcs + cdv_intel_crt_connector_helper_funcs = { + .mode_valid = cdv_intel_crt_mode_valid, + .get_modes = cdv_intel_crt_get_modes, + .best_encoder = psb_intel_best_encoder, +}; + +static void cdv_intel_crt_enc_destroy(struct drm_encoder *encoder) +{ + drm_encoder_cleanup(encoder); +} + +static const struct drm_encoder_funcs cdv_intel_crt_enc_funcs = { + .destroy = cdv_intel_crt_enc_destroy, +}; + +void cdv_intel_crt_init(struct drm_device *dev, + struct psb_intel_mode_device *mode_dev) +{ + + struct psb_intel_connector *psb_intel_connector; + struct psb_intel_encoder *psb_intel_encoder; + struct drm_connector *connector; + struct drm_encoder *encoder; + + u32 i2c_reg; + + psb_intel_encoder = kzalloc(sizeof(struct psb_intel_encoder), GFP_KERNEL); + if (!psb_intel_encoder) + return; + + psb_intel_connector = kzalloc(sizeof(struct psb_intel_connector), GFP_KERNEL); + if (!psb_intel_connector) + goto failed_connector; + + connector = &psb_intel_connector->base; + drm_connector_init(dev, connector, + &cdv_intel_crt_connector_funcs, DRM_MODE_CONNECTOR_VGA); + + encoder = &psb_intel_encoder->base; + drm_encoder_init(dev, encoder, + &cdv_intel_crt_enc_funcs, DRM_MODE_ENCODER_DAC); + + psb_intel_connector_attach_encoder(psb_intel_connector, + psb_intel_encoder); + + /* Set up the DDC bus. */ + i2c_reg = GPIOA; + /* Remove the following code for CDV */ + /* + if (dev_priv->crt_ddc_bus != 0) + i2c_reg = dev_priv->crt_ddc_bus; + }*/ + psb_intel_encoder->ddc_bus = psb_intel_i2c_create(dev, + i2c_reg, "CRTDDC_A"); + if (!psb_intel_encoder->ddc_bus) { + dev_printk(KERN_ERR, &dev->pdev->dev, "DDC bus registration " + "failed.\n"); + goto failed_ddc; + } + + psb_intel_encoder->type = INTEL_OUTPUT_ANALOG; + /* + psb_intel_output->clone_mask = (1 << INTEL_ANALOG_CLONE_BIT); + psb_intel_output->crtc_mask = (1 << 0) | (1 << 1); + */ + connector->interlace_allowed = 0; + connector->doublescan_allowed = 0; + + drm_encoder_helper_add(encoder, &cdv_intel_crt_helper_funcs); + drm_connector_helper_add(connector, + &cdv_intel_crt_connector_helper_funcs); + + drm_sysfs_connector_add(connector); + + return; +failed_ddc: + drm_encoder_cleanup(&psb_intel_encoder->base); + drm_connector_cleanup(&psb_intel_connector->base); + kfree(psb_intel_connector); +failed_connector: + kfree(psb_intel_encoder); + return; +} diff --git a/drivers/gpu/drm/gma500/cdv_intel_display.c b/drivers/gpu/drm/gma500/cdv_intel_display.c new file mode 100644 index 0000000..be84559 --- /dev/null +++ b/drivers/gpu/drm/gma500/cdv_intel_display.c @@ -0,0 +1,1459 @@ +/* + * Copyright © 2006-2011 Intel Corporation + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + * Authors: + * Eric Anholt + */ + +#include +#include + +#include +#include "framebuffer.h" +#include "psb_drv.h" +#include "psb_intel_drv.h" +#include "psb_intel_reg.h" +#include "psb_intel_display.h" +#include "power.h" +#include "cdv_device.h" + + +struct cdv_intel_range_t { + int min, max; +}; + +struct cdv_intel_p2_t { + int dot_limit; + int p2_slow, p2_fast; +}; + +struct cdv_intel_clock_t { + /* given values */ + int n; + int m1, m2; + int p1, p2; + /* derived values */ + int dot; + int vco; + int m; + int p; +}; + +#define INTEL_P2_NUM 2 + +struct cdv_intel_limit_t { + struct cdv_intel_range_t dot, vco, n, m, m1, m2, p, p1; + struct cdv_intel_p2_t p2; +}; + +#define CDV_LIMIT_SINGLE_LVDS_96 0 +#define CDV_LIMIT_SINGLE_LVDS_100 1 +#define CDV_LIMIT_DAC_HDMI_27 2 +#define CDV_LIMIT_DAC_HDMI_96 3 + +static const struct cdv_intel_limit_t cdv_intel_limits[] = { + { /* CDV_SIGNLE_LVDS_96MHz */ + .dot = {.min = 20000, .max = 115500}, + .vco = {.min = 1800000, .max = 3600000}, + .n = {.min = 2, .max = 6}, + .m = {.min = 60, .max = 160}, + .m1 = {.min = 0, .max = 0}, + .m2 = {.min = 58, .max = 158}, + .p = {.min = 28, .max = 140}, + .p1 = {.min = 2, .max = 10}, + .p2 = {.dot_limit = 200000, + .p2_slow = 14, .p2_fast = 14}, + }, + { /* CDV_SINGLE_LVDS_100MHz */ + .dot = {.min = 20000, .max = 115500}, + .vco = {.min = 1800000, .max = 3600000}, + .n = {.min = 2, .max = 6}, + .m = {.min = 60, .max = 160}, + .m1 = {.min = 0, .max = 0}, + .m2 = {.min = 58, .max = 158}, + .p = {.min = 28, .max = 140}, + .p1 = {.min = 2, .max = 10}, + /* The single-channel range is 25-112Mhz, and dual-channel + * is 80-224Mhz. Prefer single channel as much as possible. + */ + .p2 = {.dot_limit = 200000, .p2_slow = 14, .p2_fast = 14}, + }, + { /* CDV_DAC_HDMI_27MHz */ + .dot = {.min = 20000, .max = 400000}, + .vco = {.min = 1809000, .max = 3564000}, + .n = {.min = 1, .max = 1}, + .m = {.min = 67, .max = 132}, + .m1 = {.min = 0, .max = 0}, + .m2 = {.min = 65, .max = 130}, + .p = {.min = 5, .max = 90}, + .p1 = {.min = 1, .max = 9}, + .p2 = {.dot_limit = 225000, .p2_slow = 10, .p2_fast = 5}, + }, + { /* CDV_DAC_HDMI_96MHz */ + .dot = {.min = 20000, .max = 400000}, + .vco = {.min = 1800000, .max = 3600000}, + .n = {.min = 2, .max = 6}, + .m = {.min = 60, .max = 160}, + .m1 = {.min = 0, .max = 0}, + .m2 = {.min = 58, .max = 158}, + .p = {.min = 5, .max = 100}, + .p1 = {.min = 1, .max = 10}, + .p2 = {.dot_limit = 225000, .p2_slow = 10, .p2_fast = 5}, + }, +}; + +#define _wait_for(COND, MS, W) ({ \ + unsigned long timeout__ = jiffies + msecs_to_jiffies(MS); \ + int ret__ = 0; \ + while (!(COND)) { \ + if (time_after(jiffies, timeout__)) { \ + ret__ = -ETIMEDOUT; \ + break; \ + } \ + if (W && !in_dbg_master()) \ + msleep(W); \ + } \ + ret__; \ +}) + +#define wait_for(COND, MS) _wait_for(COND, MS, 1) + + +static int cdv_sb_read(struct drm_device *dev, u32 reg, u32 *val) +{ + int ret; + + ret = wait_for((REG_READ(SB_PCKT) & SB_BUSY) == 0, 1000); + if (ret) { + DRM_ERROR("timeout waiting for SB to idle before read\n"); + return ret; + } + + REG_WRITE(SB_ADDR, reg); + REG_WRITE(SB_PCKT, + SET_FIELD(SB_OPCODE_READ, SB_OPCODE) | + SET_FIELD(SB_DEST_DPLL, SB_DEST) | + SET_FIELD(0xf, SB_BYTE_ENABLE)); + + ret = wait_for((REG_READ(SB_PCKT) & SB_BUSY) == 0, 1000); + if (ret) { + DRM_ERROR("timeout waiting for SB to idle after read\n"); + return ret; + } + + *val = REG_READ(SB_DATA); + + return 0; +} + +static int cdv_sb_write(struct drm_device *dev, u32 reg, u32 val) +{ + int ret; + static bool dpio_debug = true; + u32 temp; + + if (dpio_debug) { + if (cdv_sb_read(dev, reg, &temp) == 0) + DRM_DEBUG_KMS("0x%08x: 0x%08x (before)\n", reg, temp); + DRM_DEBUG_KMS("0x%08x: 0x%08x\n", reg, val); + } + + ret = wait_for((REG_READ(SB_PCKT) & SB_BUSY) == 0, 1000); + if (ret) { + DRM_ERROR("timeout waiting for SB to idle before write\n"); + return ret; + } + + REG_WRITE(SB_ADDR, reg); + REG_WRITE(SB_DATA, val); + REG_WRITE(SB_PCKT, + SET_FIELD(SB_OPCODE_WRITE, SB_OPCODE) | + SET_FIELD(SB_DEST_DPLL, SB_DEST) | + SET_FIELD(0xf, SB_BYTE_ENABLE)); + + ret = wait_for((REG_READ(SB_PCKT) & SB_BUSY) == 0, 1000); + if (ret) { + DRM_ERROR("timeout waiting for SB to idle after write\n"); + return ret; + } + + if (dpio_debug) { + if (cdv_sb_read(dev, reg, &temp) == 0) + DRM_DEBUG_KMS("0x%08x: 0x%08x (after)\n", reg, temp); + } + + return 0; +} + +/* Reset the DPIO configuration register. The BIOS does this at every + * mode set. + */ +static void cdv_sb_reset(struct drm_device *dev) +{ + + REG_WRITE(DPIO_CFG, 0); + REG_READ(DPIO_CFG); + REG_WRITE(DPIO_CFG, DPIO_MODE_SELECT_0 | DPIO_CMN_RESET_N); +} + +/* Unlike most Intel display engines, on Cedarview the DPLL registers + * are behind this sideband bus. They must be programmed while the + * DPLL reference clock is on in the DPLL control register, but before + * the DPLL is enabled in the DPLL control register. + */ +static int +cdv_dpll_set_clock_cdv(struct drm_device *dev, struct drm_crtc *crtc, + struct cdv_intel_clock_t *clock) +{ + struct psb_intel_crtc *psb_crtc = + to_psb_intel_crtc(crtc); + int pipe = psb_crtc->pipe; + u32 m, n_vco, p; + int ret = 0; + int dpll_reg = (pipe == 0) ? DPLL_A : DPLL_B; + u32 ref_value; + + cdv_sb_reset(dev); + + if ((REG_READ(dpll_reg) & DPLL_SYNCLOCK_ENABLE) == 0) { + DRM_ERROR("Attempting to set DPLL with refclk disabled\n"); + return -EBUSY; + } + + /* Follow the BIOS and write the REF/SFR Register. Hardcoded value */ + ref_value = 0x68A701; + + cdv_sb_write(dev, SB_REF_SFR(pipe), ref_value); + + /* We don't know what the other fields of these regs are, so + * leave them in place. + */ + ret = cdv_sb_read(dev, SB_M(pipe), &m); + if (ret) + return ret; + m &= ~SB_M_DIVIDER_MASK; + m |= ((clock->m2) << SB_M_DIVIDER_SHIFT); + ret = cdv_sb_write(dev, SB_M(pipe), m); + if (ret) + return ret; + + ret = cdv_sb_read(dev, SB_N_VCO(pipe), &n_vco); + if (ret) + return ret; + + /* Follow the BIOS to program the N_DIVIDER REG */ + n_vco &= 0xFFFF; + n_vco |= 0x107; + n_vco &= ~(SB_N_VCO_SEL_MASK | + SB_N_DIVIDER_MASK | + SB_N_CB_TUNE_MASK); + + n_vco |= ((clock->n) << SB_N_DIVIDER_SHIFT); + + if (clock->vco < 2250000) { + n_vco |= (2 << SB_N_CB_TUNE_SHIFT); + n_vco |= (0 << SB_N_VCO_SEL_SHIFT); + } else if (clock->vco < 2750000) { + n_vco |= (1 << SB_N_CB_TUNE_SHIFT); + n_vco |= (1 << SB_N_VCO_SEL_SHIFT); + } else if (clock->vco < 3300000) { + n_vco |= (0 << SB_N_CB_TUNE_SHIFT); + n_vco |= (2 << SB_N_VCO_SEL_SHIFT); + } else { + n_vco |= (0 << SB_N_CB_TUNE_SHIFT); + n_vco |= (3 << SB_N_VCO_SEL_SHIFT); + } + + ret = cdv_sb_write(dev, SB_N_VCO(pipe), n_vco); + if (ret) + return ret; + + ret = cdv_sb_read(dev, SB_P(pipe), &p); + if (ret) + return ret; + p &= ~(SB_P2_DIVIDER_MASK | SB_P1_DIVIDER_MASK); + p |= SET_FIELD(clock->p1, SB_P1_DIVIDER); + switch (clock->p2) { + case 5: + p |= SET_FIELD(SB_P2_5, SB_P2_DIVIDER); + break; + case 10: + p |= SET_FIELD(SB_P2_10, SB_P2_DIVIDER); + break; + case 14: + p |= SET_FIELD(SB_P2_14, SB_P2_DIVIDER); + break; + case 7: + p |= SET_FIELD(SB_P2_7, SB_P2_DIVIDER); + break; + default: + DRM_ERROR("Bad P2 clock: %d\n", clock->p2); + return -EINVAL; + } + ret = cdv_sb_write(dev, SB_P(pipe), p); + if (ret) + return ret; + + /* always Program the Lane Register for the Pipe A*/ + if (pipe == 0) { + /* Program the Lane0/1 for HDMI B */ + u32 lane_reg, lane_value; + + lane_reg = PSB_LANE0; + cdv_sb_read(dev, lane_reg, &lane_value); + lane_value &= ~(LANE_PLL_MASK); + lane_value |= LANE_PLL_ENABLE; + cdv_sb_write(dev, lane_reg, lane_value); + + lane_reg = PSB_LANE1; + cdv_sb_read(dev, lane_reg, &lane_value); + lane_value &= ~(LANE_PLL_MASK); + lane_value |= LANE_PLL_ENABLE; + cdv_sb_write(dev, lane_reg, lane_value); + + /* Program the Lane2/3 for HDMI C */ + lane_reg = PSB_LANE2; + cdv_sb_read(dev, lane_reg, &lane_value); + lane_value &= ~(LANE_PLL_MASK); + lane_value |= LANE_PLL_ENABLE; + cdv_sb_write(dev, lane_reg, lane_value); + + lane_reg = PSB_LANE3; + cdv_sb_read(dev, lane_reg, &lane_value); + lane_value &= ~(LANE_PLL_MASK); + lane_value |= LANE_PLL_ENABLE; + cdv_sb_write(dev, lane_reg, lane_value); + } + + return 0; +} + +/* + * Returns whether any encoder on the specified pipe is of the specified type + */ +static bool cdv_intel_pipe_has_type(struct drm_crtc *crtc, int type) +{ + struct drm_device *dev = crtc->dev; + struct drm_mode_config *mode_config = &dev->mode_config; + struct drm_connector *l_entry; + + list_for_each_entry(l_entry, &mode_config->connector_list, head) { + if (l_entry->encoder && l_entry->encoder->crtc == crtc) { + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(l_entry); + if (psb_intel_encoder->type == type) + return true; + } + } + return false; +} + +static const struct cdv_intel_limit_t *cdv_intel_limit(struct drm_crtc *crtc, + int refclk) +{ + const struct cdv_intel_limit_t *limit; + if (cdv_intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) { + /* + * Now only single-channel LVDS is supported on CDV. If it is + * incorrect, please add the dual-channel LVDS. + */ + if (refclk == 96000) + limit = &cdv_intel_limits[CDV_LIMIT_SINGLE_LVDS_96]; + else + limit = &cdv_intel_limits[CDV_LIMIT_SINGLE_LVDS_100]; + } else { + if (refclk == 27000) + limit = &cdv_intel_limits[CDV_LIMIT_DAC_HDMI_27]; + else + limit = &cdv_intel_limits[CDV_LIMIT_DAC_HDMI_96]; + } + return limit; +} + +/* m1 is reserved as 0 in CDV, n is a ring counter */ +static void cdv_intel_clock(struct drm_device *dev, + int refclk, struct cdv_intel_clock_t *clock) +{ + clock->m = clock->m2 + 2; + clock->p = clock->p1 * clock->p2; + clock->vco = (refclk * clock->m) / clock->n; + clock->dot = clock->vco / clock->p; +} + + +#define INTELPllInvalid(s) { /* ErrorF (s) */; return false; } +static bool cdv_intel_PLL_is_valid(struct drm_crtc *crtc, + const struct cdv_intel_limit_t *limit, + struct cdv_intel_clock_t *clock) +{ + if (clock->p1 < limit->p1.min || limit->p1.max < clock->p1) + INTELPllInvalid("p1 out of range\n"); + if (clock->p < limit->p.min || limit->p.max < clock->p) + INTELPllInvalid("p out of range\n"); + /* unnecessary to check the range of m(m1/M2)/n again */ + if (clock->vco < limit->vco.min || limit->vco.max < clock->vco) + INTELPllInvalid("vco out of range\n"); + /* XXX: We may need to be checking "Dot clock" + * depending on the multiplier, connector, etc., + * rather than just a single range. + */ + if (clock->dot < limit->dot.min || limit->dot.max < clock->dot) + INTELPllInvalid("dot out of range\n"); + + return true; +} + +static bool cdv_intel_find_best_PLL(struct drm_crtc *crtc, int target, + int refclk, + struct cdv_intel_clock_t *best_clock) +{ + struct drm_device *dev = crtc->dev; + struct cdv_intel_clock_t clock; + const struct cdv_intel_limit_t *limit = cdv_intel_limit(crtc, refclk); + int err = target; + + + if (cdv_intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) && + (REG_READ(LVDS) & LVDS_PORT_EN) != 0) { + /* + * For LVDS, if the panel is on, just rely on its current + * settings for dual-channel. We haven't figured out how to + * reliably set up different single/dual channel state, if we + * even can. + */ + if ((REG_READ(LVDS) & LVDS_CLKB_POWER_MASK) == + LVDS_CLKB_POWER_UP) + clock.p2 = limit->p2.p2_fast; + else + clock.p2 = limit->p2.p2_slow; + } else { + if (target < limit->p2.dot_limit) + clock.p2 = limit->p2.p2_slow; + else + clock.p2 = limit->p2.p2_fast; + } + + memset(best_clock, 0, sizeof(*best_clock)); + clock.m1 = 0; + /* m1 is reserved as 0 in CDV, n is a ring counter. + So skip the m1 loop */ + for (clock.n = limit->n.min; clock.n <= limit->n.max; clock.n++) { + for (clock.m2 = limit->m2.min; clock.m2 <= limit->m2.max; + clock.m2++) { + for (clock.p1 = limit->p1.min; + clock.p1 <= limit->p1.max; + clock.p1++) { + int this_err; + + cdv_intel_clock(dev, refclk, &clock); + + if (!cdv_intel_PLL_is_valid(crtc, + limit, &clock)) + continue; + + this_err = abs(clock.dot - target); + if (this_err < err) { + *best_clock = clock; + err = this_err; + } + } + } + } + + return err != target; +} + +static int cdv_intel_pipe_set_base(struct drm_crtc *crtc, + int x, int y, struct drm_framebuffer *old_fb) +{ + struct drm_device *dev = crtc->dev; + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + struct psb_framebuffer *psbfb = to_psb_fb(crtc->fb); + int pipe = psb_intel_crtc->pipe; + unsigned long start, offset; + int dspbase = (pipe == 0 ? DSPABASE : DSPBBASE); + int dspsurf = (pipe == 0 ? DSPASURF : DSPBSURF); + int dspstride = (pipe == 0) ? DSPASTRIDE : DSPBSTRIDE; + int dspcntr_reg = (pipe == 0) ? DSPACNTR : DSPBCNTR; + u32 dspcntr; + int ret = 0; + + if (!gma_power_begin(dev, true)) + return 0; + + /* no fb bound */ + if (!crtc->fb) { + dev_err(dev->dev, "No FB bound\n"); + goto psb_intel_pipe_cleaner; + } + + + /* We are displaying this buffer, make sure it is actually loaded + into the GTT */ + ret = psb_gtt_pin(psbfb->gtt); + if (ret < 0) + goto psb_intel_pipe_set_base_exit; + start = psbfb->gtt->offset; + offset = y * crtc->fb->pitches[0] + x * (crtc->fb->bits_per_pixel / 8); + + REG_WRITE(dspstride, crtc->fb->pitches[0]); + + dspcntr = REG_READ(dspcntr_reg); + dspcntr &= ~DISPPLANE_PIXFORMAT_MASK; + + switch (crtc->fb->bits_per_pixel) { + case 8: + dspcntr |= DISPPLANE_8BPP; + break; + case 16: + if (crtc->fb->depth == 15) + dspcntr |= DISPPLANE_15_16BPP; + else + dspcntr |= DISPPLANE_16BPP; + break; + case 24: + case 32: + dspcntr |= DISPPLANE_32BPP_NO_ALPHA; + break; + default: + dev_err(dev->dev, "Unknown color depth\n"); + ret = -EINVAL; + goto psb_intel_pipe_set_base_exit; + } + REG_WRITE(dspcntr_reg, dspcntr); + + dev_dbg(dev->dev, + "Writing base %08lX %08lX %d %d\n", start, offset, x, y); + + REG_WRITE(dspbase, offset); + REG_READ(dspbase); + REG_WRITE(dspsurf, start); + REG_READ(dspsurf); + +psb_intel_pipe_cleaner: + /* If there was a previous display we can now unpin it */ + if (old_fb) + psb_gtt_unpin(to_psb_fb(old_fb)->gtt); + +psb_intel_pipe_set_base_exit: + gma_power_end(dev); + return ret; +} + +/** + * Sets the power management mode of the pipe and plane. + * + * This code should probably grow support for turning the cursor off and back + * on appropriately at the same time as we're turning the pipe off/on. + */ +static void cdv_intel_crtc_dpms(struct drm_crtc *crtc, int mode) +{ + struct drm_device *dev = crtc->dev; + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + int pipe = psb_intel_crtc->pipe; + int dpll_reg = (pipe == 0) ? DPLL_A : DPLL_B; + int dspcntr_reg = (pipe == 0) ? DSPACNTR : DSPBCNTR; + int dspbase_reg = (pipe == 0) ? DSPABASE : DSPBBASE; + int pipeconf_reg = (pipe == 0) ? PIPEACONF : PIPEBCONF; + u32 temp; + + /* XXX: When our outputs are all unaware of DPMS modes other than off + * and on, we should map those modes to DRM_MODE_DPMS_OFF in the CRTC. + */ + switch (mode) { + case DRM_MODE_DPMS_ON: + case DRM_MODE_DPMS_STANDBY: + case DRM_MODE_DPMS_SUSPEND: + /* Enable the DPLL */ + temp = REG_READ(dpll_reg); + if ((temp & DPLL_VCO_ENABLE) == 0) { + REG_WRITE(dpll_reg, temp); + REG_READ(dpll_reg); + /* Wait for the clocks to stabilize. */ + udelay(150); + REG_WRITE(dpll_reg, temp | DPLL_VCO_ENABLE); + REG_READ(dpll_reg); + /* Wait for the clocks to stabilize. */ + udelay(150); + REG_WRITE(dpll_reg, temp | DPLL_VCO_ENABLE); + REG_READ(dpll_reg); + /* Wait for the clocks to stabilize. */ + udelay(150); + } + + /* Jim Bish - switch plan and pipe per scott */ + /* Enable the plane */ + temp = REG_READ(dspcntr_reg); + if ((temp & DISPLAY_PLANE_ENABLE) == 0) { + REG_WRITE(dspcntr_reg, + temp | DISPLAY_PLANE_ENABLE); + /* Flush the plane changes */ + REG_WRITE(dspbase_reg, REG_READ(dspbase_reg)); + } + + udelay(150); + + /* Enable the pipe */ + temp = REG_READ(pipeconf_reg); + if ((temp & PIPEACONF_ENABLE) == 0) + REG_WRITE(pipeconf_reg, temp | PIPEACONF_ENABLE); + + psb_intel_crtc_load_lut(crtc); + + /* Give the overlay scaler a chance to enable + * if it's on this pipe */ + /* psb_intel_crtc_dpms_video(crtc, true); TODO */ + break; + case DRM_MODE_DPMS_OFF: + /* Give the overlay scaler a chance to disable + * if it's on this pipe */ + /* psb_intel_crtc_dpms_video(crtc, FALSE); TODO */ + + /* Disable the VGA plane that we never use */ + REG_WRITE(VGACNTRL, VGA_DISP_DISABLE); + + /* Jim Bish - changed pipe/plane here as well. */ + + /* Wait for vblank for the disable to take effect */ + cdv_intel_wait_for_vblank(dev); + + /* Next, disable display pipes */ + temp = REG_READ(pipeconf_reg); + if ((temp & PIPEACONF_ENABLE) != 0) { + REG_WRITE(pipeconf_reg, temp & ~PIPEACONF_ENABLE); + REG_READ(pipeconf_reg); + } + + /* Wait for vblank for the disable to take effect. */ + cdv_intel_wait_for_vblank(dev); + + udelay(150); + + /* Disable display plane */ + temp = REG_READ(dspcntr_reg); + if ((temp & DISPLAY_PLANE_ENABLE) != 0) { + REG_WRITE(dspcntr_reg, + temp & ~DISPLAY_PLANE_ENABLE); + /* Flush the plane changes */ + REG_WRITE(dspbase_reg, REG_READ(dspbase_reg)); + REG_READ(dspbase_reg); + } + + temp = REG_READ(dpll_reg); + if ((temp & DPLL_VCO_ENABLE) != 0) { + REG_WRITE(dpll_reg, temp & ~DPLL_VCO_ENABLE); + REG_READ(dpll_reg); + } + + /* Wait for the clocks to turn off. */ + udelay(150); + break; + } + /*Set FIFO Watermarks*/ + REG_WRITE(DSPARB, 0x3F3E); +} + +static void cdv_intel_crtc_prepare(struct drm_crtc *crtc) +{ + struct drm_crtc_helper_funcs *crtc_funcs = crtc->helper_private; + crtc_funcs->dpms(crtc, DRM_MODE_DPMS_OFF); +} + +static void cdv_intel_crtc_commit(struct drm_crtc *crtc) +{ + struct drm_crtc_helper_funcs *crtc_funcs = crtc->helper_private; + crtc_funcs->dpms(crtc, DRM_MODE_DPMS_ON); +} + +static bool cdv_intel_crtc_mode_fixup(struct drm_crtc *crtc, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + return true; +} + + +/** + * Return the pipe currently connected to the panel fitter, + * or -1 if the panel fitter is not present or not in use + */ +static int cdv_intel_panel_fitter_pipe(struct drm_device *dev) +{ + u32 pfit_control; + + pfit_control = REG_READ(PFIT_CONTROL); + + /* See if the panel fitter is in use */ + if ((pfit_control & PFIT_ENABLE) == 0) + return -1; + return (pfit_control >> 29) & 0x3; +} + +static int cdv_intel_crtc_mode_set(struct drm_crtc *crtc, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode, + int x, int y, + struct drm_framebuffer *old_fb) +{ + struct drm_device *dev = crtc->dev; + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + int pipe = psb_intel_crtc->pipe; + int dpll_reg = (pipe == 0) ? DPLL_A : DPLL_B; + int dpll_md_reg = (psb_intel_crtc->pipe == 0) ? DPLL_A_MD : DPLL_B_MD; + int dspcntr_reg = (pipe == 0) ? DSPACNTR : DSPBCNTR; + int pipeconf_reg = (pipe == 0) ? PIPEACONF : PIPEBCONF; + int htot_reg = (pipe == 0) ? HTOTAL_A : HTOTAL_B; + int hblank_reg = (pipe == 0) ? HBLANK_A : HBLANK_B; + int hsync_reg = (pipe == 0) ? HSYNC_A : HSYNC_B; + int vtot_reg = (pipe == 0) ? VTOTAL_A : VTOTAL_B; + int vblank_reg = (pipe == 0) ? VBLANK_A : VBLANK_B; + int vsync_reg = (pipe == 0) ? VSYNC_A : VSYNC_B; + int dspsize_reg = (pipe == 0) ? DSPASIZE : DSPBSIZE; + int dsppos_reg = (pipe == 0) ? DSPAPOS : DSPBPOS; + int pipesrc_reg = (pipe == 0) ? PIPEASRC : PIPEBSRC; + int refclk; + struct cdv_intel_clock_t clock; + u32 dpll = 0, dspcntr, pipeconf; + bool ok; + bool is_crt = false, is_lvds = false, is_tv = false; + bool is_hdmi = false; + struct drm_mode_config *mode_config = &dev->mode_config; + struct drm_connector *connector; + + list_for_each_entry(connector, &mode_config->connector_list, head) { + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(connector); + + if (!connector->encoder + || connector->encoder->crtc != crtc) + continue; + + switch (psb_intel_encoder->type) { + case INTEL_OUTPUT_LVDS: + is_lvds = true; + break; + case INTEL_OUTPUT_TVOUT: + is_tv = true; + break; + case INTEL_OUTPUT_ANALOG: + is_crt = true; + break; + case INTEL_OUTPUT_HDMI: + is_hdmi = true; + break; + } + } + + refclk = 96000; + + /* Hack selection about ref clk for CRT */ + /* Select 27MHz as the reference clk for HDMI */ + if (is_crt || is_hdmi) + refclk = 27000; + + drm_mode_debug_printmodeline(adjusted_mode); + + ok = cdv_intel_find_best_PLL(crtc, adjusted_mode->clock, refclk, + &clock); + if (!ok) { + dev_err(dev->dev, "Couldn't find PLL settings for mode!\n"); + return 0; + } + + dpll = DPLL_VGA_MODE_DIS; + if (is_tv) { + /* XXX: just matching BIOS for now */ +/* dpll |= PLL_REF_INPUT_TVCLKINBC; */ + dpll |= 3; + } + dpll |= PLL_REF_INPUT_DREFCLK; + + dpll |= DPLL_SYNCLOCK_ENABLE; + dpll |= DPLL_VGA_MODE_DIS; + if (is_lvds) + dpll |= DPLLB_MODE_LVDS; + else + dpll |= DPLLB_MODE_DAC_SERIAL; + /* dpll |= (2 << 11); */ + + /* setup pipeconf */ + pipeconf = REG_READ(pipeconf_reg); + + /* Set up the display plane register */ + dspcntr = DISPPLANE_GAMMA_ENABLE; + + if (pipe == 0) + dspcntr |= DISPPLANE_SEL_PIPE_A; + else + dspcntr |= DISPPLANE_SEL_PIPE_B; + + dspcntr |= DISPLAY_PLANE_ENABLE; + pipeconf |= PIPEACONF_ENABLE; + + REG_WRITE(dpll_reg, dpll | DPLL_VGA_MODE_DIS | DPLL_SYNCLOCK_ENABLE); + REG_READ(dpll_reg); + + cdv_dpll_set_clock_cdv(dev, crtc, &clock); + + udelay(150); + + + /* The LVDS pin pair needs to be on before the DPLLs are enabled. + * This is an exception to the general rule that mode_set doesn't turn + * things on. + */ + if (is_lvds) { + u32 lvds = REG_READ(LVDS); + + lvds |= + LVDS_PORT_EN | LVDS_A0A2_CLKA_POWER_UP | + LVDS_PIPEB_SELECT; + /* Set the B0-B3 data pairs corresponding to + * whether we're going to + * set the DPLLs for dual-channel mode or not. + */ + if (clock.p2 == 7) + lvds |= LVDS_B0B3_POWER_UP | LVDS_CLKB_POWER_UP; + else + lvds &= ~(LVDS_B0B3_POWER_UP | LVDS_CLKB_POWER_UP); + + /* It would be nice to set 24 vs 18-bit mode (LVDS_A3_POWER_UP) + * appropriately here, but we need to look more + * thoroughly into how panels behave in the two modes. + */ + + REG_WRITE(LVDS, lvds); + REG_READ(LVDS); + } + + dpll |= DPLL_VCO_ENABLE; + + /* Disable the panel fitter if it was on our pipe */ + if (cdv_intel_panel_fitter_pipe(dev) == pipe) + REG_WRITE(PFIT_CONTROL, 0); + + DRM_DEBUG_KMS("Mode for pipe %c:\n", pipe == 0 ? 'A' : 'B'); + drm_mode_debug_printmodeline(mode); + + REG_WRITE(dpll_reg, + (REG_READ(dpll_reg) & ~DPLL_LOCK) | DPLL_VCO_ENABLE); + REG_READ(dpll_reg); + /* Wait for the clocks to stabilize. */ + udelay(150); /* 42 usec w/o calibration, 110 with. rounded up. */ + + if (!(REG_READ(dpll_reg) & DPLL_LOCK)) { + dev_err(dev->dev, "Failed to get DPLL lock\n"); + return -EBUSY; + } + + { + int sdvo_pixel_multiply = adjusted_mode->clock / mode->clock; + REG_WRITE(dpll_md_reg, (0 << DPLL_MD_UDI_DIVIDER_SHIFT) | ((sdvo_pixel_multiply - 1) << DPLL_MD_UDI_MULTIPLIER_SHIFT)); + } + + REG_WRITE(htot_reg, (adjusted_mode->crtc_hdisplay - 1) | + ((adjusted_mode->crtc_htotal - 1) << 16)); + REG_WRITE(hblank_reg, (adjusted_mode->crtc_hblank_start - 1) | + ((adjusted_mode->crtc_hblank_end - 1) << 16)); + REG_WRITE(hsync_reg, (adjusted_mode->crtc_hsync_start - 1) | + ((adjusted_mode->crtc_hsync_end - 1) << 16)); + REG_WRITE(vtot_reg, (adjusted_mode->crtc_vdisplay - 1) | + ((adjusted_mode->crtc_vtotal - 1) << 16)); + REG_WRITE(vblank_reg, (adjusted_mode->crtc_vblank_start - 1) | + ((adjusted_mode->crtc_vblank_end - 1) << 16)); + REG_WRITE(vsync_reg, (adjusted_mode->crtc_vsync_start - 1) | + ((adjusted_mode->crtc_vsync_end - 1) << 16)); + /* pipesrc and dspsize control the size that is scaled from, + * which should always be the user's requested size. + */ + REG_WRITE(dspsize_reg, + ((mode->vdisplay - 1) << 16) | (mode->hdisplay - 1)); + REG_WRITE(dsppos_reg, 0); + REG_WRITE(pipesrc_reg, + ((mode->hdisplay - 1) << 16) | (mode->vdisplay - 1)); + REG_WRITE(pipeconf_reg, pipeconf); + REG_READ(pipeconf_reg); + + cdv_intel_wait_for_vblank(dev); + + REG_WRITE(dspcntr_reg, dspcntr); + + /* Flush the plane changes */ + { + struct drm_crtc_helper_funcs *crtc_funcs = + crtc->helper_private; + crtc_funcs->mode_set_base(crtc, x, y, old_fb); + } + + cdv_intel_wait_for_vblank(dev); + + return 0; +} + +/** Loads the palette/gamma unit for the CRTC with the prepared values */ +static void cdv_intel_crtc_load_lut(struct drm_crtc *crtc) +{ + struct drm_device *dev = crtc->dev; + struct drm_psb_private *dev_priv = + (struct drm_psb_private *)dev->dev_private; + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + int palreg = PALETTE_A; + int i; + + /* The clocks have to be on to load the palette. */ + if (!crtc->enabled) + return; + + switch (psb_intel_crtc->pipe) { + case 0: + break; + case 1: + palreg = PALETTE_B; + break; + case 2: + palreg = PALETTE_C; + break; + default: + dev_err(dev->dev, "Illegal Pipe Number.\n"); + return; + } + + if (gma_power_begin(dev, false)) { + for (i = 0; i < 256; i++) { + REG_WRITE(palreg + 4 * i, + ((psb_intel_crtc->lut_r[i] + + psb_intel_crtc->lut_adj[i]) << 16) | + ((psb_intel_crtc->lut_g[i] + + psb_intel_crtc->lut_adj[i]) << 8) | + (psb_intel_crtc->lut_b[i] + + psb_intel_crtc->lut_adj[i])); + } + gma_power_end(dev); + } else { + for (i = 0; i < 256; i++) { + dev_priv->regs.psb.save_palette_a[i] = + ((psb_intel_crtc->lut_r[i] + + psb_intel_crtc->lut_adj[i]) << 16) | + ((psb_intel_crtc->lut_g[i] + + psb_intel_crtc->lut_adj[i]) << 8) | + (psb_intel_crtc->lut_b[i] + + psb_intel_crtc->lut_adj[i]); + } + + } +} + +/** + * Save HW states of giving crtc + */ +static void cdv_intel_crtc_save(struct drm_crtc *crtc) +{ + struct drm_device *dev = crtc->dev; + /* struct drm_psb_private *dev_priv = + (struct drm_psb_private *)dev->dev_private; */ + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + struct psb_intel_crtc_state *crtc_state = psb_intel_crtc->crtc_state; + int pipeA = (psb_intel_crtc->pipe == 0); + uint32_t paletteReg; + int i; + + if (!crtc_state) { + dev_dbg(dev->dev, "No CRTC state found\n"); + return; + } + + crtc_state->saveDSPCNTR = REG_READ(pipeA ? DSPACNTR : DSPBCNTR); + crtc_state->savePIPECONF = REG_READ(pipeA ? PIPEACONF : PIPEBCONF); + crtc_state->savePIPESRC = REG_READ(pipeA ? PIPEASRC : PIPEBSRC); + crtc_state->saveFP0 = REG_READ(pipeA ? FPA0 : FPB0); + crtc_state->saveFP1 = REG_READ(pipeA ? FPA1 : FPB1); + crtc_state->saveDPLL = REG_READ(pipeA ? DPLL_A : DPLL_B); + crtc_state->saveHTOTAL = REG_READ(pipeA ? HTOTAL_A : HTOTAL_B); + crtc_state->saveHBLANK = REG_READ(pipeA ? HBLANK_A : HBLANK_B); + crtc_state->saveHSYNC = REG_READ(pipeA ? HSYNC_A : HSYNC_B); + crtc_state->saveVTOTAL = REG_READ(pipeA ? VTOTAL_A : VTOTAL_B); + crtc_state->saveVBLANK = REG_READ(pipeA ? VBLANK_A : VBLANK_B); + crtc_state->saveVSYNC = REG_READ(pipeA ? VSYNC_A : VSYNC_B); + crtc_state->saveDSPSTRIDE = REG_READ(pipeA ? DSPASTRIDE : DSPBSTRIDE); + + /*NOTE: DSPSIZE DSPPOS only for psb*/ + crtc_state->saveDSPSIZE = REG_READ(pipeA ? DSPASIZE : DSPBSIZE); + crtc_state->saveDSPPOS = REG_READ(pipeA ? DSPAPOS : DSPBPOS); + + crtc_state->saveDSPBASE = REG_READ(pipeA ? DSPABASE : DSPBBASE); + + DRM_DEBUG("(%x %x %x %x %x %x %x %x %x %x %x %x %x %x %x %x)\n", + crtc_state->saveDSPCNTR, + crtc_state->savePIPECONF, + crtc_state->savePIPESRC, + crtc_state->saveFP0, + crtc_state->saveFP1, + crtc_state->saveDPLL, + crtc_state->saveHTOTAL, + crtc_state->saveHBLANK, + crtc_state->saveHSYNC, + crtc_state->saveVTOTAL, + crtc_state->saveVBLANK, + crtc_state->saveVSYNC, + crtc_state->saveDSPSTRIDE, + crtc_state->saveDSPSIZE, + crtc_state->saveDSPPOS, + crtc_state->saveDSPBASE + ); + + paletteReg = pipeA ? PALETTE_A : PALETTE_B; + for (i = 0; i < 256; ++i) + crtc_state->savePalette[i] = REG_READ(paletteReg + (i << 2)); +} + +/** + * Restore HW states of giving crtc + */ +static void cdv_intel_crtc_restore(struct drm_crtc *crtc) +{ + struct drm_device *dev = crtc->dev; + /* struct drm_psb_private * dev_priv = + (struct drm_psb_private *)dev->dev_private; */ + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + struct psb_intel_crtc_state *crtc_state = psb_intel_crtc->crtc_state; + /* struct drm_crtc_helper_funcs * crtc_funcs = crtc->helper_private; */ + int pipeA = (psb_intel_crtc->pipe == 0); + uint32_t paletteReg; + int i; + + if (!crtc_state) { + dev_dbg(dev->dev, "No crtc state\n"); + return; + } + + DRM_DEBUG( + "current:(%x %x %x %x %x %x %x %x %x %x %x %x %x %x %x %x)\n", + REG_READ(pipeA ? DSPACNTR : DSPBCNTR), + REG_READ(pipeA ? PIPEACONF : PIPEBCONF), + REG_READ(pipeA ? PIPEASRC : PIPEBSRC), + REG_READ(pipeA ? FPA0 : FPB0), + REG_READ(pipeA ? FPA1 : FPB1), + REG_READ(pipeA ? DPLL_A : DPLL_B), + REG_READ(pipeA ? HTOTAL_A : HTOTAL_B), + REG_READ(pipeA ? HBLANK_A : HBLANK_B), + REG_READ(pipeA ? HSYNC_A : HSYNC_B), + REG_READ(pipeA ? VTOTAL_A : VTOTAL_B), + REG_READ(pipeA ? VBLANK_A : VBLANK_B), + REG_READ(pipeA ? VSYNC_A : VSYNC_B), + REG_READ(pipeA ? DSPASTRIDE : DSPBSTRIDE), + REG_READ(pipeA ? DSPASIZE : DSPBSIZE), + REG_READ(pipeA ? DSPAPOS : DSPBPOS), + REG_READ(pipeA ? DSPABASE : DSPBBASE) + ); + + DRM_DEBUG( + "saved: (%x %x %x %x %x %x %x %x %x %x %x %x %x %x %x %x)\n", + crtc_state->saveDSPCNTR, + crtc_state->savePIPECONF, + crtc_state->savePIPESRC, + crtc_state->saveFP0, + crtc_state->saveFP1, + crtc_state->saveDPLL, + crtc_state->saveHTOTAL, + crtc_state->saveHBLANK, + crtc_state->saveHSYNC, + crtc_state->saveVTOTAL, + crtc_state->saveVBLANK, + crtc_state->saveVSYNC, + crtc_state->saveDSPSTRIDE, + crtc_state->saveDSPSIZE, + crtc_state->saveDSPPOS, + crtc_state->saveDSPBASE + ); + + + if (crtc_state->saveDPLL & DPLL_VCO_ENABLE) { + REG_WRITE(pipeA ? DPLL_A : DPLL_B, + crtc_state->saveDPLL & ~DPLL_VCO_ENABLE); + REG_READ(pipeA ? DPLL_A : DPLL_B); + DRM_DEBUG("write dpll: %x\n", + REG_READ(pipeA ? DPLL_A : DPLL_B)); + udelay(150); + } + + REG_WRITE(pipeA ? FPA0 : FPB0, crtc_state->saveFP0); + REG_READ(pipeA ? FPA0 : FPB0); + + REG_WRITE(pipeA ? FPA1 : FPB1, crtc_state->saveFP1); + REG_READ(pipeA ? FPA1 : FPB1); + + REG_WRITE(pipeA ? DPLL_A : DPLL_B, crtc_state->saveDPLL); + REG_READ(pipeA ? DPLL_A : DPLL_B); + udelay(150); + + REG_WRITE(pipeA ? HTOTAL_A : HTOTAL_B, crtc_state->saveHTOTAL); + REG_WRITE(pipeA ? HBLANK_A : HBLANK_B, crtc_state->saveHBLANK); + REG_WRITE(pipeA ? HSYNC_A : HSYNC_B, crtc_state->saveHSYNC); + REG_WRITE(pipeA ? VTOTAL_A : VTOTAL_B, crtc_state->saveVTOTAL); + REG_WRITE(pipeA ? VBLANK_A : VBLANK_B, crtc_state->saveVBLANK); + REG_WRITE(pipeA ? VSYNC_A : VSYNC_B, crtc_state->saveVSYNC); + REG_WRITE(pipeA ? DSPASTRIDE : DSPBSTRIDE, crtc_state->saveDSPSTRIDE); + + REG_WRITE(pipeA ? DSPASIZE : DSPBSIZE, crtc_state->saveDSPSIZE); + REG_WRITE(pipeA ? DSPAPOS : DSPBPOS, crtc_state->saveDSPPOS); + + REG_WRITE(pipeA ? PIPEASRC : PIPEBSRC, crtc_state->savePIPESRC); + REG_WRITE(pipeA ? DSPABASE : DSPBBASE, crtc_state->saveDSPBASE); + REG_WRITE(pipeA ? PIPEACONF : PIPEBCONF, crtc_state->savePIPECONF); + + cdv_intel_wait_for_vblank(dev); + + REG_WRITE(pipeA ? DSPACNTR : DSPBCNTR, crtc_state->saveDSPCNTR); + REG_WRITE(pipeA ? DSPABASE : DSPBBASE, crtc_state->saveDSPBASE); + + cdv_intel_wait_for_vblank(dev); + + paletteReg = pipeA ? PALETTE_A : PALETTE_B; + for (i = 0; i < 256; ++i) + REG_WRITE(paletteReg + (i << 2), crtc_state->savePalette[i]); +} + +static int cdv_intel_crtc_cursor_set(struct drm_crtc *crtc, + struct drm_file *file_priv, + uint32_t handle, + uint32_t width, uint32_t height) +{ + struct drm_device *dev = crtc->dev; + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + int pipe = psb_intel_crtc->pipe; + uint32_t control = (pipe == 0) ? CURACNTR : CURBCNTR; + uint32_t base = (pipe == 0) ? CURABASE : CURBBASE; + uint32_t temp; + size_t addr = 0; + struct gtt_range *gt; + struct drm_gem_object *obj; + int ret; + + /* if we want to turn of the cursor ignore width and height */ + if (!handle) { + /* turn off the cursor */ + temp = CURSOR_MODE_DISABLE; + + if (gma_power_begin(dev, false)) { + REG_WRITE(control, temp); + REG_WRITE(base, 0); + gma_power_end(dev); + } + + /* unpin the old GEM object */ + if (psb_intel_crtc->cursor_obj) { + gt = container_of(psb_intel_crtc->cursor_obj, + struct gtt_range, gem); + psb_gtt_unpin(gt); + drm_gem_object_unreference(psb_intel_crtc->cursor_obj); + psb_intel_crtc->cursor_obj = NULL; + } + + return 0; + } + + /* Currently we only support 64x64 cursors */ + if (width != 64 || height != 64) { + dev_dbg(dev->dev, "we currently only support 64x64 cursors\n"); + return -EINVAL; + } + + obj = drm_gem_object_lookup(dev, file_priv, handle); + if (!obj) + return -ENOENT; + + if (obj->size < width * height * 4) { + dev_dbg(dev->dev, "buffer is to small\n"); + return -ENOMEM; + } + + gt = container_of(obj, struct gtt_range, gem); + + /* Pin the memory into the GTT */ + ret = psb_gtt_pin(gt); + if (ret) { + dev_err(dev->dev, "Can not pin down handle 0x%x\n", handle); + return ret; + } + + addr = gt->offset; /* Or resource.start ??? */ + + psb_intel_crtc->cursor_addr = addr; + + temp = 0; + /* set the pipe for the cursor */ + temp |= (pipe << 28); + temp |= CURSOR_MODE_64_ARGB_AX | MCURSOR_GAMMA_ENABLE; + + if (gma_power_begin(dev, false)) { + REG_WRITE(control, temp); + REG_WRITE(base, addr); + gma_power_end(dev); + } + + /* unpin the old GEM object */ + if (psb_intel_crtc->cursor_obj) { + gt = container_of(psb_intel_crtc->cursor_obj, + struct gtt_range, gem); + psb_gtt_unpin(gt); + drm_gem_object_unreference(psb_intel_crtc->cursor_obj); + psb_intel_crtc->cursor_obj = obj; + } + return 0; +} + +static int cdv_intel_crtc_cursor_move(struct drm_crtc *crtc, int x, int y) +{ + struct drm_device *dev = crtc->dev; + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + int pipe = psb_intel_crtc->pipe; + uint32_t temp = 0; + uint32_t adder; + + + if (x < 0) { + temp |= (CURSOR_POS_SIGN << CURSOR_X_SHIFT); + x = -x; + } + if (y < 0) { + temp |= (CURSOR_POS_SIGN << CURSOR_Y_SHIFT); + y = -y; + } + + temp |= ((x & CURSOR_POS_MASK) << CURSOR_X_SHIFT); + temp |= ((y & CURSOR_POS_MASK) << CURSOR_Y_SHIFT); + + adder = psb_intel_crtc->cursor_addr; + + if (gma_power_begin(dev, false)) { + REG_WRITE((pipe == 0) ? CURAPOS : CURBPOS, temp); + REG_WRITE((pipe == 0) ? CURABASE : CURBBASE, adder); + gma_power_end(dev); + } + return 0; +} + +static void cdv_intel_crtc_gamma_set(struct drm_crtc *crtc, u16 *red, + u16 *green, u16 *blue, uint32_t start, uint32_t size) +{ + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + int i; + int end = (start + size > 256) ? 256 : start + size; + + for (i = start; i < end; i++) { + psb_intel_crtc->lut_r[i] = red[i] >> 8; + psb_intel_crtc->lut_g[i] = green[i] >> 8; + psb_intel_crtc->lut_b[i] = blue[i] >> 8; + } + + cdv_intel_crtc_load_lut(crtc); +} + +static int cdv_crtc_set_config(struct drm_mode_set *set) +{ + int ret = 0; + struct drm_device *dev = set->crtc->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + + if (!dev_priv->rpm_enabled) + return drm_crtc_helper_set_config(set); + + pm_runtime_forbid(&dev->pdev->dev); + + ret = drm_crtc_helper_set_config(set); + + pm_runtime_allow(&dev->pdev->dev); + + return ret; +} + +/** Derive the pixel clock for the given refclk and divisors for 8xx chips. */ + +/* FIXME: why are we using this, should it be cdv_ in this tree ? */ + +static void i8xx_clock(int refclk, struct cdv_intel_clock_t *clock) +{ + clock->m = 5 * (clock->m1 + 2) + (clock->m2 + 2); + clock->p = clock->p1 * clock->p2; + clock->vco = refclk * clock->m / (clock->n + 2); + clock->dot = clock->vco / clock->p; +} + +/* Returns the clock of the currently programmed mode of the given pipe. */ +static int cdv_intel_crtc_clock_get(struct drm_device *dev, + struct drm_crtc *crtc) +{ + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + int pipe = psb_intel_crtc->pipe; + u32 dpll; + u32 fp; + struct cdv_intel_clock_t clock; + bool is_lvds; + struct drm_psb_private *dev_priv = dev->dev_private; + + if (gma_power_begin(dev, false)) { + dpll = REG_READ((pipe == 0) ? DPLL_A : DPLL_B); + if ((dpll & DISPLAY_RATE_SELECT_FPA1) == 0) + fp = REG_READ((pipe == 0) ? FPA0 : FPB0); + else + fp = REG_READ((pipe == 0) ? FPA1 : FPB1); + is_lvds = (pipe == 1) && (REG_READ(LVDS) & LVDS_PORT_EN); + gma_power_end(dev); + } else { + dpll = (pipe == 0) ? + dev_priv->regs.psb.saveDPLL_A : + dev_priv->regs.psb.saveDPLL_B; + + if ((dpll & DISPLAY_RATE_SELECT_FPA1) == 0) + fp = (pipe == 0) ? + dev_priv->regs.psb.saveFPA0 : + dev_priv->regs.psb.saveFPB0; + else + fp = (pipe == 0) ? + dev_priv->regs.psb.saveFPA1 : + dev_priv->regs.psb.saveFPB1; + + is_lvds = (pipe == 1) && + (dev_priv->regs.psb.saveLVDS & LVDS_PORT_EN); + } + + clock.m1 = (fp & FP_M1_DIV_MASK) >> FP_M1_DIV_SHIFT; + clock.m2 = (fp & FP_M2_DIV_MASK) >> FP_M2_DIV_SHIFT; + clock.n = (fp & FP_N_DIV_MASK) >> FP_N_DIV_SHIFT; + + if (is_lvds) { + clock.p1 = + ffs((dpll & + DPLL_FPA01_P1_POST_DIV_MASK_I830_LVDS) >> + DPLL_FPA01_P1_POST_DIV_SHIFT); + if (clock.p1 == 0) { + clock.p1 = 4; + dev_err(dev->dev, "PLL %d\n", dpll); + } + clock.p2 = 14; + + if ((dpll & PLL_REF_INPUT_MASK) == + PLLB_REF_INPUT_SPREADSPECTRUMIN) { + /* XXX: might not be 66MHz */ + i8xx_clock(66000, &clock); + } else + i8xx_clock(48000, &clock); + } else { + if (dpll & PLL_P1_DIVIDE_BY_TWO) + clock.p1 = 2; + else { + clock.p1 = + ((dpll & + DPLL_FPA01_P1_POST_DIV_MASK_I830) >> + DPLL_FPA01_P1_POST_DIV_SHIFT) + 2; + } + if (dpll & PLL_P2_DIVIDE_BY_4) + clock.p2 = 4; + else + clock.p2 = 2; + + i8xx_clock(48000, &clock); + } + + /* XXX: It would be nice to validate the clocks, but we can't reuse + * i830PllIsValid() because it relies on the xf86_config connector + * configuration being accurate, which it isn't necessarily. + */ + + return clock.dot; +} + +/** Returns the currently programmed mode of the given pipe. */ +struct drm_display_mode *cdv_intel_crtc_mode_get(struct drm_device *dev, + struct drm_crtc *crtc) +{ + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + int pipe = psb_intel_crtc->pipe; + struct drm_display_mode *mode; + int htot; + int hsync; + int vtot; + int vsync; + struct drm_psb_private *dev_priv = dev->dev_private; + + if (gma_power_begin(dev, false)) { + htot = REG_READ((pipe == 0) ? HTOTAL_A : HTOTAL_B); + hsync = REG_READ((pipe == 0) ? HSYNC_A : HSYNC_B); + vtot = REG_READ((pipe == 0) ? VTOTAL_A : VTOTAL_B); + vsync = REG_READ((pipe == 0) ? VSYNC_A : VSYNC_B); + gma_power_end(dev); + } else { + htot = (pipe == 0) ? + dev_priv->regs.psb.saveHTOTAL_A : + dev_priv->regs.psb.saveHTOTAL_B; + hsync = (pipe == 0) ? + dev_priv->regs.psb.saveHSYNC_A : + dev_priv->regs.psb.saveHSYNC_B; + vtot = (pipe == 0) ? + dev_priv->regs.psb.saveVTOTAL_A : + dev_priv->regs.psb.saveVTOTAL_B; + vsync = (pipe == 0) ? + dev_priv->regs.psb.saveVSYNC_A : + dev_priv->regs.psb.saveVSYNC_B; + } + + mode = kzalloc(sizeof(*mode), GFP_KERNEL); + if (!mode) + return NULL; + + mode->clock = cdv_intel_crtc_clock_get(dev, crtc); + mode->hdisplay = (htot & 0xffff) + 1; + mode->htotal = ((htot & 0xffff0000) >> 16) + 1; + mode->hsync_start = (hsync & 0xffff) + 1; + mode->hsync_end = ((hsync & 0xffff0000) >> 16) + 1; + mode->vdisplay = (vtot & 0xffff) + 1; + mode->vtotal = ((vtot & 0xffff0000) >> 16) + 1; + mode->vsync_start = (vsync & 0xffff) + 1; + mode->vsync_end = ((vsync & 0xffff0000) >> 16) + 1; + + drm_mode_set_name(mode); + drm_mode_set_crtcinfo(mode, 0); + + return mode; +} + +static void cdv_intel_crtc_destroy(struct drm_crtc *crtc) +{ + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + + kfree(psb_intel_crtc->crtc_state); + drm_crtc_cleanup(crtc); + kfree(psb_intel_crtc); +} + +const struct drm_crtc_helper_funcs cdv_intel_helper_funcs = { + .dpms = cdv_intel_crtc_dpms, + .mode_fixup = cdv_intel_crtc_mode_fixup, + .mode_set = cdv_intel_crtc_mode_set, + .mode_set_base = cdv_intel_pipe_set_base, + .prepare = cdv_intel_crtc_prepare, + .commit = cdv_intel_crtc_commit, +}; + +const struct drm_crtc_funcs cdv_intel_crtc_funcs = { + .save = cdv_intel_crtc_save, + .restore = cdv_intel_crtc_restore, + .cursor_set = cdv_intel_crtc_cursor_set, + .cursor_move = cdv_intel_crtc_cursor_move, + .gamma_set = cdv_intel_crtc_gamma_set, + .set_config = cdv_crtc_set_config, + .destroy = cdv_intel_crtc_destroy, +}; diff --git a/drivers/gpu/drm/gma500/cdv_intel_hdmi.c b/drivers/gpu/drm/gma500/cdv_intel_hdmi.c new file mode 100644 index 0000000..8d52695 --- /dev/null +++ b/drivers/gpu/drm/gma500/cdv_intel_hdmi.c @@ -0,0 +1,393 @@ +/* + * Copyright © 2006-2011 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * jim liu + * + * FIXME: + * We should probably make this generic and share it with Medfield + */ + +#include +#include +#include +#include +#include "psb_intel_drv.h" +#include "psb_drv.h" +#include "psb_intel_reg.h" +#include "cdv_device.h" +#include + +/* hdmi control bits */ +#define HDMI_NULL_PACKETS_DURING_VSYNC (1 << 9) +#define HDMI_BORDER_ENABLE (1 << 7) +#define HDMI_AUDIO_ENABLE (1 << 6) +#define HDMI_VSYNC_ACTIVE_HIGH (1 << 4) +#define HDMI_HSYNC_ACTIVE_HIGH (1 << 3) +/* hdmi-b control bits */ +#define HDMIB_PIPE_B_SELECT (1 << 30) + + +struct mid_intel_hdmi_priv { + u32 hdmi_reg; + u32 save_HDMIB; + bool has_hdmi_sink; + bool has_hdmi_audio; + /* Should set this when detect hotplug */ + bool hdmi_device_connected; + struct mdfld_hdmi_i2c *i2c_bus; + struct i2c_adapter *hdmi_i2c_adapter; /* for control functions */ + struct drm_device *dev; +}; + +static void cdv_hdmi_mode_set(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + struct drm_device *dev = encoder->dev; + struct psb_intel_encoder *psb_intel_encoder = to_psb_intel_encoder(encoder); + struct mid_intel_hdmi_priv *hdmi_priv = psb_intel_encoder->dev_priv; + u32 hdmib; + struct drm_crtc *crtc = encoder->crtc; + struct psb_intel_crtc *intel_crtc = to_psb_intel_crtc(crtc); + + hdmib = (2 << 10); + + if (adjusted_mode->flags & DRM_MODE_FLAG_PVSYNC) + hdmib |= HDMI_VSYNC_ACTIVE_HIGH; + if (adjusted_mode->flags & DRM_MODE_FLAG_PHSYNC) + hdmib |= HDMI_HSYNC_ACTIVE_HIGH; + + if (intel_crtc->pipe == 1) + hdmib |= HDMIB_PIPE_B_SELECT; + + if (hdmi_priv->has_hdmi_audio) { + hdmib |= HDMI_AUDIO_ENABLE; + hdmib |= HDMI_NULL_PACKETS_DURING_VSYNC; + } + + REG_WRITE(hdmi_priv->hdmi_reg, hdmib); + REG_READ(hdmi_priv->hdmi_reg); +} + +static bool cdv_hdmi_mode_fixup(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + return true; +} + +static void cdv_hdmi_dpms(struct drm_encoder *encoder, int mode) +{ + struct drm_device *dev = encoder->dev; + struct psb_intel_encoder *psb_intel_encoder = + to_psb_intel_encoder(encoder); + struct mid_intel_hdmi_priv *hdmi_priv = psb_intel_encoder->dev_priv; + u32 hdmib; + + hdmib = REG_READ(hdmi_priv->hdmi_reg); + + if (mode != DRM_MODE_DPMS_ON) + REG_WRITE(hdmi_priv->hdmi_reg, hdmib & ~HDMIB_PORT_EN); + else + REG_WRITE(hdmi_priv->hdmi_reg, hdmib | HDMIB_PORT_EN); + REG_READ(hdmi_priv->hdmi_reg); +} + +static void cdv_hdmi_save(struct drm_connector *connector) +{ + struct drm_device *dev = connector->dev; + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(connector); + struct mid_intel_hdmi_priv *hdmi_priv = psb_intel_encoder->dev_priv; + + hdmi_priv->save_HDMIB = REG_READ(hdmi_priv->hdmi_reg); +} + +static void cdv_hdmi_restore(struct drm_connector *connector) +{ + struct drm_device *dev = connector->dev; + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(connector); + struct mid_intel_hdmi_priv *hdmi_priv = psb_intel_encoder->dev_priv; + + REG_WRITE(hdmi_priv->hdmi_reg, hdmi_priv->save_HDMIB); + REG_READ(hdmi_priv->hdmi_reg); +} + +static enum drm_connector_status cdv_hdmi_detect( + struct drm_connector *connector, bool force) +{ + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(connector); + struct psb_intel_connector *psb_intel_connector = + to_psb_intel_connector(connector); + struct mid_intel_hdmi_priv *hdmi_priv = psb_intel_encoder->dev_priv; + struct edid *edid = NULL; + enum drm_connector_status status = connector_status_disconnected; + + edid = drm_get_edid(connector, &psb_intel_encoder->i2c_bus->adapter); + + hdmi_priv->has_hdmi_sink = false; + hdmi_priv->has_hdmi_audio = false; + if (edid) { + if (edid->input & DRM_EDID_INPUT_DIGITAL) { + status = connector_status_connected; + hdmi_priv->has_hdmi_sink = + drm_detect_hdmi_monitor(edid); + hdmi_priv->has_hdmi_audio = + drm_detect_monitor_audio(edid); + } + + psb_intel_connector->base.display_info.raw_edid = NULL; + kfree(edid); + } + return status; +} + +static int cdv_hdmi_set_property(struct drm_connector *connector, + struct drm_property *property, + uint64_t value) +{ + struct drm_encoder *encoder = connector->encoder; + + if (!strcmp(property->name, "scaling mode") && encoder) { + struct psb_intel_crtc *crtc = to_psb_intel_crtc(encoder->crtc); + bool centre; + uint64_t curValue; + + if (!crtc) + return -1; + + switch (value) { + case DRM_MODE_SCALE_FULLSCREEN: + break; + case DRM_MODE_SCALE_NO_SCALE: + break; + case DRM_MODE_SCALE_ASPECT: + break; + default: + return -1; + } + + if (drm_connector_property_get_value(connector, + property, &curValue)) + return -1; + + if (curValue == value) + return 0; + + if (drm_connector_property_set_value(connector, + property, value)) + return -1; + + centre = (curValue == DRM_MODE_SCALE_NO_SCALE) || + (value == DRM_MODE_SCALE_NO_SCALE); + + if (crtc->saved_mode.hdisplay != 0 && + crtc->saved_mode.vdisplay != 0) { + if (centre) { + if (!drm_crtc_helper_set_mode(encoder->crtc, &crtc->saved_mode, + encoder->crtc->x, encoder->crtc->y, encoder->crtc->fb)) + return -1; + } else { + struct drm_encoder_helper_funcs *helpers + = encoder->helper_private; + helpers->mode_set(encoder, &crtc->saved_mode, + &crtc->saved_adjusted_mode); + } + } + } + return 0; +} + +/* + * Return the list of HDMI DDC modes if available. + */ +static int cdv_hdmi_get_modes(struct drm_connector *connector) +{ + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(connector); + struct edid *edid = NULL; + int ret = 0; + + edid = drm_get_edid(connector, &psb_intel_encoder->i2c_bus->adapter); + if (edid) { + drm_mode_connector_update_edid_property(connector, edid); + ret = drm_add_edid_modes(connector, edid); + kfree(edid); + } + return ret; +} + +static int cdv_hdmi_mode_valid(struct drm_connector *connector, + struct drm_display_mode *mode) +{ + struct drm_psb_private *dev_priv = connector->dev->dev_private; + + if (mode->clock > 165000) + return MODE_CLOCK_HIGH; + if (mode->clock < 20000) + return MODE_CLOCK_HIGH; + + /* just in case */ + if (mode->flags & DRM_MODE_FLAG_DBLSCAN) + return MODE_NO_DBLESCAN; + + /* just in case */ + if (mode->flags & DRM_MODE_FLAG_INTERLACE) + return MODE_NO_INTERLACE; + + /* We assume worst case scenario of 32 bpp here, since we don't know */ + if ((ALIGN(mode->hdisplay * 4, 64) * mode->vdisplay) > + dev_priv->vram_stolen_size) + return MODE_MEM; + + return MODE_OK; +} + +static void cdv_hdmi_destroy(struct drm_connector *connector) +{ + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(connector); + + if (psb_intel_encoder->i2c_bus) + psb_intel_i2c_destroy(psb_intel_encoder->i2c_bus); + drm_sysfs_connector_remove(connector); + drm_connector_cleanup(connector); + kfree(connector); +} + +static const struct drm_encoder_helper_funcs cdv_hdmi_helper_funcs = { + .dpms = cdv_hdmi_dpms, + .mode_fixup = cdv_hdmi_mode_fixup, + .prepare = psb_intel_encoder_prepare, + .mode_set = cdv_hdmi_mode_set, + .commit = psb_intel_encoder_commit, +}; + +static const struct drm_connector_helper_funcs + cdv_hdmi_connector_helper_funcs = { + .get_modes = cdv_hdmi_get_modes, + .mode_valid = cdv_hdmi_mode_valid, + .best_encoder = psb_intel_best_encoder, +}; + +static const struct drm_connector_funcs cdv_hdmi_connector_funcs = { + .dpms = drm_helper_connector_dpms, + .save = cdv_hdmi_save, + .restore = cdv_hdmi_restore, + .detect = cdv_hdmi_detect, + .fill_modes = drm_helper_probe_single_connector_modes, + .set_property = cdv_hdmi_set_property, + .destroy = cdv_hdmi_destroy, +}; + +void cdv_hdmi_init(struct drm_device *dev, + struct psb_intel_mode_device *mode_dev, int reg) +{ + struct psb_intel_encoder *psb_intel_encoder; + struct psb_intel_connector *psb_intel_connector; + struct drm_connector *connector; + struct drm_encoder *encoder; + struct mid_intel_hdmi_priv *hdmi_priv; + int ddc_bus; + + psb_intel_encoder = kzalloc(sizeof(struct psb_intel_encoder), + GFP_KERNEL); + + if (!psb_intel_encoder) + return; + + psb_intel_connector = kzalloc(sizeof(struct psb_intel_connector), + GFP_KERNEL); + + if (!psb_intel_connector) + goto err_connector; + + hdmi_priv = kzalloc(sizeof(struct mid_intel_hdmi_priv), GFP_KERNEL); + + if (!hdmi_priv) + goto err_priv; + + connector = &psb_intel_connector->base; + encoder = &psb_intel_encoder->base; + drm_connector_init(dev, connector, + &cdv_hdmi_connector_funcs, + DRM_MODE_CONNECTOR_DVID); + + drm_encoder_init(dev, encoder, &psb_intel_lvds_enc_funcs, + DRM_MODE_ENCODER_TMDS); + + psb_intel_connector_attach_encoder(psb_intel_connector, + psb_intel_encoder); + psb_intel_encoder->type = INTEL_OUTPUT_HDMI; + hdmi_priv->hdmi_reg = reg; + hdmi_priv->has_hdmi_sink = false; + psb_intel_encoder->dev_priv = hdmi_priv; + + drm_encoder_helper_add(encoder, &cdv_hdmi_helper_funcs); + drm_connector_helper_add(connector, + &cdv_hdmi_connector_helper_funcs); + connector->display_info.subpixel_order = SubPixelHorizontalRGB; + connector->interlace_allowed = false; + connector->doublescan_allowed = false; + + drm_connector_attach_property(connector, + dev->mode_config.scaling_mode_property, + DRM_MODE_SCALE_FULLSCREEN); + + switch (reg) { + case SDVOB: + ddc_bus = GPIOE; + break; + case SDVOC: + ddc_bus = GPIOD; + break; + default: + DRM_ERROR("unknown reg 0x%x for HDMI\n", reg); + goto failed_ddc; + break; + } + + psb_intel_encoder->i2c_bus = psb_intel_i2c_create(dev, + ddc_bus, (reg == SDVOB) ? "HDMIB" : "HDMIC"); + + if (!psb_intel_encoder->i2c_bus) { + dev_err(dev->dev, "No ddc adapter available!\n"); + goto failed_ddc; + } + + hdmi_priv->hdmi_i2c_adapter = + &(psb_intel_encoder->i2c_bus->adapter); + hdmi_priv->dev = dev; + drm_sysfs_connector_add(connector); + return; + +failed_ddc: + drm_encoder_cleanup(encoder); + drm_connector_cleanup(connector); +err_priv: + kfree(psb_intel_connector); +err_connector: + kfree(psb_intel_encoder); +} diff --git a/drivers/gpu/drm/gma500/cdv_intel_lvds.c b/drivers/gpu/drm/gma500/cdv_intel_lvds.c new file mode 100644 index 0000000..8359c1a --- /dev/null +++ b/drivers/gpu/drm/gma500/cdv_intel_lvds.c @@ -0,0 +1,734 @@ +/* + * Copyright © 2006-2011 Intel Corporation + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + * Authors: + * Eric Anholt + * Dave Airlie + * Jesse Barnes + */ + +#include +#include +#include + +#include "intel_bios.h" +#include "psb_drv.h" +#include "psb_intel_drv.h" +#include "psb_intel_reg.h" +#include "power.h" +#include +#include "cdv_device.h" + +/** + * LVDS I2C backlight control macros + */ +#define BRIGHTNESS_MAX_LEVEL 100 +#define BRIGHTNESS_MASK 0xFF +#define BLC_I2C_TYPE 0x01 +#define BLC_PWM_TYPT 0x02 + +#define BLC_POLARITY_NORMAL 0 +#define BLC_POLARITY_INVERSE 1 + +#define PSB_BLC_MAX_PWM_REG_FREQ (0xFFFE) +#define PSB_BLC_MIN_PWM_REG_FREQ (0x2) +#define PSB_BLC_PWM_PRECISION_FACTOR (10) +#define PSB_BACKLIGHT_PWM_CTL_SHIFT (16) +#define PSB_BACKLIGHT_PWM_POLARITY_BIT_CLEAR (0xFFFE) + +struct cdv_intel_lvds_priv { + /** + * Saved LVDO output states + */ + uint32_t savePP_ON; + uint32_t savePP_OFF; + uint32_t saveLVDS; + uint32_t savePP_CONTROL; + uint32_t savePP_CYCLE; + uint32_t savePFIT_CONTROL; + uint32_t savePFIT_PGM_RATIOS; + uint32_t saveBLC_PWM_CTL; +}; + +/* + * Returns the maximum level of the backlight duty cycle field. + */ +static u32 cdv_intel_lvds_get_max_backlight(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + u32 retval; + + if (gma_power_begin(dev, false)) { + retval = ((REG_READ(BLC_PWM_CTL) & + BACKLIGHT_MODULATION_FREQ_MASK) >> + BACKLIGHT_MODULATION_FREQ_SHIFT) * 2; + + gma_power_end(dev); + } else + retval = ((dev_priv->regs.saveBLC_PWM_CTL & + BACKLIGHT_MODULATION_FREQ_MASK) >> + BACKLIGHT_MODULATION_FREQ_SHIFT) * 2; + + return retval; +} + +#if 0 +/* + * Set LVDS backlight level by I2C command + */ +static int cdv_lvds_i2c_set_brightness(struct drm_device *dev, + unsigned int level) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_intel_i2c_chan *lvds_i2c_bus = dev_priv->lvds_i2c_bus; + u8 out_buf[2]; + unsigned int blc_i2c_brightness; + + struct i2c_msg msgs[] = { + { + .addr = lvds_i2c_bus->slave_addr, + .flags = 0, + .len = 2, + .buf = out_buf, + } + }; + + blc_i2c_brightness = BRIGHTNESS_MASK & ((unsigned int)level * + BRIGHTNESS_MASK / + BRIGHTNESS_MAX_LEVEL); + + if (dev_priv->lvds_bl->pol == BLC_POLARITY_INVERSE) + blc_i2c_brightness = BRIGHTNESS_MASK - blc_i2c_brightness; + + out_buf[0] = dev_priv->lvds_bl->brightnesscmd; + out_buf[1] = (u8)blc_i2c_brightness; + + if (i2c_transfer(&lvds_i2c_bus->adapter, msgs, 1) == 1) + return 0; + + DRM_ERROR("I2C transfer error\n"); + return -1; +} + + +static int cdv_lvds_pwm_set_brightness(struct drm_device *dev, int level) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + + u32 max_pwm_blc; + u32 blc_pwm_duty_cycle; + + max_pwm_blc = cdv_intel_lvds_get_max_backlight(dev); + + /*BLC_PWM_CTL Should be initiated while backlight device init*/ + BUG_ON((max_pwm_blc & PSB_BLC_MAX_PWM_REG_FREQ) == 0); + + blc_pwm_duty_cycle = level * max_pwm_blc / BRIGHTNESS_MAX_LEVEL; + + if (dev_priv->lvds_bl->pol == BLC_POLARITY_INVERSE) + blc_pwm_duty_cycle = max_pwm_blc - blc_pwm_duty_cycle; + + blc_pwm_duty_cycle &= PSB_BACKLIGHT_PWM_POLARITY_BIT_CLEAR; + REG_WRITE(BLC_PWM_CTL, + (max_pwm_blc << PSB_BACKLIGHT_PWM_CTL_SHIFT) | + (blc_pwm_duty_cycle)); + + return 0; +} + +/* + * Set LVDS backlight level either by I2C or PWM + */ +void cdv_intel_lvds_set_brightness(struct drm_device *dev, int level) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + + if (!dev_priv->lvds_bl) { + DRM_ERROR("NO LVDS Backlight Info\n"); + return; + } + + if (dev_priv->lvds_bl->type == BLC_I2C_TYPE) + cdv_lvds_i2c_set_brightness(dev, level); + else + cdv_lvds_pwm_set_brightness(dev, level); +} +#endif + +/** + * Sets the backlight level. + * + * level backlight level, from 0 to cdv_intel_lvds_get_max_backlight(). + */ +static void cdv_intel_lvds_set_backlight(struct drm_device *dev, int level) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + u32 blc_pwm_ctl; + + if (gma_power_begin(dev, false)) { + blc_pwm_ctl = + REG_READ(BLC_PWM_CTL) & ~BACKLIGHT_DUTY_CYCLE_MASK; + REG_WRITE(BLC_PWM_CTL, + (blc_pwm_ctl | + (level << BACKLIGHT_DUTY_CYCLE_SHIFT))); + gma_power_end(dev); + } else { + blc_pwm_ctl = dev_priv->regs.saveBLC_PWM_CTL & + ~BACKLIGHT_DUTY_CYCLE_MASK; + dev_priv->regs.saveBLC_PWM_CTL = (blc_pwm_ctl | + (level << BACKLIGHT_DUTY_CYCLE_SHIFT)); + } +} + +/** + * Sets the power state for the panel. + */ +static void cdv_intel_lvds_set_power(struct drm_device *dev, + struct drm_encoder *encoder, bool on) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + u32 pp_status; + + if (!gma_power_begin(dev, true)) + return; + + if (on) { + REG_WRITE(PP_CONTROL, REG_READ(PP_CONTROL) | + POWER_TARGET_ON); + do { + pp_status = REG_READ(PP_STATUS); + } while ((pp_status & PP_ON) == 0); + + cdv_intel_lvds_set_backlight(dev, + dev_priv->mode_dev.backlight_duty_cycle); + } else { + cdv_intel_lvds_set_backlight(dev, 0); + + REG_WRITE(PP_CONTROL, REG_READ(PP_CONTROL) & + ~POWER_TARGET_ON); + do { + pp_status = REG_READ(PP_STATUS); + } while (pp_status & PP_ON); + } + gma_power_end(dev); +} + +static void cdv_intel_lvds_encoder_dpms(struct drm_encoder *encoder, int mode) +{ + struct drm_device *dev = encoder->dev; + if (mode == DRM_MODE_DPMS_ON) + cdv_intel_lvds_set_power(dev, encoder, true); + else + cdv_intel_lvds_set_power(dev, encoder, false); + /* XXX: We never power down the LVDS pairs. */ +} + +static void cdv_intel_lvds_save(struct drm_connector *connector) +{ +} + +static void cdv_intel_lvds_restore(struct drm_connector *connector) +{ +} + +static int cdv_intel_lvds_mode_valid(struct drm_connector *connector, + struct drm_display_mode *mode) +{ + struct drm_device *dev = connector->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + struct drm_display_mode *fixed_mode = + dev_priv->mode_dev.panel_fixed_mode; + + /* just in case */ + if (mode->flags & DRM_MODE_FLAG_DBLSCAN) + return MODE_NO_DBLESCAN; + + /* just in case */ + if (mode->flags & DRM_MODE_FLAG_INTERLACE) + return MODE_NO_INTERLACE; + + if (fixed_mode) { + if (mode->hdisplay > fixed_mode->hdisplay) + return MODE_PANEL; + if (mode->vdisplay > fixed_mode->vdisplay) + return MODE_PANEL; + } + return MODE_OK; +} + +static bool cdv_intel_lvds_mode_fixup(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + struct drm_device *dev = encoder->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_intel_mode_device *mode_dev = &dev_priv->mode_dev; + struct drm_encoder *tmp_encoder; + struct drm_display_mode *panel_fixed_mode = mode_dev->panel_fixed_mode; + + /* Should never happen!! */ + list_for_each_entry(tmp_encoder, &dev->mode_config.encoder_list, + head) { + if (tmp_encoder != encoder + && tmp_encoder->crtc == encoder->crtc) { + printk(KERN_ERR "Can't enable LVDS and another " + "encoder on the same pipe\n"); + return false; + } + } + + /* + * If we have timings from the BIOS for the panel, put them in + * to the adjusted mode. The CRTC will be set up for this mode, + * with the panel scaling set up to source from the H/VDisplay + * of the original mode. + */ + if (panel_fixed_mode != NULL) { + adjusted_mode->hdisplay = panel_fixed_mode->hdisplay; + adjusted_mode->hsync_start = panel_fixed_mode->hsync_start; + adjusted_mode->hsync_end = panel_fixed_mode->hsync_end; + adjusted_mode->htotal = panel_fixed_mode->htotal; + adjusted_mode->vdisplay = panel_fixed_mode->vdisplay; + adjusted_mode->vsync_start = panel_fixed_mode->vsync_start; + adjusted_mode->vsync_end = panel_fixed_mode->vsync_end; + adjusted_mode->vtotal = panel_fixed_mode->vtotal; + adjusted_mode->clock = panel_fixed_mode->clock; + drm_mode_set_crtcinfo(adjusted_mode, + CRTC_INTERLACE_HALVE_V); + } + + /* + * XXX: It would be nice to support lower refresh rates on the + * panels to reduce power consumption, and perhaps match the + * user's requested refresh rate. + */ + + return true; +} + +static void cdv_intel_lvds_prepare(struct drm_encoder *encoder) +{ + struct drm_device *dev = encoder->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_intel_mode_device *mode_dev = &dev_priv->mode_dev; + + if (!gma_power_begin(dev, true)) + return; + + mode_dev->saveBLC_PWM_CTL = REG_READ(BLC_PWM_CTL); + mode_dev->backlight_duty_cycle = (mode_dev->saveBLC_PWM_CTL & + BACKLIGHT_DUTY_CYCLE_MASK); + + cdv_intel_lvds_set_power(dev, encoder, false); + + gma_power_end(dev); +} + +static void cdv_intel_lvds_commit(struct drm_encoder *encoder) +{ + struct drm_device *dev = encoder->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_intel_mode_device *mode_dev = &dev_priv->mode_dev; + + if (mode_dev->backlight_duty_cycle == 0) + mode_dev->backlight_duty_cycle = + cdv_intel_lvds_get_max_backlight(dev); + + cdv_intel_lvds_set_power(dev, encoder, true); +} + +static void cdv_intel_lvds_mode_set(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + struct drm_device *dev = encoder->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + u32 pfit_control; + + /* + * The LVDS pin pair will already have been turned on in the + * cdv_intel_crtc_mode_set since it has a large impact on the DPLL + * settings. + */ + + /* + * Enable automatic panel scaling so that non-native modes fill the + * screen. Should be enabled before the pipe is enabled, according to + * register description and PRM. + */ + if (mode->hdisplay != adjusted_mode->hdisplay || + mode->vdisplay != adjusted_mode->vdisplay) + pfit_control = (PFIT_ENABLE | VERT_AUTO_SCALE | + HORIZ_AUTO_SCALE | VERT_INTERP_BILINEAR | + HORIZ_INTERP_BILINEAR); + else + pfit_control = 0; + + if (dev_priv->lvds_dither) + pfit_control |= PANEL_8TO6_DITHER_ENABLE; + + REG_WRITE(PFIT_CONTROL, pfit_control); +} + +/** + * Detect the LVDS connection. + * + * This always returns CONNECTOR_STATUS_CONNECTED. + * This connector should only have + * been set up if the LVDS was actually connected anyway. + */ +static enum drm_connector_status cdv_intel_lvds_detect( + struct drm_connector *connector, bool force) +{ + return connector_status_connected; +} + +/** + * Return the list of DDC modes if available, or the BIOS fixed mode otherwise. + */ +static int cdv_intel_lvds_get_modes(struct drm_connector *connector) +{ + struct drm_device *dev = connector->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(connector); + struct psb_intel_mode_device *mode_dev = &dev_priv->mode_dev; + int ret; + + ret = psb_intel_ddc_get_modes(connector, &psb_intel_encoder->i2c_bus->adapter); + + if (ret) + return ret; + + /* Didn't get an EDID, so + * Set wide sync ranges so we get all modes + * handed to valid_mode for checking + */ + connector->display_info.min_vfreq = 0; + connector->display_info.max_vfreq = 200; + connector->display_info.min_hfreq = 0; + connector->display_info.max_hfreq = 200; + if (mode_dev->panel_fixed_mode != NULL) { + struct drm_display_mode *mode = + drm_mode_duplicate(dev, mode_dev->panel_fixed_mode); + drm_mode_probed_add(connector, mode); + return 1; + } + + return 0; +} + +/** + * cdv_intel_lvds_destroy - unregister and free LVDS structures + * @connector: connector to free + * + * Unregister the DDC bus for this connector then free the driver private + * structure. + */ +static void cdv_intel_lvds_destroy(struct drm_connector *connector) +{ + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(connector); + + if (psb_intel_encoder->i2c_bus) + psb_intel_i2c_destroy(psb_intel_encoder->i2c_bus); + drm_sysfs_connector_remove(connector); + drm_connector_cleanup(connector); + kfree(connector); +} + +static int cdv_intel_lvds_set_property(struct drm_connector *connector, + struct drm_property *property, + uint64_t value) +{ + struct drm_encoder *encoder = connector->encoder; + + if (!strcmp(property->name, "scaling mode") && encoder) { + struct psb_intel_crtc *crtc = + to_psb_intel_crtc(encoder->crtc); + uint64_t curValue; + + if (!crtc) + return -1; + + switch (value) { + case DRM_MODE_SCALE_FULLSCREEN: + break; + case DRM_MODE_SCALE_NO_SCALE: + break; + case DRM_MODE_SCALE_ASPECT: + break; + default: + return -1; + } + + if (drm_connector_property_get_value(connector, + property, + &curValue)) + return -1; + + if (curValue == value) + return 0; + + if (drm_connector_property_set_value(connector, + property, + value)) + return -1; + + if (crtc->saved_mode.hdisplay != 0 && + crtc->saved_mode.vdisplay != 0) { + if (!drm_crtc_helper_set_mode(encoder->crtc, + &crtc->saved_mode, + encoder->crtc->x, + encoder->crtc->y, + encoder->crtc->fb)) + return -1; + } + } else if (!strcmp(property->name, "backlight") && encoder) { + if (drm_connector_property_set_value(connector, + property, + value)) + return -1; + else { +#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE + struct drm_psb_private *dev_priv = + encoder->dev->dev_private; + struct backlight_device *bd = + dev_priv->backlight_device; + bd->props.brightness = value; + backlight_update_status(bd); +#endif + } + } else if (!strcmp(property->name, "DPMS") && encoder) { + struct drm_encoder_helper_funcs *helpers = + encoder->helper_private; + helpers->dpms(encoder, value); + } + return 0; +} + +static const struct drm_encoder_helper_funcs + cdv_intel_lvds_helper_funcs = { + .dpms = cdv_intel_lvds_encoder_dpms, + .mode_fixup = cdv_intel_lvds_mode_fixup, + .prepare = cdv_intel_lvds_prepare, + .mode_set = cdv_intel_lvds_mode_set, + .commit = cdv_intel_lvds_commit, +}; + +static const struct drm_connector_helper_funcs + cdv_intel_lvds_connector_helper_funcs = { + .get_modes = cdv_intel_lvds_get_modes, + .mode_valid = cdv_intel_lvds_mode_valid, + .best_encoder = psb_intel_best_encoder, +}; + +static const struct drm_connector_funcs cdv_intel_lvds_connector_funcs = { + .dpms = drm_helper_connector_dpms, + .save = cdv_intel_lvds_save, + .restore = cdv_intel_lvds_restore, + .detect = cdv_intel_lvds_detect, + .fill_modes = drm_helper_probe_single_connector_modes, + .set_property = cdv_intel_lvds_set_property, + .destroy = cdv_intel_lvds_destroy, +}; + + +static void cdv_intel_lvds_enc_destroy(struct drm_encoder *encoder) +{ + drm_encoder_cleanup(encoder); +} + +const struct drm_encoder_funcs cdv_intel_lvds_enc_funcs = { + .destroy = cdv_intel_lvds_enc_destroy, +}; + +/** + * cdv_intel_lvds_init - setup LVDS connectors on this device + * @dev: drm device + * + * Create the connector, register the LVDS DDC bus, and try to figure out what + * modes we can display on the LVDS panel (if present). + */ +void cdv_intel_lvds_init(struct drm_device *dev, + struct psb_intel_mode_device *mode_dev) +{ + struct psb_intel_encoder *psb_intel_encoder; + struct psb_intel_connector *psb_intel_connector; + struct cdv_intel_lvds_priv *lvds_priv; + struct drm_connector *connector; + struct drm_encoder *encoder; + struct drm_display_mode *scan; + struct drm_crtc *crtc; + struct drm_psb_private *dev_priv = dev->dev_private; + u32 lvds; + int pipe; + + psb_intel_encoder = kzalloc(sizeof(struct psb_intel_encoder), + GFP_KERNEL); + if (!psb_intel_encoder) + return; + + psb_intel_connector = kzalloc(sizeof(struct psb_intel_connector), + GFP_KERNEL); + if (!psb_intel_connector) + goto failed_connector; + + lvds_priv = kzalloc(sizeof(struct cdv_intel_lvds_priv), GFP_KERNEL); + if (!lvds_priv) + goto failed_lvds_priv; + + psb_intel_encoder->dev_priv = lvds_priv; + + connector = &psb_intel_connector->base; + encoder = &psb_intel_encoder->base; + + + drm_connector_init(dev, connector, + &cdv_intel_lvds_connector_funcs, + DRM_MODE_CONNECTOR_LVDS); + + drm_encoder_init(dev, encoder, + &cdv_intel_lvds_enc_funcs, + DRM_MODE_ENCODER_LVDS); + + + psb_intel_connector_attach_encoder(psb_intel_connector, + psb_intel_encoder); + psb_intel_encoder->type = INTEL_OUTPUT_LVDS; + + drm_encoder_helper_add(encoder, &cdv_intel_lvds_helper_funcs); + drm_connector_helper_add(connector, + &cdv_intel_lvds_connector_helper_funcs); + connector->display_info.subpixel_order = SubPixelHorizontalRGB; + connector->interlace_allowed = false; + connector->doublescan_allowed = false; + + /*Attach connector properties*/ + drm_connector_attach_property(connector, + dev->mode_config.scaling_mode_property, + DRM_MODE_SCALE_FULLSCREEN); + drm_connector_attach_property(connector, + dev_priv->backlight_property, + BRIGHTNESS_MAX_LEVEL); + + /** + * Set up I2C bus + * FIXME: distroy i2c_bus when exit + */ + psb_intel_encoder->i2c_bus = psb_intel_i2c_create(dev, + GPIOB, + "LVDSBLC_B"); + if (!psb_intel_encoder->i2c_bus) { + dev_printk(KERN_ERR, + &dev->pdev->dev, "I2C bus registration failed.\n"); + goto failed_blc_i2c; + } + psb_intel_encoder->i2c_bus->slave_addr = 0x2C; + dev_priv->lvds_i2c_bus = psb_intel_encoder->i2c_bus; + + /* + * LVDS discovery: + * 1) check for EDID on DDC + * 2) check for VBT data + * 3) check to see if LVDS is already on + * if none of the above, no panel + * 4) make sure lid is open + * if closed, act like it's not there for now + */ + + /* Set up the DDC bus. */ + psb_intel_encoder->ddc_bus = psb_intel_i2c_create(dev, + GPIOC, + "LVDSDDC_C"); + if (!psb_intel_encoder->ddc_bus) { + dev_printk(KERN_ERR, &dev->pdev->dev, + "DDC bus registration " "failed.\n"); + goto failed_ddc; + } + + /* + * Attempt to get the fixed panel mode from DDC. Assume that the + * preferred mode is the right one. + */ + psb_intel_ddc_get_modes(connector, + &psb_intel_encoder->ddc_bus->adapter); + list_for_each_entry(scan, &connector->probed_modes, head) { + if (scan->type & DRM_MODE_TYPE_PREFERRED) { + mode_dev->panel_fixed_mode = + drm_mode_duplicate(dev, scan); + goto out; /* FIXME: check for quirks */ + } + } + + /* Failed to get EDID, what about VBT? do we need this?*/ + if (dev_priv->lfp_lvds_vbt_mode) { + mode_dev->panel_fixed_mode = + drm_mode_duplicate(dev, dev_priv->lfp_lvds_vbt_mode); + if (mode_dev->panel_fixed_mode) { + mode_dev->panel_fixed_mode->type |= + DRM_MODE_TYPE_PREFERRED; + goto out; /* FIXME: check for quirks */ + } + } + /* + * If we didn't get EDID, try checking if the panel is already turned + * on. If so, assume that whatever is currently programmed is the + * correct mode. + */ + lvds = REG_READ(LVDS); + pipe = (lvds & LVDS_PIPEB_SELECT) ? 1 : 0; + crtc = psb_intel_get_crtc_from_pipe(dev, pipe); + + if (crtc && (lvds & LVDS_PORT_EN)) { + mode_dev->panel_fixed_mode = + cdv_intel_crtc_mode_get(dev, crtc); + if (mode_dev->panel_fixed_mode) { + mode_dev->panel_fixed_mode->type |= + DRM_MODE_TYPE_PREFERRED; + goto out; /* FIXME: check for quirks */ + } + } + + /* If we still don't have a mode after all that, give up. */ + if (!mode_dev->panel_fixed_mode) { + DRM_DEBUG + ("Found no modes on the lvds, ignoring the LVDS\n"); + goto failed_find; + } + +out: + drm_sysfs_connector_add(connector); + return; + +failed_find: + printk(KERN_ERR "Failed find\n"); + if (psb_intel_encoder->ddc_bus) + psb_intel_i2c_destroy(psb_intel_encoder->ddc_bus); +failed_ddc: + printk(KERN_ERR "Failed DDC\n"); + if (psb_intel_encoder->i2c_bus) + psb_intel_i2c_destroy(psb_intel_encoder->i2c_bus); +failed_blc_i2c: + printk(KERN_ERR "Failed BLC\n"); + drm_encoder_cleanup(encoder); + drm_connector_cleanup(connector); + kfree(lvds_priv); +failed_lvds_priv: + kfree(psb_intel_connector); +failed_connector: + kfree(psb_intel_encoder); +} diff --git a/drivers/gpu/drm/gma500/framebuffer.c b/drivers/gpu/drm/gma500/framebuffer.c new file mode 100644 index 0000000..8ea202f --- /dev/null +++ b/drivers/gpu/drm/gma500/framebuffer.c @@ -0,0 +1,800 @@ +/************************************************************************** + * Copyright (c) 2007-2011, Intel Corporation. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + **************************************************************************/ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include + +#include "psb_drv.h" +#include "psb_intel_reg.h" +#include "psb_intel_drv.h" +#include "framebuffer.h" +#include "gtt.h" + +static void psb_user_framebuffer_destroy(struct drm_framebuffer *fb); +static int psb_user_framebuffer_create_handle(struct drm_framebuffer *fb, + struct drm_file *file_priv, + unsigned int *handle); + +static const struct drm_framebuffer_funcs psb_fb_funcs = { + .destroy = psb_user_framebuffer_destroy, + .create_handle = psb_user_framebuffer_create_handle, +}; + +#define CMAP_TOHW(_val, _width) ((((_val) << (_width)) + 0x7FFF - (_val)) >> 16) + +static int psbfb_setcolreg(unsigned regno, unsigned red, unsigned green, + unsigned blue, unsigned transp, + struct fb_info *info) +{ + struct psb_fbdev *fbdev = info->par; + struct drm_framebuffer *fb = fbdev->psb_fb_helper.fb; + uint32_t v; + + if (!fb) + return -ENOMEM; + + if (regno > 255) + return 1; + + red = CMAP_TOHW(red, info->var.red.length); + blue = CMAP_TOHW(blue, info->var.blue.length); + green = CMAP_TOHW(green, info->var.green.length); + transp = CMAP_TOHW(transp, info->var.transp.length); + + v = (red << info->var.red.offset) | + (green << info->var.green.offset) | + (blue << info->var.blue.offset) | + (transp << info->var.transp.offset); + + if (regno < 16) { + switch (fb->bits_per_pixel) { + case 16: + ((uint32_t *) info->pseudo_palette)[regno] = v; + break; + case 24: + case 32: + ((uint32_t *) info->pseudo_palette)[regno] = v; + break; + } + } + + return 0; +} + +static int psbfb_pan(struct fb_var_screeninfo *var, struct fb_info *info) +{ + struct psb_fbdev *fbdev = info->par; + struct psb_framebuffer *psbfb = &fbdev->pfb; + struct drm_device *dev = psbfb->base.dev; + + /* + * We have to poke our nose in here. The core fb code assumes + * panning is part of the hardware that can be invoked before + * the actual fb is mapped. In our case that isn't quite true. + */ + if (psbfb->gtt->npage) { + /* GTT roll shifts in 4K pages, we need to shift the right + number of pages */ + int pages = info->fix.line_length >> 12; + psb_gtt_roll(dev, psbfb->gtt, var->yoffset * pages); + } + return 0; +} + +static int psbfb_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf) +{ + struct psb_framebuffer *psbfb = vma->vm_private_data; + struct drm_device *dev = psbfb->base.dev; + struct drm_psb_private *dev_priv = dev->dev_private; + int page_num; + int i; + unsigned long address; + int ret; + unsigned long pfn; + /* FIXME: assumes fb at stolen base which may not be true */ + unsigned long phys_addr = (unsigned long)dev_priv->stolen_base; + + page_num = (vma->vm_end - vma->vm_start) >> PAGE_SHIFT; + address = (unsigned long)vmf->virtual_address - (vmf->pgoff << PAGE_SHIFT); + + vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); + + for (i = 0; i < page_num; i++) { + pfn = (phys_addr >> PAGE_SHIFT); + + ret = vm_insert_mixed(vma, address, pfn); + if (unlikely((ret == -EBUSY) || (ret != 0 && i > 0))) + break; + else if (unlikely(ret != 0)) { + ret = (ret == -ENOMEM) ? VM_FAULT_OOM : VM_FAULT_SIGBUS; + return ret; + } + address += PAGE_SIZE; + phys_addr += PAGE_SIZE; + } + return VM_FAULT_NOPAGE; +} + +static void psbfb_vm_open(struct vm_area_struct *vma) +{ +} + +static void psbfb_vm_close(struct vm_area_struct *vma) +{ +} + +static struct vm_operations_struct psbfb_vm_ops = { + .fault = psbfb_vm_fault, + .open = psbfb_vm_open, + .close = psbfb_vm_close +}; + +static int psbfb_mmap(struct fb_info *info, struct vm_area_struct *vma) +{ + struct psb_fbdev *fbdev = info->par; + struct psb_framebuffer *psbfb = &fbdev->pfb; + + if (vma->vm_pgoff != 0) + return -EINVAL; + if (vma->vm_pgoff > (~0UL >> PAGE_SHIFT)) + return -EINVAL; + + if (!psbfb->addr_space) + psbfb->addr_space = vma->vm_file->f_mapping; + /* + * If this is a GEM object then info->screen_base is the virtual + * kernel remapping of the object. FIXME: Review if this is + * suitable for our mmap work + */ + vma->vm_ops = &psbfb_vm_ops; + vma->vm_private_data = (void *)psbfb; + vma->vm_flags |= VM_RESERVED | VM_IO | + VM_MIXEDMAP | VM_DONTEXPAND; + return 0; +} + +static int psbfb_ioctl(struct fb_info *info, unsigned int cmd, + unsigned long arg) +{ + return -ENOTTY; +} + +static struct fb_ops psbfb_ops = { + .owner = THIS_MODULE, + .fb_check_var = drm_fb_helper_check_var, + .fb_set_par = drm_fb_helper_set_par, + .fb_blank = drm_fb_helper_blank, + .fb_setcolreg = psbfb_setcolreg, + .fb_fillrect = cfb_fillrect, + .fb_copyarea = psbfb_copyarea, + .fb_imageblit = cfb_imageblit, + .fb_mmap = psbfb_mmap, + .fb_sync = psbfb_sync, + .fb_ioctl = psbfb_ioctl, +}; + +static struct fb_ops psbfb_roll_ops = { + .owner = THIS_MODULE, + .fb_check_var = drm_fb_helper_check_var, + .fb_set_par = drm_fb_helper_set_par, + .fb_blank = drm_fb_helper_blank, + .fb_setcolreg = psbfb_setcolreg, + .fb_fillrect = cfb_fillrect, + .fb_copyarea = cfb_copyarea, + .fb_imageblit = cfb_imageblit, + .fb_pan_display = psbfb_pan, + .fb_mmap = psbfb_mmap, + .fb_ioctl = psbfb_ioctl, +}; + +static struct fb_ops psbfb_unaccel_ops = { + .owner = THIS_MODULE, + .fb_check_var = drm_fb_helper_check_var, + .fb_set_par = drm_fb_helper_set_par, + .fb_blank = drm_fb_helper_blank, + .fb_setcolreg = psbfb_setcolreg, + .fb_fillrect = cfb_fillrect, + .fb_copyarea = cfb_copyarea, + .fb_imageblit = cfb_imageblit, + .fb_mmap = psbfb_mmap, + .fb_ioctl = psbfb_ioctl, +}; + +/** + * psb_framebuffer_init - initialize a framebuffer + * @dev: our DRM device + * @fb: framebuffer to set up + * @mode_cmd: mode description + * @gt: backing object + * + * Configure and fill in the boilerplate for our frame buffer. Return + * 0 on success or an error code if we fail. + */ +static int psb_framebuffer_init(struct drm_device *dev, + struct psb_framebuffer *fb, + struct drm_mode_fb_cmd2 *mode_cmd, + struct gtt_range *gt) +{ + u32 bpp, depth; + int ret; + + drm_fb_get_bpp_depth(mode_cmd->pixel_format, &depth, &bpp); + + if (mode_cmd->pitches[0] & 63) + return -EINVAL; + switch (bpp) { + case 8: + case 16: + case 24: + case 32: + break; + default: + return -EINVAL; + } + ret = drm_framebuffer_init(dev, &fb->base, &psb_fb_funcs); + if (ret) { + dev_err(dev->dev, "framebuffer init failed: %d\n", ret); + return ret; + } + drm_helper_mode_fill_fb_struct(&fb->base, mode_cmd); + fb->gtt = gt; + return 0; +} + +/** + * psb_framebuffer_create - create a framebuffer backed by gt + * @dev: our DRM device + * @mode_cmd: the description of the requested mode + * @gt: the backing object + * + * Create a framebuffer object backed by the gt, and fill in the + * boilerplate required + * + * TODO: review object references + */ + +static struct drm_framebuffer *psb_framebuffer_create + (struct drm_device *dev, + struct drm_mode_fb_cmd2 *mode_cmd, + struct gtt_range *gt) +{ + struct psb_framebuffer *fb; + int ret; + + fb = kzalloc(sizeof(*fb), GFP_KERNEL); + if (!fb) + return ERR_PTR(-ENOMEM); + + ret = psb_framebuffer_init(dev, fb, mode_cmd, gt); + if (ret) { + kfree(fb); + return ERR_PTR(ret); + } + return &fb->base; +} + +/** + * psbfb_alloc - allocate frame buffer memory + * @dev: the DRM device + * @aligned_size: space needed + * @force: fall back to GEM buffers if need be + * + * Allocate the frame buffer. In the usual case we get a GTT range that + * is stolen memory backed and life is simple. If there isn't sufficient + * we fail as we don't have the virtual mapping space to really vmap it + * and the kernel console code can't handle non linear framebuffers. + * + * Re-address this as and if the framebuffer layer grows this ability. + */ +static struct gtt_range *psbfb_alloc(struct drm_device *dev, int aligned_size) +{ + struct gtt_range *backing; + /* Begin by trying to use stolen memory backing */ + backing = psb_gtt_alloc_range(dev, aligned_size, "fb", 1); + if (backing) { + if (drm_gem_private_object_init(dev, + &backing->gem, aligned_size) == 0) + return backing; + psb_gtt_free_range(dev, backing); + } + return NULL; +} + +/** + * psbfb_create - create a framebuffer + * @fbdev: the framebuffer device + * @sizes: specification of the layout + * + * Create a framebuffer to the specifications provided + */ +static int psbfb_create(struct psb_fbdev *fbdev, + struct drm_fb_helper_surface_size *sizes) +{ + struct drm_device *dev = fbdev->psb_fb_helper.dev; + struct drm_psb_private *dev_priv = dev->dev_private; + struct fb_info *info; + struct drm_framebuffer *fb; + struct psb_framebuffer *psbfb = &fbdev->pfb; + struct drm_mode_fb_cmd2 mode_cmd; + struct device *device = &dev->pdev->dev; + int size; + int ret; + struct gtt_range *backing; + u32 bpp, depth; + int gtt_roll = 0; + int pitch_lines = 0; + + mode_cmd.width = sizes->surface_width; + mode_cmd.height = sizes->surface_height; + bpp = sizes->surface_bpp; + depth = sizes->surface_depth; + + /* No 24bit packed */ + if (bpp == 24) + bpp = 32; + + do { + /* + * Acceleration via the GTT requires pitch to be + * power of two aligned. Preferably page but less + * is ok with some fonts + */ + mode_cmd.pitches[0] = ALIGN(mode_cmd.width * ((bpp + 7) / 8), 4096 >> pitch_lines); + + size = mode_cmd.pitches[0] * mode_cmd.height; + size = ALIGN(size, PAGE_SIZE); + + /* Allocate the fb in the GTT with stolen page backing */ + backing = psbfb_alloc(dev, size); + + if (pitch_lines) + pitch_lines *= 2; + else + pitch_lines = 1; + gtt_roll++; + } while (backing == NULL && pitch_lines <= 16); + + /* The final pitch we accepted if we succeeded */ + pitch_lines /= 2; + + if (backing == NULL) { + /* + * We couldn't get the space we wanted, fall back to the + * display engine requirement instead. The HW requires + * the pitch to be 64 byte aligned + */ + + gtt_roll = 0; /* Don't use GTT accelerated scrolling */ + pitch_lines = 64; + + mode_cmd.pitches[0] = ALIGN(mode_cmd.width * ((bpp + 7) / 8), 64); + + size = mode_cmd.pitches[0] * mode_cmd.height; + size = ALIGN(size, PAGE_SIZE); + + /* Allocate the framebuffer in the GTT with stolen page backing */ + backing = psbfb_alloc(dev, size); + if (backing == NULL) + return -ENOMEM; + } + + mutex_lock(&dev->struct_mutex); + + info = framebuffer_alloc(0, device); + if (!info) { + ret = -ENOMEM; + goto out_err1; + } + info->par = fbdev; + + mode_cmd.pixel_format = drm_mode_legacy_fb_format(bpp, depth); + + ret = psb_framebuffer_init(dev, psbfb, &mode_cmd, backing); + if (ret) + goto out_unref; + + fb = &psbfb->base; + psbfb->fbdev = info; + + fbdev->psb_fb_helper.fb = fb; + fbdev->psb_fb_helper.fbdev = info; + + drm_fb_helper_fill_fix(info, fb->pitches[0], fb->depth); + strcpy(info->fix.id, "psbfb"); + + info->flags = FBINFO_DEFAULT; + if (dev_priv->ops->accel_2d && pitch_lines > 8) /* 2D engine */ + info->fbops = &psbfb_ops; + else if (gtt_roll) { /* GTT rolling seems best */ + info->fbops = &psbfb_roll_ops; + info->flags |= FBINFO_HWACCEL_YPAN; + } else /* Software */ + info->fbops = &psbfb_unaccel_ops; + + ret = fb_alloc_cmap(&info->cmap, 256, 0); + if (ret) { + ret = -ENOMEM; + goto out_unref; + } + + info->fix.smem_start = dev->mode_config.fb_base; + info->fix.smem_len = size; + info->fix.ywrapstep = gtt_roll; + info->fix.ypanstep = 0; + + /* Accessed stolen memory directly */ + info->screen_base = (char *)dev_priv->vram_addr + + backing->offset; + info->screen_size = size; + + if (dev_priv->gtt.stolen_size) { + info->apertures = alloc_apertures(1); + if (!info->apertures) { + ret = -ENOMEM; + goto out_unref; + } + info->apertures->ranges[0].base = dev->mode_config.fb_base; + info->apertures->ranges[0].size = dev_priv->gtt.stolen_size; + } + + drm_fb_helper_fill_var(info, &fbdev->psb_fb_helper, + sizes->fb_width, sizes->fb_height); + + info->fix.mmio_start = pci_resource_start(dev->pdev, 0); + info->fix.mmio_len = pci_resource_len(dev->pdev, 0); + + /* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */ + + dev_info(dev->dev, "allocated %dx%d fb\n", + psbfb->base.width, psbfb->base.height); + + mutex_unlock(&dev->struct_mutex); + return 0; +out_unref: + if (backing->stolen) + psb_gtt_free_range(dev, backing); + else + drm_gem_object_unreference(&backing->gem); +out_err1: + mutex_unlock(&dev->struct_mutex); + psb_gtt_free_range(dev, backing); + return ret; +} + +/** + * psb_user_framebuffer_create - create framebuffer + * @dev: our DRM device + * @filp: client file + * @cmd: mode request + * + * Create a new framebuffer backed by a userspace GEM object + */ +static struct drm_framebuffer *psb_user_framebuffer_create + (struct drm_device *dev, struct drm_file *filp, + struct drm_mode_fb_cmd2 *cmd) +{ + struct gtt_range *r; + struct drm_gem_object *obj; + + /* + * Find the GEM object and thus the gtt range object that is + * to back this space + */ + obj = drm_gem_object_lookup(dev, filp, cmd->handles[0]); + if (obj == NULL) + return ERR_PTR(-ENOENT); + + /* Let the core code do all the work */ + r = container_of(obj, struct gtt_range, gem); + return psb_framebuffer_create(dev, cmd, r); +} + +static void psbfb_gamma_set(struct drm_crtc *crtc, u16 red, u16 green, + u16 blue, int regno) +{ + struct psb_intel_crtc *intel_crtc = to_psb_intel_crtc(crtc); + + intel_crtc->lut_r[regno] = red >> 8; + intel_crtc->lut_g[regno] = green >> 8; + intel_crtc->lut_b[regno] = blue >> 8; +} + +static void psbfb_gamma_get(struct drm_crtc *crtc, u16 *red, + u16 *green, u16 *blue, int regno) +{ + struct psb_intel_crtc *intel_crtc = to_psb_intel_crtc(crtc); + + *red = intel_crtc->lut_r[regno] << 8; + *green = intel_crtc->lut_g[regno] << 8; + *blue = intel_crtc->lut_b[regno] << 8; +} + +static int psbfb_probe(struct drm_fb_helper *helper, + struct drm_fb_helper_surface_size *sizes) +{ + struct psb_fbdev *psb_fbdev = (struct psb_fbdev *)helper; + int new_fb = 0; + int ret; + + if (!helper->fb) { + ret = psbfb_create(psb_fbdev, sizes); + if (ret) + return ret; + new_fb = 1; + } + return new_fb; +} + +struct drm_fb_helper_funcs psb_fb_helper_funcs = { + .gamma_set = psbfb_gamma_set, + .gamma_get = psbfb_gamma_get, + .fb_probe = psbfb_probe, +}; + +static int psb_fbdev_destroy(struct drm_device *dev, struct psb_fbdev *fbdev) +{ + struct fb_info *info; + struct psb_framebuffer *psbfb = &fbdev->pfb; + + if (fbdev->psb_fb_helper.fbdev) { + info = fbdev->psb_fb_helper.fbdev; + unregister_framebuffer(info); + if (info->cmap.len) + fb_dealloc_cmap(&info->cmap); + framebuffer_release(info); + } + drm_fb_helper_fini(&fbdev->psb_fb_helper); + drm_framebuffer_cleanup(&psbfb->base); + + if (psbfb->gtt) + drm_gem_object_unreference(&psbfb->gtt->gem); + return 0; +} + +int psb_fbdev_init(struct drm_device *dev) +{ + struct psb_fbdev *fbdev; + struct drm_psb_private *dev_priv = dev->dev_private; + + fbdev = kzalloc(sizeof(struct psb_fbdev), GFP_KERNEL); + if (!fbdev) { + dev_err(dev->dev, "no memory\n"); + return -ENOMEM; + } + + dev_priv->fbdev = fbdev; + fbdev->psb_fb_helper.funcs = &psb_fb_helper_funcs; + + drm_fb_helper_init(dev, &fbdev->psb_fb_helper, dev_priv->ops->crtcs, + INTELFB_CONN_LIMIT); + + drm_fb_helper_single_add_all_connectors(&fbdev->psb_fb_helper); + drm_fb_helper_initial_config(&fbdev->psb_fb_helper, 32); + return 0; +} + +static void psb_fbdev_fini(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + + if (!dev_priv->fbdev) + return; + + psb_fbdev_destroy(dev, dev_priv->fbdev); + kfree(dev_priv->fbdev); + dev_priv->fbdev = NULL; +} + +static void psbfb_output_poll_changed(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_fbdev *fbdev = (struct psb_fbdev *)dev_priv->fbdev; + drm_fb_helper_hotplug_event(&fbdev->psb_fb_helper); +} + +/** + * psb_user_framebuffer_create_handle - add hamdle to a framebuffer + * @fb: framebuffer + * @file_priv: our DRM file + * @handle: returned handle + * + * Our framebuffer object is a GTT range which also contains a GEM + * object. We need to turn it into a handle for userspace. GEM will do + * the work for us + */ +static int psb_user_framebuffer_create_handle(struct drm_framebuffer *fb, + struct drm_file *file_priv, + unsigned int *handle) +{ + struct psb_framebuffer *psbfb = to_psb_fb(fb); + struct gtt_range *r = psbfb->gtt; + return drm_gem_handle_create(file_priv, &r->gem, handle); +} + +/** + * psb_user_framebuffer_destroy - destruct user created fb + * @fb: framebuffer + * + * User framebuffers are backed by GEM objects so all we have to do is + * clean up a bit and drop the reference, GEM will handle the fallout + */ +static void psb_user_framebuffer_destroy(struct drm_framebuffer *fb) +{ + struct psb_framebuffer *psbfb = to_psb_fb(fb); + struct gtt_range *r = psbfb->gtt; + struct drm_device *dev = fb->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_fbdev *fbdev = dev_priv->fbdev; + struct drm_crtc *crtc; + int reset = 0; + + /* Should never get stolen memory for a user fb */ + WARN_ON(r->stolen); + + /* Check if we are erroneously live */ + list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) + if (crtc->fb == fb) + reset = 1; + + if (reset) + /* + * Now force a sane response before we permit the DRM CRTC + * layer to do stupid things like blank the display. Instead + * we reset this framebuffer as if the user had forced a reset. + * We must do this before the cleanup so that the DRM layer + * doesn't get a chance to stick its oar in where it isn't + * wanted. + */ + drm_fb_helper_restore_fbdev_mode(&fbdev->psb_fb_helper); + + /* Let DRM do its clean up */ + drm_framebuffer_cleanup(fb); + /* We are no longer using the resource in GEM */ + drm_gem_object_unreference_unlocked(&r->gem); + kfree(fb); +} + +static const struct drm_mode_config_funcs psb_mode_funcs = { + .fb_create = psb_user_framebuffer_create, + .output_poll_changed = psbfb_output_poll_changed, +}; + +static int psb_create_backlight_property(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct drm_property *backlight; + + if (dev_priv->backlight_property) + return 0; + + backlight = drm_property_create_range(dev, 0, "backlight", 0, 100); + + dev_priv->backlight_property = backlight; + + return 0; +} + +static void psb_setup_outputs(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct drm_connector *connector; + + drm_mode_create_scaling_mode_property(dev); + psb_create_backlight_property(dev); + + dev_priv->ops->output_init(dev); + + list_for_each_entry(connector, &dev->mode_config.connector_list, + head) { + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(connector); + struct drm_encoder *encoder = &psb_intel_encoder->base; + int crtc_mask = 0, clone_mask = 0; + + /* valid crtcs */ + switch (psb_intel_encoder->type) { + case INTEL_OUTPUT_ANALOG: + crtc_mask = (1 << 0); + clone_mask = (1 << INTEL_OUTPUT_ANALOG); + break; + case INTEL_OUTPUT_SDVO: + crtc_mask = ((1 << 0) | (1 << 1)); + clone_mask = (1 << INTEL_OUTPUT_SDVO); + break; + case INTEL_OUTPUT_LVDS: + if (IS_MRST(dev)) + crtc_mask = (1 << 0); + else + crtc_mask = (1 << 1); + clone_mask = (1 << INTEL_OUTPUT_LVDS); + break; + case INTEL_OUTPUT_MIPI: + crtc_mask = (1 << 0); + clone_mask = (1 << INTEL_OUTPUT_MIPI); + break; + case INTEL_OUTPUT_MIPI2: + crtc_mask = (1 << 2); + clone_mask = (1 << INTEL_OUTPUT_MIPI2); + break; + case INTEL_OUTPUT_HDMI: + if (IS_MFLD(dev)) + crtc_mask = (1 << 1); + else + crtc_mask = (1 << 0); + clone_mask = (1 << INTEL_OUTPUT_HDMI); + break; + } + encoder->possible_crtcs = crtc_mask; + encoder->possible_clones = + psb_intel_connector_clones(dev, clone_mask); + } +} + +void psb_modeset_init(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_intel_mode_device *mode_dev = &dev_priv->mode_dev; + int i; + + drm_mode_config_init(dev); + + dev->mode_config.min_width = 0; + dev->mode_config.min_height = 0; + + dev->mode_config.funcs = (void *) &psb_mode_funcs; + + /* set memory base */ + /* Oaktrail and Poulsbo should use BAR 2*/ + pci_read_config_dword(dev->pdev, PSB_BSM, (u32 *) + &(dev->mode_config.fb_base)); + + /* num pipes is 2 for PSB but 1 for Mrst */ + for (i = 0; i < dev_priv->num_pipe; i++) + psb_intel_crtc_init(dev, i, mode_dev); + + dev->mode_config.max_width = 2048; + dev->mode_config.max_height = 2048; + + psb_setup_outputs(dev); +} + +void psb_modeset_cleanup(struct drm_device *dev) +{ + mutex_lock(&dev->struct_mutex); + + drm_kms_helper_poll_fini(dev); + psb_fbdev_fini(dev); + drm_mode_config_cleanup(dev); + + mutex_unlock(&dev->struct_mutex); +} diff --git a/drivers/gpu/drm/gma500/framebuffer.h b/drivers/gpu/drm/gma500/framebuffer.h new file mode 100644 index 0000000..989558a --- /dev/null +++ b/drivers/gpu/drm/gma500/framebuffer.h @@ -0,0 +1,47 @@ +/* + * Copyright (c) 2008-2011, Intel Corporation + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + * Authors: + * Eric Anholt + * + */ + +#ifndef _FRAMEBUFFER_H_ +#define _FRAMEBUFFER_H_ + +#include +#include + +#include "psb_drv.h" + +struct psb_framebuffer { + struct drm_framebuffer base; + struct address_space *addr_space; + struct fb_info *fbdev; + struct gtt_range *gtt; +}; + +struct psb_fbdev { + struct drm_fb_helper psb_fb_helper; + struct psb_framebuffer pfb; +}; + +#define to_psb_fb(x) container_of(x, struct psb_framebuffer, base) + +extern int psb_intel_connector_clones(struct drm_device *dev, int type_mask); + +#endif + diff --git a/drivers/gpu/drm/gma500/gem.c b/drivers/gpu/drm/gma500/gem.c new file mode 100644 index 0000000..9fbb868 --- /dev/null +++ b/drivers/gpu/drm/gma500/gem.c @@ -0,0 +1,292 @@ +/* + * psb GEM interface + * + * Copyright (c) 2011, Intel Corporation. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + * Authors: Alan Cox + * + * TODO: + * - we need to work out if the MMU is relevant (eg for + * accelerated operations on a GEM object) + */ + +#include +#include +#include "gma_drm.h" +#include "psb_drv.h" + +int psb_gem_init_object(struct drm_gem_object *obj) +{ + return -EINVAL; +} + +void psb_gem_free_object(struct drm_gem_object *obj) +{ + struct gtt_range *gtt = container_of(obj, struct gtt_range, gem); + drm_gem_object_release_wrap(obj); + /* This must occur last as it frees up the memory of the GEM object */ + psb_gtt_free_range(obj->dev, gtt); +} + +int psb_gem_get_aperture(struct drm_device *dev, void *data, + struct drm_file *file) +{ + return -EINVAL; +} + +/** + * psb_gem_dumb_map_gtt - buffer mapping for dumb interface + * @file: our drm client file + * @dev: drm device + * @handle: GEM handle to the object (from dumb_create) + * + * Do the necessary setup to allow the mapping of the frame buffer + * into user memory. We don't have to do much here at the moment. + */ +int psb_gem_dumb_map_gtt(struct drm_file *file, struct drm_device *dev, + uint32_t handle, uint64_t *offset) +{ + int ret = 0; + struct drm_gem_object *obj; + + if (!(dev->driver->driver_features & DRIVER_GEM)) + return -ENODEV; + + mutex_lock(&dev->struct_mutex); + + /* GEM does all our handle to object mapping */ + obj = drm_gem_object_lookup(dev, file, handle); + if (obj == NULL) { + ret = -ENOENT; + goto unlock; + } + /* What validation is needed here ? */ + + /* Make it mmapable */ + if (!obj->map_list.map) { + ret = gem_create_mmap_offset(obj); + if (ret) + goto out; + } + /* GEM should really work out the hash offsets for us */ + *offset = (u64)obj->map_list.hash.key << PAGE_SHIFT; +out: + drm_gem_object_unreference(obj); +unlock: + mutex_unlock(&dev->struct_mutex); + return ret; +} + +/** + * psb_gem_create - create a mappable object + * @file: the DRM file of the client + * @dev: our device + * @size: the size requested + * @handlep: returned handle (opaque number) + * + * Create a GEM object, fill in the boilerplate and attach a handle to + * it so that userspace can speak about it. This does the core work + * for the various methods that do/will create GEM objects for things + */ +static int psb_gem_create(struct drm_file *file, + struct drm_device *dev, uint64_t size, uint32_t *handlep) +{ + struct gtt_range *r; + int ret; + u32 handle; + + size = roundup(size, PAGE_SIZE); + + /* Allocate our object - for now a direct gtt range which is not + stolen memory backed */ + r = psb_gtt_alloc_range(dev, size, "gem", 0); + if (r == NULL) { + dev_err(dev->dev, "no memory for %lld byte GEM object\n", size); + return -ENOSPC; + } + /* Initialize the extra goodies GEM needs to do all the hard work */ + if (drm_gem_object_init(dev, &r->gem, size) != 0) { + psb_gtt_free_range(dev, r); + /* GEM doesn't give an error code so use -ENOMEM */ + dev_err(dev->dev, "GEM init failed for %lld\n", size); + return -ENOMEM; + } + /* Give the object a handle so we can carry it more easily */ + ret = drm_gem_handle_create(file, &r->gem, &handle); + if (ret) { + dev_err(dev->dev, "GEM handle failed for %p, %lld\n", + &r->gem, size); + drm_gem_object_release(&r->gem); + psb_gtt_free_range(dev, r); + return ret; + } + /* We have the initial and handle reference but need only one now */ + drm_gem_object_unreference(&r->gem); + *handlep = handle; + return 0; +} + +/** + * psb_gem_dumb_create - create a dumb buffer + * @drm_file: our client file + * @dev: our device + * @args: the requested arguments copied from userspace + * + * Allocate a buffer suitable for use for a frame buffer of the + * form described by user space. Give userspace a handle by which + * to reference it. + */ +int psb_gem_dumb_create(struct drm_file *file, struct drm_device *dev, + struct drm_mode_create_dumb *args) +{ + args->pitch = ALIGN(args->width * ((args->bpp + 7) / 8), 64); + args->size = args->pitch * args->height; + return psb_gem_create(file, dev, args->size, &args->handle); +} + +/** + * psb_gem_dumb_destroy - destroy a dumb buffer + * @file: client file + * @dev: our DRM device + * @handle: the object handle + * + * Destroy a handle that was created via psb_gem_dumb_create, at least + * we hope it was created that way. i915 seems to assume the caller + * does the checking but that might be worth review ! FIXME + */ +int psb_gem_dumb_destroy(struct drm_file *file, struct drm_device *dev, + uint32_t handle) +{ + /* No special work needed, drop the reference and see what falls out */ + return drm_gem_handle_delete(file, handle); +} + +/** + * psb_gem_fault - pagefault handler for GEM objects + * @vma: the VMA of the GEM object + * @vmf: fault detail + * + * Invoked when a fault occurs on an mmap of a GEM managed area. GEM + * does most of the work for us including the actual map/unmap calls + * but we need to do the actual page work. + * + * This code eventually needs to handle faulting objects in and out + * of the GTT and repacking it when we run out of space. We can put + * that off for now and for our simple uses + * + * The VMA was set up by GEM. In doing so it also ensured that the + * vma->vm_private_data points to the GEM object that is backing this + * mapping. + */ +int psb_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf) +{ + struct drm_gem_object *obj; + struct gtt_range *r; + int ret; + unsigned long pfn; + pgoff_t page_offset; + struct drm_device *dev; + struct drm_psb_private *dev_priv; + + obj = vma->vm_private_data; /* GEM object */ + dev = obj->dev; + dev_priv = dev->dev_private; + + r = container_of(obj, struct gtt_range, gem); /* Get the gtt range */ + + /* Make sure we don't parallel update on a fault, nor move or remove + something from beneath our feet */ + mutex_lock(&dev->struct_mutex); + + /* For now the mmap pins the object and it stays pinned. As things + stand that will do us no harm */ + if (r->mmapping == 0) { + ret = psb_gtt_pin(r); + if (ret < 0) { + dev_err(dev->dev, "gma500: pin failed: %d\n", ret); + goto fail; + } + r->mmapping = 1; + } + + /* Page relative to the VMA start - we must calculate this ourselves + because vmf->pgoff is the fake GEM offset */ + page_offset = ((unsigned long) vmf->virtual_address - vma->vm_start) + >> PAGE_SHIFT; + + /* CPU view of the page, don't go via the GART for CPU writes */ + if (r->stolen) + pfn = (dev_priv->stolen_base + r->offset) >> PAGE_SHIFT; + else + pfn = page_to_pfn(r->pages[page_offset]); + ret = vm_insert_pfn(vma, (unsigned long)vmf->virtual_address, pfn); + +fail: + mutex_unlock(&dev->struct_mutex); + switch (ret) { + case 0: + case -ERESTARTSYS: + case -EINTR: + return VM_FAULT_NOPAGE; + case -ENOMEM: + return VM_FAULT_OOM; + default: + return VM_FAULT_SIGBUS; + } +} + +static int psb_gem_create_stolen(struct drm_file *file, struct drm_device *dev, + int size, u32 *handle) +{ + struct gtt_range *gtt = psb_gtt_alloc_range(dev, size, "gem", 1); + if (gtt == NULL) + return -ENOMEM; + if (drm_gem_private_object_init(dev, >t->gem, size) != 0) + goto free_gtt; + if (drm_gem_handle_create(file, >t->gem, handle) == 0) + return 0; +free_gtt: + psb_gtt_free_range(dev, gtt); + return -ENOMEM; +} + +/* + * GEM interfaces for our specific client + */ +int psb_gem_create_ioctl(struct drm_device *dev, void *data, + struct drm_file *file) +{ + struct drm_psb_gem_create *args = data; + int ret; + if (args->flags & GMA_GEM_CREATE_STOLEN) { + ret = psb_gem_create_stolen(file, dev, args->size, + &args->handle); + if (ret == 0) + return 0; + /* Fall throguh */ + args->flags &= ~GMA_GEM_CREATE_STOLEN; + } + return psb_gem_create(file, dev, args->size, &args->handle); +} + +int psb_gem_mmap_ioctl(struct drm_device *dev, void *data, + struct drm_file *file) +{ + struct drm_psb_gem_mmap *args = data; + return dev->driver->dumb_map_offset(file, dev, + args->handle, &args->offset); +} + diff --git a/drivers/gpu/drm/gma500/gem_glue.c b/drivers/gpu/drm/gma500/gem_glue.c new file mode 100644 index 0000000..3c17634 --- /dev/null +++ b/drivers/gpu/drm/gma500/gem_glue.c @@ -0,0 +1,90 @@ +/************************************************************************** + * Copyright (c) 2011, Intel Corporation. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + **************************************************************************/ + +#include +#include +#include "gem_glue.h" + +void drm_gem_object_release_wrap(struct drm_gem_object *obj) +{ + /* Remove the list map if one is present */ + if (obj->map_list.map) { + struct drm_gem_mm *mm = obj->dev->mm_private; + struct drm_map_list *list = &obj->map_list; + drm_ht_remove_item(&mm->offset_hash, &list->hash); + drm_mm_put_block(list->file_offset_node); + kfree(list->map); + list->map = NULL; + } + drm_gem_object_release(obj); +} + +/** + * gem_create_mmap_offset - invent an mmap offset + * @obj: our object + * + * Standard implementation of offset generation for mmap as is + * duplicated in several drivers. This belongs in GEM. + */ +int gem_create_mmap_offset(struct drm_gem_object *obj) +{ + struct drm_device *dev = obj->dev; + struct drm_gem_mm *mm = dev->mm_private; + struct drm_map_list *list; + struct drm_local_map *map; + int ret; + + list = &obj->map_list; + list->map = kzalloc(sizeof(struct drm_map_list), GFP_KERNEL); + if (list->map == NULL) + return -ENOMEM; + map = list->map; + map->type = _DRM_GEM; + map->size = obj->size; + map->handle = obj; + + list->file_offset_node = drm_mm_search_free(&mm->offset_manager, + obj->size / PAGE_SIZE, 0, 0); + if (!list->file_offset_node) { + dev_err(dev->dev, "failed to allocate offset for bo %d\n", + obj->name); + ret = -ENOSPC; + goto free_it; + } + list->file_offset_node = drm_mm_get_block(list->file_offset_node, + obj->size / PAGE_SIZE, 0); + if (!list->file_offset_node) { + ret = -ENOMEM; + goto free_it; + } + list->hash.key = list->file_offset_node->start; + ret = drm_ht_insert_item(&mm->offset_hash, &list->hash); + if (ret) { + dev_err(dev->dev, "failed to add to map hash\n"); + goto free_mm; + } + return 0; + +free_mm: + drm_mm_put_block(list->file_offset_node); +free_it: + kfree(list->map); + list->map = NULL; + return ret; +} diff --git a/drivers/gpu/drm/gma500/gem_glue.h b/drivers/gpu/drm/gma500/gem_glue.h new file mode 100644 index 0000000..ce5ce30 --- /dev/null +++ b/drivers/gpu/drm/gma500/gem_glue.h @@ -0,0 +1,2 @@ +extern void drm_gem_object_release_wrap(struct drm_gem_object *obj); +extern int gem_create_mmap_offset(struct drm_gem_object *obj); diff --git a/drivers/gpu/drm/gma500/gtt.c b/drivers/gpu/drm/gma500/gtt.c new file mode 100644 index 0000000..c6465b4 --- /dev/null +++ b/drivers/gpu/drm/gma500/gtt.c @@ -0,0 +1,551 @@ +/* + * Copyright (c) 2007, Intel Corporation. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + * Authors: Thomas Hellstrom + * Alan Cox + */ + +#include +#include +#include "psb_drv.h" + + +/* + * GTT resource allocator - manage page mappings in GTT space + */ + +/** + * psb_gtt_mask_pte - generate GTT pte entry + * @pfn: page number to encode + * @type: type of memory in the GTT + * + * Set the GTT entry for the appropriate memory type. + */ +static inline uint32_t psb_gtt_mask_pte(uint32_t pfn, int type) +{ + uint32_t mask = PSB_PTE_VALID; + + if (type & PSB_MMU_CACHED_MEMORY) + mask |= PSB_PTE_CACHED; + if (type & PSB_MMU_RO_MEMORY) + mask |= PSB_PTE_RO; + if (type & PSB_MMU_WO_MEMORY) + mask |= PSB_PTE_WO; + + return (pfn << PAGE_SHIFT) | mask; +} + +/** + * psb_gtt_entry - find the GTT entries for a gtt_range + * @dev: our DRM device + * @r: our GTT range + * + * Given a gtt_range object return the GTT offset of the page table + * entries for this gtt_range + */ +static u32 *psb_gtt_entry(struct drm_device *dev, struct gtt_range *r) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + unsigned long offset; + + offset = r->resource.start - dev_priv->gtt_mem->start; + + return dev_priv->gtt_map + (offset >> PAGE_SHIFT); +} + +/** + * psb_gtt_insert - put an object into the GTT + * @dev: our DRM device + * @r: our GTT range + * + * Take our preallocated GTT range and insert the GEM object into + * the GTT. This is protected via the gtt mutex which the caller + * must hold. + */ +static int psb_gtt_insert(struct drm_device *dev, struct gtt_range *r) +{ + u32 *gtt_slot, pte; + struct page **pages; + int i; + + if (r->pages == NULL) { + WARN_ON(1); + return -EINVAL; + } + + WARN_ON(r->stolen); /* refcount these maybe ? */ + + gtt_slot = psb_gtt_entry(dev, r); + pages = r->pages; + + /* Make sure changes are visible to the GPU */ + set_pages_array_uc(pages, r->npage); + + /* Write our page entries into the GTT itself */ + for (i = r->roll; i < r->npage; i++) { + pte = psb_gtt_mask_pte(page_to_pfn(r->pages[i]), 0); + iowrite32(pte, gtt_slot++); + } + for (i = 0; i < r->roll; i++) { + pte = psb_gtt_mask_pte(page_to_pfn(r->pages[i]), 0); + iowrite32(pte, gtt_slot++); + } + /* Make sure all the entries are set before we return */ + ioread32(gtt_slot - 1); + + return 0; +} + +/** + * psb_gtt_remove - remove an object from the GTT + * @dev: our DRM device + * @r: our GTT range + * + * Remove a preallocated GTT range from the GTT. Overwrite all the + * page table entries with the dummy page. This is protected via the gtt + * mutex which the caller must hold. + */ +static void psb_gtt_remove(struct drm_device *dev, struct gtt_range *r) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + u32 *gtt_slot, pte; + int i; + + WARN_ON(r->stolen); + + gtt_slot = psb_gtt_entry(dev, r); + pte = psb_gtt_mask_pte(page_to_pfn(dev_priv->scratch_page), 0); + + for (i = 0; i < r->npage; i++) + iowrite32(pte, gtt_slot++); + ioread32(gtt_slot - 1); + set_pages_array_wb(r->pages, r->npage); +} + +/** + * psb_gtt_roll - set scrolling position + * @dev: our DRM device + * @r: the gtt mapping we are using + * @roll: roll offset + * + * Roll an existing pinned mapping by moving the pages through the GTT. + * This allows us to implement hardware scrolling on the consoles without + * a 2D engine + */ +void psb_gtt_roll(struct drm_device *dev, struct gtt_range *r, int roll) +{ + u32 *gtt_slot, pte; + int i; + + if (roll >= r->npage) { + WARN_ON(1); + return; + } + + r->roll = roll; + + /* Not currently in the GTT - no worry we will write the mapping at + the right position when it gets pinned */ + if (!r->stolen && !r->in_gart) + return; + + gtt_slot = psb_gtt_entry(dev, r); + + for (i = r->roll; i < r->npage; i++) { + pte = psb_gtt_mask_pte(page_to_pfn(r->pages[i]), 0); + iowrite32(pte, gtt_slot++); + } + for (i = 0; i < r->roll; i++) { + pte = psb_gtt_mask_pte(page_to_pfn(r->pages[i]), 0); + iowrite32(pte, gtt_slot++); + } + ioread32(gtt_slot - 1); +} + +/** + * psb_gtt_attach_pages - attach and pin GEM pages + * @gt: the gtt range + * + * Pin and build an in kernel list of the pages that back our GEM object. + * While we hold this the pages cannot be swapped out. This is protected + * via the gtt mutex which the caller must hold. + */ +static int psb_gtt_attach_pages(struct gtt_range *gt) +{ + struct inode *inode; + struct address_space *mapping; + int i; + struct page *p; + int pages = gt->gem.size / PAGE_SIZE; + + WARN_ON(gt->pages); + + /* This is the shared memory object that backs the GEM resource */ + inode = gt->gem.filp->f_path.dentry->d_inode; + mapping = inode->i_mapping; + + gt->pages = kmalloc(pages * sizeof(struct page *), GFP_KERNEL); + if (gt->pages == NULL) + return -ENOMEM; + gt->npage = pages; + + for (i = 0; i < pages; i++) { + p = shmem_read_mapping_page(mapping, i); + if (IS_ERR(p)) + goto err; + gt->pages[i] = p; + } + return 0; + +err: + while (i--) + page_cache_release(gt->pages[i]); + kfree(gt->pages); + gt->pages = NULL; + return PTR_ERR(p); +} + +/** + * psb_gtt_detach_pages - attach and pin GEM pages + * @gt: the gtt range + * + * Undo the effect of psb_gtt_attach_pages. At this point the pages + * must have been removed from the GTT as they could now be paged out + * and move bus address. This is protected via the gtt mutex which the + * caller must hold. + */ +static void psb_gtt_detach_pages(struct gtt_range *gt) +{ + int i; + for (i = 0; i < gt->npage; i++) { + /* FIXME: do we need to force dirty */ + set_page_dirty(gt->pages[i]); + page_cache_release(gt->pages[i]); + } + kfree(gt->pages); + gt->pages = NULL; +} + +/** + * psb_gtt_pin - pin pages into the GTT + * @gt: range to pin + * + * Pin a set of pages into the GTT. The pins are refcounted so that + * multiple pins need multiple unpins to undo. + * + * Non GEM backed objects treat this as a no-op as they are always GTT + * backed objects. + */ +int psb_gtt_pin(struct gtt_range *gt) +{ + int ret = 0; + struct drm_device *dev = gt->gem.dev; + struct drm_psb_private *dev_priv = dev->dev_private; + + mutex_lock(&dev_priv->gtt_mutex); + + if (gt->in_gart == 0 && gt->stolen == 0) { + ret = psb_gtt_attach_pages(gt); + if (ret < 0) + goto out; + ret = psb_gtt_insert(dev, gt); + if (ret < 0) { + psb_gtt_detach_pages(gt); + goto out; + } + } + gt->in_gart++; +out: + mutex_unlock(&dev_priv->gtt_mutex); + return ret; +} + +/** + * psb_gtt_unpin - Drop a GTT pin requirement + * @gt: range to pin + * + * Undoes the effect of psb_gtt_pin. On the last drop the GEM object + * will be removed from the GTT which will also drop the page references + * and allow the VM to clean up or page stuff. + * + * Non GEM backed objects treat this as a no-op as they are always GTT + * backed objects. + */ +void psb_gtt_unpin(struct gtt_range *gt) +{ + struct drm_device *dev = gt->gem.dev; + struct drm_psb_private *dev_priv = dev->dev_private; + + mutex_lock(&dev_priv->gtt_mutex); + + WARN_ON(!gt->in_gart); + + gt->in_gart--; + if (gt->in_gart == 0 && gt->stolen == 0) { + psb_gtt_remove(dev, gt); + psb_gtt_detach_pages(gt); + } + mutex_unlock(&dev_priv->gtt_mutex); +} + +/* + * GTT resource allocator - allocate and manage GTT address space + */ + +/** + * psb_gtt_alloc_range - allocate GTT address space + * @dev: Our DRM device + * @len: length (bytes) of address space required + * @name: resource name + * @backed: resource should be backed by stolen pages + * + * Ask the kernel core to find us a suitable range of addresses + * to use for a GTT mapping. + * + * Returns a gtt_range structure describing the object, or NULL on + * error. On successful return the resource is both allocated and marked + * as in use. + */ +struct gtt_range *psb_gtt_alloc_range(struct drm_device *dev, int len, + const char *name, int backed) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct gtt_range *gt; + struct resource *r = dev_priv->gtt_mem; + int ret; + unsigned long start, end; + + if (backed) { + /* The start of the GTT is the stolen pages */ + start = r->start; + end = r->start + dev_priv->gtt.stolen_size - 1; + } else { + /* The rest we will use for GEM backed objects */ + start = r->start + dev_priv->gtt.stolen_size; + end = r->end; + } + + gt = kzalloc(sizeof(struct gtt_range), GFP_KERNEL); + if (gt == NULL) + return NULL; + gt->resource.name = name; + gt->stolen = backed; + gt->in_gart = backed; + gt->roll = 0; + /* Ensure this is set for non GEM objects */ + gt->gem.dev = dev; + ret = allocate_resource(dev_priv->gtt_mem, >->resource, + len, start, end, PAGE_SIZE, NULL, NULL); + if (ret == 0) { + gt->offset = gt->resource.start - r->start; + return gt; + } + kfree(gt); + return NULL; +} + +/** + * psb_gtt_free_range - release GTT address space + * @dev: our DRM device + * @gt: a mapping created with psb_gtt_alloc_range + * + * Release a resource that was allocated with psb_gtt_alloc_range. If the + * object has been pinned by mmap users we clean this up here currently. + */ +void psb_gtt_free_range(struct drm_device *dev, struct gtt_range *gt) +{ + /* Undo the mmap pin if we are destroying the object */ + if (gt->mmapping) { + psb_gtt_unpin(gt); + gt->mmapping = 0; + } + WARN_ON(gt->in_gart && !gt->stolen); + release_resource(>->resource); + kfree(gt); +} + +static void psb_gtt_alloc(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + init_rwsem(&dev_priv->gtt.sem); +} + +void psb_gtt_takedown(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + + if (dev_priv->gtt_map) { + iounmap(dev_priv->gtt_map); + dev_priv->gtt_map = NULL; + } + if (dev_priv->gtt_initialized) { + pci_write_config_word(dev->pdev, PSB_GMCH_CTRL, + dev_priv->gmch_ctrl); + PSB_WVDC32(dev_priv->pge_ctl, PSB_PGETBL_CTL); + (void) PSB_RVDC32(PSB_PGETBL_CTL); + } + if (dev_priv->vram_addr) + iounmap(dev_priv->gtt_map); +} + +int psb_gtt_init(struct drm_device *dev, int resume) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + unsigned gtt_pages; + unsigned long stolen_size, vram_stolen_size; + unsigned i, num_pages; + unsigned pfn_base; + uint32_t vram_pages; + uint32_t dvmt_mode = 0; + struct psb_gtt *pg; + + int ret = 0; + uint32_t pte; + + mutex_init(&dev_priv->gtt_mutex); + + psb_gtt_alloc(dev); + pg = &dev_priv->gtt; + + /* Enable the GTT */ + pci_read_config_word(dev->pdev, PSB_GMCH_CTRL, &dev_priv->gmch_ctrl); + pci_write_config_word(dev->pdev, PSB_GMCH_CTRL, + dev_priv->gmch_ctrl | _PSB_GMCH_ENABLED); + + dev_priv->pge_ctl = PSB_RVDC32(PSB_PGETBL_CTL); + PSB_WVDC32(dev_priv->pge_ctl | _PSB_PGETBL_ENABLED, PSB_PGETBL_CTL); + (void) PSB_RVDC32(PSB_PGETBL_CTL); + + /* The root resource we allocate address space from */ + dev_priv->gtt_initialized = 1; + + pg->gtt_phys_start = dev_priv->pge_ctl & PAGE_MASK; + + /* + * The video mmu has a hw bug when accessing 0x0D0000000. + * Make gatt start at 0x0e000,0000. This doesn't actually + * matter for us but may do if the video acceleration ever + * gets opened up. + */ + pg->mmu_gatt_start = 0xE0000000; + + pg->gtt_start = pci_resource_start(dev->pdev, PSB_GTT_RESOURCE); + gtt_pages = pci_resource_len(dev->pdev, PSB_GTT_RESOURCE) + >> PAGE_SHIFT; + /* CDV doesn't report this. In which case the system has 64 gtt pages */ + if (pg->gtt_start == 0 || gtt_pages == 0) { + dev_dbg(dev->dev, "GTT PCI BAR not initialized.\n"); + gtt_pages = 64; + pg->gtt_start = dev_priv->pge_ctl; + } + + pg->gatt_start = pci_resource_start(dev->pdev, PSB_GATT_RESOURCE); + pg->gatt_pages = pci_resource_len(dev->pdev, PSB_GATT_RESOURCE) + >> PAGE_SHIFT; + dev_priv->gtt_mem = &dev->pdev->resource[PSB_GATT_RESOURCE]; + + if (pg->gatt_pages == 0 || pg->gatt_start == 0) { + static struct resource fudge; /* Preferably peppermint */ + /* This can occur on CDV systems. Fudge it in this case. + We really don't care what imaginary space is being allocated + at this point */ + dev_dbg(dev->dev, "GATT PCI BAR not initialized.\n"); + pg->gatt_start = 0x40000000; + pg->gatt_pages = (128 * 1024 * 1024) >> PAGE_SHIFT; + /* This is a little confusing but in fact the GTT is providing + a view from the GPU into memory and not vice versa. As such + this is really allocating space that is not the same as the + CPU address space on CDV */ + fudge.start = 0x40000000; + fudge.end = 0x40000000 + 128 * 1024 * 1024 - 1; + fudge.name = "fudge"; + fudge.flags = IORESOURCE_MEM; + dev_priv->gtt_mem = &fudge; + } + + pci_read_config_dword(dev->pdev, PSB_BSM, &dev_priv->stolen_base); + vram_stolen_size = pg->gtt_phys_start - dev_priv->stolen_base + - PAGE_SIZE; + + stolen_size = vram_stolen_size; + + printk(KERN_INFO "Stolen memory information\n"); + printk(KERN_INFO " base in RAM: 0x%x\n", dev_priv->stolen_base); + printk(KERN_INFO " size: %luK, calculated by (GTT RAM base) - (Stolen base), seems wrong\n", + vram_stolen_size/1024); + dvmt_mode = (dev_priv->gmch_ctrl >> 4) & 0x7; + printk(KERN_INFO " the correct size should be: %dM(dvmt mode=%d)\n", + (dvmt_mode == 1) ? 1 : (2 << (dvmt_mode - 1)), dvmt_mode); + + if (resume && (gtt_pages != pg->gtt_pages) && + (stolen_size != pg->stolen_size)) { + dev_err(dev->dev, "GTT resume error.\n"); + ret = -EINVAL; + goto out_err; + } + + pg->gtt_pages = gtt_pages; + pg->stolen_size = stolen_size; + dev_priv->vram_stolen_size = vram_stolen_size; + + /* + * Map the GTT and the stolen memory area + */ + dev_priv->gtt_map = ioremap_nocache(pg->gtt_phys_start, + gtt_pages << PAGE_SHIFT); + if (!dev_priv->gtt_map) { + dev_err(dev->dev, "Failure to map gtt.\n"); + ret = -ENOMEM; + goto out_err; + } + + dev_priv->vram_addr = ioremap_wc(dev_priv->stolen_base, stolen_size); + if (!dev_priv->vram_addr) { + dev_err(dev->dev, "Failure to map stolen base.\n"); + ret = -ENOMEM; + goto out_err; + } + + /* + * Insert vram stolen pages into the GTT + */ + + pfn_base = dev_priv->stolen_base >> PAGE_SHIFT; + vram_pages = num_pages = vram_stolen_size >> PAGE_SHIFT; + printk(KERN_INFO"Set up %d stolen pages starting at 0x%08x, GTT offset %dK\n", + num_pages, pfn_base << PAGE_SHIFT, 0); + for (i = 0; i < num_pages; ++i) { + pte = psb_gtt_mask_pte(pfn_base + i, 0); + iowrite32(pte, dev_priv->gtt_map + i); + } + + /* + * Init rest of GTT to the scratch page to avoid accidents or scribbles + */ + + pfn_base = page_to_pfn(dev_priv->scratch_page); + pte = psb_gtt_mask_pte(pfn_base, 0); + for (; i < gtt_pages; ++i) + iowrite32(pte, dev_priv->gtt_map + i); + + (void) ioread32(dev_priv->gtt_map + i - 1); + return 0; + +out_err: + psb_gtt_takedown(dev); + return ret; +} diff --git a/drivers/gpu/drm/gma500/gtt.h b/drivers/gpu/drm/gma500/gtt.h new file mode 100644 index 0000000..aa17423 --- /dev/null +++ b/drivers/gpu/drm/gma500/gtt.h @@ -0,0 +1,64 @@ +/************************************************************************** + * Copyright (c) 2007-2008, Intel Corporation. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + **************************************************************************/ + +#ifndef _PSB_GTT_H_ +#define _PSB_GTT_H_ + +#include + +/* This wants cleaning up with respect to the psb_dev and un-needed stuff */ +struct psb_gtt { + uint32_t gatt_start; + uint32_t mmu_gatt_start; + uint32_t gtt_start; + uint32_t gtt_phys_start; + unsigned gtt_pages; + unsigned gatt_pages; + unsigned long stolen_size; + unsigned long vram_stolen_size; + struct rw_semaphore sem; +}; + +/* Exported functions */ +extern int psb_gtt_init(struct drm_device *dev, int resume); +extern void psb_gtt_takedown(struct drm_device *dev); + +/* Each gtt_range describes an allocation in the GTT area */ +struct gtt_range { + struct resource resource; /* Resource for our allocation */ + u32 offset; /* GTT offset of our object */ + struct drm_gem_object gem; /* GEM high level stuff */ + int in_gart; /* Currently in the GART (ref ct) */ + bool stolen; /* Backed from stolen RAM */ + bool mmapping; /* Is mmappable */ + struct page **pages; /* Backing pages if present */ + int npage; /* Number of backing pages */ + int roll; /* Roll applied to the GTT entries */ +}; + +extern struct gtt_range *psb_gtt_alloc_range(struct drm_device *dev, int len, + const char *name, int backed); +extern void psb_gtt_kref_put(struct gtt_range *gt); +extern void psb_gtt_free_range(struct drm_device *dev, struct gtt_range *gt); +extern int psb_gtt_pin(struct gtt_range *gt); +extern void psb_gtt_unpin(struct gtt_range *gt); +extern void psb_gtt_roll(struct drm_device *dev, + struct gtt_range *gt, int roll); + +#endif diff --git a/drivers/gpu/drm/gma500/intel_bios.c b/drivers/gpu/drm/gma500/intel_bios.c new file mode 100644 index 0000000..d4d0c5b --- /dev/null +++ b/drivers/gpu/drm/gma500/intel_bios.c @@ -0,0 +1,303 @@ +/* + * Copyright (c) 2006 Intel Corporation + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + * Authors: + * Eric Anholt + * + */ +#include +#include +#include "gma_drm.h" +#include "psb_drv.h" +#include "psb_intel_drv.h" +#include "psb_intel_reg.h" +#include "intel_bios.h" + + +static void *find_section(struct bdb_header *bdb, int section_id) +{ + u8 *base = (u8 *)bdb; + int index = 0; + u16 total, current_size; + u8 current_id; + + /* skip to first section */ + index += bdb->header_size; + total = bdb->bdb_size; + + /* walk the sections looking for section_id */ + while (index < total) { + current_id = *(base + index); + index++; + current_size = *((u16 *)(base + index)); + index += 2; + if (current_id == section_id) + return base + index; + index += current_size; + } + + return NULL; +} + +static void fill_detail_timing_data(struct drm_display_mode *panel_fixed_mode, + struct lvds_dvo_timing *dvo_timing) +{ + panel_fixed_mode->hdisplay = (dvo_timing->hactive_hi << 8) | + dvo_timing->hactive_lo; + panel_fixed_mode->hsync_start = panel_fixed_mode->hdisplay + + ((dvo_timing->hsync_off_hi << 8) | dvo_timing->hsync_off_lo); + panel_fixed_mode->hsync_end = panel_fixed_mode->hsync_start + + dvo_timing->hsync_pulse_width; + panel_fixed_mode->htotal = panel_fixed_mode->hdisplay + + ((dvo_timing->hblank_hi << 8) | dvo_timing->hblank_lo); + + panel_fixed_mode->vdisplay = (dvo_timing->vactive_hi << 8) | + dvo_timing->vactive_lo; + panel_fixed_mode->vsync_start = panel_fixed_mode->vdisplay + + dvo_timing->vsync_off; + panel_fixed_mode->vsync_end = panel_fixed_mode->vsync_start + + dvo_timing->vsync_pulse_width; + panel_fixed_mode->vtotal = panel_fixed_mode->vdisplay + + ((dvo_timing->vblank_hi << 8) | dvo_timing->vblank_lo); + panel_fixed_mode->clock = dvo_timing->clock * 10; + panel_fixed_mode->type = DRM_MODE_TYPE_PREFERRED; + + /* Some VBTs have bogus h/vtotal values */ + if (panel_fixed_mode->hsync_end > panel_fixed_mode->htotal) + panel_fixed_mode->htotal = panel_fixed_mode->hsync_end + 1; + if (panel_fixed_mode->vsync_end > panel_fixed_mode->vtotal) + panel_fixed_mode->vtotal = panel_fixed_mode->vsync_end + 1; + + drm_mode_set_name(panel_fixed_mode); +} + +static void parse_backlight_data(struct drm_psb_private *dev_priv, + struct bdb_header *bdb) +{ + struct bdb_lvds_backlight *vbt_lvds_bl = NULL; + struct bdb_lvds_backlight *lvds_bl; + u8 p_type = 0; + void *bl_start = NULL; + struct bdb_lvds_options *lvds_opts + = find_section(bdb, BDB_LVDS_OPTIONS); + + dev_priv->lvds_bl = NULL; + + if (lvds_opts) + p_type = lvds_opts->panel_type; + else + return; + + bl_start = find_section(bdb, BDB_LVDS_BACKLIGHT); + vbt_lvds_bl = (struct bdb_lvds_backlight *)(bl_start + 1) + p_type; + + lvds_bl = kzalloc(sizeof(*vbt_lvds_bl), GFP_KERNEL); + if (!lvds_bl) { + dev_err(dev_priv->dev->dev, "out of memory for backlight data\n"); + return; + } + memcpy(lvds_bl, vbt_lvds_bl, sizeof(*vbt_lvds_bl)); + dev_priv->lvds_bl = lvds_bl; +} + +/* Try to find integrated panel data */ +static void parse_lfp_panel_data(struct drm_psb_private *dev_priv, + struct bdb_header *bdb) +{ + struct bdb_lvds_options *lvds_options; + struct bdb_lvds_lfp_data *lvds_lfp_data; + struct bdb_lvds_lfp_data_entry *entry; + struct lvds_dvo_timing *dvo_timing; + struct drm_display_mode *panel_fixed_mode; + + /* Defaults if we can't find VBT info */ + dev_priv->lvds_dither = 0; + dev_priv->lvds_vbt = 0; + + lvds_options = find_section(bdb, BDB_LVDS_OPTIONS); + if (!lvds_options) + return; + + dev_priv->lvds_dither = lvds_options->pixel_dither; + if (lvds_options->panel_type == 0xff) + return; + + lvds_lfp_data = find_section(bdb, BDB_LVDS_LFP_DATA); + if (!lvds_lfp_data) + return; + + + entry = &lvds_lfp_data->data[lvds_options->panel_type]; + dvo_timing = &entry->dvo_timing; + + panel_fixed_mode = kzalloc(sizeof(*panel_fixed_mode), + GFP_KERNEL); + if (panel_fixed_mode == NULL) { + dev_err(dev_priv->dev->dev, "out of memory for fixed panel mode\n"); + return; + } + + dev_priv->lvds_vbt = 1; + fill_detail_timing_data(panel_fixed_mode, dvo_timing); + + if (panel_fixed_mode->htotal > 0 && panel_fixed_mode->vtotal > 0) { + dev_priv->lfp_lvds_vbt_mode = panel_fixed_mode; + drm_mode_debug_printmodeline(panel_fixed_mode); + } else { + dev_dbg(dev_priv->dev->dev, "ignoring invalid LVDS VBT\n"); + dev_priv->lvds_vbt = 0; + kfree(panel_fixed_mode); + } + return; +} + +/* Try to find sdvo panel data */ +static void parse_sdvo_panel_data(struct drm_psb_private *dev_priv, + struct bdb_header *bdb) +{ + struct bdb_sdvo_lvds_options *sdvo_lvds_options; + struct lvds_dvo_timing *dvo_timing; + struct drm_display_mode *panel_fixed_mode; + + dev_priv->sdvo_lvds_vbt_mode = NULL; + + sdvo_lvds_options = find_section(bdb, BDB_SDVO_LVDS_OPTIONS); + if (!sdvo_lvds_options) + return; + + dvo_timing = find_section(bdb, BDB_SDVO_PANEL_DTDS); + if (!dvo_timing) + return; + + panel_fixed_mode = kzalloc(sizeof(*panel_fixed_mode), GFP_KERNEL); + + if (!panel_fixed_mode) + return; + + fill_detail_timing_data(panel_fixed_mode, + dvo_timing + sdvo_lvds_options->panel_type); + + dev_priv->sdvo_lvds_vbt_mode = panel_fixed_mode; + + return; +} + +static void parse_general_features(struct drm_psb_private *dev_priv, + struct bdb_header *bdb) +{ + struct bdb_general_features *general; + + /* Set sensible defaults in case we can't find the general block */ + dev_priv->int_tv_support = 1; + dev_priv->int_crt_support = 1; + + general = find_section(bdb, BDB_GENERAL_FEATURES); + if (general) { + dev_priv->int_tv_support = general->int_tv_support; + dev_priv->int_crt_support = general->int_crt_support; + dev_priv->lvds_use_ssc = general->enable_ssc; + + if (dev_priv->lvds_use_ssc) { + dev_priv->lvds_ssc_freq + = general->ssc_freq ? 100 : 96; + } + } +} + +/** + * psb_intel_init_bios - initialize VBIOS settings & find VBT + * @dev: DRM device + * + * Loads the Video BIOS and checks that the VBT exists. Sets scratch registers + * to appropriate values. + * + * VBT existence is a sanity check that is relied on by other i830_bios.c code. + * Note that it would be better to use a BIOS call to get the VBT, as BIOSes may + * feed an updated VBT back through that, compared to what we'll fetch using + * this method of groping around in the BIOS data. + * + * Returns 0 on success, nonzero on failure. + */ +bool psb_intel_init_bios(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct pci_dev *pdev = dev->pdev; + struct vbt_header *vbt = NULL; + struct bdb_header *bdb; + u8 __iomem *bios; + size_t size; + int i; + + bios = pci_map_rom(pdev, &size); + if (!bios) + return -1; + + /* Scour memory looking for the VBT signature */ + for (i = 0; i + 4 < size; i++) { + if (!memcmp(bios + i, "$VBT", 4)) { + vbt = (struct vbt_header *)(bios + i); + break; + } + } + + if (!vbt) { + dev_err(dev->dev, "VBT signature missing\n"); + pci_unmap_rom(pdev, bios); + return -1; + } + + bdb = (struct bdb_header *)(bios + i + vbt->bdb_offset); + + /* Grab useful general definitions */ + parse_general_features(dev_priv, bdb); + parse_lfp_panel_data(dev_priv, bdb); + parse_sdvo_panel_data(dev_priv, bdb); + parse_backlight_data(dev_priv, bdb); + + pci_unmap_rom(pdev, bios); + + return 0; +} + +/** + * Destroy and free VBT data + */ +void psb_intel_destroy_bios(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct drm_display_mode *sdvo_lvds_vbt_mode = + dev_priv->sdvo_lvds_vbt_mode; + struct drm_display_mode *lfp_lvds_vbt_mode = + dev_priv->lfp_lvds_vbt_mode; + struct bdb_lvds_backlight *lvds_bl = + dev_priv->lvds_bl; + + /*free sdvo panel mode*/ + if (sdvo_lvds_vbt_mode) { + dev_priv->sdvo_lvds_vbt_mode = NULL; + kfree(sdvo_lvds_vbt_mode); + } + + if (lfp_lvds_vbt_mode) { + dev_priv->lfp_lvds_vbt_mode = NULL; + kfree(lfp_lvds_vbt_mode); + } + + if (lvds_bl) { + dev_priv->lvds_bl = NULL; + kfree(lvds_bl); + } +} diff --git a/drivers/gpu/drm/gma500/intel_bios.h b/drivers/gpu/drm/gma500/intel_bios.h new file mode 100644 index 0000000..70f1bf0 --- /dev/null +++ b/drivers/gpu/drm/gma500/intel_bios.h @@ -0,0 +1,430 @@ +/* + * Copyright (c) 2006 Intel Corporation + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + * Authors: + * Eric Anholt + * + */ + +#ifndef _I830_BIOS_H_ +#define _I830_BIOS_H_ + +#include + +struct vbt_header { + u8 signature[20]; /**< Always starts with 'VBT$' */ + u16 version; /**< decimal */ + u16 header_size; /**< in bytes */ + u16 vbt_size; /**< in bytes */ + u8 vbt_checksum; + u8 reserved0; + u32 bdb_offset; /**< from beginning of VBT */ + u32 aim_offset[4]; /**< from beginning of VBT */ +} __attribute__((packed)); + + +struct bdb_header { + u8 signature[16]; /**< Always 'BIOS_DATA_BLOCK' */ + u16 version; /**< decimal */ + u16 header_size; /**< in bytes */ + u16 bdb_size; /**< in bytes */ +}; + +/* strictly speaking, this is a "skip" block, but it has interesting info */ +struct vbios_data { + u8 type; /* 0 == desktop, 1 == mobile */ + u8 relstage; + u8 chipset; + u8 lvds_present:1; + u8 tv_present:1; + u8 rsvd2:6; /* finish byte */ + u8 rsvd3[4]; + u8 signon[155]; + u8 copyright[61]; + u16 code_segment; + u8 dos_boot_mode; + u8 bandwidth_percent; + u8 rsvd4; /* popup memory size */ + u8 resize_pci_bios; + u8 rsvd5; /* is crt already on ddc2 */ +} __attribute__((packed)); + +/* + * There are several types of BIOS data blocks (BDBs), each block has + * an ID and size in the first 3 bytes (ID in first, size in next 2). + * Known types are listed below. + */ +#define BDB_GENERAL_FEATURES 1 +#define BDB_GENERAL_DEFINITIONS 2 +#define BDB_OLD_TOGGLE_LIST 3 +#define BDB_MODE_SUPPORT_LIST 4 +#define BDB_GENERIC_MODE_TABLE 5 +#define BDB_EXT_MMIO_REGS 6 +#define BDB_SWF_IO 7 +#define BDB_SWF_MMIO 8 +#define BDB_DOT_CLOCK_TABLE 9 +#define BDB_MODE_REMOVAL_TABLE 10 +#define BDB_CHILD_DEVICE_TABLE 11 +#define BDB_DRIVER_FEATURES 12 +#define BDB_DRIVER_PERSISTENCE 13 +#define BDB_EXT_TABLE_PTRS 14 +#define BDB_DOT_CLOCK_OVERRIDE 15 +#define BDB_DISPLAY_SELECT 16 +/* 17 rsvd */ +#define BDB_DRIVER_ROTATION 18 +#define BDB_DISPLAY_REMOVE 19 +#define BDB_OEM_CUSTOM 20 +#define BDB_EFP_LIST 21 /* workarounds for VGA hsync/vsync */ +#define BDB_SDVO_LVDS_OPTIONS 22 +#define BDB_SDVO_PANEL_DTDS 23 +#define BDB_SDVO_LVDS_PNP_IDS 24 +#define BDB_SDVO_LVDS_POWER_SEQ 25 +#define BDB_TV_OPTIONS 26 +#define BDB_LVDS_OPTIONS 40 +#define BDB_LVDS_LFP_DATA_PTRS 41 +#define BDB_LVDS_LFP_DATA 42 +#define BDB_LVDS_BACKLIGHT 43 +#define BDB_LVDS_POWER 44 +#define BDB_SKIP 254 /* VBIOS private block, ignore */ + +struct bdb_general_features { + /* bits 1 */ + u8 panel_fitting:2; + u8 flexaim:1; + u8 msg_enable:1; + u8 clear_screen:3; + u8 color_flip:1; + + /* bits 2 */ + u8 download_ext_vbt:1; + u8 enable_ssc:1; + u8 ssc_freq:1; + u8 enable_lfp_on_override:1; + u8 disable_ssc_ddt:1; + u8 rsvd8:3; /* finish byte */ + + /* bits 3 */ + u8 disable_smooth_vision:1; + u8 single_dvi:1; + u8 rsvd9:6; /* finish byte */ + + /* bits 4 */ + u8 legacy_monitor_detect; + + /* bits 5 */ + u8 int_crt_support:1; + u8 int_tv_support:1; + u8 rsvd11:6; /* finish byte */ +} __attribute__((packed)); + +struct bdb_general_definitions { + /* DDC GPIO */ + u8 crt_ddc_gmbus_pin; + + /* DPMS bits */ + u8 dpms_acpi:1; + u8 skip_boot_crt_detect:1; + u8 dpms_aim:1; + u8 rsvd1:5; /* finish byte */ + + /* boot device bits */ + u8 boot_display[2]; + u8 child_dev_size; + + /* device info */ + u8 tv_or_lvds_info[33]; + u8 dev1[33]; + u8 dev2[33]; + u8 dev3[33]; + u8 dev4[33]; + /* may be another device block here on some platforms */ +}; + +struct bdb_lvds_options { + u8 panel_type; + u8 rsvd1; + /* LVDS capabilities, stored in a dword */ + u8 pfit_mode:2; + u8 pfit_text_mode_enhanced:1; + u8 pfit_gfx_mode_enhanced:1; + u8 pfit_ratio_auto:1; + u8 pixel_dither:1; + u8 lvds_edid:1; + u8 rsvd2:1; + u8 rsvd4; +} __attribute__((packed)); + +struct bdb_lvds_backlight { + u8 type:2; + u8 pol:1; + u8 gpio:3; + u8 gmbus:2; + u16 freq; + u8 minbrightness; + u8 i2caddr; + u8 brightnesscmd; + /*FIXME: more...*/ +} __attribute__((packed)); + +/* LFP pointer table contains entries to the struct below */ +struct bdb_lvds_lfp_data_ptr { + u16 fp_timing_offset; /* offsets are from start of bdb */ + u8 fp_table_size; + u16 dvo_timing_offset; + u8 dvo_table_size; + u16 panel_pnp_id_offset; + u8 pnp_table_size; +} __attribute__((packed)); + +struct bdb_lvds_lfp_data_ptrs { + u8 lvds_entries; /* followed by one or more lvds_data_ptr structs */ + struct bdb_lvds_lfp_data_ptr ptr[16]; +} __attribute__((packed)); + +/* LFP data has 3 blocks per entry */ +struct lvds_fp_timing { + u16 x_res; + u16 y_res; + u32 lvds_reg; + u32 lvds_reg_val; + u32 pp_on_reg; + u32 pp_on_reg_val; + u32 pp_off_reg; + u32 pp_off_reg_val; + u32 pp_cycle_reg; + u32 pp_cycle_reg_val; + u32 pfit_reg; + u32 pfit_reg_val; + u16 terminator; +} __attribute__((packed)); + +struct lvds_dvo_timing { + u16 clock; /**< In 10khz */ + u8 hactive_lo; + u8 hblank_lo; + u8 hblank_hi:4; + u8 hactive_hi:4; + u8 vactive_lo; + u8 vblank_lo; + u8 vblank_hi:4; + u8 vactive_hi:4; + u8 hsync_off_lo; + u8 hsync_pulse_width; + u8 vsync_pulse_width:4; + u8 vsync_off:4; + u8 rsvd0:6; + u8 hsync_off_hi:2; + u8 h_image; + u8 v_image; + u8 max_hv; + u8 h_border; + u8 v_border; + u8 rsvd1:3; + u8 digital:2; + u8 vsync_positive:1; + u8 hsync_positive:1; + u8 rsvd2:1; +} __attribute__((packed)); + +struct lvds_pnp_id { + u16 mfg_name; + u16 product_code; + u32 serial; + u8 mfg_week; + u8 mfg_year; +} __attribute__((packed)); + +struct bdb_lvds_lfp_data_entry { + struct lvds_fp_timing fp_timing; + struct lvds_dvo_timing dvo_timing; + struct lvds_pnp_id pnp_id; +} __attribute__((packed)); + +struct bdb_lvds_lfp_data { + struct bdb_lvds_lfp_data_entry data[16]; +} __attribute__((packed)); + +struct aimdb_header { + char signature[16]; + char oem_device[20]; + u16 aimdb_version; + u16 aimdb_header_size; + u16 aimdb_size; +} __attribute__((packed)); + +struct aimdb_block { + u8 aimdb_id; + u16 aimdb_size; +} __attribute__((packed)); + +struct vch_panel_data { + u16 fp_timing_offset; + u8 fp_timing_size; + u16 dvo_timing_offset; + u8 dvo_timing_size; + u16 text_fitting_offset; + u8 text_fitting_size; + u16 graphics_fitting_offset; + u8 graphics_fitting_size; +} __attribute__((packed)); + +struct vch_bdb_22 { + struct aimdb_block aimdb_block; + struct vch_panel_data panels[16]; +} __attribute__((packed)); + +struct bdb_sdvo_lvds_options { + u8 panel_backlight; + u8 h40_set_panel_type; + u8 panel_type; + u8 ssc_clk_freq; + u16 als_low_trip; + u16 als_high_trip; + u8 sclalarcoeff_tab_row_num; + u8 sclalarcoeff_tab_row_size; + u8 coefficient[8]; + u8 panel_misc_bits_1; + u8 panel_misc_bits_2; + u8 panel_misc_bits_3; + u8 panel_misc_bits_4; +} __attribute__((packed)); + + +extern bool psb_intel_init_bios(struct drm_device *dev); +extern void psb_intel_destroy_bios(struct drm_device *dev); + +/* + * Driver<->VBIOS interaction occurs through scratch bits in + * GR18 & SWF*. + */ + +/* GR18 bits are set on display switch and hotkey events */ +#define GR18_DRIVER_SWITCH_EN (1<<7) /* 0: VBIOS control, 1: driver control */ +#define GR18_HOTKEY_MASK 0x78 /* See also SWF4 15:0 */ +#define GR18_HK_NONE (0x0<<3) +#define GR18_HK_LFP_STRETCH (0x1<<3) +#define GR18_HK_TOGGLE_DISP (0x2<<3) +#define GR18_HK_DISP_SWITCH (0x4<<3) /* see SWF14 15:0 for what to enable */ +#define GR18_HK_POPUP_DISABLED (0x6<<3) +#define GR18_HK_POPUP_ENABLED (0x7<<3) +#define GR18_HK_PFIT (0x8<<3) +#define GR18_HK_APM_CHANGE (0xa<<3) +#define GR18_HK_MULTIPLE (0xc<<3) +#define GR18_USER_INT_EN (1<<2) +#define GR18_A0000_FLUSH_EN (1<<1) +#define GR18_SMM_EN (1<<0) + +/* Set by driver, cleared by VBIOS */ +#define SWF00_YRES_SHIFT 16 +#define SWF00_XRES_SHIFT 0 +#define SWF00_RES_MASK 0xffff + +/* Set by VBIOS at boot time and driver at runtime */ +#define SWF01_TV2_FORMAT_SHIFT 8 +#define SWF01_TV1_FORMAT_SHIFT 0 +#define SWF01_TV_FORMAT_MASK 0xffff + +#define SWF10_VBIOS_BLC_I2C_EN (1<<29) +#define SWF10_GTT_OVERRIDE_EN (1<<28) +#define SWF10_LFP_DPMS_OVR (1<<27) /* override DPMS on display switch */ +#define SWF10_ACTIVE_TOGGLE_LIST_MASK (7<<24) +#define SWF10_OLD_TOGGLE 0x0 +#define SWF10_TOGGLE_LIST_1 0x1 +#define SWF10_TOGGLE_LIST_2 0x2 +#define SWF10_TOGGLE_LIST_3 0x3 +#define SWF10_TOGGLE_LIST_4 0x4 +#define SWF10_PANNING_EN (1<<23) +#define SWF10_DRIVER_LOADED (1<<22) +#define SWF10_EXTENDED_DESKTOP (1<<21) +#define SWF10_EXCLUSIVE_MODE (1<<20) +#define SWF10_OVERLAY_EN (1<<19) +#define SWF10_PLANEB_HOLDOFF (1<<18) +#define SWF10_PLANEA_HOLDOFF (1<<17) +#define SWF10_VGA_HOLDOFF (1<<16) +#define SWF10_ACTIVE_DISP_MASK 0xffff +#define SWF10_PIPEB_LFP2 (1<<15) +#define SWF10_PIPEB_EFP2 (1<<14) +#define SWF10_PIPEB_TV2 (1<<13) +#define SWF10_PIPEB_CRT2 (1<<12) +#define SWF10_PIPEB_LFP (1<<11) +#define SWF10_PIPEB_EFP (1<<10) +#define SWF10_PIPEB_TV (1<<9) +#define SWF10_PIPEB_CRT (1<<8) +#define SWF10_PIPEA_LFP2 (1<<7) +#define SWF10_PIPEA_EFP2 (1<<6) +#define SWF10_PIPEA_TV2 (1<<5) +#define SWF10_PIPEA_CRT2 (1<<4) +#define SWF10_PIPEA_LFP (1<<3) +#define SWF10_PIPEA_EFP (1<<2) +#define SWF10_PIPEA_TV (1<<1) +#define SWF10_PIPEA_CRT (1<<0) + +#define SWF11_MEMORY_SIZE_SHIFT 16 +#define SWF11_SV_TEST_EN (1<<15) +#define SWF11_IS_AGP (1<<14) +#define SWF11_DISPLAY_HOLDOFF (1<<13) +#define SWF11_DPMS_REDUCED (1<<12) +#define SWF11_IS_VBE_MODE (1<<11) +#define SWF11_PIPEB_ACCESS (1<<10) /* 0 here means pipe a */ +#define SWF11_DPMS_MASK 0x07 +#define SWF11_DPMS_OFF (1<<2) +#define SWF11_DPMS_SUSPEND (1<<1) +#define SWF11_DPMS_STANDBY (1<<0) +#define SWF11_DPMS_ON 0 + +#define SWF14_GFX_PFIT_EN (1<<31) +#define SWF14_TEXT_PFIT_EN (1<<30) +#define SWF14_LID_STATUS_CLOSED (1<<29) /* 0 here means open */ +#define SWF14_POPUP_EN (1<<28) +#define SWF14_DISPLAY_HOLDOFF (1<<27) +#define SWF14_DISP_DETECT_EN (1<<26) +#define SWF14_DOCKING_STATUS_DOCKED (1<<25) /* 0 here means undocked */ +#define SWF14_DRIVER_STATUS (1<<24) +#define SWF14_OS_TYPE_WIN9X (1<<23) +#define SWF14_OS_TYPE_WINNT (1<<22) +/* 21:19 rsvd */ +#define SWF14_PM_TYPE_MASK 0x00070000 +#define SWF14_PM_ACPI_VIDEO (0x4 << 16) +#define SWF14_PM_ACPI (0x3 << 16) +#define SWF14_PM_APM_12 (0x2 << 16) +#define SWF14_PM_APM_11 (0x1 << 16) +#define SWF14_HK_REQUEST_MASK 0x0000ffff /* see GR18 6:3 for event type */ + /* if GR18 indicates a display switch */ +#define SWF14_DS_PIPEB_LFP2_EN (1<<15) +#define SWF14_DS_PIPEB_EFP2_EN (1<<14) +#define SWF14_DS_PIPEB_TV2_EN (1<<13) +#define SWF14_DS_PIPEB_CRT2_EN (1<<12) +#define SWF14_DS_PIPEB_LFP_EN (1<<11) +#define SWF14_DS_PIPEB_EFP_EN (1<<10) +#define SWF14_DS_PIPEB_TV_EN (1<<9) +#define SWF14_DS_PIPEB_CRT_EN (1<<8) +#define SWF14_DS_PIPEA_LFP2_EN (1<<7) +#define SWF14_DS_PIPEA_EFP2_EN (1<<6) +#define SWF14_DS_PIPEA_TV2_EN (1<<5) +#define SWF14_DS_PIPEA_CRT2_EN (1<<4) +#define SWF14_DS_PIPEA_LFP_EN (1<<3) +#define SWF14_DS_PIPEA_EFP_EN (1<<2) +#define SWF14_DS_PIPEA_TV_EN (1<<1) +#define SWF14_DS_PIPEA_CRT_EN (1<<0) + /* if GR18 indicates a panel fitting request */ +#define SWF14_PFIT_EN (1<<0) /* 0 means disable */ + /* if GR18 indicates an APM change request */ +#define SWF14_APM_HIBERNATE 0x4 +#define SWF14_APM_SUSPEND 0x3 +#define SWF14_APM_STANDBY 0x1 +#define SWF14_APM_RESTORE 0x0 + +#endif /* _I830_BIOS_H_ */ diff --git a/drivers/gpu/drm/gma500/intel_gmbus.c b/drivers/gpu/drm/gma500/intel_gmbus.c new file mode 100644 index 0000000..9db9052 --- /dev/null +++ b/drivers/gpu/drm/gma500/intel_gmbus.c @@ -0,0 +1,493 @@ +/* + * Copyright (c) 2006 Dave Airlie + * Copyright © 2006-2008,2010 Intel Corporation + * Jesse Barnes + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * Eric Anholt + * Chris Wilson + */ +#include +#include +#include +#include "drmP.h" +#include "drm.h" +#include "psb_intel_drv.h" +#include "gma_drm.h" +#include "psb_drv.h" +#include "psb_intel_reg.h" + +#define _wait_for(COND, MS, W) ({ \ + unsigned long timeout__ = jiffies + msecs_to_jiffies(MS); \ + int ret__ = 0; \ + while (! (COND)) { \ + if (time_after(jiffies, timeout__)) { \ + ret__ = -ETIMEDOUT; \ + break; \ + } \ + if (W && !(in_atomic() || in_dbg_master())) msleep(W); \ + } \ + ret__; \ +}) + +#define wait_for(COND, MS) _wait_for(COND, MS, 1) +#define wait_for_atomic(COND, MS) _wait_for(COND, MS, 0) + +/* Intel GPIO access functions */ + +#define I2C_RISEFALL_TIME 20 + +static inline struct intel_gmbus * +to_intel_gmbus(struct i2c_adapter *i2c) +{ + return container_of(i2c, struct intel_gmbus, adapter); +} + +struct intel_gpio { + struct i2c_adapter adapter; + struct i2c_algo_bit_data algo; + struct drm_psb_private *dev_priv; + u32 reg; +}; + +void +gma_intel_i2c_reset(struct drm_device *dev) +{ + REG_WRITE(GMBUS0, 0); +} + +static void intel_i2c_quirk_set(struct drm_psb_private *dev_priv, bool enable) +{ + /* When using bit bashing for I2C, this bit needs to be set to 1 */ + /* FIXME: We are never Pineview, right? + + u32 val; + + if (!IS_PINEVIEW(dev_priv->dev)) + return; + + val = REG_READ(DSPCLK_GATE_D); + if (enable) + val |= DPCUNIT_CLOCK_GATE_DISABLE; + else + val &= ~DPCUNIT_CLOCK_GATE_DISABLE; + REG_WRITE(DSPCLK_GATE_D, val); + + return; + */ +} + +static u32 get_reserved(struct intel_gpio *gpio) +{ + struct drm_psb_private *dev_priv = gpio->dev_priv; + struct drm_device *dev = dev_priv->dev; + u32 reserved = 0; + + /* On most chips, these bits must be preserved in software. */ + reserved = REG_READ(gpio->reg) & + (GPIO_DATA_PULLUP_DISABLE | + GPIO_CLOCK_PULLUP_DISABLE); + + return reserved; +} + +static int get_clock(void *data) +{ + struct intel_gpio *gpio = data; + struct drm_psb_private *dev_priv = gpio->dev_priv; + struct drm_device *dev = dev_priv->dev; + u32 reserved = get_reserved(gpio); + REG_WRITE(gpio->reg, reserved | GPIO_CLOCK_DIR_MASK); + REG_WRITE(gpio->reg, reserved); + return (REG_READ(gpio->reg) & GPIO_CLOCK_VAL_IN) != 0; +} + +static int get_data(void *data) +{ + struct intel_gpio *gpio = data; + struct drm_psb_private *dev_priv = gpio->dev_priv; + struct drm_device *dev = dev_priv->dev; + u32 reserved = get_reserved(gpio); + REG_WRITE(gpio->reg, reserved | GPIO_DATA_DIR_MASK); + REG_WRITE(gpio->reg, reserved); + return (REG_READ(gpio->reg) & GPIO_DATA_VAL_IN) != 0; +} + +static void set_clock(void *data, int state_high) +{ + struct intel_gpio *gpio = data; + struct drm_psb_private *dev_priv = gpio->dev_priv; + struct drm_device *dev = dev_priv->dev; + u32 reserved = get_reserved(gpio); + u32 clock_bits; + + if (state_high) + clock_bits = GPIO_CLOCK_DIR_IN | GPIO_CLOCK_DIR_MASK; + else + clock_bits = GPIO_CLOCK_DIR_OUT | GPIO_CLOCK_DIR_MASK | + GPIO_CLOCK_VAL_MASK; + + REG_WRITE(gpio->reg, reserved | clock_bits); + REG_READ(gpio->reg); /* Posting */ +} + +static void set_data(void *data, int state_high) +{ + struct intel_gpio *gpio = data; + struct drm_psb_private *dev_priv = gpio->dev_priv; + struct drm_device *dev = dev_priv->dev; + u32 reserved = get_reserved(gpio); + u32 data_bits; + + if (state_high) + data_bits = GPIO_DATA_DIR_IN | GPIO_DATA_DIR_MASK; + else + data_bits = GPIO_DATA_DIR_OUT | GPIO_DATA_DIR_MASK | + GPIO_DATA_VAL_MASK; + + REG_WRITE(gpio->reg, reserved | data_bits); + REG_READ(gpio->reg); +} + +static struct i2c_adapter * +intel_gpio_create(struct drm_psb_private *dev_priv, u32 pin) +{ + static const int map_pin_to_reg[] = { + 0, + GPIOB, + GPIOA, + GPIOC, + GPIOD, + GPIOE, + 0, + GPIOF, + }; + struct intel_gpio *gpio; + + if (pin >= ARRAY_SIZE(map_pin_to_reg) || !map_pin_to_reg[pin]) + return NULL; + + gpio = kzalloc(sizeof(struct intel_gpio), GFP_KERNEL); + if (gpio == NULL) + return NULL; + + gpio->reg = map_pin_to_reg[pin]; + gpio->dev_priv = dev_priv; + + snprintf(gpio->adapter.name, sizeof(gpio->adapter.name), + "gma500 GPIO%c", "?BACDE?F"[pin]); + gpio->adapter.owner = THIS_MODULE; + gpio->adapter.algo_data = &gpio->algo; + gpio->adapter.dev.parent = &dev_priv->dev->pdev->dev; + gpio->algo.setsda = set_data; + gpio->algo.setscl = set_clock; + gpio->algo.getsda = get_data; + gpio->algo.getscl = get_clock; + gpio->algo.udelay = I2C_RISEFALL_TIME; + gpio->algo.timeout = usecs_to_jiffies(2200); + gpio->algo.data = gpio; + + if (i2c_bit_add_bus(&gpio->adapter)) + goto out_free; + + return &gpio->adapter; + +out_free: + kfree(gpio); + return NULL; +} + +static int +intel_i2c_quirk_xfer(struct drm_psb_private *dev_priv, + struct i2c_adapter *adapter, + struct i2c_msg *msgs, + int num) +{ + struct intel_gpio *gpio = container_of(adapter, + struct intel_gpio, + adapter); + int ret; + + gma_intel_i2c_reset(dev_priv->dev); + + intel_i2c_quirk_set(dev_priv, true); + set_data(gpio, 1); + set_clock(gpio, 1); + udelay(I2C_RISEFALL_TIME); + + ret = adapter->algo->master_xfer(adapter, msgs, num); + + set_data(gpio, 1); + set_clock(gpio, 1); + intel_i2c_quirk_set(dev_priv, false); + + return ret; +} + +static int +gmbus_xfer(struct i2c_adapter *adapter, + struct i2c_msg *msgs, + int num) +{ + struct intel_gmbus *bus = container_of(adapter, + struct intel_gmbus, + adapter); + struct drm_psb_private *dev_priv = adapter->algo_data; + struct drm_device *dev = dev_priv->dev; + int i, reg_offset; + + if (bus->force_bit) + return intel_i2c_quirk_xfer(dev_priv, + bus->force_bit, msgs, num); + + reg_offset = 0; + + REG_WRITE(GMBUS0 + reg_offset, bus->reg0); + + for (i = 0; i < num; i++) { + u16 len = msgs[i].len; + u8 *buf = msgs[i].buf; + + if (msgs[i].flags & I2C_M_RD) { + REG_WRITE(GMBUS1 + reg_offset, + GMBUS_CYCLE_WAIT | (i + 1 == num ? GMBUS_CYCLE_STOP : 0) | + (len << GMBUS_BYTE_COUNT_SHIFT) | + (msgs[i].addr << GMBUS_SLAVE_ADDR_SHIFT) | + GMBUS_SLAVE_READ | GMBUS_SW_RDY); + REG_READ(GMBUS2+reg_offset); + do { + u32 val, loop = 0; + + if (wait_for(REG_READ(GMBUS2 + reg_offset) & (GMBUS_SATOER | GMBUS_HW_RDY), 50)) + goto timeout; + if (REG_READ(GMBUS2 + reg_offset) & GMBUS_SATOER) + goto clear_err; + + val = REG_READ(GMBUS3 + reg_offset); + do { + *buf++ = val & 0xff; + val >>= 8; + } while (--len && ++loop < 4); + } while (len); + } else { + u32 val, loop; + + val = loop = 0; + do { + val |= *buf++ << (8 * loop); + } while (--len && ++loop < 4); + + REG_WRITE(GMBUS3 + reg_offset, val); + REG_WRITE(GMBUS1 + reg_offset, + (i + 1 == num ? GMBUS_CYCLE_STOP : GMBUS_CYCLE_WAIT) | + (msgs[i].len << GMBUS_BYTE_COUNT_SHIFT) | + (msgs[i].addr << GMBUS_SLAVE_ADDR_SHIFT) | + GMBUS_SLAVE_WRITE | GMBUS_SW_RDY); + REG_READ(GMBUS2+reg_offset); + + while (len) { + if (wait_for(REG_READ(GMBUS2 + reg_offset) & (GMBUS_SATOER | GMBUS_HW_RDY), 50)) + goto timeout; + if (REG_READ(GMBUS2 + reg_offset) & GMBUS_SATOER) + goto clear_err; + + val = loop = 0; + do { + val |= *buf++ << (8 * loop); + } while (--len && ++loop < 4); + + REG_WRITE(GMBUS3 + reg_offset, val); + REG_READ(GMBUS2+reg_offset); + } + } + + if (i + 1 < num && wait_for(REG_READ(GMBUS2 + reg_offset) & (GMBUS_SATOER | GMBUS_HW_WAIT_PHASE), 50)) + goto timeout; + if (REG_READ(GMBUS2 + reg_offset) & GMBUS_SATOER) + goto clear_err; + } + + goto done; + +clear_err: + /* Toggle the Software Clear Interrupt bit. This has the effect + * of resetting the GMBUS controller and so clearing the + * BUS_ERROR raised by the slave's NAK. + */ + REG_WRITE(GMBUS1 + reg_offset, GMBUS_SW_CLR_INT); + REG_WRITE(GMBUS1 + reg_offset, 0); + +done: + /* Mark the GMBUS interface as disabled. We will re-enable it at the + * start of the next xfer, till then let it sleep. + */ + REG_WRITE(GMBUS0 + reg_offset, 0); + return i; + +timeout: + DRM_INFO("GMBUS timed out, falling back to bit banging on pin %d [%s]\n", + bus->reg0 & 0xff, bus->adapter.name); + REG_WRITE(GMBUS0 + reg_offset, 0); + + /* Hardware may not support GMBUS over these pins? Try GPIO bitbanging instead. */ + bus->force_bit = intel_gpio_create(dev_priv, bus->reg0 & 0xff); + if (!bus->force_bit) + return -ENOMEM; + + return intel_i2c_quirk_xfer(dev_priv, bus->force_bit, msgs, num); +} + +static u32 gmbus_func(struct i2c_adapter *adapter) +{ + struct intel_gmbus *bus = container_of(adapter, + struct intel_gmbus, + adapter); + + if (bus->force_bit) + bus->force_bit->algo->functionality(bus->force_bit); + + return (I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL | + /* I2C_FUNC_10BIT_ADDR | */ + I2C_FUNC_SMBUS_READ_BLOCK_DATA | + I2C_FUNC_SMBUS_BLOCK_PROC_CALL); +} + +static const struct i2c_algorithm gmbus_algorithm = { + .master_xfer = gmbus_xfer, + .functionality = gmbus_func +}; + +/** + * intel_gmbus_setup - instantiate all Intel i2c GMBuses + * @dev: DRM device + */ +int gma_intel_setup_gmbus(struct drm_device *dev) +{ + static const char *names[GMBUS_NUM_PORTS] = { + "disabled", + "ssc", + "vga", + "panel", + "dpc", + "dpb", + "reserved", + "dpd", + }; + struct drm_psb_private *dev_priv = dev->dev_private; + int ret, i; + + dev_priv->gmbus = kcalloc(GMBUS_NUM_PORTS, sizeof(struct intel_gmbus), + GFP_KERNEL); + if (dev_priv->gmbus == NULL) + return -ENOMEM; + + for (i = 0; i < GMBUS_NUM_PORTS; i++) { + struct intel_gmbus *bus = &dev_priv->gmbus[i]; + + bus->adapter.owner = THIS_MODULE; + bus->adapter.class = I2C_CLASS_DDC; + snprintf(bus->adapter.name, + sizeof(bus->adapter.name), + "gma500 gmbus %s", + names[i]); + + bus->adapter.dev.parent = &dev->pdev->dev; + bus->adapter.algo_data = dev_priv; + + bus->adapter.algo = &gmbus_algorithm; + ret = i2c_add_adapter(&bus->adapter); + if (ret) + goto err; + + /* By default use a conservative clock rate */ + bus->reg0 = i | GMBUS_RATE_100KHZ; + + /* XXX force bit banging until GMBUS is fully debugged */ + bus->force_bit = intel_gpio_create(dev_priv, i); + } + + gma_intel_i2c_reset(dev_priv->dev); + + return 0; + +err: + while (--i) { + struct intel_gmbus *bus = &dev_priv->gmbus[i]; + i2c_del_adapter(&bus->adapter); + } + kfree(dev_priv->gmbus); + dev_priv->gmbus = NULL; + return ret; +} + +void gma_intel_gmbus_set_speed(struct i2c_adapter *adapter, int speed) +{ + struct intel_gmbus *bus = to_intel_gmbus(adapter); + + /* speed: + * 0x0 = 100 KHz + * 0x1 = 50 KHz + * 0x2 = 400 KHz + * 0x3 = 1000 Khz + */ + bus->reg0 = (bus->reg0 & ~(0x3 << 8)) | (speed << 8); +} + +void gma_intel_gmbus_force_bit(struct i2c_adapter *adapter, bool force_bit) +{ + struct intel_gmbus *bus = to_intel_gmbus(adapter); + + if (force_bit) { + if (bus->force_bit == NULL) { + struct drm_psb_private *dev_priv = adapter->algo_data; + bus->force_bit = intel_gpio_create(dev_priv, + bus->reg0 & 0xff); + } + } else { + if (bus->force_bit) { + i2c_del_adapter(bus->force_bit); + kfree(bus->force_bit); + bus->force_bit = NULL; + } + } +} + +void gma_intel_teardown_gmbus(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + int i; + + if (dev_priv->gmbus == NULL) + return; + + for (i = 0; i < GMBUS_NUM_PORTS; i++) { + struct intel_gmbus *bus = &dev_priv->gmbus[i]; + if (bus->force_bit) { + i2c_del_adapter(bus->force_bit); + kfree(bus->force_bit); + } + i2c_del_adapter(&bus->adapter); + } + + kfree(dev_priv->gmbus); + dev_priv->gmbus = NULL; +} diff --git a/drivers/gpu/drm/gma500/intel_i2c.c b/drivers/gpu/drm/gma500/intel_i2c.c new file mode 100644 index 0000000..98a28c2 --- /dev/null +++ b/drivers/gpu/drm/gma500/intel_i2c.c @@ -0,0 +1,169 @@ +/* + * Copyright © 2006-2007 Intel Corporation + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + * Authors: + * Eric Anholt + */ +#include +#include +#include + +#include "psb_drv.h" +#include "psb_intel_reg.h" + +/* + * Intel GPIO access functions + */ + +#define I2C_RISEFALL_TIME 20 + +static int get_clock(void *data) +{ + struct psb_intel_i2c_chan *chan = data; + struct drm_device *dev = chan->drm_dev; + u32 val; + + val = REG_READ(chan->reg); + return (val & GPIO_CLOCK_VAL_IN) != 0; +} + +static int get_data(void *data) +{ + struct psb_intel_i2c_chan *chan = data; + struct drm_device *dev = chan->drm_dev; + u32 val; + + val = REG_READ(chan->reg); + return (val & GPIO_DATA_VAL_IN) != 0; +} + +static void set_clock(void *data, int state_high) +{ + struct psb_intel_i2c_chan *chan = data; + struct drm_device *dev = chan->drm_dev; + u32 reserved = 0, clock_bits; + + /* On most chips, these bits must be preserved in software. */ + reserved = + REG_READ(chan->reg) & (GPIO_DATA_PULLUP_DISABLE | + GPIO_CLOCK_PULLUP_DISABLE); + + if (state_high) + clock_bits = GPIO_CLOCK_DIR_IN | GPIO_CLOCK_DIR_MASK; + else + clock_bits = GPIO_CLOCK_DIR_OUT | GPIO_CLOCK_DIR_MASK | + GPIO_CLOCK_VAL_MASK; + REG_WRITE(chan->reg, reserved | clock_bits); + udelay(I2C_RISEFALL_TIME); /* wait for the line to change state */ +} + +static void set_data(void *data, int state_high) +{ + struct psb_intel_i2c_chan *chan = data; + struct drm_device *dev = chan->drm_dev; + u32 reserved = 0, data_bits; + + /* On most chips, these bits must be preserved in software. */ + reserved = + REG_READ(chan->reg) & (GPIO_DATA_PULLUP_DISABLE | + GPIO_CLOCK_PULLUP_DISABLE); + + if (state_high) + data_bits = GPIO_DATA_DIR_IN | GPIO_DATA_DIR_MASK; + else + data_bits = + GPIO_DATA_DIR_OUT | GPIO_DATA_DIR_MASK | + GPIO_DATA_VAL_MASK; + + REG_WRITE(chan->reg, reserved | data_bits); + udelay(I2C_RISEFALL_TIME); /* wait for the line to change state */ +} + +/** + * psb_intel_i2c_create - instantiate an Intel i2c bus using the specified GPIO reg + * @dev: DRM device + * @output: driver specific output device + * @reg: GPIO reg to use + * @name: name for this bus + * + * Creates and registers a new i2c bus with the Linux i2c layer, for use + * in output probing and control (e.g. DDC or SDVO control functions). + * + * Possible values for @reg include: + * %GPIOA + * %GPIOB + * %GPIOC + * %GPIOD + * %GPIOE + * %GPIOF + * %GPIOG + * %GPIOH + * see PRM for details on how these different busses are used. + */ +struct psb_intel_i2c_chan *psb_intel_i2c_create(struct drm_device *dev, + const u32 reg, const char *name) +{ + struct psb_intel_i2c_chan *chan; + + chan = kzalloc(sizeof(struct psb_intel_i2c_chan), GFP_KERNEL); + if (!chan) + goto out_free; + + chan->drm_dev = dev; + chan->reg = reg; + snprintf(chan->adapter.name, I2C_NAME_SIZE, "intel drm %s", name); + chan->adapter.owner = THIS_MODULE; + chan->adapter.algo_data = &chan->algo; + chan->adapter.dev.parent = &dev->pdev->dev; + chan->algo.setsda = set_data; + chan->algo.setscl = set_clock; + chan->algo.getsda = get_data; + chan->algo.getscl = get_clock; + chan->algo.udelay = 20; + chan->algo.timeout = usecs_to_jiffies(2200); + chan->algo.data = chan; + + i2c_set_adapdata(&chan->adapter, chan); + + if (i2c_bit_add_bus(&chan->adapter)) + goto out_free; + + /* JJJ: raise SCL and SDA? */ + set_data(chan, 1); + set_clock(chan, 1); + udelay(20); + + return chan; + +out_free: + kfree(chan); + return NULL; +} + +/** + * psb_intel_i2c_destroy - unregister and free i2c bus resources + * @output: channel to free + * + * Unregister the adapter from the i2c layer, then free the structure. + */ +void psb_intel_i2c_destroy(struct psb_intel_i2c_chan *chan) +{ + if (!chan) + return; + + i2c_del_adapter(&chan->adapter); + kfree(chan); +} diff --git a/drivers/gpu/drm/gma500/intel_opregion.c b/drivers/gpu/drm/gma500/intel_opregion.c new file mode 100644 index 0000000..d946bc1 --- /dev/null +++ b/drivers/gpu/drm/gma500/intel_opregion.c @@ -0,0 +1,81 @@ +/* + * Copyright 2010 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * FIXME: resolve with the i915 version + */ + +#include "psb_drv.h" + +struct opregion_header { + u8 signature[16]; + u32 size; + u32 opregion_ver; + u8 bios_ver[32]; + u8 vbios_ver[16]; + u8 driver_ver[16]; + u32 mboxes; + u8 reserved[164]; +} __packed; + +struct opregion_apci { + /*FIXME: add it later*/ +} __packed; + +struct opregion_swsci { + /*FIXME: add it later*/ +} __packed; + +struct opregion_acpi { + /*FIXME: add it later*/ +} __packed; + +int gma_intel_opregion_init(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + u32 opregion_phy; + void *base; + u32 *lid_state; + + dev_priv->lid_state = NULL; + + pci_read_config_dword(dev->pdev, 0xfc, &opregion_phy); + if (opregion_phy == 0) + return -ENOTSUPP; + + base = ioremap(opregion_phy, 8*1024); + if (!base) + return -ENOMEM; + + lid_state = base + 0x01ac; + + dev_priv->lid_state = lid_state; + dev_priv->lid_last_state = readl(lid_state); + return 0; +} + +int gma_intel_opregion_exit(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + if (dev_priv->lid_state) + iounmap(dev_priv->lid_state); + return 0; +} diff --git a/drivers/gpu/drm/gma500/mdfld_device.c b/drivers/gpu/drm/gma500/mdfld_device.c new file mode 100644 index 0000000..af65678 --- /dev/null +++ b/drivers/gpu/drm/gma500/mdfld_device.c @@ -0,0 +1,691 @@ +/************************************************************************** + * Copyright (c) 2011, Intel Corporation. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + **************************************************************************/ + +#include "psb_drv.h" +#include "mid_bios.h" +#include "mdfld_output.h" +#include "mdfld_dsi_output.h" +#include "tc35876x-dsi-lvds.h" + +#include + +#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE + +#define MRST_BLC_MAX_PWM_REG_FREQ 0xFFFF +#define BLC_PWM_PRECISION_FACTOR 100 /* 10000000 */ +#define BLC_PWM_FREQ_CALC_CONSTANT 32 +#define MHz 1000000 +#define BRIGHTNESS_MIN_LEVEL 1 +#define BRIGHTNESS_MAX_LEVEL 100 +#define BRIGHTNESS_MASK 0xFF +#define BLC_POLARITY_NORMAL 0 +#define BLC_POLARITY_INVERSE 1 +#define BLC_ADJUSTMENT_MAX 100 + +#define MDFLD_BLC_PWM_PRECISION_FACTOR 10 +#define MDFLD_BLC_MAX_PWM_REG_FREQ 0xFFFE +#define MDFLD_BLC_MIN_PWM_REG_FREQ 0x2 + +#define MDFLD_BACKLIGHT_PWM_POLARITY_BIT_CLEAR (0xFFFE) +#define MDFLD_BACKLIGHT_PWM_CTL_SHIFT (16) + +static struct backlight_device *mdfld_backlight_device; + +int mdfld_set_brightness(struct backlight_device *bd) +{ + struct drm_device *dev = + (struct drm_device *)bl_get_data(mdfld_backlight_device); + struct drm_psb_private *dev_priv = dev->dev_private; + int level = bd->props.brightness; + + DRM_DEBUG_DRIVER("backlight level set to %d\n", level); + + /* Perform value bounds checking */ + if (level < BRIGHTNESS_MIN_LEVEL) + level = BRIGHTNESS_MIN_LEVEL; + + if (gma_power_begin(dev, false)) { + u32 adjusted_level = 0; + + /* + * Adjust the backlight level with the percent in + * dev_priv->blc_adj2 + */ + adjusted_level = level * dev_priv->blc_adj2; + adjusted_level = adjusted_level / BLC_ADJUSTMENT_MAX; + dev_priv->brightness_adjusted = adjusted_level; + + if (mdfld_get_panel_type(dev, 0) == TC35876X) { + if (dev_priv->dpi_panel_on[0] || + dev_priv->dpi_panel_on[2]) + tc35876x_brightness_control(dev, + dev_priv->brightness_adjusted); + } else { + if (dev_priv->dpi_panel_on[0]) + mdfld_dsi_brightness_control(dev, 0, + dev_priv->brightness_adjusted); + } + + if (dev_priv->dpi_panel_on[2]) + mdfld_dsi_brightness_control(dev, 2, + dev_priv->brightness_adjusted); + gma_power_end(dev); + } + + /* cache the brightness for later use */ + dev_priv->brightness = level; + return 0; +} + +static int mdfld_get_brightness(struct backlight_device *bd) +{ + struct drm_device *dev = + (struct drm_device *)bl_get_data(mdfld_backlight_device); + struct drm_psb_private *dev_priv = dev->dev_private; + + DRM_DEBUG_DRIVER("brightness = 0x%x \n", dev_priv->brightness); + + /* return locally cached var instead of HW read (due to DPST etc.) */ + return dev_priv->brightness; +} + +static const struct backlight_ops mdfld_ops = { + .get_brightness = mdfld_get_brightness, + .update_status = mdfld_set_brightness, +}; + +static int device_backlight_init(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = (struct drm_psb_private *) + dev->dev_private; + + dev_priv->blc_adj1 = BLC_ADJUSTMENT_MAX; + dev_priv->blc_adj2 = BLC_ADJUSTMENT_MAX; + + return 0; +} + +static int mdfld_backlight_init(struct drm_device *dev) +{ + struct backlight_properties props; + int ret = 0; + + memset(&props, 0, sizeof(struct backlight_properties)); + props.max_brightness = BRIGHTNESS_MAX_LEVEL; + props.type = BACKLIGHT_PLATFORM; + mdfld_backlight_device = backlight_device_register("mdfld-bl", + NULL, (void *)dev, &mdfld_ops, &props); + + if (IS_ERR(mdfld_backlight_device)) + return PTR_ERR(mdfld_backlight_device); + + ret = device_backlight_init(dev); + if (ret) + return ret; + + mdfld_backlight_device->props.brightness = BRIGHTNESS_MAX_LEVEL; + mdfld_backlight_device->props.max_brightness = BRIGHTNESS_MAX_LEVEL; + backlight_update_status(mdfld_backlight_device); + return 0; +} +#endif + +struct backlight_device *mdfld_get_backlight_device(void) +{ +#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE + return mdfld_backlight_device; +#else + return NULL; +#endif +} + +/* + * mdfld_save_display_registers + * + * Description: We are going to suspend so save current display + * register state. + * + * Notes: FIXME_JLIU7 need to add the support for DPI MIPI & HDMI audio + */ +static int mdfld_save_display_registers(struct drm_device *dev, int pipe) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct medfield_state *regs = &dev_priv->regs.mdfld; + int i; + + /* register */ + u32 dpll_reg = MRST_DPLL_A; + u32 fp_reg = MRST_FPA0; + u32 pipeconf_reg = PIPEACONF; + u32 htot_reg = HTOTAL_A; + u32 hblank_reg = HBLANK_A; + u32 hsync_reg = HSYNC_A; + u32 vtot_reg = VTOTAL_A; + u32 vblank_reg = VBLANK_A; + u32 vsync_reg = VSYNC_A; + u32 pipesrc_reg = PIPEASRC; + u32 dspstride_reg = DSPASTRIDE; + u32 dsplinoff_reg = DSPALINOFF; + u32 dsptileoff_reg = DSPATILEOFF; + u32 dspsize_reg = DSPASIZE; + u32 dsppos_reg = DSPAPOS; + u32 dspsurf_reg = DSPASURF; + u32 mipi_reg = MIPI; + u32 dspcntr_reg = DSPACNTR; + u32 dspstatus_reg = PIPEASTAT; + u32 palette_reg = PALETTE_A; + + /* pointer to values */ + u32 *dpll_val = ®s->saveDPLL_A; + u32 *fp_val = ®s->saveFPA0; + u32 *pipeconf_val = ®s->savePIPEACONF; + u32 *htot_val = ®s->saveHTOTAL_A; + u32 *hblank_val = ®s->saveHBLANK_A; + u32 *hsync_val = ®s->saveHSYNC_A; + u32 *vtot_val = ®s->saveVTOTAL_A; + u32 *vblank_val = ®s->saveVBLANK_A; + u32 *vsync_val = ®s->saveVSYNC_A; + u32 *pipesrc_val = ®s->savePIPEASRC; + u32 *dspstride_val = ®s->saveDSPASTRIDE; + u32 *dsplinoff_val = ®s->saveDSPALINOFF; + u32 *dsptileoff_val = ®s->saveDSPATILEOFF; + u32 *dspsize_val = ®s->saveDSPASIZE; + u32 *dsppos_val = ®s->saveDSPAPOS; + u32 *dspsurf_val = ®s->saveDSPASURF; + u32 *mipi_val = ®s->saveMIPI; + u32 *dspcntr_val = ®s->saveDSPACNTR; + u32 *dspstatus_val = ®s->saveDSPASTATUS; + u32 *palette_val = regs->save_palette_a; + + switch (pipe) { + case 0: + break; + case 1: + /* regester */ + dpll_reg = MDFLD_DPLL_B; + fp_reg = MDFLD_DPLL_DIV0; + pipeconf_reg = PIPEBCONF; + htot_reg = HTOTAL_B; + hblank_reg = HBLANK_B; + hsync_reg = HSYNC_B; + vtot_reg = VTOTAL_B; + vblank_reg = VBLANK_B; + vsync_reg = VSYNC_B; + pipesrc_reg = PIPEBSRC; + dspstride_reg = DSPBSTRIDE; + dsplinoff_reg = DSPBLINOFF; + dsptileoff_reg = DSPBTILEOFF; + dspsize_reg = DSPBSIZE; + dsppos_reg = DSPBPOS; + dspsurf_reg = DSPBSURF; + dspcntr_reg = DSPBCNTR; + dspstatus_reg = PIPEBSTAT; + palette_reg = PALETTE_B; + + /* values */ + dpll_val = ®s->saveDPLL_B; + fp_val = ®s->saveFPB0; + pipeconf_val = ®s->savePIPEBCONF; + htot_val = ®s->saveHTOTAL_B; + hblank_val = ®s->saveHBLANK_B; + hsync_val = ®s->saveHSYNC_B; + vtot_val = ®s->saveVTOTAL_B; + vblank_val = ®s->saveVBLANK_B; + vsync_val = ®s->saveVSYNC_B; + pipesrc_val = ®s->savePIPEBSRC; + dspstride_val = ®s->saveDSPBSTRIDE; + dsplinoff_val = ®s->saveDSPBLINOFF; + dsptileoff_val = ®s->saveDSPBTILEOFF; + dspsize_val = ®s->saveDSPBSIZE; + dsppos_val = ®s->saveDSPBPOS; + dspsurf_val = ®s->saveDSPBSURF; + dspcntr_val = ®s->saveDSPBCNTR; + dspstatus_val = ®s->saveDSPBSTATUS; + palette_val = regs->save_palette_b; + break; + case 2: + /* register */ + pipeconf_reg = PIPECCONF; + htot_reg = HTOTAL_C; + hblank_reg = HBLANK_C; + hsync_reg = HSYNC_C; + vtot_reg = VTOTAL_C; + vblank_reg = VBLANK_C; + vsync_reg = VSYNC_C; + pipesrc_reg = PIPECSRC; + dspstride_reg = DSPCSTRIDE; + dsplinoff_reg = DSPCLINOFF; + dsptileoff_reg = DSPCTILEOFF; + dspsize_reg = DSPCSIZE; + dsppos_reg = DSPCPOS; + dspsurf_reg = DSPCSURF; + mipi_reg = MIPI_C; + dspcntr_reg = DSPCCNTR; + dspstatus_reg = PIPECSTAT; + palette_reg = PALETTE_C; + + /* pointer to values */ + pipeconf_val = ®s->savePIPECCONF; + htot_val = ®s->saveHTOTAL_C; + hblank_val = ®s->saveHBLANK_C; + hsync_val = ®s->saveHSYNC_C; + vtot_val = ®s->saveVTOTAL_C; + vblank_val = ®s->saveVBLANK_C; + vsync_val = ®s->saveVSYNC_C; + pipesrc_val = ®s->savePIPECSRC; + dspstride_val = ®s->saveDSPCSTRIDE; + dsplinoff_val = ®s->saveDSPCLINOFF; + dsptileoff_val = ®s->saveDSPCTILEOFF; + dspsize_val = ®s->saveDSPCSIZE; + dsppos_val = ®s->saveDSPCPOS; + dspsurf_val = ®s->saveDSPCSURF; + mipi_val = ®s->saveMIPI_C; + dspcntr_val = ®s->saveDSPCCNTR; + dspstatus_val = ®s->saveDSPCSTATUS; + palette_val = regs->save_palette_c; + break; + default: + DRM_ERROR("%s, invalid pipe number.\n", __func__); + return -EINVAL; + } + + /* Pipe & plane A info */ + *dpll_val = PSB_RVDC32(dpll_reg); + *fp_val = PSB_RVDC32(fp_reg); + *pipeconf_val = PSB_RVDC32(pipeconf_reg); + *htot_val = PSB_RVDC32(htot_reg); + *hblank_val = PSB_RVDC32(hblank_reg); + *hsync_val = PSB_RVDC32(hsync_reg); + *vtot_val = PSB_RVDC32(vtot_reg); + *vblank_val = PSB_RVDC32(vblank_reg); + *vsync_val = PSB_RVDC32(vsync_reg); + *pipesrc_val = PSB_RVDC32(pipesrc_reg); + *dspstride_val = PSB_RVDC32(dspstride_reg); + *dsplinoff_val = PSB_RVDC32(dsplinoff_reg); + *dsptileoff_val = PSB_RVDC32(dsptileoff_reg); + *dspsize_val = PSB_RVDC32(dspsize_reg); + *dsppos_val = PSB_RVDC32(dsppos_reg); + *dspsurf_val = PSB_RVDC32(dspsurf_reg); + *dspcntr_val = PSB_RVDC32(dspcntr_reg); + *dspstatus_val = PSB_RVDC32(dspstatus_reg); + + /*save palette (gamma) */ + for (i = 0; i < 256; i++) + palette_val[i] = PSB_RVDC32(palette_reg + (i << 2)); + + if (pipe == 1) { + regs->savePFIT_CONTROL = PSB_RVDC32(PFIT_CONTROL); + regs->savePFIT_PGM_RATIOS = PSB_RVDC32(PFIT_PGM_RATIOS); + + regs->saveHDMIPHYMISCCTL = PSB_RVDC32(HDMIPHYMISCCTL); + regs->saveHDMIB_CONTROL = PSB_RVDC32(HDMIB_CONTROL); + return 0; + } + + *mipi_val = PSB_RVDC32(mipi_reg); + return 0; +} + +/* + * mdfld_restore_display_registers + * + * Description: We are going to resume so restore display register state. + * + * Notes: FIXME_JLIU7 need to add the support for DPI MIPI & HDMI audio + */ +static int mdfld_restore_display_registers(struct drm_device *dev, int pipe) +{ + /* To get panel out of ULPS mode. */ + u32 temp = 0; + u32 device_ready_reg = DEVICE_READY_REG; + struct drm_psb_private *dev_priv = dev->dev_private; + struct mdfld_dsi_config *dsi_config = NULL; + struct medfield_state *regs = &dev_priv->regs.mdfld; + u32 i = 0; + u32 dpll = 0; + u32 timeout = 0; + + /* regester */ + u32 dpll_reg = MRST_DPLL_A; + u32 fp_reg = MRST_FPA0; + u32 pipeconf_reg = PIPEACONF; + u32 htot_reg = HTOTAL_A; + u32 hblank_reg = HBLANK_A; + u32 hsync_reg = HSYNC_A; + u32 vtot_reg = VTOTAL_A; + u32 vblank_reg = VBLANK_A; + u32 vsync_reg = VSYNC_A; + u32 pipesrc_reg = PIPEASRC; + u32 dspstride_reg = DSPASTRIDE; + u32 dsplinoff_reg = DSPALINOFF; + u32 dsptileoff_reg = DSPATILEOFF; + u32 dspsize_reg = DSPASIZE; + u32 dsppos_reg = DSPAPOS; + u32 dspsurf_reg = DSPASURF; + u32 dspstatus_reg = PIPEASTAT; + u32 mipi_reg = MIPI; + u32 dspcntr_reg = DSPACNTR; + u32 palette_reg = PALETTE_A; + + /* values */ + u32 dpll_val = regs->saveDPLL_A & ~DPLL_VCO_ENABLE; + u32 fp_val = regs->saveFPA0; + u32 pipeconf_val = regs->savePIPEACONF; + u32 htot_val = regs->saveHTOTAL_A; + u32 hblank_val = regs->saveHBLANK_A; + u32 hsync_val = regs->saveHSYNC_A; + u32 vtot_val = regs->saveVTOTAL_A; + u32 vblank_val = regs->saveVBLANK_A; + u32 vsync_val = regs->saveVSYNC_A; + u32 pipesrc_val = regs->savePIPEASRC; + u32 dspstride_val = regs->saveDSPASTRIDE; + u32 dsplinoff_val = regs->saveDSPALINOFF; + u32 dsptileoff_val = regs->saveDSPATILEOFF; + u32 dspsize_val = regs->saveDSPASIZE; + u32 dsppos_val = regs->saveDSPAPOS; + u32 dspsurf_val = regs->saveDSPASURF; + u32 dspstatus_val = regs->saveDSPASTATUS; + u32 mipi_val = regs->saveMIPI; + u32 dspcntr_val = regs->saveDSPACNTR; + u32 *palette_val = regs->save_palette_a; + + switch (pipe) { + case 0: + dsi_config = dev_priv->dsi_configs[0]; + break; + case 1: + /* regester */ + dpll_reg = MDFLD_DPLL_B; + fp_reg = MDFLD_DPLL_DIV0; + pipeconf_reg = PIPEBCONF; + htot_reg = HTOTAL_B; + hblank_reg = HBLANK_B; + hsync_reg = HSYNC_B; + vtot_reg = VTOTAL_B; + vblank_reg = VBLANK_B; + vsync_reg = VSYNC_B; + pipesrc_reg = PIPEBSRC; + dspstride_reg = DSPBSTRIDE; + dsplinoff_reg = DSPBLINOFF; + dsptileoff_reg = DSPBTILEOFF; + dspsize_reg = DSPBSIZE; + dsppos_reg = DSPBPOS; + dspsurf_reg = DSPBSURF; + dspcntr_reg = DSPBCNTR; + dspstatus_reg = PIPEBSTAT; + palette_reg = PALETTE_B; + + /* values */ + dpll_val = regs->saveDPLL_B & ~DPLL_VCO_ENABLE; + fp_val = regs->saveFPB0; + pipeconf_val = regs->savePIPEBCONF; + htot_val = regs->saveHTOTAL_B; + hblank_val = regs->saveHBLANK_B; + hsync_val = regs->saveHSYNC_B; + vtot_val = regs->saveVTOTAL_B; + vblank_val = regs->saveVBLANK_B; + vsync_val = regs->saveVSYNC_B; + pipesrc_val = regs->savePIPEBSRC; + dspstride_val = regs->saveDSPBSTRIDE; + dsplinoff_val = regs->saveDSPBLINOFF; + dsptileoff_val = regs->saveDSPBTILEOFF; + dspsize_val = regs->saveDSPBSIZE; + dsppos_val = regs->saveDSPBPOS; + dspsurf_val = regs->saveDSPBSURF; + dspcntr_val = regs->saveDSPBCNTR; + dspstatus_val = regs->saveDSPBSTATUS; + palette_val = regs->save_palette_b; + break; + case 2: + /* regester */ + pipeconf_reg = PIPECCONF; + htot_reg = HTOTAL_C; + hblank_reg = HBLANK_C; + hsync_reg = HSYNC_C; + vtot_reg = VTOTAL_C; + vblank_reg = VBLANK_C; + vsync_reg = VSYNC_C; + pipesrc_reg = PIPECSRC; + dspstride_reg = DSPCSTRIDE; + dsplinoff_reg = DSPCLINOFF; + dsptileoff_reg = DSPCTILEOFF; + dspsize_reg = DSPCSIZE; + dsppos_reg = DSPCPOS; + dspsurf_reg = DSPCSURF; + mipi_reg = MIPI_C; + dspcntr_reg = DSPCCNTR; + dspstatus_reg = PIPECSTAT; + palette_reg = PALETTE_C; + + /* values */ + pipeconf_val = regs->savePIPECCONF; + htot_val = regs->saveHTOTAL_C; + hblank_val = regs->saveHBLANK_C; + hsync_val = regs->saveHSYNC_C; + vtot_val = regs->saveVTOTAL_C; + vblank_val = regs->saveVBLANK_C; + vsync_val = regs->saveVSYNC_C; + pipesrc_val = regs->savePIPECSRC; + dspstride_val = regs->saveDSPCSTRIDE; + dsplinoff_val = regs->saveDSPCLINOFF; + dsptileoff_val = regs->saveDSPCTILEOFF; + dspsize_val = regs->saveDSPCSIZE; + dsppos_val = regs->saveDSPCPOS; + dspsurf_val = regs->saveDSPCSURF; + mipi_val = regs->saveMIPI_C; + dspcntr_val = regs->saveDSPCCNTR; + dspstatus_val = regs->saveDSPCSTATUS; + palette_val = regs->save_palette_c; + + dsi_config = dev_priv->dsi_configs[1]; + break; + default: + DRM_ERROR("%s, invalid pipe number.\n", __func__); + return -EINVAL; + } + + /*make sure VGA plane is off. it initializes to on after reset!*/ + PSB_WVDC32(0x80000000, VGACNTRL); + + if (pipe == 1) { + PSB_WVDC32(dpll_val & ~DPLL_VCO_ENABLE, dpll_reg); + PSB_RVDC32(dpll_reg); + + PSB_WVDC32(fp_val, fp_reg); + } else { + + dpll = PSB_RVDC32(dpll_reg); + + if (!(dpll & DPLL_VCO_ENABLE)) { + + /* When ungating power of DPLL, needs to wait 0.5us + before enable the VCO */ + if (dpll & MDFLD_PWR_GATE_EN) { + dpll &= ~MDFLD_PWR_GATE_EN; + PSB_WVDC32(dpll, dpll_reg); + /* FIXME_MDFLD PO - change 500 to 1 after PO */ + udelay(500); + } + + PSB_WVDC32(fp_val, fp_reg); + PSB_WVDC32(dpll_val, dpll_reg); + /* FIXME_MDFLD PO - change 500 to 1 after PO */ + udelay(500); + + dpll_val |= DPLL_VCO_ENABLE; + PSB_WVDC32(dpll_val, dpll_reg); + PSB_RVDC32(dpll_reg); + + /* wait for DSI PLL to lock */ + while (timeout < 20000 && + !(PSB_RVDC32(pipeconf_reg) & PIPECONF_DSIPLL_LOCK)) { + udelay(150); + timeout++; + } + + if (timeout == 20000) { + DRM_ERROR("%s, can't lock DSIPLL.\n", + __func__); + return -EINVAL; + } + } + } + /* Restore mode */ + PSB_WVDC32(htot_val, htot_reg); + PSB_WVDC32(hblank_val, hblank_reg); + PSB_WVDC32(hsync_val, hsync_reg); + PSB_WVDC32(vtot_val, vtot_reg); + PSB_WVDC32(vblank_val, vblank_reg); + PSB_WVDC32(vsync_val, vsync_reg); + PSB_WVDC32(pipesrc_val, pipesrc_reg); + PSB_WVDC32(dspstatus_val, dspstatus_reg); + + /*set up the plane*/ + PSB_WVDC32(dspstride_val, dspstride_reg); + PSB_WVDC32(dsplinoff_val, dsplinoff_reg); + PSB_WVDC32(dsptileoff_val, dsptileoff_reg); + PSB_WVDC32(dspsize_val, dspsize_reg); + PSB_WVDC32(dsppos_val, dsppos_reg); + PSB_WVDC32(dspsurf_val, dspsurf_reg); + + if (pipe == 1) { + /* restore palette (gamma) */ + /*DRM_UDELAY(50000); */ + for (i = 0; i < 256; i++) + PSB_WVDC32(palette_val[i], palette_reg + (i << 2)); + + PSB_WVDC32(regs->savePFIT_CONTROL, PFIT_CONTROL); + PSB_WVDC32(regs->savePFIT_PGM_RATIOS, PFIT_PGM_RATIOS); + + /*TODO: resume HDMI port */ + + /*TODO: resume pipe*/ + + /*enable the plane*/ + PSB_WVDC32(dspcntr_val & ~DISPLAY_PLANE_ENABLE, dspcntr_reg); + + return 0; + } + + /*set up pipe related registers*/ + PSB_WVDC32(mipi_val, mipi_reg); + + /*setup MIPI adapter + MIPI IP registers*/ + if (dsi_config) + mdfld_dsi_controller_init(dsi_config, pipe); + + if (in_atomic() || in_interrupt()) + mdelay(20); + else + msleep(20); + + /*enable the plane*/ + PSB_WVDC32(dspcntr_val, dspcntr_reg); + + if (in_atomic() || in_interrupt()) + mdelay(20); + else + msleep(20); + + /* LP Hold Release */ + temp = REG_READ(mipi_reg); + temp |= LP_OUTPUT_HOLD_RELEASE; + REG_WRITE(mipi_reg, temp); + mdelay(1); + + + /* Set DSI host to exit from Utra Low Power State */ + temp = REG_READ(device_ready_reg); + temp &= ~ULPS_MASK; + temp |= 0x3; + temp |= EXIT_ULPS_DEV_READY; + REG_WRITE(device_ready_reg, temp); + mdelay(1); + + temp = REG_READ(device_ready_reg); + temp &= ~ULPS_MASK; + temp |= EXITING_ULPS; + REG_WRITE(device_ready_reg, temp); + mdelay(1); + + /*enable the pipe*/ + PSB_WVDC32(pipeconf_val, pipeconf_reg); + + /* restore palette (gamma) */ + /*DRM_UDELAY(50000); */ + for (i = 0; i < 256; i++) + PSB_WVDC32(palette_val[i], palette_reg + (i << 2)); + + return 0; +} + +static int mdfld_save_registers(struct drm_device *dev) +{ + /* mdfld_save_cursor_overlay_registers(dev); */ + mdfld_save_display_registers(dev, 0); + mdfld_save_display_registers(dev, 2); + mdfld_disable_crtc(dev, 0); + mdfld_disable_crtc(dev, 2); + + return 0; +} + +static int mdfld_restore_registers(struct drm_device *dev) +{ + mdfld_restore_display_registers(dev, 2); + mdfld_restore_display_registers(dev, 0); + /* mdfld_restore_cursor_overlay_registers(dev); */ + + return 0; +} + +static int mdfld_power_down(struct drm_device *dev) +{ + /* FIXME */ + return 0; +} + +static int mdfld_power_up(struct drm_device *dev) +{ + /* FIXME */ + return 0; +} + +const struct psb_ops mdfld_chip_ops = { + .name = "mdfld", + .accel_2d = 0, + .pipes = 3, + .crtcs = 3, + .sgx_offset = MRST_SGX_OFFSET, + + .chip_setup = mid_chip_setup, + .crtc_helper = &mdfld_helper_funcs, + .crtc_funcs = &psb_intel_crtc_funcs, + + .output_init = mdfld_output_init, + +#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE + .backlight_init = mdfld_backlight_init, +#endif + + .save_regs = mdfld_save_registers, + .restore_regs = mdfld_restore_registers, + .power_down = mdfld_power_down, + .power_up = mdfld_power_up, +}; diff --git a/drivers/gpu/drm/gma500/mdfld_dsi_dpi.c b/drivers/gpu/drm/gma500/mdfld_dsi_dpi.c new file mode 100644 index 0000000..d52358b --- /dev/null +++ b/drivers/gpu/drm/gma500/mdfld_dsi_dpi.c @@ -0,0 +1,1017 @@ +/* + * Copyright © 2010 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * jim liu + * Jackie Li + */ + +#include "mdfld_dsi_dpi.h" +#include "mdfld_output.h" +#include "mdfld_dsi_pkg_sender.h" +#include "psb_drv.h" +#include "tc35876x-dsi-lvds.h" + +static void mdfld_dsi_dpi_shut_down(struct mdfld_dsi_dpi_output *output, + int pipe); + +static void mdfld_wait_for_HS_DATA_FIFO(struct drm_device *dev, u32 pipe) +{ + u32 gen_fifo_stat_reg = MIPI_GEN_FIFO_STAT_REG(pipe); + int timeout = 0; + + udelay(500); + + /* This will time out after approximately 2+ seconds */ + while ((timeout < 20000) && + (REG_READ(gen_fifo_stat_reg) & DSI_FIFO_GEN_HS_DATA_FULL)) { + udelay(100); + timeout++; + } + + if (timeout == 20000) + DRM_INFO("MIPI: HS Data FIFO was never cleared!\n"); +} + +static void mdfld_wait_for_HS_CTRL_FIFO(struct drm_device *dev, u32 pipe) +{ + u32 gen_fifo_stat_reg = MIPI_GEN_FIFO_STAT_REG(pipe); + int timeout = 0; + + udelay(500); + + /* This will time out after approximately 2+ seconds */ + while ((timeout < 20000) && (REG_READ(gen_fifo_stat_reg) + & DSI_FIFO_GEN_HS_CTRL_FULL)) { + udelay(100); + timeout++; + } + if (timeout == 20000) + DRM_INFO("MIPI: HS CMD FIFO was never cleared!\n"); +} + +static void mdfld_wait_for_DPI_CTRL_FIFO(struct drm_device *dev, u32 pipe) +{ + u32 gen_fifo_stat_reg = MIPI_GEN_FIFO_STAT_REG(pipe); + int timeout = 0; + + udelay(500); + + /* This will time out after approximately 2+ seconds */ + while ((timeout < 20000) && ((REG_READ(gen_fifo_stat_reg) & + DPI_FIFO_EMPTY) != DPI_FIFO_EMPTY)) { + udelay(100); + timeout++; + } + + if (timeout == 20000) + DRM_ERROR("MIPI: DPI FIFO was never cleared\n"); +} + +static void mdfld_wait_for_SPL_PKG_SENT(struct drm_device *dev, u32 pipe) +{ + u32 intr_stat_reg = MIPI_INTR_STAT_REG(pipe); + int timeout = 0; + + udelay(500); + + /* This will time out after approximately 2+ seconds */ + while ((timeout < 20000) && (!(REG_READ(intr_stat_reg) + & DSI_INTR_STATE_SPL_PKG_SENT))) { + udelay(100); + timeout++; + } + + if (timeout == 20000) + DRM_ERROR("MIPI: SPL_PKT_SENT_INTERRUPT was not sent successfully!\n"); +} + +/* For TC35876X */ + +static void dsi_set_device_ready_state(struct drm_device *dev, int state, + int pipe) +{ + REG_FLD_MOD(MIPI_DEVICE_READY_REG(pipe), !!state, 0, 0); +} + +static void dsi_set_pipe_plane_enable_state(struct drm_device *dev, + int state, int pipe) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + u32 pipeconf_reg = PIPEACONF; + u32 dspcntr_reg = DSPACNTR; + + u32 dspcntr = dev_priv->dspcntr[pipe]; + u32 mipi = MIPI_PORT_EN | PASS_FROM_SPHY_TO_AFE | SEL_FLOPPED_HSTX; + + if (pipe) { + pipeconf_reg = PIPECCONF; + dspcntr_reg = DSPCCNTR; + } else + mipi &= (~0x03); + + if (state) { + /*Set up pipe */ + REG_WRITE(pipeconf_reg, BIT(31)); + + if (REG_BIT_WAIT(pipeconf_reg, 1, 30)) + dev_err(&dev->pdev->dev, "%s: Pipe enable timeout\n", + __func__); + + /*Set up display plane */ + REG_WRITE(dspcntr_reg, dspcntr); + } else { + u32 dspbase_reg = pipe ? MDFLD_DSPCBASE : MRST_DSPABASE; + + /* Put DSI lanes to ULPS to disable pipe */ + REG_FLD_MOD(MIPI_DEVICE_READY_REG(pipe), 2, 2, 1); + REG_READ(MIPI_DEVICE_READY_REG(pipe)); /* posted write? */ + + /* LP Hold */ + REG_FLD_MOD(MIPI_PORT_CONTROL(pipe), 0, 16, 16); + REG_READ(MIPI_PORT_CONTROL(pipe)); /* posted write? */ + + /* Disable display plane */ + REG_FLD_MOD(dspcntr_reg, 0, 31, 31); + + /* Flush the plane changes ??? posted write? */ + REG_WRITE(dspbase_reg, REG_READ(dspbase_reg)); + REG_READ(dspbase_reg); + + /* Disable PIPE */ + REG_FLD_MOD(pipeconf_reg, 0, 31, 31); + + if (REG_BIT_WAIT(pipeconf_reg, 0, 30)) + dev_err(&dev->pdev->dev, "%s: Pipe disable timeout\n", + __func__); + + if (REG_BIT_WAIT(MIPI_GEN_FIFO_STAT_REG(pipe), 1, 28)) + dev_err(&dev->pdev->dev, "%s: FIFO not empty\n", + __func__); + } +} + +static void mdfld_dsi_configure_down(struct mdfld_dsi_encoder *dsi_encoder, + int pipe) +{ + struct mdfld_dsi_dpi_output *dpi_output = + MDFLD_DSI_DPI_OUTPUT(dsi_encoder); + struct mdfld_dsi_config *dsi_config = + mdfld_dsi_encoder_get_config(dsi_encoder); + struct drm_device *dev = dsi_config->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + + if (!dev_priv->dpi_panel_on[pipe]) { + dev_err(dev->dev, "DPI panel is already off\n"); + return; + } + tc35876x_toshiba_bridge_panel_off(dev); + tc35876x_set_bridge_reset_state(dev, 1); + dsi_set_pipe_plane_enable_state(dev, 0, pipe); + mdfld_dsi_dpi_shut_down(dpi_output, pipe); + dsi_set_device_ready_state(dev, 0, pipe); +} + +static void mdfld_dsi_configure_up(struct mdfld_dsi_encoder *dsi_encoder, + int pipe) +{ + struct mdfld_dsi_dpi_output *dpi_output = + MDFLD_DSI_DPI_OUTPUT(dsi_encoder); + struct mdfld_dsi_config *dsi_config = + mdfld_dsi_encoder_get_config(dsi_encoder); + struct drm_device *dev = dsi_config->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + + if (dev_priv->dpi_panel_on[pipe]) { + dev_err(dev->dev, "DPI panel is already on\n"); + return; + } + + /* For resume path sequence */ + mdfld_dsi_dpi_shut_down(dpi_output, pipe); + dsi_set_device_ready_state(dev, 0, pipe); + + dsi_set_device_ready_state(dev, 1, pipe); + tc35876x_set_bridge_reset_state(dev, 0); + tc35876x_configure_lvds_bridge(dev); + mdfld_dsi_dpi_turn_on(dpi_output, pipe); /* Send turn on command */ + dsi_set_pipe_plane_enable_state(dev, 1, pipe); +} +/* End for TC35876X */ + +/* ************************************************************************* *\ + * FUNCTION: mdfld_dsi_tpo_ic_init + * + * DESCRIPTION: This function is called only by mrst_dsi_mode_set and + * restore_display_registers. since this function does not + * acquire the mutex, it is important that the calling function + * does! +\* ************************************************************************* */ +static void mdfld_dsi_tpo_ic_init(struct mdfld_dsi_config *dsi_config, u32 pipe) +{ + struct drm_device *dev = dsi_config->dev; + u32 dcsChannelNumber = dsi_config->channel_num; + u32 gen_data_reg = MIPI_HS_GEN_DATA_REG(pipe); + u32 gen_ctrl_reg = MIPI_HS_GEN_CTRL_REG(pipe); + u32 gen_ctrl_val = GEN_LONG_WRITE; + + DRM_INFO("Enter mrst init TPO MIPI display.\n"); + + gen_ctrl_val |= dcsChannelNumber << DCS_CHANNEL_NUMBER_POS; + + /* Flip page order */ + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x00008036); + mdfld_wait_for_HS_CTRL_FIFO(dev, pipe); + REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x02 << WORD_COUNTS_POS)); + + /* 0xF0 */ + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x005a5af0); + mdfld_wait_for_HS_CTRL_FIFO(dev, pipe); + REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x03 << WORD_COUNTS_POS)); + + /* Write protection key */ + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x005a5af1); + mdfld_wait_for_HS_CTRL_FIFO(dev, pipe); + REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x03 << WORD_COUNTS_POS)); + + /* 0xFC */ + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x005a5afc); + mdfld_wait_for_HS_CTRL_FIFO(dev, pipe); + REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x03 << WORD_COUNTS_POS)); + + /* 0xB7 */ + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x770000b7); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x00000044); + mdfld_wait_for_HS_CTRL_FIFO(dev, pipe); + REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x05 << WORD_COUNTS_POS)); + + /* 0xB6 */ + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x000a0ab6); + mdfld_wait_for_HS_CTRL_FIFO(dev, pipe); + REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x03 << WORD_COUNTS_POS)); + + /* 0xF2 */ + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x081010f2); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x4a070708); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x000000c5); + mdfld_wait_for_HS_CTRL_FIFO(dev, pipe); + REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x09 << WORD_COUNTS_POS)); + + /* 0xF8 */ + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x024003f8); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x01030a04); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x0e020220); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x00000004); + mdfld_wait_for_HS_CTRL_FIFO(dev, pipe); + REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x0d << WORD_COUNTS_POS)); + + /* 0xE2 */ + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x398fc3e2); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x0000916f); + mdfld_wait_for_HS_CTRL_FIFO(dev, pipe); + REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x06 << WORD_COUNTS_POS)); + + /* 0xB0 */ + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x000000b0); + mdfld_wait_for_HS_CTRL_FIFO(dev, pipe); + REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x02 << WORD_COUNTS_POS)); + + /* 0xF4 */ + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x240242f4); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x78ee2002); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x2a071050); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x507fee10); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x10300710); + mdfld_wait_for_HS_CTRL_FIFO(dev, pipe); + REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x14 << WORD_COUNTS_POS)); + + /* 0xBA */ + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x19fe07ba); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x101c0a31); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x00000010); + mdfld_wait_for_HS_CTRL_FIFO(dev, pipe); + REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x09 << WORD_COUNTS_POS)); + + /* 0xBB */ + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x28ff07bb); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x24280a31); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x00000034); + mdfld_wait_for_HS_CTRL_FIFO(dev, pipe); + REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x09 << WORD_COUNTS_POS)); + + /* 0xFB */ + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x535d05fb); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x1b1a2130); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x221e180e); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x131d2120); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x535d0508); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x1c1a2131); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x231f160d); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x111b2220); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x535c2008); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x1f1d2433); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x2c251a10); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x2c34372d); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x00000023); + mdfld_wait_for_HS_CTRL_FIFO(dev, pipe); + REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x31 << WORD_COUNTS_POS)); + + /* 0xFA */ + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x525c0bfa); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x1c1c232f); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x2623190e); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x18212625); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x545d0d0e); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x1e1d2333); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x26231a10); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x1a222725); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x545d280f); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x21202635); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x31292013); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x31393d33); + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x00000029); + mdfld_wait_for_HS_CTRL_FIFO(dev, pipe); + REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x31 << WORD_COUNTS_POS)); + + /* Set DM */ + mdfld_wait_for_HS_DATA_FIFO(dev, pipe); + REG_WRITE(gen_data_reg, 0x000100f7); + mdfld_wait_for_HS_CTRL_FIFO(dev, pipe); + REG_WRITE(gen_ctrl_reg, gen_ctrl_val | (0x03 << WORD_COUNTS_POS)); +} + +static u16 mdfld_dsi_dpi_to_byte_clock_count(int pixel_clock_count, + int num_lane, int bpp) +{ + return (u16)((pixel_clock_count * bpp) / (num_lane * 8)); +} + +/* + * Calculate the dpi time basing on a given drm mode @mode + * return 0 on success. + * FIXME: I was using proposed mode value for calculation, may need to + * use crtc mode values later + */ +int mdfld_dsi_dpi_timing_calculation(struct drm_display_mode *mode, + struct mdfld_dsi_dpi_timing *dpi_timing, + int num_lane, int bpp) +{ + int pclk_hsync, pclk_hfp, pclk_hbp, pclk_hactive; + int pclk_vsync, pclk_vfp, pclk_vbp; + + pclk_hactive = mode->hdisplay; + pclk_hfp = mode->hsync_start - mode->hdisplay; + pclk_hsync = mode->hsync_end - mode->hsync_start; + pclk_hbp = mode->htotal - mode->hsync_end; + + pclk_vfp = mode->vsync_start - mode->vdisplay; + pclk_vsync = mode->vsync_end - mode->vsync_start; + pclk_vbp = mode->vtotal - mode->vsync_end; + + /* + * byte clock counts were calculated by following formula + * bclock_count = pclk_count * bpp / num_lane / 8 + */ + dpi_timing->hsync_count = mdfld_dsi_dpi_to_byte_clock_count( + pclk_hsync, num_lane, bpp); + dpi_timing->hbp_count = mdfld_dsi_dpi_to_byte_clock_count( + pclk_hbp, num_lane, bpp); + dpi_timing->hfp_count = mdfld_dsi_dpi_to_byte_clock_count( + pclk_hfp, num_lane, bpp); + dpi_timing->hactive_count = mdfld_dsi_dpi_to_byte_clock_count( + pclk_hactive, num_lane, bpp); + dpi_timing->vsync_count = mdfld_dsi_dpi_to_byte_clock_count( + pclk_vsync, num_lane, bpp); + dpi_timing->vbp_count = mdfld_dsi_dpi_to_byte_clock_count( + pclk_vbp, num_lane, bpp); + dpi_timing->vfp_count = mdfld_dsi_dpi_to_byte_clock_count( + pclk_vfp, num_lane, bpp); + + return 0; +} + +void mdfld_dsi_dpi_controller_init(struct mdfld_dsi_config *dsi_config, + int pipe) +{ + struct drm_device *dev = dsi_config->dev; + int lane_count = dsi_config->lane_count; + struct mdfld_dsi_dpi_timing dpi_timing; + struct drm_display_mode *mode = dsi_config->mode; + u32 val; + + /*un-ready device*/ + REG_FLD_MOD(MIPI_DEVICE_READY_REG(pipe), 0, 0, 0); + + /*init dsi adapter before kicking off*/ + REG_WRITE(MIPI_CTRL_REG(pipe), 0x00000018); + + /*enable all interrupts*/ + REG_WRITE(MIPI_INTR_EN_REG(pipe), 0xffffffff); + + /*set up func_prg*/ + val = lane_count; + val |= dsi_config->channel_num << DSI_DPI_VIRT_CHANNEL_OFFSET; + + switch (dsi_config->bpp) { + case 16: + val |= DSI_DPI_COLOR_FORMAT_RGB565; + break; + case 18: + val |= DSI_DPI_COLOR_FORMAT_RGB666; + break; + case 24: + val |= DSI_DPI_COLOR_FORMAT_RGB888; + break; + default: + DRM_ERROR("unsupported color format, bpp = %d\n", + dsi_config->bpp); + } + REG_WRITE(MIPI_DSI_FUNC_PRG_REG(pipe), val); + + REG_WRITE(MIPI_HS_TX_TIMEOUT_REG(pipe), + (mode->vtotal * mode->htotal * dsi_config->bpp / + (8 * lane_count)) & DSI_HS_TX_TIMEOUT_MASK); + REG_WRITE(MIPI_LP_RX_TIMEOUT_REG(pipe), + 0xffff & DSI_LP_RX_TIMEOUT_MASK); + + /*max value: 20 clock cycles of txclkesc*/ + REG_WRITE(MIPI_TURN_AROUND_TIMEOUT_REG(pipe), + 0x14 & DSI_TURN_AROUND_TIMEOUT_MASK); + + /*min 21 txclkesc, max: ffffh*/ + REG_WRITE(MIPI_DEVICE_RESET_TIMER_REG(pipe), + 0xffff & DSI_RESET_TIMER_MASK); + + REG_WRITE(MIPI_DPI_RESOLUTION_REG(pipe), + mode->vdisplay << 16 | mode->hdisplay); + + /*set DPI timing registers*/ + mdfld_dsi_dpi_timing_calculation(mode, &dpi_timing, + dsi_config->lane_count, dsi_config->bpp); + + REG_WRITE(MIPI_HSYNC_COUNT_REG(pipe), + dpi_timing.hsync_count & DSI_DPI_TIMING_MASK); + REG_WRITE(MIPI_HBP_COUNT_REG(pipe), + dpi_timing.hbp_count & DSI_DPI_TIMING_MASK); + REG_WRITE(MIPI_HFP_COUNT_REG(pipe), + dpi_timing.hfp_count & DSI_DPI_TIMING_MASK); + REG_WRITE(MIPI_HACTIVE_COUNT_REG(pipe), + dpi_timing.hactive_count & DSI_DPI_TIMING_MASK); + REG_WRITE(MIPI_VSYNC_COUNT_REG(pipe), + dpi_timing.vsync_count & DSI_DPI_TIMING_MASK); + REG_WRITE(MIPI_VBP_COUNT_REG(pipe), + dpi_timing.vbp_count & DSI_DPI_TIMING_MASK); + REG_WRITE(MIPI_VFP_COUNT_REG(pipe), + dpi_timing.vfp_count & DSI_DPI_TIMING_MASK); + + REG_WRITE(MIPI_HIGH_LOW_SWITCH_COUNT_REG(pipe), 0x46); + + /*min: 7d0 max: 4e20*/ + REG_WRITE(MIPI_INIT_COUNT_REG(pipe), 0x000007d0); + + /*set up video mode*/ + val = dsi_config->video_mode | DSI_DPI_COMPLETE_LAST_LINE; + REG_WRITE(MIPI_VIDEO_MODE_FORMAT_REG(pipe), val); + + REG_WRITE(MIPI_EOT_DISABLE_REG(pipe), 0x00000000); + + REG_WRITE(MIPI_LP_BYTECLK_REG(pipe), 0x00000004); + + /*TODO: figure out how to setup these registers*/ + if (mdfld_get_panel_type(dev, pipe) == TC35876X) + REG_WRITE(MIPI_DPHY_PARAM_REG(pipe), 0x2A0c6008); + else + REG_WRITE(MIPI_DPHY_PARAM_REG(pipe), 0x150c3408); + + REG_WRITE(MIPI_CLK_LANE_SWITCH_TIME_CNT_REG(pipe), (0xa << 16) | 0x14); + + if (mdfld_get_panel_type(dev, pipe) == TC35876X) + tc35876x_set_bridge_reset_state(dev, 0); /*Pull High Reset */ + + /*set device ready*/ + REG_FLD_MOD(MIPI_DEVICE_READY_REG(pipe), 1, 0, 0); +} + +void mdfld_dsi_dpi_turn_on(struct mdfld_dsi_dpi_output *output, int pipe) +{ + struct drm_device *dev = output->dev; + + /* clear special packet sent bit */ + if (REG_READ(MIPI_INTR_STAT_REG(pipe)) & DSI_INTR_STATE_SPL_PKG_SENT) + REG_WRITE(MIPI_INTR_STAT_REG(pipe), + DSI_INTR_STATE_SPL_PKG_SENT); + + /*send turn on package*/ + REG_WRITE(MIPI_DPI_CONTROL_REG(pipe), DSI_DPI_CTRL_HS_TURN_ON); + + /*wait for SPL_PKG_SENT interrupt*/ + mdfld_wait_for_SPL_PKG_SENT(dev, pipe); + + if (REG_READ(MIPI_INTR_STAT_REG(pipe)) & DSI_INTR_STATE_SPL_PKG_SENT) + REG_WRITE(MIPI_INTR_STAT_REG(pipe), + DSI_INTR_STATE_SPL_PKG_SENT); + + output->panel_on = 1; + + /* FIXME the following is disabled to WA the X slow start issue + for TMD panel + if (pipe == 2) + dev_priv->dpi_panel_on2 = true; + else if (pipe == 0) + dev_priv->dpi_panel_on = true; */ +} + +static void mdfld_dsi_dpi_shut_down(struct mdfld_dsi_dpi_output *output, + int pipe) +{ + struct drm_device *dev = output->dev; + + /*if output is on, or mode setting didn't happen, ignore this*/ + if ((!output->panel_on) || output->first_boot) { + output->first_boot = 0; + return; + } + + /* Wait for dpi fifo to empty */ + mdfld_wait_for_DPI_CTRL_FIFO(dev, pipe); + + /* Clear the special packet interrupt bit if set */ + if (REG_READ(MIPI_INTR_STAT_REG(pipe)) & DSI_INTR_STATE_SPL_PKG_SENT) + REG_WRITE(MIPI_INTR_STAT_REG(pipe), + DSI_INTR_STATE_SPL_PKG_SENT); + + if (REG_READ(MIPI_DPI_CONTROL_REG(pipe)) == DSI_DPI_CTRL_HS_SHUTDOWN) + goto shutdown_out; + + REG_WRITE(MIPI_DPI_CONTROL_REG(pipe), DSI_DPI_CTRL_HS_SHUTDOWN); + +shutdown_out: + output->panel_on = 0; + output->first_boot = 0; + + /* FIXME the following is disabled to WA the X slow start issue + for TMD panel + if (pipe == 2) + dev_priv->dpi_panel_on2 = false; + else if (pipe == 0) + dev_priv->dpi_panel_on = false; */ +} + +static void mdfld_dsi_dpi_set_power(struct drm_encoder *encoder, bool on) +{ + struct mdfld_dsi_encoder *dsi_encoder = mdfld_dsi_encoder(encoder); + struct mdfld_dsi_dpi_output *dpi_output = + MDFLD_DSI_DPI_OUTPUT(dsi_encoder); + struct mdfld_dsi_config *dsi_config = + mdfld_dsi_encoder_get_config(dsi_encoder); + int pipe = mdfld_dsi_encoder_get_pipe(dsi_encoder); + struct drm_device *dev = dsi_config->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + + /*start up display island if it was shutdown*/ + if (!gma_power_begin(dev, true)) + return; + + if (on) { + if (mdfld_get_panel_type(dev, pipe) == TMD_VID) + mdfld_dsi_dpi_turn_on(dpi_output, pipe); + else if (mdfld_get_panel_type(dev, pipe) == TC35876X) + mdfld_dsi_configure_up(dsi_encoder, pipe); + else { + /*enable mipi port*/ + REG_WRITE(MIPI_PORT_CONTROL(pipe), + REG_READ(MIPI_PORT_CONTROL(pipe)) | BIT(31)); + REG_READ(MIPI_PORT_CONTROL(pipe)); + + mdfld_dsi_dpi_turn_on(dpi_output, pipe); + mdfld_dsi_tpo_ic_init(dsi_config, pipe); + } + dev_priv->dpi_panel_on[pipe] = true; + } else { + if (mdfld_get_panel_type(dev, pipe) == TMD_VID) + mdfld_dsi_dpi_shut_down(dpi_output, pipe); + else if (mdfld_get_panel_type(dev, pipe) == TC35876X) + mdfld_dsi_configure_down(dsi_encoder, pipe); + else { + mdfld_dsi_dpi_shut_down(dpi_output, pipe); + + /*disable mipi port*/ + REG_WRITE(MIPI_PORT_CONTROL(pipe), + REG_READ(MIPI_PORT_CONTROL(pipe)) & ~BIT(31)); + REG_READ(MIPI_PORT_CONTROL(pipe)); + } + dev_priv->dpi_panel_on[pipe] = false; + } + gma_power_end(dev); +} + +void mdfld_dsi_dpi_dpms(struct drm_encoder *encoder, int mode) +{ + mdfld_dsi_dpi_set_power(encoder, mode == DRM_MODE_DPMS_ON); +} + +bool mdfld_dsi_dpi_mode_fixup(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + struct mdfld_dsi_encoder *dsi_encoder = mdfld_dsi_encoder(encoder); + struct mdfld_dsi_config *dsi_config = + mdfld_dsi_encoder_get_config(dsi_encoder); + struct drm_display_mode *fixed_mode = dsi_config->fixed_mode; + + if (fixed_mode) { + adjusted_mode->hdisplay = fixed_mode->hdisplay; + adjusted_mode->hsync_start = fixed_mode->hsync_start; + adjusted_mode->hsync_end = fixed_mode->hsync_end; + adjusted_mode->htotal = fixed_mode->htotal; + adjusted_mode->vdisplay = fixed_mode->vdisplay; + adjusted_mode->vsync_start = fixed_mode->vsync_start; + adjusted_mode->vsync_end = fixed_mode->vsync_end; + adjusted_mode->vtotal = fixed_mode->vtotal; + adjusted_mode->clock = fixed_mode->clock; + drm_mode_set_crtcinfo(adjusted_mode, CRTC_INTERLACE_HALVE_V); + } + return true; +} + +void mdfld_dsi_dpi_prepare(struct drm_encoder *encoder) +{ + mdfld_dsi_dpi_set_power(encoder, false); +} + +void mdfld_dsi_dpi_commit(struct drm_encoder *encoder) +{ + mdfld_dsi_dpi_set_power(encoder, true); +} + +/* For TC35876X */ +/* This functionality was implemented in FW in iCDK */ +/* But removed in DV0 and later. So need to add here. */ +static void mipi_set_properties(struct mdfld_dsi_config *dsi_config, int pipe) +{ + struct drm_device *dev = dsi_config->dev; + + REG_WRITE(MIPI_CTRL_REG(pipe), 0x00000018); + REG_WRITE(MIPI_INTR_EN_REG(pipe), 0xffffffff); + REG_WRITE(MIPI_HS_TX_TIMEOUT_REG(pipe), 0xffffff); + REG_WRITE(MIPI_LP_RX_TIMEOUT_REG(pipe), 0xffffff); + REG_WRITE(MIPI_TURN_AROUND_TIMEOUT_REG(pipe), 0x14); + REG_WRITE(MIPI_DEVICE_RESET_TIMER_REG(pipe), 0xff); + REG_WRITE(MIPI_HIGH_LOW_SWITCH_COUNT_REG(pipe), 0x25); + REG_WRITE(MIPI_INIT_COUNT_REG(pipe), 0xf0); + REG_WRITE(MIPI_EOT_DISABLE_REG(pipe), 0x00000000); + REG_WRITE(MIPI_LP_BYTECLK_REG(pipe), 0x00000004); + REG_WRITE(MIPI_DBI_BW_CTRL_REG(pipe), 0x00000820); + REG_WRITE(MIPI_CLK_LANE_SWITCH_TIME_CNT_REG(pipe), (0xa << 16) | 0x14); +} + +static void mdfld_mipi_set_video_timing(struct mdfld_dsi_config *dsi_config, + int pipe) +{ + struct drm_device *dev = dsi_config->dev; + struct mdfld_dsi_dpi_timing dpi_timing; + struct drm_display_mode *mode = dsi_config->mode; + + mdfld_dsi_dpi_timing_calculation(mode, &dpi_timing, + dsi_config->lane_count, + dsi_config->bpp); + + REG_WRITE(MIPI_DPI_RESOLUTION_REG(pipe), + mode->vdisplay << 16 | mode->hdisplay); + REG_WRITE(MIPI_HSYNC_COUNT_REG(pipe), + dpi_timing.hsync_count & DSI_DPI_TIMING_MASK); + REG_WRITE(MIPI_HBP_COUNT_REG(pipe), + dpi_timing.hbp_count & DSI_DPI_TIMING_MASK); + REG_WRITE(MIPI_HFP_COUNT_REG(pipe), + dpi_timing.hfp_count & DSI_DPI_TIMING_MASK); + REG_WRITE(MIPI_HACTIVE_COUNT_REG(pipe), + dpi_timing.hactive_count & DSI_DPI_TIMING_MASK); + REG_WRITE(MIPI_VSYNC_COUNT_REG(pipe), + dpi_timing.vsync_count & DSI_DPI_TIMING_MASK); + REG_WRITE(MIPI_VBP_COUNT_REG(pipe), + dpi_timing.vbp_count & DSI_DPI_TIMING_MASK); + REG_WRITE(MIPI_VFP_COUNT_REG(pipe), + dpi_timing.vfp_count & DSI_DPI_TIMING_MASK); +} + +static void mdfld_mipi_config(struct mdfld_dsi_config *dsi_config, int pipe) +{ + struct drm_device *dev = dsi_config->dev; + int lane_count = dsi_config->lane_count; + + if (pipe) { + REG_WRITE(MIPI_PORT_CONTROL(0), 0x00000002); + REG_WRITE(MIPI_PORT_CONTROL(2), 0x80000000); + } else { + REG_WRITE(MIPI_PORT_CONTROL(0), 0x80010000); + REG_WRITE(MIPI_PORT_CONTROL(2), 0x00); + } + + REG_WRITE(MIPI_DPHY_PARAM_REG(pipe), 0x150A600F); + REG_WRITE(MIPI_VIDEO_MODE_FORMAT_REG(pipe), 0x0000000F); + + /* lane_count = 3 */ + REG_WRITE(MIPI_DSI_FUNC_PRG_REG(pipe), 0x00000200 | lane_count); + + mdfld_mipi_set_video_timing(dsi_config, pipe); +} + +static void mdfld_set_pipe_timing(struct mdfld_dsi_config *dsi_config, int pipe) +{ + struct drm_device *dev = dsi_config->dev; + struct drm_display_mode *mode = dsi_config->mode; + + REG_WRITE(HTOTAL_A, ((mode->htotal - 1) << 16) | (mode->hdisplay - 1)); + REG_WRITE(HBLANK_A, ((mode->htotal - 1) << 16) | (mode->hdisplay - 1)); + REG_WRITE(HSYNC_A, + ((mode->hsync_end - 1) << 16) | (mode->hsync_start - 1)); + + REG_WRITE(VTOTAL_A, ((mode->vtotal - 1) << 16) | (mode->vdisplay - 1)); + REG_WRITE(VBLANK_A, ((mode->vtotal - 1) << 16) | (mode->vdisplay - 1)); + REG_WRITE(VSYNC_A, + ((mode->vsync_end - 1) << 16) | (mode->vsync_start - 1)); + + REG_WRITE(PIPEASRC, + ((mode->hdisplay - 1) << 16) | (mode->vdisplay - 1)); +} +/* End for TC35876X */ + +void mdfld_dsi_dpi_mode_set(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + struct mdfld_dsi_encoder *dsi_encoder = mdfld_dsi_encoder(encoder); + struct mdfld_dsi_dpi_output *dpi_output = + MDFLD_DSI_DPI_OUTPUT(dsi_encoder); + struct mdfld_dsi_config *dsi_config = + mdfld_dsi_encoder_get_config(dsi_encoder); + struct drm_device *dev = dsi_config->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + int pipe = mdfld_dsi_encoder_get_pipe(dsi_encoder); + + u32 pipeconf_reg = PIPEACONF; + u32 dspcntr_reg = DSPACNTR; + + u32 pipeconf = dev_priv->pipeconf[pipe]; + u32 dspcntr = dev_priv->dspcntr[pipe]; + u32 mipi = MIPI_PORT_EN | PASS_FROM_SPHY_TO_AFE | SEL_FLOPPED_HSTX; + + if (pipe) { + pipeconf_reg = PIPECCONF; + dspcntr_reg = DSPCCNTR; + } else { + if (mdfld_get_panel_type(dev, pipe) == TC35876X) + mipi &= (~0x03); /* Use all four lanes */ + else + mipi |= 2; + } + + /*start up display island if it was shutdown*/ + if (!gma_power_begin(dev, true)) + return; + + if (mdfld_get_panel_type(dev, pipe) == TC35876X) { + /* + * The following logic is required to reset the bridge and + * configure. This also starts the DSI clock at 200MHz. + */ + tc35876x_set_bridge_reset_state(dev, 0); /*Pull High Reset */ + tc35876x_toshiba_bridge_panel_on(dev); + udelay(100); + /* Now start the DSI clock */ + REG_WRITE(MRST_DPLL_A, 0x00); + REG_WRITE(MRST_FPA0, 0xC1); + REG_WRITE(MRST_DPLL_A, 0x00800000); + udelay(500); + REG_WRITE(MRST_DPLL_A, 0x80800000); + + if (REG_BIT_WAIT(pipeconf_reg, 1, 29)) + dev_err(&dev->pdev->dev, "%s: DSI PLL lock timeout\n", + __func__); + + REG_WRITE(MIPI_DPHY_PARAM_REG(pipe), 0x2A0c6008); + + mipi_set_properties(dsi_config, pipe); + mdfld_mipi_config(dsi_config, pipe); + mdfld_set_pipe_timing(dsi_config, pipe); + + REG_WRITE(DSPABASE, 0x00); + REG_WRITE(DSPASTRIDE, (mode->hdisplay * 4)); + REG_WRITE(DSPASIZE, + ((mode->vdisplay - 1) << 16) | (mode->hdisplay - 1)); + + REG_WRITE(DSPACNTR, 0x98000000); + REG_WRITE(DSPASURF, 0x00); + + REG_WRITE(VGACNTRL, 0x80000000); + REG_WRITE(DEVICE_READY_REG, 0x00000001); + + REG_WRITE(MIPI_PORT_CONTROL(pipe), 0x80810000); + } else { + /*set up mipi port FIXME: do at init time */ + REG_WRITE(MIPI_PORT_CONTROL(pipe), mipi); + } + REG_READ(MIPI_PORT_CONTROL(pipe)); + + if (mdfld_get_panel_type(dev, pipe) == TMD_VID) { + /* NOP */ + } else if (mdfld_get_panel_type(dev, pipe) == TC35876X) { + /* set up DSI controller DPI interface */ + mdfld_dsi_dpi_controller_init(dsi_config, pipe); + + /* Configure MIPI Bridge and Panel */ + tc35876x_configure_lvds_bridge(dev); + dev_priv->dpi_panel_on[pipe] = true; + } else { + /*turn on DPI interface*/ + mdfld_dsi_dpi_turn_on(dpi_output, pipe); + } + + /*set up pipe*/ + REG_WRITE(pipeconf_reg, pipeconf); + REG_READ(pipeconf_reg); + + /*set up display plane*/ + REG_WRITE(dspcntr_reg, dspcntr); + REG_READ(dspcntr_reg); + + msleep(20); /* FIXME: this should wait for vblank */ + + if (mdfld_get_panel_type(dev, pipe) == TMD_VID) { + /* NOP */ + } else if (mdfld_get_panel_type(dev, pipe) == TC35876X) { + mdfld_dsi_dpi_turn_on(dpi_output, pipe); + } else { + /* init driver ic */ + mdfld_dsi_tpo_ic_init(dsi_config, pipe); + /*init backlight*/ + mdfld_dsi_brightness_init(dsi_config, pipe); + } + + gma_power_end(dev); +} + +/* + * Init DSI DPI encoder. + * Allocate an mdfld_dsi_encoder and attach it to given @dsi_connector + * return pointer of newly allocated DPI encoder, NULL on error + */ +struct mdfld_dsi_encoder *mdfld_dsi_dpi_init(struct drm_device *dev, + struct mdfld_dsi_connector *dsi_connector, + const struct panel_funcs *p_funcs) +{ + struct mdfld_dsi_dpi_output *dpi_output = NULL; + struct mdfld_dsi_config *dsi_config; + struct drm_connector *connector = NULL; + struct drm_encoder *encoder = NULL; + int pipe; + u32 data; + int ret; + + pipe = dsi_connector->pipe; + + if (mdfld_get_panel_type(dev, pipe) != TC35876X) { + dsi_config = mdfld_dsi_get_config(dsi_connector); + + /* panel hard-reset */ + if (p_funcs->reset) { + ret = p_funcs->reset(pipe); + if (ret) { + DRM_ERROR("Panel %d hard-reset failed\n", pipe); + return NULL; + } + } + + /* panel drvIC init */ + if (p_funcs->drv_ic_init) + p_funcs->drv_ic_init(dsi_config, pipe); + + /* panel power mode detect */ + ret = mdfld_dsi_get_power_mode(dsi_config, &data, false); + if (ret) { + DRM_ERROR("Panel %d get power mode failed\n", pipe); + dsi_connector->status = connector_status_disconnected; + } else { + DRM_INFO("pipe %d power mode 0x%x\n", pipe, data); + dsi_connector->status = connector_status_connected; + } + } + + dpi_output = kzalloc(sizeof(struct mdfld_dsi_dpi_output), GFP_KERNEL); + if (!dpi_output) { + DRM_ERROR("No memory\n"); + return NULL; + } + + if (dsi_connector->pipe) + dpi_output->panel_on = 0; + else + dpi_output->panel_on = 0; + + dpi_output->dev = dev; + if (mdfld_get_panel_type(dev, pipe) != TC35876X) + dpi_output->p_funcs = p_funcs; + dpi_output->first_boot = 1; + + /*get fixed mode*/ + dsi_config = mdfld_dsi_get_config(dsi_connector); + + /*create drm encoder object*/ + connector = &dsi_connector->base.base; + encoder = &dpi_output->base.base.base; + drm_encoder_init(dev, + encoder, + p_funcs->encoder_funcs, + DRM_MODE_ENCODER_LVDS); + drm_encoder_helper_add(encoder, + p_funcs->encoder_helper_funcs); + + /*attach to given connector*/ + drm_mode_connector_attach_encoder(connector, encoder); + + /*set possible crtcs and clones*/ + if (dsi_connector->pipe) { + encoder->possible_crtcs = (1 << 2); + encoder->possible_clones = (1 << 1); + } else { + encoder->possible_crtcs = (1 << 0); + encoder->possible_clones = (1 << 0); + } + + dsi_connector->base.encoder = &dpi_output->base.base; + + return &dpi_output->base; +} diff --git a/drivers/gpu/drm/gma500/mdfld_dsi_dpi.h b/drivers/gpu/drm/gma500/mdfld_dsi_dpi.h new file mode 100644 index 0000000..6f76247 --- /dev/null +++ b/drivers/gpu/drm/gma500/mdfld_dsi_dpi.h @@ -0,0 +1,79 @@ +/* + * Copyright © 2010 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * jim liu + * Jackie Li + */ + +#ifndef __MDFLD_DSI_DPI_H__ +#define __MDFLD_DSI_DPI_H__ + +#include "mdfld_dsi_output.h" +#include "mdfld_output.h" + +struct mdfld_dsi_dpi_timing { + u16 hsync_count; + u16 hbp_count; + u16 hfp_count; + u16 hactive_count; + u16 vsync_count; + u16 vbp_count; + u16 vfp_count; +}; + +struct mdfld_dsi_dpi_output { + struct mdfld_dsi_encoder base; + struct drm_device *dev; + + int panel_on; + int first_boot; + + const struct panel_funcs *p_funcs; +}; + +#define MDFLD_DSI_DPI_OUTPUT(dsi_encoder)\ + container_of(dsi_encoder, struct mdfld_dsi_dpi_output, base) + +/* Export functions */ +extern int mdfld_dsi_dpi_timing_calculation(struct drm_display_mode *mode, + struct mdfld_dsi_dpi_timing *dpi_timing, + int num_lane, int bpp); +extern struct mdfld_dsi_encoder *mdfld_dsi_dpi_init(struct drm_device *dev, + struct mdfld_dsi_connector *dsi_connector, + const struct panel_funcs *p_funcs); + +/* MDFLD DPI helper functions */ +extern void mdfld_dsi_dpi_dpms(struct drm_encoder *encoder, int mode); +extern bool mdfld_dsi_dpi_mode_fixup(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode); +extern void mdfld_dsi_dpi_prepare(struct drm_encoder *encoder); +extern void mdfld_dsi_dpi_commit(struct drm_encoder *encoder); +extern void mdfld_dsi_dpi_mode_set(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode); +extern void mdfld_dsi_dpi_turn_on(struct mdfld_dsi_dpi_output *output, + int pipe); +extern void mdfld_dsi_dpi_controller_init(struct mdfld_dsi_config *dsi_config, + int pipe); +#endif /*__MDFLD_DSI_DPI_H__*/ diff --git a/drivers/gpu/drm/gma500/mdfld_dsi_output.c b/drivers/gpu/drm/gma500/mdfld_dsi_output.c new file mode 100644 index 0000000..5675d93 --- /dev/null +++ b/drivers/gpu/drm/gma500/mdfld_dsi_output.c @@ -0,0 +1,621 @@ +/* + * Copyright © 2010 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * jim liu + * Jackie Li + */ + +#include + +#include "mdfld_dsi_output.h" +#include "mdfld_dsi_dpi.h" +#include "mdfld_output.h" +#include "mdfld_dsi_pkg_sender.h" +#include "tc35876x-dsi-lvds.h" +#include +#include + +/* get the LABC from command line. */ +static int LABC_control = 1; + +#ifdef MODULE +module_param(LABC_control, int, 0644); +#else + +static int __init parse_LABC_control(char *arg) +{ + /* LABC control can be passed in as a cmdline parameter */ + /* to enable this feature add LABC=1 to cmdline */ + /* to disable this feature add LABC=0 to cmdline */ + if (!arg) + return -EINVAL; + + if (!strcasecmp(arg, "0")) + LABC_control = 0; + else if (!strcasecmp(arg, "1")) + LABC_control = 1; + + return 0; +} +early_param("LABC", parse_LABC_control); +#endif + +/** + * Check and see if the generic control or data buffer is empty and ready. + */ +void mdfld_dsi_gen_fifo_ready(struct drm_device *dev, u32 gen_fifo_stat_reg, + u32 fifo_stat) +{ + u32 GEN_BF_time_out_count; + + /* Check MIPI Adatper command registers */ + for (GEN_BF_time_out_count = 0; + GEN_BF_time_out_count < GEN_FB_TIME_OUT; + GEN_BF_time_out_count++) { + if ((REG_READ(gen_fifo_stat_reg) & fifo_stat) == fifo_stat) + break; + udelay(100); + } + + if (GEN_BF_time_out_count == GEN_FB_TIME_OUT) + DRM_ERROR("mdfld_dsi_gen_fifo_ready, Timeout. gen_fifo_stat_reg = 0x%x.\n", + gen_fifo_stat_reg); +} + +/** + * Manage the DSI MIPI keyboard and display brightness. + * FIXME: this is exported to OSPM code. should work out an specific + * display interface to OSPM. + */ + +void mdfld_dsi_brightness_init(struct mdfld_dsi_config *dsi_config, int pipe) +{ + struct mdfld_dsi_pkg_sender *sender = + mdfld_dsi_get_pkg_sender(dsi_config); + struct drm_device *dev = sender->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + u32 gen_ctrl_val; + + if (!sender) { + DRM_ERROR("No sender found\n"); + return; + } + + /* Set default display backlight value to 85% (0xd8)*/ + mdfld_dsi_send_mcs_short(sender, write_display_brightness, 0xd8, 1, + true); + + /* Set minimum brightness setting of CABC function to 20% (0x33)*/ + mdfld_dsi_send_mcs_short(sender, write_cabc_min_bright, 0x33, 1, true); + + /* Enable backlight or/and LABC */ + gen_ctrl_val = BRIGHT_CNTL_BLOCK_ON | DISPLAY_DIMMING_ON | + BACKLIGHT_ON; + if (LABC_control == 1) + gen_ctrl_val |= DISPLAY_DIMMING_ON | DISPLAY_BRIGHTNESS_AUTO + | GAMMA_AUTO; + + if (LABC_control == 1) + gen_ctrl_val |= AMBIENT_LIGHT_SENSE_ON; + + dev_priv->mipi_ctrl_display = gen_ctrl_val; + + mdfld_dsi_send_mcs_short(sender, write_ctrl_display, (u8)gen_ctrl_val, + 1, true); + + mdfld_dsi_send_mcs_short(sender, write_ctrl_cabc, UI_IMAGE, 1, true); +} + +void mdfld_dsi_brightness_control(struct drm_device *dev, int pipe, int level) +{ + struct mdfld_dsi_pkg_sender *sender; + struct drm_psb_private *dev_priv; + struct mdfld_dsi_config *dsi_config; + u32 gen_ctrl_val = 0; + int p_type = TMD_VID; + + if (!dev || (pipe != 0 && pipe != 2)) { + DRM_ERROR("Invalid parameter\n"); + return; + } + + p_type = mdfld_get_panel_type(dev, 0); + + dev_priv = dev->dev_private; + + if (pipe) + dsi_config = dev_priv->dsi_configs[1]; + else + dsi_config = dev_priv->dsi_configs[0]; + + sender = mdfld_dsi_get_pkg_sender(dsi_config); + + if (!sender) { + DRM_ERROR("No sender found\n"); + return; + } + + gen_ctrl_val = (level * 0xff / MDFLD_DSI_BRIGHTNESS_MAX_LEVEL) & 0xff; + + dev_dbg(sender->dev->dev, "pipe = %d, gen_ctrl_val = %d.\n", + pipe, gen_ctrl_val); + + if (p_type == TMD_VID) { + /* Set display backlight value */ + mdfld_dsi_send_mcs_short(sender, tmd_write_display_brightness, + (u8)gen_ctrl_val, 1, true); + } else { + /* Set display backlight value */ + mdfld_dsi_send_mcs_short(sender, write_display_brightness, + (u8)gen_ctrl_val, 1, true); + + /* Enable backlight control */ + if (level == 0) + gen_ctrl_val = 0; + else + gen_ctrl_val = dev_priv->mipi_ctrl_display; + + mdfld_dsi_send_mcs_short(sender, write_ctrl_display, + (u8)gen_ctrl_val, 1, true); + } +} + +static int mdfld_dsi_get_panel_status(struct mdfld_dsi_config *dsi_config, + u8 dcs, u32 *data, bool hs) +{ + struct mdfld_dsi_pkg_sender *sender + = mdfld_dsi_get_pkg_sender(dsi_config); + + if (!sender || !data) { + DRM_ERROR("Invalid parameter\n"); + return -EINVAL; + } + + return mdfld_dsi_read_mcs(sender, dcs, data, 1, hs); +} + +int mdfld_dsi_get_power_mode(struct mdfld_dsi_config *dsi_config, u32 *mode, + bool hs) +{ + if (!dsi_config || !mode) { + DRM_ERROR("Invalid parameter\n"); + return -EINVAL; + } + + return mdfld_dsi_get_panel_status(dsi_config, 0x0a, mode, hs); +} + +/* + * NOTE: this function was used by OSPM. + * TODO: will be removed later, should work out display interfaces for OSPM + */ +void mdfld_dsi_controller_init(struct mdfld_dsi_config *dsi_config, int pipe) +{ + if (!dsi_config || ((pipe != 0) && (pipe != 2))) { + DRM_ERROR("Invalid parameters\n"); + return; + } + + mdfld_dsi_dpi_controller_init(dsi_config, pipe); +} + +static void mdfld_dsi_connector_save(struct drm_connector *connector) +{ +} + +static void mdfld_dsi_connector_restore(struct drm_connector *connector) +{ +} + +/* FIXME: start using the force parameter */ +static enum drm_connector_status +mdfld_dsi_connector_detect(struct drm_connector *connector, bool force) +{ + struct mdfld_dsi_connector *dsi_connector + = mdfld_dsi_connector(connector); + + dsi_connector->status = connector_status_connected; + + return dsi_connector->status; +} + +static int mdfld_dsi_connector_set_property(struct drm_connector *connector, + struct drm_property *property, + uint64_t value) +{ + struct drm_encoder *encoder = connector->encoder; + + if (!strcmp(property->name, "scaling mode") && encoder) { + struct psb_intel_crtc *psb_crtc = + to_psb_intel_crtc(encoder->crtc); + bool centerechange; + uint64_t val; + + if (!psb_crtc) + goto set_prop_error; + + switch (value) { + case DRM_MODE_SCALE_FULLSCREEN: + break; + case DRM_MODE_SCALE_NO_SCALE: + break; + case DRM_MODE_SCALE_ASPECT: + break; + default: + goto set_prop_error; + } + + if (drm_connector_property_get_value(connector, property, &val)) + goto set_prop_error; + + if (val == value) + goto set_prop_done; + + if (drm_connector_property_set_value(connector, + property, value)) + goto set_prop_error; + + centerechange = (val == DRM_MODE_SCALE_NO_SCALE) || + (value == DRM_MODE_SCALE_NO_SCALE); + + if (psb_crtc->saved_mode.hdisplay != 0 && + psb_crtc->saved_mode.vdisplay != 0) { + if (centerechange) { + if (!drm_crtc_helper_set_mode(encoder->crtc, + &psb_crtc->saved_mode, + encoder->crtc->x, + encoder->crtc->y, + encoder->crtc->fb)) + goto set_prop_error; + } else { + struct drm_encoder_helper_funcs *funcs = + encoder->helper_private; + funcs->mode_set(encoder, + &psb_crtc->saved_mode, + &psb_crtc->saved_adjusted_mode); + } + } + } else if (!strcmp(property->name, "backlight") && encoder) { + if (drm_connector_property_set_value(connector, property, + value)) + goto set_prop_error; + else { +#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE + struct backlight_device *psb_bd; + + psb_bd = mdfld_get_backlight_device(); + if (psb_bd) { + psb_bd->props.brightness = value; + mdfld_set_brightness(psb_bd); + } +#endif + } + } +set_prop_done: + return 0; +set_prop_error: + return -1; +} + +static void mdfld_dsi_connector_destroy(struct drm_connector *connector) +{ + struct mdfld_dsi_connector *dsi_connector = + mdfld_dsi_connector(connector); + struct mdfld_dsi_pkg_sender *sender; + + if (!dsi_connector) + return; + drm_sysfs_connector_remove(connector); + drm_connector_cleanup(connector); + sender = dsi_connector->pkg_sender; + mdfld_dsi_pkg_sender_destroy(sender); + kfree(dsi_connector); +} + +static int mdfld_dsi_connector_get_modes(struct drm_connector *connector) +{ + struct mdfld_dsi_connector *dsi_connector = + mdfld_dsi_connector(connector); + struct mdfld_dsi_config *dsi_config = + mdfld_dsi_get_config(dsi_connector); + struct drm_display_mode *fixed_mode = dsi_config->fixed_mode; + struct drm_display_mode *dup_mode = NULL; + struct drm_device *dev = connector->dev; + + connector->display_info.min_vfreq = 0; + connector->display_info.max_vfreq = 200; + connector->display_info.min_hfreq = 0; + connector->display_info.max_hfreq = 200; + + if (fixed_mode) { + dev_dbg(dev->dev, "fixed_mode %dx%d\n", + fixed_mode->hdisplay, fixed_mode->vdisplay); + dup_mode = drm_mode_duplicate(dev, fixed_mode); + drm_mode_probed_add(connector, dup_mode); + return 1; + } + DRM_ERROR("Didn't get any modes!\n"); + return 0; +} + +static int mdfld_dsi_connector_mode_valid(struct drm_connector *connector, + struct drm_display_mode *mode) +{ + struct mdfld_dsi_connector *dsi_connector = + mdfld_dsi_connector(connector); + struct mdfld_dsi_config *dsi_config = + mdfld_dsi_get_config(dsi_connector); + struct drm_display_mode *fixed_mode = dsi_config->fixed_mode; + + if (mode->flags & DRM_MODE_FLAG_DBLSCAN) + return MODE_NO_DBLESCAN; + + if (mode->flags & DRM_MODE_FLAG_INTERLACE) + return MODE_NO_INTERLACE; + + /** + * FIXME: current DC has no fitting unit, reject any mode setting + * request + * Will figure out a way to do up-scaling(pannel fitting) later. + **/ + if (fixed_mode) { + if (mode->hdisplay != fixed_mode->hdisplay) + return MODE_PANEL; + + if (mode->vdisplay != fixed_mode->vdisplay) + return MODE_PANEL; + } + + return MODE_OK; +} + +static void mdfld_dsi_connector_dpms(struct drm_connector *connector, int mode) +{ + if (mode == connector->dpms) + return; + + /*first, execute dpms*/ + + drm_helper_connector_dpms(connector, mode); +} + +static struct drm_encoder *mdfld_dsi_connector_best_encoder( + struct drm_connector *connector) +{ + struct mdfld_dsi_connector *dsi_connector = + mdfld_dsi_connector(connector); + struct mdfld_dsi_config *dsi_config = + mdfld_dsi_get_config(dsi_connector); + return &dsi_config->encoder->base.base; +} + +/*DSI connector funcs*/ +static const struct drm_connector_funcs mdfld_dsi_connector_funcs = { + .dpms = /*drm_helper_connector_dpms*/mdfld_dsi_connector_dpms, + .save = mdfld_dsi_connector_save, + .restore = mdfld_dsi_connector_restore, + .detect = mdfld_dsi_connector_detect, + .fill_modes = drm_helper_probe_single_connector_modes, + .set_property = mdfld_dsi_connector_set_property, + .destroy = mdfld_dsi_connector_destroy, +}; + +/*DSI connector helper funcs*/ +static const struct drm_connector_helper_funcs + mdfld_dsi_connector_helper_funcs = { + .get_modes = mdfld_dsi_connector_get_modes, + .mode_valid = mdfld_dsi_connector_mode_valid, + .best_encoder = mdfld_dsi_connector_best_encoder, +}; + +static int mdfld_dsi_get_default_config(struct drm_device *dev, + struct mdfld_dsi_config *config, int pipe) +{ + if (!dev || !config) { + DRM_ERROR("Invalid parameters"); + return -EINVAL; + } + + config->bpp = 24; + if (mdfld_get_panel_type(dev, pipe) == TC35876X) + config->lane_count = 4; + else + config->lane_count = 2; + config->channel_num = 0; + + if (mdfld_get_panel_type(dev, pipe) == TMD_VID) + config->video_mode = MDFLD_DSI_VIDEO_NON_BURST_MODE_SYNC_PULSE; + else if (mdfld_get_panel_type(dev, pipe) == TC35876X) + config->video_mode = + MDFLD_DSI_VIDEO_NON_BURST_MODE_SYNC_EVENTS; + else + config->video_mode = MDFLD_DSI_VIDEO_BURST_MODE; + + return 0; +} + +int mdfld_dsi_panel_reset(int pipe) +{ + unsigned gpio; + int ret = 0; + + switch (pipe) { + case 0: + gpio = 128; + break; + case 2: + gpio = 34; + break; + default: + DRM_ERROR("Invalid output\n"); + return -EINVAL; + } + + ret = gpio_request(gpio, "gfx"); + if (ret) { + DRM_ERROR("gpio_rqueset failed\n"); + return ret; + } + + ret = gpio_direction_output(gpio, 1); + if (ret) { + DRM_ERROR("gpio_direction_output failed\n"); + goto gpio_error; + } + + gpio_get_value(128); + +gpio_error: + if (gpio_is_valid(gpio)) + gpio_free(gpio); + + return ret; +} + +/* + * MIPI output init + * @dev drm device + * @pipe pipe number. 0 or 2 + * @config + * + * Do the initialization of a MIPI output, including create DRM mode objects + * initialization of DSI output on @pipe + */ +void mdfld_dsi_output_init(struct drm_device *dev, + int pipe, + const struct panel_funcs *p_vid_funcs) +{ + struct mdfld_dsi_config *dsi_config; + struct mdfld_dsi_connector *dsi_connector; + struct drm_connector *connector; + struct mdfld_dsi_encoder *encoder; + struct drm_psb_private *dev_priv = dev->dev_private; + struct panel_info dsi_panel_info; + u32 width_mm, height_mm; + + dev_dbg(dev->dev, "init DSI output on pipe %d\n", pipe); + + if (!dev || ((pipe != 0) && (pipe != 2))) { + DRM_ERROR("Invalid parameter\n"); + return; + } + + /*create a new connetor*/ + dsi_connector = kzalloc(sizeof(struct mdfld_dsi_connector), GFP_KERNEL); + if (!dsi_connector) { + DRM_ERROR("No memory"); + return; + } + + dsi_connector->pipe = pipe; + + dsi_config = kzalloc(sizeof(struct mdfld_dsi_config), + GFP_KERNEL); + if (!dsi_config) { + DRM_ERROR("cannot allocate memory for DSI config\n"); + goto dsi_init_err0; + } + mdfld_dsi_get_default_config(dev, dsi_config, pipe); + + dsi_connector->private = dsi_config; + + dsi_config->changed = 1; + dsi_config->dev = dev; + + dsi_config->fixed_mode = p_vid_funcs->get_config_mode(dev); + if (p_vid_funcs->get_panel_info(dev, pipe, &dsi_panel_info)) + goto dsi_init_err0; + + width_mm = dsi_panel_info.width_mm; + height_mm = dsi_panel_info.height_mm; + + dsi_config->mode = dsi_config->fixed_mode; + dsi_config->connector = dsi_connector; + + if (!dsi_config->fixed_mode) { + DRM_ERROR("No pannel fixed mode was found\n"); + goto dsi_init_err0; + } + + if (pipe && dev_priv->dsi_configs[0]) { + dsi_config->dvr_ic_inited = 0; + dev_priv->dsi_configs[1] = dsi_config; + } else if (pipe == 0) { + dsi_config->dvr_ic_inited = 1; + dev_priv->dsi_configs[0] = dsi_config; + } else { + DRM_ERROR("Trying to init MIPI1 before MIPI0\n"); + goto dsi_init_err0; + } + + + connector = &dsi_connector->base.base; + drm_connector_init(dev, connector, &mdfld_dsi_connector_funcs, + DRM_MODE_CONNECTOR_LVDS); + drm_connector_helper_add(connector, &mdfld_dsi_connector_helper_funcs); + + connector->display_info.subpixel_order = SubPixelHorizontalRGB; + connector->display_info.width_mm = width_mm; + connector->display_info.height_mm = height_mm; + connector->interlace_allowed = false; + connector->doublescan_allowed = false; + + /*attach properties*/ + drm_connector_attach_property(connector, + dev->mode_config.scaling_mode_property, + DRM_MODE_SCALE_FULLSCREEN); + drm_connector_attach_property(connector, + dev_priv->backlight_property, + MDFLD_DSI_BRIGHTNESS_MAX_LEVEL); + + /*init DSI package sender on this output*/ + if (mdfld_dsi_pkg_sender_init(dsi_connector, pipe)) { + DRM_ERROR("Package Sender initialization failed on pipe %d\n", + pipe); + goto dsi_init_err0; + } + + encoder = mdfld_dsi_dpi_init(dev, dsi_connector, p_vid_funcs); + if (!encoder) { + DRM_ERROR("Create DPI encoder failed\n"); + goto dsi_init_err1; + } + encoder->private = dsi_config; + dsi_config->encoder = encoder; + encoder->base.type = (pipe == 0) ? INTEL_OUTPUT_MIPI : + INTEL_OUTPUT_MIPI2; + drm_sysfs_connector_add(connector); + return; + + /*TODO: add code to destroy outputs on error*/ +dsi_init_err1: + /*destroy sender*/ + mdfld_dsi_pkg_sender_destroy(dsi_connector->pkg_sender); + + drm_connector_cleanup(connector); + + kfree(dsi_config->fixed_mode); + kfree(dsi_config); +dsi_init_err0: + kfree(dsi_connector); +} diff --git a/drivers/gpu/drm/gma500/mdfld_dsi_output.h b/drivers/gpu/drm/gma500/mdfld_dsi_output.h new file mode 100644 index 0000000..36eb074 --- /dev/null +++ b/drivers/gpu/drm/gma500/mdfld_dsi_output.h @@ -0,0 +1,377 @@ +/* + * Copyright © 2010 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * jim liu + * Jackie Li + */ + +#ifndef __MDFLD_DSI_OUTPUT_H__ +#define __MDFLD_DSI_OUTPUT_H__ + +#include +#include +#include +#include +#include + +#include "psb_drv.h" +#include "psb_intel_drv.h" +#include "psb_intel_reg.h" +#include "mdfld_output.h" + +#include + +#define FLD_MASK(start, end) (((1 << ((start) - (end) + 1)) - 1) << (end)) +#define FLD_VAL(val, start, end) (((val) << (end)) & FLD_MASK(start, end)) +#define FLD_GET(val, start, end) (((val) & FLD_MASK(start, end)) >> (end)) +#define FLD_MOD(orig, val, start, end) \ + (((orig) & ~FLD_MASK(start, end)) | FLD_VAL(val, start, end)) + +#define REG_FLD_MOD(reg, val, start, end) \ + REG_WRITE(reg, FLD_MOD(REG_READ(reg), val, start, end)) + +static inline int REGISTER_FLD_WAIT(struct drm_device *dev, u32 reg, + u32 val, int start, int end) +{ + int t = 100000; + + while (FLD_GET(REG_READ(reg), start, end) != val) { + if (--t == 0) + return 1; + } + + return 0; +} + +#define REG_FLD_WAIT(reg, val, start, end) \ + REGISTER_FLD_WAIT(dev, reg, val, start, end) + +#define REG_BIT_WAIT(reg, val, bitnum) \ + REGISTER_FLD_WAIT(dev, reg, val, bitnum, bitnum) + +#define MDFLD_DSI_BRIGHTNESS_MAX_LEVEL 100 + +#ifdef DEBUG +#define CHECK_PIPE(pipe) ({ \ + const typeof(pipe) __pipe = (pipe); \ + BUG_ON(__pipe != 0 && __pipe != 2); \ + __pipe; }) +#else +#define CHECK_PIPE(pipe) (pipe) +#endif + +/* + * Actual MIPIA->MIPIC reg offset is 0x800, value 0x400 is valid for 0 and 2 + */ +#define REG_OFFSET(pipe) (CHECK_PIPE(pipe) * 0x400) + +/* mdfld DSI controller registers */ +#define MIPI_DEVICE_READY_REG(pipe) (0xb000 + REG_OFFSET(pipe)) +#define MIPI_INTR_STAT_REG(pipe) (0xb004 + REG_OFFSET(pipe)) +#define MIPI_INTR_EN_REG(pipe) (0xb008 + REG_OFFSET(pipe)) +#define MIPI_DSI_FUNC_PRG_REG(pipe) (0xb00c + REG_OFFSET(pipe)) +#define MIPI_HS_TX_TIMEOUT_REG(pipe) (0xb010 + REG_OFFSET(pipe)) +#define MIPI_LP_RX_TIMEOUT_REG(pipe) (0xb014 + REG_OFFSET(pipe)) +#define MIPI_TURN_AROUND_TIMEOUT_REG(pipe) (0xb018 + REG_OFFSET(pipe)) +#define MIPI_DEVICE_RESET_TIMER_REG(pipe) (0xb01c + REG_OFFSET(pipe)) +#define MIPI_DPI_RESOLUTION_REG(pipe) (0xb020 + REG_OFFSET(pipe)) +#define MIPI_DBI_FIFO_THROTTLE_REG(pipe) (0xb024 + REG_OFFSET(pipe)) +#define MIPI_HSYNC_COUNT_REG(pipe) (0xb028 + REG_OFFSET(pipe)) +#define MIPI_HBP_COUNT_REG(pipe) (0xb02c + REG_OFFSET(pipe)) +#define MIPI_HFP_COUNT_REG(pipe) (0xb030 + REG_OFFSET(pipe)) +#define MIPI_HACTIVE_COUNT_REG(pipe) (0xb034 + REG_OFFSET(pipe)) +#define MIPI_VSYNC_COUNT_REG(pipe) (0xb038 + REG_OFFSET(pipe)) +#define MIPI_VBP_COUNT_REG(pipe) (0xb03c + REG_OFFSET(pipe)) +#define MIPI_VFP_COUNT_REG(pipe) (0xb040 + REG_OFFSET(pipe)) +#define MIPI_HIGH_LOW_SWITCH_COUNT_REG(pipe) (0xb044 + REG_OFFSET(pipe)) +#define MIPI_DPI_CONTROL_REG(pipe) (0xb048 + REG_OFFSET(pipe)) +#define MIPI_DPI_DATA_REG(pipe) (0xb04c + REG_OFFSET(pipe)) +#define MIPI_INIT_COUNT_REG(pipe) (0xb050 + REG_OFFSET(pipe)) +#define MIPI_MAX_RETURN_PACK_SIZE_REG(pipe) (0xb054 + REG_OFFSET(pipe)) +#define MIPI_VIDEO_MODE_FORMAT_REG(pipe) (0xb058 + REG_OFFSET(pipe)) +#define MIPI_EOT_DISABLE_REG(pipe) (0xb05c + REG_OFFSET(pipe)) +#define MIPI_LP_BYTECLK_REG(pipe) (0xb060 + REG_OFFSET(pipe)) +#define MIPI_LP_GEN_DATA_REG(pipe) (0xb064 + REG_OFFSET(pipe)) +#define MIPI_HS_GEN_DATA_REG(pipe) (0xb068 + REG_OFFSET(pipe)) +#define MIPI_LP_GEN_CTRL_REG(pipe) (0xb06c + REG_OFFSET(pipe)) +#define MIPI_HS_GEN_CTRL_REG(pipe) (0xb070 + REG_OFFSET(pipe)) +#define MIPI_GEN_FIFO_STAT_REG(pipe) (0xb074 + REG_OFFSET(pipe)) +#define MIPI_HS_LS_DBI_ENABLE_REG(pipe) (0xb078 + REG_OFFSET(pipe)) +#define MIPI_DPHY_PARAM_REG(pipe) (0xb080 + REG_OFFSET(pipe)) +#define MIPI_DBI_BW_CTRL_REG(pipe) (0xb084 + REG_OFFSET(pipe)) +#define MIPI_CLK_LANE_SWITCH_TIME_CNT_REG(pipe) (0xb088 + REG_OFFSET(pipe)) + +#define MIPI_CTRL_REG(pipe) (0xb104 + REG_OFFSET(pipe)) +#define MIPI_DATA_ADD_REG(pipe) (0xb108 + REG_OFFSET(pipe)) +#define MIPI_DATA_LEN_REG(pipe) (0xb10c + REG_OFFSET(pipe)) +#define MIPI_CMD_ADD_REG(pipe) (0xb110 + REG_OFFSET(pipe)) +#define MIPI_CMD_LEN_REG(pipe) (0xb114 + REG_OFFSET(pipe)) + +/* non-uniform reg offset */ +#define MIPI_PORT_CONTROL(pipe) (CHECK_PIPE(pipe) ? MIPI_C : MIPI) + +#define DSI_DEVICE_READY (0x1) +#define DSI_POWER_STATE_ULPS_ENTER (0x2 << 1) +#define DSI_POWER_STATE_ULPS_EXIT (0x1 << 1) +#define DSI_POWER_STATE_ULPS_OFFSET (0x1) + + +#define DSI_ONE_DATA_LANE (0x1) +#define DSI_TWO_DATA_LANE (0x2) +#define DSI_THREE_DATA_LANE (0X3) +#define DSI_FOUR_DATA_LANE (0x4) +#define DSI_DPI_VIRT_CHANNEL_OFFSET (0x3) +#define DSI_DBI_VIRT_CHANNEL_OFFSET (0x5) +#define DSI_DPI_COLOR_FORMAT_RGB565 (0x01 << 7) +#define DSI_DPI_COLOR_FORMAT_RGB666 (0x02 << 7) +#define DSI_DPI_COLOR_FORMAT_RGB666_UNPACK (0x03 << 7) +#define DSI_DPI_COLOR_FORMAT_RGB888 (0x04 << 7) +#define DSI_DBI_COLOR_FORMAT_OPTION2 (0x05 << 13) + +#define DSI_INTR_STATE_RXSOTERROR BIT(0) + +#define DSI_INTR_STATE_SPL_PKG_SENT BIT(30) +#define DSI_INTR_STATE_TE BIT(31) + +#define DSI_HS_TX_TIMEOUT_MASK (0xffffff) + +#define DSI_LP_RX_TIMEOUT_MASK (0xffffff) + +#define DSI_TURN_AROUND_TIMEOUT_MASK (0x3f) + +#define DSI_RESET_TIMER_MASK (0xffff) + +#define DSI_DBI_FIFO_WM_HALF (0x0) +#define DSI_DBI_FIFO_WM_QUARTER (0x1) +#define DSI_DBI_FIFO_WM_LOW (0x2) + +#define DSI_DPI_TIMING_MASK (0xffff) + +#define DSI_INIT_TIMER_MASK (0xffff) + +#define DSI_DBI_RETURN_PACK_SIZE_MASK (0x3ff) + +#define DSI_LP_BYTECLK_MASK (0x0ffff) + +#define DSI_HS_CTRL_GEN_SHORT_W0 (0x03) +#define DSI_HS_CTRL_GEN_SHORT_W1 (0x13) +#define DSI_HS_CTRL_GEN_SHORT_W2 (0x23) +#define DSI_HS_CTRL_GEN_R0 (0x04) +#define DSI_HS_CTRL_GEN_R1 (0x14) +#define DSI_HS_CTRL_GEN_R2 (0x24) +#define DSI_HS_CTRL_GEN_LONG_W (0x29) +#define DSI_HS_CTRL_MCS_SHORT_W0 (0x05) +#define DSI_HS_CTRL_MCS_SHORT_W1 (0x15) +#define DSI_HS_CTRL_MCS_R0 (0x06) +#define DSI_HS_CTRL_MCS_LONG_W (0x39) +#define DSI_HS_CTRL_VC_OFFSET (0x06) +#define DSI_HS_CTRL_WC_OFFSET (0x08) + +#define DSI_FIFO_GEN_HS_DATA_FULL BIT(0) +#define DSI_FIFO_GEN_HS_DATA_HALF_EMPTY BIT(1) +#define DSI_FIFO_GEN_HS_DATA_EMPTY BIT(2) +#define DSI_FIFO_GEN_LP_DATA_FULL BIT(8) +#define DSI_FIFO_GEN_LP_DATA_HALF_EMPTY BIT(9) +#define DSI_FIFO_GEN_LP_DATA_EMPTY BIT(10) +#define DSI_FIFO_GEN_HS_CTRL_FULL BIT(16) +#define DSI_FIFO_GEN_HS_CTRL_HALF_EMPTY BIT(17) +#define DSI_FIFO_GEN_HS_CTRL_EMPTY BIT(18) +#define DSI_FIFO_GEN_LP_CTRL_FULL BIT(24) +#define DSI_FIFO_GEN_LP_CTRL_HALF_EMPTY BIT(25) +#define DSI_FIFO_GEN_LP_CTRL_EMPTY BIT(26) +#define DSI_FIFO_DBI_EMPTY BIT(27) +#define DSI_FIFO_DPI_EMPTY BIT(28) + +#define DSI_DBI_HS_LP_SWITCH_MASK (0x1) + +#define DSI_HS_LP_SWITCH_COUNTER_OFFSET (0x0) +#define DSI_LP_HS_SWITCH_COUNTER_OFFSET (0x16) + +#define DSI_DPI_CTRL_HS_SHUTDOWN (0x00000001) +#define DSI_DPI_CTRL_HS_TURN_ON (0x00000002) + +/*dsi power modes*/ +#define DSI_POWER_MODE_DISPLAY_ON BIT(2) +#define DSI_POWER_MODE_NORMAL_ON BIT(3) +#define DSI_POWER_MODE_SLEEP_OUT BIT(4) +#define DSI_POWER_MODE_PARTIAL_ON BIT(5) +#define DSI_POWER_MODE_IDLE_ON BIT(6) + +enum { + MDFLD_DSI_VIDEO_NON_BURST_MODE_SYNC_PULSE = 1, + MDFLD_DSI_VIDEO_NON_BURST_MODE_SYNC_EVENTS = 2, + MDFLD_DSI_VIDEO_BURST_MODE = 3, +}; + +#define DSI_DPI_COMPLETE_LAST_LINE BIT(2) +#define DSI_DPI_DISABLE_BTA BIT(3) + +struct mdfld_dsi_connector { + struct psb_intel_connector base; + + int pipe; + void *private; + void *pkg_sender; + + /* Connection status */ + enum drm_connector_status status; +}; + +struct mdfld_dsi_encoder { + struct psb_intel_encoder base; + void *private; +}; + +/* + * DSI config, consists of one DSI connector, two DSI encoders. + * DRM will pick up on DSI encoder basing on differents configs. + */ +struct mdfld_dsi_config { + struct drm_device *dev; + struct drm_display_mode *fixed_mode; + struct drm_display_mode *mode; + + struct mdfld_dsi_connector *connector; + struct mdfld_dsi_encoder *encoder; + + int changed; + + int bpp; + int lane_count; + /*Virtual channel number for this encoder*/ + int channel_num; + /*video mode configure*/ + int video_mode; + + int dvr_ic_inited; +}; + +static inline struct mdfld_dsi_connector *mdfld_dsi_connector( + struct drm_connector *connector) +{ + struct psb_intel_connector *psb_connector; + + psb_connector = to_psb_intel_connector(connector); + + return container_of(psb_connector, struct mdfld_dsi_connector, base); +} + +static inline struct mdfld_dsi_encoder *mdfld_dsi_encoder( + struct drm_encoder *encoder) +{ + struct psb_intel_encoder *psb_encoder; + + psb_encoder = to_psb_intel_encoder(encoder); + + return container_of(psb_encoder, struct mdfld_dsi_encoder, base); +} + +static inline struct mdfld_dsi_config * + mdfld_dsi_get_config(struct mdfld_dsi_connector *connector) +{ + if (!connector) + return NULL; + return (struct mdfld_dsi_config *)connector->private; +} + +static inline void *mdfld_dsi_get_pkg_sender(struct mdfld_dsi_config *config) +{ + struct mdfld_dsi_connector *dsi_connector; + + if (!config) + return NULL; + + dsi_connector = config->connector; + + if (!dsi_connector) + return NULL; + + return dsi_connector->pkg_sender; +} + +static inline struct mdfld_dsi_config * + mdfld_dsi_encoder_get_config(struct mdfld_dsi_encoder *encoder) +{ + if (!encoder) + return NULL; + return (struct mdfld_dsi_config *)encoder->private; +} + +static inline struct mdfld_dsi_connector * + mdfld_dsi_encoder_get_connector(struct mdfld_dsi_encoder *encoder) +{ + struct mdfld_dsi_config *config; + + if (!encoder) + return NULL; + + config = mdfld_dsi_encoder_get_config(encoder); + if (!config) + return NULL; + + return config->connector; +} + +static inline void *mdfld_dsi_encoder_get_pkg_sender( + struct mdfld_dsi_encoder *encoder) +{ + struct mdfld_dsi_config *dsi_config; + + dsi_config = mdfld_dsi_encoder_get_config(encoder); + if (!dsi_config) + return NULL; + + return mdfld_dsi_get_pkg_sender(dsi_config); +} + +static inline int mdfld_dsi_encoder_get_pipe(struct mdfld_dsi_encoder *encoder) +{ + struct mdfld_dsi_connector *connector; + + if (!encoder) + return -1; + + connector = mdfld_dsi_encoder_get_connector(encoder); + if (!connector) + return -1; + return connector->pipe; +} + +/* Export functions */ +extern void mdfld_dsi_gen_fifo_ready(struct drm_device *dev, + u32 gen_fifo_stat_reg, u32 fifo_stat); +extern void mdfld_dsi_brightness_init(struct mdfld_dsi_config *dsi_config, + int pipe); +extern void mdfld_dsi_brightness_control(struct drm_device *dev, int pipe, + int level); +extern void mdfld_dsi_output_init(struct drm_device *dev, + int pipe, + const struct panel_funcs *p_vid_funcs); +extern void mdfld_dsi_controller_init(struct mdfld_dsi_config *dsi_config, + int pipe); + +extern int mdfld_dsi_get_power_mode(struct mdfld_dsi_config *dsi_config, + u32 *mode, bool hs); +extern int mdfld_dsi_panel_reset(int pipe); + +#endif /*__MDFLD_DSI_OUTPUT_H__*/ diff --git a/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.c b/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.c new file mode 100644 index 0000000..baa0e14 --- /dev/null +++ b/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.c @@ -0,0 +1,694 @@ +/* + * Copyright © 2010 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * Jackie Li + */ + +#include + +#include "mdfld_dsi_output.h" +#include "mdfld_dsi_pkg_sender.h" +#include "mdfld_dsi_dpi.h" + +#define MDFLD_DSI_READ_MAX_COUNT 5000 + +enum data_type { + DSI_DT_GENERIC_SHORT_WRITE_0 = 0x03, + DSI_DT_GENERIC_SHORT_WRITE_1 = 0x13, + DSI_DT_GENERIC_SHORT_WRITE_2 = 0x23, + DSI_DT_GENERIC_READ_0 = 0x04, + DSI_DT_GENERIC_READ_1 = 0x14, + DSI_DT_GENERIC_READ_2 = 0x24, + DSI_DT_GENERIC_LONG_WRITE = 0x29, + DSI_DT_DCS_SHORT_WRITE_0 = 0x05, + DSI_DT_DCS_SHORT_WRITE_1 = 0x15, + DSI_DT_DCS_READ = 0x06, + DSI_DT_DCS_LONG_WRITE = 0x39, +}; + +enum { + MDFLD_DSI_PANEL_MODE_SLEEP = 0x1, +}; + +enum { + MDFLD_DSI_PKG_SENDER_FREE = 0x0, + MDFLD_DSI_PKG_SENDER_BUSY = 0x1, +}; + +static const char *const dsi_errors[] = { + "RX SOT Error", + "RX SOT Sync Error", + "RX EOT Sync Error", + "RX Escape Mode Entry Error", + "RX LP TX Sync Error", + "RX HS Receive Timeout Error", + "RX False Control Error", + "RX ECC Single Bit Error", + "RX ECC Multibit Error", + "RX Checksum Error", + "RX DSI Data Type Not Recognised", + "RX DSI VC ID Invalid", + "TX False Control Error", + "TX ECC Single Bit Error", + "TX ECC Multibit Error", + "TX Checksum Error", + "TX DSI Data Type Not Recognised", + "TX DSI VC ID invalid", + "High Contention", + "Low contention", + "DPI FIFO Under run", + "HS TX Timeout", + "LP RX Timeout", + "Turn Around ACK Timeout", + "ACK With No Error", + "RX Invalid TX Length", + "RX Prot Violation", + "HS Generic Write FIFO Full", + "LP Generic Write FIFO Full", + "Generic Read Data Avail" + "Special Packet Sent", + "Tearing Effect", +}; + +static inline int wait_for_gen_fifo_empty(struct mdfld_dsi_pkg_sender *sender, + u32 mask) +{ + struct drm_device *dev = sender->dev; + u32 gen_fifo_stat_reg = sender->mipi_gen_fifo_stat_reg; + int retry = 0xffff; + + while (retry--) { + if ((mask & REG_READ(gen_fifo_stat_reg)) == mask) + return 0; + udelay(100); + } + DRM_ERROR("fifo is NOT empty 0x%08x\n", REG_READ(gen_fifo_stat_reg)); + return -EIO; +} + +static int wait_for_all_fifos_empty(struct mdfld_dsi_pkg_sender *sender) +{ + return wait_for_gen_fifo_empty(sender, (BIT(2) | BIT(10) | BIT(18) | + BIT(26) | BIT(27) | BIT(28))); +} + +static int wait_for_lp_fifos_empty(struct mdfld_dsi_pkg_sender *sender) +{ + return wait_for_gen_fifo_empty(sender, (BIT(10) | BIT(26))); +} + +static int wait_for_hs_fifos_empty(struct mdfld_dsi_pkg_sender *sender) +{ + return wait_for_gen_fifo_empty(sender, (BIT(2) | BIT(18))); +} + +static int handle_dsi_error(struct mdfld_dsi_pkg_sender *sender, u32 mask) +{ + u32 intr_stat_reg = sender->mipi_intr_stat_reg; + struct drm_device *dev = sender->dev; + + dev_dbg(sender->dev->dev, "Handling error 0x%08x\n", mask); + + switch (mask) { + case BIT(0): + case BIT(1): + case BIT(2): + case BIT(3): + case BIT(4): + case BIT(5): + case BIT(6): + case BIT(7): + case BIT(8): + case BIT(9): + case BIT(10): + case BIT(11): + case BIT(12): + case BIT(13): + dev_dbg(sender->dev->dev, "No Action required\n"); + break; + case BIT(14): + /*wait for all fifo empty*/ + /*wait_for_all_fifos_empty(sender)*/; + break; + case BIT(15): + dev_dbg(sender->dev->dev, "No Action required\n"); + break; + case BIT(16): + break; + case BIT(17): + break; + case BIT(18): + case BIT(19): + dev_dbg(sender->dev->dev, "High/Low contention detected\n"); + /*wait for contention recovery time*/ + /*mdelay(10);*/ + /*wait for all fifo empty*/ + if (0) + wait_for_all_fifos_empty(sender); + break; + case BIT(20): + dev_dbg(sender->dev->dev, "No Action required\n"); + break; + case BIT(21): + /*wait for all fifo empty*/ + /*wait_for_all_fifos_empty(sender);*/ + break; + case BIT(22): + break; + case BIT(23): + case BIT(24): + case BIT(25): + case BIT(26): + case BIT(27): + dev_dbg(sender->dev->dev, "HS Gen fifo full\n"); + REG_WRITE(intr_stat_reg, mask); + wait_for_hs_fifos_empty(sender); + break; + case BIT(28): + dev_dbg(sender->dev->dev, "LP Gen fifo full\n"); + REG_WRITE(intr_stat_reg, mask); + wait_for_lp_fifos_empty(sender); + break; + case BIT(29): + case BIT(30): + case BIT(31): + dev_dbg(sender->dev->dev, "No Action required\n"); + break; + } + + if (mask & REG_READ(intr_stat_reg)) + dev_dbg(sender->dev->dev, + "Cannot clean interrupt 0x%08x\n", mask); + return 0; +} + +static int dsi_error_handler(struct mdfld_dsi_pkg_sender *sender) +{ + struct drm_device *dev = sender->dev; + u32 intr_stat_reg = sender->mipi_intr_stat_reg; + u32 mask; + u32 intr_stat; + int i; + int err = 0; + + intr_stat = REG_READ(intr_stat_reg); + + for (i = 0; i < 32; i++) { + mask = (0x00000001UL) << i; + if (intr_stat & mask) { + dev_dbg(sender->dev->dev, "[DSI]: %s\n", dsi_errors[i]); + err = handle_dsi_error(sender, mask); + if (err) + DRM_ERROR("Cannot handle error\n"); + } + } + return err; +} + +static int send_short_pkg(struct mdfld_dsi_pkg_sender *sender, u8 data_type, + u8 cmd, u8 param, bool hs) +{ + struct drm_device *dev = sender->dev; + u32 ctrl_reg; + u32 val; + u8 virtual_channel = 0; + + if (hs) { + ctrl_reg = sender->mipi_hs_gen_ctrl_reg; + + /* FIXME: wait_for_hs_fifos_empty(sender); */ + } else { + ctrl_reg = sender->mipi_lp_gen_ctrl_reg; + + /* FIXME: wait_for_lp_fifos_empty(sender); */ + } + + val = FLD_VAL(param, 23, 16) | FLD_VAL(cmd, 15, 8) | + FLD_VAL(virtual_channel, 7, 6) | FLD_VAL(data_type, 5, 0); + + REG_WRITE(ctrl_reg, val); + + return 0; +} + +static int send_long_pkg(struct mdfld_dsi_pkg_sender *sender, u8 data_type, + u8 *data, int len, bool hs) +{ + struct drm_device *dev = sender->dev; + u32 ctrl_reg; + u32 data_reg; + u32 val; + u8 *p; + u8 b1, b2, b3, b4; + u8 virtual_channel = 0; + int i; + + if (hs) { + ctrl_reg = sender->mipi_hs_gen_ctrl_reg; + data_reg = sender->mipi_hs_gen_data_reg; + + /* FIXME: wait_for_hs_fifos_empty(sender); */ + } else { + ctrl_reg = sender->mipi_lp_gen_ctrl_reg; + data_reg = sender->mipi_lp_gen_data_reg; + + /* FIXME: wait_for_lp_fifos_empty(sender); */ + } + + p = data; + for (i = 0; i < len / 4; i++) { + b1 = *p++; + b2 = *p++; + b3 = *p++; + b4 = *p++; + + REG_WRITE(data_reg, b4 << 24 | b3 << 16 | b2 << 8 | b1); + } + + i = len % 4; + if (i) { + b1 = 0; b2 = 0; b3 = 0; + + switch (i) { + case 3: + b1 = *p++; + b2 = *p++; + b3 = *p++; + break; + case 2: + b1 = *p++; + b2 = *p++; + break; + case 1: + b1 = *p++; + break; + } + + REG_WRITE(data_reg, b3 << 16 | b2 << 8 | b1); + } + + val = FLD_VAL(len, 23, 8) | FLD_VAL(virtual_channel, 7, 6) | + FLD_VAL(data_type, 5, 0); + + REG_WRITE(ctrl_reg, val); + + return 0; +} + +static int send_pkg_prepare(struct mdfld_dsi_pkg_sender *sender, u8 data_type, + u8 *data, u16 len) +{ + u8 cmd; + + switch (data_type) { + case DSI_DT_DCS_SHORT_WRITE_0: + case DSI_DT_DCS_SHORT_WRITE_1: + case DSI_DT_DCS_LONG_WRITE: + cmd = *data; + break; + default: + return 0; + } + + /*this prevents other package sending while doing msleep*/ + sender->status = MDFLD_DSI_PKG_SENDER_BUSY; + + /*wait for 120 milliseconds in case exit_sleep_mode just be sent*/ + if (unlikely(cmd == DCS_ENTER_SLEEP_MODE)) { + /*TODO: replace it with msleep later*/ + mdelay(120); + } + + if (unlikely(cmd == DCS_EXIT_SLEEP_MODE)) { + /*TODO: replace it with msleep later*/ + mdelay(120); + } + return 0; +} + +static int send_pkg_done(struct mdfld_dsi_pkg_sender *sender, u8 data_type, + u8 *data, u16 len) +{ + u8 cmd; + + switch (data_type) { + case DSI_DT_DCS_SHORT_WRITE_0: + case DSI_DT_DCS_SHORT_WRITE_1: + case DSI_DT_DCS_LONG_WRITE: + cmd = *data; + break; + default: + return 0; + } + + /*update panel status*/ + if (unlikely(cmd == DCS_ENTER_SLEEP_MODE)) { + sender->panel_mode |= MDFLD_DSI_PANEL_MODE_SLEEP; + /*TODO: replace it with msleep later*/ + mdelay(120); + } else if (unlikely(cmd == DCS_EXIT_SLEEP_MODE)) { + sender->panel_mode &= ~MDFLD_DSI_PANEL_MODE_SLEEP; + /*TODO: replace it with msleep later*/ + mdelay(120); + } else if (unlikely(cmd == DCS_SOFT_RESET)) { + /*TODO: replace it with msleep later*/ + mdelay(5); + } + + sender->status = MDFLD_DSI_PKG_SENDER_FREE; + + return 0; +} + +static int send_pkg(struct mdfld_dsi_pkg_sender *sender, u8 data_type, + u8 *data, u16 len, bool hs) +{ + int ret; + + /*handle DSI error*/ + ret = dsi_error_handler(sender); + if (ret) { + DRM_ERROR("Error handling failed\n"); + return -EAGAIN; + } + + /* send pkg */ + if (sender->status == MDFLD_DSI_PKG_SENDER_BUSY) { + DRM_ERROR("sender is busy\n"); + return -EAGAIN; + } + + ret = send_pkg_prepare(sender, data_type, data, len); + if (ret) { + DRM_ERROR("send_pkg_prepare error\n"); + return ret; + } + + switch (data_type) { + case DSI_DT_GENERIC_SHORT_WRITE_0: + case DSI_DT_GENERIC_SHORT_WRITE_1: + case DSI_DT_GENERIC_SHORT_WRITE_2: + case DSI_DT_GENERIC_READ_0: + case DSI_DT_GENERIC_READ_1: + case DSI_DT_GENERIC_READ_2: + case DSI_DT_DCS_SHORT_WRITE_0: + case DSI_DT_DCS_SHORT_WRITE_1: + case DSI_DT_DCS_READ: + ret = send_short_pkg(sender, data_type, data[0], data[1], hs); + break; + case DSI_DT_GENERIC_LONG_WRITE: + case DSI_DT_DCS_LONG_WRITE: + ret = send_long_pkg(sender, data_type, data, len, hs); + break; + } + + send_pkg_done(sender, data_type, data, len); + + /*FIXME: should I query complete and fifo empty here?*/ + + return ret; +} + +int mdfld_dsi_send_mcs_long(struct mdfld_dsi_pkg_sender *sender, u8 *data, + u32 len, bool hs) +{ + unsigned long flags; + + if (!sender || !data || !len) { + DRM_ERROR("Invalid parameters\n"); + return -EINVAL; + } + + spin_lock_irqsave(&sender->lock, flags); + send_pkg(sender, DSI_DT_DCS_LONG_WRITE, data, len, hs); + spin_unlock_irqrestore(&sender->lock, flags); + + return 0; +} + +int mdfld_dsi_send_mcs_short(struct mdfld_dsi_pkg_sender *sender, u8 cmd, + u8 param, u8 param_num, bool hs) +{ + u8 data[2]; + unsigned long flags; + u8 data_type; + + if (!sender) { + DRM_ERROR("Invalid parameter\n"); + return -EINVAL; + } + + data[0] = cmd; + + if (param_num) { + data_type = DSI_DT_DCS_SHORT_WRITE_1; + data[1] = param; + } else { + data_type = DSI_DT_DCS_SHORT_WRITE_0; + data[1] = 0; + } + + spin_lock_irqsave(&sender->lock, flags); + send_pkg(sender, data_type, data, sizeof(data), hs); + spin_unlock_irqrestore(&sender->lock, flags); + + return 0; +} + +int mdfld_dsi_send_gen_short(struct mdfld_dsi_pkg_sender *sender, u8 param0, + u8 param1, u8 param_num, bool hs) +{ + u8 data[2]; + unsigned long flags; + u8 data_type; + + if (!sender || param_num > 2) { + DRM_ERROR("Invalid parameter\n"); + return -EINVAL; + } + + switch (param_num) { + case 0: + data_type = DSI_DT_GENERIC_SHORT_WRITE_0; + data[0] = 0; + data[1] = 0; + break; + case 1: + data_type = DSI_DT_GENERIC_SHORT_WRITE_1; + data[0] = param0; + data[1] = 0; + break; + case 2: + data_type = DSI_DT_GENERIC_SHORT_WRITE_2; + data[0] = param0; + data[1] = param1; + break; + } + + spin_lock_irqsave(&sender->lock, flags); + send_pkg(sender, data_type, data, sizeof(data), hs); + spin_unlock_irqrestore(&sender->lock, flags); + + return 0; +} + +int mdfld_dsi_send_gen_long(struct mdfld_dsi_pkg_sender *sender, u8 *data, + u32 len, bool hs) +{ + unsigned long flags; + + if (!sender || !data || !len) { + DRM_ERROR("Invalid parameters\n"); + return -EINVAL; + } + + spin_lock_irqsave(&sender->lock, flags); + send_pkg(sender, DSI_DT_GENERIC_LONG_WRITE, data, len, hs); + spin_unlock_irqrestore(&sender->lock, flags); + + return 0; +} + +static int __read_panel_data(struct mdfld_dsi_pkg_sender *sender, u8 data_type, + u8 *data, u16 len, u32 *data_out, u16 len_out, bool hs) +{ + unsigned long flags; + struct drm_device *dev = sender->dev; + int i; + u32 gen_data_reg; + int retry = MDFLD_DSI_READ_MAX_COUNT; + + if (!sender || !data_out || !len_out) { + DRM_ERROR("Invalid parameters\n"); + return -EINVAL; + } + + /** + * do reading. + * 0) send out generic read request + * 1) polling read data avail interrupt + * 2) read data + */ + spin_lock_irqsave(&sender->lock, flags); + + REG_WRITE(sender->mipi_intr_stat_reg, BIT(29)); + + if ((REG_READ(sender->mipi_intr_stat_reg) & BIT(29))) + DRM_ERROR("Can NOT clean read data valid interrupt\n"); + + /*send out read request*/ + send_pkg(sender, data_type, data, len, hs); + + /*polling read data avail interrupt*/ + while (retry && !(REG_READ(sender->mipi_intr_stat_reg) & BIT(29))) { + udelay(100); + retry--; + } + + if (!retry) { + spin_unlock_irqrestore(&sender->lock, flags); + return -ETIMEDOUT; + } + + REG_WRITE(sender->mipi_intr_stat_reg, BIT(29)); + + /*read data*/ + if (hs) + gen_data_reg = sender->mipi_hs_gen_data_reg; + else + gen_data_reg = sender->mipi_lp_gen_data_reg; + + for (i = 0; i < len_out; i++) + *(data_out + i) = REG_READ(gen_data_reg); + + spin_unlock_irqrestore(&sender->lock, flags); + + return 0; +} + +int mdfld_dsi_read_mcs(struct mdfld_dsi_pkg_sender *sender, u8 cmd, + u32 *data, u16 len, bool hs) +{ + if (!sender || !data || !len) { + DRM_ERROR("Invalid parameters\n"); + return -EINVAL; + } + + return __read_panel_data(sender, DSI_DT_DCS_READ, &cmd, 1, + data, len, hs); +} + +int mdfld_dsi_pkg_sender_init(struct mdfld_dsi_connector *dsi_connector, + int pipe) +{ + struct mdfld_dsi_pkg_sender *pkg_sender; + struct mdfld_dsi_config *dsi_config = + mdfld_dsi_get_config(dsi_connector); + struct drm_device *dev = dsi_config->dev; + u32 mipi_val = 0; + + if (!dsi_connector) { + DRM_ERROR("Invalid parameter\n"); + return -EINVAL; + } + + pkg_sender = dsi_connector->pkg_sender; + + if (!pkg_sender || IS_ERR(pkg_sender)) { + pkg_sender = kzalloc(sizeof(struct mdfld_dsi_pkg_sender), + GFP_KERNEL); + if (!pkg_sender) { + DRM_ERROR("Create DSI pkg sender failed\n"); + return -ENOMEM; + } + dsi_connector->pkg_sender = (void *)pkg_sender; + } + + pkg_sender->dev = dev; + pkg_sender->dsi_connector = dsi_connector; + pkg_sender->pipe = pipe; + pkg_sender->pkg_num = 0; + pkg_sender->panel_mode = 0; + pkg_sender->status = MDFLD_DSI_PKG_SENDER_FREE; + + /*init regs*/ + if (pipe == 0) { + pkg_sender->dpll_reg = MRST_DPLL_A; + pkg_sender->dspcntr_reg = DSPACNTR; + pkg_sender->pipeconf_reg = PIPEACONF; + pkg_sender->dsplinoff_reg = DSPALINOFF; + pkg_sender->dspsurf_reg = DSPASURF; + pkg_sender->pipestat_reg = PIPEASTAT; + } else if (pipe == 2) { + pkg_sender->dpll_reg = MRST_DPLL_A; + pkg_sender->dspcntr_reg = DSPCCNTR; + pkg_sender->pipeconf_reg = PIPECCONF; + pkg_sender->dsplinoff_reg = DSPCLINOFF; + pkg_sender->dspsurf_reg = DSPCSURF; + pkg_sender->pipestat_reg = PIPECSTAT; + } + + pkg_sender->mipi_intr_stat_reg = MIPI_INTR_STAT_REG(pipe); + pkg_sender->mipi_lp_gen_data_reg = MIPI_LP_GEN_DATA_REG(pipe); + pkg_sender->mipi_hs_gen_data_reg = MIPI_HS_GEN_DATA_REG(pipe); + pkg_sender->mipi_lp_gen_ctrl_reg = MIPI_LP_GEN_CTRL_REG(pipe); + pkg_sender->mipi_hs_gen_ctrl_reg = MIPI_HS_GEN_CTRL_REG(pipe); + pkg_sender->mipi_gen_fifo_stat_reg = MIPI_GEN_FIFO_STAT_REG(pipe); + pkg_sender->mipi_data_addr_reg = MIPI_DATA_ADD_REG(pipe); + pkg_sender->mipi_data_len_reg = MIPI_DATA_LEN_REG(pipe); + pkg_sender->mipi_cmd_addr_reg = MIPI_CMD_ADD_REG(pipe); + pkg_sender->mipi_cmd_len_reg = MIPI_CMD_LEN_REG(pipe); + + /*init lock*/ + spin_lock_init(&pkg_sender->lock); + + if (mdfld_get_panel_type(dev, pipe) != TC35876X) { + /** + * For video mode, don't enable DPI timing output here, + * will init the DPI timing output during mode setting. + */ + mipi_val = PASS_FROM_SPHY_TO_AFE | SEL_FLOPPED_HSTX; + + if (pipe == 0) + mipi_val |= 0x2; + + REG_WRITE(MIPI_PORT_CONTROL(pipe), mipi_val); + REG_READ(MIPI_PORT_CONTROL(pipe)); + + /* do dsi controller init */ + mdfld_dsi_controller_init(dsi_config, pipe); + } + + return 0; +} + +void mdfld_dsi_pkg_sender_destroy(struct mdfld_dsi_pkg_sender *sender) +{ + if (!sender || IS_ERR(sender)) + return; + + /*free*/ + kfree(sender); +} + + diff --git a/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.h b/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.h new file mode 100644 index 0000000..459cd7e --- /dev/null +++ b/drivers/gpu/drm/gma500/mdfld_dsi_pkg_sender.h @@ -0,0 +1,92 @@ +/* + * Copyright © 2010 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * Jackie Li + */ +#ifndef __MDFLD_DSI_PKG_SENDER_H__ +#define __MDFLD_DSI_PKG_SENDER_H__ + +#include + +#define MDFLD_MAX_DCS_PARAM 8 + +struct mdfld_dsi_pkg_sender { + struct drm_device *dev; + struct mdfld_dsi_connector *dsi_connector; + u32 status; + u32 panel_mode; + + int pipe; + + spinlock_t lock; + + u32 pkg_num; + + /* Registers */ + u32 dpll_reg; + u32 dspcntr_reg; + u32 pipeconf_reg; + u32 pipestat_reg; + u32 dsplinoff_reg; + u32 dspsurf_reg; + + u32 mipi_intr_stat_reg; + u32 mipi_lp_gen_data_reg; + u32 mipi_hs_gen_data_reg; + u32 mipi_lp_gen_ctrl_reg; + u32 mipi_hs_gen_ctrl_reg; + u32 mipi_gen_fifo_stat_reg; + u32 mipi_data_addr_reg; + u32 mipi_data_len_reg; + u32 mipi_cmd_addr_reg; + u32 mipi_cmd_len_reg; +}; + +/* DCS definitions */ +#define DCS_SOFT_RESET 0x01 +#define DCS_ENTER_SLEEP_MODE 0x10 +#define DCS_EXIT_SLEEP_MODE 0x11 +#define DCS_SET_DISPLAY_OFF 0x28 +#define DCS_SET_DISPLAY_ON 0x29 +#define DCS_SET_COLUMN_ADDRESS 0x2a +#define DCS_SET_PAGE_ADDRESS 0x2b +#define DCS_WRITE_MEM_START 0x2c +#define DCS_SET_TEAR_OFF 0x34 +#define DCS_SET_TEAR_ON 0x35 + +extern int mdfld_dsi_pkg_sender_init(struct mdfld_dsi_connector *dsi_connector, + int pipe); +extern void mdfld_dsi_pkg_sender_destroy(struct mdfld_dsi_pkg_sender *sender); +int mdfld_dsi_send_mcs_short(struct mdfld_dsi_pkg_sender *sender, u8 cmd, + u8 param, u8 param_num, bool hs); +int mdfld_dsi_send_mcs_long(struct mdfld_dsi_pkg_sender *sender, u8 *data, + u32 len, bool hs); +int mdfld_dsi_send_gen_short(struct mdfld_dsi_pkg_sender *sender, u8 param0, + u8 param1, u8 param_num, bool hs); +int mdfld_dsi_send_gen_long(struct mdfld_dsi_pkg_sender *sender, u8 *data, + u32 len, bool hs); +/* Read interfaces */ +int mdfld_dsi_read_mcs(struct mdfld_dsi_pkg_sender *sender, u8 cmd, + u32 *data, u16 len, bool hs); + +#endif diff --git a/drivers/gpu/drm/gma500/mdfld_intel_display.c b/drivers/gpu/drm/gma500/mdfld_intel_display.c new file mode 100644 index 0000000..a35a292 --- /dev/null +++ b/drivers/gpu/drm/gma500/mdfld_intel_display.c @@ -0,0 +1,1180 @@ +/* + * Copyright © 2006-2007 Intel Corporation + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + * Authors: + * Eric Anholt + */ + +#include +#include + +#include +#include "psb_intel_reg.h" +#include "psb_intel_display.h" +#include "framebuffer.h" +#include "mdfld_output.h" +#include "mdfld_dsi_output.h" + +/* Hardcoded currently */ +static int ksel = KSEL_CRYSTAL_19; + +struct psb_intel_range_t { + int min, max; +}; + +struct mrst_limit_t { + struct psb_intel_range_t dot, m, p1; +}; + +struct mrst_clock_t { + /* derived values */ + int dot; + int m; + int p1; +}; + +#define COUNT_MAX 0x10000000 + +void mdfldWaitForPipeDisable(struct drm_device *dev, int pipe) +{ + int count, temp; + u32 pipeconf_reg = PIPEACONF; + + switch (pipe) { + case 0: + break; + case 1: + pipeconf_reg = PIPEBCONF; + break; + case 2: + pipeconf_reg = PIPECCONF; + break; + default: + DRM_ERROR("Illegal Pipe Number.\n"); + return; + } + + /* FIXME JLIU7_PO */ + psb_intel_wait_for_vblank(dev); + return; + + /* Wait for for the pipe disable to take effect. */ + for (count = 0; count < COUNT_MAX; count++) { + temp = REG_READ(pipeconf_reg); + if ((temp & PIPEACONF_PIPE_STATE) == 0) + break; + } +} + +void mdfldWaitForPipeEnable(struct drm_device *dev, int pipe) +{ + int count, temp; + u32 pipeconf_reg = PIPEACONF; + + switch (pipe) { + case 0: + break; + case 1: + pipeconf_reg = PIPEBCONF; + break; + case 2: + pipeconf_reg = PIPECCONF; + break; + default: + DRM_ERROR("Illegal Pipe Number.\n"); + return; + } + + /* FIXME JLIU7_PO */ + psb_intel_wait_for_vblank(dev); + return; + + /* Wait for for the pipe enable to take effect. */ + for (count = 0; count < COUNT_MAX; count++) { + temp = REG_READ(pipeconf_reg); + if ((temp & PIPEACONF_PIPE_STATE) == 1) + break; + } +} + +static void psb_intel_crtc_prepare(struct drm_crtc *crtc) +{ + struct drm_crtc_helper_funcs *crtc_funcs = crtc->helper_private; + crtc_funcs->dpms(crtc, DRM_MODE_DPMS_OFF); +} + +static void psb_intel_crtc_commit(struct drm_crtc *crtc) +{ + struct drm_crtc_helper_funcs *crtc_funcs = crtc->helper_private; + crtc_funcs->dpms(crtc, DRM_MODE_DPMS_ON); +} + +static bool psb_intel_crtc_mode_fixup(struct drm_crtc *crtc, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + return true; +} + +/** + * Return the pipe currently connected to the panel fitter, + * or -1 if the panel fitter is not present or not in use + */ +static int psb_intel_panel_fitter_pipe(struct drm_device *dev) +{ + u32 pfit_control; + + pfit_control = REG_READ(PFIT_CONTROL); + + /* See if the panel fitter is in use */ + if ((pfit_control & PFIT_ENABLE) == 0) + return -1; + + /* 965 can place panel fitter on either pipe */ + return (pfit_control >> 29) & 0x3; +} + +static struct drm_device globle_dev; + +void mdfld__intel_plane_set_alpha(int enable) +{ + struct drm_device *dev = &globle_dev; + int dspcntr_reg = DSPACNTR; + u32 dspcntr; + + dspcntr = REG_READ(dspcntr_reg); + + if (enable) { + dspcntr &= ~DISPPLANE_32BPP_NO_ALPHA; + dspcntr |= DISPPLANE_32BPP; + } else { + dspcntr &= ~DISPPLANE_32BPP; + dspcntr |= DISPPLANE_32BPP_NO_ALPHA; + } + + REG_WRITE(dspcntr_reg, dspcntr); +} + +static int check_fb(struct drm_framebuffer *fb) +{ + if (!fb) + return 0; + + switch (fb->bits_per_pixel) { + case 8: + case 16: + case 24: + case 32: + return 0; + default: + DRM_ERROR("Unknown color depth\n"); + return -EINVAL; + } +} + +static int mdfld__intel_pipe_set_base(struct drm_crtc *crtc, int x, int y, + struct drm_framebuffer *old_fb) +{ + struct drm_device *dev = crtc->dev; + /* struct drm_i915_master_private *master_priv; */ + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + struct psb_framebuffer *psbfb = to_psb_fb(crtc->fb); + int pipe = psb_intel_crtc->pipe; + unsigned long start, offset; + int dsplinoff = DSPALINOFF; + int dspsurf = DSPASURF; + int dspstride = DSPASTRIDE; + int dspcntr_reg = DSPACNTR; + u32 dspcntr; + int ret; + + memcpy(&globle_dev, dev, sizeof(struct drm_device)); + + dev_dbg(dev->dev, "pipe = 0x%x.\n", pipe); + + /* no fb bound */ + if (!crtc->fb) { + dev_dbg(dev->dev, "No FB bound\n"); + return 0; + } + + ret = check_fb(crtc->fb); + if (ret) + return ret; + + switch (pipe) { + case 0: + dsplinoff = DSPALINOFF; + break; + case 1: + dsplinoff = DSPBLINOFF; + dspsurf = DSPBSURF; + dspstride = DSPBSTRIDE; + dspcntr_reg = DSPBCNTR; + break; + case 2: + dsplinoff = DSPCLINOFF; + dspsurf = DSPCSURF; + dspstride = DSPCSTRIDE; + dspcntr_reg = DSPCCNTR; + break; + default: + DRM_ERROR("Illegal Pipe Number.\n"); + return -EINVAL; + } + + if (!gma_power_begin(dev, true)) + return 0; + + start = psbfb->gtt->offset; + offset = y * crtc->fb->pitches[0] + x * (crtc->fb->bits_per_pixel / 8); + + REG_WRITE(dspstride, crtc->fb->pitches[0]); + dspcntr = REG_READ(dspcntr_reg); + dspcntr &= ~DISPPLANE_PIXFORMAT_MASK; + + switch (crtc->fb->bits_per_pixel) { + case 8: + dspcntr |= DISPPLANE_8BPP; + break; + case 16: + if (crtc->fb->depth == 15) + dspcntr |= DISPPLANE_15_16BPP; + else + dspcntr |= DISPPLANE_16BPP; + break; + case 24: + case 32: + dspcntr |= DISPPLANE_32BPP_NO_ALPHA; + break; + } + REG_WRITE(dspcntr_reg, dspcntr); + + dev_dbg(dev->dev, "Writing base %08lX %08lX %d %d\n", + start, offset, x, y); + REG_WRITE(dsplinoff, offset); + REG_READ(dsplinoff); + REG_WRITE(dspsurf, start); + REG_READ(dspsurf); + + gma_power_end(dev); + + return 0; +} + +/* + * Disable the pipe, plane and pll. + * + */ +void mdfld_disable_crtc(struct drm_device *dev, int pipe) +{ + int dpll_reg = MRST_DPLL_A; + int dspcntr_reg = DSPACNTR; + int dspbase_reg = MRST_DSPABASE; + int pipeconf_reg = PIPEACONF; + u32 temp; + + dev_dbg(dev->dev, "pipe = %d\n", pipe); + + + switch (pipe) { + case 0: + break; + case 1: + dpll_reg = MDFLD_DPLL_B; + dspcntr_reg = DSPBCNTR; + dspbase_reg = DSPBSURF; + pipeconf_reg = PIPEBCONF; + break; + case 2: + dpll_reg = MRST_DPLL_A; + dspcntr_reg = DSPCCNTR; + dspbase_reg = MDFLD_DSPCBASE; + pipeconf_reg = PIPECCONF; + break; + default: + DRM_ERROR("Illegal Pipe Number.\n"); + return; + } + + if (pipe != 1) + mdfld_dsi_gen_fifo_ready(dev, MIPI_GEN_FIFO_STAT_REG(pipe), + HS_CTRL_FIFO_EMPTY | HS_DATA_FIFO_EMPTY); + + /* Disable display plane */ + temp = REG_READ(dspcntr_reg); + if ((temp & DISPLAY_PLANE_ENABLE) != 0) { + REG_WRITE(dspcntr_reg, + temp & ~DISPLAY_PLANE_ENABLE); + /* Flush the plane changes */ + REG_WRITE(dspbase_reg, REG_READ(dspbase_reg)); + REG_READ(dspbase_reg); + } + + /* FIXME_JLIU7 MDFLD_PO revisit */ + + /* Next, disable display pipes */ + temp = REG_READ(pipeconf_reg); + if ((temp & PIPEACONF_ENABLE) != 0) { + temp &= ~PIPEACONF_ENABLE; + temp |= PIPECONF_PLANE_OFF | PIPECONF_CURSOR_OFF; + REG_WRITE(pipeconf_reg, temp); + REG_READ(pipeconf_reg); + + /* Wait for for the pipe disable to take effect. */ + mdfldWaitForPipeDisable(dev, pipe); + } + + temp = REG_READ(dpll_reg); + if (temp & DPLL_VCO_ENABLE) { + if ((pipe != 1 && + !((REG_READ(PIPEACONF) | REG_READ(PIPECCONF)) + & PIPEACONF_ENABLE)) || pipe == 1) { + temp &= ~(DPLL_VCO_ENABLE); + REG_WRITE(dpll_reg, temp); + REG_READ(dpll_reg); + /* Wait for the clocks to turn off. */ + /* FIXME_MDFLD PO may need more delay */ + udelay(500); + + if (!(temp & MDFLD_PWR_GATE_EN)) { + /* gating power of DPLL */ + REG_WRITE(dpll_reg, temp | MDFLD_PWR_GATE_EN); + /* FIXME_MDFLD PO - change 500 to 1 after PO */ + udelay(5000); + } + } + } + +} + +/** + * Sets the power management mode of the pipe and plane. + * + * This code should probably grow support for turning the cursor off and back + * on appropriately at the same time as we're turning the pipe off/on. + */ +static void mdfld_crtc_dpms(struct drm_crtc *crtc, int mode) +{ + struct drm_device *dev = crtc->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + int pipe = psb_intel_crtc->pipe; + int dpll_reg = MRST_DPLL_A; + int dspcntr_reg = DSPACNTR; + int dspbase_reg = MRST_DSPABASE; + int pipeconf_reg = PIPEACONF; + u32 pipestat_reg = PIPEASTAT; + u32 pipeconf = dev_priv->pipeconf[pipe]; + u32 temp; + int timeout = 0; + + dev_dbg(dev->dev, "mode = %d, pipe = %d\n", mode, pipe); + +/* FIXME_JLIU7 MDFLD_PO replaced w/ the following function */ +/* mdfld_dbi_dpms (struct drm_device *dev, int pipe, bool enabled) */ + + switch (pipe) { + case 0: + break; + case 1: + dpll_reg = DPLL_B; + dspcntr_reg = DSPBCNTR; + dspbase_reg = MRST_DSPBBASE; + pipeconf_reg = PIPEBCONF; + dpll_reg = MDFLD_DPLL_B; + break; + case 2: + dpll_reg = MRST_DPLL_A; + dspcntr_reg = DSPCCNTR; + dspbase_reg = MDFLD_DSPCBASE; + pipeconf_reg = PIPECCONF; + pipestat_reg = PIPECSTAT; + break; + default: + DRM_ERROR("Illegal Pipe Number.\n"); + return; + } + + if (!gma_power_begin(dev, true)) + return; + + /* XXX: When our outputs are all unaware of DPMS modes other than off + * and on, we should map those modes to DRM_MODE_DPMS_OFF in the CRTC. + */ + switch (mode) { + case DRM_MODE_DPMS_ON: + case DRM_MODE_DPMS_STANDBY: + case DRM_MODE_DPMS_SUSPEND: + /* Enable the DPLL */ + temp = REG_READ(dpll_reg); + + if ((temp & DPLL_VCO_ENABLE) == 0) { + /* When ungating power of DPLL, needs to wait 0.5us + before enable the VCO */ + if (temp & MDFLD_PWR_GATE_EN) { + temp &= ~MDFLD_PWR_GATE_EN; + REG_WRITE(dpll_reg, temp); + /* FIXME_MDFLD PO - change 500 to 1 after PO */ + udelay(500); + } + + REG_WRITE(dpll_reg, temp); + REG_READ(dpll_reg); + /* FIXME_MDFLD PO - change 500 to 1 after PO */ + udelay(500); + + REG_WRITE(dpll_reg, temp | DPLL_VCO_ENABLE); + REG_READ(dpll_reg); + + /** + * wait for DSI PLL to lock + * NOTE: only need to poll status of pipe 0 and pipe 1, + * since both MIPI pipes share the same PLL. + */ + while ((pipe != 2) && (timeout < 20000) && + !(REG_READ(pipeconf_reg) & PIPECONF_DSIPLL_LOCK)) { + udelay(150); + timeout++; + } + } + + /* Enable the plane */ + temp = REG_READ(dspcntr_reg); + if ((temp & DISPLAY_PLANE_ENABLE) == 0) { + REG_WRITE(dspcntr_reg, + temp | DISPLAY_PLANE_ENABLE); + /* Flush the plane changes */ + REG_WRITE(dspbase_reg, REG_READ(dspbase_reg)); + } + + /* Enable the pipe */ + temp = REG_READ(pipeconf_reg); + if ((temp & PIPEACONF_ENABLE) == 0) { + REG_WRITE(pipeconf_reg, pipeconf); + + /* Wait for for the pipe enable to take effect. */ + mdfldWaitForPipeEnable(dev, pipe); + } + + /*workaround for sighting 3741701 Random X blank display*/ + /*perform w/a in video mode only on pipe A or C*/ + if (pipe == 0 || pipe == 2) { + REG_WRITE(pipestat_reg, REG_READ(pipestat_reg)); + msleep(100); + if (PIPE_VBLANK_STATUS & REG_READ(pipestat_reg)) + dev_dbg(dev->dev, "OK"); + else { + dev_dbg(dev->dev, "STUCK!!!!"); + /*shutdown controller*/ + temp = REG_READ(dspcntr_reg); + REG_WRITE(dspcntr_reg, + temp & ~DISPLAY_PLANE_ENABLE); + REG_WRITE(dspbase_reg, REG_READ(dspbase_reg)); + /*mdfld_dsi_dpi_shut_down(dev, pipe);*/ + REG_WRITE(0xb048, 1); + msleep(100); + temp = REG_READ(pipeconf_reg); + temp &= ~PIPEACONF_ENABLE; + REG_WRITE(pipeconf_reg, temp); + msleep(100); /*wait for pipe disable*/ + REG_WRITE(MIPI_DEVICE_READY_REG(pipe), 0); + msleep(100); + REG_WRITE(0xb004, REG_READ(0xb004)); + /* try to bring the controller back up again*/ + REG_WRITE(MIPI_DEVICE_READY_REG(pipe), 1); + temp = REG_READ(dspcntr_reg); + REG_WRITE(dspcntr_reg, + temp | DISPLAY_PLANE_ENABLE); + REG_WRITE(dspbase_reg, REG_READ(dspbase_reg)); + /*mdfld_dsi_dpi_turn_on(dev, pipe);*/ + REG_WRITE(0xb048, 2); + msleep(100); + temp = REG_READ(pipeconf_reg); + temp |= PIPEACONF_ENABLE; + REG_WRITE(pipeconf_reg, temp); + } + } + + psb_intel_crtc_load_lut(crtc); + + /* Give the overlay scaler a chance to enable + if it's on this pipe */ + /* psb_intel_crtc_dpms_video(crtc, true); TODO */ + + break; + case DRM_MODE_DPMS_OFF: + /* Give the overlay scaler a chance to disable + * if it's on this pipe */ + /* psb_intel_crtc_dpms_video(crtc, FALSE); TODO */ + if (pipe != 1) + mdfld_dsi_gen_fifo_ready(dev, + MIPI_GEN_FIFO_STAT_REG(pipe), + HS_CTRL_FIFO_EMPTY | HS_DATA_FIFO_EMPTY); + + /* Disable the VGA plane that we never use */ + REG_WRITE(VGACNTRL, VGA_DISP_DISABLE); + + /* Disable display plane */ + temp = REG_READ(dspcntr_reg); + if ((temp & DISPLAY_PLANE_ENABLE) != 0) { + REG_WRITE(dspcntr_reg, + temp & ~DISPLAY_PLANE_ENABLE); + /* Flush the plane changes */ + REG_WRITE(dspbase_reg, REG_READ(dspbase_reg)); + REG_READ(dspbase_reg); + } + + /* Next, disable display pipes */ + temp = REG_READ(pipeconf_reg); + if ((temp & PIPEACONF_ENABLE) != 0) { + temp &= ~PIPEACONF_ENABLE; + temp |= PIPECONF_PLANE_OFF | PIPECONF_CURSOR_OFF; + REG_WRITE(pipeconf_reg, temp); + REG_READ(pipeconf_reg); + + /* Wait for for the pipe disable to take effect. */ + mdfldWaitForPipeDisable(dev, pipe); + } + + temp = REG_READ(dpll_reg); + if (temp & DPLL_VCO_ENABLE) { + if ((pipe != 1 && !((REG_READ(PIPEACONF) + | REG_READ(PIPECCONF)) & PIPEACONF_ENABLE)) + || pipe == 1) { + temp &= ~(DPLL_VCO_ENABLE); + REG_WRITE(dpll_reg, temp); + REG_READ(dpll_reg); + /* Wait for the clocks to turn off. */ + /* FIXME_MDFLD PO may need more delay */ + udelay(500); + } + } + break; + } + gma_power_end(dev); +} + + +#define MDFLD_LIMT_DPLL_19 0 +#define MDFLD_LIMT_DPLL_25 1 +#define MDFLD_LIMT_DPLL_83 2 +#define MDFLD_LIMT_DPLL_100 3 +#define MDFLD_LIMT_DSIPLL_19 4 +#define MDFLD_LIMT_DSIPLL_25 5 +#define MDFLD_LIMT_DSIPLL_83 6 +#define MDFLD_LIMT_DSIPLL_100 7 + +#define MDFLD_DOT_MIN 19750 +#define MDFLD_DOT_MAX 120000 +#define MDFLD_DPLL_M_MIN_19 113 +#define MDFLD_DPLL_M_MAX_19 155 +#define MDFLD_DPLL_P1_MIN_19 2 +#define MDFLD_DPLL_P1_MAX_19 10 +#define MDFLD_DPLL_M_MIN_25 101 +#define MDFLD_DPLL_M_MAX_25 130 +#define MDFLD_DPLL_P1_MIN_25 2 +#define MDFLD_DPLL_P1_MAX_25 10 +#define MDFLD_DPLL_M_MIN_83 64 +#define MDFLD_DPLL_M_MAX_83 64 +#define MDFLD_DPLL_P1_MIN_83 2 +#define MDFLD_DPLL_P1_MAX_83 2 +#define MDFLD_DPLL_M_MIN_100 64 +#define MDFLD_DPLL_M_MAX_100 64 +#define MDFLD_DPLL_P1_MIN_100 2 +#define MDFLD_DPLL_P1_MAX_100 2 +#define MDFLD_DSIPLL_M_MIN_19 131 +#define MDFLD_DSIPLL_M_MAX_19 175 +#define MDFLD_DSIPLL_P1_MIN_19 3 +#define MDFLD_DSIPLL_P1_MAX_19 8 +#define MDFLD_DSIPLL_M_MIN_25 97 +#define MDFLD_DSIPLL_M_MAX_25 140 +#define MDFLD_DSIPLL_P1_MIN_25 3 +#define MDFLD_DSIPLL_P1_MAX_25 9 +#define MDFLD_DSIPLL_M_MIN_83 33 +#define MDFLD_DSIPLL_M_MAX_83 92 +#define MDFLD_DSIPLL_P1_MIN_83 2 +#define MDFLD_DSIPLL_P1_MAX_83 3 +#define MDFLD_DSIPLL_M_MIN_100 97 +#define MDFLD_DSIPLL_M_MAX_100 140 +#define MDFLD_DSIPLL_P1_MIN_100 3 +#define MDFLD_DSIPLL_P1_MAX_100 9 + +static const struct mrst_limit_t mdfld_limits[] = { + { /* MDFLD_LIMT_DPLL_19 */ + .dot = {.min = MDFLD_DOT_MIN, .max = MDFLD_DOT_MAX}, + .m = {.min = MDFLD_DPLL_M_MIN_19, .max = MDFLD_DPLL_M_MAX_19}, + .p1 = {.min = MDFLD_DPLL_P1_MIN_19, .max = MDFLD_DPLL_P1_MAX_19}, + }, + { /* MDFLD_LIMT_DPLL_25 */ + .dot = {.min = MDFLD_DOT_MIN, .max = MDFLD_DOT_MAX}, + .m = {.min = MDFLD_DPLL_M_MIN_25, .max = MDFLD_DPLL_M_MAX_25}, + .p1 = {.min = MDFLD_DPLL_P1_MIN_25, .max = MDFLD_DPLL_P1_MAX_25}, + }, + { /* MDFLD_LIMT_DPLL_83 */ + .dot = {.min = MDFLD_DOT_MIN, .max = MDFLD_DOT_MAX}, + .m = {.min = MDFLD_DPLL_M_MIN_83, .max = MDFLD_DPLL_M_MAX_83}, + .p1 = {.min = MDFLD_DPLL_P1_MIN_83, .max = MDFLD_DPLL_P1_MAX_83}, + }, + { /* MDFLD_LIMT_DPLL_100 */ + .dot = {.min = MDFLD_DOT_MIN, .max = MDFLD_DOT_MAX}, + .m = {.min = MDFLD_DPLL_M_MIN_100, .max = MDFLD_DPLL_M_MAX_100}, + .p1 = {.min = MDFLD_DPLL_P1_MIN_100, .max = MDFLD_DPLL_P1_MAX_100}, + }, + { /* MDFLD_LIMT_DSIPLL_19 */ + .dot = {.min = MDFLD_DOT_MIN, .max = MDFLD_DOT_MAX}, + .m = {.min = MDFLD_DSIPLL_M_MIN_19, .max = MDFLD_DSIPLL_M_MAX_19}, + .p1 = {.min = MDFLD_DSIPLL_P1_MIN_19, .max = MDFLD_DSIPLL_P1_MAX_19}, + }, + { /* MDFLD_LIMT_DSIPLL_25 */ + .dot = {.min = MDFLD_DOT_MIN, .max = MDFLD_DOT_MAX}, + .m = {.min = MDFLD_DSIPLL_M_MIN_25, .max = MDFLD_DSIPLL_M_MAX_25}, + .p1 = {.min = MDFLD_DSIPLL_P1_MIN_25, .max = MDFLD_DSIPLL_P1_MAX_25}, + }, + { /* MDFLD_LIMT_DSIPLL_83 */ + .dot = {.min = MDFLD_DOT_MIN, .max = MDFLD_DOT_MAX}, + .m = {.min = MDFLD_DSIPLL_M_MIN_83, .max = MDFLD_DSIPLL_M_MAX_83}, + .p1 = {.min = MDFLD_DSIPLL_P1_MIN_83, .max = MDFLD_DSIPLL_P1_MAX_83}, + }, + { /* MDFLD_LIMT_DSIPLL_100 */ + .dot = {.min = MDFLD_DOT_MIN, .max = MDFLD_DOT_MAX}, + .m = {.min = MDFLD_DSIPLL_M_MIN_100, .max = MDFLD_DSIPLL_M_MAX_100}, + .p1 = {.min = MDFLD_DSIPLL_P1_MIN_100, .max = MDFLD_DSIPLL_P1_MAX_100}, + }, +}; + +#define MDFLD_M_MIN 21 +#define MDFLD_M_MAX 180 +static const u32 mdfld_m_converts[] = { +/* M configuration table from 9-bit LFSR table */ + 224, 368, 440, 220, 366, 439, 219, 365, 182, 347, /* 21 - 30 */ + 173, 342, 171, 85, 298, 149, 74, 37, 18, 265, /* 31 - 40 */ + 388, 194, 353, 432, 216, 108, 310, 155, 333, 166, /* 41 - 50 */ + 83, 41, 276, 138, 325, 162, 337, 168, 340, 170, /* 51 - 60 */ + 341, 426, 469, 234, 373, 442, 221, 110, 311, 411, /* 61 - 70 */ + 461, 486, 243, 377, 188, 350, 175, 343, 427, 213, /* 71 - 80 */ + 106, 53, 282, 397, 354, 227, 113, 56, 284, 142, /* 81 - 90 */ + 71, 35, 273, 136, 324, 418, 465, 488, 500, 506, /* 91 - 100 */ + 253, 126, 63, 287, 399, 455, 483, 241, 376, 444, /* 101 - 110 */ + 478, 495, 503, 251, 381, 446, 479, 239, 375, 443, /* 111 - 120 */ + 477, 238, 119, 315, 157, 78, 295, 147, 329, 420, /* 121 - 130 */ + 210, 105, 308, 154, 77, 38, 275, 137, 68, 290, /* 131 - 140 */ + 145, 328, 164, 82, 297, 404, 458, 485, 498, 249, /* 141 - 150 */ + 380, 190, 351, 431, 471, 235, 117, 314, 413, 206, /* 151 - 160 */ + 103, 51, 25, 12, 262, 387, 193, 96, 48, 280, /* 161 - 170 */ + 396, 198, 99, 305, 152, 76, 294, 403, 457, 228, /* 171 - 180 */ +}; + +static const struct mrst_limit_t *mdfld_limit(struct drm_crtc *crtc) +{ + const struct mrst_limit_t *limit = NULL; + struct drm_device *dev = crtc->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + + if (psb_intel_pipe_has_type(crtc, INTEL_OUTPUT_MIPI) + || psb_intel_pipe_has_type(crtc, INTEL_OUTPUT_MIPI2)) { + if ((ksel == KSEL_CRYSTAL_19) || (ksel == KSEL_BYPASS_19)) + limit = &mdfld_limits[MDFLD_LIMT_DSIPLL_19]; + else if (ksel == KSEL_BYPASS_25) + limit = &mdfld_limits[MDFLD_LIMT_DSIPLL_25]; + else if ((ksel == KSEL_BYPASS_83_100) && + (dev_priv->core_freq == 166)) + limit = &mdfld_limits[MDFLD_LIMT_DSIPLL_83]; + else if ((ksel == KSEL_BYPASS_83_100) && + (dev_priv->core_freq == 100 || + dev_priv->core_freq == 200)) + limit = &mdfld_limits[MDFLD_LIMT_DSIPLL_100]; + } else if (psb_intel_pipe_has_type(crtc, INTEL_OUTPUT_HDMI)) { + if ((ksel == KSEL_CRYSTAL_19) || (ksel == KSEL_BYPASS_19)) + limit = &mdfld_limits[MDFLD_LIMT_DPLL_19]; + else if (ksel == KSEL_BYPASS_25) + limit = &mdfld_limits[MDFLD_LIMT_DPLL_25]; + else if ((ksel == KSEL_BYPASS_83_100) && + (dev_priv->core_freq == 166)) + limit = &mdfld_limits[MDFLD_LIMT_DPLL_83]; + else if ((ksel == KSEL_BYPASS_83_100) && + (dev_priv->core_freq == 100 || + dev_priv->core_freq == 200)) + limit = &mdfld_limits[MDFLD_LIMT_DPLL_100]; + } else { + limit = NULL; + dev_dbg(dev->dev, "mdfld_limit Wrong display type.\n"); + } + + return limit; +} + +/** Derive the pixel clock for the given refclk and divisors for 8xx chips. */ +static void mdfld_clock(int refclk, struct mrst_clock_t *clock) +{ + clock->dot = (refclk * clock->m) / clock->p1; +} + +/** + * Returns a set of divisors for the desired target clock with the given refclk, + * or FALSE. Divisor values are the actual divisors for + */ +static bool +mdfldFindBestPLL(struct drm_crtc *crtc, int target, int refclk, + struct mrst_clock_t *best_clock) +{ + struct mrst_clock_t clock; + const struct mrst_limit_t *limit = mdfld_limit(crtc); + int err = target; + + memset(best_clock, 0, sizeof(*best_clock)); + + for (clock.m = limit->m.min; clock.m <= limit->m.max; clock.m++) { + for (clock.p1 = limit->p1.min; clock.p1 <= limit->p1.max; + clock.p1++) { + int this_err; + + mdfld_clock(refclk, &clock); + + this_err = abs(clock.dot - target); + if (this_err < err) { + *best_clock = clock; + err = this_err; + } + } + } + return err != target; +} + +static int mdfld_crtc_mode_set(struct drm_crtc *crtc, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode, + int x, int y, + struct drm_framebuffer *old_fb) +{ + struct drm_device *dev = crtc->dev; + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + struct drm_psb_private *dev_priv = dev->dev_private; + int pipe = psb_intel_crtc->pipe; + int fp_reg = MRST_FPA0; + int dpll_reg = MRST_DPLL_A; + int dspcntr_reg = DSPACNTR; + int pipeconf_reg = PIPEACONF; + int htot_reg = HTOTAL_A; + int hblank_reg = HBLANK_A; + int hsync_reg = HSYNC_A; + int vtot_reg = VTOTAL_A; + int vblank_reg = VBLANK_A; + int vsync_reg = VSYNC_A; + int dspsize_reg = DSPASIZE; + int dsppos_reg = DSPAPOS; + int pipesrc_reg = PIPEASRC; + u32 *pipeconf = &dev_priv->pipeconf[pipe]; + u32 *dspcntr = &dev_priv->dspcntr[pipe]; + int refclk = 0; + int clk_n = 0, clk_p2 = 0, clk_byte = 1, clk = 0, m_conv = 0, + clk_tmp = 0; + struct mrst_clock_t clock; + bool ok; + u32 dpll = 0, fp = 0; + bool is_mipi = false, is_mipi2 = false, is_hdmi = false; + struct drm_mode_config *mode_config = &dev->mode_config; + struct psb_intel_encoder *psb_intel_encoder = NULL; + uint64_t scalingType = DRM_MODE_SCALE_FULLSCREEN; + struct drm_encoder *encoder; + struct drm_connector *connector; + int timeout = 0; + int ret; + + dev_dbg(dev->dev, "pipe = 0x%x\n", pipe); + +#if 0 + if (pipe == 1) { + if (!gma_power_begin(dev, true)) + return 0; + android_hdmi_crtc_mode_set(crtc, mode, adjusted_mode, + x, y, old_fb); + goto mrst_crtc_mode_set_exit; + } +#endif + + switch (pipe) { + case 0: + break; + case 1: + fp_reg = FPB0; + dpll_reg = DPLL_B; + dspcntr_reg = DSPBCNTR; + pipeconf_reg = PIPEBCONF; + htot_reg = HTOTAL_B; + hblank_reg = HBLANK_B; + hsync_reg = HSYNC_B; + vtot_reg = VTOTAL_B; + vblank_reg = VBLANK_B; + vsync_reg = VSYNC_B; + dspsize_reg = DSPBSIZE; + dsppos_reg = DSPBPOS; + pipesrc_reg = PIPEBSRC; + fp_reg = MDFLD_DPLL_DIV0; + dpll_reg = MDFLD_DPLL_B; + break; + case 2: + dpll_reg = MRST_DPLL_A; + dspcntr_reg = DSPCCNTR; + pipeconf_reg = PIPECCONF; + htot_reg = HTOTAL_C; + hblank_reg = HBLANK_C; + hsync_reg = HSYNC_C; + vtot_reg = VTOTAL_C; + vblank_reg = VBLANK_C; + vsync_reg = VSYNC_C; + dspsize_reg = DSPCSIZE; + dsppos_reg = DSPCPOS; + pipesrc_reg = PIPECSRC; + break; + default: + DRM_ERROR("Illegal Pipe Number.\n"); + return 0; + } + + ret = check_fb(crtc->fb); + if (ret) + return ret; + + dev_dbg(dev->dev, "adjusted_hdisplay = %d\n", + adjusted_mode->hdisplay); + dev_dbg(dev->dev, "adjusted_vdisplay = %d\n", + adjusted_mode->vdisplay); + dev_dbg(dev->dev, "adjusted_hsync_start = %d\n", + adjusted_mode->hsync_start); + dev_dbg(dev->dev, "adjusted_hsync_end = %d\n", + adjusted_mode->hsync_end); + dev_dbg(dev->dev, "adjusted_htotal = %d\n", + adjusted_mode->htotal); + dev_dbg(dev->dev, "adjusted_vsync_start = %d\n", + adjusted_mode->vsync_start); + dev_dbg(dev->dev, "adjusted_vsync_end = %d\n", + adjusted_mode->vsync_end); + dev_dbg(dev->dev, "adjusted_vtotal = %d\n", + adjusted_mode->vtotal); + dev_dbg(dev->dev, "adjusted_clock = %d\n", + adjusted_mode->clock); + dev_dbg(dev->dev, "hdisplay = %d\n", + mode->hdisplay); + dev_dbg(dev->dev, "vdisplay = %d\n", + mode->vdisplay); + + if (!gma_power_begin(dev, true)) + return 0; + + memcpy(&psb_intel_crtc->saved_mode, mode, + sizeof(struct drm_display_mode)); + memcpy(&psb_intel_crtc->saved_adjusted_mode, adjusted_mode, + sizeof(struct drm_display_mode)); + + list_for_each_entry(connector, &mode_config->connector_list, head) { + if (!connector) + continue; + + encoder = connector->encoder; + + if (!encoder) + continue; + + if (encoder->crtc != crtc) + continue; + + psb_intel_encoder = psb_intel_attached_encoder(connector); + + switch (psb_intel_encoder->type) { + case INTEL_OUTPUT_MIPI: + is_mipi = true; + break; + case INTEL_OUTPUT_MIPI2: + is_mipi2 = true; + break; + case INTEL_OUTPUT_HDMI: + is_hdmi = true; + break; + } + } + + /* Disable the VGA plane that we never use */ + REG_WRITE(VGACNTRL, VGA_DISP_DISABLE); + + /* Disable the panel fitter if it was on our pipe */ + if (psb_intel_panel_fitter_pipe(dev) == pipe) + REG_WRITE(PFIT_CONTROL, 0); + + /* pipesrc and dspsize control the size that is scaled from, + * which should always be the user's requested size. + */ + if (pipe == 1) { + /* FIXME: To make HDMI display with 864x480 (TPO), 480x864 + * (PYR) or 480x854 (TMD), set the sprite width/height and + * souce image size registers with the adjusted mode for + * pipe B. + */ + + /* + * The defined sprite rectangle must always be completely + * contained within the displayable area of the screen image + * (frame buffer). + */ + REG_WRITE(dspsize_reg, ((min(mode->crtc_vdisplay, adjusted_mode->crtc_vdisplay) - 1) << 16) + | (min(mode->crtc_hdisplay, adjusted_mode->crtc_hdisplay) - 1)); + /* Set the CRTC with encoder mode. */ + REG_WRITE(pipesrc_reg, ((mode->crtc_hdisplay - 1) << 16) + | (mode->crtc_vdisplay - 1)); + } else { + REG_WRITE(dspsize_reg, + ((mode->crtc_vdisplay - 1) << 16) | + (mode->crtc_hdisplay - 1)); + REG_WRITE(pipesrc_reg, + ((mode->crtc_hdisplay - 1) << 16) | + (mode->crtc_vdisplay - 1)); + } + + REG_WRITE(dsppos_reg, 0); + + if (psb_intel_encoder) + drm_connector_property_get_value(connector, + dev->mode_config.scaling_mode_property, &scalingType); + + if (scalingType == DRM_MODE_SCALE_NO_SCALE) { + /* Medfield doesn't have register support for centering so we + * need to mess with the h/vblank and h/vsync start and ends + * to get centering + */ + int offsetX = 0, offsetY = 0; + + offsetX = (adjusted_mode->crtc_hdisplay - + mode->crtc_hdisplay) / 2; + offsetY = (adjusted_mode->crtc_vdisplay - + mode->crtc_vdisplay) / 2; + + REG_WRITE(htot_reg, (mode->crtc_hdisplay - 1) | + ((adjusted_mode->crtc_htotal - 1) << 16)); + REG_WRITE(vtot_reg, (mode->crtc_vdisplay - 1) | + ((adjusted_mode->crtc_vtotal - 1) << 16)); + REG_WRITE(hblank_reg, (adjusted_mode->crtc_hblank_start - + offsetX - 1) | + ((adjusted_mode->crtc_hblank_end - offsetX - 1) << 16)); + REG_WRITE(hsync_reg, (adjusted_mode->crtc_hsync_start - + offsetX - 1) | + ((adjusted_mode->crtc_hsync_end - offsetX - 1) << 16)); + REG_WRITE(vblank_reg, (adjusted_mode->crtc_vblank_start - + offsetY - 1) | + ((adjusted_mode->crtc_vblank_end - offsetY - 1) << 16)); + REG_WRITE(vsync_reg, (adjusted_mode->crtc_vsync_start - + offsetY - 1) | + ((adjusted_mode->crtc_vsync_end - offsetY - 1) << 16)); + } else { + REG_WRITE(htot_reg, (adjusted_mode->crtc_hdisplay - 1) | + ((adjusted_mode->crtc_htotal - 1) << 16)); + REG_WRITE(vtot_reg, (adjusted_mode->crtc_vdisplay - 1) | + ((adjusted_mode->crtc_vtotal - 1) << 16)); + REG_WRITE(hblank_reg, (adjusted_mode->crtc_hblank_start - 1) | + ((adjusted_mode->crtc_hblank_end - 1) << 16)); + REG_WRITE(hsync_reg, (adjusted_mode->crtc_hsync_start - 1) | + ((adjusted_mode->crtc_hsync_end - 1) << 16)); + REG_WRITE(vblank_reg, (adjusted_mode->crtc_vblank_start - 1) | + ((adjusted_mode->crtc_vblank_end - 1) << 16)); + REG_WRITE(vsync_reg, (adjusted_mode->crtc_vsync_start - 1) | + ((adjusted_mode->crtc_vsync_end - 1) << 16)); + } + + /* Flush the plane changes */ + { + struct drm_crtc_helper_funcs *crtc_funcs = + crtc->helper_private; + crtc_funcs->mode_set_base(crtc, x, y, old_fb); + } + + /* setup pipeconf */ + *pipeconf = PIPEACONF_ENABLE; /* FIXME_JLIU7 REG_READ(pipeconf_reg); */ + + /* Set up the display plane register */ + *dspcntr = REG_READ(dspcntr_reg); + *dspcntr |= pipe << DISPPLANE_SEL_PIPE_POS; + *dspcntr |= DISPLAY_PLANE_ENABLE; + + if (is_mipi2) + goto mrst_crtc_mode_set_exit; + clk = adjusted_mode->clock; + + if (is_hdmi) { + if ((ksel == KSEL_CRYSTAL_19) || (ksel == KSEL_BYPASS_19)) { + refclk = 19200; + + if (is_mipi || is_mipi2) + clk_n = 1, clk_p2 = 8; + else if (is_hdmi) + clk_n = 1, clk_p2 = 10; + } else if (ksel == KSEL_BYPASS_25) { + refclk = 25000; + + if (is_mipi || is_mipi2) + clk_n = 1, clk_p2 = 8; + else if (is_hdmi) + clk_n = 1, clk_p2 = 10; + } else if ((ksel == KSEL_BYPASS_83_100) && + dev_priv->core_freq == 166) { + refclk = 83000; + + if (is_mipi || is_mipi2) + clk_n = 4, clk_p2 = 8; + else if (is_hdmi) + clk_n = 4, clk_p2 = 10; + } else if ((ksel == KSEL_BYPASS_83_100) && + (dev_priv->core_freq == 100 || + dev_priv->core_freq == 200)) { + refclk = 100000; + if (is_mipi || is_mipi2) + clk_n = 4, clk_p2 = 8; + else if (is_hdmi) + clk_n = 4, clk_p2 = 10; + } + + if (is_mipi) + clk_byte = dev_priv->bpp / 8; + else if (is_mipi2) + clk_byte = dev_priv->bpp2 / 8; + + clk_tmp = clk * clk_n * clk_p2 * clk_byte; + + dev_dbg(dev->dev, "clk = %d, clk_n = %d, clk_p2 = %d.\n", + clk, clk_n, clk_p2); + dev_dbg(dev->dev, "adjusted_mode->clock = %d, clk_tmp = %d.\n", + adjusted_mode->clock, clk_tmp); + + ok = mdfldFindBestPLL(crtc, clk_tmp, refclk, &clock); + + if (!ok) { + DRM_ERROR + ("mdfldFindBestPLL fail in mdfld_crtc_mode_set.\n"); + } else { + m_conv = mdfld_m_converts[(clock.m - MDFLD_M_MIN)]; + + dev_dbg(dev->dev, "dot clock = %d," + "m = %d, p1 = %d, m_conv = %d.\n", + clock.dot, clock.m, + clock.p1, m_conv); + } + + dpll = REG_READ(dpll_reg); + + if (dpll & DPLL_VCO_ENABLE) { + dpll &= ~DPLL_VCO_ENABLE; + REG_WRITE(dpll_reg, dpll); + REG_READ(dpll_reg); + + /* FIXME jliu7 check the DPLL lock bit PIPEACONF[29] */ + /* FIXME_MDFLD PO - change 500 to 1 after PO */ + udelay(500); + + /* reset M1, N1 & P1 */ + REG_WRITE(fp_reg, 0); + dpll &= ~MDFLD_P1_MASK; + REG_WRITE(dpll_reg, dpll); + /* FIXME_MDFLD PO - change 500 to 1 after PO */ + udelay(500); + } + + /* When ungating power of DPLL, needs to wait 0.5us before + * enable the VCO */ + if (dpll & MDFLD_PWR_GATE_EN) { + dpll &= ~MDFLD_PWR_GATE_EN; + REG_WRITE(dpll_reg, dpll); + /* FIXME_MDFLD PO - change 500 to 1 after PO */ + udelay(500); + } + dpll = 0; + +#if 0 /* FIXME revisit later */ + if (ksel == KSEL_CRYSTAL_19 || ksel == KSEL_BYPASS_19 || + ksel == KSEL_BYPASS_25) + dpll &= ~MDFLD_INPUT_REF_SEL; + else if (ksel == KSEL_BYPASS_83_100) + dpll |= MDFLD_INPUT_REF_SEL; +#endif /* FIXME revisit later */ + + if (is_hdmi) + dpll |= MDFLD_VCO_SEL; + + fp = (clk_n / 2) << 16; + fp |= m_conv; + + /* compute bitmask from p1 value */ + dpll |= (1 << (clock.p1 - 2)) << 17; + +#if 0 /* 1080p30 & 720p */ + dpll = 0x00050000; + fp = 0x000001be; +#endif +#if 0 /* 480p */ + dpll = 0x02010000; + fp = 0x000000d2; +#endif + } else { +#if 0 /*DBI_TPO_480x864*/ + dpll = 0x00020000; + fp = 0x00000156; +#endif /* DBI_TPO_480x864 */ /* get from spec. */ + + dpll = 0x00800000; + fp = 0x000000c1; + } + + REG_WRITE(fp_reg, fp); + REG_WRITE(dpll_reg, dpll); + /* FIXME_MDFLD PO - change 500 to 1 after PO */ + udelay(500); + + dpll |= DPLL_VCO_ENABLE; + REG_WRITE(dpll_reg, dpll); + REG_READ(dpll_reg); + + /* wait for DSI PLL to lock */ + while (timeout < 20000 && + !(REG_READ(pipeconf_reg) & PIPECONF_DSIPLL_LOCK)) { + udelay(150); + timeout++; + } + + if (is_mipi) + goto mrst_crtc_mode_set_exit; + + dev_dbg(dev->dev, "is_mipi = 0x%x\n", is_mipi); + + REG_WRITE(pipeconf_reg, *pipeconf); + REG_READ(pipeconf_reg); + + /* Wait for for the pipe enable to take effect. */ + REG_WRITE(dspcntr_reg, *dspcntr); + psb_intel_wait_for_vblank(dev); + +mrst_crtc_mode_set_exit: + + gma_power_end(dev); + + return 0; +} + +const struct drm_crtc_helper_funcs mdfld_helper_funcs = { + .dpms = mdfld_crtc_dpms, + .mode_fixup = psb_intel_crtc_mode_fixup, + .mode_set = mdfld_crtc_mode_set, + .mode_set_base = mdfld__intel_pipe_set_base, + .prepare = psb_intel_crtc_prepare, + .commit = psb_intel_crtc_commit, +}; + diff --git a/drivers/gpu/drm/gma500/mdfld_output.c b/drivers/gpu/drm/gma500/mdfld_output.c new file mode 100644 index 0000000..c95966b --- /dev/null +++ b/drivers/gpu/drm/gma500/mdfld_output.c @@ -0,0 +1,74 @@ +/* + * Copyright (c) 2010 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicensen + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * Thomas Eaton + * Scott Rowe +*/ + +#include "mdfld_output.h" +#include "mdfld_dsi_dpi.h" +#include "mdfld_dsi_output.h" + +#include "tc35876x-dsi-lvds.h" + +int mdfld_get_panel_type(struct drm_device *dev, int pipe) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + return dev_priv->mdfld_panel_id; +} + +static void mdfld_init_panel(struct drm_device *dev, int mipi_pipe, + int p_type) +{ + switch (p_type) { + case TPO_VID: + mdfld_dsi_output_init(dev, mipi_pipe, &mdfld_tpo_vid_funcs); + break; + case TC35876X: + tc35876x_init(dev); + mdfld_dsi_output_init(dev, mipi_pipe, &mdfld_tc35876x_funcs); + break; + case TMD_VID: + mdfld_dsi_output_init(dev, mipi_pipe, &mdfld_tmd_vid_funcs); + break; + case HDMI: +/* if (dev_priv->mdfld_hdmi_present) + mdfld_hdmi_init(dev, &dev_priv->mode_dev); */ + break; + } +} + + +int mdfld_output_init(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + + /* FIXME: hardcoded for now */ + dev_priv->mdfld_panel_id = TC35876X; + /* MIPI panel 1 */ + mdfld_init_panel(dev, 0, dev_priv->mdfld_panel_id); + /* HDMI panel */ + mdfld_init_panel(dev, 1, HDMI); + return 0; +} + diff --git a/drivers/gpu/drm/gma500/mdfld_output.h b/drivers/gpu/drm/gma500/mdfld_output.h new file mode 100644 index 0000000..ab2b27c --- /dev/null +++ b/drivers/gpu/drm/gma500/mdfld_output.h @@ -0,0 +1,77 @@ +/* + * Copyright (c) 2010 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicensen + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * Thomas Eaton + * Scott Rowe +*/ + +#ifndef MDFLD_OUTPUT_H +#define MDFLD_OUTPUT_H + +#include "psb_drv.h" + +#define TPO_PANEL_WIDTH 84 +#define TPO_PANEL_HEIGHT 46 +#define TMD_PANEL_WIDTH 39 +#define TMD_PANEL_HEIGHT 71 + +struct mdfld_dsi_config; + +enum panel_type { + TPO_VID, + TMD_VID, + HDMI, + TC35876X, +}; + +struct panel_info { + u32 width_mm; + u32 height_mm; + /* Other info */ +}; + +struct panel_funcs { + const struct drm_encoder_funcs *encoder_funcs; + const struct drm_encoder_helper_funcs *encoder_helper_funcs; + struct drm_display_mode * (*get_config_mode)(struct drm_device *); + int (*get_panel_info)(struct drm_device *, int, struct panel_info *); + int (*reset)(int pipe); + void (*drv_ic_init)(struct mdfld_dsi_config *dsi_config, int pipe); +}; + +int mdfld_output_init(struct drm_device *dev); + +struct backlight_device *mdfld_get_backlight_device(void); +int mdfld_set_brightness(struct backlight_device *bd); + +int mdfld_get_panel_type(struct drm_device *dev, int pipe); + +extern const struct drm_crtc_helper_funcs mdfld_helper_funcs; + +extern const struct panel_funcs mdfld_tmd_vid_funcs; +extern const struct panel_funcs mdfld_tpo_vid_funcs; + +extern void mdfld_disable_crtc(struct drm_device *dev, int pipe); +extern void mdfldWaitForPipeEnable(struct drm_device *dev, int pipe); +extern void mdfldWaitForPipeDisable(struct drm_device *dev, int pipe); +#endif diff --git a/drivers/gpu/drm/gma500/mdfld_tmd_vid.c b/drivers/gpu/drm/gma500/mdfld_tmd_vid.c new file mode 100644 index 0000000..dc0c6c3 --- /dev/null +++ b/drivers/gpu/drm/gma500/mdfld_tmd_vid.c @@ -0,0 +1,201 @@ +/* + * Copyright © 2010 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * Jim Liu + * Jackie Li + * Gideon Eaton + */ + +#include "mdfld_dsi_dpi.h" +#include "mdfld_dsi_pkg_sender.h" + +static struct drm_display_mode *tmd_vid_get_config_mode(struct drm_device *dev) +{ + struct drm_display_mode *mode; + struct drm_psb_private *dev_priv = dev->dev_private; + struct oaktrail_timing_info *ti = &dev_priv->gct_data.DTD; + bool use_gct = false; /*Disable GCT for now*/ + + mode = kzalloc(sizeof(*mode), GFP_KERNEL); + if (!mode) + return NULL; + + if (use_gct) { + mode->hdisplay = (ti->hactive_hi << 8) | ti->hactive_lo; + mode->vdisplay = (ti->vactive_hi << 8) | ti->vactive_lo; + mode->hsync_start = mode->hdisplay + \ + ((ti->hsync_offset_hi << 8) | \ + ti->hsync_offset_lo); + mode->hsync_end = mode->hsync_start + \ + ((ti->hsync_pulse_width_hi << 8) | \ + ti->hsync_pulse_width_lo); + mode->htotal = mode->hdisplay + ((ti->hblank_hi << 8) | \ + ti->hblank_lo); + mode->vsync_start = \ + mode->vdisplay + ((ti->vsync_offset_hi << 8) | \ + ti->vsync_offset_lo); + mode->vsync_end = \ + mode->vsync_start + ((ti->vsync_pulse_width_hi << 8) | \ + ti->vsync_pulse_width_lo); + mode->vtotal = mode->vdisplay + \ + ((ti->vblank_hi << 8) | ti->vblank_lo); + mode->clock = ti->pixel_clock * 10; + + dev_dbg(dev->dev, "hdisplay is %d\n", mode->hdisplay); + dev_dbg(dev->dev, "vdisplay is %d\n", mode->vdisplay); + dev_dbg(dev->dev, "HSS is %d\n", mode->hsync_start); + dev_dbg(dev->dev, "HSE is %d\n", mode->hsync_end); + dev_dbg(dev->dev, "htotal is %d\n", mode->htotal); + dev_dbg(dev->dev, "VSS is %d\n", mode->vsync_start); + dev_dbg(dev->dev, "VSE is %d\n", mode->vsync_end); + dev_dbg(dev->dev, "vtotal is %d\n", mode->vtotal); + dev_dbg(dev->dev, "clock is %d\n", mode->clock); + } else { + mode->hdisplay = 480; + mode->vdisplay = 854; + mode->hsync_start = 487; + mode->hsync_end = 490; + mode->htotal = 499; + mode->vsync_start = 861; + mode->vsync_end = 865; + mode->vtotal = 873; + mode->clock = 33264; + } + + drm_mode_set_name(mode); + drm_mode_set_crtcinfo(mode, 0); + + mode->type |= DRM_MODE_TYPE_PREFERRED; + + return mode; +} + +static int tmd_vid_get_panel_info(struct drm_device *dev, + int pipe, + struct panel_info *pi) +{ + if (!dev || !pi) + return -EINVAL; + + pi->width_mm = TMD_PANEL_WIDTH; + pi->height_mm = TMD_PANEL_HEIGHT; + + return 0; +} + +/* ************************************************************************* *\ + * FUNCTION: mdfld_init_TMD_MIPI + * + * DESCRIPTION: This function is called only by mrst_dsi_mode_set and + * restore_display_registers. since this function does not + * acquire the mutex, it is important that the calling function + * does! +\* ************************************************************************* */ + +/* FIXME: make the below data u8 instead of u32; note byte order! */ +static u32 tmd_cmd_mcap_off[] = {0x000000b2}; +static u32 tmd_cmd_enable_lane_switch[] = {0x000101ef}; +static u32 tmd_cmd_set_lane_num[] = {0x006360ef}; +static u32 tmd_cmd_pushing_clock0[] = {0x00cc2fef}; +static u32 tmd_cmd_pushing_clock1[] = {0x00dd6eef}; +static u32 tmd_cmd_set_mode[] = {0x000000b3}; +static u32 tmd_cmd_set_sync_pulse_mode[] = {0x000961ef}; +static u32 tmd_cmd_set_column[] = {0x0100002a, 0x000000df}; +static u32 tmd_cmd_set_page[] = {0x0300002b, 0x00000055}; +static u32 tmd_cmd_set_video_mode[] = {0x00000153}; +/*no auto_bl,need add in furture*/ +static u32 tmd_cmd_enable_backlight[] = {0x00005ab4}; +static u32 tmd_cmd_set_backlight_dimming[] = {0x00000ebd}; + +static void mdfld_dsi_tmd_drv_ic_init(struct mdfld_dsi_config *dsi_config, + int pipe) +{ + struct mdfld_dsi_pkg_sender *sender + = mdfld_dsi_get_pkg_sender(dsi_config); + + DRM_INFO("Enter mdfld init TMD MIPI display.\n"); + + if (!sender) { + DRM_ERROR("Cannot get sender\n"); + return; + } + + if (dsi_config->dvr_ic_inited) + return; + + msleep(3); + + /* FIXME: make the below data u8 instead of u32; note byte order! */ + + mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_mcap_off, + sizeof(tmd_cmd_mcap_off), false); + mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_enable_lane_switch, + sizeof(tmd_cmd_enable_lane_switch), false); + mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_set_lane_num, + sizeof(tmd_cmd_set_lane_num), false); + mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_pushing_clock0, + sizeof(tmd_cmd_pushing_clock0), false); + mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_pushing_clock1, + sizeof(tmd_cmd_pushing_clock1), false); + mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_set_mode, + sizeof(tmd_cmd_set_mode), false); + mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_set_sync_pulse_mode, + sizeof(tmd_cmd_set_sync_pulse_mode), false); + mdfld_dsi_send_mcs_long(sender, (u8 *) tmd_cmd_set_column, + sizeof(tmd_cmd_set_column), false); + mdfld_dsi_send_mcs_long(sender, (u8 *) tmd_cmd_set_page, + sizeof(tmd_cmd_set_page), false); + mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_set_video_mode, + sizeof(tmd_cmd_set_video_mode), false); + mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_enable_backlight, + sizeof(tmd_cmd_enable_backlight), false); + mdfld_dsi_send_gen_long(sender, (u8 *) tmd_cmd_set_backlight_dimming, + sizeof(tmd_cmd_set_backlight_dimming), false); + + dsi_config->dvr_ic_inited = 1; +} + +/*TPO DPI encoder helper funcs*/ +static const struct drm_encoder_helper_funcs + mdfld_tpo_dpi_encoder_helper_funcs = { + .dpms = mdfld_dsi_dpi_dpms, + .mode_fixup = mdfld_dsi_dpi_mode_fixup, + .prepare = mdfld_dsi_dpi_prepare, + .mode_set = mdfld_dsi_dpi_mode_set, + .commit = mdfld_dsi_dpi_commit, +}; + +/*TPO DPI encoder funcs*/ +static const struct drm_encoder_funcs mdfld_tpo_dpi_encoder_funcs = { + .destroy = drm_encoder_cleanup, +}; + +const struct panel_funcs mdfld_tmd_vid_funcs = { + .encoder_funcs = &mdfld_tpo_dpi_encoder_funcs, + .encoder_helper_funcs = &mdfld_tpo_dpi_encoder_helper_funcs, + .get_config_mode = &tmd_vid_get_config_mode, + .get_panel_info = tmd_vid_get_panel_info, + .reset = mdfld_dsi_panel_reset, + .drv_ic_init = mdfld_dsi_tmd_drv_ic_init, +}; diff --git a/drivers/gpu/drm/gma500/mdfld_tpo_vid.c b/drivers/gpu/drm/gma500/mdfld_tpo_vid.c new file mode 100644 index 0000000..d8d4170 --- /dev/null +++ b/drivers/gpu/drm/gma500/mdfld_tpo_vid.c @@ -0,0 +1,124 @@ +/* + * Copyright © 2010 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * jim liu + * Jackie Li + */ + +#include "mdfld_dsi_dpi.h" + +static struct drm_display_mode *tpo_vid_get_config_mode(struct drm_device *dev) +{ + struct drm_display_mode *mode; + struct drm_psb_private *dev_priv = dev->dev_private; + struct oaktrail_timing_info *ti = &dev_priv->gct_data.DTD; + bool use_gct = false; + + mode = kzalloc(sizeof(*mode), GFP_KERNEL); + if (!mode) + return NULL; + + if (use_gct) { + mode->hdisplay = (ti->hactive_hi << 8) | ti->hactive_lo; + mode->vdisplay = (ti->vactive_hi << 8) | ti->vactive_lo; + mode->hsync_start = mode->hdisplay + + ((ti->hsync_offset_hi << 8) | + ti->hsync_offset_lo); + mode->hsync_end = mode->hsync_start + + ((ti->hsync_pulse_width_hi << 8) | + ti->hsync_pulse_width_lo); + mode->htotal = mode->hdisplay + ((ti->hblank_hi << 8) | + ti->hblank_lo); + mode->vsync_start = + mode->vdisplay + ((ti->vsync_offset_hi << 8) | + ti->vsync_offset_lo); + mode->vsync_end = + mode->vsync_start + ((ti->vsync_pulse_width_hi << 8) | + ti->vsync_pulse_width_lo); + mode->vtotal = mode->vdisplay + + ((ti->vblank_hi << 8) | ti->vblank_lo); + mode->clock = ti->pixel_clock * 10; + + dev_dbg(dev->dev, "hdisplay is %d\n", mode->hdisplay); + dev_dbg(dev->dev, "vdisplay is %d\n", mode->vdisplay); + dev_dbg(dev->dev, "HSS is %d\n", mode->hsync_start); + dev_dbg(dev->dev, "HSE is %d\n", mode->hsync_end); + dev_dbg(dev->dev, "htotal is %d\n", mode->htotal); + dev_dbg(dev->dev, "VSS is %d\n", mode->vsync_start); + dev_dbg(dev->dev, "VSE is %d\n", mode->vsync_end); + dev_dbg(dev->dev, "vtotal is %d\n", mode->vtotal); + dev_dbg(dev->dev, "clock is %d\n", mode->clock); + } else { + mode->hdisplay = 864; + mode->vdisplay = 480; + mode->hsync_start = 873; + mode->hsync_end = 876; + mode->htotal = 887; + mode->vsync_start = 487; + mode->vsync_end = 490; + mode->vtotal = 499; + mode->clock = 33264; + } + + drm_mode_set_name(mode); + drm_mode_set_crtcinfo(mode, 0); + + mode->type |= DRM_MODE_TYPE_PREFERRED; + + return mode; +} + +static int tpo_vid_get_panel_info(struct drm_device *dev, + int pipe, + struct panel_info *pi) +{ + if (!dev || !pi) + return -EINVAL; + + pi->width_mm = TPO_PANEL_WIDTH; + pi->height_mm = TPO_PANEL_HEIGHT; + + return 0; +} + +/*TPO DPI encoder helper funcs*/ +static const struct drm_encoder_helper_funcs + mdfld_tpo_dpi_encoder_helper_funcs = { + .dpms = mdfld_dsi_dpi_dpms, + .mode_fixup = mdfld_dsi_dpi_mode_fixup, + .prepare = mdfld_dsi_dpi_prepare, + .mode_set = mdfld_dsi_dpi_mode_set, + .commit = mdfld_dsi_dpi_commit, +}; + +/*TPO DPI encoder funcs*/ +static const struct drm_encoder_funcs mdfld_tpo_dpi_encoder_funcs = { + .destroy = drm_encoder_cleanup, +}; + +const struct panel_funcs mdfld_tpo_vid_funcs = { + .encoder_funcs = &mdfld_tpo_dpi_encoder_funcs, + .encoder_helper_funcs = &mdfld_tpo_dpi_encoder_helper_funcs, + .get_config_mode = &tpo_vid_get_config_mode, + .get_panel_info = tpo_vid_get_panel_info, +}; diff --git a/drivers/gpu/drm/gma500/mid_bios.c b/drivers/gpu/drm/gma500/mid_bios.c new file mode 100644 index 0000000..5eee9ad --- /dev/null +++ b/drivers/gpu/drm/gma500/mid_bios.c @@ -0,0 +1,263 @@ +/************************************************************************** + * Copyright (c) 2011, Intel Corporation. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + **************************************************************************/ + +/* TODO + * - Split functions by vbt type + * - Make them all take drm_device + * - Check ioremap failures + */ + +#include +#include +#include "gma_drm.h" +#include "psb_drv.h" +#include "mid_bios.h" + +static void mid_get_fuse_settings(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct pci_dev *pci_root = pci_get_bus_and_slot(0, 0); + uint32_t fuse_value = 0; + uint32_t fuse_value_tmp = 0; + +#define FB_REG06 0xD0810600 +#define FB_MIPI_DISABLE (1 << 11) +#define FB_REG09 0xD0810900 +#define FB_REG09 0xD0810900 +#define FB_SKU_MASK 0x7000 +#define FB_SKU_SHIFT 12 +#define FB_SKU_100 0 +#define FB_SKU_100L 1 +#define FB_SKU_83 2 + if (pci_root == NULL) { + WARN_ON(1); + return; + } + + + pci_write_config_dword(pci_root, 0xD0, FB_REG06); + pci_read_config_dword(pci_root, 0xD4, &fuse_value); + + /* FB_MIPI_DISABLE doesn't mean LVDS on with Medfield */ + if (IS_MRST(dev)) + dev_priv->iLVDS_enable = fuse_value & FB_MIPI_DISABLE; + + DRM_INFO("internal display is %s\n", + dev_priv->iLVDS_enable ? "LVDS display" : "MIPI display"); + + /* Prevent runtime suspend at start*/ + if (dev_priv->iLVDS_enable) { + dev_priv->is_lvds_on = true; + dev_priv->is_mipi_on = false; + } else { + dev_priv->is_mipi_on = true; + dev_priv->is_lvds_on = false; + } + + dev_priv->video_device_fuse = fuse_value; + + pci_write_config_dword(pci_root, 0xD0, FB_REG09); + pci_read_config_dword(pci_root, 0xD4, &fuse_value); + + dev_dbg(dev->dev, "SKU values is 0x%x.\n", fuse_value); + fuse_value_tmp = (fuse_value & FB_SKU_MASK) >> FB_SKU_SHIFT; + + dev_priv->fuse_reg_value = fuse_value; + + switch (fuse_value_tmp) { + case FB_SKU_100: + dev_priv->core_freq = 200; + break; + case FB_SKU_100L: + dev_priv->core_freq = 100; + break; + case FB_SKU_83: + dev_priv->core_freq = 166; + break; + default: + dev_warn(dev->dev, "Invalid SKU values, SKU value = 0x%08x\n", + fuse_value_tmp); + dev_priv->core_freq = 0; + } + dev_dbg(dev->dev, "LNC core clk is %dMHz.\n", dev_priv->core_freq); + pci_dev_put(pci_root); +} + +/* + * Get the revison ID, B0:D2:F0;0x08 + */ +static void mid_get_pci_revID(struct drm_psb_private *dev_priv) +{ + uint32_t platform_rev_id = 0; + struct pci_dev *pci_gfx_root = pci_get_bus_and_slot(0, PCI_DEVFN(2, 0)); + + if (pci_gfx_root == NULL) { + WARN_ON(1); + return; + } + pci_read_config_dword(pci_gfx_root, 0x08, &platform_rev_id); + dev_priv->platform_rev_id = (uint8_t) platform_rev_id; + pci_dev_put(pci_gfx_root); + dev_dbg(dev_priv->dev->dev, "platform_rev_id is %x\n", + dev_priv->platform_rev_id); +} + +static void mid_get_vbt_data(struct drm_psb_private *dev_priv) +{ + struct drm_device *dev = dev_priv->dev; + struct oaktrail_vbt *vbt = &dev_priv->vbt_data; + u32 addr; + u16 new_size; + u8 *vbt_virtual; + u8 bpi; + u8 number_desc = 0; + struct oaktrail_timing_info *dp_ti = &dev_priv->gct_data.DTD; + struct gct_r10_timing_info ti; + void *pGCT; + struct pci_dev *pci_gfx_root = pci_get_bus_and_slot(0, PCI_DEVFN(2, 0)); + + /* Get the address of the platform config vbt, B0:D2:F0;0xFC */ + pci_read_config_dword(pci_gfx_root, 0xFC, &addr); + pci_dev_put(pci_gfx_root); + + dev_dbg(dev->dev, "drm platform config address is %x\n", addr); + + /* check for platform config address == 0. */ + /* this means fw doesn't support vbt */ + + if (addr == 0) { + vbt->size = 0; + return; + } + + /* get the virtual address of the vbt */ + vbt_virtual = ioremap(addr, sizeof(*vbt)); + if (vbt_virtual == NULL) { + vbt->size = 0; + return; + } + + memcpy(vbt, vbt_virtual, sizeof(*vbt)); + iounmap(vbt_virtual); /* Free virtual address space */ + + /* No matching signature don't process the data */ + if (memcmp(vbt->signature, "$GCT", 4)) { + vbt->size = 0; + return; + } + + dev_dbg(dev->dev, "GCT revision is %x\n", vbt->revision); + + switch (vbt->revision) { + case 0: + vbt->oaktrail_gct = ioremap(addr + sizeof(*vbt) - 4, + vbt->size - sizeof(*vbt) + 4); + pGCT = vbt->oaktrail_gct; + bpi = ((struct oaktrail_gct_v1 *)pGCT)->PD.BootPanelIndex; + dev_priv->gct_data.bpi = bpi; + dev_priv->gct_data.pt = + ((struct oaktrail_gct_v1 *)pGCT)->PD.PanelType; + memcpy(&dev_priv->gct_data.DTD, + &((struct oaktrail_gct_v1 *)pGCT)->panel[bpi].DTD, + sizeof(struct oaktrail_timing_info)); + dev_priv->gct_data.Panel_Port_Control = + ((struct oaktrail_gct_v1 *)pGCT)->panel[bpi].Panel_Port_Control; + dev_priv->gct_data.Panel_MIPI_Display_Descriptor = + ((struct oaktrail_gct_v1 *)pGCT)->panel[bpi].Panel_MIPI_Display_Descriptor; + break; + case 1: + vbt->oaktrail_gct = ioremap(addr + sizeof(*vbt) - 4, + vbt->size - sizeof(*vbt) + 4); + pGCT = vbt->oaktrail_gct; + bpi = ((struct oaktrail_gct_v2 *)pGCT)->PD.BootPanelIndex; + dev_priv->gct_data.bpi = bpi; + dev_priv->gct_data.pt = + ((struct oaktrail_gct_v2 *)pGCT)->PD.PanelType; + memcpy(&dev_priv->gct_data.DTD, + &((struct oaktrail_gct_v2 *)pGCT)->panel[bpi].DTD, + sizeof(struct oaktrail_timing_info)); + dev_priv->gct_data.Panel_Port_Control = + ((struct oaktrail_gct_v2 *)pGCT)->panel[bpi].Panel_Port_Control; + dev_priv->gct_data.Panel_MIPI_Display_Descriptor = + ((struct oaktrail_gct_v2 *)pGCT)->panel[bpi].Panel_MIPI_Display_Descriptor; + break; + case 0x10: + /*header definition changed from rev 01 (v2) to rev 10h. */ + /*so, some values have changed location*/ + new_size = vbt->checksum; /*checksum contains lo size byte*/ + /*LSB of oaktrail_gct contains hi size byte*/ + new_size |= ((0xff & (unsigned int)(long)vbt->oaktrail_gct)) << 8; + + vbt->checksum = vbt->size; /*size contains the checksum*/ + if (new_size > 0xff) + vbt->size = 0xff; /*restrict size to 255*/ + else + vbt->size = new_size; + + /* number of descriptors defined in the GCT */ + number_desc = ((0xff00 & (unsigned int)(long)vbt->oaktrail_gct)) >> 8; + bpi = ((0xff0000 & (unsigned int)(long)vbt->oaktrail_gct)) >> 16; + vbt->oaktrail_gct = ioremap(addr + GCT_R10_HEADER_SIZE, + GCT_R10_DISPLAY_DESC_SIZE * number_desc); + pGCT = vbt->oaktrail_gct; + pGCT = (u8 *)pGCT + (bpi*GCT_R10_DISPLAY_DESC_SIZE); + dev_priv->gct_data.bpi = bpi; /*save boot panel id*/ + + /*copy the GCT display timings into a temp structure*/ + memcpy(&ti, pGCT, sizeof(struct gct_r10_timing_info)); + + /*now copy the temp struct into the dev_priv->gct_data*/ + dp_ti->pixel_clock = ti.pixel_clock; + dp_ti->hactive_hi = ti.hactive_hi; + dp_ti->hactive_lo = ti.hactive_lo; + dp_ti->hblank_hi = ti.hblank_hi; + dp_ti->hblank_lo = ti.hblank_lo; + dp_ti->hsync_offset_hi = ti.hsync_offset_hi; + dp_ti->hsync_offset_lo = ti.hsync_offset_lo; + dp_ti->hsync_pulse_width_hi = ti.hsync_pulse_width_hi; + dp_ti->hsync_pulse_width_lo = ti.hsync_pulse_width_lo; + dp_ti->vactive_hi = ti.vactive_hi; + dp_ti->vactive_lo = ti.vactive_lo; + dp_ti->vblank_hi = ti.vblank_hi; + dp_ti->vblank_lo = ti.vblank_lo; + dp_ti->vsync_offset_hi = ti.vsync_offset_hi; + dp_ti->vsync_offset_lo = ti.vsync_offset_lo; + dp_ti->vsync_pulse_width_hi = ti.vsync_pulse_width_hi; + dp_ti->vsync_pulse_width_lo = ti.vsync_pulse_width_lo; + + /* Move the MIPI_Display_Descriptor data from GCT to dev priv */ + dev_priv->gct_data.Panel_MIPI_Display_Descriptor = + *((u8 *)pGCT + 0x0d); + dev_priv->gct_data.Panel_MIPI_Display_Descriptor |= + (*((u8 *)pGCT + 0x0e)) << 8; + break; + default: + dev_err(dev->dev, "Unknown revision of GCT!\n"); + vbt->size = 0; + } +} + +int mid_chip_setup(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + mid_get_fuse_settings(dev); + mid_get_vbt_data(dev_priv); + mid_get_pci_revID(dev_priv); + return 0; +} diff --git a/drivers/gpu/drm/gma500/mid_bios.h b/drivers/gpu/drm/gma500/mid_bios.h new file mode 100644 index 0000000..00e7d56 --- /dev/null +++ b/drivers/gpu/drm/gma500/mid_bios.h @@ -0,0 +1,21 @@ +/************************************************************************** + * Copyright (c) 2011, Intel Corporation. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + **************************************************************************/ + +extern int mid_chip_setup(struct drm_device *dev); + diff --git a/drivers/gpu/drm/gma500/mmu.c b/drivers/gpu/drm/gma500/mmu.c new file mode 100644 index 0000000..49bac41 --- /dev/null +++ b/drivers/gpu/drm/gma500/mmu.c @@ -0,0 +1,849 @@ +/************************************************************************** + * Copyright (c) 2007, Intel Corporation. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + **************************************************************************/ +#include +#include "psb_drv.h" +#include "psb_reg.h" + +/* + * Code for the SGX MMU: + */ + +/* + * clflush on one processor only: + * clflush should apparently flush the cache line on all processors in an + * SMP system. + */ + +/* + * kmap atomic: + * The usage of the slots must be completely encapsulated within a spinlock, and + * no other functions that may be using the locks for other purposed may be + * called from within the locked region. + * Since the slots are per processor, this will guarantee that we are the only + * user. + */ + +/* + * TODO: Inserting ptes from an interrupt handler: + * This may be desirable for some SGX functionality where the GPU can fault in + * needed pages. For that, we need to make an atomic insert_pages function, that + * may fail. + * If it fails, the caller need to insert the page using a workqueue function, + * but on average it should be fast. + */ + +struct psb_mmu_driver { + /* protects driver- and pd structures. Always take in read mode + * before taking the page table spinlock. + */ + struct rw_semaphore sem; + + /* protects page tables, directory tables and pt tables. + * and pt structures. + */ + spinlock_t lock; + + atomic_t needs_tlbflush; + + uint8_t __iomem *register_map; + struct psb_mmu_pd *default_pd; + /*uint32_t bif_ctrl;*/ + int has_clflush; + int clflush_add; + unsigned long clflush_mask; + + struct drm_psb_private *dev_priv; +}; + +struct psb_mmu_pd; + +struct psb_mmu_pt { + struct psb_mmu_pd *pd; + uint32_t index; + uint32_t count; + struct page *p; + uint32_t *v; +}; + +struct psb_mmu_pd { + struct psb_mmu_driver *driver; + int hw_context; + struct psb_mmu_pt **tables; + struct page *p; + struct page *dummy_pt; + struct page *dummy_page; + uint32_t pd_mask; + uint32_t invalid_pde; + uint32_t invalid_pte; +}; + +static inline uint32_t psb_mmu_pt_index(uint32_t offset) +{ + return (offset >> PSB_PTE_SHIFT) & 0x3FF; +} + +static inline uint32_t psb_mmu_pd_index(uint32_t offset) +{ + return offset >> PSB_PDE_SHIFT; +} + +static inline void psb_clflush(void *addr) +{ + __asm__ __volatile__("clflush (%0)\n" : : "r"(addr) : "memory"); +} + +static inline void psb_mmu_clflush(struct psb_mmu_driver *driver, + void *addr) +{ + if (!driver->has_clflush) + return; + + mb(); + psb_clflush(addr); + mb(); +} + +static void psb_page_clflush(struct psb_mmu_driver *driver, struct page* page) +{ + uint32_t clflush_add = driver->clflush_add >> PAGE_SHIFT; + uint32_t clflush_count = PAGE_SIZE / clflush_add; + int i; + uint8_t *clf; + + clf = kmap_atomic(page); + mb(); + for (i = 0; i < clflush_count; ++i) { + psb_clflush(clf); + clf += clflush_add; + } + mb(); + kunmap_atomic(clf); +} + +static void psb_pages_clflush(struct psb_mmu_driver *driver, + struct page *page[], unsigned long num_pages) +{ + int i; + + if (!driver->has_clflush) + return ; + + for (i = 0; i < num_pages; i++) + psb_page_clflush(driver, *page++); +} + +static void psb_mmu_flush_pd_locked(struct psb_mmu_driver *driver, + int force) +{ + atomic_set(&driver->needs_tlbflush, 0); +} + +static void psb_mmu_flush_pd(struct psb_mmu_driver *driver, int force) +{ + down_write(&driver->sem); + psb_mmu_flush_pd_locked(driver, force); + up_write(&driver->sem); +} + +void psb_mmu_flush(struct psb_mmu_driver *driver, int rc_prot) +{ + if (rc_prot) + down_write(&driver->sem); + if (rc_prot) + up_write(&driver->sem); +} + +void psb_mmu_set_pd_context(struct psb_mmu_pd *pd, int hw_context) +{ + /*ttm_tt_cache_flush(&pd->p, 1);*/ + psb_pages_clflush(pd->driver, &pd->p, 1); + down_write(&pd->driver->sem); + wmb(); + psb_mmu_flush_pd_locked(pd->driver, 1); + pd->hw_context = hw_context; + up_write(&pd->driver->sem); + +} + +static inline unsigned long psb_pd_addr_end(unsigned long addr, + unsigned long end) +{ + + addr = (addr + PSB_PDE_MASK + 1) & ~PSB_PDE_MASK; + return (addr < end) ? addr : end; +} + +static inline uint32_t psb_mmu_mask_pte(uint32_t pfn, int type) +{ + uint32_t mask = PSB_PTE_VALID; + + if (type & PSB_MMU_CACHED_MEMORY) + mask |= PSB_PTE_CACHED; + if (type & PSB_MMU_RO_MEMORY) + mask |= PSB_PTE_RO; + if (type & PSB_MMU_WO_MEMORY) + mask |= PSB_PTE_WO; + + return (pfn << PAGE_SHIFT) | mask; +} + +struct psb_mmu_pd *psb_mmu_alloc_pd(struct psb_mmu_driver *driver, + int trap_pagefaults, int invalid_type) +{ + struct psb_mmu_pd *pd = kmalloc(sizeof(*pd), GFP_KERNEL); + uint32_t *v; + int i; + + if (!pd) + return NULL; + + pd->p = alloc_page(GFP_DMA32); + if (!pd->p) + goto out_err1; + pd->dummy_pt = alloc_page(GFP_DMA32); + if (!pd->dummy_pt) + goto out_err2; + pd->dummy_page = alloc_page(GFP_DMA32); + if (!pd->dummy_page) + goto out_err3; + + if (!trap_pagefaults) { + pd->invalid_pde = + psb_mmu_mask_pte(page_to_pfn(pd->dummy_pt), + invalid_type); + pd->invalid_pte = + psb_mmu_mask_pte(page_to_pfn(pd->dummy_page), + invalid_type); + } else { + pd->invalid_pde = 0; + pd->invalid_pte = 0; + } + + v = kmap(pd->dummy_pt); + for (i = 0; i < (PAGE_SIZE / sizeof(uint32_t)); ++i) + v[i] = pd->invalid_pte; + + kunmap(pd->dummy_pt); + + v = kmap(pd->p); + for (i = 0; i < (PAGE_SIZE / sizeof(uint32_t)); ++i) + v[i] = pd->invalid_pde; + + kunmap(pd->p); + + clear_page(kmap(pd->dummy_page)); + kunmap(pd->dummy_page); + + pd->tables = vmalloc_user(sizeof(struct psb_mmu_pt *) * 1024); + if (!pd->tables) + goto out_err4; + + pd->hw_context = -1; + pd->pd_mask = PSB_PTE_VALID; + pd->driver = driver; + + return pd; + +out_err4: + __free_page(pd->dummy_page); +out_err3: + __free_page(pd->dummy_pt); +out_err2: + __free_page(pd->p); +out_err1: + kfree(pd); + return NULL; +} + +static void psb_mmu_free_pt(struct psb_mmu_pt *pt) +{ + __free_page(pt->p); + kfree(pt); +} + +void psb_mmu_free_pagedir(struct psb_mmu_pd *pd) +{ + struct psb_mmu_driver *driver = pd->driver; + struct psb_mmu_pt *pt; + int i; + + down_write(&driver->sem); + if (pd->hw_context != -1) + psb_mmu_flush_pd_locked(driver, 1); + + /* Should take the spinlock here, but we don't need to do that + since we have the semaphore in write mode. */ + + for (i = 0; i < 1024; ++i) { + pt = pd->tables[i]; + if (pt) + psb_mmu_free_pt(pt); + } + + vfree(pd->tables); + __free_page(pd->dummy_page); + __free_page(pd->dummy_pt); + __free_page(pd->p); + kfree(pd); + up_write(&driver->sem); +} + +static struct psb_mmu_pt *psb_mmu_alloc_pt(struct psb_mmu_pd *pd) +{ + struct psb_mmu_pt *pt = kmalloc(sizeof(*pt), GFP_KERNEL); + void *v; + uint32_t clflush_add = pd->driver->clflush_add >> PAGE_SHIFT; + uint32_t clflush_count = PAGE_SIZE / clflush_add; + spinlock_t *lock = &pd->driver->lock; + uint8_t *clf; + uint32_t *ptes; + int i; + + if (!pt) + return NULL; + + pt->p = alloc_page(GFP_DMA32); + if (!pt->p) { + kfree(pt); + return NULL; + } + + spin_lock(lock); + + v = kmap_atomic(pt->p); + clf = (uint8_t *) v; + ptes = (uint32_t *) v; + for (i = 0; i < (PAGE_SIZE / sizeof(uint32_t)); ++i) + *ptes++ = pd->invalid_pte; + + + if (pd->driver->has_clflush && pd->hw_context != -1) { + mb(); + for (i = 0; i < clflush_count; ++i) { + psb_clflush(clf); + clf += clflush_add; + } + mb(); + } + + kunmap_atomic(v); + spin_unlock(lock); + + pt->count = 0; + pt->pd = pd; + pt->index = 0; + + return pt; +} + +static struct psb_mmu_pt *psb_mmu_pt_alloc_map_lock(struct psb_mmu_pd *pd, + unsigned long addr) +{ + uint32_t index = psb_mmu_pd_index(addr); + struct psb_mmu_pt *pt; + uint32_t *v; + spinlock_t *lock = &pd->driver->lock; + + spin_lock(lock); + pt = pd->tables[index]; + while (!pt) { + spin_unlock(lock); + pt = psb_mmu_alloc_pt(pd); + if (!pt) + return NULL; + spin_lock(lock); + + if (pd->tables[index]) { + spin_unlock(lock); + psb_mmu_free_pt(pt); + spin_lock(lock); + pt = pd->tables[index]; + continue; + } + + v = kmap_atomic(pd->p); + pd->tables[index] = pt; + v[index] = (page_to_pfn(pt->p) << 12) | pd->pd_mask; + pt->index = index; + kunmap_atomic((void *) v); + + if (pd->hw_context != -1) { + psb_mmu_clflush(pd->driver, (void *) &v[index]); + atomic_set(&pd->driver->needs_tlbflush, 1); + } + } + pt->v = kmap_atomic(pt->p); + return pt; +} + +static struct psb_mmu_pt *psb_mmu_pt_map_lock(struct psb_mmu_pd *pd, + unsigned long addr) +{ + uint32_t index = psb_mmu_pd_index(addr); + struct psb_mmu_pt *pt; + spinlock_t *lock = &pd->driver->lock; + + spin_lock(lock); + pt = pd->tables[index]; + if (!pt) { + spin_unlock(lock); + return NULL; + } + pt->v = kmap_atomic(pt->p); + return pt; +} + +static void psb_mmu_pt_unmap_unlock(struct psb_mmu_pt *pt) +{ + struct psb_mmu_pd *pd = pt->pd; + uint32_t *v; + + kunmap_atomic(pt->v); + if (pt->count == 0) { + v = kmap_atomic(pd->p); + v[pt->index] = pd->invalid_pde; + pd->tables[pt->index] = NULL; + + if (pd->hw_context != -1) { + psb_mmu_clflush(pd->driver, + (void *) &v[pt->index]); + atomic_set(&pd->driver->needs_tlbflush, 1); + } + kunmap_atomic(pt->v); + spin_unlock(&pd->driver->lock); + psb_mmu_free_pt(pt); + return; + } + spin_unlock(&pd->driver->lock); +} + +static inline void psb_mmu_set_pte(struct psb_mmu_pt *pt, + unsigned long addr, uint32_t pte) +{ + pt->v[psb_mmu_pt_index(addr)] = pte; +} + +static inline void psb_mmu_invalidate_pte(struct psb_mmu_pt *pt, + unsigned long addr) +{ + pt->v[psb_mmu_pt_index(addr)] = pt->pd->invalid_pte; +} + + +void psb_mmu_mirror_gtt(struct psb_mmu_pd *pd, + uint32_t mmu_offset, uint32_t gtt_start, + uint32_t gtt_pages) +{ + uint32_t *v; + uint32_t start = psb_mmu_pd_index(mmu_offset); + struct psb_mmu_driver *driver = pd->driver; + int num_pages = gtt_pages; + + down_read(&driver->sem); + spin_lock(&driver->lock); + + v = kmap_atomic(pd->p); + v += start; + + while (gtt_pages--) { + *v++ = gtt_start | pd->pd_mask; + gtt_start += PAGE_SIZE; + } + + /*ttm_tt_cache_flush(&pd->p, num_pages);*/ + psb_pages_clflush(pd->driver, &pd->p, num_pages); + kunmap_atomic(v); + spin_unlock(&driver->lock); + + if (pd->hw_context != -1) + atomic_set(&pd->driver->needs_tlbflush, 1); + + up_read(&pd->driver->sem); + psb_mmu_flush_pd(pd->driver, 0); +} + +struct psb_mmu_pd *psb_mmu_get_default_pd(struct psb_mmu_driver *driver) +{ + struct psb_mmu_pd *pd; + + /* down_read(&driver->sem); */ + pd = driver->default_pd; + /* up_read(&driver->sem); */ + + return pd; +} + +void psb_mmu_driver_takedown(struct psb_mmu_driver *driver) +{ + psb_mmu_free_pagedir(driver->default_pd); + kfree(driver); +} + +struct psb_mmu_driver *psb_mmu_driver_init(uint8_t __iomem * registers, + int trap_pagefaults, + int invalid_type, + struct drm_psb_private *dev_priv) +{ + struct psb_mmu_driver *driver; + + driver = kmalloc(sizeof(*driver), GFP_KERNEL); + + if (!driver) + return NULL; + driver->dev_priv = dev_priv; + + driver->default_pd = psb_mmu_alloc_pd(driver, trap_pagefaults, + invalid_type); + if (!driver->default_pd) + goto out_err1; + + spin_lock_init(&driver->lock); + init_rwsem(&driver->sem); + down_write(&driver->sem); + driver->register_map = registers; + atomic_set(&driver->needs_tlbflush, 1); + + driver->has_clflush = 0; + + if (boot_cpu_has(X86_FEATURE_CLFLSH)) { + uint32_t tfms, misc, cap0, cap4, clflush_size; + + /* + * clflush size is determined at kernel setup for x86_64 + * but not for i386. We have to do it here. + */ + + cpuid(0x00000001, &tfms, &misc, &cap0, &cap4); + clflush_size = ((misc >> 8) & 0xff) * 8; + driver->has_clflush = 1; + driver->clflush_add = + PAGE_SIZE * clflush_size / sizeof(uint32_t); + driver->clflush_mask = driver->clflush_add - 1; + driver->clflush_mask = ~driver->clflush_mask; + } + + up_write(&driver->sem); + return driver; + +out_err1: + kfree(driver); + return NULL; +} + +static void psb_mmu_flush_ptes(struct psb_mmu_pd *pd, + unsigned long address, uint32_t num_pages, + uint32_t desired_tile_stride, + uint32_t hw_tile_stride) +{ + struct psb_mmu_pt *pt; + uint32_t rows = 1; + uint32_t i; + unsigned long addr; + unsigned long end; + unsigned long next; + unsigned long add; + unsigned long row_add; + unsigned long clflush_add = pd->driver->clflush_add; + unsigned long clflush_mask = pd->driver->clflush_mask; + + if (!pd->driver->has_clflush) { + /*ttm_tt_cache_flush(&pd->p, num_pages);*/ + psb_pages_clflush(pd->driver, &pd->p, num_pages); + return; + } + + if (hw_tile_stride) + rows = num_pages / desired_tile_stride; + else + desired_tile_stride = num_pages; + + add = desired_tile_stride << PAGE_SHIFT; + row_add = hw_tile_stride << PAGE_SHIFT; + mb(); + for (i = 0; i < rows; ++i) { + + addr = address; + end = addr + add; + + do { + next = psb_pd_addr_end(addr, end); + pt = psb_mmu_pt_map_lock(pd, addr); + if (!pt) + continue; + do { + psb_clflush(&pt->v + [psb_mmu_pt_index(addr)]); + } while (addr += + clflush_add, + (addr & clflush_mask) < next); + + psb_mmu_pt_unmap_unlock(pt); + } while (addr = next, next != end); + address += row_add; + } + mb(); +} + +void psb_mmu_remove_pfn_sequence(struct psb_mmu_pd *pd, + unsigned long address, uint32_t num_pages) +{ + struct psb_mmu_pt *pt; + unsigned long addr; + unsigned long end; + unsigned long next; + unsigned long f_address = address; + + down_read(&pd->driver->sem); + + addr = address; + end = addr + (num_pages << PAGE_SHIFT); + + do { + next = psb_pd_addr_end(addr, end); + pt = psb_mmu_pt_alloc_map_lock(pd, addr); + if (!pt) + goto out; + do { + psb_mmu_invalidate_pte(pt, addr); + --pt->count; + } while (addr += PAGE_SIZE, addr < next); + psb_mmu_pt_unmap_unlock(pt); + + } while (addr = next, next != end); + +out: + if (pd->hw_context != -1) + psb_mmu_flush_ptes(pd, f_address, num_pages, 1, 1); + + up_read(&pd->driver->sem); + + if (pd->hw_context != -1) + psb_mmu_flush(pd->driver, 0); + + return; +} + +void psb_mmu_remove_pages(struct psb_mmu_pd *pd, unsigned long address, + uint32_t num_pages, uint32_t desired_tile_stride, + uint32_t hw_tile_stride) +{ + struct psb_mmu_pt *pt; + uint32_t rows = 1; + uint32_t i; + unsigned long addr; + unsigned long end; + unsigned long next; + unsigned long add; + unsigned long row_add; + unsigned long f_address = address; + + if (hw_tile_stride) + rows = num_pages / desired_tile_stride; + else + desired_tile_stride = num_pages; + + add = desired_tile_stride << PAGE_SHIFT; + row_add = hw_tile_stride << PAGE_SHIFT; + + /* down_read(&pd->driver->sem); */ + + /* Make sure we only need to flush this processor's cache */ + + for (i = 0; i < rows; ++i) { + + addr = address; + end = addr + add; + + do { + next = psb_pd_addr_end(addr, end); + pt = psb_mmu_pt_map_lock(pd, addr); + if (!pt) + continue; + do { + psb_mmu_invalidate_pte(pt, addr); + --pt->count; + + } while (addr += PAGE_SIZE, addr < next); + psb_mmu_pt_unmap_unlock(pt); + + } while (addr = next, next != end); + address += row_add; + } + if (pd->hw_context != -1) + psb_mmu_flush_ptes(pd, f_address, num_pages, + desired_tile_stride, hw_tile_stride); + + /* up_read(&pd->driver->sem); */ + + if (pd->hw_context != -1) + psb_mmu_flush(pd->driver, 0); +} + +int psb_mmu_insert_pfn_sequence(struct psb_mmu_pd *pd, uint32_t start_pfn, + unsigned long address, uint32_t num_pages, + int type) +{ + struct psb_mmu_pt *pt; + uint32_t pte; + unsigned long addr; + unsigned long end; + unsigned long next; + unsigned long f_address = address; + int ret = 0; + + down_read(&pd->driver->sem); + + addr = address; + end = addr + (num_pages << PAGE_SHIFT); + + do { + next = psb_pd_addr_end(addr, end); + pt = psb_mmu_pt_alloc_map_lock(pd, addr); + if (!pt) { + ret = -ENOMEM; + goto out; + } + do { + pte = psb_mmu_mask_pte(start_pfn++, type); + psb_mmu_set_pte(pt, addr, pte); + pt->count++; + } while (addr += PAGE_SIZE, addr < next); + psb_mmu_pt_unmap_unlock(pt); + + } while (addr = next, next != end); + +out: + if (pd->hw_context != -1) + psb_mmu_flush_ptes(pd, f_address, num_pages, 1, 1); + + up_read(&pd->driver->sem); + + if (pd->hw_context != -1) + psb_mmu_flush(pd->driver, 1); + + return ret; +} + +int psb_mmu_insert_pages(struct psb_mmu_pd *pd, struct page **pages, + unsigned long address, uint32_t num_pages, + uint32_t desired_tile_stride, + uint32_t hw_tile_stride, int type) +{ + struct psb_mmu_pt *pt; + uint32_t rows = 1; + uint32_t i; + uint32_t pte; + unsigned long addr; + unsigned long end; + unsigned long next; + unsigned long add; + unsigned long row_add; + unsigned long f_address = address; + int ret = 0; + + if (hw_tile_stride) { + if (num_pages % desired_tile_stride != 0) + return -EINVAL; + rows = num_pages / desired_tile_stride; + } else { + desired_tile_stride = num_pages; + } + + add = desired_tile_stride << PAGE_SHIFT; + row_add = hw_tile_stride << PAGE_SHIFT; + + down_read(&pd->driver->sem); + + for (i = 0; i < rows; ++i) { + + addr = address; + end = addr + add; + + do { + next = psb_pd_addr_end(addr, end); + pt = psb_mmu_pt_alloc_map_lock(pd, addr); + if (!pt) { + ret = -ENOMEM; + goto out; + } + do { + pte = + psb_mmu_mask_pte(page_to_pfn(*pages++), + type); + psb_mmu_set_pte(pt, addr, pte); + pt->count++; + } while (addr += PAGE_SIZE, addr < next); + psb_mmu_pt_unmap_unlock(pt); + + } while (addr = next, next != end); + + address += row_add; + } +out: + if (pd->hw_context != -1) + psb_mmu_flush_ptes(pd, f_address, num_pages, + desired_tile_stride, hw_tile_stride); + + up_read(&pd->driver->sem); + + if (pd->hw_context != -1) + psb_mmu_flush(pd->driver, 1); + + return ret; +} + +int psb_mmu_virtual_to_pfn(struct psb_mmu_pd *pd, uint32_t virtual, + unsigned long *pfn) +{ + int ret; + struct psb_mmu_pt *pt; + uint32_t tmp; + spinlock_t *lock = &pd->driver->lock; + + down_read(&pd->driver->sem); + pt = psb_mmu_pt_map_lock(pd, virtual); + if (!pt) { + uint32_t *v; + + spin_lock(lock); + v = kmap_atomic(pd->p); + tmp = v[psb_mmu_pd_index(virtual)]; + kunmap_atomic(v); + spin_unlock(lock); + + if (tmp != pd->invalid_pde || !(tmp & PSB_PTE_VALID) || + !(pd->invalid_pte & PSB_PTE_VALID)) { + ret = -EINVAL; + goto out; + } + ret = 0; + *pfn = pd->invalid_pte >> PAGE_SHIFT; + goto out; + } + tmp = pt->v[psb_mmu_pt_index(virtual)]; + if (!(tmp & PSB_PTE_VALID)) { + ret = -EINVAL; + } else { + ret = 0; + *pfn = tmp >> PAGE_SHIFT; + } + psb_mmu_pt_unmap_unlock(pt); +out: + up_read(&pd->driver->sem); + return ret; +} diff --git a/drivers/gpu/drm/gma500/oaktrail.h b/drivers/gpu/drm/gma500/oaktrail.h new file mode 100644 index 0000000..2da1f36 --- /dev/null +++ b/drivers/gpu/drm/gma500/oaktrail.h @@ -0,0 +1,252 @@ +/************************************************************************** + * Copyright (c) 2007-2011, Intel Corporation. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + **************************************************************************/ + +/* MID device specific descriptors */ + +struct oaktrail_vbt { + s8 signature[4]; /*4 bytes,"$GCT" */ + u8 revision; + u8 size; + u8 checksum; + void *oaktrail_gct; +} __packed; + +struct oaktrail_timing_info { + u16 pixel_clock; + u8 hactive_lo; + u8 hblank_lo; + u8 hblank_hi:4; + u8 hactive_hi:4; + u8 vactive_lo; + u8 vblank_lo; + u8 vblank_hi:4; + u8 vactive_hi:4; + u8 hsync_offset_lo; + u8 hsync_pulse_width_lo; + u8 vsync_pulse_width_lo:4; + u8 vsync_offset_lo:4; + u8 vsync_pulse_width_hi:2; + u8 vsync_offset_hi:2; + u8 hsync_pulse_width_hi:2; + u8 hsync_offset_hi:2; + u8 width_mm_lo; + u8 height_mm_lo; + u8 height_mm_hi:4; + u8 width_mm_hi:4; + u8 hborder; + u8 vborder; + u8 unknown0:1; + u8 hsync_positive:1; + u8 vsync_positive:1; + u8 separate_sync:2; + u8 stereo:1; + u8 unknown6:1; + u8 interlaced:1; +} __packed; + +struct gct_r10_timing_info { + u16 pixel_clock; + u32 hactive_lo:8; + u32 hactive_hi:4; + u32 hblank_lo:8; + u32 hblank_hi:4; + u32 hsync_offset_lo:8; + u16 hsync_offset_hi:2; + u16 hsync_pulse_width_lo:8; + u16 hsync_pulse_width_hi:2; + u16 hsync_positive:1; + u16 rsvd_1:3; + u8 vactive_lo:8; + u16 vactive_hi:4; + u16 vblank_lo:8; + u16 vblank_hi:4; + u16 vsync_offset_lo:4; + u16 vsync_offset_hi:2; + u16 vsync_pulse_width_lo:4; + u16 vsync_pulse_width_hi:2; + u16 vsync_positive:1; + u16 rsvd_2:3; +} __packed; + +struct oaktrail_panel_descriptor_v1 { + u32 Panel_Port_Control; /* 1 dword, Register 0x61180 if LVDS */ + /* 0x61190 if MIPI */ + u32 Panel_Power_On_Sequencing;/*1 dword,Register 0x61208,*/ + u32 Panel_Power_Off_Sequencing;/*1 dword,Register 0x6120C,*/ + u32 Panel_Power_Cycle_Delay_and_Reference_Divisor;/* 1 dword */ + /* Register 0x61210 */ + struct oaktrail_timing_info DTD;/*18 bytes, Standard definition */ + u16 Panel_Backlight_Inverter_Descriptor;/* 16 bits, as follows */ + /* Bit 0, Frequency, 15 bits,0 - 32767Hz */ + /* Bit 15, Polarity, 1 bit, 0: Normal, 1: Inverted */ + u16 Panel_MIPI_Display_Descriptor; + /*16 bits, Defined as follows: */ + /* if MIPI, 0x0000 if LVDS */ + /* Bit 0, Type, 2 bits, */ + /* 0: Type-1, */ + /* 1: Type-2, */ + /* 2: Type-3, */ + /* 3: Type-4 */ + /* Bit 2, Pixel Format, 4 bits */ + /* Bit0: 16bpp (not supported in LNC), */ + /* Bit1: 18bpp loosely packed, */ + /* Bit2: 18bpp packed, */ + /* Bit3: 24bpp */ + /* Bit 6, Reserved, 2 bits, 00b */ + /* Bit 8, Minimum Supported Frame Rate, 6 bits, 0 - 63Hz */ + /* Bit 14, Reserved, 2 bits, 00b */ +} __packed; + +struct oaktrail_panel_descriptor_v2 { + u32 Panel_Port_Control; /* 1 dword, Register 0x61180 if LVDS */ + /* 0x61190 if MIPI */ + u32 Panel_Power_On_Sequencing;/*1 dword,Register 0x61208,*/ + u32 Panel_Power_Off_Sequencing;/*1 dword,Register 0x6120C,*/ + u8 Panel_Power_Cycle_Delay_and_Reference_Divisor;/* 1 byte */ + /* Register 0x61210 */ + struct oaktrail_timing_info DTD;/*18 bytes, Standard definition */ + u16 Panel_Backlight_Inverter_Descriptor;/*16 bits, as follows*/ + /*Bit 0, Frequency, 16 bits, 0 - 32767Hz*/ + u8 Panel_Initial_Brightness;/* [7:0] 0 - 100% */ + /*Bit 7, Polarity, 1 bit,0: Normal, 1: Inverted*/ + u16 Panel_MIPI_Display_Descriptor; + /*16 bits, Defined as follows: */ + /* if MIPI, 0x0000 if LVDS */ + /* Bit 0, Type, 2 bits, */ + /* 0: Type-1, */ + /* 1: Type-2, */ + /* 2: Type-3, */ + /* 3: Type-4 */ + /* Bit 2, Pixel Format, 4 bits */ + /* Bit0: 16bpp (not supported in LNC), */ + /* Bit1: 18bpp loosely packed, */ + /* Bit2: 18bpp packed, */ + /* Bit3: 24bpp */ + /* Bit 6, Reserved, 2 bits, 00b */ + /* Bit 8, Minimum Supported Frame Rate, 6 bits, 0 - 63Hz */ + /* Bit 14, Reserved, 2 bits, 00b */ +} __packed; + +union oaktrail_panel_rx { + struct { + u16 NumberOfLanes:2; /*Num of Lanes, 2 bits,0 = 1 lane,*/ + /* 1 = 2 lanes, 2 = 3 lanes, 3 = 4 lanes. */ + u16 MaxLaneFreq:3; /* 0: 100MHz, 1: 200MHz, 2: 300MHz, */ + /*3: 400MHz, 4: 500MHz, 5: 600MHz, 6: 700MHz, 7: 800MHz.*/ + u16 SupportedVideoTransferMode:2; /*0: Non-burst only */ + /* 1: Burst and non-burst */ + /* 2/3: Reserved */ + u16 HSClkBehavior:1; /*0: Continuous, 1: Non-continuous*/ + u16 DuoDisplaySupport:1; /*1 bit,0: No, 1: Yes*/ + u16 ECC_ChecksumCapabilities:1;/*1 bit,0: No, 1: Yes*/ + u16 BidirectionalCommunication:1;/*1 bit,0: No, 1: Yes */ + u16 Rsvd:5;/*5 bits,00000b */ + } panelrx; + u16 panel_receiver; +} __packed; + +struct oaktrail_gct_v1 { + union { /*8 bits,Defined as follows: */ + struct { + u8 PanelType:4; /*4 bits, Bit field for panels*/ + /* 0 - 3: 0 = LVDS, 1 = MIPI*/ + /*2 bits,Specifies which of the*/ + u8 BootPanelIndex:2; + /* 4 panels to use by default*/ + u8 BootMIPI_DSI_RxIndex:2;/*Specifies which of*/ + /* the 4 MIPI DSI receivers to use*/ + } PD; + u8 PanelDescriptor; + }; + struct oaktrail_panel_descriptor_v1 panel[4];/*panel descrs,38 bytes each*/ + union oaktrail_panel_rx panelrx[4]; /* panel receivers*/ +} __packed; + +struct oaktrail_gct_v2 { + union { /*8 bits,Defined as follows: */ + struct { + u8 PanelType:4; /*4 bits, Bit field for panels*/ + /* 0 - 3: 0 = LVDS, 1 = MIPI*/ + /*2 bits,Specifies which of the*/ + u8 BootPanelIndex:2; + /* 4 panels to use by default*/ + u8 BootMIPI_DSI_RxIndex:2;/*Specifies which of*/ + /* the 4 MIPI DSI receivers to use*/ + } PD; + u8 PanelDescriptor; + }; + struct oaktrail_panel_descriptor_v2 panel[4];/*panel descrs,38 bytes each*/ + union oaktrail_panel_rx panelrx[4]; /* panel receivers*/ +} __packed; + +struct oaktrail_gct_data { + u8 bpi; /* boot panel index, number of panel used during boot */ + u8 pt; /* panel type, 4 bit field, 0=lvds, 1=mipi */ + struct oaktrail_timing_info DTD; /* timing info for the selected panel */ + u32 Panel_Port_Control; + u32 PP_On_Sequencing;/*1 dword,Register 0x61208,*/ + u32 PP_Off_Sequencing;/*1 dword,Register 0x6120C,*/ + u32 PP_Cycle_Delay; + u16 Panel_Backlight_Inverter_Descriptor; + u16 Panel_MIPI_Display_Descriptor; +} __packed; + +#define MODE_SETTING_IN_CRTC 0x1 +#define MODE_SETTING_IN_ENCODER 0x2 +#define MODE_SETTING_ON_GOING 0x3 +#define MODE_SETTING_IN_DSR 0x4 +#define MODE_SETTING_ENCODER_DONE 0x8 + +#define GCT_R10_HEADER_SIZE 16 +#define GCT_R10_DISPLAY_DESC_SIZE 28 + +/* + * Moorestown HDMI interfaces + */ + +struct oaktrail_hdmi_dev { + struct pci_dev *dev; + void __iomem *regs; + unsigned int mmio, mmio_len; + int dpms_mode; + struct hdmi_i2c_dev *i2c_dev; + + /* register state */ + u32 saveDPLL_CTRL; + u32 saveDPLL_DIV_CTRL; + u32 saveDPLL_ADJUST; + u32 saveDPLL_UPDATE; + u32 saveDPLL_CLK_ENABLE; + u32 savePCH_HTOTAL_B; + u32 savePCH_HBLANK_B; + u32 savePCH_HSYNC_B; + u32 savePCH_VTOTAL_B; + u32 savePCH_VBLANK_B; + u32 savePCH_VSYNC_B; + u32 savePCH_PIPEBCONF; + u32 savePCH_PIPEBSRC; +}; + +extern void oaktrail_hdmi_setup(struct drm_device *dev); +extern void oaktrail_hdmi_teardown(struct drm_device *dev); +extern int oaktrail_hdmi_i2c_init(struct pci_dev *dev); +extern void oaktrail_hdmi_i2c_exit(struct pci_dev *dev); +extern void oaktrail_hdmi_save(struct drm_device *dev); +extern void oaktrail_hdmi_restore(struct drm_device *dev); +extern void oaktrail_hdmi_init(struct drm_device *dev, struct psb_intel_mode_device *mode_dev); diff --git a/drivers/gpu/drm/gma500/oaktrail_crtc.c b/drivers/gpu/drm/gma500/oaktrail_crtc.c new file mode 100644 index 0000000..a39b0d0 --- /dev/null +++ b/drivers/gpu/drm/gma500/oaktrail_crtc.c @@ -0,0 +1,592 @@ +/* + * Copyright © 2009 Intel Corporation + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#include +#include + +#include +#include "framebuffer.h" +#include "psb_drv.h" +#include "psb_intel_drv.h" +#include "psb_intel_reg.h" +#include "psb_intel_display.h" +#include "power.h" + +struct psb_intel_range_t { + int min, max; +}; + +struct oaktrail_limit_t { + struct psb_intel_range_t dot, m, p1; +}; + +struct oaktrail_clock_t { + /* derived values */ + int dot; + int m; + int p1; +}; + +#define MRST_LIMIT_LVDS_100L 0 +#define MRST_LIMIT_LVDS_83 1 +#define MRST_LIMIT_LVDS_100 2 + +#define MRST_DOT_MIN 19750 +#define MRST_DOT_MAX 120000 +#define MRST_M_MIN_100L 20 +#define MRST_M_MIN_100 10 +#define MRST_M_MIN_83 12 +#define MRST_M_MAX_100L 34 +#define MRST_M_MAX_100 17 +#define MRST_M_MAX_83 20 +#define MRST_P1_MIN 2 +#define MRST_P1_MAX_0 7 +#define MRST_P1_MAX_1 8 + +static const struct oaktrail_limit_t oaktrail_limits[] = { + { /* MRST_LIMIT_LVDS_100L */ + .dot = {.min = MRST_DOT_MIN, .max = MRST_DOT_MAX}, + .m = {.min = MRST_M_MIN_100L, .max = MRST_M_MAX_100L}, + .p1 = {.min = MRST_P1_MIN, .max = MRST_P1_MAX_1}, + }, + { /* MRST_LIMIT_LVDS_83L */ + .dot = {.min = MRST_DOT_MIN, .max = MRST_DOT_MAX}, + .m = {.min = MRST_M_MIN_83, .max = MRST_M_MAX_83}, + .p1 = {.min = MRST_P1_MIN, .max = MRST_P1_MAX_0}, + }, + { /* MRST_LIMIT_LVDS_100 */ + .dot = {.min = MRST_DOT_MIN, .max = MRST_DOT_MAX}, + .m = {.min = MRST_M_MIN_100, .max = MRST_M_MAX_100}, + .p1 = {.min = MRST_P1_MIN, .max = MRST_P1_MAX_1}, + }, +}; + +#define MRST_M_MIN 10 +static const u32 oaktrail_m_converts[] = { + 0x2B, 0x15, 0x2A, 0x35, 0x1A, 0x0D, 0x26, 0x33, 0x19, 0x2C, + 0x36, 0x3B, 0x1D, 0x2E, 0x37, 0x1B, 0x2D, 0x16, 0x0B, 0x25, + 0x12, 0x09, 0x24, 0x32, 0x39, 0x1c, +}; + +static const struct oaktrail_limit_t *oaktrail_limit(struct drm_crtc *crtc) +{ + const struct oaktrail_limit_t *limit = NULL; + struct drm_device *dev = crtc->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + + if (psb_intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) + || psb_intel_pipe_has_type(crtc, INTEL_OUTPUT_MIPI)) { + switch (dev_priv->core_freq) { + case 100: + limit = &oaktrail_limits[MRST_LIMIT_LVDS_100L]; + break; + case 166: + limit = &oaktrail_limits[MRST_LIMIT_LVDS_83]; + break; + case 200: + limit = &oaktrail_limits[MRST_LIMIT_LVDS_100]; + break; + } + } else { + limit = NULL; + dev_err(dev->dev, "oaktrail_limit Wrong display type.\n"); + } + + return limit; +} + +/** Derive the pixel clock for the given refclk and divisors for 8xx chips. */ +static void oaktrail_clock(int refclk, struct oaktrail_clock_t *clock) +{ + clock->dot = (refclk * clock->m) / (14 * clock->p1); +} + +static void mrstPrintPll(char *prefix, struct oaktrail_clock_t *clock) +{ + pr_debug("%s: dotclock = %d, m = %d, p1 = %d.\n", + prefix, clock->dot, clock->m, clock->p1); +} + +/** + * Returns a set of divisors for the desired target clock with the given refclk, + * or FALSE. Divisor values are the actual divisors for + */ +static bool +mrstFindBestPLL(struct drm_crtc *crtc, int target, int refclk, + struct oaktrail_clock_t *best_clock) +{ + struct oaktrail_clock_t clock; + const struct oaktrail_limit_t *limit = oaktrail_limit(crtc); + int err = target; + + memset(best_clock, 0, sizeof(*best_clock)); + + for (clock.m = limit->m.min; clock.m <= limit->m.max; clock.m++) { + for (clock.p1 = limit->p1.min; clock.p1 <= limit->p1.max; + clock.p1++) { + int this_err; + + oaktrail_clock(refclk, &clock); + + this_err = abs(clock.dot - target); + if (this_err < err) { + *best_clock = clock; + err = this_err; + } + } + } + dev_dbg(crtc->dev->dev, "mrstFindBestPLL err = %d.\n", err); + return err != target; +} + +/** + * Sets the power management mode of the pipe and plane. + * + * This code should probably grow support for turning the cursor off and back + * on appropriately at the same time as we're turning the pipe off/on. + */ +static void oaktrail_crtc_dpms(struct drm_crtc *crtc, int mode) +{ + struct drm_device *dev = crtc->dev; + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + int pipe = psb_intel_crtc->pipe; + int dpll_reg = (pipe == 0) ? MRST_DPLL_A : DPLL_B; + int dspcntr_reg = (pipe == 0) ? DSPACNTR : DSPBCNTR; + int dspbase_reg = (pipe == 0) ? MRST_DSPABASE : DSPBBASE; + int pipeconf_reg = (pipe == 0) ? PIPEACONF : PIPEBCONF; + u32 temp; + + if (!gma_power_begin(dev, true)) + return; + + /* XXX: When our outputs are all unaware of DPMS modes other than off + * and on, we should map those modes to DRM_MODE_DPMS_OFF in the CRTC. + */ + switch (mode) { + case DRM_MODE_DPMS_ON: + case DRM_MODE_DPMS_STANDBY: + case DRM_MODE_DPMS_SUSPEND: + /* Enable the DPLL */ + temp = REG_READ(dpll_reg); + if ((temp & DPLL_VCO_ENABLE) == 0) { + REG_WRITE(dpll_reg, temp); + REG_READ(dpll_reg); + /* Wait for the clocks to stabilize. */ + udelay(150); + REG_WRITE(dpll_reg, temp | DPLL_VCO_ENABLE); + REG_READ(dpll_reg); + /* Wait for the clocks to stabilize. */ + udelay(150); + REG_WRITE(dpll_reg, temp | DPLL_VCO_ENABLE); + REG_READ(dpll_reg); + /* Wait for the clocks to stabilize. */ + udelay(150); + } + /* Enable the pipe */ + temp = REG_READ(pipeconf_reg); + if ((temp & PIPEACONF_ENABLE) == 0) + REG_WRITE(pipeconf_reg, temp | PIPEACONF_ENABLE); + /* Enable the plane */ + temp = REG_READ(dspcntr_reg); + if ((temp & DISPLAY_PLANE_ENABLE) == 0) { + REG_WRITE(dspcntr_reg, + temp | DISPLAY_PLANE_ENABLE); + /* Flush the plane changes */ + REG_WRITE(dspbase_reg, REG_READ(dspbase_reg)); + } + + psb_intel_crtc_load_lut(crtc); + + /* Give the overlay scaler a chance to enable + if it's on this pipe */ + /* psb_intel_crtc_dpms_video(crtc, true); TODO */ + break; + case DRM_MODE_DPMS_OFF: + /* Give the overlay scaler a chance to disable + * if it's on this pipe */ + /* psb_intel_crtc_dpms_video(crtc, FALSE); TODO */ + + /* Disable the VGA plane that we never use */ + REG_WRITE(VGACNTRL, VGA_DISP_DISABLE); + /* Disable display plane */ + temp = REG_READ(dspcntr_reg); + if ((temp & DISPLAY_PLANE_ENABLE) != 0) { + REG_WRITE(dspcntr_reg, + temp & ~DISPLAY_PLANE_ENABLE); + /* Flush the plane changes */ + REG_WRITE(dspbase_reg, REG_READ(dspbase_reg)); + REG_READ(dspbase_reg); + } + + /* Next, disable display pipes */ + temp = REG_READ(pipeconf_reg); + if ((temp & PIPEACONF_ENABLE) != 0) { + REG_WRITE(pipeconf_reg, temp & ~PIPEACONF_ENABLE); + REG_READ(pipeconf_reg); + } + /* Wait for for the pipe disable to take effect. */ + psb_intel_wait_for_vblank(dev); + + temp = REG_READ(dpll_reg); + if ((temp & DPLL_VCO_ENABLE) != 0) { + REG_WRITE(dpll_reg, temp & ~DPLL_VCO_ENABLE); + REG_READ(dpll_reg); + } + + /* Wait for the clocks to turn off. */ + udelay(150); + break; + } + + /*Set FIFO Watermarks*/ + REG_WRITE(DSPARB, 0x3FFF); + REG_WRITE(DSPFW1, 0x3F88080A); + REG_WRITE(DSPFW2, 0x0b060808); + REG_WRITE(DSPFW3, 0x0); + REG_WRITE(DSPFW4, 0x08030404); + REG_WRITE(DSPFW5, 0x04040404); + REG_WRITE(DSPFW6, 0x78); + REG_WRITE(0x70400, REG_READ(0x70400) | 0x4000); + /* Must write Bit 14 of the Chicken Bit Register */ + + gma_power_end(dev); +} + +/** + * Return the pipe currently connected to the panel fitter, + * or -1 if the panel fitter is not present or not in use + */ +static int oaktrail_panel_fitter_pipe(struct drm_device *dev) +{ + u32 pfit_control; + + pfit_control = REG_READ(PFIT_CONTROL); + + /* See if the panel fitter is in use */ + if ((pfit_control & PFIT_ENABLE) == 0) + return -1; + return (pfit_control >> 29) & 3; +} + +static int oaktrail_crtc_mode_set(struct drm_crtc *crtc, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode, + int x, int y, + struct drm_framebuffer *old_fb) +{ + struct drm_device *dev = crtc->dev; + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + struct drm_psb_private *dev_priv = dev->dev_private; + int pipe = psb_intel_crtc->pipe; + int fp_reg = (pipe == 0) ? MRST_FPA0 : FPB0; + int dpll_reg = (pipe == 0) ? MRST_DPLL_A : DPLL_B; + int dspcntr_reg = (pipe == 0) ? DSPACNTR : DSPBCNTR; + int pipeconf_reg = (pipe == 0) ? PIPEACONF : PIPEBCONF; + int htot_reg = (pipe == 0) ? HTOTAL_A : HTOTAL_B; + int hblank_reg = (pipe == 0) ? HBLANK_A : HBLANK_B; + int hsync_reg = (pipe == 0) ? HSYNC_A : HSYNC_B; + int vtot_reg = (pipe == 0) ? VTOTAL_A : VTOTAL_B; + int vblank_reg = (pipe == 0) ? VBLANK_A : VBLANK_B; + int vsync_reg = (pipe == 0) ? VSYNC_A : VSYNC_B; + int pipesrc_reg = (pipe == 0) ? PIPEASRC : PIPEBSRC; + int refclk = 0; + struct oaktrail_clock_t clock; + u32 dpll = 0, fp = 0, dspcntr, pipeconf; + bool ok, is_sdvo = false; + bool is_lvds = false; + bool is_mipi = false; + struct drm_mode_config *mode_config = &dev->mode_config; + struct psb_intel_encoder *psb_intel_encoder = NULL; + uint64_t scalingType = DRM_MODE_SCALE_FULLSCREEN; + struct drm_connector *connector; + + if (!gma_power_begin(dev, true)) + return 0; + + memcpy(&psb_intel_crtc->saved_mode, + mode, + sizeof(struct drm_display_mode)); + memcpy(&psb_intel_crtc->saved_adjusted_mode, + adjusted_mode, + sizeof(struct drm_display_mode)); + + list_for_each_entry(connector, &mode_config->connector_list, head) { + if (!connector->encoder || connector->encoder->crtc != crtc) + continue; + + psb_intel_encoder = psb_intel_attached_encoder(connector); + + switch (psb_intel_encoder->type) { + case INTEL_OUTPUT_LVDS: + is_lvds = true; + break; + case INTEL_OUTPUT_SDVO: + is_sdvo = true; + break; + case INTEL_OUTPUT_MIPI: + is_mipi = true; + break; + } + } + + /* Disable the VGA plane that we never use */ + REG_WRITE(VGACNTRL, VGA_DISP_DISABLE); + + /* Disable the panel fitter if it was on our pipe */ + if (oaktrail_panel_fitter_pipe(dev) == pipe) + REG_WRITE(PFIT_CONTROL, 0); + + REG_WRITE(pipesrc_reg, + ((mode->crtc_hdisplay - 1) << 16) | + (mode->crtc_vdisplay - 1)); + + if (psb_intel_encoder) + drm_connector_property_get_value(connector, + dev->mode_config.scaling_mode_property, &scalingType); + + if (scalingType == DRM_MODE_SCALE_NO_SCALE) { + /* Moorestown doesn't have register support for centering so + * we need to mess with the h/vblank and h/vsync start and + * ends to get centering */ + int offsetX = 0, offsetY = 0; + + offsetX = (adjusted_mode->crtc_hdisplay - + mode->crtc_hdisplay) / 2; + offsetY = (adjusted_mode->crtc_vdisplay - + mode->crtc_vdisplay) / 2; + + REG_WRITE(htot_reg, (mode->crtc_hdisplay - 1) | + ((adjusted_mode->crtc_htotal - 1) << 16)); + REG_WRITE(vtot_reg, (mode->crtc_vdisplay - 1) | + ((adjusted_mode->crtc_vtotal - 1) << 16)); + REG_WRITE(hblank_reg, + (adjusted_mode->crtc_hblank_start - offsetX - 1) | + ((adjusted_mode->crtc_hblank_end - offsetX - 1) << 16)); + REG_WRITE(hsync_reg, + (adjusted_mode->crtc_hsync_start - offsetX - 1) | + ((adjusted_mode->crtc_hsync_end - offsetX - 1) << 16)); + REG_WRITE(vblank_reg, + (adjusted_mode->crtc_vblank_start - offsetY - 1) | + ((adjusted_mode->crtc_vblank_end - offsetY - 1) << 16)); + REG_WRITE(vsync_reg, + (adjusted_mode->crtc_vsync_start - offsetY - 1) | + ((adjusted_mode->crtc_vsync_end - offsetY - 1) << 16)); + } else { + REG_WRITE(htot_reg, (adjusted_mode->crtc_hdisplay - 1) | + ((adjusted_mode->crtc_htotal - 1) << 16)); + REG_WRITE(vtot_reg, (adjusted_mode->crtc_vdisplay - 1) | + ((adjusted_mode->crtc_vtotal - 1) << 16)); + REG_WRITE(hblank_reg, (adjusted_mode->crtc_hblank_start - 1) | + ((adjusted_mode->crtc_hblank_end - 1) << 16)); + REG_WRITE(hsync_reg, (adjusted_mode->crtc_hsync_start - 1) | + ((adjusted_mode->crtc_hsync_end - 1) << 16)); + REG_WRITE(vblank_reg, (adjusted_mode->crtc_vblank_start - 1) | + ((adjusted_mode->crtc_vblank_end - 1) << 16)); + REG_WRITE(vsync_reg, (adjusted_mode->crtc_vsync_start - 1) | + ((adjusted_mode->crtc_vsync_end - 1) << 16)); + } + + /* Flush the plane changes */ + { + struct drm_crtc_helper_funcs *crtc_funcs = + crtc->helper_private; + crtc_funcs->mode_set_base(crtc, x, y, old_fb); + } + + /* setup pipeconf */ + pipeconf = REG_READ(pipeconf_reg); + + /* Set up the display plane register */ + dspcntr = REG_READ(dspcntr_reg); + dspcntr |= DISPPLANE_GAMMA_ENABLE; + + if (pipe == 0) + dspcntr |= DISPPLANE_SEL_PIPE_A; + else + dspcntr |= DISPPLANE_SEL_PIPE_B; + + if (is_mipi) + goto oaktrail_crtc_mode_set_exit; + + refclk = dev_priv->core_freq * 1000; + + dpll = 0; /*BIT16 = 0 for 100MHz reference */ + + ok = mrstFindBestPLL(crtc, adjusted_mode->clock, refclk, &clock); + + if (!ok) { + dev_dbg(dev->dev, "mrstFindBestPLL fail in oaktrail_crtc_mode_set.\n"); + } else { + dev_dbg(dev->dev, "oaktrail_crtc_mode_set pixel clock = %d," + "m = %x, p1 = %x.\n", clock.dot, clock.m, + clock.p1); + } + + fp = oaktrail_m_converts[(clock.m - MRST_M_MIN)] << 8; + + dpll |= DPLL_VGA_MODE_DIS; + + + dpll |= DPLL_VCO_ENABLE; + + if (is_lvds) + dpll |= DPLLA_MODE_LVDS; + else + dpll |= DPLLB_MODE_DAC_SERIAL; + + if (is_sdvo) { + int sdvo_pixel_multiply = + adjusted_mode->clock / mode->clock; + + dpll |= DPLL_DVO_HIGH_SPEED; + dpll |= + (sdvo_pixel_multiply - + 1) << SDVO_MULTIPLIER_SHIFT_HIRES; + } + + + /* compute bitmask from p1 value */ + dpll |= (1 << (clock.p1 - 2)) << 17; + + dpll |= DPLL_VCO_ENABLE; + + mrstPrintPll("chosen", &clock); + + if (dpll & DPLL_VCO_ENABLE) { + REG_WRITE(fp_reg, fp); + REG_WRITE(dpll_reg, dpll & ~DPLL_VCO_ENABLE); + REG_READ(dpll_reg); + /* Check the DPLLA lock bit PIPEACONF[29] */ + udelay(150); + } + + REG_WRITE(fp_reg, fp); + REG_WRITE(dpll_reg, dpll); + REG_READ(dpll_reg); + /* Wait for the clocks to stabilize. */ + udelay(150); + + /* write it again -- the BIOS does, after all */ + REG_WRITE(dpll_reg, dpll); + REG_READ(dpll_reg); + /* Wait for the clocks to stabilize. */ + udelay(150); + + REG_WRITE(pipeconf_reg, pipeconf); + REG_READ(pipeconf_reg); + psb_intel_wait_for_vblank(dev); + + REG_WRITE(dspcntr_reg, dspcntr); + psb_intel_wait_for_vblank(dev); + +oaktrail_crtc_mode_set_exit: + gma_power_end(dev); + return 0; +} + +static bool oaktrail_crtc_mode_fixup(struct drm_crtc *crtc, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + return true; +} + +static int oaktrail_pipe_set_base(struct drm_crtc *crtc, + int x, int y, struct drm_framebuffer *old_fb) +{ + struct drm_device *dev = crtc->dev; + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + struct psb_framebuffer *psbfb = to_psb_fb(crtc->fb); + int pipe = psb_intel_crtc->pipe; + unsigned long start, offset; + + int dspbase = (pipe == 0 ? DSPALINOFF : DSPBBASE); + int dspsurf = (pipe == 0 ? DSPASURF : DSPBSURF); + int dspstride = (pipe == 0) ? DSPASTRIDE : DSPBSTRIDE; + int dspcntr_reg = (pipe == 0) ? DSPACNTR : DSPBCNTR; + u32 dspcntr; + int ret = 0; + + /* no fb bound */ + if (!crtc->fb) { + dev_dbg(dev->dev, "No FB bound\n"); + return 0; + } + + if (!gma_power_begin(dev, true)) + return 0; + + start = psbfb->gtt->offset; + offset = y * crtc->fb->pitches[0] + x * (crtc->fb->bits_per_pixel / 8); + + REG_WRITE(dspstride, crtc->fb->pitches[0]); + + dspcntr = REG_READ(dspcntr_reg); + dspcntr &= ~DISPPLANE_PIXFORMAT_MASK; + + switch (crtc->fb->bits_per_pixel) { + case 8: + dspcntr |= DISPPLANE_8BPP; + break; + case 16: + if (crtc->fb->depth == 15) + dspcntr |= DISPPLANE_15_16BPP; + else + dspcntr |= DISPPLANE_16BPP; + break; + case 24: + case 32: + dspcntr |= DISPPLANE_32BPP_NO_ALPHA; + break; + default: + dev_err(dev->dev, "Unknown color depth\n"); + ret = -EINVAL; + goto pipe_set_base_exit; + } + REG_WRITE(dspcntr_reg, dspcntr); + + REG_WRITE(dspbase, offset); + REG_READ(dspbase); + REG_WRITE(dspsurf, start); + REG_READ(dspsurf); + +pipe_set_base_exit: + gma_power_end(dev); + return ret; +} + +static void oaktrail_crtc_prepare(struct drm_crtc *crtc) +{ + struct drm_crtc_helper_funcs *crtc_funcs = crtc->helper_private; + crtc_funcs->dpms(crtc, DRM_MODE_DPMS_OFF); +} + +static void oaktrail_crtc_commit(struct drm_crtc *crtc) +{ + struct drm_crtc_helper_funcs *crtc_funcs = crtc->helper_private; + crtc_funcs->dpms(crtc, DRM_MODE_DPMS_ON); +} + +const struct drm_crtc_helper_funcs oaktrail_helper_funcs = { + .dpms = oaktrail_crtc_dpms, + .mode_fixup = oaktrail_crtc_mode_fixup, + .mode_set = oaktrail_crtc_mode_set, + .mode_set_base = oaktrail_pipe_set_base, + .prepare = oaktrail_crtc_prepare, + .commit = oaktrail_crtc_commit, +}; + diff --git a/drivers/gpu/drm/gma500/oaktrail_device.c b/drivers/gpu/drm/gma500/oaktrail_device.c new file mode 100644 index 0000000..41d1924 --- /dev/null +++ b/drivers/gpu/drm/gma500/oaktrail_device.c @@ -0,0 +1,509 @@ +/************************************************************************** + * Copyright (c) 2011, Intel Corporation. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + **************************************************************************/ + +#include +#include +#include +#include +#include +#include "gma_drm.h" +#include "psb_drv.h" +#include "psb_reg.h" +#include "psb_intel_reg.h" +#include +#include +#include "mid_bios.h" +#include "intel_bios.h" + +static int oaktrail_output_init(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + if (dev_priv->iLVDS_enable) + oaktrail_lvds_init(dev, &dev_priv->mode_dev); + else + dev_err(dev->dev, "DSI is not supported\n"); + if (dev_priv->hdmi_priv) + oaktrail_hdmi_init(dev, &dev_priv->mode_dev); + return 0; +} + +/* + * Provide the low level interfaces for the Moorestown backlight + */ + +#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE + +#define MRST_BLC_MAX_PWM_REG_FREQ 0xFFFF +#define BLC_PWM_PRECISION_FACTOR 100 /* 10000000 */ +#define BLC_PWM_FREQ_CALC_CONSTANT 32 +#define MHz 1000000 +#define BLC_ADJUSTMENT_MAX 100 + +static struct backlight_device *oaktrail_backlight_device; +static int oaktrail_brightness; + +static int oaktrail_set_brightness(struct backlight_device *bd) +{ + struct drm_device *dev = bl_get_data(oaktrail_backlight_device); + struct drm_psb_private *dev_priv = dev->dev_private; + int level = bd->props.brightness; + u32 blc_pwm_ctl; + u32 max_pwm_blc; + + /* Percentage 1-100% being valid */ + if (level < 1) + level = 1; + + if (gma_power_begin(dev, 0)) { + /* Calculate and set the brightness value */ + max_pwm_blc = REG_READ(BLC_PWM_CTL) >> 16; + blc_pwm_ctl = level * max_pwm_blc / 100; + + /* Adjust the backlight level with the percent in + * dev_priv->blc_adj1; + */ + blc_pwm_ctl = blc_pwm_ctl * dev_priv->blc_adj1; + blc_pwm_ctl = blc_pwm_ctl / 100; + + /* Adjust the backlight level with the percent in + * dev_priv->blc_adj2; + */ + blc_pwm_ctl = blc_pwm_ctl * dev_priv->blc_adj2; + blc_pwm_ctl = blc_pwm_ctl / 100; + + /* force PWM bit on */ + REG_WRITE(BLC_PWM_CTL2, (0x80000000 | REG_READ(BLC_PWM_CTL2))); + REG_WRITE(BLC_PWM_CTL, (max_pwm_blc << 16) | blc_pwm_ctl); + gma_power_end(dev); + } + oaktrail_brightness = level; + return 0; +} + +static int oaktrail_get_brightness(struct backlight_device *bd) +{ + /* return locally cached var instead of HW read (due to DPST etc.) */ + /* FIXME: ideally return actual value in case firmware fiddled with + it */ + return oaktrail_brightness; +} + +static int device_backlight_init(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + unsigned long core_clock; + u16 bl_max_freq; + uint32_t value; + uint32_t blc_pwm_precision_factor; + + dev_priv->blc_adj1 = BLC_ADJUSTMENT_MAX; + dev_priv->blc_adj2 = BLC_ADJUSTMENT_MAX; + bl_max_freq = 256; + /* this needs to be set elsewhere */ + blc_pwm_precision_factor = BLC_PWM_PRECISION_FACTOR; + + core_clock = dev_priv->core_freq; + + value = (core_clock * MHz) / BLC_PWM_FREQ_CALC_CONSTANT; + value *= blc_pwm_precision_factor; + value /= bl_max_freq; + value /= blc_pwm_precision_factor; + + if (value > (unsigned long long)MRST_BLC_MAX_PWM_REG_FREQ) + return -ERANGE; + + if (gma_power_begin(dev, false)) { + REG_WRITE(BLC_PWM_CTL2, (0x80000000 | REG_READ(BLC_PWM_CTL2))); + REG_WRITE(BLC_PWM_CTL, value | (value << 16)); + gma_power_end(dev); + } + return 0; +} + +static const struct backlight_ops oaktrail_ops = { + .get_brightness = oaktrail_get_brightness, + .update_status = oaktrail_set_brightness, +}; + +static int oaktrail_backlight_init(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + int ret; + struct backlight_properties props; + + memset(&props, 0, sizeof(struct backlight_properties)); + props.max_brightness = 100; + props.type = BACKLIGHT_PLATFORM; + + oaktrail_backlight_device = backlight_device_register("oaktrail-bl", + NULL, (void *)dev, &oaktrail_ops, &props); + + if (IS_ERR(oaktrail_backlight_device)) + return PTR_ERR(oaktrail_backlight_device); + + ret = device_backlight_init(dev); + if (ret < 0) { + backlight_device_unregister(oaktrail_backlight_device); + return ret; + } + oaktrail_backlight_device->props.brightness = 100; + oaktrail_backlight_device->props.max_brightness = 100; + backlight_update_status(oaktrail_backlight_device); + dev_priv->backlight_device = oaktrail_backlight_device; + return 0; +} + +#endif + +/* + * Provide the Moorestown specific chip logic and low level methods + * for power management + */ + +/** + * oaktrail_save_display_registers - save registers lost on suspend + * @dev: our DRM device + * + * Save the state we need in order to be able to restore the interface + * upon resume from suspend + */ +static int oaktrail_save_display_registers(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_save_area *regs = &dev_priv->regs; + int i; + u32 pp_stat; + + /* Display arbitration control + watermarks */ + regs->psb.saveDSPARB = PSB_RVDC32(DSPARB); + regs->psb.saveDSPFW1 = PSB_RVDC32(DSPFW1); + regs->psb.saveDSPFW2 = PSB_RVDC32(DSPFW2); + regs->psb.saveDSPFW3 = PSB_RVDC32(DSPFW3); + regs->psb.saveDSPFW4 = PSB_RVDC32(DSPFW4); + regs->psb.saveDSPFW5 = PSB_RVDC32(DSPFW5); + regs->psb.saveDSPFW6 = PSB_RVDC32(DSPFW6); + regs->psb.saveCHICKENBIT = PSB_RVDC32(DSPCHICKENBIT); + + /* Pipe & plane A info */ + regs->psb.savePIPEACONF = PSB_RVDC32(PIPEACONF); + regs->psb.savePIPEASRC = PSB_RVDC32(PIPEASRC); + regs->psb.saveFPA0 = PSB_RVDC32(MRST_FPA0); + regs->psb.saveFPA1 = PSB_RVDC32(MRST_FPA1); + regs->psb.saveDPLL_A = PSB_RVDC32(MRST_DPLL_A); + regs->psb.saveHTOTAL_A = PSB_RVDC32(HTOTAL_A); + regs->psb.saveHBLANK_A = PSB_RVDC32(HBLANK_A); + regs->psb.saveHSYNC_A = PSB_RVDC32(HSYNC_A); + regs->psb.saveVTOTAL_A = PSB_RVDC32(VTOTAL_A); + regs->psb.saveVBLANK_A = PSB_RVDC32(VBLANK_A); + regs->psb.saveVSYNC_A = PSB_RVDC32(VSYNC_A); + regs->psb.saveBCLRPAT_A = PSB_RVDC32(BCLRPAT_A); + regs->psb.saveDSPACNTR = PSB_RVDC32(DSPACNTR); + regs->psb.saveDSPASTRIDE = PSB_RVDC32(DSPASTRIDE); + regs->psb.saveDSPAADDR = PSB_RVDC32(DSPABASE); + regs->psb.saveDSPASURF = PSB_RVDC32(DSPASURF); + regs->psb.saveDSPALINOFF = PSB_RVDC32(DSPALINOFF); + regs->psb.saveDSPATILEOFF = PSB_RVDC32(DSPATILEOFF); + + /* Save cursor regs */ + regs->psb.saveDSPACURSOR_CTRL = PSB_RVDC32(CURACNTR); + regs->psb.saveDSPACURSOR_BASE = PSB_RVDC32(CURABASE); + regs->psb.saveDSPACURSOR_POS = PSB_RVDC32(CURAPOS); + + /* Save palette (gamma) */ + for (i = 0; i < 256; i++) + regs->psb.save_palette_a[i] = PSB_RVDC32(PALETTE_A + (i << 2)); + + if (dev_priv->hdmi_priv) + oaktrail_hdmi_save(dev); + + /* Save performance state */ + regs->psb.savePERF_MODE = PSB_RVDC32(MRST_PERF_MODE); + + /* LVDS state */ + regs->psb.savePP_CONTROL = PSB_RVDC32(PP_CONTROL); + regs->psb.savePFIT_PGM_RATIOS = PSB_RVDC32(PFIT_PGM_RATIOS); + regs->psb.savePFIT_AUTO_RATIOS = PSB_RVDC32(PFIT_AUTO_RATIOS); + regs->saveBLC_PWM_CTL = PSB_RVDC32(BLC_PWM_CTL); + regs->saveBLC_PWM_CTL2 = PSB_RVDC32(BLC_PWM_CTL2); + regs->psb.saveLVDS = PSB_RVDC32(LVDS); + regs->psb.savePFIT_CONTROL = PSB_RVDC32(PFIT_CONTROL); + regs->psb.savePP_ON_DELAYS = PSB_RVDC32(LVDSPP_ON); + regs->psb.savePP_OFF_DELAYS = PSB_RVDC32(LVDSPP_OFF); + regs->psb.savePP_DIVISOR = PSB_RVDC32(PP_CYCLE); + + /* HW overlay */ + regs->psb.saveOV_OVADD = PSB_RVDC32(OV_OVADD); + regs->psb.saveOV_OGAMC0 = PSB_RVDC32(OV_OGAMC0); + regs->psb.saveOV_OGAMC1 = PSB_RVDC32(OV_OGAMC1); + regs->psb.saveOV_OGAMC2 = PSB_RVDC32(OV_OGAMC2); + regs->psb.saveOV_OGAMC3 = PSB_RVDC32(OV_OGAMC3); + regs->psb.saveOV_OGAMC4 = PSB_RVDC32(OV_OGAMC4); + regs->psb.saveOV_OGAMC5 = PSB_RVDC32(OV_OGAMC5); + + /* DPST registers */ + regs->psb.saveHISTOGRAM_INT_CONTROL_REG = + PSB_RVDC32(HISTOGRAM_INT_CONTROL); + regs->psb.saveHISTOGRAM_LOGIC_CONTROL_REG = + PSB_RVDC32(HISTOGRAM_LOGIC_CONTROL); + regs->psb.savePWM_CONTROL_LOGIC = PSB_RVDC32(PWM_CONTROL_LOGIC); + + if (dev_priv->iLVDS_enable) { + /* Shut down the panel */ + PSB_WVDC32(0, PP_CONTROL); + + do { + pp_stat = PSB_RVDC32(PP_STATUS); + } while (pp_stat & 0x80000000); + + /* Turn off the plane */ + PSB_WVDC32(0x58000000, DSPACNTR); + /* Trigger the plane disable */ + PSB_WVDC32(0, DSPASURF); + + /* Wait ~4 ticks */ + msleep(4); + + /* Turn off pipe */ + PSB_WVDC32(0x0, PIPEACONF); + /* Wait ~8 ticks */ + msleep(8); + + /* Turn off PLLs */ + PSB_WVDC32(0, MRST_DPLL_A); + } + return 0; +} + +/** + * oaktrail_restore_display_registers - restore lost register state + * @dev: our DRM device + * + * Restore register state that was lost during suspend and resume. + */ +static int oaktrail_restore_display_registers(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_save_area *regs = &dev_priv->regs; + u32 pp_stat; + int i; + + /* Display arbitration + watermarks */ + PSB_WVDC32(regs->psb.saveDSPARB, DSPARB); + PSB_WVDC32(regs->psb.saveDSPFW1, DSPFW1); + PSB_WVDC32(regs->psb.saveDSPFW2, DSPFW2); + PSB_WVDC32(regs->psb.saveDSPFW3, DSPFW3); + PSB_WVDC32(regs->psb.saveDSPFW4, DSPFW4); + PSB_WVDC32(regs->psb.saveDSPFW5, DSPFW5); + PSB_WVDC32(regs->psb.saveDSPFW6, DSPFW6); + PSB_WVDC32(regs->psb.saveCHICKENBIT, DSPCHICKENBIT); + + /* Make sure VGA plane is off. it initializes to on after reset!*/ + PSB_WVDC32(0x80000000, VGACNTRL); + + /* set the plls */ + PSB_WVDC32(regs->psb.saveFPA0, MRST_FPA0); + PSB_WVDC32(regs->psb.saveFPA1, MRST_FPA1); + + /* Actually enable it */ + PSB_WVDC32(regs->psb.saveDPLL_A, MRST_DPLL_A); + DRM_UDELAY(150); + + /* Restore mode */ + PSB_WVDC32(regs->psb.saveHTOTAL_A, HTOTAL_A); + PSB_WVDC32(regs->psb.saveHBLANK_A, HBLANK_A); + PSB_WVDC32(regs->psb.saveHSYNC_A, HSYNC_A); + PSB_WVDC32(regs->psb.saveVTOTAL_A, VTOTAL_A); + PSB_WVDC32(regs->psb.saveVBLANK_A, VBLANK_A); + PSB_WVDC32(regs->psb.saveVSYNC_A, VSYNC_A); + PSB_WVDC32(regs->psb.savePIPEASRC, PIPEASRC); + PSB_WVDC32(regs->psb.saveBCLRPAT_A, BCLRPAT_A); + + /* Restore performance mode*/ + PSB_WVDC32(regs->psb.savePERF_MODE, MRST_PERF_MODE); + + /* Enable the pipe*/ + if (dev_priv->iLVDS_enable) + PSB_WVDC32(regs->psb.savePIPEACONF, PIPEACONF); + + /* Set up the plane*/ + PSB_WVDC32(regs->psb.saveDSPALINOFF, DSPALINOFF); + PSB_WVDC32(regs->psb.saveDSPASTRIDE, DSPASTRIDE); + PSB_WVDC32(regs->psb.saveDSPATILEOFF, DSPATILEOFF); + + /* Enable the plane */ + PSB_WVDC32(regs->psb.saveDSPACNTR, DSPACNTR); + PSB_WVDC32(regs->psb.saveDSPASURF, DSPASURF); + + /* Enable Cursor A */ + PSB_WVDC32(regs->psb.saveDSPACURSOR_CTRL, CURACNTR); + PSB_WVDC32(regs->psb.saveDSPACURSOR_POS, CURAPOS); + PSB_WVDC32(regs->psb.saveDSPACURSOR_BASE, CURABASE); + + /* Restore palette (gamma) */ + for (i = 0; i < 256; i++) + PSB_WVDC32(regs->psb.save_palette_a[i], PALETTE_A + (i << 2)); + + if (dev_priv->hdmi_priv) + oaktrail_hdmi_restore(dev); + + if (dev_priv->iLVDS_enable) { + PSB_WVDC32(regs->saveBLC_PWM_CTL2, BLC_PWM_CTL2); + PSB_WVDC32(regs->psb.saveLVDS, LVDS); /*port 61180h*/ + PSB_WVDC32(regs->psb.savePFIT_CONTROL, PFIT_CONTROL); + PSB_WVDC32(regs->psb.savePFIT_PGM_RATIOS, PFIT_PGM_RATIOS); + PSB_WVDC32(regs->psb.savePFIT_AUTO_RATIOS, PFIT_AUTO_RATIOS); + PSB_WVDC32(regs->saveBLC_PWM_CTL, BLC_PWM_CTL); + PSB_WVDC32(regs->psb.savePP_ON_DELAYS, LVDSPP_ON); + PSB_WVDC32(regs->psb.savePP_OFF_DELAYS, LVDSPP_OFF); + PSB_WVDC32(regs->psb.savePP_DIVISOR, PP_CYCLE); + PSB_WVDC32(regs->psb.savePP_CONTROL, PP_CONTROL); + } + + /* Wait for cycle delay */ + do { + pp_stat = PSB_RVDC32(PP_STATUS); + } while (pp_stat & 0x08000000); + + /* Wait for panel power up */ + do { + pp_stat = PSB_RVDC32(PP_STATUS); + } while (pp_stat & 0x10000000); + + /* Restore HW overlay */ + PSB_WVDC32(regs->psb.saveOV_OVADD, OV_OVADD); + PSB_WVDC32(regs->psb.saveOV_OGAMC0, OV_OGAMC0); + PSB_WVDC32(regs->psb.saveOV_OGAMC1, OV_OGAMC1); + PSB_WVDC32(regs->psb.saveOV_OGAMC2, OV_OGAMC2); + PSB_WVDC32(regs->psb.saveOV_OGAMC3, OV_OGAMC3); + PSB_WVDC32(regs->psb.saveOV_OGAMC4, OV_OGAMC4); + PSB_WVDC32(regs->psb.saveOV_OGAMC5, OV_OGAMC5); + + /* DPST registers */ + PSB_WVDC32(regs->psb.saveHISTOGRAM_INT_CONTROL_REG, + HISTOGRAM_INT_CONTROL); + PSB_WVDC32(regs->psb.saveHISTOGRAM_LOGIC_CONTROL_REG, + HISTOGRAM_LOGIC_CONTROL); + PSB_WVDC32(regs->psb.savePWM_CONTROL_LOGIC, PWM_CONTROL_LOGIC); + + return 0; +} + +/** + * oaktrail_power_down - power down the display island + * @dev: our DRM device + * + * Power down the display interface of our device + */ +static int oaktrail_power_down(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + u32 pwr_mask ; + u32 pwr_sts; + + pwr_mask = PSB_PWRGT_DISPLAY_MASK; + outl(pwr_mask, dev_priv->ospm_base + PSB_PM_SSC); + + while (true) { + pwr_sts = inl(dev_priv->ospm_base + PSB_PM_SSS); + if ((pwr_sts & pwr_mask) == pwr_mask) + break; + else + udelay(10); + } + return 0; +} + +/* + * oaktrail_power_up + * + * Restore power to the specified island(s) (powergating) + */ +static int oaktrail_power_up(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + u32 pwr_mask = PSB_PWRGT_DISPLAY_MASK; + u32 pwr_sts, pwr_cnt; + + pwr_cnt = inl(dev_priv->ospm_base + PSB_PM_SSC); + pwr_cnt &= ~pwr_mask; + outl(pwr_cnt, (dev_priv->ospm_base + PSB_PM_SSC)); + + while (true) { + pwr_sts = inl(dev_priv->ospm_base + PSB_PM_SSS); + if ((pwr_sts & pwr_mask) == 0) + break; + else + udelay(10); + } + return 0; +} + + +static int oaktrail_chip_setup(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct oaktrail_vbt *vbt = &dev_priv->vbt_data; + int ret; + + ret = mid_chip_setup(dev); + if (ret < 0) + return ret; + if (vbt->size == 0) { + /* Now pull the BIOS data */ + gma_intel_opregion_init(dev); + psb_intel_init_bios(dev); + } + return 0; +} + +static void oaktrail_teardown(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct oaktrail_vbt *vbt = &dev_priv->vbt_data; + + oaktrail_hdmi_teardown(dev); + if (vbt->size == 0) + psb_intel_destroy_bios(dev); +} + +const struct psb_ops oaktrail_chip_ops = { + .name = "Oaktrail", + .accel_2d = 1, + .pipes = 2, + .crtcs = 2, + .sgx_offset = MRST_SGX_OFFSET, + + .chip_setup = oaktrail_chip_setup, + .chip_teardown = oaktrail_teardown, + .crtc_helper = &oaktrail_helper_funcs, + .crtc_funcs = &psb_intel_crtc_funcs, + + .output_init = oaktrail_output_init, + +#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE + .backlight_init = oaktrail_backlight_init, +#endif + + .save_regs = oaktrail_save_display_registers, + .restore_regs = oaktrail_restore_display_registers, + .power_down = oaktrail_power_down, + .power_up = oaktrail_power_up, + + .i2c_bus = 1, +}; diff --git a/drivers/gpu/drm/gma500/oaktrail_hdmi.c b/drivers/gpu/drm/gma500/oaktrail_hdmi.c new file mode 100644 index 0000000..f8b367b --- /dev/null +++ b/drivers/gpu/drm/gma500/oaktrail_hdmi.c @@ -0,0 +1,540 @@ +/* + * Copyright © 2010 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * Li Peng + */ + +#include +#include +#include "psb_intel_drv.h" +#include "psb_intel_reg.h" +#include "psb_drv.h" + +#define HDMI_READ(reg) readl(hdmi_dev->regs + (reg)) +#define HDMI_WRITE(reg, val) writel(val, hdmi_dev->regs + (reg)) + +#define HDMI_HCR 0x1000 +#define HCR_ENABLE_HDCP (1 << 5) +#define HCR_ENABLE_AUDIO (1 << 2) +#define HCR_ENABLE_PIXEL (1 << 1) +#define HCR_ENABLE_TMDS (1 << 0) + +#define HDMI_HICR 0x1004 +#define HDMI_HSR 0x1008 +#define HDMI_HISR 0x100C +#define HDMI_DETECT_HDP (1 << 0) + +#define HDMI_VIDEO_REG 0x3000 +#define HDMI_UNIT_EN (1 << 7) +#define HDMI_MODE_OUTPUT (1 << 0) +#define HDMI_HBLANK_A 0x3100 + +#define HDMI_AUDIO_CTRL 0x4000 +#define HDMI_ENABLE_AUDIO (1 << 0) + +#define PCH_HTOTAL_B 0x3100 +#define PCH_HBLANK_B 0x3104 +#define PCH_HSYNC_B 0x3108 +#define PCH_VTOTAL_B 0x310C +#define PCH_VBLANK_B 0x3110 +#define PCH_VSYNC_B 0x3114 +#define PCH_PIPEBSRC 0x311C + +#define PCH_PIPEB_DSL 0x3800 +#define PCH_PIPEB_SLC 0x3804 +#define PCH_PIPEBCONF 0x3808 +#define PCH_PIPEBSTAT 0x3824 + +#define CDVO_DFT 0x5000 +#define CDVO_SLEWRATE 0x5004 +#define CDVO_STRENGTH 0x5008 +#define CDVO_RCOMP 0x500C + +#define DPLL_CTRL 0x6000 +#define DPLL_PDIV_SHIFT 16 +#define DPLL_PDIV_MASK (0xf << 16) +#define DPLL_PWRDN (1 << 4) +#define DPLL_RESET (1 << 3) +#define DPLL_FASTEN (1 << 2) +#define DPLL_ENSTAT (1 << 1) +#define DPLL_DITHEN (1 << 0) + +#define DPLL_DIV_CTRL 0x6004 +#define DPLL_CLKF_MASK 0xffffffc0 +#define DPLL_CLKR_MASK (0x3f) + +#define DPLL_CLK_ENABLE 0x6008 +#define DPLL_EN_DISP (1 << 31) +#define DPLL_SEL_HDMI (1 << 8) +#define DPLL_EN_HDMI (1 << 1) +#define DPLL_EN_VGA (1 << 0) + +#define DPLL_ADJUST 0x600C +#define DPLL_STATUS 0x6010 +#define DPLL_UPDATE 0x6014 +#define DPLL_DFT 0x6020 + +struct intel_range { + int min, max; +}; + +struct oaktrail_hdmi_limit { + struct intel_range vco, np, nr, nf; +}; + +struct oaktrail_hdmi_clock { + int np; + int nr; + int nf; + int dot; +}; + +#define VCO_MIN 320000 +#define VCO_MAX 1650000 +#define NP_MIN 1 +#define NP_MAX 15 +#define NR_MIN 1 +#define NR_MAX 64 +#define NF_MIN 2 +#define NF_MAX 4095 + +static const struct oaktrail_hdmi_limit oaktrail_hdmi_limit = { + .vco = { .min = VCO_MIN, .max = VCO_MAX }, + .np = { .min = NP_MIN, .max = NP_MAX }, + .nr = { .min = NR_MIN, .max = NR_MAX }, + .nf = { .min = NF_MIN, .max = NF_MAX }, +}; + +static void oaktrail_hdmi_audio_enable(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct oaktrail_hdmi_dev *hdmi_dev = dev_priv->hdmi_priv; + + HDMI_WRITE(HDMI_HCR, 0x67); + HDMI_READ(HDMI_HCR); + + HDMI_WRITE(0x51a8, 0x10); + HDMI_READ(0x51a8); + + HDMI_WRITE(HDMI_AUDIO_CTRL, 0x1); + HDMI_READ(HDMI_AUDIO_CTRL); +} + +static void oaktrail_hdmi_audio_disable(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct oaktrail_hdmi_dev *hdmi_dev = dev_priv->hdmi_priv; + + HDMI_WRITE(0x51a8, 0x0); + HDMI_READ(0x51a8); + + HDMI_WRITE(HDMI_AUDIO_CTRL, 0x0); + HDMI_READ(HDMI_AUDIO_CTRL); + + HDMI_WRITE(HDMI_HCR, 0x47); + HDMI_READ(HDMI_HCR); +} + +static void oaktrail_hdmi_dpms(struct drm_encoder *encoder, int mode) +{ + static int dpms_mode = -1; + + struct drm_device *dev = encoder->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + struct oaktrail_hdmi_dev *hdmi_dev = dev_priv->hdmi_priv; + u32 temp; + + if (dpms_mode == mode) + return; + + if (mode != DRM_MODE_DPMS_ON) + temp = 0x0; + else + temp = 0x99; + + dpms_mode = mode; + HDMI_WRITE(HDMI_VIDEO_REG, temp); +} + +static int oaktrail_hdmi_mode_valid(struct drm_connector *connector, + struct drm_display_mode *mode) +{ + struct drm_psb_private *dev_priv = connector->dev->dev_private; + if (mode->clock > 165000) + return MODE_CLOCK_HIGH; + if (mode->clock < 20000) + return MODE_CLOCK_LOW; + + if (mode->flags & DRM_MODE_FLAG_DBLSCAN) + return MODE_NO_DBLESCAN; + + /* We assume worst case scenario of 32 bpp here, since we don't know */ + if ((ALIGN(mode->hdisplay * 4, 64) * mode->vdisplay) > + dev_priv->vram_stolen_size) + return MODE_MEM; + + return MODE_OK; +} + +static bool oaktrail_hdmi_mode_fixup(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + return true; +} + +static enum drm_connector_status +oaktrail_hdmi_detect(struct drm_connector *connector, bool force) +{ + enum drm_connector_status status; + struct drm_device *dev = connector->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + struct oaktrail_hdmi_dev *hdmi_dev = dev_priv->hdmi_priv; + u32 temp; + + temp = HDMI_READ(HDMI_HSR); + DRM_DEBUG_KMS("HDMI_HSR %x\n", temp); + + if ((temp & HDMI_DETECT_HDP) != 0) + status = connector_status_connected; + else + status = connector_status_disconnected; + + return status; +} + +static const unsigned char raw_edid[] = { + 0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00, 0x10, 0xac, 0x2f, 0xa0, + 0x53, 0x55, 0x33, 0x30, 0x16, 0x13, 0x01, 0x03, 0x0e, 0x3a, 0x24, 0x78, + 0xea, 0xe9, 0xf5, 0xac, 0x51, 0x30, 0xb4, 0x25, 0x11, 0x50, 0x54, 0xa5, + 0x4b, 0x00, 0x81, 0x80, 0xa9, 0x40, 0x71, 0x4f, 0xb3, 0x00, 0x01, 0x01, + 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x28, 0x3c, 0x80, 0xa0, 0x70, 0xb0, + 0x23, 0x40, 0x30, 0x20, 0x36, 0x00, 0x46, 0x6c, 0x21, 0x00, 0x00, 0x1a, + 0x00, 0x00, 0x00, 0xff, 0x00, 0x47, 0x4e, 0x37, 0x32, 0x31, 0x39, 0x35, + 0x52, 0x30, 0x33, 0x55, 0x53, 0x0a, 0x00, 0x00, 0x00, 0xfc, 0x00, 0x44, + 0x45, 0x4c, 0x4c, 0x20, 0x32, 0x37, 0x30, 0x39, 0x57, 0x0a, 0x20, 0x20, + 0x00, 0x00, 0x00, 0xfd, 0x00, 0x38, 0x4c, 0x1e, 0x53, 0x11, 0x00, 0x0a, + 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x00, 0x8d +}; + +static int oaktrail_hdmi_get_modes(struct drm_connector *connector) +{ + struct drm_device *dev = connector->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + struct i2c_adapter *i2c_adap; + struct edid *edid; + struct drm_display_mode *mode, *t; + int i = 0, ret = 0; + + i2c_adap = i2c_get_adapter(3); + if (i2c_adap == NULL) { + DRM_ERROR("No ddc adapter available!\n"); + edid = (struct edid *)raw_edid; + } else { + edid = (struct edid *)raw_edid; + /* FIXME ? edid = drm_get_edid(connector, i2c_adap); */ + } + + if (edid) { + drm_mode_connector_update_edid_property(connector, edid); + ret = drm_add_edid_modes(connector, edid); + connector->display_info.raw_edid = NULL; + } + + /* + * prune modes that require frame buffer bigger than stolen mem + */ + list_for_each_entry_safe(mode, t, &connector->probed_modes, head) { + if ((mode->hdisplay * mode->vdisplay * 4) >= dev_priv->vram_stolen_size) { + i++; + drm_mode_remove(connector, mode); + } + } + return ret - i; +} + +static void oaktrail_hdmi_mode_set(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + struct drm_device *dev = encoder->dev; + + oaktrail_hdmi_audio_enable(dev); + return; +} + +static void oaktrail_hdmi_destroy(struct drm_connector *connector) +{ + return; +} + +static const struct drm_encoder_helper_funcs oaktrail_hdmi_helper_funcs = { + .dpms = oaktrail_hdmi_dpms, + .mode_fixup = oaktrail_hdmi_mode_fixup, + .prepare = psb_intel_encoder_prepare, + .mode_set = oaktrail_hdmi_mode_set, + .commit = psb_intel_encoder_commit, +}; + +static const struct drm_connector_helper_funcs + oaktrail_hdmi_connector_helper_funcs = { + .get_modes = oaktrail_hdmi_get_modes, + .mode_valid = oaktrail_hdmi_mode_valid, + .best_encoder = psb_intel_best_encoder, +}; + +static const struct drm_connector_funcs oaktrail_hdmi_connector_funcs = { + .dpms = drm_helper_connector_dpms, + .detect = oaktrail_hdmi_detect, + .fill_modes = drm_helper_probe_single_connector_modes, + .destroy = oaktrail_hdmi_destroy, +}; + +static void oaktrail_hdmi_enc_destroy(struct drm_encoder *encoder) +{ + drm_encoder_cleanup(encoder); +} + +static const struct drm_encoder_funcs oaktrail_hdmi_enc_funcs = { + .destroy = oaktrail_hdmi_enc_destroy, +}; + +void oaktrail_hdmi_init(struct drm_device *dev, + struct psb_intel_mode_device *mode_dev) +{ + struct psb_intel_encoder *psb_intel_encoder; + struct psb_intel_connector *psb_intel_connector; + struct drm_connector *connector; + struct drm_encoder *encoder; + + psb_intel_encoder = kzalloc(sizeof(struct psb_intel_encoder), GFP_KERNEL); + if (!psb_intel_encoder) + return; + + psb_intel_connector = kzalloc(sizeof(struct psb_intel_connector), GFP_KERNEL); + if (!psb_intel_connector) + goto failed_connector; + + connector = &psb_intel_connector->base; + encoder = &psb_intel_encoder->base; + drm_connector_init(dev, connector, + &oaktrail_hdmi_connector_funcs, + DRM_MODE_CONNECTOR_DVID); + + drm_encoder_init(dev, encoder, + &oaktrail_hdmi_enc_funcs, + DRM_MODE_ENCODER_TMDS); + + psb_intel_connector_attach_encoder(psb_intel_connector, + psb_intel_encoder); + + psb_intel_encoder->type = INTEL_OUTPUT_HDMI; + drm_encoder_helper_add(encoder, &oaktrail_hdmi_helper_funcs); + drm_connector_helper_add(connector, &oaktrail_hdmi_connector_helper_funcs); + + connector->display_info.subpixel_order = SubPixelHorizontalRGB; + connector->interlace_allowed = false; + connector->doublescan_allowed = false; + drm_sysfs_connector_add(connector); + + return; + +failed_connector: + kfree(psb_intel_encoder); +} + +static DEFINE_PCI_DEVICE_TABLE(hdmi_ids) = { + { PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x080d) }, + { 0 } +}; + +void oaktrail_hdmi_setup(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct pci_dev *pdev; + struct oaktrail_hdmi_dev *hdmi_dev; + int ret; + + pdev = pci_get_device(PCI_VENDOR_ID_INTEL, 0x080d, NULL); + if (!pdev) + return; + + hdmi_dev = kzalloc(sizeof(struct oaktrail_hdmi_dev), GFP_KERNEL); + if (!hdmi_dev) { + dev_err(dev->dev, "failed to allocate memory\n"); + goto out; + } + + + ret = pci_enable_device(pdev); + if (ret) { + dev_err(dev->dev, "failed to enable hdmi controller\n"); + goto free; + } + + hdmi_dev->mmio = pci_resource_start(pdev, 0); + hdmi_dev->mmio_len = pci_resource_len(pdev, 0); + hdmi_dev->regs = ioremap(hdmi_dev->mmio, hdmi_dev->mmio_len); + if (!hdmi_dev->regs) { + dev_err(dev->dev, "failed to map hdmi mmio\n"); + goto free; + } + + hdmi_dev->dev = pdev; + pci_set_drvdata(pdev, hdmi_dev); + + /* Initialize i2c controller */ + ret = oaktrail_hdmi_i2c_init(hdmi_dev->dev); + if (ret) + dev_err(dev->dev, "HDMI I2C initialization failed\n"); + + dev_priv->hdmi_priv = hdmi_dev; + oaktrail_hdmi_audio_disable(dev); + return; + +free: + kfree(hdmi_dev); +out: + return; +} + +void oaktrail_hdmi_teardown(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct oaktrail_hdmi_dev *hdmi_dev = dev_priv->hdmi_priv; + struct pci_dev *pdev; + + if (hdmi_dev) { + pdev = hdmi_dev->dev; + pci_set_drvdata(pdev, NULL); + oaktrail_hdmi_i2c_exit(pdev); + iounmap(hdmi_dev->regs); + kfree(hdmi_dev); + pci_dev_put(pdev); + } +} + +/* save HDMI register state */ +void oaktrail_hdmi_save(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct oaktrail_hdmi_dev *hdmi_dev = dev_priv->hdmi_priv; + struct psb_state *regs = &dev_priv->regs.psb; + int i; + + /* dpll */ + hdmi_dev->saveDPLL_CTRL = PSB_RVDC32(DPLL_CTRL); + hdmi_dev->saveDPLL_DIV_CTRL = PSB_RVDC32(DPLL_DIV_CTRL); + hdmi_dev->saveDPLL_ADJUST = PSB_RVDC32(DPLL_ADJUST); + hdmi_dev->saveDPLL_UPDATE = PSB_RVDC32(DPLL_UPDATE); + hdmi_dev->saveDPLL_CLK_ENABLE = PSB_RVDC32(DPLL_CLK_ENABLE); + + /* pipe B */ + regs->savePIPEBCONF = PSB_RVDC32(PIPEBCONF); + regs->savePIPEBSRC = PSB_RVDC32(PIPEBSRC); + regs->saveHTOTAL_B = PSB_RVDC32(HTOTAL_B); + regs->saveHBLANK_B = PSB_RVDC32(HBLANK_B); + regs->saveHSYNC_B = PSB_RVDC32(HSYNC_B); + regs->saveVTOTAL_B = PSB_RVDC32(VTOTAL_B); + regs->saveVBLANK_B = PSB_RVDC32(VBLANK_B); + regs->saveVSYNC_B = PSB_RVDC32(VSYNC_B); + + hdmi_dev->savePCH_PIPEBCONF = PSB_RVDC32(PCH_PIPEBCONF); + hdmi_dev->savePCH_PIPEBSRC = PSB_RVDC32(PCH_PIPEBSRC); + hdmi_dev->savePCH_HTOTAL_B = PSB_RVDC32(PCH_HTOTAL_B); + hdmi_dev->savePCH_HBLANK_B = PSB_RVDC32(PCH_HBLANK_B); + hdmi_dev->savePCH_HSYNC_B = PSB_RVDC32(PCH_HSYNC_B); + hdmi_dev->savePCH_VTOTAL_B = PSB_RVDC32(PCH_VTOTAL_B); + hdmi_dev->savePCH_VBLANK_B = PSB_RVDC32(PCH_VBLANK_B); + hdmi_dev->savePCH_VSYNC_B = PSB_RVDC32(PCH_VSYNC_B); + + /* plane */ + regs->saveDSPBCNTR = PSB_RVDC32(DSPBCNTR); + regs->saveDSPBSTRIDE = PSB_RVDC32(DSPBSTRIDE); + regs->saveDSPBADDR = PSB_RVDC32(DSPBBASE); + regs->saveDSPBSURF = PSB_RVDC32(DSPBSURF); + regs->saveDSPBLINOFF = PSB_RVDC32(DSPBLINOFF); + regs->saveDSPBTILEOFF = PSB_RVDC32(DSPBTILEOFF); + + /* cursor B */ + regs->saveDSPBCURSOR_CTRL = PSB_RVDC32(CURBCNTR); + regs->saveDSPBCURSOR_BASE = PSB_RVDC32(CURBBASE); + regs->saveDSPBCURSOR_POS = PSB_RVDC32(CURBPOS); + + /* save palette */ + for (i = 0; i < 256; i++) + regs->save_palette_b[i] = PSB_RVDC32(PALETTE_B + (i << 2)); +} + +/* restore HDMI register state */ +void oaktrail_hdmi_restore(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct oaktrail_hdmi_dev *hdmi_dev = dev_priv->hdmi_priv; + struct psb_state *regs = &dev_priv->regs.psb; + int i; + + /* dpll */ + PSB_WVDC32(hdmi_dev->saveDPLL_CTRL, DPLL_CTRL); + PSB_WVDC32(hdmi_dev->saveDPLL_DIV_CTRL, DPLL_DIV_CTRL); + PSB_WVDC32(hdmi_dev->saveDPLL_ADJUST, DPLL_ADJUST); + PSB_WVDC32(hdmi_dev->saveDPLL_UPDATE, DPLL_UPDATE); + PSB_WVDC32(hdmi_dev->saveDPLL_CLK_ENABLE, DPLL_CLK_ENABLE); + DRM_UDELAY(150); + + /* pipe */ + PSB_WVDC32(regs->savePIPEBSRC, PIPEBSRC); + PSB_WVDC32(regs->saveHTOTAL_B, HTOTAL_B); + PSB_WVDC32(regs->saveHBLANK_B, HBLANK_B); + PSB_WVDC32(regs->saveHSYNC_B, HSYNC_B); + PSB_WVDC32(regs->saveVTOTAL_B, VTOTAL_B); + PSB_WVDC32(regs->saveVBLANK_B, VBLANK_B); + PSB_WVDC32(regs->saveVSYNC_B, VSYNC_B); + + PSB_WVDC32(hdmi_dev->savePCH_PIPEBSRC, PCH_PIPEBSRC); + PSB_WVDC32(hdmi_dev->savePCH_HTOTAL_B, PCH_HTOTAL_B); + PSB_WVDC32(hdmi_dev->savePCH_HBLANK_B, PCH_HBLANK_B); + PSB_WVDC32(hdmi_dev->savePCH_HSYNC_B, PCH_HSYNC_B); + PSB_WVDC32(hdmi_dev->savePCH_VTOTAL_B, PCH_VTOTAL_B); + PSB_WVDC32(hdmi_dev->savePCH_VBLANK_B, PCH_VBLANK_B); + PSB_WVDC32(hdmi_dev->savePCH_VSYNC_B, PCH_VSYNC_B); + + PSB_WVDC32(regs->savePIPEBCONF, PIPEBCONF); + PSB_WVDC32(hdmi_dev->savePCH_PIPEBCONF, PCH_PIPEBCONF); + + /* plane */ + PSB_WVDC32(regs->saveDSPBLINOFF, DSPBLINOFF); + PSB_WVDC32(regs->saveDSPBSTRIDE, DSPBSTRIDE); + PSB_WVDC32(regs->saveDSPBTILEOFF, DSPBTILEOFF); + PSB_WVDC32(regs->saveDSPBCNTR, DSPBCNTR); + PSB_WVDC32(regs->saveDSPBSURF, DSPBSURF); + + /* cursor B */ + PSB_WVDC32(regs->saveDSPBCURSOR_CTRL, CURBCNTR); + PSB_WVDC32(regs->saveDSPBCURSOR_POS, CURBPOS); + PSB_WVDC32(regs->saveDSPBCURSOR_BASE, CURBBASE); + + /* restore palette */ + for (i = 0; i < 256; i++) + PSB_WVDC32(regs->save_palette_b[i], PALETTE_B + (i << 2)); +} diff --git a/drivers/gpu/drm/gma500/oaktrail_hdmi_i2c.c b/drivers/gpu/drm/gma500/oaktrail_hdmi_i2c.c new file mode 100644 index 0000000..5e84fbd --- /dev/null +++ b/drivers/gpu/drm/gma500/oaktrail_hdmi_i2c.c @@ -0,0 +1,328 @@ +/* + * Copyright © 2010 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * Li Peng + */ + +#include +#include +#include +#include +#include +#include +#include "psb_drv.h" + +#define HDMI_READ(reg) readl(hdmi_dev->regs + (reg)) +#define HDMI_WRITE(reg, val) writel(val, hdmi_dev->regs + (reg)) + +#define HDMI_HCR 0x1000 +#define HCR_DETECT_HDP (1 << 6) +#define HCR_ENABLE_HDCP (1 << 5) +#define HCR_ENABLE_AUDIO (1 << 2) +#define HCR_ENABLE_PIXEL (1 << 1) +#define HCR_ENABLE_TMDS (1 << 0) +#define HDMI_HICR 0x1004 +#define HDMI_INTR_I2C_ERROR (1 << 4) +#define HDMI_INTR_I2C_FULL (1 << 3) +#define HDMI_INTR_I2C_DONE (1 << 2) +#define HDMI_INTR_HPD (1 << 0) +#define HDMI_HSR 0x1008 +#define HDMI_HISR 0x100C +#define HDMI_HI2CRDB0 0x1200 +#define HDMI_HI2CHCR 0x1240 +#define HI2C_HDCP_WRITE (0 << 2) +#define HI2C_HDCP_RI_READ (1 << 2) +#define HI2C_HDCP_READ (2 << 2) +#define HI2C_EDID_READ (3 << 2) +#define HI2C_READ_CONTINUE (1 << 1) +#define HI2C_ENABLE_TRANSACTION (1 << 0) + +#define HDMI_ICRH 0x1100 +#define HDMI_HI2CTDR0 0x1244 +#define HDMI_HI2CTDR1 0x1248 + +#define I2C_STAT_INIT 0 +#define I2C_READ_DONE 1 +#define I2C_TRANSACTION_DONE 2 + +struct hdmi_i2c_dev { + struct i2c_adapter *adap; + struct mutex i2c_lock; + struct completion complete; + int status; + struct i2c_msg *msg; + int buf_offset; +}; + +static void hdmi_i2c_irq_enable(struct oaktrail_hdmi_dev *hdmi_dev) +{ + u32 temp; + + temp = HDMI_READ(HDMI_HICR); + temp |= (HDMI_INTR_I2C_ERROR | HDMI_INTR_I2C_FULL | HDMI_INTR_I2C_DONE); + HDMI_WRITE(HDMI_HICR, temp); + HDMI_READ(HDMI_HICR); +} + +static void hdmi_i2c_irq_disable(struct oaktrail_hdmi_dev *hdmi_dev) +{ + HDMI_WRITE(HDMI_HICR, 0x0); + HDMI_READ(HDMI_HICR); +} + +static int xfer_read(struct i2c_adapter *adap, struct i2c_msg *pmsg) +{ + struct oaktrail_hdmi_dev *hdmi_dev = i2c_get_adapdata(adap); + struct hdmi_i2c_dev *i2c_dev = hdmi_dev->i2c_dev; + u32 temp; + + i2c_dev->status = I2C_STAT_INIT; + i2c_dev->msg = pmsg; + i2c_dev->buf_offset = 0; + INIT_COMPLETION(i2c_dev->complete); + + /* Enable I2C transaction */ + temp = ((pmsg->len) << 20) | HI2C_EDID_READ | HI2C_ENABLE_TRANSACTION; + HDMI_WRITE(HDMI_HI2CHCR, temp); + HDMI_READ(HDMI_HI2CHCR); + + while (i2c_dev->status != I2C_TRANSACTION_DONE) + wait_for_completion_interruptible_timeout(&i2c_dev->complete, + 10 * HZ); + + return 0; +} + +static int xfer_write(struct i2c_adapter *adap, struct i2c_msg *pmsg) +{ + /* + * XXX: i2c write seems isn't useful for EDID probe, don't do anything + */ + return 0; +} + +static int oaktrail_hdmi_i2c_access(struct i2c_adapter *adap, + struct i2c_msg *pmsg, + int num) +{ + struct oaktrail_hdmi_dev *hdmi_dev = i2c_get_adapdata(adap); + struct hdmi_i2c_dev *i2c_dev = hdmi_dev->i2c_dev; + int i; + + mutex_lock(&i2c_dev->i2c_lock); + + /* Enable i2c unit */ + HDMI_WRITE(HDMI_ICRH, 0x00008760); + + /* Enable irq */ + hdmi_i2c_irq_enable(hdmi_dev); + for (i = 0; i < num; i++) { + if (pmsg->len && pmsg->buf) { + if (pmsg->flags & I2C_M_RD) + xfer_read(adap, pmsg); + else + xfer_write(adap, pmsg); + } + pmsg++; /* next message */ + } + + /* Disable irq */ + hdmi_i2c_irq_disable(hdmi_dev); + + mutex_unlock(&i2c_dev->i2c_lock); + + return i; +} + +static u32 oaktrail_hdmi_i2c_func(struct i2c_adapter *adapter) +{ + return I2C_FUNC_I2C | I2C_FUNC_10BIT_ADDR; +} + +static const struct i2c_algorithm oaktrail_hdmi_i2c_algorithm = { + .master_xfer = oaktrail_hdmi_i2c_access, + .functionality = oaktrail_hdmi_i2c_func, +}; + +static struct i2c_adapter oaktrail_hdmi_i2c_adapter = { + .name = "oaktrail_hdmi_i2c", + .nr = 3, + .owner = THIS_MODULE, + .class = I2C_CLASS_DDC, + .algo = &oaktrail_hdmi_i2c_algorithm, +}; + +static void hdmi_i2c_read(struct oaktrail_hdmi_dev *hdmi_dev) +{ + struct hdmi_i2c_dev *i2c_dev = hdmi_dev->i2c_dev; + struct i2c_msg *msg = i2c_dev->msg; + u8 *buf = msg->buf; + u32 temp; + int i, offset; + + offset = i2c_dev->buf_offset; + for (i = 0; i < 0x10; i++) { + temp = HDMI_READ(HDMI_HI2CRDB0 + (i * 4)); + memcpy(buf + (offset + i * 4), &temp, 4); + } + i2c_dev->buf_offset += (0x10 * 4); + + /* clearing read buffer full intr */ + temp = HDMI_READ(HDMI_HISR); + HDMI_WRITE(HDMI_HISR, temp | HDMI_INTR_I2C_FULL); + HDMI_READ(HDMI_HISR); + + /* continue read transaction */ + temp = HDMI_READ(HDMI_HI2CHCR); + HDMI_WRITE(HDMI_HI2CHCR, temp | HI2C_READ_CONTINUE); + HDMI_READ(HDMI_HI2CHCR); + + i2c_dev->status = I2C_READ_DONE; + return; +} + +static void hdmi_i2c_transaction_done(struct oaktrail_hdmi_dev *hdmi_dev) +{ + struct hdmi_i2c_dev *i2c_dev = hdmi_dev->i2c_dev; + u32 temp; + + /* clear transaction done intr */ + temp = HDMI_READ(HDMI_HISR); + HDMI_WRITE(HDMI_HISR, temp | HDMI_INTR_I2C_DONE); + HDMI_READ(HDMI_HISR); + + + temp = HDMI_READ(HDMI_HI2CHCR); + HDMI_WRITE(HDMI_HI2CHCR, temp & ~HI2C_ENABLE_TRANSACTION); + HDMI_READ(HDMI_HI2CHCR); + + i2c_dev->status = I2C_TRANSACTION_DONE; + return; +} + +static irqreturn_t oaktrail_hdmi_i2c_handler(int this_irq, void *dev) +{ + struct oaktrail_hdmi_dev *hdmi_dev = dev; + struct hdmi_i2c_dev *i2c_dev = hdmi_dev->i2c_dev; + u32 stat; + + stat = HDMI_READ(HDMI_HISR); + + if (stat & HDMI_INTR_HPD) { + HDMI_WRITE(HDMI_HISR, stat | HDMI_INTR_HPD); + HDMI_READ(HDMI_HISR); + } + + if (stat & HDMI_INTR_I2C_FULL) + hdmi_i2c_read(hdmi_dev); + + if (stat & HDMI_INTR_I2C_DONE) + hdmi_i2c_transaction_done(hdmi_dev); + + complete(&i2c_dev->complete); + + return IRQ_HANDLED; +} + +/* + * choose alternate function 2 of GPIO pin 52, 53, + * which is used by HDMI I2C logic + */ +static void oaktrail_hdmi_i2c_gpio_fix(void) +{ + void *base; + unsigned int gpio_base = 0xff12c000; + int gpio_len = 0x1000; + u32 temp; + + base = ioremap((resource_size_t)gpio_base, gpio_len); + if (base == NULL) { + DRM_ERROR("gpio ioremap fail\n"); + return; + } + + temp = readl(base + 0x44); + DRM_DEBUG_DRIVER("old gpio val %x\n", temp); + writel((temp | 0x00000a00), (base + 0x44)); + temp = readl(base + 0x44); + DRM_DEBUG_DRIVER("new gpio val %x\n", temp); + + iounmap(base); +} + +int oaktrail_hdmi_i2c_init(struct pci_dev *dev) +{ + struct oaktrail_hdmi_dev *hdmi_dev; + struct hdmi_i2c_dev *i2c_dev; + int ret; + + hdmi_dev = pci_get_drvdata(dev); + + i2c_dev = kzalloc(sizeof(struct hdmi_i2c_dev), GFP_KERNEL); + if (i2c_dev == NULL) { + DRM_ERROR("Can't allocate interface\n"); + ret = -ENOMEM; + goto exit; + } + + i2c_dev->adap = &oaktrail_hdmi_i2c_adapter; + i2c_dev->status = I2C_STAT_INIT; + init_completion(&i2c_dev->complete); + mutex_init(&i2c_dev->i2c_lock); + i2c_set_adapdata(&oaktrail_hdmi_i2c_adapter, hdmi_dev); + hdmi_dev->i2c_dev = i2c_dev; + + /* Enable HDMI I2C function on gpio */ + oaktrail_hdmi_i2c_gpio_fix(); + + /* request irq */ + ret = request_irq(dev->irq, oaktrail_hdmi_i2c_handler, IRQF_SHARED, + oaktrail_hdmi_i2c_adapter.name, hdmi_dev); + if (ret) { + DRM_ERROR("Failed to request IRQ for I2C controller\n"); + goto err; + } + + /* Adapter registration */ + ret = i2c_add_numbered_adapter(&oaktrail_hdmi_i2c_adapter); + return ret; + +err: + kfree(i2c_dev); +exit: + return ret; +} + +void oaktrail_hdmi_i2c_exit(struct pci_dev *dev) +{ + struct oaktrail_hdmi_dev *hdmi_dev; + struct hdmi_i2c_dev *i2c_dev; + + hdmi_dev = pci_get_drvdata(dev); + if (i2c_del_adapter(&oaktrail_hdmi_i2c_adapter)) + DRM_DEBUG_DRIVER("Failed to delete hdmi-i2c adapter\n"); + + i2c_dev = hdmi_dev->i2c_dev; + kfree(i2c_dev); + free_irq(dev->irq, hdmi_dev); +} diff --git a/drivers/gpu/drm/gma500/oaktrail_lvds.c b/drivers/gpu/drm/gma500/oaktrail_lvds.c new file mode 100644 index 0000000..654f32b --- /dev/null +++ b/drivers/gpu/drm/gma500/oaktrail_lvds.c @@ -0,0 +1,448 @@ +/* + * Copyright © 2006-2009 Intel Corporation + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + * Authors: + * Eric Anholt + * Dave Airlie + * Jesse Barnes + */ + +#include +#include +#include + +#include "intel_bios.h" +#include "psb_drv.h" +#include "psb_intel_drv.h" +#include "psb_intel_reg.h" +#include "power.h" +#include + +/* The max/min PWM frequency in BPCR[31:17] - */ +/* The smallest number is 1 (not 0) that can fit in the + * 15-bit field of the and then*/ +/* shifts to the left by one bit to get the actual 16-bit + * value that the 15-bits correspond to.*/ +#define MRST_BLC_MAX_PWM_REG_FREQ 0xFFFF +#define BRIGHTNESS_MAX_LEVEL 100 + +/** + * Sets the power state for the panel. + */ +static void oaktrail_lvds_set_power(struct drm_device *dev, + struct psb_intel_encoder *psb_intel_encoder, + bool on) +{ + u32 pp_status; + struct drm_psb_private *dev_priv = dev->dev_private; + + if (!gma_power_begin(dev, true)) + return; + + if (on) { + REG_WRITE(PP_CONTROL, REG_READ(PP_CONTROL) | + POWER_TARGET_ON); + do { + pp_status = REG_READ(PP_STATUS); + } while ((pp_status & (PP_ON | PP_READY)) == PP_READY); + dev_priv->is_lvds_on = true; + if (dev_priv->ops->lvds_bl_power) + dev_priv->ops->lvds_bl_power(dev, true); + } else { + if (dev_priv->ops->lvds_bl_power) + dev_priv->ops->lvds_bl_power(dev, false); + REG_WRITE(PP_CONTROL, REG_READ(PP_CONTROL) & + ~POWER_TARGET_ON); + do { + pp_status = REG_READ(PP_STATUS); + } while (pp_status & PP_ON); + dev_priv->is_lvds_on = false; + pm_request_idle(&dev->pdev->dev); + } + gma_power_end(dev); +} + +static void oaktrail_lvds_dpms(struct drm_encoder *encoder, int mode) +{ + struct drm_device *dev = encoder->dev; + struct psb_intel_encoder *psb_intel_encoder = + to_psb_intel_encoder(encoder); + + if (mode == DRM_MODE_DPMS_ON) + oaktrail_lvds_set_power(dev, psb_intel_encoder, true); + else + oaktrail_lvds_set_power(dev, psb_intel_encoder, false); + + /* XXX: We never power down the LVDS pairs. */ +} + +static void oaktrail_lvds_mode_set(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + struct drm_device *dev = encoder->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_intel_mode_device *mode_dev = &dev_priv->mode_dev; + struct drm_mode_config *mode_config = &dev->mode_config; + struct drm_connector *connector = NULL; + struct drm_crtc *crtc = encoder->crtc; + u32 lvds_port; + uint64_t v = DRM_MODE_SCALE_FULLSCREEN; + + if (!gma_power_begin(dev, true)) + return; + + /* + * The LVDS pin pair will already have been turned on in the + * psb_intel_crtc_mode_set since it has a large impact on the DPLL + * settings. + */ + lvds_port = (REG_READ(LVDS) & + (~LVDS_PIPEB_SELECT)) | + LVDS_PORT_EN | + LVDS_BORDER_EN; + + /* If the firmware says dither on Moorestown, or the BIOS does + on Oaktrail then enable dithering */ + if (mode_dev->panel_wants_dither || dev_priv->lvds_dither) + lvds_port |= MRST_PANEL_8TO6_DITHER_ENABLE; + + REG_WRITE(LVDS, lvds_port); + + /* Find the connector we're trying to set up */ + list_for_each_entry(connector, &mode_config->connector_list, head) { + if (!connector->encoder || connector->encoder->crtc != crtc) + continue; + } + + if (!connector) { + DRM_ERROR("Couldn't find connector when setting mode"); + return; + } + + drm_connector_property_get_value( + connector, + dev->mode_config.scaling_mode_property, + &v); + + if (v == DRM_MODE_SCALE_NO_SCALE) + REG_WRITE(PFIT_CONTROL, 0); + else if (v == DRM_MODE_SCALE_ASPECT) { + if ((mode->vdisplay != adjusted_mode->crtc_vdisplay) || + (mode->hdisplay != adjusted_mode->crtc_hdisplay)) { + if ((adjusted_mode->crtc_hdisplay * mode->vdisplay) == + (mode->hdisplay * adjusted_mode->crtc_vdisplay)) + REG_WRITE(PFIT_CONTROL, PFIT_ENABLE); + else if ((adjusted_mode->crtc_hdisplay * + mode->vdisplay) > (mode->hdisplay * + adjusted_mode->crtc_vdisplay)) + REG_WRITE(PFIT_CONTROL, PFIT_ENABLE | + PFIT_SCALING_MODE_PILLARBOX); + else + REG_WRITE(PFIT_CONTROL, PFIT_ENABLE | + PFIT_SCALING_MODE_LETTERBOX); + } else + REG_WRITE(PFIT_CONTROL, PFIT_ENABLE); + } else /*(v == DRM_MODE_SCALE_FULLSCREEN)*/ + REG_WRITE(PFIT_CONTROL, PFIT_ENABLE); + + gma_power_end(dev); +} + +static void oaktrail_lvds_prepare(struct drm_encoder *encoder) +{ + struct drm_device *dev = encoder->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_intel_encoder *psb_intel_encoder = + to_psb_intel_encoder(encoder); + struct psb_intel_mode_device *mode_dev = &dev_priv->mode_dev; + + if (!gma_power_begin(dev, true)) + return; + + mode_dev->saveBLC_PWM_CTL = REG_READ(BLC_PWM_CTL); + mode_dev->backlight_duty_cycle = (mode_dev->saveBLC_PWM_CTL & + BACKLIGHT_DUTY_CYCLE_MASK); + oaktrail_lvds_set_power(dev, psb_intel_encoder, false); + gma_power_end(dev); +} + +static u32 oaktrail_lvds_get_max_backlight(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + u32 ret; + + if (gma_power_begin(dev, false)) { + ret = ((REG_READ(BLC_PWM_CTL) & + BACKLIGHT_MODULATION_FREQ_MASK) >> + BACKLIGHT_MODULATION_FREQ_SHIFT) * 2; + + gma_power_end(dev); + } else + ret = ((dev_priv->regs.saveBLC_PWM_CTL & + BACKLIGHT_MODULATION_FREQ_MASK) >> + BACKLIGHT_MODULATION_FREQ_SHIFT) * 2; + + return ret; +} + +static void oaktrail_lvds_commit(struct drm_encoder *encoder) +{ + struct drm_device *dev = encoder->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_intel_encoder *psb_intel_encoder = + to_psb_intel_encoder(encoder); + struct psb_intel_mode_device *mode_dev = &dev_priv->mode_dev; + + if (mode_dev->backlight_duty_cycle == 0) + mode_dev->backlight_duty_cycle = + oaktrail_lvds_get_max_backlight(dev); + oaktrail_lvds_set_power(dev, psb_intel_encoder, true); +} + +static const struct drm_encoder_helper_funcs oaktrail_lvds_helper_funcs = { + .dpms = oaktrail_lvds_dpms, + .mode_fixup = psb_intel_lvds_mode_fixup, + .prepare = oaktrail_lvds_prepare, + .mode_set = oaktrail_lvds_mode_set, + .commit = oaktrail_lvds_commit, +}; + +static struct drm_display_mode lvds_configuration_modes[] = { + /* hard coded fixed mode for TPO LTPS LPJ040K001A */ + { DRM_MODE("800x480", DRM_MODE_TYPE_DRIVER, 33264, 800, 836, + 846, 1056, 0, 480, 489, 491, 525, 0, 0) }, + /* hard coded fixed mode for LVDS 800x480 */ + { DRM_MODE("800x480", DRM_MODE_TYPE_DRIVER, 30994, 800, 801, + 802, 1024, 0, 480, 481, 482, 525, 0, 0) }, + /* hard coded fixed mode for Samsung 480wsvga LVDS 1024x600@75 */ + { DRM_MODE("1024x600", DRM_MODE_TYPE_DRIVER, 53990, 1024, 1072, + 1104, 1184, 0, 600, 603, 604, 608, 0, 0) }, + /* hard coded fixed mode for Samsung 480wsvga LVDS 1024x600@75 */ + { DRM_MODE("1024x600", DRM_MODE_TYPE_DRIVER, 53990, 1024, 1104, + 1136, 1184, 0, 600, 603, 604, 608, 0, 0) }, + /* hard coded fixed mode for Sharp wsvga LVDS 1024x600 */ + { DRM_MODE("1024x600", DRM_MODE_TYPE_DRIVER, 48885, 1024, 1124, + 1204, 1312, 0, 600, 607, 610, 621, 0, 0) }, + /* hard coded fixed mode for LVDS 1024x768 */ + { DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 65000, 1024, 1048, + 1184, 1344, 0, 768, 771, 777, 806, 0, 0) }, + /* hard coded fixed mode for LVDS 1366x768 */ + { DRM_MODE("1366x768", DRM_MODE_TYPE_DRIVER, 77500, 1366, 1430, + 1558, 1664, 0, 768, 769, 770, 776, 0, 0) }, +}; + +/* Returns the panel fixed mode from configuration. */ + +static void oaktrail_lvds_get_configuration_mode(struct drm_device *dev, + struct psb_intel_mode_device *mode_dev) +{ + struct drm_display_mode *mode = NULL; + struct drm_psb_private *dev_priv = dev->dev_private; + struct oaktrail_timing_info *ti = &dev_priv->gct_data.DTD; + + mode_dev->panel_fixed_mode = NULL; + + /* Use the firmware provided data on Moorestown */ + if (dev_priv->vbt_data.size != 0x00) { /*if non-zero, then use vbt*/ + mode = kzalloc(sizeof(*mode), GFP_KERNEL); + if (!mode) + return; + + mode->hdisplay = (ti->hactive_hi << 8) | ti->hactive_lo; + mode->vdisplay = (ti->vactive_hi << 8) | ti->vactive_lo; + mode->hsync_start = mode->hdisplay + \ + ((ti->hsync_offset_hi << 8) | \ + ti->hsync_offset_lo); + mode->hsync_end = mode->hsync_start + \ + ((ti->hsync_pulse_width_hi << 8) | \ + ti->hsync_pulse_width_lo); + mode->htotal = mode->hdisplay + ((ti->hblank_hi << 8) | \ + ti->hblank_lo); + mode->vsync_start = \ + mode->vdisplay + ((ti->vsync_offset_hi << 4) | \ + ti->vsync_offset_lo); + mode->vsync_end = \ + mode->vsync_start + ((ti->vsync_pulse_width_hi << 4) | \ + ti->vsync_pulse_width_lo); + mode->vtotal = mode->vdisplay + \ + ((ti->vblank_hi << 8) | ti->vblank_lo); + mode->clock = ti->pixel_clock * 10; +#if 0 + printk(KERN_INFO "hdisplay is %d\n", mode->hdisplay); + printk(KERN_INFO "vdisplay is %d\n", mode->vdisplay); + printk(KERN_INFO "HSS is %d\n", mode->hsync_start); + printk(KERN_INFO "HSE is %d\n", mode->hsync_end); + printk(KERN_INFO "htotal is %d\n", mode->htotal); + printk(KERN_INFO "VSS is %d\n", mode->vsync_start); + printk(KERN_INFO "VSE is %d\n", mode->vsync_end); + printk(KERN_INFO "vtotal is %d\n", mode->vtotal); + printk(KERN_INFO "clock is %d\n", mode->clock); +#endif + mode_dev->panel_fixed_mode = mode; + } + + /* Use the BIOS VBT mode if available */ + if (mode_dev->panel_fixed_mode == NULL && mode_dev->vbt_mode) + mode_dev->panel_fixed_mode = drm_mode_duplicate(dev, + mode_dev->vbt_mode); + + /* Then try the LVDS VBT mode */ + if (mode_dev->panel_fixed_mode == NULL) + if (dev_priv->lfp_lvds_vbt_mode) + mode_dev->panel_fixed_mode = + drm_mode_duplicate(dev, + dev_priv->lfp_lvds_vbt_mode); + /* Then guess */ + if (mode_dev->panel_fixed_mode == NULL) + mode_dev->panel_fixed_mode + = drm_mode_duplicate(dev, &lvds_configuration_modes[2]); + + drm_mode_set_name(mode_dev->panel_fixed_mode); + drm_mode_set_crtcinfo(mode_dev->panel_fixed_mode, 0); +} + +/** + * oaktrail_lvds_init - setup LVDS connectors on this device + * @dev: drm device + * + * Create the connector, register the LVDS DDC bus, and try to figure out what + * modes we can display on the LVDS panel (if present). + */ +void oaktrail_lvds_init(struct drm_device *dev, + struct psb_intel_mode_device *mode_dev) +{ + struct psb_intel_encoder *psb_intel_encoder; + struct psb_intel_connector *psb_intel_connector; + struct drm_connector *connector; + struct drm_encoder *encoder; + struct drm_psb_private *dev_priv = dev->dev_private; + struct edid *edid; + struct i2c_adapter *i2c_adap; + struct drm_display_mode *scan; /* *modes, *bios_mode; */ + + psb_intel_encoder = kzalloc(sizeof(struct psb_intel_encoder), GFP_KERNEL); + if (!psb_intel_encoder) + return; + + psb_intel_connector = kzalloc(sizeof(struct psb_intel_connector), GFP_KERNEL); + if (!psb_intel_connector) + goto failed_connector; + + connector = &psb_intel_connector->base; + encoder = &psb_intel_encoder->base; + dev_priv->is_lvds_on = true; + drm_connector_init(dev, connector, + &psb_intel_lvds_connector_funcs, + DRM_MODE_CONNECTOR_LVDS); + + drm_encoder_init(dev, encoder, &psb_intel_lvds_enc_funcs, + DRM_MODE_ENCODER_LVDS); + + psb_intel_connector_attach_encoder(psb_intel_connector, + psb_intel_encoder); + psb_intel_encoder->type = INTEL_OUTPUT_LVDS; + + drm_encoder_helper_add(encoder, &oaktrail_lvds_helper_funcs); + drm_connector_helper_add(connector, + &psb_intel_lvds_connector_helper_funcs); + connector->display_info.subpixel_order = SubPixelHorizontalRGB; + connector->interlace_allowed = false; + connector->doublescan_allowed = false; + + drm_connector_attach_property(connector, + dev->mode_config.scaling_mode_property, + DRM_MODE_SCALE_FULLSCREEN); + drm_connector_attach_property(connector, + dev_priv->backlight_property, + BRIGHTNESS_MAX_LEVEL); + + mode_dev->panel_wants_dither = false; + if (dev_priv->vbt_data.size != 0x00) + mode_dev->panel_wants_dither = (dev_priv->gct_data. + Panel_Port_Control & MRST_PANEL_8TO6_DITHER_ENABLE); + if (dev_priv->lvds_dither) + mode_dev->panel_wants_dither = 1; + + /* + * LVDS discovery: + * 1) check for EDID on DDC + * 2) check for VBT data + * 3) check to see if LVDS is already on + * if none of the above, no panel + * 4) make sure lid is open + * if closed, act like it's not there for now + */ + + i2c_adap = i2c_get_adapter(dev_priv->ops->i2c_bus); + if (i2c_adap == NULL) + dev_err(dev->dev, "No ddc adapter available!\n"); + /* + * Attempt to get the fixed panel mode from DDC. Assume that the + * preferred mode is the right one. + */ + if (i2c_adap) { + edid = drm_get_edid(connector, i2c_adap); + if (edid) { + drm_mode_connector_update_edid_property(connector, + edid); + drm_add_edid_modes(connector, edid); + kfree(edid); + } + + list_for_each_entry(scan, &connector->probed_modes, head) { + if (scan->type & DRM_MODE_TYPE_PREFERRED) { + mode_dev->panel_fixed_mode = + drm_mode_duplicate(dev, scan); + goto out; /* FIXME: check for quirks */ + } + } + } + /* + * If we didn't get EDID, try geting panel timing + * from configuration data + */ + oaktrail_lvds_get_configuration_mode(dev, mode_dev); + + if (mode_dev->panel_fixed_mode) { + mode_dev->panel_fixed_mode->type |= DRM_MODE_TYPE_PREFERRED; + goto out; /* FIXME: check for quirks */ + } + + /* If we still don't have a mode after all that, give up. */ + if (!mode_dev->panel_fixed_mode) { + dev_err(dev->dev, "Found no modes on the lvds, ignoring the LVDS\n"); + goto failed_find; + } + +out: + drm_sysfs_connector_add(connector); + return; + +failed_find: + dev_dbg(dev->dev, "No LVDS modes found, disabling.\n"); + if (psb_intel_encoder->ddc_bus) + psb_intel_i2c_destroy(psb_intel_encoder->ddc_bus); + +/* failed_ddc: */ + + drm_encoder_cleanup(encoder); + drm_connector_cleanup(connector); + kfree(psb_intel_connector); +failed_connector: + kfree(psb_intel_encoder); +} + diff --git a/drivers/gpu/drm/gma500/power.c b/drivers/gpu/drm/gma500/power.c new file mode 100644 index 0000000..889b854 --- /dev/null +++ b/drivers/gpu/drm/gma500/power.c @@ -0,0 +1,315 @@ +/************************************************************************** + * Copyright (c) 2009-2011, Intel Corporation. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + * Authors: + * Benjamin Defnet + * Rajesh Poornachandran + * Massively reworked + * Alan Cox + */ + +#include "power.h" +#include "psb_drv.h" +#include "psb_reg.h" +#include "psb_intel_reg.h" +#include +#include + +static struct mutex power_mutex; /* Serialize power ops */ +static spinlock_t power_ctrl_lock; /* Serialize power claim */ + +/** + * gma_power_init - initialise power manager + * @dev: our device + * + * Set up for power management tracking of our hardware. + */ +void gma_power_init(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + + /* FIXME: Move APM/OSPM base into relevant device code */ + dev_priv->apm_base = dev_priv->apm_reg & 0xffff; + dev_priv->ospm_base &= 0xffff; + + dev_priv->display_power = true; /* We start active */ + dev_priv->display_count = 0; /* Currently no users */ + dev_priv->suspended = false; /* And not suspended */ + spin_lock_init(&power_ctrl_lock); + mutex_init(&power_mutex); + + if (dev_priv->ops->init_pm) + dev_priv->ops->init_pm(dev); +} + +/** + * gma_power_uninit - end power manager + * @dev: device to end for + * + * Undo the effects of gma_power_init + */ +void gma_power_uninit(struct drm_device *dev) +{ + pm_runtime_disable(&dev->pdev->dev); + pm_runtime_set_suspended(&dev->pdev->dev); +} + +/** + * gma_suspend_display - suspend the display logic + * @dev: our DRM device + * + * Suspend the display logic of the graphics interface + */ +static void gma_suspend_display(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + + if (dev_priv->suspended) + return; + dev_priv->ops->save_regs(dev); + dev_priv->ops->power_down(dev); + dev_priv->display_power = false; +} + +/** + * gma_resume_display - resume display side logic + * + * Resume the display hardware restoring state and enabling + * as necessary. + */ +static void gma_resume_display(struct pci_dev *pdev) +{ + struct drm_device *dev = pci_get_drvdata(pdev); + struct drm_psb_private *dev_priv = dev->dev_private; + + /* turn on the display power island */ + dev_priv->ops->power_up(dev); + dev_priv->suspended = false; + dev_priv->display_power = true; + + PSB_WVDC32(dev_priv->pge_ctl | _PSB_PGETBL_ENABLED, PSB_PGETBL_CTL); + pci_write_config_word(pdev, PSB_GMCH_CTRL, + dev_priv->gmch_ctrl | _PSB_GMCH_ENABLED); + dev_priv->ops->restore_regs(dev); +} + +/** + * gma_suspend_pci - suspend PCI side + * @pdev: PCI device + * + * Perform the suspend processing on our PCI device state + */ +static void gma_suspend_pci(struct pci_dev *pdev) +{ + struct drm_device *dev = pci_get_drvdata(pdev); + struct drm_psb_private *dev_priv = dev->dev_private; + int bsm, vbt; + + if (dev_priv->suspended) + return; + + pci_save_state(pdev); + pci_read_config_dword(pdev, 0x5C, &bsm); + dev_priv->regs.saveBSM = bsm; + pci_read_config_dword(pdev, 0xFC, &vbt); + dev_priv->regs.saveVBT = vbt; + pci_read_config_dword(pdev, PSB_PCIx_MSI_ADDR_LOC, &dev_priv->msi_addr); + pci_read_config_dword(pdev, PSB_PCIx_MSI_DATA_LOC, &dev_priv->msi_data); + + pci_disable_device(pdev); + pci_set_power_state(pdev, PCI_D3hot); + + dev_priv->suspended = true; +} + +/** + * gma_resume_pci - resume helper + * @dev: our PCI device + * + * Perform the resume processing on our PCI device state - rewrite + * register state and re-enable the PCI device + */ +static bool gma_resume_pci(struct pci_dev *pdev) +{ + struct drm_device *dev = pci_get_drvdata(pdev); + struct drm_psb_private *dev_priv = dev->dev_private; + int ret; + + if (!dev_priv->suspended) + return true; + + pci_set_power_state(pdev, PCI_D0); + pci_restore_state(pdev); + pci_write_config_dword(pdev, 0x5c, dev_priv->regs.saveBSM); + pci_write_config_dword(pdev, 0xFC, dev_priv->regs.saveVBT); + /* restoring MSI address and data in PCIx space */ + pci_write_config_dword(pdev, PSB_PCIx_MSI_ADDR_LOC, dev_priv->msi_addr); + pci_write_config_dword(pdev, PSB_PCIx_MSI_DATA_LOC, dev_priv->msi_data); + ret = pci_enable_device(pdev); + + if (ret != 0) + dev_err(&pdev->dev, "pci_enable failed: %d\n", ret); + else + dev_priv->suspended = false; + return !dev_priv->suspended; +} + +/** + * gma_power_suspend - bus callback for suspend + * @pdev: our PCI device + * @state: suspend type + * + * Called back by the PCI layer during a suspend of the system. We + * perform the necessary shut down steps and save enough state that + * we can undo this when resume is called. + */ +int gma_power_suspend(struct device *_dev) +{ + struct pci_dev *pdev = container_of(_dev, struct pci_dev, dev); + struct drm_device *dev = pci_get_drvdata(pdev); + struct drm_psb_private *dev_priv = dev->dev_private; + + mutex_lock(&power_mutex); + if (!dev_priv->suspended) { + if (dev_priv->display_count) { + mutex_unlock(&power_mutex); + dev_err(dev->dev, "GPU hardware busy, cannot suspend\n"); + return -EBUSY; + } + psb_irq_uninstall(dev); + gma_suspend_display(dev); + gma_suspend_pci(pdev); + } + mutex_unlock(&power_mutex); + return 0; +} + +/** + * gma_power_resume - resume power + * @pdev: PCI device + * + * Resume the PCI side of the graphics and then the displays + */ +int gma_power_resume(struct device *_dev) +{ + struct pci_dev *pdev = container_of(_dev, struct pci_dev, dev); + struct drm_device *dev = pci_get_drvdata(pdev); + + mutex_lock(&power_mutex); + gma_resume_pci(pdev); + gma_resume_display(pdev); + psb_irq_preinstall(dev); + psb_irq_postinstall(dev); + mutex_unlock(&power_mutex); + return 0; +} + +/** + * gma_power_is_on - returne true if power is on + * @dev: our DRM device + * + * Returns true if the display island power is on at this moment + */ +bool gma_power_is_on(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + return dev_priv->display_power; +} + +/** + * gma_power_begin - begin requiring power + * @dev: our DRM device + * @force_on: true to force power on + * + * Begin an action that requires the display power island is enabled. + * We refcount the islands. + */ +bool gma_power_begin(struct drm_device *dev, bool force_on) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + int ret; + unsigned long flags; + + spin_lock_irqsave(&power_ctrl_lock, flags); + /* Power already on ? */ + if (dev_priv->display_power) { + dev_priv->display_count++; + pm_runtime_get(&dev->pdev->dev); + spin_unlock_irqrestore(&power_ctrl_lock, flags); + return true; + } + if (force_on == false) + goto out_false; + + /* Ok power up needed */ + ret = gma_resume_pci(dev->pdev); + if (ret == 0) { + psb_irq_preinstall(dev); + psb_irq_postinstall(dev); + pm_runtime_get(&dev->pdev->dev); + dev_priv->display_count++; + spin_unlock_irqrestore(&power_ctrl_lock, flags); + return true; + } +out_false: + spin_unlock_irqrestore(&power_ctrl_lock, flags); + return false; +} + +/** + * gma_power_end - end use of power + * @dev: Our DRM device + * + * Indicate that one of our gma_power_begin() requested periods when + * the diplay island power is needed has completed. + */ +void gma_power_end(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + unsigned long flags; + spin_lock_irqsave(&power_ctrl_lock, flags); + dev_priv->display_count--; + WARN_ON(dev_priv->display_count < 0); + spin_unlock_irqrestore(&power_ctrl_lock, flags); + pm_runtime_put(&dev->pdev->dev); +} + +int psb_runtime_suspend(struct device *dev) +{ + return gma_power_suspend(dev); +} + +int psb_runtime_resume(struct device *dev) +{ + return gma_power_resume(dev); +} + +int psb_runtime_idle(struct device *dev) +{ + struct drm_device *drmdev = pci_get_drvdata(to_pci_dev(dev)); + struct drm_psb_private *dev_priv = drmdev->dev_private; + if (dev_priv->display_count) + return 0; + else + return 1; +} diff --git a/drivers/gpu/drm/gma500/power.h b/drivers/gpu/drm/gma500/power.h new file mode 100644 index 0000000..1969d2e --- /dev/null +++ b/drivers/gpu/drm/gma500/power.h @@ -0,0 +1,67 @@ +/************************************************************************** + * Copyright (c) 2009-2011, Intel Corporation. + * All Rights Reserved. + + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + * Authors: + * Benjamin Defnet + * Rajesh Poornachandran + * Massively reworked + * Alan Cox + */ +#ifndef _PSB_POWERMGMT_H_ +#define _PSB_POWERMGMT_H_ + +#include +#include + +void gma_power_init(struct drm_device *dev); +void gma_power_uninit(struct drm_device *dev); + +/* + * The kernel bus power management will call these functions + */ +int gma_power_suspend(struct device *dev); +int gma_power_resume(struct device *dev); + +/* + * These are the functions the driver should use to wrap all hw access + * (i.e. register reads and writes) + */ +bool gma_power_begin(struct drm_device *dev, bool force); +void gma_power_end(struct drm_device *dev); + +/* + * Use this function to do an instantaneous check for if the hw is on. + * Only use this in cases where you know the mutex is already held such + * as in irq install/uninstall and you need to + * prevent a deadlock situation. Otherwise use gma_power_begin(). + */ +bool gma_power_is_on(struct drm_device *dev); + +/* + * GFX-Runtime PM callbacks + */ +int psb_runtime_suspend(struct device *dev); +int psb_runtime_resume(struct device *dev); +int psb_runtime_idle(struct device *dev); + +#endif /*_PSB_POWERMGMT_H_*/ diff --git a/drivers/gpu/drm/gma500/psb_device.c b/drivers/gpu/drm/gma500/psb_device.c new file mode 100644 index 0000000..328a193 --- /dev/null +++ b/drivers/gpu/drm/gma500/psb_device.c @@ -0,0 +1,332 @@ +/************************************************************************** + * Copyright (c) 2011, Intel Corporation. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + **************************************************************************/ + +#include +#include +#include +#include "gma_drm.h" +#include "psb_drv.h" +#include "psb_reg.h" +#include "psb_intel_reg.h" +#include "intel_bios.h" + + +static int psb_output_init(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + psb_intel_lvds_init(dev, &dev_priv->mode_dev); + psb_intel_sdvo_init(dev, SDVOB); + return 0; +} + +#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE + +/* + * Poulsbo Backlight Interfaces + */ + +#define BLC_PWM_PRECISION_FACTOR 100 /* 10000000 */ +#define BLC_PWM_FREQ_CALC_CONSTANT 32 +#define MHz 1000000 + +#define PSB_BLC_PWM_PRECISION_FACTOR 10 +#define PSB_BLC_MAX_PWM_REG_FREQ 0xFFFE +#define PSB_BLC_MIN_PWM_REG_FREQ 0x2 + +#define PSB_BACKLIGHT_PWM_POLARITY_BIT_CLEAR (0xFFFE) +#define PSB_BACKLIGHT_PWM_CTL_SHIFT (16) + +static int psb_brightness; +static struct backlight_device *psb_backlight_device; + +static int psb_get_brightness(struct backlight_device *bd) +{ + /* return locally cached var instead of HW read (due to DPST etc.) */ + /* FIXME: ideally return actual value in case firmware fiddled with + it */ + return psb_brightness; +} + + +static int psb_backlight_setup(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + unsigned long core_clock; + /* u32 bl_max_freq; */ + /* unsigned long value; */ + u16 bl_max_freq; + uint32_t value; + uint32_t blc_pwm_precision_factor; + + /* get bl_max_freq and pol from dev_priv*/ + if (!dev_priv->lvds_bl) { + dev_err(dev->dev, "Has no valid LVDS backlight info\n"); + return -ENOENT; + } + bl_max_freq = dev_priv->lvds_bl->freq; + blc_pwm_precision_factor = PSB_BLC_PWM_PRECISION_FACTOR; + + core_clock = dev_priv->core_freq; + + value = (core_clock * MHz) / BLC_PWM_FREQ_CALC_CONSTANT; + value *= blc_pwm_precision_factor; + value /= bl_max_freq; + value /= blc_pwm_precision_factor; + + if (value > (unsigned long long)PSB_BLC_MAX_PWM_REG_FREQ || + value < (unsigned long long)PSB_BLC_MIN_PWM_REG_FREQ) + return -ERANGE; + else { + value &= PSB_BACKLIGHT_PWM_POLARITY_BIT_CLEAR; + REG_WRITE(BLC_PWM_CTL, + (value << PSB_BACKLIGHT_PWM_CTL_SHIFT) | (value)); + } + return 0; +} + +static int psb_set_brightness(struct backlight_device *bd) +{ + struct drm_device *dev = bl_get_data(psb_backlight_device); + int level = bd->props.brightness; + + /* Percentage 1-100% being valid */ + if (level < 1) + level = 1; + + psb_intel_lvds_set_brightness(dev, level); + psb_brightness = level; + return 0; +} + +static const struct backlight_ops psb_ops = { + .get_brightness = psb_get_brightness, + .update_status = psb_set_brightness, +}; + +static int psb_backlight_init(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + int ret; + struct backlight_properties props; + + memset(&props, 0, sizeof(struct backlight_properties)); + props.max_brightness = 100; + props.type = BACKLIGHT_PLATFORM; + + psb_backlight_device = backlight_device_register("psb-bl", + NULL, (void *)dev, &psb_ops, &props); + if (IS_ERR(psb_backlight_device)) + return PTR_ERR(psb_backlight_device); + + ret = psb_backlight_setup(dev); + if (ret < 0) { + backlight_device_unregister(psb_backlight_device); + psb_backlight_device = NULL; + return ret; + } + psb_backlight_device->props.brightness = 100; + psb_backlight_device->props.max_brightness = 100; + backlight_update_status(psb_backlight_device); + dev_priv->backlight_device = psb_backlight_device; + return 0; +} + +#endif + +/* + * Provide the Poulsbo specific chip logic and low level methods + * for power management + */ + +static void psb_init_pm(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + + u32 gating = PSB_RSGX32(PSB_CR_CLKGATECTL); + gating &= ~3; /* Disable 2D clock gating */ + gating |= 1; + PSB_WSGX32(gating, PSB_CR_CLKGATECTL); + PSB_RSGX32(PSB_CR_CLKGATECTL); +} + +/** + * psb_save_display_registers - save registers lost on suspend + * @dev: our DRM device + * + * Save the state we need in order to be able to restore the interface + * upon resume from suspend + */ +static int psb_save_display_registers(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct drm_crtc *crtc; + struct drm_connector *connector; + struct psb_state *regs = &dev_priv->regs.psb; + + /* Display arbitration control + watermarks */ + regs->saveDSPARB = PSB_RVDC32(DSPARB); + regs->saveDSPFW1 = PSB_RVDC32(DSPFW1); + regs->saveDSPFW2 = PSB_RVDC32(DSPFW2); + regs->saveDSPFW3 = PSB_RVDC32(DSPFW3); + regs->saveDSPFW4 = PSB_RVDC32(DSPFW4); + regs->saveDSPFW5 = PSB_RVDC32(DSPFW5); + regs->saveDSPFW6 = PSB_RVDC32(DSPFW6); + regs->saveCHICKENBIT = PSB_RVDC32(DSPCHICKENBIT); + + /* Save crtc and output state */ + mutex_lock(&dev->mode_config.mutex); + list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { + if (drm_helper_crtc_in_use(crtc)) + crtc->funcs->save(crtc); + } + + list_for_each_entry(connector, &dev->mode_config.connector_list, head) + if (connector->funcs->save) + connector->funcs->save(connector); + + mutex_unlock(&dev->mode_config.mutex); + return 0; +} + +/** + * psb_restore_display_registers - restore lost register state + * @dev: our DRM device + * + * Restore register state that was lost during suspend and resume. + */ +static int psb_restore_display_registers(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct drm_crtc *crtc; + struct drm_connector *connector; + struct psb_state *regs = &dev_priv->regs.psb; + + /* Display arbitration + watermarks */ + PSB_WVDC32(regs->saveDSPARB, DSPARB); + PSB_WVDC32(regs->saveDSPFW1, DSPFW1); + PSB_WVDC32(regs->saveDSPFW2, DSPFW2); + PSB_WVDC32(regs->saveDSPFW3, DSPFW3); + PSB_WVDC32(regs->saveDSPFW4, DSPFW4); + PSB_WVDC32(regs->saveDSPFW5, DSPFW5); + PSB_WVDC32(regs->saveDSPFW6, DSPFW6); + PSB_WVDC32(regs->saveCHICKENBIT, DSPCHICKENBIT); + + /*make sure VGA plane is off. it initializes to on after reset!*/ + PSB_WVDC32(0x80000000, VGACNTRL); + + mutex_lock(&dev->mode_config.mutex); + list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) + if (drm_helper_crtc_in_use(crtc)) + crtc->funcs->restore(crtc); + + list_for_each_entry(connector, &dev->mode_config.connector_list, head) + if (connector->funcs->restore) + connector->funcs->restore(connector); + + mutex_unlock(&dev->mode_config.mutex); + return 0; +} + +static int psb_power_down(struct drm_device *dev) +{ + return 0; +} + +static int psb_power_up(struct drm_device *dev) +{ + return 0; +} + +static void psb_get_core_freq(struct drm_device *dev) +{ + uint32_t clock; + struct pci_dev *pci_root = pci_get_bus_and_slot(0, 0); + struct drm_psb_private *dev_priv = dev->dev_private; + + /*pci_write_config_dword(pci_root, 0xD4, 0x00C32004);*/ + /*pci_write_config_dword(pci_root, 0xD0, 0xE0033000);*/ + + pci_write_config_dword(pci_root, 0xD0, 0xD0050300); + pci_read_config_dword(pci_root, 0xD4, &clock); + pci_dev_put(pci_root); + + switch (clock & 0x07) { + case 0: + dev_priv->core_freq = 100; + break; + case 1: + dev_priv->core_freq = 133; + break; + case 2: + dev_priv->core_freq = 150; + break; + case 3: + dev_priv->core_freq = 178; + break; + case 4: + dev_priv->core_freq = 200; + break; + case 5: + case 6: + case 7: + dev_priv->core_freq = 266; + default: + dev_priv->core_freq = 0; + } +} + +static int psb_chip_setup(struct drm_device *dev) +{ + psb_get_core_freq(dev); + gma_intel_setup_gmbus(dev); + gma_intel_opregion_init(dev); + psb_intel_init_bios(dev); + return 0; +} + +static void psb_chip_teardown(struct drm_device *dev) +{ + gma_intel_teardown_gmbus(dev); +} + +const struct psb_ops psb_chip_ops = { + .name = "Poulsbo", + .accel_2d = 1, + .pipes = 2, + .crtcs = 2, + .sgx_offset = PSB_SGX_OFFSET, + .chip_setup = psb_chip_setup, + .chip_teardown = psb_chip_teardown, + + .crtc_helper = &psb_intel_helper_funcs, + .crtc_funcs = &psb_intel_crtc_funcs, + + .output_init = psb_output_init, + +#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE + .backlight_init = psb_backlight_init, +#endif + + .init_pm = psb_init_pm, + .save_regs = psb_save_display_registers, + .restore_regs = psb_restore_display_registers, + .power_down = psb_power_down, + .power_up = psb_power_up, +}; + diff --git a/drivers/gpu/drm/gma500/psb_drv.c b/drivers/gpu/drm/gma500/psb_drv.c new file mode 100644 index 0000000..09af2ff --- /dev/null +++ b/drivers/gpu/drm/gma500/psb_drv.c @@ -0,0 +1,706 @@ +/************************************************************************** + * Copyright (c) 2007-2011, Intel Corporation. + * All Rights Reserved. + * Copyright (c) 2008, Tungsten Graphics, Inc. Cedar Park, TX., USA. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + **************************************************************************/ + +#include +#include +#include "gma_drm.h" +#include "psb_drv.h" +#include "framebuffer.h" +#include "psb_reg.h" +#include "psb_intel_reg.h" +#include "intel_bios.h" +#include "mid_bios.h" +#include +#include "power.h" +#include +#include +#include +#include +#include +#include + +static int drm_psb_trap_pagefaults; + +static int psb_probe(struct pci_dev *pdev, const struct pci_device_id *ent); + +MODULE_PARM_DESC(trap_pagefaults, "Error and reset on MMU pagefaults"); +module_param_named(trap_pagefaults, drm_psb_trap_pagefaults, int, 0600); + + +static DEFINE_PCI_DEVICE_TABLE(pciidlist) = { + { 0x8086, 0x8108, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &psb_chip_ops }, + { 0x8086, 0x8109, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &psb_chip_ops }, +#if defined(CONFIG_DRM_GMA600) + { 0x8086, 0x4100, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &oaktrail_chip_ops}, + { 0x8086, 0x4101, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &oaktrail_chip_ops}, + { 0x8086, 0x4102, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &oaktrail_chip_ops}, + { 0x8086, 0x4103, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &oaktrail_chip_ops}, + { 0x8086, 0x4104, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &oaktrail_chip_ops}, + { 0x8086, 0x4105, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &oaktrail_chip_ops}, + { 0x8086, 0x4106, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &oaktrail_chip_ops}, + { 0x8086, 0x4107, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &oaktrail_chip_ops}, + /* Atom E620 */ + { 0x8086, 0x4108, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &oaktrail_chip_ops}, +#endif +#if defined(CONFIG_DRM_MEDFIELD) + {0x8086, 0x0130, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &mdfld_chip_ops}, + {0x8086, 0x0131, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &mdfld_chip_ops}, + {0x8086, 0x0132, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &mdfld_chip_ops}, + {0x8086, 0x0133, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &mdfld_chip_ops}, + {0x8086, 0x0134, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &mdfld_chip_ops}, + {0x8086, 0x0135, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &mdfld_chip_ops}, + {0x8086, 0x0136, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &mdfld_chip_ops}, + {0x8086, 0x0137, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &mdfld_chip_ops}, +#endif +#if defined(CONFIG_DRM_GMA3600) + { 0x8086, 0x0be0, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &cdv_chip_ops}, + { 0x8086, 0x0be1, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &cdv_chip_ops}, + { 0x8086, 0x0be2, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &cdv_chip_ops}, + { 0x8086, 0x0be3, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &cdv_chip_ops}, + { 0x8086, 0x0be4, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &cdv_chip_ops}, + { 0x8086, 0x0be5, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &cdv_chip_ops}, + { 0x8086, 0x0be6, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &cdv_chip_ops}, + { 0x8086, 0x0be7, PCI_ANY_ID, PCI_ANY_ID, 0, 0, (long) &cdv_chip_ops}, +#endif + { 0, } +}; +MODULE_DEVICE_TABLE(pci, pciidlist); + +/* + * Standard IOCTLs. + */ + +#define DRM_IOCTL_GMA_ADB \ + DRM_IOWR(DRM_GMA_ADB + DRM_COMMAND_BASE, uint32_t) +#define DRM_IOCTL_GMA_MODE_OPERATION \ + DRM_IOWR(DRM_GMA_MODE_OPERATION + DRM_COMMAND_BASE, \ + struct drm_psb_mode_operation_arg) +#define DRM_IOCTL_GMA_STOLEN_MEMORY \ + DRM_IOWR(DRM_GMA_STOLEN_MEMORY + DRM_COMMAND_BASE, \ + struct drm_psb_stolen_memory_arg) +#define DRM_IOCTL_GMA_GAMMA \ + DRM_IOWR(DRM_GMA_GAMMA + DRM_COMMAND_BASE, \ + struct drm_psb_dpst_lut_arg) +#define DRM_IOCTL_GMA_DPST_BL \ + DRM_IOWR(DRM_GMA_DPST_BL + DRM_COMMAND_BASE, \ + uint32_t) +#define DRM_IOCTL_GMA_GET_PIPE_FROM_CRTC_ID \ + DRM_IOWR(DRM_GMA_GET_PIPE_FROM_CRTC_ID + DRM_COMMAND_BASE, \ + struct drm_psb_get_pipe_from_crtc_id_arg) +#define DRM_IOCTL_GMA_GEM_CREATE \ + DRM_IOWR(DRM_GMA_GEM_CREATE + DRM_COMMAND_BASE, \ + struct drm_psb_gem_create) +#define DRM_IOCTL_GMA_GEM_MMAP \ + DRM_IOWR(DRM_GMA_GEM_MMAP + DRM_COMMAND_BASE, \ + struct drm_psb_gem_mmap) + +static int psb_adb_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv); +static int psb_mode_operation_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv); +static int psb_stolen_memory_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv); +static int psb_gamma_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv); +static int psb_dpst_bl_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv); + +static struct drm_ioctl_desc psb_ioctls[] = { + DRM_IOCTL_DEF_DRV(GMA_ADB, psb_adb_ioctl, DRM_AUTH), + DRM_IOCTL_DEF_DRV(GMA_MODE_OPERATION, psb_mode_operation_ioctl, + DRM_AUTH), + DRM_IOCTL_DEF_DRV(GMA_STOLEN_MEMORY, psb_stolen_memory_ioctl, + DRM_AUTH), + DRM_IOCTL_DEF_DRV(GMA_GAMMA, psb_gamma_ioctl, DRM_AUTH), + DRM_IOCTL_DEF_DRV(GMA_DPST_BL, psb_dpst_bl_ioctl, DRM_AUTH), + DRM_IOCTL_DEF_DRV(GMA_GET_PIPE_FROM_CRTC_ID, + psb_intel_get_pipe_from_crtc_id, 0), + DRM_IOCTL_DEF_DRV(GMA_GEM_CREATE, psb_gem_create_ioctl, + DRM_UNLOCKED | DRM_AUTH), + DRM_IOCTL_DEF_DRV(GMA_GEM_MMAP, psb_gem_mmap_ioctl, + DRM_UNLOCKED | DRM_AUTH), +}; + +static void psb_lastclose(struct drm_device *dev) +{ + return; +} + +static void psb_do_takedown(struct drm_device *dev) +{ +} + +static int psb_do_init(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_gtt *pg = &dev_priv->gtt; + + uint32_t stolen_gtt; + + int ret = -ENOMEM; + + if (pg->mmu_gatt_start & 0x0FFFFFFF) { + dev_err(dev->dev, "Gatt must be 256M aligned. This is a bug.\n"); + ret = -EINVAL; + goto out_err; + } + + + stolen_gtt = (pg->stolen_size >> PAGE_SHIFT) * 4; + stolen_gtt = (stolen_gtt + PAGE_SIZE - 1) >> PAGE_SHIFT; + stolen_gtt = + (stolen_gtt < pg->gtt_pages) ? stolen_gtt : pg->gtt_pages; + + dev_priv->gatt_free_offset = pg->mmu_gatt_start + + (stolen_gtt << PAGE_SHIFT) * 1024; + + if (1 || drm_debug) { + uint32_t core_id = PSB_RSGX32(PSB_CR_CORE_ID); + uint32_t core_rev = PSB_RSGX32(PSB_CR_CORE_REVISION); + DRM_INFO("SGX core id = 0x%08x\n", core_id); + DRM_INFO("SGX core rev major = 0x%02x, minor = 0x%02x\n", + (core_rev & _PSB_CC_REVISION_MAJOR_MASK) >> + _PSB_CC_REVISION_MAJOR_SHIFT, + (core_rev & _PSB_CC_REVISION_MINOR_MASK) >> + _PSB_CC_REVISION_MINOR_SHIFT); + DRM_INFO + ("SGX core rev maintenance = 0x%02x, designer = 0x%02x\n", + (core_rev & _PSB_CC_REVISION_MAINTENANCE_MASK) >> + _PSB_CC_REVISION_MAINTENANCE_SHIFT, + (core_rev & _PSB_CC_REVISION_DESIGNER_MASK) >> + _PSB_CC_REVISION_DESIGNER_SHIFT); + } + + + spin_lock_init(&dev_priv->irqmask_lock); + spin_lock_init(&dev_priv->lock_2d); + + PSB_WSGX32(0x00000000, PSB_CR_BIF_BANK0); + PSB_WSGX32(0x00000000, PSB_CR_BIF_BANK1); + PSB_RSGX32(PSB_CR_BIF_BANK1); + PSB_WSGX32(PSB_RSGX32(PSB_CR_BIF_CTRL) | _PSB_MMU_ER_MASK, + PSB_CR_BIF_CTRL); + psb_spank(dev_priv); + + /* mmu_gatt ?? */ + PSB_WSGX32(pg->gatt_start, PSB_CR_BIF_TWOD_REQ_BASE); + return 0; +out_err: + psb_do_takedown(dev); + return ret; +} + +static int psb_driver_unload(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + + /* Kill vblank etc here */ + + gma_backlight_exit(dev); + + psb_modeset_cleanup(dev); + + if (dev_priv) { + psb_lid_timer_takedown(dev_priv); + gma_intel_opregion_exit(dev); + + if (dev_priv->ops->chip_teardown) + dev_priv->ops->chip_teardown(dev); + psb_do_takedown(dev); + + + if (dev_priv->pf_pd) { + psb_mmu_free_pagedir(dev_priv->pf_pd); + dev_priv->pf_pd = NULL; + } + if (dev_priv->mmu) { + struct psb_gtt *pg = &dev_priv->gtt; + + down_read(&pg->sem); + psb_mmu_remove_pfn_sequence( + psb_mmu_get_default_pd + (dev_priv->mmu), + pg->mmu_gatt_start, + dev_priv->vram_stolen_size >> PAGE_SHIFT); + up_read(&pg->sem); + psb_mmu_driver_takedown(dev_priv->mmu); + dev_priv->mmu = NULL; + } + psb_gtt_takedown(dev); + if (dev_priv->scratch_page) { + __free_page(dev_priv->scratch_page); + dev_priv->scratch_page = NULL; + } + if (dev_priv->vdc_reg) { + iounmap(dev_priv->vdc_reg); + dev_priv->vdc_reg = NULL; + } + if (dev_priv->sgx_reg) { + iounmap(dev_priv->sgx_reg); + dev_priv->sgx_reg = NULL; + } + + kfree(dev_priv); + dev->dev_private = NULL; + + /*destroy VBT data*/ + psb_intel_destroy_bios(dev); + } + + gma_power_uninit(dev); + + return 0; +} + + +static int psb_driver_load(struct drm_device *dev, unsigned long chipset) +{ + struct drm_psb_private *dev_priv; + unsigned long resource_start; + unsigned long irqflags; + int ret = -ENOMEM; + struct drm_connector *connector; + struct psb_intel_encoder *psb_intel_encoder; + + dev_priv = kzalloc(sizeof(*dev_priv), GFP_KERNEL); + if (dev_priv == NULL) + return -ENOMEM; + + dev_priv->ops = (struct psb_ops *)chipset; + dev_priv->dev = dev; + dev->dev_private = (void *) dev_priv; + + pci_set_master(dev->pdev); + + if (!IS_PSB(dev)) { + if (pci_enable_msi(dev->pdev)) + dev_warn(dev->dev, "Enabling MSI failed!\n"); + } + + dev_priv->num_pipe = dev_priv->ops->pipes; + + resource_start = pci_resource_start(dev->pdev, PSB_MMIO_RESOURCE); + + dev_priv->vdc_reg = + ioremap(resource_start + PSB_VDC_OFFSET, PSB_VDC_SIZE); + if (!dev_priv->vdc_reg) + goto out_err; + + dev_priv->sgx_reg = ioremap(resource_start + dev_priv->ops->sgx_offset, + PSB_SGX_SIZE); + if (!dev_priv->sgx_reg) + goto out_err; + + ret = dev_priv->ops->chip_setup(dev); + if (ret) + goto out_err; + + /* Init OSPM support */ + gma_power_init(dev); + + ret = -ENOMEM; + + dev_priv->scratch_page = alloc_page(GFP_DMA32 | __GFP_ZERO); + if (!dev_priv->scratch_page) + goto out_err; + + set_pages_uc(dev_priv->scratch_page, 1); + + ret = psb_gtt_init(dev, 0); + if (ret) + goto out_err; + + dev_priv->mmu = psb_mmu_driver_init((void *)0, + drm_psb_trap_pagefaults, 0, + dev_priv); + if (!dev_priv->mmu) + goto out_err; + + dev_priv->pf_pd = psb_mmu_alloc_pd(dev_priv->mmu, 1, 0); + if (!dev_priv->pf_pd) + goto out_err; + + psb_mmu_set_pd_context(psb_mmu_get_default_pd(dev_priv->mmu), 0); + psb_mmu_set_pd_context(dev_priv->pf_pd, 1); + + ret = psb_do_init(dev); + if (ret) + return ret; + + PSB_WSGX32(0x20000000, PSB_CR_PDS_EXEC_BASE); + PSB_WSGX32(0x30000000, PSB_CR_BIF_3D_REQ_BASE); + +/* igd_opregion_init(&dev_priv->opregion_dev); */ +/* acpi_video_register(); */ + if (dev_priv->lid_state) + psb_lid_timer_init(dev_priv); + + ret = drm_vblank_init(dev, dev_priv->num_pipe); + if (ret) + goto out_err; + + /* + * Install interrupt handlers prior to powering off SGX or else we will + * crash. + */ + dev_priv->vdc_irq_mask = 0; + dev_priv->pipestat[0] = 0; + dev_priv->pipestat[1] = 0; + dev_priv->pipestat[2] = 0; + spin_lock_irqsave(&dev_priv->irqmask_lock, irqflags); + PSB_WVDC32(0xFFFFFFFF, PSB_HWSTAM); + PSB_WVDC32(0x00000000, PSB_INT_ENABLE_R); + PSB_WVDC32(0xFFFFFFFF, PSB_INT_MASK_R); + spin_unlock_irqrestore(&dev_priv->irqmask_lock, irqflags); + if (IS_PSB(dev) && drm_core_check_feature(dev, DRIVER_MODESET)) + drm_irq_install(dev); + + dev->vblank_disable_allowed = 1; + + dev->max_vblank_count = 0xffffff; /* only 24 bits of frame count */ + + dev->driver->get_vblank_counter = psb_get_vblank_counter; + + psb_modeset_init(dev); + psb_fbdev_init(dev); + drm_kms_helper_poll_init(dev); + + /* Only add backlight support if we have LVDS output */ + list_for_each_entry(connector, &dev->mode_config.connector_list, + head) { + psb_intel_encoder = psb_intel_attached_encoder(connector); + + switch (psb_intel_encoder->type) { + case INTEL_OUTPUT_LVDS: + case INTEL_OUTPUT_MIPI: + ret = gma_backlight_init(dev); + break; + } + } + + if (ret) + return ret; +#if 0 + /*enable runtime pm at last*/ + pm_runtime_enable(&dev->pdev->dev); + pm_runtime_set_active(&dev->pdev->dev); +#endif + /*Intel drm driver load is done, continue doing pvr load*/ + return 0; +out_err: + psb_driver_unload(dev); + return ret; +} + +static int psb_driver_device_is_agp(struct drm_device *dev) +{ + return 0; +} + +static inline void get_brightness(struct backlight_device *bd) +{ +#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE + if (bd) { + bd->props.brightness = bd->ops->get_brightness(bd); + backlight_update_status(bd); + } +#endif +} + +static int psb_dpst_bl_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + struct drm_psb_private *dev_priv = psb_priv(dev); + uint32_t *arg = data; + + dev_priv->blc_adj2 = *arg; + get_brightness(dev_priv->backlight_device); + return 0; +} + +static int psb_adb_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + struct drm_psb_private *dev_priv = psb_priv(dev); + uint32_t *arg = data; + + dev_priv->blc_adj1 = *arg; + get_brightness(dev_priv->backlight_device); + return 0; +} + +static int psb_gamma_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + struct drm_psb_dpst_lut_arg *lut_arg = data; + struct drm_mode_object *obj; + struct drm_crtc *crtc; + struct drm_connector *connector; + struct psb_intel_crtc *psb_intel_crtc; + int i = 0; + int32_t obj_id; + + obj_id = lut_arg->output_id; + obj = drm_mode_object_find(dev, obj_id, DRM_MODE_OBJECT_CONNECTOR); + if (!obj) { + dev_dbg(dev->dev, "Invalid Connector object.\n"); + return -EINVAL; + } + + connector = obj_to_connector(obj); + crtc = connector->encoder->crtc; + psb_intel_crtc = to_psb_intel_crtc(crtc); + + for (i = 0; i < 256; i++) + psb_intel_crtc->lut_adj[i] = lut_arg->lut[i]; + + psb_intel_crtc_load_lut(crtc); + + return 0; +} + +static int psb_mode_operation_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + uint32_t obj_id; + uint16_t op; + struct drm_mode_modeinfo *umode; + struct drm_display_mode *mode = NULL; + struct drm_psb_mode_operation_arg *arg; + struct drm_mode_object *obj; + struct drm_connector *connector; + struct drm_connector_helper_funcs *connector_funcs; + int ret = 0; + int resp = MODE_OK; + + arg = (struct drm_psb_mode_operation_arg *)data; + obj_id = arg->obj_id; + op = arg->operation; + + switch (op) { + case PSB_MODE_OPERATION_MODE_VALID: + umode = &arg->mode; + + mutex_lock(&dev->mode_config.mutex); + + obj = drm_mode_object_find(dev, obj_id, + DRM_MODE_OBJECT_CONNECTOR); + if (!obj) { + ret = -EINVAL; + goto mode_op_out; + } + + connector = obj_to_connector(obj); + + mode = drm_mode_create(dev); + if (!mode) { + ret = -ENOMEM; + goto mode_op_out; + } + + /* drm_crtc_convert_umode(mode, umode); */ + { + mode->clock = umode->clock; + mode->hdisplay = umode->hdisplay; + mode->hsync_start = umode->hsync_start; + mode->hsync_end = umode->hsync_end; + mode->htotal = umode->htotal; + mode->hskew = umode->hskew; + mode->vdisplay = umode->vdisplay; + mode->vsync_start = umode->vsync_start; + mode->vsync_end = umode->vsync_end; + mode->vtotal = umode->vtotal; + mode->vscan = umode->vscan; + mode->vrefresh = umode->vrefresh; + mode->flags = umode->flags; + mode->type = umode->type; + strncpy(mode->name, umode->name, DRM_DISPLAY_MODE_LEN); + mode->name[DRM_DISPLAY_MODE_LEN-1] = 0; + } + + connector_funcs = (struct drm_connector_helper_funcs *) + connector->helper_private; + + if (connector_funcs->mode_valid) { + resp = connector_funcs->mode_valid(connector, mode); + arg->data = resp; + } + + /*do some clean up work*/ + if (mode) + drm_mode_destroy(dev, mode); +mode_op_out: + mutex_unlock(&dev->mode_config.mutex); + return ret; + + default: + dev_dbg(dev->dev, "Unsupported psb mode operation\n"); + return -EOPNOTSUPP; + } + + return 0; +} + +static int psb_stolen_memory_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + struct drm_psb_private *dev_priv = psb_priv(dev); + struct drm_psb_stolen_memory_arg *arg = data; + + arg->base = dev_priv->stolen_base; + arg->size = dev_priv->vram_stolen_size; + + return 0; +} + +static int psb_driver_open(struct drm_device *dev, struct drm_file *priv) +{ + return 0; +} + +static void psb_driver_close(struct drm_device *dev, struct drm_file *priv) +{ +} + +static long psb_unlocked_ioctl(struct file *filp, unsigned int cmd, + unsigned long arg) +{ + struct drm_file *file_priv = filp->private_data; + struct drm_device *dev = file_priv->minor->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + static unsigned int runtime_allowed; + + if (runtime_allowed == 1 && dev_priv->is_lvds_on) { + runtime_allowed++; + pm_runtime_allow(&dev->pdev->dev); + dev_priv->rpm_enabled = 1; + } + return drm_ioctl(filp, cmd, arg); + /* FIXME: do we need to wrap the other side of this */ +} + + +/* When a client dies: + * - Check for and clean up flipped page state + */ +static void psb_driver_preclose(struct drm_device *dev, struct drm_file *priv) +{ +} + +static void psb_remove(struct pci_dev *pdev) +{ + struct drm_device *dev = pci_get_drvdata(pdev); + drm_put_dev(dev); +} + +static const struct dev_pm_ops psb_pm_ops = { + .resume = gma_power_resume, + .suspend = gma_power_suspend, + .runtime_suspend = psb_runtime_suspend, + .runtime_resume = psb_runtime_resume, + .runtime_idle = psb_runtime_idle, +}; + +static struct vm_operations_struct psb_gem_vm_ops = { + .fault = psb_gem_fault, + .open = drm_gem_vm_open, + .close = drm_gem_vm_close, +}; + +static const struct file_operations psb_gem_fops = { + .owner = THIS_MODULE, + .open = drm_open, + .release = drm_release, + .unlocked_ioctl = psb_unlocked_ioctl, + .mmap = drm_gem_mmap, + .poll = drm_poll, + .fasync = drm_fasync, + .read = drm_read, +}; + +static struct drm_driver driver = { + .driver_features = DRIVER_HAVE_IRQ | DRIVER_IRQ_SHARED | \ + DRIVER_IRQ_VBL | DRIVER_MODESET | DRIVER_GEM , + .load = psb_driver_load, + .unload = psb_driver_unload, + + .ioctls = psb_ioctls, + .num_ioctls = DRM_ARRAY_SIZE(psb_ioctls), + .device_is_agp = psb_driver_device_is_agp, + .irq_preinstall = psb_irq_preinstall, + .irq_postinstall = psb_irq_postinstall, + .irq_uninstall = psb_irq_uninstall, + .irq_handler = psb_irq_handler, + .enable_vblank = psb_enable_vblank, + .disable_vblank = psb_disable_vblank, + .get_vblank_counter = psb_get_vblank_counter, + .lastclose = psb_lastclose, + .open = psb_driver_open, + .preclose = psb_driver_preclose, + .postclose = psb_driver_close, + .reclaim_buffers = drm_core_reclaim_buffers, + + .gem_init_object = psb_gem_init_object, + .gem_free_object = psb_gem_free_object, + .gem_vm_ops = &psb_gem_vm_ops, + .dumb_create = psb_gem_dumb_create, + .dumb_map_offset = psb_gem_dumb_map_gtt, + .dumb_destroy = psb_gem_dumb_destroy, + .fops = &psb_gem_fops, + .name = DRIVER_NAME, + .desc = DRIVER_DESC, + .date = PSB_DRM_DRIVER_DATE, + .major = PSB_DRM_DRIVER_MAJOR, + .minor = PSB_DRM_DRIVER_MINOR, + .patchlevel = PSB_DRM_DRIVER_PATCHLEVEL +}; + +static struct pci_driver psb_pci_driver = { + .name = DRIVER_NAME, + .id_table = pciidlist, + .probe = psb_probe, + .remove = psb_remove, + .driver = { + .pm = &psb_pm_ops, + } +}; + +static int psb_probe(struct pci_dev *pdev, const struct pci_device_id *ent) +{ + return drm_get_pci_dev(pdev, ent, &driver); +} + +static int __init psb_init(void) +{ + return drm_pci_init(&driver, &psb_pci_driver); +} + +static void __exit psb_exit(void) +{ + drm_pci_exit(&driver, &psb_pci_driver); +} + +late_initcall(psb_init); +module_exit(psb_exit); + +MODULE_AUTHOR("Alan Cox and others"); +MODULE_DESCRIPTION(DRIVER_DESC); +MODULE_LICENSE("GPL"); diff --git a/drivers/gpu/drm/gma500/psb_drv.h b/drivers/gpu/drm/gma500/psb_drv.h new file mode 100644 index 0000000..40ce2c9 --- /dev/null +++ b/drivers/gpu/drm/gma500/psb_drv.h @@ -0,0 +1,993 @@ +/************************************************************************** + * Copyright (c) 2007-2011, Intel Corporation. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + **************************************************************************/ + +#ifndef _PSB_DRV_H_ +#define _PSB_DRV_H_ + +#include + +#include +#include "drm_global.h" +#include "gem_glue.h" +#include "gma_drm.h" +#include "psb_reg.h" +#include "psb_intel_drv.h" +#include "gtt.h" +#include "power.h" +#include "oaktrail.h" + +/* Append new drm mode definition here, align with libdrm definition */ +#define DRM_MODE_SCALE_NO_SCALE 2 + +enum { + CHIP_PSB_8108 = 0, /* Poulsbo */ + CHIP_PSB_8109 = 1, /* Poulsbo */ + CHIP_MRST_4100 = 2, /* Moorestown/Oaktrail */ + CHIP_MFLD_0130 = 3, /* Medfield */ +}; + +#define IS_PSB(dev) (((dev)->pci_device & 0xfffe) == 0x8108) +#define IS_MRST(dev) (((dev)->pci_device & 0xfffc) == 0x4100) +#define IS_MFLD(dev) (((dev)->pci_device & 0xfff8) == 0x0130) + +/* + * Driver definitions + */ + +#define DRIVER_NAME "gma500" +#define DRIVER_DESC "DRM driver for the Intel GMA500" + +#define PSB_DRM_DRIVER_DATE "2011-06-06" +#define PSB_DRM_DRIVER_MAJOR 1 +#define PSB_DRM_DRIVER_MINOR 0 +#define PSB_DRM_DRIVER_PATCHLEVEL 0 + +/* + * Hardware offsets + */ +#define PSB_VDC_OFFSET 0x00000000 +#define PSB_VDC_SIZE 0x000080000 +#define MRST_MMIO_SIZE 0x0000C0000 +#define MDFLD_MMIO_SIZE 0x000100000 +#define PSB_SGX_SIZE 0x8000 +#define PSB_SGX_OFFSET 0x00040000 +#define MRST_SGX_OFFSET 0x00080000 +/* + * PCI resource identifiers + */ +#define PSB_MMIO_RESOURCE 0 +#define PSB_GATT_RESOURCE 2 +#define PSB_GTT_RESOURCE 3 +/* + * PCI configuration + */ +#define PSB_GMCH_CTRL 0x52 +#define PSB_BSM 0x5C +#define _PSB_GMCH_ENABLED 0x4 +#define PSB_PGETBL_CTL 0x2020 +#define _PSB_PGETBL_ENABLED 0x00000001 +#define PSB_SGX_2D_SLAVE_PORT 0x4000 + +/* To get rid of */ +#define PSB_TT_PRIV0_LIMIT (256*1024*1024) +#define PSB_TT_PRIV0_PLIMIT (PSB_TT_PRIV0_LIMIT >> PAGE_SHIFT) + +/* + * SGX side MMU definitions (these can probably go) + */ + +/* + * Flags for external memory type field. + */ +#define PSB_MMU_CACHED_MEMORY 0x0001 /* Bind to MMU only */ +#define PSB_MMU_RO_MEMORY 0x0002 /* MMU RO memory */ +#define PSB_MMU_WO_MEMORY 0x0004 /* MMU WO memory */ +/* + * PTE's and PDE's + */ +#define PSB_PDE_MASK 0x003FFFFF +#define PSB_PDE_SHIFT 22 +#define PSB_PTE_SHIFT 12 +/* + * Cache control + */ +#define PSB_PTE_VALID 0x0001 /* PTE / PDE valid */ +#define PSB_PTE_WO 0x0002 /* Write only */ +#define PSB_PTE_RO 0x0004 /* Read only */ +#define PSB_PTE_CACHED 0x0008 /* CPU cache coherent */ + +/* + * VDC registers and bits + */ +#define PSB_MSVDX_CLOCKGATING 0x2064 +#define PSB_TOPAZ_CLOCKGATING 0x2068 +#define PSB_HWSTAM 0x2098 +#define PSB_INSTPM 0x20C0 +#define PSB_INT_IDENTITY_R 0x20A4 +#define _MDFLD_PIPEC_EVENT_FLAG (1<<2) +#define _MDFLD_PIPEC_VBLANK_FLAG (1<<3) +#define _PSB_DPST_PIPEB_FLAG (1<<4) +#define _MDFLD_PIPEB_EVENT_FLAG (1<<4) +#define _PSB_VSYNC_PIPEB_FLAG (1<<5) +#define _PSB_DPST_PIPEA_FLAG (1<<6) +#define _PSB_PIPEA_EVENT_FLAG (1<<6) +#define _PSB_VSYNC_PIPEA_FLAG (1<<7) +#define _MDFLD_MIPIA_FLAG (1<<16) +#define _MDFLD_MIPIC_FLAG (1<<17) +#define _PSB_IRQ_SGX_FLAG (1<<18) +#define _PSB_IRQ_MSVDX_FLAG (1<<19) +#define _LNC_IRQ_TOPAZ_FLAG (1<<20) + +#define _PSB_PIPE_EVENT_FLAG (_PSB_VSYNC_PIPEA_FLAG | \ + _PSB_VSYNC_PIPEB_FLAG) + +/* This flag includes all the display IRQ bits excepts the vblank irqs. */ +#define _MDFLD_DISP_ALL_IRQ_FLAG (_MDFLD_PIPEC_EVENT_FLAG | \ + _MDFLD_PIPEB_EVENT_FLAG | \ + _PSB_PIPEA_EVENT_FLAG | \ + _PSB_VSYNC_PIPEA_FLAG | \ + _MDFLD_MIPIA_FLAG | \ + _MDFLD_MIPIC_FLAG) +#define PSB_INT_IDENTITY_R 0x20A4 +#define PSB_INT_MASK_R 0x20A8 +#define PSB_INT_ENABLE_R 0x20A0 + +#define _PSB_MMU_ER_MASK 0x0001FF00 +#define _PSB_MMU_ER_HOST (1 << 16) +#define GPIOA 0x5010 +#define GPIOB 0x5014 +#define GPIOC 0x5018 +#define GPIOD 0x501c +#define GPIOE 0x5020 +#define GPIOF 0x5024 +#define GPIOG 0x5028 +#define GPIOH 0x502c +#define GPIO_CLOCK_DIR_MASK (1 << 0) +#define GPIO_CLOCK_DIR_IN (0 << 1) +#define GPIO_CLOCK_DIR_OUT (1 << 1) +#define GPIO_CLOCK_VAL_MASK (1 << 2) +#define GPIO_CLOCK_VAL_OUT (1 << 3) +#define GPIO_CLOCK_VAL_IN (1 << 4) +#define GPIO_CLOCK_PULLUP_DISABLE (1 << 5) +#define GPIO_DATA_DIR_MASK (1 << 8) +#define GPIO_DATA_DIR_IN (0 << 9) +#define GPIO_DATA_DIR_OUT (1 << 9) +#define GPIO_DATA_VAL_MASK (1 << 10) +#define GPIO_DATA_VAL_OUT (1 << 11) +#define GPIO_DATA_VAL_IN (1 << 12) +#define GPIO_DATA_PULLUP_DISABLE (1 << 13) + +#define VCLK_DIVISOR_VGA0 0x6000 +#define VCLK_DIVISOR_VGA1 0x6004 +#define VCLK_POST_DIV 0x6010 + +#define PSB_COMM_2D (PSB_ENGINE_2D << 4) +#define PSB_COMM_3D (PSB_ENGINE_3D << 4) +#define PSB_COMM_TA (PSB_ENGINE_TA << 4) +#define PSB_COMM_HP (PSB_ENGINE_HP << 4) +#define PSB_COMM_USER_IRQ (1024 >> 2) +#define PSB_COMM_USER_IRQ_LOST (PSB_COMM_USER_IRQ + 1) +#define PSB_COMM_FW (2048 >> 2) + +#define PSB_UIRQ_VISTEST 1 +#define PSB_UIRQ_OOM_REPLY 2 +#define PSB_UIRQ_FIRE_TA_REPLY 3 +#define PSB_UIRQ_FIRE_RASTER_REPLY 4 + +#define PSB_2D_SIZE (256*1024*1024) +#define PSB_MAX_RELOC_PAGES 1024 + +#define PSB_LOW_REG_OFFS 0x0204 +#define PSB_HIGH_REG_OFFS 0x0600 + +#define PSB_NUM_VBLANKS 2 + + +#define PSB_2D_SIZE (256*1024*1024) +#define PSB_MAX_RELOC_PAGES 1024 + +#define PSB_LOW_REG_OFFS 0x0204 +#define PSB_HIGH_REG_OFFS 0x0600 + +#define PSB_NUM_VBLANKS 2 +#define PSB_WATCHDOG_DELAY (DRM_HZ * 2) +#define PSB_LID_DELAY (DRM_HZ / 10) + +#define MDFLD_PNW_B0 0x04 +#define MDFLD_PNW_C0 0x08 + +#define MDFLD_DSR_2D_3D_0 (1 << 0) +#define MDFLD_DSR_2D_3D_2 (1 << 1) +#define MDFLD_DSR_CURSOR_0 (1 << 2) +#define MDFLD_DSR_CURSOR_2 (1 << 3) +#define MDFLD_DSR_OVERLAY_0 (1 << 4) +#define MDFLD_DSR_OVERLAY_2 (1 << 5) +#define MDFLD_DSR_MIPI_CONTROL (1 << 6) +#define MDFLD_DSR_DAMAGE_MASK_0 ((1 << 0) | (1 << 2) | (1 << 4)) +#define MDFLD_DSR_DAMAGE_MASK_2 ((1 << 1) | (1 << 3) | (1 << 5)) +#define MDFLD_DSR_2D_3D (MDFLD_DSR_2D_3D_0 | MDFLD_DSR_2D_3D_2) + +#define MDFLD_DSR_RR 45 +#define MDFLD_DPU_ENABLE (1 << 31) +#define MDFLD_DSR_FULLSCREEN (1 << 30) +#define MDFLD_DSR_DELAY (DRM_HZ / MDFLD_DSR_RR) + +#define PSB_PWR_STATE_ON 1 +#define PSB_PWR_STATE_OFF 2 + +#define PSB_PMPOLICY_NOPM 0 +#define PSB_PMPOLICY_CLOCKGATING 1 +#define PSB_PMPOLICY_POWERDOWN 2 + +#define PSB_PMSTATE_POWERUP 0 +#define PSB_PMSTATE_CLOCKGATED 1 +#define PSB_PMSTATE_POWERDOWN 2 +#define PSB_PCIx_MSI_ADDR_LOC 0x94 +#define PSB_PCIx_MSI_DATA_LOC 0x98 + +/* Medfield crystal settings */ +#define KSEL_CRYSTAL_19 1 +#define KSEL_BYPASS_19 5 +#define KSEL_BYPASS_25 6 +#define KSEL_BYPASS_83_100 7 + +struct opregion_header; +struct opregion_acpi; +struct opregion_swsci; +struct opregion_asle; + +struct psb_intel_opregion { + struct opregion_header *header; + struct opregion_acpi *acpi; + struct opregion_swsci *swsci; + struct opregion_asle *asle; + int enabled; +}; + +struct sdvo_device_mapping { + u8 initialized; + u8 dvo_port; + u8 slave_addr; + u8 dvo_wiring; + u8 i2c_pin; + u8 i2c_speed; + u8 ddc_pin; +}; + +struct intel_gmbus { + struct i2c_adapter adapter; + struct i2c_adapter *force_bit; + u32 reg0; +}; + +/* + * Register save state. This is used to hold the context when the + * device is powered off. In the case of Oaktrail this can (but does not + * yet) include screen blank. Operations occuring during the save + * update the register cache instead. + */ +struct psb_state { + uint32_t saveDSPACNTR; + uint32_t saveDSPBCNTR; + uint32_t savePIPEACONF; + uint32_t savePIPEBCONF; + uint32_t savePIPEASRC; + uint32_t savePIPEBSRC; + uint32_t saveFPA0; + uint32_t saveFPA1; + uint32_t saveDPLL_A; + uint32_t saveDPLL_A_MD; + uint32_t saveHTOTAL_A; + uint32_t saveHBLANK_A; + uint32_t saveHSYNC_A; + uint32_t saveVTOTAL_A; + uint32_t saveVBLANK_A; + uint32_t saveVSYNC_A; + uint32_t saveDSPASTRIDE; + uint32_t saveDSPASIZE; + uint32_t saveDSPAPOS; + uint32_t saveDSPABASE; + uint32_t saveDSPASURF; + uint32_t saveDSPASTATUS; + uint32_t saveFPB0; + uint32_t saveFPB1; + uint32_t saveDPLL_B; + uint32_t saveDPLL_B_MD; + uint32_t saveHTOTAL_B; + uint32_t saveHBLANK_B; + uint32_t saveHSYNC_B; + uint32_t saveVTOTAL_B; + uint32_t saveVBLANK_B; + uint32_t saveVSYNC_B; + uint32_t saveDSPBSTRIDE; + uint32_t saveDSPBSIZE; + uint32_t saveDSPBPOS; + uint32_t saveDSPBBASE; + uint32_t saveDSPBSURF; + uint32_t saveDSPBSTATUS; + uint32_t saveVCLK_DIVISOR_VGA0; + uint32_t saveVCLK_DIVISOR_VGA1; + uint32_t saveVCLK_POST_DIV; + uint32_t saveVGACNTRL; + uint32_t saveADPA; + uint32_t saveLVDS; + uint32_t saveDVOA; + uint32_t saveDVOB; + uint32_t saveDVOC; + uint32_t savePP_ON; + uint32_t savePP_OFF; + uint32_t savePP_CONTROL; + uint32_t savePP_CYCLE; + uint32_t savePFIT_CONTROL; + uint32_t savePaletteA[256]; + uint32_t savePaletteB[256]; + uint32_t saveCLOCKGATING; + uint32_t saveDSPARB; + uint32_t saveDSPATILEOFF; + uint32_t saveDSPBTILEOFF; + uint32_t saveDSPAADDR; + uint32_t saveDSPBADDR; + uint32_t savePFIT_AUTO_RATIOS; + uint32_t savePFIT_PGM_RATIOS; + uint32_t savePP_ON_DELAYS; + uint32_t savePP_OFF_DELAYS; + uint32_t savePP_DIVISOR; + uint32_t saveBCLRPAT_A; + uint32_t saveBCLRPAT_B; + uint32_t saveDSPALINOFF; + uint32_t saveDSPBLINOFF; + uint32_t savePERF_MODE; + uint32_t saveDSPFW1; + uint32_t saveDSPFW2; + uint32_t saveDSPFW3; + uint32_t saveDSPFW4; + uint32_t saveDSPFW5; + uint32_t saveDSPFW6; + uint32_t saveCHICKENBIT; + uint32_t saveDSPACURSOR_CTRL; + uint32_t saveDSPBCURSOR_CTRL; + uint32_t saveDSPACURSOR_BASE; + uint32_t saveDSPBCURSOR_BASE; + uint32_t saveDSPACURSOR_POS; + uint32_t saveDSPBCURSOR_POS; + uint32_t save_palette_a[256]; + uint32_t save_palette_b[256]; + uint32_t saveOV_OVADD; + uint32_t saveOV_OGAMC0; + uint32_t saveOV_OGAMC1; + uint32_t saveOV_OGAMC2; + uint32_t saveOV_OGAMC3; + uint32_t saveOV_OGAMC4; + uint32_t saveOV_OGAMC5; + uint32_t saveOVC_OVADD; + uint32_t saveOVC_OGAMC0; + uint32_t saveOVC_OGAMC1; + uint32_t saveOVC_OGAMC2; + uint32_t saveOVC_OGAMC3; + uint32_t saveOVC_OGAMC4; + uint32_t saveOVC_OGAMC5; + + /* DPST register save */ + uint32_t saveHISTOGRAM_INT_CONTROL_REG; + uint32_t saveHISTOGRAM_LOGIC_CONTROL_REG; + uint32_t savePWM_CONTROL_LOGIC; +}; + +struct medfield_state { + uint32_t saveDPLL_A; + uint32_t saveFPA0; + uint32_t savePIPEACONF; + uint32_t saveHTOTAL_A; + uint32_t saveHBLANK_A; + uint32_t saveHSYNC_A; + uint32_t saveVTOTAL_A; + uint32_t saveVBLANK_A; + uint32_t saveVSYNC_A; + uint32_t savePIPEASRC; + uint32_t saveDSPASTRIDE; + uint32_t saveDSPALINOFF; + uint32_t saveDSPATILEOFF; + uint32_t saveDSPASIZE; + uint32_t saveDSPAPOS; + uint32_t saveDSPASURF; + uint32_t saveDSPACNTR; + uint32_t saveDSPASTATUS; + uint32_t save_palette_a[256]; + uint32_t saveMIPI; + + uint32_t saveDPLL_B; + uint32_t saveFPB0; + uint32_t savePIPEBCONF; + uint32_t saveHTOTAL_B; + uint32_t saveHBLANK_B; + uint32_t saveHSYNC_B; + uint32_t saveVTOTAL_B; + uint32_t saveVBLANK_B; + uint32_t saveVSYNC_B; + uint32_t savePIPEBSRC; + uint32_t saveDSPBSTRIDE; + uint32_t saveDSPBLINOFF; + uint32_t saveDSPBTILEOFF; + uint32_t saveDSPBSIZE; + uint32_t saveDSPBPOS; + uint32_t saveDSPBSURF; + uint32_t saveDSPBCNTR; + uint32_t saveDSPBSTATUS; + uint32_t save_palette_b[256]; + + uint32_t savePIPECCONF; + uint32_t saveHTOTAL_C; + uint32_t saveHBLANK_C; + uint32_t saveHSYNC_C; + uint32_t saveVTOTAL_C; + uint32_t saveVBLANK_C; + uint32_t saveVSYNC_C; + uint32_t savePIPECSRC; + uint32_t saveDSPCSTRIDE; + uint32_t saveDSPCLINOFF; + uint32_t saveDSPCTILEOFF; + uint32_t saveDSPCSIZE; + uint32_t saveDSPCPOS; + uint32_t saveDSPCSURF; + uint32_t saveDSPCCNTR; + uint32_t saveDSPCSTATUS; + uint32_t save_palette_c[256]; + uint32_t saveMIPI_C; + + uint32_t savePFIT_CONTROL; + uint32_t savePFIT_PGM_RATIOS; + uint32_t saveHDMIPHYMISCCTL; + uint32_t saveHDMIB_CONTROL; +}; + +struct cdv_state { + uint32_t saveDSPCLK_GATE_D; + uint32_t saveRAMCLK_GATE_D; + uint32_t saveDSPARB; + uint32_t saveDSPFW[6]; + uint32_t saveADPA; + uint32_t savePP_CONTROL; + uint32_t savePFIT_PGM_RATIOS; + uint32_t saveLVDS; + uint32_t savePFIT_CONTROL; + uint32_t savePP_ON_DELAYS; + uint32_t savePP_OFF_DELAYS; + uint32_t savePP_CYCLE; + uint32_t saveVGACNTRL; + uint32_t saveIER; + uint32_t saveIMR; + u8 saveLBB; +}; + +struct psb_save_area { + uint32_t saveBSM; + uint32_t saveVBT; + union { + struct psb_state psb; + struct medfield_state mdfld; + struct cdv_state cdv; + }; + uint32_t saveBLC_PWM_CTL2; + uint32_t saveBLC_PWM_CTL; +}; + +struct psb_ops; + +#define PSB_NUM_PIPE 3 + +struct drm_psb_private { + struct drm_device *dev; + const struct psb_ops *ops; + + struct psb_gtt gtt; + + /* GTT Memory manager */ + struct psb_gtt_mm *gtt_mm; + struct page *scratch_page; + u32 *gtt_map; + uint32_t stolen_base; + void *vram_addr; + unsigned long vram_stolen_size; + int gtt_initialized; + u16 gmch_ctrl; /* Saved GTT setup */ + u32 pge_ctl; + + struct mutex gtt_mutex; + struct resource *gtt_mem; /* Our PCI resource */ + + struct psb_mmu_driver *mmu; + struct psb_mmu_pd *pf_pd; + + /* + * Register base + */ + + uint8_t *sgx_reg; + uint8_t *vdc_reg; + uint32_t gatt_free_offset; + + /* + * Fencing / irq. + */ + + uint32_t vdc_irq_mask; + uint32_t pipestat[PSB_NUM_PIPE]; + + spinlock_t irqmask_lock; + + /* + * Power + */ + + bool suspended; + bool display_power; + int display_count; + + /* + * Modesetting + */ + struct psb_intel_mode_device mode_dev; + + struct drm_crtc *plane_to_crtc_mapping[PSB_NUM_PIPE]; + struct drm_crtc *pipe_to_crtc_mapping[PSB_NUM_PIPE]; + uint32_t num_pipe; + + /* + * OSPM info (Power management base) (can go ?) + */ + uint32_t ospm_base; + + /* + * Sizes info + */ + + u32 fuse_reg_value; + u32 video_device_fuse; + + /* PCI revision ID for B0:D2:F0 */ + uint8_t platform_rev_id; + + /* gmbus */ + struct intel_gmbus *gmbus; + + /* Used by SDVO */ + int crt_ddc_pin; + /* FIXME: The mappings should be parsed from bios but for now we can + pretend there are no mappings available */ + struct sdvo_device_mapping sdvo_mappings[2]; + u32 hotplug_supported_mask; + struct drm_property *broadcast_rgb_property; + struct drm_property *force_audio_property; + + /* + * LVDS info + */ + int backlight_duty_cycle; /* restore backlight to this value */ + bool panel_wants_dither; + struct drm_display_mode *panel_fixed_mode; + struct drm_display_mode *lfp_lvds_vbt_mode; + struct drm_display_mode *sdvo_lvds_vbt_mode; + + struct bdb_lvds_backlight *lvds_bl; /* LVDS backlight info from VBT */ + struct psb_intel_i2c_chan *lvds_i2c_bus; /* FIXME: Remove this? */ + + /* Feature bits from the VBIOS */ + unsigned int int_tv_support:1; + unsigned int lvds_dither:1; + unsigned int lvds_vbt:1; + unsigned int int_crt_support:1; + unsigned int lvds_use_ssc:1; + int lvds_ssc_freq; + bool is_lvds_on; + bool is_mipi_on; + u32 mipi_ctrl_display; + + unsigned int core_freq; + uint32_t iLVDS_enable; + + /* Runtime PM state */ + int rpm_enabled; + + /* MID specific */ + struct oaktrail_vbt vbt_data; + struct oaktrail_gct_data gct_data; + + /* Oaktrail HDMI state */ + struct oaktrail_hdmi_dev *hdmi_priv; + + /* + * Register state + */ + + struct psb_save_area regs; + + /* MSI reg save */ + uint32_t msi_addr; + uint32_t msi_data; + + + /* + * LID-Switch + */ + spinlock_t lid_lock; + struct timer_list lid_timer; + struct psb_intel_opregion opregion; + u32 *lid_state; + u32 lid_last_state; + + /* + * Watchdog + */ + + uint32_t apm_reg; + uint16_t apm_base; + + /* + * Used for modifying backlight from + * xrandr -- consider removing and using HAL instead + */ + struct backlight_device *backlight_device; + struct drm_property *backlight_property; + uint32_t blc_adj1; + uint32_t blc_adj2; + + void *fbdev; + + /* 2D acceleration */ + spinlock_t lock_2d; + + /* + * Panel brightness + */ + int brightness; + int brightness_adjusted; + + bool dsr_enable; + u32 dsr_fb_update; + bool dpi_panel_on[3]; + void *dsi_configs[2]; + u32 bpp; + u32 bpp2; + + u32 pipeconf[3]; + u32 dspcntr[3]; + + int mdfld_panel_id; +}; + + +/* + * Operations for each board type + */ + +struct psb_ops { + const char *name; + unsigned int accel_2d:1; + int pipes; /* Number of output pipes */ + int crtcs; /* Number of CRTCs */ + int sgx_offset; /* Base offset of SGX device */ + + /* Sub functions */ + struct drm_crtc_helper_funcs const *crtc_helper; + struct drm_crtc_funcs const *crtc_funcs; + + /* Setup hooks */ + int (*chip_setup)(struct drm_device *dev); + void (*chip_teardown)(struct drm_device *dev); + + /* Display management hooks */ + int (*output_init)(struct drm_device *dev); + /* Power management hooks */ + void (*init_pm)(struct drm_device *dev); + int (*save_regs)(struct drm_device *dev); + int (*restore_regs)(struct drm_device *dev); + int (*power_up)(struct drm_device *dev); + int (*power_down)(struct drm_device *dev); + + void (*lvds_bl_power)(struct drm_device *dev, bool on); +#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE + /* Backlight */ + int (*backlight_init)(struct drm_device *dev); +#endif + int i2c_bus; /* I2C bus identifier for Moorestown */ +}; + + + +struct psb_mmu_driver; + +extern int drm_crtc_probe_output_modes(struct drm_device *dev, int, int); +extern int drm_pick_crtcs(struct drm_device *dev); + +static inline struct drm_psb_private *psb_priv(struct drm_device *dev) +{ + return (struct drm_psb_private *) dev->dev_private; +} + +/* + * MMU stuff. + */ + +extern struct psb_mmu_driver *psb_mmu_driver_init(uint8_t __iomem * registers, + int trap_pagefaults, + int invalid_type, + struct drm_psb_private *dev_priv); +extern void psb_mmu_driver_takedown(struct psb_mmu_driver *driver); +extern struct psb_mmu_pd *psb_mmu_get_default_pd(struct psb_mmu_driver + *driver); +extern void psb_mmu_mirror_gtt(struct psb_mmu_pd *pd, uint32_t mmu_offset, + uint32_t gtt_start, uint32_t gtt_pages); +extern struct psb_mmu_pd *psb_mmu_alloc_pd(struct psb_mmu_driver *driver, + int trap_pagefaults, + int invalid_type); +extern void psb_mmu_free_pagedir(struct psb_mmu_pd *pd); +extern void psb_mmu_flush(struct psb_mmu_driver *driver, int rc_prot); +extern void psb_mmu_remove_pfn_sequence(struct psb_mmu_pd *pd, + unsigned long address, + uint32_t num_pages); +extern int psb_mmu_insert_pfn_sequence(struct psb_mmu_pd *pd, + uint32_t start_pfn, + unsigned long address, + uint32_t num_pages, int type); +extern int psb_mmu_virtual_to_pfn(struct psb_mmu_pd *pd, uint32_t virtual, + unsigned long *pfn); + +/* + * Enable / disable MMU for different requestors. + */ + + +extern void psb_mmu_set_pd_context(struct psb_mmu_pd *pd, int hw_context); +extern int psb_mmu_insert_pages(struct psb_mmu_pd *pd, struct page **pages, + unsigned long address, uint32_t num_pages, + uint32_t desired_tile_stride, + uint32_t hw_tile_stride, int type); +extern void psb_mmu_remove_pages(struct psb_mmu_pd *pd, + unsigned long address, uint32_t num_pages, + uint32_t desired_tile_stride, + uint32_t hw_tile_stride); +/* + *psb_irq.c + */ + +extern irqreturn_t psb_irq_handler(DRM_IRQ_ARGS); +extern int psb_irq_enable_dpst(struct drm_device *dev); +extern int psb_irq_disable_dpst(struct drm_device *dev); +extern void psb_irq_preinstall(struct drm_device *dev); +extern int psb_irq_postinstall(struct drm_device *dev); +extern void psb_irq_uninstall(struct drm_device *dev); +extern void psb_irq_turn_on_dpst(struct drm_device *dev); +extern void psb_irq_turn_off_dpst(struct drm_device *dev); + +extern void psb_irq_uninstall_islands(struct drm_device *dev, int hw_islands); +extern int psb_vblank_wait2(struct drm_device *dev, unsigned int *sequence); +extern int psb_vblank_wait(struct drm_device *dev, unsigned int *sequence); +extern int psb_enable_vblank(struct drm_device *dev, int crtc); +extern void psb_disable_vblank(struct drm_device *dev, int crtc); +void +psb_enable_pipestat(struct drm_psb_private *dev_priv, int pipe, u32 mask); + +void +psb_disable_pipestat(struct drm_psb_private *dev_priv, int pipe, u32 mask); + +extern u32 psb_get_vblank_counter(struct drm_device *dev, int crtc); + +/* + * intel_opregion.c + */ +extern int gma_intel_opregion_init(struct drm_device *dev); +extern int gma_intel_opregion_exit(struct drm_device *dev); + +/* + * framebuffer.c + */ +extern int psbfb_probed(struct drm_device *dev); +extern int psbfb_remove(struct drm_device *dev, + struct drm_framebuffer *fb); +/* + * accel_2d.c + */ +extern void psbfb_copyarea(struct fb_info *info, + const struct fb_copyarea *region); +extern int psbfb_sync(struct fb_info *info); +extern void psb_spank(struct drm_psb_private *dev_priv); + +/* + * psb_reset.c + */ + +extern void psb_lid_timer_init(struct drm_psb_private *dev_priv); +extern void psb_lid_timer_takedown(struct drm_psb_private *dev_priv); +extern void psb_print_pagefault(struct drm_psb_private *dev_priv); + +/* modesetting */ +extern void psb_modeset_init(struct drm_device *dev); +extern void psb_modeset_cleanup(struct drm_device *dev); +extern int psb_fbdev_init(struct drm_device *dev); + +/* backlight.c */ +int gma_backlight_init(struct drm_device *dev); +void gma_backlight_exit(struct drm_device *dev); + +/* oaktrail_crtc.c */ +extern const struct drm_crtc_helper_funcs oaktrail_helper_funcs; + +/* oaktrail_lvds.c */ +extern void oaktrail_lvds_init(struct drm_device *dev, + struct psb_intel_mode_device *mode_dev); + +/* psb_intel_display.c */ +extern const struct drm_crtc_helper_funcs psb_intel_helper_funcs; +extern const struct drm_crtc_funcs psb_intel_crtc_funcs; + +/* psb_intel_lvds.c */ +extern const struct drm_connector_helper_funcs + psb_intel_lvds_connector_helper_funcs; +extern const struct drm_connector_funcs psb_intel_lvds_connector_funcs; + +/* gem.c */ +extern int psb_gem_init_object(struct drm_gem_object *obj); +extern void psb_gem_free_object(struct drm_gem_object *obj); +extern int psb_gem_get_aperture(struct drm_device *dev, void *data, + struct drm_file *file); +extern int psb_gem_dumb_create(struct drm_file *file, struct drm_device *dev, + struct drm_mode_create_dumb *args); +extern int psb_gem_dumb_destroy(struct drm_file *file, struct drm_device *dev, + uint32_t handle); +extern int psb_gem_dumb_map_gtt(struct drm_file *file, struct drm_device *dev, + uint32_t handle, uint64_t *offset); +extern int psb_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf); +extern int psb_gem_create_ioctl(struct drm_device *dev, void *data, + struct drm_file *file); +extern int psb_gem_mmap_ioctl(struct drm_device *dev, void *data, + struct drm_file *file); + +/* psb_device.c */ +extern const struct psb_ops psb_chip_ops; + +/* oaktrail_device.c */ +extern const struct psb_ops oaktrail_chip_ops; + +/* mdlfd_device.c */ +extern const struct psb_ops mdfld_chip_ops; + +/* cdv_device.c */ +extern const struct psb_ops cdv_chip_ops; + +/* + * Debug print bits setting + */ +#define PSB_D_GENERAL (1 << 0) +#define PSB_D_INIT (1 << 1) +#define PSB_D_IRQ (1 << 2) +#define PSB_D_ENTRY (1 << 3) +/* debug the get H/V BP/FP count */ +#define PSB_D_HV (1 << 4) +#define PSB_D_DBI_BF (1 << 5) +#define PSB_D_PM (1 << 6) +#define PSB_D_RENDER (1 << 7) +#define PSB_D_REG (1 << 8) +#define PSB_D_MSVDX (1 << 9) +#define PSB_D_TOPAZ (1 << 10) + +extern int drm_psb_no_fb; +extern int drm_idle_check_interval; + +/* + * Utilities + */ + +static inline u32 MRST_MSG_READ32(uint port, uint offset) +{ + int mcr = (0xD0<<24) | (port << 16) | (offset << 8); + uint32_t ret_val = 0; + struct pci_dev *pci_root = pci_get_bus_and_slot(0, 0); + pci_write_config_dword(pci_root, 0xD0, mcr); + pci_read_config_dword(pci_root, 0xD4, &ret_val); + pci_dev_put(pci_root); + return ret_val; +} +static inline void MRST_MSG_WRITE32(uint port, uint offset, u32 value) +{ + int mcr = (0xE0<<24) | (port << 16) | (offset << 8) | 0xF0; + struct pci_dev *pci_root = pci_get_bus_and_slot(0, 0); + pci_write_config_dword(pci_root, 0xD4, value); + pci_write_config_dword(pci_root, 0xD0, mcr); + pci_dev_put(pci_root); +} +static inline u32 MDFLD_MSG_READ32(uint port, uint offset) +{ + int mcr = (0x10<<24) | (port << 16) | (offset << 8); + uint32_t ret_val = 0; + struct pci_dev *pci_root = pci_get_bus_and_slot(0, 0); + pci_write_config_dword(pci_root, 0xD0, mcr); + pci_read_config_dword(pci_root, 0xD4, &ret_val); + pci_dev_put(pci_root); + return ret_val; +} +static inline void MDFLD_MSG_WRITE32(uint port, uint offset, u32 value) +{ + int mcr = (0x11<<24) | (port << 16) | (offset << 8) | 0xF0; + struct pci_dev *pci_root = pci_get_bus_and_slot(0, 0); + pci_write_config_dword(pci_root, 0xD4, value); + pci_write_config_dword(pci_root, 0xD0, mcr); + pci_dev_put(pci_root); +} + +static inline uint32_t REGISTER_READ(struct drm_device *dev, uint32_t reg) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + return ioread32(dev_priv->vdc_reg + reg); +} + +#define REG_READ(reg) REGISTER_READ(dev, (reg)) + +static inline void REGISTER_WRITE(struct drm_device *dev, uint32_t reg, + uint32_t val) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + iowrite32((val), dev_priv->vdc_reg + (reg)); +} + +#define REG_WRITE(reg, val) REGISTER_WRITE(dev, (reg), (val)) + +static inline void REGISTER_WRITE16(struct drm_device *dev, + uint32_t reg, uint32_t val) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + iowrite16((val), dev_priv->vdc_reg + (reg)); +} + +#define REG_WRITE16(reg, val) REGISTER_WRITE16(dev, (reg), (val)) + +static inline void REGISTER_WRITE8(struct drm_device *dev, + uint32_t reg, uint32_t val) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + iowrite8((val), dev_priv->vdc_reg + (reg)); +} + +#define REG_WRITE8(reg, val) REGISTER_WRITE8(dev, (reg), (val)) + +#define PSB_WVDC32(_val, _offs) iowrite32(_val, dev_priv->vdc_reg + (_offs)) +#define PSB_RVDC32(_offs) ioread32(dev_priv->vdc_reg + (_offs)) + +/* #define TRAP_SGX_PM_FAULT 1 */ +#ifdef TRAP_SGX_PM_FAULT +#define PSB_RSGX32(_offs) \ +({ \ + if (inl(dev_priv->apm_base + PSB_APM_STS) & 0x3) { \ + printk(KERN_ERR \ + "access sgx when it's off!! (READ) %s, %d\n", \ + __FILE__, __LINE__); \ + melay(1000); \ + } \ + ioread32(dev_priv->sgx_reg + (_offs)); \ +}) +#else +#define PSB_RSGX32(_offs) ioread32(dev_priv->sgx_reg + (_offs)) +#endif +#define PSB_WSGX32(_val, _offs) iowrite32(_val, dev_priv->sgx_reg + (_offs)) + +#define MSVDX_REG_DUMP 0 + +#define PSB_WMSVDX32(_val, _offs) iowrite32(_val, dev_priv->msvdx_reg + (_offs)) +#define PSB_RMSVDX32(_offs) ioread32(dev_priv->msvdx_reg + (_offs)) + +#endif diff --git a/drivers/gpu/drm/gma500/psb_intel_display.c b/drivers/gpu/drm/gma500/psb_intel_display.c new file mode 100644 index 0000000..2616558 --- /dev/null +++ b/drivers/gpu/drm/gma500/psb_intel_display.c @@ -0,0 +1,1436 @@ +/* + * Copyright © 2006-2011 Intel Corporation + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + * Authors: + * Eric Anholt + */ + +#include +#include + +#include +#include "framebuffer.h" +#include "psb_drv.h" +#include "psb_intel_drv.h" +#include "psb_intel_reg.h" +#include "psb_intel_display.h" +#include "power.h" + +struct psb_intel_clock_t { + /* given values */ + int n; + int m1, m2; + int p1, p2; + /* derived values */ + int dot; + int vco; + int m; + int p; +}; + +struct psb_intel_range_t { + int min, max; +}; + +struct psb_intel_p2_t { + int dot_limit; + int p2_slow, p2_fast; +}; + +#define INTEL_P2_NUM 2 + +struct psb_intel_limit_t { + struct psb_intel_range_t dot, vco, n, m, m1, m2, p, p1; + struct psb_intel_p2_t p2; +}; + +#define I8XX_DOT_MIN 25000 +#define I8XX_DOT_MAX 350000 +#define I8XX_VCO_MIN 930000 +#define I8XX_VCO_MAX 1400000 +#define I8XX_N_MIN 3 +#define I8XX_N_MAX 16 +#define I8XX_M_MIN 96 +#define I8XX_M_MAX 140 +#define I8XX_M1_MIN 18 +#define I8XX_M1_MAX 26 +#define I8XX_M2_MIN 6 +#define I8XX_M2_MAX 16 +#define I8XX_P_MIN 4 +#define I8XX_P_MAX 128 +#define I8XX_P1_MIN 2 +#define I8XX_P1_MAX 33 +#define I8XX_P1_LVDS_MIN 1 +#define I8XX_P1_LVDS_MAX 6 +#define I8XX_P2_SLOW 4 +#define I8XX_P2_FAST 2 +#define I8XX_P2_LVDS_SLOW 14 +#define I8XX_P2_LVDS_FAST 14 /* No fast option */ +#define I8XX_P2_SLOW_LIMIT 165000 + +#define I9XX_DOT_MIN 20000 +#define I9XX_DOT_MAX 400000 +#define I9XX_VCO_MIN 1400000 +#define I9XX_VCO_MAX 2800000 +#define I9XX_N_MIN 3 +#define I9XX_N_MAX 8 +#define I9XX_M_MIN 70 +#define I9XX_M_MAX 120 +#define I9XX_M1_MIN 10 +#define I9XX_M1_MAX 20 +#define I9XX_M2_MIN 5 +#define I9XX_M2_MAX 9 +#define I9XX_P_SDVO_DAC_MIN 5 +#define I9XX_P_SDVO_DAC_MAX 80 +#define I9XX_P_LVDS_MIN 7 +#define I9XX_P_LVDS_MAX 98 +#define I9XX_P1_MIN 1 +#define I9XX_P1_MAX 8 +#define I9XX_P2_SDVO_DAC_SLOW 10 +#define I9XX_P2_SDVO_DAC_FAST 5 +#define I9XX_P2_SDVO_DAC_SLOW_LIMIT 200000 +#define I9XX_P2_LVDS_SLOW 14 +#define I9XX_P2_LVDS_FAST 7 +#define I9XX_P2_LVDS_SLOW_LIMIT 112000 + +#define INTEL_LIMIT_I8XX_DVO_DAC 0 +#define INTEL_LIMIT_I8XX_LVDS 1 +#define INTEL_LIMIT_I9XX_SDVO_DAC 2 +#define INTEL_LIMIT_I9XX_LVDS 3 + +static const struct psb_intel_limit_t psb_intel_limits[] = { + { /* INTEL_LIMIT_I8XX_DVO_DAC */ + .dot = {.min = I8XX_DOT_MIN, .max = I8XX_DOT_MAX}, + .vco = {.min = I8XX_VCO_MIN, .max = I8XX_VCO_MAX}, + .n = {.min = I8XX_N_MIN, .max = I8XX_N_MAX}, + .m = {.min = I8XX_M_MIN, .max = I8XX_M_MAX}, + .m1 = {.min = I8XX_M1_MIN, .max = I8XX_M1_MAX}, + .m2 = {.min = I8XX_M2_MIN, .max = I8XX_M2_MAX}, + .p = {.min = I8XX_P_MIN, .max = I8XX_P_MAX}, + .p1 = {.min = I8XX_P1_MIN, .max = I8XX_P1_MAX}, + .p2 = {.dot_limit = I8XX_P2_SLOW_LIMIT, + .p2_slow = I8XX_P2_SLOW, .p2_fast = I8XX_P2_FAST}, + }, + { /* INTEL_LIMIT_I8XX_LVDS */ + .dot = {.min = I8XX_DOT_MIN, .max = I8XX_DOT_MAX}, + .vco = {.min = I8XX_VCO_MIN, .max = I8XX_VCO_MAX}, + .n = {.min = I8XX_N_MIN, .max = I8XX_N_MAX}, + .m = {.min = I8XX_M_MIN, .max = I8XX_M_MAX}, + .m1 = {.min = I8XX_M1_MIN, .max = I8XX_M1_MAX}, + .m2 = {.min = I8XX_M2_MIN, .max = I8XX_M2_MAX}, + .p = {.min = I8XX_P_MIN, .max = I8XX_P_MAX}, + .p1 = {.min = I8XX_P1_LVDS_MIN, .max = I8XX_P1_LVDS_MAX}, + .p2 = {.dot_limit = I8XX_P2_SLOW_LIMIT, + .p2_slow = I8XX_P2_LVDS_SLOW, .p2_fast = I8XX_P2_LVDS_FAST}, + }, + { /* INTEL_LIMIT_I9XX_SDVO_DAC */ + .dot = {.min = I9XX_DOT_MIN, .max = I9XX_DOT_MAX}, + .vco = {.min = I9XX_VCO_MIN, .max = I9XX_VCO_MAX}, + .n = {.min = I9XX_N_MIN, .max = I9XX_N_MAX}, + .m = {.min = I9XX_M_MIN, .max = I9XX_M_MAX}, + .m1 = {.min = I9XX_M1_MIN, .max = I9XX_M1_MAX}, + .m2 = {.min = I9XX_M2_MIN, .max = I9XX_M2_MAX}, + .p = {.min = I9XX_P_SDVO_DAC_MIN, .max = I9XX_P_SDVO_DAC_MAX}, + .p1 = {.min = I9XX_P1_MIN, .max = I9XX_P1_MAX}, + .p2 = {.dot_limit = I9XX_P2_SDVO_DAC_SLOW_LIMIT, + .p2_slow = I9XX_P2_SDVO_DAC_SLOW, .p2_fast = + I9XX_P2_SDVO_DAC_FAST}, + }, + { /* INTEL_LIMIT_I9XX_LVDS */ + .dot = {.min = I9XX_DOT_MIN, .max = I9XX_DOT_MAX}, + .vco = {.min = I9XX_VCO_MIN, .max = I9XX_VCO_MAX}, + .n = {.min = I9XX_N_MIN, .max = I9XX_N_MAX}, + .m = {.min = I9XX_M_MIN, .max = I9XX_M_MAX}, + .m1 = {.min = I9XX_M1_MIN, .max = I9XX_M1_MAX}, + .m2 = {.min = I9XX_M2_MIN, .max = I9XX_M2_MAX}, + .p = {.min = I9XX_P_LVDS_MIN, .max = I9XX_P_LVDS_MAX}, + .p1 = {.min = I9XX_P1_MIN, .max = I9XX_P1_MAX}, + /* The single-channel range is 25-112Mhz, and dual-channel + * is 80-224Mhz. Prefer single channel as much as possible. + */ + .p2 = {.dot_limit = I9XX_P2_LVDS_SLOW_LIMIT, + .p2_slow = I9XX_P2_LVDS_SLOW, .p2_fast = I9XX_P2_LVDS_FAST}, + }, +}; + +static const struct psb_intel_limit_t *psb_intel_limit(struct drm_crtc *crtc) +{ + const struct psb_intel_limit_t *limit; + + if (psb_intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) + limit = &psb_intel_limits[INTEL_LIMIT_I9XX_LVDS]; + else + limit = &psb_intel_limits[INTEL_LIMIT_I9XX_SDVO_DAC]; + return limit; +} + +/** Derive the pixel clock for the given refclk and divisors for 8xx chips. */ + +static void i8xx_clock(int refclk, struct psb_intel_clock_t *clock) +{ + clock->m = 5 * (clock->m1 + 2) + (clock->m2 + 2); + clock->p = clock->p1 * clock->p2; + clock->vco = refclk * clock->m / (clock->n + 2); + clock->dot = clock->vco / clock->p; +} + +/** Derive the pixel clock for the given refclk and divisors for 9xx chips. */ + +static void i9xx_clock(int refclk, struct psb_intel_clock_t *clock) +{ + clock->m = 5 * (clock->m1 + 2) + (clock->m2 + 2); + clock->p = clock->p1 * clock->p2; + clock->vco = refclk * clock->m / (clock->n + 2); + clock->dot = clock->vco / clock->p; +} + +static void psb_intel_clock(struct drm_device *dev, int refclk, + struct psb_intel_clock_t *clock) +{ + return i9xx_clock(refclk, clock); +} + +/** + * Returns whether any output on the specified pipe is of the specified type + */ +bool psb_intel_pipe_has_type(struct drm_crtc *crtc, int type) +{ + struct drm_device *dev = crtc->dev; + struct drm_mode_config *mode_config = &dev->mode_config; + struct drm_connector *l_entry; + + list_for_each_entry(l_entry, &mode_config->connector_list, head) { + if (l_entry->encoder && l_entry->encoder->crtc == crtc) { + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(l_entry); + if (psb_intel_encoder->type == type) + return true; + } + } + return false; +} + +#define INTELPllInvalid(s) { /* ErrorF (s) */; return false; } +/** + * Returns whether the given set of divisors are valid for a given refclk with + * the given connectors. + */ + +static bool psb_intel_PLL_is_valid(struct drm_crtc *crtc, + struct psb_intel_clock_t *clock) +{ + const struct psb_intel_limit_t *limit = psb_intel_limit(crtc); + + if (clock->p1 < limit->p1.min || limit->p1.max < clock->p1) + INTELPllInvalid("p1 out of range\n"); + if (clock->p < limit->p.min || limit->p.max < clock->p) + INTELPllInvalid("p out of range\n"); + if (clock->m2 < limit->m2.min || limit->m2.max < clock->m2) + INTELPllInvalid("m2 out of range\n"); + if (clock->m1 < limit->m1.min || limit->m1.max < clock->m1) + INTELPllInvalid("m1 out of range\n"); + if (clock->m1 <= clock->m2) + INTELPllInvalid("m1 <= m2\n"); + if (clock->m < limit->m.min || limit->m.max < clock->m) + INTELPllInvalid("m out of range\n"); + if (clock->n < limit->n.min || limit->n.max < clock->n) + INTELPllInvalid("n out of range\n"); + if (clock->vco < limit->vco.min || limit->vco.max < clock->vco) + INTELPllInvalid("vco out of range\n"); + /* XXX: We may need to be checking "Dot clock" + * depending on the multiplier, connector, etc., + * rather than just a single range. + */ + if (clock->dot < limit->dot.min || limit->dot.max < clock->dot) + INTELPllInvalid("dot out of range\n"); + + return true; +} + +/** + * Returns a set of divisors for the desired target clock with the given + * refclk, or FALSE. The returned values represent the clock equation: + * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2. + */ +static bool psb_intel_find_best_PLL(struct drm_crtc *crtc, int target, + int refclk, + struct psb_intel_clock_t *best_clock) +{ + struct drm_device *dev = crtc->dev; + struct psb_intel_clock_t clock; + const struct psb_intel_limit_t *limit = psb_intel_limit(crtc); + int err = target; + + if (psb_intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) && + (REG_READ(LVDS) & LVDS_PORT_EN) != 0) { + /* + * For LVDS, if the panel is on, just rely on its current + * settings for dual-channel. We haven't figured out how to + * reliably set up different single/dual channel state, if we + * even can. + */ + if ((REG_READ(LVDS) & LVDS_CLKB_POWER_MASK) == + LVDS_CLKB_POWER_UP) + clock.p2 = limit->p2.p2_fast; + else + clock.p2 = limit->p2.p2_slow; + } else { + if (target < limit->p2.dot_limit) + clock.p2 = limit->p2.p2_slow; + else + clock.p2 = limit->p2.p2_fast; + } + + memset(best_clock, 0, sizeof(*best_clock)); + + for (clock.m1 = limit->m1.min; clock.m1 <= limit->m1.max; + clock.m1++) { + for (clock.m2 = limit->m2.min; + clock.m2 < clock.m1 && clock.m2 <= limit->m2.max; + clock.m2++) { + for (clock.n = limit->n.min; + clock.n <= limit->n.max; clock.n++) { + for (clock.p1 = limit->p1.min; + clock.p1 <= limit->p1.max; + clock.p1++) { + int this_err; + + psb_intel_clock(dev, refclk, &clock); + + if (!psb_intel_PLL_is_valid + (crtc, &clock)) + continue; + + this_err = abs(clock.dot - target); + if (this_err < err) { + *best_clock = clock; + err = this_err; + } + } + } + } + } + + return err != target; +} + +void psb_intel_wait_for_vblank(struct drm_device *dev) +{ + /* Wait for 20ms, i.e. one cycle at 50hz. */ + mdelay(20); +} + +static int psb_intel_pipe_set_base(struct drm_crtc *crtc, + int x, int y, struct drm_framebuffer *old_fb) +{ + struct drm_device *dev = crtc->dev; + /* struct drm_i915_master_private *master_priv; */ + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + struct psb_framebuffer *psbfb = to_psb_fb(crtc->fb); + int pipe = psb_intel_crtc->pipe; + unsigned long start, offset; + int dspbase = (pipe == 0 ? DSPABASE : DSPBBASE); + int dspsurf = (pipe == 0 ? DSPASURF : DSPBSURF); + int dspstride = (pipe == 0) ? DSPASTRIDE : DSPBSTRIDE; + int dspcntr_reg = (pipe == 0) ? DSPACNTR : DSPBCNTR; + u32 dspcntr; + int ret = 0; + + if (!gma_power_begin(dev, true)) + return 0; + + /* no fb bound */ + if (!crtc->fb) { + dev_dbg(dev->dev, "No FB bound\n"); + goto psb_intel_pipe_cleaner; + } + + /* We are displaying this buffer, make sure it is actually loaded + into the GTT */ + ret = psb_gtt_pin(psbfb->gtt); + if (ret < 0) + goto psb_intel_pipe_set_base_exit; + start = psbfb->gtt->offset; + + offset = y * crtc->fb->pitches[0] + x * (crtc->fb->bits_per_pixel / 8); + + REG_WRITE(dspstride, crtc->fb->pitches[0]); + + dspcntr = REG_READ(dspcntr_reg); + dspcntr &= ~DISPPLANE_PIXFORMAT_MASK; + + switch (crtc->fb->bits_per_pixel) { + case 8: + dspcntr |= DISPPLANE_8BPP; + break; + case 16: + if (crtc->fb->depth == 15) + dspcntr |= DISPPLANE_15_16BPP; + else + dspcntr |= DISPPLANE_16BPP; + break; + case 24: + case 32: + dspcntr |= DISPPLANE_32BPP_NO_ALPHA; + break; + default: + dev_err(dev->dev, "Unknown color depth\n"); + ret = -EINVAL; + psb_gtt_unpin(psbfb->gtt); + goto psb_intel_pipe_set_base_exit; + } + REG_WRITE(dspcntr_reg, dspcntr); + + + if (0 /* FIXMEAC - check what PSB needs */) { + REG_WRITE(dspbase, offset); + REG_READ(dspbase); + REG_WRITE(dspsurf, start); + REG_READ(dspsurf); + } else { + REG_WRITE(dspbase, start + offset); + REG_READ(dspbase); + } + +psb_intel_pipe_cleaner: + /* If there was a previous display we can now unpin it */ + if (old_fb) + psb_gtt_unpin(to_psb_fb(old_fb)->gtt); + +psb_intel_pipe_set_base_exit: + gma_power_end(dev); + return ret; +} + +/** + * Sets the power management mode of the pipe and plane. + * + * This code should probably grow support for turning the cursor off and back + * on appropriately at the same time as we're turning the pipe off/on. + */ +static void psb_intel_crtc_dpms(struct drm_crtc *crtc, int mode) +{ + struct drm_device *dev = crtc->dev; + /* struct drm_i915_master_private *master_priv; */ + /* struct drm_i915_private *dev_priv = dev->dev_private; */ + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + int pipe = psb_intel_crtc->pipe; + int dpll_reg = (pipe == 0) ? DPLL_A : DPLL_B; + int dspcntr_reg = (pipe == 0) ? DSPACNTR : DSPBCNTR; + int dspbase_reg = (pipe == 0) ? DSPABASE : DSPBBASE; + int pipeconf_reg = (pipe == 0) ? PIPEACONF : PIPEBCONF; + u32 temp; + + /* XXX: When our outputs are all unaware of DPMS modes other than off + * and on, we should map those modes to DRM_MODE_DPMS_OFF in the CRTC. + */ + switch (mode) { + case DRM_MODE_DPMS_ON: + case DRM_MODE_DPMS_STANDBY: + case DRM_MODE_DPMS_SUSPEND: + /* Enable the DPLL */ + temp = REG_READ(dpll_reg); + if ((temp & DPLL_VCO_ENABLE) == 0) { + REG_WRITE(dpll_reg, temp); + REG_READ(dpll_reg); + /* Wait for the clocks to stabilize. */ + udelay(150); + REG_WRITE(dpll_reg, temp | DPLL_VCO_ENABLE); + REG_READ(dpll_reg); + /* Wait for the clocks to stabilize. */ + udelay(150); + REG_WRITE(dpll_reg, temp | DPLL_VCO_ENABLE); + REG_READ(dpll_reg); + /* Wait for the clocks to stabilize. */ + udelay(150); + } + + /* Enable the pipe */ + temp = REG_READ(pipeconf_reg); + if ((temp & PIPEACONF_ENABLE) == 0) + REG_WRITE(pipeconf_reg, temp | PIPEACONF_ENABLE); + + /* Enable the plane */ + temp = REG_READ(dspcntr_reg); + if ((temp & DISPLAY_PLANE_ENABLE) == 0) { + REG_WRITE(dspcntr_reg, + temp | DISPLAY_PLANE_ENABLE); + /* Flush the plane changes */ + REG_WRITE(dspbase_reg, REG_READ(dspbase_reg)); + } + + psb_intel_crtc_load_lut(crtc); + + /* Give the overlay scaler a chance to enable + * if it's on this pipe */ + /* psb_intel_crtc_dpms_video(crtc, true); TODO */ + break; + case DRM_MODE_DPMS_OFF: + /* Give the overlay scaler a chance to disable + * if it's on this pipe */ + /* psb_intel_crtc_dpms_video(crtc, FALSE); TODO */ + + /* Disable the VGA plane that we never use */ + REG_WRITE(VGACNTRL, VGA_DISP_DISABLE); + + /* Disable display plane */ + temp = REG_READ(dspcntr_reg); + if ((temp & DISPLAY_PLANE_ENABLE) != 0) { + REG_WRITE(dspcntr_reg, + temp & ~DISPLAY_PLANE_ENABLE); + /* Flush the plane changes */ + REG_WRITE(dspbase_reg, REG_READ(dspbase_reg)); + REG_READ(dspbase_reg); + } + + /* Next, disable display pipes */ + temp = REG_READ(pipeconf_reg); + if ((temp & PIPEACONF_ENABLE) != 0) { + REG_WRITE(pipeconf_reg, temp & ~PIPEACONF_ENABLE); + REG_READ(pipeconf_reg); + } + + /* Wait for vblank for the disable to take effect. */ + psb_intel_wait_for_vblank(dev); + + temp = REG_READ(dpll_reg); + if ((temp & DPLL_VCO_ENABLE) != 0) { + REG_WRITE(dpll_reg, temp & ~DPLL_VCO_ENABLE); + REG_READ(dpll_reg); + } + + /* Wait for the clocks to turn off. */ + udelay(150); + break; + } + + /*Set FIFO Watermarks*/ + REG_WRITE(DSPARB, 0x3F3E); +} + +static void psb_intel_crtc_prepare(struct drm_crtc *crtc) +{ + struct drm_crtc_helper_funcs *crtc_funcs = crtc->helper_private; + crtc_funcs->dpms(crtc, DRM_MODE_DPMS_OFF); +} + +static void psb_intel_crtc_commit(struct drm_crtc *crtc) +{ + struct drm_crtc_helper_funcs *crtc_funcs = crtc->helper_private; + crtc_funcs->dpms(crtc, DRM_MODE_DPMS_ON); +} + +void psb_intel_encoder_prepare(struct drm_encoder *encoder) +{ + struct drm_encoder_helper_funcs *encoder_funcs = + encoder->helper_private; + /* lvds has its own version of prepare see psb_intel_lvds_prepare */ + encoder_funcs->dpms(encoder, DRM_MODE_DPMS_OFF); +} + +void psb_intel_encoder_commit(struct drm_encoder *encoder) +{ + struct drm_encoder_helper_funcs *encoder_funcs = + encoder->helper_private; + /* lvds has its own version of commit see psb_intel_lvds_commit */ + encoder_funcs->dpms(encoder, DRM_MODE_DPMS_ON); +} + +void psb_intel_encoder_destroy(struct drm_encoder *encoder) +{ + struct psb_intel_encoder *intel_encoder = to_psb_intel_encoder(encoder); + + drm_encoder_cleanup(encoder); + kfree(intel_encoder); +} + +static bool psb_intel_crtc_mode_fixup(struct drm_crtc *crtc, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + return true; +} + + +/** + * Return the pipe currently connected to the panel fitter, + * or -1 if the panel fitter is not present or not in use + */ +static int psb_intel_panel_fitter_pipe(struct drm_device *dev) +{ + u32 pfit_control; + + pfit_control = REG_READ(PFIT_CONTROL); + + /* See if the panel fitter is in use */ + if ((pfit_control & PFIT_ENABLE) == 0) + return -1; + /* Must be on PIPE 1 for PSB */ + return 1; +} + +static int psb_intel_crtc_mode_set(struct drm_crtc *crtc, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode, + int x, int y, + struct drm_framebuffer *old_fb) +{ + struct drm_device *dev = crtc->dev; + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + struct drm_crtc_helper_funcs *crtc_funcs = crtc->helper_private; + int pipe = psb_intel_crtc->pipe; + int fp_reg = (pipe == 0) ? FPA0 : FPB0; + int dpll_reg = (pipe == 0) ? DPLL_A : DPLL_B; + int dspcntr_reg = (pipe == 0) ? DSPACNTR : DSPBCNTR; + int pipeconf_reg = (pipe == 0) ? PIPEACONF : PIPEBCONF; + int htot_reg = (pipe == 0) ? HTOTAL_A : HTOTAL_B; + int hblank_reg = (pipe == 0) ? HBLANK_A : HBLANK_B; + int hsync_reg = (pipe == 0) ? HSYNC_A : HSYNC_B; + int vtot_reg = (pipe == 0) ? VTOTAL_A : VTOTAL_B; + int vblank_reg = (pipe == 0) ? VBLANK_A : VBLANK_B; + int vsync_reg = (pipe == 0) ? VSYNC_A : VSYNC_B; + int dspsize_reg = (pipe == 0) ? DSPASIZE : DSPBSIZE; + int dsppos_reg = (pipe == 0) ? DSPAPOS : DSPBPOS; + int pipesrc_reg = (pipe == 0) ? PIPEASRC : PIPEBSRC; + int refclk; + struct psb_intel_clock_t clock; + u32 dpll = 0, fp = 0, dspcntr, pipeconf; + bool ok, is_sdvo = false; + bool is_lvds = false, is_tv = false; + struct drm_mode_config *mode_config = &dev->mode_config; + struct drm_connector *connector; + + /* No scan out no play */ + if (crtc->fb == NULL) { + crtc_funcs->mode_set_base(crtc, x, y, old_fb); + return 0; + } + + list_for_each_entry(connector, &mode_config->connector_list, head) { + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(connector); + + if (!connector->encoder + || connector->encoder->crtc != crtc) + continue; + + switch (psb_intel_encoder->type) { + case INTEL_OUTPUT_LVDS: + is_lvds = true; + break; + case INTEL_OUTPUT_SDVO: + is_sdvo = true; + break; + case INTEL_OUTPUT_TVOUT: + is_tv = true; + break; + } + } + + refclk = 96000; + + ok = psb_intel_find_best_PLL(crtc, adjusted_mode->clock, refclk, + &clock); + if (!ok) { + dev_err(dev->dev, "Couldn't find PLL settings for mode!\n"); + return 0; + } + + fp = clock.n << 16 | clock.m1 << 8 | clock.m2; + + dpll = DPLL_VGA_MODE_DIS; + if (is_lvds) { + dpll |= DPLLB_MODE_LVDS; + dpll |= DPLL_DVO_HIGH_SPEED; + } else + dpll |= DPLLB_MODE_DAC_SERIAL; + if (is_sdvo) { + int sdvo_pixel_multiply = + adjusted_mode->clock / mode->clock; + dpll |= DPLL_DVO_HIGH_SPEED; + dpll |= + (sdvo_pixel_multiply - 1) << SDVO_MULTIPLIER_SHIFT_HIRES; + } + + /* compute bitmask from p1 value */ + dpll |= (1 << (clock.p1 - 1)) << 16; + switch (clock.p2) { + case 5: + dpll |= DPLL_DAC_SERIAL_P2_CLOCK_DIV_5; + break; + case 7: + dpll |= DPLLB_LVDS_P2_CLOCK_DIV_7; + break; + case 10: + dpll |= DPLL_DAC_SERIAL_P2_CLOCK_DIV_10; + break; + case 14: + dpll |= DPLLB_LVDS_P2_CLOCK_DIV_14; + break; + } + + if (is_tv) { + /* XXX: just matching BIOS for now */ +/* dpll |= PLL_REF_INPUT_TVCLKINBC; */ + dpll |= 3; + } + dpll |= PLL_REF_INPUT_DREFCLK; + + /* setup pipeconf */ + pipeconf = REG_READ(pipeconf_reg); + + /* Set up the display plane register */ + dspcntr = DISPPLANE_GAMMA_ENABLE; + + if (pipe == 0) + dspcntr |= DISPPLANE_SEL_PIPE_A; + else + dspcntr |= DISPPLANE_SEL_PIPE_B; + + dspcntr |= DISPLAY_PLANE_ENABLE; + pipeconf |= PIPEACONF_ENABLE; + dpll |= DPLL_VCO_ENABLE; + + + /* Disable the panel fitter if it was on our pipe */ + if (psb_intel_panel_fitter_pipe(dev) == pipe) + REG_WRITE(PFIT_CONTROL, 0); + + drm_mode_debug_printmodeline(mode); + + if (dpll & DPLL_VCO_ENABLE) { + REG_WRITE(fp_reg, fp); + REG_WRITE(dpll_reg, dpll & ~DPLL_VCO_ENABLE); + REG_READ(dpll_reg); + udelay(150); + } + + /* The LVDS pin pair needs to be on before the DPLLs are enabled. + * This is an exception to the general rule that mode_set doesn't turn + * things on. + */ + if (is_lvds) { + u32 lvds = REG_READ(LVDS); + + lvds &= ~LVDS_PIPEB_SELECT; + if (pipe == 1) + lvds |= LVDS_PIPEB_SELECT; + + lvds |= LVDS_PORT_EN | LVDS_A0A2_CLKA_POWER_UP; + /* Set the B0-B3 data pairs corresponding to + * whether we're going to + * set the DPLLs for dual-channel mode or not. + */ + lvds &= ~(LVDS_B0B3_POWER_UP | LVDS_CLKB_POWER_UP); + if (clock.p2 == 7) + lvds |= LVDS_B0B3_POWER_UP | LVDS_CLKB_POWER_UP; + + /* It would be nice to set 24 vs 18-bit mode (LVDS_A3_POWER_UP) + * appropriately here, but we need to look more + * thoroughly into how panels behave in the two modes. + */ + + REG_WRITE(LVDS, lvds); + REG_READ(LVDS); + } + + REG_WRITE(fp_reg, fp); + REG_WRITE(dpll_reg, dpll); + REG_READ(dpll_reg); + /* Wait for the clocks to stabilize. */ + udelay(150); + + /* write it again -- the BIOS does, after all */ + REG_WRITE(dpll_reg, dpll); + + REG_READ(dpll_reg); + /* Wait for the clocks to stabilize. */ + udelay(150); + + REG_WRITE(htot_reg, (adjusted_mode->crtc_hdisplay - 1) | + ((adjusted_mode->crtc_htotal - 1) << 16)); + REG_WRITE(hblank_reg, (adjusted_mode->crtc_hblank_start - 1) | + ((adjusted_mode->crtc_hblank_end - 1) << 16)); + REG_WRITE(hsync_reg, (adjusted_mode->crtc_hsync_start - 1) | + ((adjusted_mode->crtc_hsync_end - 1) << 16)); + REG_WRITE(vtot_reg, (adjusted_mode->crtc_vdisplay - 1) | + ((adjusted_mode->crtc_vtotal - 1) << 16)); + REG_WRITE(vblank_reg, (adjusted_mode->crtc_vblank_start - 1) | + ((adjusted_mode->crtc_vblank_end - 1) << 16)); + REG_WRITE(vsync_reg, (adjusted_mode->crtc_vsync_start - 1) | + ((adjusted_mode->crtc_vsync_end - 1) << 16)); + /* pipesrc and dspsize control the size that is scaled from, + * which should always be the user's requested size. + */ + REG_WRITE(dspsize_reg, + ((mode->vdisplay - 1) << 16) | (mode->hdisplay - 1)); + REG_WRITE(dsppos_reg, 0); + REG_WRITE(pipesrc_reg, + ((mode->hdisplay - 1) << 16) | (mode->vdisplay - 1)); + REG_WRITE(pipeconf_reg, pipeconf); + REG_READ(pipeconf_reg); + + psb_intel_wait_for_vblank(dev); + + REG_WRITE(dspcntr_reg, dspcntr); + + /* Flush the plane changes */ + crtc_funcs->mode_set_base(crtc, x, y, old_fb); + + psb_intel_wait_for_vblank(dev); + + return 0; +} + +/** Loads the palette/gamma unit for the CRTC with the prepared values */ +void psb_intel_crtc_load_lut(struct drm_crtc *crtc) +{ + struct drm_device *dev = crtc->dev; + struct drm_psb_private *dev_priv = + (struct drm_psb_private *)dev->dev_private; + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + int palreg = PALETTE_A; + int i; + + /* The clocks have to be on to load the palette. */ + if (!crtc->enabled) + return; + + switch (psb_intel_crtc->pipe) { + case 0: + break; + case 1: + palreg = PALETTE_B; + break; + case 2: + palreg = PALETTE_C; + break; + default: + dev_err(dev->dev, "Illegal Pipe Number.\n"); + return; + } + + if (gma_power_begin(dev, false)) { + for (i = 0; i < 256; i++) { + REG_WRITE(palreg + 4 * i, + ((psb_intel_crtc->lut_r[i] + + psb_intel_crtc->lut_adj[i]) << 16) | + ((psb_intel_crtc->lut_g[i] + + psb_intel_crtc->lut_adj[i]) << 8) | + (psb_intel_crtc->lut_b[i] + + psb_intel_crtc->lut_adj[i])); + } + gma_power_end(dev); + } else { + for (i = 0; i < 256; i++) { + dev_priv->regs.psb.save_palette_a[i] = + ((psb_intel_crtc->lut_r[i] + + psb_intel_crtc->lut_adj[i]) << 16) | + ((psb_intel_crtc->lut_g[i] + + psb_intel_crtc->lut_adj[i]) << 8) | + (psb_intel_crtc->lut_b[i] + + psb_intel_crtc->lut_adj[i]); + } + + } +} + +/** + * Save HW states of giving crtc + */ +static void psb_intel_crtc_save(struct drm_crtc *crtc) +{ + struct drm_device *dev = crtc->dev; + /* struct drm_psb_private *dev_priv = + (struct drm_psb_private *)dev->dev_private; */ + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + struct psb_intel_crtc_state *crtc_state = psb_intel_crtc->crtc_state; + int pipeA = (psb_intel_crtc->pipe == 0); + uint32_t paletteReg; + int i; + + if (!crtc_state) { + dev_err(dev->dev, "No CRTC state found\n"); + return; + } + + crtc_state->saveDSPCNTR = REG_READ(pipeA ? DSPACNTR : DSPBCNTR); + crtc_state->savePIPECONF = REG_READ(pipeA ? PIPEACONF : PIPEBCONF); + crtc_state->savePIPESRC = REG_READ(pipeA ? PIPEASRC : PIPEBSRC); + crtc_state->saveFP0 = REG_READ(pipeA ? FPA0 : FPB0); + crtc_state->saveFP1 = REG_READ(pipeA ? FPA1 : FPB1); + crtc_state->saveDPLL = REG_READ(pipeA ? DPLL_A : DPLL_B); + crtc_state->saveHTOTAL = REG_READ(pipeA ? HTOTAL_A : HTOTAL_B); + crtc_state->saveHBLANK = REG_READ(pipeA ? HBLANK_A : HBLANK_B); + crtc_state->saveHSYNC = REG_READ(pipeA ? HSYNC_A : HSYNC_B); + crtc_state->saveVTOTAL = REG_READ(pipeA ? VTOTAL_A : VTOTAL_B); + crtc_state->saveVBLANK = REG_READ(pipeA ? VBLANK_A : VBLANK_B); + crtc_state->saveVSYNC = REG_READ(pipeA ? VSYNC_A : VSYNC_B); + crtc_state->saveDSPSTRIDE = REG_READ(pipeA ? DSPASTRIDE : DSPBSTRIDE); + + /*NOTE: DSPSIZE DSPPOS only for psb*/ + crtc_state->saveDSPSIZE = REG_READ(pipeA ? DSPASIZE : DSPBSIZE); + crtc_state->saveDSPPOS = REG_READ(pipeA ? DSPAPOS : DSPBPOS); + + crtc_state->saveDSPBASE = REG_READ(pipeA ? DSPABASE : DSPBBASE); + + paletteReg = pipeA ? PALETTE_A : PALETTE_B; + for (i = 0; i < 256; ++i) + crtc_state->savePalette[i] = REG_READ(paletteReg + (i << 2)); +} + +/** + * Restore HW states of giving crtc + */ +static void psb_intel_crtc_restore(struct drm_crtc *crtc) +{ + struct drm_device *dev = crtc->dev; + /* struct drm_psb_private * dev_priv = + (struct drm_psb_private *)dev->dev_private; */ + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + struct psb_intel_crtc_state *crtc_state = psb_intel_crtc->crtc_state; + /* struct drm_crtc_helper_funcs * crtc_funcs = crtc->helper_private; */ + int pipeA = (psb_intel_crtc->pipe == 0); + uint32_t paletteReg; + int i; + + if (!crtc_state) { + dev_err(dev->dev, "No crtc state\n"); + return; + } + + if (crtc_state->saveDPLL & DPLL_VCO_ENABLE) { + REG_WRITE(pipeA ? DPLL_A : DPLL_B, + crtc_state->saveDPLL & ~DPLL_VCO_ENABLE); + REG_READ(pipeA ? DPLL_A : DPLL_B); + udelay(150); + } + + REG_WRITE(pipeA ? FPA0 : FPB0, crtc_state->saveFP0); + REG_READ(pipeA ? FPA0 : FPB0); + + REG_WRITE(pipeA ? FPA1 : FPB1, crtc_state->saveFP1); + REG_READ(pipeA ? FPA1 : FPB1); + + REG_WRITE(pipeA ? DPLL_A : DPLL_B, crtc_state->saveDPLL); + REG_READ(pipeA ? DPLL_A : DPLL_B); + udelay(150); + + REG_WRITE(pipeA ? HTOTAL_A : HTOTAL_B, crtc_state->saveHTOTAL); + REG_WRITE(pipeA ? HBLANK_A : HBLANK_B, crtc_state->saveHBLANK); + REG_WRITE(pipeA ? HSYNC_A : HSYNC_B, crtc_state->saveHSYNC); + REG_WRITE(pipeA ? VTOTAL_A : VTOTAL_B, crtc_state->saveVTOTAL); + REG_WRITE(pipeA ? VBLANK_A : VBLANK_B, crtc_state->saveVBLANK); + REG_WRITE(pipeA ? VSYNC_A : VSYNC_B, crtc_state->saveVSYNC); + REG_WRITE(pipeA ? DSPASTRIDE : DSPBSTRIDE, crtc_state->saveDSPSTRIDE); + + REG_WRITE(pipeA ? DSPASIZE : DSPBSIZE, crtc_state->saveDSPSIZE); + REG_WRITE(pipeA ? DSPAPOS : DSPBPOS, crtc_state->saveDSPPOS); + + REG_WRITE(pipeA ? PIPEASRC : PIPEBSRC, crtc_state->savePIPESRC); + REG_WRITE(pipeA ? DSPABASE : DSPBBASE, crtc_state->saveDSPBASE); + REG_WRITE(pipeA ? PIPEACONF : PIPEBCONF, crtc_state->savePIPECONF); + + psb_intel_wait_for_vblank(dev); + + REG_WRITE(pipeA ? DSPACNTR : DSPBCNTR, crtc_state->saveDSPCNTR); + REG_WRITE(pipeA ? DSPABASE : DSPBBASE, crtc_state->saveDSPBASE); + + psb_intel_wait_for_vblank(dev); + + paletteReg = pipeA ? PALETTE_A : PALETTE_B; + for (i = 0; i < 256; ++i) + REG_WRITE(paletteReg + (i << 2), crtc_state->savePalette[i]); +} + +static int psb_intel_crtc_cursor_set(struct drm_crtc *crtc, + struct drm_file *file_priv, + uint32_t handle, + uint32_t width, uint32_t height) +{ + struct drm_device *dev = crtc->dev; + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + int pipe = psb_intel_crtc->pipe; + uint32_t control = (pipe == 0) ? CURACNTR : CURBCNTR; + uint32_t base = (pipe == 0) ? CURABASE : CURBBASE; + uint32_t temp; + size_t addr = 0; + struct gtt_range *gt; + struct drm_gem_object *obj; + int ret; + + /* if we want to turn of the cursor ignore width and height */ + if (!handle) { + /* turn off the cursor */ + temp = CURSOR_MODE_DISABLE; + + if (gma_power_begin(dev, false)) { + REG_WRITE(control, temp); + REG_WRITE(base, 0); + gma_power_end(dev); + } + + /* Unpin the old GEM object */ + if (psb_intel_crtc->cursor_obj) { + gt = container_of(psb_intel_crtc->cursor_obj, + struct gtt_range, gem); + psb_gtt_unpin(gt); + drm_gem_object_unreference(psb_intel_crtc->cursor_obj); + psb_intel_crtc->cursor_obj = NULL; + } + + return 0; + } + + /* Currently we only support 64x64 cursors */ + if (width != 64 || height != 64) { + dev_dbg(dev->dev, "we currently only support 64x64 cursors\n"); + return -EINVAL; + } + + obj = drm_gem_object_lookup(dev, file_priv, handle); + if (!obj) + return -ENOENT; + + if (obj->size < width * height * 4) { + dev_dbg(dev->dev, "buffer is to small\n"); + return -ENOMEM; + } + + gt = container_of(obj, struct gtt_range, gem); + + /* Pin the memory into the GTT */ + ret = psb_gtt_pin(gt); + if (ret) { + dev_err(dev->dev, "Can not pin down handle 0x%x\n", handle); + return ret; + } + + + addr = gt->offset; /* Or resource.start ??? */ + + psb_intel_crtc->cursor_addr = addr; + + temp = 0; + /* set the pipe for the cursor */ + temp |= (pipe << 28); + temp |= CURSOR_MODE_64_ARGB_AX | MCURSOR_GAMMA_ENABLE; + + if (gma_power_begin(dev, false)) { + REG_WRITE(control, temp); + REG_WRITE(base, addr); + gma_power_end(dev); + } + + /* unpin the old bo */ + if (psb_intel_crtc->cursor_obj) { + gt = container_of(psb_intel_crtc->cursor_obj, + struct gtt_range, gem); + psb_gtt_unpin(gt); + drm_gem_object_unreference(psb_intel_crtc->cursor_obj); + psb_intel_crtc->cursor_obj = obj; + } + return 0; +} + +static int psb_intel_crtc_cursor_move(struct drm_crtc *crtc, int x, int y) +{ + struct drm_device *dev = crtc->dev; + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + int pipe = psb_intel_crtc->pipe; + uint32_t temp = 0; + uint32_t addr; + + + if (x < 0) { + temp |= (CURSOR_POS_SIGN << CURSOR_X_SHIFT); + x = -x; + } + if (y < 0) { + temp |= (CURSOR_POS_SIGN << CURSOR_Y_SHIFT); + y = -y; + } + + temp |= ((x & CURSOR_POS_MASK) << CURSOR_X_SHIFT); + temp |= ((y & CURSOR_POS_MASK) << CURSOR_Y_SHIFT); + + addr = psb_intel_crtc->cursor_addr; + + if (gma_power_begin(dev, false)) { + REG_WRITE((pipe == 0) ? CURAPOS : CURBPOS, temp); + REG_WRITE((pipe == 0) ? CURABASE : CURBBASE, addr); + gma_power_end(dev); + } + return 0; +} + +void psb_intel_crtc_gamma_set(struct drm_crtc *crtc, u16 *red, + u16 *green, u16 *blue, uint32_t type, uint32_t size) +{ + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + int i; + + if (size != 256) + return; + + for (i = 0; i < 256; i++) { + psb_intel_crtc->lut_r[i] = red[i] >> 8; + psb_intel_crtc->lut_g[i] = green[i] >> 8; + psb_intel_crtc->lut_b[i] = blue[i] >> 8; + } + + psb_intel_crtc_load_lut(crtc); +} + +static int psb_crtc_set_config(struct drm_mode_set *set) +{ + int ret; + struct drm_device *dev = set->crtc->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + + if (!dev_priv->rpm_enabled) + return drm_crtc_helper_set_config(set); + + pm_runtime_forbid(&dev->pdev->dev); + ret = drm_crtc_helper_set_config(set); + pm_runtime_allow(&dev->pdev->dev); + return ret; +} + +/* Returns the clock of the currently programmed mode of the given pipe. */ +static int psb_intel_crtc_clock_get(struct drm_device *dev, + struct drm_crtc *crtc) +{ + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + int pipe = psb_intel_crtc->pipe; + u32 dpll; + u32 fp; + struct psb_intel_clock_t clock; + bool is_lvds; + struct drm_psb_private *dev_priv = dev->dev_private; + + if (gma_power_begin(dev, false)) { + dpll = REG_READ((pipe == 0) ? DPLL_A : DPLL_B); + if ((dpll & DISPLAY_RATE_SELECT_FPA1) == 0) + fp = REG_READ((pipe == 0) ? FPA0 : FPB0); + else + fp = REG_READ((pipe == 0) ? FPA1 : FPB1); + is_lvds = (pipe == 1) && (REG_READ(LVDS) & LVDS_PORT_EN); + gma_power_end(dev); + } else { + dpll = (pipe == 0) ? + dev_priv->regs.psb.saveDPLL_A : + dev_priv->regs.psb.saveDPLL_B; + + if ((dpll & DISPLAY_RATE_SELECT_FPA1) == 0) + fp = (pipe == 0) ? + dev_priv->regs.psb.saveFPA0 : + dev_priv->regs.psb.saveFPB0; + else + fp = (pipe == 0) ? + dev_priv->regs.psb.saveFPA1 : + dev_priv->regs.psb.saveFPB1; + + is_lvds = (pipe == 1) && (dev_priv->regs.psb.saveLVDS & + LVDS_PORT_EN); + } + + clock.m1 = (fp & FP_M1_DIV_MASK) >> FP_M1_DIV_SHIFT; + clock.m2 = (fp & FP_M2_DIV_MASK) >> FP_M2_DIV_SHIFT; + clock.n = (fp & FP_N_DIV_MASK) >> FP_N_DIV_SHIFT; + + if (is_lvds) { + clock.p1 = + ffs((dpll & + DPLL_FPA01_P1_POST_DIV_MASK_I830_LVDS) >> + DPLL_FPA01_P1_POST_DIV_SHIFT); + clock.p2 = 14; + + if ((dpll & PLL_REF_INPUT_MASK) == + PLLB_REF_INPUT_SPREADSPECTRUMIN) { + /* XXX: might not be 66MHz */ + i8xx_clock(66000, &clock); + } else + i8xx_clock(48000, &clock); + } else { + if (dpll & PLL_P1_DIVIDE_BY_TWO) + clock.p1 = 2; + else { + clock.p1 = + ((dpll & + DPLL_FPA01_P1_POST_DIV_MASK_I830) >> + DPLL_FPA01_P1_POST_DIV_SHIFT) + 2; + } + if (dpll & PLL_P2_DIVIDE_BY_4) + clock.p2 = 4; + else + clock.p2 = 2; + + i8xx_clock(48000, &clock); + } + + /* XXX: It would be nice to validate the clocks, but we can't reuse + * i830PllIsValid() because it relies on the xf86_config connector + * configuration being accurate, which it isn't necessarily. + */ + + return clock.dot; +} + +/** Returns the currently programmed mode of the given pipe. */ +struct drm_display_mode *psb_intel_crtc_mode_get(struct drm_device *dev, + struct drm_crtc *crtc) +{ + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + int pipe = psb_intel_crtc->pipe; + struct drm_display_mode *mode; + int htot; + int hsync; + int vtot; + int vsync; + struct drm_psb_private *dev_priv = dev->dev_private; + + if (gma_power_begin(dev, false)) { + htot = REG_READ((pipe == 0) ? HTOTAL_A : HTOTAL_B); + hsync = REG_READ((pipe == 0) ? HSYNC_A : HSYNC_B); + vtot = REG_READ((pipe == 0) ? VTOTAL_A : VTOTAL_B); + vsync = REG_READ((pipe == 0) ? VSYNC_A : VSYNC_B); + gma_power_end(dev); + } else { + htot = (pipe == 0) ? + dev_priv->regs.psb.saveHTOTAL_A : + dev_priv->regs.psb.saveHTOTAL_B; + hsync = (pipe == 0) ? + dev_priv->regs.psb.saveHSYNC_A : + dev_priv->regs.psb.saveHSYNC_B; + vtot = (pipe == 0) ? + dev_priv->regs.psb.saveVTOTAL_A : + dev_priv->regs.psb.saveVTOTAL_B; + vsync = (pipe == 0) ? + dev_priv->regs.psb.saveVSYNC_A : + dev_priv->regs.psb.saveVSYNC_B; + } + + mode = kzalloc(sizeof(*mode), GFP_KERNEL); + if (!mode) + return NULL; + + mode->clock = psb_intel_crtc_clock_get(dev, crtc); + mode->hdisplay = (htot & 0xffff) + 1; + mode->htotal = ((htot & 0xffff0000) >> 16) + 1; + mode->hsync_start = (hsync & 0xffff) + 1; + mode->hsync_end = ((hsync & 0xffff0000) >> 16) + 1; + mode->vdisplay = (vtot & 0xffff) + 1; + mode->vtotal = ((vtot & 0xffff0000) >> 16) + 1; + mode->vsync_start = (vsync & 0xffff) + 1; + mode->vsync_end = ((vsync & 0xffff0000) >> 16) + 1; + + drm_mode_set_name(mode); + drm_mode_set_crtcinfo(mode, 0); + + return mode; +} + +void psb_intel_crtc_destroy(struct drm_crtc *crtc) +{ + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + struct gtt_range *gt; + + /* Unpin the old GEM object */ + if (psb_intel_crtc->cursor_obj) { + gt = container_of(psb_intel_crtc->cursor_obj, + struct gtt_range, gem); + psb_gtt_unpin(gt); + drm_gem_object_unreference(psb_intel_crtc->cursor_obj); + psb_intel_crtc->cursor_obj = NULL; + } + kfree(psb_intel_crtc->crtc_state); + drm_crtc_cleanup(crtc); + kfree(psb_intel_crtc); +} + +const struct drm_crtc_helper_funcs psb_intel_helper_funcs = { + .dpms = psb_intel_crtc_dpms, + .mode_fixup = psb_intel_crtc_mode_fixup, + .mode_set = psb_intel_crtc_mode_set, + .mode_set_base = psb_intel_pipe_set_base, + .prepare = psb_intel_crtc_prepare, + .commit = psb_intel_crtc_commit, +}; + +const struct drm_crtc_funcs psb_intel_crtc_funcs = { + .save = psb_intel_crtc_save, + .restore = psb_intel_crtc_restore, + .cursor_set = psb_intel_crtc_cursor_set, + .cursor_move = psb_intel_crtc_cursor_move, + .gamma_set = psb_intel_crtc_gamma_set, + .set_config = psb_crtc_set_config, + .destroy = psb_intel_crtc_destroy, +}; + +/* + * Set the default value of cursor control and base register + * to zero. This is a workaround for h/w defect on Oaktrail + */ +static void psb_intel_cursor_init(struct drm_device *dev, int pipe) +{ + u32 control[3] = { CURACNTR, CURBCNTR, CURCCNTR }; + u32 base[3] = { CURABASE, CURBBASE, CURCBASE }; + + REG_WRITE(control[pipe], 0); + REG_WRITE(base[pipe], 0); +} + +void psb_intel_crtc_init(struct drm_device *dev, int pipe, + struct psb_intel_mode_device *mode_dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_intel_crtc *psb_intel_crtc; + int i; + uint16_t *r_base, *g_base, *b_base; + + /* We allocate a extra array of drm_connector pointers + * for fbdev after the crtc */ + psb_intel_crtc = + kzalloc(sizeof(struct psb_intel_crtc) + + (INTELFB_CONN_LIMIT * sizeof(struct drm_connector *)), + GFP_KERNEL); + if (psb_intel_crtc == NULL) + return; + + psb_intel_crtc->crtc_state = + kzalloc(sizeof(struct psb_intel_crtc_state), GFP_KERNEL); + if (!psb_intel_crtc->crtc_state) { + dev_err(dev->dev, "Crtc state error: No memory\n"); + kfree(psb_intel_crtc); + return; + } + + /* Set the CRTC operations from the chip specific data */ + drm_crtc_init(dev, &psb_intel_crtc->base, dev_priv->ops->crtc_funcs); + + drm_mode_crtc_set_gamma_size(&psb_intel_crtc->base, 256); + psb_intel_crtc->pipe = pipe; + psb_intel_crtc->plane = pipe; + + r_base = psb_intel_crtc->base.gamma_store; + g_base = r_base + 256; + b_base = g_base + 256; + for (i = 0; i < 256; i++) { + psb_intel_crtc->lut_r[i] = i; + psb_intel_crtc->lut_g[i] = i; + psb_intel_crtc->lut_b[i] = i; + r_base[i] = i << 8; + g_base[i] = i << 8; + b_base[i] = i << 8; + + psb_intel_crtc->lut_adj[i] = 0; + } + + psb_intel_crtc->mode_dev = mode_dev; + psb_intel_crtc->cursor_addr = 0; + + drm_crtc_helper_add(&psb_intel_crtc->base, + dev_priv->ops->crtc_helper); + + /* Setup the array of drm_connector pointer array */ + psb_intel_crtc->mode_set.crtc = &psb_intel_crtc->base; + BUG_ON(pipe >= ARRAY_SIZE(dev_priv->plane_to_crtc_mapping) || + dev_priv->plane_to_crtc_mapping[psb_intel_crtc->plane] != NULL); + dev_priv->plane_to_crtc_mapping[psb_intel_crtc->plane] = + &psb_intel_crtc->base; + dev_priv->pipe_to_crtc_mapping[psb_intel_crtc->pipe] = + &psb_intel_crtc->base; + psb_intel_crtc->mode_set.connectors = + (struct drm_connector **) (psb_intel_crtc + 1); + psb_intel_crtc->mode_set.num_connectors = 0; + psb_intel_cursor_init(dev, pipe); +} + +int psb_intel_get_pipe_from_crtc_id(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct drm_psb_get_pipe_from_crtc_id_arg *pipe_from_crtc_id = data; + struct drm_mode_object *drmmode_obj; + struct psb_intel_crtc *crtc; + + if (!dev_priv) { + dev_err(dev->dev, "called with no initialization\n"); + return -EINVAL; + } + + drmmode_obj = drm_mode_object_find(dev, pipe_from_crtc_id->crtc_id, + DRM_MODE_OBJECT_CRTC); + + if (!drmmode_obj) { + dev_err(dev->dev, "no such CRTC id\n"); + return -EINVAL; + } + + crtc = to_psb_intel_crtc(obj_to_crtc(drmmode_obj)); + pipe_from_crtc_id->pipe = crtc->pipe; + + return 0; +} + +struct drm_crtc *psb_intel_get_crtc_from_pipe(struct drm_device *dev, int pipe) +{ + struct drm_crtc *crtc = NULL; + + list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + if (psb_intel_crtc->pipe == pipe) + break; + } + return crtc; +} + +int psb_intel_connector_clones(struct drm_device *dev, int type_mask) +{ + int index_mask = 0; + struct drm_connector *connector; + int entry = 0; + + list_for_each_entry(connector, &dev->mode_config.connector_list, + head) { + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(connector); + if (type_mask & (1 << psb_intel_encoder->type)) + index_mask |= (1 << entry); + entry++; + } + return index_mask; +} + +/* current intel driver doesn't take advantage of encoders + always give back the encoder for the connector +*/ +struct drm_encoder *psb_intel_best_encoder(struct drm_connector *connector) +{ + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(connector); + + return &psb_intel_encoder->base; +} + +void psb_intel_connector_attach_encoder(struct psb_intel_connector *connector, + struct psb_intel_encoder *encoder) +{ + connector->encoder = encoder; + drm_mode_connector_attach_encoder(&connector->base, + &encoder->base); +} diff --git a/drivers/gpu/drm/gma500/psb_intel_display.h b/drivers/gpu/drm/gma500/psb_intel_display.h new file mode 100644 index 0000000..535b49a --- /dev/null +++ b/drivers/gpu/drm/gma500/psb_intel_display.h @@ -0,0 +1,28 @@ +/* copyright (c) 2008, Intel Corporation + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + * Authors: + * Eric Anholt + */ + +#ifndef _INTEL_DISPLAY_H_ +#define _INTEL_DISPLAY_H_ + +bool psb_intel_pipe_has_type(struct drm_crtc *crtc, int type); +void psb_intel_crtc_gamma_set(struct drm_crtc *crtc, u16 *red, + u16 *green, u16 *blue, uint32_t type, uint32_t size); +void psb_intel_crtc_destroy(struct drm_crtc *crtc); + +#endif diff --git a/drivers/gpu/drm/gma500/psb_intel_drv.h b/drivers/gpu/drm/gma500/psb_intel_drv.h new file mode 100644 index 0000000..f40535e --- /dev/null +++ b/drivers/gpu/drm/gma500/psb_intel_drv.h @@ -0,0 +1,289 @@ +/* + * Copyright (c) 2009-2011, Intel Corporation. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + */ + +#ifndef __INTEL_DRV_H__ +#define __INTEL_DRV_H__ + +#include +#include +#include +#include +#include + +/* + * Display related stuff + */ + +/* store information about an Ixxx DVO */ +/* The i830->i865 use multiple DVOs with multiple i2cs */ +/* the i915, i945 have a single sDVO i2c bus - which is different */ +#define MAX_OUTPUTS 6 +/* maximum connectors per crtcs in the mode set */ +#define INTELFB_CONN_LIMIT 4 + +#define INTEL_I2C_BUS_DVO 1 +#define INTEL_I2C_BUS_SDVO 2 + +/* Intel Pipe Clone Bit */ +#define INTEL_HDMIB_CLONE_BIT 1 +#define INTEL_HDMIC_CLONE_BIT 2 +#define INTEL_HDMID_CLONE_BIT 3 +#define INTEL_HDMIE_CLONE_BIT 4 +#define INTEL_HDMIF_CLONE_BIT 5 +#define INTEL_SDVO_NON_TV_CLONE_BIT 6 +#define INTEL_SDVO_TV_CLONE_BIT 7 +#define INTEL_SDVO_LVDS_CLONE_BIT 8 +#define INTEL_ANALOG_CLONE_BIT 9 +#define INTEL_TV_CLONE_BIT 10 +#define INTEL_DP_B_CLONE_BIT 11 +#define INTEL_DP_C_CLONE_BIT 12 +#define INTEL_DP_D_CLONE_BIT 13 +#define INTEL_LVDS_CLONE_BIT 14 +#define INTEL_DVO_TMDS_CLONE_BIT 15 +#define INTEL_DVO_LVDS_CLONE_BIT 16 +#define INTEL_EDP_CLONE_BIT 17 + +/* these are outputs from the chip - integrated only + * external chips are via DVO or SDVO output */ +#define INTEL_OUTPUT_UNUSED 0 +#define INTEL_OUTPUT_ANALOG 1 +#define INTEL_OUTPUT_DVO 2 +#define INTEL_OUTPUT_SDVO 3 +#define INTEL_OUTPUT_LVDS 4 +#define INTEL_OUTPUT_TVOUT 5 +#define INTEL_OUTPUT_HDMI 6 +#define INTEL_OUTPUT_MIPI 7 +#define INTEL_OUTPUT_MIPI2 8 + +#define INTEL_DVO_CHIP_NONE 0 +#define INTEL_DVO_CHIP_LVDS 1 +#define INTEL_DVO_CHIP_TMDS 2 +#define INTEL_DVO_CHIP_TVOUT 4 + +#define INTEL_MODE_PIXEL_MULTIPLIER_SHIFT (0x0) +#define INTEL_MODE_PIXEL_MULTIPLIER_MASK (0xf << INTEL_MODE_PIXEL_MULTIPLIER_SHIFT) + +static inline void +psb_intel_mode_set_pixel_multiplier(struct drm_display_mode *mode, + int multiplier) +{ + mode->clock *= multiplier; + mode->private_flags |= multiplier; +} + +static inline int +psb_intel_mode_get_pixel_multiplier(const struct drm_display_mode *mode) +{ + return (mode->private_flags & INTEL_MODE_PIXEL_MULTIPLIER_MASK) + >> INTEL_MODE_PIXEL_MULTIPLIER_SHIFT; +} + + +/* + * Hold information useally put on the device driver privates here, + * since it needs to be shared across multiple of devices drivers privates. + */ +struct psb_intel_mode_device { + + /* + * Abstracted memory manager operations + */ + size_t(*bo_offset) (struct drm_device *dev, void *bo); + + /* + * Cursor (Can go ?) + */ + int cursor_needs_physical; + + /* + * LVDS info + */ + int backlight_duty_cycle; /* restore backlight to this value */ + bool panel_wants_dither; + struct drm_display_mode *panel_fixed_mode; + struct drm_display_mode *panel_fixed_mode2; + struct drm_display_mode *vbt_mode; /* if any */ + + uint32_t saveBLC_PWM_CTL; +}; + +struct psb_intel_i2c_chan { + /* for getting at dev. private (mmio etc.) */ + struct drm_device *drm_dev; + u32 reg; /* GPIO reg */ + struct i2c_adapter adapter; + struct i2c_algo_bit_data algo; + u8 slave_addr; +}; + +struct psb_intel_encoder { + struct drm_encoder base; + int type; + bool needs_tv_clock; + void (*hot_plug)(struct psb_intel_encoder *); + int crtc_mask; + int clone_mask; + void *dev_priv; /* For sdvo_priv, lvds_priv, etc... */ + + /* FIXME: Either make SDVO and LVDS store it's i2c here or give CDV it's + own set of output privates */ + struct psb_intel_i2c_chan *i2c_bus; + struct psb_intel_i2c_chan *ddc_bus; +}; + +struct psb_intel_connector { + struct drm_connector base; + struct psb_intel_encoder *encoder; +}; + +struct psb_intel_crtc_state { + uint32_t saveDSPCNTR; + uint32_t savePIPECONF; + uint32_t savePIPESRC; + uint32_t saveDPLL; + uint32_t saveFP0; + uint32_t saveFP1; + uint32_t saveHTOTAL; + uint32_t saveHBLANK; + uint32_t saveHSYNC; + uint32_t saveVTOTAL; + uint32_t saveVBLANK; + uint32_t saveVSYNC; + uint32_t saveDSPSTRIDE; + uint32_t saveDSPSIZE; + uint32_t saveDSPPOS; + uint32_t saveDSPBASE; + uint32_t savePalette[256]; +}; + +struct psb_intel_crtc { + struct drm_crtc base; + int pipe; + int plane; + uint32_t cursor_addr; + u8 lut_r[256], lut_g[256], lut_b[256]; + u8 lut_adj[256]; + struct psb_intel_framebuffer *fbdev_fb; + /* a mode_set for fbdev users on this crtc */ + struct drm_mode_set mode_set; + + /* GEM object that holds our cursor */ + struct drm_gem_object *cursor_obj; + + struct drm_display_mode saved_mode; + struct drm_display_mode saved_adjusted_mode; + + struct psb_intel_mode_device *mode_dev; + + /*crtc mode setting flags*/ + u32 mode_flags; + + /* Saved Crtc HW states */ + struct psb_intel_crtc_state *crtc_state; +}; + +#define to_psb_intel_crtc(x) \ + container_of(x, struct psb_intel_crtc, base) +#define to_psb_intel_connector(x) \ + container_of(x, struct psb_intel_connector, base) +#define to_psb_intel_encoder(x) \ + container_of(x, struct psb_intel_encoder, base) +#define to_psb_intel_framebuffer(x) \ + container_of(x, struct psb_intel_framebuffer, base) + +struct psb_intel_i2c_chan *psb_intel_i2c_create(struct drm_device *dev, + const u32 reg, const char *name); +void psb_intel_i2c_destroy(struct psb_intel_i2c_chan *chan); +int psb_intel_ddc_get_modes(struct drm_connector *connector, + struct i2c_adapter *adapter); +extern bool psb_intel_ddc_probe(struct i2c_adapter *adapter); + +extern void psb_intel_crtc_init(struct drm_device *dev, int pipe, + struct psb_intel_mode_device *mode_dev); +extern void psb_intel_crt_init(struct drm_device *dev); +extern bool psb_intel_sdvo_init(struct drm_device *dev, int output_device); +extern void psb_intel_dvo_init(struct drm_device *dev); +extern void psb_intel_tv_init(struct drm_device *dev); +extern void psb_intel_lvds_init(struct drm_device *dev, + struct psb_intel_mode_device *mode_dev); +extern void psb_intel_lvds_set_brightness(struct drm_device *dev, int level); +extern void oaktrail_lvds_init(struct drm_device *dev, + struct psb_intel_mode_device *mode_dev); +extern void oaktrail_wait_for_INTR_PKT_SENT(struct drm_device *dev); +extern void oaktrail_dsi_init(struct drm_device *dev, + struct psb_intel_mode_device *mode_dev); +extern void mid_dsi_init(struct drm_device *dev, + struct psb_intel_mode_device *mode_dev, int dsi_num); + +extern void psb_intel_crtc_load_lut(struct drm_crtc *crtc); +extern void psb_intel_encoder_prepare(struct drm_encoder *encoder); +extern void psb_intel_encoder_commit(struct drm_encoder *encoder); +extern void psb_intel_encoder_destroy(struct drm_encoder *encoder); + +static inline struct psb_intel_encoder *psb_intel_attached_encoder( + struct drm_connector *connector) +{ + return to_psb_intel_connector(connector)->encoder; +} + +extern void psb_intel_connector_attach_encoder( + struct psb_intel_connector *connector, + struct psb_intel_encoder *encoder); + +extern struct drm_encoder *psb_intel_best_encoder(struct drm_connector + *connector); + +extern struct drm_display_mode *psb_intel_crtc_mode_get(struct drm_device *dev, + struct drm_crtc *crtc); +extern void psb_intel_wait_for_vblank(struct drm_device *dev); +extern int psb_intel_get_pipe_from_crtc_id(struct drm_device *dev, void *data, + struct drm_file *file_priv); +extern struct drm_crtc *psb_intel_get_crtc_from_pipe(struct drm_device *dev, + int pipe); +extern struct drm_connector *psb_intel_sdvo_find(struct drm_device *dev, + int sdvoB); +extern int psb_intel_sdvo_supports_hotplug(struct drm_connector *connector); +extern void psb_intel_sdvo_set_hotplug(struct drm_connector *connector, + int enable); +extern int intelfb_probe(struct drm_device *dev); +extern int intelfb_remove(struct drm_device *dev, + struct drm_framebuffer *fb); +extern struct drm_framebuffer *psb_intel_framebuffer_create(struct drm_device + *dev, struct + drm_mode_fb_cmd + *mode_cmd, + void *mm_private); +extern bool psb_intel_lvds_mode_fixup(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode); +extern int psb_intel_lvds_mode_valid(struct drm_connector *connector, + struct drm_display_mode *mode); +extern int psb_intel_lvds_set_property(struct drm_connector *connector, + struct drm_property *property, + uint64_t value); +extern void psb_intel_lvds_destroy(struct drm_connector *connector); +extern const struct drm_encoder_funcs psb_intel_lvds_enc_funcs; + +/* intel_gmbus.c */ +extern void gma_intel_i2c_reset(struct drm_device *dev); +extern int gma_intel_setup_gmbus(struct drm_device *dev); +extern void gma_intel_gmbus_set_speed(struct i2c_adapter *adapter, int speed); +extern void gma_intel_gmbus_force_bit(struct i2c_adapter *adapter, bool force_bit); +extern void gma_intel_teardown_gmbus(struct drm_device *dev); + +#endif /* __INTEL_DRV_H__ */ diff --git a/drivers/gpu/drm/gma500/psb_intel_lvds.c b/drivers/gpu/drm/gma500/psb_intel_lvds.c new file mode 100644 index 0000000..c83f5b5 --- /dev/null +++ b/drivers/gpu/drm/gma500/psb_intel_lvds.c @@ -0,0 +1,867 @@ +/* + * Copyright © 2006-2007 Intel Corporation + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + * Authors: + * Eric Anholt + * Dave Airlie + * Jesse Barnes + */ + +#include +#include + +#include "intel_bios.h" +#include "psb_drv.h" +#include "psb_intel_drv.h" +#include "psb_intel_reg.h" +#include "power.h" +#include + +/* + * LVDS I2C backlight control macros + */ +#define BRIGHTNESS_MAX_LEVEL 100 +#define BRIGHTNESS_MASK 0xFF +#define BLC_I2C_TYPE 0x01 +#define BLC_PWM_TYPT 0x02 + +#define BLC_POLARITY_NORMAL 0 +#define BLC_POLARITY_INVERSE 1 + +#define PSB_BLC_MAX_PWM_REG_FREQ (0xFFFE) +#define PSB_BLC_MIN_PWM_REG_FREQ (0x2) +#define PSB_BLC_PWM_PRECISION_FACTOR (10) +#define PSB_BACKLIGHT_PWM_CTL_SHIFT (16) +#define PSB_BACKLIGHT_PWM_POLARITY_BIT_CLEAR (0xFFFE) + +struct psb_intel_lvds_priv { + /* + * Saved LVDO output states + */ + uint32_t savePP_ON; + uint32_t savePP_OFF; + uint32_t saveLVDS; + uint32_t savePP_CONTROL; + uint32_t savePP_CYCLE; + uint32_t savePFIT_CONTROL; + uint32_t savePFIT_PGM_RATIOS; + uint32_t saveBLC_PWM_CTL; + + struct psb_intel_i2c_chan *i2c_bus; + struct psb_intel_i2c_chan *ddc_bus; +}; + + +/* + * Returns the maximum level of the backlight duty cycle field. + */ +static u32 psb_intel_lvds_get_max_backlight(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + u32 ret; + + if (gma_power_begin(dev, false)) { + ret = REG_READ(BLC_PWM_CTL); + gma_power_end(dev); + } else /* Powered off, use the saved value */ + ret = dev_priv->regs.saveBLC_PWM_CTL; + + /* Top 15bits hold the frequency mask */ + ret = (ret & BACKLIGHT_MODULATION_FREQ_MASK) >> + BACKLIGHT_MODULATION_FREQ_SHIFT; + + ret *= 2; /* Return a 16bit range as needed for setting */ + if (ret == 0) + dev_err(dev->dev, "BL bug: Reg %08x save %08X\n", + REG_READ(BLC_PWM_CTL), dev_priv->regs.saveBLC_PWM_CTL); + return ret; +} + +/* + * Set LVDS backlight level by I2C command + * + * FIXME: at some point we need to both track this for PM and also + * disable runtime pm on MRST if the brightness is nil (ie blanked) + */ +static int psb_lvds_i2c_set_brightness(struct drm_device *dev, + unsigned int level) +{ + struct drm_psb_private *dev_priv = + (struct drm_psb_private *)dev->dev_private; + + struct psb_intel_i2c_chan *lvds_i2c_bus = dev_priv->lvds_i2c_bus; + u8 out_buf[2]; + unsigned int blc_i2c_brightness; + + struct i2c_msg msgs[] = { + { + .addr = lvds_i2c_bus->slave_addr, + .flags = 0, + .len = 2, + .buf = out_buf, + } + }; + + blc_i2c_brightness = BRIGHTNESS_MASK & ((unsigned int)level * + BRIGHTNESS_MASK / + BRIGHTNESS_MAX_LEVEL); + + if (dev_priv->lvds_bl->pol == BLC_POLARITY_INVERSE) + blc_i2c_brightness = BRIGHTNESS_MASK - blc_i2c_brightness; + + out_buf[0] = dev_priv->lvds_bl->brightnesscmd; + out_buf[1] = (u8)blc_i2c_brightness; + + if (i2c_transfer(&lvds_i2c_bus->adapter, msgs, 1) == 1) { + dev_dbg(dev->dev, "I2C set brightness.(command, value) (%d, %d)\n", + dev_priv->lvds_bl->brightnesscmd, + blc_i2c_brightness); + return 0; + } + + dev_err(dev->dev, "I2C transfer error\n"); + return -1; +} + + +static int psb_lvds_pwm_set_brightness(struct drm_device *dev, int level) +{ + struct drm_psb_private *dev_priv = + (struct drm_psb_private *)dev->dev_private; + + u32 max_pwm_blc; + u32 blc_pwm_duty_cycle; + + max_pwm_blc = psb_intel_lvds_get_max_backlight(dev); + + /*BLC_PWM_CTL Should be initiated while backlight device init*/ + BUG_ON(max_pwm_blc == 0); + + blc_pwm_duty_cycle = level * max_pwm_blc / BRIGHTNESS_MAX_LEVEL; + + if (dev_priv->lvds_bl->pol == BLC_POLARITY_INVERSE) + blc_pwm_duty_cycle = max_pwm_blc - blc_pwm_duty_cycle; + + blc_pwm_duty_cycle &= PSB_BACKLIGHT_PWM_POLARITY_BIT_CLEAR; + REG_WRITE(BLC_PWM_CTL, + (max_pwm_blc << PSB_BACKLIGHT_PWM_CTL_SHIFT) | + (blc_pwm_duty_cycle)); + + dev_info(dev->dev, "Backlight lvds set brightness %08x\n", + (max_pwm_blc << PSB_BACKLIGHT_PWM_CTL_SHIFT) | + (blc_pwm_duty_cycle)); + + return 0; +} + +/* + * Set LVDS backlight level either by I2C or PWM + */ +void psb_intel_lvds_set_brightness(struct drm_device *dev, int level) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + + dev_dbg(dev->dev, "backlight level is %d\n", level); + + if (!dev_priv->lvds_bl) { + dev_err(dev->dev, "NO LVDS backlight info\n"); + return; + } + + if (dev_priv->lvds_bl->type == BLC_I2C_TYPE) + psb_lvds_i2c_set_brightness(dev, level); + else + psb_lvds_pwm_set_brightness(dev, level); +} + +/* + * Sets the backlight level. + * + * level: backlight level, from 0 to psb_intel_lvds_get_max_backlight(). + */ +static void psb_intel_lvds_set_backlight(struct drm_device *dev, int level) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + u32 blc_pwm_ctl; + + if (gma_power_begin(dev, false)) { + blc_pwm_ctl = REG_READ(BLC_PWM_CTL); + blc_pwm_ctl &= ~BACKLIGHT_DUTY_CYCLE_MASK; + REG_WRITE(BLC_PWM_CTL, + (blc_pwm_ctl | + (level << BACKLIGHT_DUTY_CYCLE_SHIFT))); + dev_priv->regs.saveBLC_PWM_CTL = (blc_pwm_ctl | + (level << BACKLIGHT_DUTY_CYCLE_SHIFT)); + gma_power_end(dev); + } else { + blc_pwm_ctl = dev_priv->regs.saveBLC_PWM_CTL & + ~BACKLIGHT_DUTY_CYCLE_MASK; + dev_priv->regs.saveBLC_PWM_CTL = (blc_pwm_ctl | + (level << BACKLIGHT_DUTY_CYCLE_SHIFT)); + } +} + +/* + * Sets the power state for the panel. + */ +static void psb_intel_lvds_set_power(struct drm_device *dev, bool on) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_intel_mode_device *mode_dev = &dev_priv->mode_dev; + u32 pp_status; + + if (!gma_power_begin(dev, true)) { + dev_err(dev->dev, "set power, chip off!\n"); + return; + } + + if (on) { + REG_WRITE(PP_CONTROL, REG_READ(PP_CONTROL) | + POWER_TARGET_ON); + do { + pp_status = REG_READ(PP_STATUS); + } while ((pp_status & PP_ON) == 0); + + psb_intel_lvds_set_backlight(dev, + mode_dev->backlight_duty_cycle); + } else { + psb_intel_lvds_set_backlight(dev, 0); + + REG_WRITE(PP_CONTROL, REG_READ(PP_CONTROL) & + ~POWER_TARGET_ON); + do { + pp_status = REG_READ(PP_STATUS); + } while (pp_status & PP_ON); + } + + gma_power_end(dev); +} + +static void psb_intel_lvds_encoder_dpms(struct drm_encoder *encoder, int mode) +{ + struct drm_device *dev = encoder->dev; + + if (mode == DRM_MODE_DPMS_ON) + psb_intel_lvds_set_power(dev, true); + else + psb_intel_lvds_set_power(dev, false); + + /* XXX: We never power down the LVDS pairs. */ +} + +static void psb_intel_lvds_save(struct drm_connector *connector) +{ + struct drm_device *dev = connector->dev; + struct drm_psb_private *dev_priv = + (struct drm_psb_private *)dev->dev_private; + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(connector); + struct psb_intel_lvds_priv *lvds_priv = + (struct psb_intel_lvds_priv *)psb_intel_encoder->dev_priv; + + lvds_priv->savePP_ON = REG_READ(LVDSPP_ON); + lvds_priv->savePP_OFF = REG_READ(LVDSPP_OFF); + lvds_priv->saveLVDS = REG_READ(LVDS); + lvds_priv->savePP_CONTROL = REG_READ(PP_CONTROL); + lvds_priv->savePP_CYCLE = REG_READ(PP_CYCLE); + /*lvds_priv->savePP_DIVISOR = REG_READ(PP_DIVISOR);*/ + lvds_priv->saveBLC_PWM_CTL = REG_READ(BLC_PWM_CTL); + lvds_priv->savePFIT_CONTROL = REG_READ(PFIT_CONTROL); + lvds_priv->savePFIT_PGM_RATIOS = REG_READ(PFIT_PGM_RATIOS); + + /*TODO: move backlight_duty_cycle to psb_intel_lvds_priv*/ + dev_priv->backlight_duty_cycle = (dev_priv->regs.saveBLC_PWM_CTL & + BACKLIGHT_DUTY_CYCLE_MASK); + + /* + * If the light is off at server startup, + * just make it full brightness + */ + if (dev_priv->backlight_duty_cycle == 0) + dev_priv->backlight_duty_cycle = + psb_intel_lvds_get_max_backlight(dev); + + dev_dbg(dev->dev, "(0x%x, 0x%x, 0x%x, 0x%x, 0x%x, 0x%x)\n", + lvds_priv->savePP_ON, + lvds_priv->savePP_OFF, + lvds_priv->saveLVDS, + lvds_priv->savePP_CONTROL, + lvds_priv->savePP_CYCLE, + lvds_priv->saveBLC_PWM_CTL); +} + +static void psb_intel_lvds_restore(struct drm_connector *connector) +{ + struct drm_device *dev = connector->dev; + u32 pp_status; + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(connector); + struct psb_intel_lvds_priv *lvds_priv = + (struct psb_intel_lvds_priv *)psb_intel_encoder->dev_priv; + + dev_dbg(dev->dev, "(0x%x, 0x%x, 0x%x, 0x%x, 0x%x, 0x%x)\n", + lvds_priv->savePP_ON, + lvds_priv->savePP_OFF, + lvds_priv->saveLVDS, + lvds_priv->savePP_CONTROL, + lvds_priv->savePP_CYCLE, + lvds_priv->saveBLC_PWM_CTL); + + REG_WRITE(BLC_PWM_CTL, lvds_priv->saveBLC_PWM_CTL); + REG_WRITE(PFIT_CONTROL, lvds_priv->savePFIT_CONTROL); + REG_WRITE(PFIT_PGM_RATIOS, lvds_priv->savePFIT_PGM_RATIOS); + REG_WRITE(LVDSPP_ON, lvds_priv->savePP_ON); + REG_WRITE(LVDSPP_OFF, lvds_priv->savePP_OFF); + /*REG_WRITE(PP_DIVISOR, lvds_priv->savePP_DIVISOR);*/ + REG_WRITE(PP_CYCLE, lvds_priv->savePP_CYCLE); + REG_WRITE(PP_CONTROL, lvds_priv->savePP_CONTROL); + REG_WRITE(LVDS, lvds_priv->saveLVDS); + + if (lvds_priv->savePP_CONTROL & POWER_TARGET_ON) { + REG_WRITE(PP_CONTROL, REG_READ(PP_CONTROL) | + POWER_TARGET_ON); + do { + pp_status = REG_READ(PP_STATUS); + } while ((pp_status & PP_ON) == 0); + } else { + REG_WRITE(PP_CONTROL, REG_READ(PP_CONTROL) & + ~POWER_TARGET_ON); + do { + pp_status = REG_READ(PP_STATUS); + } while (pp_status & PP_ON); + } +} + +int psb_intel_lvds_mode_valid(struct drm_connector *connector, + struct drm_display_mode *mode) +{ + struct drm_psb_private *dev_priv = connector->dev->dev_private; + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(connector); + struct drm_display_mode *fixed_mode = + dev_priv->mode_dev.panel_fixed_mode; + + if (psb_intel_encoder->type == INTEL_OUTPUT_MIPI2) + fixed_mode = dev_priv->mode_dev.panel_fixed_mode2; + + /* just in case */ + if (mode->flags & DRM_MODE_FLAG_DBLSCAN) + return MODE_NO_DBLESCAN; + + /* just in case */ + if (mode->flags & DRM_MODE_FLAG_INTERLACE) + return MODE_NO_INTERLACE; + + if (fixed_mode) { + if (mode->hdisplay > fixed_mode->hdisplay) + return MODE_PANEL; + if (mode->vdisplay > fixed_mode->vdisplay) + return MODE_PANEL; + } + return MODE_OK; +} + +bool psb_intel_lvds_mode_fixup(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + struct drm_device *dev = encoder->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_intel_mode_device *mode_dev = &dev_priv->mode_dev; + struct psb_intel_crtc *psb_intel_crtc = + to_psb_intel_crtc(encoder->crtc); + struct drm_encoder *tmp_encoder; + struct drm_display_mode *panel_fixed_mode = mode_dev->panel_fixed_mode; + struct psb_intel_encoder *psb_intel_encoder = + to_psb_intel_encoder(encoder); + + if (psb_intel_encoder->type == INTEL_OUTPUT_MIPI2) + panel_fixed_mode = mode_dev->panel_fixed_mode2; + + /* PSB requires the LVDS is on pipe B, MRST has only one pipe anyway */ + if (!IS_MRST(dev) && psb_intel_crtc->pipe == 0) { + printk(KERN_ERR "Can't support LVDS on pipe A\n"); + return false; + } + if (IS_MRST(dev) && psb_intel_crtc->pipe != 0) { + printk(KERN_ERR "Must use PIPE A\n"); + return false; + } + /* Should never happen!! */ + list_for_each_entry(tmp_encoder, &dev->mode_config.encoder_list, + head) { + if (tmp_encoder != encoder + && tmp_encoder->crtc == encoder->crtc) { + printk(KERN_ERR "Can't enable LVDS and another " + "encoder on the same pipe\n"); + return false; + } + } + + /* + * If we have timings from the BIOS for the panel, put them in + * to the adjusted mode. The CRTC will be set up for this mode, + * with the panel scaling set up to source from the H/VDisplay + * of the original mode. + */ + if (panel_fixed_mode != NULL) { + adjusted_mode->hdisplay = panel_fixed_mode->hdisplay; + adjusted_mode->hsync_start = panel_fixed_mode->hsync_start; + adjusted_mode->hsync_end = panel_fixed_mode->hsync_end; + adjusted_mode->htotal = panel_fixed_mode->htotal; + adjusted_mode->vdisplay = panel_fixed_mode->vdisplay; + adjusted_mode->vsync_start = panel_fixed_mode->vsync_start; + adjusted_mode->vsync_end = panel_fixed_mode->vsync_end; + adjusted_mode->vtotal = panel_fixed_mode->vtotal; + adjusted_mode->clock = panel_fixed_mode->clock; + drm_mode_set_crtcinfo(adjusted_mode, + CRTC_INTERLACE_HALVE_V); + } + + /* + * XXX: It would be nice to support lower refresh rates on the + * panels to reduce power consumption, and perhaps match the + * user's requested refresh rate. + */ + + return true; +} + +static void psb_intel_lvds_prepare(struct drm_encoder *encoder) +{ + struct drm_device *dev = encoder->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_intel_mode_device *mode_dev = &dev_priv->mode_dev; + + if (!gma_power_begin(dev, true)) + return; + + mode_dev->saveBLC_PWM_CTL = REG_READ(BLC_PWM_CTL); + mode_dev->backlight_duty_cycle = (mode_dev->saveBLC_PWM_CTL & + BACKLIGHT_DUTY_CYCLE_MASK); + + psb_intel_lvds_set_power(dev, false); + + gma_power_end(dev); +} + +static void psb_intel_lvds_commit(struct drm_encoder *encoder) +{ + struct drm_device *dev = encoder->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_intel_mode_device *mode_dev = &dev_priv->mode_dev; + + if (mode_dev->backlight_duty_cycle == 0) + mode_dev->backlight_duty_cycle = + psb_intel_lvds_get_max_backlight(dev); + + psb_intel_lvds_set_power(dev, true); +} + +static void psb_intel_lvds_mode_set(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + struct drm_device *dev = encoder->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + u32 pfit_control; + + /* + * The LVDS pin pair will already have been turned on in the + * psb_intel_crtc_mode_set since it has a large impact on the DPLL + * settings. + */ + + /* + * Enable automatic panel scaling so that non-native modes fill the + * screen. Should be enabled before the pipe is enabled, according to + * register description and PRM. + */ + if (mode->hdisplay != adjusted_mode->hdisplay || + mode->vdisplay != adjusted_mode->vdisplay) + pfit_control = (PFIT_ENABLE | VERT_AUTO_SCALE | + HORIZ_AUTO_SCALE | VERT_INTERP_BILINEAR | + HORIZ_INTERP_BILINEAR); + else + pfit_control = 0; + + if (dev_priv->lvds_dither) + pfit_control |= PANEL_8TO6_DITHER_ENABLE; + + REG_WRITE(PFIT_CONTROL, pfit_control); +} + +/* + * Detect the LVDS connection. + * + * This always returns CONNECTOR_STATUS_CONNECTED. + * This connector should only have + * been set up if the LVDS was actually connected anyway. + */ +static enum drm_connector_status psb_intel_lvds_detect(struct drm_connector + *connector, bool force) +{ + return connector_status_connected; +} + +/* + * Return the list of DDC modes if available, or the BIOS fixed mode otherwise. + */ +static int psb_intel_lvds_get_modes(struct drm_connector *connector) +{ + struct drm_device *dev = connector->dev; + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_intel_mode_device *mode_dev = &dev_priv->mode_dev; + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(connector); + struct psb_intel_lvds_priv *lvds_priv = psb_intel_encoder->dev_priv; + int ret = 0; + + if (!IS_MRST(dev)) + ret = psb_intel_ddc_get_modes(connector, &lvds_priv->i2c_bus->adapter); + + if (ret) + return ret; + + /* Didn't get an EDID, so + * Set wide sync ranges so we get all modes + * handed to valid_mode for checking + */ + connector->display_info.min_vfreq = 0; + connector->display_info.max_vfreq = 200; + connector->display_info.min_hfreq = 0; + connector->display_info.max_hfreq = 200; + + if (mode_dev->panel_fixed_mode != NULL) { + struct drm_display_mode *mode = + drm_mode_duplicate(dev, mode_dev->panel_fixed_mode); + drm_mode_probed_add(connector, mode); + return 1; + } + + return 0; +} + +/** + * psb_intel_lvds_destroy - unregister and free LVDS structures + * @connector: connector to free + * + * Unregister the DDC bus for this connector then free the driver private + * structure. + */ +void psb_intel_lvds_destroy(struct drm_connector *connector) +{ + struct psb_intel_encoder *psb_intel_encoder = + psb_intel_attached_encoder(connector); + struct psb_intel_lvds_priv *lvds_priv = psb_intel_encoder->dev_priv; + + if (lvds_priv->ddc_bus) + psb_intel_i2c_destroy(lvds_priv->ddc_bus); + drm_sysfs_connector_remove(connector); + drm_connector_cleanup(connector); + kfree(connector); +} + +int psb_intel_lvds_set_property(struct drm_connector *connector, + struct drm_property *property, + uint64_t value) +{ + struct drm_encoder *encoder = connector->encoder; + + if (!encoder) + return -1; + + if (!strcmp(property->name, "scaling mode")) { + struct psb_intel_crtc *crtc = + to_psb_intel_crtc(encoder->crtc); + uint64_t curval; + + if (!crtc) + goto set_prop_error; + + switch (value) { + case DRM_MODE_SCALE_FULLSCREEN: + break; + case DRM_MODE_SCALE_NO_SCALE: + break; + case DRM_MODE_SCALE_ASPECT: + break; + default: + goto set_prop_error; + } + + if (drm_connector_property_get_value(connector, + property, + &curval)) + goto set_prop_error; + + if (curval == value) + goto set_prop_done; + + if (drm_connector_property_set_value(connector, + property, + value)) + goto set_prop_error; + + if (crtc->saved_mode.hdisplay != 0 && + crtc->saved_mode.vdisplay != 0) { + if (!drm_crtc_helper_set_mode(encoder->crtc, + &crtc->saved_mode, + encoder->crtc->x, + encoder->crtc->y, + encoder->crtc->fb)) + goto set_prop_error; + } + } else if (!strcmp(property->name, "backlight")) { + if (drm_connector_property_set_value(connector, + property, + value)) + goto set_prop_error; + else { +#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE + struct drm_psb_private *devp = + encoder->dev->dev_private; + struct backlight_device *bd = devp->backlight_device; + if (bd) { + bd->props.brightness = value; + backlight_update_status(bd); + } +#endif + } + } else if (!strcmp(property->name, "DPMS")) { + struct drm_encoder_helper_funcs *hfuncs + = encoder->helper_private; + hfuncs->dpms(encoder, value); + } + +set_prop_done: + return 0; +set_prop_error: + return -1; +} + +static const struct drm_encoder_helper_funcs psb_intel_lvds_helper_funcs = { + .dpms = psb_intel_lvds_encoder_dpms, + .mode_fixup = psb_intel_lvds_mode_fixup, + .prepare = psb_intel_lvds_prepare, + .mode_set = psb_intel_lvds_mode_set, + .commit = psb_intel_lvds_commit, +}; + +const struct drm_connector_helper_funcs + psb_intel_lvds_connector_helper_funcs = { + .get_modes = psb_intel_lvds_get_modes, + .mode_valid = psb_intel_lvds_mode_valid, + .best_encoder = psb_intel_best_encoder, +}; + +const struct drm_connector_funcs psb_intel_lvds_connector_funcs = { + .dpms = drm_helper_connector_dpms, + .save = psb_intel_lvds_save, + .restore = psb_intel_lvds_restore, + .detect = psb_intel_lvds_detect, + .fill_modes = drm_helper_probe_single_connector_modes, + .set_property = psb_intel_lvds_set_property, + .destroy = psb_intel_lvds_destroy, +}; + + +static void psb_intel_lvds_enc_destroy(struct drm_encoder *encoder) +{ + drm_encoder_cleanup(encoder); +} + +const struct drm_encoder_funcs psb_intel_lvds_enc_funcs = { + .destroy = psb_intel_lvds_enc_destroy, +}; + + + +/** + * psb_intel_lvds_init - setup LVDS connectors on this device + * @dev: drm device + * + * Create the connector, register the LVDS DDC bus, and try to figure out what + * modes we can display on the LVDS panel (if present). + */ +void psb_intel_lvds_init(struct drm_device *dev, + struct psb_intel_mode_device *mode_dev) +{ + struct psb_intel_encoder *psb_intel_encoder; + struct psb_intel_connector *psb_intel_connector; + struct psb_intel_lvds_priv *lvds_priv; + struct drm_connector *connector; + struct drm_encoder *encoder; + struct drm_display_mode *scan; /* *modes, *bios_mode; */ + struct drm_crtc *crtc; + struct drm_psb_private *dev_priv = dev->dev_private; + u32 lvds; + int pipe; + + psb_intel_encoder = + kzalloc(sizeof(struct psb_intel_encoder), GFP_KERNEL); + if (!psb_intel_encoder) { + dev_err(dev->dev, "psb_intel_encoder allocation error\n"); + return; + } + + psb_intel_connector = + kzalloc(sizeof(struct psb_intel_connector), GFP_KERNEL); + if (!psb_intel_connector) { + dev_err(dev->dev, "psb_intel_connector allocation error\n"); + goto failed_encoder; + } + + lvds_priv = kzalloc(sizeof(struct psb_intel_lvds_priv), GFP_KERNEL); + if (!lvds_priv) { + dev_err(dev->dev, "LVDS private allocation error\n"); + goto failed_connector; + } + + psb_intel_encoder->dev_priv = lvds_priv; + + connector = &psb_intel_connector->base; + encoder = &psb_intel_encoder->base; + drm_connector_init(dev, connector, + &psb_intel_lvds_connector_funcs, + DRM_MODE_CONNECTOR_LVDS); + + drm_encoder_init(dev, encoder, + &psb_intel_lvds_enc_funcs, + DRM_MODE_ENCODER_LVDS); + + psb_intel_connector_attach_encoder(psb_intel_connector, + psb_intel_encoder); + psb_intel_encoder->type = INTEL_OUTPUT_LVDS; + + drm_encoder_helper_add(encoder, &psb_intel_lvds_helper_funcs); + drm_connector_helper_add(connector, + &psb_intel_lvds_connector_helper_funcs); + connector->display_info.subpixel_order = SubPixelHorizontalRGB; + connector->interlace_allowed = false; + connector->doublescan_allowed = false; + + /*Attach connector properties*/ + drm_connector_attach_property(connector, + dev->mode_config.scaling_mode_property, + DRM_MODE_SCALE_FULLSCREEN); + drm_connector_attach_property(connector, + dev_priv->backlight_property, + BRIGHTNESS_MAX_LEVEL); + + /* + * Set up I2C bus + * FIXME: distroy i2c_bus when exit + */ + lvds_priv->i2c_bus = psb_intel_i2c_create(dev, GPIOB, "LVDSBLC_B"); + if (!lvds_priv->i2c_bus) { + dev_printk(KERN_ERR, + &dev->pdev->dev, "I2C bus registration failed.\n"); + goto failed_blc_i2c; + } + lvds_priv->i2c_bus->slave_addr = 0x2C; + dev_priv->lvds_i2c_bus = lvds_priv->i2c_bus; + + /* + * LVDS discovery: + * 1) check for EDID on DDC + * 2) check for VBT data + * 3) check to see if LVDS is already on + * if none of the above, no panel + * 4) make sure lid is open + * if closed, act like it's not there for now + */ + + /* Set up the DDC bus. */ + lvds_priv->ddc_bus = psb_intel_i2c_create(dev, GPIOC, "LVDSDDC_C"); + if (!lvds_priv->ddc_bus) { + dev_printk(KERN_ERR, &dev->pdev->dev, + "DDC bus registration " "failed.\n"); + goto failed_ddc; + } + + /* + * Attempt to get the fixed panel mode from DDC. Assume that the + * preferred mode is the right one. + */ + psb_intel_ddc_get_modes(connector, &lvds_priv->ddc_bus->adapter); + list_for_each_entry(scan, &connector->probed_modes, head) { + if (scan->type & DRM_MODE_TYPE_PREFERRED) { + mode_dev->panel_fixed_mode = + drm_mode_duplicate(dev, scan); + goto out; /* FIXME: check for quirks */ + } + } + + /* Failed to get EDID, what about VBT? do we need this? */ + if (mode_dev->vbt_mode) + mode_dev->panel_fixed_mode = + drm_mode_duplicate(dev, mode_dev->vbt_mode); + + if (!mode_dev->panel_fixed_mode) + if (dev_priv->lfp_lvds_vbt_mode) + mode_dev->panel_fixed_mode = + drm_mode_duplicate(dev, + dev_priv->lfp_lvds_vbt_mode); + + /* + * If we didn't get EDID, try checking if the panel is already turned + * on. If so, assume that whatever is currently programmed is the + * correct mode. + */ + lvds = REG_READ(LVDS); + pipe = (lvds & LVDS_PIPEB_SELECT) ? 1 : 0; + crtc = psb_intel_get_crtc_from_pipe(dev, pipe); + + if (crtc && (lvds & LVDS_PORT_EN)) { + mode_dev->panel_fixed_mode = + psb_intel_crtc_mode_get(dev, crtc); + if (mode_dev->panel_fixed_mode) { + mode_dev->panel_fixed_mode->type |= + DRM_MODE_TYPE_PREFERRED; + goto out; /* FIXME: check for quirks */ + } + } + + /* If we still don't have a mode after all that, give up. */ + if (!mode_dev->panel_fixed_mode) { + dev_err(dev->dev, "Found no modes on the lvds, ignoring the LVDS\n"); + goto failed_find; + } + + /* + * Blacklist machines with BIOSes that list an LVDS panel without + * actually having one. + */ +out: + drm_sysfs_connector_add(connector); + return; + +failed_find: + if (lvds_priv->ddc_bus) + psb_intel_i2c_destroy(lvds_priv->ddc_bus); +failed_ddc: + if (lvds_priv->i2c_bus) + psb_intel_i2c_destroy(lvds_priv->i2c_bus); +failed_blc_i2c: + drm_encoder_cleanup(encoder); + drm_connector_cleanup(connector); +failed_connector: + kfree(psb_intel_connector); +failed_encoder: + kfree(psb_intel_encoder); +} + diff --git a/drivers/gpu/drm/gma500/psb_intel_modes.c b/drivers/gpu/drm/gma500/psb_intel_modes.c new file mode 100644 index 0000000..4fca0d6 --- /dev/null +++ b/drivers/gpu/drm/gma500/psb_intel_modes.c @@ -0,0 +1,75 @@ +/* + * Copyright (c) 2007 Intel Corporation + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + * Authers: Jesse Barnes + */ + +#include +#include +#include +#include "psb_intel_drv.h" + +/** + * psb_intel_ddc_probe + * + */ +bool psb_intel_ddc_probe(struct i2c_adapter *adapter) +{ + u8 out_buf[] = { 0x0, 0x0 }; + u8 buf[2]; + int ret; + struct i2c_msg msgs[] = { + { + .addr = 0x50, + .flags = 0, + .len = 1, + .buf = out_buf, + }, + { + .addr = 0x50, + .flags = I2C_M_RD, + .len = 1, + .buf = buf, + } + }; + + ret = i2c_transfer(adapter, msgs, 2); + if (ret == 2) + return true; + + return false; +} + +/** + * psb_intel_ddc_get_modes - get modelist from monitor + * @connector: DRM connector device to use + * + * Fetch the EDID information from @connector using the DDC bus. + */ +int psb_intel_ddc_get_modes(struct drm_connector *connector, + struct i2c_adapter *adapter) +{ + struct edid *edid; + int ret = 0; + + edid = drm_get_edid(connector, adapter); + if (edid) { + drm_mode_connector_update_edid_property(connector, edid); + ret = drm_add_edid_modes(connector, edid); + kfree(edid); + } + return ret; +} diff --git a/drivers/gpu/drm/gma500/psb_intel_reg.h b/drivers/gpu/drm/gma500/psb_intel_reg.h new file mode 100644 index 0000000..e89d3a2 --- /dev/null +++ b/drivers/gpu/drm/gma500/psb_intel_reg.h @@ -0,0 +1,1318 @@ +/* + * Copyright (c) 2009, Intel Corporation. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + */ +#ifndef __PSB_INTEL_REG_H__ +#define __PSB_INTEL_REG_H__ + +/* + * GPIO regs + */ +#define GPIOA 0x5010 +#define GPIOB 0x5014 +#define GPIOC 0x5018 +#define GPIOD 0x501c +#define GPIOE 0x5020 +#define GPIOF 0x5024 +#define GPIOG 0x5028 +#define GPIOH 0x502c +# define GPIO_CLOCK_DIR_MASK (1 << 0) +# define GPIO_CLOCK_DIR_IN (0 << 1) +# define GPIO_CLOCK_DIR_OUT (1 << 1) +# define GPIO_CLOCK_VAL_MASK (1 << 2) +# define GPIO_CLOCK_VAL_OUT (1 << 3) +# define GPIO_CLOCK_VAL_IN (1 << 4) +# define GPIO_CLOCK_PULLUP_DISABLE (1 << 5) +# define GPIO_DATA_DIR_MASK (1 << 8) +# define GPIO_DATA_DIR_IN (0 << 9) +# define GPIO_DATA_DIR_OUT (1 << 9) +# define GPIO_DATA_VAL_MASK (1 << 10) +# define GPIO_DATA_VAL_OUT (1 << 11) +# define GPIO_DATA_VAL_IN (1 << 12) +# define GPIO_DATA_PULLUP_DISABLE (1 << 13) + +#define GMBUS0 0x5100 /* clock/port select */ +#define GMBUS_RATE_100KHZ (0<<8) +#define GMBUS_RATE_50KHZ (1<<8) +#define GMBUS_RATE_400KHZ (2<<8) /* reserved on Pineview */ +#define GMBUS_RATE_1MHZ (3<<8) /* reserved on Pineview */ +#define GMBUS_HOLD_EXT (1<<7) /* 300ns hold time, rsvd on Pineview */ +#define GMBUS_PORT_DISABLED 0 +#define GMBUS_PORT_SSC 1 +#define GMBUS_PORT_VGADDC 2 +#define GMBUS_PORT_PANEL 3 +#define GMBUS_PORT_DPC 4 /* HDMIC */ +#define GMBUS_PORT_DPB 5 /* SDVO, HDMIB */ + /* 6 reserved */ +#define GMBUS_PORT_DPD 7 /* HDMID */ +#define GMBUS_NUM_PORTS 8 +#define GMBUS1 0x5104 /* command/status */ +#define GMBUS_SW_CLR_INT (1<<31) +#define GMBUS_SW_RDY (1<<30) +#define GMBUS_ENT (1<<29) /* enable timeout */ +#define GMBUS_CYCLE_NONE (0<<25) +#define GMBUS_CYCLE_WAIT (1<<25) +#define GMBUS_CYCLE_INDEX (2<<25) +#define GMBUS_CYCLE_STOP (4<<25) +#define GMBUS_BYTE_COUNT_SHIFT 16 +#define GMBUS_SLAVE_INDEX_SHIFT 8 +#define GMBUS_SLAVE_ADDR_SHIFT 1 +#define GMBUS_SLAVE_READ (1<<0) +#define GMBUS_SLAVE_WRITE (0<<0) +#define GMBUS2 0x5108 /* status */ +#define GMBUS_INUSE (1<<15) +#define GMBUS_HW_WAIT_PHASE (1<<14) +#define GMBUS_STALL_TIMEOUT (1<<13) +#define GMBUS_INT (1<<12) +#define GMBUS_HW_RDY (1<<11) +#define GMBUS_SATOER (1<<10) +#define GMBUS_ACTIVE (1<<9) +#define GMBUS3 0x510c /* data buffer bytes 3-0 */ +#define GMBUS4 0x5110 /* interrupt mask (Pineview+) */ +#define GMBUS_SLAVE_TIMEOUT_EN (1<<4) +#define GMBUS_NAK_EN (1<<3) +#define GMBUS_IDLE_EN (1<<2) +#define GMBUS_HW_WAIT_EN (1<<1) +#define GMBUS_HW_RDY_EN (1<<0) +#define GMBUS5 0x5120 /* byte index */ +#define GMBUS_2BYTE_INDEX_EN (1<<31) + +#define BLC_PWM_CTL 0x61254 +#define BLC_PWM_CTL2 0x61250 +#define BLC_PWM_CTL_C 0x62254 +#define BLC_PWM_CTL2_C 0x62250 +#define BACKLIGHT_MODULATION_FREQ_SHIFT (17) +/* + * This is the most significant 15 bits of the number of backlight cycles in a + * complete cycle of the modulated backlight control. + * + * The actual value is this field multiplied by two. + */ +#define BACKLIGHT_MODULATION_FREQ_MASK (0x7fff << 17) +#define BLM_LEGACY_MODE (1 << 16) +/* + * This is the number of cycles out of the backlight modulation cycle for which + * the backlight is on. + * + * This field must be no greater than the number of cycles in the complete + * backlight modulation cycle. + */ +#define BACKLIGHT_DUTY_CYCLE_SHIFT (0) +#define BACKLIGHT_DUTY_CYCLE_MASK (0xffff) + +#define I915_GCFGC 0xf0 +#define I915_LOW_FREQUENCY_ENABLE (1 << 7) +#define I915_DISPLAY_CLOCK_190_200_MHZ (0 << 4) +#define I915_DISPLAY_CLOCK_333_MHZ (4 << 4) +#define I915_DISPLAY_CLOCK_MASK (7 << 4) + +#define I855_HPLLCC 0xc0 +#define I855_CLOCK_CONTROL_MASK (3 << 0) +#define I855_CLOCK_133_200 (0 << 0) +#define I855_CLOCK_100_200 (1 << 0) +#define I855_CLOCK_100_133 (2 << 0) +#define I855_CLOCK_166_250 (3 << 0) + +/* I830 CRTC registers */ +#define HTOTAL_A 0x60000 +#define HBLANK_A 0x60004 +#define HSYNC_A 0x60008 +#define VTOTAL_A 0x6000c +#define VBLANK_A 0x60010 +#define VSYNC_A 0x60014 +#define PIPEASRC 0x6001c +#define BCLRPAT_A 0x60020 +#define VSYNCSHIFT_A 0x60028 + +#define HTOTAL_B 0x61000 +#define HBLANK_B 0x61004 +#define HSYNC_B 0x61008 +#define VTOTAL_B 0x6100c +#define VBLANK_B 0x61010 +#define VSYNC_B 0x61014 +#define PIPEBSRC 0x6101c +#define BCLRPAT_B 0x61020 +#define VSYNCSHIFT_B 0x61028 + +#define HTOTAL_C 0x62000 +#define HBLANK_C 0x62004 +#define HSYNC_C 0x62008 +#define VTOTAL_C 0x6200c +#define VBLANK_C 0x62010 +#define VSYNC_C 0x62014 +#define PIPECSRC 0x6201c +#define BCLRPAT_C 0x62020 +#define VSYNCSHIFT_C 0x62028 + +#define PP_STATUS 0x61200 +# define PP_ON (1 << 31) +/* + * Indicates that all dependencies of the panel are on: + * + * - PLL enabled + * - pipe enabled + * - LVDS/DVOB/DVOC on + */ +#define PP_READY (1 << 30) +#define PP_SEQUENCE_NONE (0 << 28) +#define PP_SEQUENCE_ON (1 << 28) +#define PP_SEQUENCE_OFF (2 << 28) +#define PP_SEQUENCE_MASK 0x30000000 +#define PP_CONTROL 0x61204 +#define POWER_TARGET_ON (1 << 0) + +#define LVDSPP_ON 0x61208 +#define LVDSPP_OFF 0x6120c +#define PP_CYCLE 0x61210 + +#define PP_ON_DELAYS 0x61208 /* Cedartrail */ +#define PP_OFF_DELAYS 0x6120c /* Cedartrail */ + +#define PFIT_CONTROL 0x61230 +#define PFIT_ENABLE (1 << 31) +#define PFIT_PIPE_MASK (3 << 29) +#define PFIT_PIPE_SHIFT 29 +#define PFIT_SCALING_MODE_PILLARBOX (1 << 27) +#define PFIT_SCALING_MODE_LETTERBOX (3 << 26) +#define VERT_INTERP_DISABLE (0 << 10) +#define VERT_INTERP_BILINEAR (1 << 10) +#define VERT_INTERP_MASK (3 << 10) +#define VERT_AUTO_SCALE (1 << 9) +#define HORIZ_INTERP_DISABLE (0 << 6) +#define HORIZ_INTERP_BILINEAR (1 << 6) +#define HORIZ_INTERP_MASK (3 << 6) +#define HORIZ_AUTO_SCALE (1 << 5) +#define PANEL_8TO6_DITHER_ENABLE (1 << 3) + +#define PFIT_PGM_RATIOS 0x61234 +#define PFIT_VERT_SCALE_MASK 0xfff00000 +#define PFIT_HORIZ_SCALE_MASK 0x0000fff0 + +#define PFIT_AUTO_RATIOS 0x61238 + +#define DPLL_A 0x06014 +#define DPLL_B 0x06018 +#define DPLL_VCO_ENABLE (1 << 31) +#define DPLL_DVO_HIGH_SPEED (1 << 30) +#define DPLL_SYNCLOCK_ENABLE (1 << 29) +#define DPLL_VGA_MODE_DIS (1 << 28) +#define DPLLB_MODE_DAC_SERIAL (1 << 26) /* i915 */ +#define DPLLB_MODE_LVDS (2 << 26) /* i915 */ +#define DPLL_MODE_MASK (3 << 26) +#define DPLL_DAC_SERIAL_P2_CLOCK_DIV_10 (0 << 24) /* i915 */ +#define DPLL_DAC_SERIAL_P2_CLOCK_DIV_5 (1 << 24) /* i915 */ +#define DPLLB_LVDS_P2_CLOCK_DIV_14 (0 << 24) /* i915 */ +#define DPLLB_LVDS_P2_CLOCK_DIV_7 (1 << 24) /* i915 */ +#define DPLL_P2_CLOCK_DIV_MASK 0x03000000 /* i915 */ +#define DPLL_FPA01_P1_POST_DIV_MASK 0x00ff0000 /* i915 */ +#define DPLL_LOCK (1 << 15) /* CDV */ + +/* + * The i830 generation, in DAC/serial mode, defines p1 as two plus this + * bitfield, or just 2 if PLL_P1_DIVIDE_BY_TWO is set. + */ +# define DPLL_FPA01_P1_POST_DIV_MASK_I830 0x001f0000 +/* + * The i830 generation, in LVDS mode, defines P1 as the bit number set within + * this field (only one bit may be set). + */ +#define DPLL_FPA01_P1_POST_DIV_MASK_I830_LVDS 0x003f0000 +#define DPLL_FPA01_P1_POST_DIV_SHIFT 16 +#define PLL_P2_DIVIDE_BY_4 (1 << 23) /* i830, required + * in DVO non-gang */ +# define PLL_P1_DIVIDE_BY_TWO (1 << 21) /* i830 */ +#define PLL_REF_INPUT_DREFCLK (0 << 13) +#define PLL_REF_INPUT_TVCLKINA (1 << 13) /* i830 */ +#define PLL_REF_INPUT_TVCLKINBC (2 << 13) /* SDVO + * TVCLKIN */ +#define PLLB_REF_INPUT_SPREADSPECTRUMIN (3 << 13) +#define PLL_REF_INPUT_MASK (3 << 13) +#define PLL_LOAD_PULSE_PHASE_SHIFT 9 +/* + * Parallel to Serial Load Pulse phase selection. + * Selects the phase for the 10X DPLL clock for the PCIe + * digital display port. The range is 4 to 13; 10 or more + * is just a flip delay. The default is 6 + */ +#define PLL_LOAD_PULSE_PHASE_MASK (0xf << PLL_LOAD_PULSE_PHASE_SHIFT) +#define DISPLAY_RATE_SELECT_FPA1 (1 << 8) + +/* + * SDVO multiplier for 945G/GM. Not used on 965. + * + * DPLL_MD_UDI_MULTIPLIER_MASK + */ +#define SDVO_MULTIPLIER_MASK 0x000000ff +#define SDVO_MULTIPLIER_SHIFT_HIRES 4 +#define SDVO_MULTIPLIER_SHIFT_VGA 0 + +/* + * PLL_MD + */ +/* Pipe A SDVO/UDI clock multiplier/divider register for G965. */ +#define DPLL_A_MD 0x0601c +/* Pipe B SDVO/UDI clock multiplier/divider register for G965. */ +#define DPLL_B_MD 0x06020 +/* + * UDI pixel divider, controlling how many pixels are stuffed into a packet. + * + * Value is pixels minus 1. Must be set to 1 pixel for SDVO. + */ +#define DPLL_MD_UDI_DIVIDER_MASK 0x3f000000 +#define DPLL_MD_UDI_DIVIDER_SHIFT 24 +/* UDI pixel divider for VGA, same as DPLL_MD_UDI_DIVIDER_MASK. */ +#define DPLL_MD_VGA_UDI_DIVIDER_MASK 0x003f0000 +#define DPLL_MD_VGA_UDI_DIVIDER_SHIFT 16 +/* + * SDVO/UDI pixel multiplier. + * + * SDVO requires that the bus clock rate be between 1 and 2 Ghz, and the bus + * clock rate is 10 times the DPLL clock. At low resolution/refresh rate + * modes, the bus rate would be below the limits, so SDVO allows for stuffing + * dummy bytes in the datastream at an increased clock rate, with both sides of + * the link knowing how many bytes are fill. + * + * So, for a mode with a dotclock of 65Mhz, we would want to double the clock + * rate to 130Mhz to get a bus rate of 1.30Ghz. The DPLL clock rate would be + * set to 130Mhz, and the SDVO multiplier set to 2x in this register and + * through an SDVO command. + * + * This register field has values of multiplication factor minus 1, with + * a maximum multiplier of 5 for SDVO. + */ +#define DPLL_MD_UDI_MULTIPLIER_MASK 0x00003f00 +#define DPLL_MD_UDI_MULTIPLIER_SHIFT 8 +/* + * SDVO/UDI pixel multiplier for VGA, same as DPLL_MD_UDI_MULTIPLIER_MASK. + * This best be set to the default value (3) or the CRT won't work. No, + * I don't entirely understand what this does... + */ +#define DPLL_MD_VGA_UDI_MULTIPLIER_MASK 0x0000003f +#define DPLL_MD_VGA_UDI_MULTIPLIER_SHIFT 0 + +#define DPLL_TEST 0x606c +#define DPLLB_TEST_SDVO_DIV_1 (0 << 22) +#define DPLLB_TEST_SDVO_DIV_2 (1 << 22) +#define DPLLB_TEST_SDVO_DIV_4 (2 << 22) +#define DPLLB_TEST_SDVO_DIV_MASK (3 << 22) +#define DPLLB_TEST_N_BYPASS (1 << 19) +#define DPLLB_TEST_M_BYPASS (1 << 18) +#define DPLLB_INPUT_BUFFER_ENABLE (1 << 16) +#define DPLLA_TEST_N_BYPASS (1 << 3) +#define DPLLA_TEST_M_BYPASS (1 << 2) +#define DPLLA_INPUT_BUFFER_ENABLE (1 << 0) + +#define ADPA 0x61100 +#define ADPA_DAC_ENABLE (1 << 31) +#define ADPA_DAC_DISABLE 0 +#define ADPA_PIPE_SELECT_MASK (1 << 30) +#define ADPA_PIPE_A_SELECT 0 +#define ADPA_PIPE_B_SELECT (1 << 30) +#define ADPA_USE_VGA_HVPOLARITY (1 << 15) +#define ADPA_SETS_HVPOLARITY 0 +#define ADPA_VSYNC_CNTL_DISABLE (1 << 11) +#define ADPA_VSYNC_CNTL_ENABLE 0 +#define ADPA_HSYNC_CNTL_DISABLE (1 << 10) +#define ADPA_HSYNC_CNTL_ENABLE 0 +#define ADPA_VSYNC_ACTIVE_HIGH (1 << 4) +#define ADPA_VSYNC_ACTIVE_LOW 0 +#define ADPA_HSYNC_ACTIVE_HIGH (1 << 3) +#define ADPA_HSYNC_ACTIVE_LOW 0 + +#define FPA0 0x06040 +#define FPA1 0x06044 +#define FPB0 0x06048 +#define FPB1 0x0604c +#define FP_N_DIV_MASK 0x003f0000 +#define FP_N_DIV_SHIFT 16 +#define FP_M1_DIV_MASK 0x00003f00 +#define FP_M1_DIV_SHIFT 8 +#define FP_M2_DIV_MASK 0x0000003f +#define FP_M2_DIV_SHIFT 0 + +#define PORT_HOTPLUG_EN 0x61110 +#define SDVOB_HOTPLUG_INT_EN (1 << 26) +#define SDVOC_HOTPLUG_INT_EN (1 << 25) +#define TV_HOTPLUG_INT_EN (1 << 18) +#define CRT_HOTPLUG_INT_EN (1 << 9) +#define CRT_HOTPLUG_FORCE_DETECT (1 << 3) +/* CDV.. */ +#define CRT_HOTPLUG_ACTIVATION_PERIOD_64 (1 << 8) +#define CRT_HOTPLUG_DAC_ON_TIME_2M (0 << 7) +#define CRT_HOTPLUG_DAC_ON_TIME_4M (1 << 7) +#define CRT_HOTPLUG_VOLTAGE_COMPARE_40 (0 << 5) +#define CRT_HOTPLUG_VOLTAGE_COMPARE_50 (1 << 5) +#define CRT_HOTPLUG_VOLTAGE_COMPARE_60 (2 << 5) +#define CRT_HOTPLUG_VOLTAGE_COMPARE_70 (3 << 5) +#define CRT_HOTPLUG_VOLTAGE_COMPARE_MASK (3 << 5) +#define CRT_HOTPLUG_DETECT_DELAY_1G (0 << 4) +#define CRT_HOTPLUG_DETECT_DELAY_2G (1 << 4) +#define CRT_HOTPLUG_DETECT_VOLTAGE_325MV (0 << 2) +#define CRT_HOTPLUG_DETECT_VOLTAGE_475MV (1 << 2) +#define CRT_HOTPLUG_DETECT_MASK 0x000000F8 + +#define PORT_HOTPLUG_STAT 0x61114 +#define CRT_HOTPLUG_INT_STATUS (1 << 11) +#define TV_HOTPLUG_INT_STATUS (1 << 10) +#define CRT_HOTPLUG_MONITOR_MASK (3 << 8) +#define CRT_HOTPLUG_MONITOR_COLOR (3 << 8) +#define CRT_HOTPLUG_MONITOR_MONO (2 << 8) +#define CRT_HOTPLUG_MONITOR_NONE (0 << 8) +#define SDVOC_HOTPLUG_INT_STATUS (1 << 7) +#define SDVOB_HOTPLUG_INT_STATUS (1 << 6) + +#define SDVOB 0x61140 +#define SDVOC 0x61160 +#define SDVO_ENABLE (1 << 31) +#define SDVO_PIPE_B_SELECT (1 << 30) +#define SDVO_STALL_SELECT (1 << 29) +#define SDVO_INTERRUPT_ENABLE (1 << 26) +#define SDVO_COLOR_RANGE_16_235 (1 << 8) +#define SDVO_AUDIO_ENABLE (1 << 6) + +/** + * 915G/GM SDVO pixel multiplier. + * + * Programmed value is multiplier - 1, up to 5x. + * + * DPLL_MD_UDI_MULTIPLIER_MASK + */ +#define SDVO_PORT_MULTIPLY_MASK (7 << 23) +#define SDVO_PORT_MULTIPLY_SHIFT 23 +#define SDVO_PHASE_SELECT_MASK (15 << 19) +#define SDVO_PHASE_SELECT_DEFAULT (6 << 19) +#define SDVO_CLOCK_OUTPUT_INVERT (1 << 18) +#define SDVOC_GANG_MODE (1 << 16) +#define SDVO_BORDER_ENABLE (1 << 7) +#define SDVOB_PCIE_CONCURRENCY (1 << 3) +#define SDVO_DETECTED (1 << 2) +/* Bits to be preserved when writing */ +#define SDVOB_PRESERVE_MASK ((1 << 17) | (1 << 16) | (1 << 14)) +#define SDVOC_PRESERVE_MASK (1 << 17) + +/* + * This register controls the LVDS output enable, pipe selection, and data + * format selection. + * + * All of the clock/data pairs are force powered down by power sequencing. + */ +#define LVDS 0x61180 +/* + * Enables the LVDS port. This bit must be set before DPLLs are enabled, as + * the DPLL semantics change when the LVDS is assigned to that pipe. + */ +#define LVDS_PORT_EN (1 << 31) +/* Selects pipe B for LVDS data. Must be set on pre-965. */ +#define LVDS_PIPEB_SELECT (1 << 30) + +/* Turns on border drawing to allow centered display. */ +#define LVDS_BORDER_EN (1 << 15) + +/* + * Enables the A0-A2 data pairs and CLKA, containing 18 bits of color data per + * pixel. + */ +#define LVDS_A0A2_CLKA_POWER_MASK (3 << 8) +#define LVDS_A0A2_CLKA_POWER_DOWN (0 << 8) +#define LVDS_A0A2_CLKA_POWER_UP (3 << 8) +/* + * Controls the A3 data pair, which contains the additional LSBs for 24 bit + * mode. Only enabled if LVDS_A0A2_CLKA_POWER_UP also indicates it should be + * on. + */ +#define LVDS_A3_POWER_MASK (3 << 6) +#define LVDS_A3_POWER_DOWN (0 << 6) +#define LVDS_A3_POWER_UP (3 << 6) +/* + * Controls the CLKB pair. This should only be set when LVDS_B0B3_POWER_UP + * is set. + */ +#define LVDS_CLKB_POWER_MASK (3 << 4) +#define LVDS_CLKB_POWER_DOWN (0 << 4) +#define LVDS_CLKB_POWER_UP (3 << 4) +/* + * Controls the B0-B3 data pairs. This must be set to match the DPLL p2 + * setting for whether we are in dual-channel mode. The B3 pair will + * additionally only be powered up when LVDS_A3_POWER_UP is set. + */ +#define LVDS_B0B3_POWER_MASK (3 << 2) +#define LVDS_B0B3_POWER_DOWN (0 << 2) +#define LVDS_B0B3_POWER_UP (3 << 2) + +#define PIPEACONF 0x70008 +#define PIPEACONF_ENABLE (1 << 31) +#define PIPEACONF_DISABLE 0 +#define PIPEACONF_DOUBLE_WIDE (1 << 30) +#define PIPECONF_ACTIVE (1 << 30) +#define I965_PIPECONF_ACTIVE (1 << 30) +#define PIPECONF_DSIPLL_LOCK (1 << 29) +#define PIPEACONF_SINGLE_WIDE 0 +#define PIPEACONF_PIPE_UNLOCKED 0 +#define PIPEACONF_DSR (1 << 26) +#define PIPEACONF_PIPE_LOCKED (1 << 25) +#define PIPEACONF_PALETTE 0 +#define PIPECONF_FORCE_BORDER (1 << 25) +#define PIPEACONF_GAMMA (1 << 24) +#define PIPECONF_PROGRESSIVE (0 << 21) +#define PIPECONF_INTERLACE_W_FIELD_INDICATION (6 << 21) +#define PIPECONF_INTERLACE_FIELD_0_ONLY (7 << 21) +#define PIPECONF_PLANE_OFF (1 << 19) +#define PIPECONF_CURSOR_OFF (1 << 18) + +#define PIPEBCONF 0x71008 +#define PIPEBCONF_ENABLE (1 << 31) +#define PIPEBCONF_DISABLE 0 +#define PIPEBCONF_DOUBLE_WIDE (1 << 30) +#define PIPEBCONF_DISABLE 0 +#define PIPEBCONF_GAMMA (1 << 24) +#define PIPEBCONF_PALETTE 0 + +#define PIPECCONF 0x72008 + +#define PIPEBGCMAXRED 0x71010 +#define PIPEBGCMAXGREEN 0x71014 +#define PIPEBGCMAXBLUE 0x71018 + +#define PIPEASTAT 0x70024 +#define PIPEBSTAT 0x71024 +#define PIPECSTAT 0x72024 +#define PIPE_VBLANK_INTERRUPT_STATUS (1UL << 1) +#define PIPE_START_VBLANK_INTERRUPT_STATUS (1UL << 2) +#define PIPE_VBLANK_CLEAR (1 << 1) +#define PIPE_VBLANK_STATUS (1 << 1) +#define PIPE_TE_STATUS (1UL << 6) +#define PIPE_DPST_EVENT_STATUS (1UL << 7) +#define PIPE_VSYNC_CLEAR (1UL << 9) +#define PIPE_VSYNC_STATUS (1UL << 9) +#define PIPE_HDMI_AUDIO_UNDERRUN_STATUS (1UL << 10) +#define PIPE_HDMI_AUDIO_BUFFER_DONE_STATUS (1UL << 11) +#define PIPE_VBLANK_INTERRUPT_ENABLE (1UL << 17) +#define PIPE_START_VBLANK_INTERRUPT_ENABLE (1UL << 18) +#define PIPE_TE_ENABLE (1UL << 22) +#define PIPE_DPST_EVENT_ENABLE (1UL << 23) +#define PIPE_VSYNC_ENABL (1UL << 25) +#define PIPE_HDMI_AUDIO_UNDERRUN (1UL << 26) +#define PIPE_HDMI_AUDIO_BUFFER_DONE (1UL << 27) +#define PIPE_HDMI_AUDIO_INT_MASK (PIPE_HDMI_AUDIO_UNDERRUN | \ + PIPE_HDMI_AUDIO_BUFFER_DONE) +#define PIPE_EVENT_MASK ((1 << 29)|(1 << 28)|(1 << 27)|(1 << 26)|(1 << 24)|(1 << 23)|(1 << 22)|(1 << 21)|(1 << 20)|(1 << 16)) +#define PIPE_VBLANK_MASK ((1 << 25)|(1 << 24)|(1 << 18)|(1 << 17)) +#define HISTOGRAM_INT_CONTROL 0x61268 +#define HISTOGRAM_BIN_DATA 0X61264 +#define HISTOGRAM_LOGIC_CONTROL 0x61260 +#define PWM_CONTROL_LOGIC 0x61250 +#define PIPE_HOTPLUG_INTERRUPT_STATUS (1UL << 10) +#define HISTOGRAM_INTERRUPT_ENABLE (1UL << 31) +#define HISTOGRAM_LOGIC_ENABLE (1UL << 31) +#define PWM_LOGIC_ENABLE (1UL << 31) +#define PWM_PHASEIN_ENABLE (1UL << 25) +#define PWM_PHASEIN_INT_ENABLE (1UL << 24) +#define PWM_PHASEIN_VB_COUNT 0x00001f00 +#define PWM_PHASEIN_INC 0x0000001f +#define HISTOGRAM_INT_CTRL_CLEAR (1UL << 30) +#define DPST_YUV_LUMA_MODE 0 + +struct dpst_ie_histogram_control { + union { + uint32_t data; + struct { + uint32_t bin_reg_index:7; + uint32_t reserved:4; + uint32_t bin_reg_func_select:1; + uint32_t sync_to_phase_in:1; + uint32_t alt_enhancement_mode:2; + uint32_t reserved1:1; + uint32_t sync_to_phase_in_count:8; + uint32_t histogram_mode_select:1; + uint32_t reserved2:4; + uint32_t ie_pipe_assignment:1; + uint32_t ie_mode_table_enabled:1; + uint32_t ie_histogram_enable:1; + }; + }; +}; + +struct dpst_guardband { + union { + uint32_t data; + struct { + uint32_t guardband:22; + uint32_t guardband_interrupt_delay:8; + uint32_t interrupt_status:1; + uint32_t interrupt_enable:1; + }; + }; +}; + +#define PIPEAFRAMEHIGH 0x70040 +#define PIPEAFRAMEPIXEL 0x70044 +#define PIPEBFRAMEHIGH 0x71040 +#define PIPEBFRAMEPIXEL 0x71044 +#define PIPECFRAMEHIGH 0x72040 +#define PIPECFRAMEPIXEL 0x72044 +#define PIPE_FRAME_HIGH_MASK 0x0000ffff +#define PIPE_FRAME_HIGH_SHIFT 0 +#define PIPE_FRAME_LOW_MASK 0xff000000 +#define PIPE_FRAME_LOW_SHIFT 24 +#define PIPE_PIXEL_MASK 0x00ffffff +#define PIPE_PIXEL_SHIFT 0 + +#define DSPARB 0x70030 +#define DSPFW1 0x70034 +#define DSPFW2 0x70038 +#define DSPFW3 0x7003c +#define DSPFW4 0x70050 +#define DSPFW5 0x70054 +#define DSPFW6 0x70058 +#define DSPCHICKENBIT 0x70400 +#define DSPACNTR 0x70180 +#define DSPBCNTR 0x71180 +#define DSPCCNTR 0x72180 +#define DISPLAY_PLANE_ENABLE (1 << 31) +#define DISPLAY_PLANE_DISABLE 0 +#define DISPPLANE_GAMMA_ENABLE (1 << 30) +#define DISPPLANE_GAMMA_DISABLE 0 +#define DISPPLANE_PIXFORMAT_MASK (0xf << 26) +#define DISPPLANE_8BPP (0x2 << 26) +#define DISPPLANE_15_16BPP (0x4 << 26) +#define DISPPLANE_16BPP (0x5 << 26) +#define DISPPLANE_32BPP_NO_ALPHA (0x6 << 26) +#define DISPPLANE_32BPP (0x7 << 26) +#define DISPPLANE_STEREO_ENABLE (1 << 25) +#define DISPPLANE_STEREO_DISABLE 0 +#define DISPPLANE_SEL_PIPE_MASK (1 << 24) +#define DISPPLANE_SEL_PIPE_POS 24 +#define DISPPLANE_SEL_PIPE_A 0 +#define DISPPLANE_SEL_PIPE_B (1 << 24) +#define DISPPLANE_SRC_KEY_ENABLE (1 << 22) +#define DISPPLANE_SRC_KEY_DISABLE 0 +#define DISPPLANE_LINE_DOUBLE (1 << 20) +#define DISPPLANE_NO_LINE_DOUBLE 0 +#define DISPPLANE_STEREO_POLARITY_FIRST 0 +#define DISPPLANE_STEREO_POLARITY_SECOND (1 << 18) +/* plane B only */ +#define DISPPLANE_ALPHA_TRANS_ENABLE (1 << 15) +#define DISPPLANE_ALPHA_TRANS_DISABLE 0 +#define DISPPLANE_SPRITE_ABOVE_DISPLAYA 0 +#define DISPPLANE_SPRITE_ABOVE_OVERLAY (1) +#define DISPPLANE_BOTTOM (4) + +#define DSPABASE 0x70184 +#define DSPALINOFF 0x70184 +#define DSPASTRIDE 0x70188 + +#define DSPBBASE 0x71184 +#define DSPBLINOFF 0X71184 +#define DSPBADDR DSPBBASE +#define DSPBSTRIDE 0x71188 + +#define DSPCBASE 0x72184 +#define DSPCLINOFF 0x72184 +#define DSPCSTRIDE 0x72188 + +#define DSPAKEYVAL 0x70194 +#define DSPAKEYMASK 0x70198 + +#define DSPAPOS 0x7018C /* reserved */ +#define DSPASIZE 0x70190 +#define DSPBPOS 0x7118C +#define DSPBSIZE 0x71190 +#define DSPCPOS 0x7218C +#define DSPCSIZE 0x72190 + +#define DSPASURF 0x7019C +#define DSPATILEOFF 0x701A4 + +#define DSPBSURF 0x7119C +#define DSPBTILEOFF 0x711A4 + +#define DSPCSURF 0x7219C +#define DSPCTILEOFF 0x721A4 +#define DSPCKEYMAXVAL 0x721A0 +#define DSPCKEYMINVAL 0x72194 +#define DSPCKEYMSK 0x72198 + +#define VGACNTRL 0x71400 +#define VGA_DISP_DISABLE (1 << 31) +#define VGA_2X_MODE (1 << 30) +#define VGA_PIPE_B_SELECT (1 << 29) + +/* + * Overlay registers + */ +#define OV_C_OFFSET 0x08000 +#define OV_OVADD 0x30000 +#define OV_DOVASTA 0x30008 +# define OV_PIPE_SELECT ((1 << 6)|(1 << 7)) +# define OV_PIPE_SELECT_POS 6 +# define OV_PIPE_A 0 +# define OV_PIPE_C 1 +#define OV_OGAMC5 0x30010 +#define OV_OGAMC4 0x30014 +#define OV_OGAMC3 0x30018 +#define OV_OGAMC2 0x3001C +#define OV_OGAMC1 0x30020 +#define OV_OGAMC0 0x30024 +#define OVC_OVADD 0x38000 +#define OVC_DOVCSTA 0x38008 +#define OVC_OGAMC5 0x38010 +#define OVC_OGAMC4 0x38014 +#define OVC_OGAMC3 0x38018 +#define OVC_OGAMC2 0x3801C +#define OVC_OGAMC1 0x38020 +#define OVC_OGAMC0 0x38024 + +/* + * Some BIOS scratch area registers. The 845 (and 830?) store the amount + * of video memory available to the BIOS in SWF1. + */ +#define SWF0 0x71410 +#define SWF1 0x71414 +#define SWF2 0x71418 +#define SWF3 0x7141c +#define SWF4 0x71420 +#define SWF5 0x71424 +#define SWF6 0x71428 + +/* + * 855 scratch registers. + */ +#define SWF00 0x70410 +#define SWF01 0x70414 +#define SWF02 0x70418 +#define SWF03 0x7041c +#define SWF04 0x70420 +#define SWF05 0x70424 +#define SWF06 0x70428 + +#define SWF10 SWF0 +#define SWF11 SWF1 +#define SWF12 SWF2 +#define SWF13 SWF3 +#define SWF14 SWF4 +#define SWF15 SWF5 +#define SWF16 SWF6 + +#define SWF30 0x72414 +#define SWF31 0x72418 +#define SWF32 0x7241c + + +/* + * Palette registers + */ +#define PALETTE_A 0x0a000 +#define PALETTE_B 0x0a800 +#define PALETTE_C 0x0ac00 + +/* Cursor A & B regs */ +#define CURACNTR 0x70080 +#define CURSOR_MODE_DISABLE 0x00 +#define CURSOR_MODE_64_32B_AX 0x07 +#define CURSOR_MODE_64_ARGB_AX ((1 << 5) | CURSOR_MODE_64_32B_AX) +#define MCURSOR_GAMMA_ENABLE (1 << 26) +#define CURABASE 0x70084 +#define CURAPOS 0x70088 +#define CURSOR_POS_MASK 0x007FF +#define CURSOR_POS_SIGN 0x8000 +#define CURSOR_X_SHIFT 0 +#define CURSOR_Y_SHIFT 16 +#define CURBCNTR 0x700c0 +#define CURBBASE 0x700c4 +#define CURBPOS 0x700c8 +#define CURCCNTR 0x700e0 +#define CURCBASE 0x700e4 +#define CURCPOS 0x700e8 + +/* + * Interrupt Registers + */ +#define IER 0x020a0 +#define IIR 0x020a4 +#define IMR 0x020a8 +#define ISR 0x020ac + +/* + * MOORESTOWN delta registers + */ +#define MRST_DPLL_A 0x0f014 +#define MDFLD_DPLL_B 0x0f018 +#define MDFLD_INPUT_REF_SEL (1 << 14) +#define MDFLD_VCO_SEL (1 << 16) +#define DPLLA_MODE_LVDS (2 << 26) /* mrst */ +#define MDFLD_PLL_LATCHEN (1 << 28) +#define MDFLD_PWR_GATE_EN (1 << 30) +#define MDFLD_P1_MASK (0x1FF << 17) +#define MRST_FPA0 0x0f040 +#define MRST_FPA1 0x0f044 +#define MDFLD_DPLL_DIV0 0x0f048 +#define MDFLD_DPLL_DIV1 0x0f04c +#define MRST_PERF_MODE 0x020f4 + +/* + * MEDFIELD HDMI registers + */ +#define HDMIPHYMISCCTL 0x61134 +#define HDMI_PHY_POWER_DOWN 0x7f +#define HDMIB_CONTROL 0x61140 +#define HDMIB_PORT_EN (1 << 31) +#define HDMIB_PIPE_B_SELECT (1 << 30) +#define HDMIB_NULL_PACKET (1 << 9) +#define HDMIB_HDCP_PORT (1 << 5) + +/* #define LVDS 0x61180 */ +#define MRST_PANEL_8TO6_DITHER_ENABLE (1 << 25) +#define MRST_PANEL_24_DOT_1_FORMAT (1 << 24) +#define LVDS_A3_POWER_UP_0_OUTPUT (1 << 6) + +#define MIPI 0x61190 +#define MIPI_C 0x62190 +#define MIPI_PORT_EN (1 << 31) +/* Turns on border drawing to allow centered display. */ +#define SEL_FLOPPED_HSTX (1 << 23) +#define PASS_FROM_SPHY_TO_AFE (1 << 16) +#define MIPI_BORDER_EN (1 << 15) +#define MIPIA_3LANE_MIPIC_1LANE 0x1 +#define MIPIA_2LANE_MIPIC_2LANE 0x2 +#define TE_TRIGGER_DSI_PROTOCOL (1 << 2) +#define TE_TRIGGER_GPIO_PIN (1 << 3) +#define MIPI_TE_COUNT 0x61194 + +/* #define PP_CONTROL 0x61204 */ +#define POWER_DOWN_ON_RESET (1 << 1) + +/* #define PFIT_CONTROL 0x61230 */ +#define PFIT_PIPE_SELECT (3 << 29) +#define PFIT_PIPE_SELECT_SHIFT (29) + +/* #define BLC_PWM_CTL 0x61254 */ +#define MRST_BACKLIGHT_MODULATION_FREQ_SHIFT (16) +#define MRST_BACKLIGHT_MODULATION_FREQ_MASK (0xffff << 16) + +/* #define PIPEACONF 0x70008 */ +#define PIPEACONF_PIPE_STATE (1 << 30) +/* #define DSPACNTR 0x70180 */ + +#define MRST_DSPABASE 0x7019c +#define MRST_DSPBBASE 0x7119c +#define MDFLD_DSPCBASE 0x7219c + +/* + * Moorestown registers. + */ + +/* + * MIPI IP registers + */ +#define MIPIC_REG_OFFSET 0x800 + +#define DEVICE_READY_REG 0xb000 +#define LP_OUTPUT_HOLD (1 << 16) +#define EXIT_ULPS_DEV_READY 0x3 +#define LP_OUTPUT_HOLD_RELEASE 0x810000 +# define ENTERING_ULPS (2 << 1) +# define EXITING_ULPS (1 << 1) +# define ULPS_MASK (3 << 1) +# define BUS_POSSESSION (1 << 3) +#define INTR_STAT_REG 0xb004 +#define RX_SOT_ERROR (1 << 0) +#define RX_SOT_SYNC_ERROR (1 << 1) +#define RX_ESCAPE_MODE_ENTRY_ERROR (1 << 3) +#define RX_LP_TX_SYNC_ERROR (1 << 4) +#define RX_HS_RECEIVE_TIMEOUT_ERROR (1 << 5) +#define RX_FALSE_CONTROL_ERROR (1 << 6) +#define RX_ECC_SINGLE_BIT_ERROR (1 << 7) +#define RX_ECC_MULTI_BIT_ERROR (1 << 8) +#define RX_CHECKSUM_ERROR (1 << 9) +#define RX_DSI_DATA_TYPE_NOT_RECOGNIZED (1 << 10) +#define RX_DSI_VC_ID_INVALID (1 << 11) +#define TX_FALSE_CONTROL_ERROR (1 << 12) +#define TX_ECC_SINGLE_BIT_ERROR (1 << 13) +#define TX_ECC_MULTI_BIT_ERROR (1 << 14) +#define TX_CHECKSUM_ERROR (1 << 15) +#define TX_DSI_DATA_TYPE_NOT_RECOGNIZED (1 << 16) +#define TX_DSI_VC_ID_INVALID (1 << 17) +#define HIGH_CONTENTION (1 << 18) +#define LOW_CONTENTION (1 << 19) +#define DPI_FIFO_UNDER_RUN (1 << 20) +#define HS_TX_TIMEOUT (1 << 21) +#define LP_RX_TIMEOUT (1 << 22) +#define TURN_AROUND_ACK_TIMEOUT (1 << 23) +#define ACK_WITH_NO_ERROR (1 << 24) +#define HS_GENERIC_WR_FIFO_FULL (1 << 27) +#define LP_GENERIC_WR_FIFO_FULL (1 << 28) +#define SPL_PKT_SENT (1 << 30) +#define INTR_EN_REG 0xb008 +#define DSI_FUNC_PRG_REG 0xb00c +#define DPI_CHANNEL_NUMBER_POS 0x03 +#define DBI_CHANNEL_NUMBER_POS 0x05 +#define FMT_DPI_POS 0x07 +#define FMT_DBI_POS 0x0A +#define DBI_DATA_WIDTH_POS 0x0D + +/* DPI PIXEL FORMATS */ +#define RGB_565_FMT 0x01 /* RGB 565 FORMAT */ +#define RGB_666_FMT 0x02 /* RGB 666 FORMAT */ +#define LRGB_666_FMT 0x03 /* RGB LOOSELY PACKED + * 666 FORMAT + */ +#define RGB_888_FMT 0x04 /* RGB 888 FORMAT */ +#define VIRTUAL_CHANNEL_NUMBER_0 0x00 /* Virtual channel 0 */ +#define VIRTUAL_CHANNEL_NUMBER_1 0x01 /* Virtual channel 1 */ +#define VIRTUAL_CHANNEL_NUMBER_2 0x02 /* Virtual channel 2 */ +#define VIRTUAL_CHANNEL_NUMBER_3 0x03 /* Virtual channel 3 */ + +#define DBI_NOT_SUPPORTED 0x00 /* command mode + * is not supported + */ +#define DBI_DATA_WIDTH_16BIT 0x01 /* 16 bit data */ +#define DBI_DATA_WIDTH_9BIT 0x02 /* 9 bit data */ +#define DBI_DATA_WIDTH_8BIT 0x03 /* 8 bit data */ +#define DBI_DATA_WIDTH_OPT1 0x04 /* option 1 */ +#define DBI_DATA_WIDTH_OPT2 0x05 /* option 2 */ + +#define HS_TX_TIMEOUT_REG 0xb010 +#define LP_RX_TIMEOUT_REG 0xb014 +#define TURN_AROUND_TIMEOUT_REG 0xb018 +#define DEVICE_RESET_REG 0xb01C +#define DPI_RESOLUTION_REG 0xb020 +#define RES_V_POS 0x10 +#define DBI_RESOLUTION_REG 0xb024 /* Reserved for MDFLD */ +#define HORIZ_SYNC_PAD_COUNT_REG 0xb028 +#define HORIZ_BACK_PORCH_COUNT_REG 0xb02C +#define HORIZ_FRONT_PORCH_COUNT_REG 0xb030 +#define HORIZ_ACTIVE_AREA_COUNT_REG 0xb034 +#define VERT_SYNC_PAD_COUNT_REG 0xb038 +#define VERT_BACK_PORCH_COUNT_REG 0xb03c +#define VERT_FRONT_PORCH_COUNT_REG 0xb040 +#define HIGH_LOW_SWITCH_COUNT_REG 0xb044 +#define DPI_CONTROL_REG 0xb048 +#define DPI_SHUT_DOWN (1 << 0) +#define DPI_TURN_ON (1 << 1) +#define DPI_COLOR_MODE_ON (1 << 2) +#define DPI_COLOR_MODE_OFF (1 << 3) +#define DPI_BACK_LIGHT_ON (1 << 4) +#define DPI_BACK_LIGHT_OFF (1 << 5) +#define DPI_LP (1 << 6) +#define DPI_DATA_REG 0xb04c +#define DPI_BACK_LIGHT_ON_DATA 0x07 +#define DPI_BACK_LIGHT_OFF_DATA 0x17 +#define INIT_COUNT_REG 0xb050 +#define MAX_RET_PAK_REG 0xb054 +#define VIDEO_FMT_REG 0xb058 +#define COMPLETE_LAST_PCKT (1 << 2) +#define EOT_DISABLE_REG 0xb05c +#define ENABLE_CLOCK_STOPPING (1 << 1) +#define LP_BYTECLK_REG 0xb060 +#define LP_GEN_DATA_REG 0xb064 +#define HS_GEN_DATA_REG 0xb068 +#define LP_GEN_CTRL_REG 0xb06C +#define HS_GEN_CTRL_REG 0xb070 +#define DCS_CHANNEL_NUMBER_POS 0x6 +#define MCS_COMMANDS_POS 0x8 +#define WORD_COUNTS_POS 0x8 +#define MCS_PARAMETER_POS 0x10 +#define GEN_FIFO_STAT_REG 0xb074 +#define HS_DATA_FIFO_FULL (1 << 0) +#define HS_DATA_FIFO_HALF_EMPTY (1 << 1) +#define HS_DATA_FIFO_EMPTY (1 << 2) +#define LP_DATA_FIFO_FULL (1 << 8) +#define LP_DATA_FIFO_HALF_EMPTY (1 << 9) +#define LP_DATA_FIFO_EMPTY (1 << 10) +#define HS_CTRL_FIFO_FULL (1 << 16) +#define HS_CTRL_FIFO_HALF_EMPTY (1 << 17) +#define HS_CTRL_FIFO_EMPTY (1 << 18) +#define LP_CTRL_FIFO_FULL (1 << 24) +#define LP_CTRL_FIFO_HALF_EMPTY (1 << 25) +#define LP_CTRL_FIFO_EMPTY (1 << 26) +#define DBI_FIFO_EMPTY (1 << 27) +#define DPI_FIFO_EMPTY (1 << 28) +#define HS_LS_DBI_ENABLE_REG 0xb078 +#define TXCLKESC_REG 0xb07c +#define DPHY_PARAM_REG 0xb080 +#define DBI_BW_CTRL_REG 0xb084 +#define CLK_LANE_SWT_REG 0xb088 + +/* + * MIPI Adapter registers + */ +#define MIPI_CONTROL_REG 0xb104 +#define MIPI_2X_CLOCK_BITS ((1 << 0) | (1 << 1)) +#define MIPI_DATA_ADDRESS_REG 0xb108 +#define MIPI_DATA_LENGTH_REG 0xb10C +#define MIPI_COMMAND_ADDRESS_REG 0xb110 +#define MIPI_COMMAND_LENGTH_REG 0xb114 +#define MIPI_READ_DATA_RETURN_REG0 0xb118 +#define MIPI_READ_DATA_RETURN_REG1 0xb11C +#define MIPI_READ_DATA_RETURN_REG2 0xb120 +#define MIPI_READ_DATA_RETURN_REG3 0xb124 +#define MIPI_READ_DATA_RETURN_REG4 0xb128 +#define MIPI_READ_DATA_RETURN_REG5 0xb12C +#define MIPI_READ_DATA_RETURN_REG6 0xb130 +#define MIPI_READ_DATA_RETURN_REG7 0xb134 +#define MIPI_READ_DATA_VALID_REG 0xb138 + +/* DBI COMMANDS */ +#define soft_reset 0x01 +/* + * The display module performs a software reset. + * Registers are written with their SW Reset default values. + */ +#define get_power_mode 0x0a +/* + * The display module returns the current power mode + */ +#define get_address_mode 0x0b +/* + * The display module returns the current status. + */ +#define get_pixel_format 0x0c +/* + * This command gets the pixel format for the RGB image data + * used by the interface. + */ +#define get_display_mode 0x0d +/* + * The display module returns the Display Image Mode status. + */ +#define get_signal_mode 0x0e +/* + * The display module returns the Display Signal Mode. + */ +#define get_diagnostic_result 0x0f +/* + * The display module returns the self-diagnostic results following + * a Sleep Out command. + */ +#define enter_sleep_mode 0x10 +/* + * This command causes the display module to enter the Sleep mode. + * In this mode, all unnecessary blocks inside the display module are + * disabled except interface communication. This is the lowest power + * mode the display module supports. + */ +#define exit_sleep_mode 0x11 +/* + * This command causes the display module to exit Sleep mode. + * All blocks inside the display module are enabled. + */ +#define enter_partial_mode 0x12 +/* + * This command causes the display module to enter the Partial Display + * Mode. The Partial Display Mode window is described by the + * set_partial_area command. + */ +#define enter_normal_mode 0x13 +/* + * This command causes the display module to enter the Normal mode. + * Normal Mode is defined as Partial Display mode and Scroll mode are off + */ +#define exit_invert_mode 0x20 +/* + * This command causes the display module to stop inverting the image + * data on the display device. The frame memory contents remain unchanged. + * No status bits are changed. + */ +#define enter_invert_mode 0x21 +/* + * This command causes the display module to invert the image data only on + * the display device. The frame memory contents remain unchanged. + * No status bits are changed. + */ +#define set_gamma_curve 0x26 +/* + * This command selects the desired gamma curve for the display device. + * Four fixed gamma curves are defined in section DCS spec. + */ +#define set_display_off 0x28 +/* ************************************************************************* *\ +This command causes the display module to stop displaying the image data +on the display device. The frame memory contents remain unchanged. +No status bits are changed. +\* ************************************************************************* */ +#define set_display_on 0x29 +/* ************************************************************************* *\ +This command causes the display module to start displaying the image data +on the display device. The frame memory contents remain unchanged. +No status bits are changed. +\* ************************************************************************* */ +#define set_column_address 0x2a +/* + * This command defines the column extent of the frame memory accessed by + * the hostprocessor with the read_memory_continue and + * write_memory_continue commands. + * No status bits are changed. + */ +#define set_page_addr 0x2b +/* + * This command defines the page extent of the frame memory accessed by + * the host processor with the write_memory_continue and + * read_memory_continue command. + * No status bits are changed. + */ +#define write_mem_start 0x2c +/* + * This command transfers image data from the host processor to the + * display modules frame memory starting at the pixel location specified + * by preceding set_column_address and set_page_address commands. + */ +#define set_partial_area 0x30 +/* + * This command defines the Partial Display mode s display area. + * There are two parameters associated with this command, the first + * defines the Start Row (SR) and the second the End Row (ER). SR and ER + * refer to the Frame Memory Line Pointer. + */ +#define set_scroll_area 0x33 +/* + * This command defines the display modules Vertical Scrolling Area. + */ +#define set_tear_off 0x34 +/* + * This command turns off the display modules Tearing Effect output + * signal on the TE signal line. + */ +#define set_tear_on 0x35 +/* + * This command turns on the display modules Tearing Effect output signal + * on the TE signal line. + */ +#define set_address_mode 0x36 +/* + * This command sets the data order for transfers from the host processor + * to display modules frame memory,bits B[7:5] and B3, and from the + * display modules frame memory to the display device, bits B[2:0] and B4. + */ +#define set_scroll_start 0x37 +/* + * This command sets the start of the vertical scrolling area in the frame + * memory. The vertical scrolling area is fully defined when this command + * is used with the set_scroll_area command The set_scroll_start command + * has one parameter, the Vertical Scroll Pointer. The VSP defines the + * line in the frame memory that is written to the display device as the + * first line of the vertical scroll area. + */ +#define exit_idle_mode 0x38 +/* + * This command causes the display module to exit Idle mode. + */ +#define enter_idle_mode 0x39 +/* + * This command causes the display module to enter Idle Mode. + * In Idle Mode, color expression is reduced. Colors are shown on the + * display device using the MSB of each of the R, G and B color + * components in the frame memory + */ +#define set_pixel_format 0x3a +/* + * This command sets the pixel format for the RGB image data used by the + * interface. + * Bits D[6:4] DPI Pixel Format Definition + * Bits D[2:0] DBI Pixel Format Definition + * Bits D7 and D3 are not used. + */ +#define DCS_PIXEL_FORMAT_3bpp 0x1 +#define DCS_PIXEL_FORMAT_8bpp 0x2 +#define DCS_PIXEL_FORMAT_12bpp 0x3 +#define DCS_PIXEL_FORMAT_16bpp 0x5 +#define DCS_PIXEL_FORMAT_18bpp 0x6 +#define DCS_PIXEL_FORMAT_24bpp 0x7 + +#define write_mem_cont 0x3c + +/* + * This command transfers image data from the host processor to the + * display module's frame memory continuing from the pixel location + * following the previous write_memory_continue or write_memory_start + * command. + */ +#define set_tear_scanline 0x44 +/* + * This command turns on the display modules Tearing Effect output signal + * on the TE signal line when the display module reaches line N. + */ +#define get_scanline 0x45 +/* + * The display module returns the current scanline, N, used to update the + * display device. The total number of scanlines on a display device is + * defined as VSYNC + VBP + VACT + VFP.The first scanline is defined as + * the first line of V Sync and is denoted as Line 0. + * When in Sleep Mode, the value returned by get_scanline is undefined. + */ + +/* MCS or Generic COMMANDS */ +/* MCS/generic data type */ +#define GEN_SHORT_WRITE_0 0x03 /* generic short write, no parameters */ +#define GEN_SHORT_WRITE_1 0x13 /* generic short write, 1 parameters */ +#define GEN_SHORT_WRITE_2 0x23 /* generic short write, 2 parameters */ +#define GEN_READ_0 0x04 /* generic read, no parameters */ +#define GEN_READ_1 0x14 /* generic read, 1 parameters */ +#define GEN_READ_2 0x24 /* generic read, 2 parameters */ +#define GEN_LONG_WRITE 0x29 /* generic long write */ +#define MCS_SHORT_WRITE_0 0x05 /* MCS short write, no parameters */ +#define MCS_SHORT_WRITE_1 0x15 /* MCS short write, 1 parameters */ +#define MCS_READ 0x06 /* MCS read, no parameters */ +#define MCS_LONG_WRITE 0x39 /* MCS long write */ +/* MCS/generic commands */ +/* TPO MCS */ +#define write_display_profile 0x50 +#define write_display_brightness 0x51 +#define write_ctrl_display 0x53 +#define write_ctrl_cabc 0x55 + #define UI_IMAGE 0x01 + #define STILL_IMAGE 0x02 + #define MOVING_IMAGE 0x03 +#define write_hysteresis 0x57 +#define write_gamma_setting 0x58 +#define write_cabc_min_bright 0x5e +#define write_kbbc_profile 0x60 +/* TMD MCS */ +#define tmd_write_display_brightness 0x8c + +/* + * This command is used to control ambient light, panel backlight + * brightness and gamma settings. + */ +#define BRIGHT_CNTL_BLOCK_ON (1 << 5) +#define AMBIENT_LIGHT_SENSE_ON (1 << 4) +#define DISPLAY_DIMMING_ON (1 << 3) +#define BACKLIGHT_ON (1 << 2) +#define DISPLAY_BRIGHTNESS_AUTO (1 << 1) +#define GAMMA_AUTO (1 << 0) + +/* DCS Interface Pixel Formats */ +#define DCS_PIXEL_FORMAT_3BPP 0x1 +#define DCS_PIXEL_FORMAT_8BPP 0x2 +#define DCS_PIXEL_FORMAT_12BPP 0x3 +#define DCS_PIXEL_FORMAT_16BPP 0x5 +#define DCS_PIXEL_FORMAT_18BPP 0x6 +#define DCS_PIXEL_FORMAT_24BPP 0x7 +/* ONE PARAMETER READ DATA */ +#define addr_mode_data 0xfc +#define diag_res_data 0x00 +#define disp_mode_data 0x23 +#define pxl_fmt_data 0x77 +#define pwr_mode_data 0x74 +#define sig_mode_data 0x00 +/* TWO PARAMETERS READ DATA */ +#define scanline_data1 0xff +#define scanline_data2 0xff +#define NON_BURST_MODE_SYNC_PULSE 0x01 /* Non Burst Mode + * with Sync Pulse + */ +#define NON_BURST_MODE_SYNC_EVENTS 0x02 /* Non Burst Mode + * with Sync events + */ +#define BURST_MODE 0x03 /* Burst Mode */ +#define DBI_COMMAND_BUFFER_SIZE 0x240 /* 0x32 */ /* 0x120 */ + /* Allocate at least + * 0x100 Byte with 32 + * byte alignment + */ +#define DBI_DATA_BUFFER_SIZE 0x120 /* Allocate at least + * 0x100 Byte with 32 + * byte alignment + */ +#define DBI_CB_TIME_OUT 0xFFFF + +#define GEN_FB_TIME_OUT 2000 + +#define SKU_83 0x01 +#define SKU_100 0x02 +#define SKU_100L 0x04 +#define SKU_BYPASS 0x08 + +/* Some handy macros for playing with bitfields. */ +#define PSB_MASK(high, low) (((1<<((high)-(low)+1))-1)<<(low)) +#define SET_FIELD(value, field) (((value) << field ## _SHIFT) & field ## _MASK) +#define GET_FIELD(word, field) (((word) & field ## _MASK) >> field ## _SHIFT) + +#define _PIPE(pipe, a, b) ((a) + (pipe)*((b)-(a))) + +/* PCI config space */ + +#define SB_PCKT 0x02100 /* cedarview */ +# define SB_OPCODE_MASK PSB_MASK(31, 16) +# define SB_OPCODE_SHIFT 16 +# define SB_OPCODE_READ 0 +# define SB_OPCODE_WRITE 1 +# define SB_DEST_MASK PSB_MASK(15, 8) +# define SB_DEST_SHIFT 8 +# define SB_DEST_DPLL 0x88 +# define SB_BYTE_ENABLE_MASK PSB_MASK(7, 4) +# define SB_BYTE_ENABLE_SHIFT 4 +# define SB_BUSY (1 << 0) + +#define DSPCLK_GATE_D 0x6200 +# define VRHUNIT_CLOCK_GATE_DISABLE (1 << 28) /* Fixed value on CDV */ +# define DPOUNIT_CLOCK_GATE_DISABLE (1 << 11) +# define DPIOUNIT_CLOCK_GATE_DISABLE (1 << 6) + +#define RAMCLK_GATE_D 0x6210 + +/* 32-bit value read/written from the DPIO reg. */ +#define SB_DATA 0x02104 /* cedarview */ +/* 32-bit address of the DPIO reg to be read/written. */ +#define SB_ADDR 0x02108 /* cedarview */ +#define DPIO_CFG 0x02110 /* cedarview */ +# define DPIO_MODE_SELECT_1 (1 << 3) +# define DPIO_MODE_SELECT_0 (1 << 2) +# define DPIO_SFR_BYPASS (1 << 1) +/* reset is active low */ +# define DPIO_CMN_RESET_N (1 << 0) + +/* Cedarview sideband registers */ +#define _SB_M_A 0x8008 +#define _SB_M_B 0x8028 +#define SB_M(pipe) _PIPE(pipe, _SB_M_A, _SB_M_B) +# define SB_M_DIVIDER_MASK (0xFF << 24) +# define SB_M_DIVIDER_SHIFT 24 + +#define _SB_N_VCO_A 0x8014 +#define _SB_N_VCO_B 0x8034 +#define SB_N_VCO(pipe) _PIPE(pipe, _SB_N_VCO_A, _SB_N_VCO_B) +#define SB_N_VCO_SEL_MASK PSB_MASK(31, 30) +#define SB_N_VCO_SEL_SHIFT 30 +#define SB_N_DIVIDER_MASK PSB_MASK(29, 26) +#define SB_N_DIVIDER_SHIFT 26 +#define SB_N_CB_TUNE_MASK PSB_MASK(25, 24) +#define SB_N_CB_TUNE_SHIFT 24 + +#define _SB_REF_A 0x8018 +#define _SB_REF_B 0x8038 +#define SB_REF_SFR(pipe) _PIPE(pipe, _SB_REF_A, _SB_REF_B) + +#define _SB_P_A 0x801c +#define _SB_P_B 0x803c +#define SB_P(pipe) _PIPE(pipe, _SB_P_A, _SB_P_B) +#define SB_P2_DIVIDER_MASK PSB_MASK(31, 30) +#define SB_P2_DIVIDER_SHIFT 30 +#define SB_P2_10 0 /* HDMI, DP, DAC */ +#define SB_P2_5 1 /* DAC */ +#define SB_P2_14 2 /* LVDS single */ +#define SB_P2_7 3 /* LVDS double */ +#define SB_P1_DIVIDER_MASK PSB_MASK(15, 12) +#define SB_P1_DIVIDER_SHIFT 12 + +#define PSB_LANE0 0x120 +#define PSB_LANE1 0x220 +#define PSB_LANE2 0x2320 +#define PSB_LANE3 0x2420 + +#define LANE_PLL_MASK (0x7 << 20) +#define LANE_PLL_ENABLE (0x3 << 20) + + +#endif diff --git a/drivers/gpu/drm/gma500/psb_intel_sdvo.c b/drivers/gpu/drm/gma500/psb_intel_sdvo.c new file mode 100644 index 0000000..36330ca --- /dev/null +++ b/drivers/gpu/drm/gma500/psb_intel_sdvo.c @@ -0,0 +1,2607 @@ +/* + * Copyright 2006 Dave Airlie + * Copyright © 2006-2007 Intel Corporation + * Jesse Barnes + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * Eric Anholt + */ +#include +#include +#include +#include +#include "drmP.h" +#include "drm.h" +#include "drm_crtc.h" +#include "drm_edid.h" +#include "psb_intel_drv.h" +#include "gma_drm.h" +#include "psb_drv.h" +#include "psb_intel_sdvo_regs.h" +#include "psb_intel_reg.h" + +#define SDVO_TMDS_MASK (SDVO_OUTPUT_TMDS0 | SDVO_OUTPUT_TMDS1) +#define SDVO_RGB_MASK (SDVO_OUTPUT_RGB0 | SDVO_OUTPUT_RGB1) +#define SDVO_LVDS_MASK (SDVO_OUTPUT_LVDS0 | SDVO_OUTPUT_LVDS1) +#define SDVO_TV_MASK (SDVO_OUTPUT_CVBS0 | SDVO_OUTPUT_SVID0) + +#define SDVO_OUTPUT_MASK (SDVO_TMDS_MASK | SDVO_RGB_MASK | SDVO_LVDS_MASK |\ + SDVO_TV_MASK) + +#define IS_TV(c) (c->output_flag & SDVO_TV_MASK) +#define IS_TMDS(c) (c->output_flag & SDVO_TMDS_MASK) +#define IS_LVDS(c) (c->output_flag & SDVO_LVDS_MASK) +#define IS_TV_OR_LVDS(c) (c->output_flag & (SDVO_TV_MASK | SDVO_LVDS_MASK)) + + +static const char *tv_format_names[] = { + "NTSC_M" , "NTSC_J" , "NTSC_443", + "PAL_B" , "PAL_D" , "PAL_G" , + "PAL_H" , "PAL_I" , "PAL_M" , + "PAL_N" , "PAL_NC" , "PAL_60" , + "SECAM_B" , "SECAM_D" , "SECAM_G" , + "SECAM_K" , "SECAM_K1", "SECAM_L" , + "SECAM_60" +}; + +#define TV_FORMAT_NUM (sizeof(tv_format_names) / sizeof(*tv_format_names)) + +struct psb_intel_sdvo { + struct psb_intel_encoder base; + + struct i2c_adapter *i2c; + u8 slave_addr; + + struct i2c_adapter ddc; + + /* Register for the SDVO device: SDVOB or SDVOC */ + int sdvo_reg; + + /* Active outputs controlled by this SDVO output */ + uint16_t controlled_output; + + /* + * Capabilities of the SDVO device returned by + * i830_sdvo_get_capabilities() + */ + struct psb_intel_sdvo_caps caps; + + /* Pixel clock limitations reported by the SDVO device, in kHz */ + int pixel_clock_min, pixel_clock_max; + + /* + * For multiple function SDVO device, + * this is for current attached outputs. + */ + uint16_t attached_output; + + /** + * This is used to select the color range of RBG outputs in HDMI mode. + * It is only valid when using TMDS encoding and 8 bit per color mode. + */ + uint32_t color_range; + + /** + * This is set if we're going to treat the device as TV-out. + * + * While we have these nice friendly flags for output types that ought + * to decide this for us, the S-Video output on our HDMI+S-Video card + * shows up as RGB1 (VGA). + */ + bool is_tv; + + /* This is for current tv format name */ + int tv_format_index; + + /** + * This is set if we treat the device as HDMI, instead of DVI. + */ + bool is_hdmi; + bool has_hdmi_monitor; + bool has_hdmi_audio; + + /** + * This is set if we detect output of sdvo device as LVDS and + * have a valid fixed mode to use with the panel. + */ + bool is_lvds; + + /** + * This is sdvo fixed pannel mode pointer + */ + struct drm_display_mode *sdvo_lvds_fixed_mode; + + /* DDC bus used by this SDVO encoder */ + uint8_t ddc_bus; + + /* Input timings for adjusted_mode */ + struct psb_intel_sdvo_dtd input_dtd; +}; + +struct psb_intel_sdvo_connector { + struct psb_intel_connector base; + + /* Mark the type of connector */ + uint16_t output_flag; + + int force_audio; + + /* This contains all current supported TV format */ + u8 tv_format_supported[TV_FORMAT_NUM]; + int format_supported_num; + struct drm_property *tv_format; + + /* add the property for the SDVO-TV */ + struct drm_property *left; + struct drm_property *right; + struct drm_property *top; + struct drm_property *bottom; + struct drm_property *hpos; + struct drm_property *vpos; + struct drm_property *contrast; + struct drm_property *saturation; + struct drm_property *hue; + struct drm_property *sharpness; + struct drm_property *flicker_filter; + struct drm_property *flicker_filter_adaptive; + struct drm_property *flicker_filter_2d; + struct drm_property *tv_chroma_filter; + struct drm_property *tv_luma_filter; + struct drm_property *dot_crawl; + + /* add the property for the SDVO-TV/LVDS */ + struct drm_property *brightness; + + /* Add variable to record current setting for the above property */ + u32 left_margin, right_margin, top_margin, bottom_margin; + + /* this is to get the range of margin.*/ + u32 max_hscan, max_vscan; + u32 max_hpos, cur_hpos; + u32 max_vpos, cur_vpos; + u32 cur_brightness, max_brightness; + u32 cur_contrast, max_contrast; + u32 cur_saturation, max_saturation; + u32 cur_hue, max_hue; + u32 cur_sharpness, max_sharpness; + u32 cur_flicker_filter, max_flicker_filter; + u32 cur_flicker_filter_adaptive, max_flicker_filter_adaptive; + u32 cur_flicker_filter_2d, max_flicker_filter_2d; + u32 cur_tv_chroma_filter, max_tv_chroma_filter; + u32 cur_tv_luma_filter, max_tv_luma_filter; + u32 cur_dot_crawl, max_dot_crawl; +}; + +static struct psb_intel_sdvo *to_psb_intel_sdvo(struct drm_encoder *encoder) +{ + return container_of(encoder, struct psb_intel_sdvo, base.base); +} + +static struct psb_intel_sdvo *intel_attached_sdvo(struct drm_connector *connector) +{ + return container_of(psb_intel_attached_encoder(connector), + struct psb_intel_sdvo, base); +} + +static struct psb_intel_sdvo_connector *to_psb_intel_sdvo_connector(struct drm_connector *connector) +{ + return container_of(to_psb_intel_connector(connector), struct psb_intel_sdvo_connector, base); +} + +static bool +psb_intel_sdvo_output_setup(struct psb_intel_sdvo *psb_intel_sdvo, uint16_t flags); +static bool +psb_intel_sdvo_tv_create_property(struct psb_intel_sdvo *psb_intel_sdvo, + struct psb_intel_sdvo_connector *psb_intel_sdvo_connector, + int type); +static bool +psb_intel_sdvo_create_enhance_property(struct psb_intel_sdvo *psb_intel_sdvo, + struct psb_intel_sdvo_connector *psb_intel_sdvo_connector); + +/** + * Writes the SDVOB or SDVOC with the given value, but always writes both + * SDVOB and SDVOC to work around apparent hardware issues (according to + * comments in the BIOS). + */ +static void psb_intel_sdvo_write_sdvox(struct psb_intel_sdvo *psb_intel_sdvo, u32 val) +{ + struct drm_device *dev = psb_intel_sdvo->base.base.dev; + u32 bval = val, cval = val; + int i; + + if (psb_intel_sdvo->sdvo_reg == SDVOB) { + cval = REG_READ(SDVOC); + } else { + bval = REG_READ(SDVOB); + } + /* + * Write the registers twice for luck. Sometimes, + * writing them only once doesn't appear to 'stick'. + * The BIOS does this too. Yay, magic + */ + for (i = 0; i < 2; i++) + { + REG_WRITE(SDVOB, bval); + REG_READ(SDVOB); + REG_WRITE(SDVOC, cval); + REG_READ(SDVOC); + } +} + +static bool psb_intel_sdvo_read_byte(struct psb_intel_sdvo *psb_intel_sdvo, u8 addr, u8 *ch) +{ + struct i2c_msg msgs[] = { + { + .addr = psb_intel_sdvo->slave_addr, + .flags = 0, + .len = 1, + .buf = &addr, + }, + { + .addr = psb_intel_sdvo->slave_addr, + .flags = I2C_M_RD, + .len = 1, + .buf = ch, + } + }; + int ret; + + if ((ret = i2c_transfer(psb_intel_sdvo->i2c, msgs, 2)) == 2) + return true; + + DRM_DEBUG_KMS("i2c transfer returned %d\n", ret); + return false; +} + +#define SDVO_CMD_NAME_ENTRY(cmd) {cmd, #cmd} +/** Mapping of command numbers to names, for debug output */ +static const struct _sdvo_cmd_name { + u8 cmd; + const char *name; +} sdvo_cmd_names[] = { + SDVO_CMD_NAME_ENTRY(SDVO_CMD_RESET), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_DEVICE_CAPS), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_FIRMWARE_REV), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_TRAINED_INPUTS), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_ACTIVE_OUTPUTS), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_ACTIVE_OUTPUTS), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_IN_OUT_MAP), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_IN_OUT_MAP), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_ATTACHED_DISPLAYS), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HOT_PLUG_SUPPORT), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_ACTIVE_HOT_PLUG), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_ACTIVE_HOT_PLUG), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_INTERRUPT_EVENT_SOURCE), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_TARGET_INPUT), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_TARGET_OUTPUT), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_INPUT_TIMINGS_PART1), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_INPUT_TIMINGS_PART2), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_INPUT_TIMINGS_PART1), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_INPUT_TIMINGS_PART2), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_INPUT_TIMINGS_PART1), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_OUTPUT_TIMINGS_PART1), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_OUTPUT_TIMINGS_PART2), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_OUTPUT_TIMINGS_PART1), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_OUTPUT_TIMINGS_PART2), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_CREATE_PREFERRED_INPUT_TIMING), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_PREFERRED_INPUT_TIMING_PART1), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_PREFERRED_INPUT_TIMING_PART2), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_INPUT_PIXEL_CLOCK_RANGE), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_OUTPUT_PIXEL_CLOCK_RANGE), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SUPPORTED_CLOCK_RATE_MULTS), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_CLOCK_RATE_MULT), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_CLOCK_RATE_MULT), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SUPPORTED_TV_FORMATS), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_TV_FORMAT), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_TV_FORMAT), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SUPPORTED_POWER_STATES), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_POWER_STATE), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_ENCODER_POWER_STATE), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_DISPLAY_POWER_STATE), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_CONTROL_BUS_SWITCH), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SDTV_RESOLUTION_SUPPORT), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SCALED_HDTV_RESOLUTION_SUPPORT), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SUPPORTED_ENHANCEMENTS), + + /* Add the op code for SDVO enhancements */ + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_HPOS), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HPOS), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_HPOS), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_VPOS), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_VPOS), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_VPOS), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_SATURATION), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SATURATION), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_SATURATION), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_HUE), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HUE), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_HUE), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_CONTRAST), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_CONTRAST), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_CONTRAST), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_BRIGHTNESS), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_BRIGHTNESS), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_BRIGHTNESS), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_OVERSCAN_H), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_OVERSCAN_H), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_OVERSCAN_H), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_OVERSCAN_V), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_OVERSCAN_V), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_OVERSCAN_V), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_FLICKER_FILTER), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_FLICKER_FILTER), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_FLICKER_FILTER), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_FLICKER_FILTER_ADAPTIVE), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_FLICKER_FILTER_ADAPTIVE), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_FLICKER_FILTER_ADAPTIVE), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_FLICKER_FILTER_2D), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_FLICKER_FILTER_2D), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_FLICKER_FILTER_2D), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_SHARPNESS), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SHARPNESS), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_SHARPNESS), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_DOT_CRAWL), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_DOT_CRAWL), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_TV_CHROMA_FILTER), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_TV_CHROMA_FILTER), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_TV_CHROMA_FILTER), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_TV_LUMA_FILTER), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_TV_LUMA_FILTER), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_TV_LUMA_FILTER), + + /* HDMI op code */ + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SUPP_ENCODE), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_ENCODE), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_ENCODE), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_PIXEL_REPLI), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_PIXEL_REPLI), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_COLORIMETRY_CAP), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_COLORIMETRY), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_COLORIMETRY), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_AUDIO_ENCRYPT_PREFER), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_AUDIO_STAT), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_AUDIO_STAT), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HBUF_INDEX), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_HBUF_INDEX), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HBUF_INFO), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HBUF_AV_SPLIT), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_HBUF_AV_SPLIT), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HBUF_TXRATE), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_HBUF_TXRATE), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_HBUF_DATA), + SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HBUF_DATA), +}; + +#define IS_SDVOB(reg) (reg == SDVOB) +#define SDVO_NAME(svdo) (IS_SDVOB((svdo)->sdvo_reg) ? "SDVOB" : "SDVOC") + +static void psb_intel_sdvo_debug_write(struct psb_intel_sdvo *psb_intel_sdvo, u8 cmd, + const void *args, int args_len) +{ + int i; + + DRM_DEBUG_KMS("%s: W: %02X ", + SDVO_NAME(psb_intel_sdvo), cmd); + for (i = 0; i < args_len; i++) + DRM_LOG_KMS("%02X ", ((u8 *)args)[i]); + for (; i < 8; i++) + DRM_LOG_KMS(" "); + for (i = 0; i < ARRAY_SIZE(sdvo_cmd_names); i++) { + if (cmd == sdvo_cmd_names[i].cmd) { + DRM_LOG_KMS("(%s)", sdvo_cmd_names[i].name); + break; + } + } + if (i == ARRAY_SIZE(sdvo_cmd_names)) + DRM_LOG_KMS("(%02X)", cmd); + DRM_LOG_KMS("\n"); +} + +static const char *cmd_status_names[] = { + "Power on", + "Success", + "Not supported", + "Invalid arg", + "Pending", + "Target not specified", + "Scaling not supported" +}; + +static bool psb_intel_sdvo_write_cmd(struct psb_intel_sdvo *psb_intel_sdvo, u8 cmd, + const void *args, int args_len) +{ + u8 buf[args_len*2 + 2], status; + struct i2c_msg msgs[args_len + 3]; + int i, ret; + + psb_intel_sdvo_debug_write(psb_intel_sdvo, cmd, args, args_len); + + for (i = 0; i < args_len; i++) { + msgs[i].addr = psb_intel_sdvo->slave_addr; + msgs[i].flags = 0; + msgs[i].len = 2; + msgs[i].buf = buf + 2 *i; + buf[2*i + 0] = SDVO_I2C_ARG_0 - i; + buf[2*i + 1] = ((u8*)args)[i]; + } + msgs[i].addr = psb_intel_sdvo->slave_addr; + msgs[i].flags = 0; + msgs[i].len = 2; + msgs[i].buf = buf + 2*i; + buf[2*i + 0] = SDVO_I2C_OPCODE; + buf[2*i + 1] = cmd; + + /* the following two are to read the response */ + status = SDVO_I2C_CMD_STATUS; + msgs[i+1].addr = psb_intel_sdvo->slave_addr; + msgs[i+1].flags = 0; + msgs[i+1].len = 1; + msgs[i+1].buf = &status; + + msgs[i+2].addr = psb_intel_sdvo->slave_addr; + msgs[i+2].flags = I2C_M_RD; + msgs[i+2].len = 1; + msgs[i+2].buf = &status; + + ret = i2c_transfer(psb_intel_sdvo->i2c, msgs, i+3); + if (ret < 0) { + DRM_DEBUG_KMS("I2c transfer returned %d\n", ret); + return false; + } + if (ret != i+3) { + /* failure in I2C transfer */ + DRM_DEBUG_KMS("I2c transfer returned %d/%d\n", ret, i+3); + return false; + } + + return true; +} + +static bool psb_intel_sdvo_read_response(struct psb_intel_sdvo *psb_intel_sdvo, + void *response, int response_len) +{ + u8 retry = 5; + u8 status; + int i; + + DRM_DEBUG_KMS("%s: R: ", SDVO_NAME(psb_intel_sdvo)); + + /* + * The documentation states that all commands will be + * processed within 15µs, and that we need only poll + * the status byte a maximum of 3 times in order for the + * command to be complete. + * + * Check 5 times in case the hardware failed to read the docs. + */ + if (!psb_intel_sdvo_read_byte(psb_intel_sdvo, + SDVO_I2C_CMD_STATUS, + &status)) + goto log_fail; + + while (status == SDVO_CMD_STATUS_PENDING && retry--) { + udelay(15); + if (!psb_intel_sdvo_read_byte(psb_intel_sdvo, + SDVO_I2C_CMD_STATUS, + &status)) + goto log_fail; + } + + if (status <= SDVO_CMD_STATUS_SCALING_NOT_SUPP) + DRM_LOG_KMS("(%s)", cmd_status_names[status]); + else + DRM_LOG_KMS("(??? %d)", status); + + if (status != SDVO_CMD_STATUS_SUCCESS) + goto log_fail; + + /* Read the command response */ + for (i = 0; i < response_len; i++) { + if (!psb_intel_sdvo_read_byte(psb_intel_sdvo, + SDVO_I2C_RETURN_0 + i, + &((u8 *)response)[i])) + goto log_fail; + DRM_LOG_KMS(" %02X", ((u8 *)response)[i]); + } + DRM_LOG_KMS("\n"); + return true; + +log_fail: + DRM_LOG_KMS("... failed\n"); + return false; +} + +static int psb_intel_sdvo_get_pixel_multiplier(struct drm_display_mode *mode) +{ + if (mode->clock >= 100000) + return 1; + else if (mode->clock >= 50000) + return 2; + else + return 4; +} + +static bool psb_intel_sdvo_set_control_bus_switch(struct psb_intel_sdvo *psb_intel_sdvo, + u8 ddc_bus) +{ + /* This must be the immediately preceding write before the i2c xfer */ + return psb_intel_sdvo_write_cmd(psb_intel_sdvo, + SDVO_CMD_SET_CONTROL_BUS_SWITCH, + &ddc_bus, 1); +} + +static bool psb_intel_sdvo_set_value(struct psb_intel_sdvo *psb_intel_sdvo, u8 cmd, const void *data, int len) +{ + if (!psb_intel_sdvo_write_cmd(psb_intel_sdvo, cmd, data, len)) + return false; + + return psb_intel_sdvo_read_response(psb_intel_sdvo, NULL, 0); +} + +static bool +psb_intel_sdvo_get_value(struct psb_intel_sdvo *psb_intel_sdvo, u8 cmd, void *value, int len) +{ + if (!psb_intel_sdvo_write_cmd(psb_intel_sdvo, cmd, NULL, 0)) + return false; + + return psb_intel_sdvo_read_response(psb_intel_sdvo, value, len); +} + +static bool psb_intel_sdvo_set_target_input(struct psb_intel_sdvo *psb_intel_sdvo) +{ + struct psb_intel_sdvo_set_target_input_args targets = {0}; + return psb_intel_sdvo_set_value(psb_intel_sdvo, + SDVO_CMD_SET_TARGET_INPUT, + &targets, sizeof(targets)); +} + +/** + * Return whether each input is trained. + * + * This function is making an assumption about the layout of the response, + * which should be checked against the docs. + */ +static bool psb_intel_sdvo_get_trained_inputs(struct psb_intel_sdvo *psb_intel_sdvo, bool *input_1, bool *input_2) +{ + struct psb_intel_sdvo_get_trained_inputs_response response; + + BUILD_BUG_ON(sizeof(response) != 1); + if (!psb_intel_sdvo_get_value(psb_intel_sdvo, SDVO_CMD_GET_TRAINED_INPUTS, + &response, sizeof(response))) + return false; + + *input_1 = response.input0_trained; + *input_2 = response.input1_trained; + return true; +} + +static bool psb_intel_sdvo_set_active_outputs(struct psb_intel_sdvo *psb_intel_sdvo, + u16 outputs) +{ + return psb_intel_sdvo_set_value(psb_intel_sdvo, + SDVO_CMD_SET_ACTIVE_OUTPUTS, + &outputs, sizeof(outputs)); +} + +static bool psb_intel_sdvo_set_encoder_power_state(struct psb_intel_sdvo *psb_intel_sdvo, + int mode) +{ + u8 state = SDVO_ENCODER_STATE_ON; + + switch (mode) { + case DRM_MODE_DPMS_ON: + state = SDVO_ENCODER_STATE_ON; + break; + case DRM_MODE_DPMS_STANDBY: + state = SDVO_ENCODER_STATE_STANDBY; + break; + case DRM_MODE_DPMS_SUSPEND: + state = SDVO_ENCODER_STATE_SUSPEND; + break; + case DRM_MODE_DPMS_OFF: + state = SDVO_ENCODER_STATE_OFF; + break; + } + + return psb_intel_sdvo_set_value(psb_intel_sdvo, + SDVO_CMD_SET_ENCODER_POWER_STATE, &state, sizeof(state)); +} + +static bool psb_intel_sdvo_get_input_pixel_clock_range(struct psb_intel_sdvo *psb_intel_sdvo, + int *clock_min, + int *clock_max) +{ + struct psb_intel_sdvo_pixel_clock_range clocks; + + BUILD_BUG_ON(sizeof(clocks) != 4); + if (!psb_intel_sdvo_get_value(psb_intel_sdvo, + SDVO_CMD_GET_INPUT_PIXEL_CLOCK_RANGE, + &clocks, sizeof(clocks))) + return false; + + /* Convert the values from units of 10 kHz to kHz. */ + *clock_min = clocks.min * 10; + *clock_max = clocks.max * 10; + return true; +} + +static bool psb_intel_sdvo_set_target_output(struct psb_intel_sdvo *psb_intel_sdvo, + u16 outputs) +{ + return psb_intel_sdvo_set_value(psb_intel_sdvo, + SDVO_CMD_SET_TARGET_OUTPUT, + &outputs, sizeof(outputs)); +} + +static bool psb_intel_sdvo_set_timing(struct psb_intel_sdvo *psb_intel_sdvo, u8 cmd, + struct psb_intel_sdvo_dtd *dtd) +{ + return psb_intel_sdvo_set_value(psb_intel_sdvo, cmd, &dtd->part1, sizeof(dtd->part1)) && + psb_intel_sdvo_set_value(psb_intel_sdvo, cmd + 1, &dtd->part2, sizeof(dtd->part2)); +} + +static bool psb_intel_sdvo_set_input_timing(struct psb_intel_sdvo *psb_intel_sdvo, + struct psb_intel_sdvo_dtd *dtd) +{ + return psb_intel_sdvo_set_timing(psb_intel_sdvo, + SDVO_CMD_SET_INPUT_TIMINGS_PART1, dtd); +} + +static bool psb_intel_sdvo_set_output_timing(struct psb_intel_sdvo *psb_intel_sdvo, + struct psb_intel_sdvo_dtd *dtd) +{ + return psb_intel_sdvo_set_timing(psb_intel_sdvo, + SDVO_CMD_SET_OUTPUT_TIMINGS_PART1, dtd); +} + +static bool +psb_intel_sdvo_create_preferred_input_timing(struct psb_intel_sdvo *psb_intel_sdvo, + uint16_t clock, + uint16_t width, + uint16_t height) +{ + struct psb_intel_sdvo_preferred_input_timing_args args; + + memset(&args, 0, sizeof(args)); + args.clock = clock; + args.width = width; + args.height = height; + args.interlace = 0; + + if (psb_intel_sdvo->is_lvds && + (psb_intel_sdvo->sdvo_lvds_fixed_mode->hdisplay != width || + psb_intel_sdvo->sdvo_lvds_fixed_mode->vdisplay != height)) + args.scaled = 1; + + return psb_intel_sdvo_set_value(psb_intel_sdvo, + SDVO_CMD_CREATE_PREFERRED_INPUT_TIMING, + &args, sizeof(args)); +} + +static bool psb_intel_sdvo_get_preferred_input_timing(struct psb_intel_sdvo *psb_intel_sdvo, + struct psb_intel_sdvo_dtd *dtd) +{ + BUILD_BUG_ON(sizeof(dtd->part1) != 8); + BUILD_BUG_ON(sizeof(dtd->part2) != 8); + return psb_intel_sdvo_get_value(psb_intel_sdvo, SDVO_CMD_GET_PREFERRED_INPUT_TIMING_PART1, + &dtd->part1, sizeof(dtd->part1)) && + psb_intel_sdvo_get_value(psb_intel_sdvo, SDVO_CMD_GET_PREFERRED_INPUT_TIMING_PART2, + &dtd->part2, sizeof(dtd->part2)); +} + +static bool psb_intel_sdvo_set_clock_rate_mult(struct psb_intel_sdvo *psb_intel_sdvo, u8 val) +{ + return psb_intel_sdvo_set_value(psb_intel_sdvo, SDVO_CMD_SET_CLOCK_RATE_MULT, &val, 1); +} + +static void psb_intel_sdvo_get_dtd_from_mode(struct psb_intel_sdvo_dtd *dtd, + const struct drm_display_mode *mode) +{ + uint16_t width, height; + uint16_t h_blank_len, h_sync_len, v_blank_len, v_sync_len; + uint16_t h_sync_offset, v_sync_offset; + + width = mode->crtc_hdisplay; + height = mode->crtc_vdisplay; + + /* do some mode translations */ + h_blank_len = mode->crtc_hblank_end - mode->crtc_hblank_start; + h_sync_len = mode->crtc_hsync_end - mode->crtc_hsync_start; + + v_blank_len = mode->crtc_vblank_end - mode->crtc_vblank_start; + v_sync_len = mode->crtc_vsync_end - mode->crtc_vsync_start; + + h_sync_offset = mode->crtc_hsync_start - mode->crtc_hblank_start; + v_sync_offset = mode->crtc_vsync_start - mode->crtc_vblank_start; + + dtd->part1.clock = mode->clock / 10; + dtd->part1.h_active = width & 0xff; + dtd->part1.h_blank = h_blank_len & 0xff; + dtd->part1.h_high = (((width >> 8) & 0xf) << 4) | + ((h_blank_len >> 8) & 0xf); + dtd->part1.v_active = height & 0xff; + dtd->part1.v_blank = v_blank_len & 0xff; + dtd->part1.v_high = (((height >> 8) & 0xf) << 4) | + ((v_blank_len >> 8) & 0xf); + + dtd->part2.h_sync_off = h_sync_offset & 0xff; + dtd->part2.h_sync_width = h_sync_len & 0xff; + dtd->part2.v_sync_off_width = (v_sync_offset & 0xf) << 4 | + (v_sync_len & 0xf); + dtd->part2.sync_off_width_high = ((h_sync_offset & 0x300) >> 2) | + ((h_sync_len & 0x300) >> 4) | ((v_sync_offset & 0x30) >> 2) | + ((v_sync_len & 0x30) >> 4); + + dtd->part2.dtd_flags = 0x18; + if (mode->flags & DRM_MODE_FLAG_PHSYNC) + dtd->part2.dtd_flags |= 0x2; + if (mode->flags & DRM_MODE_FLAG_PVSYNC) + dtd->part2.dtd_flags |= 0x4; + + dtd->part2.sdvo_flags = 0; + dtd->part2.v_sync_off_high = v_sync_offset & 0xc0; + dtd->part2.reserved = 0; +} + +static void psb_intel_sdvo_get_mode_from_dtd(struct drm_display_mode * mode, + const struct psb_intel_sdvo_dtd *dtd) +{ + mode->hdisplay = dtd->part1.h_active; + mode->hdisplay += ((dtd->part1.h_high >> 4) & 0x0f) << 8; + mode->hsync_start = mode->hdisplay + dtd->part2.h_sync_off; + mode->hsync_start += (dtd->part2.sync_off_width_high & 0xc0) << 2; + mode->hsync_end = mode->hsync_start + dtd->part2.h_sync_width; + mode->hsync_end += (dtd->part2.sync_off_width_high & 0x30) << 4; + mode->htotal = mode->hdisplay + dtd->part1.h_blank; + mode->htotal += (dtd->part1.h_high & 0xf) << 8; + + mode->vdisplay = dtd->part1.v_active; + mode->vdisplay += ((dtd->part1.v_high >> 4) & 0x0f) << 8; + mode->vsync_start = mode->vdisplay; + mode->vsync_start += (dtd->part2.v_sync_off_width >> 4) & 0xf; + mode->vsync_start += (dtd->part2.sync_off_width_high & 0x0c) << 2; + mode->vsync_start += dtd->part2.v_sync_off_high & 0xc0; + mode->vsync_end = mode->vsync_start + + (dtd->part2.v_sync_off_width & 0xf); + mode->vsync_end += (dtd->part2.sync_off_width_high & 0x3) << 4; + mode->vtotal = mode->vdisplay + dtd->part1.v_blank; + mode->vtotal += (dtd->part1.v_high & 0xf) << 8; + + mode->clock = dtd->part1.clock * 10; + + mode->flags &= ~(DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC); + if (dtd->part2.dtd_flags & 0x2) + mode->flags |= DRM_MODE_FLAG_PHSYNC; + if (dtd->part2.dtd_flags & 0x4) + mode->flags |= DRM_MODE_FLAG_PVSYNC; +} + +static bool psb_intel_sdvo_check_supp_encode(struct psb_intel_sdvo *psb_intel_sdvo) +{ + struct psb_intel_sdvo_encode encode; + + BUILD_BUG_ON(sizeof(encode) != 2); + return psb_intel_sdvo_get_value(psb_intel_sdvo, + SDVO_CMD_GET_SUPP_ENCODE, + &encode, sizeof(encode)); +} + +static bool psb_intel_sdvo_set_encode(struct psb_intel_sdvo *psb_intel_sdvo, + uint8_t mode) +{ + return psb_intel_sdvo_set_value(psb_intel_sdvo, SDVO_CMD_SET_ENCODE, &mode, 1); +} + +static bool psb_intel_sdvo_set_colorimetry(struct psb_intel_sdvo *psb_intel_sdvo, + uint8_t mode) +{ + return psb_intel_sdvo_set_value(psb_intel_sdvo, SDVO_CMD_SET_COLORIMETRY, &mode, 1); +} + +#if 0 +static void psb_intel_sdvo_dump_hdmi_buf(struct psb_intel_sdvo *psb_intel_sdvo) +{ + int i, j; + uint8_t set_buf_index[2]; + uint8_t av_split; + uint8_t buf_size; + uint8_t buf[48]; + uint8_t *pos; + + psb_intel_sdvo_get_value(encoder, SDVO_CMD_GET_HBUF_AV_SPLIT, &av_split, 1); + + for (i = 0; i <= av_split; i++) { + set_buf_index[0] = i; set_buf_index[1] = 0; + psb_intel_sdvo_write_cmd(encoder, SDVO_CMD_SET_HBUF_INDEX, + set_buf_index, 2); + psb_intel_sdvo_write_cmd(encoder, SDVO_CMD_GET_HBUF_INFO, NULL, 0); + psb_intel_sdvo_read_response(encoder, &buf_size, 1); + + pos = buf; + for (j = 0; j <= buf_size; j += 8) { + psb_intel_sdvo_write_cmd(encoder, SDVO_CMD_GET_HBUF_DATA, + NULL, 0); + psb_intel_sdvo_read_response(encoder, pos, 8); + pos += 8; + } + } +} +#endif + +static bool psb_intel_sdvo_set_avi_infoframe(struct psb_intel_sdvo *psb_intel_sdvo) +{ + DRM_INFO("HDMI is not supported yet"); + + return false; +#if 0 + struct dip_infoframe avi_if = { + .type = DIP_TYPE_AVI, + .ver = DIP_VERSION_AVI, + .len = DIP_LEN_AVI, + }; + uint8_t tx_rate = SDVO_HBUF_TX_VSYNC; + uint8_t set_buf_index[2] = { 1, 0 }; + uint64_t *data = (uint64_t *)&avi_if; + unsigned i; + + intel_dip_infoframe_csum(&avi_if); + + if (!psb_intel_sdvo_set_value(psb_intel_sdvo, + SDVO_CMD_SET_HBUF_INDEX, + set_buf_index, 2)) + return false; + + for (i = 0; i < sizeof(avi_if); i += 8) { + if (!psb_intel_sdvo_set_value(psb_intel_sdvo, + SDVO_CMD_SET_HBUF_DATA, + data, 8)) + return false; + data++; + } + + return psb_intel_sdvo_set_value(psb_intel_sdvo, + SDVO_CMD_SET_HBUF_TXRATE, + &tx_rate, 1); +#endif +} + +static bool psb_intel_sdvo_set_tv_format(struct psb_intel_sdvo *psb_intel_sdvo) +{ + struct psb_intel_sdvo_tv_format format; + uint32_t format_map; + + format_map = 1 << psb_intel_sdvo->tv_format_index; + memset(&format, 0, sizeof(format)); + memcpy(&format, &format_map, min(sizeof(format), sizeof(format_map))); + + BUILD_BUG_ON(sizeof(format) != 6); + return psb_intel_sdvo_set_value(psb_intel_sdvo, + SDVO_CMD_SET_TV_FORMAT, + &format, sizeof(format)); +} + +static bool +psb_intel_sdvo_set_output_timings_from_mode(struct psb_intel_sdvo *psb_intel_sdvo, + struct drm_display_mode *mode) +{ + struct psb_intel_sdvo_dtd output_dtd; + + if (!psb_intel_sdvo_set_target_output(psb_intel_sdvo, + psb_intel_sdvo->attached_output)) + return false; + + psb_intel_sdvo_get_dtd_from_mode(&output_dtd, mode); + if (!psb_intel_sdvo_set_output_timing(psb_intel_sdvo, &output_dtd)) + return false; + + return true; +} + +static bool +psb_intel_sdvo_set_input_timings_for_mode(struct psb_intel_sdvo *psb_intel_sdvo, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + /* Reset the input timing to the screen. Assume always input 0. */ + if (!psb_intel_sdvo_set_target_input(psb_intel_sdvo)) + return false; + + if (!psb_intel_sdvo_create_preferred_input_timing(psb_intel_sdvo, + mode->clock / 10, + mode->hdisplay, + mode->vdisplay)) + return false; + + if (!psb_intel_sdvo_get_preferred_input_timing(psb_intel_sdvo, + &psb_intel_sdvo->input_dtd)) + return false; + + psb_intel_sdvo_get_mode_from_dtd(adjusted_mode, &psb_intel_sdvo->input_dtd); + + drm_mode_set_crtcinfo(adjusted_mode, 0); + return true; +} + +static bool psb_intel_sdvo_mode_fixup(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + struct psb_intel_sdvo *psb_intel_sdvo = to_psb_intel_sdvo(encoder); + int multiplier; + + /* We need to construct preferred input timings based on our + * output timings. To do that, we have to set the output + * timings, even though this isn't really the right place in + * the sequence to do it. Oh well. + */ + if (psb_intel_sdvo->is_tv) { + if (!psb_intel_sdvo_set_output_timings_from_mode(psb_intel_sdvo, mode)) + return false; + + (void) psb_intel_sdvo_set_input_timings_for_mode(psb_intel_sdvo, + mode, + adjusted_mode); + } else if (psb_intel_sdvo->is_lvds) { + if (!psb_intel_sdvo_set_output_timings_from_mode(psb_intel_sdvo, + psb_intel_sdvo->sdvo_lvds_fixed_mode)) + return false; + + (void) psb_intel_sdvo_set_input_timings_for_mode(psb_intel_sdvo, + mode, + adjusted_mode); + } + + /* Make the CRTC code factor in the SDVO pixel multiplier. The + * SDVO device will factor out the multiplier during mode_set. + */ + multiplier = psb_intel_sdvo_get_pixel_multiplier(adjusted_mode); + psb_intel_mode_set_pixel_multiplier(adjusted_mode, multiplier); + + return true; +} + +static void psb_intel_sdvo_mode_set(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + struct drm_device *dev = encoder->dev; + struct drm_crtc *crtc = encoder->crtc; + struct psb_intel_crtc *psb_intel_crtc = to_psb_intel_crtc(crtc); + struct psb_intel_sdvo *psb_intel_sdvo = to_psb_intel_sdvo(encoder); + u32 sdvox; + struct psb_intel_sdvo_in_out_map in_out; + struct psb_intel_sdvo_dtd input_dtd; + int pixel_multiplier = psb_intel_mode_get_pixel_multiplier(adjusted_mode); + int rate; + + if (!mode) + return; + + /* First, set the input mapping for the first input to our controlled + * output. This is only correct if we're a single-input device, in + * which case the first input is the output from the appropriate SDVO + * channel on the motherboard. In a two-input device, the first input + * will be SDVOB and the second SDVOC. + */ + in_out.in0 = psb_intel_sdvo->attached_output; + in_out.in1 = 0; + + psb_intel_sdvo_set_value(psb_intel_sdvo, + SDVO_CMD_SET_IN_OUT_MAP, + &in_out, sizeof(in_out)); + + /* Set the output timings to the screen */ + if (!psb_intel_sdvo_set_target_output(psb_intel_sdvo, + psb_intel_sdvo->attached_output)) + return; + + /* We have tried to get input timing in mode_fixup, and filled into + * adjusted_mode. + */ + if (psb_intel_sdvo->is_tv || psb_intel_sdvo->is_lvds) { + input_dtd = psb_intel_sdvo->input_dtd; + } else { + /* Set the output timing to the screen */ + if (!psb_intel_sdvo_set_target_output(psb_intel_sdvo, + psb_intel_sdvo->attached_output)) + return; + + psb_intel_sdvo_get_dtd_from_mode(&input_dtd, adjusted_mode); + (void) psb_intel_sdvo_set_output_timing(psb_intel_sdvo, &input_dtd); + } + + /* Set the input timing to the screen. Assume always input 0. */ + if (!psb_intel_sdvo_set_target_input(psb_intel_sdvo)) + return; + + if (psb_intel_sdvo->has_hdmi_monitor) { + psb_intel_sdvo_set_encode(psb_intel_sdvo, SDVO_ENCODE_HDMI); + psb_intel_sdvo_set_colorimetry(psb_intel_sdvo, + SDVO_COLORIMETRY_RGB256); + psb_intel_sdvo_set_avi_infoframe(psb_intel_sdvo); + } else + psb_intel_sdvo_set_encode(psb_intel_sdvo, SDVO_ENCODE_DVI); + + if (psb_intel_sdvo->is_tv && + !psb_intel_sdvo_set_tv_format(psb_intel_sdvo)) + return; + + (void) psb_intel_sdvo_set_input_timing(psb_intel_sdvo, &input_dtd); + + switch (pixel_multiplier) { + default: + case 1: rate = SDVO_CLOCK_RATE_MULT_1X; break; + case 2: rate = SDVO_CLOCK_RATE_MULT_2X; break; + case 4: rate = SDVO_CLOCK_RATE_MULT_4X; break; + } + if (!psb_intel_sdvo_set_clock_rate_mult(psb_intel_sdvo, rate)) + return; + + /* Set the SDVO control regs. */ + sdvox = REG_READ(psb_intel_sdvo->sdvo_reg); + switch (psb_intel_sdvo->sdvo_reg) { + case SDVOB: + sdvox &= SDVOB_PRESERVE_MASK; + break; + case SDVOC: + sdvox &= SDVOC_PRESERVE_MASK; + break; + } + sdvox |= (9 << 19) | SDVO_BORDER_ENABLE; + + if (psb_intel_crtc->pipe == 1) + sdvox |= SDVO_PIPE_B_SELECT; + if (psb_intel_sdvo->has_hdmi_audio) + sdvox |= SDVO_AUDIO_ENABLE; + + /* FIXME: Check if this is needed for PSB + sdvox |= (pixel_multiplier - 1) << SDVO_PORT_MULTIPLY_SHIFT; + */ + + if (input_dtd.part2.sdvo_flags & SDVO_NEED_TO_STALL) + sdvox |= SDVO_STALL_SELECT; + psb_intel_sdvo_write_sdvox(psb_intel_sdvo, sdvox); +} + +static void psb_intel_sdvo_dpms(struct drm_encoder *encoder, int mode) +{ + struct drm_device *dev = encoder->dev; + struct psb_intel_sdvo *psb_intel_sdvo = to_psb_intel_sdvo(encoder); + u32 temp; + + switch (mode) { + case DRM_MODE_DPMS_ON: + DRM_DEBUG("DPMS_ON"); + break; + case DRM_MODE_DPMS_OFF: + DRM_DEBUG("DPMS_OFF"); + break; + default: + DRM_DEBUG("DPMS: %d", mode); + } + + if (mode != DRM_MODE_DPMS_ON) { + psb_intel_sdvo_set_active_outputs(psb_intel_sdvo, 0); + if (0) + psb_intel_sdvo_set_encoder_power_state(psb_intel_sdvo, mode); + + if (mode == DRM_MODE_DPMS_OFF) { + temp = REG_READ(psb_intel_sdvo->sdvo_reg); + if ((temp & SDVO_ENABLE) != 0) { + psb_intel_sdvo_write_sdvox(psb_intel_sdvo, temp & ~SDVO_ENABLE); + } + } + } else { + bool input1, input2; + int i; + u8 status; + + temp = REG_READ(psb_intel_sdvo->sdvo_reg); + if ((temp & SDVO_ENABLE) == 0) + psb_intel_sdvo_write_sdvox(psb_intel_sdvo, temp | SDVO_ENABLE); + for (i = 0; i < 2; i++) + psb_intel_wait_for_vblank(dev); + + status = psb_intel_sdvo_get_trained_inputs(psb_intel_sdvo, &input1, &input2); + /* Warn if the device reported failure to sync. + * A lot of SDVO devices fail to notify of sync, but it's + * a given it the status is a success, we succeeded. + */ + if (status == SDVO_CMD_STATUS_SUCCESS && !input1) { + DRM_DEBUG_KMS("First %s output reported failure to " + "sync\n", SDVO_NAME(psb_intel_sdvo)); + } + + if (0) + psb_intel_sdvo_set_encoder_power_state(psb_intel_sdvo, mode); + psb_intel_sdvo_set_active_outputs(psb_intel_sdvo, psb_intel_sdvo->attached_output); + } + return; +} + +static int psb_intel_sdvo_mode_valid(struct drm_connector *connector, + struct drm_display_mode *mode) +{ + struct drm_psb_private *dev_priv = connector->dev->dev_private; + struct psb_intel_sdvo *psb_intel_sdvo = intel_attached_sdvo(connector); + + if (mode->flags & DRM_MODE_FLAG_DBLSCAN) + return MODE_NO_DBLESCAN; + + if (psb_intel_sdvo->pixel_clock_min > mode->clock) + return MODE_CLOCK_LOW; + + if (psb_intel_sdvo->pixel_clock_max < mode->clock) + return MODE_CLOCK_HIGH; + + if (psb_intel_sdvo->is_lvds) { + if (mode->hdisplay > psb_intel_sdvo->sdvo_lvds_fixed_mode->hdisplay) + return MODE_PANEL; + + if (mode->vdisplay > psb_intel_sdvo->sdvo_lvds_fixed_mode->vdisplay) + return MODE_PANEL; + } + + /* We assume worst case scenario of 32 bpp here, since we don't know */ + if ((ALIGN(mode->hdisplay * 4, 64) * mode->vdisplay) > + dev_priv->vram_stolen_size) + return MODE_MEM; + + return MODE_OK; +} + +static bool psb_intel_sdvo_get_capabilities(struct psb_intel_sdvo *psb_intel_sdvo, struct psb_intel_sdvo_caps *caps) +{ + BUILD_BUG_ON(sizeof(*caps) != 8); + if (!psb_intel_sdvo_get_value(psb_intel_sdvo, + SDVO_CMD_GET_DEVICE_CAPS, + caps, sizeof(*caps))) + return false; + + DRM_DEBUG_KMS("SDVO capabilities:\n" + " vendor_id: %d\n" + " device_id: %d\n" + " device_rev_id: %d\n" + " sdvo_version_major: %d\n" + " sdvo_version_minor: %d\n" + " sdvo_inputs_mask: %d\n" + " smooth_scaling: %d\n" + " sharp_scaling: %d\n" + " up_scaling: %d\n" + " down_scaling: %d\n" + " stall_support: %d\n" + " output_flags: %d\n", + caps->vendor_id, + caps->device_id, + caps->device_rev_id, + caps->sdvo_version_major, + caps->sdvo_version_minor, + caps->sdvo_inputs_mask, + caps->smooth_scaling, + caps->sharp_scaling, + caps->up_scaling, + caps->down_scaling, + caps->stall_support, + caps->output_flags); + + return true; +} + +/* No use! */ +#if 0 +struct drm_connector* psb_intel_sdvo_find(struct drm_device *dev, int sdvoB) +{ + struct drm_connector *connector = NULL; + struct psb_intel_sdvo *iout = NULL; + struct psb_intel_sdvo *sdvo; + + /* find the sdvo connector */ + list_for_each_entry(connector, &dev->mode_config.connector_list, head) { + iout = to_psb_intel_sdvo(connector); + + if (iout->type != INTEL_OUTPUT_SDVO) + continue; + + sdvo = iout->dev_priv; + + if (sdvo->sdvo_reg == SDVOB && sdvoB) + return connector; + + if (sdvo->sdvo_reg == SDVOC && !sdvoB) + return connector; + + } + + return NULL; +} + +int psb_intel_sdvo_supports_hotplug(struct drm_connector *connector) +{ + u8 response[2]; + u8 status; + struct psb_intel_sdvo *psb_intel_sdvo; + DRM_DEBUG_KMS("\n"); + + if (!connector) + return 0; + + psb_intel_sdvo = to_psb_intel_sdvo(connector); + + return psb_intel_sdvo_get_value(psb_intel_sdvo, SDVO_CMD_GET_HOT_PLUG_SUPPORT, + &response, 2) && response[0]; +} + +void psb_intel_sdvo_set_hotplug(struct drm_connector *connector, int on) +{ + u8 response[2]; + u8 status; + struct psb_intel_sdvo *psb_intel_sdvo = to_psb_intel_sdvo(connector); + + psb_intel_sdvo_write_cmd(psb_intel_sdvo, SDVO_CMD_GET_ACTIVE_HOT_PLUG, NULL, 0); + psb_intel_sdvo_read_response(psb_intel_sdvo, &response, 2); + + if (on) { + psb_intel_sdvo_write_cmd(psb_intel_sdvo, SDVO_CMD_GET_HOT_PLUG_SUPPORT, NULL, 0); + status = psb_intel_sdvo_read_response(psb_intel_sdvo, &response, 2); + + psb_intel_sdvo_write_cmd(psb_intel_sdvo, SDVO_CMD_SET_ACTIVE_HOT_PLUG, &response, 2); + } else { + response[0] = 0; + response[1] = 0; + psb_intel_sdvo_write_cmd(psb_intel_sdvo, SDVO_CMD_SET_ACTIVE_HOT_PLUG, &response, 2); + } + + psb_intel_sdvo_write_cmd(psb_intel_sdvo, SDVO_CMD_GET_ACTIVE_HOT_PLUG, NULL, 0); + psb_intel_sdvo_read_response(psb_intel_sdvo, &response, 2); +} +#endif + +static bool +psb_intel_sdvo_multifunc_encoder(struct psb_intel_sdvo *psb_intel_sdvo) +{ + /* Is there more than one type of output? */ + int caps = psb_intel_sdvo->caps.output_flags & 0xf; + return caps & -caps; +} + +static struct edid * +psb_intel_sdvo_get_edid(struct drm_connector *connector) +{ + struct psb_intel_sdvo *sdvo = intel_attached_sdvo(connector); + return drm_get_edid(connector, &sdvo->ddc); +} + +/* Mac mini hack -- use the same DDC as the analog connector */ +static struct edid * +psb_intel_sdvo_get_analog_edid(struct drm_connector *connector) +{ + struct drm_psb_private *dev_priv = connector->dev->dev_private; + + return drm_get_edid(connector, + &dev_priv->gmbus[dev_priv->crt_ddc_pin].adapter); + return NULL; +} + +static enum drm_connector_status +psb_intel_sdvo_hdmi_sink_detect(struct drm_connector *connector) +{ + struct psb_intel_sdvo *psb_intel_sdvo = intel_attached_sdvo(connector); + enum drm_connector_status status; + struct edid *edid; + + edid = psb_intel_sdvo_get_edid(connector); + + if (edid == NULL && psb_intel_sdvo_multifunc_encoder(psb_intel_sdvo)) { + u8 ddc, saved_ddc = psb_intel_sdvo->ddc_bus; + + /* + * Don't use the 1 as the argument of DDC bus switch to get + * the EDID. It is used for SDVO SPD ROM. + */ + for (ddc = psb_intel_sdvo->ddc_bus >> 1; ddc > 1; ddc >>= 1) { + psb_intel_sdvo->ddc_bus = ddc; + edid = psb_intel_sdvo_get_edid(connector); + if (edid) + break; + } + /* + * If we found the EDID on the other bus, + * assume that is the correct DDC bus. + */ + if (edid == NULL) + psb_intel_sdvo->ddc_bus = saved_ddc; + } + + /* + * When there is no edid and no monitor is connected with VGA + * port, try to use the CRT ddc to read the EDID for DVI-connector. + */ + if (edid == NULL) + edid = psb_intel_sdvo_get_analog_edid(connector); + + status = connector_status_unknown; + if (edid != NULL) { + /* DDC bus is shared, match EDID to connector type */ + if (edid->input & DRM_EDID_INPUT_DIGITAL) { + status = connector_status_connected; + if (psb_intel_sdvo->is_hdmi) { + psb_intel_sdvo->has_hdmi_monitor = drm_detect_hdmi_monitor(edid); + psb_intel_sdvo->has_hdmi_audio = drm_detect_monitor_audio(edid); + } + } else + status = connector_status_disconnected; + connector->display_info.raw_edid = NULL; + kfree(edid); + } + + if (status == connector_status_connected) { + struct psb_intel_sdvo_connector *psb_intel_sdvo_connector = to_psb_intel_sdvo_connector(connector); + if (psb_intel_sdvo_connector->force_audio) + psb_intel_sdvo->has_hdmi_audio = psb_intel_sdvo_connector->force_audio > 0; + } + + return status; +} + +static enum drm_connector_status +psb_intel_sdvo_detect(struct drm_connector *connector, bool force) +{ + uint16_t response; + struct psb_intel_sdvo *psb_intel_sdvo = intel_attached_sdvo(connector); + struct psb_intel_sdvo_connector *psb_intel_sdvo_connector = to_psb_intel_sdvo_connector(connector); + enum drm_connector_status ret; + + if (!psb_intel_sdvo_write_cmd(psb_intel_sdvo, + SDVO_CMD_GET_ATTACHED_DISPLAYS, NULL, 0)) + return connector_status_unknown; + + /* add 30ms delay when the output type might be TV */ + if (psb_intel_sdvo->caps.output_flags & + (SDVO_OUTPUT_SVID0 | SDVO_OUTPUT_CVBS0)) + mdelay(30); + + if (!psb_intel_sdvo_read_response(psb_intel_sdvo, &response, 2)) + return connector_status_unknown; + + DRM_DEBUG_KMS("SDVO response %d %d [%x]\n", + response & 0xff, response >> 8, + psb_intel_sdvo_connector->output_flag); + + if (response == 0) + return connector_status_disconnected; + + psb_intel_sdvo->attached_output = response; + + psb_intel_sdvo->has_hdmi_monitor = false; + psb_intel_sdvo->has_hdmi_audio = false; + + if ((psb_intel_sdvo_connector->output_flag & response) == 0) + ret = connector_status_disconnected; + else if (IS_TMDS(psb_intel_sdvo_connector)) + ret = psb_intel_sdvo_hdmi_sink_detect(connector); + else { + struct edid *edid; + + /* if we have an edid check it matches the connection */ + edid = psb_intel_sdvo_get_edid(connector); + if (edid == NULL) + edid = psb_intel_sdvo_get_analog_edid(connector); + if (edid != NULL) { + if (edid->input & DRM_EDID_INPUT_DIGITAL) + ret = connector_status_disconnected; + else + ret = connector_status_connected; + connector->display_info.raw_edid = NULL; + kfree(edid); + } else + ret = connector_status_connected; + } + + /* May update encoder flag for like clock for SDVO TV, etc.*/ + if (ret == connector_status_connected) { + psb_intel_sdvo->is_tv = false; + psb_intel_sdvo->is_lvds = false; + psb_intel_sdvo->base.needs_tv_clock = false; + + if (response & SDVO_TV_MASK) { + psb_intel_sdvo->is_tv = true; + psb_intel_sdvo->base.needs_tv_clock = true; + } + if (response & SDVO_LVDS_MASK) + psb_intel_sdvo->is_lvds = psb_intel_sdvo->sdvo_lvds_fixed_mode != NULL; + } + + return ret; +} + +static void psb_intel_sdvo_get_ddc_modes(struct drm_connector *connector) +{ + struct edid *edid; + + /* set the bus switch and get the modes */ + edid = psb_intel_sdvo_get_edid(connector); + + /* + * Mac mini hack. On this device, the DVI-I connector shares one DDC + * link between analog and digital outputs. So, if the regular SDVO + * DDC fails, check to see if the analog output is disconnected, in + * which case we'll look there for the digital DDC data. + */ + if (edid == NULL) + edid = psb_intel_sdvo_get_analog_edid(connector); + + if (edid != NULL) { + struct psb_intel_sdvo_connector *psb_intel_sdvo_connector = to_psb_intel_sdvo_connector(connector); + bool monitor_is_digital = !!(edid->input & DRM_EDID_INPUT_DIGITAL); + bool connector_is_digital = !!IS_TMDS(psb_intel_sdvo_connector); + + if (connector_is_digital == monitor_is_digital) { + drm_mode_connector_update_edid_property(connector, edid); + drm_add_edid_modes(connector, edid); + } + + connector->display_info.raw_edid = NULL; + kfree(edid); + } +} + +/* + * Set of SDVO TV modes. + * Note! This is in reply order (see loop in get_tv_modes). + * XXX: all 60Hz refresh? + */ +static const struct drm_display_mode sdvo_tv_modes[] = { + { DRM_MODE("320x200", DRM_MODE_TYPE_DRIVER, 5815, 320, 321, 384, + 416, 0, 200, 201, 232, 233, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + { DRM_MODE("320x240", DRM_MODE_TYPE_DRIVER, 6814, 320, 321, 384, + 416, 0, 240, 241, 272, 273, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + { DRM_MODE("400x300", DRM_MODE_TYPE_DRIVER, 9910, 400, 401, 464, + 496, 0, 300, 301, 332, 333, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + { DRM_MODE("640x350", DRM_MODE_TYPE_DRIVER, 16913, 640, 641, 704, + 736, 0, 350, 351, 382, 383, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + { DRM_MODE("640x400", DRM_MODE_TYPE_DRIVER, 19121, 640, 641, 704, + 736, 0, 400, 401, 432, 433, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + { DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 22654, 640, 641, 704, + 736, 0, 480, 481, 512, 513, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + { DRM_MODE("704x480", DRM_MODE_TYPE_DRIVER, 24624, 704, 705, 768, + 800, 0, 480, 481, 512, 513, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + { DRM_MODE("704x576", DRM_MODE_TYPE_DRIVER, 29232, 704, 705, 768, + 800, 0, 576, 577, 608, 609, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + { DRM_MODE("720x350", DRM_MODE_TYPE_DRIVER, 18751, 720, 721, 784, + 816, 0, 350, 351, 382, 383, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + { DRM_MODE("720x400", DRM_MODE_TYPE_DRIVER, 21199, 720, 721, 784, + 816, 0, 400, 401, 432, 433, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + { DRM_MODE("720x480", DRM_MODE_TYPE_DRIVER, 25116, 720, 721, 784, + 816, 0, 480, 481, 512, 513, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + { DRM_MODE("720x540", DRM_MODE_TYPE_DRIVER, 28054, 720, 721, 784, + 816, 0, 540, 541, 572, 573, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + { DRM_MODE("720x576", DRM_MODE_TYPE_DRIVER, 29816, 720, 721, 784, + 816, 0, 576, 577, 608, 609, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + { DRM_MODE("768x576", DRM_MODE_TYPE_DRIVER, 31570, 768, 769, 832, + 864, 0, 576, 577, 608, 609, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + { DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 34030, 800, 801, 864, + 896, 0, 600, 601, 632, 633, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + { DRM_MODE("832x624", DRM_MODE_TYPE_DRIVER, 36581, 832, 833, 896, + 928, 0, 624, 625, 656, 657, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + { DRM_MODE("920x766", DRM_MODE_TYPE_DRIVER, 48707, 920, 921, 984, + 1016, 0, 766, 767, 798, 799, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + { DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 53827, 1024, 1025, 1088, + 1120, 0, 768, 769, 800, 801, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, + { DRM_MODE("1280x1024", DRM_MODE_TYPE_DRIVER, 87265, 1280, 1281, 1344, + 1376, 0, 1024, 1025, 1056, 1057, 0, + DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) }, +}; + +static void psb_intel_sdvo_get_tv_modes(struct drm_connector *connector) +{ + struct psb_intel_sdvo *psb_intel_sdvo = intel_attached_sdvo(connector); + struct psb_intel_sdvo_sdtv_resolution_request tv_res; + uint32_t reply = 0, format_map = 0; + int i; + + /* Read the list of supported input resolutions for the selected TV + * format. + */ + format_map = 1 << psb_intel_sdvo->tv_format_index; + memcpy(&tv_res, &format_map, + min(sizeof(format_map), sizeof(struct psb_intel_sdvo_sdtv_resolution_request))); + + if (!psb_intel_sdvo_set_target_output(psb_intel_sdvo, psb_intel_sdvo->attached_output)) + return; + + BUILD_BUG_ON(sizeof(tv_res) != 3); + if (!psb_intel_sdvo_write_cmd(psb_intel_sdvo, + SDVO_CMD_GET_SDTV_RESOLUTION_SUPPORT, + &tv_res, sizeof(tv_res))) + return; + if (!psb_intel_sdvo_read_response(psb_intel_sdvo, &reply, 3)) + return; + + for (i = 0; i < ARRAY_SIZE(sdvo_tv_modes); i++) + if (reply & (1 << i)) { + struct drm_display_mode *nmode; + nmode = drm_mode_duplicate(connector->dev, + &sdvo_tv_modes[i]); + if (nmode) + drm_mode_probed_add(connector, nmode); + } +} + +static void psb_intel_sdvo_get_lvds_modes(struct drm_connector *connector) +{ + struct psb_intel_sdvo *psb_intel_sdvo = intel_attached_sdvo(connector); + struct drm_psb_private *dev_priv = connector->dev->dev_private; + struct drm_display_mode *newmode; + + /* + * Attempt to get the mode list from DDC. + * Assume that the preferred modes are + * arranged in priority order. + */ + psb_intel_ddc_get_modes(connector, psb_intel_sdvo->i2c); + if (list_empty(&connector->probed_modes) == false) + goto end; + + /* Fetch modes from VBT */ + if (dev_priv->sdvo_lvds_vbt_mode != NULL) { + newmode = drm_mode_duplicate(connector->dev, + dev_priv->sdvo_lvds_vbt_mode); + if (newmode != NULL) { + /* Guarantee the mode is preferred */ + newmode->type = (DRM_MODE_TYPE_PREFERRED | + DRM_MODE_TYPE_DRIVER); + drm_mode_probed_add(connector, newmode); + } + } + +end: + list_for_each_entry(newmode, &connector->probed_modes, head) { + if (newmode->type & DRM_MODE_TYPE_PREFERRED) { + psb_intel_sdvo->sdvo_lvds_fixed_mode = + drm_mode_duplicate(connector->dev, newmode); + + drm_mode_set_crtcinfo(psb_intel_sdvo->sdvo_lvds_fixed_mode, + 0); + + psb_intel_sdvo->is_lvds = true; + break; + } + } + +} + +static int psb_intel_sdvo_get_modes(struct drm_connector *connector) +{ + struct psb_intel_sdvo_connector *psb_intel_sdvo_connector = to_psb_intel_sdvo_connector(connector); + + if (IS_TV(psb_intel_sdvo_connector)) + psb_intel_sdvo_get_tv_modes(connector); + else if (IS_LVDS(psb_intel_sdvo_connector)) + psb_intel_sdvo_get_lvds_modes(connector); + else + psb_intel_sdvo_get_ddc_modes(connector); + + return !list_empty(&connector->probed_modes); +} + +static void +psb_intel_sdvo_destroy_enhance_property(struct drm_connector *connector) +{ + struct psb_intel_sdvo_connector *psb_intel_sdvo_connector = to_psb_intel_sdvo_connector(connector); + struct drm_device *dev = connector->dev; + + if (psb_intel_sdvo_connector->left) + drm_property_destroy(dev, psb_intel_sdvo_connector->left); + if (psb_intel_sdvo_connector->right) + drm_property_destroy(dev, psb_intel_sdvo_connector->right); + if (psb_intel_sdvo_connector->top) + drm_property_destroy(dev, psb_intel_sdvo_connector->top); + if (psb_intel_sdvo_connector->bottom) + drm_property_destroy(dev, psb_intel_sdvo_connector->bottom); + if (psb_intel_sdvo_connector->hpos) + drm_property_destroy(dev, psb_intel_sdvo_connector->hpos); + if (psb_intel_sdvo_connector->vpos) + drm_property_destroy(dev, psb_intel_sdvo_connector->vpos); + if (psb_intel_sdvo_connector->saturation) + drm_property_destroy(dev, psb_intel_sdvo_connector->saturation); + if (psb_intel_sdvo_connector->contrast) + drm_property_destroy(dev, psb_intel_sdvo_connector->contrast); + if (psb_intel_sdvo_connector->hue) + drm_property_destroy(dev, psb_intel_sdvo_connector->hue); + if (psb_intel_sdvo_connector->sharpness) + drm_property_destroy(dev, psb_intel_sdvo_connector->sharpness); + if (psb_intel_sdvo_connector->flicker_filter) + drm_property_destroy(dev, psb_intel_sdvo_connector->flicker_filter); + if (psb_intel_sdvo_connector->flicker_filter_2d) + drm_property_destroy(dev, psb_intel_sdvo_connector->flicker_filter_2d); + if (psb_intel_sdvo_connector->flicker_filter_adaptive) + drm_property_destroy(dev, psb_intel_sdvo_connector->flicker_filter_adaptive); + if (psb_intel_sdvo_connector->tv_luma_filter) + drm_property_destroy(dev, psb_intel_sdvo_connector->tv_luma_filter); + if (psb_intel_sdvo_connector->tv_chroma_filter) + drm_property_destroy(dev, psb_intel_sdvo_connector->tv_chroma_filter); + if (psb_intel_sdvo_connector->dot_crawl) + drm_property_destroy(dev, psb_intel_sdvo_connector->dot_crawl); + if (psb_intel_sdvo_connector->brightness) + drm_property_destroy(dev, psb_intel_sdvo_connector->brightness); +} + +static void psb_intel_sdvo_destroy(struct drm_connector *connector) +{ + struct psb_intel_sdvo_connector *psb_intel_sdvo_connector = to_psb_intel_sdvo_connector(connector); + + if (psb_intel_sdvo_connector->tv_format) + drm_property_destroy(connector->dev, + psb_intel_sdvo_connector->tv_format); + + psb_intel_sdvo_destroy_enhance_property(connector); + drm_sysfs_connector_remove(connector); + drm_connector_cleanup(connector); + kfree(connector); +} + +static bool psb_intel_sdvo_detect_hdmi_audio(struct drm_connector *connector) +{ + struct psb_intel_sdvo *psb_intel_sdvo = intel_attached_sdvo(connector); + struct edid *edid; + bool has_audio = false; + + if (!psb_intel_sdvo->is_hdmi) + return false; + + edid = psb_intel_sdvo_get_edid(connector); + if (edid != NULL && edid->input & DRM_EDID_INPUT_DIGITAL) + has_audio = drm_detect_monitor_audio(edid); + + return has_audio; +} + +static int +psb_intel_sdvo_set_property(struct drm_connector *connector, + struct drm_property *property, + uint64_t val) +{ + struct psb_intel_sdvo *psb_intel_sdvo = intel_attached_sdvo(connector); + struct psb_intel_sdvo_connector *psb_intel_sdvo_connector = to_psb_intel_sdvo_connector(connector); + struct drm_psb_private *dev_priv = connector->dev->dev_private; + uint16_t temp_value; + uint8_t cmd; + int ret; + + ret = drm_connector_property_set_value(connector, property, val); + if (ret) + return ret; + + if (property == dev_priv->force_audio_property) { + int i = val; + bool has_audio; + + if (i == psb_intel_sdvo_connector->force_audio) + return 0; + + psb_intel_sdvo_connector->force_audio = i; + + if (i == 0) + has_audio = psb_intel_sdvo_detect_hdmi_audio(connector); + else + has_audio = i > 0; + + if (has_audio == psb_intel_sdvo->has_hdmi_audio) + return 0; + + psb_intel_sdvo->has_hdmi_audio = has_audio; + goto done; + } + + if (property == dev_priv->broadcast_rgb_property) { + if (val == !!psb_intel_sdvo->color_range) + return 0; + + psb_intel_sdvo->color_range = val ? SDVO_COLOR_RANGE_16_235 : 0; + goto done; + } + +#define CHECK_PROPERTY(name, NAME) \ + if (psb_intel_sdvo_connector->name == property) { \ + if (psb_intel_sdvo_connector->cur_##name == temp_value) return 0; \ + if (psb_intel_sdvo_connector->max_##name < temp_value) return -EINVAL; \ + cmd = SDVO_CMD_SET_##NAME; \ + psb_intel_sdvo_connector->cur_##name = temp_value; \ + goto set_value; \ + } + + if (property == psb_intel_sdvo_connector->tv_format) { + if (val >= TV_FORMAT_NUM) + return -EINVAL; + + if (psb_intel_sdvo->tv_format_index == + psb_intel_sdvo_connector->tv_format_supported[val]) + return 0; + + psb_intel_sdvo->tv_format_index = psb_intel_sdvo_connector->tv_format_supported[val]; + goto done; + } else if (IS_TV_OR_LVDS(psb_intel_sdvo_connector)) { + temp_value = val; + if (psb_intel_sdvo_connector->left == property) { + drm_connector_property_set_value(connector, + psb_intel_sdvo_connector->right, val); + if (psb_intel_sdvo_connector->left_margin == temp_value) + return 0; + + psb_intel_sdvo_connector->left_margin = temp_value; + psb_intel_sdvo_connector->right_margin = temp_value; + temp_value = psb_intel_sdvo_connector->max_hscan - + psb_intel_sdvo_connector->left_margin; + cmd = SDVO_CMD_SET_OVERSCAN_H; + goto set_value; + } else if (psb_intel_sdvo_connector->right == property) { + drm_connector_property_set_value(connector, + psb_intel_sdvo_connector->left, val); + if (psb_intel_sdvo_connector->right_margin == temp_value) + return 0; + + psb_intel_sdvo_connector->left_margin = temp_value; + psb_intel_sdvo_connector->right_margin = temp_value; + temp_value = psb_intel_sdvo_connector->max_hscan - + psb_intel_sdvo_connector->left_margin; + cmd = SDVO_CMD_SET_OVERSCAN_H; + goto set_value; + } else if (psb_intel_sdvo_connector->top == property) { + drm_connector_property_set_value(connector, + psb_intel_sdvo_connector->bottom, val); + if (psb_intel_sdvo_connector->top_margin == temp_value) + return 0; + + psb_intel_sdvo_connector->top_margin = temp_value; + psb_intel_sdvo_connector->bottom_margin = temp_value; + temp_value = psb_intel_sdvo_connector->max_vscan - + psb_intel_sdvo_connector->top_margin; + cmd = SDVO_CMD_SET_OVERSCAN_V; + goto set_value; + } else if (psb_intel_sdvo_connector->bottom == property) { + drm_connector_property_set_value(connector, + psb_intel_sdvo_connector->top, val); + if (psb_intel_sdvo_connector->bottom_margin == temp_value) + return 0; + + psb_intel_sdvo_connector->top_margin = temp_value; + psb_intel_sdvo_connector->bottom_margin = temp_value; + temp_value = psb_intel_sdvo_connector->max_vscan - + psb_intel_sdvo_connector->top_margin; + cmd = SDVO_CMD_SET_OVERSCAN_V; + goto set_value; + } + CHECK_PROPERTY(hpos, HPOS) + CHECK_PROPERTY(vpos, VPOS) + CHECK_PROPERTY(saturation, SATURATION) + CHECK_PROPERTY(contrast, CONTRAST) + CHECK_PROPERTY(hue, HUE) + CHECK_PROPERTY(brightness, BRIGHTNESS) + CHECK_PROPERTY(sharpness, SHARPNESS) + CHECK_PROPERTY(flicker_filter, FLICKER_FILTER) + CHECK_PROPERTY(flicker_filter_2d, FLICKER_FILTER_2D) + CHECK_PROPERTY(flicker_filter_adaptive, FLICKER_FILTER_ADAPTIVE) + CHECK_PROPERTY(tv_chroma_filter, TV_CHROMA_FILTER) + CHECK_PROPERTY(tv_luma_filter, TV_LUMA_FILTER) + CHECK_PROPERTY(dot_crawl, DOT_CRAWL) + } + + return -EINVAL; /* unknown property */ + +set_value: + if (!psb_intel_sdvo_set_value(psb_intel_sdvo, cmd, &temp_value, 2)) + return -EIO; + + +done: + if (psb_intel_sdvo->base.base.crtc) { + struct drm_crtc *crtc = psb_intel_sdvo->base.base.crtc; + drm_crtc_helper_set_mode(crtc, &crtc->mode, crtc->x, + crtc->y, crtc->fb); + } + + return 0; +#undef CHECK_PROPERTY +} + +static const struct drm_encoder_helper_funcs psb_intel_sdvo_helper_funcs = { + .dpms = psb_intel_sdvo_dpms, + .mode_fixup = psb_intel_sdvo_mode_fixup, + .prepare = psb_intel_encoder_prepare, + .mode_set = psb_intel_sdvo_mode_set, + .commit = psb_intel_encoder_commit, +}; + +static const struct drm_connector_funcs psb_intel_sdvo_connector_funcs = { + .dpms = drm_helper_connector_dpms, + .detect = psb_intel_sdvo_detect, + .fill_modes = drm_helper_probe_single_connector_modes, + .set_property = psb_intel_sdvo_set_property, + .destroy = psb_intel_sdvo_destroy, +}; + +static const struct drm_connector_helper_funcs psb_intel_sdvo_connector_helper_funcs = { + .get_modes = psb_intel_sdvo_get_modes, + .mode_valid = psb_intel_sdvo_mode_valid, + .best_encoder = psb_intel_best_encoder, +}; + +static void psb_intel_sdvo_enc_destroy(struct drm_encoder *encoder) +{ + struct psb_intel_sdvo *psb_intel_sdvo = to_psb_intel_sdvo(encoder); + + if (psb_intel_sdvo->sdvo_lvds_fixed_mode != NULL) + drm_mode_destroy(encoder->dev, + psb_intel_sdvo->sdvo_lvds_fixed_mode); + + i2c_del_adapter(&psb_intel_sdvo->ddc); + psb_intel_encoder_destroy(encoder); +} + +static const struct drm_encoder_funcs psb_intel_sdvo_enc_funcs = { + .destroy = psb_intel_sdvo_enc_destroy, +}; + +static void +psb_intel_sdvo_guess_ddc_bus(struct psb_intel_sdvo *sdvo) +{ + /* FIXME: At the moment, ddc_bus = 2 is the only thing that works. + * We need to figure out if this is true for all available poulsbo + * hardware, or if we need to fiddle with the guessing code above. + * The problem might go away if we can parse sdvo mappings from bios */ + sdvo->ddc_bus = 2; + +#if 0 + uint16_t mask = 0; + unsigned int num_bits; + + /* Make a mask of outputs less than or equal to our own priority in the + * list. + */ + switch (sdvo->controlled_output) { + case SDVO_OUTPUT_LVDS1: + mask |= SDVO_OUTPUT_LVDS1; + case SDVO_OUTPUT_LVDS0: + mask |= SDVO_OUTPUT_LVDS0; + case SDVO_OUTPUT_TMDS1: + mask |= SDVO_OUTPUT_TMDS1; + case SDVO_OUTPUT_TMDS0: + mask |= SDVO_OUTPUT_TMDS0; + case SDVO_OUTPUT_RGB1: + mask |= SDVO_OUTPUT_RGB1; + case SDVO_OUTPUT_RGB0: + mask |= SDVO_OUTPUT_RGB0; + break; + } + + /* Count bits to find what number we are in the priority list. */ + mask &= sdvo->caps.output_flags; + num_bits = hweight16(mask); + /* If more than 3 outputs, default to DDC bus 3 for now. */ + if (num_bits > 3) + num_bits = 3; + + /* Corresponds to SDVO_CONTROL_BUS_DDCx */ + sdvo->ddc_bus = 1 << num_bits; +#endif +} + +/** + * Choose the appropriate DDC bus for control bus switch command for this + * SDVO output based on the controlled output. + * + * DDC bus number assignment is in a priority order of RGB outputs, then TMDS + * outputs, then LVDS outputs. + */ +static void +psb_intel_sdvo_select_ddc_bus(struct drm_psb_private *dev_priv, + struct psb_intel_sdvo *sdvo, u32 reg) +{ + struct sdvo_device_mapping *mapping; + + if (IS_SDVOB(reg)) + mapping = &(dev_priv->sdvo_mappings[0]); + else + mapping = &(dev_priv->sdvo_mappings[1]); + + if (mapping->initialized) + sdvo->ddc_bus = 1 << ((mapping->ddc_pin & 0xf0) >> 4); + else + psb_intel_sdvo_guess_ddc_bus(sdvo); +} + +static void +psb_intel_sdvo_select_i2c_bus(struct drm_psb_private *dev_priv, + struct psb_intel_sdvo *sdvo, u32 reg) +{ + struct sdvo_device_mapping *mapping; + u8 pin, speed; + + if (IS_SDVOB(reg)) + mapping = &dev_priv->sdvo_mappings[0]; + else + mapping = &dev_priv->sdvo_mappings[1]; + + pin = GMBUS_PORT_DPB; + speed = GMBUS_RATE_1MHZ >> 8; + if (mapping->initialized) { + pin = mapping->i2c_pin; + speed = mapping->i2c_speed; + } + + if (pin < GMBUS_NUM_PORTS) { + sdvo->i2c = &dev_priv->gmbus[pin].adapter; + gma_intel_gmbus_set_speed(sdvo->i2c, speed); + gma_intel_gmbus_force_bit(sdvo->i2c, true); + } else + sdvo->i2c = &dev_priv->gmbus[GMBUS_PORT_DPB].adapter; +} + +static bool +psb_intel_sdvo_is_hdmi_connector(struct psb_intel_sdvo *psb_intel_sdvo, int device) +{ + return psb_intel_sdvo_check_supp_encode(psb_intel_sdvo); +} + +static u8 +psb_intel_sdvo_get_slave_addr(struct drm_device *dev, int sdvo_reg) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct sdvo_device_mapping *my_mapping, *other_mapping; + + if (IS_SDVOB(sdvo_reg)) { + my_mapping = &dev_priv->sdvo_mappings[0]; + other_mapping = &dev_priv->sdvo_mappings[1]; + } else { + my_mapping = &dev_priv->sdvo_mappings[1]; + other_mapping = &dev_priv->sdvo_mappings[0]; + } + + /* If the BIOS described our SDVO device, take advantage of it. */ + if (my_mapping->slave_addr) + return my_mapping->slave_addr; + + /* If the BIOS only described a different SDVO device, use the + * address that it isn't using. + */ + if (other_mapping->slave_addr) { + if (other_mapping->slave_addr == 0x70) + return 0x72; + else + return 0x70; + } + + /* No SDVO device info is found for another DVO port, + * so use mapping assumption we had before BIOS parsing. + */ + if (IS_SDVOB(sdvo_reg)) + return 0x70; + else + return 0x72; +} + +static void +psb_intel_sdvo_connector_init(struct psb_intel_sdvo_connector *connector, + struct psb_intel_sdvo *encoder) +{ + drm_connector_init(encoder->base.base.dev, + &connector->base.base, + &psb_intel_sdvo_connector_funcs, + connector->base.base.connector_type); + + drm_connector_helper_add(&connector->base.base, + &psb_intel_sdvo_connector_helper_funcs); + + connector->base.base.interlace_allowed = 0; + connector->base.base.doublescan_allowed = 0; + connector->base.base.display_info.subpixel_order = SubPixelHorizontalRGB; + + psb_intel_connector_attach_encoder(&connector->base, &encoder->base); + drm_sysfs_connector_add(&connector->base.base); +} + +static void +psb_intel_sdvo_add_hdmi_properties(struct psb_intel_sdvo_connector *connector) +{ + /* FIXME: We don't support HDMI at the moment + struct drm_device *dev = connector->base.base.dev; + + intel_attach_force_audio_property(&connector->base.base); + if (INTEL_INFO(dev)->gen >= 4 && IS_MOBILE(dev)) + intel_attach_broadcast_rgb_property(&connector->base.base); + */ +} + +static bool +psb_intel_sdvo_dvi_init(struct psb_intel_sdvo *psb_intel_sdvo, int device) +{ + struct drm_encoder *encoder = &psb_intel_sdvo->base.base; + struct drm_connector *connector; + struct psb_intel_connector *intel_connector; + struct psb_intel_sdvo_connector *psb_intel_sdvo_connector; + + psb_intel_sdvo_connector = kzalloc(sizeof(struct psb_intel_sdvo_connector), GFP_KERNEL); + if (!psb_intel_sdvo_connector) + return false; + + if (device == 0) { + psb_intel_sdvo->controlled_output |= SDVO_OUTPUT_TMDS0; + psb_intel_sdvo_connector->output_flag = SDVO_OUTPUT_TMDS0; + } else if (device == 1) { + psb_intel_sdvo->controlled_output |= SDVO_OUTPUT_TMDS1; + psb_intel_sdvo_connector->output_flag = SDVO_OUTPUT_TMDS1; + } + + intel_connector = &psb_intel_sdvo_connector->base; + connector = &intel_connector->base; + // connector->polled = DRM_CONNECTOR_POLL_CONNECT | DRM_CONNECTOR_POLL_DISCONNECT; + encoder->encoder_type = DRM_MODE_ENCODER_TMDS; + connector->connector_type = DRM_MODE_CONNECTOR_DVID; + + if (psb_intel_sdvo_is_hdmi_connector(psb_intel_sdvo, device)) { + connector->connector_type = DRM_MODE_CONNECTOR_HDMIA; + psb_intel_sdvo->is_hdmi = true; + } + psb_intel_sdvo->base.clone_mask = ((1 << INTEL_SDVO_NON_TV_CLONE_BIT) | + (1 << INTEL_ANALOG_CLONE_BIT)); + + psb_intel_sdvo_connector_init(psb_intel_sdvo_connector, psb_intel_sdvo); + if (psb_intel_sdvo->is_hdmi) + psb_intel_sdvo_add_hdmi_properties(psb_intel_sdvo_connector); + + return true; +} + +static bool +psb_intel_sdvo_tv_init(struct psb_intel_sdvo *psb_intel_sdvo, int type) +{ + struct drm_encoder *encoder = &psb_intel_sdvo->base.base; + struct drm_connector *connector; + struct psb_intel_connector *intel_connector; + struct psb_intel_sdvo_connector *psb_intel_sdvo_connector; + + psb_intel_sdvo_connector = kzalloc(sizeof(struct psb_intel_sdvo_connector), GFP_KERNEL); + if (!psb_intel_sdvo_connector) + return false; + + intel_connector = &psb_intel_sdvo_connector->base; + connector = &intel_connector->base; + encoder->encoder_type = DRM_MODE_ENCODER_TVDAC; + connector->connector_type = DRM_MODE_CONNECTOR_SVIDEO; + + psb_intel_sdvo->controlled_output |= type; + psb_intel_sdvo_connector->output_flag = type; + + psb_intel_sdvo->is_tv = true; + psb_intel_sdvo->base.needs_tv_clock = true; + psb_intel_sdvo->base.clone_mask = 1 << INTEL_SDVO_TV_CLONE_BIT; + + psb_intel_sdvo_connector_init(psb_intel_sdvo_connector, psb_intel_sdvo); + + if (!psb_intel_sdvo_tv_create_property(psb_intel_sdvo, psb_intel_sdvo_connector, type)) + goto err; + + if (!psb_intel_sdvo_create_enhance_property(psb_intel_sdvo, psb_intel_sdvo_connector)) + goto err; + + return true; + +err: + psb_intel_sdvo_destroy(connector); + return false; +} + +static bool +psb_intel_sdvo_analog_init(struct psb_intel_sdvo *psb_intel_sdvo, int device) +{ + struct drm_encoder *encoder = &psb_intel_sdvo->base.base; + struct drm_connector *connector; + struct psb_intel_connector *intel_connector; + struct psb_intel_sdvo_connector *psb_intel_sdvo_connector; + + psb_intel_sdvo_connector = kzalloc(sizeof(struct psb_intel_sdvo_connector), GFP_KERNEL); + if (!psb_intel_sdvo_connector) + return false; + + intel_connector = &psb_intel_sdvo_connector->base; + connector = &intel_connector->base; + connector->polled = DRM_CONNECTOR_POLL_CONNECT; + encoder->encoder_type = DRM_MODE_ENCODER_DAC; + connector->connector_type = DRM_MODE_CONNECTOR_VGA; + + if (device == 0) { + psb_intel_sdvo->controlled_output |= SDVO_OUTPUT_RGB0; + psb_intel_sdvo_connector->output_flag = SDVO_OUTPUT_RGB0; + } else if (device == 1) { + psb_intel_sdvo->controlled_output |= SDVO_OUTPUT_RGB1; + psb_intel_sdvo_connector->output_flag = SDVO_OUTPUT_RGB1; + } + + psb_intel_sdvo->base.clone_mask = ((1 << INTEL_SDVO_NON_TV_CLONE_BIT) | + (1 << INTEL_ANALOG_CLONE_BIT)); + + psb_intel_sdvo_connector_init(psb_intel_sdvo_connector, + psb_intel_sdvo); + return true; +} + +static bool +psb_intel_sdvo_lvds_init(struct psb_intel_sdvo *psb_intel_sdvo, int device) +{ + struct drm_encoder *encoder = &psb_intel_sdvo->base.base; + struct drm_connector *connector; + struct psb_intel_connector *intel_connector; + struct psb_intel_sdvo_connector *psb_intel_sdvo_connector; + + psb_intel_sdvo_connector = kzalloc(sizeof(struct psb_intel_sdvo_connector), GFP_KERNEL); + if (!psb_intel_sdvo_connector) + return false; + + intel_connector = &psb_intel_sdvo_connector->base; + connector = &intel_connector->base; + encoder->encoder_type = DRM_MODE_ENCODER_LVDS; + connector->connector_type = DRM_MODE_CONNECTOR_LVDS; + + if (device == 0) { + psb_intel_sdvo->controlled_output |= SDVO_OUTPUT_LVDS0; + psb_intel_sdvo_connector->output_flag = SDVO_OUTPUT_LVDS0; + } else if (device == 1) { + psb_intel_sdvo->controlled_output |= SDVO_OUTPUT_LVDS1; + psb_intel_sdvo_connector->output_flag = SDVO_OUTPUT_LVDS1; + } + + psb_intel_sdvo->base.clone_mask = ((1 << INTEL_ANALOG_CLONE_BIT) | + (1 << INTEL_SDVO_LVDS_CLONE_BIT)); + + psb_intel_sdvo_connector_init(psb_intel_sdvo_connector, psb_intel_sdvo); + if (!psb_intel_sdvo_create_enhance_property(psb_intel_sdvo, psb_intel_sdvo_connector)) + goto err; + + return true; + +err: + psb_intel_sdvo_destroy(connector); + return false; +} + +static bool +psb_intel_sdvo_output_setup(struct psb_intel_sdvo *psb_intel_sdvo, uint16_t flags) +{ + psb_intel_sdvo->is_tv = false; + psb_intel_sdvo->base.needs_tv_clock = false; + psb_intel_sdvo->is_lvds = false; + + /* SDVO requires XXX1 function may not exist unless it has XXX0 function.*/ + + if (flags & SDVO_OUTPUT_TMDS0) + if (!psb_intel_sdvo_dvi_init(psb_intel_sdvo, 0)) + return false; + + if ((flags & SDVO_TMDS_MASK) == SDVO_TMDS_MASK) + if (!psb_intel_sdvo_dvi_init(psb_intel_sdvo, 1)) + return false; + + /* TV has no XXX1 function block */ + if (flags & SDVO_OUTPUT_SVID0) + if (!psb_intel_sdvo_tv_init(psb_intel_sdvo, SDVO_OUTPUT_SVID0)) + return false; + + if (flags & SDVO_OUTPUT_CVBS0) + if (!psb_intel_sdvo_tv_init(psb_intel_sdvo, SDVO_OUTPUT_CVBS0)) + return false; + + if (flags & SDVO_OUTPUT_RGB0) + if (!psb_intel_sdvo_analog_init(psb_intel_sdvo, 0)) + return false; + + if ((flags & SDVO_RGB_MASK) == SDVO_RGB_MASK) + if (!psb_intel_sdvo_analog_init(psb_intel_sdvo, 1)) + return false; + + if (flags & SDVO_OUTPUT_LVDS0) + if (!psb_intel_sdvo_lvds_init(psb_intel_sdvo, 0)) + return false; + + if ((flags & SDVO_LVDS_MASK) == SDVO_LVDS_MASK) + if (!psb_intel_sdvo_lvds_init(psb_intel_sdvo, 1)) + return false; + + if ((flags & SDVO_OUTPUT_MASK) == 0) { + unsigned char bytes[2]; + + psb_intel_sdvo->controlled_output = 0; + memcpy(bytes, &psb_intel_sdvo->caps.output_flags, 2); + DRM_DEBUG_KMS("%s: Unknown SDVO output type (0x%02x%02x)\n", + SDVO_NAME(psb_intel_sdvo), + bytes[0], bytes[1]); + return false; + } + psb_intel_sdvo->base.crtc_mask = (1 << 0) | (1 << 1); + + return true; +} + +static bool psb_intel_sdvo_tv_create_property(struct psb_intel_sdvo *psb_intel_sdvo, + struct psb_intel_sdvo_connector *psb_intel_sdvo_connector, + int type) +{ + struct drm_device *dev = psb_intel_sdvo->base.base.dev; + struct psb_intel_sdvo_tv_format format; + uint32_t format_map, i; + + if (!psb_intel_sdvo_set_target_output(psb_intel_sdvo, type)) + return false; + + BUILD_BUG_ON(sizeof(format) != 6); + if (!psb_intel_sdvo_get_value(psb_intel_sdvo, + SDVO_CMD_GET_SUPPORTED_TV_FORMATS, + &format, sizeof(format))) + return false; + + memcpy(&format_map, &format, min(sizeof(format_map), sizeof(format))); + + if (format_map == 0) + return false; + + psb_intel_sdvo_connector->format_supported_num = 0; + for (i = 0 ; i < TV_FORMAT_NUM; i++) + if (format_map & (1 << i)) + psb_intel_sdvo_connector->tv_format_supported[psb_intel_sdvo_connector->format_supported_num++] = i; + + + psb_intel_sdvo_connector->tv_format = + drm_property_create(dev, DRM_MODE_PROP_ENUM, + "mode", psb_intel_sdvo_connector->format_supported_num); + if (!psb_intel_sdvo_connector->tv_format) + return false; + + for (i = 0; i < psb_intel_sdvo_connector->format_supported_num; i++) + drm_property_add_enum( + psb_intel_sdvo_connector->tv_format, i, + i, tv_format_names[psb_intel_sdvo_connector->tv_format_supported[i]]); + + psb_intel_sdvo->tv_format_index = psb_intel_sdvo_connector->tv_format_supported[0]; + drm_connector_attach_property(&psb_intel_sdvo_connector->base.base, + psb_intel_sdvo_connector->tv_format, 0); + return true; + +} + +#define ENHANCEMENT(name, NAME) do { \ + if (enhancements.name) { \ + if (!psb_intel_sdvo_get_value(psb_intel_sdvo, SDVO_CMD_GET_MAX_##NAME, &data_value, 4) || \ + !psb_intel_sdvo_get_value(psb_intel_sdvo, SDVO_CMD_GET_##NAME, &response, 2)) \ + return false; \ + psb_intel_sdvo_connector->max_##name = data_value[0]; \ + psb_intel_sdvo_connector->cur_##name = response; \ + psb_intel_sdvo_connector->name = \ + drm_property_create_range(dev, 0, #name, 0, data_value[0]); \ + if (!psb_intel_sdvo_connector->name) return false; \ + drm_connector_attach_property(connector, \ + psb_intel_sdvo_connector->name, \ + psb_intel_sdvo_connector->cur_##name); \ + DRM_DEBUG_KMS(#name ": max %d, default %d, current %d\n", \ + data_value[0], data_value[1], response); \ + } \ +} while(0) + +static bool +psb_intel_sdvo_create_enhance_property_tv(struct psb_intel_sdvo *psb_intel_sdvo, + struct psb_intel_sdvo_connector *psb_intel_sdvo_connector, + struct psb_intel_sdvo_enhancements_reply enhancements) +{ + struct drm_device *dev = psb_intel_sdvo->base.base.dev; + struct drm_connector *connector = &psb_intel_sdvo_connector->base.base; + uint16_t response, data_value[2]; + + /* when horizontal overscan is supported, Add the left/right property */ + if (enhancements.overscan_h) { + if (!psb_intel_sdvo_get_value(psb_intel_sdvo, + SDVO_CMD_GET_MAX_OVERSCAN_H, + &data_value, 4)) + return false; + + if (!psb_intel_sdvo_get_value(psb_intel_sdvo, + SDVO_CMD_GET_OVERSCAN_H, + &response, 2)) + return false; + + psb_intel_sdvo_connector->max_hscan = data_value[0]; + psb_intel_sdvo_connector->left_margin = data_value[0] - response; + psb_intel_sdvo_connector->right_margin = psb_intel_sdvo_connector->left_margin; + psb_intel_sdvo_connector->left = + drm_property_create_range(dev, 0, "left_margin", 0, data_value[0]); + if (!psb_intel_sdvo_connector->left) + return false; + + drm_connector_attach_property(connector, + psb_intel_sdvo_connector->left, + psb_intel_sdvo_connector->left_margin); + + psb_intel_sdvo_connector->right = + drm_property_create_range(dev, 0, "right_margin", 0, data_value[0]); + if (!psb_intel_sdvo_connector->right) + return false; + + drm_connector_attach_property(connector, + psb_intel_sdvo_connector->right, + psb_intel_sdvo_connector->right_margin); + DRM_DEBUG_KMS("h_overscan: max %d, " + "default %d, current %d\n", + data_value[0], data_value[1], response); + } + + if (enhancements.overscan_v) { + if (!psb_intel_sdvo_get_value(psb_intel_sdvo, + SDVO_CMD_GET_MAX_OVERSCAN_V, + &data_value, 4)) + return false; + + if (!psb_intel_sdvo_get_value(psb_intel_sdvo, + SDVO_CMD_GET_OVERSCAN_V, + &response, 2)) + return false; + + psb_intel_sdvo_connector->max_vscan = data_value[0]; + psb_intel_sdvo_connector->top_margin = data_value[0] - response; + psb_intel_sdvo_connector->bottom_margin = psb_intel_sdvo_connector->top_margin; + psb_intel_sdvo_connector->top = + drm_property_create_range(dev, 0, "top_margin", 0, data_value[0]); + if (!psb_intel_sdvo_connector->top) + return false; + + drm_connector_attach_property(connector, + psb_intel_sdvo_connector->top, + psb_intel_sdvo_connector->top_margin); + + psb_intel_sdvo_connector->bottom = + drm_property_create_range(dev, 0, "bottom_margin", 0, data_value[0]); + if (!psb_intel_sdvo_connector->bottom) + return false; + + drm_connector_attach_property(connector, + psb_intel_sdvo_connector->bottom, + psb_intel_sdvo_connector->bottom_margin); + DRM_DEBUG_KMS("v_overscan: max %d, " + "default %d, current %d\n", + data_value[0], data_value[1], response); + } + + ENHANCEMENT(hpos, HPOS); + ENHANCEMENT(vpos, VPOS); + ENHANCEMENT(saturation, SATURATION); + ENHANCEMENT(contrast, CONTRAST); + ENHANCEMENT(hue, HUE); + ENHANCEMENT(sharpness, SHARPNESS); + ENHANCEMENT(brightness, BRIGHTNESS); + ENHANCEMENT(flicker_filter, FLICKER_FILTER); + ENHANCEMENT(flicker_filter_adaptive, FLICKER_FILTER_ADAPTIVE); + ENHANCEMENT(flicker_filter_2d, FLICKER_FILTER_2D); + ENHANCEMENT(tv_chroma_filter, TV_CHROMA_FILTER); + ENHANCEMENT(tv_luma_filter, TV_LUMA_FILTER); + + if (enhancements.dot_crawl) { + if (!psb_intel_sdvo_get_value(psb_intel_sdvo, SDVO_CMD_GET_DOT_CRAWL, &response, 2)) + return false; + + psb_intel_sdvo_connector->max_dot_crawl = 1; + psb_intel_sdvo_connector->cur_dot_crawl = response & 0x1; + psb_intel_sdvo_connector->dot_crawl = + drm_property_create_range(dev, 0, "dot_crawl", 0, 1); + if (!psb_intel_sdvo_connector->dot_crawl) + return false; + + drm_connector_attach_property(connector, + psb_intel_sdvo_connector->dot_crawl, + psb_intel_sdvo_connector->cur_dot_crawl); + DRM_DEBUG_KMS("dot crawl: current %d\n", response); + } + + return true; +} + +static bool +psb_intel_sdvo_create_enhance_property_lvds(struct psb_intel_sdvo *psb_intel_sdvo, + struct psb_intel_sdvo_connector *psb_intel_sdvo_connector, + struct psb_intel_sdvo_enhancements_reply enhancements) +{ + struct drm_device *dev = psb_intel_sdvo->base.base.dev; + struct drm_connector *connector = &psb_intel_sdvo_connector->base.base; + uint16_t response, data_value[2]; + + ENHANCEMENT(brightness, BRIGHTNESS); + + return true; +} +#undef ENHANCEMENT + +static bool psb_intel_sdvo_create_enhance_property(struct psb_intel_sdvo *psb_intel_sdvo, + struct psb_intel_sdvo_connector *psb_intel_sdvo_connector) +{ + union { + struct psb_intel_sdvo_enhancements_reply reply; + uint16_t response; + } enhancements; + + BUILD_BUG_ON(sizeof(enhancements) != 2); + + enhancements.response = 0; + psb_intel_sdvo_get_value(psb_intel_sdvo, + SDVO_CMD_GET_SUPPORTED_ENHANCEMENTS, + &enhancements, sizeof(enhancements)); + if (enhancements.response == 0) { + DRM_DEBUG_KMS("No enhancement is supported\n"); + return true; + } + + if (IS_TV(psb_intel_sdvo_connector)) + return psb_intel_sdvo_create_enhance_property_tv(psb_intel_sdvo, psb_intel_sdvo_connector, enhancements.reply); + else if(IS_LVDS(psb_intel_sdvo_connector)) + return psb_intel_sdvo_create_enhance_property_lvds(psb_intel_sdvo, psb_intel_sdvo_connector, enhancements.reply); + else + return true; +} + +static int psb_intel_sdvo_ddc_proxy_xfer(struct i2c_adapter *adapter, + struct i2c_msg *msgs, + int num) +{ + struct psb_intel_sdvo *sdvo = adapter->algo_data; + + if (!psb_intel_sdvo_set_control_bus_switch(sdvo, sdvo->ddc_bus)) + return -EIO; + + return sdvo->i2c->algo->master_xfer(sdvo->i2c, msgs, num); +} + +static u32 psb_intel_sdvo_ddc_proxy_func(struct i2c_adapter *adapter) +{ + struct psb_intel_sdvo *sdvo = adapter->algo_data; + return sdvo->i2c->algo->functionality(sdvo->i2c); +} + +static const struct i2c_algorithm psb_intel_sdvo_ddc_proxy = { + .master_xfer = psb_intel_sdvo_ddc_proxy_xfer, + .functionality = psb_intel_sdvo_ddc_proxy_func +}; + +static bool +psb_intel_sdvo_init_ddc_proxy(struct psb_intel_sdvo *sdvo, + struct drm_device *dev) +{ + sdvo->ddc.owner = THIS_MODULE; + sdvo->ddc.class = I2C_CLASS_DDC; + snprintf(sdvo->ddc.name, I2C_NAME_SIZE, "SDVO DDC proxy"); + sdvo->ddc.dev.parent = &dev->pdev->dev; + sdvo->ddc.algo_data = sdvo; + sdvo->ddc.algo = &psb_intel_sdvo_ddc_proxy; + + return i2c_add_adapter(&sdvo->ddc) == 0; +} + +bool psb_intel_sdvo_init(struct drm_device *dev, int sdvo_reg) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + struct psb_intel_encoder *psb_intel_encoder; + struct psb_intel_sdvo *psb_intel_sdvo; + int i; + + psb_intel_sdvo = kzalloc(sizeof(struct psb_intel_sdvo), GFP_KERNEL); + if (!psb_intel_sdvo) + return false; + + psb_intel_sdvo->sdvo_reg = sdvo_reg; + psb_intel_sdvo->slave_addr = psb_intel_sdvo_get_slave_addr(dev, sdvo_reg) >> 1; + psb_intel_sdvo_select_i2c_bus(dev_priv, psb_intel_sdvo, sdvo_reg); + if (!psb_intel_sdvo_init_ddc_proxy(psb_intel_sdvo, dev)) { + kfree(psb_intel_sdvo); + return false; + } + + /* encoder type will be decided later */ + psb_intel_encoder = &psb_intel_sdvo->base; + psb_intel_encoder->type = INTEL_OUTPUT_SDVO; + drm_encoder_init(dev, &psb_intel_encoder->base, &psb_intel_sdvo_enc_funcs, 0); + + /* Read the regs to test if we can talk to the device */ + for (i = 0; i < 0x40; i++) { + u8 byte; + + if (!psb_intel_sdvo_read_byte(psb_intel_sdvo, i, &byte)) { + DRM_DEBUG_KMS("No SDVO device found on SDVO%c\n", + IS_SDVOB(sdvo_reg) ? 'B' : 'C'); + goto err; + } + } + + if (IS_SDVOB(sdvo_reg)) + dev_priv->hotplug_supported_mask |= SDVOB_HOTPLUG_INT_STATUS; + else + dev_priv->hotplug_supported_mask |= SDVOC_HOTPLUG_INT_STATUS; + + drm_encoder_helper_add(&psb_intel_encoder->base, &psb_intel_sdvo_helper_funcs); + + /* In default case sdvo lvds is false */ + if (!psb_intel_sdvo_get_capabilities(psb_intel_sdvo, &psb_intel_sdvo->caps)) + goto err; + + if (psb_intel_sdvo_output_setup(psb_intel_sdvo, + psb_intel_sdvo->caps.output_flags) != true) { + DRM_DEBUG_KMS("SDVO output failed to setup on SDVO%c\n", + IS_SDVOB(sdvo_reg) ? 'B' : 'C'); + goto err; + } + + psb_intel_sdvo_select_ddc_bus(dev_priv, psb_intel_sdvo, sdvo_reg); + + /* Set the input timing to the screen. Assume always input 0. */ + if (!psb_intel_sdvo_set_target_input(psb_intel_sdvo)) + goto err; + + if (!psb_intel_sdvo_get_input_pixel_clock_range(psb_intel_sdvo, + &psb_intel_sdvo->pixel_clock_min, + &psb_intel_sdvo->pixel_clock_max)) + goto err; + + DRM_DEBUG_KMS("%s device VID/DID: %02X:%02X.%02X, " + "clock range %dMHz - %dMHz, " + "input 1: %c, input 2: %c, " + "output 1: %c, output 2: %c\n", + SDVO_NAME(psb_intel_sdvo), + psb_intel_sdvo->caps.vendor_id, psb_intel_sdvo->caps.device_id, + psb_intel_sdvo->caps.device_rev_id, + psb_intel_sdvo->pixel_clock_min / 1000, + psb_intel_sdvo->pixel_clock_max / 1000, + (psb_intel_sdvo->caps.sdvo_inputs_mask & 0x1) ? 'Y' : 'N', + (psb_intel_sdvo->caps.sdvo_inputs_mask & 0x2) ? 'Y' : 'N', + /* check currently supported outputs */ + psb_intel_sdvo->caps.output_flags & + (SDVO_OUTPUT_TMDS0 | SDVO_OUTPUT_RGB0) ? 'Y' : 'N', + psb_intel_sdvo->caps.output_flags & + (SDVO_OUTPUT_TMDS1 | SDVO_OUTPUT_RGB1) ? 'Y' : 'N'); + return true; + +err: + drm_encoder_cleanup(&psb_intel_encoder->base); + i2c_del_adapter(&psb_intel_sdvo->ddc); + kfree(psb_intel_sdvo); + + return false; +} diff --git a/drivers/gpu/drm/gma500/psb_intel_sdvo_regs.h b/drivers/gpu/drm/gma500/psb_intel_sdvo_regs.h new file mode 100644 index 0000000..600e797 --- /dev/null +++ b/drivers/gpu/drm/gma500/psb_intel_sdvo_regs.h @@ -0,0 +1,723 @@ +/* + * Copyright ? 2006-2007 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * Eric Anholt + */ + +/** + * @file SDVO command definitions and structures. + */ + +#define SDVO_OUTPUT_FIRST (0) +#define SDVO_OUTPUT_TMDS0 (1 << 0) +#define SDVO_OUTPUT_RGB0 (1 << 1) +#define SDVO_OUTPUT_CVBS0 (1 << 2) +#define SDVO_OUTPUT_SVID0 (1 << 3) +#define SDVO_OUTPUT_YPRPB0 (1 << 4) +#define SDVO_OUTPUT_SCART0 (1 << 5) +#define SDVO_OUTPUT_LVDS0 (1 << 6) +#define SDVO_OUTPUT_TMDS1 (1 << 8) +#define SDVO_OUTPUT_RGB1 (1 << 9) +#define SDVO_OUTPUT_CVBS1 (1 << 10) +#define SDVO_OUTPUT_SVID1 (1 << 11) +#define SDVO_OUTPUT_YPRPB1 (1 << 12) +#define SDVO_OUTPUT_SCART1 (1 << 13) +#define SDVO_OUTPUT_LVDS1 (1 << 14) +#define SDVO_OUTPUT_LAST (14) + +struct psb_intel_sdvo_caps { + u8 vendor_id; + u8 device_id; + u8 device_rev_id; + u8 sdvo_version_major; + u8 sdvo_version_minor; + unsigned int sdvo_inputs_mask:2; + unsigned int smooth_scaling:1; + unsigned int sharp_scaling:1; + unsigned int up_scaling:1; + unsigned int down_scaling:1; + unsigned int stall_support:1; + unsigned int pad:1; + u16 output_flags; +} __attribute__((packed)); + +/** This matches the EDID DTD structure, more or less */ +struct psb_intel_sdvo_dtd { + struct { + u16 clock; /**< pixel clock, in 10kHz units */ + u8 h_active; /**< lower 8 bits (pixels) */ + u8 h_blank; /**< lower 8 bits (pixels) */ + u8 h_high; /**< upper 4 bits each h_active, h_blank */ + u8 v_active; /**< lower 8 bits (lines) */ + u8 v_blank; /**< lower 8 bits (lines) */ + u8 v_high; /**< upper 4 bits each v_active, v_blank */ + } part1; + + struct { + u8 h_sync_off; /**< lower 8 bits, from hblank start */ + u8 h_sync_width; /**< lower 8 bits (pixels) */ + /** lower 4 bits each vsync offset, vsync width */ + u8 v_sync_off_width; + /** + * 2 high bits of hsync offset, 2 high bits of hsync width, + * bits 4-5 of vsync offset, and 2 high bits of vsync width. + */ + u8 sync_off_width_high; + u8 dtd_flags; + u8 sdvo_flags; + /** bits 6-7 of vsync offset at bits 6-7 */ + u8 v_sync_off_high; + u8 reserved; + } part2; +} __attribute__((packed)); + +struct psb_intel_sdvo_pixel_clock_range { + u16 min; /**< pixel clock, in 10kHz units */ + u16 max; /**< pixel clock, in 10kHz units */ +} __attribute__((packed)); + +struct psb_intel_sdvo_preferred_input_timing_args { + u16 clock; + u16 width; + u16 height; + u8 interlace:1; + u8 scaled:1; + u8 pad:6; +} __attribute__((packed)); + +/* I2C registers for SDVO */ +#define SDVO_I2C_ARG_0 0x07 +#define SDVO_I2C_ARG_1 0x06 +#define SDVO_I2C_ARG_2 0x05 +#define SDVO_I2C_ARG_3 0x04 +#define SDVO_I2C_ARG_4 0x03 +#define SDVO_I2C_ARG_5 0x02 +#define SDVO_I2C_ARG_6 0x01 +#define SDVO_I2C_ARG_7 0x00 +#define SDVO_I2C_OPCODE 0x08 +#define SDVO_I2C_CMD_STATUS 0x09 +#define SDVO_I2C_RETURN_0 0x0a +#define SDVO_I2C_RETURN_1 0x0b +#define SDVO_I2C_RETURN_2 0x0c +#define SDVO_I2C_RETURN_3 0x0d +#define SDVO_I2C_RETURN_4 0x0e +#define SDVO_I2C_RETURN_5 0x0f +#define SDVO_I2C_RETURN_6 0x10 +#define SDVO_I2C_RETURN_7 0x11 +#define SDVO_I2C_VENDOR_BEGIN 0x20 + +/* Status results */ +#define SDVO_CMD_STATUS_POWER_ON 0x0 +#define SDVO_CMD_STATUS_SUCCESS 0x1 +#define SDVO_CMD_STATUS_NOTSUPP 0x2 +#define SDVO_CMD_STATUS_INVALID_ARG 0x3 +#define SDVO_CMD_STATUS_PENDING 0x4 +#define SDVO_CMD_STATUS_TARGET_NOT_SPECIFIED 0x5 +#define SDVO_CMD_STATUS_SCALING_NOT_SUPP 0x6 + +/* SDVO commands, argument/result registers */ + +#define SDVO_CMD_RESET 0x01 + +/** Returns a struct intel_sdvo_caps */ +#define SDVO_CMD_GET_DEVICE_CAPS 0x02 + +#define SDVO_CMD_GET_FIRMWARE_REV 0x86 +# define SDVO_DEVICE_FIRMWARE_MINOR SDVO_I2C_RETURN_0 +# define SDVO_DEVICE_FIRMWARE_MAJOR SDVO_I2C_RETURN_1 +# define SDVO_DEVICE_FIRMWARE_PATCH SDVO_I2C_RETURN_2 + +/** + * Reports which inputs are trained (managed to sync). + * + * Devices must have trained within 2 vsyncs of a mode change. + */ +#define SDVO_CMD_GET_TRAINED_INPUTS 0x03 +struct psb_intel_sdvo_get_trained_inputs_response { + unsigned int input0_trained:1; + unsigned int input1_trained:1; + unsigned int pad:6; +} __attribute__((packed)); + +/** Returns a struct intel_sdvo_output_flags of active outputs. */ +#define SDVO_CMD_GET_ACTIVE_OUTPUTS 0x04 + +/** + * Sets the current set of active outputs. + * + * Takes a struct intel_sdvo_output_flags. Must be preceded by a SET_IN_OUT_MAP + * on multi-output devices. + */ +#define SDVO_CMD_SET_ACTIVE_OUTPUTS 0x05 + +/** + * Returns the current mapping of SDVO inputs to outputs on the device. + * + * Returns two struct intel_sdvo_output_flags structures. + */ +#define SDVO_CMD_GET_IN_OUT_MAP 0x06 +struct psb_intel_sdvo_in_out_map { + u16 in0, in1; +}; + +/** + * Sets the current mapping of SDVO inputs to outputs on the device. + * + * Takes two struct i380_sdvo_output_flags structures. + */ +#define SDVO_CMD_SET_IN_OUT_MAP 0x07 + +/** + * Returns a struct intel_sdvo_output_flags of attached displays. + */ +#define SDVO_CMD_GET_ATTACHED_DISPLAYS 0x0b + +/** + * Returns a struct intel_sdvo_ouptut_flags of displays supporting hot plugging. + */ +#define SDVO_CMD_GET_HOT_PLUG_SUPPORT 0x0c + +/** + * Takes a struct intel_sdvo_output_flags. + */ +#define SDVO_CMD_SET_ACTIVE_HOT_PLUG 0x0d + +/** + * Returns a struct intel_sdvo_output_flags of displays with hot plug + * interrupts enabled. + */ +#define SDVO_CMD_GET_ACTIVE_HOT_PLUG 0x0e + +#define SDVO_CMD_GET_INTERRUPT_EVENT_SOURCE 0x0f +struct intel_sdvo_get_interrupt_event_source_response { + u16 interrupt_status; + unsigned int ambient_light_interrupt:1; + unsigned int hdmi_audio_encrypt_change:1; + unsigned int pad:6; +} __attribute__((packed)); + +/** + * Selects which input is affected by future input commands. + * + * Commands affected include SET_INPUT_TIMINGS_PART[12], + * GET_INPUT_TIMINGS_PART[12], GET_PREFERRED_INPUT_TIMINGS_PART[12], + * GET_INPUT_PIXEL_CLOCK_RANGE, and CREATE_PREFERRED_INPUT_TIMINGS. + */ +#define SDVO_CMD_SET_TARGET_INPUT 0x10 +struct psb_intel_sdvo_set_target_input_args { + unsigned int target_1:1; + unsigned int pad:7; +} __attribute__((packed)); + +/** + * Takes a struct intel_sdvo_output_flags of which outputs are targeted by + * future output commands. + * + * Affected commands inclue SET_OUTPUT_TIMINGS_PART[12], + * GET_OUTPUT_TIMINGS_PART[12], and GET_OUTPUT_PIXEL_CLOCK_RANGE. + */ +#define SDVO_CMD_SET_TARGET_OUTPUT 0x11 + +#define SDVO_CMD_GET_INPUT_TIMINGS_PART1 0x12 +#define SDVO_CMD_GET_INPUT_TIMINGS_PART2 0x13 +#define SDVO_CMD_SET_INPUT_TIMINGS_PART1 0x14 +#define SDVO_CMD_SET_INPUT_TIMINGS_PART2 0x15 +#define SDVO_CMD_SET_OUTPUT_TIMINGS_PART1 0x16 +#define SDVO_CMD_SET_OUTPUT_TIMINGS_PART2 0x17 +#define SDVO_CMD_GET_OUTPUT_TIMINGS_PART1 0x18 +#define SDVO_CMD_GET_OUTPUT_TIMINGS_PART2 0x19 +/* Part 1 */ +# define SDVO_DTD_CLOCK_LOW SDVO_I2C_ARG_0 +# define SDVO_DTD_CLOCK_HIGH SDVO_I2C_ARG_1 +# define SDVO_DTD_H_ACTIVE SDVO_I2C_ARG_2 +# define SDVO_DTD_H_BLANK SDVO_I2C_ARG_3 +# define SDVO_DTD_H_HIGH SDVO_I2C_ARG_4 +# define SDVO_DTD_V_ACTIVE SDVO_I2C_ARG_5 +# define SDVO_DTD_V_BLANK SDVO_I2C_ARG_6 +# define SDVO_DTD_V_HIGH SDVO_I2C_ARG_7 +/* Part 2 */ +# define SDVO_DTD_HSYNC_OFF SDVO_I2C_ARG_0 +# define SDVO_DTD_HSYNC_WIDTH SDVO_I2C_ARG_1 +# define SDVO_DTD_VSYNC_OFF_WIDTH SDVO_I2C_ARG_2 +# define SDVO_DTD_SYNC_OFF_WIDTH_HIGH SDVO_I2C_ARG_3 +# define SDVO_DTD_DTD_FLAGS SDVO_I2C_ARG_4 +# define SDVO_DTD_DTD_FLAG_INTERLACED (1 << 7) +# define SDVO_DTD_DTD_FLAG_STEREO_MASK (3 << 5) +# define SDVO_DTD_DTD_FLAG_INPUT_MASK (3 << 3) +# define SDVO_DTD_DTD_FLAG_SYNC_MASK (3 << 1) +# define SDVO_DTD_SDVO_FLAS SDVO_I2C_ARG_5 +# define SDVO_DTD_SDVO_FLAG_STALL (1 << 7) +# define SDVO_DTD_SDVO_FLAG_CENTERED (0 << 6) +# define SDVO_DTD_SDVO_FLAG_UPPER_LEFT (1 << 6) +# define SDVO_DTD_SDVO_FLAG_SCALING_MASK (3 << 4) +# define SDVO_DTD_SDVO_FLAG_SCALING_NONE (0 << 4) +# define SDVO_DTD_SDVO_FLAG_SCALING_SHARP (1 << 4) +# define SDVO_DTD_SDVO_FLAG_SCALING_SMOOTH (2 << 4) +# define SDVO_DTD_VSYNC_OFF_HIGH SDVO_I2C_ARG_6 + +/** + * Generates a DTD based on the given width, height, and flags. + * + * This will be supported by any device supporting scaling or interlaced + * modes. + */ +#define SDVO_CMD_CREATE_PREFERRED_INPUT_TIMING 0x1a +# define SDVO_PREFERRED_INPUT_TIMING_CLOCK_LOW SDVO_I2C_ARG_0 +# define SDVO_PREFERRED_INPUT_TIMING_CLOCK_HIGH SDVO_I2C_ARG_1 +# define SDVO_PREFERRED_INPUT_TIMING_WIDTH_LOW SDVO_I2C_ARG_2 +# define SDVO_PREFERRED_INPUT_TIMING_WIDTH_HIGH SDVO_I2C_ARG_3 +# define SDVO_PREFERRED_INPUT_TIMING_HEIGHT_LOW SDVO_I2C_ARG_4 +# define SDVO_PREFERRED_INPUT_TIMING_HEIGHT_HIGH SDVO_I2C_ARG_5 +# define SDVO_PREFERRED_INPUT_TIMING_FLAGS SDVO_I2C_ARG_6 +# define SDVO_PREFERRED_INPUT_TIMING_FLAGS_INTERLACED (1 << 0) +# define SDVO_PREFERRED_INPUT_TIMING_FLAGS_SCALED (1 << 1) + +#define SDVO_CMD_GET_PREFERRED_INPUT_TIMING_PART1 0x1b +#define SDVO_CMD_GET_PREFERRED_INPUT_TIMING_PART2 0x1c + +/** Returns a struct intel_sdvo_pixel_clock_range */ +#define SDVO_CMD_GET_INPUT_PIXEL_CLOCK_RANGE 0x1d +/** Returns a struct intel_sdvo_pixel_clock_range */ +#define SDVO_CMD_GET_OUTPUT_PIXEL_CLOCK_RANGE 0x1e + +/** Returns a byte bitfield containing SDVO_CLOCK_RATE_MULT_* flags */ +#define SDVO_CMD_GET_SUPPORTED_CLOCK_RATE_MULTS 0x1f + +/** Returns a byte containing a SDVO_CLOCK_RATE_MULT_* flag */ +#define SDVO_CMD_GET_CLOCK_RATE_MULT 0x20 +/** Takes a byte containing a SDVO_CLOCK_RATE_MULT_* flag */ +#define SDVO_CMD_SET_CLOCK_RATE_MULT 0x21 +# define SDVO_CLOCK_RATE_MULT_1X (1 << 0) +# define SDVO_CLOCK_RATE_MULT_2X (1 << 1) +# define SDVO_CLOCK_RATE_MULT_4X (1 << 3) + +#define SDVO_CMD_GET_SUPPORTED_TV_FORMATS 0x27 +/** 6 bytes of bit flags for TV formats shared by all TV format functions */ +struct psb_intel_sdvo_tv_format { + unsigned int ntsc_m:1; + unsigned int ntsc_j:1; + unsigned int ntsc_443:1; + unsigned int pal_b:1; + unsigned int pal_d:1; + unsigned int pal_g:1; + unsigned int pal_h:1; + unsigned int pal_i:1; + + unsigned int pal_m:1; + unsigned int pal_n:1; + unsigned int pal_nc:1; + unsigned int pal_60:1; + unsigned int secam_b:1; + unsigned int secam_d:1; + unsigned int secam_g:1; + unsigned int secam_k:1; + + unsigned int secam_k1:1; + unsigned int secam_l:1; + unsigned int secam_60:1; + unsigned int hdtv_std_smpte_240m_1080i_59:1; + unsigned int hdtv_std_smpte_240m_1080i_60:1; + unsigned int hdtv_std_smpte_260m_1080i_59:1; + unsigned int hdtv_std_smpte_260m_1080i_60:1; + unsigned int hdtv_std_smpte_274m_1080i_50:1; + + unsigned int hdtv_std_smpte_274m_1080i_59:1; + unsigned int hdtv_std_smpte_274m_1080i_60:1; + unsigned int hdtv_std_smpte_274m_1080p_23:1; + unsigned int hdtv_std_smpte_274m_1080p_24:1; + unsigned int hdtv_std_smpte_274m_1080p_25:1; + unsigned int hdtv_std_smpte_274m_1080p_29:1; + unsigned int hdtv_std_smpte_274m_1080p_30:1; + unsigned int hdtv_std_smpte_274m_1080p_50:1; + + unsigned int hdtv_std_smpte_274m_1080p_59:1; + unsigned int hdtv_std_smpte_274m_1080p_60:1; + unsigned int hdtv_std_smpte_295m_1080i_50:1; + unsigned int hdtv_std_smpte_295m_1080p_50:1; + unsigned int hdtv_std_smpte_296m_720p_59:1; + unsigned int hdtv_std_smpte_296m_720p_60:1; + unsigned int hdtv_std_smpte_296m_720p_50:1; + unsigned int hdtv_std_smpte_293m_480p_59:1; + + unsigned int hdtv_std_smpte_170m_480i_59:1; + unsigned int hdtv_std_iturbt601_576i_50:1; + unsigned int hdtv_std_iturbt601_576p_50:1; + unsigned int hdtv_std_eia_7702a_480i_60:1; + unsigned int hdtv_std_eia_7702a_480p_60:1; + unsigned int pad:3; +} __attribute__((packed)); + +#define SDVO_CMD_GET_TV_FORMAT 0x28 + +#define SDVO_CMD_SET_TV_FORMAT 0x29 + +/** Returns the resolutiosn that can be used with the given TV format */ +#define SDVO_CMD_GET_SDTV_RESOLUTION_SUPPORT 0x83 +struct psb_intel_sdvo_sdtv_resolution_request { + unsigned int ntsc_m:1; + unsigned int ntsc_j:1; + unsigned int ntsc_443:1; + unsigned int pal_b:1; + unsigned int pal_d:1; + unsigned int pal_g:1; + unsigned int pal_h:1; + unsigned int pal_i:1; + + unsigned int pal_m:1; + unsigned int pal_n:1; + unsigned int pal_nc:1; + unsigned int pal_60:1; + unsigned int secam_b:1; + unsigned int secam_d:1; + unsigned int secam_g:1; + unsigned int secam_k:1; + + unsigned int secam_k1:1; + unsigned int secam_l:1; + unsigned int secam_60:1; + unsigned int pad:5; +} __attribute__((packed)); + +struct psb_intel_sdvo_sdtv_resolution_reply { + unsigned int res_320x200:1; + unsigned int res_320x240:1; + unsigned int res_400x300:1; + unsigned int res_640x350:1; + unsigned int res_640x400:1; + unsigned int res_640x480:1; + unsigned int res_704x480:1; + unsigned int res_704x576:1; + + unsigned int res_720x350:1; + unsigned int res_720x400:1; + unsigned int res_720x480:1; + unsigned int res_720x540:1; + unsigned int res_720x576:1; + unsigned int res_768x576:1; + unsigned int res_800x600:1; + unsigned int res_832x624:1; + + unsigned int res_920x766:1; + unsigned int res_1024x768:1; + unsigned int res_1280x1024:1; + unsigned int pad:5; +} __attribute__((packed)); + +/* Get supported resolution with squire pixel aspect ratio that can be + scaled for the requested HDTV format */ +#define SDVO_CMD_GET_SCALED_HDTV_RESOLUTION_SUPPORT 0x85 + +struct psb_intel_sdvo_hdtv_resolution_request { + unsigned int hdtv_std_smpte_240m_1080i_59:1; + unsigned int hdtv_std_smpte_240m_1080i_60:1; + unsigned int hdtv_std_smpte_260m_1080i_59:1; + unsigned int hdtv_std_smpte_260m_1080i_60:1; + unsigned int hdtv_std_smpte_274m_1080i_50:1; + unsigned int hdtv_std_smpte_274m_1080i_59:1; + unsigned int hdtv_std_smpte_274m_1080i_60:1; + unsigned int hdtv_std_smpte_274m_1080p_23:1; + + unsigned int hdtv_std_smpte_274m_1080p_24:1; + unsigned int hdtv_std_smpte_274m_1080p_25:1; + unsigned int hdtv_std_smpte_274m_1080p_29:1; + unsigned int hdtv_std_smpte_274m_1080p_30:1; + unsigned int hdtv_std_smpte_274m_1080p_50:1; + unsigned int hdtv_std_smpte_274m_1080p_59:1; + unsigned int hdtv_std_smpte_274m_1080p_60:1; + unsigned int hdtv_std_smpte_295m_1080i_50:1; + + unsigned int hdtv_std_smpte_295m_1080p_50:1; + unsigned int hdtv_std_smpte_296m_720p_59:1; + unsigned int hdtv_std_smpte_296m_720p_60:1; + unsigned int hdtv_std_smpte_296m_720p_50:1; + unsigned int hdtv_std_smpte_293m_480p_59:1; + unsigned int hdtv_std_smpte_170m_480i_59:1; + unsigned int hdtv_std_iturbt601_576i_50:1; + unsigned int hdtv_std_iturbt601_576p_50:1; + + unsigned int hdtv_std_eia_7702a_480i_60:1; + unsigned int hdtv_std_eia_7702a_480p_60:1; + unsigned int pad:6; +} __attribute__((packed)); + +struct psb_intel_sdvo_hdtv_resolution_reply { + unsigned int res_640x480:1; + unsigned int res_800x600:1; + unsigned int res_1024x768:1; + unsigned int res_1280x960:1; + unsigned int res_1400x1050:1; + unsigned int res_1600x1200:1; + unsigned int res_1920x1440:1; + unsigned int res_2048x1536:1; + + unsigned int res_2560x1920:1; + unsigned int res_3200x2400:1; + unsigned int res_3840x2880:1; + unsigned int pad1:5; + + unsigned int res_848x480:1; + unsigned int res_1064x600:1; + unsigned int res_1280x720:1; + unsigned int res_1360x768:1; + unsigned int res_1704x960:1; + unsigned int res_1864x1050:1; + unsigned int res_1920x1080:1; + unsigned int res_2128x1200:1; + + unsigned int res_2560x1400:1; + unsigned int res_2728x1536:1; + unsigned int res_3408x1920:1; + unsigned int res_4264x2400:1; + unsigned int res_5120x2880:1; + unsigned int pad2:3; + + unsigned int res_768x480:1; + unsigned int res_960x600:1; + unsigned int res_1152x720:1; + unsigned int res_1124x768:1; + unsigned int res_1536x960:1; + unsigned int res_1680x1050:1; + unsigned int res_1728x1080:1; + unsigned int res_1920x1200:1; + + unsigned int res_2304x1440:1; + unsigned int res_2456x1536:1; + unsigned int res_3072x1920:1; + unsigned int res_3840x2400:1; + unsigned int res_4608x2880:1; + unsigned int pad3:3; + + unsigned int res_1280x1024:1; + unsigned int pad4:7; + + unsigned int res_1280x768:1; + unsigned int pad5:7; +} __attribute__((packed)); + +/* Get supported power state returns info for encoder and monitor, rely on + last SetTargetInput and SetTargetOutput calls */ +#define SDVO_CMD_GET_SUPPORTED_POWER_STATES 0x2a +/* Get power state returns info for encoder and monitor, rely on last + SetTargetInput and SetTargetOutput calls */ +#define SDVO_CMD_GET_POWER_STATE 0x2b +#define SDVO_CMD_GET_ENCODER_POWER_STATE 0x2b +#define SDVO_CMD_SET_ENCODER_POWER_STATE 0x2c +# define SDVO_ENCODER_STATE_ON (1 << 0) +# define SDVO_ENCODER_STATE_STANDBY (1 << 1) +# define SDVO_ENCODER_STATE_SUSPEND (1 << 2) +# define SDVO_ENCODER_STATE_OFF (1 << 3) +# define SDVO_MONITOR_STATE_ON (1 << 4) +# define SDVO_MONITOR_STATE_STANDBY (1 << 5) +# define SDVO_MONITOR_STATE_SUSPEND (1 << 6) +# define SDVO_MONITOR_STATE_OFF (1 << 7) + +#define SDVO_CMD_GET_MAX_PANEL_POWER_SEQUENCING 0x2d +#define SDVO_CMD_GET_PANEL_POWER_SEQUENCING 0x2e +#define SDVO_CMD_SET_PANEL_POWER_SEQUENCING 0x2f +/** + * The panel power sequencing parameters are in units of milliseconds. + * The high fields are bits 8:9 of the 10-bit values. + */ +struct psb_sdvo_panel_power_sequencing { + u8 t0; + u8 t1; + u8 t2; + u8 t3; + u8 t4; + + unsigned int t0_high:2; + unsigned int t1_high:2; + unsigned int t2_high:2; + unsigned int t3_high:2; + + unsigned int t4_high:2; + unsigned int pad:6; +} __attribute__((packed)); + +#define SDVO_CMD_GET_MAX_BACKLIGHT_LEVEL 0x30 +struct sdvo_max_backlight_reply { + u8 max_value; + u8 default_value; +} __attribute__((packed)); + +#define SDVO_CMD_GET_BACKLIGHT_LEVEL 0x31 +#define SDVO_CMD_SET_BACKLIGHT_LEVEL 0x32 + +#define SDVO_CMD_GET_AMBIENT_LIGHT 0x33 +struct sdvo_get_ambient_light_reply { + u16 trip_low; + u16 trip_high; + u16 value; +} __attribute__((packed)); +#define SDVO_CMD_SET_AMBIENT_LIGHT 0x34 +struct sdvo_set_ambient_light_reply { + u16 trip_low; + u16 trip_high; + unsigned int enable:1; + unsigned int pad:7; +} __attribute__((packed)); + +/* Set display power state */ +#define SDVO_CMD_SET_DISPLAY_POWER_STATE 0x7d +# define SDVO_DISPLAY_STATE_ON (1 << 0) +# define SDVO_DISPLAY_STATE_STANDBY (1 << 1) +# define SDVO_DISPLAY_STATE_SUSPEND (1 << 2) +# define SDVO_DISPLAY_STATE_OFF (1 << 3) + +#define SDVO_CMD_GET_SUPPORTED_ENHANCEMENTS 0x84 +struct psb_intel_sdvo_enhancements_reply { + unsigned int flicker_filter:1; + unsigned int flicker_filter_adaptive:1; + unsigned int flicker_filter_2d:1; + unsigned int saturation:1; + unsigned int hue:1; + unsigned int brightness:1; + unsigned int contrast:1; + unsigned int overscan_h:1; + + unsigned int overscan_v:1; + unsigned int hpos:1; + unsigned int vpos:1; + unsigned int sharpness:1; + unsigned int dot_crawl:1; + unsigned int dither:1; + unsigned int tv_chroma_filter:1; + unsigned int tv_luma_filter:1; +} __attribute__((packed)); + +/* Picture enhancement limits below are dependent on the current TV format, + * and thus need to be queried and set after it. + */ +#define SDVO_CMD_GET_MAX_FLICKER_FILTER 0x4d +#define SDVO_CMD_GET_MAX_FLICKER_FILTER_ADAPTIVE 0x7b +#define SDVO_CMD_GET_MAX_FLICKER_FILTER_2D 0x52 +#define SDVO_CMD_GET_MAX_SATURATION 0x55 +#define SDVO_CMD_GET_MAX_HUE 0x58 +#define SDVO_CMD_GET_MAX_BRIGHTNESS 0x5b +#define SDVO_CMD_GET_MAX_CONTRAST 0x5e +#define SDVO_CMD_GET_MAX_OVERSCAN_H 0x61 +#define SDVO_CMD_GET_MAX_OVERSCAN_V 0x64 +#define SDVO_CMD_GET_MAX_HPOS 0x67 +#define SDVO_CMD_GET_MAX_VPOS 0x6a +#define SDVO_CMD_GET_MAX_SHARPNESS 0x6d +#define SDVO_CMD_GET_MAX_TV_CHROMA_FILTER 0x74 +#define SDVO_CMD_GET_MAX_TV_LUMA_FILTER 0x77 +struct psb_intel_sdvo_enhancement_limits_reply { + u16 max_value; + u16 default_value; +} __attribute__((packed)); + +#define SDVO_CMD_GET_LVDS_PANEL_INFORMATION 0x7f +#define SDVO_CMD_SET_LVDS_PANEL_INFORMATION 0x80 +# define SDVO_LVDS_COLOR_DEPTH_18 (0 << 0) +# define SDVO_LVDS_COLOR_DEPTH_24 (1 << 0) +# define SDVO_LVDS_CONNECTOR_SPWG (0 << 2) +# define SDVO_LVDS_CONNECTOR_OPENLDI (1 << 2) +# define SDVO_LVDS_SINGLE_CHANNEL (0 << 4) +# define SDVO_LVDS_DUAL_CHANNEL (1 << 4) + +#define SDVO_CMD_GET_FLICKER_FILTER 0x4e +#define SDVO_CMD_SET_FLICKER_FILTER 0x4f +#define SDVO_CMD_GET_FLICKER_FILTER_ADAPTIVE 0x50 +#define SDVO_CMD_SET_FLICKER_FILTER_ADAPTIVE 0x51 +#define SDVO_CMD_GET_FLICKER_FILTER_2D 0x53 +#define SDVO_CMD_SET_FLICKER_FILTER_2D 0x54 +#define SDVO_CMD_GET_SATURATION 0x56 +#define SDVO_CMD_SET_SATURATION 0x57 +#define SDVO_CMD_GET_HUE 0x59 +#define SDVO_CMD_SET_HUE 0x5a +#define SDVO_CMD_GET_BRIGHTNESS 0x5c +#define SDVO_CMD_SET_BRIGHTNESS 0x5d +#define SDVO_CMD_GET_CONTRAST 0x5f +#define SDVO_CMD_SET_CONTRAST 0x60 +#define SDVO_CMD_GET_OVERSCAN_H 0x62 +#define SDVO_CMD_SET_OVERSCAN_H 0x63 +#define SDVO_CMD_GET_OVERSCAN_V 0x65 +#define SDVO_CMD_SET_OVERSCAN_V 0x66 +#define SDVO_CMD_GET_HPOS 0x68 +#define SDVO_CMD_SET_HPOS 0x69 +#define SDVO_CMD_GET_VPOS 0x6b +#define SDVO_CMD_SET_VPOS 0x6c +#define SDVO_CMD_GET_SHARPNESS 0x6e +#define SDVO_CMD_SET_SHARPNESS 0x6f +#define SDVO_CMD_GET_TV_CHROMA_FILTER 0x75 +#define SDVO_CMD_SET_TV_CHROMA_FILTER 0x76 +#define SDVO_CMD_GET_TV_LUMA_FILTER 0x78 +#define SDVO_CMD_SET_TV_LUMA_FILTER 0x79 +struct psb_intel_sdvo_enhancements_arg { + u16 value; +}__attribute__((packed)); + +#define SDVO_CMD_GET_DOT_CRAWL 0x70 +#define SDVO_CMD_SET_DOT_CRAWL 0x71 +# define SDVO_DOT_CRAWL_ON (1 << 0) +# define SDVO_DOT_CRAWL_DEFAULT_ON (1 << 1) + +#define SDVO_CMD_GET_DITHER 0x72 +#define SDVO_CMD_SET_DITHER 0x73 +# define SDVO_DITHER_ON (1 << 0) +# define SDVO_DITHER_DEFAULT_ON (1 << 1) + +#define SDVO_CMD_SET_CONTROL_BUS_SWITCH 0x7a +# define SDVO_CONTROL_BUS_PROM (1 << 0) +# define SDVO_CONTROL_BUS_DDC1 (1 << 1) +# define SDVO_CONTROL_BUS_DDC2 (1 << 2) +# define SDVO_CONTROL_BUS_DDC3 (1 << 3) + +/* HDMI op codes */ +#define SDVO_CMD_GET_SUPP_ENCODE 0x9d +#define SDVO_CMD_GET_ENCODE 0x9e +#define SDVO_CMD_SET_ENCODE 0x9f + #define SDVO_ENCODE_DVI 0x0 + #define SDVO_ENCODE_HDMI 0x1 +#define SDVO_CMD_SET_PIXEL_REPLI 0x8b +#define SDVO_CMD_GET_PIXEL_REPLI 0x8c +#define SDVO_CMD_GET_COLORIMETRY_CAP 0x8d +#define SDVO_CMD_SET_COLORIMETRY 0x8e + #define SDVO_COLORIMETRY_RGB256 0x0 + #define SDVO_COLORIMETRY_RGB220 0x1 + #define SDVO_COLORIMETRY_YCrCb422 0x3 + #define SDVO_COLORIMETRY_YCrCb444 0x4 +#define SDVO_CMD_GET_COLORIMETRY 0x8f +#define SDVO_CMD_GET_AUDIO_ENCRYPT_PREFER 0x90 +#define SDVO_CMD_SET_AUDIO_STAT 0x91 +#define SDVO_CMD_GET_AUDIO_STAT 0x92 +#define SDVO_CMD_SET_HBUF_INDEX 0x93 +#define SDVO_CMD_GET_HBUF_INDEX 0x94 +#define SDVO_CMD_GET_HBUF_INFO 0x95 +#define SDVO_CMD_SET_HBUF_AV_SPLIT 0x96 +#define SDVO_CMD_GET_HBUF_AV_SPLIT 0x97 +#define SDVO_CMD_SET_HBUF_DATA 0x98 +#define SDVO_CMD_GET_HBUF_DATA 0x99 +#define SDVO_CMD_SET_HBUF_TXRATE 0x9a +#define SDVO_CMD_GET_HBUF_TXRATE 0x9b + #define SDVO_HBUF_TX_DISABLED (0 << 6) + #define SDVO_HBUF_TX_ONCE (2 << 6) + #define SDVO_HBUF_TX_VSYNC (3 << 6) +#define SDVO_CMD_GET_AUDIO_TX_INFO 0x9c +#define SDVO_NEED_TO_STALL (1 << 7) + +struct psb_intel_sdvo_encode { + u8 dvi_rev; + u8 hdmi_rev; +} __attribute__ ((packed)); diff --git a/drivers/gpu/drm/gma500/psb_irq.c b/drivers/gpu/drm/gma500/psb_irq.c new file mode 100644 index 0000000..1869586 --- /dev/null +++ b/drivers/gpu/drm/gma500/psb_irq.c @@ -0,0 +1,622 @@ +/************************************************************************** + * Copyright (c) 2007, Intel Corporation. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + * Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to + * develop this driver. + * + **************************************************************************/ +/* + */ + +#include +#include "psb_drv.h" +#include "psb_reg.h" +#include "psb_intel_reg.h" +#include "power.h" +#include "psb_irq.h" +#include "mdfld_output.h" + +/* + * inline functions + */ + +static inline u32 +psb_pipestat(int pipe) +{ + if (pipe == 0) + return PIPEASTAT; + if (pipe == 1) + return PIPEBSTAT; + if (pipe == 2) + return PIPECSTAT; + BUG(); +} + +static inline u32 +mid_pipe_event(int pipe) +{ + if (pipe == 0) + return _PSB_PIPEA_EVENT_FLAG; + if (pipe == 1) + return _MDFLD_PIPEB_EVENT_FLAG; + if (pipe == 2) + return _MDFLD_PIPEC_EVENT_FLAG; + BUG(); +} + +static inline u32 +mid_pipe_vsync(int pipe) +{ + if (pipe == 0) + return _PSB_VSYNC_PIPEA_FLAG; + if (pipe == 1) + return _PSB_VSYNC_PIPEB_FLAG; + if (pipe == 2) + return _MDFLD_PIPEC_VBLANK_FLAG; + BUG(); +} + +static inline u32 +mid_pipeconf(int pipe) +{ + if (pipe == 0) + return PIPEACONF; + if (pipe == 1) + return PIPEBCONF; + if (pipe == 2) + return PIPECCONF; + BUG(); +} + +void +psb_enable_pipestat(struct drm_psb_private *dev_priv, int pipe, u32 mask) +{ + if ((dev_priv->pipestat[pipe] & mask) != mask) { + u32 reg = psb_pipestat(pipe); + dev_priv->pipestat[pipe] |= mask; + /* Enable the interrupt, clear any pending status */ + if (gma_power_begin(dev_priv->dev, false)) { + u32 writeVal = PSB_RVDC32(reg); + writeVal |= (mask | (mask >> 16)); + PSB_WVDC32(writeVal, reg); + (void) PSB_RVDC32(reg); + gma_power_end(dev_priv->dev); + } + } +} + +void +psb_disable_pipestat(struct drm_psb_private *dev_priv, int pipe, u32 mask) +{ + if ((dev_priv->pipestat[pipe] & mask) != 0) { + u32 reg = psb_pipestat(pipe); + dev_priv->pipestat[pipe] &= ~mask; + if (gma_power_begin(dev_priv->dev, false)) { + u32 writeVal = PSB_RVDC32(reg); + writeVal &= ~mask; + PSB_WVDC32(writeVal, reg); + (void) PSB_RVDC32(reg); + gma_power_end(dev_priv->dev); + } + } +} + +static void mid_enable_pipe_event(struct drm_psb_private *dev_priv, int pipe) +{ + if (gma_power_begin(dev_priv->dev, false)) { + u32 pipe_event = mid_pipe_event(pipe); + dev_priv->vdc_irq_mask |= pipe_event; + PSB_WVDC32(~dev_priv->vdc_irq_mask, PSB_INT_MASK_R); + PSB_WVDC32(dev_priv->vdc_irq_mask, PSB_INT_ENABLE_R); + gma_power_end(dev_priv->dev); + } +} + +static void mid_disable_pipe_event(struct drm_psb_private *dev_priv, int pipe) +{ + if (dev_priv->pipestat[pipe] == 0) { + if (gma_power_begin(dev_priv->dev, false)) { + u32 pipe_event = mid_pipe_event(pipe); + dev_priv->vdc_irq_mask &= ~pipe_event; + PSB_WVDC32(~dev_priv->vdc_irq_mask, PSB_INT_MASK_R); + PSB_WVDC32(dev_priv->vdc_irq_mask, PSB_INT_ENABLE_R); + gma_power_end(dev_priv->dev); + } + } +} + +/** + * Display controller interrupt handler for pipe event. + * + */ +static void mid_pipe_event_handler(struct drm_device *dev, int pipe) +{ + struct drm_psb_private *dev_priv = + (struct drm_psb_private *) dev->dev_private; + + uint32_t pipe_stat_val = 0; + uint32_t pipe_stat_reg = psb_pipestat(pipe); + uint32_t pipe_enable = dev_priv->pipestat[pipe]; + uint32_t pipe_status = dev_priv->pipestat[pipe] >> 16; + uint32_t pipe_clear; + uint32_t i = 0; + + spin_lock(&dev_priv->irqmask_lock); + + pipe_stat_val = PSB_RVDC32(pipe_stat_reg); + pipe_stat_val &= pipe_enable | pipe_status; + pipe_stat_val &= pipe_stat_val >> 16; + + spin_unlock(&dev_priv->irqmask_lock); + + /* Clear the 2nd level interrupt status bits + * Sometimes the bits are very sticky so we repeat until they unstick */ + for (i = 0; i < 0xffff; i++) { + PSB_WVDC32(PSB_RVDC32(pipe_stat_reg), pipe_stat_reg); + pipe_clear = PSB_RVDC32(pipe_stat_reg) & pipe_status; + + if (pipe_clear == 0) + break; + } + + if (pipe_clear) + dev_err(dev->dev, + "%s, can't clear status bits for pipe %d, its value = 0x%x.\n", + __func__, pipe, PSB_RVDC32(pipe_stat_reg)); + + if (pipe_stat_val & PIPE_VBLANK_STATUS) + drm_handle_vblank(dev, pipe); + + if (pipe_stat_val & PIPE_TE_STATUS) + drm_handle_vblank(dev, pipe); +} + +/* + * Display controller interrupt handler. + */ +static void psb_vdc_interrupt(struct drm_device *dev, uint32_t vdc_stat) +{ + if (vdc_stat & _PSB_VSYNC_PIPEA_FLAG) + mid_pipe_event_handler(dev, 0); + + if (vdc_stat & _PSB_VSYNC_PIPEB_FLAG) + mid_pipe_event_handler(dev, 1); +} + +irqreturn_t psb_irq_handler(DRM_IRQ_ARGS) +{ + struct drm_device *dev = (struct drm_device *) arg; + struct drm_psb_private *dev_priv = + (struct drm_psb_private *) dev->dev_private; + + uint32_t vdc_stat, dsp_int = 0, sgx_int = 0; + int handled = 0; + + spin_lock(&dev_priv->irqmask_lock); + + vdc_stat = PSB_RVDC32(PSB_INT_IDENTITY_R); + + if (vdc_stat & _PSB_PIPE_EVENT_FLAG) + dsp_int = 1; + + /* FIXME: Handle Medfield + if (vdc_stat & _MDFLD_DISP_ALL_IRQ_FLAG) + dsp_int = 1; + */ + + if (vdc_stat & _PSB_IRQ_SGX_FLAG) + sgx_int = 1; + + vdc_stat &= dev_priv->vdc_irq_mask; + spin_unlock(&dev_priv->irqmask_lock); + + if (dsp_int && gma_power_is_on(dev)) { + psb_vdc_interrupt(dev, vdc_stat); + handled = 1; + } + + if (sgx_int) { + /* Not expected - we have it masked, shut it up */ + u32 s, s2; + s = PSB_RSGX32(PSB_CR_EVENT_STATUS); + s2 = PSB_RSGX32(PSB_CR_EVENT_STATUS2); + PSB_WSGX32(s, PSB_CR_EVENT_HOST_CLEAR); + PSB_WSGX32(s2, PSB_CR_EVENT_HOST_CLEAR2); + /* if s & _PSB_CE_TWOD_COMPLETE we have 2D done but + we may as well poll even if we add that ! */ + handled = 1; + } + + PSB_WVDC32(vdc_stat, PSB_INT_IDENTITY_R); + (void) PSB_RVDC32(PSB_INT_IDENTITY_R); + DRM_READMEMORYBARRIER(); + + if (!handled) + return IRQ_NONE; + + return IRQ_HANDLED; +} + +void psb_irq_preinstall(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = + (struct drm_psb_private *) dev->dev_private; + unsigned long irqflags; + + spin_lock_irqsave(&dev_priv->irqmask_lock, irqflags); + + if (gma_power_is_on(dev)) + PSB_WVDC32(0xFFFFFFFF, PSB_HWSTAM); + if (dev->vblank_enabled[0]) + dev_priv->vdc_irq_mask |= _PSB_VSYNC_PIPEA_FLAG; + if (dev->vblank_enabled[1]) + dev_priv->vdc_irq_mask |= _PSB_VSYNC_PIPEB_FLAG; + + /* FIXME: Handle Medfield irq mask + if (dev->vblank_enabled[1]) + dev_priv->vdc_irq_mask |= _MDFLD_PIPEB_EVENT_FLAG; + if (dev->vblank_enabled[2]) + dev_priv->vdc_irq_mask |= _MDFLD_PIPEC_EVENT_FLAG; + */ + + /* This register is safe even if display island is off */ + PSB_WVDC32(~dev_priv->vdc_irq_mask, PSB_INT_MASK_R); + spin_unlock_irqrestore(&dev_priv->irqmask_lock, irqflags); +} + +int psb_irq_postinstall(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = + (struct drm_psb_private *) dev->dev_private; + unsigned long irqflags; + + spin_lock_irqsave(&dev_priv->irqmask_lock, irqflags); + + /* This register is safe even if display island is off */ + PSB_WVDC32(dev_priv->vdc_irq_mask, PSB_INT_ENABLE_R); + PSB_WVDC32(0xFFFFFFFF, PSB_HWSTAM); + + if (dev->vblank_enabled[0]) + psb_enable_pipestat(dev_priv, 0, PIPE_VBLANK_INTERRUPT_ENABLE); + else + psb_disable_pipestat(dev_priv, 0, PIPE_VBLANK_INTERRUPT_ENABLE); + + if (dev->vblank_enabled[1]) + psb_enable_pipestat(dev_priv, 1, PIPE_VBLANK_INTERRUPT_ENABLE); + else + psb_disable_pipestat(dev_priv, 1, PIPE_VBLANK_INTERRUPT_ENABLE); + + if (dev->vblank_enabled[2]) + psb_enable_pipestat(dev_priv, 2, PIPE_VBLANK_INTERRUPT_ENABLE); + else + psb_disable_pipestat(dev_priv, 2, PIPE_VBLANK_INTERRUPT_ENABLE); + + spin_unlock_irqrestore(&dev_priv->irqmask_lock, irqflags); + return 0; +} + +void psb_irq_uninstall(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = + (struct drm_psb_private *) dev->dev_private; + unsigned long irqflags; + + spin_lock_irqsave(&dev_priv->irqmask_lock, irqflags); + + PSB_WVDC32(0xFFFFFFFF, PSB_HWSTAM); + + if (dev->vblank_enabled[0]) + psb_disable_pipestat(dev_priv, 0, PIPE_VBLANK_INTERRUPT_ENABLE); + + if (dev->vblank_enabled[1]) + psb_disable_pipestat(dev_priv, 1, PIPE_VBLANK_INTERRUPT_ENABLE); + + if (dev->vblank_enabled[2]) + psb_disable_pipestat(dev_priv, 2, PIPE_VBLANK_INTERRUPT_ENABLE); + + dev_priv->vdc_irq_mask &= _PSB_IRQ_SGX_FLAG | + _PSB_IRQ_MSVDX_FLAG | + _LNC_IRQ_TOPAZ_FLAG; + + /* These two registers are safe even if display island is off */ + PSB_WVDC32(~dev_priv->vdc_irq_mask, PSB_INT_MASK_R); + PSB_WVDC32(dev_priv->vdc_irq_mask, PSB_INT_ENABLE_R); + + wmb(); + + /* This register is safe even if display island is off */ + PSB_WVDC32(PSB_RVDC32(PSB_INT_IDENTITY_R), PSB_INT_IDENTITY_R); + spin_unlock_irqrestore(&dev_priv->irqmask_lock, irqflags); +} + +void psb_irq_turn_on_dpst(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = + (struct drm_psb_private *) dev->dev_private; + u32 hist_reg; + u32 pwm_reg; + + if (gma_power_begin(dev, false)) { + PSB_WVDC32(1 << 31, HISTOGRAM_LOGIC_CONTROL); + hist_reg = PSB_RVDC32(HISTOGRAM_LOGIC_CONTROL); + PSB_WVDC32(1 << 31, HISTOGRAM_INT_CONTROL); + hist_reg = PSB_RVDC32(HISTOGRAM_INT_CONTROL); + + PSB_WVDC32(0x80010100, PWM_CONTROL_LOGIC); + pwm_reg = PSB_RVDC32(PWM_CONTROL_LOGIC); + PSB_WVDC32(pwm_reg | PWM_PHASEIN_ENABLE + | PWM_PHASEIN_INT_ENABLE, + PWM_CONTROL_LOGIC); + pwm_reg = PSB_RVDC32(PWM_CONTROL_LOGIC); + + psb_enable_pipestat(dev_priv, 0, PIPE_DPST_EVENT_ENABLE); + + hist_reg = PSB_RVDC32(HISTOGRAM_INT_CONTROL); + PSB_WVDC32(hist_reg | HISTOGRAM_INT_CTRL_CLEAR, + HISTOGRAM_INT_CONTROL); + pwm_reg = PSB_RVDC32(PWM_CONTROL_LOGIC); + PSB_WVDC32(pwm_reg | 0x80010100 | PWM_PHASEIN_ENABLE, + PWM_CONTROL_LOGIC); + + gma_power_end(dev); + } +} + +int psb_irq_enable_dpst(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = + (struct drm_psb_private *) dev->dev_private; + unsigned long irqflags; + + spin_lock_irqsave(&dev_priv->irqmask_lock, irqflags); + + /* enable DPST */ + mid_enable_pipe_event(dev_priv, 0); + psb_irq_turn_on_dpst(dev); + + spin_unlock_irqrestore(&dev_priv->irqmask_lock, irqflags); + return 0; +} + +void psb_irq_turn_off_dpst(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = + (struct drm_psb_private *) dev->dev_private; + u32 hist_reg; + u32 pwm_reg; + + if (gma_power_begin(dev, false)) { + PSB_WVDC32(0x00000000, HISTOGRAM_INT_CONTROL); + hist_reg = PSB_RVDC32(HISTOGRAM_INT_CONTROL); + + psb_disable_pipestat(dev_priv, 0, PIPE_DPST_EVENT_ENABLE); + + pwm_reg = PSB_RVDC32(PWM_CONTROL_LOGIC); + PSB_WVDC32(pwm_reg & !(PWM_PHASEIN_INT_ENABLE), + PWM_CONTROL_LOGIC); + pwm_reg = PSB_RVDC32(PWM_CONTROL_LOGIC); + + gma_power_end(dev); + } +} + +int psb_irq_disable_dpst(struct drm_device *dev) +{ + struct drm_psb_private *dev_priv = + (struct drm_psb_private *) dev->dev_private; + unsigned long irqflags; + + spin_lock_irqsave(&dev_priv->irqmask_lock, irqflags); + + mid_disable_pipe_event(dev_priv, 0); + psb_irq_turn_off_dpst(dev); + + spin_unlock_irqrestore(&dev_priv->irqmask_lock, irqflags); + + return 0; +} + +#ifdef PSB_FIXME +static int psb_vblank_do_wait(struct drm_device *dev, + unsigned int *sequence, atomic_t *counter) +{ + unsigned int cur_vblank; + int ret = 0; + DRM_WAIT_ON(ret, dev->vbl_queue, 3 * DRM_HZ, + (((cur_vblank = atomic_read(counter)) + - *sequence) <= (1 << 23))); + *sequence = cur_vblank; + + return ret; +} +#endif + +/* + * It is used to enable VBLANK interrupt + */ +int psb_enable_vblank(struct drm_device *dev, int pipe) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + unsigned long irqflags; + uint32_t reg_val = 0; + uint32_t pipeconf_reg = mid_pipeconf(pipe); + + /* Medfield is different - we should perhaps extract out vblank + and blacklight etc ops */ + if (IS_MFLD(dev)) + return mdfld_enable_te(dev, pipe); + + if (gma_power_begin(dev, false)) { + reg_val = REG_READ(pipeconf_reg); + gma_power_end(dev); + } + + if (!(reg_val & PIPEACONF_ENABLE)) + return -EINVAL; + + spin_lock_irqsave(&dev_priv->irqmask_lock, irqflags); + + if (pipe == 0) + dev_priv->vdc_irq_mask |= _PSB_VSYNC_PIPEA_FLAG; + else if (pipe == 1) + dev_priv->vdc_irq_mask |= _PSB_VSYNC_PIPEB_FLAG; + + PSB_WVDC32(~dev_priv->vdc_irq_mask, PSB_INT_MASK_R); + PSB_WVDC32(dev_priv->vdc_irq_mask, PSB_INT_ENABLE_R); + psb_enable_pipestat(dev_priv, pipe, PIPE_VBLANK_INTERRUPT_ENABLE); + + spin_unlock_irqrestore(&dev_priv->irqmask_lock, irqflags); + + return 0; +} + +/* + * It is used to disable VBLANK interrupt + */ +void psb_disable_vblank(struct drm_device *dev, int pipe) +{ + struct drm_psb_private *dev_priv = dev->dev_private; + unsigned long irqflags; + + if (IS_MFLD(dev)) + mdfld_disable_te(dev, pipe); + spin_lock_irqsave(&dev_priv->irqmask_lock, irqflags); + + if (pipe == 0) + dev_priv->vdc_irq_mask &= ~_PSB_VSYNC_PIPEA_FLAG; + else if (pipe == 1) + dev_priv->vdc_irq_mask &= ~_PSB_VSYNC_PIPEB_FLAG; + + PSB_WVDC32(~dev_priv->vdc_irq_mask, PSB_INT_MASK_R); + PSB_WVDC32(dev_priv->vdc_irq_mask, PSB_INT_ENABLE_R); + psb_disable_pipestat(dev_priv, pipe, PIPE_VBLANK_INTERRUPT_ENABLE); + + spin_unlock_irqrestore(&dev_priv->irqmask_lock, irqflags); +} + +/* + * It is used to enable TE interrupt + */ +int mdfld_enable_te(struct drm_device *dev, int pipe) +{ + struct drm_psb_private *dev_priv = + (struct drm_psb_private *) dev->dev_private; + unsigned long irqflags; + uint32_t reg_val = 0; + uint32_t pipeconf_reg = mid_pipeconf(pipe); + + if (gma_power_begin(dev, false)) { + reg_val = REG_READ(pipeconf_reg); + gma_power_end(dev); + } + + if (!(reg_val & PIPEACONF_ENABLE)) + return -EINVAL; + + spin_lock_irqsave(&dev_priv->irqmask_lock, irqflags); + + mid_enable_pipe_event(dev_priv, pipe); + psb_enable_pipestat(dev_priv, pipe, PIPE_TE_ENABLE); + + spin_unlock_irqrestore(&dev_priv->irqmask_lock, irqflags); + + return 0; +} + +/* + * It is used to disable TE interrupt + */ +void mdfld_disable_te(struct drm_device *dev, int pipe) +{ + struct drm_psb_private *dev_priv = + (struct drm_psb_private *) dev->dev_private; + unsigned long irqflags; + + if (!dev_priv->dsr_enable) + return; + + spin_lock_irqsave(&dev_priv->irqmask_lock, irqflags); + + mid_disable_pipe_event(dev_priv, pipe); + psb_disable_pipestat(dev_priv, pipe, PIPE_TE_ENABLE); + + spin_unlock_irqrestore(&dev_priv->irqmask_lock, irqflags); +} + +/* Called from drm generic code, passed a 'crtc', which + * we use as a pipe index + */ +u32 psb_get_vblank_counter(struct drm_device *dev, int pipe) +{ + uint32_t high_frame = PIPEAFRAMEHIGH; + uint32_t low_frame = PIPEAFRAMEPIXEL; + uint32_t pipeconf_reg = PIPEACONF; + uint32_t reg_val = 0; + uint32_t high1 = 0, high2 = 0, low = 0, count = 0; + + switch (pipe) { + case 0: + break; + case 1: + high_frame = PIPEBFRAMEHIGH; + low_frame = PIPEBFRAMEPIXEL; + pipeconf_reg = PIPEBCONF; + break; + case 2: + high_frame = PIPECFRAMEHIGH; + low_frame = PIPECFRAMEPIXEL; + pipeconf_reg = PIPECCONF; + break; + default: + dev_err(dev->dev, "%s, invalid pipe.\n", __func__); + return 0; + } + + if (!gma_power_begin(dev, false)) + return 0; + + reg_val = REG_READ(pipeconf_reg); + + if (!(reg_val & PIPEACONF_ENABLE)) { + dev_err(dev->dev, "trying to get vblank count for disabled pipe %d\n", + pipe); + goto psb_get_vblank_counter_exit; + } + + /* + * High & low register fields aren't synchronized, so make sure + * we get a low value that's stable across two reads of the high + * register. + */ + do { + high1 = ((REG_READ(high_frame) & PIPE_FRAME_HIGH_MASK) >> + PIPE_FRAME_HIGH_SHIFT); + low = ((REG_READ(low_frame) & PIPE_FRAME_LOW_MASK) >> + PIPE_FRAME_LOW_SHIFT); + high2 = ((REG_READ(high_frame) & PIPE_FRAME_HIGH_MASK) >> + PIPE_FRAME_HIGH_SHIFT); + } while (high1 != high2); + + count = (high1 << 8) | low; + +psb_get_vblank_counter_exit: + + gma_power_end(dev); + + return count; +} + diff --git a/drivers/gpu/drm/gma500/psb_irq.h b/drivers/gpu/drm/gma500/psb_irq.h new file mode 100644 index 0000000..603045b --- /dev/null +++ b/drivers/gpu/drm/gma500/psb_irq.h @@ -0,0 +1,47 @@ +/************************************************************************** + * Copyright (c) 2009-2011, Intel Corporation. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + * Authors: + * Benjamin Defnet + * Rajesh Poornachandran + * + **************************************************************************/ + +#ifndef _SYSIRQ_H_ +#define _SYSIRQ_H_ + +#include + +bool sysirq_init(struct drm_device *dev); +void sysirq_uninit(struct drm_device *dev); + +void psb_irq_preinstall(struct drm_device *dev); +int psb_irq_postinstall(struct drm_device *dev); +void psb_irq_uninstall(struct drm_device *dev); +irqreturn_t psb_irq_handler(DRM_IRQ_ARGS); + +int psb_irq_enable_dpst(struct drm_device *dev); +int psb_irq_disable_dpst(struct drm_device *dev); +void psb_irq_turn_on_dpst(struct drm_device *dev); +void psb_irq_turn_off_dpst(struct drm_device *dev); +int psb_enable_vblank(struct drm_device *dev, int pipe); +void psb_disable_vblank(struct drm_device *dev, int pipe); +u32 psb_get_vblank_counter(struct drm_device *dev, int pipe); + +int mdfld_enable_te(struct drm_device *dev, int pipe); +void mdfld_disable_te(struct drm_device *dev, int pipe); +#endif /* _SYSIRQ_H_ */ diff --git a/drivers/gpu/drm/gma500/psb_lid.c b/drivers/gpu/drm/gma500/psb_lid.c new file mode 100644 index 0000000..b867aabe --- /dev/null +++ b/drivers/gpu/drm/gma500/psb_lid.c @@ -0,0 +1,88 @@ +/************************************************************************** + * Copyright (c) 2007, Intel Corporation. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + * Authors: Thomas Hellstrom + **************************************************************************/ + +#include +#include "psb_drv.h" +#include "psb_reg.h" +#include "psb_intel_reg.h" +#include + +static void psb_lid_timer_func(unsigned long data) +{ + struct drm_psb_private * dev_priv = (struct drm_psb_private *)data; + struct drm_device *dev = (struct drm_device *)dev_priv->dev; + struct timer_list *lid_timer = &dev_priv->lid_timer; + unsigned long irq_flags; + u32 *lid_state = dev_priv->lid_state; + u32 pp_status; + + if (readl(lid_state) == dev_priv->lid_last_state) + goto lid_timer_schedule; + + if ((readl(lid_state)) & 0x01) { + /*lid state is open*/ + REG_WRITE(PP_CONTROL, REG_READ(PP_CONTROL) | POWER_TARGET_ON); + do { + pp_status = REG_READ(PP_STATUS); + } while ((pp_status & PP_ON) == 0); + + /*FIXME: should be backlight level before*/ + psb_intel_lvds_set_brightness(dev, 100); + } else { + psb_intel_lvds_set_brightness(dev, 0); + + REG_WRITE(PP_CONTROL, REG_READ(PP_CONTROL) & ~POWER_TARGET_ON); + do { + pp_status = REG_READ(PP_STATUS); + } while ((pp_status & PP_ON) == 0); + } + dev_priv->lid_last_state = readl(lid_state); + +lid_timer_schedule: + spin_lock_irqsave(&dev_priv->lid_lock, irq_flags); + if (!timer_pending(lid_timer)) { + lid_timer->expires = jiffies + PSB_LID_DELAY; + add_timer(lid_timer); + } + spin_unlock_irqrestore(&dev_priv->lid_lock, irq_flags); +} + +void psb_lid_timer_init(struct drm_psb_private *dev_priv) +{ + struct timer_list *lid_timer = &dev_priv->lid_timer; + unsigned long irq_flags; + + spin_lock_init(&dev_priv->lid_lock); + spin_lock_irqsave(&dev_priv->lid_lock, irq_flags); + + init_timer(lid_timer); + + lid_timer->data = (unsigned long)dev_priv; + lid_timer->function = psb_lid_timer_func; + lid_timer->expires = jiffies + PSB_LID_DELAY; + + add_timer(lid_timer); + spin_unlock_irqrestore(&dev_priv->lid_lock, irq_flags); +} + +void psb_lid_timer_takedown(struct drm_psb_private *dev_priv) +{ + del_timer_sync(&dev_priv->lid_timer); +} + diff --git a/drivers/gpu/drm/gma500/psb_reg.h b/drivers/gpu/drm/gma500/psb_reg.h new file mode 100644 index 0000000..b81c7c1 --- /dev/null +++ b/drivers/gpu/drm/gma500/psb_reg.h @@ -0,0 +1,582 @@ +/************************************************************************** + * + * Copyright (c) (2005-2007) Imagination Technologies Limited. + * Copyright (c) 2007, Intel Corporation. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.. + * + **************************************************************************/ + +#ifndef _PSB_REG_H_ +#define _PSB_REG_H_ + +#define PSB_CR_CLKGATECTL 0x0000 +#define _PSB_C_CLKGATECTL_AUTO_MAN_REG (1 << 24) +#define _PSB_C_CLKGATECTL_USE_CLKG_SHIFT (20) +#define _PSB_C_CLKGATECTL_USE_CLKG_MASK (0x3 << 20) +#define _PSB_C_CLKGATECTL_DPM_CLKG_SHIFT (16) +#define _PSB_C_CLKGATECTL_DPM_CLKG_MASK (0x3 << 16) +#define _PSB_C_CLKGATECTL_TA_CLKG_SHIFT (12) +#define _PSB_C_CLKGATECTL_TA_CLKG_MASK (0x3 << 12) +#define _PSB_C_CLKGATECTL_TSP_CLKG_SHIFT (8) +#define _PSB_C_CLKGATECTL_TSP_CLKG_MASK (0x3 << 8) +#define _PSB_C_CLKGATECTL_ISP_CLKG_SHIFT (4) +#define _PSB_C_CLKGATECTL_ISP_CLKG_MASK (0x3 << 4) +#define _PSB_C_CLKGATECTL_2D_CLKG_SHIFT (0) +#define _PSB_C_CLKGATECTL_2D_CLKG_MASK (0x3 << 0) +#define _PSB_C_CLKGATECTL_CLKG_ENABLED (0) +#define _PSB_C_CLKGATECTL_CLKG_DISABLED (1) +#define _PSB_C_CLKGATECTL_CLKG_AUTO (2) + +#define PSB_CR_CORE_ID 0x0010 +#define _PSB_CC_ID_ID_SHIFT (16) +#define _PSB_CC_ID_ID_MASK (0xFFFF << 16) +#define _PSB_CC_ID_CONFIG_SHIFT (0) +#define _PSB_CC_ID_CONFIG_MASK (0xFFFF << 0) + +#define PSB_CR_CORE_REVISION 0x0014 +#define _PSB_CC_REVISION_DESIGNER_SHIFT (24) +#define _PSB_CC_REVISION_DESIGNER_MASK (0xFF << 24) +#define _PSB_CC_REVISION_MAJOR_SHIFT (16) +#define _PSB_CC_REVISION_MAJOR_MASK (0xFF << 16) +#define _PSB_CC_REVISION_MINOR_SHIFT (8) +#define _PSB_CC_REVISION_MINOR_MASK (0xFF << 8) +#define _PSB_CC_REVISION_MAINTENANCE_SHIFT (0) +#define _PSB_CC_REVISION_MAINTENANCE_MASK (0xFF << 0) + +#define PSB_CR_DESIGNER_REV_FIELD1 0x0018 + +#define PSB_CR_SOFT_RESET 0x0080 +#define _PSB_CS_RESET_TSP_RESET (1 << 6) +#define _PSB_CS_RESET_ISP_RESET (1 << 5) +#define _PSB_CS_RESET_USE_RESET (1 << 4) +#define _PSB_CS_RESET_TA_RESET (1 << 3) +#define _PSB_CS_RESET_DPM_RESET (1 << 2) +#define _PSB_CS_RESET_TWOD_RESET (1 << 1) +#define _PSB_CS_RESET_BIF_RESET (1 << 0) + +#define PSB_CR_DESIGNER_REV_FIELD2 0x001C + +#define PSB_CR_EVENT_HOST_ENABLE2 0x0110 + +#define PSB_CR_EVENT_STATUS2 0x0118 + +#define PSB_CR_EVENT_HOST_CLEAR2 0x0114 +#define _PSB_CE2_BIF_REQUESTER_FAULT (1 << 4) + +#define PSB_CR_EVENT_STATUS 0x012C + +#define PSB_CR_EVENT_HOST_ENABLE 0x0130 + +#define PSB_CR_EVENT_HOST_CLEAR 0x0134 +#define _PSB_CE_MASTER_INTERRUPT (1 << 31) +#define _PSB_CE_TA_DPM_FAULT (1 << 28) +#define _PSB_CE_TWOD_COMPLETE (1 << 27) +#define _PSB_CE_DPM_OUT_OF_MEMORY_ZLS (1 << 25) +#define _PSB_CE_DPM_TA_MEM_FREE (1 << 24) +#define _PSB_CE_PIXELBE_END_RENDER (1 << 18) +#define _PSB_CE_SW_EVENT (1 << 14) +#define _PSB_CE_TA_FINISHED (1 << 13) +#define _PSB_CE_TA_TERMINATE (1 << 12) +#define _PSB_CE_DPM_REACHED_MEM_THRESH (1 << 3) +#define _PSB_CE_DPM_OUT_OF_MEMORY_GBL (1 << 2) +#define _PSB_CE_DPM_OUT_OF_MEMORY_MT (1 << 1) +#define _PSB_CE_DPM_3D_MEM_FREE (1 << 0) + + +#define PSB_USE_OFFSET_MASK 0x0007FFFF +#define PSB_USE_OFFSET_SIZE (PSB_USE_OFFSET_MASK + 1) +#define PSB_CR_USE_CODE_BASE0 0x0A0C +#define PSB_CR_USE_CODE_BASE1 0x0A10 +#define PSB_CR_USE_CODE_BASE2 0x0A14 +#define PSB_CR_USE_CODE_BASE3 0x0A18 +#define PSB_CR_USE_CODE_BASE4 0x0A1C +#define PSB_CR_USE_CODE_BASE5 0x0A20 +#define PSB_CR_USE_CODE_BASE6 0x0A24 +#define PSB_CR_USE_CODE_BASE7 0x0A28 +#define PSB_CR_USE_CODE_BASE8 0x0A2C +#define PSB_CR_USE_CODE_BASE9 0x0A30 +#define PSB_CR_USE_CODE_BASE10 0x0A34 +#define PSB_CR_USE_CODE_BASE11 0x0A38 +#define PSB_CR_USE_CODE_BASE12 0x0A3C +#define PSB_CR_USE_CODE_BASE13 0x0A40 +#define PSB_CR_USE_CODE_BASE14 0x0A44 +#define PSB_CR_USE_CODE_BASE15 0x0A48 +#define PSB_CR_USE_CODE_BASE(_i) (0x0A0C + ((_i) << 2)) +#define _PSB_CUC_BASE_DM_SHIFT (25) +#define _PSB_CUC_BASE_DM_MASK (0x3 << 25) +#define _PSB_CUC_BASE_ADDR_SHIFT (0) /* 1024-bit aligned address? */ +#define _PSB_CUC_BASE_ADDR_ALIGNSHIFT (7) +#define _PSB_CUC_BASE_ADDR_MASK (0x1FFFFFF << 0) +#define _PSB_CUC_DM_VERTEX (0) +#define _PSB_CUC_DM_PIXEL (1) +#define _PSB_CUC_DM_RESERVED (2) +#define _PSB_CUC_DM_EDM (3) + +#define PSB_CR_PDS_EXEC_BASE 0x0AB8 +#define _PSB_CR_PDS_EXEC_BASE_ADDR_SHIFT (20) /* 1MB aligned address */ +#define _PSB_CR_PDS_EXEC_BASE_ADDR_ALIGNSHIFT (20) + +#define PSB_CR_EVENT_KICKER 0x0AC4 +#define _PSB_CE_KICKER_ADDRESS_SHIFT (4) /* 128-bit aligned address */ + +#define PSB_CR_EVENT_KICK 0x0AC8 +#define _PSB_CE_KICK_NOW (1 << 0) + +#define PSB_CR_BIF_DIR_LIST_BASE1 0x0C38 + +#define PSB_CR_BIF_CTRL 0x0C00 +#define _PSB_CB_CTRL_CLEAR_FAULT (1 << 4) +#define _PSB_CB_CTRL_INVALDC (1 << 3) +#define _PSB_CB_CTRL_FLUSH (1 << 2) + +#define PSB_CR_BIF_INT_STAT 0x0C04 + +#define PSB_CR_BIF_FAULT 0x0C08 +#define _PSB_CBI_STAT_PF_N_RW (1 << 14) +#define _PSB_CBI_STAT_FAULT_SHIFT (0) +#define _PSB_CBI_STAT_FAULT_MASK (0x3FFF << 0) +#define _PSB_CBI_STAT_FAULT_CACHE (1 << 1) +#define _PSB_CBI_STAT_FAULT_TA (1 << 2) +#define _PSB_CBI_STAT_FAULT_VDM (1 << 3) +#define _PSB_CBI_STAT_FAULT_2D (1 << 4) +#define _PSB_CBI_STAT_FAULT_PBE (1 << 5) +#define _PSB_CBI_STAT_FAULT_TSP (1 << 6) +#define _PSB_CBI_STAT_FAULT_ISP (1 << 7) +#define _PSB_CBI_STAT_FAULT_USSEPDS (1 << 8) +#define _PSB_CBI_STAT_FAULT_HOST (1 << 9) + +#define PSB_CR_BIF_BANK0 0x0C78 +#define PSB_CR_BIF_BANK1 0x0C7C +#define PSB_CR_BIF_DIR_LIST_BASE0 0x0C84 +#define PSB_CR_BIF_TWOD_REQ_BASE 0x0C88 +#define PSB_CR_BIF_3D_REQ_BASE 0x0CAC + +#define PSB_CR_2D_SOCIF 0x0E18 +#define _PSB_C2_SOCIF_FREESPACE_SHIFT (0) +#define _PSB_C2_SOCIF_FREESPACE_MASK (0xFF << 0) +#define _PSB_C2_SOCIF_EMPTY (0x80 << 0) + +#define PSB_CR_2D_BLIT_STATUS 0x0E04 +#define _PSB_C2B_STATUS_BUSY (1 << 24) +#define _PSB_C2B_STATUS_COMPLETE_SHIFT (0) +#define _PSB_C2B_STATUS_COMPLETE_MASK (0xFFFFFF << 0) + +/* + * 2D defs. + */ + +/* + * 2D Slave Port Data : Block Header's Object Type + */ + +#define PSB_2D_CLIP_BH (0x00000000) +#define PSB_2D_PAT_BH (0x10000000) +#define PSB_2D_CTRL_BH (0x20000000) +#define PSB_2D_SRC_OFF_BH (0x30000000) +#define PSB_2D_MASK_OFF_BH (0x40000000) +#define PSB_2D_RESERVED1_BH (0x50000000) +#define PSB_2D_RESERVED2_BH (0x60000000) +#define PSB_2D_FENCE_BH (0x70000000) +#define PSB_2D_BLIT_BH (0x80000000) +#define PSB_2D_SRC_SURF_BH (0x90000000) +#define PSB_2D_DST_SURF_BH (0xA0000000) +#define PSB_2D_PAT_SURF_BH (0xB0000000) +#define PSB_2D_SRC_PAL_BH (0xC0000000) +#define PSB_2D_PAT_PAL_BH (0xD0000000) +#define PSB_2D_MASK_SURF_BH (0xE0000000) +#define PSB_2D_FLUSH_BH (0xF0000000) + +/* + * Clip Definition block (PSB_2D_CLIP_BH) + */ +#define PSB_2D_CLIPCOUNT_MAX (1) +#define PSB_2D_CLIPCOUNT_MASK (0x00000000) +#define PSB_2D_CLIPCOUNT_CLRMASK (0xFFFFFFFF) +#define PSB_2D_CLIPCOUNT_SHIFT (0) +/* clip rectangle min & max */ +#define PSB_2D_CLIP_XMAX_MASK (0x00FFF000) +#define PSB_2D_CLIP_XMAX_CLRMASK (0xFF000FFF) +#define PSB_2D_CLIP_XMAX_SHIFT (12) +#define PSB_2D_CLIP_XMIN_MASK (0x00000FFF) +#define PSB_2D_CLIP_XMIN_CLRMASK (0x00FFF000) +#define PSB_2D_CLIP_XMIN_SHIFT (0) +/* clip rectangle offset */ +#define PSB_2D_CLIP_YMAX_MASK (0x00FFF000) +#define PSB_2D_CLIP_YMAX_CLRMASK (0xFF000FFF) +#define PSB_2D_CLIP_YMAX_SHIFT (12) +#define PSB_2D_CLIP_YMIN_MASK (0x00000FFF) +#define PSB_2D_CLIP_YMIN_CLRMASK (0x00FFF000) +#define PSB_2D_CLIP_YMIN_SHIFT (0) + +/* + * Pattern Control (PSB_2D_PAT_BH) + */ +#define PSB_2D_PAT_HEIGHT_MASK (0x0000001F) +#define PSB_2D_PAT_HEIGHT_SHIFT (0) +#define PSB_2D_PAT_WIDTH_MASK (0x000003E0) +#define PSB_2D_PAT_WIDTH_SHIFT (5) +#define PSB_2D_PAT_YSTART_MASK (0x00007C00) +#define PSB_2D_PAT_YSTART_SHIFT (10) +#define PSB_2D_PAT_XSTART_MASK (0x000F8000) +#define PSB_2D_PAT_XSTART_SHIFT (15) + +/* + * 2D Control block (PSB_2D_CTRL_BH) + */ +/* Present Flags */ +#define PSB_2D_SRCCK_CTRL (0x00000001) +#define PSB_2D_DSTCK_CTRL (0x00000002) +#define PSB_2D_ALPHA_CTRL (0x00000004) +/* Colour Key Colour (SRC/DST)*/ +#define PSB_2D_CK_COL_MASK (0xFFFFFFFF) +#define PSB_2D_CK_COL_CLRMASK (0x00000000) +#define PSB_2D_CK_COL_SHIFT (0) +/* Colour Key Mask (SRC/DST)*/ +#define PSB_2D_CK_MASK_MASK (0xFFFFFFFF) +#define PSB_2D_CK_MASK_CLRMASK (0x00000000) +#define PSB_2D_CK_MASK_SHIFT (0) +/* Alpha Control (Alpha/RGB)*/ +#define PSB_2D_GBLALPHA_MASK (0x000FF000) +#define PSB_2D_GBLALPHA_CLRMASK (0xFFF00FFF) +#define PSB_2D_GBLALPHA_SHIFT (12) +#define PSB_2D_SRCALPHA_OP_MASK (0x00700000) +#define PSB_2D_SRCALPHA_OP_CLRMASK (0xFF8FFFFF) +#define PSB_2D_SRCALPHA_OP_SHIFT (20) +#define PSB_2D_SRCALPHA_OP_ONE (0x00000000) +#define PSB_2D_SRCALPHA_OP_SRC (0x00100000) +#define PSB_2D_SRCALPHA_OP_DST (0x00200000) +#define PSB_2D_SRCALPHA_OP_SG (0x00300000) +#define PSB_2D_SRCALPHA_OP_DG (0x00400000) +#define PSB_2D_SRCALPHA_OP_GBL (0x00500000) +#define PSB_2D_SRCALPHA_OP_ZERO (0x00600000) +#define PSB_2D_SRCALPHA_INVERT (0x00800000) +#define PSB_2D_SRCALPHA_INVERT_CLR (0xFF7FFFFF) +#define PSB_2D_DSTALPHA_OP_MASK (0x07000000) +#define PSB_2D_DSTALPHA_OP_CLRMASK (0xF8FFFFFF) +#define PSB_2D_DSTALPHA_OP_SHIFT (24) +#define PSB_2D_DSTALPHA_OP_ONE (0x00000000) +#define PSB_2D_DSTALPHA_OP_SRC (0x01000000) +#define PSB_2D_DSTALPHA_OP_DST (0x02000000) +#define PSB_2D_DSTALPHA_OP_SG (0x03000000) +#define PSB_2D_DSTALPHA_OP_DG (0x04000000) +#define PSB_2D_DSTALPHA_OP_GBL (0x05000000) +#define PSB_2D_DSTALPHA_OP_ZERO (0x06000000) +#define PSB_2D_DSTALPHA_INVERT (0x08000000) +#define PSB_2D_DSTALPHA_INVERT_CLR (0xF7FFFFFF) + +#define PSB_2D_PRE_MULTIPLICATION_ENABLE (0x10000000) +#define PSB_2D_PRE_MULTIPLICATION_CLRMASK (0xEFFFFFFF) +#define PSB_2D_ZERO_SOURCE_ALPHA_ENABLE (0x20000000) +#define PSB_2D_ZERO_SOURCE_ALPHA_CLRMASK (0xDFFFFFFF) + +/* + *Source Offset (PSB_2D_SRC_OFF_BH) + */ +#define PSB_2D_SRCOFF_XSTART_MASK ((0x00000FFF) << 12) +#define PSB_2D_SRCOFF_XSTART_SHIFT (12) +#define PSB_2D_SRCOFF_YSTART_MASK (0x00000FFF) +#define PSB_2D_SRCOFF_YSTART_SHIFT (0) + +/* + * Mask Offset (PSB_2D_MASK_OFF_BH) + */ +#define PSB_2D_MASKOFF_XSTART_MASK ((0x00000FFF) << 12) +#define PSB_2D_MASKOFF_XSTART_SHIFT (12) +#define PSB_2D_MASKOFF_YSTART_MASK (0x00000FFF) +#define PSB_2D_MASKOFF_YSTART_SHIFT (0) + +/* + * 2D Fence (see PSB_2D_FENCE_BH): bits 0:27 are ignored + */ + +/* + *Blit Rectangle (PSB_2D_BLIT_BH) + */ + +#define PSB_2D_ROT_MASK (3 << 25) +#define PSB_2D_ROT_CLRMASK (~PSB_2D_ROT_MASK) +#define PSB_2D_ROT_NONE (0 << 25) +#define PSB_2D_ROT_90DEGS (1 << 25) +#define PSB_2D_ROT_180DEGS (2 << 25) +#define PSB_2D_ROT_270DEGS (3 << 25) + +#define PSB_2D_COPYORDER_MASK (3 << 23) +#define PSB_2D_COPYORDER_CLRMASK (~PSB_2D_COPYORDER_MASK) +#define PSB_2D_COPYORDER_TL2BR (0 << 23) +#define PSB_2D_COPYORDER_BR2TL (1 << 23) +#define PSB_2D_COPYORDER_TR2BL (2 << 23) +#define PSB_2D_COPYORDER_BL2TR (3 << 23) + +#define PSB_2D_DSTCK_CLRMASK (0xFF9FFFFF) +#define PSB_2D_DSTCK_DISABLE (0x00000000) +#define PSB_2D_DSTCK_PASS (0x00200000) +#define PSB_2D_DSTCK_REJECT (0x00400000) + +#define PSB_2D_SRCCK_CLRMASK (0xFFE7FFFF) +#define PSB_2D_SRCCK_DISABLE (0x00000000) +#define PSB_2D_SRCCK_PASS (0x00080000) +#define PSB_2D_SRCCK_REJECT (0x00100000) + +#define PSB_2D_CLIP_ENABLE (0x00040000) + +#define PSB_2D_ALPHA_ENABLE (0x00020000) + +#define PSB_2D_PAT_CLRMASK (0xFFFEFFFF) +#define PSB_2D_PAT_MASK (0x00010000) +#define PSB_2D_USE_PAT (0x00010000) +#define PSB_2D_USE_FILL (0x00000000) +/* + * Tungsten Graphics note on rop codes: If rop A and rop B are + * identical, the mask surface will not be read and need not be + * set up. + */ + +#define PSB_2D_ROP3B_MASK (0x0000FF00) +#define PSB_2D_ROP3B_CLRMASK (0xFFFF00FF) +#define PSB_2D_ROP3B_SHIFT (8) +/* rop code A */ +#define PSB_2D_ROP3A_MASK (0x000000FF) +#define PSB_2D_ROP3A_CLRMASK (0xFFFFFF00) +#define PSB_2D_ROP3A_SHIFT (0) + +#define PSB_2D_ROP4_MASK (0x0000FFFF) +/* + * DWORD0: (Only pass if Pattern control == Use Fill Colour) + * Fill Colour RGBA8888 + */ +#define PSB_2D_FILLCOLOUR_MASK (0xFFFFFFFF) +#define PSB_2D_FILLCOLOUR_SHIFT (0) +/* + * DWORD1: (Always Present) + * X Start (Dest) + * Y Start (Dest) + */ +#define PSB_2D_DST_XSTART_MASK (0x00FFF000) +#define PSB_2D_DST_XSTART_CLRMASK (0xFF000FFF) +#define PSB_2D_DST_XSTART_SHIFT (12) +#define PSB_2D_DST_YSTART_MASK (0x00000FFF) +#define PSB_2D_DST_YSTART_CLRMASK (0xFFFFF000) +#define PSB_2D_DST_YSTART_SHIFT (0) +/* + * DWORD2: (Always Present) + * X Size (Dest) + * Y Size (Dest) + */ +#define PSB_2D_DST_XSIZE_MASK (0x00FFF000) +#define PSB_2D_DST_XSIZE_CLRMASK (0xFF000FFF) +#define PSB_2D_DST_XSIZE_SHIFT (12) +#define PSB_2D_DST_YSIZE_MASK (0x00000FFF) +#define PSB_2D_DST_YSIZE_CLRMASK (0xFFFFF000) +#define PSB_2D_DST_YSIZE_SHIFT (0) + +/* + * Source Surface (PSB_2D_SRC_SURF_BH) + */ +/* + * WORD 0 + */ + +#define PSB_2D_SRC_FORMAT_MASK (0x00078000) +#define PSB_2D_SRC_1_PAL (0x00000000) +#define PSB_2D_SRC_2_PAL (0x00008000) +#define PSB_2D_SRC_4_PAL (0x00010000) +#define PSB_2D_SRC_8_PAL (0x00018000) +#define PSB_2D_SRC_8_ALPHA (0x00020000) +#define PSB_2D_SRC_4_ALPHA (0x00028000) +#define PSB_2D_SRC_332RGB (0x00030000) +#define PSB_2D_SRC_4444ARGB (0x00038000) +#define PSB_2D_SRC_555RGB (0x00040000) +#define PSB_2D_SRC_1555ARGB (0x00048000) +#define PSB_2D_SRC_565RGB (0x00050000) +#define PSB_2D_SRC_0888ARGB (0x00058000) +#define PSB_2D_SRC_8888ARGB (0x00060000) +#define PSB_2D_SRC_8888UYVY (0x00068000) +#define PSB_2D_SRC_RESERVED (0x00070000) +#define PSB_2D_SRC_1555ARGB_LOOKUP (0x00078000) + + +#define PSB_2D_SRC_STRIDE_MASK (0x00007FFF) +#define PSB_2D_SRC_STRIDE_CLRMASK (0xFFFF8000) +#define PSB_2D_SRC_STRIDE_SHIFT (0) +/* + * WORD 1 - Base Address + */ +#define PSB_2D_SRC_ADDR_MASK (0x0FFFFFFC) +#define PSB_2D_SRC_ADDR_CLRMASK (0x00000003) +#define PSB_2D_SRC_ADDR_SHIFT (2) +#define PSB_2D_SRC_ADDR_ALIGNSHIFT (2) + +/* + * Pattern Surface (PSB_2D_PAT_SURF_BH) + */ +/* + * WORD 0 + */ + +#define PSB_2D_PAT_FORMAT_MASK (0x00078000) +#define PSB_2D_PAT_1_PAL (0x00000000) +#define PSB_2D_PAT_2_PAL (0x00008000) +#define PSB_2D_PAT_4_PAL (0x00010000) +#define PSB_2D_PAT_8_PAL (0x00018000) +#define PSB_2D_PAT_8_ALPHA (0x00020000) +#define PSB_2D_PAT_4_ALPHA (0x00028000) +#define PSB_2D_PAT_332RGB (0x00030000) +#define PSB_2D_PAT_4444ARGB (0x00038000) +#define PSB_2D_PAT_555RGB (0x00040000) +#define PSB_2D_PAT_1555ARGB (0x00048000) +#define PSB_2D_PAT_565RGB (0x00050000) +#define PSB_2D_PAT_0888ARGB (0x00058000) +#define PSB_2D_PAT_8888ARGB (0x00060000) + +#define PSB_2D_PAT_STRIDE_MASK (0x00007FFF) +#define PSB_2D_PAT_STRIDE_CLRMASK (0xFFFF8000) +#define PSB_2D_PAT_STRIDE_SHIFT (0) +/* + * WORD 1 - Base Address + */ +#define PSB_2D_PAT_ADDR_MASK (0x0FFFFFFC) +#define PSB_2D_PAT_ADDR_CLRMASK (0x00000003) +#define PSB_2D_PAT_ADDR_SHIFT (2) +#define PSB_2D_PAT_ADDR_ALIGNSHIFT (2) + +/* + * Destination Surface (PSB_2D_DST_SURF_BH) + */ +/* + * WORD 0 + */ + +#define PSB_2D_DST_FORMAT_MASK (0x00078000) +#define PSB_2D_DST_332RGB (0x00030000) +#define PSB_2D_DST_4444ARGB (0x00038000) +#define PSB_2D_DST_555RGB (0x00040000) +#define PSB_2D_DST_1555ARGB (0x00048000) +#define PSB_2D_DST_565RGB (0x00050000) +#define PSB_2D_DST_0888ARGB (0x00058000) +#define PSB_2D_DST_8888ARGB (0x00060000) +#define PSB_2D_DST_8888AYUV (0x00070000) + +#define PSB_2D_DST_STRIDE_MASK (0x00007FFF) +#define PSB_2D_DST_STRIDE_CLRMASK (0xFFFF8000) +#define PSB_2D_DST_STRIDE_SHIFT (0) +/* + * WORD 1 - Base Address + */ +#define PSB_2D_DST_ADDR_MASK (0x0FFFFFFC) +#define PSB_2D_DST_ADDR_CLRMASK (0x00000003) +#define PSB_2D_DST_ADDR_SHIFT (2) +#define PSB_2D_DST_ADDR_ALIGNSHIFT (2) + +/* + * Mask Surface (PSB_2D_MASK_SURF_BH) + */ +/* + * WORD 0 + */ +#define PSB_2D_MASK_STRIDE_MASK (0x00007FFF) +#define PSB_2D_MASK_STRIDE_CLRMASK (0xFFFF8000) +#define PSB_2D_MASK_STRIDE_SHIFT (0) +/* + * WORD 1 - Base Address + */ +#define PSB_2D_MASK_ADDR_MASK (0x0FFFFFFC) +#define PSB_2D_MASK_ADDR_CLRMASK (0x00000003) +#define PSB_2D_MASK_ADDR_SHIFT (2) +#define PSB_2D_MASK_ADDR_ALIGNSHIFT (2) + +/* + * Source Palette (PSB_2D_SRC_PAL_BH) + */ + +#define PSB_2D_SRCPAL_ADDR_SHIFT (0) +#define PSB_2D_SRCPAL_ADDR_CLRMASK (0xF0000007) +#define PSB_2D_SRCPAL_ADDR_MASK (0x0FFFFFF8) +#define PSB_2D_SRCPAL_BYTEALIGN (1024) + +/* + * Pattern Palette (PSB_2D_PAT_PAL_BH) + */ + +#define PSB_2D_PATPAL_ADDR_SHIFT (0) +#define PSB_2D_PATPAL_ADDR_CLRMASK (0xF0000007) +#define PSB_2D_PATPAL_ADDR_MASK (0x0FFFFFF8) +#define PSB_2D_PATPAL_BYTEALIGN (1024) + +/* + * Rop3 Codes (2 LS bytes) + */ + +#define PSB_2D_ROP3_SRCCOPY (0xCCCC) +#define PSB_2D_ROP3_PATCOPY (0xF0F0) +#define PSB_2D_ROP3_WHITENESS (0xFFFF) +#define PSB_2D_ROP3_BLACKNESS (0x0000) +#define PSB_2D_ROP3_SRC (0xCC) +#define PSB_2D_ROP3_PAT (0xF0) +#define PSB_2D_ROP3_DST (0xAA) + +/* + * Sizes. + */ + +#define PSB_SCENE_HW_COOKIE_SIZE 16 +#define PSB_TA_MEM_HW_COOKIE_SIZE 16 + +/* + * Scene stuff. + */ + +#define PSB_NUM_HW_SCENES 2 + +/* + * Scheduler completion actions. + */ + +#define PSB_RASTER_BLOCK 0 +#define PSB_RASTER 1 +#define PSB_RETURN 2 +#define PSB_TA 3 + +/* Power management */ +#define PSB_PUNIT_PORT 0x04 +#define PSB_OSPMBA 0x78 +#define PSB_APMBA 0x7a +#define PSB_APM_CMD 0x0 +#define PSB_APM_STS 0x04 +#define PSB_PWRGT_VID_ENC_MASK 0x30 +#define PSB_PWRGT_VID_DEC_MASK 0xc +#define PSB_PWRGT_GL3_MASK 0xc0 + +#define PSB_PM_SSC 0x20 +#define PSB_PM_SSS 0x30 +#define PSB_PWRGT_DISPLAY_MASK 0xc /*on a different BA than video/gfx*/ +#define MDFLD_PWRGT_DISPLAY_A_CNTR 0x0000000c +#define MDFLD_PWRGT_DISPLAY_B_CNTR 0x0000c000 +#define MDFLD_PWRGT_DISPLAY_C_CNTR 0x00030000 +#define MDFLD_PWRGT_DISP_MIPI_CNTR 0x000c0000 +#define MDFLD_PWRGT_DISPLAY_CNTR (MDFLD_PWRGT_DISPLAY_A_CNTR | MDFLD_PWRGT_DISPLAY_B_CNTR | MDFLD_PWRGT_DISPLAY_C_CNTR | MDFLD_PWRGT_DISP_MIPI_CNTR) /* 0x000fc00c */ +/* Display SSS register bits are different in A0 vs. B0 */ +#define PSB_PWRGT_GFX_MASK 0x3 +#define MDFLD_PWRGT_DISPLAY_A_STS 0x000000c0 +#define MDFLD_PWRGT_DISPLAY_B_STS 0x00000300 +#define MDFLD_PWRGT_DISPLAY_C_STS 0x00000c00 +#define PSB_PWRGT_GFX_MASK_B0 0xc3 +#define MDFLD_PWRGT_DISPLAY_A_STS_B0 0x0000000c +#define MDFLD_PWRGT_DISPLAY_B_STS_B0 0x0000c000 +#define MDFLD_PWRGT_DISPLAY_C_STS_B0 0x00030000 +#define MDFLD_PWRGT_DISP_MIPI_STS 0x000c0000 +#define MDFLD_PWRGT_DISPLAY_STS_A0 (MDFLD_PWRGT_DISPLAY_A_STS | MDFLD_PWRGT_DISPLAY_B_STS | MDFLD_PWRGT_DISPLAY_C_STS | MDFLD_PWRGT_DISP_MIPI_STS) /* 0x000fc00c */ +#define MDFLD_PWRGT_DISPLAY_STS_B0 (MDFLD_PWRGT_DISPLAY_A_STS_B0 | MDFLD_PWRGT_DISPLAY_B_STS_B0 | MDFLD_PWRGT_DISPLAY_C_STS_B0 | MDFLD_PWRGT_DISP_MIPI_STS) /* 0x000fc00c */ +#endif diff --git a/drivers/gpu/drm/gma500/tc35876x-dsi-lvds.c b/drivers/gpu/drm/gma500/tc35876x-dsi-lvds.c new file mode 100644 index 0000000..4a07ab5 --- /dev/null +++ b/drivers/gpu/drm/gma500/tc35876x-dsi-lvds.c @@ -0,0 +1,829 @@ +/* + * Copyright © 2011 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + */ + +#include "mdfld_dsi_dpi.h" +#include "mdfld_output.h" +#include "mdfld_dsi_pkg_sender.h" +#include "tc35876x-dsi-lvds.h" +#include +#include +#include +#include + +static struct i2c_client *tc35876x_client; +static struct i2c_client *cmi_lcd_i2c_client; + +#define FLD_MASK(start, end) (((1 << ((start) - (end) + 1)) - 1) << (end)) +#define FLD_VAL(val, start, end) (((val) << (end)) & FLD_MASK(start, end)) + +/* DSI D-PHY Layer Registers */ +#define D0W_DPHYCONTTX 0x0004 +#define CLW_DPHYCONTRX 0x0020 +#define D0W_DPHYCONTRX 0x0024 +#define D1W_DPHYCONTRX 0x0028 +#define D2W_DPHYCONTRX 0x002C +#define D3W_DPHYCONTRX 0x0030 +#define COM_DPHYCONTRX 0x0038 +#define CLW_CNTRL 0x0040 +#define D0W_CNTRL 0x0044 +#define D1W_CNTRL 0x0048 +#define D2W_CNTRL 0x004C +#define D3W_CNTRL 0x0050 +#define DFTMODE_CNTRL 0x0054 + +/* DSI PPI Layer Registers */ +#define PPI_STARTPPI 0x0104 +#define PPI_BUSYPPI 0x0108 +#define PPI_LINEINITCNT 0x0110 +#define PPI_LPTXTIMECNT 0x0114 +#define PPI_LANEENABLE 0x0134 +#define PPI_TX_RX_TA 0x013C +#define PPI_CLS_ATMR 0x0140 +#define PPI_D0S_ATMR 0x0144 +#define PPI_D1S_ATMR 0x0148 +#define PPI_D2S_ATMR 0x014C +#define PPI_D3S_ATMR 0x0150 +#define PPI_D0S_CLRSIPOCOUNT 0x0164 +#define PPI_D1S_CLRSIPOCOUNT 0x0168 +#define PPI_D2S_CLRSIPOCOUNT 0x016C +#define PPI_D3S_CLRSIPOCOUNT 0x0170 +#define CLS_PRE 0x0180 +#define D0S_PRE 0x0184 +#define D1S_PRE 0x0188 +#define D2S_PRE 0x018C +#define D3S_PRE 0x0190 +#define CLS_PREP 0x01A0 +#define D0S_PREP 0x01A4 +#define D1S_PREP 0x01A8 +#define D2S_PREP 0x01AC +#define D3S_PREP 0x01B0 +#define CLS_ZERO 0x01C0 +#define D0S_ZERO 0x01C4 +#define D1S_ZERO 0x01C8 +#define D2S_ZERO 0x01CC +#define D3S_ZERO 0x01D0 +#define PPI_CLRFLG 0x01E0 +#define PPI_CLRSIPO 0x01E4 +#define HSTIMEOUT 0x01F0 +#define HSTIMEOUTENABLE 0x01F4 + +/* DSI Protocol Layer Registers */ +#define DSI_STARTDSI 0x0204 +#define DSI_BUSYDSI 0x0208 +#define DSI_LANEENABLE 0x0210 +#define DSI_LANESTATUS0 0x0214 +#define DSI_LANESTATUS1 0x0218 +#define DSI_INTSTATUS 0x0220 +#define DSI_INTMASK 0x0224 +#define DSI_INTCLR 0x0228 +#define DSI_LPTXTO 0x0230 + +/* DSI General Registers */ +#define DSIERRCNT 0x0300 + +/* DSI Application Layer Registers */ +#define APLCTRL 0x0400 +#define RDPKTLN 0x0404 + +/* Video Path Registers */ +#define VPCTRL 0x0450 +#define HTIM1 0x0454 +#define HTIM2 0x0458 +#define VTIM1 0x045C +#define VTIM2 0x0460 +#define VFUEN 0x0464 + +/* LVDS Registers */ +#define LVMX0003 0x0480 +#define LVMX0407 0x0484 +#define LVMX0811 0x0488 +#define LVMX1215 0x048C +#define LVMX1619 0x0490 +#define LVMX2023 0x0494 +#define LVMX2427 0x0498 +#define LVCFG 0x049C +#define LVPHY0 0x04A0 +#define LVPHY1 0x04A4 + +/* System Registers */ +#define SYSSTAT 0x0500 +#define SYSRST 0x0504 + +/* GPIO Registers */ +/*#define GPIOC 0x0520*/ +#define GPIOO 0x0524 +#define GPIOI 0x0528 + +/* I2C Registers */ +#define I2CTIMCTRL 0x0540 +#define I2CMADDR 0x0544 +#define WDATAQ 0x0548 +#define RDATAQ 0x054C + +/* Chip/Rev Registers */ +#define IDREG 0x0580 + +/* Debug Registers */ +#define DEBUG00 0x05A0 +#define DEBUG01 0x05A4 + +/* Panel CABC registers */ +#define PANEL_PWM_CONTROL 0x90 +#define PANEL_FREQ_DIVIDER_HI 0x91 +#define PANEL_FREQ_DIVIDER_LO 0x92 +#define PANEL_DUTY_CONTROL 0x93 +#define PANEL_MODIFY_RGB 0x94 +#define PANEL_FRAMERATE_CONTROL 0x96 +#define PANEL_PWM_MIN 0x97 +#define PANEL_PWM_REF 0x98 +#define PANEL_PWM_MAX 0x99 +#define PANEL_ALLOW_DISTORT 0x9A +#define PANEL_BYPASS_PWMI 0x9B + +/* Panel color management registers */ +#define PANEL_CM_ENABLE 0x700 +#define PANEL_CM_HUE 0x701 +#define PANEL_CM_SATURATION 0x702 +#define PANEL_CM_INTENSITY 0x703 +#define PANEL_CM_BRIGHTNESS 0x704 +#define PANEL_CM_CE_ENABLE 0x705 +#define PANEL_CM_PEAK_EN 0x710 +#define PANEL_CM_GAIN 0x711 +#define PANEL_CM_HUETABLE_START 0x730 +#define PANEL_CM_HUETABLE_END 0x747 /* inclusive */ + +/* Input muxing for registers LVMX0003...LVMX2427 */ +enum { + INPUT_R0, /* 0 */ + INPUT_R1, + INPUT_R2, + INPUT_R3, + INPUT_R4, + INPUT_R5, + INPUT_R6, + INPUT_R7, + INPUT_G0, /* 8 */ + INPUT_G1, + INPUT_G2, + INPUT_G3, + INPUT_G4, + INPUT_G5, + INPUT_G6, + INPUT_G7, + INPUT_B0, /* 16 */ + INPUT_B1, + INPUT_B2, + INPUT_B3, + INPUT_B4, + INPUT_B5, + INPUT_B6, + INPUT_B7, + INPUT_HSYNC, /* 24 */ + INPUT_VSYNC, + INPUT_DE, + LOGIC_0, + /* 28...31 undefined */ +}; + +#define INPUT_MUX(lvmx03, lvmx02, lvmx01, lvmx00) \ + (FLD_VAL(lvmx03, 29, 24) | FLD_VAL(lvmx02, 20, 16) | \ + FLD_VAL(lvmx01, 12, 8) | FLD_VAL(lvmx00, 4, 0)) + +/** + * tc35876x_regw - Write DSI-LVDS bridge register using I2C + * @client: struct i2c_client to use + * @reg: register address + * @value: value to write + * + * Returns 0 on success, or a negative error value. + */ +static int tc35876x_regw(struct i2c_client *client, u16 reg, u32 value) +{ + int r; + u8 tx_data[] = { + /* NOTE: Register address big-endian, data little-endian. */ + (reg >> 8) & 0xff, + reg & 0xff, + value & 0xff, + (value >> 8) & 0xff, + (value >> 16) & 0xff, + (value >> 24) & 0xff, + }; + struct i2c_msg msgs[] = { + { + .addr = client->addr, + .flags = 0, + .buf = tx_data, + .len = ARRAY_SIZE(tx_data), + }, + }; + + r = i2c_transfer(client->adapter, msgs, ARRAY_SIZE(msgs)); + if (r < 0) { + dev_err(&client->dev, "%s: reg 0x%04x val 0x%08x error %d\n", + __func__, reg, value, r); + return r; + } + + if (r < ARRAY_SIZE(msgs)) { + dev_err(&client->dev, "%s: reg 0x%04x val 0x%08x msgs %d\n", + __func__, reg, value, r); + return -EAGAIN; + } + + dev_dbg(&client->dev, "%s: reg 0x%04x val 0x%08x\n", + __func__, reg, value); + + return 0; +} + +/** + * tc35876x_regr - Read DSI-LVDS bridge register using I2C + * @client: struct i2c_client to use + * @reg: register address + * @value: pointer for storing the value + * + * Returns 0 on success, or a negative error value. + */ +static int tc35876x_regr(struct i2c_client *client, u16 reg, u32 *value) +{ + int r; + u8 tx_data[] = { + (reg >> 8) & 0xff, + reg & 0xff, + }; + u8 rx_data[4]; + struct i2c_msg msgs[] = { + { + .addr = client->addr, + .flags = 0, + .buf = tx_data, + .len = ARRAY_SIZE(tx_data), + }, + { + .addr = client->addr, + .flags = I2C_M_RD, + .buf = rx_data, + .len = ARRAY_SIZE(rx_data), + }, + }; + + r = i2c_transfer(client->adapter, msgs, ARRAY_SIZE(msgs)); + if (r < 0) { + dev_err(&client->dev, "%s: reg 0x%04x error %d\n", __func__, + reg, r); + return r; + } + + if (r < ARRAY_SIZE(msgs)) { + dev_err(&client->dev, "%s: reg 0x%04x msgs %d\n", __func__, + reg, r); + return -EAGAIN; + } + + *value = rx_data[0] << 24 | rx_data[1] << 16 | + rx_data[2] << 8 | rx_data[3]; + + dev_dbg(&client->dev, "%s: reg 0x%04x value 0x%08x\n", __func__, + reg, *value); + + return 0; +} + +void tc35876x_set_bridge_reset_state(struct drm_device *dev, int state) +{ + struct tc35876x_platform_data *pdata; + + if (WARN(!tc35876x_client, "%s called before probe", __func__)) + return; + + dev_dbg(&tc35876x_client->dev, "%s: state %d\n", __func__, state); + + pdata = dev_get_platdata(&tc35876x_client->dev); + + if (pdata->gpio_bridge_reset == -1) + return; + + if (state) { + gpio_set_value_cansleep(pdata->gpio_bridge_reset, 0); + mdelay(10); + } else { + /* Pull MIPI Bridge reset pin to Low */ + gpio_set_value_cansleep(pdata->gpio_bridge_reset, 0); + mdelay(20); + /* Pull MIPI Bridge reset pin to High */ + gpio_set_value_cansleep(pdata->gpio_bridge_reset, 1); + mdelay(40); + } +} + +void tc35876x_configure_lvds_bridge(struct drm_device *dev) +{ + struct i2c_client *i2c = tc35876x_client; + u32 ppi_lptxtimecnt; + u32 txtagocnt; + u32 txtasurecnt; + u32 id; + + if (WARN(!tc35876x_client, "%s called before probe", __func__)) + return; + + dev_dbg(&tc35876x_client->dev, "%s\n", __func__); + + if (!tc35876x_regr(i2c, IDREG, &id)) + dev_info(&tc35876x_client->dev, "tc35876x ID 0x%08x\n", id); + else + dev_err(&tc35876x_client->dev, "Cannot read ID\n"); + + ppi_lptxtimecnt = 4; + txtagocnt = (5 * ppi_lptxtimecnt - 3) / 4; + txtasurecnt = 3 * ppi_lptxtimecnt / 2; + tc35876x_regw(i2c, PPI_TX_RX_TA, FLD_VAL(txtagocnt, 26, 16) | + FLD_VAL(txtasurecnt, 10, 0)); + tc35876x_regw(i2c, PPI_LPTXTIMECNT, FLD_VAL(ppi_lptxtimecnt, 10, 0)); + + tc35876x_regw(i2c, PPI_D0S_CLRSIPOCOUNT, FLD_VAL(1, 5, 0)); + tc35876x_regw(i2c, PPI_D1S_CLRSIPOCOUNT, FLD_VAL(1, 5, 0)); + tc35876x_regw(i2c, PPI_D2S_CLRSIPOCOUNT, FLD_VAL(1, 5, 0)); + tc35876x_regw(i2c, PPI_D3S_CLRSIPOCOUNT, FLD_VAL(1, 5, 0)); + + /* Enabling MIPI & PPI lanes, Enable 4 lanes */ + tc35876x_regw(i2c, PPI_LANEENABLE, + BIT(4) | BIT(3) | BIT(2) | BIT(1) | BIT(0)); + tc35876x_regw(i2c, DSI_LANEENABLE, + BIT(4) | BIT(3) | BIT(2) | BIT(1) | BIT(0)); + tc35876x_regw(i2c, PPI_STARTPPI, BIT(0)); + tc35876x_regw(i2c, DSI_STARTDSI, BIT(0)); + + /* Setting LVDS output frequency */ + tc35876x_regw(i2c, LVPHY0, FLD_VAL(1, 20, 16) | + FLD_VAL(2, 15, 14) | FLD_VAL(6, 4, 0)); /* 0x00048006 */ + + /* Setting video panel control register,0x00000120 VTGen=ON ?!?!? */ + tc35876x_regw(i2c, VPCTRL, BIT(8) | BIT(5)); + + /* Horizontal back porch and horizontal pulse width. 0x00280028 */ + tc35876x_regw(i2c, HTIM1, FLD_VAL(40, 24, 16) | FLD_VAL(40, 8, 0)); + + /* Horizontal front porch and horizontal active video size. 0x00500500*/ + tc35876x_regw(i2c, HTIM2, FLD_VAL(80, 24, 16) | FLD_VAL(1280, 10, 0)); + + /* Vertical back porch and vertical sync pulse width. 0x000e000a */ + tc35876x_regw(i2c, VTIM1, FLD_VAL(14, 23, 16) | FLD_VAL(10, 7, 0)); + + /* Vertical front porch and vertical display size. 0x000e0320 */ + tc35876x_regw(i2c, VTIM2, FLD_VAL(14, 23, 16) | FLD_VAL(800, 10, 0)); + + /* Set above HTIM1, HTIM2, VTIM1, and VTIM2 at next VSYNC. */ + tc35876x_regw(i2c, VFUEN, BIT(0)); + + /* Soft reset LCD controller. */ + tc35876x_regw(i2c, SYSRST, BIT(2)); + + /* LVDS-TX input muxing */ + tc35876x_regw(i2c, LVMX0003, + INPUT_MUX(INPUT_R5, INPUT_R4, INPUT_R3, INPUT_R2)); + tc35876x_regw(i2c, LVMX0407, + INPUT_MUX(INPUT_G2, INPUT_R7, INPUT_R1, INPUT_R6)); + tc35876x_regw(i2c, LVMX0811, + INPUT_MUX(INPUT_G1, INPUT_G0, INPUT_G4, INPUT_G3)); + tc35876x_regw(i2c, LVMX1215, + INPUT_MUX(INPUT_B2, INPUT_G7, INPUT_G6, INPUT_G5)); + tc35876x_regw(i2c, LVMX1619, + INPUT_MUX(INPUT_B4, INPUT_B3, INPUT_B1, INPUT_B0)); + tc35876x_regw(i2c, LVMX2023, + INPUT_MUX(LOGIC_0, INPUT_B7, INPUT_B6, INPUT_B5)); + tc35876x_regw(i2c, LVMX2427, + INPUT_MUX(INPUT_R0, INPUT_DE, INPUT_VSYNC, INPUT_HSYNC)); + + /* Enable LVDS transmitter. */ + tc35876x_regw(i2c, LVCFG, BIT(0)); + + /* Clear notifications. Don't write reserved bits. Was write 0xffffffff + * to 0x0288, must be in error?! */ + tc35876x_regw(i2c, DSI_INTCLR, FLD_MASK(31, 30) | FLD_MASK(22, 0)); +} + +#define GPIOPWMCTRL 0x38F +#define PWM0CLKDIV0 0x62 /* low byte */ +#define PWM0CLKDIV1 0x61 /* high byte */ + +#define SYSTEMCLK 19200000UL /* 19.2 MHz */ +#define PWM_FREQUENCY 9600 /* Hz */ + +/* f = baseclk / (clkdiv + 1) => clkdiv = (baseclk - f) / f */ +static inline u16 calc_clkdiv(unsigned long baseclk, unsigned int f) +{ + return (baseclk - f) / f; +} + +static void tc35876x_brightness_init(struct drm_device *dev) +{ + int ret; + u8 pwmctrl; + u16 clkdiv; + + /* Make sure the PWM reference is the 19.2 MHz system clock. Read first + * instead of setting directly to catch potential conflicts between PWM + * users. */ + ret = intel_scu_ipc_ioread8(GPIOPWMCTRL, &pwmctrl); + if (ret || pwmctrl != 0x01) { + if (ret) + dev_err(&dev->pdev->dev, "GPIOPWMCTRL read failed\n"); + else + dev_warn(&dev->pdev->dev, "GPIOPWMCTRL was not set to system clock (pwmctrl = 0x%02x)\n", pwmctrl); + + ret = intel_scu_ipc_iowrite8(GPIOPWMCTRL, 0x01); + if (ret) + dev_err(&dev->pdev->dev, "GPIOPWMCTRL set failed\n"); + } + + clkdiv = calc_clkdiv(SYSTEMCLK, PWM_FREQUENCY); + + ret = intel_scu_ipc_iowrite8(PWM0CLKDIV1, (clkdiv >> 8) & 0xff); + if (!ret) + ret = intel_scu_ipc_iowrite8(PWM0CLKDIV0, clkdiv & 0xff); + + if (ret) + dev_err(&dev->pdev->dev, "PWM0CLKDIV set failed\n"); + else + dev_dbg(&dev->pdev->dev, "PWM0CLKDIV set to 0x%04x (%d Hz)\n", + clkdiv, PWM_FREQUENCY); +} + +#define PWM0DUTYCYCLE 0x67 + +void tc35876x_brightness_control(struct drm_device *dev, int level) +{ + int ret; + u8 duty_val; + u8 panel_duty_val; + + level = clamp(level, 0, MDFLD_DSI_BRIGHTNESS_MAX_LEVEL); + + /* PWM duty cycle 0x00...0x63 corresponds to 0...99% */ + duty_val = level * 0x63 / MDFLD_DSI_BRIGHTNESS_MAX_LEVEL; + + /* I won't pretend to understand this formula. The panel spec is quite + * bad engrish. + */ + panel_duty_val = (2 * level - 100) * 0xA9 / + MDFLD_DSI_BRIGHTNESS_MAX_LEVEL + 0x56; + + ret = intel_scu_ipc_iowrite8(PWM0DUTYCYCLE, duty_val); + if (ret) + dev_err(&tc35876x_client->dev, "%s: ipc write fail\n", + __func__); + + if (cmi_lcd_i2c_client) { + ret = i2c_smbus_write_byte_data(cmi_lcd_i2c_client, + PANEL_PWM_MAX, panel_duty_val); + if (ret < 0) + dev_err(&cmi_lcd_i2c_client->dev, "%s: i2c write failed\n", + __func__); + } +} + +void tc35876x_toshiba_bridge_panel_off(struct drm_device *dev) +{ + struct tc35876x_platform_data *pdata; + + if (WARN(!tc35876x_client, "%s called before probe", __func__)) + return; + + dev_dbg(&tc35876x_client->dev, "%s\n", __func__); + + pdata = dev_get_platdata(&tc35876x_client->dev); + + if (pdata->gpio_panel_bl_en != -1) + gpio_set_value_cansleep(pdata->gpio_panel_bl_en, 0); + + if (pdata->gpio_panel_vadd != -1) + gpio_set_value_cansleep(pdata->gpio_panel_vadd, 0); +} + +void tc35876x_toshiba_bridge_panel_on(struct drm_device *dev) +{ + struct tc35876x_platform_data *pdata; + struct drm_psb_private *dev_priv = dev->dev_private; + + if (WARN(!tc35876x_client, "%s called before probe", __func__)) + return; + + dev_dbg(&tc35876x_client->dev, "%s\n", __func__); + + pdata = dev_get_platdata(&tc35876x_client->dev); + + if (pdata->gpio_panel_vadd != -1) { + gpio_set_value_cansleep(pdata->gpio_panel_vadd, 1); + msleep(260); + } + + if (cmi_lcd_i2c_client) { + int ret; + dev_dbg(&cmi_lcd_i2c_client->dev, "setting TCON\n"); + /* Bit 4 is average_saving. Setting it to 1, the brightness is + * referenced to the average of the frame content. 0 means + * reference to the maximum of frame contents. Bits 3:0 are + * allow_distort. When set to a nonzero value, all color values + * between 255-allow_distort*2 and 255 are mapped to the + * 255-allow_distort*2 value. + */ + ret = i2c_smbus_write_byte_data(cmi_lcd_i2c_client, + PANEL_ALLOW_DISTORT, 0x10); + if (ret < 0) + dev_err(&cmi_lcd_i2c_client->dev, + "i2c write failed (%d)\n", ret); + ret = i2c_smbus_write_byte_data(cmi_lcd_i2c_client, + PANEL_BYPASS_PWMI, 0); + if (ret < 0) + dev_err(&cmi_lcd_i2c_client->dev, + "i2c write failed (%d)\n", ret); + /* Set minimum brightness value - this is tunable */ + ret = i2c_smbus_write_byte_data(cmi_lcd_i2c_client, + PANEL_PWM_MIN, 0x35); + if (ret < 0) + dev_err(&cmi_lcd_i2c_client->dev, + "i2c write failed (%d)\n", ret); + } + + if (pdata->gpio_panel_bl_en != -1) + gpio_set_value_cansleep(pdata->gpio_panel_bl_en, 1); + + tc35876x_brightness_control(dev, dev_priv->brightness_adjusted); +} + +static struct drm_display_mode *tc35876x_get_config_mode(struct drm_device *dev) +{ + struct drm_display_mode *mode; + + dev_dbg(&dev->pdev->dev, "%s\n", __func__); + + mode = kzalloc(sizeof(*mode), GFP_KERNEL); + if (!mode) + return NULL; + + /* FIXME: do this properly. */ + mode->hdisplay = 1280; + mode->vdisplay = 800; + mode->hsync_start = 1360; + mode->hsync_end = 1400; + mode->htotal = 1440; + mode->vsync_start = 814; + mode->vsync_end = 824; + mode->vtotal = 838; + mode->clock = 33324 << 1; + + dev_info(&dev->pdev->dev, "hdisplay(w) = %d\n", mode->hdisplay); + dev_info(&dev->pdev->dev, "vdisplay(h) = %d\n", mode->vdisplay); + dev_info(&dev->pdev->dev, "HSS = %d\n", mode->hsync_start); + dev_info(&dev->pdev->dev, "HSE = %d\n", mode->hsync_end); + dev_info(&dev->pdev->dev, "htotal = %d\n", mode->htotal); + dev_info(&dev->pdev->dev, "VSS = %d\n", mode->vsync_start); + dev_info(&dev->pdev->dev, "VSE = %d\n", mode->vsync_end); + dev_info(&dev->pdev->dev, "vtotal = %d\n", mode->vtotal); + dev_info(&dev->pdev->dev, "clock = %d\n", mode->clock); + + drm_mode_set_name(mode); + drm_mode_set_crtcinfo(mode, 0); + + mode->type |= DRM_MODE_TYPE_PREFERRED; + + return mode; +} + +/* DV1 Active area 216.96 x 135.6 mm */ +#define DV1_PANEL_WIDTH 217 +#define DV1_PANEL_HEIGHT 136 + +static int tc35876x_get_panel_info(struct drm_device *dev, int pipe, + struct panel_info *pi) +{ + if (!dev || !pi) + return -EINVAL; + + pi->width_mm = DV1_PANEL_WIDTH; + pi->height_mm = DV1_PANEL_HEIGHT; + + return 0; +} + +static int tc35876x_bridge_probe(struct i2c_client *client, + const struct i2c_device_id *id) +{ + struct tc35876x_platform_data *pdata; + + dev_info(&client->dev, "%s\n", __func__); + + if (!i2c_check_functionality(client->adapter, I2C_FUNC_I2C)) { + dev_err(&client->dev, "%s: i2c_check_functionality() failed\n", + __func__); + return -ENODEV; + } + + pdata = dev_get_platdata(&client->dev); + if (!pdata) { + dev_err(&client->dev, "%s: no platform data\n", __func__); + return -ENODEV; + } + + if (pdata->gpio_bridge_reset != -1) { + gpio_request(pdata->gpio_bridge_reset, "tc35876x bridge reset"); + gpio_direction_output(pdata->gpio_bridge_reset, 0); + } + + if (pdata->gpio_panel_bl_en != -1) { + gpio_request(pdata->gpio_panel_bl_en, "tc35876x panel bl en"); + gpio_direction_output(pdata->gpio_panel_bl_en, 0); + } + + if (pdata->gpio_panel_vadd != -1) { + gpio_request(pdata->gpio_panel_vadd, "tc35876x panel vadd"); + gpio_direction_output(pdata->gpio_panel_vadd, 0); + } + + tc35876x_client = client; + + return 0; +} + +static int tc35876x_bridge_remove(struct i2c_client *client) +{ + struct tc35876x_platform_data *pdata = dev_get_platdata(&client->dev); + + dev_dbg(&client->dev, "%s\n", __func__); + + if (pdata->gpio_bridge_reset != -1) + gpio_free(pdata->gpio_bridge_reset); + + if (pdata->gpio_panel_bl_en != -1) + gpio_free(pdata->gpio_panel_bl_en); + + if (pdata->gpio_panel_vadd != -1) + gpio_free(pdata->gpio_panel_vadd); + + tc35876x_client = NULL; + + return 0; +} + +static const struct i2c_device_id tc35876x_bridge_id[] = { + { "i2c_disp_brig", 0 }, + { } +}; +MODULE_DEVICE_TABLE(i2c, tc35876x_bridge_id); + +static struct i2c_driver tc35876x_bridge_i2c_driver = { + .driver = { + .name = "i2c_disp_brig", + }, + .id_table = tc35876x_bridge_id, + .probe = tc35876x_bridge_probe, + .remove = __devexit_p(tc35876x_bridge_remove), +}; + +/* LCD panel I2C */ +static int cmi_lcd_i2c_probe(struct i2c_client *client, + const struct i2c_device_id *id) +{ + dev_info(&client->dev, "%s\n", __func__); + + if (!i2c_check_functionality(client->adapter, I2C_FUNC_I2C)) { + dev_err(&client->dev, "%s: i2c_check_functionality() failed\n", + __func__); + return -ENODEV; + } + + cmi_lcd_i2c_client = client; + + return 0; +} + +static int cmi_lcd_i2c_remove(struct i2c_client *client) +{ + dev_dbg(&client->dev, "%s\n", __func__); + + cmi_lcd_i2c_client = NULL; + + return 0; +} + +static const struct i2c_device_id cmi_lcd_i2c_id[] = { + { "cmi-lcd", 0 }, + { } +}; +MODULE_DEVICE_TABLE(i2c, cmi_lcd_i2c_id); + +static struct i2c_driver cmi_lcd_i2c_driver = { + .driver = { + .name = "cmi-lcd", + }, + .id_table = cmi_lcd_i2c_id, + .probe = cmi_lcd_i2c_probe, + .remove = __devexit_p(cmi_lcd_i2c_remove), +}; + +/* HACK to create I2C device while it's not created by platform code */ +#define CMI_LCD_I2C_ADAPTER 2 +#define CMI_LCD_I2C_ADDR 0x60 + +static int cmi_lcd_hack_create_device(void) +{ + struct i2c_adapter *adapter; + struct i2c_client *client; + struct i2c_board_info info = { + .type = "cmi-lcd", + .addr = CMI_LCD_I2C_ADDR, + }; + + pr_debug("%s\n", __func__); + + adapter = i2c_get_adapter(CMI_LCD_I2C_ADAPTER); + if (!adapter) { + pr_err("%s: i2c_get_adapter(%d) failed\n", __func__, + CMI_LCD_I2C_ADAPTER); + return -EINVAL; + } + + client = i2c_new_device(adapter, &info); + if (!client) { + pr_err("%s: i2c_new_device() failed\n", __func__); + i2c_put_adapter(adapter); + return -EINVAL; + } + + return 0; +} + +static const struct drm_encoder_helper_funcs tc35876x_encoder_helper_funcs = { + .dpms = mdfld_dsi_dpi_dpms, + .mode_fixup = mdfld_dsi_dpi_mode_fixup, + .prepare = mdfld_dsi_dpi_prepare, + .mode_set = mdfld_dsi_dpi_mode_set, + .commit = mdfld_dsi_dpi_commit, +}; + +static const struct drm_encoder_funcs tc35876x_encoder_funcs = { + .destroy = drm_encoder_cleanup, +}; + +const struct panel_funcs mdfld_tc35876x_funcs = { + .encoder_funcs = &tc35876x_encoder_funcs, + .encoder_helper_funcs = &tc35876x_encoder_helper_funcs, + .get_config_mode = tc35876x_get_config_mode, + .get_panel_info = tc35876x_get_panel_info, +}; + +void tc35876x_init(struct drm_device *dev) +{ + int r; + + dev_dbg(&dev->pdev->dev, "%s\n", __func__); + + cmi_lcd_hack_create_device(); + + r = i2c_add_driver(&cmi_lcd_i2c_driver); + if (r < 0) + dev_err(&dev->pdev->dev, + "%s: i2c_add_driver() for %s failed (%d)\n", + __func__, cmi_lcd_i2c_driver.driver.name, r); + + r = i2c_add_driver(&tc35876x_bridge_i2c_driver); + if (r < 0) + dev_err(&dev->pdev->dev, + "%s: i2c_add_driver() for %s failed (%d)\n", + __func__, tc35876x_bridge_i2c_driver.driver.name, r); + + tc35876x_brightness_init(dev); +} + +void tc35876x_exit(void) +{ + pr_debug("%s\n", __func__); + + i2c_del_driver(&tc35876x_bridge_i2c_driver); + + if (cmi_lcd_i2c_client) + i2c_del_driver(&cmi_lcd_i2c_driver); +} diff --git a/drivers/gpu/drm/gma500/tc35876x-dsi-lvds.h b/drivers/gpu/drm/gma500/tc35876x-dsi-lvds.h new file mode 100644 index 0000000..b14b7f9 --- /dev/null +++ b/drivers/gpu/drm/gma500/tc35876x-dsi-lvds.h @@ -0,0 +1,38 @@ +/* + * Copyright © 2011 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + */ + +#ifndef __MDFLD_DSI_LVDS_BRIDGE_H__ +#define __MDFLD_DSI_LVDS_BRIDGE_H__ + +void tc35876x_set_bridge_reset_state(struct drm_device *dev, int state); +void tc35876x_configure_lvds_bridge(struct drm_device *dev); +void tc35876x_brightness_control(struct drm_device *dev, int level); +void tc35876x_toshiba_bridge_panel_off(struct drm_device *dev); +void tc35876x_toshiba_bridge_panel_on(struct drm_device *dev); +void tc35876x_init(struct drm_device *dev); +void tc35876x_exit(void); + +extern const struct panel_funcs mdfld_tc35876x_funcs; + +#endif /*__MDFLD_DSI_LVDS_BRIDGE_H__*/ diff --git a/drivers/gpu/drm/i2c/ch7006_drv.c b/drivers/gpu/drm/i2c/ch7006_drv.c index 07d55df..d3f2e87 100644 --- a/drivers/gpu/drm/i2c/ch7006_drv.c +++ b/drivers/gpu/drm/i2c/ch7006_drv.c @@ -252,10 +252,7 @@ static int ch7006_encoder_create_resources(struct drm_encoder *encoder, drm_mode_create_tv_properties(dev, NUM_TV_NORMS, ch7006_tv_norm_names); - priv->scale_property = drm_property_create(dev, DRM_MODE_PROP_RANGE, - "scale", 2); - priv->scale_property->values[0] = 0; - priv->scale_property->values[1] = 2; + priv->scale_property = drm_property_create_range(dev, 0, "scale", 0, 2); drm_connector_attach_property(connector, conf->tv_select_subconnector_property, priv->select_subconnector); diff --git a/drivers/gpu/drm/i810/i810_dma.c b/drivers/gpu/drm/i810/i810_dma.c index 8f371e8..f920fb5 100644 --- a/drivers/gpu/drm/i810/i810_dma.c +++ b/drivers/gpu/drm/i810/i810_dma.c @@ -99,7 +99,6 @@ static int i810_mmap_buffers(struct file *filp, struct vm_area_struct *vma) buf_priv = buf->dev_private; vma->vm_flags |= (VM_IO | VM_DONTCOPY); - vma->vm_file = filp; buf_priv->currently_mapped = I810_BUF_MAPPED; @@ -130,6 +129,7 @@ static int i810_map_buffer(struct drm_buf *buf, struct drm_file *file_priv) if (buf_priv->currently_mapped == I810_BUF_MAPPED) return -EINVAL; + /* This is all entirely broken */ down_write(¤t->mm->mmap_sem); old_fops = file_priv->filp->f_op; file_priv->filp->f_op = &i810_buffer_fops; @@ -158,11 +158,8 @@ static int i810_unmap_buffer(struct drm_buf *buf) if (buf_priv->currently_mapped != I810_BUF_MAPPED) return -EINVAL; - down_write(¤t->mm->mmap_sem); - retcode = do_munmap(current->mm, - (unsigned long)buf_priv->virtual, + retcode = vm_munmap((unsigned long)buf_priv->virtual, (size_t) buf->total); - up_write(¤t->mm->mmap_sem); buf_priv->currently_mapped = I810_BUF_UNMAPPED; buf_priv->virtual = NULL; @@ -222,8 +219,6 @@ static int i810_dma_cleanup(struct drm_device *dev) pci_free_consistent(dev->pdev, PAGE_SIZE, dev_priv->hw_status_page, dev_priv->dma_status_page); - /* Need to rewrite hardware status page */ - I810_WRITE(0x02080, 0x1ffff000); } kfree(dev->dev_private); dev->dev_private = NULL; @@ -1210,6 +1205,8 @@ int i810_driver_load(struct drm_device *dev, unsigned long flags) dev->types[8] = _DRM_STAT_SECONDARY; dev->types[9] = _DRM_STAT_DMA; + pci_set_master(dev->pdev); + return 0; } diff --git a/drivers/gpu/drm/i810/i810_drv.c b/drivers/gpu/drm/i810/i810_drv.c index d4266bd..ec12f7d 100644 --- a/drivers/gpu/drm/i810/i810_drv.c +++ b/drivers/gpu/drm/i810/i810_drv.c @@ -43,6 +43,17 @@ static struct pci_device_id pciidlist[] = { i810_PCI_IDS }; +static const struct file_operations i810_driver_fops = { + .owner = THIS_MODULE, + .open = drm_open, + .release = drm_release, + .unlocked_ioctl = drm_ioctl, + .mmap = drm_mmap, + .poll = drm_poll, + .fasync = drm_fasync, + .llseek = noop_llseek, +}; + static struct drm_driver driver = { .driver_features = DRIVER_USE_AGP | DRIVER_REQUIRE_AGP | DRIVER_USE_MTRR | @@ -55,17 +66,7 @@ static struct drm_driver driver = { .reclaim_buffers_locked = i810_driver_reclaim_buffers_locked, .dma_quiescent = i810_driver_dma_quiescent, .ioctls = i810_ioctls, - .fops = { - .owner = THIS_MODULE, - .open = drm_open, - .release = drm_release, - .unlocked_ioctl = drm_ioctl, - .mmap = drm_mmap, - .poll = drm_poll, - .fasync = drm_fasync, - .llseek = noop_llseek, - }, - + .fops = &i810_driver_fops, .name = DRIVER_NAME, .desc = DRIVER_DESC, .date = DRIVER_DATE, diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile index 0ae6a7c..ce7fc77 100644 --- a/drivers/gpu/drm/i915/Makefile +++ b/drivers/gpu/drm/i915/Makefile @@ -3,7 +3,7 @@ # Direct Rendering Infrastructure (DRI) in XFree86 4.1.0 and higher. ccflags-y := -Iinclude/drm -i915-y := i915_drv.o i915_dma.o i915_irq.o i915_mem.o \ +i915-y := i915_drv.o i915_dma.o i915_irq.o \ i915_debugfs.o \ i915_suspend.o \ i915_gem.o \ @@ -28,6 +28,7 @@ i915-y := i915_drv.o i915_dma.o i915_irq.o i915_mem.o \ intel_dvo.o \ intel_ringbuffer.o \ intel_overlay.o \ + intel_sprite.o \ intel_opregion.o \ dvo_ch7xxx.o \ dvo_ch7017.o \ diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 10fe480..e6162a1 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -30,7 +30,6 @@ #include #include #include -#include #include "drmP.h" #include "drm.h" #include "intel_drv.h" @@ -84,6 +83,7 @@ static int i915_capabilities(struct seq_file *m, void *data) B(supports_tv); B(has_bsd_ring); B(has_blt_ring); + B(has_llc); #undef B return 0; @@ -122,11 +122,11 @@ static const char *cache_level_str(int type) static void describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj) { - seq_printf(m, "%p: %s%s %8zd %04x %04x %d %d%s%s%s", + seq_printf(m, "%p: %s%s %8zdKiB %04x %04x %d %d%s%s%s", &obj->base, get_pin_flag(obj), get_tiling_flag(obj), - obj->base.size, + obj->base.size / 1024, obj->base.read_domains, obj->base.write_domain, obj->last_rendering_seqno, @@ -340,7 +340,7 @@ static int i915_gem_pageflip_info(struct seq_file *m, void *data) seq_printf(m, "No flip due on pipe %c (plane %c)\n", pipe, plane); } else { - if (atomic_read(&work->pending) < INTEL_FLIP_COMPLETE) { + if (!work->pending) { seq_printf(m, "Flip queued on pipe %c (plane %c)\n", pipe, plane); } else { @@ -351,7 +351,7 @@ static int i915_gem_pageflip_info(struct seq_file *m, void *data) seq_printf(m, "Stall check enabled, "); else seq_printf(m, "Stall check waiting for page flip ioctl, "); - seq_printf(m, "%d prepares\n", atomic_read(&work->pending)); + seq_printf(m, "%d prepares\n", work->pending); if (work->old_fb_obj) { struct drm_i915_gem_object *obj = work->old_fb_obj; @@ -564,45 +564,6 @@ static int i915_hws_info(struct seq_file *m, void *data) return 0; } -static void i915_dump_object(struct seq_file *m, - struct io_mapping *mapping, - struct drm_i915_gem_object *obj) -{ - int page, page_count, i; - - page_count = obj->base.size / PAGE_SIZE; - for (page = 0; page < page_count; page++) { - u32 *mem = io_mapping_map_wc(mapping, - obj->gtt_offset + page * PAGE_SIZE); - for (i = 0; i < PAGE_SIZE; i += 4) - seq_printf(m, "%08x : %08x\n", i, mem[i / 4]); - io_mapping_unmap(mem); - } -} - -static int i915_batchbuffer_info(struct seq_file *m, void *data) -{ - struct drm_info_node *node = (struct drm_info_node *) m->private; - struct drm_device *dev = node->minor->dev; - drm_i915_private_t *dev_priv = dev->dev_private; - struct drm_i915_gem_object *obj; - int ret; - - ret = mutex_lock_interruptible(&dev->struct_mutex); - if (ret) - return ret; - - list_for_each_entry(obj, &dev_priv->mm.active_list, mm_list) { - if (obj->base.read_domains & I915_GEM_DOMAIN_COMMAND) { - seq_printf(m, "--- gtt_offset = 0x%08x\n", obj->gtt_offset); - i915_dump_object(m, dev_priv->mm.gtt_mapping, obj); - } - } - - mutex_unlock(&dev->struct_mutex); - return 0; -} - static int i915_ringbuffer_data(struct seq_file *m, void *data) { struct drm_info_node *node = (struct drm_info_node *) m->private; @@ -654,7 +615,7 @@ static int i915_ringbuffer_info(struct seq_file *m, void *data) seq_printf(m, " Size : %08x\n", ring->size); seq_printf(m, " Active : %08x\n", intel_ring_get_active_head(ring)); seq_printf(m, " NOPID : %08x\n", I915_READ_NOPID(ring)); - if (IS_GEN6(dev)) { + if (IS_GEN6(dev) || IS_GEN7(dev)) { seq_printf(m, " Sync 0 : %08x\n", I915_READ_SYNC_0(ring)); seq_printf(m, " Sync 1 : %08x\n", I915_READ_SYNC_1(ring)); } @@ -669,9 +630,9 @@ static int i915_ringbuffer_info(struct seq_file *m, void *data) static const char *ring_str(int ring) { switch (ring) { - case RING_RENDER: return " render"; - case RING_BSD: return " bsd"; - case RING_BLT: return " blt"; + case RCS: return "render"; + case VCS: return "bsd"; + case BCS: return "blt"; default: return ""; } } @@ -714,7 +675,7 @@ static void print_error_buffers(struct seq_file *m, seq_printf(m, "%s [%d]:\n", name, count); while (count--) { - seq_printf(m, " %08x %8u %04x %04x %08x%s%s%s%s%s%s", + seq_printf(m, " %08x %8u %04x %04x %08x%s%s%s%s%s%s%s", err->gtt_offset, err->size, err->read_domains, @@ -724,6 +685,7 @@ static void print_error_buffers(struct seq_file *m, tiling_flag(err->tiling), dirty_flag(err->dirty), purgeable_flag(err->purgeable), + err->ring != -1 ? " " : "", ring_str(err->ring), cache_level_str(err->cache_level)); @@ -737,6 +699,38 @@ static void print_error_buffers(struct seq_file *m, } } +static void i915_ring_error_state(struct seq_file *m, + struct drm_device *dev, + struct drm_i915_error_state *error, + unsigned ring) +{ + seq_printf(m, "%s command stream:\n", ring_str(ring)); + seq_printf(m, " HEAD: 0x%08x\n", error->head[ring]); + seq_printf(m, " TAIL: 0x%08x\n", error->tail[ring]); + seq_printf(m, " ACTHD: 0x%08x\n", error->acthd[ring]); + seq_printf(m, " IPEIR: 0x%08x\n", error->ipeir[ring]); + seq_printf(m, " IPEHR: 0x%08x\n", error->ipehr[ring]); + seq_printf(m, " INSTDONE: 0x%08x\n", error->instdone[ring]); + if (ring == RCS && INTEL_INFO(dev)->gen >= 4) { + seq_printf(m, " INSTDONE1: 0x%08x\n", error->instdone1); + seq_printf(m, " BBADDR: 0x%08llx\n", error->bbaddr); + } + if (INTEL_INFO(dev)->gen >= 4) + seq_printf(m, " INSTPS: 0x%08x\n", error->instps[ring]); + seq_printf(m, " INSTPM: 0x%08x\n", error->instpm[ring]); + if (INTEL_INFO(dev)->gen >= 6) { + seq_printf(m, " FADDR: 0x%08x\n", error->faddr[ring]); + seq_printf(m, " FAULT_REG: 0x%08x\n", error->fault_reg[ring]); + seq_printf(m, " SYNC_0: 0x%08x\n", + error->semaphore_mboxes[ring][0]); + seq_printf(m, " SYNC_1: 0x%08x\n", + error->semaphore_mboxes[ring][1]); + } + seq_printf(m, " seqno: 0x%08x\n", error->seqno[ring]); + seq_printf(m, " ring->head: 0x%08x\n", error->cpu_ring_head[ring]); + seq_printf(m, " ring->tail: 0x%08x\n", error->cpu_ring_tail[ring]); +} + static int i915_error_state(struct seq_file *m, void *unused) { struct drm_info_node *node = (struct drm_info_node *) m->private; @@ -744,7 +738,7 @@ static int i915_error_state(struct seq_file *m, void *unused) drm_i915_private_t *dev_priv = dev->dev_private; struct drm_i915_error_state *error; unsigned long flags; - int i, page, offset, elt; + int i, j, page, offset, elt; spin_lock_irqsave(&dev_priv->error_lock, flags); if (!dev_priv->first_error) { @@ -756,39 +750,23 @@ static int i915_error_state(struct seq_file *m, void *unused) seq_printf(m, "Time: %ld s %ld us\n", error->time.tv_sec, error->time.tv_usec); - seq_printf(m, "Kernel: " UTS_RELEASE); seq_printf(m, "PCI ID: 0x%04x\n", dev->pci_device); seq_printf(m, "EIR: 0x%08x\n", error->eir); seq_printf(m, "PGTBL_ER: 0x%08x\n", error->pgtbl_er); + + for (i = 0; i < dev_priv->num_fence_regs; i++) + seq_printf(m, " fence[%d] = %08llx\n", i, error->fence[i]); + if (INTEL_INFO(dev)->gen >= 6) { seq_printf(m, "ERROR: 0x%08x\n", error->error); - seq_printf(m, "Blitter command stream:\n"); - seq_printf(m, " ACTHD: 0x%08x\n", error->bcs_acthd); - seq_printf(m, " IPEIR: 0x%08x\n", error->bcs_ipeir); - seq_printf(m, " IPEHR: 0x%08x\n", error->bcs_ipehr); - seq_printf(m, " INSTDONE: 0x%08x\n", error->bcs_instdone); - seq_printf(m, " seqno: 0x%08x\n", error->bcs_seqno); - seq_printf(m, "Video (BSD) command stream:\n"); - seq_printf(m, " ACTHD: 0x%08x\n", error->vcs_acthd); - seq_printf(m, " IPEIR: 0x%08x\n", error->vcs_ipeir); - seq_printf(m, " IPEHR: 0x%08x\n", error->vcs_ipehr); - seq_printf(m, " INSTDONE: 0x%08x\n", error->vcs_instdone); - seq_printf(m, " seqno: 0x%08x\n", error->vcs_seqno); + seq_printf(m, "DONE_REG: 0x%08x\n", error->done_reg); } - seq_printf(m, "Render command stream:\n"); - seq_printf(m, " ACTHD: 0x%08x\n", error->acthd); - seq_printf(m, " IPEIR: 0x%08x\n", error->ipeir); - seq_printf(m, " IPEHR: 0x%08x\n", error->ipehr); - seq_printf(m, " INSTDONE: 0x%08x\n", error->instdone); - if (INTEL_INFO(dev)->gen >= 4) { - seq_printf(m, " INSTDONE1: 0x%08x\n", error->instdone1); - seq_printf(m, " INSTPS: 0x%08x\n", error->instps); - } - seq_printf(m, " INSTPM: 0x%08x\n", error->instpm); - seq_printf(m, " seqno: 0x%08x\n", error->seqno); - for (i = 0; i < dev_priv->num_fence_regs; i++) - seq_printf(m, " fence[%d] = %08llx\n", i, error->fence[i]); + i915_ring_error_state(m, dev, error, RCS); + if (HAS_BLT(dev)) + i915_ring_error_state(m, dev, error, BCS); + if (HAS_BSD(dev)) + i915_ring_error_state(m, dev, error, VCS); if (error->active_bo) print_error_buffers(m, "Active", @@ -800,10 +778,10 @@ static int i915_error_state(struct seq_file *m, void *unused) error->pinned_bo, error->pinned_bo_count); - for (i = 0; i < ARRAY_SIZE(error->batchbuffer); i++) { - if (error->batchbuffer[i]) { - struct drm_i915_error_object *obj = error->batchbuffer[i]; + for (i = 0; i < ARRAY_SIZE(error->ring); i++) { + struct drm_i915_error_object *obj; + if ((obj = error->ring[i].batchbuffer)) { seq_printf(m, "%s --- gtt_offset = 0x%08x\n", dev_priv->ring[i].name, obj->gtt_offset); @@ -815,11 +793,20 @@ static int i915_error_state(struct seq_file *m, void *unused) } } } - } - for (i = 0; i < ARRAY_SIZE(error->ringbuffer); i++) { - if (error->ringbuffer[i]) { - struct drm_i915_error_object *obj = error->ringbuffer[i]; + if (error->ring[i].num_requests) { + seq_printf(m, "%s --- %d requests\n", + dev_priv->ring[i].name, + error->ring[i].num_requests); + for (j = 0; j < error->ring[i].num_requests; j++) { + seq_printf(m, " seqno 0x%08x, emitted %ld, tail 0x%08x\n", + error->ring[i].requests[j].seqno, + error->ring[i].requests[j].jiffies, + error->ring[i].requests[j].tail); + } + } + + if ((obj = error->ring[i].ringbuffer)) { seq_printf(m, "%s --- ringbuffer = 0x%08x\n", dev_priv->ring[i].name, obj->gtt_offset); @@ -1003,7 +990,7 @@ static int i915_inttoext_table(struct seq_file *m, void *unused) return 0; } -static int i915_drpc_info(struct seq_file *m, void *unused) +static int ironlake_drpc_info(struct seq_file *m) { struct drm_info_node *node = (struct drm_info_node *) m->private; struct drm_device *dev = node->minor->dev; @@ -1070,6 +1057,95 @@ static int i915_drpc_info(struct seq_file *m, void *unused) return 0; } +static int gen6_drpc_info(struct seq_file *m) +{ + + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + u32 rpmodectl1, gt_core_status, rcctl1; + unsigned forcewake_count; + int count=0, ret; + + + ret = mutex_lock_interruptible(&dev->struct_mutex); + if (ret) + return ret; + + spin_lock_irq(&dev_priv->gt_lock); + forcewake_count = dev_priv->forcewake_count; + spin_unlock_irq(&dev_priv->gt_lock); + + if (forcewake_count) { + seq_printf(m, "RC information inaccurate because somebody " + "holds a forcewake reference \n"); + } else { + /* NB: we cannot use forcewake, else we read the wrong values */ + while (count++ < 50 && (I915_READ_NOTRACE(FORCEWAKE_ACK) & 1)) + udelay(10); + seq_printf(m, "RC information accurate: %s\n", yesno(count < 51)); + } + + gt_core_status = readl(dev_priv->regs + GEN6_GT_CORE_STATUS); + trace_i915_reg_rw(false, GEN6_GT_CORE_STATUS, gt_core_status, 4); + + rpmodectl1 = I915_READ(GEN6_RP_CONTROL); + rcctl1 = I915_READ(GEN6_RC_CONTROL); + mutex_unlock(&dev->struct_mutex); + + seq_printf(m, "Video Turbo Mode: %s\n", + yesno(rpmodectl1 & GEN6_RP_MEDIA_TURBO)); + seq_printf(m, "HW control enabled: %s\n", + yesno(rpmodectl1 & GEN6_RP_ENABLE)); + seq_printf(m, "SW control enabled: %s\n", + yesno((rpmodectl1 & GEN6_RP_MEDIA_MODE_MASK) == + GEN6_RP_MEDIA_SW_MODE)); + seq_printf(m, "RC1e Enabled: %s\n", + yesno(rcctl1 & GEN6_RC_CTL_RC1e_ENABLE)); + seq_printf(m, "RC6 Enabled: %s\n", + yesno(rcctl1 & GEN6_RC_CTL_RC6_ENABLE)); + seq_printf(m, "Deep RC6 Enabled: %s\n", + yesno(rcctl1 & GEN6_RC_CTL_RC6p_ENABLE)); + seq_printf(m, "Deepest RC6 Enabled: %s\n", + yesno(rcctl1 & GEN6_RC_CTL_RC6pp_ENABLE)); + seq_printf(m, "Current RC state: "); + switch (gt_core_status & GEN6_RCn_MASK) { + case GEN6_RC0: + if (gt_core_status & GEN6_CORE_CPD_STATE_MASK) + seq_printf(m, "Core Power Down\n"); + else + seq_printf(m, "on\n"); + break; + case GEN6_RC3: + seq_printf(m, "RC3\n"); + break; + case GEN6_RC6: + seq_printf(m, "RC6\n"); + break; + case GEN6_RC7: + seq_printf(m, "RC7\n"); + break; + default: + seq_printf(m, "Unknown\n"); + break; + } + + seq_printf(m, "Core Power Down: %s\n", + yesno(gt_core_status & GEN6_CORE_CPD_STATE_MASK)); + return 0; +} + +static int i915_drpc_info(struct seq_file *m, void *unused) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + + if (IS_GEN6(dev) || IS_GEN7(dev)) + return gen6_drpc_info(m); + else + return ironlake_drpc_info(m); +} + static int i915_fbc_status(struct seq_file *m, void *unused) { struct drm_info_node *node = (struct drm_info_node *) m->private; @@ -1148,6 +1224,9 @@ static int i915_emon_status(struct seq_file *m, void *unused) unsigned long temp, chipset, gfx; int ret; + if (!IS_GEN5(dev)) + return -ENODEV; + ret = mutex_lock_interruptible(&dev->struct_mutex); if (ret) return ret; @@ -1327,11 +1406,102 @@ static int i915_gen6_forcewake_count_info(struct seq_file *m, void *data) return 0; } -static int -i915_wedged_open(struct inode *inode, - struct file *filp) +static const char *swizzle_string(unsigned swizzle) { - filp->private_data = inode->i_private; + switch(swizzle) { + case I915_BIT_6_SWIZZLE_NONE: + return "none"; + case I915_BIT_6_SWIZZLE_9: + return "bit9"; + case I915_BIT_6_SWIZZLE_9_10: + return "bit9/bit10"; + case I915_BIT_6_SWIZZLE_9_11: + return "bit9/bit11"; + case I915_BIT_6_SWIZZLE_9_10_11: + return "bit9/bit10/bit11"; + case I915_BIT_6_SWIZZLE_9_17: + return "bit9/bit17"; + case I915_BIT_6_SWIZZLE_9_10_17: + return "bit9/bit10/bit17"; + case I915_BIT_6_SWIZZLE_UNKNOWN: + return "unkown"; + } + + return "bug"; +} + +static int i915_swizzle_info(struct seq_file *m, void *data) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + + mutex_lock(&dev->struct_mutex); + seq_printf(m, "bit6 swizzle for X-tiling = %s\n", + swizzle_string(dev_priv->mm.bit_6_swizzle_x)); + seq_printf(m, "bit6 swizzle for Y-tiling = %s\n", + swizzle_string(dev_priv->mm.bit_6_swizzle_y)); + + if (IS_GEN3(dev) || IS_GEN4(dev)) { + seq_printf(m, "DDC = 0x%08x\n", + I915_READ(DCC)); + seq_printf(m, "C0DRB3 = 0x%04x\n", + I915_READ16(C0DRB3)); + seq_printf(m, "C1DRB3 = 0x%04x\n", + I915_READ16(C1DRB3)); + } else if (IS_GEN6(dev) || IS_GEN7(dev)) { + seq_printf(m, "MAD_DIMM_C0 = 0x%08x\n", + I915_READ(MAD_DIMM_C0)); + seq_printf(m, "MAD_DIMM_C1 = 0x%08x\n", + I915_READ(MAD_DIMM_C1)); + seq_printf(m, "MAD_DIMM_C2 = 0x%08x\n", + I915_READ(MAD_DIMM_C2)); + seq_printf(m, "TILECTL = 0x%08x\n", + I915_READ(TILECTL)); + seq_printf(m, "ARB_MODE = 0x%08x\n", + I915_READ(ARB_MODE)); + seq_printf(m, "DISP_ARB_CTL = 0x%08x\n", + I915_READ(DISP_ARB_CTL)); + } + mutex_unlock(&dev->struct_mutex); + + return 0; +} + +static int i915_ppgtt_info(struct seq_file *m, void *data) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_ring_buffer *ring; + int i, ret; + + + ret = mutex_lock_interruptible(&dev->struct_mutex); + if (ret) + return ret; + if (INTEL_INFO(dev)->gen == 6) + seq_printf(m, "GFX_MODE: 0x%08x\n", I915_READ(GFX_MODE)); + + for (i = 0; i < I915_NUM_RINGS; i++) { + ring = &dev_priv->ring[i]; + + seq_printf(m, "%s\n", ring->name); + if (INTEL_INFO(dev)->gen == 7) + seq_printf(m, "GFX_MODE: 0x%08x\n", I915_READ(RING_MODE_GEN7(ring))); + seq_printf(m, "PP_DIR_BASE: 0x%08x\n", I915_READ(RING_PP_DIR_BASE(ring))); + seq_printf(m, "PP_DIR_BASE_READ: 0x%08x\n", I915_READ(RING_PP_DIR_BASE_READ(ring))); + seq_printf(m, "PP_DIR_DCLV: 0x%08x\n", I915_READ(RING_PP_DIR_DCLV(ring))); + } + if (dev_priv->mm.aliasing_ppgtt) { + struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt; + + seq_printf(m, "aliasing PPGTT:\n"); + seq_printf(m, "pd gtt offset: 0x%08x\n", ppgtt->pd_offset); + } + seq_printf(m, "ECOCHK: 0x%08x\n", I915_READ(GAM_ECOCHK)); + mutex_unlock(&dev->struct_mutex); + return 0; } @@ -1385,20 +1555,12 @@ i915_wedged_write(struct file *filp, static const struct file_operations i915_wedged_fops = { .owner = THIS_MODULE, - .open = i915_wedged_open, + .open = simple_open, .read = i915_wedged_read, .write = i915_wedged_write, .llseek = default_llseek, }; -static int -i915_max_freq_open(struct inode *inode, - struct file *filp) -{ - filp->private_data = inode->i_private; - return 0; -} - static ssize_t i915_max_freq_read(struct file *filp, char __user *ubuf, @@ -1455,20 +1617,12 @@ i915_max_freq_write(struct file *filp, static const struct file_operations i915_max_freq_fops = { .owner = THIS_MODULE, - .open = i915_max_freq_open, + .open = simple_open, .read = i915_max_freq_read, .write = i915_max_freq_write, .llseek = default_llseek, }; -static int -i915_cache_sharing_open(struct inode *inode, - struct file *filp) -{ - filp->private_data = inode->i_private; - return 0; -} - static ssize_t i915_cache_sharing_read(struct file *filp, char __user *ubuf, @@ -1534,7 +1688,7 @@ i915_cache_sharing_write(struct file *filp, static const struct file_operations i915_cache_sharing_fops = { .owner = THIS_MODULE, - .open = i915_cache_sharing_open, + .open = simple_open, .read = i915_cache_sharing_read, .write = i915_cache_sharing_write, .llseek = default_llseek, @@ -1566,28 +1720,13 @@ drm_add_fake_info_node(struct drm_minor *minor, return 0; } -static int i915_wedged_create(struct dentry *root, struct drm_minor *minor) -{ - struct drm_device *dev = minor->dev; - struct dentry *ent; - - ent = debugfs_create_file("i915_wedged", - S_IRUGO | S_IWUSR, - root, dev, - &i915_wedged_fops); - if (IS_ERR(ent)) - return PTR_ERR(ent); - - return drm_add_fake_info_node(minor, ent, &i915_wedged_fops); -} - static int i915_forcewake_open(struct inode *inode, struct file *file) { struct drm_device *dev = inode->i_private; struct drm_i915_private *dev_priv = dev->dev_private; int ret; - if (!IS_GEN6(dev)) + if (INTEL_INFO(dev)->gen < 6) return 0; ret = mutex_lock_interruptible(&dev->struct_mutex); @@ -1604,7 +1743,7 @@ int i915_forcewake_release(struct inode *inode, struct file *file) struct drm_device *dev = inode->i_private; struct drm_i915_private *dev_priv = dev->dev_private; - if (!IS_GEN6(dev)) + if (INTEL_INFO(dev)->gen < 6) return 0; /* @@ -1642,34 +1781,22 @@ static int i915_forcewake_create(struct dentry *root, struct drm_minor *minor) return drm_add_fake_info_node(minor, ent, &i915_forcewake_fops); } -static int i915_max_freq_create(struct dentry *root, struct drm_minor *minor) -{ - struct drm_device *dev = minor->dev; - struct dentry *ent; - - ent = debugfs_create_file("i915_max_freq", - S_IRUGO | S_IWUSR, - root, dev, - &i915_max_freq_fops); - if (IS_ERR(ent)) - return PTR_ERR(ent); - - return drm_add_fake_info_node(minor, ent, &i915_max_freq_fops); -} - -static int i915_cache_sharing_create(struct dentry *root, struct drm_minor *minor) +static int i915_debugfs_create(struct dentry *root, + struct drm_minor *minor, + const char *name, + const struct file_operations *fops) { struct drm_device *dev = minor->dev; struct dentry *ent; - ent = debugfs_create_file("i915_cache_sharing", + ent = debugfs_create_file(name, S_IRUGO | S_IWUSR, root, dev, - &i915_cache_sharing_fops); + fops); if (IS_ERR(ent)) return PTR_ERR(ent); - return drm_add_fake_info_node(minor, ent, &i915_cache_sharing_fops); + return drm_add_fake_info_node(minor, ent, fops); } static struct drm_info_list i915_debugfs_list[] = { @@ -1695,7 +1822,6 @@ static struct drm_info_list i915_debugfs_list[] = { {"i915_bsd_ringbuffer_info", i915_ringbuffer_info, 0, (void *)VCS}, {"i915_blt_ringbuffer_data", i915_ringbuffer_data, 0, (void *)BCS}, {"i915_blt_ringbuffer_info", i915_ringbuffer_info, 0, (void *)BCS}, - {"i915_batchbuffers", i915_batchbuffer_info, 0}, {"i915_error_state", i915_error_state, 0}, {"i915_rstdby_delays", i915_rstdby_delays, 0}, {"i915_cur_delayinfo", i915_cur_delayinfo, 0}, @@ -1711,6 +1837,8 @@ static struct drm_info_list i915_debugfs_list[] = { {"i915_gem_framebuffer", i915_gem_framebuffer_info, 0}, {"i915_context_status", i915_context_status, 0}, {"i915_gen6_forcewake_count", i915_gen6_forcewake_count_info, 0}, + {"i915_swizzle_info", i915_swizzle_info, 0}, + {"i915_ppgtt_info", i915_ppgtt_info, 0}, }; #define I915_DEBUGFS_ENTRIES ARRAY_SIZE(i915_debugfs_list) @@ -1718,17 +1846,25 @@ int i915_debugfs_init(struct drm_minor *minor) { int ret; - ret = i915_wedged_create(minor->debugfs_root, minor); + ret = i915_debugfs_create(minor->debugfs_root, minor, + "i915_wedged", + &i915_wedged_fops); if (ret) return ret; ret = i915_forcewake_create(minor->debugfs_root, minor); if (ret) return ret; - ret = i915_max_freq_create(minor->debugfs_root, minor); + + ret = i915_debugfs_create(minor->debugfs_root, minor, + "i915_max_freq", + &i915_max_freq_fops); if (ret) return ret; - ret = i915_cache_sharing_create(minor->debugfs_root, minor); + + ret = i915_debugfs_create(minor->debugfs_root, minor, + "i915_cache_sharing", + &i915_cache_sharing_fops); if (ret) return ret; diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c index ca67338..ba60f3c 100644 --- a/drivers/gpu/drm/i915/i915_dma.c +++ b/drivers/gpu/drm/i915/i915_dma.c @@ -781,6 +781,12 @@ static int i915_getparam(struct drm_device *dev, void *data, case I915_PARAM_HAS_RELAXED_DELTA: value = 1; break; + case I915_PARAM_HAS_GEN7_SOL_RESET: + value = 1; + break; + case I915_PARAM_HAS_LLC: + value = HAS_LLC(dev); + break; default: DRM_DEBUG_DRIVER("Unknown parameter %d\n", param->param); @@ -1177,6 +1183,21 @@ static bool i915_switcheroo_can_switch(struct pci_dev *pdev) return can_switch; } +static bool +intel_enable_ppgtt(struct drm_device *dev) +{ + if (i915_enable_ppgtt >= 0) + return i915_enable_ppgtt; + +#ifdef CONFIG_INTEL_IOMMU + /* Disable ppgtt on SNB if VT-d is on. */ + if (INTEL_INFO(dev)->gen == 6 && intel_iommu_gfx_mapped) + return false; +#endif + + return true; +} + static int i915_load_gem_init(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; @@ -1190,22 +1211,41 @@ static int i915_load_gem_init(struct drm_device *dev) /* Basic memrange allocator for stolen space */ drm_mm_init(&dev_priv->mm.stolen, 0, prealloc_size); - /* Let GEM Manage all of the aperture. - * - * However, leave one page at the end still bound to the scratch page. - * There are a number of places where the hardware apparently - * prefetches past the end of the object, and we've seen multiple - * hangs with the GPU head pointer stuck in a batchbuffer bound - * at the last page of the aperture. One page should be enough to - * keep any prefetching inside of the aperture. - */ - i915_gem_do_init(dev, 0, mappable_size, gtt_size - PAGE_SIZE); - mutex_lock(&dev->struct_mutex); - ret = i915_gem_init_ringbuffer(dev); + if (intel_enable_ppgtt(dev) && HAS_ALIASING_PPGTT(dev)) { + /* PPGTT pdes are stolen from global gtt ptes, so shrink the + * aperture accordingly when using aliasing ppgtt. */ + gtt_size -= I915_PPGTT_PD_ENTRIES*PAGE_SIZE; + /* For paranoia keep the guard page in between. */ + gtt_size -= PAGE_SIZE; + + i915_gem_do_init(dev, 0, mappable_size, gtt_size); + + ret = i915_gem_init_aliasing_ppgtt(dev); + if (ret) { + mutex_unlock(&dev->struct_mutex); + return ret; + } + } else { + /* Let GEM Manage all of the aperture. + * + * However, leave one page at the end still bound to the scratch + * page. There are a number of places where the hardware + * apparently prefetches past the end of the object, and we've + * seen multiple hangs with the GPU head pointer stuck in a + * batchbuffer bound at the last page of the aperture. One page + * should be enough to keep any prefetching inside of the + * aperture. + */ + i915_gem_do_init(dev, 0, mappable_size, gtt_size - PAGE_SIZE); + } + + ret = i915_gem_init_hw(dev); mutex_unlock(&dev->struct_mutex); - if (ret) + if (ret) { + i915_gem_cleanup_aliasing_ppgtt(dev); return ret; + } /* Try to set up FBC with a reasonable compressed buffer size */ if (I915_HAS_FBC(dev) && i915_powersave) { @@ -1292,6 +1332,7 @@ cleanup_gem: mutex_lock(&dev->struct_mutex); i915_gem_cleanup_ringbuffer(dev); mutex_unlock(&dev->struct_mutex); + i915_gem_cleanup_aliasing_ppgtt(dev); cleanup_vga_switcheroo: vga_switcheroo_unregister_client(dev->pdev); cleanup_vga_client: @@ -1660,6 +1701,9 @@ void i915_update_gfx_val(struct drm_i915_private *dev_priv) unsigned long diffms; u32 count; + if (dev_priv->info->gen != 5) + return; + getrawmonotonic(&now); diff1 = timespec_sub(now, dev_priv->last_time2); @@ -1890,27 +1934,6 @@ ips_ping_for_i915_load(void) } } -static void i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv) -{ - struct apertures_struct *ap; - struct pci_dev *pdev = dev_priv->dev->pdev; - bool primary; - - ap = alloc_apertures(1); - if (!ap) - return; - - ap->ranges[0].base = dev_priv->dev->agp->base; - ap->ranges[0].size = - dev_priv->mm.gtt->gtt_mappable_entries << PAGE_SHIFT; - primary = - pdev->resource[PCI_ROM_RESOURCE].flags & IORESOURCE_ROM_SHADOW; - - remove_conflicting_framebuffers(ap, "inteldrmfb", primary); - - kfree(ap); -} - /** * i915_driver_load - setup chip and create an initial config * @dev: DRM device @@ -1948,14 +1971,7 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags) goto free_priv; } - dev_priv->mm.gtt = intel_gtt_get(); - if (!dev_priv->mm.gtt) { - DRM_ERROR("Failed to initialize GTT\n"); - ret = -ENODEV; - goto put_bridge; - } - - i915_kick_out_firmware_fb(dev_priv); + pci_set_master(dev->pdev); /* overlay on gen2 is broken and can't address above 1G */ if (IS_GEN2(dev)) @@ -1980,6 +1996,13 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags) goto put_bridge; } + dev_priv->mm.gtt = intel_gtt_get(); + if (!dev_priv->mm.gtt) { + DRM_ERROR("Failed to initialize GTT\n"); + ret = -ENODEV; + goto out_rmmap; + } + agp_size = dev_priv->mm.gtt->gtt_mappable_entries << PAGE_SHIFT; dev_priv->mm.gtt_mapping = @@ -2101,12 +2124,14 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags) setup_timer(&dev_priv->hangcheck_timer, i915_hangcheck_elapsed, (unsigned long) dev); - spin_lock(&mchdev_lock); - i915_mch_dev = dev_priv; - dev_priv->mchdev_lock = &mchdev_lock; - spin_unlock(&mchdev_lock); + if (IS_GEN5(dev)) { + spin_lock(&mchdev_lock); + i915_mch_dev = dev_priv; + dev_priv->mchdev_lock = &mchdev_lock; + spin_unlock(&mchdev_lock); - ips_ping_for_i915_load(); + ips_ping_for_i915_load(); + } return 0; @@ -2149,7 +2174,7 @@ int i915_driver_unload(struct drm_device *dev) unregister_shrinker(&dev_priv->mm.inactive_shrinker); mutex_lock(&dev->struct_mutex); - ret = i915_gpu_idle(dev); + ret = i915_gpu_idle(dev, true); if (ret) DRM_ERROR("failed to idle hardware: %d\n", ret); mutex_unlock(&dev->struct_mutex); @@ -2202,6 +2227,7 @@ int i915_driver_unload(struct drm_device *dev) i915_gem_free_all_phys_object(dev); i915_gem_cleanup_ringbuffer(dev); mutex_unlock(&dev->struct_mutex); + i915_gem_cleanup_aliasing_ppgtt(dev); if (I915_HAS_FBC(dev) && i915_powersave) i915_cleanup_compression(dev); drm_mm_takedown(&dev_priv->mm.stolen); @@ -2267,18 +2293,12 @@ void i915_driver_lastclose(struct drm_device * dev) i915_gem_lastclose(dev); - if (dev_priv->agp_heap) - i915_mem_takedown(&(dev_priv->agp_heap)); - i915_dma_cleanup(dev); } void i915_driver_preclose(struct drm_device * dev, struct drm_file *file_priv) { - drm_i915_private_t *dev_priv = dev->dev_private; i915_gem_release(dev, file_priv); - if (!drm_core_check_feature(dev, DRIVER_MODESET)) - i915_mem_release(dev, file_priv, dev_priv->agp_heap); } void i915_driver_postclose(struct drm_device *dev, struct drm_file *file) @@ -2297,11 +2317,11 @@ struct drm_ioctl_desc i915_ioctls[] = { DRM_IOCTL_DEF_DRV(I915_IRQ_WAIT, i915_irq_wait, DRM_AUTH), DRM_IOCTL_DEF_DRV(I915_GETPARAM, i915_getparam, DRM_AUTH), DRM_IOCTL_DEF_DRV(I915_SETPARAM, i915_setparam, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY), - DRM_IOCTL_DEF_DRV(I915_ALLOC, i915_mem_alloc, DRM_AUTH), - DRM_IOCTL_DEF_DRV(I915_FREE, i915_mem_free, DRM_AUTH), - DRM_IOCTL_DEF_DRV(I915_INIT_HEAP, i915_mem_init_heap, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY), + DRM_IOCTL_DEF_DRV(I915_ALLOC, drm_noop, DRM_AUTH), + DRM_IOCTL_DEF_DRV(I915_FREE, drm_noop, DRM_AUTH), + DRM_IOCTL_DEF_DRV(I915_INIT_HEAP, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY), DRM_IOCTL_DEF_DRV(I915_CMDBUFFER, i915_cmdbuffer, DRM_AUTH), - DRM_IOCTL_DEF_DRV(I915_DESTROY_HEAP, i915_mem_destroy_heap, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY), + DRM_IOCTL_DEF_DRV(I915_DESTROY_HEAP, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY), DRM_IOCTL_DEF_DRV(I915_SET_VBLANK_PIPE, i915_vblank_pipe_set, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY), DRM_IOCTL_DEF_DRV(I915_GET_VBLANK_PIPE, i915_vblank_pipe_get, DRM_AUTH), DRM_IOCTL_DEF_DRV(I915_VBLANK_SWAP, i915_vblank_swap, DRM_AUTH), @@ -2329,6 +2349,8 @@ struct drm_ioctl_desc i915_ioctls[] = { DRM_IOCTL_DEF_DRV(I915_GEM_MADVISE, i915_gem_madvise_ioctl, DRM_UNLOCKED), DRM_IOCTL_DEF_DRV(I915_OVERLAY_PUT_IMAGE, intel_overlay_put_image, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), DRM_IOCTL_DEF_DRV(I915_OVERLAY_ATTRS, intel_overlay_attrs, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF_DRV(I915_SET_SPRITE_COLORKEY, intel_sprite_set_colorkey, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), + DRM_IOCTL_DEF_DRV(I915_GET_SPRITE_COLORKEY, intel_sprite_get_colorkey, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED), }; int i915_max_ioctl = DRM_ARRAY_SIZE(i915_ioctls); diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 452bc51..89f3d4a 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -66,7 +66,11 @@ MODULE_PARM_DESC(semaphores, int i915_enable_rc6 __read_mostly = -1; module_param_named(i915_enable_rc6, i915_enable_rc6, int, 0400); MODULE_PARM_DESC(i915_enable_rc6, - "Enable power-saving render C-state 6 (default: -1 (use per-chip default)"); + "Enable power-saving render C-state 6. " + "Different stages can be selected via bitmask values " + "(0 = disable; 1 = enable rc6; 2 = enable deep rc6; 4 = enable deepest rc6). " + "For example, 3 would enable rc6 and deep rc6, and 7 would enable everything. " + "default: -1 (use per-chip default)"); int i915_enable_fbc __read_mostly = -1; module_param_named(i915_enable_fbc, i915_enable_fbc, int, 0600); @@ -103,6 +107,11 @@ MODULE_PARM_DESC(enable_hangcheck, "WARNING: Disabling this can cause system wide hangs. " "(default: true)"); +int i915_enable_ppgtt __read_mostly = -1; +module_param_named(i915_enable_ppgtt, i915_enable_ppgtt, int, 0600); +MODULE_PARM_DESC(i915_enable_ppgtt, + "Enable PPGTT (default: true)"); + static struct drm_driver driver; extern int intel_agp_enabled; @@ -198,7 +207,7 @@ static const struct intel_device_info intel_pineview_info = { static const struct intel_device_info intel_ironlake_d_info = { .gen = 5, - .need_gfx_hws = 1, .has_pipe_cxsr = 1, .has_hotplug = 1, + .need_gfx_hws = 1, .has_hotplug = 1, .has_bsd_ring = 1, }; @@ -214,6 +223,7 @@ static const struct intel_device_info intel_sandybridge_d_info = { .need_gfx_hws = 1, .has_hotplug = 1, .has_bsd_ring = 1, .has_blt_ring = 1, + .has_llc = 1, .has_force_wake = 1, }; @@ -223,6 +233,7 @@ static const struct intel_device_info intel_sandybridge_m_info = { .has_fbc = 1, .has_bsd_ring = 1, .has_blt_ring = 1, + .has_llc = 1, .has_force_wake = 1, }; @@ -231,6 +242,7 @@ static const struct intel_device_info intel_ivybridge_d_info = { .need_gfx_hws = 1, .has_hotplug = 1, .has_bsd_ring = 1, .has_blt_ring = 1, + .has_llc = 1, .has_force_wake = 1, }; @@ -240,6 +252,7 @@ static const struct intel_device_info intel_ivybridge_m_info = { .has_fbc = 0, /* FBC is not enabled on Ivybridge mobile yet */ .has_bsd_ring = 1, .has_blt_ring = 1, + .has_llc = 1, .has_force_wake = 1, }; @@ -381,16 +394,27 @@ void gen6_gt_force_wake_get(struct drm_i915_private *dev_priv) spin_unlock_irqrestore(&dev_priv->gt_lock, irqflags); } +static void gen6_gt_check_fifodbg(struct drm_i915_private *dev_priv) +{ + u32 gtfifodbg; + gtfifodbg = I915_READ_NOTRACE(GTFIFODBG); + if (WARN(gtfifodbg & GT_FIFO_CPU_ERROR_MASK, + "MMIO read or write has been dropped %x\n", gtfifodbg)) + I915_WRITE_NOTRACE(GTFIFODBG, GT_FIFO_CPU_ERROR_MASK); +} + void __gen6_gt_force_wake_put(struct drm_i915_private *dev_priv) { I915_WRITE_NOTRACE(FORCEWAKE, 0); - POSTING_READ(FORCEWAKE); + /* The below doubles as a POSTING_READ */ + gen6_gt_check_fifodbg(dev_priv); } void __gen6_gt_force_wake_mt_put(struct drm_i915_private *dev_priv) { I915_WRITE_NOTRACE(FORCEWAKE_MT, (1<<16) | 0); - POSTING_READ(FORCEWAKE_MT); + /* The below doubles as a POSTING_READ */ + gen6_gt_check_fifodbg(dev_priv); } /* @@ -406,8 +430,10 @@ void gen6_gt_force_wake_put(struct drm_i915_private *dev_priv) spin_unlock_irqrestore(&dev_priv->gt_lock, irqflags); } -void __gen6_gt_wait_for_fifo(struct drm_i915_private *dev_priv) +int __gen6_gt_wait_for_fifo(struct drm_i915_private *dev_priv) { + int ret = 0; + if (dev_priv->gt_fifo_count < GT_FIFO_NUM_RESERVED_ENTRIES) { int loop = 500; u32 fifo = I915_READ_NOTRACE(GT_FIFO_FREE_ENTRIES); @@ -415,10 +441,13 @@ void __gen6_gt_wait_for_fifo(struct drm_i915_private *dev_priv) udelay(10); fifo = I915_READ_NOTRACE(GT_FIFO_FREE_ENTRIES); } - WARN_ON(loop < 0 && fifo <= GT_FIFO_NUM_RESERVED_ENTRIES); + if (WARN_ON(loop < 0 && fifo <= GT_FIFO_NUM_RESERVED_ENTRIES)) + ++ret; dev_priv->gt_fifo_count = fifo; } dev_priv->gt_fifo_count--; + + return ret; } static int i915_drm_freeze(struct drm_device *dev) @@ -503,7 +532,7 @@ static int i915_drm_thaw(struct drm_device *dev) mutex_lock(&dev->struct_mutex); dev_priv->mm.suspended = 0; - error = i915_gem_init_ringbuffer(dev); + error = i915_gem_init_hw(dev); mutex_unlock(&dev->struct_mutex); if (HAS_PCH_SPLIT(dev)) @@ -614,13 +643,40 @@ static int ironlake_do_reset(struct drm_device *dev, u8 flags) static int gen6_do_reset(struct drm_device *dev, u8 flags) { struct drm_i915_private *dev_priv = dev->dev_private; + int ret; + unsigned long irqflags; + + /* Hold gt_lock across reset to prevent any register access + * with forcewake not set correctly + */ + spin_lock_irqsave(&dev_priv->gt_lock, irqflags); + + /* Reset the chip */ + + /* GEN6_GDRST is not in the gt power well, no need to check + * for fifo space for the write or forcewake the chip for + * the read + */ + I915_WRITE_NOTRACE(GEN6_GDRST, GEN6_GRDOM_FULL); + + /* Spin waiting for the device to ack the reset request */ + ret = wait_for((I915_READ_NOTRACE(GEN6_GDRST) & GEN6_GRDOM_FULL) == 0, 500); - I915_WRITE(GEN6_GDRST, GEN6_GRDOM_FULL); - return wait_for((I915_READ(GEN6_GDRST) & GEN6_GRDOM_FULL) == 0, 500); + /* If reset with a user forcewake, try to restore, otherwise turn it off */ + if (dev_priv->forcewake_count) + dev_priv->display.force_wake_get(dev_priv); + else + dev_priv->display.force_wake_put(dev_priv); + + /* Restore fifo count */ + dev_priv->gt_fifo_count = I915_READ_NOTRACE(GT_FIFO_FREE_ENTRIES); + + spin_unlock_irqrestore(&dev_priv->gt_lock, irqflags); + return ret; } /** - * i965_reset - reset chip after a hang + * i915_reset - reset chip after a hang * @dev: drm device to reset * @flags: reset domains * @@ -643,7 +699,6 @@ int i915_reset(struct drm_device *dev, u8 flags) * need to */ bool need_display = true; - unsigned long irqflags; int ret; if (!i915_try_reset) @@ -661,11 +716,6 @@ int i915_reset(struct drm_device *dev, u8 flags) case 7: case 6: ret = gen6_do_reset(dev, flags); - /* If reset with a user forcewake, try to restore */ - spin_lock_irqsave(&dev_priv->gt_lock, irqflags); - if (dev_priv->forcewake_count) - dev_priv->display.force_wake_get(dev_priv); - spin_unlock_irqrestore(&dev_priv->gt_lock, irqflags); break; case 5: ret = ironlake_do_reset(dev, flags); @@ -702,12 +752,16 @@ int i915_reset(struct drm_device *dev, u8 flags) !dev_priv->mm.suspended) { dev_priv->mm.suspended = 0; + i915_gem_init_swizzling(dev); + dev_priv->ring[RCS].init(&dev_priv->ring[RCS]); if (HAS_BSD(dev)) dev_priv->ring[VCS].init(&dev_priv->ring[VCS]); if (HAS_BLT(dev)) dev_priv->ring[BCS].init(&dev_priv->ring[BCS]); + i915_gem_init_ppgtt(dev); + mutex_unlock(&dev->struct_mutex); drm_irq_uninstall(dev); drm_mode_config_reset(dev); @@ -830,6 +884,21 @@ static struct vm_operations_struct i915_gem_vm_ops = { .close = drm_gem_vm_close, }; +static const struct file_operations i915_driver_fops = { + .owner = THIS_MODULE, + .open = drm_open, + .release = drm_release, + .unlocked_ioctl = drm_ioctl, + .mmap = drm_gem_mmap, + .poll = drm_poll, + .fasync = drm_fasync, + .read = drm_read, +#ifdef CONFIG_COMPAT + .compat_ioctl = i915_compat_ioctl, +#endif + .llseek = noop_llseek, +}; + static struct drm_driver driver = { /* Don't use MTRRs here; the Xserver or userspace app should * deal with them for Intel hardware. @@ -863,21 +932,7 @@ static struct drm_driver driver = { .dumb_map_offset = i915_gem_mmap_gtt, .dumb_destroy = i915_gem_dumb_destroy, .ioctls = i915_ioctls, - .fops = { - .owner = THIS_MODULE, - .open = drm_open, - .release = drm_release, - .unlocked_ioctl = drm_ioctl, - .mmap = drm_gem_mmap, - .poll = drm_poll, - .fasync = drm_fasync, - .read = drm_read, -#ifdef CONFIG_COMPAT - .compat_ioctl = i915_compat_ioctl, -#endif - .llseek = noop_llseek, - }, - + .fops = &i915_driver_fops, .name = DRIVER_NAME, .desc = DRIVER_DESC, .date = DRIVER_DATE, @@ -945,17 +1000,21 @@ MODULE_LICENSE("GPL and additional rights"); /* We give fast paths for the really cool registers */ #define NEEDS_FORCE_WAKE(dev_priv, reg) \ ((HAS_FORCE_WAKE((dev_priv)->dev)) && \ - ((reg) < 0x40000) && \ - ((reg) != FORCEWAKE) && \ - ((reg) != ECOBUS)) + ((reg) < 0x40000) && \ + ((reg) != FORCEWAKE)) #define __i915_read(x, y) \ u##x i915_read##x(struct drm_i915_private *dev_priv, u32 reg) { \ u##x val = 0; \ if (NEEDS_FORCE_WAKE((dev_priv), (reg))) { \ - gen6_gt_force_wake_get(dev_priv); \ + unsigned long irqflags; \ + spin_lock_irqsave(&dev_priv->gt_lock, irqflags); \ + if (dev_priv->forcewake_count == 0) \ + dev_priv->display.force_wake_get(dev_priv); \ val = read##y(dev_priv->regs + reg); \ - gen6_gt_force_wake_put(dev_priv); \ + if (dev_priv->forcewake_count == 0) \ + dev_priv->display.force_wake_put(dev_priv); \ + spin_unlock_irqrestore(&dev_priv->gt_lock, irqflags); \ } else { \ val = read##y(dev_priv->regs + reg); \ } \ @@ -971,11 +1030,15 @@ __i915_read(64, q) #define __i915_write(x, y) \ void i915_write##x(struct drm_i915_private *dev_priv, u32 reg, u##x val) { \ + u32 __fifo_ret = 0; \ trace_i915_reg_rw(true, reg, val, sizeof(val)); \ if (NEEDS_FORCE_WAKE((dev_priv), (reg))) { \ - __gen6_gt_wait_for_fifo(dev_priv); \ + __fifo_ret = __gen6_gt_wait_for_fifo(dev_priv); \ } \ write##y(val, dev_priv->regs + reg); \ + if (unlikely(__fifo_ret)) { \ + gen6_gt_check_fifodbg(dev_priv); \ + } \ } __i915_write(8, b) __i915_write(16, w) diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 012a9d2..45c5cf8 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -35,6 +35,7 @@ #include "intel_ringbuffer.h" #include #include +#include #include #include @@ -135,6 +136,7 @@ struct drm_i915_fence_reg { struct list_head lru_list; struct drm_i915_gem_object *obj; uint32_t setup_seqno; + int pin_count; }; struct sdvo_device_mapping { @@ -152,33 +154,40 @@ struct drm_i915_error_state { u32 eir; u32 pgtbl_er; u32 pipestat[I915_MAX_PIPES]; - u32 ipeir; - u32 ipehr; - u32 instdone; - u32 acthd; + u32 tail[I915_NUM_RINGS]; + u32 head[I915_NUM_RINGS]; + u32 ipeir[I915_NUM_RINGS]; + u32 ipehr[I915_NUM_RINGS]; + u32 instdone[I915_NUM_RINGS]; + u32 acthd[I915_NUM_RINGS]; + u32 semaphore_mboxes[I915_NUM_RINGS][I915_NUM_RINGS - 1]; + /* our own tracking of ring head and tail */ + u32 cpu_ring_head[I915_NUM_RINGS]; + u32 cpu_ring_tail[I915_NUM_RINGS]; u32 error; /* gen6+ */ - u32 bcs_acthd; /* gen6+ blt engine */ - u32 bcs_ipehr; - u32 bcs_ipeir; - u32 bcs_instdone; - u32 bcs_seqno; - u32 vcs_acthd; /* gen6+ bsd engine */ - u32 vcs_ipehr; - u32 vcs_ipeir; - u32 vcs_instdone; - u32 vcs_seqno; - u32 instpm; - u32 instps; + u32 instpm[I915_NUM_RINGS]; + u32 instps[I915_NUM_RINGS]; u32 instdone1; - u32 seqno; + u32 seqno[I915_NUM_RINGS]; u64 bbaddr; + u32 fault_reg[I915_NUM_RINGS]; + u32 done_reg; + u32 faddr[I915_NUM_RINGS]; u64 fence[I915_MAX_NUM_FENCES]; struct timeval time; - struct drm_i915_error_object { - int page_count; - u32 gtt_offset; - u32 *pages[0]; - } *ringbuffer[I915_NUM_RINGS], *batchbuffer[I915_NUM_RINGS]; + struct drm_i915_error_ring { + struct drm_i915_error_object { + int page_count; + u32 gtt_offset; + u32 *pages[0]; + } *ringbuffer, *batchbuffer; + struct drm_i915_error_request { + long jiffies; + u32 seqno; + u32 tail; + } *requests; + int num_requests; + } ring[I915_NUM_RINGS]; struct drm_i915_error_buffer { u32 size; u32 name; @@ -191,7 +200,7 @@ struct drm_i915_error_state { u32 tiling:2; u32 dirty:1; u32 purgeable:1; - u32 ring:4; + s32 ring:4; u32 cache_level:2; } *active_bo, *pinned_bo; u32 active_bo_count, pinned_bo_count; @@ -207,6 +216,8 @@ struct drm_i915_display_funcs { int (*get_display_clock_speed)(struct drm_device *dev); int (*get_fifo_size)(struct drm_device *dev, int plane); void (*update_wm)(struct drm_device *dev); + void (*update_sprite_wm)(struct drm_device *dev, int pipe, + uint32_t sprite_width, int pixel_size); int (*crtc_mode_set)(struct drm_crtc *crtc, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode, @@ -254,6 +265,17 @@ struct intel_device_info { u8 supports_tv:1; u8 has_bsd_ring:1; u8 has_blt_ring:1; + u8 has_llc:1; +}; + +#define I915_PPGTT_PD_ENTRIES 512 +#define I915_PPGTT_PT_ENTRIES 1024 +struct i915_hw_ppgtt { + unsigned num_pd_entries; + struct page **pt_pages; + uint32_t pd_offset; + dma_addr_t *pt_dma_addr; + dma_addr_t scratch_page_dma_addr; }; enum no_fbc_reason { @@ -278,6 +300,16 @@ enum intel_pch { struct intel_fbdev; struct intel_fbc_work; +struct intel_gmbus { + struct i2c_adapter adapter; + bool force_bit; + bool has_gpio; + u32 reg0; + u32 gpio_reg; + struct i2c_algo_bit_data bit_algo; + struct drm_i915_private *dev_priv; +}; + typedef struct drm_i915_private { struct drm_device *dev; @@ -295,11 +327,11 @@ typedef struct drm_i915_private { /** gt_lock is also taken in irq contexts. */ struct spinlock gt_lock; - struct intel_gmbus { - struct i2c_adapter adapter; - struct i2c_adapter *force_bit; - u32 reg0; - } *gmbus; + struct intel_gmbus *gmbus; + + /** gmbus_mutex protects against concurrent usage of the single hw gmbus + * controller on different i2c buses. */ + struct mutex gmbus_mutex; struct pci_dev *bridge_dev; struct intel_ring_buffer ring[I915_NUM_RINGS]; @@ -334,7 +366,6 @@ typedef struct drm_i915_private { int tex_lru_log_granularity; int allow_batchbuffer; - struct mem_block *agp_heap; unsigned int sr01, adpa, ppcr, dvob, dvoc, lvds; int vblank_pipe; int num_pipe; @@ -359,6 +390,7 @@ typedef struct drm_i915_private { /* overlay */ struct intel_overlay *overlay; + bool sprite_scaling_enabled; /* LVDS info */ int backlight_level; /* restore backlight to this value */ @@ -584,6 +616,9 @@ typedef struct drm_i915_private { struct io_mapping *gtt_mapping; int gtt_mtrr; + /** PPGTT used for aliasing the PPGTT with the GTT */ + struct i915_hw_ppgtt *aliasing_ppgtt; + struct shrinker inactive_shrinker; /** @@ -749,6 +784,13 @@ typedef struct drm_i915_private { struct drm_property *force_audio_property; } drm_i915_private_t; +enum hdmi_force_audio { + HDMI_AUDIO_OFF_DVI = -2, /* no aux data for HDMI-DVI converter */ + HDMI_AUDIO_OFF, /* force turn off HDMI audio */ + HDMI_AUDIO_AUTO, /* trust EDID */ + HDMI_AUDIO_ON, /* force turn on HDMI audio */ +}; + enum i915_cache_level { I915_CACHE_NONE, I915_CACHE_LLC, @@ -841,6 +883,8 @@ struct drm_i915_gem_object { unsigned int cache_level:2; + unsigned int has_aliasing_ppgtt_mapping:1; + struct page **pages; /** @@ -918,6 +962,9 @@ struct drm_i915_gem_request { /** GEM sequence number associated with this request. */ uint32_t seqno; + /** Postion in the ringbuffer of the end of the request */ + u32 tail; + /** Time at which this request was emitted, in jiffies. */ unsigned long emitted_jiffies; @@ -974,8 +1021,11 @@ struct drm_i915_file_private { #define HAS_BSD(dev) (INTEL_INFO(dev)->has_bsd_ring) #define HAS_BLT(dev) (INTEL_INFO(dev)->has_blt_ring) +#define HAS_LLC(dev) (INTEL_INFO(dev)->has_llc) #define I915_NEED_GFX_HWS(dev) (INTEL_INFO(dev)->need_gfx_hws) +#define HAS_ALIASING_PPGTT(dev) (INTEL_INFO(dev)->gen >=6) + #define HAS_OVERLAY(dev) (INTEL_INFO(dev)->has_overlay) #define OVERLAY_NEEDS_PHYSICAL(dev) (INTEL_INFO(dev)->overlay_needs_physical) @@ -1008,6 +1058,27 @@ struct drm_i915_file_private { #include "i915_trace.h" +/** + * RC6 is a special power stage which allows the GPU to enter an very + * low-voltage mode when idle, using down to 0V while at this stage. This + * stage is entered automatically when the GPU is idle when RC6 support is + * enabled, and as soon as new workload arises GPU wakes up automatically as well. + * + * There are different RC6 modes available in Intel GPU, which differentiate + * among each other with the latency required to enter and leave RC6 and + * voltage consumed by the GPU in different states. + * + * The combination of the following flags define which states GPU is allowed + * to enter, while RC6 is the normal RC6 state, RC6p is the deep RC6, and + * RC6pp is deepest RC6. Their support by hardware varies according to the + * GPU, BIOS, chipset and platform. RC6 is usually the safest one and the one + * which brings the most power savings; deeper states save more power, but + * require higher latency to switch to and wake up. + */ +#define INTEL_RC6_ENABLE (1<<0) +#define INTEL_RC6p_ENABLE (1<<1) +#define INTEL_RC6pp_ENABLE (1<<2) + extern struct drm_ioctl_desc i915_ioctls[]; extern int i915_max_ioctl; extern unsigned int i915_fbpercrtc __always_unused; @@ -1020,6 +1091,7 @@ extern int i915_vbt_sdvo_panel_type __read_mostly; extern int i915_enable_rc6 __read_mostly; extern int i915_enable_fbc __read_mostly; extern bool i915_enable_hangcheck __read_mostly; +extern int i915_enable_ppgtt __read_mostly; extern int i915_suspend(struct drm_device *dev, pm_message_t state); extern int i915_resume(struct drm_device *dev); @@ -1081,18 +1153,6 @@ extern void i915_destroy_error_state(struct drm_device *dev); #endif -/* i915_mem.c */ -extern int i915_mem_alloc(struct drm_device *dev, void *data, - struct drm_file *file_priv); -extern int i915_mem_free(struct drm_device *dev, void *data, - struct drm_file *file_priv); -extern int i915_mem_init_heap(struct drm_device *dev, void *data, - struct drm_file *file_priv); -extern int i915_mem_destroy_heap(struct drm_device *dev, void *data, - struct drm_file *file_priv); -extern void i915_mem_takedown(struct mem_block **heap); -extern void i915_mem_release(struct drm_device * dev, - struct drm_file *file_priv, struct mem_block *heap); /* i915_gem.c */ int i915_gem_init_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); @@ -1178,26 +1238,49 @@ int __must_check i915_gem_object_get_fence(struct drm_i915_gem_object *obj, struct intel_ring_buffer *pipelined); int __must_check i915_gem_object_put_fence(struct drm_i915_gem_object *obj); +static inline void +i915_gem_object_pin_fence(struct drm_i915_gem_object *obj) +{ + if (obj->fence_reg != I915_FENCE_REG_NONE) { + struct drm_i915_private *dev_priv = obj->base.dev->dev_private; + dev_priv->fence_regs[obj->fence_reg].pin_count++; + } +} + +static inline void +i915_gem_object_unpin_fence(struct drm_i915_gem_object *obj) +{ + if (obj->fence_reg != I915_FENCE_REG_NONE) { + struct drm_i915_private *dev_priv = obj->base.dev->dev_private; + dev_priv->fence_regs[obj->fence_reg].pin_count--; + } +} + void i915_gem_retire_requests(struct drm_device *dev); +void i915_gem_retire_requests_ring(struct intel_ring_buffer *ring); + void i915_gem_reset(struct drm_device *dev); void i915_gem_clflush_object(struct drm_i915_gem_object *obj); int __must_check i915_gem_object_set_domain(struct drm_i915_gem_object *obj, uint32_t read_domains, uint32_t write_domain); int __must_check i915_gem_object_finish_gpu(struct drm_i915_gem_object *obj); -int __must_check i915_gem_init_ringbuffer(struct drm_device *dev); +int __must_check i915_gem_init_hw(struct drm_device *dev); +void i915_gem_init_swizzling(struct drm_device *dev); +void i915_gem_init_ppgtt(struct drm_device *dev); void i915_gem_cleanup_ringbuffer(struct drm_device *dev); void i915_gem_do_init(struct drm_device *dev, unsigned long start, unsigned long mappable_end, unsigned long end); -int __must_check i915_gpu_idle(struct drm_device *dev); +int __must_check i915_gpu_idle(struct drm_device *dev, bool do_retire); int __must_check i915_gem_idle(struct drm_device *dev); int __must_check i915_add_request(struct intel_ring_buffer *ring, struct drm_file *file, struct drm_i915_gem_request *request); int __must_check i915_wait_request(struct intel_ring_buffer *ring, - uint32_t seqno); + uint32_t seqno, + bool do_retire); int i915_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf); int __must_check i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj, @@ -1224,6 +1307,14 @@ int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj, enum i915_cache_level cache_level); /* i915_gem_gtt.c */ +int __must_check i915_gem_init_aliasing_ppgtt(struct drm_device *dev); +void i915_gem_cleanup_aliasing_ppgtt(struct drm_device *dev); +void i915_ppgtt_bind_object(struct i915_hw_ppgtt *ppgtt, + struct drm_i915_gem_object *obj, + enum i915_cache_level cache_level); +void i915_ppgtt_unbind_object(struct i915_hw_ppgtt *ppgtt, + struct drm_i915_gem_object *obj); + void i915_gem_restore_gtt_mappings(struct drm_device *dev); int __must_check i915_gem_gtt_bind_object(struct drm_i915_gem_object *obj); void i915_gem_gtt_rebind_object(struct drm_i915_gem_object *obj, @@ -1305,7 +1396,6 @@ static inline void intel_unregister_dsm_handler(void) { return; } #endif /* CONFIG_ACPI */ /* modesetting */ -extern void i915_redisable_vga(struct drm_device *dev); extern void intel_modeset_init(struct drm_device *dev); extern void intel_modeset_gem_init(struct drm_device *dev); extern void intel_modeset_cleanup(struct drm_device *dev); @@ -1363,14 +1453,7 @@ extern void intel_display_print_error_state(struct seq_file *m, */ void gen6_gt_force_wake_get(struct drm_i915_private *dev_priv); void gen6_gt_force_wake_put(struct drm_i915_private *dev_priv); -void __gen6_gt_wait_for_fifo(struct drm_i915_private *dev_priv); - -/* We give fast paths for the really cool registers */ -#define NEEDS_FORCE_WAKE(dev_priv, reg) \ - (((dev_priv)->info->gen >= 6) && \ - ((reg) < 0x40000) && \ - ((reg) != FORCEWAKE) && \ - ((reg) != ECOBUS)) +int __gen6_gt_wait_for_fifo(struct drm_i915_private *dev_priv); #define __i915_read(x, y) \ u##x i915_read##x(struct drm_i915_private *dev_priv, u32 reg); diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index b0186b8..a230a93 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -58,6 +58,7 @@ static void i915_gem_free_object_tail(struct drm_i915_gem_object *obj); static int i915_gem_inactive_shrink(struct shrinker *shrinker, struct shrink_control *sc); +static void i915_gem_object_truncate(struct drm_i915_gem_object *obj); /* some bookkeeping */ static void i915_gem_info_add_obj(struct drm_i915_private *dev_priv, @@ -258,73 +259,6 @@ static int i915_gem_object_needs_bit17_swizzle(struct drm_i915_gem_object *obj) obj->tiling_mode != I915_TILING_NONE; } -static inline void -slow_shmem_copy(struct page *dst_page, - int dst_offset, - struct page *src_page, - int src_offset, - int length) -{ - char *dst_vaddr, *src_vaddr; - - dst_vaddr = kmap(dst_page); - src_vaddr = kmap(src_page); - - memcpy(dst_vaddr + dst_offset, src_vaddr + src_offset, length); - - kunmap(src_page); - kunmap(dst_page); -} - -static inline void -slow_shmem_bit17_copy(struct page *gpu_page, - int gpu_offset, - struct page *cpu_page, - int cpu_offset, - int length, - int is_read) -{ - char *gpu_vaddr, *cpu_vaddr; - - /* Use the unswizzled path if this page isn't affected. */ - if ((page_to_phys(gpu_page) & (1 << 17)) == 0) { - if (is_read) - return slow_shmem_copy(cpu_page, cpu_offset, - gpu_page, gpu_offset, length); - else - return slow_shmem_copy(gpu_page, gpu_offset, - cpu_page, cpu_offset, length); - } - - gpu_vaddr = kmap(gpu_page); - cpu_vaddr = kmap(cpu_page); - - /* Copy the data, XORing A6 with A17 (1). The user already knows he's - * XORing with the other bits (A9 for Y, A9 and A10 for X) - */ - while (length > 0) { - int cacheline_end = ALIGN(gpu_offset + 1, 64); - int this_length = min(cacheline_end - gpu_offset, length); - int swizzled_gpu_offset = gpu_offset ^ 64; - - if (is_read) { - memcpy(cpu_vaddr + cpu_offset, - gpu_vaddr + swizzled_gpu_offset, - this_length); - } else { - memcpy(gpu_vaddr + swizzled_gpu_offset, - cpu_vaddr + cpu_offset, - this_length); - } - cpu_offset += this_length; - gpu_offset += this_length; - length -= this_length; - } - - kunmap(cpu_page); - kunmap(gpu_page); -} - /** * This is the fast shmem pread path, which attempts to copy_from_user directly * from the backing pages of the object to the user's address space. On a @@ -385,6 +319,58 @@ i915_gem_shmem_pread_fast(struct drm_device *dev, return 0; } +static inline int +__copy_to_user_swizzled(char __user *cpu_vaddr, + const char *gpu_vaddr, int gpu_offset, + int length) +{ + int ret, cpu_offset = 0; + + while (length > 0) { + int cacheline_end = ALIGN(gpu_offset + 1, 64); + int this_length = min(cacheline_end - gpu_offset, length); + int swizzled_gpu_offset = gpu_offset ^ 64; + + ret = __copy_to_user(cpu_vaddr + cpu_offset, + gpu_vaddr + swizzled_gpu_offset, + this_length); + if (ret) + return ret + length; + + cpu_offset += this_length; + gpu_offset += this_length; + length -= this_length; + } + + return 0; +} + +static inline int +__copy_from_user_swizzled(char __user *gpu_vaddr, int gpu_offset, + const char *cpu_vaddr, + int length) +{ + int ret, cpu_offset = 0; + + while (length > 0) { + int cacheline_end = ALIGN(gpu_offset + 1, 64); + int this_length = min(cacheline_end - gpu_offset, length); + int swizzled_gpu_offset = gpu_offset ^ 64; + + ret = __copy_from_user(gpu_vaddr + swizzled_gpu_offset, + cpu_vaddr + cpu_offset, + this_length); + if (ret) + return ret + length; + + cpu_offset += this_length; + gpu_offset += this_length; + length -= this_length; + } + + return 0; +} + /** * This is the fallback shmem pread path, which allocates temporary storage * in kernel space to copy_to_user into outside of the struct_mutex, so we @@ -398,72 +384,34 @@ i915_gem_shmem_pread_slow(struct drm_device *dev, struct drm_file *file) { struct address_space *mapping = obj->base.filp->f_path.dentry->d_inode->i_mapping; - struct mm_struct *mm = current->mm; - struct page **user_pages; + char __user *user_data; ssize_t remain; - loff_t offset, pinned_pages, i; - loff_t first_data_page, last_data_page, num_pages; - int shmem_page_offset; - int data_page_index, data_page_offset; - int page_length; - int ret; - uint64_t data_ptr = args->data_ptr; - int do_bit17_swizzling; + loff_t offset; + int shmem_page_offset, page_length, ret; + int obj_do_bit17_swizzling, page_do_bit17_swizzling; + user_data = (char __user *) (uintptr_t) args->data_ptr; remain = args->size; - /* Pin the user pages containing the data. We can't fault while - * holding the struct mutex, yet we want to hold it while - * dereferencing the user data. - */ - first_data_page = data_ptr / PAGE_SIZE; - last_data_page = (data_ptr + args->size - 1) / PAGE_SIZE; - num_pages = last_data_page - first_data_page + 1; + obj_do_bit17_swizzling = i915_gem_object_needs_bit17_swizzle(obj); - user_pages = drm_malloc_ab(num_pages, sizeof(struct page *)); - if (user_pages == NULL) - return -ENOMEM; + offset = args->offset; mutex_unlock(&dev->struct_mutex); - down_read(&mm->mmap_sem); - pinned_pages = get_user_pages(current, mm, (uintptr_t)args->data_ptr, - num_pages, 1, 0, user_pages, NULL); - up_read(&mm->mmap_sem); - mutex_lock(&dev->struct_mutex); - if (pinned_pages < num_pages) { - ret = -EFAULT; - goto out; - } - - ret = i915_gem_object_set_cpu_read_domain_range(obj, - args->offset, - args->size); - if (ret) - goto out; - - do_bit17_swizzling = i915_gem_object_needs_bit17_swizzle(obj); - - offset = args->offset; while (remain > 0) { struct page *page; + char *vaddr; /* Operation in this page * * shmem_page_offset = offset within page in shmem file - * data_page_index = page number in get_user_pages return - * data_page_offset = offset with data_page_index page. * page_length = bytes to copy for this page */ shmem_page_offset = offset_in_page(offset); - data_page_index = data_ptr / PAGE_SIZE - first_data_page; - data_page_offset = offset_in_page(data_ptr); - page_length = remain; if ((shmem_page_offset + page_length) > PAGE_SIZE) page_length = PAGE_SIZE - shmem_page_offset; - if ((data_page_offset + page_length) > PAGE_SIZE) - page_length = PAGE_SIZE - data_page_offset; page = shmem_read_mapping_page(mapping, offset >> PAGE_SHIFT); if (IS_ERR(page)) { @@ -471,36 +419,38 @@ i915_gem_shmem_pread_slow(struct drm_device *dev, goto out; } - if (do_bit17_swizzling) { - slow_shmem_bit17_copy(page, - shmem_page_offset, - user_pages[data_page_index], - data_page_offset, - page_length, - 1); - } else { - slow_shmem_copy(user_pages[data_page_index], - data_page_offset, - page, - shmem_page_offset, - page_length); - } + page_do_bit17_swizzling = obj_do_bit17_swizzling && + (page_to_phys(page) & (1 << 17)) != 0; + + vaddr = kmap(page); + if (page_do_bit17_swizzling) + ret = __copy_to_user_swizzled(user_data, + vaddr, shmem_page_offset, + page_length); + else + ret = __copy_to_user(user_data, + vaddr + shmem_page_offset, + page_length); + kunmap(page); mark_page_accessed(page); page_cache_release(page); + if (ret) { + ret = -EFAULT; + goto out; + } + remain -= page_length; - data_ptr += page_length; + user_data += page_length; offset += page_length; } out: - for (i = 0; i < pinned_pages; i++) { - SetPageDirty(user_pages[i]); - mark_page_accessed(user_pages[i]); - page_cache_release(user_pages[i]); - } - drm_free_large(user_pages); + mutex_lock(&dev->struct_mutex); + /* Fixup: Kill any reinstated backing storage pages */ + if (obj->madv == __I915_MADV_PURGED) + i915_gem_object_truncate(obj); return ret; } @@ -841,71 +791,36 @@ i915_gem_shmem_pwrite_slow(struct drm_device *dev, struct drm_file *file) { struct address_space *mapping = obj->base.filp->f_path.dentry->d_inode->i_mapping; - struct mm_struct *mm = current->mm; - struct page **user_pages; ssize_t remain; - loff_t offset, pinned_pages, i; - loff_t first_data_page, last_data_page, num_pages; - int shmem_page_offset; - int data_page_index, data_page_offset; - int page_length; - int ret; - uint64_t data_ptr = args->data_ptr; - int do_bit17_swizzling; + loff_t offset; + char __user *user_data; + int shmem_page_offset, page_length, ret; + int obj_do_bit17_swizzling, page_do_bit17_swizzling; + user_data = (char __user *) (uintptr_t) args->data_ptr; remain = args->size; - /* Pin the user pages containing the data. We can't fault while - * holding the struct mutex, and all of the pwrite implementations - * want to hold it while dereferencing the user data. - */ - first_data_page = data_ptr / PAGE_SIZE; - last_data_page = (data_ptr + args->size - 1) / PAGE_SIZE; - num_pages = last_data_page - first_data_page + 1; - - user_pages = drm_malloc_ab(num_pages, sizeof(struct page *)); - if (user_pages == NULL) - return -ENOMEM; - - mutex_unlock(&dev->struct_mutex); - down_read(&mm->mmap_sem); - pinned_pages = get_user_pages(current, mm, (uintptr_t)args->data_ptr, - num_pages, 0, 0, user_pages, NULL); - up_read(&mm->mmap_sem); - mutex_lock(&dev->struct_mutex); - if (pinned_pages < num_pages) { - ret = -EFAULT; - goto out; - } - - ret = i915_gem_object_set_to_cpu_domain(obj, 1); - if (ret) - goto out; - - do_bit17_swizzling = i915_gem_object_needs_bit17_swizzle(obj); + obj_do_bit17_swizzling = i915_gem_object_needs_bit17_swizzle(obj); offset = args->offset; obj->dirty = 1; + mutex_unlock(&dev->struct_mutex); + while (remain > 0) { struct page *page; + char *vaddr; /* Operation in this page * * shmem_page_offset = offset within page in shmem file - * data_page_index = page number in get_user_pages return - * data_page_offset = offset with data_page_index page. * page_length = bytes to copy for this page */ shmem_page_offset = offset_in_page(offset); - data_page_index = data_ptr / PAGE_SIZE - first_data_page; - data_page_offset = offset_in_page(data_ptr); page_length = remain; if ((shmem_page_offset + page_length) > PAGE_SIZE) page_length = PAGE_SIZE - shmem_page_offset; - if ((data_page_offset + page_length) > PAGE_SIZE) - page_length = PAGE_SIZE - data_page_offset; page = shmem_read_mapping_page(mapping, offset >> PAGE_SHIFT); if (IS_ERR(page)) { @@ -913,34 +828,45 @@ i915_gem_shmem_pwrite_slow(struct drm_device *dev, goto out; } - if (do_bit17_swizzling) { - slow_shmem_bit17_copy(page, - shmem_page_offset, - user_pages[data_page_index], - data_page_offset, - page_length, - 0); - } else { - slow_shmem_copy(page, - shmem_page_offset, - user_pages[data_page_index], - data_page_offset, - page_length); - } + page_do_bit17_swizzling = obj_do_bit17_swizzling && + (page_to_phys(page) & (1 << 17)) != 0; + + vaddr = kmap(page); + if (page_do_bit17_swizzling) + ret = __copy_from_user_swizzled(vaddr, shmem_page_offset, + user_data, + page_length); + else + ret = __copy_from_user(vaddr + shmem_page_offset, + user_data, + page_length); + kunmap(page); set_page_dirty(page); mark_page_accessed(page); page_cache_release(page); + if (ret) { + ret = -EFAULT; + goto out; + } + remain -= page_length; - data_ptr += page_length; + user_data += page_length; offset += page_length; } out: - for (i = 0; i < pinned_pages; i++) - page_cache_release(user_pages[i]); - drm_free_large(user_pages); + mutex_lock(&dev->struct_mutex); + /* Fixup: Kill any reinstated backing storage pages */ + if (obj->madv == __I915_MADV_PURGED) + i915_gem_object_truncate(obj); + /* and flush dirty cachelines in case the object isn't in the cpu write + * domain anymore. */ + if (obj->base.write_domain != I915_GEM_DOMAIN_CPU) { + i915_gem_clflush_object(obj); + intel_gtt_chipset_flush(); + } return ret; } @@ -996,11 +922,14 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data, * pread/pwrite currently are reading and writing from the CPU * perspective, requiring manual detiling by the client. */ - if (obj->phys_obj) + if (obj->phys_obj) { ret = i915_gem_phys_pwrite(dev, obj, args, file); - else if (obj->gtt_space && - obj->tiling_mode == I915_TILING_NONE && - obj->base.write_domain != I915_GEM_DOMAIN_CPU) { + goto out; + } + + if (obj->gtt_space && + obj->tiling_mode == I915_TILING_NONE && + obj->base.write_domain != I915_GEM_DOMAIN_CPU) { ret = i915_gem_object_pin(obj, 0, true); if (ret) goto out; @@ -1019,18 +948,24 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data, out_unpin: i915_gem_object_unpin(obj); - } else { - ret = i915_gem_object_set_to_cpu_domain(obj, 1); - if (ret) - goto out; - ret = -EFAULT; - if (!i915_gem_object_needs_bit17_swizzle(obj)) - ret = i915_gem_shmem_pwrite_fast(dev, obj, args, file); - if (ret == -EFAULT) - ret = i915_gem_shmem_pwrite_slow(dev, obj, args, file); + if (ret != -EFAULT) + goto out; + /* Fall through to the shmfs paths because the gtt paths might + * fail with non-page-backed user pointers (e.g. gtt mappings + * when moving data between textures). */ } + ret = i915_gem_object_set_to_cpu_domain(obj, 1); + if (ret) + goto out; + + ret = -EFAULT; + if (!i915_gem_object_needs_bit17_swizzle(obj)) + ret = i915_gem_shmem_pwrite_fast(dev, obj, args, file); + if (ret == -EFAULT) + ret = i915_gem_shmem_pwrite_slow(dev, obj, args, file); + out: drm_gem_object_unreference(&obj->base); unlock: @@ -1142,7 +1077,6 @@ int i915_gem_mmap_ioctl(struct drm_device *dev, void *data, struct drm_file *file) { - struct drm_i915_private *dev_priv = dev->dev_private; struct drm_i915_gem_mmap *args = data; struct drm_gem_object *obj; unsigned long addr; @@ -1154,16 +1088,9 @@ i915_gem_mmap_ioctl(struct drm_device *dev, void *data, if (obj == NULL) return -ENOENT; - if (obj->size > dev_priv->mm.gtt_mappable_end) { - drm_gem_object_unreference_unlocked(obj); - return -E2BIG; - } - - down_write(¤t->mm->mmap_sem); - addr = do_mmap(obj->filp, 0, args->size, + addr = vm_mmap(obj->filp, 0, args->size, PROT_READ | PROT_WRITE, MAP_SHARED, args->offset); - up_write(¤t->mm->mmap_sem); drm_gem_object_unreference_unlocked(obj); if (IS_ERR((void *)addr)) return addr; @@ -1259,11 +1186,6 @@ out: case 0: case -ERESTARTSYS: case -EINTR: - case -EBUSY: - /* - * EBUSY is ok: this just means that another thread - * already did the job. - */ return VM_FAULT_NOPAGE; case -ENOMEM: return VM_FAULT_OOM; @@ -1687,12 +1609,20 @@ i915_add_request(struct intel_ring_buffer *ring, { drm_i915_private_t *dev_priv = ring->dev->dev_private; uint32_t seqno; + u32 request_ring_position; int was_empty; int ret; BUG_ON(request == NULL); seqno = i915_gem_next_request_seqno(ring); + /* Record the position of the start of the request so that + * should we detect the updated seqno part-way through the + * GPU processing the request, we never over-estimate the + * position of the head. + */ + request_ring_position = intel_ring_get_tail(ring); + ret = ring->add_request(ring, &seqno); if (ret) return ret; @@ -1701,6 +1631,7 @@ i915_add_request(struct intel_ring_buffer *ring, request->seqno = seqno; request->ring = ring; + request->tail = request_ring_position; request->emitted_jiffies = jiffies; was_empty = list_empty(&ring->request_list); list_add_tail(&request->list, &ring->request_list); @@ -1715,7 +1646,7 @@ i915_add_request(struct intel_ring_buffer *ring, spin_unlock(&file_priv->mm.lock); } - ring->outstanding_lazy_request = false; + ring->outstanding_lazy_request = 0; if (!dev_priv->mm.suspended) { if (i915_enable_hangcheck) { @@ -1837,7 +1768,7 @@ void i915_gem_reset(struct drm_device *dev) /** * This function clears the request list as sequence numbers are passed. */ -static void +void i915_gem_retire_requests_ring(struct intel_ring_buffer *ring) { uint32_t seqno; @@ -1865,6 +1796,12 @@ i915_gem_retire_requests_ring(struct intel_ring_buffer *ring) break; trace_i915_gem_request_retire(ring, request->seqno); + /* We know the GPU must have read the request to have + * sent us the seqno + interrupt, so use the position + * of tail of the request to update the last known position + * of the GPU head. + */ + ring->last_retired_head = request->tail; list_del(&request->list); i915_gem_request_remove_from_client(request); @@ -1977,7 +1914,8 @@ i915_gem_retire_work_handler(struct work_struct *work) */ int i915_wait_request(struct intel_ring_buffer *ring, - uint32_t seqno) + uint32_t seqno, + bool do_retire) { drm_i915_private_t *dev_priv = ring->dev->dev_private; u32 ier; @@ -2040,9 +1978,9 @@ i915_wait_request(struct intel_ring_buffer *ring, || atomic_read(&dev_priv->mm.wedged)); ring->irq_put(ring); - } else if (wait_for(i915_seqno_passed(ring->get_seqno(ring), - seqno) || - atomic_read(&dev_priv->mm.wedged), 3000)) + } else if (wait_for_atomic(i915_seqno_passed(ring->get_seqno(ring), + seqno) || + atomic_read(&dev_priv->mm.wedged), 3000)) ret = -EBUSY; ring->waiting_seqno = 0; @@ -2051,17 +1989,12 @@ i915_wait_request(struct intel_ring_buffer *ring, if (atomic_read(&dev_priv->mm.wedged)) ret = -EAGAIN; - if (ret && ret != -ERESTARTSYS) - DRM_ERROR("%s returns %d (awaiting %d at %d, next %d)\n", - __func__, ret, seqno, ring->get_seqno(ring), - dev_priv->next_seqno); - /* Directly dispatch request retiring. While we have the work queue * to handle this, the waiter on a request often wants an associated * buffer to have made it to the inactive list, and we would need * a separate wait queue to handle that. */ - if (ret == 0) + if (ret == 0 && do_retire) i915_gem_retire_requests_ring(ring); return ret; @@ -2085,7 +2018,8 @@ i915_gem_object_wait_rendering(struct drm_i915_gem_object *obj) * it. */ if (obj->active) { - ret = i915_wait_request(obj->ring, obj->last_rendering_seqno); + ret = i915_wait_request(obj->ring, obj->last_rendering_seqno, + true); if (ret) return ret; } @@ -2123,6 +2057,7 @@ static void i915_gem_object_finish_gtt(struct drm_i915_gem_object *obj) int i915_gem_object_unbind(struct drm_i915_gem_object *obj) { + drm_i915_private_t *dev_priv = obj->base.dev->dev_private; int ret = 0; if (obj->gtt_space == NULL) @@ -2167,6 +2102,11 @@ i915_gem_object_unbind(struct drm_i915_gem_object *obj) trace_i915_gem_object_unbind(obj); i915_gem_gtt_unbind_object(obj); + if (obj->has_aliasing_ppgtt_mapping) { + i915_ppgtt_unbind_object(dev_priv->mm.aliasing_ppgtt, obj); + obj->has_aliasing_ppgtt_mapping = 0; + } + i915_gem_object_put_pages_gtt(obj); list_del_init(&obj->gtt_list); @@ -2206,7 +2146,7 @@ i915_gem_flush_ring(struct intel_ring_buffer *ring, return 0; } -static int i915_ring_idle(struct intel_ring_buffer *ring) +static int i915_ring_idle(struct intel_ring_buffer *ring, bool do_retire) { int ret; @@ -2220,18 +2160,18 @@ static int i915_ring_idle(struct intel_ring_buffer *ring) return ret; } - return i915_wait_request(ring, i915_gem_next_request_seqno(ring)); + return i915_wait_request(ring, i915_gem_next_request_seqno(ring), + do_retire); } -int -i915_gpu_idle(struct drm_device *dev) +int i915_gpu_idle(struct drm_device *dev, bool do_retire) { drm_i915_private_t *dev_priv = dev->dev_private; int ret, i; /* Flush everything onto the inactive list. */ for (i = 0; i < I915_NUM_RINGS; i++) { - ret = i915_ring_idle(&dev_priv->ring[i]); + ret = i915_ring_idle(&dev_priv->ring[i], do_retire); if (ret) return ret; } @@ -2434,7 +2374,8 @@ i915_gem_object_flush_fence(struct drm_i915_gem_object *obj, if (!ring_passed_seqno(obj->last_fenced_ring, obj->last_fenced_seqno)) { ret = i915_wait_request(obj->last_fenced_ring, - obj->last_fenced_seqno); + obj->last_fenced_seqno, + true); if (ret) return ret; } @@ -2466,6 +2407,8 @@ i915_gem_object_put_fence(struct drm_i915_gem_object *obj) if (obj->fence_reg != I915_FENCE_REG_NONE) { struct drm_i915_private *dev_priv = obj->base.dev->dev_private; + + WARN_ON(dev_priv->fence_regs[obj->fence_reg].pin_count); i915_gem_clear_fence_reg(obj->base.dev, &dev_priv->fence_regs[obj->fence_reg]); @@ -2490,7 +2433,7 @@ i915_find_fence_reg(struct drm_device *dev, if (!reg->obj) return reg; - if (!reg->obj->pin_count) + if (!reg->pin_count) avail = reg; } @@ -2500,7 +2443,7 @@ i915_find_fence_reg(struct drm_device *dev, /* None available, try to steal one or wait for a user to finish */ avail = first = NULL; list_for_each_entry(reg, &dev_priv->mm.fence_list, lru_list) { - if (reg->obj->pin_count) + if (reg->pin_count) continue; if (first == NULL) @@ -2575,7 +2518,8 @@ i915_gem_object_get_fence(struct drm_i915_gem_object *obj, if (!ring_passed_seqno(obj->last_fenced_ring, reg->setup_seqno)) { ret = i915_wait_request(obj->last_fenced_ring, - reg->setup_seqno); + reg->setup_seqno, + true); if (ret) return ret; } @@ -2594,7 +2538,7 @@ i915_gem_object_get_fence(struct drm_i915_gem_object *obj, reg = i915_find_fence_reg(dev, pipelined); if (reg == NULL) - return -ENOSPC; + return -EDEADLK; ret = i915_gem_object_flush_fence(obj, pipelined); if (ret) @@ -2694,6 +2638,7 @@ i915_gem_clear_fence_reg(struct drm_device *dev, list_del_init(®->lru_list); reg->obj = NULL; reg->setup_seqno = 0; + reg->pin_count = 0; } /** @@ -2980,6 +2925,8 @@ i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj, bool write) int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj, enum i915_cache_level cache_level) { + struct drm_device *dev = obj->base.dev; + drm_i915_private_t *dev_priv = dev->dev_private; int ret; if (obj->cache_level == cache_level) @@ -3008,6 +2955,9 @@ int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj, } i915_gem_gtt_rebind_object(obj, cache_level); + if (obj->has_aliasing_ppgtt_mapping) + i915_ppgtt_bind_object(dev_priv->mm.aliasing_ppgtt, + obj, cache_level); } if (cache_level == I915_CACHE_NONE) { @@ -3346,8 +3296,8 @@ i915_gem_ring_throttle(struct drm_device *dev, struct drm_file *file) if (ret == 0 && atomic_read(&dev_priv->mm.wedged)) ret = -EIO; - } else if (wait_for(i915_seqno_passed(ring->get_seqno(ring), - seqno) || + } else if (wait_for_atomic(i915_seqno_passed(ring->get_seqno(ring), + seqno) || atomic_read(&dev_priv->mm.wedged), 3000)) { ret = -EBUSY; } @@ -3456,15 +3406,14 @@ i915_gem_pin_ioctl(struct drm_device *dev, void *data, goto out; } - if (obj->user_pin_count == 0) { + obj->user_pin_count++; + obj->pin_filp = file; + if (obj->user_pin_count == 1) { ret = i915_gem_object_pin(obj, args->alignment, true); if (ret) goto out; } - obj->user_pin_count++; - obj->pin_filp = file; - /* XXX - flush the CPU caches for pinned objects * as the X server doesn't manage domains yet */ @@ -3658,8 +3607,8 @@ struct drm_i915_gem_object *i915_gem_alloc_object(struct drm_device *dev, obj->base.write_domain = I915_GEM_DOMAIN_CPU; obj->base.read_domains = I915_GEM_DOMAIN_CPU; - if (IS_GEN6(dev) || IS_GEN7(dev)) { - /* On Gen6, we can have the GPU use the LLC (the CPU + if (HAS_LLC(dev)) { + /* On some devices, we can have the GPU use the LLC (the CPU * cache) for about a 10% performance improvement * compared to uncached. Graphics requests other than * display scanout are coherent with the CPU in @@ -3749,7 +3698,7 @@ i915_gem_idle(struct drm_device *dev) return 0; } - ret = i915_gpu_idle(dev); + ret = i915_gpu_idle(dev, true); if (ret) { mutex_unlock(&dev->struct_mutex); return ret; @@ -3784,12 +3733,91 @@ i915_gem_idle(struct drm_device *dev) return 0; } +void i915_gem_init_swizzling(struct drm_device *dev) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + + if (INTEL_INFO(dev)->gen < 5 || + dev_priv->mm.bit_6_swizzle_x == I915_BIT_6_SWIZZLE_NONE) + return; + + I915_WRITE(DISP_ARB_CTL, I915_READ(DISP_ARB_CTL) | + DISP_TILE_SURFACE_SWIZZLING); + + if (IS_GEN5(dev)) + return; + + I915_WRITE(TILECTL, I915_READ(TILECTL) | TILECTL_SWZCTL); + if (IS_GEN6(dev)) + I915_WRITE(ARB_MODE, ARB_MODE_ENABLE(ARB_MODE_SWIZZLE_SNB)); + else + I915_WRITE(ARB_MODE, ARB_MODE_ENABLE(ARB_MODE_SWIZZLE_IVB)); +} + +void i915_gem_init_ppgtt(struct drm_device *dev) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + uint32_t pd_offset; + struct intel_ring_buffer *ring; + struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt; + uint32_t __iomem *pd_addr; + uint32_t pd_entry; + int i; + + if (!dev_priv->mm.aliasing_ppgtt) + return; + + + pd_addr = dev_priv->mm.gtt->gtt + ppgtt->pd_offset/sizeof(uint32_t); + for (i = 0; i < ppgtt->num_pd_entries; i++) { + dma_addr_t pt_addr; + + if (dev_priv->mm.gtt->needs_dmar) + pt_addr = ppgtt->pt_dma_addr[i]; + else + pt_addr = page_to_phys(ppgtt->pt_pages[i]); + + pd_entry = GEN6_PDE_ADDR_ENCODE(pt_addr); + pd_entry |= GEN6_PDE_VALID; + + writel(pd_entry, pd_addr + i); + } + readl(pd_addr); + + pd_offset = ppgtt->pd_offset; + pd_offset /= 64; /* in cachelines, */ + pd_offset <<= 16; + + if (INTEL_INFO(dev)->gen == 6) { + uint32_t ecochk = I915_READ(GAM_ECOCHK); + I915_WRITE(GAM_ECOCHK, ecochk | ECOCHK_SNB_BIT | + ECOCHK_PPGTT_CACHE64B); + I915_WRITE(GFX_MODE, GFX_MODE_ENABLE(GFX_PPGTT_ENABLE)); + } else if (INTEL_INFO(dev)->gen >= 7) { + I915_WRITE(GAM_ECOCHK, ECOCHK_PPGTT_CACHE64B); + /* GFX_MODE is per-ring on gen7+ */ + } + + for (i = 0; i < I915_NUM_RINGS; i++) { + ring = &dev_priv->ring[i]; + + if (INTEL_INFO(dev)->gen >= 7) + I915_WRITE(RING_MODE_GEN7(ring), + GFX_MODE_ENABLE(GFX_PPGTT_ENABLE)); + + I915_WRITE(RING_PP_DIR_DCLV(ring), PP_DIR_DCLV_2G); + I915_WRITE(RING_PP_DIR_BASE(ring), pd_offset); + } +} + int -i915_gem_init_ringbuffer(struct drm_device *dev) +i915_gem_init_hw(struct drm_device *dev) { drm_i915_private_t *dev_priv = dev->dev_private; int ret; + i915_gem_init_swizzling(dev); + ret = intel_init_render_ring_buffer(dev); if (ret) return ret; @@ -3808,6 +3836,8 @@ i915_gem_init_ringbuffer(struct drm_device *dev) dev_priv->next_seqno = 1; + i915_gem_init_ppgtt(dev); + return 0; cleanup_bsd_ring: @@ -3845,7 +3875,7 @@ i915_gem_entervt_ioctl(struct drm_device *dev, void *data, mutex_lock(&dev->struct_mutex); dev_priv->mm.suspended = 0; - ret = i915_gem_init_ringbuffer(dev); + ret = i915_gem_init_hw(dev); if (ret != 0) { mutex_unlock(&dev->struct_mutex); return ret; @@ -4240,7 +4270,7 @@ rescan: * This has a dramatic impact to reduce the number of * OOM-killer events whilst running the GPU aggressively. */ - if (i915_gpu_idle(dev) == 0) + if (i915_gpu_idle(dev, true) == 0) goto rescan; } mutex_unlock(&dev->struct_mutex); diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c index ead5d00..21a8271 100644 --- a/drivers/gpu/drm/i915/i915_gem_evict.c +++ b/drivers/gpu/drm/i915/i915_gem_evict.c @@ -36,7 +36,6 @@ static bool mark_free(struct drm_i915_gem_object *obj, struct list_head *unwind) { list_add(&obj->exec_list, unwind); - drm_gem_object_reference(&obj->base); return drm_mm_scan_add_block(obj->gtt_space); } @@ -49,21 +48,6 @@ i915_gem_evict_something(struct drm_device *dev, int min_size, struct drm_i915_gem_object *obj; int ret = 0; - i915_gem_retire_requests(dev); - - /* Re-check for free space after retiring requests */ - if (mappable) { - if (drm_mm_search_free_in_range(&dev_priv->mm.gtt_space, - min_size, alignment, 0, - dev_priv->mm.gtt_mappable_end, - 0)) - return 0; - } else { - if (drm_mm_search_free(&dev_priv->mm.gtt_space, - min_size, alignment, 0)) - return 0; - } - trace_i915_gem_evict(dev, min_size, alignment, mappable); /* @@ -139,7 +123,6 @@ i915_gem_evict_something(struct drm_device *dev, int min_size, BUG_ON(ret); list_del_init(&obj->exec_list); - drm_gem_object_unreference(&obj->base); } /* We expect the caller to unpin, evict all and try again, or give up. @@ -158,10 +141,10 @@ found: exec_list); if (drm_mm_scan_remove_block(obj->gtt_space)) { list_move(&obj->exec_list, &eviction_list); + drm_gem_object_reference(&obj->base); continue; } list_del_init(&obj->exec_list); - drm_gem_object_unreference(&obj->base); } /* Unbinding will emit any required flushes */ @@ -195,7 +178,7 @@ i915_gem_evict_everything(struct drm_device *dev, bool purgeable_only) trace_i915_gem_evict_everything(dev, purgeable_only); /* Flush everything (on to the inactive lists) and evict */ - ret = i915_gpu_idle(dev); + ret = i915_gpu_idle(dev, true); if (ret) return ret; diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c index 878b989..d4417e3 100644 --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c @@ -203,9 +203,9 @@ i915_gem_object_set_to_gpu_domain(struct drm_i915_gem_object *obj, cd->invalidate_domains |= invalidate_domains; cd->flush_domains |= flush_domains; if (flush_domains & I915_GEM_GPU_DOMAINS) - cd->flush_rings |= obj->ring->id; + cd->flush_rings |= intel_ring_flag(obj->ring); if (invalidate_domains & I915_GEM_GPU_DOMAINS) - cd->flush_rings |= ring->id; + cd->flush_rings |= intel_ring_flag(ring); } struct eb_objects { @@ -287,14 +287,14 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj, * exec_object list, so it should have a GTT space bound by now. */ if (unlikely(target_offset == 0)) { - DRM_ERROR("No GTT space found for object %d\n", + DRM_DEBUG("No GTT space found for object %d\n", reloc->target_handle); return ret; } /* Validate that the target is in a valid r/w GPU domain */ if (unlikely(reloc->write_domain & (reloc->write_domain - 1))) { - DRM_ERROR("reloc with multiple write domains: " + DRM_DEBUG("reloc with multiple write domains: " "obj %p target %d offset %d " "read %08x write %08x", obj, reloc->target_handle, @@ -303,8 +303,9 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj, reloc->write_domain); return ret; } - if (unlikely((reloc->write_domain | reloc->read_domains) & I915_GEM_DOMAIN_CPU)) { - DRM_ERROR("reloc with read/write CPU domains: " + if (unlikely((reloc->write_domain | reloc->read_domains) + & ~I915_GEM_GPU_DOMAINS)) { + DRM_DEBUG("reloc with read/write non-GPU domains: " "obj %p target %d offset %d " "read %08x write %08x", obj, reloc->target_handle, @@ -315,7 +316,7 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj, } if (unlikely(reloc->write_domain && target_obj->pending_write_domain && reloc->write_domain != target_obj->pending_write_domain)) { - DRM_ERROR("Write domain conflict: " + DRM_DEBUG("Write domain conflict: " "obj %p target %d offset %d " "new %08x old %08x\n", obj, reloc->target_handle, @@ -336,7 +337,7 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj, /* Check that the relocation address is valid... */ if (unlikely(reloc->offset > obj->base.size - 4)) { - DRM_ERROR("Relocation beyond object bounds: " + DRM_DEBUG("Relocation beyond object bounds: " "obj %p target %d offset %d size %d.\n", obj, reloc->target_handle, (int) reloc->offset, @@ -344,7 +345,7 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj, return ret; } if (unlikely(reloc->offset & 3)) { - DRM_ERROR("Relocation not 4-byte aligned: " + DRM_DEBUG("Relocation not 4-byte aligned: " "obj %p target %d offset %d.\n", obj, reloc->target_handle, (int) reloc->offset); @@ -461,11 +462,60 @@ i915_gem_execbuffer_relocate(struct drm_device *dev, return ret; } +#define __EXEC_OBJECT_HAS_FENCE (1<<31) + +static int +pin_and_fence_object(struct drm_i915_gem_object *obj, + struct intel_ring_buffer *ring) +{ + struct drm_i915_gem_exec_object2 *entry = obj->exec_entry; + bool has_fenced_gpu_access = INTEL_INFO(ring->dev)->gen < 4; + bool need_fence, need_mappable; + int ret; + + need_fence = + has_fenced_gpu_access && + entry->flags & EXEC_OBJECT_NEEDS_FENCE && + obj->tiling_mode != I915_TILING_NONE; + need_mappable = + entry->relocation_count ? true : need_fence; + + ret = i915_gem_object_pin(obj, entry->alignment, need_mappable); + if (ret) + return ret; + + if (has_fenced_gpu_access) { + if (entry->flags & EXEC_OBJECT_NEEDS_FENCE) { + if (obj->tiling_mode) { + ret = i915_gem_object_get_fence(obj, ring); + if (ret) + goto err_unpin; + + entry->flags |= __EXEC_OBJECT_HAS_FENCE; + i915_gem_object_pin_fence(obj); + } else { + ret = i915_gem_object_put_fence(obj); + if (ret) + goto err_unpin; + } + obj->pending_fenced_gpu_access = true; + } + } + + entry->offset = obj->gtt_offset; + return 0; + +err_unpin: + i915_gem_object_unpin(obj); + return ret; +} + static int i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring, struct drm_file *file, struct list_head *objects) { + drm_i915_private_t *dev_priv = ring->dev->dev_private; struct drm_i915_gem_object *obj; int ret, retry; bool has_fenced_gpu_access = INTEL_INFO(ring->dev)->gen < 4; @@ -518,6 +568,7 @@ i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring, list_for_each_entry(obj, objects, exec_list) { struct drm_i915_gem_exec_object2 *entry = obj->exec_entry; bool need_fence, need_mappable; + if (!obj->gtt_space) continue; @@ -532,59 +583,55 @@ i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring, (need_mappable && !obj->map_and_fenceable)) ret = i915_gem_object_unbind(obj); else - ret = i915_gem_object_pin(obj, - entry->alignment, - need_mappable); + ret = pin_and_fence_object(obj, ring); if (ret) goto err; - - entry++; } /* Bind fresh objects */ list_for_each_entry(obj, objects, exec_list) { - struct drm_i915_gem_exec_object2 *entry = obj->exec_entry; - bool need_fence; + if (obj->gtt_space) + continue; - need_fence = - has_fenced_gpu_access && - entry->flags & EXEC_OBJECT_NEEDS_FENCE && - obj->tiling_mode != I915_TILING_NONE; + ret = pin_and_fence_object(obj, ring); + if (ret) { + int ret_ignore; + + /* This can potentially raise a harmless + * -EINVAL if we failed to bind in the above + * call. It cannot raise -EINTR since we know + * that the bo is freshly bound and so will + * not need to be flushed or waited upon. + */ + ret_ignore = i915_gem_object_unbind(obj); + (void)ret_ignore; + WARN_ON(obj->gtt_space); + break; + } + } - if (!obj->gtt_space) { - bool need_mappable = - entry->relocation_count ? true : need_fence; + /* Decrement pin count for bound objects */ + list_for_each_entry(obj, objects, exec_list) { + struct drm_i915_gem_exec_object2 *entry; - ret = i915_gem_object_pin(obj, - entry->alignment, - need_mappable); - if (ret) - break; - } + if (!obj->gtt_space) + continue; - if (has_fenced_gpu_access) { - if (need_fence) { - ret = i915_gem_object_get_fence(obj, ring); - if (ret) - break; - } else if (entry->flags & EXEC_OBJECT_NEEDS_FENCE && - obj->tiling_mode == I915_TILING_NONE) { - /* XXX pipelined! */ - ret = i915_gem_object_put_fence(obj); - if (ret) - break; - } - obj->pending_fenced_gpu_access = - !!(entry->flags & EXEC_OBJECT_NEEDS_FENCE); + entry = obj->exec_entry; + if (entry->flags & __EXEC_OBJECT_HAS_FENCE) { + i915_gem_object_unpin_fence(obj); + entry->flags &= ~__EXEC_OBJECT_HAS_FENCE; } - entry->offset = obj->gtt_offset; - } + i915_gem_object_unpin(obj); - /* Decrement pin count for bound objects */ - list_for_each_entry(obj, objects, exec_list) { - if (obj->gtt_space) - i915_gem_object_unpin(obj); + /* ... and ensure ppgtt mapping exist if needed. */ + if (dev_priv->mm.aliasing_ppgtt && !obj->has_aliasing_ppgtt_mapping) { + i915_ppgtt_bind_object(dev_priv->mm.aliasing_ppgtt, + obj, obj->cache_level); + + obj->has_aliasing_ppgtt_mapping = 1; + } } if (ret != -ENOSPC || retry > 1) @@ -601,16 +648,19 @@ i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring, } while (1); err: - obj = list_entry(obj->exec_list.prev, - struct drm_i915_gem_object, - exec_list); - while (objects != &obj->exec_list) { - if (obj->gtt_space) - i915_gem_object_unpin(obj); + list_for_each_entry_continue_reverse(obj, objects, exec_list) { + struct drm_i915_gem_exec_object2 *entry; + + if (!obj->gtt_space) + continue; - obj = list_entry(obj->exec_list.prev, - struct drm_i915_gem_object, - exec_list); + entry = obj->exec_entry; + if (entry->flags & __EXEC_OBJECT_HAS_FENCE) { + i915_gem_object_unpin_fence(obj); + entry->flags &= ~__EXEC_OBJECT_HAS_FENCE; + } + + i915_gem_object_unpin(obj); } return ret; @@ -704,7 +754,7 @@ i915_gem_execbuffer_relocate_slow(struct drm_device *dev, obj = to_intel_bo(drm_gem_object_lookup(dev, file, exec[i].handle)); if (&obj->base == NULL) { - DRM_ERROR("Invalid object handle %d at index %d\n", + DRM_DEBUG("Invalid object handle %d at index %d\n", exec[i].handle, i); ret = -ENOENT; goto err; @@ -993,6 +1043,31 @@ i915_gem_execbuffer_retire_commands(struct drm_device *dev, } static int +i915_reset_gen7_sol_offsets(struct drm_device *dev, + struct intel_ring_buffer *ring) +{ + drm_i915_private_t *dev_priv = dev->dev_private; + int ret, i; + + if (!IS_GEN7(dev) || ring != &dev_priv->ring[RCS]) + return 0; + + ret = intel_ring_begin(ring, 4 * 3); + if (ret) + return ret; + + for (i = 0; i < 4; i++) { + intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1)); + intel_ring_emit(ring, GEN7_SO_WRITE_OFFSET(i)); + intel_ring_emit(ring, 0); + } + + intel_ring_advance(ring); + + return 0; +} + +static int i915_gem_do_execbuffer(struct drm_device *dev, void *data, struct drm_file *file, struct drm_i915_gem_execbuffer2 *args, @@ -1010,7 +1085,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, int ret, mode, i; if (!i915_gem_check_execbuffer(args)) { - DRM_ERROR("execbuf with invalid offset/length\n"); + DRM_DEBUG("execbuf with invalid offset/length\n"); return -EINVAL; } @@ -1025,20 +1100,20 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, break; case I915_EXEC_BSD: if (!HAS_BSD(dev)) { - DRM_ERROR("execbuf with invalid ring (BSD)\n"); + DRM_DEBUG("execbuf with invalid ring (BSD)\n"); return -EINVAL; } ring = &dev_priv->ring[VCS]; break; case I915_EXEC_BLT: if (!HAS_BLT(dev)) { - DRM_ERROR("execbuf with invalid ring (BLT)\n"); + DRM_DEBUG("execbuf with invalid ring (BLT)\n"); return -EINVAL; } ring = &dev_priv->ring[BCS]; break; default: - DRM_ERROR("execbuf with unknown ring: %d\n", + DRM_DEBUG("execbuf with unknown ring: %d\n", (int)(args->flags & I915_EXEC_RING_MASK)); return -EINVAL; } @@ -1064,18 +1139,18 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, } break; default: - DRM_ERROR("execbuf with unknown constants: %d\n", mode); + DRM_DEBUG("execbuf with unknown constants: %d\n", mode); return -EINVAL; } if (args->buffer_count < 1) { - DRM_ERROR("execbuf with %d buffers\n", args->buffer_count); + DRM_DEBUG("execbuf with %d buffers\n", args->buffer_count); return -EINVAL; } if (args->num_cliprects != 0) { if (ring != &dev_priv->ring[RCS]) { - DRM_ERROR("clip rectangles are only valid with the render ring\n"); + DRM_DEBUG("clip rectangles are only valid with the render ring\n"); return -EINVAL; } @@ -1125,7 +1200,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, obj = to_intel_bo(drm_gem_object_lookup(dev, file, exec[i].handle)); if (&obj->base == NULL) { - DRM_ERROR("Invalid object handle %d at index %d\n", + DRM_DEBUG("Invalid object handle %d at index %d\n", exec[i].handle, i); /* prevent error path from reading uninitialized data */ ret = -ENOENT; @@ -1133,7 +1208,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, } if (!list_empty(&obj->exec_list)) { - DRM_ERROR("Object %p [handle %d, index %d] appears more than once in object list\n", + DRM_DEBUG("Object %p [handle %d, index %d] appears more than once in object list\n", obj, exec[i].handle, i); ret = -EINVAL; goto err; @@ -1171,7 +1246,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, /* Set the pending read domains for the batch buffer to COMMAND */ if (batch_obj->base.pending_write_domain) { - DRM_ERROR("Attempting to use self-modifying batch buffer\n"); + DRM_DEBUG("Attempting to use self-modifying batch buffer\n"); ret = -EINVAL; goto err; } @@ -1188,7 +1263,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, * so every billion or so execbuffers, we need to stall * the GPU in order to reset the counters. */ - ret = i915_gpu_idle(dev); + ret = i915_gpu_idle(dev, true); if (ret) goto err; @@ -1211,6 +1286,12 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, dev_priv->relative_constants_mode = mode; } + if (args->flags & I915_EXEC_GEN7_SOL_RESET) { + ret = i915_reset_gen7_sol_offsets(dev, ring); + if (ret) + goto err; + } + trace_i915_gem_ring_dispatch(ring, seqno); exec_start = batch_obj->gtt_offset + args->batch_start_offset; @@ -1270,7 +1351,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data, int ret, i; if (args->buffer_count < 1) { - DRM_ERROR("execbuf with %d buffers\n", args->buffer_count); + DRM_DEBUG("execbuf with %d buffers\n", args->buffer_count); return -EINVAL; } @@ -1278,7 +1359,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data, exec_list = drm_malloc_ab(sizeof(*exec_list), args->buffer_count); exec2_list = drm_malloc_ab(sizeof(*exec2_list), args->buffer_count); if (exec_list == NULL || exec2_list == NULL) { - DRM_ERROR("Failed to allocate exec list for %d buffers\n", + DRM_DEBUG("Failed to allocate exec list for %d buffers\n", args->buffer_count); drm_free_large(exec_list); drm_free_large(exec2_list); @@ -1289,7 +1370,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data, (uintptr_t) args->buffers_ptr, sizeof(*exec_list) * args->buffer_count); if (ret != 0) { - DRM_ERROR("copy %d exec entries failed %d\n", + DRM_DEBUG("copy %d exec entries failed %d\n", args->buffer_count, ret); drm_free_large(exec_list); drm_free_large(exec2_list); @@ -1330,7 +1411,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data, sizeof(*exec_list) * args->buffer_count); if (ret) { ret = -EFAULT; - DRM_ERROR("failed to copy %d exec entries " + DRM_DEBUG("failed to copy %d exec entries " "back to user (%d)\n", args->buffer_count, ret); } @@ -1351,7 +1432,7 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data, if (args->buffer_count < 1 || args->buffer_count > UINT_MAX / sizeof(*exec2_list)) { - DRM_ERROR("execbuf2 with %d buffers\n", args->buffer_count); + DRM_DEBUG("execbuf2 with %d buffers\n", args->buffer_count); return -EINVAL; } @@ -1361,7 +1442,7 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data, exec2_list = drm_malloc_ab(sizeof(*exec2_list), args->buffer_count); if (exec2_list == NULL) { - DRM_ERROR("Failed to allocate exec list for %d buffers\n", + DRM_DEBUG("Failed to allocate exec list for %d buffers\n", args->buffer_count); return -ENOMEM; } @@ -1370,7 +1451,7 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data, (uintptr_t) args->buffers_ptr, sizeof(*exec2_list) * args->buffer_count); if (ret != 0) { - DRM_ERROR("copy %d exec entries failed %d\n", + DRM_DEBUG("copy %d exec entries failed %d\n", args->buffer_count, ret); drm_free_large(exec2_list); return -EFAULT; @@ -1385,7 +1466,7 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data, sizeof(*exec2_list) * args->buffer_count); if (ret) { ret = -EFAULT; - DRM_ERROR("failed to copy %d exec entries " + DRM_DEBUG("failed to copy %d exec entries " "back to user (%d)\n", args->buffer_count, ret); } diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c index 6042c5e..a135c61 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.c +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c @@ -29,6 +29,270 @@ #include "i915_trace.h" #include "intel_drv.h" +/* PPGTT support for Sandybdrige/Gen6 and later */ +static void i915_ppgtt_clear_range(struct i915_hw_ppgtt *ppgtt, + unsigned first_entry, + unsigned num_entries) +{ + uint32_t *pt_vaddr; + uint32_t scratch_pte; + unsigned act_pd = first_entry / I915_PPGTT_PT_ENTRIES; + unsigned first_pte = first_entry % I915_PPGTT_PT_ENTRIES; + unsigned last_pte, i; + + scratch_pte = GEN6_PTE_ADDR_ENCODE(ppgtt->scratch_page_dma_addr); + scratch_pte |= GEN6_PTE_VALID | GEN6_PTE_CACHE_LLC; + + while (num_entries) { + last_pte = first_pte + num_entries; + if (last_pte > I915_PPGTT_PT_ENTRIES) + last_pte = I915_PPGTT_PT_ENTRIES; + + pt_vaddr = kmap_atomic(ppgtt->pt_pages[act_pd]); + + for (i = first_pte; i < last_pte; i++) + pt_vaddr[i] = scratch_pte; + + kunmap_atomic(pt_vaddr); + + num_entries -= last_pte - first_pte; + first_pte = 0; + act_pd++; + } +} + +int i915_gem_init_aliasing_ppgtt(struct drm_device *dev) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + struct i915_hw_ppgtt *ppgtt; + unsigned first_pd_entry_in_global_pt; + int i; + int ret = -ENOMEM; + + /* ppgtt PDEs reside in the global gtt pagetable, which has 512*1024 + * entries. For aliasing ppgtt support we just steal them at the end for + * now. */ + first_pd_entry_in_global_pt = 512*1024 - I915_PPGTT_PD_ENTRIES; + + ppgtt = kzalloc(sizeof(*ppgtt), GFP_KERNEL); + if (!ppgtt) + return ret; + + ppgtt->num_pd_entries = I915_PPGTT_PD_ENTRIES; + ppgtt->pt_pages = kzalloc(sizeof(struct page *)*ppgtt->num_pd_entries, + GFP_KERNEL); + if (!ppgtt->pt_pages) + goto err_ppgtt; + + for (i = 0; i < ppgtt->num_pd_entries; i++) { + ppgtt->pt_pages[i] = alloc_page(GFP_KERNEL); + if (!ppgtt->pt_pages[i]) + goto err_pt_alloc; + } + + if (dev_priv->mm.gtt->needs_dmar) { + ppgtt->pt_dma_addr = kzalloc(sizeof(dma_addr_t) + *ppgtt->num_pd_entries, + GFP_KERNEL); + if (!ppgtt->pt_dma_addr) + goto err_pt_alloc; + } + + for (i = 0; i < ppgtt->num_pd_entries; i++) { + dma_addr_t pt_addr; + if (dev_priv->mm.gtt->needs_dmar) { + pt_addr = pci_map_page(dev->pdev, ppgtt->pt_pages[i], + 0, 4096, + PCI_DMA_BIDIRECTIONAL); + + if (pci_dma_mapping_error(dev->pdev, + pt_addr)) { + ret = -EIO; + goto err_pd_pin; + + } + ppgtt->pt_dma_addr[i] = pt_addr; + } else + pt_addr = page_to_phys(ppgtt->pt_pages[i]); + } + + ppgtt->scratch_page_dma_addr = dev_priv->mm.gtt->scratch_page_dma; + + i915_ppgtt_clear_range(ppgtt, 0, + ppgtt->num_pd_entries*I915_PPGTT_PT_ENTRIES); + + ppgtt->pd_offset = (first_pd_entry_in_global_pt)*sizeof(uint32_t); + + dev_priv->mm.aliasing_ppgtt = ppgtt; + + return 0; + +err_pd_pin: + if (ppgtt->pt_dma_addr) { + for (i--; i >= 0; i--) + pci_unmap_page(dev->pdev, ppgtt->pt_dma_addr[i], + 4096, PCI_DMA_BIDIRECTIONAL); + } +err_pt_alloc: + kfree(ppgtt->pt_dma_addr); + for (i = 0; i < ppgtt->num_pd_entries; i++) { + if (ppgtt->pt_pages[i]) + __free_page(ppgtt->pt_pages[i]); + } + kfree(ppgtt->pt_pages); +err_ppgtt: + kfree(ppgtt); + + return ret; +} + +void i915_gem_cleanup_aliasing_ppgtt(struct drm_device *dev) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt; + int i; + + if (!ppgtt) + return; + + if (ppgtt->pt_dma_addr) { + for (i = 0; i < ppgtt->num_pd_entries; i++) + pci_unmap_page(dev->pdev, ppgtt->pt_dma_addr[i], + 4096, PCI_DMA_BIDIRECTIONAL); + } + + kfree(ppgtt->pt_dma_addr); + for (i = 0; i < ppgtt->num_pd_entries; i++) + __free_page(ppgtt->pt_pages[i]); + kfree(ppgtt->pt_pages); + kfree(ppgtt); +} + +static void i915_ppgtt_insert_sg_entries(struct i915_hw_ppgtt *ppgtt, + struct scatterlist *sg_list, + unsigned sg_len, + unsigned first_entry, + uint32_t pte_flags) +{ + uint32_t *pt_vaddr, pte; + unsigned act_pd = first_entry / I915_PPGTT_PT_ENTRIES; + unsigned first_pte = first_entry % I915_PPGTT_PT_ENTRIES; + unsigned i, j, m, segment_len; + dma_addr_t page_addr; + struct scatterlist *sg; + + /* init sg walking */ + sg = sg_list; + i = 0; + segment_len = sg_dma_len(sg) >> PAGE_SHIFT; + m = 0; + + while (i < sg_len) { + pt_vaddr = kmap_atomic(ppgtt->pt_pages[act_pd]); + + for (j = first_pte; j < I915_PPGTT_PT_ENTRIES; j++) { + page_addr = sg_dma_address(sg) + (m << PAGE_SHIFT); + pte = GEN6_PTE_ADDR_ENCODE(page_addr); + pt_vaddr[j] = pte | pte_flags; + + /* grab the next page */ + m++; + if (m == segment_len) { + sg = sg_next(sg); + i++; + if (i == sg_len) + break; + + segment_len = sg_dma_len(sg) >> PAGE_SHIFT; + m = 0; + } + } + + kunmap_atomic(pt_vaddr); + + first_pte = 0; + act_pd++; + } +} + +static void i915_ppgtt_insert_pages(struct i915_hw_ppgtt *ppgtt, + unsigned first_entry, unsigned num_entries, + struct page **pages, uint32_t pte_flags) +{ + uint32_t *pt_vaddr, pte; + unsigned act_pd = first_entry / I915_PPGTT_PT_ENTRIES; + unsigned first_pte = first_entry % I915_PPGTT_PT_ENTRIES; + unsigned last_pte, i; + dma_addr_t page_addr; + + while (num_entries) { + last_pte = first_pte + num_entries; + last_pte = min_t(unsigned, last_pte, I915_PPGTT_PT_ENTRIES); + + pt_vaddr = kmap_atomic(ppgtt->pt_pages[act_pd]); + + for (i = first_pte; i < last_pte; i++) { + page_addr = page_to_phys(*pages); + pte = GEN6_PTE_ADDR_ENCODE(page_addr); + pt_vaddr[i] = pte | pte_flags; + + pages++; + } + + kunmap_atomic(pt_vaddr); + + num_entries -= last_pte - first_pte; + first_pte = 0; + act_pd++; + } +} + +void i915_ppgtt_bind_object(struct i915_hw_ppgtt *ppgtt, + struct drm_i915_gem_object *obj, + enum i915_cache_level cache_level) +{ + struct drm_device *dev = obj->base.dev; + struct drm_i915_private *dev_priv = dev->dev_private; + uint32_t pte_flags = GEN6_PTE_VALID; + + switch (cache_level) { + case I915_CACHE_LLC_MLC: + pte_flags |= GEN6_PTE_CACHE_LLC_MLC; + break; + case I915_CACHE_LLC: + pte_flags |= GEN6_PTE_CACHE_LLC; + break; + case I915_CACHE_NONE: + pte_flags |= GEN6_PTE_UNCACHED; + break; + default: + BUG(); + } + + if (dev_priv->mm.gtt->needs_dmar) { + BUG_ON(!obj->sg_list); + + i915_ppgtt_insert_sg_entries(ppgtt, + obj->sg_list, + obj->num_sg, + obj->gtt_space->start >> PAGE_SHIFT, + pte_flags); + } else + i915_ppgtt_insert_pages(ppgtt, + obj->gtt_space->start >> PAGE_SHIFT, + obj->base.size >> PAGE_SHIFT, + obj->pages, + pte_flags); +} + +void i915_ppgtt_unbind_object(struct i915_hw_ppgtt *ppgtt, + struct drm_i915_gem_object *obj) +{ + i915_ppgtt_clear_range(ppgtt, + obj->gtt_space->start >> PAGE_SHIFT, + obj->base.size >> PAGE_SHIFT); +} + /* XXX kill agp_type! */ static unsigned int cache_level_to_agp_type(struct drm_device *dev, enum i915_cache_level cache_level) @@ -55,7 +319,7 @@ static bool do_idling(struct drm_i915_private *dev_priv) if (unlikely(dev_priv->mm.gtt->do_idle_maps)) { dev_priv->mm.interruptible = false; - if (i915_gpu_idle(dev_priv->dev)) { + if (i915_gpu_idle(dev_priv->dev, false)) { DRM_ERROR("Couldn't idle GPU\n"); /* Wait a bit, in hopes it avoids the hang */ udelay(10); diff --git a/drivers/gpu/drm/i915/i915_gem_tiling.c b/drivers/gpu/drm/i915/i915_gem_tiling.c index 861223b..1a93066 100644 --- a/drivers/gpu/drm/i915/i915_gem_tiling.c +++ b/drivers/gpu/drm/i915/i915_gem_tiling.c @@ -93,8 +93,23 @@ i915_gem_detect_bit_6_swizzle(struct drm_device *dev) uint32_t swizzle_y = I915_BIT_6_SWIZZLE_UNKNOWN; if (INTEL_INFO(dev)->gen >= 6) { - swizzle_x = I915_BIT_6_SWIZZLE_NONE; - swizzle_y = I915_BIT_6_SWIZZLE_NONE; + uint32_t dimm_c0, dimm_c1; + dimm_c0 = I915_READ(MAD_DIMM_C0); + dimm_c1 = I915_READ(MAD_DIMM_C1); + dimm_c0 &= MAD_DIMM_A_SIZE_MASK | MAD_DIMM_B_SIZE_MASK; + dimm_c1 &= MAD_DIMM_A_SIZE_MASK | MAD_DIMM_B_SIZE_MASK; + /* Enable swizzling when the channels are populated with + * identically sized dimms. We don't need to check the 3rd + * channel because no cpu with gpu attached ships in that + * configuration. Also, swizzling only makes sense for 2 + * channels anyway. */ + if (dimm_c0 == dimm_c1) { + swizzle_x = I915_BIT_6_SWIZZLE_9_10; + swizzle_y = I915_BIT_6_SWIZZLE_9; + } else { + swizzle_x = I915_BIT_6_SWIZZLE_NONE; + swizzle_y = I915_BIT_6_SWIZZLE_NONE; + } } else if (IS_GEN5(dev)) { /* On Ironlake whatever DRAM config, GPU always do * same swizzling setup. diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index 93e74fb..8bca2d2 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -716,7 +716,6 @@ i915_error_object_create(struct drm_i915_private *dev_priv, reloc_offset = src->gtt_offset; for (page = 0; page < page_count; page++) { unsigned long flags; - void __iomem *s; void *d; d = kmalloc(PAGE_SIZE, GFP_ATOMIC); @@ -724,10 +723,29 @@ i915_error_object_create(struct drm_i915_private *dev_priv, goto unwind; local_irq_save(flags); - s = io_mapping_map_atomic_wc(dev_priv->mm.gtt_mapping, - reloc_offset); - memcpy_fromio(d, s, PAGE_SIZE); - io_mapping_unmap_atomic(s); + if (reloc_offset < dev_priv->mm.gtt_mappable_end) { + void __iomem *s; + + /* Simply ignore tiling or any overlapping fence. + * It's part of the error state, and this hopefully + * captures what the GPU read. + */ + + s = io_mapping_map_atomic_wc(dev_priv->mm.gtt_mapping, + reloc_offset); + memcpy_fromio(d, s, PAGE_SIZE); + io_mapping_unmap_atomic(s); + } else { + void *s; + + drm_clflush_pages(&src->pages[page], 1); + + s = kmap_atomic(src->pages[page]); + memcpy(d, s, PAGE_SIZE); + kunmap_atomic(s); + + drm_clflush_pages(&src->pages[page], 1); + } local_irq_restore(flags); dst->pages[page] = d; @@ -766,11 +784,11 @@ i915_error_state_free(struct drm_device *dev, { int i; - for (i = 0; i < ARRAY_SIZE(error->batchbuffer); i++) - i915_error_object_free(error->batchbuffer[i]); - - for (i = 0; i < ARRAY_SIZE(error->ringbuffer); i++) - i915_error_object_free(error->ringbuffer[i]); + for (i = 0; i < ARRAY_SIZE(error->ring); i++) { + i915_error_object_free(error->ring[i].batchbuffer); + i915_error_object_free(error->ring[i].ringbuffer); + kfree(error->ring[i].requests); + } kfree(error->active_bo); kfree(error->overlay); @@ -800,7 +818,7 @@ static u32 capture_bo_list(struct drm_i915_error_buffer *err, err->tiling = obj->tiling_mode; err->dirty = obj->dirty; err->purgeable = obj->madv != I915_MADV_WILLNEED; - err->ring = obj->ring ? obj->ring->id : 0; + err->ring = obj->ring ? obj->ring->id : -1; err->cache_level = obj->cache_level; if (++i == count) @@ -872,6 +890,92 @@ i915_error_first_batchbuffer(struct drm_i915_private *dev_priv, return NULL; } +static void i915_record_ring_state(struct drm_device *dev, + struct drm_i915_error_state *error, + struct intel_ring_buffer *ring) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + + if (INTEL_INFO(dev)->gen >= 6) { + error->faddr[ring->id] = I915_READ(RING_DMA_FADD(ring->mmio_base)); + error->fault_reg[ring->id] = I915_READ(RING_FAULT_REG(ring)); + error->semaphore_mboxes[ring->id][0] + = I915_READ(RING_SYNC_0(ring->mmio_base)); + error->semaphore_mboxes[ring->id][1] + = I915_READ(RING_SYNC_1(ring->mmio_base)); + } + + if (INTEL_INFO(dev)->gen >= 4) { + error->ipeir[ring->id] = I915_READ(RING_IPEIR(ring->mmio_base)); + error->ipehr[ring->id] = I915_READ(RING_IPEHR(ring->mmio_base)); + error->instdone[ring->id] = I915_READ(RING_INSTDONE(ring->mmio_base)); + error->instps[ring->id] = I915_READ(RING_INSTPS(ring->mmio_base)); + if (ring->id == RCS) { + error->instdone1 = I915_READ(INSTDONE1); + error->bbaddr = I915_READ64(BB_ADDR); + } + } else { + error->ipeir[ring->id] = I915_READ(IPEIR); + error->ipehr[ring->id] = I915_READ(IPEHR); + error->instdone[ring->id] = I915_READ(INSTDONE); + } + + error->instpm[ring->id] = I915_READ(RING_INSTPM(ring->mmio_base)); + error->seqno[ring->id] = ring->get_seqno(ring); + error->acthd[ring->id] = intel_ring_get_active_head(ring); + error->head[ring->id] = I915_READ_HEAD(ring); + error->tail[ring->id] = I915_READ_TAIL(ring); + + error->cpu_ring_head[ring->id] = ring->head; + error->cpu_ring_tail[ring->id] = ring->tail; +} + +static void i915_gem_record_rings(struct drm_device *dev, + struct drm_i915_error_state *error) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + struct drm_i915_gem_request *request; + int i, count; + + for (i = 0; i < I915_NUM_RINGS; i++) { + struct intel_ring_buffer *ring = &dev_priv->ring[i]; + + if (ring->obj == NULL) + continue; + + i915_record_ring_state(dev, error, ring); + + error->ring[i].batchbuffer = + i915_error_first_batchbuffer(dev_priv, ring); + + error->ring[i].ringbuffer = + i915_error_object_create(dev_priv, ring->obj); + + count = 0; + list_for_each_entry(request, &ring->request_list, list) + count++; + + error->ring[i].num_requests = count; + error->ring[i].requests = + kmalloc(count*sizeof(struct drm_i915_error_request), + GFP_ATOMIC); + if (error->ring[i].requests == NULL) { + error->ring[i].num_requests = 0; + continue; + } + + count = 0; + list_for_each_entry(request, &ring->request_list, list) { + struct drm_i915_error_request *erq; + + erq = &error->ring[i].requests[count++]; + erq->seqno = request->seqno; + erq->jiffies = request->emitted_jiffies; + erq->tail = request->tail; + } + } +} + /** * i915_capture_error_state - capture an error record for later analysis * @dev: drm device @@ -896,7 +1000,7 @@ static void i915_capture_error_state(struct drm_device *dev) return; /* Account for pipe specific data like PIPE*STAT */ - error = kmalloc(sizeof(*error), GFP_ATOMIC); + error = kzalloc(sizeof(*error), GFP_ATOMIC); if (!error) { DRM_DEBUG_DRIVER("out of memory, not capturing error state\n"); return; @@ -905,59 +1009,18 @@ static void i915_capture_error_state(struct drm_device *dev) DRM_INFO("capturing error event; look for more information in /debug/dri/%d/i915_error_state\n", dev->primary->index); - error->seqno = dev_priv->ring[RCS].get_seqno(&dev_priv->ring[RCS]); error->eir = I915_READ(EIR); error->pgtbl_er = I915_READ(PGTBL_ER); for_each_pipe(pipe) error->pipestat[pipe] = I915_READ(PIPESTAT(pipe)); - error->instpm = I915_READ(INSTPM); - error->error = 0; + if (INTEL_INFO(dev)->gen >= 6) { error->error = I915_READ(ERROR_GEN6); - - error->bcs_acthd = I915_READ(BCS_ACTHD); - error->bcs_ipehr = I915_READ(BCS_IPEHR); - error->bcs_ipeir = I915_READ(BCS_IPEIR); - error->bcs_instdone = I915_READ(BCS_INSTDONE); - error->bcs_seqno = 0; - if (dev_priv->ring[BCS].get_seqno) - error->bcs_seqno = dev_priv->ring[BCS].get_seqno(&dev_priv->ring[BCS]); - - error->vcs_acthd = I915_READ(VCS_ACTHD); - error->vcs_ipehr = I915_READ(VCS_IPEHR); - error->vcs_ipeir = I915_READ(VCS_IPEIR); - error->vcs_instdone = I915_READ(VCS_INSTDONE); - error->vcs_seqno = 0; - if (dev_priv->ring[VCS].get_seqno) - error->vcs_seqno = dev_priv->ring[VCS].get_seqno(&dev_priv->ring[VCS]); + error->done_reg = I915_READ(DONE_REG); } - if (INTEL_INFO(dev)->gen >= 4) { - error->ipeir = I915_READ(IPEIR_I965); - error->ipehr = I915_READ(IPEHR_I965); - error->instdone = I915_READ(INSTDONE_I965); - error->instps = I915_READ(INSTPS); - error->instdone1 = I915_READ(INSTDONE1); - error->acthd = I915_READ(ACTHD_I965); - error->bbaddr = I915_READ64(BB_ADDR); - } else { - error->ipeir = I915_READ(IPEIR); - error->ipehr = I915_READ(IPEHR); - error->instdone = I915_READ(INSTDONE); - error->acthd = I915_READ(ACTHD); - error->bbaddr = 0; - } - i915_gem_record_fences(dev, error); - /* Record the active batch and ring buffers */ - for (i = 0; i < I915_NUM_RINGS; i++) { - error->batchbuffer[i] = - i915_error_first_batchbuffer(dev_priv, - &dev_priv->ring[i]); - - error->ringbuffer[i] = - i915_error_object_create(dev_priv, - dev_priv->ring[i].obj); - } + i915_gem_record_fences(dev, error); + i915_gem_record_rings(dev, error); /* Record buffers on the active and pinned lists. */ error->active_bo = NULL; @@ -1013,11 +1076,12 @@ void i915_destroy_error_state(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; struct drm_i915_error_state *error; + unsigned long flags; - spin_lock(&dev_priv->error_lock); + spin_lock_irqsave(&dev_priv->error_lock, flags); error = dev_priv->first_error; dev_priv->first_error = NULL; - spin_unlock(&dev_priv->error_lock); + spin_unlock_irqrestore(&dev_priv->error_lock, flags); if (error) i915_error_state_free(dev, error); @@ -1187,9 +1251,7 @@ static void i915_pageflip_stall_check(struct drm_device *dev, int pipe) spin_lock_irqsave(&dev->event_lock, flags); work = intel_crtc->unpin_work; - if (work == NULL || - atomic_read(&work->pending) >= INTEL_FLIP_COMPLETE || - !work->enable_stall_check) { + if (work == NULL || work->pending || !work->enable_stall_check) { /* Either the pending flip IRQ arrived, or we're too early. Don't check */ spin_unlock_irqrestore(&dev->event_lock, flags); return; @@ -1203,7 +1265,7 @@ static void i915_pageflip_stall_check(struct drm_device *dev, int pipe) } else { int dspaddr = DSPADDR(intel_crtc->plane); stall_detected = I915_READ(dspaddr) == (obj->gtt_offset + - crtc->y * crtc->fb->pitch + + crtc->y * crtc->fb->pitches[0] + crtc->x * crtc->fb->bits_per_pixel/8); } @@ -1647,13 +1709,6 @@ static bool kick_ring(struct intel_ring_buffer *ring) I915_WRITE_CTL(ring, tmp); return true; } - if (IS_GEN6(dev) && - (tmp & RING_WAIT_SEMAPHORE)) { - DRM_ERROR("Kicking stuck semaphore on %s\n", - ring->name); - I915_WRITE_CTL(ring, tmp); - return true; - } return false; } @@ -1703,6 +1758,7 @@ void i915_hangcheck_elapsed(unsigned long data) dev_priv->last_instdone1 == instdone1) { if (dev_priv->hangcheck_count++ > 1) { DRM_ERROR("Hangcheck timer elapsed... GPU hung\n"); + i915_handle_error(dev, true); if (!IS_GEN2(dev)) { /* Is the chip hanging on a WAIT_FOR_EVENT? @@ -1710,7 +1766,6 @@ void i915_hangcheck_elapsed(unsigned long data) * and break the hang. This should work on * all but the second generation chipsets. */ - if (kick_ring(&dev_priv->ring[RCS])) goto repeat; @@ -1723,7 +1778,6 @@ void i915_hangcheck_elapsed(unsigned long data) goto repeat; } - i915_handle_error(dev, true); return; } } else { @@ -1756,17 +1810,6 @@ static void ironlake_irq_preinstall(struct drm_device *dev) INIT_WORK(&dev_priv->rps_work, gen6_pm_rps_work); I915_WRITE(HWSTAM, 0xeffe); - if (IS_GEN6(dev) || IS_GEN7(dev)) { - /* Workaround stalls observed on Sandy Bridge GPUs by - * making the blitter command streamer generate a - * write to the Hardware Status Page for - * MI_USER_INTERRUPT. This appears to serialize the - * previous seqno write out before the interrupt - * happens. - */ - I915_WRITE(GEN6_BLITTER_HWSTAM, ~GEN6_BLITTER_USER_INTERRUPT); - I915_WRITE(GEN6_BSD_HWSTAM, ~GEN6_BSD_USER_INTERRUPT); - } /* XXX hotplug from PCH */ diff --git a/drivers/gpu/drm/i915/i915_mem.c b/drivers/gpu/drm/i915/i915_mem.c deleted file mode 100644 index cc8f6d4..0000000 --- a/drivers/gpu/drm/i915/i915_mem.c +++ /dev/null @@ -1,387 +0,0 @@ -/* i915_mem.c -- Simple agp/fb memory manager for i915 -*- linux-c -*- - */ -/* - * Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - */ - -#include "drmP.h" -#include "drm.h" -#include "i915_drm.h" -#include "i915_drv.h" - -/* This memory manager is integrated into the global/local lru - * mechanisms used by the clients. Specifically, it operates by - * setting the 'in_use' fields of the global LRU to indicate whether - * this region is privately allocated to a client. - * - * This does require the client to actually respect that field. - * - * Currently no effort is made to allocate 'private' memory in any - * clever way - the LRU information isn't used to determine which - * block to allocate, and the ring is drained prior to allocations -- - * in other words allocation is expensive. - */ -static void mark_block(struct drm_device * dev, struct mem_block *p, int in_use) -{ - drm_i915_private_t *dev_priv = dev->dev_private; - struct drm_i915_master_private *master_priv = dev->primary->master->driver_priv; - drm_i915_sarea_t *sarea_priv = master_priv->sarea_priv; - struct drm_tex_region *list; - unsigned shift, nr; - unsigned start; - unsigned end; - unsigned i; - int age; - - shift = dev_priv->tex_lru_log_granularity; - nr = I915_NR_TEX_REGIONS; - - start = p->start >> shift; - end = (p->start + p->size - 1) >> shift; - - age = ++sarea_priv->texAge; - list = sarea_priv->texList; - - /* Mark the regions with the new flag and update their age. Move - * them to head of list to preserve LRU semantics. - */ - for (i = start; i <= end; i++) { - list[i].in_use = in_use; - list[i].age = age; - - /* remove_from_list(i) - */ - list[(unsigned)list[i].next].prev = list[i].prev; - list[(unsigned)list[i].prev].next = list[i].next; - - /* insert_at_head(list, i) - */ - list[i].prev = nr; - list[i].next = list[nr].next; - list[(unsigned)list[nr].next].prev = i; - list[nr].next = i; - } -} - -/* Very simple allocator for agp memory, working on a static range - * already mapped into each client's address space. - */ - -static struct mem_block *split_block(struct mem_block *p, int start, int size, - struct drm_file *file_priv) -{ - /* Maybe cut off the start of an existing block */ - if (start > p->start) { - struct mem_block *newblock = kmalloc(sizeof(*newblock), - GFP_KERNEL); - if (!newblock) - goto out; - newblock->start = start; - newblock->size = p->size - (start - p->start); - newblock->file_priv = NULL; - newblock->next = p->next; - newblock->prev = p; - p->next->prev = newblock; - p->next = newblock; - p->size -= newblock->size; - p = newblock; - } - - /* Maybe cut off the end of an existing block */ - if (size < p->size) { - struct mem_block *newblock = kmalloc(sizeof(*newblock), - GFP_KERNEL); - if (!newblock) - goto out; - newblock->start = start + size; - newblock->size = p->size - size; - newblock->file_priv = NULL; - newblock->next = p->next; - newblock->prev = p; - p->next->prev = newblock; - p->next = newblock; - p->size = size; - } - - out: - /* Our block is in the middle */ - p->file_priv = file_priv; - return p; -} - -static struct mem_block *alloc_block(struct mem_block *heap, int size, - int align2, struct drm_file *file_priv) -{ - struct mem_block *p; - int mask = (1 << align2) - 1; - - for (p = heap->next; p != heap; p = p->next) { - int start = (p->start + mask) & ~mask; - if (p->file_priv == NULL && start + size <= p->start + p->size) - return split_block(p, start, size, file_priv); - } - - return NULL; -} - -static struct mem_block *find_block(struct mem_block *heap, int start) -{ - struct mem_block *p; - - for (p = heap->next; p != heap; p = p->next) - if (p->start == start) - return p; - - return NULL; -} - -static void free_block(struct mem_block *p) -{ - p->file_priv = NULL; - - /* Assumes a single contiguous range. Needs a special file_priv in - * 'heap' to stop it being subsumed. - */ - if (p->next->file_priv == NULL) { - struct mem_block *q = p->next; - p->size += q->size; - p->next = q->next; - p->next->prev = p; - kfree(q); - } - - if (p->prev->file_priv == NULL) { - struct mem_block *q = p->prev; - q->size += p->size; - q->next = p->next; - q->next->prev = q; - kfree(p); - } -} - -/* Initialize. How to check for an uninitialized heap? - */ -static int init_heap(struct mem_block **heap, int start, int size) -{ - struct mem_block *blocks = kmalloc(sizeof(*blocks), GFP_KERNEL); - - if (!blocks) - return -ENOMEM; - - *heap = kmalloc(sizeof(**heap), GFP_KERNEL); - if (!*heap) { - kfree(blocks); - return -ENOMEM; - } - - blocks->start = start; - blocks->size = size; - blocks->file_priv = NULL; - blocks->next = blocks->prev = *heap; - - memset(*heap, 0, sizeof(**heap)); - (*heap)->file_priv = (struct drm_file *) -1; - (*heap)->next = (*heap)->prev = blocks; - return 0; -} - -/* Free all blocks associated with the releasing file. - */ -void i915_mem_release(struct drm_device * dev, struct drm_file *file_priv, - struct mem_block *heap) -{ - struct mem_block *p; - - if (!heap || !heap->next) - return; - - for (p = heap->next; p != heap; p = p->next) { - if (p->file_priv == file_priv) { - p->file_priv = NULL; - mark_block(dev, p, 0); - } - } - - /* Assumes a single contiguous range. Needs a special file_priv in - * 'heap' to stop it being subsumed. - */ - for (p = heap->next; p != heap; p = p->next) { - while (p->file_priv == NULL && p->next->file_priv == NULL) { - struct mem_block *q = p->next; - p->size += q->size; - p->next = q->next; - p->next->prev = p; - kfree(q); - } - } -} - -/* Shutdown. - */ -void i915_mem_takedown(struct mem_block **heap) -{ - struct mem_block *p; - - if (!*heap) - return; - - for (p = (*heap)->next; p != *heap;) { - struct mem_block *q = p; - p = p->next; - kfree(q); - } - - kfree(*heap); - *heap = NULL; -} - -static struct mem_block **get_heap(drm_i915_private_t * dev_priv, int region) -{ - switch (region) { - case I915_MEM_REGION_AGP: - return &dev_priv->agp_heap; - default: - return NULL; - } -} - -/* IOCTL HANDLERS */ - -int i915_mem_alloc(struct drm_device *dev, void *data, - struct drm_file *file_priv) -{ - drm_i915_private_t *dev_priv = dev->dev_private; - drm_i915_mem_alloc_t *alloc = data; - struct mem_block *block, **heap; - - if (!dev_priv) { - DRM_ERROR("called with no initialization\n"); - return -EINVAL; - } - - heap = get_heap(dev_priv, alloc->region); - if (!heap || !*heap) - return -EFAULT; - - /* Make things easier on ourselves: all allocations at least - * 4k aligned. - */ - if (alloc->alignment < 12) - alloc->alignment = 12; - - block = alloc_block(*heap, alloc->size, alloc->alignment, file_priv); - - if (!block) - return -ENOMEM; - - mark_block(dev, block, 1); - - if (DRM_COPY_TO_USER(alloc->region_offset, &block->start, - sizeof(int))) { - DRM_ERROR("copy_to_user\n"); - return -EFAULT; - } - - return 0; -} - -int i915_mem_free(struct drm_device *dev, void *data, - struct drm_file *file_priv) -{ - drm_i915_private_t *dev_priv = dev->dev_private; - drm_i915_mem_free_t *memfree = data; - struct mem_block *block, **heap; - - if (!dev_priv) { - DRM_ERROR("called with no initialization\n"); - return -EINVAL; - } - - heap = get_heap(dev_priv, memfree->region); - if (!heap || !*heap) - return -EFAULT; - - block = find_block(*heap, memfree->region_offset); - if (!block) - return -EFAULT; - - if (block->file_priv != file_priv) - return -EPERM; - - mark_block(dev, block, 0); - free_block(block); - return 0; -} - -int i915_mem_init_heap(struct drm_device *dev, void *data, - struct drm_file *file_priv) -{ - drm_i915_private_t *dev_priv = dev->dev_private; - drm_i915_mem_init_heap_t *initheap = data; - struct mem_block **heap; - - if (!dev_priv) { - DRM_ERROR("called with no initialization\n"); - return -EINVAL; - } - - heap = get_heap(dev_priv, initheap->region); - if (!heap) - return -EFAULT; - - if (*heap) { - DRM_ERROR("heap already initialized?"); - return -EFAULT; - } - - return init_heap(heap, initheap->start, initheap->size); -} - -int i915_mem_destroy_heap(struct drm_device *dev, void *data, - struct drm_file *file_priv) -{ - drm_i915_private_t *dev_priv = dev->dev_private; - drm_i915_mem_destroy_heap_t *destroyheap = data; - struct mem_block **heap; - - if (!dev_priv) { - DRM_ERROR("called with no initialization\n"); - return -EINVAL; - } - - heap = get_heap(dev_priv, destroyheap->region); - if (!heap) { - DRM_ERROR("get_heap failed"); - return -EFAULT; - } - - if (!*heap) { - DRM_ERROR("heap not initialized?"); - return -EFAULT; - } - - i915_mem_takedown(heap); - return 0; -} diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 124dd87..dde62bf 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -88,12 +88,45 @@ #define GEN6_MBC_SNPCR_LOW (2<<21) #define GEN6_MBC_SNPCR_MIN (3<<21) /* only 1/16th of the cache is shared */ +#define GEN6_MBCTL 0x0907c +#define GEN6_MBCTL_ENABLE_BOOT_FETCH (1 << 4) +#define GEN6_MBCTL_CTX_FETCH_NEEDED (1 << 3) +#define GEN6_MBCTL_BME_UPDATE_ENABLE (1 << 2) +#define GEN6_MBCTL_MAE_UPDATE_ENABLE (1 << 1) +#define GEN6_MBCTL_BOOT_FETCH_MECH (1 << 0) + #define GEN6_GDRST 0x941c #define GEN6_GRDOM_FULL (1 << 0) #define GEN6_GRDOM_RENDER (1 << 1) #define GEN6_GRDOM_MEDIA (1 << 2) #define GEN6_GRDOM_BLT (1 << 3) +/* PPGTT stuff */ +#define GEN6_GTT_ADDR_ENCODE(addr) ((addr) | (((addr) >> 28) & 0xff0)) + +#define GEN6_PDE_VALID (1 << 0) +#define GEN6_PDE_LARGE_PAGE (2 << 0) /* use 32kb pages */ +/* gen6+ has bit 11-4 for physical addr bit 39-32 */ +#define GEN6_PDE_ADDR_ENCODE(addr) GEN6_GTT_ADDR_ENCODE(addr) + +#define GEN6_PTE_VALID (1 << 0) +#define GEN6_PTE_UNCACHED (1 << 1) +#define GEN6_PTE_CACHE_LLC (2 << 1) +#define GEN6_PTE_CACHE_LLC_MLC (3 << 1) +#define GEN6_PTE_CACHE_BITS (3 << 1) +#define GEN6_PTE_GFDT (1 << 3) +#define GEN6_PTE_ADDR_ENCODE(addr) GEN6_GTT_ADDR_ENCODE(addr) + +#define RING_PP_DIR_BASE(ring) ((ring)->mmio_base+0x228) +#define RING_PP_DIR_BASE_READ(ring) ((ring)->mmio_base+0x518) +#define RING_PP_DIR_DCLV(ring) ((ring)->mmio_base+0x220) +#define PP_DIR_DCLV_2G 0xffffffff + +#define GAM_ECOCHK 0x4090 +#define ECOCHK_SNB_BIT (1<<10) +#define ECOCHK_PPGTT_CACHE64B (0x3<<3) +#define ECOCHK_PPGTT_CACHE4B (0x0<<3) + /* VGA stuff */ #define VGA_ST01_MDA 0x3ba @@ -305,6 +338,12 @@ #define FENCE_REG_SANDYBRIDGE_0 0x100000 #define SANDYBRIDGE_FENCE_PITCH_SHIFT 32 +/* control register for cpu gtt access */ +#define TILECTL 0x101000 +#define TILECTL_SWZCTL (1 << 0) +#define TILECTL_TLB_PREFETCH_DIS (1 << 2) +#define TILECTL_BACKSNOOP_DIS (1 << 3) + /* * Instruction and interrupt control regs */ @@ -328,7 +367,14 @@ #define RING_MAX_IDLE(base) ((base)+0x54) #define RING_HWS_PGA(base) ((base)+0x80) #define RING_HWS_PGA_GEN6(base) ((base)+0x2080) +#define ARB_MODE 0x04030 +#define ARB_MODE_SWIZZLE_SNB (1<<4) +#define ARB_MODE_SWIZZLE_IVB (1<<5) +#define ARB_MODE_ENABLE(x) GFX_MODE_ENABLE(x) +#define ARB_MODE_DISABLE(x) GFX_MODE_DISABLE(x) #define RENDER_HWS_PGA_GEN7 (0x04080) +#define RING_FAULT_REG(ring) (0x4094 + 0x100*(ring)->id) +#define DONE_REG 0x40b0 #define BSD_HWS_PGA_GEN7 (0x04180) #define BLT_HWS_PGA_GEN7 (0x04280) #define RING_ACTHD(base) ((base)+0x74) @@ -362,6 +408,12 @@ #define IPEIR_I965 0x02064 #define IPEHR_I965 0x02068 #define INSTDONE_I965 0x0206c +#define RING_IPEIR(base) ((base)+0x64) +#define RING_IPEHR(base) ((base)+0x68) +#define RING_INSTDONE(base) ((base)+0x6c) +#define RING_INSTPS(base) ((base)+0x70) +#define RING_DMA_FADD(base) ((base)+0x78) +#define RING_INSTPM(base) ((base)+0xc0) #define INSTPS 0x02070 /* 965+ only */ #define INSTDONE1 0x0207c /* 965+ only */ #define ACTHD_I965 0x02074 @@ -375,14 +427,6 @@ #define INSTDONE 0x02090 #define NOPID 0x02094 #define HWSTAM 0x02098 -#define VCS_INSTDONE 0x1206C -#define VCS_IPEIR 0x12064 -#define VCS_IPEHR 0x12068 -#define VCS_ACTHD 0x12074 -#define BCS_INSTDONE 0x2206C -#define BCS_IPEIR 0x22064 -#define BCS_IPEHR 0x22068 -#define BCS_ACTHD 0x22074 #define ERROR_GEN6 0x040a0 @@ -403,13 +447,13 @@ #define MI_MODE 0x0209c # define VS_TIMER_DISPATCH (1 << 6) # define MI_FLUSH_ENABLE (1 << 12) -# define ASYNC_FLIP_PERF_DISABLE (1 << 14) #define GEN6_GT_MODE 0x20d0 #define GEN6_GT_MODE_HI (1 << 9) #define GFX_MODE 0x02520 #define GFX_MODE_GEN7 0x0229c +#define RING_MODE_GEN7(ring) ((ring)->mmio_base+0x29c) #define GFX_RUN_LIST_ENABLE (1<<15) #define GFX_TLB_INVALIDATE_ALWAYS (1<<13) #define GFX_SURFACE_FAULT_ENABLE (1<<12) @@ -1068,6 +1112,29 @@ #define C0DRB3 0x10206 #define C1DRB3 0x10606 +/** snb MCH registers for reading the DRAM channel configuration */ +#define MAD_DIMM_C0 (MCHBAR_MIRROR_BASE_SNB + 0x5004) +#define MAD_DIMM_C1 (MCHBAR_MIRROR_BASE_SNB + 0x5008) +#define MAD_DIMM_C2 (MCHBAR_MIRROR_BASE_SNB + 0x500C) +#define MAD_DIMM_ECC_MASK (0x3 << 24) +#define MAD_DIMM_ECC_OFF (0x0 << 24) +#define MAD_DIMM_ECC_IO_ON_LOGIC_OFF (0x1 << 24) +#define MAD_DIMM_ECC_IO_OFF_LOGIC_ON (0x2 << 24) +#define MAD_DIMM_ECC_ON (0x3 << 24) +#define MAD_DIMM_ENH_INTERLEAVE (0x1 << 22) +#define MAD_DIMM_RANK_INTERLEAVE (0x1 << 21) +#define MAD_DIMM_B_WIDTH_X16 (0x1 << 20) /* X8 chips if unset */ +#define MAD_DIMM_A_WIDTH_X16 (0x1 << 19) /* X8 chips if unset */ +#define MAD_DIMM_B_DUAL_RANK (0x1 << 18) +#define MAD_DIMM_A_DUAL_RANK (0x1 << 17) +#define MAD_DIMM_A_SELECT (0x1 << 16) +/* DIMM sizes are in multiples of 256mb. */ +#define MAD_DIMM_B_SIZE_SHIFT 8 +#define MAD_DIMM_B_SIZE_MASK (0xff << MAD_DIMM_B_SIZE_SHIFT) +#define MAD_DIMM_A_SIZE_SHIFT 0 +#define MAD_DIMM_A_SIZE_MASK (0xff << MAD_DIMM_A_SIZE_SHIFT) + + /* Clocking configuration register */ #define CLKCFG 0x10c00 #define CLKCFG_FSB_400 (5 << 0) /* hrawclk 100 */ @@ -1347,6 +1414,7 @@ #define _VSYNC_A 0x60014 #define _PIPEASRC 0x6001c #define _BCLRPAT_A 0x60020 +#define _VSYNCSHIFT_A 0x60028 /* Pipe B timing regs */ #define _HTOTAL_B 0x61000 @@ -1357,6 +1425,8 @@ #define _VSYNC_B 0x61014 #define _PIPEBSRC 0x6101c #define _BCLRPAT_B 0x61020 +#define _VSYNCSHIFT_B 0x61028 + #define HTOTAL(pipe) _PIPE(pipe, _HTOTAL_A, _HTOTAL_B) #define HBLANK(pipe) _PIPE(pipe, _HBLANK_A, _HBLANK_B) @@ -1365,6 +1435,7 @@ #define VBLANK(pipe) _PIPE(pipe, _VBLANK_A, _VBLANK_B) #define VSYNC(pipe) _PIPE(pipe, _VSYNC_A, _VSYNC_B) #define BCLRPAT(pipe) _PIPE(pipe, _BCLRPAT_A, _BCLRPAT_B) +#define VSYNCSHIFT(pipe) _PIPE(pipe, _VSYNCSHIFT_A, _VSYNCSHIFT_B) /* VGA port control */ #define ADPA 0x61100 @@ -1564,10 +1635,6 @@ /* Video Data Island Packet control */ #define VIDEO_DIP_DATA 0x61178 -/* Read the description of VIDEO_DIP_DATA (before Haswel) or VIDEO_DIP_ECC - * (Haswell and newer) to see which VIDEO_DIP_DATA byte corresponds to each byte - * of the infoframe structure specified by CEA-861. */ -#define VIDEO_DIP_DATA_SIZE 32 #define VIDEO_DIP_CTL 0x61170 #define VIDEO_DIP_ENABLE (1 << 31) #define VIDEO_DIP_PORT_B (1 << 29) @@ -2361,9 +2428,21 @@ #define PIPECONF_PALETTE 0 #define PIPECONF_GAMMA (1<<24) #define PIPECONF_FORCE_BORDER (1<<25) -#define PIPECONF_PROGRESSIVE (0 << 21) +#define PIPECONF_INTERLACE_MASK (7 << 21) +/* Note that pre-gen3 does not support interlaced display directly. Panel + * fitting must be disabled on pre-ilk for interlaced. */ +#define PIPECONF_PROGRESSIVE (0 << 21) +#define PIPECONF_INTERLACE_W_SYNC_SHIFT_PANEL (4 << 21) /* gen4 only */ +#define PIPECONF_INTERLACE_W_SYNC_SHIFT (5 << 21) /* gen4 only */ #define PIPECONF_INTERLACE_W_FIELD_INDICATION (6 << 21) -#define PIPECONF_INTERLACE_FIELD_0_ONLY (7 << 21) +#define PIPECONF_INTERLACE_FIELD_0_ONLY (7 << 21) /* gen3 only */ +/* Ironlake and later have a complete new set of values for interlaced. PFIT + * means panel fitter required, PF means progressive fetch, DBL means power + * saving pixel doubling. */ +#define PIPECONF_PFIT_PF_INTERLACED_ILK (1 << 21) +#define PIPECONF_INTERLACED_ILK (3 << 21) +#define PIPECONF_INTERLACED_DBL_ILK (4 << 21) /* ilk/snb only */ +#define PIPECONF_PFIT_PF_INTERLACED_DBL_ILK (5 << 21) /* ilk/snb only */ #define PIPECONF_CXSR_DOWNCLOCK (1<<16) #define PIPECONF_BPP_MASK (0x000000e0) #define PIPECONF_BPP_8 (0<<5) @@ -2502,6 +2581,8 @@ #define WM3_LP_ILK 0x45110 #define WM3_LP_EN (1<<31) #define WM1S_LP_ILK 0x45120 +#define WM2S_LP_IVB 0x45124 +#define WM3S_LP_IVB 0x45128 #define WM1S_LP_EN (1<<31) /* Memory latency timer register */ @@ -2718,6 +2799,140 @@ #define _DSPBSURF 0x7119C #define _DSPBTILEOFF 0x711A4 +/* Sprite A control */ +#define _DVSACNTR 0x72180 +#define DVS_ENABLE (1<<31) +#define DVS_GAMMA_ENABLE (1<<30) +#define DVS_PIXFORMAT_MASK (3<<25) +#define DVS_FORMAT_YUV422 (0<<25) +#define DVS_FORMAT_RGBX101010 (1<<25) +#define DVS_FORMAT_RGBX888 (2<<25) +#define DVS_FORMAT_RGBX161616 (3<<25) +#define DVS_SOURCE_KEY (1<<22) +#define DVS_RGB_ORDER_XBGR (1<<20) +#define DVS_YUV_BYTE_ORDER_MASK (3<<16) +#define DVS_YUV_ORDER_YUYV (0<<16) +#define DVS_YUV_ORDER_UYVY (1<<16) +#define DVS_YUV_ORDER_YVYU (2<<16) +#define DVS_YUV_ORDER_VYUY (3<<16) +#define DVS_DEST_KEY (1<<2) +#define DVS_TRICKLE_FEED_DISABLE (1<<14) +#define DVS_TILED (1<<10) +#define _DVSALINOFF 0x72184 +#define _DVSASTRIDE 0x72188 +#define _DVSAPOS 0x7218c +#define _DVSASIZE 0x72190 +#define _DVSAKEYVAL 0x72194 +#define _DVSAKEYMSK 0x72198 +#define _DVSASURF 0x7219c +#define _DVSAKEYMAXVAL 0x721a0 +#define _DVSATILEOFF 0x721a4 +#define _DVSASURFLIVE 0x721ac +#define _DVSASCALE 0x72204 +#define DVS_SCALE_ENABLE (1<<31) +#define DVS_FILTER_MASK (3<<29) +#define DVS_FILTER_MEDIUM (0<<29) +#define DVS_FILTER_ENHANCING (1<<29) +#define DVS_FILTER_SOFTENING (2<<29) +#define DVS_VERTICAL_OFFSET_HALF (1<<28) /* must be enabled below */ +#define DVS_VERTICAL_OFFSET_ENABLE (1<<27) +#define _DVSAGAMC 0x72300 + +#define _DVSBCNTR 0x73180 +#define _DVSBLINOFF 0x73184 +#define _DVSBSTRIDE 0x73188 +#define _DVSBPOS 0x7318c +#define _DVSBSIZE 0x73190 +#define _DVSBKEYVAL 0x73194 +#define _DVSBKEYMSK 0x73198 +#define _DVSBSURF 0x7319c +#define _DVSBKEYMAXVAL 0x731a0 +#define _DVSBTILEOFF 0x731a4 +#define _DVSBSURFLIVE 0x731ac +#define _DVSBSCALE 0x73204 +#define _DVSBGAMC 0x73300 + +#define DVSCNTR(pipe) _PIPE(pipe, _DVSACNTR, _DVSBCNTR) +#define DVSLINOFF(pipe) _PIPE(pipe, _DVSALINOFF, _DVSBLINOFF) +#define DVSSTRIDE(pipe) _PIPE(pipe, _DVSASTRIDE, _DVSBSTRIDE) +#define DVSPOS(pipe) _PIPE(pipe, _DVSAPOS, _DVSBPOS) +#define DVSSURF(pipe) _PIPE(pipe, _DVSASURF, _DVSBSURF) +#define DVSKEYMAX(pipe) _PIPE(pipe, _DVSAKEYMAXVAL, _DVSBKEYMAXVAL) +#define DVSSIZE(pipe) _PIPE(pipe, _DVSASIZE, _DVSBSIZE) +#define DVSSCALE(pipe) _PIPE(pipe, _DVSASCALE, _DVSBSCALE) +#define DVSTILEOFF(pipe) _PIPE(pipe, _DVSATILEOFF, _DVSBTILEOFF) +#define DVSKEYVAL(pipe) _PIPE(pipe, _DVSAKEYVAL, _DVSBKEYVAL) +#define DVSKEYMSK(pipe) _PIPE(pipe, _DVSAKEYMSK, _DVSBKEYMSK) + +#define _SPRA_CTL 0x70280 +#define SPRITE_ENABLE (1<<31) +#define SPRITE_GAMMA_ENABLE (1<<30) +#define SPRITE_PIXFORMAT_MASK (7<<25) +#define SPRITE_FORMAT_YUV422 (0<<25) +#define SPRITE_FORMAT_RGBX101010 (1<<25) +#define SPRITE_FORMAT_RGBX888 (2<<25) +#define SPRITE_FORMAT_RGBX161616 (3<<25) +#define SPRITE_FORMAT_YUV444 (4<<25) +#define SPRITE_FORMAT_XR_BGR101010 (5<<25) /* Extended range */ +#define SPRITE_CSC_ENABLE (1<<24) +#define SPRITE_SOURCE_KEY (1<<22) +#define SPRITE_RGB_ORDER_RGBX (1<<20) /* only for 888 and 161616 */ +#define SPRITE_YUV_TO_RGB_CSC_DISABLE (1<<19) +#define SPRITE_YUV_CSC_FORMAT_BT709 (1<<18) /* 0 is BT601 */ +#define SPRITE_YUV_BYTE_ORDER_MASK (3<<16) +#define SPRITE_YUV_ORDER_YUYV (0<<16) +#define SPRITE_YUV_ORDER_UYVY (1<<16) +#define SPRITE_YUV_ORDER_YVYU (2<<16) +#define SPRITE_YUV_ORDER_VYUY (3<<16) +#define SPRITE_TRICKLE_FEED_DISABLE (1<<14) +#define SPRITE_INT_GAMMA_ENABLE (1<<13) +#define SPRITE_TILED (1<<10) +#define SPRITE_DEST_KEY (1<<2) +#define _SPRA_LINOFF 0x70284 +#define _SPRA_STRIDE 0x70288 +#define _SPRA_POS 0x7028c +#define _SPRA_SIZE 0x70290 +#define _SPRA_KEYVAL 0x70294 +#define _SPRA_KEYMSK 0x70298 +#define _SPRA_SURF 0x7029c +#define _SPRA_KEYMAX 0x702a0 +#define _SPRA_TILEOFF 0x702a4 +#define _SPRA_SCALE 0x70304 +#define SPRITE_SCALE_ENABLE (1<<31) +#define SPRITE_FILTER_MASK (3<<29) +#define SPRITE_FILTER_MEDIUM (0<<29) +#define SPRITE_FILTER_ENHANCING (1<<29) +#define SPRITE_FILTER_SOFTENING (2<<29) +#define SPRITE_VERTICAL_OFFSET_HALF (1<<28) /* must be enabled below */ +#define SPRITE_VERTICAL_OFFSET_ENABLE (1<<27) +#define _SPRA_GAMC 0x70400 + +#define _SPRB_CTL 0x71280 +#define _SPRB_LINOFF 0x71284 +#define _SPRB_STRIDE 0x71288 +#define _SPRB_POS 0x7128c +#define _SPRB_SIZE 0x71290 +#define _SPRB_KEYVAL 0x71294 +#define _SPRB_KEYMSK 0x71298 +#define _SPRB_SURF 0x7129c +#define _SPRB_KEYMAX 0x712a0 +#define _SPRB_TILEOFF 0x712a4 +#define _SPRB_SCALE 0x71304 +#define _SPRB_GAMC 0x71400 + +#define SPRCTL(pipe) _PIPE(pipe, _SPRA_CTL, _SPRB_CTL) +#define SPRLINOFF(pipe) _PIPE(pipe, _SPRA_LINOFF, _SPRB_LINOFF) +#define SPRSTRIDE(pipe) _PIPE(pipe, _SPRA_STRIDE, _SPRB_STRIDE) +#define SPRPOS(pipe) _PIPE(pipe, _SPRA_POS, _SPRB_POS) +#define SPRSIZE(pipe) _PIPE(pipe, _SPRA_SIZE, _SPRB_SIZE) +#define SPRKEYVAL(pipe) _PIPE(pipe, _SPRA_KEYVAL, _SPRB_KEYVAL) +#define SPRKEYMSK(pipe) _PIPE(pipe, _SPRA_KEYMSK, _SPRB_KEYMSK) +#define SPRSURF(pipe) _PIPE(pipe, _SPRA_SURF, _SPRB_SURF) +#define SPRKEYMAX(pipe) _PIPE(pipe, _SPRA_KEYMAX, _SPRB_KEYMAX) +#define SPRTILEOFF(pipe) _PIPE(pipe, _SPRA_TILEOFF, _SPRB_TILEOFF) +#define SPRSCALE(pipe) _PIPE(pipe, _SPRA_SCALE, _SPRB_SCALE) +#define SPRGAMC(pipe) _PIPE(pipe, _SPRA_GAMC, _SPRB_GAMC) + /* VBIOS regs */ #define VGACNTRL 0x71400 # define VGA_DISP_DISABLE (1 << 31) @@ -2927,6 +3142,10 @@ #define ILK_DPFC_DIS1 (1<<8) #define ILK_DPFC_DIS2 (1<<9) +#define IVB_CHICKEN3 0x4200c +# define CHICKEN3_DGMG_REQ_OUT_FIX_DISABLE (1 << 5) +# define CHICKEN3_DGMG_DONE_FIX_DISABLE (1 << 2) + #define DISP_ARB_CTL 0x45000 #define DISP_TILE_SURFACE_SWIZZLING (1<<13) #define DISP_FBC_WM_DIS (1<<15) @@ -3122,6 +3341,7 @@ #define _TRANS_VSYNC_A 0xe0014 #define TRANS_VSYNC_END_SHIFT 16 #define TRANS_VSYNC_START_SHIFT 0 +#define _TRANS_VSYNCSHIFT_A 0xe0028 #define _TRANSA_DATA_M1 0xe0030 #define _TRANSA_DATA_N1 0xe0034 @@ -3152,6 +3372,7 @@ #define _TRANS_VTOTAL_B 0xe100c #define _TRANS_VBLANK_B 0xe1010 #define _TRANS_VSYNC_B 0xe1014 +#define _TRANS_VSYNCSHIFT_B 0xe1028 #define TRANS_HTOTAL(pipe) _PIPE(pipe, _TRANS_HTOTAL_A, _TRANS_HTOTAL_B) #define TRANS_HBLANK(pipe) _PIPE(pipe, _TRANS_HBLANK_A, _TRANS_HBLANK_B) @@ -3159,6 +3380,8 @@ #define TRANS_VTOTAL(pipe) _PIPE(pipe, _TRANS_VTOTAL_A, _TRANS_VTOTAL_B) #define TRANS_VBLANK(pipe) _PIPE(pipe, _TRANS_VBLANK_A, _TRANS_VBLANK_B) #define TRANS_VSYNC(pipe) _PIPE(pipe, _TRANS_VSYNC_A, _TRANS_VSYNC_B) +#define TRANS_VSYNCSHIFT(pipe) _PIPE(pipe, _TRANS_VSYNCSHIFT_A, \ + _TRANS_VSYNCSHIFT_B) #define _TRANSB_DATA_M1 0xe1030 #define _TRANSB_DATA_N1 0xe1034 @@ -3192,7 +3415,10 @@ #define TRANS_FSYNC_DELAY_HB4 (3<<27) #define TRANS_DP_AUDIO_ONLY (1<<26) #define TRANS_DP_VIDEO_AUDIO (0<<26) +#define TRANS_INTERLACE_MASK (7<<21) #define TRANS_PROGRESSIVE (0<<21) +#define TRANS_INTERLACED (3<<21) +#define TRANS_LEGACY_INTERLACED_ILK (2<<21) #define TRANS_8BPC (0<<5) #define TRANS_10BPC (1<<5) #define TRANS_6BPC (2<<5) @@ -3531,9 +3757,18 @@ #define ECOBUS 0xa180 #define FORCEWAKE_MT_ENABLE (1<<5) +#define GTFIFODBG 0x120000 +#define GT_FIFO_CPU_ERROR_MASK 7 +#define GT_FIFO_OVFERR (1<<2) +#define GT_FIFO_IAWRERR (1<<1) +#define GT_FIFO_IARDERR (1<<0) + #define GT_FIFO_FREE_ENTRIES 0x120008 #define GT_FIFO_NUM_RESERVED_ENTRIES 20 +#define GEN6_UCGCTL1 0x9400 +# define GEN6_BLBUNIT_CLOCK_GATE_DISABLE (1 << 5) + #define GEN6_UCGCTL2 0x9404 # define GEN6_RCZUNIT_CLOCK_GATE_DISABLE (1 << 13) # define GEN6_RCPBUNIT_CLOCK_GATE_DISABLE (1 << 12) @@ -3621,6 +3856,14 @@ #define GEN6_PCODE_DATA 0x138128 #define GEN6_PCODE_FREQ_IA_RATIO_SHIFT 8 +#define GEN6_GT_CORE_STATUS 0x138060 +#define GEN6_CORE_CPD_STATE_MASK (7<<4) +#define GEN6_RCn_MASK 7 +#define GEN6_RC0 0 +#define GEN6_RC3 2 +#define GEN6_RC6 3 +#define GEN6_RC7 4 + #define G4X_AUD_VID_DID 0x62020 #define INTEL_AUDIO_DEVCL 0x808629FB #define INTEL_AUDIO_DEVBLC 0x80862801 @@ -3633,17 +3876,35 @@ #define G4X_ELD_ACK (1 << 4) #define G4X_HDMIW_HDMIEDID 0x6210C -#define GEN5_HDMIW_HDMIEDID_A 0xE2050 -#define GEN5_AUD_CNTL_ST_A 0xE20B4 -#define GEN5_ELD_BUFFER_SIZE (0x1f << 10) -#define GEN5_ELD_ADDRESS (0x1f << 5) -#define GEN5_ELD_ACK (1 << 4) -#define GEN5_AUD_CNTL_ST2 0xE20C0 -#define GEN5_ELD_VALIDB (1 << 0) -#define GEN5_CP_READYB (1 << 1) - -#define GEN7_HDMIW_HDMIEDID_A 0xE5050 -#define GEN7_AUD_CNTRL_ST_A 0xE50B4 -#define GEN7_AUD_CNTRL_ST2 0xE50C0 +#define IBX_HDMIW_HDMIEDID_A 0xE2050 +#define IBX_AUD_CNTL_ST_A 0xE20B4 +#define IBX_ELD_BUFFER_SIZE (0x1f << 10) +#define IBX_ELD_ADDRESS (0x1f << 5) +#define IBX_ELD_ACK (1 << 4) +#define IBX_AUD_CNTL_ST2 0xE20C0 +#define IBX_ELD_VALIDB (1 << 0) +#define IBX_CP_READYB (1 << 1) + +#define CPT_HDMIW_HDMIEDID_A 0xE5050 +#define CPT_AUD_CNTL_ST_A 0xE50B4 +#define CPT_AUD_CNTRL_ST2 0xE50C0 + +/* These are the 4 32-bit write offset registers for each stream + * output buffer. It determines the offset from the + * 3DSTATE_SO_BUFFERs that the next streamed vertex output goes to. + */ +#define GEN7_SO_WRITE_OFFSET(n) (0x5280 + (n) * 4) + +#define IBX_AUD_CONFIG_A 0xe2000 +#define CPT_AUD_CONFIG_A 0xe5000 +#define AUD_CONFIG_N_VALUE_INDEX (1 << 29) +#define AUD_CONFIG_N_PROG_ENABLE (1 << 28) +#define AUD_CONFIG_UPPER_N_SHIFT 20 +#define AUD_CONFIG_UPPER_N_VALUE (0xff << 20) +#define AUD_CONFIG_LOWER_N_SHIFT 4 +#define AUD_CONFIG_LOWER_N_VALUE (0xfff << 4) +#define AUD_CONFIG_PIXEL_CLOCK_HDMI_SHIFT 16 +#define AUD_CONFIG_PIXEL_CLOCK_HDMI (0xf << 16) +#define AUD_CONFIG_DISABLE_NCTS (1 << 3) #endif /* _I915_REG_H_ */ diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c index f38d196..0d13778 100644 --- a/drivers/gpu/drm/i915/i915_suspend.c +++ b/drivers/gpu/drm/i915/i915_suspend.c @@ -28,6 +28,7 @@ #include "drm.h" #include "i915_drm.h" #include "intel_drv.h" +#include "i915_reg.h" static bool i915_pipe_enabled(struct drm_device *dev, enum pipe pipe) { @@ -39,7 +40,7 @@ static bool i915_pipe_enabled(struct drm_device *dev, enum pipe pipe) return false; if (HAS_PCH_SPLIT(dev)) - dpll_reg = (pipe == PIPE_A) ? _PCH_DPLL_A : _PCH_DPLL_B; + dpll_reg = PCH_DPLL(pipe); else dpll_reg = (pipe == PIPE_A) ? _DPLL_A : _DPLL_B; diff --git a/drivers/gpu/drm/i915/intel_acpi.c b/drivers/gpu/drm/i915/intel_acpi.c index cb91210..bae3edf 100644 --- a/drivers/gpu/drm/i915/intel_acpi.c +++ b/drivers/gpu/drm/i915/intel_acpi.c @@ -208,7 +208,7 @@ static bool intel_dsm_pci_probe(struct pci_dev *pdev) ret = intel_dsm(dhandle, INTEL_DSM_FN_SUPPORTED_FUNCTIONS, 0); if (ret < 0) { - DRM_ERROR("failed to get supported _DSM functions\n"); + DRM_DEBUG_KMS("failed to get supported _DSM functions\n"); return false; } diff --git a/drivers/gpu/drm/i915/intel_bios.c b/drivers/gpu/drm/i915/intel_bios.c index 0016fee..a2c9e56 100644 --- a/drivers/gpu/drm/i915/intel_bios.c +++ b/drivers/gpu/drm/i915/intel_bios.c @@ -605,7 +605,7 @@ parse_device_mapping(struct drm_i915_private *dev_priv, DRM_DEBUG_KMS("no child dev is parsed from VBT\n"); return; } - dev_priv->child_dev = kzalloc(sizeof(*p_child) * count, GFP_KERNEL); + dev_priv->child_dev = kcalloc(count, sizeof(*p_child), GFP_KERNEL); if (!dev_priv->child_dev) { DRM_DEBUG_KMS("No memory space for child device\n"); return; @@ -719,7 +719,7 @@ intel_parse_bios(struct drm_device *dev) } if (!vbt) { - DRM_ERROR("VBT signature missing\n"); + DRM_DEBUG_DRIVER("VBT signature missing\n"); pci_unmap_rom(pdev, bios); return -1; } diff --git a/drivers/gpu/drm/i915/intel_bios.h b/drivers/gpu/drm/i915/intel_bios.h index 8af3735..dbda6e3 100644 --- a/drivers/gpu/drm/i915/intel_bios.h +++ b/drivers/gpu/drm/i915/intel_bios.h @@ -467,8 +467,12 @@ struct edp_link_params { struct bdb_edp { struct edp_power_seq power_seqs[16]; u32 color_depth; - u32 sdrrs_msa_timing_delay; struct edp_link_params link_params[16]; + u32 sdrrs_msa_timing_delay; + + /* ith bit indicates enabled/disabled for (i+1)th panel */ + u16 edp_s3d_feature; + u16 edp_t3_optimization; } __attribute__ ((packed)); void intel_setup_bios(struct drm_device *dev); diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c index fee0ad0..342ffb7 100644 --- a/drivers/gpu/drm/i915/intel_crt.c +++ b/drivers/gpu/drm/i915/intel_crt.c @@ -24,6 +24,7 @@ * Eric Anholt */ +#include #include #include #include "drmP.h" @@ -265,6 +266,36 @@ static bool intel_crt_detect_hotplug(struct drm_connector *connector) return ret; } +static struct edid *intel_crt_get_edid(struct drm_connector *connector, + struct i2c_adapter *i2c) +{ + struct edid *edid; + + edid = drm_get_edid(connector, i2c); + + if (!edid && !intel_gmbus_is_forced_bit(i2c)) { + DRM_DEBUG_KMS("CRT GMBUS EDID read failed, retry using GPIO bit-banging\n"); + intel_gmbus_force_bit(i2c, true); + edid = drm_get_edid(connector, i2c); + intel_gmbus_force_bit(i2c, false); + } + + return edid; +} + +/* local version of intel_ddc_get_modes() to use intel_crt_get_edid() */ +static int intel_crt_ddc_get_modes(struct drm_connector *connector, + struct i2c_adapter *adapter) +{ + struct edid *edid; + + edid = intel_crt_get_edid(connector, adapter); + if (!edid) + return 0; + + return intel_connector_update_modes(connector, edid); +} + static bool intel_crt_detect_ddc(struct drm_connector *connector) { struct intel_crt *crt = intel_attached_crt(connector); @@ -278,7 +309,7 @@ static bool intel_crt_detect_ddc(struct drm_connector *connector) struct edid *edid; bool is_digital = false; - edid = drm_get_edid(connector, + edid = intel_crt_get_edid(connector, &dev_priv->gmbus[dev_priv->crt_ddc_pin].adapter); /* * This may be a DVI-I connector with a shared DDC @@ -429,8 +460,8 @@ intel_crt_detect(struct drm_connector *connector, bool force) { struct drm_device *dev = connector->dev; struct intel_crt *crt = intel_attached_crt(connector); - struct drm_crtc *crtc; enum drm_connector_status status; + struct intel_load_detect_pipe tmp; if (I915_HAS_HOTPLUG(dev)) { if (intel_crt_detect_hotplug(connector)) { @@ -449,23 +480,16 @@ intel_crt_detect(struct drm_connector *connector, bool force) return connector->status; /* for pre-945g platforms use load detect */ - crtc = crt->base.base.crtc; - if (crtc && crtc->enabled) { - status = intel_crt_load_detect(crt); - } else { - struct intel_load_detect_pipe tmp; - - if (intel_get_load_detect_pipe(&crt->base, connector, NULL, - &tmp)) { - if (intel_crt_detect_ddc(connector)) - status = connector_status_connected; - else - status = intel_crt_load_detect(crt); - intel_release_load_detect_pipe(&crt->base, connector, - &tmp); - } else - status = connector_status_unknown; - } + if (intel_get_load_detect_pipe(&crt->base, connector, NULL, + &tmp)) { + if (intel_crt_detect_ddc(connector)) + status = connector_status_connected; + else + status = intel_crt_load_detect(crt); + intel_release_load_detect_pipe(&crt->base, connector, + &tmp); + } else + status = connector_status_unknown; return status; } @@ -483,13 +507,13 @@ static int intel_crt_get_modes(struct drm_connector *connector) struct drm_i915_private *dev_priv = dev->dev_private; int ret; - ret = intel_ddc_get_modes(connector, + ret = intel_crt_ddc_get_modes(connector, &dev_priv->gmbus[dev_priv->crt_ddc_pin].adapter); if (ret || !IS_G4X(dev)) return ret; /* Try to probe digital port for output in DVI-I -> VGA mode. */ - return intel_ddc_get_modes(connector, + return intel_crt_ddc_get_modes(connector, &dev_priv->gmbus[GMBUS_PORT_DPB].adapter); } @@ -540,6 +564,24 @@ static const struct drm_encoder_funcs intel_crt_enc_funcs = { .destroy = intel_encoder_destroy, }; +static int __init intel_no_crt_dmi_callback(const struct dmi_system_id *id) +{ + DRM_DEBUG_KMS("Skipping CRT initialization for %s\n", id->ident); + return 1; +} + +static const struct dmi_system_id intel_no_crt[] = { + { + .callback = intel_no_crt_dmi_callback, + .ident = "ACER ZGB", + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "ACER"), + DMI_MATCH(DMI_PRODUCT_NAME, "ZGB"), + }, + }, + { } +}; + void intel_crt_init(struct drm_device *dev) { struct drm_connector *connector; @@ -547,6 +589,10 @@ void intel_crt_init(struct drm_device *dev) struct intel_connector *intel_connector; struct drm_i915_private *dev_priv = dev->dev_private; + /* Skip machines without VGA that falsely report hotplug events */ + if (dmi_check_system(intel_no_crt)) + return; + crt = kzalloc(sizeof(struct intel_crt), GFP_KERNEL); if (!crt) return; @@ -571,7 +617,10 @@ void intel_crt_init(struct drm_device *dev) 1 << INTEL_ANALOG_CLONE_BIT | 1 << INTEL_SDVO_LVDS_CLONE_BIT); crt->base.crtc_mask = (1 << 0) | (1 << 1); - connector->interlace_allowed = 1; + if (IS_GEN2(dev)) + connector->interlace_allowed = 0; + else + connector->interlace_allowed = 1; connector->doublescan_allowed = 0; drm_encoder_helper_add(&crt->base.base, &intel_crt_helper_funcs); diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 7817429..3c9b9c5 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -75,7 +75,7 @@ struct intel_limit { intel_range_t dot, vco, n, m, m1, m2, p, p1; intel_p2_t p2; bool (* find_pll)(const intel_limit_t *, struct drm_crtc *, - int, int, intel_clock_t *); + int, int, intel_clock_t *, intel_clock_t *); }; /* FDI */ @@ -83,17 +83,21 @@ struct intel_limit { static bool intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc, - int target, int refclk, intel_clock_t *best_clock); + int target, int refclk, intel_clock_t *match_clock, + intel_clock_t *best_clock); static bool intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc, - int target, int refclk, intel_clock_t *best_clock); + int target, int refclk, intel_clock_t *match_clock, + intel_clock_t *best_clock); static bool intel_find_pll_g4x_dp(const intel_limit_t *, struct drm_crtc *crtc, - int target, int refclk, intel_clock_t *best_clock); + int target, int refclk, intel_clock_t *match_clock, + intel_clock_t *best_clock); static bool intel_find_pll_ironlake_dp(const intel_limit_t *, struct drm_crtc *crtc, - int target, int refclk, intel_clock_t *best_clock); + int target, int refclk, intel_clock_t *match_clock, + intel_clock_t *best_clock); static inline u32 /* units of 100MHz */ intel_fdi_link_freq(struct drm_device *dev) @@ -534,7 +538,8 @@ static bool intel_PLL_is_valid(struct drm_device *dev, static bool intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc, - int target, int refclk, intel_clock_t *best_clock) + int target, int refclk, intel_clock_t *match_clock, + intel_clock_t *best_clock) { struct drm_device *dev = crtc->dev; @@ -580,6 +585,9 @@ intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc, if (!intel_PLL_is_valid(dev, limit, &clock)) continue; + if (match_clock && + clock.p != match_clock->p) + continue; this_err = abs(clock.dot - target); if (this_err < err) { @@ -596,7 +604,8 @@ intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc, static bool intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc, - int target, int refclk, intel_clock_t *best_clock) + int target, int refclk, intel_clock_t *match_clock, + intel_clock_t *best_clock) { struct drm_device *dev = crtc->dev; struct drm_i915_private *dev_priv = dev->dev_private; @@ -643,6 +652,9 @@ intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc, if (!intel_PLL_is_valid(dev, limit, &clock)) continue; + if (match_clock && + clock.p != match_clock->p) + continue; this_err = abs(clock.dot - target); if (this_err < err_most) { @@ -660,7 +672,8 @@ intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc, static bool intel_find_pll_ironlake_dp(const intel_limit_t *limit, struct drm_crtc *crtc, - int target, int refclk, intel_clock_t *best_clock) + int target, int refclk, intel_clock_t *match_clock, + intel_clock_t *best_clock) { struct drm_device *dev = crtc->dev; intel_clock_t clock; @@ -686,7 +699,8 @@ intel_find_pll_ironlake_dp(const intel_limit_t *limit, struct drm_crtc *crtc, /* DisplayPort has only two frequencies, 162MHz and 270MHz */ static bool intel_find_pll_g4x_dp(const intel_limit_t *limit, struct drm_crtc *crtc, - int target, int refclk, intel_clock_t *best_clock) + int target, int refclk, intel_clock_t *match_clock, + intel_clock_t *best_clock) { intel_clock_t clock; if (target < 200000) { @@ -933,13 +947,17 @@ static void assert_panel_unlocked(struct drm_i915_private *dev_priv, pipe_name(pipe)); } -static void assert_pipe(struct drm_i915_private *dev_priv, - enum pipe pipe, bool state) +void assert_pipe(struct drm_i915_private *dev_priv, + enum pipe pipe, bool state) { int reg; u32 val; bool cur_state; + /* if we need the pipe A quirk it must be always on */ + if (pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE) + state = true; + reg = PIPECONF(pipe); val = I915_READ(reg); cur_state = !!(val & PIPECONF_ENABLE); @@ -947,22 +965,25 @@ static void assert_pipe(struct drm_i915_private *dev_priv, "pipe %c assertion failure (expected %s, current %s)\n", pipe_name(pipe), state_string(state), state_string(cur_state)); } -#define assert_pipe_enabled(d, p) assert_pipe(d, p, true) -#define assert_pipe_disabled(d, p) assert_pipe(d, p, false) -static void assert_plane_enabled(struct drm_i915_private *dev_priv, - enum plane plane) +static void assert_plane(struct drm_i915_private *dev_priv, + enum plane plane, bool state) { int reg; u32 val; + bool cur_state; reg = DSPCNTR(plane); val = I915_READ(reg); - WARN(!(val & DISPLAY_PLANE_ENABLE), - "plane %c assertion failure, should be active but is disabled\n", - plane_name(plane)); + cur_state = !!(val & DISPLAY_PLANE_ENABLE); + WARN(cur_state != state, + "plane %c assertion failure (expected %s, current %s)\n", + plane_name(plane), state_string(state), state_string(cur_state)); } +#define assert_plane_enabled(d, p) assert_plane(d, p, true) +#define assert_plane_disabled(d, p) assert_plane(d, p, false) + static void assert_planes_disabled(struct drm_i915_private *dev_priv, enum pipe pipe) { @@ -971,8 +992,14 @@ static void assert_planes_disabled(struct drm_i915_private *dev_priv, int cur_pipe; /* Planes are fixed to pipes on ILK+ */ - if (HAS_PCH_SPLIT(dev_priv->dev)) + if (HAS_PCH_SPLIT(dev_priv->dev)) { + reg = DSPCNTR(pipe); + val = I915_READ(reg); + WARN((val & DISPLAY_PLANE_ENABLE), + "plane %c assertion failure, should be disabled but not\n", + plane_name(pipe)); return; + } /* Need to check both planes against the pipe */ for (i = 0; i < 2; i++) { @@ -1224,7 +1251,8 @@ static void intel_disable_pch_pll(struct drm_i915_private *dev_priv, enum pipe pipe) { int reg; - u32 val; + u32 val, pll_mask = TRANSC_DPLL_ENABLE | TRANSC_DPLLB_SEL, + pll_sel = TRANSC_DPLL_ENABLE; if (pipe > 1) return; @@ -1235,6 +1263,15 @@ static void intel_disable_pch_pll(struct drm_i915_private *dev_priv, /* Make sure transcoder isn't still depending on us */ assert_transcoder_disabled(dev_priv, pipe); + if (pipe == 0) + pll_sel |= TRANSC_DPLLA_SEL; + else if (pipe == 1) + pll_sel |= TRANSC_DPLLB_SEL; + + + if ((I915_READ(PCH_DPLL_SEL) & pll_mask) == pll_sel) + return; + reg = PCH_DPLL(pipe); val = I915_READ(reg); val &= ~DPLL_VCO_ENABLE; @@ -1247,7 +1284,8 @@ static void intel_enable_transcoder(struct drm_i915_private *dev_priv, enum pipe pipe) { int reg; - u32 val; + u32 val, pipeconf_val; + struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe]; /* PCH only available on ILK+ */ BUG_ON(dev_priv->info->gen < 5); @@ -1261,6 +1299,7 @@ static void intel_enable_transcoder(struct drm_i915_private *dev_priv, reg = TRANSCONF(pipe); val = I915_READ(reg); + pipeconf_val = I915_READ(PIPECONF(pipe)); if (HAS_PCH_IBX(dev_priv->dev)) { /* @@ -1268,8 +1307,19 @@ static void intel_enable_transcoder(struct drm_i915_private *dev_priv, * that in pipeconf reg. */ val &= ~PIPE_BPC_MASK; - val |= I915_READ(PIPECONF(pipe)) & PIPE_BPC_MASK; + val |= pipeconf_val & PIPE_BPC_MASK; } + + val &= ~TRANS_INTERLACE_MASK; + if ((pipeconf_val & PIPECONF_INTERLACE_MASK) == PIPECONF_INTERLACED_ILK) + if (HAS_PCH_IBX(dev_priv->dev) && + intel_pipe_has_type(crtc, INTEL_OUTPUT_SDVO)) + val |= TRANS_LEGACY_INTERLACED_ILK; + else + val |= TRANS_INTERLACED; + else + val |= TRANS_PROGRESSIVE; + I915_WRITE(reg, val | TRANS_ENABLE); if (wait_for(I915_READ(reg) & TRANS_STATE_ENABLE, 100)) DRM_ERROR("failed to enable transcoder %d\n", pipe); @@ -1529,8 +1579,8 @@ static void i8xx_enable_fbc(struct drm_crtc *crtc, unsigned long interval) u32 fbc_ctl, fbc_ctl2; cfb_pitch = dev_priv->cfb_size / FBC_LL_SIZE; - if (fb->pitch < cfb_pitch) - cfb_pitch = fb->pitch; + if (fb->pitches[0] < cfb_pitch) + cfb_pitch = fb->pitches[0]; /* FBC_CTL wants 64B units */ cfb_pitch = (cfb_pitch / 64) - 1; @@ -2022,6 +2072,8 @@ intel_pin_and_fence_fb_obj(struct drm_device *dev, ret = i915_gem_object_get_fence(obj, pipelined); if (ret) goto err_unpin; + + i915_gem_object_pin_fence(obj); } dev_priv->mm.interruptible = true; @@ -2034,6 +2086,12 @@ err_interruptible: return ret; } +void intel_unpin_fb_obj(struct drm_i915_gem_object *obj) +{ + i915_gem_object_unpin_fence(obj); + i915_gem_object_unpin(obj); +} + static int i9xx_update_plane(struct drm_crtc *crtc, struct drm_framebuffer *fb, int x, int y) { @@ -2091,11 +2149,11 @@ static int i9xx_update_plane(struct drm_crtc *crtc, struct drm_framebuffer *fb, I915_WRITE(reg, dspcntr); Start = obj->gtt_offset; - Offset = y * fb->pitch + x * (fb->bits_per_pixel / 8); + Offset = y * fb->pitches[0] + x * (fb->bits_per_pixel / 8); DRM_DEBUG_KMS("Writing base %08lX %08lX %d %d %d\n", - Start, Offset, x, y, fb->pitch); - I915_WRITE(DSPSTRIDE(plane), fb->pitch); + Start, Offset, x, y, fb->pitches[0]); + I915_WRITE(DSPSTRIDE(plane), fb->pitches[0]); if (INTEL_INFO(dev)->gen >= 4) { I915_WRITE(DSPSURF(plane), Start); I915_WRITE(DSPTILEOFF(plane), (y << 16) | x); @@ -2172,11 +2230,11 @@ static int ironlake_update_plane(struct drm_crtc *crtc, I915_WRITE(reg, dspcntr); Start = obj->gtt_offset; - Offset = y * fb->pitch + x * (fb->bits_per_pixel / 8); + Offset = y * fb->pitches[0] + x * (fb->bits_per_pixel / 8); DRM_DEBUG_KMS("Writing base %08lX %08lX %d %d %d\n", - Start, Offset, x, y, fb->pitch); - I915_WRITE(DSPSTRIDE(plane), fb->pitch); + Start, Offset, x, y, fb->pitches[0]); + I915_WRITE(DSPSTRIDE(plane), fb->pitches[0]); I915_WRITE(DSPSURF(plane), Start); I915_WRITE(DSPTILEOFF(plane), (y << 16) | x); I915_WRITE(DSPADDR(plane), Offset); @@ -2275,7 +2333,7 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y, ret = intel_pipe_set_base_atomic(crtc, crtc->fb, x, y, LEAVE_ATOMIC_MODE_SET); if (ret) { - i915_gem_object_unpin(to_intel_framebuffer(crtc->fb)->obj); + intel_unpin_fb_obj(to_intel_framebuffer(crtc->fb)->obj); mutex_unlock(&dev->struct_mutex); DRM_ERROR("failed to update base address\n"); return ret; @@ -2283,7 +2341,7 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y, if (old_fb) { intel_wait_for_vblank(dev, intel_crtc->pipe); - i915_gem_object_unpin(to_intel_framebuffer(old_fb)->obj); + intel_unpin_fb_obj(to_intel_framebuffer(old_fb)->obj); } mutex_unlock(&dev->struct_mutex); @@ -2427,9 +2485,11 @@ static void ironlake_fdi_link_train(struct drm_crtc *crtc) udelay(150); /* Ironlake workaround, enable clock pointer after FDI enable*/ - I915_WRITE(FDI_RX_CHICKEN(pipe), FDI_RX_PHASE_SYNC_POINTER_OVR); - I915_WRITE(FDI_RX_CHICKEN(pipe), FDI_RX_PHASE_SYNC_POINTER_OVR | - FDI_RX_PHASE_SYNC_POINTER_EN); + if (HAS_PCH_IBX(dev)) { + I915_WRITE(FDI_RX_CHICKEN(pipe), FDI_RX_PHASE_SYNC_POINTER_OVR); + I915_WRITE(FDI_RX_CHICKEN(pipe), FDI_RX_PHASE_SYNC_POINTER_OVR | + FDI_RX_PHASE_SYNC_POINTER_EN); + } reg = FDI_RX_IIR(pipe); for (tries = 0; tries < 5; tries++) { @@ -2942,6 +3002,7 @@ static void ironlake_pch_enable(struct drm_crtc *crtc) I915_WRITE(TRANS_VTOTAL(pipe), I915_READ(VTOTAL(pipe))); I915_WRITE(TRANS_VBLANK(pipe), I915_READ(VBLANK(pipe))); I915_WRITE(TRANS_VSYNC(pipe), I915_READ(VSYNC(pipe))); + I915_WRITE(TRANS_VSYNCSHIFT(pipe), I915_READ(VSYNCSHIFT(pipe))); intel_fdi_normal_train(crtc); @@ -3331,10 +3392,12 @@ static void intel_crtc_disable(struct drm_crtc *crtc) struct drm_device *dev = crtc->dev; crtc_funcs->dpms(crtc, DRM_MODE_DPMS_OFF); + assert_plane_disabled(dev->dev_private, to_intel_crtc(crtc)->plane); + assert_pipe_disabled(dev->dev_private, to_intel_crtc(crtc)->pipe); if (crtc->fb) { mutex_lock(&dev->struct_mutex); - i915_gem_object_unpin(to_intel_framebuffer(crtc->fb)->obj); + intel_unpin_fb_obj(to_intel_framebuffer(crtc->fb)->obj); mutex_unlock(&dev->struct_mutex); } } @@ -3408,10 +3471,10 @@ static bool intel_crtc_mode_fixup(struct drm_crtc *crtc, return false; } - /* XXX some encoders set the crtcinfo, others don't. - * Obviously we need some form of conflict resolution here... - */ - if (adjusted_mode->crtc_htotal == 0) + /* All interlaced capable intel hw wants timings in frames. Note though + * that intel_lvds_mode_fixup does some funny tricks with the crtc + * timings, so we need to be careful not to clobber these.*/ + if (!(adjusted_mode->private_flags & INTEL_MODE_CRTC_TIMINGS_SET)) drm_mode_set_crtcinfo(adjusted_mode, 0); return true; @@ -4527,10 +4590,11 @@ static void ironlake_update_wm(struct drm_device *dev) */ } -static void sandybridge_update_wm(struct drm_device *dev) +void sandybridge_update_wm(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; int latency = SNB_READ_WM0_LATENCY() * 100; /* In unit 0.1us */ + u32 val; int fbc_wm, plane_wm, cursor_wm; unsigned int enabled; @@ -4539,8 +4603,10 @@ static void sandybridge_update_wm(struct drm_device *dev) &sandybridge_display_wm_info, latency, &sandybridge_cursor_wm_info, latency, &plane_wm, &cursor_wm)) { - I915_WRITE(WM0_PIPEA_ILK, - (plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm); + val = I915_READ(WM0_PIPEA_ILK); + val &= ~(WM0_PIPE_PLANE_MASK | WM0_PIPE_CURSOR_MASK); + I915_WRITE(WM0_PIPEA_ILK, val | + ((plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm)); DRM_DEBUG_KMS("FIFO watermarks For pipe A -" " plane %d, " "cursor: %d\n", plane_wm, cursor_wm); @@ -4551,8 +4617,10 @@ static void sandybridge_update_wm(struct drm_device *dev) &sandybridge_display_wm_info, latency, &sandybridge_cursor_wm_info, latency, &plane_wm, &cursor_wm)) { - I915_WRITE(WM0_PIPEB_ILK, - (plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm); + val = I915_READ(WM0_PIPEB_ILK); + val &= ~(WM0_PIPE_PLANE_MASK | WM0_PIPE_CURSOR_MASK); + I915_WRITE(WM0_PIPEB_ILK, val | + ((plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm)); DRM_DEBUG_KMS("FIFO watermarks For pipe B -" " plane %d, cursor: %d\n", plane_wm, cursor_wm); @@ -4565,8 +4633,10 @@ static void sandybridge_update_wm(struct drm_device *dev) &sandybridge_display_wm_info, latency, &sandybridge_cursor_wm_info, latency, &plane_wm, &cursor_wm)) { - I915_WRITE(WM0_PIPEC_IVB, - (plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm); + val = I915_READ(WM0_PIPEC_IVB); + val &= ~(WM0_PIPE_PLANE_MASK | WM0_PIPE_CURSOR_MASK); + I915_WRITE(WM0_PIPEC_IVB, val | + ((plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm)); DRM_DEBUG_KMS("FIFO watermarks For pipe C -" " plane %d, cursor: %d\n", plane_wm, cursor_wm); @@ -4587,7 +4657,8 @@ static void sandybridge_update_wm(struct drm_device *dev) I915_WRITE(WM2_LP_ILK, 0); I915_WRITE(WM1_LP_ILK, 0); - if (!single_plane_enabled(enabled)) + if (!single_plane_enabled(enabled) || + dev_priv->sprite_scaling_enabled) return; enabled = ffs(enabled) - 1; @@ -4637,6 +4708,161 @@ static void sandybridge_update_wm(struct drm_device *dev) cursor_wm); } +static bool +sandybridge_compute_sprite_wm(struct drm_device *dev, int plane, + uint32_t sprite_width, int pixel_size, + const struct intel_watermark_params *display, + int display_latency_ns, int *sprite_wm) +{ + struct drm_crtc *crtc; + int clock; + int entries, tlb_miss; + + crtc = intel_get_crtc_for_plane(dev, plane); + if (crtc->fb == NULL || !crtc->enabled) { + *sprite_wm = display->guard_size; + return false; + } + + clock = crtc->mode.clock; + + /* Use the small buffer method to calculate the sprite watermark */ + entries = ((clock * pixel_size / 1000) * display_latency_ns) / 1000; + tlb_miss = display->fifo_size*display->cacheline_size - + sprite_width * 8; + if (tlb_miss > 0) + entries += tlb_miss; + entries = DIV_ROUND_UP(entries, display->cacheline_size); + *sprite_wm = entries + display->guard_size; + if (*sprite_wm > (int)display->max_wm) + *sprite_wm = display->max_wm; + + return true; +} + +static bool +sandybridge_compute_sprite_srwm(struct drm_device *dev, int plane, + uint32_t sprite_width, int pixel_size, + const struct intel_watermark_params *display, + int latency_ns, int *sprite_wm) +{ + struct drm_crtc *crtc; + unsigned long line_time_us; + int clock; + int line_count, line_size; + int small, large; + int entries; + + if (!latency_ns) { + *sprite_wm = 0; + return false; + } + + crtc = intel_get_crtc_for_plane(dev, plane); + clock = crtc->mode.clock; + if (!clock) { + *sprite_wm = 0; + return false; + } + + line_time_us = (sprite_width * 1000) / clock; + if (!line_time_us) { + *sprite_wm = 0; + return false; + } + + line_count = (latency_ns / line_time_us + 1000) / 1000; + line_size = sprite_width * pixel_size; + + /* Use the minimum of the small and large buffer method for primary */ + small = ((clock * pixel_size / 1000) * latency_ns) / 1000; + large = line_count * line_size; + + entries = DIV_ROUND_UP(min(small, large), display->cacheline_size); + *sprite_wm = entries + display->guard_size; + + return *sprite_wm > 0x3ff ? false : true; +} + +static void sandybridge_update_sprite_wm(struct drm_device *dev, int pipe, + uint32_t sprite_width, int pixel_size) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + int latency = SNB_READ_WM0_LATENCY() * 100; /* In unit 0.1us */ + u32 val; + int sprite_wm, reg; + int ret; + + switch (pipe) { + case 0: + reg = WM0_PIPEA_ILK; + break; + case 1: + reg = WM0_PIPEB_ILK; + break; + case 2: + reg = WM0_PIPEC_IVB; + break; + default: + return; /* bad pipe */ + } + + ret = sandybridge_compute_sprite_wm(dev, pipe, sprite_width, pixel_size, + &sandybridge_display_wm_info, + latency, &sprite_wm); + if (!ret) { + DRM_DEBUG_KMS("failed to compute sprite wm for pipe %d\n", + pipe); + return; + } + + val = I915_READ(reg); + val &= ~WM0_PIPE_SPRITE_MASK; + I915_WRITE(reg, val | (sprite_wm << WM0_PIPE_SPRITE_SHIFT)); + DRM_DEBUG_KMS("sprite watermarks For pipe %d - %d\n", pipe, sprite_wm); + + + ret = sandybridge_compute_sprite_srwm(dev, pipe, sprite_width, + pixel_size, + &sandybridge_display_srwm_info, + SNB_READ_WM1_LATENCY() * 500, + &sprite_wm); + if (!ret) { + DRM_DEBUG_KMS("failed to compute sprite lp1 wm on pipe %d\n", + pipe); + return; + } + I915_WRITE(WM1S_LP_ILK, sprite_wm); + + /* Only IVB has two more LP watermarks for sprite */ + if (!IS_IVYBRIDGE(dev)) + return; + + ret = sandybridge_compute_sprite_srwm(dev, pipe, sprite_width, + pixel_size, + &sandybridge_display_srwm_info, + SNB_READ_WM2_LATENCY() * 500, + &sprite_wm); + if (!ret) { + DRM_DEBUG_KMS("failed to compute sprite lp2 wm on pipe %d\n", + pipe); + return; + } + I915_WRITE(WM2S_LP_IVB, sprite_wm); + + ret = sandybridge_compute_sprite_srwm(dev, pipe, sprite_width, + pixel_size, + &sandybridge_display_srwm_info, + SNB_READ_WM3_LATENCY() * 500, + &sprite_wm); + if (!ret) { + DRM_DEBUG_KMS("failed to compute sprite lp3 wm on pipe %d\n", + pipe); + return; + } + I915_WRITE(WM3S_LP_IVB, sprite_wm); +} + /** * intel_update_watermarks - update FIFO watermark values based on current modes * @@ -4677,6 +4903,16 @@ static void intel_update_watermarks(struct drm_device *dev) dev_priv->display.update_wm(dev); } +void intel_update_sprite_watermarks(struct drm_device *dev, int pipe, + uint32_t sprite_width, int pixel_size) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + + if (dev_priv->display.update_sprite_wm) + dev_priv->display.update_sprite_wm(dev, pipe, sprite_width, + pixel_size); +} + static inline bool intel_panel_use_ssc(struct drm_i915_private *dev_priv) { if (i915_panel_use_ssc >= 0) @@ -4824,6 +5060,82 @@ static bool intel_choose_pipe_bpp_dither(struct drm_crtc *crtc, return display_bpc != bpc; } +static int i9xx_get_refclk(struct drm_crtc *crtc, int num_connectors) +{ + struct drm_device *dev = crtc->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + int refclk; + + if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) && + intel_panel_use_ssc(dev_priv) && num_connectors < 2) { + refclk = dev_priv->lvds_ssc_freq * 1000; + DRM_DEBUG_KMS("using SSC reference clock of %d MHz\n", + refclk / 1000); + } else if (!IS_GEN2(dev)) { + refclk = 96000; + } else { + refclk = 48000; + } + + return refclk; +} + +static void i9xx_adjust_sdvo_tv_clock(struct drm_display_mode *adjusted_mode, + intel_clock_t *clock) +{ + /* SDVO TV has fixed PLL values depend on its clock range, + this mirrors vbios setting. */ + if (adjusted_mode->clock >= 100000 + && adjusted_mode->clock < 140500) { + clock->p1 = 2; + clock->p2 = 10; + clock->n = 3; + clock->m1 = 16; + clock->m2 = 8; + } else if (adjusted_mode->clock >= 140500 + && adjusted_mode->clock <= 200000) { + clock->p1 = 1; + clock->p2 = 10; + clock->n = 6; + clock->m1 = 12; + clock->m2 = 8; + } +} + +static void i9xx_update_pll_dividers(struct drm_crtc *crtc, + intel_clock_t *clock, + intel_clock_t *reduced_clock) +{ + struct drm_device *dev = crtc->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_crtc *intel_crtc = to_intel_crtc(crtc); + int pipe = intel_crtc->pipe; + u32 fp, fp2 = 0; + + if (IS_PINEVIEW(dev)) { + fp = (1 << clock->n) << 16 | clock->m1 << 8 | clock->m2; + if (reduced_clock) + fp2 = (1 << reduced_clock->n) << 16 | + reduced_clock->m1 << 8 | reduced_clock->m2; + } else { + fp = clock->n << 16 | clock->m1 << 8 | clock->m2; + if (reduced_clock) + fp2 = reduced_clock->n << 16 | reduced_clock->m1 << 8 | + reduced_clock->m2; + } + + I915_WRITE(FP0(pipe), fp); + + intel_crtc->lowfreq_avail = false; + if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) && + reduced_clock && i915_powersave) { + I915_WRITE(FP1(pipe), fp2); + intel_crtc->lowfreq_avail = true; + } else { + I915_WRITE(FP1(pipe), fp); + } +} + static int i9xx_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode, @@ -4837,7 +5149,7 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc, int plane = intel_crtc->plane; int refclk, num_connectors = 0; intel_clock_t clock, reduced_clock; - u32 dpll, fp = 0, fp2 = 0, dspcntr, pipeconf; + u32 dpll, dspcntr, pipeconf, vsyncshift; bool ok, has_reduced_clock = false, is_sdvo = false, is_dvo = false; bool is_crt = false, is_lvds = false, is_tv = false, is_dp = false; struct drm_mode_config *mode_config = &dev->mode_config; @@ -4878,15 +5190,7 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc, num_connectors++; } - if (is_lvds && intel_panel_use_ssc(dev_priv) && num_connectors < 2) { - refclk = dev_priv->lvds_ssc_freq * 1000; - DRM_DEBUG_KMS("using SSC reference clock of %d MHz\n", - refclk / 1000); - } else if (!IS_GEN2(dev)) { - refclk = 96000; - } else { - refclk = 48000; - } + refclk = i9xx_get_refclk(crtc, num_connectors); /* * Returns a set of divisors for the desired target clock with the given @@ -4894,7 +5198,8 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc, * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2. */ limit = intel_limit(crtc, refclk); - ok = limit->find_pll(limit, crtc, adjusted_mode->clock, refclk, &clock); + ok = limit->find_pll(limit, crtc, adjusted_mode->clock, refclk, NULL, + &clock); if (!ok) { DRM_ERROR("Couldn't find PLL settings for mode!\n"); return -EINVAL; @@ -4904,53 +5209,24 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc, intel_crtc_update_cursor(crtc, true); if (is_lvds && dev_priv->lvds_downclock_avail) { + /* + * Ensure we match the reduced clock's P to the target clock. + * If the clocks don't match, we can't switch the display clock + * by using the FP0/FP1. In such case we will disable the LVDS + * downclock feature. + */ has_reduced_clock = limit->find_pll(limit, crtc, dev_priv->lvds_downclock, refclk, + &clock, &reduced_clock); - if (has_reduced_clock && (clock.p != reduced_clock.p)) { - /* - * If the different P is found, it means that we can't - * switch the display clock by using the FP0/FP1. - * In such case we will disable the LVDS downclock - * feature. - */ - DRM_DEBUG_KMS("Different P is found for " - "LVDS clock/downclock\n"); - has_reduced_clock = 0; - } - } - /* SDVO TV has fixed PLL values depend on its clock range, - this mirrors vbios setting. */ - if (is_sdvo && is_tv) { - if (adjusted_mode->clock >= 100000 - && adjusted_mode->clock < 140500) { - clock.p1 = 2; - clock.p2 = 10; - clock.n = 3; - clock.m1 = 16; - clock.m2 = 8; - } else if (adjusted_mode->clock >= 140500 - && adjusted_mode->clock <= 200000) { - clock.p1 = 1; - clock.p2 = 10; - clock.n = 6; - clock.m1 = 12; - clock.m2 = 8; - } } - if (IS_PINEVIEW(dev)) { - fp = (1 << clock.n) << 16 | clock.m1 << 8 | clock.m2; - if (has_reduced_clock) - fp2 = (1 << reduced_clock.n) << 16 | - reduced_clock.m1 << 8 | reduced_clock.m2; - } else { - fp = clock.n << 16 | clock.m1 << 8 | clock.m2; - if (has_reduced_clock) - fp2 = reduced_clock.n << 16 | reduced_clock.m1 << 8 | - reduced_clock.m2; - } + if (is_sdvo && is_tv) + i9xx_adjust_sdvo_tv_clock(adjusted_mode, &clock); + + i9xx_update_pll_dividers(crtc, &clock, has_reduced_clock ? + &reduced_clock : NULL); dpll = DPLL_VGA_MODE_DIS; @@ -5024,8 +5300,6 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc, /* Set up the display plane register */ dspcntr = DISPPLANE_GAMMA_ENABLE; - /* Ironlake's plane is forced to pipe, bit 24 is to - enable color space conversion */ if (pipe == 0) dspcntr &= ~DISPPLANE_SEL_PIPE_MASK; else @@ -5060,7 +5334,6 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc, DRM_DEBUG_KMS("Mode for pipe %c:\n", pipe == 0 ? 'A' : 'B'); drm_mode_debug_printmodeline(mode); - I915_WRITE(FP0(pipe), fp); I915_WRITE(DPLL(pipe), dpll & ~DPLL_VCO_ENABLE); POSTING_READ(DPLL(pipe)); @@ -5147,33 +5420,32 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc, I915_WRITE(DPLL(pipe), dpll); } - intel_crtc->lowfreq_avail = false; - if (is_lvds && has_reduced_clock && i915_powersave) { - I915_WRITE(FP1(pipe), fp2); - intel_crtc->lowfreq_avail = true; - if (HAS_PIPE_CXSR(dev)) { + if (HAS_PIPE_CXSR(dev)) { + if (intel_crtc->lowfreq_avail) { DRM_DEBUG_KMS("enabling CxSR downclocking\n"); pipeconf |= PIPECONF_CXSR_DOWNCLOCK; - } - } else { - I915_WRITE(FP1(pipe), fp); - if (HAS_PIPE_CXSR(dev)) { + } else { DRM_DEBUG_KMS("disabling CxSR downclocking\n"); pipeconf &= ~PIPECONF_CXSR_DOWNCLOCK; } } - if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) { + pipeconf &= ~PIPECONF_INTERLACE_MASK; + if (!IS_GEN2(dev) && + adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) { pipeconf |= PIPECONF_INTERLACE_W_FIELD_INDICATION; /* the chip adds 2 halflines automatically */ - adjusted_mode->crtc_vdisplay -= 1; adjusted_mode->crtc_vtotal -= 1; - adjusted_mode->crtc_vblank_start -= 1; adjusted_mode->crtc_vblank_end -= 1; - adjusted_mode->crtc_vsync_end -= 1; - adjusted_mode->crtc_vsync_start -= 1; - } else - pipeconf &= ~PIPECONF_INTERLACE_W_FIELD_INDICATION; /* progressive */ + vsyncshift = adjusted_mode->crtc_hsync_start + - adjusted_mode->crtc_htotal/2; + } else { + pipeconf |= PIPECONF_PROGRESSIVE; + vsyncshift = 0; + } + + if (!IS_GEN3(dev)) + I915_WRITE(VSYNCSHIFT(pipe), vsyncshift); I915_WRITE(HTOTAL(pipe), (adjusted_mode->crtc_hdisplay - 1) | @@ -5290,7 +5562,8 @@ void ironlake_init_pch_refclk(struct drm_device *dev) if (intel_panel_use_ssc(dev_priv) && can_ssc) { DRM_DEBUG_KMS("Using SSC on panel\n"); temp |= DREF_SSC1_ENABLE; - } + } else + temp &= ~DREF_SSC1_ENABLE; /* Get SSC going before enabling the outputs */ I915_WRITE(PCH_DREF_CONTROL, temp); @@ -5439,7 +5712,8 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc, * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2. */ limit = intel_limit(crtc, refclk); - ok = limit->find_pll(limit, crtc, adjusted_mode->clock, refclk, &clock); + ok = limit->find_pll(limit, crtc, adjusted_mode->clock, refclk, NULL, + &clock); if (!ok) { DRM_ERROR("Couldn't find PLL settings for mode!\n"); return -EINVAL; @@ -5449,21 +5723,17 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc, intel_crtc_update_cursor(crtc, true); if (is_lvds && dev_priv->lvds_downclock_avail) { + /* + * Ensure we match the reduced clock's P to the target clock. + * If the clocks don't match, we can't switch the display clock + * by using the FP0/FP1. In such case we will disable the LVDS + * downclock feature. + */ has_reduced_clock = limit->find_pll(limit, crtc, dev_priv->lvds_downclock, refclk, + &clock, &reduced_clock); - if (has_reduced_clock && (clock.p != reduced_clock.p)) { - /* - * If the different P is found, it means that we can't - * switch the display clock by using the FP0/FP1. - * In such case we will disable the LVDS downclock - * feature. - */ - DRM_DEBUG_KMS("Different P is found for " - "LVDS clock/downclock\n"); - has_reduced_clock = 0; - } } /* SDVO TV has fixed PLL values depend on its clock range, this mirrors vbios setting. */ @@ -5758,17 +6028,19 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc, } } + pipeconf &= ~PIPECONF_INTERLACE_MASK; if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) { - pipeconf |= PIPECONF_INTERLACE_W_FIELD_INDICATION; + pipeconf |= PIPECONF_INTERLACED_ILK; /* the chip adds 2 halflines automatically */ - adjusted_mode->crtc_vdisplay -= 1; adjusted_mode->crtc_vtotal -= 1; - adjusted_mode->crtc_vblank_start -= 1; adjusted_mode->crtc_vblank_end -= 1; - adjusted_mode->crtc_vsync_end -= 1; - adjusted_mode->crtc_vsync_start -= 1; - } else - pipeconf &= ~PIPECONF_INTERLACE_W_FIELD_INDICATION; /* progressive */ + I915_WRITE(VSYNCSHIFT(pipe), + adjusted_mode->crtc_hsync_start + - adjusted_mode->crtc_htotal/2); + } else { + pipeconf |= PIPECONF_PROGRESSIVE; + I915_WRITE(VSYNCSHIFT(pipe), 0); + } I915_WRITE(HTOTAL(pipe), (adjusted_mode->crtc_hdisplay - 1) | @@ -5811,12 +6083,6 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc, intel_wait_for_vblank(dev, pipe); - if (IS_GEN5(dev)) { - /* enable address swizzle for tiling buffer */ - temp = I915_READ(DISP_ARB_CTL); - I915_WRITE(DISP_ARB_CTL, temp | DISP_TILE_SURFACE_SWIZZLING); - } - I915_WRITE(DSPCNTR(plane), dspcntr); POSTING_READ(DSPCNTR(plane)); @@ -5843,14 +6109,45 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc, ret = dev_priv->display.crtc_mode_set(crtc, mode, adjusted_mode, x, y, old_fb); - drm_vblank_post_modeset(dev, pipe); - intel_crtc->dpms_mode = DRM_MODE_DPMS_ON; + if (ret) + intel_crtc->dpms_mode = DRM_MODE_DPMS_OFF; + else + intel_crtc->dpms_mode = DRM_MODE_DPMS_ON; return ret; } +static bool intel_eld_uptodate(struct drm_connector *connector, + int reg_eldv, uint32_t bits_eldv, + int reg_elda, uint32_t bits_elda, + int reg_edid) +{ + struct drm_i915_private *dev_priv = connector->dev->dev_private; + uint8_t *eld = connector->eld; + uint32_t i; + + i = I915_READ(reg_eldv); + i &= bits_eldv; + + if (!eld[0]) + return !i; + + if (!i) + return false; + + i = I915_READ(reg_elda); + i &= ~bits_elda; + I915_WRITE(reg_elda, i); + + for (i = 0; i < eld[2]; i++) + if (I915_READ(reg_edid) != *((uint32_t *)eld + i)) + return false; + + return true; +} + static void g4x_write_eld(struct drm_connector *connector, struct drm_crtc *crtc) { @@ -5867,6 +6164,12 @@ static void g4x_write_eld(struct drm_connector *connector, else eldv = G4X_ELDV_DEVCTG; + if (intel_eld_uptodate(connector, + G4X_AUD_CNTL_ST, eldv, + G4X_AUD_CNTL_ST, G4X_ELD_ADDR, + G4X_HDMIW_HDMIEDID)) + return; + i = I915_READ(G4X_AUD_CNTL_ST); i &= ~(eldv | G4X_ELD_ADDR); len = (i >> 9) & 0x1f; /* ELD buffer size */ @@ -5894,22 +6197,26 @@ static void ironlake_write_eld(struct drm_connector *connector, uint32_t i; int len; int hdmiw_hdmiedid; + int aud_config; int aud_cntl_st; int aud_cntrl_st2; if (HAS_PCH_IBX(connector->dev)) { - hdmiw_hdmiedid = GEN5_HDMIW_HDMIEDID_A; - aud_cntl_st = GEN5_AUD_CNTL_ST_A; - aud_cntrl_st2 = GEN5_AUD_CNTL_ST2; + hdmiw_hdmiedid = IBX_HDMIW_HDMIEDID_A; + aud_config = IBX_AUD_CONFIG_A; + aud_cntl_st = IBX_AUD_CNTL_ST_A; + aud_cntrl_st2 = IBX_AUD_CNTL_ST2; } else { - hdmiw_hdmiedid = GEN7_HDMIW_HDMIEDID_A; - aud_cntl_st = GEN7_AUD_CNTRL_ST_A; - aud_cntrl_st2 = GEN7_AUD_CNTRL_ST2; + hdmiw_hdmiedid = CPT_HDMIW_HDMIEDID_A; + aud_config = CPT_AUD_CONFIG_A; + aud_cntl_st = CPT_AUD_CNTL_ST_A; + aud_cntrl_st2 = CPT_AUD_CNTRL_ST2; } i = to_intel_crtc(crtc)->pipe; hdmiw_hdmiedid += i * 0x100; aud_cntl_st += i * 0x100; + aud_config += i * 0x100; DRM_DEBUG_DRIVER("ELD on pipe %c\n", pipe_name(i)); @@ -5918,14 +6225,27 @@ static void ironlake_write_eld(struct drm_connector *connector, if (!i) { DRM_DEBUG_DRIVER("Audio directed to unknown port\n"); /* operate blindly on all ports */ - eldv = GEN5_ELD_VALIDB; - eldv |= GEN5_ELD_VALIDB << 4; - eldv |= GEN5_ELD_VALIDB << 8; + eldv = IBX_ELD_VALIDB; + eldv |= IBX_ELD_VALIDB << 4; + eldv |= IBX_ELD_VALIDB << 8; } else { DRM_DEBUG_DRIVER("ELD on port %c\n", 'A' + i); - eldv = GEN5_ELD_VALIDB << ((i - 1) * 4); + eldv = IBX_ELD_VALIDB << ((i - 1) * 4); } + if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT)) { + DRM_DEBUG_DRIVER("ELD: DisplayPort detected\n"); + eld[5] |= (1 << 2); /* Conn_Type, 0x1 = DisplayPort */ + I915_WRITE(aud_config, AUD_CONFIG_N_VALUE_INDEX); /* 0x1 = DP */ + } else + I915_WRITE(aud_config, 0); + + if (intel_eld_uptodate(connector, + aud_cntrl_st2, eldv, + aud_cntl_st, IBX_ELD_ADDRESS, + hdmiw_hdmiedid)) + return; + i = I915_READ(aud_cntrl_st2); i &= ~eldv; I915_WRITE(aud_cntrl_st2, i); @@ -5933,13 +6253,8 @@ static void ironlake_write_eld(struct drm_connector *connector, if (!eld[0]) return; - if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT)) { - DRM_DEBUG_DRIVER("ELD: DisplayPort detected\n"); - eld[5] |= (1 << 2); /* Conn_Type, 0x1 = DisplayPort */ - } - i = I915_READ(aud_cntl_st); - i &= ~GEN5_ELD_ADDRESS; + i &= ~IBX_ELD_ADDRESS; I915_WRITE(aud_cntl_st, i); len = min_t(uint8_t, eld[2], 21); /* 84 bytes of hw ELD buffer */ @@ -6319,7 +6634,7 @@ static struct drm_display_mode load_detect_mode = { static struct drm_framebuffer * intel_framebuffer_create(struct drm_device *dev, - struct drm_mode_fb_cmd *mode_cmd, + struct drm_mode_fb_cmd2 *mode_cmd, struct drm_i915_gem_object *obj) { struct intel_framebuffer *intel_fb; @@ -6361,7 +6676,7 @@ intel_framebuffer_create_for_mode(struct drm_device *dev, int depth, int bpp) { struct drm_i915_gem_object *obj; - struct drm_mode_fb_cmd mode_cmd; + struct drm_mode_fb_cmd2 mode_cmd; obj = i915_gem_alloc_object(dev, intel_framebuffer_size_for_mode(mode, bpp)); @@ -6370,9 +6685,9 @@ intel_framebuffer_create_for_mode(struct drm_device *dev, mode_cmd.width = mode->hdisplay; mode_cmd.height = mode->vdisplay; - mode_cmd.depth = depth; - mode_cmd.bpp = bpp; - mode_cmd.pitch = intel_framebuffer_pitch_for_width(mode_cmd.width, bpp); + mode_cmd.pitches[0] = intel_framebuffer_pitch_for_width(mode_cmd.width, + bpp); + mode_cmd.pixel_format = drm_mode_legacy_fb_format(bpp, depth); return intel_framebuffer_create(dev, &mode_cmd, obj); } @@ -6393,11 +6708,11 @@ mode_fits_in_fbdev(struct drm_device *dev, return NULL; fb = &dev_priv->fbdev->ifb.base; - if (fb->pitch < intel_framebuffer_pitch_for_width(mode->hdisplay, - fb->bits_per_pixel)) + if (fb->pitches[0] < intel_framebuffer_pitch_for_width(mode->hdisplay, + fb->bits_per_pixel)) return NULL; - if (obj->base.size < mode->vdisplay * fb->pitch) + if (obj->base.size < mode->vdisplay * fb->pitches[0]) return NULL; return fb; @@ -6729,9 +7044,7 @@ static void intel_increase_pllclock(struct drm_crtc *crtc) if (!HAS_PIPE_CXSR(dev) && (dpll & DISPLAY_RATE_SELECT_FPA1)) { DRM_DEBUG_DRIVER("upclocking LVDS\n"); - /* Unlock panel regs */ - I915_WRITE(PP_CONTROL, - I915_READ(PP_CONTROL) | PANEL_UNLOCK_REGS); + assert_panel_unlocked(dev_priv, pipe); dpll &= ~DISPLAY_RATE_SELECT_FPA1; I915_WRITE(dpll_reg, dpll); @@ -6740,9 +7053,6 @@ static void intel_increase_pllclock(struct drm_crtc *crtc) dpll = I915_READ(dpll_reg); if (dpll & DISPLAY_RATE_SELECT_FPA1) DRM_DEBUG_DRIVER("failed to upclock LVDS!\n"); - - /* ...and lock them again */ - I915_WRITE(PP_CONTROL, I915_READ(PP_CONTROL) & 0x3); } /* Schedule downclock */ @@ -6755,9 +7065,6 @@ static void intel_decrease_pllclock(struct drm_crtc *crtc) struct drm_device *dev = crtc->dev; drm_i915_private_t *dev_priv = dev->dev_private; struct intel_crtc *intel_crtc = to_intel_crtc(crtc); - int pipe = intel_crtc->pipe; - int dpll_reg = DPLL(pipe); - int dpll = I915_READ(dpll_reg); if (HAS_PCH_SPLIT(dev)) return; @@ -6770,23 +7077,22 @@ static void intel_decrease_pllclock(struct drm_crtc *crtc) * the manual case. */ if (!HAS_PIPE_CXSR(dev) && intel_crtc->lowfreq_avail) { + int pipe = intel_crtc->pipe; + int dpll_reg = DPLL(pipe); + u32 dpll; + DRM_DEBUG_DRIVER("downclocking LVDS\n"); - /* Unlock panel regs */ - I915_WRITE(PP_CONTROL, I915_READ(PP_CONTROL) | - PANEL_UNLOCK_REGS); + assert_panel_unlocked(dev_priv, pipe); + dpll = I915_READ(dpll_reg); dpll |= DISPLAY_RATE_SELECT_FPA1; I915_WRITE(dpll_reg, dpll); intel_wait_for_vblank(dev, pipe); dpll = I915_READ(dpll_reg); if (!(dpll & DISPLAY_RATE_SELECT_FPA1)) DRM_DEBUG_DRIVER("failed to downclock LVDS!\n"); - - /* ...and lock them again */ - I915_WRITE(PP_CONTROL, I915_READ(PP_CONTROL) & 0x3); } - } /** @@ -6899,7 +7205,7 @@ static void intel_unpin_work_fn(struct work_struct *__work) container_of(__work, struct intel_unpin_work, work); mutex_lock(&work->dev->struct_mutex); - i915_gem_object_unpin(work->old_fb_obj); + intel_unpin_fb_obj(work->old_fb_obj); drm_gem_object_unreference(&work->pending_flip_obj->base); drm_gem_object_unreference(&work->old_fb_obj->base); @@ -6927,18 +7233,11 @@ static void do_intel_finish_page_flip(struct drm_device *dev, spin_lock_irqsave(&dev->event_lock, flags); work = intel_crtc->unpin_work; - - /* Ensure we don't miss a work->pending update ... */ - smp_rmb(); - - if (work == NULL || atomic_read(&work->pending) < INTEL_FLIP_COMPLETE) { + if (work == NULL || !work->pending) { spin_unlock_irqrestore(&dev->event_lock, flags); return; } - /* and that the unpin work is consistent wrt ->pending. */ - smp_rmb(); - intel_crtc->unpin_work = NULL; if (work->event) { @@ -7010,25 +7309,16 @@ void intel_prepare_page_flip(struct drm_device *dev, int plane) to_intel_crtc(dev_priv->plane_to_crtc_mapping[plane]); unsigned long flags; - /* NB: An MMIO update of the plane base pointer will also - * generate a page-flip completion irq, i.e. every modeset - * is also accompanied by a spurious intel_prepare_page_flip(). - */ spin_lock_irqsave(&dev->event_lock, flags); - if (intel_crtc->unpin_work) - atomic_inc_not_zero(&intel_crtc->unpin_work->pending); + if (intel_crtc->unpin_work) { + if ((++intel_crtc->unpin_work->pending) > 1) + DRM_ERROR("Prepared flip multiple times\n"); + } else { + DRM_DEBUG_DRIVER("preparing flip with no unpin work?\n"); + } spin_unlock_irqrestore(&dev->event_lock, flags); } -inline static void intel_mark_page_flip_active(struct intel_crtc *intel_crtc) -{ - /* Ensure that the work item is consistent when activating it ... */ - smp_wmb(); - atomic_set(&intel_crtc->unpin_work->pending, INTEL_FLIP_PENDING); - /* and that it is marked active as soon as the irq could fire. */ - smp_wmb(); -} - static int intel_gen2_queue_flip(struct drm_device *dev, struct drm_crtc *crtc, struct drm_framebuffer *fb, @@ -7045,7 +7335,7 @@ static int intel_gen2_queue_flip(struct drm_device *dev, goto err; /* Offset into the new buffer for cases of shared fbs between CRTCs */ - offset = crtc->y * fb->pitch + crtc->x * fb->bits_per_pixel/8; + offset = crtc->y * fb->pitches[0] + crtc->x * fb->bits_per_pixel/8; ret = BEGIN_LP_RING(6); if (ret) @@ -7062,16 +7352,14 @@ static int intel_gen2_queue_flip(struct drm_device *dev, OUT_RING(MI_NOOP); OUT_RING(MI_DISPLAY_FLIP | MI_DISPLAY_FLIP_PLANE(intel_crtc->plane)); - OUT_RING(fb->pitch); + OUT_RING(fb->pitches[0]); OUT_RING(obj->gtt_offset + offset); - OUT_RING(MI_NOOP); - - intel_mark_page_flip_active(intel_crtc); + OUT_RING(0); /* aux display base address, unused */ ADVANCE_LP_RING(); return 0; err_unpin: - i915_gem_object_unpin(obj); + intel_unpin_fb_obj(obj); err: return ret; } @@ -7092,7 +7380,7 @@ static int intel_gen3_queue_flip(struct drm_device *dev, goto err; /* Offset into the new buffer for cases of shared fbs between CRTCs */ - offset = crtc->y * fb->pitch + crtc->x * fb->bits_per_pixel/8; + offset = crtc->y * fb->pitches[0] + crtc->x * fb->bits_per_pixel/8; ret = BEGIN_LP_RING(6); if (ret) @@ -7106,16 +7394,15 @@ static int intel_gen3_queue_flip(struct drm_device *dev, OUT_RING(MI_NOOP); OUT_RING(MI_DISPLAY_FLIP_I915 | MI_DISPLAY_FLIP_PLANE(intel_crtc->plane)); - OUT_RING(fb->pitch); + OUT_RING(fb->pitches[0]); OUT_RING(obj->gtt_offset + offset); OUT_RING(MI_NOOP); - intel_mark_page_flip_active(intel_crtc); ADVANCE_LP_RING(); return 0; err_unpin: - i915_gem_object_unpin(obj); + intel_unpin_fb_obj(obj); err: return ret; } @@ -7144,7 +7431,7 @@ static int intel_gen4_queue_flip(struct drm_device *dev, */ OUT_RING(MI_DISPLAY_FLIP | MI_DISPLAY_FLIP_PLANE(intel_crtc->plane)); - OUT_RING(fb->pitch); + OUT_RING(fb->pitches[0]); OUT_RING(obj->gtt_offset | obj->tiling_mode); /* XXX Enabling the panel-fitter across page-flip is so far @@ -7154,13 +7441,11 @@ static int intel_gen4_queue_flip(struct drm_device *dev, pf = 0; pipesrc = I915_READ(PIPESRC(intel_crtc->pipe)) & 0x0fff0fff; OUT_RING(pf | pipesrc); - - intel_mark_page_flip_active(intel_crtc); ADVANCE_LP_RING(); return 0; err_unpin: - i915_gem_object_unpin(obj); + intel_unpin_fb_obj(obj); err: return ret; } @@ -7185,19 +7470,23 @@ static int intel_gen6_queue_flip(struct drm_device *dev, OUT_RING(MI_DISPLAY_FLIP | MI_DISPLAY_FLIP_PLANE(intel_crtc->plane)); - OUT_RING(fb->pitch | obj->tiling_mode); + OUT_RING(fb->pitches[0] | obj->tiling_mode); OUT_RING(obj->gtt_offset); - pf = I915_READ(PF_CTL(intel_crtc->pipe)) & PF_ENABLE; + /* Contrary to the suggestions in the documentation, + * "Enable Panel Fitter" does not seem to be required when page + * flipping with a non-native mode, and worse causes a normal + * modeset to fail. + * pf = I915_READ(PF_CTL(intel_crtc->pipe)) & PF_ENABLE; + */ + pf = 0; pipesrc = I915_READ(PIPESRC(intel_crtc->pipe)) & 0x0fff0fff; OUT_RING(pf | pipesrc); - - intel_mark_page_flip_active(intel_crtc); ADVANCE_LP_RING(); return 0; err_unpin: - i915_gem_object_unpin(obj); + intel_unpin_fb_obj(obj); err: return ret; } @@ -7244,16 +7533,14 @@ static int intel_gen7_queue_flip(struct drm_device *dev, goto err_unpin; intel_ring_emit(ring, MI_DISPLAY_FLIP_I915 | plane_bit); - intel_ring_emit(ring, (fb->pitch | obj->tiling_mode)); + intel_ring_emit(ring, (fb->pitches[0] | obj->tiling_mode)); intel_ring_emit(ring, (obj->gtt_offset)); intel_ring_emit(ring, (MI_NOOP)); - - intel_mark_page_flip_active(intel_crtc); intel_ring_advance(ring); return 0; err_unpin: - i915_gem_object_unpin(obj); + intel_unpin_fb_obj(obj); err: return ret; } @@ -7551,10 +7838,9 @@ static void intel_setup_outputs(struct drm_device *dev) struct drm_i915_private *dev_priv = dev->dev_private; struct intel_encoder *encoder; bool dpd_is_edp = false; - bool has_lvds = false; + bool has_lvds; - if (IS_MOBILE(dev) && !IS_I830(dev)) - has_lvds = intel_lvds_init(dev); + has_lvds = intel_lvds_init(dev); if (!has_lvds && !HAS_PCH_SPLIT(dev)) { /* disable the panel fitter on everything but LVDS */ I915_WRITE(PFIT_CONTROL, 0); @@ -7683,7 +7969,7 @@ static const struct drm_framebuffer_funcs intel_fb_funcs = { int intel_framebuffer_init(struct drm_device *dev, struct intel_framebuffer *intel_fb, - struct drm_mode_fb_cmd *mode_cmd, + struct drm_mode_fb_cmd2 *mode_cmd, struct drm_i915_gem_object *obj) { int ret; @@ -7691,21 +7977,27 @@ int intel_framebuffer_init(struct drm_device *dev, if (obj->tiling_mode == I915_TILING_Y) return -EINVAL; - if (mode_cmd->pitch & 63) + if (mode_cmd->pitches[0] & 63) return -EINVAL; - switch (mode_cmd->bpp) { - case 8: - case 16: - /* Only pre-ILK can handle 5:5:5 */ - if (mode_cmd->depth == 15 && !HAS_PCH_SPLIT(dev)) - return -EINVAL; + switch (mode_cmd->pixel_format) { + case DRM_FORMAT_RGB332: + case DRM_FORMAT_RGB565: + case DRM_FORMAT_XRGB8888: + case DRM_FORMAT_XBGR8888: + case DRM_FORMAT_ARGB8888: + case DRM_FORMAT_XRGB2101010: + case DRM_FORMAT_ARGB2101010: + /* RGB formats are common across chipsets */ break; - - case 24: - case 32: + case DRM_FORMAT_YUYV: + case DRM_FORMAT_UYVY: + case DRM_FORMAT_YVYU: + case DRM_FORMAT_VYUY: break; default: + DRM_DEBUG_KMS("unsupported pixel format %u\n", + mode_cmd->pixel_format); return -EINVAL; } @@ -7723,11 +8015,12 @@ int intel_framebuffer_init(struct drm_device *dev, static struct drm_framebuffer * intel_user_framebuffer_create(struct drm_device *dev, struct drm_file *filp, - struct drm_mode_fb_cmd *mode_cmd) + struct drm_mode_fb_cmd2 *mode_cmd) { struct drm_i915_gem_object *obj; - obj = to_intel_bo(drm_gem_object_lookup(dev, filp, mode_cmd->handle)); + obj = to_intel_bo(drm_gem_object_lookup(dev, filp, + mode_cmd->handles[0])); if (&obj->base == NULL) return ERR_PTR(-ENOENT); @@ -7996,7 +8289,7 @@ void intel_init_emon(struct drm_device *dev) dev_priv->corr = (lcfuse & LCFUSE_HIV_MASK); } -static bool intel_enable_rc6(struct drm_device *dev) +static int intel_enable_rc6(struct drm_device *dev) { /* * Respect the kernel parameter if it is set @@ -8014,11 +8307,11 @@ static bool intel_enable_rc6(struct drm_device *dev) * Disable rc6 on Sandybridge */ if (INTEL_INFO(dev)->gen == 6) { - DRM_DEBUG_DRIVER("Sandybridge: RC6 disabled\n"); - return 0; + DRM_DEBUG_DRIVER("Sandybridge: deep RC6 disabled\n"); + return INTEL_RC6_ENABLE; } - DRM_DEBUG_DRIVER("RC6 enabled\n"); - return 1; + DRM_DEBUG_DRIVER("RC6 and deep RC6 enabled\n"); + return (INTEL_RC6_ENABLE | INTEL_RC6p_ENABLE); } void gen6_enable_rps(struct drm_i915_private *dev_priv) @@ -8026,7 +8319,9 @@ void gen6_enable_rps(struct drm_i915_private *dev_priv) u32 rp_state_cap = I915_READ(GEN6_RP_STATE_CAP); u32 gt_perf_status = I915_READ(GEN6_GT_PERF_STATUS); u32 pcu_mbox, rc6_mask = 0; + u32 gtfifodbg; int cur_freq, min_freq, max_freq; + int rc6_mode; int i; /* Here begins a magic sequence of register writes to enable @@ -8037,6 +8332,13 @@ void gen6_enable_rps(struct drm_i915_private *dev_priv) */ I915_WRITE(GEN6_RC_STATE, 0); mutex_lock(&dev_priv->dev->struct_mutex); + + /* Clear the DBG now so we don't confuse earlier errors */ + if ((gtfifodbg = I915_READ(GTFIFODBG))) { + DRM_ERROR("GT fifo had a previous error %x\n", gtfifodbg); + I915_WRITE(GTFIFODBG, gtfifodbg); + } + gen6_gt_force_wake_get(dev_priv); /* disable the counters and set deterministic thresholds */ @@ -8057,9 +8359,20 @@ void gen6_enable_rps(struct drm_i915_private *dev_priv) I915_WRITE(GEN6_RC6p_THRESHOLD, 100000); I915_WRITE(GEN6_RC6pp_THRESHOLD, 64000); /* unused */ - if (intel_enable_rc6(dev_priv->dev)) - rc6_mask = GEN6_RC_CTL_RC6_ENABLE | - ((IS_GEN7(dev_priv->dev)) ? GEN6_RC_CTL_RC6p_ENABLE : 0); + rc6_mode = intel_enable_rc6(dev_priv->dev); + if (rc6_mode & INTEL_RC6_ENABLE) + rc6_mask |= GEN6_RC_CTL_RC6_ENABLE; + + if (rc6_mode & INTEL_RC6p_ENABLE) + rc6_mask |= GEN6_RC_CTL_RC6p_ENABLE; + + if (rc6_mode & INTEL_RC6pp_ENABLE) + rc6_mask |= GEN6_RC_CTL_RC6pp_ENABLE; + + DRM_INFO("Enabling RC6 states: RC6 %s, RC6p %s, RC6pp %s\n", + (rc6_mode & INTEL_RC6_ENABLE) ? "on" : "off", + (rc6_mode & INTEL_RC6p_ENABLE) ? "on" : "off", + (rc6_mode & INTEL_RC6pp_ENABLE) ? "on" : "off"); I915_WRITE(GEN6_RC_CONTROL, rc6_mask | @@ -8287,6 +8600,10 @@ static void gen6_init_clock_gating(struct drm_device *dev) I915_WRITE(WM2_LP_ILK, 0); I915_WRITE(WM1_LP_ILK, 0); + I915_WRITE(GEN6_UCGCTL1, + I915_READ(GEN6_UCGCTL1) | + GEN6_BLBUNIT_CLOCK_GATE_DISABLE); + /* According to the BSpec vol1g, bit 12 (RCPBUNIT) clock * gating disable must be set. Failure to set it results in * flickering pixels due to Z write ordering failures after @@ -8365,6 +8682,10 @@ static void ivybridge_init_clock_gating(struct drm_device *dev) I915_WRITE(ILK_DSPCLK_GATE, IVB_VRHUNIT_CLK_GATE); + I915_WRITE(IVB_CHICKEN3, + CHICKEN3_DGMG_REQ_OUT_FIX_DISABLE | + CHICKEN3_DGMG_DONE_FIX_DISABLE); + /* Apply the WaDisableRHWOOptimizationForRenderHang workaround. */ I915_WRITE(GEN7_COMMON_SLICE_CHICKEN1, GEN7_CSC1_RHWO_OPT_DISABLE_IN_RCC); @@ -8675,9 +8996,15 @@ static void intel_init_display(struct drm_device *dev) if (IS_IVYBRIDGE(dev)) { u32 ecobus; + /* A small trick here - if the bios hasn't configured MT forcewake, + * and if the device is in RC6, then force_wake_mt_get will not wake + * the device and the ECOBUS read will return zero. Which will be + * (correctly) interpreted by the test below as MT forcewake being + * disabled. + */ mutex_lock(&dev->struct_mutex); __gen6_gt_force_wake_mt_get(dev_priv); - ecobus = I915_READ(ECOBUS); + ecobus = I915_READ_NOTRACE(ECOBUS); __gen6_gt_force_wake_mt_put(dev_priv); mutex_unlock(&dev->struct_mutex); @@ -8709,6 +9036,7 @@ static void intel_init_display(struct drm_device *dev) } else if (IS_GEN6(dev)) { if (SNB_READ_WM0_LATENCY()) { dev_priv->display.update_wm = sandybridge_update_wm; + dev_priv->display.update_sprite_wm = sandybridge_update_sprite_wm; } else { DRM_DEBUG_KMS("Failed to read display plane latency. " "Disable CxSR\n"); @@ -8722,6 +9050,7 @@ static void intel_init_display(struct drm_device *dev) dev_priv->display.fdi_link_train = ivb_manual_fdi_link_train; if (SNB_READ_WM0_LATENCY()) { dev_priv->display.update_wm = sandybridge_update_wm; + dev_priv->display.update_sprite_wm = sandybridge_update_sprite_wm; } else { DRM_DEBUG_KMS("Failed to read display plane latency. " "Disable CxSR\n"); @@ -8834,8 +9163,6 @@ struct intel_quirk { }; struct intel_quirk intel_quirks[] = { - /* HP Compaq 2730p needs pipe A force quirk (LP: #291555) */ - { 0x2a42, 0x103c, 0x30eb, quirk_pipea_force }, /* HP Mini needs pipe A force quirk (LP: #322104) */ { 0x27ae, 0x103c, 0x361a, quirk_pipea_force }, @@ -8902,33 +9229,19 @@ static void i915_disable_vga(struct drm_device *dev) POSTING_READ(vga_reg); } -void i915_redisable_vga(struct drm_device *dev) -{ - struct drm_i915_private *dev_priv = dev->dev_private; - u32 vga_reg; - - if (HAS_PCH_SPLIT(dev)) - vga_reg = CPU_VGACNTRL; - else - vga_reg = VGACNTRL; - - if (I915_READ(vga_reg) != VGA_DISP_DISABLE) { - DRM_DEBUG_KMS("Something enabled VGA plane, disabling it\n"); - I915_WRITE(vga_reg, VGA_DISP_DISABLE); - POSTING_READ(vga_reg); - } -} - void intel_modeset_init(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; - int i; + int i, ret; drm_mode_config_init(dev); dev->mode_config.min_width = 0; dev->mode_config.min_height = 0; + dev->mode_config.preferred_depth = 24; + dev->mode_config.prefer_shadow = 1; + dev->mode_config.funcs = (void *)&intel_mode_funcs; intel_init_quirks(dev); @@ -8952,6 +9265,9 @@ void intel_modeset_init(struct drm_device *dev) for (i = 0; i < dev_priv->num_pipe; i++) { intel_crtc_init(dev, i); + ret = intel_plane_init(dev, i); + if (ret) + DRM_DEBUG_KMS("plane %d init failed: %d\n", i, ret); } /* Just disable it once at startup */ diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index c8ecaab..069725c 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -49,7 +49,7 @@ struct intel_dp { uint32_t DP; uint8_t link_configuration[DP_LINK_CONFIGURATION_SIZE]; bool has_audio; - int force_audio; + enum hdmi_force_audio force_audio; uint32_t color_range; int dpms_mode; uint8_t link_bw; @@ -384,7 +384,7 @@ intel_dp_aux_ch(struct intel_dp *intel_dp, else aux_clock_divider = 225; /* eDP input clock at 450Mhz */ } else if (HAS_PCH_SPLIT(dev)) - aux_clock_divider = 62; /* IRL input clock fixed at 125Mhz */ + aux_clock_divider = 63; /* IRL input clock fixed at 125Mhz */ else aux_clock_divider = intel_hrawclk(dev) / 2; @@ -437,6 +437,10 @@ intel_dp_aux_ch(struct intel_dp *intel_dp, DP_AUX_CH_CTL_DONE | DP_AUX_CH_CTL_TIME_OUT_ERROR | DP_AUX_CH_CTL_RECEIVE_ERROR); + + if (status & (DP_AUX_CH_CTL_TIME_OUT_ERROR | + DP_AUX_CH_CTL_RECEIVE_ERROR)) + continue; if (status & DP_AUX_CH_CTL_DONE) break; } @@ -2137,8 +2141,8 @@ intel_dp_detect(struct drm_connector *connector, bool force) if (status != connector_status_connected) return status; - if (intel_dp->force_audio) { - intel_dp->has_audio = intel_dp->force_audio > 0; + if (intel_dp->force_audio != HDMI_AUDIO_AUTO) { + intel_dp->has_audio = (intel_dp->force_audio == HDMI_AUDIO_ON); } else { edid = intel_dp_get_edid(connector, &intel_dp->adapter); if (edid) { @@ -2238,10 +2242,10 @@ intel_dp_set_property(struct drm_connector *connector, intel_dp->force_audio = i; - if (i == 0) + if (i == HDMI_AUDIO_AUTO) has_audio = intel_dp_detect_audio(connector); else - has_audio = i > 0; + has_audio = (i == HDMI_AUDIO_ON); if (has_audio == intel_dp->has_audio) return 0; diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index 5212284..cd623e8 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -26,6 +26,7 @@ #define __INTEL_DRV_H__ #include +#include "i915_drm.h" #include "i915_drv.h" #include "drm_crtc.h" #include "drm_crtc_helper.h" @@ -39,7 +40,7 @@ ret__ = -ETIMEDOUT; \ break; \ } \ - if (W && !(in_atomic() || in_dbg_master())) msleep(W); \ + if (W && drm_can_sleep()) msleep(W); \ } \ ret__; \ }) @@ -47,13 +48,6 @@ #define wait_for(COND, MS) _wait_for(COND, MS, 1) #define wait_for_atomic(COND, MS) _wait_for(COND, MS, 0) -#define MSLEEP(x) do { \ - if (in_dbg_master()) \ - mdelay(x); \ - else \ - msleep(x); \ -} while (0) - #define KHz(x) (1000*x) #define MHz(x) KHz(1000*x) @@ -111,6 +105,10 @@ #define INTEL_MODE_PIXEL_MULTIPLIER_SHIFT (0x0) #define INTEL_MODE_PIXEL_MULTIPLIER_MASK (0xf << INTEL_MODE_PIXEL_MULTIPLIER_SHIFT) #define INTEL_MODE_DP_FORCE_6BPC (0x10) +/* This flag must be set by the encoder's mode_fixup if it changes the crtc + * timings in the mode to prevent the crtc fixup from overwriting them. + * Currently only lvds needs that. */ +#define INTEL_MODE_CRTC_TIMINGS_SET (0x20) static inline void intel_mode_set_pixel_multiplier(struct drm_display_mode *mode, @@ -177,10 +175,32 @@ struct intel_crtc { bool use_pll_a; }; +struct intel_plane { + struct drm_plane base; + enum pipe pipe; + struct drm_i915_gem_object *obj; + bool primary_disabled; + int max_downscale; + u32 lut_r[1024], lut_g[1024], lut_b[1024]; + void (*update_plane)(struct drm_plane *plane, + struct drm_framebuffer *fb, + struct drm_i915_gem_object *obj, + int crtc_x, int crtc_y, + unsigned int crtc_w, unsigned int crtc_h, + uint32_t x, uint32_t y, + uint32_t src_w, uint32_t src_h); + void (*disable_plane)(struct drm_plane *plane); + int (*update_colorkey)(struct drm_plane *plane, + struct drm_intel_sprite_colorkey *key); + void (*get_colorkey)(struct drm_plane *plane, + struct drm_intel_sprite_colorkey *key); +}; + #define to_intel_crtc(x) container_of(x, struct intel_crtc, base) #define to_intel_connector(x) container_of(x, struct intel_connector, base) #define to_intel_encoder(x) container_of(x, struct intel_encoder, base) #define to_intel_framebuffer(x) container_of(x, struct intel_framebuffer, base) +#define to_intel_plane(x) container_of(x, struct intel_plane, base) #define DIP_HEADER_SIZE 5 @@ -257,10 +277,7 @@ struct intel_unpin_work { struct drm_i915_gem_object *old_fb_obj; struct drm_i915_gem_object *pending_flip_obj; struct drm_pending_vblank_event *event; - atomic_t pending; -#define INTEL_FLIP_INACTIVE 0 -#define INTEL_FLIP_PENDING 1 -#define INTEL_FLIP_COMPLETE 2 + int pending; bool enable_stall_check; }; @@ -271,6 +288,8 @@ struct intel_fbc_work { int interval; }; +int intel_connector_update_modes(struct drm_connector *connector, + struct edid *edid); int intel_ddc_get_modes(struct drm_connector *c, struct i2c_adapter *adapter); extern bool intel_ddc_probe(struct intel_encoder *intel_encoder, int ddc_bus); @@ -293,6 +312,7 @@ intel_dp_set_m_n(struct drm_crtc *crtc, struct drm_display_mode *mode, extern bool intel_dpd_is_edp(struct drm_device *dev); extern void intel_edp_link_config(struct intel_encoder *, int *, int *); extern bool intel_encoder_is_pch_edp(struct drm_encoder *encoder); +extern int intel_plane_init(struct drm_device *dev, enum pipe pipe); /* intel_panel.c */ extern void intel_fixed_panel_mode(struct drm_display_mode *fixed_mode, @@ -360,10 +380,11 @@ extern void intel_init_emon(struct drm_device *dev); extern int intel_pin_and_fence_fb_obj(struct drm_device *dev, struct drm_i915_gem_object *obj, struct intel_ring_buffer *pipelined); +extern void intel_unpin_fb_obj(struct drm_i915_gem_object *obj); extern int intel_framebuffer_init(struct drm_device *dev, struct intel_framebuffer *ifb, - struct drm_mode_fb_cmd *mode_cmd, + struct drm_mode_fb_cmd2 *mode_cmd, struct drm_i915_gem_object *obj); extern int intel_fbdev_init(struct drm_device *dev); extern void intel_fbdev_fini(struct drm_device *dev); @@ -383,9 +404,25 @@ extern int intel_overlay_attrs(struct drm_device *dev, void *data, extern void intel_fb_output_poll_changed(struct drm_device *dev); extern void intel_fb_restore_mode(struct drm_device *dev); +extern void assert_pipe(struct drm_i915_private *dev_priv, enum pipe pipe, + bool state); +#define assert_pipe_enabled(d, p) assert_pipe(d, p, true) +#define assert_pipe_disabled(d, p) assert_pipe(d, p, false) + extern void intel_init_clock_gating(struct drm_device *dev); extern void intel_write_eld(struct drm_encoder *encoder, struct drm_display_mode *mode); extern void intel_cpt_verify_modeset(struct drm_device *dev, int pipe); +/* For use by IVB LP watermark workaround in intel_sprite.c */ +extern void sandybridge_update_wm(struct drm_device *dev); +extern void intel_update_sprite_watermarks(struct drm_device *dev, int pipe, + uint32_t sprite_width, + int pixel_size); + +extern int intel_sprite_set_colorkey(struct drm_device *dev, void *data, + struct drm_file *file_priv); +extern int intel_sprite_get_colorkey(struct drm_device *dev, void *data, + struct drm_file *file_priv); + #endif /* __INTEL_DRV_H__ */ diff --git a/drivers/gpu/drm/i915/intel_dvo.c b/drivers/gpu/drm/i915/intel_dvo.c index 6eda1b5..020a7d7 100644 --- a/drivers/gpu/drm/i915/intel_dvo.c +++ b/drivers/gpu/drm/i915/intel_dvo.c @@ -157,7 +157,6 @@ static bool intel_dvo_mode_fixup(struct drm_encoder *encoder, C(vsync_end); C(vtotal); C(clock); - drm_mode_set_crtcinfo(adjusted_mode, CRTC_INTERLACE_HALVE_V); #undef C } diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c index d0ce34b..6e9ee33 100644 --- a/drivers/gpu/drm/i915/intel_fb.c +++ b/drivers/gpu/drm/i915/intel_fb.c @@ -65,7 +65,7 @@ static int intelfb_create(struct intel_fbdev *ifbdev, struct drm_i915_private *dev_priv = dev->dev_private; struct fb_info *info; struct drm_framebuffer *fb; - struct drm_mode_fb_cmd mode_cmd; + struct drm_mode_fb_cmd2 mode_cmd; struct drm_i915_gem_object *obj; struct device *device = &dev->pdev->dev; int size, ret; @@ -77,11 +77,12 @@ static int intelfb_create(struct intel_fbdev *ifbdev, mode_cmd.width = sizes->surface_width; mode_cmd.height = sizes->surface_height; - mode_cmd.bpp = sizes->surface_bpp; - mode_cmd.pitch = ALIGN(mode_cmd.width * ((mode_cmd.bpp + 7) / 8), 64); - mode_cmd.depth = sizes->surface_depth; + mode_cmd.pitches[0] = ALIGN(mode_cmd.width * ((sizes->surface_bpp + 7) / + 8), 64); + mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp, + sizes->surface_depth); - size = mode_cmd.pitch * mode_cmd.height; + size = mode_cmd.pitches[0] * mode_cmd.height; size = ALIGN(size, PAGE_SIZE); obj = i915_gem_alloc_object(dev, size); if (!obj) { @@ -148,14 +149,10 @@ static int intelfb_create(struct intel_fbdev *ifbdev, // memset(info->screen_base, 0, size); - drm_fb_helper_fill_fix(info, fb->pitch, fb->depth); + drm_fb_helper_fill_fix(info, fb->pitches[0], fb->depth); drm_fb_helper_fill_var(info, &ifbdev->helper, sizes->fb_width, sizes->fb_height); - info->pixmap.size = 64*1024; - info->pixmap.buf_align = 8; - info->pixmap.access_align = 32; - info->pixmap.flags = FB_PIXMAP_SYSTEM; - info->pixmap.scan_align = 1; + /* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */ DRM_DEBUG_KMS("allocated %dx%d fb: 0x%08x, bo %p\n", fb->width, fb->height, @@ -279,8 +276,18 @@ void intel_fb_restore_mode(struct drm_device *dev) { int ret; drm_i915_private_t *dev_priv = dev->dev_private; + struct drm_mode_config *config = &dev->mode_config; + struct drm_plane *plane; + + mutex_lock(&dev->mode_config.mutex); ret = drm_fb_helper_restore_fbdev_mode(&dev_priv->fbdev->helper); if (ret) DRM_DEBUG("failed to restore crtc mode\n"); + + /* Be sure to shut off any planes that may be active */ + list_for_each_entry(plane, &config->plane_list, head) + plane->funcs->disable_plane(plane); + + mutex_unlock(&dev->mode_config.mutex); } diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c index 497da2a..c60100d 100644 --- a/drivers/gpu/drm/i915/intel_hdmi.c +++ b/drivers/gpu/drm/i915/intel_hdmi.c @@ -44,7 +44,7 @@ struct intel_hdmi { uint32_t color_range; bool has_hdmi_sink; bool has_audio; - int force_audio; + enum hdmi_force_audio force_audio; void (*write_infoframe)(struct drm_encoder *encoder, struct dip_infoframe *frame); }; @@ -143,9 +143,6 @@ static void i9xx_write_infoframe(struct drm_encoder *encoder, I915_WRITE(VIDEO_DIP_DATA, *data); data++; } - /* Write every possible data byte to force correct ECC calculation. */ - for (; i < VIDEO_DIP_DATA_SIZE; i += 4) - I915_WRITE(VIDEO_DIP_DATA, 0); mmiowb(); flags |= intel_infoframe_flags(frame); @@ -179,9 +176,6 @@ static void ironlake_write_infoframe(struct drm_encoder *encoder, I915_WRITE(TVIDEO_DIP_DATA(intel_crtc->pipe), *data); data++; } - /* Write every possible data byte to force correct ECC calculation. */ - for (; i < VIDEO_DIP_DATA_SIZE; i += 4) - I915_WRITE(TVIDEO_DIP_DATA(intel_crtc->pipe), 0); mmiowb(); flags |= intel_infoframe_flags(frame); @@ -351,7 +345,9 @@ intel_hdmi_detect(struct drm_connector *connector, bool force) if (edid) { if (edid->input & DRM_EDID_INPUT_DIGITAL) { status = connector_status_connected; - intel_hdmi->has_hdmi_sink = drm_detect_hdmi_monitor(edid); + if (intel_hdmi->force_audio != HDMI_AUDIO_OFF_DVI) + intel_hdmi->has_hdmi_sink = + drm_detect_hdmi_monitor(edid); intel_hdmi->has_audio = drm_detect_monitor_audio(edid); } connector->display_info.raw_edid = NULL; @@ -359,8 +355,9 @@ intel_hdmi_detect(struct drm_connector *connector, bool force) } if (status == connector_status_connected) { - if (intel_hdmi->force_audio) - intel_hdmi->has_audio = intel_hdmi->force_audio > 0; + if (intel_hdmi->force_audio != HDMI_AUDIO_AUTO) + intel_hdmi->has_audio = + (intel_hdmi->force_audio == HDMI_AUDIO_ON); } return status; @@ -414,7 +411,7 @@ intel_hdmi_set_property(struct drm_connector *connector, return ret; if (property == dev_priv->force_audio_property) { - int i = val; + enum hdmi_force_audio i = val; bool has_audio; if (i == intel_hdmi->force_audio) @@ -422,13 +419,13 @@ intel_hdmi_set_property(struct drm_connector *connector, intel_hdmi->force_audio = i; - if (i == 0) + if (i == HDMI_AUDIO_AUTO) has_audio = intel_hdmi_detect_audio(connector); else - has_audio = i > 0; + has_audio = (i == HDMI_AUDIO_ON); - if (has_audio == intel_hdmi->has_audio) - return 0; + if (i == HDMI_AUDIO_OFF_DVI) + intel_hdmi->has_hdmi_sink = 0; intel_hdmi->has_audio = has_audio; goto done; @@ -526,7 +523,7 @@ void intel_hdmi_init(struct drm_device *dev, int sdvox_reg) intel_encoder->type = INTEL_OUTPUT_HDMI; connector->polled = DRM_CONNECTOR_POLL_HPD; - connector->interlace_allowed = 0; + connector->interlace_allowed = 1; connector->doublescan_allowed = 0; intel_encoder->crtc_mask = (1 << 0) | (1 << 1) | (1 << 2); diff --git a/drivers/gpu/drm/i915/intel_i2c.c b/drivers/gpu/drm/i915/intel_i2c.c index d30cccc..8fdc957 100644 --- a/drivers/gpu/drm/i915/intel_i2c.c +++ b/drivers/gpu/drm/i915/intel_i2c.c @@ -37,7 +37,7 @@ /* Intel GPIO access functions */ -#define I2C_RISEFALL_TIME 20 +#define I2C_RISEFALL_TIME 10 static inline struct intel_gmbus * to_intel_gmbus(struct i2c_adapter *i2c) @@ -45,13 +45,6 @@ to_intel_gmbus(struct i2c_adapter *i2c) return container_of(i2c, struct intel_gmbus, adapter); } -struct intel_gpio { - struct i2c_adapter adapter; - struct i2c_algo_bit_data algo; - struct drm_i915_private *dev_priv; - u32 reg; -}; - void intel_i2c_reset(struct drm_device *dev) { @@ -78,15 +71,15 @@ static void intel_i2c_quirk_set(struct drm_i915_private *dev_priv, bool enable) I915_WRITE(DSPCLK_GATE_D, val); } -static u32 get_reserved(struct intel_gpio *gpio) +static u32 get_reserved(struct intel_gmbus *bus) { - struct drm_i915_private *dev_priv = gpio->dev_priv; + struct drm_i915_private *dev_priv = bus->dev_priv; struct drm_device *dev = dev_priv->dev; u32 reserved = 0; /* On most chips, these bits must be preserved in software. */ if (!IS_I830(dev) && !IS_845G(dev)) - reserved = I915_READ_NOTRACE(gpio->reg) & + reserved = I915_READ_NOTRACE(bus->gpio_reg) & (GPIO_DATA_PULLUP_DISABLE | GPIO_CLOCK_PULLUP_DISABLE); @@ -95,29 +88,29 @@ static u32 get_reserved(struct intel_gpio *gpio) static int get_clock(void *data) { - struct intel_gpio *gpio = data; - struct drm_i915_private *dev_priv = gpio->dev_priv; - u32 reserved = get_reserved(gpio); - I915_WRITE_NOTRACE(gpio->reg, reserved | GPIO_CLOCK_DIR_MASK); - I915_WRITE_NOTRACE(gpio->reg, reserved); - return (I915_READ_NOTRACE(gpio->reg) & GPIO_CLOCK_VAL_IN) != 0; + struct intel_gmbus *bus = data; + struct drm_i915_private *dev_priv = bus->dev_priv; + u32 reserved = get_reserved(bus); + I915_WRITE_NOTRACE(bus->gpio_reg, reserved | GPIO_CLOCK_DIR_MASK); + I915_WRITE_NOTRACE(bus->gpio_reg, reserved); + return (I915_READ_NOTRACE(bus->gpio_reg) & GPIO_CLOCK_VAL_IN) != 0; } static int get_data(void *data) { - struct intel_gpio *gpio = data; - struct drm_i915_private *dev_priv = gpio->dev_priv; - u32 reserved = get_reserved(gpio); - I915_WRITE_NOTRACE(gpio->reg, reserved | GPIO_DATA_DIR_MASK); - I915_WRITE_NOTRACE(gpio->reg, reserved); - return (I915_READ_NOTRACE(gpio->reg) & GPIO_DATA_VAL_IN) != 0; + struct intel_gmbus *bus = data; + struct drm_i915_private *dev_priv = bus->dev_priv; + u32 reserved = get_reserved(bus); + I915_WRITE_NOTRACE(bus->gpio_reg, reserved | GPIO_DATA_DIR_MASK); + I915_WRITE_NOTRACE(bus->gpio_reg, reserved); + return (I915_READ_NOTRACE(bus->gpio_reg) & GPIO_DATA_VAL_IN) != 0; } static void set_clock(void *data, int state_high) { - struct intel_gpio *gpio = data; - struct drm_i915_private *dev_priv = gpio->dev_priv; - u32 reserved = get_reserved(gpio); + struct intel_gmbus *bus = data; + struct drm_i915_private *dev_priv = bus->dev_priv; + u32 reserved = get_reserved(bus); u32 clock_bits; if (state_high) @@ -126,15 +119,15 @@ static void set_clock(void *data, int state_high) clock_bits = GPIO_CLOCK_DIR_OUT | GPIO_CLOCK_DIR_MASK | GPIO_CLOCK_VAL_MASK; - I915_WRITE_NOTRACE(gpio->reg, reserved | clock_bits); - POSTING_READ(gpio->reg); + I915_WRITE_NOTRACE(bus->gpio_reg, reserved | clock_bits); + POSTING_READ(bus->gpio_reg); } static void set_data(void *data, int state_high) { - struct intel_gpio *gpio = data; - struct drm_i915_private *dev_priv = gpio->dev_priv; - u32 reserved = get_reserved(gpio); + struct intel_gmbus *bus = data; + struct drm_i915_private *dev_priv = bus->dev_priv; + u32 reserved = get_reserved(bus); u32 data_bits; if (state_high) @@ -143,13 +136,14 @@ static void set_data(void *data, int state_high) data_bits = GPIO_DATA_DIR_OUT | GPIO_DATA_DIR_MASK | GPIO_DATA_VAL_MASK; - I915_WRITE_NOTRACE(gpio->reg, reserved | data_bits); - POSTING_READ(gpio->reg); + I915_WRITE_NOTRACE(bus->gpio_reg, reserved | data_bits); + POSTING_READ(bus->gpio_reg); } -static struct i2c_adapter * -intel_gpio_create(struct drm_i915_private *dev_priv, u32 pin) +static bool +intel_gpio_setup(struct intel_gmbus *bus, u32 pin) { + struct drm_i915_private *dev_priv = bus->dev_priv; static const int map_pin_to_reg[] = { 0, GPIOB, @@ -160,65 +154,48 @@ intel_gpio_create(struct drm_i915_private *dev_priv, u32 pin) 0, GPIOF, }; - struct intel_gpio *gpio; + struct i2c_algo_bit_data *algo; if (pin >= ARRAY_SIZE(map_pin_to_reg) || !map_pin_to_reg[pin]) - return NULL; + return false; - gpio = kzalloc(sizeof(struct intel_gpio), GFP_KERNEL); - if (gpio == NULL) - return NULL; + algo = &bus->bit_algo; - gpio->reg = map_pin_to_reg[pin]; + bus->gpio_reg = map_pin_to_reg[pin]; if (HAS_PCH_SPLIT(dev_priv->dev)) - gpio->reg += PCH_GPIOA - GPIOA; - gpio->dev_priv = dev_priv; - - snprintf(gpio->adapter.name, sizeof(gpio->adapter.name), - "i915 GPIO%c", "?BACDE?F"[pin]); - gpio->adapter.owner = THIS_MODULE; - gpio->adapter.algo_data = &gpio->algo; - gpio->adapter.dev.parent = &dev_priv->dev->pdev->dev; - gpio->algo.setsda = set_data; - gpio->algo.setscl = set_clock; - gpio->algo.getsda = get_data; - gpio->algo.getscl = get_clock; - gpio->algo.udelay = I2C_RISEFALL_TIME; - gpio->algo.timeout = usecs_to_jiffies(2200); - gpio->algo.data = gpio; - - if (i2c_bit_add_bus(&gpio->adapter)) - goto out_free; - - return &gpio->adapter; - -out_free: - kfree(gpio); - return NULL; + bus->gpio_reg += PCH_GPIOA - GPIOA; + + bus->adapter.algo_data = algo; + algo->setsda = set_data; + algo->setscl = set_clock; + algo->getsda = get_data; + algo->getscl = get_clock; + algo->udelay = I2C_RISEFALL_TIME; + algo->timeout = usecs_to_jiffies(2200); + algo->data = bus; + + return true; } static int -intel_i2c_quirk_xfer(struct drm_i915_private *dev_priv, - struct i2c_adapter *adapter, +intel_i2c_quirk_xfer(struct intel_gmbus *bus, struct i2c_msg *msgs, int num) { - struct intel_gpio *gpio = container_of(adapter, - struct intel_gpio, - adapter); + struct drm_i915_private *dev_priv = bus->dev_priv; int ret; intel_i2c_reset(dev_priv->dev); intel_i2c_quirk_set(dev_priv, true); - set_data(gpio, 1); - set_clock(gpio, 1); + set_data(bus, 1); + set_clock(bus, 1); udelay(I2C_RISEFALL_TIME); - ret = adapter->algo->master_xfer(adapter, msgs, num); + ret = i2c_bit_algo.master_xfer(&bus->adapter, msgs, num); - set_data(gpio, 1); - set_clock(gpio, 1); + set_data(bus, 1); + set_clock(bus, 1); intel_i2c_quirk_set(dev_priv, false); return ret; @@ -232,12 +209,15 @@ gmbus_xfer(struct i2c_adapter *adapter, struct intel_gmbus *bus = container_of(adapter, struct intel_gmbus, adapter); - struct drm_i915_private *dev_priv = adapter->algo_data; - int i, reg_offset; + struct drm_i915_private *dev_priv = bus->dev_priv; + int i, reg_offset, ret; - if (bus->force_bit) - return intel_i2c_quirk_xfer(dev_priv, - bus->force_bit, msgs, num); + mutex_lock(&dev_priv->gmbus_mutex); + + if (bus->force_bit) { + ret = intel_i2c_quirk_xfer(bus, msgs, num); + goto out; + } reg_offset = HAS_PCH_SPLIT(dev_priv->dev) ? PCH_GMBUS0 - GMBUS0 : 0; @@ -249,7 +229,8 @@ gmbus_xfer(struct i2c_adapter *adapter, if (msgs[i].flags & I2C_M_RD) { I915_WRITE(GMBUS1 + reg_offset, - GMBUS_CYCLE_WAIT | (i + 1 == num ? GMBUS_CYCLE_STOP : 0) | + GMBUS_CYCLE_WAIT | + (i + 1 == num ? GMBUS_CYCLE_STOP : 0) | (len << GMBUS_BYTE_COUNT_SHIFT) | (msgs[i].addr << GMBUS_SLAVE_ADDR_SHIFT) | GMBUS_SLAVE_READ | GMBUS_SW_RDY); @@ -278,7 +259,8 @@ gmbus_xfer(struct i2c_adapter *adapter, I915_WRITE(GMBUS3 + reg_offset, val); I915_WRITE(GMBUS1 + reg_offset, - (i + 1 == num ? GMBUS_CYCLE_STOP : GMBUS_CYCLE_WAIT) | + GMBUS_CYCLE_WAIT | + (i + 1 == num ? GMBUS_CYCLE_STOP : 0) | (msgs[i].len << GMBUS_BYTE_COUNT_SHIFT) | (msgs[i].addr << GMBUS_SLAVE_ADDR_SHIFT) | GMBUS_SLAVE_WRITE | GMBUS_SW_RDY); @@ -317,11 +299,15 @@ clear_err: I915_WRITE(GMBUS1 + reg_offset, 0); done: - /* Mark the GMBUS interface as disabled. We will re-enable it at the - * start of the next xfer, till then let it sleep. + /* Mark the GMBUS interface as disabled after waiting for idle. + * We will re-enable it at the start of the next xfer, + * till then let it sleep. */ + if (wait_for((I915_READ(GMBUS2 + reg_offset) & GMBUS_ACTIVE) == 0, 10)) + DRM_INFO("GMBUS timed out waiting for idle\n"); I915_WRITE(GMBUS0 + reg_offset, 0); - return i; + ret = i; + goto out; timeout: DRM_INFO("GMBUS timed out, falling back to bit banging on pin %d [%s]\n", @@ -329,23 +315,21 @@ timeout: I915_WRITE(GMBUS0 + reg_offset, 0); /* Hardware may not support GMBUS over these pins? Try GPIO bitbanging instead. */ - bus->force_bit = intel_gpio_create(dev_priv, bus->reg0 & 0xff); - if (!bus->force_bit) - return -ENOMEM; - - return intel_i2c_quirk_xfer(dev_priv, bus->force_bit, msgs, num); + if (!bus->has_gpio) { + ret = -EIO; + } else { + bus->force_bit = true; + ret = intel_i2c_quirk_xfer(bus, msgs, num); + } +out: + mutex_unlock(&dev_priv->gmbus_mutex); + return ret; } static u32 gmbus_func(struct i2c_adapter *adapter) { - struct intel_gmbus *bus = container_of(adapter, - struct intel_gmbus, - adapter); - - if (bus->force_bit) - bus->force_bit->algo->functionality(bus->force_bit); - - return (I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL | + return i2c_bit_algo.functionality(adapter) & + (I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL | /* I2C_FUNC_10BIT_ADDR | */ I2C_FUNC_SMBUS_READ_BLOCK_DATA | I2C_FUNC_SMBUS_BLOCK_PROC_CALL); @@ -375,11 +359,13 @@ int intel_setup_gmbus(struct drm_device *dev) struct drm_i915_private *dev_priv = dev->dev_private; int ret, i; - dev_priv->gmbus = kcalloc(sizeof(struct intel_gmbus), GMBUS_NUM_PORTS, + dev_priv->gmbus = kcalloc(GMBUS_NUM_PORTS, sizeof(struct intel_gmbus), GFP_KERNEL); if (dev_priv->gmbus == NULL) return -ENOMEM; + mutex_init(&dev_priv->gmbus_mutex); + for (i = 0; i < GMBUS_NUM_PORTS; i++) { struct intel_gmbus *bus = &dev_priv->gmbus[i]; @@ -391,7 +377,7 @@ int intel_setup_gmbus(struct drm_device *dev) names[i]); bus->adapter.dev.parent = &dev->pdev->dev; - bus->adapter.algo_data = dev_priv; + bus->dev_priv = dev_priv; bus->adapter.algo = &gmbus_algorithm; ret = i2c_add_adapter(&bus->adapter); @@ -401,8 +387,11 @@ int intel_setup_gmbus(struct drm_device *dev) /* By default use a conservative clock rate */ bus->reg0 = i | GMBUS_RATE_100KHZ; + bus->has_gpio = intel_gpio_setup(bus, i); + /* XXX force bit banging until GMBUS is fully debugged */ - bus->force_bit = intel_gpio_create(dev_priv, i); + if (bus->has_gpio) + bus->force_bit = true; } intel_i2c_reset(dev_priv->dev); @@ -430,19 +419,8 @@ void intel_gmbus_force_bit(struct i2c_adapter *adapter, bool force_bit) { struct intel_gmbus *bus = to_intel_gmbus(adapter); - if (force_bit) { - if (bus->force_bit == NULL) { - struct drm_i915_private *dev_priv = adapter->algo_data; - bus->force_bit = intel_gpio_create(dev_priv, - bus->reg0 & 0xff); - } - } else { - if (bus->force_bit) { - i2c_del_adapter(bus->force_bit); - kfree(bus->force_bit); - bus->force_bit = NULL; - } - } + if (bus->has_gpio) + bus->force_bit = force_bit; } void intel_teardown_gmbus(struct drm_device *dev) @@ -455,10 +433,6 @@ void intel_teardown_gmbus(struct drm_device *dev) for (i = 0; i < GMBUS_NUM_PORTS; i++) { struct intel_gmbus *bus = &dev_priv->gmbus[i]; - if (bus->force_bit) { - i2c_del_adapter(bus->force_bit); - kfree(bus->force_bit); - } i2c_del_adapter(&bus->adapter); } diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c index 6601d21..a8b28c4 100644 --- a/drivers/gpu/drm/i915/intel_lvds.c +++ b/drivers/gpu/drm/i915/intel_lvds.c @@ -187,6 +187,8 @@ centre_horizontally(struct drm_display_mode *mode, mode->crtc_hsync_start = mode->crtc_hblank_start + sync_pos; mode->crtc_hsync_end = mode->crtc_hsync_start + sync_width; + + mode->private_flags |= INTEL_MODE_CRTC_TIMINGS_SET; } static void @@ -208,6 +210,8 @@ centre_vertically(struct drm_display_mode *mode, mode->crtc_vsync_start = mode->crtc_vblank_start + sync_pos; mode->crtc_vsync_end = mode->crtc_vsync_start + sync_width; + + mode->private_flags |= INTEL_MODE_CRTC_TIMINGS_SET; } static inline u32 panel_fitter_scaling(u32 source, u32 target) @@ -283,6 +287,8 @@ static bool intel_lvds_mode_fixup(struct drm_encoder *encoder, for_each_pipe(pipe) I915_WRITE(BCLRPAT(pipe), 0); + drm_mode_set_crtcinfo(adjusted_mode, 0); + switch (intel_lvds->fitting_mode) { case DRM_MODE_SCALE_CENTER: /* @@ -535,7 +541,6 @@ static int intel_lid_notify(struct notifier_block *nb, unsigned long val, mutex_lock(&dev->mode_config.mutex); drm_helper_resume_force_mode(dev); - i915_redisable_vga(dev); mutex_unlock(&dev->mode_config.mutex); return NOTIFY_OK; @@ -893,6 +898,18 @@ static bool lvds_is_present_in_vbt(struct drm_device *dev, return false; } +static bool intel_lvds_supported(struct drm_device *dev) +{ + /* With the introduction of the PCH we gained a dedicated + * LVDS presence pin, use it. */ + if (HAS_PCH_SPLIT(dev)) + return true; + + /* Otherwise LVDS was only attached to mobile products, + * except for the inglorious 830gm */ + return IS_MOBILE(dev) && !IS_I830(dev); +} + /** * intel_lvds_init - setup LVDS connectors on this device * @dev: drm device @@ -914,6 +931,9 @@ bool intel_lvds_init(struct drm_device *dev) int pipe; u8 pin; + if (!intel_lvds_supported(dev)) + return false; + /* Skip init on machines we know falsely report LVDS */ if (dmi_check_system(intel_no_lvds)) return false; diff --git a/drivers/gpu/drm/i915/intel_modes.c b/drivers/gpu/drm/i915/intel_modes.c index be2c6fe..9a2b270 100644 --- a/drivers/gpu/drm/i915/intel_modes.c +++ b/drivers/gpu/drm/i915/intel_modes.c @@ -28,6 +28,7 @@ #include #include #include "drmP.h" +#include "drm_edid.h" #include "intel_drv.h" #include "i915_drv.h" @@ -42,13 +43,13 @@ bool intel_ddc_probe(struct intel_encoder *intel_encoder, int ddc_bus) u8 buf[2]; struct i2c_msg msgs[] = { { - .addr = 0x50, + .addr = DDC_ADDR, .flags = 0, .len = 1, .buf = out_buf, }, { - .addr = 0x50, + .addr = DDC_ADDR, .flags = I2C_M_RD, .len = 1, .buf = buf, @@ -59,6 +60,25 @@ bool intel_ddc_probe(struct intel_encoder *intel_encoder, int ddc_bus) } /** + * intel_connector_update_modes - update connector from edid + * @connector: DRM connector device to use + * @edid: previously read EDID information + */ +int intel_connector_update_modes(struct drm_connector *connector, + struct edid *edid) +{ + int ret; + + drm_mode_connector_update_edid_property(connector, edid); + ret = drm_add_edid_modes(connector, edid); + drm_edid_to_eld(connector, edid); + connector->display_info.raw_edid = NULL; + kfree(edid); + + return ret; +} + +/** * intel_ddc_get_modes - get modelist from monitor * @connector: DRM connector device to use * @adapter: i2c adapter @@ -69,24 +89,19 @@ int intel_ddc_get_modes(struct drm_connector *connector, struct i2c_adapter *adapter) { struct edid *edid; - int ret = 0; edid = drm_get_edid(connector, adapter); - if (edid) { - drm_mode_connector_update_edid_property(connector, edid); - ret = drm_add_edid_modes(connector, edid); - drm_edid_to_eld(connector, edid); - connector->display_info.raw_edid = NULL; - kfree(edid); - } + if (!edid) + return 0; - return ret; + return intel_connector_update_modes(connector, edid); } -static const char *force_audio_names[] = { - "off", - "auto", - "on", +static const struct drm_prop_enum_list force_audio_names[] = { + { HDMI_AUDIO_OFF_DVI, "force-dvi" }, + { HDMI_AUDIO_OFF, "off" }, + { HDMI_AUDIO_AUTO, "auto" }, + { HDMI_AUDIO_ON, "on" }, }; void @@ -95,27 +110,24 @@ intel_attach_force_audio_property(struct drm_connector *connector) struct drm_device *dev = connector->dev; struct drm_i915_private *dev_priv = dev->dev_private; struct drm_property *prop; - int i; prop = dev_priv->force_audio_property; if (prop == NULL) { - prop = drm_property_create(dev, DRM_MODE_PROP_ENUM, + prop = drm_property_create_enum(dev, 0, "audio", + force_audio_names, ARRAY_SIZE(force_audio_names)); if (prop == NULL) return; - for (i = 0; i < ARRAY_SIZE(force_audio_names); i++) - drm_property_add_enum(prop, i, i-1, force_audio_names[i]); - dev_priv->force_audio_property = prop; } drm_connector_attach_property(connector, prop, 0); } -static const char *broadcast_rgb_names[] = { - "Full", - "Limited 16:235", +static const struct drm_prop_enum_list broadcast_rgb_names[] = { + { 0, "Full" }, + { 1, "Limited 16:235" }, }; void @@ -124,19 +136,16 @@ intel_attach_broadcast_rgb_property(struct drm_connector *connector) struct drm_device *dev = connector->dev; struct drm_i915_private *dev_priv = dev->dev_private; struct drm_property *prop; - int i; prop = dev_priv->broadcast_rgb_property; if (prop == NULL) { - prop = drm_property_create(dev, DRM_MODE_PROP_ENUM, + prop = drm_property_create_enum(dev, DRM_MODE_PROP_ENUM, "Broadcast RGB", + broadcast_rgb_names, ARRAY_SIZE(broadcast_rgb_names)); if (prop == NULL) return; - for (i = 0; i < ARRAY_SIZE(broadcast_rgb_names); i++) - drm_property_add_enum(prop, i, i, broadcast_rgb_names[i]); - dev_priv->broadcast_rgb_property = prop; } diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c index 478b51f..5ba5e66 100644 --- a/drivers/gpu/drm/i915/intel_overlay.c +++ b/drivers/gpu/drm/i915/intel_overlay.c @@ -25,8 +25,6 @@ * * Derived from Xorg ddx, xf86-video-intel, src/i830_video.c */ - -#include #include "drmP.h" #include "drm.h" #include "i915_drm.h" @@ -227,7 +225,8 @@ static int intel_overlay_do_wait_request(struct intel_overlay *overlay, } overlay->last_flip_req = request->seqno; overlay->flip_tail = tail; - ret = i915_wait_request(LP_RING(dev_priv), overlay->last_flip_req); + ret = i915_wait_request(LP_RING(dev_priv), overlay->last_flip_req, + true); if (ret) return ret; @@ -263,7 +262,7 @@ i830_activate_pipe_a(struct drm_device *dev) DRM_DEBUG_DRIVER("Enabling pipe A in order to enable overlay\n"); mode = drm_mode_duplicate(dev, &vesa_640x480); - drm_mode_set_crtcinfo(mode, CRTC_INTERLACE_HALVE_V); + drm_mode_set_crtcinfo(mode, 0); if (!drm_crtc_helper_set_mode(&crtc->base, mode, crtc->base.x, crtc->base.y, crtc->base.fb)) @@ -456,7 +455,8 @@ static int intel_overlay_recover_from_interrupt(struct intel_overlay *overlay) if (overlay->last_flip_req == 0) return 0; - ret = i915_wait_request(LP_RING(dev_priv), overlay->last_flip_req); + ret = i915_wait_request(LP_RING(dev_priv), overlay->last_flip_req, + true); if (ret) return ret; @@ -943,10 +943,10 @@ static int check_overlay_dst(struct intel_overlay *overlay, { struct drm_display_mode *mode = &overlay->crtc->base.mode; - if (rec->dst_x < mode->crtc_hdisplay && - rec->dst_x + rec->dst_width <= mode->crtc_hdisplay && - rec->dst_y < mode->crtc_vdisplay && - rec->dst_y + rec->dst_height <= mode->crtc_vdisplay) + if (rec->dst_x < mode->hdisplay && + rec->dst_x + rec->dst_width <= mode->hdisplay && + rec->dst_y < mode->vdisplay && + rec->dst_y + rec->dst_height <= mode->vdisplay) return 0; else return -EINVAL; diff --git a/drivers/gpu/drm/i915/intel_panel.c b/drivers/gpu/drm/i915/intel_panel.c index 04d79fd..48177ec 100644 --- a/drivers/gpu/drm/i915/intel_panel.c +++ b/drivers/gpu/drm/i915/intel_panel.c @@ -47,8 +47,6 @@ intel_fixed_panel_mode(struct drm_display_mode *fixed_mode, adjusted_mode->vtotal = fixed_mode->vtotal; adjusted_mode->clock = fixed_mode->clock; - - drm_mode_set_crtcinfo(adjusted_mode, CRTC_INTERLACE_HALVE_V); } /* adjusted_mode has been preset to be the panel's fixed mode */ @@ -141,8 +139,8 @@ static u32 i915_read_blc_pwm_ctl(struct drm_i915_private *dev_priv) dev_priv->saveBLC_PWM_CTL2 = val; } else if (val == 0) { I915_WRITE(BLC_PWM_PCH_CTL2, - dev_priv->saveBLC_PWM_CTL); - val = dev_priv->saveBLC_PWM_CTL; + dev_priv->saveBLC_PWM_CTL2); + val = dev_priv->saveBLC_PWM_CTL2; } } else { val = I915_READ(BLC_PWM_CTL); diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 4fddd21..c17325c 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -263,6 +263,8 @@ static int init_ring_common(struct intel_ring_buffer *ring) I915_WRITE_HEAD(ring, 0); ring->write_tail(ring, 0); + /* Initialize the ring. */ + I915_WRITE_START(ring, obj->gtt_offset); head = I915_READ_HEAD(ring) & HEAD_ADDR; /* G45 ring initialization fails to reset head to zero */ @@ -288,11 +290,6 @@ static int init_ring_common(struct intel_ring_buffer *ring) } } - /* Initialize the ring. This must happen _after_ we've cleared the ring - * registers with the above sequence (the readback of the HEAD registers - * also enforces ordering), otherwise the hw might lose the new ring - * register values. */ - I915_WRITE_START(ring, obj->gtt_offset); I915_WRITE_CTL(ring, ((ring->size - PAGE_SIZE) & RING_NR_PAGES) | RING_VALID); @@ -318,6 +315,7 @@ static int init_ring_common(struct intel_ring_buffer *ring) ring->head = I915_READ_HEAD(ring); ring->tail = I915_READ_TAIL(ring) & TAIL_ADDR; ring->space = ring_space(ring); + ring->last_retired_head = -1; } out: @@ -399,25 +397,12 @@ static int init_render_ring(struct intel_ring_buffer *ring) if (INTEL_INFO(dev)->gen > 3) { int mode = VS_TIMER_DISPATCH << 16 | VS_TIMER_DISPATCH; I915_WRITE(MI_MODE, mode); + if (IS_GEN7(dev)) + I915_WRITE(GFX_MODE_GEN7, + GFX_MODE_DISABLE(GFX_TLB_INVALIDATE_ALWAYS) | + GFX_MODE_ENABLE(GFX_REPLAY_MODE)); } - /* We need to disable the AsyncFlip performance optimisations in order - * to use MI_WAIT_FOR_EVENT within the CS. It should already be - * programmed to '1' on all products. - */ - if (INTEL_INFO(dev)->gen >= 6) - I915_WRITE(MI_MODE, GFX_MODE_ENABLE(ASYNC_FLIP_PERF_DISABLE)); - - /* Required for the hardware to program scanline values for waiting */ - if (INTEL_INFO(dev)->gen == 6) - I915_WRITE(GFX_MODE, - GFX_MODE_ENABLE(GFX_TLB_INVALIDATE_ALWAYS)); - - if (IS_GEN7(dev)) - I915_WRITE(GFX_MODE_GEN7, - GFX_MODE_DISABLE(GFX_TLB_INVALIDATE_ALWAYS) | - GFX_MODE_ENABLE(GFX_REPLAY_MODE)); - if (INTEL_INFO(dev)->gen >= 5) { ret = init_pipe_control(ring); if (ret) @@ -620,6 +605,7 @@ pc_render_add_request(struct intel_ring_buffer *ring, PIPE_CONTROL_FLUSH(ring, scratch_addr); scratch_addr += 128; PIPE_CONTROL_FLUSH(ring, scratch_addr); + intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(4) | PIPE_CONTROL_QW_WRITE | PIPE_CONTROL_WRITE_FLUSH | PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE | @@ -662,7 +648,7 @@ gen6_ring_get_seqno(struct intel_ring_buffer *ring) /* Workaround to force correct ordering between irq and seqno writes on * ivb (and maybe also on snb) by reading from a CS register (like * ACTHD) before reading the status page. */ - if (IS_GEN7(dev)) + if (IS_GEN6(dev) || IS_GEN7(dev)) intel_ring_get_active_head(ring); return intel_read_status_page(ring, I915_GEM_HWS_INDEX); } @@ -763,13 +749,13 @@ void intel_ring_setup_status_page(struct intel_ring_buffer *ring) */ if (IS_GEN7(dev)) { switch (ring->id) { - case RING_RENDER: + case RCS: mmio = RENDER_HWS_PGA_GEN7; break; - case RING_BLT: + case BCS: mmio = BLT_HWS_PGA_GEN7; break; - case RING_BSD: + case VCS: mmio = BSD_HWS_PGA_GEN7; break; } @@ -835,8 +821,7 @@ gen6_ring_get_irq(struct intel_ring_buffer *ring, u32 gflag, u32 rflag) /* It looks like we need to prevent the gt from suspending while waiting * for an notifiy irq, otherwise irqs seem to get lost on at least the * blt/bsd rings on ivb. */ - if (IS_GEN7(dev)) - gen6_gt_force_wake_get(dev_priv); + gen6_gt_force_wake_get(dev_priv); spin_lock(&ring->irq_lock); if (ring->irq_refcount++ == 0) { @@ -863,8 +848,7 @@ gen6_ring_put_irq(struct intel_ring_buffer *ring, u32 gflag, u32 rflag) } spin_unlock(&ring->irq_lock); - if (IS_GEN7(dev)) - gen6_gt_force_wake_put(dev_priv); + gen6_gt_force_wake_put(dev_priv); } static bool @@ -1150,14 +1134,105 @@ static int intel_wrap_ring_buffer(struct intel_ring_buffer *ring) return 0; } +static int intel_ring_wait_seqno(struct intel_ring_buffer *ring, u32 seqno) +{ + struct drm_i915_private *dev_priv = ring->dev->dev_private; + bool was_interruptible; + int ret; + + /* XXX As we have not yet audited all the paths to check that + * they are ready for ERESTARTSYS from intel_ring_begin, do not + * allow us to be interruptible by a signal. + */ + was_interruptible = dev_priv->mm.interruptible; + dev_priv->mm.interruptible = false; + + ret = i915_wait_request(ring, seqno, true); + + dev_priv->mm.interruptible = was_interruptible; + + return ret; +} + +static int intel_ring_wait_request(struct intel_ring_buffer *ring, int n) +{ + struct drm_i915_gem_request *request; + u32 seqno = 0; + int ret; + + i915_gem_retire_requests_ring(ring); + + if (ring->last_retired_head != -1) { + ring->head = ring->last_retired_head; + ring->last_retired_head = -1; + ring->space = ring_space(ring); + if (ring->space >= n) + return 0; + } + + list_for_each_entry(request, &ring->request_list, list) { + int space; + + if (request->tail == -1) + continue; + + space = request->tail - (ring->tail + 8); + if (space < 0) + space += ring->size; + if (space >= n) { + seqno = request->seqno; + break; + } + + /* Consume this request in case we need more space than + * is available and so need to prevent a race between + * updating last_retired_head and direct reads of + * I915_RING_HEAD. It also provides a nice sanity check. + */ + request->tail = -1; + } + + if (seqno == 0) + return -ENOSPC; + + ret = intel_ring_wait_seqno(ring, seqno); + if (ret) + return ret; + + if (WARN_ON(ring->last_retired_head == -1)) + return -ENOSPC; + + ring->head = ring->last_retired_head; + ring->last_retired_head = -1; + ring->space = ring_space(ring); + if (WARN_ON(ring->space < n)) + return -ENOSPC; + + return 0; +} + int intel_wait_ring_buffer(struct intel_ring_buffer *ring, int n) { struct drm_device *dev = ring->dev; struct drm_i915_private *dev_priv = dev->dev_private; unsigned long end; + int ret; + + ret = intel_ring_wait_request(ring, n); + if (ret != -ENOSPC) + return ret; trace_i915_ring_wait_begin(ring); - end = jiffies + 3 * HZ; + if (drm_core_check_feature(dev, DRIVER_GEM)) + /* With GEM the hangcheck timer should kick us out of the loop, + * leaving it early runs the risk of corrupting GEM state (due + * to running on almost untested codepaths). But on resume + * timers don't work yet, so prevent a complete hang in that + * case by choosing an insanely large timeout. */ + end = jiffies + 60 * HZ; + else + end = jiffies + 3 * HZ; + do { ring->head = I915_READ_HEAD(ring); ring->space = ring_space(ring); @@ -1214,7 +1289,7 @@ void intel_ring_advance(struct intel_ring_buffer *ring) static const struct intel_ring_buffer render_ring = { .name = "render ring", - .id = RING_RENDER, + .id = RCS, .mmio_base = RENDER_RING_BASE, .size = 32 * PAGE_SIZE, .init = init_render_ring, @@ -1237,7 +1312,7 @@ static const struct intel_ring_buffer render_ring = { static const struct intel_ring_buffer bsd_ring = { .name = "bsd ring", - .id = RING_BSD, + .id = VCS, .mmio_base = BSD_RING_BASE, .size = 32 * PAGE_SIZE, .init = init_ring_common, @@ -1347,7 +1422,7 @@ gen6_bsd_ring_put_irq(struct intel_ring_buffer *ring) /* ring buffer for Video Codec for Gen6+ */ static const struct intel_ring_buffer gen6_bsd_ring = { .name = "gen6 bsd ring", - .id = RING_BSD, + .id = VCS, .mmio_base = GEN6_BSD_RING_BASE, .size = 32 * PAGE_SIZE, .init = init_ring_common, @@ -1383,79 +1458,13 @@ blt_ring_put_irq(struct intel_ring_buffer *ring) GEN6_BLITTER_USER_INTERRUPT); } - -/* Workaround for some stepping of SNB, - * each time when BLT engine ring tail moved, - * the first command in the ring to be parsed - * should be MI_BATCH_BUFFER_START - */ -#define NEED_BLT_WORKAROUND(dev) \ - (IS_GEN6(dev) && (dev->pdev->revision < 8)) - -static inline struct drm_i915_gem_object * -to_blt_workaround(struct intel_ring_buffer *ring) -{ - return ring->private; -} - -static int blt_ring_init(struct intel_ring_buffer *ring) -{ - if (NEED_BLT_WORKAROUND(ring->dev)) { - struct drm_i915_gem_object *obj; - u32 *ptr; - int ret; - - obj = i915_gem_alloc_object(ring->dev, 4096); - if (obj == NULL) - return -ENOMEM; - - ret = i915_gem_object_pin(obj, 4096, true); - if (ret) { - drm_gem_object_unreference(&obj->base); - return ret; - } - - ptr = kmap(obj->pages[0]); - *ptr++ = MI_BATCH_BUFFER_END; - *ptr++ = MI_NOOP; - kunmap(obj->pages[0]); - - ret = i915_gem_object_set_to_gtt_domain(obj, false); - if (ret) { - i915_gem_object_unpin(obj); - drm_gem_object_unreference(&obj->base); - return ret; - } - - ring->private = obj; - } - - return init_ring_common(ring); -} - -static int blt_ring_begin(struct intel_ring_buffer *ring, - int num_dwords) -{ - if (ring->private) { - int ret = intel_ring_begin(ring, num_dwords+2); - if (ret) - return ret; - - intel_ring_emit(ring, MI_BATCH_BUFFER_START); - intel_ring_emit(ring, to_blt_workaround(ring)->gtt_offset); - - return 0; - } else - return intel_ring_begin(ring, 4); -} - static int blt_ring_flush(struct intel_ring_buffer *ring, u32 invalidate, u32 flush) { uint32_t cmd; int ret; - ret = blt_ring_begin(ring, 4); + ret = intel_ring_begin(ring, 4); if (ret) return ret; @@ -1470,22 +1479,12 @@ static int blt_ring_flush(struct intel_ring_buffer *ring, return 0; } -static void blt_ring_cleanup(struct intel_ring_buffer *ring) -{ - if (!ring->private) - return; - - i915_gem_object_unpin(ring->private); - drm_gem_object_unreference(ring->private); - ring->private = NULL; -} - static const struct intel_ring_buffer gen6_blt_ring = { .name = "blt ring", - .id = RING_BLT, + .id = BCS, .mmio_base = BLT_RING_BASE, .size = 32 * PAGE_SIZE, - .init = blt_ring_init, + .init = init_ring_common, .write_tail = ring_write_tail, .flush = blt_ring_flush, .add_request = gen6_add_request, @@ -1493,7 +1492,6 @@ static const struct intel_ring_buffer gen6_blt_ring = { .irq_get = blt_ring_get_irq, .irq_put = blt_ring_put_irq, .dispatch_execbuffer = gen6_ring_dispatch_execbuffer, - .cleanup = blt_ring_cleanup, .sync_to = gen6_blt_ring_sync_to, .semaphore_register = {MI_SEMAPHORE_SYNC_BR, MI_SEMAPHORE_SYNC_BV, diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index 68281c9..bc0365b 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -1,13 +1,6 @@ #ifndef _INTEL_RINGBUFFER_H_ #define _INTEL_RINGBUFFER_H_ -enum { - RCS = 0x0, - VCS, - BCS, - I915_NUM_RINGS, -}; - struct intel_hw_status_page { u32 __iomem *page_addr; unsigned int gfx_addr; @@ -36,10 +29,11 @@ struct intel_hw_status_page { struct intel_ring_buffer { const char *name; enum intel_ring_id { - RING_RENDER = 0x1, - RING_BSD = 0x2, - RING_BLT = 0x4, + RCS = 0x0, + VCS, + BCS, } id; +#define I915_NUM_RINGS 3 u32 mmio_base; void __iomem *virtual_start; struct drm_device *dev; @@ -52,6 +46,16 @@ struct intel_ring_buffer { int effective_size; struct intel_hw_status_page status_page; + /** We track the position of the requests in the ring buffer, and + * when each is retired we increment last_retired_head as the GPU + * must have finished processing the request and so we know we + * can advance the ringbuffer up to that position. + * + * last_retired_head is set to -1 after the value is consumed so + * we can detect new retirements. + */ + u32 last_retired_head; + spinlock_t irq_lock; u32 irq_refcount; u32 irq_mask; @@ -119,6 +123,12 @@ struct intel_ring_buffer { void *private; }; +static inline unsigned +intel_ring_flag(struct intel_ring_buffer *ring) +{ + return 1 << ring->id; +} + static inline u32 intel_ring_sync_index(struct intel_ring_buffer *ring, struct intel_ring_buffer *other) @@ -193,6 +203,11 @@ int intel_init_blt_ring_buffer(struct drm_device *dev); u32 intel_ring_get_active_head(struct intel_ring_buffer *ring); void intel_ring_setup_status_page(struct intel_ring_buffer *ring); +static inline u32 intel_ring_get_tail(struct intel_ring_buffer *ring) +{ + return ring->tail; +} + static inline void i915_trace_irq_get(struct intel_ring_buffer *ring, u32 seqno) { if (ring->trace_irq_seqno == 0 && ring->irq_get(ring)) diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c index 9e24670..1b6b157 100644 --- a/drivers/gpu/drm/i915/intel_sdvo.c +++ b/drivers/gpu/drm/i915/intel_sdvo.c @@ -148,7 +148,7 @@ struct intel_sdvo_connector { /* Mark the type of connector */ uint16_t output_flag; - int force_audio; + enum hdmi_force_audio force_audio; /* This contains all current supported TV format */ u8 tv_format_supported[TV_FORMAT_NUM]; @@ -982,7 +982,6 @@ intel_sdvo_set_input_timings_for_mode(struct intel_sdvo *intel_sdvo, intel_sdvo_get_mode_from_dtd(adjusted_mode, &intel_sdvo->input_dtd); - drm_mode_set_crtcinfo(adjusted_mode, 0); return true; } @@ -1351,8 +1350,8 @@ intel_sdvo_tmds_sink_detect(struct drm_connector *connector) if (status == connector_status_connected) { struct intel_sdvo_connector *intel_sdvo_connector = to_intel_sdvo_connector(connector); - if (intel_sdvo_connector->force_audio) - intel_sdvo->has_hdmi_audio = intel_sdvo_connector->force_audio > 0; + if (intel_sdvo_connector->force_audio != HDMI_AUDIO_AUTO) + intel_sdvo->has_hdmi_audio = (intel_sdvo_connector->force_audio == HDMI_AUDIO_ON); } return status; @@ -1725,10 +1724,10 @@ intel_sdvo_set_property(struct drm_connector *connector, intel_sdvo_connector->force_audio = i; - if (i == 0) + if (i == HDMI_AUDIO_AUTO) has_audio = intel_sdvo_detect_hdmi_audio(connector); else - has_audio = i > 0; + has_audio = (i == HDMI_AUDIO_ON); if (has_audio == intel_sdvo->has_hdmi_audio) return 0; @@ -2026,7 +2025,7 @@ intel_sdvo_connector_init(struct intel_sdvo_connector *connector, drm_connector_helper_add(&connector->base.base, &intel_sdvo_connector_helper_funcs); - connector->base.base.interlace_allowed = 0; + connector->base.base.interlace_allowed = 1; connector->base.base.doublescan_allowed = 0; connector->base.base.display_info.subpixel_order = SubPixelHorizontalRGB; @@ -2264,18 +2263,6 @@ intel_sdvo_output_setup(struct intel_sdvo *intel_sdvo, uint16_t flags) return true; } -static void intel_sdvo_output_cleanup(struct intel_sdvo *intel_sdvo) -{ - struct drm_device *dev = intel_sdvo->base.base.dev; - struct drm_connector *connector, *tmp; - - list_for_each_entry_safe(connector, tmp, - &dev->mode_config.connector_list, head) { - if (intel_attached_encoder(connector) == &intel_sdvo->base) - intel_sdvo_destroy(connector); - } -} - static bool intel_sdvo_tv_create_property(struct intel_sdvo *intel_sdvo, struct intel_sdvo_connector *intel_sdvo_connector, int type) @@ -2330,10 +2317,8 @@ static bool intel_sdvo_tv_create_property(struct intel_sdvo *intel_sdvo, intel_sdvo_connector->max_##name = data_value[0]; \ intel_sdvo_connector->cur_##name = response; \ intel_sdvo_connector->name = \ - drm_property_create(dev, DRM_MODE_PROP_RANGE, #name, 2); \ + drm_property_create_range(dev, 0, #name, 0, data_value[0]); \ if (!intel_sdvo_connector->name) return false; \ - intel_sdvo_connector->name->values[0] = 0; \ - intel_sdvo_connector->name->values[1] = data_value[0]; \ drm_connector_attach_property(connector, \ intel_sdvo_connector->name, \ intel_sdvo_connector->cur_##name); \ @@ -2367,25 +2352,19 @@ intel_sdvo_create_enhance_property_tv(struct intel_sdvo *intel_sdvo, intel_sdvo_connector->left_margin = data_value[0] - response; intel_sdvo_connector->right_margin = intel_sdvo_connector->left_margin; intel_sdvo_connector->left = - drm_property_create(dev, DRM_MODE_PROP_RANGE, - "left_margin", 2); + drm_property_create_range(dev, 0, "left_margin", 0, data_value[0]); if (!intel_sdvo_connector->left) return false; - intel_sdvo_connector->left->values[0] = 0; - intel_sdvo_connector->left->values[1] = data_value[0]; drm_connector_attach_property(connector, intel_sdvo_connector->left, intel_sdvo_connector->left_margin); intel_sdvo_connector->right = - drm_property_create(dev, DRM_MODE_PROP_RANGE, - "right_margin", 2); + drm_property_create_range(dev, 0, "right_margin", 0, data_value[0]); if (!intel_sdvo_connector->right) return false; - intel_sdvo_connector->right->values[0] = 0; - intel_sdvo_connector->right->values[1] = data_value[0]; drm_connector_attach_property(connector, intel_sdvo_connector->right, intel_sdvo_connector->right_margin); @@ -2409,25 +2388,21 @@ intel_sdvo_create_enhance_property_tv(struct intel_sdvo *intel_sdvo, intel_sdvo_connector->top_margin = data_value[0] - response; intel_sdvo_connector->bottom_margin = intel_sdvo_connector->top_margin; intel_sdvo_connector->top = - drm_property_create(dev, DRM_MODE_PROP_RANGE, - "top_margin", 2); + drm_property_create_range(dev, 0, + "top_margin", 0, data_value[0]); if (!intel_sdvo_connector->top) return false; - intel_sdvo_connector->top->values[0] = 0; - intel_sdvo_connector->top->values[1] = data_value[0]; drm_connector_attach_property(connector, intel_sdvo_connector->top, intel_sdvo_connector->top_margin); intel_sdvo_connector->bottom = - drm_property_create(dev, DRM_MODE_PROP_RANGE, - "bottom_margin", 2); + drm_property_create_range(dev, 0, + "bottom_margin", 0, data_value[0]); if (!intel_sdvo_connector->bottom) return false; - intel_sdvo_connector->bottom->values[0] = 0; - intel_sdvo_connector->bottom->values[1] = data_value[0]; drm_connector_attach_property(connector, intel_sdvo_connector->bottom, intel_sdvo_connector->bottom_margin); @@ -2456,12 +2431,10 @@ intel_sdvo_create_enhance_property_tv(struct intel_sdvo *intel_sdvo, intel_sdvo_connector->max_dot_crawl = 1; intel_sdvo_connector->cur_dot_crawl = response & 0x1; intel_sdvo_connector->dot_crawl = - drm_property_create(dev, DRM_MODE_PROP_RANGE, "dot_crawl", 2); + drm_property_create_range(dev, 0, "dot_crawl", 0, 1); if (!intel_sdvo_connector->dot_crawl) return false; - intel_sdvo_connector->dot_crawl->values[0] = 0; - intel_sdvo_connector->dot_crawl->values[1] = 1; drm_connector_attach_property(connector, intel_sdvo_connector->dot_crawl, intel_sdvo_connector->cur_dot_crawl); @@ -2608,8 +2581,7 @@ bool intel_sdvo_init(struct drm_device *dev, int sdvo_reg) intel_sdvo->caps.output_flags) != true) { DRM_DEBUG_KMS("SDVO output failed to setup on SDVO%c\n", IS_SDVOB(sdvo_reg) ? 'B' : 'C'); - /* Output_setup can leave behind connectors! */ - goto err_output; + goto err; } /* Only enable the hotplug irq if we need it, to work around noisy @@ -2622,12 +2594,12 @@ bool intel_sdvo_init(struct drm_device *dev, int sdvo_reg) /* Set the input timing to the screen. Assume always input 0. */ if (!intel_sdvo_set_target_input(intel_sdvo)) - goto err_output; + goto err; if (!intel_sdvo_get_input_pixel_clock_range(intel_sdvo, &intel_sdvo->pixel_clock_min, &intel_sdvo->pixel_clock_max)) - goto err_output; + goto err; DRM_DEBUG_KMS("%s device VID/DID: %02X:%02X.%02X, " "clock range %dMHz - %dMHz, " @@ -2647,9 +2619,6 @@ bool intel_sdvo_init(struct drm_device *dev, int sdvo_reg) (SDVO_OUTPUT_TMDS1 | SDVO_OUTPUT_RGB1) ? 'Y' : 'N'); return true; -err_output: - intel_sdvo_output_cleanup(intel_sdvo); - err: drm_encoder_cleanup(&intel_encoder->base); i2c_del_adapter(&intel_sdvo->ddc); diff --git a/drivers/gpu/drm/i915/intel_sdvo_regs.h b/drivers/gpu/drm/i915/intel_sdvo_regs.h index 4193c54..770bdd6 100644 --- a/drivers/gpu/drm/i915/intel_sdvo_regs.h +++ b/drivers/gpu/drm/i915/intel_sdvo_regs.h @@ -1,5 +1,5 @@ /* - * Copyright © 2006-2007 Intel Corporation + * Copyright © 2006-2007 Intel Corporation * * Permission is hereby granted, free of charge, to any person obtaining a * copy of this software and associated documentation files (the "Software"), diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c new file mode 100644 index 0000000..e90dfb6 --- /dev/null +++ b/drivers/gpu/drm/i915/intel_sprite.c @@ -0,0 +1,666 @@ +/* + * Copyright © 2011 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + * Authors: + * Jesse Barnes + * + * New plane/sprite handling. + * + * The older chips had a separate interface for programming plane related + * registers; newer ones are much simpler and we can use the new DRM plane + * support. + */ +#include "drmP.h" +#include "drm_crtc.h" +#include "drm_fourcc.h" +#include "intel_drv.h" +#include "i915_drm.h" +#include "i915_drv.h" + +static void +ivb_update_plane(struct drm_plane *plane, struct drm_framebuffer *fb, + struct drm_i915_gem_object *obj, int crtc_x, int crtc_y, + unsigned int crtc_w, unsigned int crtc_h, + uint32_t x, uint32_t y, + uint32_t src_w, uint32_t src_h) +{ + struct drm_device *dev = plane->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_plane *intel_plane = to_intel_plane(plane); + int pipe = intel_plane->pipe; + u32 sprctl, sprscale = 0; + int pixel_size; + + sprctl = I915_READ(SPRCTL(pipe)); + + /* Mask out pixel format bits in case we change it */ + sprctl &= ~SPRITE_PIXFORMAT_MASK; + sprctl &= ~SPRITE_RGB_ORDER_RGBX; + sprctl &= ~SPRITE_YUV_BYTE_ORDER_MASK; + + switch (fb->pixel_format) { + case DRM_FORMAT_XBGR8888: + sprctl |= SPRITE_FORMAT_RGBX888; + pixel_size = 4; + break; + case DRM_FORMAT_XRGB8888: + sprctl |= SPRITE_FORMAT_RGBX888 | SPRITE_RGB_ORDER_RGBX; + pixel_size = 4; + break; + case DRM_FORMAT_YUYV: + sprctl |= SPRITE_FORMAT_YUV422 | SPRITE_YUV_ORDER_YUYV; + pixel_size = 2; + break; + case DRM_FORMAT_YVYU: + sprctl |= SPRITE_FORMAT_YUV422 | SPRITE_YUV_ORDER_YVYU; + pixel_size = 2; + break; + case DRM_FORMAT_UYVY: + sprctl |= SPRITE_FORMAT_YUV422 | SPRITE_YUV_ORDER_UYVY; + pixel_size = 2; + break; + case DRM_FORMAT_VYUY: + sprctl |= SPRITE_FORMAT_YUV422 | SPRITE_YUV_ORDER_VYUY; + pixel_size = 2; + break; + default: + DRM_DEBUG_DRIVER("bad pixel format, assuming RGBX888\n"); + sprctl |= DVS_FORMAT_RGBX888; + pixel_size = 4; + break; + } + + if (obj->tiling_mode != I915_TILING_NONE) + sprctl |= SPRITE_TILED; + + /* must disable */ + sprctl |= SPRITE_TRICKLE_FEED_DISABLE; + sprctl |= SPRITE_ENABLE; + + /* Sizes are 0 based */ + src_w--; + src_h--; + crtc_w--; + crtc_h--; + + intel_update_sprite_watermarks(dev, pipe, crtc_w, pixel_size); + + /* + * IVB workaround: must disable low power watermarks for at least + * one frame before enabling scaling. LP watermarks can be re-enabled + * when scaling is disabled. + */ + if (crtc_w != src_w || crtc_h != src_h) { + dev_priv->sprite_scaling_enabled = true; + sandybridge_update_wm(dev); + intel_wait_for_vblank(dev, pipe); + sprscale = SPRITE_SCALE_ENABLE | (src_w << 16) | src_h; + } else { + dev_priv->sprite_scaling_enabled = false; + /* potentially re-enable LP watermarks */ + sandybridge_update_wm(dev); + } + + I915_WRITE(SPRSTRIDE(pipe), fb->pitches[0]); + I915_WRITE(SPRPOS(pipe), (crtc_y << 16) | crtc_x); + if (obj->tiling_mode != I915_TILING_NONE) { + I915_WRITE(SPRTILEOFF(pipe), (y << 16) | x); + } else { + unsigned long offset; + + offset = y * fb->pitches[0] + x * (fb->bits_per_pixel / 8); + I915_WRITE(SPRLINOFF(pipe), offset); + } + I915_WRITE(SPRSIZE(pipe), (crtc_h << 16) | crtc_w); + I915_WRITE(SPRSCALE(pipe), sprscale); + I915_WRITE(SPRCTL(pipe), sprctl); + I915_WRITE(SPRSURF(pipe), obj->gtt_offset); + POSTING_READ(SPRSURF(pipe)); +} + +static void +ivb_disable_plane(struct drm_plane *plane) +{ + struct drm_device *dev = plane->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_plane *intel_plane = to_intel_plane(plane); + int pipe = intel_plane->pipe; + + I915_WRITE(SPRCTL(pipe), I915_READ(SPRCTL(pipe)) & ~SPRITE_ENABLE); + /* Can't leave the scaler enabled... */ + I915_WRITE(SPRSCALE(pipe), 0); + /* Activate double buffered register update */ + I915_WRITE(SPRSURF(pipe), 0); + POSTING_READ(SPRSURF(pipe)); +} + +static int +ivb_update_colorkey(struct drm_plane *plane, + struct drm_intel_sprite_colorkey *key) +{ + struct drm_device *dev = plane->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_plane *intel_plane; + u32 sprctl; + int ret = 0; + + intel_plane = to_intel_plane(plane); + + I915_WRITE(SPRKEYVAL(intel_plane->pipe), key->min_value); + I915_WRITE(SPRKEYMAX(intel_plane->pipe), key->max_value); + I915_WRITE(SPRKEYMSK(intel_plane->pipe), key->channel_mask); + + sprctl = I915_READ(SPRCTL(intel_plane->pipe)); + sprctl &= ~(SPRITE_SOURCE_KEY | SPRITE_DEST_KEY); + if (key->flags & I915_SET_COLORKEY_DESTINATION) + sprctl |= SPRITE_DEST_KEY; + else if (key->flags & I915_SET_COLORKEY_SOURCE) + sprctl |= SPRITE_SOURCE_KEY; + I915_WRITE(SPRCTL(intel_plane->pipe), sprctl); + + POSTING_READ(SPRKEYMSK(intel_plane->pipe)); + + return ret; +} + +static void +ivb_get_colorkey(struct drm_plane *plane, struct drm_intel_sprite_colorkey *key) +{ + struct drm_device *dev = plane->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_plane *intel_plane; + u32 sprctl; + + intel_plane = to_intel_plane(plane); + + key->min_value = I915_READ(SPRKEYVAL(intel_plane->pipe)); + key->max_value = I915_READ(SPRKEYMAX(intel_plane->pipe)); + key->channel_mask = I915_READ(SPRKEYMSK(intel_plane->pipe)); + key->flags = 0; + + sprctl = I915_READ(SPRCTL(intel_plane->pipe)); + + if (sprctl & SPRITE_DEST_KEY) + key->flags = I915_SET_COLORKEY_DESTINATION; + else if (sprctl & SPRITE_SOURCE_KEY) + key->flags = I915_SET_COLORKEY_SOURCE; + else + key->flags = I915_SET_COLORKEY_NONE; +} + +static void +snb_update_plane(struct drm_plane *plane, struct drm_framebuffer *fb, + struct drm_i915_gem_object *obj, int crtc_x, int crtc_y, + unsigned int crtc_w, unsigned int crtc_h, + uint32_t x, uint32_t y, + uint32_t src_w, uint32_t src_h) +{ + struct drm_device *dev = plane->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_plane *intel_plane = to_intel_plane(plane); + int pipe = intel_plane->pipe, pixel_size; + u32 dvscntr, dvsscale = 0; + + dvscntr = I915_READ(DVSCNTR(pipe)); + + /* Mask out pixel format bits in case we change it */ + dvscntr &= ~DVS_PIXFORMAT_MASK; + dvscntr &= ~DVS_RGB_ORDER_XBGR; + dvscntr &= ~DVS_YUV_BYTE_ORDER_MASK; + + switch (fb->pixel_format) { + case DRM_FORMAT_XBGR8888: + dvscntr |= DVS_FORMAT_RGBX888 | DVS_RGB_ORDER_XBGR; + pixel_size = 4; + break; + case DRM_FORMAT_XRGB8888: + dvscntr |= DVS_FORMAT_RGBX888; + pixel_size = 4; + break; + case DRM_FORMAT_YUYV: + dvscntr |= DVS_FORMAT_YUV422 | DVS_YUV_ORDER_YUYV; + pixel_size = 2; + break; + case DRM_FORMAT_YVYU: + dvscntr |= DVS_FORMAT_YUV422 | DVS_YUV_ORDER_YVYU; + pixel_size = 2; + break; + case DRM_FORMAT_UYVY: + dvscntr |= DVS_FORMAT_YUV422 | DVS_YUV_ORDER_UYVY; + pixel_size = 2; + break; + case DRM_FORMAT_VYUY: + dvscntr |= DVS_FORMAT_YUV422 | DVS_YUV_ORDER_VYUY; + pixel_size = 2; + break; + default: + DRM_DEBUG_DRIVER("bad pixel format, assuming RGBX888\n"); + dvscntr |= DVS_FORMAT_RGBX888; + pixel_size = 4; + break; + } + + if (obj->tiling_mode != I915_TILING_NONE) + dvscntr |= DVS_TILED; + + /* must disable */ + dvscntr |= DVS_TRICKLE_FEED_DISABLE; + dvscntr |= DVS_ENABLE; + + /* Sizes are 0 based */ + src_w--; + src_h--; + crtc_w--; + crtc_h--; + + intel_update_sprite_watermarks(dev, pipe, crtc_w, pixel_size); + + if (crtc_w != src_w || crtc_h != src_h) + dvsscale = DVS_SCALE_ENABLE | (src_w << 16) | src_h; + + I915_WRITE(DVSSTRIDE(pipe), fb->pitches[0]); + I915_WRITE(DVSPOS(pipe), (crtc_y << 16) | crtc_x); + if (obj->tiling_mode != I915_TILING_NONE) { + I915_WRITE(DVSTILEOFF(pipe), (y << 16) | x); + } else { + unsigned long offset; + + offset = y * fb->pitches[0] + x * (fb->bits_per_pixel / 8); + I915_WRITE(DVSLINOFF(pipe), offset); + } + I915_WRITE(DVSSIZE(pipe), (crtc_h << 16) | crtc_w); + I915_WRITE(DVSSCALE(pipe), dvsscale); + I915_WRITE(DVSCNTR(pipe), dvscntr); + I915_WRITE(DVSSURF(pipe), obj->gtt_offset); + POSTING_READ(DVSSURF(pipe)); +} + +static void +snb_disable_plane(struct drm_plane *plane) +{ + struct drm_device *dev = plane->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_plane *intel_plane = to_intel_plane(plane); + int pipe = intel_plane->pipe; + + I915_WRITE(DVSCNTR(pipe), I915_READ(DVSCNTR(pipe)) & ~DVS_ENABLE); + /* Disable the scaler */ + I915_WRITE(DVSSCALE(pipe), 0); + /* Flush double buffered register updates */ + I915_WRITE(DVSSURF(pipe), 0); + POSTING_READ(DVSSURF(pipe)); +} + +static void +intel_enable_primary(struct drm_crtc *crtc) +{ + struct drm_device *dev = crtc->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_crtc *intel_crtc = to_intel_crtc(crtc); + int reg = DSPCNTR(intel_crtc->plane); + + I915_WRITE(reg, I915_READ(reg) | DISPLAY_PLANE_ENABLE); +} + +static void +intel_disable_primary(struct drm_crtc *crtc) +{ + struct drm_device *dev = crtc->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_crtc *intel_crtc = to_intel_crtc(crtc); + int reg = DSPCNTR(intel_crtc->plane); + + I915_WRITE(reg, I915_READ(reg) & ~DISPLAY_PLANE_ENABLE); +} + +static int +snb_update_colorkey(struct drm_plane *plane, + struct drm_intel_sprite_colorkey *key) +{ + struct drm_device *dev = plane->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_plane *intel_plane; + u32 dvscntr; + int ret = 0; + + intel_plane = to_intel_plane(plane); + + I915_WRITE(DVSKEYVAL(intel_plane->pipe), key->min_value); + I915_WRITE(DVSKEYMAX(intel_plane->pipe), key->max_value); + I915_WRITE(DVSKEYMSK(intel_plane->pipe), key->channel_mask); + + dvscntr = I915_READ(DVSCNTR(intel_plane->pipe)); + dvscntr &= ~(DVS_SOURCE_KEY | DVS_DEST_KEY); + if (key->flags & I915_SET_COLORKEY_DESTINATION) + dvscntr |= DVS_DEST_KEY; + else if (key->flags & I915_SET_COLORKEY_SOURCE) + dvscntr |= DVS_SOURCE_KEY; + I915_WRITE(DVSCNTR(intel_plane->pipe), dvscntr); + + POSTING_READ(DVSKEYMSK(intel_plane->pipe)); + + return ret; +} + +static void +snb_get_colorkey(struct drm_plane *plane, struct drm_intel_sprite_colorkey *key) +{ + struct drm_device *dev = plane->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_plane *intel_plane; + u32 dvscntr; + + intel_plane = to_intel_plane(plane); + + key->min_value = I915_READ(DVSKEYVAL(intel_plane->pipe)); + key->max_value = I915_READ(DVSKEYMAX(intel_plane->pipe)); + key->channel_mask = I915_READ(DVSKEYMSK(intel_plane->pipe)); + key->flags = 0; + + dvscntr = I915_READ(DVSCNTR(intel_plane->pipe)); + + if (dvscntr & DVS_DEST_KEY) + key->flags = I915_SET_COLORKEY_DESTINATION; + else if (dvscntr & DVS_SOURCE_KEY) + key->flags = I915_SET_COLORKEY_SOURCE; + else + key->flags = I915_SET_COLORKEY_NONE; +} + +static int +intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc, + struct drm_framebuffer *fb, int crtc_x, int crtc_y, + unsigned int crtc_w, unsigned int crtc_h, + uint32_t src_x, uint32_t src_y, + uint32_t src_w, uint32_t src_h) +{ + struct drm_device *dev = plane->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_crtc *intel_crtc = to_intel_crtc(crtc); + struct intel_plane *intel_plane = to_intel_plane(plane); + struct intel_framebuffer *intel_fb; + struct drm_i915_gem_object *obj, *old_obj; + int pipe = intel_plane->pipe; + int ret = 0; + int x = src_x >> 16, y = src_y >> 16; + int primary_w = crtc->mode.hdisplay, primary_h = crtc->mode.vdisplay; + bool disable_primary = false; + + intel_fb = to_intel_framebuffer(fb); + obj = intel_fb->obj; + + old_obj = intel_plane->obj; + + src_w = src_w >> 16; + src_h = src_h >> 16; + + /* Pipe must be running... */ + if (!(I915_READ(PIPECONF(pipe)) & PIPECONF_ENABLE)) + return -EINVAL; + + if (crtc_x >= primary_w || crtc_y >= primary_h) + return -EINVAL; + + /* Don't modify another pipe's plane */ + if (intel_plane->pipe != intel_crtc->pipe) + return -EINVAL; + + /* + * Clamp the width & height into the visible area. Note we don't + * try to scale the source if part of the visible region is offscreen. + * The caller must handle that by adjusting source offset and size. + */ + if ((crtc_x < 0) && ((crtc_x + crtc_w) > 0)) { + crtc_w += crtc_x; + crtc_x = 0; + } + if ((crtc_x + crtc_w) <= 0) /* Nothing to display */ + goto out; + if ((crtc_x + crtc_w) > primary_w) + crtc_w = primary_w - crtc_x; + + if ((crtc_y < 0) && ((crtc_y + crtc_h) > 0)) { + crtc_h += crtc_y; + crtc_y = 0; + } + if ((crtc_y + crtc_h) <= 0) /* Nothing to display */ + goto out; + if (crtc_y + crtc_h > primary_h) + crtc_h = primary_h - crtc_y; + + if (!crtc_w || !crtc_h) /* Again, nothing to display */ + goto out; + + /* + * We can take a larger source and scale it down, but + * only so much... 16x is the max on SNB. + */ + if (((src_w * src_h) / (crtc_w * crtc_h)) > intel_plane->max_downscale) + return -EINVAL; + + /* + * If the sprite is completely covering the primary plane, + * we can disable the primary and save power. + */ + if ((crtc_x == 0) && (crtc_y == 0) && + (crtc_w == primary_w) && (crtc_h == primary_h)) + disable_primary = true; + + mutex_lock(&dev->struct_mutex); + + ret = intel_pin_and_fence_fb_obj(dev, obj, NULL); + if (ret) + goto out_unlock; + + intel_plane->obj = obj; + + /* + * Be sure to re-enable the primary before the sprite is no longer + * covering it fully. + */ + if (!disable_primary && intel_plane->primary_disabled) { + intel_enable_primary(crtc); + intel_plane->primary_disabled = false; + } + + intel_plane->update_plane(plane, fb, obj, crtc_x, crtc_y, + crtc_w, crtc_h, x, y, src_w, src_h); + + if (disable_primary) { + intel_disable_primary(crtc); + intel_plane->primary_disabled = true; + } + + /* Unpin old obj after new one is active to avoid ugliness */ + if (old_obj) { + /* + * It's fairly common to simply update the position of + * an existing object. In that case, we don't need to + * wait for vblank to avoid ugliness, we only need to + * do the pin & ref bookkeeping. + */ + if (old_obj != obj) { + mutex_unlock(&dev->struct_mutex); + intel_wait_for_vblank(dev, to_intel_crtc(crtc)->pipe); + mutex_lock(&dev->struct_mutex); + } + intel_unpin_fb_obj(old_obj); + } + +out_unlock: + mutex_unlock(&dev->struct_mutex); +out: + return ret; +} + +static int +intel_disable_plane(struct drm_plane *plane) +{ + struct drm_device *dev = plane->dev; + struct intel_plane *intel_plane = to_intel_plane(plane); + int ret = 0; + + if (intel_plane->primary_disabled) { + intel_enable_primary(plane->crtc); + intel_plane->primary_disabled = false; + } + + intel_plane->disable_plane(plane); + + if (!intel_plane->obj) + goto out; + + mutex_lock(&dev->struct_mutex); + intel_unpin_fb_obj(intel_plane->obj); + intel_plane->obj = NULL; + mutex_unlock(&dev->struct_mutex); +out: + + return ret; +} + +static void intel_destroy_plane(struct drm_plane *plane) +{ + struct intel_plane *intel_plane = to_intel_plane(plane); + intel_disable_plane(plane); + drm_plane_cleanup(plane); + kfree(intel_plane); +} + +int intel_sprite_set_colorkey(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + struct drm_intel_sprite_colorkey *set = data; + struct drm_i915_private *dev_priv = dev->dev_private; + struct drm_mode_object *obj; + struct drm_plane *plane; + struct intel_plane *intel_plane; + int ret = 0; + + if (!dev_priv) + return -EINVAL; + + /* Make sure we don't try to enable both src & dest simultaneously */ + if ((set->flags & (I915_SET_COLORKEY_DESTINATION | I915_SET_COLORKEY_SOURCE)) == (I915_SET_COLORKEY_DESTINATION | I915_SET_COLORKEY_SOURCE)) + return -EINVAL; + + mutex_lock(&dev->mode_config.mutex); + + obj = drm_mode_object_find(dev, set->plane_id, DRM_MODE_OBJECT_PLANE); + if (!obj) { + ret = -EINVAL; + goto out_unlock; + } + + plane = obj_to_plane(obj); + intel_plane = to_intel_plane(plane); + ret = intel_plane->update_colorkey(plane, set); + +out_unlock: + mutex_unlock(&dev->mode_config.mutex); + return ret; +} + +int intel_sprite_get_colorkey(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + struct drm_intel_sprite_colorkey *get = data; + struct drm_i915_private *dev_priv = dev->dev_private; + struct drm_mode_object *obj; + struct drm_plane *plane; + struct intel_plane *intel_plane; + int ret = 0; + + if (!dev_priv) + return -EINVAL; + + mutex_lock(&dev->mode_config.mutex); + + obj = drm_mode_object_find(dev, get->plane_id, DRM_MODE_OBJECT_PLANE); + if (!obj) { + ret = -EINVAL; + goto out_unlock; + } + + plane = obj_to_plane(obj); + intel_plane = to_intel_plane(plane); + intel_plane->get_colorkey(plane, get); + +out_unlock: + mutex_unlock(&dev->mode_config.mutex); + return ret; +} + +static const struct drm_plane_funcs intel_plane_funcs = { + .update_plane = intel_update_plane, + .disable_plane = intel_disable_plane, + .destroy = intel_destroy_plane, +}; + +static uint32_t snb_plane_formats[] = { + DRM_FORMAT_XBGR8888, + DRM_FORMAT_XRGB8888, + DRM_FORMAT_YUYV, + DRM_FORMAT_YVYU, + DRM_FORMAT_UYVY, + DRM_FORMAT_VYUY, +}; + +int +intel_plane_init(struct drm_device *dev, enum pipe pipe) +{ + struct intel_plane *intel_plane; + unsigned long possible_crtcs; + int ret; + + if (!(IS_GEN6(dev) || IS_GEN7(dev))) + return -ENODEV; + + intel_plane = kzalloc(sizeof(struct intel_plane), GFP_KERNEL); + if (!intel_plane) + return -ENOMEM; + + if (IS_GEN6(dev)) { + intel_plane->max_downscale = 16; + intel_plane->update_plane = snb_update_plane; + intel_plane->disable_plane = snb_disable_plane; + intel_plane->update_colorkey = snb_update_colorkey; + intel_plane->get_colorkey = snb_get_colorkey; + } else if (IS_GEN7(dev)) { + intel_plane->max_downscale = 2; + intel_plane->update_plane = ivb_update_plane; + intel_plane->disable_plane = ivb_disable_plane; + intel_plane->update_colorkey = ivb_update_colorkey; + intel_plane->get_colorkey = ivb_get_colorkey; + } + + intel_plane->pipe = pipe; + possible_crtcs = (1 << pipe); + ret = drm_plane_init(dev, &intel_plane->base, possible_crtcs, + &intel_plane_funcs, snb_plane_formats, + ARRAY_SIZE(snb_plane_formats), false); + if (ret) + kfree(intel_plane); + + return ret; +} + diff --git a/drivers/gpu/drm/i915/intel_tv.c b/drivers/gpu/drm/i915/intel_tv.c index 12041fa..c82b1d4 100644 --- a/drivers/gpu/drm/i915/intel_tv.c +++ b/drivers/gpu/drm/i915/intel_tv.c @@ -674,7 +674,7 @@ static const struct tv_mode tv_modes[] = { .filter_table = filter_table, }, { - .name = "480p@59.94Hz", + .name = "480p", .clock = 107520, .refresh = 59940, .oversample = TV_OVERSAMPLE_4X, @@ -698,30 +698,6 @@ static const struct tv_mode tv_modes[] = { .filter_table = filter_table, }, { - .name = "480p@60Hz", - .clock = 107520, - .refresh = 60000, - .oversample = TV_OVERSAMPLE_4X, - .component_only = 1, - - .hsync_end = 64, .hblank_end = 122, - .hblank_start = 842, .htotal = 856, - - .progressive = true, .trilevel_sync = false, - - .vsync_start_f1 = 12, .vsync_start_f2 = 12, - .vsync_len = 12, - - .veq_ena = false, - - .vi_end_f1 = 44, .vi_end_f2 = 44, - .nbr_end = 479, - - .burst_ena = false, - - .filter_table = filter_table, - }, - { .name = "576p", .clock = 107520, .refresh = 50000, @@ -731,7 +707,7 @@ static const struct tv_mode tv_modes[] = { .hsync_end = 64, .hblank_end = 139, .hblank_start = 859, .htotal = 863, - .progressive = true, .trilevel_sync = false, + .progressive = true, .trilevel_sync = false, .vsync_start_f1 = 10, .vsync_start_f2 = 10, .vsync_len = 10, @@ -770,30 +746,6 @@ static const struct tv_mode tv_modes[] = { .filter_table = filter_table, }, { - .name = "720p@59.94Hz", - .clock = 148800, - .refresh = 59940, - .oversample = TV_OVERSAMPLE_2X, - .component_only = 1, - - .hsync_end = 80, .hblank_end = 300, - .hblank_start = 1580, .htotal = 1651, - - .progressive = true, .trilevel_sync = true, - - .vsync_start_f1 = 10, .vsync_start_f2 = 10, - .vsync_len = 10, - - .veq_ena = false, - - .vi_end_f1 = 29, .vi_end_f2 = 29, - .nbr_end = 719, - - .burst_ena = false, - - .filter_table = filter_table, - }, - { .name = "720p@50Hz", .clock = 148800, .refresh = 50000, @@ -870,32 +822,6 @@ static const struct tv_mode tv_modes[] = { .filter_table = filter_table, }, - { - .name = "1080i@59.94Hz", - .clock = 148800, - .refresh = 29970, - .oversample = TV_OVERSAMPLE_2X, - .component_only = 1, - - .hsync_end = 88, .hblank_end = 235, - .hblank_start = 2155, .htotal = 2201, - - .progressive = false, .trilevel_sync = true, - - .vsync_start_f1 = 4, .vsync_start_f2 = 5, - .vsync_len = 10, - - .veq_ena = true, .veq_start_f1 = 4, - .veq_start_f2 = 4, .veq_len = 10, - - - .vi_end_f1 = 21, .vi_end_f2 = 22, - .nbr_end = 539, - - .burst_ena = false, - - .filter_table = filter_table, - }, }; static struct intel_tv *enc_to_intel_tv(struct drm_encoder *encoder) @@ -1367,7 +1293,7 @@ intel_tv_detect(struct drm_connector *connector, bool force) int type; mode = reported_modes[0]; - drm_mode_set_crtcinfo(&mode, CRTC_INTERLACE_HALVE_V); + drm_mode_set_crtcinfo(&mode, 0); if (intel_tv->base.base.crtc && intel_tv->base.base.crtc->enabled) { type = intel_tv_detect_type(intel_tv, connector); diff --git a/drivers/gpu/drm/mga/mga_dma.c b/drivers/gpu/drm/mga/mga_dma.c index 5ccb65de..507aa3d 100644 --- a/drivers/gpu/drm/mga/mga_dma.c +++ b/drivers/gpu/drm/mga/mga_dma.c @@ -403,6 +403,8 @@ int mga_driver_load(struct drm_device *dev, unsigned long flags) dev_priv->usec_timeout = MGA_DEFAULT_USEC_TIMEOUT; dev_priv->chipset = flags; + pci_set_master(dev->pdev); + dev_priv->mmio_base = pci_resource_start(dev->pdev, 1); dev_priv->mmio_size = pci_resource_len(dev->pdev, 1); diff --git a/drivers/gpu/drm/mga/mga_drv.c b/drivers/gpu/drm/mga/mga_drv.c index 33daa29..f9a925d 100644 --- a/drivers/gpu/drm/mga/mga_drv.c +++ b/drivers/gpu/drm/mga/mga_drv.c @@ -44,6 +44,20 @@ static struct pci_device_id pciidlist[] = { mga_PCI_IDS }; +static const struct file_operations mga_driver_fops = { + .owner = THIS_MODULE, + .open = drm_open, + .release = drm_release, + .unlocked_ioctl = drm_ioctl, + .mmap = drm_mmap, + .poll = drm_poll, + .fasync = drm_fasync, +#ifdef CONFIG_COMPAT + .compat_ioctl = mga_compat_ioctl, +#endif + .llseek = noop_llseek, +}; + static struct drm_driver driver = { .driver_features = DRIVER_USE_AGP | DRIVER_USE_MTRR | DRIVER_PCI_DMA | @@ -64,20 +78,7 @@ static struct drm_driver driver = { .reclaim_buffers = drm_core_reclaim_buffers, .ioctls = mga_ioctls, .dma_ioctl = mga_dma_buffers, - .fops = { - .owner = THIS_MODULE, - .open = drm_open, - .release = drm_release, - .unlocked_ioctl = drm_ioctl, - .mmap = drm_mmap, - .poll = drm_poll, - .fasync = drm_fasync, -#ifdef CONFIG_COMPAT - .compat_ioctl = mga_compat_ioctl, -#endif - .llseek = noop_llseek, - }, - + .fops = &mga_driver_fops, .name = DRIVER_NAME, .desc = DRIVER_DESC, .date = DRIVER_DATE, diff --git a/drivers/gpu/drm/nouveau/Kconfig b/drivers/gpu/drm/nouveau/Kconfig index ca16399..97a8126 100644 --- a/drivers/gpu/drm/nouveau/Kconfig +++ b/drivers/gpu/drm/nouveau/Kconfig @@ -13,6 +13,7 @@ config DRM_NOUVEAU select ACPI_VIDEO if ACPI && X86 && BACKLIGHT_CLASS_DEVICE && VIDEO_OUTPUT_CONTROL && INPUT select ACPI_WMI if ACPI select MXM_WMI if ACPI + select POWER_SUPPLY help Choose this option for open-source nVidia support. diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile index 35ef5b1..1a2ad7e 100644 --- a/drivers/gpu/drm/nouveau/Makefile +++ b/drivers/gpu/drm/nouveau/Makefile @@ -9,19 +9,23 @@ nouveau-y := nouveau_drv.o nouveau_state.o nouveau_channel.o nouveau_mem.o \ nouveau_bo.o nouveau_fence.o nouveau_gem.o nouveau_ttm.o \ nouveau_hw.o nouveau_calc.o nouveau_bios.o nouveau_i2c.o \ nouveau_display.o nouveau_connector.o nouveau_fbcon.o \ - nouveau_dp.o nouveau_ramht.o \ + nouveau_hdmi.o nouveau_dp.o nouveau_ramht.o \ nouveau_pm.o nouveau_volt.o nouveau_perf.o nouveau_temp.o \ - nouveau_mm.o nouveau_vm.o \ + nouveau_mm.o nouveau_vm.o nouveau_mxm.o nouveau_gpio.o \ nv04_timer.o \ nv04_mc.o nv40_mc.o nv50_mc.o \ - nv04_fb.o nv10_fb.o nv30_fb.o nv40_fb.o nv50_fb.o nvc0_fb.o \ + nv04_fb.o nv10_fb.o nv20_fb.o nv30_fb.o nv40_fb.o \ + nv50_fb.o nvc0_fb.o \ nv04_fifo.o nv10_fifo.o nv40_fifo.o nv50_fifo.o nvc0_fifo.o \ nv04_graph.o nv10_graph.o nv20_graph.o \ nv40_graph.o nv50_graph.o nvc0_graph.o \ nv40_grctx.o nv50_grctx.o nvc0_grctx.o \ - nv84_crypt.o \ + nv84_crypt.o nv98_crypt.o \ nva3_copy.o nvc0_copy.o \ nv31_mpeg.o nv50_mpeg.o \ + nv84_bsp.o \ + nv84_vp.o \ + nv98_ppp.o \ nv04_instmem.o nv50_instmem.o nvc0_instmem.o \ nv04_dac.o nv04_dfp.o nv04_tv.o nv17_tv.o nv17_tv_modes.o \ nv04_crtc.o nv04_display.o nv04_cursor.o \ diff --git a/drivers/gpu/drm/nouveau/nouveau_acpi.c b/drivers/gpu/drm/nouveau/nouveau_acpi.c index 3df56c7..284bd25 100644 --- a/drivers/gpu/drm/nouveau/nouveau_acpi.c +++ b/drivers/gpu/drm/nouveau/nouveau_acpi.c @@ -18,12 +18,6 @@ #include -#define NOUVEAU_DSM_SUPPORTED 0x00 -#define NOUVEAU_DSM_SUPPORTED_FUNCTIONS 0x00 - -#define NOUVEAU_DSM_ACTIVE 0x01 -#define NOUVEAU_DSM_ACTIVE_QUERY 0x00 - #define NOUVEAU_DSM_LED 0x02 #define NOUVEAU_DSM_LED_STATE 0x00 #define NOUVEAU_DSM_LED_OFF 0x10 @@ -35,6 +29,9 @@ #define NOUVEAU_DSM_POWER_SPEED 0x01 #define NOUVEAU_DSM_POWER_STAMINA 0x02 +#define NOUVEAU_DSM_OPTIMUS_FN 0x1A +#define NOUVEAU_DSM_OPTIMUS_ARGS 0x03000001 + static struct nouveau_dsm_priv { bool dsm_detected; bool optimus_detected; @@ -61,7 +58,8 @@ static int nouveau_optimus_dsm(acpi_handle handle, int func, int arg, uint32_t * struct acpi_object_list input; union acpi_object params[4]; union acpi_object *obj; - int err; + int i, err; + char args_buff[4]; input.count = 4; input.pointer = params; @@ -73,7 +71,11 @@ static int nouveau_optimus_dsm(acpi_handle handle, int func, int arg, uint32_t * params[2].type = ACPI_TYPE_INTEGER; params[2].integer.value = func; params[3].type = ACPI_TYPE_BUFFER; - params[3].buffer.length = 0; + params[3].buffer.length = 4; + /* ACPI is little endian, AABBCCDD becomes {DD,CC,BB,AA} */ + for (i = 0; i < 4; i++) + args_buff[i] = (arg >> i * 8) & 0xFF; + params[3].buffer.pointer = args_buff; err = acpi_evaluate_object(handle, "_DSM", &input, &output); if (err) { @@ -148,6 +150,23 @@ static int nouveau_dsm(acpi_handle handle, int func, int arg, uint32_t *result) return 0; } +/* Returns 1 if a DSM function is usable and 0 otherwise */ +static int nouveau_test_dsm(acpi_handle test_handle, + int (*dsm_func)(acpi_handle, int, int, uint32_t *), + int sfnc) +{ + u32 result = 0; + + /* Function 0 returns a Buffer containing available functions. The args + * parameter is ignored for function 0, so just put 0 in it */ + if (dsm_func(test_handle, 0, 0, &result)) + return 0; + + /* ACPI Spec v4 9.14.1: if bit 0 is zero, no function is supported. If + * the n-th bit is enabled, function n is supported */ + return result & 1 && result & (1 << sfnc); +} + static int nouveau_dsm_switch_mux(acpi_handle handle, int mux_id) { mxm_wmi_call_mxmx(mux_id == NOUVEAU_DSM_LED_STAMINA ? MXM_MXDS_ADAPTER_IGD : MXM_MXDS_ADAPTER_0); @@ -168,6 +187,10 @@ static int nouveau_dsm_set_discrete_state(acpi_handle handle, enum vga_switchero static int nouveau_dsm_switchto(enum vga_switcheroo_client_id id) { + /* perhaps the _DSM functions are mutually exclusive, but prepare for + * the future */ + if (!nouveau_dsm_priv.dsm_detected && nouveau_dsm_priv.optimus_detected) + return 0; if (id == VGA_SWITCHEROO_IGD) return nouveau_dsm_switch_mux(nouveau_dsm_priv.dhandle, NOUVEAU_DSM_LED_STAMINA); else @@ -180,6 +203,11 @@ static int nouveau_dsm_power_state(enum vga_switcheroo_client_id id, if (id == VGA_SWITCHEROO_IGD) return 0; + /* Optimus laptops have the card already disabled in + * nouveau_switcheroo_set_state */ + if (!nouveau_dsm_priv.dsm_detected && nouveau_dsm_priv.optimus_detected) + return 0; + return nouveau_dsm_set_discrete_state(nouveau_dsm_priv.dhandle, state); } @@ -212,8 +240,7 @@ static int nouveau_dsm_pci_probe(struct pci_dev *pdev) { acpi_handle dhandle, nvidia_handle; acpi_status status; - int ret, retval = 0; - uint32_t result; + int retval = 0; dhandle = DEVICE_ACPI_HANDLE(&pdev->dev); if (!dhandle) @@ -224,13 +251,11 @@ static int nouveau_dsm_pci_probe(struct pci_dev *pdev) return false; } - ret = nouveau_dsm(dhandle, NOUVEAU_DSM_SUPPORTED, - NOUVEAU_DSM_SUPPORTED_FUNCTIONS, &result); - if (ret == 0) + if (nouveau_test_dsm(dhandle, nouveau_dsm, NOUVEAU_DSM_POWER)) retval |= NOUVEAU_DSM_HAS_MUX; - ret = nouveau_optimus_dsm(dhandle, 0, 0, &result); - if (ret == 0) + if (nouveau_test_dsm(dhandle, nouveau_optimus_dsm, + NOUVEAU_DSM_OPTIMUS_FN)) retval |= NOUVEAU_DSM_HAS_OPT; if (retval) @@ -269,15 +294,22 @@ static bool nouveau_dsm_detect(void) } if (vga_count == 2 && has_dsm && guid_valid) { - acpi_get_name(nouveau_dsm_priv.dhandle, ACPI_FULL_PATHNAME, &buffer); + acpi_get_name(nouveau_dsm_priv.dhandle, ACPI_FULL_PATHNAME, + &buffer); printk(KERN_INFO "VGA switcheroo: detected DSM switching method %s handle\n", - acpi_method_name); + acpi_method_name); nouveau_dsm_priv.dsm_detected = true; ret = true; } - if (has_optimus == 1) + if (has_optimus == 1) { + acpi_get_name(nouveau_dsm_priv.dhandle, ACPI_FULL_PATHNAME, + &buffer); + printk(KERN_INFO "VGA switcheroo: detected Optimus DSM method %s handle\n", + acpi_method_name); nouveau_dsm_priv.optimus_detected = true; + ret = true; + } return ret; } @@ -293,6 +325,17 @@ void nouveau_register_dsm_handler(void) vga_switcheroo_register_handler(&nouveau_dsm_handler); } +/* Must be called for Optimus models before the card can be turned off */ +void nouveau_switcheroo_optimus_dsm(void) +{ + u32 result = 0; + if (!nouveau_dsm_priv.optimus_detected) + return; + + nouveau_optimus_dsm(nouveau_dsm_priv.dhandle, NOUVEAU_DSM_OPTIMUS_FN, + NOUVEAU_DSM_OPTIMUS_ARGS, &result); +} + void nouveau_unregister_dsm_handler(void) { vga_switcheroo_unregister_handler(); diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.c b/drivers/gpu/drm/nouveau/nouveau_bios.c index 5fc201b..0be4a81 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bios.c +++ b/drivers/gpu/drm/nouveau/nouveau_bios.c @@ -27,6 +27,7 @@ #include "nouveau_drv.h" #include "nouveau_hw.h" #include "nouveau_encoder.h" +#include "nouveau_gpio.h" #include @@ -34,9 +35,6 @@ #define NV_CIO_CRE_44_HEADA 0x0 #define NV_CIO_CRE_44_HEADB 0x3 #define FEATURE_MOBILE 0x10 /* also FEATURE_QUADRO for BMP */ -#define LEGACY_I2C_CRT 0x80 -#define LEGACY_I2C_PANEL 0x81 -#define LEGACY_I2C_TV 0x82 #define EDID1_LEN 128 @@ -67,195 +65,233 @@ static bool nv_cksum(const uint8_t *data, unsigned int length) } static int -score_vbios(struct drm_device *dev, const uint8_t *data, const bool writeable) +score_vbios(struct nvbios *bios, const bool writeable) { - if (!(data[0] == 0x55 && data[1] == 0xAA)) { - NV_TRACEWARN(dev, "... BIOS signature not found\n"); + if (!bios->data || bios->data[0] != 0x55 || bios->data[1] != 0xAA) { + NV_TRACEWARN(bios->dev, "... BIOS signature not found\n"); return 0; } - if (nv_cksum(data, data[2] * 512)) { - NV_TRACEWARN(dev, "... BIOS checksum invalid\n"); + if (nv_cksum(bios->data, bios->data[2] * 512)) { + NV_TRACEWARN(bios->dev, "... BIOS checksum invalid\n"); /* if a ro image is somewhat bad, it's probably all rubbish */ return writeable ? 2 : 1; - } else - NV_TRACE(dev, "... appears to be valid\n"); + } + NV_TRACE(bios->dev, "... appears to be valid\n"); return 3; } -static void load_vbios_prom(struct drm_device *dev, uint8_t *data) +static void +bios_shadow_prom(struct nvbios *bios) { + struct drm_device *dev = bios->dev; struct drm_nouveau_private *dev_priv = dev->dev_private; - uint32_t pci_nv_20, save_pci_nv_20; - int pcir_ptr; + u32 pcireg, access; + u16 pcir; int i; + /* enable access to rom */ if (dev_priv->card_type >= NV_50) - pci_nv_20 = 0x88050; + pcireg = 0x088050; else - pci_nv_20 = NV_PBUS_PCI_NV_20; + pcireg = NV_PBUS_PCI_NV_20; + access = nv_mask(dev, pcireg, 0x00000001, 0x00000000); - /* enable ROM access */ - save_pci_nv_20 = nvReadMC(dev, pci_nv_20); - nvWriteMC(dev, pci_nv_20, - save_pci_nv_20 & ~NV_PBUS_PCI_NV_20_ROM_SHADOW_ENABLED); + /* bail if no rom signature, with a workaround for a PROM reading + * issue on some chipsets. the first read after a period of + * inactivity returns the wrong result, so retry the first header + * byte a few times before giving up as a workaround + */ + i = 16; + do { + if (nv_rd08(dev, NV_PROM_OFFSET + 0) == 0x55) + break; + } while (i--); - /* bail if no rom signature */ - if (nv_rd08(dev, NV_PROM_OFFSET) != 0x55 || - nv_rd08(dev, NV_PROM_OFFSET + 1) != 0xaa) + if (!i || nv_rd08(dev, NV_PROM_OFFSET + 1) != 0xaa) goto out; /* additional check (see note below) - read PCI record header */ - pcir_ptr = nv_rd08(dev, NV_PROM_OFFSET + 0x18) | - nv_rd08(dev, NV_PROM_OFFSET + 0x19) << 8; - if (nv_rd08(dev, NV_PROM_OFFSET + pcir_ptr) != 'P' || - nv_rd08(dev, NV_PROM_OFFSET + pcir_ptr + 1) != 'C' || - nv_rd08(dev, NV_PROM_OFFSET + pcir_ptr + 2) != 'I' || - nv_rd08(dev, NV_PROM_OFFSET + pcir_ptr + 3) != 'R') + pcir = nv_rd08(dev, NV_PROM_OFFSET + 0x18) | + nv_rd08(dev, NV_PROM_OFFSET + 0x19) << 8; + if (nv_rd08(dev, NV_PROM_OFFSET + pcir + 0) != 'P' || + nv_rd08(dev, NV_PROM_OFFSET + pcir + 1) != 'C' || + nv_rd08(dev, NV_PROM_OFFSET + pcir + 2) != 'I' || + nv_rd08(dev, NV_PROM_OFFSET + pcir + 3) != 'R') goto out; - /* on some 6600GT/6800LE prom reads are messed up. nvclock alleges a - * a good read may be obtained by waiting or re-reading (cargocult: 5x) - * each byte. we'll hope pramin has something usable instead - */ - for (i = 0; i < NV_PROM_SIZE; i++) - data[i] = nv_rd08(dev, NV_PROM_OFFSET + i); + /* read entire bios image to system memory */ + bios->length = nv_rd08(dev, NV_PROM_OFFSET + 2) * 512; + bios->data = kmalloc(bios->length, GFP_KERNEL); + if (bios->data) { + for (i = 0; i < bios->length; i++) + bios->data[i] = nv_rd08(dev, NV_PROM_OFFSET + i); + } out: - /* disable ROM access */ - nvWriteMC(dev, pci_nv_20, - save_pci_nv_20 | NV_PBUS_PCI_NV_20_ROM_SHADOW_ENABLED); + /* disable access to rom */ + nv_wr32(dev, pcireg, access); } -static void load_vbios_pramin(struct drm_device *dev, uint8_t *data) +static void +bios_shadow_pramin(struct nvbios *bios) { + struct drm_device *dev = bios->dev; struct drm_nouveau_private *dev_priv = dev->dev_private; - uint32_t old_bar0_pramin = 0; + u32 bar0 = 0; int i; if (dev_priv->card_type >= NV_50) { u64 addr = (u64)(nv_rd32(dev, 0x619f04) & 0xffffff00) << 8; if (!addr) { - addr = (u64)nv_rd32(dev, 0x1700) << 16; + addr = (u64)nv_rd32(dev, 0x001700) << 16; addr += 0xf0000; } - old_bar0_pramin = nv_rd32(dev, 0x1700); - nv_wr32(dev, 0x1700, addr >> 16); + bar0 = nv_mask(dev, 0x001700, 0xffffffff, addr >> 16); } /* bail if no rom signature */ - if (nv_rd08(dev, NV_PRAMIN_OFFSET) != 0x55 || + if (nv_rd08(dev, NV_PRAMIN_OFFSET + 0) != 0x55 || nv_rd08(dev, NV_PRAMIN_OFFSET + 1) != 0xaa) goto out; - for (i = 0; i < NV_PROM_SIZE; i++) - data[i] = nv_rd08(dev, NV_PRAMIN_OFFSET + i); + bios->length = nv_rd08(dev, NV_PRAMIN_OFFSET + 2) * 512; + bios->data = kmalloc(bios->length, GFP_KERNEL); + if (bios->data) { + for (i = 0; i < bios->length; i++) + bios->data[i] = nv_rd08(dev, NV_PRAMIN_OFFSET + i); + } out: if (dev_priv->card_type >= NV_50) - nv_wr32(dev, 0x1700, old_bar0_pramin); + nv_wr32(dev, 0x001700, bar0); +} + +static void +bios_shadow_pci(struct nvbios *bios) +{ + struct pci_dev *pdev = bios->dev->pdev; + size_t length; + + if (!pci_enable_rom(pdev)) { + void __iomem *rom = pci_map_rom(pdev, &length); + if (rom && length) { + bios->data = kmalloc(length, GFP_KERNEL); + if (bios->data) { + memcpy_fromio(bios->data, rom, length); + bios->length = length; + } + } + if (rom) + pci_unmap_rom(pdev, rom); + + pci_disable_rom(pdev); + } } -static void load_vbios_pci(struct drm_device *dev, uint8_t *data) +static void +bios_shadow_acpi(struct nvbios *bios) { - void __iomem *rom = NULL; - size_t rom_len; - int ret; + struct pci_dev *pdev = bios->dev->pdev; + int ptr, len, ret; + u8 data[3]; - ret = pci_enable_rom(dev->pdev); - if (ret) + if (!nouveau_acpi_rom_supported(pdev)) return; - rom = pci_map_rom(dev->pdev, &rom_len); - if (!rom) - goto out; - memcpy_fromio(data, rom, rom_len); - pci_unmap_rom(dev->pdev, rom); + ret = nouveau_acpi_get_bios_chunk(data, 0, sizeof(data)); + if (ret != sizeof(data)) + return; -out: - pci_disable_rom(dev->pdev); -} + bios->length = min(data[2] * 512, 65536); + bios->data = kmalloc(bios->length, GFP_KERNEL); + if (!bios->data) + return; -static void load_vbios_acpi(struct drm_device *dev, uint8_t *data) -{ - int i; - int ret; - int size = 64 * 1024; + len = bios->length; + ptr = 0; + while (len) { + int size = (len > ROM_BIOS_PAGE) ? ROM_BIOS_PAGE : len; - if (!nouveau_acpi_rom_supported(dev->pdev)) - return; + ret = nouveau_acpi_get_bios_chunk(bios->data, ptr, size); + if (ret != size) { + kfree(bios->data); + bios->data = NULL; + return; + } - for (i = 0; i < (size / ROM_BIOS_PAGE); i++) { - ret = nouveau_acpi_get_bios_chunk(data, - (i * ROM_BIOS_PAGE), - ROM_BIOS_PAGE); - if (ret <= 0) - break; + len -= size; + ptr += size; } - return; } struct methods { const char desc[8]; - void (*loadbios)(struct drm_device *, uint8_t *); + void (*shadow)(struct nvbios *); const bool rw; + int score; + u32 size; + u8 *data; }; -static struct methods shadow_methods[] = { - { "PRAMIN", load_vbios_pramin, true }, - { "PROM", load_vbios_prom, false }, - { "PCIROM", load_vbios_pci, true }, - { "ACPI", load_vbios_acpi, true }, -}; -#define NUM_SHADOW_METHODS ARRAY_SIZE(shadow_methods) - -static bool NVShadowVBIOS(struct drm_device *dev, uint8_t *data) -{ - struct methods *methods = shadow_methods; - int testscore = 3; - int scores[NUM_SHADOW_METHODS], i; +static bool +bios_shadow(struct drm_device *dev) +{ + struct methods shadow_methods[] = { + { "PRAMIN", bios_shadow_pramin, true, 0, 0, NULL }, + { "PROM", bios_shadow_prom, false, 0, 0, NULL }, + { "ACPI", bios_shadow_acpi, true, 0, 0, NULL }, + { "PCIROM", bios_shadow_pci, true, 0, 0, NULL }, + {} + }; + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nvbios *bios = &dev_priv->vbios; + struct methods *mthd, *best; if (nouveau_vbios) { - for (i = 0; i < NUM_SHADOW_METHODS; i++) - if (!strcasecmp(nouveau_vbios, methods[i].desc)) - break; - - if (i < NUM_SHADOW_METHODS) { - NV_INFO(dev, "Attempting to use BIOS image from %s\n", - methods[i].desc); + mthd = shadow_methods; + do { + if (strcasecmp(nouveau_vbios, mthd->desc)) + continue; + NV_INFO(dev, "VBIOS source: %s\n", mthd->desc); - methods[i].loadbios(dev, data); - if (score_vbios(dev, data, methods[i].rw)) + mthd->shadow(bios); + mthd->score = score_vbios(bios, mthd->rw); + if (mthd->score) return true; - } + } while ((++mthd)->shadow); NV_ERROR(dev, "VBIOS source \'%s\' invalid\n", nouveau_vbios); } - for (i = 0; i < NUM_SHADOW_METHODS; i++) { - NV_TRACE(dev, "Attempting to load BIOS image from %s\n", - methods[i].desc); - data[0] = data[1] = 0; /* avoid reuse of previous image */ - methods[i].loadbios(dev, data); - scores[i] = score_vbios(dev, data, methods[i].rw); - if (scores[i] == testscore) - return true; - } - - while (--testscore > 0) { - for (i = 0; i < NUM_SHADOW_METHODS; i++) { - if (scores[i] == testscore) { - NV_TRACE(dev, "Using BIOS image from %s\n", - methods[i].desc); - methods[i].loadbios(dev, data); - return true; - } + mthd = shadow_methods; + do { + NV_TRACE(dev, "Checking %s for VBIOS\n", mthd->desc); + mthd->shadow(bios); + mthd->score = score_vbios(bios, mthd->rw); + mthd->size = bios->length; + mthd->data = bios->data; + } while (mthd->score != 3 && (++mthd)->shadow); + + mthd = shadow_methods; + best = mthd; + do { + if (mthd->score > best->score) { + kfree(best->data); + best = mthd; } + } while ((++mthd)->shadow); + + if (best->score) { + NV_TRACE(dev, "Using VBIOS from %s\n", best->desc); + bios->length = best->size; + bios->data = best->data; + return true; } - NV_ERROR(dev, "No valid BIOS image found\n"); + NV_ERROR(dev, "No valid VBIOS image found\n"); return false; } @@ -723,115 +759,19 @@ static int dcb_entry_idx_from_crtchead(struct drm_device *dev) return dcb_entry; } -static int -read_dcb_i2c_entry(struct drm_device *dev, int dcb_version, uint8_t *i2ctable, int index, struct dcb_i2c_entry *i2c) -{ - uint8_t dcb_i2c_ver = dcb_version, headerlen = 0, entry_len = 4; - int i2c_entries = DCB_MAX_NUM_I2C_ENTRIES; - int recordoffset = 0, rdofs = 1, wrofs = 0; - uint8_t port_type = 0; - - if (!i2ctable) - return -EINVAL; - - if (dcb_version >= 0x30) { - if (i2ctable[0] != dcb_version) /* necessary? */ - NV_WARN(dev, - "DCB I2C table version mismatch (%02X vs %02X)\n", - i2ctable[0], dcb_version); - dcb_i2c_ver = i2ctable[0]; - headerlen = i2ctable[1]; - if (i2ctable[2] <= DCB_MAX_NUM_I2C_ENTRIES) - i2c_entries = i2ctable[2]; - else - NV_WARN(dev, - "DCB I2C table has more entries than indexable " - "(%d entries, max %d)\n", i2ctable[2], - DCB_MAX_NUM_I2C_ENTRIES); - entry_len = i2ctable[3]; - /* [4] is i2c_default_indices, read in parse_dcb_table() */ - } - /* - * It's your own fault if you call this function on a DCB 1.1 BIOS -- - * the test below is for DCB 1.2 - */ - if (dcb_version < 0x14) { - recordoffset = 2; - rdofs = 0; - wrofs = 1; - } - - if (index == 0xf) - return 0; - if (index >= i2c_entries) { - NV_ERROR(dev, "DCB I2C index too big (%d >= %d)\n", - index, i2ctable[2]); - return -ENOENT; - } - if (i2ctable[headerlen + entry_len * index + 3] == 0xff) { - NV_ERROR(dev, "DCB I2C entry invalid\n"); - return -EINVAL; - } - - if (dcb_i2c_ver >= 0x30) { - port_type = i2ctable[headerlen + recordoffset + 3 + entry_len * index]; - - /* - * Fixup for chips using same address offset for read and - * write. - */ - if (port_type == 4) /* seen on C51 */ - rdofs = wrofs = 1; - if (port_type >= 5) /* G80+ */ - rdofs = wrofs = 0; - } - - if (dcb_i2c_ver >= 0x40) { - if (port_type != 5 && port_type != 6) - NV_WARN(dev, "DCB I2C table has port type %d\n", port_type); - - i2c->entry = ROM32(i2ctable[headerlen + recordoffset + entry_len * index]); - } - - i2c->port_type = port_type; - i2c->read = i2ctable[headerlen + recordoffset + rdofs + entry_len * index]; - i2c->write = i2ctable[headerlen + recordoffset + wrofs + entry_len * index]; - - return 0; -} - static struct nouveau_i2c_chan * init_i2c_device_find(struct drm_device *dev, int i2c_index) { - struct drm_nouveau_private *dev_priv = dev->dev_private; - struct dcb_table *dcb = &dev_priv->vbios.dcb; - if (i2c_index == 0xff) { + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct dcb_table *dcb = &dev_priv->vbios.dcb; /* note: dcb_entry_idx_from_crtchead needs pre-script set-up */ - int idx = dcb_entry_idx_from_crtchead(dev), shift = 0; - int default_indices = dcb->i2c_default_indices; + int idx = dcb_entry_idx_from_crtchead(dev); + i2c_index = NV_I2C_DEFAULT(0); if (idx != 0x7f && dcb->entry[idx].i2c_upper_default) - shift = 4; - - i2c_index = (default_indices >> shift) & 0xf; + i2c_index = NV_I2C_DEFAULT(1); } - if (i2c_index == 0x80) /* g80+ */ - i2c_index = dcb->i2c_default_indices & 0xf; - else - if (i2c_index == 0x81) - i2c_index = (dcb->i2c_default_indices & 0xf0) >> 4; - - if (i2c_index >= DCB_MAX_NUM_I2C_ENTRIES) { - NV_ERROR(dev, "invalid i2c_index 0x%x\n", i2c_index); - return NULL; - } - - /* Make sure i2c table entry has been parsed, it may not - * have been if this is a bus not referenced by a DCB encoder - */ - read_dcb_i2c_entry(dev, dcb->version, dcb->i2c_table, - i2c_index, &dcb->i2c[i2c_index]); return nouveau_i2c_find(dev, i2c_index); } @@ -1199,17 +1139,14 @@ init_dp_condition(struct nvbios *bios, uint16_t offset, struct init_exec *iexec) switch (cond) { case 0: - { - struct dcb_connector_table_entry *ent = - &bios->dcb.connector.entry[dcb->connector]; - - if (ent->type != DCB_CONNECTOR_eDP) + entry = dcb_conn(dev, dcb->connector); + if (!entry || entry[0] != DCB_CONNECTOR_eDP) iexec->execute = false; - } break; case 1: case 2: - if (!(entry[5] & cond)) + if ((table[0] < 0x40 && !(entry[5] & cond)) || + (table[0] == 0x40 && !(entry[4] & cond))) iexec->execute = false; break; case 5: @@ -3227,49 +3164,6 @@ init_8d(struct nvbios *bios, uint16_t offset, struct init_exec *iexec) return 1; } -static void -init_gpio_unknv50(struct nvbios *bios, struct dcb_gpio_entry *gpio) -{ - const uint32_t nv50_gpio_ctl[2] = { 0xe100, 0xe28c }; - u32 r, s, v; - - /* Not a clue, needs de-magicing */ - r = nv50_gpio_ctl[gpio->line >> 4]; - s = (gpio->line & 0x0f); - v = bios_rd32(bios, r) & ~(0x00010001 << s); - switch ((gpio->entry & 0x06000000) >> 25) { - case 1: - v |= (0x00000001 << s); - break; - case 2: - v |= (0x00010000 << s); - break; - default: - break; - } - - bios_wr32(bios, r, v); -} - -static void -init_gpio_unknvd0(struct nvbios *bios, struct dcb_gpio_entry *gpio) -{ - u32 v, i; - - v = bios_rd32(bios, 0x00d610 + (gpio->line * 4)); - v &= 0xffffff00; - v |= (gpio->entry & 0x00ff0000) >> 16; - bios_wr32(bios, 0x00d610 + (gpio->line * 4), v); - - i = (gpio->entry & 0x1f000000) >> 24; - if (i) { - v = bios_rd32(bios, 0x00d640 + ((i - 1) * 4)); - v &= 0xffffff00; - v |= gpio->line; - bios_wr32(bios, 0x00d640 + ((i - 1) * 4), v); - } -} - static int init_gpio(struct nvbios *bios, uint16_t offset, struct init_exec *iexec) { @@ -3282,35 +3176,8 @@ init_gpio(struct nvbios *bios, uint16_t offset, struct init_exec *iexec) * each GPIO according to various values listed in each entry */ - struct drm_nouveau_private *dev_priv = bios->dev->dev_private; - struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; - int i; - - if (dev_priv->card_type < NV_50) { - NV_ERROR(bios->dev, "INIT_GPIO on unsupported chipset\n"); - return 1; - } - - if (!iexec->execute) - return 1; - - for (i = 0; i < bios->dcb.gpio.entries; i++) { - struct dcb_gpio_entry *gpio = &bios->dcb.gpio.entry[i]; - - BIOSLOG(bios, "0x%04X: Entry: 0x%08X\n", offset, gpio->entry); - - BIOSLOG(bios, "0x%04X: set gpio 0x%02x, state %d\n", - offset, gpio->tag, gpio->state_default); - - if (!bios->execute) - continue; - - pgpio->set(bios->dev, gpio->tag, gpio->state_default); - if (dev_priv->card_type < NV_D0) - init_gpio_unknv50(bios, gpio); - else - init_gpio_unknvd0(bios, gpio); - } + if (iexec->execute && bios->execute) + nouveau_gpio_reset(bios->dev); return 1; } @@ -4407,18 +4274,6 @@ int nouveau_bios_parse_lvds_table(struct drm_device *dev, int pxclk, bool *dl, b break; } - /* Dell Latitude D620 reports a too-high value for the dual-link - * transition freq, causing us to program the panel incorrectly. - * - * It doesn't appear the VBIOS actually uses its transition freq - * (90000kHz), instead it uses the "Number of LVDS channels" field - * out of the panel ID structure (http://www.spwg.org/). - * - * For the moment, a quirk will do :) - */ - if (nv_match_device(dev, 0x01d7, 0x1028, 0x01c2)) - bios->fp.duallink_transition_clk = 80000; - /* set dual_link flag for EDID case */ if (pxclk && (chip_version < 0x25 || chip_version > 0x28)) bios->fp.dual_link = (pxclk >= bios->fp.duallink_transition_clk); @@ -4541,7 +4396,7 @@ nouveau_bios_run_display_table(struct drm_device *dev, u16 type, int pclk, NV_DEBUG_KMS(dev, "Searching for output entry for %d %d %d\n", dcbent->type, dcbent->location, dcbent->or); for (i = 0; i < table[3]; i++) { - otable = ROMPTR(bios, table[table[1] + (i * table[2])]); + otable = ROMPTR(dev, table[table[1] + (i * table[2])]); if (otable && bios_encoder_match(dcbent, ROM32(otable[0]))) break; } @@ -4719,7 +4574,7 @@ static struct pll_mapping nv84_pll_mapping[] = { { PLL_CORE , 0x004028 }, { PLL_SHADER, 0x004020 }, { PLL_MEMORY, 0x004008 }, - { PLL_UNK05 , 0x004030 }, + { PLL_VDEC , 0x004030 }, { PLL_UNK41 , 0x00e818 }, { PLL_VPLL0 , 0x614100 }, { PLL_VPLL1 , 0x614900 }, @@ -5485,6 +5340,9 @@ bit_table(struct drm_device *dev, u8 id, struct bit_entry *bit) struct nvbios *bios = &dev_priv->vbios; u8 entries, *entry; + if (bios->type != NVBIOS_BIT) + return -ENODEV; + entries = bios->data[bios->offset + 10]; entry = &bios->data[bios->offset + 12]; while (entries--) { @@ -5493,7 +5351,7 @@ bit_table(struct drm_device *dev, u8 id, struct bit_entry *bit) bit->version = entry[1]; bit->length = ROM16(entry[2]); bit->offset = ROM16(entry[4]); - bit->data = ROMPTR(bios, entry[4]); + bit->data = ROMPTR(dev, entry[4]); return 0; } @@ -5598,10 +5456,6 @@ static int parse_bmp_structure(struct drm_device *dev, struct nvbios *bios, unsi uint16_t legacy_scripts_offset, legacy_i2c_offset; /* load needed defaults in case we can't parse this info */ - bios->dcb.i2c[0].write = NV_CIO_CRE_DDC_WR__INDEX; - bios->dcb.i2c[0].read = NV_CIO_CRE_DDC_STATUS__INDEX; - bios->dcb.i2c[1].write = NV_CIO_CRE_DDC0_WR__INDEX; - bios->dcb.i2c[1].read = NV_CIO_CRE_DDC0_STATUS__INDEX; bios->digital_min_front_porch = 0x4b; bios->fmaxvco = 256000; bios->fminvco = 128000; @@ -5709,14 +5563,6 @@ static int parse_bmp_structure(struct drm_device *dev, struct nvbios *bios, unsi bios->legacy.i2c_indices.crt = bios->data[legacy_i2c_offset]; bios->legacy.i2c_indices.tv = bios->data[legacy_i2c_offset + 1]; bios->legacy.i2c_indices.panel = bios->data[legacy_i2c_offset + 2]; - if (bios->data[legacy_i2c_offset + 4]) - bios->dcb.i2c[0].write = bios->data[legacy_i2c_offset + 4]; - if (bios->data[legacy_i2c_offset + 5]) - bios->dcb.i2c[0].read = bios->data[legacy_i2c_offset + 5]; - if (bios->data[legacy_i2c_offset + 6]) - bios->dcb.i2c[1].write = bios->data[legacy_i2c_offset + 6]; - if (bios->data[legacy_i2c_offset + 7]) - bios->dcb.i2c[1].read = bios->data[legacy_i2c_offset + 7]; if (bmplength > 74) { bios->fmaxvco = ROM32(bmp[67]); @@ -5767,286 +5613,128 @@ static uint16_t findstr(uint8_t *data, int n, const uint8_t *str, int len) return 0; } -static struct dcb_gpio_entry * -new_gpio_entry(struct nvbios *bios) -{ - struct drm_device *dev = bios->dev; - struct dcb_gpio_table *gpio = &bios->dcb.gpio; - - if (gpio->entries >= DCB_MAX_NUM_GPIO_ENTRIES) { - NV_ERROR(dev, "exceeded maximum number of gpio entries!!\n"); - return NULL; - } - - return &gpio->entry[gpio->entries++]; -} - -struct dcb_gpio_entry * -nouveau_bios_gpio_entry(struct drm_device *dev, enum dcb_gpio_tag tag) +void * +dcb_table(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->vbios; - int i; - - for (i = 0; i < bios->dcb.gpio.entries; i++) { - if (bios->dcb.gpio.entry[i].tag != tag) - continue; + u8 *dcb = NULL; - return &bios->dcb.gpio.entry[i]; + if (dev_priv->card_type > NV_04) + dcb = ROMPTR(dev, dev_priv->vbios.data[0x36]); + if (!dcb) { + NV_WARNONCE(dev, "No DCB data found in VBIOS\n"); + return NULL; } - return NULL; -} - -static void -parse_dcb_gpio_table(struct nvbios *bios) -{ - struct drm_device *dev = bios->dev; - struct dcb_gpio_entry *e; - u8 headerlen, entries, recordlen; - u8 *dcb, *gpio = NULL, *entry; - int i; - - dcb = ROMPTR(bios, bios->data[0x36]); + if (dcb[0] >= 0x41) { + NV_WARNONCE(dev, "DCB version 0x%02x unknown\n", dcb[0]); + return NULL; + } else if (dcb[0] >= 0x30) { - gpio = ROMPTR(bios, dcb[10]); - if (!gpio) - goto no_table; - - headerlen = gpio[1]; - entries = gpio[2]; - recordlen = gpio[3]; + if (ROM32(dcb[6]) == 0x4edcbdcb) + return dcb; } else - if (dcb[0] >= 0x22 && dcb[-1] >= 0x13) { - gpio = ROMPTR(bios, dcb[-15]); - if (!gpio) - goto no_table; - - headerlen = 3; - entries = gpio[2]; - recordlen = gpio[1]; + if (dcb[0] >= 0x20) { + if (ROM32(dcb[4]) == 0x4edcbdcb) + return dcb; } else - if (dcb[0] >= 0x22) { - /* No GPIO table present, parse the TVDAC GPIO data. */ - uint8_t *tvdac_gpio = &dcb[-5]; - - if (tvdac_gpio[0] & 1) { - e = new_gpio_entry(bios); - e->tag = DCB_GPIO_TVDAC0; - e->line = tvdac_gpio[1] >> 4; - e->invert = tvdac_gpio[0] & 2; - } - - goto no_table; + if (dcb[0] >= 0x15) { + if (!memcmp(&dcb[-7], "DEV_REC", 7)) + return dcb; } else { - NV_DEBUG(dev, "no/unknown gpio table on DCB 0x%02x\n", dcb[0]); - goto no_table; - } - - entry = gpio + headerlen; - for (i = 0; i < entries; i++, entry += recordlen) { - e = new_gpio_entry(bios); - if (!e) - break; - - if (gpio[0] < 0x40) { - e->entry = ROM16(entry[0]); - e->tag = (e->entry & 0x07e0) >> 5; - if (e->tag == 0x3f) { - bios->dcb.gpio.entries--; - continue; - } - - e->line = (e->entry & 0x001f); - e->invert = ((e->entry & 0xf800) >> 11) != 4; - } else { - e->entry = ROM32(entry[0]); - e->tag = (e->entry & 0x0000ff00) >> 8; - if (e->tag == 0xff) { - bios->dcb.gpio.entries--; - continue; - } - - e->line = (e->entry & 0x0000001f) >> 0; - if (gpio[0] == 0x40) { - e->state_default = (e->entry & 0x01000000) >> 24; - e->state[0] = (e->entry & 0x18000000) >> 27; - e->state[1] = (e->entry & 0x60000000) >> 29; - } else { - e->state_default = (e->entry & 0x00000080) >> 7; - e->state[0] = (entry[4] >> 4) & 3; - e->state[1] = (entry[4] >> 6) & 3; - } - } - } - -no_table: - /* Apple iMac G4 NV18 */ - if (nv_match_device(dev, 0x0189, 0x10de, 0x0010)) { - e = new_gpio_entry(bios); - if (e) { - e->tag = DCB_GPIO_TVDAC0; - e->line = 4; - } - } -} - -struct dcb_connector_table_entry * -nouveau_bios_connector_entry(struct drm_device *dev, int index) -{ - struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->vbios; - struct dcb_connector_table_entry *cte; - - if (index >= bios->dcb.connector.entries) - return NULL; - - cte = &bios->dcb.connector.entry[index]; - if (cte->type == 0xff) + /* + * v1.4 (some NV15/16, NV11+) seems the same as v1.5, but + * always has the same single (crt) entry, even when tv-out + * present, so the conclusion is this version cannot really + * be used. + * + * v1.2 tables (some NV6/10, and NV15+) normally have the + * same 5 entries, which are not specific to the card and so + * no use. + * + * v1.2 does have an I2C table that read_dcb_i2c_table can + * handle, but cards exist (nv11 in #14821) with a bad i2c + * table pointer, so use the indices parsed in + * parse_bmp_structure. + * + * v1.1 (NV5+, maybe some NV4) is entirely unhelpful + */ + NV_WARNONCE(dev, "No useful DCB data in VBIOS\n"); return NULL; + } - return cte; + NV_WARNONCE(dev, "DCB header validation failed\n"); + return NULL; } -static enum dcb_connector_type -divine_connector_type(struct nvbios *bios, int index) +void * +dcb_outp(struct drm_device *dev, u8 idx) { - struct dcb_table *dcb = &bios->dcb; - unsigned encoders = 0, type = DCB_CONNECTOR_NONE; - int i; - - for (i = 0; i < dcb->entries; i++) { - if (dcb->entry[i].connector == index) - encoders |= (1 << dcb->entry[i].type); - } - - if (encoders & (1 << OUTPUT_DP)) { - if (encoders & (1 << OUTPUT_TMDS)) - type = DCB_CONNECTOR_DP; - else - type = DCB_CONNECTOR_eDP; + u8 *dcb = dcb_table(dev); + if (dcb && dcb[0] >= 0x30) { + if (idx < dcb[2]) + return dcb + dcb[1] + (idx * dcb[3]); } else - if (encoders & (1 << OUTPUT_TMDS)) { - if (encoders & (1 << OUTPUT_ANALOG)) - type = DCB_CONNECTOR_DVI_I; - else - type = DCB_CONNECTOR_DVI_D; - } else - if (encoders & (1 << OUTPUT_ANALOG)) { - type = DCB_CONNECTOR_VGA; - } else - if (encoders & (1 << OUTPUT_LVDS)) { - type = DCB_CONNECTOR_LVDS; + if (dcb && dcb[0] >= 0x20) { + u8 *i2c = ROMPTR(dev, dcb[2]); + u8 *ent = dcb + 8 + (idx * 8); + if (i2c && ent < i2c) + return ent; } else - if (encoders & (1 << OUTPUT_TV)) { - type = DCB_CONNECTOR_TV_0; + if (dcb && dcb[0] >= 0x15) { + u8 *i2c = ROMPTR(dev, dcb[2]); + u8 *ent = dcb + 4 + (idx * 10); + if (i2c && ent < i2c) + return ent; } - return type; + return NULL; } -static void -apply_dcb_connector_quirks(struct nvbios *bios, int idx) -{ - struct dcb_connector_table_entry *cte = &bios->dcb.connector.entry[idx]; - struct drm_device *dev = bios->dev; +int +dcb_outp_foreach(struct drm_device *dev, void *data, + int (*exec)(struct drm_device *, void *, int idx, u8 *outp)) +{ + int ret, idx = -1; + u8 *outp = NULL; + while ((outp = dcb_outp(dev, ++idx))) { + if (ROM32(outp[0]) == 0x00000000) + break; /* seen on an NV11 with DCB v1.5 */ + if (ROM32(outp[0]) == 0xffffffff) + break; /* seen on an NV17 with DCB v2.0 */ + + if ((outp[0] & 0x0f) == OUTPUT_UNUSED) + continue; + if ((outp[0] & 0x0f) == OUTPUT_EOL) + break; - /* Gigabyte NX85T */ - if (nv_match_device(dev, 0x0421, 0x1458, 0x344c)) { - if (cte->type == DCB_CONNECTOR_HDMI_1) - cte->type = DCB_CONNECTOR_DVI_I; + ret = exec(dev, data, idx, outp); + if (ret) + return ret; } - /* Gigabyte GV-NX86T512H */ - if (nv_match_device(dev, 0x0402, 0x1458, 0x3455)) { - if (cte->type == DCB_CONNECTOR_HDMI_1) - cte->type = DCB_CONNECTOR_DVI_I; - } + return 0; } -static const u8 hpd_gpio[16] = { - 0xff, 0x07, 0x08, 0xff, 0xff, 0x51, 0x52, 0xff, - 0xff, 0xff, 0xff, 0xff, 0xff, 0x5e, 0x5f, 0x60, -}; - -static void -parse_dcb_connector_table(struct nvbios *bios) +u8 * +dcb_conntab(struct drm_device *dev) { - struct drm_device *dev = bios->dev; - struct dcb_connector_table *ct = &bios->dcb.connector; - struct dcb_connector_table_entry *cte; - uint8_t *conntab = &bios->data[bios->dcb.connector_table_ptr]; - uint8_t *entry; - int i; - - if (!bios->dcb.connector_table_ptr) { - NV_DEBUG_KMS(dev, "No DCB connector table present\n"); - return; - } - - NV_INFO(dev, "DCB connector table: VHER 0x%02x %d %d %d\n", - conntab[0], conntab[1], conntab[2], conntab[3]); - if ((conntab[0] != 0x30 && conntab[0] != 0x40) || - (conntab[3] != 2 && conntab[3] != 4)) { - NV_ERROR(dev, " Unknown! Please report.\n"); - return; + u8 *dcb = dcb_table(dev); + if (dcb && dcb[0] >= 0x30 && dcb[1] >= 0x16) { + u8 *conntab = ROMPTR(dev, dcb[0x14]); + if (conntab && conntab[0] >= 0x30 && conntab[0] <= 0x40) + return conntab; } + return NULL; +} - ct->entries = conntab[2]; - - entry = conntab + conntab[1]; - cte = &ct->entry[0]; - for (i = 0; i < conntab[2]; i++, entry += conntab[3], cte++) { - cte->index = i; - if (conntab[3] == 2) - cte->entry = ROM16(entry[0]); - else - cte->entry = ROM32(entry[0]); - - cte->type = (cte->entry & 0x000000ff) >> 0; - cte->index2 = (cte->entry & 0x00000f00) >> 8; - - cte->gpio_tag = ffs((cte->entry & 0x07033000) >> 12); - cte->gpio_tag = hpd_gpio[cte->gpio_tag]; - - if (cte->type == 0xff) - continue; - - apply_dcb_connector_quirks(bios, i); - - NV_INFO(dev, " %d: 0x%08x: type 0x%02x idx %d tag 0x%02x\n", - i, cte->entry, cte->type, cte->index, cte->gpio_tag); - - /* check for known types, fallback to guessing the type - * from attached encoders if we hit an unknown. - */ - switch (cte->type) { - case DCB_CONNECTOR_VGA: - case DCB_CONNECTOR_TV_0: - case DCB_CONNECTOR_TV_1: - case DCB_CONNECTOR_TV_3: - case DCB_CONNECTOR_DVI_I: - case DCB_CONNECTOR_DVI_D: - case DCB_CONNECTOR_LVDS: - case DCB_CONNECTOR_LVDS_SPWG: - case DCB_CONNECTOR_DP: - case DCB_CONNECTOR_eDP: - case DCB_CONNECTOR_HDMI_0: - case DCB_CONNECTOR_HDMI_1: - break; - default: - cte->type = divine_connector_type(bios, cte->index); - NV_WARN(dev, "unknown type, using 0x%02x\n", cte->type); - break; - } - - if (nouveau_override_conntype) { - int type = divine_connector_type(bios, cte->index); - if (type != cte->type) - NV_WARN(dev, " -> type 0x%02x\n", cte->type); - } - - } +u8 * +dcb_conn(struct drm_device *dev, u8 idx) +{ + u8 *conntab = dcb_conntab(dev); + if (conntab && idx < conntab[2]) + return conntab + conntab[1] + (idx * conntab[3]); + return NULL; } static struct dcb_entry *new_dcb_entry(struct dcb_table *dcb) @@ -6079,8 +5767,7 @@ parse_dcb20_entry(struct drm_device *dev, struct dcb_table *dcb, entry->type = conn & 0xf; entry->i2c_index = (conn >> 4) & 0xf; entry->heads = (conn >> 8) & 0xf; - if (dcb->version >= 0x40) - entry->connector = (conn >> 12) & 0xf; + entry->connector = (conn >> 12) & 0xf; entry->bus = (conn >> 16) & 0xf; entry->location = (conn >> 20) & 0x3; entry->or = (conn >> 24) & 0xf; @@ -6252,25 +5939,6 @@ parse_dcb15_entry(struct drm_device *dev, struct dcb_table *dcb, return true; } -static bool parse_dcb_entry(struct drm_device *dev, struct dcb_table *dcb, - uint32_t conn, uint32_t conf) -{ - struct dcb_entry *entry = new_dcb_entry(dcb); - bool ret; - - if (dcb->version >= 0x20) - ret = parse_dcb20_entry(dev, dcb, conn, conf, entry); - else - ret = parse_dcb15_entry(dev, dcb, conn, conf, entry); - if (!ret) - return ret; - - read_dcb_i2c_entry(dev, dcb->version, dcb->i2c_table, - entry->i2c_index, &dcb->i2c[entry->i2c_index]); - - return true; -} - static void merge_like_dcb_entries(struct drm_device *dev, struct dcb_table *dcb) { @@ -6431,154 +6099,122 @@ fabricate_dcb_encoder_table(struct drm_device *dev, struct nvbios *bios) #endif /* Make up some sane defaults */ - fabricate_dcb_output(dcb, OUTPUT_ANALOG, LEGACY_I2C_CRT, 1, 1); + fabricate_dcb_output(dcb, OUTPUT_ANALOG, + bios->legacy.i2c_indices.crt, 1, 1); if (nv04_tv_identify(dev, bios->legacy.i2c_indices.tv) >= 0) - fabricate_dcb_output(dcb, OUTPUT_TV, LEGACY_I2C_TV, + fabricate_dcb_output(dcb, OUTPUT_TV, + bios->legacy.i2c_indices.tv, all_heads, 0); else if (bios->tmds.output0_script_ptr || bios->tmds.output1_script_ptr) - fabricate_dcb_output(dcb, OUTPUT_TMDS, LEGACY_I2C_PANEL, + fabricate_dcb_output(dcb, OUTPUT_TMDS, + bios->legacy.i2c_indices.panel, all_heads, 1); } static int -parse_dcb_table(struct drm_device *dev, struct nvbios *bios) +parse_dcb_entry(struct drm_device *dev, void *data, int idx, u8 *outp) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct dcb_table *dcb = &bios->dcb; - uint16_t dcbptr = 0, i2ctabptr = 0; - uint8_t *dcbtable; - uint8_t headerlen = 0x4, entries = DCB_MAX_NUM_ENTRIES; - bool configblock = true; - int recordlength = 8, confofs = 4; - int i; - - /* get the offset from 0x36 */ - if (dev_priv->card_type > NV_04) { - dcbptr = ROM16(bios->data[0x36]); - if (dcbptr == 0x0000) - NV_WARN(dev, "No output data (DCB) found in BIOS\n"); - } - - /* this situation likely means a really old card, pre DCB */ - if (dcbptr == 0x0) { - fabricate_dcb_encoder_table(dev, bios); - return 0; - } - - dcbtable = &bios->data[dcbptr]; - - /* get DCB version */ - dcb->version = dcbtable[0]; - NV_TRACE(dev, "Found Display Configuration Block version %d.%d\n", - dcb->version >> 4, dcb->version & 0xf); - - if (dcb->version >= 0x20) { /* NV17+ */ - uint32_t sig; + struct dcb_table *dcb = &dev_priv->vbios.dcb; + u32 conf = (dcb->version >= 0x20) ? ROM32(outp[4]) : ROM32(outp[6]); + u32 conn = ROM32(outp[0]); + bool ret; - if (dcb->version >= 0x30) { /* NV40+ */ - headerlen = dcbtable[1]; - entries = dcbtable[2]; - recordlength = dcbtable[3]; - i2ctabptr = ROM16(dcbtable[4]); - sig = ROM32(dcbtable[6]); - dcb->gpio_table_ptr = ROM16(dcbtable[10]); - dcb->connector_table_ptr = ROM16(dcbtable[20]); - } else { - i2ctabptr = ROM16(dcbtable[2]); - sig = ROM32(dcbtable[4]); - headerlen = 8; - } + if (apply_dcb_encoder_quirks(dev, idx, &conn, &conf)) { + struct dcb_entry *entry = new_dcb_entry(dcb); - if (sig != 0x4edcbdcb) { - NV_ERROR(dev, "Bad Display Configuration Block " - "signature (%08X)\n", sig); - return -EINVAL; - } - } else if (dcb->version >= 0x15) { /* some NV11 and NV20 */ - char sig[8] = { 0 }; + NV_TRACEWARN(dev, "DCB outp %02d: %08x %08x\n", idx, conn, conf); - strncpy(sig, (char *)&dcbtable[-7], 7); - i2ctabptr = ROM16(dcbtable[2]); - recordlength = 10; - confofs = 6; + if (dcb->version >= 0x20) + ret = parse_dcb20_entry(dev, dcb, conn, conf, entry); + else + ret = parse_dcb15_entry(dev, dcb, conn, conf, entry); + if (!ret) + return 1; /* stop parsing */ - if (strcmp(sig, "DEV_REC")) { - NV_ERROR(dev, "Bad Display Configuration Block " - "signature (%s)\n", sig); - return -EINVAL; - } - } else { - /* - * v1.4 (some NV15/16, NV11+) seems the same as v1.5, but always - * has the same single (crt) entry, even when tv-out present, so - * the conclusion is this version cannot really be used. - * v1.2 tables (some NV6/10, and NV15+) normally have the same - * 5 entries, which are not specific to the card and so no use. - * v1.2 does have an I2C table that read_dcb_i2c_table can - * handle, but cards exist (nv11 in #14821) with a bad i2c table - * pointer, so use the indices parsed in parse_bmp_structure. - * v1.1 (NV5+, maybe some NV4) is entirely unhelpful + /* Ignore the I2C index for on-chip TV-out, as there + * are cards with bogus values (nv31m in bug 23212), + * and it's otherwise useless. */ - NV_TRACEWARN(dev, "No useful information in BIOS output table; " - "adding all possible outputs\n"); - fabricate_dcb_encoder_table(dev, bios); - return 0; + if (entry->type == OUTPUT_TV && + entry->location == DCB_LOC_ON_CHIP) + entry->i2c_index = 0x0f; } - if (!i2ctabptr) - NV_WARN(dev, "No pointer to DCB I2C port table\n"); - else { - dcb->i2c_table = &bios->data[i2ctabptr]; - if (dcb->version >= 0x30) - dcb->i2c_default_indices = dcb->i2c_table[4]; + return 0; +} - /* - * Parse the "management" I2C bus, used for hardware - * monitoring and some external TMDS transmitters. - */ - if (dcb->version >= 0x22) { - int idx = (dcb->version >= 0x40 ? - dcb->i2c_default_indices & 0xf : - 2); +static void +dcb_fake_connectors(struct nvbios *bios) +{ + struct dcb_table *dcbt = &bios->dcb; + u8 map[16] = { }; + int i, idx = 0; - read_dcb_i2c_entry(dev, dcb->version, dcb->i2c_table, - idx, &dcb->i2c[idx]); + /* heuristic: if we ever get a non-zero connector field, assume + * that all the indices are valid and we don't need fake them. + * + * and, as usual, a blacklist of boards with bad bios data.. + */ + if (!nv_match_device(bios->dev, 0x0392, 0x107d, 0x20a2)) { + for (i = 0; i < dcbt->entries; i++) { + if (dcbt->entry[i].connector) + return; } } - if (entries > DCB_MAX_NUM_ENTRIES) - entries = DCB_MAX_NUM_ENTRIES; - - for (i = 0; i < entries; i++) { - uint32_t connection, config = 0; - - connection = ROM32(dcbtable[headerlen + recordlength * i]); - if (configblock) - config = ROM32(dcbtable[headerlen + confofs + recordlength * i]); - - /* seen on an NV11 with DCB v1.5 */ - if (connection == 0x00000000) - break; + /* no useful connector info available, we need to make it up + * ourselves. the rule here is: anything on the same i2c bus + * is considered to be on the same connector. any output + * without an associated i2c bus is assigned its own unique + * connector index. + */ + for (i = 0; i < dcbt->entries; i++) { + u8 i2c = dcbt->entry[i].i2c_index; + if (i2c == 0x0f) { + dcbt->entry[i].connector = idx++; + } else { + if (!map[i2c]) + map[i2c] = ++idx; + dcbt->entry[i].connector = map[i2c] - 1; + } + } - /* seen on an NV17 with DCB v2.0 */ - if (connection == 0xffffffff) - break; + /* if we created more than one connector, destroy the connector + * table - just in case it has random, rather than stub, entries. + */ + if (i > 1) { + u8 *conntab = dcb_conntab(bios->dev); + if (conntab) + conntab[0] = 0x00; + } +} - if ((connection & 0x0000000f) == 0x0000000f) - continue; +static int +parse_dcb_table(struct drm_device *dev, struct nvbios *bios) +{ + struct dcb_table *dcb = &bios->dcb; + u8 *dcbt, *conn; + int idx; + + dcbt = dcb_table(dev); + if (!dcbt) { + /* handle pre-DCB boards */ + if (bios->type == NVBIOS_BMP) { + fabricate_dcb_encoder_table(dev, bios); + return 0; + } - if (!apply_dcb_encoder_quirks(dev, i, &connection, &config)) - continue; + return -EINVAL; + } - NV_TRACEWARN(dev, "Raw DCB entry %d: %08x %08x\n", - dcb->entries, connection, config); + NV_TRACE(dev, "DCB version %d.%d\n", dcbt[0] >> 4, dcbt[0] & 0xf); - if (!parse_dcb_entry(dev, dcb, connection, config)) - break; - } + dcb->version = dcbt[0]; + dcb_outp_foreach(dev, NULL, parse_dcb_entry); /* * apart for v2.1+ not being known for requiring merging, this @@ -6590,77 +6226,19 @@ parse_dcb_table(struct drm_device *dev, struct nvbios *bios) if (!dcb->entries) return -ENXIO; - parse_dcb_gpio_table(bios); - parse_dcb_connector_table(bios); - return 0; -} - -static void -fixup_legacy_connector(struct nvbios *bios) -{ - struct dcb_table *dcb = &bios->dcb; - int i, i2c, i2c_conn[DCB_MAX_NUM_I2C_ENTRIES] = { }; - - /* - * DCB 3.0 also has the table in most cases, but there are some cards - * where the table is filled with stub entries, and the DCB entriy - * indices are all 0. We don't need the connector indices on pre-G80 - * chips (yet?) so limit the use to DCB 4.0 and above. - */ - if (dcb->version >= 0x40) - return; - - dcb->connector.entries = 0; - - /* - * No known connector info before v3.0, so make it up. the rule here - * is: anything on the same i2c bus is considered to be on the same - * connector. any output without an associated i2c bus is assigned - * its own unique connector index. - */ - for (i = 0; i < dcb->entries; i++) { - /* - * Ignore the I2C index for on-chip TV-out, as there - * are cards with bogus values (nv31m in bug 23212), - * and it's otherwise useless. - */ - if (dcb->entry[i].type == OUTPUT_TV && - dcb->entry[i].location == DCB_LOC_ON_CHIP) - dcb->entry[i].i2c_index = 0xf; - i2c = dcb->entry[i].i2c_index; - - if (i2c_conn[i2c]) { - dcb->entry[i].connector = i2c_conn[i2c] - 1; - continue; + /* dump connector table entries to log, if any exist */ + idx = -1; + while ((conn = dcb_conn(dev, ++idx))) { + if (conn[0] != 0xff) { + NV_TRACE(dev, "DCB conn %02d: ", idx); + if (dcb_conntab(dev)[3] < 4) + printk("%04x\n", ROM16(conn[0])); + else + printk("%08x\n", ROM32(conn[0])); } - - dcb->entry[i].connector = dcb->connector.entries++; - if (i2c != 0xf) - i2c_conn[i2c] = dcb->connector.entries; - } - - /* Fake the connector table as well as just connector indices */ - for (i = 0; i < dcb->connector.entries; i++) { - dcb->connector.entry[i].index = i; - dcb->connector.entry[i].type = divine_connector_type(bios, i); - dcb->connector.entry[i].gpio_tag = 0xff; - } -} - -static void -fixup_legacy_i2c(struct nvbios *bios) -{ - struct dcb_table *dcb = &bios->dcb; - int i; - - for (i = 0; i < dcb->entries; i++) { - if (dcb->entry[i].i2c_index == LEGACY_I2C_CRT) - dcb->entry[i].i2c_index = bios->legacy.i2c_indices.crt; - if (dcb->entry[i].i2c_index == LEGACY_I2C_PANEL) - dcb->entry[i].i2c_index = bios->legacy.i2c_indices.panel; - if (dcb->entry[i].i2c_index == LEGACY_I2C_TV) - dcb->entry[i].i2c_index = bios->legacy.i2c_indices.tv; } + dcb_fake_connectors(bios); + return 0; } static int load_nv17_hwsq_ucode_entry(struct drm_device *dev, struct nvbios *bios, uint16_t hwsq_offset, int entry) @@ -6799,11 +6377,7 @@ static bool NVInitVBIOS(struct drm_device *dev) spin_lock_init(&bios->lock); bios->dev = dev; - if (!NVShadowVBIOS(dev, bios->data)) - return false; - - bios->length = NV_PROM_SIZE; - return true; + return bios_shadow(dev); } static int nouveau_parse_vbios_struct(struct drm_device *dev) @@ -6879,19 +6453,6 @@ nouveau_run_vbios_init(struct drm_device *dev) return ret; } -static void -nouveau_bios_i2c_devices_takedown(struct drm_device *dev) -{ - struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->vbios; - struct dcb_i2c_entry *entry; - int i; - - entry = &bios->dcb.i2c[0]; - for (i = 0; i < DCB_MAX_NUM_I2C_ENTRIES; i++, entry++) - nouveau_i2c_fini(dev, entry); -} - static bool nouveau_bios_posted(struct drm_device *dev) { @@ -6928,12 +6489,17 @@ nouveau_bios_init(struct drm_device *dev) if (ret) return ret; - ret = parse_dcb_table(dev, bios); + ret = nouveau_i2c_init(dev); if (ret) return ret; - fixup_legacy_i2c(bios); - fixup_legacy_connector(bios); + ret = nouveau_mxm_init(dev); + if (ret) + return ret; + + ret = parse_dcb_table(dev, bios); + if (ret) + return ret; if (!bios->major_version) /* we don't run version 0 bios */ return 0; @@ -6971,5 +6537,10 @@ nouveau_bios_init(struct drm_device *dev) void nouveau_bios_takedown(struct drm_device *dev) { - nouveau_bios_i2c_devices_takedown(dev); + struct drm_nouveau_private *dev_priv = dev->dev_private; + + nouveau_mxm_fini(dev); + nouveau_i2c_fini(dev); + + kfree(dev_priv->vbios.data); } diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.h b/drivers/gpu/drm/nouveau/nouveau_bios.h index 8adb69e..298a3af 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bios.h +++ b/drivers/gpu/drm/nouveau/nouveau_bios.h @@ -34,9 +34,14 @@ #define DCB_LOC_ON_CHIP 0 -#define ROM16(x) le16_to_cpu(*(uint16_t *)&(x)) -#define ROM32(x) le32_to_cpu(*(uint32_t *)&(x)) -#define ROMPTR(bios, x) (ROM16(x) ? &(bios)->data[ROM16(x)] : NULL) +#define ROM16(x) le16_to_cpu(*(u16 *)&(x)) +#define ROM32(x) le32_to_cpu(*(u32 *)&(x)) +#define ROM48(x) ({ u8 *p = &(x); (u64)ROM16(p[4]) << 32 | ROM32(p[0]); }) +#define ROM64(x) le64_to_cpu(*(u64 *)&(x)) +#define ROMPTR(d,x) ({ \ + struct drm_nouveau_private *dev_priv = (d)->dev_private; \ + ROM16(x) ? &dev_priv->vbios.data[ROM16(x)] : NULL; \ +}) struct bit_entry { uint8_t id; @@ -48,30 +53,13 @@ struct bit_entry { int bit_table(struct drm_device *, u8 id, struct bit_entry *); -struct dcb_i2c_entry { - uint32_t entry; - uint8_t port_type; - uint8_t read, write; - struct nouveau_i2c_chan *chan; -}; - enum dcb_gpio_tag { - DCB_GPIO_TVDAC0 = 0xc, + DCB_GPIO_PANEL_POWER = 0x01, + DCB_GPIO_TVDAC0 = 0x0c, DCB_GPIO_TVDAC1 = 0x2d, -}; - -struct dcb_gpio_entry { - enum dcb_gpio_tag tag; - int line; - bool invert; - uint32_t entry; - uint8_t state_default; - uint8_t state[2]; -}; - -struct dcb_gpio_table { - int entries; - struct dcb_gpio_entry entry[DCB_MAX_NUM_GPIO_ENTRIES]; + DCB_GPIO_PWM_FAN = 0x09, + DCB_GPIO_FAN_SENSE = 0x3d, + DCB_GPIO_UNUSED = 0xff }; enum dcb_connector_type { @@ -81,29 +69,19 @@ enum dcb_connector_type { DCB_CONNECTOR_TV_3 = 0x13, DCB_CONNECTOR_DVI_I = 0x30, DCB_CONNECTOR_DVI_D = 0x31, + DCB_CONNECTOR_DMS59_0 = 0x38, + DCB_CONNECTOR_DMS59_1 = 0x39, DCB_CONNECTOR_LVDS = 0x40, DCB_CONNECTOR_LVDS_SPWG = 0x41, DCB_CONNECTOR_DP = 0x46, DCB_CONNECTOR_eDP = 0x47, DCB_CONNECTOR_HDMI_0 = 0x60, DCB_CONNECTOR_HDMI_1 = 0x61, + DCB_CONNECTOR_DMS59_DP0 = 0x64, + DCB_CONNECTOR_DMS59_DP1 = 0x65, DCB_CONNECTOR_NONE = 0xff }; -struct dcb_connector_table_entry { - uint8_t index; - uint32_t entry; - enum dcb_connector_type type; - uint8_t index2; - uint8_t gpio_tag; - void *drm; -}; - -struct dcb_connector_table { - int entries; - struct dcb_connector_table_entry entry[DCB_MAX_NUM_CONNECTOR_ENTRIES]; -}; - enum dcb_type { OUTPUT_ANALOG = 0, OUTPUT_TV = 1, @@ -111,6 +89,7 @@ enum dcb_type { OUTPUT_LVDS = 3, OUTPUT_DP = 6, OUTPUT_EOL = 14, /* DCB 4.0+, appears to be end-of-list */ + OUTPUT_UNUSED = 15, OUTPUT_ANY = -1 }; @@ -155,18 +134,8 @@ struct dcb_entry { struct dcb_table { uint8_t version; - int entries; struct dcb_entry entry[DCB_MAX_NUM_ENTRIES]; - - uint8_t *i2c_table; - uint8_t i2c_default_indices; - struct dcb_i2c_entry i2c[DCB_MAX_NUM_I2C_ENTRIES]; - - uint16_t gpio_table_ptr; - struct dcb_gpio_table gpio; - uint16_t connector_table_ptr; - struct dcb_connector_table connector; }; enum nouveau_or { @@ -195,7 +164,7 @@ enum pll_types { PLL_SHADER = 0x02, PLL_UNK03 = 0x03, PLL_MEMORY = 0x04, - PLL_UNK05 = 0x05, + PLL_VDEC = 0x05, PLL_UNK40 = 0x40, PLL_UNK41 = 0x41, PLL_UNK42 = 0x42, @@ -244,6 +213,8 @@ struct nvbios { NVBIOS_BIT } type; uint16_t offset; + uint32_t length; + uint8_t *data; uint8_t chip_version; @@ -254,8 +225,6 @@ struct nvbios { spinlock_t lock; - uint8_t data[NV_PROM_SIZE]; - unsigned int length; bool execute; uint8_t major_version; @@ -333,4 +302,11 @@ struct nvbios { } legacy; }; +void *dcb_table(struct drm_device *); +void *dcb_outp(struct drm_device *, u8 idx); +int dcb_outp_foreach(struct drm_device *, void *data, + int (*)(struct drm_device *, void *, int idx, u8 *outp)); +u8 *dcb_conntab(struct drm_device *); +u8 *dcb_conn(struct drm_device *, u8 idx); + #endif diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c index 2bb29c9..12ce044 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bo.c +++ b/drivers/gpu/drm/nouveau/nouveau_bo.c @@ -28,6 +28,7 @@ */ #include "drmP.h" +#include "ttm/ttm_page_alloc.h" #include "nouveau_drm.h" #include "nouveau_drv.h" @@ -92,6 +93,7 @@ nouveau_bo_new(struct drm_device *dev, int size, int align, { struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_bo *nvbo; + size_t acc_size; int ret; nvbo = kzalloc(sizeof(struct nouveau_bo), GFP_KERNEL); @@ -114,9 +116,12 @@ nouveau_bo_new(struct drm_device *dev, int size, int align, nvbo->bo.mem.num_pages = size >> PAGE_SHIFT; nouveau_bo_placement_set(nvbo, flags, 0); + acc_size = ttm_bo_dma_acc_size(&dev_priv->ttm.bdev, size, + sizeof(struct nouveau_bo)); + ret = ttm_bo_init(&dev_priv->ttm.bdev, &nvbo->bo, size, ttm_bo_type_device, &nvbo->placement, - align >> PAGE_SHIFT, 0, false, NULL, size, + align >> PAGE_SHIFT, 0, false, NULL, acc_size, nouveau_bo_del_ttm); if (ret) { /* ttm will call nouveau_bo_del_ttm if it fails.. */ @@ -343,8 +348,10 @@ nouveau_bo_wr32(struct nouveau_bo *nvbo, unsigned index, u32 val) *mem = val; } -static struct ttm_backend * -nouveau_bo_create_ttm_backend_entry(struct ttm_bo_device *bdev) +static struct ttm_tt * +nouveau_ttm_tt_create(struct ttm_bo_device *bdev, + unsigned long size, uint32_t page_flags, + struct page *dummy_read_page) { struct drm_nouveau_private *dev_priv = nouveau_bdev(bdev); struct drm_device *dev = dev_priv->dev; @@ -352,11 +359,13 @@ nouveau_bo_create_ttm_backend_entry(struct ttm_bo_device *bdev) switch (dev_priv->gart_info.type) { #if __OS_HAS_AGP case NOUVEAU_GART_AGP: - return ttm_agp_backend_init(bdev, dev->agp->bridge); + return ttm_agp_tt_create(bdev, dev->agp->bridge, + size, page_flags, dummy_read_page); #endif case NOUVEAU_GART_PDMA: case NOUVEAU_GART_HW: - return nouveau_sgdma_init_ttm(dev); + return nouveau_sgdma_create_ttm(bdev, size, page_flags, + dummy_read_page); default: NV_ERROR(dev, "Unknown GART type %d\n", dev_priv->gart_info.type); @@ -673,8 +682,7 @@ nouveau_vma_getmap(struct nouveau_channel *chan, struct nouveau_bo *nvbo, if (mem->mem_type == TTM_PL_VRAM) nouveau_vm_map(vma, node); else - nouveau_vm_map_sg(vma, 0, mem->num_pages << PAGE_SHIFT, - node, node->pages); + nouveau_vm_map_sg(vma, 0, mem->num_pages << PAGE_SHIFT, node); return 0; } @@ -685,16 +693,12 @@ nouveau_bo_move_m2mf(struct ttm_buffer_object *bo, int evict, bool intr, struct ttm_mem_reg *new_mem) { struct drm_nouveau_private *dev_priv = nouveau_bdev(bo->bdev); + struct nouveau_channel *chan = chan = dev_priv->channel; struct nouveau_bo *nvbo = nouveau_bo(bo); struct ttm_mem_reg *old_mem = &bo->mem; - struct nouveau_channel *chan; int ret; - chan = nvbo->channel; - if (!chan) { - chan = dev_priv->channel; - mutex_lock_nested(&chan->mutex, NOUVEAU_KCHANNEL_MUTEX); - } + mutex_lock_nested(&chan->mutex, NOUVEAU_KCHANNEL_MUTEX); /* create temporary vmas for the transfer and attach them to the * old nouveau_mem node, these will get cleaned up after ttm has @@ -726,8 +730,7 @@ nouveau_bo_move_m2mf(struct ttm_buffer_object *bo, int evict, bool intr, } out: - if (chan == dev_priv->channel) - mutex_unlock(&chan->mutex); + mutex_unlock(&chan->mutex); return ret; } @@ -801,19 +804,22 @@ out: static void nouveau_bo_move_ntfy(struct ttm_buffer_object *bo, struct ttm_mem_reg *new_mem) { - struct nouveau_mem *node = new_mem->mm_node; struct nouveau_bo *nvbo = nouveau_bo(bo); struct nouveau_vma *vma; + /* ttm can now (stupidly) pass the driver bos it didn't create... */ + if (bo->destroy != nouveau_bo_del_ttm) + return; + list_for_each_entry(vma, &nvbo->vma_list, head) { - if (new_mem->mem_type == TTM_PL_VRAM) { + if (new_mem && new_mem->mem_type == TTM_PL_VRAM) { nouveau_vm_map(vma, new_mem->mm_node); } else - if (new_mem->mem_type == TTM_PL_TT && + if (new_mem && new_mem->mem_type == TTM_PL_TT && nvbo->page_shift == vma->vm->spg_shift) { nouveau_vm_map_sg(vma, 0, new_mem-> num_pages << PAGE_SHIFT, - node, node->pages); + new_mem->mm_node); } else { nouveau_vm_unmap(vma); } @@ -940,7 +946,7 @@ nouveau_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem) if (dev_priv->gart_info.type == NOUVEAU_GART_AGP) { mem->bus.offset = mem->start << PAGE_SHIFT; mem->bus.base = dev_priv->gart_info.aper_base; - mem->bus.is_iomem = !dev->agp->cant_use_aperture; + mem->bus.is_iomem = true; } #endif break; @@ -1044,8 +1050,94 @@ nouveau_bo_fence(struct nouveau_bo *nvbo, struct nouveau_fence *fence) nouveau_fence_unref(&old_fence); } +static int +nouveau_ttm_tt_populate(struct ttm_tt *ttm) +{ + struct ttm_dma_tt *ttm_dma = (void *)ttm; + struct drm_nouveau_private *dev_priv; + struct drm_device *dev; + unsigned i; + int r; + + if (ttm->state != tt_unpopulated) + return 0; + + dev_priv = nouveau_bdev(ttm->bdev); + dev = dev_priv->dev; + +#if __OS_HAS_AGP + if (dev_priv->gart_info.type == NOUVEAU_GART_AGP) { + return ttm_agp_tt_populate(ttm); + } +#endif + +#ifdef CONFIG_SWIOTLB + if (swiotlb_nr_tbl()) { + return ttm_dma_populate((void *)ttm, dev->dev); + } +#endif + + r = ttm_pool_populate(ttm); + if (r) { + return r; + } + + for (i = 0; i < ttm->num_pages; i++) { + ttm_dma->dma_address[i] = pci_map_page(dev->pdev, ttm->pages[i], + 0, PAGE_SIZE, + PCI_DMA_BIDIRECTIONAL); + if (pci_dma_mapping_error(dev->pdev, ttm_dma->dma_address[i])) { + while (--i) { + pci_unmap_page(dev->pdev, ttm_dma->dma_address[i], + PAGE_SIZE, PCI_DMA_BIDIRECTIONAL); + ttm_dma->dma_address[i] = 0; + } + ttm_pool_unpopulate(ttm); + return -EFAULT; + } + } + return 0; +} + +static void +nouveau_ttm_tt_unpopulate(struct ttm_tt *ttm) +{ + struct ttm_dma_tt *ttm_dma = (void *)ttm; + struct drm_nouveau_private *dev_priv; + struct drm_device *dev; + unsigned i; + + dev_priv = nouveau_bdev(ttm->bdev); + dev = dev_priv->dev; + +#if __OS_HAS_AGP + if (dev_priv->gart_info.type == NOUVEAU_GART_AGP) { + ttm_agp_tt_unpopulate(ttm); + return; + } +#endif + +#ifdef CONFIG_SWIOTLB + if (swiotlb_nr_tbl()) { + ttm_dma_unpopulate((void *)ttm, dev->dev); + return; + } +#endif + + for (i = 0; i < ttm->num_pages; i++) { + if (ttm_dma->dma_address[i]) { + pci_unmap_page(dev->pdev, ttm_dma->dma_address[i], + PAGE_SIZE, PCI_DMA_BIDIRECTIONAL); + } + } + + ttm_pool_unpopulate(ttm); +} + struct ttm_bo_driver nouveau_bo_driver = { - .create_ttm_backend_entry = nouveau_bo_create_ttm_backend_entry, + .ttm_tt_create = &nouveau_ttm_tt_create, + .ttm_tt_populate = &nouveau_ttm_tt_populate, + .ttm_tt_unpopulate = &nouveau_ttm_tt_unpopulate, .invalidate_caches = nouveau_bo_invalidate_caches, .init_mem_type = nouveau_bo_init_mem_type, .evict_flags = nouveau_bo_evict_flags, @@ -1091,7 +1183,7 @@ nouveau_bo_vma_add(struct nouveau_bo *nvbo, struct nouveau_vm *vm, nouveau_vm_map(vma, nvbo->bo.mem.mm_node); else if (nvbo->bo.mem.mem_type == TTM_PL_TT) - nouveau_vm_map_sg(vma, 0, size, node, node->pages); + nouveau_vm_map_sg(vma, 0, size, node); list_add_tail(&vma->head, &nvbo->vma_list); vma->refcount = 1; diff --git a/drivers/gpu/drm/nouveau/nouveau_channel.c b/drivers/gpu/drm/nouveau/nouveau_channel.c index bb6ec9e..846afb0 100644 --- a/drivers/gpu/drm/nouveau/nouveau_channel.c +++ b/drivers/gpu/drm/nouveau/nouveau_channel.c @@ -122,7 +122,7 @@ nouveau_channel_alloc(struct drm_device *dev, struct nouveau_channel **chan_ret, struct nouveau_fpriv *fpriv = nouveau_fpriv(file_priv); struct nouveau_channel *chan; unsigned long flags; - int ret; + int ret, i; /* allocate and lock channel structure */ chan = kzalloc(sizeof(*chan), GFP_KERNEL); @@ -184,9 +184,11 @@ nouveau_channel_alloc(struct drm_device *dev, struct nouveau_channel **chan_ret, return ret; } - nouveau_dma_pre_init(chan); + nouveau_dma_init(chan); chan->user_put = 0x40; chan->user_get = 0x44; + if (dev_priv->card_type >= NV_50) + chan->user_get_hi = 0x60; /* disable the fifo caches */ pfifo->reassign(dev, false); @@ -200,9 +202,18 @@ nouveau_channel_alloc(struct drm_device *dev, struct nouveau_channel **chan_ret, pfifo->reassign(dev, true); - ret = nouveau_dma_init(chan); - if (!ret) - ret = nouveau_fence_channel_init(chan); + /* Insert NOPs for NOUVEAU_DMA_SKIPS */ + ret = RING_SPACE(chan, NOUVEAU_DMA_SKIPS); + if (ret) { + nouveau_channel_put(&chan); + return ret; + } + + for (i = 0; i < NOUVEAU_DMA_SKIPS; i++) + OUT_RING (chan, 0x00000000); + FIRE_RING(chan); + + ret = nouveau_fence_channel_init(chan); if (ret) { nouveau_channel_put(&chan); return ret; @@ -425,18 +436,11 @@ nouveau_ioctl_fifo_alloc(struct drm_device *dev, void *data, } if (dev_priv->card_type < NV_C0) { - init->subchan[0].handle = NvM2MF; - if (dev_priv->card_type < NV_50) - init->subchan[0].grclass = 0x0039; - else - init->subchan[0].grclass = 0x5039; + init->subchan[0].handle = 0x00000000; + init->subchan[0].grclass = 0x0000; init->subchan[1].handle = NvSw; init->subchan[1].grclass = NV_SW; init->nr_subchan = 2; - } else { - init->subchan[0].handle = 0x9039; - init->subchan[0].grclass = 0x9039; - init->nr_subchan = 1; } /* Named memory object area */ diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.c b/drivers/gpu/drm/nouveau/nouveau_connector.c index 1e72db5..7b11edb 100644 --- a/drivers/gpu/drm/nouveau/nouveau_connector.c +++ b/drivers/gpu/drm/nouveau/nouveau_connector.c @@ -35,6 +35,7 @@ #include "nouveau_encoder.h" #include "nouveau_crtc.h" #include "nouveau_connector.h" +#include "nouveau_gpio.h" #include "nouveau_hw.h" static void nouveau_connector_hotplug(void *, int); @@ -78,29 +79,11 @@ nouveau_encoder_connector_get(struct nouveau_encoder *encoder) return NULL; } -/*TODO: This could use improvement, and learn to handle the fixed - * BIOS tables etc. It's fine currently, for its only user. - */ -int -nouveau_connector_bpp(struct drm_connector *connector) -{ - struct nouveau_connector *nv_connector = nouveau_connector(connector); - - if (nv_connector->edid && nv_connector->edid->revision >= 4) { - u8 bpc = ((nv_connector->edid->input & 0x70) >> 3) + 4; - if (bpc > 4) - return bpc; - } - - return 18; -} - static void nouveau_connector_destroy(struct drm_connector *connector) { struct nouveau_connector *nv_connector = nouveau_connector(connector); struct drm_nouveau_private *dev_priv; - struct nouveau_gpio_engine *pgpio; struct drm_device *dev; if (!nv_connector) @@ -110,10 +93,9 @@ nouveau_connector_destroy(struct drm_connector *connector) dev_priv = dev->dev_private; NV_DEBUG_KMS(dev, "\n"); - pgpio = &dev_priv->engine.gpio; - if (pgpio->irq_unregister) { - pgpio->irq_unregister(dev, nv_connector->dcb->gpio_tag, - nouveau_connector_hotplug, connector); + if (nv_connector->hpd != DCB_GPIO_UNUSED) { + nouveau_gpio_isr_del(dev, 0, nv_connector->hpd, 0xff, + nouveau_connector_hotplug, connector); } kfree(nv_connector->edid); @@ -198,6 +180,10 @@ nouveau_connector_set_encoder(struct drm_connector *connector, return; nv_connector->detected_encoder = nv_encoder; + if (dev_priv->card_type >= NV_50) { + connector->interlace_allowed = true; + connector->doublescan_allowed = true; + } else if (nv_encoder->dcb->type == OUTPUT_LVDS || nv_encoder->dcb->type == OUTPUT_TMDS) { connector->doublescan_allowed = false; @@ -214,7 +200,7 @@ nouveau_connector_set_encoder(struct drm_connector *connector, connector->interlace_allowed = true; } - if (nv_connector->dcb->type == DCB_CONNECTOR_DVI_I) { + if (nv_connector->type == DCB_CONNECTOR_DVI_I) { drm_connector_property_set_value(connector, dev->mode_config.dvi_i_subconnector_property, nv_encoder->dcb->type == OUTPUT_TMDS ? @@ -397,7 +383,7 @@ nouveau_connector_force(struct drm_connector *connector) struct nouveau_encoder *nv_encoder; int type; - if (nv_connector->dcb->type == DCB_CONNECTOR_DVI_I) { + if (nv_connector->type == DCB_CONNECTOR_DVI_I) { if (connector->force == DRM_FORCE_ON_DIGITAL) type = OUTPUT_TMDS; else @@ -420,15 +406,21 @@ static int nouveau_connector_set_property(struct drm_connector *connector, struct drm_property *property, uint64_t value) { + struct drm_nouveau_private *dev_priv = connector->dev->dev_private; + struct nouveau_display_engine *disp = &dev_priv->engine.display; struct nouveau_connector *nv_connector = nouveau_connector(connector); struct nouveau_encoder *nv_encoder = nv_connector->detected_encoder; struct drm_encoder *encoder = to_drm_encoder(nv_encoder); struct drm_device *dev = connector->dev; + struct nouveau_crtc *nv_crtc; int ret; + nv_crtc = NULL; + if (connector->encoder && connector->encoder->crtc) + nv_crtc = nouveau_crtc(connector->encoder->crtc); + /* Scaling mode */ if (property == dev->mode_config.scaling_mode_property) { - struct nouveau_crtc *nv_crtc = NULL; bool modeset = false; switch (value) { @@ -454,8 +446,6 @@ nouveau_connector_set_property(struct drm_connector *connector, modeset = true; nv_connector->scaling_mode = value; - if (connector->encoder && connector->encoder->crtc) - nv_crtc = nouveau_crtc(connector->encoder->crtc); if (!nv_crtc) return 0; @@ -467,7 +457,7 @@ nouveau_connector_set_property(struct drm_connector *connector, if (!ret) return -EINVAL; } else { - ret = nv_crtc->set_scale(nv_crtc, value, true); + ret = nv_crtc->set_scale(nv_crtc, true); if (ret) return ret; } @@ -475,23 +465,71 @@ nouveau_connector_set_property(struct drm_connector *connector, return 0; } - /* Dithering */ - if (property == dev->mode_config.dithering_mode_property) { - struct nouveau_crtc *nv_crtc = NULL; + /* Underscan */ + if (property == disp->underscan_property) { + if (nv_connector->underscan != value) { + nv_connector->underscan = value; + if (!nv_crtc || !nv_crtc->set_scale) + return 0; - if (value == DRM_MODE_DITHERING_ON) - nv_connector->use_dithering = true; - else - nv_connector->use_dithering = false; + return nv_crtc->set_scale(nv_crtc, true); + } + + return 0; + } + + if (property == disp->underscan_hborder_property) { + if (nv_connector->underscan_hborder != value) { + nv_connector->underscan_hborder = value; + if (!nv_crtc || !nv_crtc->set_scale) + return 0; + + return nv_crtc->set_scale(nv_crtc, true); + } + + return 0; + } + + if (property == disp->underscan_vborder_property) { + if (nv_connector->underscan_vborder != value) { + nv_connector->underscan_vborder = value; + if (!nv_crtc || !nv_crtc->set_scale) + return 0; + + return nv_crtc->set_scale(nv_crtc, true); + } + + return 0; + } - if (connector->encoder && connector->encoder->crtc) - nv_crtc = nouveau_crtc(connector->encoder->crtc); + /* Dithering */ + if (property == disp->dithering_mode) { + nv_connector->dithering_mode = value; + if (!nv_crtc || !nv_crtc->set_dither) + return 0; + + return nv_crtc->set_dither(nv_crtc, true); + } + if (property == disp->dithering_depth) { + nv_connector->dithering_depth = value; if (!nv_crtc || !nv_crtc->set_dither) return 0; - return nv_crtc->set_dither(nv_crtc, nv_connector->use_dithering, - true); + return nv_crtc->set_dither(nv_crtc, true); + } + + if (nv_crtc && nv_crtc->set_color_vibrance) { + /* Hue */ + if (property == disp->vibrant_hue_property) { + nv_crtc->vibrant_hue = value - 90; + return nv_crtc->set_color_vibrance(nv_crtc, true); + } + /* Saturation */ + if (property == disp->color_vibrance_property) { + nv_crtc->color_vibrance = value - 100; + return nv_crtc->set_color_vibrance(nv_crtc, true); + } } if (nv_encoder && nv_encoder->dcb->type == OUTPUT_TV) @@ -617,7 +655,7 @@ nouveau_connector_detect_depth(struct drm_connector *connector) return; /* EDID 1.4 is *supposed* to be supported on eDP, but, Apple... */ - if (nv_connector->dcb->type == DCB_CONNECTOR_eDP) { + if (nv_connector->type == DCB_CONNECTOR_eDP) { connector->display_info.bpc = 6; return; } @@ -641,7 +679,7 @@ nouveau_connector_detect_depth(struct drm_connector *connector) * know which if_is_24bit flag to check... */ if (nv_connector->edid && - nv_connector->dcb->type == DCB_CONNECTOR_LVDS_SPWG) + nv_connector->type == DCB_CONNECTOR_LVDS_SPWG) duallink = ((u8 *)nv_connector->edid)[121] == 2; else duallink = mode->clock >= bios->fp.duallink_transition_clk; @@ -680,6 +718,12 @@ nouveau_connector_get_modes(struct drm_connector *connector) nv_connector->native_mode = drm_mode_duplicate(dev, &mode); } + /* Determine display colour depth for everything except LVDS now, + * DP requires this before mode_valid() is called. + */ + if (connector->connector_type != DRM_MODE_CONNECTOR_LVDS) + nouveau_connector_detect_depth(connector); + /* Find the native mode if this is a digital panel, if we didn't * find any modes through DDC previously add the native mode to * the list of modes. @@ -695,18 +739,19 @@ nouveau_connector_get_modes(struct drm_connector *connector) ret = 1; } - /* Attempt to determine display colour depth, this has to happen after - * we've determined the "native" mode for LVDS, as the VBIOS tables - * require us to compare against a pixel clock in some cases.. + /* Determine LVDS colour depth, must happen after determining + * "native" mode as some VBIOS tables require us to use the + * pixel clock as part of the lookup... */ - nouveau_connector_detect_depth(connector); + if (connector->connector_type == DRM_MODE_CONNECTOR_LVDS) + nouveau_connector_detect_depth(connector); if (nv_encoder->dcb->type == OUTPUT_TV) ret = get_slave_funcs(encoder)->get_modes(encoder, connector); - if (nv_connector->dcb->type == DCB_CONNECTOR_LVDS || - nv_connector->dcb->type == DCB_CONNECTOR_LVDS_SPWG || - nv_connector->dcb->type == DCB_CONNECTOR_eDP) + if (nv_connector->type == DCB_CONNECTOR_LVDS || + nv_connector->type == DCB_CONNECTOR_LVDS_SPWG || + nv_connector->type == DCB_CONNECTOR_eDP) ret += nouveau_connector_scaler_modes_add(connector); return ret; @@ -765,7 +810,7 @@ nouveau_connector_mode_valid(struct drm_connector *connector, case OUTPUT_DP: max_clock = nv_encoder->dp.link_nr; max_clock *= nv_encoder->dp.link_bw; - clock = clock * nouveau_connector_bpp(connector) / 10; + clock = clock * (connector->display_info.bpc * 3) / 10; break; default: BUG_ON(1); @@ -823,96 +868,188 @@ nouveau_connector_funcs_lvds = { .force = nouveau_connector_force }; +static int +drm_conntype_from_dcb(enum dcb_connector_type dcb) +{ + switch (dcb) { + case DCB_CONNECTOR_VGA : return DRM_MODE_CONNECTOR_VGA; + case DCB_CONNECTOR_TV_0 : + case DCB_CONNECTOR_TV_1 : + case DCB_CONNECTOR_TV_3 : return DRM_MODE_CONNECTOR_TV; + case DCB_CONNECTOR_DMS59_0 : + case DCB_CONNECTOR_DMS59_1 : + case DCB_CONNECTOR_DVI_I : return DRM_MODE_CONNECTOR_DVII; + case DCB_CONNECTOR_DVI_D : return DRM_MODE_CONNECTOR_DVID; + case DCB_CONNECTOR_LVDS : + case DCB_CONNECTOR_LVDS_SPWG: return DRM_MODE_CONNECTOR_LVDS; + case DCB_CONNECTOR_DMS59_DP0: + case DCB_CONNECTOR_DMS59_DP1: + case DCB_CONNECTOR_DP : return DRM_MODE_CONNECTOR_DisplayPort; + case DCB_CONNECTOR_eDP : return DRM_MODE_CONNECTOR_eDP; + case DCB_CONNECTOR_HDMI_0 : + case DCB_CONNECTOR_HDMI_1 : return DRM_MODE_CONNECTOR_HDMIA; + default: + break; + } + + return DRM_MODE_CONNECTOR_Unknown; +} + struct drm_connector * nouveau_connector_create(struct drm_device *dev, int index) { const struct drm_connector_funcs *funcs = &nouveau_connector_funcs; struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; + struct nouveau_display_engine *disp = &dev_priv->engine.display; struct nouveau_connector *nv_connector = NULL; - struct dcb_connector_table_entry *dcb = NULL; struct drm_connector *connector; int type, ret = 0; + bool dummy; NV_DEBUG_KMS(dev, "\n"); - if (index >= dev_priv->vbios.dcb.connector.entries) - return ERR_PTR(-EINVAL); - - dcb = &dev_priv->vbios.dcb.connector.entry[index]; - if (dcb->drm) - return dcb->drm; - - switch (dcb->type) { - case DCB_CONNECTOR_VGA: - type = DRM_MODE_CONNECTOR_VGA; - break; - case DCB_CONNECTOR_TV_0: - case DCB_CONNECTOR_TV_1: - case DCB_CONNECTOR_TV_3: - type = DRM_MODE_CONNECTOR_TV; - break; - case DCB_CONNECTOR_DVI_I: - type = DRM_MODE_CONNECTOR_DVII; - break; - case DCB_CONNECTOR_DVI_D: - type = DRM_MODE_CONNECTOR_DVID; - break; - case DCB_CONNECTOR_HDMI_0: - case DCB_CONNECTOR_HDMI_1: - type = DRM_MODE_CONNECTOR_HDMIA; - break; - case DCB_CONNECTOR_LVDS: - case DCB_CONNECTOR_LVDS_SPWG: - type = DRM_MODE_CONNECTOR_LVDS; - funcs = &nouveau_connector_funcs_lvds; - break; - case DCB_CONNECTOR_DP: - type = DRM_MODE_CONNECTOR_DisplayPort; - break; - case DCB_CONNECTOR_eDP: - type = DRM_MODE_CONNECTOR_eDP; - break; - default: - NV_ERROR(dev, "unknown connector type: 0x%02x!!\n", dcb->type); - return ERR_PTR(-EINVAL); + list_for_each_entry(connector, &dev->mode_config.connector_list, head) { + nv_connector = nouveau_connector(connector); + if (nv_connector->index == index) + return connector; } nv_connector = kzalloc(sizeof(*nv_connector), GFP_KERNEL); if (!nv_connector) return ERR_PTR(-ENOMEM); - nv_connector->dcb = dcb; + connector = &nv_connector->base; + nv_connector->index = index; + + /* attempt to parse vbios connector type and hotplug gpio */ + nv_connector->dcb = dcb_conn(dev, index); + if (nv_connector->dcb) { + static const u8 hpd[16] = { + 0xff, 0x07, 0x08, 0xff, 0xff, 0x51, 0x52, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0x5e, 0x5f, 0x60, + }; + + u32 entry = ROM16(nv_connector->dcb[0]); + if (dcb_conntab(dev)[3] >= 4) + entry |= (u32)ROM16(nv_connector->dcb[2]) << 16; + + nv_connector->hpd = ffs((entry & 0x07033000) >> 12); + nv_connector->hpd = hpd[nv_connector->hpd]; + + nv_connector->type = nv_connector->dcb[0]; + if (drm_conntype_from_dcb(nv_connector->type) == + DRM_MODE_CONNECTOR_Unknown) { + NV_WARN(dev, "unknown connector type %02x\n", + nv_connector->type); + nv_connector->type = DCB_CONNECTOR_NONE; + } - /* defaults, will get overridden in detect() */ - connector->interlace_allowed = false; - connector->doublescan_allowed = false; + /* Gigabyte NX85T */ + if (nv_match_device(dev, 0x0421, 0x1458, 0x344c)) { + if (nv_connector->type == DCB_CONNECTOR_HDMI_1) + nv_connector->type = DCB_CONNECTOR_DVI_I; + } - drm_connector_init(dev, connector, funcs, type); - drm_connector_helper_add(connector, &nouveau_connector_helper_funcs); + /* Gigabyte GV-NX86T512H */ + if (nv_match_device(dev, 0x0402, 0x1458, 0x3455)) { + if (nv_connector->type == DCB_CONNECTOR_HDMI_1) + nv_connector->type = DCB_CONNECTOR_DVI_I; + } + } else { + nv_connector->type = DCB_CONNECTOR_NONE; + nv_connector->hpd = DCB_GPIO_UNUSED; + } + + /* no vbios data, or an unknown dcb connector type - attempt to + * figure out something suitable ourselves + */ + if (nv_connector->type == DCB_CONNECTOR_NONE) { + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct dcb_table *dcbt = &dev_priv->vbios.dcb; + u32 encoders = 0; + int i; + + for (i = 0; i < dcbt->entries; i++) { + if (dcbt->entry[i].connector == nv_connector->index) + encoders |= (1 << dcbt->entry[i].type); + } - /* Check if we need dithering enabled */ - if (connector->connector_type == DRM_MODE_CONNECTOR_LVDS) { - bool dummy, is_24bit = false; + if (encoders & (1 << OUTPUT_DP)) { + if (encoders & (1 << OUTPUT_TMDS)) + nv_connector->type = DCB_CONNECTOR_DP; + else + nv_connector->type = DCB_CONNECTOR_eDP; + } else + if (encoders & (1 << OUTPUT_TMDS)) { + if (encoders & (1 << OUTPUT_ANALOG)) + nv_connector->type = DCB_CONNECTOR_DVI_I; + else + nv_connector->type = DCB_CONNECTOR_DVI_D; + } else + if (encoders & (1 << OUTPUT_ANALOG)) { + nv_connector->type = DCB_CONNECTOR_VGA; + } else + if (encoders & (1 << OUTPUT_LVDS)) { + nv_connector->type = DCB_CONNECTOR_LVDS; + } else + if (encoders & (1 << OUTPUT_TV)) { + nv_connector->type = DCB_CONNECTOR_TV_0; + } + } - ret = nouveau_bios_parse_lvds_table(dev, 0, &dummy, &is_24bit); + type = drm_conntype_from_dcb(nv_connector->type); + if (type == DRM_MODE_CONNECTOR_LVDS) { + ret = nouveau_bios_parse_lvds_table(dev, 0, &dummy, &dummy); if (ret) { - NV_ERROR(dev, "Error parsing LVDS table, disabling " - "LVDS\n"); - goto fail; + NV_ERROR(dev, "Error parsing LVDS table, disabling\n"); + kfree(nv_connector); + return ERR_PTR(ret); } - nv_connector->use_dithering = !is_24bit; + funcs = &nouveau_connector_funcs_lvds; + } else { + funcs = &nouveau_connector_funcs; } + /* defaults, will get overridden in detect() */ + connector->interlace_allowed = false; + connector->doublescan_allowed = false; + + drm_connector_init(dev, connector, funcs, type); + drm_connector_helper_add(connector, &nouveau_connector_helper_funcs); + /* Init DVI-I specific properties */ - if (dcb->type == DCB_CONNECTOR_DVI_I) { - drm_mode_create_dvi_i_properties(dev); + if (nv_connector->type == DCB_CONNECTOR_DVI_I) drm_connector_attach_property(connector, dev->mode_config.dvi_i_subconnector_property, 0); - drm_connector_attach_property(connector, dev->mode_config.dvi_i_select_subconnector_property, 0); + + /* Add overscan compensation options to digital outputs */ + if (disp->underscan_property && + (type == DRM_MODE_CONNECTOR_DVID || + type == DRM_MODE_CONNECTOR_DVII || + type == DRM_MODE_CONNECTOR_HDMIA || + type == DRM_MODE_CONNECTOR_DisplayPort)) { + drm_connector_attach_property(connector, + disp->underscan_property, + UNDERSCAN_OFF); + drm_connector_attach_property(connector, + disp->underscan_hborder_property, + 0); + drm_connector_attach_property(connector, + disp->underscan_vborder_property, + 0); } - switch (dcb->type) { + /* Add hue and saturation options */ + if (disp->vibrant_hue_property) + drm_connector_attach_property(connector, + disp->vibrant_hue_property, + 90); + if (disp->color_vibrance_property) + drm_connector_attach_property(connector, + disp->color_vibrance_property, + 150); + + switch (nv_connector->type) { case DCB_CONNECTOR_VGA: if (dev_priv->card_type >= NV_50) { drm_connector_attach_property(connector, @@ -931,32 +1068,32 @@ nouveau_connector_create(struct drm_device *dev, int index) drm_connector_attach_property(connector, dev->mode_config.scaling_mode_property, nv_connector->scaling_mode); - drm_connector_attach_property(connector, - dev->mode_config.dithering_mode_property, - nv_connector->use_dithering ? - DRM_MODE_DITHERING_ON : DRM_MODE_DITHERING_OFF); + if (disp->dithering_mode) { + nv_connector->dithering_mode = DITHERING_MODE_AUTO; + drm_connector_attach_property(connector, + disp->dithering_mode, + nv_connector->dithering_mode); + } + if (disp->dithering_depth) { + nv_connector->dithering_depth = DITHERING_DEPTH_AUTO; + drm_connector_attach_property(connector, + disp->dithering_depth, + nv_connector->dithering_depth); + } break; } - if (nv_connector->dcb->gpio_tag != 0xff && pgpio->irq_register) { - pgpio->irq_register(dev, nv_connector->dcb->gpio_tag, - nouveau_connector_hotplug, connector); - - connector->polled = DRM_CONNECTOR_POLL_HPD; - } else { - connector->polled = DRM_CONNECTOR_POLL_CONNECT; + connector->polled = DRM_CONNECTOR_POLL_CONNECT; + if (nv_connector->hpd != DCB_GPIO_UNUSED) { + ret = nouveau_gpio_isr_add(dev, 0, nv_connector->hpd, 0xff, + nouveau_connector_hotplug, + connector); + if (ret == 0) + connector->polled = DRM_CONNECTOR_POLL_HPD; } drm_sysfs_connector_add(connector); - - dcb->drm = connector; - return dcb->drm; - -fail: - drm_connector_cleanup(connector); - kfree(connector); - return ERR_PTR(ret); - + return connector; } static void diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.h b/drivers/gpu/drm/nouveau/nouveau_connector.h index 711b1e9..e485702 100644 --- a/drivers/gpu/drm/nouveau/nouveau_connector.h +++ b/drivers/gpu/drm/nouveau/nouveau_connector.h @@ -30,13 +30,43 @@ #include "drm_edid.h" #include "nouveau_i2c.h" +enum nouveau_underscan_type { + UNDERSCAN_OFF, + UNDERSCAN_ON, + UNDERSCAN_AUTO, +}; + +/* the enum values specifically defined here match nv50/nvd0 hw values, and + * the code relies on this + */ +enum nouveau_dithering_mode { + DITHERING_MODE_OFF = 0x00, + DITHERING_MODE_ON = 0x01, + DITHERING_MODE_DYNAMIC2X2 = 0x10 | DITHERING_MODE_ON, + DITHERING_MODE_STATIC2X2 = 0x18 | DITHERING_MODE_ON, + DITHERING_MODE_TEMPORAL = 0x20 | DITHERING_MODE_ON, + DITHERING_MODE_AUTO +}; + +enum nouveau_dithering_depth { + DITHERING_DEPTH_6BPC = 0x00, + DITHERING_DEPTH_8BPC = 0x02, + DITHERING_DEPTH_AUTO +}; + struct nouveau_connector { struct drm_connector base; + enum dcb_connector_type type; + u8 index; + u8 *dcb; + u8 hpd; - struct dcb_connector_table_entry *dcb; - + int dithering_mode; + int dithering_depth; int scaling_mode; - bool use_dithering; + enum nouveau_underscan_type underscan; + u32 underscan_hborder; + u32 underscan_vborder; struct nouveau_encoder *detected_encoder; struct edid *edid; diff --git a/drivers/gpu/drm/nouveau/nouveau_crtc.h b/drivers/gpu/drm/nouveau/nouveau_crtc.h index bf8e128..e6d0d1e 100644 --- a/drivers/gpu/drm/nouveau/nouveau_crtc.h +++ b/drivers/gpu/drm/nouveau/nouveau_crtc.h @@ -32,11 +32,11 @@ struct nouveau_crtc { int index; - struct drm_display_mode *mode; - uint32_t dpms_saved_fp_control; uint32_t fp_users; int saturation; + int color_vibrance; + int vibrant_hue; int sharpness; int last_dpms; @@ -67,8 +67,9 @@ struct nouveau_crtc { int depth; } lut; - int (*set_dither)(struct nouveau_crtc *crtc, bool on, bool update); - int (*set_scale)(struct nouveau_crtc *crtc, int mode, bool update); + int (*set_dither)(struct nouveau_crtc *crtc, bool update); + int (*set_scale)(struct nouveau_crtc *crtc, bool update); + int (*set_color_vibrance)(struct nouveau_crtc *crtc, bool update); }; static inline struct nouveau_crtc *nouveau_crtc(struct drm_crtc *crtc) diff --git a/drivers/gpu/drm/nouveau/nouveau_debugfs.c b/drivers/gpu/drm/nouveau/nouveau_debugfs.c index 8e15923..fa2ec49 100644 --- a/drivers/gpu/drm/nouveau/nouveau_debugfs.c +++ b/drivers/gpu/drm/nouveau/nouveau_debugfs.c @@ -44,7 +44,7 @@ nouveau_debugfs_channel_info(struct seq_file *m, void *data) seq_printf(m, "channel id : %d\n", chan->id); seq_printf(m, "cpu fifo state:\n"); - seq_printf(m, " base: 0x%08x\n", chan->pushbuf_base); + seq_printf(m, " base: 0x%10llx\n", chan->pushbuf_base); seq_printf(m, " max: 0x%08x\n", chan->dma.max << 2); seq_printf(m, " cur: 0x%08x\n", chan->dma.cur << 2); seq_printf(m, " put: 0x%08x\n", chan->dma.put << 2); @@ -178,6 +178,7 @@ static struct drm_info_list nouveau_debugfs_list[] = { { "memory", nouveau_debugfs_memory_info, 0, NULL }, { "vbios.rom", nouveau_debugfs_vbios_image, 0, NULL }, { "ttm_page_pool", ttm_page_alloc_debugfs, 0, NULL }, + { "ttm_dma_page_pool", ttm_dma_page_alloc_debugfs, 0, NULL }, }; #define NOUVEAU_DEBUGFS_ENTRIES ARRAY_SIZE(nouveau_debugfs_list) diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c index 6adef06..f233b8f 100644 --- a/drivers/gpu/drm/nouveau/nouveau_display.c +++ b/drivers/gpu/drm/nouveau/nouveau_display.c @@ -32,6 +32,8 @@ #include "nouveau_hw.h" #include "nouveau_crtc.h" #include "nouveau_dma.h" +#include "nouveau_connector.h" +#include "nouveau_gpio.h" #include "nv50_display.h" static void @@ -64,7 +66,7 @@ static const struct drm_framebuffer_funcs nouveau_framebuffer_funcs = { int nouveau_framebuffer_init(struct drm_device *dev, struct nouveau_framebuffer *nv_fb, - struct drm_mode_fb_cmd *mode_cmd, + struct drm_mode_fb_cmd2 *mode_cmd, struct nouveau_bo *nvbo) { struct drm_nouveau_private *dev_priv = dev->dev_private; @@ -107,14 +109,14 @@ nouveau_framebuffer_init(struct drm_device *dev, if (!tile_flags) { if (dev_priv->card_type < NV_D0) - nv_fb->r_pitch = 0x00100000 | fb->pitch; + nv_fb->r_pitch = 0x00100000 | fb->pitches[0]; else - nv_fb->r_pitch = 0x01000000 | fb->pitch; + nv_fb->r_pitch = 0x01000000 | fb->pitches[0]; } else { u32 mode = nvbo->tile_mode; if (dev_priv->card_type >= NV_C0) mode >>= 4; - nv_fb->r_pitch = ((fb->pitch / 4) << 4) | mode; + nv_fb->r_pitch = ((fb->pitches[0] / 4) << 4) | mode; } } @@ -124,13 +126,13 @@ nouveau_framebuffer_init(struct drm_device *dev, static struct drm_framebuffer * nouveau_user_framebuffer_create(struct drm_device *dev, struct drm_file *file_priv, - struct drm_mode_fb_cmd *mode_cmd) + struct drm_mode_fb_cmd2 *mode_cmd) { struct nouveau_framebuffer *nouveau_fb; struct drm_gem_object *gem; int ret; - gem = drm_gem_object_lookup(dev, file_priv, mode_cmd->handle); + gem = drm_gem_object_lookup(dev, file_priv, mode_cmd->handles[0]); if (!gem) return ERR_PTR(-ENOENT); @@ -147,11 +149,207 @@ nouveau_user_framebuffer_create(struct drm_device *dev, return &nouveau_fb->base; } -const struct drm_mode_config_funcs nouveau_mode_config_funcs = { +static const struct drm_mode_config_funcs nouveau_mode_config_funcs = { .fb_create = nouveau_user_framebuffer_create, .output_poll_changed = nouveau_fbcon_output_poll_changed, }; + +struct nouveau_drm_prop_enum_list { + u8 gen_mask; + int type; + char *name; +}; + +static struct nouveau_drm_prop_enum_list underscan[] = { + { 6, UNDERSCAN_AUTO, "auto" }, + { 6, UNDERSCAN_OFF, "off" }, + { 6, UNDERSCAN_ON, "on" }, + {} +}; + +static struct nouveau_drm_prop_enum_list dither_mode[] = { + { 7, DITHERING_MODE_AUTO, "auto" }, + { 7, DITHERING_MODE_OFF, "off" }, + { 1, DITHERING_MODE_ON, "on" }, + { 6, DITHERING_MODE_STATIC2X2, "static 2x2" }, + { 6, DITHERING_MODE_DYNAMIC2X2, "dynamic 2x2" }, + { 4, DITHERING_MODE_TEMPORAL, "temporal" }, + {} +}; + +static struct nouveau_drm_prop_enum_list dither_depth[] = { + { 6, DITHERING_DEPTH_AUTO, "auto" }, + { 6, DITHERING_DEPTH_6BPC, "6 bpc" }, + { 6, DITHERING_DEPTH_8BPC, "8 bpc" }, + {} +}; + +#define PROP_ENUM(p,gen,n,list) do { \ + struct nouveau_drm_prop_enum_list *l = (list); \ + int c = 0; \ + while (l->gen_mask) { \ + if (l->gen_mask & (1 << (gen))) \ + c++; \ + l++; \ + } \ + if (c) { \ + p = drm_property_create(dev, DRM_MODE_PROP_ENUM, n, c); \ + l = (list); \ + c = 0; \ + while (p && l->gen_mask) { \ + if (l->gen_mask & (1 << (gen))) { \ + drm_property_add_enum(p, c, l->type, l->name); \ + c++; \ + } \ + l++; \ + } \ + } \ +} while(0) + +int +nouveau_display_init(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_display_engine *disp = &dev_priv->engine.display; + struct drm_connector *connector; + int ret; + + ret = disp->init(dev); + if (ret) + return ret; + + /* power on internal panel if it's not already. the init tables of + * some vbios default this to off for some reason, causing the + * panel to not work after resume + */ + if (nouveau_gpio_func_get(dev, DCB_GPIO_PANEL_POWER) == 0) { + nouveau_gpio_func_set(dev, DCB_GPIO_PANEL_POWER, true); + msleep(300); + } + + /* enable polling for external displays */ + drm_kms_helper_poll_enable(dev); + + /* enable hotplug interrupts */ + list_for_each_entry(connector, &dev->mode_config.connector_list, head) { + struct nouveau_connector *conn = nouveau_connector(connector); + nouveau_gpio_irq(dev, 0, conn->hpd, 0xff, true); + } + + return ret; +} + +void +nouveau_display_fini(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_display_engine *disp = &dev_priv->engine.display; + struct drm_connector *connector; + + /* disable hotplug interrupts */ + list_for_each_entry(connector, &dev->mode_config.connector_list, head) { + struct nouveau_connector *conn = nouveau_connector(connector); + nouveau_gpio_irq(dev, 0, conn->hpd, 0xff, false); + } + + drm_kms_helper_poll_disable(dev); + disp->fini(dev); +} + +int +nouveau_display_create(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_display_engine *disp = &dev_priv->engine.display; + int ret, gen; + + drm_mode_config_init(dev); + drm_mode_create_scaling_mode_property(dev); + drm_mode_create_dvi_i_properties(dev); + + if (dev_priv->card_type < NV_50) + gen = 0; + else + if (dev_priv->card_type < NV_D0) + gen = 1; + else + gen = 2; + + PROP_ENUM(disp->dithering_mode, gen, "dithering mode", dither_mode); + PROP_ENUM(disp->dithering_depth, gen, "dithering depth", dither_depth); + PROP_ENUM(disp->underscan_property, gen, "underscan", underscan); + + disp->underscan_hborder_property = + drm_property_create_range(dev, 0, "underscan hborder", 0, 128); + + disp->underscan_vborder_property = + drm_property_create_range(dev, 0, "underscan vborder", 0, 128); + + if (gen == 1) { + disp->vibrant_hue_property = + drm_property_create(dev, DRM_MODE_PROP_RANGE, + "vibrant hue", 2); + disp->vibrant_hue_property->values[0] = 0; + disp->vibrant_hue_property->values[1] = 180; /* -90..+90 */ + + disp->color_vibrance_property = + drm_property_create(dev, DRM_MODE_PROP_RANGE, + "color vibrance", 2); + disp->color_vibrance_property->values[0] = 0; + disp->color_vibrance_property->values[1] = 200; /* -100..+100 */ + } + + dev->mode_config.funcs = (void *)&nouveau_mode_config_funcs; + dev->mode_config.fb_base = pci_resource_start(dev->pdev, 1); + + dev->mode_config.min_width = 0; + dev->mode_config.min_height = 0; + if (dev_priv->card_type < NV_10) { + dev->mode_config.max_width = 2048; + dev->mode_config.max_height = 2048; + } else + if (dev_priv->card_type < NV_50) { + dev->mode_config.max_width = 4096; + dev->mode_config.max_height = 4096; + } else { + dev->mode_config.max_width = 8192; + dev->mode_config.max_height = 8192; + } + + dev->mode_config.preferred_depth = 24; + dev->mode_config.prefer_shadow = 1; + + drm_kms_helper_poll_init(dev); + drm_kms_helper_poll_disable(dev); + + ret = disp->create(dev); + if (ret) + return ret; + + if (dev->mode_config.num_crtc) { + ret = drm_vblank_init(dev, dev->mode_config.num_crtc); + if (ret) + return ret; + } + + return ret; +} + +void +nouveau_display_destroy(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_display_engine *disp = &dev_priv->engine.display; + + drm_vblank_cleanup(dev); + + disp->destroy(dev); + + drm_kms_helper_poll_fini(dev); + drm_mode_config_cleanup(dev); +} + int nouveau_vblank_enable(struct drm_device *dev, int crtc) { @@ -243,15 +441,19 @@ nouveau_page_flip_emit(struct nouveau_channel *chan, goto fail; /* Emit the pageflip */ - ret = RING_SPACE(chan, 2); + ret = RING_SPACE(chan, 3); if (ret) goto fail; - if (dev_priv->card_type < NV_C0) + if (dev_priv->card_type < NV_C0) { BEGIN_RING(chan, NvSubSw, NV_SW_PAGE_FLIP, 1); - else - BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0500, 1); - OUT_RING (chan, 0); + OUT_RING (chan, 0x00000000); + OUT_RING (chan, 0x00000000); + } else { + BEGIN_NVC0(chan, 2, 0, NV10_SUBCHAN_REF_CNT, 1); + OUT_RING (chan, ++chan->fence.sequence); + BEGIN_NVC0(chan, 8, 0, NVSW_SUBCHAN_PAGE_FLIP, 0x0000); + } FIRE_RING (chan); ret = nouveau_fence_new(chan, pfence, true); @@ -294,7 +496,7 @@ nouveau_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb, /* Initialize a page flip struct */ *s = (struct nouveau_page_flip_state) { { }, event, nouveau_crtc(crtc)->index, - fb->bits_per_pixel, fb->pitch, crtc->x, crtc->y, + fb->bits_per_pixel, fb->pitches[0], crtc->x, crtc->y, new_bo->bo.offset }; /* Choose the channel the flip will be handled in */ @@ -305,7 +507,10 @@ nouveau_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb, /* Emit a page flip */ if (dev_priv->card_type >= NV_50) { - ret = nv50_display_flip_next(crtc, fb, chan); + if (dev_priv->card_type >= NV_D0) + ret = nvd0_display_flip_next(crtc, fb, chan, 0); + else + ret = nv50_display_flip_next(crtc, fb, chan); if (ret) { nouveau_channel_put(&chan); goto fail_unreserve; diff --git a/drivers/gpu/drm/nouveau/nouveau_dma.c b/drivers/gpu/drm/nouveau/nouveau_dma.c index 00bc6ea..295932e 100644 --- a/drivers/gpu/drm/nouveau/nouveau_dma.c +++ b/drivers/gpu/drm/nouveau/nouveau_dma.c @@ -31,7 +31,7 @@ #include "nouveau_ramht.h" void -nouveau_dma_pre_init(struct nouveau_channel *chan) +nouveau_dma_init(struct nouveau_channel *chan) { struct drm_nouveau_private *dev_priv = chan->dev->dev_private; struct nouveau_bo *pushbuf = chan->pushbuf_bo; @@ -54,65 +54,6 @@ nouveau_dma_pre_init(struct nouveau_channel *chan) chan->dma.free = chan->dma.max - chan->dma.cur; } -int -nouveau_dma_init(struct nouveau_channel *chan) -{ - struct drm_device *dev = chan->dev; - struct drm_nouveau_private *dev_priv = dev->dev_private; - int ret, i; - - if (dev_priv->card_type >= NV_C0) { - ret = nouveau_gpuobj_gr_new(chan, 0x9039, 0x9039); - if (ret) - return ret; - - ret = RING_SPACE(chan, 2); - if (ret) - return ret; - - BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0000, 1); - OUT_RING (chan, 0x00009039); - FIRE_RING (chan); - return 0; - } - - /* Create NV_MEMORY_TO_MEMORY_FORMAT for buffer moves */ - ret = nouveau_gpuobj_gr_new(chan, NvM2MF, dev_priv->card_type < NV_50 ? - 0x0039 : 0x5039); - if (ret) - return ret; - - /* NV_MEMORY_TO_MEMORY_FORMAT requires a notifier object */ - ret = nouveau_notifier_alloc(chan, NvNotify0, 32, 0xfe0, 0x1000, - &chan->m2mf_ntfy); - if (ret) - return ret; - - /* Insert NOPS for NOUVEAU_DMA_SKIPS */ - ret = RING_SPACE(chan, NOUVEAU_DMA_SKIPS); - if (ret) - return ret; - - for (i = 0; i < NOUVEAU_DMA_SKIPS; i++) - OUT_RING(chan, 0); - - /* Initialise NV_MEMORY_TO_MEMORY_FORMAT */ - ret = RING_SPACE(chan, 6); - if (ret) - return ret; - BEGIN_RING(chan, NvSubM2MF, NV_MEMORY_TO_MEMORY_FORMAT_NAME, 1); - OUT_RING (chan, NvM2MF); - BEGIN_RING(chan, NvSubM2MF, NV_MEMORY_TO_MEMORY_FORMAT_DMA_NOTIFY, 3); - OUT_RING (chan, NvNotify0); - OUT_RING (chan, chan->vram_handle); - OUT_RING (chan, chan->gart_handle); - - /* Sit back and pray the channel works.. */ - FIRE_RING(chan); - - return 0; -} - void OUT_RINGp(struct nouveau_channel *chan, const void *data, unsigned nr_dwords) { @@ -134,11 +75,13 @@ OUT_RINGp(struct nouveau_channel *chan, const void *data, unsigned nr_dwords) * -EBUSY if timeout exceeded */ static inline int -READ_GET(struct nouveau_channel *chan, uint32_t *prev_get, uint32_t *timeout) +READ_GET(struct nouveau_channel *chan, uint64_t *prev_get, int *timeout) { - uint32_t val; + uint64_t val; val = nvchan_rd32(chan, chan->user_get); + if (chan->user_get_hi) + val |= (uint64_t)nvchan_rd32(chan, chan->user_get_hi) << 32; /* reset counter as long as GET is still advancing, this is * to avoid misdetecting a GPU lockup if the GPU happens to @@ -218,8 +161,8 @@ nv50_dma_push_wait(struct nouveau_channel *chan, int count) static int nv50_dma_wait(struct nouveau_channel *chan, int slots, int count) { - uint32_t cnt = 0, prev_get = 0; - int ret; + uint64_t prev_get = 0; + int ret, cnt = 0; ret = nv50_dma_push_wait(chan, slots + 1); if (unlikely(ret)) @@ -261,8 +204,8 @@ nv50_dma_wait(struct nouveau_channel *chan, int slots, int count) int nouveau_dma_wait(struct nouveau_channel *chan, int slots, int size) { - uint32_t prev_get = 0, cnt = 0; - int get; + uint64_t prev_get = 0; + int cnt = 0, get; if (chan->dma.ib_max) return nv50_dma_wait(chan, slots, size); diff --git a/drivers/gpu/drm/nouveau/nouveau_dp.c b/drivers/gpu/drm/nouveau/nouveau_dp.c index de5efe7..d996134 100644 --- a/drivers/gpu/drm/nouveau/nouveau_dp.c +++ b/drivers/gpu/drm/nouveau/nouveau_dp.c @@ -29,6 +29,7 @@ #include "nouveau_connector.h" #include "nouveau_encoder.h" #include "nouveau_crtc.h" +#include "nouveau_gpio.h" /****************************************************************************** * aux channel util functions @@ -160,121 +161,9 @@ out: return ret; } -static u32 -dp_link_bw_get(struct drm_device *dev, int or, int link) -{ - u32 ctrl = nv_rd32(dev, 0x614300 + (or * 0x800)); - if (!(ctrl & 0x000c0000)) - return 162000; - return 270000; -} - -static int -dp_lane_count_get(struct drm_device *dev, int or, int link) -{ - u32 ctrl = nv_rd32(dev, NV50_SOR_DP_CTRL(or, link)); - switch (ctrl & 0x000f0000) { - case 0x00010000: return 1; - case 0x00030000: return 2; - default: - return 4; - } -} - -void -nouveau_dp_tu_update(struct drm_device *dev, int or, int link, u32 clk, u32 bpp) -{ - const u32 symbol = 100000; - int bestTU = 0, bestVTUi = 0, bestVTUf = 0, bestVTUa = 0; - int TU, VTUi, VTUf, VTUa; - u64 link_data_rate, link_ratio, unk; - u32 best_diff = 64 * symbol; - u32 link_nr, link_bw, r; - - /* calculate packed data rate for each lane */ - link_nr = dp_lane_count_get(dev, or, link); - link_data_rate = (clk * bpp / 8) / link_nr; - - /* calculate ratio of packed data rate to link symbol rate */ - link_bw = dp_link_bw_get(dev, or, link); - link_ratio = link_data_rate * symbol; - r = do_div(link_ratio, link_bw); - - for (TU = 64; TU >= 32; TU--) { - /* calculate average number of valid symbols in each TU */ - u32 tu_valid = link_ratio * TU; - u32 calc, diff; - - /* find a hw representation for the fraction.. */ - VTUi = tu_valid / symbol; - calc = VTUi * symbol; - diff = tu_valid - calc; - if (diff) { - if (diff >= (symbol / 2)) { - VTUf = symbol / (symbol - diff); - if (symbol - (VTUf * diff)) - VTUf++; - - if (VTUf <= 15) { - VTUa = 1; - calc += symbol - (symbol / VTUf); - } else { - VTUa = 0; - VTUf = 1; - calc += symbol; - } - } else { - VTUa = 0; - VTUf = min((int)(symbol / diff), 15); - calc += symbol / VTUf; - } - - diff = calc - tu_valid; - } else { - /* no remainder, but the hw doesn't like the fractional - * part to be zero. decrement the integer part and - * have the fraction add a whole symbol back - */ - VTUa = 0; - VTUf = 1; - VTUi--; - } - - if (diff < best_diff) { - best_diff = diff; - bestTU = TU; - bestVTUa = VTUa; - bestVTUf = VTUf; - bestVTUi = VTUi; - if (diff == 0) - break; - } - } - - if (!bestTU) { - NV_ERROR(dev, "DP: unable to find suitable config\n"); - return; - } - - /* XXX close to vbios numbers, but not right */ - unk = (symbol - link_ratio) * bestTU; - unk *= link_ratio; - r = do_div(unk, symbol); - r = do_div(unk, symbol); - unk += 6; - - nv_mask(dev, NV50_SOR_DP_CTRL(or, link), 0x000001fc, bestTU << 2); - nv_mask(dev, NV50_SOR_DP_SCFG(or, link), 0x010f7f3f, bestVTUa << 24 | - bestVTUf << 16 | - bestVTUi << 8 | - unk); -} - u8 * nouveau_dp_bios_data(struct drm_device *dev, struct dcb_entry *dcb, u8 **entry) { - struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->vbios; struct bit_entry d; u8 *table; int i; @@ -289,7 +178,7 @@ nouveau_dp_bios_data(struct drm_device *dev, struct dcb_entry *dcb, u8 **entry) return NULL; } - table = ROMPTR(bios, d.data[0]); + table = ROMPTR(dev, d.data[0]); if (!table) { NV_ERROR(dev, "displayport table pointer invalid\n"); return NULL; @@ -299,6 +188,7 @@ nouveau_dp_bios_data(struct drm_device *dev, struct dcb_entry *dcb, u8 **entry) case 0x20: case 0x21: case 0x30: + case 0x40: break; default: NV_ERROR(dev, "displayport table 0x%02x unknown\n", table[0]); @@ -306,7 +196,7 @@ nouveau_dp_bios_data(struct drm_device *dev, struct dcb_entry *dcb, u8 **entry) } for (i = 0; i < table[3]; i++) { - *entry = ROMPTR(bios, table[table[1] + (i * table[2])]); + *entry = ROMPTR(dev, table[table[1] + (i * table[2])]); if (*entry && bios_encoder_match(dcb, ROM32((*entry)[0]))) return table; } @@ -319,13 +209,10 @@ nouveau_dp_bios_data(struct drm_device *dev, struct dcb_entry *dcb, u8 **entry) * link training *****************************************************************************/ struct dp_state { + struct dp_train_func *func; struct dcb_entry *dcb; - u8 *table; - u8 *entry; int auxch; int crtc; - int or; - int link; u8 *dpcd; int link_nr; u32 link_bw; @@ -336,143 +223,58 @@ struct dp_state { static void dp_set_link_config(struct drm_device *dev, struct dp_state *dp) { - struct drm_nouveau_private *dev_priv = dev->dev_private; - int or = dp->or, link = dp->link; - u8 *entry, sink[2]; - u32 dp_ctrl; - u16 script; + u8 sink[2]; NV_DEBUG_KMS(dev, "%d lanes at %d KB/s\n", dp->link_nr, dp->link_bw); - /* set selected link rate on source */ - switch (dp->link_bw) { - case 270000: - nv_mask(dev, 0x614300 + (or * 0x800), 0x000c0000, 0x00040000); - sink[0] = DP_LINK_BW_2_7; - break; - default: - nv_mask(dev, 0x614300 + (or * 0x800), 0x000c0000, 0x00000000); - sink[0] = DP_LINK_BW_1_62; - break; - } - - /* offset +0x0a of each dp encoder table entry is a pointer to another - * table, that has (among other things) pointers to more scripts that - * need to be executed, this time depending on link speed. - */ - entry = ROMPTR(&dev_priv->vbios, dp->entry[10]); - if (entry) { - if (dp->table[0] < 0x30) { - while (dp->link_bw < (ROM16(entry[0]) * 10)) - entry += 4; - script = ROM16(entry[2]); - } else { - while (dp->link_bw < (entry[0] * 27000)) - entry += 3; - script = ROM16(entry[1]); - } - - nouveau_bios_run_init_table(dev, script, dp->dcb, dp->crtc); - } + /* set desired link configuration on the source */ + dp->func->link_set(dev, dp->dcb, dp->crtc, dp->link_nr, dp->link_bw, + dp->dpcd[2] & DP_ENHANCED_FRAME_CAP); - /* configure lane count on the source */ - dp_ctrl = ((1 << dp->link_nr) - 1) << 16; + /* inform the sink of the new configuration */ + sink[0] = dp->link_bw / 27000; sink[1] = dp->link_nr; - if (dp->dpcd[2] & DP_ENHANCED_FRAME_CAP) { - dp_ctrl |= 0x00004000; + if (dp->dpcd[2] & DP_ENHANCED_FRAME_CAP) sink[1] |= DP_LANE_COUNT_ENHANCED_FRAME_EN; - } - - nv_mask(dev, NV50_SOR_DP_CTRL(or, link), 0x001f4000, dp_ctrl); - /* inform the sink of the new configuration */ auxch_tx(dev, dp->auxch, 8, DP_LINK_BW_SET, sink, 2); } static void -dp_set_training_pattern(struct drm_device *dev, struct dp_state *dp, u8 tp) +dp_set_training_pattern(struct drm_device *dev, struct dp_state *dp, u8 pattern) { u8 sink_tp; - NV_DEBUG_KMS(dev, "training pattern %d\n", tp); + NV_DEBUG_KMS(dev, "training pattern %d\n", pattern); - nv_mask(dev, NV50_SOR_DP_CTRL(dp->or, dp->link), 0x0f000000, tp << 24); + dp->func->train_set(dev, dp->dcb, pattern); auxch_tx(dev, dp->auxch, 9, DP_TRAINING_PATTERN_SET, &sink_tp, 1); sink_tp &= ~DP_TRAINING_PATTERN_MASK; - sink_tp |= tp; + sink_tp |= pattern; auxch_tx(dev, dp->auxch, 8, DP_TRAINING_PATTERN_SET, &sink_tp, 1); } -static const u8 nv50_lane_map[] = { 16, 8, 0, 24 }; -static const u8 nvaf_lane_map[] = { 24, 16, 8, 0 }; - static int dp_link_train_commit(struct drm_device *dev, struct dp_state *dp) { - struct drm_nouveau_private *dev_priv = dev->dev_private; - u32 mask = 0, drv = 0, pre = 0, unk = 0; - const u8 *shifts; - int link = dp->link; - int or = dp->or; int i; - if (dev_priv->chipset != 0xaf) - shifts = nv50_lane_map; - else - shifts = nvaf_lane_map; - for (i = 0; i < dp->link_nr; i++) { - u8 *conf = dp->entry + dp->table[4]; u8 lane = (dp->stat[4 + (i >> 1)] >> ((i & 1) * 4)) & 0xf; u8 lpre = (lane & 0x0c) >> 2; u8 lvsw = (lane & 0x03) >> 0; - mask |= 0xff << shifts[i]; - unk |= 1 << (shifts[i] >> 3); - dp->conf[i] = (lpre << 3) | lvsw; if (lvsw == DP_TRAIN_VOLTAGE_SWING_1200) dp->conf[i] |= DP_TRAIN_MAX_SWING_REACHED; - if (lpre == DP_TRAIN_PRE_EMPHASIS_9_5) + if ((lpre << 3) == DP_TRAIN_PRE_EMPHASIS_9_5) dp->conf[i] |= DP_TRAIN_MAX_PRE_EMPHASIS_REACHED; NV_DEBUG_KMS(dev, "config lane %d %02x\n", i, dp->conf[i]); - - if (dp->table[0] < 0x30) { - u8 *last = conf + (dp->entry[4] * dp->table[5]); - while (lvsw != conf[0] || lpre != conf[1]) { - conf += dp->table[5]; - if (conf >= last) - return -EINVAL; - } - - conf += 2; - } else { - /* no lookup table anymore, set entries for each - * combination of voltage swing and pre-emphasis - * level allowed by the DP spec. - */ - switch (lvsw) { - case 0: lpre += 0; break; - case 1: lpre += 4; break; - case 2: lpre += 7; break; - case 3: lpre += 9; break; - } - - conf = conf + (lpre * dp->table[5]); - conf++; - } - - drv |= conf[0] << shifts[i]; - pre |= conf[1] << shifts[i]; - unk = (unk & ~0x0000ff00) | (conf[2] << 8); + dp->func->train_adj(dev, dp->dcb, i, lvsw, lpre); } - nv_mask(dev, NV50_SOR_DP_UNK118(or, link), mask, drv); - nv_mask(dev, NV50_SOR_DP_UNK120(or, link), mask, pre); - nv_mask(dev, NV50_SOR_DP_UNK130(or, link), 0x0000ff0f, unk); - return auxch_tx(dev, dp->auxch, 8, DP_TRAINING_LANE0_SET, dp->conf, 4); } @@ -556,11 +358,61 @@ dp_link_train_eq(struct drm_device *dev, struct dp_state *dp) return eq_done ? 0 : -1; } +static void +dp_set_downspread(struct drm_device *dev, struct dp_state *dp, bool enable) +{ + u16 script = 0x0000; + u8 *entry, *table = nouveau_dp_bios_data(dev, dp->dcb, &entry); + if (table) { + if (table[0] >= 0x20 && table[0] <= 0x30) { + if (enable) script = ROM16(entry[12]); + else script = ROM16(entry[14]); + } else + if (table[0] == 0x40) { + if (enable) script = ROM16(entry[11]); + else script = ROM16(entry[13]); + } + } + + nouveau_bios_run_init_table(dev, script, dp->dcb, dp->crtc); +} + +static void +dp_link_train_init(struct drm_device *dev, struct dp_state *dp) +{ + u16 script = 0x0000; + u8 *entry, *table = nouveau_dp_bios_data(dev, dp->dcb, &entry); + if (table) { + if (table[0] >= 0x20 && table[0] <= 0x30) + script = ROM16(entry[6]); + else + if (table[0] == 0x40) + script = ROM16(entry[5]); + } + + nouveau_bios_run_init_table(dev, script, dp->dcb, dp->crtc); +} + +static void +dp_link_train_fini(struct drm_device *dev, struct dp_state *dp) +{ + u16 script = 0x0000; + u8 *entry, *table = nouveau_dp_bios_data(dev, dp->dcb, &entry); + if (table) { + if (table[0] >= 0x20 && table[0] <= 0x30) + script = ROM16(entry[8]); + else + if (table[0] == 0x40) + script = ROM16(entry[7]); + } + + nouveau_bios_run_init_table(dev, script, dp->dcb, dp->crtc); +} + bool -nouveau_dp_link_train(struct drm_encoder *encoder, u32 datarate) +nouveau_dp_link_train(struct drm_encoder *encoder, u32 datarate, + struct dp_train_func *func) { - struct drm_nouveau_private *dev_priv = encoder->dev->dev_private; - struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc); struct nouveau_connector *nv_connector = @@ -575,34 +427,26 @@ nouveau_dp_link_train(struct drm_encoder *encoder, u32 datarate) if (!auxch) return false; - dp.table = nouveau_dp_bios_data(dev, nv_encoder->dcb, &dp.entry); - if (!dp.table) - return -EINVAL; - + dp.func = func; dp.dcb = nv_encoder->dcb; dp.crtc = nv_crtc->index; - dp.auxch = auxch->rd; - dp.or = nv_encoder->or; - dp.link = !(nv_encoder->dcb->sorconf.link & 1); + dp.auxch = auxch->drive; dp.dpcd = nv_encoder->dp.dpcd; + /* adjust required bandwidth for 8B/10B coding overhead */ + datarate = (datarate / 8) * 10; + /* some sinks toggle hotplug in response to some of the actions * we take during link training (DP_SET_POWER is one), we need * to ignore them for the moment to avoid races. */ - pgpio->irq_enable(dev, nv_connector->dcb->gpio_tag, false); + nouveau_gpio_irq(dev, 0, nv_connector->hpd, 0xff, false); /* enable down-spreading, if possible */ - if (dp.table[1] >= 16) { - u16 script = ROM16(dp.entry[14]); - if (nv_encoder->dp.dpcd[3] & 1) - script = ROM16(dp.entry[12]); - - nouveau_bios_run_init_table(dev, script, dp.dcb, dp.crtc); - } + dp_set_downspread(dev, &dp, nv_encoder->dp.dpcd[3] & 1); /* execute pre-train script from vbios */ - nouveau_bios_run_init_table(dev, ROM16(dp.entry[6]), dp.dcb, dp.crtc); + dp_link_train_init(dev, &dp); /* start off at highest link rate supported by encoder and display */ while (*link_bw > nv_encoder->dp.link_bw) @@ -636,13 +480,36 @@ nouveau_dp_link_train(struct drm_encoder *encoder, u32 datarate) dp_set_training_pattern(dev, &dp, DP_TRAINING_PATTERN_DISABLE); /* execute post-train script from vbios */ - nouveau_bios_run_init_table(dev, ROM16(dp.entry[8]), dp.dcb, dp.crtc); + dp_link_train_fini(dev, &dp); /* re-enable hotplug detect */ - pgpio->irq_enable(dev, nv_connector->dcb->gpio_tag, true); + nouveau_gpio_irq(dev, 0, nv_connector->hpd, 0xff, true); return true; } +void +nouveau_dp_dpms(struct drm_encoder *encoder, int mode, u32 datarate, + struct dp_train_func *func) +{ + struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); + struct nouveau_i2c_chan *auxch; + u8 status; + + auxch = nouveau_i2c_find(encoder->dev, nv_encoder->dcb->i2c_index); + if (!auxch) + return; + + if (mode == DRM_MODE_DPMS_ON) + status = DP_SET_POWER_D0; + else + status = DP_SET_POWER_D3; + + nouveau_dp_auxch(auxch, 8, DP_SET_POWER, &status, 1); + + if (mode == DRM_MODE_DPMS_ON) + nouveau_dp_link_train(encoder, datarate, func); +} + bool nouveau_dp_detect(struct drm_encoder *encoder) { @@ -656,7 +523,7 @@ nouveau_dp_detect(struct drm_encoder *encoder) if (!auxch) return false; - ret = auxch_tx(dev, auxch->rd, 9, DP_DPCD_REV, dpcd, 8); + ret = auxch_tx(dev, auxch->drive, 9, DP_DPCD_REV, dpcd, 8); if (ret) return false; @@ -684,7 +551,7 @@ int nouveau_dp_auxch(struct nouveau_i2c_chan *auxch, int cmd, int addr, uint8_t *data, int data_nr) { - return auxch_tx(auxch->dev, auxch->rd, cmd, addr, data, data_nr); + return auxch_tx(auxch->dev, auxch->drive, cmd, addr, data, data_nr); } static int diff --git a/drivers/gpu/drm/nouveau/nouveau_drv.c b/drivers/gpu/drm/nouveau/nouveau_drv.c index 8c084c0..05091c2 100644 --- a/drivers/gpu/drm/nouveau/nouveau_drv.c +++ b/drivers/gpu/drm/nouveau/nouveau_drv.c @@ -57,6 +57,10 @@ MODULE_PARM_DESC(vram_notify, "Force DMA notifiers to be in VRAM"); int nouveau_vram_notify = 0; module_param_named(vram_notify, nouveau_vram_notify, int, 0400); +MODULE_PARM_DESC(vram_type, "Override detected VRAM type"); +char *nouveau_vram_type; +module_param_named(vram_type, nouveau_vram_type, charp, 0400); + MODULE_PARM_DESC(duallink, "Allow dual-link TMDS (>=GeForce 8)"); int nouveau_duallink = 1; module_param_named(duallink, nouveau_duallink, int, 0400); @@ -89,7 +93,7 @@ MODULE_PARM_DESC(override_conntype, "Ignore DCB connector type"); int nouveau_override_conntype = 0; module_param_named(override_conntype, nouveau_override_conntype, int, 0400); -MODULE_PARM_DESC(tv_disable, "Disable TV-out detection\n"); +MODULE_PARM_DESC(tv_disable, "Disable TV-out detection"); int nouveau_tv_disable = 0; module_param_named(tv_disable, nouveau_tv_disable, int, 0400); @@ -104,26 +108,30 @@ module_param_named(tv_norm, nouveau_tv_norm, charp, 0400); MODULE_PARM_DESC(reg_debug, "Register access debug bitmask:\n" "\t\t0x1 mc, 0x2 video, 0x4 fb, 0x8 extdev,\n" "\t\t0x10 crtc, 0x20 ramdac, 0x40 vgacrtc, 0x80 rmvio,\n" - "\t\t0x100 vgaattr, 0x200 EVO (G80+). "); + "\t\t0x100 vgaattr, 0x200 EVO (G80+)"); int nouveau_reg_debug; module_param_named(reg_debug, nouveau_reg_debug, int, 0600); -MODULE_PARM_DESC(perflvl, "Performance level (default: boot)\n"); +MODULE_PARM_DESC(perflvl, "Performance level (default: boot)"); char *nouveau_perflvl; module_param_named(perflvl, nouveau_perflvl, charp, 0400); -MODULE_PARM_DESC(perflvl_wr, "Allow perflvl changes (warning: dangerous!)\n"); +MODULE_PARM_DESC(perflvl_wr, "Allow perflvl changes (warning: dangerous!)"); int nouveau_perflvl_wr; module_param_named(perflvl_wr, nouveau_perflvl_wr, int, 0400); -MODULE_PARM_DESC(msi, "Enable MSI (default: off)\n"); +MODULE_PARM_DESC(msi, "Enable MSI (default: off)"); int nouveau_msi; module_param_named(msi, nouveau_msi, int, 0400); -MODULE_PARM_DESC(ctxfw, "Use external HUB/GPC ucode (fermi)\n"); +MODULE_PARM_DESC(ctxfw, "Use external HUB/GPC ucode (fermi)"); int nouveau_ctxfw; module_param_named(ctxfw, nouveau_ctxfw, int, 0400); +MODULE_PARM_DESC(mxmdcb, "Santise DCB table according to MXM-SIS"); +int nouveau_mxmdcb = 1; +module_param_named(mxmdcb, nouveau_mxmdcb, int, 0400); + int nouveau_fbpercrtc; #if 0 module_param_named(fbpercrtc, nouveau_fbpercrtc, int, 0400); @@ -179,8 +187,11 @@ nouveau_pci_suspend(struct pci_dev *pdev, pm_message_t pm_state) return 0; if (dev->mode_config.num_crtc) { - NV_INFO(dev, "Disabling fbcon acceleration...\n"); - nouveau_fbcon_save_disable_accel(dev); + NV_INFO(dev, "Disabling display...\n"); + nouveau_display_fini(dev); + + NV_INFO(dev, "Disabling fbcon...\n"); + nouveau_fbcon_set_suspend(dev, 1); } NV_INFO(dev, "Unpinning framebuffer(s)...\n"); @@ -222,7 +233,7 @@ nouveau_pci_suspend(struct pci_dev *pdev, pm_message_t pm_state) ret = dev_priv->eng[e]->fini(dev, e, true); if (ret) { - NV_ERROR(dev, "... engine %d failed: %d\n", i, ret); + NV_ERROR(dev, "... engine %d failed: %d\n", e, ret); goto out_abort; } } @@ -248,12 +259,6 @@ nouveau_pci_suspend(struct pci_dev *pdev, pm_message_t pm_state) pci_set_power_state(pdev, PCI_D3hot); } - if (dev->mode_config.num_crtc) { - console_lock(); - nouveau_fbcon_set_suspend(dev, 1); - console_unlock(); - nouveau_fbcon_restore_accel(dev); - } return 0; out_abort: @@ -279,9 +284,6 @@ nouveau_pci_resume(struct pci_dev *pdev) if (dev->switch_power_state == DRM_SWITCH_POWER_OFF) return 0; - if (dev->mode_config.num_crtc) - nouveau_fbcon_save_disable_accel(dev); - NV_INFO(dev, "We're back, enabling device...\n"); pci_set_power_state(pdev, PCI_D0); pci_restore_state(pdev); @@ -301,8 +303,6 @@ nouveau_pci_resume(struct pci_dev *pdev) if (ret) return ret; - nouveau_pm_resume(dev); - if (dev_priv->gart_info.type == NOUVEAU_GART_AGP) { ret = nouveau_mem_init_agp(dev); if (ret) { @@ -342,6 +342,8 @@ nouveau_pci_resume(struct pci_dev *pdev) } } + nouveau_pm_resume(dev); + NV_INFO(dev, "Restoring mode...\n"); list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { struct nouveau_framebuffer *nouveau_fb; @@ -363,15 +365,11 @@ nouveau_pci_resume(struct pci_dev *pdev) NV_ERROR(dev, "Could not pin/map cursor.\n"); } - engine->display.init(dev); - - list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { - struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc); - u32 offset = nv_crtc->cursor.nvbo->bo.offset; + if (dev->mode_config.num_crtc) { + nouveau_fbcon_set_suspend(dev, 0); + nouveau_fbcon_zfill_all(dev); - nv_crtc->cursor.set_offset(nv_crtc, offset); - nv_crtc->cursor.set_pos(nv_crtc, nv_crtc->cursor_saved_x, - nv_crtc->cursor_saved_y); + nouveau_display_init(dev); } /* Force CLUT to get re-loaded during modeset */ @@ -381,21 +379,35 @@ nouveau_pci_resume(struct pci_dev *pdev) nv_crtc->lut.depth = 0; } - if (dev->mode_config.num_crtc) { - console_lock(); - nouveau_fbcon_set_suspend(dev, 0); - console_unlock(); + drm_helper_resume_force_mode(dev); - nouveau_fbcon_zfill_all(dev); - } + list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { + struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc); + u32 offset = nv_crtc->cursor.nvbo->bo.offset; - drm_helper_resume_force_mode(dev); + nv_crtc->cursor.set_offset(nv_crtc, offset); + nv_crtc->cursor.set_pos(nv_crtc, nv_crtc->cursor_saved_x, + nv_crtc->cursor_saved_y); + } - if (dev->mode_config.num_crtc) - nouveau_fbcon_restore_accel(dev); return 0; } +static const struct file_operations nouveau_driver_fops = { + .owner = THIS_MODULE, + .open = drm_open, + .release = drm_release, + .unlocked_ioctl = drm_ioctl, + .mmap = nouveau_ttm_mmap, + .poll = drm_poll, + .fasync = drm_fasync, + .read = drm_read, +#if defined(CONFIG_COMPAT) + .compat_ioctl = nouveau_compat_ioctl, +#endif + .llseek = noop_llseek, +}; + static struct drm_driver driver = { .driver_features = DRIVER_USE_AGP | DRIVER_PCI_DMA | DRIVER_SG | @@ -421,21 +433,7 @@ static struct drm_driver driver = { .disable_vblank = nouveau_vblank_disable, .reclaim_buffers = drm_core_reclaim_buffers, .ioctls = nouveau_ioctls, - .fops = { - .owner = THIS_MODULE, - .open = drm_open, - .release = drm_release, - .unlocked_ioctl = drm_ioctl, - .mmap = nouveau_ttm_mmap, - .poll = drm_poll, - .fasync = drm_fasync, - .read = drm_read, -#if defined(CONFIG_COMPAT) - .compat_ioctl = nouveau_compat_ioctl, -#endif - .llseek = noop_llseek, - }, - + .fops = &nouveau_driver_fops, .gem_init_object = nouveau_gem_object_new, .gem_free_object = nouveau_gem_object_del, .gem_open_object = nouveau_gem_object_open, diff --git a/drivers/gpu/drm/nouveau/nouveau_drv.h b/drivers/gpu/drm/nouveau/nouveau_drv.h index 4c0be3a..3aef353 100644 --- a/drivers/gpu/drm/nouveau/nouveau_drv.h +++ b/drivers/gpu/drm/nouveau/nouveau_drv.h @@ -26,15 +26,15 @@ #define __NOUVEAU_DRV_H__ #define DRIVER_AUTHOR "Stephane Marchesin" -#define DRIVER_EMAIL "dri-devel@lists.sourceforge.net" +#define DRIVER_EMAIL "nouveau@lists.freedesktop.org" #define DRIVER_NAME "nouveau" #define DRIVER_DESC "nVidia Riva/TNT/GeForce" -#define DRIVER_DATE "20090420" +#define DRIVER_DATE "20120316" -#define DRIVER_MAJOR 0 +#define DRIVER_MAJOR 1 #define DRIVER_MINOR 0 -#define DRIVER_PATCHLEVEL 16 +#define DRIVER_PATCHLEVEL 0 #define NOUVEAU_FAMILY 0x0000FFFF #define NOUVEAU_FLAGS 0xFFFF0000 @@ -113,8 +113,6 @@ struct nouveau_bo { int pbbo_index; bool validate_mapped; - struct nouveau_channel *channel; - struct list_head vma_list; unsigned page_shift; @@ -163,6 +161,9 @@ enum nouveau_flags { #define NVOBJ_ENGINE_COPY0 3 #define NVOBJ_ENGINE_COPY1 4 #define NVOBJ_ENGINE_MPEG 5 +#define NVOBJ_ENGINE_PPP NVOBJ_ENGINE_MPEG +#define NVOBJ_ENGINE_BSP 6 +#define NVOBJ_ENGINE_VP 7 #define NVOBJ_ENGINE_DISPLAY 15 #define NVOBJ_ENGINE_NR 16 @@ -229,6 +230,7 @@ struct nouveau_channel { /* mapping of the regs controlling the fifo */ void __iomem *user; uint32_t user_get; + uint32_t user_get_hi; uint32_t user_put; /* Fencing */ @@ -246,7 +248,7 @@ struct nouveau_channel { struct nouveau_gpuobj *pushbuf; struct nouveau_bo *pushbuf_bo; struct nouveau_vma pushbuf_vma; - uint32_t pushbuf_base; + uint64_t pushbuf_base; /* Notifier memory */ struct nouveau_bo *notifier_bo; @@ -292,7 +294,7 @@ struct nouveau_channel { uint32_t sw_subchannel[8]; - struct nouveau_vma dispc_vma[2]; + struct nouveau_vma dispc_vma[4]; struct { struct nouveau_gpuobj *vblsem; uint32_t vblsem_head; @@ -393,24 +395,28 @@ struct nouveau_display_engine { int (*early_init)(struct drm_device *); void (*late_takedown)(struct drm_device *); int (*create)(struct drm_device *); - int (*init)(struct drm_device *); void (*destroy)(struct drm_device *); + int (*init)(struct drm_device *); + void (*fini)(struct drm_device *); + + struct drm_property *dithering_mode; + struct drm_property *dithering_depth; + struct drm_property *underscan_property; + struct drm_property *underscan_hborder_property; + struct drm_property *underscan_vborder_property; + /* not really hue and saturation: */ + struct drm_property *vibrant_hue_property; + struct drm_property *color_vibrance_property; }; struct nouveau_gpio_engine { - void *priv; - - int (*init)(struct drm_device *); - void (*takedown)(struct drm_device *); - - int (*get)(struct drm_device *, enum dcb_gpio_tag); - int (*set)(struct drm_device *, enum dcb_gpio_tag, int state); - - int (*irq_register)(struct drm_device *, enum dcb_gpio_tag, - void (*)(void *, int), void *); - void (*irq_unregister)(struct drm_device *, enum dcb_gpio_tag, - void (*)(void *, int), void *); - bool (*irq_enable)(struct drm_device *, enum dcb_gpio_tag, bool on); + spinlock_t lock; + struct list_head isr; + int (*init)(struct drm_device *); + void (*fini)(struct drm_device *); + int (*drive)(struct drm_device *, int line, int dir, int out); + int (*sense)(struct drm_device *, int line); + void (*irq_enable)(struct drm_device *, int line, bool); }; struct nouveau_pm_voltage_level { @@ -427,64 +433,91 @@ struct nouveau_pm_voltage { int nr_level; }; +/* Exclusive upper limits */ +#define NV_MEM_CL_DDR2_MAX 8 +#define NV_MEM_WR_DDR2_MAX 9 +#define NV_MEM_CL_DDR3_MAX 17 +#define NV_MEM_WR_DDR3_MAX 17 +#define NV_MEM_CL_GDDR3_MAX 16 +#define NV_MEM_WR_GDDR3_MAX 18 +#define NV_MEM_CL_GDDR5_MAX 21 +#define NV_MEM_WR_GDDR5_MAX 20 + struct nouveau_pm_memtiming { int id; - u32 reg_0; /* 0x10f290 on Fermi, 0x100220 for older */ - u32 reg_1; - u32 reg_2; - u32 reg_3; - u32 reg_4; - u32 reg_5; - u32 reg_6; - u32 reg_7; - u32 reg_8; - /* To be written to 0x1002c0 */ - u8 CL; - u8 WR; + + u32 reg[9]; + u32 mr[4]; + + u8 tCWL; + + u8 odt; + u8 drive_strength; }; -struct nouveau_pm_tbl_header{ +struct nouveau_pm_tbl_header { u8 version; u8 header_len; u8 entry_cnt; u8 entry_len; }; -struct nouveau_pm_tbl_entry{ +struct nouveau_pm_tbl_entry { u8 tWR; - u8 tUNK_1; + u8 tWTR; u8 tCL; - u8 tRP; /* Byte 3 */ + u8 tRC; u8 empty_4; - u8 tRAS; /* Byte 5 */ + u8 tRFC; /* Byte 5 */ u8 empty_6; - u8 tRFC; /* Byte 7 */ + u8 tRAS; /* Byte 7 */ u8 empty_8; - u8 tRC; /* Byte 9 */ - u8 tUNK_10, tUNK_11, tUNK_12, tUNK_13, tUNK_14; - u8 empty_15,empty_16,empty_17; - u8 tUNK_18, tUNK_19, tUNK_20, tUNK_21; + u8 tRP; /* Byte 9 */ + u8 tRCDRD; + u8 tRCDWR; + u8 tRRD; + u8 tUNK_13; + u8 RAM_FT1; /* 14, a bitmask of random RAM features */ + u8 empty_15; + u8 tUNK_16; + u8 empty_17; + u8 tUNK_18; + u8 tCWL; + u8 tUNK_20, tUNK_21; }; -/* nouveau_mem.c */ -void nv30_mem_timing_entry(struct drm_device *dev, struct nouveau_pm_tbl_header *hdr, - struct nouveau_pm_tbl_entry *e, uint8_t magic_number, - struct nouveau_pm_memtiming *timing); +struct nouveau_pm_profile; +struct nouveau_pm_profile_func { + void (*destroy)(struct nouveau_pm_profile *); + void (*init)(struct nouveau_pm_profile *); + void (*fini)(struct nouveau_pm_profile *); + struct nouveau_pm_level *(*select)(struct nouveau_pm_profile *); +}; + +struct nouveau_pm_profile { + const struct nouveau_pm_profile_func *func; + struct list_head head; + char name[8]; +}; #define NOUVEAU_PM_MAX_LEVEL 8 struct nouveau_pm_level { + struct nouveau_pm_profile profile; struct device_attribute dev_attr; char name[32]; int id; - u32 core; + struct nouveau_pm_memtiming timing; u32 memory; + u16 memscript; + + u32 core; u32 shader; u32 rop; u32 copy; u32 daemon; u32 vdec; - u32 unk05; /* nv50:nva3, roughly.. */ + u32 dom6; u32 unka0; /* nva3:nvc0 */ u32 hub01; /* nvc0- */ u32 hub06; /* nvc0- */ @@ -493,9 +526,6 @@ struct nouveau_pm_level { u32 volt_min; /* microvolts */ u32 volt_max; u8 fanspeed; - - u16 memscript; - struct nouveau_pm_memtiming *timing; }; struct nouveau_pm_temp_sensor_constants { @@ -512,19 +542,26 @@ struct nouveau_pm_threshold_temp { s16 fan_boost; }; -struct nouveau_pm_memtimings { - bool supported; - struct nouveau_pm_memtiming *timing; - int nr_timing; +struct nouveau_pm_fan { + u32 percent; + u32 min_duty; + u32 max_duty; + u32 pwm_freq; + u32 pwm_divisor; }; struct nouveau_pm_engine { struct nouveau_pm_voltage voltage; struct nouveau_pm_level perflvl[NOUVEAU_PM_MAX_LEVEL]; int nr_perflvl; - struct nouveau_pm_memtimings memtimings; struct nouveau_pm_temp_sensor_constants sensor_constants; struct nouveau_pm_threshold_temp threshold_temp; + struct nouveau_pm_fan fan; + + struct nouveau_pm_profile *profile_ac; + struct nouveau_pm_profile *profile_dc; + struct nouveau_pm_profile *profile; + struct list_head profiles; struct nouveau_pm_level boot; struct nouveau_pm_level *cur; @@ -532,19 +569,14 @@ struct nouveau_pm_engine { struct device *hwmon; struct notifier_block acpi_nb; - int (*clock_get)(struct drm_device *, u32 id); - void *(*clock_pre)(struct drm_device *, struct nouveau_pm_level *, - u32 id, int khz); - void (*clock_set)(struct drm_device *, void *); - int (*clocks_get)(struct drm_device *, struct nouveau_pm_level *); void *(*clocks_pre)(struct drm_device *, struct nouveau_pm_level *); - void (*clocks_set)(struct drm_device *, void *); + int (*clocks_set)(struct drm_device *, void *); int (*voltage_get)(struct drm_device *); int (*voltage_set)(struct drm_device *, int voltage); - int (*fanspeed_get)(struct drm_device *); - int (*fanspeed_set)(struct drm_device *, int fanspeed); + int (*pwm_get)(struct drm_device *, int line, u32*, u32*); + int (*pwm_set)(struct drm_device *, int line, u32, u32); int (*temp_get)(struct drm_device *); }; @@ -661,14 +693,15 @@ struct nv04_mode_state { }; enum nouveau_card_type { - NV_04 = 0x00, + NV_04 = 0x04, NV_10 = 0x10, NV_20 = 0x20, NV_30 = 0x30, NV_40 = 0x40, NV_50 = 0x50, NV_C0 = 0xc0, - NV_D0 = 0xd0 + NV_D0 = 0xd0, + NV_E0 = 0xe0, }; struct drm_nouveau_private { @@ -764,8 +797,22 @@ struct drm_nouveau_private { } tile; /* VRAM/fb configuration */ + enum { + NV_MEM_TYPE_UNKNOWN = 0, + NV_MEM_TYPE_STOLEN, + NV_MEM_TYPE_SGRAM, + NV_MEM_TYPE_SDRAM, + NV_MEM_TYPE_DDR1, + NV_MEM_TYPE_DDR2, + NV_MEM_TYPE_DDR3, + NV_MEM_TYPE_GDDR2, + NV_MEM_TYPE_GDDR3, + NV_MEM_TYPE_GDDR4, + NV_MEM_TYPE_GDDR5 + } vram_type; uint64_t vram_size; uint64_t vram_sys_base; + bool vram_rank_B; uint64_t fb_available_size; uint64_t fb_mappable_pages; @@ -780,6 +827,8 @@ struct drm_nouveau_private { struct nouveau_vm *chan_vm; struct nvbios vbios; + u8 *mxms; + struct list_head i2c_ports; struct nv04_mode_state mode_reg; struct nv04_mode_state saved_reg; @@ -836,6 +885,7 @@ extern int nouveau_uscript_lvds; extern int nouveau_uscript_tmds; extern int nouveau_vram_pushbuf; extern int nouveau_vram_notify; +extern char *nouveau_vram_type; extern int nouveau_fbpercrtc; extern int nouveau_tv_disable; extern char *nouveau_tv_norm; @@ -850,6 +900,7 @@ extern char *nouveau_perflvl; extern int nouveau_perflvl_wr; extern int nouveau_msi; extern int nouveau_ctxfw; +extern int nouveau_mxmdcb; extern int nouveau_pci_suspend(struct pci_dev *pdev, pm_message_t pm_state); extern int nouveau_pci_resume(struct pci_dev *pdev); @@ -883,8 +934,12 @@ extern void nouveau_mem_gart_fini(struct drm_device *); extern int nouveau_mem_init_agp(struct drm_device *); extern int nouveau_mem_reset_agp(struct drm_device *); extern void nouveau_mem_close(struct drm_device *); -extern int nouveau_mem_detect(struct drm_device *); extern bool nouveau_mem_flags_valid(struct drm_device *, u32 tile_flags); +extern int nouveau_mem_timing_calc(struct drm_device *, u32 freq, + struct nouveau_pm_memtiming *); +extern void nouveau_mem_timing_read(struct drm_device *, + struct nouveau_pm_memtiming *); +extern int nouveau_mem_vbios_type(struct drm_device *); extern struct nouveau_tile_reg *nv10_mem_set_tiling( struct drm_device *dev, uint32_t addr, uint32_t size, uint32_t pitch, uint32_t flags); @@ -1000,7 +1055,10 @@ extern int nouveau_sgdma_init(struct drm_device *); extern void nouveau_sgdma_takedown(struct drm_device *); extern uint32_t nouveau_sgdma_get_physical(struct drm_device *, uint32_t offset); -extern struct ttm_backend *nouveau_sgdma_init_ttm(struct drm_device *); +extern struct ttm_tt *nouveau_sgdma_create_ttm(struct ttm_bo_device *bdev, + unsigned long size, + uint32_t page_flags, + struct page *dummy_read_page); /* nouveau_debugfs.c */ #if defined(CONFIG_DRM_NOUVEAU_DEBUG) @@ -1032,8 +1090,7 @@ nouveau_debugfs_channel_fini(struct nouveau_channel *chan) #endif /* nouveau_dma.c */ -extern void nouveau_dma_pre_init(struct nouveau_channel *); -extern int nouveau_dma_init(struct nouveau_channel *); +extern void nouveau_dma_init(struct nouveau_channel *); extern int nouveau_dma_wait(struct nouveau_channel *, int slots, int size); /* nouveau_acpi.c */ @@ -1041,12 +1098,14 @@ extern int nouveau_dma_wait(struct nouveau_channel *, int slots, int size); #if defined(CONFIG_ACPI) void nouveau_register_dsm_handler(void); void nouveau_unregister_dsm_handler(void); +void nouveau_switcheroo_optimus_dsm(void); int nouveau_acpi_get_bios_chunk(uint8_t *bios, int offset, int len); bool nouveau_acpi_rom_supported(struct pci_dev *pdev); int nouveau_acpi_edid(struct drm_device *, struct drm_connector *); #else static inline void nouveau_register_dsm_handler(void) {} static inline void nouveau_unregister_dsm_handler(void) {} +static inline void nouveau_switcheroo_optimus_dsm(void) {} static inline bool nouveau_acpi_rom_supported(struct pci_dev *pdev) { return false; } static inline int nouveau_acpi_get_bios_chunk(uint8_t *bios, int offset, int len) { return -EINVAL; } static inline int nouveau_acpi_edid(struct drm_device *dev, struct drm_connector *connector) { return -EINVAL; } @@ -1072,8 +1131,6 @@ extern int nouveau_run_vbios_init(struct drm_device *); extern void nouveau_bios_run_init_table(struct drm_device *, uint16_t table, struct dcb_entry *, int crtc); extern void nouveau_bios_init_exec(struct drm_device *, uint16_t table); -extern struct dcb_gpio_entry *nouveau_bios_gpio_entry(struct drm_device *, - enum dcb_gpio_tag); extern struct dcb_connector_table_entry * nouveau_bios_connector_entry(struct drm_device *, int index); extern u32 get_pll_register(struct drm_device *, enum pll_types); @@ -1091,24 +1148,26 @@ extern int call_lvds_script(struct drm_device *, struct dcb_entry *, int head, enum LVDS_script, int pxclk); bool bios_encoder_match(struct dcb_entry *, u32 hash); +/* nouveau_mxm.c */ +int nouveau_mxm_init(struct drm_device *dev); +void nouveau_mxm_fini(struct drm_device *dev); + /* nouveau_ttm.c */ int nouveau_ttm_global_init(struct drm_nouveau_private *); void nouveau_ttm_global_release(struct drm_nouveau_private *); int nouveau_ttm_mmap(struct file *, struct vm_area_struct *); -/* nouveau_dp.c */ -int nouveau_dp_auxch(struct nouveau_i2c_chan *auxch, int cmd, int addr, - uint8_t *data, int data_nr); -bool nouveau_dp_detect(struct drm_encoder *); -bool nouveau_dp_link_train(struct drm_encoder *, u32 datarate); -void nouveau_dp_tu_update(struct drm_device *, int, int, u32, u32); -u8 *nouveau_dp_bios_data(struct drm_device *, struct dcb_entry *, u8 **); +/* nouveau_hdmi.c */ +void nouveau_hdmi_mode_set(struct drm_encoder *, struct drm_display_mode *); /* nv04_fb.c */ +extern int nv04_fb_vram_init(struct drm_device *); extern int nv04_fb_init(struct drm_device *); extern void nv04_fb_takedown(struct drm_device *); /* nv10_fb.c */ +extern int nv10_fb_vram_init(struct drm_device *dev); +extern int nv1a_fb_vram_init(struct drm_device *dev); extern int nv10_fb_init(struct drm_device *); extern void nv10_fb_takedown(struct drm_device *); extern void nv10_fb_init_tile_region(struct drm_device *dev, int i, @@ -1117,6 +1176,16 @@ extern void nv10_fb_init_tile_region(struct drm_device *dev, int i, extern void nv10_fb_set_tile_region(struct drm_device *dev, int i); extern void nv10_fb_free_tile_region(struct drm_device *dev, int i); +/* nv20_fb.c */ +extern int nv20_fb_vram_init(struct drm_device *dev); +extern int nv20_fb_init(struct drm_device *); +extern void nv20_fb_takedown(struct drm_device *); +extern void nv20_fb_init_tile_region(struct drm_device *dev, int i, + uint32_t addr, uint32_t size, + uint32_t pitch, uint32_t flags); +extern void nv20_fb_set_tile_region(struct drm_device *dev, int i); +extern void nv20_fb_free_tile_region(struct drm_device *dev, int i); + /* nv30_fb.c */ extern int nv30_fb_init(struct drm_device *); extern void nv30_fb_takedown(struct drm_device *); @@ -1126,6 +1195,7 @@ extern void nv30_fb_init_tile_region(struct drm_device *dev, int i, extern void nv30_fb_free_tile_region(struct drm_device *dev, int i); /* nv40_fb.c */ +extern int nv40_fb_vram_init(struct drm_device *dev); extern int nv40_fb_init(struct drm_device *); extern void nv40_fb_takedown(struct drm_device *); extern void nv40_fb_set_tile_region(struct drm_device *dev, int i); @@ -1222,6 +1292,9 @@ extern int nvc0_graph_isr_chid(struct drm_device *dev, u64 inst); /* nv84_crypt.c */ extern int nv84_crypt_create(struct drm_device *); +/* nv98_crypt.c */ +extern int nv98_crypt_create(struct drm_device *dev); + /* nva3_copy.c */ extern int nva3_copy_create(struct drm_device *dev); @@ -1234,6 +1307,17 @@ extern int nv31_mpeg_create(struct drm_device *dev); /* nv50_mpeg.c */ extern int nv50_mpeg_create(struct drm_device *dev); +/* nv84_bsp.c */ +/* nv98_bsp.c */ +extern int nv84_bsp_create(struct drm_device *dev); + +/* nv84_vp.c */ +/* nv98_vp.c */ +extern int nv84_vp_create(struct drm_device *dev); + +/* nv98_ppp.c */ +extern int nv98_ppp_create(struct drm_device *dev); + /* nv04_instmem.c */ extern int nv04_instmem_init(struct drm_device *); extern void nv04_instmem_takedown(struct drm_device *); @@ -1311,13 +1395,19 @@ extern int nv17_tv_create(struct drm_connector *, struct dcb_entry *); extern int nv04_display_early_init(struct drm_device *); extern void nv04_display_late_takedown(struct drm_device *); extern int nv04_display_create(struct drm_device *); -extern int nv04_display_init(struct drm_device *); extern void nv04_display_destroy(struct drm_device *); +extern int nv04_display_init(struct drm_device *); +extern void nv04_display_fini(struct drm_device *); /* nvd0_display.c */ extern int nvd0_display_create(struct drm_device *); -extern int nvd0_display_init(struct drm_device *); extern void nvd0_display_destroy(struct drm_device *); +extern int nvd0_display_init(struct drm_device *); +extern void nvd0_display_fini(struct drm_device *); +struct nouveau_bo *nvd0_display_crtc_sema(struct drm_device *, int crtc); +void nvd0_display_flip_stop(struct drm_crtc *); +int nvd0_display_flip_next(struct drm_crtc *, struct drm_framebuffer *, + struct nouveau_channel *, u32 swap_interval); /* nv04_crtc.c */ extern int nv04_crtc_create(struct drm_device *, int index); @@ -1412,6 +1502,10 @@ extern int nouveau_gem_ioctl_info(struct drm_device *, void *, struct drm_file *); /* nouveau_display.c */ +int nouveau_display_create(struct drm_device *dev); +void nouveau_display_destroy(struct drm_device *dev); +int nouveau_display_init(struct drm_device *dev); +void nouveau_display_fini(struct drm_device *dev); int nouveau_vblank_enable(struct drm_device *dev, int crtc); void nouveau_vblank_disable(struct drm_device *dev, int crtc); int nouveau_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb, @@ -1426,23 +1520,22 @@ int nouveau_display_dumb_destroy(struct drm_file *, struct drm_device *, uint32_t handle); /* nv10_gpio.c */ -int nv10_gpio_get(struct drm_device *dev, enum dcb_gpio_tag tag); -int nv10_gpio_set(struct drm_device *dev, enum dcb_gpio_tag tag, int state); +int nv10_gpio_init(struct drm_device *dev); +void nv10_gpio_fini(struct drm_device *dev); +int nv10_gpio_drive(struct drm_device *dev, int line, int dir, int out); +int nv10_gpio_sense(struct drm_device *dev, int line); +void nv10_gpio_irq_enable(struct drm_device *, int line, bool on); /* nv50_gpio.c */ int nv50_gpio_init(struct drm_device *dev); void nv50_gpio_fini(struct drm_device *dev); -int nv50_gpio_get(struct drm_device *dev, enum dcb_gpio_tag tag); -int nv50_gpio_set(struct drm_device *dev, enum dcb_gpio_tag tag, int state); -int nvd0_gpio_get(struct drm_device *dev, enum dcb_gpio_tag tag); -int nvd0_gpio_set(struct drm_device *dev, enum dcb_gpio_tag tag, int state); -int nv50_gpio_irq_register(struct drm_device *, enum dcb_gpio_tag, - void (*)(void *, int), void *); -void nv50_gpio_irq_unregister(struct drm_device *, enum dcb_gpio_tag, - void (*)(void *, int), void *); -bool nv50_gpio_irq_enable(struct drm_device *, enum dcb_gpio_tag, bool on); - -/* nv50_calc. */ +int nv50_gpio_drive(struct drm_device *dev, int line, int dir, int out); +int nv50_gpio_sense(struct drm_device *dev, int line); +void nv50_gpio_irq_enable(struct drm_device *, int line, bool on); +int nvd0_gpio_drive(struct drm_device *dev, int line, int dir, int out); +int nvd0_gpio_sense(struct drm_device *dev, int line); + +/* nv50_calc.c */ int nv50_calc_pll(struct drm_device *, struct pll_lims *, int clk, int *N1, int *M1, int *N2, int *M2, int *P); int nva3_calc_pll(struct drm_device *, struct pll_lims *, @@ -1565,6 +1658,13 @@ extern void nv_wo32(struct nouveau_gpuobj *, u32 offset, u32 val); #define NV_TRACEWARN(d, fmt, arg...) NV_PRINTK(KERN_NOTICE, d, fmt, ##arg) #define NV_TRACE(d, fmt, arg...) NV_PRINTK(KERN_INFO, d, fmt, ##arg) #define NV_WARN(d, fmt, arg...) NV_PRINTK(KERN_WARNING, d, fmt, ##arg) +#define NV_WARNONCE(d, fmt, arg...) do { \ + static int _warned = 0; \ + if (!_warned) { \ + NV_WARN(d, fmt, ##arg); \ + _warned = 1; \ + } \ +} while(0) /* nouveau_reg_debug bitmask */ enum { @@ -1652,6 +1752,7 @@ nv44_graph_class(struct drm_device *dev) #define NV_MEM_ACCESS_RW (NV_MEM_ACCESS_RO | NV_MEM_ACCESS_WO) #define NV_MEM_ACCESS_SYS 4 #define NV_MEM_ACCESS_VM 8 +#define NV_MEM_ACCESS_NOSNOOP 16 #define NV_MEM_TARGET_VRAM 0 #define NV_MEM_TARGET_PCI 1 @@ -1662,13 +1763,27 @@ nv44_graph_class(struct drm_device *dev) #define NV_MEM_TYPE_VM 0x7f #define NV_MEM_COMP_VM 0x03 +/* FIFO methods */ +#define NV01_SUBCHAN_OBJECT 0x00000000 +#define NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH 0x00000010 +#define NV84_SUBCHAN_SEMAPHORE_ADDRESS_LOW 0x00000014 +#define NV84_SUBCHAN_SEMAPHORE_SEQUENCE 0x00000018 +#define NV84_SUBCHAN_SEMAPHORE_TRIGGER 0x0000001c +#define NV84_SUBCHAN_SEMAPHORE_TRIGGER_ACQUIRE_EQUAL 0x00000001 +#define NV84_SUBCHAN_SEMAPHORE_TRIGGER_WRITE_LONG 0x00000002 +#define NV84_SUBCHAN_SEMAPHORE_TRIGGER_ACQUIRE_GEQUAL 0x00000004 +#define NV84_SUBCHAN_NOTIFY_INTR 0x00000020 +#define NV84_SUBCHAN_WRCACHE_FLUSH 0x00000024 +#define NV10_SUBCHAN_REF_CNT 0x00000050 +#define NVSW_SUBCHAN_PAGE_FLIP 0x00000054 +#define NV11_SUBCHAN_DMA_SEMAPHORE 0x00000060 +#define NV11_SUBCHAN_SEMAPHORE_OFFSET 0x00000064 +#define NV11_SUBCHAN_SEMAPHORE_ACQUIRE 0x00000068 +#define NV11_SUBCHAN_SEMAPHORE_RELEASE 0x0000006c +#define NV40_SUBCHAN_YIELD 0x00000080 + /* NV_SW object class */ #define NV_SW 0x0000506e -#define NV_SW_DMA_SEMAPHORE 0x00000060 -#define NV_SW_SEMAPHORE_OFFSET 0x00000064 -#define NV_SW_SEMAPHORE_ACQUIRE 0x00000068 -#define NV_SW_SEMAPHORE_RELEASE 0x0000006c -#define NV_SW_YIELD 0x00000080 #define NV_SW_DMA_VBLSEM 0x0000018c #define NV_SW_VBLSEM_OFFSET 0x00000400 #define NV_SW_VBLSEM_RELEASE_VALUE 0x00000404 diff --git a/drivers/gpu/drm/nouveau/nouveau_encoder.h b/drivers/gpu/drm/nouveau/nouveau_encoder.h index e5d6e3f..3dc14a3 100644 --- a/drivers/gpu/drm/nouveau/nouveau_encoder.h +++ b/drivers/gpu/drm/nouveau/nouveau_encoder.h @@ -32,6 +32,14 @@ #define NV_DPMS_CLEARED 0x80 +struct dp_train_func { + void (*link_set)(struct drm_device *, struct dcb_entry *, int crtc, + int nr, u32 bw, bool enhframe); + void (*train_set)(struct drm_device *, struct dcb_entry *, u8 pattern); + void (*train_adj)(struct drm_device *, struct dcb_entry *, + u8 lane, u8 swing, u8 preem); +}; + struct nouveau_encoder { struct drm_encoder_slave base; @@ -78,9 +86,19 @@ get_slave_funcs(struct drm_encoder *enc) return to_encoder_slave(enc)->slave_funcs; } +/* nouveau_dp.c */ +int nouveau_dp_auxch(struct nouveau_i2c_chan *auxch, int cmd, int addr, + uint8_t *data, int data_nr); +bool nouveau_dp_detect(struct drm_encoder *); +void nouveau_dp_dpms(struct drm_encoder *, int mode, u32 datarate, + struct dp_train_func *); +u8 *nouveau_dp_bios_data(struct drm_device *, struct dcb_entry *, u8 **); + struct nouveau_connector * nouveau_encoder_connector_get(struct nouveau_encoder *encoder); int nv50_sor_create(struct drm_connector *, struct dcb_entry *); +void nv50_sor_dp_calc_tu(struct drm_device *, int, int, u32, u32); int nv50_dac_create(struct drm_connector *, struct dcb_entry *); + #endif /* __NOUVEAU_ENCODER_H__ */ diff --git a/drivers/gpu/drm/nouveau/nouveau_fb.h b/drivers/gpu/drm/nouveau/nouveau_fb.h index 95c843e..f3fb649 100644 --- a/drivers/gpu/drm/nouveau/nouveau_fb.h +++ b/drivers/gpu/drm/nouveau/nouveau_fb.h @@ -42,8 +42,6 @@ nouveau_framebuffer(struct drm_framebuffer *fb) return container_of(fb, struct nouveau_framebuffer, base); } -extern const struct drm_mode_config_funcs nouveau_mode_config_funcs; - int nouveau_framebuffer_init(struct drm_device *dev, struct nouveau_framebuffer *nouveau_fb, - struct drm_mode_fb_cmd *mode_cmd, struct nouveau_bo *nvbo); + struct drm_mode_fb_cmd2 *mode_cmd, struct nouveau_bo *nvbo); #endif /* __NOUVEAU_FB_H__ */ diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.c b/drivers/gpu/drm/nouveau/nouveau_fbcon.c index cc0801d..6fd2211 100644 --- a/drivers/gpu/drm/nouveau/nouveau_fbcon.c +++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.c @@ -36,6 +36,7 @@ #include #include #include +#include #include "drmP.h" #include "drm.h" @@ -281,7 +282,7 @@ nouveau_fbcon_create(struct nouveau_fbdev *nfbdev, struct nouveau_framebuffer *nouveau_fb; struct nouveau_channel *chan; struct nouveau_bo *nvbo; - struct drm_mode_fb_cmd mode_cmd; + struct drm_mode_fb_cmd2 mode_cmd; struct pci_dev *pdev = dev->pdev; struct device *device = &pdev->dev; int size, ret; @@ -289,12 +290,13 @@ nouveau_fbcon_create(struct nouveau_fbdev *nfbdev, mode_cmd.width = sizes->surface_width; mode_cmd.height = sizes->surface_height; - mode_cmd.bpp = sizes->surface_bpp; - mode_cmd.pitch = mode_cmd.width * (mode_cmd.bpp >> 3); - mode_cmd.pitch = roundup(mode_cmd.pitch, 256); - mode_cmd.depth = sizes->surface_depth; + mode_cmd.pitches[0] = mode_cmd.width * (sizes->surface_bpp >> 3); + mode_cmd.pitches[0] = roundup(mode_cmd.pitches[0], 256); - size = mode_cmd.pitch * mode_cmd.height; + mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp, + sizes->surface_depth); + + size = mode_cmd.pitches[0] * mode_cmd.height; size = roundup(size, PAGE_SIZE); ret = nouveau_gem_new(dev, size, 0, NOUVEAU_GEM_DOMAIN_VRAM, @@ -369,7 +371,7 @@ nouveau_fbcon_create(struct nouveau_fbdev *nfbdev, info->screen_base = nvbo_kmap_obj_iovirtual(nouveau_fb->nvbo); info->screen_size = size; - drm_fb_helper_fill_fix(info, fb->pitch, fb->depth); + drm_fb_helper_fill_fix(info, fb->pitches[0], fb->depth); drm_fb_helper_fill_var(info, &nfbdev->helper, sizes->fb_width, sizes->fb_height); /* Set aperture base/size for vesafb takeover */ @@ -379,11 +381,7 @@ nouveau_fbcon_create(struct nouveau_fbdev *nfbdev, goto out_unref; } - info->pixmap.size = 64*1024; - info->pixmap.buf_align = 8; - info->pixmap.access_align = 32; - info->pixmap.flags = FB_PIXMAP_SYSTEM; - info->pixmap.scan_align = 1; + /* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */ mutex_unlock(&dev->struct_mutex); @@ -547,7 +545,13 @@ void nouveau_fbcon_restore_accel(struct drm_device *dev) void nouveau_fbcon_set_suspend(struct drm_device *dev, int state) { struct drm_nouveau_private *dev_priv = dev->dev_private; + console_lock(); + if (state == 0) + nouveau_fbcon_save_disable_accel(dev); fb_set_suspend(dev_priv->nfbdev->helper.fbdev, state); + if (state == 1) + nouveau_fbcon_restore_accel(dev); + console_unlock(); } void nouveau_fbcon_zfill_all(struct drm_device *dev) diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.c b/drivers/gpu/drm/nouveau/nouveau_fence.c index 2f6daae..c1dc20f 100644 --- a/drivers/gpu/drm/nouveau/nouveau_fence.c +++ b/drivers/gpu/drm/nouveau/nouveau_fence.c @@ -93,18 +93,17 @@ nouveau_fence_update(struct nouveau_channel *chan) } list_for_each_entry_safe(fence, tmp, &chan->fence.pending, entry) { - sequence = fence->sequence; + if (fence->sequence > chan->fence.sequence_ack) + break; + fence->signalled = true; list_del(&fence->entry); - - if (unlikely(fence->work)) + if (fence->work) fence->work(fence->priv, true); kref_put(&fence->refcount, nouveau_fence_del); - - if (sequence == chan->fence.sequence_ack) - break; } + out: spin_unlock(&chan->fence.lock); } @@ -165,9 +164,9 @@ nouveau_fence_emit(struct nouveau_fence *fence) if (USE_REFCNT(dev)) { if (dev_priv->card_type < NV_C0) - BEGIN_RING(chan, NvSubSw, 0x0050, 1); + BEGIN_RING(chan, 0, NV10_SUBCHAN_REF_CNT, 1); else - BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0050, 1); + BEGIN_NVC0(chan, 2, 0, NV10_SUBCHAN_REF_CNT, 1); } else { BEGIN_RING(chan, NvSubSw, 0x0150, 1); } @@ -344,7 +343,7 @@ semaphore_acquire(struct nouveau_channel *chan, struct nouveau_semaphore *sema) if (ret) return ret; - BEGIN_RING(chan, NvSubSw, NV_SW_DMA_SEMAPHORE, 3); + BEGIN_RING(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 3); OUT_RING (chan, NvSema); OUT_RING (chan, offset); OUT_RING (chan, 1); @@ -354,9 +353,9 @@ semaphore_acquire(struct nouveau_channel *chan, struct nouveau_semaphore *sema) if (ret) return ret; - BEGIN_RING(chan, NvSubSw, NV_SW_DMA_SEMAPHORE, 1); + BEGIN_RING(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 1); OUT_RING (chan, chan->vram_handle); - BEGIN_RING(chan, NvSubSw, 0x0010, 4); + BEGIN_RING(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4); OUT_RING (chan, upper_32_bits(offset)); OUT_RING (chan, lower_32_bits(offset)); OUT_RING (chan, 1); @@ -366,7 +365,7 @@ semaphore_acquire(struct nouveau_channel *chan, struct nouveau_semaphore *sema) if (ret) return ret; - BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0010, 4); + BEGIN_NVC0(chan, 2, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4); OUT_RING (chan, upper_32_bits(offset)); OUT_RING (chan, lower_32_bits(offset)); OUT_RING (chan, 1); @@ -397,10 +396,10 @@ semaphore_release(struct nouveau_channel *chan, struct nouveau_semaphore *sema) if (ret) return ret; - BEGIN_RING(chan, NvSubSw, NV_SW_DMA_SEMAPHORE, 2); + BEGIN_RING(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 2); OUT_RING (chan, NvSema); OUT_RING (chan, offset); - BEGIN_RING(chan, NvSubSw, NV_SW_SEMAPHORE_RELEASE, 1); + BEGIN_RING(chan, 0, NV11_SUBCHAN_SEMAPHORE_RELEASE, 1); OUT_RING (chan, 1); } else if (dev_priv->chipset < 0xc0) { @@ -408,9 +407,9 @@ semaphore_release(struct nouveau_channel *chan, struct nouveau_semaphore *sema) if (ret) return ret; - BEGIN_RING(chan, NvSubSw, NV_SW_DMA_SEMAPHORE, 1); + BEGIN_RING(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 1); OUT_RING (chan, chan->vram_handle); - BEGIN_RING(chan, NvSubSw, 0x0010, 4); + BEGIN_RING(chan, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4); OUT_RING (chan, upper_32_bits(offset)); OUT_RING (chan, lower_32_bits(offset)); OUT_RING (chan, 1); @@ -420,7 +419,7 @@ semaphore_release(struct nouveau_channel *chan, struct nouveau_semaphore *sema) if (ret) return ret; - BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0010, 4); + BEGIN_NVC0(chan, 2, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4); OUT_RING (chan, upper_32_bits(offset)); OUT_RING (chan, lower_32_bits(offset)); OUT_RING (chan, 1); @@ -510,7 +509,7 @@ nouveau_fence_channel_init(struct nouveau_channel *chan) if (ret) return ret; - BEGIN_RING(chan, NvSubSw, 0, 1); + BEGIN_RING(chan, NvSubSw, NV01_SUBCHAN_OBJECT, 1); OUT_RING (chan, NvSw); FIRE_RING (chan); } diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c index 7ce3fde..ed52a6f 100644 --- a/drivers/gpu/drm/nouveau/nouveau_gem.c +++ b/drivers/gpu/drm/nouveau/nouveau_gem.c @@ -426,9 +426,7 @@ validate_list(struct nouveau_channel *chan, struct list_head *list, return ret; } - nvbo->channel = (b->read_domains & (1 << 31)) ? NULL : chan; ret = nouveau_bo_validate(nvbo, true, false, false); - nvbo->channel = NULL; if (unlikely(ret)) { if (ret != -ERESTARTSYS) NV_ERROR(dev, "fail ttm_validate\n"); @@ -678,19 +676,13 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data, return PTR_ERR(bo); } - /* Mark push buffers as being used on PFIFO, the validation code - * will then make sure that if the pushbuf bo moves, that they - * happen on the kernel channel, which will in turn cause a sync - * to happen before we try and submit the push buffer. - */ + /* Ensure all push buffers are on validate list */ for (i = 0; i < req->nr_push; i++) { if (push[i].bo_index >= req->nr_buffers) { NV_ERROR(dev, "push %d buffer not in list\n", i); ret = -EINVAL; goto out_prevalid; } - - bo[push[i].bo_index].read_domains |= (1 << 31); } /* Validate buffer list */ diff --git a/drivers/gpu/drm/nouveau/nouveau_gpio.c b/drivers/gpu/drm/nouveau/nouveau_gpio.c new file mode 100644 index 0000000..a580cc6 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nouveau_gpio.c @@ -0,0 +1,400 @@ +/* + * Copyright 2011 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs + */ + +#include "drmP.h" +#include "nouveau_drv.h" +#include "nouveau_i2c.h" +#include "nouveau_gpio.h" + +static u8 * +dcb_gpio_table(struct drm_device *dev) +{ + u8 *dcb = dcb_table(dev); + if (dcb) { + if (dcb[0] >= 0x30 && dcb[1] >= 0x0c) + return ROMPTR(dev, dcb[0x0a]); + if (dcb[0] >= 0x22 && dcb[-1] >= 0x13) + return ROMPTR(dev, dcb[-15]); + } + return NULL; +} + +static u8 * +dcb_gpio_entry(struct drm_device *dev, int idx, int ent, u8 *version) +{ + u8 *table = dcb_gpio_table(dev); + if (table) { + *version = table[0]; + if (*version < 0x30 && ent < table[2]) + return table + 3 + (ent * table[1]); + else if (ent < table[2]) + return table + table[1] + (ent * table[3]); + } + return NULL; +} + +int +nouveau_gpio_drive(struct drm_device *dev, int idx, int line, int dir, int out) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; + + return pgpio->drive ? pgpio->drive(dev, line, dir, out) : -ENODEV; +} + +int +nouveau_gpio_sense(struct drm_device *dev, int idx, int line) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; + + return pgpio->sense ? pgpio->sense(dev, line) : -ENODEV; +} + +int +nouveau_gpio_find(struct drm_device *dev, int idx, u8 func, u8 line, + struct gpio_func *gpio) +{ + u8 *table, *entry, version; + int i = -1; + + if (line == 0xff && func == 0xff) + return -EINVAL; + + while ((entry = dcb_gpio_entry(dev, idx, ++i, &version))) { + if (version < 0x40) { + u16 data = ROM16(entry[0]); + *gpio = (struct gpio_func) { + .line = (data & 0x001f) >> 0, + .func = (data & 0x07e0) >> 5, + .log[0] = (data & 0x1800) >> 11, + .log[1] = (data & 0x6000) >> 13, + }; + } else + if (version < 0x41) { + *gpio = (struct gpio_func) { + .line = entry[0] & 0x1f, + .func = entry[1], + .log[0] = (entry[3] & 0x18) >> 3, + .log[1] = (entry[3] & 0x60) >> 5, + }; + } else { + *gpio = (struct gpio_func) { + .line = entry[0] & 0x3f, + .func = entry[1], + .log[0] = (entry[4] & 0x30) >> 4, + .log[1] = (entry[4] & 0xc0) >> 6, + }; + } + + if ((line == 0xff || line == gpio->line) && + (func == 0xff || func == gpio->func)) + return 0; + } + + /* DCB 2.2, fixed TVDAC GPIO data */ + if ((table = dcb_table(dev)) && table[0] >= 0x22) { + if (func == DCB_GPIO_TVDAC0) { + *gpio = (struct gpio_func) { + .func = DCB_GPIO_TVDAC0, + .line = table[-4] >> 4, + .log[0] = !!(table[-5] & 2), + .log[1] = !(table[-5] & 2), + }; + return 0; + } + } + + /* Apple iMac G4 NV18 */ + if (nv_match_device(dev, 0x0189, 0x10de, 0x0010)) { + if (func == DCB_GPIO_TVDAC0) { + *gpio = (struct gpio_func) { + .func = DCB_GPIO_TVDAC0, + .line = 4, + .log[0] = 0, + .log[1] = 1, + }; + return 0; + } + } + + return -EINVAL; +} + +int +nouveau_gpio_set(struct drm_device *dev, int idx, u8 tag, u8 line, int state) +{ + struct gpio_func gpio; + int ret; + + ret = nouveau_gpio_find(dev, idx, tag, line, &gpio); + if (ret == 0) { + int dir = !!(gpio.log[state] & 0x02); + int out = !!(gpio.log[state] & 0x01); + ret = nouveau_gpio_drive(dev, idx, gpio.line, dir, out); + } + + return ret; +} + +int +nouveau_gpio_get(struct drm_device *dev, int idx, u8 tag, u8 line) +{ + struct gpio_func gpio; + int ret; + + ret = nouveau_gpio_find(dev, idx, tag, line, &gpio); + if (ret == 0) { + ret = nouveau_gpio_sense(dev, idx, gpio.line); + if (ret >= 0) + ret = (ret == (gpio.log[1] & 1)); + } + + return ret; +} + +int +nouveau_gpio_irq(struct drm_device *dev, int idx, u8 tag, u8 line, bool on) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; + struct gpio_func gpio; + int ret; + + ret = nouveau_gpio_find(dev, idx, tag, line, &gpio); + if (ret == 0) { + if (idx == 0 && pgpio->irq_enable) + pgpio->irq_enable(dev, gpio.line, on); + else + ret = -ENODEV; + } + + return ret; +} + +struct gpio_isr { + struct drm_device *dev; + struct list_head head; + struct work_struct work; + int idx; + struct gpio_func func; + void (*handler)(void *, int); + void *data; + bool inhibit; +}; + +static void +nouveau_gpio_isr_bh(struct work_struct *work) +{ + struct gpio_isr *isr = container_of(work, struct gpio_isr, work); + struct drm_device *dev = isr->dev; + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; + unsigned long flags; + int state; + + state = nouveau_gpio_get(dev, isr->idx, isr->func.func, isr->func.line); + if (state >= 0) + isr->handler(isr->data, state); + + spin_lock_irqsave(&pgpio->lock, flags); + isr->inhibit = false; + spin_unlock_irqrestore(&pgpio->lock, flags); +} + +void +nouveau_gpio_isr(struct drm_device *dev, int idx, u32 line_mask) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; + struct gpio_isr *isr; + + if (idx != 0) + return; + + spin_lock(&pgpio->lock); + list_for_each_entry(isr, &pgpio->isr, head) { + if (line_mask & (1 << isr->func.line)) { + if (isr->inhibit) + continue; + isr->inhibit = true; + schedule_work(&isr->work); + } + } + spin_unlock(&pgpio->lock); +} + +int +nouveau_gpio_isr_add(struct drm_device *dev, int idx, u8 tag, u8 line, + void (*handler)(void *, int), void *data) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; + struct gpio_isr *isr; + unsigned long flags; + int ret; + + isr = kzalloc(sizeof(*isr), GFP_KERNEL); + if (!isr) + return -ENOMEM; + + ret = nouveau_gpio_find(dev, idx, tag, line, &isr->func); + if (ret) { + kfree(isr); + return ret; + } + + INIT_WORK(&isr->work, nouveau_gpio_isr_bh); + isr->dev = dev; + isr->handler = handler; + isr->data = data; + isr->idx = idx; + + spin_lock_irqsave(&pgpio->lock, flags); + list_add(&isr->head, &pgpio->isr); + spin_unlock_irqrestore(&pgpio->lock, flags); + return 0; +} + +void +nouveau_gpio_isr_del(struct drm_device *dev, int idx, u8 tag, u8 line, + void (*handler)(void *, int), void *data) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; + struct gpio_isr *isr, *tmp; + struct gpio_func func; + unsigned long flags; + LIST_HEAD(tofree); + int ret; + + ret = nouveau_gpio_find(dev, idx, tag, line, &func); + if (ret == 0) { + spin_lock_irqsave(&pgpio->lock, flags); + list_for_each_entry_safe(isr, tmp, &pgpio->isr, head) { + if (memcmp(&isr->func, &func, sizeof(func)) || + isr->idx != idx || + isr->handler != handler || isr->data != data) + continue; + list_move(&isr->head, &tofree); + } + spin_unlock_irqrestore(&pgpio->lock, flags); + + list_for_each_entry_safe(isr, tmp, &tofree, head) { + flush_work_sync(&isr->work); + kfree(isr); + } + } +} + +int +nouveau_gpio_create(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; + + INIT_LIST_HEAD(&pgpio->isr); + spin_lock_init(&pgpio->lock); + + return nouveau_gpio_init(dev); +} + +void +nouveau_gpio_destroy(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; + + nouveau_gpio_fini(dev); + BUG_ON(!list_empty(&pgpio->isr)); +} + +int +nouveau_gpio_init(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; + int ret = 0; + + if (pgpio->init) + ret = pgpio->init(dev); + + return ret; +} + +void +nouveau_gpio_fini(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; + + if (pgpio->fini) + pgpio->fini(dev); +} + +void +nouveau_gpio_reset(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + u8 *entry, version; + int ent = -1; + + while ((entry = dcb_gpio_entry(dev, 0, ++ent, &version))) { + u8 func = 0xff, line, defs, unk0, unk1; + if (version >= 0x41) { + defs = !!(entry[0] & 0x80); + line = entry[0] & 0x3f; + func = entry[1]; + unk0 = entry[2]; + unk1 = entry[3] & 0x1f; + } else + if (version >= 0x40) { + line = entry[0] & 0x1f; + func = entry[1]; + defs = !!(entry[3] & 0x01); + unk0 = !!(entry[3] & 0x02); + unk1 = !!(entry[3] & 0x04); + } else { + break; + } + + if (func == 0xff) + continue; + + nouveau_gpio_func_set(dev, func, defs); + + if (dev_priv->card_type >= NV_D0) { + nv_mask(dev, 0x00d610 + (line * 4), 0xff, unk0); + if (unk1--) + nv_mask(dev, 0x00d640 + (unk1 * 4), 0xff, line); + } else + if (dev_priv->card_type >= NV_50) { + static const u32 regs[] = { 0xe100, 0xe28c }; + u32 val = (unk1 << 16) | unk0; + u32 reg = regs[line >> 4]; line &= 0x0f; + + nv_mask(dev, reg, 0x00010001 << line, val << line); + } + } +} diff --git a/drivers/gpu/drm/nouveau/nouveau_gpio.h b/drivers/gpu/drm/nouveau/nouveau_gpio.h new file mode 100644 index 0000000..64c5cb0 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nouveau_gpio.h @@ -0,0 +1,71 @@ +/* + * Copyright 2011 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ + +#ifndef __NOUVEAU_GPIO_H__ +#define __NOUVEAU_GPIO_H__ + +struct gpio_func { + u8 func; + u8 line; + u8 log[2]; +}; + +/* nouveau_gpio.c */ +int nouveau_gpio_create(struct drm_device *); +void nouveau_gpio_destroy(struct drm_device *); +int nouveau_gpio_init(struct drm_device *); +void nouveau_gpio_fini(struct drm_device *); +void nouveau_gpio_reset(struct drm_device *); +int nouveau_gpio_drive(struct drm_device *, int idx, int line, + int dir, int out); +int nouveau_gpio_sense(struct drm_device *, int idx, int line); +int nouveau_gpio_find(struct drm_device *, int idx, u8 tag, u8 line, + struct gpio_func *); +int nouveau_gpio_set(struct drm_device *, int idx, u8 tag, u8 line, int state); +int nouveau_gpio_get(struct drm_device *, int idx, u8 tag, u8 line); +int nouveau_gpio_irq(struct drm_device *, int idx, u8 tag, u8 line, bool on); +void nouveau_gpio_isr(struct drm_device *, int idx, u32 mask); +int nouveau_gpio_isr_add(struct drm_device *, int idx, u8 tag, u8 line, + void (*)(void *, int state), void *data); +void nouveau_gpio_isr_del(struct drm_device *, int idx, u8 tag, u8 line, + void (*)(void *, int state), void *data); + +static inline bool +nouveau_gpio_func_valid(struct drm_device *dev, u8 tag) +{ + struct gpio_func func; + return (nouveau_gpio_find(dev, 0, tag, 0xff, &func)) == 0; +} + +static inline int +nouveau_gpio_func_set(struct drm_device *dev, u8 tag, int state) +{ + return nouveau_gpio_set(dev, 0, tag, 0xff, state); +} + +static inline int +nouveau_gpio_func_get(struct drm_device *dev, u8 tag) +{ + return nouveau_gpio_get(dev, 0, tag, 0xff); +} + +#endif diff --git a/drivers/gpu/drm/nouveau/nouveau_hdmi.c b/drivers/gpu/drm/nouveau/nouveau_hdmi.c new file mode 100644 index 0000000..c3de363 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nouveau_hdmi.c @@ -0,0 +1,260 @@ +/* + * Copyright 2011 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs + */ + +#include "drmP.h" +#include "nouveau_drv.h" +#include "nouveau_connector.h" +#include "nouveau_encoder.h" +#include "nouveau_crtc.h" + +static bool +hdmi_sor(struct drm_encoder *encoder) +{ + struct drm_nouveau_private *dev_priv = encoder->dev->dev_private; + if (dev_priv->chipset < 0xa3 || + dev_priv->chipset == 0xaa || + dev_priv->chipset == 0xac) + return false; + return true; +} + +static inline u32 +hdmi_base(struct drm_encoder *encoder) +{ + struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); + struct nouveau_crtc *nv_crtc = nouveau_crtc(nv_encoder->crtc); + if (!hdmi_sor(encoder)) + return 0x616500 + (nv_crtc->index * 0x800); + return 0x61c500 + (nv_encoder->or * 0x800); +} + +static void +hdmi_wr32(struct drm_encoder *encoder, u32 reg, u32 val) +{ + nv_wr32(encoder->dev, hdmi_base(encoder) + reg, val); +} + +static u32 +hdmi_rd32(struct drm_encoder *encoder, u32 reg) +{ + return nv_rd32(encoder->dev, hdmi_base(encoder) + reg); +} + +static u32 +hdmi_mask(struct drm_encoder *encoder, u32 reg, u32 mask, u32 val) +{ + u32 tmp = hdmi_rd32(encoder, reg); + hdmi_wr32(encoder, reg, (tmp & ~mask) | val); + return tmp; +} + +static void +nouveau_audio_disconnect(struct drm_encoder *encoder) +{ + struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); + struct drm_device *dev = encoder->dev; + u32 or = nv_encoder->or * 0x800; + + if (hdmi_sor(encoder)) { + nv_mask(dev, 0x61c448 + or, 0x00000003, 0x00000000); + } +} + +static void +nouveau_audio_mode_set(struct drm_encoder *encoder, + struct drm_display_mode *mode) +{ + struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); + struct nouveau_connector *nv_connector; + struct drm_device *dev = encoder->dev; + u32 or = nv_encoder->or * 0x800; + int i; + + nv_connector = nouveau_encoder_connector_get(nv_encoder); + if (!drm_detect_monitor_audio(nv_connector->edid)) { + nouveau_audio_disconnect(encoder); + return; + } + + if (hdmi_sor(encoder)) { + nv_mask(dev, 0x61c448 + or, 0x00000001, 0x00000001); + + drm_edid_to_eld(&nv_connector->base, nv_connector->edid); + if (nv_connector->base.eld[0]) { + u8 *eld = nv_connector->base.eld; + for (i = 0; i < eld[2] * 4; i++) + nv_wr32(dev, 0x61c440 + or, (i << 8) | eld[i]); + for (i = eld[2] * 4; i < 0x60; i++) + nv_wr32(dev, 0x61c440 + or, (i << 8) | 0x00); + nv_mask(dev, 0x61c448 + or, 0x00000002, 0x00000002); + } + } +} + +static void +nouveau_hdmi_infoframe(struct drm_encoder *encoder, u32 ctrl, u8 *frame) +{ + /* calculate checksum for the infoframe */ + u8 sum = 0, i; + for (i = 0; i < frame[2]; i++) + sum += frame[i]; + frame[3] = 256 - sum; + + /* disable infoframe, and write header */ + hdmi_mask(encoder, ctrl + 0x00, 0x00000001, 0x00000000); + hdmi_wr32(encoder, ctrl + 0x08, *(u32 *)frame & 0xffffff); + + /* register scans tell me the audio infoframe has only one set of + * subpack regs, according to tegra (gee nvidia, it'd be nice if we + * could get those docs too!), the hdmi block pads out the rest of + * the packet on its own. + */ + if (ctrl == 0x020) + frame[2] = 6; + + /* write out checksum and data, weird weird 7 byte register pairs */ + for (i = 0; i < frame[2] + 1; i += 7) { + u32 rsubpack = ctrl + 0x0c + ((i / 7) * 8); + u32 *subpack = (u32 *)&frame[3 + i]; + hdmi_wr32(encoder, rsubpack + 0, subpack[0]); + hdmi_wr32(encoder, rsubpack + 4, subpack[1] & 0xffffff); + } + + /* enable the infoframe */ + hdmi_mask(encoder, ctrl, 0x00000001, 0x00000001); +} + +static void +nouveau_hdmi_video_infoframe(struct drm_encoder *encoder, + struct drm_display_mode *mode) +{ + const u8 Y = 0, A = 0, B = 0, S = 0, C = 0, M = 0, R = 0; + const u8 ITC = 0, EC = 0, Q = 0, SC = 0, VIC = 0, PR = 0; + const u8 bar_top = 0, bar_bottom = 0, bar_left = 0, bar_right = 0; + u8 frame[20]; + + frame[0x00] = 0x82; /* AVI infoframe */ + frame[0x01] = 0x02; /* version */ + frame[0x02] = 0x0d; /* length */ + frame[0x03] = 0x00; + frame[0x04] = (Y << 5) | (A << 4) | (B << 2) | S; + frame[0x05] = (C << 6) | (M << 4) | R; + frame[0x06] = (ITC << 7) | (EC << 4) | (Q << 2) | SC; + frame[0x07] = VIC; + frame[0x08] = PR; + frame[0x09] = bar_top & 0xff; + frame[0x0a] = bar_top >> 8; + frame[0x0b] = bar_bottom & 0xff; + frame[0x0c] = bar_bottom >> 8; + frame[0x0d] = bar_left & 0xff; + frame[0x0e] = bar_left >> 8; + frame[0x0f] = bar_right & 0xff; + frame[0x10] = bar_right >> 8; + frame[0x11] = 0x00; + frame[0x12] = 0x00; + frame[0x13] = 0x00; + + nouveau_hdmi_infoframe(encoder, 0x020, frame); +} + +static void +nouveau_hdmi_audio_infoframe(struct drm_encoder *encoder, + struct drm_display_mode *mode) +{ + const u8 CT = 0x00, CC = 0x01, ceaSS = 0x00, SF = 0x00, FMT = 0x00; + const u8 CA = 0x00, DM_INH = 0, LSV = 0x00; + u8 frame[12]; + + frame[0x00] = 0x84; /* Audio infoframe */ + frame[0x01] = 0x01; /* version */ + frame[0x02] = 0x0a; /* length */ + frame[0x03] = 0x00; + frame[0x04] = (CT << 4) | CC; + frame[0x05] = (SF << 2) | ceaSS; + frame[0x06] = FMT; + frame[0x07] = CA; + frame[0x08] = (DM_INH << 7) | (LSV << 3); + frame[0x09] = 0x00; + frame[0x0a] = 0x00; + frame[0x0b] = 0x00; + + nouveau_hdmi_infoframe(encoder, 0x000, frame); +} + +static void +nouveau_hdmi_disconnect(struct drm_encoder *encoder) +{ + nouveau_audio_disconnect(encoder); + + /* disable audio and avi infoframes */ + hdmi_mask(encoder, 0x000, 0x00000001, 0x00000000); + hdmi_mask(encoder, 0x020, 0x00000001, 0x00000000); + + /* disable hdmi */ + hdmi_mask(encoder, 0x0a4, 0x40000000, 0x00000000); +} + +void +nouveau_hdmi_mode_set(struct drm_encoder *encoder, + struct drm_display_mode *mode) +{ + struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); + struct nouveau_connector *nv_connector; + struct drm_device *dev = encoder->dev; + u32 max_ac_packet, rekey; + + nv_connector = nouveau_encoder_connector_get(nv_encoder); + if (!mode || !nv_connector || !nv_connector->edid || + !drm_detect_hdmi_monitor(nv_connector->edid)) { + nouveau_hdmi_disconnect(encoder); + return; + } + + nouveau_hdmi_video_infoframe(encoder, mode); + nouveau_hdmi_audio_infoframe(encoder, mode); + + hdmi_mask(encoder, 0x0d0, 0x00070001, 0x00010001); /* SPARE, HW_CTS */ + hdmi_mask(encoder, 0x068, 0x00010101, 0x00000000); /* ACR_CTRL, ?? */ + hdmi_mask(encoder, 0x078, 0x80000000, 0x80000000); /* ACR_0441_ENABLE */ + + nv_mask(dev, 0x61733c, 0x00100000, 0x00100000); /* RESETF */ + nv_mask(dev, 0x61733c, 0x10000000, 0x10000000); /* LOOKUP_EN */ + nv_mask(dev, 0x61733c, 0x00100000, 0x00000000); /* !RESETF */ + + /* value matches nvidia binary driver, and tegra constant */ + rekey = 56; + + max_ac_packet = mode->htotal - mode->hdisplay; + max_ac_packet -= rekey; + max_ac_packet -= 18; /* constant from tegra */ + max_ac_packet /= 32; + + /* enable hdmi */ + hdmi_mask(encoder, 0x0a4, 0x5f1f003f, 0x40000000 | /* enable */ + 0x1f000000 | /* unknown */ + max_ac_packet << 16 | + rekey); + + nouveau_audio_mode_set(encoder, mode); +} diff --git a/drivers/gpu/drm/nouveau/nouveau_hwsq.h b/drivers/gpu/drm/nouveau/nouveau_hwsq.h new file mode 100644 index 0000000..6976875 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nouveau_hwsq.h @@ -0,0 +1,115 @@ +/* + * Copyright 2010 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs + */ + +#ifndef __NOUVEAU_HWSQ_H__ +#define __NOUVEAU_HWSQ_H__ + +struct hwsq_ucode { + u8 data[0x200]; + union { + u8 *u08; + u16 *u16; + u32 *u32; + } ptr; + u16 len; + + u32 reg; + u32 val; +}; + +static inline void +hwsq_init(struct hwsq_ucode *hwsq) +{ + hwsq->ptr.u08 = hwsq->data; + hwsq->reg = 0xffffffff; + hwsq->val = 0xffffffff; +} + +static inline void +hwsq_fini(struct hwsq_ucode *hwsq) +{ + do { + *hwsq->ptr.u08++ = 0x7f; + hwsq->len = hwsq->ptr.u08 - hwsq->data; + } while (hwsq->len & 3); + hwsq->ptr.u08 = hwsq->data; +} + +static inline void +hwsq_usec(struct hwsq_ucode *hwsq, u8 usec) +{ + u32 shift = 0; + while (usec & ~3) { + usec >>= 2; + shift++; + } + + *hwsq->ptr.u08++ = (shift << 2) | usec; +} + +static inline void +hwsq_setf(struct hwsq_ucode *hwsq, u8 flag, int val) +{ + flag += 0x80; + if (val >= 0) + flag += 0x20; + if (val >= 1) + flag += 0x20; + *hwsq->ptr.u08++ = flag; +} + +static inline void +hwsq_op5f(struct hwsq_ucode *hwsq, u8 v0, u8 v1) +{ + *hwsq->ptr.u08++ = 0x5f; + *hwsq->ptr.u08++ = v0; + *hwsq->ptr.u08++ = v1; +} + +static inline void +hwsq_wr32(struct hwsq_ucode *hwsq, u32 reg, u32 val) +{ + if (val != hwsq->val) { + if ((val & 0xffff0000) == (hwsq->val & 0xffff0000)) { + *hwsq->ptr.u08++ = 0x42; + *hwsq->ptr.u16++ = (val & 0x0000ffff); + } else { + *hwsq->ptr.u08++ = 0xe2; + *hwsq->ptr.u32++ = val; + } + + hwsq->val = val; + } + + if ((reg & 0xffff0000) == (hwsq->reg & 0xffff0000)) { + *hwsq->ptr.u08++ = 0x40; + *hwsq->ptr.u16++ = (reg & 0x0000ffff); + } else { + *hwsq->ptr.u08++ = 0xe0; + *hwsq->ptr.u32++ = reg; + } + hwsq->reg = reg; +} + +#endif diff --git a/drivers/gpu/drm/nouveau/nouveau_i2c.c b/drivers/gpu/drm/nouveau/nouveau_i2c.c index d39b220..77e5646 100644 --- a/drivers/gpu/drm/nouveau/nouveau_i2c.c +++ b/drivers/gpu/drm/nouveau/nouveau_i2c.c @@ -30,83 +30,83 @@ #include "nouveau_hw.h" static void -nv04_i2c_setscl(void *data, int state) +i2c_drive_scl(void *data, int state) { - struct nouveau_i2c_chan *i2c = data; - struct drm_device *dev = i2c->dev; - uint8_t val; - - val = (NVReadVgaCrtc(dev, 0, i2c->wr) & 0xd0) | (state ? 0x20 : 0); - NVWriteVgaCrtc(dev, 0, i2c->wr, val | 0x01); -} - -static void -nv04_i2c_setsda(void *data, int state) -{ - struct nouveau_i2c_chan *i2c = data; - struct drm_device *dev = i2c->dev; - uint8_t val; - - val = (NVReadVgaCrtc(dev, 0, i2c->wr) & 0xe0) | (state ? 0x10 : 0); - NVWriteVgaCrtc(dev, 0, i2c->wr, val | 0x01); -} - -static int -nv04_i2c_getscl(void *data) -{ - struct nouveau_i2c_chan *i2c = data; - struct drm_device *dev = i2c->dev; - - return !!(NVReadVgaCrtc(dev, 0, i2c->rd) & 4); -} - -static int -nv04_i2c_getsda(void *data) -{ - struct nouveau_i2c_chan *i2c = data; - struct drm_device *dev = i2c->dev; - - return !!(NVReadVgaCrtc(dev, 0, i2c->rd) & 8); -} - -static void -nv4e_i2c_setscl(void *data, int state) -{ - struct nouveau_i2c_chan *i2c = data; - struct drm_device *dev = i2c->dev; - uint8_t val; - - val = (nv_rd32(dev, i2c->wr) & 0xd0) | (state ? 0x20 : 0); - nv_wr32(dev, i2c->wr, val | 0x01); + struct nouveau_i2c_chan *port = data; + if (port->type == 0) { + u8 val = NVReadVgaCrtc(port->dev, 0, port->drive); + if (state) val |= 0x20; + else val &= 0xdf; + NVWriteVgaCrtc(port->dev, 0, port->drive, val | 0x01); + } else + if (port->type == 4) { + nv_mask(port->dev, port->drive, 0x2f, state ? 0x21 : 0x01); + } else + if (port->type == 5) { + if (state) port->state |= 0x01; + else port->state &= 0xfe; + nv_wr32(port->dev, port->drive, 4 | port->state); + } } static void -nv4e_i2c_setsda(void *data, int state) +i2c_drive_sda(void *data, int state) { - struct nouveau_i2c_chan *i2c = data; - struct drm_device *dev = i2c->dev; - uint8_t val; - - val = (nv_rd32(dev, i2c->wr) & 0xe0) | (state ? 0x10 : 0); - nv_wr32(dev, i2c->wr, val | 0x01); + struct nouveau_i2c_chan *port = data; + if (port->type == 0) { + u8 val = NVReadVgaCrtc(port->dev, 0, port->drive); + if (state) val |= 0x10; + else val &= 0xef; + NVWriteVgaCrtc(port->dev, 0, port->drive, val | 0x01); + } else + if (port->type == 4) { + nv_mask(port->dev, port->drive, 0x1f, state ? 0x11 : 0x01); + } else + if (port->type == 5) { + if (state) port->state |= 0x02; + else port->state &= 0xfd; + nv_wr32(port->dev, port->drive, 4 | port->state); + } } static int -nv4e_i2c_getscl(void *data) +i2c_sense_scl(void *data) { - struct nouveau_i2c_chan *i2c = data; - struct drm_device *dev = i2c->dev; - - return !!((nv_rd32(dev, i2c->rd) >> 16) & 4); + struct nouveau_i2c_chan *port = data; + struct drm_nouveau_private *dev_priv = port->dev->dev_private; + if (port->type == 0) { + return !!(NVReadVgaCrtc(port->dev, 0, port->sense) & 0x04); + } else + if (port->type == 4) { + return !!(nv_rd32(port->dev, port->sense) & 0x00040000); + } else + if (port->type == 5) { + if (dev_priv->card_type < NV_D0) + return !!(nv_rd32(port->dev, port->sense) & 0x01); + else + return !!(nv_rd32(port->dev, port->sense) & 0x10); + } + return 0; } static int -nv4e_i2c_getsda(void *data) +i2c_sense_sda(void *data) { - struct nouveau_i2c_chan *i2c = data; - struct drm_device *dev = i2c->dev; - - return !!((nv_rd32(dev, i2c->rd) >> 16) & 8); + struct nouveau_i2c_chan *port = data; + struct drm_nouveau_private *dev_priv = port->dev->dev_private; + if (port->type == 0) { + return !!(NVReadVgaCrtc(port->dev, 0, port->sense) & 0x08); + } else + if (port->type == 4) { + return !!(nv_rd32(port->dev, port->sense) & 0x00080000); + } else + if (port->type == 5) { + if (dev_priv->card_type < NV_D0) + return !!(nv_rd32(port->dev, port->sense) & 0x02); + else + return !!(nv_rd32(port->dev, port->sense) & 0x20); + } + return 0; } static const uint32_t nv50_i2c_port[] = { @@ -114,177 +114,223 @@ static const uint32_t nv50_i2c_port[] = { 0x00e254, 0x00e274, 0x00e764, 0x00e780, 0x00e79c, 0x00e7b8 }; -#define NV50_I2C_PORTS ARRAY_SIZE(nv50_i2c_port) -static int -nv50_i2c_getscl(void *data) -{ - struct nouveau_i2c_chan *i2c = data; - struct drm_device *dev = i2c->dev; - - return !!(nv_rd32(dev, i2c->rd) & 1); -} - - -static int -nv50_i2c_getsda(void *data) +static u8 * +i2c_table(struct drm_device *dev, u8 *version) { - struct nouveau_i2c_chan *i2c = data; - struct drm_device *dev = i2c->dev; - - return !!(nv_rd32(dev, i2c->rd) & 2); -} + u8 *dcb = dcb_table(dev), *i2c = NULL; + if (dcb) { + if (dcb[0] >= 0x15) + i2c = ROMPTR(dev, dcb[2]); + if (dcb[0] >= 0x30) + i2c = ROMPTR(dev, dcb[4]); + } -static void -nv50_i2c_setscl(void *data, int state) -{ - struct nouveau_i2c_chan *i2c = data; + /* early revisions had no version number, use dcb version */ + if (i2c) { + *version = dcb[0]; + if (*version >= 0x30) + *version = i2c[0]; + } - nv_wr32(i2c->dev, i2c->wr, 4 | (i2c->data ? 2 : 0) | (state ? 1 : 0)); + return i2c; } -static void -nv50_i2c_setsda(void *data, int state) +int +nouveau_i2c_init(struct drm_device *dev) { - struct nouveau_i2c_chan *i2c = data; + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nvbios *bios = &dev_priv->vbios; + struct nouveau_i2c_chan *port; + u8 version = 0x00, entries, recordlen; + u8 *i2c, *entry, legacy[2][4] = {}; + int ret, i; + + INIT_LIST_HEAD(&dev_priv->i2c_ports); + + i2c = i2c_table(dev, &version); + if (!i2c) { + u8 *bmp = &bios->data[bios->offset]; + if (bios->type != NVBIOS_BMP) + return -ENODEV; + + legacy[0][0] = NV_CIO_CRE_DDC_WR__INDEX; + legacy[0][1] = NV_CIO_CRE_DDC_STATUS__INDEX; + legacy[1][0] = NV_CIO_CRE_DDC0_WR__INDEX; + legacy[1][1] = NV_CIO_CRE_DDC0_STATUS__INDEX; + + /* BMP (from v4.0) has i2c info in the structure, it's in a + * fixed location on earlier VBIOS + */ + if (bmp[5] < 4) + i2c = &bios->data[0x48]; + else + i2c = &bmp[0x36]; + + if (i2c[4]) legacy[0][0] = i2c[4]; + if (i2c[5]) legacy[0][1] = i2c[5]; + if (i2c[6]) legacy[1][0] = i2c[6]; + if (i2c[7]) legacy[1][1] = i2c[7]; + } - nv_mask(i2c->dev, i2c->wr, 0x00000006, 4 | (state ? 2 : 0)); - i2c->data = state; -} + if (version >= 0x30) { + entry = i2c[1] + i2c; + entries = i2c[2]; + recordlen = i2c[3]; + } else + if (version) { + entry = i2c; + entries = 16; + recordlen = 4; + } else { + entry = legacy[0]; + entries = 2; + recordlen = 4; + } -static int -nvd0_i2c_getscl(void *data) -{ - struct nouveau_i2c_chan *i2c = data; - return !!(nv_rd32(i2c->dev, i2c->rd) & 0x10); -} + for (i = 0; i < entries; i++, entry += recordlen) { + port = kzalloc(sizeof(*port), GFP_KERNEL); + if (port == NULL) { + nouveau_i2c_fini(dev); + return -ENOMEM; + } -static int -nvd0_i2c_getsda(void *data) -{ - struct nouveau_i2c_chan *i2c = data; - return !!(nv_rd32(i2c->dev, i2c->rd) & 0x20); -} + port->type = entry[3]; + if (version < 0x30) { + port->type &= 0x07; + if (port->type == 0x07) + port->type = 0xff; + } -int -nouveau_i2c_init(struct drm_device *dev, struct dcb_i2c_entry *entry, int index) -{ - struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_i2c_chan *i2c; - int ret; + if (port->type == 0xff) { + kfree(port); + continue; + } - if (entry->chan) - return -EEXIST; + switch (port->type) { + case 0: /* NV04:NV50 */ + port->drive = entry[0]; + port->sense = entry[1]; + break; + case 4: /* NV4E */ + port->drive = 0x600800 + entry[1]; + port->sense = port->drive; + break; + case 5: /* NV50- */ + port->drive = entry[0] & 0x0f; + if (dev_priv->card_type < NV_D0) { + if (port->drive >= ARRAY_SIZE(nv50_i2c_port)) + break; + port->drive = nv50_i2c_port[port->drive]; + port->sense = port->drive; + } else { + port->drive = 0x00d014 + (port->drive * 0x20); + port->sense = port->drive; + } + break; + case 6: /* NV50- DP AUX */ + port->drive = entry[0]; + port->sense = port->drive; + port->adapter.algo = &nouveau_dp_i2c_algo; + break; + default: + break; + } - if (dev_priv->card_type >= NV_50 && - dev_priv->card_type <= NV_C0 && entry->read >= NV50_I2C_PORTS) { - NV_ERROR(dev, "unknown i2c port %d\n", entry->read); - return -EINVAL; - } + if (!port->adapter.algo && !port->drive) { + NV_ERROR(dev, "I2C%d: type %d index %x/%x unknown\n", + i, port->type, port->drive, port->sense); + kfree(port); + continue; + } - i2c = kzalloc(sizeof(*i2c), GFP_KERNEL); - if (i2c == NULL) - return -ENOMEM; - - switch (entry->port_type) { - case 0: - i2c->bit.setsda = nv04_i2c_setsda; - i2c->bit.setscl = nv04_i2c_setscl; - i2c->bit.getsda = nv04_i2c_getsda; - i2c->bit.getscl = nv04_i2c_getscl; - i2c->rd = entry->read; - i2c->wr = entry->write; - break; - case 4: - i2c->bit.setsda = nv4e_i2c_setsda; - i2c->bit.setscl = nv4e_i2c_setscl; - i2c->bit.getsda = nv4e_i2c_getsda; - i2c->bit.getscl = nv4e_i2c_getscl; - i2c->rd = 0x600800 + entry->read; - i2c->wr = 0x600800 + entry->write; - break; - case 5: - i2c->bit.setsda = nv50_i2c_setsda; - i2c->bit.setscl = nv50_i2c_setscl; - if (dev_priv->card_type < NV_D0) { - i2c->bit.getsda = nv50_i2c_getsda; - i2c->bit.getscl = nv50_i2c_getscl; - i2c->rd = nv50_i2c_port[entry->read]; - i2c->wr = i2c->rd; + snprintf(port->adapter.name, sizeof(port->adapter.name), + "nouveau-%s-%d", pci_name(dev->pdev), i); + port->adapter.owner = THIS_MODULE; + port->adapter.dev.parent = &dev->pdev->dev; + port->dev = dev; + port->index = i; + port->dcb = ROM32(entry[0]); + i2c_set_adapdata(&port->adapter, i2c); + + if (port->adapter.algo != &nouveau_dp_i2c_algo) { + port->adapter.algo_data = &port->bit; + port->bit.udelay = 10; + port->bit.timeout = usecs_to_jiffies(2200); + port->bit.data = port; + port->bit.setsda = i2c_drive_sda; + port->bit.setscl = i2c_drive_scl; + port->bit.getsda = i2c_sense_sda; + port->bit.getscl = i2c_sense_scl; + + i2c_drive_scl(port, 0); + i2c_drive_sda(port, 1); + i2c_drive_scl(port, 1); + + ret = i2c_bit_add_bus(&port->adapter); } else { - i2c->bit.getsda = nvd0_i2c_getsda; - i2c->bit.getscl = nvd0_i2c_getscl; - i2c->rd = 0x00d014 + (entry->read * 0x20); - i2c->wr = i2c->rd; + port->adapter.algo = &nouveau_dp_i2c_algo; + ret = i2c_add_adapter(&port->adapter); } - break; - case 6: - i2c->rd = entry->read; - i2c->wr = entry->write; - break; - default: - NV_ERROR(dev, "DCB I2C port type %d unknown\n", - entry->port_type); - kfree(i2c); - return -EINVAL; - } - snprintf(i2c->adapter.name, sizeof(i2c->adapter.name), - "nouveau-%s-%d", pci_name(dev->pdev), index); - i2c->adapter.owner = THIS_MODULE; - i2c->adapter.dev.parent = &dev->pdev->dev; - i2c->dev = dev; - i2c_set_adapdata(&i2c->adapter, i2c); - - if (entry->port_type < 6) { - i2c->adapter.algo_data = &i2c->bit; - i2c->bit.udelay = 40; - i2c->bit.timeout = usecs_to_jiffies(5000); - i2c->bit.data = i2c; - ret = i2c_bit_add_bus(&i2c->adapter); - } else { - i2c->adapter.algo = &nouveau_dp_i2c_algo; - ret = i2c_add_adapter(&i2c->adapter); - } + if (ret) { + NV_ERROR(dev, "I2C%d: failed register: %d\n", i, ret); + kfree(port); + continue; + } - if (ret) { - NV_ERROR(dev, "Failed to register i2c %d\n", index); - kfree(i2c); - return ret; + list_add_tail(&port->head, &dev_priv->i2c_ports); } - entry->chan = i2c; return 0; } void -nouveau_i2c_fini(struct drm_device *dev, struct dcb_i2c_entry *entry) +nouveau_i2c_fini(struct drm_device *dev) { - if (!entry->chan) - return; + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_i2c_chan *port, *tmp; - i2c_del_adapter(&entry->chan->adapter); - kfree(entry->chan); - entry->chan = NULL; + list_for_each_entry_safe(port, tmp, &dev_priv->i2c_ports, head) { + i2c_del_adapter(&port->adapter); + kfree(port); + } } struct nouveau_i2c_chan * -nouveau_i2c_find(struct drm_device *dev, int index) +nouveau_i2c_find(struct drm_device *dev, u8 index) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct dcb_i2c_entry *i2c = &dev_priv->vbios.dcb.i2c[index]; + struct nouveau_i2c_chan *port; + + if (index == NV_I2C_DEFAULT(0) || + index == NV_I2C_DEFAULT(1)) { + u8 version, *i2c = i2c_table(dev, &version); + if (i2c && version >= 0x30) { + if (index == NV_I2C_DEFAULT(0)) + index = (i2c[4] & 0x0f); + else + index = (i2c[4] & 0xf0) >> 4; + } else { + index = 2; + } + } - if (index >= DCB_MAX_NUM_I2C_ENTRIES) - return NULL; + list_for_each_entry(port, &dev_priv->i2c_ports, head) { + if (port->index == index) + break; + } - if (dev_priv->card_type >= NV_50 && (i2c->entry & 0x00000100)) { - uint32_t reg = 0xe500, val; + if (&port->head == &dev_priv->i2c_ports) + return NULL; - if (i2c->port_type == 6) { - reg += i2c->read * 0x50; + if (dev_priv->card_type >= NV_50 && (port->dcb & 0x00000100)) { + u32 reg = 0x00e500, val; + if (port->type == 6) { + reg += port->drive * 0x50; val = 0x2002; } else { - reg += ((i2c->entry & 0x1e00) >> 9) * 0x50; + reg += ((port->dcb & 0x1e00) >> 9) * 0x50; val = 0xe001; } @@ -294,9 +340,7 @@ nouveau_i2c_find(struct drm_device *dev, int index) nv_mask(dev, reg + 0x00, 0x0000f003, val); } - if (!i2c->chan && nouveau_i2c_init(dev, i2c, index)) - return NULL; - return i2c->chan; + return port; } bool @@ -331,9 +375,13 @@ nouveau_i2c_identify(struct drm_device *dev, const char *what, struct nouveau_i2c_chan *i2c = nouveau_i2c_find(dev, index); int i; - NV_DEBUG(dev, "Probing %ss on I2C bus: %d\n", what, index); + if (!i2c) { + NV_DEBUG(dev, "No bus when probing %s on %d\n", what, index); + return -ENODEV; + } - for (i = 0; i2c && info[i].addr; i++) { + NV_DEBUG(dev, "Probing %ss on I2C bus: %d\n", what, i2c->index); + for (i = 0; info[i].addr; i++) { if (nouveau_probe_i2c_addr(i2c, info[i].addr) && (!match || match(i2c, &info[i]))) { NV_INFO(dev, "Detected %s: %s\n", what, info[i].type); @@ -342,6 +390,5 @@ nouveau_i2c_identify(struct drm_device *dev, const char *what, } NV_DEBUG(dev, "No devices found.\n"); - return -ENODEV; } diff --git a/drivers/gpu/drm/nouveau/nouveau_i2c.h b/drivers/gpu/drm/nouveau/nouveau_i2c.h index 422b62f..1d08389 100644 --- a/drivers/gpu/drm/nouveau/nouveau_i2c.h +++ b/drivers/gpu/drm/nouveau/nouveau_i2c.h @@ -27,20 +27,26 @@ #include #include "drm_dp_helper.h" -struct dcb_i2c_entry; +#define NV_I2C_PORT(n) (0x00 + (n)) +#define NV_I2C_PORT_NUM 0x10 +#define NV_I2C_DEFAULT(n) (0x80 + (n)) struct nouveau_i2c_chan { struct i2c_adapter adapter; struct drm_device *dev; struct i2c_algo_bit_data bit; - unsigned rd; - unsigned wr; - unsigned data; + struct list_head head; + u8 index; + u8 type; + u32 dcb; + u32 drive; + u32 sense; + u32 state; }; -int nouveau_i2c_init(struct drm_device *, struct dcb_i2c_entry *, int index); -void nouveau_i2c_fini(struct drm_device *, struct dcb_i2c_entry *); -struct nouveau_i2c_chan *nouveau_i2c_find(struct drm_device *, int index); +int nouveau_i2c_init(struct drm_device *); +void nouveau_i2c_fini(struct drm_device *); +struct nouveau_i2c_chan *nouveau_i2c_find(struct drm_device *, u8 index); bool nouveau_probe_i2c_addr(struct nouveau_i2c_chan *i2c, int addr); int nouveau_i2c_identify(struct drm_device *dev, const char *what, struct i2c_board_info *info, diff --git a/drivers/gpu/drm/nouveau/nouveau_mem.c b/drivers/gpu/drm/nouveau/nouveau_mem.c index 36bec48..b08065f 100644 --- a/drivers/gpu/drm/nouveau/nouveau_mem.c +++ b/drivers/gpu/drm/nouveau/nouveau_mem.c @@ -26,7 +26,8 @@ * DEALINGS IN THE SOFTWARE. * * Authors: - * Keith Whitwell + * Ben Skeggs + * Roy Spliet */ @@ -192,75 +193,6 @@ nouveau_mem_gart_fini(struct drm_device *dev) } } -static uint32_t -nouveau_mem_detect_nv04(struct drm_device *dev) -{ - uint32_t boot0 = nv_rd32(dev, NV04_PFB_BOOT_0); - - if (boot0 & 0x00000100) - return (((boot0 >> 12) & 0xf) * 2 + 2) * 1024 * 1024; - - switch (boot0 & NV04_PFB_BOOT_0_RAM_AMOUNT) { - case NV04_PFB_BOOT_0_RAM_AMOUNT_32MB: - return 32 * 1024 * 1024; - case NV04_PFB_BOOT_0_RAM_AMOUNT_16MB: - return 16 * 1024 * 1024; - case NV04_PFB_BOOT_0_RAM_AMOUNT_8MB: - return 8 * 1024 * 1024; - case NV04_PFB_BOOT_0_RAM_AMOUNT_4MB: - return 4 * 1024 * 1024; - } - - return 0; -} - -static uint32_t -nouveau_mem_detect_nforce(struct drm_device *dev) -{ - struct drm_nouveau_private *dev_priv = dev->dev_private; - struct pci_dev *bridge; - uint32_t mem; - - bridge = pci_get_bus_and_slot(0, PCI_DEVFN(0, 1)); - if (!bridge) { - NV_ERROR(dev, "no bridge device\n"); - return 0; - } - - if (dev_priv->flags & NV_NFORCE) { - pci_read_config_dword(bridge, 0x7C, &mem); - return (uint64_t)(((mem >> 6) & 31) + 1)*1024*1024; - } else - if (dev_priv->flags & NV_NFORCE2) { - pci_read_config_dword(bridge, 0x84, &mem); - return (uint64_t)(((mem >> 4) & 127) + 1)*1024*1024; - } - - NV_ERROR(dev, "impossible!\n"); - return 0; -} - -int -nouveau_mem_detect(struct drm_device *dev) -{ - struct drm_nouveau_private *dev_priv = dev->dev_private; - - if (dev_priv->card_type == NV_04) { - dev_priv->vram_size = nouveau_mem_detect_nv04(dev); - } else - if (dev_priv->flags & (NV_NFORCE | NV_NFORCE2)) { - dev_priv->vram_size = nouveau_mem_detect_nforce(dev); - } else - if (dev_priv->card_type < NV_50) { - dev_priv->vram_size = nv_rd32(dev, NV04_PFB_FIFO_DATA); - dev_priv->vram_size &= NV10_PFB_FIFO_DATA_RAM_AMOUNT_MB_MASK; - } - - if (dev_priv->vram_size) - return 0; - return -ENOMEM; -} - bool nouveau_mem_flags_valid(struct drm_device *dev, u32 tile_flags) { @@ -385,11 +317,29 @@ nouveau_mem_init_agp(struct drm_device *dev) return 0; } +static const struct vram_types { + int value; + const char *name; +} vram_type_map[] = { + { NV_MEM_TYPE_STOLEN , "stolen system memory" }, + { NV_MEM_TYPE_SGRAM , "SGRAM" }, + { NV_MEM_TYPE_SDRAM , "SDRAM" }, + { NV_MEM_TYPE_DDR1 , "DDR1" }, + { NV_MEM_TYPE_DDR2 , "DDR2" }, + { NV_MEM_TYPE_DDR3 , "DDR3" }, + { NV_MEM_TYPE_GDDR2 , "GDDR2" }, + { NV_MEM_TYPE_GDDR3 , "GDDR3" }, + { NV_MEM_TYPE_GDDR4 , "GDDR4" }, + { NV_MEM_TYPE_GDDR5 , "GDDR5" }, + { NV_MEM_TYPE_UNKNOWN, "unknown type" } +}; + int nouveau_mem_vram_init(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; struct ttm_bo_device *bdev = &dev_priv->ttm.bdev; + const struct vram_types *vram_type; int ret, dma_bits; dma_bits = 32; @@ -407,6 +357,12 @@ nouveau_mem_vram_init(struct drm_device *dev) ret = pci_set_dma_mask(dev->pdev, DMA_BIT_MASK(dma_bits)); if (ret) return ret; + ret = pci_set_consistent_dma_mask(dev->pdev, DMA_BIT_MASK(dma_bits)); + if (ret) { + /* Reset to default value. */ + pci_set_consistent_dma_mask(dev->pdev, DMA_BIT_MASK(32)); + } + ret = nouveau_ttm_global_init(dev_priv); if (ret) @@ -421,7 +377,21 @@ nouveau_mem_vram_init(struct drm_device *dev) return ret; } - NV_INFO(dev, "Detected %dMiB VRAM\n", (int)(dev_priv->vram_size >> 20)); + vram_type = vram_type_map; + while (vram_type->value != NV_MEM_TYPE_UNKNOWN) { + if (nouveau_vram_type) { + if (!strcasecmp(nouveau_vram_type, vram_type->name)) + break; + dev_priv->vram_type = vram_type->value; + } else { + if (vram_type->value == dev_priv->vram_type) + break; + } + vram_type++; + } + + NV_INFO(dev, "Detected %dMiB VRAM (%s)\n", + (int)(dev_priv->vram_size >> 20), vram_type->name); if (dev_priv->vram_sys_base) { NV_INFO(dev, "Stolen system memory at: 0x%010llx\n", dev_priv->vram_sys_base); @@ -502,216 +472,617 @@ nouveau_mem_gart_init(struct drm_device *dev) return 0; } -/* XXX: For now a dummy. More samples required, possibly even a card - * Called from nouveau_perf.c */ -void nv30_mem_timing_entry(struct drm_device *dev, struct nouveau_pm_tbl_header *hdr, - struct nouveau_pm_tbl_entry *e, uint8_t magic_number, - struct nouveau_pm_memtiming *timing) { - - NV_DEBUG(dev,"Timing entry format unknown, please contact nouveau developers"); -} - -void nv40_mem_timing_entry(struct drm_device *dev, struct nouveau_pm_tbl_header *hdr, - struct nouveau_pm_tbl_entry *e, uint8_t magic_number, - struct nouveau_pm_memtiming *timing) { - - timing->reg_0 = (e->tRC << 24 | e->tRFC << 16 | e->tRAS << 8 | e->tRP); +static int +nv40_mem_timing_calc(struct drm_device *dev, u32 freq, + struct nouveau_pm_tbl_entry *e, u8 len, + struct nouveau_pm_memtiming *boot, + struct nouveau_pm_memtiming *t) +{ + t->reg[0] = (e->tRP << 24 | e->tRAS << 16 | e->tRFC << 8 | e->tRC); /* XXX: I don't trust the -1's and +1's... they must come * from somewhere! */ - timing->reg_1 = (e->tWR + 2 + magic_number) << 24 | - 1 << 16 | - (e->tUNK_1 + 2 + magic_number) << 8 | - (e->tCL + 2 - magic_number); - timing->reg_2 = (magic_number << 24 | e->tUNK_12 << 16 | e->tUNK_11 << 8 | e->tUNK_10); - timing->reg_2 |= 0x20200000; - - NV_DEBUG(dev, "Entry %d: 220: %08x %08x %08x\n", timing->id, - timing->reg_0, timing->reg_1,timing->reg_2); + t->reg[1] = (e->tWR + 2 + (t->tCWL - 1)) << 24 | + 1 << 16 | + (e->tWTR + 2 + (t->tCWL - 1)) << 8 | + (e->tCL + 2 - (t->tCWL - 1)); + + t->reg[2] = 0x20200000 | + ((t->tCWL - 1) << 24 | + e->tRRD << 16 | + e->tRCDWR << 8 | + e->tRCDRD); + + NV_DEBUG(dev, "Entry %d: 220: %08x %08x %08x\n", t->id, + t->reg[0], t->reg[1], t->reg[2]); + return 0; } -void nv50_mem_timing_entry(struct drm_device *dev, struct bit_entry *P, struct nouveau_pm_tbl_header *hdr, - struct nouveau_pm_tbl_entry *e, uint8_t magic_number,struct nouveau_pm_memtiming *timing) { +static int +nv50_mem_timing_calc(struct drm_device *dev, u32 freq, + struct nouveau_pm_tbl_entry *e, u8 len, + struct nouveau_pm_memtiming *boot, + struct nouveau_pm_memtiming *t) +{ struct drm_nouveau_private *dev_priv = dev->dev_private; + struct bit_entry P; + uint8_t unk18 = 1, unk20 = 0, unk21 = 0, tmp7_3; - uint8_t unk18 = 1, - unk19 = 1, - unk20 = 0, - unk21 = 0; + if (bit_table(dev, 'P', &P)) + return -EINVAL; - switch (min(hdr->entry_len, (u8) 22)) { + switch (min(len, (u8) 22)) { case 22: unk21 = e->tUNK_21; case 21: unk20 = e->tUNK_20; case 20: - unk19 = e->tUNK_19; + if (e->tCWL > 0) + t->tCWL = e->tCWL; case 19: unk18 = e->tUNK_18; break; } - timing->reg_0 = (e->tRC << 24 | e->tRFC << 16 | e->tRAS << 8 | e->tRP); + t->reg[0] = (e->tRP << 24 | e->tRAS << 16 | e->tRFC << 8 | e->tRC); - /* XXX: I don't trust the -1's and +1's... they must come - * from somewhere! */ - timing->reg_1 = (e->tWR + unk19 + 1 + magic_number) << 24 | - max(unk18, (u8) 1) << 16 | - (e->tUNK_1 + unk19 + 1 + magic_number) << 8; - if (dev_priv->chipset == 0xa8) { - timing->reg_1 |= (e->tCL - 1); - } else { - timing->reg_1 |= (e->tCL + 2 - magic_number); - } - timing->reg_2 = (e->tUNK_12 << 16 | e->tUNK_11 << 8 | e->tUNK_10); - - timing->reg_5 = (e->tRAS << 24 | e->tRC); - timing->reg_5 += max(e->tUNK_10, e->tUNK_11) << 16; - - if (P->version == 1) { - timing->reg_2 |= magic_number << 24; - timing->reg_3 = (0x14 + e->tCL) << 24 | - 0x16 << 16 | - (e->tCL - 1) << 8 | - (e->tCL - 1); - timing->reg_4 = (nv_rd32(dev,0x10022c) & 0xffff0000) | e->tUNK_13 << 8 | e->tUNK_13; - timing->reg_5 |= (e->tCL + 2) << 8; - timing->reg_7 = 0x4000202 | (e->tCL - 1) << 16; + t->reg[1] = (e->tWR + 2 + (t->tCWL - 1)) << 24 | + max(unk18, (u8) 1) << 16 | + (e->tWTR + 2 + (t->tCWL - 1)) << 8; + + t->reg[2] = ((t->tCWL - 1) << 24 | + e->tRRD << 16 | + e->tRCDWR << 8 | + e->tRCDRD); + + t->reg[4] = e->tUNK_13 << 8 | e->tUNK_13; + + t->reg[5] = (e->tRFC << 24 | max(e->tRCDRD, e->tRCDWR) << 16 | e->tRP); + + t->reg[8] = boot->reg[8] & 0xffffff00; + + if (P.version == 1) { + t->reg[1] |= (e->tCL + 2 - (t->tCWL - 1)); + + t->reg[3] = (0x14 + e->tCL) << 24 | + 0x16 << 16 | + (e->tCL - 1) << 8 | + (e->tCL - 1); + + t->reg[4] |= boot->reg[4] & 0xffff0000; + + t->reg[6] = (0x33 - t->tCWL) << 16 | + t->tCWL << 8 | + (0x2e + e->tCL - t->tCWL); + + t->reg[7] = 0x4000202 | (e->tCL - 1) << 16; + + /* XXX: P.version == 1 only has DDR2 and GDDR3? */ + if (dev_priv->vram_type == NV_MEM_TYPE_DDR2) { + t->reg[5] |= (e->tCL + 3) << 8; + t->reg[6] |= (t->tCWL - 2) << 8; + t->reg[8] |= (e->tCL - 4); + } else { + t->reg[5] |= (e->tCL + 2) << 8; + t->reg[6] |= t->tCWL << 8; + t->reg[8] |= (e->tCL - 2); + } } else { - timing->reg_2 |= (unk19 - 1) << 24; - /* XXX: reg_10022c for recentish cards pretty much unknown*/ - timing->reg_3 = e->tCL - 1; - timing->reg_4 = (unk20 << 24 | unk21 << 16 | - e->tUNK_13 << 8 | e->tUNK_13); + t->reg[1] |= (5 + e->tCL - (t->tCWL)); + + /* XXX: 0xb? 0x30? */ + t->reg[3] = (0x30 + e->tCL) << 24 | + (boot->reg[3] & 0x00ff0000)| + (0xb + e->tCL) << 8 | + (e->tCL - 1); + + t->reg[4] |= (unk20 << 24 | unk21 << 16); + /* XXX: +6? */ - timing->reg_5 |= (unk19 + 6) << 8; + t->reg[5] |= (t->tCWL + 6) << 8; - /* XXX: reg_10023c currently unknown - * 10023c seen as 06xxxxxx, 0bxxxxxx or 0fxxxxxx */ - timing->reg_7 = 0x202; + t->reg[6] = (0x5a + e->tCL) << 16 | + (6 - e->tCL + t->tCWL) << 8 | + (0x50 + e->tCL - t->tCWL); + + tmp7_3 = (boot->reg[7] & 0xff000000) >> 24; + t->reg[7] = (tmp7_3 << 24) | + ((tmp7_3 - 6 + e->tCL) << 16) | + 0x202; } - NV_DEBUG(dev, "Entry %d: 220: %08x %08x %08x %08x\n", timing->id, - timing->reg_0, timing->reg_1, - timing->reg_2, timing->reg_3); + NV_DEBUG(dev, "Entry %d: 220: %08x %08x %08x %08x\n", t->id, + t->reg[0], t->reg[1], t->reg[2], t->reg[3]); NV_DEBUG(dev, " 230: %08x %08x %08x %08x\n", - timing->reg_4, timing->reg_5, - timing->reg_6, timing->reg_7); - NV_DEBUG(dev, " 240: %08x\n", timing->reg_8); -} - -void nvc0_mem_timing_entry(struct drm_device *dev, struct nouveau_pm_tbl_header *hdr, - struct nouveau_pm_tbl_entry *e, struct nouveau_pm_memtiming *timing) { - timing->reg_0 = (e->tRC << 24 | (e->tRFC & 0x7f) << 17 | e->tRAS << 8 | e->tRP); - timing->reg_1 = (nv_rd32(dev,0x10f294) & 0xff000000) | (e->tUNK_11&0x0f) << 20 | (e->tUNK_19 << 7) | (e->tCL & 0x0f); - timing->reg_2 = (nv_rd32(dev,0x10f298) & 0xff0000ff) | e->tWR << 16 | e->tUNK_1 << 8; - timing->reg_3 = e->tUNK_20 << 9 | e->tUNK_13; - timing->reg_4 = (nv_rd32(dev,0x10f2a0) & 0xfff000ff) | e->tUNK_12 << 15; - NV_DEBUG(dev, "Entry %d: 290: %08x %08x %08x %08x\n", timing->id, - timing->reg_0, timing->reg_1, - timing->reg_2, timing->reg_3); - NV_DEBUG(dev, " 2a0: %08x %08x %08x %08x\n", - timing->reg_4, timing->reg_5, - timing->reg_6, timing->reg_7); + t->reg[4], t->reg[5], t->reg[6], t->reg[7]); + NV_DEBUG(dev, " 240: %08x\n", t->reg[8]); + return 0; +} + +static int +nvc0_mem_timing_calc(struct drm_device *dev, u32 freq, + struct nouveau_pm_tbl_entry *e, u8 len, + struct nouveau_pm_memtiming *boot, + struct nouveau_pm_memtiming *t) +{ + if (e->tCWL > 0) + t->tCWL = e->tCWL; + + t->reg[0] = (e->tRP << 24 | (e->tRAS & 0x7f) << 17 | + e->tRFC << 8 | e->tRC); + + t->reg[1] = (boot->reg[1] & 0xff000000) | + (e->tRCDWR & 0x0f) << 20 | + (e->tRCDRD & 0x0f) << 14 | + (t->tCWL << 7) | + (e->tCL & 0x0f); + + t->reg[2] = (boot->reg[2] & 0xff0000ff) | + e->tWR << 16 | e->tWTR << 8; + + t->reg[3] = (e->tUNK_20 & 0x1f) << 9 | + (e->tUNK_21 & 0xf) << 5 | + (e->tUNK_13 & 0x1f); + + t->reg[4] = (boot->reg[4] & 0xfff00fff) | + (e->tRRD&0x1f) << 15; + + NV_DEBUG(dev, "Entry %d: 290: %08x %08x %08x %08x\n", t->id, + t->reg[0], t->reg[1], t->reg[2], t->reg[3]); + NV_DEBUG(dev, " 2a0: %08x\n", t->reg[4]); + return 0; } /** - * Processes the Memory Timing BIOS table, stores generated - * register values - * @pre init scripts were run, memtiming regs are initialized + * MR generation methods */ -void -nouveau_mem_timing_init(struct drm_device *dev) + +static int +nouveau_mem_ddr2_mr(struct drm_device *dev, u32 freq, + struct nouveau_pm_tbl_entry *e, u8 len, + struct nouveau_pm_memtiming *boot, + struct nouveau_pm_memtiming *t) +{ + t->drive_strength = 0; + if (len < 15) { + t->odt = boot->odt; + } else { + t->odt = e->RAM_FT1 & 0x07; + } + + if (e->tCL >= NV_MEM_CL_DDR2_MAX) { + NV_WARN(dev, "(%u) Invalid tCL: %u", t->id, e->tCL); + return -ERANGE; + } + + if (e->tWR >= NV_MEM_WR_DDR2_MAX) { + NV_WARN(dev, "(%u) Invalid tWR: %u", t->id, e->tWR); + return -ERANGE; + } + + if (t->odt > 3) { + NV_WARN(dev, "(%u) Invalid odt value, assuming disabled: %x", + t->id, t->odt); + t->odt = 0; + } + + t->mr[0] = (boot->mr[0] & 0x100f) | + (e->tCL) << 4 | + (e->tWR - 1) << 9; + t->mr[1] = (boot->mr[1] & 0x101fbb) | + (t->odt & 0x1) << 2 | + (t->odt & 0x2) << 5; + + NV_DEBUG(dev, "(%u) MR: %08x", t->id, t->mr[0]); + return 0; +} + +uint8_t nv_mem_wr_lut_ddr3[NV_MEM_WR_DDR3_MAX] = { + 0, 0, 0, 0, 0, 1, 2, 3, 4, 5, 5, 6, 6, 7, 7, 0, 0}; + +static int +nouveau_mem_ddr3_mr(struct drm_device *dev, u32 freq, + struct nouveau_pm_tbl_entry *e, u8 len, + struct nouveau_pm_memtiming *boot, + struct nouveau_pm_memtiming *t) +{ + u8 cl = e->tCL - 4; + + t->drive_strength = 0; + if (len < 15) { + t->odt = boot->odt; + } else { + t->odt = e->RAM_FT1 & 0x07; + } + + if (e->tCL >= NV_MEM_CL_DDR3_MAX || e->tCL < 4) { + NV_WARN(dev, "(%u) Invalid tCL: %u", t->id, e->tCL); + return -ERANGE; + } + + if (e->tWR >= NV_MEM_WR_DDR3_MAX || e->tWR < 4) { + NV_WARN(dev, "(%u) Invalid tWR: %u", t->id, e->tWR); + return -ERANGE; + } + + if (e->tCWL < 5) { + NV_WARN(dev, "(%u) Invalid tCWL: %u", t->id, e->tCWL); + return -ERANGE; + } + + t->mr[0] = (boot->mr[0] & 0x180b) | + /* CAS */ + (cl & 0x7) << 4 | + (cl & 0x8) >> 1 | + (nv_mem_wr_lut_ddr3[e->tWR]) << 9; + t->mr[1] = (boot->mr[1] & 0x101dbb) | + (t->odt & 0x1) << 2 | + (t->odt & 0x2) << 5 | + (t->odt & 0x4) << 7; + t->mr[2] = (boot->mr[2] & 0x20ffb7) | (e->tCWL - 5) << 3; + + NV_DEBUG(dev, "(%u) MR: %08x %08x", t->id, t->mr[0], t->mr[2]); + return 0; +} + +uint8_t nv_mem_cl_lut_gddr3[NV_MEM_CL_GDDR3_MAX] = { + 0, 0, 0, 0, 4, 5, 6, 7, 0, 1, 2, 3, 8, 9, 10, 11}; +uint8_t nv_mem_wr_lut_gddr3[NV_MEM_WR_GDDR3_MAX] = { + 0, 0, 0, 0, 0, 2, 3, 8, 9, 10, 11, 0, 0, 1, 1, 0, 3}; + +static int +nouveau_mem_gddr3_mr(struct drm_device *dev, u32 freq, + struct nouveau_pm_tbl_entry *e, u8 len, + struct nouveau_pm_memtiming *boot, + struct nouveau_pm_memtiming *t) +{ + if (len < 15) { + t->drive_strength = boot->drive_strength; + t->odt = boot->odt; + } else { + t->drive_strength = (e->RAM_FT1 & 0x30) >> 4; + t->odt = e->RAM_FT1 & 0x07; + } + + if (e->tCL >= NV_MEM_CL_GDDR3_MAX) { + NV_WARN(dev, "(%u) Invalid tCL: %u", t->id, e->tCL); + return -ERANGE; + } + + if (e->tWR >= NV_MEM_WR_GDDR3_MAX) { + NV_WARN(dev, "(%u) Invalid tWR: %u", t->id, e->tWR); + return -ERANGE; + } + + if (t->odt > 3) { + NV_WARN(dev, "(%u) Invalid odt value, assuming autocal: %x", + t->id, t->odt); + t->odt = 0; + } + + t->mr[0] = (boot->mr[0] & 0xe0b) | + /* CAS */ + ((nv_mem_cl_lut_gddr3[e->tCL] & 0x7) << 4) | + ((nv_mem_cl_lut_gddr3[e->tCL] & 0x8) >> 2); + t->mr[1] = (boot->mr[1] & 0x100f40) | t->drive_strength | + (t->odt << 2) | + (nv_mem_wr_lut_gddr3[e->tWR] & 0xf) << 4; + t->mr[2] = boot->mr[2]; + + NV_DEBUG(dev, "(%u) MR: %08x %08x %08x", t->id, + t->mr[0], t->mr[1], t->mr[2]); + return 0; +} + +static int +nouveau_mem_gddr5_mr(struct drm_device *dev, u32 freq, + struct nouveau_pm_tbl_entry *e, u8 len, + struct nouveau_pm_memtiming *boot, + struct nouveau_pm_memtiming *t) +{ + if (len < 15) { + t->drive_strength = boot->drive_strength; + t->odt = boot->odt; + } else { + t->drive_strength = (e->RAM_FT1 & 0x30) >> 4; + t->odt = e->RAM_FT1 & 0x03; + } + + if (e->tCL >= NV_MEM_CL_GDDR5_MAX) { + NV_WARN(dev, "(%u) Invalid tCL: %u", t->id, e->tCL); + return -ERANGE; + } + + if (e->tWR >= NV_MEM_WR_GDDR5_MAX) { + NV_WARN(dev, "(%u) Invalid tWR: %u", t->id, e->tWR); + return -ERANGE; + } + + if (t->odt > 3) { + NV_WARN(dev, "(%u) Invalid odt value, assuming autocal: %x", + t->id, t->odt); + t->odt = 0; + } + + t->mr[0] = (boot->mr[0] & 0x007) | + ((e->tCL - 5) << 3) | + ((e->tWR - 4) << 8); + t->mr[1] = (boot->mr[1] & 0x1007f0) | + t->drive_strength | + (t->odt << 2); + + NV_DEBUG(dev, "(%u) MR: %08x %08x", t->id, t->mr[0], t->mr[1]); + return 0; +} + +int +nouveau_mem_timing_calc(struct drm_device *dev, u32 freq, + struct nouveau_pm_memtiming *t) { struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_pm_engine *pm = &dev_priv->engine.pm; - struct nouveau_pm_memtimings *memtimings = &pm->memtimings; - struct nvbios *bios = &dev_priv->vbios; - struct bit_entry P; - struct nouveau_pm_tbl_header *hdr = NULL; - uint8_t magic_number; - u8 *entry; - int i; + struct nouveau_pm_memtiming *boot = &pm->boot.timing; + struct nouveau_pm_tbl_entry *e; + u8 ver, len, *ptr, *ramcfg; + int ret; + + ptr = nouveau_perf_timing(dev, freq, &ver, &len); + if (!ptr || ptr[0] == 0x00) { + *t = *boot; + return 0; + } + e = (struct nouveau_pm_tbl_entry *)ptr; + + t->tCWL = boot->tCWL; + + switch (dev_priv->card_type) { + case NV_40: + ret = nv40_mem_timing_calc(dev, freq, e, len, boot, t); + break; + case NV_50: + ret = nv50_mem_timing_calc(dev, freq, e, len, boot, t); + break; + case NV_C0: + ret = nvc0_mem_timing_calc(dev, freq, e, len, boot, t); + break; + default: + ret = -ENODEV; + break; + } - if (bios->type == NVBIOS_BIT) { - if (bit_table(dev, 'P', &P)) - return; + switch (dev_priv->vram_type * !ret) { + case NV_MEM_TYPE_GDDR3: + ret = nouveau_mem_gddr3_mr(dev, freq, e, len, boot, t); + break; + case NV_MEM_TYPE_GDDR5: + ret = nouveau_mem_gddr5_mr(dev, freq, e, len, boot, t); + break; + case NV_MEM_TYPE_DDR2: + ret = nouveau_mem_ddr2_mr(dev, freq, e, len, boot, t); + break; + case NV_MEM_TYPE_DDR3: + ret = nouveau_mem_ddr3_mr(dev, freq, e, len, boot, t); + break; + default: + ret = -EINVAL; + break; + } - if (P.version == 1) - hdr = (struct nouveau_pm_tbl_header *) ROMPTR(bios, P.data[4]); + ramcfg = nouveau_perf_ramcfg(dev, freq, &ver, &len); + if (ramcfg) { + int dll_off; + + if (ver == 0x00) + dll_off = !!(ramcfg[3] & 0x04); else - if (P.version == 2) - hdr = (struct nouveau_pm_tbl_header *) ROMPTR(bios, P.data[8]); - else { - NV_WARN(dev, "unknown mem for BIT P %d\n", P.version); + dll_off = !!(ramcfg[2] & 0x40); + + switch (dev_priv->vram_type) { + case NV_MEM_TYPE_GDDR3: + t->mr[1] &= ~0x00000040; + t->mr[1] |= 0x00000040 * dll_off; + break; + default: + t->mr[1] &= ~0x00000001; + t->mr[1] |= 0x00000001 * dll_off; + break; } + } + + return ret; +} + +void +nouveau_mem_timing_read(struct drm_device *dev, struct nouveau_pm_memtiming *t) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + u32 timing_base, timing_regs, mr_base; + int i; + + if (dev_priv->card_type >= 0xC0) { + timing_base = 0x10f290; + mr_base = 0x10f300; } else { - NV_DEBUG(dev, "BMP version too old for memory\n"); - return; + timing_base = 0x100220; + mr_base = 0x1002c0; } - if (!hdr) { - NV_DEBUG(dev, "memory timing table pointer invalid\n"); + t->id = -1; + + switch (dev_priv->card_type) { + case NV_50: + timing_regs = 9; + break; + case NV_C0: + case NV_D0: + timing_regs = 5; + break; + case NV_30: + case NV_40: + timing_regs = 3; + break; + default: + timing_regs = 0; return; } + for(i = 0; i < timing_regs; i++) + t->reg[i] = nv_rd32(dev, timing_base + (0x04 * i)); + + t->tCWL = 0; + if (dev_priv->card_type < NV_C0) { + t->tCWL = ((nv_rd32(dev, 0x100228) & 0x0f000000) >> 24) + 1; + } else if (dev_priv->card_type <= NV_D0) { + t->tCWL = ((nv_rd32(dev, 0x10f294) & 0x00000f80) >> 7); + } - if (hdr->version != 0x10) { - NV_WARN(dev, "memory timing table 0x%02x unknown\n", hdr->version); - return; + t->mr[0] = nv_rd32(dev, mr_base); + t->mr[1] = nv_rd32(dev, mr_base + 0x04); + t->mr[2] = nv_rd32(dev, mr_base + 0x20); + t->mr[3] = nv_rd32(dev, mr_base + 0x24); + + t->odt = 0; + t->drive_strength = 0; + + switch (dev_priv->vram_type) { + case NV_MEM_TYPE_DDR3: + t->odt |= (t->mr[1] & 0x200) >> 7; + case NV_MEM_TYPE_DDR2: + t->odt |= (t->mr[1] & 0x04) >> 2 | + (t->mr[1] & 0x40) >> 5; + break; + case NV_MEM_TYPE_GDDR3: + case NV_MEM_TYPE_GDDR5: + t->drive_strength = t->mr[1] & 0x03; + t->odt = (t->mr[1] & 0x0c) >> 2; + break; + default: + break; } +} - /* validate record length */ - if (hdr->entry_len < 15) { - NV_ERROR(dev, "mem timing table length unknown: %d\n", hdr->entry_len); - return; +int +nouveau_mem_exec(struct nouveau_mem_exec_func *exec, + struct nouveau_pm_level *perflvl) +{ + struct drm_nouveau_private *dev_priv = exec->dev->dev_private; + struct nouveau_pm_memtiming *info = &perflvl->timing; + u32 tMRD = 1000, tCKSRE = 0, tCKSRX = 0, tXS = 0, tDLLK = 0; + u32 mr[3] = { info->mr[0], info->mr[1], info->mr[2] }; + u32 mr1_dlloff; + + switch (dev_priv->vram_type) { + case NV_MEM_TYPE_DDR2: + tDLLK = 2000; + mr1_dlloff = 0x00000001; + break; + case NV_MEM_TYPE_DDR3: + tDLLK = 12000; + mr1_dlloff = 0x00000001; + break; + case NV_MEM_TYPE_GDDR3: + tDLLK = 40000; + mr1_dlloff = 0x00000040; + break; + default: + NV_ERROR(exec->dev, "cannot reclock unsupported memtype\n"); + return -ENODEV; } - /* parse vbios entries into common format */ - memtimings->timing = - kcalloc(hdr->entry_cnt, sizeof(*memtimings->timing), GFP_KERNEL); - if (!memtimings->timing) - return; + /* fetch current MRs */ + switch (dev_priv->vram_type) { + case NV_MEM_TYPE_GDDR3: + case NV_MEM_TYPE_DDR3: + mr[2] = exec->mrg(exec, 2); + default: + mr[1] = exec->mrg(exec, 1); + mr[0] = exec->mrg(exec, 0); + break; + } - /* Get "some number" from the timing reg for NV_40 and NV_50 - * Used in calculations later... source unknown */ - magic_number = 0; - if (P.version == 1) { - magic_number = (nv_rd32(dev, 0x100228) & 0x0f000000) >> 24; + /* DLL 'on' -> DLL 'off' mode, disable before entering self-refresh */ + if (!(mr[1] & mr1_dlloff) && (info->mr[1] & mr1_dlloff)) { + exec->precharge(exec); + exec->mrs (exec, 1, mr[1] | mr1_dlloff); + exec->wait(exec, tMRD); } - entry = (u8*) hdr + hdr->header_len; - for (i = 0; i < hdr->entry_cnt; i++, entry += hdr->entry_len) { - struct nouveau_pm_memtiming *timing = &pm->memtimings.timing[i]; - if (entry[0] == 0) - continue; + /* enter self-refresh mode */ + exec->precharge(exec); + exec->refresh(exec); + exec->refresh(exec); + exec->refresh_auto(exec, false); + exec->refresh_self(exec, true); + exec->wait(exec, tCKSRE); + + /* modify input clock frequency */ + exec->clock_set(exec); + + /* exit self-refresh mode */ + exec->wait(exec, tCKSRX); + exec->precharge(exec); + exec->refresh_self(exec, false); + exec->refresh_auto(exec, true); + exec->wait(exec, tXS); + + /* update MRs */ + if (mr[2] != info->mr[2]) { + exec->mrs (exec, 2, info->mr[2]); + exec->wait(exec, tMRD); + } + + if (mr[1] != info->mr[1]) { + /* need to keep DLL off until later, at least on GDDR3 */ + exec->mrs (exec, 1, info->mr[1] | (mr[1] & mr1_dlloff)); + exec->wait(exec, tMRD); + } + + if (mr[0] != info->mr[0]) { + exec->mrs (exec, 0, info->mr[0]); + exec->wait(exec, tMRD); + } - timing->id = i; - timing->WR = entry[0]; - timing->CL = entry[2]; + /* update PFB timing registers */ + exec->timing_set(exec); - if(dev_priv->card_type <= NV_40) { - nv40_mem_timing_entry(dev,hdr,(struct nouveau_pm_tbl_entry*) entry,magic_number,&pm->memtimings.timing[i]); - } else if(dev_priv->card_type == NV_50){ - nv50_mem_timing_entry(dev,&P,hdr,(struct nouveau_pm_tbl_entry*) entry,magic_number,&pm->memtimings.timing[i]); - } else if(dev_priv->card_type == NV_C0) { - nvc0_mem_timing_entry(dev,hdr,(struct nouveau_pm_tbl_entry*) entry,&pm->memtimings.timing[i]); + /* DLL (enable + ) reset */ + if (!(info->mr[1] & mr1_dlloff)) { + if (mr[1] & mr1_dlloff) { + exec->mrs (exec, 1, info->mr[1]); + exec->wait(exec, tMRD); } + exec->mrs (exec, 0, info->mr[0] | 0x00000100); + exec->wait(exec, tMRD); + exec->mrs (exec, 0, info->mr[0] | 0x00000000); + exec->wait(exec, tMRD); + exec->wait(exec, tDLLK); + if (dev_priv->vram_type == NV_MEM_TYPE_GDDR3) + exec->precharge(exec); } - memtimings->nr_timing = hdr->entry_cnt; - memtimings->supported = P.version == 1; + return 0; } -void -nouveau_mem_timing_fini(struct drm_device *dev) +int +nouveau_mem_vbios_type(struct drm_device *dev) { - struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_pm_memtimings *mem = &dev_priv->engine.pm.memtimings; + struct bit_entry M; + u8 ramcfg = (nv_rd32(dev, 0x101000) & 0x0000003c) >> 2; + if (!bit_table(dev, 'M', &M) || M.version != 2 || M.length < 5) { + u8 *table = ROMPTR(dev, M.data[3]); + if (table && table[0] == 0x10 && ramcfg < table[3]) { + u8 *entry = table + table[1] + (ramcfg * table[2]); + switch (entry[0] & 0x0f) { + case 0: return NV_MEM_TYPE_DDR2; + case 1: return NV_MEM_TYPE_DDR3; + case 2: return NV_MEM_TYPE_GDDR3; + case 3: return NV_MEM_TYPE_GDDR5; + default: + break; + } - if(mem->timing) { - kfree(mem->timing); - mem->timing = NULL; + } } + return NV_MEM_TYPE_UNKNOWN; } static int diff --git a/drivers/gpu/drm/nouveau/nouveau_mxm.c b/drivers/gpu/drm/nouveau/nouveau_mxm.c new file mode 100644 index 0000000..07d0d1e --- /dev/null +++ b/drivers/gpu/drm/nouveau/nouveau_mxm.c @@ -0,0 +1,723 @@ +/* + * Copyright 2011 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs + */ + +#include + +#include "drmP.h" +#include "nouveau_drv.h" + +#define MXM_DBG(dev, fmt, args...) NV_DEBUG((dev), "MXM: " fmt, ##args) +#define MXM_MSG(dev, fmt, args...) NV_INFO((dev), "MXM: " fmt, ##args) + +static u8 * +mxms_data(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + return dev_priv->mxms; + +} + +static u16 +mxms_version(struct drm_device *dev) +{ + u8 *mxms = mxms_data(dev); + u16 version = (mxms[4] << 8) | mxms[5]; + switch (version ) { + case 0x0200: + case 0x0201: + case 0x0300: + return version; + default: + break; + } + + MXM_DBG(dev, "unknown version %d.%d\n", mxms[4], mxms[5]); + return 0x0000; +} + +static u16 +mxms_headerlen(struct drm_device *dev) +{ + return 8; +} + +static u16 +mxms_structlen(struct drm_device *dev) +{ + return *(u16 *)&mxms_data(dev)[6]; +} + +static bool +mxms_checksum(struct drm_device *dev) +{ + u16 size = mxms_headerlen(dev) + mxms_structlen(dev); + u8 *mxms = mxms_data(dev), sum = 0; + while (size--) + sum += *mxms++; + if (sum) { + MXM_DBG(dev, "checksum invalid\n"); + return false; + } + return true; +} + +static bool +mxms_valid(struct drm_device *dev) +{ + u8 *mxms = mxms_data(dev); + if (*(u32 *)mxms != 0x5f4d584d) { + MXM_DBG(dev, "signature invalid\n"); + return false; + } + + if (!mxms_version(dev) || !mxms_checksum(dev)) + return false; + + return true; +} + +static bool +mxms_foreach(struct drm_device *dev, u8 types, + bool (*exec)(struct drm_device *, u8 *, void *), void *info) +{ + u8 *mxms = mxms_data(dev); + u8 *desc = mxms + mxms_headerlen(dev); + u8 *fini = desc + mxms_structlen(dev) - 1; + while (desc < fini) { + u8 type = desc[0] & 0x0f; + u8 headerlen = 0; + u8 recordlen = 0; + u8 entries = 0; + + switch (type) { + case 0: /* Output Device Structure */ + if (mxms_version(dev) >= 0x0300) + headerlen = 8; + else + headerlen = 6; + break; + case 1: /* System Cooling Capability Structure */ + case 2: /* Thermal Structure */ + case 3: /* Input Power Structure */ + headerlen = 4; + break; + case 4: /* GPIO Device Structure */ + headerlen = 4; + recordlen = 2; + entries = (ROM32(desc[0]) & 0x01f00000) >> 20; + break; + case 5: /* Vendor Specific Structure */ + headerlen = 8; + break; + case 6: /* Backlight Control Structure */ + if (mxms_version(dev) >= 0x0300) { + headerlen = 4; + recordlen = 8; + entries = (desc[1] & 0xf0) >> 4; + } else { + headerlen = 8; + } + break; + case 7: /* Fan Control Structure */ + headerlen = 8; + recordlen = 4; + entries = desc[1] & 0x07; + break; + default: + MXM_DBG(dev, "unknown descriptor type %d\n", type); + return false; + } + + if ((drm_debug & DRM_UT_DRIVER) && (exec == NULL)) { + static const char * mxms_desc_name[] = { + "ODS", "SCCS", "TS", "IPS", + "GSD", "VSS", "BCS", "FCS", + }; + u8 *dump = desc; + int i, j; + + MXM_DBG(dev, "%4s: ", mxms_desc_name[type]); + for (j = headerlen - 1; j >= 0; j--) + printk("%02x", dump[j]); + printk("\n"); + dump += headerlen; + + for (i = 0; i < entries; i++, dump += recordlen) { + MXM_DBG(dev, " "); + for (j = recordlen - 1; j >= 0; j--) + printk("%02x", dump[j]); + printk("\n"); + } + } + + if (types & (1 << type)) { + if (!exec(dev, desc, info)) + return false; + } + + desc += headerlen + (entries * recordlen); + } + + return true; +} + +static u8 * +mxm_table(struct drm_device *dev, u8 *size) +{ + struct bit_entry x; + + if (bit_table(dev, 'x', &x)) { + MXM_DBG(dev, "BIT 'x' table not present\n"); + return NULL; + } + + if (x.version != 1 || x.length < 3) { + MXM_MSG(dev, "BIT x table %d/%d unknown\n", + x.version, x.length); + return NULL; + } + + *size = x.length; + return x.data; +} + +/* These map MXM v2.x digital connection values to the appropriate SOR/link, + * hopefully they're correct for all boards within the same chipset... + * + * MXM v3.x VBIOS are nicer and provide pointers to these tables. + */ +static u8 nv84_sor_map[16] = { + 0x00, 0x12, 0x22, 0x11, 0x32, 0x31, 0x11, 0x31, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 +}; + +static u8 nv92_sor_map[16] = { + 0x00, 0x12, 0x22, 0x11, 0x32, 0x31, 0x11, 0x31, + 0x11, 0x31, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 +}; + +static u8 nv94_sor_map[16] = { + 0x00, 0x14, 0x24, 0x11, 0x34, 0x31, 0x11, 0x31, + 0x11, 0x31, 0x12, 0x00, 0x00, 0x00, 0x00, 0x00 +}; + +static u8 nv96_sor_map[16] = { + 0x00, 0x14, 0x24, 0x00, 0x34, 0x00, 0x11, 0x31, + 0x11, 0x31, 0x12, 0x00, 0x00, 0x00, 0x00, 0x00 +}; + +static u8 nv98_sor_map[16] = { + 0x00, 0x14, 0x12, 0x11, 0x00, 0x31, 0x11, 0x31, + 0x11, 0x31, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 +}; + +static u8 +mxm_sor_map(struct drm_device *dev, u8 conn) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + u8 len, *mxm = mxm_table(dev, &len); + if (mxm && len >= 6) { + u8 *map = ROMPTR(dev, mxm[4]); + if (map) { + if (map[0] == 0x10) { + if (conn < map[3]) + return map[map[1] + conn]; + return 0x00; + } + + MXM_MSG(dev, "unknown sor map 0x%02x\n", map[0]); + } + } + + if (dev_priv->chipset == 0x84 || dev_priv->chipset == 0x86) + return nv84_sor_map[conn]; + if (dev_priv->chipset == 0x92) + return nv92_sor_map[conn]; + if (dev_priv->chipset == 0x94) + return nv94_sor_map[conn]; + if (dev_priv->chipset == 0x96) + return nv96_sor_map[conn]; + if (dev_priv->chipset == 0x98) + return nv98_sor_map[conn]; + + MXM_MSG(dev, "missing sor map\n"); + return 0x00; +} + +static u8 +mxm_ddc_map(struct drm_device *dev, u8 port) +{ + u8 len, *mxm = mxm_table(dev, &len); + if (mxm && len >= 8) { + u8 *map = ROMPTR(dev, mxm[6]); + if (map) { + if (map[0] == 0x10) { + if (port < map[3]) + return map[map[1] + port]; + return 0x00; + } + + MXM_MSG(dev, "unknown ddc map 0x%02x\n", map[0]); + } + } + + /* v2.x: directly write port as dcb i2cidx */ + return (port << 4) | port; +} + +struct mxms_odev { + u8 outp_type; + u8 conn_type; + u8 ddc_port; + u8 dig_conn; +}; + +static void +mxms_output_device(struct drm_device *dev, u8 *pdata, struct mxms_odev *desc) +{ + u64 data = ROM32(pdata[0]); + if (mxms_version(dev) >= 0x0300) + data |= (u64)ROM16(pdata[4]) << 32; + + desc->outp_type = (data & 0x00000000000000f0ULL) >> 4; + desc->ddc_port = (data & 0x0000000000000f00ULL) >> 8; + desc->conn_type = (data & 0x000000000001f000ULL) >> 12; + desc->dig_conn = (data & 0x0000000000780000ULL) >> 19; +} + +struct context { + u32 *outp; + struct mxms_odev desc; +}; + +static bool +mxm_match_tmds_partner(struct drm_device *dev, u8 *data, void *info) +{ + struct context *ctx = info; + struct mxms_odev desc; + + mxms_output_device(dev, data, &desc); + if (desc.outp_type == 2 && + desc.dig_conn == ctx->desc.dig_conn) + return false; + return true; +} + +static bool +mxm_match_dcb(struct drm_device *dev, u8 *data, void *info) +{ + struct context *ctx = info; + u64 desc = *(u64 *)data; + + mxms_output_device(dev, data, &ctx->desc); + + /* match dcb encoder type to mxm-ods device type */ + if ((ctx->outp[0] & 0x0000000f) != ctx->desc.outp_type) + return true; + + /* digital output, have some extra stuff to match here, there's a + * table in the vbios that provides a mapping from the mxm digital + * connection enum values to SOR/link + */ + if ((desc & 0x00000000000000f0) >= 0x20) { + /* check against sor index */ + u8 link = mxm_sor_map(dev, ctx->desc.dig_conn); + if ((ctx->outp[0] & 0x0f000000) != (link & 0x0f) << 24) + return true; + + /* check dcb entry has a compatible link field */ + link = (link & 0x30) >> 4; + if ((link & ((ctx->outp[1] & 0x00000030) >> 4)) != link) + return true; + } + + /* mark this descriptor accounted for by setting invalid device type, + * except of course some manufactures don't follow specs properly and + * we need to avoid killing off the TMDS function on DP connectors + * if MXM-SIS is missing an entry for it. + */ + data[0] &= ~0xf0; + if (ctx->desc.outp_type == 6 && ctx->desc.conn_type == 6 && + mxms_foreach(dev, 0x01, mxm_match_tmds_partner, ctx)) { + data[0] |= 0x20; /* modify descriptor to match TMDS now */ + } else { + data[0] |= 0xf0; + } + + return false; +} + +static int +mxm_dcb_sanitise_entry(struct drm_device *dev, void *data, int idx, u8 *dcbe) +{ + struct context ctx = { .outp = (u32 *)dcbe }; + u8 type, i2cidx, link; + u8 *conn; + + /* look for an output device structure that matches this dcb entry. + * if one isn't found, disable it. + */ + if (mxms_foreach(dev, 0x01, mxm_match_dcb, &ctx)) { + MXM_DBG(dev, "disable %d: 0x%08x 0x%08x\n", + idx, ctx.outp[0], ctx.outp[1]); + ctx.outp[0] |= 0x0000000f; + return 0; + } + + /* modify the output's ddc/aux port, there's a pointer to a table + * with the mapping from mxm ddc/aux port to dcb i2c_index in the + * vbios mxm table + */ + i2cidx = mxm_ddc_map(dev, ctx.desc.ddc_port); + if ((ctx.outp[0] & 0x0000000f) != OUTPUT_DP) + i2cidx = (i2cidx & 0x0f) << 4; + else + i2cidx = (i2cidx & 0xf0); + + if (i2cidx != 0xf0) { + ctx.outp[0] &= ~0x000000f0; + ctx.outp[0] |= i2cidx; + } + + /* override dcb sorconf.link, based on what mxm data says */ + switch (ctx.desc.outp_type) { + case 0x00: /* Analog CRT */ + case 0x01: /* Analog TV/HDTV */ + break; + default: + link = mxm_sor_map(dev, ctx.desc.dig_conn) & 0x30; + ctx.outp[1] &= ~0x00000030; + ctx.outp[1] |= link; + break; + } + + /* we may need to fixup various other vbios tables based on what + * the descriptor says the connector type should be. + * + * in a lot of cases, the vbios tables will claim DVI-I is possible, + * and the mxm data says the connector is really HDMI. another + * common example is DP->eDP. + */ + conn = dcb_conn(dev, (ctx.outp[0] & 0x0000f000) >> 12); + type = conn[0]; + switch (ctx.desc.conn_type) { + case 0x01: /* LVDS */ + ctx.outp[1] |= 0x00000004; /* use_power_scripts */ + /* XXX: modify default link width in LVDS table */ + break; + case 0x02: /* HDMI */ + type = DCB_CONNECTOR_HDMI_1; + break; + case 0x03: /* DVI-D */ + type = DCB_CONNECTOR_DVI_D; + break; + case 0x0e: /* eDP, falls through to DPint */ + ctx.outp[1] |= 0x00010000; + case 0x07: /* DP internal, wtf is this?? HP8670w */ + ctx.outp[1] |= 0x00000004; /* use_power_scripts? */ + type = DCB_CONNECTOR_eDP; + break; + default: + break; + } + + if (mxms_version(dev) >= 0x0300) + conn[0] = type; + + return 0; +} + +static bool +mxm_show_unmatched(struct drm_device *dev, u8 *data, void *info) +{ + u64 desc = *(u64 *)data; + if ((desc & 0xf0) != 0xf0) + MXM_MSG(dev, "unmatched output device 0x%016llx\n", desc); + return true; +} + +static void +mxm_dcb_sanitise(struct drm_device *dev) +{ + u8 *dcb = dcb_table(dev); + if (!dcb || dcb[0] != 0x40) { + MXM_DBG(dev, "unsupported DCB version\n"); + return; + } + + dcb_outp_foreach(dev, NULL, mxm_dcb_sanitise_entry); + mxms_foreach(dev, 0x01, mxm_show_unmatched, NULL); +} + +static bool +mxm_shadow_rom_fetch(struct nouveau_i2c_chan *i2c, u8 addr, + u8 offset, u8 size, u8 *data) +{ + struct i2c_msg msgs[] = { + { .addr = addr, .flags = 0, .len = 1, .buf = &offset }, + { .addr = addr, .flags = I2C_M_RD, .len = size, .buf = data, }, + }; + + return i2c_transfer(&i2c->adapter, msgs, 2) == 2; +} + +static bool +mxm_shadow_rom(struct drm_device *dev, u8 version) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_i2c_chan *i2c = NULL; + u8 i2cidx, mxms[6], addr, size; + + i2cidx = mxm_ddc_map(dev, 1 /* LVDS_DDC */) & 0x0f; + if (i2cidx < 0x0f) + i2c = nouveau_i2c_find(dev, i2cidx); + if (!i2c) + return false; + + addr = 0x54; + if (!mxm_shadow_rom_fetch(i2c, addr, 0, 6, mxms)) { + addr = 0x56; + if (!mxm_shadow_rom_fetch(i2c, addr, 0, 6, mxms)) + return false; + } + + dev_priv->mxms = mxms; + size = mxms_headerlen(dev) + mxms_structlen(dev); + dev_priv->mxms = kmalloc(size, GFP_KERNEL); + + if (dev_priv->mxms && + mxm_shadow_rom_fetch(i2c, addr, 0, size, dev_priv->mxms)) + return true; + + kfree(dev_priv->mxms); + dev_priv->mxms = NULL; + return false; +} + +#if defined(CONFIG_ACPI) +static bool +mxm_shadow_dsm(struct drm_device *dev, u8 version) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + static char muid[] = { + 0x00, 0xA4, 0x04, 0x40, 0x7D, 0x91, 0xF2, 0x4C, + 0xB8, 0x9C, 0x79, 0xB6, 0x2F, 0xD5, 0x56, 0x65 + }; + u32 mxms_args[] = { 0x00000000 }; + union acpi_object args[4] = { + /* _DSM MUID */ + { .buffer.type = 3, + .buffer.length = sizeof(muid), + .buffer.pointer = muid, + }, + /* spec says this can be zero to mean "highest revision", but + * of course there's at least one bios out there which fails + * unless you pass in exactly the version it supports.. + */ + { .integer.type = ACPI_TYPE_INTEGER, + .integer.value = (version & 0xf0) << 4 | (version & 0x0f), + }, + /* MXMS function */ + { .integer.type = ACPI_TYPE_INTEGER, + .integer.value = 0x00000010, + }, + /* Pointer to MXMS arguments */ + { .buffer.type = ACPI_TYPE_BUFFER, + .buffer.length = sizeof(mxms_args), + .buffer.pointer = (char *)mxms_args, + }, + }; + struct acpi_object_list list = { ARRAY_SIZE(args), args }; + struct acpi_buffer retn = { ACPI_ALLOCATE_BUFFER, NULL }; + union acpi_object *obj; + acpi_handle handle; + int ret; + + handle = DEVICE_ACPI_HANDLE(&dev->pdev->dev); + if (!handle) + return false; + + ret = acpi_evaluate_object(handle, "_DSM", &list, &retn); + if (ret) { + MXM_DBG(dev, "DSM MXMS failed: %d\n", ret); + return false; + } + + obj = retn.pointer; + if (obj->type == ACPI_TYPE_BUFFER) { + dev_priv->mxms = kmemdup(obj->buffer.pointer, + obj->buffer.length, GFP_KERNEL); + } else + if (obj->type == ACPI_TYPE_INTEGER) { + MXM_DBG(dev, "DSM MXMS returned 0x%llx\n", obj->integer.value); + } + + kfree(obj); + return dev_priv->mxms != NULL; +} +#endif + +#if defined(CONFIG_ACPI_WMI) || defined(CONFIG_ACPI_WMI_MODULE) + +#define WMI_WMMX_GUID "F6CB5C3C-9CAE-4EBD-B577-931EA32A2CC0" + +static u8 +wmi_wmmx_mxmi(struct drm_device *dev, u8 version) +{ + u32 mxmi_args[] = { 0x494D584D /* MXMI */, version, 0 }; + struct acpi_buffer args = { sizeof(mxmi_args), mxmi_args }; + struct acpi_buffer retn = { ACPI_ALLOCATE_BUFFER, NULL }; + union acpi_object *obj; + acpi_status status; + + status = wmi_evaluate_method(WMI_WMMX_GUID, 0, 0, &args, &retn); + if (ACPI_FAILURE(status)) { + MXM_DBG(dev, "WMMX MXMI returned %d\n", status); + return 0x00; + } + + obj = retn.pointer; + if (obj->type == ACPI_TYPE_INTEGER) { + version = obj->integer.value; + MXM_DBG(dev, "WMMX MXMI version %d.%d\n", + (version >> 4), version & 0x0f); + } else { + version = 0; + MXM_DBG(dev, "WMMX MXMI returned non-integer\n"); + } + + kfree(obj); + return version; +} + +static bool +mxm_shadow_wmi(struct drm_device *dev, u8 version) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + u32 mxms_args[] = { 0x534D584D /* MXMS */, version, 0 }; + struct acpi_buffer args = { sizeof(mxms_args), mxms_args }; + struct acpi_buffer retn = { ACPI_ALLOCATE_BUFFER, NULL }; + union acpi_object *obj; + acpi_status status; + + if (!wmi_has_guid(WMI_WMMX_GUID)) { + MXM_DBG(dev, "WMMX GUID not found\n"); + return false; + } + + mxms_args[1] = wmi_wmmx_mxmi(dev, 0x00); + if (!mxms_args[1]) + mxms_args[1] = wmi_wmmx_mxmi(dev, version); + if (!mxms_args[1]) + return false; + + status = wmi_evaluate_method(WMI_WMMX_GUID, 0, 0, &args, &retn); + if (ACPI_FAILURE(status)) { + MXM_DBG(dev, "WMMX MXMS returned %d\n", status); + return false; + } + + obj = retn.pointer; + if (obj->type == ACPI_TYPE_BUFFER) { + dev_priv->mxms = kmemdup(obj->buffer.pointer, + obj->buffer.length, GFP_KERNEL); + } + + kfree(obj); + return dev_priv->mxms != NULL; +} +#endif + +struct mxm_shadow_h { + const char *name; + bool (*exec)(struct drm_device *, u8 version); +} _mxm_shadow[] = { + { "ROM", mxm_shadow_rom }, +#if defined(CONFIG_ACPI) + { "DSM", mxm_shadow_dsm }, +#endif +#if defined(CONFIG_ACPI_WMI) || defined(CONFIG_ACPI_WMI_MODULE) + { "WMI", mxm_shadow_wmi }, +#endif + {} +}; + +static int +mxm_shadow(struct drm_device *dev, u8 version) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct mxm_shadow_h *shadow = _mxm_shadow; + do { + MXM_DBG(dev, "checking %s\n", shadow->name); + if (shadow->exec(dev, version)) { + if (mxms_valid(dev)) + return 0; + kfree(dev_priv->mxms); + dev_priv->mxms = NULL; + } + } while ((++shadow)->name); + return -ENOENT; +} + +int +nouveau_mxm_init(struct drm_device *dev) +{ + u8 mxm_size, *mxm = mxm_table(dev, &mxm_size); + if (!mxm || !mxm[0]) { + MXM_MSG(dev, "no VBIOS data, nothing to do\n"); + return 0; + } + + MXM_MSG(dev, "BIOS version %d.%d\n", mxm[0] >> 4, mxm[0] & 0x0f); + + if (mxm_shadow(dev, mxm[0])) { + MXM_MSG(dev, "failed to locate valid SIS\n"); +#if 0 + /* we should, perhaps, fall back to some kind of limited + * mode here if the x86 vbios hasn't already done the + * work for us (so we prevent loading with completely + * whacked vbios tables). + */ + return -EINVAL; +#else + return 0; +#endif + } + + MXM_MSG(dev, "MXMS Version %d.%d\n", + mxms_version(dev) >> 8, mxms_version(dev) & 0xff); + mxms_foreach(dev, 0, NULL, NULL); + + if (nouveau_mxmdcb) + mxm_dcb_sanitise(dev); + return 0; +} + +void +nouveau_mxm_fini(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + kfree(dev_priv->mxms); + dev_priv->mxms = NULL; +} diff --git a/drivers/gpu/drm/nouveau/nouveau_notifier.c b/drivers/gpu/drm/nouveau/nouveau_notifier.c index 6abdbe6..2ef883c 100644 --- a/drivers/gpu/drm/nouveau/nouveau_notifier.c +++ b/drivers/gpu/drm/nouveau/nouveau_notifier.c @@ -115,7 +115,7 @@ nouveau_notifier_alloc(struct nouveau_channel *chan, uint32_t handle, struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_gpuobj *nobj = NULL; struct drm_mm_node *mem; - uint32_t offset; + uint64_t offset; int target, ret; mem = drm_mm_search_free_in_range(&chan->notifier_heap, size, 0, diff --git a/drivers/gpu/drm/nouveau/nouveau_object.c b/drivers/gpu/drm/nouveau/nouveau_object.c index 960c0ae..cc419fae 100644 --- a/drivers/gpu/drm/nouveau/nouveau_object.c +++ b/drivers/gpu/drm/nouveau/nouveau_object.c @@ -723,14 +723,14 @@ nvc0_gpuobj_channel_init(struct nouveau_channel *chan, struct nouveau_vm *vm) nv_wo32(chan->ramin, 0x020c, 0x000000ff); /* map display semaphore buffers into channel's vm */ - if (dev_priv->card_type >= NV_D0) - return 0; - - for (i = 0; i < 2; i++) { - struct nv50_display_crtc *dispc = &nv50_display(dev)->crtc[i]; - - ret = nouveau_bo_vma_add(dispc->sem.bo, chan->vm, - &chan->dispc_vma[i]); + for (i = 0; i < dev->mode_config.num_crtc; i++) { + struct nouveau_bo *bo; + if (dev_priv->card_type >= NV_D0) + bo = nvd0_display_crtc_sema(dev, i); + else + bo = nv50_display(dev)->crtc[i].sem.bo; + + ret = nouveau_bo_vma_add(bo, chan->vm, &chan->dispc_vma[i]); if (ret) return ret; } @@ -879,9 +879,14 @@ nouveau_gpuobj_channel_takedown(struct nouveau_channel *chan) NV_DEBUG(dev, "ch%d\n", chan->id); - if (dev_priv->card_type >= NV_50 && dev_priv->card_type <= NV_C0) { + if (dev_priv->card_type >= NV_D0) { + for (i = 0; i < dev->mode_config.num_crtc; i++) { + struct nouveau_bo *bo = nvd0_display_crtc_sema(dev, i); + nouveau_bo_vma_del(bo, &chan->dispc_vma[i]); + } + } else + if (dev_priv->card_type >= NV_50) { struct nv50_display *disp = nv50_display(dev); - for (i = 0; i < dev->mode_config.num_crtc; i++) { struct nv50_display_crtc *dispc = &disp->crtc[i]; nouveau_bo_vma_del(dispc->sem.bo, &chan->dispc_vma[i]); diff --git a/drivers/gpu/drm/nouveau/nouveau_perf.c b/drivers/gpu/drm/nouveau/nouveau_perf.c index 33d03fb..69a528d 100644 --- a/drivers/gpu/drm/nouveau/nouveau_perf.c +++ b/drivers/gpu/drm/nouveau/nouveau_perf.c @@ -27,6 +27,178 @@ #include "nouveau_drv.h" #include "nouveau_pm.h" +static u8 * +nouveau_perf_table(struct drm_device *dev, u8 *ver) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nvbios *bios = &dev_priv->vbios; + struct bit_entry P; + + if (!bit_table(dev, 'P', &P) && P.version && P.version <= 2) { + u8 *perf = ROMPTR(dev, P.data[0]); + if (perf) { + *ver = perf[0]; + return perf; + } + } + + if (bios->type == NVBIOS_BMP) { + if (bios->data[bios->offset + 6] >= 0x25) { + u8 *perf = ROMPTR(dev, bios->data[bios->offset + 0x94]); + if (perf) { + *ver = perf[1]; + return perf; + } + } + } + + return NULL; +} + +static u8 * +nouveau_perf_entry(struct drm_device *dev, int idx, + u8 *ver, u8 *hdr, u8 *cnt, u8 *len) +{ + u8 *perf = nouveau_perf_table(dev, ver); + if (perf) { + if (*ver >= 0x12 && *ver < 0x20 && idx < perf[2]) { + *hdr = perf[3]; + *cnt = 0; + *len = 0; + return perf + perf[0] + idx * perf[3]; + } else + if (*ver >= 0x20 && *ver < 0x40 && idx < perf[2]) { + *hdr = perf[3]; + *cnt = perf[4]; + *len = perf[5]; + return perf + perf[1] + idx * (*hdr + (*cnt * *len)); + } else + if (*ver >= 0x40 && *ver < 0x41 && idx < perf[5]) { + *hdr = perf[2]; + *cnt = perf[4]; + *len = perf[3]; + return perf + perf[1] + idx * (*hdr + (*cnt * *len)); + } + } + return NULL; +} + +static u8 * +nouveau_perf_rammap(struct drm_device *dev, u32 freq, + u8 *ver, u8 *hdr, u8 *cnt, u8 *len) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct bit_entry P; + u8 *perf, i = 0; + + if (!bit_table(dev, 'P', &P) && P.version == 2) { + u8 *rammap = ROMPTR(dev, P.data[4]); + if (rammap) { + u8 *ramcfg = rammap + rammap[1]; + + *ver = rammap[0]; + *hdr = rammap[2]; + *cnt = rammap[4]; + *len = rammap[3]; + + freq /= 1000; + for (i = 0; i < rammap[5]; i++) { + if (freq >= ROM16(ramcfg[0]) && + freq <= ROM16(ramcfg[2])) + return ramcfg; + + ramcfg += *hdr + (*cnt * *len); + } + } + + return NULL; + } + + if (dev_priv->chipset == 0x49 || + dev_priv->chipset == 0x4b) + freq /= 2; + + while ((perf = nouveau_perf_entry(dev, i++, ver, hdr, cnt, len))) { + if (*ver >= 0x20 && *ver < 0x25) { + if (perf[0] != 0xff && freq <= ROM16(perf[11]) * 1000) + break; + } else + if (*ver >= 0x25 && *ver < 0x40) { + if (perf[0] != 0xff && freq <= ROM16(perf[12]) * 1000) + break; + } + } + + if (perf) { + u8 *ramcfg = perf + *hdr; + *ver = 0x00; + *hdr = 0; + return ramcfg; + } + + return NULL; +} + +u8 * +nouveau_perf_ramcfg(struct drm_device *dev, u32 freq, u8 *ver, u8 *len) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nvbios *bios = &dev_priv->vbios; + u8 strap, hdr, cnt; + u8 *rammap; + + strap = (nv_rd32(dev, 0x101000) & 0x0000003c) >> 2; + if (bios->ram_restrict_tbl_ptr) + strap = bios->data[bios->ram_restrict_tbl_ptr + strap]; + + rammap = nouveau_perf_rammap(dev, freq, ver, &hdr, &cnt, len); + if (rammap && strap < cnt) + return rammap + hdr + (strap * *len); + + return NULL; +} + +u8 * +nouveau_perf_timing(struct drm_device *dev, u32 freq, u8 *ver, u8 *len) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nvbios *bios = &dev_priv->vbios; + struct bit_entry P; + u8 *perf, *timing = NULL; + u8 i = 0, hdr, cnt; + + if (bios->type == NVBIOS_BMP) { + while ((perf = nouveau_perf_entry(dev, i++, ver, &hdr, &cnt, + len)) && *ver == 0x15) { + if (freq <= ROM32(perf[5]) * 20) { + *ver = 0x00; + *len = 14; + return perf + 41; + } + } + return NULL; + } + + if (!bit_table(dev, 'P', &P)) { + if (P.version == 1) + timing = ROMPTR(dev, P.data[4]); + else + if (P.version == 2) + timing = ROMPTR(dev, P.data[8]); + } + + if (timing && timing[0] == 0x10) { + u8 *ramcfg = nouveau_perf_ramcfg(dev, freq, ver, len); + if (ramcfg && ramcfg[1] < timing[2]) { + *ver = timing[0]; + *len = timing[3]; + return timing + timing[1] + (ramcfg[1] * timing[3]); + } + } + + return NULL; +} + static void legacy_perf_init(struct drm_device *dev) { @@ -41,7 +213,7 @@ legacy_perf_init(struct drm_device *dev) return; } - perf = ROMPTR(bios, bmp[0x73]); + perf = ROMPTR(dev, bmp[0x73]); if (!perf) { NV_DEBUG(dev, "No memclock table pointer found.\n"); return; @@ -72,75 +244,11 @@ legacy_perf_init(struct drm_device *dev) pm->nr_perflvl = 1; } -static struct nouveau_pm_memtiming * -nouveau_perf_timing(struct drm_device *dev, struct bit_entry *P, - u16 memclk, u8 *entry, u8 recordlen, u8 entries) -{ - struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_pm_engine *pm = &dev_priv->engine.pm; - struct nvbios *bios = &dev_priv->vbios; - u8 ramcfg; - int i; - - /* perf v2 has a separate "timing map" table, we have to match - * the target memory clock to a specific entry, *then* use - * ramcfg to select the correct subentry - */ - if (P->version == 2) { - u8 *tmap = ROMPTR(bios, P->data[4]); - if (!tmap) { - NV_DEBUG(dev, "no timing map pointer\n"); - return NULL; - } - - if (tmap[0] != 0x10) { - NV_WARN(dev, "timing map 0x%02x unknown\n", tmap[0]); - return NULL; - } - - entry = tmap + tmap[1]; - recordlen = tmap[2] + (tmap[4] * tmap[3]); - for (i = 0; i < tmap[5]; i++, entry += recordlen) { - if (memclk >= ROM16(entry[0]) && - memclk <= ROM16(entry[2])) - break; - } - - if (i == tmap[5]) { - NV_WARN(dev, "no match in timing map table\n"); - return NULL; - } - - entry += tmap[2]; - recordlen = tmap[3]; - entries = tmap[4]; - } - - ramcfg = (nv_rd32(dev, NV_PEXTDEV_BOOT_0) & 0x0000003c) >> 2; - if (bios->ram_restrict_tbl_ptr) - ramcfg = bios->data[bios->ram_restrict_tbl_ptr + ramcfg]; - - if (ramcfg >= entries) { - NV_WARN(dev, "ramcfg strap out of bounds!\n"); - return NULL; - } - - entry += ramcfg * recordlen; - if (entry[1] >= pm->memtimings.nr_timing) { - if (entry[1] != 0xff) - NV_WARN(dev, "timingset %d does not exist\n", entry[1]); - return NULL; - } - - return &pm->memtimings.timing[entry[1]]; -} - static void -nouveau_perf_voltage(struct drm_device *dev, struct bit_entry *P, - struct nouveau_pm_level *perflvl) +nouveau_perf_voltage(struct drm_device *dev, struct nouveau_pm_level *perflvl) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nvbios *bios = &dev_priv->vbios; + struct bit_entry P; u8 *vmap; int id; @@ -159,13 +267,13 @@ nouveau_perf_voltage(struct drm_device *dev, struct bit_entry *P, /* on newer ones, the perflvl stores an index into yet another * vbios table containing a min/max voltage value for the perflvl */ - if (P->version != 2 || P->length < 34) { + if (bit_table(dev, 'P', &P) || P.version != 2 || P.length < 34) { NV_DEBUG(dev, "where's our volt map table ptr? %d %d\n", - P->version, P->length); + P.version, P.length); return; } - vmap = ROMPTR(bios, P->data[32]); + vmap = ROMPTR(dev, P.data[32]); if (!vmap) { NV_DEBUG(dev, "volt map table pointer invalid\n"); return; @@ -184,129 +292,70 @@ nouveau_perf_init(struct drm_device *dev) struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_pm_engine *pm = &dev_priv->engine.pm; struct nvbios *bios = &dev_priv->vbios; - struct bit_entry P; - struct nouveau_pm_memtimings *memtimings = &pm->memtimings; - struct nouveau_pm_tbl_header mt_hdr; - u8 version, headerlen, recordlen, entries; - u8 *perf, *entry; - int vid, i; - - if (bios->type == NVBIOS_BIT) { - if (bit_table(dev, 'P', &P)) - return; - - if (P.version != 1 && P.version != 2) { - NV_WARN(dev, "unknown perf for BIT P %d\n", P.version); - return; - } - - perf = ROMPTR(bios, P.data[0]); - version = perf[0]; - headerlen = perf[1]; - if (version < 0x40) { - recordlen = perf[3] + (perf[4] * perf[5]); - entries = perf[2]; - } else { - recordlen = perf[2] + (perf[3] * perf[4]); - entries = perf[5]; - } - } else { - if (bios->data[bios->offset + 6] < 0x25) { - legacy_perf_init(dev); - return; - } - - perf = ROMPTR(bios, bios->data[bios->offset + 0x94]); - if (!perf) { - NV_DEBUG(dev, "perf table pointer invalid\n"); - return; - } - - version = perf[1]; - headerlen = perf[0]; - recordlen = perf[3]; - entries = perf[2]; - } + u8 *perf, ver, hdr, cnt, len; + int ret, vid, i = -1; - if (entries > NOUVEAU_PM_MAX_LEVEL) { - NV_DEBUG(dev, "perf table has too many entries - buggy vbios?\n"); - entries = NOUVEAU_PM_MAX_LEVEL; + if (bios->type == NVBIOS_BMP && bios->data[bios->offset + 6] < 0x25) { + legacy_perf_init(dev); + return; } - entry = perf + headerlen; - - /* For version 0x15, initialize memtiming table */ - if(version == 0x15) { - memtimings->timing = - kcalloc(entries, sizeof(*memtimings->timing), GFP_KERNEL); - if (!memtimings->timing) { - NV_WARN(dev,"Could not allocate memtiming table\n"); - return; - } + perf = nouveau_perf_table(dev, &ver); + if (ver >= 0x20 && ver < 0x40) + pm->fan.pwm_divisor = ROM16(perf[6]); - mt_hdr.entry_cnt = entries; - mt_hdr.entry_len = 14; - mt_hdr.version = version; - mt_hdr.header_len = 4; - } - - for (i = 0; i < entries; i++) { + while ((perf = nouveau_perf_entry(dev, ++i, &ver, &hdr, &cnt, &len))) { struct nouveau_pm_level *perflvl = &pm->perflvl[pm->nr_perflvl]; - perflvl->timing = NULL; - - if (entry[0] == 0xff) { - entry += recordlen; + if (perf[0] == 0xff) continue; - } - switch (version) { + switch (ver) { case 0x12: case 0x13: case 0x15: - perflvl->fanspeed = entry[55]; - if (recordlen > 56) - perflvl->volt_min = entry[56]; - perflvl->core = ROM32(entry[1]) * 10; - perflvl->memory = ROM32(entry[5]) * 20; + perflvl->fanspeed = perf[55]; + if (hdr > 56) + perflvl->volt_min = perf[56]; + perflvl->core = ROM32(perf[1]) * 10; + perflvl->memory = ROM32(perf[5]) * 20; break; case 0x21: case 0x23: case 0x24: - perflvl->fanspeed = entry[4]; - perflvl->volt_min = entry[5]; - perflvl->shader = ROM16(entry[6]) * 1000; + perflvl->fanspeed = perf[4]; + perflvl->volt_min = perf[5]; + perflvl->shader = ROM16(perf[6]) * 1000; perflvl->core = perflvl->shader; - perflvl->core += (signed char)entry[8] * 1000; + perflvl->core += (signed char)perf[8] * 1000; if (dev_priv->chipset == 0x49 || dev_priv->chipset == 0x4b) - perflvl->memory = ROM16(entry[11]) * 1000; + perflvl->memory = ROM16(perf[11]) * 1000; else - perflvl->memory = ROM16(entry[11]) * 2000; - + perflvl->memory = ROM16(perf[11]) * 2000; break; case 0x25: - perflvl->fanspeed = entry[4]; - perflvl->volt_min = entry[5]; - perflvl->core = ROM16(entry[6]) * 1000; - perflvl->shader = ROM16(entry[10]) * 1000; - perflvl->memory = ROM16(entry[12]) * 1000; + perflvl->fanspeed = perf[4]; + perflvl->volt_min = perf[5]; + perflvl->core = ROM16(perf[6]) * 1000; + perflvl->shader = ROM16(perf[10]) * 1000; + perflvl->memory = ROM16(perf[12]) * 1000; break; case 0x30: - perflvl->memscript = ROM16(entry[2]); + perflvl->memscript = ROM16(perf[2]); case 0x35: - perflvl->fanspeed = entry[6]; - perflvl->volt_min = entry[7]; - perflvl->core = ROM16(entry[8]) * 1000; - perflvl->shader = ROM16(entry[10]) * 1000; - perflvl->memory = ROM16(entry[12]) * 1000; - /*XXX: confirm on 0x35 */ - perflvl->unk05 = ROM16(entry[16]) * 1000; + perflvl->fanspeed = perf[6]; + perflvl->volt_min = perf[7]; + perflvl->core = ROM16(perf[8]) * 1000; + perflvl->shader = ROM16(perf[10]) * 1000; + perflvl->memory = ROM16(perf[12]) * 1000; + perflvl->vdec = ROM16(perf[16]) * 1000; + perflvl->dom6 = ROM16(perf[20]) * 1000; break; case 0x40: -#define subent(n) (ROM16(entry[perf[2] + ((n) * perf[3])]) & 0xfff) * 1000 +#define subent(n) ((ROM16(perf[hdr + (n) * len]) & 0xfff) * 1000) perflvl->fanspeed = 0; /*XXX*/ - perflvl->volt_min = entry[2]; + perflvl->volt_min = perf[2]; if (dev_priv->card_type == NV_50) { perflvl->core = subent(0); perflvl->shader = subent(1); @@ -329,36 +378,34 @@ nouveau_perf_init(struct drm_device *dev) } /* make sure vid is valid */ - nouveau_perf_voltage(dev, &P, perflvl); + nouveau_perf_voltage(dev, perflvl); if (pm->voltage.supported && perflvl->volt_min) { vid = nouveau_volt_vid_lookup(dev, perflvl->volt_min); if (vid < 0) { - NV_DEBUG(dev, "drop perflvl %d, bad vid\n", i); - entry += recordlen; + NV_DEBUG(dev, "perflvl %d, bad vid\n", i); continue; } } /* get the corresponding memory timings */ - if (version == 0x15) { - memtimings->timing[i].id = i; - nv30_mem_timing_entry(dev,&mt_hdr,(struct nouveau_pm_tbl_entry*) &entry[41],0,&memtimings->timing[i]); - perflvl->timing = &memtimings->timing[i]; - } else if (version > 0x15) { - /* last 3 args are for < 0x40, ignored for >= 0x40 */ - perflvl->timing = - nouveau_perf_timing(dev, &P, - perflvl->memory / 1000, - entry + perf[3], - perf[5], perf[4]); + ret = nouveau_mem_timing_calc(dev, perflvl->memory, + &perflvl->timing); + if (ret) { + NV_DEBUG(dev, "perflvl %d, bad timing: %d\n", i, ret); + continue; } snprintf(perflvl->name, sizeof(perflvl->name), "performance_level_%d", i); perflvl->id = i; - pm->nr_perflvl++; - entry += recordlen; + snprintf(perflvl->profile.name, sizeof(perflvl->profile.name), + "%d", perflvl->id); + perflvl->profile.func = &nouveau_pm_static_profile_func; + list_add_tail(&perflvl->profile.head, &pm->profiles); + + + pm->nr_perflvl++; } } diff --git a/drivers/gpu/drm/nouveau/nouveau_pm.c b/drivers/gpu/drm/nouveau/nouveau_pm.c index a539fd2..da3e7c3 100644 --- a/drivers/gpu/drm/nouveau/nouveau_pm.c +++ b/drivers/gpu/drm/nouveau/nouveau_pm.c @@ -26,6 +26,7 @@ #include "nouveau_drv.h" #include "nouveau_pm.h" +#include "nouveau_gpio.h" #ifdef CONFIG_ACPI #include @@ -35,22 +36,98 @@ #include static int -nouveau_pm_clock_set(struct drm_device *dev, struct nouveau_pm_level *perflvl, - u8 id, u32 khz) +nouveau_pwmfan_get(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_pm_engine *pm = &dev_priv->engine.pm; - void *pre_state; + struct gpio_func gpio; + u32 divs, duty; + int ret; - if (khz == 0) - return 0; + if (!pm->pwm_get) + return -ENODEV; + + ret = nouveau_gpio_find(dev, 0, DCB_GPIO_PWM_FAN, 0xff, &gpio); + if (ret == 0) { + ret = pm->pwm_get(dev, gpio.line, &divs, &duty); + if (ret == 0 && divs) { + divs = max(divs, duty); + if (dev_priv->card_type <= NV_40 || (gpio.log[0] & 1)) + duty = divs - duty; + return (duty * 100) / divs; + } + + return nouveau_gpio_func_get(dev, gpio.func) * 100; + } + + return -ENODEV; +} + +static int +nouveau_pwmfan_set(struct drm_device *dev, int percent) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_pm_engine *pm = &dev_priv->engine.pm; + struct gpio_func gpio; + u32 divs, duty; + int ret; + + if (!pm->pwm_set) + return -ENODEV; + + ret = nouveau_gpio_find(dev, 0, DCB_GPIO_PWM_FAN, 0xff, &gpio); + if (ret == 0) { + divs = pm->fan.pwm_divisor; + if (pm->fan.pwm_freq) { + /*XXX: PNVIO clock more than likely... */ + divs = 135000 / pm->fan.pwm_freq; + if (dev_priv->chipset < 0xa3) + divs /= 4; + } + + duty = ((divs * percent) + 99) / 100; + if (dev_priv->card_type <= NV_40 || (gpio.log[0] & 1)) + duty = divs - duty; + + ret = pm->pwm_set(dev, gpio.line, divs, duty); + if (!ret) + pm->fan.percent = percent; + return ret; + } + + return -ENODEV; +} + +static int +nouveau_pm_perflvl_aux(struct drm_device *dev, struct nouveau_pm_level *perflvl, + struct nouveau_pm_level *a, struct nouveau_pm_level *b) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_pm_engine *pm = &dev_priv->engine.pm; + int ret; - pre_state = pm->clock_pre(dev, perflvl, id, khz); - if (IS_ERR(pre_state)) - return PTR_ERR(pre_state); + /*XXX: not on all boards, we should control based on temperature + * on recent boards.. or maybe on some other factor we don't + * know about? + */ + if (a->fanspeed && b->fanspeed && b->fanspeed > a->fanspeed) { + ret = nouveau_pwmfan_set(dev, perflvl->fanspeed); + if (ret && ret != -ENODEV) { + NV_ERROR(dev, "fanspeed set failed: %d\n", ret); + return ret; + } + } + + if (pm->voltage.supported && pm->voltage_set) { + if (perflvl->volt_min && b->volt_min > a->volt_min) { + ret = pm->voltage_set(dev, perflvl->volt_min); + if (ret) { + NV_ERROR(dev, "voltage set failed: %d\n", ret); + return ret; + } + } + } - if (pre_state) - pm->clock_set(dev, pre_state); return 0; } @@ -59,34 +136,90 @@ nouveau_pm_perflvl_set(struct drm_device *dev, struct nouveau_pm_level *perflvl) { struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_pm_engine *pm = &dev_priv->engine.pm; + void *state; int ret; if (perflvl == pm->cur) return 0; - if (pm->voltage.supported && pm->voltage_set && perflvl->volt_min) { - ret = pm->voltage_set(dev, perflvl->volt_min); - if (ret) { - NV_ERROR(dev, "voltage_set %d failed: %d\n", - perflvl->volt_min, ret); - } - } + ret = nouveau_pm_perflvl_aux(dev, perflvl, pm->cur, perflvl); + if (ret) + return ret; - if (pm->clocks_pre) { - void *state = pm->clocks_pre(dev, perflvl); - if (IS_ERR(state)) - return PTR_ERR(state); - pm->clocks_set(dev, state); - } else - if (pm->clock_set) { - nouveau_pm_clock_set(dev, perflvl, PLL_CORE, perflvl->core); - nouveau_pm_clock_set(dev, perflvl, PLL_SHADER, perflvl->shader); - nouveau_pm_clock_set(dev, perflvl, PLL_MEMORY, perflvl->memory); - nouveau_pm_clock_set(dev, perflvl, PLL_UNK05, perflvl->unk05); + state = pm->clocks_pre(dev, perflvl); + if (IS_ERR(state)) { + ret = PTR_ERR(state); + goto error; } + ret = pm->clocks_set(dev, state); + if (ret) + goto error; + + ret = nouveau_pm_perflvl_aux(dev, perflvl, perflvl, pm->cur); + if (ret) + return ret; pm->cur = perflvl; return 0; + +error: + /* restore the fan speed and voltage before leaving */ + nouveau_pm_perflvl_aux(dev, perflvl, perflvl, pm->cur); + return ret; +} + +void +nouveau_pm_trigger(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_pm_engine *pm = &dev_priv->engine.pm; + struct nouveau_pm_profile *profile = NULL; + struct nouveau_pm_level *perflvl = NULL; + int ret; + + /* select power profile based on current power source */ + if (power_supply_is_system_supplied()) + profile = pm->profile_ac; + else + profile = pm->profile_dc; + + if (profile != pm->profile) { + pm->profile->func->fini(pm->profile); + pm->profile = profile; + pm->profile->func->init(pm->profile); + } + + /* select performance level based on profile */ + perflvl = profile->func->select(profile); + + /* change perflvl, if necessary */ + if (perflvl != pm->cur) { + struct nouveau_timer_engine *ptimer = &dev_priv->engine.timer; + u64 time0 = ptimer->read(dev); + + NV_INFO(dev, "setting performance level: %d", perflvl->id); + ret = nouveau_pm_perflvl_set(dev, perflvl); + if (ret) + NV_INFO(dev, "> reclocking failed: %d\n\n", ret); + + NV_INFO(dev, "> reclocking took %lluns\n\n", + ptimer->read(dev) - time0); + } +} + +static struct nouveau_pm_profile * +profile_find(struct drm_device *dev, const char *string) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_pm_engine *pm = &dev_priv->engine.pm; + struct nouveau_pm_profile *profile; + + list_for_each_entry(profile, &pm->profiles, head) { + if (!strncmp(profile->name, string, sizeof(profile->name))) + return profile; + } + + return NULL; } static int @@ -94,33 +227,55 @@ nouveau_pm_profile_set(struct drm_device *dev, const char *profile) { struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_pm_engine *pm = &dev_priv->engine.pm; - struct nouveau_pm_level *perflvl = NULL; + struct nouveau_pm_profile *ac = NULL, *dc = NULL; + char string[16], *cur = string, *ptr; /* safety precaution, for now */ if (nouveau_perflvl_wr != 7777) return -EPERM; - if (!strncmp(profile, "boot", 4)) - perflvl = &pm->boot; - else { - int pl = simple_strtol(profile, NULL, 10); - int i; + strncpy(string, profile, sizeof(string)); + string[sizeof(string) - 1] = 0; + if ((ptr = strchr(string, '\n'))) + *ptr = '\0'; - for (i = 0; i < pm->nr_perflvl; i++) { - if (pm->perflvl[i].id == pl) { - perflvl = &pm->perflvl[i]; - break; - } - } + ptr = strsep(&cur, ","); + if (ptr) + ac = profile_find(dev, ptr); - if (!perflvl) - return -EINVAL; - } + ptr = strsep(&cur, ","); + if (ptr) + dc = profile_find(dev, ptr); + else + dc = ac; + + if (ac == NULL || dc == NULL) + return -EINVAL; - NV_INFO(dev, "setting performance level: %s\n", profile); - return nouveau_pm_perflvl_set(dev, perflvl); + pm->profile_ac = ac; + pm->profile_dc = dc; + nouveau_pm_trigger(dev); + return 0; } +static void +nouveau_pm_static_dummy(struct nouveau_pm_profile *profile) +{ +} + +static struct nouveau_pm_level * +nouveau_pm_static_select(struct nouveau_pm_profile *profile) +{ + return container_of(profile, struct nouveau_pm_level, profile); +} + +const struct nouveau_pm_profile_func nouveau_pm_static_profile_func = { + .destroy = nouveau_pm_static_dummy, + .init = nouveau_pm_static_dummy, + .fini = nouveau_pm_static_dummy, + .select = nouveau_pm_static_select, +}; + static int nouveau_pm_perflvl_get(struct drm_device *dev, struct nouveau_pm_level *perflvl) { @@ -134,23 +289,6 @@ nouveau_pm_perflvl_get(struct drm_device *dev, struct nouveau_pm_level *perflvl) ret = pm->clocks_get(dev, perflvl); if (ret) return ret; - } else - if (pm->clock_get) { - ret = pm->clock_get(dev, PLL_CORE); - if (ret > 0) - perflvl->core = ret; - - ret = pm->clock_get(dev, PLL_MEMORY); - if (ret > 0) - perflvl->memory = ret; - - ret = pm->clock_get(dev, PLL_SHADER); - if (ret > 0) - perflvl->shader = ret; - - ret = pm->clock_get(dev, PLL_UNK05); - if (ret > 0) - perflvl->unk05 = ret; } if (pm->voltage.supported && pm->voltage_get) { @@ -161,13 +299,18 @@ nouveau_pm_perflvl_get(struct drm_device *dev, struct nouveau_pm_level *perflvl) } } + ret = nouveau_pwmfan_get(dev); + if (ret > 0) + perflvl->fanspeed = ret; + + nouveau_mem_timing_read(dev, &perflvl->timing); return 0; } static void nouveau_pm_perflvl_info(struct nouveau_pm_level *perflvl, char *ptr, int len) { - char c[16], s[16], v[32], f[16], t[16], m[16]; + char c[16], s[16], v[32], f[16], m[16]; c[0] = '\0'; if (perflvl->core) @@ -195,18 +338,15 @@ nouveau_pm_perflvl_info(struct nouveau_pm_level *perflvl, char *ptr, int len) if (perflvl->fanspeed) snprintf(f, sizeof(f), " fanspeed %d%%", perflvl->fanspeed); - t[0] = '\0'; - if (perflvl->timing) - snprintf(t, sizeof(t), " timing %d", perflvl->timing->id); - - snprintf(ptr, len, "%s%s%s%s%s%s\n", c, s, m, t, v, f); + snprintf(ptr, len, "%s%s%s%s%s\n", c, s, m, v, f); } static ssize_t nouveau_pm_get_perflvl_info(struct device *d, struct device_attribute *a, char *buf) { - struct nouveau_pm_level *perflvl = (struct nouveau_pm_level *)a; + struct nouveau_pm_level *perflvl = + container_of(a, struct nouveau_pm_level, dev_attr); char *ptr = buf; int len = PAGE_SIZE; @@ -228,12 +368,8 @@ nouveau_pm_get_perflvl(struct device *d, struct device_attribute *a, char *buf) int len = PAGE_SIZE, ret; char *ptr = buf; - if (!pm->cur) - snprintf(ptr, len, "setting: boot\n"); - else if (pm->cur == &pm->boot) - snprintf(ptr, len, "setting: boot\nc:"); - else - snprintf(ptr, len, "setting: static %d\nc:", pm->cur->id); + snprintf(ptr, len, "profile: %s, %s\nc:", + pm->profile_ac->name, pm->profile_dc->name); ptr += strlen(buf); len -= strlen(buf); @@ -345,7 +481,7 @@ nouveau_hwmon_set_max_temp(struct device *d, struct device_attribute *a, struct nouveau_pm_threshold_temp *temp = &pm->threshold_temp; long value; - if (strict_strtol(buf, 10, &value) == -EINVAL) + if (kstrtol(buf, 10, &value) == -EINVAL) return count; temp->down_clock = value/1000; @@ -380,7 +516,7 @@ nouveau_hwmon_set_critical_temp(struct device *d, struct device_attribute *a, struct nouveau_pm_threshold_temp *temp = &pm->threshold_temp; long value; - if (strict_strtol(buf, 10, &value) == -EINVAL) + if (kstrtol(buf, 10, &value) == -EINVAL) return count; temp->critical = value/1000; @@ -412,6 +548,172 @@ static SENSOR_DEVICE_ATTR(update_rate, S_IRUGO, nouveau_hwmon_show_update_rate, NULL, 0); +static ssize_t +nouveau_hwmon_show_fan0_input(struct device *d, struct device_attribute *attr, + char *buf) +{ + struct drm_device *dev = dev_get_drvdata(d); + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_timer_engine *ptimer = &dev_priv->engine.timer; + struct gpio_func gpio; + u32 cycles, cur, prev; + u64 start; + int ret; + + ret = nouveau_gpio_find(dev, 0, DCB_GPIO_FAN_SENSE, 0xff, &gpio); + if (ret) + return ret; + + /* Monitor the GPIO input 0x3b for 250ms. + * When the fan spins, it changes the value of GPIO FAN_SENSE. + * We get 4 changes (0 -> 1 -> 0 -> 1 -> [...]) per complete rotation. + */ + start = ptimer->read(dev); + prev = nouveau_gpio_sense(dev, 0, gpio.line); + cycles = 0; + do { + cur = nouveau_gpio_sense(dev, 0, gpio.line); + if (prev != cur) { + cycles++; + prev = cur; + } + + usleep_range(500, 1000); /* supports 0 < rpm < 7500 */ + } while (ptimer->read(dev) - start < 250000000); + + /* interpolate to get rpm */ + return sprintf(buf, "%i\n", cycles / 4 * 4 * 60); +} +static SENSOR_DEVICE_ATTR(fan0_input, S_IRUGO, nouveau_hwmon_show_fan0_input, + NULL, 0); + +static ssize_t +nouveau_hwmon_get_pwm0(struct device *d, struct device_attribute *a, char *buf) +{ + struct drm_device *dev = dev_get_drvdata(d); + int ret; + + ret = nouveau_pwmfan_get(dev); + if (ret < 0) + return ret; + + return sprintf(buf, "%i\n", ret); +} + +static ssize_t +nouveau_hwmon_set_pwm0(struct device *d, struct device_attribute *a, + const char *buf, size_t count) +{ + struct drm_device *dev = dev_get_drvdata(d); + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_pm_engine *pm = &dev_priv->engine.pm; + int ret = -ENODEV; + long value; + + if (nouveau_perflvl_wr != 7777) + return -EPERM; + + if (kstrtol(buf, 10, &value) == -EINVAL) + return -EINVAL; + + if (value < pm->fan.min_duty) + value = pm->fan.min_duty; + if (value > pm->fan.max_duty) + value = pm->fan.max_duty; + + ret = nouveau_pwmfan_set(dev, value); + if (ret) + return ret; + + return count; +} + +static SENSOR_DEVICE_ATTR(pwm0, S_IRUGO | S_IWUSR, + nouveau_hwmon_get_pwm0, + nouveau_hwmon_set_pwm0, 0); + +static ssize_t +nouveau_hwmon_get_pwm0_min(struct device *d, + struct device_attribute *a, char *buf) +{ + struct drm_device *dev = dev_get_drvdata(d); + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_pm_engine *pm = &dev_priv->engine.pm; + + return sprintf(buf, "%i\n", pm->fan.min_duty); +} + +static ssize_t +nouveau_hwmon_set_pwm0_min(struct device *d, struct device_attribute *a, + const char *buf, size_t count) +{ + struct drm_device *dev = dev_get_drvdata(d); + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_pm_engine *pm = &dev_priv->engine.pm; + long value; + + if (kstrtol(buf, 10, &value) == -EINVAL) + return -EINVAL; + + if (value < 0) + value = 0; + + if (pm->fan.max_duty - value < 10) + value = pm->fan.max_duty - 10; + + if (value < 10) + pm->fan.min_duty = 10; + else + pm->fan.min_duty = value; + + return count; +} + +static SENSOR_DEVICE_ATTR(pwm0_min, S_IRUGO | S_IWUSR, + nouveau_hwmon_get_pwm0_min, + nouveau_hwmon_set_pwm0_min, 0); + +static ssize_t +nouveau_hwmon_get_pwm0_max(struct device *d, + struct device_attribute *a, char *buf) +{ + struct drm_device *dev = dev_get_drvdata(d); + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_pm_engine *pm = &dev_priv->engine.pm; + + return sprintf(buf, "%i\n", pm->fan.max_duty); +} + +static ssize_t +nouveau_hwmon_set_pwm0_max(struct device *d, struct device_attribute *a, + const char *buf, size_t count) +{ + struct drm_device *dev = dev_get_drvdata(d); + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_pm_engine *pm = &dev_priv->engine.pm; + long value; + + if (kstrtol(buf, 10, &value) == -EINVAL) + return -EINVAL; + + if (value < 0) + value = 0; + + if (value - pm->fan.min_duty < 10) + value = pm->fan.min_duty + 10; + + if (value > 100) + pm->fan.max_duty = 100; + else + pm->fan.max_duty = value; + + return count; +} + +static SENSOR_DEVICE_ATTR(pwm0_max, S_IRUGO | S_IWUSR, + nouveau_hwmon_get_pwm0_max, + nouveau_hwmon_set_pwm0_max, 0); + static struct attribute *hwmon_attributes[] = { &sensor_dev_attr_temp1_input.dev_attr.attr, &sensor_dev_attr_temp1_max.dev_attr.attr, @@ -420,20 +722,36 @@ static struct attribute *hwmon_attributes[] = { &sensor_dev_attr_update_rate.dev_attr.attr, NULL }; +static struct attribute *hwmon_fan_rpm_attributes[] = { + &sensor_dev_attr_fan0_input.dev_attr.attr, + NULL +}; +static struct attribute *hwmon_pwm_fan_attributes[] = { + &sensor_dev_attr_pwm0.dev_attr.attr, + &sensor_dev_attr_pwm0_min.dev_attr.attr, + &sensor_dev_attr_pwm0_max.dev_attr.attr, + NULL +}; static const struct attribute_group hwmon_attrgroup = { .attrs = hwmon_attributes, }; +static const struct attribute_group hwmon_fan_rpm_attrgroup = { + .attrs = hwmon_fan_rpm_attributes, +}; +static const struct attribute_group hwmon_pwm_fan_attrgroup = { + .attrs = hwmon_pwm_fan_attributes, +}; #endif static int nouveau_hwmon_init(struct drm_device *dev) { -#if defined(CONFIG_HWMON) || (defined(MODULE) && defined(CONFIG_HWMON_MODULE)) struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_pm_engine *pm = &dev_priv->engine.pm; +#if defined(CONFIG_HWMON) || (defined(MODULE) && defined(CONFIG_HWMON_MODULE)) struct device *hwmon_dev; - int ret; + int ret = 0; if (!pm->temp_get) return -ENODEV; @@ -446,17 +764,46 @@ nouveau_hwmon_init(struct drm_device *dev) return ret; } dev_set_drvdata(hwmon_dev, dev); + + /* default sysfs entries */ ret = sysfs_create_group(&dev->pdev->dev.kobj, &hwmon_attrgroup); if (ret) { - NV_ERROR(dev, - "Unable to create hwmon sysfs file: %d\n", ret); - hwmon_device_unregister(hwmon_dev); - return ret; + if (ret) + goto error; + } + + /* if the card has a pwm fan */ + /*XXX: incorrect, need better detection for this, some boards have + * the gpio entries for pwm fan control even when there's no + * actual fan connected to it... therm table? */ + if (nouveau_pwmfan_get(dev) >= 0) { + ret = sysfs_create_group(&dev->pdev->dev.kobj, + &hwmon_pwm_fan_attrgroup); + if (ret) + goto error; + } + + /* if the card can read the fan rpm */ + if (nouveau_gpio_func_valid(dev, DCB_GPIO_FAN_SENSE)) { + ret = sysfs_create_group(&dev->pdev->dev.kobj, + &hwmon_fan_rpm_attrgroup); + if (ret) + goto error; } pm->hwmon = hwmon_dev; -#endif + + return 0; + +error: + NV_ERROR(dev, "Unable to create some hwmon sysfs files: %d\n", ret); + hwmon_device_unregister(hwmon_dev); + pm->hwmon = NULL; + return ret; +#else + pm->hwmon = NULL; return 0; +#endif } static void @@ -468,6 +815,11 @@ nouveau_hwmon_fini(struct drm_device *dev) if (pm->hwmon) { sysfs_remove_group(&dev->pdev->dev.kobj, &hwmon_attrgroup); + sysfs_remove_group(&dev->pdev->dev.kobj, + &hwmon_pwm_fan_attrgroup); + sysfs_remove_group(&dev->pdev->dev.kobj, + &hwmon_fan_rpm_attrgroup); + hwmon_device_unregister(pm->hwmon); } #endif @@ -486,6 +838,7 @@ nouveau_pm_acpi_event(struct notifier_block *nb, unsigned long val, void *data) bool ac = power_supply_is_system_supplied(); NV_DEBUG(dev, "power supply changed: %s\n", ac ? "AC" : "DC"); + nouveau_pm_trigger(dev); } return NOTIFY_OK; @@ -500,35 +853,48 @@ nouveau_pm_init(struct drm_device *dev) char info[256]; int ret, i; - nouveau_mem_timing_init(dev); + /* parse aux tables from vbios */ nouveau_volt_init(dev); - nouveau_perf_init(dev); nouveau_temp_init(dev); + /* determine current ("boot") performance level */ + ret = nouveau_pm_perflvl_get(dev, &pm->boot); + if (ret) { + NV_ERROR(dev, "failed to determine boot perflvl\n"); + return ret; + } + + strncpy(pm->boot.name, "boot", 4); + strncpy(pm->boot.profile.name, "boot", 4); + pm->boot.profile.func = &nouveau_pm_static_profile_func; + + INIT_LIST_HEAD(&pm->profiles); + list_add(&pm->boot.profile.head, &pm->profiles); + + pm->profile_ac = &pm->boot.profile; + pm->profile_dc = &pm->boot.profile; + pm->profile = &pm->boot.profile; + pm->cur = &pm->boot; + + /* add performance levels from vbios */ + nouveau_perf_init(dev); + + /* display available performance levels */ NV_INFO(dev, "%d available performance level(s)\n", pm->nr_perflvl); for (i = 0; i < pm->nr_perflvl; i++) { nouveau_pm_perflvl_info(&pm->perflvl[i], info, sizeof(info)); NV_INFO(dev, "%d:%s", pm->perflvl[i].id, info); } - /* determine current ("boot") performance level */ - ret = nouveau_pm_perflvl_get(dev, &pm->boot); - if (ret == 0) { - strncpy(pm->boot.name, "boot", 4); - pm->cur = &pm->boot; - - nouveau_pm_perflvl_info(&pm->boot, info, sizeof(info)); - NV_INFO(dev, "c:%s", info); - } + nouveau_pm_perflvl_info(&pm->boot, info, sizeof(info)); + NV_INFO(dev, "c:%s", info); /* switch performance levels now if requested */ - if (nouveau_perflvl != NULL) { - ret = nouveau_pm_profile_set(dev, nouveau_perflvl); - if (ret) { - NV_ERROR(dev, "error setting perflvl \"%s\": %d\n", - nouveau_perflvl, ret); - } - } + if (nouveau_perflvl != NULL) + nouveau_pm_profile_set(dev, nouveau_perflvl); + + /* determine the current fan speed */ + pm->fan.percent = nouveau_pwmfan_get(dev); nouveau_sysfs_init(dev); nouveau_hwmon_init(dev); @@ -545,6 +911,12 @@ nouveau_pm_fini(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_pm_engine *pm = &dev_priv->engine.pm; + struct nouveau_pm_profile *profile, *tmp; + + list_for_each_entry_safe(profile, tmp, &pm->profiles, head) { + list_del(&profile->head); + profile->func->destroy(profile); + } if (pm->cur != &pm->boot) nouveau_pm_perflvl_set(dev, &pm->boot); @@ -552,7 +924,6 @@ nouveau_pm_fini(struct drm_device *dev) nouveau_temp_fini(dev); nouveau_perf_fini(dev); nouveau_volt_fini(dev); - nouveau_mem_timing_fini(dev); #if defined(CONFIG_ACPI) && defined(CONFIG_POWER_SUPPLY) unregister_acpi_notifier(&pm->acpi_nb); @@ -574,4 +945,5 @@ nouveau_pm_resume(struct drm_device *dev) perflvl = pm->cur; pm->cur = &pm->boot; nouveau_pm_perflvl_set(dev, perflvl); + nouveau_pwmfan_set(dev, pm->fan.percent); } diff --git a/drivers/gpu/drm/nouveau/nouveau_pm.h b/drivers/gpu/drm/nouveau/nouveau_pm.h index 8ac02cd..3f82dfe 100644 --- a/drivers/gpu/drm/nouveau/nouveau_pm.h +++ b/drivers/gpu/drm/nouveau/nouveau_pm.h @@ -25,10 +25,30 @@ #ifndef __NOUVEAU_PM_H__ #define __NOUVEAU_PM_H__ +struct nouveau_mem_exec_func { + struct drm_device *dev; + void (*precharge)(struct nouveau_mem_exec_func *); + void (*refresh)(struct nouveau_mem_exec_func *); + void (*refresh_auto)(struct nouveau_mem_exec_func *, bool); + void (*refresh_self)(struct nouveau_mem_exec_func *, bool); + void (*wait)(struct nouveau_mem_exec_func *, u32 nsec); + u32 (*mrg)(struct nouveau_mem_exec_func *, int mr); + void (*mrs)(struct nouveau_mem_exec_func *, int mr, u32 data); + void (*clock_set)(struct nouveau_mem_exec_func *); + void (*timing_set)(struct nouveau_mem_exec_func *); + void *priv; +}; + +/* nouveau_mem.c */ +int nouveau_mem_exec(struct nouveau_mem_exec_func *, + struct nouveau_pm_level *); + /* nouveau_pm.c */ int nouveau_pm_init(struct drm_device *dev); void nouveau_pm_fini(struct drm_device *dev); void nouveau_pm_resume(struct drm_device *dev); +extern const struct nouveau_pm_profile_func nouveau_pm_static_profile_func; +void nouveau_pm_trigger(struct drm_device *dev); /* nouveau_volt.c */ void nouveau_volt_init(struct drm_device *); @@ -41,35 +61,41 @@ int nouveau_voltage_gpio_set(struct drm_device *, int voltage); /* nouveau_perf.c */ void nouveau_perf_init(struct drm_device *); void nouveau_perf_fini(struct drm_device *); +u8 *nouveau_perf_timing(struct drm_device *, u32 freq, u8 *ver, u8 *len); +u8 *nouveau_perf_ramcfg(struct drm_device *, u32 freq, u8 *ver, u8 *len); /* nouveau_mem.c */ void nouveau_mem_timing_init(struct drm_device *); void nouveau_mem_timing_fini(struct drm_device *); /* nv04_pm.c */ -int nv04_pm_clock_get(struct drm_device *, u32 id); -void *nv04_pm_clock_pre(struct drm_device *, struct nouveau_pm_level *, - u32 id, int khz); -void nv04_pm_clock_set(struct drm_device *, void *); +int nv04_pm_clocks_get(struct drm_device *, struct nouveau_pm_level *); +void *nv04_pm_clocks_pre(struct drm_device *, struct nouveau_pm_level *); +int nv04_pm_clocks_set(struct drm_device *, void *); /* nv40_pm.c */ int nv40_pm_clocks_get(struct drm_device *, struct nouveau_pm_level *); void *nv40_pm_clocks_pre(struct drm_device *, struct nouveau_pm_level *); -void nv40_pm_clocks_set(struct drm_device *, void *); +int nv40_pm_clocks_set(struct drm_device *, void *); +int nv40_pm_pwm_get(struct drm_device *, int, u32 *, u32 *); +int nv40_pm_pwm_set(struct drm_device *, int, u32, u32); /* nv50_pm.c */ -int nv50_pm_clock_get(struct drm_device *, u32 id); -void *nv50_pm_clock_pre(struct drm_device *, struct nouveau_pm_level *, - u32 id, int khz); -void nv50_pm_clock_set(struct drm_device *, void *); +int nv50_pm_clocks_get(struct drm_device *, struct nouveau_pm_level *); +void *nv50_pm_clocks_pre(struct drm_device *, struct nouveau_pm_level *); +int nv50_pm_clocks_set(struct drm_device *, void *); +int nv50_pm_pwm_get(struct drm_device *, int, u32 *, u32 *); +int nv50_pm_pwm_set(struct drm_device *, int, u32, u32); /* nva3_pm.c */ int nva3_pm_clocks_get(struct drm_device *, struct nouveau_pm_level *); void *nva3_pm_clocks_pre(struct drm_device *, struct nouveau_pm_level *); -void nva3_pm_clocks_set(struct drm_device *, void *); +int nva3_pm_clocks_set(struct drm_device *, void *); /* nvc0_pm.c */ int nvc0_pm_clocks_get(struct drm_device *, struct nouveau_pm_level *); +void *nvc0_pm_clocks_pre(struct drm_device *, struct nouveau_pm_level *); +int nvc0_pm_clocks_set(struct drm_device *, void *); /* nouveau_temp.c */ void nouveau_temp_init(struct drm_device *dev); diff --git a/drivers/gpu/drm/nouveau/nouveau_sgdma.c b/drivers/gpu/drm/nouveau/nouveau_sgdma.c index c8a463b..47f245e 100644 --- a/drivers/gpu/drm/nouveau/nouveau_sgdma.c +++ b/drivers/gpu/drm/nouveau/nouveau_sgdma.c @@ -8,91 +8,30 @@ #define NV_CTXDMA_PAGE_MASK (NV_CTXDMA_PAGE_SIZE - 1) struct nouveau_sgdma_be { - struct ttm_backend backend; + /* this has to be the first field so populate/unpopulated in + * nouve_bo.c works properly, otherwise have to move them here + */ + struct ttm_dma_tt ttm; struct drm_device *dev; - - dma_addr_t *pages; - unsigned nr_pages; - bool unmap_pages; - u64 offset; - bool bound; }; -static int -nouveau_sgdma_populate(struct ttm_backend *be, unsigned long num_pages, - struct page **pages, struct page *dummy_read_page, - dma_addr_t *dma_addrs) -{ - struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be; - struct drm_device *dev = nvbe->dev; - int i; - - NV_DEBUG(nvbe->dev, "num_pages = %ld\n", num_pages); - - nvbe->pages = dma_addrs; - nvbe->nr_pages = num_pages; - nvbe->unmap_pages = true; - - /* this code path isn't called and is incorrect anyways */ - if (0) { /* dma_addrs[0] != DMA_ERROR_CODE) { */ - nvbe->unmap_pages = false; - return 0; - } - - for (i = 0; i < num_pages; i++) { - nvbe->pages[i] = pci_map_page(dev->pdev, pages[i], 0, - PAGE_SIZE, PCI_DMA_BIDIRECTIONAL); - if (pci_dma_mapping_error(dev->pdev, nvbe->pages[i])) { - nvbe->nr_pages = --i; - be->func->clear(be); - return -EFAULT; - } - } - - return 0; -} - static void -nouveau_sgdma_clear(struct ttm_backend *be) +nouveau_sgdma_destroy(struct ttm_tt *ttm) { - struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be; - struct drm_device *dev = nvbe->dev; - - if (nvbe->bound) - be->func->unbind(be); + struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm; - if (nvbe->unmap_pages) { - while (nvbe->nr_pages--) { - pci_unmap_page(dev->pdev, nvbe->pages[nvbe->nr_pages], - PAGE_SIZE, PCI_DMA_BIDIRECTIONAL); - } - nvbe->unmap_pages = false; - } - - nvbe->pages = NULL; -} - -static void -nouveau_sgdma_destroy(struct ttm_backend *be) -{ - struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be; - - if (be) { + if (ttm) { NV_DEBUG(nvbe->dev, "\n"); - - if (nvbe) { - if (nvbe->pages) - be->func->clear(be); - kfree(nvbe); - } + ttm_dma_tt_fini(&nvbe->ttm); + kfree(nvbe); } } static int -nv04_sgdma_bind(struct ttm_backend *be, struct ttm_mem_reg *mem) +nv04_sgdma_bind(struct ttm_tt *ttm, struct ttm_mem_reg *mem) { - struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be; + struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm; struct drm_device *dev = nvbe->dev; struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_gpuobj *gpuobj = dev_priv->gart_info.sg_ctxdma; @@ -102,8 +41,8 @@ nv04_sgdma_bind(struct ttm_backend *be, struct ttm_mem_reg *mem) nvbe->offset = mem->start << PAGE_SHIFT; pte = (nvbe->offset >> NV_CTXDMA_PAGE_SHIFT) + 2; - for (i = 0; i < nvbe->nr_pages; i++) { - dma_addr_t dma_offset = nvbe->pages[i]; + for (i = 0; i < ttm->num_pages; i++) { + dma_addr_t dma_offset = nvbe->ttm.dma_address[i]; uint32_t offset_l = lower_32_bits(dma_offset); for (j = 0; j < PAGE_SIZE / NV_CTXDMA_PAGE_SIZE; j++, pte++) { @@ -112,14 +51,13 @@ nv04_sgdma_bind(struct ttm_backend *be, struct ttm_mem_reg *mem) } } - nvbe->bound = true; return 0; } static int -nv04_sgdma_unbind(struct ttm_backend *be) +nv04_sgdma_unbind(struct ttm_tt *ttm) { - struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be; + struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm; struct drm_device *dev = nvbe->dev; struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_gpuobj *gpuobj = dev_priv->gart_info.sg_ctxdma; @@ -127,22 +65,19 @@ nv04_sgdma_unbind(struct ttm_backend *be) NV_DEBUG(dev, "\n"); - if (!nvbe->bound) + if (ttm->state != tt_bound) return 0; pte = (nvbe->offset >> NV_CTXDMA_PAGE_SHIFT) + 2; - for (i = 0; i < nvbe->nr_pages; i++) { + for (i = 0; i < ttm->num_pages; i++) { for (j = 0; j < PAGE_SIZE / NV_CTXDMA_PAGE_SIZE; j++, pte++) nv_wo32(gpuobj, (pte * 4) + 0, 0x00000000); } - nvbe->bound = false; return 0; } static struct ttm_backend_func nv04_sgdma_backend = { - .populate = nouveau_sgdma_populate, - .clear = nouveau_sgdma_clear, .bind = nv04_sgdma_bind, .unbind = nv04_sgdma_unbind, .destroy = nouveau_sgdma_destroy @@ -161,14 +96,14 @@ nv41_sgdma_flush(struct nouveau_sgdma_be *nvbe) } static int -nv41_sgdma_bind(struct ttm_backend *be, struct ttm_mem_reg *mem) +nv41_sgdma_bind(struct ttm_tt *ttm, struct ttm_mem_reg *mem) { - struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be; + struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm; struct drm_nouveau_private *dev_priv = nvbe->dev->dev_private; struct nouveau_gpuobj *pgt = dev_priv->gart_info.sg_ctxdma; - dma_addr_t *list = nvbe->pages; + dma_addr_t *list = nvbe->ttm.dma_address; u32 pte = mem->start << 2; - u32 cnt = nvbe->nr_pages; + u32 cnt = ttm->num_pages; nvbe->offset = mem->start << PAGE_SHIFT; @@ -178,18 +113,17 @@ nv41_sgdma_bind(struct ttm_backend *be, struct ttm_mem_reg *mem) } nv41_sgdma_flush(nvbe); - nvbe->bound = true; return 0; } static int -nv41_sgdma_unbind(struct ttm_backend *be) +nv41_sgdma_unbind(struct ttm_tt *ttm) { - struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be; + struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm; struct drm_nouveau_private *dev_priv = nvbe->dev->dev_private; struct nouveau_gpuobj *pgt = dev_priv->gart_info.sg_ctxdma; u32 pte = (nvbe->offset >> 12) << 2; - u32 cnt = nvbe->nr_pages; + u32 cnt = ttm->num_pages; while (cnt--) { nv_wo32(pgt, pte, 0x00000000); @@ -197,24 +131,22 @@ nv41_sgdma_unbind(struct ttm_backend *be) } nv41_sgdma_flush(nvbe); - nvbe->bound = false; return 0; } static struct ttm_backend_func nv41_sgdma_backend = { - .populate = nouveau_sgdma_populate, - .clear = nouveau_sgdma_clear, .bind = nv41_sgdma_bind, .unbind = nv41_sgdma_unbind, .destroy = nouveau_sgdma_destroy }; static void -nv44_sgdma_flush(struct nouveau_sgdma_be *nvbe) +nv44_sgdma_flush(struct ttm_tt *ttm) { + struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm; struct drm_device *dev = nvbe->dev; - nv_wr32(dev, 0x100814, (nvbe->nr_pages - 1) << 12); + nv_wr32(dev, 0x100814, (ttm->num_pages - 1) << 12); nv_wr32(dev, 0x100808, nvbe->offset | 0x20); if (!nv_wait(dev, 0x100808, 0x00000001, 0x00000001)) NV_ERROR(dev, "gart flush timeout: 0x%08x\n", @@ -273,14 +205,14 @@ nv44_sgdma_fill(struct nouveau_gpuobj *pgt, dma_addr_t *list, u32 base, u32 cnt) } static int -nv44_sgdma_bind(struct ttm_backend *be, struct ttm_mem_reg *mem) +nv44_sgdma_bind(struct ttm_tt *ttm, struct ttm_mem_reg *mem) { - struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be; + struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm; struct drm_nouveau_private *dev_priv = nvbe->dev->dev_private; struct nouveau_gpuobj *pgt = dev_priv->gart_info.sg_ctxdma; - dma_addr_t *list = nvbe->pages; + dma_addr_t *list = nvbe->ttm.dma_address; u32 pte = mem->start << 2, tmp[4]; - u32 cnt = nvbe->nr_pages; + u32 cnt = ttm->num_pages; int i; nvbe->offset = mem->start << PAGE_SHIFT; @@ -308,19 +240,18 @@ nv44_sgdma_bind(struct ttm_backend *be, struct ttm_mem_reg *mem) if (cnt) nv44_sgdma_fill(pgt, list, pte, cnt); - nv44_sgdma_flush(nvbe); - nvbe->bound = true; + nv44_sgdma_flush(ttm); return 0; } static int -nv44_sgdma_unbind(struct ttm_backend *be) +nv44_sgdma_unbind(struct ttm_tt *ttm) { - struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be; + struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm; struct drm_nouveau_private *dev_priv = nvbe->dev->dev_private; struct nouveau_gpuobj *pgt = dev_priv->gart_info.sg_ctxdma; u32 pte = (nvbe->offset >> 12) << 2; - u32 cnt = nvbe->nr_pages; + u32 cnt = ttm->num_pages; if (pte & 0x0000000c) { u32 max = 4 - ((pte >> 2) & 0x3); @@ -342,55 +273,47 @@ nv44_sgdma_unbind(struct ttm_backend *be) if (cnt) nv44_sgdma_fill(pgt, NULL, pte, cnt); - nv44_sgdma_flush(nvbe); - nvbe->bound = false; + nv44_sgdma_flush(ttm); return 0; } static struct ttm_backend_func nv44_sgdma_backend = { - .populate = nouveau_sgdma_populate, - .clear = nouveau_sgdma_clear, .bind = nv44_sgdma_bind, .unbind = nv44_sgdma_unbind, .destroy = nouveau_sgdma_destroy }; static int -nv50_sgdma_bind(struct ttm_backend *be, struct ttm_mem_reg *mem) +nv50_sgdma_bind(struct ttm_tt *ttm, struct ttm_mem_reg *mem) { - struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be; + struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm; struct nouveau_mem *node = mem->mm_node; + /* noop: bound in move_notify() */ - node->pages = nvbe->pages; - nvbe->pages = (dma_addr_t *)node; - nvbe->bound = true; + node->pages = nvbe->ttm.dma_address; return 0; } static int -nv50_sgdma_unbind(struct ttm_backend *be) +nv50_sgdma_unbind(struct ttm_tt *ttm) { - struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be; - struct nouveau_mem *node = (struct nouveau_mem *)nvbe->pages; /* noop: unbound in move_notify() */ - nvbe->pages = node->pages; - node->pages = NULL; - nvbe->bound = false; return 0; } static struct ttm_backend_func nv50_sgdma_backend = { - .populate = nouveau_sgdma_populate, - .clear = nouveau_sgdma_clear, .bind = nv50_sgdma_bind, .unbind = nv50_sgdma_unbind, .destroy = nouveau_sgdma_destroy }; -struct ttm_backend * -nouveau_sgdma_init_ttm(struct drm_device *dev) +struct ttm_tt * +nouveau_sgdma_create_ttm(struct ttm_bo_device *bdev, + unsigned long size, uint32_t page_flags, + struct page *dummy_read_page) { - struct drm_nouveau_private *dev_priv = dev->dev_private; + struct drm_nouveau_private *dev_priv = nouveau_bdev(bdev); + struct drm_device *dev = dev_priv->dev; struct nouveau_sgdma_be *nvbe; nvbe = kzalloc(sizeof(*nvbe), GFP_KERNEL); @@ -398,9 +321,13 @@ nouveau_sgdma_init_ttm(struct drm_device *dev) return NULL; nvbe->dev = dev; + nvbe->ttm.ttm.func = dev_priv->gart_info.func; - nvbe->backend.func = dev_priv->gart_info.func; - return &nvbe->backend; + if (ttm_dma_tt_init(&nvbe->ttm, bdev, size, page_flags, dummy_read_page)) { + kfree(nvbe); + return NULL; + } + return &nvbe->ttm.ttm; } int diff --git a/drivers/gpu/drm/nouveau/nouveau_state.c b/drivers/gpu/drm/nouveau/nouveau_state.c index 01adcfb..b096cf2 100644 --- a/drivers/gpu/drm/nouveau/nouveau_state.c +++ b/drivers/gpu/drm/nouveau/nouveau_state.c @@ -36,6 +36,7 @@ #include "nouveau_drm.h" #include "nouveau_fbcon.h" #include "nouveau_ramht.h" +#include "nouveau_gpio.h" #include "nouveau_pm.h" #include "nv50_display.h" @@ -81,17 +82,13 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev) engine->display.early_init = nv04_display_early_init; engine->display.late_takedown = nv04_display_late_takedown; engine->display.create = nv04_display_create; - engine->display.init = nv04_display_init; engine->display.destroy = nv04_display_destroy; - engine->gpio.init = nouveau_stub_init; - engine->gpio.takedown = nouveau_stub_takedown; - engine->gpio.get = NULL; - engine->gpio.set = NULL; - engine->gpio.irq_enable = NULL; - engine->pm.clock_get = nv04_pm_clock_get; - engine->pm.clock_pre = nv04_pm_clock_pre; - engine->pm.clock_set = nv04_pm_clock_set; - engine->vram.init = nouveau_mem_detect; + engine->display.init = nv04_display_init; + engine->display.fini = nv04_display_fini; + engine->pm.clocks_get = nv04_pm_clocks_get; + engine->pm.clocks_pre = nv04_pm_clocks_pre; + engine->pm.clocks_set = nv04_pm_clocks_set; + engine->vram.init = nv04_fb_vram_init; engine->vram.takedown = nouveau_stub_takedown; engine->vram.flags_valid = nouveau_mem_flags_valid; break; @@ -130,17 +127,19 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev) engine->display.early_init = nv04_display_early_init; engine->display.late_takedown = nv04_display_late_takedown; engine->display.create = nv04_display_create; - engine->display.init = nv04_display_init; engine->display.destroy = nv04_display_destroy; - engine->gpio.init = nouveau_stub_init; - engine->gpio.takedown = nouveau_stub_takedown; - engine->gpio.get = nv10_gpio_get; - engine->gpio.set = nv10_gpio_set; - engine->gpio.irq_enable = NULL; - engine->pm.clock_get = nv04_pm_clock_get; - engine->pm.clock_pre = nv04_pm_clock_pre; - engine->pm.clock_set = nv04_pm_clock_set; - engine->vram.init = nouveau_mem_detect; + engine->display.init = nv04_display_init; + engine->display.fini = nv04_display_fini; + engine->gpio.drive = nv10_gpio_drive; + engine->gpio.sense = nv10_gpio_sense; + engine->pm.clocks_get = nv04_pm_clocks_get; + engine->pm.clocks_pre = nv04_pm_clocks_pre; + engine->pm.clocks_set = nv04_pm_clocks_set; + if (dev_priv->chipset == 0x1a || + dev_priv->chipset == 0x1f) + engine->vram.init = nv1a_fb_vram_init; + else + engine->vram.init = nv10_fb_vram_init; engine->vram.takedown = nouveau_stub_takedown; engine->vram.flags_valid = nouveau_mem_flags_valid; break; @@ -159,11 +158,11 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev) engine->timer.init = nv04_timer_init; engine->timer.read = nv04_timer_read; engine->timer.takedown = nv04_timer_takedown; - engine->fb.init = nv10_fb_init; - engine->fb.takedown = nv10_fb_takedown; - engine->fb.init_tile_region = nv10_fb_init_tile_region; - engine->fb.set_tile_region = nv10_fb_set_tile_region; - engine->fb.free_tile_region = nv10_fb_free_tile_region; + engine->fb.init = nv20_fb_init; + engine->fb.takedown = nv20_fb_takedown; + engine->fb.init_tile_region = nv20_fb_init_tile_region; + engine->fb.set_tile_region = nv20_fb_set_tile_region; + engine->fb.free_tile_region = nv20_fb_free_tile_region; engine->fifo.channels = 32; engine->fifo.init = nv10_fifo_init; engine->fifo.takedown = nv04_fifo_fini; @@ -179,17 +178,15 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev) engine->display.early_init = nv04_display_early_init; engine->display.late_takedown = nv04_display_late_takedown; engine->display.create = nv04_display_create; - engine->display.init = nv04_display_init; engine->display.destroy = nv04_display_destroy; - engine->gpio.init = nouveau_stub_init; - engine->gpio.takedown = nouveau_stub_takedown; - engine->gpio.get = nv10_gpio_get; - engine->gpio.set = nv10_gpio_set; - engine->gpio.irq_enable = NULL; - engine->pm.clock_get = nv04_pm_clock_get; - engine->pm.clock_pre = nv04_pm_clock_pre; - engine->pm.clock_set = nv04_pm_clock_set; - engine->vram.init = nouveau_mem_detect; + engine->display.init = nv04_display_init; + engine->display.fini = nv04_display_fini; + engine->gpio.drive = nv10_gpio_drive; + engine->gpio.sense = nv10_gpio_sense; + engine->pm.clocks_get = nv04_pm_clocks_get; + engine->pm.clocks_pre = nv04_pm_clocks_pre; + engine->pm.clocks_set = nv04_pm_clocks_set; + engine->vram.init = nv20_fb_vram_init; engine->vram.takedown = nouveau_stub_takedown; engine->vram.flags_valid = nouveau_mem_flags_valid; break; @@ -228,19 +225,17 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev) engine->display.early_init = nv04_display_early_init; engine->display.late_takedown = nv04_display_late_takedown; engine->display.create = nv04_display_create; - engine->display.init = nv04_display_init; engine->display.destroy = nv04_display_destroy; - engine->gpio.init = nouveau_stub_init; - engine->gpio.takedown = nouveau_stub_takedown; - engine->gpio.get = nv10_gpio_get; - engine->gpio.set = nv10_gpio_set; - engine->gpio.irq_enable = NULL; - engine->pm.clock_get = nv04_pm_clock_get; - engine->pm.clock_pre = nv04_pm_clock_pre; - engine->pm.clock_set = nv04_pm_clock_set; + engine->display.init = nv04_display_init; + engine->display.fini = nv04_display_fini; + engine->gpio.drive = nv10_gpio_drive; + engine->gpio.sense = nv10_gpio_sense; + engine->pm.clocks_get = nv04_pm_clocks_get; + engine->pm.clocks_pre = nv04_pm_clocks_pre; + engine->pm.clocks_set = nv04_pm_clocks_set; engine->pm.voltage_get = nouveau_voltage_gpio_get; engine->pm.voltage_set = nouveau_voltage_gpio_set; - engine->vram.init = nouveau_mem_detect; + engine->vram.init = nv20_fb_vram_init; engine->vram.takedown = nouveau_stub_takedown; engine->vram.flags_valid = nouveau_mem_flags_valid; break; @@ -280,20 +275,23 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev) engine->display.early_init = nv04_display_early_init; engine->display.late_takedown = nv04_display_late_takedown; engine->display.create = nv04_display_create; - engine->display.init = nv04_display_init; engine->display.destroy = nv04_display_destroy; - engine->gpio.init = nouveau_stub_init; - engine->gpio.takedown = nouveau_stub_takedown; - engine->gpio.get = nv10_gpio_get; - engine->gpio.set = nv10_gpio_set; - engine->gpio.irq_enable = NULL; + engine->display.init = nv04_display_init; + engine->display.fini = nv04_display_fini; + engine->gpio.init = nv10_gpio_init; + engine->gpio.fini = nv10_gpio_fini; + engine->gpio.drive = nv10_gpio_drive; + engine->gpio.sense = nv10_gpio_sense; + engine->gpio.irq_enable = nv10_gpio_irq_enable; engine->pm.clocks_get = nv40_pm_clocks_get; engine->pm.clocks_pre = nv40_pm_clocks_pre; engine->pm.clocks_set = nv40_pm_clocks_set; engine->pm.voltage_get = nouveau_voltage_gpio_get; engine->pm.voltage_set = nouveau_voltage_gpio_set; engine->pm.temp_get = nv40_temp_get; - engine->vram.init = nouveau_mem_detect; + engine->pm.pwm_get = nv40_pm_pwm_get; + engine->pm.pwm_set = nv40_pm_pwm_set; + engine->vram.init = nv40_fb_vram_init; engine->vram.takedown = nouveau_stub_takedown; engine->vram.flags_valid = nouveau_mem_flags_valid; break; @@ -335,14 +333,13 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev) engine->display.early_init = nv50_display_early_init; engine->display.late_takedown = nv50_display_late_takedown; engine->display.create = nv50_display_create; - engine->display.init = nv50_display_init; engine->display.destroy = nv50_display_destroy; + engine->display.init = nv50_display_init; + engine->display.fini = nv50_display_fini; engine->gpio.init = nv50_gpio_init; - engine->gpio.takedown = nv50_gpio_fini; - engine->gpio.get = nv50_gpio_get; - engine->gpio.set = nv50_gpio_set; - engine->gpio.irq_register = nv50_gpio_irq_register; - engine->gpio.irq_unregister = nv50_gpio_irq_unregister; + engine->gpio.fini = nv50_gpio_fini; + engine->gpio.drive = nv50_gpio_drive; + engine->gpio.sense = nv50_gpio_sense; engine->gpio.irq_enable = nv50_gpio_irq_enable; switch (dev_priv->chipset) { case 0x84: @@ -355,9 +352,9 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev) case 0xaa: case 0xac: case 0x50: - engine->pm.clock_get = nv50_pm_clock_get; - engine->pm.clock_pre = nv50_pm_clock_pre; - engine->pm.clock_set = nv50_pm_clock_set; + engine->pm.clocks_get = nv50_pm_clocks_get; + engine->pm.clocks_pre = nv50_pm_clocks_pre; + engine->pm.clocks_set = nv50_pm_clocks_set; break; default: engine->pm.clocks_get = nva3_pm_clocks_get; @@ -371,6 +368,8 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev) engine->pm.temp_get = nv84_temp_get; else engine->pm.temp_get = nv40_temp_get; + engine->pm.pwm_get = nv50_pm_pwm_get; + engine->pm.pwm_set = nv50_pm_pwm_set; engine->vram.init = nv50_vram_init; engine->vram.takedown = nv50_vram_fini; engine->vram.get = nv50_vram_new; @@ -408,14 +407,13 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev) engine->display.early_init = nv50_display_early_init; engine->display.late_takedown = nv50_display_late_takedown; engine->display.create = nv50_display_create; - engine->display.init = nv50_display_init; engine->display.destroy = nv50_display_destroy; + engine->display.init = nv50_display_init; + engine->display.fini = nv50_display_fini; engine->gpio.init = nv50_gpio_init; - engine->gpio.takedown = nouveau_stub_takedown; - engine->gpio.get = nv50_gpio_get; - engine->gpio.set = nv50_gpio_set; - engine->gpio.irq_register = nv50_gpio_irq_register; - engine->gpio.irq_unregister = nv50_gpio_irq_unregister; + engine->gpio.fini = nv50_gpio_fini; + engine->gpio.drive = nv50_gpio_drive; + engine->gpio.sense = nv50_gpio_sense; engine->gpio.irq_enable = nv50_gpio_irq_enable; engine->vram.init = nvc0_vram_init; engine->vram.takedown = nv50_vram_fini; @@ -424,8 +422,12 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev) engine->vram.flags_valid = nvc0_vram_flags_valid; engine->pm.temp_get = nv84_temp_get; engine->pm.clocks_get = nvc0_pm_clocks_get; + engine->pm.clocks_pre = nvc0_pm_clocks_pre; + engine->pm.clocks_set = nvc0_pm_clocks_set; engine->pm.voltage_get = nouveau_voltage_gpio_get; engine->pm.voltage_set = nouveau_voltage_gpio_set; + engine->pm.pwm_get = nv50_pm_pwm_get; + engine->pm.pwm_set = nv50_pm_pwm_set; break; case 0xd0: engine->instmem.init = nvc0_instmem_init; @@ -458,24 +460,67 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev) engine->display.early_init = nouveau_stub_init; engine->display.late_takedown = nouveau_stub_takedown; engine->display.create = nvd0_display_create; - engine->display.init = nvd0_display_init; engine->display.destroy = nvd0_display_destroy; + engine->display.init = nvd0_display_init; + engine->display.fini = nvd0_display_fini; engine->gpio.init = nv50_gpio_init; - engine->gpio.takedown = nouveau_stub_takedown; - engine->gpio.get = nvd0_gpio_get; - engine->gpio.set = nvd0_gpio_set; - engine->gpio.irq_register = nv50_gpio_irq_register; - engine->gpio.irq_unregister = nv50_gpio_irq_unregister; + engine->gpio.fini = nv50_gpio_fini; + engine->gpio.drive = nvd0_gpio_drive; + engine->gpio.sense = nvd0_gpio_sense; engine->gpio.irq_enable = nv50_gpio_irq_enable; engine->vram.init = nvc0_vram_init; engine->vram.takedown = nv50_vram_fini; engine->vram.get = nvc0_vram_new; engine->vram.put = nv50_vram_del; engine->vram.flags_valid = nvc0_vram_flags_valid; + engine->pm.temp_get = nv84_temp_get; engine->pm.clocks_get = nvc0_pm_clocks_get; + engine->pm.clocks_pre = nvc0_pm_clocks_pre; + engine->pm.clocks_set = nvc0_pm_clocks_set; engine->pm.voltage_get = nouveau_voltage_gpio_get; engine->pm.voltage_set = nouveau_voltage_gpio_set; break; + case 0xe0: + engine->instmem.init = nvc0_instmem_init; + engine->instmem.takedown = nvc0_instmem_takedown; + engine->instmem.suspend = nvc0_instmem_suspend; + engine->instmem.resume = nvc0_instmem_resume; + engine->instmem.get = nv50_instmem_get; + engine->instmem.put = nv50_instmem_put; + engine->instmem.map = nv50_instmem_map; + engine->instmem.unmap = nv50_instmem_unmap; + engine->instmem.flush = nv84_instmem_flush; + engine->mc.init = nv50_mc_init; + engine->mc.takedown = nv50_mc_takedown; + engine->timer.init = nv04_timer_init; + engine->timer.read = nv04_timer_read; + engine->timer.takedown = nv04_timer_takedown; + engine->fb.init = nvc0_fb_init; + engine->fb.takedown = nvc0_fb_takedown; + engine->fifo.channels = 0; + engine->fifo.init = nouveau_stub_init; + engine->fifo.takedown = nouveau_stub_takedown; + engine->fifo.disable = nvc0_fifo_disable; + engine->fifo.enable = nvc0_fifo_enable; + engine->fifo.reassign = nvc0_fifo_reassign; + engine->fifo.unload_context = nouveau_stub_init; + engine->display.early_init = nouveau_stub_init; + engine->display.late_takedown = nouveau_stub_takedown; + engine->display.create = nvd0_display_create; + engine->display.destroy = nvd0_display_destroy; + engine->display.init = nvd0_display_init; + engine->display.fini = nvd0_display_fini; + engine->gpio.init = nv50_gpio_init; + engine->gpio.fini = nv50_gpio_fini; + engine->gpio.drive = nvd0_gpio_drive; + engine->gpio.sense = nvd0_gpio_sense; + engine->gpio.irq_enable = nv50_gpio_irq_enable; + engine->vram.init = nvc0_vram_init; + engine->vram.takedown = nv50_vram_fini; + engine->vram.get = nvc0_vram_new; + engine->vram.put = nv50_vram_del; + engine->vram.flags_valid = nvc0_vram_flags_valid; + break; default: NV_ERROR(dev, "NV%02x unsupported\n", dev_priv->chipset); return 1; @@ -527,6 +572,7 @@ static void nouveau_switcheroo_set_state(struct pci_dev *pdev, printk(KERN_ERR "VGA switcheroo: switched nouveau off\n"); dev->switch_power_state = DRM_SWITCH_POWER_CHANGING; drm_kms_helper_poll_disable(dev); + nouveau_switcheroo_optimus_dsm(); nouveau_pci_suspend(pdev, pmm); dev->switch_power_state = DRM_SWITCH_POWER_OFF; } @@ -549,6 +595,75 @@ static bool nouveau_switcheroo_can_switch(struct pci_dev *pdev) return can_switch; } +static void +nouveau_card_channel_fini(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + + if (dev_priv->channel) + nouveau_channel_put_unlocked(&dev_priv->channel); +} + +static int +nouveau_card_channel_init(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_channel *chan; + int ret, oclass; + + ret = nouveau_channel_alloc(dev, &chan, NULL, NvDmaFB, NvDmaTT); + dev_priv->channel = chan; + if (ret) + return ret; + + mutex_unlock(&dev_priv->channel->mutex); + + if (dev_priv->card_type <= NV_50) { + if (dev_priv->card_type < NV_50) + oclass = 0x0039; + else + oclass = 0x5039; + + ret = nouveau_gpuobj_gr_new(chan, NvM2MF, oclass); + if (ret) + goto error; + + ret = nouveau_notifier_alloc(chan, NvNotify0, 32, 0xfe0, 0x1000, + &chan->m2mf_ntfy); + if (ret) + goto error; + + ret = RING_SPACE(chan, 6); + if (ret) + goto error; + + BEGIN_RING(chan, NvSubM2MF, NV_MEMORY_TO_MEMORY_FORMAT_NAME, 1); + OUT_RING (chan, NvM2MF); + BEGIN_RING(chan, NvSubM2MF, NV_MEMORY_TO_MEMORY_FORMAT_DMA_NOTIFY, 3); + OUT_RING (chan, NvNotify0); + OUT_RING (chan, chan->vram_handle); + OUT_RING (chan, chan->gart_handle); + } else + if (dev_priv->card_type <= NV_D0) { + ret = nouveau_gpuobj_gr_new(chan, 0x9039, 0x9039); + if (ret) + goto error; + + ret = RING_SPACE(chan, 2); + if (ret) + goto error; + + BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0000, 1); + OUT_RING (chan, 0x00009039); + } + + FIRE_RING (chan); +error: + if (ret) + nouveau_card_channel_fini(dev); + return ret; +} + int nouveau_card_init(struct drm_device *dev) { @@ -589,47 +704,45 @@ nouveau_card_init(struct drm_device *dev) nv_mask(dev, 0x00088080, 0x00000800, 0x00000000); } - nouveau_pm_init(dev); - - ret = engine->vram.init(dev); + /* PMC */ + ret = engine->mc.init(dev); if (ret) goto out_bios; - ret = nouveau_gpuobj_init(dev); + /* PTIMER */ + ret = engine->timer.init(dev); if (ret) - goto out_vram; + goto out_mc; - ret = engine->instmem.init(dev); + /* PFB */ + ret = engine->fb.init(dev); if (ret) - goto out_gpuobj; + goto out_timer; - ret = nouveau_mem_vram_init(dev); + ret = engine->vram.init(dev); if (ret) - goto out_instmem; + goto out_fb; - ret = nouveau_mem_gart_init(dev); + /* PGPIO */ + ret = nouveau_gpio_create(dev); if (ret) - goto out_ttmvram; + goto out_vram; - /* PMC */ - ret = engine->mc.init(dev); + ret = nouveau_gpuobj_init(dev); if (ret) - goto out_gart; + goto out_gpio; - /* PGPIO */ - ret = engine->gpio.init(dev); + ret = engine->instmem.init(dev); if (ret) - goto out_mc; + goto out_gpuobj; - /* PTIMER */ - ret = engine->timer.init(dev); + ret = nouveau_mem_vram_init(dev); if (ret) - goto out_gpio; + goto out_instmem; - /* PFB */ - ret = engine->fb.init(dev); + ret = nouveau_mem_gart_init(dev); if (ret) - goto out_timer; + goto out_ttmvram; if (!dev_priv->noaccel) { switch (dev_priv->card_type) { @@ -650,6 +763,7 @@ nouveau_card_init(struct drm_device *dev) nv50_graph_create(dev); break; case NV_C0: + case NV_D0: nvc0_graph_create(dev); break; default: @@ -665,6 +779,11 @@ nouveau_card_init(struct drm_device *dev) case 0xa0: nv84_crypt_create(dev); break; + case 0x98: + case 0xaa: + case 0xac: + nv98_crypt_create(dev); + break; } switch (dev_priv->card_type) { @@ -686,15 +805,25 @@ nouveau_card_init(struct drm_device *dev) break; } + if (dev_priv->chipset >= 0xa3 || dev_priv->chipset == 0x98) { + nv84_bsp_create(dev); + nv84_vp_create(dev); + nv98_ppp_create(dev); + } else + if (dev_priv->chipset >= 0x84) { + nv50_mpeg_create(dev); + nv84_bsp_create(dev); + nv84_vp_create(dev); + } else + if (dev_priv->chipset >= 0x50) { + nv50_mpeg_create(dev); + } else if (dev_priv->card_type == NV_40 || dev_priv->chipset == 0x31 || dev_priv->chipset == 0x34 || - dev_priv->chipset == 0x36) + dev_priv->chipset == 0x36) { nv31_mpeg_create(dev); - else - if (dev_priv->card_type == NV_50 && - (dev_priv->chipset < 0x98 || dev_priv->chipset == 0xa0)) - nv50_mpeg_create(dev); + } for (e = 0; e < NVOBJ_ENGINE_NR; e++) { if (dev_priv->eng[e]) { @@ -714,63 +843,41 @@ nouveau_card_init(struct drm_device *dev) if (ret) goto out_fifo; - /* initialise general modesetting */ - drm_mode_config_init(dev); - drm_mode_create_scaling_mode_property(dev); - drm_mode_create_dithering_property(dev); - dev->mode_config.funcs = (void *)&nouveau_mode_config_funcs; - dev->mode_config.fb_base = pci_resource_start(dev->pdev, 1); - dev->mode_config.min_width = 0; - dev->mode_config.min_height = 0; - if (dev_priv->card_type < NV_10) { - dev->mode_config.max_width = 2048; - dev->mode_config.max_height = 2048; - } else - if (dev_priv->card_type < NV_50) { - dev->mode_config.max_width = 4096; - dev->mode_config.max_height = 4096; - } else { - dev->mode_config.max_width = 8192; - dev->mode_config.max_height = 8192; - } - - ret = engine->display.create(dev); + ret = nouveau_display_create(dev); if (ret) goto out_irq; nouveau_backlight_init(dev); + nouveau_pm_init(dev); - if (dev_priv->eng[NVOBJ_ENGINE_GR]) { - ret = nouveau_fence_init(dev); - if (ret) - goto out_disp; + ret = nouveau_fence_init(dev); + if (ret) + goto out_pm; - ret = nouveau_channel_alloc(dev, &dev_priv->channel, NULL, - NvDmaFB, NvDmaTT); + if (dev_priv->eng[NVOBJ_ENGINE_GR]) { + ret = nouveau_card_channel_init(dev); if (ret) goto out_fence; - - mutex_unlock(&dev_priv->channel->mutex); } if (dev->mode_config.num_crtc) { - ret = drm_vblank_init(dev, dev->mode_config.num_crtc); + ret = nouveau_display_init(dev); if (ret) goto out_chan; nouveau_fbcon_init(dev); - drm_kms_helper_poll_init(dev); } return 0; out_chan: - nouveau_channel_put_unlocked(&dev_priv->channel); + nouveau_card_channel_fini(dev); out_fence: nouveau_fence_fini(dev); -out_disp: +out_pm: + nouveau_pm_fini(dev); nouveau_backlight_exit(dev); - engine->display.destroy(dev); + nouveau_display_destroy(dev); out_irq: nouveau_irq_fini(dev); out_fifo: @@ -785,15 +892,6 @@ out_engine: dev_priv->eng[e]->destroy(dev,e ); } } - - engine->fb.takedown(dev); -out_timer: - engine->timer.takedown(dev); -out_gpio: - engine->gpio.takedown(dev); -out_mc: - engine->mc.takedown(dev); -out_gart: nouveau_mem_gart_fini(dev); out_ttmvram: nouveau_mem_vram_fini(dev); @@ -801,10 +899,17 @@ out_instmem: engine->instmem.takedown(dev); out_gpuobj: nouveau_gpuobj_takedown(dev); +out_gpio: + nouveau_gpio_destroy(dev); out_vram: engine->vram.takedown(dev); +out_fb: + engine->fb.takedown(dev); +out_timer: + engine->timer.takedown(dev); +out_mc: + engine->mc.takedown(dev); out_bios: - nouveau_pm_fini(dev); nouveau_bios_takedown(dev); out_display_early: engine->display.late_takedown(dev); @@ -820,19 +925,15 @@ static void nouveau_card_takedown(struct drm_device *dev) int e; if (dev->mode_config.num_crtc) { - drm_kms_helper_poll_fini(dev); nouveau_fbcon_fini(dev); - drm_vblank_cleanup(dev); - } - - if (dev_priv->channel) { - nouveau_channel_put_unlocked(&dev_priv->channel); - nouveau_fence_fini(dev); + nouveau_display_fini(dev); } + nouveau_card_channel_fini(dev); + nouveau_fence_fini(dev); + nouveau_pm_fini(dev); nouveau_backlight_exit(dev); - engine->display.destroy(dev); - drm_mode_config_cleanup(dev); + nouveau_display_destroy(dev); if (!dev_priv->noaccel) { engine->fifo.takedown(dev); @@ -843,11 +944,6 @@ static void nouveau_card_takedown(struct drm_device *dev) } } } - engine->fb.takedown(dev); - engine->timer.takedown(dev); - engine->gpio.takedown(dev); - engine->mc.takedown(dev); - engine->display.late_takedown(dev); if (dev_priv->vga_ram) { nouveau_bo_unpin(dev_priv->vga_ram); @@ -863,12 +959,17 @@ static void nouveau_card_takedown(struct drm_device *dev) engine->instmem.takedown(dev); nouveau_gpuobj_takedown(dev); - engine->vram.takedown(dev); - nouveau_irq_fini(dev); + nouveau_gpio_destroy(dev); + engine->vram.takedown(dev); + engine->fb.takedown(dev); + engine->timer.takedown(dev); + engine->mc.takedown(dev); - nouveau_pm_fini(dev); nouveau_bios_takedown(dev); + engine->display.late_takedown(dev); + + nouveau_irq_fini(dev); vga_client_register(dev->pdev, NULL, NULL, NULL); } @@ -998,8 +1099,8 @@ static int nouveau_remove_conflicting_drivers(struct drm_device *dev) int nouveau_load(struct drm_device *dev, unsigned long flags) { struct drm_nouveau_private *dev_priv; - uint32_t reg0, strap; - resource_size_t mmio_start_offs; + unsigned long long offset, length; + uint32_t reg0 = ~0, strap; int ret; dev_priv = kzalloc(sizeof(*dev_priv), GFP_KERNEL); @@ -1010,83 +1111,90 @@ int nouveau_load(struct drm_device *dev, unsigned long flags) dev->dev_private = dev_priv; dev_priv->dev = dev; + pci_set_master(dev->pdev); + dev_priv->flags = flags & NOUVEAU_FLAGS; NV_DEBUG(dev, "vendor: 0x%X device: 0x%X class: 0x%X\n", dev->pci_vendor, dev->pci_device, dev->pdev->class); - /* resource 0 is mmio regs */ - /* resource 1 is linear FB */ - /* resource 2 is RAMIN (mmio regs + 0x1000000) */ - /* resource 6 is bios */ + /* first up, map the start of mmio and determine the chipset */ + dev_priv->mmio = ioremap(pci_resource_start(dev->pdev, 0), PAGE_SIZE); + if (dev_priv->mmio) { +#ifdef __BIG_ENDIAN + /* put the card into big-endian mode if it's not */ + if (nv_rd32(dev, NV03_PMC_BOOT_1) != 0x01000001) + nv_wr32(dev, NV03_PMC_BOOT_1, 0x01000001); + DRM_MEMORYBARRIER(); +#endif - /* map the mmio regs */ - mmio_start_offs = pci_resource_start(dev->pdev, 0); - dev_priv->mmio = ioremap(mmio_start_offs, 0x00800000); - if (!dev_priv->mmio) { - NV_ERROR(dev, "Unable to initialize the mmio mapping. " - "Please report your setup to " DRIVER_EMAIL "\n"); + /* determine chipset and derive architecture from it */ + reg0 = nv_rd32(dev, NV03_PMC_BOOT_0); + if ((reg0 & 0x0f000000) > 0) { + dev_priv->chipset = (reg0 & 0xff00000) >> 20; + switch (dev_priv->chipset & 0xf0) { + case 0x10: + case 0x20: + case 0x30: + dev_priv->card_type = dev_priv->chipset & 0xf0; + break; + case 0x40: + case 0x60: + dev_priv->card_type = NV_40; + break; + case 0x50: + case 0x80: + case 0x90: + case 0xa0: + dev_priv->card_type = NV_50; + break; + case 0xc0: + dev_priv->card_type = NV_C0; + break; + case 0xd0: + dev_priv->card_type = NV_D0; + break; + case 0xe0: + dev_priv->card_type = NV_E0; + break; + default: + break; + } + } else + if ((reg0 & 0xff00fff0) == 0x20004000) { + if (reg0 & 0x00f00000) + dev_priv->chipset = 0x05; + else + dev_priv->chipset = 0x04; + dev_priv->card_type = NV_04; + } + + iounmap(dev_priv->mmio); + } + + if (!dev_priv->card_type) { + NV_ERROR(dev, "unsupported chipset 0x%08x\n", reg0); ret = -EINVAL; goto err_priv; } - NV_DEBUG(dev, "regs mapped ok at 0x%llx\n", - (unsigned long long)mmio_start_offs); -#ifdef __BIG_ENDIAN - /* Put the card in BE mode if it's not */ - if (nv_rd32(dev, NV03_PMC_BOOT_1) != 0x01000001) - nv_wr32(dev, NV03_PMC_BOOT_1, 0x01000001); + NV_INFO(dev, "Detected an NV%2x generation card (0x%08x)\n", + dev_priv->card_type, reg0); - DRM_MEMORYBARRIER(); -#endif + /* map the mmio regs, limiting the amount to preserve vmap space */ + offset = pci_resource_start(dev->pdev, 0); + length = pci_resource_len(dev->pdev, 0); + if (dev_priv->card_type < NV_E0) + length = min(length, (unsigned long long)0x00800000); - /* Time to determine the card architecture */ - reg0 = nv_rd32(dev, NV03_PMC_BOOT_0); - - /* We're dealing with >=NV10 */ - if ((reg0 & 0x0f000000) > 0) { - /* Bit 27-20 contain the architecture in hex */ - dev_priv->chipset = (reg0 & 0xff00000) >> 20; - /* NV04 or NV05 */ - } else if ((reg0 & 0xff00fff0) == 0x20004000) { - if (reg0 & 0x00f00000) - dev_priv->chipset = 0x05; - else - dev_priv->chipset = 0x04; - } else - dev_priv->chipset = 0xff; - - switch (dev_priv->chipset & 0xf0) { - case 0x00: - case 0x10: - case 0x20: - case 0x30: - dev_priv->card_type = dev_priv->chipset & 0xf0; - break; - case 0x40: - case 0x60: - dev_priv->card_type = NV_40; - break; - case 0x50: - case 0x80: - case 0x90: - case 0xa0: - dev_priv->card_type = NV_50; - break; - case 0xc0: - dev_priv->card_type = NV_C0; - break; - case 0xd0: - dev_priv->card_type = NV_D0; - break; - default: - NV_INFO(dev, "Unsupported chipset 0x%08x\n", reg0); + dev_priv->mmio = ioremap(offset, length); + if (!dev_priv->mmio) { + NV_ERROR(dev, "Unable to initialize the mmio mapping. " + "Please report your setup to " DRIVER_EMAIL "\n"); ret = -EINVAL; - goto err_mmio; + goto err_priv; } - - NV_INFO(dev, "Detected an NV%2x generation card (0x%08x)\n", - dev_priv->card_type, reg0); + NV_DEBUG(dev, "regs mapped ok at 0x%llx\n", offset); /* determine frequency of timing crystal */ strap = nv_rd32(dev, 0x101000); @@ -1112,13 +1220,11 @@ int nouveau_load(struct drm_device *dev, unsigned long flags) dev_priv->noaccel = !!nouveau_noaccel; if (nouveau_noaccel == -1) { switch (dev_priv->chipset) { -#if 0 - case 0xXX: /* known broken */ + case 0xd9: /* known broken */ NV_INFO(dev, "acceleration disabled by default, pass " "noaccel=0 to force enable\n"); dev_priv->noaccel = true; break; -#endif default: dev_priv->noaccel = false; break; @@ -1146,7 +1252,7 @@ int nouveau_load(struct drm_device *dev, unsigned long flags) } } else { dev_priv->ramin_size = 1 * 1024 * 1024; - dev_priv->ramin = ioremap(mmio_start_offs + NV_RAMIN, + dev_priv->ramin = ioremap(offset + NV_RAMIN, dev_priv->ramin_size); if (!dev_priv->ramin) { NV_ERROR(dev, "Failed to map BAR0 PRAMIN.\n"); @@ -1240,7 +1346,7 @@ int nouveau_ioctl_getparam(struct drm_device *dev, void *data, getparam->value = 1; break; case NOUVEAU_GETPARAM_HAS_PAGEFLIP: - getparam->value = dev_priv->card_type < NV_D0; + getparam->value = 1; break; case NOUVEAU_GETPARAM_GRAPH_UNITS: /* NV40 and NV50 versions are quite different, but register diff --git a/drivers/gpu/drm/nouveau/nouveau_temp.c b/drivers/gpu/drm/nouveau/nouveau_temp.c index 5a46446..0f5a301 100644 --- a/drivers/gpu/drm/nouveau/nouveau_temp.c +++ b/drivers/gpu/drm/nouveau/nouveau_temp.c @@ -55,6 +55,10 @@ nouveau_temp_vbios_parse(struct drm_device *dev, u8 *temp) temps->down_clock = 100; temps->fan_boost = 90; + /* Set the default range for the pwm fan */ + pm->fan.min_duty = 30; + pm->fan.max_duty = 100; + /* Set the known default values to setup the temperature sensor */ if (dev_priv->card_type >= NV_40) { switch (dev_priv->chipset) { @@ -156,11 +160,26 @@ nouveau_temp_vbios_parse(struct drm_device *dev, u8 *temp) case 0x13: sensor->slope_div = value; break; + case 0x22: + pm->fan.min_duty = value & 0xff; + pm->fan.max_duty = (value & 0xff00) >> 8; + break; + case 0x26: + pm->fan.pwm_freq = value; + break; } temp += recordlen; } nouveau_temp_safety_checks(dev); + + /* check the fan min/max settings */ + if (pm->fan.min_duty < 10) + pm->fan.min_duty = 10; + if (pm->fan.max_duty > 100) + pm->fan.max_duty = 100; + if (pm->fan.max_duty < pm->fan.min_duty) + pm->fan.max_duty = pm->fan.min_duty; } static int @@ -267,8 +286,6 @@ probe_monitoring_device(struct nouveau_i2c_chan *i2c, static void nouveau_temp_probe_i2c(struct drm_device *dev) { - struct drm_nouveau_private *dev_priv = dev->dev_private; - struct dcb_table *dcb = &dev_priv->vbios.dcb; struct i2c_board_info info[] = { { I2C_BOARD_INFO("w83l785ts", 0x2d) }, { I2C_BOARD_INFO("w83781d", 0x2d) }, @@ -277,11 +294,9 @@ nouveau_temp_probe_i2c(struct drm_device *dev) { I2C_BOARD_INFO("lm99", 0x4c) }, { } }; - int idx = (dcb->version >= 0x40 ? - dcb->i2c_default_indices & 0xf : 2); nouveau_i2c_identify(dev, "monitoring device", info, - probe_monitoring_device, idx); + probe_monitoring_device, NV_I2C_DEFAULT(0)); } void @@ -297,9 +312,9 @@ nouveau_temp_init(struct drm_device *dev) return; if (P.version == 1) - temp = ROMPTR(bios, P.data[12]); + temp = ROMPTR(dev, P.data[12]); else if (P.version == 2) - temp = ROMPTR(bios, P.data[16]); + temp = ROMPTR(dev, P.data[16]); else NV_WARN(dev, "unknown temp for BIT P %d\n", P.version); diff --git a/drivers/gpu/drm/nouveau/nouveau_vm.c b/drivers/gpu/drm/nouveau/nouveau_vm.c index ef0832b..2bf6c03 100644 --- a/drivers/gpu/drm/nouveau/nouveau_vm.c +++ b/drivers/gpu/drm/nouveau/nouveau_vm.c @@ -78,9 +78,10 @@ nouveau_vm_map(struct nouveau_vma *vma, struct nouveau_mem *node) void nouveau_vm_map_sg(struct nouveau_vma *vma, u64 delta, u64 length, - struct nouveau_mem *mem, dma_addr_t *list) + struct nouveau_mem *mem) { struct nouveau_vm *vm = vma->vm; + dma_addr_t *list = mem->pages; int big = vma->node->type != vm->spg_shift; u32 offset = vma->node->offset + (delta >> 12); u32 bits = vma->node->type - 12; diff --git a/drivers/gpu/drm/nouveau/nouveau_vm.h b/drivers/gpu/drm/nouveau/nouveau_vm.h index 6ce995f..4fb6e72 100644 --- a/drivers/gpu/drm/nouveau/nouveau_vm.h +++ b/drivers/gpu/drm/nouveau/nouveau_vm.h @@ -89,7 +89,7 @@ void nouveau_vm_map_at(struct nouveau_vma *, u64 offset, struct nouveau_mem *); void nouveau_vm_unmap(struct nouveau_vma *); void nouveau_vm_unmap_at(struct nouveau_vma *, u64 offset, u64 length); void nouveau_vm_map_sg(struct nouveau_vma *, u64 offset, u64 length, - struct nouveau_mem *, dma_addr_t *); + struct nouveau_mem *); /* nv50_vm.c */ void nv50_vm_map_pgt(struct nouveau_gpuobj *pgd, u32 pde, diff --git a/drivers/gpu/drm/nouveau/nouveau_volt.c b/drivers/gpu/drm/nouveau/nouveau_volt.c index 86d03e1..b010cb9 100644 --- a/drivers/gpu/drm/nouveau/nouveau_volt.c +++ b/drivers/gpu/drm/nouveau/nouveau_volt.c @@ -26,6 +26,7 @@ #include "nouveau_drv.h" #include "nouveau_pm.h" +#include "nouveau_gpio.h" static const enum dcb_gpio_tag vidtag[] = { 0x04, 0x05, 0x06, 0x1a, 0x73 }; static int nr_vidtag = sizeof(vidtag) / sizeof(vidtag[0]); @@ -34,7 +35,6 @@ int nouveau_voltage_gpio_get(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_gpio_engine *gpio = &dev_priv->engine.gpio; struct nouveau_pm_voltage *volt = &dev_priv->engine.pm.voltage; u8 vid = 0; int i; @@ -43,7 +43,7 @@ nouveau_voltage_gpio_get(struct drm_device *dev) if (!(volt->vid_mask & (1 << i))) continue; - vid |= gpio->get(dev, vidtag[i]) << i; + vid |= nouveau_gpio_func_get(dev, vidtag[i]) << i; } return nouveau_volt_lvl_lookup(dev, vid); @@ -53,7 +53,6 @@ int nouveau_voltage_gpio_set(struct drm_device *dev, int voltage) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_gpio_engine *gpio = &dev_priv->engine.gpio; struct nouveau_pm_voltage *volt = &dev_priv->engine.pm.voltage; int vid, i; @@ -65,7 +64,7 @@ nouveau_voltage_gpio_set(struct drm_device *dev, int voltage) if (!(volt->vid_mask & (1 << i))) continue; - gpio->set(dev, vidtag[i], !!(vid & (1 << i))); + nouveau_gpio_func_set(dev, vidtag[i], !!(vid & (1 << i))); } return 0; @@ -117,10 +116,10 @@ nouveau_volt_init(struct drm_device *dev) return; if (P.version == 1) - volt = ROMPTR(bios, P.data[16]); + volt = ROMPTR(dev, P.data[16]); else if (P.version == 2) - volt = ROMPTR(bios, P.data[12]); + volt = ROMPTR(dev, P.data[12]); else { NV_WARN(dev, "unknown volt for BIT P %d\n", P.version); } @@ -130,7 +129,7 @@ nouveau_volt_init(struct drm_device *dev) return; } - volt = ROMPTR(bios, bios->data[bios->offset + 0x98]); + volt = ROMPTR(dev, bios->data[bios->offset + 0x98]); } if (!volt) { @@ -194,7 +193,7 @@ nouveau_volt_init(struct drm_device *dev) return; } - if (!nouveau_bios_gpio_entry(dev, vidtag[i])) { + if (!nouveau_gpio_func_valid(dev, vidtag[i])) { NV_DEBUG(dev, "vid bit %d has no gpio tag\n", i); return; } diff --git a/drivers/gpu/drm/nouveau/nv04_crtc.c b/drivers/gpu/drm/nouveau/nv04_crtc.c index 5e45398..728d075 100644 --- a/drivers/gpu/drm/nouveau/nv04_crtc.c +++ b/drivers/gpu/drm/nouveau/nv04_crtc.c @@ -364,7 +364,7 @@ nv_crtc_mode_set_vga(struct drm_crtc *crtc, struct drm_display_mode *mode) regp->CRTC[NV_CIO_CR_VRE_INDEX] = 1 << 5 | XLATE(vertEnd, 0, NV_CIO_CR_VRE_3_0); regp->CRTC[NV_CIO_CR_VDE_INDEX] = vertDisplay; /* framebuffer can be larger than crtc scanout area. */ - regp->CRTC[NV_CIO_CR_OFFSET_INDEX] = fb->pitch / 8; + regp->CRTC[NV_CIO_CR_OFFSET_INDEX] = fb->pitches[0] / 8; regp->CRTC[NV_CIO_CR_ULINE_INDEX] = 0x00; regp->CRTC[NV_CIO_CR_VBS_INDEX] = vertBlankStart; regp->CRTC[NV_CIO_CR_VBE_INDEX] = vertBlankEnd; @@ -377,9 +377,9 @@ nv_crtc_mode_set_vga(struct drm_crtc *crtc, struct drm_display_mode *mode) /* framebuffer can be larger than crtc scanout area. */ regp->CRTC[NV_CIO_CRE_RPC0_INDEX] = - XLATE(fb->pitch / 8, 8, NV_CIO_CRE_RPC0_OFFSET_10_8); + XLATE(fb->pitches[0] / 8, 8, NV_CIO_CRE_RPC0_OFFSET_10_8); regp->CRTC[NV_CIO_CRE_42] = - XLATE(fb->pitch / 8, 11, NV_CIO_CRE_42_OFFSET_11); + XLATE(fb->pitches[0] / 8, 11, NV_CIO_CRE_42_OFFSET_11); regp->CRTC[NV_CIO_CRE_RPC1_INDEX] = mode->crtc_hdisplay < 1280 ? MASK(NV_CIO_CRE_RPC1_LARGE) : 0x00; regp->CRTC[NV_CIO_CRE_LSR_INDEX] = XLATE(horizBlankEnd, 6, NV_CIO_CRE_LSR_HBE_6) | @@ -835,18 +835,18 @@ nv04_crtc_do_mode_set_base(struct drm_crtc *crtc, NVWriteRAMDAC(dev, nv_crtc->index, NV_PRAMDAC_GENERAL_CONTROL, regp->ramdac_gen_ctrl); - regp->CRTC[NV_CIO_CR_OFFSET_INDEX] = drm_fb->pitch >> 3; + regp->CRTC[NV_CIO_CR_OFFSET_INDEX] = drm_fb->pitches[0] >> 3; regp->CRTC[NV_CIO_CRE_RPC0_INDEX] = - XLATE(drm_fb->pitch >> 3, 8, NV_CIO_CRE_RPC0_OFFSET_10_8); + XLATE(drm_fb->pitches[0] >> 3, 8, NV_CIO_CRE_RPC0_OFFSET_10_8); regp->CRTC[NV_CIO_CRE_42] = - XLATE(drm_fb->pitch / 8, 11, NV_CIO_CRE_42_OFFSET_11); + XLATE(drm_fb->pitches[0] / 8, 11, NV_CIO_CRE_42_OFFSET_11); crtc_wr_cio_state(crtc, regp, NV_CIO_CRE_RPC0_INDEX); crtc_wr_cio_state(crtc, regp, NV_CIO_CR_OFFSET_INDEX); crtc_wr_cio_state(crtc, regp, NV_CIO_CRE_42); /* Update the framebuffer location. */ regp->fb_start = nv_crtc->fb.offset & ~3; - regp->fb_start += (y * drm_fb->pitch) + (x * drm_fb->bits_per_pixel / 8); + regp->fb_start += (y * drm_fb->pitches[0]) + (x * drm_fb->bits_per_pixel / 8); nv_set_crtc_base(dev, nv_crtc->index, regp->fb_start); /* Update the arbitration parameters. */ diff --git a/drivers/gpu/drm/nouveau/nv04_dac.c b/drivers/gpu/drm/nouveau/nv04_dac.c index 2d6bfd0..f180dcf 100644 --- a/drivers/gpu/drm/nouveau/nv04_dac.c +++ b/drivers/gpu/drm/nouveau/nv04_dac.c @@ -32,6 +32,7 @@ #include "nouveau_connector.h" #include "nouveau_crtc.h" #include "nouveau_hw.h" +#include "nouveau_gpio.h" #include "nvreg.h" int nv04_dac_output_offset(struct drm_encoder *encoder) @@ -220,7 +221,6 @@ uint32_t nv17_dac_sample_load(struct drm_encoder *encoder) { struct drm_device *dev = encoder->dev; struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_gpio_engine *gpio = &dev_priv->engine.gpio; struct dcb_entry *dcb = nouveau_encoder(encoder)->dcb; uint32_t sample, testval, regoffset = nv04_dac_output_offset(encoder); uint32_t saved_powerctrl_2 = 0, saved_powerctrl_4 = 0, saved_routput, @@ -252,11 +252,11 @@ uint32_t nv17_dac_sample_load(struct drm_encoder *encoder) nvWriteMC(dev, NV_PBUS_POWERCTRL_4, saved_powerctrl_4 & 0xffffffcf); } - saved_gpio1 = gpio->get(dev, DCB_GPIO_TVDAC1); - saved_gpio0 = gpio->get(dev, DCB_GPIO_TVDAC0); + saved_gpio1 = nouveau_gpio_func_get(dev, DCB_GPIO_TVDAC1); + saved_gpio0 = nouveau_gpio_func_get(dev, DCB_GPIO_TVDAC0); - gpio->set(dev, DCB_GPIO_TVDAC1, dcb->type == OUTPUT_TV); - gpio->set(dev, DCB_GPIO_TVDAC0, dcb->type == OUTPUT_TV); + nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC1, dcb->type == OUTPUT_TV); + nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC0, dcb->type == OUTPUT_TV); msleep(4); @@ -306,8 +306,8 @@ uint32_t nv17_dac_sample_load(struct drm_encoder *encoder) nvWriteMC(dev, NV_PBUS_POWERCTRL_4, saved_powerctrl_4); nvWriteMC(dev, NV_PBUS_POWERCTRL_2, saved_powerctrl_2); - gpio->set(dev, DCB_GPIO_TVDAC1, saved_gpio1); - gpio->set(dev, DCB_GPIO_TVDAC0, saved_gpio0); + nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC1, saved_gpio1); + nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC0, saved_gpio0); return sample; } diff --git a/drivers/gpu/drm/nouveau/nv04_dfp.c b/drivers/gpu/drm/nouveau/nv04_dfp.c index 752440c..71b6235 100644 --- a/drivers/gpu/drm/nouveau/nv04_dfp.c +++ b/drivers/gpu/drm/nouveau/nv04_dfp.c @@ -289,6 +289,7 @@ static void nv04_dfp_mode_set(struct drm_encoder *encoder, struct nouveau_connector *nv_connector = nouveau_crtc_connector_get(nv_crtc); struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); struct drm_display_mode *output_mode = &nv_encoder->mode; + struct drm_connector *connector = &nv_connector->base; uint32_t mode_ratio, panel_ratio; NV_DEBUG_KMS(dev, "Output mode on CRTC %d:\n", nv_crtc->index); @@ -340,10 +341,15 @@ static void nv04_dfp_mode_set(struct drm_encoder *encoder, output_mode->clock > 165000) regp->fp_control |= (2 << 24); if (nv_encoder->dcb->type == OUTPUT_LVDS) { - bool duallink, dummy; + bool duallink = false, dummy; + if (nv_connector->edid && + nv_connector->type == DCB_CONNECTOR_LVDS_SPWG) { + duallink = (((u8 *)nv_connector->edid)[121] == 2); + } else { + nouveau_bios_parse_lvds_table(dev, output_mode->clock, + &duallink, &dummy); + } - nouveau_bios_parse_lvds_table(dev, output_mode->clock, - &duallink, &dummy); if (duallink) regp->fp_control |= (8 << 28); } else @@ -407,7 +413,9 @@ static void nv04_dfp_mode_set(struct drm_encoder *encoder, } /* Output property. */ - if (nv_connector->use_dithering) { + if ((nv_connector->dithering_mode == DITHERING_MODE_ON) || + (nv_connector->dithering_mode == DITHERING_MODE_AUTO && + encoder->crtc->fb->depth > connector->display_info.bpc * 3)) { if (dev_priv->chipset == 0x11) regp->dither = savep->dither | 0x00010000; else { @@ -496,7 +504,7 @@ static void nv04_dfp_update_backlight(struct drm_encoder *encoder, int mode) static inline bool is_powersaving_dpms(int mode) { - return (mode != DRM_MODE_DPMS_ON); + return mode != DRM_MODE_DPMS_ON && mode != NV_DPMS_CLEARED; } static void nv04_lvds_dpms(struct drm_encoder *encoder, int mode) diff --git a/drivers/gpu/drm/nouveau/nv04_display.c b/drivers/gpu/drm/nouveau/nv04_display.c index 6bd8518..7047d37 100644 --- a/drivers/gpu/drm/nouveau/nv04_display.c +++ b/drivers/gpu/drm/nouveau/nv04_display.c @@ -243,6 +243,11 @@ nv04_display_init(struct drm_device *dev) return 0; } +void +nv04_display_fini(struct drm_device *dev) +{ +} + static void nv04_vblank_crtc0_isr(struct drm_device *dev) { diff --git a/drivers/gpu/drm/nouveau/nv04_fb.c b/drivers/gpu/drm/nouveau/nv04_fb.c index 638cf60..d5eedd6 100644 --- a/drivers/gpu/drm/nouveau/nv04_fb.c +++ b/drivers/gpu/drm/nouveau/nv04_fb.c @@ -4,6 +4,40 @@ #include "nouveau_drm.h" int +nv04_fb_vram_init(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + u32 boot0 = nv_rd32(dev, NV04_PFB_BOOT_0); + + if (boot0 & 0x00000100) { + dev_priv->vram_size = ((boot0 >> 12) & 0xf) * 2 + 2; + dev_priv->vram_size *= 1024 * 1024; + } else { + switch (boot0 & NV04_PFB_BOOT_0_RAM_AMOUNT) { + case NV04_PFB_BOOT_0_RAM_AMOUNT_32MB: + dev_priv->vram_size = 32 * 1024 * 1024; + break; + case NV04_PFB_BOOT_0_RAM_AMOUNT_16MB: + dev_priv->vram_size = 16 * 1024 * 1024; + break; + case NV04_PFB_BOOT_0_RAM_AMOUNT_8MB: + dev_priv->vram_size = 8 * 1024 * 1024; + break; + case NV04_PFB_BOOT_0_RAM_AMOUNT_4MB: + dev_priv->vram_size = 4 * 1024 * 1024; + break; + } + } + + if ((boot0 & 0x00000038) <= 0x10) + dev_priv->vram_type = NV_MEM_TYPE_SGRAM; + else + dev_priv->vram_type = NV_MEM_TYPE_SDRAM; + + return 0; +} + +int nv04_fb_init(struct drm_device *dev) { /* This is what the DDX did for NV_ARCH_04, but a mmio-trace shows diff --git a/drivers/gpu/drm/nouveau/nv04_pm.c b/drivers/gpu/drm/nouveau/nv04_pm.c index 9ae92a8..6e75899 100644 --- a/drivers/gpu/drm/nouveau/nv04_pm.c +++ b/drivers/gpu/drm/nouveau/nv04_pm.c @@ -27,68 +27,111 @@ #include "nouveau_hw.h" #include "nouveau_pm.h" -struct nv04_pm_state { +int +nv04_pm_clocks_get(struct drm_device *dev, struct nouveau_pm_level *perflvl) +{ + int ret; + + ret = nouveau_hw_get_clock(dev, PLL_CORE); + if (ret < 0) + return ret; + perflvl->core = ret; + + ret = nouveau_hw_get_clock(dev, PLL_MEMORY); + if (ret < 0) + return ret; + perflvl->memory = ret; + + return 0; +} + +struct nv04_pm_clock { struct pll_lims pll; struct nouveau_pll_vals calc; }; -int -nv04_pm_clock_get(struct drm_device *dev, u32 id) +struct nv04_pm_state { + struct nv04_pm_clock core; + struct nv04_pm_clock memory; +}; + +static int +calc_pll(struct drm_device *dev, u32 id, int khz, struct nv04_pm_clock *clk) { - return nouveau_hw_get_clock(dev, id); + int ret; + + ret = get_pll_limits(dev, id, &clk->pll); + if (ret) + return ret; + + ret = nouveau_calc_pll_mnp(dev, &clk->pll, khz, &clk->calc); + if (!ret) + return -EINVAL; + + return 0; } void * -nv04_pm_clock_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl, - u32 id, int khz) +nv04_pm_clocks_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl) { - struct nv04_pm_state *state; + struct nv04_pm_state *info; int ret; - state = kzalloc(sizeof(*state), GFP_KERNEL); - if (!state) + info = kzalloc(sizeof(*info), GFP_KERNEL); + if (!info) return ERR_PTR(-ENOMEM); - ret = get_pll_limits(dev, id, &state->pll); - if (ret) { - kfree(state); - return (ret == -ENOENT) ? NULL : ERR_PTR(ret); - } + ret = calc_pll(dev, PLL_CORE, perflvl->core, &info->core); + if (ret) + goto error; - ret = nouveau_calc_pll_mnp(dev, &state->pll, khz, &state->calc); - if (!ret) { - kfree(state); - return ERR_PTR(-EINVAL); + if (perflvl->memory) { + ret = calc_pll(dev, PLL_MEMORY, perflvl->memory, &info->memory); + if (ret) + goto error; } - return state; + return info; +error: + kfree(info); + return ERR_PTR(ret); } -void -nv04_pm_clock_set(struct drm_device *dev, void *pre_state) +static void +prog_pll(struct drm_device *dev, struct nv04_pm_clock *clk) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_timer_engine *ptimer = &dev_priv->engine.timer; - struct nv04_pm_state *state = pre_state; - u32 reg = state->pll.reg; + u32 reg = clk->pll.reg; /* thank the insane nouveau_hw_setpll() interface for this */ if (dev_priv->card_type >= NV_40) reg += 4; - nouveau_hw_setpll(dev, reg, &state->calc); + nouveau_hw_setpll(dev, reg, &clk->calc); +} + +int +nv04_pm_clocks_set(struct drm_device *dev, void *pre_state) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_timer_engine *ptimer = &dev_priv->engine.timer; + struct nv04_pm_state *state = pre_state; + + prog_pll(dev, &state->core); - if (dev_priv->card_type < NV_30 && reg == NV_PRAMDAC_MPLL_COEFF) { - if (dev_priv->card_type == NV_20) - nv_mask(dev, 0x1002c4, 0, 1 << 20); + if (state->memory.pll.reg) { + prog_pll(dev, &state->memory); + if (dev_priv->card_type < NV_30) { + if (dev_priv->card_type == NV_20) + nv_mask(dev, 0x1002c4, 0, 1 << 20); - /* Reset the DLLs */ - nv_mask(dev, 0x1002c0, 0, 1 << 8); + /* Reset the DLLs */ + nv_mask(dev, 0x1002c0, 0, 1 << 8); + } } - if (reg == NV_PRAMDAC_NVPLL_COEFF) - ptimer->init(dev); + ptimer->init(dev); kfree(state); + return 0; } - diff --git a/drivers/gpu/drm/nouveau/nv04_timer.c b/drivers/gpu/drm/nouveau/nv04_timer.c index 263301b..55c9452 100644 --- a/drivers/gpu/drm/nouveau/nv04_timer.c +++ b/drivers/gpu/drm/nouveau/nv04_timer.c @@ -2,6 +2,7 @@ #include "drm.h" #include "nouveau_drv.h" #include "nouveau_drm.h" +#include "nouveau_hw.h" int nv04_timer_init(struct drm_device *dev) @@ -17,7 +18,7 @@ nv04_timer_init(struct drm_device *dev) /* determine base clock for timer source */ if (dev_priv->chipset < 0x40) { - n = dev_priv->engine.pm.clock_get(dev, PLL_CORE); + n = nouveau_hw_get_clock(dev, PLL_CORE); } else if (dev_priv->chipset == 0x40) { /*XXX: figure this out */ diff --git a/drivers/gpu/drm/nouveau/nv10_fb.c b/drivers/gpu/drm/nouveau/nv10_fb.c index f78181a..420b1608 100644 --- a/drivers/gpu/drm/nouveau/nv10_fb.c +++ b/drivers/gpu/drm/nouveau/nv10_fb.c @@ -3,81 +3,16 @@ #include "nouveau_drv.h" #include "nouveau_drm.h" -static struct drm_mm_node * -nv20_fb_alloc_tag(struct drm_device *dev, uint32_t size) -{ - struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_fb_engine *pfb = &dev_priv->engine.fb; - struct drm_mm_node *mem; - int ret; - - ret = drm_mm_pre_get(&pfb->tag_heap); - if (ret) - return NULL; - - spin_lock(&dev_priv->tile.lock); - mem = drm_mm_search_free(&pfb->tag_heap, size, 0, 0); - if (mem) - mem = drm_mm_get_block_atomic(mem, size, 0); - spin_unlock(&dev_priv->tile.lock); - - return mem; -} - -static void -nv20_fb_free_tag(struct drm_device *dev, struct drm_mm_node *mem) -{ - struct drm_nouveau_private *dev_priv = dev->dev_private; - - spin_lock(&dev_priv->tile.lock); - drm_mm_put_block(mem); - spin_unlock(&dev_priv->tile.lock); -} - void nv10_fb_init_tile_region(struct drm_device *dev, int i, uint32_t addr, uint32_t size, uint32_t pitch, uint32_t flags) { struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_tile_reg *tile = &dev_priv->tile.reg[i]; - int bpp = (flags & NOUVEAU_GEM_TILE_32BPP ? 32 : 16); - tile->addr = addr; + tile->addr = 0x80000000 | addr; tile->limit = max(1u, addr + size) - 1; tile->pitch = pitch; - - if (dev_priv->card_type == NV_20) { - if (flags & NOUVEAU_GEM_TILE_ZETA) { - /* - * Allocate some of the on-die tag memory, - * used to store Z compression meta-data (most - * likely just a bitmap determining if a given - * tile is compressed or not). - */ - tile->tag_mem = nv20_fb_alloc_tag(dev, size / 256); - - if (tile->tag_mem) { - /* Enable Z compression */ - if (dev_priv->chipset >= 0x25) - tile->zcomp = tile->tag_mem->start | - (bpp == 16 ? - NV25_PFB_ZCOMP_MODE_16 : - NV25_PFB_ZCOMP_MODE_32); - else - tile->zcomp = tile->tag_mem->start | - NV20_PFB_ZCOMP_EN | - (bpp == 16 ? 0 : - NV20_PFB_ZCOMP_MODE_32); - } - - tile->addr |= 3; - } else { - tile->addr |= 1; - } - - } else { - tile->addr |= 1 << 31; - } } void @@ -86,11 +21,6 @@ nv10_fb_free_tile_region(struct drm_device *dev, int i) struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_tile_reg *tile = &dev_priv->tile.reg[i]; - if (tile->tag_mem) { - nv20_fb_free_tag(dev, tile->tag_mem); - tile->tag_mem = NULL; - } - tile->addr = tile->limit = tile->pitch = tile->zcomp = 0; } @@ -103,9 +33,48 @@ nv10_fb_set_tile_region(struct drm_device *dev, int i) nv_wr32(dev, NV10_PFB_TLIMIT(i), tile->limit); nv_wr32(dev, NV10_PFB_TSIZE(i), tile->pitch); nv_wr32(dev, NV10_PFB_TILE(i), tile->addr); +} + +int +nv1a_fb_vram_init(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct pci_dev *bridge; + uint32_t mem, mib; + + bridge = pci_get_bus_and_slot(0, PCI_DEVFN(0, 1)); + if (!bridge) { + NV_ERROR(dev, "no bridge device\n"); + return 0; + } + + if (dev_priv->chipset == 0x1a) { + pci_read_config_dword(bridge, 0x7c, &mem); + mib = ((mem >> 6) & 31) + 1; + } else { + pci_read_config_dword(bridge, 0x84, &mem); + mib = ((mem >> 4) & 127) + 1; + } + + dev_priv->vram_size = mib * 1024 * 1024; + return 0; +} + +int +nv10_fb_vram_init(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + u32 fifo_data = nv_rd32(dev, NV04_PFB_FIFO_DATA); + u32 cfg0 = nv_rd32(dev, 0x100200); - if (dev_priv->card_type == NV_20) - nv_wr32(dev, NV20_PFB_ZCOMP(i), tile->zcomp); + dev_priv->vram_size = fifo_data & NV10_PFB_FIFO_DATA_RAM_AMOUNT_MB_MASK; + + if (cfg0 & 0x00000001) + dev_priv->vram_type = NV_MEM_TYPE_DDR1; + else + dev_priv->vram_type = NV_MEM_TYPE_SDRAM; + + return 0; } int @@ -115,14 +84,8 @@ nv10_fb_init(struct drm_device *dev) struct nouveau_fb_engine *pfb = &dev_priv->engine.fb; int i; - pfb->num_tiles = NV10_PFB_TILE__SIZE; - - if (dev_priv->card_type == NV_20) - drm_mm_init(&pfb->tag_heap, 0, - (dev_priv->chipset >= 0x25 ? - 64 * 1024 : 32 * 1024)); - /* Turn all the tiling regions off. */ + pfb->num_tiles = NV10_PFB_TILE__SIZE; for (i = 0; i < pfb->num_tiles; i++) pfb->set_tile_region(dev, i); @@ -138,7 +101,4 @@ nv10_fb_takedown(struct drm_device *dev) for (i = 0; i < pfb->num_tiles; i++) pfb->free_tile_region(dev, i); - - if (dev_priv->card_type == NV_20) - drm_mm_takedown(&pfb->tag_heap); } diff --git a/drivers/gpu/drm/nouveau/nv10_gpio.c b/drivers/gpu/drm/nouveau/nv10_gpio.c index 007fc29..9d79180 100644 --- a/drivers/gpu/drm/nouveau/nv10_gpio.c +++ b/drivers/gpu/drm/nouveau/nv10_gpio.c @@ -27,66 +27,97 @@ #include "drmP.h" #include "nouveau_drv.h" #include "nouveau_hw.h" +#include "nouveau_gpio.h" -static bool -get_gpio_location(struct dcb_gpio_entry *ent, uint32_t *reg, uint32_t *shift, - uint32_t *mask) +int +nv10_gpio_sense(struct drm_device *dev, int line) { - if (ent->line < 2) { - *reg = NV_PCRTC_GPIO; - *shift = ent->line * 16; - *mask = 0x11; - - } else if (ent->line < 10) { - *reg = NV_PCRTC_GPIO_EXT; - *shift = (ent->line - 2) * 4; - *mask = 0x3; + if (line < 2) { + line = line * 16; + line = NVReadCRTC(dev, 0, NV_PCRTC_GPIO) >> line; + return !!(line & 0x0100); + } else + if (line < 10) { + line = (line - 2) * 4; + line = NVReadCRTC(dev, 0, NV_PCRTC_GPIO_EXT) >> line; + return !!(line & 0x04); + } else + if (line < 14) { + line = (line - 10) * 4; + line = NVReadCRTC(dev, 0, NV_PCRTC_850) >> line; + return !!(line & 0x04); + } - } else if (ent->line < 14) { - *reg = NV_PCRTC_850; - *shift = (ent->line - 10) * 4; - *mask = 0x3; + return -EINVAL; +} +int +nv10_gpio_drive(struct drm_device *dev, int line, int dir, int out) +{ + u32 reg, mask, data; + + if (line < 2) { + line = line * 16; + reg = NV_PCRTC_GPIO; + mask = 0x00000011; + data = (dir << 4) | out; + } else + if (line < 10) { + line = (line - 2) * 4; + reg = NV_PCRTC_GPIO_EXT; + mask = 0x00000003; + data = (dir << 1) | out; + } else + if (line < 14) { + line = (line - 10) * 4; + reg = NV_PCRTC_850; + mask = 0x00000003; + data = (dir << 1) | out; } else { - return false; + return -EINVAL; } - return true; + mask = NVReadCRTC(dev, 0, reg) & ~(mask << line); + NVWriteCRTC(dev, 0, reg, mask | (data << line)); + return 0; } -int -nv10_gpio_get(struct drm_device *dev, enum dcb_gpio_tag tag) +void +nv10_gpio_irq_enable(struct drm_device *dev, int line, bool on) { - struct dcb_gpio_entry *ent = nouveau_bios_gpio_entry(dev, tag); - uint32_t reg, shift, mask, value; + u32 mask = 0x00010001 << line; - if (!ent) - return -ENODEV; + nv_wr32(dev, 0x001104, mask); + nv_mask(dev, 0x001144, mask, on ? mask : 0); +} - if (!get_gpio_location(ent, ®, &shift, &mask)) - return -ENODEV; +static void +nv10_gpio_isr(struct drm_device *dev) +{ + u32 intr = nv_rd32(dev, 0x1104); + u32 hi = (intr & 0x0000ffff) >> 0; + u32 lo = (intr & 0xffff0000) >> 16; - value = NVReadCRTC(dev, 0, reg) >> shift; + nouveau_gpio_isr(dev, 0, hi | lo); - return (ent->invert ? 1 : 0) ^ (value & 1); + nv_wr32(dev, 0x001104, intr); } int -nv10_gpio_set(struct drm_device *dev, enum dcb_gpio_tag tag, int state) +nv10_gpio_init(struct drm_device *dev) { - struct dcb_gpio_entry *ent = nouveau_bios_gpio_entry(dev, tag); - uint32_t reg, shift, mask, value; - - if (!ent) - return -ENODEV; - - if (!get_gpio_location(ent, ®, &shift, &mask)) - return -ENODEV; - - value = ((ent->invert ? 1 : 0) ^ (state ? 1 : 0)) << shift; - mask = ~(mask << shift); - - NVWriteCRTC(dev, 0, reg, value | (NVReadCRTC(dev, 0, reg) & mask)); - + nv_wr32(dev, 0x001140, 0x00000000); + nv_wr32(dev, 0x001100, 0xffffffff); + nv_wr32(dev, 0x001144, 0x00000000); + nv_wr32(dev, 0x001104, 0xffffffff); + nouveau_irq_register(dev, 28, nv10_gpio_isr); /* PBUS */ return 0; } + +void +nv10_gpio_fini(struct drm_device *dev) +{ + nv_wr32(dev, 0x001140, 0x00000000); + nv_wr32(dev, 0x001144, 0x00000000); + nouveau_irq_unregister(dev, 28); +} diff --git a/drivers/gpu/drm/nouveau/nv17_tv.c b/drivers/gpu/drm/nouveau/nv17_tv.c index 3900ceb..696d7e7 100644 --- a/drivers/gpu/drm/nouveau/nv17_tv.c +++ b/drivers/gpu/drm/nouveau/nv17_tv.c @@ -30,6 +30,7 @@ #include "nouveau_encoder.h" #include "nouveau_connector.h" #include "nouveau_crtc.h" +#include "nouveau_gpio.h" #include "nouveau_hw.h" #include "nv17_tv.h" @@ -37,7 +38,6 @@ static uint32_t nv42_tv_sample_load(struct drm_encoder *encoder) { struct drm_device *dev = encoder->dev; struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_gpio_engine *gpio = &dev_priv->engine.gpio; uint32_t testval, regoffset = nv04_dac_output_offset(encoder); uint32_t gpio0, gpio1, fp_htotal, fp_hsync_start, fp_hsync_end, fp_control, test_ctrl, dacclk, ctv_14, ctv_1c, ctv_6c; @@ -53,8 +53,8 @@ static uint32_t nv42_tv_sample_load(struct drm_encoder *encoder) head = (dacclk & 0x100) >> 8; /* Save the previous state. */ - gpio1 = gpio->get(dev, DCB_GPIO_TVDAC1); - gpio0 = gpio->get(dev, DCB_GPIO_TVDAC0); + gpio1 = nouveau_gpio_func_get(dev, DCB_GPIO_TVDAC1); + gpio0 = nouveau_gpio_func_get(dev, DCB_GPIO_TVDAC0); fp_htotal = NVReadRAMDAC(dev, head, NV_PRAMDAC_FP_HTOTAL); fp_hsync_start = NVReadRAMDAC(dev, head, NV_PRAMDAC_FP_HSYNC_START); fp_hsync_end = NVReadRAMDAC(dev, head, NV_PRAMDAC_FP_HSYNC_END); @@ -65,8 +65,8 @@ static uint32_t nv42_tv_sample_load(struct drm_encoder *encoder) ctv_6c = NVReadRAMDAC(dev, head, 0x680c6c); /* Prepare the DAC for load detection. */ - gpio->set(dev, DCB_GPIO_TVDAC1, true); - gpio->set(dev, DCB_GPIO_TVDAC0, true); + nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC1, true); + nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC0, true); NVWriteRAMDAC(dev, head, NV_PRAMDAC_FP_HTOTAL, 1343); NVWriteRAMDAC(dev, head, NV_PRAMDAC_FP_HSYNC_START, 1047); @@ -111,8 +111,8 @@ static uint32_t nv42_tv_sample_load(struct drm_encoder *encoder) NVWriteRAMDAC(dev, head, NV_PRAMDAC_FP_HSYNC_END, fp_hsync_end); NVWriteRAMDAC(dev, head, NV_PRAMDAC_FP_HSYNC_START, fp_hsync_start); NVWriteRAMDAC(dev, head, NV_PRAMDAC_FP_HTOTAL, fp_htotal); - gpio->set(dev, DCB_GPIO_TVDAC1, gpio1); - gpio->set(dev, DCB_GPIO_TVDAC0, gpio0); + nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC1, gpio1); + nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC0, gpio0); return sample; } @@ -357,8 +357,6 @@ static bool nv17_tv_mode_fixup(struct drm_encoder *encoder, static void nv17_tv_dpms(struct drm_encoder *encoder, int mode) { struct drm_device *dev = encoder->dev; - struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_gpio_engine *gpio = &dev_priv->engine.gpio; struct nv17_tv_state *regs = &to_tv_enc(encoder)->state; struct nv17_tv_norm_params *tv_norm = get_tv_norm(encoder); @@ -383,8 +381,8 @@ static void nv17_tv_dpms(struct drm_encoder *encoder, int mode) nv_load_ptv(dev, regs, 200); - gpio->set(dev, DCB_GPIO_TVDAC1, mode == DRM_MODE_DPMS_ON); - gpio->set(dev, DCB_GPIO_TVDAC0, mode == DRM_MODE_DPMS_ON); + nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC1, mode == DRM_MODE_DPMS_ON); + nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC0, mode == DRM_MODE_DPMS_ON); nv04_dac_update_dacclk(encoder, mode == DRM_MODE_DPMS_ON); } diff --git a/drivers/gpu/drm/nouveau/nv20_fb.c b/drivers/gpu/drm/nouveau/nv20_fb.c new file mode 100644 index 0000000..19bd640 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nv20_fb.c @@ -0,0 +1,148 @@ +#include "drmP.h" +#include "drm.h" +#include "nouveau_drv.h" +#include "nouveau_drm.h" + +static struct drm_mm_node * +nv20_fb_alloc_tag(struct drm_device *dev, uint32_t size) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_fb_engine *pfb = &dev_priv->engine.fb; + struct drm_mm_node *mem; + int ret; + + ret = drm_mm_pre_get(&pfb->tag_heap); + if (ret) + return NULL; + + spin_lock(&dev_priv->tile.lock); + mem = drm_mm_search_free(&pfb->tag_heap, size, 0, 0); + if (mem) + mem = drm_mm_get_block_atomic(mem, size, 0); + spin_unlock(&dev_priv->tile.lock); + + return mem; +} + +static void +nv20_fb_free_tag(struct drm_device *dev, struct drm_mm_node **pmem) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct drm_mm_node *mem = *pmem; + if (mem) { + spin_lock(&dev_priv->tile.lock); + drm_mm_put_block(mem); + spin_unlock(&dev_priv->tile.lock); + *pmem = NULL; + } +} + +void +nv20_fb_init_tile_region(struct drm_device *dev, int i, uint32_t addr, + uint32_t size, uint32_t pitch, uint32_t flags) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_tile_reg *tile = &dev_priv->tile.reg[i]; + int bpp = (flags & NOUVEAU_GEM_TILE_32BPP ? 32 : 16); + + tile->addr = 0x00000001 | addr; + tile->limit = max(1u, addr + size) - 1; + tile->pitch = pitch; + + /* Allocate some of the on-die tag memory, used to store Z + * compression meta-data (most likely just a bitmap determining + * if a given tile is compressed or not). + */ + if (flags & NOUVEAU_GEM_TILE_ZETA) { + tile->tag_mem = nv20_fb_alloc_tag(dev, size / 256); + if (tile->tag_mem) { + /* Enable Z compression */ + tile->zcomp = tile->tag_mem->start; + if (dev_priv->chipset >= 0x25) { + if (bpp == 16) + tile->zcomp |= NV25_PFB_ZCOMP_MODE_16; + else + tile->zcomp |= NV25_PFB_ZCOMP_MODE_32; + } else { + tile->zcomp |= NV20_PFB_ZCOMP_EN; + if (bpp != 16) + tile->zcomp |= NV20_PFB_ZCOMP_MODE_32; + } + } + + tile->addr |= 2; + } +} + +void +nv20_fb_free_tile_region(struct drm_device *dev, int i) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_tile_reg *tile = &dev_priv->tile.reg[i]; + + tile->addr = tile->limit = tile->pitch = tile->zcomp = 0; + nv20_fb_free_tag(dev, &tile->tag_mem); +} + +void +nv20_fb_set_tile_region(struct drm_device *dev, int i) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_tile_reg *tile = &dev_priv->tile.reg[i]; + + nv_wr32(dev, NV10_PFB_TLIMIT(i), tile->limit); + nv_wr32(dev, NV10_PFB_TSIZE(i), tile->pitch); + nv_wr32(dev, NV10_PFB_TILE(i), tile->addr); + nv_wr32(dev, NV20_PFB_ZCOMP(i), tile->zcomp); +} + +int +nv20_fb_vram_init(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + u32 mem_size = nv_rd32(dev, 0x10020c); + u32 pbus1218 = nv_rd32(dev, 0x001218); + + dev_priv->vram_size = mem_size & 0xff000000; + switch (pbus1218 & 0x00000300) { + case 0x00000000: dev_priv->vram_type = NV_MEM_TYPE_SDRAM; break; + case 0x00000100: dev_priv->vram_type = NV_MEM_TYPE_DDR1; break; + case 0x00000200: dev_priv->vram_type = NV_MEM_TYPE_GDDR3; break; + case 0x00000300: dev_priv->vram_type = NV_MEM_TYPE_GDDR2; break; + } + + return 0; +} + +int +nv20_fb_init(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_fb_engine *pfb = &dev_priv->engine.fb; + int i; + + if (dev_priv->chipset >= 0x25) + drm_mm_init(&pfb->tag_heap, 0, 64 * 1024); + else + drm_mm_init(&pfb->tag_heap, 0, 32 * 1024); + + /* Turn all the tiling regions off. */ + pfb->num_tiles = NV10_PFB_TILE__SIZE; + for (i = 0; i < pfb->num_tiles; i++) + pfb->set_tile_region(dev, i); + + return 0; +} + +void +nv20_fb_takedown(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_fb_engine *pfb = &dev_priv->engine.fb; + int i; + + for (i = 0; i < pfb->num_tiles; i++) + pfb->free_tile_region(dev, i); + + drm_mm_takedown(&pfb->tag_heap); +} diff --git a/drivers/gpu/drm/nouveau/nv40_fb.c b/drivers/gpu/drm/nouveau/nv40_fb.c index f0ac2a7..7fbcb33 100644 --- a/drivers/gpu/drm/nouveau/nv40_fb.c +++ b/drivers/gpu/drm/nouveau/nv40_fb.c @@ -72,6 +72,51 @@ nv44_fb_init_gart(struct drm_device *dev) } int +nv40_fb_vram_init(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + + /* 0x001218 is actually present on a few other NV4X I looked at, + * and even contains sane values matching 0x100474. From looking + * at various vbios images however, this isn't the case everywhere. + * So, I chose to use the same regs I've seen NVIDIA reading around + * the memory detection, hopefully that'll get us the right numbers + */ + if (dev_priv->chipset == 0x40) { + u32 pbus1218 = nv_rd32(dev, 0x001218); + switch (pbus1218 & 0x00000300) { + case 0x00000000: dev_priv->vram_type = NV_MEM_TYPE_SDRAM; break; + case 0x00000100: dev_priv->vram_type = NV_MEM_TYPE_DDR1; break; + case 0x00000200: dev_priv->vram_type = NV_MEM_TYPE_GDDR3; break; + case 0x00000300: dev_priv->vram_type = NV_MEM_TYPE_DDR2; break; + } + } else + if (dev_priv->chipset == 0x49 || dev_priv->chipset == 0x4b) { + u32 pfb914 = nv_rd32(dev, 0x100914); + switch (pfb914 & 0x00000003) { + case 0x00000000: dev_priv->vram_type = NV_MEM_TYPE_DDR1; break; + case 0x00000001: dev_priv->vram_type = NV_MEM_TYPE_DDR2; break; + case 0x00000002: dev_priv->vram_type = NV_MEM_TYPE_GDDR3; break; + case 0x00000003: break; + } + } else + if (dev_priv->chipset != 0x4e) { + u32 pfb474 = nv_rd32(dev, 0x100474); + if (pfb474 & 0x00000004) + dev_priv->vram_type = NV_MEM_TYPE_GDDR3; + if (pfb474 & 0x00000002) + dev_priv->vram_type = NV_MEM_TYPE_DDR2; + if (pfb474 & 0x00000001) + dev_priv->vram_type = NV_MEM_TYPE_DDR1; + } else { + dev_priv->vram_type = NV_MEM_TYPE_STOLEN; + } + + dev_priv->vram_size = nv_rd32(dev, 0x10020c) & 0xff000000; + return 0; +} + +int nv40_fb_init(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; diff --git a/drivers/gpu/drm/nouveau/nv40_pm.c b/drivers/gpu/drm/nouveau/nv40_pm.c index e676b0d..c761538 100644 --- a/drivers/gpu/drm/nouveau/nv40_pm.c +++ b/drivers/gpu/drm/nouveau/nv40_pm.c @@ -222,7 +222,7 @@ nv40_pm_gr_idle(void *data) return true; } -void +int nv40_pm_clocks_set(struct drm_device *dev, void *pre_state) { struct drm_nouveau_private *dev_priv = dev->dev_private; @@ -231,7 +231,7 @@ nv40_pm_clocks_set(struct drm_device *dev, void *pre_state) struct bit_entry M; u32 crtc_mask = 0; u8 sr1[2]; - int i; + int i, ret = -EAGAIN; /* determine which CRTCs are active, fetch VGA_SR1 for each */ for (i = 0; i < 2; i++) { @@ -263,6 +263,8 @@ nv40_pm_clocks_set(struct drm_device *dev, void *pre_state) if (!nv_wait_cb(dev, nv40_pm_gr_idle, dev)) goto resume; + ret = 0; + /* set engine clocks */ nv_mask(dev, 0x00c040, 0x00000333, 0x00000000); nv_wr32(dev, 0x004004, info->npll_coef); @@ -345,4 +347,48 @@ resume: spin_unlock_irqrestore(&dev_priv->context_switch_lock, flags); kfree(info); + return ret; +} + +int +nv40_pm_pwm_get(struct drm_device *dev, int line, u32 *divs, u32 *duty) +{ + if (line == 2) { + u32 reg = nv_rd32(dev, 0x0010f0); + if (reg & 0x80000000) { + *duty = (reg & 0x7fff0000) >> 16; + *divs = (reg & 0x00007fff); + return 0; + } + } else + if (line == 9) { + u32 reg = nv_rd32(dev, 0x0015f4); + if (reg & 0x80000000) { + *divs = nv_rd32(dev, 0x0015f8); + *duty = (reg & 0x7fffffff); + return 0; + } + } else { + NV_ERROR(dev, "unknown pwm ctrl for gpio %d\n", line); + return -ENODEV; + } + + return -EINVAL; +} + +int +nv40_pm_pwm_set(struct drm_device *dev, int line, u32 divs, u32 duty) +{ + if (line == 2) { + nv_wr32(dev, 0x0010f0, 0x80000000 | (duty << 16) | divs); + } else + if (line == 9) { + nv_wr32(dev, 0x0015f8, divs); + nv_wr32(dev, 0x0015f4, duty | 0x80000000); + } else { + NV_ERROR(dev, "unknown pwm ctrl for gpio %d\n", line); + return -ENODEV; + } + + return 0; } diff --git a/drivers/gpu/drm/nouveau/nv50_crtc.c b/drivers/gpu/drm/nouveau/nv50_crtc.c index 882080e..701b927 100644 --- a/drivers/gpu/drm/nouveau/nv50_crtc.c +++ b/drivers/gpu/drm/nouveau/nv50_crtc.c @@ -132,30 +132,74 @@ nv50_crtc_blank(struct nouveau_crtc *nv_crtc, bool blanked) } static int -nv50_crtc_set_dither(struct nouveau_crtc *nv_crtc, bool on, bool update) +nv50_crtc_set_dither(struct nouveau_crtc *nv_crtc, bool update) +{ + struct nouveau_channel *evo = nv50_display(nv_crtc->base.dev)->master; + struct nouveau_connector *nv_connector; + struct drm_connector *connector; + int head = nv_crtc->index, ret; + u32 mode = 0x00; + + nv_connector = nouveau_crtc_connector_get(nv_crtc); + connector = &nv_connector->base; + if (nv_connector->dithering_mode == DITHERING_MODE_AUTO) { + if (nv_crtc->base.fb->depth > connector->display_info.bpc * 3) + mode = DITHERING_MODE_DYNAMIC2X2; + } else { + mode = nv_connector->dithering_mode; + } + + if (nv_connector->dithering_depth == DITHERING_DEPTH_AUTO) { + if (connector->display_info.bpc >= 8) + mode |= DITHERING_DEPTH_8BPC; + } else { + mode |= nv_connector->dithering_depth; + } + + ret = RING_SPACE(evo, 2 + (update ? 2 : 0)); + if (ret == 0) { + BEGIN_RING(evo, 0, NV50_EVO_CRTC(head, DITHER_CTRL), 1); + OUT_RING (evo, mode); + if (update) { + BEGIN_RING(evo, 0, NV50_EVO_UPDATE, 1); + OUT_RING (evo, 0); + FIRE_RING (evo); + } + } + + return ret; +} + +static int +nv50_crtc_set_color_vibrance(struct nouveau_crtc *nv_crtc, bool update) { struct drm_device *dev = nv_crtc->base.dev; struct nouveau_channel *evo = nv50_display(dev)->master; int ret; + int adj; + u32 hue, vib; - NV_DEBUG_KMS(dev, "\n"); + NV_DEBUG_KMS(dev, "vibrance = %i, hue = %i\n", + nv_crtc->color_vibrance, nv_crtc->vibrant_hue); ret = RING_SPACE(evo, 2 + (update ? 2 : 0)); if (ret) { - NV_ERROR(dev, "no space while setting dither\n"); + NV_ERROR(dev, "no space while setting color vibrance\n"); return ret; } - BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, DITHER_CTRL), 1); - if (on) - OUT_RING(evo, NV50_EVO_CRTC_DITHER_CTRL_ON); - else - OUT_RING(evo, NV50_EVO_CRTC_DITHER_CTRL_OFF); + adj = (nv_crtc->color_vibrance > 0) ? 50 : 0; + vib = ((nv_crtc->color_vibrance * 2047 + adj) / 100) & 0xfff; + + hue = ((nv_crtc->vibrant_hue * 2047) / 100) & 0xfff; + + BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, COLOR_CTRL), 1); + OUT_RING (evo, (hue << 20) | (vib << 8)); if (update) { BEGIN_RING(evo, 0, NV50_EVO_UPDATE, 1); - OUT_RING(evo, 0); - FIRE_RING(evo); + OUT_RING (evo, 0); + FIRE_RING (evo); } return 0; @@ -180,80 +224,103 @@ nouveau_crtc_connector_get(struct nouveau_crtc *nv_crtc) } static int -nv50_crtc_set_scale(struct nouveau_crtc *nv_crtc, int scaling_mode, bool update) +nv50_crtc_set_scale(struct nouveau_crtc *nv_crtc, bool update) { - struct nouveau_connector *nv_connector = - nouveau_crtc_connector_get(nv_crtc); - struct drm_device *dev = nv_crtc->base.dev; + struct nouveau_connector *nv_connector; + struct drm_crtc *crtc = &nv_crtc->base; + struct drm_device *dev = crtc->dev; struct nouveau_channel *evo = nv50_display(dev)->master; - struct drm_display_mode *native_mode = NULL; - struct drm_display_mode *mode = &nv_crtc->base.mode; - uint32_t outX, outY, horiz, vert; - int ret; + struct drm_display_mode *umode = &crtc->mode; + struct drm_display_mode *omode; + int scaling_mode, ret; + u32 ctrl = 0, oX, oY; NV_DEBUG_KMS(dev, "\n"); - switch (scaling_mode) { - case DRM_MODE_SCALE_NONE: - break; - default: - if (!nv_connector || !nv_connector->native_mode) { - NV_ERROR(dev, "No native mode, forcing panel scaling\n"); - scaling_mode = DRM_MODE_SCALE_NONE; + nv_connector = nouveau_crtc_connector_get(nv_crtc); + if (!nv_connector || !nv_connector->native_mode) { + NV_ERROR(dev, "no native mode, forcing panel scaling\n"); + scaling_mode = DRM_MODE_SCALE_NONE; + } else { + scaling_mode = nv_connector->scaling_mode; + } + + /* start off at the resolution we programmed the crtc for, this + * effectively handles NONE/FULL scaling + */ + if (scaling_mode != DRM_MODE_SCALE_NONE) + omode = nv_connector->native_mode; + else + omode = umode; + + oX = omode->hdisplay; + oY = omode->vdisplay; + if (omode->flags & DRM_MODE_FLAG_DBLSCAN) + oY *= 2; + + /* add overscan compensation if necessary, will keep the aspect + * ratio the same as the backend mode unless overridden by the + * user setting both hborder and vborder properties. + */ + if (nv_connector && ( nv_connector->underscan == UNDERSCAN_ON || + (nv_connector->underscan == UNDERSCAN_AUTO && + nv_connector->edid && + drm_detect_hdmi_monitor(nv_connector->edid)))) { + u32 bX = nv_connector->underscan_hborder; + u32 bY = nv_connector->underscan_vborder; + u32 aspect = (oY << 19) / oX; + + if (bX) { + oX -= (bX * 2); + if (bY) oY -= (bY * 2); + else oY = ((oX * aspect) + (aspect / 2)) >> 19; } else { - native_mode = nv_connector->native_mode; + oX -= (oX >> 4) + 32; + if (bY) oY -= (bY * 2); + else oY = ((oX * aspect) + (aspect / 2)) >> 19; } - break; } + /* handle CENTER/ASPECT scaling, taking into account the areas + * removed already for overscan compensation + */ switch (scaling_mode) { + case DRM_MODE_SCALE_CENTER: + oX = min((u32)umode->hdisplay, oX); + oY = min((u32)umode->vdisplay, oY); + /* fall-through */ case DRM_MODE_SCALE_ASPECT: - horiz = (native_mode->hdisplay << 19) / mode->hdisplay; - vert = (native_mode->vdisplay << 19) / mode->vdisplay; - - if (vert > horiz) { - outX = (mode->hdisplay * horiz) >> 19; - outY = (mode->vdisplay * horiz) >> 19; + if (oY < oX) { + u32 aspect = (umode->hdisplay << 19) / umode->vdisplay; + oX = ((oY * aspect) + (aspect / 2)) >> 19; } else { - outX = (mode->hdisplay * vert) >> 19; - outY = (mode->vdisplay * vert) >> 19; + u32 aspect = (umode->vdisplay << 19) / umode->hdisplay; + oY = ((oX * aspect) + (aspect / 2)) >> 19; } break; - case DRM_MODE_SCALE_FULLSCREEN: - outX = native_mode->hdisplay; - outY = native_mode->vdisplay; - break; - case DRM_MODE_SCALE_CENTER: - case DRM_MODE_SCALE_NONE: default: - outX = mode->hdisplay; - outY = mode->vdisplay; break; } - ret = RING_SPACE(evo, update ? 7 : 5); + if (umode->hdisplay != oX || umode->vdisplay != oY || + umode->flags & DRM_MODE_FLAG_INTERLACE || + umode->flags & DRM_MODE_FLAG_DBLSCAN) + ctrl |= NV50_EVO_CRTC_SCALE_CTRL_ACTIVE; + + ret = RING_SPACE(evo, 5); if (ret) return ret; - /* Got a better name for SCALER_ACTIVE? */ - /* One day i've got to really figure out why this is needed. */ BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, SCALE_CTRL), 1); - if ((mode->flags & DRM_MODE_FLAG_DBLSCAN) || - (mode->flags & DRM_MODE_FLAG_INTERLACE) || - mode->hdisplay != outX || mode->vdisplay != outY) { - OUT_RING(evo, NV50_EVO_CRTC_SCALE_CTRL_ACTIVE); - } else { - OUT_RING(evo, NV50_EVO_CRTC_SCALE_CTRL_INACTIVE); - } - + OUT_RING (evo, ctrl); BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, SCALE_RES1), 2); - OUT_RING(evo, outY << 16 | outX); - OUT_RING(evo, outY << 16 | outX); + OUT_RING (evo, oY << 16 | oX); + OUT_RING (evo, oY << 16 | oX); if (update) { - BEGIN_RING(evo, 0, NV50_EVO_UPDATE, 1); - OUT_RING(evo, 0); - FIRE_RING(evo); + nv50_display_flip_stop(crtc); + nv50_display_sync(dev); + nv50_display_flip_next(crtc, crtc->fb, NULL); } return 0; @@ -333,7 +400,6 @@ nv50_crtc_destroy(struct drm_crtc *crtc) nouveau_bo_ref(NULL, &nv_crtc->lut.nvbo); nouveau_bo_unmap(nv_crtc->cursor.nvbo); nouveau_bo_ref(NULL, &nv_crtc->cursor.nvbo); - kfree(nv_crtc->mode); kfree(nv_crtc); } @@ -441,39 +507,6 @@ nv50_crtc_dpms(struct drm_crtc *crtc, int mode) { } -static int -nv50_crtc_wait_complete(struct drm_crtc *crtc) -{ - struct drm_device *dev = crtc->dev; - struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_timer_engine *ptimer = &dev_priv->engine.timer; - struct nv50_display *disp = nv50_display(dev); - struct nouveau_channel *evo = disp->master; - u64 start; - int ret; - - ret = RING_SPACE(evo, 6); - if (ret) - return ret; - BEGIN_RING(evo, 0, 0x0084, 1); - OUT_RING (evo, 0x80000000); - BEGIN_RING(evo, 0, 0x0080, 1); - OUT_RING (evo, 0); - BEGIN_RING(evo, 0, 0x0084, 1); - OUT_RING (evo, 0x00000000); - - nv_wo32(disp->ntfy, 0x000, 0x00000000); - FIRE_RING (evo); - - start = ptimer->read(dev); - do { - if (nv_ro32(disp->ntfy, 0x000)) - return 0; - } while (ptimer->read(dev) - start < 2000000000ULL); - - return -EBUSY; -} - static void nv50_crtc_prepare(struct drm_crtc *crtc) { @@ -497,7 +530,7 @@ nv50_crtc_commit(struct drm_crtc *crtc) nv50_crtc_blank(nv_crtc, false); drm_vblank_post_modeset(dev, nv_crtc->index); - nv50_crtc_wait_complete(crtc); + nv50_display_sync(dev); nv50_display_flip_next(crtc, crtc->fb, NULL); } @@ -579,8 +612,6 @@ nv50_crtc_do_mode_set_base(struct drm_crtc *crtc, OUT_RING (evo, fb->base.depth == 8 ? NV50_EVO_CRTC_CLUT_MODE_OFF : NV50_EVO_CRTC_CLUT_MODE_ON); - BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, COLOR_CTRL), 1); - OUT_RING (evo, NV50_EVO_CRTC_COLOR_CTRL_COLOR); BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, FB_POS), 1); OUT_RING (evo, (y << 16) | x); @@ -593,90 +624,77 @@ nv50_crtc_do_mode_set_base(struct drm_crtc *crtc, } static int -nv50_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode, int x, int y, +nv50_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *umode, + struct drm_display_mode *mode, int x, int y, struct drm_framebuffer *old_fb) { struct drm_device *dev = crtc->dev; struct nouveau_channel *evo = nv50_display(dev)->master; struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc); - struct nouveau_connector *nv_connector = NULL; - uint32_t hsync_dur, vsync_dur, hsync_start_to_end, vsync_start_to_end; - uint32_t hunk1, vunk1, vunk2a, vunk2b; + u32 head = nv_crtc->index * 0x400; + u32 ilace = (mode->flags & DRM_MODE_FLAG_INTERLACE) ? 2 : 1; + u32 vscan = (mode->flags & DRM_MODE_FLAG_DBLSCAN) ? 2 : 1; + u32 hactive, hsynce, hbackp, hfrontp, hblanke, hblanks; + u32 vactive, vsynce, vbackp, vfrontp, vblanke, vblanks; + u32 vblan2e = 0, vblan2s = 1; int ret; - /* Find the connector attached to this CRTC */ - nv_connector = nouveau_crtc_connector_get(nv_crtc); - - *nv_crtc->mode = *adjusted_mode; - - NV_DEBUG_KMS(dev, "index %d\n", nv_crtc->index); + /* hw timing description looks like this: + * + * <---------display---------> + * ______ + * |____________|---------------------------|____________| + * + * ^ synce ^ blanke ^ blanks ^ active + * + * interlaced modes also have 2 additional values pointing at the end + * and start of the next field's blanking period. + */ - hsync_dur = adjusted_mode->hsync_end - adjusted_mode->hsync_start; - vsync_dur = adjusted_mode->vsync_end - adjusted_mode->vsync_start; - hsync_start_to_end = adjusted_mode->htotal - adjusted_mode->hsync_start; - vsync_start_to_end = adjusted_mode->vtotal - adjusted_mode->vsync_start; - /* I can't give this a proper name, anyone else can? */ - hunk1 = adjusted_mode->htotal - - adjusted_mode->hsync_start + adjusted_mode->hdisplay; - vunk1 = adjusted_mode->vtotal - - adjusted_mode->vsync_start + adjusted_mode->vdisplay; - /* Another strange value, this time only for interlaced adjusted_modes. */ - vunk2a = 2 * adjusted_mode->vtotal - - adjusted_mode->vsync_start + adjusted_mode->vdisplay; - vunk2b = adjusted_mode->vtotal - - adjusted_mode->vsync_start + adjusted_mode->vtotal; - - if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) { - vsync_dur /= 2; - vsync_start_to_end /= 2; - vunk1 /= 2; - vunk2a /= 2; - vunk2b /= 2; - /* magic */ - if (adjusted_mode->flags & DRM_MODE_FLAG_DBLSCAN) { - vsync_start_to_end -= 1; - vunk1 -= 1; - vunk2a -= 1; - vunk2b -= 1; - } + hactive = mode->htotal; + hsynce = mode->hsync_end - mode->hsync_start - 1; + hbackp = mode->htotal - mode->hsync_end; + hblanke = hsynce + hbackp; + hfrontp = mode->hsync_start - mode->hdisplay; + hblanks = mode->htotal - hfrontp - 1; + + vactive = mode->vtotal * vscan / ilace; + vsynce = ((mode->vsync_end - mode->vsync_start) * vscan / ilace) - 1; + vbackp = (mode->vtotal - mode->vsync_end) * vscan / ilace; + vblanke = vsynce + vbackp; + vfrontp = (mode->vsync_start - mode->vdisplay) * vscan / ilace; + vblanks = vactive - vfrontp - 1; + if (mode->flags & DRM_MODE_FLAG_INTERLACE) { + vblan2e = vactive + vsynce + vbackp; + vblan2s = vblan2e + (mode->vdisplay * vscan / ilace); + vactive = (vactive * 2) + 1; } - ret = RING_SPACE(evo, 17); - if (ret) - return ret; - - BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, CLOCK), 2); - OUT_RING(evo, adjusted_mode->clock | 0x800000); - OUT_RING(evo, (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) ? 2 : 0); - - BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, DISPLAY_START), 5); - OUT_RING(evo, 0); - OUT_RING(evo, (adjusted_mode->vtotal << 16) | adjusted_mode->htotal); - OUT_RING(evo, (vsync_dur - 1) << 16 | (hsync_dur - 1)); - OUT_RING(evo, (vsync_start_to_end - 1) << 16 | - (hsync_start_to_end - 1)); - OUT_RING(evo, (vunk1 - 1) << 16 | (hunk1 - 1)); - - if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) { - BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, UNK0824), 1); - OUT_RING(evo, (vunk2b - 1) << 16 | (vunk2a - 1)); - } else { - OUT_RING(evo, 0); - OUT_RING(evo, 0); + ret = RING_SPACE(evo, 18); + if (ret == 0) { + BEGIN_RING(evo, 0, 0x0804 + head, 2); + OUT_RING (evo, 0x00800000 | mode->clock); + OUT_RING (evo, (ilace == 2) ? 2 : 0); + BEGIN_RING(evo, 0, 0x0810 + head, 6); + OUT_RING (evo, 0x00000000); /* border colour */ + OUT_RING (evo, (vactive << 16) | hactive); + OUT_RING (evo, ( vsynce << 16) | hsynce); + OUT_RING (evo, (vblanke << 16) | hblanke); + OUT_RING (evo, (vblanks << 16) | hblanks); + OUT_RING (evo, (vblan2e << 16) | vblan2s); + BEGIN_RING(evo, 0, 0x082c + head, 1); + OUT_RING (evo, 0x00000000); + BEGIN_RING(evo, 0, 0x0900 + head, 1); + OUT_RING (evo, 0x00000311); /* makes sync channel work */ + BEGIN_RING(evo, 0, 0x08c8 + head, 1); + OUT_RING (evo, (umode->vdisplay << 16) | umode->hdisplay); + BEGIN_RING(evo, 0, 0x08d4 + head, 1); + OUT_RING (evo, 0x00000000); /* screen position */ } - BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, UNK082C), 1); - OUT_RING(evo, 0); - - /* This is the actual resolution of the mode. */ - BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, REAL_RES), 1); - OUT_RING(evo, (mode->vdisplay << 16) | mode->hdisplay); - BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, SCALE_CENTER_OFFSET), 1); - OUT_RING(evo, NV50_EVO_CRTC_SCALE_CENTER_OFFSET_VAL(0, 0)); - - nv_crtc->set_dither(nv_crtc, nv_connector->use_dithering, false); - nv_crtc->set_scale(nv_crtc, nv_connector->scaling_mode, false); + nv_crtc->set_dither(nv_crtc, false); + nv_crtc->set_scale(nv_crtc, false); + nv_crtc->set_color_vibrance(nv_crtc, false); return nv50_crtc_do_mode_set_base(crtc, old_fb, x, y, false); } @@ -692,7 +710,7 @@ nv50_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y, if (ret) return ret; - ret = nv50_crtc_wait_complete(crtc); + ret = nv50_display_sync(crtc->dev); if (ret) return ret; @@ -711,7 +729,7 @@ nv50_crtc_mode_set_base_atomic(struct drm_crtc *crtc, if (ret) return ret; - return nv50_crtc_wait_complete(crtc); + return nv50_display_sync(crtc->dev); } static const struct drm_crtc_helper_funcs nv50_crtc_helper_funcs = { @@ -737,11 +755,8 @@ nv50_crtc_create(struct drm_device *dev, int index) if (!nv_crtc) return -ENOMEM; - nv_crtc->mode = kzalloc(sizeof(*nv_crtc->mode), GFP_KERNEL); - if (!nv_crtc->mode) { - kfree(nv_crtc); - return -ENOMEM; - } + nv_crtc->color_vibrance = 50; + nv_crtc->vibrant_hue = 0; /* Default CLUT parameters, will be activated on the hw upon * first mode set. @@ -764,7 +779,6 @@ nv50_crtc_create(struct drm_device *dev, int index) } if (ret) { - kfree(nv_crtc->mode); kfree(nv_crtc); return ret; } @@ -774,6 +788,7 @@ nv50_crtc_create(struct drm_device *dev, int index) /* set function pointers */ nv_crtc->set_dither = nv50_crtc_set_dither; nv_crtc->set_scale = nv50_crtc_set_scale; + nv_crtc->set_color_vibrance = nv50_crtc_set_color_vibrance; drm_crtc_init(dev, &nv_crtc->base, &nv50_crtc_funcs); drm_crtc_helper_add(&nv_crtc->base, &nv50_crtc_helper_funcs); diff --git a/drivers/gpu/drm/nouveau/nv50_dac.c b/drivers/gpu/drm/nouveau/nv50_dac.c index 808f3ec..55c5633 100644 --- a/drivers/gpu/drm/nouveau/nv50_dac.c +++ b/drivers/gpu/drm/nouveau/nv50_dac.c @@ -190,21 +190,13 @@ nv50_dac_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode, } if (connector->scaling_mode != DRM_MODE_SCALE_NONE && - connector->native_mode) { - int id = adjusted_mode->base.id; - *adjusted_mode = *connector->native_mode; - adjusted_mode->base.id = id; - } + connector->native_mode) + drm_mode_copy(adjusted_mode, connector->native_mode); return true; } static void -nv50_dac_prepare(struct drm_encoder *encoder) -{ -} - -static void nv50_dac_commit(struct drm_encoder *encoder) { } @@ -266,7 +258,7 @@ static const struct drm_encoder_helper_funcs nv50_dac_helper_funcs = { .save = nv50_dac_save, .restore = nv50_dac_restore, .mode_fixup = nv50_dac_mode_fixup, - .prepare = nv50_dac_prepare, + .prepare = nv50_dac_disconnect, .commit = nv50_dac_commit, .mode_set = nv50_dac_mode_set, .get_crtc = nv50_dac_crtc_get, diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c index 06de250..8b78b9c 100644 --- a/drivers/gpu/drm/nouveau/nv50_display.c +++ b/drivers/gpu/drm/nouveau/nv50_display.c @@ -50,9 +50,76 @@ nv50_sor_nr(struct drm_device *dev) return 4; } +u32 +nv50_display_active_crtcs(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + u32 mask = 0; + int i; + + if (dev_priv->chipset < 0x90 || + dev_priv->chipset == 0x92 || + dev_priv->chipset == 0xa0) { + for (i = 0; i < 2; i++) + mask |= nv_rd32(dev, NV50_PDISPLAY_SOR_MODE_CTRL_C(i)); + } else { + for (i = 0; i < 4; i++) + mask |= nv_rd32(dev, NV90_PDISPLAY_SOR_MODE_CTRL_C(i)); + } + + for (i = 0; i < 3; i++) + mask |= nv_rd32(dev, NV50_PDISPLAY_DAC_MODE_CTRL_C(i)); + + return mask & 3; +} + +static int +evo_icmd(struct drm_device *dev, int ch, u32 mthd, u32 data) +{ + int ret = 0; + nv_mask(dev, 0x610300 + (ch * 0x08), 0x00000001, 0x00000001); + nv_wr32(dev, 0x610304 + (ch * 0x08), data); + nv_wr32(dev, 0x610300 + (ch * 0x08), 0x80000001 | mthd); + if (!nv_wait(dev, 0x610300 + (ch * 0x08), 0x80000000, 0x00000000)) + ret = -EBUSY; + if (ret || (nouveau_reg_debug & NOUVEAU_REG_DEBUG_EVO)) + NV_INFO(dev, "EvoPIO: %d 0x%04x 0x%08x\n", ch, mthd, data); + nv_mask(dev, 0x610300 + (ch * 0x08), 0x00000001, 0x00000000); + return ret; +} + int nv50_display_early_init(struct drm_device *dev) { + u32 ctrl = nv_rd32(dev, 0x610200); + int i; + + /* check if master evo channel is already active, a good a sign as any + * that the display engine is in a weird state (hibernate/kexec), if + * it is, do our best to reset the display engine... + */ + if ((ctrl & 0x00000003) == 0x00000003) { + NV_INFO(dev, "PDISP: EVO(0) 0x%08x, resetting...\n", ctrl); + + /* deactivate both heads first, PDISP will disappear forever + * (well, until you power cycle) on some boards as soon as + * PMC_ENABLE is hit unless they are.. + */ + for (i = 0; i < 2; i++) { + evo_icmd(dev, 0, 0x0880 + (i * 0x400), 0x05000000); + evo_icmd(dev, 0, 0x089c + (i * 0x400), 0); + evo_icmd(dev, 0, 0x0840 + (i * 0x400), 0); + evo_icmd(dev, 0, 0x0844 + (i * 0x400), 0); + evo_icmd(dev, 0, 0x085c + (i * 0x400), 0); + evo_icmd(dev, 0, 0x0874 + (i * 0x400), 0); + } + evo_icmd(dev, 0, 0x0080, 0); + + /* reset PDISP */ + nv_mask(dev, 0x000200, 0x40000000, 0x00000000); + nv_mask(dev, 0x000200, 0x40000000, 0x40000000); + } + return 0; } @@ -62,11 +129,40 @@ nv50_display_late_takedown(struct drm_device *dev) } int -nv50_display_init(struct drm_device *dev) +nv50_display_sync(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; - struct drm_connector *connector; + struct nouveau_timer_engine *ptimer = &dev_priv->engine.timer; + struct nv50_display *disp = nv50_display(dev); + struct nouveau_channel *evo = disp->master; + u64 start; + int ret; + + ret = RING_SPACE(evo, 6); + if (ret == 0) { + BEGIN_RING(evo, 0, 0x0084, 1); + OUT_RING (evo, 0x80000000); + BEGIN_RING(evo, 0, 0x0080, 1); + OUT_RING (evo, 0); + BEGIN_RING(evo, 0, 0x0084, 1); + OUT_RING (evo, 0x00000000); + + nv_wo32(disp->ntfy, 0x000, 0x00000000); + FIRE_RING (evo); + + start = ptimer->read(dev); + do { + if (nv_ro32(disp->ntfy, 0x000)) + return 0; + } while (ptimer->read(dev) - start < 2000000000ULL); + } + + return -EBUSY; +} + +int +nv50_display_init(struct drm_device *dev) +{ struct nouveau_channel *evo; int ret, i; u32 val; @@ -161,16 +257,6 @@ nv50_display_init(struct drm_device *dev) NV50_PDISPLAY_INTR_EN_1_CLK_UNK20 | NV50_PDISPLAY_INTR_EN_1_CLK_UNK40); - /* enable hotplug interrupts */ - list_for_each_entry(connector, &dev->mode_config.connector_list, head) { - struct nouveau_connector *conn = nouveau_connector(connector); - - if (conn->dcb->gpio_tag == 0xff) - continue; - - pgpio->irq_enable(dev, conn->dcb->gpio_tag, true); - } - ret = nv50_evo_init(dev); if (ret) return ret; @@ -178,36 +264,19 @@ nv50_display_init(struct drm_device *dev) nv_wr32(dev, NV50_PDISPLAY_OBJECTS, (evo->ramin->vinst >> 8) | 9); - ret = RING_SPACE(evo, 15); + ret = RING_SPACE(evo, 3); if (ret) return ret; BEGIN_RING(evo, 0, NV50_EVO_UNK84, 2); - OUT_RING(evo, NV50_EVO_UNK84_NOTIFY_DISABLED); - OUT_RING(evo, NvEvoSync); - BEGIN_RING(evo, 0, NV50_EVO_CRTC(0, FB_DMA), 1); - OUT_RING(evo, NV50_EVO_CRTC_FB_DMA_HANDLE_NONE); - BEGIN_RING(evo, 0, NV50_EVO_CRTC(0, UNK0800), 1); - OUT_RING(evo, 0); - BEGIN_RING(evo, 0, NV50_EVO_CRTC(0, DISPLAY_START), 1); - OUT_RING(evo, 0); - BEGIN_RING(evo, 0, NV50_EVO_CRTC(0, UNK082C), 1); - OUT_RING(evo, 0); - /* required to make display sync channels not hate life */ - BEGIN_RING(evo, 0, NV50_EVO_CRTC(0, UNK900), 1); - OUT_RING (evo, 0x00000311); - BEGIN_RING(evo, 0, NV50_EVO_CRTC(1, UNK900), 1); - OUT_RING (evo, 0x00000311); - FIRE_RING(evo); - if (!nv_wait(dev, 0x640004, 0xffffffff, evo->dma.put << 2)) - NV_ERROR(dev, "evo pushbuf stalled\n"); - + OUT_RING (evo, NV50_EVO_UNK84_NOTIFY_DISABLED); + OUT_RING (evo, NvEvoSync); - return 0; + return nv50_display_sync(dev); } -static int nv50_display_disable(struct drm_device *dev) +void +nv50_display_fini(struct drm_device *dev) { - struct drm_nouveau_private *dev_priv = dev->dev_private; struct nv50_display *disp = nv50_display(dev); struct nouveau_channel *evo = disp->master; struct drm_crtc *drm_crtc; @@ -270,18 +339,10 @@ static int nv50_display_disable(struct drm_device *dev) /* disable interrupts. */ nv_wr32(dev, NV50_PDISPLAY_INTR_EN_1, 0x00000000); - - /* disable hotplug interrupts */ - nv_wr32(dev, 0xe054, 0xffffffff); - nv_wr32(dev, 0xe050, 0x00000000); - if (dev_priv->chipset >= 0x90) { - nv_wr32(dev, 0xe074, 0xffffffff); - nv_wr32(dev, 0xe070, 0x00000000); - } - return 0; } -int nv50_display_create(struct drm_device *dev) +int +nv50_display_create(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; struct dcb_table *dcb = &dev_priv->vbios.dcb; @@ -341,7 +402,7 @@ int nv50_display_create(struct drm_device *dev) tasklet_init(&priv->tasklet, nv50_display_bh, (unsigned long)dev); nouveau_irq_register(dev, 26, nv50_display_isr); - ret = nv50_display_init(dev); + ret = nv50_evo_create(dev); if (ret) { nv50_display_destroy(dev); return ret; @@ -357,7 +418,7 @@ nv50_display_destroy(struct drm_device *dev) NV_DEBUG_KMS(dev, "\n"); - nv50_display_disable(dev); + nv50_evo_destroy(dev); nouveau_irq_unregister(dev, 26); kfree(disp); } @@ -413,15 +474,15 @@ nv50_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb, } if (dev_priv->chipset < 0xc0) { - BEGIN_RING(chan, NvSubSw, 0x0060, 2); + BEGIN_RING(chan, 0, 0x0060, 2); OUT_RING (chan, NvEvoSema0 + nv_crtc->index); OUT_RING (chan, dispc->sem.offset); - BEGIN_RING(chan, NvSubSw, 0x006c, 1); + BEGIN_RING(chan, 0, 0x006c, 1); OUT_RING (chan, 0xf00d0000 | dispc->sem.value); - BEGIN_RING(chan, NvSubSw, 0x0064, 2); + BEGIN_RING(chan, 0, 0x0064, 2); OUT_RING (chan, dispc->sem.offset ^ 0x10); OUT_RING (chan, 0x74b1e000); - BEGIN_RING(chan, NvSubSw, 0x0060, 1); + BEGIN_RING(chan, 0, 0x0060, 1); if (dev_priv->chipset < 0x84) OUT_RING (chan, NvSema); else @@ -429,12 +490,12 @@ nv50_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb, } else { u64 offset = chan->dispc_vma[nv_crtc->index].offset; offset += dispc->sem.offset; - BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0010, 4); + BEGIN_NVC0(chan, 2, 0, 0x0010, 4); OUT_RING (chan, upper_32_bits(offset)); OUT_RING (chan, lower_32_bits(offset)); OUT_RING (chan, 0xf00d0000 | dispc->sem.value); OUT_RING (chan, 0x1002); - BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0010, 4); + BEGIN_NVC0(chan, 2, 0, 0x0010, 4); OUT_RING (chan, upper_32_bits(offset)); OUT_RING (chan, lower_32_bits(offset ^ 0x10)); OUT_RING (chan, 0x74b1e000); @@ -521,7 +582,7 @@ nv50_display_script_select(struct drm_device *dev, struct dcb_entry *dcb, } else { /* determine number of lvds links */ if (nv_connector && nv_connector->edid && - nv_connector->dcb->type == DCB_CONNECTOR_LVDS_SPWG) { + nv_connector->type == DCB_CONNECTOR_LVDS_SPWG) { /* http://www.spwg.org */ if (((u8 *)nv_connector->edid)[121] == 2) script |= 0x0100; @@ -722,8 +783,8 @@ nv50_display_unk20_handler(struct drm_device *dev) if (crtc >= 0) { pclk = nv_rd32(dev, NV50_PDISPLAY_CRTC_P(crtc, CLOCK)); pclk &= 0x003fffff; - - nv50_crtc_set_clock(dev, crtc, pclk); + if (pclk) + nv50_crtc_set_clock(dev, crtc, pclk); tmp = nv_rd32(dev, NV50_PDISPLAY_CRTC_CLK_CTRL2(crtc)); tmp &= ~0x000000f; @@ -802,9 +863,9 @@ nv50_display_unk20_handler(struct drm_device *dev) if (type == OUTPUT_DP) { int link = !(dcb->dpconf.sor.link & 1); if ((mc & 0x000f0000) == 0x00020000) - nouveau_dp_tu_update(dev, or, link, pclk, 18); + nv50_sor_dp_calc_tu(dev, or, link, pclk, 18); else - nouveau_dp_tu_update(dev, or, link, pclk, 24); + nv50_sor_dp_calc_tu(dev, or, link, pclk, 24); } if (dcb->type != OUTPUT_ANALOG) { diff --git a/drivers/gpu/drm/nouveau/nv50_display.h b/drivers/gpu/drm/nouveau/nv50_display.h index c2da503..5d3dd14 100644 --- a/drivers/gpu/drm/nouveau/nv50_display.h +++ b/drivers/gpu/drm/nouveau/nv50_display.h @@ -69,14 +69,20 @@ int nv50_display_early_init(struct drm_device *dev); void nv50_display_late_takedown(struct drm_device *dev); int nv50_display_create(struct drm_device *dev); int nv50_display_init(struct drm_device *dev); +void nv50_display_fini(struct drm_device *dev); void nv50_display_destroy(struct drm_device *dev); int nv50_crtc_blank(struct nouveau_crtc *, bool blank); int nv50_crtc_set_clock(struct drm_device *, int head, int pclk); +u32 nv50_display_active_crtcs(struct drm_device *); + +int nv50_display_sync(struct drm_device *); int nv50_display_flip_next(struct drm_crtc *, struct drm_framebuffer *, struct nouveau_channel *chan); void nv50_display_flip_stop(struct drm_crtc *); +int nv50_evo_create(struct drm_device *dev); +void nv50_evo_destroy(struct drm_device *dev); int nv50_evo_init(struct drm_device *dev); void nv50_evo_fini(struct drm_device *dev); void nv50_evo_dmaobj_init(struct nouveau_gpuobj *, u32 memtype, u64 base, diff --git a/drivers/gpu/drm/nouveau/nv50_evo.c b/drivers/gpu/drm/nouveau/nv50_evo.c index c99d975..9b962e9 100644 --- a/drivers/gpu/drm/nouveau/nv50_evo.c +++ b/drivers/gpu/drm/nouveau/nv50_evo.c @@ -218,7 +218,7 @@ nv50_evo_channel_fini(struct nouveau_channel *evo) } } -static void +void nv50_evo_destroy(struct drm_device *dev) { struct nv50_display *disp = nv50_display(dev); @@ -235,7 +235,7 @@ nv50_evo_destroy(struct drm_device *dev) nv50_evo_channel_del(&disp->master); } -static int +int nv50_evo_create(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; @@ -388,12 +388,6 @@ nv50_evo_init(struct drm_device *dev) struct nv50_display *disp = nv50_display(dev); int ret, i; - if (!disp->master) { - ret = nv50_evo_create(dev); - if (ret) - return ret; - } - ret = nv50_evo_channel_init(disp->master); if (ret) return ret; @@ -420,6 +414,4 @@ nv50_evo_fini(struct drm_device *dev) if (disp->master) nv50_evo_channel_fini(disp->master); - - nv50_evo_destroy(dev); } diff --git a/drivers/gpu/drm/nouveau/nv50_evo.h b/drivers/gpu/drm/nouveau/nv50_evo.h index 3860ca6..771d879 100644 --- a/drivers/gpu/drm/nouveau/nv50_evo.h +++ b/drivers/gpu/drm/nouveau/nv50_evo.h @@ -104,7 +104,8 @@ #define NV50_EVO_CRTC_SCALE_CTRL_INACTIVE 0x00000000 #define NV50_EVO_CRTC_SCALE_CTRL_ACTIVE 0x00000009 #define NV50_EVO_CRTC_COLOR_CTRL 0x000008a8 -#define NV50_EVO_CRTC_COLOR_CTRL_COLOR 0x00040000 +#define NV50_EVO_CRTC_COLOR_CTRL_VIBRANCE 0x000fff00 +#define NV50_EVO_CRTC_COLOR_CTRL_HUE 0xfff00000 #define NV50_EVO_CRTC_FB_POS 0x000008c0 #define NV50_EVO_CRTC_REAL_RES 0x000008c8 #define NV50_EVO_CRTC_SCALE_CENTER_OFFSET 0x000008d4 diff --git a/drivers/gpu/drm/nouveau/nv50_fifo.c b/drivers/gpu/drm/nouveau/nv50_fifo.c index c34a074..3bc2a56 100644 --- a/drivers/gpu/drm/nouveau/nv50_fifo.c +++ b/drivers/gpu/drm/nouveau/nv50_fifo.c @@ -230,6 +230,7 @@ nv50_fifo_create_context(struct nouveau_channel *chan) struct drm_device *dev = chan->dev; struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_gpuobj *ramfc = NULL; + uint64_t ib_offset = chan->pushbuf_base + chan->dma.ib_base * 4; unsigned long flags; int ret; @@ -280,8 +281,9 @@ nv50_fifo_create_context(struct nouveau_channel *chan) nv_wo32(ramfc, 0x7c, 0x30000001); nv_wo32(ramfc, 0x78, 0x00000000); nv_wo32(ramfc, 0x3c, 0x403f6078); - nv_wo32(ramfc, 0x50, chan->pushbuf_base + chan->dma.ib_base * 4); - nv_wo32(ramfc, 0x54, drm_order(chan->dma.ib_max + 1) << 16); + nv_wo32(ramfc, 0x50, lower_32_bits(ib_offset)); + nv_wo32(ramfc, 0x54, upper_32_bits(ib_offset) | + drm_order(chan->dma.ib_max + 1) << 16); if (dev_priv->chipset != 0x50) { nv_wo32(chan->ramin, 0, chan->id); diff --git a/drivers/gpu/drm/nouveau/nv50_gpio.c b/drivers/gpu/drm/nouveau/nv50_gpio.c index 793a5cc..f429e6a 100644 --- a/drivers/gpu/drm/nouveau/nv50_gpio.c +++ b/drivers/gpu/drm/nouveau/nv50_gpio.c @@ -25,229 +25,95 @@ #include "drmP.h" #include "nouveau_drv.h" #include "nouveau_hw.h" +#include "nouveau_gpio.h" #include "nv50_display.h" -static void nv50_gpio_isr(struct drm_device *dev); -static void nv50_gpio_isr_bh(struct work_struct *work); - -struct nv50_gpio_priv { - struct list_head handlers; - spinlock_t lock; -}; - -struct nv50_gpio_handler { - struct drm_device *dev; - struct list_head head; - struct work_struct work; - bool inhibit; - - struct dcb_gpio_entry *gpio; - - void (*handler)(void *data, int state); - void *data; -}; - static int -nv50_gpio_location(struct dcb_gpio_entry *gpio, uint32_t *reg, uint32_t *shift) +nv50_gpio_location(int line, u32 *reg, u32 *shift) { const uint32_t nv50_gpio_reg[4] = { 0xe104, 0xe108, 0xe280, 0xe284 }; - if (gpio->line >= 32) + if (line >= 32) return -EINVAL; - *reg = nv50_gpio_reg[gpio->line >> 3]; - *shift = (gpio->line & 7) << 2; + *reg = nv50_gpio_reg[line >> 3]; + *shift = (line & 7) << 2; return 0; } int -nv50_gpio_get(struct drm_device *dev, enum dcb_gpio_tag tag) +nv50_gpio_drive(struct drm_device *dev, int line, int dir, int out) { - struct dcb_gpio_entry *gpio; - uint32_t r, s, v; - - gpio = nouveau_bios_gpio_entry(dev, tag); - if (!gpio) - return -ENOENT; + u32 reg, shift; - if (nv50_gpio_location(gpio, &r, &s)) + if (nv50_gpio_location(line, ®, &shift)) return -EINVAL; - v = nv_rd32(dev, r) >> (s + 2); - return ((v & 1) == (gpio->state[1] & 1)); + nv_mask(dev, reg, 7 << shift, (((dir ^ 1) << 1) | out) << shift); + return 0; } int -nv50_gpio_set(struct drm_device *dev, enum dcb_gpio_tag tag, int state) +nv50_gpio_sense(struct drm_device *dev, int line) { - struct dcb_gpio_entry *gpio; - uint32_t r, s, v; - - gpio = nouveau_bios_gpio_entry(dev, tag); - if (!gpio) - return -ENOENT; + u32 reg, shift; - if (nv50_gpio_location(gpio, &r, &s)) + if (nv50_gpio_location(line, ®, &shift)) return -EINVAL; - v = nv_rd32(dev, r) & ~(0x3 << s); - v |= (gpio->state[state] ^ 2) << s; - nv_wr32(dev, r, v); - return 0; + return !!(nv_rd32(dev, reg) & (4 << shift)); } -int -nvd0_gpio_get(struct drm_device *dev, enum dcb_gpio_tag tag) +void +nv50_gpio_irq_enable(struct drm_device *dev, int line, bool on) { - struct dcb_gpio_entry *gpio; - u32 v; - - gpio = nouveau_bios_gpio_entry(dev, tag); - if (!gpio) - return -ENOENT; + u32 reg = line < 16 ? 0xe050 : 0xe070; + u32 mask = 0x00010001 << (line & 0xf); - v = nv_rd32(dev, 0x00d610 + (gpio->line * 4)); - v &= 0x00004000; - return (!!v == (gpio->state[1] & 1)); + nv_wr32(dev, reg + 4, mask); + nv_mask(dev, reg + 0, mask, on ? mask : 0); } int -nvd0_gpio_set(struct drm_device *dev, enum dcb_gpio_tag tag, int state) +nvd0_gpio_drive(struct drm_device *dev, int line, int dir, int out) { - struct dcb_gpio_entry *gpio; - u32 v; - - gpio = nouveau_bios_gpio_entry(dev, tag); - if (!gpio) - return -ENOENT; - - v = gpio->state[state] ^ 2; - - nv_mask(dev, 0x00d610 + (gpio->line * 4), 0x00003000, v << 12); + u32 data = ((dir ^ 1) << 13) | (out << 12); + nv_mask(dev, 0x00d610 + (line * 4), 0x00003000, data); + nv_mask(dev, 0x00d604, 0x00000001, 0x00000001); /* update? */ return 0; } int -nv50_gpio_irq_register(struct drm_device *dev, enum dcb_gpio_tag tag, - void (*handler)(void *, int), void *data) +nvd0_gpio_sense(struct drm_device *dev, int line) { - struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; - struct nv50_gpio_priv *priv = pgpio->priv; - struct nv50_gpio_handler *gpioh; - struct dcb_gpio_entry *gpio; - unsigned long flags; - - gpio = nouveau_bios_gpio_entry(dev, tag); - if (!gpio) - return -ENOENT; - - gpioh = kzalloc(sizeof(*gpioh), GFP_KERNEL); - if (!gpioh) - return -ENOMEM; - - INIT_WORK(&gpioh->work, nv50_gpio_isr_bh); - gpioh->dev = dev; - gpioh->gpio = gpio; - gpioh->handler = handler; - gpioh->data = data; - - spin_lock_irqsave(&priv->lock, flags); - list_add(&gpioh->head, &priv->handlers); - spin_unlock_irqrestore(&priv->lock, flags); - return 0; + return !!(nv_rd32(dev, 0x00d610 + (line * 4)) & 0x00004000); } -void -nv50_gpio_irq_unregister(struct drm_device *dev, enum dcb_gpio_tag tag, - void (*handler)(void *, int), void *data) -{ - struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; - struct nv50_gpio_priv *priv = pgpio->priv; - struct nv50_gpio_handler *gpioh, *tmp; - struct dcb_gpio_entry *gpio; - LIST_HEAD(tofree); - unsigned long flags; - - gpio = nouveau_bios_gpio_entry(dev, tag); - if (!gpio) - return; - - spin_lock_irqsave(&priv->lock, flags); - list_for_each_entry_safe(gpioh, tmp, &priv->handlers, head) { - if (gpioh->gpio != gpio || - gpioh->handler != handler || - gpioh->data != data) - continue; - list_move(&gpioh->head, &tofree); - } - spin_unlock_irqrestore(&priv->lock, flags); - - list_for_each_entry_safe(gpioh, tmp, &tofree, head) { - flush_work_sync(&gpioh->work); - kfree(gpioh); - } -} - -bool -nv50_gpio_irq_enable(struct drm_device *dev, enum dcb_gpio_tag tag, bool on) -{ - struct dcb_gpio_entry *gpio; - u32 reg, mask; - - gpio = nouveau_bios_gpio_entry(dev, tag); - if (!gpio) - return false; - - reg = gpio->line < 16 ? 0xe050 : 0xe070; - mask = 0x00010001 << (gpio->line & 0xf); - - nv_wr32(dev, reg + 4, mask); - reg = nv_mask(dev, reg + 0, mask, on ? mask : 0); - return (reg & mask) == mask; -} - -static int -nv50_gpio_create(struct drm_device *dev) +static void +nv50_gpio_isr(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; - struct nv50_gpio_priv *priv; - - priv = kzalloc(sizeof(*priv), GFP_KERNEL); - if (!priv) - return -ENOMEM; + u32 intr0, intr1 = 0; + u32 hi, lo; - INIT_LIST_HEAD(&priv->handlers); - spin_lock_init(&priv->lock); - pgpio->priv = priv; - return 0; -} + intr0 = nv_rd32(dev, 0xe054) & nv_rd32(dev, 0xe050); + if (dev_priv->chipset >= 0x90) + intr1 = nv_rd32(dev, 0xe074) & nv_rd32(dev, 0xe070); -static void -nv50_gpio_destroy(struct drm_device *dev) -{ - struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; + hi = (intr0 & 0x0000ffff) | (intr1 << 16); + lo = (intr0 >> 16) | (intr1 & 0xffff0000); + nouveau_gpio_isr(dev, 0, hi | lo); - kfree(pgpio->priv); - pgpio->priv = NULL; + nv_wr32(dev, 0xe054, intr0); + if (dev_priv->chipset >= 0x90) + nv_wr32(dev, 0xe074, intr1); } int nv50_gpio_init(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; - int ret; - - if (!pgpio->priv) { - ret = nv50_gpio_create(dev); - if (ret) - return ret; - } /* disable, and ack any pending gpio interrupts */ nv_wr32(dev, 0xe050, 0x00000000); @@ -270,64 +136,4 @@ nv50_gpio_fini(struct drm_device *dev) if (dev_priv->chipset >= 0x90) nv_wr32(dev, 0xe070, 0x00000000); nouveau_irq_unregister(dev, 21); - - nv50_gpio_destroy(dev); -} - -static void -nv50_gpio_isr_bh(struct work_struct *work) -{ - struct nv50_gpio_handler *gpioh = - container_of(work, struct nv50_gpio_handler, work); - struct drm_nouveau_private *dev_priv = gpioh->dev->dev_private; - struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; - struct nv50_gpio_priv *priv = pgpio->priv; - unsigned long flags; - int state; - - state = pgpio->get(gpioh->dev, gpioh->gpio->tag); - if (state < 0) - return; - - gpioh->handler(gpioh->data, state); - - spin_lock_irqsave(&priv->lock, flags); - gpioh->inhibit = false; - spin_unlock_irqrestore(&priv->lock, flags); -} - -static void -nv50_gpio_isr(struct drm_device *dev) -{ - struct drm_nouveau_private *dev_priv = dev->dev_private; - struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio; - struct nv50_gpio_priv *priv = pgpio->priv; - struct nv50_gpio_handler *gpioh; - u32 intr0, intr1 = 0; - u32 hi, lo, ch; - - intr0 = nv_rd32(dev, 0xe054) & nv_rd32(dev, 0xe050); - if (dev_priv->chipset >= 0x90) - intr1 = nv_rd32(dev, 0xe074) & nv_rd32(dev, 0xe070); - - hi = (intr0 & 0x0000ffff) | (intr1 << 16); - lo = (intr0 >> 16) | (intr1 & 0xffff0000); - ch = hi | lo; - - nv_wr32(dev, 0xe054, intr0); - if (dev_priv->chipset >= 0x90) - nv_wr32(dev, 0xe074, intr1); - - spin_lock(&priv->lock); - list_for_each_entry(gpioh, &priv->handlers, head) { - if (!(ch & (1 << gpioh->gpio->line))) - continue; - - if (gpioh->inhibit) - continue; - gpioh->inhibit = true; - - schedule_work(&gpioh->work); - } - spin_unlock(&priv->lock); } diff --git a/drivers/gpu/drm/nouveau/nv50_graph.c b/drivers/gpu/drm/nouveau/nv50_graph.c index ac601f7..33d5711 100644 --- a/drivers/gpu/drm/nouveau/nv50_graph.c +++ b/drivers/gpu/drm/nouveau/nv50_graph.c @@ -616,9 +616,9 @@ nv50_pgraph_tp_trap(struct drm_device *dev, int type, uint32_t ustatus_old, } break; case 7: /* MP error */ - if (ustatus & 0x00010000) { + if (ustatus & 0x04030000) { nv50_pgraph_mp_trap(dev, i, display); - ustatus &= ~0x00010000; + ustatus &= ~0x04030000; } break; case 8: /* TPDMA error */ diff --git a/drivers/gpu/drm/nouveau/nv50_pm.c b/drivers/gpu/drm/nouveau/nv50_pm.c index 3d5a86b..d020ed4 100644 --- a/drivers/gpu/drm/nouveau/nv50_pm.c +++ b/drivers/gpu/drm/nouveau/nv50_pm.c @@ -25,122 +25,862 @@ #include "drmP.h" #include "nouveau_drv.h" #include "nouveau_bios.h" +#include "nouveau_hw.h" #include "nouveau_pm.h" +#include "nouveau_hwsq.h" +#include "nv50_display.h" + +enum clk_src { + clk_src_crystal, + clk_src_href, + clk_src_hclk, + clk_src_hclkm3, + clk_src_hclkm3d2, + clk_src_host, + clk_src_nvclk, + clk_src_sclk, + clk_src_mclk, + clk_src_vdec, + clk_src_dom6 +}; + +static u32 read_clk(struct drm_device *, enum clk_src); + +static u32 +read_div(struct drm_device *dev) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + + switch (dev_priv->chipset) { + case 0x50: /* it exists, but only has bit 31, not the dividers.. */ + case 0x84: + case 0x86: + case 0x98: + case 0xa0: + return nv_rd32(dev, 0x004700); + case 0x92: + case 0x94: + case 0x96: + return nv_rd32(dev, 0x004800); + default: + return 0x00000000; + } +} + +static u32 +read_pll_src(struct drm_device *dev, u32 base) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + u32 coef, ref = read_clk(dev, clk_src_crystal); + u32 rsel = nv_rd32(dev, 0x00e18c); + int P, N, M, id; + + switch (dev_priv->chipset) { + case 0x50: + case 0xa0: + switch (base) { + case 0x4020: + case 0x4028: id = !!(rsel & 0x00000004); break; + case 0x4008: id = !!(rsel & 0x00000008); break; + case 0x4030: id = 0; break; + default: + NV_ERROR(dev, "ref: bad pll 0x%06x\n", base); + return 0; + } + + coef = nv_rd32(dev, 0x00e81c + (id * 0x0c)); + ref *= (coef & 0x01000000) ? 2 : 4; + P = (coef & 0x00070000) >> 16; + N = ((coef & 0x0000ff00) >> 8) + 1; + M = ((coef & 0x000000ff) >> 0) + 1; + break; + case 0x84: + case 0x86: + case 0x92: + coef = nv_rd32(dev, 0x00e81c); + P = (coef & 0x00070000) >> 16; + N = (coef & 0x0000ff00) >> 8; + M = (coef & 0x000000ff) >> 0; + break; + case 0x94: + case 0x96: + case 0x98: + rsel = nv_rd32(dev, 0x00c050); + switch (base) { + case 0x4020: rsel = (rsel & 0x00000003) >> 0; break; + case 0x4008: rsel = (rsel & 0x0000000c) >> 2; break; + case 0x4028: rsel = (rsel & 0x00001800) >> 11; break; + case 0x4030: rsel = 3; break; + default: + NV_ERROR(dev, "ref: bad pll 0x%06x\n", base); + return 0; + } + + switch (rsel) { + case 0: id = 1; break; + case 1: return read_clk(dev, clk_src_crystal); + case 2: return read_clk(dev, clk_src_href); + case 3: id = 0; break; + } + + coef = nv_rd32(dev, 0x00e81c + (id * 0x28)); + P = (nv_rd32(dev, 0x00e824 + (id * 0x28)) >> 16) & 7; + P += (coef & 0x00070000) >> 16; + N = (coef & 0x0000ff00) >> 8; + M = (coef & 0x000000ff) >> 0; + break; + default: + BUG_ON(1); + } + + if (M) + return (ref * N / M) >> P; + return 0; +} + +static u32 +read_pll_ref(struct drm_device *dev, u32 base) +{ + u32 src, mast = nv_rd32(dev, 0x00c040); + + switch (base) { + case 0x004028: + src = !!(mast & 0x00200000); + break; + case 0x004020: + src = !!(mast & 0x00400000); + break; + case 0x004008: + src = !!(mast & 0x00010000); + break; + case 0x004030: + src = !!(mast & 0x02000000); + break; + case 0x00e810: + return read_clk(dev, clk_src_crystal); + default: + NV_ERROR(dev, "bad pll 0x%06x\n", base); + return 0; + } + + if (src) + return read_clk(dev, clk_src_href); + return read_pll_src(dev, base); +} + +static u32 +read_pll(struct drm_device *dev, u32 base) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + u32 mast = nv_rd32(dev, 0x00c040); + u32 ctrl = nv_rd32(dev, base + 0); + u32 coef = nv_rd32(dev, base + 4); + u32 ref = read_pll_ref(dev, base); + u32 clk = 0; + int N1, N2, M1, M2; + + if (base == 0x004028 && (mast & 0x00100000)) { + /* wtf, appears to only disable post-divider on nva0 */ + if (dev_priv->chipset != 0xa0) + return read_clk(dev, clk_src_dom6); + } + + N2 = (coef & 0xff000000) >> 24; + M2 = (coef & 0x00ff0000) >> 16; + N1 = (coef & 0x0000ff00) >> 8; + M1 = (coef & 0x000000ff); + if ((ctrl & 0x80000000) && M1) { + clk = ref * N1 / M1; + if ((ctrl & 0x40000100) == 0x40000000) { + if (M2) + clk = clk * N2 / M2; + else + clk = 0; + } + } + + return clk; +} + +static u32 +read_clk(struct drm_device *dev, enum clk_src src) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + u32 mast = nv_rd32(dev, 0x00c040); + u32 P = 0; + + switch (src) { + case clk_src_crystal: + return dev_priv->crystal; + case clk_src_href: + return 100000; /* PCIE reference clock */ + case clk_src_hclk: + return read_clk(dev, clk_src_href) * 27778 / 10000; + case clk_src_hclkm3: + return read_clk(dev, clk_src_hclk) * 3; + case clk_src_hclkm3d2: + return read_clk(dev, clk_src_hclk) * 3 / 2; + case clk_src_host: + switch (mast & 0x30000000) { + case 0x00000000: return read_clk(dev, clk_src_href); + case 0x10000000: break; + case 0x20000000: /* !0x50 */ + case 0x30000000: return read_clk(dev, clk_src_hclk); + } + break; + case clk_src_nvclk: + if (!(mast & 0x00100000)) + P = (nv_rd32(dev, 0x004028) & 0x00070000) >> 16; + switch (mast & 0x00000003) { + case 0x00000000: return read_clk(dev, clk_src_crystal) >> P; + case 0x00000001: return read_clk(dev, clk_src_dom6); + case 0x00000002: return read_pll(dev, 0x004020) >> P; + case 0x00000003: return read_pll(dev, 0x004028) >> P; + } + break; + case clk_src_sclk: + P = (nv_rd32(dev, 0x004020) & 0x00070000) >> 16; + switch (mast & 0x00000030) { + case 0x00000000: + if (mast & 0x00000080) + return read_clk(dev, clk_src_host) >> P; + return read_clk(dev, clk_src_crystal) >> P; + case 0x00000010: break; + case 0x00000020: return read_pll(dev, 0x004028) >> P; + case 0x00000030: return read_pll(dev, 0x004020) >> P; + } + break; + case clk_src_mclk: + P = (nv_rd32(dev, 0x004008) & 0x00070000) >> 16; + if (nv_rd32(dev, 0x004008) & 0x00000200) { + switch (mast & 0x0000c000) { + case 0x00000000: + return read_clk(dev, clk_src_crystal) >> P; + case 0x00008000: + case 0x0000c000: + return read_clk(dev, clk_src_href) >> P; + } + } else { + return read_pll(dev, 0x004008) >> P; + } + break; + case clk_src_vdec: + P = (read_div(dev) & 0x00000700) >> 8; + switch (dev_priv->chipset) { + case 0x84: + case 0x86: + case 0x92: + case 0x94: + case 0x96: + case 0xa0: + switch (mast & 0x00000c00) { + case 0x00000000: + if (dev_priv->chipset == 0xa0) /* wtf?? */ + return read_clk(dev, clk_src_nvclk) >> P; + return read_clk(dev, clk_src_crystal) >> P; + case 0x00000400: + return 0; + case 0x00000800: + if (mast & 0x01000000) + return read_pll(dev, 0x004028) >> P; + return read_pll(dev, 0x004030) >> P; + case 0x00000c00: + return read_clk(dev, clk_src_nvclk) >> P; + } + break; + case 0x98: + switch (mast & 0x00000c00) { + case 0x00000000: + return read_clk(dev, clk_src_nvclk) >> P; + case 0x00000400: + return 0; + case 0x00000800: + return read_clk(dev, clk_src_hclkm3d2) >> P; + case 0x00000c00: + return read_clk(dev, clk_src_mclk) >> P; + } + break; + } + break; + case clk_src_dom6: + switch (dev_priv->chipset) { + case 0x50: + case 0xa0: + return read_pll(dev, 0x00e810) >> 2; + case 0x84: + case 0x86: + case 0x92: + case 0x94: + case 0x96: + case 0x98: + P = (read_div(dev) & 0x00000007) >> 0; + switch (mast & 0x0c000000) { + case 0x00000000: return read_clk(dev, clk_src_href); + case 0x04000000: break; + case 0x08000000: return read_clk(dev, clk_src_hclk); + case 0x0c000000: + return read_clk(dev, clk_src_hclkm3) >> P; + } + break; + default: + break; + } + default: + break; + } + + NV_DEBUG(dev, "unknown clock source %d 0x%08x\n", src, mast); + return 0; +} + +int +nv50_pm_clocks_get(struct drm_device *dev, struct nouveau_pm_level *perflvl) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + if (dev_priv->chipset == 0xaa || + dev_priv->chipset == 0xac) + return 0; + + perflvl->core = read_clk(dev, clk_src_nvclk); + perflvl->shader = read_clk(dev, clk_src_sclk); + perflvl->memory = read_clk(dev, clk_src_mclk); + if (dev_priv->chipset != 0x50) { + perflvl->vdec = read_clk(dev, clk_src_vdec); + perflvl->dom6 = read_clk(dev, clk_src_dom6); + } + + return 0; +} struct nv50_pm_state { struct nouveau_pm_level *perflvl; - struct pll_lims pll; - enum pll_types type; - int N, M, P; + struct hwsq_ucode eclk_hwsq; + struct hwsq_ucode mclk_hwsq; + u32 mscript; + u32 mmast; + u32 mctrl; + u32 mcoef; }; -int -nv50_pm_clock_get(struct drm_device *dev, u32 id) +static u32 +calc_pll(struct drm_device *dev, u32 reg, struct pll_lims *pll, + u32 clk, int *N1, int *M1, int *log2P) { - struct pll_lims pll; - int P, N, M, ret; - u32 reg0, reg1; + struct nouveau_pll_vals coef; + int ret; - ret = get_pll_limits(dev, id, &pll); + ret = get_pll_limits(dev, reg, pll); if (ret) - return ret; + return 0; + + pll->vco2.maxfreq = 0; + pll->refclk = read_pll_ref(dev, reg); + if (!pll->refclk) + return 0; + + ret = nouveau_calc_pll_mnp(dev, pll, clk, &coef); + if (ret == 0) + return 0; - reg0 = nv_rd32(dev, pll.reg + 0); - reg1 = nv_rd32(dev, pll.reg + 4); - - if ((reg0 & 0x80000000) == 0) { - if (id == PLL_SHADER) { - NV_DEBUG(dev, "Shader PLL is disabled. " - "Shader clock is twice the core\n"); - ret = nv50_pm_clock_get(dev, PLL_CORE); - if (ret > 0) - return ret << 1; - } else if (id == PLL_MEMORY) { - NV_DEBUG(dev, "Memory PLL is disabled. " - "Memory clock is equal to the ref_clk\n"); - return pll.refclk; + *N1 = coef.N1; + *M1 = coef.M1; + *log2P = coef.log2P; + return ret; +} + +static inline u32 +calc_div(u32 src, u32 target, int *div) +{ + u32 clk0 = src, clk1 = src; + for (*div = 0; *div <= 7; (*div)++) { + if (clk0 <= target) { + clk1 = clk0 << (*div ? 1 : 0); + break; } + clk0 >>= 1; + } + + if (target - clk0 <= clk1 - target) + return clk0; + (*div)--; + return clk1; +} + +static inline u32 +clk_same(u32 a, u32 b) +{ + return ((a / 1000) == (b / 1000)); +} + +static void +mclk_precharge(struct nouveau_mem_exec_func *exec) +{ + struct nv50_pm_state *info = exec->priv; + struct hwsq_ucode *hwsq = &info->mclk_hwsq; + + hwsq_wr32(hwsq, 0x1002d4, 0x00000001); +} + +static void +mclk_refresh(struct nouveau_mem_exec_func *exec) +{ + struct nv50_pm_state *info = exec->priv; + struct hwsq_ucode *hwsq = &info->mclk_hwsq; + + hwsq_wr32(hwsq, 0x1002d0, 0x00000001); +} + +static void +mclk_refresh_auto(struct nouveau_mem_exec_func *exec, bool enable) +{ + struct nv50_pm_state *info = exec->priv; + struct hwsq_ucode *hwsq = &info->mclk_hwsq; + + hwsq_wr32(hwsq, 0x100210, enable ? 0x80000000 : 0x00000000); +} + +static void +mclk_refresh_self(struct nouveau_mem_exec_func *exec, bool enable) +{ + struct nv50_pm_state *info = exec->priv; + struct hwsq_ucode *hwsq = &info->mclk_hwsq; + + hwsq_wr32(hwsq, 0x1002dc, enable ? 0x00000001 : 0x00000000); +} + +static void +mclk_wait(struct nouveau_mem_exec_func *exec, u32 nsec) +{ + struct nv50_pm_state *info = exec->priv; + struct hwsq_ucode *hwsq = &info->mclk_hwsq; + + if (nsec > 1000) + hwsq_usec(hwsq, (nsec + 500) / 1000); +} + +static u32 +mclk_mrg(struct nouveau_mem_exec_func *exec, int mr) +{ + if (mr <= 1) + return nv_rd32(exec->dev, 0x1002c0 + ((mr - 0) * 4)); + if (mr <= 3) + return nv_rd32(exec->dev, 0x1002e0 + ((mr - 2) * 4)); + return 0; +} + +static void +mclk_mrs(struct nouveau_mem_exec_func *exec, int mr, u32 data) +{ + struct drm_nouveau_private *dev_priv = exec->dev->dev_private; + struct nv50_pm_state *info = exec->priv; + struct hwsq_ucode *hwsq = &info->mclk_hwsq; + + if (mr <= 1) { + if (dev_priv->vram_rank_B) + hwsq_wr32(hwsq, 0x1002c8 + ((mr - 0) * 4), data); + hwsq_wr32(hwsq, 0x1002c0 + ((mr - 0) * 4), data); + } else + if (mr <= 3) { + if (dev_priv->vram_rank_B) + hwsq_wr32(hwsq, 0x1002e8 + ((mr - 2) * 4), data); + hwsq_wr32(hwsq, 0x1002e0 + ((mr - 2) * 4), data); + } +} + +static void +mclk_clock_set(struct nouveau_mem_exec_func *exec) +{ + struct nv50_pm_state *info = exec->priv; + struct hwsq_ucode *hwsq = &info->mclk_hwsq; + u32 ctrl = nv_rd32(exec->dev, 0x004008); + + info->mmast = nv_rd32(exec->dev, 0x00c040); + info->mmast &= ~0xc0000000; /* get MCLK_2 from HREF */ + info->mmast |= 0x0000c000; /* use MCLK_2 as MPLL_BYPASS clock */ + + hwsq_wr32(hwsq, 0xc040, info->mmast); + hwsq_wr32(hwsq, 0x4008, ctrl | 0x00000200); /* bypass MPLL */ + if (info->mctrl & 0x80000000) + hwsq_wr32(hwsq, 0x400c, info->mcoef); + hwsq_wr32(hwsq, 0x4008, info->mctrl); +} + +static void +mclk_timing_set(struct nouveau_mem_exec_func *exec) +{ + struct drm_device *dev = exec->dev; + struct nv50_pm_state *info = exec->priv; + struct nouveau_pm_level *perflvl = info->perflvl; + struct hwsq_ucode *hwsq = &info->mclk_hwsq; + int i; + + for (i = 0; i < 9; i++) { + u32 reg = 0x100220 + (i * 4); + u32 val = nv_rd32(dev, reg); + if (val != perflvl->timing.reg[i]) + hwsq_wr32(hwsq, reg, perflvl->timing.reg[i]); + } +} + +static int +calc_mclk(struct drm_device *dev, struct nouveau_pm_level *perflvl, + struct nv50_pm_state *info) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + u32 crtc_mask = nv50_display_active_crtcs(dev); + struct nouveau_mem_exec_func exec = { + .dev = dev, + .precharge = mclk_precharge, + .refresh = mclk_refresh, + .refresh_auto = mclk_refresh_auto, + .refresh_self = mclk_refresh_self, + .wait = mclk_wait, + .mrg = mclk_mrg, + .mrs = mclk_mrs, + .clock_set = mclk_clock_set, + .timing_set = mclk_timing_set, + .priv = info + }; + struct hwsq_ucode *hwsq = &info->mclk_hwsq; + struct pll_lims pll; + int N, M, P; + int ret; + + /* use pcie refclock if possible, otherwise use mpll */ + info->mctrl = nv_rd32(dev, 0x004008); + info->mctrl &= ~0x81ff0200; + if (clk_same(perflvl->memory, read_clk(dev, clk_src_href))) { + info->mctrl |= 0x00000200 | (pll.log2p_bias << 19); + } else { + ret = calc_pll(dev, 0x4008, &pll, perflvl->memory, &N, &M, &P); + if (ret == 0) + return -EINVAL; + + info->mctrl |= 0x80000000 | (P << 22) | (P << 16); + info->mctrl |= pll.log2p_bias << 19; + info->mcoef = (N << 8) | M; } - P = (reg0 & 0x00070000) >> 16; - N = (reg1 & 0x0000ff00) >> 8; - M = (reg1 & 0x000000ff); + /* build the ucode which will reclock the memory for us */ + hwsq_init(hwsq); + if (crtc_mask) { + hwsq_op5f(hwsq, crtc_mask, 0x00); /* wait for scanout */ + hwsq_op5f(hwsq, crtc_mask, 0x01); /* wait for vblank */ + } + if (dev_priv->chipset >= 0x92) + hwsq_wr32(hwsq, 0x611200, 0x00003300); /* disable scanout */ + hwsq_setf(hwsq, 0x10, 0); /* disable bus access */ + hwsq_op5f(hwsq, 0x00, 0x01); /* no idea :s */ - return ((pll.refclk * N / M) >> P); + ret = nouveau_mem_exec(&exec, perflvl); + if (ret) + return ret; + + hwsq_setf(hwsq, 0x10, 1); /* enable bus access */ + hwsq_op5f(hwsq, 0x00, 0x00); /* no idea, reverse of 0x00, 0x01? */ + if (dev_priv->chipset >= 0x92) + hwsq_wr32(hwsq, 0x611200, 0x00003330); /* enable scanout */ + hwsq_fini(hwsq); + return 0; } void * -nv50_pm_clock_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl, - u32 id, int khz) +nv50_pm_clocks_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl) { - struct nv50_pm_state *state; - int dummy, ret; + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nv50_pm_state *info; + struct hwsq_ucode *hwsq; + struct pll_lims pll; + u32 out, mast, divs, ctrl; + int clk, ret = -EINVAL; + int N, M, P1, P2; - state = kzalloc(sizeof(*state), GFP_KERNEL); - if (!state) + if (dev_priv->chipset == 0xaa || + dev_priv->chipset == 0xac) + return ERR_PTR(-ENODEV); + + info = kmalloc(sizeof(*info), GFP_KERNEL); + if (!info) return ERR_PTR(-ENOMEM); - state->type = id; - state->perflvl = perflvl; + info->perflvl = perflvl; + + /* memory: build hwsq ucode which we'll use to reclock memory. + * use pcie refclock if possible, otherwise use mpll */ + info->mclk_hwsq.len = 0; + if (perflvl->memory) { + ret = calc_mclk(dev, perflvl, info); + if (ret) + goto error; + info->mscript = perflvl->memscript; + } + + divs = read_div(dev); + mast = info->mmast; + + /* start building HWSQ script for engine reclocking */ + hwsq = &info->eclk_hwsq; + hwsq_init(hwsq); + hwsq_setf(hwsq, 0x10, 0); /* disable bus access */ + hwsq_op5f(hwsq, 0x00, 0x01); /* wait for access disabled? */ + + /* vdec/dom6: switch to "safe" clocks temporarily */ + if (perflvl->vdec) { + mast &= ~0x00000c00; + divs &= ~0x00000700; + } + + if (perflvl->dom6) { + mast &= ~0x0c000000; + divs &= ~0x00000007; + } + + hwsq_wr32(hwsq, 0x00c040, mast); + + /* vdec: avoid modifying xpll until we know exactly how the other + * clock domains work, i suspect at least some of them can also be + * tied to xpll... + */ + if (perflvl->vdec) { + /* see how close we can get using nvclk as a source */ + clk = calc_div(perflvl->core, perflvl->vdec, &P1); + + /* see how close we can get using xpll/hclk as a source */ + if (dev_priv->chipset != 0x98) + out = read_pll(dev, 0x004030); + else + out = read_clk(dev, clk_src_hclkm3d2); + out = calc_div(out, perflvl->vdec, &P2); + + /* select whichever gets us closest */ + if (abs((int)perflvl->vdec - clk) <= + abs((int)perflvl->vdec - out)) { + if (dev_priv->chipset != 0x98) + mast |= 0x00000c00; + divs |= P1 << 8; + } else { + mast |= 0x00000800; + divs |= P2 << 8; + } + } + + /* dom6: nfi what this is, but we're limited to various combinations + * of the host clock frequency + */ + if (perflvl->dom6) { + if (clk_same(perflvl->dom6, read_clk(dev, clk_src_href))) { + mast |= 0x00000000; + } else + if (clk_same(perflvl->dom6, read_clk(dev, clk_src_hclk))) { + mast |= 0x08000000; + } else { + clk = read_clk(dev, clk_src_hclk) * 3; + clk = calc_div(clk, perflvl->dom6, &P1); + + mast |= 0x0c000000; + divs |= P1; + } + } + + /* vdec/dom6: complete switch to new clocks */ + switch (dev_priv->chipset) { + case 0x92: + case 0x94: + case 0x96: + hwsq_wr32(hwsq, 0x004800, divs); + break; + default: + hwsq_wr32(hwsq, 0x004700, divs); + break; + } + + hwsq_wr32(hwsq, 0x00c040, mast); + + /* core/shader: make sure sclk/nvclk are disconnected from their + * PLLs (nvclk to dom6, sclk to hclk) + */ + if (dev_priv->chipset < 0x92) + mast = (mast & ~0x001000b0) | 0x00100080; + else + mast = (mast & ~0x000000b3) | 0x00000081; + + hwsq_wr32(hwsq, 0x00c040, mast); + + /* core: for the moment at least, always use nvpll */ + clk = calc_pll(dev, 0x4028, &pll, perflvl->core, &N, &M, &P1); + if (clk == 0) + goto error; + + ctrl = nv_rd32(dev, 0x004028) & ~0xc03f0100; + mast &= ~0x00100000; + mast |= 3; + + hwsq_wr32(hwsq, 0x004028, 0x80000000 | (P1 << 19) | (P1 << 16) | ctrl); + hwsq_wr32(hwsq, 0x00402c, (N << 8) | M); + + /* shader: tie to nvclk if possible, otherwise use spll. have to be + * very careful that the shader clock is at least twice the core, or + * some chipsets will be very unhappy. i expect most or all of these + * cases will be handled by tying to nvclk, but it's possible there's + * corners + */ + ctrl = nv_rd32(dev, 0x004020) & ~0xc03f0100; + + if (P1-- && perflvl->shader == (perflvl->core << 1)) { + hwsq_wr32(hwsq, 0x004020, (P1 << 19) | (P1 << 16) | ctrl); + hwsq_wr32(hwsq, 0x00c040, 0x00000020 | mast); + } else { + clk = calc_pll(dev, 0x4020, &pll, perflvl->shader, &N, &M, &P1); + if (clk == 0) + goto error; + ctrl |= 0x80000000; + + hwsq_wr32(hwsq, 0x004020, (P1 << 19) | (P1 << 16) | ctrl); + hwsq_wr32(hwsq, 0x004024, (N << 8) | M); + hwsq_wr32(hwsq, 0x00c040, 0x00000030 | mast); + } + + hwsq_setf(hwsq, 0x10, 1); /* enable bus access */ + hwsq_op5f(hwsq, 0x00, 0x00); /* wait for access enabled? */ + hwsq_fini(hwsq); + + return info; +error: + kfree(info); + return ERR_PTR(ret); +} + +static int +prog_hwsq(struct drm_device *dev, struct hwsq_ucode *hwsq) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + u32 hwsq_data, hwsq_kick; + int i; - ret = get_pll_limits(dev, id, &state->pll); - if (ret < 0) { - kfree(state); - return (ret == -ENOENT) ? NULL : ERR_PTR(ret); + if (dev_priv->chipset < 0x94) { + hwsq_data = 0x001400; + hwsq_kick = 0x00000003; + } else { + hwsq_data = 0x080000; + hwsq_kick = 0x00000001; } + /* upload hwsq ucode */ + nv_mask(dev, 0x001098, 0x00000008, 0x00000000); + nv_wr32(dev, 0x001304, 0x00000000); + if (dev_priv->chipset >= 0x92) + nv_wr32(dev, 0x001318, 0x00000000); + for (i = 0; i < hwsq->len / 4; i++) + nv_wr32(dev, hwsq_data + (i * 4), hwsq->ptr.u32[i]); + nv_mask(dev, 0x001098, 0x00000018, 0x00000018); + + /* launch, and wait for completion */ + nv_wr32(dev, 0x00130c, hwsq_kick); + if (!nv_wait(dev, 0x001308, 0x00000100, 0x00000000)) { + NV_ERROR(dev, "hwsq ucode exec timed out\n"); + NV_ERROR(dev, "0x001308: 0x%08x\n", nv_rd32(dev, 0x001308)); + for (i = 0; i < hwsq->len / 4; i++) { + NV_ERROR(dev, "0x%06x: 0x%08x\n", 0x1400 + (i * 4), + nv_rd32(dev, 0x001400 + (i * 4))); + } - ret = nv50_calc_pll(dev, &state->pll, khz, &state->N, &state->M, - &dummy, &dummy, &state->P); - if (ret < 0) { - kfree(state); - return ERR_PTR(ret); + return -EIO; } - return state; + return 0; } -void -nv50_pm_clock_set(struct drm_device *dev, void *pre_state) +int +nv50_pm_clocks_set(struct drm_device *dev, void *data) { - struct nv50_pm_state *state = pre_state; - struct nouveau_pm_level *perflvl = state->perflvl; - u32 reg = state->pll.reg, tmp; - struct bit_entry BIT_M; - u16 script; - int N = state->N; - int M = state->M; - int P = state->P; + struct nv50_pm_state *info = data; + struct bit_entry M; + int ret = -EBUSY; + + /* halt and idle execution engines */ + nv_mask(dev, 0x002504, 0x00000001, 0x00000001); + if (!nv_wait(dev, 0x002504, 0x00000010, 0x00000010)) + goto resume; + if (!nv_wait(dev, 0x00251c, 0x0000003f, 0x0000003f)) + goto resume; - if (state->type == PLL_MEMORY && perflvl->memscript && - bit_table(dev, 'M', &BIT_M) == 0 && - BIT_M.version == 1 && BIT_M.length >= 0x0b) { - script = ROM16(BIT_M.data[0x05]); - if (script) - nouveau_bios_run_init_table(dev, script, NULL, -1); - script = ROM16(BIT_M.data[0x07]); - if (script) - nouveau_bios_run_init_table(dev, script, NULL, -1); - script = ROM16(BIT_M.data[0x09]); - if (script) - nouveau_bios_run_init_table(dev, script, NULL, -1); + /* program memory clock, if necessary - must come before engine clock + * reprogramming due to how we construct the hwsq scripts in pre() + */ + if (info->mclk_hwsq.len) { + /* execute some scripts that do ??? from the vbios.. */ + if (!bit_table(dev, 'M', &M) && M.version == 1) { + if (M.length >= 6) + nouveau_bios_init_exec(dev, ROM16(M.data[5])); + if (M.length >= 8) + nouveau_bios_init_exec(dev, ROM16(M.data[7])); + if (M.length >= 10) + nouveau_bios_init_exec(dev, ROM16(M.data[9])); + nouveau_bios_init_exec(dev, info->mscript); + } - nouveau_bios_run_init_table(dev, perflvl->memscript, NULL, -1); + ret = prog_hwsq(dev, &info->mclk_hwsq); + if (ret) + goto resume; } - if (state->type == PLL_MEMORY) { - nv_wr32(dev, 0x100210, 0); - nv_wr32(dev, 0x1002dc, 1); + /* program engine clocks */ + ret = prog_hwsq(dev, &info->eclk_hwsq); + +resume: + nv_mask(dev, 0x002504, 0x00000001, 0x00000000); + kfree(info); + return ret; +} + +static int +pwm_info(struct drm_device *dev, int *line, int *ctrl, int *indx) +{ + if (*line == 0x04) { + *ctrl = 0x00e100; + *line = 4; + *indx = 0; + } else + if (*line == 0x09) { + *ctrl = 0x00e100; + *line = 9; + *indx = 1; + } else + if (*line == 0x10) { + *ctrl = 0x00e28c; + *line = 0; + *indx = 0; + } else { + NV_ERROR(dev, "unknown pwm ctrl for gpio %d\n", *line); + return -ENODEV; } - tmp = nv_rd32(dev, reg + 0) & 0xfff8ffff; - tmp |= 0x80000000 | (P << 16); - nv_wr32(dev, reg + 0, tmp); - nv_wr32(dev, reg + 4, (N << 8) | M); + return 0; +} + +int +nv50_pm_pwm_get(struct drm_device *dev, int line, u32 *divs, u32 *duty) +{ + int ctrl, id, ret = pwm_info(dev, &line, &ctrl, &id); + if (ret) + return ret; - if (state->type == PLL_MEMORY) { - nv_wr32(dev, 0x1002dc, 0); - nv_wr32(dev, 0x100210, 0x80000000); + if (nv_rd32(dev, ctrl) & (1 << line)) { + *divs = nv_rd32(dev, 0x00e114 + (id * 8)); + *duty = nv_rd32(dev, 0x00e118 + (id * 8)); + return 0; } - kfree(state); + return -EINVAL; } +int +nv50_pm_pwm_set(struct drm_device *dev, int line, u32 divs, u32 duty) +{ + int ctrl, id, ret = pwm_info(dev, &line, &ctrl, &id); + if (ret) + return ret; + + nv_mask(dev, ctrl, 0x00010001 << line, 0x00000001 << line); + nv_wr32(dev, 0x00e114 + (id * 8), divs); + nv_wr32(dev, 0x00e118 + (id * 8), duty | 0x80000000); + return 0; +} diff --git a/drivers/gpu/drm/nouveau/nv50_sor.c b/drivers/gpu/drm/nouveau/nv50_sor.c index 2633aa8..2746402 100644 --- a/drivers/gpu/drm/nouveau/nv50_sor.c +++ b/drivers/gpu/drm/nouveau/nv50_sor.c @@ -36,6 +36,193 @@ #include "nouveau_crtc.h" #include "nv50_display.h" +static u32 +nv50_sor_dp_lane_map(struct drm_device *dev, struct dcb_entry *dcb, u8 lane) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + static const u8 nvaf[] = { 24, 16, 8, 0 }; /* thanks, apple.. */ + static const u8 nv50[] = { 16, 8, 0, 24 }; + if (dev_priv->chipset == 0xaf) + return nvaf[lane]; + return nv50[lane]; +} + +static void +nv50_sor_dp_train_set(struct drm_device *dev, struct dcb_entry *dcb, u8 pattern) +{ + u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1); + nv_mask(dev, NV50_SOR_DP_CTRL(or, link), 0x0f000000, pattern << 24); +} + +static void +nv50_sor_dp_train_adj(struct drm_device *dev, struct dcb_entry *dcb, + u8 lane, u8 swing, u8 preem) +{ + u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1); + u32 shift = nv50_sor_dp_lane_map(dev, dcb, lane); + u32 mask = 0x000000ff << shift; + u8 *table, *entry, *config; + + table = nouveau_dp_bios_data(dev, dcb, &entry); + if (!table || (table[0] != 0x20 && table[0] != 0x21)) { + NV_ERROR(dev, "PDISP: unsupported DP table for chipset\n"); + return; + } + + config = entry + table[4]; + while (config[0] != swing || config[1] != preem) { + config += table[5]; + if (config >= entry + table[4] + entry[4] * table[5]) + return; + } + + nv_mask(dev, NV50_SOR_DP_UNK118(or, link), mask, config[2] << shift); + nv_mask(dev, NV50_SOR_DP_UNK120(or, link), mask, config[3] << shift); + nv_mask(dev, NV50_SOR_DP_UNK130(or, link), 0x0000ff00, config[4] << 8); +} + +static void +nv50_sor_dp_link_set(struct drm_device *dev, struct dcb_entry *dcb, int crtc, + int link_nr, u32 link_bw, bool enhframe) +{ + u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1); + u32 dpctrl = nv_rd32(dev, NV50_SOR_DP_CTRL(or, link)) & ~0x001f4000; + u32 clksor = nv_rd32(dev, 0x614300 + (or * 0x800)) & ~0x000c0000; + u8 *table, *entry, mask; + int i; + + table = nouveau_dp_bios_data(dev, dcb, &entry); + if (!table || (table[0] != 0x20 && table[0] != 0x21)) { + NV_ERROR(dev, "PDISP: unsupported DP table for chipset\n"); + return; + } + + entry = ROMPTR(dev, entry[10]); + if (entry) { + while (link_bw < ROM16(entry[0]) * 10) + entry += 4; + + nouveau_bios_run_init_table(dev, ROM16(entry[2]), dcb, crtc); + } + + dpctrl |= ((1 << link_nr) - 1) << 16; + if (enhframe) + dpctrl |= 0x00004000; + + if (link_bw > 162000) + clksor |= 0x00040000; + + nv_wr32(dev, 0x614300 + (or * 0x800), clksor); + nv_wr32(dev, NV50_SOR_DP_CTRL(or, link), dpctrl); + + mask = 0; + for (i = 0; i < link_nr; i++) + mask |= 1 << (nv50_sor_dp_lane_map(dev, dcb, i) >> 3); + nv_mask(dev, NV50_SOR_DP_UNK130(or, link), 0x0000000f, mask); +} + +static void +nv50_sor_dp_link_get(struct drm_device *dev, u32 or, u32 link, u32 *nr, u32 *bw) +{ + u32 dpctrl = nv_rd32(dev, NV50_SOR_DP_CTRL(or, link)) & 0x000f0000; + u32 clksor = nv_rd32(dev, 0x614300 + (or * 0x800)); + if (clksor & 0x000c0000) + *bw = 270000; + else + *bw = 162000; + + if (dpctrl > 0x00030000) *nr = 4; + else if (dpctrl > 0x00010000) *nr = 2; + else *nr = 1; +} + +void +nv50_sor_dp_calc_tu(struct drm_device *dev, int or, int link, u32 clk, u32 bpp) +{ + const u32 symbol = 100000; + int bestTU = 0, bestVTUi = 0, bestVTUf = 0, bestVTUa = 0; + int TU, VTUi, VTUf, VTUa; + u64 link_data_rate, link_ratio, unk; + u32 best_diff = 64 * symbol; + u32 link_nr, link_bw, r; + + /* calculate packed data rate for each lane */ + nv50_sor_dp_link_get(dev, or, link, &link_nr, &link_bw); + link_data_rate = (clk * bpp / 8) / link_nr; + + /* calculate ratio of packed data rate to link symbol rate */ + link_ratio = link_data_rate * symbol; + r = do_div(link_ratio, link_bw); + + for (TU = 64; TU >= 32; TU--) { + /* calculate average number of valid symbols in each TU */ + u32 tu_valid = link_ratio * TU; + u32 calc, diff; + + /* find a hw representation for the fraction.. */ + VTUi = tu_valid / symbol; + calc = VTUi * symbol; + diff = tu_valid - calc; + if (diff) { + if (diff >= (symbol / 2)) { + VTUf = symbol / (symbol - diff); + if (symbol - (VTUf * diff)) + VTUf++; + + if (VTUf <= 15) { + VTUa = 1; + calc += symbol - (symbol / VTUf); + } else { + VTUa = 0; + VTUf = 1; + calc += symbol; + } + } else { + VTUa = 0; + VTUf = min((int)(symbol / diff), 15); + calc += symbol / VTUf; + } + + diff = calc - tu_valid; + } else { + /* no remainder, but the hw doesn't like the fractional + * part to be zero. decrement the integer part and + * have the fraction add a whole symbol back + */ + VTUa = 0; + VTUf = 1; + VTUi--; + } + + if (diff < best_diff) { + best_diff = diff; + bestTU = TU; + bestVTUa = VTUa; + bestVTUf = VTUf; + bestVTUi = VTUi; + if (diff == 0) + break; + } + } + + if (!bestTU) { + NV_ERROR(dev, "DP: unable to find suitable config\n"); + return; + } + + /* XXX close to vbios numbers, but not right */ + unk = (symbol - link_ratio) * bestTU; + unk *= link_ratio; + r = do_div(unk, symbol); + r = do_div(unk, symbol); + unk += 6; + + nv_mask(dev, NV50_SOR_DP_CTRL(or, link), 0x000001fc, bestTU << 2); + nv_mask(dev, NV50_SOR_DP_SCFG(or, link), 0x010f7f3f, bestVTUa << 24 | + bestVTUf << 16 | + bestVTUi << 8 | + unk); +} static void nv50_sor_disconnect(struct drm_encoder *encoder) { @@ -60,6 +247,8 @@ nv50_sor_disconnect(struct drm_encoder *encoder) BEGIN_RING(evo, 0, NV50_EVO_UPDATE, 1); OUT_RING (evo, 0); + nouveau_hdmi_mode_set(encoder, NULL); + nv_encoder->crtc = NULL; nv_encoder->last_dpms = DRM_MODE_DPMS_OFF; } @@ -115,20 +304,13 @@ nv50_sor_dpms(struct drm_encoder *encoder, int mode) } if (nv_encoder->dcb->type == OUTPUT_DP) { - struct nouveau_i2c_chan *auxch; - - auxch = nouveau_i2c_find(dev, nv_encoder->dcb->i2c_index); - if (!auxch) - return; + struct dp_train_func func = { + .link_set = nv50_sor_dp_link_set, + .train_set = nv50_sor_dp_train_set, + .train_adj = nv50_sor_dp_train_adj + }; - if (mode == DRM_MODE_DPMS_ON) { - u8 status = DP_SET_POWER_D0; - nouveau_dp_auxch(auxch, 8, DP_SET_POWER, &status, 1); - nouveau_dp_link_train(encoder, nv_encoder->dp.datarate); - } else { - u8 status = DP_SET_POWER_D3; - nouveau_dp_auxch(auxch, 8, DP_SET_POWER, &status, 1); - } + nouveau_dp_dpms(encoder, mode, nv_encoder->dp.datarate, &func); } } @@ -160,11 +342,8 @@ nv50_sor_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode, } if (connector->scaling_mode != DRM_MODE_SCALE_NONE && - connector->native_mode) { - int id = adjusted_mode->base.id; - *adjusted_mode = *connector->native_mode; - adjusted_mode->base.id = id; - } + connector->native_mode) + drm_mode_copy(adjusted_mode, connector->native_mode); return true; } @@ -172,6 +351,12 @@ nv50_sor_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode, static void nv50_sor_prepare(struct drm_encoder *encoder) { + struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); + nv50_sor_disconnect(encoder); + if (nv_encoder->dcb->type == OUTPUT_DP) { + /* avoid race between link training and supervisor intr */ + nv50_display_sync(encoder->dev); + } } static void @@ -180,8 +365,8 @@ nv50_sor_commit(struct drm_encoder *encoder) } static void -nv50_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) +nv50_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *umode, + struct drm_display_mode *mode) { struct nouveau_channel *evo = nv50_display(encoder->dev)->master; struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); @@ -193,24 +378,27 @@ nv50_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, NV_DEBUG_KMS(dev, "or %d type %d -> crtc %d\n", nv_encoder->or, nv_encoder->dcb->type, crtc->index); + nv_encoder->crtc = encoder->crtc; switch (nv_encoder->dcb->type) { case OUTPUT_TMDS: if (nv_encoder->dcb->sorconf.link & 1) { - if (adjusted_mode->clock < 165000) + if (mode->clock < 165000) mode_ctl = 0x0100; else mode_ctl = 0x0500; } else mode_ctl = 0x0200; + + nouveau_hdmi_mode_set(encoder, mode); break; case OUTPUT_DP: nv_connector = nouveau_encoder_connector_get(nv_encoder); if (nv_connector && nv_connector->base.display_info.bpc == 6) { - nv_encoder->dp.datarate = crtc->mode->clock * 18 / 8; + nv_encoder->dp.datarate = mode->clock * 18 / 8; mode_ctl |= 0x00020000; } else { - nv_encoder->dp.datarate = crtc->mode->clock * 24 / 8; + nv_encoder->dp.datarate = mode->clock * 24 / 8; mode_ctl |= 0x00050000; } @@ -228,10 +416,10 @@ nv50_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, else mode_ctl |= NV50_EVO_SOR_MODE_CTRL_CRTC0; - if (adjusted_mode->flags & DRM_MODE_FLAG_NHSYNC) + if (mode->flags & DRM_MODE_FLAG_NHSYNC) mode_ctl |= NV50_EVO_SOR_MODE_CTRL_NHSYNC; - if (adjusted_mode->flags & DRM_MODE_FLAG_NVSYNC) + if (mode->flags & DRM_MODE_FLAG_NVSYNC) mode_ctl |= NV50_EVO_SOR_MODE_CTRL_NVSYNC; nv50_sor_dpms(encoder, DRM_MODE_DPMS_ON); @@ -239,12 +427,11 @@ nv50_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, ret = RING_SPACE(evo, 2); if (ret) { NV_ERROR(dev, "no space while connecting SOR\n"); + nv_encoder->crtc = NULL; return; } BEGIN_RING(evo, 0, NV50_EVO_SOR(nv_encoder->or, MODE_CTRL), 1); OUT_RING(evo, mode_ctl); - - nv_encoder->crtc = encoder->crtc; } static struct drm_crtc * diff --git a/drivers/gpu/drm/nouveau/nv50_vm.c b/drivers/gpu/drm/nouveau/nv50_vm.c index 40b84f2..44fbac9 100644 --- a/drivers/gpu/drm/nouveau/nv50_vm.c +++ b/drivers/gpu/drm/nouveau/nv50_vm.c @@ -48,7 +48,7 @@ nv50_vm_map_pgt(struct nouveau_gpuobj *pgd, u32 pde, phys |= 0x60; else if (coverage <= 64 * 1024 * 1024) phys |= 0x40; - else if (coverage < 128 * 1024 * 1024) + else if (coverage <= 128 * 1024 * 1024) phys |= 0x20; } @@ -57,27 +57,15 @@ nv50_vm_map_pgt(struct nouveau_gpuobj *pgd, u32 pde, } static inline u64 -nv50_vm_addr(struct nouveau_vma *vma, u64 phys, u32 memtype, u32 target) +vm_addr(struct nouveau_vma *vma, u64 phys, u32 memtype, u32 target) { - struct drm_nouveau_private *dev_priv = vma->vm->dev->dev_private; - phys |= 1; /* present */ phys |= (u64)memtype << 40; - - /* IGPs don't have real VRAM, re-target to stolen system memory */ - if (target == 0 && dev_priv->vram_sys_base) { - phys += dev_priv->vram_sys_base; - target = 3; - } - phys |= target << 4; - if (vma->access & NV_MEM_ACCESS_SYS) phys |= (1 << 6); - if (!(vma->access & NV_MEM_ACCESS_WO)) phys |= (1 << 3); - return phys; } @@ -85,11 +73,19 @@ void nv50_vm_map(struct nouveau_vma *vma, struct nouveau_gpuobj *pgt, struct nouveau_mem *mem, u32 pte, u32 cnt, u64 phys, u64 delta) { + struct drm_nouveau_private *dev_priv = vma->vm->dev->dev_private; u32 comp = (mem->memtype & 0x180) >> 7; - u32 block; + u32 block, target; int i; - phys = nv50_vm_addr(vma, phys, mem->memtype, 0); + /* IGPs don't have real VRAM, re-target to stolen system memory */ + target = 0; + if (dev_priv->vram_sys_base) { + phys += dev_priv->vram_sys_base; + target = 3; + } + + phys = vm_addr(vma, phys, mem->memtype, target); pte <<= 3; cnt <<= 3; @@ -125,9 +121,10 @@ void nv50_vm_map_sg(struct nouveau_vma *vma, struct nouveau_gpuobj *pgt, struct nouveau_mem *mem, u32 pte, u32 cnt, dma_addr_t *list) { + u32 target = (vma->access & NV_MEM_ACCESS_NOSNOOP) ? 3 : 2; pte <<= 3; while (cnt--) { - u64 phys = nv50_vm_addr(vma, (u64)*list++, mem->memtype, 2); + u64 phys = vm_addr(vma, (u64)*list++, mem->memtype, target); nv_wo32(pgt, pte + 0, lower_32_bits(phys)); nv_wo32(pgt, pte + 4, upper_32_bits(phys)); pte += 8; diff --git a/drivers/gpu/drm/nouveau/nv50_vram.c b/drivers/gpu/drm/nouveau/nv50_vram.c index 2e45e57..9ed9ae39 100644 --- a/drivers/gpu/drm/nouveau/nv50_vram.c +++ b/drivers/gpu/drm/nouveau/nv50_vram.c @@ -189,8 +189,25 @@ nv50_vram_init(struct drm_device *dev) struct nouveau_vram_engine *vram = &dev_priv->engine.vram; const u32 rsvd_head = ( 256 * 1024) >> 12; /* vga memory */ const u32 rsvd_tail = (1024 * 1024) >> 12; /* vbios etc */ + u32 pfb714 = nv_rd32(dev, 0x100714); u32 rblock, length; + switch (pfb714 & 0x00000007) { + case 0: dev_priv->vram_type = NV_MEM_TYPE_DDR1; break; + case 1: + if (nouveau_mem_vbios_type(dev) == NV_MEM_TYPE_DDR3) + dev_priv->vram_type = NV_MEM_TYPE_DDR3; + else + dev_priv->vram_type = NV_MEM_TYPE_DDR2; + break; + case 2: dev_priv->vram_type = NV_MEM_TYPE_GDDR3; break; + case 3: dev_priv->vram_type = NV_MEM_TYPE_GDDR4; break; + case 4: dev_priv->vram_type = NV_MEM_TYPE_GDDR5; break; + default: + break; + } + + dev_priv->vram_rank_B = !!(nv_rd32(dev, 0x100200) & 0x4); dev_priv->vram_size = nv_rd32(dev, 0x10020c); dev_priv->vram_size |= (dev_priv->vram_size & 0xff) << 32; dev_priv->vram_size &= 0xffffffff00ULL; diff --git a/drivers/gpu/drm/nouveau/nv84_bsp.c b/drivers/gpu/drm/nouveau/nv84_bsp.c new file mode 100644 index 0000000..7487573 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nv84_bsp.c @@ -0,0 +1,83 @@ +/* + * Copyright 2011 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs + */ + +#include "drmP.h" +#include "nouveau_drv.h" +#include "nouveau_util.h" +#include "nouveau_vm.h" +#include "nouveau_ramht.h" + +/*XXX: This stub is currently used on NV98+ also, as soon as this becomes + * more than just an enable/disable stub this needs to be split out to + * nv98_bsp.c... + */ + +struct nv84_bsp_engine { + struct nouveau_exec_engine base; +}; + +static int +nv84_bsp_fini(struct drm_device *dev, int engine, bool suspend) +{ + if (!(nv_rd32(dev, 0x000200) & 0x00008000)) + return 0; + + nv_mask(dev, 0x000200, 0x00008000, 0x00000000); + return 0; +} + +static int +nv84_bsp_init(struct drm_device *dev, int engine) +{ + nv_mask(dev, 0x000200, 0x00008000, 0x00000000); + nv_mask(dev, 0x000200, 0x00008000, 0x00008000); + return 0; +} + +static void +nv84_bsp_destroy(struct drm_device *dev, int engine) +{ + struct nv84_bsp_engine *pbsp = nv_engine(dev, engine); + + NVOBJ_ENGINE_DEL(dev, BSP); + + kfree(pbsp); +} + +int +nv84_bsp_create(struct drm_device *dev) +{ + struct nv84_bsp_engine *pbsp; + + pbsp = kzalloc(sizeof(*pbsp), GFP_KERNEL); + if (!pbsp) + return -ENOMEM; + + pbsp->base.destroy = nv84_bsp_destroy; + pbsp->base.init = nv84_bsp_init; + pbsp->base.fini = nv84_bsp_fini; + + NVOBJ_ENGINE_ADD(dev, BSP, &pbsp->base); + return 0; +} diff --git a/drivers/gpu/drm/nouveau/nv84_vp.c b/drivers/gpu/drm/nouveau/nv84_vp.c new file mode 100644 index 0000000..6570d30 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nv84_vp.c @@ -0,0 +1,83 @@ +/* + * Copyright 2011 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs + */ + +#include "drmP.h" +#include "nouveau_drv.h" +#include "nouveau_util.h" +#include "nouveau_vm.h" +#include "nouveau_ramht.h" + +/*XXX: This stub is currently used on NV98+ also, as soon as this becomes + * more than just an enable/disable stub this needs to be split out to + * nv98_vp.c... + */ + +struct nv84_vp_engine { + struct nouveau_exec_engine base; +}; + +static int +nv84_vp_fini(struct drm_device *dev, int engine, bool suspend) +{ + if (!(nv_rd32(dev, 0x000200) & 0x00020000)) + return 0; + + nv_mask(dev, 0x000200, 0x00020000, 0x00000000); + return 0; +} + +static int +nv84_vp_init(struct drm_device *dev, int engine) +{ + nv_mask(dev, 0x000200, 0x00020000, 0x00000000); + nv_mask(dev, 0x000200, 0x00020000, 0x00020000); + return 0; +} + +static void +nv84_vp_destroy(struct drm_device *dev, int engine) +{ + struct nv84_vp_engine *pvp = nv_engine(dev, engine); + + NVOBJ_ENGINE_DEL(dev, VP); + + kfree(pvp); +} + +int +nv84_vp_create(struct drm_device *dev) +{ + struct nv84_vp_engine *pvp; + + pvp = kzalloc(sizeof(*pvp), GFP_KERNEL); + if (!pvp) + return -ENOMEM; + + pvp->base.destroy = nv84_vp_destroy; + pvp->base.init = nv84_vp_init; + pvp->base.fini = nv84_vp_fini; + + NVOBJ_ENGINE_ADD(dev, VP, &pvp->base); + return 0; +} diff --git a/drivers/gpu/drm/nouveau/nv98_crypt.c b/drivers/gpu/drm/nouveau/nv98_crypt.c new file mode 100644 index 0000000..db94ff0 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nv98_crypt.c @@ -0,0 +1,78 @@ +/* + * Copyright 2011 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs + */ + +#include "drmP.h" +#include "nouveau_drv.h" +#include "nouveau_util.h" +#include "nouveau_vm.h" +#include "nouveau_ramht.h" + +struct nv98_crypt_engine { + struct nouveau_exec_engine base; +}; + +static int +nv98_crypt_fini(struct drm_device *dev, int engine, bool suspend) +{ + if (!(nv_rd32(dev, 0x000200) & 0x00004000)) + return 0; + + nv_mask(dev, 0x000200, 0x00004000, 0x00000000); + return 0; +} + +static int +nv98_crypt_init(struct drm_device *dev, int engine) +{ + nv_mask(dev, 0x000200, 0x00004000, 0x00000000); + nv_mask(dev, 0x000200, 0x00004000, 0x00004000); + return 0; +} + +static void +nv98_crypt_destroy(struct drm_device *dev, int engine) +{ + struct nv98_crypt_engine *pcrypt = nv_engine(dev, engine); + + NVOBJ_ENGINE_DEL(dev, CRYPT); + + kfree(pcrypt); +} + +int +nv98_crypt_create(struct drm_device *dev) +{ + struct nv98_crypt_engine *pcrypt; + + pcrypt = kzalloc(sizeof(*pcrypt), GFP_KERNEL); + if (!pcrypt) + return -ENOMEM; + + pcrypt->base.destroy = nv98_crypt_destroy; + pcrypt->base.init = nv98_crypt_init; + pcrypt->base.fini = nv98_crypt_fini; + + NVOBJ_ENGINE_ADD(dev, CRYPT, &pcrypt->base); + return 0; +} diff --git a/drivers/gpu/drm/nouveau/nv98_ppp.c b/drivers/gpu/drm/nouveau/nv98_ppp.c new file mode 100644 index 0000000..a987dd6 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nv98_ppp.c @@ -0,0 +1,78 @@ +/* + * Copyright 2011 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs + */ + +#include "drmP.h" +#include "nouveau_drv.h" +#include "nouveau_util.h" +#include "nouveau_vm.h" +#include "nouveau_ramht.h" + +struct nv98_ppp_engine { + struct nouveau_exec_engine base; +}; + +static int +nv98_ppp_fini(struct drm_device *dev, int engine, bool suspend) +{ + if (!(nv_rd32(dev, 0x000200) & 0x00000002)) + return 0; + + nv_mask(dev, 0x000200, 0x00000002, 0x00000000); + return 0; +} + +static int +nv98_ppp_init(struct drm_device *dev, int engine) +{ + nv_mask(dev, 0x000200, 0x00000002, 0x00000000); + nv_mask(dev, 0x000200, 0x00000002, 0x00000002); + return 0; +} + +static void +nv98_ppp_destroy(struct drm_device *dev, int engine) +{ + struct nv98_ppp_engine *pppp = nv_engine(dev, engine); + + NVOBJ_ENGINE_DEL(dev, PPP); + + kfree(pppp); +} + +int +nv98_ppp_create(struct drm_device *dev) +{ + struct nv98_ppp_engine *pppp; + + pppp = kzalloc(sizeof(*pppp), GFP_KERNEL); + if (!pppp) + return -ENOMEM; + + pppp->base.destroy = nv98_ppp_destroy; + pppp->base.init = nv98_ppp_init; + pppp->base.fini = nv98_ppp_fini; + + NVOBJ_ENGINE_ADD(dev, PPP, &pppp->base); + return 0; +} diff --git a/drivers/gpu/drm/nouveau/nva3_copy.fuc b/drivers/gpu/drm/nouveau/nva3_copy.fuc index d894731..219850d 100644 --- a/drivers/gpu/drm/nouveau/nva3_copy.fuc +++ b/drivers/gpu/drm/nouveau/nva3_copy.fuc @@ -31,8 +31,9 @@ */ ifdef(`NVA3', -.section nva3_pcopy_data, -.section nvc0_pcopy_data +.section #nva3_pcopy_data +, +.section #nvc0_pcopy_data ) ctx_object: .b32 0 @@ -42,7 +43,7 @@ ctx_dma_query: .b32 0 ctx_dma_src: .b32 0 ctx_dma_dst: .b32 0 ,) -.equ ctx_dma_count 3 +.equ #ctx_dma_count 3 ctx_query_address_high: .b32 0 ctx_query_address_low: .b32 0 ctx_query_counter: .b32 0 @@ -78,64 +79,65 @@ ctx_ycnt: .b32 0 dispatch_table: // mthd 0x0000, NAME .b16 0x000 1 -.b32 ctx_object ~0xffffffff +.b32 #ctx_object ~0xffffffff // mthd 0x0100, NOP .b16 0x040 1 -.b32 0x00010000 + cmd_nop ~0xffffffff +.b32 0x00010000 + #cmd_nop ~0xffffffff // mthd 0x0140, PM_TRIGGER .b16 0x050 1 -.b32 0x00010000 + cmd_pm_trigger ~0xffffffff +.b32 0x00010000 + #cmd_pm_trigger ~0xffffffff ifdef(`NVA3', ` // mthd 0x0180-0x018c, DMA_ -.b16 0x060 ctx_dma_count +.b16 0x060 #ctx_dma_count dispatch_dma: -.b32 0x00010000 + cmd_dma ~0xffffffff -.b32 0x00010000 + cmd_dma ~0xffffffff -.b32 0x00010000 + cmd_dma ~0xffffffff +.b32 0x00010000 + #cmd_dma ~0xffffffff +.b32 0x00010000 + #cmd_dma ~0xffffffff +.b32 0x00010000 + #cmd_dma ~0xffffffff ',) // mthd 0x0200-0x0218, SRC_TILE .b16 0x80 7 -.b32 ctx_src_tile_mode ~0x00000fff -.b32 ctx_src_xsize ~0x0007ffff -.b32 ctx_src_ysize ~0x00001fff -.b32 ctx_src_zsize ~0x000007ff -.b32 ctx_src_zoff ~0x00000fff -.b32 ctx_src_xoff ~0x0007ffff -.b32 ctx_src_yoff ~0x00001fff +.b32 #ctx_src_tile_mode ~0x00000fff +.b32 #ctx_src_xsize ~0x0007ffff +.b32 #ctx_src_ysize ~0x00001fff +.b32 #ctx_src_zsize ~0x000007ff +.b32 #ctx_src_zoff ~0x00000fff +.b32 #ctx_src_xoff ~0x0007ffff +.b32 #ctx_src_yoff ~0x00001fff // mthd 0x0220-0x0238, DST_TILE .b16 0x88 7 -.b32 ctx_dst_tile_mode ~0x00000fff -.b32 ctx_dst_xsize ~0x0007ffff -.b32 ctx_dst_ysize ~0x00001fff -.b32 ctx_dst_zsize ~0x000007ff -.b32 ctx_dst_zoff ~0x00000fff -.b32 ctx_dst_xoff ~0x0007ffff -.b32 ctx_dst_yoff ~0x00001fff +.b32 #ctx_dst_tile_mode ~0x00000fff +.b32 #ctx_dst_xsize ~0x0007ffff +.b32 #ctx_dst_ysize ~0x00001fff +.b32 #ctx_dst_zsize ~0x000007ff +.b32 #ctx_dst_zoff ~0x00000fff +.b32 #ctx_dst_xoff ~0x0007ffff +.b32 #ctx_dst_yoff ~0x00001fff // mthd 0x0300-0x0304, EXEC, WRCACHE_FLUSH .b16 0xc0 2 -.b32 0x00010000 + cmd_exec ~0xffffffff -.b32 0x00010000 + cmd_wrcache_flush ~0xffffffff +.b32 0x00010000 + #cmd_exec ~0xffffffff +.b32 0x00010000 + #cmd_wrcache_flush ~0xffffffff // mthd 0x030c-0x0340, various stuff .b16 0xc3 14 -.b32 ctx_src_address_high ~0x000000ff -.b32 ctx_src_address_low ~0xffffffff -.b32 ctx_dst_address_high ~0x000000ff -.b32 ctx_dst_address_low ~0xffffffff -.b32 ctx_src_pitch ~0x0007ffff -.b32 ctx_dst_pitch ~0x0007ffff -.b32 ctx_xcnt ~0x0000ffff -.b32 ctx_ycnt ~0x00001fff -.b32 ctx_format ~0x0333ffff -.b32 ctx_swz_const0 ~0xffffffff -.b32 ctx_swz_const1 ~0xffffffff -.b32 ctx_query_address_high ~0x000000ff -.b32 ctx_query_address_low ~0xffffffff -.b32 ctx_query_counter ~0xffffffff +.b32 #ctx_src_address_high ~0x000000ff +.b32 #ctx_src_address_low ~0xffffffff +.b32 #ctx_dst_address_high ~0x000000ff +.b32 #ctx_dst_address_low ~0xffffffff +.b32 #ctx_src_pitch ~0x0007ffff +.b32 #ctx_dst_pitch ~0x0007ffff +.b32 #ctx_xcnt ~0x0000ffff +.b32 #ctx_ycnt ~0x00001fff +.b32 #ctx_format ~0x0333ffff +.b32 #ctx_swz_const0 ~0xffffffff +.b32 #ctx_swz_const1 ~0xffffffff +.b32 #ctx_query_address_high ~0x000000ff +.b32 #ctx_query_address_low ~0xffffffff +.b32 #ctx_query_counter ~0xffffffff .b16 0x800 0 ifdef(`NVA3', -.section nva3_pcopy_code, -.section nvc0_pcopy_code +.section #nva3_pcopy_code +, +.section #nvc0_pcopy_code ) main: @@ -143,12 +145,12 @@ main: mov $sp $r0 // setup i0 handler and route fifo and ctxswitch to it - mov $r1 ih + mov $r1 #ih mov $iv0 $r1 mov $r1 0x400 movw $r2 0xfff3 sethi $r2 0 - iowr I[$r2 + 0x300] $r2 + iowr I[$r1 + 0x300] $r2 // enable interrupts or $r2 0xc @@ -164,19 +166,19 @@ main: bset $flags $p0 spin: sleep $p0 - bra spin + bra #spin // i0 handler ih: iord $r1 I[$r0 + 0x200] and $r2 $r1 0x00000008 - bra e ih_no_chsw - call chsw + bra e #ih_no_chsw + call #chsw ih_no_chsw: and $r2 $r1 0x00000004 - bra e ih_no_cmd - call dispatch + bra e #ih_no_cmd + call #dispatch ih_no_cmd: and $r1 $r1 0x0000000c @@ -235,9 +237,9 @@ ifdef(`NVA3', ` sethi $r4 0x60000 // swap! - bra $p1 swctx_load + bra $p1 #swctx_load xdst $r0 $r4 - bra swctx_done + bra #swctx_done swctx_load: xdld $r0 $r4 swctx_done: @@ -251,9 +253,9 @@ chsw: // if it's active, unload it and return xbit $r15 $r3 0x1e - bra e chsw_no_unload + bra e #chsw_no_unload bclr $flags $p1 - call swctx + call #swctx bclr $r3 0x1e iowr I[$r2] $r3 mov $r4 1 @@ -266,20 +268,20 @@ chsw: // is there a channel waiting to be loaded? xbit $r13 $r3 0x1e - bra e chsw_finish_load + bra e #chsw_finish_load bset $flags $p1 - call swctx + call #swctx ifdef(`NVA3', // load dma objects back into TARGET regs - mov $r5 ctx_dma - mov $r6 ctx_dma_count + mov $r5 #ctx_dma + mov $r6 #ctx_dma_count chsw_load_ctx_dma: ld b32 $r7 D[$r5 + $r6 * 4] add b32 $r8 $r6 0x180 shl b32 $r8 8 iowr I[$r8] $r7 sub b32 $r6 1 - bra nc chsw_load_ctx_dma + bra nc #chsw_load_ctx_dma ,) chsw_finish_load: @@ -297,7 +299,7 @@ dispatch: shl b32 $r2 0x10 // lookup method in the dispatch table, ILLEGAL_MTHD if not found - mov $r5 dispatch_table + mov $r5 #dispatch_table clear b32 $r6 clear b32 $r7 dispatch_loop: @@ -305,14 +307,14 @@ dispatch: ld b16 $r7 D[$r5 + 2] add b32 $r5 4 cmpu b32 $r4 $r6 - bra c dispatch_illegal_mthd + bra c #dispatch_illegal_mthd add b32 $r7 $r6 cmpu b32 $r4 $r7 - bra c dispatch_valid_mthd + bra c #dispatch_valid_mthd sub b32 $r7 $r6 shl b32 $r7 3 add b32 $r5 $r7 - bra dispatch_loop + bra #dispatch_loop // ensure no bits set in reserved fields, INVALID_BITFIELD dispatch_valid_mthd: @@ -322,20 +324,20 @@ dispatch: ld b32 $r5 D[$r4 + 4] and $r5 $r3 cmpu b32 $r5 0 - bra ne dispatch_invalid_bitfield + bra ne #dispatch_invalid_bitfield // depending on dispatch flags: execute method, or save data as state ld b16 $r5 D[$r4 + 0] ld b16 $r6 D[$r4 + 2] cmpu b32 $r6 0 - bra ne dispatch_cmd + bra ne #dispatch_cmd st b32 D[$r5] $r3 - bra dispatch_done + bra #dispatch_done dispatch_cmd: bclr $flags $p1 call $r5 - bra $p1 dispatch_error - bra dispatch_done + bra $p1 #dispatch_error + bra #dispatch_done dispatch_invalid_bitfield: or $r2 2 @@ -353,7 +355,7 @@ dispatch: iord $r2 I[$r0 + 0x200] and $r2 0x40 cmpu b32 $r2 0 - bra ne hostirq_wait + bra ne #hostirq_wait dispatch_done: mov $r2 0x1d00 @@ -409,10 +411,10 @@ ifdef(`NVA3', // $r2: hostirq state // $r3: data cmd_dma: - sub b32 $r4 dispatch_dma + sub b32 $r4 #dispatch_dma shr b32 $r4 1 bset $r3 0x1e - st b32 D[$r4 + ctx_dma] $r3 + st b32 D[$r4 + #ctx_dma] $r3 add b32 $r4 0x600 shl b32 $r4 6 iowr I[$r4] $r3 @@ -430,7 +432,7 @@ cmd_exec_set_format: st b32 D[$sp + 0x0c] $r0 // extract cpp, src_ncomp and dst_ncomp from FORMAT - ld b32 $r4 D[$r0 + ctx_format] + ld b32 $r4 D[$r0 + #ctx_format] extr $r5 $r4 16:17 add b32 $r5 1 extr $r6 $r4 20:21 @@ -448,22 +450,22 @@ cmd_exec_set_format: clear b32 $r11 bpc_loop: cmpu b8 $r10 4 - bra nc cmp_c0 + bra nc #cmp_c0 mulu $r12 $r10 $r5 add b32 $r12 $r11 bset $flags $p2 - bra bpc_next + bra #bpc_next cmp_c0: - bra ne cmp_c1 + bra ne #cmp_c1 mov $r12 0x10 add b32 $r12 $r11 - bra bpc_next + bra #bpc_next cmp_c1: cmpu b8 $r10 6 - bra nc cmp_zero + bra nc #cmp_zero mov $r12 0x14 add b32 $r12 $r11 - bra bpc_next + bra #bpc_next cmp_zero: mov $r12 0x80 bpc_next: @@ -471,22 +473,22 @@ cmd_exec_set_format: add b32 $r8 1 add b32 $r11 1 cmpu b32 $r11 $r5 - bra c bpc_loop + bra c #bpc_loop add b32 $r9 1 cmpu b32 $r9 $r7 - bra c ncomp_loop + bra c #ncomp_loop // SRC_XCNT = (xcnt * src_cpp), or 0 if no src ref in swz (hw will hang) mulu $r6 $r5 - st b32 D[$r0 + ctx_src_cpp] $r6 - ld b32 $r8 D[$r0 + ctx_xcnt] + st b32 D[$r0 + #ctx_src_cpp] $r6 + ld b32 $r8 D[$r0 + #ctx_xcnt] mulu $r6 $r8 - bra $p2 dst_xcnt + bra $p2 #dst_xcnt clear b32 $r6 dst_xcnt: mulu $r7 $r5 - st b32 D[$r0 + ctx_dst_cpp] $r7 + st b32 D[$r0 + #ctx_dst_cpp] $r7 mulu $r7 $r8 mov $r5 0x810 @@ -494,10 +496,10 @@ cmd_exec_set_format: iowr I[$r5 + 0x000] $r6 iowr I[$r5 + 0x100] $r7 add b32 $r5 0x800 - ld b32 $r6 D[$r0 + ctx_dst_cpp] + ld b32 $r6 D[$r0 + #ctx_dst_cpp] sub b32 $r6 1 shl b32 $r6 8 - ld b32 $r7 D[$r0 + ctx_src_cpp] + ld b32 $r7 D[$r0 + #ctx_src_cpp] sub b32 $r7 1 or $r6 $r7 iowr I[$r5 + 0x000] $r6 @@ -511,9 +513,9 @@ cmd_exec_set_format: ld b32 $r6 D[$sp + 0x0c] iowr I[$r5 + 0x300] $r6 add b32 $r5 0x400 - ld b32 $r6 D[$r0 + ctx_swz_const0] + ld b32 $r6 D[$r0 + #ctx_swz_const0] iowr I[$r5 + 0x000] $r6 - ld b32 $r6 D[$r0 + ctx_swz_const1] + ld b32 $r6 D[$r0 + #ctx_swz_const1] iowr I[$r5 + 0x100] $r6 add $sp 0x10 ret @@ -543,7 +545,7 @@ cmd_exec_set_format: // cmd_exec_set_surface_tiled: // translate TILE_MODE into Tp, Th, Td shift values - ld b32 $r7 D[$r5 + ctx_src_tile_mode] + ld b32 $r7 D[$r5 + #ctx_src_tile_mode] extr $r9 $r7 8:11 extr $r8 $r7 4:7 ifdef(`NVA3', @@ -553,9 +555,9 @@ ifdef(`NVA3', ) extr $r7 $r7 0:3 cmp b32 $r7 0xe - bra ne xtile64 + bra ne #xtile64 mov $r7 4 - bra xtileok + bra #xtileok xtile64: xbit $r7 $flags $p2 add b32 $r7 17 @@ -565,8 +567,8 @@ ifdef(`NVA3', // Op = (x * cpp) & ((1 << Tp) - 1) // Tx = (x * cpp) >> Tp - ld b32 $r10 D[$r5 + ctx_src_xoff] - ld b32 $r11 D[$r5 + ctx_src_cpp] + ld b32 $r10 D[$r5 + #ctx_src_xoff] + ld b32 $r11 D[$r5 + #ctx_src_cpp] mulu $r10 $r11 mov $r11 1 shl b32 $r11 $r7 @@ -576,7 +578,7 @@ ifdef(`NVA3', // Tyo = y & ((1 << Th) - 1) // Ty = y >> Th - ld b32 $r13 D[$r5 + ctx_src_yoff] + ld b32 $r13 D[$r5 + #ctx_src_yoff] mov $r14 1 shl b32 $r14 $r8 sub b32 $r14 1 @@ -598,8 +600,8 @@ ifdef(`NVA3', add b32 $r12 $r11 // nTx = ((w * cpp) + ((1 << Tp) - 1) >> Tp) - ld b32 $r15 D[$r5 + ctx_src_xsize] - ld b32 $r11 D[$r5 + ctx_src_cpp] + ld b32 $r15 D[$r5 + #ctx_src_xsize] + ld b32 $r11 D[$r5 + #ctx_src_cpp] mulu $r15 $r11 mov $r11 1 shl b32 $r11 $r7 @@ -609,7 +611,7 @@ ifdef(`NVA3', push $r15 // nTy = (h + ((1 << Th) - 1)) >> Th - ld b32 $r15 D[$r5 + ctx_src_ysize] + ld b32 $r15 D[$r5 + #ctx_src_ysize] mov $r11 1 shl b32 $r11 $r8 sub b32 $r11 1 @@ -629,7 +631,7 @@ ifdef(`NVA3', // Tz = z >> Td // Op += Tzo << Tys // Ts = Tys + Td - ld b32 $r8 D[$r5 + ctx_src_zoff] + ld b32 $r8 D[$r5 + #ctx_src_zoff] mov $r14 1 shl b32 $r14 $r9 sub b32 $r14 1 @@ -656,8 +658,8 @@ ifdef(`NVA3', // SRC_ADDRESS_LOW = (Ot + Op) & 0xffffffff // CFG_ADDRESS_HIGH |= ((Ot + Op) >> 32) << 16 - ld b32 $r7 D[$r5 + ctx_src_address_low] - ld b32 $r8 D[$r5 + ctx_src_address_high] + ld b32 $r7 D[$r5 + #ctx_src_address_low] + ld b32 $r8 D[$r5 + #ctx_src_address_high] add b32 $r10 $r12 add b32 $r7 $r10 adc b32 $r8 0 @@ -677,14 +679,14 @@ cmd_exec_set_surface_linear: xbit $r6 $flags $p2 add b32 $r6 0x202 shl b32 $r6 8 - ld b32 $r7 D[$r5 + ctx_src_address_low] + ld b32 $r7 D[$r5 + #ctx_src_address_low] iowr I[$r6 + 0x000] $r7 add b32 $r6 0x400 - ld b32 $r7 D[$r5 + ctx_src_address_high] + ld b32 $r7 D[$r5 + #ctx_src_address_high] shl b32 $r7 16 iowr I[$r6 + 0x000] $r7 add b32 $r6 0x400 - ld b32 $r7 D[$r5 + ctx_src_pitch] + ld b32 $r7 D[$r5 + #ctx_src_pitch] iowr I[$r6 + 0x000] $r7 ret @@ -697,7 +699,7 @@ cmd_exec_wait: loop: iord $r1 I[$r0] and $r1 1 - bra ne loop + bra ne #loop pop $r1 pop $r0 ret @@ -705,18 +707,18 @@ cmd_exec_wait: cmd_exec_query: // if QUERY_SHORT not set, write out { -, 0, TIME_LO, TIME_HI } xbit $r4 $r3 13 - bra ne query_counter - call cmd_exec_wait + bra ne #query_counter + call #cmd_exec_wait mov $r4 0x80c shl b32 $r4 6 - ld b32 $r5 D[$r0 + ctx_query_address_low] + ld b32 $r5 D[$r0 + #ctx_query_address_low] add b32 $r5 4 iowr I[$r4 + 0x000] $r5 iowr I[$r4 + 0x100] $r0 mov $r5 0xc iowr I[$r4 + 0x200] $r5 add b32 $r4 0x400 - ld b32 $r5 D[$r0 + ctx_query_address_high] + ld b32 $r5 D[$r0 + #ctx_query_address_high] shl b32 $r5 16 iowr I[$r4 + 0x000] $r5 add b32 $r4 0x500 @@ -741,16 +743,16 @@ cmd_exec_query: // write COUNTER query_counter: - call cmd_exec_wait + call #cmd_exec_wait mov $r4 0x80c shl b32 $r4 6 - ld b32 $r5 D[$r0 + ctx_query_address_low] + ld b32 $r5 D[$r0 + #ctx_query_address_low] iowr I[$r4 + 0x000] $r5 iowr I[$r4 + 0x100] $r0 mov $r5 0x4 iowr I[$r4 + 0x200] $r5 add b32 $r4 0x400 - ld b32 $r5 D[$r0 + ctx_query_address_high] + ld b32 $r5 D[$r0 + #ctx_query_address_high] shl b32 $r5 16 iowr I[$r4 + 0x000] $r5 add b32 $r4 0x500 @@ -759,7 +761,7 @@ cmd_exec_query: mov $r5 0x00001110 sethi $r5 0x13120000 iowr I[$r4 + 0x100] $r5 - ld b32 $r5 D[$r0 + ctx_query_counter] + ld b32 $r5 D[$r0 + #ctx_query_counter] add b32 $r4 0x500 iowr I[$r4 + 0x000] $r5 mov $r5 0x00002601 @@ -787,22 +789,22 @@ cmd_exec_query: // $r2: hostirq state // $r3: data cmd_exec: - call cmd_exec_wait + call #cmd_exec_wait // if format requested, call function to calculate it, otherwise // fill in cpp/xcnt for both surfaces as if (cpp == 1) xbit $r15 $r3 0 - bra e cmd_exec_no_format - call cmd_exec_set_format + bra e #cmd_exec_no_format + call #cmd_exec_set_format mov $r4 0x200 - bra cmd_exec_init_src_surface + bra #cmd_exec_init_src_surface cmd_exec_no_format: mov $r6 0x810 shl b32 $r6 6 mov $r7 1 - st b32 D[$r0 + ctx_src_cpp] $r7 - st b32 D[$r0 + ctx_dst_cpp] $r7 - ld b32 $r7 D[$r0 + ctx_xcnt] + st b32 D[$r0 + #ctx_src_cpp] $r7 + st b32 D[$r0 + #ctx_dst_cpp] $r7 + ld b32 $r7 D[$r0 + #ctx_xcnt] iowr I[$r6 + 0x000] $r7 iowr I[$r6 + 0x100] $r7 clear b32 $r4 @@ -811,28 +813,28 @@ cmd_exec: bclr $flags $p2 clear b32 $r5 xbit $r15 $r3 4 - bra e src_tiled - call cmd_exec_set_surface_linear - bra cmd_exec_init_dst_surface + bra e #src_tiled + call #cmd_exec_set_surface_linear + bra #cmd_exec_init_dst_surface src_tiled: - call cmd_exec_set_surface_tiled + call #cmd_exec_set_surface_tiled bset $r4 7 cmd_exec_init_dst_surface: bset $flags $p2 - mov $r5 ctx_dst_address_high - ctx_src_address_high + mov $r5 #ctx_dst_address_high - #ctx_src_address_high xbit $r15 $r3 8 - bra e dst_tiled - call cmd_exec_set_surface_linear - bra cmd_exec_kick + bra e #dst_tiled + call #cmd_exec_set_surface_linear + bra #cmd_exec_kick dst_tiled: - call cmd_exec_set_surface_tiled + call #cmd_exec_set_surface_tiled bset $r4 8 cmd_exec_kick: mov $r5 0x800 shl b32 $r5 6 - ld b32 $r6 D[$r0 + ctx_ycnt] + ld b32 $r6 D[$r0 + #ctx_ycnt] iowr I[$r5 + 0x100] $r6 mov $r6 0x0041 // SRC_TARGET = 1, DST_TARGET = 2 @@ -842,8 +844,8 @@ cmd_exec: // if requested, queue up a QUERY write after the copy has completed xbit $r15 $r3 12 - bra e cmd_exec_done - call cmd_exec_query + bra e #cmd_exec_done + call #cmd_exec_query cmd_exec_done: ret diff --git a/drivers/gpu/drm/nouveau/nva3_copy.fuc.h b/drivers/gpu/drm/nouveau/nva3_copy.fuc.h index e2a0e88..37d6de3 100644 --- a/drivers/gpu/drm/nouveau/nva3_copy.fuc.h +++ b/drivers/gpu/drm/nouveau/nva3_copy.fuc.h @@ -190,7 +190,7 @@ u32 nva3_pcopy_code[] = { 0xf10010fe, 0xf1040017, 0xf0fff327, - 0x22d00023, + 0x12d00023, 0x0c25f0c0, 0xf40012d0, 0x17f11031, diff --git a/drivers/gpu/drm/nouveau/nva3_pm.c b/drivers/gpu/drm/nouveau/nva3_pm.c index 618c144..9e636e6 100644 --- a/drivers/gpu/drm/nouveau/nva3_pm.c +++ b/drivers/gpu/drm/nouveau/nva3_pm.c @@ -287,12 +287,13 @@ nva3_pm_grcp_idle(void *data) return false; } -void +int nva3_pm_clocks_set(struct drm_device *dev, void *pre_state) { struct drm_nouveau_private *dev_priv = dev->dev_private; struct nva3_pm_state *info = pre_state; unsigned long flags; + int ret = -EAGAIN; /* prevent any new grctx switches from starting */ spin_lock_irqsave(&dev_priv->context_switch_lock, flags); @@ -328,6 +329,8 @@ nva3_pm_clocks_set(struct drm_device *dev, void *pre_state) nv_wr32(dev, 0x100210, 0x80000000); } + ret = 0; + cleanup: /* unfreeze PFIFO */ nv_mask(dev, 0x002504, 0x00000001, 0x00000000); @@ -339,4 +342,5 @@ cleanup: nv_mask(dev, 0x400824, 0x10000000, 0x10000000); spin_unlock_irqrestore(&dev_priv->context_switch_lock, flags); kfree(info); + return ret; } diff --git a/drivers/gpu/drm/nouveau/nvc0_copy.fuc.h b/drivers/gpu/drm/nouveau/nvc0_copy.fuc.h index 9e87036..cd879f3 100644 --- a/drivers/gpu/drm/nouveau/nvc0_copy.fuc.h +++ b/drivers/gpu/drm/nouveau/nvc0_copy.fuc.h @@ -178,7 +178,7 @@ u32 nvc0_pcopy_code[] = { 0xf10010fe, 0xf1040017, 0xf0fff327, - 0x22d00023, + 0x12d00023, 0x0c25f0c0, 0xf40012d0, 0x17f11031, diff --git a/drivers/gpu/drm/nouveau/nvc0_fb.c b/drivers/gpu/drm/nouveau/nvc0_fb.c index 5bf5503..f704e94 100644 --- a/drivers/gpu/drm/nouveau/nvc0_fb.c +++ b/drivers/gpu/drm/nouveau/nvc0_fb.c @@ -54,6 +54,11 @@ nvc0_mfb_isr(struct drm_device *dev) nvc0_mfb_subp_isr(dev, unit, subp); units &= ~(1 << unit); } + + /* we do something horribly wrong and upset PMFB a lot, so mask off + * interrupts from it after the first one until it's fixed + */ + nv_mask(dev, 0x000640, 0x02000000, 0x00000000); } static void diff --git a/drivers/gpu/drm/nouveau/nvc0_fifo.c b/drivers/gpu/drm/nouveau/nvc0_fifo.c index dcbe0d5..50d68a7 100644 --- a/drivers/gpu/drm/nouveau/nvc0_fifo.c +++ b/drivers/gpu/drm/nouveau/nvc0_fifo.c @@ -436,6 +436,24 @@ nvc0_fifo_isr_vm_fault(struct drm_device *dev, int unit) printk(" on channel 0x%010llx\n", (u64)inst << 12); } +static int +nvc0_fifo_page_flip(struct drm_device *dev, u32 chid) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nouveau_channel *chan = NULL; + unsigned long flags; + int ret = -EINVAL; + + spin_lock_irqsave(&dev_priv->channels.lock, flags); + if (likely(chid >= 0 && chid < dev_priv->engine.fifo.channels)) { + chan = dev_priv->channels.ptr[chid]; + if (likely(chan)) + ret = nouveau_finish_page_flip(chan, NULL); + } + spin_unlock_irqrestore(&dev_priv->channels.lock, flags); + return ret; +} + static void nvc0_fifo_isr_subfifo_intr(struct drm_device *dev, int unit) { @@ -445,11 +463,21 @@ nvc0_fifo_isr_subfifo_intr(struct drm_device *dev, int unit) u32 chid = nv_rd32(dev, 0x040120 + (unit * 0x2000)) & 0x7f; u32 subc = (addr & 0x00070000); u32 mthd = (addr & 0x00003ffc); + u32 show = stat; - NV_INFO(dev, "PSUBFIFO %d:", unit); - nouveau_bitfield_print(nvc0_fifo_subfifo_intr, stat); - NV_INFO(dev, "PSUBFIFO %d: ch %d subc %d mthd 0x%04x data 0x%08x\n", - unit, chid, subc, mthd, data); + if (stat & 0x00200000) { + if (mthd == 0x0054) { + if (!nvc0_fifo_page_flip(dev, chid)) + show &= ~0x00200000; + } + } + + if (show) { + NV_INFO(dev, "PFIFO%d:", unit); + nouveau_bitfield_print(nvc0_fifo_subfifo_intr, show); + NV_INFO(dev, "PFIFO%d: ch %d subc %d mthd 0x%04x data 0x%08x\n", + unit, chid, subc, mthd, data); + } nv_wr32(dev, 0x0400c0 + (unit * 0x2000), 0x80600008); nv_wr32(dev, 0x040108 + (unit * 0x2000), stat); diff --git a/drivers/gpu/drm/nouveau/nvc0_graph.c b/drivers/gpu/drm/nouveau/nvc0_graph.c index ecfafd7..9066102 100644 --- a/drivers/gpu/drm/nouveau/nvc0_graph.c +++ b/drivers/gpu/drm/nouveau/nvc0_graph.c @@ -333,14 +333,6 @@ nvc0_graph_fini(struct drm_device *dev, int engine, bool suspend) return 0; } -static int -nvc0_graph_mthd_page_flip(struct nouveau_channel *chan, - u32 class, u32 mthd, u32 data) -{ - nouveau_finish_page_flip(chan, NULL); - return 0; -} - static void nvc0_graph_init_obj418880(struct drm_device *dev) { @@ -875,19 +867,20 @@ nvc0_graph_create(struct drm_device *dev) case 0xcf: /* 4/0/0/0, 3 */ priv->magic_not_rop_nr = 0x03; break; + case 0xd9: /* 1/0/0/0, 1 */ + priv->magic_not_rop_nr = 0x01; + break; } if (!priv->magic_not_rop_nr) { NV_ERROR(dev, "PGRAPH: unknown config: %d/%d/%d/%d, %d\n", priv->tp_nr[0], priv->tp_nr[1], priv->tp_nr[2], priv->tp_nr[3], priv->rop_nr); - /* use 0xc3's values... */ - priv->magic_not_rop_nr = 0x03; + priv->magic_not_rop_nr = 0x00; } NVOBJ_CLASS(dev, 0x902d, GR); /* 2D */ NVOBJ_CLASS(dev, 0x9039, GR); /* M2MF */ - NVOBJ_MTHD (dev, 0x9039, 0x0500, nvc0_graph_mthd_page_flip); NVOBJ_CLASS(dev, 0x9097, GR); /* 3D */ if (fermi >= 0x9197) NVOBJ_CLASS(dev, 0x9197, GR); /* 3D (NVC1-) */ diff --git a/drivers/gpu/drm/nouveau/nvc0_graph.fuc b/drivers/gpu/drm/nouveau/nvc0_graph.fuc index 2a4b6dc..e6b2288 100644 --- a/drivers/gpu/drm/nouveau/nvc0_graph.fuc +++ b/drivers/gpu/drm/nouveau/nvc0_graph.fuc @@ -71,9 +71,9 @@ queue_put: ld b32 $r9 D[$r13 + 0x4] // PUT xor $r8 8 cmpu b32 $r8 $r9 - bra ne queue_put_next + bra ne #queue_put_next mov $r15 E_CMD_OVERFLOW - call error + call #error ret // store cmd/data on queue @@ -104,7 +104,7 @@ queue_get: ld b32 $r8 D[$r13 + 0x0] // GET ld b32 $r9 D[$r13 + 0x4] // PUT cmpu b32 $r8 $r9 - bra e queue_get_done + bra e #queue_get_done // fetch first cmd/data pair and $r9 $r8 7 shl b32 $r9 3 @@ -135,9 +135,9 @@ nv_rd32: nv_rd32_wait: iord $r12 I[$r11 + 0x000] xbit $r12 $r12 31 - bra ne nv_rd32_wait + bra ne #nv_rd32_wait mov $r10 6 // DONE_MMIO_RD - call wait_doneo + call #wait_doneo iord $r15 I[$r11 + 0x100] // MMIO_RDVAL ret @@ -157,7 +157,7 @@ nv_wr32: nv_wr32_wait: iord $r12 I[$r11 + 0x000] xbit $r12 $r12 31 - bra ne nv_wr32_wait + bra ne #nv_wr32_wait ret // (re)set watchdog timer @@ -193,7 +193,7 @@ $1: shl b32 $r8 6 iord $r8 I[$r8 + 0x000] // DONE xbit $r8 $r8 $r10 - bra $2 wait_done_$1 + bra $2 #wait_done_$1 trace_clr(T_WAIT) ret ') @@ -216,7 +216,7 @@ mmctx_size: add b32 $r9 $r8 add b32 $r14 4 cmpu b32 $r14 $r15 - bra ne nv_mmctx_size_loop + bra ne #nv_mmctx_size_loop mov b32 $r15 $r9 ret @@ -238,12 +238,12 @@ mmctx_xfer: shl b32 $r8 6 clear b32 $r9 or $r11 $r11 - bra e mmctx_base_disabled + bra e #mmctx_base_disabled iowr I[$r8 + 0x000] $r11 // MMCTX_BASE bset $r9 0 // BASE_EN mmctx_base_disabled: or $r14 $r14 - bra e mmctx_multi_disabled + bra e #mmctx_multi_disabled iowr I[$r8 + 0x200] $r14 // MMCTX_MULTI_STRIDE iowr I[$r8 + 0x300] $r15 // MMCTX_MULTI_MASK bset $r9 1 // MULTI_EN @@ -264,7 +264,7 @@ mmctx_xfer: mmctx_wait_free: iord $r14 I[$r8 + 0x000] // MMCTX_CTRL and $r14 0x1f - bra e mmctx_wait_free + bra e #mmctx_wait_free // queue up an entry ld b32 $r14 D[$r12] @@ -272,19 +272,19 @@ mmctx_xfer: iowr I[$r8 + 0x300] $r14 add b32 $r12 4 cmpu b32 $r12 $r13 - bra ne mmctx_exec_loop + bra ne #mmctx_exec_loop xbit $r11 $r10 2 - bra ne mmctx_stop + bra ne #mmctx_stop // wait for queue to empty mmctx_fini_wait: iord $r11 I[$r8 + 0x000] // MMCTX_CTRL and $r11 0x1f cmpu b32 $r11 0x10 - bra ne mmctx_fini_wait + bra ne #mmctx_fini_wait mov $r10 2 // DONE_MMCTX - call wait_donez - bra mmctx_done + call #wait_donez + bra #mmctx_done mmctx_stop: xbit $r11 $r10 0 shl b32 $r11 16 // DIR @@ -295,7 +295,7 @@ mmctx_xfer: // wait for STOP_TRIGGER to clear iord $r11 I[$r8 + 0x000] // MMCTX_CTRL xbit $r11 $r11 18 - bra ne mmctx_stop_wait + bra ne #mmctx_stop_wait mmctx_done: trace_clr(T_MMCTX) ret @@ -305,7 +305,7 @@ mmctx_xfer: strand_wait: push $r10 mov $r10 2 - call wait_donez + call #wait_donez pop $r10 ret @@ -316,7 +316,7 @@ strand_pre: sethi $r8 0x20000 mov $r9 0xc iowr I[$r8] $r9 - call strand_wait + call #strand_wait ret // unknown - call after issuing strand commands @@ -326,7 +326,7 @@ strand_post: sethi $r8 0x20000 mov $r9 0xd iowr I[$r8] $r9 - call strand_wait + call #strand_wait ret // Selects strand set?! @@ -341,11 +341,11 @@ strand_set: iowr I[$r10 + 0x000] $r12 // 0x93c = 0xf mov $r12 0xb iowr I[$r11 + 0x000] $r12 // 0x928 = 0xb - call strand_wait + call #strand_wait iowr I[$r10 + 0x000] $r14 // 0x93c = mov $r12 0xa iowr I[$r11 + 0x000] $r12 // 0x928 = 0xa - call strand_wait + call #strand_wait ret // Initialise strand context data @@ -357,22 +357,22 @@ strand_set: // strand_ctx_init: trace_set(T_STRINIT) - call strand_pre + call #strand_pre mov $r14 3 - call strand_set + call #strand_set mov $r10 0x46fc sethi $r10 0x20000 add b32 $r11 $r10 0x400 iowr I[$r10 + 0x100] $r0 // STRAND_FIRST_GENE = 0 mov $r12 1 iowr I[$r11 + 0x000] $r12 // STRAND_CMD = LATCH_FIRST_GENE - call strand_wait + call #strand_wait sub b32 $r12 $r0 1 iowr I[$r10 + 0x000] $r12 // STRAND_GENE_CNT = 0xffffffff mov $r12 2 iowr I[$r11 + 0x000] $r12 // STRAND_CMD = LATCH_GENE_CNT - call strand_wait - call strand_post + call #strand_wait + call #strand_post // read the size of each strand, poke the context offset of // each into STRAND_{SAVE,LOAD}_SWBASE now, no need to worry @@ -391,7 +391,7 @@ strand_ctx_init: add b32 $r14 $r10 add b32 $r8 4 sub b32 $r9 1 - bra ne ctx_init_strand_loop + bra ne #ctx_init_strand_loop shl b32 $r14 8 sub b32 $r15 $r14 $r15 diff --git a/drivers/gpu/drm/nouveau/nvc0_graph.h b/drivers/gpu/drm/nouveau/nvc0_graph.h index 636fe98..91d44ea 100644 --- a/drivers/gpu/drm/nouveau/nvc0_graph.h +++ b/drivers/gpu/drm/nouveau/nvc0_graph.h @@ -87,6 +87,7 @@ nvc0_graph_class(struct drm_device *dev) case 0xc1: return 0x9197; case 0xc8: + case 0xd9: return 0x9297; default: return 0; diff --git a/drivers/gpu/drm/nouveau/nvc0_grctx.c b/drivers/gpu/drm/nouveau/nvc0_grctx.c index 96b0b93d..de77842 100644 --- a/drivers/gpu/drm/nouveau/nvc0_grctx.c +++ b/drivers/gpu/drm/nouveau/nvc0_grctx.c @@ -1268,6 +1268,17 @@ nvc0_grctx_generate_9039(struct drm_device *dev) static void nvc0_grctx_generate_90c0(struct drm_device *dev) { + struct drm_nouveau_private *dev_priv = dev->dev_private; + int i; + + for (i = 0; dev_priv->chipset == 0xd9 && i < 4; i++) { + nv_mthd(dev, 0x90c0, 0x2700 + (i * 0x40), 0x00000000); + nv_mthd(dev, 0x90c0, 0x2720 + (i * 0x40), 0x00000000); + nv_mthd(dev, 0x90c0, 0x2704 + (i * 0x40), 0x00000000); + nv_mthd(dev, 0x90c0, 0x2724 + (i * 0x40), 0x00000000); + nv_mthd(dev, 0x90c0, 0x2708 + (i * 0x40), 0x00000000); + nv_mthd(dev, 0x90c0, 0x2728 + (i * 0x40), 0x00000000); + } nv_mthd(dev, 0x90c0, 0x270c, 0x00000000); nv_mthd(dev, 0x90c0, 0x272c, 0x00000000); nv_mthd(dev, 0x90c0, 0x274c, 0x00000000); @@ -1276,6 +1287,12 @@ nvc0_grctx_generate_90c0(struct drm_device *dev) nv_mthd(dev, 0x90c0, 0x27ac, 0x00000000); nv_mthd(dev, 0x90c0, 0x27cc, 0x00000000); nv_mthd(dev, 0x90c0, 0x27ec, 0x00000000); + for (i = 0; dev_priv->chipset == 0xd9 && i < 4; i++) { + nv_mthd(dev, 0x90c0, 0x2710 + (i * 0x40), 0x00014000); + nv_mthd(dev, 0x90c0, 0x2730 + (i * 0x40), 0x00014000); + nv_mthd(dev, 0x90c0, 0x2714 + (i * 0x40), 0x00000040); + nv_mthd(dev, 0x90c0, 0x2734 + (i * 0x40), 0x00000040); + } nv_mthd(dev, 0x90c0, 0x030c, 0x00000001); nv_mthd(dev, 0x90c0, 0x1944, 0x00000000); nv_mthd(dev, 0x90c0, 0x0758, 0x00000100); @@ -1471,14 +1488,20 @@ nvc0_grctx_generate_shaders(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private; - if (dev_priv->chipset != 0xc1) { - nv_wr32(dev, 0x405800, 0x078000bf); - nv_wr32(dev, 0x405830, 0x02180000); - } else { + if (dev_priv->chipset == 0xd9) { nv_wr32(dev, 0x405800, 0x0f8000bf); nv_wr32(dev, 0x405830, 0x02180218); + nv_wr32(dev, 0x405834, 0x08000000); + } else + if (dev_priv->chipset == 0xc1) { + nv_wr32(dev, 0x405800, 0x0f8000bf); + nv_wr32(dev, 0x405830, 0x02180218); + nv_wr32(dev, 0x405834, 0x00000000); + } else { + nv_wr32(dev, 0x405800, 0x078000bf); + nv_wr32(dev, 0x405830, 0x02180000); + nv_wr32(dev, 0x405834, 0x00000000); } - nv_wr32(dev, 0x405834, 0x00000000); nv_wr32(dev, 0x405838, 0x00000000); nv_wr32(dev, 0x405854, 0x00000000); nv_wr32(dev, 0x405870, 0x00000001); @@ -1509,7 +1532,10 @@ nvc0_grctx_generate_unk64xx(struct drm_device *dev) nv_wr32(dev, 0x4064ac, 0x00003fff); nv_wr32(dev, 0x4064b4, 0x00000000); nv_wr32(dev, 0x4064b8, 0x00000000); - if (dev_priv->chipset == 0xc1) { + if (dev_priv->chipset == 0xd9) + nv_wr32(dev, 0x4064bc, 0x00000000); + if (dev_priv->chipset == 0xc1 || + dev_priv->chipset == 0xd9) { nv_wr32(dev, 0x4064c0, 0x80140078); nv_wr32(dev, 0x4064c4, 0x0086ffff); } @@ -1550,10 +1576,23 @@ nvc0_grctx_generate_rop(struct drm_device *dev) /* ROPC_BROADCAST */ nv_wr32(dev, 0x408800, 0x02802a3c); nv_wr32(dev, 0x408804, 0x00000040); - nv_wr32(dev, 0x408808, chipset != 0xc1 ? 0x0003e00d : 0x1003e005); - nv_wr32(dev, 0x408900, 0x3080b801); - nv_wr32(dev, 0x408904, chipset != 0xc1 ? 0x02000001 : 0x62000001); - nv_wr32(dev, 0x408908, 0x00c80929); + if (chipset == 0xd9) { + nv_wr32(dev, 0x408808, 0x1043e005); + nv_wr32(dev, 0x408900, 0x3080b801); + nv_wr32(dev, 0x408904, 0x1043e005); + nv_wr32(dev, 0x408908, 0x00c8102f); + } else + if (chipset == 0xc1) { + nv_wr32(dev, 0x408808, 0x1003e005); + nv_wr32(dev, 0x408900, 0x3080b801); + nv_wr32(dev, 0x408904, 0x62000001); + nv_wr32(dev, 0x408908, 0x00c80929); + } else { + nv_wr32(dev, 0x408808, 0x0003e00d); + nv_wr32(dev, 0x408900, 0x3080b801); + nv_wr32(dev, 0x408904, 0x02000001); + nv_wr32(dev, 0x408908, 0x00c80929); + } nv_wr32(dev, 0x40890c, 0x00000000); nv_wr32(dev, 0x408980, 0x0000011d); } @@ -1572,7 +1611,7 @@ nvc0_grctx_generate_gpc(struct drm_device *dev) nv_wr32(dev, 0x418408, 0x00000000); nv_wr32(dev, 0x41840c, 0x00001008); nv_wr32(dev, 0x418410, 0x0fff0fff); - nv_wr32(dev, 0x418414, 0x00200fff); + nv_wr32(dev, 0x418414, chipset != 0xd9 ? 0x00200fff : 0x02200fff); nv_wr32(dev, 0x418450, 0x00000000); nv_wr32(dev, 0x418454, 0x00000000); nv_wr32(dev, 0x418458, 0x00000000); @@ -1587,14 +1626,17 @@ nvc0_grctx_generate_gpc(struct drm_device *dev) nv_wr32(dev, 0x418700, 0x00000002); nv_wr32(dev, 0x418704, 0x00000080); nv_wr32(dev, 0x418708, 0x00000000); - nv_wr32(dev, 0x41870c, 0x07c80000); + nv_wr32(dev, 0x41870c, chipset != 0xd9 ? 0x07c80000 : 0x00000000); nv_wr32(dev, 0x418710, 0x00000000); - nv_wr32(dev, 0x418800, 0x0006860a); + nv_wr32(dev, 0x418800, chipset != 0xd9 ? 0x0006860a : 0x7006860a); nv_wr32(dev, 0x418808, 0x00000000); nv_wr32(dev, 0x41880c, 0x00000000); nv_wr32(dev, 0x418810, 0x00000000); nv_wr32(dev, 0x418828, 0x00008442); - nv_wr32(dev, 0x418830, chipset != 0xc1 ? 0x00000001 : 0x10000001); + if (chipset == 0xc1 || chipset == 0xd9) + nv_wr32(dev, 0x418830, 0x10000001); + else + nv_wr32(dev, 0x418830, 0x00000001); nv_wr32(dev, 0x4188d8, 0x00000008); nv_wr32(dev, 0x4188e0, 0x01000000); nv_wr32(dev, 0x4188e8, 0x00000000); @@ -1602,7 +1644,12 @@ nvc0_grctx_generate_gpc(struct drm_device *dev) nv_wr32(dev, 0x4188f0, 0x00000000); nv_wr32(dev, 0x4188f4, 0x00000000); nv_wr32(dev, 0x4188f8, 0x00000000); - nv_wr32(dev, 0x4188fc, chipset != 0xc1 ? 0x00100000 : 0x00100018); + if (chipset == 0xd9) + nv_wr32(dev, 0x4188fc, 0x20100008); + else if (chipset == 0xc1) + nv_wr32(dev, 0x4188fc, 0x00100018); + else + nv_wr32(dev, 0x4188fc, 0x00100000); nv_wr32(dev, 0x41891c, 0x00ff00ff); nv_wr32(dev, 0x418924, 0x00000000); nv_wr32(dev, 0x418928, 0x00ffff00); @@ -1616,7 +1663,7 @@ nvc0_grctx_generate_gpc(struct drm_device *dev) nv_wr32(dev, 0x418a14 + (i * 0x20), 0x00000000); nv_wr32(dev, 0x418a18 + (i * 0x20), 0x00000000); } - nv_wr32(dev, 0x418b00, 0x00000000); + nv_wr32(dev, 0x418b00, chipset != 0xd9 ? 0x00000000 : 0x00000006); nv_wr32(dev, 0x418b08, 0x0a418820); nv_wr32(dev, 0x418b0c, 0x062080e6); nv_wr32(dev, 0x418b10, 0x020398a4); @@ -1633,7 +1680,7 @@ nvc0_grctx_generate_gpc(struct drm_device *dev) nv_wr32(dev, 0x418c24, 0x00000000); nv_wr32(dev, 0x418c28, 0x00000000); nv_wr32(dev, 0x418c2c, 0x00000000); - if (chipset == 0xc1) + if (chipset == 0xc1 || chipset == 0xd9) nv_wr32(dev, 0x418c6c, 0x00000001); nv_wr32(dev, 0x418c80, 0x20200004); nv_wr32(dev, 0x418c8c, 0x00000001); @@ -1653,7 +1700,10 @@ nvc0_grctx_generate_tp(struct drm_device *dev) nv_wr32(dev, 0x419818, 0x00000000); nv_wr32(dev, 0x41983c, 0x00038bc7); nv_wr32(dev, 0x419848, 0x00000000); - nv_wr32(dev, 0x419864, chipset != 0xc1 ? 0x0000012a : 0x00000129); + if (chipset == 0xc1 || chipset == 0xd9) + nv_wr32(dev, 0x419864, 0x00000129); + else + nv_wr32(dev, 0x419864, 0x0000012a); nv_wr32(dev, 0x419888, 0x00000000); nv_wr32(dev, 0x419a00, 0x000001f0); nv_wr32(dev, 0x419a04, 0x00000001); @@ -1663,7 +1713,9 @@ nvc0_grctx_generate_tp(struct drm_device *dev) nv_wr32(dev, 0x419a14, 0x00000200); nv_wr32(dev, 0x419a1c, 0x00000000); nv_wr32(dev, 0x419a20, 0x00000800); - if (chipset != 0xc0 && chipset != 0xc8) + if (chipset == 0xd9) + nv_wr32(dev, 0x00419ac4, 0x0017f440); + else if (chipset != 0xc0 && chipset != 0xc8) nv_wr32(dev, 0x00419ac4, 0x0007f440); nv_wr32(dev, 0x419b00, 0x0a418820); nv_wr32(dev, 0x419b04, 0x062080e6); @@ -1672,21 +1724,33 @@ nvc0_grctx_generate_tp(struct drm_device *dev) nv_wr32(dev, 0x419b10, 0x0a418820); nv_wr32(dev, 0x419b14, 0x000000e6); nv_wr32(dev, 0x419bd0, 0x00900103); - nv_wr32(dev, 0x419be0, chipset != 0xc1 ? 0x00000001 : 0x00400001); + if (chipset == 0xc1 || chipset == 0xd9) + nv_wr32(dev, 0x419be0, 0x00400001); + else + nv_wr32(dev, 0x419be0, 0x00000001); nv_wr32(dev, 0x419be4, 0x00000000); - nv_wr32(dev, 0x419c00, 0x00000002); + nv_wr32(dev, 0x419c00, chipset != 0xd9 ? 0x00000002 : 0x0000000a); nv_wr32(dev, 0x419c04, 0x00000006); nv_wr32(dev, 0x419c08, 0x00000002); nv_wr32(dev, 0x419c20, 0x00000000); - if (chipset == 0xce || chipset == 0xcf) + if (dev_priv->chipset == 0xd9) { + nv_wr32(dev, 0x419c24, 0x00084210); + nv_wr32(dev, 0x419c28, 0x3cf3cf3c); nv_wr32(dev, 0x419cb0, 0x00020048); - else + } else + if (chipset == 0xce || chipset == 0xcf) { + nv_wr32(dev, 0x419cb0, 0x00020048); + } else { nv_wr32(dev, 0x419cb0, 0x00060048); + } nv_wr32(dev, 0x419ce8, 0x00000000); nv_wr32(dev, 0x419cf4, 0x00000183); - nv_wr32(dev, 0x419d20, chipset != 0xc1 ? 0x02180000 : 0x12180000); + if (chipset == 0xc1 || chipset == 0xd9) + nv_wr32(dev, 0x419d20, 0x12180000); + else + nv_wr32(dev, 0x419d20, 0x02180000); nv_wr32(dev, 0x419d24, 0x00001fff); - if (chipset == 0xc1) + if (chipset == 0xc1 || chipset == 0xd9) nv_wr32(dev, 0x419d44, 0x02180218); nv_wr32(dev, 0x419e04, 0x00000000); nv_wr32(dev, 0x419e08, 0x00000000); @@ -1986,6 +2050,10 @@ nvc0_grctx_generate(struct nouveau_channel *chan) nv_icmd(dev, 0x00000215, 0x00000040); nv_icmd(dev, 0x00000216, 0x00000040); nv_icmd(dev, 0x00000217, 0x00000040); + if (dev_priv->chipset == 0xd9) { + for (i = 0x0400; i <= 0x0417; i++) + nv_icmd(dev, i, 0x00000040); + } nv_icmd(dev, 0x00000218, 0x0000c080); nv_icmd(dev, 0x00000219, 0x0000c080); nv_icmd(dev, 0x0000021a, 0x0000c080); @@ -1994,6 +2062,10 @@ nvc0_grctx_generate(struct nouveau_channel *chan) nv_icmd(dev, 0x0000021d, 0x0000c080); nv_icmd(dev, 0x0000021e, 0x0000c080); nv_icmd(dev, 0x0000021f, 0x0000c080); + if (dev_priv->chipset == 0xd9) { + for (i = 0x0440; i <= 0x0457; i++) + nv_icmd(dev, i, 0x0000c080); + } nv_icmd(dev, 0x000000ad, 0x0000013e); nv_icmd(dev, 0x000000e1, 0x00000010); nv_icmd(dev, 0x00000290, 0x00000000); @@ -2556,7 +2628,8 @@ nvc0_grctx_generate(struct nouveau_channel *chan) nv_icmd(dev, 0x0000053f, 0xffff0000); nv_icmd(dev, 0x00000585, 0x0000003f); nv_icmd(dev, 0x00000576, 0x00000003); - if (dev_priv->chipset == 0xc1) + if (dev_priv->chipset == 0xc1 || + dev_priv->chipset == 0xd9) nv_icmd(dev, 0x0000057b, 0x00000059); nv_icmd(dev, 0x00000586, 0x00000040); nv_icmd(dev, 0x00000582, 0x00000080); @@ -2658,6 +2731,8 @@ nvc0_grctx_generate(struct nouveau_channel *chan) nv_icmd(dev, 0x00000957, 0x00000003); nv_icmd(dev, 0x0000095e, 0x20164010); nv_icmd(dev, 0x0000095f, 0x00000020); + if (dev_priv->chipset == 0xd9) + nv_icmd(dev, 0x0000097d, 0x00000020); nv_icmd(dev, 0x00000683, 0x00000006); nv_icmd(dev, 0x00000685, 0x003fffff); nv_icmd(dev, 0x00000687, 0x00000c48); diff --git a/drivers/gpu/drm/nouveau/nvc0_grgpc.fuc b/drivers/gpu/drm/nouveau/nvc0_grgpc.fuc index 06f5e26..15272be 100644 --- a/drivers/gpu/drm/nouveau/nvc0_grgpc.fuc +++ b/drivers/gpu/drm/nouveau/nvc0_grgpc.fuc @@ -32,7 +32,7 @@ * - watchdog timer around ctx operations */ -.section nvc0_grgpc_data +.section #nvc0_grgpc_data include(`nvc0_graph.fuc') gpc_id: .b32 0 gpc_mmio_list_head: .b32 0 @@ -48,40 +48,45 @@ cmd_queue: queue_init // chipset descriptions chipsets: .b8 0xc0 0 0 0 -.b16 nvc0_gpc_mmio_head -.b16 nvc0_gpc_mmio_tail -.b16 nvc0_tpc_mmio_head -.b16 nvc0_tpc_mmio_tail +.b16 #nvc0_gpc_mmio_head +.b16 #nvc0_gpc_mmio_tail +.b16 #nvc0_tpc_mmio_head +.b16 #nvc0_tpc_mmio_tail .b8 0xc1 0 0 0 -.b16 nvc0_gpc_mmio_head -.b16 nvc1_gpc_mmio_tail -.b16 nvc0_tpc_mmio_head -.b16 nvc1_tpc_mmio_tail +.b16 #nvc0_gpc_mmio_head +.b16 #nvc1_gpc_mmio_tail +.b16 #nvc0_tpc_mmio_head +.b16 #nvc1_tpc_mmio_tail .b8 0xc3 0 0 0 -.b16 nvc0_gpc_mmio_head -.b16 nvc0_gpc_mmio_tail -.b16 nvc0_tpc_mmio_head -.b16 nvc3_tpc_mmio_tail +.b16 #nvc0_gpc_mmio_head +.b16 #nvc0_gpc_mmio_tail +.b16 #nvc0_tpc_mmio_head +.b16 #nvc3_tpc_mmio_tail .b8 0xc4 0 0 0 -.b16 nvc0_gpc_mmio_head -.b16 nvc0_gpc_mmio_tail -.b16 nvc0_tpc_mmio_head -.b16 nvc3_tpc_mmio_tail +.b16 #nvc0_gpc_mmio_head +.b16 #nvc0_gpc_mmio_tail +.b16 #nvc0_tpc_mmio_head +.b16 #nvc3_tpc_mmio_tail .b8 0xc8 0 0 0 -.b16 nvc0_gpc_mmio_head -.b16 nvc0_gpc_mmio_tail -.b16 nvc0_tpc_mmio_head -.b16 nvc0_tpc_mmio_tail +.b16 #nvc0_gpc_mmio_head +.b16 #nvc0_gpc_mmio_tail +.b16 #nvc0_tpc_mmio_head +.b16 #nvc0_tpc_mmio_tail .b8 0xce 0 0 0 -.b16 nvc0_gpc_mmio_head -.b16 nvc0_gpc_mmio_tail -.b16 nvc0_tpc_mmio_head -.b16 nvc3_tpc_mmio_tail +.b16 #nvc0_gpc_mmio_head +.b16 #nvc0_gpc_mmio_tail +.b16 #nvc0_tpc_mmio_head +.b16 #nvc3_tpc_mmio_tail .b8 0xcf 0 0 0 -.b16 nvc0_gpc_mmio_head -.b16 nvc0_gpc_mmio_tail -.b16 nvc0_tpc_mmio_head -.b16 nvcf_tpc_mmio_tail +.b16 #nvc0_gpc_mmio_head +.b16 #nvc0_gpc_mmio_tail +.b16 #nvc0_tpc_mmio_head +.b16 #nvcf_tpc_mmio_tail +.b8 0xd9 0 0 0 +.b16 #nvd9_gpc_mmio_head +.b16 #nvd9_gpc_mmio_tail +.b16 #nvd9_tpc_mmio_head +.b16 #nvd9_tpc_mmio_tail .b8 0 0 0 0 // GPC mmio lists @@ -114,6 +119,35 @@ nvc0_gpc_mmio_tail: mmctx_data(0x000c6c, 1); nvc1_gpc_mmio_tail: +nvd9_gpc_mmio_head: +mmctx_data(0x000380, 1) +mmctx_data(0x000400, 2) +mmctx_data(0x00040c, 3) +mmctx_data(0x000450, 9) +mmctx_data(0x000600, 1) +mmctx_data(0x000684, 1) +mmctx_data(0x000700, 5) +mmctx_data(0x000800, 1) +mmctx_data(0x000808, 3) +mmctx_data(0x000828, 1) +mmctx_data(0x000830, 1) +mmctx_data(0x0008d8, 1) +mmctx_data(0x0008e0, 1) +mmctx_data(0x0008e8, 6) +mmctx_data(0x00091c, 1) +mmctx_data(0x000924, 3) +mmctx_data(0x000b00, 1) +mmctx_data(0x000b08, 6) +mmctx_data(0x000bb8, 1) +mmctx_data(0x000c08, 1) +mmctx_data(0x000c10, 8) +mmctx_data(0x000c6c, 1) +mmctx_data(0x000c80, 1) +mmctx_data(0x000c8c, 1) +mmctx_data(0x001000, 3) +mmctx_data(0x001014, 1) +nvd9_gpc_mmio_tail: + // TPC mmio lists nvc0_tpc_mmio_head: mmctx_data(0x000018, 1) @@ -146,9 +180,34 @@ nvc3_tpc_mmio_tail: mmctx_data(0x000544, 1) nvc1_tpc_mmio_tail: +nvd9_tpc_mmio_head: +mmctx_data(0x000018, 1) +mmctx_data(0x00003c, 1) +mmctx_data(0x000048, 1) +mmctx_data(0x000064, 1) +mmctx_data(0x000088, 1) +mmctx_data(0x000200, 6) +mmctx_data(0x00021c, 2) +mmctx_data(0x0002c4, 1) +mmctx_data(0x000300, 6) +mmctx_data(0x0003d0, 1) +mmctx_data(0x0003e0, 2) +mmctx_data(0x000400, 3) +mmctx_data(0x000420, 3) +mmctx_data(0x0004b0, 1) +mmctx_data(0x0004e8, 1) +mmctx_data(0x0004f4, 1) +mmctx_data(0x000520, 2) +mmctx_data(0x000544, 1) +mmctx_data(0x000604, 4) +mmctx_data(0x000644, 20) +mmctx_data(0x000698, 1) +mmctx_data(0x0006e0, 1) +mmctx_data(0x000750, 3) +nvd9_tpc_mmio_tail: -.section nvc0_grgpc_code -bra init +.section #nvc0_grgpc_code +bra #init define(`include_code') include(`nvc0_graph.fuc') @@ -160,10 +219,10 @@ error: push $r14 mov $r14 -0x67ec // 0x9814 sethi $r14 0x400000 - call nv_wr32 // HUB_CTXCTL_CC_SCRATCH[5] = error code + call #nv_wr32 // HUB_CTXCTL_CC_SCRATCH[5] = error code add b32 $r14 0x41c mov $r15 1 - call nv_wr32 // HUB_CTXCTL_INTR_UP_SET + call #nv_wr32 // HUB_CTXCTL_INTR_UP_SET pop $r14 ret @@ -190,7 +249,7 @@ init: iowr I[$r1 + 0x000] $r2 // FIFO_ENABLE // setup i0 handler, and route all interrupts to it - mov $r1 ih + mov $r1 #ih mov $iv0 $r1 mov $r1 0x400 iowr I[$r1 + 0x300] $r0 // INTR_DISPATCH @@ -210,24 +269,24 @@ init: and $r2 0x1f shl b32 $r3 $r2 sub b32 $r3 1 - st b32 D[$r0 + tpc_count] $r2 - st b32 D[$r0 + tpc_mask] $r3 + st b32 D[$r0 + #tpc_count] $r2 + st b32 D[$r0 + #tpc_mask] $r3 add b32 $r1 0x400 iord $r2 I[$r1 + 0x000] // MYINDEX - st b32 D[$r0 + gpc_id] $r2 + st b32 D[$r0 + #gpc_id] $r2 // find context data for this chipset mov $r2 0x800 shl b32 $r2 6 iord $r2 I[$r2 + 0x000] // CC_SCRATCH[0] - mov $r1 chipsets - 12 + mov $r1 #chipsets - 12 init_find_chipset: add b32 $r1 12 ld b32 $r3 D[$r1 + 0x00] cmpu b32 $r3 $r2 - bra e init_context + bra e #init_context cmpu b32 $r3 0 - bra ne init_find_chipset + bra ne #init_find_chipset // unknown chipset ret @@ -253,19 +312,19 @@ init: clear b32 $r15 ld b16 $r14 D[$r1 + 4] ld b16 $r15 D[$r1 + 6] - st b16 D[$r0 + gpc_mmio_list_head] $r14 - st b16 D[$r0 + gpc_mmio_list_tail] $r15 - call mmctx_size + st b16 D[$r0 + #gpc_mmio_list_head] $r14 + st b16 D[$r0 + #gpc_mmio_list_tail] $r15 + call #mmctx_size add b32 $r2 $r15 add b32 $r3 $r15 // calculate per-TPC mmio context size, store the list pointers ld b16 $r14 D[$r1 + 8] ld b16 $r15 D[$r1 + 10] - st b16 D[$r0 + tpc_mmio_list_head] $r14 - st b16 D[$r0 + tpc_mmio_list_tail] $r15 - call mmctx_size - ld b32 $r14 D[$r0 + tpc_count] + st b16 D[$r0 + #tpc_mmio_list_head] $r14 + st b16 D[$r0 + #tpc_mmio_list_tail] $r15 + call #mmctx_size + ld b32 $r14 D[$r0 + #tpc_count] mulu $r14 $r15 add b32 $r2 $r14 add b32 $r3 $r14 @@ -283,7 +342,7 @@ init: // calculate size of strand context data mov b32 $r15 $r2 - call strand_ctx_init + call #strand_ctx_init add b32 $r3 $r15 // save context size, and tell HUB we're done @@ -301,13 +360,13 @@ init: main: bset $flags $p0 sleep $p0 - mov $r13 cmd_queue - call queue_get - bra $p1 main + mov $r13 #cmd_queue + call #queue_get + bra $p1 #main // 0x0000-0x0003 are all context transfers cmpu b32 $r14 0x04 - bra nc main_not_ctx_xfer + bra nc #main_not_ctx_xfer // fetch $flags and mask off $p1/$p2 mov $r1 $flags mov $r2 0x0006 @@ -318,14 +377,14 @@ main: or $r1 $r14 mov $flags $r1 // transfer context data - call ctx_xfer - bra main + call #ctx_xfer + bra #main main_not_ctx_xfer: shl b32 $r15 $r14 16 or $r15 E_BAD_COMMAND - call error - bra main + call #error + bra #main // interrupt handler ih: @@ -342,13 +401,13 @@ ih: // incoming fifo command? iord $r10 I[$r0 + 0x200] // INTR and $r11 $r10 0x00000004 - bra e ih_no_fifo + bra e #ih_no_fifo // queue incoming fifo command for later processing mov $r11 0x1900 - mov $r13 cmd_queue + mov $r13 #cmd_queue iord $r14 I[$r11 + 0x100] // FIFO_CMD iord $r15 I[$r11 + 0x000] // FIFO_DATA - call queue_put + call #queue_put add b32 $r11 0x400 mov $r14 1 iowr I[$r11 + 0x000] $r14 // FIFO_ACK @@ -374,11 +433,11 @@ ih: // hub_barrier_done: mov $r15 1 - ld b32 $r14 D[$r0 + gpc_id] + ld b32 $r14 D[$r0 + #gpc_id] shl b32 $r15 $r14 mov $r14 -0x6be8 // 0x409418 - HUB_BAR_SET sethi $r14 0x400000 - call nv_wr32 + call #nv_wr32 ret // Disables various things, waits a bit, and re-enables them.. @@ -395,7 +454,7 @@ ctx_redswitch: mov $r15 8 ctx_redswitch_delay: sub b32 $r15 1 - bra ne ctx_redswitch_delay + bra ne #ctx_redswitch_delay mov $r15 0xa20 iowr I[$r14] $r15 // GPC_RED_SWITCH = UNK11, ENABLE, POWER ret @@ -413,8 +472,8 @@ ctx_xfer: mov $r1 0xa04 shl b32 $r1 6 iowr I[$r1 + 0x000] $r15// MEM_BASE - bra not $p1 ctx_xfer_not_load - call ctx_redswitch + bra not $p1 #ctx_xfer_not_load + call #ctx_redswitch ctx_xfer_not_load: // strands @@ -422,7 +481,7 @@ ctx_xfer: sethi $r1 0x20000 mov $r2 0xc iowr I[$r1] $r2 // STRAND_CMD(0x3f) = 0x0c - call strand_wait + call #strand_wait mov $r2 0x47fc sethi $r2 0x20000 iowr I[$r2] $r0 // STRAND_FIRST_GENE(0x3f) = 0x00 @@ -435,46 +494,46 @@ ctx_xfer: or $r10 2 // first mov $r11 0x0000 sethi $r11 0x500000 - ld b32 $r12 D[$r0 + gpc_id] + ld b32 $r12 D[$r0 + #gpc_id] shl b32 $r12 15 add b32 $r11 $r12 // base = NV_PGRAPH_GPCn - ld b32 $r12 D[$r0 + gpc_mmio_list_head] - ld b32 $r13 D[$r0 + gpc_mmio_list_tail] + ld b32 $r12 D[$r0 + #gpc_mmio_list_head] + ld b32 $r13 D[$r0 + #gpc_mmio_list_tail] mov $r14 0 // not multi - call mmctx_xfer + call #mmctx_xfer // per-TPC mmio context xbit $r10 $flags $p1 // direction or $r10 4 // last mov $r11 0x4000 sethi $r11 0x500000 // base = NV_PGRAPH_GPC0_TPC0 - ld b32 $r12 D[$r0 + gpc_id] + ld b32 $r12 D[$r0 + #gpc_id] shl b32 $r12 15 add b32 $r11 $r12 // base = NV_PGRAPH_GPCn_TPC0 - ld b32 $r12 D[$r0 + tpc_mmio_list_head] - ld b32 $r13 D[$r0 + tpc_mmio_list_tail] - ld b32 $r15 D[$r0 + tpc_mask] + ld b32 $r12 D[$r0 + #tpc_mmio_list_head] + ld b32 $r13 D[$r0 + #tpc_mmio_list_tail] + ld b32 $r15 D[$r0 + #tpc_mask] mov $r14 0x800 // stride = 0x800 - call mmctx_xfer + call #mmctx_xfer // wait for strands to finish - call strand_wait + call #strand_wait // if load, or a save without a load following, do some // unknown stuff that's done after finishing a block of // strand commands - bra $p1 ctx_xfer_post - bra not $p2 ctx_xfer_done + bra $p1 #ctx_xfer_post + bra not $p2 #ctx_xfer_done ctx_xfer_post: mov $r1 0x4afc sethi $r1 0x20000 mov $r2 0xd iowr I[$r1] $r2 // STRAND_CMD(0x3f) = 0x0d - call strand_wait + call #strand_wait // mark completion in HUB's barrier ctx_xfer_done: - call hub_barrier_done + call #hub_barrier_done ret .align 256 diff --git a/drivers/gpu/drm/nouveau/nvc0_grgpc.fuc.h b/drivers/gpu/drm/nouveau/nvc0_grgpc.fuc.h index 6f82032..a988b8a 100644 --- a/drivers/gpu/drm/nouveau/nvc0_grgpc.fuc.h +++ b/drivers/gpu/drm/nouveau/nvc0_grgpc.fuc.h @@ -25,26 +25,29 @@ uint32_t nvc0_grgpc_data[] = { 0x00000000, 0x00000000, 0x000000c0, - 0x011c00bc, - 0x01700120, + 0x012800c8, + 0x01e40194, 0x000000c1, - 0x012000bc, - 0x01840120, + 0x012c00c8, + 0x01f80194, 0x000000c3, - 0x011c00bc, - 0x01800120, + 0x012800c8, + 0x01f40194, 0x000000c4, - 0x011c00bc, - 0x01800120, + 0x012800c8, + 0x01f40194, 0x000000c8, - 0x011c00bc, - 0x01700120, + 0x012800c8, + 0x01e40194, 0x000000ce, - 0x011c00bc, - 0x01800120, + 0x012800c8, + 0x01f40194, 0x000000cf, - 0x011c00bc, - 0x017c0120, + 0x012800c8, + 0x01f00194, + 0x000000d9, + 0x0194012c, + 0x025401f8, 0x00000000, 0x00000380, 0x14000400, @@ -71,6 +74,32 @@ uint32_t nvc0_grgpc_data[] = { 0x08001000, 0x00001014, 0x00000c6c, + 0x00000380, + 0x04000400, + 0x0800040c, + 0x20000450, + 0x00000600, + 0x00000684, + 0x10000700, + 0x00000800, + 0x08000808, + 0x00000828, + 0x00000830, + 0x000008d8, + 0x000008e0, + 0x140008e8, + 0x0000091c, + 0x08000924, + 0x00000b00, + 0x14000b08, + 0x00000bb8, + 0x00000c08, + 0x1c000c10, + 0x00000c6c, + 0x00000c80, + 0x00000c8c, + 0x08001000, + 0x00001014, 0x00000018, 0x0000003c, 0x00000048, @@ -96,6 +125,29 @@ uint32_t nvc0_grgpc_data[] = { 0x000006e0, 0x000004bc, 0x00000544, + 0x00000018, + 0x0000003c, + 0x00000048, + 0x00000064, + 0x00000088, + 0x14000200, + 0x0400021c, + 0x000002c4, + 0x14000300, + 0x000003d0, + 0x040003e0, + 0x08000400, + 0x08000420, + 0x000004b0, + 0x000004e8, + 0x000004f4, + 0x04000520, + 0x00000544, + 0x0c000604, + 0x4c000644, + 0x00000698, + 0x000006e0, + 0x08000750, }; uint32_t nvc0_grgpc_code[] = { diff --git a/drivers/gpu/drm/nouveau/nvc0_grhub.fuc b/drivers/gpu/drm/nouveau/nvc0_grhub.fuc index e4f8c7e..98acddb 100644 --- a/drivers/gpu/drm/nouveau/nvc0_grhub.fuc +++ b/drivers/gpu/drm/nouveau/nvc0_grhub.fuc @@ -27,7 +27,7 @@ * m4 nvc0_grhub.fuc | envyas -a -w -m fuc -V nva3 -o nvc0_grhub.fuc.h */ -.section nvc0_grhub_data +.section #nvc0_grhub_data include(`nvc0_graph.fuc') gpc_count: .b32 0 rop_count: .b32 0 @@ -39,26 +39,29 @@ ctx_current: .b32 0 chipsets: .b8 0xc0 0 0 0 -.b16 nvc0_hub_mmio_head -.b16 nvc0_hub_mmio_tail +.b16 #nvc0_hub_mmio_head +.b16 #nvc0_hub_mmio_tail .b8 0xc1 0 0 0 -.b16 nvc0_hub_mmio_head -.b16 nvc1_hub_mmio_tail +.b16 #nvc0_hub_mmio_head +.b16 #nvc1_hub_mmio_tail .b8 0xc3 0 0 0 -.b16 nvc0_hub_mmio_head -.b16 nvc0_hub_mmio_tail +.b16 #nvc0_hub_mmio_head +.b16 #nvc0_hub_mmio_tail .b8 0xc4 0 0 0 -.b16 nvc0_hub_mmio_head -.b16 nvc0_hub_mmio_tail +.b16 #nvc0_hub_mmio_head +.b16 #nvc0_hub_mmio_tail .b8 0xc8 0 0 0 -.b16 nvc0_hub_mmio_head -.b16 nvc0_hub_mmio_tail +.b16 #nvc0_hub_mmio_head +.b16 #nvc0_hub_mmio_tail .b8 0xce 0 0 0 -.b16 nvc0_hub_mmio_head -.b16 nvc0_hub_mmio_tail +.b16 #nvc0_hub_mmio_head +.b16 #nvc0_hub_mmio_tail .b8 0xcf 0 0 0 -.b16 nvc0_hub_mmio_head -.b16 nvc0_hub_mmio_tail +.b16 #nvc0_hub_mmio_head +.b16 #nvc0_hub_mmio_tail +.b8 0xd9 0 0 0 +.b16 #nvd9_hub_mmio_head +.b16 #nvd9_hub_mmio_tail .b8 0 0 0 0 nvc0_hub_mmio_head: @@ -105,6 +108,48 @@ nvc0_hub_mmio_tail: mmctx_data(0x4064c0, 2) nvc1_hub_mmio_tail: +nvd9_hub_mmio_head: +mmctx_data(0x17e91c, 2) +mmctx_data(0x400204, 2) +mmctx_data(0x404004, 10) +mmctx_data(0x404044, 1) +mmctx_data(0x404094, 14) +mmctx_data(0x4040d0, 7) +mmctx_data(0x4040f8, 1) +mmctx_data(0x404130, 3) +mmctx_data(0x404150, 3) +mmctx_data(0x404164, 2) +mmctx_data(0x404178, 2) +mmctx_data(0x404200, 8) +mmctx_data(0x404404, 14) +mmctx_data(0x404460, 4) +mmctx_data(0x404480, 1) +mmctx_data(0x404498, 1) +mmctx_data(0x404604, 4) +mmctx_data(0x404618, 32) +mmctx_data(0x404698, 21) +mmctx_data(0x4046f0, 2) +mmctx_data(0x404700, 22) +mmctx_data(0x405800, 1) +mmctx_data(0x405830, 3) +mmctx_data(0x405854, 1) +mmctx_data(0x405870, 4) +mmctx_data(0x405a00, 2) +mmctx_data(0x405a18, 1) +mmctx_data(0x406020, 1) +mmctx_data(0x406028, 4) +mmctx_data(0x4064a8, 2) +mmctx_data(0x4064b4, 5) +mmctx_data(0x407804, 1) +mmctx_data(0x40780c, 6) +mmctx_data(0x4078bc, 1) +mmctx_data(0x408000, 7) +mmctx_data(0x408064, 1) +mmctx_data(0x408800, 3) +mmctx_data(0x408900, 4) +mmctx_data(0x408980, 1) +nvd9_hub_mmio_tail: + .align 256 chan_data: chan_mmio_count: .b32 0 @@ -113,8 +158,8 @@ chan_mmio_address: .b32 0 .align 256 xfer_data: .b32 0 -.section nvc0_grhub_code -bra init +.section #nvc0_grhub_code +bra #init define(`include_code') include(`nvc0_graph.fuc') @@ -157,7 +202,7 @@ init: iowr I[$r1 + 0x000] $r2 // FIFO_ENABLE // setup i0 handler, and route all interrupts to it - mov $r1 ih + mov $r1 #ih mov $iv0 $r1 mov $r1 0x400 iowr I[$r1 + 0x300] $r0 // INTR_DISPATCH @@ -201,11 +246,11 @@ init: // fetch enabled GPC/ROP counts mov $r14 -0x69fc // 0x409604 sethi $r14 0x400000 - call nv_rd32 + call #nv_rd32 extr $r1 $r15 16:20 - st b32 D[$r0 + rop_count] $r1 + st b32 D[$r0 + #rop_count] $r1 and $r15 0x1f - st b32 D[$r0 + gpc_count] $r15 + st b32 D[$r0 + #gpc_count] $r15 // set BAR_REQMASK to GPC mask mov $r1 1 @@ -220,14 +265,14 @@ init: mov $r2 0x800 shl b32 $r2 6 iord $r2 I[$r2 + 0x000] // CC_SCRATCH[0] - mov $r15 chipsets - 8 + mov $r15 #chipsets - 8 init_find_chipset: add b32 $r15 8 ld b32 $r3 D[$r15 + 0x00] cmpu b32 $r3 $r2 - bra e init_context + bra e #init_context cmpu b32 $r3 0 - bra ne init_find_chipset + bra ne #init_find_chipset // unknown chipset ret @@ -239,9 +284,9 @@ init: ld b16 $r14 D[$r15 + 4] ld b16 $r15 D[$r15 + 6] sethi $r14 0 - st b32 D[$r0 + hub_mmio_list_head] $r14 - st b32 D[$r0 + hub_mmio_list_tail] $r15 - call mmctx_size + st b32 D[$r0 + #hub_mmio_list_head] $r14 + st b32 D[$r0 + #hub_mmio_list_tail] $r15 + call #mmctx_size // set mmctx base addresses now so we don't have to do it later, // they don't (currently) ever change @@ -260,7 +305,7 @@ init: add b32 $r1 1 shl b32 $r1 8 mov b32 $r15 $r1 - call strand_ctx_init + call #strand_ctx_init add b32 $r1 $r15 // initialise each GPC in sequence by passing in the offset of its @@ -271,40 +316,40 @@ init: // when it has completed, and return the size of its context data // in GPCn_CC_SCRATCH[1] // - ld b32 $r3 D[$r0 + gpc_count] + ld b32 $r3 D[$r0 + #gpc_count] mov $r4 0x2000 sethi $r4 0x500000 init_gpc: // setup, and start GPC ucode running add b32 $r14 $r4 0x804 mov b32 $r15 $r1 - call nv_wr32 // CC_SCRATCH[1] = ctx offset + call #nv_wr32 // CC_SCRATCH[1] = ctx offset add b32 $r14 $r4 0x800 mov b32 $r15 $r2 - call nv_wr32 // CC_SCRATCH[0] = chipset + call #nv_wr32 // CC_SCRATCH[0] = chipset add b32 $r14 $r4 0x10c clear b32 $r15 - call nv_wr32 + call #nv_wr32 add b32 $r14 $r4 0x104 - call nv_wr32 // ENTRY + call #nv_wr32 // ENTRY add b32 $r14 $r4 0x100 mov $r15 2 // CTRL_START_TRIGGER - call nv_wr32 // CTRL + call #nv_wr32 // CTRL // wait for it to complete, and adjust context size add b32 $r14 $r4 0x800 init_gpc_wait: - call nv_rd32 + call #nv_rd32 xbit $r15 $r15 31 - bra e init_gpc_wait + bra e #init_gpc_wait add b32 $r14 $r4 0x804 - call nv_rd32 + call #nv_rd32 add b32 $r1 $r15 // next! add b32 $r4 0x8000 sub b32 $r3 1 - bra ne init_gpc + bra ne #init_gpc // save context size, and tell host we're ready mov $r2 0x800 @@ -322,13 +367,13 @@ main: // sleep until we have something to do bset $flags $p0 sleep $p0 - mov $r13 cmd_queue - call queue_get - bra $p1 main + mov $r13 #cmd_queue + call #queue_get + bra $p1 #main // context switch, requested by GPU? cmpu b32 $r14 0x4001 - bra ne main_not_ctx_switch + bra ne #main_not_ctx_switch trace_set(T_AUTO) mov $r1 0xb00 shl b32 $r1 6 @@ -336,39 +381,39 @@ main: iord $r1 I[$r1 + 0x000] // CHAN_CUR xbit $r3 $r1 31 - bra e chsw_no_prev + bra e #chsw_no_prev xbit $r3 $r2 31 - bra e chsw_prev_no_next + bra e #chsw_prev_no_next push $r2 mov b32 $r2 $r1 trace_set(T_SAVE) bclr $flags $p1 bset $flags $p2 - call ctx_xfer + call #ctx_xfer trace_clr(T_SAVE); pop $r2 trace_set(T_LOAD); bset $flags $p1 - call ctx_xfer + call #ctx_xfer trace_clr(T_LOAD); - bra chsw_done + bra #chsw_done chsw_prev_no_next: push $r2 mov b32 $r2 $r1 bclr $flags $p1 bclr $flags $p2 - call ctx_xfer + call #ctx_xfer pop $r2 mov $r1 0xb00 shl b32 $r1 6 iowr I[$r1] $r2 - bra chsw_done + bra #chsw_done chsw_no_prev: xbit $r3 $r2 31 - bra e chsw_done + bra e #chsw_done bset $flags $p1 bclr $flags $p2 - call ctx_xfer + call #ctx_xfer // ack the context switch request chsw_done: @@ -377,32 +422,32 @@ main: mov $r2 1 iowr I[$r1 + 0x000] $r2 // 0x409b0c trace_clr(T_AUTO) - bra main + bra #main // request to set current channel? (*not* a context switch) main_not_ctx_switch: cmpu b32 $r14 0x0001 - bra ne main_not_ctx_chan + bra ne #main_not_ctx_chan mov b32 $r2 $r15 - call ctx_chan - bra main_done + call #ctx_chan + bra #main_done // request to store current channel context? main_not_ctx_chan: cmpu b32 $r14 0x0002 - bra ne main_not_ctx_save + bra ne #main_not_ctx_save trace_set(T_SAVE) bclr $flags $p1 bclr $flags $p2 - call ctx_xfer + call #ctx_xfer trace_clr(T_SAVE) - bra main_done + bra #main_done main_not_ctx_save: shl b32 $r15 $r14 16 or $r15 E_BAD_COMMAND - call error - bra main + call #error + bra #main main_done: mov $r1 0x820 @@ -410,7 +455,7 @@ main: clear b32 $r2 bset $r2 31 iowr I[$r1 + 0x000] $r2 // CC_SCRATCH[0] |= 0x80000000 - bra main + bra #main // interrupt handler ih: @@ -427,13 +472,13 @@ ih: // incoming fifo command? iord $r10 I[$r0 + 0x200] // INTR and $r11 $r10 0x00000004 - bra e ih_no_fifo + bra e #ih_no_fifo // queue incoming fifo command for later processing mov $r11 0x1900 - mov $r13 cmd_queue + mov $r13 #cmd_queue iord $r14 I[$r11 + 0x100] // FIFO_CMD iord $r15 I[$r11 + 0x000] // FIFO_DATA - call queue_put + call #queue_put add b32 $r11 0x400 mov $r14 1 iowr I[$r11 + 0x000] $r14 // FIFO_ACK @@ -441,18 +486,18 @@ ih: // context switch request? ih_no_fifo: and $r11 $r10 0x00000100 - bra e ih_no_ctxsw + bra e #ih_no_ctxsw // enqueue a context switch for later processing - mov $r13 cmd_queue + mov $r13 #cmd_queue mov $r14 0x4001 - call queue_put + call #queue_put // anything we didn't handle, bring it to the host's attention ih_no_ctxsw: mov $r11 0x104 not b32 $r11 and $r11 $r10 $r11 - bra e ih_no_other + bra e #ih_no_other mov $r10 0xc1c shl b32 $r10 6 iowr I[$r10] $r11 // INTR_UP_SET @@ -478,11 +523,11 @@ ctx_4160s: mov $r14 0x4160 sethi $r14 0x400000 mov $r15 1 - call nv_wr32 + call #nv_wr32 ctx_4160s_wait: - call nv_rd32 + call #nv_rd32 xbit $r15 $r15 4 - bra e ctx_4160s_wait + bra e #ctx_4160s_wait ret // Without clearing again at end of xfer, some things cause PGRAPH @@ -492,7 +537,7 @@ ctx_4160c: mov $r14 0x4160 sethi $r14 0x400000 clear b32 $r15 - call nv_wr32 + call #nv_wr32 ret // Again, not real sure @@ -503,7 +548,7 @@ ctx_4170s: mov $r14 0x4170 sethi $r14 0x400000 or $r15 0x10 - call nv_wr32 + call #nv_wr32 ret // Waits for a ctx_4170s() call to complete @@ -511,9 +556,9 @@ ctx_4170s: ctx_4170w: mov $r14 0x4170 sethi $r14 0x400000 - call nv_rd32 + call #nv_rd32 and $r15 0x10 - bra ne ctx_4170w + bra ne #ctx_4170w ret // Disables various things, waits a bit, and re-enables them.. @@ -530,7 +575,7 @@ ctx_redswitch: mov $r15 8 ctx_redswitch_delay: sub b32 $r15 1 - bra ne ctx_redswitch_delay + bra ne #ctx_redswitch_delay mov $r15 0x770 iowr I[$r14] $r15 // HUB_RED_SWITCH = ENABLE_ALL, POWER_ALL ret @@ -546,10 +591,10 @@ ctx_86c: iowr I[$r14] $r15 // HUB(0x86c) = val mov $r14 -0x75ec sethi $r14 0x400000 - call nv_wr32 // ROP(0xa14) = val + call #nv_wr32 // ROP(0xa14) = val mov $r14 -0x5794 sethi $r14 0x410000 - call nv_wr32 // GPC(0x86c) = val + call #nv_wr32 // GPC(0x86c) = val ret // ctx_load - load's a channel's ctxctl data, and selects its vm @@ -561,7 +606,7 @@ ctx_load: // switch to channel, somewhat magic in parts.. mov $r10 12 // DONE_UNK12 - call wait_donez + call #wait_donez mov $r1 0xa24 shl b32 $r1 6 iowr I[$r1 + 0x000] $r0 // 0x409a24 @@ -576,7 +621,7 @@ ctx_load: ctx_chan_wait_0: iord $r4 I[$r1 + 0x100] and $r4 0x1f - bra ne ctx_chan_wait_0 + bra ne #ctx_chan_wait_0 iowr I[$r3 + 0x000] $r2 // CHAN_CUR // load channel header, fetch PGRAPH context pointer @@ -595,19 +640,19 @@ ctx_load: sethi $r2 0x80000000 iowr I[$r1 + 0x000] $r2 // MEM_TARGET = vram mov $r1 0x10 // chan + 0x0210 - mov $r2 xfer_data + mov $r2 #xfer_data sethi $r2 0x00020000 // 16 bytes xdld $r1 $r2 xdwait trace_clr(T_LCHAN) // update current context - ld b32 $r1 D[$r0 + xfer_data + 4] + ld b32 $r1 D[$r0 + #xfer_data + 4] shl b32 $r1 24 - ld b32 $r2 D[$r0 + xfer_data + 0] + ld b32 $r2 D[$r0 + #xfer_data + 0] shr b32 $r2 8 or $r1 $r2 - st b32 D[$r0 + ctx_current] $r1 + st b32 D[$r0 + #ctx_current] $r1 // set transfer base to start of context, and fetch context header trace_set(T_LCTXH) @@ -618,7 +663,7 @@ ctx_load: mov $r1 0xa20 shl b32 $r1 6 iowr I[$r1 + 0x000] $r2 // MEM_TARGET = vm - mov $r1 chan_data + mov $r1 #chan_data sethi $r1 0x00060000 // 256 bytes xdld $r0 $r1 xdwait @@ -635,10 +680,10 @@ ctx_load: // In: $r2 channel address // ctx_chan: - call ctx_4160s - call ctx_load + call #ctx_4160s + call #ctx_load mov $r10 12 // DONE_UNK12 - call wait_donez + call #wait_donez mov $r1 0xa10 shl b32 $r1 6 mov $r2 5 @@ -646,8 +691,8 @@ ctx_chan: ctx_chan_wait: iord $r2 I[$r1 + 0x000] or $r2 $r2 - bra ne ctx_chan_wait - call ctx_4160c + bra ne #ctx_chan_wait + call #ctx_4160c ret // Execute per-context state overrides list @@ -661,7 +706,7 @@ ctx_chan: // ctx_mmio_exec: // set transfer base to be the mmio list - ld b32 $r3 D[$r0 + chan_mmio_address] + ld b32 $r3 D[$r0 + #chan_mmio_address] mov $r2 0xa04 shl b32 $r2 6 iowr I[$r2 + 0x000] $r3 // MEM_BASE @@ -670,31 +715,31 @@ ctx_mmio_exec: ctx_mmio_loop: // fetch next 256 bytes of mmio list if necessary and $r4 $r3 0xff - bra ne ctx_mmio_pull - mov $r5 xfer_data + bra ne #ctx_mmio_pull + mov $r5 #xfer_data sethi $r5 0x00060000 // 256 bytes xdld $r3 $r5 xdwait // execute a single list entry ctx_mmio_pull: - ld b32 $r14 D[$r4 + xfer_data + 0x00] - ld b32 $r15 D[$r4 + xfer_data + 0x04] - call nv_wr32 + ld b32 $r14 D[$r4 + #xfer_data + 0x00] + ld b32 $r15 D[$r4 + #xfer_data + 0x04] + call #nv_wr32 // next! add b32 $r3 8 sub b32 $r1 1 - bra ne ctx_mmio_loop + bra ne #ctx_mmio_loop // set transfer base back to the current context ctx_mmio_done: - ld b32 $r3 D[$r0 + ctx_current] + ld b32 $r3 D[$r0 + #ctx_current] iowr I[$r2 + 0x000] $r3 // MEM_BASE // disable the mmio list now, we don't need/want to execute it again - st b32 D[$r0 + chan_mmio_count] $r0 - mov $r1 chan_data + st b32 D[$r0 + #chan_mmio_count] $r0 + mov $r1 #chan_data sethi $r1 0x00060000 // 256 bytes xdst $r0 $r1 xdwait @@ -709,46 +754,46 @@ ctx_mmio_exec: // on load it means: "a save preceeded this load" // ctx_xfer: - bra not $p1 ctx_xfer_pre - bra $p2 ctx_xfer_pre_load + bra not $p1 #ctx_xfer_pre + bra $p2 #ctx_xfer_pre_load ctx_xfer_pre: mov $r15 0x10 - call ctx_86c - call ctx_4160s - bra not $p1 ctx_xfer_exec + call #ctx_86c + call #ctx_4160s + bra not $p1 #ctx_xfer_exec ctx_xfer_pre_load: mov $r15 2 - call ctx_4170s - call ctx_4170w - call ctx_redswitch + call #ctx_4170s + call #ctx_4170w + call #ctx_redswitch clear b32 $r15 - call ctx_4170s - call ctx_load + call #ctx_4170s + call #ctx_load // fetch context pointer, and initiate xfer on all GPCs ctx_xfer_exec: - ld b32 $r1 D[$r0 + ctx_current] + ld b32 $r1 D[$r0 + #ctx_current] mov $r2 0x414 shl b32 $r2 6 iowr I[$r2 + 0x000] $r0 // BAR_STATUS = reset mov $r14 -0x5b00 sethi $r14 0x410000 mov b32 $r15 $r1 - call nv_wr32 // GPC_BCAST_WRCMD_DATA = ctx pointer + call #nv_wr32 // GPC_BCAST_WRCMD_DATA = ctx pointer add b32 $r14 4 xbit $r15 $flags $p1 xbit $r2 $flags $p2 shl b32 $r2 1 or $r15 $r2 - call nv_wr32 // GPC_BCAST_WRCMD_CMD = GPC_XFER(type) + call #nv_wr32 // GPC_BCAST_WRCMD_CMD = GPC_XFER(type) // strands mov $r1 0x4afc sethi $r1 0x20000 mov $r2 0xc iowr I[$r1] $r2 // STRAND_CMD(0x3f) = 0x0c - call strand_wait + call #strand_wait mov $r2 0x47fc sethi $r2 0x20000 iowr I[$r2] $r0 // STRAND_FIRST_GENE(0x3f) = 0x00 @@ -760,22 +805,22 @@ ctx_xfer: xbit $r10 $flags $p1 // direction or $r10 6 // first, last mov $r11 0 // base = 0 - ld b32 $r12 D[$r0 + hub_mmio_list_head] - ld b32 $r13 D[$r0 + hub_mmio_list_tail] + ld b32 $r12 D[$r0 + #hub_mmio_list_head] + ld b32 $r13 D[$r0 + #hub_mmio_list_tail] mov $r14 0 // not multi - call mmctx_xfer + call #mmctx_xfer // wait for GPCs to all complete mov $r10 8 // DONE_BAR - call wait_doneo + call #wait_doneo // wait for strand xfer to complete - call strand_wait + call #strand_wait // post-op - bra $p1 ctx_xfer_post + bra $p1 #ctx_xfer_post mov $r10 12 // DONE_UNK12 - call wait_donez + call #wait_donez mov $r1 0xa10 shl b32 $r1 6 mov $r2 5 @@ -783,27 +828,27 @@ ctx_xfer: ctx_xfer_post_save_wait: iord $r2 I[$r1] or $r2 $r2 - bra ne ctx_xfer_post_save_wait + bra ne #ctx_xfer_post_save_wait - bra $p2 ctx_xfer_done + bra $p2 #ctx_xfer_done ctx_xfer_post: mov $r15 2 - call ctx_4170s + call #ctx_4170s clear b32 $r15 - call ctx_86c - call strand_post - call ctx_4170w + call #ctx_86c + call #strand_post + call #ctx_4170w clear b32 $r15 - call ctx_4170s + call #ctx_4170s - bra not $p1 ctx_xfer_no_post_mmio - ld b32 $r1 D[$r0 + chan_mmio_count] + bra not $p1 #ctx_xfer_no_post_mmio + ld b32 $r1 D[$r0 + #chan_mmio_count] or $r1 $r1 - bra e ctx_xfer_no_post_mmio - call ctx_mmio_exec + bra e #ctx_xfer_no_post_mmio + call #ctx_mmio_exec ctx_xfer_no_post_mmio: - call ctx_4160c + call #ctx_4160c ctx_xfer_done: ret diff --git a/drivers/gpu/drm/nouveau/nvc0_grhub.fuc.h b/drivers/gpu/drm/nouveau/nvc0_grhub.fuc.h index 241d326..c5ed307 100644 --- a/drivers/gpu/drm/nouveau/nvc0_grhub.fuc.h +++ b/drivers/gpu/drm/nouveau/nvc0_grhub.fuc.h @@ -23,19 +23,21 @@ uint32_t nvc0_grhub_data[] = { 0x00000000, 0x00000000, 0x000000c0, - 0x01340098, + 0x013c00a0, 0x000000c1, - 0x01380098, + 0x014000a0, 0x000000c3, - 0x01340098, + 0x013c00a0, 0x000000c4, - 0x01340098, + 0x013c00a0, 0x000000c8, - 0x01340098, + 0x013c00a0, 0x000000ce, - 0x01340098, + 0x013c00a0, 0x000000cf, - 0x01340098, + 0x013c00a0, + 0x000000d9, + 0x01dc0140, 0x00000000, 0x0417e91c, 0x04400204, @@ -77,47 +79,45 @@ uint32_t nvc0_grhub_data[] = { 0x0c408900, 0x00408980, 0x044064c0, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, - 0x00000000, + 0x0417e91c, + 0x04400204, + 0x24404004, + 0x00404044, + 0x34404094, + 0x184040d0, + 0x004040f8, + 0x08404130, + 0x08404150, + 0x04404164, + 0x04404178, + 0x1c404200, + 0x34404404, + 0x0c404460, + 0x00404480, + 0x00404498, + 0x0c404604, + 0x7c404618, + 0x50404698, + 0x044046f0, + 0x54404700, + 0x00405800, + 0x08405830, + 0x00405854, + 0x0c405870, + 0x04405a00, + 0x00405a18, + 0x00406020, + 0x0c406028, + 0x044064a8, + 0x104064b4, + 0x00407804, + 0x1440780c, + 0x004078bc, + 0x18408000, + 0x00408064, + 0x08408800, + 0x0c408900, + 0x00408980, 0x00000000, 0x00000000, 0x00000000, diff --git a/drivers/gpu/drm/nouveau/nvc0_pm.c b/drivers/gpu/drm/nouveau/nvc0_pm.c index 929aded..ce65f81 100644 --- a/drivers/gpu/drm/nouveau/nvc0_pm.c +++ b/drivers/gpu/drm/nouveau/nvc0_pm.c @@ -153,3 +153,240 @@ nvc0_pm_clocks_get(struct drm_device *dev, struct nouveau_pm_level *perflvl) perflvl->vdec = read_clk(dev, 0x0e); return 0; } + +struct nvc0_pm_clock { + u32 freq; + u32 ssel; + u32 mdiv; + u32 dsrc; + u32 ddiv; + u32 coef; +}; + +struct nvc0_pm_state { + struct nvc0_pm_clock eng[16]; +}; + +static u32 +calc_div(struct drm_device *dev, int clk, u32 ref, u32 freq, u32 *ddiv) +{ + u32 div = min((ref * 2) / freq, (u32)65); + if (div < 2) + div = 2; + + *ddiv = div - 2; + return (ref * 2) / div; +} + +static u32 +calc_src(struct drm_device *dev, int clk, u32 freq, u32 *dsrc, u32 *ddiv) +{ + u32 sclk; + + /* use one of the fixed frequencies if possible */ + *ddiv = 0x00000000; + switch (freq) { + case 27000: + case 108000: + *dsrc = 0x00000000; + if (freq == 108000) + *dsrc |= 0x00030000; + return freq; + case 100000: + *dsrc = 0x00000002; + return freq; + default: + *dsrc = 0x00000003; + break; + } + + /* otherwise, calculate the closest divider */ + sclk = read_vco(dev, clk); + if (clk < 7) + sclk = calc_div(dev, clk, sclk, freq, ddiv); + return sclk; +} + +static u32 +calc_pll(struct drm_device *dev, int clk, u32 freq, u32 *coef) +{ + struct pll_lims limits; + int N, M, P, ret; + + ret = get_pll_limits(dev, 0x137000 + (clk * 0x20), &limits); + if (ret) + return 0; + + limits.refclk = read_div(dev, clk, 0x137120, 0x137140); + if (!limits.refclk) + return 0; + + ret = nva3_calc_pll(dev, &limits, freq, &N, NULL, &M, &P); + if (ret <= 0) + return 0; + + *coef = (P << 16) | (N << 8) | M; + return ret; +} + +/* A (likely rather simplified and incomplete) view of the clock tree + * + * Key: + * + * S: source select + * D: divider + * P: pll + * F: switch + * + * Engine clocks: + * + * 137250(D) ---- 137100(F0) ---- 137160(S)/1371d0(D) ------------------- ref + * (F1) ---- 1370X0(P) ---- 137120(S)/137140(D) ---- ref + * + * Not all registers exist for all clocks. For example: clocks >= 8 don't + * have their own PLL (all tied to clock 7's PLL when in PLL mode), nor do + * they have the divider at 1371d0, though the source selection at 137160 + * still exists. You must use the divider at 137250 for these instead. + * + * Memory clock: + * + * TBD, read_mem() above is likely very wrong... + * + */ + +static int +calc_clk(struct drm_device *dev, int clk, struct nvc0_pm_clock *info, u32 freq) +{ + u32 src0, div0, div1D, div1P = 0; + u32 clk0, clk1 = 0; + + /* invalid clock domain */ + if (!freq) + return 0; + + /* first possible path, using only dividers */ + clk0 = calc_src(dev, clk, freq, &src0, &div0); + clk0 = calc_div(dev, clk, clk0, freq, &div1D); + + /* see if we can get any closer using PLLs */ + if (clk0 != freq && (0x00004387 & (1 << clk))) { + if (clk < 7) + clk1 = calc_pll(dev, clk, freq, &info->coef); + else + clk1 = read_pll(dev, 0x1370e0); + clk1 = calc_div(dev, clk, clk1, freq, &div1P); + } + + /* select the method which gets closest to target freq */ + if (abs((int)freq - clk0) <= abs((int)freq - clk1)) { + info->dsrc = src0; + if (div0) { + info->ddiv |= 0x80000000; + info->ddiv |= div0 << 8; + info->ddiv |= div0; + } + if (div1D) { + info->mdiv |= 0x80000000; + info->mdiv |= div1D; + } + info->ssel = 0; + info->freq = clk0; + } else { + if (div1P) { + info->mdiv |= 0x80000000; + info->mdiv |= div1P << 8; + } + info->ssel = (1 << clk); + info->freq = clk1; + } + + return 0; +} + +void * +nvc0_pm_clocks_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl) +{ + struct drm_nouveau_private *dev_priv = dev->dev_private; + struct nvc0_pm_state *info; + int ret; + + info = kzalloc(sizeof(*info), GFP_KERNEL); + if (!info) + return ERR_PTR(-ENOMEM); + + /* NFI why this is still in the performance table, the ROPCs appear + * to get their clock from clock 2 ("hub07", actually hub05 on this + * chip, but, anyway...) as well. nvatiming confirms hub05 and ROP + * are always the same freq with the binary driver even when the + * performance table says they should differ. + */ + if (dev_priv->chipset == 0xd9) + perflvl->rop = 0; + + if ((ret = calc_clk(dev, 0x00, &info->eng[0x00], perflvl->shader)) || + (ret = calc_clk(dev, 0x01, &info->eng[0x01], perflvl->rop)) || + (ret = calc_clk(dev, 0x02, &info->eng[0x02], perflvl->hub07)) || + (ret = calc_clk(dev, 0x07, &info->eng[0x07], perflvl->hub06)) || + (ret = calc_clk(dev, 0x08, &info->eng[0x08], perflvl->hub01)) || + (ret = calc_clk(dev, 0x09, &info->eng[0x09], perflvl->copy)) || + (ret = calc_clk(dev, 0x0c, &info->eng[0x0c], perflvl->daemon)) || + (ret = calc_clk(dev, 0x0e, &info->eng[0x0e], perflvl->vdec))) { + kfree(info); + return ERR_PTR(ret); + } + + return info; +} + +static void +prog_clk(struct drm_device *dev, int clk, struct nvc0_pm_clock *info) +{ + /* program dividers at 137160/1371d0 first */ + if (clk < 7 && !info->ssel) { + nv_mask(dev, 0x1371d0 + (clk * 0x04), 0x80003f3f, info->ddiv); + nv_wr32(dev, 0x137160 + (clk * 0x04), info->dsrc); + } + + /* switch clock to non-pll mode */ + nv_mask(dev, 0x137100, (1 << clk), 0x00000000); + nv_wait(dev, 0x137100, (1 << clk), 0x00000000); + + /* reprogram pll */ + if (clk < 7) { + /* make sure it's disabled first... */ + u32 base = 0x137000 + (clk * 0x20); + u32 ctrl = nv_rd32(dev, base + 0x00); + if (ctrl & 0x00000001) { + nv_mask(dev, base + 0x00, 0x00000004, 0x00000000); + nv_mask(dev, base + 0x00, 0x00000001, 0x00000000); + } + /* program it to new values, if necessary */ + if (info->ssel) { + nv_wr32(dev, base + 0x04, info->coef); + nv_mask(dev, base + 0x00, 0x00000001, 0x00000001); + nv_wait(dev, base + 0x00, 0x00020000, 0x00020000); + nv_mask(dev, base + 0x00, 0x00020004, 0x00000004); + } + } + + /* select pll/non-pll mode, and program final clock divider */ + nv_mask(dev, 0x137100, (1 << clk), info->ssel); + nv_wait(dev, 0x137100, (1 << clk), info->ssel); + nv_mask(dev, 0x137250 + (clk * 0x04), 0x00003f3f, info->mdiv); +} + +int +nvc0_pm_clocks_set(struct drm_device *dev, void *data) +{ + struct nvc0_pm_state *info = data; + int i; + + for (i = 0; i < 16; i++) { + if (!info->eng[i].freq) + continue; + prog_clk(dev, i, &info->eng[i]); + } + + kfree(info); + return 0; +} diff --git a/drivers/gpu/drm/nouveau/nvc0_vm.c b/drivers/gpu/drm/nouveau/nvc0_vm.c index 9e35294..30d2bd5 100644 --- a/drivers/gpu/drm/nouveau/nvc0_vm.c +++ b/drivers/gpu/drm/nouveau/nvc0_vm.c @@ -77,9 +77,11 @@ void nvc0_vm_map_sg(struct nouveau_vma *vma, struct nouveau_gpuobj *pgt, struct nouveau_mem *mem, u32 pte, u32 cnt, dma_addr_t *list) { + u32 target = (vma->access & NV_MEM_ACCESS_NOSNOOP) ? 7 : 5; + pte <<= 3; while (cnt--) { - u64 phys = nvc0_vm_addr(vma, *list++, mem->memtype, 5); + u64 phys = nvc0_vm_addr(vma, *list++, mem->memtype, target); nv_wo32(pgt, pte + 0, lower_32_bits(phys)); nv_wo32(pgt, pte + 4, upper_32_bits(phys)); pte += 8; diff --git a/drivers/gpu/drm/nouveau/nvc0_vram.c b/drivers/gpu/drm/nouveau/nvc0_vram.c index ce984d5..a7eef89 100644 --- a/drivers/gpu/drm/nouveau/nvc0_vram.c +++ b/drivers/gpu/drm/nouveau/nvc0_vram.c @@ -106,31 +106,32 @@ nvc0_vram_init(struct drm_device *dev) struct nouveau_vram_engine *vram = &dev_priv->engine.vram; const u32 rsvd_head = ( 256 * 1024) >> 12; /* vga memory */ const u32 rsvd_tail = (1024 * 1024) >> 12; /* vbios etc */ - u32 parts = nv_rd32(dev, 0x121c74); + u32 parts = nv_rd32(dev, 0x022438); + u32 pmask = nv_rd32(dev, 0x022554); u32 bsize = nv_rd32(dev, 0x10f20c); u32 offset, length; bool uniform = true; int ret, part; NV_DEBUG(dev, "0x100800: 0x%08x\n", nv_rd32(dev, 0x100800)); - NV_DEBUG(dev, "parts 0x%08x bcast_mem_amount 0x%08x\n", parts, bsize); + NV_DEBUG(dev, "parts 0x%08x mask 0x%08x\n", parts, pmask); + + dev_priv->vram_type = nouveau_mem_vbios_type(dev); + dev_priv->vram_rank_B = !!(nv_rd32(dev, 0x10f200) & 0x00000004); /* read amount of vram attached to each memory controller */ - part = 0; - while (parts) { - u32 psize = nv_rd32(dev, 0x11020c + (part++ * 0x1000)); - if (psize == 0) - continue; - parts--; - - if (psize != bsize) { - if (psize < bsize) - bsize = psize; - uniform = false; + for (part = 0; part < parts; part++) { + if (!(pmask & (1 << part))) { + u32 psize = nv_rd32(dev, 0x11020c + (part * 0x1000)); + if (psize != bsize) { + if (psize < bsize) + bsize = psize; + uniform = false; + } + + NV_DEBUG(dev, "%d: mem_amount 0x%08x\n", part, psize); + dev_priv->vram_size += (u64)psize << 20; } - - NV_DEBUG(dev, "%d: mem_amount 0x%08x\n", part, psize); - dev_priv->vram_size += (u64)psize << 20; } /* if all controllers have the same amount attached, there's no holes */ diff --git a/drivers/gpu/drm/nouveau/nvd0_display.c b/drivers/gpu/drm/nouveau/nvd0_display.c index 3002d82..8a555fb 100644 --- a/drivers/gpu/drm/nouveau/nvd0_display.c +++ b/drivers/gpu/drm/nouveau/nvd0_display.c @@ -35,12 +35,34 @@ #include "nouveau_fb.h" #include "nv50_display.h" +#define EVO_DMA_NR 9 + +#define EVO_MASTER (0x00) +#define EVO_FLIP(c) (0x01 + (c)) +#define EVO_OVLY(c) (0x05 + (c)) +#define EVO_OIMM(c) (0x09 + (c)) +#define EVO_CURS(c) (0x0d + (c)) + +/* offsets in shared sync bo of various structures */ +#define EVO_SYNC(c, o) ((c) * 0x0100 + (o)) +#define EVO_MAST_NTFY EVO_SYNC( 0, 0x00) +#define EVO_FLIP_SEM0(c) EVO_SYNC((c), 0x00) +#define EVO_FLIP_SEM1(c) EVO_SYNC((c), 0x10) + +struct evo { + int idx; + dma_addr_t handle; + u32 *ptr; + struct { + u32 offset; + u16 value; + } sem; +}; + struct nvd0_display { struct nouveau_gpuobj *mem; - struct { - dma_addr_t handle; - u32 *ptr; - } evo[1]; + struct nouveau_bo *sync; + struct evo evo[9]; struct tasklet_struct tasklet; u32 modeset; @@ -53,6 +75,15 @@ nvd0_display(struct drm_device *dev) return dev_priv->engine.display.priv; } +static struct drm_crtc * +nvd0_display_crtc_get(struct drm_encoder *encoder) +{ + return nouveau_encoder(encoder)->crtc; +} + +/****************************************************************************** + * EVO channel helpers + *****************************************************************************/ static inline int evo_icmd(struct drm_device *dev, int id, u32 mthd, u32 data) { @@ -84,6 +115,9 @@ evo_wait(struct drm_device *dev, int id, int nr) put = 0; } + if (nouveau_reg_debug & NOUVEAU_REG_DEBUG_EVO) + NV_INFO(dev, "Evo%d: %p START\n", id, disp->evo[id].ptr + put); + return disp->evo[id].ptr + put; } @@ -91,104 +125,372 @@ static void evo_kick(u32 *push, struct drm_device *dev, int id) { struct nvd0_display *disp = nvd0_display(dev); + + if (nouveau_reg_debug & NOUVEAU_REG_DEBUG_EVO) { + u32 curp = nv_rd32(dev, 0x640000 + (id * 0x1000)) >> 2; + u32 *cur = disp->evo[id].ptr + curp; + + while (cur < push) + NV_INFO(dev, "Evo%d: 0x%08x\n", id, *cur++); + NV_INFO(dev, "Evo%d: %p KICK!\n", id, push); + } + nv_wr32(dev, 0x640000 + (id * 0x1000), (push - disp->evo[id].ptr) << 2); } #define evo_mthd(p,m,s) *((p)++) = (((s) << 18) | (m)) #define evo_data(p,d) *((p)++) = (d) -static struct drm_crtc * -nvd0_display_crtc_get(struct drm_encoder *encoder) +static int +evo_init_dma(struct drm_device *dev, int ch) { - return nouveau_encoder(encoder)->crtc; + struct nvd0_display *disp = nvd0_display(dev); + u32 flags; + + flags = 0x00000000; + if (ch == EVO_MASTER) + flags |= 0x01000000; + + nv_wr32(dev, 0x610494 + (ch * 0x0010), (disp->evo[ch].handle >> 8) | 3); + nv_wr32(dev, 0x610498 + (ch * 0x0010), 0x00010000); + nv_wr32(dev, 0x61049c + (ch * 0x0010), 0x00000001); + nv_mask(dev, 0x610490 + (ch * 0x0010), 0x00000010, 0x00000010); + nv_wr32(dev, 0x640000 + (ch * 0x1000), 0x00000000); + nv_wr32(dev, 0x610490 + (ch * 0x0010), 0x00000013 | flags); + if (!nv_wait(dev, 0x610490 + (ch * 0x0010), 0x80000000, 0x00000000)) { + NV_ERROR(dev, "PDISP: ch%d 0x%08x\n", ch, + nv_rd32(dev, 0x610490 + (ch * 0x0010))); + return -EBUSY; + } + + nv_mask(dev, 0x610090, (1 << ch), (1 << ch)); + nv_mask(dev, 0x6100a0, (1 << ch), (1 << ch)); + return 0; +} + +static void +evo_fini_dma(struct drm_device *dev, int ch) +{ + if (!(nv_rd32(dev, 0x610490 + (ch * 0x0010)) & 0x00000010)) + return; + + nv_mask(dev, 0x610490 + (ch * 0x0010), 0x00000010, 0x00000000); + nv_mask(dev, 0x610490 + (ch * 0x0010), 0x00000003, 0x00000000); + nv_wait(dev, 0x610490 + (ch * 0x0010), 0x80000000, 0x00000000); + nv_mask(dev, 0x610090, (1 << ch), 0x00000000); + nv_mask(dev, 0x6100a0, (1 << ch), 0x00000000); +} + +static inline void +evo_piow(struct drm_device *dev, int ch, u16 mthd, u32 data) +{ + nv_wr32(dev, 0x640000 + (ch * 0x1000) + mthd, data); +} + +static int +evo_init_pio(struct drm_device *dev, int ch) +{ + nv_wr32(dev, 0x610490 + (ch * 0x0010), 0x00000001); + if (!nv_wait(dev, 0x610490 + (ch * 0x0010), 0x00010000, 0x00010000)) { + NV_ERROR(dev, "PDISP: ch%d 0x%08x\n", ch, + nv_rd32(dev, 0x610490 + (ch * 0x0010))); + return -EBUSY; + } + + nv_mask(dev, 0x610090, (1 << ch), (1 << ch)); + nv_mask(dev, 0x6100a0, (1 << ch), (1 << ch)); + return 0; +} + +static void +evo_fini_pio(struct drm_device *dev, int ch) +{ + if (!(nv_rd32(dev, 0x610490 + (ch * 0x0010)) & 0x00000001)) + return; + + nv_mask(dev, 0x610490 + (ch * 0x0010), 0x00000010, 0x00000010); + nv_mask(dev, 0x610490 + (ch * 0x0010), 0x00000001, 0x00000000); + nv_wait(dev, 0x610490 + (ch * 0x0010), 0x00010000, 0x00000000); + nv_mask(dev, 0x610090, (1 << ch), 0x00000000); + nv_mask(dev, 0x6100a0, (1 << ch), 0x00000000); +} + +static bool +evo_sync_wait(void *data) +{ + return nouveau_bo_rd32(data, EVO_MAST_NTFY) != 0x00000000; +} + +static int +evo_sync(struct drm_device *dev, int ch) +{ + struct nvd0_display *disp = nvd0_display(dev); + u32 *push = evo_wait(dev, ch, 8); + if (push) { + nouveau_bo_wr32(disp->sync, EVO_MAST_NTFY, 0x00000000); + evo_mthd(push, 0x0084, 1); + evo_data(push, 0x80000000 | EVO_MAST_NTFY); + evo_mthd(push, 0x0080, 2); + evo_data(push, 0x00000000); + evo_data(push, 0x00000000); + evo_kick(push, dev, ch); + if (nv_wait_cb(dev, evo_sync_wait, disp->sync)) + return 0; + } + + return -EBUSY; +} + +/****************************************************************************** + * Page flipping channel + *****************************************************************************/ +struct nouveau_bo * +nvd0_display_crtc_sema(struct drm_device *dev, int crtc) +{ + return nvd0_display(dev)->sync; +} + +void +nvd0_display_flip_stop(struct drm_crtc *crtc) +{ + struct nvd0_display *disp = nvd0_display(crtc->dev); + struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc); + struct evo *evo = &disp->evo[EVO_FLIP(nv_crtc->index)]; + u32 *push; + + push = evo_wait(crtc->dev, evo->idx, 8); + if (push) { + evo_mthd(push, 0x0084, 1); + evo_data(push, 0x00000000); + evo_mthd(push, 0x0094, 1); + evo_data(push, 0x00000000); + evo_mthd(push, 0x00c0, 1); + evo_data(push, 0x00000000); + evo_mthd(push, 0x0080, 1); + evo_data(push, 0x00000000); + evo_kick(push, crtc->dev, evo->idx); + } +} + +int +nvd0_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb, + struct nouveau_channel *chan, u32 swap_interval) +{ + struct nouveau_framebuffer *nv_fb = nouveau_framebuffer(fb); + struct nvd0_display *disp = nvd0_display(crtc->dev); + struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc); + struct evo *evo = &disp->evo[EVO_FLIP(nv_crtc->index)]; + u64 offset; + u32 *push; + int ret; + + evo_sync(crtc->dev, EVO_MASTER); + + swap_interval <<= 4; + if (swap_interval == 0) + swap_interval |= 0x100; + + push = evo_wait(crtc->dev, evo->idx, 128); + if (unlikely(push == NULL)) + return -EBUSY; + + /* synchronise with the rendering channel, if necessary */ + if (likely(chan)) { + ret = RING_SPACE(chan, 10); + if (ret) + return ret; + + offset = chan->dispc_vma[nv_crtc->index].offset; + offset += evo->sem.offset; + + BEGIN_NVC0(chan, 2, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4); + OUT_RING (chan, upper_32_bits(offset)); + OUT_RING (chan, lower_32_bits(offset)); + OUT_RING (chan, 0xf00d0000 | evo->sem.value); + OUT_RING (chan, 0x1002); + BEGIN_NVC0(chan, 2, 0, NV84_SUBCHAN_SEMAPHORE_ADDRESS_HIGH, 4); + OUT_RING (chan, upper_32_bits(offset)); + OUT_RING (chan, lower_32_bits(offset ^ 0x10)); + OUT_RING (chan, 0x74b1e000); + OUT_RING (chan, 0x1001); + FIRE_RING (chan); + } else { + nouveau_bo_wr32(disp->sync, evo->sem.offset / 4, + 0xf00d0000 | evo->sem.value); + evo_sync(crtc->dev, EVO_MASTER); + } + + /* queue the flip */ + evo_mthd(push, 0x0100, 1); + evo_data(push, 0xfffe0000); + evo_mthd(push, 0x0084, 1); + evo_data(push, swap_interval); + if (!(swap_interval & 0x00000100)) { + evo_mthd(push, 0x00e0, 1); + evo_data(push, 0x40000000); + } + evo_mthd(push, 0x0088, 4); + evo_data(push, evo->sem.offset); + evo_data(push, 0xf00d0000 | evo->sem.value); + evo_data(push, 0x74b1e000); + evo_data(push, NvEvoSync); + evo_mthd(push, 0x00a0, 2); + evo_data(push, 0x00000000); + evo_data(push, 0x00000000); + evo_mthd(push, 0x00c0, 1); + evo_data(push, nv_fb->r_dma); + evo_mthd(push, 0x0110, 2); + evo_data(push, 0x00000000); + evo_data(push, 0x00000000); + evo_mthd(push, 0x0400, 5); + evo_data(push, nv_fb->nvbo->bo.offset >> 8); + evo_data(push, 0); + evo_data(push, (fb->height << 16) | fb->width); + evo_data(push, nv_fb->r_pitch); + evo_data(push, nv_fb->r_format); + evo_mthd(push, 0x0080, 1); + evo_data(push, 0x00000000); + evo_kick(push, crtc->dev, evo->idx); + + evo->sem.offset ^= 0x10; + evo->sem.value++; + return 0; } /****************************************************************************** * CRTC *****************************************************************************/ static int -nvd0_crtc_set_dither(struct nouveau_crtc *nv_crtc, bool on, bool update) +nvd0_crtc_set_dither(struct nouveau_crtc *nv_crtc, bool update) { + struct drm_nouveau_private *dev_priv = nv_crtc->base.dev->dev_private; struct drm_device *dev = nv_crtc->base.dev; - u32 *push, mode; + struct nouveau_connector *nv_connector; + struct drm_connector *connector; + u32 *push, mode = 0x00; + u32 mthd; - mode = 0x00000000; - if (on) { - /* 0x11: 6bpc dynamic 2x2 - * 0x13: 8bpc dynamic 2x2 - * 0x19: 6bpc static 2x2 - * 0x1b: 8bpc static 2x2 - * 0x21: 6bpc temporal - * 0x23: 8bpc temporal - */ - mode = 0x00000011; + nv_connector = nouveau_crtc_connector_get(nv_crtc); + connector = &nv_connector->base; + if (nv_connector->dithering_mode == DITHERING_MODE_AUTO) { + if (nv_crtc->base.fb->depth > connector->display_info.bpc * 3) + mode = DITHERING_MODE_DYNAMIC2X2; + } else { + mode = nv_connector->dithering_mode; } - push = evo_wait(dev, 0, 4); + if (nv_connector->dithering_depth == DITHERING_DEPTH_AUTO) { + if (connector->display_info.bpc >= 8) + mode |= DITHERING_DEPTH_8BPC; + } else { + mode |= nv_connector->dithering_depth; + } + + if (dev_priv->card_type < NV_E0) + mthd = 0x0490 + (nv_crtc->index * 0x0300); + else + mthd = 0x04a0 + (nv_crtc->index * 0x0300); + + push = evo_wait(dev, EVO_MASTER, 4); if (push) { - evo_mthd(push, 0x0490 + (nv_crtc->index * 0x300), 1); + evo_mthd(push, mthd, 1); evo_data(push, mode); if (update) { evo_mthd(push, 0x0080, 1); evo_data(push, 0x00000000); } - evo_kick(push, dev, 0); + evo_kick(push, dev, EVO_MASTER); } return 0; } static int -nvd0_crtc_set_scale(struct nouveau_crtc *nv_crtc, int type, bool update) +nvd0_crtc_set_scale(struct nouveau_crtc *nv_crtc, bool update) { - struct drm_display_mode *mode = &nv_crtc->base.mode; + struct drm_display_mode *omode, *umode = &nv_crtc->base.mode; struct drm_device *dev = nv_crtc->base.dev; + struct drm_crtc *crtc = &nv_crtc->base; struct nouveau_connector *nv_connector; - u32 *push, outX, outY; - - outX = mode->hdisplay; - outY = mode->vdisplay; + int mode = DRM_MODE_SCALE_NONE; + u32 oX, oY, *push; + /* start off at the resolution we programmed the crtc for, this + * effectively handles NONE/FULL scaling + */ nv_connector = nouveau_crtc_connector_get(nv_crtc); - if (nv_connector && nv_connector->native_mode) { - struct drm_display_mode *native = nv_connector->native_mode; - u32 xratio = (native->hdisplay << 19) / mode->hdisplay; - u32 yratio = (native->vdisplay << 19) / mode->vdisplay; - - switch (type) { - case DRM_MODE_SCALE_ASPECT: - if (xratio > yratio) { - outX = (mode->hdisplay * yratio) >> 19; - outY = (mode->vdisplay * yratio) >> 19; - } else { - outX = (mode->hdisplay * xratio) >> 19; - outY = (mode->vdisplay * xratio) >> 19; - } - break; - case DRM_MODE_SCALE_FULLSCREEN: - outX = native->hdisplay; - outY = native->vdisplay; - break; - default: - break; + if (nv_connector && nv_connector->native_mode) + mode = nv_connector->scaling_mode; + + if (mode != DRM_MODE_SCALE_NONE) + omode = nv_connector->native_mode; + else + omode = umode; + + oX = omode->hdisplay; + oY = omode->vdisplay; + if (omode->flags & DRM_MODE_FLAG_DBLSCAN) + oY *= 2; + + /* add overscan compensation if necessary, will keep the aspect + * ratio the same as the backend mode unless overridden by the + * user setting both hborder and vborder properties. + */ + if (nv_connector && ( nv_connector->underscan == UNDERSCAN_ON || + (nv_connector->underscan == UNDERSCAN_AUTO && + nv_connector->edid && + drm_detect_hdmi_monitor(nv_connector->edid)))) { + u32 bX = nv_connector->underscan_hborder; + u32 bY = nv_connector->underscan_vborder; + u32 aspect = (oY << 19) / oX; + + if (bX) { + oX -= (bX * 2); + if (bY) oY -= (bY * 2); + else oY = ((oX * aspect) + (aspect / 2)) >> 19; + } else { + oX -= (oX >> 4) + 32; + if (bY) oY -= (bY * 2); + else oY = ((oX * aspect) + (aspect / 2)) >> 19; } } - push = evo_wait(dev, 0, 16); + /* handle CENTER/ASPECT scaling, taking into account the areas + * removed already for overscan compensation + */ + switch (mode) { + case DRM_MODE_SCALE_CENTER: + oX = min((u32)umode->hdisplay, oX); + oY = min((u32)umode->vdisplay, oY); + /* fall-through */ + case DRM_MODE_SCALE_ASPECT: + if (oY < oX) { + u32 aspect = (umode->hdisplay << 19) / umode->vdisplay; + oX = ((oY * aspect) + (aspect / 2)) >> 19; + } else { + u32 aspect = (umode->vdisplay << 19) / umode->hdisplay; + oY = ((oX * aspect) + (aspect / 2)) >> 19; + } + break; + default: + break; + } + + push = evo_wait(dev, EVO_MASTER, 8); if (push) { evo_mthd(push, 0x04c0 + (nv_crtc->index * 0x300), 3); - evo_data(push, (outY << 16) | outX); - evo_data(push, (outY << 16) | outX); - evo_data(push, (outY << 16) | outX); + evo_data(push, (oY << 16) | oX); + evo_data(push, (oY << 16) | oX); + evo_data(push, (oY << 16) | oX); evo_mthd(push, 0x0494 + (nv_crtc->index * 0x300), 1); evo_data(push, 0x00000000); evo_mthd(push, 0x04b8 + (nv_crtc->index * 0x300), 1); - evo_data(push, (mode->vdisplay << 16) | mode->hdisplay); + evo_data(push, (umode->vdisplay << 16) | umode->hdisplay); + evo_kick(push, dev, EVO_MASTER); if (update) { - evo_mthd(push, 0x0080, 1); - evo_data(push, 0x00000000); + nvd0_display_flip_stop(crtc); + nvd0_display_flip_next(crtc, crtc->fb, NULL, 1); } - evo_kick(push, dev, 0); } return 0; @@ -201,7 +503,7 @@ nvd0_crtc_set_image(struct nouveau_crtc *nv_crtc, struct drm_framebuffer *fb, struct nouveau_framebuffer *nvfb = nouveau_framebuffer(fb); u32 *push; - push = evo_wait(fb->dev, 0, 16); + push = evo_wait(fb->dev, EVO_MASTER, 16); if (push) { evo_mthd(push, 0x0460 + (nv_crtc->index * 0x300), 1); evo_data(push, nvfb->nvbo->bo.offset >> 8); @@ -216,7 +518,7 @@ nvd0_crtc_set_image(struct nouveau_crtc *nv_crtc, struct drm_framebuffer *fb, evo_mthd(push, 0x0080, 1); evo_data(push, 0x00000000); } - evo_kick(push, fb->dev, 0); + evo_kick(push, fb->dev, EVO_MASTER); } nv_crtc->fb.tile_flags = nvfb->r_dma; @@ -227,7 +529,7 @@ static void nvd0_crtc_cursor_show(struct nouveau_crtc *nv_crtc, bool show, bool update) { struct drm_device *dev = nv_crtc->base.dev; - u32 *push = evo_wait(dev, 0, 16); + u32 *push = evo_wait(dev, EVO_MASTER, 16); if (push) { if (show) { evo_mthd(push, 0x0480 + (nv_crtc->index * 0x300), 2); @@ -247,7 +549,7 @@ nvd0_crtc_cursor_show(struct nouveau_crtc *nv_crtc, bool show, bool update) evo_data(push, 0x00000000); } - evo_kick(push, dev, 0); + evo_kick(push, dev, EVO_MASTER); } } @@ -262,7 +564,9 @@ nvd0_crtc_prepare(struct drm_crtc *crtc) struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc); u32 *push; - push = evo_wait(crtc->dev, 0, 2); + nvd0_display_flip_stop(crtc); + + push = evo_wait(crtc->dev, EVO_MASTER, 2); if (push) { evo_mthd(push, 0x0474 + (nv_crtc->index * 0x300), 1); evo_data(push, 0x00000000); @@ -270,7 +574,7 @@ nvd0_crtc_prepare(struct drm_crtc *crtc) evo_data(push, 0x03000000); evo_mthd(push, 0x045c + (nv_crtc->index * 0x300), 1); evo_data(push, 0x00000000); - evo_kick(push, crtc->dev, 0); + evo_kick(push, crtc->dev, EVO_MASTER); } nvd0_crtc_cursor_show(nv_crtc, false, false); @@ -282,7 +586,7 @@ nvd0_crtc_commit(struct drm_crtc *crtc) struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc); u32 *push; - push = evo_wait(crtc->dev, 0, 32); + push = evo_wait(crtc->dev, EVO_MASTER, 32); if (push) { evo_mthd(push, 0x0474 + (nv_crtc->index * 0x300), 1); evo_data(push, nv_crtc->fb.tile_flags); @@ -295,10 +599,11 @@ nvd0_crtc_commit(struct drm_crtc *crtc) evo_data(push, NvEvoVRAM); evo_mthd(push, 0x0430 + (nv_crtc->index * 0x300), 1); evo_data(push, 0xffffff00); - evo_kick(push, crtc->dev, 0); + evo_kick(push, crtc->dev, EVO_MASTER); } nvd0_crtc_cursor_show(nv_crtc, nv_crtc->cursor.visible, true); + nvd0_display_flip_next(crtc, crtc->fb, NULL, 1); } static bool @@ -333,53 +638,61 @@ nvd0_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *umode, { struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc); struct nouveau_connector *nv_connector; - u32 htotal = mode->htotal; - u32 vtotal = mode->vtotal; - u32 hsyncw = mode->hsync_end - mode->hsync_start - 1; - u32 vsyncw = mode->vsync_end - mode->vsync_start - 1; - u32 hfrntp = mode->hsync_start - mode->hdisplay; - u32 vfrntp = mode->vsync_start - mode->vdisplay; - u32 hbackp = mode->htotal - mode->hsync_end; - u32 vbackp = mode->vtotal - mode->vsync_end; - u32 hss2be = hsyncw + hbackp; - u32 vss2be = vsyncw + vbackp; - u32 hss2de = htotal - hfrntp; - u32 vss2de = vtotal - vfrntp; - u32 syncs, *push; + u32 ilace = (mode->flags & DRM_MODE_FLAG_INTERLACE) ? 2 : 1; + u32 vscan = (mode->flags & DRM_MODE_FLAG_DBLSCAN) ? 2 : 1; + u32 hactive, hsynce, hbackp, hfrontp, hblanke, hblanks; + u32 vactive, vsynce, vbackp, vfrontp, vblanke, vblanks; + u32 vblan2e = 0, vblan2s = 1; + u32 *push; int ret; - syncs = 0x00000001; - if (mode->flags & DRM_MODE_FLAG_NHSYNC) - syncs |= 0x00000008; - if (mode->flags & DRM_MODE_FLAG_NVSYNC) - syncs |= 0x00000010; + hactive = mode->htotal; + hsynce = mode->hsync_end - mode->hsync_start - 1; + hbackp = mode->htotal - mode->hsync_end; + hblanke = hsynce + hbackp; + hfrontp = mode->hsync_start - mode->hdisplay; + hblanks = mode->htotal - hfrontp - 1; + + vactive = mode->vtotal * vscan / ilace; + vsynce = ((mode->vsync_end - mode->vsync_start) * vscan / ilace) - 1; + vbackp = (mode->vtotal - mode->vsync_end) * vscan / ilace; + vblanke = vsynce + vbackp; + vfrontp = (mode->vsync_start - mode->vdisplay) * vscan / ilace; + vblanks = vactive - vfrontp - 1; + if (mode->flags & DRM_MODE_FLAG_INTERLACE) { + vblan2e = vactive + vsynce + vbackp; + vblan2s = vblan2e + (mode->vdisplay * vscan / ilace); + vactive = (vactive * 2) + 1; + } ret = nvd0_crtc_swap_fbs(crtc, old_fb); if (ret) return ret; - push = evo_wait(crtc->dev, 0, 64); + push = evo_wait(crtc->dev, EVO_MASTER, 64); if (push) { - evo_mthd(push, 0x0410 + (nv_crtc->index * 0x300), 5); + evo_mthd(push, 0x0410 + (nv_crtc->index * 0x300), 6); evo_data(push, 0x00000000); - evo_data(push, (vtotal << 16) | htotal); - evo_data(push, (vsyncw << 16) | hsyncw); - evo_data(push, (vss2be << 16) | hss2be); - evo_data(push, (vss2de << 16) | hss2de); + evo_data(push, (vactive << 16) | hactive); + evo_data(push, ( vsynce << 16) | hsynce); + evo_data(push, (vblanke << 16) | hblanke); + evo_data(push, (vblanks << 16) | hblanks); + evo_data(push, (vblan2e << 16) | vblan2s); evo_mthd(push, 0x042c + (nv_crtc->index * 0x300), 1); evo_data(push, 0x00000000); /* ??? */ evo_mthd(push, 0x0450 + (nv_crtc->index * 0x300), 3); evo_data(push, mode->clock * 1000); evo_data(push, 0x00200000); /* ??? */ evo_data(push, mode->clock * 1000); - evo_mthd(push, 0x0404 + (nv_crtc->index * 0x300), 1); - evo_data(push, syncs); - evo_kick(push, crtc->dev, 0); + evo_mthd(push, 0x04d0 + (nv_crtc->index * 0x300), 2); + evo_data(push, 0x00000311); + evo_data(push, 0x00000100); + evo_kick(push, crtc->dev, EVO_MASTER); } nv_connector = nouveau_crtc_connector_get(nv_crtc); - nvd0_crtc_set_dither(nv_crtc, nv_connector->use_dithering, false); - nvd0_crtc_set_scale(nv_crtc, nv_connector->scaling_mode, false); + nvd0_crtc_set_dither(nv_crtc, false); + nvd0_crtc_set_scale(nv_crtc, false); nvd0_crtc_set_image(nv_crtc, crtc->fb, x, y, false); return 0; } @@ -400,7 +713,9 @@ nvd0_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y, if (ret) return ret; + nvd0_display_flip_stop(crtc); nvd0_crtc_set_image(nv_crtc, crtc->fb, x, y, true); + nvd0_display_flip_next(crtc, crtc->fb, NULL, 1); return 0; } @@ -410,6 +725,7 @@ nvd0_crtc_mode_set_base_atomic(struct drm_crtc *crtc, enum mode_set_atomic state) { struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc); + nvd0_display_flip_stop(crtc); nvd0_crtc_set_image(nv_crtc, fb, x, y, true); return 0; } @@ -472,10 +788,10 @@ static int nvd0_crtc_cursor_move(struct drm_crtc *crtc, int x, int y) { struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc); - const u32 data = (y << 16) | (x & 0xffff); + int ch = EVO_CURS(nv_crtc->index); - nv_wr32(crtc->dev, 0x64d084 + (nv_crtc->index * 0x1000), data); - nv_wr32(crtc->dev, 0x64d080 + (nv_crtc->index * 0x1000), 0x00000000); + evo_piow(crtc->dev, ch, 0x0084, (y << 16) | (x & 0xffff)); + evo_piow(crtc->dev, ch, 0x0080, 0x00000000); return 0; } @@ -525,6 +841,7 @@ static const struct drm_crtc_funcs nvd0_crtc_func = { .gamma_set = nvd0_crtc_gamma_set, .set_config = drm_crtc_helper_set_config, .destroy = nvd0_crtc_destroy, + .page_flip = nouveau_crtc_page_flip, }; static void @@ -640,11 +957,6 @@ nvd0_dac_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode, } static void -nvd0_dac_prepare(struct drm_encoder *encoder) -{ -} - -static void nvd0_dac_commit(struct drm_encoder *encoder) { } @@ -655,16 +967,29 @@ nvd0_dac_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, { struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc); - u32 *push; + u32 syncs, magic, *push; + + syncs = 0x00000001; + if (mode->flags & DRM_MODE_FLAG_NHSYNC) + syncs |= 0x00000008; + if (mode->flags & DRM_MODE_FLAG_NVSYNC) + syncs |= 0x00000010; + + magic = 0x31ec6000 | (nv_crtc->index << 25); + if (mode->flags & DRM_MODE_FLAG_INTERLACE) + magic |= 0x00000001; nvd0_dac_dpms(encoder, DRM_MODE_DPMS_ON); - push = evo_wait(encoder->dev, 0, 4); + push = evo_wait(encoder->dev, EVO_MASTER, 8); if (push) { - evo_mthd(push, 0x0180 + (nv_encoder->or * 0x20), 2); + evo_mthd(push, 0x0404 + (nv_crtc->index * 0x300), 2); + evo_data(push, syncs); + evo_data(push, magic); + evo_mthd(push, 0x0180 + (nv_encoder->or * 0x020), 2); evo_data(push, 1 << nv_crtc->index); evo_data(push, 0x00ff); - evo_kick(push, encoder->dev, 0); + evo_kick(push, encoder->dev, EVO_MASTER); } nv_encoder->crtc = encoder->crtc; @@ -680,13 +1005,13 @@ nvd0_dac_disconnect(struct drm_encoder *encoder) if (nv_encoder->crtc) { nvd0_crtc_prepare(nv_encoder->crtc); - push = evo_wait(dev, 0, 4); + push = evo_wait(dev, EVO_MASTER, 4); if (push) { evo_mthd(push, 0x0180 + (nv_encoder->or * 0x20), 1); evo_data(push, 0x00000000); evo_mthd(push, 0x0080, 1); evo_data(push, 0x00000000); - evo_kick(push, dev, 0); + evo_kick(push, dev, EVO_MASTER); } nv_encoder->crtc = NULL; @@ -724,7 +1049,7 @@ nvd0_dac_destroy(struct drm_encoder *encoder) static const struct drm_encoder_helper_funcs nvd0_dac_hfunc = { .dpms = nvd0_dac_dpms, .mode_fixup = nvd0_dac_mode_fixup, - .prepare = nvd0_dac_prepare, + .prepare = nvd0_dac_disconnect, .commit = nvd0_dac_commit, .mode_set = nvd0_dac_mode_set, .disable = nvd0_dac_disconnect, @@ -760,8 +1085,253 @@ nvd0_dac_create(struct drm_connector *connector, struct dcb_entry *dcbe) } /****************************************************************************** + * Audio + *****************************************************************************/ +static void +nvd0_audio_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode) +{ + struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); + struct nouveau_connector *nv_connector; + struct drm_device *dev = encoder->dev; + int i, or = nv_encoder->or * 0x30; + + nv_connector = nouveau_encoder_connector_get(nv_encoder); + if (!drm_detect_monitor_audio(nv_connector->edid)) + return; + + nv_mask(dev, 0x10ec10 + or, 0x80000003, 0x80000001); + + drm_edid_to_eld(&nv_connector->base, nv_connector->edid); + if (nv_connector->base.eld[0]) { + u8 *eld = nv_connector->base.eld; + + for (i = 0; i < eld[2] * 4; i++) + nv_wr32(dev, 0x10ec00 + or, (i << 8) | eld[i]); + for (i = eld[2] * 4; i < 0x60; i++) + nv_wr32(dev, 0x10ec00 + or, (i << 8) | 0x00); + + nv_mask(dev, 0x10ec10 + or, 0x80000002, 0x80000002); + } +} + +static void +nvd0_audio_disconnect(struct drm_encoder *encoder) +{ + struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); + struct drm_device *dev = encoder->dev; + int or = nv_encoder->or * 0x30; + + nv_mask(dev, 0x10ec10 + or, 0x80000003, 0x80000000); +} + +/****************************************************************************** + * HDMI + *****************************************************************************/ +static void +nvd0_hdmi_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode) +{ + struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); + struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc); + struct nouveau_connector *nv_connector; + struct drm_device *dev = encoder->dev; + int head = nv_crtc->index * 0x800; + u32 rekey = 56; /* binary driver, and tegra constant */ + u32 max_ac_packet; + + nv_connector = nouveau_encoder_connector_get(nv_encoder); + if (!drm_detect_hdmi_monitor(nv_connector->edid)) + return; + + max_ac_packet = mode->htotal - mode->hdisplay; + max_ac_packet -= rekey; + max_ac_packet -= 18; /* constant from tegra */ + max_ac_packet /= 32; + + /* AVI InfoFrame */ + nv_mask(dev, 0x616714 + head, 0x00000001, 0x00000000); + nv_wr32(dev, 0x61671c + head, 0x000d0282); + nv_wr32(dev, 0x616720 + head, 0x0000006f); + nv_wr32(dev, 0x616724 + head, 0x00000000); + nv_wr32(dev, 0x616728 + head, 0x00000000); + nv_wr32(dev, 0x61672c + head, 0x00000000); + nv_mask(dev, 0x616714 + head, 0x00000001, 0x00000001); + + /* ??? InfoFrame? */ + nv_mask(dev, 0x6167a4 + head, 0x00000001, 0x00000000); + nv_wr32(dev, 0x6167ac + head, 0x00000010); + nv_mask(dev, 0x6167a4 + head, 0x00000001, 0x00000001); + + /* HDMI_CTRL */ + nv_mask(dev, 0x616798 + head, 0x401f007f, 0x40000000 | rekey | + max_ac_packet << 16); + + /* NFI, audio doesn't work without it though.. */ + nv_mask(dev, 0x616548 + head, 0x00000070, 0x00000000); + + nvd0_audio_mode_set(encoder, mode); +} + +static void +nvd0_hdmi_disconnect(struct drm_encoder *encoder) +{ + struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); + struct nouveau_crtc *nv_crtc = nouveau_crtc(nv_encoder->crtc); + struct drm_device *dev = encoder->dev; + int head = nv_crtc->index * 0x800; + + nvd0_audio_disconnect(encoder); + + nv_mask(dev, 0x616798 + head, 0x40000000, 0x00000000); + nv_mask(dev, 0x6167a4 + head, 0x00000001, 0x00000000); + nv_mask(dev, 0x616714 + head, 0x00000001, 0x00000000); +} + +/****************************************************************************** * SOR *****************************************************************************/ +static inline u32 +nvd0_sor_dp_lane_map(struct drm_device *dev, struct dcb_entry *dcb, u8 lane) +{ + static const u8 nvd0[] = { 16, 8, 0, 24 }; + return nvd0[lane]; +} + +static void +nvd0_sor_dp_train_set(struct drm_device *dev, struct dcb_entry *dcb, u8 pattern) +{ + const u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1); + const u32 loff = (or * 0x800) + (link * 0x80); + nv_mask(dev, 0x61c110 + loff, 0x0f0f0f0f, 0x01010101 * pattern); +} + +static void +nvd0_sor_dp_train_adj(struct drm_device *dev, struct dcb_entry *dcb, + u8 lane, u8 swing, u8 preem) +{ + const u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1); + const u32 loff = (or * 0x800) + (link * 0x80); + u32 shift = nvd0_sor_dp_lane_map(dev, dcb, lane); + u32 mask = 0x000000ff << shift; + u8 *table, *entry, *config = NULL; + + switch (swing) { + case 0: preem += 0; break; + case 1: preem += 4; break; + case 2: preem += 7; break; + case 3: preem += 9; break; + } + + table = nouveau_dp_bios_data(dev, dcb, &entry); + if (table) { + if (table[0] == 0x30) { + config = entry + table[4]; + config += table[5] * preem; + } else + if (table[0] == 0x40) { + config = table + table[1]; + config += table[2] * table[3]; + config += table[6] * preem; + } + } + + if (!config) { + NV_ERROR(dev, "PDISP: unsupported DP table for chipset\n"); + return; + } + + nv_mask(dev, 0x61c118 + loff, mask, config[1] << shift); + nv_mask(dev, 0x61c120 + loff, mask, config[2] << shift); + nv_mask(dev, 0x61c130 + loff, 0x0000ff00, config[3] << 8); + nv_mask(dev, 0x61c13c + loff, 0x00000000, 0x00000000); +} + +static void +nvd0_sor_dp_link_set(struct drm_device *dev, struct dcb_entry *dcb, int crtc, + int link_nr, u32 link_bw, bool enhframe) +{ + const u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1); + const u32 loff = (or * 0x800) + (link * 0x80); + const u32 soff = (or * 0x800); + u32 dpctrl = nv_rd32(dev, 0x61c10c + loff) & ~0x001f4000; + u32 clksor = nv_rd32(dev, 0x612300 + soff) & ~0x007c0000; + u32 script = 0x0000, lane_mask = 0; + u8 *table, *entry; + int i; + + link_bw /= 27000; + + table = nouveau_dp_bios_data(dev, dcb, &entry); + if (table) { + if (table[0] == 0x30) entry = ROMPTR(dev, entry[10]); + else if (table[0] == 0x40) entry = ROMPTR(dev, entry[9]); + else entry = NULL; + + while (entry) { + if (entry[0] >= link_bw) + break; + entry += 3; + } + + nouveau_bios_run_init_table(dev, script, dcb, crtc); + } + + clksor |= link_bw << 18; + dpctrl |= ((1 << link_nr) - 1) << 16; + if (enhframe) + dpctrl |= 0x00004000; + + for (i = 0; i < link_nr; i++) + lane_mask |= 1 << (nvd0_sor_dp_lane_map(dev, dcb, i) >> 3); + + nv_wr32(dev, 0x612300 + soff, clksor); + nv_wr32(dev, 0x61c10c + loff, dpctrl); + nv_mask(dev, 0x61c130 + loff, 0x0000000f, lane_mask); +} + +static void +nvd0_sor_dp_link_get(struct drm_device *dev, struct dcb_entry *dcb, + u32 *link_nr, u32 *link_bw) +{ + const u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1); + const u32 loff = (or * 0x800) + (link * 0x80); + const u32 soff = (or * 0x800); + u32 dpctrl = nv_rd32(dev, 0x61c10c + loff) & 0x000f0000; + u32 clksor = nv_rd32(dev, 0x612300 + soff); + + if (dpctrl > 0x00030000) *link_nr = 4; + else if (dpctrl > 0x00010000) *link_nr = 2; + else *link_nr = 1; + + *link_bw = (clksor & 0x007c0000) >> 18; + *link_bw *= 27000; +} + +static void +nvd0_sor_dp_calc_tu(struct drm_device *dev, struct dcb_entry *dcb, + u32 crtc, u32 datarate) +{ + const u32 symbol = 100000; + const u32 TU = 64; + u32 link_nr, link_bw; + u64 ratio, value; + + nvd0_sor_dp_link_get(dev, dcb, &link_nr, &link_bw); + + ratio = datarate; + ratio *= symbol; + do_div(ratio, link_nr * link_bw); + + value = (symbol - ratio) * TU; + value *= ratio; + do_div(value, symbol); + do_div(value, symbol); + + value += 5; + value |= 0x08000000; + + nv_wr32(dev, 0x616610 + (crtc * 0x800), value); +} + static void nvd0_sor_dpms(struct drm_encoder *encoder, int mode) { @@ -794,6 +1364,16 @@ nvd0_sor_dpms(struct drm_encoder *encoder, int mode) nv_mask(dev, 0x61c004 + (or * 0x0800), 0x80000001, dpms_ctrl); nv_wait(dev, 0x61c004 + (or * 0x0800), 0x80000000, 0x00000000); nv_wait(dev, 0x61c030 + (or * 0x0800), 0x10000000, 0x00000000); + + if (nv_encoder->dcb->type == OUTPUT_DP) { + struct dp_train_func func = { + .link_set = nvd0_sor_dp_link_set, + .train_set = nvd0_sor_dp_train_set, + .train_adj = nvd0_sor_dp_train_adj + }; + + nouveau_dp_dpms(encoder, mode, nv_encoder->dp.datarate, &func); + } } static bool @@ -816,8 +1396,37 @@ nvd0_sor_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode, } static void +nvd0_sor_disconnect(struct drm_encoder *encoder) +{ + struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); + struct drm_device *dev = encoder->dev; + u32 *push; + + if (nv_encoder->crtc) { + nvd0_crtc_prepare(nv_encoder->crtc); + + push = evo_wait(dev, EVO_MASTER, 4); + if (push) { + evo_mthd(push, 0x0200 + (nv_encoder->or * 0x20), 1); + evo_data(push, 0x00000000); + evo_mthd(push, 0x0080, 1); + evo_data(push, 0x00000000); + evo_kick(push, dev, EVO_MASTER); + } + + nvd0_hdmi_disconnect(encoder); + + nv_encoder->crtc = NULL; + nv_encoder->last_dpms = DRM_MODE_DPMS_OFF; + } +} + +static void nvd0_sor_prepare(struct drm_encoder *encoder) { + nvd0_sor_disconnect(encoder); + if (nouveau_encoder(encoder)->dcb->type == OUTPUT_DP) + evo_sync(encoder->dev, EVO_MASTER); } static void @@ -829,13 +1438,25 @@ static void nvd0_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *umode, struct drm_display_mode *mode) { - struct drm_nouveau_private *dev_priv = encoder->dev->dev_private; + struct drm_device *dev = encoder->dev; + struct drm_nouveau_private *dev_priv = dev->dev_private; struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc); struct nouveau_connector *nv_connector; struct nvbios *bios = &dev_priv->vbios; u32 mode_ctrl = (1 << nv_crtc->index); - u32 *push, or_config; + u32 syncs, magic, *push; + u32 or_config; + + syncs = 0x00000001; + if (mode->flags & DRM_MODE_FLAG_NHSYNC) + syncs |= 0x00000008; + if (mode->flags & DRM_MODE_FLAG_NVSYNC) + syncs |= 0x00000010; + + magic = 0x31ec6000 | (nv_crtc->index << 25); + if (mode->flags & DRM_MODE_FLAG_INTERLACE) + magic |= 0x00000001; nv_connector = nouveau_encoder_connector_get(nv_encoder); switch (nv_encoder->dcb->type) { @@ -852,6 +1473,8 @@ nvd0_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *umode, or_config = (mode_ctrl & 0x00000f00) >> 8; if (mode->clock >= 165000) or_config |= 0x0100; + + nvd0_hdmi_mode_set(encoder, mode); break; case OUTPUT_LVDS: or_config = (mode_ctrl & 0x00000f00) >> 8; @@ -861,7 +1484,7 @@ nvd0_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *umode, if (bios->fp.if_is_24bit) or_config |= 0x0200; } else { - if (nv_connector->dcb->type == DCB_CONNECTOR_LVDS_SPWG) { + if (nv_connector->type == DCB_CONNECTOR_LVDS_SPWG) { if (((u8 *)nv_connector->edid)[121] == 2) or_config |= 0x0100; } else @@ -882,6 +1505,22 @@ nvd0_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *umode, } break; + case OUTPUT_DP: + if (nv_connector->base.display_info.bpc == 6) { + nv_encoder->dp.datarate = mode->clock * 18 / 8; + syncs |= 0x00000140; + } else { + nv_encoder->dp.datarate = mode->clock * 24 / 8; + syncs |= 0x00000180; + } + + if (nv_encoder->dcb->sorconf.link & 1) + mode_ctrl |= 0x00000800; + else + mode_ctrl |= 0x00000900; + + or_config = (mode_ctrl & 0x00000f00) >> 8; + break; default: BUG_ON(1); break; @@ -889,42 +1528,26 @@ nvd0_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *umode, nvd0_sor_dpms(encoder, DRM_MODE_DPMS_ON); - push = evo_wait(encoder->dev, 0, 4); + if (nv_encoder->dcb->type == OUTPUT_DP) { + nvd0_sor_dp_calc_tu(dev, nv_encoder->dcb, nv_crtc->index, + nv_encoder->dp.datarate); + } + + push = evo_wait(dev, EVO_MASTER, 8); if (push) { - evo_mthd(push, 0x0200 + (nv_encoder->or * 0x20), 2); + evo_mthd(push, 0x0404 + (nv_crtc->index * 0x300), 2); + evo_data(push, syncs); + evo_data(push, magic); + evo_mthd(push, 0x0200 + (nv_encoder->or * 0x020), 2); evo_data(push, mode_ctrl); evo_data(push, or_config); - evo_kick(push, encoder->dev, 0); + evo_kick(push, dev, EVO_MASTER); } nv_encoder->crtc = encoder->crtc; } static void -nvd0_sor_disconnect(struct drm_encoder *encoder) -{ - struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder); - struct drm_device *dev = encoder->dev; - u32 *push; - - if (nv_encoder->crtc) { - nvd0_crtc_prepare(nv_encoder->crtc); - - push = evo_wait(dev, 0, 4); - if (push) { - evo_mthd(push, 0x0200 + (nv_encoder->or * 0x20), 1); - evo_data(push, 0x00000000); - evo_mthd(push, 0x0080, 1); - evo_data(push, 0x00000000); - evo_kick(push, dev, 0); - } - - nv_encoder->crtc = NULL; - nv_encoder->last_dpms = DRM_MODE_DPMS_OFF; - } -} - -static void nvd0_sor_destroy(struct drm_encoder *encoder) { drm_encoder_cleanup(encoder); @@ -976,17 +1599,19 @@ static struct dcb_entry * lookup_dcb(struct drm_device *dev, int id, u32 mc) { struct drm_nouveau_private *dev_priv = dev->dev_private; - int type, or, i; + int type, or, i, link = -1; if (id < 4) { type = OUTPUT_ANALOG; or = id; } else { switch (mc & 0x00000f00) { - case 0x00000000: type = OUTPUT_LVDS; break; - case 0x00000100: type = OUTPUT_TMDS; break; - case 0x00000200: type = OUTPUT_TMDS; break; - case 0x00000500: type = OUTPUT_TMDS; break; + case 0x00000000: link = 0; type = OUTPUT_LVDS; break; + case 0x00000100: link = 0; type = OUTPUT_TMDS; break; + case 0x00000200: link = 1; type = OUTPUT_TMDS; break; + case 0x00000500: link = 0; type = OUTPUT_TMDS; break; + case 0x00000800: link = 0; type = OUTPUT_DP; break; + case 0x00000900: link = 1; type = OUTPUT_DP; break; default: NV_ERROR(dev, "PDISP: unknown SOR mc 0x%08x\n", mc); return NULL; @@ -997,7 +1622,8 @@ lookup_dcb(struct drm_device *dev, int id, u32 mc) for (i = 0; i < dev_priv->vbios.dcb.entries; i++) { struct dcb_entry *dcb = &dev_priv->vbios.dcb.entry[i]; - if (dcb->type == type && (dcb->or & (1 << or))) + if (dcb->type == type && (dcb->or & (1 << or)) && + (link < 0 || link == !(dcb->sorconf.link & 1))) return dcb; } @@ -1048,7 +1674,9 @@ nvd0_display_unk2_handler(struct drm_device *dev, u32 crtc, u32 mask) } pclk = nv_rd32(dev, 0x660450 + (crtc * 0x300)) / 1000; - if (mask & 0x00010000) { + NV_DEBUG_KMS(dev, "PDISP: crtc %d pclk %d mask 0x%08x\n", + crtc, pclk, mask); + if (pclk && (mask & 0x00010000)) { nv50_crtc_set_clock(dev, crtc, pclk); } @@ -1072,6 +1700,7 @@ nvd0_display_unk2_handler(struct drm_device *dev, u32 crtc, u32 mask) break; case OUTPUT_TMDS: case OUTPUT_LVDS: + case OUTPUT_DP: if (cfg & 0x00000100) tmp = 0x00000101; else @@ -1122,7 +1751,7 @@ nvd0_display_bh(unsigned long data) { struct drm_device *dev = (struct drm_device *)data; struct nvd0_display *disp = nvd0_display(dev); - u32 mask, crtc; + u32 mask = 0, crtc = ~0; int i; if (drm_debug & (DRM_UT_DRIVER | DRM_UT_KMS)) { @@ -1138,12 +1767,8 @@ nvd0_display_bh(unsigned long data) } } - mask = nv_rd32(dev, 0x6101d4); - crtc = 0; - if (!mask) { - mask = nv_rd32(dev, 0x6109d4); - crtc = 1; - } + while (!mask && ++crtc < dev->mode_config.num_crtc) + mask = nv_rd32(dev, 0x6101d4 + (crtc * 0x800)); if (disp->modeset & 0x00000001) nvd0_display_unk1_handler(dev, crtc, mask); @@ -1158,6 +1783,13 @@ nvd0_display_intr(struct drm_device *dev) { struct nvd0_display *disp = nvd0_display(dev); u32 intr = nv_rd32(dev, 0x610088); + int i; + + if (intr & 0x00000001) { + u32 stat = nv_rd32(dev, 0x61008c); + nv_wr32(dev, 0x61008c, stat); + intr &= ~0x00000001; + } if (intr & 0x00000002) { u32 stat = nv_rd32(dev, 0x61009c); @@ -1196,16 +1828,13 @@ nvd0_display_intr(struct drm_device *dev) intr &= ~0x00100000; } - if (intr & 0x01000000) { - u32 stat = nv_rd32(dev, 0x6100bc); - nv_wr32(dev, 0x6100bc, stat); - intr &= ~0x01000000; - } - - if (intr & 0x02000000) { - u32 stat = nv_rd32(dev, 0x6108bc); - nv_wr32(dev, 0x6108bc, stat); - intr &= ~0x02000000; + for (i = 0; i < dev->mode_config.num_crtc; i++) { + u32 mask = 0x01000000 << i; + if (intr & mask) { + u32 stat = nv_rd32(dev, 0x6100bc + (i * 0x800)); + nv_wr32(dev, 0x6100bc + (i * 0x800), stat); + intr &= ~mask; + } } if (intr) @@ -1215,38 +1844,29 @@ nvd0_display_intr(struct drm_device *dev) /****************************************************************************** * Init *****************************************************************************/ -static void +void nvd0_display_fini(struct drm_device *dev) { int i; - /* fini cursors */ - for (i = 14; i >= 13; i--) { - if (!(nv_rd32(dev, 0x610490 + (i * 0x10)) & 0x00000001)) - continue; - - nv_mask(dev, 0x610490 + (i * 0x10), 0x00000001, 0x00000000); - nv_wait(dev, 0x610490 + (i * 0x10), 0x00010000, 0x00000000); - nv_mask(dev, 0x610090, 1 << i, 0x00000000); - nv_mask(dev, 0x6100a0, 1 << i, 0x00000000); + /* fini cursors + overlays + flips */ + for (i = 1; i >= 0; i--) { + evo_fini_pio(dev, EVO_CURS(i)); + evo_fini_pio(dev, EVO_OIMM(i)); + evo_fini_dma(dev, EVO_OVLY(i)); + evo_fini_dma(dev, EVO_FLIP(i)); } /* fini master */ - if (nv_rd32(dev, 0x610490) & 0x00000010) { - nv_mask(dev, 0x610490, 0x00000010, 0x00000000); - nv_mask(dev, 0x610490, 0x00000003, 0x00000000); - nv_wait(dev, 0x610490, 0x80000000, 0x00000000); - nv_mask(dev, 0x610090, 0x00000001, 0x00000000); - nv_mask(dev, 0x6100a0, 0x00000001, 0x00000000); - } + evo_fini_dma(dev, EVO_MASTER); } int nvd0_display_init(struct drm_device *dev) { struct nvd0_display *disp = nvd0_display(dev); + int ret, i; u32 *push; - int i; if (nv_rd32(dev, 0x6100ac) & 0x00000100) { nv_wr32(dev, 0x6100ac, 0x00000100); @@ -1271,7 +1891,7 @@ nvd0_display_init(struct drm_device *dev) nv_wr32(dev, 0x6301c4 + (i * 0x800), sor); } - for (i = 0; i < 2; i++) { + for (i = 0; i < dev->mode_config.num_crtc; i++) { u32 crtc0 = nv_rd32(dev, 0x616104 + (i * 0x800)); u32 crtc1 = nv_rd32(dev, 0x616108 + (i * 0x800)); u32 crtc2 = nv_rd32(dev, 0x61610c + (i * 0x800)); @@ -1285,36 +1905,24 @@ nvd0_display_init(struct drm_device *dev) nv_mask(dev, 0x6100b0, 0x00000307, 0x00000307); /* init master */ - nv_wr32(dev, 0x610494, (disp->evo[0].handle >> 8) | 3); - nv_wr32(dev, 0x610498, 0x00010000); - nv_wr32(dev, 0x61049c, 0x00000001); - nv_mask(dev, 0x610490, 0x00000010, 0x00000010); - nv_wr32(dev, 0x640000, 0x00000000); - nv_wr32(dev, 0x610490, 0x01000013); - if (!nv_wait(dev, 0x610490, 0x80000000, 0x00000000)) { - NV_ERROR(dev, "PDISP: master 0x%08x\n", - nv_rd32(dev, 0x610490)); - return -EBUSY; + ret = evo_init_dma(dev, EVO_MASTER); + if (ret) + goto error; + + /* init flips + overlays + cursors */ + for (i = 0; i < dev->mode_config.num_crtc; i++) { + if ((ret = evo_init_dma(dev, EVO_FLIP(i))) || + (ret = evo_init_dma(dev, EVO_OVLY(i))) || + (ret = evo_init_pio(dev, EVO_OIMM(i))) || + (ret = evo_init_pio(dev, EVO_CURS(i)))) + goto error; } - nv_mask(dev, 0x610090, 0x00000001, 0x00000001); - nv_mask(dev, 0x6100a0, 0x00000001, 0x00000001); - /* init cursors */ - for (i = 13; i <= 14; i++) { - nv_wr32(dev, 0x610490 + (i * 0x10), 0x00000001); - if (!nv_wait(dev, 0x610490 + (i * 0x10), 0x00010000, 0x00010000)) { - NV_ERROR(dev, "PDISP: curs%d 0x%08x\n", i, - nv_rd32(dev, 0x610490 + (i * 0x10))); - return -EBUSY; - } - - nv_mask(dev, 0x610090, 1 << i, 1 << i); - nv_mask(dev, 0x6100a0, 1 << i, 1 << i); + push = evo_wait(dev, EVO_MASTER, 32); + if (!push) { + ret = -EBUSY; + goto error; } - - push = evo_wait(dev, 0, 32); - if (!push) - return -EBUSY; evo_mthd(push, 0x0088, 1); evo_data(push, NvEvoSync); evo_mthd(push, 0x0084, 1); @@ -1323,9 +1931,12 @@ nvd0_display_init(struct drm_device *dev) evo_data(push, 0x80000000); evo_mthd(push, 0x008c, 1); evo_data(push, 0x00000000); - evo_kick(push, dev, 0); + evo_kick(push, dev, EVO_MASTER); - return 0; +error: + if (ret) + nvd0_display_fini(dev); + return ret; } void @@ -1334,11 +1945,16 @@ nvd0_display_destroy(struct drm_device *dev) struct drm_nouveau_private *dev_priv = dev->dev_private; struct nvd0_display *disp = nvd0_display(dev); struct pci_dev *pdev = dev->pdev; + int i; - nvd0_display_fini(dev); + for (i = 0; i < EVO_DMA_NR; i++) { + struct evo *evo = &disp->evo[i]; + pci_free_consistent(pdev, PAGE_SIZE, evo->ptr, evo->handle); + } - pci_free_consistent(pdev, PAGE_SIZE, disp->evo[0].ptr, disp->evo[0].handle); nouveau_gpuobj_ref(NULL, &disp->mem); + nouveau_bo_unmap(disp->sync); + nouveau_bo_ref(NULL, &disp->sync); nouveau_irq_unregister(dev, 26); dev_priv->engine.display.priv = NULL; @@ -1355,7 +1971,7 @@ nvd0_display_create(struct drm_device *dev) struct pci_dev *pdev = dev->pdev; struct nvd0_display *disp; struct dcb_entry *dcbe; - int ret, i; + int crtcs, ret, i; disp = kzalloc(sizeof(*disp), GFP_KERNEL); if (!disp) @@ -1363,7 +1979,8 @@ nvd0_display_create(struct drm_device *dev) dev_priv->engine.display.priv = disp; /* create crtc objects to represent the hw heads */ - for (i = 0; i < 2; i++) { + crtcs = nv_rd32(dev, 0x022448); + for (i = 0; i < crtcs; i++) { ret = nvd0_crtc_create(dev, i); if (ret) goto out; @@ -1384,6 +2001,7 @@ nvd0_display_create(struct drm_device *dev) switch (dcbe->type) { case OUTPUT_TMDS: case OUTPUT_LVDS: + case OUTPUT_DP: nvd0_sor_create(connector, dcbe); break; case OUTPUT_ANALOG: @@ -1410,61 +2028,83 @@ nvd0_display_create(struct drm_device *dev) tasklet_init(&disp->tasklet, nvd0_display_bh, (unsigned long)dev); nouveau_irq_register(dev, 26, nvd0_display_intr); + /* small shared memory area we use for notifiers and semaphores */ + ret = nouveau_bo_new(dev, 4096, 0x1000, TTM_PL_FLAG_VRAM, + 0, 0x0000, &disp->sync); + if (!ret) { + ret = nouveau_bo_pin(disp->sync, TTM_PL_FLAG_VRAM); + if (!ret) + ret = nouveau_bo_map(disp->sync); + if (ret) + nouveau_bo_ref(NULL, &disp->sync); + } + + if (ret) + goto out; + /* hash table and dma objects for the memory areas we care about */ ret = nouveau_gpuobj_new(dev, NULL, 0x4000, 0x10000, NVOBJ_FLAG_ZERO_ALLOC, &disp->mem); if (ret) goto out; - nv_wo32(disp->mem, 0x1000, 0x00000049); - nv_wo32(disp->mem, 0x1004, (disp->mem->vinst + 0x2000) >> 8); - nv_wo32(disp->mem, 0x1008, (disp->mem->vinst + 0x2fff) >> 8); - nv_wo32(disp->mem, 0x100c, 0x00000000); - nv_wo32(disp->mem, 0x1010, 0x00000000); - nv_wo32(disp->mem, 0x1014, 0x00000000); - nv_wo32(disp->mem, 0x0000, NvEvoSync); - nv_wo32(disp->mem, 0x0004, (0x1000 << 9) | 0x00000001); - - nv_wo32(disp->mem, 0x1020, 0x00000049); - nv_wo32(disp->mem, 0x1024, 0x00000000); - nv_wo32(disp->mem, 0x1028, (dev_priv->vram_size - 1) >> 8); - nv_wo32(disp->mem, 0x102c, 0x00000000); - nv_wo32(disp->mem, 0x1030, 0x00000000); - nv_wo32(disp->mem, 0x1034, 0x00000000); - nv_wo32(disp->mem, 0x0008, NvEvoVRAM); - nv_wo32(disp->mem, 0x000c, (0x1020 << 9) | 0x00000001); - - nv_wo32(disp->mem, 0x1040, 0x00000009); - nv_wo32(disp->mem, 0x1044, 0x00000000); - nv_wo32(disp->mem, 0x1048, (dev_priv->vram_size - 1) >> 8); - nv_wo32(disp->mem, 0x104c, 0x00000000); - nv_wo32(disp->mem, 0x1050, 0x00000000); - nv_wo32(disp->mem, 0x1054, 0x00000000); - nv_wo32(disp->mem, 0x0010, NvEvoVRAM_LP); - nv_wo32(disp->mem, 0x0014, (0x1040 << 9) | 0x00000001); - - nv_wo32(disp->mem, 0x1060, 0x0fe00009); - nv_wo32(disp->mem, 0x1064, 0x00000000); - nv_wo32(disp->mem, 0x1068, (dev_priv->vram_size - 1) >> 8); - nv_wo32(disp->mem, 0x106c, 0x00000000); - nv_wo32(disp->mem, 0x1070, 0x00000000); - nv_wo32(disp->mem, 0x1074, 0x00000000); - nv_wo32(disp->mem, 0x0018, NvEvoFB32); - nv_wo32(disp->mem, 0x001c, (0x1060 << 9) | 0x00000001); - - pinstmem->flush(dev); + /* create evo dma channels */ + for (i = 0; i < EVO_DMA_NR; i++) { + struct evo *evo = &disp->evo[i]; + u64 offset = disp->sync->bo.offset; + u32 dmao = 0x1000 + (i * 0x100); + u32 hash = 0x0000 + (i * 0x040); + + evo->idx = i; + evo->sem.offset = EVO_SYNC(evo->idx, 0x00); + evo->ptr = pci_alloc_consistent(pdev, PAGE_SIZE, &evo->handle); + if (!evo->ptr) { + ret = -ENOMEM; + goto out; + } - /* push buffers for evo channels */ - disp->evo[0].ptr = - pci_alloc_consistent(pdev, PAGE_SIZE, &disp->evo[0].handle); - if (!disp->evo[0].ptr) { - ret = -ENOMEM; - goto out; + nv_wo32(disp->mem, dmao + 0x00, 0x00000049); + nv_wo32(disp->mem, dmao + 0x04, (offset + 0x0000) >> 8); + nv_wo32(disp->mem, dmao + 0x08, (offset + 0x0fff) >> 8); + nv_wo32(disp->mem, dmao + 0x0c, 0x00000000); + nv_wo32(disp->mem, dmao + 0x10, 0x00000000); + nv_wo32(disp->mem, dmao + 0x14, 0x00000000); + nv_wo32(disp->mem, hash + 0x00, NvEvoSync); + nv_wo32(disp->mem, hash + 0x04, 0x00000001 | (i << 27) | + ((dmao + 0x00) << 9)); + + nv_wo32(disp->mem, dmao + 0x20, 0x00000049); + nv_wo32(disp->mem, dmao + 0x24, 0x00000000); + nv_wo32(disp->mem, dmao + 0x28, (dev_priv->vram_size - 1) >> 8); + nv_wo32(disp->mem, dmao + 0x2c, 0x00000000); + nv_wo32(disp->mem, dmao + 0x30, 0x00000000); + nv_wo32(disp->mem, dmao + 0x34, 0x00000000); + nv_wo32(disp->mem, hash + 0x08, NvEvoVRAM); + nv_wo32(disp->mem, hash + 0x0c, 0x00000001 | (i << 27) | + ((dmao + 0x20) << 9)); + + nv_wo32(disp->mem, dmao + 0x40, 0x00000009); + nv_wo32(disp->mem, dmao + 0x44, 0x00000000); + nv_wo32(disp->mem, dmao + 0x48, (dev_priv->vram_size - 1) >> 8); + nv_wo32(disp->mem, dmao + 0x4c, 0x00000000); + nv_wo32(disp->mem, dmao + 0x50, 0x00000000); + nv_wo32(disp->mem, dmao + 0x54, 0x00000000); + nv_wo32(disp->mem, hash + 0x10, NvEvoVRAM_LP); + nv_wo32(disp->mem, hash + 0x14, 0x00000001 | (i << 27) | + ((dmao + 0x40) << 9)); + + nv_wo32(disp->mem, dmao + 0x60, 0x0fe00009); + nv_wo32(disp->mem, dmao + 0x64, 0x00000000); + nv_wo32(disp->mem, dmao + 0x68, (dev_priv->vram_size - 1) >> 8); + nv_wo32(disp->mem, dmao + 0x6c, 0x00000000); + nv_wo32(disp->mem, dmao + 0x70, 0x00000000); + nv_wo32(disp->mem, dmao + 0x74, 0x00000000); + nv_wo32(disp->mem, hash + 0x18, NvEvoFB32); + nv_wo32(disp->mem, hash + 0x1c, 0x00000001 | (i << 27) | + ((dmao + 0x60) << 9)); } - ret = nvd0_display_init(dev); - if (ret) - goto out; + pinstmem->flush(dev); out: if (ret) diff --git a/drivers/gpu/drm/r128/r128_drv.c b/drivers/gpu/drm/r128/r128_drv.c index 4c8796b..88718fa 100644 --- a/drivers/gpu/drm/r128/r128_drv.c +++ b/drivers/gpu/drm/r128/r128_drv.c @@ -42,6 +42,20 @@ static struct pci_device_id pciidlist[] = { r128_PCI_IDS }; +static const struct file_operations r128_driver_fops = { + .owner = THIS_MODULE, + .open = drm_open, + .release = drm_release, + .unlocked_ioctl = drm_ioctl, + .mmap = drm_mmap, + .poll = drm_poll, + .fasync = drm_fasync, +#ifdef CONFIG_COMPAT + .compat_ioctl = r128_compat_ioctl, +#endif + .llseek = noop_llseek, +}; + static struct drm_driver driver = { .driver_features = DRIVER_USE_AGP | DRIVER_USE_MTRR | DRIVER_PCI_DMA | DRIVER_SG | @@ -60,21 +74,7 @@ static struct drm_driver driver = { .reclaim_buffers = drm_core_reclaim_buffers, .ioctls = r128_ioctls, .dma_ioctl = r128_cce_buffers, - .fops = { - .owner = THIS_MODULE, - .open = drm_open, - .release = drm_release, - .unlocked_ioctl = drm_ioctl, - .mmap = drm_mmap, - .poll = drm_poll, - .fasync = drm_fasync, -#ifdef CONFIG_COMPAT - .compat_ioctl = r128_compat_ioctl, -#endif - .llseek = noop_llseek, - }, - - + .fops = &r128_driver_fops, .name = DRIVER_NAME, .desc = DRIVER_DESC, .date = DRIVER_DATE, @@ -85,6 +85,7 @@ static struct drm_driver driver = { int r128_driver_load(struct drm_device *dev, unsigned long flags) { + pci_set_master(dev->pdev); return drm_vblank_init(dev, 1); } diff --git a/drivers/gpu/drm/radeon/Makefile b/drivers/gpu/drm/radeon/Makefile index cf8b4bc..9d83729 100644 --- a/drivers/gpu/drm/radeon/Makefile +++ b/drivers/gpu/drm/radeon/Makefile @@ -70,7 +70,8 @@ radeon-y += radeon_device.o radeon_asic.o radeon_kms.o \ r200.o radeon_legacy_tv.o r600_cs.o r600_blit.o r600_blit_shaders.o \ r600_blit_kms.o radeon_pm.o atombios_dp.o r600_audio.o r600_hdmi.o \ evergreen.o evergreen_cs.o evergreen_blit_shaders.o evergreen_blit_kms.o \ - radeon_trace_points.o ni.o cayman_blit_shaders.o atombios_encoders.o + radeon_trace_points.o ni.o cayman_blit_shaders.o atombios_encoders.o \ + radeon_semaphore.o radeon_sa.o atombios_i2c.o si.o si_blit_shaders.o radeon-$(CONFIG_COMPAT) += radeon_ioc32.o radeon-$(CONFIG_VGA_SWITCHEROO) += radeon_atpx_handler.o @@ -78,4 +79,4 @@ radeon-$(CONFIG_ACPI) += radeon_acpi.o obj-$(CONFIG_DRM_RADEON)+= radeon.o -CFLAGS_radeon_trace_points.o := -I$(src) \ No newline at end of file +CFLAGS_radeon_trace_points.o := -I$(src) diff --git a/drivers/gpu/drm/radeon/ObjectID.h b/drivers/gpu/drm/radeon/ObjectID.h index c61c3fe..ca4b038 100644 --- a/drivers/gpu/drm/radeon/ObjectID.h +++ b/drivers/gpu/drm/radeon/ObjectID.h @@ -85,6 +85,7 @@ #define ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA 0x1F #define ENCODER_OBJECT_ID_INTERNAL_UNIPHY1 0x20 #define ENCODER_OBJECT_ID_INTERNAL_UNIPHY2 0x21 +#define ENCODER_OBJECT_ID_INTERNAL_VCE 0x24 #define ENCODER_OBJECT_ID_GENERAL_EXTERNAL_DVO 0xFF @@ -387,6 +388,10 @@ GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\ ENCODER_OBJECT_ID_NUTMEG << OBJECT_ID_SHIFT) +#define ENCODER_VCE_ENUM_ID1 ( GRAPH_OBJECT_TYPE_ENCODER << OBJECT_TYPE_SHIFT |\ + GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\ + ENCODER_OBJECT_ID_INTERNAL_VCE << OBJECT_ID_SHIFT) + /****************************************************/ /* Connector Object ID definition - Shared with BIOS */ /****************************************************/ diff --git a/drivers/gpu/drm/radeon/atom.c b/drivers/gpu/drm/radeon/atom.c index 3a05cdb..5ce9bf5 100644 --- a/drivers/gpu/drm/radeon/atom.c +++ b/drivers/gpu/drm/radeon/atom.c @@ -665,6 +665,8 @@ static void atom_op_delay(atom_exec_context *ctx, int *ptr, int arg) SDEBUG(" count: %d\n", count); if (arg == ATOM_UNIT_MICROSEC) udelay(count); + else if (!drm_can_sleep()) + mdelay(count); else msleep(count); } diff --git a/drivers/gpu/drm/radeon/atombios.h b/drivers/gpu/drm/radeon/atombios.h index 4760466..4b04ba3 100644 --- a/drivers/gpu/drm/radeon/atombios.h +++ b/drivers/gpu/drm/radeon/atombios.h @@ -7270,6 +7270,8 @@ typedef struct _ATOM_PPLIB_THERMALCONTROLLER #define ATOM_PP_THERMALCONTROLLER_EMC2103 13 /* 0x0D */ // Only fan control will be implemented, do NOT show this in PPGen. #define ATOM_PP_THERMALCONTROLLER_SUMO 14 /* 0x0E */ // Sumo type, used internally #define ATOM_PP_THERMALCONTROLLER_NISLANDS 15 +#define ATOM_PP_THERMALCONTROLLER_SISLANDS 16 +#define ATOM_PP_THERMALCONTROLLER_LM96163 17 // Thermal controller 'combo type' to use an external controller for Fan control and an internal controller for thermal. // We probably should reserve the bit 0x80 for this use. @@ -7285,6 +7287,7 @@ typedef struct _ATOM_PPLIB_STATE UCHAR ucClockStateIndices[1]; // variable-sized } ATOM_PPLIB_STATE; + typedef struct _ATOM_PPLIB_FANTABLE { UCHAR ucFanTableFormat; // Change this if the table format changes or version changes so that the other fields are not the same. @@ -7297,12 +7300,20 @@ typedef struct _ATOM_PPLIB_FANTABLE USHORT usPWMHigh; // The PWM value at THigh. } ATOM_PPLIB_FANTABLE; +typedef struct _ATOM_PPLIB_FANTABLE2 +{ + ATOM_PPLIB_FANTABLE basicTable; + USHORT usTMax; // The max temperature +} ATOM_PPLIB_FANTABLE2; + typedef struct _ATOM_PPLIB_EXTENDEDHEADER { USHORT usSize; ULONG ulMaxEngineClock; // For Overdrive. ULONG ulMaxMemoryClock; // For Overdrive. // Add extra system parameters here, always adjust size to include all fields. + USHORT usVCETableOffset; //points to ATOM_PPLIB_VCE_Table + USHORT usUVDTableOffset; //points to ATOM_PPLIB_UVD_Table } ATOM_PPLIB_EXTENDEDHEADER; //// ATOM_PPLIB_POWERPLAYTABLE::ulPlatformCaps @@ -7325,6 +7336,7 @@ typedef struct _ATOM_PPLIB_EXTENDEDHEADER #define ATOM_PP_PLATFORM_CAP_REGULATOR_HOT 0x00010000 // Enable the 'regulator hot' feature. #define ATOM_PP_PLATFORM_CAP_BACO 0x00020000 // Does the driver supports BACO state. + typedef struct _ATOM_PPLIB_POWERPLAYTABLE { ATOM_COMMON_TABLE_HEADER sHeader; @@ -7383,7 +7395,8 @@ typedef struct _ATOM_PPLIB_POWERPLAYTABLE4 USHORT usVddciDependencyOnMCLKOffset; USHORT usVddcDependencyOnMCLKOffset; USHORT usMaxClockVoltageOnDCOffset; - USHORT usReserved[2]; + USHORT usVddcPhaseShedLimitsTableOffset; // Points to ATOM_PPLIB_PhaseSheddingLimits_Table + USHORT usReserved; } ATOM_PPLIB_POWERPLAYTABLE4, *LPATOM_PPLIB_POWERPLAYTABLE4; typedef struct _ATOM_PPLIB_POWERPLAYTABLE5 @@ -7393,8 +7406,9 @@ typedef struct _ATOM_PPLIB_POWERPLAYTABLE5 ULONG ulNearTDPLimit; ULONG ulSQRampingThreshold; USHORT usCACLeakageTableOffset; // Points to ATOM_PPLIB_CAC_Leakage_Table - ULONG ulCACLeakage; // TBD, this parameter is still under discussion. Change to ulReserved if not needed. - ULONG ulReserved; + ULONG ulCACLeakage; // The iLeakage for driver calculated CAC leakage table + USHORT usTDPODLimit; + USHORT usLoadLineSlope; // in milliOhms * 100 } ATOM_PPLIB_POWERPLAYTABLE5, *LPATOM_PPLIB_POWERPLAYTABLE5; //// ATOM_PPLIB_NONCLOCK_INFO::usClassification @@ -7423,6 +7437,7 @@ typedef struct _ATOM_PPLIB_POWERPLAYTABLE5 //// ATOM_PPLIB_NONCLOCK_INFO::usClassification2 #define ATOM_PPLIB_CLASSIFICATION2_LIMITEDPOWERSOURCE_2 0x0001 #define ATOM_PPLIB_CLASSIFICATION2_ULV 0x0002 +#define ATOM_PPLIB_CLASSIFICATION2_MVC 0x0004 //Multi-View Codec (BD-3D) //// ATOM_PPLIB_NONCLOCK_INFO::ulCapsAndSettings #define ATOM_PPLIB_SINGLE_DISPLAY_ONLY 0x00000001 @@ -7446,7 +7461,9 @@ typedef struct _ATOM_PPLIB_POWERPLAYTABLE5 #define ATOM_PPLIB_SOFTWARE_DISABLE_LOADBALANCING 0x00001000 #define ATOM_PPLIB_SOFTWARE_ENABLE_SLEEP_FOR_TIMESTAMPS 0x00002000 -#define ATOM_PPLIB_DISALLOW_ON_DC 0x00004000 + +#define ATOM_PPLIB_DISALLOW_ON_DC 0x00004000 + #define ATOM_PPLIB_ENABLE_VARIBRIGHT 0x00008000 //memory related flags @@ -7508,7 +7525,7 @@ typedef struct _ATOM_PPLIB_R600_CLOCK_INFO #define ATOM_PPLIB_R600_FLAGS_UVDSAFE 2 #define ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE 4 #define ATOM_PPLIB_R600_FLAGS_MEMORY_ODT_OFF 8 -#define ATOM_PPLIB_R600_FLAGS_MEMORY_DLL_OFF 16 +#define ATOM_PPLIB_R600_FLAGS_MEMORY_DLL_OFF 16 #define ATOM_PPLIB_R600_FLAGS_LOWPOWER 32 // On the RV770 use 'low power' setting (sequencer S0). typedef struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO @@ -7527,6 +7544,24 @@ typedef struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO } ATOM_PPLIB_EVERGREEN_CLOCK_INFO; +typedef struct _ATOM_PPLIB_SI_CLOCK_INFO +{ + USHORT usEngineClockLow; + UCHAR ucEngineClockHigh; + + USHORT usMemoryClockLow; + UCHAR ucMemoryClockHigh; + + USHORT usVDDC; + USHORT usVDDCI; + UCHAR ucPCIEGen; + UCHAR ucUnused1; + + ULONG ulFlags; // ATOM_PPLIB_SI_FLAGS_*, no flag is necessary for now + +} ATOM_PPLIB_SI_CLOCK_INFO; + + typedef struct _ATOM_PPLIB_RS780_CLOCK_INFO { @@ -7539,7 +7574,7 @@ typedef struct _ATOM_PPLIB_RS780_CLOCK_INFO UCHAR ucPadding; // For proper alignment and size. USHORT usVDDC; // For the 780, use: None, Low, High, Variable UCHAR ucMaxHTLinkWidth; // From SBIOS - {2, 4, 8, 16} - UCHAR ucMinHTLinkWidth; // From SBIOS - {2, 4, 8, 16}. Effective only if CDLW enabled. Minimum down stream width could be bigger as display BW requirement. + UCHAR ucMinHTLinkWidth; // From SBIOS - {2, 4, 8, 16}. Effective only if CDLW enabled. Minimum down stream width could be bigger as display BW requriement. USHORT usHTLinkFreq; // See definition ATOM_PPLIB_RS780_HTLINKFREQ_xxx or in MHz(>=200). ULONG ulFlags; } ATOM_PPLIB_RS780_CLOCK_INFO; @@ -7561,9 +7596,7 @@ typedef struct _ATOM_PPLIB_SUMO_CLOCK_INFO{ USHORT usEngineClockLow; //clockfrequency & 0xFFFF. The unit is in 10khz UCHAR ucEngineClockHigh; //clockfrequency >> 16. UCHAR vddcIndex; //2-bit vddc index; - UCHAR leakage; //please use 8-bit absolute value, not the 6-bit % value - //please initalize to 0 - UCHAR rsv; + USHORT tdpLimit; //please initalize to 0 USHORT rsv1; //please initialize to 0s @@ -7586,7 +7619,7 @@ typedef struct _ATOM_PPLIB_STATE_V2 UCHAR clockInfoIndex[1]; } ATOM_PPLIB_STATE_V2; -typedef struct StateArray{ +typedef struct _StateArray{ //how many states we have UCHAR ucNumEntries; @@ -7594,18 +7627,17 @@ typedef struct StateArray{ }StateArray; -typedef struct ClockInfoArray{ +typedef struct _ClockInfoArray{ //how many clock levels we have UCHAR ucNumEntries; - //sizeof(ATOM_PPLIB_SUMO_CLOCK_INFO) + //sizeof(ATOM_PPLIB_CLOCK_INFO) UCHAR ucEntrySize; - //this is for Sumo - ATOM_PPLIB_SUMO_CLOCK_INFO clockInfo[1]; + UCHAR clockInfo[1]; }ClockInfoArray; -typedef struct NonClockInfoArray{ +typedef struct _NonClockInfoArray{ //how many non-clock levels we have. normally should be same as number of states UCHAR ucNumEntries; @@ -7644,6 +7676,124 @@ typedef struct _ATOM_PPLIB_Clock_Voltage_Limit_Table ATOM_PPLIB_Clock_Voltage_Limit_Record entries[1]; // Dynamically allocate entries. }ATOM_PPLIB_Clock_Voltage_Limit_Table; +typedef struct _ATOM_PPLIB_CAC_Leakage_Record +{ + USHORT usVddc; // We use this field for the "fake" standardized VDDC for power calculations + ULONG ulLeakageValue; +}ATOM_PPLIB_CAC_Leakage_Record; + +typedef struct _ATOM_PPLIB_CAC_Leakage_Table +{ + UCHAR ucNumEntries; // Number of entries. + ATOM_PPLIB_CAC_Leakage_Record entries[1]; // Dynamically allocate entries. +}ATOM_PPLIB_CAC_Leakage_Table; + +typedef struct _ATOM_PPLIB_PhaseSheddingLimits_Record +{ + USHORT usVoltage; + USHORT usSclkLow; + UCHAR ucSclkHigh; + USHORT usMclkLow; + UCHAR ucMclkHigh; +}ATOM_PPLIB_PhaseSheddingLimits_Record; + +typedef struct _ATOM_PPLIB_PhaseSheddingLimits_Table +{ + UCHAR ucNumEntries; // Number of entries. + ATOM_PPLIB_PhaseSheddingLimits_Record entries[1]; // Dynamically allocate entries. +}ATOM_PPLIB_PhaseSheddingLimits_Table; + +typedef struct _VCEClockInfo{ + USHORT usEVClkLow; + UCHAR ucEVClkHigh; + USHORT usECClkLow; + UCHAR ucECClkHigh; +}VCEClockInfo; + +typedef struct _VCEClockInfoArray{ + UCHAR ucNumEntries; + VCEClockInfo entries[1]; +}VCEClockInfoArray; + +typedef struct _ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record +{ + USHORT usVoltage; + UCHAR ucVCEClockInfoIndex; +}ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record; + +typedef struct _ATOM_PPLIB_VCE_Clock_Voltage_Limit_Table +{ + UCHAR numEntries; + ATOM_PPLIB_VCE_Clock_Voltage_Limit_Record entries[1]; +}ATOM_PPLIB_VCE_Clock_Voltage_Limit_Table; + +typedef struct _ATOM_PPLIB_VCE_State_Record +{ + UCHAR ucVCEClockInfoIndex; + UCHAR ucClockInfoIndex; //highest 2 bits indicates memory p-states, lower 6bits indicates index to ClockInfoArrary +}ATOM_PPLIB_VCE_State_Record; + +typedef struct _ATOM_PPLIB_VCE_State_Table +{ + UCHAR numEntries; + ATOM_PPLIB_VCE_State_Record entries[1]; +}ATOM_PPLIB_VCE_State_Table; + + +typedef struct _ATOM_PPLIB_VCE_Table +{ + UCHAR revid; +// VCEClockInfoArray array; +// ATOM_PPLIB_VCE_Clock_Voltage_Limit_Table limits; +// ATOM_PPLIB_VCE_State_Table states; +}ATOM_PPLIB_VCE_Table; + + +typedef struct _UVDClockInfo{ + USHORT usVClkLow; + UCHAR ucVClkHigh; + USHORT usDClkLow; + UCHAR ucDClkHigh; +}UVDClockInfo; + +typedef struct _UVDClockInfoArray{ + UCHAR ucNumEntries; + UVDClockInfo entries[1]; +}UVDClockInfoArray; + +typedef struct _ATOM_PPLIB_UVD_Clock_Voltage_Limit_Record +{ + USHORT usVoltage; + UCHAR ucUVDClockInfoIndex; +}ATOM_PPLIB_UVD_Clock_Voltage_Limit_Record; + +typedef struct _ATOM_PPLIB_UVD_Clock_Voltage_Limit_Table +{ + UCHAR numEntries; + ATOM_PPLIB_UVD_Clock_Voltage_Limit_Record entries[1]; +}ATOM_PPLIB_UVD_Clock_Voltage_Limit_Table; + +typedef struct _ATOM_PPLIB_UVD_State_Record +{ + UCHAR ucUVDClockInfoIndex; + UCHAR ucClockInfoIndex; //highest 2 bits indicates memory p-states, lower 6bits indicates index to ClockInfoArrary +}ATOM_PPLIB_UVD_State_Record; + +typedef struct _ATOM_PPLIB_UVD_State_Table +{ + UCHAR numEntries; + ATOM_PPLIB_UVD_State_Record entries[1]; +}ATOM_PPLIB_UVD_State_Table; + + +typedef struct _ATOM_PPLIB_UVD_Table +{ + UCHAR revid; +// UVDClockInfoArray array; +// ATOM_PPLIB_UVD_Clock_Voltage_Limit_Table limits; +// ATOM_PPLIB_UVD_State_Table states; +}ATOM_PPLIB_UVD_Table; + /**************************************************************************/ diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c index a25d08a..15594a3 100644 --- a/drivers/gpu/drm/radeon/atombios_crtc.c +++ b/drivers/gpu/drm/radeon/atombios_crtc.c @@ -231,6 +231,22 @@ static void atombios_blank_crtc(struct drm_crtc *crtc, int state) atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); } +static void atombios_powergate_crtc(struct drm_crtc *crtc, int state) +{ + struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); + struct drm_device *dev = crtc->dev; + struct radeon_device *rdev = dev->dev_private; + int index = GetIndexIntoMasterTable(COMMAND, EnableDispPowerGating); + ENABLE_DISP_POWER_GATING_PARAMETERS_V2_1 args; + + memset(&args, 0, sizeof(args)); + + args.ucDispPipeId = radeon_crtc->crtc_id; + args.ucEnable = state; + + atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); +} + void atombios_crtc_dpms(struct drm_crtc *crtc, int mode) { struct drm_device *dev = crtc->dev; @@ -242,8 +258,10 @@ void atombios_crtc_dpms(struct drm_crtc *crtc, int mode) radeon_crtc->enabled = true; /* adjust pm to dpms changes BEFORE enabling crtcs */ radeon_pm_compute_clocks(rdev); + if (ASIC_IS_DCE6(rdev) && !radeon_crtc->in_mode_set) + atombios_powergate_crtc(crtc, ATOM_DISABLE); atombios_enable_crtc(crtc, ATOM_ENABLE); - if (ASIC_IS_DCE3(rdev)) + if (ASIC_IS_DCE3(rdev) && !ASIC_IS_DCE6(rdev)) atombios_enable_crtc_memreq(crtc, ATOM_ENABLE); atombios_blank_crtc(crtc, ATOM_DISABLE); drm_vblank_post_modeset(dev, radeon_crtc->crtc_id); @@ -255,10 +273,12 @@ void atombios_crtc_dpms(struct drm_crtc *crtc, int mode) drm_vblank_pre_modeset(dev, radeon_crtc->crtc_id); if (radeon_crtc->enabled) atombios_blank_crtc(crtc, ATOM_ENABLE); - if (ASIC_IS_DCE3(rdev)) + if (ASIC_IS_DCE3(rdev) && !ASIC_IS_DCE6(rdev)) atombios_enable_crtc_memreq(crtc, ATOM_DISABLE); atombios_enable_crtc(crtc, ATOM_DISABLE); radeon_crtc->enabled = false; + if (ASIC_IS_DCE6(rdev) && !radeon_crtc->in_mode_set) + atombios_powergate_crtc(crtc, ATOM_ENABLE); /* adjust pm to dpms changes AFTER disabling crtcs */ radeon_pm_compute_clocks(rdev); break; @@ -355,15 +375,12 @@ static void atombios_crtc_set_timing(struct drm_crtc *crtc, atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); } -static void atombios_disable_ss(struct drm_crtc *crtc) +static void atombios_disable_ss(struct radeon_device *rdev, int pll_id) { - struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); - struct drm_device *dev = crtc->dev; - struct radeon_device *rdev = dev->dev_private; u32 ss_cntl; if (ASIC_IS_DCE4(rdev)) { - switch (radeon_crtc->pll_id) { + switch (pll_id) { case ATOM_PPLL1: ss_cntl = RREG32(EVERGREEN_P1PLL_SS_CNTL); ss_cntl &= ~EVERGREEN_PxPLL_SS_EN; @@ -379,7 +396,7 @@ static void atombios_disable_ss(struct drm_crtc *crtc) return; } } else if (ASIC_IS_AVIVO(rdev)) { - switch (radeon_crtc->pll_id) { + switch (pll_id) { case ATOM_PPLL1: ss_cntl = RREG32(AVIVO_P1PLL_INT_SS_CNTL); ss_cntl &= ~1; @@ -406,16 +423,31 @@ union atom_enable_ss { ENABLE_SPREAD_SPECTRUM_ON_PPLL_V3 v3; }; -static void atombios_crtc_program_ss(struct drm_crtc *crtc, +static void atombios_crtc_program_ss(struct radeon_device *rdev, int enable, int pll_id, + int crtc_id, struct radeon_atom_ss *ss) { - struct drm_device *dev = crtc->dev; - struct radeon_device *rdev = dev->dev_private; + unsigned i; int index = GetIndexIntoMasterTable(COMMAND, EnableSpreadSpectrumOnPPLL); union atom_enable_ss args; + if (!enable) { + for (i = 0; i < rdev->num_crtc; i++) { + if (rdev->mode_info.crtcs[i] && + rdev->mode_info.crtcs[i]->enabled && + i != crtc_id && + pll_id == rdev->mode_info.crtcs[i]->pll_id) { + /* one other crtc is using this pll don't turn + * off spread spectrum as it might turn off + * display on active crtc + */ + return; + } + } + } + memset(&args, 0, sizeof(args)); if (ASIC_IS_DCE5(rdev)) { @@ -441,7 +473,7 @@ static void atombios_crtc_program_ss(struct drm_crtc *crtc, return; } args.v3.ucEnable = enable; - if ((ss->percentage == 0) || (ss->type & ATOM_EXTERNAL_SS_MASK)) + if ((ss->percentage == 0) || (ss->type & ATOM_EXTERNAL_SS_MASK) || ASIC_IS_DCE61(rdev)) args.v3.ucEnable = ATOM_DISABLE; } else if (ASIC_IS_DCE4(rdev)) { args.v2.usSpreadSpectrumPercentage = cpu_to_le16(ss->percentage); @@ -479,7 +511,7 @@ static void atombios_crtc_program_ss(struct drm_crtc *crtc, } else if (ASIC_IS_AVIVO(rdev)) { if ((enable == ATOM_DISABLE) || (ss->percentage == 0) || (ss->type & ATOM_EXTERNAL_SS_MASK)) { - atombios_disable_ss(crtc); + atombios_disable_ss(rdev, pll_id); return; } args.lvds_ss_2.usSpreadSpectrumPercentage = cpu_to_le16(ss->percentage); @@ -491,7 +523,7 @@ static void atombios_crtc_program_ss(struct drm_crtc *crtc, } else { if ((enable == ATOM_DISABLE) || (ss->percentage == 0) || (ss->type & ATOM_EXTERNAL_SS_MASK)) { - atombios_disable_ss(crtc); + atombios_disable_ss(rdev, pll_id); return; } args.lvds_ss.usSpreadSpectrumPercentage = cpu_to_le16(ss->percentage); @@ -523,6 +555,7 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc, int encoder_mode = 0; u32 dp_clock = mode->clock; int bpc = 8; + bool is_duallink = false; /* reset the pll flags */ pll->flags = 0; @@ -542,9 +575,7 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc, if (rdev->family < CHIP_RV770) pll->flags |= RADEON_PLL_PREFER_MINM_OVER_MAXP; /* use frac fb div on APUs */ - if (ASIC_IS_DCE41(rdev)) - pll->flags |= RADEON_PLL_USE_FRAC_FB_DIV; - if (ASIC_IS_DCE32(rdev) && mode->clock > 165000) + if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE61(rdev)) pll->flags |= RADEON_PLL_USE_FRAC_FB_DIV; } else { pll->flags |= RADEON_PLL_LEGACY; @@ -559,9 +590,10 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc, if (encoder->crtc == crtc) { radeon_encoder = to_radeon_encoder(encoder); connector = radeon_get_connector_for_encoder(encoder); - if (connector) - bpc = connector->display_info.bpc; + /* if (connector && connector->display_info.bpc) + bpc = connector->display_info.bpc; */ encoder_mode = atombios_get_encoder_mode(encoder); + is_duallink = radeon_dig_monitor_is_duallink(encoder, mode->clock); if ((radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT | ATOM_DEVICE_DFP_SUPPORT)) || (radeon_encoder_get_dp_bridge_encoder_id(encoder) != ENCODER_OBJECT_ID_NONE)) { if (connector) { @@ -657,7 +689,7 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc, if (dig->coherent_mode) args.v3.sInput.ucDispPllConfig |= DISPPLL_CONFIG_COHERENT_MODE; - if (mode->clock > 165000) + if (is_duallink) args.v3.sInput.ucDispPllConfig |= DISPPLL_CONFIG_DUAL_LINK; } @@ -707,11 +739,9 @@ union set_pixel_clock { /* on DCE5, make sure the voltage is high enough to support the * required disp clk. */ -static void atombios_crtc_set_dcpll(struct drm_crtc *crtc, +static void atombios_crtc_set_disp_eng_pll(struct radeon_device *rdev, u32 dispclk) { - struct drm_device *dev = crtc->dev; - struct radeon_device *rdev = dev->dev_private; u8 frev, crev; int index; union set_pixel_clock args; @@ -739,7 +769,12 @@ static void atombios_crtc_set_dcpll(struct drm_crtc *crtc, * SetPixelClock provides the dividers */ args.v6.ulDispEngClkFreq = cpu_to_le32(dispclk); - args.v6.ucPpll = ATOM_DCPLL; + if (ASIC_IS_DCE61(rdev)) + args.v6.ucPpll = ATOM_EXT_PLL1; + else if (ASIC_IS_DCE6(rdev)) + args.v6.ucPpll = ATOM_PPLL0; + else + args.v6.ucPpll = ATOM_DCPLL; break; default: DRM_ERROR("Unknown table version %d %d\n", frev, crev); @@ -932,7 +967,9 @@ static void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode struct radeon_connector_atom_dig *dig_connector = radeon_connector->con_priv; int dp_clock; - bpc = connector->display_info.bpc; + + /* if (connector->display_info.bpc) + bpc = connector->display_info.bpc; */ switch (encoder_mode) { case ATOM_ENCODER_MODE_DP_MST: @@ -1001,7 +1038,7 @@ static void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode radeon_compute_pll_legacy(pll, adjusted_clock, &pll_clock, &fb_div, &frac_fb_div, &ref_div, &post_div); - atombios_crtc_program_ss(crtc, ATOM_DISABLE, radeon_crtc->pll_id, &ss); + atombios_crtc_program_ss(rdev, ATOM_DISABLE, radeon_crtc->pll_id, radeon_crtc->crtc_id, &ss); atombios_crtc_program_pll(crtc, radeon_crtc->crtc_id, radeon_crtc->pll_id, encoder_mode, radeon_encoder->encoder_id, mode->clock, @@ -1024,7 +1061,7 @@ static void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode ss.step = step_size; } - atombios_crtc_program_ss(crtc, ATOM_ENABLE, radeon_crtc->pll_id, &ss); + atombios_crtc_program_ss(rdev, ATOM_ENABLE, radeon_crtc->pll_id, radeon_crtc->crtc_id, &ss); } } @@ -1041,6 +1078,7 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc, struct radeon_bo *rbo; uint64_t fb_location; uint32_t fb_format, fb_pitch_pixels, tiling_flags; + unsigned bankw, bankh, mtaspect, tile_split; u32 fb_swap = EVERGREEN_GRPH_ENDIAN_SWAP(EVERGREEN_GRPH_ENDIAN_NONE); u32 tmp, viewport_w, viewport_h; int r; @@ -1131,20 +1169,13 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc, break; } - switch ((tmp & 0xf000) >> 12) { - case 0: /* 1KB rows */ - default: - fb_format |= EVERGREEN_GRPH_TILE_SPLIT(EVERGREEN_ADDR_SURF_TILE_SPLIT_1KB); - break; - case 1: /* 2KB rows */ - fb_format |= EVERGREEN_GRPH_TILE_SPLIT(EVERGREEN_ADDR_SURF_TILE_SPLIT_2KB); - break; - case 2: /* 4KB rows */ - fb_format |= EVERGREEN_GRPH_TILE_SPLIT(EVERGREEN_ADDR_SURF_TILE_SPLIT_4KB); - break; - } - fb_format |= EVERGREEN_GRPH_ARRAY_MODE(EVERGREEN_GRPH_ARRAY_2D_TILED_THIN1); + + evergreen_tiling_fields(tiling_flags, &bankw, &bankh, &mtaspect, &tile_split); + fb_format |= EVERGREEN_GRPH_TILE_SPLIT(tile_split); + fb_format |= EVERGREEN_GRPH_BANK_WIDTH(bankw); + fb_format |= EVERGREEN_GRPH_BANK_HEIGHT(bankh); + fb_format |= EVERGREEN_GRPH_MACRO_TILE_ASPECT(mtaspect); } else if (tiling_flags & RADEON_TILING_MICRO) fb_format |= EVERGREEN_GRPH_ARRAY_MODE(EVERGREEN_GRPH_ARRAY_1D_TILED_THIN1); @@ -1189,7 +1220,7 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc, WREG32(EVERGREEN_GRPH_X_END + radeon_crtc->crtc_offset, target_fb->width); WREG32(EVERGREEN_GRPH_Y_END + radeon_crtc->crtc_offset, target_fb->height); - fb_pitch_pixels = target_fb->pitch / (target_fb->bits_per_pixel / 8); + fb_pitch_pixels = target_fb->pitches[0] / (target_fb->bits_per_pixel / 8); WREG32(EVERGREEN_GRPH_PITCH + radeon_crtc->crtc_offset, fb_pitch_pixels); WREG32(EVERGREEN_GRPH_ENABLE + radeon_crtc->crtc_offset, 1); @@ -1358,7 +1389,7 @@ static int avivo_crtc_do_set_base(struct drm_crtc *crtc, WREG32(AVIVO_D1GRPH_X_END + radeon_crtc->crtc_offset, target_fb->width); WREG32(AVIVO_D1GRPH_Y_END + radeon_crtc->crtc_offset, target_fb->height); - fb_pitch_pixels = target_fb->pitch / (target_fb->bits_per_pixel / 8); + fb_pitch_pixels = target_fb->pitches[0] / (target_fb->bits_per_pixel / 8); WREG32(AVIVO_D1GRPH_PITCH + radeon_crtc->crtc_offset, fb_pitch_pixels); WREG32(AVIVO_D1GRPH_ENABLE + radeon_crtc->crtc_offset, 1); @@ -1460,7 +1491,36 @@ static int radeon_atom_pick_pll(struct drm_crtc *crtc) struct drm_crtc *test_crtc; uint32_t pll_in_use = 0; - if (ASIC_IS_DCE4(rdev)) { + if (ASIC_IS_DCE61(rdev)) { + list_for_each_entry(test_encoder, &dev->mode_config.encoder_list, head) { + if (test_encoder->crtc && (test_encoder->crtc == crtc)) { + struct radeon_encoder *test_radeon_encoder = + to_radeon_encoder(test_encoder); + struct radeon_encoder_atom_dig *dig = + test_radeon_encoder->enc_priv; + + if ((test_radeon_encoder->encoder_id == + ENCODER_OBJECT_ID_INTERNAL_UNIPHY) && + (dig->linkb == false)) /* UNIPHY A uses PPLL2 */ + return ATOM_PPLL2; + } + } + /* UNIPHY B/C/D/E/F */ + list_for_each_entry(test_crtc, &dev->mode_config.crtc_list, head) { + struct radeon_crtc *radeon_test_crtc; + + if (crtc == test_crtc) + continue; + + radeon_test_crtc = to_radeon_crtc(test_crtc); + if ((radeon_test_crtc->pll_id == ATOM_PPLL0) || + (radeon_test_crtc->pll_id == ATOM_PPLL1)) + pll_in_use |= (1 << radeon_test_crtc->pll_id); + } + if (!(pll_in_use & 4)) + return ATOM_PPLL0; + return ATOM_PPLL1; + } else if (ASIC_IS_DCE4(rdev)) { list_for_each_entry(test_encoder, &dev->mode_config.encoder_list, head) { if (test_encoder->crtc && (test_encoder->crtc == crtc)) { /* in DP mode, the DP ref clock can come from PPLL, DCPLL, or ext clock, @@ -1475,6 +1535,8 @@ static int radeon_atom_pick_pll(struct drm_crtc *crtc) if (ENCODER_MODE_IS_DP(atombios_get_encoder_mode(test_encoder))) { if (rdev->clock.dp_extclk) return ATOM_PPLL_INVALID; + else if (ASIC_IS_DCE6(rdev)) + return ATOM_PPLL0; else if (ASIC_IS_DCE5(rdev)) return ATOM_DCPLL; } @@ -1501,6 +1563,26 @@ static int radeon_atom_pick_pll(struct drm_crtc *crtc) } +void radeon_atom_disp_eng_pll_init(struct radeon_device *rdev) +{ + /* always set DCPLL */ + if (ASIC_IS_DCE6(rdev)) + atombios_crtc_set_disp_eng_pll(rdev, rdev->clock.default_dispclk); + else if (ASIC_IS_DCE4(rdev)) { + struct radeon_atom_ss ss; + bool ss_enabled = radeon_atombios_get_asic_ss_info(rdev, &ss, + ASIC_INTERNAL_SS_ON_DCPLL, + rdev->clock.default_dispclk); + if (ss_enabled) + atombios_crtc_program_ss(rdev, ATOM_DISABLE, ATOM_DCPLL, -1, &ss); + /* XXX: DCE5, make sure voltage, dispclk is high enough */ + atombios_crtc_set_disp_eng_pll(rdev, rdev->clock.default_dispclk); + if (ss_enabled) + atombios_crtc_program_ss(rdev, ATOM_ENABLE, ATOM_DCPLL, -1, &ss); + } + +} + int atombios_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode, @@ -1522,19 +1604,6 @@ int atombios_crtc_mode_set(struct drm_crtc *crtc, } } - /* always set DCPLL */ - if (ASIC_IS_DCE4(rdev)) { - struct radeon_atom_ss ss; - bool ss_enabled = radeon_atombios_get_asic_ss_info(rdev, &ss, - ASIC_INTERNAL_SS_ON_DCPLL, - rdev->clock.default_dispclk); - if (ss_enabled) - atombios_crtc_program_ss(crtc, ATOM_DISABLE, ATOM_DCPLL, &ss); - /* XXX: DCE5, make sure voltage, dispclk is high enough */ - atombios_crtc_set_dcpll(crtc, rdev->clock.default_dispclk); - if (ss_enabled) - atombios_crtc_program_ss(crtc, ATOM_ENABLE, ATOM_DCPLL, &ss); - } atombios_crtc_set_pll(crtc, adjusted_mode); if (ASIC_IS_DCE4(rdev)) @@ -1568,18 +1637,28 @@ static bool atombios_crtc_mode_fixup(struct drm_crtc *crtc, static void atombios_crtc_prepare(struct drm_crtc *crtc) { struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); + struct drm_device *dev = crtc->dev; + struct radeon_device *rdev = dev->dev_private; + radeon_crtc->in_mode_set = true; /* pick pll */ radeon_crtc->pll_id = radeon_atom_pick_pll(crtc); + /* disable crtc pair power gating before programming */ + if (ASIC_IS_DCE6(rdev)) + atombios_powergate_crtc(crtc, ATOM_DISABLE); + atombios_lock_crtc(crtc, ATOM_ENABLE); atombios_crtc_dpms(crtc, DRM_MODE_DPMS_OFF); } static void atombios_crtc_commit(struct drm_crtc *crtc) { + struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); + atombios_crtc_dpms(crtc, DRM_MODE_DPMS_ON); atombios_lock_crtc(crtc, ATOM_DISABLE); + radeon_crtc->in_mode_set = false; } static void atombios_crtc_disable(struct drm_crtc *crtc) @@ -1611,6 +1690,12 @@ static void atombios_crtc_disable(struct drm_crtc *crtc) atombios_crtc_program_pll(crtc, radeon_crtc->crtc_id, radeon_crtc->pll_id, 0, 0, ATOM_DISABLE, 0, 0, 0, 0, 0, false, &ss); break; + case ATOM_PPLL0: + /* disable the ppll */ + if (ASIC_IS_DCE61(rdev)) + atombios_crtc_program_pll(crtc, radeon_crtc->crtc_id, radeon_crtc->pll_id, + 0, 0, ATOM_DISABLE, 0, 0, 0, 0, 0, false, &ss); + break; default: break; } diff --git a/drivers/gpu/drm/radeon/atombios_dp.c b/drivers/gpu/drm/radeon/atombios_dp.c index 3254d51..886b41f 100644 --- a/drivers/gpu/drm/radeon/atombios_dp.c +++ b/drivers/gpu/drm/radeon/atombios_dp.c @@ -64,12 +64,12 @@ static int radeon_process_aux_ch(struct radeon_i2c_chan *chan, memset(&args, 0, sizeof(args)); - base = (unsigned char *)rdev->mode_info.atom_context->scratch; + base = (unsigned char *)(rdev->mode_info.atom_context->scratch + 1); memcpy(base, send, send_bytes); - args.v1.lpAuxRequest = 0; - args.v1.lpDataOut = 16; + args.v1.lpAuxRequest = 0 + 4; + args.v1.lpDataOut = 16 + 4; args.v1.ucDataOutLen = 0; args.v1.ucChannelID = chan->rec.i2c_id; args.v1.ucDelay = delay / 10; @@ -406,10 +406,13 @@ static void dp_get_adjust_train(u8 link_status[DP_LINK_STATUS_SIZE], /* get bpc from the EDID */ static int convert_bpc_to_bpp(int bpc) { +#if 0 if (bpc == 0) return 24; else return bpc * 3; +#endif + return 24; } /* get the max pix clock supported by the link rate and lane num */ @@ -746,7 +749,8 @@ static int radeon_dp_link_train_init(struct radeon_dp_link_train_info *dp_info) /* set the lane count on the sink */ tmp = dp_info->dp_lane_count; - if (dp_info->dpcd[0] >= 0x11) + if (dp_info->dpcd[DP_DPCD_REV] >= 0x11 && + dp_info->dpcd[DP_MAX_LANE_COUNT] & DP_ENHANCED_FRAME_CAP) tmp |= DP_LANE_COUNT_ENHANCED_FRAME_EN; radeon_write_dpcd_reg(dp_info->radeon_connector, DP_LANE_COUNT_SET, tmp); diff --git a/drivers/gpu/drm/radeon/atombios_encoders.c b/drivers/gpu/drm/radeon/atombios_encoders.c index f0dc04b..23e3ea6 100644 --- a/drivers/gpu/drm/radeon/atombios_encoders.c +++ b/drivers/gpu/drm/radeon/atombios_encoders.c @@ -57,22 +57,6 @@ static inline bool radeon_encoder_is_digital(struct drm_encoder *encoder) } } -static struct drm_connector * -radeon_get_connector_for_encoder_init(struct drm_encoder *encoder) -{ - struct drm_device *dev = encoder->dev; - struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); - struct drm_connector *connector; - struct radeon_connector *radeon_connector; - - list_for_each_entry(connector, &dev->mode_config.connector_list, head) { - radeon_connector = to_radeon_connector(connector); - if (radeon_encoder->devices & radeon_connector->devices) - return connector; - } - return NULL; -} - static bool radeon_atom_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) @@ -257,7 +241,7 @@ atombios_dvo_setup(struct drm_encoder *encoder, int action) /* R4xx, R5xx */ args.ext_tmds.sXTmdsEncoder.ucEnable = action; - if (radeon_encoder->pixel_clock > 165000) + if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.ext_tmds.sXTmdsEncoder.ucMisc |= PANEL_ENCODER_MISC_DUAL; args.ext_tmds.sXTmdsEncoder.ucMisc |= ATOM_PANEL_MISC_888RGB; @@ -269,7 +253,7 @@ atombios_dvo_setup(struct drm_encoder *encoder, int action) /* DFP1, CRT1, TV1 depending on the type of port */ args.dvo.sDVOEncoder.ucDeviceType = ATOM_DEVICE_DFP1_INDEX; - if (radeon_encoder->pixel_clock > 165000) + if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.dvo.sDVOEncoder.usDevAttr.sDigAttrib.ucAttribute |= PANEL_ENCODER_MISC_DUAL; break; case 3: @@ -353,7 +337,7 @@ atombios_digital_setup(struct drm_encoder *encoder, int action) } else { if (dig->linkb) args.v1.ucMisc |= PANEL_ENCODER_MISC_TMDS_LINKB; - if (radeon_encoder->pixel_clock > 165000) + if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.v1.ucMisc |= PANEL_ENCODER_MISC_DUAL; /*if (pScrn->rgbBits == 8) */ args.v1.ucMisc |= ATOM_PANEL_MISC_888RGB; @@ -392,7 +376,7 @@ atombios_digital_setup(struct drm_encoder *encoder, int action) } else { if (dig->linkb) args.v2.ucMisc |= PANEL_ENCODER_MISC_TMDS_LINKB; - if (radeon_encoder->pixel_clock > 165000) + if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.v2.ucMisc |= PANEL_ENCODER_MISC_DUAL; } break; @@ -413,8 +397,6 @@ int atombios_get_encoder_mode(struct drm_encoder *encoder) { struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); - struct drm_device *dev = encoder->dev; - struct radeon_device *rdev = dev->dev_private; struct drm_connector *connector; struct radeon_connector *radeon_connector; struct radeon_connector_atom_dig *dig_connector; @@ -438,13 +420,10 @@ atombios_get_encoder_mode(struct drm_encoder *encoder) switch (connector->connector_type) { case DRM_MODE_CONNECTOR_DVII: case DRM_MODE_CONNECTOR_HDMIB: /* HDMI-B is basically DL-DVI; analog works fine */ - if (drm_detect_monitor_audio(radeon_connector->edid) && radeon_audio) { - /* fix me */ - if (ASIC_IS_DCE4(rdev)) - return ATOM_ENCODER_MODE_DVI; - else - return ATOM_ENCODER_MODE_HDMI; - } else if (radeon_connector->use_digital) + if (drm_detect_hdmi_monitor(radeon_connector->edid) && + radeon_audio) + return ATOM_ENCODER_MODE_HDMI; + else if (radeon_connector->use_digital) return ATOM_ENCODER_MODE_DVI; else return ATOM_ENCODER_MODE_CRT; @@ -452,13 +431,10 @@ atombios_get_encoder_mode(struct drm_encoder *encoder) case DRM_MODE_CONNECTOR_DVID: case DRM_MODE_CONNECTOR_HDMIA: default: - if (drm_detect_monitor_audio(radeon_connector->edid) && radeon_audio) { - /* fix me */ - if (ASIC_IS_DCE4(rdev)) - return ATOM_ENCODER_MODE_DVI; - else - return ATOM_ENCODER_MODE_HDMI; - } else + if (drm_detect_hdmi_monitor(radeon_connector->edid) && + radeon_audio) + return ATOM_ENCODER_MODE_HDMI; + else return ATOM_ENCODER_MODE_DVI; break; case DRM_MODE_CONNECTOR_LVDS: @@ -469,13 +445,10 @@ atombios_get_encoder_mode(struct drm_encoder *encoder) if ((dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_DISPLAYPORT) || (dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_eDP)) return ATOM_ENCODER_MODE_DP; - else if (drm_detect_monitor_audio(radeon_connector->edid) && radeon_audio) { - /* fix me */ - if (ASIC_IS_DCE4(rdev)) - return ATOM_ENCODER_MODE_DVI; - else - return ATOM_ENCODER_MODE_HDMI; - } else + else if (drm_detect_hdmi_monitor(radeon_connector->edid) && + radeon_audio) + return ATOM_ENCODER_MODE_HDMI; + else return ATOM_ENCODER_MODE_DVI; break; case DRM_MODE_CONNECTOR_eDP: @@ -510,7 +483,7 @@ atombios_get_encoder_mode(struct drm_encoder *encoder) * - 2 DIG encoder blocks. * DIG1/2 can drive UNIPHY0/1/2 link A or link B * - * DCE 4.0/5.0 + * DCE 4.0/5.0/6.0 * - 3 DIG transmitter blocks UNIPHY0/1/2 (links A and B). * Supports up to 6 digital outputs * - 6 DIG encoder blocks. @@ -526,7 +499,11 @@ atombios_get_encoder_mode(struct drm_encoder *encoder) * - 3 DIG transmitter blocks UNIPHY0/1/2 (links A and B). * Supports up to 6 digital outputs * - 2 DIG encoder blocks. + * llano * DIG1/2 can drive UNIPHY0/1/2 link A or link B + * ontario + * DIG1 drives UNIPHY0/1/2 link A + * DIG2 drives UNIPHY0/1/2 link B * * Routing * crtc -> dig encoder -> UNIPHY/LVTMA (1 or 2 links) @@ -568,7 +545,7 @@ atombios_dig_encoder_setup(struct drm_encoder *encoder, int action, int panel_mo dp_clock = dig_connector->dp_clock; dp_lane_count = dig_connector->dp_lane_count; hpd_id = radeon_connector->hpd.hpd; - bpc = connector->display_info.bpc; + /* bpc = connector->display_info.bpc; */ } /* no dig encoder assigned */ @@ -602,7 +579,7 @@ atombios_dig_encoder_setup(struct drm_encoder *encoder, int action, int panel_mo if (ENCODER_MODE_IS_DP(args.v1.ucEncoderMode)) args.v1.ucLaneNum = dp_lane_count; - else if (radeon_encoder->pixel_clock > 165000) + else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.v1.ucLaneNum = 8; else args.v1.ucLaneNum = 4; @@ -637,7 +614,7 @@ atombios_dig_encoder_setup(struct drm_encoder *encoder, int action, int panel_mo if (ENCODER_MODE_IS_DP(args.v1.ucEncoderMode)) args.v3.ucLaneNum = dp_lane_count; - else if (radeon_encoder->pixel_clock > 165000) + else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.v3.ucLaneNum = 8; else args.v3.ucLaneNum = 4; @@ -677,7 +654,7 @@ atombios_dig_encoder_setup(struct drm_encoder *encoder, int action, int panel_mo if (ENCODER_MODE_IS_DP(args.v1.ucEncoderMode)) args.v4.ucLaneNum = dp_lane_count; - else if (radeon_encoder->pixel_clock > 165000) + else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.v4.ucLaneNum = 8; else args.v4.ucLaneNum = 4; @@ -734,6 +711,7 @@ union dig_transmitter_control { DIG_TRANSMITTER_CONTROL_PARAMETERS_V2 v2; DIG_TRANSMITTER_CONTROL_PARAMETERS_V3 v3; DIG_TRANSMITTER_CONTROL_PARAMETERS_V4 v4; + DIG_TRANSMITTER_CONTROL_PARAMETERS_V1_5 v5; }; void @@ -754,6 +732,7 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t int connector_object_id = 0; int igp_lane_info = 0; int dig_encoder = dig->dig_encoder; + int hpd_id = RADEON_HPD_NONE; if (action == ATOM_TRANSMITTER_ACTION_INIT) { connector = radeon_get_connector_for_encoder_init(encoder); @@ -769,6 +748,7 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t struct radeon_connector_atom_dig *dig_connector = radeon_connector->con_priv; + hpd_id = radeon_connector->hpd.hpd; dp_clock = dig_connector->dp_clock; dp_lane_count = dig_connector->dp_lane_count; connector_object_id = @@ -821,7 +801,7 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t if (is_dp) args.v1.usPixelClock = cpu_to_le16(dp_clock / 10); - else if (radeon_encoder->pixel_clock > 165000) + else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.v1.usPixelClock = cpu_to_le16((radeon_encoder->pixel_clock / 2) / 10); else args.v1.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10); @@ -836,7 +816,8 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t if ((rdev->flags & RADEON_IS_IGP) && (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_UNIPHY)) { - if (is_dp || (radeon_encoder->pixel_clock <= 165000)) { + if (is_dp || + !radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) { if (igp_lane_info & 0x1) args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LANE_0_3; else if (igp_lane_info & 0x2) @@ -863,7 +844,7 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t else if (radeon_encoder->devices & (ATOM_DEVICE_DFP_SUPPORT)) { if (dig->coherent_mode) args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_COHERENT; - if (radeon_encoder->pixel_clock > 165000) + if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_8LANE_LINK; } break; @@ -878,7 +859,7 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t if (is_dp) args.v2.usPixelClock = cpu_to_le16(dp_clock / 10); - else if (radeon_encoder->pixel_clock > 165000) + else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.v2.usPixelClock = cpu_to_le16((radeon_encoder->pixel_clock / 2) / 10); else args.v2.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10); @@ -906,7 +887,7 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t } else if (radeon_encoder->devices & (ATOM_DEVICE_DFP_SUPPORT)) { if (dig->coherent_mode) args.v2.acConfig.fCoherentMode = 1; - if (radeon_encoder->pixel_clock > 165000) + if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.v2.acConfig.fDualLinkConnector = 1; } break; @@ -921,7 +902,7 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t if (is_dp) args.v3.usPixelClock = cpu_to_le16(dp_clock / 10); - else if (radeon_encoder->pixel_clock > 165000) + else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.v3.usPixelClock = cpu_to_le16((radeon_encoder->pixel_clock / 2) / 10); else args.v3.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10); @@ -929,7 +910,7 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t if (is_dp) args.v3.ucLaneNum = dp_lane_count; - else if (radeon_encoder->pixel_clock > 165000) + else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.v3.ucLaneNum = 8; else args.v3.ucLaneNum = 4; @@ -966,7 +947,7 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t else if (radeon_encoder->devices & (ATOM_DEVICE_DFP_SUPPORT)) { if (dig->coherent_mode) args.v3.acConfig.fCoherentMode = 1; - if (radeon_encoder->pixel_clock > 165000) + if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.v3.acConfig.fDualLinkConnector = 1; } break; @@ -981,7 +962,7 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t if (is_dp) args.v4.usPixelClock = cpu_to_le16(dp_clock / 10); - else if (radeon_encoder->pixel_clock > 165000) + else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.v4.usPixelClock = cpu_to_le16((radeon_encoder->pixel_clock / 2) / 10); else args.v4.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10); @@ -989,7 +970,7 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t if (is_dp) args.v4.ucLaneNum = dp_lane_count; - else if (radeon_encoder->pixel_clock > 165000) + else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.v4.ucLaneNum = 8; else args.v4.ucLaneNum = 4; @@ -1029,10 +1010,64 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t else if (radeon_encoder->devices & (ATOM_DEVICE_DFP_SUPPORT)) { if (dig->coherent_mode) args.v4.acConfig.fCoherentMode = 1; - if (radeon_encoder->pixel_clock > 165000) + if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.v4.acConfig.fDualLinkConnector = 1; } break; + case 5: + args.v5.ucAction = action; + if (is_dp) + args.v5.usSymClock = cpu_to_le16(dp_clock / 10); + else + args.v5.usSymClock = cpu_to_le16(radeon_encoder->pixel_clock / 10); + + switch (radeon_encoder->encoder_id) { + case ENCODER_OBJECT_ID_INTERNAL_UNIPHY: + if (dig->linkb) + args.v5.ucPhyId = ATOM_PHY_ID_UNIPHYB; + else + args.v5.ucPhyId = ATOM_PHY_ID_UNIPHYA; + break; + case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1: + if (dig->linkb) + args.v5.ucPhyId = ATOM_PHY_ID_UNIPHYD; + else + args.v5.ucPhyId = ATOM_PHY_ID_UNIPHYC; + break; + case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2: + if (dig->linkb) + args.v5.ucPhyId = ATOM_PHY_ID_UNIPHYF; + else + args.v5.ucPhyId = ATOM_PHY_ID_UNIPHYE; + break; + } + if (is_dp) + args.v5.ucLaneNum = dp_lane_count; + else if (radeon_encoder->pixel_clock > 165000) + args.v5.ucLaneNum = 8; + else + args.v5.ucLaneNum = 4; + args.v5.ucConnObjId = connector_object_id; + args.v5.ucDigMode = atombios_get_encoder_mode(encoder); + + if (is_dp && rdev->clock.dp_extclk) + args.v5.asConfig.ucPhyClkSrcId = ENCODER_REFCLK_SRC_EXTCLK; + else + args.v5.asConfig.ucPhyClkSrcId = pll_id; + + if (is_dp) + args.v5.asConfig.ucCoherentMode = 1; /* DP requires coherent */ + else if (radeon_encoder->devices & (ATOM_DEVICE_DFP_SUPPORT)) { + if (dig->coherent_mode) + args.v5.asConfig.ucCoherentMode = 1; + } + if (hpd_id == RADEON_HPD_NONE) + args.v5.asConfig.ucHPDSel = 0; + else + args.v5.asConfig.ucHPDSel = hpd_id + 1; + args.v5.ucDigEncoderSel = 1 << dig_encoder; + args.v5.ucDPLaneSet = lane_set; + break; default: DRM_ERROR("Unknown table version %d, %d\n", frev, crev); break; @@ -1128,7 +1163,7 @@ atombios_external_encoder_setup(struct drm_encoder *encoder, dp_lane_count = dig_connector->dp_lane_count; connector_object_id = (radeon_connector->connector_object_id & OBJECT_ID_MASK) >> OBJECT_ID_SHIFT; - bpc = connector->display_info.bpc; + /* bpc = connector->display_info.bpc; */ } memset(&args, 0, sizeof(args)); @@ -1152,7 +1187,7 @@ atombios_external_encoder_setup(struct drm_encoder *encoder, if (dp_clock == 270000) args.v1.sDigEncoder.ucConfig |= ATOM_ENCODER_CONFIG_DPLINKRATE_2_70GHZ; args.v1.sDigEncoder.ucLaneNum = dp_lane_count; - } else if (radeon_encoder->pixel_clock > 165000) + } else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.v1.sDigEncoder.ucLaneNum = 8; else args.v1.sDigEncoder.ucLaneNum = 4; @@ -1171,7 +1206,7 @@ atombios_external_encoder_setup(struct drm_encoder *encoder, else if (dp_clock == 540000) args.v3.sExtEncoder.ucConfig |= EXTERNAL_ENCODER_CONFIG_V3_DPLINKRATE_5_40GHZ; args.v3.sExtEncoder.ucLaneNum = dp_lane_count; - } else if (radeon_encoder->pixel_clock > 165000) + } else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) args.v3.sExtEncoder.ucLaneNum = 8; else args.v3.sExtEncoder.ucLaneNum = 4; @@ -1369,7 +1404,7 @@ radeon_atom_encoder_dpms_dig(struct drm_encoder *encoder, int mode) ATOM_ENCODER_CMD_SETUP_PANEL_MODE, dig->panel_mode); if (ext_encoder) { - if (ASIC_IS_DCE41(rdev)) + if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE61(rdev)) atombios_external_encoder_setup(encoder, ext_encoder, EXTERNAL_ENCODER_ACTION_V3_ENCODER_SETUP); } @@ -1444,7 +1479,7 @@ radeon_atom_encoder_dpms_ext(struct drm_encoder *encoder, switch (mode) { case DRM_MODE_DPMS_ON: default: - if (ASIC_IS_DCE41(rdev)) { + if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE61(rdev)) { atombios_external_encoder_setup(encoder, ext_encoder, EXTERNAL_ENCODER_ACTION_V3_ENABLE_OUTPUT); atombios_external_encoder_setup(encoder, ext_encoder, @@ -1455,7 +1490,7 @@ radeon_atom_encoder_dpms_ext(struct drm_encoder *encoder, case DRM_MODE_DPMS_STANDBY: case DRM_MODE_DPMS_SUSPEND: case DRM_MODE_DPMS_OFF: - if (ASIC_IS_DCE41(rdev)) { + if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE61(rdev)) { atombios_external_encoder_setup(encoder, ext_encoder, EXTERNAL_ENCODER_ACTION_V3_ENCODER_BLANKING); atombios_external_encoder_setup(encoder, ext_encoder, @@ -1734,13 +1769,34 @@ static int radeon_atom_pick_dig_encoder(struct drm_encoder *encoder) struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc); struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); struct drm_encoder *test_encoder; - struct radeon_encoder_atom_dig *dig; + struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; uint32_t dig_enc_in_use = 0; - /* DCE4/5 */ - if (ASIC_IS_DCE4(rdev)) { - dig = radeon_encoder->enc_priv; - if (ASIC_IS_DCE41(rdev)) { + if (ASIC_IS_DCE6(rdev)) { + /* DCE6 */ + switch (radeon_encoder->encoder_id) { + case ENCODER_OBJECT_ID_INTERNAL_UNIPHY: + if (dig->linkb) + return 1; + else + return 0; + break; + case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1: + if (dig->linkb) + return 3; + else + return 2; + break; + case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2: + if (dig->linkb) + return 5; + else + return 4; + break; + } + } else if (ASIC_IS_DCE4(rdev)) { + /* DCE4/5 */ + if (ASIC_IS_DCE41(rdev) && !ASIC_IS_DCE61(rdev)) { /* ontario follows DCE4 */ if (rdev->family == CHIP_PALM) { if (dig->linkb) @@ -1828,7 +1884,7 @@ radeon_atom_encoder_init(struct radeon_device *rdev) break; } - if (ext_encoder && ASIC_IS_DCE41(rdev)) + if (ext_encoder && (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE61(rdev))) atombios_external_encoder_setup(encoder, ext_encoder, EXTERNAL_ENCODER_ACTION_V3_ENCODER_INIT); } diff --git a/drivers/gpu/drm/radeon/atombios_i2c.c b/drivers/gpu/drm/radeon/atombios_i2c.c new file mode 100644 index 0000000..44d87b6 --- /dev/null +++ b/drivers/gpu/drm/radeon/atombios_i2c.c @@ -0,0 +1,139 @@ +/* + * Copyright 2011 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Alex Deucher + * + */ +#include "drmP.h" +#include "radeon_drm.h" +#include "radeon.h" +#include "atom.h" + +#define TARGET_HW_I2C_CLOCK 50 + +/* these are a limitation of ProcessI2cChannelTransaction not the hw */ +#define ATOM_MAX_HW_I2C_WRITE 2 +#define ATOM_MAX_HW_I2C_READ 255 + +static int radeon_process_i2c_ch(struct radeon_i2c_chan *chan, + u8 slave_addr, u8 flags, + u8 *buf, u8 num) +{ + struct drm_device *dev = chan->dev; + struct radeon_device *rdev = dev->dev_private; + PROCESS_I2C_CHANNEL_TRANSACTION_PS_ALLOCATION args; + int index = GetIndexIntoMasterTable(COMMAND, ProcessI2cChannelTransaction); + unsigned char *base; + u16 out; + + memset(&args, 0, sizeof(args)); + + base = (unsigned char *)rdev->mode_info.atom_context->scratch; + + if (flags & HW_I2C_WRITE) { + if (num > ATOM_MAX_HW_I2C_WRITE) { + DRM_ERROR("hw i2c: tried to write too many bytes (%d vs 2)\n", num); + return -EINVAL; + } + memcpy(&out, buf, num); + args.lpI2CDataOut = cpu_to_le16(out); + } else { + if (num > ATOM_MAX_HW_I2C_READ) { + DRM_ERROR("hw i2c: tried to read too many bytes (%d vs 255)\n", num); + return -EINVAL; + } + } + + args.ucI2CSpeed = TARGET_HW_I2C_CLOCK; + args.ucRegIndex = 0; + args.ucTransBytes = num; + args.ucSlaveAddr = slave_addr << 1; + args.ucLineNumber = chan->rec.i2c_id; + + atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); + + /* error */ + if (args.ucStatus != HW_ASSISTED_I2C_STATUS_SUCCESS) { + DRM_DEBUG_KMS("hw_i2c error\n"); + return -EIO; + } + + if (!(flags & HW_I2C_WRITE)) + memcpy(buf, base, num); + + return 0; +} + +int radeon_atom_hw_i2c_xfer(struct i2c_adapter *i2c_adap, + struct i2c_msg *msgs, int num) +{ + struct radeon_i2c_chan *i2c = i2c_get_adapdata(i2c_adap); + struct i2c_msg *p; + int i, remaining, current_count, buffer_offset, max_bytes, ret; + u8 buf = 0, flags; + + /* check for bus probe */ + p = &msgs[0]; + if ((num == 1) && (p->len == 0)) { + ret = radeon_process_i2c_ch(i2c, + p->addr, HW_I2C_WRITE, + &buf, 1); + if (ret) + return ret; + else + return num; + } + + for (i = 0; i < num; i++) { + p = &msgs[i]; + remaining = p->len; + buffer_offset = 0; + /* max_bytes are a limitation of ProcessI2cChannelTransaction not the hw */ + if (p->flags & I2C_M_RD) { + max_bytes = ATOM_MAX_HW_I2C_READ; + flags = HW_I2C_READ; + } else { + max_bytes = ATOM_MAX_HW_I2C_WRITE; + flags = HW_I2C_WRITE; + } + while (remaining) { + if (remaining > max_bytes) + current_count = max_bytes; + else + current_count = remaining; + ret = radeon_process_i2c_ch(i2c, + p->addr, flags, + &p->buf[buffer_offset], current_count); + if (ret) + return ret; + remaining -= current_count; + buffer_offset += current_count; + } + } + + return num; +} + +u32 radeon_atom_hw_i2c_func(struct i2c_adapter *adap) +{ + return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL; +} + diff --git a/drivers/gpu/drm/radeon/cayman_blit_shaders.c b/drivers/gpu/drm/radeon/cayman_blit_shaders.c index 7b4eeb7..19a0114 100644 --- a/drivers/gpu/drm/radeon/cayman_blit_shaders.c +++ b/drivers/gpu/drm/radeon/cayman_blit_shaders.c @@ -24,6 +24,7 @@ * Alex Deucher */ +#include #include #include diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c index 0977849..c62132c 100644 --- a/drivers/gpu/drm/radeon/evergreen.c +++ b/drivers/gpu/drm/radeon/evergreen.c @@ -50,6 +50,39 @@ static const u32 crtc_offsets[6] = static void evergreen_gpu_init(struct radeon_device *rdev); void evergreen_fini(struct radeon_device *rdev); void evergreen_pcie_gen2_enable(struct radeon_device *rdev); +extern void cayman_cp_int_cntl_setup(struct radeon_device *rdev, + int ring, u32 cp_int_cntl); + +void evergreen_tiling_fields(unsigned tiling_flags, unsigned *bankw, + unsigned *bankh, unsigned *mtaspect, + unsigned *tile_split) +{ + *bankw = (tiling_flags >> RADEON_TILING_EG_BANKW_SHIFT) & RADEON_TILING_EG_BANKW_MASK; + *bankh = (tiling_flags >> RADEON_TILING_EG_BANKH_SHIFT) & RADEON_TILING_EG_BANKH_MASK; + *mtaspect = (tiling_flags >> RADEON_TILING_EG_MACRO_TILE_ASPECT_SHIFT) & RADEON_TILING_EG_MACRO_TILE_ASPECT_MASK; + *tile_split = (tiling_flags >> RADEON_TILING_EG_TILE_SPLIT_SHIFT) & RADEON_TILING_EG_TILE_SPLIT_MASK; + switch (*bankw) { + default: + case 1: *bankw = EVERGREEN_ADDR_SURF_BANK_WIDTH_1; break; + case 2: *bankw = EVERGREEN_ADDR_SURF_BANK_WIDTH_2; break; + case 4: *bankw = EVERGREEN_ADDR_SURF_BANK_WIDTH_4; break; + case 8: *bankw = EVERGREEN_ADDR_SURF_BANK_WIDTH_8; break; + } + switch (*bankh) { + default: + case 1: *bankh = EVERGREEN_ADDR_SURF_BANK_HEIGHT_1; break; + case 2: *bankh = EVERGREEN_ADDR_SURF_BANK_HEIGHT_2; break; + case 4: *bankh = EVERGREEN_ADDR_SURF_BANK_HEIGHT_4; break; + case 8: *bankh = EVERGREEN_ADDR_SURF_BANK_HEIGHT_8; break; + } + switch (*mtaspect) { + default: + case 1: *mtaspect = EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_1; break; + case 2: *mtaspect = EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_2; break; + case 4: *mtaspect = EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_4; break; + case 8: *mtaspect = EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_8; break; + } +} void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev) { @@ -560,7 +593,7 @@ static u32 evergreen_line_buffer_adjust(struct radeon_device *rdev, return 0; } -static u32 evergreen_get_number_of_dram_channels(struct radeon_device *rdev) +u32 evergreen_get_number_of_dram_channels(struct radeon_device *rdev) { u32 tmp = RREG32(MC_SHARED_CHMAP); @@ -1109,11 +1142,24 @@ void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *sav crtc_enabled = RREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[i]) & EVERGREEN_CRTC_MASTER_EN; if (crtc_enabled) { save->crtc_enabled[i] = true; - tmp = RREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[i]); - if (!(tmp & EVERGREEN_CRTC_DISP_READ_REQUEST_DISABLE)) { - dce4_wait_for_vblank(rdev, i); - tmp |= EVERGREEN_CRTC_DISP_READ_REQUEST_DISABLE; - WREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[i], tmp); + if (ASIC_IS_DCE6(rdev)) { + tmp = RREG32(EVERGREEN_CRTC_BLANK_CONTROL + crtc_offsets[i]); + if (!(tmp & EVERGREEN_CRTC_BLANK_DATA_EN)) { + radeon_wait_for_vblank(rdev, i); + tmp |= EVERGREEN_CRTC_BLANK_DATA_EN; + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 1); + WREG32(EVERGREEN_CRTC_BLANK_CONTROL + crtc_offsets[i], tmp); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 0); + } + } else { + tmp = RREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[i]); + if (!(tmp & EVERGREEN_CRTC_DISP_READ_REQUEST_DISABLE)) { + radeon_wait_for_vblank(rdev, i); + tmp |= EVERGREEN_CRTC_DISP_READ_REQUEST_DISABLE; + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 1); + WREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[i], tmp); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 0); + } } /* wait for the next frame */ frame_count = radeon_get_vblank_counter(rdev, i); @@ -1127,7 +1173,7 @@ void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *sav } } - evergreen_mc_wait_for_idle(rdev); + radeon_mc_wait_for_idle(rdev); blackout = RREG32(MC_SHARED_BLACKOUT_CNTL); if ((blackout & BLACKOUT_MODE_MASK) != 1) { @@ -1137,6 +1183,8 @@ void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *sav blackout &= ~BLACKOUT_MODE_MASK; WREG32(MC_SHARED_BLACKOUT_CNTL, blackout | 1); } + /* wait for the MC to settle */ + udelay(100); } void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *save) @@ -1166,10 +1214,20 @@ void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *s WREG32(BIF_FB_EN, FB_READ_EN | FB_WRITE_EN); for (i = 0; i < rdev->num_crtc; i++) { - if (save->crtc_enabled[i]) { - tmp = RREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[i]); - tmp &= ~EVERGREEN_CRTC_DISP_READ_REQUEST_DISABLE; - WREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[i], tmp); + if (save->crtc_enabled) { + if (ASIC_IS_DCE6(rdev)) { + tmp = RREG32(EVERGREEN_CRTC_BLANK_CONTROL + crtc_offsets[i]); + tmp |= EVERGREEN_CRTC_BLANK_DATA_EN; + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 1); + WREG32(EVERGREEN_CRTC_BLANK_CONTROL + crtc_offsets[i], tmp); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 0); + } else { + tmp = RREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[i]); + tmp &= ~EVERGREEN_CRTC_DISP_READ_REQUEST_DISABLE; + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 1); + WREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[i], tmp); + WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 0); + } /* wait for the next frame */ frame_count = radeon_get_vblank_counter(rdev, i); for (j = 0; j < rdev->usec_timeout; j++) { @@ -1229,7 +1287,10 @@ void evergreen_mc_program(struct radeon_device *rdev) rdev->mc.vram_end >> 12); } WREG32(MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR, rdev->vram_scratch.gpu_addr >> 12); - if (rdev->flags & RADEON_IS_IGP) { + /* llano/ontario only */ + if ((rdev->family == CHIP_PALM) || + (rdev->family == CHIP_SUMO) || + (rdev->family == CHIP_SUMO2)) { tmp = RREG32(MC_FUS_VM_FB_OFFSET) & 0x000FFFFF; tmp |= ((rdev->mc.vram_end >> 20) & 0xF) << 24; tmp |= ((rdev->mc.vram_start >> 20) & 0xF) << 20; @@ -1264,18 +1325,20 @@ void evergreen_mc_program(struct radeon_device *rdev) */ void evergreen_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) { + struct radeon_ring *ring = &rdev->ring[ib->fence->ring]; + /* set to DX10/11 mode */ - radeon_ring_write(rdev, PACKET3(PACKET3_MODE_CONTROL, 0)); - radeon_ring_write(rdev, 1); + radeon_ring_write(ring, PACKET3(PACKET3_MODE_CONTROL, 0)); + radeon_ring_write(ring, 1); /* FIXME: implement */ - radeon_ring_write(rdev, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); - radeon_ring_write(rdev, + radeon_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); + radeon_ring_write(ring, #ifdef __BIG_ENDIAN (2 << 0) | #endif (ib->gpu_addr & 0xFFFFFFFC)); - radeon_ring_write(rdev, upper_32_bits(ib->gpu_addr) & 0xFF); - radeon_ring_write(rdev, ib->length_dw); + radeon_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xFF); + radeon_ring_write(ring, ib->length_dw); } @@ -1313,71 +1376,73 @@ static int evergreen_cp_load_microcode(struct radeon_device *rdev) static int evergreen_cp_start(struct radeon_device *rdev) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; int r, i; uint32_t cp_me; - r = radeon_ring_lock(rdev, 7); + r = radeon_ring_lock(rdev, ring, 7); if (r) { DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r); return r; } - radeon_ring_write(rdev, PACKET3(PACKET3_ME_INITIALIZE, 5)); - radeon_ring_write(rdev, 0x1); - radeon_ring_write(rdev, 0x0); - radeon_ring_write(rdev, rdev->config.evergreen.max_hw_contexts - 1); - radeon_ring_write(rdev, PACKET3_ME_INITIALIZE_DEVICE_ID(1)); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, 0); - radeon_ring_unlock_commit(rdev); + radeon_ring_write(ring, PACKET3(PACKET3_ME_INITIALIZE, 5)); + radeon_ring_write(ring, 0x1); + radeon_ring_write(ring, 0x0); + radeon_ring_write(ring, rdev->config.evergreen.max_hw_contexts - 1); + radeon_ring_write(ring, PACKET3_ME_INITIALIZE_DEVICE_ID(1)); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 0); + radeon_ring_unlock_commit(rdev, ring); cp_me = 0xff; WREG32(CP_ME_CNTL, cp_me); - r = radeon_ring_lock(rdev, evergreen_default_size + 19); + r = radeon_ring_lock(rdev, ring, evergreen_default_size + 19); if (r) { DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r); return r; } /* setup clear context state */ - radeon_ring_write(rdev, PACKET3(PACKET3_PREAMBLE_CNTL, 0)); - radeon_ring_write(rdev, PACKET3_PREAMBLE_BEGIN_CLEAR_STATE); + radeon_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0)); + radeon_ring_write(ring, PACKET3_PREAMBLE_BEGIN_CLEAR_STATE); for (i = 0; i < evergreen_default_size; i++) - radeon_ring_write(rdev, evergreen_default_state[i]); + radeon_ring_write(ring, evergreen_default_state[i]); - radeon_ring_write(rdev, PACKET3(PACKET3_PREAMBLE_CNTL, 0)); - radeon_ring_write(rdev, PACKET3_PREAMBLE_END_CLEAR_STATE); + radeon_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0)); + radeon_ring_write(ring, PACKET3_PREAMBLE_END_CLEAR_STATE); /* set clear context state */ - radeon_ring_write(rdev, PACKET3(PACKET3_CLEAR_STATE, 0)); - radeon_ring_write(rdev, 0); + radeon_ring_write(ring, PACKET3(PACKET3_CLEAR_STATE, 0)); + radeon_ring_write(ring, 0); /* SQ_VTX_BASE_VTX_LOC */ - radeon_ring_write(rdev, 0xc0026f00); - radeon_ring_write(rdev, 0x00000000); - radeon_ring_write(rdev, 0x00000000); - radeon_ring_write(rdev, 0x00000000); + radeon_ring_write(ring, 0xc0026f00); + radeon_ring_write(ring, 0x00000000); + radeon_ring_write(ring, 0x00000000); + radeon_ring_write(ring, 0x00000000); /* Clear consts */ - radeon_ring_write(rdev, 0xc0036f00); - radeon_ring_write(rdev, 0x00000bc4); - radeon_ring_write(rdev, 0xffffffff); - radeon_ring_write(rdev, 0xffffffff); - radeon_ring_write(rdev, 0xffffffff); + radeon_ring_write(ring, 0xc0036f00); + radeon_ring_write(ring, 0x00000bc4); + radeon_ring_write(ring, 0xffffffff); + radeon_ring_write(ring, 0xffffffff); + radeon_ring_write(ring, 0xffffffff); - radeon_ring_write(rdev, 0xc0026900); - radeon_ring_write(rdev, 0x00000316); - radeon_ring_write(rdev, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */ - radeon_ring_write(rdev, 0x00000010); /* */ + radeon_ring_write(ring, 0xc0026900); + radeon_ring_write(ring, 0x00000316); + radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */ + radeon_ring_write(ring, 0x00000010); /* */ - radeon_ring_unlock_commit(rdev); + radeon_ring_unlock_commit(rdev, ring); return 0; } int evergreen_cp_resume(struct radeon_device *rdev) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; u32 tmp; u32 rb_bufsz; int r; @@ -1395,13 +1460,14 @@ int evergreen_cp_resume(struct radeon_device *rdev) RREG32(GRBM_SOFT_RESET); /* Set ring buffer size */ - rb_bufsz = drm_order(rdev->cp.ring_size / 8); + rb_bufsz = drm_order(ring->ring_size / 8); tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; #ifdef __BIG_ENDIAN tmp |= BUF_SWAP_32BIT; #endif WREG32(CP_RB_CNTL, tmp); - WREG32(CP_SEM_WAIT_TIMER, 0x4); + WREG32(CP_SEM_WAIT_TIMER, 0x0); + WREG32(CP_SEM_INCOMPLETE_TIMER_CNTL, 0x0); /* Set the write pointer delay */ WREG32(CP_RB_WPTR_DELAY, 0); @@ -1409,8 +1475,8 @@ int evergreen_cp_resume(struct radeon_device *rdev) /* Initialize the ring buffer's read and write pointers */ WREG32(CP_RB_CNTL, tmp | RB_RPTR_WR_ENA); WREG32(CP_RB_RPTR_WR, 0); - rdev->cp.wptr = 0; - WREG32(CP_RB_WPTR, rdev->cp.wptr); + ring->wptr = 0; + WREG32(CP_RB_WPTR, ring->wptr); /* set the wb address wether it's enabled or not */ WREG32(CP_RB_RPTR_ADDR, @@ -1428,16 +1494,16 @@ int evergreen_cp_resume(struct radeon_device *rdev) mdelay(1); WREG32(CP_RB_CNTL, tmp); - WREG32(CP_RB_BASE, rdev->cp.gpu_addr >> 8); + WREG32(CP_RB_BASE, ring->gpu_addr >> 8); WREG32(CP_DEBUG, (1 << 27) | (1 << 28)); - rdev->cp.rptr = RREG32(CP_RB_RPTR); + ring->rptr = RREG32(CP_RB_RPTR); evergreen_cp_start(rdev); - rdev->cp.ready = true; - r = radeon_ring_test(rdev); + ring->ready = true; + r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, ring); if (r) { - rdev->cp.ready = false; + ring->ready = false; return r; } return 0; @@ -1730,7 +1796,7 @@ static void evergreen_gpu_init(struct radeon_device *rdev) case CHIP_SUMO: rdev->config.evergreen.num_ses = 1; rdev->config.evergreen.max_pipes = 4; - rdev->config.evergreen.max_tile_pipes = 4; + rdev->config.evergreen.max_tile_pipes = 2; if (rdev->pdev->device == 0x9648) rdev->config.evergreen.max_simds = 3; else if ((rdev->pdev->device == 0x9647) || @@ -1819,7 +1885,7 @@ static void evergreen_gpu_init(struct radeon_device *rdev) break; case CHIP_CAICOS: rdev->config.evergreen.num_ses = 1; - rdev->config.evergreen.max_pipes = 2; + rdev->config.evergreen.max_pipes = 4; rdev->config.evergreen.max_tile_pipes = 2; rdev->config.evergreen.max_simds = 2; rdev->config.evergreen.max_backends = 1 * rdev->config.evergreen.num_ses; @@ -1868,7 +1934,9 @@ static void evergreen_gpu_init(struct radeon_device *rdev) mc_shared_chmap = RREG32(MC_SHARED_CHMAP); - if (rdev->flags & RADEON_IS_IGP) + if ((rdev->family == CHIP_PALM) || + (rdev->family == CHIP_SUMO) || + (rdev->family == CHIP_SUMO2)) mc_arb_ramcfg = RREG32(FUS_MC_ARB_RAMCFG); else mc_arb_ramcfg = RREG32(MC_ARB_RAMCFG); @@ -2272,7 +2340,9 @@ int evergreen_mc_init(struct radeon_device *rdev) /* Get VRAM informations */ rdev->mc.vram_is_ddr = true; - if (rdev->flags & RADEON_IS_IGP) + if ((rdev->family == CHIP_PALM) || + (rdev->family == CHIP_SUMO) || + (rdev->family == CHIP_SUMO2)) tmp = RREG32(FUS_MC_ARB_RAMCFG); else tmp = RREG32(MC_ARB_RAMCFG); @@ -2304,12 +2374,14 @@ int evergreen_mc_init(struct radeon_device *rdev) rdev->mc.aper_base = pci_resource_start(rdev->pdev, 0); rdev->mc.aper_size = pci_resource_len(rdev->pdev, 0); /* Setup GPU memory space */ - if (rdev->flags & RADEON_IS_IGP) { + if ((rdev->family == CHIP_PALM) || + (rdev->family == CHIP_SUMO) || + (rdev->family == CHIP_SUMO2)) { /* size in bytes on fusion */ rdev->mc.mc_vram_size = RREG32(CONFIG_MEMSIZE); rdev->mc.real_vram_size = RREG32(CONFIG_MEMSIZE); } else { - /* size in MB on evergreen */ + /* size in MB on evergreen/cayman/tn */ rdev->mc.mc_vram_size = RREG32(CONFIG_MEMSIZE) * 1024 * 1024; rdev->mc.real_vram_size = RREG32(CONFIG_MEMSIZE) * 1024 * 1024; } @@ -2320,7 +2392,7 @@ int evergreen_mc_init(struct radeon_device *rdev) return 0; } -bool evergreen_gpu_is_lockup(struct radeon_device *rdev) +bool evergreen_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) { u32 srbm_status; u32 grbm_status; @@ -2333,19 +2405,19 @@ bool evergreen_gpu_is_lockup(struct radeon_device *rdev) grbm_status_se0 = RREG32(GRBM_STATUS_SE0); grbm_status_se1 = RREG32(GRBM_STATUS_SE1); if (!(grbm_status & GUI_ACTIVE)) { - r100_gpu_lockup_update(lockup, &rdev->cp); + r100_gpu_lockup_update(lockup, ring); return false; } /* force CP activities */ - r = radeon_ring_lock(rdev, 2); + r = radeon_ring_lock(rdev, ring, 2); if (!r) { /* PACKET2 NOP */ - radeon_ring_write(rdev, 0x80000000); - radeon_ring_write(rdev, 0x80000000); - radeon_ring_unlock_commit(rdev); + radeon_ring_write(ring, 0x80000000); + radeon_ring_write(ring, 0x80000000); + radeon_ring_unlock_commit(rdev, ring); } - rdev->cp.rptr = RREG32(CP_RB_RPTR); - return r100_gpu_cp_is_lockup(rdev, lockup, &rdev->cp); + ring->rptr = RREG32(CP_RB_RPTR); + return r100_gpu_cp_is_lockup(rdev, lockup, ring); } static int evergreen_gpu_soft_reset(struct radeon_device *rdev) @@ -2437,7 +2509,13 @@ void evergreen_disable_interrupt_state(struct radeon_device *rdev) { u32 tmp; - WREG32(CP_INT_CNTL, CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE); + if (rdev->family >= CHIP_CAYMAN) { + cayman_cp_int_cntl_setup(rdev, 0, + CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE); + cayman_cp_int_cntl_setup(rdev, 1, 0); + cayman_cp_int_cntl_setup(rdev, 2, 0); + } else + WREG32(CP_INT_CNTL, CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE); WREG32(GRBM_INT_CNTL, 0); WREG32(INT_MASK + EVERGREEN_CRTC0_REGISTER_OFFSET, 0); WREG32(INT_MASK + EVERGREEN_CRTC1_REGISTER_OFFSET, 0); @@ -2461,7 +2539,9 @@ void evergreen_disable_interrupt_state(struct radeon_device *rdev) WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, 0); } - WREG32(DACA_AUTODETECT_INT_CONTROL, 0); + /* only one DAC on DCE6 */ + if (!ASIC_IS_DCE6(rdev)) + WREG32(DACA_AUTODETECT_INT_CONTROL, 0); WREG32(DACB_AUTODETECT_INT_CONTROL, 0); tmp = RREG32(DC_HPD1_INT_CONTROL) & DC_HPDx_INT_POLARITY; @@ -2482,6 +2562,7 @@ void evergreen_disable_interrupt_state(struct radeon_device *rdev) int evergreen_irq_set(struct radeon_device *rdev) { u32 cp_int_cntl = CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE; + u32 cp_int_cntl1 = 0, cp_int_cntl2 = 0; u32 crtc1 = 0, crtc2 = 0, crtc3 = 0, crtc4 = 0, crtc5 = 0, crtc6 = 0; u32 hpd1, hpd2, hpd3, hpd4, hpd5, hpd6; u32 grbm_int_cntl = 0; @@ -2506,11 +2587,28 @@ int evergreen_irq_set(struct radeon_device *rdev) hpd5 = RREG32(DC_HPD5_INT_CONTROL) & ~DC_HPDx_INT_EN; hpd6 = RREG32(DC_HPD6_INT_CONTROL) & ~DC_HPDx_INT_EN; - if (rdev->irq.sw_int) { - DRM_DEBUG("evergreen_irq_set: sw int\n"); - cp_int_cntl |= RB_INT_ENABLE; - cp_int_cntl |= TIME_STAMP_INT_ENABLE; + if (rdev->family >= CHIP_CAYMAN) { + /* enable CP interrupts on all rings */ + if (rdev->irq.sw_int[RADEON_RING_TYPE_GFX_INDEX]) { + DRM_DEBUG("evergreen_irq_set: sw int gfx\n"); + cp_int_cntl |= TIME_STAMP_INT_ENABLE; + } + if (rdev->irq.sw_int[CAYMAN_RING_TYPE_CP1_INDEX]) { + DRM_DEBUG("evergreen_irq_set: sw int cp1\n"); + cp_int_cntl1 |= TIME_STAMP_INT_ENABLE; + } + if (rdev->irq.sw_int[CAYMAN_RING_TYPE_CP2_INDEX]) { + DRM_DEBUG("evergreen_irq_set: sw int cp2\n"); + cp_int_cntl2 |= TIME_STAMP_INT_ENABLE; + } + } else { + if (rdev->irq.sw_int[RADEON_RING_TYPE_GFX_INDEX]) { + DRM_DEBUG("evergreen_irq_set: sw int gfx\n"); + cp_int_cntl |= RB_INT_ENABLE; + cp_int_cntl |= TIME_STAMP_INT_ENABLE; + } } + if (rdev->irq.crtc_vblank_int[0] || rdev->irq.pflip[0]) { DRM_DEBUG("evergreen_irq_set: vblank 0\n"); @@ -2570,7 +2668,12 @@ int evergreen_irq_set(struct radeon_device *rdev) grbm_int_cntl |= GUI_IDLE_INT_ENABLE; } - WREG32(CP_INT_CNTL, cp_int_cntl); + if (rdev->family >= CHIP_CAYMAN) { + cayman_cp_int_cntl_setup(rdev, 0, cp_int_cntl); + cayman_cp_int_cntl_setup(rdev, 1, cp_int_cntl1); + cayman_cp_int_cntl_setup(rdev, 2, cp_int_cntl2); + } else + WREG32(CP_INT_CNTL, cp_int_cntl); WREG32(GRBM_INT_CNTL, grbm_int_cntl); WREG32(INT_MASK + EVERGREEN_CRTC0_REGISTER_OFFSET, crtc1); @@ -2985,11 +3088,24 @@ restart_ih: case 177: /* CP_INT in IB1 */ case 178: /* CP_INT in IB2 */ DRM_DEBUG("IH: CP int: 0x%08x\n", src_data); - radeon_fence_process(rdev); + radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX); break; case 181: /* CP EOP event */ DRM_DEBUG("IH: CP EOP\n"); - radeon_fence_process(rdev); + if (rdev->family >= CHIP_CAYMAN) { + switch (src_data) { + case 0: + radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX); + break; + case 1: + radeon_fence_process(rdev, CAYMAN_RING_TYPE_CP1_INDEX); + break; + case 2: + radeon_fence_process(rdev, CAYMAN_RING_TYPE_CP2_INDEX); + break; + } + } else + radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX); break; case 233: /* GUI IDLE */ DRM_DEBUG("IH: GUI idle\n"); @@ -3019,6 +3135,7 @@ restart_ih: static int evergreen_startup(struct radeon_device *rdev) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; int r; /* enable pcie gen2 link */ @@ -3064,7 +3181,7 @@ static int evergreen_startup(struct radeon_device *rdev) r = evergreen_blit_init(rdev); if (r) { r600_blit_fini(rdev); - rdev->asic->copy = NULL; + rdev->asic->copy.copy = NULL; dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r); } @@ -3073,6 +3190,12 @@ static int evergreen_startup(struct radeon_device *rdev) if (r) return r; + r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r); + return r; + } + /* Enable IRQ */ r = r600_irq_init(rdev); if (r) { @@ -3082,7 +3205,9 @@ static int evergreen_startup(struct radeon_device *rdev) } evergreen_irq_set(rdev); - r = radeon_ring_init(rdev, rdev->cp.ring_size); + r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET, + R600_CP_RB_RPTR, R600_CP_RB_WPTR, + 0, 0xfffff, RADEON_CP_PACKET2); if (r) return r; r = evergreen_cp_load_microcode(rdev); @@ -3092,6 +3217,23 @@ static int evergreen_startup(struct radeon_device *rdev) if (r) return r; + r = radeon_ib_pool_start(rdev); + if (r) + return r; + + r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); + if (r) { + DRM_ERROR("radeon: failed testing IB (%d).\n", r); + rdev->accel_working = false; + return r; + } + + r = r600_audio_init(rdev); + if (r) { + DRM_ERROR("radeon: audio init failed\n"); + return r; + } + return 0; } @@ -3111,15 +3253,11 @@ int evergreen_resume(struct radeon_device *rdev) /* post card */ atom_asic_init(rdev->mode_info.atom_context); + rdev->accel_working = true; r = evergreen_startup(rdev); if (r) { DRM_ERROR("evergreen startup failed on resume\n"); - return r; - } - - r = r600_ib_test(rdev); - if (r) { - DRM_ERROR("radeon: failed testing IB (%d).\n", r); + rdev->accel_working = false; return r; } @@ -3129,13 +3267,17 @@ int evergreen_resume(struct radeon_device *rdev) int evergreen_suspend(struct radeon_device *rdev) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; + + r600_audio_fini(rdev); /* FIXME: we should wait for ring to be empty */ + radeon_ib_pool_suspend(rdev); + r600_blit_suspend(rdev); r700_cp_stop(rdev); - rdev->cp.ready = false; + ring->ready = false; evergreen_irq_suspend(rdev); radeon_wb_disable(rdev); evergreen_pcie_gart_disable(rdev); - r600_blit_suspend(rdev); return 0; } @@ -3210,8 +3352,8 @@ int evergreen_init(struct radeon_device *rdev) if (r) return r; - rdev->cp.ring_obj = NULL; - r600_ring_init(rdev, 1024 * 1024); + rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ring_obj = NULL; + r600_ring_init(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX], 1024 * 1024); rdev->ih.ring_obj = NULL; r600_ih_ring_init(rdev, 64 * 1024); @@ -3220,29 +3362,24 @@ int evergreen_init(struct radeon_device *rdev) if (r) return r; + r = radeon_ib_pool_init(rdev); rdev->accel_working = true; + if (r) { + dev_err(rdev->dev, "IB initialization failed (%d).\n", r); + rdev->accel_working = false; + } + r = evergreen_startup(rdev); if (r) { dev_err(rdev->dev, "disabling GPU acceleration\n"); r700_cp_fini(rdev); r600_irq_fini(rdev); radeon_wb_fini(rdev); + r100_ib_fini(rdev); radeon_irq_kms_fini(rdev); evergreen_pcie_gart_fini(rdev); rdev->accel_working = false; } - if (rdev->accel_working) { - r = radeon_ib_pool_init(rdev); - if (r) { - DRM_ERROR("radeon: failed initializing IB pool (%d).\n", r); - rdev->accel_working = false; - } - r = r600_ib_test(rdev); - if (r) { - DRM_ERROR("radeon: failed testing IB (%d).\n", r); - rdev->accel_working = false; - } - } /* Don't start up if the MC ucode is missing on BTC parts. * The default clocks and voltages before the MC ucode @@ -3260,15 +3397,17 @@ int evergreen_init(struct radeon_device *rdev) void evergreen_fini(struct radeon_device *rdev) { + r600_audio_fini(rdev); r600_blit_fini(rdev); r700_cp_fini(rdev); r600_irq_fini(rdev); radeon_wb_fini(rdev); - radeon_ib_pool_fini(rdev); + r100_ib_fini(rdev); radeon_irq_kms_fini(rdev); evergreen_pcie_gart_fini(rdev); r600_vram_scratch_fini(rdev); radeon_gem_fini(rdev); + radeon_semaphore_driver_fini(rdev); radeon_fence_driver_fini(rdev); radeon_agp_fini(rdev); radeon_bo_fini(rdev); diff --git a/drivers/gpu/drm/radeon/evergreen_blit_kms.c b/drivers/gpu/drm/radeon/evergreen_blit_kms.c index 914e5af..222acd2 100644 --- a/drivers/gpu/drm/radeon/evergreen_blit_kms.c +++ b/drivers/gpu/drm/radeon/evergreen_blit_kms.c @@ -32,23 +32,14 @@ #include "evergreend.h" #include "evergreen_blit_shaders.h" #include "cayman_blit_shaders.h" - -#define DI_PT_RECTLIST 0x11 -#define DI_INDEX_SIZE_16_BIT 0x0 -#define DI_SRC_SEL_AUTO_INDEX 0x2 - -#define FMT_8 0x1 -#define FMT_5_6_5 0x8 -#define FMT_8_8_8_8 0x1a -#define COLOR_8 0x1 -#define COLOR_5_6_5 0x8 -#define COLOR_8_8_8_8 0x1a +#include "radeon_blit_common.h" /* emits 17 */ static void set_render_target(struct radeon_device *rdev, int format, int w, int h, u64 gpu_addr) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; u32 cb_color_info; int pitch, slice; @@ -62,23 +53,23 @@ set_render_target(struct radeon_device *rdev, int format, pitch = (w / 8) - 1; slice = ((w * h) / 64) - 1; - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 15)); - radeon_ring_write(rdev, (CB_COLOR0_BASE - PACKET3_SET_CONTEXT_REG_START) >> 2); - radeon_ring_write(rdev, gpu_addr >> 8); - radeon_ring_write(rdev, pitch); - radeon_ring_write(rdev, slice); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, cb_color_info); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, (w - 1) | ((h - 1) << 16)); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, 0); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 15)); + radeon_ring_write(ring, (CB_COLOR0_BASE - PACKET3_SET_CONTEXT_REG_START) >> 2); + radeon_ring_write(ring, gpu_addr >> 8); + radeon_ring_write(ring, pitch); + radeon_ring_write(ring, slice); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, cb_color_info); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, (w - 1) | ((h - 1) << 16)); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 0); } /* emits 5dw */ @@ -87,6 +78,7 @@ cp_set_surface_sync(struct radeon_device *rdev, u32 sync_type, u32 size, u64 mc_addr) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; u32 cp_coher_size; if (size == 0xffffffff) @@ -99,39 +91,40 @@ cp_set_surface_sync(struct radeon_device *rdev, * to the RB directly. For IBs, the CP programs this as part of the * surface_sync packet. */ - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1)); - radeon_ring_write(rdev, (0x85e8 - PACKET3_SET_CONFIG_REG_START) >> 2); - radeon_ring_write(rdev, 0); /* CP_COHER_CNTL2 */ + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + radeon_ring_write(ring, (0x85e8 - PACKET3_SET_CONFIG_REG_START) >> 2); + radeon_ring_write(ring, 0); /* CP_COHER_CNTL2 */ } - radeon_ring_write(rdev, PACKET3(PACKET3_SURFACE_SYNC, 3)); - radeon_ring_write(rdev, sync_type); - radeon_ring_write(rdev, cp_coher_size); - radeon_ring_write(rdev, mc_addr >> 8); - radeon_ring_write(rdev, 10); /* poll interval */ + radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); + radeon_ring_write(ring, sync_type); + radeon_ring_write(ring, cp_coher_size); + radeon_ring_write(ring, mc_addr >> 8); + radeon_ring_write(ring, 10); /* poll interval */ } /* emits 11dw + 1 surface sync = 16dw */ static void set_shaders(struct radeon_device *rdev) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; u64 gpu_addr; /* VS */ gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.vs_offset; - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 3)); - radeon_ring_write(rdev, (SQ_PGM_START_VS - PACKET3_SET_CONTEXT_REG_START) >> 2); - radeon_ring_write(rdev, gpu_addr >> 8); - radeon_ring_write(rdev, 2); - radeon_ring_write(rdev, 0); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 3)); + radeon_ring_write(ring, (SQ_PGM_START_VS - PACKET3_SET_CONTEXT_REG_START) >> 2); + radeon_ring_write(ring, gpu_addr >> 8); + radeon_ring_write(ring, 2); + radeon_ring_write(ring, 0); /* PS */ gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.ps_offset; - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 4)); - radeon_ring_write(rdev, (SQ_PGM_START_PS - PACKET3_SET_CONTEXT_REG_START) >> 2); - radeon_ring_write(rdev, gpu_addr >> 8); - radeon_ring_write(rdev, 1); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, 2); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 4)); + radeon_ring_write(ring, (SQ_PGM_START_PS - PACKET3_SET_CONTEXT_REG_START) >> 2); + radeon_ring_write(ring, gpu_addr >> 8); + radeon_ring_write(ring, 1); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 2); gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.vs_offset; cp_set_surface_sync(rdev, PACKET3_SH_ACTION_ENA, 512, gpu_addr); @@ -141,6 +134,7 @@ set_shaders(struct radeon_device *rdev) static void set_vtx_resource(struct radeon_device *rdev, u64 gpu_addr) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; u32 sq_vtx_constant_word2, sq_vtx_constant_word3; /* high addr, stride */ @@ -155,16 +149,16 @@ set_vtx_resource(struct radeon_device *rdev, u64 gpu_addr) SQ_VTCX_SEL_Z(SQ_SEL_Z) | SQ_VTCX_SEL_W(SQ_SEL_W); - radeon_ring_write(rdev, PACKET3(PACKET3_SET_RESOURCE, 8)); - radeon_ring_write(rdev, 0x580); - radeon_ring_write(rdev, gpu_addr & 0xffffffff); - radeon_ring_write(rdev, 48 - 1); /* size */ - radeon_ring_write(rdev, sq_vtx_constant_word2); - radeon_ring_write(rdev, sq_vtx_constant_word3); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, S__SQ_CONSTANT_TYPE(SQ_TEX_VTX_VALID_BUFFER)); + radeon_ring_write(ring, PACKET3(PACKET3_SET_RESOURCE, 8)); + radeon_ring_write(ring, 0x580); + radeon_ring_write(ring, gpu_addr & 0xffffffff); + radeon_ring_write(ring, 48 - 1); /* size */ + radeon_ring_write(ring, sq_vtx_constant_word2); + radeon_ring_write(ring, sq_vtx_constant_word3); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, S__SQ_CONSTANT_TYPE(SQ_TEX_VTX_VALID_BUFFER)); if ((rdev->family == CHIP_CEDAR) || (rdev->family == CHIP_PALM) || @@ -185,6 +179,7 @@ set_tex_resource(struct radeon_device *rdev, int format, int w, int h, int pitch, u64 gpu_addr, u32 size) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; u32 sq_tex_resource_word0, sq_tex_resource_word1; u32 sq_tex_resource_word4, sq_tex_resource_word7; @@ -208,16 +203,16 @@ set_tex_resource(struct radeon_device *rdev, cp_set_surface_sync(rdev, PACKET3_TC_ACTION_ENA, size, gpu_addr); - radeon_ring_write(rdev, PACKET3(PACKET3_SET_RESOURCE, 8)); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, sq_tex_resource_word0); - radeon_ring_write(rdev, sq_tex_resource_word1); - radeon_ring_write(rdev, gpu_addr >> 8); - radeon_ring_write(rdev, gpu_addr >> 8); - radeon_ring_write(rdev, sq_tex_resource_word4); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, sq_tex_resource_word7); + radeon_ring_write(ring, PACKET3(PACKET3_SET_RESOURCE, 8)); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, sq_tex_resource_word0); + radeon_ring_write(ring, sq_tex_resource_word1); + radeon_ring_write(ring, gpu_addr >> 8); + radeon_ring_write(ring, gpu_addr >> 8); + radeon_ring_write(ring, sq_tex_resource_word4); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, sq_tex_resource_word7); } /* emits 12 */ @@ -225,53 +220,55 @@ static void set_scissors(struct radeon_device *rdev, int x1, int y1, int x2, int y2) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; /* workaround some hw bugs */ if (x2 == 0) x1 = 1; if (y2 == 0) y1 = 1; - if (rdev->family == CHIP_CAYMAN) { + if (rdev->family >= CHIP_CAYMAN) { if ((x2 == 1) && (y2 == 1)) x2 = 2; } - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); - radeon_ring_write(rdev, (PA_SC_SCREEN_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_START) >> 2); - radeon_ring_write(rdev, (x1 << 0) | (y1 << 16)); - radeon_ring_write(rdev, (x2 << 0) | (y2 << 16)); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); + radeon_ring_write(ring, (PA_SC_SCREEN_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_START) >> 2); + radeon_ring_write(ring, (x1 << 0) | (y1 << 16)); + radeon_ring_write(ring, (x2 << 0) | (y2 << 16)); - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); - radeon_ring_write(rdev, (PA_SC_GENERIC_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_START) >> 2); - radeon_ring_write(rdev, (x1 << 0) | (y1 << 16) | (1 << 31)); - radeon_ring_write(rdev, (x2 << 0) | (y2 << 16)); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); + radeon_ring_write(ring, (PA_SC_GENERIC_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_START) >> 2); + radeon_ring_write(ring, (x1 << 0) | (y1 << 16) | (1 << 31)); + radeon_ring_write(ring, (x2 << 0) | (y2 << 16)); - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); - radeon_ring_write(rdev, (PA_SC_WINDOW_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_START) >> 2); - radeon_ring_write(rdev, (x1 << 0) | (y1 << 16) | (1 << 31)); - radeon_ring_write(rdev, (x2 << 0) | (y2 << 16)); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); + radeon_ring_write(ring, (PA_SC_WINDOW_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_START) >> 2); + radeon_ring_write(ring, (x1 << 0) | (y1 << 16) | (1 << 31)); + radeon_ring_write(ring, (x2 << 0) | (y2 << 16)); } /* emits 10 */ static void draw_auto(struct radeon_device *rdev) { - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1)); - radeon_ring_write(rdev, (VGT_PRIMITIVE_TYPE - PACKET3_SET_CONFIG_REG_START) >> 2); - radeon_ring_write(rdev, DI_PT_RECTLIST); + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + radeon_ring_write(ring, (VGT_PRIMITIVE_TYPE - PACKET3_SET_CONFIG_REG_START) >> 2); + radeon_ring_write(ring, DI_PT_RECTLIST); - radeon_ring_write(rdev, PACKET3(PACKET3_INDEX_TYPE, 0)); - radeon_ring_write(rdev, + radeon_ring_write(ring, PACKET3(PACKET3_INDEX_TYPE, 0)); + radeon_ring_write(ring, #ifdef __BIG_ENDIAN (2 << 2) | #endif DI_INDEX_SIZE_16_BIT); - radeon_ring_write(rdev, PACKET3(PACKET3_NUM_INSTANCES, 0)); - radeon_ring_write(rdev, 1); + radeon_ring_write(ring, PACKET3(PACKET3_NUM_INSTANCES, 0)); + radeon_ring_write(ring, 1); - radeon_ring_write(rdev, PACKET3(PACKET3_DRAW_INDEX_AUTO, 1)); - radeon_ring_write(rdev, 3); - radeon_ring_write(rdev, DI_SRC_SEL_AUTO_INDEX); + radeon_ring_write(ring, PACKET3(PACKET3_DRAW_INDEX_AUTO, 1)); + radeon_ring_write(ring, 3); + radeon_ring_write(ring, DI_SRC_SEL_AUTO_INDEX); } @@ -279,6 +276,7 @@ draw_auto(struct radeon_device *rdev) static void set_default_state(struct radeon_device *rdev) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; u32 sq_config, sq_gpr_resource_mgmt_1, sq_gpr_resource_mgmt_2, sq_gpr_resource_mgmt_3; u32 sq_thread_resource_mgmt, sq_thread_resource_mgmt_2; u32 sq_stack_resource_mgmt_1, sq_stack_resource_mgmt_2, sq_stack_resource_mgmt_3; @@ -292,8 +290,8 @@ set_default_state(struct radeon_device *rdev) int dwords; /* set clear context state */ - radeon_ring_write(rdev, PACKET3(PACKET3_CLEAR_STATE, 0)); - radeon_ring_write(rdev, 0); + radeon_ring_write(ring, PACKET3(PACKET3_CLEAR_STATE, 0)); + radeon_ring_write(ring, 0); if (rdev->family < CHIP_CAYMAN) { switch (rdev->family) { @@ -550,60 +548,60 @@ set_default_state(struct radeon_device *rdev) NUM_LS_STACK_ENTRIES(num_ls_stack_entries)); /* disable dyn gprs */ - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1)); - radeon_ring_write(rdev, (SQ_DYN_GPR_CNTL_PS_FLUSH_REQ - PACKET3_SET_CONFIG_REG_START) >> 2); - radeon_ring_write(rdev, 0); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + radeon_ring_write(ring, (SQ_DYN_GPR_CNTL_PS_FLUSH_REQ - PACKET3_SET_CONFIG_REG_START) >> 2); + radeon_ring_write(ring, 0); /* setup LDS */ - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1)); - radeon_ring_write(rdev, (SQ_LDS_RESOURCE_MGMT - PACKET3_SET_CONFIG_REG_START) >> 2); - radeon_ring_write(rdev, 0x10001000); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + radeon_ring_write(ring, (SQ_LDS_RESOURCE_MGMT - PACKET3_SET_CONFIG_REG_START) >> 2); + radeon_ring_write(ring, 0x10001000); /* SQ config */ - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 11)); - radeon_ring_write(rdev, (SQ_CONFIG - PACKET3_SET_CONFIG_REG_START) >> 2); - radeon_ring_write(rdev, sq_config); - radeon_ring_write(rdev, sq_gpr_resource_mgmt_1); - radeon_ring_write(rdev, sq_gpr_resource_mgmt_2); - radeon_ring_write(rdev, sq_gpr_resource_mgmt_3); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, sq_thread_resource_mgmt); - radeon_ring_write(rdev, sq_thread_resource_mgmt_2); - radeon_ring_write(rdev, sq_stack_resource_mgmt_1); - radeon_ring_write(rdev, sq_stack_resource_mgmt_2); - radeon_ring_write(rdev, sq_stack_resource_mgmt_3); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 11)); + radeon_ring_write(ring, (SQ_CONFIG - PACKET3_SET_CONFIG_REG_START) >> 2); + radeon_ring_write(ring, sq_config); + radeon_ring_write(ring, sq_gpr_resource_mgmt_1); + radeon_ring_write(ring, sq_gpr_resource_mgmt_2); + radeon_ring_write(ring, sq_gpr_resource_mgmt_3); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, sq_thread_resource_mgmt); + radeon_ring_write(ring, sq_thread_resource_mgmt_2); + radeon_ring_write(ring, sq_stack_resource_mgmt_1); + radeon_ring_write(ring, sq_stack_resource_mgmt_2); + radeon_ring_write(ring, sq_stack_resource_mgmt_3); } /* CONTEXT_CONTROL */ - radeon_ring_write(rdev, 0xc0012800); - radeon_ring_write(rdev, 0x80000000); - radeon_ring_write(rdev, 0x80000000); + radeon_ring_write(ring, 0xc0012800); + radeon_ring_write(ring, 0x80000000); + radeon_ring_write(ring, 0x80000000); /* SQ_VTX_BASE_VTX_LOC */ - radeon_ring_write(rdev, 0xc0026f00); - radeon_ring_write(rdev, 0x00000000); - radeon_ring_write(rdev, 0x00000000); - radeon_ring_write(rdev, 0x00000000); + radeon_ring_write(ring, 0xc0026f00); + radeon_ring_write(ring, 0x00000000); + radeon_ring_write(ring, 0x00000000); + radeon_ring_write(ring, 0x00000000); /* SET_SAMPLER */ - radeon_ring_write(rdev, 0xc0036e00); - radeon_ring_write(rdev, 0x00000000); - radeon_ring_write(rdev, 0x00000012); - radeon_ring_write(rdev, 0x00000000); - radeon_ring_write(rdev, 0x00000000); + radeon_ring_write(ring, 0xc0036e00); + radeon_ring_write(ring, 0x00000000); + radeon_ring_write(ring, 0x00000012); + radeon_ring_write(ring, 0x00000000); + radeon_ring_write(ring, 0x00000000); /* set to DX10/11 mode */ - radeon_ring_write(rdev, PACKET3(PACKET3_MODE_CONTROL, 0)); - radeon_ring_write(rdev, 1); + radeon_ring_write(ring, PACKET3(PACKET3_MODE_CONTROL, 0)); + radeon_ring_write(ring, 1); /* emit an IB pointing at default state */ dwords = ALIGN(rdev->r600_blit.state_len, 0x10); gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.state_offset; - radeon_ring_write(rdev, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); - radeon_ring_write(rdev, gpu_addr & 0xFFFFFFFC); - radeon_ring_write(rdev, upper_32_bits(gpu_addr) & 0xFF); - radeon_ring_write(rdev, dwords); + radeon_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); + radeon_ring_write(ring, gpu_addr & 0xFFFFFFFC); + radeon_ring_write(ring, upper_32_bits(gpu_addr) & 0xFF); + radeon_ring_write(ring, dwords); } diff --git a/drivers/gpu/drm/radeon/evergreen_blit_shaders.c b/drivers/gpu/drm/radeon/evergreen_blit_shaders.c index 3a10399..f85c0af 100644 --- a/drivers/gpu/drm/radeon/evergreen_blit_shaders.c +++ b/drivers/gpu/drm/radeon/evergreen_blit_shaders.c @@ -24,6 +24,7 @@ * Alex Deucher */ +#include #include #include diff --git a/drivers/gpu/drm/radeon/evergreen_cs.c b/drivers/gpu/drm/radeon/evergreen_cs.c index cd4590a..2cbd369 100644 --- a/drivers/gpu/drm/radeon/evergreen_cs.c +++ b/drivers/gpu/drm/radeon/evergreen_cs.c @@ -31,6 +31,9 @@ #include "evergreen_reg_safe.h" #include "cayman_reg_safe.h" +#define MAX(a,b) (((a)>(b))?(a):(b)) +#define MIN(a,b) (((a)<(b))?(a):(b)) + static int evergreen_cs_packet_next_reloc(struct radeon_cs_parser *p, struct radeon_cs_reloc **cs_reloc); @@ -40,42 +43,47 @@ struct evergreen_cs_track { u32 npipes; u32 row_size; /* value we track */ - u32 nsamples; - u32 cb_color_base_last[12]; + u32 nsamples; /* unused */ struct radeon_bo *cb_color_bo[12]; u32 cb_color_bo_offset[12]; - struct radeon_bo *cb_color_fmask_bo[8]; - struct radeon_bo *cb_color_cmask_bo[8]; + struct radeon_bo *cb_color_fmask_bo[8]; /* unused */ + struct radeon_bo *cb_color_cmask_bo[8]; /* unused */ u32 cb_color_info[12]; u32 cb_color_view[12]; - u32 cb_color_pitch_idx[12]; - u32 cb_color_slice_idx[12]; - u32 cb_color_dim_idx[12]; - u32 cb_color_dim[12]; u32 cb_color_pitch[12]; u32 cb_color_slice[12]; - u32 cb_color_cmask_slice[8]; - u32 cb_color_fmask_slice[8]; + u32 cb_color_slice_idx[12]; + u32 cb_color_attrib[12]; + u32 cb_color_cmask_slice[8];/* unused */ + u32 cb_color_fmask_slice[8];/* unused */ u32 cb_target_mask; - u32 cb_shader_mask; + u32 cb_shader_mask; /* unused */ u32 vgt_strmout_config; u32 vgt_strmout_buffer_config; + struct radeon_bo *vgt_strmout_bo[4]; + u32 vgt_strmout_bo_offset[4]; + u32 vgt_strmout_size[4]; u32 db_depth_control; u32 db_depth_view; + u32 db_depth_slice; u32 db_depth_size; - u32 db_depth_size_idx; u32 db_z_info; - u32 db_z_idx; u32 db_z_read_offset; u32 db_z_write_offset; struct radeon_bo *db_z_read_bo; struct radeon_bo *db_z_write_bo; u32 db_s_info; - u32 db_s_idx; u32 db_s_read_offset; u32 db_s_write_offset; struct radeon_bo *db_s_read_bo; struct radeon_bo *db_s_write_bo; + bool sx_misc_kill_all_prims; + bool cb_dirty; + bool db_dirty; + bool streamout_dirty; + u32 htile_offset; + u32 htile_surface; + struct radeon_bo *htile_bo; }; static u32 evergreen_cs_get_aray_mode(u32 tiling_flags) @@ -103,19 +111,6 @@ static u32 evergreen_cs_get_num_banks(u32 nbanks) } } -static u32 evergreen_cs_get_tile_split(u32 row_size) -{ - switch (row_size) { - case 1: - default: - return ADDR_SURF_TILE_SPLIT_1KB; - case 2: - return ADDR_SURF_TILE_SPLIT_2KB; - case 4: - return ADDR_SURF_TILE_SPLIT_4KB; - } -} - static void evergreen_cs_track_init(struct evergreen_cs_track *track) { int i; @@ -128,50 +123,858 @@ static void evergreen_cs_track_init(struct evergreen_cs_track *track) } for (i = 0; i < 12; i++) { - track->cb_color_base_last[i] = 0; track->cb_color_bo[i] = NULL; track->cb_color_bo_offset[i] = 0xFFFFFFFF; track->cb_color_info[i] = 0; - track->cb_color_view[i] = 0; - track->cb_color_pitch_idx[i] = 0; - track->cb_color_slice_idx[i] = 0; - track->cb_color_dim[i] = 0; + track->cb_color_view[i] = 0xFFFFFFFF; track->cb_color_pitch[i] = 0; - track->cb_color_slice[i] = 0; - track->cb_color_dim[i] = 0; + track->cb_color_slice[i] = 0xfffffff; + track->cb_color_slice_idx[i] = 0; } track->cb_target_mask = 0xFFFFFFFF; track->cb_shader_mask = 0xFFFFFFFF; + track->cb_dirty = true; + track->db_depth_slice = 0xffffffff; track->db_depth_view = 0xFFFFC000; track->db_depth_size = 0xFFFFFFFF; - track->db_depth_size_idx = 0; track->db_depth_control = 0xFFFFFFFF; track->db_z_info = 0xFFFFFFFF; - track->db_z_idx = 0xFFFFFFFF; track->db_z_read_offset = 0xFFFFFFFF; track->db_z_write_offset = 0xFFFFFFFF; track->db_z_read_bo = NULL; track->db_z_write_bo = NULL; track->db_s_info = 0xFFFFFFFF; - track->db_s_idx = 0xFFFFFFFF; track->db_s_read_offset = 0xFFFFFFFF; track->db_s_write_offset = 0xFFFFFFFF; track->db_s_read_bo = NULL; track->db_s_write_bo = NULL; + track->db_dirty = true; + track->htile_bo = NULL; + track->htile_offset = 0xFFFFFFFF; + track->htile_surface = 0; + + for (i = 0; i < 4; i++) { + track->vgt_strmout_size[i] = 0; + track->vgt_strmout_bo[i] = NULL; + track->vgt_strmout_bo_offset[i] = 0xFFFFFFFF; + } + track->streamout_dirty = true; + track->sx_misc_kill_all_prims = false; } -static int evergreen_cs_track_check(struct radeon_cs_parser *p) +struct eg_surface { + /* value gathered from cs */ + unsigned nbx; + unsigned nby; + unsigned format; + unsigned mode; + unsigned nbanks; + unsigned bankw; + unsigned bankh; + unsigned tsplit; + unsigned mtilea; + unsigned nsamples; + /* output value */ + unsigned bpe; + unsigned layer_size; + unsigned palign; + unsigned halign; + unsigned long base_align; +}; + +static int evergreen_surface_check_linear(struct radeon_cs_parser *p, + struct eg_surface *surf, + const char *prefix) +{ + surf->layer_size = surf->nbx * surf->nby * surf->bpe * surf->nsamples; + surf->base_align = surf->bpe; + surf->palign = 1; + surf->halign = 1; + return 0; +} + +static int evergreen_surface_check_linear_aligned(struct radeon_cs_parser *p, + struct eg_surface *surf, + const char *prefix) +{ + struct evergreen_cs_track *track = p->track; + unsigned palign; + + palign = MAX(64, track->group_size / surf->bpe); + surf->layer_size = surf->nbx * surf->nby * surf->bpe * surf->nsamples; + surf->base_align = track->group_size; + surf->palign = palign; + surf->halign = 1; + if (surf->nbx & (palign - 1)) { + if (prefix) { + dev_warn(p->dev, "%s:%d %s pitch %d invalid must be aligned with %d\n", + __func__, __LINE__, prefix, surf->nbx, palign); + } + return -EINVAL; + } + return 0; +} + +static int evergreen_surface_check_1d(struct radeon_cs_parser *p, + struct eg_surface *surf, + const char *prefix) +{ + struct evergreen_cs_track *track = p->track; + unsigned palign; + + palign = track->group_size / (8 * surf->bpe * surf->nsamples); + palign = MAX(8, palign); + surf->layer_size = surf->nbx * surf->nby * surf->bpe; + surf->base_align = track->group_size; + surf->palign = palign; + surf->halign = 8; + if ((surf->nbx & (palign - 1))) { + if (prefix) { + dev_warn(p->dev, "%s:%d %s pitch %d invalid must be aligned with %d (%d %d %d)\n", + __func__, __LINE__, prefix, surf->nbx, palign, + track->group_size, surf->bpe, surf->nsamples); + } + return -EINVAL; + } + if ((surf->nby & (8 - 1))) { + if (prefix) { + dev_warn(p->dev, "%s:%d %s height %d invalid must be aligned with 8\n", + __func__, __LINE__, prefix, surf->nby); + } + return -EINVAL; + } + return 0; +} + +static int evergreen_surface_check_2d(struct radeon_cs_parser *p, + struct eg_surface *surf, + const char *prefix) +{ + struct evergreen_cs_track *track = p->track; + unsigned palign, halign, tileb, slice_pt; + unsigned mtile_pr, mtile_ps, mtileb; + + tileb = 64 * surf->bpe * surf->nsamples; + slice_pt = 1; + if (tileb > surf->tsplit) { + slice_pt = tileb / surf->tsplit; + } + tileb = tileb / slice_pt; + /* macro tile width & height */ + palign = (8 * surf->bankw * track->npipes) * surf->mtilea; + halign = (8 * surf->bankh * surf->nbanks) / surf->mtilea; + mtileb = (palign / 8) * (halign / 8) * tileb;; + mtile_pr = surf->nbx / palign; + mtile_ps = (mtile_pr * surf->nby) / halign; + surf->layer_size = mtile_ps * mtileb * slice_pt; + surf->base_align = (palign / 8) * (halign / 8) * tileb; + surf->palign = palign; + surf->halign = halign; + + if ((surf->nbx & (palign - 1))) { + if (prefix) { + dev_warn(p->dev, "%s:%d %s pitch %d invalid must be aligned with %d\n", + __func__, __LINE__, prefix, surf->nbx, palign); + } + return -EINVAL; + } + if ((surf->nby & (halign - 1))) { + if (prefix) { + dev_warn(p->dev, "%s:%d %s height %d invalid must be aligned with %d\n", + __func__, __LINE__, prefix, surf->nby, halign); + } + return -EINVAL; + } + + return 0; +} + +static int evergreen_surface_check(struct radeon_cs_parser *p, + struct eg_surface *surf, + const char *prefix) +{ + /* some common value computed here */ + surf->bpe = r600_fmt_get_blocksize(surf->format); + + switch (surf->mode) { + case ARRAY_LINEAR_GENERAL: + return evergreen_surface_check_linear(p, surf, prefix); + case ARRAY_LINEAR_ALIGNED: + return evergreen_surface_check_linear_aligned(p, surf, prefix); + case ARRAY_1D_TILED_THIN1: + return evergreen_surface_check_1d(p, surf, prefix); + case ARRAY_2D_TILED_THIN1: + return evergreen_surface_check_2d(p, surf, prefix); + default: + dev_warn(p->dev, "%s:%d %s invalid array mode %d\n", + __func__, __LINE__, prefix, surf->mode); + return -EINVAL; + } + return -EINVAL; +} + +static int evergreen_surface_value_conv_check(struct radeon_cs_parser *p, + struct eg_surface *surf, + const char *prefix) +{ + switch (surf->mode) { + case ARRAY_2D_TILED_THIN1: + break; + case ARRAY_LINEAR_GENERAL: + case ARRAY_LINEAR_ALIGNED: + case ARRAY_1D_TILED_THIN1: + return 0; + default: + dev_warn(p->dev, "%s:%d %s invalid array mode %d\n", + __func__, __LINE__, prefix, surf->mode); + return -EINVAL; + } + + switch (surf->nbanks) { + case 0: surf->nbanks = 2; break; + case 1: surf->nbanks = 4; break; + case 2: surf->nbanks = 8; break; + case 3: surf->nbanks = 16; break; + default: + dev_warn(p->dev, "%s:%d %s invalid number of banks %d\n", + __func__, __LINE__, prefix, surf->nbanks); + return -EINVAL; + } + switch (surf->bankw) { + case 0: surf->bankw = 1; break; + case 1: surf->bankw = 2; break; + case 2: surf->bankw = 4; break; + case 3: surf->bankw = 8; break; + default: + dev_warn(p->dev, "%s:%d %s invalid bankw %d\n", + __func__, __LINE__, prefix, surf->bankw); + return -EINVAL; + } + switch (surf->bankh) { + case 0: surf->bankh = 1; break; + case 1: surf->bankh = 2; break; + case 2: surf->bankh = 4; break; + case 3: surf->bankh = 8; break; + default: + dev_warn(p->dev, "%s:%d %s invalid bankh %d\n", + __func__, __LINE__, prefix, surf->bankh); + return -EINVAL; + } + switch (surf->mtilea) { + case 0: surf->mtilea = 1; break; + case 1: surf->mtilea = 2; break; + case 2: surf->mtilea = 4; break; + case 3: surf->mtilea = 8; break; + default: + dev_warn(p->dev, "%s:%d %s invalid macro tile aspect %d\n", + __func__, __LINE__, prefix, surf->mtilea); + return -EINVAL; + } + switch (surf->tsplit) { + case 0: surf->tsplit = 64; break; + case 1: surf->tsplit = 128; break; + case 2: surf->tsplit = 256; break; + case 3: surf->tsplit = 512; break; + case 4: surf->tsplit = 1024; break; + case 5: surf->tsplit = 2048; break; + case 6: surf->tsplit = 4096; break; + default: + dev_warn(p->dev, "%s:%d %s invalid tile split %d\n", + __func__, __LINE__, prefix, surf->tsplit); + return -EINVAL; + } + return 0; +} + +static int evergreen_cs_track_validate_cb(struct radeon_cs_parser *p, unsigned id) +{ + struct evergreen_cs_track *track = p->track; + struct eg_surface surf; + unsigned pitch, slice, mslice; + unsigned long offset; + int r; + + mslice = G_028C6C_SLICE_MAX(track->cb_color_view[id]) + 1; + pitch = track->cb_color_pitch[id]; + slice = track->cb_color_slice[id]; + surf.nbx = (pitch + 1) * 8; + surf.nby = ((slice + 1) * 64) / surf.nbx; + surf.mode = G_028C70_ARRAY_MODE(track->cb_color_info[id]); + surf.format = G_028C70_FORMAT(track->cb_color_info[id]); + surf.tsplit = G_028C74_TILE_SPLIT(track->cb_color_attrib[id]); + surf.nbanks = G_028C74_NUM_BANKS(track->cb_color_attrib[id]); + surf.bankw = G_028C74_BANK_WIDTH(track->cb_color_attrib[id]); + surf.bankh = G_028C74_BANK_HEIGHT(track->cb_color_attrib[id]); + surf.mtilea = G_028C74_MACRO_TILE_ASPECT(track->cb_color_attrib[id]); + surf.nsamples = 1; + + if (!r600_fmt_is_valid_color(surf.format)) { + dev_warn(p->dev, "%s:%d cb invalid format %d for %d (0x%08x)\n", + __func__, __LINE__, surf.format, + id, track->cb_color_info[id]); + return -EINVAL; + } + + r = evergreen_surface_value_conv_check(p, &surf, "cb"); + if (r) { + return r; + } + + r = evergreen_surface_check(p, &surf, "cb"); + if (r) { + dev_warn(p->dev, "%s:%d cb[%d] invalid (0x%08x 0x%08x 0x%08x 0x%08x)\n", + __func__, __LINE__, id, track->cb_color_pitch[id], + track->cb_color_slice[id], track->cb_color_attrib[id], + track->cb_color_info[id]); + return r; + } + + offset = track->cb_color_bo_offset[id] << 8; + if (offset & (surf.base_align - 1)) { + dev_warn(p->dev, "%s:%d cb[%d] bo base %ld not aligned with %ld\n", + __func__, __LINE__, id, offset, surf.base_align); + return -EINVAL; + } + + offset += surf.layer_size * mslice; + if (offset > radeon_bo_size(track->cb_color_bo[id])) { + /* old ddx are broken they allocate bo with w*h*bpp but + * program slice with ALIGN(h, 8), catch this and patch + * command stream. + */ + if (!surf.mode) { + volatile u32 *ib = p->ib->ptr; + unsigned long tmp, nby, bsize, size, min = 0; + + /* find the height the ddx wants */ + if (surf.nby > 8) { + min = surf.nby - 8; + } + bsize = radeon_bo_size(track->cb_color_bo[id]); + tmp = track->cb_color_bo_offset[id] << 8; + for (nby = surf.nby; nby > min; nby--) { + size = nby * surf.nbx * surf.bpe * surf.nsamples; + if ((tmp + size * mslice) <= bsize) { + break; + } + } + if (nby > min) { + surf.nby = nby; + slice = ((nby * surf.nbx) / 64) - 1; + if (!evergreen_surface_check(p, &surf, "cb")) { + /* check if this one works */ + tmp += surf.layer_size * mslice; + if (tmp <= bsize) { + ib[track->cb_color_slice_idx[id]] = slice; + goto old_ddx_ok; + } + } + } + } + dev_warn(p->dev, "%s:%d cb[%d] bo too small (layer size %d, " + "offset %d, max layer %d, bo size %ld, slice %d)\n", + __func__, __LINE__, id, surf.layer_size, + track->cb_color_bo_offset[id] << 8, mslice, + radeon_bo_size(track->cb_color_bo[id]), slice); + dev_warn(p->dev, "%s:%d problematic surf: (%d %d) (%d %d %d %d %d %d %d)\n", + __func__, __LINE__, surf.nbx, surf.nby, + surf.mode, surf.bpe, surf.nsamples, + surf.bankw, surf.bankh, + surf.tsplit, surf.mtilea); + return -EINVAL; + } +old_ddx_ok: + + return 0; +} + +static int evergreen_cs_track_validate_htile(struct radeon_cs_parser *p, + unsigned nbx, unsigned nby) +{ + struct evergreen_cs_track *track = p->track; + unsigned long size; + + if (track->htile_bo == NULL) { + dev_warn(p->dev, "%s:%d htile enabled without htile surface 0x%08x\n", + __func__, __LINE__, track->db_z_info); + return -EINVAL; + } + + if (G_028ABC_LINEAR(track->htile_surface)) { + /* pitch must be 16 htiles aligned == 16 * 8 pixel aligned */ + nbx = round_up(nbx, 16 * 8); + /* height is npipes htiles aligned == npipes * 8 pixel aligned */ + nby = round_up(nby, track->npipes * 8); + } else { + switch (track->npipes) { + case 8: + nbx = round_up(nbx, 64 * 8); + nby = round_up(nby, 64 * 8); + break; + case 4: + nbx = round_up(nbx, 64 * 8); + nby = round_up(nby, 32 * 8); + break; + case 2: + nbx = round_up(nbx, 32 * 8); + nby = round_up(nby, 32 * 8); + break; + case 1: + nbx = round_up(nbx, 32 * 8); + nby = round_up(nby, 16 * 8); + break; + default: + dev_warn(p->dev, "%s:%d invalid num pipes %d\n", + __func__, __LINE__, track->npipes); + return -EINVAL; + } + } + /* compute number of htile */ + nbx = nbx / 8; + nby = nby / 8; + size = nbx * nby * 4; + size += track->htile_offset; + + if (size > radeon_bo_size(track->htile_bo)) { + dev_warn(p->dev, "%s:%d htile surface too small %ld for %ld (%d %d)\n", + __func__, __LINE__, radeon_bo_size(track->htile_bo), + size, nbx, nby); + return -EINVAL; + } + return 0; +} + +static int evergreen_cs_track_validate_stencil(struct radeon_cs_parser *p) +{ + struct evergreen_cs_track *track = p->track; + struct eg_surface surf; + unsigned pitch, slice, mslice; + unsigned long offset; + int r; + + mslice = G_028008_SLICE_MAX(track->db_depth_view) + 1; + pitch = G_028058_PITCH_TILE_MAX(track->db_depth_size); + slice = track->db_depth_slice; + surf.nbx = (pitch + 1) * 8; + surf.nby = ((slice + 1) * 64) / surf.nbx; + surf.mode = G_028040_ARRAY_MODE(track->db_z_info); + surf.format = G_028044_FORMAT(track->db_s_info); + surf.tsplit = G_028044_TILE_SPLIT(track->db_s_info); + surf.nbanks = G_028040_NUM_BANKS(track->db_z_info); + surf.bankw = G_028040_BANK_WIDTH(track->db_z_info); + surf.bankh = G_028040_BANK_HEIGHT(track->db_z_info); + surf.mtilea = G_028040_MACRO_TILE_ASPECT(track->db_z_info); + surf.nsamples = 1; + + if (surf.format != 1) { + dev_warn(p->dev, "%s:%d stencil invalid format %d\n", + __func__, __LINE__, surf.format); + return -EINVAL; + } + /* replace by color format so we can use same code */ + surf.format = V_028C70_COLOR_8; + + r = evergreen_surface_value_conv_check(p, &surf, "stencil"); + if (r) { + return r; + } + + r = evergreen_surface_check(p, &surf, NULL); + if (r) { + /* old userspace doesn't compute proper depth/stencil alignment + * check that alignment against a bigger byte per elements and + * only report if that alignment is wrong too. + */ + surf.format = V_028C70_COLOR_8_8_8_8; + r = evergreen_surface_check(p, &surf, "stencil"); + if (r) { + dev_warn(p->dev, "%s:%d stencil invalid (0x%08x 0x%08x 0x%08x 0x%08x)\n", + __func__, __LINE__, track->db_depth_size, + track->db_depth_slice, track->db_s_info, track->db_z_info); + } + return r; + } + + offset = track->db_s_read_offset << 8; + if (offset & (surf.base_align - 1)) { + dev_warn(p->dev, "%s:%d stencil read bo base %ld not aligned with %ld\n", + __func__, __LINE__, offset, surf.base_align); + return -EINVAL; + } + offset += surf.layer_size * mslice; + if (offset > radeon_bo_size(track->db_s_read_bo)) { + dev_warn(p->dev, "%s:%d stencil read bo too small (layer size %d, " + "offset %ld, max layer %d, bo size %ld)\n", + __func__, __LINE__, surf.layer_size, + (unsigned long)track->db_s_read_offset << 8, mslice, + radeon_bo_size(track->db_s_read_bo)); + dev_warn(p->dev, "%s:%d stencil invalid (0x%08x 0x%08x 0x%08x 0x%08x)\n", + __func__, __LINE__, track->db_depth_size, + track->db_depth_slice, track->db_s_info, track->db_z_info); + return -EINVAL; + } + + offset = track->db_s_write_offset << 8; + if (offset & (surf.base_align - 1)) { + dev_warn(p->dev, "%s:%d stencil write bo base %ld not aligned with %ld\n", + __func__, __LINE__, offset, surf.base_align); + return -EINVAL; + } + offset += surf.layer_size * mslice; + if (offset > radeon_bo_size(track->db_s_write_bo)) { + dev_warn(p->dev, "%s:%d stencil write bo too small (layer size %d, " + "offset %ld, max layer %d, bo size %ld)\n", + __func__, __LINE__, surf.layer_size, + (unsigned long)track->db_s_write_offset << 8, mslice, + radeon_bo_size(track->db_s_write_bo)); + return -EINVAL; + } + + /* hyperz */ + if (G_028040_TILE_SURFACE_ENABLE(track->db_z_info)) { + r = evergreen_cs_track_validate_htile(p, surf.nbx, surf.nby); + if (r) { + return r; + } + } + + return 0; +} + +static int evergreen_cs_track_validate_depth(struct radeon_cs_parser *p) { struct evergreen_cs_track *track = p->track; + struct eg_surface surf; + unsigned pitch, slice, mslice; + unsigned long offset; + int r; + + mslice = G_028008_SLICE_MAX(track->db_depth_view) + 1; + pitch = G_028058_PITCH_TILE_MAX(track->db_depth_size); + slice = track->db_depth_slice; + surf.nbx = (pitch + 1) * 8; + surf.nby = ((slice + 1) * 64) / surf.nbx; + surf.mode = G_028040_ARRAY_MODE(track->db_z_info); + surf.format = G_028040_FORMAT(track->db_z_info); + surf.tsplit = G_028040_TILE_SPLIT(track->db_z_info); + surf.nbanks = G_028040_NUM_BANKS(track->db_z_info); + surf.bankw = G_028040_BANK_WIDTH(track->db_z_info); + surf.bankh = G_028040_BANK_HEIGHT(track->db_z_info); + surf.mtilea = G_028040_MACRO_TILE_ASPECT(track->db_z_info); + surf.nsamples = 1; + + switch (surf.format) { + case V_028040_Z_16: + surf.format = V_028C70_COLOR_16; + break; + case V_028040_Z_24: + case V_028040_Z_32_FLOAT: + surf.format = V_028C70_COLOR_8_8_8_8; + break; + default: + dev_warn(p->dev, "%s:%d depth invalid format %d\n", + __func__, __LINE__, surf.format); + return -EINVAL; + } + + r = evergreen_surface_value_conv_check(p, &surf, "depth"); + if (r) { + dev_warn(p->dev, "%s:%d depth invalid (0x%08x 0x%08x 0x%08x)\n", + __func__, __LINE__, track->db_depth_size, + track->db_depth_slice, track->db_z_info); + return r; + } - /* we don't support stream out buffer yet */ - if (track->vgt_strmout_config || track->vgt_strmout_buffer_config) { - dev_warn(p->dev, "this kernel doesn't support SMX output buffer\n"); + r = evergreen_surface_check(p, &surf, "depth"); + if (r) { + dev_warn(p->dev, "%s:%d depth invalid (0x%08x 0x%08x 0x%08x)\n", + __func__, __LINE__, track->db_depth_size, + track->db_depth_slice, track->db_z_info); + return r; + } + + offset = track->db_z_read_offset << 8; + if (offset & (surf.base_align - 1)) { + dev_warn(p->dev, "%s:%d stencil read bo base %ld not aligned with %ld\n", + __func__, __LINE__, offset, surf.base_align); + return -EINVAL; + } + offset += surf.layer_size * mslice; + if (offset > radeon_bo_size(track->db_z_read_bo)) { + dev_warn(p->dev, "%s:%d depth read bo too small (layer size %d, " + "offset %ld, max layer %d, bo size %ld)\n", + __func__, __LINE__, surf.layer_size, + (unsigned long)track->db_z_read_offset << 8, mslice, + radeon_bo_size(track->db_z_read_bo)); + return -EINVAL; + } + + offset = track->db_z_write_offset << 8; + if (offset & (surf.base_align - 1)) { + dev_warn(p->dev, "%s:%d stencil write bo base %ld not aligned with %ld\n", + __func__, __LINE__, offset, surf.base_align); + return -EINVAL; + } + offset += surf.layer_size * mslice; + if (offset > radeon_bo_size(track->db_z_write_bo)) { + dev_warn(p->dev, "%s:%d depth write bo too small (layer size %d, " + "offset %ld, max layer %d, bo size %ld)\n", + __func__, __LINE__, surf.layer_size, + (unsigned long)track->db_z_write_offset << 8, mslice, + radeon_bo_size(track->db_z_write_bo)); return -EINVAL; } - /* XXX fill in */ + /* hyperz */ + if (G_028040_TILE_SURFACE_ENABLE(track->db_z_info)) { + r = evergreen_cs_track_validate_htile(p, surf.nbx, surf.nby); + if (r) { + return r; + } + } + + return 0; +} + +static int evergreen_cs_track_validate_texture(struct radeon_cs_parser *p, + struct radeon_bo *texture, + struct radeon_bo *mipmap, + unsigned idx) +{ + struct eg_surface surf; + unsigned long toffset, moffset; + unsigned dim, llevel, mslice, width, height, depth, i; + u32 texdw[8]; + int r; + + texdw[0] = radeon_get_ib_value(p, idx + 0); + texdw[1] = radeon_get_ib_value(p, idx + 1); + texdw[2] = radeon_get_ib_value(p, idx + 2); + texdw[3] = radeon_get_ib_value(p, idx + 3); + texdw[4] = radeon_get_ib_value(p, idx + 4); + texdw[5] = radeon_get_ib_value(p, idx + 5); + texdw[6] = radeon_get_ib_value(p, idx + 6); + texdw[7] = radeon_get_ib_value(p, idx + 7); + dim = G_030000_DIM(texdw[0]); + llevel = G_030014_LAST_LEVEL(texdw[5]); + mslice = G_030014_LAST_ARRAY(texdw[5]) + 1; + width = G_030000_TEX_WIDTH(texdw[0]) + 1; + height = G_030004_TEX_HEIGHT(texdw[1]) + 1; + depth = G_030004_TEX_DEPTH(texdw[1]) + 1; + surf.format = G_03001C_DATA_FORMAT(texdw[7]); + surf.nbx = (G_030000_PITCH(texdw[0]) + 1) * 8; + surf.nbx = r600_fmt_get_nblocksx(surf.format, surf.nbx); + surf.nby = r600_fmt_get_nblocksy(surf.format, height); + surf.mode = G_030004_ARRAY_MODE(texdw[1]); + surf.tsplit = G_030018_TILE_SPLIT(texdw[6]); + surf.nbanks = G_03001C_NUM_BANKS(texdw[7]); + surf.bankw = G_03001C_BANK_WIDTH(texdw[7]); + surf.bankh = G_03001C_BANK_HEIGHT(texdw[7]); + surf.mtilea = G_03001C_MACRO_TILE_ASPECT(texdw[7]); + surf.nsamples = 1; + toffset = texdw[2] << 8; + moffset = texdw[3] << 8; + + if (!r600_fmt_is_valid_texture(surf.format, p->family)) { + dev_warn(p->dev, "%s:%d texture invalid format %d\n", + __func__, __LINE__, surf.format); + return -EINVAL; + } + switch (dim) { + case V_030000_SQ_TEX_DIM_1D: + case V_030000_SQ_TEX_DIM_2D: + case V_030000_SQ_TEX_DIM_CUBEMAP: + case V_030000_SQ_TEX_DIM_1D_ARRAY: + case V_030000_SQ_TEX_DIM_2D_ARRAY: + depth = 1; + case V_030000_SQ_TEX_DIM_3D: + break; + default: + dev_warn(p->dev, "%s:%d texture invalid dimension %d\n", + __func__, __LINE__, dim); + return -EINVAL; + } + + r = evergreen_surface_value_conv_check(p, &surf, "texture"); + if (r) { + return r; + } + + /* align height */ + evergreen_surface_check(p, &surf, NULL); + surf.nby = ALIGN(surf.nby, surf.halign); + + r = evergreen_surface_check(p, &surf, "texture"); + if (r) { + dev_warn(p->dev, "%s:%d texture invalid 0x%08x 0x%08x 0x%08x 0x%08x 0x%08x 0x%08x\n", + __func__, __LINE__, texdw[0], texdw[1], texdw[4], + texdw[5], texdw[6], texdw[7]); + return r; + } + + /* check texture size */ + if (toffset & (surf.base_align - 1)) { + dev_warn(p->dev, "%s:%d texture bo base %ld not aligned with %ld\n", + __func__, __LINE__, toffset, surf.base_align); + return -EINVAL; + } + if (moffset & (surf.base_align - 1)) { + dev_warn(p->dev, "%s:%d mipmap bo base %ld not aligned with %ld\n", + __func__, __LINE__, moffset, surf.base_align); + return -EINVAL; + } + if (dim == SQ_TEX_DIM_3D) { + toffset += surf.layer_size * depth; + } else { + toffset += surf.layer_size * mslice; + } + if (toffset > radeon_bo_size(texture)) { + dev_warn(p->dev, "%s:%d texture bo too small (layer size %d, " + "offset %ld, max layer %d, depth %d, bo size %ld) (%d %d)\n", + __func__, __LINE__, surf.layer_size, + (unsigned long)texdw[2] << 8, mslice, + depth, radeon_bo_size(texture), + surf.nbx, surf.nby); + return -EINVAL; + } + + /* check mipmap size */ + for (i = 1; i <= llevel; i++) { + unsigned w, h, d; + + w = r600_mip_minify(width, i); + h = r600_mip_minify(height, i); + d = r600_mip_minify(depth, i); + surf.nbx = r600_fmt_get_nblocksx(surf.format, w); + surf.nby = r600_fmt_get_nblocksy(surf.format, h); + + switch (surf.mode) { + case ARRAY_2D_TILED_THIN1: + if (surf.nbx < surf.palign || surf.nby < surf.halign) { + surf.mode = ARRAY_1D_TILED_THIN1; + } + /* recompute alignment */ + evergreen_surface_check(p, &surf, NULL); + break; + case ARRAY_LINEAR_GENERAL: + case ARRAY_LINEAR_ALIGNED: + case ARRAY_1D_TILED_THIN1: + break; + default: + dev_warn(p->dev, "%s:%d invalid array mode %d\n", + __func__, __LINE__, surf.mode); + return -EINVAL; + } + surf.nbx = ALIGN(surf.nbx, surf.palign); + surf.nby = ALIGN(surf.nby, surf.halign); + + r = evergreen_surface_check(p, &surf, "mipmap"); + if (r) { + return r; + } + + if (dim == SQ_TEX_DIM_3D) { + moffset += surf.layer_size * d; + } else { + moffset += surf.layer_size * mslice; + } + if (moffset > radeon_bo_size(mipmap)) { + dev_warn(p->dev, "%s:%d mipmap [%d] bo too small (layer size %d, " + "offset %ld, coffset %ld, max layer %d, depth %d, " + "bo size %ld) level0 (%d %d %d)\n", + __func__, __LINE__, i, surf.layer_size, + (unsigned long)texdw[3] << 8, moffset, mslice, + d, radeon_bo_size(mipmap), + width, height, depth); + dev_warn(p->dev, "%s:%d problematic surf: (%d %d) (%d %d %d %d %d %d %d)\n", + __func__, __LINE__, surf.nbx, surf.nby, + surf.mode, surf.bpe, surf.nsamples, + surf.bankw, surf.bankh, + surf.tsplit, surf.mtilea); + return -EINVAL; + } + } + + return 0; +} + +static int evergreen_cs_track_check(struct radeon_cs_parser *p) +{ + struct evergreen_cs_track *track = p->track; + unsigned tmp, i; + int r; + unsigned buffer_mask = 0; + + /* check streamout */ + if (track->streamout_dirty && track->vgt_strmout_config) { + for (i = 0; i < 4; i++) { + if (track->vgt_strmout_config & (1 << i)) { + buffer_mask |= (track->vgt_strmout_buffer_config >> (i * 4)) & 0xf; + } + } + + for (i = 0; i < 4; i++) { + if (buffer_mask & (1 << i)) { + if (track->vgt_strmout_bo[i]) { + u64 offset = (u64)track->vgt_strmout_bo_offset[i] + + (u64)track->vgt_strmout_size[i]; + if (offset > radeon_bo_size(track->vgt_strmout_bo[i])) { + DRM_ERROR("streamout %d bo too small: 0x%llx, 0x%lx\n", + i, offset, + radeon_bo_size(track->vgt_strmout_bo[i])); + return -EINVAL; + } + } else { + dev_warn(p->dev, "No buffer for streamout %d\n", i); + return -EINVAL; + } + } + } + track->streamout_dirty = false; + } + + if (track->sx_misc_kill_all_prims) + return 0; + + /* check that we have a cb for each enabled target + */ + if (track->cb_dirty) { + tmp = track->cb_target_mask; + for (i = 0; i < 8; i++) { + if ((tmp >> (i * 4)) & 0xF) { + /* at least one component is enabled */ + if (track->cb_color_bo[i] == NULL) { + dev_warn(p->dev, "%s:%d mask 0x%08X | 0x%08X no cb for %d\n", + __func__, __LINE__, track->cb_target_mask, track->cb_shader_mask, i); + return -EINVAL; + } + /* check cb */ + r = evergreen_cs_track_validate_cb(p, i); + if (r) { + return r; + } + } + } + track->cb_dirty = false; + } + + if (track->db_dirty) { + /* Check stencil buffer */ + if (G_028800_STENCIL_ENABLE(track->db_depth_control)) { + r = evergreen_cs_track_validate_stencil(p); + if (r) + return r; + } + /* Check depth buffer */ + if (G_028800_Z_ENABLE(track->db_depth_control)) { + r = evergreen_cs_track_validate_depth(p); + if (r) + return r; + } + track->db_dirty = false; + } + return 0; } @@ -503,6 +1306,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) break; case DB_DEPTH_CONTROL: track->db_depth_control = radeon_get_ib_value(p, idx); + track->db_dirty = true; break; case CAYMAN_DB_EQAA: if (p->rdev->family < CHIP_CAYMAN) { @@ -520,7 +1324,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) break; case DB_Z_INFO: track->db_z_info = radeon_get_ib_value(p, idx); - if (!p->keep_tiling_flags) { + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { r = evergreen_cs_packet_next_reloc(p, &reloc); if (r) { dev_warn(p->dev, "bad SET_CONTEXT_REG " @@ -532,20 +1336,35 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) ib[idx] |= Z_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); track->db_z_info |= Z_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { + unsigned bankw, bankh, mtaspect, tile_split; + + evergreen_tiling_fields(reloc->lobj.tiling_flags, + &bankw, &bankh, &mtaspect, + &tile_split); ib[idx] |= DB_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks)); - ib[idx] |= DB_TILE_SPLIT(evergreen_cs_get_tile_split(track->row_size)); + ib[idx] |= DB_TILE_SPLIT(tile_split) | + DB_BANK_WIDTH(bankw) | + DB_BANK_HEIGHT(bankh) | + DB_MACRO_TILE_ASPECT(mtaspect); } } + track->db_dirty = true; break; case DB_STENCIL_INFO: track->db_s_info = radeon_get_ib_value(p, idx); + track->db_dirty = true; break; case DB_DEPTH_VIEW: track->db_depth_view = radeon_get_ib_value(p, idx); + track->db_dirty = true; break; case DB_DEPTH_SIZE: track->db_depth_size = radeon_get_ib_value(p, idx); - track->db_depth_size_idx = idx; + track->db_dirty = true; + break; + case R_02805C_DB_DEPTH_SLICE: + track->db_depth_slice = radeon_get_ib_value(p, idx); + track->db_dirty = true; break; case DB_Z_READ_BASE: r = evergreen_cs_packet_next_reloc(p, &reloc); @@ -557,6 +1376,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) track->db_z_read_offset = radeon_get_ib_value(p, idx); ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); track->db_z_read_bo = reloc->robj; + track->db_dirty = true; break; case DB_Z_WRITE_BASE: r = evergreen_cs_packet_next_reloc(p, &reloc); @@ -568,6 +1388,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) track->db_z_write_offset = radeon_get_ib_value(p, idx); ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); track->db_z_write_bo = reloc->robj; + track->db_dirty = true; break; case DB_STENCIL_READ_BASE: r = evergreen_cs_packet_next_reloc(p, &reloc); @@ -579,6 +1400,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) track->db_s_read_offset = radeon_get_ib_value(p, idx); ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); track->db_s_read_bo = reloc->robj; + track->db_dirty = true; break; case DB_STENCIL_WRITE_BASE: r = evergreen_cs_packet_next_reloc(p, &reloc); @@ -590,18 +1412,56 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) track->db_s_write_offset = radeon_get_ib_value(p, idx); ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); track->db_s_write_bo = reloc->robj; + track->db_dirty = true; break; case VGT_STRMOUT_CONFIG: track->vgt_strmout_config = radeon_get_ib_value(p, idx); + track->streamout_dirty = true; break; case VGT_STRMOUT_BUFFER_CONFIG: track->vgt_strmout_buffer_config = radeon_get_ib_value(p, idx); + track->streamout_dirty = true; break; + case VGT_STRMOUT_BUFFER_BASE_0: + case VGT_STRMOUT_BUFFER_BASE_1: + case VGT_STRMOUT_BUFFER_BASE_2: + case VGT_STRMOUT_BUFFER_BASE_3: + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "bad SET_CONTEXT_REG " + "0x%04X\n", reg); + return -EINVAL; + } + tmp = (reg - VGT_STRMOUT_BUFFER_BASE_0) / 16; + track->vgt_strmout_bo_offset[tmp] = radeon_get_ib_value(p, idx) << 8; + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + track->vgt_strmout_bo[tmp] = reloc->robj; + track->streamout_dirty = true; + break; + case VGT_STRMOUT_BUFFER_SIZE_0: + case VGT_STRMOUT_BUFFER_SIZE_1: + case VGT_STRMOUT_BUFFER_SIZE_2: + case VGT_STRMOUT_BUFFER_SIZE_3: + tmp = (reg - VGT_STRMOUT_BUFFER_SIZE_0) / 16; + /* size in register is DWs, convert to bytes */ + track->vgt_strmout_size[tmp] = radeon_get_ib_value(p, idx) * 4; + track->streamout_dirty = true; + break; + case CP_COHER_BASE: + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "missing reloc for CP_COHER_BASE " + "0x%04X\n", reg); + return -EINVAL; + } + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); case CB_TARGET_MASK: track->cb_target_mask = radeon_get_ib_value(p, idx); + track->cb_dirty = true; break; case CB_SHADER_MASK: track->cb_shader_mask = radeon_get_ib_value(p, idx); + track->cb_dirty = true; break; case PA_SC_AA_CONFIG: if (p->rdev->family >= CHIP_CAYMAN) { @@ -631,6 +1491,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) case CB_COLOR7_VIEW: tmp = (reg - CB_COLOR0_VIEW) / 0x3c; track->cb_color_view[tmp] = radeon_get_ib_value(p, idx); + track->cb_dirty = true; break; case CB_COLOR8_VIEW: case CB_COLOR9_VIEW: @@ -638,6 +1499,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) case CB_COLOR11_VIEW: tmp = ((reg - CB_COLOR8_VIEW) / 0x1c) + 8; track->cb_color_view[tmp] = radeon_get_ib_value(p, idx); + track->cb_dirty = true; break; case CB_COLOR0_INFO: case CB_COLOR1_INFO: @@ -649,7 +1511,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) case CB_COLOR7_INFO: tmp = (reg - CB_COLOR0_INFO) / 0x3c; track->cb_color_info[tmp] = radeon_get_ib_value(p, idx); - if (!p->keep_tiling_flags) { + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { r = evergreen_cs_packet_next_reloc(p, &reloc); if (r) { dev_warn(p->dev, "bad SET_CONTEXT_REG " @@ -659,6 +1521,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) ib[idx] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); track->cb_color_info[tmp] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); } + track->cb_dirty = true; break; case CB_COLOR8_INFO: case CB_COLOR9_INFO: @@ -666,7 +1529,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) case CB_COLOR11_INFO: tmp = ((reg - CB_COLOR8_INFO) / 0x1c) + 8; track->cb_color_info[tmp] = radeon_get_ib_value(p, idx); - if (!p->keep_tiling_flags) { + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { r = evergreen_cs_packet_next_reloc(p, &reloc); if (r) { dev_warn(p->dev, "bad SET_CONTEXT_REG " @@ -676,6 +1539,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) ib[idx] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); track->cb_color_info[tmp] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); } + track->cb_dirty = true; break; case CB_COLOR0_PITCH: case CB_COLOR1_PITCH: @@ -687,7 +1551,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) case CB_COLOR7_PITCH: tmp = (reg - CB_COLOR0_PITCH) / 0x3c; track->cb_color_pitch[tmp] = radeon_get_ib_value(p, idx); - track->cb_color_pitch_idx[tmp] = idx; + track->cb_dirty = true; break; case CB_COLOR8_PITCH: case CB_COLOR9_PITCH: @@ -695,7 +1559,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) case CB_COLOR11_PITCH: tmp = ((reg - CB_COLOR8_PITCH) / 0x1c) + 8; track->cb_color_pitch[tmp] = radeon_get_ib_value(p, idx); - track->cb_color_pitch_idx[tmp] = idx; + track->cb_dirty = true; break; case CB_COLOR0_SLICE: case CB_COLOR1_SLICE: @@ -708,6 +1572,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) tmp = (reg - CB_COLOR0_SLICE) / 0x3c; track->cb_color_slice[tmp] = radeon_get_ib_value(p, idx); track->cb_color_slice_idx[tmp] = idx; + track->cb_dirty = true; break; case CB_COLOR8_SLICE: case CB_COLOR9_SLICE: @@ -716,6 +1581,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) tmp = ((reg - CB_COLOR8_SLICE) / 0x1c) + 8; track->cb_color_slice[tmp] = radeon_get_ib_value(p, idx); track->cb_color_slice_idx[tmp] = idx; + track->cb_dirty = true; break; case CB_COLOR0_ATTRIB: case CB_COLOR1_ATTRIB: @@ -725,6 +1591,30 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) case CB_COLOR5_ATTRIB: case CB_COLOR6_ATTRIB: case CB_COLOR7_ATTRIB: + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "bad SET_CONTEXT_REG " + "0x%04X\n", reg); + return -EINVAL; + } + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { + if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { + unsigned bankw, bankh, mtaspect, tile_split; + + evergreen_tiling_fields(reloc->lobj.tiling_flags, + &bankw, &bankh, &mtaspect, + &tile_split); + ib[idx] |= CB_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks)); + ib[idx] |= CB_TILE_SPLIT(tile_split) | + CB_BANK_WIDTH(bankw) | + CB_BANK_HEIGHT(bankh) | + CB_MACRO_TILE_ASPECT(mtaspect); + } + } + tmp = ((reg - CB_COLOR0_ATTRIB) / 0x3c); + track->cb_color_attrib[tmp] = ib[idx]; + track->cb_dirty = true; + break; case CB_COLOR8_ATTRIB: case CB_COLOR9_ATTRIB: case CB_COLOR10_ATTRIB: @@ -735,30 +1625,23 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) "0x%04X\n", reg); return -EINVAL; } - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { - ib[idx] |= CB_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks)); - ib[idx] |= CB_TILE_SPLIT(evergreen_cs_get_tile_split(track->row_size)); + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { + if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { + unsigned bankw, bankh, mtaspect, tile_split; + + evergreen_tiling_fields(reloc->lobj.tiling_flags, + &bankw, &bankh, &mtaspect, + &tile_split); + ib[idx] |= CB_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks)); + ib[idx] |= CB_TILE_SPLIT(tile_split) | + CB_BANK_WIDTH(bankw) | + CB_BANK_HEIGHT(bankh) | + CB_MACRO_TILE_ASPECT(mtaspect); + } } - break; - case CB_COLOR0_DIM: - case CB_COLOR1_DIM: - case CB_COLOR2_DIM: - case CB_COLOR3_DIM: - case CB_COLOR4_DIM: - case CB_COLOR5_DIM: - case CB_COLOR6_DIM: - case CB_COLOR7_DIM: - tmp = (reg - CB_COLOR0_DIM) / 0x3c; - track->cb_color_dim[tmp] = radeon_get_ib_value(p, idx); - track->cb_color_dim_idx[tmp] = idx; - break; - case CB_COLOR8_DIM: - case CB_COLOR9_DIM: - case CB_COLOR10_DIM: - case CB_COLOR11_DIM: - tmp = ((reg - CB_COLOR8_DIM) / 0x1c) + 8; - track->cb_color_dim[tmp] = radeon_get_ib_value(p, idx); - track->cb_color_dim_idx[tmp] = idx; + tmp = ((reg - CB_COLOR8_ATTRIB) / 0x1c) + 8; + track->cb_color_attrib[tmp] = ib[idx]; + track->cb_dirty = true; break; case CB_COLOR0_FMASK: case CB_COLOR1_FMASK: @@ -833,8 +1716,8 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) tmp = (reg - CB_COLOR0_BASE) / 0x3c; track->cb_color_bo_offset[tmp] = radeon_get_ib_value(p, idx); ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); - track->cb_color_base_last[tmp] = ib[idx]; track->cb_color_bo[tmp] = reloc->robj; + track->cb_dirty = true; break; case CB_COLOR8_BASE: case CB_COLOR9_BASE: @@ -849,8 +1732,25 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) tmp = ((reg - CB_COLOR8_BASE) / 0x1c) + 8; track->cb_color_bo_offset[tmp] = radeon_get_ib_value(p, idx); ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); - track->cb_color_base_last[tmp] = ib[idx]; track->cb_color_bo[tmp] = reloc->robj; + track->cb_dirty = true; + break; + case DB_HTILE_DATA_BASE: + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "bad SET_CONTEXT_REG " + "0x%04X\n", reg); + return -EINVAL; + } + track->htile_offset = radeon_get_ib_value(p, idx); + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + track->htile_bo = reloc->robj; + track->db_dirty = true; + break; + case DB_HTILE_SURFACE: + /* 8x8 only */ + track->htile_surface = radeon_get_ib_value(p, idx); + track->db_dirty = true; break; case CB_IMMED0_BASE: case CB_IMMED1_BASE: @@ -864,7 +1764,6 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) case CB_IMMED9_BASE: case CB_IMMED10_BASE: case CB_IMMED11_BASE: - case DB_HTILE_DATA_BASE: case SQ_PGM_START_FS: case SQ_PGM_START_ES: case SQ_PGM_START_VS: @@ -989,6 +1888,9 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); break; + case SX_MISC: + track->sx_misc_kill_all_prims = (radeon_get_ib_value(p, idx) & 0x1) != 0; + break; default: dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx); return -EINVAL; @@ -996,22 +1898,30 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return 0; } -/** - * evergreen_check_texture_resource() - check if register is authorized or not - * @p: parser structure holding parsing context - * @idx: index into the cs buffer - * @texture: texture's bo structure - * @mipmap: mipmap's bo structure - * - * This function will check that the resource has valid field and that - * the texture and mipmap bo object are big enough to cover this resource. - */ -static int evergreen_check_texture_resource(struct radeon_cs_parser *p, u32 idx, - struct radeon_bo *texture, - struct radeon_bo *mipmap) +static bool evergreen_is_safe_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) { - /* XXX fill in */ - return 0; + u32 last_reg, m, i; + + if (p->rdev->family >= CHIP_CAYMAN) + last_reg = ARRAY_SIZE(cayman_reg_safe_bm); + else + last_reg = ARRAY_SIZE(evergreen_reg_safe_bm); + + i = (reg >> 7); + if (i >= last_reg) { + dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx); + return false; + } + m = 1 << ((reg >> 2) & 31); + if (p->rdev->family >= CHIP_CAYMAN) { + if (!(cayman_reg_safe_bm[i] & m)) + return true; + } else { + if (!(evergreen_reg_safe_bm[i] & m)) + return true; + } + dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx); + return false; } static int evergreen_packet3_check(struct radeon_cs_parser *p, @@ -1036,6 +1946,8 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, { int pred_op; int tmp; + uint64_t offset; + if (pkt->count != 1) { DRM_ERROR("bad SET PREDICATION\n"); return -EINVAL; @@ -1059,8 +1971,12 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - ib[idx + 0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff); - ib[idx + 1] = tmp + (upper_32_bits(reloc->lobj.gpu_offset) & 0xff); + offset = reloc->lobj.gpu_offset + + (idx_value & 0xfffffff0) + + ((u64)(tmp & 0xff) << 32); + + ib[idx + 0] = offset; + ib[idx + 1] = (tmp & 0xffffff00) | (upper_32_bits(offset) & 0xff); } break; case PACKET3_CONTEXT_CONTROL: @@ -1088,6 +2004,9 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, } break; case PACKET3_INDEX_BASE: + { + uint64_t offset; + if (pkt->count != 1) { DRM_ERROR("bad INDEX_BASE\n"); return -EINVAL; @@ -1097,15 +2016,24 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad INDEX_BASE\n"); return -EINVAL; } - ib[idx+0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff); - ib[idx+1] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + + offset = reloc->lobj.gpu_offset + + idx_value + + ((u64)(radeon_get_ib_value(p, idx+1) & 0xff) << 32); + + ib[idx+0] = offset; + ib[idx+1] = upper_32_bits(offset) & 0xff; + r = evergreen_cs_track_check(p); if (r) { dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__); return r; } break; + } case PACKET3_DRAW_INDEX: + { + uint64_t offset; if (pkt->count != 3) { DRM_ERROR("bad DRAW_INDEX\n"); return -EINVAL; @@ -1115,15 +2043,25 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad DRAW_INDEX\n"); return -EINVAL; } - ib[idx+0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff); - ib[idx+1] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + + offset = reloc->lobj.gpu_offset + + idx_value + + ((u64)(radeon_get_ib_value(p, idx+1) & 0xff) << 32); + + ib[idx+0] = offset; + ib[idx+1] = upper_32_bits(offset) & 0xff; + r = evergreen_cs_track_check(p); if (r) { dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__); return r; } break; + } case PACKET3_DRAW_INDEX_2: + { + uint64_t offset; + if (pkt->count != 4) { DRM_ERROR("bad DRAW_INDEX_2\n"); return -EINVAL; @@ -1133,14 +2071,21 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad DRAW_INDEX_2\n"); return -EINVAL; } - ib[idx+1] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff); - ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + + offset = reloc->lobj.gpu_offset + + radeon_get_ib_value(p, idx+1) + + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); + + ib[idx+1] = offset; + ib[idx+2] = upper_32_bits(offset) & 0xff; + r = evergreen_cs_track_check(p); if (r) { dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__); return r; } break; + } case PACKET3_DRAW_INDEX_AUTO: if (pkt->count != 1) { DRM_ERROR("bad DRAW_INDEX_AUTO\n"); @@ -1231,13 +2176,20 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, } /* bit 4 is reg (0) or mem (1) */ if (idx_value & 0x10) { + uint64_t offset; + r = evergreen_cs_packet_next_reloc(p, &reloc); if (r) { DRM_ERROR("bad WAIT_REG_MEM\n"); return -EINVAL; } - ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff); - ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + + offset = reloc->lobj.gpu_offset + + (radeon_get_ib_value(p, idx+1) & 0xfffffffc) + + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); + + ib[idx+1] = (ib[idx+1] & 0x3) | (offset & 0xfffffffc); + ib[idx+2] = upper_32_bits(offset) & 0xff; } break; case PACKET3_SURFACE_SYNC: @@ -1262,16 +2214,25 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } if (pkt->count) { + uint64_t offset; + r = evergreen_cs_packet_next_reloc(p, &reloc); if (r) { DRM_ERROR("bad EVENT_WRITE\n"); return -EINVAL; } - ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff); - ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + offset = reloc->lobj.gpu_offset + + (radeon_get_ib_value(p, idx+1) & 0xfffffff8) + + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); + + ib[idx+1] = offset & 0xfffffff8; + ib[idx+2] = upper_32_bits(offset) & 0xff; } break; case PACKET3_EVENT_WRITE_EOP: + { + uint64_t offset; + if (pkt->count != 4) { DRM_ERROR("bad EVENT_WRITE_EOP\n"); return -EINVAL; @@ -1281,10 +2242,19 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad EVENT_WRITE_EOP\n"); return -EINVAL; } - ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff); - ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + + offset = reloc->lobj.gpu_offset + + (radeon_get_ib_value(p, idx+1) & 0xfffffffc) + + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); + + ib[idx+1] = offset & 0xfffffffc; + ib[idx+2] = (ib[idx+2] & 0xffffff00) | (upper_32_bits(offset) & 0xff); break; + } case PACKET3_EVENT_WRITE_EOS: + { + uint64_t offset; + if (pkt->count != 3) { DRM_ERROR("bad EVENT_WRITE_EOS\n"); return -EINVAL; @@ -1294,9 +2264,15 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad EVENT_WRITE_EOS\n"); return -EINVAL; } - ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff); - ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + + offset = reloc->lobj.gpu_offset + + (radeon_get_ib_value(p, idx+1) & 0xfffffffc) + + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); + + ib[idx+1] = offset & 0xfffffffc; + ib[idx+2] = (ib[idx+2] & 0xffffff00) | (upper_32_bits(offset) & 0xff); break; + } case PACKET3_SET_CONFIG_REG: start_reg = (idx_value << 2) + PACKET3_SET_CONFIG_REG_START; end_reg = 4 * pkt->count + start_reg - 4; @@ -1344,6 +2320,7 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, } for (i = 0; i < (pkt->count / 8); i++) { struct radeon_bo *texture, *mipmap; + u32 toffset, moffset; u32 size, offset; switch (G__SQ_CONSTANT_TYPE(radeon_get_ib_value(p, idx+1+(i*8)+7))) { @@ -1354,32 +2331,42 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad SET_RESOURCE (tex)\n"); return -EINVAL; } - ib[idx+1+(i*8)+2] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); - if (!p->keep_tiling_flags) { + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { ib[idx+1+(i*8)+1] |= TEX_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags)); if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) { - ib[idx+1+(i*8)+6] |= - TEX_TILE_SPLIT(evergreen_cs_get_tile_split(track->row_size)); + unsigned bankw, bankh, mtaspect, tile_split; + + evergreen_tiling_fields(reloc->lobj.tiling_flags, + &bankw, &bankh, &mtaspect, + &tile_split); + ib[idx+1+(i*8)+6] |= TEX_TILE_SPLIT(tile_split); ib[idx+1+(i*8)+7] |= + TEX_BANK_WIDTH(bankw) | + TEX_BANK_HEIGHT(bankh) | + MACRO_TILE_ASPECT(mtaspect) | TEX_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks)); } } texture = reloc->robj; + toffset = (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); /* tex mip base */ r = evergreen_cs_packet_next_reloc(p, &reloc); if (r) { DRM_ERROR("bad SET_RESOURCE (tex)\n"); return -EINVAL; } - ib[idx+1+(i*8)+3] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + moffset = (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); mipmap = reloc->robj; - r = evergreen_check_texture_resource(p, idx+1+(i*8), - texture, mipmap); + r = evergreen_cs_track_validate_texture(p, texture, mipmap, idx+1+(i*8)); if (r) return r; + ib[idx+1+(i*8)+2] += toffset; + ib[idx+1+(i*8)+3] += moffset; break; case SQ_TEX_VTX_VALID_BUFFER: + { + uint64_t offset64; /* vtx base */ r = evergreen_cs_packet_next_reloc(p, &reloc); if (r) { @@ -1391,11 +2378,15 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, if (p->rdev && (size + offset) > radeon_bo_size(reloc->robj)) { /* force size to size of the buffer */ dev_warn(p->dev, "vbo resource seems too big for the bo\n"); - ib[idx+1+(i*8)+1] = radeon_bo_size(reloc->robj); + ib[idx+1+(i*8)+1] = radeon_bo_size(reloc->robj) - offset; } - ib[idx+1+(i*8)+0] += (u32)((reloc->lobj.gpu_offset) & 0xffffffff); - ib[idx+1+(i*8)+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + + offset64 = reloc->lobj.gpu_offset + offset; + ib[idx+1+(i*8)+0] = offset64; + ib[idx+1+(i*8)+2] = (ib[idx+1+(i*8)+2] & 0xffffff00) | + (upper_32_bits(offset64) & 0xff); break; + } case SQ_TEX_VTX_INVALID_TEXTURE: case SQ_TEX_VTX_INVALID_BUFFER: default: @@ -1451,6 +2442,104 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } break; + case PACKET3_STRMOUT_BUFFER_UPDATE: + if (pkt->count != 4) { + DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (invalid count)\n"); + return -EINVAL; + } + /* Updating memory at DST_ADDRESS. */ + if (idx_value & 0x1) { + u64 offset; + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (missing dst reloc)\n"); + return -EINVAL; + } + offset = radeon_get_ib_value(p, idx+1); + offset += ((u64)(radeon_get_ib_value(p, idx+2) & 0xff)) << 32; + if ((offset + 4) > radeon_bo_size(reloc->robj)) { + DRM_ERROR("bad STRMOUT_BUFFER_UPDATE dst bo too small: 0x%llx, 0x%lx\n", + offset + 4, radeon_bo_size(reloc->robj)); + return -EINVAL; + } + offset += reloc->lobj.gpu_offset; + ib[idx+1] = offset; + ib[idx+2] = upper_32_bits(offset) & 0xff; + } + /* Reading data from SRC_ADDRESS. */ + if (((idx_value >> 1) & 0x3) == 2) { + u64 offset; + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (missing src reloc)\n"); + return -EINVAL; + } + offset = radeon_get_ib_value(p, idx+3); + offset += ((u64)(radeon_get_ib_value(p, idx+4) & 0xff)) << 32; + if ((offset + 4) > radeon_bo_size(reloc->robj)) { + DRM_ERROR("bad STRMOUT_BUFFER_UPDATE src bo too small: 0x%llx, 0x%lx\n", + offset + 4, radeon_bo_size(reloc->robj)); + return -EINVAL; + } + offset += reloc->lobj.gpu_offset; + ib[idx+3] = offset; + ib[idx+4] = upper_32_bits(offset) & 0xff; + } + break; + case PACKET3_COPY_DW: + if (pkt->count != 4) { + DRM_ERROR("bad COPY_DW (invalid count)\n"); + return -EINVAL; + } + if (idx_value & 0x1) { + u64 offset; + /* SRC is memory. */ + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("bad COPY_DW (missing src reloc)\n"); + return -EINVAL; + } + offset = radeon_get_ib_value(p, idx+1); + offset += ((u64)(radeon_get_ib_value(p, idx+2) & 0xff)) << 32; + if ((offset + 4) > radeon_bo_size(reloc->robj)) { + DRM_ERROR("bad COPY_DW src bo too small: 0x%llx, 0x%lx\n", + offset + 4, radeon_bo_size(reloc->robj)); + return -EINVAL; + } + offset += reloc->lobj.gpu_offset; + ib[idx+1] = offset; + ib[idx+2] = upper_32_bits(offset) & 0xff; + } else { + /* SRC is a reg. */ + reg = radeon_get_ib_value(p, idx+1) << 2; + if (!evergreen_is_safe_reg(p, reg, idx+1)) + return -EINVAL; + } + if (idx_value & 0x2) { + u64 offset; + /* DST is memory. */ + r = evergreen_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("bad COPY_DW (missing dst reloc)\n"); + return -EINVAL; + } + offset = radeon_get_ib_value(p, idx+3); + offset += ((u64)(radeon_get_ib_value(p, idx+4) & 0xff)) << 32; + if ((offset + 4) > radeon_bo_size(reloc->robj)) { + DRM_ERROR("bad COPY_DW dst bo too small: 0x%llx, 0x%lx\n", + offset + 4, radeon_bo_size(reloc->robj)); + return -EINVAL; + } + offset += reloc->lobj.gpu_offset; + ib[idx+3] = offset; + ib[idx+4] = upper_32_bits(offset) & 0xff; + } else { + /* DST is a reg. */ + reg = radeon_get_ib_value(p, idx+3) << 2; + if (!evergreen_is_safe_reg(p, reg, idx+3)) + return -EINVAL; + } + break; case PACKET3_NOP: break; default: @@ -1572,3 +2661,247 @@ int evergreen_cs_parse(struct radeon_cs_parser *p) return 0; } +/* vm parser */ +static bool evergreen_vm_reg_valid(u32 reg) +{ + /* context regs are fine */ + if (reg >= 0x28000) + return true; + + /* check config regs */ + switch (reg) { + case WAIT_UNTIL: + case GRBM_GFX_INDEX: + case CP_STRMOUT_CNTL: + case CP_COHER_CNTL: + case CP_COHER_SIZE: + case VGT_VTX_VECT_EJECT_REG: + case VGT_CACHE_INVALIDATION: + case VGT_GS_VERTEX_REUSE: + case VGT_PRIMITIVE_TYPE: + case VGT_INDEX_TYPE: + case VGT_NUM_INDICES: + case VGT_NUM_INSTANCES: + case VGT_COMPUTE_DIM_X: + case VGT_COMPUTE_DIM_Y: + case VGT_COMPUTE_DIM_Z: + case VGT_COMPUTE_START_X: + case VGT_COMPUTE_START_Y: + case VGT_COMPUTE_START_Z: + case VGT_COMPUTE_INDEX: + case VGT_COMPUTE_THREAD_GROUP_SIZE: + case VGT_HS_OFFCHIP_PARAM: + case PA_CL_ENHANCE: + case PA_SU_LINE_STIPPLE_VALUE: + case PA_SC_LINE_STIPPLE_STATE: + case PA_SC_ENHANCE: + case SQ_DYN_GPR_CNTL_PS_FLUSH_REQ: + case SQ_DYN_GPR_SIMD_LOCK_EN: + case SQ_CONFIG: + case SQ_GPR_RESOURCE_MGMT_1: + case SQ_GLOBAL_GPR_RESOURCE_MGMT_1: + case SQ_GLOBAL_GPR_RESOURCE_MGMT_2: + case SQ_CONST_MEM_BASE: + case SQ_STATIC_THREAD_MGMT_1: + case SQ_STATIC_THREAD_MGMT_2: + case SQ_STATIC_THREAD_MGMT_3: + case SPI_CONFIG_CNTL: + case SPI_CONFIG_CNTL_1: + case TA_CNTL_AUX: + case DB_DEBUG: + case DB_DEBUG2: + case DB_DEBUG3: + case DB_DEBUG4: + case DB_WATERMARKS: + case TD_PS_BORDER_COLOR_INDEX: + case TD_PS_BORDER_COLOR_RED: + case TD_PS_BORDER_COLOR_GREEN: + case TD_PS_BORDER_COLOR_BLUE: + case TD_PS_BORDER_COLOR_ALPHA: + case TD_VS_BORDER_COLOR_INDEX: + case TD_VS_BORDER_COLOR_RED: + case TD_VS_BORDER_COLOR_GREEN: + case TD_VS_BORDER_COLOR_BLUE: + case TD_VS_BORDER_COLOR_ALPHA: + case TD_GS_BORDER_COLOR_INDEX: + case TD_GS_BORDER_COLOR_RED: + case TD_GS_BORDER_COLOR_GREEN: + case TD_GS_BORDER_COLOR_BLUE: + case TD_GS_BORDER_COLOR_ALPHA: + case TD_HS_BORDER_COLOR_INDEX: + case TD_HS_BORDER_COLOR_RED: + case TD_HS_BORDER_COLOR_GREEN: + case TD_HS_BORDER_COLOR_BLUE: + case TD_HS_BORDER_COLOR_ALPHA: + case TD_LS_BORDER_COLOR_INDEX: + case TD_LS_BORDER_COLOR_RED: + case TD_LS_BORDER_COLOR_GREEN: + case TD_LS_BORDER_COLOR_BLUE: + case TD_LS_BORDER_COLOR_ALPHA: + case TD_CS_BORDER_COLOR_INDEX: + case TD_CS_BORDER_COLOR_RED: + case TD_CS_BORDER_COLOR_GREEN: + case TD_CS_BORDER_COLOR_BLUE: + case TD_CS_BORDER_COLOR_ALPHA: + case SQ_ESGS_RING_SIZE: + case SQ_GSVS_RING_SIZE: + case SQ_ESTMP_RING_SIZE: + case SQ_GSTMP_RING_SIZE: + case SQ_HSTMP_RING_SIZE: + case SQ_LSTMP_RING_SIZE: + case SQ_PSTMP_RING_SIZE: + case SQ_VSTMP_RING_SIZE: + case SQ_ESGS_RING_ITEMSIZE: + case SQ_ESTMP_RING_ITEMSIZE: + case SQ_GSTMP_RING_ITEMSIZE: + case SQ_GSVS_RING_ITEMSIZE: + case SQ_GS_VERT_ITEMSIZE: + case SQ_GS_VERT_ITEMSIZE_1: + case SQ_GS_VERT_ITEMSIZE_2: + case SQ_GS_VERT_ITEMSIZE_3: + case SQ_GSVS_RING_OFFSET_1: + case SQ_GSVS_RING_OFFSET_2: + case SQ_GSVS_RING_OFFSET_3: + case SQ_HSTMP_RING_ITEMSIZE: + case SQ_LSTMP_RING_ITEMSIZE: + case SQ_PSTMP_RING_ITEMSIZE: + case SQ_VSTMP_RING_ITEMSIZE: + case VGT_TF_RING_SIZE: + case SQ_ESGS_RING_BASE: + case SQ_GSVS_RING_BASE: + case SQ_ESTMP_RING_BASE: + case SQ_GSTMP_RING_BASE: + case SQ_HSTMP_RING_BASE: + case SQ_LSTMP_RING_BASE: + case SQ_PSTMP_RING_BASE: + case SQ_VSTMP_RING_BASE: + case CAYMAN_VGT_OFFCHIP_LDS_BASE: + case CAYMAN_SQ_EX_ALLOC_TABLE_SLOTS: + return true; + default: + DRM_ERROR("Invalid register 0x%x in CS\n", reg); + return false; + } +} + +static int evergreen_vm_packet3_check(struct radeon_device *rdev, + u32 *ib, struct radeon_cs_packet *pkt) +{ + u32 idx = pkt->idx + 1; + u32 idx_value = ib[idx]; + u32 start_reg, end_reg, reg, i; + + switch (pkt->opcode) { + case PACKET3_NOP: + case PACKET3_SET_BASE: + case PACKET3_CLEAR_STATE: + case PACKET3_INDEX_BUFFER_SIZE: + case PACKET3_DISPATCH_DIRECT: + case PACKET3_DISPATCH_INDIRECT: + case PACKET3_MODE_CONTROL: + case PACKET3_SET_PREDICATION: + case PACKET3_COND_EXEC: + case PACKET3_PRED_EXEC: + case PACKET3_DRAW_INDIRECT: + case PACKET3_DRAW_INDEX_INDIRECT: + case PACKET3_INDEX_BASE: + case PACKET3_DRAW_INDEX_2: + case PACKET3_CONTEXT_CONTROL: + case PACKET3_DRAW_INDEX_OFFSET: + case PACKET3_INDEX_TYPE: + case PACKET3_DRAW_INDEX: + case PACKET3_DRAW_INDEX_AUTO: + case PACKET3_DRAW_INDEX_IMMD: + case PACKET3_NUM_INSTANCES: + case PACKET3_DRAW_INDEX_MULTI_AUTO: + case PACKET3_STRMOUT_BUFFER_UPDATE: + case PACKET3_DRAW_INDEX_OFFSET_2: + case PACKET3_DRAW_INDEX_MULTI_ELEMENT: + case PACKET3_MPEG_INDEX: + case PACKET3_WAIT_REG_MEM: + case PACKET3_MEM_WRITE: + case PACKET3_SURFACE_SYNC: + case PACKET3_EVENT_WRITE: + case PACKET3_EVENT_WRITE_EOP: + case PACKET3_EVENT_WRITE_EOS: + case PACKET3_SET_CONTEXT_REG: + case PACKET3_SET_BOOL_CONST: + case PACKET3_SET_LOOP_CONST: + case PACKET3_SET_RESOURCE: + case PACKET3_SET_SAMPLER: + case PACKET3_SET_CTL_CONST: + case PACKET3_SET_RESOURCE_OFFSET: + case PACKET3_SET_CONTEXT_REG_INDIRECT: + case PACKET3_SET_RESOURCE_INDIRECT: + case CAYMAN_PACKET3_DEALLOC_STATE: + break; + case PACKET3_COND_WRITE: + if (idx_value & 0x100) { + reg = ib[idx + 5] * 4; + if (!evergreen_vm_reg_valid(reg)) + return -EINVAL; + } + break; + case PACKET3_COPY_DW: + if (idx_value & 0x2) { + reg = ib[idx + 3] * 4; + if (!evergreen_vm_reg_valid(reg)) + return -EINVAL; + } + break; + case PACKET3_SET_CONFIG_REG: + start_reg = (idx_value << 2) + PACKET3_SET_CONFIG_REG_START; + end_reg = 4 * pkt->count + start_reg - 4; + if ((start_reg < PACKET3_SET_CONFIG_REG_START) || + (start_reg >= PACKET3_SET_CONFIG_REG_END) || + (end_reg >= PACKET3_SET_CONFIG_REG_END)) { + DRM_ERROR("bad PACKET3_SET_CONFIG_REG\n"); + return -EINVAL; + } + for (i = 0; i < pkt->count; i++) { + reg = start_reg + (4 * i); + if (!evergreen_vm_reg_valid(reg)) + return -EINVAL; + } + break; + default: + return -EINVAL; + } + return 0; +} + +int evergreen_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib) +{ + int ret = 0; + u32 idx = 0; + struct radeon_cs_packet pkt; + + do { + pkt.idx = idx; + pkt.type = CP_PACKET_GET_TYPE(ib->ptr[idx]); + pkt.count = CP_PACKET_GET_COUNT(ib->ptr[idx]); + pkt.one_reg_wr = 0; + switch (pkt.type) { + case PACKET_TYPE0: + dev_err(rdev->dev, "Packet0 not allowed!\n"); + ret = -EINVAL; + break; + case PACKET_TYPE2: + idx += 1; + break; + case PACKET_TYPE3: + pkt.opcode = CP_PACKET3_GET_OPCODE(ib->ptr[idx]); + ret = evergreen_vm_packet3_check(rdev, ib->ptr, &pkt); + idx += pkt.count + 2; + break; + default: + dev_err(rdev->dev, "Unknown packet type %d !\n", pkt.type); + ret = -EINVAL; + break; + } + if (ret) + break; + } while (idx < ib->length_dw); + + return ret; +} diff --git a/drivers/gpu/drm/radeon/evergreen_reg.h b/drivers/gpu/drm/radeon/evergreen_reg.h index e022776..34a0e85 100644 --- a/drivers/gpu/drm/radeon/evergreen_reg.h +++ b/drivers/gpu/drm/radeon/evergreen_reg.h @@ -35,6 +35,14 @@ #define EVERGREEN_P1PLL_SS_CNTL 0x414 #define EVERGREEN_P2PLL_SS_CNTL 0x454 # define EVERGREEN_PxPLL_SS_EN (1 << 12) + +#define EVERGREEN_AUDIO_PLL1_MUL 0x5b0 +#define EVERGREEN_AUDIO_PLL1_DIV 0x5b4 +#define EVERGREEN_AUDIO_PLL1_UNK 0x5bc + +#define EVERGREEN_AUDIO_ENABLE 0x5e78 +#define EVERGREEN_AUDIO_VENDOR_ID 0x5ec0 + /* GRPH blocks at 0x6800, 0x7400, 0x10000, 0x10c00, 0x11800, 0x12400 */ #define EVERGREEN_GRPH_ENABLE 0x6800 #define EVERGREEN_GRPH_CONTROL 0x6804 @@ -223,4 +231,9 @@ #define EVERGREEN_DC_GPIO_HPD_EN 0x64b8 #define EVERGREEN_DC_GPIO_HPD_Y 0x64bc +/* HDMI blocks at 0x7030, 0x7c30, 0x10830, 0x11430, 0x12030, 0x12c30 */ +#define EVERGREEN_HDMI_BASE 0x7030 + +#define EVERGREEN_HDMI_CONFIG_OFFSET 0xf0 + #endif diff --git a/drivers/gpu/drm/radeon/evergreend.h b/drivers/gpu/drm/radeon/evergreend.h index fe44a95..81e744f 100644 --- a/drivers/gpu/drm/radeon/evergreend.h +++ b/drivers/gpu/drm/radeon/evergreend.h @@ -81,6 +81,11 @@ #define FB_READ_EN (1 << 0) #define FB_WRITE_EN (1 << 1) +#define CP_STRMOUT_CNTL 0x84FC + +#define CP_COHER_CNTL 0x85F0 +#define CP_COHER_SIZE 0x85F4 +#define CP_COHER_BASE 0x85F8 #define CP_ME_CNTL 0x86D8 #define CP_ME_HALT (1 << 28) #define CP_PFP_HALT (1 << 26) @@ -112,6 +117,7 @@ #define CP_RB_WPTR_ADDR_HI 0xC11C #define CP_RB_WPTR_DELAY 0x8704 #define CP_SEM_WAIT_TIMER 0x85BC +#define CP_SEM_INCOMPLETE_TIMER_CNTL 0x85C8 #define CP_DEBUG 0xC1FC @@ -250,6 +256,7 @@ #define PA_CL_ENHANCE 0x8A14 #define CLIP_VTX_REORDER_ENA (1 << 0) #define NUM_CLIP_SEQ(x) ((x) << 1) +#define PA_SC_ENHANCE 0x8BF0 #define PA_SC_AA_CONFIG 0x28C04 #define MSAA_NUM_SAMPLES_SHIFT 0 #define MSAA_NUM_SAMPLES_MASK 0x3 @@ -328,6 +335,8 @@ #define SQ_GPR_RESOURCE_MGMT_3 0x8C0C #define NUM_HS_GPRS(x) ((x) << 0) #define NUM_LS_GPRS(x) ((x) << 16) +#define SQ_GLOBAL_GPR_RESOURCE_MGMT_1 0x8C10 +#define SQ_GLOBAL_GPR_RESOURCE_MGMT_2 0x8C14 #define SQ_THREAD_RESOURCE_MGMT 0x8C18 #define NUM_PS_THREADS(x) ((x) << 0) #define NUM_VS_THREADS(x) ((x) << 8) @@ -346,6 +355,10 @@ #define NUM_HS_STACK_ENTRIES(x) ((x) << 0) #define NUM_LS_STACK_ENTRIES(x) ((x) << 16) #define SQ_DYN_GPR_CNTL_PS_FLUSH_REQ 0x8D8C +#define SQ_DYN_GPR_SIMD_LOCK_EN 0x8D94 +#define SQ_STATIC_THREAD_MGMT_1 0x8E20 +#define SQ_STATIC_THREAD_MGMT_2 0x8E24 +#define SQ_STATIC_THREAD_MGMT_3 0x8E28 #define SQ_LDS_RESOURCE_MGMT 0x8E2C #define SQ_MS_FIFO_SIZES 0x8CF0 @@ -700,6 +713,7 @@ #define PACKET3_DRAW_INDEX_MULTI_ELEMENT 0x36 #define PACKET3_MEM_SEMAPHORE 0x39 #define PACKET3_MPEG_INDEX 0x3A +#define PACKET3_COPY_DW 0x3B #define PACKET3_WAIT_REG_MEM 0x3C #define PACKET3_MEM_WRITE 0x3D #define PACKET3_INDIRECT_BUFFER 0x32 @@ -777,6 +791,8 @@ #define SQ_TEX_VTX_VALID_TEXTURE 0x2 #define SQ_TEX_VTX_VALID_BUFFER 0x3 +#define VGT_VTX_VECT_EJECT_REG 0x88b0 + #define SQ_CONST_MEM_BASE 0x8df8 #define SQ_ESGS_RING_BASE 0x8c40 @@ -901,23 +917,160 @@ #define PA_SC_SCREEN_SCISSOR_TL 0x28030 #define PA_SC_GENERIC_SCISSOR_TL 0x28240 #define PA_SC_WINDOW_SCISSOR_TL 0x28204 -#define VGT_PRIMITIVE_TYPE 0x8958 +#define VGT_PRIMITIVE_TYPE 0x8958 +#define VGT_INDEX_TYPE 0x895C + +#define VGT_NUM_INDICES 0x8970 + +#define VGT_COMPUTE_DIM_X 0x8990 +#define VGT_COMPUTE_DIM_Y 0x8994 +#define VGT_COMPUTE_DIM_Z 0x8998 +#define VGT_COMPUTE_START_X 0x899C +#define VGT_COMPUTE_START_Y 0x89A0 +#define VGT_COMPUTE_START_Z 0x89A4 +#define VGT_COMPUTE_INDEX 0x89A8 +#define VGT_COMPUTE_THREAD_GROUP_SIZE 0x89AC +#define VGT_HS_OFFCHIP_PARAM 0x89B0 + +#define DB_DEBUG 0x9830 +#define DB_DEBUG2 0x9834 +#define DB_DEBUG3 0x9838 +#define DB_DEBUG4 0x983C +#define DB_WATERMARKS 0x9854 #define DB_DEPTH_CONTROL 0x28800 +#define R_028800_DB_DEPTH_CONTROL 0x028800 +#define S_028800_STENCIL_ENABLE(x) (((x) & 0x1) << 0) +#define G_028800_STENCIL_ENABLE(x) (((x) >> 0) & 0x1) +#define C_028800_STENCIL_ENABLE 0xFFFFFFFE +#define S_028800_Z_ENABLE(x) (((x) & 0x1) << 1) +#define G_028800_Z_ENABLE(x) (((x) >> 1) & 0x1) +#define C_028800_Z_ENABLE 0xFFFFFFFD +#define S_028800_Z_WRITE_ENABLE(x) (((x) & 0x1) << 2) +#define G_028800_Z_WRITE_ENABLE(x) (((x) >> 2) & 0x1) +#define C_028800_Z_WRITE_ENABLE 0xFFFFFFFB +#define S_028800_ZFUNC(x) (((x) & 0x7) << 4) +#define G_028800_ZFUNC(x) (((x) >> 4) & 0x7) +#define C_028800_ZFUNC 0xFFFFFF8F +#define S_028800_BACKFACE_ENABLE(x) (((x) & 0x1) << 7) +#define G_028800_BACKFACE_ENABLE(x) (((x) >> 7) & 0x1) +#define C_028800_BACKFACE_ENABLE 0xFFFFFF7F +#define S_028800_STENCILFUNC(x) (((x) & 0x7) << 8) +#define G_028800_STENCILFUNC(x) (((x) >> 8) & 0x7) +#define C_028800_STENCILFUNC 0xFFFFF8FF +#define V_028800_STENCILFUNC_NEVER 0x00000000 +#define V_028800_STENCILFUNC_LESS 0x00000001 +#define V_028800_STENCILFUNC_EQUAL 0x00000002 +#define V_028800_STENCILFUNC_LEQUAL 0x00000003 +#define V_028800_STENCILFUNC_GREATER 0x00000004 +#define V_028800_STENCILFUNC_NOTEQUAL 0x00000005 +#define V_028800_STENCILFUNC_GEQUAL 0x00000006 +#define V_028800_STENCILFUNC_ALWAYS 0x00000007 +#define S_028800_STENCILFAIL(x) (((x) & 0x7) << 11) +#define G_028800_STENCILFAIL(x) (((x) >> 11) & 0x7) +#define C_028800_STENCILFAIL 0xFFFFC7FF +#define V_028800_STENCIL_KEEP 0x00000000 +#define V_028800_STENCIL_ZERO 0x00000001 +#define V_028800_STENCIL_REPLACE 0x00000002 +#define V_028800_STENCIL_INCR 0x00000003 +#define V_028800_STENCIL_DECR 0x00000004 +#define V_028800_STENCIL_INVERT 0x00000005 +#define V_028800_STENCIL_INCR_WRAP 0x00000006 +#define V_028800_STENCIL_DECR_WRAP 0x00000007 +#define S_028800_STENCILZPASS(x) (((x) & 0x7) << 14) +#define G_028800_STENCILZPASS(x) (((x) >> 14) & 0x7) +#define C_028800_STENCILZPASS 0xFFFE3FFF +#define S_028800_STENCILZFAIL(x) (((x) & 0x7) << 17) +#define G_028800_STENCILZFAIL(x) (((x) >> 17) & 0x7) +#define C_028800_STENCILZFAIL 0xFFF1FFFF +#define S_028800_STENCILFUNC_BF(x) (((x) & 0x7) << 20) +#define G_028800_STENCILFUNC_BF(x) (((x) >> 20) & 0x7) +#define C_028800_STENCILFUNC_BF 0xFF8FFFFF +#define S_028800_STENCILFAIL_BF(x) (((x) & 0x7) << 23) +#define G_028800_STENCILFAIL_BF(x) (((x) >> 23) & 0x7) +#define C_028800_STENCILFAIL_BF 0xFC7FFFFF +#define S_028800_STENCILZPASS_BF(x) (((x) & 0x7) << 26) +#define G_028800_STENCILZPASS_BF(x) (((x) >> 26) & 0x7) +#define C_028800_STENCILZPASS_BF 0xE3FFFFFF +#define S_028800_STENCILZFAIL_BF(x) (((x) & 0x7) << 29) +#define G_028800_STENCILZFAIL_BF(x) (((x) >> 29) & 0x7) +#define C_028800_STENCILZFAIL_BF 0x1FFFFFFF #define DB_DEPTH_VIEW 0x28008 +#define R_028008_DB_DEPTH_VIEW 0x00028008 +#define S_028008_SLICE_START(x) (((x) & 0x7FF) << 0) +#define G_028008_SLICE_START(x) (((x) >> 0) & 0x7FF) +#define C_028008_SLICE_START 0xFFFFF800 +#define S_028008_SLICE_MAX(x) (((x) & 0x7FF) << 13) +#define G_028008_SLICE_MAX(x) (((x) >> 13) & 0x7FF) +#define C_028008_SLICE_MAX 0xFF001FFF #define DB_HTILE_DATA_BASE 0x28014 +#define DB_HTILE_SURFACE 0x28abc +#define S_028ABC_HTILE_WIDTH(x) (((x) & 0x1) << 0) +#define G_028ABC_HTILE_WIDTH(x) (((x) >> 0) & 0x1) +#define C_028ABC_HTILE_WIDTH 0xFFFFFFFE +#define S_028ABC_HTILE_HEIGHT(x) (((x) & 0x1) << 1) +#define G_028ABC_HTILE_HEIGHT(x) (((x) >> 1) & 0x1) +#define C_028ABC_HTILE_HEIGHT 0xFFFFFFFD +#define G_028ABC_LINEAR(x) (((x) >> 2) & 0x1) #define DB_Z_INFO 0x28040 # define Z_ARRAY_MODE(x) ((x) << 4) # define DB_TILE_SPLIT(x) (((x) & 0x7) << 8) # define DB_NUM_BANKS(x) (((x) & 0x3) << 12) # define DB_BANK_WIDTH(x) (((x) & 0x3) << 16) # define DB_BANK_HEIGHT(x) (((x) & 0x3) << 20) +# define DB_MACRO_TILE_ASPECT(x) (((x) & 0x3) << 24) +#define R_028040_DB_Z_INFO 0x028040 +#define S_028040_FORMAT(x) (((x) & 0x3) << 0) +#define G_028040_FORMAT(x) (((x) >> 0) & 0x3) +#define C_028040_FORMAT 0xFFFFFFFC +#define V_028040_Z_INVALID 0x00000000 +#define V_028040_Z_16 0x00000001 +#define V_028040_Z_24 0x00000002 +#define V_028040_Z_32_FLOAT 0x00000003 +#define S_028040_ARRAY_MODE(x) (((x) & 0xF) << 4) +#define G_028040_ARRAY_MODE(x) (((x) >> 4) & 0xF) +#define C_028040_ARRAY_MODE 0xFFFFFF0F +#define S_028040_READ_SIZE(x) (((x) & 0x1) << 28) +#define G_028040_READ_SIZE(x) (((x) >> 28) & 0x1) +#define C_028040_READ_SIZE 0xEFFFFFFF +#define S_028040_TILE_SURFACE_ENABLE(x) (((x) & 0x1) << 29) +#define G_028040_TILE_SURFACE_ENABLE(x) (((x) >> 29) & 0x1) +#define C_028040_TILE_SURFACE_ENABLE 0xDFFFFFFF +#define S_028040_ZRANGE_PRECISION(x) (((x) & 0x1) << 31) +#define G_028040_ZRANGE_PRECISION(x) (((x) >> 31) & 0x1) +#define C_028040_ZRANGE_PRECISION 0x7FFFFFFF +#define S_028040_TILE_SPLIT(x) (((x) & 0x7) << 8) +#define G_028040_TILE_SPLIT(x) (((x) >> 8) & 0x7) +#define S_028040_NUM_BANKS(x) (((x) & 0x3) << 12) +#define G_028040_NUM_BANKS(x) (((x) >> 12) & 0x3) +#define S_028040_BANK_WIDTH(x) (((x) & 0x3) << 16) +#define G_028040_BANK_WIDTH(x) (((x) >> 16) & 0x3) +#define S_028040_BANK_HEIGHT(x) (((x) & 0x3) << 20) +#define G_028040_BANK_HEIGHT(x) (((x) >> 20) & 0x3) +#define S_028040_MACRO_TILE_ASPECT(x) (((x) & 0x3) << 24) +#define G_028040_MACRO_TILE_ASPECT(x) (((x) >> 24) & 0x3) #define DB_STENCIL_INFO 0x28044 +#define R_028044_DB_STENCIL_INFO 0x028044 +#define S_028044_FORMAT(x) (((x) & 0x1) << 0) +#define G_028044_FORMAT(x) (((x) >> 0) & 0x1) +#define C_028044_FORMAT 0xFFFFFFFE +#define G_028044_TILE_SPLIT(x) (((x) >> 8) & 0x7) #define DB_Z_READ_BASE 0x28048 #define DB_STENCIL_READ_BASE 0x2804c #define DB_Z_WRITE_BASE 0x28050 #define DB_STENCIL_WRITE_BASE 0x28054 #define DB_DEPTH_SIZE 0x28058 +#define R_028058_DB_DEPTH_SIZE 0x028058 +#define S_028058_PITCH_TILE_MAX(x) (((x) & 0x7FF) << 0) +#define G_028058_PITCH_TILE_MAX(x) (((x) >> 0) & 0x7FF) +#define C_028058_PITCH_TILE_MAX 0xFFFFF800 +#define S_028058_HEIGHT_TILE_MAX(x) (((x) & 0x7FF) << 11) +#define G_028058_HEIGHT_TILE_MAX(x) (((x) >> 11) & 0x7FF) +#define C_028058_HEIGHT_TILE_MAX 0xFFC007FF +#define R_02805C_DB_DEPTH_SLICE 0x02805C +#define S_02805C_SLICE_TILE_MAX(x) (((x) & 0x3FFFFF) << 0) +#define G_02805C_SLICE_TILE_MAX(x) (((x) >> 0) & 0x3FFFFF) +#define C_02805C_SLICE_TILE_MAX 0xFFC00000 #define SQ_PGM_START_PS 0x28840 #define SQ_PGM_START_VS 0x2885c @@ -927,6 +1080,14 @@ #define SQ_PGM_START_HS 0x288b8 #define SQ_PGM_START_LS 0x288d0 +#define VGT_STRMOUT_BUFFER_BASE_0 0x28AD8 +#define VGT_STRMOUT_BUFFER_BASE_1 0x28AE8 +#define VGT_STRMOUT_BUFFER_BASE_2 0x28AF8 +#define VGT_STRMOUT_BUFFER_BASE_3 0x28B08 +#define VGT_STRMOUT_BUFFER_SIZE_0 0x28AD0 +#define VGT_STRMOUT_BUFFER_SIZE_1 0x28AE0 +#define VGT_STRMOUT_BUFFER_SIZE_2 0x28AF0 +#define VGT_STRMOUT_BUFFER_SIZE_3 0x28B00 #define VGT_STRMOUT_CONFIG 0x28b94 #define VGT_STRMOUT_BUFFER_CONFIG 0x28b98 @@ -953,6 +1114,114 @@ #define CB_COLOR0_PITCH 0x28c64 #define CB_COLOR0_SLICE 0x28c68 #define CB_COLOR0_VIEW 0x28c6c +#define R_028C6C_CB_COLOR0_VIEW 0x00028C6C +#define S_028C6C_SLICE_START(x) (((x) & 0x7FF) << 0) +#define G_028C6C_SLICE_START(x) (((x) >> 0) & 0x7FF) +#define C_028C6C_SLICE_START 0xFFFFF800 +#define S_028C6C_SLICE_MAX(x) (((x) & 0x7FF) << 13) +#define G_028C6C_SLICE_MAX(x) (((x) >> 13) & 0x7FF) +#define C_028C6C_SLICE_MAX 0xFF001FFF +#define R_028C70_CB_COLOR0_INFO 0x028C70 +#define S_028C70_ENDIAN(x) (((x) & 0x3) << 0) +#define G_028C70_ENDIAN(x) (((x) >> 0) & 0x3) +#define C_028C70_ENDIAN 0xFFFFFFFC +#define S_028C70_FORMAT(x) (((x) & 0x3F) << 2) +#define G_028C70_FORMAT(x) (((x) >> 2) & 0x3F) +#define C_028C70_FORMAT 0xFFFFFF03 +#define V_028C70_COLOR_INVALID 0x00000000 +#define V_028C70_COLOR_8 0x00000001 +#define V_028C70_COLOR_4_4 0x00000002 +#define V_028C70_COLOR_3_3_2 0x00000003 +#define V_028C70_COLOR_16 0x00000005 +#define V_028C70_COLOR_16_FLOAT 0x00000006 +#define V_028C70_COLOR_8_8 0x00000007 +#define V_028C70_COLOR_5_6_5 0x00000008 +#define V_028C70_COLOR_6_5_5 0x00000009 +#define V_028C70_COLOR_1_5_5_5 0x0000000A +#define V_028C70_COLOR_4_4_4_4 0x0000000B +#define V_028C70_COLOR_5_5_5_1 0x0000000C +#define V_028C70_COLOR_32 0x0000000D +#define V_028C70_COLOR_32_FLOAT 0x0000000E +#define V_028C70_COLOR_16_16 0x0000000F +#define V_028C70_COLOR_16_16_FLOAT 0x00000010 +#define V_028C70_COLOR_8_24 0x00000011 +#define V_028C70_COLOR_8_24_FLOAT 0x00000012 +#define V_028C70_COLOR_24_8 0x00000013 +#define V_028C70_COLOR_24_8_FLOAT 0x00000014 +#define V_028C70_COLOR_10_11_11 0x00000015 +#define V_028C70_COLOR_10_11_11_FLOAT 0x00000016 +#define V_028C70_COLOR_11_11_10 0x00000017 +#define V_028C70_COLOR_11_11_10_FLOAT 0x00000018 +#define V_028C70_COLOR_2_10_10_10 0x00000019 +#define V_028C70_COLOR_8_8_8_8 0x0000001A +#define V_028C70_COLOR_10_10_10_2 0x0000001B +#define V_028C70_COLOR_X24_8_32_FLOAT 0x0000001C +#define V_028C70_COLOR_32_32 0x0000001D +#define V_028C70_COLOR_32_32_FLOAT 0x0000001E +#define V_028C70_COLOR_16_16_16_16 0x0000001F +#define V_028C70_COLOR_16_16_16_16_FLOAT 0x00000020 +#define V_028C70_COLOR_32_32_32_32 0x00000022 +#define V_028C70_COLOR_32_32_32_32_FLOAT 0x00000023 +#define V_028C70_COLOR_32_32_32_FLOAT 0x00000030 +#define S_028C70_ARRAY_MODE(x) (((x) & 0xF) << 8) +#define G_028C70_ARRAY_MODE(x) (((x) >> 8) & 0xF) +#define C_028C70_ARRAY_MODE 0xFFFFF0FF +#define V_028C70_ARRAY_LINEAR_GENERAL 0x00000000 +#define V_028C70_ARRAY_LINEAR_ALIGNED 0x00000001 +#define V_028C70_ARRAY_1D_TILED_THIN1 0x00000002 +#define V_028C70_ARRAY_2D_TILED_THIN1 0x00000004 +#define S_028C70_NUMBER_TYPE(x) (((x) & 0x7) << 12) +#define G_028C70_NUMBER_TYPE(x) (((x) >> 12) & 0x7) +#define C_028C70_NUMBER_TYPE 0xFFFF8FFF +#define V_028C70_NUMBER_UNORM 0x00000000 +#define V_028C70_NUMBER_SNORM 0x00000001 +#define V_028C70_NUMBER_USCALED 0x00000002 +#define V_028C70_NUMBER_SSCALED 0x00000003 +#define V_028C70_NUMBER_UINT 0x00000004 +#define V_028C70_NUMBER_SINT 0x00000005 +#define V_028C70_NUMBER_SRGB 0x00000006 +#define V_028C70_NUMBER_FLOAT 0x00000007 +#define S_028C70_COMP_SWAP(x) (((x) & 0x3) << 15) +#define G_028C70_COMP_SWAP(x) (((x) >> 15) & 0x3) +#define C_028C70_COMP_SWAP 0xFFFE7FFF +#define V_028C70_SWAP_STD 0x00000000 +#define V_028C70_SWAP_ALT 0x00000001 +#define V_028C70_SWAP_STD_REV 0x00000002 +#define V_028C70_SWAP_ALT_REV 0x00000003 +#define S_028C70_FAST_CLEAR(x) (((x) & 0x1) << 17) +#define G_028C70_FAST_CLEAR(x) (((x) >> 17) & 0x1) +#define C_028C70_FAST_CLEAR 0xFFFDFFFF +#define S_028C70_COMPRESSION(x) (((x) & 0x3) << 18) +#define G_028C70_COMPRESSION(x) (((x) >> 18) & 0x3) +#define C_028C70_COMPRESSION 0xFFF3FFFF +#define S_028C70_BLEND_CLAMP(x) (((x) & 0x1) << 19) +#define G_028C70_BLEND_CLAMP(x) (((x) >> 19) & 0x1) +#define C_028C70_BLEND_CLAMP 0xFFF7FFFF +#define S_028C70_BLEND_BYPASS(x) (((x) & 0x1) << 20) +#define G_028C70_BLEND_BYPASS(x) (((x) >> 20) & 0x1) +#define C_028C70_BLEND_BYPASS 0xFFEFFFFF +#define S_028C70_SIMPLE_FLOAT(x) (((x) & 0x1) << 21) +#define G_028C70_SIMPLE_FLOAT(x) (((x) >> 21) & 0x1) +#define C_028C70_SIMPLE_FLOAT 0xFFDFFFFF +#define S_028C70_ROUND_MODE(x) (((x) & 0x1) << 22) +#define G_028C70_ROUND_MODE(x) (((x) >> 22) & 0x1) +#define C_028C70_ROUND_MODE 0xFFBFFFFF +#define S_028C70_TILE_COMPACT(x) (((x) & 0x1) << 23) +#define G_028C70_TILE_COMPACT(x) (((x) >> 23) & 0x1) +#define C_028C70_TILE_COMPACT 0xFF7FFFFF +#define S_028C70_SOURCE_FORMAT(x) (((x) & 0x3) << 24) +#define G_028C70_SOURCE_FORMAT(x) (((x) >> 24) & 0x3) +#define C_028C70_SOURCE_FORMAT 0xFCFFFFFF +#define V_028C70_EXPORT_4C_32BPC 0x0 +#define V_028C70_EXPORT_4C_16BPC 0x1 +#define V_028C70_EXPORT_2C_32BPC 0x2 /* Do not use */ +#define S_028C70_RAT(x) (((x) & 0x1) << 26) +#define G_028C70_RAT(x) (((x) >> 26) & 0x1) +#define C_028C70_RAT 0xFBFFFFFF +#define S_028C70_RESOURCE_TYPE(x) (((x) & 0x7) << 27) +#define G_028C70_RESOURCE_TYPE(x) (((x) >> 27) & 0x7) +#define C_028C70_RESOURCE_TYPE 0xC7FFFFFF + #define CB_COLOR0_INFO 0x28c70 # define CB_FORMAT(x) ((x) << 2) # define CB_ARRAY_MODE(x) ((x) << 8) @@ -963,6 +1232,20 @@ # define CB_SOURCE_FORMAT(x) ((x) << 24) # define CB_SF_EXPORT_FULL 0 # define CB_SF_EXPORT_NORM 1 +#define R_028C74_CB_COLOR0_ATTRIB 0x028C74 +#define S_028C74_NON_DISP_TILING_ORDER(x) (((x) & 0x1) << 4) +#define G_028C74_NON_DISP_TILING_ORDER(x) (((x) >> 4) & 0x1) +#define C_028C74_NON_DISP_TILING_ORDER 0xFFFFFFEF +#define S_028C74_TILE_SPLIT(x) (((x) & 0xf) << 5) +#define G_028C74_TILE_SPLIT(x) (((x) >> 5) & 0xf) +#define S_028C74_NUM_BANKS(x) (((x) & 0x3) << 10) +#define G_028C74_NUM_BANKS(x) (((x) >> 10) & 0x3) +#define S_028C74_BANK_WIDTH(x) (((x) & 0x3) << 13) +#define G_028C74_BANK_WIDTH(x) (((x) >> 13) & 0x3) +#define S_028C74_BANK_HEIGHT(x) (((x) & 0x3) << 16) +#define G_028C74_BANK_HEIGHT(x) (((x) >> 16) & 0x3) +#define S_028C74_MACRO_TILE_ASPECT(x) (((x) & 0x3) << 19) +#define G_028C74_MACRO_TILE_ASPECT(x) (((x) >> 19) & 0x3) #define CB_COLOR0_ATTRIB 0x28c74 # define CB_TILE_SPLIT(x) (((x) & 0x7) << 5) # define ADDR_SURF_TILE_SPLIT_64B 0 @@ -987,6 +1270,7 @@ # define ADDR_SURF_BANK_HEIGHT_2 1 # define ADDR_SURF_BANK_HEIGHT_4 2 # define ADDR_SURF_BANK_HEIGHT_8 3 +# define CB_MACRO_TILE_ASPECT(x) (((x) & 0x3) << 19) #define CB_COLOR0_DIM 0x28c78 /* only CB0-7 blocks have these regs */ #define CB_COLOR0_CMASK 0x28c7c @@ -1175,9 +1459,144 @@ #define SQ_TEX_RESOURCE_WORD6_0 0x30018 # define TEX_TILE_SPLIT(x) (((x) & 0x7) << 29) #define SQ_TEX_RESOURCE_WORD7_0 0x3001c +# define MACRO_TILE_ASPECT(x) (((x) & 0x3) << 6) # define TEX_BANK_WIDTH(x) (((x) & 0x3) << 8) # define TEX_BANK_HEIGHT(x) (((x) & 0x3) << 10) # define TEX_NUM_BANKS(x) (((x) & 0x3) << 16) +#define R_030000_SQ_TEX_RESOURCE_WORD0_0 0x030000 +#define S_030000_DIM(x) (((x) & 0x7) << 0) +#define G_030000_DIM(x) (((x) >> 0) & 0x7) +#define C_030000_DIM 0xFFFFFFF8 +#define V_030000_SQ_TEX_DIM_1D 0x00000000 +#define V_030000_SQ_TEX_DIM_2D 0x00000001 +#define V_030000_SQ_TEX_DIM_3D 0x00000002 +#define V_030000_SQ_TEX_DIM_CUBEMAP 0x00000003 +#define V_030000_SQ_TEX_DIM_1D_ARRAY 0x00000004 +#define V_030000_SQ_TEX_DIM_2D_ARRAY 0x00000005 +#define V_030000_SQ_TEX_DIM_2D_MSAA 0x00000006 +#define V_030000_SQ_TEX_DIM_2D_ARRAY_MSAA 0x00000007 +#define S_030000_NON_DISP_TILING_ORDER(x) (((x) & 0x1) << 5) +#define G_030000_NON_DISP_TILING_ORDER(x) (((x) >> 5) & 0x1) +#define C_030000_NON_DISP_TILING_ORDER 0xFFFFFFDF +#define S_030000_PITCH(x) (((x) & 0xFFF) << 6) +#define G_030000_PITCH(x) (((x) >> 6) & 0xFFF) +#define C_030000_PITCH 0xFFFC003F +#define S_030000_TEX_WIDTH(x) (((x) & 0x3FFF) << 18) +#define G_030000_TEX_WIDTH(x) (((x) >> 18) & 0x3FFF) +#define C_030000_TEX_WIDTH 0x0003FFFF +#define R_030004_SQ_TEX_RESOURCE_WORD1_0 0x030004 +#define S_030004_TEX_HEIGHT(x) (((x) & 0x3FFF) << 0) +#define G_030004_TEX_HEIGHT(x) (((x) >> 0) & 0x3FFF) +#define C_030004_TEX_HEIGHT 0xFFFFC000 +#define S_030004_TEX_DEPTH(x) (((x) & 0x1FFF) << 14) +#define G_030004_TEX_DEPTH(x) (((x) >> 14) & 0x1FFF) +#define C_030004_TEX_DEPTH 0xF8003FFF +#define S_030004_ARRAY_MODE(x) (((x) & 0xF) << 28) +#define G_030004_ARRAY_MODE(x) (((x) >> 28) & 0xF) +#define C_030004_ARRAY_MODE 0x0FFFFFFF +#define R_030008_SQ_TEX_RESOURCE_WORD2_0 0x030008 +#define S_030008_BASE_ADDRESS(x) (((x) & 0xFFFFFFFF) << 0) +#define G_030008_BASE_ADDRESS(x) (((x) >> 0) & 0xFFFFFFFF) +#define C_030008_BASE_ADDRESS 0x00000000 +#define R_03000C_SQ_TEX_RESOURCE_WORD3_0 0x03000C +#define S_03000C_MIP_ADDRESS(x) (((x) & 0xFFFFFFFF) << 0) +#define G_03000C_MIP_ADDRESS(x) (((x) >> 0) & 0xFFFFFFFF) +#define C_03000C_MIP_ADDRESS 0x00000000 +#define R_030010_SQ_TEX_RESOURCE_WORD4_0 0x030010 +#define S_030010_FORMAT_COMP_X(x) (((x) & 0x3) << 0) +#define G_030010_FORMAT_COMP_X(x) (((x) >> 0) & 0x3) +#define C_030010_FORMAT_COMP_X 0xFFFFFFFC +#define V_030010_SQ_FORMAT_COMP_UNSIGNED 0x00000000 +#define V_030010_SQ_FORMAT_COMP_SIGNED 0x00000001 +#define V_030010_SQ_FORMAT_COMP_UNSIGNED_BIASED 0x00000002 +#define S_030010_FORMAT_COMP_Y(x) (((x) & 0x3) << 2) +#define G_030010_FORMAT_COMP_Y(x) (((x) >> 2) & 0x3) +#define C_030010_FORMAT_COMP_Y 0xFFFFFFF3 +#define S_030010_FORMAT_COMP_Z(x) (((x) & 0x3) << 4) +#define G_030010_FORMAT_COMP_Z(x) (((x) >> 4) & 0x3) +#define C_030010_FORMAT_COMP_Z 0xFFFFFFCF +#define S_030010_FORMAT_COMP_W(x) (((x) & 0x3) << 6) +#define G_030010_FORMAT_COMP_W(x) (((x) >> 6) & 0x3) +#define C_030010_FORMAT_COMP_W 0xFFFFFF3F +#define S_030010_NUM_FORMAT_ALL(x) (((x) & 0x3) << 8) +#define G_030010_NUM_FORMAT_ALL(x) (((x) >> 8) & 0x3) +#define C_030010_NUM_FORMAT_ALL 0xFFFFFCFF +#define V_030010_SQ_NUM_FORMAT_NORM 0x00000000 +#define V_030010_SQ_NUM_FORMAT_INT 0x00000001 +#define V_030010_SQ_NUM_FORMAT_SCALED 0x00000002 +#define S_030010_SRF_MODE_ALL(x) (((x) & 0x1) << 10) +#define G_030010_SRF_MODE_ALL(x) (((x) >> 10) & 0x1) +#define C_030010_SRF_MODE_ALL 0xFFFFFBFF +#define V_030010_SRF_MODE_ZERO_CLAMP_MINUS_ONE 0x00000000 +#define V_030010_SRF_MODE_NO_ZERO 0x00000001 +#define S_030010_FORCE_DEGAMMA(x) (((x) & 0x1) << 11) +#define G_030010_FORCE_DEGAMMA(x) (((x) >> 11) & 0x1) +#define C_030010_FORCE_DEGAMMA 0xFFFFF7FF +#define S_030010_ENDIAN_SWAP(x) (((x) & 0x3) << 12) +#define G_030010_ENDIAN_SWAP(x) (((x) >> 12) & 0x3) +#define C_030010_ENDIAN_SWAP 0xFFFFCFFF +#define S_030010_DST_SEL_X(x) (((x) & 0x7) << 16) +#define G_030010_DST_SEL_X(x) (((x) >> 16) & 0x7) +#define C_030010_DST_SEL_X 0xFFF8FFFF +#define V_030010_SQ_SEL_X 0x00000000 +#define V_030010_SQ_SEL_Y 0x00000001 +#define V_030010_SQ_SEL_Z 0x00000002 +#define V_030010_SQ_SEL_W 0x00000003 +#define V_030010_SQ_SEL_0 0x00000004 +#define V_030010_SQ_SEL_1 0x00000005 +#define S_030010_DST_SEL_Y(x) (((x) & 0x7) << 19) +#define G_030010_DST_SEL_Y(x) (((x) >> 19) & 0x7) +#define C_030010_DST_SEL_Y 0xFFC7FFFF +#define S_030010_DST_SEL_Z(x) (((x) & 0x7) << 22) +#define G_030010_DST_SEL_Z(x) (((x) >> 22) & 0x7) +#define C_030010_DST_SEL_Z 0xFE3FFFFF +#define S_030010_DST_SEL_W(x) (((x) & 0x7) << 25) +#define G_030010_DST_SEL_W(x) (((x) >> 25) & 0x7) +#define C_030010_DST_SEL_W 0xF1FFFFFF +#define S_030010_BASE_LEVEL(x) (((x) & 0xF) << 28) +#define G_030010_BASE_LEVEL(x) (((x) >> 28) & 0xF) +#define C_030010_BASE_LEVEL 0x0FFFFFFF +#define R_030014_SQ_TEX_RESOURCE_WORD5_0 0x030014 +#define S_030014_LAST_LEVEL(x) (((x) & 0xF) << 0) +#define G_030014_LAST_LEVEL(x) (((x) >> 0) & 0xF) +#define C_030014_LAST_LEVEL 0xFFFFFFF0 +#define S_030014_BASE_ARRAY(x) (((x) & 0x1FFF) << 4) +#define G_030014_BASE_ARRAY(x) (((x) >> 4) & 0x1FFF) +#define C_030014_BASE_ARRAY 0xFFFE000F +#define S_030014_LAST_ARRAY(x) (((x) & 0x1FFF) << 17) +#define G_030014_LAST_ARRAY(x) (((x) >> 17) & 0x1FFF) +#define C_030014_LAST_ARRAY 0xC001FFFF +#define R_030018_SQ_TEX_RESOURCE_WORD6_0 0x030018 +#define S_030018_MAX_ANISO(x) (((x) & 0x7) << 0) +#define G_030018_MAX_ANISO(x) (((x) >> 0) & 0x7) +#define C_030018_MAX_ANISO 0xFFFFFFF8 +#define S_030018_PERF_MODULATION(x) (((x) & 0x7) << 3) +#define G_030018_PERF_MODULATION(x) (((x) >> 3) & 0x7) +#define C_030018_PERF_MODULATION 0xFFFFFFC7 +#define S_030018_INTERLACED(x) (((x) & 0x1) << 6) +#define G_030018_INTERLACED(x) (((x) >> 6) & 0x1) +#define C_030018_INTERLACED 0xFFFFFFBF +#define S_030018_TILE_SPLIT(x) (((x) & 0x7) << 29) +#define G_030018_TILE_SPLIT(x) (((x) >> 29) & 0x7) +#define R_03001C_SQ_TEX_RESOURCE_WORD7_0 0x03001C +#define S_03001C_MACRO_TILE_ASPECT(x) (((x) & 0x3) << 6) +#define G_03001C_MACRO_TILE_ASPECT(x) (((x) >> 6) & 0x3) +#define S_03001C_BANK_WIDTH(x) (((x) & 0x3) << 8) +#define G_03001C_BANK_WIDTH(x) (((x) >> 8) & 0x3) +#define S_03001C_BANK_HEIGHT(x) (((x) & 0x3) << 10) +#define G_03001C_BANK_HEIGHT(x) (((x) >> 10) & 0x3) +#define S_03001C_NUM_BANKS(x) (((x) & 0x3) << 16) +#define G_03001C_NUM_BANKS(x) (((x) >> 16) & 0x3) +#define S_03001C_TYPE(x) (((x) & 0x3) << 30) +#define G_03001C_TYPE(x) (((x) >> 30) & 0x3) +#define C_03001C_TYPE 0x3FFFFFFF +#define V_03001C_SQ_TEX_VTX_INVALID_TEXTURE 0x00000000 +#define V_03001C_SQ_TEX_VTX_INVALID_BUFFER 0x00000001 +#define V_03001C_SQ_TEX_VTX_VALID_TEXTURE 0x00000002 +#define V_03001C_SQ_TEX_VTX_VALID_BUFFER 0x00000003 +#define S_03001C_DATA_FORMAT(x) (((x) & 0x3F) << 0) +#define G_03001C_DATA_FORMAT(x) (((x) >> 0) & 0x3F) +#define C_03001C_DATA_FORMAT 0xFFFFFFC0 #define SQ_VTX_CONSTANT_WORD0_0 0x30000 #define SQ_VTX_CONSTANT_WORD1_0 0x30004 @@ -1198,8 +1617,40 @@ #define SQ_VTX_CONSTANT_WORD6_0 0x30018 #define SQ_VTX_CONSTANT_WORD7_0 0x3001c +#define TD_PS_BORDER_COLOR_INDEX 0xA400 +#define TD_PS_BORDER_COLOR_RED 0xA404 +#define TD_PS_BORDER_COLOR_GREEN 0xA408 +#define TD_PS_BORDER_COLOR_BLUE 0xA40C +#define TD_PS_BORDER_COLOR_ALPHA 0xA410 +#define TD_VS_BORDER_COLOR_INDEX 0xA414 +#define TD_VS_BORDER_COLOR_RED 0xA418 +#define TD_VS_BORDER_COLOR_GREEN 0xA41C +#define TD_VS_BORDER_COLOR_BLUE 0xA420 +#define TD_VS_BORDER_COLOR_ALPHA 0xA424 +#define TD_GS_BORDER_COLOR_INDEX 0xA428 +#define TD_GS_BORDER_COLOR_RED 0xA42C +#define TD_GS_BORDER_COLOR_GREEN 0xA430 +#define TD_GS_BORDER_COLOR_BLUE 0xA434 +#define TD_GS_BORDER_COLOR_ALPHA 0xA438 +#define TD_HS_BORDER_COLOR_INDEX 0xA43C +#define TD_HS_BORDER_COLOR_RED 0xA440 +#define TD_HS_BORDER_COLOR_GREEN 0xA444 +#define TD_HS_BORDER_COLOR_BLUE 0xA448 +#define TD_HS_BORDER_COLOR_ALPHA 0xA44C +#define TD_LS_BORDER_COLOR_INDEX 0xA450 +#define TD_LS_BORDER_COLOR_RED 0xA454 +#define TD_LS_BORDER_COLOR_GREEN 0xA458 +#define TD_LS_BORDER_COLOR_BLUE 0xA45C +#define TD_LS_BORDER_COLOR_ALPHA 0xA460 +#define TD_CS_BORDER_COLOR_INDEX 0xA464 +#define TD_CS_BORDER_COLOR_RED 0xA468 +#define TD_CS_BORDER_COLOR_GREEN 0xA46C +#define TD_CS_BORDER_COLOR_BLUE 0xA470 +#define TD_CS_BORDER_COLOR_ALPHA 0xA474 + /* cayman 3D regs */ -#define CAYMAN_VGT_OFFCHIP_LDS_BASE 0x89B0 +#define CAYMAN_VGT_OFFCHIP_LDS_BASE 0x89B4 +#define CAYMAN_SQ_EX_ALLOC_TABLE_SLOTS 0x8E48 #define CAYMAN_DB_EQAA 0x28804 #define CAYMAN_DB_DEPTH_INFO 0x2803C #define CAYMAN_PA_SC_AA_CONFIG 0x28BE0 diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c index 636255b..9934c9d 100644 --- a/drivers/gpu/drm/radeon/ni.c +++ b/drivers/gpu/drm/radeon/ni.c @@ -42,6 +42,8 @@ extern void evergreen_irq_suspend(struct radeon_device *rdev); extern int evergreen_mc_init(struct radeon_device *rdev); extern void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev); extern void evergreen_pcie_gen2_enable(struct radeon_device *rdev); +extern void si_rlc_fini(struct radeon_device *rdev); +extern int si_rlc_init(struct radeon_device *rdev); #define EVERGREEN_PFP_UCODE_SIZE 1120 #define EVERGREEN_PM4_UCODE_SIZE 1376 @@ -53,6 +55,8 @@ extern void evergreen_pcie_gen2_enable(struct radeon_device *rdev); #define CAYMAN_RLC_UCODE_SIZE 1024 #define CAYMAN_MC_UCODE_SIZE 6037 +#define ARUBA_RLC_UCODE_SIZE 1536 + /* Firmware Names */ MODULE_FIRMWARE("radeon/BARTS_pfp.bin"); MODULE_FIRMWARE("radeon/BARTS_me.bin"); @@ -68,6 +72,9 @@ MODULE_FIRMWARE("radeon/CAYMAN_pfp.bin"); MODULE_FIRMWARE("radeon/CAYMAN_me.bin"); MODULE_FIRMWARE("radeon/CAYMAN_mc.bin"); MODULE_FIRMWARE("radeon/CAYMAN_rlc.bin"); +MODULE_FIRMWARE("radeon/ARUBA_pfp.bin"); +MODULE_FIRMWARE("radeon/ARUBA_me.bin"); +MODULE_FIRMWARE("radeon/ARUBA_rlc.bin"); #define BTC_IO_MC_REGS_SIZE 29 @@ -326,6 +333,15 @@ int ni_init_microcode(struct radeon_device *rdev) rlc_req_size = CAYMAN_RLC_UCODE_SIZE * 4; mc_req_size = CAYMAN_MC_UCODE_SIZE * 4; break; + case CHIP_ARUBA: + chip_name = "ARUBA"; + rlc_chip_name = "ARUBA"; + /* pfp/me same size as CAYMAN */ + pfp_req_size = CAYMAN_PFP_UCODE_SIZE * 4; + me_req_size = CAYMAN_PM4_UCODE_SIZE * 4; + rlc_req_size = ARUBA_RLC_UCODE_SIZE * 4; + mc_req_size = 0; + break; default: BUG(); } @@ -365,15 +381,18 @@ int ni_init_microcode(struct radeon_device *rdev) err = -EINVAL; } - snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name); - err = request_firmware(&rdev->mc_fw, fw_name, &pdev->dev); - if (err) - goto out; - if (rdev->mc_fw->size != mc_req_size) { - printk(KERN_ERR - "ni_mc: Bogus length %zu in firmware \"%s\"\n", - rdev->mc_fw->size, fw_name); - err = -EINVAL; + /* no MC ucode on TN */ + if (!(rdev->flags & RADEON_IS_IGP)) { + snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name); + err = request_firmware(&rdev->mc_fw, fw_name, &pdev->dev); + if (err) + goto out; + if (rdev->mc_fw->size != mc_req_size) { + printk(KERN_ERR + "ni_mc: Bogus length %zu in firmware \"%s\"\n", + rdev->mc_fw->size, fw_name); + err = -EINVAL; + } } out: platform_device_unregister(pdev); @@ -478,6 +497,7 @@ static u32 cayman_get_tile_pipe_to_backend_map(struct radeon_device *rdev, memset((uint8_t *)&swizzle_pipe[0], 0, sizeof(u32) * CAYMAN_MAX_PIPES); switch (rdev->family) { case CHIP_CAYMAN: + case CHIP_ARUBA: force_no_swizzle = true; break; default: @@ -610,7 +630,6 @@ static void cayman_gpu_init(struct radeon_device *rdev) switch (rdev->family) { case CHIP_CAYMAN: - default: rdev->config.cayman.max_shader_engines = 2; rdev->config.cayman.max_pipes_per_simd = 4; rdev->config.cayman.max_tile_pipes = 8; @@ -632,6 +651,56 @@ static void cayman_gpu_init(struct radeon_device *rdev) rdev->config.cayman.sc_hiz_tile_fifo_size = 0x30; rdev->config.cayman.sc_earlyz_tile_fifo_size = 0x130; break; + case CHIP_ARUBA: + default: + rdev->config.cayman.max_shader_engines = 1; + rdev->config.cayman.max_pipes_per_simd = 4; + rdev->config.cayman.max_tile_pipes = 2; + if ((rdev->pdev->device == 0x9900) || + (rdev->pdev->device == 0x9901) || + (rdev->pdev->device == 0x9905) || + (rdev->pdev->device == 0x9906) || + (rdev->pdev->device == 0x9907) || + (rdev->pdev->device == 0x9908) || + (rdev->pdev->device == 0x9909) || + (rdev->pdev->device == 0x9910) || + (rdev->pdev->device == 0x9917)) { + rdev->config.cayman.max_simds_per_se = 6; + rdev->config.cayman.max_backends_per_se = 2; + } else if ((rdev->pdev->device == 0x9903) || + (rdev->pdev->device == 0x9904) || + (rdev->pdev->device == 0x990A) || + (rdev->pdev->device == 0x9913) || + (rdev->pdev->device == 0x9918)) { + rdev->config.cayman.max_simds_per_se = 4; + rdev->config.cayman.max_backends_per_se = 2; + } else if ((rdev->pdev->device == 0x9919) || + (rdev->pdev->device == 0x9990) || + (rdev->pdev->device == 0x9991) || + (rdev->pdev->device == 0x9994) || + (rdev->pdev->device == 0x99A0)) { + rdev->config.cayman.max_simds_per_se = 3; + rdev->config.cayman.max_backends_per_se = 1; + } else { + rdev->config.cayman.max_simds_per_se = 2; + rdev->config.cayman.max_backends_per_se = 1; + } + rdev->config.cayman.max_texture_channel_caches = 2; + rdev->config.cayman.max_gprs = 256; + rdev->config.cayman.max_threads = 256; + rdev->config.cayman.max_gs_threads = 32; + rdev->config.cayman.max_stack_entries = 512; + rdev->config.cayman.sx_num_of_sets = 8; + rdev->config.cayman.sx_max_export_size = 256; + rdev->config.cayman.sx_max_export_pos_size = 64; + rdev->config.cayman.sx_max_export_smx_size = 192; + rdev->config.cayman.max_hw_contexts = 8; + rdev->config.cayman.sq_num_cf_insts = 2; + + rdev->config.cayman.sc_prim_fifo_size = 0x40; + rdev->config.cayman.sc_hiz_tile_fifo_size = 0x30; + rdev->config.cayman.sc_earlyz_tile_fifo_size = 0x130; + break; } /* Initialize HDP */ @@ -652,7 +721,9 @@ static void cayman_gpu_init(struct radeon_device *rdev) cc_rb_backend_disable = RREG32(CC_RB_BACKEND_DISABLE); cc_gc_shader_pipe_config = RREG32(CC_GC_SHADER_PIPE_CONFIG); - cgts_tcc_disable = 0xff000000; + cgts_tcc_disable = 0xffff0000; + for (i = 0; i < rdev->config.cayman.max_texture_channel_caches; i++) + cgts_tcc_disable &= ~(1 << (16 + i)); gc_user_rb_backend_disable = RREG32(GC_USER_RB_BACKEND_DISABLE); gc_user_shader_pipe_config = RREG32(GC_USER_SHADER_PIPE_CONFIG); cgts_user_tcc_disable = RREG32(CGTS_USER_TCC_DISABLE); @@ -804,17 +875,23 @@ static void cayman_gpu_init(struct radeon_device *rdev) rdev->config.cayman.tile_config |= (3 << 0); break; } - switch ((mc_arb_ramcfg & NOOFBANK_MASK) >> NOOFBANK_SHIFT) { - case 0: /* four banks */ - rdev->config.cayman.tile_config |= 0 << 4; - break; - case 1: /* eight banks */ + + /* num banks is 8 on all fusion asics. 0 = 4, 1 = 8, 2 = 16 */ + if (rdev->flags & RADEON_IS_IGP) rdev->config.cayman.tile_config |= 1 << 4; - break; - case 2: /* sixteen banks */ - default: - rdev->config.cayman.tile_config |= 2 << 4; - break; + else { + switch ((mc_arb_ramcfg & NOOFBANK_MASK) >> NOOFBANK_SHIFT) { + case 0: /* four banks */ + rdev->config.cayman.tile_config |= 0 << 4; + break; + case 1: /* eight banks */ + rdev->config.cayman.tile_config |= 1 << 4; + break; + case 2: /* sixteen banks */ + default: + rdev->config.cayman.tile_config |= 2 << 4; + break; + } } rdev->config.cayman.tile_config |= ((gb_addr_config & PIPE_INTERLEAVE_SIZE_MASK) >> PIPE_INTERLEAVE_SIZE_SHIFT) << 8; @@ -944,7 +1021,7 @@ void cayman_pcie_gart_tlb_flush(struct radeon_device *rdev) int cayman_pcie_gart_enable(struct radeon_device *rdev) { - int r; + int i, r; if (rdev->gart.robj == NULL) { dev_err(rdev->dev, "No VRAM object for PCIE GART.\n"); @@ -955,9 +1032,12 @@ int cayman_pcie_gart_enable(struct radeon_device *rdev) return r; radeon_gart_restore(rdev); /* Setup TLB control */ - WREG32(MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB | + WREG32(MC_VM_MX_L1_TLB_CNTL, + (0xA << 7) | + ENABLE_L1_TLB | ENABLE_L1_FRAGMENT_PROCESSING | SYSTEM_ACCESS_MODE_NOT_IN_SYS | + ENABLE_ADVANCED_DRIVER_MODEL | SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU); /* Setup L2 cache */ WREG32(VM_L2_CNTL, ENABLE_L2_CACHE | @@ -977,9 +1057,26 @@ int cayman_pcie_gart_enable(struct radeon_device *rdev) WREG32(VM_CONTEXT0_CNTL2, 0); WREG32(VM_CONTEXT0_CNTL, ENABLE_CONTEXT | PAGE_TABLE_DEPTH(0) | RANGE_PROTECTION_FAULT_ENABLE_DEFAULT); - /* disable context1-7 */ + + WREG32(0x15D4, 0); + WREG32(0x15D8, 0); + WREG32(0x15DC, 0); + + /* empty context1-7 */ + for (i = 1; i < 8; i++) { + WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR + (i << 2), 0); + WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR + (i << 2), 0); + WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2), + rdev->gart.table_addr >> 12); + } + + /* enable context1-7 */ + WREG32(VM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR, + (u32)(rdev->dummy_page.addr >> 12)); WREG32(VM_CONTEXT1_CNTL2, 0); WREG32(VM_CONTEXT1_CNTL, 0); + WREG32(VM_CONTEXT1_CNTL, ENABLE_CONTEXT | PAGE_TABLE_DEPTH(0) | + RANGE_PROTECTION_FAULT_ENABLE_DEFAULT); cayman_pcie_gart_tlb_flush(rdev); DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n", @@ -1016,9 +1113,69 @@ void cayman_pcie_gart_fini(struct radeon_device *rdev) radeon_gart_fini(rdev); } +void cayman_cp_int_cntl_setup(struct radeon_device *rdev, + int ring, u32 cp_int_cntl) +{ + u32 srbm_gfx_cntl = RREG32(SRBM_GFX_CNTL) & ~3; + + WREG32(SRBM_GFX_CNTL, srbm_gfx_cntl | (ring & 3)); + WREG32(CP_INT_CNTL, cp_int_cntl); +} + /* * CP. */ +void cayman_fence_ring_emit(struct radeon_device *rdev, + struct radeon_fence *fence) +{ + struct radeon_ring *ring = &rdev->ring[fence->ring]; + u64 addr = rdev->fence_drv[fence->ring].gpu_addr; + + /* flush read cache over gart for this vmid */ + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + radeon_ring_write(ring, (CP_COHER_CNTL2 - PACKET3_SET_CONFIG_REG_START) >> 2); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); + radeon_ring_write(ring, PACKET3_TC_ACTION_ENA | PACKET3_SH_ACTION_ENA); + radeon_ring_write(ring, 0xFFFFFFFF); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 10); /* poll interval */ + /* EVENT_WRITE_EOP - flush caches, send int */ + radeon_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE_EOP, 4)); + radeon_ring_write(ring, EVENT_TYPE(CACHE_FLUSH_AND_INV_EVENT_TS) | EVENT_INDEX(5)); + radeon_ring_write(ring, addr & 0xffffffff); + radeon_ring_write(ring, (upper_32_bits(addr) & 0xff) | DATA_SEL(1) | INT_SEL(2)); + radeon_ring_write(ring, fence->seq); + radeon_ring_write(ring, 0); +} + +void cayman_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) +{ + struct radeon_ring *ring = &rdev->ring[ib->fence->ring]; + + /* set to DX10/11 mode */ + radeon_ring_write(ring, PACKET3(PACKET3_MODE_CONTROL, 0)); + radeon_ring_write(ring, 1); + radeon_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); + radeon_ring_write(ring, +#ifdef __BIG_ENDIAN + (2 << 0) | +#endif + (ib->gpu_addr & 0xFFFFFFFC)); + radeon_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xFF); + radeon_ring_write(ring, ib->length_dw | (ib->vm_id << 24)); + + /* flush read cache over gart for this vmid */ + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + radeon_ring_write(ring, (CP_COHER_CNTL2 - PACKET3_SET_CONFIG_REG_START) >> 2); + radeon_ring_write(ring, ib->vm_id); + radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); + radeon_ring_write(ring, PACKET3_TC_ACTION_ENA | PACKET3_SH_ACTION_ENA); + radeon_ring_write(ring, 0xFFFFFFFF); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 10); /* poll interval */ +} + static void cayman_cp_enable(struct radeon_device *rdev, bool enable) { if (enable) @@ -1059,63 +1216,64 @@ static int cayman_cp_load_microcode(struct radeon_device *rdev) static int cayman_cp_start(struct radeon_device *rdev) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; int r, i; - r = radeon_ring_lock(rdev, 7); + r = radeon_ring_lock(rdev, ring, 7); if (r) { DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r); return r; } - radeon_ring_write(rdev, PACKET3(PACKET3_ME_INITIALIZE, 5)); - radeon_ring_write(rdev, 0x1); - radeon_ring_write(rdev, 0x0); - radeon_ring_write(rdev, rdev->config.cayman.max_hw_contexts - 1); - radeon_ring_write(rdev, PACKET3_ME_INITIALIZE_DEVICE_ID(1)); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, 0); - radeon_ring_unlock_commit(rdev); + radeon_ring_write(ring, PACKET3(PACKET3_ME_INITIALIZE, 5)); + radeon_ring_write(ring, 0x1); + radeon_ring_write(ring, 0x0); + radeon_ring_write(ring, rdev->config.cayman.max_hw_contexts - 1); + radeon_ring_write(ring, PACKET3_ME_INITIALIZE_DEVICE_ID(1)); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 0); + radeon_ring_unlock_commit(rdev, ring); cayman_cp_enable(rdev, true); - r = radeon_ring_lock(rdev, cayman_default_size + 19); + r = radeon_ring_lock(rdev, ring, cayman_default_size + 19); if (r) { DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r); return r; } /* setup clear context state */ - radeon_ring_write(rdev, PACKET3(PACKET3_PREAMBLE_CNTL, 0)); - radeon_ring_write(rdev, PACKET3_PREAMBLE_BEGIN_CLEAR_STATE); + radeon_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0)); + radeon_ring_write(ring, PACKET3_PREAMBLE_BEGIN_CLEAR_STATE); for (i = 0; i < cayman_default_size; i++) - radeon_ring_write(rdev, cayman_default_state[i]); + radeon_ring_write(ring, cayman_default_state[i]); - radeon_ring_write(rdev, PACKET3(PACKET3_PREAMBLE_CNTL, 0)); - radeon_ring_write(rdev, PACKET3_PREAMBLE_END_CLEAR_STATE); + radeon_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0)); + radeon_ring_write(ring, PACKET3_PREAMBLE_END_CLEAR_STATE); /* set clear context state */ - radeon_ring_write(rdev, PACKET3(PACKET3_CLEAR_STATE, 0)); - radeon_ring_write(rdev, 0); + radeon_ring_write(ring, PACKET3(PACKET3_CLEAR_STATE, 0)); + radeon_ring_write(ring, 0); /* SQ_VTX_BASE_VTX_LOC */ - radeon_ring_write(rdev, 0xc0026f00); - radeon_ring_write(rdev, 0x00000000); - radeon_ring_write(rdev, 0x00000000); - radeon_ring_write(rdev, 0x00000000); + radeon_ring_write(ring, 0xc0026f00); + radeon_ring_write(ring, 0x00000000); + radeon_ring_write(ring, 0x00000000); + radeon_ring_write(ring, 0x00000000); /* Clear consts */ - radeon_ring_write(rdev, 0xc0036f00); - radeon_ring_write(rdev, 0x00000bc4); - radeon_ring_write(rdev, 0xffffffff); - radeon_ring_write(rdev, 0xffffffff); - radeon_ring_write(rdev, 0xffffffff); + radeon_ring_write(ring, 0xc0036f00); + radeon_ring_write(ring, 0x00000bc4); + radeon_ring_write(ring, 0xffffffff); + radeon_ring_write(ring, 0xffffffff); + radeon_ring_write(ring, 0xffffffff); - radeon_ring_write(rdev, 0xc0026900); - radeon_ring_write(rdev, 0x00000316); - radeon_ring_write(rdev, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */ - radeon_ring_write(rdev, 0x00000010); /* */ + radeon_ring_write(ring, 0xc0026900); + radeon_ring_write(ring, 0x00000316); + radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */ + radeon_ring_write(ring, 0x00000010); /* */ - radeon_ring_unlock_commit(rdev); + radeon_ring_unlock_commit(rdev, ring); /* XXX init other rings */ @@ -1125,11 +1283,12 @@ static int cayman_cp_start(struct radeon_device *rdev) static void cayman_cp_fini(struct radeon_device *rdev) { cayman_cp_enable(rdev, false); - radeon_ring_fini(rdev); + radeon_ring_fini(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); } int cayman_cp_resume(struct radeon_device *rdev) { + struct radeon_ring *ring; u32 tmp; u32 rb_bufsz; int r; @@ -1146,7 +1305,8 @@ int cayman_cp_resume(struct radeon_device *rdev) WREG32(GRBM_SOFT_RESET, 0); RREG32(GRBM_SOFT_RESET); - WREG32(CP_SEM_WAIT_TIMER, 0x4); + WREG32(CP_SEM_WAIT_TIMER, 0x0); + WREG32(CP_SEM_INCOMPLETE_TIMER_CNTL, 0x0); /* Set the write pointer delay */ WREG32(CP_RB_WPTR_DELAY, 0); @@ -1155,7 +1315,8 @@ int cayman_cp_resume(struct radeon_device *rdev) /* ring 0 - compute and gfx */ /* Set ring buffer size */ - rb_bufsz = drm_order(rdev->cp.ring_size / 8); + ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; + rb_bufsz = drm_order(ring->ring_size / 8); tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; #ifdef __BIG_ENDIAN tmp |= BUF_SWAP_32BIT; @@ -1164,8 +1325,8 @@ int cayman_cp_resume(struct radeon_device *rdev) /* Initialize the ring buffer's read and write pointers */ WREG32(CP_RB0_CNTL, tmp | RB_RPTR_WR_ENA); - rdev->cp.wptr = 0; - WREG32(CP_RB0_WPTR, rdev->cp.wptr); + ring->wptr = 0; + WREG32(CP_RB0_WPTR, ring->wptr); /* set the wb address wether it's enabled or not */ WREG32(CP_RB0_RPTR_ADDR, (rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFFFFFFFC); @@ -1182,13 +1343,14 @@ int cayman_cp_resume(struct radeon_device *rdev) mdelay(1); WREG32(CP_RB0_CNTL, tmp); - WREG32(CP_RB0_BASE, rdev->cp.gpu_addr >> 8); + WREG32(CP_RB0_BASE, ring->gpu_addr >> 8); - rdev->cp.rptr = RREG32(CP_RB0_RPTR); + ring->rptr = RREG32(CP_RB0_RPTR); /* ring1 - compute only */ /* Set ring buffer size */ - rb_bufsz = drm_order(rdev->cp1.ring_size / 8); + ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]; + rb_bufsz = drm_order(ring->ring_size / 8); tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; #ifdef __BIG_ENDIAN tmp |= BUF_SWAP_32BIT; @@ -1197,8 +1359,8 @@ int cayman_cp_resume(struct radeon_device *rdev) /* Initialize the ring buffer's read and write pointers */ WREG32(CP_RB1_CNTL, tmp | RB_RPTR_WR_ENA); - rdev->cp1.wptr = 0; - WREG32(CP_RB1_WPTR, rdev->cp1.wptr); + ring->wptr = 0; + WREG32(CP_RB1_WPTR, ring->wptr); /* set the wb address wether it's enabled or not */ WREG32(CP_RB1_RPTR_ADDR, (rdev->wb.gpu_addr + RADEON_WB_CP1_RPTR_OFFSET) & 0xFFFFFFFC); @@ -1207,13 +1369,14 @@ int cayman_cp_resume(struct radeon_device *rdev) mdelay(1); WREG32(CP_RB1_CNTL, tmp); - WREG32(CP_RB1_BASE, rdev->cp1.gpu_addr >> 8); + WREG32(CP_RB1_BASE, ring->gpu_addr >> 8); - rdev->cp1.rptr = RREG32(CP_RB1_RPTR); + ring->rptr = RREG32(CP_RB1_RPTR); /* ring2 - compute only */ /* Set ring buffer size */ - rb_bufsz = drm_order(rdev->cp2.ring_size / 8); + ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]; + rb_bufsz = drm_order(ring->ring_size / 8); tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; #ifdef __BIG_ENDIAN tmp |= BUF_SWAP_32BIT; @@ -1222,8 +1385,8 @@ int cayman_cp_resume(struct radeon_device *rdev) /* Initialize the ring buffer's read and write pointers */ WREG32(CP_RB2_CNTL, tmp | RB_RPTR_WR_ENA); - rdev->cp2.wptr = 0; - WREG32(CP_RB2_WPTR, rdev->cp2.wptr); + ring->wptr = 0; + WREG32(CP_RB2_WPTR, ring->wptr); /* set the wb address wether it's enabled or not */ WREG32(CP_RB2_RPTR_ADDR, (rdev->wb.gpu_addr + RADEON_WB_CP2_RPTR_OFFSET) & 0xFFFFFFFC); @@ -1232,28 +1395,28 @@ int cayman_cp_resume(struct radeon_device *rdev) mdelay(1); WREG32(CP_RB2_CNTL, tmp); - WREG32(CP_RB2_BASE, rdev->cp2.gpu_addr >> 8); + WREG32(CP_RB2_BASE, ring->gpu_addr >> 8); - rdev->cp2.rptr = RREG32(CP_RB2_RPTR); + ring->rptr = RREG32(CP_RB2_RPTR); /* start the rings */ cayman_cp_start(rdev); - rdev->cp.ready = true; - rdev->cp1.ready = true; - rdev->cp2.ready = true; + rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = true; + rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX].ready = false; + rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX].ready = false; /* this only test cp0 */ - r = radeon_ring_test(rdev); + r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); if (r) { - rdev->cp.ready = false; - rdev->cp1.ready = false; - rdev->cp2.ready = false; + rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false; + rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX].ready = false; + rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX].ready = false; return r; } return 0; } -bool cayman_gpu_is_lockup(struct radeon_device *rdev) +bool cayman_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) { u32 srbm_status; u32 grbm_status; @@ -1266,20 +1429,20 @@ bool cayman_gpu_is_lockup(struct radeon_device *rdev) grbm_status_se0 = RREG32(GRBM_STATUS_SE0); grbm_status_se1 = RREG32(GRBM_STATUS_SE1); if (!(grbm_status & GUI_ACTIVE)) { - r100_gpu_lockup_update(lockup, &rdev->cp); + r100_gpu_lockup_update(lockup, ring); return false; } /* force CP activities */ - r = radeon_ring_lock(rdev, 2); + r = radeon_ring_lock(rdev, ring, 2); if (!r) { /* PACKET2 NOP */ - radeon_ring_write(rdev, 0x80000000); - radeon_ring_write(rdev, 0x80000000); - radeon_ring_unlock_commit(rdev); + radeon_ring_write(ring, 0x80000000); + radeon_ring_write(ring, 0x80000000); + radeon_ring_unlock_commit(rdev, ring); } /* XXX deal with CP0,1,2 */ - rdev->cp.rptr = RREG32(CP_RB0_RPTR); - return r100_gpu_cp_is_lockup(rdev, lockup, &rdev->cp); + ring->rptr = RREG32(ring->rptr_reg); + return r100_gpu_cp_is_lockup(rdev, lockup, ring); } static int cayman_gpu_soft_reset(struct radeon_device *rdev) @@ -1299,6 +1462,15 @@ static int cayman_gpu_soft_reset(struct radeon_device *rdev) RREG32(GRBM_STATUS_SE1)); dev_info(rdev->dev, " SRBM_STATUS=0x%08X\n", RREG32(SRBM_STATUS)); + dev_info(rdev->dev, " VM_CONTEXT0_PROTECTION_FAULT_ADDR 0x%08X\n", + RREG32(0x14F8)); + dev_info(rdev->dev, " VM_CONTEXT0_PROTECTION_FAULT_STATUS 0x%08X\n", + RREG32(0x14D8)); + dev_info(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_ADDR 0x%08X\n", + RREG32(0x14FC)); + dev_info(rdev->dev, " VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n", + RREG32(0x14DC)); + evergreen_mc_stop(rdev, &save); if (evergreen_mc_wait_for_idle(rdev)) { dev_warn(rdev->dev, "Wait for MC idle timedout !\n"); @@ -1329,6 +1501,7 @@ static int cayman_gpu_soft_reset(struct radeon_device *rdev) (void)RREG32(GRBM_SOFT_RESET); /* Wait a little for things to settle down */ udelay(50); + dev_info(rdev->dev, " GRBM_STATUS=0x%08X\n", RREG32(GRBM_STATUS)); dev_info(rdev->dev, " GRBM_STATUS_SE0=0x%08X\n", @@ -1348,23 +1521,35 @@ int cayman_asic_reset(struct radeon_device *rdev) static int cayman_startup(struct radeon_device *rdev) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; int r; /* enable pcie gen2 link */ evergreen_pcie_gen2_enable(rdev); - if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw || !rdev->mc_fw) { - r = ni_init_microcode(rdev); + if (rdev->flags & RADEON_IS_IGP) { + if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw) { + r = ni_init_microcode(rdev); + if (r) { + DRM_ERROR("Failed to load firmware!\n"); + return r; + } + } + } else { + if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw || !rdev->mc_fw) { + r = ni_init_microcode(rdev); + if (r) { + DRM_ERROR("Failed to load firmware!\n"); + return r; + } + } + + r = ni_mc_load_microcode(rdev); if (r) { - DRM_ERROR("Failed to load firmware!\n"); + DRM_ERROR("Failed to load MC firmware!\n"); return r; } } - r = ni_mc_load_microcode(rdev); - if (r) { - DRM_ERROR("Failed to load MC firmware!\n"); - return r; - } r = r600_vram_scratch_init(rdev); if (r) @@ -1379,15 +1564,42 @@ static int cayman_startup(struct radeon_device *rdev) r = evergreen_blit_init(rdev); if (r) { r600_blit_fini(rdev); - rdev->asic->copy = NULL; + rdev->asic->copy.copy = NULL; dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r); } + /* allocate rlc buffers */ + if (rdev->flags & RADEON_IS_IGP) { + r = si_rlc_init(rdev); + if (r) { + DRM_ERROR("Failed to init rlc BOs!\n"); + return r; + } + } + /* allocate wb buffer */ r = radeon_wb_init(rdev); if (r) return r; + r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r); + return r; + } + + r = radeon_fence_driver_start_ring(rdev, CAYMAN_RING_TYPE_CP1_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r); + return r; + } + + r = radeon_fence_driver_start_ring(rdev, CAYMAN_RING_TYPE_CP2_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r); + return r; + } + /* Enable IRQ */ r = r600_irq_init(rdev); if (r) { @@ -1397,7 +1609,9 @@ static int cayman_startup(struct radeon_device *rdev) } evergreen_irq_set(rdev); - r = radeon_ring_init(rdev, rdev->cp.ring_size); + r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET, + CP_RB0_RPTR, CP_RB0_WPTR, + 0, 0xfffff, RADEON_CP_PACKET2); if (r) return r; r = cayman_cp_load_microcode(rdev); @@ -1407,6 +1621,21 @@ static int cayman_startup(struct radeon_device *rdev) if (r) return r; + r = radeon_ib_pool_start(rdev); + if (r) + return r; + + r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); + if (r) { + DRM_ERROR("radeon: failed testing IB (%d).\n", r); + rdev->accel_working = false; + return r; + } + + r = radeon_vm_manager_start(rdev); + if (r) + return r; + return 0; } @@ -1421,32 +1650,27 @@ int cayman_resume(struct radeon_device *rdev) /* post card */ atom_asic_init(rdev->mode_info.atom_context); + rdev->accel_working = true; r = cayman_startup(rdev); if (r) { DRM_ERROR("cayman startup failed on resume\n"); + rdev->accel_working = false; return r; } - - r = r600_ib_test(rdev); - if (r) { - DRM_ERROR("radeon: failled testing IB (%d).\n", r); - return r; - } - return r; - } int cayman_suspend(struct radeon_device *rdev) { /* FIXME: we should wait for ring to be empty */ + radeon_ib_pool_suspend(rdev); + radeon_vm_manager_suspend(rdev); + r600_blit_suspend(rdev); cayman_cp_enable(rdev, false); - rdev->cp.ready = false; + rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false; evergreen_irq_suspend(rdev); radeon_wb_disable(rdev); cayman_pcie_gart_disable(rdev); - r600_blit_suspend(rdev); - return 0; } @@ -1458,6 +1682,7 @@ int cayman_suspend(struct radeon_device *rdev) */ int cayman_init(struct radeon_device *rdev) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; int r; /* This don't do much */ @@ -1510,8 +1735,8 @@ int cayman_init(struct radeon_device *rdev) if (r) return r; - rdev->cp.ring_obj = NULL; - r600_ring_init(rdev, 1024 * 1024); + ring->ring_obj = NULL; + r600_ring_init(rdev, ring, 1024 * 1024); rdev->ih.ring_obj = NULL; r600_ih_ring_init(rdev, 64 * 1024); @@ -1520,35 +1745,40 @@ int cayman_init(struct radeon_device *rdev) if (r) return r; + r = radeon_ib_pool_init(rdev); rdev->accel_working = true; + if (r) { + dev_err(rdev->dev, "IB initialization failed (%d).\n", r); + rdev->accel_working = false; + } + r = radeon_vm_manager_init(rdev); + if (r) { + dev_err(rdev->dev, "vm manager initialization failed (%d).\n", r); + } + r = cayman_startup(rdev); if (r) { dev_err(rdev->dev, "disabling GPU acceleration\n"); cayman_cp_fini(rdev); r600_irq_fini(rdev); + if (rdev->flags & RADEON_IS_IGP) + si_rlc_fini(rdev); radeon_wb_fini(rdev); + r100_ib_fini(rdev); + radeon_vm_manager_fini(rdev); radeon_irq_kms_fini(rdev); cayman_pcie_gart_fini(rdev); rdev->accel_working = false; } - if (rdev->accel_working) { - r = radeon_ib_pool_init(rdev); - if (r) { - DRM_ERROR("radeon: failed initializing IB pool (%d).\n", r); - rdev->accel_working = false; - } - r = r600_ib_test(rdev); - if (r) { - DRM_ERROR("radeon: failed testing IB (%d).\n", r); - rdev->accel_working = false; - } - } /* Don't start up if the MC ucode is missing. * The default clocks and voltages before the MC ucode * is loaded are not suffient for advanced operations. + * + * We can skip this check for TN, because there is no MC + * ucode. */ - if (!rdev->mc_fw) { + if (!rdev->mc_fw && !(rdev->flags & RADEON_IS_IGP)) { DRM_ERROR("radeon: MC ucode required for NI+.\n"); return -EINVAL; } @@ -1561,12 +1791,16 @@ void cayman_fini(struct radeon_device *rdev) r600_blit_fini(rdev); cayman_cp_fini(rdev); r600_irq_fini(rdev); + if (rdev->flags & RADEON_IS_IGP) + si_rlc_fini(rdev); radeon_wb_fini(rdev); - radeon_ib_pool_fini(rdev); + radeon_vm_manager_fini(rdev); + r100_ib_fini(rdev); radeon_irq_kms_fini(rdev); cayman_pcie_gart_fini(rdev); r600_vram_scratch_fini(rdev); radeon_gem_fini(rdev); + radeon_semaphore_driver_fini(rdev); radeon_fence_driver_fini(rdev); radeon_bo_fini(rdev); radeon_atombios_fini(rdev); @@ -1574,3 +1808,89 @@ void cayman_fini(struct radeon_device *rdev) rdev->bios = NULL; } +/* + * vm + */ +int cayman_vm_init(struct radeon_device *rdev) +{ + /* number of VMs */ + rdev->vm_manager.nvm = 8; + /* base offset of vram pages */ + if (rdev->flags & RADEON_IS_IGP) { + u64 tmp = RREG32(FUS_MC_VM_FB_OFFSET); + tmp <<= 22; + rdev->vm_manager.vram_base_offset = tmp; + } else + rdev->vm_manager.vram_base_offset = 0; + return 0; +} + +void cayman_vm_fini(struct radeon_device *rdev) +{ +} + +int cayman_vm_bind(struct radeon_device *rdev, struct radeon_vm *vm, int id) +{ + WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR + (id << 2), 0); + WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR + (id << 2), vm->last_pfn); + WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (id << 2), vm->pt_gpu_addr >> 12); + /* flush hdp cache */ + WREG32(HDP_MEM_COHERENCY_FLUSH_CNTL, 0x1); + /* bits 0-7 are the VM contexts0-7 */ + WREG32(VM_INVALIDATE_REQUEST, 1 << id); + return 0; +} + +void cayman_vm_unbind(struct radeon_device *rdev, struct radeon_vm *vm) +{ + WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR + (vm->id << 2), 0); + WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR + (vm->id << 2), 0); + WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2), 0); + /* flush hdp cache */ + WREG32(HDP_MEM_COHERENCY_FLUSH_CNTL, 0x1); + /* bits 0-7 are the VM contexts0-7 */ + WREG32(VM_INVALIDATE_REQUEST, 1 << vm->id); +} + +void cayman_vm_tlb_flush(struct radeon_device *rdev, struct radeon_vm *vm) +{ + if (vm->id == -1) + return; + + /* flush hdp cache */ + WREG32(HDP_MEM_COHERENCY_FLUSH_CNTL, 0x1); + /* bits 0-7 are the VM contexts0-7 */ + WREG32(VM_INVALIDATE_REQUEST, 1 << vm->id); +} + +#define R600_PTE_VALID (1 << 0) +#define R600_PTE_SYSTEM (1 << 1) +#define R600_PTE_SNOOPED (1 << 2) +#define R600_PTE_READABLE (1 << 5) +#define R600_PTE_WRITEABLE (1 << 6) + +uint32_t cayman_vm_page_flags(struct radeon_device *rdev, + struct radeon_vm *vm, + uint32_t flags) +{ + uint32_t r600_flags = 0; + + r600_flags |= (flags & RADEON_VM_PAGE_VALID) ? R600_PTE_VALID : 0; + r600_flags |= (flags & RADEON_VM_PAGE_READABLE) ? R600_PTE_READABLE : 0; + r600_flags |= (flags & RADEON_VM_PAGE_WRITEABLE) ? R600_PTE_WRITEABLE : 0; + if (flags & RADEON_VM_PAGE_SYSTEM) { + r600_flags |= R600_PTE_SYSTEM; + r600_flags |= (flags & RADEON_VM_PAGE_SNOOPED) ? R600_PTE_SNOOPED : 0; + } + return r600_flags; +} + +void cayman_vm_set_page(struct radeon_device *rdev, struct radeon_vm *vm, + unsigned pfn, uint64_t addr, uint32_t flags) +{ + void __iomem *ptr = (void *)vm->pt; + + addr = addr & 0xFFFFFFFFFFFFF000ULL; + addr |= flags; + writeq(addr, ptr + (pfn * 8)); +} diff --git a/drivers/gpu/drm/radeon/nid.h b/drivers/gpu/drm/radeon/nid.h index 4672869..2aa7046 100644 --- a/drivers/gpu/drm/radeon/nid.h +++ b/drivers/gpu/drm/radeon/nid.h @@ -42,6 +42,9 @@ #define CAYMAN_MAX_TCC_MASK 0xFF #define DMIF_ADDR_CONFIG 0xBD4 +#define SRBM_GFX_CNTL 0x0E44 +#define RINGID(x) (((x) & 0x3) << 0) +#define VMID(x) (((x) & 0x7) << 0) #define SRBM_STATUS 0x0E50 #define VM_CONTEXT0_REQUEST_RESPONSE 0x1470 @@ -103,6 +106,7 @@ #define SYSTEM_ACCESS_MODE_NOT_IN_SYS (3 << 3) #define SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU (0 << 5) #define ENABLE_ADVANCED_DRIVER_MODEL (1 << 6) +#define FUS_MC_VM_FB_OFFSET 0x2068 #define MC_SHARED_BLACKOUT_CNTL 0x20ac #define MC_ARB_RAMCFG 0x2760 @@ -219,6 +223,8 @@ #define SCRATCH_UMSK 0x8540 #define SCRATCH_ADDR 0x8544 #define CP_SEM_WAIT_TIMER 0x85BC +#define CP_SEM_INCOMPLETE_TIMER_CNTL 0x85C8 +#define CP_COHER_CNTL2 0x85E8 #define CP_ME_CNTL 0x86D8 #define CP_ME_HALT (1 << 28) #define CP_PFP_HALT (1 << 26) @@ -394,6 +400,12 @@ #define CP_RB0_RPTR_ADDR 0xC10C #define CP_RB0_RPTR_ADDR_HI 0xC110 #define CP_RB0_WPTR 0xC114 + +#define CP_INT_CNTL 0xC124 +# define CNTX_BUSY_INT_ENABLE (1 << 19) +# define CNTX_EMPTY_INT_ENABLE (1 << 20) +# define TIME_STAMP_INT_ENABLE (1 << 26) + #define CP_RB1_BASE 0xC180 #define CP_RB1_CNTL 0xC184 #define CP_RB1_RPTR_ADDR 0xC188 @@ -411,6 +423,10 @@ #define CP_ME_RAM_DATA 0xC160 #define CP_DEBUG 0xC1FC +#define VGT_EVENT_INITIATOR 0x28a90 +# define CACHE_FLUSH_AND_INV_EVENT_TS (0x14 << 0) +# define CACHE_FLUSH_AND_INV_EVENT (0x16 << 0) + /* * PM4 */ @@ -445,6 +461,7 @@ #define PACKET3_DISPATCH_DIRECT 0x15 #define PACKET3_DISPATCH_INDIRECT 0x16 #define PACKET3_INDIRECT_BUFFER_END 0x17 +#define PACKET3_MODE_CONTROL 0x18 #define PACKET3_SET_PREDICATION 0x20 #define PACKET3_REG_RMW 0x21 #define PACKET3_COND_EXEC 0x22 @@ -494,7 +511,27 @@ #define PACKET3_ME_INITIALIZE_DEVICE_ID(x) ((x) << 16) #define PACKET3_COND_WRITE 0x45 #define PACKET3_EVENT_WRITE 0x46 +#define EVENT_TYPE(x) ((x) << 0) +#define EVENT_INDEX(x) ((x) << 8) + /* 0 - any non-TS event + * 1 - ZPASS_DONE + * 2 - SAMPLE_PIPELINESTAT + * 3 - SAMPLE_STREAMOUTSTAT* + * 4 - *S_PARTIAL_FLUSH + * 5 - TS events + */ #define PACKET3_EVENT_WRITE_EOP 0x47 +#define DATA_SEL(x) ((x) << 29) + /* 0 - discard + * 1 - send low 32bit data + * 2 - send 64bit data + * 3 - send 64bit counter value + */ +#define INT_SEL(x) ((x) << 24) + /* 0 - none + * 1 - interrupt only (DATA_SEL = 0) + * 2 - interrupt when data write is confirmed + */ #define PACKET3_EVENT_WRITE_EOS 0x48 #define PACKET3_PREAMBLE_CNTL 0x4A # define PACKET3_PREAMBLE_BEGIN_CLEAR_STATE (2 << 28) diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c index fad7cd1..fe33d35 100644 --- a/drivers/gpu/drm/radeon/r100.c +++ b/drivers/gpu/drm/radeon/r100.c @@ -65,6 +65,40 @@ MODULE_FIRMWARE(FIRMWARE_R520); #include "r100_track.h" +void r100_wait_for_vblank(struct radeon_device *rdev, int crtc) +{ + struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc]; + int i; + + if (radeon_crtc->crtc_id == 0) { + if (RREG32(RADEON_CRTC_GEN_CNTL) & RADEON_CRTC_EN) { + for (i = 0; i < rdev->usec_timeout; i++) { + if (!(RREG32(RADEON_CRTC_STATUS) & RADEON_CRTC_VBLANK_CUR)) + break; + udelay(1); + } + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32(RADEON_CRTC_STATUS) & RADEON_CRTC_VBLANK_CUR) + break; + udelay(1); + } + } + } else { + if (RREG32(RADEON_CRTC2_GEN_CNTL) & RADEON_CRTC2_EN) { + for (i = 0; i < rdev->usec_timeout; i++) { + if (!(RREG32(RADEON_CRTC2_STATUS) & RADEON_CRTC2_VBLANK_CUR)) + break; + udelay(1); + } + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32(RADEON_CRTC2_STATUS) & RADEON_CRTC2_VBLANK_CUR) + break; + udelay(1); + } + } + } +} + /* This files gather functions specifics to: * r100,rv100,rs100,rv200,rs200,r200,rv250,rs300,rv280 */ @@ -87,23 +121,27 @@ int r100_reloc_pitch_offset(struct radeon_cs_parser *p, r100_cs_dump_packet(p, pkt); return r; } + value = radeon_get_ib_value(p, idx); tmp = value & 0x003fffff; tmp += (((u32)reloc->lobj.gpu_offset) >> 10); - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) - tile_flags |= RADEON_DST_TILE_MACRO; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) { - if (reg == RADEON_SRC_PITCH_OFFSET) { - DRM_ERROR("Cannot src blit from microtiled surface\n"); - r100_cs_dump_packet(p, pkt); - return -EINVAL; + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { + if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + tile_flags |= RADEON_DST_TILE_MACRO; + if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) { + if (reg == RADEON_SRC_PITCH_OFFSET) { + DRM_ERROR("Cannot src blit from microtiled surface\n"); + r100_cs_dump_packet(p, pkt); + return -EINVAL; + } + tile_flags |= RADEON_DST_TILE_MICRO; } - tile_flags |= RADEON_DST_TILE_MICRO; - } - tmp |= tile_flags; - p->ib->ptr[idx] = (value & 0x3fc00000) | tmp; + tmp |= tile_flags; + p->ib->ptr[idx] = (value & 0x3fc00000) | tmp; + } else + p->ib->ptr[idx] = (value & 0xffc00000) | tmp; return 0; } @@ -412,7 +450,7 @@ void r100_pm_misc(struct radeon_device *rdev) /* set pcie lanes */ if ((rdev->flags & RADEON_IS_PCIE) && !(rdev->flags & RADEON_IS_IGP) && - rdev->asic->set_pcie_lanes && + rdev->asic->pm.set_pcie_lanes && (ps->pcie_lanes != rdev->pm.power_state[rdev->pm.current_power_state_index].pcie_lanes)) { radeon_set_pcie_lanes(rdev, @@ -592,8 +630,8 @@ int r100_pci_gart_init(struct radeon_device *rdev) if (r) return r; rdev->gart.table_size = rdev->gart.num_gpu_pages * 4; - rdev->asic->gart_tlb_flush = &r100_pci_gart_tlb_flush; - rdev->asic->gart_set_page = &r100_pci_gart_set_page; + rdev->asic->gart.tlb_flush = &r100_pci_gart_tlb_flush; + rdev->asic->gart.set_page = &r100_pci_gart_set_page; return radeon_gart_table_ram_alloc(rdev); } @@ -667,7 +705,7 @@ int r100_irq_set(struct radeon_device *rdev) WREG32(R_000040_GEN_INT_CNTL, 0); return -EINVAL; } - if (rdev->irq.sw_int) { + if (rdev->irq.sw_int[RADEON_RING_TYPE_GFX_INDEX]) { tmp |= RADEON_SW_INT_ENABLE; } if (rdev->irq.gui_idle) { @@ -739,7 +777,7 @@ int r100_irq_process(struct radeon_device *rdev) while (status) { /* SW interrupt */ if (status & RADEON_SW_INT_TEST) { - radeon_fence_process(rdev); + radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX); } /* gui idle interrupt */ if (status & RADEON_GUI_IDLE_STAT) { @@ -809,25 +847,36 @@ u32 r100_get_vblank_counter(struct radeon_device *rdev, int crtc) void r100_fence_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence) { + struct radeon_ring *ring = &rdev->ring[fence->ring]; + /* We have to make sure that caches are flushed before * CPU might read something from VRAM. */ - radeon_ring_write(rdev, PACKET0(RADEON_RB3D_DSTCACHE_CTLSTAT, 0)); - radeon_ring_write(rdev, RADEON_RB3D_DC_FLUSH_ALL); - radeon_ring_write(rdev, PACKET0(RADEON_RB3D_ZCACHE_CTLSTAT, 0)); - radeon_ring_write(rdev, RADEON_RB3D_ZC_FLUSH_ALL); + radeon_ring_write(ring, PACKET0(RADEON_RB3D_DSTCACHE_CTLSTAT, 0)); + radeon_ring_write(ring, RADEON_RB3D_DC_FLUSH_ALL); + radeon_ring_write(ring, PACKET0(RADEON_RB3D_ZCACHE_CTLSTAT, 0)); + radeon_ring_write(ring, RADEON_RB3D_ZC_FLUSH_ALL); /* Wait until IDLE & CLEAN */ - radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0)); - radeon_ring_write(rdev, RADEON_WAIT_2D_IDLECLEAN | RADEON_WAIT_3D_IDLECLEAN); - radeon_ring_write(rdev, PACKET0(RADEON_HOST_PATH_CNTL, 0)); - radeon_ring_write(rdev, rdev->config.r100.hdp_cntl | + radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0)); + radeon_ring_write(ring, RADEON_WAIT_2D_IDLECLEAN | RADEON_WAIT_3D_IDLECLEAN); + radeon_ring_write(ring, PACKET0(RADEON_HOST_PATH_CNTL, 0)); + radeon_ring_write(ring, rdev->config.r100.hdp_cntl | RADEON_HDP_READ_BUFFER_INVALIDATE); - radeon_ring_write(rdev, PACKET0(RADEON_HOST_PATH_CNTL, 0)); - radeon_ring_write(rdev, rdev->config.r100.hdp_cntl); + radeon_ring_write(ring, PACKET0(RADEON_HOST_PATH_CNTL, 0)); + radeon_ring_write(ring, rdev->config.r100.hdp_cntl); /* Emit fence sequence & fire IRQ */ - radeon_ring_write(rdev, PACKET0(rdev->fence_drv.scratch_reg, 0)); - radeon_ring_write(rdev, fence->seq); - radeon_ring_write(rdev, PACKET0(RADEON_GEN_INT_STATUS, 0)); - radeon_ring_write(rdev, RADEON_SW_INT_FIRE); + radeon_ring_write(ring, PACKET0(rdev->fence_drv[fence->ring].scratch_reg, 0)); + radeon_ring_write(ring, fence->seq); + radeon_ring_write(ring, PACKET0(RADEON_GEN_INT_STATUS, 0)); + radeon_ring_write(ring, RADEON_SW_INT_FIRE); +} + +void r100_semaphore_ring_emit(struct radeon_device *rdev, + struct radeon_ring *ring, + struct radeon_semaphore *semaphore, + bool emit_wait) +{ + /* Unused on older asics, since we don't have semaphores or multiple rings */ + BUG(); } int r100_copy_blit(struct radeon_device *rdev, @@ -836,6 +885,7 @@ int r100_copy_blit(struct radeon_device *rdev, unsigned num_gpu_pages, struct radeon_fence *fence) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; uint32_t cur_pages; uint32_t stride_bytes = RADEON_GPU_PAGE_SIZE; uint32_t pitch; @@ -853,7 +903,7 @@ int r100_copy_blit(struct radeon_device *rdev, /* Ask for enough room for blit + flush + fence */ ndw = 64 + (10 * num_loops); - r = radeon_ring_lock(rdev, ndw); + r = radeon_ring_lock(rdev, ring, ndw); if (r) { DRM_ERROR("radeon: moving bo (%d) asking for %u dw.\n", r, ndw); return -EINVAL; @@ -867,8 +917,8 @@ int r100_copy_blit(struct radeon_device *rdev, /* pages are in Y direction - height page width in X direction - width */ - radeon_ring_write(rdev, PACKET3(PACKET3_BITBLT_MULTI, 8)); - radeon_ring_write(rdev, + radeon_ring_write(ring, PACKET3(PACKET3_BITBLT_MULTI, 8)); + radeon_ring_write(ring, RADEON_GMC_SRC_PITCH_OFFSET_CNTL | RADEON_GMC_DST_PITCH_OFFSET_CNTL | RADEON_GMC_SRC_CLIPPING | @@ -880,26 +930,26 @@ int r100_copy_blit(struct radeon_device *rdev, RADEON_DP_SRC_SOURCE_MEMORY | RADEON_GMC_CLR_CMP_CNTL_DIS | RADEON_GMC_WR_MSK_DIS); - radeon_ring_write(rdev, (pitch << 22) | (src_offset >> 10)); - radeon_ring_write(rdev, (pitch << 22) | (dst_offset >> 10)); - radeon_ring_write(rdev, (0x1fff) | (0x1fff << 16)); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, (0x1fff) | (0x1fff << 16)); - radeon_ring_write(rdev, num_gpu_pages); - radeon_ring_write(rdev, num_gpu_pages); - radeon_ring_write(rdev, cur_pages | (stride_pixels << 16)); - } - radeon_ring_write(rdev, PACKET0(RADEON_DSTCACHE_CTLSTAT, 0)); - radeon_ring_write(rdev, RADEON_RB2D_DC_FLUSH_ALL); - radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0)); - radeon_ring_write(rdev, + radeon_ring_write(ring, (pitch << 22) | (src_offset >> 10)); + radeon_ring_write(ring, (pitch << 22) | (dst_offset >> 10)); + radeon_ring_write(ring, (0x1fff) | (0x1fff << 16)); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, (0x1fff) | (0x1fff << 16)); + radeon_ring_write(ring, num_gpu_pages); + radeon_ring_write(ring, num_gpu_pages); + radeon_ring_write(ring, cur_pages | (stride_pixels << 16)); + } + radeon_ring_write(ring, PACKET0(RADEON_DSTCACHE_CTLSTAT, 0)); + radeon_ring_write(ring, RADEON_RB2D_DC_FLUSH_ALL); + radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0)); + radeon_ring_write(ring, RADEON_WAIT_2D_IDLECLEAN | RADEON_WAIT_HOST_IDLECLEAN | RADEON_WAIT_DMA_GUI_IDLE); if (fence) { r = radeon_fence_emit(rdev, fence); } - radeon_ring_unlock_commit(rdev); + radeon_ring_unlock_commit(rdev, ring); return r; } @@ -918,21 +968,21 @@ static int r100_cp_wait_for_idle(struct radeon_device *rdev) return -1; } -void r100_ring_start(struct radeon_device *rdev) +void r100_ring_start(struct radeon_device *rdev, struct radeon_ring *ring) { int r; - r = radeon_ring_lock(rdev, 2); + r = radeon_ring_lock(rdev, ring, 2); if (r) { return; } - radeon_ring_write(rdev, PACKET0(RADEON_ISYNC_CNTL, 0)); - radeon_ring_write(rdev, + radeon_ring_write(ring, PACKET0(RADEON_ISYNC_CNTL, 0)); + radeon_ring_write(ring, RADEON_ISYNC_ANY2D_IDLE3D | RADEON_ISYNC_ANY3D_IDLE2D | RADEON_ISYNC_WAIT_IDLEGUI | RADEON_ISYNC_CPSCRATCH_IDLEGUI); - radeon_ring_unlock_commit(rdev); + radeon_ring_unlock_commit(rdev, ring); } @@ -1033,6 +1083,7 @@ static void r100_cp_load_microcode(struct radeon_device *rdev) int r100_cp_init(struct radeon_device *rdev, unsigned ring_size) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; unsigned rb_bufsz; unsigned rb_blksz; unsigned max_fetch; @@ -1058,7 +1109,9 @@ int r100_cp_init(struct radeon_device *rdev, unsigned ring_size) rb_bufsz = drm_order(ring_size / 8); ring_size = (1 << (rb_bufsz + 1)) * 4; r100_cp_load_microcode(rdev); - r = radeon_ring_init(rdev, ring_size); + r = radeon_ring_init(rdev, ring, ring_size, RADEON_WB_CP_RPTR_OFFSET, + RADEON_CP_RB_RPTR, RADEON_CP_RB_WPTR, + 0, 0x7fffff, RADEON_CP_PACKET2); if (r) { return r; } @@ -1067,7 +1120,7 @@ int r100_cp_init(struct radeon_device *rdev, unsigned ring_size) rb_blksz = 9; /* cp will read 128bytes at a time (4 dwords) */ max_fetch = 1; - rdev->cp.align_mask = 16 - 1; + ring->align_mask = 16 - 1; /* Write to CP_RB_WPTR will be delayed for pre_write_timer clocks */ pre_write_timer = 64; /* Force CP_RB_WPTR write if written more than one time before the @@ -1097,13 +1150,13 @@ int r100_cp_init(struct radeon_device *rdev, unsigned ring_size) WREG32(RADEON_CP_RB_CNTL, tmp | RADEON_RB_NO_UPDATE); /* Set ring address */ - DRM_INFO("radeon: ring at 0x%016lX\n", (unsigned long)rdev->cp.gpu_addr); - WREG32(RADEON_CP_RB_BASE, rdev->cp.gpu_addr); + DRM_INFO("radeon: ring at 0x%016lX\n", (unsigned long)ring->gpu_addr); + WREG32(RADEON_CP_RB_BASE, ring->gpu_addr); /* Force read & write ptr to 0 */ WREG32(RADEON_CP_RB_CNTL, tmp | RADEON_RB_RPTR_WR_ENA | RADEON_RB_NO_UPDATE); WREG32(RADEON_CP_RB_RPTR_WR, 0); - rdev->cp.wptr = 0; - WREG32(RADEON_CP_RB_WPTR, rdev->cp.wptr); + ring->wptr = 0; + WREG32(RADEON_CP_RB_WPTR, ring->wptr); /* set the wb address whether it's enabled or not */ WREG32(R_00070C_CP_RB_RPTR_ADDR, @@ -1119,7 +1172,7 @@ int r100_cp_init(struct radeon_device *rdev, unsigned ring_size) WREG32(RADEON_CP_RB_CNTL, tmp); udelay(10); - rdev->cp.rptr = RREG32(RADEON_CP_RB_RPTR); + ring->rptr = RREG32(RADEON_CP_RB_RPTR); /* Set cp mode to bus mastering & enable cp*/ WREG32(RADEON_CP_CSQ_MODE, REG_SET(RADEON_INDIRECT2_START, indirect2_start) | @@ -1127,13 +1180,13 @@ int r100_cp_init(struct radeon_device *rdev, unsigned ring_size) WREG32(RADEON_CP_RB_WPTR_DELAY, 0); WREG32(RADEON_CP_CSQ_MODE, 0x00004D4D); WREG32(RADEON_CP_CSQ_CNTL, RADEON_CSQ_PRIBM_INDBM); - radeon_ring_start(rdev); - r = radeon_ring_test(rdev); + radeon_ring_start(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); + r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, ring); if (r) { DRM_ERROR("radeon: cp isn't working (%d).\n", r); return r; } - rdev->cp.ready = true; + ring->ready = true; radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size); return 0; } @@ -1145,7 +1198,7 @@ void r100_cp_fini(struct radeon_device *rdev) } /* Disable ring */ r100_cp_disable(rdev); - radeon_ring_fini(rdev); + radeon_ring_fini(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); DRM_INFO("radeon: cp finalized\n"); } @@ -1153,7 +1206,7 @@ void r100_cp_disable(struct radeon_device *rdev) { /* Disable ring */ radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); - rdev->cp.ready = false; + rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false; WREG32(RADEON_CP_CSQ_MODE, 0); WREG32(RADEON_CP_CSQ_CNTL, 0); WREG32(R_000770_SCRATCH_UMSK, 0); @@ -1163,13 +1216,6 @@ void r100_cp_disable(struct radeon_device *rdev) } } -void r100_cp_commit(struct radeon_device *rdev) -{ - WREG32(RADEON_CP_RB_WPTR, rdev->cp.wptr); - (void)RREG32(RADEON_CP_RB_WPTR); -} - - /* * CS functions */ @@ -1543,7 +1589,17 @@ static int r100_packet0_check(struct radeon_cs_parser *p, r100_cs_dump_packet(p, pkt); return r; } - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { + if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + tile_flags |= RADEON_TXO_MACRO_TILE; + if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + tile_flags |= RADEON_TXO_MICRO_TILE_X2; + + tmp = idx_value & ~(0x7 << 2); + tmp |= tile_flags; + ib[idx] = tmp + ((u32)reloc->lobj.gpu_offset); + } else + ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); track->textures[i].robj = reloc->robj; track->tex_dirty = true; break; @@ -1614,15 +1670,17 @@ static int r100_packet0_check(struct radeon_cs_parser *p, r100_cs_dump_packet(p, pkt); return r; } - - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) - tile_flags |= RADEON_COLOR_TILE_ENABLE; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) - tile_flags |= RADEON_COLOR_MICROTILE_ENABLE; - - tmp = idx_value & ~(0x7 << 16); - tmp |= tile_flags; - ib[idx] = tmp; + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { + if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + tile_flags |= RADEON_COLOR_TILE_ENABLE; + if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + tile_flags |= RADEON_COLOR_MICROTILE_ENABLE; + + tmp = idx_value & ~(0x7 << 16); + tmp |= tile_flags; + ib[idx] = tmp; + } else + ib[idx] = idx_value; track->cb[0].pitch = idx_value & RADEON_COLORPITCH_MASK; track->cb_dirty = true; @@ -2097,9 +2155,9 @@ int r100_mc_wait_for_idle(struct radeon_device *rdev) return -1; } -void r100_gpu_lockup_update(struct r100_gpu_lockup *lockup, struct radeon_cp *cp) +void r100_gpu_lockup_update(struct r100_gpu_lockup *lockup, struct radeon_ring *ring) { - lockup->last_cp_rptr = cp->rptr; + lockup->last_cp_rptr = ring->rptr; lockup->last_jiffies = jiffies; } @@ -2124,20 +2182,20 @@ void r100_gpu_lockup_update(struct r100_gpu_lockup *lockup, struct radeon_cp *cp * false positive when CP is just gived nothing to do. * **/ -bool r100_gpu_cp_is_lockup(struct radeon_device *rdev, struct r100_gpu_lockup *lockup, struct radeon_cp *cp) +bool r100_gpu_cp_is_lockup(struct radeon_device *rdev, struct r100_gpu_lockup *lockup, struct radeon_ring *ring) { unsigned long cjiffies, elapsed; cjiffies = jiffies; if (!time_after(cjiffies, lockup->last_jiffies)) { /* likely a wrap around */ - lockup->last_cp_rptr = cp->rptr; + lockup->last_cp_rptr = ring->rptr; lockup->last_jiffies = jiffies; return false; } - if (cp->rptr != lockup->last_cp_rptr) { + if (ring->rptr != lockup->last_cp_rptr) { /* CP is still working no lockup */ - lockup->last_cp_rptr = cp->rptr; + lockup->last_cp_rptr = ring->rptr; lockup->last_jiffies = jiffies; return false; } @@ -2150,32 +2208,31 @@ bool r100_gpu_cp_is_lockup(struct radeon_device *rdev, struct r100_gpu_lockup *l return false; } -bool r100_gpu_is_lockup(struct radeon_device *rdev) +bool r100_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) { u32 rbbm_status; int r; rbbm_status = RREG32(R_000E40_RBBM_STATUS); if (!G_000E40_GUI_ACTIVE(rbbm_status)) { - r100_gpu_lockup_update(&rdev->config.r100.lockup, &rdev->cp); + r100_gpu_lockup_update(&rdev->config.r100.lockup, ring); return false; } /* force CP activities */ - r = radeon_ring_lock(rdev, 2); + r = radeon_ring_lock(rdev, ring, 2); if (!r) { /* PACKET2 NOP */ - radeon_ring_write(rdev, 0x80000000); - radeon_ring_write(rdev, 0x80000000); - radeon_ring_unlock_commit(rdev); + radeon_ring_write(ring, 0x80000000); + radeon_ring_write(ring, 0x80000000); + radeon_ring_unlock_commit(rdev, ring); } - rdev->cp.rptr = RREG32(RADEON_CP_RB_RPTR); - return r100_gpu_cp_is_lockup(rdev, &rdev->config.r100.lockup, &rdev->cp); + ring->rptr = RREG32(ring->rptr_reg); + return r100_gpu_cp_is_lockup(rdev, &rdev->config.r100.lockup, ring); } void r100_bm_disable(struct radeon_device *rdev) { u32 tmp; - u16 tmp16; /* disable bus mastering */ tmp = RREG32(R_000030_BUS_CNTL); @@ -2186,8 +2243,7 @@ void r100_bm_disable(struct radeon_device *rdev) WREG32(R_000030_BUS_CNTL, (tmp & 0xFFFFFFFF) | 0x00000040); tmp = RREG32(RADEON_BUS_CNTL); mdelay(1); - pci_read_config_word(rdev->pdev, 0x4, &tmp16); - pci_write_config_word(rdev->pdev, 0x4, tmp16 & 0xFFFB); + pci_clear_master(rdev->pdev); mdelay(1); } @@ -2497,7 +2553,7 @@ static void r100_pll_errata_after_data(struct radeon_device *rdev) * or the chip could hang on a subsequent access */ if (rdev->pll_errata & CHIP_ERRATA_PLL_DELAY) { - udelay(5000); + mdelay(5); } /* This function is required to workaround a hardware bug in some (all?) @@ -2578,21 +2634,22 @@ static int r100_debugfs_cp_ring_info(struct seq_file *m, void *data) struct drm_info_node *node = (struct drm_info_node *) m->private; struct drm_device *dev = node->minor->dev; struct radeon_device *rdev = dev->dev_private; + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; uint32_t rdp, wdp; unsigned count, i, j; - radeon_ring_free_size(rdev); + radeon_ring_free_size(rdev, ring); rdp = RREG32(RADEON_CP_RB_RPTR); wdp = RREG32(RADEON_CP_RB_WPTR); - count = (rdp + rdev->cp.ring_size - wdp) & rdev->cp.ptr_mask; + count = (rdp + ring->ring_size - wdp) & ring->ptr_mask; seq_printf(m, "CP_STAT 0x%08x\n", RREG32(RADEON_CP_STAT)); seq_printf(m, "CP_RB_WPTR 0x%08x\n", wdp); seq_printf(m, "CP_RB_RPTR 0x%08x\n", rdp); - seq_printf(m, "%u free dwords in ring\n", rdev->cp.ring_free_dw); + seq_printf(m, "%u free dwords in ring\n", ring->ring_free_dw); seq_printf(m, "%u dwords in ring\n", count); for (j = 0; j <= count; j++) { - i = (rdp + j) & rdev->cp.ptr_mask; - seq_printf(m, "r[%04d]=0x%08x\n", i, rdev->cp.ring[i]); + i = (rdp + j) & ring->ptr_mask; + seq_printf(m, "r[%04d]=0x%08x\n", i, ring->ring[i]); } return 0; } @@ -3634,7 +3691,7 @@ void r100_cs_track_clear(struct radeon_device *rdev, struct r100_cs_track *track } } -int r100_ring_test(struct radeon_device *rdev) +int r100_ring_test(struct radeon_device *rdev, struct radeon_ring *ring) { uint32_t scratch; uint32_t tmp = 0; @@ -3647,15 +3704,15 @@ int r100_ring_test(struct radeon_device *rdev) return r; } WREG32(scratch, 0xCAFEDEAD); - r = radeon_ring_lock(rdev, 2); + r = radeon_ring_lock(rdev, ring, 2); if (r) { DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r); radeon_scratch_free(rdev, scratch); return r; } - radeon_ring_write(rdev, PACKET0(scratch, 0)); - radeon_ring_write(rdev, 0xDEADBEEF); - radeon_ring_unlock_commit(rdev); + radeon_ring_write(ring, PACKET0(scratch, 0)); + radeon_ring_write(ring, 0xDEADBEEF); + radeon_ring_unlock_commit(rdev, ring); for (i = 0; i < rdev->usec_timeout; i++) { tmp = RREG32(scratch); if (tmp == 0xDEADBEEF) { @@ -3676,12 +3733,14 @@ int r100_ring_test(struct radeon_device *rdev) void r100_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) { - radeon_ring_write(rdev, PACKET0(RADEON_CP_IB_BASE, 1)); - radeon_ring_write(rdev, ib->gpu_addr); - radeon_ring_write(rdev, ib->length_dw); + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; + + radeon_ring_write(ring, PACKET0(RADEON_CP_IB_BASE, 1)); + radeon_ring_write(ring, ib->gpu_addr); + radeon_ring_write(ring, ib->length_dw); } -int r100_ib_test(struct radeon_device *rdev) +int r100_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) { struct radeon_ib *ib; uint32_t scratch; @@ -3695,7 +3754,7 @@ int r100_ib_test(struct radeon_device *rdev) return r; } WREG32(scratch, 0xCAFEDEAD); - r = radeon_ib_get(rdev, &ib); + r = radeon_ib_get(rdev, RADEON_RING_TYPE_GFX_INDEX, &ib, 256); if (r) { return r; } @@ -3739,34 +3798,16 @@ int r100_ib_test(struct radeon_device *rdev) void r100_ib_fini(struct radeon_device *rdev) { + radeon_ib_pool_suspend(rdev); radeon_ib_pool_fini(rdev); } -int r100_ib_init(struct radeon_device *rdev) -{ - int r; - - r = radeon_ib_pool_init(rdev); - if (r) { - dev_err(rdev->dev, "failed initializing IB pool (%d).\n", r); - r100_ib_fini(rdev); - return r; - } - r = r100_ib_test(rdev); - if (r) { - dev_err(rdev->dev, "failed testing IB (%d).\n", r); - r100_ib_fini(rdev); - return r; - } - return 0; -} - void r100_mc_stop(struct radeon_device *rdev, struct r100_mc_save *save) { /* Shutdown CP we shouldn't need to do that but better be safe than * sorry */ - rdev->cp.ready = false; + rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false; WREG32(R_000740_CP_CSQ_CNTL, 0); /* Save few CRTC registers */ @@ -3904,6 +3945,12 @@ static int r100_startup(struct radeon_device *rdev) if (r) return r; + r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r); + return r; + } + /* Enable IRQ */ r100_irq_set(rdev); rdev->config.r100.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL); @@ -3913,16 +3960,25 @@ static int r100_startup(struct radeon_device *rdev) dev_err(rdev->dev, "failed initializing CP (%d).\n", r); return r; } - r = r100_ib_init(rdev); + + r = radeon_ib_pool_start(rdev); + if (r) + return r; + + r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); if (r) { - dev_err(rdev->dev, "failed initializing IB (%d).\n", r); + dev_err(rdev->dev, "failed testing IB (%d).\n", r); + rdev->accel_working = false; return r; } + return 0; } int r100_resume(struct radeon_device *rdev) { + int r; + /* Make sur GART are not working */ if (rdev->flags & RADEON_IS_PCI) r100_pci_gart_disable(rdev); @@ -3940,11 +3996,18 @@ int r100_resume(struct radeon_device *rdev) r100_clock_startup(rdev); /* Initialize surface registers */ radeon_surface_init(rdev); - return r100_startup(rdev); + + rdev->accel_working = true; + r = r100_startup(rdev); + if (r) { + rdev->accel_working = false; + } + return r; } int r100_suspend(struct radeon_device *rdev) { + radeon_ib_pool_suspend(rdev); r100_cp_disable(rdev); radeon_wb_disable(rdev); r100_irq_disable(rdev); @@ -4063,7 +4126,14 @@ int r100_init(struct radeon_device *rdev) return r; } r100_set_safe_registers(rdev); + + r = radeon_ib_pool_init(rdev); rdev->accel_working = true; + if (r) { + dev_err(rdev->dev, "IB initialization failed (%d).\n", r); + rdev->accel_working = false; + } + r = r100_startup(rdev); if (r) { /* Somethings want wront with the accel init stop accel */ diff --git a/drivers/gpu/drm/radeon/r200.c b/drivers/gpu/drm/radeon/r200.c index a1f3ba0..a59cc47 100644 --- a/drivers/gpu/drm/radeon/r200.c +++ b/drivers/gpu/drm/radeon/r200.c @@ -87,6 +87,7 @@ int r200_copy_dma(struct radeon_device *rdev, unsigned num_gpu_pages, struct radeon_fence *fence) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; uint32_t size; uint32_t cur_size; int i, num_loops; @@ -95,33 +96,33 @@ int r200_copy_dma(struct radeon_device *rdev, /* radeon pitch is /64 */ size = num_gpu_pages << RADEON_GPU_PAGE_SHIFT; num_loops = DIV_ROUND_UP(size, 0x1FFFFF); - r = radeon_ring_lock(rdev, num_loops * 4 + 64); + r = radeon_ring_lock(rdev, ring, num_loops * 4 + 64); if (r) { DRM_ERROR("radeon: moving bo (%d).\n", r); return r; } /* Must wait for 2D idle & clean before DMA or hangs might happen */ - radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0)); - radeon_ring_write(rdev, (1 << 16)); + radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0)); + radeon_ring_write(ring, (1 << 16)); for (i = 0; i < num_loops; i++) { cur_size = size; if (cur_size > 0x1FFFFF) { cur_size = 0x1FFFFF; } size -= cur_size; - radeon_ring_write(rdev, PACKET0(0x720, 2)); - radeon_ring_write(rdev, src_offset); - radeon_ring_write(rdev, dst_offset); - radeon_ring_write(rdev, cur_size | (1 << 31) | (1 << 30)); + radeon_ring_write(ring, PACKET0(0x720, 2)); + radeon_ring_write(ring, src_offset); + radeon_ring_write(ring, dst_offset); + radeon_ring_write(ring, cur_size | (1 << 31) | (1 << 30)); src_offset += cur_size; dst_offset += cur_size; } - radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0)); - radeon_ring_write(rdev, RADEON_WAIT_DMA_GUI_IDLE); + radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0)); + radeon_ring_write(ring, RADEON_WAIT_DMA_GUI_IDLE); if (fence) { r = radeon_fence_emit(rdev, fence); } - radeon_ring_unlock_commit(rdev); + radeon_ring_unlock_commit(rdev, ring); return r; } @@ -214,7 +215,17 @@ int r200_packet0_check(struct radeon_cs_parser *p, r100_cs_dump_packet(p, pkt); return r; } - ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { + if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + tile_flags |= R200_TXO_MACRO_TILE; + if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + tile_flags |= R200_TXO_MICRO_TILE; + + tmp = idx_value & ~(0x7 << 2); + tmp |= tile_flags; + ib[idx] = tmp + ((u32)reloc->lobj.gpu_offset); + } else + ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); track->textures[i].robj = reloc->robj; track->tex_dirty = true; break; @@ -276,14 +287,17 @@ int r200_packet0_check(struct radeon_cs_parser *p, return r; } - if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) - tile_flags |= RADEON_COLOR_TILE_ENABLE; - if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) - tile_flags |= RADEON_COLOR_MICROTILE_ENABLE; + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { + if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) + tile_flags |= RADEON_COLOR_TILE_ENABLE; + if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) + tile_flags |= RADEON_COLOR_MICROTILE_ENABLE; - tmp = idx_value & ~(0x7 << 16); - tmp |= tile_flags; - ib[idx] = tmp; + tmp = idx_value & ~(0x7 << 16); + tmp |= tile_flags; + ib[idx] = tmp; + } else + ib[idx] = idx_value; track->cb[0].pitch = idx_value & RADEON_COLORPITCH_MASK; track->cb_dirty = true; diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c index c93bc64..fa14383 100644 --- a/drivers/gpu/drm/radeon/r300.c +++ b/drivers/gpu/drm/radeon/r300.c @@ -105,8 +105,8 @@ int rv370_pcie_gart_init(struct radeon_device *rdev) if (r) DRM_ERROR("Failed to register debugfs file for PCIE gart !\n"); rdev->gart.table_size = rdev->gart.num_gpu_pages * 4; - rdev->asic->gart_tlb_flush = &rv370_pcie_gart_tlb_flush; - rdev->asic->gart_set_page = &rv370_pcie_gart_set_page; + rdev->asic->gart.tlb_flush = &rv370_pcie_gart_tlb_flush; + rdev->asic->gart.set_page = &rv370_pcie_gart_set_page; return radeon_gart_table_vram_alloc(rdev); } @@ -175,36 +175,38 @@ void rv370_pcie_gart_fini(struct radeon_device *rdev) void r300_fence_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence) { + struct radeon_ring *ring = &rdev->ring[fence->ring]; + /* Who ever call radeon_fence_emit should call ring_lock and ask * for enough space (today caller are ib schedule and buffer move) */ /* Write SC register so SC & US assert idle */ - radeon_ring_write(rdev, PACKET0(R300_RE_SCISSORS_TL, 0)); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, PACKET0(R300_RE_SCISSORS_BR, 0)); - radeon_ring_write(rdev, 0); + radeon_ring_write(ring, PACKET0(R300_RE_SCISSORS_TL, 0)); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, PACKET0(R300_RE_SCISSORS_BR, 0)); + radeon_ring_write(ring, 0); /* Flush 3D cache */ - radeon_ring_write(rdev, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0)); - radeon_ring_write(rdev, R300_RB3D_DC_FLUSH); - radeon_ring_write(rdev, PACKET0(R300_RB3D_ZCACHE_CTLSTAT, 0)); - radeon_ring_write(rdev, R300_ZC_FLUSH); + radeon_ring_write(ring, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0)); + radeon_ring_write(ring, R300_RB3D_DC_FLUSH); + radeon_ring_write(ring, PACKET0(R300_RB3D_ZCACHE_CTLSTAT, 0)); + radeon_ring_write(ring, R300_ZC_FLUSH); /* Wait until IDLE & CLEAN */ - radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0)); - radeon_ring_write(rdev, (RADEON_WAIT_3D_IDLECLEAN | + radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0)); + radeon_ring_write(ring, (RADEON_WAIT_3D_IDLECLEAN | RADEON_WAIT_2D_IDLECLEAN | RADEON_WAIT_DMA_GUI_IDLE)); - radeon_ring_write(rdev, PACKET0(RADEON_HOST_PATH_CNTL, 0)); - radeon_ring_write(rdev, rdev->config.r300.hdp_cntl | + radeon_ring_write(ring, PACKET0(RADEON_HOST_PATH_CNTL, 0)); + radeon_ring_write(ring, rdev->config.r300.hdp_cntl | RADEON_HDP_READ_BUFFER_INVALIDATE); - radeon_ring_write(rdev, PACKET0(RADEON_HOST_PATH_CNTL, 0)); - radeon_ring_write(rdev, rdev->config.r300.hdp_cntl); + radeon_ring_write(ring, PACKET0(RADEON_HOST_PATH_CNTL, 0)); + radeon_ring_write(ring, rdev->config.r300.hdp_cntl); /* Emit fence sequence & fire IRQ */ - radeon_ring_write(rdev, PACKET0(rdev->fence_drv.scratch_reg, 0)); - radeon_ring_write(rdev, fence->seq); - radeon_ring_write(rdev, PACKET0(RADEON_GEN_INT_STATUS, 0)); - radeon_ring_write(rdev, RADEON_SW_INT_FIRE); + radeon_ring_write(ring, PACKET0(rdev->fence_drv[fence->ring].scratch_reg, 0)); + radeon_ring_write(ring, fence->seq); + radeon_ring_write(ring, PACKET0(RADEON_GEN_INT_STATUS, 0)); + radeon_ring_write(ring, RADEON_SW_INT_FIRE); } -void r300_ring_start(struct radeon_device *rdev) +void r300_ring_start(struct radeon_device *rdev, struct radeon_ring *ring) { unsigned gb_tile_config; int r; @@ -227,44 +229,44 @@ void r300_ring_start(struct radeon_device *rdev) break; } - r = radeon_ring_lock(rdev, 64); + r = radeon_ring_lock(rdev, ring, 64); if (r) { return; } - radeon_ring_write(rdev, PACKET0(RADEON_ISYNC_CNTL, 0)); - radeon_ring_write(rdev, + radeon_ring_write(ring, PACKET0(RADEON_ISYNC_CNTL, 0)); + radeon_ring_write(ring, RADEON_ISYNC_ANY2D_IDLE3D | RADEON_ISYNC_ANY3D_IDLE2D | RADEON_ISYNC_WAIT_IDLEGUI | RADEON_ISYNC_CPSCRATCH_IDLEGUI); - radeon_ring_write(rdev, PACKET0(R300_GB_TILE_CONFIG, 0)); - radeon_ring_write(rdev, gb_tile_config); - radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0)); - radeon_ring_write(rdev, + radeon_ring_write(ring, PACKET0(R300_GB_TILE_CONFIG, 0)); + radeon_ring_write(ring, gb_tile_config); + radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0)); + radeon_ring_write(ring, RADEON_WAIT_2D_IDLECLEAN | RADEON_WAIT_3D_IDLECLEAN); - radeon_ring_write(rdev, PACKET0(R300_DST_PIPE_CONFIG, 0)); - radeon_ring_write(rdev, R300_PIPE_AUTO_CONFIG); - radeon_ring_write(rdev, PACKET0(R300_GB_SELECT, 0)); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, PACKET0(R300_GB_ENABLE, 0)); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0)); - radeon_ring_write(rdev, R300_RB3D_DC_FLUSH | R300_RB3D_DC_FREE); - radeon_ring_write(rdev, PACKET0(R300_RB3D_ZCACHE_CTLSTAT, 0)); - radeon_ring_write(rdev, R300_ZC_FLUSH | R300_ZC_FREE); - radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0)); - radeon_ring_write(rdev, + radeon_ring_write(ring, PACKET0(R300_DST_PIPE_CONFIG, 0)); + radeon_ring_write(ring, R300_PIPE_AUTO_CONFIG); + radeon_ring_write(ring, PACKET0(R300_GB_SELECT, 0)); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, PACKET0(R300_GB_ENABLE, 0)); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0)); + radeon_ring_write(ring, R300_RB3D_DC_FLUSH | R300_RB3D_DC_FREE); + radeon_ring_write(ring, PACKET0(R300_RB3D_ZCACHE_CTLSTAT, 0)); + radeon_ring_write(ring, R300_ZC_FLUSH | R300_ZC_FREE); + radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0)); + radeon_ring_write(ring, RADEON_WAIT_2D_IDLECLEAN | RADEON_WAIT_3D_IDLECLEAN); - radeon_ring_write(rdev, PACKET0(R300_GB_AA_CONFIG, 0)); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0)); - radeon_ring_write(rdev, R300_RB3D_DC_FLUSH | R300_RB3D_DC_FREE); - radeon_ring_write(rdev, PACKET0(R300_RB3D_ZCACHE_CTLSTAT, 0)); - radeon_ring_write(rdev, R300_ZC_FLUSH | R300_ZC_FREE); - radeon_ring_write(rdev, PACKET0(R300_GB_MSPOS0, 0)); - radeon_ring_write(rdev, + radeon_ring_write(ring, PACKET0(R300_GB_AA_CONFIG, 0)); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0)); + radeon_ring_write(ring, R300_RB3D_DC_FLUSH | R300_RB3D_DC_FREE); + radeon_ring_write(ring, PACKET0(R300_RB3D_ZCACHE_CTLSTAT, 0)); + radeon_ring_write(ring, R300_ZC_FLUSH | R300_ZC_FREE); + radeon_ring_write(ring, PACKET0(R300_GB_MSPOS0, 0)); + radeon_ring_write(ring, ((6 << R300_MS_X0_SHIFT) | (6 << R300_MS_Y0_SHIFT) | (6 << R300_MS_X1_SHIFT) | @@ -273,8 +275,8 @@ void r300_ring_start(struct radeon_device *rdev) (6 << R300_MS_Y2_SHIFT) | (6 << R300_MSBD0_Y_SHIFT) | (6 << R300_MSBD0_X_SHIFT))); - radeon_ring_write(rdev, PACKET0(R300_GB_MSPOS1, 0)); - radeon_ring_write(rdev, + radeon_ring_write(ring, PACKET0(R300_GB_MSPOS1, 0)); + radeon_ring_write(ring, ((6 << R300_MS_X3_SHIFT) | (6 << R300_MS_Y3_SHIFT) | (6 << R300_MS_X4_SHIFT) | @@ -282,16 +284,16 @@ void r300_ring_start(struct radeon_device *rdev) (6 << R300_MS_X5_SHIFT) | (6 << R300_MS_Y5_SHIFT) | (6 << R300_MSBD1_SHIFT))); - radeon_ring_write(rdev, PACKET0(R300_GA_ENHANCE, 0)); - radeon_ring_write(rdev, R300_GA_DEADLOCK_CNTL | R300_GA_FASTSYNC_CNTL); - radeon_ring_write(rdev, PACKET0(R300_GA_POLY_MODE, 0)); - radeon_ring_write(rdev, + radeon_ring_write(ring, PACKET0(R300_GA_ENHANCE, 0)); + radeon_ring_write(ring, R300_GA_DEADLOCK_CNTL | R300_GA_FASTSYNC_CNTL); + radeon_ring_write(ring, PACKET0(R300_GA_POLY_MODE, 0)); + radeon_ring_write(ring, R300_FRONT_PTYPE_TRIANGE | R300_BACK_PTYPE_TRIANGE); - radeon_ring_write(rdev, PACKET0(R300_GA_ROUND_MODE, 0)); - radeon_ring_write(rdev, + radeon_ring_write(ring, PACKET0(R300_GA_ROUND_MODE, 0)); + radeon_ring_write(ring, R300_GEOMETRY_ROUND_NEAREST | R300_COLOR_ROUND_NEAREST); - radeon_ring_unlock_commit(rdev); + radeon_ring_unlock_commit(rdev, ring); } void r300_errata(struct radeon_device *rdev) @@ -375,26 +377,26 @@ void r300_gpu_init(struct radeon_device *rdev) rdev->num_gb_pipes, rdev->num_z_pipes); } -bool r300_gpu_is_lockup(struct radeon_device *rdev) +bool r300_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) { u32 rbbm_status; int r; rbbm_status = RREG32(R_000E40_RBBM_STATUS); if (!G_000E40_GUI_ACTIVE(rbbm_status)) { - r100_gpu_lockup_update(&rdev->config.r300.lockup, &rdev->cp); + r100_gpu_lockup_update(&rdev->config.r300.lockup, ring); return false; } /* force CP activities */ - r = radeon_ring_lock(rdev, 2); + r = radeon_ring_lock(rdev, ring, 2); if (!r) { /* PACKET2 NOP */ - radeon_ring_write(rdev, 0x80000000); - radeon_ring_write(rdev, 0x80000000); - radeon_ring_unlock_commit(rdev); + radeon_ring_write(ring, 0x80000000); + radeon_ring_write(ring, 0x80000000); + radeon_ring_unlock_commit(rdev, ring); } - rdev->cp.rptr = RREG32(RADEON_CP_RB_RPTR); - return r100_gpu_cp_is_lockup(rdev, &rdev->config.r300.lockup, &rdev->cp); + ring->rptr = RREG32(RADEON_CP_RB_RPTR); + return r100_gpu_cp_is_lockup(rdev, &rdev->config.r300.lockup, ring); } int r300_asic_reset(struct radeon_device *rdev) @@ -701,7 +703,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, return r; } - if (p->keep_tiling_flags) { + if (p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS) { ib[idx] = (idx_value & 31) | /* keep the 1st 5 bits */ ((idx_value & ~31) + (u32)reloc->lobj.gpu_offset); } else { @@ -765,7 +767,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, /* RB3D_COLORPITCH1 */ /* RB3D_COLORPITCH2 */ /* RB3D_COLORPITCH3 */ - if (!p->keep_tiling_flags) { + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { r = r100_cs_packet_next_reloc(p, &reloc); if (r) { DRM_ERROR("No reloc for ib[%d]=0x%04X\n", @@ -850,7 +852,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, break; case 0x4F24: /* ZB_DEPTHPITCH */ - if (!p->keep_tiling_flags) { + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { r = r100_cs_packet_next_reloc(p, &reloc); if (r) { DRM_ERROR("No reloc for ib[%d]=0x%04X\n", @@ -1396,6 +1398,12 @@ static int r300_startup(struct radeon_device *rdev) if (r) return r; + r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r); + return r; + } + /* Enable IRQ */ r100_irq_set(rdev); rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL); @@ -1405,16 +1413,25 @@ static int r300_startup(struct radeon_device *rdev) dev_err(rdev->dev, "failed initializing CP (%d).\n", r); return r; } - r = r100_ib_init(rdev); + + r = radeon_ib_pool_start(rdev); + if (r) + return r; + + r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); if (r) { - dev_err(rdev->dev, "failed initializing IB (%d).\n", r); + dev_err(rdev->dev, "failed testing IB (%d).\n", r); + rdev->accel_working = false; return r; } + return 0; } int r300_resume(struct radeon_device *rdev) { + int r; + /* Make sur GART are not working */ if (rdev->flags & RADEON_IS_PCIE) rv370_pcie_gart_disable(rdev); @@ -1434,11 +1451,18 @@ int r300_resume(struct radeon_device *rdev) r300_clock_startup(rdev); /* Initialize surface registers */ radeon_surface_init(rdev); - return r300_startup(rdev); + + rdev->accel_working = true; + r = r300_startup(rdev); + if (r) { + rdev->accel_working = false; + } + return r; } int r300_suspend(struct radeon_device *rdev) { + radeon_ib_pool_suspend(rdev); r100_cp_disable(rdev); radeon_wb_disable(rdev); r100_irq_disable(rdev); @@ -1539,7 +1563,14 @@ int r300_init(struct radeon_device *rdev) return r; } r300_set_reg_safe(rdev); + + r = radeon_ib_pool_init(rdev); rdev->accel_working = true; + if (r) { + dev_err(rdev->dev, "IB initialization failed (%d).\n", r); + rdev->accel_working = false; + } + r = r300_startup(rdev); if (r) { /* Somethings want wront with the accel init stop accel */ diff --git a/drivers/gpu/drm/radeon/r420.c b/drivers/gpu/drm/radeon/r420.c index 417fab8..f3fcaac 100644 --- a/drivers/gpu/drm/radeon/r420.c +++ b/drivers/gpu/drm/radeon/r420.c @@ -199,6 +199,8 @@ static void r420_clock_resume(struct radeon_device *rdev) static void r420_cp_errata_init(struct radeon_device *rdev) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; + /* RV410 and R420 can lock up if CP DMA to host memory happens * while the 2D engine is busy. * @@ -206,22 +208,24 @@ static void r420_cp_errata_init(struct radeon_device *rdev) * of the CP init, apparently. */ radeon_scratch_get(rdev, &rdev->config.r300.resync_scratch); - radeon_ring_lock(rdev, 8); - radeon_ring_write(rdev, PACKET0(R300_CP_RESYNC_ADDR, 1)); - radeon_ring_write(rdev, rdev->config.r300.resync_scratch); - radeon_ring_write(rdev, 0xDEADBEEF); - radeon_ring_unlock_commit(rdev); + radeon_ring_lock(rdev, ring, 8); + radeon_ring_write(ring, PACKET0(R300_CP_RESYNC_ADDR, 1)); + radeon_ring_write(ring, rdev->config.r300.resync_scratch); + radeon_ring_write(ring, 0xDEADBEEF); + radeon_ring_unlock_commit(rdev, ring); } static void r420_cp_errata_fini(struct radeon_device *rdev) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; + /* Catch the RESYNC we dispatched all the way back, * at the very beginning of the CP init. */ - radeon_ring_lock(rdev, 8); - radeon_ring_write(rdev, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0)); - radeon_ring_write(rdev, R300_RB3D_DC_FINISH); - radeon_ring_unlock_commit(rdev); + radeon_ring_lock(rdev, ring, 8); + radeon_ring_write(ring, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0)); + radeon_ring_write(ring, R300_RB3D_DC_FINISH); + radeon_ring_unlock_commit(rdev, ring); radeon_scratch_free(rdev, rdev->config.r300.resync_scratch); } @@ -254,6 +258,12 @@ static int r420_startup(struct radeon_device *rdev) if (r) return r; + r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r); + return r; + } + /* Enable IRQ */ r100_irq_set(rdev); rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL); @@ -264,16 +274,25 @@ static int r420_startup(struct radeon_device *rdev) return r; } r420_cp_errata_init(rdev); - r = r100_ib_init(rdev); + + r = radeon_ib_pool_start(rdev); + if (r) + return r; + + r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); if (r) { - dev_err(rdev->dev, "failed initializing IB (%d).\n", r); + dev_err(rdev->dev, "failed testing IB (%d).\n", r); + rdev->accel_working = false; return r; } + return 0; } int r420_resume(struct radeon_device *rdev) { + int r; + /* Make sur GART are not working */ if (rdev->flags & RADEON_IS_PCIE) rv370_pcie_gart_disable(rdev); @@ -297,11 +316,18 @@ int r420_resume(struct radeon_device *rdev) r420_clock_resume(rdev); /* Initialize surface registers */ radeon_surface_init(rdev); - return r420_startup(rdev); + + rdev->accel_working = true; + r = r420_startup(rdev); + if (r) { + rdev->accel_working = false; + } + return r; } int r420_suspend(struct radeon_device *rdev) { + radeon_ib_pool_suspend(rdev); r420_cp_errata_fini(rdev); r100_cp_disable(rdev); radeon_wb_disable(rdev); @@ -414,7 +440,14 @@ int r420_init(struct radeon_device *rdev) return r; } r420_set_reg_safe(rdev); + + r = radeon_ib_pool_init(rdev); rdev->accel_working = true; + if (r) { + dev_err(rdev->dev, "IB initialization failed (%d).\n", r); + rdev->accel_working = false; + } + r = r420_startup(rdev); if (r) { /* Somethings want wront with the accel init stop accel */ diff --git a/drivers/gpu/drm/radeon/r500_reg.h b/drivers/gpu/drm/radeon/r500_reg.h index fc43705..ec576aa 100644 --- a/drivers/gpu/drm/radeon/r500_reg.h +++ b/drivers/gpu/drm/radeon/r500_reg.h @@ -351,6 +351,8 @@ #define AVIVO_D1CRTC_BLANK_CONTROL 0x6084 #define AVIVO_D1CRTC_INTERLACE_CONTROL 0x6088 #define AVIVO_D1CRTC_INTERLACE_STATUS 0x608c +#define AVIVO_D1CRTC_STATUS 0x609c +# define AVIVO_D1CRTC_V_BLANK (1 << 0) #define AVIVO_D1CRTC_STATUS_POSITION 0x60a0 #define AVIVO_D1CRTC_FRAME_COUNT 0x60a4 #define AVIVO_D1CRTC_STEREO_CONTROL 0x60c4 @@ -573,6 +575,7 @@ #define AVIVO_TMDSA_CNTL 0x7880 # define AVIVO_TMDSA_CNTL_ENABLE (1 << 0) +# define AVIVO_TMDSA_CNTL_HDMI_EN (1 << 2) # define AVIVO_TMDSA_CNTL_HPD_MASK (1 << 4) # define AVIVO_TMDSA_CNTL_HPD_SELECT (1 << 8) # define AVIVO_TMDSA_CNTL_SYNC_PHASE (1 << 12) @@ -633,6 +636,7 @@ #define AVIVO_LVTMA_CNTL 0x7a80 # define AVIVO_LVTMA_CNTL_ENABLE (1 << 0) +# define AVIVO_LVTMA_CNTL_HDMI_EN (1 << 2) # define AVIVO_LVTMA_CNTL_HPD_MASK (1 << 4) # define AVIVO_LVTMA_CNTL_HPD_SELECT (1 << 8) # define AVIVO_LVTMA_CNTL_SYNC_PHASE (1 << 12) diff --git a/drivers/gpu/drm/radeon/r520.c b/drivers/gpu/drm/radeon/r520.c index 3081d07..ebcc15b 100644 --- a/drivers/gpu/drm/radeon/r520.c +++ b/drivers/gpu/drm/radeon/r520.c @@ -33,7 +33,7 @@ /* This files gather functions specifics to: r520,rv530,rv560,rv570,r580 */ -static int r520_mc_wait_for_idle(struct radeon_device *rdev) +int r520_mc_wait_for_idle(struct radeon_device *rdev) { unsigned i; uint32_t tmp; @@ -187,6 +187,12 @@ static int r520_startup(struct radeon_device *rdev) if (r) return r; + r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r); + return r; + } + /* Enable IRQ */ rs600_irq_set(rdev); rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL); @@ -196,9 +202,15 @@ static int r520_startup(struct radeon_device *rdev) dev_err(rdev->dev, "failed initializing CP (%d).\n", r); return r; } - r = r100_ib_init(rdev); + + r = radeon_ib_pool_start(rdev); + if (r) + return r; + + r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); if (r) { - dev_err(rdev->dev, "failed initializing IB (%d).\n", r); + dev_err(rdev->dev, "failed testing IB (%d).\n", r); + rdev->accel_working = false; return r; } return 0; @@ -206,6 +218,8 @@ static int r520_startup(struct radeon_device *rdev) int r520_resume(struct radeon_device *rdev) { + int r; + /* Make sur GART are not working */ if (rdev->flags & RADEON_IS_PCIE) rv370_pcie_gart_disable(rdev); @@ -223,7 +237,13 @@ int r520_resume(struct radeon_device *rdev) rv515_clock_startup(rdev); /* Initialize surface registers */ radeon_surface_init(rdev); - return r520_startup(rdev); + + rdev->accel_working = true; + r = r520_startup(rdev); + if (r) { + rdev->accel_working = false; + } + return r; } int r520_init(struct radeon_device *rdev) @@ -292,7 +312,14 @@ int r520_init(struct radeon_device *rdev) if (r) return r; rv515_set_safe_registers(rdev); + + r = radeon_ib_pool_init(rdev); rdev->accel_working = true; + if (r) { + dev_err(rdev->dev, "IB initialization failed (%d).\n", r); + rdev->accel_working = false; + } + r = r520_startup(rdev); if (r) { /* Somethings want wront with the accel init stop accel */ diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c index bdfa82a..b1ff9cc 100644 --- a/drivers/gpu/drm/radeon/r600.c +++ b/drivers/gpu/drm/radeon/r600.c @@ -49,6 +49,7 @@ #define EVERGREEN_PM4_UCODE_SIZE 1376 #define EVERGREEN_RLC_UCODE_SIZE 768 #define CAYMAN_RLC_UCODE_SIZE 1024 +#define ARUBA_RLC_UCODE_SIZE 1536 /* Firmware Names */ MODULE_FIRMWARE("radeon/R600_pfp.bin"); @@ -1134,7 +1135,7 @@ static void r600_vram_gtt_location(struct radeon_device *rdev, struct radeon_mc } if (rdev->flags & RADEON_IS_AGP) { size_bf = mc->gtt_start; - size_af = 0xFFFFFFFF - mc->gtt_end + 1; + size_af = 0xFFFFFFFF - mc->gtt_end; if (size_bf > size_af) { if (mc->mc_vram_size > size_bf) { dev_warn(rdev->dev, "limiting VRAM\n"); @@ -1148,7 +1149,7 @@ static void r600_vram_gtt_location(struct radeon_device *rdev, struct radeon_mc mc->real_vram_size = size_af; mc->mc_vram_size = size_af; } - mc->vram_start = mc->gtt_end; + mc->vram_start = mc->gtt_end + 1; } mc->vram_end = mc->vram_start + mc->mc_vram_size - 1; dev_info(rdev->dev, "VRAM: %lluM 0x%08llX - 0x%08llX (%lluM used)\n", @@ -1344,7 +1345,7 @@ int r600_gpu_soft_reset(struct radeon_device *rdev) return 0; } -bool r600_gpu_is_lockup(struct radeon_device *rdev) +bool r600_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) { u32 srbm_status; u32 grbm_status; @@ -1361,19 +1362,19 @@ bool r600_gpu_is_lockup(struct radeon_device *rdev) grbm_status = RREG32(R_008010_GRBM_STATUS); grbm_status2 = RREG32(R_008014_GRBM_STATUS2); if (!G_008010_GUI_ACTIVE(grbm_status)) { - r100_gpu_lockup_update(lockup, &rdev->cp); + r100_gpu_lockup_update(lockup, ring); return false; } /* force CP activities */ - r = radeon_ring_lock(rdev, 2); + r = radeon_ring_lock(rdev, ring, 2); if (!r) { /* PACKET2 NOP */ - radeon_ring_write(rdev, 0x80000000); - radeon_ring_write(rdev, 0x80000000); - radeon_ring_unlock_commit(rdev); + radeon_ring_write(ring, 0x80000000); + radeon_ring_write(ring, 0x80000000); + radeon_ring_unlock_commit(rdev, ring); } - rdev->cp.rptr = RREG32(R600_CP_RB_RPTR); - return r100_gpu_cp_is_lockup(rdev, lockup, &rdev->cp); + ring->rptr = RREG32(ring->rptr_reg); + return r100_gpu_cp_is_lockup(rdev, lockup, ring); } int r600_asic_reset(struct radeon_device *rdev) @@ -2145,27 +2146,28 @@ static int r600_cp_load_microcode(struct radeon_device *rdev) int r600_cp_start(struct radeon_device *rdev) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; int r; uint32_t cp_me; - r = radeon_ring_lock(rdev, 7); + r = radeon_ring_lock(rdev, ring, 7); if (r) { DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r); return r; } - radeon_ring_write(rdev, PACKET3(PACKET3_ME_INITIALIZE, 5)); - radeon_ring_write(rdev, 0x1); + radeon_ring_write(ring, PACKET3(PACKET3_ME_INITIALIZE, 5)); + radeon_ring_write(ring, 0x1); if (rdev->family >= CHIP_RV770) { - radeon_ring_write(rdev, 0x0); - radeon_ring_write(rdev, rdev->config.rv770.max_hw_contexts - 1); + radeon_ring_write(ring, 0x0); + radeon_ring_write(ring, rdev->config.rv770.max_hw_contexts - 1); } else { - radeon_ring_write(rdev, 0x3); - radeon_ring_write(rdev, rdev->config.r600.max_hw_contexts - 1); + radeon_ring_write(ring, 0x3); + radeon_ring_write(ring, rdev->config.r600.max_hw_contexts - 1); } - radeon_ring_write(rdev, PACKET3_ME_INITIALIZE_DEVICE_ID(1)); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, 0); - radeon_ring_unlock_commit(rdev); + radeon_ring_write(ring, PACKET3_ME_INITIALIZE_DEVICE_ID(1)); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 0); + radeon_ring_unlock_commit(rdev, ring); cp_me = 0xff; WREG32(R_0086D8_CP_ME_CNTL, cp_me); @@ -2174,6 +2176,7 @@ int r600_cp_start(struct radeon_device *rdev) int r600_cp_resume(struct radeon_device *rdev) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; u32 tmp; u32 rb_bufsz; int r; @@ -2185,13 +2188,13 @@ int r600_cp_resume(struct radeon_device *rdev) WREG32(GRBM_SOFT_RESET, 0); /* Set ring buffer size */ - rb_bufsz = drm_order(rdev->cp.ring_size / 8); + rb_bufsz = drm_order(ring->ring_size / 8); tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; #ifdef __BIG_ENDIAN tmp |= BUF_SWAP_32BIT; #endif WREG32(CP_RB_CNTL, tmp); - WREG32(CP_SEM_WAIT_TIMER, 0x4); + WREG32(CP_SEM_WAIT_TIMER, 0x0); /* Set the write pointer delay */ WREG32(CP_RB_WPTR_DELAY, 0); @@ -2199,8 +2202,8 @@ int r600_cp_resume(struct radeon_device *rdev) /* Initialize the ring buffer's read and write pointers */ WREG32(CP_RB_CNTL, tmp | RB_RPTR_WR_ENA); WREG32(CP_RB_RPTR_WR, 0); - rdev->cp.wptr = 0; - WREG32(CP_RB_WPTR, rdev->cp.wptr); + ring->wptr = 0; + WREG32(CP_RB_WPTR, ring->wptr); /* set the wb address whether it's enabled or not */ WREG32(CP_RB_RPTR_ADDR, @@ -2218,42 +2221,36 @@ int r600_cp_resume(struct radeon_device *rdev) mdelay(1); WREG32(CP_RB_CNTL, tmp); - WREG32(CP_RB_BASE, rdev->cp.gpu_addr >> 8); + WREG32(CP_RB_BASE, ring->gpu_addr >> 8); WREG32(CP_DEBUG, (1 << 27) | (1 << 28)); - rdev->cp.rptr = RREG32(CP_RB_RPTR); + ring->rptr = RREG32(CP_RB_RPTR); r600_cp_start(rdev); - rdev->cp.ready = true; - r = radeon_ring_test(rdev); + ring->ready = true; + r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, ring); if (r) { - rdev->cp.ready = false; + ring->ready = false; return r; } return 0; } -void r600_cp_commit(struct radeon_device *rdev) -{ - WREG32(CP_RB_WPTR, rdev->cp.wptr); - (void)RREG32(CP_RB_WPTR); -} - -void r600_ring_init(struct radeon_device *rdev, unsigned ring_size) +void r600_ring_init(struct radeon_device *rdev, struct radeon_ring *ring, unsigned ring_size) { u32 rb_bufsz; /* Align ring size */ rb_bufsz = drm_order(ring_size / 8); ring_size = (1 << (rb_bufsz + 1)) * 4; - rdev->cp.ring_size = ring_size; - rdev->cp.align_mask = 16 - 1; + ring->ring_size = ring_size; + ring->align_mask = 16 - 1; } void r600_cp_fini(struct radeon_device *rdev) { r600_cp_stop(rdev); - radeon_ring_fini(rdev); + radeon_ring_fini(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); } @@ -2272,11 +2269,11 @@ void r600_scratch_init(struct radeon_device *rdev) } } -int r600_ring_test(struct radeon_device *rdev) +int r600_ring_test(struct radeon_device *rdev, struct radeon_ring *ring) { uint32_t scratch; uint32_t tmp = 0; - unsigned i; + unsigned i, ridx = radeon_ring_index(rdev, ring); int r; r = radeon_scratch_get(rdev, &scratch); @@ -2285,16 +2282,16 @@ int r600_ring_test(struct radeon_device *rdev) return r; } WREG32(scratch, 0xCAFEDEAD); - r = radeon_ring_lock(rdev, 3); + r = radeon_ring_lock(rdev, ring, 3); if (r) { - DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r); + DRM_ERROR("radeon: cp failed to lock ring %d (%d).\n", ridx, r); radeon_scratch_free(rdev, scratch); return r; } - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1)); - radeon_ring_write(rdev, ((scratch - PACKET3_SET_CONFIG_REG_OFFSET) >> 2)); - radeon_ring_write(rdev, 0xDEADBEEF); - radeon_ring_unlock_commit(rdev); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + radeon_ring_write(ring, ((scratch - PACKET3_SET_CONFIG_REG_OFFSET) >> 2)); + radeon_ring_write(ring, 0xDEADBEEF); + radeon_ring_unlock_commit(rdev, ring); for (i = 0; i < rdev->usec_timeout; i++) { tmp = RREG32(scratch); if (tmp == 0xDEADBEEF) @@ -2302,10 +2299,10 @@ int r600_ring_test(struct radeon_device *rdev) DRM_UDELAY(1); } if (i < rdev->usec_timeout) { - DRM_INFO("ring test succeeded in %d usecs\n", i); + DRM_INFO("ring test on %d succeeded in %d usecs\n", ridx, i); } else { - DRM_ERROR("radeon: ring test failed (scratch(0x%04X)=0x%08X)\n", - scratch, tmp); + DRM_ERROR("radeon: ring %d test failed (scratch(0x%04X)=0x%08X)\n", + ridx, scratch, tmp); r = -EINVAL; } radeon_scratch_free(rdev, scratch); @@ -2315,49 +2312,66 @@ int r600_ring_test(struct radeon_device *rdev) void r600_fence_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence) { + struct radeon_ring *ring = &rdev->ring[fence->ring]; + if (rdev->wb.use_event) { - u64 addr = rdev->wb.gpu_addr + R600_WB_EVENT_OFFSET + - (u64)(rdev->fence_drv.scratch_reg - rdev->scratch.reg_base); + u64 addr = rdev->fence_drv[fence->ring].gpu_addr; /* flush read cache over gart */ - radeon_ring_write(rdev, PACKET3(PACKET3_SURFACE_SYNC, 3)); - radeon_ring_write(rdev, PACKET3_TC_ACTION_ENA | + radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); + radeon_ring_write(ring, PACKET3_TC_ACTION_ENA | PACKET3_VC_ACTION_ENA | PACKET3_SH_ACTION_ENA); - radeon_ring_write(rdev, 0xFFFFFFFF); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, 10); /* poll interval */ + radeon_ring_write(ring, 0xFFFFFFFF); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 10); /* poll interval */ /* EVENT_WRITE_EOP - flush caches, send int */ - radeon_ring_write(rdev, PACKET3(PACKET3_EVENT_WRITE_EOP, 4)); - radeon_ring_write(rdev, EVENT_TYPE(CACHE_FLUSH_AND_INV_EVENT_TS) | EVENT_INDEX(5)); - radeon_ring_write(rdev, addr & 0xffffffff); - radeon_ring_write(rdev, (upper_32_bits(addr) & 0xff) | DATA_SEL(1) | INT_SEL(2)); - radeon_ring_write(rdev, fence->seq); - radeon_ring_write(rdev, 0); + radeon_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE_EOP, 4)); + radeon_ring_write(ring, EVENT_TYPE(CACHE_FLUSH_AND_INV_EVENT_TS) | EVENT_INDEX(5)); + radeon_ring_write(ring, addr & 0xffffffff); + radeon_ring_write(ring, (upper_32_bits(addr) & 0xff) | DATA_SEL(1) | INT_SEL(2)); + radeon_ring_write(ring, fence->seq); + radeon_ring_write(ring, 0); } else { /* flush read cache over gart */ - radeon_ring_write(rdev, PACKET3(PACKET3_SURFACE_SYNC, 3)); - radeon_ring_write(rdev, PACKET3_TC_ACTION_ENA | + radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); + radeon_ring_write(ring, PACKET3_TC_ACTION_ENA | PACKET3_VC_ACTION_ENA | PACKET3_SH_ACTION_ENA); - radeon_ring_write(rdev, 0xFFFFFFFF); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, 10); /* poll interval */ - radeon_ring_write(rdev, PACKET3(PACKET3_EVENT_WRITE, 0)); - radeon_ring_write(rdev, EVENT_TYPE(CACHE_FLUSH_AND_INV_EVENT) | EVENT_INDEX(0)); + radeon_ring_write(ring, 0xFFFFFFFF); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 10); /* poll interval */ + radeon_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE, 0)); + radeon_ring_write(ring, EVENT_TYPE(CACHE_FLUSH_AND_INV_EVENT) | EVENT_INDEX(0)); /* wait for 3D idle clean */ - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1)); - radeon_ring_write(rdev, (WAIT_UNTIL - PACKET3_SET_CONFIG_REG_OFFSET) >> 2); - radeon_ring_write(rdev, WAIT_3D_IDLE_bit | WAIT_3D_IDLECLEAN_bit); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + radeon_ring_write(ring, (WAIT_UNTIL - PACKET3_SET_CONFIG_REG_OFFSET) >> 2); + radeon_ring_write(ring, WAIT_3D_IDLE_bit | WAIT_3D_IDLECLEAN_bit); /* Emit fence sequence & fire IRQ */ - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1)); - radeon_ring_write(rdev, ((rdev->fence_drv.scratch_reg - PACKET3_SET_CONFIG_REG_OFFSET) >> 2)); - radeon_ring_write(rdev, fence->seq); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + radeon_ring_write(ring, ((rdev->fence_drv[fence->ring].scratch_reg - PACKET3_SET_CONFIG_REG_OFFSET) >> 2)); + radeon_ring_write(ring, fence->seq); /* CP_INTERRUPT packet 3 no longer exists, use packet 0 */ - radeon_ring_write(rdev, PACKET0(CP_INT_STATUS, 0)); - radeon_ring_write(rdev, RB_INT_STAT); + radeon_ring_write(ring, PACKET0(CP_INT_STATUS, 0)); + radeon_ring_write(ring, RB_INT_STAT); } } +void r600_semaphore_ring_emit(struct radeon_device *rdev, + struct radeon_ring *ring, + struct radeon_semaphore *semaphore, + bool emit_wait) +{ + uint64_t addr = semaphore->gpu_addr; + unsigned sel = emit_wait ? PACKET3_SEM_SEL_WAIT : PACKET3_SEM_SEL_SIGNAL; + + if (rdev->family < CHIP_CAYMAN) + sel |= PACKET3_SEM_WAIT_ON_SIGNAL; + + radeon_ring_write(ring, PACKET3(PACKET3_MEM_SEMAPHORE, 1)); + radeon_ring_write(ring, addr & 0xffffffff); + radeon_ring_write(ring, (upper_32_bits(addr) & 0xff) | sel); +} + int r600_copy_blit(struct radeon_device *rdev, uint64_t src_offset, uint64_t dst_offset, @@ -2410,6 +2424,7 @@ void r600_clear_surface_reg(struct radeon_device *rdev, int reg) int r600_startup(struct radeon_device *rdev) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; int r; /* enable pcie gen2 link */ @@ -2439,7 +2454,7 @@ int r600_startup(struct radeon_device *rdev) r = r600_blit_init(rdev); if (r) { r600_blit_fini(rdev); - rdev->asic->copy = NULL; + rdev->asic->copy.copy = NULL; dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r); } @@ -2448,6 +2463,12 @@ int r600_startup(struct radeon_device *rdev) if (r) return r; + r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r); + return r; + } + /* Enable IRQ */ r = r600_irq_init(rdev); if (r) { @@ -2457,7 +2478,10 @@ int r600_startup(struct radeon_device *rdev) } r600_irq_set(rdev); - r = radeon_ring_init(rdev, rdev->cp.ring_size); + r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET, + R600_CP_RB_RPTR, R600_CP_RB_WPTR, + 0, 0xfffff, RADEON_CP_PACKET2); + if (r) return r; r = r600_cp_load_microcode(rdev); @@ -2467,6 +2491,17 @@ int r600_startup(struct radeon_device *rdev) if (r) return r; + r = radeon_ib_pool_start(rdev); + if (r) + return r; + + r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); + if (r) { + DRM_ERROR("radeon: failed testing IB (%d).\n", r); + rdev->accel_working = false; + return r; + } + return 0; } @@ -2495,15 +2530,11 @@ int r600_resume(struct radeon_device *rdev) /* post card */ atom_asic_init(rdev->mode_info.atom_context); + rdev->accel_working = true; r = r600_startup(rdev); if (r) { DRM_ERROR("r600 startup failed on resume\n"); - return r; - } - - r = r600_ib_test(rdev); - if (r) { - DRM_ERROR("radeon: failed testing IB (%d).\n", r); + rdev->accel_working = false; return r; } @@ -2519,13 +2550,14 @@ int r600_resume(struct radeon_device *rdev) int r600_suspend(struct radeon_device *rdev) { r600_audio_fini(rdev); + radeon_ib_pool_suspend(rdev); + r600_blit_suspend(rdev); /* FIXME: we should wait for ring to be empty */ r600_cp_stop(rdev); - rdev->cp.ready = false; + rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false; r600_irq_suspend(rdev); radeon_wb_disable(rdev); r600_pcie_gart_disable(rdev); - r600_blit_suspend(rdev); return 0; } @@ -2596,8 +2628,8 @@ int r600_init(struct radeon_device *rdev) if (r) return r; - rdev->cp.ring_obj = NULL; - r600_ring_init(rdev, 1024 * 1024); + rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ring_obj = NULL; + r600_ring_init(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX], 1024 * 1024); rdev->ih.ring_obj = NULL; r600_ih_ring_init(rdev, 64 * 1024); @@ -2606,30 +2638,24 @@ int r600_init(struct radeon_device *rdev) if (r) return r; + r = radeon_ib_pool_init(rdev); rdev->accel_working = true; + if (r) { + dev_err(rdev->dev, "IB initialization failed (%d).\n", r); + rdev->accel_working = false; + } + r = r600_startup(rdev); if (r) { dev_err(rdev->dev, "disabling GPU acceleration\n"); r600_cp_fini(rdev); r600_irq_fini(rdev); radeon_wb_fini(rdev); + r100_ib_fini(rdev); radeon_irq_kms_fini(rdev); r600_pcie_gart_fini(rdev); rdev->accel_working = false; } - if (rdev->accel_working) { - r = radeon_ib_pool_init(rdev); - if (r) { - dev_err(rdev->dev, "IB initialization failed (%d).\n", r); - rdev->accel_working = false; - } else { - r = r600_ib_test(rdev); - if (r) { - dev_err(rdev->dev, "IB test failed (%d).\n", r); - rdev->accel_working = false; - } - } - } r = r600_audio_init(rdev); if (r) @@ -2644,12 +2670,13 @@ void r600_fini(struct radeon_device *rdev) r600_cp_fini(rdev); r600_irq_fini(rdev); radeon_wb_fini(rdev); - radeon_ib_pool_fini(rdev); + r100_ib_fini(rdev); radeon_irq_kms_fini(rdev); r600_pcie_gart_fini(rdev); r600_vram_scratch_fini(rdev); radeon_agp_fini(rdev); radeon_gem_fini(rdev); + radeon_semaphore_driver_fini(rdev); radeon_fence_driver_fini(rdev); radeon_bo_fini(rdev); radeon_atombios_fini(rdev); @@ -2663,24 +2690,27 @@ void r600_fini(struct radeon_device *rdev) */ void r600_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) { + struct radeon_ring *ring = &rdev->ring[ib->fence->ring]; + /* FIXME: implement */ - radeon_ring_write(rdev, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); - radeon_ring_write(rdev, + radeon_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); + radeon_ring_write(ring, #ifdef __BIG_ENDIAN (2 << 0) | #endif (ib->gpu_addr & 0xFFFFFFFC)); - radeon_ring_write(rdev, upper_32_bits(ib->gpu_addr) & 0xFF); - radeon_ring_write(rdev, ib->length_dw); + radeon_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xFF); + radeon_ring_write(ring, ib->length_dw); } -int r600_ib_test(struct radeon_device *rdev) +int r600_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) { struct radeon_ib *ib; uint32_t scratch; uint32_t tmp = 0; unsigned i; int r; + int ring_index = radeon_ring_index(rdev, ring); r = radeon_scratch_get(rdev, &scratch); if (r) { @@ -2688,7 +2718,7 @@ int r600_ib_test(struct radeon_device *rdev) return r; } WREG32(scratch, 0xCAFEDEAD); - r = radeon_ib_get(rdev, &ib); + r = radeon_ib_get(rdev, ring_index, &ib, 256); if (r) { DRM_ERROR("radeon: failed to get ib (%d).\n", r); return r; @@ -2696,20 +2726,7 @@ int r600_ib_test(struct radeon_device *rdev) ib->ptr[0] = PACKET3(PACKET3_SET_CONFIG_REG, 1); ib->ptr[1] = ((scratch - PACKET3_SET_CONFIG_REG_OFFSET) >> 2); ib->ptr[2] = 0xDEADBEEF; - ib->ptr[3] = PACKET2(0); - ib->ptr[4] = PACKET2(0); - ib->ptr[5] = PACKET2(0); - ib->ptr[6] = PACKET2(0); - ib->ptr[7] = PACKET2(0); - ib->ptr[8] = PACKET2(0); - ib->ptr[9] = PACKET2(0); - ib->ptr[10] = PACKET2(0); - ib->ptr[11] = PACKET2(0); - ib->ptr[12] = PACKET2(0); - ib->ptr[13] = PACKET2(0); - ib->ptr[14] = PACKET2(0); - ib->ptr[15] = PACKET2(0); - ib->length_dw = 16; + ib->length_dw = 3; r = radeon_ib_schedule(rdev, ib); if (r) { radeon_scratch_free(rdev, scratch); @@ -2729,7 +2746,7 @@ int r600_ib_test(struct radeon_device *rdev) DRM_UDELAY(1); } if (i < rdev->usec_timeout) { - DRM_INFO("ib test succeeded in %u usecs\n", i); + DRM_INFO("ib test on ring %d succeeded in %u usecs\n", ib->fence->ring, i); } else { DRM_ERROR("radeon: ib test failed (scratch(0x%04X)=0x%08X)\n", scratch, tmp); @@ -2763,7 +2780,7 @@ void r600_ih_ring_init(struct radeon_device *rdev, unsigned ring_size) rdev->ih.rptr = 0; } -static int r600_ih_ring_alloc(struct radeon_device *rdev) +int r600_ih_ring_alloc(struct radeon_device *rdev) { int r; @@ -2799,7 +2816,7 @@ static int r600_ih_ring_alloc(struct radeon_device *rdev) return 0; } -static void r600_ih_ring_fini(struct radeon_device *rdev) +void r600_ih_ring_fini(struct radeon_device *rdev) { int r; if (rdev->ih.ring_obj) { @@ -2823,7 +2840,7 @@ void r600_rlc_stop(struct radeon_device *rdev) /* r7xx asics need to soft reset RLC before halting */ WREG32(SRBM_SOFT_RESET, SOFT_RESET_RLC); RREG32(SRBM_SOFT_RESET); - udelay(15000); + mdelay(15); WREG32(SRBM_SOFT_RESET, 0); RREG32(SRBM_SOFT_RESET); } @@ -2846,10 +2863,17 @@ static int r600_rlc_init(struct radeon_device *rdev) r600_rlc_stop(rdev); - WREG32(RLC_HB_BASE, 0); WREG32(RLC_HB_CNTL, 0); - WREG32(RLC_HB_RPTR, 0); - WREG32(RLC_HB_WPTR, 0); + + if (rdev->family == CHIP_ARUBA) { + WREG32(TN_RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8); + WREG32(TN_RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8); + } + if (rdev->family <= CHIP_CAYMAN) { + WREG32(RLC_HB_BASE, 0); + WREG32(RLC_HB_RPTR, 0); + WREG32(RLC_HB_WPTR, 0); + } if (rdev->family <= CHIP_CAICOS) { WREG32(RLC_HB_WPTR_LSB_ADDR, 0); WREG32(RLC_HB_WPTR_MSB_ADDR, 0); @@ -2858,7 +2882,12 @@ static int r600_rlc_init(struct radeon_device *rdev) WREG32(RLC_UCODE_CNTL, 0); fw_data = (const __be32 *)rdev->rlc_fw->data; - if (rdev->family >= CHIP_CAYMAN) { + if (rdev->family >= CHIP_ARUBA) { + for (i = 0; i < ARUBA_RLC_UCODE_SIZE; i++) { + WREG32(RLC_UCODE_ADDR, i); + WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++)); + } + } else if (rdev->family >= CHIP_CAYMAN) { for (i = 0; i < CAYMAN_RLC_UCODE_SIZE; i++) { WREG32(RLC_UCODE_ADDR, i); WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++)); @@ -3076,7 +3105,7 @@ int r600_irq_set(struct radeon_device *rdev) hpd3 = RREG32(DC_HOT_PLUG_DETECT3_INT_CONTROL) & ~DC_HPDx_INT_EN; } - if (rdev->irq.sw_int) { + if (rdev->irq.sw_int[RADEON_RING_TYPE_GFX_INDEX]) { DRM_DEBUG("r600_irq_set: sw int\n"); cp_int_cntl |= RB_INT_ENABLE; cp_int_cntl |= TIME_STAMP_INT_ENABLE; @@ -3460,11 +3489,11 @@ restart_ih: case 177: /* CP_INT in IB1 */ case 178: /* CP_INT in IB2 */ DRM_DEBUG("IH: CP int: 0x%08x\n", src_data); - radeon_fence_process(rdev); + radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX); break; case 181: /* CP EOP event */ DRM_DEBUG("IH: CP EOP\n"); - radeon_fence_process(rdev); + radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX); break; case 233: /* GUI IDLE */ DRM_DEBUG("IH: GUI idle\n"); @@ -3497,30 +3526,6 @@ restart_ih: */ #if defined(CONFIG_DEBUG_FS) -static int r600_debugfs_cp_ring_info(struct seq_file *m, void *data) -{ - struct drm_info_node *node = (struct drm_info_node *) m->private; - struct drm_device *dev = node->minor->dev; - struct radeon_device *rdev = dev->dev_private; - unsigned count, i, j; - - radeon_ring_free_size(rdev); - count = (rdev->cp.ring_size / 4) - rdev->cp.ring_free_dw; - seq_printf(m, "CP_STAT 0x%08x\n", RREG32(CP_STAT)); - seq_printf(m, "CP_RB_WPTR 0x%08x\n", RREG32(CP_RB_WPTR)); - seq_printf(m, "CP_RB_RPTR 0x%08x\n", RREG32(CP_RB_RPTR)); - seq_printf(m, "driver's copy of the CP_RB_WPTR 0x%08x\n", rdev->cp.wptr); - seq_printf(m, "driver's copy of the CP_RB_RPTR 0x%08x\n", rdev->cp.rptr); - seq_printf(m, "%u free dwords in ring\n", rdev->cp.ring_free_dw); - seq_printf(m, "%u dwords in ring\n", count); - i = rdev->cp.rptr; - for (j = 0; j <= count; j++) { - seq_printf(m, "r[%04d]=0x%08x\n", i, rdev->cp.ring[i]); - i = (i + 1) & rdev->cp.ptr_mask; - } - return 0; -} - static int r600_debugfs_mc_info(struct seq_file *m, void *data) { struct drm_info_node *node = (struct drm_info_node *) m->private; @@ -3534,7 +3539,6 @@ static int r600_debugfs_mc_info(struct seq_file *m, void *data) static struct drm_info_list r600_mc_info_list[] = { {"r600_mc_info", r600_debugfs_mc_info, 0, NULL}, - {"r600_ring_info", r600_debugfs_cp_ring_info, 0, NULL}, }; #endif diff --git a/drivers/gpu/drm/radeon/r600_audio.c b/drivers/gpu/drm/radeon/r600_audio.c index 846fae5..24e3939 100644 --- a/drivers/gpu/drm/radeon/r600_audio.c +++ b/drivers/gpu/drm/radeon/r600_audio.c @@ -36,7 +36,7 @@ */ static int r600_audio_chipset_supported(struct radeon_device *rdev) { - return (rdev->family >= CHIP_R600 && rdev->family < CHIP_CEDAR) + return (rdev->family >= CHIP_R600 && !ASIC_IS_DCE5(rdev)) || rdev->family == CHIP_RS600 || rdev->family == CHIP_RS690 || rdev->family == CHIP_RS740; @@ -161,8 +161,18 @@ static void r600_audio_update_hdmi(unsigned long param) */ static void r600_audio_engine_enable(struct radeon_device *rdev, bool enable) { + u32 value = 0; DRM_INFO("%s audio support\n", enable ? "Enabling" : "Disabling"); - WREG32_P(R600_AUDIO_ENABLE, enable ? 0x81000000 : 0x0, ~0x81000000); + if (ASIC_IS_DCE4(rdev)) { + if (enable) { + value |= 0x81000000; /* Required to enable audio */ + value |= 0x0e1000f0; /* fglrx sets that too */ + } + WREG32(EVERGREEN_AUDIO_ENABLE, value); + } else { + WREG32_P(R600_AUDIO_ENABLE, + enable ? 0x81000000 : 0x0, ~0x81000000); + } rdev->audio_enabled = enable; } @@ -229,6 +239,7 @@ void r600_audio_set_clock(struct drm_encoder *encoder, int clock) struct radeon_device *rdev = dev->dev_private; struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; + struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc); int base_rate = 48000; switch (radeon_encoder->encoder_id) { @@ -248,22 +259,33 @@ void r600_audio_set_clock(struct drm_encoder *encoder, int clock) return; } - switch (dig->dig_encoder) { - case 0: - WREG32(R600_AUDIO_PLL1_MUL, base_rate * 50); - WREG32(R600_AUDIO_PLL1_DIV, clock * 100); - WREG32(R600_AUDIO_CLK_SRCSEL, 0); - break; - - case 1: - WREG32(R600_AUDIO_PLL2_MUL, base_rate * 50); - WREG32(R600_AUDIO_PLL2_DIV, clock * 100); - WREG32(R600_AUDIO_CLK_SRCSEL, 1); - break; - default: - dev_err(rdev->dev, "Unsupported DIG on encoder 0x%02X\n", - radeon_encoder->encoder_id); - return; + if (ASIC_IS_DCE4(rdev)) { + /* TODO: other PLLs? */ + WREG32(EVERGREEN_AUDIO_PLL1_MUL, base_rate * 10); + WREG32(EVERGREEN_AUDIO_PLL1_DIV, clock * 10); + WREG32(EVERGREEN_AUDIO_PLL1_UNK, 0x00000071); + + /* Select DTO source */ + WREG32(0x5ac, radeon_crtc->crtc_id); + } else { + switch (dig->dig_encoder) { + case 0: + WREG32(R600_AUDIO_PLL1_MUL, base_rate * 50); + WREG32(R600_AUDIO_PLL1_DIV, clock * 100); + WREG32(R600_AUDIO_CLK_SRCSEL, 0); + break; + + case 1: + WREG32(R600_AUDIO_PLL2_MUL, base_rate * 50); + WREG32(R600_AUDIO_PLL2_DIV, clock * 100); + WREG32(R600_AUDIO_CLK_SRCSEL, 1); + break; + default: + dev_err(rdev->dev, + "Unsupported DIG on encoder 0x%02X\n", + radeon_encoder->encoder_id); + return; + } } } diff --git a/drivers/gpu/drm/radeon/r600_blit_kms.c b/drivers/gpu/drm/radeon/r600_blit_kms.c index e09d281..db38f58 100644 --- a/drivers/gpu/drm/radeon/r600_blit_kms.c +++ b/drivers/gpu/drm/radeon/r600_blit_kms.c @@ -30,26 +30,14 @@ #include "r600d.h" #include "r600_blit_shaders.h" - -#define DI_PT_RECTLIST 0x11 -#define DI_INDEX_SIZE_16_BIT 0x0 -#define DI_SRC_SEL_AUTO_INDEX 0x2 - -#define FMT_8 0x1 -#define FMT_5_6_5 0x8 -#define FMT_8_8_8_8 0x1a -#define COLOR_8 0x1 -#define COLOR_5_6_5 0x8 -#define COLOR_8_8_8_8 0x1a - -#define RECT_UNIT_H 32 -#define RECT_UNIT_W (RADEON_GPU_PAGE_SIZE / 4 / RECT_UNIT_H) +#include "radeon_blit_common.h" /* emits 21 on rv770+, 23 on r600 */ static void set_render_target(struct radeon_device *rdev, int format, int w, int h, u64 gpu_addr) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; u32 cb_color_info; int pitch, slice; @@ -63,38 +51,38 @@ set_render_target(struct radeon_device *rdev, int format, pitch = (w / 8) - 1; slice = ((w * h) / 64) - 1; - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(rdev, (CB_COLOR0_BASE - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(rdev, gpu_addr >> 8); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); + radeon_ring_write(ring, (CB_COLOR0_BASE - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); + radeon_ring_write(ring, gpu_addr >> 8); if (rdev->family > CHIP_R600 && rdev->family < CHIP_RV770) { - radeon_ring_write(rdev, PACKET3(PACKET3_SURFACE_BASE_UPDATE, 0)); - radeon_ring_write(rdev, 2 << 0); + radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_BASE_UPDATE, 0)); + radeon_ring_write(ring, 2 << 0); } - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(rdev, (CB_COLOR0_SIZE - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(rdev, (pitch << 0) | (slice << 10)); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); + radeon_ring_write(ring, (CB_COLOR0_SIZE - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); + radeon_ring_write(ring, (pitch << 0) | (slice << 10)); - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(rdev, (CB_COLOR0_VIEW - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(rdev, 0); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); + radeon_ring_write(ring, (CB_COLOR0_VIEW - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); + radeon_ring_write(ring, 0); - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(rdev, (CB_COLOR0_INFO - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(rdev, cb_color_info); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); + radeon_ring_write(ring, (CB_COLOR0_INFO - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); + radeon_ring_write(ring, cb_color_info); - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(rdev, (CB_COLOR0_TILE - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(rdev, 0); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); + radeon_ring_write(ring, (CB_COLOR0_TILE - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); + radeon_ring_write(ring, 0); - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(rdev, (CB_COLOR0_FRAG - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(rdev, 0); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); + radeon_ring_write(ring, (CB_COLOR0_FRAG - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); + radeon_ring_write(ring, 0); - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(rdev, (CB_COLOR0_MASK - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(rdev, 0); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); + radeon_ring_write(ring, (CB_COLOR0_MASK - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); + radeon_ring_write(ring, 0); } /* emits 5dw */ @@ -103,6 +91,7 @@ cp_set_surface_sync(struct radeon_device *rdev, u32 sync_type, u32 size, u64 mc_addr) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; u32 cp_coher_size; if (size == 0xffffffff) @@ -110,17 +99,18 @@ cp_set_surface_sync(struct radeon_device *rdev, else cp_coher_size = ((size + 255) >> 8); - radeon_ring_write(rdev, PACKET3(PACKET3_SURFACE_SYNC, 3)); - radeon_ring_write(rdev, sync_type); - radeon_ring_write(rdev, cp_coher_size); - radeon_ring_write(rdev, mc_addr >> 8); - radeon_ring_write(rdev, 10); /* poll interval */ + radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); + radeon_ring_write(ring, sync_type); + radeon_ring_write(ring, cp_coher_size); + radeon_ring_write(ring, mc_addr >> 8); + radeon_ring_write(ring, 10); /* poll interval */ } /* emits 21dw + 1 surface sync = 26dw */ static void set_shaders(struct radeon_device *rdev) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; u64 gpu_addr; u32 sq_pgm_resources; @@ -129,35 +119,35 @@ set_shaders(struct radeon_device *rdev) /* VS */ gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.vs_offset; - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(rdev, (SQ_PGM_START_VS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(rdev, gpu_addr >> 8); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); + radeon_ring_write(ring, (SQ_PGM_START_VS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); + radeon_ring_write(ring, gpu_addr >> 8); - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(rdev, (SQ_PGM_RESOURCES_VS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(rdev, sq_pgm_resources); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); + radeon_ring_write(ring, (SQ_PGM_RESOURCES_VS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); + radeon_ring_write(ring, sq_pgm_resources); - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(rdev, (SQ_PGM_CF_OFFSET_VS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(rdev, 0); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); + radeon_ring_write(ring, (SQ_PGM_CF_OFFSET_VS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); + radeon_ring_write(ring, 0); /* PS */ gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.ps_offset; - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(rdev, (SQ_PGM_START_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(rdev, gpu_addr >> 8); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); + radeon_ring_write(ring, (SQ_PGM_START_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); + radeon_ring_write(ring, gpu_addr >> 8); - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(rdev, (SQ_PGM_RESOURCES_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(rdev, sq_pgm_resources | (1 << 28)); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); + radeon_ring_write(ring, (SQ_PGM_RESOURCES_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); + radeon_ring_write(ring, sq_pgm_resources | (1 << 28)); - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(rdev, (SQ_PGM_EXPORTS_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(rdev, 2); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); + radeon_ring_write(ring, (SQ_PGM_EXPORTS_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); + radeon_ring_write(ring, 2); - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); - radeon_ring_write(rdev, (SQ_PGM_CF_OFFSET_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(rdev, 0); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1)); + radeon_ring_write(ring, (SQ_PGM_CF_OFFSET_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); + radeon_ring_write(ring, 0); gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.vs_offset; cp_set_surface_sync(rdev, PACKET3_SH_ACTION_ENA, 512, gpu_addr); @@ -167,6 +157,7 @@ set_shaders(struct radeon_device *rdev) static void set_vtx_resource(struct radeon_device *rdev, u64 gpu_addr) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; u32 sq_vtx_constant_word2; sq_vtx_constant_word2 = SQ_VTXC_BASE_ADDR_HI(upper_32_bits(gpu_addr) & 0xff) | @@ -175,15 +166,15 @@ set_vtx_resource(struct radeon_device *rdev, u64 gpu_addr) sq_vtx_constant_word2 |= SQ_VTXC_ENDIAN_SWAP(SQ_ENDIAN_8IN32); #endif - radeon_ring_write(rdev, PACKET3(PACKET3_SET_RESOURCE, 7)); - radeon_ring_write(rdev, 0x460); - radeon_ring_write(rdev, gpu_addr & 0xffffffff); - radeon_ring_write(rdev, 48 - 1); - radeon_ring_write(rdev, sq_vtx_constant_word2); - radeon_ring_write(rdev, 1 << 0); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, SQ_TEX_VTX_VALID_BUFFER << 30); + radeon_ring_write(ring, PACKET3(PACKET3_SET_RESOURCE, 7)); + radeon_ring_write(ring, 0x460); + radeon_ring_write(ring, gpu_addr & 0xffffffff); + radeon_ring_write(ring, 48 - 1); + radeon_ring_write(ring, sq_vtx_constant_word2); + radeon_ring_write(ring, 1 << 0); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, SQ_TEX_VTX_VALID_BUFFER << 30); if ((rdev->family == CHIP_RV610) || (rdev->family == CHIP_RV620) || @@ -203,6 +194,7 @@ set_tex_resource(struct radeon_device *rdev, int format, int w, int h, int pitch, u64 gpu_addr, u32 size) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; uint32_t sq_tex_resource_word0, sq_tex_resource_word1, sq_tex_resource_word4; if (h < 1) @@ -225,15 +217,15 @@ set_tex_resource(struct radeon_device *rdev, cp_set_surface_sync(rdev, PACKET3_TC_ACTION_ENA, size, gpu_addr); - radeon_ring_write(rdev, PACKET3(PACKET3_SET_RESOURCE, 7)); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, sq_tex_resource_word0); - radeon_ring_write(rdev, sq_tex_resource_word1); - radeon_ring_write(rdev, gpu_addr >> 8); - radeon_ring_write(rdev, gpu_addr >> 8); - radeon_ring_write(rdev, sq_tex_resource_word4); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, SQ_TEX_VTX_VALID_TEXTURE << 30); + radeon_ring_write(ring, PACKET3(PACKET3_SET_RESOURCE, 7)); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, sq_tex_resource_word0); + radeon_ring_write(ring, sq_tex_resource_word1); + radeon_ring_write(ring, gpu_addr >> 8); + radeon_ring_write(ring, gpu_addr >> 8); + radeon_ring_write(ring, sq_tex_resource_word4); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, SQ_TEX_VTX_VALID_TEXTURE << 30); } /* emits 12 */ @@ -241,43 +233,45 @@ static void set_scissors(struct radeon_device *rdev, int x1, int y1, int x2, int y2) { - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); - radeon_ring_write(rdev, (PA_SC_SCREEN_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(rdev, (x1 << 0) | (y1 << 16)); - radeon_ring_write(rdev, (x2 << 0) | (y2 << 16)); - - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); - radeon_ring_write(rdev, (PA_SC_GENERIC_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(rdev, (x1 << 0) | (y1 << 16) | (1 << 31)); - radeon_ring_write(rdev, (x2 << 0) | (y2 << 16)); - - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); - radeon_ring_write(rdev, (PA_SC_WINDOW_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); - radeon_ring_write(rdev, (x1 << 0) | (y1 << 16) | (1 << 31)); - radeon_ring_write(rdev, (x2 << 0) | (y2 << 16)); + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); + radeon_ring_write(ring, (PA_SC_SCREEN_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); + radeon_ring_write(ring, (x1 << 0) | (y1 << 16)); + radeon_ring_write(ring, (x2 << 0) | (y2 << 16)); + + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); + radeon_ring_write(ring, (PA_SC_GENERIC_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); + radeon_ring_write(ring, (x1 << 0) | (y1 << 16) | (1 << 31)); + radeon_ring_write(ring, (x2 << 0) | (y2 << 16)); + + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); + radeon_ring_write(ring, (PA_SC_WINDOW_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2); + radeon_ring_write(ring, (x1 << 0) | (y1 << 16) | (1 << 31)); + radeon_ring_write(ring, (x2 << 0) | (y2 << 16)); } /* emits 10 */ static void draw_auto(struct radeon_device *rdev) { - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1)); - radeon_ring_write(rdev, (VGT_PRIMITIVE_TYPE - PACKET3_SET_CONFIG_REG_OFFSET) >> 2); - radeon_ring_write(rdev, DI_PT_RECTLIST); + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + radeon_ring_write(ring, (VGT_PRIMITIVE_TYPE - PACKET3_SET_CONFIG_REG_OFFSET) >> 2); + radeon_ring_write(ring, DI_PT_RECTLIST); - radeon_ring_write(rdev, PACKET3(PACKET3_INDEX_TYPE, 0)); - radeon_ring_write(rdev, + radeon_ring_write(ring, PACKET3(PACKET3_INDEX_TYPE, 0)); + radeon_ring_write(ring, #ifdef __BIG_ENDIAN (2 << 2) | #endif DI_INDEX_SIZE_16_BIT); - radeon_ring_write(rdev, PACKET3(PACKET3_NUM_INSTANCES, 0)); - radeon_ring_write(rdev, 1); + radeon_ring_write(ring, PACKET3(PACKET3_NUM_INSTANCES, 0)); + radeon_ring_write(ring, 1); - radeon_ring_write(rdev, PACKET3(PACKET3_DRAW_INDEX_AUTO, 1)); - radeon_ring_write(rdev, 3); - radeon_ring_write(rdev, DI_SRC_SEL_AUTO_INDEX); + radeon_ring_write(ring, PACKET3(PACKET3_DRAW_INDEX_AUTO, 1)); + radeon_ring_write(ring, 3); + radeon_ring_write(ring, DI_SRC_SEL_AUTO_INDEX); } @@ -285,6 +279,7 @@ draw_auto(struct radeon_device *rdev) static void set_default_state(struct radeon_device *rdev) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; u32 sq_config, sq_gpr_resource_mgmt_1, sq_gpr_resource_mgmt_2; u32 sq_thread_resource_mgmt, sq_stack_resource_mgmt_1, sq_stack_resource_mgmt_2; int num_ps_gprs, num_vs_gprs, num_temp_gprs, num_gs_gprs, num_es_gprs; @@ -440,47 +435,62 @@ set_default_state(struct radeon_device *rdev) /* emit an IB pointing at default state */ dwords = ALIGN(rdev->r600_blit.state_len, 0x10); gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.state_offset; - radeon_ring_write(rdev, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); - radeon_ring_write(rdev, + radeon_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); + radeon_ring_write(ring, #ifdef __BIG_ENDIAN (2 << 0) | #endif (gpu_addr & 0xFFFFFFFC)); - radeon_ring_write(rdev, upper_32_bits(gpu_addr) & 0xFF); - radeon_ring_write(rdev, dwords); + radeon_ring_write(ring, upper_32_bits(gpu_addr) & 0xFF); + radeon_ring_write(ring, dwords); /* SQ config */ - radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 6)); - radeon_ring_write(rdev, (SQ_CONFIG - PACKET3_SET_CONFIG_REG_OFFSET) >> 2); - radeon_ring_write(rdev, sq_config); - radeon_ring_write(rdev, sq_gpr_resource_mgmt_1); - radeon_ring_write(rdev, sq_gpr_resource_mgmt_2); - radeon_ring_write(rdev, sq_thread_resource_mgmt); - radeon_ring_write(rdev, sq_stack_resource_mgmt_1); - radeon_ring_write(rdev, sq_stack_resource_mgmt_2); + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 6)); + radeon_ring_write(ring, (SQ_CONFIG - PACKET3_SET_CONFIG_REG_OFFSET) >> 2); + radeon_ring_write(ring, sq_config); + radeon_ring_write(ring, sq_gpr_resource_mgmt_1); + radeon_ring_write(ring, sq_gpr_resource_mgmt_2); + radeon_ring_write(ring, sq_thread_resource_mgmt); + radeon_ring_write(ring, sq_stack_resource_mgmt_1); + radeon_ring_write(ring, sq_stack_resource_mgmt_2); } +#define I2F_MAX_BITS 15 +#define I2F_MAX_INPUT ((1 << I2F_MAX_BITS) - 1) +#define I2F_SHIFT (24 - I2F_MAX_BITS) + +/* + * Converts unsigned integer into 32-bit IEEE floating point representation. + * Conversion is not universal and only works for the range from 0 + * to 2^I2F_MAX_BITS-1. Currently we only use it with inputs between + * 0 and 16384 (inclusive), so I2F_MAX_BITS=15 is enough. If necessary, + * I2F_MAX_BITS can be increased, but that will add to the loop iterations + * and slow us down. Conversion is done by shifting the input and counting + * down until the first 1 reaches bit position 23. The resulting counter + * and the shifted input are, respectively, the exponent and the fraction. + * The sign is always zero. + */ static uint32_t i2f(uint32_t input) { u32 result, i, exponent, fraction; - if ((input & 0x3fff) == 0) - result = 0; /* 0 is a special case */ + WARN_ON_ONCE(input > I2F_MAX_INPUT); + + if ((input & I2F_MAX_INPUT) == 0) + result = 0; else { - exponent = 140; /* exponent biased by 127; */ - fraction = (input & 0x3fff) << 10; /* cheat and only - handle numbers below 2^^15 */ - for (i = 0; i < 14; i++) { + exponent = 126 + I2F_MAX_BITS; + fraction = (input & I2F_MAX_INPUT) << I2F_SHIFT; + + for (i = 0; i < I2F_MAX_BITS; i++) { if (fraction & 0x800000) break; else { - fraction = fraction << 1; /* keep - shifting left until top bit = 1 */ + fraction = fraction << 1; exponent = exponent - 1; } } - result = exponent << 23 | (fraction & 0x7fffff); /* mask - off top bit; assumed 1 */ + result = exponent << 23 | (fraction & 0x7fffff); } return result; } @@ -611,16 +621,17 @@ void r600_blit_fini(struct radeon_device *rdev) radeon_bo_unref(&rdev->r600_blit.shader_obj); } -static int r600_vb_ib_get(struct radeon_device *rdev) +static int r600_vb_ib_get(struct radeon_device *rdev, unsigned size) { int r; - r = radeon_ib_get(rdev, &rdev->r600_blit.vb_ib); + r = radeon_ib_get(rdev, RADEON_RING_TYPE_GFX_INDEX, + &rdev->r600_blit.vb_ib, size); if (r) { DRM_ERROR("failed to get IB for vertex buffer\n"); return r; } - rdev->r600_blit.vb_total = 64*1024; + rdev->r600_blit.vb_total = size; rdev->r600_blit.vb_used = 0; return 0; } @@ -679,15 +690,12 @@ static unsigned r600_blit_create_rect(unsigned num_gpu_pages, int r600_blit_prepare_copy(struct radeon_device *rdev, unsigned num_gpu_pages) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; int r; int ring_size; int num_loops = 0; int dwords_per_loop = rdev->r600_blit.ring_size_per_loop; - r = r600_vb_ib_get(rdev); - if (r) - return r; - /* num loops */ while (num_gpu_pages) { num_gpu_pages -= @@ -696,10 +704,15 @@ int r600_blit_prepare_copy(struct radeon_device *rdev, unsigned num_gpu_pages) num_loops++; } + /* 48 bytes for vertex per loop */ + r = r600_vb_ib_get(rdev, (num_loops*48)+256); + if (r) + return r; + /* calculate number of loops correctly */ ring_size = num_loops * dwords_per_loop; ring_size += rdev->r600_blit.ring_size_common; - r = radeon_ring_lock(rdev, ring_size); + r = radeon_ring_lock(rdev, ring, ring_size); if (r) return r; @@ -718,7 +731,7 @@ void r600_blit_done_copy(struct radeon_device *rdev, struct radeon_fence *fence) if (fence) r = radeon_fence_emit(rdev, fence); - radeon_ring_unlock_commit(rdev); + radeon_ring_unlock_commit(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); } void r600_kms_blit_copy(struct radeon_device *rdev, diff --git a/drivers/gpu/drm/radeon/r600_blit_shaders.c b/drivers/gpu/drm/radeon/r600_blit_shaders.c index 73e2c7c..34c8b23 100644 --- a/drivers/gpu/drm/radeon/r600_blit_shaders.c +++ b/drivers/gpu/drm/radeon/r600_blit_shaders.c @@ -24,6 +24,7 @@ * Alex Deucher */ +#include #include #include diff --git a/drivers/gpu/drm/radeon/r600_cp.c b/drivers/gpu/drm/radeon/r600_cp.c index c9db493..75ed17c 100644 --- a/drivers/gpu/drm/radeon/r600_cp.c +++ b/drivers/gpu/drm/radeon/r600_cp.c @@ -407,7 +407,7 @@ static void r600_cp_load_microcode(drm_radeon_private_t *dev_priv) RADEON_WRITE(R600_GRBM_SOFT_RESET, R600_SOFT_RESET_CP); RADEON_READ(R600_GRBM_SOFT_RESET); - DRM_UDELAY(15000); + mdelay(15); RADEON_WRITE(R600_GRBM_SOFT_RESET, 0); fw_data = (const __be32 *)dev_priv->me_fw->data; @@ -500,7 +500,7 @@ static void r700_cp_load_microcode(drm_radeon_private_t *dev_priv) RADEON_WRITE(R600_GRBM_SOFT_RESET, R600_SOFT_RESET_CP); RADEON_READ(R600_GRBM_SOFT_RESET); - DRM_UDELAY(15000); + mdelay(15); RADEON_WRITE(R600_GRBM_SOFT_RESET, 0); fw_data = (const __be32 *)dev_priv->pfp_fw->data; @@ -1797,7 +1797,7 @@ static void r600_cp_init_ring_buffer(struct drm_device *dev, RADEON_WRITE(R600_GRBM_SOFT_RESET, R600_SOFT_RESET_CP); RADEON_READ(R600_GRBM_SOFT_RESET); - DRM_UDELAY(15000); + mdelay(15); RADEON_WRITE(R600_GRBM_SOFT_RESET, 0); @@ -1815,7 +1815,7 @@ static void r600_cp_init_ring_buffer(struct drm_device *dev, dev_priv->ring.size_l2qw); #endif - RADEON_WRITE(R600_CP_SEM_WAIT_TIMER, 0x4); + RADEON_WRITE(R600_CP_SEM_WAIT_TIMER, 0x0); /* Set the write pointer delay */ RADEON_WRITE(R600_CP_RB_WPTR_DELAY, 0); diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c index cb1acff..b8e12af 100644 --- a/drivers/gpu/drm/radeon/r600_cs.c +++ b/drivers/gpu/drm/radeon/r600_cs.c @@ -52,15 +52,20 @@ struct r600_cs_track { struct radeon_bo *cb_color_bo[8]; u64 cb_color_bo_mc[8]; u32 cb_color_bo_offset[8]; - struct radeon_bo *cb_color_frag_bo[8]; - struct radeon_bo *cb_color_tile_bo[8]; + struct radeon_bo *cb_color_frag_bo[8]; /* unused */ + struct radeon_bo *cb_color_tile_bo[8]; /* unused */ u32 cb_color_info[8]; - u32 cb_color_size_idx[8]; + u32 cb_color_view[8]; + u32 cb_color_size_idx[8]; /* unused */ u32 cb_target_mask; - u32 cb_shader_mask; + u32 cb_shader_mask; /* unused */ u32 cb_color_size[8]; u32 vgt_strmout_en; u32 vgt_strmout_buffer_en; + struct radeon_bo *vgt_strmout_bo[4]; + u64 vgt_strmout_bo_mc[4]; /* unused */ + u32 vgt_strmout_bo_offset[4]; + u32 vgt_strmout_size[4]; u32 db_depth_control; u32 db_depth_info; u32 db_depth_size_idx; @@ -69,13 +74,20 @@ struct r600_cs_track { u32 db_offset; struct radeon_bo *db_bo; u64 db_bo_mc; + bool sx_misc_kill_all_prims; + bool cb_dirty; + bool db_dirty; + bool streamout_dirty; + struct radeon_bo *htile_bo; + u64 htile_offset; + u32 htile_surface; }; #define FMT_8_BIT(fmt, vc) [fmt] = { 1, 1, 1, vc, CHIP_R600 } #define FMT_16_BIT(fmt, vc) [fmt] = { 1, 1, 2, vc, CHIP_R600 } -#define FMT_24_BIT(fmt) [fmt] = { 1, 1, 3, 0, CHIP_R600 } +#define FMT_24_BIT(fmt) [fmt] = { 1, 1, 4, 0, CHIP_R600 } #define FMT_32_BIT(fmt, vc) [fmt] = { 1, 1, 4, vc, CHIP_R600 } -#define FMT_48_BIT(fmt) [fmt] = { 1, 1, 6, 0, CHIP_R600 } +#define FMT_48_BIT(fmt) [fmt] = { 1, 1, 8, 0, CHIP_R600 } #define FMT_64_BIT(fmt, vc) [fmt] = { 1, 1, 8, vc, CHIP_R600 } #define FMT_96_BIT(fmt) [fmt] = { 1, 1, 12, 0, CHIP_R600 } #define FMT_128_BIT(fmt, vc) [fmt] = { 1, 1, 16,vc, CHIP_R600 } @@ -107,7 +119,7 @@ static const struct gpu_formats color_formats_table[] = { /* 24-bit */ FMT_24_BIT(V_038004_FMT_8_8_8), - + /* 32-bit */ FMT_32_BIT(V_038004_COLOR_32, 1), FMT_32_BIT(V_038004_COLOR_32_FLOAT, 1), @@ -162,22 +174,22 @@ static const struct gpu_formats color_formats_table[] = { [V_038004_FMT_32_AS_32_32_32_32] = { 1, 1, 4, 0, CHIP_CEDAR}, }; -static bool fmt_is_valid_color(u32 format) +bool r600_fmt_is_valid_color(u32 format) { if (format >= ARRAY_SIZE(color_formats_table)) return false; - + if (color_formats_table[format].valid_color) return true; return false; } -static bool fmt_is_valid_texture(u32 format, enum radeon_family family) +bool r600_fmt_is_valid_texture(u32 format, enum radeon_family family) { if (format >= ARRAY_SIZE(color_formats_table)) return false; - + if (family < color_formats_table[format].min_family) return false; @@ -187,7 +199,7 @@ static bool fmt_is_valid_texture(u32 format, enum radeon_family family) return false; } -static int fmt_get_blocksize(u32 format) +int r600_fmt_get_blocksize(u32 format) { if (format >= ARRAY_SIZE(color_formats_table)) return 0; @@ -195,7 +207,7 @@ static int fmt_get_blocksize(u32 format) return color_formats_table[format].blocksize; } -static int fmt_get_nblocksx(u32 format, u32 w) +int r600_fmt_get_nblocksx(u32 format, u32 w) { unsigned bw; @@ -209,7 +221,7 @@ static int fmt_get_nblocksx(u32 format, u32 w) return (w + bw - 1) / bw; } -static int fmt_get_nblocksy(u32 format, u32 h) +int r600_fmt_get_nblocksy(u32 format, u32 h) { unsigned bh; @@ -256,7 +268,7 @@ static int r600_get_array_mode_alignment(struct array_mode_checker *values, break; case ARRAY_LINEAR_ALIGNED: *pitch_align = max((u32)64, (u32)(values->group_size / values->blocksize)); - *height_align = tile_height; + *height_align = 1; *depth_align = 1; *base_align = values->group_size; break; @@ -269,10 +281,9 @@ static int r600_get_array_mode_alignment(struct array_mode_checker *values, *base_align = values->group_size; break; case ARRAY_2D_TILED_THIN1: - *pitch_align = max((u32)macro_tile_width, - (u32)(((values->group_size / tile_height) / - (values->blocksize * values->nsamples)) * - values->nbanks)) * tile_width; + *pitch_align = max((u32)macro_tile_width * tile_width, + (u32)((values->group_size * values->nbanks) / + (values->blocksize * values->nsamples * tile_width))); *height_align = macro_tile_height * tile_height; *depth_align = 1; *base_align = max(macro_tile_bytes, @@ -296,12 +307,14 @@ static void r600_cs_track_init(struct r600_cs_track *track) track->cb_color_size[i] = 0; track->cb_color_size_idx[i] = 0; track->cb_color_info[i] = 0; + track->cb_color_view[i] = 0xFFFFFFFF; track->cb_color_bo[i] = NULL; track->cb_color_bo_offset[i] = 0xFFFFFFFF; track->cb_color_bo_mc[i] = 0xFFFFFFFF; } track->cb_target_mask = 0xFFFFFFFF; track->cb_shader_mask = 0xFFFFFFFF; + track->cb_dirty = true; track->db_bo = NULL; track->db_bo_mc = 0xFFFFFFFF; /* assume the biggest format and that htile is enabled */ @@ -310,6 +323,19 @@ static void r600_cs_track_init(struct r600_cs_track *track) track->db_depth_size = 0xFFFFFFFF; track->db_depth_size_idx = 0; track->db_depth_control = 0xFFFFFFFF; + track->db_dirty = true; + track->htile_bo = NULL; + track->htile_offset = 0xFFFFFFFF; + track->htile_surface = 0; + + for (i = 0; i < 4; i++) { + track->vgt_strmout_size[i] = 0; + track->vgt_strmout_bo[i] = NULL; + track->vgt_strmout_bo_offset[i] = 0xFFFFFFFF; + track->vgt_strmout_bo_mc[i] = 0xFFFFFFFF; + } + track->streamout_dirty = true; + track->sx_misc_kill_all_prims = false; } static int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i) @@ -322,13 +348,14 @@ static int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i) volatile u32 *ib = p->ib->ptr; unsigned array_mode; u32 format; + if (G_0280A0_TILE_MODE(track->cb_color_info[i])) { dev_warn(p->dev, "FMASK or CMASK buffer are not supported by this kernel\n"); return -EINVAL; } size = radeon_bo_size(track->cb_color_bo[i]) - track->cb_color_bo_offset[i]; format = G_0280A0_FORMAT(track->cb_color_info[i]); - if (!fmt_is_valid_color(format)) { + if (!r600_fmt_is_valid_color(format)) { dev_warn(p->dev, "%s:%d cb invalid format %d for %d (0x%08X)\n", __func__, __LINE__, format, i, track->cb_color_info[i]); @@ -349,7 +376,7 @@ static int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i) array_check.nbanks = track->nbanks; array_check.npipes = track->npipes; array_check.nsamples = track->nsamples; - array_check.blocksize = fmt_get_blocksize(format); + array_check.blocksize = r600_fmt_get_blocksize(format); if (r600_get_array_mode_alignment(&array_check, &pitch_align, &height_align, &depth_align, &base_align)) { dev_warn(p->dev, "%s invalid tiling %d for %d (0x%08X)\n", __func__, @@ -393,7 +420,18 @@ static int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i) } /* check offset */ - tmp = fmt_get_nblocksy(format, height) * fmt_get_nblocksx(format, pitch) * fmt_get_blocksize(format); + tmp = r600_fmt_get_nblocksy(format, height) * r600_fmt_get_nblocksx(format, pitch) * r600_fmt_get_blocksize(format); + switch (array_mode) { + default: + case V_0280A0_ARRAY_LINEAR_GENERAL: + case V_0280A0_ARRAY_LINEAR_ALIGNED: + tmp += track->cb_color_view[i] & 0xFF; + break; + case V_0280A0_ARRAY_1D_TILED_THIN1: + case V_0280A0_ARRAY_2D_TILED_THIN1: + tmp += G_028080_SLICE_MAX(track->cb_color_view[i]) * tmp; + break; + } if ((tmp + track->cb_color_bo_offset[i]) > radeon_bo_size(track->cb_color_bo[i])) { if (array_mode == V_0280A0_ARRAY_LINEAR_GENERAL) { /* the initial DDX does bad things with the CB size occasionally */ @@ -403,10 +441,13 @@ static int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i) * broken userspace. */ } else { - dev_warn(p->dev, "%s offset[%d] %d %d %d %lu too big\n", __func__, i, - array_mode, + dev_warn(p->dev, "%s offset[%d] %d %d %d %lu too big (%d %d) (%d %d %d)\n", + __func__, i, array_mode, track->cb_color_bo_offset[i], tmp, - radeon_bo_size(track->cb_color_bo[i])); + radeon_bo_size(track->cb_color_bo[i]), + pitch, height, r600_fmt_get_nblocksx(format, pitch), + r600_fmt_get_nblocksy(format, height), + r600_fmt_get_blocksize(format)); return -EINVAL; } } @@ -420,154 +461,316 @@ static int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i) return 0; } -static int r600_cs_track_check(struct radeon_cs_parser *p) +static int r600_cs_track_validate_db(struct radeon_cs_parser *p) { struct r600_cs_track *track = p->track; - u32 tmp; - int r, i; + u32 nviews, bpe, ntiles, size, slice_tile_max, tmp; + u32 height_align, pitch_align, depth_align; + u32 pitch = 8192; + u32 height = 8192; + u64 base_offset, base_align; + struct array_mode_checker array_check; + int array_mode; volatile u32 *ib = p->ib->ptr; - /* on legacy kernel we don't perform advanced check */ - if (p->rdev == NULL) - return 0; - /* we don't support out buffer yet */ - if (track->vgt_strmout_en || track->vgt_strmout_buffer_en) { - dev_warn(p->dev, "this kernel doesn't support SMX output buffer\n"); + + if (track->db_bo == NULL) { + dev_warn(p->dev, "z/stencil with no depth buffer\n"); return -EINVAL; } - /* check that we have a cb for each enabled target, we don't check - * shader_mask because it seems mesa isn't always setting it :( - */ - tmp = track->cb_target_mask; - for (i = 0; i < 8; i++) { - if ((tmp >> (i * 4)) & 0xF) { - /* at least one component is enabled */ - if (track->cb_color_bo[i] == NULL) { - dev_warn(p->dev, "%s:%d mask 0x%08X | 0x%08X no cb for %d\n", - __func__, __LINE__, track->cb_target_mask, track->cb_shader_mask, i); - return -EINVAL; - } - /* perform rewrite of CB_COLOR[0-7]_SIZE */ - r = r600_cs_track_validate_cb(p, i); - if (r) - return r; - } + switch (G_028010_FORMAT(track->db_depth_info)) { + case V_028010_DEPTH_16: + bpe = 2; + break; + case V_028010_DEPTH_X8_24: + case V_028010_DEPTH_8_24: + case V_028010_DEPTH_X8_24_FLOAT: + case V_028010_DEPTH_8_24_FLOAT: + case V_028010_DEPTH_32_FLOAT: + bpe = 4; + break; + case V_028010_DEPTH_X24_8_32_FLOAT: + bpe = 8; + break; + default: + dev_warn(p->dev, "z/stencil with invalid format %d\n", G_028010_FORMAT(track->db_depth_info)); + return -EINVAL; } - /* Check depth buffer */ - if (G_028800_STENCIL_ENABLE(track->db_depth_control) || - G_028800_Z_ENABLE(track->db_depth_control)) { - u32 nviews, bpe, ntiles, size, slice_tile_max; - u32 height, height_align, pitch, pitch_align, depth_align; - u64 base_offset, base_align; - struct array_mode_checker array_check; - int array_mode; - - if (track->db_bo == NULL) { - dev_warn(p->dev, "z/stencil with no depth buffer\n"); + if ((track->db_depth_size & 0xFFFFFC00) == 0xFFFFFC00) { + if (!track->db_depth_size_idx) { + dev_warn(p->dev, "z/stencil buffer size not set\n"); return -EINVAL; } - if (G_028010_TILE_SURFACE_ENABLE(track->db_depth_info)) { - dev_warn(p->dev, "this kernel doesn't support z/stencil htile\n"); + tmp = radeon_bo_size(track->db_bo) - track->db_offset; + tmp = (tmp / bpe) >> 6; + if (!tmp) { + dev_warn(p->dev, "z/stencil buffer too small (0x%08X %d %d %ld)\n", + track->db_depth_size, bpe, track->db_offset, + radeon_bo_size(track->db_bo)); return -EINVAL; } - switch (G_028010_FORMAT(track->db_depth_info)) { - case V_028010_DEPTH_16: - bpe = 2; - break; - case V_028010_DEPTH_X8_24: - case V_028010_DEPTH_8_24: - case V_028010_DEPTH_X8_24_FLOAT: - case V_028010_DEPTH_8_24_FLOAT: - case V_028010_DEPTH_32_FLOAT: - bpe = 4; + ib[track->db_depth_size_idx] = S_028000_SLICE_TILE_MAX(tmp - 1) | (track->db_depth_size & 0x3FF); + } else { + size = radeon_bo_size(track->db_bo); + /* pitch in pixels */ + pitch = (G_028000_PITCH_TILE_MAX(track->db_depth_size) + 1) * 8; + slice_tile_max = G_028000_SLICE_TILE_MAX(track->db_depth_size) + 1; + slice_tile_max *= 64; + height = slice_tile_max / pitch; + if (height > 8192) + height = 8192; + base_offset = track->db_bo_mc + track->db_offset; + array_mode = G_028010_ARRAY_MODE(track->db_depth_info); + array_check.array_mode = array_mode; + array_check.group_size = track->group_size; + array_check.nbanks = track->nbanks; + array_check.npipes = track->npipes; + array_check.nsamples = track->nsamples; + array_check.blocksize = bpe; + if (r600_get_array_mode_alignment(&array_check, + &pitch_align, &height_align, &depth_align, &base_align)) { + dev_warn(p->dev, "%s invalid tiling %d (0x%08X)\n", __func__, + G_028010_ARRAY_MODE(track->db_depth_info), + track->db_depth_info); + return -EINVAL; + } + switch (array_mode) { + case V_028010_ARRAY_1D_TILED_THIN1: + /* don't break userspace */ + height &= ~0x7; break; - case V_028010_DEPTH_X24_8_32_FLOAT: - bpe = 8; + case V_028010_ARRAY_2D_TILED_THIN1: break; default: - dev_warn(p->dev, "z/stencil with invalid format %d\n", G_028010_FORMAT(track->db_depth_info)); + dev_warn(p->dev, "%s invalid tiling %d (0x%08X)\n", __func__, + G_028010_ARRAY_MODE(track->db_depth_info), + track->db_depth_info); + return -EINVAL; + } + + if (!IS_ALIGNED(pitch, pitch_align)) { + dev_warn(p->dev, "%s:%d db pitch (%d, 0x%x, %d) invalid\n", + __func__, __LINE__, pitch, pitch_align, array_mode); + return -EINVAL; + } + if (!IS_ALIGNED(height, height_align)) { + dev_warn(p->dev, "%s:%d db height (%d, 0x%x, %d) invalid\n", + __func__, __LINE__, height, height_align, array_mode); + return -EINVAL; + } + if (!IS_ALIGNED(base_offset, base_align)) { + dev_warn(p->dev, "%s offset 0x%llx, 0x%llx, %d not aligned\n", __func__, + base_offset, base_align, array_mode); + return -EINVAL; + } + + ntiles = G_028000_SLICE_TILE_MAX(track->db_depth_size) + 1; + nviews = G_028004_SLICE_MAX(track->db_depth_view) + 1; + tmp = ntiles * bpe * 64 * nviews; + if ((tmp + track->db_offset) > radeon_bo_size(track->db_bo)) { + dev_warn(p->dev, "z/stencil buffer (%d) too small (0x%08X %d %d %d -> %u have %lu)\n", + array_mode, + track->db_depth_size, ntiles, nviews, bpe, tmp + track->db_offset, + radeon_bo_size(track->db_bo)); + return -EINVAL; + } + } + + /* hyperz */ + if (G_028010_TILE_SURFACE_ENABLE(track->db_depth_info)) { + unsigned long size; + unsigned nbx, nby; + + if (track->htile_bo == NULL) { + dev_warn(p->dev, "%s:%d htile enabled without htile surface 0x%08x\n", + __func__, __LINE__, track->db_depth_info); return -EINVAL; } if ((track->db_depth_size & 0xFFFFFC00) == 0xFFFFFC00) { - if (!track->db_depth_size_idx) { - dev_warn(p->dev, "z/stencil buffer size not set\n"); - return -EINVAL; - } - tmp = radeon_bo_size(track->db_bo) - track->db_offset; - tmp = (tmp / bpe) >> 6; - if (!tmp) { - dev_warn(p->dev, "z/stencil buffer too small (0x%08X %d %d %ld)\n", - track->db_depth_size, bpe, track->db_offset, - radeon_bo_size(track->db_bo)); - return -EINVAL; - } - ib[track->db_depth_size_idx] = S_028000_SLICE_TILE_MAX(tmp - 1) | (track->db_depth_size & 0x3FF); + dev_warn(p->dev, "%s:%d htile can't be enabled with bogus db_depth_size 0x%08x\n", + __func__, __LINE__, track->db_depth_size); + return -EINVAL; + } + + nbx = pitch; + nby = height; + if (G_028D24_LINEAR(track->htile_surface)) { + /* nbx must be 16 htiles aligned == 16 * 8 pixel aligned */ + nbx = round_up(nbx, 16 * 8); + /* nby is npipes htiles aligned == npipes * 8 pixel aligned */ + nby = round_up(nby, track->npipes * 8); } else { - size = radeon_bo_size(track->db_bo); - /* pitch in pixels */ - pitch = (G_028000_PITCH_TILE_MAX(track->db_depth_size) + 1) * 8; - slice_tile_max = G_028000_SLICE_TILE_MAX(track->db_depth_size) + 1; - slice_tile_max *= 64; - height = slice_tile_max / pitch; - if (height > 8192) - height = 8192; - base_offset = track->db_bo_mc + track->db_offset; - array_mode = G_028010_ARRAY_MODE(track->db_depth_info); - array_check.array_mode = array_mode; - array_check.group_size = track->group_size; - array_check.nbanks = track->nbanks; - array_check.npipes = track->npipes; - array_check.nsamples = track->nsamples; - array_check.blocksize = bpe; - if (r600_get_array_mode_alignment(&array_check, - &pitch_align, &height_align, &depth_align, &base_align)) { - dev_warn(p->dev, "%s invalid tiling %d (0x%08X)\n", __func__, - G_028010_ARRAY_MODE(track->db_depth_info), - track->db_depth_info); - return -EINVAL; - } - switch (array_mode) { - case V_028010_ARRAY_1D_TILED_THIN1: - /* don't break userspace */ - height &= ~0x7; + /* htile widht & nby (8 or 4) make 2 bits number */ + tmp = track->htile_surface & 3; + /* align is htile align * 8, htile align vary according to + * number of pipe and tile width and nby + */ + switch (track->npipes) { + case 8: + switch (tmp) { + case 3: /* HTILE_WIDTH = 8 & HTILE_HEIGHT = 8*/ + nbx = round_up(nbx, 64 * 8); + nby = round_up(nby, 64 * 8); + break; + case 2: /* HTILE_WIDTH = 4 & HTILE_HEIGHT = 8*/ + case 1: /* HTILE_WIDTH = 8 & HTILE_HEIGHT = 4*/ + nbx = round_up(nbx, 64 * 8); + nby = round_up(nby, 32 * 8); + break; + case 0: /* HTILE_WIDTH = 4 & HTILE_HEIGHT = 4*/ + nbx = round_up(nbx, 32 * 8); + nby = round_up(nby, 32 * 8); + break; + default: + return -EINVAL; + } + break; + case 4: + switch (tmp) { + case 3: /* HTILE_WIDTH = 8 & HTILE_HEIGHT = 8*/ + nbx = round_up(nbx, 64 * 8); + nby = round_up(nby, 32 * 8); + break; + case 2: /* HTILE_WIDTH = 4 & HTILE_HEIGHT = 8*/ + case 1: /* HTILE_WIDTH = 8 & HTILE_HEIGHT = 4*/ + nbx = round_up(nbx, 32 * 8); + nby = round_up(nby, 32 * 8); + break; + case 0: /* HTILE_WIDTH = 4 & HTILE_HEIGHT = 4*/ + nbx = round_up(nbx, 32 * 8); + nby = round_up(nby, 16 * 8); + break; + default: + return -EINVAL; + } + break; + case 2: + switch (tmp) { + case 3: /* HTILE_WIDTH = 8 & HTILE_HEIGHT = 8*/ + nbx = round_up(nbx, 32 * 8); + nby = round_up(nby, 32 * 8); + break; + case 2: /* HTILE_WIDTH = 4 & HTILE_HEIGHT = 8*/ + case 1: /* HTILE_WIDTH = 8 & HTILE_HEIGHT = 4*/ + nbx = round_up(nbx, 32 * 8); + nby = round_up(nby, 16 * 8); + break; + case 0: /* HTILE_WIDTH = 4 & HTILE_HEIGHT = 4*/ + nbx = round_up(nbx, 16 * 8); + nby = round_up(nby, 16 * 8); + break; + default: + return -EINVAL; + } break; - case V_028010_ARRAY_2D_TILED_THIN1: + case 1: + switch (tmp) { + case 3: /* HTILE_WIDTH = 8 & HTILE_HEIGHT = 8*/ + nbx = round_up(nbx, 32 * 8); + nby = round_up(nby, 16 * 8); + break; + case 2: /* HTILE_WIDTH = 4 & HTILE_HEIGHT = 8*/ + case 1: /* HTILE_WIDTH = 8 & HTILE_HEIGHT = 4*/ + nbx = round_up(nbx, 16 * 8); + nby = round_up(nby, 16 * 8); + break; + case 0: /* HTILE_WIDTH = 4 & HTILE_HEIGHT = 4*/ + nbx = round_up(nbx, 16 * 8); + nby = round_up(nby, 8 * 8); + break; + default: + return -EINVAL; + } break; default: - dev_warn(p->dev, "%s invalid tiling %d (0x%08X)\n", __func__, - G_028010_ARRAY_MODE(track->db_depth_info), - track->db_depth_info); + dev_warn(p->dev, "%s:%d invalid num pipes %d\n", + __func__, __LINE__, track->npipes); return -EINVAL; } + } + /* compute number of htile */ + nbx = G_028D24_HTILE_WIDTH(track->htile_surface) ? nbx / 8 : nbx / 4; + nby = G_028D24_HTILE_HEIGHT(track->htile_surface) ? nby / 8 : nby / 4; + size = nbx * nby * 4; + size += track->htile_offset; + + if (size > radeon_bo_size(track->htile_bo)) { + dev_warn(p->dev, "%s:%d htile surface too small %ld for %ld (%d %d)\n", + __func__, __LINE__, radeon_bo_size(track->htile_bo), + size, nbx, nby); + return -EINVAL; + } + } - if (!IS_ALIGNED(pitch, pitch_align)) { - dev_warn(p->dev, "%s:%d db pitch (%d, 0x%x, %d) invalid\n", - __func__, __LINE__, pitch, pitch_align, array_mode); - return -EINVAL; - } - if (!IS_ALIGNED(height, height_align)) { - dev_warn(p->dev, "%s:%d db height (%d, 0x%x, %d) invalid\n", - __func__, __LINE__, height, height_align, array_mode); - return -EINVAL; - } - if (!IS_ALIGNED(base_offset, base_align)) { - dev_warn(p->dev, "%s offset[%d] 0x%llx, 0x%llx, %d not aligned\n", __func__, i, - base_offset, base_align, array_mode); - return -EINVAL; + track->db_dirty = false; + return 0; +} + +static int r600_cs_track_check(struct radeon_cs_parser *p) +{ + struct r600_cs_track *track = p->track; + u32 tmp; + int r, i; + + /* on legacy kernel we don't perform advanced check */ + if (p->rdev == NULL) + return 0; + + /* check streamout */ + if (track->streamout_dirty && track->vgt_strmout_en) { + for (i = 0; i < 4; i++) { + if (track->vgt_strmout_buffer_en & (1 << i)) { + if (track->vgt_strmout_bo[i]) { + u64 offset = (u64)track->vgt_strmout_bo_offset[i] + + (u64)track->vgt_strmout_size[i]; + if (offset > radeon_bo_size(track->vgt_strmout_bo[i])) { + DRM_ERROR("streamout %d bo too small: 0x%llx, 0x%lx\n", + i, offset, + radeon_bo_size(track->vgt_strmout_bo[i])); + return -EINVAL; + } + } else { + dev_warn(p->dev, "No buffer for streamout %d\n", i); + return -EINVAL; + } } + } + track->streamout_dirty = false; + } - ntiles = G_028000_SLICE_TILE_MAX(track->db_depth_size) + 1; - nviews = G_028004_SLICE_MAX(track->db_depth_view) + 1; - tmp = ntiles * bpe * 64 * nviews; - if ((tmp + track->db_offset) > radeon_bo_size(track->db_bo)) { - dev_warn(p->dev, "z/stencil buffer (%d) too small (0x%08X %d %d %d -> %u have %lu)\n", - array_mode, - track->db_depth_size, ntiles, nviews, bpe, tmp + track->db_offset, - radeon_bo_size(track->db_bo)); - return -EINVAL; + if (track->sx_misc_kill_all_prims) + return 0; + + /* check that we have a cb for each enabled target, we don't check + * shader_mask because it seems mesa isn't always setting it :( + */ + if (track->cb_dirty) { + tmp = track->cb_target_mask; + for (i = 0; i < 8; i++) { + if ((tmp >> (i * 4)) & 0xF) { + /* at least one component is enabled */ + if (track->cb_color_bo[i] == NULL) { + dev_warn(p->dev, "%s:%d mask 0x%08X | 0x%08X no cb for %d\n", + __func__, __LINE__, track->cb_target_mask, track->cb_shader_mask, i); + return -EINVAL; + } + /* perform rewrite of CB_COLOR[0-7]_SIZE */ + r = r600_cs_track_validate_cb(p, i); + if (r) + return r; } } + track->cb_dirty = false; + } + + /* Check depth buffer */ + if (track->db_dirty && (G_028800_STENCIL_ENABLE(track->db_depth_control) || + G_028800_Z_ENABLE(track->db_depth_control))) { + r = r600_cs_track_validate_db(p); + if (r) + return r; } + return 0; } @@ -939,9 +1142,10 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) break; case R_028800_DB_DEPTH_CONTROL: track->db_depth_control = radeon_get_ib_value(p, idx); + track->db_dirty = true; break; case R_028010_DB_DEPTH_INFO: - if (!p->keep_tiling_flags && + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS) && r600_cs_packet_next_is_pkt3_nop(p)) { r = r600_cs_packet_next_reloc(p, &reloc); if (r) { @@ -959,24 +1163,66 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) ib[idx] |= S_028010_ARRAY_MODE(V_028010_ARRAY_1D_TILED_THIN1); track->db_depth_info |= S_028010_ARRAY_MODE(V_028010_ARRAY_1D_TILED_THIN1); } - } else + } else { track->db_depth_info = radeon_get_ib_value(p, idx); + } + track->db_dirty = true; break; case R_028004_DB_DEPTH_VIEW: track->db_depth_view = radeon_get_ib_value(p, idx); + track->db_dirty = true; break; case R_028000_DB_DEPTH_SIZE: track->db_depth_size = radeon_get_ib_value(p, idx); track->db_depth_size_idx = idx; + track->db_dirty = true; break; case R_028AB0_VGT_STRMOUT_EN: track->vgt_strmout_en = radeon_get_ib_value(p, idx); + track->streamout_dirty = true; break; case R_028B20_VGT_STRMOUT_BUFFER_EN: track->vgt_strmout_buffer_en = radeon_get_ib_value(p, idx); + track->streamout_dirty = true; + break; + case VGT_STRMOUT_BUFFER_BASE_0: + case VGT_STRMOUT_BUFFER_BASE_1: + case VGT_STRMOUT_BUFFER_BASE_2: + case VGT_STRMOUT_BUFFER_BASE_3: + r = r600_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "bad SET_CONTEXT_REG " + "0x%04X\n", reg); + return -EINVAL; + } + tmp = (reg - VGT_STRMOUT_BUFFER_BASE_0) / 16; + track->vgt_strmout_bo_offset[tmp] = radeon_get_ib_value(p, idx) << 8; + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + track->vgt_strmout_bo[tmp] = reloc->robj; + track->vgt_strmout_bo_mc[tmp] = reloc->lobj.gpu_offset; + track->streamout_dirty = true; + break; + case VGT_STRMOUT_BUFFER_SIZE_0: + case VGT_STRMOUT_BUFFER_SIZE_1: + case VGT_STRMOUT_BUFFER_SIZE_2: + case VGT_STRMOUT_BUFFER_SIZE_3: + tmp = (reg - VGT_STRMOUT_BUFFER_SIZE_0) / 16; + /* size in register is DWs, convert to bytes */ + track->vgt_strmout_size[tmp] = radeon_get_ib_value(p, idx) * 4; + track->streamout_dirty = true; + break; + case CP_COHER_BASE: + r = r600_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "missing reloc for CP_COHER_BASE " + "0x%04X\n", reg); + return -EINVAL; + } + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); break; case R_028238_CB_TARGET_MASK: track->cb_target_mask = radeon_get_ib_value(p, idx); + track->cb_dirty = true; break; case R_02823C_CB_SHADER_MASK: track->cb_shader_mask = radeon_get_ib_value(p, idx); @@ -984,6 +1230,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) case R_028C04_PA_SC_AA_CONFIG: tmp = G_028C04_MSAA_NUM_SAMPLES(radeon_get_ib_value(p, idx)); track->nsamples = 1 << tmp; + track->cb_dirty = true; break; case R_0280A0_CB_COLOR0_INFO: case R_0280A4_CB_COLOR1_INFO: @@ -993,7 +1240,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) case R_0280B4_CB_COLOR5_INFO: case R_0280B8_CB_COLOR6_INFO: case R_0280BC_CB_COLOR7_INFO: - if (!p->keep_tiling_flags && + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS) && r600_cs_packet_next_is_pkt3_nop(p)) { r = r600_cs_packet_next_reloc(p, &reloc); if (r) { @@ -1013,6 +1260,19 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) tmp = (reg - R_0280A0_CB_COLOR0_INFO) / 4; track->cb_color_info[tmp] = radeon_get_ib_value(p, idx); } + track->cb_dirty = true; + break; + case R_028080_CB_COLOR0_VIEW: + case R_028084_CB_COLOR1_VIEW: + case R_028088_CB_COLOR2_VIEW: + case R_02808C_CB_COLOR3_VIEW: + case R_028090_CB_COLOR4_VIEW: + case R_028094_CB_COLOR5_VIEW: + case R_028098_CB_COLOR6_VIEW: + case R_02809C_CB_COLOR7_VIEW: + tmp = (reg - R_028080_CB_COLOR0_VIEW) / 4; + track->cb_color_view[tmp] = radeon_get_ib_value(p, idx); + track->cb_dirty = true; break; case R_028060_CB_COLOR0_SIZE: case R_028064_CB_COLOR1_SIZE: @@ -1025,6 +1285,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) tmp = (reg - R_028060_CB_COLOR0_SIZE) / 4; track->cb_color_size[tmp] = radeon_get_ib_value(p, idx); track->cb_color_size_idx[tmp] = idx; + track->cb_dirty = true; break; /* This register were added late, there is userspace * which does provide relocation for those but set @@ -1107,6 +1368,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) track->cb_color_base_last[tmp] = ib[idx]; track->cb_color_bo[tmp] = reloc->robj; track->cb_color_bo_mc[tmp] = reloc->lobj.gpu_offset; + track->cb_dirty = true; break; case DB_DEPTH_BASE: r = r600_cs_packet_next_reloc(p, &reloc); @@ -1119,8 +1381,24 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); track->db_bo = reloc->robj; track->db_bo_mc = reloc->lobj.gpu_offset; + track->db_dirty = true; break; case DB_HTILE_DATA_BASE: + r = r600_cs_packet_next_reloc(p, &reloc); + if (r) { + dev_warn(p->dev, "bad SET_CONTEXT_REG " + "0x%04X\n", reg); + return -EINVAL; + } + track->htile_offset = radeon_get_ib_value(p, idx) << 8; + ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); + track->htile_bo = reloc->robj; + track->db_dirty = true; + break; + case DB_HTILE_SURFACE: + track->htile_surface = radeon_get_ib_value(p, idx); + track->db_dirty = true; + break; case SQ_PGM_START_FS: case SQ_PGM_START_ES: case SQ_PGM_START_VS: @@ -1191,6 +1469,9 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) } ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); break; + case SX_MISC: + track->sx_misc_kill_all_prims = (radeon_get_ib_value(p, idx) & 0x1) != 0; + break; default: dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx); return -EINVAL; @@ -1198,7 +1479,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) return 0; } -static unsigned mip_minify(unsigned size, unsigned level) +unsigned r600_mip_minify(unsigned size, unsigned level) { unsigned val; @@ -1220,22 +1501,22 @@ static void r600_texture_size(unsigned nfaces, unsigned blevel, unsigned llevel, unsigned nlevels = llevel - blevel + 1; *l0_size = -1; - blocksize = fmt_get_blocksize(format); + blocksize = r600_fmt_get_blocksize(format); - w0 = mip_minify(w0, 0); - h0 = mip_minify(h0, 0); - d0 = mip_minify(d0, 0); + w0 = r600_mip_minify(w0, 0); + h0 = r600_mip_minify(h0, 0); + d0 = r600_mip_minify(d0, 0); for(i = 0, offset = 0, level = blevel; i < nlevels; i++, level++) { - width = mip_minify(w0, i); - nbx = fmt_get_nblocksx(format, width); + width = r600_mip_minify(w0, i); + nbx = r600_fmt_get_nblocksx(format, width); nbx = round_up(nbx, block_align); - height = mip_minify(h0, i); - nby = fmt_get_nblocksy(format, height); + height = r600_mip_minify(h0, i); + nby = r600_fmt_get_nblocksy(format, height); nby = round_up(nby, height_align); - depth = mip_minify(d0, i); + depth = r600_mip_minify(d0, i); size = nbx * nby * blocksize; if (nfaces) @@ -1293,7 +1574,7 @@ static int r600_check_texture_resource(struct radeon_cs_parser *p, u32 idx, mip_offset <<= 8; word0 = radeon_get_ib_value(p, idx + 0); - if (!p->keep_tiling_flags) { + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { if (tiling_flags & RADEON_TILING_MACRO) word0 |= S_038000_TILE_MODE(V_038000_ARRAY_2D_TILED_THIN1); else if (tiling_flags & RADEON_TILING_MICRO) @@ -1304,6 +1585,7 @@ static int r600_check_texture_resource(struct radeon_cs_parser *p, u32 idx, h0 = G_038004_TEX_HEIGHT(word1) + 1; d0 = G_038004_TEX_DEPTH(word1); nfaces = 1; + array = 0; switch (G_038000_DIM(word0)) { case V_038000_SQ_TEX_DIM_1D: case V_038000_SQ_TEX_DIM_2D: @@ -1326,7 +1608,7 @@ static int r600_check_texture_resource(struct radeon_cs_parser *p, u32 idx, return -EINVAL; } format = G_038004_DATA_FORMAT(word1); - if (!fmt_is_valid_texture(format, p->family)) { + if (!r600_fmt_is_valid_texture(format, p->family)) { dev_warn(p->dev, "%s:%d texture invalid format %d\n", __func__, __LINE__, format); return -EINVAL; @@ -1339,7 +1621,7 @@ static int r600_check_texture_resource(struct radeon_cs_parser *p, u32 idx, array_check.nbanks = track->nbanks; array_check.npipes = track->npipes; array_check.nsamples = 1; - array_check.blocksize = fmt_get_blocksize(format); + array_check.blocksize = r600_fmt_get_blocksize(format); if (r600_get_array_mode_alignment(&array_check, &pitch_align, &height_align, &depth_align, &base_align)) { dev_warn(p->dev, "%s:%d tex array mode (%d) invalid\n", @@ -1372,6 +1654,10 @@ static int r600_check_texture_resource(struct radeon_cs_parser *p, u32 idx, word1 = radeon_get_ib_value(p, idx + 5); blevel = G_038010_BASE_LEVEL(word0); llevel = G_038014_LAST_LEVEL(word1); + if (blevel > llevel) { + dev_warn(p->dev, "texture blevel %d > llevel %d\n", + blevel, llevel); + } if (array == 1) { barray = G_038014_BASE_ARRAY(word1); larray = G_038014_LAST_ARRAY(word1); @@ -1383,8 +1669,10 @@ static int r600_check_texture_resource(struct radeon_cs_parser *p, u32 idx, &l0_size, &mipmap_size); /* using get ib will give us the offset into the texture bo */ if ((l0_size + word2) > radeon_bo_size(texture)) { - dev_warn(p->dev, "texture bo too small (%d %d %d %d -> %d have %ld)\n", - w0, h0, format, word2, l0_size, radeon_bo_size(texture)); + dev_warn(p->dev, "texture bo too small ((%d %d) (%d %d) %d %d %d -> %d have %ld)\n", + w0, h0, pitch_align, height_align, + array_check.array_mode, format, word2, + l0_size, radeon_bo_size(texture)); dev_warn(p->dev, "alignments %d %d %d %lld\n", pitch, pitch_align, height_align, base_align); return -EINVAL; } @@ -1397,6 +1685,22 @@ static int r600_check_texture_resource(struct radeon_cs_parser *p, u32 idx, return 0; } +static bool r600_is_safe_reg(struct radeon_cs_parser *p, u32 reg, u32 idx) +{ + u32 m, i; + + i = (reg >> 7); + if (i >= ARRAY_SIZE(r600_reg_safe_bm)) { + dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx); + return false; + } + m = 1 << ((reg >> 2) & 31); + if (!(r600_reg_safe_bm[i] & m)) + return true; + dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx); + return false; +} + static int r600_packet3_check(struct radeon_cs_parser *p, struct radeon_cs_packet *pkt) { @@ -1419,6 +1723,8 @@ static int r600_packet3_check(struct radeon_cs_parser *p, { int pred_op; int tmp; + uint64_t offset; + if (pkt->count != 1) { DRM_ERROR("bad SET PREDICATION\n"); return -EINVAL; @@ -1442,8 +1748,12 @@ static int r600_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } - ib[idx + 0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff); - ib[idx + 1] = tmp + (upper_32_bits(reloc->lobj.gpu_offset) & 0xff); + offset = reloc->lobj.gpu_offset + + (idx_value & 0xfffffff0) + + ((u64)(tmp & 0xff) << 32); + + ib[idx + 0] = offset; + ib[idx + 1] = (tmp & 0xffffff00) | (upper_32_bits(offset) & 0xff); } break; @@ -1467,6 +1777,8 @@ static int r600_packet3_check(struct radeon_cs_parser *p, } break; case PACKET3_DRAW_INDEX: + { + uint64_t offset; if (pkt->count != 3) { DRM_ERROR("bad DRAW_INDEX\n"); return -EINVAL; @@ -1476,14 +1788,21 @@ static int r600_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad DRAW_INDEX\n"); return -EINVAL; } - ib[idx+0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff); - ib[idx+1] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + + offset = reloc->lobj.gpu_offset + + idx_value + + ((u64)(radeon_get_ib_value(p, idx+1) & 0xff) << 32); + + ib[idx+0] = offset; + ib[idx+1] = upper_32_bits(offset) & 0xff; + r = r600_cs_track_check(p); if (r) { dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__); return r; } break; + } case PACKET3_DRAW_INDEX_AUTO: if (pkt->count != 1) { DRM_ERROR("bad DRAW_INDEX_AUTO\n"); @@ -1514,13 +1833,20 @@ static int r600_packet3_check(struct radeon_cs_parser *p, } /* bit 4 is reg (0) or mem (1) */ if (idx_value & 0x10) { + uint64_t offset; + r = r600_cs_packet_next_reloc(p, &reloc); if (r) { DRM_ERROR("bad WAIT_REG_MEM\n"); return -EINVAL; } - ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff); - ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + + offset = reloc->lobj.gpu_offset + + (radeon_get_ib_value(p, idx+1) & 0xfffffff0) + + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); + + ib[idx+1] = (ib[idx+1] & 0x3) | (offset & 0xfffffff0); + ib[idx+2] = upper_32_bits(offset) & 0xff; } break; case PACKET3_SURFACE_SYNC: @@ -1545,16 +1871,25 @@ static int r600_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } if (pkt->count) { + uint64_t offset; + r = r600_cs_packet_next_reloc(p, &reloc); if (r) { DRM_ERROR("bad EVENT_WRITE\n"); return -EINVAL; } - ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff); - ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + offset = reloc->lobj.gpu_offset + + (radeon_get_ib_value(p, idx+1) & 0xfffffff8) + + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); + + ib[idx+1] = offset & 0xfffffff8; + ib[idx+2] = upper_32_bits(offset) & 0xff; } break; case PACKET3_EVENT_WRITE_EOP: + { + uint64_t offset; + if (pkt->count != 4) { DRM_ERROR("bad EVENT_WRITE_EOP\n"); return -EINVAL; @@ -1564,9 +1899,15 @@ static int r600_packet3_check(struct radeon_cs_parser *p, DRM_ERROR("bad EVENT_WRITE\n"); return -EINVAL; } - ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff); - ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + + offset = reloc->lobj.gpu_offset + + (radeon_get_ib_value(p, idx+1) & 0xfffffffc) + + ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32); + + ib[idx+1] = offset & 0xfffffffc; + ib[idx+2] = (ib[idx+2] & 0xffffff00) | (upper_32_bits(offset) & 0xff); break; + } case PACKET3_SET_CONFIG_REG: start_reg = (idx_value << 2) + PACKET3_SET_CONFIG_REG_OFFSET; end_reg = 4 * pkt->count + start_reg - 4; @@ -1625,7 +1966,7 @@ static int r600_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } base_offset = (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff); - if (!p->keep_tiling_flags) { + if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) { if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) ib[idx+1+(i*7)+0] |= S_038000_TILE_MODE(V_038000_ARRAY_2D_TILED_THIN1); else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) @@ -1651,6 +1992,8 @@ static int r600_packet3_check(struct radeon_cs_parser *p, ib[idx+1+(i*7)+3] += mip_offset; break; case SQ_TEX_VTX_VALID_BUFFER: + { + uint64_t offset64; /* vtx base */ r = r600_cs_packet_next_reloc(p, &reloc); if (r) { @@ -1663,11 +2006,15 @@ static int r600_packet3_check(struct radeon_cs_parser *p, /* force size to size of the buffer */ dev_warn(p->dev, "vbo resource seems too big (%d) for the bo (%ld)\n", size + offset, radeon_bo_size(reloc->robj)); - ib[idx+1+(i*7)+1] = radeon_bo_size(reloc->robj); + ib[idx+1+(i*7)+1] = radeon_bo_size(reloc->robj) - offset; } - ib[idx+1+(i*7)+0] += (u32)((reloc->lobj.gpu_offset) & 0xffffffff); - ib[idx+1+(i*7)+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff; + + offset64 = reloc->lobj.gpu_offset + offset; + ib[idx+1+(i*8)+0] = offset64; + ib[idx+1+(i*8)+2] = (ib[idx+1+(i*8)+2] & 0xffffff00) | + (upper_32_bits(offset64) & 0xff); break; + } case SQ_TEX_VTX_INVALID_TEXTURE: case SQ_TEX_VTX_INVALID_BUFFER: default: @@ -1742,6 +2089,104 @@ static int r600_packet3_check(struct radeon_cs_parser *p, return -EINVAL; } break; + case PACKET3_STRMOUT_BUFFER_UPDATE: + if (pkt->count != 4) { + DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (invalid count)\n"); + return -EINVAL; + } + /* Updating memory at DST_ADDRESS. */ + if (idx_value & 0x1) { + u64 offset; + r = r600_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (missing dst reloc)\n"); + return -EINVAL; + } + offset = radeon_get_ib_value(p, idx+1); + offset += ((u64)(radeon_get_ib_value(p, idx+2) & 0xff)) << 32; + if ((offset + 4) > radeon_bo_size(reloc->robj)) { + DRM_ERROR("bad STRMOUT_BUFFER_UPDATE dst bo too small: 0x%llx, 0x%lx\n", + offset + 4, radeon_bo_size(reloc->robj)); + return -EINVAL; + } + offset += reloc->lobj.gpu_offset; + ib[idx+1] = offset; + ib[idx+2] = upper_32_bits(offset) & 0xff; + } + /* Reading data from SRC_ADDRESS. */ + if (((idx_value >> 1) & 0x3) == 2) { + u64 offset; + r = r600_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (missing src reloc)\n"); + return -EINVAL; + } + offset = radeon_get_ib_value(p, idx+3); + offset += ((u64)(radeon_get_ib_value(p, idx+4) & 0xff)) << 32; + if ((offset + 4) > radeon_bo_size(reloc->robj)) { + DRM_ERROR("bad STRMOUT_BUFFER_UPDATE src bo too small: 0x%llx, 0x%lx\n", + offset + 4, radeon_bo_size(reloc->robj)); + return -EINVAL; + } + offset += reloc->lobj.gpu_offset; + ib[idx+3] = offset; + ib[idx+4] = upper_32_bits(offset) & 0xff; + } + break; + case PACKET3_COPY_DW: + if (pkt->count != 4) { + DRM_ERROR("bad COPY_DW (invalid count)\n"); + return -EINVAL; + } + if (idx_value & 0x1) { + u64 offset; + /* SRC is memory. */ + r = r600_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("bad COPY_DW (missing src reloc)\n"); + return -EINVAL; + } + offset = radeon_get_ib_value(p, idx+1); + offset += ((u64)(radeon_get_ib_value(p, idx+2) & 0xff)) << 32; + if ((offset + 4) > radeon_bo_size(reloc->robj)) { + DRM_ERROR("bad COPY_DW src bo too small: 0x%llx, 0x%lx\n", + offset + 4, radeon_bo_size(reloc->robj)); + return -EINVAL; + } + offset += reloc->lobj.gpu_offset; + ib[idx+1] = offset; + ib[idx+2] = upper_32_bits(offset) & 0xff; + } else { + /* SRC is a reg. */ + reg = radeon_get_ib_value(p, idx+1) << 2; + if (!r600_is_safe_reg(p, reg, idx+1)) + return -EINVAL; + } + if (idx_value & 0x2) { + u64 offset; + /* DST is memory. */ + r = r600_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("bad COPY_DW (missing dst reloc)\n"); + return -EINVAL; + } + offset = radeon_get_ib_value(p, idx+3); + offset += ((u64)(radeon_get_ib_value(p, idx+4) & 0xff)) << 32; + if ((offset + 4) > radeon_bo_size(reloc->robj)) { + DRM_ERROR("bad COPY_DW dst bo too small: 0x%llx, 0x%lx\n", + offset + 4, radeon_bo_size(reloc->robj)); + return -EINVAL; + } + offset += reloc->lobj.gpu_offset; + ib[idx+3] = offset; + ib[idx+4] = upper_32_bits(offset) & 0xff; + } else { + /* DST is a reg. */ + reg = radeon_get_ib_value(p, idx+3) << 2; + if (!r600_is_safe_reg(p, reg, idx+3)) + return -EINVAL; + } + break; case PACKET3_NOP: break; default: diff --git a/drivers/gpu/drm/radeon/r600_hdmi.c b/drivers/gpu/drm/radeon/r600_hdmi.c index c45d921..0b59206 100644 --- a/drivers/gpu/drm/radeon/r600_hdmi.c +++ b/drivers/gpu/drm/radeon/r600_hdmi.c @@ -320,7 +320,7 @@ void r600_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mod struct radeon_device *rdev = dev->dev_private; uint32_t offset = to_radeon_encoder(encoder)->hdmi_offset; - if (ASIC_IS_DCE4(rdev)) + if (ASIC_IS_DCE5(rdev)) return; if (!offset) @@ -462,13 +462,31 @@ static void r600_hdmi_assign_block(struct drm_encoder *encoder) struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv; + u16 eg_offsets[] = { + EVERGREEN_CRTC0_REGISTER_OFFSET, + EVERGREEN_CRTC1_REGISTER_OFFSET, + EVERGREEN_CRTC2_REGISTER_OFFSET, + EVERGREEN_CRTC3_REGISTER_OFFSET, + EVERGREEN_CRTC4_REGISTER_OFFSET, + EVERGREEN_CRTC5_REGISTER_OFFSET, + }; + if (!dig) { dev_err(rdev->dev, "Enabling HDMI on non-dig encoder\n"); return; } - if (ASIC_IS_DCE4(rdev)) { + if (ASIC_IS_DCE5(rdev)) { /* TODO */ + } else if (ASIC_IS_DCE4(rdev)) { + if (dig->dig_encoder >= ARRAY_SIZE(eg_offsets)) { + dev_err(rdev->dev, "Enabling HDMI on unknown dig\n"); + return; + } + radeon_encoder->hdmi_offset = EVERGREEN_HDMI_BASE + + eg_offsets[dig->dig_encoder]; + radeon_encoder->hdmi_config_offset = radeon_encoder->hdmi_offset + + EVERGREEN_HDMI_CONFIG_OFFSET; } else if (ASIC_IS_DCE3(rdev)) { radeon_encoder->hdmi_offset = dig->dig_encoder ? R600_HDMI_BLOCK3 : R600_HDMI_BLOCK1; @@ -491,7 +509,7 @@ void r600_hdmi_enable(struct drm_encoder *encoder) struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); uint32_t offset; - if (ASIC_IS_DCE4(rdev)) + if (ASIC_IS_DCE5(rdev)) return; if (!radeon_encoder->hdmi_offset) { @@ -504,16 +522,24 @@ void r600_hdmi_enable(struct drm_encoder *encoder) } offset = radeon_encoder->hdmi_offset; - if (ASIC_IS_DCE32(rdev) && !ASIC_IS_DCE4(rdev)) { + if (ASIC_IS_DCE5(rdev)) { + /* TODO */ + } else if (ASIC_IS_DCE4(rdev)) { + WREG32_P(radeon_encoder->hdmi_config_offset + 0xc, 0x1, ~0x1); + } else if (ASIC_IS_DCE32(rdev)) { WREG32_P(radeon_encoder->hdmi_config_offset + 0x4, 0x1, ~0x1); - } else if (rdev->family >= CHIP_R600 && !ASIC_IS_DCE3(rdev)) { + } else if (ASIC_IS_DCE3(rdev)) { + /* TODO */ + } else if (rdev->family >= CHIP_R600) { switch (radeon_encoder->encoder_id) { case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1: - WREG32_P(AVIVO_TMDSA_CNTL, 0x4, ~0x4); + WREG32_P(AVIVO_TMDSA_CNTL, AVIVO_TMDSA_CNTL_HDMI_EN, + ~AVIVO_TMDSA_CNTL_HDMI_EN); WREG32(offset + R600_HDMI_ENABLE, 0x101); break; case ENCODER_OBJECT_ID_INTERNAL_LVTM1: - WREG32_P(AVIVO_LVTMA_CNTL, 0x4, ~0x4); + WREG32_P(AVIVO_LVTMA_CNTL, AVIVO_LVTMA_CNTL_HDMI_EN, + ~AVIVO_LVTMA_CNTL_HDMI_EN); WREG32(offset + R600_HDMI_ENABLE, 0x105); break; default: @@ -525,8 +551,8 @@ void r600_hdmi_enable(struct drm_encoder *encoder) if (rdev->irq.installed && rdev->family != CHIP_RS600 && rdev->family != CHIP_RS690 - && rdev->family != CHIP_RS740) { - + && rdev->family != CHIP_RS740 + && !ASIC_IS_DCE4(rdev)) { /* if irq is available use it */ rdev->irq.hdmi[offset == R600_HDMI_BLOCK1 ? 0 : 1] = true; radeon_irq_set(rdev); @@ -551,7 +577,7 @@ void r600_hdmi_disable(struct drm_encoder *encoder) struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); uint32_t offset; - if (ASIC_IS_DCE4(rdev)) + if (ASIC_IS_DCE5(rdev)) return; offset = radeon_encoder->hdmi_offset; @@ -570,16 +596,22 @@ void r600_hdmi_disable(struct drm_encoder *encoder) /* disable polling */ r600_audio_disable_polling(encoder); - if (ASIC_IS_DCE32(rdev) && !ASIC_IS_DCE4(rdev)) { + if (ASIC_IS_DCE5(rdev)) { + /* TODO */ + } else if (ASIC_IS_DCE4(rdev)) { + WREG32_P(radeon_encoder->hdmi_config_offset + 0xc, 0, ~0x1); + } else if (ASIC_IS_DCE32(rdev)) { WREG32_P(radeon_encoder->hdmi_config_offset + 0x4, 0, ~0x1); } else if (rdev->family >= CHIP_R600 && !ASIC_IS_DCE3(rdev)) { switch (radeon_encoder->encoder_id) { case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1: - WREG32_P(AVIVO_TMDSA_CNTL, 0, ~0x4); + WREG32_P(AVIVO_TMDSA_CNTL, 0, + ~AVIVO_TMDSA_CNTL_HDMI_EN); WREG32(offset + R600_HDMI_ENABLE, 0); break; case ENCODER_OBJECT_ID_INTERNAL_LVTM1: - WREG32_P(AVIVO_LVTMA_CNTL, 0, ~0x4); + WREG32_P(AVIVO_LVTMA_CNTL, 0, + ~AVIVO_LVTMA_CNTL_HDMI_EN); WREG32(offset + R600_HDMI_ENABLE, 0); break; default: diff --git a/drivers/gpu/drm/radeon/r600d.h b/drivers/gpu/drm/radeon/r600d.h index d4d23a8d..12ceb82 100644 --- a/drivers/gpu/drm/radeon/r600d.h +++ b/drivers/gpu/drm/radeon/r600d.h @@ -78,6 +78,20 @@ #define CB_COLOR0_SIZE 0x28060 #define CB_COLOR0_VIEW 0x28080 +#define R_028080_CB_COLOR0_VIEW 0x028080 +#define S_028080_SLICE_START(x) (((x) & 0x7FF) << 0) +#define G_028080_SLICE_START(x) (((x) >> 0) & 0x7FF) +#define C_028080_SLICE_START 0xFFFFF800 +#define S_028080_SLICE_MAX(x) (((x) & 0x7FF) << 13) +#define G_028080_SLICE_MAX(x) (((x) >> 13) & 0x7FF) +#define C_028080_SLICE_MAX 0xFF001FFF +#define R_028084_CB_COLOR1_VIEW 0x028084 +#define R_028088_CB_COLOR2_VIEW 0x028088 +#define R_02808C_CB_COLOR3_VIEW 0x02808C +#define R_028090_CB_COLOR4_VIEW 0x028090 +#define R_028094_CB_COLOR5_VIEW 0x028094 +#define R_028098_CB_COLOR6_VIEW 0x028098 +#define R_02809C_CB_COLOR7_VIEW 0x02809C #define CB_COLOR0_INFO 0x280a0 # define CB_FORMAT(x) ((x) << 2) # define CB_ARRAY_MODE(x) ((x) << 8) @@ -181,6 +195,14 @@ #define PREZ_MUST_WAIT_FOR_POSTZ_DONE (1 << 31) #define DB_DEPTH_BASE 0x2800C #define DB_HTILE_DATA_BASE 0x28014 +#define DB_HTILE_SURFACE 0x28D24 +#define S_028D24_HTILE_WIDTH(x) (((x) & 0x1) << 0) +#define G_028D24_HTILE_WIDTH(x) (((x) >> 0) & 0x1) +#define C_028D24_HTILE_WIDTH 0xFFFFFFFE +#define S_028D24_HTILE_HEIGHT(x) (((x) & 0x1) << 1) +#define G_028D24_HTILE_HEIGHT(x) (((x) >> 1) & 0x1) +#define C_028D24_HTILE_HEIGHT 0xFFFFFFFD +#define G_028D24_LINEAR(x) (((x) >> 2) & 0x1) #define DB_WATERMARKS 0x9838 #define DEPTH_FREE(x) ((x) << 0) #define DEPTH_FLUSH(x) ((x) << 5) @@ -494,6 +516,11 @@ #define VGT_STRMOUT_BUFFER_OFFSET_1 0x28AEC #define VGT_STRMOUT_BUFFER_OFFSET_2 0x28AFC #define VGT_STRMOUT_BUFFER_OFFSET_3 0x28B0C +#define VGT_STRMOUT_BUFFER_SIZE_0 0x28AD0 +#define VGT_STRMOUT_BUFFER_SIZE_1 0x28AE0 +#define VGT_STRMOUT_BUFFER_SIZE_2 0x28AF0 +#define VGT_STRMOUT_BUFFER_SIZE_3 0x28B00 + #define VGT_STRMOUT_EN 0x28AB0 #define VGT_VERTEX_REUSE_BLOCK_CNTL 0x28C58 #define VTX_REUSE_DEPTH_MASK 0x000000FF @@ -575,6 +602,10 @@ #define RLC_UCODE_ADDR 0x3f2c #define RLC_UCODE_DATA 0x3f30 +/* new for TN */ +#define TN_RLC_SAVE_AND_RESTORE_BASE 0x3f10 +#define TN_RLC_CLEAR_STATE_RESTORE_BASE 0x3f20 + #define SRBM_SOFT_RESET 0xe60 # define SOFT_RESET_RLC (1 << 13) @@ -832,7 +863,11 @@ #define PACKET3_STRMOUT_BUFFER_UPDATE 0x34 #define PACKET3_INDIRECT_BUFFER_MP 0x38 #define PACKET3_MEM_SEMAPHORE 0x39 +# define PACKET3_SEM_WAIT_ON_SIGNAL (0x1 << 12) +# define PACKET3_SEM_SEL_SIGNAL (0x6 << 29) +# define PACKET3_SEM_SEL_WAIT (0x7 << 29) #define PACKET3_MPEG_INDEX 0x3A +#define PACKET3_COPY_DW 0x3B #define PACKET3_WAIT_REG_MEM 0x3C #define PACKET3_MEM_WRITE 0x3D #define PACKET3_INDIRECT_BUFFER 0x32 diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index 28e69e9..66150f0 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -107,6 +107,21 @@ extern int radeon_msi; #define RADEONFB_CONN_LIMIT 4 #define RADEON_BIOS_NUM_SCRATCH 8 +/* max number of rings */ +#define RADEON_NUM_RINGS 3 + +/* internal ring indices */ +/* r1xx+ has gfx CP ring */ +#define RADEON_RING_TYPE_GFX_INDEX 0 + +/* cayman has 2 compute CP rings */ +#define CAYMAN_RING_TYPE_CP1_INDEX 1 +#define CAYMAN_RING_TYPE_CP2_INDEX 2 + +/* hardcode those limit for now */ +#define RADEON_VA_RESERVED_SIZE (8 << 20) +#define RADEON_IB_VM_MAX_SIZE (64 << 10) + /* * Errata workarounds. */ @@ -127,6 +142,47 @@ bool radeon_get_bios(struct radeon_device *rdev); /* + * Mutex which allows recursive locking from the same process. + */ +struct radeon_mutex { + struct mutex mutex; + struct task_struct *owner; + int level; +}; + +static inline void radeon_mutex_init(struct radeon_mutex *mutex) +{ + mutex_init(&mutex->mutex); + mutex->owner = NULL; + mutex->level = 0; +} + +static inline void radeon_mutex_lock(struct radeon_mutex *mutex) +{ + if (mutex_trylock(&mutex->mutex)) { + /* The mutex was unlocked before, so it's ours now */ + mutex->owner = current; + } else if (mutex->owner != current) { + /* Another process locked the mutex, take it */ + mutex_lock(&mutex->mutex); + mutex->owner = current; + } + /* Otherwise the mutex was already locked by this process */ + + mutex->level++; +} + +static inline void radeon_mutex_unlock(struct radeon_mutex *mutex) +{ + if (--mutex->level > 0) + return; + + mutex->owner = NULL; + mutex_unlock(&mutex->mutex); +} + + +/* * Dummy page */ struct radeon_dummy_page { @@ -165,26 +221,30 @@ void radeon_pm_resume(struct radeon_device *rdev); void radeon_combios_get_power_modes(struct radeon_device *rdev); void radeon_atombios_get_power_modes(struct radeon_device *rdev); void radeon_atom_set_voltage(struct radeon_device *rdev, u16 voltage_level, u8 voltage_type); -int radeon_atom_get_max_vddc(struct radeon_device *rdev, u16 *voltage); void rs690_pm_info(struct radeon_device *rdev); extern int rv6xx_get_temp(struct radeon_device *rdev); extern int rv770_get_temp(struct radeon_device *rdev); extern int evergreen_get_temp(struct radeon_device *rdev); extern int sumo_get_temp(struct radeon_device *rdev); +extern int si_get_temp(struct radeon_device *rdev); +extern void evergreen_tiling_fields(unsigned tiling_flags, unsigned *bankw, + unsigned *bankh, unsigned *mtaspect, + unsigned *tile_split); /* * Fences. */ struct radeon_fence_driver { uint32_t scratch_reg; + uint64_t gpu_addr; + volatile uint32_t *cpu_addr; atomic_t seq; uint32_t last_seq; unsigned long last_jiffies; unsigned long last_timeout; wait_queue_head_t queue; - rwlock_t lock; struct list_head created; - struct list_head emited; + struct list_head emitted; struct list_head signaled; bool initialized; }; @@ -195,21 +255,26 @@ struct radeon_fence { struct list_head list; /* protected by radeon_fence.lock */ uint32_t seq; - bool emited; + bool emitted; bool signaled; + /* RB, DMA, etc. */ + int ring; + struct radeon_semaphore *semaphore; }; +int radeon_fence_driver_start_ring(struct radeon_device *rdev, int ring); int radeon_fence_driver_init(struct radeon_device *rdev); void radeon_fence_driver_fini(struct radeon_device *rdev); -int radeon_fence_create(struct radeon_device *rdev, struct radeon_fence **fence); +int radeon_fence_create(struct radeon_device *rdev, struct radeon_fence **fence, int ring); int radeon_fence_emit(struct radeon_device *rdev, struct radeon_fence *fence); -void radeon_fence_process(struct radeon_device *rdev); +void radeon_fence_process(struct radeon_device *rdev, int ring); bool radeon_fence_signaled(struct radeon_fence *fence); int radeon_fence_wait(struct radeon_fence *fence, bool interruptible); -int radeon_fence_wait_next(struct radeon_device *rdev); -int radeon_fence_wait_last(struct radeon_device *rdev); +int radeon_fence_wait_next(struct radeon_device *rdev, int ring); +int radeon_fence_wait_last(struct radeon_device *rdev, int ring); struct radeon_fence *radeon_fence_ref(struct radeon_fence *fence); void radeon_fence_unref(struct radeon_fence **fence); +int radeon_fence_count_emitted(struct radeon_device *rdev, int ring); /* * Tiling registers @@ -231,6 +296,21 @@ struct radeon_mman { bool initialized; }; +/* bo virtual address in a specific vm */ +struct radeon_bo_va { + /* bo list is protected by bo being reserved */ + struct list_head bo_list; + /* vm list is protected by vm mutex */ + struct list_head vm_list; + /* constant after initialization */ + struct radeon_vm *vm; + struct radeon_bo *bo; + uint64_t soffset; + uint64_t eoffset; + uint32_t flags; + bool valid; +}; + struct radeon_bo { /* Protected by gem.mutex */ struct list_head list; @@ -244,6 +324,10 @@ struct radeon_bo { u32 tiling_flags; u32 pitch; int surface_reg; + /* list of all virtual address to which this bo + * is associated to + */ + struct list_head va; /* Constant after initialization */ struct radeon_device *rdev; struct drm_gem_object gem_base; @@ -259,6 +343,48 @@ struct radeon_bo_list { u32 tiling_flags; }; +/* sub-allocation manager, it has to be protected by another lock. + * By conception this is an helper for other part of the driver + * like the indirect buffer or semaphore, which both have their + * locking. + * + * Principe is simple, we keep a list of sub allocation in offset + * order (first entry has offset == 0, last entry has the highest + * offset). + * + * When allocating new object we first check if there is room at + * the end total_size - (last_object_offset + last_object_size) >= + * alloc_size. If so we allocate new object there. + * + * When there is not enough room at the end, we start waiting for + * each sub object until we reach object_offset+object_size >= + * alloc_size, this object then become the sub object we return. + * + * Alignment can't be bigger than page size. + * + * Hole are not considered for allocation to keep things simple. + * Assumption is that there won't be hole (all object on same + * alignment). + */ +struct radeon_sa_manager { + struct radeon_bo *bo; + struct list_head sa_bo; + unsigned size; + uint64_t gpu_addr; + void *cpu_ptr; + uint32_t domain; +}; + +struct radeon_sa_bo; + +/* sub-allocation buffer */ +struct radeon_sa_bo { + struct list_head list; + struct radeon_sa_manager *manager; + unsigned offset; + unsigned size; +}; + /* * GEM objects. */ @@ -273,9 +399,6 @@ int radeon_gem_object_create(struct radeon_device *rdev, int size, int alignment, int initial_domain, bool discardable, bool kernel, struct drm_gem_object **obj); -int radeon_gem_object_pin(struct drm_gem_object *obj, uint32_t pin_domain, - uint64_t *gpu_addr); -void radeon_gem_object_unpin(struct drm_gem_object *obj); int radeon_mode_dumb_create(struct drm_file *file_priv, struct drm_device *dev, @@ -288,6 +411,46 @@ int radeon_mode_dumb_destroy(struct drm_file *file_priv, uint32_t handle); /* + * Semaphores. + */ +struct radeon_ring; + +#define RADEON_SEMAPHORE_BO_SIZE 256 + +struct radeon_semaphore_driver { + rwlock_t lock; + struct list_head bo; +}; + +struct radeon_semaphore_bo; + +/* everything here is constant */ +struct radeon_semaphore { + struct list_head list; + uint64_t gpu_addr; + uint32_t *cpu_ptr; + struct radeon_semaphore_bo *bo; +}; + +struct radeon_semaphore_bo { + struct list_head list; + struct radeon_ib *ib; + struct list_head free; + struct radeon_semaphore semaphores[RADEON_SEMAPHORE_BO_SIZE/8]; + unsigned nused; +}; + +void radeon_semaphore_driver_fini(struct radeon_device *rdev); +int radeon_semaphore_create(struct radeon_device *rdev, + struct radeon_semaphore **semaphore); +void radeon_semaphore_emit_signal(struct radeon_device *rdev, int ring, + struct radeon_semaphore *semaphore); +void radeon_semaphore_emit_wait(struct radeon_device *rdev, int ring, + struct radeon_semaphore *semaphore); +void radeon_semaphore_free(struct radeon_device *rdev, + struct radeon_semaphore *semaphore); + +/* * GART structures, functions & helpers */ struct radeon_mc; @@ -295,6 +458,7 @@ struct radeon_mc; #define RADEON_GPU_PAGE_SIZE 4096 #define RADEON_GPU_PAGE_MASK (RADEON_GPU_PAGE_SIZE - 1) #define RADEON_GPU_PAGE_SHIFT 12 +#define RADEON_GPU_PAGE_ALIGN(a) (((a) + RADEON_GPU_PAGE_MASK) & ~RADEON_GPU_PAGE_MASK) struct radeon_gart { dma_addr_t table_addr; @@ -305,7 +469,6 @@ struct radeon_gart { unsigned table_size; struct page **pages; dma_addr_t *pages_addr; - bool *ttm_alloced; bool ready; }; @@ -419,7 +582,7 @@ union radeon_irq_stat_regs { struct radeon_irq { bool installed; - bool sw_int; + bool sw_int[RADEON_NUM_RINGS]; bool crtc_vblank_int[RADEON_MAX_CRTCS]; bool pflip[RADEON_MAX_CRTCS]; wait_queue_head_t vblank_queue; @@ -429,7 +592,7 @@ struct radeon_irq { wait_queue_head_t idle_queue; bool hdmi[RADEON_MAX_HDMI_BLOCKS]; spinlock_t sw_lock; - int sw_refcount; + int sw_refcount[RADEON_NUM_RINGS]; union radeon_irq_stat_regs stat_regs; spinlock_t pflip_lock[RADEON_MAX_CRTCS]; int pflip_refcount[RADEON_MAX_CRTCS]; @@ -437,22 +600,24 @@ struct radeon_irq { int radeon_irq_kms_init(struct radeon_device *rdev); void radeon_irq_kms_fini(struct radeon_device *rdev); -void radeon_irq_kms_sw_irq_get(struct radeon_device *rdev); -void radeon_irq_kms_sw_irq_put(struct radeon_device *rdev); +void radeon_irq_kms_sw_irq_get(struct radeon_device *rdev, int ring); +void radeon_irq_kms_sw_irq_put(struct radeon_device *rdev, int ring); void radeon_irq_kms_pflip_irq_get(struct radeon_device *rdev, int crtc); void radeon_irq_kms_pflip_irq_put(struct radeon_device *rdev, int crtc); /* - * CP & ring. + * CP & rings. */ + struct radeon_ib { - struct list_head list; + struct radeon_sa_bo sa_bo; unsigned idx; + uint32_t length_dw; uint64_t gpu_addr; - struct radeon_fence *fence; uint32_t *ptr; - uint32_t length_dw; - bool free; + struct radeon_fence *fence; + unsigned vm_id; + bool is_const_ib; }; /* @@ -460,20 +625,22 @@ struct radeon_ib { * mutex protects scheduled_ibs, ready, alloc_bm */ struct radeon_ib_pool { - struct mutex mutex; - struct radeon_bo *robj; - struct list_head bogus_ib; - struct radeon_ib ibs[RADEON_IB_POOL_SIZE]; - bool ready; - unsigned head_id; + struct radeon_mutex mutex; + struct radeon_sa_manager sa_manager; + struct radeon_ib ibs[RADEON_IB_POOL_SIZE]; + bool ready; + unsigned head_id; }; -struct radeon_cp { +struct radeon_ring { struct radeon_bo *ring_obj; volatile uint32_t *ring; unsigned rptr; + unsigned rptr_offs; + unsigned rptr_reg; unsigned wptr; unsigned wptr_old; + unsigned wptr_reg; unsigned ring_size; unsigned ring_free_dw; int count_dw; @@ -482,6 +649,61 @@ struct radeon_cp { uint32_t ptr_mask; struct mutex mutex; bool ready; + u32 ptr_reg_shift; + u32 ptr_reg_mask; + u32 nop; +}; + +/* + * VM + */ +struct radeon_vm { + struct list_head list; + struct list_head va; + int id; + unsigned last_pfn; + u64 pt_gpu_addr; + u64 *pt; + struct radeon_sa_bo sa_bo; + struct mutex mutex; + /* last fence for cs using this vm */ + struct radeon_fence *fence; +}; + +struct radeon_vm_funcs { + int (*init)(struct radeon_device *rdev); + void (*fini)(struct radeon_device *rdev); + /* cs mutex must be lock for schedule_ib */ + int (*bind)(struct radeon_device *rdev, struct radeon_vm *vm, int id); + void (*unbind)(struct radeon_device *rdev, struct radeon_vm *vm); + void (*tlb_flush)(struct radeon_device *rdev, struct radeon_vm *vm); + uint32_t (*page_flags)(struct radeon_device *rdev, + struct radeon_vm *vm, + uint32_t flags); + void (*set_page)(struct radeon_device *rdev, struct radeon_vm *vm, + unsigned pfn, uint64_t addr, uint32_t flags); +}; + +struct radeon_vm_manager { + struct list_head lru_vm; + uint32_t use_bitmap; + struct radeon_sa_manager sa_manager; + uint32_t max_pfn; + /* fields constant after init */ + const struct radeon_vm_funcs *funcs; + /* number of VMIDs */ + unsigned nvm; + /* vram base address for page table entry */ + u64 vram_base_offset; + /* is vm enabled? */ + bool enabled; +}; + +/* + * file private structure + */ +struct radeon_fpriv { + struct radeon_vm vm; }; /* @@ -491,6 +713,7 @@ struct r600_ih { struct radeon_bo *ring_obj; volatile uint32_t *ring; unsigned rptr; + unsigned rptr_offs; unsigned wptr; unsigned wptr_old; unsigned ring_size; @@ -534,23 +757,40 @@ struct r600_blit { void r600_blit_suspend(struct radeon_device *rdev); -int radeon_ib_get(struct radeon_device *rdev, struct radeon_ib **ib); +/* + * SI RLC stuff + */ +struct si_rlc { + /* for power gating */ + struct radeon_bo *save_restore_obj; + uint64_t save_restore_gpu_addr; + /* for clear state */ + struct radeon_bo *clear_state_obj; + uint64_t clear_state_gpu_addr; +}; + +int radeon_ib_get(struct radeon_device *rdev, int ring, + struct radeon_ib **ib, unsigned size); void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib **ib); +bool radeon_ib_try_free(struct radeon_device *rdev, struct radeon_ib *ib); int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib); int radeon_ib_pool_init(struct radeon_device *rdev); void radeon_ib_pool_fini(struct radeon_device *rdev); -int radeon_ib_test(struct radeon_device *rdev); -extern void radeon_ib_bogus_add(struct radeon_device *rdev, struct radeon_ib *ib); +int radeon_ib_pool_start(struct radeon_device *rdev); +int radeon_ib_pool_suspend(struct radeon_device *rdev); /* Ring access between begin & end cannot sleep */ -void radeon_ring_free_size(struct radeon_device *rdev); -int radeon_ring_alloc(struct radeon_device *rdev, unsigned ndw); -int radeon_ring_lock(struct radeon_device *rdev, unsigned ndw); -void radeon_ring_commit(struct radeon_device *rdev); -void radeon_ring_unlock_commit(struct radeon_device *rdev); -void radeon_ring_unlock_undo(struct radeon_device *rdev); -int radeon_ring_test(struct radeon_device *rdev); -int radeon_ring_init(struct radeon_device *rdev, unsigned ring_size); -void radeon_ring_fini(struct radeon_device *rdev); +int radeon_ring_index(struct radeon_device *rdev, struct radeon_ring *cp); +void radeon_ring_free_size(struct radeon_device *rdev, struct radeon_ring *cp); +int radeon_ring_alloc(struct radeon_device *rdev, struct radeon_ring *cp, unsigned ndw); +int radeon_ring_lock(struct radeon_device *rdev, struct radeon_ring *cp, unsigned ndw); +void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *cp); +void radeon_ring_unlock_commit(struct radeon_device *rdev, struct radeon_ring *cp); +void radeon_ring_unlock_undo(struct radeon_device *rdev, struct radeon_ring *cp); +int radeon_ring_test(struct radeon_device *rdev, struct radeon_ring *cp); +int radeon_ring_init(struct radeon_device *rdev, struct radeon_ring *cp, unsigned ring_size, + unsigned rptr_offs, unsigned rptr_reg, unsigned wptr_reg, + u32 ptr_reg_shift, u32 ptr_reg_mask, u32 nop); +void radeon_ring_fini(struct radeon_device *rdev, struct radeon_ring *cp); /* @@ -567,12 +807,12 @@ struct radeon_cs_reloc { struct radeon_cs_chunk { uint32_t chunk_id; uint32_t length_dw; - int kpage_idx[2]; - uint32_t *kpage[2]; + int kpage_idx[2]; + uint32_t *kpage[2]; uint32_t *kdata; - void __user *user_ptr; - int last_copied_page; - int last_page_index; + void __user *user_ptr; + int last_copied_page; + int last_page_index; }; struct radeon_cs_parser { @@ -593,11 +833,16 @@ struct radeon_cs_parser { /* indices of various chunks */ int chunk_ib_idx; int chunk_relocs_idx; + int chunk_flags_idx; + int chunk_const_ib_idx; struct radeon_ib *ib; + struct radeon_ib *const_ib; void *track; unsigned family; int parser_error; - bool keep_tiling_flags; + u32 cs_flags; + u32 ring; + s32 priority; }; extern int radeon_cs_update_pages(struct radeon_cs_parser *p, int pg_idx); @@ -733,6 +978,7 @@ enum radeon_int_thermal_type { THERMAL_TYPE_EVERGREEN, THERMAL_TYPE_SUMO, THERMAL_TYPE_NI, + THERMAL_TYPE_SI, }; struct radeon_voltage { @@ -854,11 +1100,20 @@ void radeon_benchmark(struct radeon_device *rdev, int test_number); * Testing */ void radeon_test_moves(struct radeon_device *rdev); +void radeon_test_ring_sync(struct radeon_device *rdev, + struct radeon_ring *cpA, + struct radeon_ring *cpB); +void radeon_test_syncing(struct radeon_device *rdev); /* * Debugfs */ +struct radeon_debugfs { + struct drm_info_list *files; + unsigned num_files; +}; + int radeon_debugfs_add_files(struct radeon_device *rdev, struct drm_info_list *files, unsigned nfiles); @@ -874,53 +1129,8 @@ struct radeon_asic { int (*resume)(struct radeon_device *rdev); int (*suspend)(struct radeon_device *rdev); void (*vga_set_state)(struct radeon_device *rdev, bool state); - bool (*gpu_is_lockup)(struct radeon_device *rdev); + bool (*gpu_is_lockup)(struct radeon_device *rdev, struct radeon_ring *cp); int (*asic_reset)(struct radeon_device *rdev); - void (*gart_tlb_flush)(struct radeon_device *rdev); - int (*gart_set_page)(struct radeon_device *rdev, int i, uint64_t addr); - int (*cp_init)(struct radeon_device *rdev, unsigned ring_size); - void (*cp_fini)(struct radeon_device *rdev); - void (*cp_disable)(struct radeon_device *rdev); - void (*cp_commit)(struct radeon_device *rdev); - void (*ring_start)(struct radeon_device *rdev); - int (*ring_test)(struct radeon_device *rdev); - void (*ring_ib_execute)(struct radeon_device *rdev, struct radeon_ib *ib); - int (*irq_set)(struct radeon_device *rdev); - int (*irq_process)(struct radeon_device *rdev); - u32 (*get_vblank_counter)(struct radeon_device *rdev, int crtc); - void (*fence_ring_emit)(struct radeon_device *rdev, struct radeon_fence *fence); - int (*cs_parse)(struct radeon_cs_parser *p); - int (*copy_blit)(struct radeon_device *rdev, - uint64_t src_offset, - uint64_t dst_offset, - unsigned num_gpu_pages, - struct radeon_fence *fence); - int (*copy_dma)(struct radeon_device *rdev, - uint64_t src_offset, - uint64_t dst_offset, - unsigned num_gpu_pages, - struct radeon_fence *fence); - int (*copy)(struct radeon_device *rdev, - uint64_t src_offset, - uint64_t dst_offset, - unsigned num_gpu_pages, - struct radeon_fence *fence); - uint32_t (*get_engine_clock)(struct radeon_device *rdev); - void (*set_engine_clock)(struct radeon_device *rdev, uint32_t eng_clock); - uint32_t (*get_memory_clock)(struct radeon_device *rdev); - void (*set_memory_clock)(struct radeon_device *rdev, uint32_t mem_clock); - int (*get_pcie_lanes)(struct radeon_device *rdev); - void (*set_pcie_lanes)(struct radeon_device *rdev, int lanes); - void (*set_clock_gating)(struct radeon_device *rdev, int enable); - int (*set_surface_reg)(struct radeon_device *rdev, int reg, - uint32_t tiling_flags, uint32_t pitch, - uint32_t offset, uint32_t obj_size); - void (*clear_surface_reg)(struct radeon_device *rdev, int reg); - void (*bandwidth_update)(struct radeon_device *rdev); - void (*hpd_init)(struct radeon_device *rdev); - void (*hpd_fini)(struct radeon_device *rdev); - bool (*hpd_sense)(struct radeon_device *rdev, enum radeon_hpd_id hpd); - void (*hpd_set_polarity)(struct radeon_device *rdev, enum radeon_hpd_id hpd); /* ioctl hw specific callback. Some hw might want to perform special * operation on specific ioctl. For instance on wait idle some hw * might want to perform and HDP flush through MMIO as it seems that @@ -928,17 +1138,99 @@ struct radeon_asic { * through ring. */ void (*ioctl_wait_idle)(struct radeon_device *rdev, struct radeon_bo *bo); + /* check if 3D engine is idle */ bool (*gui_idle)(struct radeon_device *rdev); + /* wait for mc_idle */ + int (*mc_wait_for_idle)(struct radeon_device *rdev); + /* gart */ + struct { + void (*tlb_flush)(struct radeon_device *rdev); + int (*set_page)(struct radeon_device *rdev, int i, uint64_t addr); + } gart; + /* ring specific callbacks */ + struct { + void (*ib_execute)(struct radeon_device *rdev, struct radeon_ib *ib); + int (*ib_parse)(struct radeon_device *rdev, struct radeon_ib *ib); + void (*emit_fence)(struct radeon_device *rdev, struct radeon_fence *fence); + void (*emit_semaphore)(struct radeon_device *rdev, struct radeon_ring *cp, + struct radeon_semaphore *semaphore, bool emit_wait); + int (*cs_parse)(struct radeon_cs_parser *p); + void (*ring_start)(struct radeon_device *rdev, struct radeon_ring *cp); + int (*ring_test)(struct radeon_device *rdev, struct radeon_ring *cp); + int (*ib_test)(struct radeon_device *rdev, struct radeon_ring *cp); + } ring[RADEON_NUM_RINGS]; + /* irqs */ + struct { + int (*set)(struct radeon_device *rdev); + int (*process)(struct radeon_device *rdev); + } irq; + /* displays */ + struct { + /* display watermarks */ + void (*bandwidth_update)(struct radeon_device *rdev); + /* get frame count */ + u32 (*get_vblank_counter)(struct radeon_device *rdev, int crtc); + /* wait for vblank */ + void (*wait_for_vblank)(struct radeon_device *rdev, int crtc); + } display; + /* copy functions for bo handling */ + struct { + int (*blit)(struct radeon_device *rdev, + uint64_t src_offset, + uint64_t dst_offset, + unsigned num_gpu_pages, + struct radeon_fence *fence); + u32 blit_ring_index; + int (*dma)(struct radeon_device *rdev, + uint64_t src_offset, + uint64_t dst_offset, + unsigned num_gpu_pages, + struct radeon_fence *fence); + u32 dma_ring_index; + /* method used for bo copy */ + int (*copy)(struct radeon_device *rdev, + uint64_t src_offset, + uint64_t dst_offset, + unsigned num_gpu_pages, + struct radeon_fence *fence); + /* ring used for bo copies */ + u32 copy_ring_index; + } copy; + /* surfaces */ + struct { + int (*set_reg)(struct radeon_device *rdev, int reg, + uint32_t tiling_flags, uint32_t pitch, + uint32_t offset, uint32_t obj_size); + void (*clear_reg)(struct radeon_device *rdev, int reg); + } surface; + /* hotplug detect */ + struct { + void (*init)(struct radeon_device *rdev); + void (*fini)(struct radeon_device *rdev); + bool (*sense)(struct radeon_device *rdev, enum radeon_hpd_id hpd); + void (*set_polarity)(struct radeon_device *rdev, enum radeon_hpd_id hpd); + } hpd; /* power management */ - void (*pm_misc)(struct radeon_device *rdev); - void (*pm_prepare)(struct radeon_device *rdev); - void (*pm_finish)(struct radeon_device *rdev); - void (*pm_init_profile)(struct radeon_device *rdev); - void (*pm_get_dynpm_state)(struct radeon_device *rdev); + struct { + void (*misc)(struct radeon_device *rdev); + void (*prepare)(struct radeon_device *rdev); + void (*finish)(struct radeon_device *rdev); + void (*init_profile)(struct radeon_device *rdev); + void (*get_dynpm_state)(struct radeon_device *rdev); + uint32_t (*get_engine_clock)(struct radeon_device *rdev); + void (*set_engine_clock)(struct radeon_device *rdev, uint32_t eng_clock); + uint32_t (*get_memory_clock)(struct radeon_device *rdev); + void (*set_memory_clock)(struct radeon_device *rdev, uint32_t mem_clock); + int (*get_pcie_lanes)(struct radeon_device *rdev); + void (*set_pcie_lanes)(struct radeon_device *rdev, int lanes); + void (*set_clock_gating)(struct radeon_device *rdev, int enable); + } pm; /* pageflipping */ - void (*pre_page_flip)(struct radeon_device *rdev, int crtc); - u32 (*page_flip)(struct radeon_device *rdev, int crtc, u64 crtc_base); - void (*post_page_flip)(struct radeon_device *rdev, int crtc); + struct { + void (*pre_page_flip)(struct radeon_device *rdev, int crtc); + u32 (*page_flip)(struct radeon_device *rdev, int crtc, u64 crtc_base); + void (*post_page_flip)(struct radeon_device *rdev, int crtc); + } pflip; }; /* @@ -1078,6 +1370,37 @@ struct cayman_asic { struct r100_gpu_lockup lockup; }; +struct si_asic { + unsigned max_shader_engines; + unsigned max_pipes_per_simd; + unsigned max_tile_pipes; + unsigned max_simds_per_se; + unsigned max_backends_per_se; + unsigned max_texture_channel_caches; + unsigned max_gprs; + unsigned max_gs_threads; + unsigned max_hw_contexts; + unsigned sc_prim_fifo_size_frontend; + unsigned sc_prim_fifo_size_backend; + unsigned sc_hiz_tile_fifo_size; + unsigned sc_earlyz_tile_fifo_size; + + unsigned num_shader_engines; + unsigned num_tile_pipes; + unsigned num_backends_per_se; + unsigned backend_disable_mask_per_asic; + unsigned backend_map; + unsigned num_texture_channel_caches; + unsigned mem_max_burst_length_bytes; + unsigned mem_row_size_in_kb; + unsigned shader_engine_tile_size; + unsigned num_gpus; + unsigned multi_gpu_tile_size; + + unsigned tile_config; + struct r100_gpu_lockup lockup; +}; + union radeon_asic_config { struct r300_asic r300; struct r100_asic r100; @@ -1085,6 +1408,7 @@ union radeon_asic_config { struct rv770_asic rv770; struct evergreen_asic evergreen; struct cayman_asic cayman; + struct si_asic si; }; /* @@ -1117,6 +1441,8 @@ int radeon_gem_busy_ioctl(struct drm_device *dev, void *data, struct drm_file *filp); int radeon_gem_wait_idle_ioctl(struct drm_device *dev, void *data, struct drm_file *filp); +int radeon_gem_va_ioctl(struct drm_device *dev, void *data, + struct drm_file *filp); int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp); int radeon_gem_set_tiling_ioctl(struct drm_device *dev, void *data, struct drm_file *filp); @@ -1132,47 +1458,6 @@ struct r600_vram_scratch { /* - * Mutex which allows recursive locking from the same process. - */ -struct radeon_mutex { - struct mutex mutex; - struct task_struct *owner; - int level; -}; - -static inline void radeon_mutex_init(struct radeon_mutex *mutex) -{ - mutex_init(&mutex->mutex); - mutex->owner = NULL; - mutex->level = 0; -} - -static inline void radeon_mutex_lock(struct radeon_mutex *mutex) -{ - if (mutex_trylock(&mutex->mutex)) { - /* The mutex was unlocked before, so it's ours now */ - mutex->owner = current; - } else if (mutex->owner != current) { - /* Another process locked the mutex, take it */ - mutex_lock(&mutex->mutex); - mutex->owner = current; - } - /* Otherwise the mutex was already locked by this process */ - - mutex->level++; -} - -static inline void radeon_mutex_unlock(struct radeon_mutex *mutex) -{ - if (--mutex->level > 0) - return; - - mutex->owner = NULL; - mutex_unlock(&mutex->mutex); -} - - -/* * Core structure, functions and helpers. */ typedef uint32_t (*radeon_rreg_t)(struct radeon_device*, uint32_t); @@ -1216,11 +1501,10 @@ struct radeon_device { struct radeon_mode_info mode_info; struct radeon_scratch scratch; struct radeon_mman mman; - struct radeon_fence_driver fence_drv; - struct radeon_cp cp; - /* cayman compute rings */ - struct radeon_cp cp1; - struct radeon_cp cp2; + rwlock_t fence_lock; + struct radeon_fence_driver fence_drv[RADEON_NUM_RINGS]; + struct radeon_semaphore_driver semaphore_drv; + struct radeon_ring ring[RADEON_NUM_RINGS]; struct radeon_ib_pool ib_pool; struct radeon_irq irq; struct radeon_asic *asic; @@ -1240,10 +1524,12 @@ struct radeon_device { const struct firmware *pfp_fw; /* r6/700 PFP firmware */ const struct firmware *rlc_fw; /* r6/700 RLC firmware */ const struct firmware *mc_fw; /* NI MC firmware */ + const struct firmware *ce_fw; /* SI CE firmware */ struct r600_blit r600_blit; struct r600_vram_scratch vram_scratch; int msi_enabled; /* msi enabled */ struct r600_ih ih; /* r6/700 interrupt ring */ + struct si_rlc rlc; struct work_struct hotplug_work; int num_crtc; /* number of crtcs */ struct mutex dc_hw_i2c_mutex; /* display controller hw i2c mutex */ @@ -1264,6 +1550,11 @@ struct radeon_device { struct drm_file *cmask_filp; /* i2c buses */ struct radeon_i2c_chan *i2c_bus[RADEON_MAX_I2C_BUS]; + /* debugfs */ + struct radeon_debugfs debugfs[RADEON_DEBUGFS_MAX_COMPONENTS]; + unsigned debugfs_count; + /* virtual memory */ + struct radeon_vm_manager vm_manager; }; int radeon_device_init(struct radeon_device *rdev, @@ -1382,6 +1673,9 @@ void r100_pll_errata_after_index(struct radeon_device *rdev); #define ASIC_IS_DCE41(rdev) ((rdev->family >= CHIP_PALM) && \ (rdev->flags & RADEON_IS_IGP)) #define ASIC_IS_DCE5(rdev) ((rdev->family >= CHIP_BARTS)) +#define ASIC_IS_DCE6(rdev) ((rdev->family >= CHIP_ARUBA)) +#define ASIC_IS_DCE61(rdev) ((rdev->family >= CHIP_ARUBA) && \ + (rdev->flags & RADEON_IS_IGP)) /* * BIOS helpers. @@ -1399,18 +1693,17 @@ void radeon_atombios_fini(struct radeon_device *rdev); /* * RING helpers. */ - #if DRM_DEBUG_CODE == 0 -static inline void radeon_ring_write(struct radeon_device *rdev, uint32_t v) +static inline void radeon_ring_write(struct radeon_ring *ring, uint32_t v) { - rdev->cp.ring[rdev->cp.wptr++] = v; - rdev->cp.wptr &= rdev->cp.ptr_mask; - rdev->cp.count_dw--; - rdev->cp.ring_free_dw--; + ring->ring[ring->wptr++] = v; + ring->wptr &= ring->ptr_mask; + ring->count_dw--; + ring->ring_free_dw--; } #else /* With debugging this is just too big to inline */ -void radeon_ring_write(struct radeon_device *rdev, uint32_t v); +void radeon_ring_write(struct radeon_ring *ring, uint32_t v); #endif /* @@ -1420,46 +1713,53 @@ void radeon_ring_write(struct radeon_device *rdev, uint32_t v); #define radeon_fini(rdev) (rdev)->asic->fini((rdev)) #define radeon_resume(rdev) (rdev)->asic->resume((rdev)) #define radeon_suspend(rdev) (rdev)->asic->suspend((rdev)) -#define radeon_cs_parse(p) rdev->asic->cs_parse((p)) +#define radeon_cs_parse(rdev, r, p) (rdev)->asic->ring[(r)].cs_parse((p)) #define radeon_vga_set_state(rdev, state) (rdev)->asic->vga_set_state((rdev), (state)) -#define radeon_gpu_is_lockup(rdev) (rdev)->asic->gpu_is_lockup((rdev)) +#define radeon_gpu_is_lockup(rdev, cp) (rdev)->asic->gpu_is_lockup((rdev), (cp)) #define radeon_asic_reset(rdev) (rdev)->asic->asic_reset((rdev)) -#define radeon_gart_tlb_flush(rdev) (rdev)->asic->gart_tlb_flush((rdev)) -#define radeon_gart_set_page(rdev, i, p) (rdev)->asic->gart_set_page((rdev), (i), (p)) -#define radeon_cp_commit(rdev) (rdev)->asic->cp_commit((rdev)) -#define radeon_ring_start(rdev) (rdev)->asic->ring_start((rdev)) -#define radeon_ring_test(rdev) (rdev)->asic->ring_test((rdev)) -#define radeon_ring_ib_execute(rdev, ib) (rdev)->asic->ring_ib_execute((rdev), (ib)) -#define radeon_irq_set(rdev) (rdev)->asic->irq_set((rdev)) -#define radeon_irq_process(rdev) (rdev)->asic->irq_process((rdev)) -#define radeon_get_vblank_counter(rdev, crtc) (rdev)->asic->get_vblank_counter((rdev), (crtc)) -#define radeon_fence_ring_emit(rdev, fence) (rdev)->asic->fence_ring_emit((rdev), (fence)) -#define radeon_copy_blit(rdev, s, d, np, f) (rdev)->asic->copy_blit((rdev), (s), (d), (np), (f)) -#define radeon_copy_dma(rdev, s, d, np, f) (rdev)->asic->copy_dma((rdev), (s), (d), (np), (f)) -#define radeon_copy(rdev, s, d, np, f) (rdev)->asic->copy((rdev), (s), (d), (np), (f)) -#define radeon_get_engine_clock(rdev) (rdev)->asic->get_engine_clock((rdev)) -#define radeon_set_engine_clock(rdev, e) (rdev)->asic->set_engine_clock((rdev), (e)) -#define radeon_get_memory_clock(rdev) (rdev)->asic->get_memory_clock((rdev)) -#define radeon_set_memory_clock(rdev, e) (rdev)->asic->set_memory_clock((rdev), (e)) -#define radeon_get_pcie_lanes(rdev) (rdev)->asic->get_pcie_lanes((rdev)) -#define radeon_set_pcie_lanes(rdev, l) (rdev)->asic->set_pcie_lanes((rdev), (l)) -#define radeon_set_clock_gating(rdev, e) (rdev)->asic->set_clock_gating((rdev), (e)) -#define radeon_set_surface_reg(rdev, r, f, p, o, s) ((rdev)->asic->set_surface_reg((rdev), (r), (f), (p), (o), (s))) -#define radeon_clear_surface_reg(rdev, r) ((rdev)->asic->clear_surface_reg((rdev), (r))) -#define radeon_bandwidth_update(rdev) (rdev)->asic->bandwidth_update((rdev)) -#define radeon_hpd_init(rdev) (rdev)->asic->hpd_init((rdev)) -#define radeon_hpd_fini(rdev) (rdev)->asic->hpd_fini((rdev)) -#define radeon_hpd_sense(rdev, hpd) (rdev)->asic->hpd_sense((rdev), (hpd)) -#define radeon_hpd_set_polarity(rdev, hpd) (rdev)->asic->hpd_set_polarity((rdev), (hpd)) +#define radeon_gart_tlb_flush(rdev) (rdev)->asic->gart.tlb_flush((rdev)) +#define radeon_gart_set_page(rdev, i, p) (rdev)->asic->gart.set_page((rdev), (i), (p)) +#define radeon_ring_start(rdev, r, cp) (rdev)->asic->ring[(r)].ring_start((rdev), (cp)) +#define radeon_ring_test(rdev, r, cp) (rdev)->asic->ring[(r)].ring_test((rdev), (cp)) +#define radeon_ib_test(rdev, r, cp) (rdev)->asic->ring[(r)].ib_test((rdev), (cp)) +#define radeon_ring_ib_execute(rdev, r, ib) (rdev)->asic->ring[(r)].ib_execute((rdev), (ib)) +#define radeon_ring_ib_parse(rdev, r, ib) (rdev)->asic->ring[(r)].ib_parse((rdev), (ib)) +#define radeon_irq_set(rdev) (rdev)->asic->irq.set((rdev)) +#define radeon_irq_process(rdev) (rdev)->asic->irq.process((rdev)) +#define radeon_get_vblank_counter(rdev, crtc) (rdev)->asic->display.get_vblank_counter((rdev), (crtc)) +#define radeon_fence_ring_emit(rdev, r, fence) (rdev)->asic->ring[(r)].emit_fence((rdev), (fence)) +#define radeon_semaphore_ring_emit(rdev, r, cp, semaphore, emit_wait) (rdev)->asic->ring[(r)].emit_semaphore((rdev), (cp), (semaphore), (emit_wait)) +#define radeon_copy_blit(rdev, s, d, np, f) (rdev)->asic->copy.blit((rdev), (s), (d), (np), (f)) +#define radeon_copy_dma(rdev, s, d, np, f) (rdev)->asic->copy.dma((rdev), (s), (d), (np), (f)) +#define radeon_copy(rdev, s, d, np, f) (rdev)->asic->copy.copy((rdev), (s), (d), (np), (f)) +#define radeon_copy_blit_ring_index(rdev) (rdev)->asic->copy.blit_ring_index +#define radeon_copy_dma_ring_index(rdev) (rdev)->asic->copy.dma_ring_index +#define radeon_copy_ring_index(rdev) (rdev)->asic->copy.copy_ring_index +#define radeon_get_engine_clock(rdev) (rdev)->asic->pm.get_engine_clock((rdev)) +#define radeon_set_engine_clock(rdev, e) (rdev)->asic->pm.set_engine_clock((rdev), (e)) +#define radeon_get_memory_clock(rdev) (rdev)->asic->pm.get_memory_clock((rdev)) +#define radeon_set_memory_clock(rdev, e) (rdev)->asic->pm.set_memory_clock((rdev), (e)) +#define radeon_get_pcie_lanes(rdev) (rdev)->asic->pm.get_pcie_lanes((rdev)) +#define radeon_set_pcie_lanes(rdev, l) (rdev)->asic->pm.set_pcie_lanes((rdev), (l)) +#define radeon_set_clock_gating(rdev, e) (rdev)->asic->pm.set_clock_gating((rdev), (e)) +#define radeon_set_surface_reg(rdev, r, f, p, o, s) ((rdev)->asic->surface.set_reg((rdev), (r), (f), (p), (o), (s))) +#define radeon_clear_surface_reg(rdev, r) ((rdev)->asic->surface.clear_reg((rdev), (r))) +#define radeon_bandwidth_update(rdev) (rdev)->asic->display.bandwidth_update((rdev)) +#define radeon_hpd_init(rdev) (rdev)->asic->hpd.init((rdev)) +#define radeon_hpd_fini(rdev) (rdev)->asic->hpd.fini((rdev)) +#define radeon_hpd_sense(rdev, h) (rdev)->asic->hpd.sense((rdev), (h)) +#define radeon_hpd_set_polarity(rdev, h) (rdev)->asic->hpd.set_polarity((rdev), (h)) #define radeon_gui_idle(rdev) (rdev)->asic->gui_idle((rdev)) -#define radeon_pm_misc(rdev) (rdev)->asic->pm_misc((rdev)) -#define radeon_pm_prepare(rdev) (rdev)->asic->pm_prepare((rdev)) -#define radeon_pm_finish(rdev) (rdev)->asic->pm_finish((rdev)) -#define radeon_pm_init_profile(rdev) (rdev)->asic->pm_init_profile((rdev)) -#define radeon_pm_get_dynpm_state(rdev) (rdev)->asic->pm_get_dynpm_state((rdev)) -#define radeon_pre_page_flip(rdev, crtc) rdev->asic->pre_page_flip((rdev), (crtc)) -#define radeon_page_flip(rdev, crtc, base) rdev->asic->page_flip((rdev), (crtc), (base)) -#define radeon_post_page_flip(rdev, crtc) rdev->asic->post_page_flip((rdev), (crtc)) +#define radeon_pm_misc(rdev) (rdev)->asic->pm.misc((rdev)) +#define radeon_pm_prepare(rdev) (rdev)->asic->pm.prepare((rdev)) +#define radeon_pm_finish(rdev) (rdev)->asic->pm.finish((rdev)) +#define radeon_pm_init_profile(rdev) (rdev)->asic->pm.init_profile((rdev)) +#define radeon_pm_get_dynpm_state(rdev) (rdev)->asic->pm.get_dynpm_state((rdev)) +#define radeon_pre_page_flip(rdev, crtc) rdev->asic->pflip.pre_page_flip((rdev), (crtc)) +#define radeon_page_flip(rdev, crtc, base) rdev->asic->pflip.page_flip((rdev), (crtc), (base)) +#define radeon_post_page_flip(rdev, crtc) rdev->asic->pflip.post_page_flip((rdev), (crtc)) +#define radeon_wait_for_vblank(rdev, crtc) rdev->asic->display.wait_for_vblank((rdev), (crtc)) +#define radeon_mc_wait_for_idle(rdev) rdev->asic->mc_wait_for_idle((rdev)) /* Common functions */ /* AGP */ @@ -1488,12 +1788,49 @@ extern int radeon_suspend_kms(struct drm_device *dev, pm_message_t state); extern void radeon_ttm_set_active_vram_size(struct radeon_device *rdev, u64 size); /* + * vm + */ +int radeon_vm_manager_init(struct radeon_device *rdev); +void radeon_vm_manager_fini(struct radeon_device *rdev); +int radeon_vm_manager_start(struct radeon_device *rdev); +int radeon_vm_manager_suspend(struct radeon_device *rdev); +int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm); +void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm); +int radeon_vm_bind(struct radeon_device *rdev, struct radeon_vm *vm); +void radeon_vm_unbind(struct radeon_device *rdev, struct radeon_vm *vm); +int radeon_vm_bo_update_pte(struct radeon_device *rdev, + struct radeon_vm *vm, + struct radeon_bo *bo, + struct ttm_mem_reg *mem); +void radeon_vm_bo_invalidate(struct radeon_device *rdev, + struct radeon_bo *bo); +int radeon_vm_bo_add(struct radeon_device *rdev, + struct radeon_vm *vm, + struct radeon_bo *bo, + uint64_t offset, + uint32_t flags); +int radeon_vm_bo_rmv(struct radeon_device *rdev, + struct radeon_vm *vm, + struct radeon_bo *bo); + + +/* * R600 vram scratch functions */ int r600_vram_scratch_init(struct radeon_device *rdev); void r600_vram_scratch_fini(struct radeon_device *rdev); /* + * r600 cs checking helper + */ +unsigned r600_mip_minify(unsigned size, unsigned level); +bool r600_fmt_is_valid_color(u32 format); +bool r600_fmt_is_valid_texture(u32 format, enum radeon_family family); +int r600_fmt_get_blocksize(u32 format); +int r600_fmt_get_nblocksx(u32 format, u32 w); +int r600_fmt_get_nblocksy(u32 format, u32 h); + +/* * r600 functions used by radeon_encoder.c */ extern void r600_hdmi_enable(struct drm_encoder *encoder); diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c index a2e1eae..be4dc2f 100644 --- a/drivers/gpu/drm/radeon/radeon_asic.c +++ b/drivers/gpu/drm/radeon/radeon_asic.c @@ -114,13 +114,13 @@ void radeon_agp_disable(struct radeon_device *rdev) rdev->family == CHIP_R423) { DRM_INFO("Forcing AGP to PCIE mode\n"); rdev->flags |= RADEON_IS_PCIE; - rdev->asic->gart_tlb_flush = &rv370_pcie_gart_tlb_flush; - rdev->asic->gart_set_page = &rv370_pcie_gart_set_page; + rdev->asic->gart.tlb_flush = &rv370_pcie_gart_tlb_flush; + rdev->asic->gart.set_page = &rv370_pcie_gart_set_page; } else { DRM_INFO("Forcing AGP to PCI mode\n"); rdev->flags |= RADEON_IS_PCI; - rdev->asic->gart_tlb_flush = &r100_pci_gart_tlb_flush; - rdev->asic->gart_set_page = &r100_pci_gart_set_page; + rdev->asic->gart.tlb_flush = &r100_pci_gart_tlb_flush; + rdev->asic->gart.set_page = &r100_pci_gart_set_page; } rdev->mc.gtt_size = radeon_gart_size * 1024 * 1024; } @@ -136,44 +136,70 @@ static struct radeon_asic r100_asic = { .vga_set_state = &r100_vga_set_state, .gpu_is_lockup = &r100_gpu_is_lockup, .asic_reset = &r100_asic_reset, - .gart_tlb_flush = &r100_pci_gart_tlb_flush, - .gart_set_page = &r100_pci_gart_set_page, - .cp_commit = &r100_cp_commit, - .ring_start = &r100_ring_start, - .ring_test = &r100_ring_test, - .ring_ib_execute = &r100_ring_ib_execute, - .irq_set = &r100_irq_set, - .irq_process = &r100_irq_process, - .get_vblank_counter = &r100_get_vblank_counter, - .fence_ring_emit = &r100_fence_ring_emit, - .cs_parse = &r100_cs_parse, - .copy_blit = &r100_copy_blit, - .copy_dma = NULL, - .copy = &r100_copy_blit, - .get_engine_clock = &radeon_legacy_get_engine_clock, - .set_engine_clock = &radeon_legacy_set_engine_clock, - .get_memory_clock = &radeon_legacy_get_memory_clock, - .set_memory_clock = NULL, - .get_pcie_lanes = NULL, - .set_pcie_lanes = NULL, - .set_clock_gating = &radeon_legacy_set_clock_gating, - .set_surface_reg = r100_set_surface_reg, - .clear_surface_reg = r100_clear_surface_reg, - .bandwidth_update = &r100_bandwidth_update, - .hpd_init = &r100_hpd_init, - .hpd_fini = &r100_hpd_fini, - .hpd_sense = &r100_hpd_sense, - .hpd_set_polarity = &r100_hpd_set_polarity, .ioctl_wait_idle = NULL, .gui_idle = &r100_gui_idle, - .pm_misc = &r100_pm_misc, - .pm_prepare = &r100_pm_prepare, - .pm_finish = &r100_pm_finish, - .pm_init_profile = &r100_pm_init_profile, - .pm_get_dynpm_state = &r100_pm_get_dynpm_state, - .pre_page_flip = &r100_pre_page_flip, - .page_flip = &r100_page_flip, - .post_page_flip = &r100_post_page_flip, + .mc_wait_for_idle = &r100_mc_wait_for_idle, + .gart = { + .tlb_flush = &r100_pci_gart_tlb_flush, + .set_page = &r100_pci_gart_set_page, + }, + .ring = { + [RADEON_RING_TYPE_GFX_INDEX] = { + .ib_execute = &r100_ring_ib_execute, + .emit_fence = &r100_fence_ring_emit, + .emit_semaphore = &r100_semaphore_ring_emit, + .cs_parse = &r100_cs_parse, + .ring_start = &r100_ring_start, + .ring_test = &r100_ring_test, + .ib_test = &r100_ib_test, + } + }, + .irq = { + .set = &r100_irq_set, + .process = &r100_irq_process, + }, + .display = { + .bandwidth_update = &r100_bandwidth_update, + .get_vblank_counter = &r100_get_vblank_counter, + .wait_for_vblank = &r100_wait_for_vblank, + }, + .copy = { + .blit = &r100_copy_blit, + .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .dma = NULL, + .dma_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .copy = &r100_copy_blit, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, + }, + .surface = { + .set_reg = r100_set_surface_reg, + .clear_reg = r100_clear_surface_reg, + }, + .hpd = { + .init = &r100_hpd_init, + .fini = &r100_hpd_fini, + .sense = &r100_hpd_sense, + .set_polarity = &r100_hpd_set_polarity, + }, + .pm = { + .misc = &r100_pm_misc, + .prepare = &r100_pm_prepare, + .finish = &r100_pm_finish, + .init_profile = &r100_pm_init_profile, + .get_dynpm_state = &r100_pm_get_dynpm_state, + .get_engine_clock = &radeon_legacy_get_engine_clock, + .set_engine_clock = &radeon_legacy_set_engine_clock, + .get_memory_clock = &radeon_legacy_get_memory_clock, + .set_memory_clock = NULL, + .get_pcie_lanes = NULL, + .set_pcie_lanes = NULL, + .set_clock_gating = &radeon_legacy_set_clock_gating, + }, + .pflip = { + .pre_page_flip = &r100_pre_page_flip, + .page_flip = &r100_page_flip, + .post_page_flip = &r100_post_page_flip, + }, }; static struct radeon_asic r200_asic = { @@ -184,43 +210,70 @@ static struct radeon_asic r200_asic = { .vga_set_state = &r100_vga_set_state, .gpu_is_lockup = &r100_gpu_is_lockup, .asic_reset = &r100_asic_reset, - .gart_tlb_flush = &r100_pci_gart_tlb_flush, - .gart_set_page = &r100_pci_gart_set_page, - .cp_commit = &r100_cp_commit, - .ring_start = &r100_ring_start, - .ring_test = &r100_ring_test, - .ring_ib_execute = &r100_ring_ib_execute, - .irq_set = &r100_irq_set, - .irq_process = &r100_irq_process, - .get_vblank_counter = &r100_get_vblank_counter, - .fence_ring_emit = &r100_fence_ring_emit, - .cs_parse = &r100_cs_parse, - .copy_blit = &r100_copy_blit, - .copy_dma = &r200_copy_dma, - .copy = &r100_copy_blit, - .get_engine_clock = &radeon_legacy_get_engine_clock, - .set_engine_clock = &radeon_legacy_set_engine_clock, - .get_memory_clock = &radeon_legacy_get_memory_clock, - .set_memory_clock = NULL, - .set_pcie_lanes = NULL, - .set_clock_gating = &radeon_legacy_set_clock_gating, - .set_surface_reg = r100_set_surface_reg, - .clear_surface_reg = r100_clear_surface_reg, - .bandwidth_update = &r100_bandwidth_update, - .hpd_init = &r100_hpd_init, - .hpd_fini = &r100_hpd_fini, - .hpd_sense = &r100_hpd_sense, - .hpd_set_polarity = &r100_hpd_set_polarity, .ioctl_wait_idle = NULL, .gui_idle = &r100_gui_idle, - .pm_misc = &r100_pm_misc, - .pm_prepare = &r100_pm_prepare, - .pm_finish = &r100_pm_finish, - .pm_init_profile = &r100_pm_init_profile, - .pm_get_dynpm_state = &r100_pm_get_dynpm_state, - .pre_page_flip = &r100_pre_page_flip, - .page_flip = &r100_page_flip, - .post_page_flip = &r100_post_page_flip, + .mc_wait_for_idle = &r100_mc_wait_for_idle, + .gart = { + .tlb_flush = &r100_pci_gart_tlb_flush, + .set_page = &r100_pci_gart_set_page, + }, + .ring = { + [RADEON_RING_TYPE_GFX_INDEX] = { + .ib_execute = &r100_ring_ib_execute, + .emit_fence = &r100_fence_ring_emit, + .emit_semaphore = &r100_semaphore_ring_emit, + .cs_parse = &r100_cs_parse, + .ring_start = &r100_ring_start, + .ring_test = &r100_ring_test, + .ib_test = &r100_ib_test, + } + }, + .irq = { + .set = &r100_irq_set, + .process = &r100_irq_process, + }, + .display = { + .bandwidth_update = &r100_bandwidth_update, + .get_vblank_counter = &r100_get_vblank_counter, + .wait_for_vblank = &r100_wait_for_vblank, + }, + .copy = { + .blit = &r100_copy_blit, + .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .dma = &r200_copy_dma, + .dma_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .copy = &r100_copy_blit, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, + }, + .surface = { + .set_reg = r100_set_surface_reg, + .clear_reg = r100_clear_surface_reg, + }, + .hpd = { + .init = &r100_hpd_init, + .fini = &r100_hpd_fini, + .sense = &r100_hpd_sense, + .set_polarity = &r100_hpd_set_polarity, + }, + .pm = { + .misc = &r100_pm_misc, + .prepare = &r100_pm_prepare, + .finish = &r100_pm_finish, + .init_profile = &r100_pm_init_profile, + .get_dynpm_state = &r100_pm_get_dynpm_state, + .get_engine_clock = &radeon_legacy_get_engine_clock, + .set_engine_clock = &radeon_legacy_set_engine_clock, + .get_memory_clock = &radeon_legacy_get_memory_clock, + .set_memory_clock = NULL, + .get_pcie_lanes = NULL, + .set_pcie_lanes = NULL, + .set_clock_gating = &radeon_legacy_set_clock_gating, + }, + .pflip = { + .pre_page_flip = &r100_pre_page_flip, + .page_flip = &r100_page_flip, + .post_page_flip = &r100_post_page_flip, + }, }; static struct radeon_asic r300_asic = { @@ -231,44 +284,70 @@ static struct radeon_asic r300_asic = { .vga_set_state = &r100_vga_set_state, .gpu_is_lockup = &r300_gpu_is_lockup, .asic_reset = &r300_asic_reset, - .gart_tlb_flush = &r100_pci_gart_tlb_flush, - .gart_set_page = &r100_pci_gart_set_page, - .cp_commit = &r100_cp_commit, - .ring_start = &r300_ring_start, - .ring_test = &r100_ring_test, - .ring_ib_execute = &r100_ring_ib_execute, - .irq_set = &r100_irq_set, - .irq_process = &r100_irq_process, - .get_vblank_counter = &r100_get_vblank_counter, - .fence_ring_emit = &r300_fence_ring_emit, - .cs_parse = &r300_cs_parse, - .copy_blit = &r100_copy_blit, - .copy_dma = &r200_copy_dma, - .copy = &r100_copy_blit, - .get_engine_clock = &radeon_legacy_get_engine_clock, - .set_engine_clock = &radeon_legacy_set_engine_clock, - .get_memory_clock = &radeon_legacy_get_memory_clock, - .set_memory_clock = NULL, - .get_pcie_lanes = &rv370_get_pcie_lanes, - .set_pcie_lanes = &rv370_set_pcie_lanes, - .set_clock_gating = &radeon_legacy_set_clock_gating, - .set_surface_reg = r100_set_surface_reg, - .clear_surface_reg = r100_clear_surface_reg, - .bandwidth_update = &r100_bandwidth_update, - .hpd_init = &r100_hpd_init, - .hpd_fini = &r100_hpd_fini, - .hpd_sense = &r100_hpd_sense, - .hpd_set_polarity = &r100_hpd_set_polarity, .ioctl_wait_idle = NULL, .gui_idle = &r100_gui_idle, - .pm_misc = &r100_pm_misc, - .pm_prepare = &r100_pm_prepare, - .pm_finish = &r100_pm_finish, - .pm_init_profile = &r100_pm_init_profile, - .pm_get_dynpm_state = &r100_pm_get_dynpm_state, - .pre_page_flip = &r100_pre_page_flip, - .page_flip = &r100_page_flip, - .post_page_flip = &r100_post_page_flip, + .mc_wait_for_idle = &r300_mc_wait_for_idle, + .gart = { + .tlb_flush = &r100_pci_gart_tlb_flush, + .set_page = &r100_pci_gart_set_page, + }, + .ring = { + [RADEON_RING_TYPE_GFX_INDEX] = { + .ib_execute = &r100_ring_ib_execute, + .emit_fence = &r300_fence_ring_emit, + .emit_semaphore = &r100_semaphore_ring_emit, + .cs_parse = &r300_cs_parse, + .ring_start = &r300_ring_start, + .ring_test = &r100_ring_test, + .ib_test = &r100_ib_test, + } + }, + .irq = { + .set = &r100_irq_set, + .process = &r100_irq_process, + }, + .display = { + .bandwidth_update = &r100_bandwidth_update, + .get_vblank_counter = &r100_get_vblank_counter, + .wait_for_vblank = &r100_wait_for_vblank, + }, + .copy = { + .blit = &r100_copy_blit, + .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .dma = &r200_copy_dma, + .dma_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .copy = &r100_copy_blit, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, + }, + .surface = { + .set_reg = r100_set_surface_reg, + .clear_reg = r100_clear_surface_reg, + }, + .hpd = { + .init = &r100_hpd_init, + .fini = &r100_hpd_fini, + .sense = &r100_hpd_sense, + .set_polarity = &r100_hpd_set_polarity, + }, + .pm = { + .misc = &r100_pm_misc, + .prepare = &r100_pm_prepare, + .finish = &r100_pm_finish, + .init_profile = &r100_pm_init_profile, + .get_dynpm_state = &r100_pm_get_dynpm_state, + .get_engine_clock = &radeon_legacy_get_engine_clock, + .set_engine_clock = &radeon_legacy_set_engine_clock, + .get_memory_clock = &radeon_legacy_get_memory_clock, + .set_memory_clock = NULL, + .get_pcie_lanes = &rv370_get_pcie_lanes, + .set_pcie_lanes = &rv370_set_pcie_lanes, + .set_clock_gating = &radeon_legacy_set_clock_gating, + }, + .pflip = { + .pre_page_flip = &r100_pre_page_flip, + .page_flip = &r100_page_flip, + .post_page_flip = &r100_post_page_flip, + }, }; static struct radeon_asic r300_asic_pcie = { @@ -279,43 +358,70 @@ static struct radeon_asic r300_asic_pcie = { .vga_set_state = &r100_vga_set_state, .gpu_is_lockup = &r300_gpu_is_lockup, .asic_reset = &r300_asic_reset, - .gart_tlb_flush = &rv370_pcie_gart_tlb_flush, - .gart_set_page = &rv370_pcie_gart_set_page, - .cp_commit = &r100_cp_commit, - .ring_start = &r300_ring_start, - .ring_test = &r100_ring_test, - .ring_ib_execute = &r100_ring_ib_execute, - .irq_set = &r100_irq_set, - .irq_process = &r100_irq_process, - .get_vblank_counter = &r100_get_vblank_counter, - .fence_ring_emit = &r300_fence_ring_emit, - .cs_parse = &r300_cs_parse, - .copy_blit = &r100_copy_blit, - .copy_dma = &r200_copy_dma, - .copy = &r100_copy_blit, - .get_engine_clock = &radeon_legacy_get_engine_clock, - .set_engine_clock = &radeon_legacy_set_engine_clock, - .get_memory_clock = &radeon_legacy_get_memory_clock, - .set_memory_clock = NULL, - .set_pcie_lanes = &rv370_set_pcie_lanes, - .set_clock_gating = &radeon_legacy_set_clock_gating, - .set_surface_reg = r100_set_surface_reg, - .clear_surface_reg = r100_clear_surface_reg, - .bandwidth_update = &r100_bandwidth_update, - .hpd_init = &r100_hpd_init, - .hpd_fini = &r100_hpd_fini, - .hpd_sense = &r100_hpd_sense, - .hpd_set_polarity = &r100_hpd_set_polarity, .ioctl_wait_idle = NULL, .gui_idle = &r100_gui_idle, - .pm_misc = &r100_pm_misc, - .pm_prepare = &r100_pm_prepare, - .pm_finish = &r100_pm_finish, - .pm_init_profile = &r100_pm_init_profile, - .pm_get_dynpm_state = &r100_pm_get_dynpm_state, - .pre_page_flip = &r100_pre_page_flip, - .page_flip = &r100_page_flip, - .post_page_flip = &r100_post_page_flip, + .mc_wait_for_idle = &r300_mc_wait_for_idle, + .gart = { + .tlb_flush = &rv370_pcie_gart_tlb_flush, + .set_page = &rv370_pcie_gart_set_page, + }, + .ring = { + [RADEON_RING_TYPE_GFX_INDEX] = { + .ib_execute = &r100_ring_ib_execute, + .emit_fence = &r300_fence_ring_emit, + .emit_semaphore = &r100_semaphore_ring_emit, + .cs_parse = &r300_cs_parse, + .ring_start = &r300_ring_start, + .ring_test = &r100_ring_test, + .ib_test = &r100_ib_test, + } + }, + .irq = { + .set = &r100_irq_set, + .process = &r100_irq_process, + }, + .display = { + .bandwidth_update = &r100_bandwidth_update, + .get_vblank_counter = &r100_get_vblank_counter, + .wait_for_vblank = &r100_wait_for_vblank, + }, + .copy = { + .blit = &r100_copy_blit, + .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .dma = &r200_copy_dma, + .dma_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .copy = &r100_copy_blit, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, + }, + .surface = { + .set_reg = r100_set_surface_reg, + .clear_reg = r100_clear_surface_reg, + }, + .hpd = { + .init = &r100_hpd_init, + .fini = &r100_hpd_fini, + .sense = &r100_hpd_sense, + .set_polarity = &r100_hpd_set_polarity, + }, + .pm = { + .misc = &r100_pm_misc, + .prepare = &r100_pm_prepare, + .finish = &r100_pm_finish, + .init_profile = &r100_pm_init_profile, + .get_dynpm_state = &r100_pm_get_dynpm_state, + .get_engine_clock = &radeon_legacy_get_engine_clock, + .set_engine_clock = &radeon_legacy_set_engine_clock, + .get_memory_clock = &radeon_legacy_get_memory_clock, + .set_memory_clock = NULL, + .get_pcie_lanes = &rv370_get_pcie_lanes, + .set_pcie_lanes = &rv370_set_pcie_lanes, + .set_clock_gating = &radeon_legacy_set_clock_gating, + }, + .pflip = { + .pre_page_flip = &r100_pre_page_flip, + .page_flip = &r100_page_flip, + .post_page_flip = &r100_post_page_flip, + }, }; static struct radeon_asic r420_asic = { @@ -326,44 +432,70 @@ static struct radeon_asic r420_asic = { .vga_set_state = &r100_vga_set_state, .gpu_is_lockup = &r300_gpu_is_lockup, .asic_reset = &r300_asic_reset, - .gart_tlb_flush = &rv370_pcie_gart_tlb_flush, - .gart_set_page = &rv370_pcie_gart_set_page, - .cp_commit = &r100_cp_commit, - .ring_start = &r300_ring_start, - .ring_test = &r100_ring_test, - .ring_ib_execute = &r100_ring_ib_execute, - .irq_set = &r100_irq_set, - .irq_process = &r100_irq_process, - .get_vblank_counter = &r100_get_vblank_counter, - .fence_ring_emit = &r300_fence_ring_emit, - .cs_parse = &r300_cs_parse, - .copy_blit = &r100_copy_blit, - .copy_dma = &r200_copy_dma, - .copy = &r100_copy_blit, - .get_engine_clock = &radeon_atom_get_engine_clock, - .set_engine_clock = &radeon_atom_set_engine_clock, - .get_memory_clock = &radeon_atom_get_memory_clock, - .set_memory_clock = &radeon_atom_set_memory_clock, - .get_pcie_lanes = &rv370_get_pcie_lanes, - .set_pcie_lanes = &rv370_set_pcie_lanes, - .set_clock_gating = &radeon_atom_set_clock_gating, - .set_surface_reg = r100_set_surface_reg, - .clear_surface_reg = r100_clear_surface_reg, - .bandwidth_update = &r100_bandwidth_update, - .hpd_init = &r100_hpd_init, - .hpd_fini = &r100_hpd_fini, - .hpd_sense = &r100_hpd_sense, - .hpd_set_polarity = &r100_hpd_set_polarity, .ioctl_wait_idle = NULL, .gui_idle = &r100_gui_idle, - .pm_misc = &r100_pm_misc, - .pm_prepare = &r100_pm_prepare, - .pm_finish = &r100_pm_finish, - .pm_init_profile = &r420_pm_init_profile, - .pm_get_dynpm_state = &r100_pm_get_dynpm_state, - .pre_page_flip = &r100_pre_page_flip, - .page_flip = &r100_page_flip, - .post_page_flip = &r100_post_page_flip, + .mc_wait_for_idle = &r300_mc_wait_for_idle, + .gart = { + .tlb_flush = &rv370_pcie_gart_tlb_flush, + .set_page = &rv370_pcie_gart_set_page, + }, + .ring = { + [RADEON_RING_TYPE_GFX_INDEX] = { + .ib_execute = &r100_ring_ib_execute, + .emit_fence = &r300_fence_ring_emit, + .emit_semaphore = &r100_semaphore_ring_emit, + .cs_parse = &r300_cs_parse, + .ring_start = &r300_ring_start, + .ring_test = &r100_ring_test, + .ib_test = &r100_ib_test, + } + }, + .irq = { + .set = &r100_irq_set, + .process = &r100_irq_process, + }, + .display = { + .bandwidth_update = &r100_bandwidth_update, + .get_vblank_counter = &r100_get_vblank_counter, + .wait_for_vblank = &r100_wait_for_vblank, + }, + .copy = { + .blit = &r100_copy_blit, + .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .dma = &r200_copy_dma, + .dma_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .copy = &r100_copy_blit, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, + }, + .surface = { + .set_reg = r100_set_surface_reg, + .clear_reg = r100_clear_surface_reg, + }, + .hpd = { + .init = &r100_hpd_init, + .fini = &r100_hpd_fini, + .sense = &r100_hpd_sense, + .set_polarity = &r100_hpd_set_polarity, + }, + .pm = { + .misc = &r100_pm_misc, + .prepare = &r100_pm_prepare, + .finish = &r100_pm_finish, + .init_profile = &r420_pm_init_profile, + .get_dynpm_state = &r100_pm_get_dynpm_state, + .get_engine_clock = &radeon_atom_get_engine_clock, + .set_engine_clock = &radeon_atom_set_engine_clock, + .get_memory_clock = &radeon_atom_get_memory_clock, + .set_memory_clock = &radeon_atom_set_memory_clock, + .get_pcie_lanes = &rv370_get_pcie_lanes, + .set_pcie_lanes = &rv370_set_pcie_lanes, + .set_clock_gating = &radeon_atom_set_clock_gating, + }, + .pflip = { + .pre_page_flip = &r100_pre_page_flip, + .page_flip = &r100_page_flip, + .post_page_flip = &r100_post_page_flip, + }, }; static struct radeon_asic rs400_asic = { @@ -374,44 +506,70 @@ static struct radeon_asic rs400_asic = { .vga_set_state = &r100_vga_set_state, .gpu_is_lockup = &r300_gpu_is_lockup, .asic_reset = &r300_asic_reset, - .gart_tlb_flush = &rs400_gart_tlb_flush, - .gart_set_page = &rs400_gart_set_page, - .cp_commit = &r100_cp_commit, - .ring_start = &r300_ring_start, - .ring_test = &r100_ring_test, - .ring_ib_execute = &r100_ring_ib_execute, - .irq_set = &r100_irq_set, - .irq_process = &r100_irq_process, - .get_vblank_counter = &r100_get_vblank_counter, - .fence_ring_emit = &r300_fence_ring_emit, - .cs_parse = &r300_cs_parse, - .copy_blit = &r100_copy_blit, - .copy_dma = &r200_copy_dma, - .copy = &r100_copy_blit, - .get_engine_clock = &radeon_legacy_get_engine_clock, - .set_engine_clock = &radeon_legacy_set_engine_clock, - .get_memory_clock = &radeon_legacy_get_memory_clock, - .set_memory_clock = NULL, - .get_pcie_lanes = NULL, - .set_pcie_lanes = NULL, - .set_clock_gating = &radeon_legacy_set_clock_gating, - .set_surface_reg = r100_set_surface_reg, - .clear_surface_reg = r100_clear_surface_reg, - .bandwidth_update = &r100_bandwidth_update, - .hpd_init = &r100_hpd_init, - .hpd_fini = &r100_hpd_fini, - .hpd_sense = &r100_hpd_sense, - .hpd_set_polarity = &r100_hpd_set_polarity, .ioctl_wait_idle = NULL, .gui_idle = &r100_gui_idle, - .pm_misc = &r100_pm_misc, - .pm_prepare = &r100_pm_prepare, - .pm_finish = &r100_pm_finish, - .pm_init_profile = &r100_pm_init_profile, - .pm_get_dynpm_state = &r100_pm_get_dynpm_state, - .pre_page_flip = &r100_pre_page_flip, - .page_flip = &r100_page_flip, - .post_page_flip = &r100_post_page_flip, + .mc_wait_for_idle = &rs400_mc_wait_for_idle, + .gart = { + .tlb_flush = &rs400_gart_tlb_flush, + .set_page = &rs400_gart_set_page, + }, + .ring = { + [RADEON_RING_TYPE_GFX_INDEX] = { + .ib_execute = &r100_ring_ib_execute, + .emit_fence = &r300_fence_ring_emit, + .emit_semaphore = &r100_semaphore_ring_emit, + .cs_parse = &r300_cs_parse, + .ring_start = &r300_ring_start, + .ring_test = &r100_ring_test, + .ib_test = &r100_ib_test, + } + }, + .irq = { + .set = &r100_irq_set, + .process = &r100_irq_process, + }, + .display = { + .bandwidth_update = &r100_bandwidth_update, + .get_vblank_counter = &r100_get_vblank_counter, + .wait_for_vblank = &r100_wait_for_vblank, + }, + .copy = { + .blit = &r100_copy_blit, + .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .dma = &r200_copy_dma, + .dma_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .copy = &r100_copy_blit, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, + }, + .surface = { + .set_reg = r100_set_surface_reg, + .clear_reg = r100_clear_surface_reg, + }, + .hpd = { + .init = &r100_hpd_init, + .fini = &r100_hpd_fini, + .sense = &r100_hpd_sense, + .set_polarity = &r100_hpd_set_polarity, + }, + .pm = { + .misc = &r100_pm_misc, + .prepare = &r100_pm_prepare, + .finish = &r100_pm_finish, + .init_profile = &r100_pm_init_profile, + .get_dynpm_state = &r100_pm_get_dynpm_state, + .get_engine_clock = &radeon_legacy_get_engine_clock, + .set_engine_clock = &radeon_legacy_set_engine_clock, + .get_memory_clock = &radeon_legacy_get_memory_clock, + .set_memory_clock = NULL, + .get_pcie_lanes = NULL, + .set_pcie_lanes = NULL, + .set_clock_gating = &radeon_legacy_set_clock_gating, + }, + .pflip = { + .pre_page_flip = &r100_pre_page_flip, + .page_flip = &r100_page_flip, + .post_page_flip = &r100_post_page_flip, + }, }; static struct radeon_asic rs600_asic = { @@ -422,44 +580,70 @@ static struct radeon_asic rs600_asic = { .vga_set_state = &r100_vga_set_state, .gpu_is_lockup = &r300_gpu_is_lockup, .asic_reset = &rs600_asic_reset, - .gart_tlb_flush = &rs600_gart_tlb_flush, - .gart_set_page = &rs600_gart_set_page, - .cp_commit = &r100_cp_commit, - .ring_start = &r300_ring_start, - .ring_test = &r100_ring_test, - .ring_ib_execute = &r100_ring_ib_execute, - .irq_set = &rs600_irq_set, - .irq_process = &rs600_irq_process, - .get_vblank_counter = &rs600_get_vblank_counter, - .fence_ring_emit = &r300_fence_ring_emit, - .cs_parse = &r300_cs_parse, - .copy_blit = &r100_copy_blit, - .copy_dma = &r200_copy_dma, - .copy = &r100_copy_blit, - .get_engine_clock = &radeon_atom_get_engine_clock, - .set_engine_clock = &radeon_atom_set_engine_clock, - .get_memory_clock = &radeon_atom_get_memory_clock, - .set_memory_clock = &radeon_atom_set_memory_clock, - .get_pcie_lanes = NULL, - .set_pcie_lanes = NULL, - .set_clock_gating = &radeon_atom_set_clock_gating, - .set_surface_reg = r100_set_surface_reg, - .clear_surface_reg = r100_clear_surface_reg, - .bandwidth_update = &rs600_bandwidth_update, - .hpd_init = &rs600_hpd_init, - .hpd_fini = &rs600_hpd_fini, - .hpd_sense = &rs600_hpd_sense, - .hpd_set_polarity = &rs600_hpd_set_polarity, .ioctl_wait_idle = NULL, .gui_idle = &r100_gui_idle, - .pm_misc = &rs600_pm_misc, - .pm_prepare = &rs600_pm_prepare, - .pm_finish = &rs600_pm_finish, - .pm_init_profile = &r420_pm_init_profile, - .pm_get_dynpm_state = &r100_pm_get_dynpm_state, - .pre_page_flip = &rs600_pre_page_flip, - .page_flip = &rs600_page_flip, - .post_page_flip = &rs600_post_page_flip, + .mc_wait_for_idle = &rs600_mc_wait_for_idle, + .gart = { + .tlb_flush = &rs600_gart_tlb_flush, + .set_page = &rs600_gart_set_page, + }, + .ring = { + [RADEON_RING_TYPE_GFX_INDEX] = { + .ib_execute = &r100_ring_ib_execute, + .emit_fence = &r300_fence_ring_emit, + .emit_semaphore = &r100_semaphore_ring_emit, + .cs_parse = &r300_cs_parse, + .ring_start = &r300_ring_start, + .ring_test = &r100_ring_test, + .ib_test = &r100_ib_test, + } + }, + .irq = { + .set = &rs600_irq_set, + .process = &rs600_irq_process, + }, + .display = { + .bandwidth_update = &rs600_bandwidth_update, + .get_vblank_counter = &rs600_get_vblank_counter, + .wait_for_vblank = &avivo_wait_for_vblank, + }, + .copy = { + .blit = &r100_copy_blit, + .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .dma = &r200_copy_dma, + .dma_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .copy = &r100_copy_blit, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, + }, + .surface = { + .set_reg = r100_set_surface_reg, + .clear_reg = r100_clear_surface_reg, + }, + .hpd = { + .init = &rs600_hpd_init, + .fini = &rs600_hpd_fini, + .sense = &rs600_hpd_sense, + .set_polarity = &rs600_hpd_set_polarity, + }, + .pm = { + .misc = &rs600_pm_misc, + .prepare = &rs600_pm_prepare, + .finish = &rs600_pm_finish, + .init_profile = &r420_pm_init_profile, + .get_dynpm_state = &r100_pm_get_dynpm_state, + .get_engine_clock = &radeon_atom_get_engine_clock, + .set_engine_clock = &radeon_atom_set_engine_clock, + .get_memory_clock = &radeon_atom_get_memory_clock, + .set_memory_clock = &radeon_atom_set_memory_clock, + .get_pcie_lanes = NULL, + .set_pcie_lanes = NULL, + .set_clock_gating = &radeon_atom_set_clock_gating, + }, + .pflip = { + .pre_page_flip = &rs600_pre_page_flip, + .page_flip = &rs600_page_flip, + .post_page_flip = &rs600_post_page_flip, + }, }; static struct radeon_asic rs690_asic = { @@ -470,44 +654,70 @@ static struct radeon_asic rs690_asic = { .vga_set_state = &r100_vga_set_state, .gpu_is_lockup = &r300_gpu_is_lockup, .asic_reset = &rs600_asic_reset, - .gart_tlb_flush = &rs400_gart_tlb_flush, - .gart_set_page = &rs400_gart_set_page, - .cp_commit = &r100_cp_commit, - .ring_start = &r300_ring_start, - .ring_test = &r100_ring_test, - .ring_ib_execute = &r100_ring_ib_execute, - .irq_set = &rs600_irq_set, - .irq_process = &rs600_irq_process, - .get_vblank_counter = &rs600_get_vblank_counter, - .fence_ring_emit = &r300_fence_ring_emit, - .cs_parse = &r300_cs_parse, - .copy_blit = &r100_copy_blit, - .copy_dma = &r200_copy_dma, - .copy = &r200_copy_dma, - .get_engine_clock = &radeon_atom_get_engine_clock, - .set_engine_clock = &radeon_atom_set_engine_clock, - .get_memory_clock = &radeon_atom_get_memory_clock, - .set_memory_clock = &radeon_atom_set_memory_clock, - .get_pcie_lanes = NULL, - .set_pcie_lanes = NULL, - .set_clock_gating = &radeon_atom_set_clock_gating, - .set_surface_reg = r100_set_surface_reg, - .clear_surface_reg = r100_clear_surface_reg, - .bandwidth_update = &rs690_bandwidth_update, - .hpd_init = &rs600_hpd_init, - .hpd_fini = &rs600_hpd_fini, - .hpd_sense = &rs600_hpd_sense, - .hpd_set_polarity = &rs600_hpd_set_polarity, .ioctl_wait_idle = NULL, .gui_idle = &r100_gui_idle, - .pm_misc = &rs600_pm_misc, - .pm_prepare = &rs600_pm_prepare, - .pm_finish = &rs600_pm_finish, - .pm_init_profile = &r420_pm_init_profile, - .pm_get_dynpm_state = &r100_pm_get_dynpm_state, - .pre_page_flip = &rs600_pre_page_flip, - .page_flip = &rs600_page_flip, - .post_page_flip = &rs600_post_page_flip, + .mc_wait_for_idle = &rs690_mc_wait_for_idle, + .gart = { + .tlb_flush = &rs400_gart_tlb_flush, + .set_page = &rs400_gart_set_page, + }, + .ring = { + [RADEON_RING_TYPE_GFX_INDEX] = { + .ib_execute = &r100_ring_ib_execute, + .emit_fence = &r300_fence_ring_emit, + .emit_semaphore = &r100_semaphore_ring_emit, + .cs_parse = &r300_cs_parse, + .ring_start = &r300_ring_start, + .ring_test = &r100_ring_test, + .ib_test = &r100_ib_test, + } + }, + .irq = { + .set = &rs600_irq_set, + .process = &rs600_irq_process, + }, + .display = { + .get_vblank_counter = &rs600_get_vblank_counter, + .bandwidth_update = &rs690_bandwidth_update, + .wait_for_vblank = &avivo_wait_for_vblank, + }, + .copy = { + .blit = &r100_copy_blit, + .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .dma = &r200_copy_dma, + .dma_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .copy = &r200_copy_dma, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, + }, + .surface = { + .set_reg = r100_set_surface_reg, + .clear_reg = r100_clear_surface_reg, + }, + .hpd = { + .init = &rs600_hpd_init, + .fini = &rs600_hpd_fini, + .sense = &rs600_hpd_sense, + .set_polarity = &rs600_hpd_set_polarity, + }, + .pm = { + .misc = &rs600_pm_misc, + .prepare = &rs600_pm_prepare, + .finish = &rs600_pm_finish, + .init_profile = &r420_pm_init_profile, + .get_dynpm_state = &r100_pm_get_dynpm_state, + .get_engine_clock = &radeon_atom_get_engine_clock, + .set_engine_clock = &radeon_atom_set_engine_clock, + .get_memory_clock = &radeon_atom_get_memory_clock, + .set_memory_clock = &radeon_atom_set_memory_clock, + .get_pcie_lanes = NULL, + .set_pcie_lanes = NULL, + .set_clock_gating = &radeon_atom_set_clock_gating, + }, + .pflip = { + .pre_page_flip = &rs600_pre_page_flip, + .page_flip = &rs600_page_flip, + .post_page_flip = &rs600_post_page_flip, + }, }; static struct radeon_asic rv515_asic = { @@ -518,44 +728,70 @@ static struct radeon_asic rv515_asic = { .vga_set_state = &r100_vga_set_state, .gpu_is_lockup = &r300_gpu_is_lockup, .asic_reset = &rs600_asic_reset, - .gart_tlb_flush = &rv370_pcie_gart_tlb_flush, - .gart_set_page = &rv370_pcie_gart_set_page, - .cp_commit = &r100_cp_commit, - .ring_start = &rv515_ring_start, - .ring_test = &r100_ring_test, - .ring_ib_execute = &r100_ring_ib_execute, - .irq_set = &rs600_irq_set, - .irq_process = &rs600_irq_process, - .get_vblank_counter = &rs600_get_vblank_counter, - .fence_ring_emit = &r300_fence_ring_emit, - .cs_parse = &r300_cs_parse, - .copy_blit = &r100_copy_blit, - .copy_dma = &r200_copy_dma, - .copy = &r100_copy_blit, - .get_engine_clock = &radeon_atom_get_engine_clock, - .set_engine_clock = &radeon_atom_set_engine_clock, - .get_memory_clock = &radeon_atom_get_memory_clock, - .set_memory_clock = &radeon_atom_set_memory_clock, - .get_pcie_lanes = &rv370_get_pcie_lanes, - .set_pcie_lanes = &rv370_set_pcie_lanes, - .set_clock_gating = &radeon_atom_set_clock_gating, - .set_surface_reg = r100_set_surface_reg, - .clear_surface_reg = r100_clear_surface_reg, - .bandwidth_update = &rv515_bandwidth_update, - .hpd_init = &rs600_hpd_init, - .hpd_fini = &rs600_hpd_fini, - .hpd_sense = &rs600_hpd_sense, - .hpd_set_polarity = &rs600_hpd_set_polarity, .ioctl_wait_idle = NULL, .gui_idle = &r100_gui_idle, - .pm_misc = &rs600_pm_misc, - .pm_prepare = &rs600_pm_prepare, - .pm_finish = &rs600_pm_finish, - .pm_init_profile = &r420_pm_init_profile, - .pm_get_dynpm_state = &r100_pm_get_dynpm_state, - .pre_page_flip = &rs600_pre_page_flip, - .page_flip = &rs600_page_flip, - .post_page_flip = &rs600_post_page_flip, + .mc_wait_for_idle = &rv515_mc_wait_for_idle, + .gart = { + .tlb_flush = &rv370_pcie_gart_tlb_flush, + .set_page = &rv370_pcie_gart_set_page, + }, + .ring = { + [RADEON_RING_TYPE_GFX_INDEX] = { + .ib_execute = &r100_ring_ib_execute, + .emit_fence = &r300_fence_ring_emit, + .emit_semaphore = &r100_semaphore_ring_emit, + .cs_parse = &r300_cs_parse, + .ring_start = &rv515_ring_start, + .ring_test = &r100_ring_test, + .ib_test = &r100_ib_test, + } + }, + .irq = { + .set = &rs600_irq_set, + .process = &rs600_irq_process, + }, + .display = { + .get_vblank_counter = &rs600_get_vblank_counter, + .bandwidth_update = &rv515_bandwidth_update, + .wait_for_vblank = &avivo_wait_for_vblank, + }, + .copy = { + .blit = &r100_copy_blit, + .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .dma = &r200_copy_dma, + .dma_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .copy = &r100_copy_blit, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, + }, + .surface = { + .set_reg = r100_set_surface_reg, + .clear_reg = r100_clear_surface_reg, + }, + .hpd = { + .init = &rs600_hpd_init, + .fini = &rs600_hpd_fini, + .sense = &rs600_hpd_sense, + .set_polarity = &rs600_hpd_set_polarity, + }, + .pm = { + .misc = &rs600_pm_misc, + .prepare = &rs600_pm_prepare, + .finish = &rs600_pm_finish, + .init_profile = &r420_pm_init_profile, + .get_dynpm_state = &r100_pm_get_dynpm_state, + .get_engine_clock = &radeon_atom_get_engine_clock, + .set_engine_clock = &radeon_atom_set_engine_clock, + .get_memory_clock = &radeon_atom_get_memory_clock, + .set_memory_clock = &radeon_atom_set_memory_clock, + .get_pcie_lanes = &rv370_get_pcie_lanes, + .set_pcie_lanes = &rv370_set_pcie_lanes, + .set_clock_gating = &radeon_atom_set_clock_gating, + }, + .pflip = { + .pre_page_flip = &rs600_pre_page_flip, + .page_flip = &rs600_page_flip, + .post_page_flip = &rs600_post_page_flip, + }, }; static struct radeon_asic r520_asic = { @@ -566,44 +802,70 @@ static struct radeon_asic r520_asic = { .vga_set_state = &r100_vga_set_state, .gpu_is_lockup = &r300_gpu_is_lockup, .asic_reset = &rs600_asic_reset, - .gart_tlb_flush = &rv370_pcie_gart_tlb_flush, - .gart_set_page = &rv370_pcie_gart_set_page, - .cp_commit = &r100_cp_commit, - .ring_start = &rv515_ring_start, - .ring_test = &r100_ring_test, - .ring_ib_execute = &r100_ring_ib_execute, - .irq_set = &rs600_irq_set, - .irq_process = &rs600_irq_process, - .get_vblank_counter = &rs600_get_vblank_counter, - .fence_ring_emit = &r300_fence_ring_emit, - .cs_parse = &r300_cs_parse, - .copy_blit = &r100_copy_blit, - .copy_dma = &r200_copy_dma, - .copy = &r100_copy_blit, - .get_engine_clock = &radeon_atom_get_engine_clock, - .set_engine_clock = &radeon_atom_set_engine_clock, - .get_memory_clock = &radeon_atom_get_memory_clock, - .set_memory_clock = &radeon_atom_set_memory_clock, - .get_pcie_lanes = &rv370_get_pcie_lanes, - .set_pcie_lanes = &rv370_set_pcie_lanes, - .set_clock_gating = &radeon_atom_set_clock_gating, - .set_surface_reg = r100_set_surface_reg, - .clear_surface_reg = r100_clear_surface_reg, - .bandwidth_update = &rv515_bandwidth_update, - .hpd_init = &rs600_hpd_init, - .hpd_fini = &rs600_hpd_fini, - .hpd_sense = &rs600_hpd_sense, - .hpd_set_polarity = &rs600_hpd_set_polarity, .ioctl_wait_idle = NULL, .gui_idle = &r100_gui_idle, - .pm_misc = &rs600_pm_misc, - .pm_prepare = &rs600_pm_prepare, - .pm_finish = &rs600_pm_finish, - .pm_init_profile = &r420_pm_init_profile, - .pm_get_dynpm_state = &r100_pm_get_dynpm_state, - .pre_page_flip = &rs600_pre_page_flip, - .page_flip = &rs600_page_flip, - .post_page_flip = &rs600_post_page_flip, + .mc_wait_for_idle = &r520_mc_wait_for_idle, + .gart = { + .tlb_flush = &rv370_pcie_gart_tlb_flush, + .set_page = &rv370_pcie_gart_set_page, + }, + .ring = { + [RADEON_RING_TYPE_GFX_INDEX] = { + .ib_execute = &r100_ring_ib_execute, + .emit_fence = &r300_fence_ring_emit, + .emit_semaphore = &r100_semaphore_ring_emit, + .cs_parse = &r300_cs_parse, + .ring_start = &rv515_ring_start, + .ring_test = &r100_ring_test, + .ib_test = &r100_ib_test, + } + }, + .irq = { + .set = &rs600_irq_set, + .process = &rs600_irq_process, + }, + .display = { + .bandwidth_update = &rv515_bandwidth_update, + .get_vblank_counter = &rs600_get_vblank_counter, + .wait_for_vblank = &avivo_wait_for_vblank, + }, + .copy = { + .blit = &r100_copy_blit, + .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .dma = &r200_copy_dma, + .dma_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .copy = &r100_copy_blit, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, + }, + .surface = { + .set_reg = r100_set_surface_reg, + .clear_reg = r100_clear_surface_reg, + }, + .hpd = { + .init = &rs600_hpd_init, + .fini = &rs600_hpd_fini, + .sense = &rs600_hpd_sense, + .set_polarity = &rs600_hpd_set_polarity, + }, + .pm = { + .misc = &rs600_pm_misc, + .prepare = &rs600_pm_prepare, + .finish = &rs600_pm_finish, + .init_profile = &r420_pm_init_profile, + .get_dynpm_state = &r100_pm_get_dynpm_state, + .get_engine_clock = &radeon_atom_get_engine_clock, + .set_engine_clock = &radeon_atom_set_engine_clock, + .get_memory_clock = &radeon_atom_get_memory_clock, + .set_memory_clock = &radeon_atom_set_memory_clock, + .get_pcie_lanes = &rv370_get_pcie_lanes, + .set_pcie_lanes = &rv370_set_pcie_lanes, + .set_clock_gating = &radeon_atom_set_clock_gating, + }, + .pflip = { + .pre_page_flip = &rs600_pre_page_flip, + .page_flip = &rs600_page_flip, + .post_page_flip = &rs600_post_page_flip, + }, }; static struct radeon_asic r600_asic = { @@ -611,46 +873,72 @@ static struct radeon_asic r600_asic = { .fini = &r600_fini, .suspend = &r600_suspend, .resume = &r600_resume, - .cp_commit = &r600_cp_commit, .vga_set_state = &r600_vga_set_state, .gpu_is_lockup = &r600_gpu_is_lockup, .asic_reset = &r600_asic_reset, - .gart_tlb_flush = &r600_pcie_gart_tlb_flush, - .gart_set_page = &rs600_gart_set_page, - .ring_test = &r600_ring_test, - .ring_ib_execute = &r600_ring_ib_execute, - .irq_set = &r600_irq_set, - .irq_process = &r600_irq_process, - .get_vblank_counter = &rs600_get_vblank_counter, - .fence_ring_emit = &r600_fence_ring_emit, - .cs_parse = &r600_cs_parse, - .copy_blit = &r600_copy_blit, - .copy_dma = NULL, - .copy = &r600_copy_blit, - .get_engine_clock = &radeon_atom_get_engine_clock, - .set_engine_clock = &radeon_atom_set_engine_clock, - .get_memory_clock = &radeon_atom_get_memory_clock, - .set_memory_clock = &radeon_atom_set_memory_clock, - .get_pcie_lanes = &r600_get_pcie_lanes, - .set_pcie_lanes = &r600_set_pcie_lanes, - .set_clock_gating = NULL, - .set_surface_reg = r600_set_surface_reg, - .clear_surface_reg = r600_clear_surface_reg, - .bandwidth_update = &rv515_bandwidth_update, - .hpd_init = &r600_hpd_init, - .hpd_fini = &r600_hpd_fini, - .hpd_sense = &r600_hpd_sense, - .hpd_set_polarity = &r600_hpd_set_polarity, .ioctl_wait_idle = r600_ioctl_wait_idle, .gui_idle = &r600_gui_idle, - .pm_misc = &r600_pm_misc, - .pm_prepare = &rs600_pm_prepare, - .pm_finish = &rs600_pm_finish, - .pm_init_profile = &r600_pm_init_profile, - .pm_get_dynpm_state = &r600_pm_get_dynpm_state, - .pre_page_flip = &rs600_pre_page_flip, - .page_flip = &rs600_page_flip, - .post_page_flip = &rs600_post_page_flip, + .mc_wait_for_idle = &r600_mc_wait_for_idle, + .gart = { + .tlb_flush = &r600_pcie_gart_tlb_flush, + .set_page = &rs600_gart_set_page, + }, + .ring = { + [RADEON_RING_TYPE_GFX_INDEX] = { + .ib_execute = &r600_ring_ib_execute, + .emit_fence = &r600_fence_ring_emit, + .emit_semaphore = &r600_semaphore_ring_emit, + .cs_parse = &r600_cs_parse, + .ring_test = &r600_ring_test, + .ib_test = &r600_ib_test, + } + }, + .irq = { + .set = &r600_irq_set, + .process = &r600_irq_process, + }, + .display = { + .bandwidth_update = &rv515_bandwidth_update, + .get_vblank_counter = &rs600_get_vblank_counter, + .wait_for_vblank = &avivo_wait_for_vblank, + }, + .copy = { + .blit = &r600_copy_blit, + .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .dma = NULL, + .dma_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .copy = &r600_copy_blit, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, + }, + .surface = { + .set_reg = r600_set_surface_reg, + .clear_reg = r600_clear_surface_reg, + }, + .hpd = { + .init = &r600_hpd_init, + .fini = &r600_hpd_fini, + .sense = &r600_hpd_sense, + .set_polarity = &r600_hpd_set_polarity, + }, + .pm = { + .misc = &r600_pm_misc, + .prepare = &rs600_pm_prepare, + .finish = &rs600_pm_finish, + .init_profile = &r600_pm_init_profile, + .get_dynpm_state = &r600_pm_get_dynpm_state, + .get_engine_clock = &radeon_atom_get_engine_clock, + .set_engine_clock = &radeon_atom_set_engine_clock, + .get_memory_clock = &radeon_atom_get_memory_clock, + .set_memory_clock = &radeon_atom_set_memory_clock, + .get_pcie_lanes = &r600_get_pcie_lanes, + .set_pcie_lanes = &r600_set_pcie_lanes, + .set_clock_gating = NULL, + }, + .pflip = { + .pre_page_flip = &rs600_pre_page_flip, + .page_flip = &rs600_page_flip, + .post_page_flip = &rs600_post_page_flip, + }, }; static struct radeon_asic rs780_asic = { @@ -658,46 +946,72 @@ static struct radeon_asic rs780_asic = { .fini = &r600_fini, .suspend = &r600_suspend, .resume = &r600_resume, - .cp_commit = &r600_cp_commit, .gpu_is_lockup = &r600_gpu_is_lockup, .vga_set_state = &r600_vga_set_state, .asic_reset = &r600_asic_reset, - .gart_tlb_flush = &r600_pcie_gart_tlb_flush, - .gart_set_page = &rs600_gart_set_page, - .ring_test = &r600_ring_test, - .ring_ib_execute = &r600_ring_ib_execute, - .irq_set = &r600_irq_set, - .irq_process = &r600_irq_process, - .get_vblank_counter = &rs600_get_vblank_counter, - .fence_ring_emit = &r600_fence_ring_emit, - .cs_parse = &r600_cs_parse, - .copy_blit = &r600_copy_blit, - .copy_dma = NULL, - .copy = &r600_copy_blit, - .get_engine_clock = &radeon_atom_get_engine_clock, - .set_engine_clock = &radeon_atom_set_engine_clock, - .get_memory_clock = NULL, - .set_memory_clock = NULL, - .get_pcie_lanes = NULL, - .set_pcie_lanes = NULL, - .set_clock_gating = NULL, - .set_surface_reg = r600_set_surface_reg, - .clear_surface_reg = r600_clear_surface_reg, - .bandwidth_update = &rs690_bandwidth_update, - .hpd_init = &r600_hpd_init, - .hpd_fini = &r600_hpd_fini, - .hpd_sense = &r600_hpd_sense, - .hpd_set_polarity = &r600_hpd_set_polarity, .ioctl_wait_idle = r600_ioctl_wait_idle, .gui_idle = &r600_gui_idle, - .pm_misc = &r600_pm_misc, - .pm_prepare = &rs600_pm_prepare, - .pm_finish = &rs600_pm_finish, - .pm_init_profile = &rs780_pm_init_profile, - .pm_get_dynpm_state = &r600_pm_get_dynpm_state, - .pre_page_flip = &rs600_pre_page_flip, - .page_flip = &rs600_page_flip, - .post_page_flip = &rs600_post_page_flip, + .mc_wait_for_idle = &r600_mc_wait_for_idle, + .gart = { + .tlb_flush = &r600_pcie_gart_tlb_flush, + .set_page = &rs600_gart_set_page, + }, + .ring = { + [RADEON_RING_TYPE_GFX_INDEX] = { + .ib_execute = &r600_ring_ib_execute, + .emit_fence = &r600_fence_ring_emit, + .emit_semaphore = &r600_semaphore_ring_emit, + .cs_parse = &r600_cs_parse, + .ring_test = &r600_ring_test, + .ib_test = &r600_ib_test, + } + }, + .irq = { + .set = &r600_irq_set, + .process = &r600_irq_process, + }, + .display = { + .bandwidth_update = &rs690_bandwidth_update, + .get_vblank_counter = &rs600_get_vblank_counter, + .wait_for_vblank = &avivo_wait_for_vblank, + }, + .copy = { + .blit = &r600_copy_blit, + .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .dma = NULL, + .dma_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .copy = &r600_copy_blit, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, + }, + .surface = { + .set_reg = r600_set_surface_reg, + .clear_reg = r600_clear_surface_reg, + }, + .hpd = { + .init = &r600_hpd_init, + .fini = &r600_hpd_fini, + .sense = &r600_hpd_sense, + .set_polarity = &r600_hpd_set_polarity, + }, + .pm = { + .misc = &r600_pm_misc, + .prepare = &rs600_pm_prepare, + .finish = &rs600_pm_finish, + .init_profile = &rs780_pm_init_profile, + .get_dynpm_state = &r600_pm_get_dynpm_state, + .get_engine_clock = &radeon_atom_get_engine_clock, + .set_engine_clock = &radeon_atom_set_engine_clock, + .get_memory_clock = NULL, + .set_memory_clock = NULL, + .get_pcie_lanes = NULL, + .set_pcie_lanes = NULL, + .set_clock_gating = NULL, + }, + .pflip = { + .pre_page_flip = &rs600_pre_page_flip, + .page_flip = &rs600_page_flip, + .post_page_flip = &rs600_post_page_flip, + }, }; static struct radeon_asic rv770_asic = { @@ -705,46 +1019,72 @@ static struct radeon_asic rv770_asic = { .fini = &rv770_fini, .suspend = &rv770_suspend, .resume = &rv770_resume, - .cp_commit = &r600_cp_commit, .asic_reset = &r600_asic_reset, .gpu_is_lockup = &r600_gpu_is_lockup, .vga_set_state = &r600_vga_set_state, - .gart_tlb_flush = &r600_pcie_gart_tlb_flush, - .gart_set_page = &rs600_gart_set_page, - .ring_test = &r600_ring_test, - .ring_ib_execute = &r600_ring_ib_execute, - .irq_set = &r600_irq_set, - .irq_process = &r600_irq_process, - .get_vblank_counter = &rs600_get_vblank_counter, - .fence_ring_emit = &r600_fence_ring_emit, - .cs_parse = &r600_cs_parse, - .copy_blit = &r600_copy_blit, - .copy_dma = NULL, - .copy = &r600_copy_blit, - .get_engine_clock = &radeon_atom_get_engine_clock, - .set_engine_clock = &radeon_atom_set_engine_clock, - .get_memory_clock = &radeon_atom_get_memory_clock, - .set_memory_clock = &radeon_atom_set_memory_clock, - .get_pcie_lanes = &r600_get_pcie_lanes, - .set_pcie_lanes = &r600_set_pcie_lanes, - .set_clock_gating = &radeon_atom_set_clock_gating, - .set_surface_reg = r600_set_surface_reg, - .clear_surface_reg = r600_clear_surface_reg, - .bandwidth_update = &rv515_bandwidth_update, - .hpd_init = &r600_hpd_init, - .hpd_fini = &r600_hpd_fini, - .hpd_sense = &r600_hpd_sense, - .hpd_set_polarity = &r600_hpd_set_polarity, .ioctl_wait_idle = r600_ioctl_wait_idle, .gui_idle = &r600_gui_idle, - .pm_misc = &rv770_pm_misc, - .pm_prepare = &rs600_pm_prepare, - .pm_finish = &rs600_pm_finish, - .pm_init_profile = &r600_pm_init_profile, - .pm_get_dynpm_state = &r600_pm_get_dynpm_state, - .pre_page_flip = &rs600_pre_page_flip, - .page_flip = &rv770_page_flip, - .post_page_flip = &rs600_post_page_flip, + .mc_wait_for_idle = &r600_mc_wait_for_idle, + .gart = { + .tlb_flush = &r600_pcie_gart_tlb_flush, + .set_page = &rs600_gart_set_page, + }, + .ring = { + [RADEON_RING_TYPE_GFX_INDEX] = { + .ib_execute = &r600_ring_ib_execute, + .emit_fence = &r600_fence_ring_emit, + .emit_semaphore = &r600_semaphore_ring_emit, + .cs_parse = &r600_cs_parse, + .ring_test = &r600_ring_test, + .ib_test = &r600_ib_test, + } + }, + .irq = { + .set = &r600_irq_set, + .process = &r600_irq_process, + }, + .display = { + .bandwidth_update = &rv515_bandwidth_update, + .get_vblank_counter = &rs600_get_vblank_counter, + .wait_for_vblank = &avivo_wait_for_vblank, + }, + .copy = { + .blit = &r600_copy_blit, + .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .dma = NULL, + .dma_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .copy = &r600_copy_blit, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, + }, + .surface = { + .set_reg = r600_set_surface_reg, + .clear_reg = r600_clear_surface_reg, + }, + .hpd = { + .init = &r600_hpd_init, + .fini = &r600_hpd_fini, + .sense = &r600_hpd_sense, + .set_polarity = &r600_hpd_set_polarity, + }, + .pm = { + .misc = &rv770_pm_misc, + .prepare = &rs600_pm_prepare, + .finish = &rs600_pm_finish, + .init_profile = &r600_pm_init_profile, + .get_dynpm_state = &r600_pm_get_dynpm_state, + .get_engine_clock = &radeon_atom_get_engine_clock, + .set_engine_clock = &radeon_atom_set_engine_clock, + .get_memory_clock = &radeon_atom_get_memory_clock, + .set_memory_clock = &radeon_atom_set_memory_clock, + .get_pcie_lanes = &r600_get_pcie_lanes, + .set_pcie_lanes = &r600_set_pcie_lanes, + .set_clock_gating = &radeon_atom_set_clock_gating, + }, + .pflip = { + .pre_page_flip = &rs600_pre_page_flip, + .page_flip = &rv770_page_flip, + .post_page_flip = &rs600_post_page_flip, + }, }; static struct radeon_asic evergreen_asic = { @@ -752,46 +1092,72 @@ static struct radeon_asic evergreen_asic = { .fini = &evergreen_fini, .suspend = &evergreen_suspend, .resume = &evergreen_resume, - .cp_commit = &r600_cp_commit, .gpu_is_lockup = &evergreen_gpu_is_lockup, .asic_reset = &evergreen_asic_reset, .vga_set_state = &r600_vga_set_state, - .gart_tlb_flush = &evergreen_pcie_gart_tlb_flush, - .gart_set_page = &rs600_gart_set_page, - .ring_test = &r600_ring_test, - .ring_ib_execute = &evergreen_ring_ib_execute, - .irq_set = &evergreen_irq_set, - .irq_process = &evergreen_irq_process, - .get_vblank_counter = &evergreen_get_vblank_counter, - .fence_ring_emit = &r600_fence_ring_emit, - .cs_parse = &evergreen_cs_parse, - .copy_blit = &r600_copy_blit, - .copy_dma = NULL, - .copy = &r600_copy_blit, - .get_engine_clock = &radeon_atom_get_engine_clock, - .set_engine_clock = &radeon_atom_set_engine_clock, - .get_memory_clock = &radeon_atom_get_memory_clock, - .set_memory_clock = &radeon_atom_set_memory_clock, - .get_pcie_lanes = &r600_get_pcie_lanes, - .set_pcie_lanes = &r600_set_pcie_lanes, - .set_clock_gating = NULL, - .set_surface_reg = r600_set_surface_reg, - .clear_surface_reg = r600_clear_surface_reg, - .bandwidth_update = &evergreen_bandwidth_update, - .hpd_init = &evergreen_hpd_init, - .hpd_fini = &evergreen_hpd_fini, - .hpd_sense = &evergreen_hpd_sense, - .hpd_set_polarity = &evergreen_hpd_set_polarity, .ioctl_wait_idle = r600_ioctl_wait_idle, .gui_idle = &r600_gui_idle, - .pm_misc = &evergreen_pm_misc, - .pm_prepare = &evergreen_pm_prepare, - .pm_finish = &evergreen_pm_finish, - .pm_init_profile = &r600_pm_init_profile, - .pm_get_dynpm_state = &r600_pm_get_dynpm_state, - .pre_page_flip = &evergreen_pre_page_flip, - .page_flip = &evergreen_page_flip, - .post_page_flip = &evergreen_post_page_flip, + .mc_wait_for_idle = &evergreen_mc_wait_for_idle, + .gart = { + .tlb_flush = &evergreen_pcie_gart_tlb_flush, + .set_page = &rs600_gart_set_page, + }, + .ring = { + [RADEON_RING_TYPE_GFX_INDEX] = { + .ib_execute = &evergreen_ring_ib_execute, + .emit_fence = &r600_fence_ring_emit, + .emit_semaphore = &r600_semaphore_ring_emit, + .cs_parse = &evergreen_cs_parse, + .ring_test = &r600_ring_test, + .ib_test = &r600_ib_test, + } + }, + .irq = { + .set = &evergreen_irq_set, + .process = &evergreen_irq_process, + }, + .display = { + .bandwidth_update = &evergreen_bandwidth_update, + .get_vblank_counter = &evergreen_get_vblank_counter, + .wait_for_vblank = &dce4_wait_for_vblank, + }, + .copy = { + .blit = &r600_copy_blit, + .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .dma = NULL, + .dma_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .copy = &r600_copy_blit, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, + }, + .surface = { + .set_reg = r600_set_surface_reg, + .clear_reg = r600_clear_surface_reg, + }, + .hpd = { + .init = &evergreen_hpd_init, + .fini = &evergreen_hpd_fini, + .sense = &evergreen_hpd_sense, + .set_polarity = &evergreen_hpd_set_polarity, + }, + .pm = { + .misc = &evergreen_pm_misc, + .prepare = &evergreen_pm_prepare, + .finish = &evergreen_pm_finish, + .init_profile = &r600_pm_init_profile, + .get_dynpm_state = &r600_pm_get_dynpm_state, + .get_engine_clock = &radeon_atom_get_engine_clock, + .set_engine_clock = &radeon_atom_set_engine_clock, + .get_memory_clock = &radeon_atom_get_memory_clock, + .set_memory_clock = &radeon_atom_set_memory_clock, + .get_pcie_lanes = &r600_get_pcie_lanes, + .set_pcie_lanes = &r600_set_pcie_lanes, + .set_clock_gating = NULL, + }, + .pflip = { + .pre_page_flip = &evergreen_pre_page_flip, + .page_flip = &evergreen_page_flip, + .post_page_flip = &evergreen_post_page_flip, + }, }; static struct radeon_asic sumo_asic = { @@ -799,46 +1165,72 @@ static struct radeon_asic sumo_asic = { .fini = &evergreen_fini, .suspend = &evergreen_suspend, .resume = &evergreen_resume, - .cp_commit = &r600_cp_commit, .gpu_is_lockup = &evergreen_gpu_is_lockup, .asic_reset = &evergreen_asic_reset, .vga_set_state = &r600_vga_set_state, - .gart_tlb_flush = &evergreen_pcie_gart_tlb_flush, - .gart_set_page = &rs600_gart_set_page, - .ring_test = &r600_ring_test, - .ring_ib_execute = &evergreen_ring_ib_execute, - .irq_set = &evergreen_irq_set, - .irq_process = &evergreen_irq_process, - .get_vblank_counter = &evergreen_get_vblank_counter, - .fence_ring_emit = &r600_fence_ring_emit, - .cs_parse = &evergreen_cs_parse, - .copy_blit = &r600_copy_blit, - .copy_dma = NULL, - .copy = &r600_copy_blit, - .get_engine_clock = &radeon_atom_get_engine_clock, - .set_engine_clock = &radeon_atom_set_engine_clock, - .get_memory_clock = NULL, - .set_memory_clock = NULL, - .get_pcie_lanes = NULL, - .set_pcie_lanes = NULL, - .set_clock_gating = NULL, - .set_surface_reg = r600_set_surface_reg, - .clear_surface_reg = r600_clear_surface_reg, - .bandwidth_update = &evergreen_bandwidth_update, - .hpd_init = &evergreen_hpd_init, - .hpd_fini = &evergreen_hpd_fini, - .hpd_sense = &evergreen_hpd_sense, - .hpd_set_polarity = &evergreen_hpd_set_polarity, .ioctl_wait_idle = r600_ioctl_wait_idle, .gui_idle = &r600_gui_idle, - .pm_misc = &evergreen_pm_misc, - .pm_prepare = &evergreen_pm_prepare, - .pm_finish = &evergreen_pm_finish, - .pm_init_profile = &sumo_pm_init_profile, - .pm_get_dynpm_state = &r600_pm_get_dynpm_state, - .pre_page_flip = &evergreen_pre_page_flip, - .page_flip = &evergreen_page_flip, - .post_page_flip = &evergreen_post_page_flip, + .mc_wait_for_idle = &evergreen_mc_wait_for_idle, + .gart = { + .tlb_flush = &evergreen_pcie_gart_tlb_flush, + .set_page = &rs600_gart_set_page, + }, + .ring = { + [RADEON_RING_TYPE_GFX_INDEX] = { + .ib_execute = &evergreen_ring_ib_execute, + .emit_fence = &r600_fence_ring_emit, + .emit_semaphore = &r600_semaphore_ring_emit, + .cs_parse = &evergreen_cs_parse, + .ring_test = &r600_ring_test, + .ib_test = &r600_ib_test, + }, + }, + .irq = { + .set = &evergreen_irq_set, + .process = &evergreen_irq_process, + }, + .display = { + .bandwidth_update = &evergreen_bandwidth_update, + .get_vblank_counter = &evergreen_get_vblank_counter, + .wait_for_vblank = &dce4_wait_for_vblank, + }, + .copy = { + .blit = &r600_copy_blit, + .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .dma = NULL, + .dma_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .copy = &r600_copy_blit, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, + }, + .surface = { + .set_reg = r600_set_surface_reg, + .clear_reg = r600_clear_surface_reg, + }, + .hpd = { + .init = &evergreen_hpd_init, + .fini = &evergreen_hpd_fini, + .sense = &evergreen_hpd_sense, + .set_polarity = &evergreen_hpd_set_polarity, + }, + .pm = { + .misc = &evergreen_pm_misc, + .prepare = &evergreen_pm_prepare, + .finish = &evergreen_pm_finish, + .init_profile = &sumo_pm_init_profile, + .get_dynpm_state = &r600_pm_get_dynpm_state, + .get_engine_clock = &radeon_atom_get_engine_clock, + .set_engine_clock = &radeon_atom_set_engine_clock, + .get_memory_clock = NULL, + .set_memory_clock = NULL, + .get_pcie_lanes = NULL, + .set_pcie_lanes = NULL, + .set_clock_gating = NULL, + }, + .pflip = { + .pre_page_flip = &evergreen_pre_page_flip, + .page_flip = &evergreen_page_flip, + .post_page_flip = &evergreen_post_page_flip, + }, }; static struct radeon_asic btc_asic = { @@ -846,46 +1238,82 @@ static struct radeon_asic btc_asic = { .fini = &evergreen_fini, .suspend = &evergreen_suspend, .resume = &evergreen_resume, - .cp_commit = &r600_cp_commit, .gpu_is_lockup = &evergreen_gpu_is_lockup, .asic_reset = &evergreen_asic_reset, .vga_set_state = &r600_vga_set_state, - .gart_tlb_flush = &evergreen_pcie_gart_tlb_flush, - .gart_set_page = &rs600_gart_set_page, - .ring_test = &r600_ring_test, - .ring_ib_execute = &evergreen_ring_ib_execute, - .irq_set = &evergreen_irq_set, - .irq_process = &evergreen_irq_process, - .get_vblank_counter = &evergreen_get_vblank_counter, - .fence_ring_emit = &r600_fence_ring_emit, - .cs_parse = &evergreen_cs_parse, - .copy_blit = &r600_copy_blit, - .copy_dma = NULL, - .copy = &r600_copy_blit, - .get_engine_clock = &radeon_atom_get_engine_clock, - .set_engine_clock = &radeon_atom_set_engine_clock, - .get_memory_clock = &radeon_atom_get_memory_clock, - .set_memory_clock = &radeon_atom_set_memory_clock, - .get_pcie_lanes = NULL, - .set_pcie_lanes = NULL, - .set_clock_gating = NULL, - .set_surface_reg = r600_set_surface_reg, - .clear_surface_reg = r600_clear_surface_reg, - .bandwidth_update = &evergreen_bandwidth_update, - .hpd_init = &evergreen_hpd_init, - .hpd_fini = &evergreen_hpd_fini, - .hpd_sense = &evergreen_hpd_sense, - .hpd_set_polarity = &evergreen_hpd_set_polarity, .ioctl_wait_idle = r600_ioctl_wait_idle, .gui_idle = &r600_gui_idle, - .pm_misc = &evergreen_pm_misc, - .pm_prepare = &evergreen_pm_prepare, - .pm_finish = &evergreen_pm_finish, - .pm_init_profile = &r600_pm_init_profile, - .pm_get_dynpm_state = &r600_pm_get_dynpm_state, - .pre_page_flip = &evergreen_pre_page_flip, - .page_flip = &evergreen_page_flip, - .post_page_flip = &evergreen_post_page_flip, + .mc_wait_for_idle = &evergreen_mc_wait_for_idle, + .gart = { + .tlb_flush = &evergreen_pcie_gart_tlb_flush, + .set_page = &rs600_gart_set_page, + }, + .ring = { + [RADEON_RING_TYPE_GFX_INDEX] = { + .ib_execute = &evergreen_ring_ib_execute, + .emit_fence = &r600_fence_ring_emit, + .emit_semaphore = &r600_semaphore_ring_emit, + .cs_parse = &evergreen_cs_parse, + .ring_test = &r600_ring_test, + .ib_test = &r600_ib_test, + } + }, + .irq = { + .set = &evergreen_irq_set, + .process = &evergreen_irq_process, + }, + .display = { + .bandwidth_update = &evergreen_bandwidth_update, + .get_vblank_counter = &evergreen_get_vblank_counter, + .wait_for_vblank = &dce4_wait_for_vblank, + }, + .copy = { + .blit = &r600_copy_blit, + .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .dma = NULL, + .dma_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .copy = &r600_copy_blit, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, + }, + .surface = { + .set_reg = r600_set_surface_reg, + .clear_reg = r600_clear_surface_reg, + }, + .hpd = { + .init = &evergreen_hpd_init, + .fini = &evergreen_hpd_fini, + .sense = &evergreen_hpd_sense, + .set_polarity = &evergreen_hpd_set_polarity, + }, + .pm = { + .misc = &evergreen_pm_misc, + .prepare = &evergreen_pm_prepare, + .finish = &evergreen_pm_finish, + .init_profile = &r600_pm_init_profile, + .get_dynpm_state = &r600_pm_get_dynpm_state, + .get_engine_clock = &radeon_atom_get_engine_clock, + .set_engine_clock = &radeon_atom_set_engine_clock, + .get_memory_clock = &radeon_atom_get_memory_clock, + .set_memory_clock = &radeon_atom_set_memory_clock, + .get_pcie_lanes = NULL, + .set_pcie_lanes = NULL, + .set_clock_gating = NULL, + }, + .pflip = { + .pre_page_flip = &evergreen_pre_page_flip, + .page_flip = &evergreen_page_flip, + .post_page_flip = &evergreen_post_page_flip, + }, +}; + +static const struct radeon_vm_funcs cayman_vm_funcs = { + .init = &cayman_vm_init, + .fini = &cayman_vm_fini, + .bind = &cayman_vm_bind, + .unbind = &cayman_vm_unbind, + .tlb_flush = &cayman_vm_tlb_flush, + .page_flags = &cayman_vm_page_flags, + .set_page = &cayman_vm_set_page, }; static struct radeon_asic cayman_asic = { @@ -893,46 +1321,285 @@ static struct radeon_asic cayman_asic = { .fini = &cayman_fini, .suspend = &cayman_suspend, .resume = &cayman_resume, - .cp_commit = &r600_cp_commit, .gpu_is_lockup = &cayman_gpu_is_lockup, .asic_reset = &cayman_asic_reset, .vga_set_state = &r600_vga_set_state, - .gart_tlb_flush = &cayman_pcie_gart_tlb_flush, - .gart_set_page = &rs600_gart_set_page, - .ring_test = &r600_ring_test, - .ring_ib_execute = &evergreen_ring_ib_execute, - .irq_set = &evergreen_irq_set, - .irq_process = &evergreen_irq_process, - .get_vblank_counter = &evergreen_get_vblank_counter, - .fence_ring_emit = &r600_fence_ring_emit, - .cs_parse = &evergreen_cs_parse, - .copy_blit = &r600_copy_blit, - .copy_dma = NULL, - .copy = &r600_copy_blit, - .get_engine_clock = &radeon_atom_get_engine_clock, - .set_engine_clock = &radeon_atom_set_engine_clock, - .get_memory_clock = &radeon_atom_get_memory_clock, - .set_memory_clock = &radeon_atom_set_memory_clock, - .get_pcie_lanes = NULL, - .set_pcie_lanes = NULL, - .set_clock_gating = NULL, - .set_surface_reg = r600_set_surface_reg, - .clear_surface_reg = r600_clear_surface_reg, - .bandwidth_update = &evergreen_bandwidth_update, - .hpd_init = &evergreen_hpd_init, - .hpd_fini = &evergreen_hpd_fini, - .hpd_sense = &evergreen_hpd_sense, - .hpd_set_polarity = &evergreen_hpd_set_polarity, .ioctl_wait_idle = r600_ioctl_wait_idle, .gui_idle = &r600_gui_idle, - .pm_misc = &evergreen_pm_misc, - .pm_prepare = &evergreen_pm_prepare, - .pm_finish = &evergreen_pm_finish, - .pm_init_profile = &r600_pm_init_profile, - .pm_get_dynpm_state = &r600_pm_get_dynpm_state, - .pre_page_flip = &evergreen_pre_page_flip, - .page_flip = &evergreen_page_flip, - .post_page_flip = &evergreen_post_page_flip, + .mc_wait_for_idle = &evergreen_mc_wait_for_idle, + .gart = { + .tlb_flush = &cayman_pcie_gart_tlb_flush, + .set_page = &rs600_gart_set_page, + }, + .ring = { + [RADEON_RING_TYPE_GFX_INDEX] = { + .ib_execute = &cayman_ring_ib_execute, + .ib_parse = &evergreen_ib_parse, + .emit_fence = &cayman_fence_ring_emit, + .emit_semaphore = &r600_semaphore_ring_emit, + .cs_parse = &evergreen_cs_parse, + .ring_test = &r600_ring_test, + .ib_test = &r600_ib_test, + }, + [CAYMAN_RING_TYPE_CP1_INDEX] = { + .ib_execute = &cayman_ring_ib_execute, + .ib_parse = &evergreen_ib_parse, + .emit_fence = &cayman_fence_ring_emit, + .emit_semaphore = &r600_semaphore_ring_emit, + .cs_parse = &evergreen_cs_parse, + .ring_test = &r600_ring_test, + .ib_test = &r600_ib_test, + }, + [CAYMAN_RING_TYPE_CP2_INDEX] = { + .ib_execute = &cayman_ring_ib_execute, + .ib_parse = &evergreen_ib_parse, + .emit_fence = &cayman_fence_ring_emit, + .emit_semaphore = &r600_semaphore_ring_emit, + .cs_parse = &evergreen_cs_parse, + .ring_test = &r600_ring_test, + .ib_test = &r600_ib_test, + } + }, + .irq = { + .set = &evergreen_irq_set, + .process = &evergreen_irq_process, + }, + .display = { + .bandwidth_update = &evergreen_bandwidth_update, + .get_vblank_counter = &evergreen_get_vblank_counter, + .wait_for_vblank = &dce4_wait_for_vblank, + }, + .copy = { + .blit = &r600_copy_blit, + .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .dma = NULL, + .dma_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .copy = &r600_copy_blit, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, + }, + .surface = { + .set_reg = r600_set_surface_reg, + .clear_reg = r600_clear_surface_reg, + }, + .hpd = { + .init = &evergreen_hpd_init, + .fini = &evergreen_hpd_fini, + .sense = &evergreen_hpd_sense, + .set_polarity = &evergreen_hpd_set_polarity, + }, + .pm = { + .misc = &evergreen_pm_misc, + .prepare = &evergreen_pm_prepare, + .finish = &evergreen_pm_finish, + .init_profile = &r600_pm_init_profile, + .get_dynpm_state = &r600_pm_get_dynpm_state, + .get_engine_clock = &radeon_atom_get_engine_clock, + .set_engine_clock = &radeon_atom_set_engine_clock, + .get_memory_clock = &radeon_atom_get_memory_clock, + .set_memory_clock = &radeon_atom_set_memory_clock, + .get_pcie_lanes = NULL, + .set_pcie_lanes = NULL, + .set_clock_gating = NULL, + }, + .pflip = { + .pre_page_flip = &evergreen_pre_page_flip, + .page_flip = &evergreen_page_flip, + .post_page_flip = &evergreen_post_page_flip, + }, +}; + +static struct radeon_asic trinity_asic = { + .init = &cayman_init, + .fini = &cayman_fini, + .suspend = &cayman_suspend, + .resume = &cayman_resume, + .gpu_is_lockup = &cayman_gpu_is_lockup, + .asic_reset = &cayman_asic_reset, + .vga_set_state = &r600_vga_set_state, + .ioctl_wait_idle = r600_ioctl_wait_idle, + .gui_idle = &r600_gui_idle, + .mc_wait_for_idle = &evergreen_mc_wait_for_idle, + .gart = { + .tlb_flush = &cayman_pcie_gart_tlb_flush, + .set_page = &rs600_gart_set_page, + }, + .ring = { + [RADEON_RING_TYPE_GFX_INDEX] = { + .ib_execute = &cayman_ring_ib_execute, + .ib_parse = &evergreen_ib_parse, + .emit_fence = &cayman_fence_ring_emit, + .emit_semaphore = &r600_semaphore_ring_emit, + .cs_parse = &evergreen_cs_parse, + .ring_test = &r600_ring_test, + .ib_test = &r600_ib_test, + }, + [CAYMAN_RING_TYPE_CP1_INDEX] = { + .ib_execute = &cayman_ring_ib_execute, + .ib_parse = &evergreen_ib_parse, + .emit_fence = &cayman_fence_ring_emit, + .emit_semaphore = &r600_semaphore_ring_emit, + .cs_parse = &evergreen_cs_parse, + .ring_test = &r600_ring_test, + .ib_test = &r600_ib_test, + }, + [CAYMAN_RING_TYPE_CP2_INDEX] = { + .ib_execute = &cayman_ring_ib_execute, + .ib_parse = &evergreen_ib_parse, + .emit_fence = &cayman_fence_ring_emit, + .emit_semaphore = &r600_semaphore_ring_emit, + .cs_parse = &evergreen_cs_parse, + .ring_test = &r600_ring_test, + .ib_test = &r600_ib_test, + } + }, + .irq = { + .set = &evergreen_irq_set, + .process = &evergreen_irq_process, + }, + .display = { + .bandwidth_update = &dce6_bandwidth_update, + .get_vblank_counter = &evergreen_get_vblank_counter, + .wait_for_vblank = &dce4_wait_for_vblank, + }, + .copy = { + .blit = &r600_copy_blit, + .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .dma = NULL, + .dma_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .copy = &r600_copy_blit, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, + }, + .surface = { + .set_reg = r600_set_surface_reg, + .clear_reg = r600_clear_surface_reg, + }, + .hpd = { + .init = &evergreen_hpd_init, + .fini = &evergreen_hpd_fini, + .sense = &evergreen_hpd_sense, + .set_polarity = &evergreen_hpd_set_polarity, + }, + .pm = { + .misc = &evergreen_pm_misc, + .prepare = &evergreen_pm_prepare, + .finish = &evergreen_pm_finish, + .init_profile = &sumo_pm_init_profile, + .get_dynpm_state = &r600_pm_get_dynpm_state, + .get_engine_clock = &radeon_atom_get_engine_clock, + .set_engine_clock = &radeon_atom_set_engine_clock, + .get_memory_clock = NULL, + .set_memory_clock = NULL, + .get_pcie_lanes = NULL, + .set_pcie_lanes = NULL, + .set_clock_gating = NULL, + }, + .pflip = { + .pre_page_flip = &evergreen_pre_page_flip, + .page_flip = &evergreen_page_flip, + .post_page_flip = &evergreen_post_page_flip, + }, +}; + +static const struct radeon_vm_funcs si_vm_funcs = { + .init = &si_vm_init, + .fini = &si_vm_fini, + .bind = &si_vm_bind, + .unbind = &si_vm_unbind, + .tlb_flush = &si_vm_tlb_flush, + .page_flags = &cayman_vm_page_flags, + .set_page = &cayman_vm_set_page, +}; + +static struct radeon_asic si_asic = { + .init = &si_init, + .fini = &si_fini, + .suspend = &si_suspend, + .resume = &si_resume, + .gpu_is_lockup = &si_gpu_is_lockup, + .asic_reset = &si_asic_reset, + .vga_set_state = &r600_vga_set_state, + .ioctl_wait_idle = r600_ioctl_wait_idle, + .gui_idle = &r600_gui_idle, + .mc_wait_for_idle = &evergreen_mc_wait_for_idle, + .gart = { + .tlb_flush = &si_pcie_gart_tlb_flush, + .set_page = &rs600_gart_set_page, + }, + .ring = { + [RADEON_RING_TYPE_GFX_INDEX] = { + .ib_execute = &si_ring_ib_execute, + .ib_parse = &si_ib_parse, + .emit_fence = &si_fence_ring_emit, + .emit_semaphore = &r600_semaphore_ring_emit, + .cs_parse = NULL, + .ring_test = &r600_ring_test, + .ib_test = &r600_ib_test, + }, + [CAYMAN_RING_TYPE_CP1_INDEX] = { + .ib_execute = &si_ring_ib_execute, + .ib_parse = &si_ib_parse, + .emit_fence = &si_fence_ring_emit, + .emit_semaphore = &r600_semaphore_ring_emit, + .cs_parse = NULL, + .ring_test = &r600_ring_test, + .ib_test = &r600_ib_test, + }, + [CAYMAN_RING_TYPE_CP2_INDEX] = { + .ib_execute = &si_ring_ib_execute, + .ib_parse = &si_ib_parse, + .emit_fence = &si_fence_ring_emit, + .emit_semaphore = &r600_semaphore_ring_emit, + .cs_parse = NULL, + .ring_test = &r600_ring_test, + .ib_test = &r600_ib_test, + } + }, + .irq = { + .set = &si_irq_set, + .process = &si_irq_process, + }, + .display = { + .bandwidth_update = &dce6_bandwidth_update, + .get_vblank_counter = &evergreen_get_vblank_counter, + .wait_for_vblank = &dce4_wait_for_vblank, + }, + .copy = { + .blit = NULL, + .blit_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .dma = NULL, + .dma_ring_index = RADEON_RING_TYPE_GFX_INDEX, + .copy = NULL, + .copy_ring_index = RADEON_RING_TYPE_GFX_INDEX, + }, + .surface = { + .set_reg = r600_set_surface_reg, + .clear_reg = r600_clear_surface_reg, + }, + .hpd = { + .init = &evergreen_hpd_init, + .fini = &evergreen_hpd_fini, + .sense = &evergreen_hpd_sense, + .set_polarity = &evergreen_hpd_set_polarity, + }, + .pm = { + .misc = &evergreen_pm_misc, + .prepare = &evergreen_pm_prepare, + .finish = &evergreen_pm_finish, + .init_profile = &sumo_pm_init_profile, + .get_dynpm_state = &r600_pm_get_dynpm_state, + .get_engine_clock = &radeon_atom_get_engine_clock, + .set_engine_clock = &radeon_atom_set_engine_clock, + .get_memory_clock = &radeon_atom_get_memory_clock, + .set_memory_clock = &radeon_atom_set_memory_clock, + .get_pcie_lanes = NULL, + .set_pcie_lanes = NULL, + .set_clock_gating = NULL, + }, + .pflip = { + .pre_page_flip = &evergreen_pre_page_flip, + .page_flip = &evergreen_page_flip, + .post_page_flip = &evergreen_post_page_flip, + }, }; int radeon_asic_init(struct radeon_device *rdev) @@ -974,10 +1641,10 @@ int radeon_asic_init(struct radeon_device *rdev) rdev->asic = &r420_asic; /* handle macs */ if (rdev->bios == NULL) { - rdev->asic->get_engine_clock = &radeon_legacy_get_engine_clock; - rdev->asic->set_engine_clock = &radeon_legacy_set_engine_clock; - rdev->asic->get_memory_clock = &radeon_legacy_get_memory_clock; - rdev->asic->set_memory_clock = NULL; + rdev->asic->pm.get_engine_clock = &radeon_legacy_get_engine_clock; + rdev->asic->pm.set_engine_clock = &radeon_legacy_set_engine_clock; + rdev->asic->pm.get_memory_clock = &radeon_legacy_get_memory_clock; + rdev->asic->pm.set_memory_clock = NULL; } break; case CHIP_RS400: @@ -1050,6 +1717,21 @@ int radeon_asic_init(struct radeon_device *rdev) rdev->asic = &cayman_asic; /* set num crtcs */ rdev->num_crtc = 6; + rdev->vm_manager.funcs = &cayman_vm_funcs; + break; + case CHIP_ARUBA: + rdev->asic = &trinity_asic; + /* set num crtcs */ + rdev->num_crtc = 4; + rdev->vm_manager.funcs = &cayman_vm_funcs; + break; + case CHIP_TAHITI: + case CHIP_PITCAIRN: + case CHIP_VERDE: + rdev->asic = &si_asic; + /* set num crtcs */ + rdev->num_crtc = 6; + rdev->vm_manager.funcs = &si_vm_funcs; break; default: /* FIXME: not supported yet */ @@ -1057,8 +1739,8 @@ int radeon_asic_init(struct radeon_device *rdev) } if (rdev->flags & RADEON_IS_IGP) { - rdev->asic->get_memory_clock = NULL; - rdev->asic->set_memory_clock = NULL; + rdev->asic->pm.get_memory_clock = NULL; + rdev->asic->pm.set_memory_clock = NULL; } return 0; diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h index 5aa6670..917e49c 100644 --- a/drivers/gpu/drm/radeon/radeon_asic.h +++ b/drivers/gpu/drm/radeon/radeon_asic.h @@ -58,17 +58,20 @@ void r100_fini(struct radeon_device *rdev); int r100_suspend(struct radeon_device *rdev); int r100_resume(struct radeon_device *rdev); void r100_vga_set_state(struct radeon_device *rdev, bool state); -bool r100_gpu_is_lockup(struct radeon_device *rdev); +bool r100_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp); int r100_asic_reset(struct radeon_device *rdev); u32 r100_get_vblank_counter(struct radeon_device *rdev, int crtc); void r100_pci_gart_tlb_flush(struct radeon_device *rdev); int r100_pci_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr); -void r100_cp_commit(struct radeon_device *rdev); -void r100_ring_start(struct radeon_device *rdev); +void r100_ring_start(struct radeon_device *rdev, struct radeon_ring *ring); int r100_irq_set(struct radeon_device *rdev); int r100_irq_process(struct radeon_device *rdev); void r100_fence_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence); +void r100_semaphore_ring_emit(struct radeon_device *rdev, + struct radeon_ring *cp, + struct radeon_semaphore *semaphore, + bool emit_wait); int r100_cs_parse(struct radeon_cs_parser *p); void r100_pll_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v); uint32_t r100_pll_rreg(struct radeon_device *rdev, uint32_t reg); @@ -83,7 +86,7 @@ int r100_set_surface_reg(struct radeon_device *rdev, int reg, void r100_clear_surface_reg(struct radeon_device *rdev, int reg); void r100_bandwidth_update(struct radeon_device *rdev); void r100_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib); -int r100_ring_test(struct radeon_device *rdev); +int r100_ring_test(struct radeon_device *rdev, struct radeon_ring *cp); void r100_hpd_init(struct radeon_device *rdev); void r100_hpd_fini(struct radeon_device *rdev); bool r100_hpd_sense(struct radeon_device *rdev, enum radeon_hpd_id hpd); @@ -101,12 +104,12 @@ void r100_pci_gart_disable(struct radeon_device *rdev); int r100_debugfs_mc_info_init(struct radeon_device *rdev); int r100_gui_wait_for_idle(struct radeon_device *rdev); void r100_gpu_lockup_update(struct r100_gpu_lockup *lockup, - struct radeon_cp *cp); + struct radeon_ring *cp); bool r100_gpu_cp_is_lockup(struct radeon_device *rdev, struct r100_gpu_lockup *lockup, - struct radeon_cp *cp); + struct radeon_ring *cp); void r100_ib_fini(struct radeon_device *rdev); -int r100_ib_init(struct radeon_device *rdev); +int r100_ib_test(struct radeon_device *rdev, struct radeon_ring *ring); void r100_irq_disable(struct radeon_device *rdev); void r100_mc_stop(struct radeon_device *rdev, struct r100_mc_save *save); void r100_mc_resume(struct radeon_device *rdev, struct r100_mc_save *save); @@ -136,6 +139,8 @@ extern void r100_pm_get_dynpm_state(struct radeon_device *rdev); extern void r100_pre_page_flip(struct radeon_device *rdev, int crtc); extern u32 r100_page_flip(struct radeon_device *rdev, int crtc, u64 crtc_base); extern void r100_post_page_flip(struct radeon_device *rdev, int crtc); +extern void r100_wait_for_vblank(struct radeon_device *rdev, int crtc); +extern int r100_mc_wait_for_idle(struct radeon_device *rdev); /* * r200,rv250,rs300,rv280 @@ -154,9 +159,9 @@ extern int r300_init(struct radeon_device *rdev); extern void r300_fini(struct radeon_device *rdev); extern int r300_suspend(struct radeon_device *rdev); extern int r300_resume(struct radeon_device *rdev); -extern bool r300_gpu_is_lockup(struct radeon_device *rdev); +extern bool r300_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp); extern int r300_asic_reset(struct radeon_device *rdev); -extern void r300_ring_start(struct radeon_device *rdev); +extern void r300_ring_start(struct radeon_device *rdev, struct radeon_ring *ring); extern void r300_fence_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence); extern int r300_cs_parse(struct radeon_cs_parser *p); @@ -173,6 +178,7 @@ extern int rv370_pcie_gart_init(struct radeon_device *rdev); extern void rv370_pcie_gart_fini(struct radeon_device *rdev); extern int rv370_pcie_gart_enable(struct radeon_device *rdev); extern void rv370_pcie_gart_disable(struct radeon_device *rdev); +extern int r300_mc_wait_for_idle(struct radeon_device *rdev); /* * r420,r423,rv410 @@ -203,6 +209,7 @@ int rs400_gart_enable(struct radeon_device *rdev); void rs400_gart_adjust_size(struct radeon_device *rdev); void rs400_gart_disable(struct radeon_device *rdev); void rs400_gart_fini(struct radeon_device *rdev); +extern int rs400_mc_wait_for_idle(struct radeon_device *rdev); /* * rs600. @@ -233,7 +240,8 @@ extern void rs600_pre_page_flip(struct radeon_device *rdev, int crtc); extern u32 rs600_page_flip(struct radeon_device *rdev, int crtc, u64 crtc_base); extern void rs600_post_page_flip(struct radeon_device *rdev, int crtc); void rs600_set_safe_registers(struct radeon_device *rdev); - +extern void avivo_wait_for_vblank(struct radeon_device *rdev, int crtc); +extern int rs600_mc_wait_for_idle(struct radeon_device *rdev); /* * rs690,rs740 @@ -248,6 +256,7 @@ void rs690_bandwidth_update(struct radeon_device *rdev); void rs690_line_buffer_adjust(struct radeon_device *rdev, struct drm_display_mode *mode1, struct drm_display_mode *mode2); +extern int rs690_mc_wait_for_idle(struct radeon_device *rdev); /* * rv515 @@ -261,7 +270,7 @@ int rv515_init(struct radeon_device *rdev); void rv515_fini(struct radeon_device *rdev); uint32_t rv515_mc_rreg(struct radeon_device *rdev, uint32_t reg); void rv515_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v); -void rv515_ring_start(struct radeon_device *rdev); +void rv515_ring_start(struct radeon_device *rdev, struct radeon_ring *ring); void rv515_bandwidth_update(struct radeon_device *rdev); int rv515_resume(struct radeon_device *rdev); int rv515_suspend(struct radeon_device *rdev); @@ -272,13 +281,14 @@ void rv515_mc_stop(struct radeon_device *rdev, struct rv515_mc_save *save); void rv515_mc_resume(struct radeon_device *rdev, struct rv515_mc_save *save); void rv515_clock_startup(struct radeon_device *rdev); void rv515_debugfs(struct radeon_device *rdev); - +int rv515_mc_wait_for_idle(struct radeon_device *rdev); /* * r520,rv530,rv560,rv570,r580 */ int r520_init(struct radeon_device *rdev); int r520_resume(struct radeon_device *rdev); +int r520_mc_wait_for_idle(struct radeon_device *rdev); /* * r600,rv610,rv630,rv620,rv635,rv670,rs780,rs880 @@ -290,22 +300,25 @@ int r600_resume(struct radeon_device *rdev); void r600_vga_set_state(struct radeon_device *rdev, bool state); int r600_wb_init(struct radeon_device *rdev); void r600_wb_fini(struct radeon_device *rdev); -void r600_cp_commit(struct radeon_device *rdev); void r600_pcie_gart_tlb_flush(struct radeon_device *rdev); uint32_t r600_pciep_rreg(struct radeon_device *rdev, uint32_t reg); void r600_pciep_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v); int r600_cs_parse(struct radeon_cs_parser *p); void r600_fence_ring_emit(struct radeon_device *rdev, struct radeon_fence *fence); -bool r600_gpu_is_lockup(struct radeon_device *rdev); +void r600_semaphore_ring_emit(struct radeon_device *rdev, + struct radeon_ring *cp, + struct radeon_semaphore *semaphore, + bool emit_wait); +bool r600_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp); int r600_asic_reset(struct radeon_device *rdev); int r600_set_surface_reg(struct radeon_device *rdev, int reg, uint32_t tiling_flags, uint32_t pitch, uint32_t offset, uint32_t obj_size); void r600_clear_surface_reg(struct radeon_device *rdev, int reg); -int r600_ib_test(struct radeon_device *rdev); +int r600_ib_test(struct radeon_device *rdev, struct radeon_ring *ring); void r600_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib); -int r600_ring_test(struct radeon_device *rdev); +int r600_ring_test(struct radeon_device *rdev, struct radeon_ring *cp); int r600_copy_blit(struct radeon_device *rdev, uint64_t src_offset, uint64_t dst_offset, unsigned num_gpu_pages, struct radeon_fence *fence); @@ -325,7 +338,7 @@ extern int r600_get_pcie_lanes(struct radeon_device *rdev); bool r600_card_posted(struct radeon_device *rdev); void r600_cp_stop(struct radeon_device *rdev); int r600_cp_start(struct radeon_device *rdev); -void r600_ring_init(struct radeon_device *rdev, unsigned ring_size); +void r600_ring_init(struct radeon_device *rdev, struct radeon_ring *cp, unsigned ring_size); int r600_cp_resume(struct radeon_device *rdev); void r600_cp_fini(struct radeon_device *rdev); int r600_count_pipe_bits(uint32_t val); @@ -366,6 +379,7 @@ void r600_blit_done_copy(struct radeon_device *rdev, struct radeon_fence *fence) void r600_kms_blit_copy(struct radeon_device *rdev, u64 src_gpu_addr, u64 dst_gpu_addr, unsigned num_gpu_pages); +int r600_mc_wait_for_idle(struct radeon_device *rdev); /* * rv770,rv730,rv710,rv740 @@ -394,7 +408,7 @@ int evergreen_init(struct radeon_device *rdev); void evergreen_fini(struct radeon_device *rdev); int evergreen_suspend(struct radeon_device *rdev); int evergreen_resume(struct radeon_device *rdev); -bool evergreen_gpu_is_lockup(struct radeon_device *rdev); +bool evergreen_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp); int evergreen_asic_reset(struct radeon_device *rdev); void evergreen_bandwidth_update(struct radeon_device *rdev); void evergreen_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib); @@ -414,18 +428,59 @@ extern void sumo_pm_init_profile(struct radeon_device *rdev); extern void evergreen_pre_page_flip(struct radeon_device *rdev, int crtc); extern u32 evergreen_page_flip(struct radeon_device *rdev, int crtc, u64 crtc_base); extern void evergreen_post_page_flip(struct radeon_device *rdev, int crtc); +extern void dce4_wait_for_vblank(struct radeon_device *rdev, int crtc); void evergreen_disable_interrupt_state(struct radeon_device *rdev); int evergreen_blit_init(struct radeon_device *rdev); +int evergreen_mc_wait_for_idle(struct radeon_device *rdev); /* * cayman */ +void cayman_fence_ring_emit(struct radeon_device *rdev, + struct radeon_fence *fence); void cayman_pcie_gart_tlb_flush(struct radeon_device *rdev); int cayman_init(struct radeon_device *rdev); void cayman_fini(struct radeon_device *rdev); int cayman_suspend(struct radeon_device *rdev); int cayman_resume(struct radeon_device *rdev); -bool cayman_gpu_is_lockup(struct radeon_device *rdev); +bool cayman_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp); int cayman_asic_reset(struct radeon_device *rdev); +void cayman_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib); +int cayman_vm_init(struct radeon_device *rdev); +void cayman_vm_fini(struct radeon_device *rdev); +int cayman_vm_bind(struct radeon_device *rdev, struct radeon_vm *vm, int id); +void cayman_vm_unbind(struct radeon_device *rdev, struct radeon_vm *vm); +void cayman_vm_tlb_flush(struct radeon_device *rdev, struct radeon_vm *vm); +uint32_t cayman_vm_page_flags(struct radeon_device *rdev, + struct radeon_vm *vm, + uint32_t flags); +void cayman_vm_set_page(struct radeon_device *rdev, struct radeon_vm *vm, + unsigned pfn, uint64_t addr, uint32_t flags); +int evergreen_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib); + +/* DCE6 - SI */ +void dce6_bandwidth_update(struct radeon_device *rdev); + +/* + * si + */ +void si_fence_ring_emit(struct radeon_device *rdev, + struct radeon_fence *fence); +void si_pcie_gart_tlb_flush(struct radeon_device *rdev); +int si_init(struct radeon_device *rdev); +void si_fini(struct radeon_device *rdev); +int si_suspend(struct radeon_device *rdev); +int si_resume(struct radeon_device *rdev); +bool si_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp); +int si_asic_reset(struct radeon_device *rdev); +void si_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib); +int si_irq_set(struct radeon_device *rdev); +int si_irq_process(struct radeon_device *rdev); +int si_vm_init(struct radeon_device *rdev); +void si_vm_fini(struct radeon_device *rdev); +int si_vm_bind(struct radeon_device *rdev, struct radeon_vm *vm, int id); +void si_vm_unbind(struct radeon_device *rdev, struct radeon_vm *vm); +void si_vm_tlb_flush(struct radeon_device *rdev, struct radeon_vm *vm); +int si_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib); #endif diff --git a/drivers/gpu/drm/radeon/radeon_atombios.c b/drivers/gpu/drm/radeon/radeon_atombios.c index 38585c5..5e30e12 100644 --- a/drivers/gpu/drm/radeon/radeon_atombios.c +++ b/drivers/gpu/drm/radeon/radeon_atombios.c @@ -56,6 +56,10 @@ extern void radeon_add_legacy_encoder(struct drm_device *dev, uint32_t encoder_enum, uint32_t supported_device); +/* local */ +static int radeon_atom_get_max_vddc(struct radeon_device *rdev, u8 voltage_type, + u16 voltage_id, u16 *voltage); + union atom_supported_devices { struct _ATOM_SUPPORTED_DEVICES_INFO info; struct _ATOM_SUPPORTED_DEVICES_INFO_2 info_2; @@ -253,7 +257,9 @@ static struct radeon_hpd radeon_atom_get_hpd_info_from_gpio(struct radeon_device memset(&hpd, 0, sizeof(struct radeon_hpd)); - if (ASIC_IS_DCE4(rdev)) + if (ASIC_IS_DCE6(rdev)) + reg = SI_DC_GPIO_HPD_A; + else if (ASIC_IS_DCE4(rdev)) reg = EVERGREEN_DC_GPIO_HPD_A; else reg = AVIVO_DC_GPIO_HPD_A; @@ -1890,6 +1896,8 @@ static const char *pp_lib_thermal_controller_names[] = { "emc2103", "Sumo", "Northern Islands", + "Southern Islands", + "lm96163", }; union power_info { @@ -1906,6 +1914,7 @@ union pplib_clock_info { struct _ATOM_PPLIB_RS780_CLOCK_INFO rs780; struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen; struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo; + struct _ATOM_PPLIB_SI_CLOCK_INFO si; }; union pplib_power_state { @@ -2163,6 +2172,11 @@ static void radeon_atombios_add_pplib_thermal_controller(struct radeon_device *r (controller->ucFanParameters & ATOM_PP_FANPARAMETERS_NOFAN) ? "without" : "with"); rdev->pm.int_thermal_type = THERMAL_TYPE_NI; + } else if (controller->ucType == ATOM_PP_THERMALCONTROLLER_SISLANDS) { + DRM_INFO("Internal thermal controller %s fan control\n", + (controller->ucFanParameters & + ATOM_PP_FANPARAMETERS_NOFAN) ? "without" : "with"); + rdev->pm.int_thermal_type = THERMAL_TYPE_SI; } else if ((controller->ucType == ATOM_PP_THERMALCONTROLLER_EXTERNAL_GPIO) || (controller->ucType == @@ -2283,6 +2297,7 @@ static bool radeon_atombios_parse_pplib_clock_info(struct radeon_device *rdev, union pplib_clock_info *clock_info) { u32 sclk, mclk; + u16 vddc; if (rdev->flags & RADEON_IS_IGP) { if (rdev->family >= CHIP_PALM) { @@ -2294,6 +2309,19 @@ static bool radeon_atombios_parse_pplib_clock_info(struct radeon_device *rdev, sclk |= clock_info->rs780.ucLowEngineClockHigh << 16; rdev->pm.power_state[state_index].clock_info[mode_index].sclk = sclk; } + } else if (ASIC_IS_DCE6(rdev)) { + sclk = le16_to_cpu(clock_info->si.usEngineClockLow); + sclk |= clock_info->si.ucEngineClockHigh << 16; + mclk = le16_to_cpu(clock_info->si.usMemoryClockLow); + mclk |= clock_info->si.ucMemoryClockHigh << 16; + rdev->pm.power_state[state_index].clock_info[mode_index].mclk = mclk; + rdev->pm.power_state[state_index].clock_info[mode_index].sclk = sclk; + rdev->pm.power_state[state_index].clock_info[mode_index].voltage.type = + VOLTAGE_SW; + rdev->pm.power_state[state_index].clock_info[mode_index].voltage.voltage = + le16_to_cpu(clock_info->si.usVDDC); + rdev->pm.power_state[state_index].clock_info[mode_index].voltage.vddci = + le16_to_cpu(clock_info->si.usVDDCI); } else if (ASIC_IS_DCE4(rdev)) { sclk = le16_to_cpu(clock_info->evergreen.usEngineClockLow); sclk |= clock_info->evergreen.ucEngineClockHigh << 16; @@ -2321,11 +2349,18 @@ static bool radeon_atombios_parse_pplib_clock_info(struct radeon_device *rdev, } /* patch up vddc if necessary */ - if (rdev->pm.power_state[state_index].clock_info[mode_index].voltage.voltage == 0xff01) { - u16 vddc; - - if (radeon_atom_get_max_vddc(rdev, &vddc) == 0) + switch (rdev->pm.power_state[state_index].clock_info[mode_index].voltage.voltage) { + case ATOM_VIRTUAL_VOLTAGE_ID0: + case ATOM_VIRTUAL_VOLTAGE_ID1: + case ATOM_VIRTUAL_VOLTAGE_ID2: + case ATOM_VIRTUAL_VOLTAGE_ID3: + if (radeon_atom_get_max_vddc(rdev, VOLTAGE_TYPE_VDDC, + rdev->pm.power_state[state_index].clock_info[mode_index].voltage.voltage, + &vddc) == 0) rdev->pm.power_state[state_index].clock_info[mode_index].voltage.voltage = vddc; + break; + default: + break; } if (rdev->flags & RADEON_IS_IGP) { @@ -2435,9 +2470,9 @@ static int radeon_atombios_parse_power_table_6(struct radeon_device *rdev) int i, j, non_clock_array_index, clock_array_index; int state_index = 0, mode_index = 0; union pplib_clock_info *clock_info; - struct StateArray *state_array; - struct ClockInfoArray *clock_info_array; - struct NonClockInfoArray *non_clock_info_array; + struct _StateArray *state_array; + struct _ClockInfoArray *clock_info_array; + struct _NonClockInfoArray *non_clock_info_array; bool valid; union power_info *power_info; int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo); @@ -2450,13 +2485,13 @@ static int radeon_atombios_parse_power_table_6(struct radeon_device *rdev) power_info = (union power_info *)(mode_info->atom_context->bios + data_offset); radeon_atombios_add_pplib_thermal_controller(rdev, &power_info->pplib.sThermalController); - state_array = (struct StateArray *) + state_array = (struct _StateArray *) (mode_info->atom_context->bios + data_offset + le16_to_cpu(power_info->pplib.usStateArrayOffset)); - clock_info_array = (struct ClockInfoArray *) + clock_info_array = (struct _ClockInfoArray *) (mode_info->atom_context->bios + data_offset + le16_to_cpu(power_info->pplib.usClockInfoArrayOffset)); - non_clock_info_array = (struct NonClockInfoArray *) + non_clock_info_array = (struct _NonClockInfoArray *) (mode_info->atom_context->bios + data_offset + le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset)); rdev->pm.power_state = kzalloc(sizeof(struct radeon_power_state) * @@ -2483,7 +2518,7 @@ static int radeon_atombios_parse_power_table_6(struct radeon_device *rdev) if (clock_array_index >= clock_info_array->ucNumEntries) continue; clock_info = (union pplib_clock_info *) - &clock_info_array->clockInfo[clock_array_index]; + &clock_info_array->clockInfo[clock_array_index * clock_info_array->ucEntrySize]; valid = radeon_atombios_parse_pplib_clock_info(rdev, state_index, mode_index, clock_info); @@ -2640,6 +2675,7 @@ union set_voltage { struct _SET_VOLTAGE_PS_ALLOCATION alloc; struct _SET_VOLTAGE_PARAMETERS v1; struct _SET_VOLTAGE_PARAMETERS_V2 v2; + struct _SET_VOLTAGE_PARAMETERS_V1_3 v3; }; void radeon_atom_set_voltage(struct radeon_device *rdev, u16 voltage_level, u8 voltage_type) @@ -2666,6 +2702,11 @@ void radeon_atom_set_voltage(struct radeon_device *rdev, u16 voltage_level, u8 v args.v2.ucVoltageMode = SET_ASIC_VOLTAGE_MODE_SET_VOLTAGE; args.v2.usVoltageLevel = cpu_to_le16(voltage_level); break; + case 3: + args.v3.ucVoltageType = voltage_type; + args.v3.ucVoltageMode = ATOM_SET_VOLTAGE; + args.v3.usVoltageLevel = cpu_to_le16(voltage_level); + break; default: DRM_ERROR("Unknown table version %d, %d\n", frev, crev); return; @@ -2674,8 +2715,8 @@ void radeon_atom_set_voltage(struct radeon_device *rdev, u16 voltage_level, u8 v atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); } -int radeon_atom_get_max_vddc(struct radeon_device *rdev, - u16 *voltage) +static int radeon_atom_get_max_vddc(struct radeon_device *rdev, u8 voltage_type, + u16 voltage_id, u16 *voltage) { union set_voltage args; int index = GetIndexIntoMasterTable(COMMAND, SetVoltage); @@ -2696,6 +2737,15 @@ int radeon_atom_get_max_vddc(struct radeon_device *rdev, *voltage = le16_to_cpu(args.v2.usVoltageLevel); break; + case 3: + args.v3.ucVoltageType = voltage_type; + args.v3.ucVoltageMode = ATOM_GET_VOLTAGE_LEVEL; + args.v3.usVoltageLevel = cpu_to_le16(voltage_id); + + atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); + + *voltage = le16_to_cpu(args.v3.usVoltageLevel); + break; default: DRM_ERROR("Unknown table version %d, %d\n", frev, crev); return -EINVAL; @@ -2947,6 +2997,20 @@ radeon_atombios_connected_scratch_regs(struct drm_connector *connector, bios_6_scratch &= ~ATOM_S6_ACC_REQ_DFP5; } } + if ((radeon_encoder->devices & ATOM_DEVICE_DFP6_SUPPORT) && + (radeon_connector->devices & ATOM_DEVICE_DFP6_SUPPORT)) { + if (connected) { + DRM_DEBUG_KMS("DFP6 connected\n"); + bios_0_scratch |= ATOM_S0_DFP6; + bios_3_scratch |= ATOM_S3_DFP6_ACTIVE; + bios_6_scratch |= ATOM_S6_ACC_REQ_DFP6; + } else { + DRM_DEBUG_KMS("DFP6 disconnected\n"); + bios_0_scratch &= ~ATOM_S0_DFP6; + bios_3_scratch &= ~ATOM_S3_DFP6_ACTIVE; + bios_6_scratch &= ~ATOM_S6_ACC_REQ_DFP6; + } + } if (rdev->family >= CHIP_R600) { WREG32(R600_BIOS_0_SCRATCH, bios_0_scratch); @@ -2967,6 +3031,9 @@ radeon_atombios_encoder_crtc_scratch_regs(struct drm_encoder *encoder, int crtc) struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); uint32_t bios_3_scratch; + if (ASIC_IS_DCE4(rdev)) + return; + if (rdev->family >= CHIP_R600) bios_3_scratch = RREG32(R600_BIOS_3_SCRATCH); else @@ -3019,6 +3086,9 @@ radeon_atombios_encoder_dpms_scratch_regs(struct drm_encoder *encoder, bool on) struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); uint32_t bios_2_scratch; + if (ASIC_IS_DCE4(rdev)) + return; + if (rdev->family >= CHIP_R600) bios_2_scratch = RREG32(R600_BIOS_2_SCRATCH); else diff --git a/drivers/gpu/drm/radeon/radeon_benchmark.c b/drivers/gpu/drm/radeon/radeon_benchmark.c index 17e1a9b..fef7b72 100644 --- a/drivers/gpu/drm/radeon/radeon_benchmark.c +++ b/drivers/gpu/drm/radeon/radeon_benchmark.c @@ -43,17 +43,19 @@ static int radeon_benchmark_do_move(struct radeon_device *rdev, unsigned size, start_jiffies = jiffies; for (i = 0; i < n; i++) { - r = radeon_fence_create(rdev, &fence); - if (r) - return r; - switch (flag) { case RADEON_BENCHMARK_COPY_DMA: + r = radeon_fence_create(rdev, &fence, radeon_copy_dma_ring_index(rdev)); + if (r) + return r; r = radeon_copy_dma(rdev, saddr, daddr, size / RADEON_GPU_PAGE_SIZE, fence); break; case RADEON_BENCHMARK_COPY_BLIT: + r = radeon_fence_create(rdev, &fence, radeon_copy_blit_ring_index(rdev)); + if (r) + return r; r = radeon_copy_blit(rdev, saddr, daddr, size / RADEON_GPU_PAGE_SIZE, fence); @@ -129,7 +131,7 @@ static void radeon_benchmark_move(struct radeon_device *rdev, unsigned size, /* r100 doesn't have dma engine so skip the test */ /* also, VRAM-to-VRAM test doesn't make much sense for DMA */ /* skip it as well if domains are the same */ - if ((rdev->asic->copy_dma) && (sdomain != ddomain)) { + if ((rdev->asic->copy.dma) && (sdomain != ddomain)) { time = radeon_benchmark_do_move(rdev, size, saddr, daddr, RADEON_BENCHMARK_COPY_DMA, n); if (time < 0) @@ -208,42 +210,42 @@ void radeon_benchmark(struct radeon_device *rdev, int test_number) break; case 3: /* GTT to VRAM, buffer size sweep, powers of 2 */ - for (i = 1; i <= 65536; i <<= 1) - radeon_benchmark_move(rdev, i*1024, + for (i = 1; i <= 16384; i <<= 1) + radeon_benchmark_move(rdev, i * RADEON_GPU_PAGE_SIZE, RADEON_GEM_DOMAIN_GTT, RADEON_GEM_DOMAIN_VRAM); break; case 4: /* VRAM to GTT, buffer size sweep, powers of 2 */ - for (i = 1; i <= 65536; i <<= 1) - radeon_benchmark_move(rdev, i*1024, + for (i = 1; i <= 16384; i <<= 1) + radeon_benchmark_move(rdev, i * RADEON_GPU_PAGE_SIZE, RADEON_GEM_DOMAIN_VRAM, RADEON_GEM_DOMAIN_GTT); break; case 5: /* VRAM to VRAM, buffer size sweep, powers of 2 */ - for (i = 1; i <= 65536; i <<= 1) - radeon_benchmark_move(rdev, i*1024, + for (i = 1; i <= 16384; i <<= 1) + radeon_benchmark_move(rdev, i * RADEON_GPU_PAGE_SIZE, RADEON_GEM_DOMAIN_VRAM, RADEON_GEM_DOMAIN_VRAM); break; case 6: /* GTT to VRAM, buffer size sweep, common modes */ - for (i = 1; i < RADEON_BENCHMARK_COMMON_MODES_N; i++) + for (i = 0; i < RADEON_BENCHMARK_COMMON_MODES_N; i++) radeon_benchmark_move(rdev, common_modes[i], RADEON_GEM_DOMAIN_GTT, RADEON_GEM_DOMAIN_VRAM); break; case 7: /* VRAM to GTT, buffer size sweep, common modes */ - for (i = 1; i < RADEON_BENCHMARK_COMMON_MODES_N; i++) + for (i = 0; i < RADEON_BENCHMARK_COMMON_MODES_N; i++) radeon_benchmark_move(rdev, common_modes[i], RADEON_GEM_DOMAIN_VRAM, RADEON_GEM_DOMAIN_GTT); break; case 8: /* VRAM to VRAM, buffer size sweep, common modes */ - for (i = 1; i < RADEON_BENCHMARK_COMMON_MODES_N; i++) + for (i = 0; i < RADEON_BENCHMARK_COMMON_MODES_N; i++) radeon_benchmark_move(rdev, common_modes[i], RADEON_GEM_DOMAIN_VRAM, RADEON_GEM_DOMAIN_VRAM); diff --git a/drivers/gpu/drm/radeon/radeon_blit_common.h b/drivers/gpu/drm/radeon/radeon_blit_common.h new file mode 100644 index 0000000..4ecbe72 --- /dev/null +++ b/drivers/gpu/drm/radeon/radeon_blit_common.h @@ -0,0 +1,44 @@ +/* + * Copyright 2009 Advanced Micro Devices, Inc. + * Copyright 2009 Red Hat Inc. + * Copyright 2012 Alcatel-Lucent, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + */ + +#ifndef __RADEON_BLIT_COMMON_H__ + +#define DI_PT_RECTLIST 0x11 +#define DI_INDEX_SIZE_16_BIT 0x0 +#define DI_SRC_SEL_AUTO_INDEX 0x2 + +#define FMT_8 0x1 +#define FMT_5_6_5 0x8 +#define FMT_8_8_8_8 0x1a +#define COLOR_8 0x1 +#define COLOR_5_6_5 0x8 +#define COLOR_8_8_8_8 0x1a + +#define RECT_UNIT_H 32 +#define RECT_UNIT_W (RADEON_GPU_PAGE_SIZE / 4 / RECT_UNIT_H) + +#define __RADEON_BLIT_COMMON_H__ +#endif diff --git a/drivers/gpu/drm/radeon/radeon_clocks.c b/drivers/gpu/drm/radeon/radeon_clocks.c index b6e18c8..9c6b29a 100644 --- a/drivers/gpu/drm/radeon/radeon_clocks.c +++ b/drivers/gpu/drm/radeon/radeon_clocks.c @@ -334,7 +334,7 @@ void radeon_get_clock_info(struct drm_device *dev) if (!rdev->clock.default_sclk) rdev->clock.default_sclk = radeon_get_engine_clock(rdev); - if ((!rdev->clock.default_mclk) && rdev->asic->get_memory_clock) + if ((!rdev->clock.default_mclk) && rdev->asic->pm.get_memory_clock) rdev->clock.default_mclk = radeon_get_memory_clock(rdev); rdev->pm.current_sclk = rdev->clock.default_sclk; @@ -633,7 +633,7 @@ void radeon_legacy_set_clock_gating(struct radeon_device *rdev, int enable) tmp &= ~(R300_SCLK_FORCE_VAP); tmp |= RADEON_SCLK_FORCE_CP; WREG32_PLL(RADEON_SCLK_CNTL, tmp); - udelay(15000); + mdelay(15); tmp = RREG32_PLL(R300_SCLK_CNTL2); tmp &= ~(R300_SCLK_FORCE_TCL | @@ -651,12 +651,12 @@ void radeon_legacy_set_clock_gating(struct radeon_device *rdev, int enable) tmp |= (RADEON_ENGIN_DYNCLK_MODE | (0x01 << RADEON_ACTIVE_HILO_LAT_SHIFT)); WREG32_PLL(RADEON_CLK_PWRMGT_CNTL, tmp); - udelay(15000); + mdelay(15); tmp = RREG32_PLL(RADEON_CLK_PIN_CNTL); tmp |= RADEON_SCLK_DYN_START_CNTL; WREG32_PLL(RADEON_CLK_PIN_CNTL, tmp); - udelay(15000); + mdelay(15); /* When DRI is enabled, setting DYN_STOP_LAT to zero can cause some R200 to lockup randomly, leave them as set by BIOS. @@ -696,7 +696,7 @@ void radeon_legacy_set_clock_gating(struct radeon_device *rdev, int enable) tmp |= RADEON_SCLK_MORE_FORCEON; } WREG32_PLL(RADEON_SCLK_MORE_CNTL, tmp); - udelay(15000); + mdelay(15); } /* RV200::A11 A12, RV250::A11 A12 */ @@ -709,7 +709,7 @@ void radeon_legacy_set_clock_gating(struct radeon_device *rdev, int enable) tmp |= RADEON_TCL_BYPASS_DISABLE; WREG32_PLL(RADEON_PLL_PWRMGT_CNTL, tmp); } - udelay(15000); + mdelay(15); /*enable dynamic mode for display clocks (PIXCLK and PIX2CLK) */ tmp = RREG32_PLL(RADEON_PIXCLKS_CNTL); @@ -722,14 +722,14 @@ void radeon_legacy_set_clock_gating(struct radeon_device *rdev, int enable) RADEON_PIXCLK_TMDS_ALWAYS_ONb); WREG32_PLL(RADEON_PIXCLKS_CNTL, tmp); - udelay(15000); + mdelay(15); tmp = RREG32_PLL(RADEON_VCLK_ECP_CNTL); tmp |= (RADEON_PIXCLK_ALWAYS_ONb | RADEON_PIXCLK_DAC_ALWAYS_ONb); WREG32_PLL(RADEON_VCLK_ECP_CNTL, tmp); - udelay(15000); + mdelay(15); } } else { /* Turn everything OFF (ForceON to everything) */ @@ -861,7 +861,7 @@ void radeon_legacy_set_clock_gating(struct radeon_device *rdev, int enable) } WREG32_PLL(RADEON_SCLK_CNTL, tmp); - udelay(16000); + mdelay(16); if ((rdev->family == CHIP_R300) || (rdev->family == CHIP_R350)) { @@ -870,7 +870,7 @@ void radeon_legacy_set_clock_gating(struct radeon_device *rdev, int enable) R300_SCLK_FORCE_GA | R300_SCLK_FORCE_CBA); WREG32_PLL(R300_SCLK_CNTL2, tmp); - udelay(16000); + mdelay(16); } if (rdev->flags & RADEON_IS_IGP) { @@ -878,7 +878,7 @@ void radeon_legacy_set_clock_gating(struct radeon_device *rdev, int enable) tmp &= ~(RADEON_FORCEON_MCLKA | RADEON_FORCEON_YCLKA); WREG32_PLL(RADEON_MCLK_CNTL, tmp); - udelay(16000); + mdelay(16); } if ((rdev->family == CHIP_RV200) || @@ -887,7 +887,7 @@ void radeon_legacy_set_clock_gating(struct radeon_device *rdev, int enable) tmp = RREG32_PLL(RADEON_SCLK_MORE_CNTL); tmp |= RADEON_SCLK_MORE_FORCEON; WREG32_PLL(RADEON_SCLK_MORE_CNTL, tmp); - udelay(16000); + mdelay(16); } tmp = RREG32_PLL(RADEON_PIXCLKS_CNTL); @@ -900,7 +900,7 @@ void radeon_legacy_set_clock_gating(struct radeon_device *rdev, int enable) RADEON_PIXCLK_TMDS_ALWAYS_ONb); WREG32_PLL(RADEON_PIXCLKS_CNTL, tmp); - udelay(16000); + mdelay(16); tmp = RREG32_PLL(RADEON_VCLK_ECP_CNTL); tmp &= ~(RADEON_PIXCLK_ALWAYS_ONb | diff --git a/drivers/gpu/drm/radeon/radeon_combios.c b/drivers/gpu/drm/radeon/radeon_combios.c index ec36dd9..a2470d9 100644 --- a/drivers/gpu/drm/radeon/radeon_combios.c +++ b/drivers/gpu/drm/radeon/radeon_combios.c @@ -1536,9 +1536,6 @@ bool radeon_get_legacy_connector_info_from_table(struct drm_device *dev) of_machine_is_compatible("PowerBook6,7")) { /* ibook */ rdev->mode_info.connector_table = CT_IBOOK; - } else if (of_machine_is_compatible("PowerMac3,5")) { - /* PowerMac G4 Silver radeon 7500 */ - rdev->mode_info.connector_table = CT_MAC_G4_SILVER; } else if (of_machine_is_compatible("PowerMac4,4")) { /* emac */ rdev->mode_info.connector_table = CT_EMAC; @@ -1564,11 +1561,6 @@ bool radeon_get_legacy_connector_info_from_table(struct drm_device *dev) (rdev->pdev->subsystem_device == 0x4150)) { /* Mac G5 tower 9600 */ rdev->mode_info.connector_table = CT_MAC_G5_9600; - } else if ((rdev->pdev->device == 0x4c66) && - (rdev->pdev->subsystem_vendor == 0x1002) && - (rdev->pdev->subsystem_device == 0x4c66)) { - /* SAM440ep RV250 embedded board */ - rdev->mode_info.connector_table = CT_SAM440EP; } else #endif /* CONFIG_PPC_PMAC */ #ifdef CONFIG_PPC64 @@ -2142,115 +2134,6 @@ bool radeon_get_legacy_connector_info_from_table(struct drm_device *dev) CONNECTOR_OBJECT_ID_SVIDEO, &hpd); break; - case CT_SAM440EP: - DRM_INFO("Connector Table: %d (SAM440ep embedded board)\n", - rdev->mode_info.connector_table); - /* LVDS */ - ddc_i2c = combios_setup_i2c_bus(rdev, DDC_NONE_DETECTED, 0, 0); - hpd.hpd = RADEON_HPD_NONE; - radeon_add_legacy_encoder(dev, - radeon_get_encoder_enum(dev, - ATOM_DEVICE_LCD1_SUPPORT, - 0), - ATOM_DEVICE_LCD1_SUPPORT); - radeon_add_legacy_connector(dev, 0, ATOM_DEVICE_LCD1_SUPPORT, - DRM_MODE_CONNECTOR_LVDS, &ddc_i2c, - CONNECTOR_OBJECT_ID_LVDS, - &hpd); - /* DVI-I - secondary dac, int tmds */ - ddc_i2c = combios_setup_i2c_bus(rdev, DDC_DVI, 0, 0); - hpd.hpd = RADEON_HPD_1; /* ??? */ - radeon_add_legacy_encoder(dev, - radeon_get_encoder_enum(dev, - ATOM_DEVICE_DFP1_SUPPORT, - 0), - ATOM_DEVICE_DFP1_SUPPORT); - radeon_add_legacy_encoder(dev, - radeon_get_encoder_enum(dev, - ATOM_DEVICE_CRT2_SUPPORT, - 2), - ATOM_DEVICE_CRT2_SUPPORT); - radeon_add_legacy_connector(dev, 1, - ATOM_DEVICE_DFP1_SUPPORT | - ATOM_DEVICE_CRT2_SUPPORT, - DRM_MODE_CONNECTOR_DVII, &ddc_i2c, - CONNECTOR_OBJECT_ID_SINGLE_LINK_DVI_I, - &hpd); - /* VGA - primary dac */ - ddc_i2c = combios_setup_i2c_bus(rdev, DDC_VGA, 0, 0); - hpd.hpd = RADEON_HPD_NONE; - radeon_add_legacy_encoder(dev, - radeon_get_encoder_enum(dev, - ATOM_DEVICE_CRT1_SUPPORT, - 1), - ATOM_DEVICE_CRT1_SUPPORT); - radeon_add_legacy_connector(dev, 2, - ATOM_DEVICE_CRT1_SUPPORT, - DRM_MODE_CONNECTOR_VGA, &ddc_i2c, - CONNECTOR_OBJECT_ID_VGA, - &hpd); - /* TV - TV DAC */ - ddc_i2c.valid = false; - hpd.hpd = RADEON_HPD_NONE; - radeon_add_legacy_encoder(dev, - radeon_get_encoder_enum(dev, - ATOM_DEVICE_TV1_SUPPORT, - 2), - ATOM_DEVICE_TV1_SUPPORT); - radeon_add_legacy_connector(dev, 3, ATOM_DEVICE_TV1_SUPPORT, - DRM_MODE_CONNECTOR_SVIDEO, - &ddc_i2c, - CONNECTOR_OBJECT_ID_SVIDEO, - &hpd); - break; - case CT_MAC_G4_SILVER: - DRM_INFO("Connector Table: %d (mac g4 silver)\n", - rdev->mode_info.connector_table); - /* DVI-I - tv dac, int tmds */ - ddc_i2c = combios_setup_i2c_bus(rdev, DDC_DVI, 0, 0); - hpd.hpd = RADEON_HPD_1; /* ??? */ - radeon_add_legacy_encoder(dev, - radeon_get_encoder_enum(dev, - ATOM_DEVICE_DFP1_SUPPORT, - 0), - ATOM_DEVICE_DFP1_SUPPORT); - radeon_add_legacy_encoder(dev, - radeon_get_encoder_enum(dev, - ATOM_DEVICE_CRT2_SUPPORT, - 2), - ATOM_DEVICE_CRT2_SUPPORT); - radeon_add_legacy_connector(dev, 0, - ATOM_DEVICE_DFP1_SUPPORT | - ATOM_DEVICE_CRT2_SUPPORT, - DRM_MODE_CONNECTOR_DVII, &ddc_i2c, - CONNECTOR_OBJECT_ID_SINGLE_LINK_DVI_I, - &hpd); - /* VGA - primary dac */ - ddc_i2c = combios_setup_i2c_bus(rdev, DDC_VGA, 0, 0); - hpd.hpd = RADEON_HPD_NONE; - radeon_add_legacy_encoder(dev, - radeon_get_encoder_enum(dev, - ATOM_DEVICE_CRT1_SUPPORT, - 1), - ATOM_DEVICE_CRT1_SUPPORT); - radeon_add_legacy_connector(dev, 1, ATOM_DEVICE_CRT1_SUPPORT, - DRM_MODE_CONNECTOR_VGA, &ddc_i2c, - CONNECTOR_OBJECT_ID_VGA, - &hpd); - /* TV - TV DAC */ - ddc_i2c.valid = false; - hpd.hpd = RADEON_HPD_NONE; - radeon_add_legacy_encoder(dev, - radeon_get_encoder_enum(dev, - ATOM_DEVICE_TV1_SUPPORT, - 2), - ATOM_DEVICE_TV1_SUPPORT); - radeon_add_legacy_connector(dev, 2, ATOM_DEVICE_TV1_SUPPORT, - DRM_MODE_CONNECTOR_SVIDEO, - &ddc_i2c, - CONNECTOR_OBJECT_ID_SVIDEO, - &hpd); - break; default: DRM_INFO("Connector table: %d (invalid)\n", rdev->mode_info.connector_table); @@ -2970,7 +2853,7 @@ bool radeon_combios_external_tmds_setup(struct drm_encoder *encoder) case 4: val = RBIOS16(index); index += 2; - udelay(val * 1000); + mdelay(val); break; case 6: slave_addr = id & 0xff; @@ -3169,7 +3052,7 @@ static void combios_parse_pll_table(struct drm_device *dev, uint16_t offset) udelay(150); break; case 2: - udelay(1000); + mdelay(1); break; case 3: while (tmp--) { @@ -3200,13 +3083,13 @@ static void combios_parse_pll_table(struct drm_device *dev, uint16_t offset) /*mclk_cntl |= 0x00001111;*//* ??? */ WREG32_PLL(RADEON_MCLK_CNTL, mclk_cntl); - udelay(10000); + mdelay(10); #endif WREG32_PLL (RADEON_CLK_PWRMGT_CNTL, tmp & ~RADEON_CG_NO1_DEBUG_0); - udelay(10000); + mdelay(10); } break; default: diff --git a/drivers/gpu/drm/radeon/radeon_connectors.c b/drivers/gpu/drm/radeon/radeon_connectors.c index 6fd53b6..ab63bcd 100644 --- a/drivers/gpu/drm/radeon/radeon_connectors.c +++ b/drivers/gpu/drm/radeon/radeon_connectors.c @@ -846,6 +846,27 @@ static int radeon_dvi_get_modes(struct drm_connector *connector) return ret; } +static bool radeon_check_hpd_status_unchanged(struct drm_connector *connector) +{ + struct drm_device *dev = connector->dev; + struct radeon_device *rdev = dev->dev_private; + struct radeon_connector *radeon_connector = to_radeon_connector(connector); + enum drm_connector_status status; + + /* We only trust HPD on R600 and newer ASICS. */ + if (rdev->family >= CHIP_R600 + && radeon_connector->hpd.hpd != RADEON_HPD_NONE) { + if (radeon_hpd_sense(rdev, radeon_connector->hpd.hpd)) + status = connector_status_connected; + else + status = connector_status_disconnected; + if (connector->status == status) + return true; + } + + return false; +} + /* * DVI is complicated * Do a DDC probe, if DDC probe passes, get the full EDID so @@ -870,6 +891,9 @@ radeon_dvi_detect(struct drm_connector *connector, bool force) enum drm_connector_status ret = connector_status_disconnected; bool dret = false; + if (!force && radeon_check_hpd_status_unchanged(connector)) + return connector->status; + if (radeon_connector->ddc_bus) dret = radeon_ddc_probe(radeon_connector, false); if (dret) { @@ -1080,7 +1104,7 @@ static int radeon_dvi_mode_valid(struct drm_connector *connector, (radeon_connector->connector_object_id == CONNECTOR_OBJECT_ID_HDMI_TYPE_B)) return MODE_OK; else if (radeon_connector->connector_object_id == CONNECTOR_OBJECT_ID_HDMI_TYPE_A) { - if (ASIC_IS_DCE3(rdev)) { + if (ASIC_IS_DCE6(rdev)) { /* HDMI 1.3+ supports max clock of 340 Mhz */ if (mode->clock > 340000) return MODE_CLOCK_HIGH; @@ -1140,13 +1164,23 @@ static int radeon_dp_get_modes(struct drm_connector *connector) (connector->connector_type == DRM_MODE_CONNECTOR_LVDS)) { struct drm_display_mode *mode; - if (!radeon_dig_connector->edp_on) - atombios_set_edp_panel_power(connector, - ATOM_TRANSMITTER_ACTION_POWER_ON); - ret = radeon_ddc_get_modes(radeon_connector); - if (!radeon_dig_connector->edp_on) - atombios_set_edp_panel_power(connector, - ATOM_TRANSMITTER_ACTION_POWER_OFF); + if (connector->connector_type == DRM_MODE_CONNECTOR_eDP) { + if (!radeon_dig_connector->edp_on) + atombios_set_edp_panel_power(connector, + ATOM_TRANSMITTER_ACTION_POWER_ON); + ret = radeon_ddc_get_modes(radeon_connector); + if (!radeon_dig_connector->edp_on) + atombios_set_edp_panel_power(connector, + ATOM_TRANSMITTER_ACTION_POWER_OFF); + } else { + /* need to setup ddc on the bridge */ + if (radeon_connector_encoder_get_dp_bridge_encoder_id(connector) != + ENCODER_OBJECT_ID_NONE) { + if (encoder) + radeon_atom_ext_encoder_setup_ddc(encoder); + } + ret = radeon_ddc_get_modes(radeon_connector); + } if (ret > 0) { if (encoder) { @@ -1157,7 +1191,6 @@ static int radeon_dp_get_modes(struct drm_connector *connector) return ret; } - encoder = radeon_best_single_encoder(connector); if (!encoder) return 0; @@ -1264,6 +1297,9 @@ radeon_dp_detect(struct drm_connector *connector, bool force) struct radeon_connector_atom_dig *radeon_dig_connector = radeon_connector->con_priv; struct drm_encoder *encoder = radeon_best_single_encoder(connector); + if (!force && radeon_check_hpd_status_unchanged(connector)) + return connector->status; + if (radeon_connector->edid) { kfree(radeon_connector->edid); radeon_connector->edid = NULL; diff --git a/drivers/gpu/drm/radeon/radeon_cp.c b/drivers/gpu/drm/radeon/radeon_cp.c index 72ae826..0ebb7d4 100644 --- a/drivers/gpu/drm/radeon/radeon_cp.c +++ b/drivers/gpu/drm/radeon/radeon_cp.c @@ -2115,6 +2115,8 @@ int radeon_driver_load(struct drm_device *dev, unsigned long flags) break; } + pci_set_master(dev->pdev); + if (drm_pci_device_is_agp(dev)) dev_priv->flags |= RADEON_IS_AGP; else if (pci_is_pcie(dev->pdev)) diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c index 1f32557..cf723c4 100644 --- a/drivers/gpu/drm/radeon/radeon_cs.c +++ b/drivers/gpu/drm/radeon/radeon_cs.c @@ -58,7 +58,7 @@ int radeon_cs_parser_relocs(struct radeon_cs_parser *p) duplicate = false; r = (struct drm_radeon_cs_reloc *)&chunk->kdata[i*4]; - for (j = 0; j < p->nrelocs; j++) { + for (j = 0; j < i; j++) { if (r->handle == p->relocs[j].handle) { p->relocs_ptr[i] = &p->relocs[j]; duplicate = true; @@ -84,16 +84,88 @@ int radeon_cs_parser_relocs(struct radeon_cs_parser *p) p->relocs[i].flags = r->flags; radeon_bo_list_add_object(&p->relocs[i].lobj, &p->validated); - } + + } else + p->relocs[i].handle = 0; } return radeon_bo_list_validate(&p->validated); } +static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority) +{ + p->priority = priority; + + switch (ring) { + default: + DRM_ERROR("unknown ring id: %d\n", ring); + return -EINVAL; + case RADEON_CS_RING_GFX: + p->ring = RADEON_RING_TYPE_GFX_INDEX; + break; + case RADEON_CS_RING_COMPUTE: + if (p->rdev->family >= CHIP_TAHITI) { + if (p->priority > 0) + p->ring = CAYMAN_RING_TYPE_CP1_INDEX; + else + p->ring = CAYMAN_RING_TYPE_CP2_INDEX; + } else + p->ring = RADEON_RING_TYPE_GFX_INDEX; + break; + } + return 0; +} + +static int radeon_cs_sync_rings(struct radeon_cs_parser *p) +{ + bool sync_to_ring[RADEON_NUM_RINGS] = { }; + int i, r; + + for (i = 0; i < p->nrelocs; i++) { + if (!p->relocs[i].robj || !p->relocs[i].robj->tbo.sync_obj) + continue; + + if (!(p->relocs[i].flags & RADEON_RELOC_DONT_SYNC)) { + struct radeon_fence *fence = p->relocs[i].robj->tbo.sync_obj; + if (!radeon_fence_signaled(fence)) { + sync_to_ring[fence->ring] = true; + } + } + } + + for (i = 0; i < RADEON_NUM_RINGS; ++i) { + /* no need to sync to our own or unused rings */ + if (i == p->ring || !sync_to_ring[i] || !p->rdev->ring[i].ready) + continue; + + if (!p->ib->fence->semaphore) { + r = radeon_semaphore_create(p->rdev, &p->ib->fence->semaphore); + if (r) + return r; + } + + r = radeon_ring_lock(p->rdev, &p->rdev->ring[i], 3); + if (r) + return r; + radeon_semaphore_emit_signal(p->rdev, i, p->ib->fence->semaphore); + radeon_ring_unlock_commit(p->rdev, &p->rdev->ring[i]); + + r = radeon_ring_lock(p->rdev, &p->rdev->ring[p->ring], 3); + if (r) + return r; + radeon_semaphore_emit_wait(p->rdev, p->ring, p->ib->fence->semaphore); + radeon_ring_unlock_commit(p->rdev, &p->rdev->ring[p->ring]); + } + return 0; +} + +/* XXX: note that this is called from the legacy UMS CS ioctl as well */ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data) { struct drm_radeon_cs *cs = data; uint64_t *chunk_array_ptr; - unsigned size, i, flags = 0; + unsigned size, i; + u32 ring = RADEON_CS_RING_GFX; + s32 priority = 0; if (!cs->num_chunks) { return 0; @@ -103,6 +175,8 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data) p->idx = 0; p->chunk_ib_idx = -1; p->chunk_relocs_idx = -1; + p->chunk_flags_idx = -1; + p->chunk_const_ib_idx = -1; p->chunks_array = kcalloc(cs->num_chunks, sizeof(uint64_t), GFP_KERNEL); if (p->chunks_array == NULL) { return -ENOMEM; @@ -112,6 +186,7 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data) sizeof(uint64_t)*cs->num_chunks)) { return -EFAULT; } + p->cs_flags = 0; p->nchunks = cs->num_chunks; p->chunks = kcalloc(p->nchunks, sizeof(struct radeon_cs_chunk), GFP_KERNEL); if (p->chunks == NULL) { @@ -140,16 +215,25 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data) if (p->chunks[i].length_dw == 0) return -EINVAL; } - if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS && - !p->chunks[i].length_dw) { - return -EINVAL; + if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_CONST_IB) { + p->chunk_const_ib_idx = i; + /* zero length CONST IB isn't useful */ + if (p->chunks[i].length_dw == 0) + return -EINVAL; + } + if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS) { + p->chunk_flags_idx = i; + /* zero length flags aren't useful */ + if (p->chunks[i].length_dw == 0) + return -EINVAL; } p->chunks[i].length_dw = user_chunk.length_dw; p->chunks[i].user_ptr = (void __user *)(unsigned long)user_chunk.chunk_data; cdata = (uint32_t *)(unsigned long)user_chunk.chunk_data; - if (p->chunks[i].chunk_id != RADEON_CHUNK_ID_IB) { + if ((p->chunks[i].chunk_id == RADEON_CHUNK_ID_RELOCS) || + (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS)) { size = p->chunks[i].length_dw * sizeof(uint32_t); p->chunks[i].kdata = kmalloc(size, GFP_KERNEL); if (p->chunks[i].kdata == NULL) { @@ -160,31 +244,55 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data) return -EFAULT; } if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS) { - flags = p->chunks[i].kdata[0]; - } - } else { - p->chunks[i].kpage[0] = kmalloc(PAGE_SIZE, GFP_KERNEL); - p->chunks[i].kpage[1] = kmalloc(PAGE_SIZE, GFP_KERNEL); - if (p->chunks[i].kpage[0] == NULL || p->chunks[i].kpage[1] == NULL) { - kfree(p->chunks[i].kpage[0]); - kfree(p->chunks[i].kpage[1]); - p->chunks[i].kpage[0] = NULL; - p->chunks[i].kpage[1] = NULL; - return -ENOMEM; + p->cs_flags = p->chunks[i].kdata[0]; + if (p->chunks[i].length_dw > 1) + ring = p->chunks[i].kdata[1]; + if (p->chunks[i].length_dw > 2) + priority = (s32)p->chunks[i].kdata[2]; } - p->chunks[i].kpage_idx[0] = -1; - p->chunks[i].kpage_idx[1] = -1; - p->chunks[i].last_copied_page = -1; - p->chunks[i].last_page_index = ((p->chunks[i].length_dw * 4) - 1) / PAGE_SIZE; } } - if (p->chunks[p->chunk_ib_idx].length_dw > (16 * 1024)) { - DRM_ERROR("cs IB too big: %d\n", - p->chunks[p->chunk_ib_idx].length_dw); - return -EINVAL; + + /* these are KMS only */ + if (p->rdev) { + if ((p->cs_flags & RADEON_CS_USE_VM) && + !p->rdev->vm_manager.enabled) { + DRM_ERROR("VM not active on asic!\n"); + return -EINVAL; + } + + /* we only support VM on SI+ */ + if ((p->rdev->family >= CHIP_TAHITI) && + ((p->cs_flags & RADEON_CS_USE_VM) == 0)) { + DRM_ERROR("VM required on SI+!\n"); + return -EINVAL; + } + + if (radeon_cs_get_ring(p, ring, priority)) + return -EINVAL; + } + + /* deal with non-vm */ + if ((p->chunk_ib_idx != -1) && + ((p->cs_flags & RADEON_CS_USE_VM) == 0) && + (p->chunks[p->chunk_ib_idx].chunk_id == RADEON_CHUNK_ID_IB)) { + if (p->chunks[p->chunk_ib_idx].length_dw > (16 * 1024)) { + DRM_ERROR("cs IB too big: %d\n", + p->chunks[p->chunk_ib_idx].length_dw); + return -EINVAL; + } + p->chunks[p->chunk_ib_idx].kpage[0] = kmalloc(PAGE_SIZE, GFP_KERNEL); + p->chunks[p->chunk_ib_idx].kpage[1] = kmalloc(PAGE_SIZE, GFP_KERNEL); + if (p->chunks[p->chunk_ib_idx].kpage[0] == NULL || + p->chunks[p->chunk_ib_idx].kpage[1] == NULL) + return -ENOMEM; + p->chunks[p->chunk_ib_idx].kpage_idx[0] = -1; + p->chunks[p->chunk_ib_idx].kpage_idx[1] = -1; + p->chunks[p->chunk_ib_idx].last_copied_page = -1; + p->chunks[p->chunk_ib_idx].last_page_index = + ((p->chunks[p->chunk_ib_idx].length_dw * 4) - 1) / PAGE_SIZE; } - p->keep_tiling_flags = (flags & RADEON_CS_KEEP_TILING_FLAGS) != 0; return 0; } @@ -226,14 +334,186 @@ static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error) radeon_ib_free(parser->rdev, &parser->ib); } +static int radeon_cs_ib_chunk(struct radeon_device *rdev, + struct radeon_cs_parser *parser) +{ + struct radeon_cs_chunk *ib_chunk; + int r; + + if (parser->chunk_ib_idx == -1) + return 0; + + if (parser->cs_flags & RADEON_CS_USE_VM) + return 0; + + ib_chunk = &parser->chunks[parser->chunk_ib_idx]; + /* Copy the packet into the IB, the parser will read from the + * input memory (cached) and write to the IB (which can be + * uncached). + */ + r = radeon_ib_get(rdev, parser->ring, &parser->ib, + ib_chunk->length_dw * 4); + if (r) { + DRM_ERROR("Failed to get ib !\n"); + return r; + } + parser->ib->length_dw = ib_chunk->length_dw; + r = radeon_cs_parse(rdev, parser->ring, parser); + if (r || parser->parser_error) { + DRM_ERROR("Invalid command stream !\n"); + return r; + } + r = radeon_cs_finish_pages(parser); + if (r) { + DRM_ERROR("Invalid command stream !\n"); + return r; + } + r = radeon_cs_sync_rings(parser); + if (r) { + DRM_ERROR("Failed to synchronize rings !\n"); + } + parser->ib->vm_id = 0; + r = radeon_ib_schedule(rdev, parser->ib); + if (r) { + DRM_ERROR("Failed to schedule IB !\n"); + } + return r; +} + +static int radeon_bo_vm_update_pte(struct radeon_cs_parser *parser, + struct radeon_vm *vm) +{ + struct radeon_bo_list *lobj; + struct radeon_bo *bo; + int r; + + list_for_each_entry(lobj, &parser->validated, tv.head) { + bo = lobj->bo; + r = radeon_vm_bo_update_pte(parser->rdev, vm, bo, &bo->tbo.mem); + if (r) { + return r; + } + } + return 0; +} + +static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev, + struct radeon_cs_parser *parser) +{ + struct radeon_cs_chunk *ib_chunk; + struct radeon_fpriv *fpriv = parser->filp->driver_priv; + struct radeon_vm *vm = &fpriv->vm; + int r; + + if (parser->chunk_ib_idx == -1) + return 0; + + if ((parser->cs_flags & RADEON_CS_USE_VM) == 0) + return 0; + + if ((rdev->family >= CHIP_TAHITI) && + (parser->chunk_const_ib_idx != -1)) { + ib_chunk = &parser->chunks[parser->chunk_const_ib_idx]; + if (ib_chunk->length_dw > RADEON_IB_VM_MAX_SIZE) { + DRM_ERROR("cs IB CONST too big: %d\n", ib_chunk->length_dw); + return -EINVAL; + } + r = radeon_ib_get(rdev, parser->ring, &parser->const_ib, + ib_chunk->length_dw * 4); + if (r) { + DRM_ERROR("Failed to get const ib !\n"); + return r; + } + parser->const_ib->is_const_ib = true; + parser->const_ib->length_dw = ib_chunk->length_dw; + /* Copy the packet into the IB */ + if (DRM_COPY_FROM_USER(parser->const_ib->ptr, ib_chunk->user_ptr, + ib_chunk->length_dw * 4)) { + return -EFAULT; + } + r = radeon_ring_ib_parse(rdev, parser->ring, parser->const_ib); + if (r) { + return r; + } + } + + ib_chunk = &parser->chunks[parser->chunk_ib_idx]; + if (ib_chunk->length_dw > RADEON_IB_VM_MAX_SIZE) { + DRM_ERROR("cs IB too big: %d\n", ib_chunk->length_dw); + return -EINVAL; + } + r = radeon_ib_get(rdev, parser->ring, &parser->ib, + ib_chunk->length_dw * 4); + if (r) { + DRM_ERROR("Failed to get ib !\n"); + return r; + } + parser->ib->length_dw = ib_chunk->length_dw; + /* Copy the packet into the IB */ + if (DRM_COPY_FROM_USER(parser->ib->ptr, ib_chunk->user_ptr, + ib_chunk->length_dw * 4)) { + return -EFAULT; + } + r = radeon_ring_ib_parse(rdev, parser->ring, parser->ib); + if (r) { + return r; + } + + mutex_lock(&vm->mutex); + r = radeon_vm_bind(rdev, vm); + if (r) { + goto out; + } + r = radeon_bo_vm_update_pte(parser, vm); + if (r) { + goto out; + } + r = radeon_cs_sync_rings(parser); + if (r) { + DRM_ERROR("Failed to synchronize rings !\n"); + } + + if ((rdev->family >= CHIP_TAHITI) && + (parser->chunk_const_ib_idx != -1)) { + parser->const_ib->vm_id = vm->id; + /* ib pool is bind at 0 in virtual address space to gpu_addr is the + * offset inside the pool bo + */ + parser->const_ib->gpu_addr = parser->const_ib->sa_bo.offset; + r = radeon_ib_schedule(rdev, parser->const_ib); + if (r) + goto out; + } + + parser->ib->vm_id = vm->id; + /* ib pool is bind at 0 in virtual address space to gpu_addr is the + * offset inside the pool bo + */ + parser->ib->gpu_addr = parser->ib->sa_bo.offset; + parser->ib->is_const_ib = false; + r = radeon_ib_schedule(rdev, parser->ib); +out: + if (!r) { + if (vm->fence) { + radeon_fence_unref(&vm->fence); + } + vm->fence = radeon_fence_ref(parser->ib->fence); + } + mutex_unlock(&fpriv->vm.mutex); + return r; +} + int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) { struct radeon_device *rdev = dev->dev_private; struct radeon_cs_parser parser; - struct radeon_cs_chunk *ib_chunk; int r; radeon_mutex_lock(&rdev->cs_mutex); + if (!rdev->accel_working) { + radeon_mutex_unlock(&rdev->cs_mutex); + return -EBUSY; + } /* initialize parser */ memset(&parser, 0, sizeof(struct radeon_cs_parser)); parser.filp = filp; @@ -247,13 +527,6 @@ int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) radeon_mutex_unlock(&rdev->cs_mutex); return r; } - r = radeon_ib_get(rdev, &parser.ib); - if (r) { - DRM_ERROR("Failed to get ib !\n"); - radeon_cs_parser_fini(&parser, r); - radeon_mutex_unlock(&rdev->cs_mutex); - return r; - } r = radeon_cs_parser_relocs(&parser); if (r) { if (r != -ERESTARTSYS) @@ -262,29 +535,15 @@ int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) radeon_mutex_unlock(&rdev->cs_mutex); return r; } - /* Copy the packet into the IB, the parser will read from the - * input memory (cached) and write to the IB (which can be - * uncached). */ - ib_chunk = &parser.chunks[parser.chunk_ib_idx]; - parser.ib->length_dw = ib_chunk->length_dw; - r = radeon_cs_parse(&parser); - if (r || parser.parser_error) { - DRM_ERROR("Invalid command stream !\n"); - radeon_cs_parser_fini(&parser, r); - radeon_mutex_unlock(&rdev->cs_mutex); - return r; - } - r = radeon_cs_finish_pages(&parser); + r = radeon_cs_ib_chunk(rdev, &parser); if (r) { - DRM_ERROR("Invalid command stream !\n"); - radeon_cs_parser_fini(&parser, r); - radeon_mutex_unlock(&rdev->cs_mutex); - return r; + goto out; } - r = radeon_ib_schedule(rdev, parser.ib); + r = radeon_cs_ib_vm_chunk(rdev, &parser); if (r) { - DRM_ERROR("Failed to schedule IB !\n"); + goto out; } +out: radeon_cs_parser_fini(&parser, r); radeon_mutex_unlock(&rdev->cs_mutex); return r; diff --git a/drivers/gpu/drm/radeon/radeon_cursor.c b/drivers/gpu/drm/radeon/radeon_cursor.c index 2132109..8fb6f41 100644 --- a/drivers/gpu/drm/radeon/radeon_cursor.c +++ b/drivers/gpu/drm/radeon/radeon_cursor.c @@ -197,7 +197,12 @@ int radeon_crtc_cursor_set(struct drm_crtc *crtc, unpin: if (radeon_crtc->cursor_bo) { - radeon_gem_object_unpin(radeon_crtc->cursor_bo); + robj = gem_to_radeon_bo(radeon_crtc->cursor_bo); + ret = radeon_bo_reserve(robj, false); + if (likely(ret == 0)) { + radeon_bo_unpin(robj); + radeon_bo_unreserve(robj); + } drm_gem_object_unreference_unlocked(radeon_crtc->cursor_bo); } @@ -233,7 +238,8 @@ int radeon_crtc_cursor_move(struct drm_crtc *crtc, y = 0; } - if (ASIC_IS_AVIVO(rdev)) { + /* fixed on DCE6 and newer */ + if (ASIC_IS_AVIVO(rdev) && !ASIC_IS_DCE6(rdev)) { int i = 0; struct drm_crtc *crtc_p; diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c index bd959c1..68c89db 100644 --- a/drivers/gpu/drm/radeon/radeon_device.c +++ b/drivers/gpu/drm/radeon/radeon_device.c @@ -89,6 +89,10 @@ static const char radeon_family_name[][16] = { "TURKS", "CAICOS", "CAYMAN", + "ARUBA", + "TAHITI", + "PITCAIRN", + "VERDE", "LAST", }; @@ -237,8 +241,8 @@ int radeon_wb_init(struct radeon_device *rdev) rdev->wb.use_event = true; } } - /* always use writeback/events on NI */ - if (ASIC_IS_DCE5(rdev)) { + /* always use writeback/events on NI, APUs */ + if (rdev->family >= CHIP_PALM) { rdev->wb.enabled = true; rdev->wb.use_event = true; } @@ -721,18 +725,25 @@ int radeon_device_init(struct radeon_device *rdev, /* mutex initialization are all done here so we * can recall function without having locking issues */ radeon_mutex_init(&rdev->cs_mutex); - mutex_init(&rdev->ib_pool.mutex); - mutex_init(&rdev->cp.mutex); + radeon_mutex_init(&rdev->ib_pool.mutex); + for (i = 0; i < RADEON_NUM_RINGS; ++i) + mutex_init(&rdev->ring[i].mutex); mutex_init(&rdev->dc_hw_i2c_mutex); if (rdev->family >= CHIP_R600) spin_lock_init(&rdev->ih.lock); mutex_init(&rdev->gem.mutex); mutex_init(&rdev->pm.mutex); mutex_init(&rdev->vram_mutex); - rwlock_init(&rdev->fence_drv.lock); + rwlock_init(&rdev->fence_lock); + rwlock_init(&rdev->semaphore_drv.lock); INIT_LIST_HEAD(&rdev->gem.objects); init_waitqueue_head(&rdev->irq.vblank_queue); init_waitqueue_head(&rdev->irq.idle_queue); + INIT_LIST_HEAD(&rdev->semaphore_drv.bo); + /* initialize vm here */ + rdev->vm_manager.use_bitmap = 1; + rdev->vm_manager.max_pfn = 1 << 20; + INIT_LIST_HEAD(&rdev->vm_manager.lru_vm); /* Set asic functions */ r = radeon_asic_init(rdev); @@ -769,8 +780,14 @@ int radeon_device_init(struct radeon_device *rdev, r = pci_set_dma_mask(rdev->pdev, DMA_BIT_MASK(dma_bits)); if (r) { rdev->need_dma32 = true; + dma_bits = 32; printk(KERN_WARNING "radeon: No suitable DMA available.\n"); } + r = pci_set_consistent_dma_mask(rdev->pdev, DMA_BIT_MASK(dma_bits)); + if (r) { + pci_set_consistent_dma_mask(rdev->pdev, DMA_BIT_MASK(32)); + printk(KERN_WARNING "radeon: No coherent DMA available.\n"); + } /* Registers mapping */ /* TODO: block userspace mapping of io register */ @@ -818,15 +835,20 @@ int radeon_device_init(struct radeon_device *rdev, if (r) return r; } - if (radeon_testing) { + if ((radeon_testing & 1)) { radeon_test_moves(rdev); } + if ((radeon_testing & 2)) { + radeon_test_syncing(rdev); + } if (radeon_benchmarking) { radeon_benchmark(rdev, radeon_benchmarking); } return 0; } +static void radeon_debugfs_remove_files(struct radeon_device *rdev); + void radeon_device_fini(struct radeon_device *rdev) { DRM_INFO("radeon: finishing device.\n"); @@ -841,6 +863,7 @@ void radeon_device_fini(struct radeon_device *rdev) rdev->rio_mem = NULL; iounmap(rdev->rmmio); rdev->rmmio = NULL; + radeon_debugfs_remove_files(rdev); } @@ -852,7 +875,7 @@ int radeon_suspend_kms(struct drm_device *dev, pm_message_t state) struct radeon_device *rdev; struct drm_crtc *crtc; struct drm_connector *connector; - int r; + int i, r; if (dev == NULL || dev->dev_private == NULL) { return -ENODEV; @@ -893,7 +916,8 @@ int radeon_suspend_kms(struct drm_device *dev, pm_message_t state) /* evict vram memory */ radeon_bo_evict_vram(rdev); /* wait for gpu to finish processing current batch */ - radeon_fence_wait_last(rdev); + for (i = 0; i < RADEON_NUM_RINGS; i++) + radeon_fence_wait_last(rdev, i); radeon_save_bios_scratch_regs(rdev); @@ -942,9 +966,11 @@ int radeon_resume_kms(struct drm_device *dev) radeon_fbdev_set_suspend(rdev, 0); console_unlock(); - /* init dig PHYs */ - if (rdev->is_atom_bios) + /* init dig PHYs, disp eng pll */ + if (rdev->is_atom_bios) { radeon_atom_encoder_init(rdev); + radeon_atom_disp_eng_pll_init(rdev); + } /* reset hpd state */ radeon_hpd_init(rdev); /* blat the mode back in */ @@ -994,36 +1020,29 @@ int radeon_gpu_reset(struct radeon_device *rdev) /* * Debugfs */ -struct radeon_debugfs { - struct drm_info_list *files; - unsigned num_files; -}; -static struct radeon_debugfs _radeon_debugfs[RADEON_DEBUGFS_MAX_COMPONENTS]; -static unsigned _radeon_debugfs_count = 0; - int radeon_debugfs_add_files(struct radeon_device *rdev, struct drm_info_list *files, unsigned nfiles) { unsigned i; - for (i = 0; i < _radeon_debugfs_count; i++) { - if (_radeon_debugfs[i].files == files) { + for (i = 0; i < rdev->debugfs_count; i++) { + if (rdev->debugfs[i].files == files) { /* Already registered */ return 0; } } - i = _radeon_debugfs_count + 1; + i = rdev->debugfs_count + 1; if (i > RADEON_DEBUGFS_MAX_COMPONENTS) { DRM_ERROR("Reached maximum number of debugfs components.\n"); DRM_ERROR("Report so we increase " "RADEON_DEBUGFS_MAX_COMPONENTS.\n"); return -EINVAL; } - _radeon_debugfs[_radeon_debugfs_count].files = files; - _radeon_debugfs[_radeon_debugfs_count].num_files = nfiles; - _radeon_debugfs_count = i; + rdev->debugfs[rdev->debugfs_count].files = files; + rdev->debugfs[rdev->debugfs_count].num_files = nfiles; + rdev->debugfs_count = i; #if defined(CONFIG_DEBUG_FS) drm_debugfs_create_files(files, nfiles, rdev->ddev->control->debugfs_root, @@ -1035,6 +1054,22 @@ int radeon_debugfs_add_files(struct radeon_device *rdev, return 0; } +static void radeon_debugfs_remove_files(struct radeon_device *rdev) +{ +#if defined(CONFIG_DEBUG_FS) + unsigned i; + + for (i = 0; i < rdev->debugfs_count; i++) { + drm_debugfs_remove_files(rdev->debugfs[i].files, + rdev->debugfs[i].num_files, + rdev->ddev->control); + drm_debugfs_remove_files(rdev->debugfs[i].files, + rdev->debugfs[i].num_files, + rdev->ddev->primary); + } +#endif +} + #if defined(CONFIG_DEBUG_FS) int radeon_debugfs_init(struct drm_minor *minor) { @@ -1043,11 +1078,5 @@ int radeon_debugfs_init(struct drm_minor *minor) void radeon_debugfs_cleanup(struct drm_minor *minor) { - unsigned i; - - for (i = 0; i < _radeon_debugfs_count; i++) { - drm_debugfs_remove_files(_radeon_debugfs[i].files, - _radeon_debugfs[i].num_files, minor); - } } #endif diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c index 63e7143..00d9cac 100644 --- a/drivers/gpu/drm/radeon/radeon_display.c +++ b/drivers/gpu/drm/radeon/radeon_display.c @@ -303,8 +303,17 @@ void radeon_crtc_handle_flip(struct radeon_device *rdev, int crtc_id) if (update_pending && (DRM_SCANOUTPOS_VALID & radeon_get_crtc_scanoutpos(rdev->ddev, crtc_id, &vpos, &hpos)) && - (vpos >=0) && - (vpos < (99 * rdev->mode_info.crtcs[crtc_id]->base.hwmode.crtc_vdisplay)/100)) { + ((vpos >= (99 * rdev->mode_info.crtcs[crtc_id]->base.hwmode.crtc_vdisplay)/100) || + (vpos < 0 && !ASIC_IS_AVIVO(rdev)))) { + /* crtc didn't flip in this target vblank interval, + * but flip is pending in crtc. Based on the current + * scanout position we know that the current frame is + * (nearly) complete and the flip will (likely) + * complete before the start of the next frame. + */ + update_pending = 0; + } + if (update_pending) { /* crtc didn't flip in this target vblank interval, * but flip is pending in crtc. It will complete it * in next vblank interval, so complete the flip at @@ -393,7 +402,9 @@ static int radeon_crtc_page_flip(struct drm_crtc *crtc, DRM_ERROR("failed to reserve new rbo buffer before flip\n"); goto pflip_cleanup; } - r = radeon_bo_pin(rbo, RADEON_GEM_DOMAIN_VRAM, &base); + /* Only 27 bit offset for legacy CRTC */ + r = radeon_bo_pin_restricted(rbo, RADEON_GEM_DOMAIN_VRAM, + ASIC_IS_AVIVO(rdev) ? 0 : 1 << 27, &base); if (unlikely(r != 0)) { radeon_bo_unreserve(rbo); r = -EINVAL; @@ -406,7 +417,7 @@ static int radeon_crtc_page_flip(struct drm_crtc *crtc, if (!ASIC_IS_AVIVO(rdev)) { /* crtc offset is from display base addr not FB location */ base -= radeon_crtc->legacy_display_base_addr; - pitch_pixels = fb->pitch / (fb->bits_per_pixel / 8); + pitch_pixels = fb->pitches[0] / (fb->bits_per_pixel / 8); if (tiling_flags & RADEON_TILING_MACRO) { if (ASIC_IS_R300(rdev)) { @@ -522,7 +533,7 @@ static void radeon_crtc_init(struct drm_device *dev, int index) radeon_legacy_init_crtc(dev, radeon_crtc); } -static const char *encoder_names[36] = { +static const char *encoder_names[37] = { "NONE", "INTERNAL_LVDS", "INTERNAL_TMDS1", @@ -559,6 +570,7 @@ static const char *encoder_names[36] = { "INTERNAL_UNIPHY2", "NUTMEG", "TRAVIS", + "INTERNAL_VCE" }; static const char *connector_names[15] = { @@ -1083,29 +1095,36 @@ static const struct drm_framebuffer_funcs radeon_fb_funcs = { .create_handle = radeon_user_framebuffer_create_handle, }; -void +int radeon_framebuffer_init(struct drm_device *dev, struct radeon_framebuffer *rfb, - struct drm_mode_fb_cmd *mode_cmd, + struct drm_mode_fb_cmd2 *mode_cmd, struct drm_gem_object *obj) { + int ret; rfb->obj = obj; - drm_framebuffer_init(dev, &rfb->base, &radeon_fb_funcs); + ret = drm_framebuffer_init(dev, &rfb->base, &radeon_fb_funcs); + if (ret) { + rfb->obj = NULL; + return ret; + } drm_helper_mode_fill_fb_struct(&rfb->base, mode_cmd); + return 0; } static struct drm_framebuffer * radeon_user_framebuffer_create(struct drm_device *dev, struct drm_file *file_priv, - struct drm_mode_fb_cmd *mode_cmd) + struct drm_mode_fb_cmd2 *mode_cmd) { struct drm_gem_object *obj; struct radeon_framebuffer *radeon_fb; + int ret; - obj = drm_gem_object_lookup(dev, file_priv, mode_cmd->handle); + obj = drm_gem_object_lookup(dev, file_priv, mode_cmd->handles[0]); if (obj == NULL) { dev_err(&dev->pdev->dev, "No GEM object associated to handle 0x%08X, " - "can't create framebuffer\n", mode_cmd->handle); + "can't create framebuffer\n", mode_cmd->handles[0]); return ERR_PTR(-ENOENT); } @@ -1115,7 +1134,12 @@ radeon_user_framebuffer_create(struct drm_device *dev, return ERR_PTR(-ENOMEM); } - radeon_framebuffer_init(dev, radeon_fb, mode_cmd, obj); + ret = radeon_framebuffer_init(dev, radeon_fb, mode_cmd, obj); + if (ret) { + kfree(radeon_fb); + drm_gem_object_unreference_unlocked(obj); + return ERR_PTR(ret); + } return &radeon_fb->base; } @@ -1131,11 +1155,6 @@ static const struct drm_mode_config_funcs radeon_mode_funcs = { .output_poll_changed = radeon_output_poll_changed }; -struct drm_prop_enum_list { - int type; - char *name; -}; - static struct drm_prop_enum_list radeon_tmds_pll_enum_list[] = { { 0, "driver" }, { 1, "bios" }, @@ -1160,86 +1179,53 @@ static struct drm_prop_enum_list radeon_underscan_enum_list[] = static int radeon_modeset_create_props(struct radeon_device *rdev) { - int i, sz; + int sz; if (rdev->is_atom_bios) { rdev->mode_info.coherent_mode_property = - drm_property_create(rdev->ddev, - DRM_MODE_PROP_RANGE, - "coherent", 2); + drm_property_create_range(rdev->ddev, 0 , "coherent", 0, 1); if (!rdev->mode_info.coherent_mode_property) return -ENOMEM; - - rdev->mode_info.coherent_mode_property->values[0] = 0; - rdev->mode_info.coherent_mode_property->values[1] = 1; } if (!ASIC_IS_AVIVO(rdev)) { sz = ARRAY_SIZE(radeon_tmds_pll_enum_list); rdev->mode_info.tmds_pll_property = - drm_property_create(rdev->ddev, - DRM_MODE_PROP_ENUM, - "tmds_pll", sz); - for (i = 0; i < sz; i++) { - drm_property_add_enum(rdev->mode_info.tmds_pll_property, - i, - radeon_tmds_pll_enum_list[i].type, - radeon_tmds_pll_enum_list[i].name); - } + drm_property_create_enum(rdev->ddev, 0, + "tmds_pll", + radeon_tmds_pll_enum_list, sz); } rdev->mode_info.load_detect_property = - drm_property_create(rdev->ddev, - DRM_MODE_PROP_RANGE, - "load detection", 2); + drm_property_create_range(rdev->ddev, 0, "load detection", 0, 1); if (!rdev->mode_info.load_detect_property) return -ENOMEM; - rdev->mode_info.load_detect_property->values[0] = 0; - rdev->mode_info.load_detect_property->values[1] = 1; drm_mode_create_scaling_mode_property(rdev->ddev); sz = ARRAY_SIZE(radeon_tv_std_enum_list); rdev->mode_info.tv_std_property = - drm_property_create(rdev->ddev, - DRM_MODE_PROP_ENUM, - "tv standard", sz); - for (i = 0; i < sz; i++) { - drm_property_add_enum(rdev->mode_info.tv_std_property, - i, - radeon_tv_std_enum_list[i].type, - radeon_tv_std_enum_list[i].name); - } + drm_property_create_enum(rdev->ddev, 0, + "tv standard", + radeon_tv_std_enum_list, sz); sz = ARRAY_SIZE(radeon_underscan_enum_list); rdev->mode_info.underscan_property = - drm_property_create(rdev->ddev, - DRM_MODE_PROP_ENUM, - "underscan", sz); - for (i = 0; i < sz; i++) { - drm_property_add_enum(rdev->mode_info.underscan_property, - i, - radeon_underscan_enum_list[i].type, - radeon_underscan_enum_list[i].name); - } + drm_property_create_enum(rdev->ddev, 0, + "underscan", + radeon_underscan_enum_list, sz); rdev->mode_info.underscan_hborder_property = - drm_property_create(rdev->ddev, - DRM_MODE_PROP_RANGE, - "underscan hborder", 2); + drm_property_create_range(rdev->ddev, 0, + "underscan hborder", 0, 128); if (!rdev->mode_info.underscan_hborder_property) return -ENOMEM; - rdev->mode_info.underscan_hborder_property->values[0] = 0; - rdev->mode_info.underscan_hborder_property->values[1] = 128; rdev->mode_info.underscan_vborder_property = - drm_property_create(rdev->ddev, - DRM_MODE_PROP_RANGE, - "underscan vborder", 2); + drm_property_create_range(rdev->ddev, 0, + "underscan vborder", 0, 128); if (!rdev->mode_info.underscan_vborder_property) return -ENOMEM; - rdev->mode_info.underscan_vborder_property->values[0] = 0; - rdev->mode_info.underscan_vborder_property->values[1] = 128; return 0; } @@ -1285,6 +1271,9 @@ int radeon_modeset_init(struct radeon_device *rdev) rdev->ddev->mode_config.max_height = 4096; } + rdev->ddev->mode_config.preferred_depth = 24; + rdev->ddev->mode_config.prefer_shadow = 1; + rdev->ddev->mode_config.fb_base = rdev->mc.aper_base; ret = radeon_modeset_create_props(rdev); @@ -1312,9 +1301,11 @@ int radeon_modeset_init(struct radeon_device *rdev) return ret; } - /* init dig PHYs */ - if (rdev->is_atom_bios) + /* init dig PHYs, disp eng pll */ + if (rdev->is_atom_bios) { radeon_atom_encoder_init(rdev); + radeon_atom_disp_eng_pll_init(rdev); + } /* initialize hpd */ radeon_hpd_init(rdev); diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c index 71499fc..15250fb 100644 --- a/drivers/gpu/drm/radeon/radeon_drv.c +++ b/drivers/gpu/drm/radeon/radeon_drv.c @@ -54,9 +54,13 @@ * 2.10.0 - fusion 2D tiling * 2.11.0 - backend map, initial compute support for the CS checker * 2.12.0 - RADEON_CS_KEEP_TILING_FLAGS + * 2.13.0 - virtual memory support, streamout + * 2.14.0 - add evergreen tiling informations + * 2.15.0 - add max_pipes query + * 2.16.0 - fix evergreen 2D tiled surface calculation */ #define KMS_DRIVER_MAJOR 2 -#define KMS_DRIVER_MINOR 12 +#define KMS_DRIVER_MINOR 16 #define KMS_DRIVER_PATCHLEVEL 0 int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags); int radeon_driver_unload_kms(struct drm_device *dev); @@ -84,6 +88,10 @@ int radeon_dma_ioctl_kms(struct drm_device *dev, void *data, struct drm_file *file_priv); int radeon_gem_object_init(struct drm_gem_object *obj); void radeon_gem_object_free(struct drm_gem_object *obj); +int radeon_gem_object_open(struct drm_gem_object *obj, + struct drm_file *file_priv); +void radeon_gem_object_close(struct drm_gem_object *obj, + struct drm_file *file_priv); extern int radeon_get_crtc_scanoutpos(struct drm_device *dev, int crtc, int *vpos, int *hpos); extern struct drm_ioctl_desc radeon_ioctls_kms[]; @@ -140,7 +148,7 @@ module_param_named(vramlimit, radeon_vram_limit, int, 0600); MODULE_PARM_DESC(agpmode, "AGP Mode (-1 == PCI)"); module_param_named(agpmode, radeon_agpmode, int, 0444); -MODULE_PARM_DESC(gartsize, "Size of PCIE/IGP gart to setup in megabytes (32,64, etc)\n"); +MODULE_PARM_DESC(gartsize, "Size of PCIE/IGP gart to setup in megabytes (32, 64, etc)"); module_param_named(gartsize, radeon_gart_size, int, 0600); MODULE_PARM_DESC(benchmark, "Run benchmark"); @@ -206,6 +214,21 @@ static struct pci_device_id pciidlist[] = { MODULE_DEVICE_TABLE(pci, pciidlist); #endif +static const struct file_operations radeon_driver_old_fops = { + .owner = THIS_MODULE, + .open = drm_open, + .release = drm_release, + .unlocked_ioctl = drm_ioctl, + .mmap = drm_mmap, + .poll = drm_poll, + .fasync = drm_fasync, + .read = drm_read, +#ifdef CONFIG_COMPAT + .compat_ioctl = radeon_compat_ioctl, +#endif + .llseek = noop_llseek, +}; + static struct drm_driver driver_old = { .driver_features = DRIVER_USE_AGP | DRIVER_USE_MTRR | DRIVER_PCI_DMA | DRIVER_SG | @@ -232,21 +255,7 @@ static struct drm_driver driver_old = { .reclaim_buffers = drm_core_reclaim_buffers, .ioctls = radeon_ioctls, .dma_ioctl = radeon_cp_buffers, - .fops = { - .owner = THIS_MODULE, - .open = drm_open, - .release = drm_release, - .unlocked_ioctl = drm_ioctl, - .mmap = drm_mmap, - .poll = drm_poll, - .fasync = drm_fasync, - .read = drm_read, -#ifdef CONFIG_COMPAT - .compat_ioctl = radeon_compat_ioctl, -#endif - .llseek = noop_llseek, - }, - + .fops = &radeon_driver_old_fops, .name = DRIVER_NAME, .desc = DRIVER_DESC, .date = DRIVER_DATE, @@ -304,6 +313,20 @@ radeon_pci_resume(struct pci_dev *pdev) return radeon_resume_kms(dev); } +static const struct file_operations radeon_driver_kms_fops = { + .owner = THIS_MODULE, + .open = drm_open, + .release = drm_release, + .unlocked_ioctl = drm_ioctl, + .mmap = radeon_mmap, + .poll = drm_poll, + .fasync = drm_fasync, + .read = drm_read, +#ifdef CONFIG_COMPAT + .compat_ioctl = radeon_kms_compat_ioctl, +#endif +}; + static struct drm_driver kms_driver = { .driver_features = DRIVER_USE_AGP | DRIVER_USE_MTRR | DRIVER_PCI_DMA | DRIVER_SG | @@ -335,24 +358,13 @@ static struct drm_driver kms_driver = { .ioctls = radeon_ioctls_kms, .gem_init_object = radeon_gem_object_init, .gem_free_object = radeon_gem_object_free, + .gem_open_object = radeon_gem_object_open, + .gem_close_object = radeon_gem_object_close, .dma_ioctl = radeon_dma_ioctl_kms, .dumb_create = radeon_mode_dumb_create, .dumb_map_offset = radeon_mode_dumb_mmap, .dumb_destroy = radeon_mode_dumb_destroy, - .fops = { - .owner = THIS_MODULE, - .open = drm_open, - .release = drm_release, - .unlocked_ioctl = drm_ioctl, - .mmap = radeon_mmap, - .poll = drm_poll, - .fasync = drm_fasync, - .read = drm_read, -#ifdef CONFIG_COMPAT - .compat_ioctl = radeon_kms_compat_ioctl, -#endif - }, - + .fops = &radeon_driver_kms_fops, .name = DRIVER_NAME, .desc = DRIVER_DESC, .date = DRIVER_DATE, diff --git a/drivers/gpu/drm/radeon/radeon_encoders.c b/drivers/gpu/drm/radeon/radeon_encoders.c index 4b27efa..7467069 100644 --- a/drivers/gpu/drm/radeon/radeon_encoders.c +++ b/drivers/gpu/drm/radeon/radeon_encoders.c @@ -202,6 +202,22 @@ radeon_get_connector_for_encoder(struct drm_encoder *encoder) return NULL; } +struct drm_connector * +radeon_get_connector_for_encoder_init(struct drm_encoder *encoder) +{ + struct drm_device *dev = encoder->dev; + struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder); + struct drm_connector *connector; + struct radeon_connector *radeon_connector; + + list_for_each_entry(connector, &dev->mode_config.connector_list, head) { + radeon_connector = to_radeon_connector(connector); + if (radeon_encoder->devices & radeon_connector->devices) + return connector; + } + return NULL; +} + struct drm_encoder *radeon_get_external_encoder(struct drm_encoder *encoder) { struct drm_device *dev = encoder->dev; @@ -288,3 +304,64 @@ void radeon_panel_mode_fixup(struct drm_encoder *encoder, } +bool radeon_dig_monitor_is_duallink(struct drm_encoder *encoder, + u32 pixel_clock) +{ + struct drm_device *dev = encoder->dev; + struct radeon_device *rdev = dev->dev_private; + struct drm_connector *connector; + struct radeon_connector *radeon_connector; + struct radeon_connector_atom_dig *dig_connector; + + connector = radeon_get_connector_for_encoder(encoder); + /* if we don't have an active device yet, just use one of + * the connectors tied to the encoder. + */ + if (!connector) + connector = radeon_get_connector_for_encoder_init(encoder); + radeon_connector = to_radeon_connector(connector); + + switch (connector->connector_type) { + case DRM_MODE_CONNECTOR_DVII: + case DRM_MODE_CONNECTOR_HDMIB: + if (radeon_connector->use_digital) { + /* HDMI 1.3 supports up to 340 Mhz over single link */ + if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector->edid)) { + if (pixel_clock > 340000) + return true; + else + return false; + } else { + if (pixel_clock > 165000) + return true; + else + return false; + } + } else + return false; + case DRM_MODE_CONNECTOR_DVID: + case DRM_MODE_CONNECTOR_HDMIA: + case DRM_MODE_CONNECTOR_DisplayPort: + dig_connector = radeon_connector->con_priv; + if ((dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_DISPLAYPORT) || + (dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_eDP)) + return false; + else { + /* HDMI 1.3 supports up to 340 Mhz over single link */ + if (ASIC_IS_DCE6(rdev) && drm_detect_hdmi_monitor(radeon_connector->edid)) { + if (pixel_clock > 340000) + return true; + else + return false; + } else { + if (pixel_clock > 165000) + return true; + else + return false; + } + } + default: + return false; + } +} + diff --git a/drivers/gpu/drm/radeon/radeon_family.h b/drivers/gpu/drm/radeon/radeon_family.h index ec2f1ea..d1fafea 100644 --- a/drivers/gpu/drm/radeon/radeon_family.h +++ b/drivers/gpu/drm/radeon/radeon_family.h @@ -87,6 +87,10 @@ enum radeon_family { CHIP_TURKS, CHIP_CAICOS, CHIP_CAYMAN, + CHIP_ARUBA, + CHIP_TAHITI, + CHIP_PITCAIRN, + CHIP_VERDE, CHIP_LAST, }; diff --git a/drivers/gpu/drm/radeon/radeon_fb.c b/drivers/gpu/drm/radeon/radeon_fb.c index 0b7b486..5906914 100644 --- a/drivers/gpu/drm/radeon/radeon_fb.c +++ b/drivers/gpu/drm/radeon/radeon_fb.c @@ -103,7 +103,7 @@ static void radeonfb_destroy_pinned_object(struct drm_gem_object *gobj) } static int radeonfb_create_pinned_object(struct radeon_fbdev *rfbdev, - struct drm_mode_fb_cmd *mode_cmd, + struct drm_mode_fb_cmd2 *mode_cmd, struct drm_gem_object **gobj_p) { struct radeon_device *rdev = rfbdev->rdev; @@ -114,13 +114,17 @@ static int radeonfb_create_pinned_object(struct radeon_fbdev *rfbdev, int ret; int aligned_size, size; int height = mode_cmd->height; + u32 bpp, depth; + + drm_fb_get_bpp_depth(mode_cmd->pixel_format, &depth, &bpp); /* need to align pitch with crtc limits */ - mode_cmd->pitch = radeon_align_pitch(rdev, mode_cmd->width, mode_cmd->bpp, fb_tiled) * ((mode_cmd->bpp + 1) / 8); + mode_cmd->pitches[0] = radeon_align_pitch(rdev, mode_cmd->width, bpp, + fb_tiled) * ((bpp + 1) / 8); if (rdev->family >= CHIP_R600) height = ALIGN(mode_cmd->height, 8); - size = mode_cmd->pitch * height; + size = mode_cmd->pitches[0] * height; aligned_size = ALIGN(size, PAGE_SIZE); ret = radeon_gem_object_create(rdev, aligned_size, 0, RADEON_GEM_DOMAIN_VRAM, @@ -137,7 +141,7 @@ static int radeonfb_create_pinned_object(struct radeon_fbdev *rfbdev, tiling_flags = RADEON_TILING_MACRO; #ifdef __BIG_ENDIAN - switch (mode_cmd->bpp) { + switch (bpp) { case 32: tiling_flags |= RADEON_TILING_SWAP_32BIT; break; @@ -151,7 +155,7 @@ static int radeonfb_create_pinned_object(struct radeon_fbdev *rfbdev, if (tiling_flags) { ret = radeon_bo_set_tiling_flags(rbo, tiling_flags | RADEON_TILING_SURFACE, - mode_cmd->pitch); + mode_cmd->pitches[0]); if (ret) dev_err(rdev->dev, "FB failed to set tiling flags\n"); } @@ -160,7 +164,10 @@ static int radeonfb_create_pinned_object(struct radeon_fbdev *rfbdev, ret = radeon_bo_reserve(rbo, false); if (unlikely(ret != 0)) goto out_unref; - ret = radeon_bo_pin(rbo, RADEON_GEM_DOMAIN_VRAM, NULL); + /* Only 27 bit offset for legacy CRTC */ + ret = radeon_bo_pin_restricted(rbo, RADEON_GEM_DOMAIN_VRAM, + ASIC_IS_AVIVO(rdev) ? 0 : 1 << 27, + NULL); if (ret) { radeon_bo_unreserve(rbo); goto out_unref; @@ -187,7 +194,7 @@ static int radeonfb_create(struct radeon_fbdev *rfbdev, struct radeon_device *rdev = rfbdev->rdev; struct fb_info *info; struct drm_framebuffer *fb = NULL; - struct drm_mode_fb_cmd mode_cmd; + struct drm_mode_fb_cmd2 mode_cmd; struct drm_gem_object *gobj = NULL; struct radeon_bo *rbo = NULL; struct device *device = &rdev->pdev->dev; @@ -201,10 +208,15 @@ static int radeonfb_create(struct radeon_fbdev *rfbdev, if ((sizes->surface_bpp == 24) && ASIC_IS_AVIVO(rdev)) sizes->surface_bpp = 32; - mode_cmd.bpp = sizes->surface_bpp; - mode_cmd.depth = sizes->surface_depth; + mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp, + sizes->surface_depth); ret = radeonfb_create_pinned_object(rfbdev, &mode_cmd, &gobj); + if (ret) { + DRM_ERROR("failed to create fbcon object %d\n", ret); + return ret; + } + rbo = gem_to_radeon_bo(gobj); /* okay we have an object now allocate the framebuffer */ @@ -216,7 +228,11 @@ static int radeonfb_create(struct radeon_fbdev *rfbdev, info->par = rfbdev; - radeon_framebuffer_init(rdev->ddev, &rfbdev->rfb, &mode_cmd, gobj); + ret = radeon_framebuffer_init(rdev->ddev, &rfbdev->rfb, &mode_cmd, gobj); + if (ret) { + DRM_ERROR("failed to initalise framebuffer %d\n", ret); + goto out_unref; + } fb = &rfbdev->rfb.base; @@ -228,7 +244,7 @@ static int radeonfb_create(struct radeon_fbdev *rfbdev, strcpy(info->fix.id, "radeondrmfb"); - drm_fb_helper_fill_fix(info, fb->pitch, fb->depth); + drm_fb_helper_fill_fix(info, fb->pitches[0], fb->depth); info->flags = FBINFO_DEFAULT | FBINFO_CAN_FORCE_OUTPUT; info->fbops = &radeonfb_ops; @@ -250,11 +266,7 @@ static int radeonfb_create(struct radeon_fbdev *rfbdev, info->apertures->ranges[0].base = rdev->ddev->mode_config.fb_base; info->apertures->ranges[0].size = rdev->mc.aper_size; - info->pixmap.size = 64*1024; - info->pixmap.buf_align = 8; - info->pixmap.access_align = 32; - info->pixmap.flags = FB_PIXMAP_SYSTEM; - info->pixmap.scan_align = 1; + /* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */ if (info->screen_base == NULL) { ret = -ENOSPC; @@ -271,7 +283,7 @@ static int radeonfb_create(struct radeon_fbdev *rfbdev, DRM_INFO("vram apper at 0x%lX\n", (unsigned long)rdev->mc.aper_base); DRM_INFO("size %lu\n", (unsigned long)radeon_bo_size(rbo)); DRM_INFO("fb depth is %d\n", fb->depth); - DRM_INFO(" pitch is %d\n", fb->pitch); + DRM_INFO(" pitch is %d\n", fb->pitches[0]); vga_switcheroo_client_fb_set(rdev->ddev->pdev, info); return 0; diff --git a/drivers/gpu/drm/radeon/radeon_fence.c b/drivers/gpu/drm/radeon/radeon_fence.c index 76ec0e9..4bd36a3 100644 --- a/drivers/gpu/drm/radeon/radeon_fence.c +++ b/drivers/gpu/drm/radeon/radeon_fence.c @@ -40,32 +40,24 @@ #include "radeon.h" #include "radeon_trace.h" -static void radeon_fence_write(struct radeon_device *rdev, u32 seq) +static void radeon_fence_write(struct radeon_device *rdev, u32 seq, int ring) { if (rdev->wb.enabled) { - u32 scratch_index; - if (rdev->wb.use_event) - scratch_index = R600_WB_EVENT_OFFSET + rdev->fence_drv.scratch_reg - rdev->scratch.reg_base; - else - scratch_index = RADEON_WB_SCRATCH_OFFSET + rdev->fence_drv.scratch_reg - rdev->scratch.reg_base; - rdev->wb.wb[scratch_index/4] = cpu_to_le32(seq); - } else - WREG32(rdev->fence_drv.scratch_reg, seq); + *rdev->fence_drv[ring].cpu_addr = cpu_to_le32(seq); + } else { + WREG32(rdev->fence_drv[ring].scratch_reg, seq); + } } -static u32 radeon_fence_read(struct radeon_device *rdev) +static u32 radeon_fence_read(struct radeon_device *rdev, int ring) { - u32 seq; + u32 seq = 0; if (rdev->wb.enabled) { - u32 scratch_index; - if (rdev->wb.use_event) - scratch_index = R600_WB_EVENT_OFFSET + rdev->fence_drv.scratch_reg - rdev->scratch.reg_base; - else - scratch_index = RADEON_WB_SCRATCH_OFFSET + rdev->fence_drv.scratch_reg - rdev->scratch.reg_base; - seq = le32_to_cpu(rdev->wb.wb[scratch_index/4]); - } else - seq = RREG32(rdev->fence_drv.scratch_reg); + seq = le32_to_cpu(*rdev->fence_drv[ring].cpu_addr); + } else { + seq = RREG32(rdev->fence_drv[ring].scratch_reg); + } return seq; } @@ -73,28 +65,28 @@ int radeon_fence_emit(struct radeon_device *rdev, struct radeon_fence *fence) { unsigned long irq_flags; - write_lock_irqsave(&rdev->fence_drv.lock, irq_flags); - if (fence->emited) { - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + write_lock_irqsave(&rdev->fence_lock, irq_flags); + if (fence->emitted) { + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); return 0; } - fence->seq = atomic_add_return(1, &rdev->fence_drv.seq); - if (!rdev->cp.ready) + fence->seq = atomic_add_return(1, &rdev->fence_drv[fence->ring].seq); + if (!rdev->ring[fence->ring].ready) /* FIXME: cp is not running assume everythings is done right * away */ - radeon_fence_write(rdev, fence->seq); + radeon_fence_write(rdev, fence->seq, fence->ring); else - radeon_fence_ring_emit(rdev, fence); + radeon_fence_ring_emit(rdev, fence->ring, fence); trace_radeon_fence_emit(rdev->ddev, fence->seq); - fence->emited = true; - list_move_tail(&fence->list, &rdev->fence_drv.emited); - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + fence->emitted = true; + list_move_tail(&fence->list, &rdev->fence_drv[fence->ring].emitted); + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); return 0; } -static bool radeon_fence_poll_locked(struct radeon_device *rdev) +static bool radeon_fence_poll_locked(struct radeon_device *rdev, int ring) { struct radeon_fence *fence; struct list_head *i, *n; @@ -102,34 +94,34 @@ static bool radeon_fence_poll_locked(struct radeon_device *rdev) bool wake = false; unsigned long cjiffies; - seq = radeon_fence_read(rdev); - if (seq != rdev->fence_drv.last_seq) { - rdev->fence_drv.last_seq = seq; - rdev->fence_drv.last_jiffies = jiffies; - rdev->fence_drv.last_timeout = RADEON_FENCE_JIFFIES_TIMEOUT; + seq = radeon_fence_read(rdev, ring); + if (seq != rdev->fence_drv[ring].last_seq) { + rdev->fence_drv[ring].last_seq = seq; + rdev->fence_drv[ring].last_jiffies = jiffies; + rdev->fence_drv[ring].last_timeout = RADEON_FENCE_JIFFIES_TIMEOUT; } else { cjiffies = jiffies; - if (time_after(cjiffies, rdev->fence_drv.last_jiffies)) { - cjiffies -= rdev->fence_drv.last_jiffies; - if (time_after(rdev->fence_drv.last_timeout, cjiffies)) { + if (time_after(cjiffies, rdev->fence_drv[ring].last_jiffies)) { + cjiffies -= rdev->fence_drv[ring].last_jiffies; + if (time_after(rdev->fence_drv[ring].last_timeout, cjiffies)) { /* update the timeout */ - rdev->fence_drv.last_timeout -= cjiffies; + rdev->fence_drv[ring].last_timeout -= cjiffies; } else { /* the 500ms timeout is elapsed we should test * for GPU lockup */ - rdev->fence_drv.last_timeout = 1; + rdev->fence_drv[ring].last_timeout = 1; } } else { /* wrap around update last jiffies, we will just wait * a little longer */ - rdev->fence_drv.last_jiffies = cjiffies; + rdev->fence_drv[ring].last_jiffies = cjiffies; } return false; } n = NULL; - list_for_each(i, &rdev->fence_drv.emited) { + list_for_each(i, &rdev->fence_drv[ring].emitted) { fence = list_entry(i, struct radeon_fence, list); if (fence->seq == seq) { n = i; @@ -141,11 +133,11 @@ static bool radeon_fence_poll_locked(struct radeon_device *rdev) i = n; do { n = i->prev; - list_move_tail(i, &rdev->fence_drv.signaled); + list_move_tail(i, &rdev->fence_drv[ring].signaled); fence = list_entry(i, struct radeon_fence, list); fence->signaled = true; i = n; - } while (i != &rdev->fence_drv.emited); + } while (i != &rdev->fence_drv[ring].emitted); wake = true; } return wake; @@ -157,14 +149,18 @@ static void radeon_fence_destroy(struct kref *kref) struct radeon_fence *fence; fence = container_of(kref, struct radeon_fence, kref); - write_lock_irqsave(&fence->rdev->fence_drv.lock, irq_flags); + write_lock_irqsave(&fence->rdev->fence_lock, irq_flags); list_del(&fence->list); - fence->emited = false; - write_unlock_irqrestore(&fence->rdev->fence_drv.lock, irq_flags); + fence->emitted = false; + write_unlock_irqrestore(&fence->rdev->fence_lock, irq_flags); + if (fence->semaphore) + radeon_semaphore_free(fence->rdev, fence->semaphore); kfree(fence); } -int radeon_fence_create(struct radeon_device *rdev, struct radeon_fence **fence) +int radeon_fence_create(struct radeon_device *rdev, + struct radeon_fence **fence, + int ring) { unsigned long irq_flags; @@ -174,18 +170,19 @@ int radeon_fence_create(struct radeon_device *rdev, struct radeon_fence **fence) } kref_init(&((*fence)->kref)); (*fence)->rdev = rdev; - (*fence)->emited = false; + (*fence)->emitted = false; (*fence)->signaled = false; (*fence)->seq = 0; + (*fence)->ring = ring; + (*fence)->semaphore = NULL; INIT_LIST_HEAD(&(*fence)->list); - write_lock_irqsave(&rdev->fence_drv.lock, irq_flags); - list_add_tail(&(*fence)->list, &rdev->fence_drv.created); - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + write_lock_irqsave(&rdev->fence_lock, irq_flags); + list_add_tail(&(*fence)->list, &rdev->fence_drv[ring].created); + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); return 0; } - bool radeon_fence_signaled(struct radeon_fence *fence) { unsigned long irq_flags; @@ -197,21 +194,21 @@ bool radeon_fence_signaled(struct radeon_fence *fence) if (fence->rdev->gpu_lockup) return true; - write_lock_irqsave(&fence->rdev->fence_drv.lock, irq_flags); + write_lock_irqsave(&fence->rdev->fence_lock, irq_flags); signaled = fence->signaled; /* if we are shuting down report all fence as signaled */ if (fence->rdev->shutdown) { signaled = true; } - if (!fence->emited) { - WARN(1, "Querying an unemited fence : %p !\n", fence); + if (!fence->emitted) { + WARN(1, "Querying an unemitted fence : %p !\n", fence); signaled = true; } if (!signaled) { - radeon_fence_poll_locked(fence->rdev); + radeon_fence_poll_locked(fence->rdev, fence->ring); signaled = fence->signaled; } - write_unlock_irqrestore(&fence->rdev->fence_drv.lock, irq_flags); + write_unlock_irqrestore(&fence->rdev->fence_lock, irq_flags); return signaled; } @@ -230,24 +227,24 @@ int radeon_fence_wait(struct radeon_fence *fence, bool intr) if (radeon_fence_signaled(fence)) { return 0; } - timeout = rdev->fence_drv.last_timeout; + timeout = rdev->fence_drv[fence->ring].last_timeout; retry: /* save current sequence used to check for GPU lockup */ - seq = rdev->fence_drv.last_seq; + seq = rdev->fence_drv[fence->ring].last_seq; trace_radeon_fence_wait_begin(rdev->ddev, seq); if (intr) { - radeon_irq_kms_sw_irq_get(rdev); - r = wait_event_interruptible_timeout(rdev->fence_drv.queue, + radeon_irq_kms_sw_irq_get(rdev, fence->ring); + r = wait_event_interruptible_timeout(rdev->fence_drv[fence->ring].queue, radeon_fence_signaled(fence), timeout); - radeon_irq_kms_sw_irq_put(rdev); + radeon_irq_kms_sw_irq_put(rdev, fence->ring); if (unlikely(r < 0)) { return r; } } else { - radeon_irq_kms_sw_irq_get(rdev); - r = wait_event_timeout(rdev->fence_drv.queue, + radeon_irq_kms_sw_irq_get(rdev, fence->ring); + r = wait_event_timeout(rdev->fence_drv[fence->ring].queue, radeon_fence_signaled(fence), timeout); - radeon_irq_kms_sw_irq_put(rdev); + radeon_irq_kms_sw_irq_put(rdev, fence->ring); } trace_radeon_fence_wait_end(rdev->ddev, seq); if (unlikely(!radeon_fence_signaled(fence))) { @@ -258,10 +255,11 @@ retry: timeout = r; goto retry; } - /* don't protect read access to rdev->fence_drv.last_seq + /* don't protect read access to rdev->fence_drv[t].last_seq * if we experiencing a lockup the value doesn't change */ - if (seq == rdev->fence_drv.last_seq && radeon_gpu_is_lockup(rdev)) { + if (seq == rdev->fence_drv[fence->ring].last_seq && + radeon_gpu_is_lockup(rdev, &rdev->ring[fence->ring])) { /* good news we believe it's a lockup */ printk(KERN_WARNING "GPU lockup (waiting for 0x%08X last fence id 0x%08X)\n", fence->seq, seq); @@ -272,20 +270,20 @@ retry: r = radeon_gpu_reset(rdev); if (r) return r; - radeon_fence_write(rdev, fence->seq); + radeon_fence_write(rdev, fence->seq, fence->ring); rdev->gpu_lockup = false; } timeout = RADEON_FENCE_JIFFIES_TIMEOUT; - write_lock_irqsave(&rdev->fence_drv.lock, irq_flags); - rdev->fence_drv.last_timeout = RADEON_FENCE_JIFFIES_TIMEOUT; - rdev->fence_drv.last_jiffies = jiffies; - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + write_lock_irqsave(&rdev->fence_lock, irq_flags); + rdev->fence_drv[fence->ring].last_timeout = RADEON_FENCE_JIFFIES_TIMEOUT; + rdev->fence_drv[fence->ring].last_jiffies = jiffies; + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); goto retry; } return 0; } -int radeon_fence_wait_next(struct radeon_device *rdev) +int radeon_fence_wait_next(struct radeon_device *rdev, int ring) { unsigned long irq_flags; struct radeon_fence *fence; @@ -294,21 +292,21 @@ int radeon_fence_wait_next(struct radeon_device *rdev) if (rdev->gpu_lockup) { return 0; } - write_lock_irqsave(&rdev->fence_drv.lock, irq_flags); - if (list_empty(&rdev->fence_drv.emited)) { - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + write_lock_irqsave(&rdev->fence_lock, irq_flags); + if (list_empty(&rdev->fence_drv[ring].emitted)) { + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); return 0; } - fence = list_entry(rdev->fence_drv.emited.next, + fence = list_entry(rdev->fence_drv[ring].emitted.next, struct radeon_fence, list); radeon_fence_ref(fence); - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); r = radeon_fence_wait(fence, false); radeon_fence_unref(&fence); return r; } -int radeon_fence_wait_last(struct radeon_device *rdev) +int radeon_fence_wait_last(struct radeon_device *rdev, int ring) { unsigned long irq_flags; struct radeon_fence *fence; @@ -317,15 +315,15 @@ int radeon_fence_wait_last(struct radeon_device *rdev) if (rdev->gpu_lockup) { return 0; } - write_lock_irqsave(&rdev->fence_drv.lock, irq_flags); - if (list_empty(&rdev->fence_drv.emited)) { - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + write_lock_irqsave(&rdev->fence_lock, irq_flags); + if (list_empty(&rdev->fence_drv[ring].emitted)) { + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); return 0; } - fence = list_entry(rdev->fence_drv.emited.prev, + fence = list_entry(rdev->fence_drv[ring].emitted.prev, struct radeon_fence, list); radeon_fence_ref(fence); - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); r = radeon_fence_wait(fence, false); radeon_fence_unref(&fence); return r; @@ -347,39 +345,97 @@ void radeon_fence_unref(struct radeon_fence **fence) } } -void radeon_fence_process(struct radeon_device *rdev) +void radeon_fence_process(struct radeon_device *rdev, int ring) { unsigned long irq_flags; bool wake; - write_lock_irqsave(&rdev->fence_drv.lock, irq_flags); - wake = radeon_fence_poll_locked(rdev); - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + write_lock_irqsave(&rdev->fence_lock, irq_flags); + wake = radeon_fence_poll_locked(rdev, ring); + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); if (wake) { - wake_up_all(&rdev->fence_drv.queue); + wake_up_all(&rdev->fence_drv[ring].queue); } } -int radeon_fence_driver_init(struct radeon_device *rdev) +int radeon_fence_count_emitted(struct radeon_device *rdev, int ring) +{ + unsigned long irq_flags; + int not_processed = 0; + + read_lock_irqsave(&rdev->fence_lock, irq_flags); + if (!rdev->fence_drv[ring].initialized) { + read_unlock_irqrestore(&rdev->fence_lock, irq_flags); + return 0; + } + + if (!list_empty(&rdev->fence_drv[ring].emitted)) { + struct list_head *ptr; + list_for_each(ptr, &rdev->fence_drv[ring].emitted) { + /* count up to 3, that's enought info */ + if (++not_processed >= 3) + break; + } + } + read_unlock_irqrestore(&rdev->fence_lock, irq_flags); + return not_processed; +} + +int radeon_fence_driver_start_ring(struct radeon_device *rdev, int ring) { unsigned long irq_flags; + uint64_t index; int r; - write_lock_irqsave(&rdev->fence_drv.lock, irq_flags); - r = radeon_scratch_get(rdev, &rdev->fence_drv.scratch_reg); - if (r) { - dev_err(rdev->dev, "fence failed to get scratch register\n"); - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); - return r; + write_lock_irqsave(&rdev->fence_lock, irq_flags); + radeon_scratch_free(rdev, rdev->fence_drv[ring].scratch_reg); + if (rdev->wb.use_event) { + rdev->fence_drv[ring].scratch_reg = 0; + index = R600_WB_EVENT_OFFSET + ring * 4; + } else { + r = radeon_scratch_get(rdev, &rdev->fence_drv[ring].scratch_reg); + if (r) { + dev_err(rdev->dev, "fence failed to get scratch register\n"); + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); + return r; + } + index = RADEON_WB_SCRATCH_OFFSET + + rdev->fence_drv[ring].scratch_reg - + rdev->scratch.reg_base; + } + rdev->fence_drv[ring].cpu_addr = &rdev->wb.wb[index/4]; + rdev->fence_drv[ring].gpu_addr = rdev->wb.gpu_addr + index; + radeon_fence_write(rdev, atomic_read(&rdev->fence_drv[ring].seq), ring); + rdev->fence_drv[ring].initialized = true; + DRM_INFO("fence driver on ring %d use gpu addr 0x%08Lx and cpu addr 0x%p\n", + ring, rdev->fence_drv[ring].gpu_addr, rdev->fence_drv[ring].cpu_addr); + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); + return 0; +} + +static void radeon_fence_driver_init_ring(struct radeon_device *rdev, int ring) +{ + rdev->fence_drv[ring].scratch_reg = -1; + rdev->fence_drv[ring].cpu_addr = NULL; + rdev->fence_drv[ring].gpu_addr = 0; + atomic_set(&rdev->fence_drv[ring].seq, 0); + INIT_LIST_HEAD(&rdev->fence_drv[ring].created); + INIT_LIST_HEAD(&rdev->fence_drv[ring].emitted); + INIT_LIST_HEAD(&rdev->fence_drv[ring].signaled); + init_waitqueue_head(&rdev->fence_drv[ring].queue); + rdev->fence_drv[ring].initialized = false; +} + +int radeon_fence_driver_init(struct radeon_device *rdev) +{ + unsigned long irq_flags; + int ring; + + write_lock_irqsave(&rdev->fence_lock, irq_flags); + for (ring = 0; ring < RADEON_NUM_RINGS; ring++) { + radeon_fence_driver_init_ring(rdev, ring); } - radeon_fence_write(rdev, 0); - atomic_set(&rdev->fence_drv.seq, 0); - INIT_LIST_HEAD(&rdev->fence_drv.created); - INIT_LIST_HEAD(&rdev->fence_drv.emited); - INIT_LIST_HEAD(&rdev->fence_drv.signaled); - init_waitqueue_head(&rdev->fence_drv.queue); - rdev->fence_drv.initialized = true; - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); if (radeon_debugfs_fence_init(rdev)) { dev_err(rdev->dev, "fence debugfs file creation failed\n"); } @@ -389,14 +445,18 @@ int radeon_fence_driver_init(struct radeon_device *rdev) void radeon_fence_driver_fini(struct radeon_device *rdev) { unsigned long irq_flags; - - if (!rdev->fence_drv.initialized) - return; - wake_up_all(&rdev->fence_drv.queue); - write_lock_irqsave(&rdev->fence_drv.lock, irq_flags); - radeon_scratch_free(rdev, rdev->fence_drv.scratch_reg); - write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); - rdev->fence_drv.initialized = false; + int ring; + + for (ring = 0; ring < RADEON_NUM_RINGS; ring++) { + if (!rdev->fence_drv[ring].initialized) + continue; + radeon_fence_wait_last(rdev, ring); + wake_up_all(&rdev->fence_drv[ring].queue); + write_lock_irqsave(&rdev->fence_lock, irq_flags); + radeon_scratch_free(rdev, rdev->fence_drv[ring].scratch_reg); + write_unlock_irqrestore(&rdev->fence_lock, irq_flags); + rdev->fence_drv[ring].initialized = false; + } } @@ -410,14 +470,21 @@ static int radeon_debugfs_fence_info(struct seq_file *m, void *data) struct drm_device *dev = node->minor->dev; struct radeon_device *rdev = dev->dev_private; struct radeon_fence *fence; - - seq_printf(m, "Last signaled fence 0x%08X\n", - radeon_fence_read(rdev)); - if (!list_empty(&rdev->fence_drv.emited)) { - fence = list_entry(rdev->fence_drv.emited.prev, - struct radeon_fence, list); - seq_printf(m, "Last emited fence %p with 0x%08X\n", - fence, fence->seq); + int i; + + for (i = 0; i < RADEON_NUM_RINGS; ++i) { + if (!rdev->fence_drv[i].initialized) + continue; + + seq_printf(m, "--- ring %d ---\n", i); + seq_printf(m, "Last signaled fence 0x%08X\n", + radeon_fence_read(rdev, i)); + if (!list_empty(&rdev->fence_drv[i].emitted)) { + fence = list_entry(rdev->fence_drv[i].emitted.prev, + struct radeon_fence, list); + seq_printf(m, "Last emitted fence %p with 0x%08X\n", + fence, fence->seq); + } } return 0; } diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c index ba7ab79..2a4c592 100644 --- a/drivers/gpu/drm/radeon/radeon_gart.c +++ b/drivers/gpu/drm/radeon/radeon_gart.c @@ -157,9 +157,6 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset, p = t / (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); for (i = 0; i < pages; i++, p++) { if (rdev->gart.pages[p]) { - if (!rdev->gart.ttm_alloced[p]) - pci_unmap_page(rdev->pdev, rdev->gart.pages_addr[p], - PAGE_SIZE, PCI_DMA_BIDIRECTIONAL); rdev->gart.pages[p] = NULL; rdev->gart.pages_addr[p] = rdev->dummy_page.addr; page_base = rdev->gart.pages_addr[p]; @@ -191,23 +188,7 @@ int radeon_gart_bind(struct radeon_device *rdev, unsigned offset, p = t / (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); for (i = 0; i < pages; i++, p++) { - /* we reverted the patch using dma_addr in TTM for now but this - * code stops building on alpha so just comment it out for now */ - if (0) { /*dma_addr[i] != DMA_ERROR_CODE) */ - rdev->gart.ttm_alloced[p] = true; - rdev->gart.pages_addr[p] = dma_addr[i]; - } else { - /* we need to support large memory configurations */ - /* assume that unbind have already been call on the range */ - rdev->gart.pages_addr[p] = pci_map_page(rdev->pdev, pagelist[i], - 0, PAGE_SIZE, - PCI_DMA_BIDIRECTIONAL); - if (pci_dma_mapping_error(rdev->pdev, rdev->gart.pages_addr[p])) { - /* FIXME: failed to map page (return -ENOMEM?) */ - radeon_gart_unbind(rdev, offset, pages); - return -ENOMEM; - } - } + rdev->gart.pages_addr[p] = dma_addr[i]; rdev->gart.pages[p] = pagelist[i]; if (rdev->gart.ptr) { page_base = rdev->gart.pages_addr[p]; @@ -274,12 +255,6 @@ int radeon_gart_init(struct radeon_device *rdev) radeon_gart_fini(rdev); return -ENOMEM; } - rdev->gart.ttm_alloced = kzalloc(sizeof(bool) * - rdev->gart.num_cpu_pages, GFP_KERNEL); - if (rdev->gart.ttm_alloced == NULL) { - radeon_gart_fini(rdev); - return -ENOMEM; - } /* set GART entry to point to the dummy page by default */ for (i = 0; i < rdev->gart.num_cpu_pages; i++) { rdev->gart.pages_addr[i] = rdev->dummy_page.addr; @@ -296,10 +271,418 @@ void radeon_gart_fini(struct radeon_device *rdev) rdev->gart.ready = false; kfree(rdev->gart.pages); kfree(rdev->gart.pages_addr); - kfree(rdev->gart.ttm_alloced); rdev->gart.pages = NULL; rdev->gart.pages_addr = NULL; - rdev->gart.ttm_alloced = NULL; radeon_dummy_page_fini(rdev); } + +/* + * vm helpers + * + * TODO bind a default page at vm initialization for default address + */ +int radeon_vm_manager_init(struct radeon_device *rdev) +{ + int r; + + rdev->vm_manager.enabled = false; + + /* mark first vm as always in use, it's the system one */ + /* allocate enough for 2 full VM pts */ + r = radeon_sa_bo_manager_init(rdev, &rdev->vm_manager.sa_manager, + rdev->vm_manager.max_pfn * 8 * 2, + RADEON_GEM_DOMAIN_VRAM); + if (r) { + dev_err(rdev->dev, "failed to allocate vm bo (%dKB)\n", + (rdev->vm_manager.max_pfn * 8) >> 10); + return r; + } + + r = rdev->vm_manager.funcs->init(rdev); + if (r == 0) + rdev->vm_manager.enabled = true; + + return r; +} + +/* cs mutex must be lock */ +static void radeon_vm_unbind_locked(struct radeon_device *rdev, + struct radeon_vm *vm) +{ + struct radeon_bo_va *bo_va; + + if (vm->id == -1) { + return; + } + + /* wait for vm use to end */ + if (vm->fence) { + radeon_fence_wait(vm->fence, false); + radeon_fence_unref(&vm->fence); + } + + /* hw unbind */ + rdev->vm_manager.funcs->unbind(rdev, vm); + rdev->vm_manager.use_bitmap &= ~(1 << vm->id); + list_del_init(&vm->list); + vm->id = -1; + radeon_sa_bo_free(rdev, &vm->sa_bo); + vm->pt = NULL; + + list_for_each_entry(bo_va, &vm->va, vm_list) { + bo_va->valid = false; + } +} + +void radeon_vm_manager_fini(struct radeon_device *rdev) +{ + if (rdev->vm_manager.sa_manager.bo == NULL) + return; + radeon_vm_manager_suspend(rdev); + rdev->vm_manager.funcs->fini(rdev); + radeon_sa_bo_manager_fini(rdev, &rdev->vm_manager.sa_manager); + rdev->vm_manager.enabled = false; +} + +int radeon_vm_manager_start(struct radeon_device *rdev) +{ + if (rdev->vm_manager.sa_manager.bo == NULL) { + return -EINVAL; + } + return radeon_sa_bo_manager_start(rdev, &rdev->vm_manager.sa_manager); +} + +int radeon_vm_manager_suspend(struct radeon_device *rdev) +{ + struct radeon_vm *vm, *tmp; + + radeon_mutex_lock(&rdev->cs_mutex); + /* unbind all active vm */ + list_for_each_entry_safe(vm, tmp, &rdev->vm_manager.lru_vm, list) { + radeon_vm_unbind_locked(rdev, vm); + } + rdev->vm_manager.funcs->fini(rdev); + radeon_mutex_unlock(&rdev->cs_mutex); + return radeon_sa_bo_manager_suspend(rdev, &rdev->vm_manager.sa_manager); +} + +/* cs mutex must be lock */ +void radeon_vm_unbind(struct radeon_device *rdev, struct radeon_vm *vm) +{ + mutex_lock(&vm->mutex); + radeon_vm_unbind_locked(rdev, vm); + mutex_unlock(&vm->mutex); +} + +/* cs mutex must be lock & vm mutex must be lock */ +int radeon_vm_bind(struct radeon_device *rdev, struct radeon_vm *vm) +{ + struct radeon_vm *vm_evict; + unsigned i; + int id = -1, r; + + if (vm == NULL) { + return -EINVAL; + } + + if (vm->id != -1) { + /* update lru */ + list_del_init(&vm->list); + list_add_tail(&vm->list, &rdev->vm_manager.lru_vm); + return 0; + } + +retry: + r = radeon_sa_bo_new(rdev, &rdev->vm_manager.sa_manager, &vm->sa_bo, + RADEON_GPU_PAGE_ALIGN(vm->last_pfn * 8), + RADEON_GPU_PAGE_SIZE); + if (r) { + if (list_empty(&rdev->vm_manager.lru_vm)) { + return r; + } + vm_evict = list_first_entry(&rdev->vm_manager.lru_vm, struct radeon_vm, list); + radeon_vm_unbind(rdev, vm_evict); + goto retry; + } + vm->pt = rdev->vm_manager.sa_manager.cpu_ptr; + vm->pt += (vm->sa_bo.offset >> 3); + vm->pt_gpu_addr = rdev->vm_manager.sa_manager.gpu_addr; + vm->pt_gpu_addr += vm->sa_bo.offset; + memset(vm->pt, 0, RADEON_GPU_PAGE_ALIGN(vm->last_pfn * 8)); + +retry_id: + /* search for free vm */ + for (i = 0; i < rdev->vm_manager.nvm; i++) { + if (!(rdev->vm_manager.use_bitmap & (1 << i))) { + id = i; + break; + } + } + /* evict vm if necessary */ + if (id == -1) { + vm_evict = list_first_entry(&rdev->vm_manager.lru_vm, struct radeon_vm, list); + radeon_vm_unbind(rdev, vm_evict); + goto retry_id; + } + + /* do hw bind */ + r = rdev->vm_manager.funcs->bind(rdev, vm, id); + if (r) { + radeon_sa_bo_free(rdev, &vm->sa_bo); + return r; + } + rdev->vm_manager.use_bitmap |= 1 << id; + vm->id = id; + list_add_tail(&vm->list, &rdev->vm_manager.lru_vm); + return radeon_vm_bo_update_pte(rdev, vm, rdev->ib_pool.sa_manager.bo, + &rdev->ib_pool.sa_manager.bo->tbo.mem); +} + +/* object have to be reserved */ +int radeon_vm_bo_add(struct radeon_device *rdev, + struct radeon_vm *vm, + struct radeon_bo *bo, + uint64_t offset, + uint32_t flags) +{ + struct radeon_bo_va *bo_va, *tmp; + struct list_head *head; + uint64_t size = radeon_bo_size(bo), last_offset = 0; + unsigned last_pfn; + + bo_va = kzalloc(sizeof(struct radeon_bo_va), GFP_KERNEL); + if (bo_va == NULL) { + return -ENOMEM; + } + bo_va->vm = vm; + bo_va->bo = bo; + bo_va->soffset = offset; + bo_va->eoffset = offset + size; + bo_va->flags = flags; + bo_va->valid = false; + INIT_LIST_HEAD(&bo_va->bo_list); + INIT_LIST_HEAD(&bo_va->vm_list); + /* make sure object fit at this offset */ + if (bo_va->soffset >= bo_va->eoffset) { + kfree(bo_va); + return -EINVAL; + } + + last_pfn = bo_va->eoffset / RADEON_GPU_PAGE_SIZE; + if (last_pfn > rdev->vm_manager.max_pfn) { + kfree(bo_va); + dev_err(rdev->dev, "va above limit (0x%08X > 0x%08X)\n", + last_pfn, rdev->vm_manager.max_pfn); + return -EINVAL; + } + + mutex_lock(&vm->mutex); + if (last_pfn > vm->last_pfn) { + /* release mutex and lock in right order */ + mutex_unlock(&vm->mutex); + radeon_mutex_lock(&rdev->cs_mutex); + mutex_lock(&vm->mutex); + /* and check again */ + if (last_pfn > vm->last_pfn) { + /* grow va space 32M by 32M */ + unsigned align = ((32 << 20) >> 12) - 1; + radeon_vm_unbind_locked(rdev, vm); + vm->last_pfn = (last_pfn + align) & ~align; + } + radeon_mutex_unlock(&rdev->cs_mutex); + } + head = &vm->va; + last_offset = 0; + list_for_each_entry(tmp, &vm->va, vm_list) { + if (bo_va->soffset >= last_offset && bo_va->eoffset < tmp->soffset) { + /* bo can be added before this one */ + break; + } + if (bo_va->soffset >= tmp->soffset && bo_va->soffset < tmp->eoffset) { + /* bo and tmp overlap, invalid offset */ + dev_err(rdev->dev, "bo %p va 0x%08X conflict with (bo %p 0x%08X 0x%08X)\n", + bo, (unsigned)bo_va->soffset, tmp->bo, + (unsigned)tmp->soffset, (unsigned)tmp->eoffset); + kfree(bo_va); + mutex_unlock(&vm->mutex); + return -EINVAL; + } + last_offset = tmp->eoffset; + head = &tmp->vm_list; + } + list_add(&bo_va->vm_list, head); + list_add_tail(&bo_va->bo_list, &bo->va); + mutex_unlock(&vm->mutex); + return 0; +} + +static u64 radeon_vm_get_addr(struct radeon_device *rdev, + struct ttm_mem_reg *mem, + unsigned pfn) +{ + u64 addr = 0; + + switch (mem->mem_type) { + case TTM_PL_VRAM: + addr = (mem->start << PAGE_SHIFT); + addr += pfn * RADEON_GPU_PAGE_SIZE; + addr += rdev->vm_manager.vram_base_offset; + break; + case TTM_PL_TT: + /* offset inside page table */ + addr = mem->start << PAGE_SHIFT; + addr += pfn * RADEON_GPU_PAGE_SIZE; + addr = addr >> PAGE_SHIFT; + /* page table offset */ + addr = rdev->gart.pages_addr[addr]; + /* in case cpu page size != gpu page size*/ + addr += (pfn * RADEON_GPU_PAGE_SIZE) & (~PAGE_MASK); + break; + default: + break; + } + return addr; +} + +/* object have to be reserved & cs mutex took & vm mutex took */ +int radeon_vm_bo_update_pte(struct radeon_device *rdev, + struct radeon_vm *vm, + struct radeon_bo *bo, + struct ttm_mem_reg *mem) +{ + struct radeon_bo_va *bo_va; + unsigned ngpu_pages, i; + uint64_t addr = 0, pfn; + uint32_t flags; + + /* nothing to do if vm isn't bound */ + if (vm->id == -1) + return 0;; + + bo_va = radeon_bo_va(bo, vm); + if (bo_va == NULL) { + dev_err(rdev->dev, "bo %p not in vm %p\n", bo, vm); + return -EINVAL; + } + + if (bo_va->valid) + return 0; + + ngpu_pages = radeon_bo_ngpu_pages(bo); + bo_va->flags &= ~RADEON_VM_PAGE_VALID; + bo_va->flags &= ~RADEON_VM_PAGE_SYSTEM; + if (mem) { + if (mem->mem_type != TTM_PL_SYSTEM) { + bo_va->flags |= RADEON_VM_PAGE_VALID; + bo_va->valid = true; + } + if (mem->mem_type == TTM_PL_TT) { + bo_va->flags |= RADEON_VM_PAGE_SYSTEM; + } + } + pfn = bo_va->soffset / RADEON_GPU_PAGE_SIZE; + flags = rdev->vm_manager.funcs->page_flags(rdev, bo_va->vm, bo_va->flags); + for (i = 0, addr = 0; i < ngpu_pages; i++) { + if (mem && bo_va->valid) { + addr = radeon_vm_get_addr(rdev, mem, i); + } + rdev->vm_manager.funcs->set_page(rdev, bo_va->vm, i + pfn, addr, flags); + } + rdev->vm_manager.funcs->tlb_flush(rdev, bo_va->vm); + return 0; +} + +/* object have to be reserved */ +int radeon_vm_bo_rmv(struct radeon_device *rdev, + struct radeon_vm *vm, + struct radeon_bo *bo) +{ + struct radeon_bo_va *bo_va; + + bo_va = radeon_bo_va(bo, vm); + if (bo_va == NULL) + return 0; + + radeon_mutex_lock(&rdev->cs_mutex); + mutex_lock(&vm->mutex); + radeon_vm_bo_update_pte(rdev, vm, bo, NULL); + radeon_mutex_unlock(&rdev->cs_mutex); + list_del(&bo_va->vm_list); + mutex_unlock(&vm->mutex); + list_del(&bo_va->bo_list); + + kfree(bo_va); + return 0; +} + +void radeon_vm_bo_invalidate(struct radeon_device *rdev, + struct radeon_bo *bo) +{ + struct radeon_bo_va *bo_va; + + BUG_ON(!atomic_read(&bo->tbo.reserved)); + list_for_each_entry(bo_va, &bo->va, bo_list) { + bo_va->valid = false; + } +} + +int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm) +{ + int r; + + vm->id = -1; + vm->fence = NULL; + mutex_init(&vm->mutex); + INIT_LIST_HEAD(&vm->list); + INIT_LIST_HEAD(&vm->va); + /* SI requires equal sized PTs for all VMs, so always set + * last_pfn to max_pfn. cayman allows variable sized + * pts so we can grow then as needed. Once we switch + * to two level pts we can unify this again. + */ + if (rdev->family >= CHIP_TAHITI) + vm->last_pfn = rdev->vm_manager.max_pfn; + else + vm->last_pfn = 0; + /* map the ib pool buffer at 0 in virtual address space, set + * read only + */ + r = radeon_vm_bo_add(rdev, vm, rdev->ib_pool.sa_manager.bo, 0, + RADEON_VM_PAGE_READABLE | RADEON_VM_PAGE_SNOOPED); + return r; +} + +void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm) +{ + struct radeon_bo_va *bo_va, *tmp; + int r; + + radeon_mutex_lock(&rdev->cs_mutex); + mutex_lock(&vm->mutex); + radeon_vm_unbind_locked(rdev, vm); + radeon_mutex_unlock(&rdev->cs_mutex); + + /* remove all bo */ + r = radeon_bo_reserve(rdev->ib_pool.sa_manager.bo, false); + if (!r) { + bo_va = radeon_bo_va(rdev->ib_pool.sa_manager.bo, vm); + list_del_init(&bo_va->bo_list); + list_del_init(&bo_va->vm_list); + radeon_bo_unreserve(rdev->ib_pool.sa_manager.bo); + kfree(bo_va); + } + if (!list_empty(&vm->va)) { + dev_err(rdev->dev, "still active bo inside vm\n"); + } + list_for_each_entry_safe(bo_va, tmp, &vm->va, vm_list) { + list_del_init(&bo_va->vm_list); + r = radeon_bo_reserve(bo_va->bo, false); + if (!r) { + list_del_init(&bo_va->bo_list); + radeon_bo_unreserve(bo_va->bo); + kfree(bo_va); + } + } + mutex_unlock(&vm->mutex); +} diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c index aa1ca2d..c7008b5 100644 --- a/drivers/gpu/drm/radeon/radeon_gem.c +++ b/drivers/gpu/drm/radeon/radeon_gem.c @@ -75,32 +75,6 @@ int radeon_gem_object_create(struct radeon_device *rdev, int size, return 0; } -int radeon_gem_object_pin(struct drm_gem_object *obj, uint32_t pin_domain, - uint64_t *gpu_addr) -{ - struct radeon_bo *robj = gem_to_radeon_bo(obj); - int r; - - r = radeon_bo_reserve(robj, false); - if (unlikely(r != 0)) - return r; - r = radeon_bo_pin(robj, pin_domain, gpu_addr); - radeon_bo_unreserve(robj); - return r; -} - -void radeon_gem_object_unpin(struct drm_gem_object *obj) -{ - struct radeon_bo *robj = gem_to_radeon_bo(obj); - int r; - - r = radeon_bo_reserve(robj, false); - if (likely(r == 0)) { - radeon_bo_unpin(robj); - radeon_bo_unreserve(robj); - } -} - int radeon_gem_set_domain(struct drm_gem_object *gobj, uint32_t rdomain, uint32_t wdomain) { @@ -142,6 +116,44 @@ void radeon_gem_fini(struct radeon_device *rdev) radeon_bo_force_delete(rdev); } +/* + * Call from drm_gem_handle_create which appear in both new and open ioctl + * case. + */ +int radeon_gem_object_open(struct drm_gem_object *obj, struct drm_file *file_priv) +{ + return 0; +} + +void radeon_gem_object_close(struct drm_gem_object *obj, + struct drm_file *file_priv) +{ + struct radeon_bo *rbo = gem_to_radeon_bo(obj); + struct radeon_device *rdev = rbo->rdev; + struct radeon_fpriv *fpriv = file_priv->driver_priv; + struct radeon_vm *vm = &fpriv->vm; + struct radeon_bo_va *bo_va, *tmp; + + if (rdev->family < CHIP_CAYMAN) { + return; + } + + if (radeon_bo_reserve(rbo, false)) { + return; + } + list_for_each_entry_safe(bo_va, tmp, &rbo->va, bo_list) { + if (bo_va->vm == vm) { + /* remove from this vm address space */ + mutex_lock(&vm->mutex); + list_del(&bo_va->vm_list); + mutex_unlock(&vm->mutex); + list_del(&bo_va->bo_list); + kfree(bo_va); + } + } + radeon_bo_unreserve(rbo); +} + /* * GEM ioctls. @@ -152,6 +164,7 @@ int radeon_gem_info_ioctl(struct drm_device *dev, void *data, struct radeon_device *rdev = dev->dev_private; struct drm_radeon_gem_info *args = data; struct ttm_mem_type_manager *man; + unsigned i; man = &rdev->mman.bdev.man[TTM_PL_VRAM]; @@ -160,8 +173,9 @@ int radeon_gem_info_ioctl(struct drm_device *dev, void *data, if (rdev->stollen_vga_memory) args->vram_visible -= radeon_bo_size(rdev->stollen_vga_memory); args->vram_visible -= radeon_fbdev_total_size(rdev); - args->gart_size = rdev->mc.gtt_size - rdev->cp.ring_size - 4096 - - RADEON_IB_POOL_SIZE*64*1024; + args->gart_size = rdev->mc.gtt_size - 4096 - RADEON_IB_POOL_SIZE*64*1024; + for(i = 0; i < RADEON_NUM_RINGS; ++i) + args->gart_size -= rdev->ring[i].ring_size; return 0; } @@ -352,6 +366,109 @@ out: return r; } +int radeon_gem_va_ioctl(struct drm_device *dev, void *data, + struct drm_file *filp) +{ + struct drm_radeon_gem_va *args = data; + struct drm_gem_object *gobj; + struct radeon_device *rdev = dev->dev_private; + struct radeon_fpriv *fpriv = filp->driver_priv; + struct radeon_bo *rbo; + struct radeon_bo_va *bo_va; + u32 invalid_flags; + int r = 0; + + if (!rdev->vm_manager.enabled) { + args->operation = RADEON_VA_RESULT_ERROR; + return -ENOTTY; + } + + /* !! DONT REMOVE !! + * We don't support vm_id yet, to be sure we don't have have broken + * userspace, reject anyone trying to use non 0 value thus moving + * forward we can use those fields without breaking existant userspace + */ + if (args->vm_id) { + args->operation = RADEON_VA_RESULT_ERROR; + return -EINVAL; + } + + if (args->offset < RADEON_VA_RESERVED_SIZE) { + dev_err(&dev->pdev->dev, + "offset 0x%lX is in reserved area 0x%X\n", + (unsigned long)args->offset, + RADEON_VA_RESERVED_SIZE); + args->operation = RADEON_VA_RESULT_ERROR; + return -EINVAL; + } + + /* don't remove, we need to enforce userspace to set the snooped flag + * otherwise we will endup with broken userspace and we won't be able + * to enable this feature without adding new interface + */ + invalid_flags = RADEON_VM_PAGE_VALID | RADEON_VM_PAGE_SYSTEM; + if ((args->flags & invalid_flags)) { + dev_err(&dev->pdev->dev, "invalid flags 0x%08X vs 0x%08X\n", + args->flags, invalid_flags); + args->operation = RADEON_VA_RESULT_ERROR; + return -EINVAL; + } + if (!(args->flags & RADEON_VM_PAGE_SNOOPED)) { + dev_err(&dev->pdev->dev, "only supported snooped mapping for now\n"); + args->operation = RADEON_VA_RESULT_ERROR; + return -EINVAL; + } + + switch (args->operation) { + case RADEON_VA_MAP: + case RADEON_VA_UNMAP: + break; + default: + dev_err(&dev->pdev->dev, "unsupported operation %d\n", + args->operation); + args->operation = RADEON_VA_RESULT_ERROR; + return -EINVAL; + } + + gobj = drm_gem_object_lookup(dev, filp, args->handle); + if (gobj == NULL) { + args->operation = RADEON_VA_RESULT_ERROR; + return -ENOENT; + } + rbo = gem_to_radeon_bo(gobj); + r = radeon_bo_reserve(rbo, false); + if (r) { + args->operation = RADEON_VA_RESULT_ERROR; + drm_gem_object_unreference_unlocked(gobj); + return r; + } + switch (args->operation) { + case RADEON_VA_MAP: + bo_va = radeon_bo_va(rbo, &fpriv->vm); + if (bo_va) { + args->operation = RADEON_VA_RESULT_VA_EXIST; + args->offset = bo_va->soffset; + goto out; + } + r = radeon_vm_bo_add(rdev, &fpriv->vm, rbo, + args->offset, args->flags); + break; + case RADEON_VA_UNMAP: + r = radeon_vm_bo_rmv(rdev, &fpriv->vm, rbo); + break; + default: + break; + } + args->operation = RADEON_VA_RESULT_OK; + if (r) { + args->operation = RADEON_VA_RESULT_ERROR; + } +out: + radeon_bo_unreserve(rbo); + drm_gem_object_unreference_unlocked(gobj); + return r; +} + int radeon_mode_dumb_create(struct drm_file *file_priv, struct drm_device *dev, struct drm_mode_create_dumb *args) diff --git a/drivers/gpu/drm/radeon/radeon_i2c.c b/drivers/gpu/drm/radeon/radeon_i2c.c index cf20351..6076e85 100644 --- a/drivers/gpu/drm/radeon/radeon_i2c.c +++ b/drivers/gpu/drm/radeon/radeon_i2c.c @@ -26,10 +26,15 @@ #include #include "drmP.h" +#include "drm_edid.h" #include "radeon_drm.h" #include "radeon.h" #include "atom.h" +extern int radeon_atom_hw_i2c_xfer(struct i2c_adapter *i2c_adap, + struct i2c_msg *msgs, int num); +extern u32 radeon_atom_hw_i2c_func(struct i2c_adapter *adap); + /** * radeon_ddc_probe * @@ -41,13 +46,13 @@ bool radeon_ddc_probe(struct radeon_connector *radeon_connector, bool use_aux) int ret; struct i2c_msg msgs[] = { { - .addr = 0x50, + .addr = DDC_ADDR, .flags = 0, .len = 1, .buf = &out, }, { - .addr = 0x50, + .addr = DDC_ADDR, .flags = I2C_M_RD, .len = 8, .buf = buf, @@ -888,6 +893,11 @@ static const struct i2c_algorithm radeon_i2c_algo = { .functionality = radeon_hw_i2c_func, }; +static const struct i2c_algorithm radeon_atom_i2c_algo = { + .master_xfer = radeon_atom_hw_i2c_xfer, + .functionality = radeon_atom_hw_i2c_func, +}; + struct radeon_i2c_chan *radeon_i2c_create(struct drm_device *dev, struct radeon_i2c_bus_rec *rec, const char *name) @@ -907,6 +917,7 @@ struct radeon_i2c_chan *radeon_i2c_create(struct drm_device *dev, i2c->rec = *rec; i2c->adapter.owner = THIS_MODULE; i2c->adapter.class = I2C_CLASS_DDC; + i2c->adapter.dev.parent = &dev->pdev->dev; i2c->dev = dev; i2c_set_adapdata(&i2c->adapter, i2c); if (rec->mm_i2c || @@ -923,6 +934,18 @@ struct radeon_i2c_chan *radeon_i2c_create(struct drm_device *dev, DRM_ERROR("Failed to register hw i2c %s\n", name); goto out_free; } + } else if (rec->hw_capable && + radeon_hw_i2c && + ASIC_IS_DCE3(rdev)) { + /* hw i2c using atom */ + snprintf(i2c->adapter.name, sizeof(i2c->adapter.name), + "Radeon i2c hw bus %s", name); + i2c->adapter.algo = &radeon_atom_i2c_algo; + ret = i2c_add_adapter(&i2c->adapter); + if (ret) { + DRM_ERROR("Failed to register hw i2c %s\n", name); + goto out_free; + } } else { /* set the radeon bit adapter */ snprintf(i2c->adapter.name, sizeof(i2c->adapter.name), @@ -934,10 +957,8 @@ struct radeon_i2c_chan *radeon_i2c_create(struct drm_device *dev, i2c->algo.bit.setscl = set_clock; i2c->algo.bit.getsda = get_data; i2c->algo.bit.getscl = get_clock; - i2c->algo.bit.udelay = 20; - /* vesa says 2.2 ms is enough, 1 jiffy doesn't seem to always - * make this, 2 jiffies is a lot more reliable */ - i2c->algo.bit.timeout = 2; + i2c->algo.bit.udelay = 10; + i2c->algo.bit.timeout = usecs_to_jiffies(2200); /* from VESA */ i2c->algo.bit.data = i2c; ret = i2c_bit_add_bus(&i2c->adapter); if (ret) { @@ -967,6 +988,7 @@ struct radeon_i2c_chan *radeon_i2c_create_dp(struct drm_device *dev, i2c->rec = *rec; i2c->adapter.owner = THIS_MODULE; i2c->adapter.class = I2C_CLASS_DDC; + i2c->adapter.dev.parent = &dev->pdev->dev; i2c->dev = dev; snprintf(i2c->adapter.name, sizeof(i2c->adapter.name), "Radeon aux bus %s", name); diff --git a/drivers/gpu/drm/radeon/radeon_irq_kms.c b/drivers/gpu/drm/radeon/radeon_irq_kms.c index 4f9496e..645dcbf 100644 --- a/drivers/gpu/drm/radeon/radeon_irq_kms.c +++ b/drivers/gpu/drm/radeon/radeon_irq_kms.c @@ -65,7 +65,8 @@ void radeon_driver_irq_preinstall_kms(struct drm_device *dev) unsigned i; /* Disable *all* interrupts */ - rdev->irq.sw_int = false; + for (i = 0; i < RADEON_NUM_RINGS; i++) + rdev->irq.sw_int[i] = false; rdev->irq.gui_idle = false; for (i = 0; i < RADEON_MAX_HPD_PINS; i++) rdev->irq.hpd[i] = false; @@ -81,9 +82,11 @@ void radeon_driver_irq_preinstall_kms(struct drm_device *dev) int radeon_driver_irq_postinstall_kms(struct drm_device *dev) { struct radeon_device *rdev = dev->dev_private; + unsigned i; dev->max_vblank_count = 0x001fffff; - rdev->irq.sw_int = true; + for (i = 0; i < RADEON_NUM_RINGS; i++) + rdev->irq.sw_int[i] = true; radeon_irq_set(rdev); return 0; } @@ -97,7 +100,8 @@ void radeon_driver_irq_uninstall_kms(struct drm_device *dev) return; } /* Disable *all* interrupts */ - rdev->irq.sw_int = false; + for (i = 0; i < RADEON_NUM_RINGS; i++) + rdev->irq.sw_int[i] = false; rdev->irq.gui_idle = false; for (i = 0; i < RADEON_MAX_HPD_PINS; i++) rdev->irq.hpd[i] = false; @@ -216,26 +220,26 @@ void radeon_irq_kms_fini(struct radeon_device *rdev) flush_work_sync(&rdev->hotplug_work); } -void radeon_irq_kms_sw_irq_get(struct radeon_device *rdev) +void radeon_irq_kms_sw_irq_get(struct radeon_device *rdev, int ring) { unsigned long irqflags; spin_lock_irqsave(&rdev->irq.sw_lock, irqflags); - if (rdev->ddev->irq_enabled && (++rdev->irq.sw_refcount == 1)) { - rdev->irq.sw_int = true; + if (rdev->ddev->irq_enabled && (++rdev->irq.sw_refcount[ring] == 1)) { + rdev->irq.sw_int[ring] = true; radeon_irq_set(rdev); } spin_unlock_irqrestore(&rdev->irq.sw_lock, irqflags); } -void radeon_irq_kms_sw_irq_put(struct radeon_device *rdev) +void radeon_irq_kms_sw_irq_put(struct radeon_device *rdev, int ring) { unsigned long irqflags; spin_lock_irqsave(&rdev->irq.sw_lock, irqflags); - BUG_ON(rdev->ddev->irq_enabled && rdev->irq.sw_refcount <= 0); - if (rdev->ddev->irq_enabled && (--rdev->irq.sw_refcount == 0)) { - rdev->irq.sw_int = false; + BUG_ON(rdev->ddev->irq_enabled && rdev->irq.sw_refcount[ring] <= 0); + if (rdev->ddev->irq_enabled && (--rdev->irq.sw_refcount[ring] == 0)) { + rdev->irq.sw_int[ring] = false; radeon_irq_set(rdev); } spin_unlock_irqrestore(&rdev->irq.sw_lock, irqflags); diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c index be2c122..3c2628b 100644 --- a/drivers/gpu/drm/radeon/radeon_kms.c +++ b/drivers/gpu/drm/radeon/radeon_kms.c @@ -57,6 +57,8 @@ int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags) } dev->dev_private = (void *)rdev; + pci_set_master(dev->pdev); + /* update BUS flag */ if (drm_pci_device_is_agp(dev)) { flags |= RADEON_IS_AGP; @@ -169,7 +171,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) value = rdev->accel_working; break; case RADEON_INFO_TILING_CONFIG: - if (rdev->family >= CHIP_CAYMAN) + if (rdev->family >= CHIP_TAHITI) + value = rdev->config.si.tile_config; + else if (rdev->family >= CHIP_CAYMAN) value = rdev->config.cayman.tile_config; else if (rdev->family >= CHIP_CEDAR) value = rdev->config.evergreen.tile_config; @@ -208,7 +212,10 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) value = rdev->clock.spll.reference_freq * 10; break; case RADEON_INFO_NUM_BACKENDS: - if (rdev->family >= CHIP_CAYMAN) + if (rdev->family >= CHIP_TAHITI) + value = rdev->config.si.max_backends_per_se * + rdev->config.si.max_shader_engines; + else if (rdev->family >= CHIP_CAYMAN) value = rdev->config.cayman.max_backends_per_se * rdev->config.cayman.max_shader_engines; else if (rdev->family >= CHIP_CEDAR) @@ -222,7 +229,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) } break; case RADEON_INFO_NUM_TILE_PIPES: - if (rdev->family >= CHIP_CAYMAN) + if (rdev->family >= CHIP_TAHITI) + value = rdev->config.si.max_tile_pipes; + else if (rdev->family >= CHIP_CAYMAN) value = rdev->config.cayman.max_tile_pipes; else if (rdev->family >= CHIP_CEDAR) value = rdev->config.evergreen.max_tile_pipes; @@ -238,7 +247,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) value = 1; break; case RADEON_INFO_BACKEND_MAP: - if (rdev->family >= CHIP_CAYMAN) + if (rdev->family >= CHIP_TAHITI) + value = rdev->config.si.backend_map; + else if (rdev->family >= CHIP_CAYMAN) value = rdev->config.cayman.backend_map; else if (rdev->family >= CHIP_CEDAR) value = rdev->config.evergreen.backend_map; @@ -250,6 +261,33 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) return -EINVAL; } break; + case RADEON_INFO_VA_START: + /* this is where we report if vm is supported or not */ + if (rdev->family < CHIP_CAYMAN) + return -EINVAL; + value = RADEON_VA_RESERVED_SIZE; + break; + case RADEON_INFO_IB_VM_MAX_SIZE: + /* this is where we report if vm is supported or not */ + if (rdev->family < CHIP_CAYMAN) + return -EINVAL; + value = RADEON_IB_VM_MAX_SIZE; + break; + case RADEON_INFO_MAX_PIPES: + if (rdev->family >= CHIP_TAHITI) + value = rdev->config.si.max_pipes_per_simd; + else if (rdev->family >= CHIP_CAYMAN) + value = rdev->config.cayman.max_pipes_per_simd; + else if (rdev->family >= CHIP_CEDAR) + value = rdev->config.evergreen.max_pipes; + else if (rdev->family >= CHIP_RV770) + value = rdev->config.rv770.max_pipes; + else if (rdev->family >= CHIP_R600) + value = rdev->config.r600.max_pipes; + else { + return -EINVAL; + } + break; default: DRM_DEBUG_KMS("Invalid request %d\n", info->request); return -EINVAL; @@ -270,7 +308,6 @@ int radeon_driver_firstopen_kms(struct drm_device *dev) return 0; } - void radeon_driver_lastclose_kms(struct drm_device *dev) { vga_switcheroo_process_delayed_switch(); @@ -278,12 +315,45 @@ void radeon_driver_lastclose_kms(struct drm_device *dev) int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv) { + struct radeon_device *rdev = dev->dev_private; + + file_priv->driver_priv = NULL; + + /* new gpu have virtual address space support */ + if (rdev->family >= CHIP_CAYMAN) { + struct radeon_fpriv *fpriv; + int r; + + fpriv = kzalloc(sizeof(*fpriv), GFP_KERNEL); + if (unlikely(!fpriv)) { + return -ENOMEM; + } + + r = radeon_vm_init(rdev, &fpriv->vm); + if (r) { + radeon_vm_fini(rdev, &fpriv->vm); + kfree(fpriv); + return r; + } + + file_priv->driver_priv = fpriv; + } return 0; } void radeon_driver_postclose_kms(struct drm_device *dev, struct drm_file *file_priv) { + struct radeon_device *rdev = dev->dev_private; + + /* new gpu have virtual address space support */ + if (rdev->family >= CHIP_CAYMAN && file_priv->driver_priv) { + struct radeon_fpriv *fpriv = file_priv->driver_priv; + + radeon_vm_fini(rdev, &fpriv->vm); + kfree(fpriv); + file_priv->driver_priv = NULL; + } } void radeon_driver_preclose_kms(struct drm_device *dev, @@ -451,5 +521,6 @@ struct drm_ioctl_desc radeon_ioctls_kms[] = { DRM_IOCTL_DEF_DRV(RADEON_GEM_SET_TILING, radeon_gem_set_tiling_ioctl, DRM_AUTH|DRM_UNLOCKED), DRM_IOCTL_DEF_DRV(RADEON_GEM_GET_TILING, radeon_gem_get_tiling_ioctl, DRM_AUTH|DRM_UNLOCKED), DRM_IOCTL_DEF_DRV(RADEON_GEM_BUSY, radeon_gem_busy_ioctl, DRM_AUTH|DRM_UNLOCKED), + DRM_IOCTL_DEF_DRV(RADEON_GEM_VA, radeon_gem_va_ioctl, DRM_AUTH|DRM_UNLOCKED), }; int radeon_max_kms_ioctl = DRM_ARRAY_SIZE(radeon_ioctls_kms); diff --git a/drivers/gpu/drm/radeon/radeon_legacy_crtc.c b/drivers/gpu/drm/radeon/radeon_legacy_crtc.c index daadf21..9760e5a 100644 --- a/drivers/gpu/drm/radeon/radeon_legacy_crtc.c +++ b/drivers/gpu/drm/radeon/radeon_legacy_crtc.c @@ -419,7 +419,9 @@ int radeon_crtc_do_set_base(struct drm_crtc *crtc, r = radeon_bo_reserve(rbo, false); if (unlikely(r != 0)) return r; - r = radeon_bo_pin(rbo, RADEON_GEM_DOMAIN_VRAM, &base); + /* Only 27 bit offset for legacy CRTC */ + r = radeon_bo_pin_restricted(rbo, RADEON_GEM_DOMAIN_VRAM, 1 << 27, + &base); if (unlikely(r != 0)) { radeon_bo_unreserve(rbo); return -EINVAL; @@ -437,7 +439,7 @@ int radeon_crtc_do_set_base(struct drm_crtc *crtc, crtc_offset_cntl = 0; - pitch_pixels = target_fb->pitch / (target_fb->bits_per_pixel / 8); + pitch_pixels = target_fb->pitches[0] / (target_fb->bits_per_pixel / 8); crtc_pitch = (((pitch_pixels * target_fb->bits_per_pixel) + ((target_fb->bits_per_pixel * 8) - 1)) / (target_fb->bits_per_pixel * 8)); @@ -1023,9 +1025,11 @@ static int radeon_crtc_mode_set(struct drm_crtc *crtc, static void radeon_crtc_prepare(struct drm_crtc *crtc) { + struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); struct drm_device *dev = crtc->dev; struct drm_crtc *crtci; + radeon_crtc->in_mode_set = true; /* * The hardware wedges sometimes if you reconfigure one CRTC * whilst another is running (see fdo bug #24611). @@ -1036,6 +1040,7 @@ static void radeon_crtc_prepare(struct drm_crtc *crtc) static void radeon_crtc_commit(struct drm_crtc *crtc) { + struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); struct drm_device *dev = crtc->dev; struct drm_crtc *crtci; @@ -1046,6 +1051,7 @@ static void radeon_crtc_commit(struct drm_crtc *crtc) if (crtci->enabled) radeon_crtc_dpms(crtci, DRM_MODE_DPMS_ON); } + radeon_crtc->in_mode_set = false; } static const struct drm_crtc_helper_funcs legacy_helper_funcs = { diff --git a/drivers/gpu/drm/radeon/radeon_legacy_encoders.c b/drivers/gpu/drm/radeon/radeon_legacy_encoders.c index a906803..53b07a3 100644 --- a/drivers/gpu/drm/radeon/radeon_legacy_encoders.c +++ b/drivers/gpu/drm/radeon/radeon_legacy_encoders.c @@ -88,7 +88,7 @@ static void radeon_legacy_lvds_update(struct drm_encoder *encoder, int mode) lvds_pll_cntl = RREG32(RADEON_LVDS_PLL_CNTL); lvds_pll_cntl |= RADEON_LVDS_PLL_EN; WREG32(RADEON_LVDS_PLL_CNTL, lvds_pll_cntl); - udelay(1000); + mdelay(1); lvds_pll_cntl = RREG32(RADEON_LVDS_PLL_CNTL); lvds_pll_cntl &= ~RADEON_LVDS_PLL_RESET; @@ -101,7 +101,7 @@ static void radeon_legacy_lvds_update(struct drm_encoder *encoder, int mode) (backlight_level << RADEON_LVDS_BL_MOD_LEVEL_SHIFT)); if (is_mac) lvds_gen_cntl |= RADEON_LVDS_BL_MOD_EN; - udelay(panel_pwr_delay * 1000); + mdelay(panel_pwr_delay); WREG32(RADEON_LVDS_GEN_CNTL, lvds_gen_cntl); break; case DRM_MODE_DPMS_STANDBY: @@ -118,10 +118,10 @@ static void radeon_legacy_lvds_update(struct drm_encoder *encoder, int mode) WREG32(RADEON_LVDS_GEN_CNTL, lvds_gen_cntl); lvds_gen_cntl &= ~(RADEON_LVDS_ON | RADEON_LVDS_BLON | RADEON_LVDS_EN | RADEON_LVDS_DIGON); } - udelay(panel_pwr_delay * 1000); + mdelay(panel_pwr_delay); WREG32(RADEON_LVDS_GEN_CNTL, lvds_gen_cntl); WREG32_PLL(RADEON_PIXCLKS_CNTL, pixclks_cntl); - udelay(panel_pwr_delay * 1000); + mdelay(panel_pwr_delay); break; } @@ -665,7 +665,7 @@ static enum drm_connector_status radeon_legacy_primary_dac_detect(struct drm_enc WREG32(RADEON_DAC_MACRO_CNTL, tmp); - udelay(2000); + mdelay(2); if (RREG32(RADEON_DAC_CNTL) & RADEON_DAC_CMP_OUTPUT) found = connector_status_connected; @@ -1504,7 +1504,7 @@ static enum drm_connector_status radeon_legacy_tv_dac_detect(struct drm_encoder tmp = dac_cntl2 | RADEON_DAC2_DAC2_CLK_SEL | RADEON_DAC2_CMP_EN; WREG32(RADEON_DAC_CNTL2, tmp); - udelay(10000); + mdelay(10); if (ASIC_IS_R300(rdev)) { if (RREG32(RADEON_DAC_CNTL2) & RADEON_DAC2_CMP_OUT_B) diff --git a/drivers/gpu/drm/radeon/radeon_mode.h b/drivers/gpu/drm/radeon/radeon_mode.h index bb42df4..dabfefd 100644 --- a/drivers/gpu/drm/radeon/radeon_mode.h +++ b/drivers/gpu/drm/radeon/radeon_mode.h @@ -210,8 +210,6 @@ enum radeon_connector_table { CT_RN50_POWER, CT_MAC_X800, CT_MAC_G5_9600, - CT_SAM440EP, - CT_MAC_G4_SILVER }; enum radeon_dvo_chip { @@ -268,6 +266,7 @@ struct radeon_crtc { u16 lut_r[256], lut_g[256], lut_b[256]; bool enabled; bool can_tile; + bool in_mode_set; uint32_t crtc_offset; struct drm_gem_object *cursor_bo; uint64_t cursor_addr; @@ -469,6 +468,10 @@ radeon_atombios_get_tv_info(struct radeon_device *rdev); extern struct drm_connector * radeon_get_connector_for_encoder(struct drm_encoder *encoder); +extern struct drm_connector * +radeon_get_connector_for_encoder_init(struct drm_encoder *encoder); +extern bool radeon_dig_monitor_is_duallink(struct drm_encoder *encoder, + u32 pixel_clock); extern u16 radeon_encoder_get_dp_bridge_encoder_id(struct drm_encoder *encoder); extern u16 radeon_connector_encoder_get_dp_bridge_encoder_id(struct drm_connector *connector); @@ -489,6 +492,7 @@ extern int radeon_dp_get_panel_mode(struct drm_encoder *encoder, struct drm_connector *connector); extern void atombios_dig_encoder_setup(struct drm_encoder *encoder, int action, int panel_mode); extern void radeon_atom_encoder_init(struct radeon_device *rdev); +extern void radeon_atom_disp_eng_pll_init(struct radeon_device *rdev); extern void atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t lane_num, uint8_t lane_set); @@ -646,9 +650,9 @@ extern void radeon_crtc_fb_gamma_set(struct drm_crtc *crtc, u16 red, u16 green, u16 blue, int regno); extern void radeon_crtc_fb_gamma_get(struct drm_crtc *crtc, u16 *red, u16 *green, u16 *blue, int regno); -void radeon_framebuffer_init(struct drm_device *dev, +int radeon_framebuffer_init(struct drm_device *dev, struct radeon_framebuffer *rfb, - struct drm_mode_fb_cmd *mode_cmd, + struct drm_mode_fb_cmd2 *mode_cmd, struct drm_gem_object *obj); int radeonfb_remove(struct drm_device *dev, struct drm_framebuffer *fb); diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c index f3ae607..df6a4db 100644 --- a/drivers/gpu/drm/radeon/radeon_object.c +++ b/drivers/gpu/drm/radeon/radeon_object.c @@ -46,6 +46,20 @@ static void radeon_bo_clear_surface_reg(struct radeon_bo *bo); * function are calling it. */ +void radeon_bo_clear_va(struct radeon_bo *bo) +{ + struct radeon_bo_va *bo_va, *tmp; + + list_for_each_entry_safe(bo_va, tmp, &bo->va, bo_list) { + /* remove from all vm address space */ + mutex_lock(&bo_va->vm->mutex); + list_del(&bo_va->vm_list); + mutex_unlock(&bo_va->vm->mutex); + list_del(&bo_va->bo_list); + kfree(bo_va); + } +} + static void radeon_ttm_bo_destroy(struct ttm_buffer_object *tbo) { struct radeon_bo *bo; @@ -55,6 +69,7 @@ static void radeon_ttm_bo_destroy(struct ttm_buffer_object *tbo) list_del_init(&bo->list); mutex_unlock(&bo->rdev->gem.mutex); radeon_bo_clear_surface_reg(bo); + radeon_bo_clear_va(bo); drm_gem_object_release(&bo->gem_base); kfree(bo); } @@ -95,6 +110,7 @@ int radeon_bo_create(struct radeon_device *rdev, enum ttm_bo_type type; unsigned long page_align = roundup(byte_align, PAGE_SIZE) >> PAGE_SHIFT; unsigned long max_size = 0; + size_t acc_size; int r; size = ALIGN(size, PAGE_SIZE); @@ -117,6 +133,9 @@ int radeon_bo_create(struct radeon_device *rdev, return -ENOMEM; } + acc_size = ttm_bo_dma_acc_size(&rdev->mman.bdev, size, + sizeof(struct radeon_bo)); + retry: bo = kzalloc(sizeof(struct radeon_bo), GFP_KERNEL); if (bo == NULL) @@ -130,12 +149,13 @@ retry: bo->gem_base.driver_private = NULL; bo->surface_reg = -1; INIT_LIST_HEAD(&bo->list); + INIT_LIST_HEAD(&bo->va); radeon_ttm_placement_from_domain(bo, domain); /* Kernel allocation are uninterruptible */ mutex_lock(&rdev->vram_mutex); r = ttm_bo_init(&rdev->mman.bdev, &bo->tbo, size, type, - &bo->placement, page_align, 0, !kernel, NULL, size, - &radeon_ttm_bo_destroy); + &bo->placement, page_align, 0, !kernel, NULL, + acc_size, &radeon_ttm_bo_destroy); mutex_unlock(&rdev->vram_mutex); if (unlikely(r != 0)) { if (r != -ERESTARTSYS) { @@ -213,7 +233,18 @@ int radeon_bo_pin_restricted(struct radeon_bo *bo, u32 domain, u64 max_offset, bo->pin_count++; if (gpu_addr) *gpu_addr = radeon_bo_gpu_offset(bo); - WARN_ON_ONCE(max_offset != 0); + + if (max_offset != 0) { + u64 domain_start; + + if (domain == RADEON_GEM_DOMAIN_VRAM) + domain_start = bo->rdev->mc.vram_start; + else + domain_start = bo->rdev->mc.gtt_start; + WARN_ON_ONCE(max_offset < + (radeon_bo_gpu_offset(bo) - domain_start)); + } + return 0; } radeon_ttm_placement_from_domain(bo, domain); @@ -441,8 +472,54 @@ static void radeon_bo_clear_surface_reg(struct radeon_bo *bo) int radeon_bo_set_tiling_flags(struct radeon_bo *bo, uint32_t tiling_flags, uint32_t pitch) { + struct radeon_device *rdev = bo->rdev; int r; + if (rdev->family >= CHIP_CEDAR) { + unsigned bankw, bankh, mtaspect, tilesplit, stilesplit; + + bankw = (tiling_flags >> RADEON_TILING_EG_BANKW_SHIFT) & RADEON_TILING_EG_BANKW_MASK; + bankh = (tiling_flags >> RADEON_TILING_EG_BANKH_SHIFT) & RADEON_TILING_EG_BANKH_MASK; + mtaspect = (tiling_flags >> RADEON_TILING_EG_MACRO_TILE_ASPECT_SHIFT) & RADEON_TILING_EG_MACRO_TILE_ASPECT_MASK; + tilesplit = (tiling_flags >> RADEON_TILING_EG_TILE_SPLIT_SHIFT) & RADEON_TILING_EG_TILE_SPLIT_MASK; + stilesplit = (tiling_flags >> RADEON_TILING_EG_STENCIL_TILE_SPLIT_SHIFT) & RADEON_TILING_EG_STENCIL_TILE_SPLIT_MASK; + switch (bankw) { + case 0: + case 1: + case 2: + case 4: + case 8: + break; + default: + return -EINVAL; + } + switch (bankh) { + case 0: + case 1: + case 2: + case 4: + case 8: + break; + default: + return -EINVAL; + } + switch (mtaspect) { + case 0: + case 1: + case 2: + case 4: + case 8: + break; + default: + return -EINVAL; + } + if (tilesplit > 6) { + return -EINVAL; + } + if (stilesplit > 6) { + return -EINVAL; + } + } r = radeon_bo_reserve(bo, false); if (unlikely(r != 0)) return r; @@ -499,6 +576,7 @@ void radeon_bo_move_notify(struct ttm_buffer_object *bo, return; rbo = container_of(bo, struct radeon_bo, tbo); radeon_bo_check_tiling(rbo, 0, 1); + radeon_vm_bo_invalidate(rbo->rdev, rbo); } int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo) @@ -572,3 +650,16 @@ int radeon_bo_reserve(struct radeon_bo *bo, bool no_wait) } return 0; } + +/* object have to be reserved */ +struct radeon_bo_va *radeon_bo_va(struct radeon_bo *rbo, struct radeon_vm *vm) +{ + struct radeon_bo_va *bo_va; + + list_for_each_entry(bo_va, &rbo->va, bo_list) { + if (bo_va->vm == vm) { + return bo_va; + } + } + return NULL; +} diff --git a/drivers/gpu/drm/radeon/radeon_object.h b/drivers/gpu/drm/radeon/radeon_object.h index fb3f433..f9104be 100644 --- a/drivers/gpu/drm/radeon/radeon_object.h +++ b/drivers/gpu/drm/radeon/radeon_object.h @@ -83,6 +83,16 @@ static inline bool radeon_bo_is_reserved(struct radeon_bo *bo) return !!atomic_read(&bo->tbo.reserved); } +static inline unsigned radeon_bo_ngpu_pages(struct radeon_bo *bo) +{ + return (bo->tbo.num_pages << PAGE_SHIFT) / RADEON_GPU_PAGE_SIZE; +} + +static inline unsigned radeon_bo_gpu_page_alignment(struct radeon_bo *bo) +{ + return (bo->tbo.mem.page_alignment << PAGE_SHIFT) / RADEON_GPU_PAGE_SIZE; +} + /** * radeon_bo_mmap_offset - return mmap offset of bo * @bo: radeon object for which we query the offset @@ -130,4 +140,26 @@ extern void radeon_bo_move_notify(struct ttm_buffer_object *bo, struct ttm_mem_reg *mem); extern int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo); extern int radeon_bo_get_surface_reg(struct radeon_bo *bo); +extern struct radeon_bo_va *radeon_bo_va(struct radeon_bo *rbo, + struct radeon_vm *vm); + +/* + * sub allocation + */ +extern int radeon_sa_bo_manager_init(struct radeon_device *rdev, + struct radeon_sa_manager *sa_manager, + unsigned size, u32 domain); +extern void radeon_sa_bo_manager_fini(struct radeon_device *rdev, + struct radeon_sa_manager *sa_manager); +extern int radeon_sa_bo_manager_start(struct radeon_device *rdev, + struct radeon_sa_manager *sa_manager); +extern int radeon_sa_bo_manager_suspend(struct radeon_device *rdev, + struct radeon_sa_manager *sa_manager); +extern int radeon_sa_bo_new(struct radeon_device *rdev, + struct radeon_sa_manager *sa_manager, + struct radeon_sa_bo *sa_bo, + unsigned size, unsigned align); +extern void radeon_sa_bo_free(struct radeon_device *rdev, + struct radeon_sa_bo *sa_bo); + #endif diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c index ebd6c51..b8459bd 100644 --- a/drivers/gpu/drm/radeon/radeon_pm.c +++ b/drivers/gpu/drm/radeon/radeon_pm.c @@ -221,7 +221,7 @@ static void radeon_set_power_state(struct radeon_device *rdev) } /* set memory clock */ - if (rdev->asic->set_memory_clock && (mclk != rdev->pm.current_mclk)) { + if (rdev->asic->pm.set_memory_clock && (mclk != rdev->pm.current_mclk)) { radeon_pm_debug_check_in_vbl(rdev, false); radeon_set_memory_clock(rdev, mclk); radeon_pm_debug_check_in_vbl(rdev, true); @@ -252,7 +252,10 @@ static void radeon_pm_set_clocks(struct radeon_device *rdev) mutex_lock(&rdev->ddev->struct_mutex); mutex_lock(&rdev->vram_mutex); - mutex_lock(&rdev->cp.mutex); + for (i = 0; i < RADEON_NUM_RINGS; ++i) { + if (rdev->ring[i].ring_obj) + mutex_lock(&rdev->ring[i].mutex); + } /* gui idle int has issues on older chips it seems */ if (rdev->family >= CHIP_R600) { @@ -268,12 +271,13 @@ static void radeon_pm_set_clocks(struct radeon_device *rdev) radeon_irq_set(rdev); } } else { - if (rdev->cp.ready) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; + if (ring->ready) { struct radeon_fence *fence; - radeon_ring_alloc(rdev, 64); - radeon_fence_create(rdev, &fence); + radeon_ring_alloc(rdev, ring, 64); + radeon_fence_create(rdev, &fence, radeon_ring_index(rdev, ring)); radeon_fence_emit(rdev, fence); - radeon_ring_commit(rdev); + radeon_ring_commit(rdev, ring); radeon_fence_wait(fence, false); radeon_fence_unref(&fence); } @@ -307,7 +311,10 @@ static void radeon_pm_set_clocks(struct radeon_device *rdev) rdev->pm.dynpm_planned_action = DYNPM_ACTION_NONE; - mutex_unlock(&rdev->cp.mutex); + for (i = 0; i < RADEON_NUM_RINGS; ++i) { + if (rdev->ring[i].ring_obj) + mutex_unlock(&rdev->ring[i].mutex); + } mutex_unlock(&rdev->vram_mutex); mutex_unlock(&rdev->ddev->struct_mutex); } @@ -467,6 +474,9 @@ static ssize_t radeon_hwmon_show_temp(struct device *dev, case THERMAL_TYPE_SUMO: temp = sumo_get_temp(rdev); break; + case THERMAL_TYPE_SI: + temp = si_get_temp(rdev); + break; default: temp = 0; break; @@ -507,6 +517,10 @@ static int radeon_hwmon_init(struct radeon_device *rdev) case THERMAL_TYPE_EVERGREEN: case THERMAL_TYPE_NI: case THERMAL_TYPE_SUMO: + case THERMAL_TYPE_SI: + /* No support for TN yet */ + if (rdev->family == CHIP_ARUBA) + return err; rdev->pm.int_hwmon_dev = hwmon_device_register(rdev->dev); if (IS_ERR(rdev->pm.int_hwmon_dev)) { err = PTR_ERR(rdev->pm.int_hwmon_dev); @@ -799,19 +813,14 @@ static void radeon_dynpm_idle_work_handler(struct work_struct *work) resched = ttm_bo_lock_delayed_workqueue(&rdev->mman.bdev); mutex_lock(&rdev->pm.mutex); if (rdev->pm.dynpm_state == DYNPM_STATE_ACTIVE) { - unsigned long irq_flags; int not_processed = 0; + int i; - read_lock_irqsave(&rdev->fence_drv.lock, irq_flags); - if (!list_empty(&rdev->fence_drv.emited)) { - struct list_head *ptr; - list_for_each(ptr, &rdev->fence_drv.emited) { - /* count up to 3, that's enought info */ - if (++not_processed >= 3) - break; - } + for (i = 0; i < RADEON_NUM_RINGS; ++i) { + not_processed += radeon_fence_count_emitted(rdev, i); + if (not_processed >= 3) + break; } - read_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags); if (not_processed >= 3) { /* should upclock */ if (rdev->pm.dynpm_planned_action == DYNPM_ACTION_DOWNCLOCK) { @@ -865,11 +874,11 @@ static int radeon_debugfs_pm_info(struct seq_file *m, void *data) seq_printf(m, "default engine clock: %u0 kHz\n", rdev->pm.default_sclk); seq_printf(m, "current engine clock: %u0 kHz\n", radeon_get_engine_clock(rdev)); seq_printf(m, "default memory clock: %u0 kHz\n", rdev->pm.default_mclk); - if (rdev->asic->get_memory_clock) + if (rdev->asic->pm.get_memory_clock) seq_printf(m, "current memory clock: %u0 kHz\n", radeon_get_memory_clock(rdev)); if (rdev->pm.current_vddc) seq_printf(m, "voltage: %u mV\n", rdev->pm.current_vddc); - if (rdev->asic->get_pcie_lanes) + if (rdev->asic->pm.get_pcie_lanes) seq_printf(m, "PCIE lanes: %d\n", radeon_get_pcie_lanes(rdev)); return 0; diff --git a/drivers/gpu/drm/radeon/radeon_reg.h b/drivers/gpu/drm/radeon/radeon_reg.h index b4ce864..5d8f735 100644 --- a/drivers/gpu/drm/radeon/radeon_reg.h +++ b/drivers/gpu/drm/radeon/radeon_reg.h @@ -56,6 +56,7 @@ #include "r600_reg.h" #include "evergreen_reg.h" #include "ni_reg.h" +#include "si_reg.h" #define RADEON_MC_AGP_LOCATION 0x014c #define RADEON_MC_AGP_START_MASK 0x0000FFFF @@ -539,9 +540,11 @@ #define RADEON_CRTC2_PITCH 0x032c #define RADEON_CRTC_STATUS 0x005c +# define RADEON_CRTC_VBLANK_CUR (1 << 0) # define RADEON_CRTC_VBLANK_SAVE (1 << 1) # define RADEON_CRTC_VBLANK_SAVE_CLEAR (1 << 1) #define RADEON_CRTC2_STATUS 0x03fc +# define RADEON_CRTC2_VBLANK_CUR (1 << 0) # define RADEON_CRTC2_VBLANK_SAVE (1 << 1) # define RADEON_CRTC2_VBLANK_SAVE_CLEAR (1 << 1) #define RADEON_CRTC_V_SYNC_STRT_WID 0x020c diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c index 65be5e8..33eff8b 100644 --- a/drivers/gpu/drm/radeon/radeon_ring.c +++ b/drivers/gpu/drm/radeon/radeon_ring.c @@ -34,6 +34,7 @@ #include "atom.h" int radeon_debugfs_ib_init(struct radeon_device *rdev); +int radeon_debugfs_ring_init(struct radeon_device *rdev); u32 radeon_get_ib_value(struct radeon_cs_parser *p, int idx) { @@ -60,105 +61,107 @@ u32 radeon_get_ib_value(struct radeon_cs_parser *p, int idx) return idx_value; } -void radeon_ring_write(struct radeon_device *rdev, uint32_t v) +void radeon_ring_write(struct radeon_ring *ring, uint32_t v) { #if DRM_DEBUG_CODE - if (rdev->cp.count_dw <= 0) { + if (ring->count_dw <= 0) { DRM_ERROR("radeon: writting more dword to ring than expected !\n"); } #endif - rdev->cp.ring[rdev->cp.wptr++] = v; - rdev->cp.wptr &= rdev->cp.ptr_mask; - rdev->cp.count_dw--; - rdev->cp.ring_free_dw--; + ring->ring[ring->wptr++] = v; + ring->wptr &= ring->ptr_mask; + ring->count_dw--; + ring->ring_free_dw--; } -void radeon_ib_bogus_cleanup(struct radeon_device *rdev) -{ - struct radeon_ib *ib, *n; - - list_for_each_entry_safe(ib, n, &rdev->ib_pool.bogus_ib, list) { - list_del(&ib->list); - vfree(ib->ptr); - kfree(ib); - } -} - -void radeon_ib_bogus_add(struct radeon_device *rdev, struct radeon_ib *ib) +/* + * IB. + */ +bool radeon_ib_try_free(struct radeon_device *rdev, struct radeon_ib *ib) { - struct radeon_ib *bib; - - bib = kmalloc(sizeof(*bib), GFP_KERNEL); - if (bib == NULL) - return; - bib->ptr = vmalloc(ib->length_dw * 4); - if (bib->ptr == NULL) { - kfree(bib); - return; + bool done = false; + + /* only free ib which have been emited */ + if (ib->fence && ib->fence->emitted) { + if (radeon_fence_signaled(ib->fence)) { + radeon_fence_unref(&ib->fence); + radeon_sa_bo_free(rdev, &ib->sa_bo); + done = true; + } } - memcpy(bib->ptr, ib->ptr, ib->length_dw * 4); - bib->length_dw = ib->length_dw; - mutex_lock(&rdev->ib_pool.mutex); - list_add_tail(&bib->list, &rdev->ib_pool.bogus_ib); - mutex_unlock(&rdev->ib_pool.mutex); + return done; } -/* - * IB. - */ -int radeon_ib_get(struct radeon_device *rdev, struct radeon_ib **ib) +int radeon_ib_get(struct radeon_device *rdev, int ring, + struct radeon_ib **ib, unsigned size) { struct radeon_fence *fence; - struct radeon_ib *nib; - int r = 0, i, c; + unsigned cretry = 0; + int r = 0, i, idx; *ib = NULL; - r = radeon_fence_create(rdev, &fence); + /* align size on 256 bytes */ + size = ALIGN(size, 256); + + r = radeon_fence_create(rdev, &fence, ring); if (r) { dev_err(rdev->dev, "failed to create fence for new IB\n"); return r; } - mutex_lock(&rdev->ib_pool.mutex); - for (i = rdev->ib_pool.head_id, c = 0, nib = NULL; c < RADEON_IB_POOL_SIZE; c++, i++) { - i &= (RADEON_IB_POOL_SIZE - 1); - if (rdev->ib_pool.ibs[i].free) { - nib = &rdev->ib_pool.ibs[i]; - break; - } - } - if (nib == NULL) { - /* This should never happen, it means we allocated all - * IB and haven't scheduled one yet, return EBUSY to - * userspace hoping that on ioctl recall we get better - * luck - */ - dev_err(rdev->dev, "no free indirect buffer !\n"); - mutex_unlock(&rdev->ib_pool.mutex); + + radeon_mutex_lock(&rdev->ib_pool.mutex); + idx = rdev->ib_pool.head_id; +retry: + if (cretry > 5) { + dev_err(rdev->dev, "failed to get an ib after 5 retry\n"); + radeon_mutex_unlock(&rdev->ib_pool.mutex); radeon_fence_unref(&fence); - return -EBUSY; + return -ENOMEM; } - rdev->ib_pool.head_id = (nib->idx + 1) & (RADEON_IB_POOL_SIZE - 1); - nib->free = false; - if (nib->fence) { - mutex_unlock(&rdev->ib_pool.mutex); - r = radeon_fence_wait(nib->fence, false); - if (r) { - dev_err(rdev->dev, "error waiting fence of IB(%u:0x%016lX:%u)\n", - nib->idx, (unsigned long)nib->gpu_addr, nib->length_dw); - mutex_lock(&rdev->ib_pool.mutex); - nib->free = true; - mutex_unlock(&rdev->ib_pool.mutex); - radeon_fence_unref(&fence); - return r; + cretry++; + for (i = 0; i < RADEON_IB_POOL_SIZE; i++) { + radeon_ib_try_free(rdev, &rdev->ib_pool.ibs[idx]); + if (rdev->ib_pool.ibs[idx].fence == NULL) { + r = radeon_sa_bo_new(rdev, &rdev->ib_pool.sa_manager, + &rdev->ib_pool.ibs[idx].sa_bo, + size, 256); + if (!r) { + *ib = &rdev->ib_pool.ibs[idx]; + (*ib)->ptr = rdev->ib_pool.sa_manager.cpu_ptr; + (*ib)->ptr += ((*ib)->sa_bo.offset >> 2); + (*ib)->gpu_addr = rdev->ib_pool.sa_manager.gpu_addr; + (*ib)->gpu_addr += (*ib)->sa_bo.offset; + (*ib)->fence = fence; + (*ib)->vm_id = 0; + (*ib)->is_const_ib = false; + /* ib are most likely to be allocated in a ring fashion + * thus rdev->ib_pool.head_id should be the id of the + * oldest ib + */ + rdev->ib_pool.head_id = (1 + idx); + rdev->ib_pool.head_id &= (RADEON_IB_POOL_SIZE - 1); + radeon_mutex_unlock(&rdev->ib_pool.mutex); + return 0; + } } - mutex_lock(&rdev->ib_pool.mutex); + idx = (idx + 1) & (RADEON_IB_POOL_SIZE - 1); } - radeon_fence_unref(&nib->fence); - nib->fence = fence; - nib->length_dw = 0; - mutex_unlock(&rdev->ib_pool.mutex); - *ib = nib; - return 0; + /* this should be rare event, ie all ib scheduled none signaled yet. + */ + for (i = 0; i < RADEON_IB_POOL_SIZE; i++) { + if (rdev->ib_pool.ibs[idx].fence && rdev->ib_pool.ibs[idx].fence->emitted) { + r = radeon_fence_wait(rdev->ib_pool.ibs[idx].fence, false); + if (!r) { + goto retry; + } + /* an error happened */ + break; + } + idx = (idx + 1) & (RADEON_IB_POOL_SIZE - 1); + } + radeon_mutex_unlock(&rdev->ib_pool.mutex); + radeon_fence_unref(&fence); + return r; } void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib **ib) @@ -169,250 +172,261 @@ void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib **ib) if (tmp == NULL) { return; } - if (!tmp->fence->emited) + radeon_mutex_lock(&rdev->ib_pool.mutex); + if (tmp->fence && !tmp->fence->emitted) { + radeon_sa_bo_free(rdev, &tmp->sa_bo); radeon_fence_unref(&tmp->fence); - mutex_lock(&rdev->ib_pool.mutex); - tmp->free = true; - mutex_unlock(&rdev->ib_pool.mutex); + } + radeon_mutex_unlock(&rdev->ib_pool.mutex); } int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib) { + struct radeon_ring *ring = &rdev->ring[ib->fence->ring]; int r = 0; - if (!ib->length_dw || !rdev->cp.ready) { + if (!ib->length_dw || !ring->ready) { /* TODO: Nothings in the ib we should report. */ DRM_ERROR("radeon: couldn't schedule IB(%u).\n", ib->idx); return -EINVAL; } /* 64 dwords should be enough for fence too */ - r = radeon_ring_lock(rdev, 64); + r = radeon_ring_lock(rdev, ring, 64); if (r) { DRM_ERROR("radeon: scheduling IB failed (%d).\n", r); return r; } - radeon_ring_ib_execute(rdev, ib); + radeon_ring_ib_execute(rdev, ib->fence->ring, ib); radeon_fence_emit(rdev, ib->fence); - mutex_lock(&rdev->ib_pool.mutex); - /* once scheduled IB is considered free and protected by the fence */ - ib->free = true; - mutex_unlock(&rdev->ib_pool.mutex); - radeon_ring_unlock_commit(rdev); + radeon_ring_unlock_commit(rdev, ring); return 0; } int radeon_ib_pool_init(struct radeon_device *rdev) { - void *ptr; - uint64_t gpu_addr; - int i; - int r = 0; + struct radeon_sa_manager tmp; + int i, r; - if (rdev->ib_pool.robj) - return 0; - INIT_LIST_HEAD(&rdev->ib_pool.bogus_ib); - /* Allocate 1M object buffer */ - r = radeon_bo_create(rdev, RADEON_IB_POOL_SIZE*64*1024, - PAGE_SIZE, true, RADEON_GEM_DOMAIN_GTT, - &rdev->ib_pool.robj); + r = radeon_sa_bo_manager_init(rdev, &tmp, + RADEON_IB_POOL_SIZE*64*1024, + RADEON_GEM_DOMAIN_GTT); if (r) { - DRM_ERROR("radeon: failed to ib pool (%d).\n", r); return r; } - r = radeon_bo_reserve(rdev->ib_pool.robj, false); - if (unlikely(r != 0)) - return r; - r = radeon_bo_pin(rdev->ib_pool.robj, RADEON_GEM_DOMAIN_GTT, &gpu_addr); - if (r) { - radeon_bo_unreserve(rdev->ib_pool.robj); - DRM_ERROR("radeon: failed to pin ib pool (%d).\n", r); - return r; - } - r = radeon_bo_kmap(rdev->ib_pool.robj, &ptr); - radeon_bo_unreserve(rdev->ib_pool.robj); - if (r) { - DRM_ERROR("radeon: failed to map ib pool (%d).\n", r); - return r; + + radeon_mutex_lock(&rdev->ib_pool.mutex); + if (rdev->ib_pool.ready) { + radeon_mutex_unlock(&rdev->ib_pool.mutex); + radeon_sa_bo_manager_fini(rdev, &tmp); + return 0; } - for (i = 0; i < RADEON_IB_POOL_SIZE; i++) { - unsigned offset; - offset = i * 64 * 1024; - rdev->ib_pool.ibs[i].gpu_addr = gpu_addr + offset; - rdev->ib_pool.ibs[i].ptr = ptr + offset; + rdev->ib_pool.sa_manager = tmp; + INIT_LIST_HEAD(&rdev->ib_pool.sa_manager.sa_bo); + for (i = 0; i < RADEON_IB_POOL_SIZE; i++) { + rdev->ib_pool.ibs[i].fence = NULL; rdev->ib_pool.ibs[i].idx = i; rdev->ib_pool.ibs[i].length_dw = 0; - rdev->ib_pool.ibs[i].free = true; + INIT_LIST_HEAD(&rdev->ib_pool.ibs[i].sa_bo.list); } rdev->ib_pool.head_id = 0; rdev->ib_pool.ready = true; DRM_INFO("radeon: ib pool ready.\n"); + if (radeon_debugfs_ib_init(rdev)) { DRM_ERROR("Failed to register debugfs file for IB !\n"); } - return r; + if (radeon_debugfs_ring_init(rdev)) { + DRM_ERROR("Failed to register debugfs file for rings !\n"); + } + radeon_mutex_unlock(&rdev->ib_pool.mutex); + return 0; } void radeon_ib_pool_fini(struct radeon_device *rdev) { - int r; - struct radeon_bo *robj; + unsigned i; - if (!rdev->ib_pool.ready) { - return; - } - mutex_lock(&rdev->ib_pool.mutex); - radeon_ib_bogus_cleanup(rdev); - robj = rdev->ib_pool.robj; - rdev->ib_pool.robj = NULL; - mutex_unlock(&rdev->ib_pool.mutex); - - if (robj) { - r = radeon_bo_reserve(robj, false); - if (likely(r == 0)) { - radeon_bo_kunmap(robj); - radeon_bo_unpin(robj); - radeon_bo_unreserve(robj); + radeon_mutex_lock(&rdev->ib_pool.mutex); + if (rdev->ib_pool.ready) { + for (i = 0; i < RADEON_IB_POOL_SIZE; i++) { + radeon_sa_bo_free(rdev, &rdev->ib_pool.ibs[i].sa_bo); + radeon_fence_unref(&rdev->ib_pool.ibs[i].fence); } - radeon_bo_unref(&robj); + radeon_sa_bo_manager_fini(rdev, &rdev->ib_pool.sa_manager); + rdev->ib_pool.ready = false; } + radeon_mutex_unlock(&rdev->ib_pool.mutex); +} + +int radeon_ib_pool_start(struct radeon_device *rdev) +{ + return radeon_sa_bo_manager_start(rdev, &rdev->ib_pool.sa_manager); } +int radeon_ib_pool_suspend(struct radeon_device *rdev) +{ + return radeon_sa_bo_manager_suspend(rdev, &rdev->ib_pool.sa_manager); +} /* * Ring. */ -void radeon_ring_free_size(struct radeon_device *rdev) +int radeon_ring_index(struct radeon_device *rdev, struct radeon_ring *ring) { - if (rdev->wb.enabled) - rdev->cp.rptr = le32_to_cpu(rdev->wb.wb[RADEON_WB_CP_RPTR_OFFSET/4]); - else { - if (rdev->family >= CHIP_R600) - rdev->cp.rptr = RREG32(R600_CP_RB_RPTR); - else - rdev->cp.rptr = RREG32(RADEON_CP_RB_RPTR); + /* r1xx-r5xx only has CP ring */ + if (rdev->family < CHIP_R600) + return RADEON_RING_TYPE_GFX_INDEX; + + if (rdev->family >= CHIP_CAYMAN) { + if (ring == &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]) + return CAYMAN_RING_TYPE_CP1_INDEX; + else if (ring == &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]) + return CAYMAN_RING_TYPE_CP2_INDEX; } + return RADEON_RING_TYPE_GFX_INDEX; +} + +void radeon_ring_free_size(struct radeon_device *rdev, struct radeon_ring *ring) +{ + u32 rptr; + + if (rdev->wb.enabled) + rptr = le32_to_cpu(rdev->wb.wb[ring->rptr_offs/4]); + else + rptr = RREG32(ring->rptr_reg); + ring->rptr = (rptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift; /* This works because ring_size is a power of 2 */ - rdev->cp.ring_free_dw = (rdev->cp.rptr + (rdev->cp.ring_size / 4)); - rdev->cp.ring_free_dw -= rdev->cp.wptr; - rdev->cp.ring_free_dw &= rdev->cp.ptr_mask; - if (!rdev->cp.ring_free_dw) { - rdev->cp.ring_free_dw = rdev->cp.ring_size / 4; + ring->ring_free_dw = (ring->rptr + (ring->ring_size / 4)); + ring->ring_free_dw -= ring->wptr; + ring->ring_free_dw &= ring->ptr_mask; + if (!ring->ring_free_dw) { + ring->ring_free_dw = ring->ring_size / 4; } } -int radeon_ring_alloc(struct radeon_device *rdev, unsigned ndw) + +int radeon_ring_alloc(struct radeon_device *rdev, struct radeon_ring *ring, unsigned ndw) { int r; /* make sure we aren't trying to allocate more space than there is on the ring */ - if (ndw > (rdev->cp.ring_size / 4)) + if (ndw > (ring->ring_size / 4)) return -ENOMEM; /* Align requested size with padding so unlock_commit can * pad safely */ - ndw = (ndw + rdev->cp.align_mask) & ~rdev->cp.align_mask; - while (ndw > (rdev->cp.ring_free_dw - 1)) { - radeon_ring_free_size(rdev); - if (ndw < rdev->cp.ring_free_dw) { + ndw = (ndw + ring->align_mask) & ~ring->align_mask; + while (ndw > (ring->ring_free_dw - 1)) { + radeon_ring_free_size(rdev, ring); + if (ndw < ring->ring_free_dw) { break; } - r = radeon_fence_wait_next(rdev); + r = radeon_fence_wait_next(rdev, radeon_ring_index(rdev, ring)); if (r) return r; } - rdev->cp.count_dw = ndw; - rdev->cp.wptr_old = rdev->cp.wptr; + ring->count_dw = ndw; + ring->wptr_old = ring->wptr; return 0; } -int radeon_ring_lock(struct radeon_device *rdev, unsigned ndw) +int radeon_ring_lock(struct radeon_device *rdev, struct radeon_ring *ring, unsigned ndw) { int r; - mutex_lock(&rdev->cp.mutex); - r = radeon_ring_alloc(rdev, ndw); + mutex_lock(&ring->mutex); + r = radeon_ring_alloc(rdev, ring, ndw); if (r) { - mutex_unlock(&rdev->cp.mutex); + mutex_unlock(&ring->mutex); return r; } return 0; } -void radeon_ring_commit(struct radeon_device *rdev) +void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *ring) { unsigned count_dw_pad; unsigned i; /* We pad to match fetch size */ - count_dw_pad = (rdev->cp.align_mask + 1) - - (rdev->cp.wptr & rdev->cp.align_mask); + count_dw_pad = (ring->align_mask + 1) - + (ring->wptr & ring->align_mask); for (i = 0; i < count_dw_pad; i++) { - radeon_ring_write(rdev, 2 << 30); + radeon_ring_write(ring, ring->nop); } DRM_MEMORYBARRIER(); - radeon_cp_commit(rdev); + WREG32(ring->wptr_reg, (ring->wptr << ring->ptr_reg_shift) & ring->ptr_reg_mask); + (void)RREG32(ring->wptr_reg); } -void radeon_ring_unlock_commit(struct radeon_device *rdev) +void radeon_ring_unlock_commit(struct radeon_device *rdev, struct radeon_ring *ring) { - radeon_ring_commit(rdev); - mutex_unlock(&rdev->cp.mutex); + radeon_ring_commit(rdev, ring); + mutex_unlock(&ring->mutex); } -void radeon_ring_unlock_undo(struct radeon_device *rdev) +void radeon_ring_unlock_undo(struct radeon_device *rdev, struct radeon_ring *ring) { - rdev->cp.wptr = rdev->cp.wptr_old; - mutex_unlock(&rdev->cp.mutex); + ring->wptr = ring->wptr_old; + mutex_unlock(&ring->mutex); } -int radeon_ring_init(struct radeon_device *rdev, unsigned ring_size) +int radeon_ring_init(struct radeon_device *rdev, struct radeon_ring *ring, unsigned ring_size, + unsigned rptr_offs, unsigned rptr_reg, unsigned wptr_reg, + u32 ptr_reg_shift, u32 ptr_reg_mask, u32 nop) { int r; - rdev->cp.ring_size = ring_size; + ring->ring_size = ring_size; + ring->rptr_offs = rptr_offs; + ring->rptr_reg = rptr_reg; + ring->wptr_reg = wptr_reg; + ring->ptr_reg_shift = ptr_reg_shift; + ring->ptr_reg_mask = ptr_reg_mask; + ring->nop = nop; /* Allocate ring buffer */ - if (rdev->cp.ring_obj == NULL) { - r = radeon_bo_create(rdev, rdev->cp.ring_size, PAGE_SIZE, true, + if (ring->ring_obj == NULL) { + r = radeon_bo_create(rdev, ring->ring_size, PAGE_SIZE, true, RADEON_GEM_DOMAIN_GTT, - &rdev->cp.ring_obj); + &ring->ring_obj); if (r) { dev_err(rdev->dev, "(%d) ring create failed\n", r); return r; } - r = radeon_bo_reserve(rdev->cp.ring_obj, false); + r = radeon_bo_reserve(ring->ring_obj, false); if (unlikely(r != 0)) return r; - r = radeon_bo_pin(rdev->cp.ring_obj, RADEON_GEM_DOMAIN_GTT, - &rdev->cp.gpu_addr); + r = radeon_bo_pin(ring->ring_obj, RADEON_GEM_DOMAIN_GTT, + &ring->gpu_addr); if (r) { - radeon_bo_unreserve(rdev->cp.ring_obj); + radeon_bo_unreserve(ring->ring_obj); dev_err(rdev->dev, "(%d) ring pin failed\n", r); return r; } - r = radeon_bo_kmap(rdev->cp.ring_obj, - (void **)&rdev->cp.ring); - radeon_bo_unreserve(rdev->cp.ring_obj); + r = radeon_bo_kmap(ring->ring_obj, + (void **)&ring->ring); + radeon_bo_unreserve(ring->ring_obj); if (r) { dev_err(rdev->dev, "(%d) ring map failed\n", r); return r; } } - rdev->cp.ptr_mask = (rdev->cp.ring_size / 4) - 1; - rdev->cp.ring_free_dw = rdev->cp.ring_size / 4; + ring->ptr_mask = (ring->ring_size / 4) - 1; + ring->ring_free_dw = ring->ring_size / 4; return 0; } -void radeon_ring_fini(struct radeon_device *rdev) +void radeon_ring_fini(struct radeon_device *rdev, struct radeon_ring *ring) { int r; struct radeon_bo *ring_obj; - mutex_lock(&rdev->cp.mutex); - ring_obj = rdev->cp.ring_obj; - rdev->cp.ring = NULL; - rdev->cp.ring_obj = NULL; - mutex_unlock(&rdev->cp.mutex); + mutex_lock(&ring->mutex); + ring_obj = ring->ring_obj; + ring->ring = NULL; + ring->ring_obj = NULL; + mutex_unlock(&ring->mutex); if (ring_obj) { r = radeon_bo_reserve(ring_obj, false); @@ -425,78 +439,96 @@ void radeon_ring_fini(struct radeon_device *rdev) } } - /* * Debugfs info */ #if defined(CONFIG_DEBUG_FS) -static int radeon_debugfs_ib_info(struct seq_file *m, void *data) + +static int radeon_debugfs_ring_info(struct seq_file *m, void *data) { struct drm_info_node *node = (struct drm_info_node *) m->private; - struct radeon_ib *ib = node->info_ent->data; - unsigned i; - - if (ib == NULL) { - return 0; - } - seq_printf(m, "IB %04u\n", ib->idx); - seq_printf(m, "IB fence %p\n", ib->fence); - seq_printf(m, "IB size %05u dwords\n", ib->length_dw); - for (i = 0; i < ib->length_dw; i++) { - seq_printf(m, "[%05u]=0x%08X\n", i, ib->ptr[i]); + struct drm_device *dev = node->minor->dev; + struct radeon_device *rdev = dev->dev_private; + int ridx = *(int*)node->info_ent->data; + struct radeon_ring *ring = &rdev->ring[ridx]; + unsigned count, i, j; + + radeon_ring_free_size(rdev, ring); + count = (ring->ring_size / 4) - ring->ring_free_dw; + seq_printf(m, "wptr(0x%04x): 0x%08x\n", ring->wptr_reg, RREG32(ring->wptr_reg)); + seq_printf(m, "rptr(0x%04x): 0x%08x\n", ring->rptr_reg, RREG32(ring->rptr_reg)); + seq_printf(m, "driver's copy of the wptr: 0x%08x\n", ring->wptr); + seq_printf(m, "driver's copy of the rptr: 0x%08x\n", ring->rptr); + seq_printf(m, "%u free dwords in ring\n", ring->ring_free_dw); + seq_printf(m, "%u dwords in ring\n", count); + i = ring->rptr; + for (j = 0; j <= count; j++) { + seq_printf(m, "r[%04d]=0x%08x\n", i, ring->ring[i]); + i = (i + 1) & ring->ptr_mask; } return 0; } -static int radeon_debugfs_ib_bogus_info(struct seq_file *m, void *data) +static int radeon_ring_type_gfx_index = RADEON_RING_TYPE_GFX_INDEX; +static int cayman_ring_type_cp1_index = CAYMAN_RING_TYPE_CP1_INDEX; +static int cayman_ring_type_cp2_index = CAYMAN_RING_TYPE_CP2_INDEX; + +static struct drm_info_list radeon_debugfs_ring_info_list[] = { + {"radeon_ring_gfx", radeon_debugfs_ring_info, 0, &radeon_ring_type_gfx_index}, + {"radeon_ring_cp1", radeon_debugfs_ring_info, 0, &cayman_ring_type_cp1_index}, + {"radeon_ring_cp2", radeon_debugfs_ring_info, 0, &cayman_ring_type_cp2_index}, +}; + +static int radeon_debugfs_ib_info(struct seq_file *m, void *data) { struct drm_info_node *node = (struct drm_info_node *) m->private; - struct radeon_device *rdev = node->info_ent->data; - struct radeon_ib *ib; + struct drm_device *dev = node->minor->dev; + struct radeon_device *rdev = dev->dev_private; + struct radeon_ib *ib = &rdev->ib_pool.ibs[*((unsigned*)node->info_ent->data)]; unsigned i; - mutex_lock(&rdev->ib_pool.mutex); - if (list_empty(&rdev->ib_pool.bogus_ib)) { - mutex_unlock(&rdev->ib_pool.mutex); - seq_printf(m, "no bogus IB recorded\n"); + if (ib == NULL) { return 0; } - ib = list_first_entry(&rdev->ib_pool.bogus_ib, struct radeon_ib, list); - list_del_init(&ib->list); - mutex_unlock(&rdev->ib_pool.mutex); + seq_printf(m, "IB %04u\n", ib->idx); + seq_printf(m, "IB fence %p\n", ib->fence); seq_printf(m, "IB size %05u dwords\n", ib->length_dw); for (i = 0; i < ib->length_dw; i++) { seq_printf(m, "[%05u]=0x%08X\n", i, ib->ptr[i]); } - vfree(ib->ptr); - kfree(ib); return 0; } static struct drm_info_list radeon_debugfs_ib_list[RADEON_IB_POOL_SIZE]; static char radeon_debugfs_ib_names[RADEON_IB_POOL_SIZE][32]; +static unsigned radeon_debugfs_ib_idx[RADEON_IB_POOL_SIZE]; +#endif -static struct drm_info_list radeon_debugfs_ib_bogus_info_list[] = { - {"radeon_ib_bogus", radeon_debugfs_ib_bogus_info, 0, NULL}, -}; +int radeon_debugfs_ring_init(struct radeon_device *rdev) +{ +#if defined(CONFIG_DEBUG_FS) + if (rdev->family >= CHIP_CAYMAN) + return radeon_debugfs_add_files(rdev, radeon_debugfs_ring_info_list, + ARRAY_SIZE(radeon_debugfs_ring_info_list)); + else + return radeon_debugfs_add_files(rdev, radeon_debugfs_ring_info_list, 1); +#else + return 0; #endif +} int radeon_debugfs_ib_init(struct radeon_device *rdev) { #if defined(CONFIG_DEBUG_FS) unsigned i; - int r; - radeon_debugfs_ib_bogus_info_list[0].data = rdev; - r = radeon_debugfs_add_files(rdev, radeon_debugfs_ib_bogus_info_list, 1); - if (r) - return r; for (i = 0; i < RADEON_IB_POOL_SIZE; i++) { sprintf(radeon_debugfs_ib_names[i], "radeon_ib_%04u", i); + radeon_debugfs_ib_idx[i] = i; radeon_debugfs_ib_list[i].name = radeon_debugfs_ib_names[i]; radeon_debugfs_ib_list[i].show = &radeon_debugfs_ib_info; radeon_debugfs_ib_list[i].driver_features = 0; - radeon_debugfs_ib_list[i].data = &rdev->ib_pool.ibs[i]; + radeon_debugfs_ib_list[i].data = &radeon_debugfs_ib_idx[i]; } return radeon_debugfs_add_files(rdev, radeon_debugfs_ib_list, RADEON_IB_POOL_SIZE); diff --git a/drivers/gpu/drm/radeon/radeon_sa.c b/drivers/gpu/drm/radeon/radeon_sa.c new file mode 100644 index 0000000..4cce47e --- /dev/null +++ b/drivers/gpu/drm/radeon/radeon_sa.c @@ -0,0 +1,189 @@ +/* + * Copyright 2011 Red Hat Inc. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, + * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR + * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE + * USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + */ +/* + * Authors: + * Jerome Glisse + */ +#include "drmP.h" +#include "drm.h" +#include "radeon.h" + +int radeon_sa_bo_manager_init(struct radeon_device *rdev, + struct radeon_sa_manager *sa_manager, + unsigned size, u32 domain) +{ + int r; + + sa_manager->bo = NULL; + sa_manager->size = size; + sa_manager->domain = domain; + INIT_LIST_HEAD(&sa_manager->sa_bo); + + r = radeon_bo_create(rdev, size, RADEON_GPU_PAGE_SIZE, true, + RADEON_GEM_DOMAIN_CPU, &sa_manager->bo); + if (r) { + dev_err(rdev->dev, "(%d) failed to allocate bo for manager\n", r); + return r; + } + + return r; +} + +void radeon_sa_bo_manager_fini(struct radeon_device *rdev, + struct radeon_sa_manager *sa_manager) +{ + struct radeon_sa_bo *sa_bo, *tmp; + + if (!list_empty(&sa_manager->sa_bo)) { + dev_err(rdev->dev, "sa_manager is not empty, clearing anyway\n"); + } + list_for_each_entry_safe(sa_bo, tmp, &sa_manager->sa_bo, list) { + list_del_init(&sa_bo->list); + } + radeon_bo_unref(&sa_manager->bo); + sa_manager->size = 0; +} + +int radeon_sa_bo_manager_start(struct radeon_device *rdev, + struct radeon_sa_manager *sa_manager) +{ + int r; + + if (sa_manager->bo == NULL) { + dev_err(rdev->dev, "no bo for sa manager\n"); + return -EINVAL; + } + + /* map the buffer */ + r = radeon_bo_reserve(sa_manager->bo, false); + if (r) { + dev_err(rdev->dev, "(%d) failed to reserve manager bo\n", r); + return r; + } + r = radeon_bo_pin(sa_manager->bo, sa_manager->domain, &sa_manager->gpu_addr); + if (r) { + radeon_bo_unreserve(sa_manager->bo); + dev_err(rdev->dev, "(%d) failed to pin manager bo\n", r); + return r; + } + r = radeon_bo_kmap(sa_manager->bo, &sa_manager->cpu_ptr); + radeon_bo_unreserve(sa_manager->bo); + return r; +} + +int radeon_sa_bo_manager_suspend(struct radeon_device *rdev, + struct radeon_sa_manager *sa_manager) +{ + int r; + + if (sa_manager->bo == NULL) { + dev_err(rdev->dev, "no bo for sa manager\n"); + return -EINVAL; + } + + r = radeon_bo_reserve(sa_manager->bo, false); + if (!r) { + radeon_bo_kunmap(sa_manager->bo); + radeon_bo_unpin(sa_manager->bo); + radeon_bo_unreserve(sa_manager->bo); + } + return r; +} + +/* + * Principe is simple, we keep a list of sub allocation in offset + * order (first entry has offset == 0, last entry has the highest + * offset). + * + * When allocating new object we first check if there is room at + * the end total_size - (last_object_offset + last_object_size) >= + * alloc_size. If so we allocate new object there. + * + * When there is not enough room at the end, we start waiting for + * each sub object until we reach object_offset+object_size >= + * alloc_size, this object then become the sub object we return. + * + * Alignment can't be bigger than page size + */ +int radeon_sa_bo_new(struct radeon_device *rdev, + struct radeon_sa_manager *sa_manager, + struct radeon_sa_bo *sa_bo, + unsigned size, unsigned align) +{ + struct radeon_sa_bo *tmp; + struct list_head *head; + unsigned offset = 0, wasted = 0; + + BUG_ON(align > RADEON_GPU_PAGE_SIZE); + BUG_ON(size > sa_manager->size); + + /* no one ? */ + head = sa_manager->sa_bo.prev; + if (list_empty(&sa_manager->sa_bo)) { + goto out; + } + + /* look for a hole big enough */ + offset = 0; + list_for_each_entry(tmp, &sa_manager->sa_bo, list) { + /* room before this object ? */ + if ((tmp->offset - offset) >= size) { + head = tmp->list.prev; + goto out; + } + offset = tmp->offset + tmp->size; + wasted = offset % align; + if (wasted) { + wasted = align - wasted; + } + offset += wasted; + } + /* room at the end ? */ + head = sa_manager->sa_bo.prev; + tmp = list_entry(head, struct radeon_sa_bo, list); + offset = tmp->offset + tmp->size; + wasted = offset % align; + if (wasted) { + wasted = align - wasted; + } + offset += wasted; + if ((sa_manager->size - offset) < size) { + /* failed to find somethings big enough */ + return -ENOMEM; + } + +out: + sa_bo->manager = sa_manager; + sa_bo->offset = offset; + sa_bo->size = size; + list_add(&sa_bo->list, head); + return 0; +} + +void radeon_sa_bo_free(struct radeon_device *rdev, struct radeon_sa_bo *sa_bo) +{ + list_del_init(&sa_bo->list); +} diff --git a/drivers/gpu/drm/radeon/radeon_semaphore.c b/drivers/gpu/drm/radeon/radeon_semaphore.c new file mode 100644 index 0000000..61dd4e3 --- /dev/null +++ b/drivers/gpu/drm/radeon/radeon_semaphore.c @@ -0,0 +1,178 @@ +/* + * Copyright 2011 Christian König. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, + * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR + * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE + * USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + */ +/* + * Authors: + * Christian König + */ +#include "drmP.h" +#include "drm.h" +#include "radeon.h" + +static int radeon_semaphore_add_bo(struct radeon_device *rdev) +{ + struct radeon_semaphore_bo *bo; + unsigned long irq_flags; + uint64_t gpu_addr; + uint32_t *cpu_ptr; + int r, i; + + + bo = kmalloc(sizeof(struct radeon_semaphore_bo), GFP_KERNEL); + if (bo == NULL) { + return -ENOMEM; + } + INIT_LIST_HEAD(&bo->free); + INIT_LIST_HEAD(&bo->list); + bo->nused = 0; + + r = radeon_ib_get(rdev, 0, &bo->ib, RADEON_SEMAPHORE_BO_SIZE); + if (r) { + dev_err(rdev->dev, "failed to get a bo after 5 retry\n"); + kfree(bo); + return r; + } + gpu_addr = rdev->ib_pool.sa_manager.gpu_addr; + gpu_addr += bo->ib->sa_bo.offset; + cpu_ptr = rdev->ib_pool.sa_manager.cpu_ptr; + cpu_ptr += (bo->ib->sa_bo.offset >> 2); + for (i = 0; i < (RADEON_SEMAPHORE_BO_SIZE/8); i++) { + bo->semaphores[i].gpu_addr = gpu_addr; + bo->semaphores[i].cpu_ptr = cpu_ptr; + bo->semaphores[i].bo = bo; + list_add_tail(&bo->semaphores[i].list, &bo->free); + gpu_addr += 8; + cpu_ptr += 2; + } + write_lock_irqsave(&rdev->semaphore_drv.lock, irq_flags); + list_add_tail(&bo->list, &rdev->semaphore_drv.bo); + write_unlock_irqrestore(&rdev->semaphore_drv.lock, irq_flags); + return 0; +} + +static void radeon_semaphore_del_bo_locked(struct radeon_device *rdev, + struct radeon_semaphore_bo *bo) +{ + radeon_sa_bo_free(rdev, &bo->ib->sa_bo); + radeon_fence_unref(&bo->ib->fence); + list_del(&bo->list); + kfree(bo); +} + +void radeon_semaphore_shrink_locked(struct radeon_device *rdev) +{ + struct radeon_semaphore_bo *bo, *n; + + if (list_empty(&rdev->semaphore_drv.bo)) { + return; + } + /* only shrink if first bo has free semaphore */ + bo = list_first_entry(&rdev->semaphore_drv.bo, struct radeon_semaphore_bo, list); + if (list_empty(&bo->free)) { + return; + } + list_for_each_entry_safe_continue(bo, n, &rdev->semaphore_drv.bo, list) { + if (bo->nused) + continue; + radeon_semaphore_del_bo_locked(rdev, bo); + } +} + +int radeon_semaphore_create(struct radeon_device *rdev, + struct radeon_semaphore **semaphore) +{ + struct radeon_semaphore_bo *bo; + unsigned long irq_flags; + bool do_retry = true; + int r; + +retry: + *semaphore = NULL; + write_lock_irqsave(&rdev->semaphore_drv.lock, irq_flags); + list_for_each_entry(bo, &rdev->semaphore_drv.bo, list) { + if (list_empty(&bo->free)) + continue; + *semaphore = list_first_entry(&bo->free, struct radeon_semaphore, list); + (*semaphore)->cpu_ptr[0] = 0; + (*semaphore)->cpu_ptr[1] = 0; + list_del(&(*semaphore)->list); + bo->nused++; + break; + } + write_unlock_irqrestore(&rdev->semaphore_drv.lock, irq_flags); + + if (*semaphore == NULL) { + if (do_retry) { + do_retry = false; + r = radeon_semaphore_add_bo(rdev); + if (r) + return r; + goto retry; + } + return -ENOMEM; + } + + return 0; +} + +void radeon_semaphore_emit_signal(struct radeon_device *rdev, int ring, + struct radeon_semaphore *semaphore) +{ + radeon_semaphore_ring_emit(rdev, ring, &rdev->ring[ring], semaphore, false); +} + +void radeon_semaphore_emit_wait(struct radeon_device *rdev, int ring, + struct radeon_semaphore *semaphore) +{ + radeon_semaphore_ring_emit(rdev, ring, &rdev->ring[ring], semaphore, true); +} + +void radeon_semaphore_free(struct radeon_device *rdev, + struct radeon_semaphore *semaphore) +{ + unsigned long irq_flags; + + write_lock_irqsave(&rdev->semaphore_drv.lock, irq_flags); + semaphore->bo->nused--; + list_add_tail(&semaphore->list, &semaphore->bo->free); + radeon_semaphore_shrink_locked(rdev); + write_unlock_irqrestore(&rdev->semaphore_drv.lock, irq_flags); +} + +void radeon_semaphore_driver_fini(struct radeon_device *rdev) +{ + struct radeon_semaphore_bo *bo, *n; + unsigned long irq_flags; + + write_lock_irqsave(&rdev->semaphore_drv.lock, irq_flags); + /* we force to free everything */ + list_for_each_entry_safe(bo, n, &rdev->semaphore_drv.bo, list) { + if (!list_empty(&bo->free)) { + dev_err(rdev->dev, "still in use semaphore\n"); + } + radeon_semaphore_del_bo_locked(rdev, bo); + } + write_unlock_irqrestore(&rdev->semaphore_drv.lock, irq_flags); +} diff --git a/drivers/gpu/drm/radeon/radeon_test.c b/drivers/gpu/drm/radeon/radeon_test.c index 602fa35..dc5dcf4 100644 --- a/drivers/gpu/drm/radeon/radeon_test.c +++ b/drivers/gpu/drm/radeon/radeon_test.c @@ -42,7 +42,9 @@ void radeon_test_moves(struct radeon_device *rdev) /* Number of tests = * (Total GTT - IB pool - writeback page - ring buffers) / test size */ - n = rdev->mc.gtt_size - RADEON_IB_POOL_SIZE*64*1024 - rdev->cp.ring_size; + n = rdev->mc.gtt_size - RADEON_IB_POOL_SIZE*64*1024; + for (i = 0; i < RADEON_NUM_RINGS; ++i) + n -= rdev->ring[i].ring_size; if (rdev->wb.wb_obj) n -= RADEON_GPU_PAGE_SIZE; if (rdev->ih.ring_obj) @@ -104,7 +106,7 @@ void radeon_test_moves(struct radeon_device *rdev) radeon_bo_kunmap(gtt_obj[i]); - r = radeon_fence_create(rdev, &fence); + r = radeon_fence_create(rdev, &fence, RADEON_RING_TYPE_GFX_INDEX); if (r) { DRM_ERROR("Failed to create GTT->VRAM fence %d\n", i); goto out_cleanup; @@ -153,7 +155,7 @@ void radeon_test_moves(struct radeon_device *rdev) radeon_bo_kunmap(vram_obj); - r = radeon_fence_create(rdev, &fence); + r = radeon_fence_create(rdev, &fence, RADEON_RING_TYPE_GFX_INDEX); if (r) { DRM_ERROR("Failed to create VRAM->GTT fence %d\n", i); goto out_cleanup; @@ -232,3 +234,264 @@ out_cleanup: printk(KERN_WARNING "Error while testing BO move.\n"); } } + +void radeon_test_ring_sync(struct radeon_device *rdev, + struct radeon_ring *ringA, + struct radeon_ring *ringB) +{ + struct radeon_fence *fence1 = NULL, *fence2 = NULL; + struct radeon_semaphore *semaphore = NULL; + int ridxA = radeon_ring_index(rdev, ringA); + int ridxB = radeon_ring_index(rdev, ringB); + int r; + + r = radeon_fence_create(rdev, &fence1, ridxA); + if (r) { + DRM_ERROR("Failed to create sync fence 1\n"); + goto out_cleanup; + } + r = radeon_fence_create(rdev, &fence2, ridxA); + if (r) { + DRM_ERROR("Failed to create sync fence 2\n"); + goto out_cleanup; + } + + r = radeon_semaphore_create(rdev, &semaphore); + if (r) { + DRM_ERROR("Failed to create semaphore\n"); + goto out_cleanup; + } + + r = radeon_ring_lock(rdev, ringA, 64); + if (r) { + DRM_ERROR("Failed to lock ring A %d\n", ridxA); + goto out_cleanup; + } + radeon_semaphore_emit_wait(rdev, ridxA, semaphore); + radeon_fence_emit(rdev, fence1); + radeon_semaphore_emit_wait(rdev, ridxA, semaphore); + radeon_fence_emit(rdev, fence2); + radeon_ring_unlock_commit(rdev, ringA); + + mdelay(1000); + + if (radeon_fence_signaled(fence1)) { + DRM_ERROR("Fence 1 signaled without waiting for semaphore.\n"); + goto out_cleanup; + } + + r = radeon_ring_lock(rdev, ringB, 64); + if (r) { + DRM_ERROR("Failed to lock ring B %p\n", ringB); + goto out_cleanup; + } + radeon_semaphore_emit_signal(rdev, ridxB, semaphore); + radeon_ring_unlock_commit(rdev, ringB); + + r = radeon_fence_wait(fence1, false); + if (r) { + DRM_ERROR("Failed to wait for sync fence 1\n"); + goto out_cleanup; + } + + mdelay(1000); + + if (radeon_fence_signaled(fence2)) { + DRM_ERROR("Fence 2 signaled without waiting for semaphore.\n"); + goto out_cleanup; + } + + r = radeon_ring_lock(rdev, ringB, 64); + if (r) { + DRM_ERROR("Failed to lock ring B %p\n", ringB); + goto out_cleanup; + } + radeon_semaphore_emit_signal(rdev, ridxB, semaphore); + radeon_ring_unlock_commit(rdev, ringB); + + r = radeon_fence_wait(fence2, false); + if (r) { + DRM_ERROR("Failed to wait for sync fence 1\n"); + goto out_cleanup; + } + +out_cleanup: + if (semaphore) + radeon_semaphore_free(rdev, semaphore); + + if (fence1) + radeon_fence_unref(&fence1); + + if (fence2) + radeon_fence_unref(&fence2); + + if (r) + printk(KERN_WARNING "Error while testing ring sync (%d).\n", r); +} + +void radeon_test_ring_sync2(struct radeon_device *rdev, + struct radeon_ring *ringA, + struct radeon_ring *ringB, + struct radeon_ring *ringC) +{ + struct radeon_fence *fenceA = NULL, *fenceB = NULL; + struct radeon_semaphore *semaphore = NULL; + int ridxA = radeon_ring_index(rdev, ringA); + int ridxB = radeon_ring_index(rdev, ringB); + int ridxC = radeon_ring_index(rdev, ringC); + bool sigA, sigB; + int i, r; + + r = radeon_fence_create(rdev, &fenceA, ridxA); + if (r) { + DRM_ERROR("Failed to create sync fence 1\n"); + goto out_cleanup; + } + r = radeon_fence_create(rdev, &fenceB, ridxB); + if (r) { + DRM_ERROR("Failed to create sync fence 2\n"); + goto out_cleanup; + } + + r = radeon_semaphore_create(rdev, &semaphore); + if (r) { + DRM_ERROR("Failed to create semaphore\n"); + goto out_cleanup; + } + + r = radeon_ring_lock(rdev, ringA, 64); + if (r) { + DRM_ERROR("Failed to lock ring A %d\n", ridxA); + goto out_cleanup; + } + radeon_semaphore_emit_wait(rdev, ridxA, semaphore); + radeon_fence_emit(rdev, fenceA); + radeon_ring_unlock_commit(rdev, ringA); + + r = radeon_ring_lock(rdev, ringB, 64); + if (r) { + DRM_ERROR("Failed to lock ring B %d\n", ridxB); + goto out_cleanup; + } + radeon_semaphore_emit_wait(rdev, ridxB, semaphore); + radeon_fence_emit(rdev, fenceB); + radeon_ring_unlock_commit(rdev, ringB); + + mdelay(1000); + + if (radeon_fence_signaled(fenceA)) { + DRM_ERROR("Fence A signaled without waiting for semaphore.\n"); + goto out_cleanup; + } + if (radeon_fence_signaled(fenceB)) { + DRM_ERROR("Fence A signaled without waiting for semaphore.\n"); + goto out_cleanup; + } + + r = radeon_ring_lock(rdev, ringC, 64); + if (r) { + DRM_ERROR("Failed to lock ring B %p\n", ringC); + goto out_cleanup; + } + radeon_semaphore_emit_signal(rdev, ridxC, semaphore); + radeon_ring_unlock_commit(rdev, ringC); + + for (i = 0; i < 30; ++i) { + mdelay(100); + sigA = radeon_fence_signaled(fenceA); + sigB = radeon_fence_signaled(fenceB); + if (sigA || sigB) + break; + } + + if (!sigA && !sigB) { + DRM_ERROR("Neither fence A nor B has been signaled\n"); + goto out_cleanup; + } else if (sigA && sigB) { + DRM_ERROR("Both fence A and B has been signaled\n"); + goto out_cleanup; + } + + DRM_INFO("Fence %c was first signaled\n", sigA ? 'A' : 'B'); + + r = radeon_ring_lock(rdev, ringC, 64); + if (r) { + DRM_ERROR("Failed to lock ring B %p\n", ringC); + goto out_cleanup; + } + radeon_semaphore_emit_signal(rdev, ridxC, semaphore); + radeon_ring_unlock_commit(rdev, ringC); + + mdelay(1000); + + r = radeon_fence_wait(fenceA, false); + if (r) { + DRM_ERROR("Failed to wait for sync fence A\n"); + goto out_cleanup; + } + r = radeon_fence_wait(fenceB, false); + if (r) { + DRM_ERROR("Failed to wait for sync fence B\n"); + goto out_cleanup; + } + +out_cleanup: + if (semaphore) + radeon_semaphore_free(rdev, semaphore); + + if (fenceA) + radeon_fence_unref(&fenceA); + + if (fenceB) + radeon_fence_unref(&fenceB); + + if (r) + printk(KERN_WARNING "Error while testing ring sync (%d).\n", r); +} + +void radeon_test_syncing(struct radeon_device *rdev) +{ + int i, j, k; + + for (i = 1; i < RADEON_NUM_RINGS; ++i) { + struct radeon_ring *ringA = &rdev->ring[i]; + if (!ringA->ready) + continue; + + for (j = 0; j < i; ++j) { + struct radeon_ring *ringB = &rdev->ring[j]; + if (!ringB->ready) + continue; + + DRM_INFO("Testing syncing between rings %d and %d...\n", i, j); + radeon_test_ring_sync(rdev, ringA, ringB); + + DRM_INFO("Testing syncing between rings %d and %d...\n", j, i); + radeon_test_ring_sync(rdev, ringB, ringA); + + for (k = 0; k < j; ++k) { + struct radeon_ring *ringC = &rdev->ring[k]; + if (!ringC->ready) + continue; + + DRM_INFO("Testing syncing between rings %d, %d and %d...\n", i, j, k); + radeon_test_ring_sync2(rdev, ringA, ringB, ringC); + + DRM_INFO("Testing syncing between rings %d, %d and %d...\n", i, k, j); + radeon_test_ring_sync2(rdev, ringA, ringC, ringB); + + DRM_INFO("Testing syncing between rings %d, %d and %d...\n", j, i, k); + radeon_test_ring_sync2(rdev, ringB, ringA, ringC); + + DRM_INFO("Testing syncing between rings %d, %d and %d...\n", j, k, i); + radeon_test_ring_sync2(rdev, ringB, ringC, ringA); + + DRM_INFO("Testing syncing between rings %d, %d and %d...\n", k, i, j); + radeon_test_ring_sync2(rdev, ringC, ringA, ringB); + + DRM_INFO("Testing syncing between rings %d, %d and %d...\n", k, j, i); + radeon_test_ring_sync2(rdev, ringC, ringB, ringA); + } + } + } +} diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c index 0b5468b..f493c64 100644 --- a/drivers/gpu/drm/radeon/radeon_ttm.c +++ b/drivers/gpu/drm/radeon/radeon_ttm.c @@ -114,24 +114,6 @@ static void radeon_ttm_global_fini(struct radeon_device *rdev) } } -struct ttm_backend *radeon_ttm_backend_create(struct radeon_device *rdev); - -static struct ttm_backend* -radeon_create_ttm_backend_entry(struct ttm_bo_device *bdev) -{ - struct radeon_device *rdev; - - rdev = radeon_get_rdev(bdev); -#if __OS_HAS_AGP - if (rdev->flags & RADEON_IS_AGP) { - return ttm_agp_backend_init(bdev, rdev->ddev->agp->bridge); - } else -#endif - { - return radeon_ttm_backend_create(rdev); - } -} - static int radeon_invalidate_caches(struct ttm_bo_device *bdev, uint32_t flags) { return 0; @@ -206,7 +188,7 @@ static void radeon_evict_flags(struct ttm_buffer_object *bo, rbo = container_of(bo, struct radeon_bo, tbo); switch (bo->mem.mem_type) { case TTM_PL_VRAM: - if (rbo->rdev->cp.ready == false) + if (rbo->rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready == false) radeon_ttm_placement_from_domain(rbo, RADEON_GEM_DOMAIN_CPU); else radeon_ttm_placement_from_domain(rbo, RADEON_GEM_DOMAIN_GTT); @@ -241,10 +223,10 @@ static int radeon_move_blit(struct ttm_buffer_object *bo, struct radeon_device *rdev; uint64_t old_start, new_start; struct radeon_fence *fence; - int r; + int r, i; rdev = radeon_get_rdev(bo->bdev); - r = radeon_fence_create(rdev, &fence); + r = radeon_fence_create(rdev, &fence, radeon_copy_ring_index(rdev)); if (unlikely(r)) { return r; } @@ -273,13 +255,43 @@ static int radeon_move_blit(struct ttm_buffer_object *bo, DRM_ERROR("Unknown placement %d\n", old_mem->mem_type); return -EINVAL; } - if (!rdev->cp.ready) { - DRM_ERROR("Trying to move memory with CP turned off.\n"); + if (!rdev->ring[radeon_copy_ring_index(rdev)].ready) { + DRM_ERROR("Trying to move memory with ring turned off.\n"); return -EINVAL; } BUILD_BUG_ON((PAGE_SIZE % RADEON_GPU_PAGE_SIZE) != 0); + /* sync other rings */ + if (rdev->family >= CHIP_R600) { + for (i = 0; i < RADEON_NUM_RINGS; ++i) { + /* no need to sync to our own or unused rings */ + if (i == radeon_copy_ring_index(rdev) || !rdev->ring[i].ready) + continue; + + if (!fence->semaphore) { + r = radeon_semaphore_create(rdev, &fence->semaphore); + /* FIXME: handle semaphore error */ + if (r) + continue; + } + + r = radeon_ring_lock(rdev, &rdev->ring[i], 3); + /* FIXME: handle ring lock error */ + if (r) + continue; + radeon_semaphore_emit_signal(rdev, i, fence->semaphore); + radeon_ring_unlock_commit(rdev, &rdev->ring[i]); + + r = radeon_ring_lock(rdev, &rdev->ring[radeon_copy_ring_index(rdev)], 3); + /* FIXME: handle ring lock error */ + if (r) + continue; + radeon_semaphore_emit_wait(rdev, radeon_copy_ring_index(rdev), fence->semaphore); + radeon_ring_unlock_commit(rdev, &rdev->ring[radeon_copy_ring_index(rdev)]); + } + } + r = radeon_copy(rdev, old_start, new_start, new_mem->num_pages * (PAGE_SIZE / RADEON_GPU_PAGE_SIZE), /* GPU pages */ fence); @@ -398,7 +410,8 @@ static int radeon_bo_move(struct ttm_buffer_object *bo, radeon_move_null(bo, new_mem); return 0; } - if (!rdev->cp.ready || rdev->asic->copy == NULL) { + if (!rdev->ring[radeon_copy_ring_index(rdev)].ready || + rdev->asic->copy.copy == NULL) { /* use memcpy */ goto memcpy; } @@ -515,8 +528,166 @@ static bool radeon_sync_obj_signaled(void *sync_obj, void *sync_arg) return radeon_fence_signaled((struct radeon_fence *)sync_obj); } +/* + * TTM backend functions. + */ +struct radeon_ttm_tt { + struct ttm_dma_tt ttm; + struct radeon_device *rdev; + u64 offset; +}; + +static int radeon_ttm_backend_bind(struct ttm_tt *ttm, + struct ttm_mem_reg *bo_mem) +{ + struct radeon_ttm_tt *gtt = (void*)ttm; + int r; + + gtt->offset = (unsigned long)(bo_mem->start << PAGE_SHIFT); + if (!ttm->num_pages) { + WARN(1, "nothing to bind %lu pages for mreg %p back %p!\n", + ttm->num_pages, bo_mem, ttm); + } + r = radeon_gart_bind(gtt->rdev, gtt->offset, + ttm->num_pages, ttm->pages, gtt->ttm.dma_address); + if (r) { + DRM_ERROR("failed to bind %lu pages at 0x%08X\n", + ttm->num_pages, (unsigned)gtt->offset); + return r; + } + return 0; +} + +static int radeon_ttm_backend_unbind(struct ttm_tt *ttm) +{ + struct radeon_ttm_tt *gtt = (void *)ttm; + + radeon_gart_unbind(gtt->rdev, gtt->offset, ttm->num_pages); + return 0; +} + +static void radeon_ttm_backend_destroy(struct ttm_tt *ttm) +{ + struct radeon_ttm_tt *gtt = (void *)ttm; + + ttm_dma_tt_fini(>t->ttm); + kfree(gtt); +} + +static struct ttm_backend_func radeon_backend_func = { + .bind = &radeon_ttm_backend_bind, + .unbind = &radeon_ttm_backend_unbind, + .destroy = &radeon_ttm_backend_destroy, +}; + +struct ttm_tt *radeon_ttm_tt_create(struct ttm_bo_device *bdev, + unsigned long size, uint32_t page_flags, + struct page *dummy_read_page) +{ + struct radeon_device *rdev; + struct radeon_ttm_tt *gtt; + + rdev = radeon_get_rdev(bdev); +#if __OS_HAS_AGP + if (rdev->flags & RADEON_IS_AGP) { + return ttm_agp_tt_create(bdev, rdev->ddev->agp->bridge, + size, page_flags, dummy_read_page); + } +#endif + + gtt = kzalloc(sizeof(struct radeon_ttm_tt), GFP_KERNEL); + if (gtt == NULL) { + return NULL; + } + gtt->ttm.ttm.func = &radeon_backend_func; + gtt->rdev = rdev; + if (ttm_dma_tt_init(>t->ttm, bdev, size, page_flags, dummy_read_page)) { + kfree(gtt); + return NULL; + } + return >t->ttm.ttm; +} + +static int radeon_ttm_tt_populate(struct ttm_tt *ttm) +{ + struct radeon_device *rdev; + struct radeon_ttm_tt *gtt = (void *)ttm; + unsigned i; + int r; + + if (ttm->state != tt_unpopulated) + return 0; + + rdev = radeon_get_rdev(ttm->bdev); +#if __OS_HAS_AGP + if (rdev->flags & RADEON_IS_AGP) { + return ttm_agp_tt_populate(ttm); + } +#endif + +#ifdef CONFIG_SWIOTLB + if (swiotlb_nr_tbl()) { + return ttm_dma_populate(>t->ttm, rdev->dev); + } +#endif + + r = ttm_pool_populate(ttm); + if (r) { + return r; + } + + for (i = 0; i < ttm->num_pages; i++) { + gtt->ttm.dma_address[i] = pci_map_page(rdev->pdev, ttm->pages[i], + 0, PAGE_SIZE, + PCI_DMA_BIDIRECTIONAL); + if (pci_dma_mapping_error(rdev->pdev, gtt->ttm.dma_address[i])) { + while (--i) { + pci_unmap_page(rdev->pdev, gtt->ttm.dma_address[i], + PAGE_SIZE, PCI_DMA_BIDIRECTIONAL); + gtt->ttm.dma_address[i] = 0; + } + ttm_pool_unpopulate(ttm); + return -EFAULT; + } + } + return 0; +} + +static void radeon_ttm_tt_unpopulate(struct ttm_tt *ttm) +{ + struct radeon_device *rdev; + struct radeon_ttm_tt *gtt = (void *)ttm; + unsigned i; + + rdev = radeon_get_rdev(ttm->bdev); +#if __OS_HAS_AGP + if (rdev->flags & RADEON_IS_AGP) { + ttm_agp_tt_unpopulate(ttm); + return; + } +#endif + +#ifdef CONFIG_SWIOTLB + if (swiotlb_nr_tbl()) { + ttm_dma_unpopulate(>t->ttm, rdev->dev); + return; + } +#endif + + for (i = 0; i < ttm->num_pages; i++) { + if (gtt->ttm.dma_address[i]) { + pci_unmap_page(rdev->pdev, gtt->ttm.dma_address[i], + PAGE_SIZE, PCI_DMA_BIDIRECTIONAL); + } + } + + ttm_pool_unpopulate(ttm); +} + static struct ttm_bo_driver radeon_bo_driver = { - .create_ttm_backend_entry = &radeon_create_ttm_backend_entry, + .ttm_tt_create = &radeon_ttm_tt_create, + .ttm_tt_populate = &radeon_ttm_tt_populate, + .ttm_tt_unpopulate = &radeon_ttm_tt_unpopulate, .invalidate_caches = &radeon_invalidate_caches, .init_mem_type = &radeon_init_mem_type, .evict_flags = &radeon_evict_flags, @@ -680,124 +851,6 @@ int radeon_mmap(struct file *filp, struct vm_area_struct *vma) } -/* - * TTM backend functions. - */ -struct radeon_ttm_backend { - struct ttm_backend backend; - struct radeon_device *rdev; - unsigned long num_pages; - struct page **pages; - struct page *dummy_read_page; - dma_addr_t *dma_addrs; - bool populated; - bool bound; - unsigned offset; -}; - -static int radeon_ttm_backend_populate(struct ttm_backend *backend, - unsigned long num_pages, - struct page **pages, - struct page *dummy_read_page, - dma_addr_t *dma_addrs) -{ - struct radeon_ttm_backend *gtt; - - gtt = container_of(backend, struct radeon_ttm_backend, backend); - gtt->pages = pages; - gtt->dma_addrs = dma_addrs; - gtt->num_pages = num_pages; - gtt->dummy_read_page = dummy_read_page; - gtt->populated = true; - return 0; -} - -static void radeon_ttm_backend_clear(struct ttm_backend *backend) -{ - struct radeon_ttm_backend *gtt; - - gtt = container_of(backend, struct radeon_ttm_backend, backend); - gtt->pages = NULL; - gtt->dma_addrs = NULL; - gtt->num_pages = 0; - gtt->dummy_read_page = NULL; - gtt->populated = false; - gtt->bound = false; -} - - -static int radeon_ttm_backend_bind(struct ttm_backend *backend, - struct ttm_mem_reg *bo_mem) -{ - struct radeon_ttm_backend *gtt; - int r; - - gtt = container_of(backend, struct radeon_ttm_backend, backend); - gtt->offset = bo_mem->start << PAGE_SHIFT; - if (!gtt->num_pages) { - WARN(1, "nothing to bind %lu pages for mreg %p back %p!\n", - gtt->num_pages, bo_mem, backend); - } - r = radeon_gart_bind(gtt->rdev, gtt->offset, - gtt->num_pages, gtt->pages, gtt->dma_addrs); - if (r) { - DRM_ERROR("failed to bind %lu pages at 0x%08X\n", - gtt->num_pages, gtt->offset); - return r; - } - gtt->bound = true; - return 0; -} - -static int radeon_ttm_backend_unbind(struct ttm_backend *backend) -{ - struct radeon_ttm_backend *gtt; - - gtt = container_of(backend, struct radeon_ttm_backend, backend); - radeon_gart_unbind(gtt->rdev, gtt->offset, gtt->num_pages); - gtt->bound = false; - return 0; -} - -static void radeon_ttm_backend_destroy(struct ttm_backend *backend) -{ - struct radeon_ttm_backend *gtt; - - gtt = container_of(backend, struct radeon_ttm_backend, backend); - if (gtt->bound) { - radeon_ttm_backend_unbind(backend); - } - kfree(gtt); -} - -static struct ttm_backend_func radeon_backend_func = { - .populate = &radeon_ttm_backend_populate, - .clear = &radeon_ttm_backend_clear, - .bind = &radeon_ttm_backend_bind, - .unbind = &radeon_ttm_backend_unbind, - .destroy = &radeon_ttm_backend_destroy, -}; - -struct ttm_backend *radeon_ttm_backend_create(struct radeon_device *rdev) -{ - struct radeon_ttm_backend *gtt; - - gtt = kzalloc(sizeof(struct radeon_ttm_backend), GFP_KERNEL); - if (gtt == NULL) { - return NULL; - } - gtt->backend.bdev = &rdev->mman.bdev; - gtt->backend.flags = 0; - gtt->backend.func = &radeon_backend_func; - gtt->rdev = rdev; - gtt->pages = NULL; - gtt->num_pages = 0; - gtt->dummy_read_page = NULL; - gtt->populated = false; - gtt->bound = false; - return >t->backend; -} - #define RADEON_DEBUGFS_MEM_TYPES 2 #if defined(CONFIG_DEBUG_FS) @@ -820,8 +873,8 @@ static int radeon_mm_dump_table(struct seq_file *m, void *data) static int radeon_ttm_debugfs_init(struct radeon_device *rdev) { #if defined(CONFIG_DEBUG_FS) - static struct drm_info_list radeon_mem_types_list[RADEON_DEBUGFS_MEM_TYPES+1]; - static char radeon_mem_types_names[RADEON_DEBUGFS_MEM_TYPES+1][32]; + static struct drm_info_list radeon_mem_types_list[RADEON_DEBUGFS_MEM_TYPES+2]; + static char radeon_mem_types_names[RADEON_DEBUGFS_MEM_TYPES+2][32]; unsigned i; for (i = 0; i < RADEON_DEBUGFS_MEM_TYPES; i++) { @@ -843,8 +896,17 @@ static int radeon_ttm_debugfs_init(struct radeon_device *rdev) radeon_mem_types_list[i].name = radeon_mem_types_names[i]; radeon_mem_types_list[i].show = &ttm_page_alloc_debugfs; radeon_mem_types_list[i].driver_features = 0; - radeon_mem_types_list[i].data = NULL; - return radeon_debugfs_add_files(rdev, radeon_mem_types_list, RADEON_DEBUGFS_MEM_TYPES+1); + radeon_mem_types_list[i++].data = NULL; +#ifdef CONFIG_SWIOTLB + if (swiotlb_nr_tbl()) { + sprintf(radeon_mem_types_names[i], "ttm_dma_page_pool"); + radeon_mem_types_list[i].name = radeon_mem_types_names[i]; + radeon_mem_types_list[i].show = &ttm_dma_page_alloc_debugfs; + radeon_mem_types_list[i].driver_features = 0; + radeon_mem_types_list[i++].data = NULL; + } +#endif + return radeon_debugfs_add_files(rdev, radeon_mem_types_list, i); #endif return 0; diff --git a/drivers/gpu/drm/radeon/reg_srcs/cayman b/drivers/gpu/drm/radeon/reg_srcs/cayman index 2316977..a072fa8 100644 --- a/drivers/gpu/drm/radeon/reg_srcs/cayman +++ b/drivers/gpu/drm/radeon/reg_srcs/cayman @@ -1,5 +1,9 @@ cayman 0x9400 0x0000802C GRBM_GFX_INDEX +0x00008040 WAIT_UNTIL +0x000084FC CP_STRMOUT_CNTL +0x000085F0 CP_COHER_CNTL +0x000085F4 CP_COHER_SIZE 0x000088B0 VGT_VTX_VECT_EJECT_REG 0x000088C4 VGT_CACHE_INVALIDATION 0x000088D4 VGT_GS_VERTEX_REUSE @@ -77,7 +81,6 @@ cayman 0x9400 0x0002802C DB_DEPTH_CLEAR 0x00028030 PA_SC_SCREEN_SCISSOR_TL 0x00028034 PA_SC_SCREEN_SCISSOR_BR -0x0002805C DB_DEPTH_SLICE 0x00028140 SQ_ALU_CONST_BUFFER_SIZE_PS_0 0x00028144 SQ_ALU_CONST_BUFFER_SIZE_PS_1 0x00028148 SQ_ALU_CONST_BUFFER_SIZE_PS_2 @@ -206,7 +209,6 @@ cayman 0x9400 0x00028344 PA_SC_VPORT_ZMAX_14 0x00028348 PA_SC_VPORT_ZMIN_15 0x0002834C PA_SC_VPORT_ZMAX_15 -0x00028350 SX_MISC 0x00028354 SX_SURFACE_SYNC 0x0002835C SX_SCATTER_EXPORT_SIZE 0x00028380 SQ_VTX_SEMANTIC_0 @@ -508,10 +510,16 @@ cayman 0x9400 0x00028AA8 IA_MULTI_VGT_PARAM 0x00028AB4 VGT_REUSE_OFF 0x00028AB8 VGT_VTX_CNT_EN -0x00028ABC DB_HTILE_SURFACE 0x00028AC0 DB_SRESULTS_COMPARE_STATE0 0x00028AC4 DB_SRESULTS_COMPARE_STATE1 0x00028AC8 DB_PRELOAD_CONTROL +0x00028AD4 VGT_STRMOUT_VTX_STRIDE_0 +0x00028AE4 VGT_STRMOUT_VTX_STRIDE_1 +0x00028AF4 VGT_STRMOUT_VTX_STRIDE_2 +0x00028B04 VGT_STRMOUT_VTX_STRIDE_3 +0x00028B28 VGT_STRMOUT_DRAW_OPAQUE_OFFSET +0x00028B2C VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE +0x00028B30 VGT_STRMOUT_DRAW_OPAQUE_VERTEX_STRIDE 0x00028B38 VGT_GS_MAX_VERT_OUT 0x00028B54 VGT_SHADER_STAGES_EN 0x00028B58 VGT_LS_HS_CONFIG @@ -551,6 +559,18 @@ cayman 0x9400 0x00028C34 PA_SC_AA_SAMPLE_LOCS_PIXEL_X1_Y1_3 0x00028C38 PA_SC_AA_MASK_X0_Y0_X1_Y0 0x00028C3C PA_SC_AA_MASK_X0_Y1_X1_Y1 +0x00028C78 CB_COLOR0_DIM +0x00028CB4 CB_COLOR1_DIM +0x00028CF0 CB_COLOR2_DIM +0x00028D2C CB_COLOR3_DIM +0x00028D68 CB_COLOR4_DIM +0x00028DA4 CB_COLOR5_DIM +0x00028DE0 CB_COLOR6_DIM +0x00028E1C CB_COLOR7_DIM +0x00028E58 CB_COLOR8_DIM +0x00028E74 CB_COLOR9_DIM +0x00028E90 CB_COLOR10_DIM +0x00028EAC CB_COLOR11_DIM 0x00028C8C CB_COLOR0_CLEAR_WORD0 0x00028C90 CB_COLOR0_CLEAR_WORD1 0x00028C94 CB_COLOR0_CLEAR_WORD2 diff --git a/drivers/gpu/drm/radeon/reg_srcs/evergreen b/drivers/gpu/drm/radeon/reg_srcs/evergreen index 161737a..b912a37 100644 --- a/drivers/gpu/drm/radeon/reg_srcs/evergreen +++ b/drivers/gpu/drm/radeon/reg_srcs/evergreen @@ -4,6 +4,9 @@ evergreen 0x9400 0x00008044 WAIT_UNTIL_POLL_CNTL 0x00008048 WAIT_UNTIL_POLL_MASK 0x0000804c WAIT_UNTIL_POLL_REFDATA +0x000084FC CP_STRMOUT_CNTL +0x000085F0 CP_COHER_CNTL +0x000085F4 CP_COHER_SIZE 0x000088B0 VGT_VTX_VECT_EJECT_REG 0x000088C4 VGT_CACHE_INVALIDATION 0x000088D4 VGT_GS_VERTEX_REUSE @@ -93,7 +96,6 @@ evergreen 0x9400 0x0002802C DB_DEPTH_CLEAR 0x00028030 PA_SC_SCREEN_SCISSOR_TL 0x00028034 PA_SC_SCREEN_SCISSOR_BR -0x0002805C DB_DEPTH_SLICE 0x00028140 SQ_ALU_CONST_BUFFER_SIZE_PS_0 0x00028144 SQ_ALU_CONST_BUFFER_SIZE_PS_1 0x00028148 SQ_ALU_CONST_BUFFER_SIZE_PS_2 @@ -222,7 +224,6 @@ evergreen 0x9400 0x00028344 PA_SC_VPORT_ZMAX_14 0x00028348 PA_SC_VPORT_ZMIN_15 0x0002834C PA_SC_VPORT_ZMAX_15 -0x00028350 SX_MISC 0x00028354 SX_SURFACE_SYNC 0x00028380 SQ_VTX_SEMANTIC_0 0x00028384 SQ_VTX_SEMANTIC_1 @@ -518,10 +519,16 @@ evergreen 0x9400 0x00028AA4 VGT_INSTANCE_STEP_RATE_1 0x00028AB4 VGT_REUSE_OFF 0x00028AB8 VGT_VTX_CNT_EN -0x00028ABC DB_HTILE_SURFACE 0x00028AC0 DB_SRESULTS_COMPARE_STATE0 0x00028AC4 DB_SRESULTS_COMPARE_STATE1 0x00028AC8 DB_PRELOAD_CONTROL +0x00028AD4 VGT_STRMOUT_VTX_STRIDE_0 +0x00028AE4 VGT_STRMOUT_VTX_STRIDE_1 +0x00028AF4 VGT_STRMOUT_VTX_STRIDE_2 +0x00028B04 VGT_STRMOUT_VTX_STRIDE_3 +0x00028B28 VGT_STRMOUT_DRAW_OPAQUE_OFFSET +0x00028B2C VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE +0x00028B30 VGT_STRMOUT_DRAW_OPAQUE_VERTEX_STRIDE 0x00028B38 VGT_GS_MAX_VERT_OUT 0x00028B54 VGT_SHADER_STAGES_EN 0x00028B58 VGT_LS_HS_CONFIG @@ -554,6 +561,18 @@ evergreen 0x9400 0x00028C34 PA_SC_AA_SAMPLE_LOCS_6 0x00028C38 PA_SC_AA_SAMPLE_LOCS_7 0x00028C3C PA_SC_AA_MASK +0x00028C78 CB_COLOR0_DIM +0x00028CB4 CB_COLOR1_DIM +0x00028CF0 CB_COLOR2_DIM +0x00028D2C CB_COLOR3_DIM +0x00028D68 CB_COLOR4_DIM +0x00028DA4 CB_COLOR5_DIM +0x00028DE0 CB_COLOR6_DIM +0x00028E1C CB_COLOR7_DIM +0x00028E58 CB_COLOR8_DIM +0x00028E74 CB_COLOR9_DIM +0x00028E90 CB_COLOR10_DIM +0x00028EAC CB_COLOR11_DIM 0x00028C8C CB_COLOR0_CLEAR_WORD0 0x00028C90 CB_COLOR0_CLEAR_WORD1 0x00028C94 CB_COLOR0_CLEAR_WORD2 diff --git a/drivers/gpu/drm/radeon/reg_srcs/r600 b/drivers/gpu/drm/radeon/reg_srcs/r600 index 0380c5c..5e659b0 100644 --- a/drivers/gpu/drm/radeon/reg_srcs/r600 +++ b/drivers/gpu/drm/radeon/reg_srcs/r600 @@ -3,6 +3,9 @@ r600 0x9400 0x00028230 R7xx_PA_SC_EDGERULE 0x000286C8 R7xx_SPI_THREAD_GROUPING 0x00008D8C R7xx_SQ_DYN_GPR_CNTL_PS_FLUSH_REQ +0x00008490 CP_STRMOUT_CNTL +0x000085F0 CP_COHER_CNTL +0x000085F4 CP_COHER_SIZE 0x000088C4 VGT_CACHE_INVALIDATION 0x00028A50 VGT_ENHANCE 0x000088CC VGT_ES_PER_GS @@ -38,6 +41,13 @@ r600 0x9400 0x00028AB4 VGT_REUSE_OFF 0x00028AB8 VGT_VTX_CNT_EN 0x000088B0 VGT_VTX_VECT_EJECT_REG +0x00028AD4 VGT_STRMOUT_VTX_STRIDE_0 +0x00028AE4 VGT_STRMOUT_VTX_STRIDE_1 +0x00028AF4 VGT_STRMOUT_VTX_STRIDE_2 +0x00028B04 VGT_STRMOUT_VTX_STRIDE_3 +0x00028B28 VGT_STRMOUT_DRAW_OPAQUE_OFFSET +0x00028B2C VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE +0x00028B30 VGT_STRMOUT_DRAW_OPAQUE_VERTEX_STRIDE 0x00028810 PA_CL_CLIP_CNTL 0x00008A14 PA_CL_ENHANCE 0x00028C14 PA_CL_GB_HORZ_CLIP_ADJ @@ -428,7 +438,7 @@ r600 0x9400 0x00028638 SPI_VS_OUT_ID_9 0x00028438 SX_ALPHA_REF 0x00028410 SX_ALPHA_TEST_CONTROL -0x00028350 SX_MISC +0x00028354 SX_SURFACE_SYNC 0x00009014 SX_MEMORY_EXPORT_SIZE 0x00009604 TC_INVALIDATE 0x00009400 TD_FILTER4 @@ -703,7 +713,6 @@ r600 0x9400 0x0000A710 TD_VS_SAMPLER17_BORDER_RED 0x00009508 TA_CNTL_AUX 0x0002802C DB_DEPTH_CLEAR -0x00028D24 DB_HTILE_SURFACE 0x00028D34 DB_PREFETCH_LIMIT 0x00028D30 DB_PRELOAD_CONTROL 0x00028D0C DB_RENDER_CONTROL @@ -743,14 +752,6 @@ r600 0x9400 0x00028114 CB_COLOR5_MASK 0x00028118 CB_COLOR6_MASK 0x0002811C CB_COLOR7_MASK -0x00028080 CB_COLOR0_VIEW -0x00028084 CB_COLOR1_VIEW -0x00028088 CB_COLOR2_VIEW -0x0002808C CB_COLOR3_VIEW -0x00028090 CB_COLOR4_VIEW -0x00028094 CB_COLOR5_VIEW -0x00028098 CB_COLOR6_VIEW -0x0002809C CB_COLOR7_VIEW 0x00028808 CB_COLOR_CONTROL 0x0002842C CB_FOG_BLUE 0x00028428 CB_FOG_GREEN diff --git a/drivers/gpu/drm/radeon/rs400.c b/drivers/gpu/drm/radeon/rs400.c index 06b90c8..4cf381b 100644 --- a/drivers/gpu/drm/radeon/rs400.c +++ b/drivers/gpu/drm/radeon/rs400.c @@ -410,6 +410,12 @@ static int rs400_startup(struct radeon_device *rdev) if (r) return r; + r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r); + return r; + } + /* Enable IRQ */ r100_irq_set(rdev); rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL); @@ -419,16 +425,25 @@ static int rs400_startup(struct radeon_device *rdev) dev_err(rdev->dev, "failed initializing CP (%d).\n", r); return r; } - r = r100_ib_init(rdev); + + r = radeon_ib_pool_start(rdev); + if (r) + return r; + + r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); if (r) { - dev_err(rdev->dev, "failed initializing IB (%d).\n", r); + dev_err(rdev->dev, "failed testing IB (%d).\n", r); + rdev->accel_working = false; return r; } + return 0; } int rs400_resume(struct radeon_device *rdev) { + int r; + /* Make sur GART are not working */ rs400_gart_disable(rdev); /* Resume clock before doing reset */ @@ -447,11 +462,18 @@ int rs400_resume(struct radeon_device *rdev) r300_clock_startup(rdev); /* Initialize surface registers */ radeon_surface_init(rdev); - return rs400_startup(rdev); + + rdev->accel_working = true; + r = rs400_startup(rdev); + if (r) { + rdev->accel_working = false; + } + return r; } int rs400_suspend(struct radeon_device *rdev) { + radeon_ib_pool_suspend(rdev); r100_cp_disable(rdev); radeon_wb_disable(rdev); r100_irq_disable(rdev); @@ -530,7 +552,14 @@ int rs400_init(struct radeon_device *rdev) if (r) return r; r300_set_reg_safe(rdev); + + r = radeon_ib_pool_init(rdev); rdev->accel_working = true; + if (r) { + dev_err(rdev->dev, "IB initialization failed (%d).\n", r); + rdev->accel_working = false; + } + r = rs400_startup(rdev); if (r) { /* Somethings want wront with the accel init stop accel */ diff --git a/drivers/gpu/drm/radeon/rs600.c b/drivers/gpu/drm/radeon/rs600.c index ee898e9..d25cf86 100644 --- a/drivers/gpu/drm/radeon/rs600.c +++ b/drivers/gpu/drm/radeon/rs600.c @@ -46,6 +46,25 @@ void rs600_gpu_init(struct radeon_device *rdev); int rs600_mc_wait_for_idle(struct radeon_device *rdev); +void avivo_wait_for_vblank(struct radeon_device *rdev, int crtc) +{ + struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc]; + int i; + + if (RREG32(AVIVO_D1CRTC_CONTROL + radeon_crtc->crtc_offset) & AVIVO_CRTC_EN) { + for (i = 0; i < rdev->usec_timeout; i++) { + if (!(RREG32(AVIVO_D1CRTC_STATUS + radeon_crtc->crtc_offset) & AVIVO_D1CRTC_V_BLANK)) + break; + udelay(1); + } + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32(AVIVO_D1CRTC_STATUS + radeon_crtc->crtc_offset) & AVIVO_D1CRTC_V_BLANK) + break; + udelay(1); + } + } +} + void rs600_pre_page_flip(struct radeon_device *rdev, int crtc) { /* enable the pflip int */ @@ -175,7 +194,7 @@ void rs600_pm_misc(struct radeon_device *rdev) /* set pcie lanes */ if ((rdev->flags & RADEON_IS_PCIE) && !(rdev->flags & RADEON_IS_IGP) && - rdev->asic->set_pcie_lanes && + rdev->asic->pm.set_pcie_lanes && (ps->pcie_lanes != rdev->pm.power_state[rdev->pm.current_power_state_index].pcie_lanes)) { radeon_set_pcie_lanes(rdev, @@ -322,16 +341,6 @@ void rs600_hpd_fini(struct radeon_device *rdev) } } -void rs600_bm_disable(struct radeon_device *rdev) -{ - u16 tmp; - - /* disable bus mastering */ - pci_read_config_word(rdev->pdev, 0x4, &tmp); - pci_write_config_word(rdev->pdev, 0x4, tmp & 0xFFFB); - mdelay(1); -} - int rs600_asic_reset(struct radeon_device *rdev) { struct rv515_mc_save save; @@ -355,7 +364,8 @@ int rs600_asic_reset(struct radeon_device *rdev) WREG32(RADEON_CP_RB_CNTL, tmp); pci_save_state(rdev->pdev); /* disable bus mastering */ - rs600_bm_disable(rdev); + pci_clear_master(rdev->pdev); + mdelay(1); /* reset GA+VAP */ WREG32(R_0000F0_RBBM_SOFT_RESET, S_0000F0_SOFT_RESET_VAP(1) | S_0000F0_SOFT_RESET_GA(1)); @@ -549,7 +559,7 @@ int rs600_irq_set(struct radeon_device *rdev) WREG32(R_000040_GEN_INT_CNTL, 0); return -EINVAL; } - if (rdev->irq.sw_int) { + if (rdev->irq.sw_int[RADEON_RING_TYPE_GFX_INDEX]) { tmp |= S_000040_SW_INT_EN(1); } if (rdev->irq.gui_idle) { @@ -642,7 +652,7 @@ int rs600_irq_process(struct radeon_device *rdev) while (status || rdev->irq.stat_regs.r500.disp_int) { /* SW interrupt */ if (G_000044_SW_INT(status)) { - radeon_fence_process(rdev); + radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX); } /* GUI idle */ if (G_000040_GUI_IDLE(status)) { @@ -847,6 +857,12 @@ static int rs600_startup(struct radeon_device *rdev) if (r) return r; + r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r); + return r; + } + /* Enable IRQ */ rs600_irq_set(rdev); rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL); @@ -856,15 +872,21 @@ static int rs600_startup(struct radeon_device *rdev) dev_err(rdev->dev, "failed initializing CP (%d).\n", r); return r; } - r = r100_ib_init(rdev); + + r = r600_audio_init(rdev); if (r) { - dev_err(rdev->dev, "failed initializing IB (%d).\n", r); + dev_err(rdev->dev, "failed initializing audio\n"); return r; } - r = r600_audio_init(rdev); + r = radeon_ib_pool_start(rdev); + if (r) + return r; + + r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); if (r) { - dev_err(rdev->dev, "failed initializing audio\n"); + dev_err(rdev->dev, "failed testing IB (%d).\n", r); + rdev->accel_working = false; return r; } @@ -873,6 +895,8 @@ static int rs600_startup(struct radeon_device *rdev) int rs600_resume(struct radeon_device *rdev) { + int r; + /* Make sur GART are not working */ rs600_gart_disable(rdev); /* Resume clock before doing reset */ @@ -889,11 +913,18 @@ int rs600_resume(struct radeon_device *rdev) rv515_clock_startup(rdev); /* Initialize surface registers */ radeon_surface_init(rdev); - return rs600_startup(rdev); + + rdev->accel_working = true; + r = rs600_startup(rdev); + if (r) { + rdev->accel_working = false; + } + return r; } int rs600_suspend(struct radeon_device *rdev) { + radeon_ib_pool_suspend(rdev); r600_audio_fini(rdev); r100_cp_disable(rdev); radeon_wb_disable(rdev); @@ -974,7 +1005,14 @@ int rs600_init(struct radeon_device *rdev) if (r) return r; rs600_set_safe_registers(rdev); + + r = radeon_ib_pool_init(rdev); rdev->accel_working = true; + if (r) { + dev_err(rdev->dev, "IB initialization failed (%d).\n", r); + rdev->accel_working = false; + } + r = rs600_startup(rdev); if (r) { /* Somethings want wront with the accel init stop accel */ diff --git a/drivers/gpu/drm/radeon/rs690.c b/drivers/gpu/drm/radeon/rs690.c index a9049ed..f2c3b9d 100644 --- a/drivers/gpu/drm/radeon/rs690.c +++ b/drivers/gpu/drm/radeon/rs690.c @@ -31,7 +31,7 @@ #include "atom.h" #include "rs690d.h" -static int rs690_mc_wait_for_idle(struct radeon_device *rdev) +int rs690_mc_wait_for_idle(struct radeon_device *rdev) { unsigned i; uint32_t tmp; @@ -621,6 +621,12 @@ static int rs690_startup(struct radeon_device *rdev) if (r) return r; + r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r); + return r; + } + /* Enable IRQ */ rs600_irq_set(rdev); rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL); @@ -630,15 +636,21 @@ static int rs690_startup(struct radeon_device *rdev) dev_err(rdev->dev, "failed initializing CP (%d).\n", r); return r; } - r = r100_ib_init(rdev); + + r = r600_audio_init(rdev); if (r) { - dev_err(rdev->dev, "failed initializing IB (%d).\n", r); + dev_err(rdev->dev, "failed initializing audio\n"); return r; } - r = r600_audio_init(rdev); + r = radeon_ib_pool_start(rdev); + if (r) + return r; + + r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); if (r) { - dev_err(rdev->dev, "failed initializing audio\n"); + dev_err(rdev->dev, "failed testing IB (%d).\n", r); + rdev->accel_working = false; return r; } @@ -647,6 +659,8 @@ static int rs690_startup(struct radeon_device *rdev) int rs690_resume(struct radeon_device *rdev) { + int r; + /* Make sur GART are not working */ rs400_gart_disable(rdev); /* Resume clock before doing reset */ @@ -663,11 +677,18 @@ int rs690_resume(struct radeon_device *rdev) rv515_clock_startup(rdev); /* Initialize surface registers */ radeon_surface_init(rdev); - return rs690_startup(rdev); + + rdev->accel_working = true; + r = rs690_startup(rdev); + if (r) { + rdev->accel_working = false; + } + return r; } int rs690_suspend(struct radeon_device *rdev) { + radeon_ib_pool_suspend(rdev); r600_audio_fini(rdev); r100_cp_disable(rdev); radeon_wb_disable(rdev); @@ -749,7 +770,14 @@ int rs690_init(struct radeon_device *rdev) if (r) return r; rs600_set_safe_registers(rdev); + + r = radeon_ib_pool_init(rdev); rdev->accel_working = true; + if (r) { + dev_err(rdev->dev, "IB initialization failed (%d).\n", r); + rdev->accel_working = false; + } + r = rs690_startup(rdev); if (r) { /* Somethings want wront with the accel init stop accel */ diff --git a/drivers/gpu/drm/radeon/rv515.c b/drivers/gpu/drm/radeon/rv515.c index d5f45b4..43af363 100644 --- a/drivers/gpu/drm/radeon/rv515.c +++ b/drivers/gpu/drm/radeon/rv515.c @@ -53,46 +53,46 @@ void rv515_debugfs(struct radeon_device *rdev) } } -void rv515_ring_start(struct radeon_device *rdev) +void rv515_ring_start(struct radeon_device *rdev, struct radeon_ring *ring) { int r; - r = radeon_ring_lock(rdev, 64); + r = radeon_ring_lock(rdev, ring, 64); if (r) { return; } - radeon_ring_write(rdev, PACKET0(ISYNC_CNTL, 0)); - radeon_ring_write(rdev, + radeon_ring_write(ring, PACKET0(ISYNC_CNTL, 0)); + radeon_ring_write(ring, ISYNC_ANY2D_IDLE3D | ISYNC_ANY3D_IDLE2D | ISYNC_WAIT_IDLEGUI | ISYNC_CPSCRATCH_IDLEGUI); - radeon_ring_write(rdev, PACKET0(WAIT_UNTIL, 0)); - radeon_ring_write(rdev, WAIT_2D_IDLECLEAN | WAIT_3D_IDLECLEAN); - radeon_ring_write(rdev, PACKET0(R300_DST_PIPE_CONFIG, 0)); - radeon_ring_write(rdev, R300_PIPE_AUTO_CONFIG); - radeon_ring_write(rdev, PACKET0(GB_SELECT, 0)); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, PACKET0(GB_ENABLE, 0)); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, PACKET0(R500_SU_REG_DEST, 0)); - radeon_ring_write(rdev, (1 << rdev->num_gb_pipes) - 1); - radeon_ring_write(rdev, PACKET0(VAP_INDEX_OFFSET, 0)); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, PACKET0(RB3D_DSTCACHE_CTLSTAT, 0)); - radeon_ring_write(rdev, RB3D_DC_FLUSH | RB3D_DC_FREE); - radeon_ring_write(rdev, PACKET0(ZB_ZCACHE_CTLSTAT, 0)); - radeon_ring_write(rdev, ZC_FLUSH | ZC_FREE); - radeon_ring_write(rdev, PACKET0(WAIT_UNTIL, 0)); - radeon_ring_write(rdev, WAIT_2D_IDLECLEAN | WAIT_3D_IDLECLEAN); - radeon_ring_write(rdev, PACKET0(GB_AA_CONFIG, 0)); - radeon_ring_write(rdev, 0); - radeon_ring_write(rdev, PACKET0(RB3D_DSTCACHE_CTLSTAT, 0)); - radeon_ring_write(rdev, RB3D_DC_FLUSH | RB3D_DC_FREE); - radeon_ring_write(rdev, PACKET0(ZB_ZCACHE_CTLSTAT, 0)); - radeon_ring_write(rdev, ZC_FLUSH | ZC_FREE); - radeon_ring_write(rdev, PACKET0(GB_MSPOS0, 0)); - radeon_ring_write(rdev, + radeon_ring_write(ring, PACKET0(WAIT_UNTIL, 0)); + radeon_ring_write(ring, WAIT_2D_IDLECLEAN | WAIT_3D_IDLECLEAN); + radeon_ring_write(ring, PACKET0(R300_DST_PIPE_CONFIG, 0)); + radeon_ring_write(ring, R300_PIPE_AUTO_CONFIG); + radeon_ring_write(ring, PACKET0(GB_SELECT, 0)); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, PACKET0(GB_ENABLE, 0)); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, PACKET0(R500_SU_REG_DEST, 0)); + radeon_ring_write(ring, (1 << rdev->num_gb_pipes) - 1); + radeon_ring_write(ring, PACKET0(VAP_INDEX_OFFSET, 0)); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, PACKET0(RB3D_DSTCACHE_CTLSTAT, 0)); + radeon_ring_write(ring, RB3D_DC_FLUSH | RB3D_DC_FREE); + radeon_ring_write(ring, PACKET0(ZB_ZCACHE_CTLSTAT, 0)); + radeon_ring_write(ring, ZC_FLUSH | ZC_FREE); + radeon_ring_write(ring, PACKET0(WAIT_UNTIL, 0)); + radeon_ring_write(ring, WAIT_2D_IDLECLEAN | WAIT_3D_IDLECLEAN); + radeon_ring_write(ring, PACKET0(GB_AA_CONFIG, 0)); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, PACKET0(RB3D_DSTCACHE_CTLSTAT, 0)); + radeon_ring_write(ring, RB3D_DC_FLUSH | RB3D_DC_FREE); + radeon_ring_write(ring, PACKET0(ZB_ZCACHE_CTLSTAT, 0)); + radeon_ring_write(ring, ZC_FLUSH | ZC_FREE); + radeon_ring_write(ring, PACKET0(GB_MSPOS0, 0)); + radeon_ring_write(ring, ((6 << MS_X0_SHIFT) | (6 << MS_Y0_SHIFT) | (6 << MS_X1_SHIFT) | @@ -101,8 +101,8 @@ void rv515_ring_start(struct radeon_device *rdev) (6 << MS_Y2_SHIFT) | (6 << MSBD0_Y_SHIFT) | (6 << MSBD0_X_SHIFT))); - radeon_ring_write(rdev, PACKET0(GB_MSPOS1, 0)); - radeon_ring_write(rdev, + radeon_ring_write(ring, PACKET0(GB_MSPOS1, 0)); + radeon_ring_write(ring, ((6 << MS_X3_SHIFT) | (6 << MS_Y3_SHIFT) | (6 << MS_X4_SHIFT) | @@ -110,15 +110,15 @@ void rv515_ring_start(struct radeon_device *rdev) (6 << MS_X5_SHIFT) | (6 << MS_Y5_SHIFT) | (6 << MSBD1_SHIFT))); - radeon_ring_write(rdev, PACKET0(GA_ENHANCE, 0)); - radeon_ring_write(rdev, GA_DEADLOCK_CNTL | GA_FASTSYNC_CNTL); - radeon_ring_write(rdev, PACKET0(GA_POLY_MODE, 0)); - radeon_ring_write(rdev, FRONT_PTYPE_TRIANGE | BACK_PTYPE_TRIANGE); - radeon_ring_write(rdev, PACKET0(GA_ROUND_MODE, 0)); - radeon_ring_write(rdev, GEOMETRY_ROUND_NEAREST | COLOR_ROUND_NEAREST); - radeon_ring_write(rdev, PACKET0(0x20C8, 0)); - radeon_ring_write(rdev, 0); - radeon_ring_unlock_commit(rdev); + radeon_ring_write(ring, PACKET0(GA_ENHANCE, 0)); + radeon_ring_write(ring, GA_DEADLOCK_CNTL | GA_FASTSYNC_CNTL); + radeon_ring_write(ring, PACKET0(GA_POLY_MODE, 0)); + radeon_ring_write(ring, FRONT_PTYPE_TRIANGE | BACK_PTYPE_TRIANGE); + radeon_ring_write(ring, PACKET0(GA_ROUND_MODE, 0)); + radeon_ring_write(ring, GEOMETRY_ROUND_NEAREST | COLOR_ROUND_NEAREST); + radeon_ring_write(ring, PACKET0(0x20C8, 0)); + radeon_ring_write(ring, 0); + radeon_ring_unlock_commit(rdev, ring); } int rv515_mc_wait_for_idle(struct radeon_device *rdev) @@ -149,7 +149,7 @@ void rv515_gpu_init(struct radeon_device *rdev) if (r100_gui_wait_for_idle(rdev)) { printk(KERN_WARNING "Failed to wait GUI idle while " - "reseting GPU. Bad things might happen.\n"); + "resetting GPU. Bad things might happen.\n"); } rv515_vga_render_disable(rdev); r420_pipes_init(rdev); @@ -161,7 +161,7 @@ void rv515_gpu_init(struct radeon_device *rdev) WREG32_PLL(0x000D, tmp); if (r100_gui_wait_for_idle(rdev)) { printk(KERN_WARNING "Failed to wait GUI idle while " - "reseting GPU. Bad things might happen.\n"); + "resetting GPU. Bad things might happen.\n"); } if (rv515_mc_wait_for_idle(rdev)) { printk(KERN_WARNING "Failed to wait MC idle while " @@ -379,6 +379,12 @@ static int rv515_startup(struct radeon_device *rdev) if (r) return r; + r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r); + return r; + } + /* Enable IRQ */ rs600_irq_set(rdev); rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL); @@ -388,9 +394,15 @@ static int rv515_startup(struct radeon_device *rdev) dev_err(rdev->dev, "failed initializing CP (%d).\n", r); return r; } - r = r100_ib_init(rdev); + + r = radeon_ib_pool_start(rdev); + if (r) + return r; + + r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); if (r) { - dev_err(rdev->dev, "failed initializing IB (%d).\n", r); + dev_err(rdev->dev, "failed testing IB (%d).\n", r); + rdev->accel_working = false; return r; } return 0; @@ -398,6 +410,8 @@ static int rv515_startup(struct radeon_device *rdev) int rv515_resume(struct radeon_device *rdev) { + int r; + /* Make sur GART are not working */ if (rdev->flags & RADEON_IS_PCIE) rv370_pcie_gart_disable(rdev); @@ -415,7 +429,13 @@ int rv515_resume(struct radeon_device *rdev) rv515_clock_startup(rdev); /* Initialize surface registers */ radeon_surface_init(rdev); - return rv515_startup(rdev); + + rdev->accel_working = true; + r = rv515_startup(rdev); + if (r) { + rdev->accel_working = false; + } + return r; } int rv515_suspend(struct radeon_device *rdev) @@ -511,7 +531,14 @@ int rv515_init(struct radeon_device *rdev) if (r) return r; rv515_set_safe_registers(rdev); + + r = radeon_ib_pool_init(rdev); rdev->accel_working = true; + if (r) { + dev_err(rdev->dev, "IB initialization failed (%d).\n", r); + rdev->accel_working = false; + } + r = rv515_startup(rdev); if (r) { /* Somethings want wront with the accel init stop accel */ diff --git a/drivers/gpu/drm/radeon/rv770.c b/drivers/gpu/drm/radeon/rv770.c index cc79449..591040b 100644 --- a/drivers/gpu/drm/radeon/rv770.c +++ b/drivers/gpu/drm/radeon/rv770.c @@ -359,7 +359,7 @@ static int rv770_cp_load_microcode(struct radeon_device *rdev) void r700_cp_fini(struct radeon_device *rdev) { r700_cp_stop(rdev); - radeon_ring_fini(rdev); + radeon_ring_fini(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); } /* @@ -978,7 +978,7 @@ void r700_vram_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc) } if (rdev->flags & RADEON_IS_AGP) { size_bf = mc->gtt_start; - size_af = 0xFFFFFFFF - mc->gtt_end + 1; + size_af = 0xFFFFFFFF - mc->gtt_end; if (size_bf > size_af) { if (mc->mc_vram_size > size_bf) { dev_warn(rdev->dev, "limiting VRAM\n"); @@ -992,7 +992,7 @@ void r700_vram_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc) mc->real_vram_size = size_af; mc->mc_vram_size = size_af; } - mc->vram_start = mc->gtt_end; + mc->vram_start = mc->gtt_end + 1; } mc->vram_end = mc->vram_start + mc->mc_vram_size - 1; dev_info(rdev->dev, "VRAM: %lluM 0x%08llX - 0x%08llX (%lluM used)\n", @@ -1052,6 +1052,7 @@ int rv770_mc_init(struct radeon_device *rdev) static int rv770_startup(struct radeon_device *rdev) { + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; int r; /* enable pcie gen2 link */ @@ -1082,7 +1083,7 @@ static int rv770_startup(struct radeon_device *rdev) r = r600_blit_init(rdev); if (r) { r600_blit_fini(rdev); - rdev->asic->copy = NULL; + rdev->asic->copy.copy = NULL; dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r); } @@ -1091,6 +1092,12 @@ static int rv770_startup(struct radeon_device *rdev) if (r) return r; + r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r); + return r; + } + /* Enable IRQ */ r = r600_irq_init(rdev); if (r) { @@ -1100,7 +1107,9 @@ static int rv770_startup(struct radeon_device *rdev) } r600_irq_set(rdev); - r = radeon_ring_init(rdev, rdev->cp.ring_size); + r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET, + R600_CP_RB_RPTR, R600_CP_RB_WPTR, + 0, 0xfffff, RADEON_CP_PACKET2); if (r) return r; r = rv770_cp_load_microcode(rdev); @@ -1110,6 +1119,17 @@ static int rv770_startup(struct radeon_device *rdev) if (r) return r; + r = radeon_ib_pool_start(rdev); + if (r) + return r; + + r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); + if (r) { + dev_err(rdev->dev, "IB test failed (%d).\n", r); + rdev->accel_working = false; + return r; + } + return 0; } @@ -1124,15 +1144,11 @@ int rv770_resume(struct radeon_device *rdev) /* post card */ atom_asic_init(rdev->mode_info.atom_context); + rdev->accel_working = true; r = rv770_startup(rdev); if (r) { DRM_ERROR("r600 startup failed on resume\n"); - return r; - } - - r = r600_ib_test(rdev); - if (r) { - DRM_ERROR("radeon: failed testing IB (%d).\n", r); + rdev->accel_working = false; return r; } @@ -1149,13 +1165,14 @@ int rv770_resume(struct radeon_device *rdev) int rv770_suspend(struct radeon_device *rdev) { r600_audio_fini(rdev); + radeon_ib_pool_suspend(rdev); + r600_blit_suspend(rdev); /* FIXME: we should wait for ring to be empty */ r700_cp_stop(rdev); - rdev->cp.ready = false; + rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false; r600_irq_suspend(rdev); radeon_wb_disable(rdev); rv770_pcie_gart_disable(rdev); - r600_blit_suspend(rdev); return 0; } @@ -1224,8 +1241,8 @@ int rv770_init(struct radeon_device *rdev) if (r) return r; - rdev->cp.ring_obj = NULL; - r600_ring_init(rdev, 1024 * 1024); + rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ring_obj = NULL; + r600_ring_init(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX], 1024 * 1024); rdev->ih.ring_obj = NULL; r600_ih_ring_init(rdev, 64 * 1024); @@ -1234,30 +1251,24 @@ int rv770_init(struct radeon_device *rdev) if (r) return r; + r = radeon_ib_pool_init(rdev); rdev->accel_working = true; + if (r) { + dev_err(rdev->dev, "IB initialization failed (%d).\n", r); + rdev->accel_working = false; + } + r = rv770_startup(rdev); if (r) { dev_err(rdev->dev, "disabling GPU acceleration\n"); r700_cp_fini(rdev); r600_irq_fini(rdev); radeon_wb_fini(rdev); + r100_ib_fini(rdev); radeon_irq_kms_fini(rdev); rv770_pcie_gart_fini(rdev); rdev->accel_working = false; } - if (rdev->accel_working) { - r = radeon_ib_pool_init(rdev); - if (r) { - dev_err(rdev->dev, "IB initialization failed (%d).\n", r); - rdev->accel_working = false; - } else { - r = r600_ib_test(rdev); - if (r) { - dev_err(rdev->dev, "IB test failed (%d).\n", r); - rdev->accel_working = false; - } - } - } r = r600_audio_init(rdev); if (r) { @@ -1274,11 +1285,12 @@ void rv770_fini(struct radeon_device *rdev) r700_cp_fini(rdev); r600_irq_fini(rdev); radeon_wb_fini(rdev); - radeon_ib_pool_fini(rdev); + r100_ib_fini(rdev); radeon_irq_kms_fini(rdev); rv770_pcie_gart_fini(rdev); r600_vram_scratch_fini(rdev); radeon_gem_fini(rdev); + radeon_semaphore_driver_fini(rdev); radeon_fence_driver_fini(rdev); radeon_agp_fini(rdev); radeon_bo_fini(rdev); diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c new file mode 100644 index 0000000..1197f21 --- /dev/null +++ b/drivers/gpu/drm/radeon/si.c @@ -0,0 +1,4128 @@ +/* + * Copyright 2011 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Alex Deucher + */ +#include +#include +#include +#include +#include "drmP.h" +#include "radeon.h" +#include "radeon_asic.h" +#include "radeon_drm.h" +#include "sid.h" +#include "atom.h" +#include "si_blit_shaders.h" + +#define SI_PFP_UCODE_SIZE 2144 +#define SI_PM4_UCODE_SIZE 2144 +#define SI_CE_UCODE_SIZE 2144 +#define SI_RLC_UCODE_SIZE 2048 +#define SI_MC_UCODE_SIZE 7769 + +MODULE_FIRMWARE("radeon/TAHITI_pfp.bin"); +MODULE_FIRMWARE("radeon/TAHITI_me.bin"); +MODULE_FIRMWARE("radeon/TAHITI_ce.bin"); +MODULE_FIRMWARE("radeon/TAHITI_mc.bin"); +MODULE_FIRMWARE("radeon/TAHITI_rlc.bin"); +MODULE_FIRMWARE("radeon/PITCAIRN_pfp.bin"); +MODULE_FIRMWARE("radeon/PITCAIRN_me.bin"); +MODULE_FIRMWARE("radeon/PITCAIRN_ce.bin"); +MODULE_FIRMWARE("radeon/PITCAIRN_mc.bin"); +MODULE_FIRMWARE("radeon/PITCAIRN_rlc.bin"); +MODULE_FIRMWARE("radeon/VERDE_pfp.bin"); +MODULE_FIRMWARE("radeon/VERDE_me.bin"); +MODULE_FIRMWARE("radeon/VERDE_ce.bin"); +MODULE_FIRMWARE("radeon/VERDE_mc.bin"); +MODULE_FIRMWARE("radeon/VERDE_rlc.bin"); + +extern int r600_ih_ring_alloc(struct radeon_device *rdev); +extern void r600_ih_ring_fini(struct radeon_device *rdev); +extern void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev); +extern void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *save); +extern void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *save); +extern u32 evergreen_get_number_of_dram_channels(struct radeon_device *rdev); + +/* get temperature in millidegrees */ +int si_get_temp(struct radeon_device *rdev) +{ + u32 temp; + int actual_temp = 0; + + temp = (RREG32(CG_MULT_THERMAL_STATUS) & CTF_TEMP_MASK) >> + CTF_TEMP_SHIFT; + + if (temp & 0x200) + actual_temp = 255; + else + actual_temp = temp & 0x1ff; + + actual_temp = (actual_temp * 1000); + + return actual_temp; +} + +#define TAHITI_IO_MC_REGS_SIZE 36 + +static const u32 tahiti_io_mc_regs[TAHITI_IO_MC_REGS_SIZE][2] = { + {0x0000006f, 0x03044000}, + {0x00000070, 0x0480c018}, + {0x00000071, 0x00000040}, + {0x00000072, 0x01000000}, + {0x00000074, 0x000000ff}, + {0x00000075, 0x00143400}, + {0x00000076, 0x08ec0800}, + {0x00000077, 0x040000cc}, + {0x00000079, 0x00000000}, + {0x0000007a, 0x21000409}, + {0x0000007c, 0x00000000}, + {0x0000007d, 0xe8000000}, + {0x0000007e, 0x044408a8}, + {0x0000007f, 0x00000003}, + {0x00000080, 0x00000000}, + {0x00000081, 0x01000000}, + {0x00000082, 0x02000000}, + {0x00000083, 0x00000000}, + {0x00000084, 0xe3f3e4f4}, + {0x00000085, 0x00052024}, + {0x00000087, 0x00000000}, + {0x00000088, 0x66036603}, + {0x00000089, 0x01000000}, + {0x0000008b, 0x1c0a0000}, + {0x0000008c, 0xff010000}, + {0x0000008e, 0xffffefff}, + {0x0000008f, 0xfff3efff}, + {0x00000090, 0xfff3efbf}, + {0x00000094, 0x00101101}, + {0x00000095, 0x00000fff}, + {0x00000096, 0x00116fff}, + {0x00000097, 0x60010000}, + {0x00000098, 0x10010000}, + {0x00000099, 0x00006000}, + {0x0000009a, 0x00001000}, + {0x0000009f, 0x00a77400} +}; + +static const u32 pitcairn_io_mc_regs[TAHITI_IO_MC_REGS_SIZE][2] = { + {0x0000006f, 0x03044000}, + {0x00000070, 0x0480c018}, + {0x00000071, 0x00000040}, + {0x00000072, 0x01000000}, + {0x00000074, 0x000000ff}, + {0x00000075, 0x00143400}, + {0x00000076, 0x08ec0800}, + {0x00000077, 0x040000cc}, + {0x00000079, 0x00000000}, + {0x0000007a, 0x21000409}, + {0x0000007c, 0x00000000}, + {0x0000007d, 0xe8000000}, + {0x0000007e, 0x044408a8}, + {0x0000007f, 0x00000003}, + {0x00000080, 0x00000000}, + {0x00000081, 0x01000000}, + {0x00000082, 0x02000000}, + {0x00000083, 0x00000000}, + {0x00000084, 0xe3f3e4f4}, + {0x00000085, 0x00052024}, + {0x00000087, 0x00000000}, + {0x00000088, 0x66036603}, + {0x00000089, 0x01000000}, + {0x0000008b, 0x1c0a0000}, + {0x0000008c, 0xff010000}, + {0x0000008e, 0xffffefff}, + {0x0000008f, 0xfff3efff}, + {0x00000090, 0xfff3efbf}, + {0x00000094, 0x00101101}, + {0x00000095, 0x00000fff}, + {0x00000096, 0x00116fff}, + {0x00000097, 0x60010000}, + {0x00000098, 0x10010000}, + {0x00000099, 0x00006000}, + {0x0000009a, 0x00001000}, + {0x0000009f, 0x00a47400} +}; + +static const u32 verde_io_mc_regs[TAHITI_IO_MC_REGS_SIZE][2] = { + {0x0000006f, 0x03044000}, + {0x00000070, 0x0480c018}, + {0x00000071, 0x00000040}, + {0x00000072, 0x01000000}, + {0x00000074, 0x000000ff}, + {0x00000075, 0x00143400}, + {0x00000076, 0x08ec0800}, + {0x00000077, 0x040000cc}, + {0x00000079, 0x00000000}, + {0x0000007a, 0x21000409}, + {0x0000007c, 0x00000000}, + {0x0000007d, 0xe8000000}, + {0x0000007e, 0x044408a8}, + {0x0000007f, 0x00000003}, + {0x00000080, 0x00000000}, + {0x00000081, 0x01000000}, + {0x00000082, 0x02000000}, + {0x00000083, 0x00000000}, + {0x00000084, 0xe3f3e4f4}, + {0x00000085, 0x00052024}, + {0x00000087, 0x00000000}, + {0x00000088, 0x66036603}, + {0x00000089, 0x01000000}, + {0x0000008b, 0x1c0a0000}, + {0x0000008c, 0xff010000}, + {0x0000008e, 0xffffefff}, + {0x0000008f, 0xfff3efff}, + {0x00000090, 0xfff3efbf}, + {0x00000094, 0x00101101}, + {0x00000095, 0x00000fff}, + {0x00000096, 0x00116fff}, + {0x00000097, 0x60010000}, + {0x00000098, 0x10010000}, + {0x00000099, 0x00006000}, + {0x0000009a, 0x00001000}, + {0x0000009f, 0x00a37400} +}; + +/* ucode loading */ +static int si_mc_load_microcode(struct radeon_device *rdev) +{ + const __be32 *fw_data; + u32 running, blackout = 0; + u32 *io_mc_regs; + int i, ucode_size, regs_size; + + if (!rdev->mc_fw) + return -EINVAL; + + switch (rdev->family) { + case CHIP_TAHITI: + io_mc_regs = (u32 *)&tahiti_io_mc_regs; + ucode_size = SI_MC_UCODE_SIZE; + regs_size = TAHITI_IO_MC_REGS_SIZE; + break; + case CHIP_PITCAIRN: + io_mc_regs = (u32 *)&pitcairn_io_mc_regs; + ucode_size = SI_MC_UCODE_SIZE; + regs_size = TAHITI_IO_MC_REGS_SIZE; + break; + case CHIP_VERDE: + default: + io_mc_regs = (u32 *)&verde_io_mc_regs; + ucode_size = SI_MC_UCODE_SIZE; + regs_size = TAHITI_IO_MC_REGS_SIZE; + break; + } + + running = RREG32(MC_SEQ_SUP_CNTL) & RUN_MASK; + + if (running == 0) { + if (running) { + blackout = RREG32(MC_SHARED_BLACKOUT_CNTL); + WREG32(MC_SHARED_BLACKOUT_CNTL, blackout | 1); + } + + /* reset the engine and set to writable */ + WREG32(MC_SEQ_SUP_CNTL, 0x00000008); + WREG32(MC_SEQ_SUP_CNTL, 0x00000010); + + /* load mc io regs */ + for (i = 0; i < regs_size; i++) { + WREG32(MC_SEQ_IO_DEBUG_INDEX, io_mc_regs[(i << 1)]); + WREG32(MC_SEQ_IO_DEBUG_DATA, io_mc_regs[(i << 1) + 1]); + } + /* load the MC ucode */ + fw_data = (const __be32 *)rdev->mc_fw->data; + for (i = 0; i < ucode_size; i++) + WREG32(MC_SEQ_SUP_PGM, be32_to_cpup(fw_data++)); + + /* put the engine back into the active state */ + WREG32(MC_SEQ_SUP_CNTL, 0x00000008); + WREG32(MC_SEQ_SUP_CNTL, 0x00000004); + WREG32(MC_SEQ_SUP_CNTL, 0x00000001); + + /* wait for training to complete */ + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32(MC_SEQ_TRAIN_WAKEUP_CNTL) & TRAIN_DONE_D0) + break; + udelay(1); + } + for (i = 0; i < rdev->usec_timeout; i++) { + if (RREG32(MC_SEQ_TRAIN_WAKEUP_CNTL) & TRAIN_DONE_D1) + break; + udelay(1); + } + + if (running) + WREG32(MC_SHARED_BLACKOUT_CNTL, blackout); + } + + return 0; +} + +static int si_init_microcode(struct radeon_device *rdev) +{ + struct platform_device *pdev; + const char *chip_name; + const char *rlc_chip_name; + size_t pfp_req_size, me_req_size, ce_req_size, rlc_req_size, mc_req_size; + char fw_name[30]; + int err; + + DRM_DEBUG("\n"); + + pdev = platform_device_register_simple("radeon_cp", 0, NULL, 0); + err = IS_ERR(pdev); + if (err) { + printk(KERN_ERR "radeon_cp: Failed to register firmware\n"); + return -EINVAL; + } + + switch (rdev->family) { + case CHIP_TAHITI: + chip_name = "TAHITI"; + rlc_chip_name = "TAHITI"; + pfp_req_size = SI_PFP_UCODE_SIZE * 4; + me_req_size = SI_PM4_UCODE_SIZE * 4; + ce_req_size = SI_CE_UCODE_SIZE * 4; + rlc_req_size = SI_RLC_UCODE_SIZE * 4; + mc_req_size = SI_MC_UCODE_SIZE * 4; + break; + case CHIP_PITCAIRN: + chip_name = "PITCAIRN"; + rlc_chip_name = "PITCAIRN"; + pfp_req_size = SI_PFP_UCODE_SIZE * 4; + me_req_size = SI_PM4_UCODE_SIZE * 4; + ce_req_size = SI_CE_UCODE_SIZE * 4; + rlc_req_size = SI_RLC_UCODE_SIZE * 4; + mc_req_size = SI_MC_UCODE_SIZE * 4; + break; + case CHIP_VERDE: + chip_name = "VERDE"; + rlc_chip_name = "VERDE"; + pfp_req_size = SI_PFP_UCODE_SIZE * 4; + me_req_size = SI_PM4_UCODE_SIZE * 4; + ce_req_size = SI_CE_UCODE_SIZE * 4; + rlc_req_size = SI_RLC_UCODE_SIZE * 4; + mc_req_size = SI_MC_UCODE_SIZE * 4; + break; + default: BUG(); + } + + DRM_INFO("Loading %s Microcode\n", chip_name); + + snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name); + err = request_firmware(&rdev->pfp_fw, fw_name, &pdev->dev); + if (err) + goto out; + if (rdev->pfp_fw->size != pfp_req_size) { + printk(KERN_ERR + "si_cp: Bogus length %zu in firmware \"%s\"\n", + rdev->pfp_fw->size, fw_name); + err = -EINVAL; + goto out; + } + + snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name); + err = request_firmware(&rdev->me_fw, fw_name, &pdev->dev); + if (err) + goto out; + if (rdev->me_fw->size != me_req_size) { + printk(KERN_ERR + "si_cp: Bogus length %zu in firmware \"%s\"\n", + rdev->me_fw->size, fw_name); + err = -EINVAL; + } + + snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", chip_name); + err = request_firmware(&rdev->ce_fw, fw_name, &pdev->dev); + if (err) + goto out; + if (rdev->ce_fw->size != ce_req_size) { + printk(KERN_ERR + "si_cp: Bogus length %zu in firmware \"%s\"\n", + rdev->ce_fw->size, fw_name); + err = -EINVAL; + } + + snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", rlc_chip_name); + err = request_firmware(&rdev->rlc_fw, fw_name, &pdev->dev); + if (err) + goto out; + if (rdev->rlc_fw->size != rlc_req_size) { + printk(KERN_ERR + "si_rlc: Bogus length %zu in firmware \"%s\"\n", + rdev->rlc_fw->size, fw_name); + err = -EINVAL; + } + + snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name); + err = request_firmware(&rdev->mc_fw, fw_name, &pdev->dev); + if (err) + goto out; + if (rdev->mc_fw->size != mc_req_size) { + printk(KERN_ERR + "si_mc: Bogus length %zu in firmware \"%s\"\n", + rdev->mc_fw->size, fw_name); + err = -EINVAL; + } + +out: + platform_device_unregister(pdev); + + if (err) { + if (err != -EINVAL) + printk(KERN_ERR + "si_cp: Failed to load firmware \"%s\"\n", + fw_name); + release_firmware(rdev->pfp_fw); + rdev->pfp_fw = NULL; + release_firmware(rdev->me_fw); + rdev->me_fw = NULL; + release_firmware(rdev->ce_fw); + rdev->ce_fw = NULL; + release_firmware(rdev->rlc_fw); + rdev->rlc_fw = NULL; + release_firmware(rdev->mc_fw); + rdev->mc_fw = NULL; + } + return err; +} + +/* watermark setup */ +static u32 dce6_line_buffer_adjust(struct radeon_device *rdev, + struct radeon_crtc *radeon_crtc, + struct drm_display_mode *mode, + struct drm_display_mode *other_mode) +{ + u32 tmp; + /* + * Line Buffer Setup + * There are 3 line buffers, each one shared by 2 display controllers. + * DC_LB_MEMORY_SPLIT controls how that line buffer is shared between + * the display controllers. The paritioning is done via one of four + * preset allocations specified in bits 21:20: + * 0 - half lb + * 2 - whole lb, other crtc must be disabled + */ + /* this can get tricky if we have two large displays on a paired group + * of crtcs. Ideally for multiple large displays we'd assign them to + * non-linked crtcs for maximum line buffer allocation. + */ + if (radeon_crtc->base.enabled && mode) { + if (other_mode) + tmp = 0; /* 1/2 */ + else + tmp = 2; /* whole */ + } else + tmp = 0; + + WREG32(DC_LB_MEMORY_SPLIT + radeon_crtc->crtc_offset, + DC_LB_MEMORY_CONFIG(tmp)); + + if (radeon_crtc->base.enabled && mode) { + switch (tmp) { + case 0: + default: + return 4096 * 2; + case 2: + return 8192 * 2; + } + } + + /* controller not enabled, so no lb used */ + return 0; +} + +static u32 si_get_number_of_dram_channels(struct radeon_device *rdev) +{ + u32 tmp = RREG32(MC_SHARED_CHMAP); + + switch ((tmp & NOOFCHAN_MASK) >> NOOFCHAN_SHIFT) { + case 0: + default: + return 1; + case 1: + return 2; + case 2: + return 4; + case 3: + return 8; + case 4: + return 3; + case 5: + return 6; + case 6: + return 10; + case 7: + return 12; + case 8: + return 16; + } +} + +struct dce6_wm_params { + u32 dram_channels; /* number of dram channels */ + u32 yclk; /* bandwidth per dram data pin in kHz */ + u32 sclk; /* engine clock in kHz */ + u32 disp_clk; /* display clock in kHz */ + u32 src_width; /* viewport width */ + u32 active_time; /* active display time in ns */ + u32 blank_time; /* blank time in ns */ + bool interlaced; /* mode is interlaced */ + fixed20_12 vsc; /* vertical scale ratio */ + u32 num_heads; /* number of active crtcs */ + u32 bytes_per_pixel; /* bytes per pixel display + overlay */ + u32 lb_size; /* line buffer allocated to pipe */ + u32 vtaps; /* vertical scaler taps */ +}; + +static u32 dce6_dram_bandwidth(struct dce6_wm_params *wm) +{ + /* Calculate raw DRAM Bandwidth */ + fixed20_12 dram_efficiency; /* 0.7 */ + fixed20_12 yclk, dram_channels, bandwidth; + fixed20_12 a; + + a.full = dfixed_const(1000); + yclk.full = dfixed_const(wm->yclk); + yclk.full = dfixed_div(yclk, a); + dram_channels.full = dfixed_const(wm->dram_channels * 4); + a.full = dfixed_const(10); + dram_efficiency.full = dfixed_const(7); + dram_efficiency.full = dfixed_div(dram_efficiency, a); + bandwidth.full = dfixed_mul(dram_channels, yclk); + bandwidth.full = dfixed_mul(bandwidth, dram_efficiency); + + return dfixed_trunc(bandwidth); +} + +static u32 dce6_dram_bandwidth_for_display(struct dce6_wm_params *wm) +{ + /* Calculate DRAM Bandwidth and the part allocated to display. */ + fixed20_12 disp_dram_allocation; /* 0.3 to 0.7 */ + fixed20_12 yclk, dram_channels, bandwidth; + fixed20_12 a; + + a.full = dfixed_const(1000); + yclk.full = dfixed_const(wm->yclk); + yclk.full = dfixed_div(yclk, a); + dram_channels.full = dfixed_const(wm->dram_channels * 4); + a.full = dfixed_const(10); + disp_dram_allocation.full = dfixed_const(3); /* XXX worse case value 0.3 */ + disp_dram_allocation.full = dfixed_div(disp_dram_allocation, a); + bandwidth.full = dfixed_mul(dram_channels, yclk); + bandwidth.full = dfixed_mul(bandwidth, disp_dram_allocation); + + return dfixed_trunc(bandwidth); +} + +static u32 dce6_data_return_bandwidth(struct dce6_wm_params *wm) +{ + /* Calculate the display Data return Bandwidth */ + fixed20_12 return_efficiency; /* 0.8 */ + fixed20_12 sclk, bandwidth; + fixed20_12 a; + + a.full = dfixed_const(1000); + sclk.full = dfixed_const(wm->sclk); + sclk.full = dfixed_div(sclk, a); + a.full = dfixed_const(10); + return_efficiency.full = dfixed_const(8); + return_efficiency.full = dfixed_div(return_efficiency, a); + a.full = dfixed_const(32); + bandwidth.full = dfixed_mul(a, sclk); + bandwidth.full = dfixed_mul(bandwidth, return_efficiency); + + return dfixed_trunc(bandwidth); +} + +static u32 dce6_get_dmif_bytes_per_request(struct dce6_wm_params *wm) +{ + return 32; +} + +static u32 dce6_dmif_request_bandwidth(struct dce6_wm_params *wm) +{ + /* Calculate the DMIF Request Bandwidth */ + fixed20_12 disp_clk_request_efficiency; /* 0.8 */ + fixed20_12 disp_clk, sclk, bandwidth; + fixed20_12 a, b1, b2; + u32 min_bandwidth; + + a.full = dfixed_const(1000); + disp_clk.full = dfixed_const(wm->disp_clk); + disp_clk.full = dfixed_div(disp_clk, a); + a.full = dfixed_const(dce6_get_dmif_bytes_per_request(wm) / 2); + b1.full = dfixed_mul(a, disp_clk); + + a.full = dfixed_const(1000); + sclk.full = dfixed_const(wm->sclk); + sclk.full = dfixed_div(sclk, a); + a.full = dfixed_const(dce6_get_dmif_bytes_per_request(wm)); + b2.full = dfixed_mul(a, sclk); + + a.full = dfixed_const(10); + disp_clk_request_efficiency.full = dfixed_const(8); + disp_clk_request_efficiency.full = dfixed_div(disp_clk_request_efficiency, a); + + min_bandwidth = min(dfixed_trunc(b1), dfixed_trunc(b2)); + + a.full = dfixed_const(min_bandwidth); + bandwidth.full = dfixed_mul(a, disp_clk_request_efficiency); + + return dfixed_trunc(bandwidth); +} + +static u32 dce6_available_bandwidth(struct dce6_wm_params *wm) +{ + /* Calculate the Available bandwidth. Display can use this temporarily but not in average. */ + u32 dram_bandwidth = dce6_dram_bandwidth(wm); + u32 data_return_bandwidth = dce6_data_return_bandwidth(wm); + u32 dmif_req_bandwidth = dce6_dmif_request_bandwidth(wm); + + return min(dram_bandwidth, min(data_return_bandwidth, dmif_req_bandwidth)); +} + +static u32 dce6_average_bandwidth(struct dce6_wm_params *wm) +{ + /* Calculate the display mode Average Bandwidth + * DisplayMode should contain the source and destination dimensions, + * timing, etc. + */ + fixed20_12 bpp; + fixed20_12 line_time; + fixed20_12 src_width; + fixed20_12 bandwidth; + fixed20_12 a; + + a.full = dfixed_const(1000); + line_time.full = dfixed_const(wm->active_time + wm->blank_time); + line_time.full = dfixed_div(line_time, a); + bpp.full = dfixed_const(wm->bytes_per_pixel); + src_width.full = dfixed_const(wm->src_width); + bandwidth.full = dfixed_mul(src_width, bpp); + bandwidth.full = dfixed_mul(bandwidth, wm->vsc); + bandwidth.full = dfixed_div(bandwidth, line_time); + + return dfixed_trunc(bandwidth); +} + +static u32 dce6_latency_watermark(struct dce6_wm_params *wm) +{ + /* First calcualte the latency in ns */ + u32 mc_latency = 2000; /* 2000 ns. */ + u32 available_bandwidth = dce6_available_bandwidth(wm); + u32 worst_chunk_return_time = (512 * 8 * 1000) / available_bandwidth; + u32 cursor_line_pair_return_time = (128 * 4 * 1000) / available_bandwidth; + u32 dc_latency = 40000000 / wm->disp_clk; /* dc pipe latency */ + u32 other_heads_data_return_time = ((wm->num_heads + 1) * worst_chunk_return_time) + + (wm->num_heads * cursor_line_pair_return_time); + u32 latency = mc_latency + other_heads_data_return_time + dc_latency; + u32 max_src_lines_per_dst_line, lb_fill_bw, line_fill_time; + u32 tmp, dmif_size = 12288; + fixed20_12 a, b, c; + + if (wm->num_heads == 0) + return 0; + + a.full = dfixed_const(2); + b.full = dfixed_const(1); + if ((wm->vsc.full > a.full) || + ((wm->vsc.full > b.full) && (wm->vtaps >= 3)) || + (wm->vtaps >= 5) || + ((wm->vsc.full >= a.full) && wm->interlaced)) + max_src_lines_per_dst_line = 4; + else + max_src_lines_per_dst_line = 2; + + a.full = dfixed_const(available_bandwidth); + b.full = dfixed_const(wm->num_heads); + a.full = dfixed_div(a, b); + + b.full = dfixed_const(mc_latency + 512); + c.full = dfixed_const(wm->disp_clk); + b.full = dfixed_div(b, c); + + c.full = dfixed_const(dmif_size); + b.full = dfixed_div(c, b); + + tmp = min(dfixed_trunc(a), dfixed_trunc(b)); + + b.full = dfixed_const(1000); + c.full = dfixed_const(wm->disp_clk); + b.full = dfixed_div(c, b); + c.full = dfixed_const(wm->bytes_per_pixel); + b.full = dfixed_mul(b, c); + + lb_fill_bw = min(tmp, dfixed_trunc(b)); + + a.full = dfixed_const(max_src_lines_per_dst_line * wm->src_width * wm->bytes_per_pixel); + b.full = dfixed_const(1000); + c.full = dfixed_const(lb_fill_bw); + b.full = dfixed_div(c, b); + a.full = dfixed_div(a, b); + line_fill_time = dfixed_trunc(a); + + if (line_fill_time < wm->active_time) + return latency; + else + return latency + (line_fill_time - wm->active_time); + +} + +static bool dce6_average_bandwidth_vs_dram_bandwidth_for_display(struct dce6_wm_params *wm) +{ + if (dce6_average_bandwidth(wm) <= + (dce6_dram_bandwidth_for_display(wm) / wm->num_heads)) + return true; + else + return false; +}; + +static bool dce6_average_bandwidth_vs_available_bandwidth(struct dce6_wm_params *wm) +{ + if (dce6_average_bandwidth(wm) <= + (dce6_available_bandwidth(wm) / wm->num_heads)) + return true; + else + return false; +}; + +static bool dce6_check_latency_hiding(struct dce6_wm_params *wm) +{ + u32 lb_partitions = wm->lb_size / wm->src_width; + u32 line_time = wm->active_time + wm->blank_time; + u32 latency_tolerant_lines; + u32 latency_hiding; + fixed20_12 a; + + a.full = dfixed_const(1); + if (wm->vsc.full > a.full) + latency_tolerant_lines = 1; + else { + if (lb_partitions <= (wm->vtaps + 1)) + latency_tolerant_lines = 1; + else + latency_tolerant_lines = 2; + } + + latency_hiding = (latency_tolerant_lines * line_time + wm->blank_time); + + if (dce6_latency_watermark(wm) <= latency_hiding) + return true; + else + return false; +} + +static void dce6_program_watermarks(struct radeon_device *rdev, + struct radeon_crtc *radeon_crtc, + u32 lb_size, u32 num_heads) +{ + struct drm_display_mode *mode = &radeon_crtc->base.mode; + struct dce6_wm_params wm; + u32 pixel_period; + u32 line_time = 0; + u32 latency_watermark_a = 0, latency_watermark_b = 0; + u32 priority_a_mark = 0, priority_b_mark = 0; + u32 priority_a_cnt = PRIORITY_OFF; + u32 priority_b_cnt = PRIORITY_OFF; + u32 tmp, arb_control3; + fixed20_12 a, b, c; + + if (radeon_crtc->base.enabled && num_heads && mode) { + pixel_period = 1000000 / (u32)mode->clock; + line_time = min((u32)mode->crtc_htotal * pixel_period, (u32)65535); + priority_a_cnt = 0; + priority_b_cnt = 0; + + wm.yclk = rdev->pm.current_mclk * 10; + wm.sclk = rdev->pm.current_sclk * 10; + wm.disp_clk = mode->clock; + wm.src_width = mode->crtc_hdisplay; + wm.active_time = mode->crtc_hdisplay * pixel_period; + wm.blank_time = line_time - wm.active_time; + wm.interlaced = false; + if (mode->flags & DRM_MODE_FLAG_INTERLACE) + wm.interlaced = true; + wm.vsc = radeon_crtc->vsc; + wm.vtaps = 1; + if (radeon_crtc->rmx_type != RMX_OFF) + wm.vtaps = 2; + wm.bytes_per_pixel = 4; /* XXX: get this from fb config */ + wm.lb_size = lb_size; + if (rdev->family == CHIP_ARUBA) + wm.dram_channels = evergreen_get_number_of_dram_channels(rdev); + else + wm.dram_channels = si_get_number_of_dram_channels(rdev); + wm.num_heads = num_heads; + + /* set for high clocks */ + latency_watermark_a = min(dce6_latency_watermark(&wm), (u32)65535); + /* set for low clocks */ + /* wm.yclk = low clk; wm.sclk = low clk */ + latency_watermark_b = min(dce6_latency_watermark(&wm), (u32)65535); + + /* possibly force display priority to high */ + /* should really do this at mode validation time... */ + if (!dce6_average_bandwidth_vs_dram_bandwidth_for_display(&wm) || + !dce6_average_bandwidth_vs_available_bandwidth(&wm) || + !dce6_check_latency_hiding(&wm) || + (rdev->disp_priority == 2)) { + DRM_DEBUG_KMS("force priority to high\n"); + priority_a_cnt |= PRIORITY_ALWAYS_ON; + priority_b_cnt |= PRIORITY_ALWAYS_ON; + } + + a.full = dfixed_const(1000); + b.full = dfixed_const(mode->clock); + b.full = dfixed_div(b, a); + c.full = dfixed_const(latency_watermark_a); + c.full = dfixed_mul(c, b); + c.full = dfixed_mul(c, radeon_crtc->hsc); + c.full = dfixed_div(c, a); + a.full = dfixed_const(16); + c.full = dfixed_div(c, a); + priority_a_mark = dfixed_trunc(c); + priority_a_cnt |= priority_a_mark & PRIORITY_MARK_MASK; + + a.full = dfixed_const(1000); + b.full = dfixed_const(mode->clock); + b.full = dfixed_div(b, a); + c.full = dfixed_const(latency_watermark_b); + c.full = dfixed_mul(c, b); + c.full = dfixed_mul(c, radeon_crtc->hsc); + c.full = dfixed_div(c, a); + a.full = dfixed_const(16); + c.full = dfixed_div(c, a); + priority_b_mark = dfixed_trunc(c); + priority_b_cnt |= priority_b_mark & PRIORITY_MARK_MASK; + } + + /* select wm A */ + arb_control3 = RREG32(DPG_PIPE_ARBITRATION_CONTROL3 + radeon_crtc->crtc_offset); + tmp = arb_control3; + tmp &= ~LATENCY_WATERMARK_MASK(3); + tmp |= LATENCY_WATERMARK_MASK(1); + WREG32(DPG_PIPE_ARBITRATION_CONTROL3 + radeon_crtc->crtc_offset, tmp); + WREG32(DPG_PIPE_LATENCY_CONTROL + radeon_crtc->crtc_offset, + (LATENCY_LOW_WATERMARK(latency_watermark_a) | + LATENCY_HIGH_WATERMARK(line_time))); + /* select wm B */ + tmp = RREG32(DPG_PIPE_ARBITRATION_CONTROL3 + radeon_crtc->crtc_offset); + tmp &= ~LATENCY_WATERMARK_MASK(3); + tmp |= LATENCY_WATERMARK_MASK(2); + WREG32(DPG_PIPE_ARBITRATION_CONTROL3 + radeon_crtc->crtc_offset, tmp); + WREG32(DPG_PIPE_LATENCY_CONTROL + radeon_crtc->crtc_offset, + (LATENCY_LOW_WATERMARK(latency_watermark_b) | + LATENCY_HIGH_WATERMARK(line_time))); + /* restore original selection */ + WREG32(DPG_PIPE_ARBITRATION_CONTROL3 + radeon_crtc->crtc_offset, arb_control3); + + /* write the priority marks */ + WREG32(PRIORITY_A_CNT + radeon_crtc->crtc_offset, priority_a_cnt); + WREG32(PRIORITY_B_CNT + radeon_crtc->crtc_offset, priority_b_cnt); + +} + +void dce6_bandwidth_update(struct radeon_device *rdev) +{ + struct drm_display_mode *mode0 = NULL; + struct drm_display_mode *mode1 = NULL; + u32 num_heads = 0, lb_size; + int i; + + radeon_update_display_priority(rdev); + + for (i = 0; i < rdev->num_crtc; i++) { + if (rdev->mode_info.crtcs[i]->base.enabled) + num_heads++; + } + for (i = 0; i < rdev->num_crtc; i += 2) { + mode0 = &rdev->mode_info.crtcs[i]->base.mode; + mode1 = &rdev->mode_info.crtcs[i+1]->base.mode; + lb_size = dce6_line_buffer_adjust(rdev, rdev->mode_info.crtcs[i], mode0, mode1); + dce6_program_watermarks(rdev, rdev->mode_info.crtcs[i], lb_size, num_heads); + lb_size = dce6_line_buffer_adjust(rdev, rdev->mode_info.crtcs[i+1], mode1, mode0); + dce6_program_watermarks(rdev, rdev->mode_info.crtcs[i+1], lb_size, num_heads); + } +} + +/* + * Core functions + */ +static u32 si_get_tile_pipe_to_backend_map(struct radeon_device *rdev, + u32 num_tile_pipes, + u32 num_backends_per_asic, + u32 *backend_disable_mask_per_asic, + u32 num_shader_engines) +{ + u32 backend_map = 0; + u32 enabled_backends_mask = 0; + u32 enabled_backends_count = 0; + u32 num_backends_per_se; + u32 cur_pipe; + u32 swizzle_pipe[SI_MAX_PIPES]; + u32 cur_backend = 0; + u32 i; + bool force_no_swizzle; + + /* force legal values */ + if (num_tile_pipes < 1) + num_tile_pipes = 1; + if (num_tile_pipes > rdev->config.si.max_tile_pipes) + num_tile_pipes = rdev->config.si.max_tile_pipes; + if (num_shader_engines < 1) + num_shader_engines = 1; + if (num_shader_engines > rdev->config.si.max_shader_engines) + num_shader_engines = rdev->config.si.max_shader_engines; + if (num_backends_per_asic < num_shader_engines) + num_backends_per_asic = num_shader_engines; + if (num_backends_per_asic > (rdev->config.si.max_backends_per_se * num_shader_engines)) + num_backends_per_asic = rdev->config.si.max_backends_per_se * num_shader_engines; + + /* make sure we have the same number of backends per se */ + num_backends_per_asic = ALIGN(num_backends_per_asic, num_shader_engines); + /* set up the number of backends per se */ + num_backends_per_se = num_backends_per_asic / num_shader_engines; + if (num_backends_per_se > rdev->config.si.max_backends_per_se) { + num_backends_per_se = rdev->config.si.max_backends_per_se; + num_backends_per_asic = num_backends_per_se * num_shader_engines; + } + + /* create enable mask and count for enabled backends */ + for (i = 0; i < SI_MAX_BACKENDS; ++i) { + if (((*backend_disable_mask_per_asic >> i) & 1) == 0) { + enabled_backends_mask |= (1 << i); + ++enabled_backends_count; + } + if (enabled_backends_count == num_backends_per_asic) + break; + } + + /* force the backends mask to match the current number of backends */ + if (enabled_backends_count != num_backends_per_asic) { + u32 this_backend_enabled; + u32 shader_engine; + u32 backend_per_se; + + enabled_backends_mask = 0; + enabled_backends_count = 0; + *backend_disable_mask_per_asic = SI_MAX_BACKENDS_MASK; + for (i = 0; i < SI_MAX_BACKENDS; ++i) { + /* calc the current se */ + shader_engine = i / rdev->config.si.max_backends_per_se; + /* calc the backend per se */ + backend_per_se = i % rdev->config.si.max_backends_per_se; + /* default to not enabled */ + this_backend_enabled = 0; + if ((shader_engine < num_shader_engines) && + (backend_per_se < num_backends_per_se)) + this_backend_enabled = 1; + if (this_backend_enabled) { + enabled_backends_mask |= (1 << i); + *backend_disable_mask_per_asic &= ~(1 << i); + ++enabled_backends_count; + } + } + } + + + memset((uint8_t *)&swizzle_pipe[0], 0, sizeof(u32) * SI_MAX_PIPES); + switch (rdev->family) { + case CHIP_TAHITI: + case CHIP_PITCAIRN: + case CHIP_VERDE: + force_no_swizzle = true; + break; + default: + force_no_swizzle = false; + break; + } + if (force_no_swizzle) { + bool last_backend_enabled = false; + + force_no_swizzle = false; + for (i = 0; i < SI_MAX_BACKENDS; ++i) { + if (((enabled_backends_mask >> i) & 1) == 1) { + if (last_backend_enabled) + force_no_swizzle = true; + last_backend_enabled = true; + } else + last_backend_enabled = false; + } + } + + switch (num_tile_pipes) { + case 1: + case 3: + case 5: + case 7: + DRM_ERROR("odd number of pipes!\n"); + break; + case 2: + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + break; + case 4: + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 1; + swizzle_pipe[3] = 3; + } + break; + case 6: + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + swizzle_pipe[4] = 4; + swizzle_pipe[5] = 5; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 4; + swizzle_pipe[3] = 1; + swizzle_pipe[4] = 3; + swizzle_pipe[5] = 5; + } + break; + case 8: + if (force_no_swizzle) { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 1; + swizzle_pipe[2] = 2; + swizzle_pipe[3] = 3; + swizzle_pipe[4] = 4; + swizzle_pipe[5] = 5; + swizzle_pipe[6] = 6; + swizzle_pipe[7] = 7; + } else { + swizzle_pipe[0] = 0; + swizzle_pipe[1] = 2; + swizzle_pipe[2] = 4; + swizzle_pipe[3] = 6; + swizzle_pipe[4] = 1; + swizzle_pipe[5] = 3; + swizzle_pipe[6] = 5; + swizzle_pipe[7] = 7; + } + break; + } + + for (cur_pipe = 0; cur_pipe < num_tile_pipes; ++cur_pipe) { + while (((1 << cur_backend) & enabled_backends_mask) == 0) + cur_backend = (cur_backend + 1) % SI_MAX_BACKENDS; + + backend_map |= (((cur_backend & 0xf) << (swizzle_pipe[cur_pipe] * 4))); + + cur_backend = (cur_backend + 1) % SI_MAX_BACKENDS; + } + + return backend_map; +} + +static u32 si_get_disable_mask_per_asic(struct radeon_device *rdev, + u32 disable_mask_per_se, + u32 max_disable_mask_per_se, + u32 num_shader_engines) +{ + u32 disable_field_width_per_se = r600_count_pipe_bits(disable_mask_per_se); + u32 disable_mask_per_asic = disable_mask_per_se & max_disable_mask_per_se; + + if (num_shader_engines == 1) + return disable_mask_per_asic; + else if (num_shader_engines == 2) + return disable_mask_per_asic | (disable_mask_per_asic << disable_field_width_per_se); + else + return 0xffffffff; +} + +static void si_tiling_mode_table_init(struct radeon_device *rdev) +{ + const u32 num_tile_mode_states = 32; + u32 reg_offset, gb_tile_moden, split_equal_to_row_size; + + switch (rdev->config.si.mem_row_size_in_kb) { + case 1: + split_equal_to_row_size = ADDR_SURF_TILE_SPLIT_1KB; + break; + case 2: + default: + split_equal_to_row_size = ADDR_SURF_TILE_SPLIT_2KB; + break; + case 4: + split_equal_to_row_size = ADDR_SURF_TILE_SPLIT_4KB; + break; + } + + if ((rdev->family == CHIP_TAHITI) || + (rdev->family == CHIP_PITCAIRN)) { + for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) { + switch (reg_offset) { + case 0: /* non-AA compressed depth or any compressed stencil */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 1: /* 2xAA/4xAA compressed depth only */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 2: /* 8xAA compressed depth only */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 3: /* 2xAA/4xAA compressed depth with stencil (for depth buffer) */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 4: /* Maps w/ a dimension less than the 2D macro-tile dimensions (for mipmapped depth textures) */ + gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 5: /* Uncompressed 16bpp depth - and stencil buffer allocated with it */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(split_equal_to_row_size) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 6: /* Uncompressed 32bpp depth - and stencil buffer allocated with it */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(split_equal_to_row_size) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); + break; + case 7: /* Uncompressed 8bpp stencil without depth (drivers typically do not use) */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(split_equal_to_row_size) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 8: /* 1D and 1D Array Surfaces */ + gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 9: /* Displayable maps. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 10: /* Display 8bpp. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 11: /* Display 16bpp. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 12: /* Display 32bpp. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); + break; + case 13: /* Thin. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 14: /* Thin 8 bpp. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); + break; + case 15: /* Thin 16 bpp. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); + break; + case 16: /* Thin 32 bpp. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); + break; + case 17: /* Thin 64 bpp. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(split_equal_to_row_size) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); + break; + case 21: /* 8 bpp PRT. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 22: /* 16 bpp PRT */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); + break; + case 23: /* 32 bpp PRT */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 24: /* 64 bpp PRT */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 25: /* 128 bpp PRT */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_1KB) | + NUM_BANKS(ADDR_SURF_8_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); + break; + default: + gb_tile_moden = 0; + break; + } + WREG32(GB_TILE_MODE0 + (reg_offset * 4), gb_tile_moden); + } + } else if (rdev->family == CHIP_VERDE) { + for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) { + switch (reg_offset) { + case 0: /* non-AA compressed depth or any compressed stencil */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); + break; + case 1: /* 2xAA/4xAA compressed depth only */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); + break; + case 2: /* 8xAA compressed depth only */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); + break; + case 3: /* 2xAA/4xAA compressed depth with stencil (for depth buffer) */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); + break; + case 4: /* Maps w/ a dimension less than the 2D macro-tile dimensions (for mipmapped depth textures) */ + gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 5: /* Uncompressed 16bpp depth - and stencil buffer allocated with it */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(split_equal_to_row_size) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 6: /* Uncompressed 32bpp depth - and stencil buffer allocated with it */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(split_equal_to_row_size) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 7: /* Uncompressed 8bpp stencil without depth (drivers typically do not use) */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(split_equal_to_row_size) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); + break; + case 8: /* 1D and 1D Array Surfaces */ + gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 9: /* Displayable maps. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 10: /* Display 8bpp. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); + break; + case 11: /* Display 16bpp. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 12: /* Display 32bpp. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 13: /* Thin. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 14: /* Thin 8 bpp. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 15: /* Thin 16 bpp. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 16: /* Thin 32 bpp. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 17: /* Thin 64 bpp. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P4_8x16) | + TILE_SPLIT(split_equal_to_row_size) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 21: /* 8 bpp PRT. */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 22: /* 16 bpp PRT */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4)); + break; + case 23: /* 32 bpp PRT */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 24: /* 64 bpp PRT */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | + NUM_BANKS(ADDR_SURF_16_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2)); + break; + case 25: /* 128 bpp PRT */ + gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) | + PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_1KB) | + NUM_BANKS(ADDR_SURF_8_BANK) | + BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1)); + break; + default: + gb_tile_moden = 0; + break; + } + WREG32(GB_TILE_MODE0 + (reg_offset * 4), gb_tile_moden); + } + } else + DRM_ERROR("unknown asic: 0x%x\n", rdev->family); +} + +static void si_gpu_init(struct radeon_device *rdev) +{ + u32 cc_rb_backend_disable = 0; + u32 cc_gc_shader_array_config; + u32 gb_addr_config = 0; + u32 mc_shared_chmap, mc_arb_ramcfg; + u32 gb_backend_map; + u32 cgts_tcc_disable; + u32 sx_debug_1; + u32 gc_user_shader_array_config; + u32 gc_user_rb_backend_disable; + u32 cgts_user_tcc_disable; + u32 hdp_host_path_cntl; + u32 tmp; + int i, j; + + switch (rdev->family) { + case CHIP_TAHITI: + rdev->config.si.max_shader_engines = 2; + rdev->config.si.max_pipes_per_simd = 4; + rdev->config.si.max_tile_pipes = 12; + rdev->config.si.max_simds_per_se = 8; + rdev->config.si.max_backends_per_se = 4; + rdev->config.si.max_texture_channel_caches = 12; + rdev->config.si.max_gprs = 256; + rdev->config.si.max_gs_threads = 32; + rdev->config.si.max_hw_contexts = 8; + + rdev->config.si.sc_prim_fifo_size_frontend = 0x20; + rdev->config.si.sc_prim_fifo_size_backend = 0x100; + rdev->config.si.sc_hiz_tile_fifo_size = 0x30; + rdev->config.si.sc_earlyz_tile_fifo_size = 0x130; + break; + case CHIP_PITCAIRN: + rdev->config.si.max_shader_engines = 2; + rdev->config.si.max_pipes_per_simd = 4; + rdev->config.si.max_tile_pipes = 8; + rdev->config.si.max_simds_per_se = 5; + rdev->config.si.max_backends_per_se = 4; + rdev->config.si.max_texture_channel_caches = 8; + rdev->config.si.max_gprs = 256; + rdev->config.si.max_gs_threads = 32; + rdev->config.si.max_hw_contexts = 8; + + rdev->config.si.sc_prim_fifo_size_frontend = 0x20; + rdev->config.si.sc_prim_fifo_size_backend = 0x100; + rdev->config.si.sc_hiz_tile_fifo_size = 0x30; + rdev->config.si.sc_earlyz_tile_fifo_size = 0x130; + break; + case CHIP_VERDE: + default: + rdev->config.si.max_shader_engines = 1; + rdev->config.si.max_pipes_per_simd = 4; + rdev->config.si.max_tile_pipes = 4; + rdev->config.si.max_simds_per_se = 2; + rdev->config.si.max_backends_per_se = 4; + rdev->config.si.max_texture_channel_caches = 4; + rdev->config.si.max_gprs = 256; + rdev->config.si.max_gs_threads = 32; + rdev->config.si.max_hw_contexts = 8; + + rdev->config.si.sc_prim_fifo_size_frontend = 0x20; + rdev->config.si.sc_prim_fifo_size_backend = 0x40; + rdev->config.si.sc_hiz_tile_fifo_size = 0x30; + rdev->config.si.sc_earlyz_tile_fifo_size = 0x130; + break; + } + + /* Initialize HDP */ + for (i = 0, j = 0; i < 32; i++, j += 0x18) { + WREG32((0x2c14 + j), 0x00000000); + WREG32((0x2c18 + j), 0x00000000); + WREG32((0x2c1c + j), 0x00000000); + WREG32((0x2c20 + j), 0x00000000); + WREG32((0x2c24 + j), 0x00000000); + } + + WREG32(GRBM_CNTL, GRBM_READ_TIMEOUT(0xff)); + + evergreen_fix_pci_max_read_req_size(rdev); + + WREG32(BIF_FB_EN, FB_READ_EN | FB_WRITE_EN); + + mc_shared_chmap = RREG32(MC_SHARED_CHMAP); + mc_arb_ramcfg = RREG32(MC_ARB_RAMCFG); + + cc_rb_backend_disable = RREG32(CC_RB_BACKEND_DISABLE); + cc_gc_shader_array_config = RREG32(CC_GC_SHADER_ARRAY_CONFIG); + cgts_tcc_disable = 0xffff0000; + for (i = 0; i < rdev->config.si.max_texture_channel_caches; i++) + cgts_tcc_disable &= ~(1 << (16 + i)); + gc_user_rb_backend_disable = RREG32(GC_USER_RB_BACKEND_DISABLE); + gc_user_shader_array_config = RREG32(GC_USER_SHADER_ARRAY_CONFIG); + cgts_user_tcc_disable = RREG32(CGTS_USER_TCC_DISABLE); + + rdev->config.si.num_shader_engines = rdev->config.si.max_shader_engines; + rdev->config.si.num_tile_pipes = rdev->config.si.max_tile_pipes; + tmp = ((~gc_user_rb_backend_disable) & BACKEND_DISABLE_MASK) >> BACKEND_DISABLE_SHIFT; + rdev->config.si.num_backends_per_se = r600_count_pipe_bits(tmp); + tmp = (gc_user_rb_backend_disable & BACKEND_DISABLE_MASK) >> BACKEND_DISABLE_SHIFT; + rdev->config.si.backend_disable_mask_per_asic = + si_get_disable_mask_per_asic(rdev, tmp, SI_MAX_BACKENDS_PER_SE_MASK, + rdev->config.si.num_shader_engines); + rdev->config.si.backend_map = + si_get_tile_pipe_to_backend_map(rdev, rdev->config.si.num_tile_pipes, + rdev->config.si.num_backends_per_se * + rdev->config.si.num_shader_engines, + &rdev->config.si.backend_disable_mask_per_asic, + rdev->config.si.num_shader_engines); + tmp = ((~cgts_user_tcc_disable) & TCC_DISABLE_MASK) >> TCC_DISABLE_SHIFT; + rdev->config.si.num_texture_channel_caches = r600_count_pipe_bits(tmp); + rdev->config.si.mem_max_burst_length_bytes = 256; + tmp = (mc_arb_ramcfg & NOOFCOLS_MASK) >> NOOFCOLS_SHIFT; + rdev->config.si.mem_row_size_in_kb = (4 * (1 << (8 + tmp))) / 1024; + if (rdev->config.si.mem_row_size_in_kb > 4) + rdev->config.si.mem_row_size_in_kb = 4; + /* XXX use MC settings? */ + rdev->config.si.shader_engine_tile_size = 32; + rdev->config.si.num_gpus = 1; + rdev->config.si.multi_gpu_tile_size = 64; + + gb_addr_config = 0; + switch (rdev->config.si.num_tile_pipes) { + case 1: + gb_addr_config |= NUM_PIPES(0); + break; + case 2: + gb_addr_config |= NUM_PIPES(1); + break; + case 4: + gb_addr_config |= NUM_PIPES(2); + break; + case 8: + default: + gb_addr_config |= NUM_PIPES(3); + break; + } + + tmp = (rdev->config.si.mem_max_burst_length_bytes / 256) - 1; + gb_addr_config |= PIPE_INTERLEAVE_SIZE(tmp); + gb_addr_config |= NUM_SHADER_ENGINES(rdev->config.si.num_shader_engines - 1); + tmp = (rdev->config.si.shader_engine_tile_size / 16) - 1; + gb_addr_config |= SHADER_ENGINE_TILE_SIZE(tmp); + switch (rdev->config.si.num_gpus) { + case 1: + default: + gb_addr_config |= NUM_GPUS(0); + break; + case 2: + gb_addr_config |= NUM_GPUS(1); + break; + case 4: + gb_addr_config |= NUM_GPUS(2); + break; + } + switch (rdev->config.si.multi_gpu_tile_size) { + case 16: + gb_addr_config |= MULTI_GPU_TILE_SIZE(0); + break; + case 32: + default: + gb_addr_config |= MULTI_GPU_TILE_SIZE(1); + break; + case 64: + gb_addr_config |= MULTI_GPU_TILE_SIZE(2); + break; + case 128: + gb_addr_config |= MULTI_GPU_TILE_SIZE(3); + break; + } + switch (rdev->config.si.mem_row_size_in_kb) { + case 1: + default: + gb_addr_config |= ROW_SIZE(0); + break; + case 2: + gb_addr_config |= ROW_SIZE(1); + break; + case 4: + gb_addr_config |= ROW_SIZE(2); + break; + } + + tmp = (gb_addr_config & NUM_PIPES_MASK) >> NUM_PIPES_SHIFT; + rdev->config.si.num_tile_pipes = (1 << tmp); + tmp = (gb_addr_config & PIPE_INTERLEAVE_SIZE_MASK) >> PIPE_INTERLEAVE_SIZE_SHIFT; + rdev->config.si.mem_max_burst_length_bytes = (tmp + 1) * 256; + tmp = (gb_addr_config & NUM_SHADER_ENGINES_MASK) >> NUM_SHADER_ENGINES_SHIFT; + rdev->config.si.num_shader_engines = tmp + 1; + tmp = (gb_addr_config & NUM_GPUS_MASK) >> NUM_GPUS_SHIFT; + rdev->config.si.num_gpus = tmp + 1; + tmp = (gb_addr_config & MULTI_GPU_TILE_SIZE_MASK) >> MULTI_GPU_TILE_SIZE_SHIFT; + rdev->config.si.multi_gpu_tile_size = 1 << tmp; + tmp = (gb_addr_config & ROW_SIZE_MASK) >> ROW_SIZE_SHIFT; + rdev->config.si.mem_row_size_in_kb = 1 << tmp; + + gb_backend_map = + si_get_tile_pipe_to_backend_map(rdev, rdev->config.si.num_tile_pipes, + rdev->config.si.num_backends_per_se * + rdev->config.si.num_shader_engines, + &rdev->config.si.backend_disable_mask_per_asic, + rdev->config.si.num_shader_engines); + + /* setup tiling info dword. gb_addr_config is not adequate since it does + * not have bank info, so create a custom tiling dword. + * bits 3:0 num_pipes + * bits 7:4 num_banks + * bits 11:8 group_size + * bits 15:12 row_size + */ + rdev->config.si.tile_config = 0; + switch (rdev->config.si.num_tile_pipes) { + case 1: + rdev->config.si.tile_config |= (0 << 0); + break; + case 2: + rdev->config.si.tile_config |= (1 << 0); + break; + case 4: + rdev->config.si.tile_config |= (2 << 0); + break; + case 8: + default: + /* XXX what about 12? */ + rdev->config.si.tile_config |= (3 << 0); + break; + } + rdev->config.si.tile_config |= + ((mc_arb_ramcfg & NOOFBANK_MASK) >> NOOFBANK_SHIFT) << 4; + rdev->config.si.tile_config |= + ((gb_addr_config & PIPE_INTERLEAVE_SIZE_MASK) >> PIPE_INTERLEAVE_SIZE_SHIFT) << 8; + rdev->config.si.tile_config |= + ((gb_addr_config & ROW_SIZE_MASK) >> ROW_SIZE_SHIFT) << 12; + + rdev->config.si.backend_map = gb_backend_map; + WREG32(GB_ADDR_CONFIG, gb_addr_config); + WREG32(DMIF_ADDR_CONFIG, gb_addr_config); + WREG32(HDP_ADDR_CONFIG, gb_addr_config); + + /* primary versions */ + WREG32(CC_RB_BACKEND_DISABLE, cc_rb_backend_disable); + WREG32(CC_SYS_RB_BACKEND_DISABLE, cc_rb_backend_disable); + WREG32(CC_GC_SHADER_ARRAY_CONFIG, cc_gc_shader_array_config); + + WREG32(CGTS_TCC_DISABLE, cgts_tcc_disable); + + /* user versions */ + WREG32(GC_USER_RB_BACKEND_DISABLE, cc_rb_backend_disable); + WREG32(GC_USER_SYS_RB_BACKEND_DISABLE, cc_rb_backend_disable); + WREG32(GC_USER_SHADER_ARRAY_CONFIG, cc_gc_shader_array_config); + + WREG32(CGTS_USER_TCC_DISABLE, cgts_tcc_disable); + + si_tiling_mode_table_init(rdev); + + /* set HW defaults for 3D engine */ + WREG32(CP_QUEUE_THRESHOLDS, (ROQ_IB1_START(0x16) | + ROQ_IB2_START(0x2b))); + WREG32(CP_MEQ_THRESHOLDS, MEQ1_START(0x30) | MEQ2_START(0x60)); + + sx_debug_1 = RREG32(SX_DEBUG_1); + WREG32(SX_DEBUG_1, sx_debug_1); + + WREG32(SPI_CONFIG_CNTL_1, VTX_DONE_DELAY(4)); + + WREG32(PA_SC_FIFO_SIZE, (SC_FRONTEND_PRIM_FIFO_SIZE(rdev->config.si.sc_prim_fifo_size_frontend) | + SC_BACKEND_PRIM_FIFO_SIZE(rdev->config.si.sc_prim_fifo_size_backend) | + SC_HIZ_TILE_FIFO_SIZE(rdev->config.si.sc_hiz_tile_fifo_size) | + SC_EARLYZ_TILE_FIFO_SIZE(rdev->config.si.sc_earlyz_tile_fifo_size))); + + WREG32(VGT_NUM_INSTANCES, 1); + + WREG32(CP_PERFMON_CNTL, 0); + + WREG32(SQ_CONFIG, 0); + + WREG32(PA_SC_FORCE_EOV_MAX_CNTS, (FORCE_EOV_MAX_CLK_CNT(4095) | + FORCE_EOV_MAX_REZ_CNT(255))); + + WREG32(VGT_CACHE_INVALIDATION, CACHE_INVALIDATION(VC_AND_TC) | + AUTO_INVLD_EN(ES_AND_GS_AUTO)); + + WREG32(VGT_GS_VERTEX_REUSE, 16); + WREG32(PA_SC_LINE_STIPPLE_STATE, 0); + + WREG32(CB_PERFCOUNTER0_SELECT0, 0); + WREG32(CB_PERFCOUNTER0_SELECT1, 0); + WREG32(CB_PERFCOUNTER1_SELECT0, 0); + WREG32(CB_PERFCOUNTER1_SELECT1, 0); + WREG32(CB_PERFCOUNTER2_SELECT0, 0); + WREG32(CB_PERFCOUNTER2_SELECT1, 0); + WREG32(CB_PERFCOUNTER3_SELECT0, 0); + WREG32(CB_PERFCOUNTER3_SELECT1, 0); + + tmp = RREG32(HDP_MISC_CNTL); + tmp |= HDP_FLUSH_INVALIDATE_CACHE; + WREG32(HDP_MISC_CNTL, tmp); + + hdp_host_path_cntl = RREG32(HDP_HOST_PATH_CNTL); + WREG32(HDP_HOST_PATH_CNTL, hdp_host_path_cntl); + + WREG32(PA_CL_ENHANCE, CLIP_VTX_REORDER_ENA | NUM_CLIP_SEQ(3)); + + udelay(50); +} + +/* + * GPU scratch registers helpers function. + */ +static void si_scratch_init(struct radeon_device *rdev) +{ + int i; + + rdev->scratch.num_reg = 7; + rdev->scratch.reg_base = SCRATCH_REG0; + for (i = 0; i < rdev->scratch.num_reg; i++) { + rdev->scratch.free[i] = true; + rdev->scratch.reg[i] = rdev->scratch.reg_base + (i * 4); + } +} + +void si_fence_ring_emit(struct radeon_device *rdev, + struct radeon_fence *fence) +{ + struct radeon_ring *ring = &rdev->ring[fence->ring]; + u64 addr = rdev->fence_drv[fence->ring].gpu_addr; + + /* flush read cache over gart */ + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + radeon_ring_write(ring, (CP_COHER_CNTL2 - PACKET3_SET_CONFIG_REG_START) >> 2); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); + radeon_ring_write(ring, PACKET3_TCL1_ACTION_ENA | + PACKET3_TC_ACTION_ENA | + PACKET3_SH_KCACHE_ACTION_ENA | + PACKET3_SH_ICACHE_ACTION_ENA); + radeon_ring_write(ring, 0xFFFFFFFF); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 10); /* poll interval */ + /* EVENT_WRITE_EOP - flush caches, send int */ + radeon_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE_EOP, 4)); + radeon_ring_write(ring, EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) | EVENT_INDEX(5)); + radeon_ring_write(ring, addr & 0xffffffff); + radeon_ring_write(ring, (upper_32_bits(addr) & 0xff) | DATA_SEL(1) | INT_SEL(2)); + radeon_ring_write(ring, fence->seq); + radeon_ring_write(ring, 0); +} + +/* + * IB stuff + */ +void si_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) +{ + struct radeon_ring *ring = &rdev->ring[ib->fence->ring]; + u32 header; + + if (ib->is_const_ib) + header = PACKET3(PACKET3_INDIRECT_BUFFER_CONST, 2); + else + header = PACKET3(PACKET3_INDIRECT_BUFFER, 2); + + radeon_ring_write(ring, header); + radeon_ring_write(ring, +#ifdef __BIG_ENDIAN + (2 << 0) | +#endif + (ib->gpu_addr & 0xFFFFFFFC)); + radeon_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xFFFF); + radeon_ring_write(ring, ib->length_dw | (ib->vm_id << 24)); + + /* flush read cache over gart for this vmid */ + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + radeon_ring_write(ring, (CP_COHER_CNTL2 - PACKET3_SET_CONFIG_REG_START) >> 2); + radeon_ring_write(ring, ib->vm_id); + radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3)); + radeon_ring_write(ring, PACKET3_TCL1_ACTION_ENA | + PACKET3_TC_ACTION_ENA | + PACKET3_SH_KCACHE_ACTION_ENA | + PACKET3_SH_ICACHE_ACTION_ENA); + radeon_ring_write(ring, 0xFFFFFFFF); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 10); /* poll interval */ +} + +/* + * CP. + */ +static void si_cp_enable(struct radeon_device *rdev, bool enable) +{ + if (enable) + WREG32(CP_ME_CNTL, 0); + else { + radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size); + WREG32(CP_ME_CNTL, (CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT)); + WREG32(SCRATCH_UMSK, 0); + } + udelay(50); +} + +static int si_cp_load_microcode(struct radeon_device *rdev) +{ + const __be32 *fw_data; + int i; + + if (!rdev->me_fw || !rdev->pfp_fw) + return -EINVAL; + + si_cp_enable(rdev, false); + + /* PFP */ + fw_data = (const __be32 *)rdev->pfp_fw->data; + WREG32(CP_PFP_UCODE_ADDR, 0); + for (i = 0; i < SI_PFP_UCODE_SIZE; i++) + WREG32(CP_PFP_UCODE_DATA, be32_to_cpup(fw_data++)); + WREG32(CP_PFP_UCODE_ADDR, 0); + + /* CE */ + fw_data = (const __be32 *)rdev->ce_fw->data; + WREG32(CP_CE_UCODE_ADDR, 0); + for (i = 0; i < SI_CE_UCODE_SIZE; i++) + WREG32(CP_CE_UCODE_DATA, be32_to_cpup(fw_data++)); + WREG32(CP_CE_UCODE_ADDR, 0); + + /* ME */ + fw_data = (const __be32 *)rdev->me_fw->data; + WREG32(CP_ME_RAM_WADDR, 0); + for (i = 0; i < SI_PM4_UCODE_SIZE; i++) + WREG32(CP_ME_RAM_DATA, be32_to_cpup(fw_data++)); + WREG32(CP_ME_RAM_WADDR, 0); + + WREG32(CP_PFP_UCODE_ADDR, 0); + WREG32(CP_CE_UCODE_ADDR, 0); + WREG32(CP_ME_RAM_WADDR, 0); + WREG32(CP_ME_RAM_RADDR, 0); + return 0; +} + +static int si_cp_start(struct radeon_device *rdev) +{ + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; + int r, i; + + r = radeon_ring_lock(rdev, ring, 7 + 4); + if (r) { + DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r); + return r; + } + /* init the CP */ + radeon_ring_write(ring, PACKET3(PACKET3_ME_INITIALIZE, 5)); + radeon_ring_write(ring, 0x1); + radeon_ring_write(ring, 0x0); + radeon_ring_write(ring, rdev->config.si.max_hw_contexts - 1); + radeon_ring_write(ring, PACKET3_ME_INITIALIZE_DEVICE_ID(1)); + radeon_ring_write(ring, 0); + radeon_ring_write(ring, 0); + + /* init the CE partitions */ + radeon_ring_write(ring, PACKET3(PACKET3_SET_BASE, 2)); + radeon_ring_write(ring, PACKET3_BASE_INDEX(CE_PARTITION_BASE)); + radeon_ring_write(ring, 0xc000); + radeon_ring_write(ring, 0xe000); + radeon_ring_unlock_commit(rdev, ring); + + si_cp_enable(rdev, true); + + r = radeon_ring_lock(rdev, ring, si_default_size + 10); + if (r) { + DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r); + return r; + } + + /* setup clear context state */ + radeon_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0)); + radeon_ring_write(ring, PACKET3_PREAMBLE_BEGIN_CLEAR_STATE); + + for (i = 0; i < si_default_size; i++) + radeon_ring_write(ring, si_default_state[i]); + + radeon_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0)); + radeon_ring_write(ring, PACKET3_PREAMBLE_END_CLEAR_STATE); + + /* set clear context state */ + radeon_ring_write(ring, PACKET3(PACKET3_CLEAR_STATE, 0)); + radeon_ring_write(ring, 0); + + radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2)); + radeon_ring_write(ring, 0x00000316); + radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */ + radeon_ring_write(ring, 0x00000010); /* VGT_OUT_DEALLOC_CNTL */ + + radeon_ring_unlock_commit(rdev, ring); + + for (i = RADEON_RING_TYPE_GFX_INDEX; i <= CAYMAN_RING_TYPE_CP2_INDEX; ++i) { + ring = &rdev->ring[i]; + r = radeon_ring_lock(rdev, ring, 2); + + /* clear the compute context state */ + radeon_ring_write(ring, PACKET3_COMPUTE(PACKET3_CLEAR_STATE, 0)); + radeon_ring_write(ring, 0); + + radeon_ring_unlock_commit(rdev, ring); + } + + return 0; +} + +static void si_cp_fini(struct radeon_device *rdev) +{ + si_cp_enable(rdev, false); + radeon_ring_fini(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); + radeon_ring_fini(rdev, &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]); + radeon_ring_fini(rdev, &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]); +} + +static int si_cp_resume(struct radeon_device *rdev) +{ + struct radeon_ring *ring; + u32 tmp; + u32 rb_bufsz; + int r; + + /* Reset cp; if cp is reset, then PA, SH, VGT also need to be reset */ + WREG32(GRBM_SOFT_RESET, (SOFT_RESET_CP | + SOFT_RESET_PA | + SOFT_RESET_VGT | + SOFT_RESET_SPI | + SOFT_RESET_SX)); + RREG32(GRBM_SOFT_RESET); + mdelay(15); + WREG32(GRBM_SOFT_RESET, 0); + RREG32(GRBM_SOFT_RESET); + + WREG32(CP_SEM_WAIT_TIMER, 0x0); + WREG32(CP_SEM_INCOMPLETE_TIMER_CNTL, 0x0); + + /* Set the write pointer delay */ + WREG32(CP_RB_WPTR_DELAY, 0); + + WREG32(CP_DEBUG, 0); + WREG32(SCRATCH_ADDR, ((rdev->wb.gpu_addr + RADEON_WB_SCRATCH_OFFSET) >> 8) & 0xFFFFFFFF); + + /* ring 0 - compute and gfx */ + /* Set ring buffer size */ + ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; + rb_bufsz = drm_order(ring->ring_size / 8); + tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; +#ifdef __BIG_ENDIAN + tmp |= BUF_SWAP_32BIT; +#endif + WREG32(CP_RB0_CNTL, tmp); + + /* Initialize the ring buffer's read and write pointers */ + WREG32(CP_RB0_CNTL, tmp | RB_RPTR_WR_ENA); + ring->wptr = 0; + WREG32(CP_RB0_WPTR, ring->wptr); + + /* set the wb address wether it's enabled or not */ + WREG32(CP_RB0_RPTR_ADDR, (rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFFFFFFFC); + WREG32(CP_RB0_RPTR_ADDR_HI, upper_32_bits(rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFF); + + if (rdev->wb.enabled) + WREG32(SCRATCH_UMSK, 0xff); + else { + tmp |= RB_NO_UPDATE; + WREG32(SCRATCH_UMSK, 0); + } + + mdelay(1); + WREG32(CP_RB0_CNTL, tmp); + + WREG32(CP_RB0_BASE, ring->gpu_addr >> 8); + + ring->rptr = RREG32(CP_RB0_RPTR); + + /* ring1 - compute only */ + /* Set ring buffer size */ + ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]; + rb_bufsz = drm_order(ring->ring_size / 8); + tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; +#ifdef __BIG_ENDIAN + tmp |= BUF_SWAP_32BIT; +#endif + WREG32(CP_RB1_CNTL, tmp); + + /* Initialize the ring buffer's read and write pointers */ + WREG32(CP_RB1_CNTL, tmp | RB_RPTR_WR_ENA); + ring->wptr = 0; + WREG32(CP_RB1_WPTR, ring->wptr); + + /* set the wb address wether it's enabled or not */ + WREG32(CP_RB1_RPTR_ADDR, (rdev->wb.gpu_addr + RADEON_WB_CP1_RPTR_OFFSET) & 0xFFFFFFFC); + WREG32(CP_RB1_RPTR_ADDR_HI, upper_32_bits(rdev->wb.gpu_addr + RADEON_WB_CP1_RPTR_OFFSET) & 0xFF); + + mdelay(1); + WREG32(CP_RB1_CNTL, tmp); + + WREG32(CP_RB1_BASE, ring->gpu_addr >> 8); + + ring->rptr = RREG32(CP_RB1_RPTR); + + /* ring2 - compute only */ + /* Set ring buffer size */ + ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]; + rb_bufsz = drm_order(ring->ring_size / 8); + tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz; +#ifdef __BIG_ENDIAN + tmp |= BUF_SWAP_32BIT; +#endif + WREG32(CP_RB2_CNTL, tmp); + + /* Initialize the ring buffer's read and write pointers */ + WREG32(CP_RB2_CNTL, tmp | RB_RPTR_WR_ENA); + ring->wptr = 0; + WREG32(CP_RB2_WPTR, ring->wptr); + + /* set the wb address wether it's enabled or not */ + WREG32(CP_RB2_RPTR_ADDR, (rdev->wb.gpu_addr + RADEON_WB_CP2_RPTR_OFFSET) & 0xFFFFFFFC); + WREG32(CP_RB2_RPTR_ADDR_HI, upper_32_bits(rdev->wb.gpu_addr + RADEON_WB_CP2_RPTR_OFFSET) & 0xFF); + + mdelay(1); + WREG32(CP_RB2_CNTL, tmp); + + WREG32(CP_RB2_BASE, ring->gpu_addr >> 8); + + ring->rptr = RREG32(CP_RB2_RPTR); + + /* start the rings */ + si_cp_start(rdev); + rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = true; + rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX].ready = true; + rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX].ready = true; + r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); + if (r) { + rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false; + rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX].ready = false; + rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX].ready = false; + return r; + } + r = radeon_ring_test(rdev, CAYMAN_RING_TYPE_CP1_INDEX, &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]); + if (r) { + rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX].ready = false; + } + r = radeon_ring_test(rdev, CAYMAN_RING_TYPE_CP2_INDEX, &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]); + if (r) { + rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX].ready = false; + } + + return 0; +} + +bool si_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring) +{ + u32 srbm_status; + u32 grbm_status, grbm_status2; + u32 grbm_status_se0, grbm_status_se1; + struct r100_gpu_lockup *lockup = &rdev->config.si.lockup; + int r; + + srbm_status = RREG32(SRBM_STATUS); + grbm_status = RREG32(GRBM_STATUS); + grbm_status2 = RREG32(GRBM_STATUS2); + grbm_status_se0 = RREG32(GRBM_STATUS_SE0); + grbm_status_se1 = RREG32(GRBM_STATUS_SE1); + if (!(grbm_status & GUI_ACTIVE)) { + r100_gpu_lockup_update(lockup, ring); + return false; + } + /* force CP activities */ + r = radeon_ring_lock(rdev, ring, 2); + if (!r) { + /* PACKET2 NOP */ + radeon_ring_write(ring, 0x80000000); + radeon_ring_write(ring, 0x80000000); + radeon_ring_unlock_commit(rdev, ring); + } + /* XXX deal with CP0,1,2 */ + ring->rptr = RREG32(ring->rptr_reg); + return r100_gpu_cp_is_lockup(rdev, lockup, ring); +} + +static int si_gpu_soft_reset(struct radeon_device *rdev) +{ + struct evergreen_mc_save save; + u32 grbm_reset = 0; + + if (!(RREG32(GRBM_STATUS) & GUI_ACTIVE)) + return 0; + + dev_info(rdev->dev, "GPU softreset \n"); + dev_info(rdev->dev, " GRBM_STATUS=0x%08X\n", + RREG32(GRBM_STATUS)); + dev_info(rdev->dev, " GRBM_STATUS2=0x%08X\n", + RREG32(GRBM_STATUS2)); + dev_info(rdev->dev, " GRBM_STATUS_SE0=0x%08X\n", + RREG32(GRBM_STATUS_SE0)); + dev_info(rdev->dev, " GRBM_STATUS_SE1=0x%08X\n", + RREG32(GRBM_STATUS_SE1)); + dev_info(rdev->dev, " SRBM_STATUS=0x%08X\n", + RREG32(SRBM_STATUS)); + evergreen_mc_stop(rdev, &save); + if (radeon_mc_wait_for_idle(rdev)) { + dev_warn(rdev->dev, "Wait for MC idle timedout !\n"); + } + /* Disable CP parsing/prefetching */ + WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT); + + /* reset all the gfx blocks */ + grbm_reset = (SOFT_RESET_CP | + SOFT_RESET_CB | + SOFT_RESET_DB | + SOFT_RESET_GDS | + SOFT_RESET_PA | + SOFT_RESET_SC | + SOFT_RESET_SPI | + SOFT_RESET_SX | + SOFT_RESET_TC | + SOFT_RESET_TA | + SOFT_RESET_VGT | + SOFT_RESET_IA); + + dev_info(rdev->dev, " GRBM_SOFT_RESET=0x%08X\n", grbm_reset); + WREG32(GRBM_SOFT_RESET, grbm_reset); + (void)RREG32(GRBM_SOFT_RESET); + udelay(50); + WREG32(GRBM_SOFT_RESET, 0); + (void)RREG32(GRBM_SOFT_RESET); + /* Wait a little for things to settle down */ + udelay(50); + dev_info(rdev->dev, " GRBM_STATUS=0x%08X\n", + RREG32(GRBM_STATUS)); + dev_info(rdev->dev, " GRBM_STATUS2=0x%08X\n", + RREG32(GRBM_STATUS2)); + dev_info(rdev->dev, " GRBM_STATUS_SE0=0x%08X\n", + RREG32(GRBM_STATUS_SE0)); + dev_info(rdev->dev, " GRBM_STATUS_SE1=0x%08X\n", + RREG32(GRBM_STATUS_SE1)); + dev_info(rdev->dev, " SRBM_STATUS=0x%08X\n", + RREG32(SRBM_STATUS)); + evergreen_mc_resume(rdev, &save); + return 0; +} + +int si_asic_reset(struct radeon_device *rdev) +{ + return si_gpu_soft_reset(rdev); +} + +/* MC */ +static void si_mc_program(struct radeon_device *rdev) +{ + struct evergreen_mc_save save; + u32 tmp; + int i, j; + + /* Initialize HDP */ + for (i = 0, j = 0; i < 32; i++, j += 0x18) { + WREG32((0x2c14 + j), 0x00000000); + WREG32((0x2c18 + j), 0x00000000); + WREG32((0x2c1c + j), 0x00000000); + WREG32((0x2c20 + j), 0x00000000); + WREG32((0x2c24 + j), 0x00000000); + } + WREG32(HDP_REG_COHERENCY_FLUSH_CNTL, 0); + + evergreen_mc_stop(rdev, &save); + if (radeon_mc_wait_for_idle(rdev)) { + dev_warn(rdev->dev, "Wait for MC idle timedout !\n"); + } + /* Lockout access through VGA aperture*/ + WREG32(VGA_HDP_CONTROL, VGA_MEMORY_DISABLE); + /* Update configuration */ + WREG32(MC_VM_SYSTEM_APERTURE_LOW_ADDR, + rdev->mc.vram_start >> 12); + WREG32(MC_VM_SYSTEM_APERTURE_HIGH_ADDR, + rdev->mc.vram_end >> 12); + WREG32(MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR, + rdev->vram_scratch.gpu_addr >> 12); + tmp = ((rdev->mc.vram_end >> 24) & 0xFFFF) << 16; + tmp |= ((rdev->mc.vram_start >> 24) & 0xFFFF); + WREG32(MC_VM_FB_LOCATION, tmp); + /* XXX double check these! */ + WREG32(HDP_NONSURFACE_BASE, (rdev->mc.vram_start >> 8)); + WREG32(HDP_NONSURFACE_INFO, (2 << 7) | (1 << 30)); + WREG32(HDP_NONSURFACE_SIZE, 0x3FFFFFFF); + WREG32(MC_VM_AGP_BASE, 0); + WREG32(MC_VM_AGP_TOP, 0x0FFFFFFF); + WREG32(MC_VM_AGP_BOT, 0x0FFFFFFF); + if (radeon_mc_wait_for_idle(rdev)) { + dev_warn(rdev->dev, "Wait for MC idle timedout !\n"); + } + evergreen_mc_resume(rdev, &save); + /* we need to own VRAM, so turn off the VGA renderer here + * to stop it overwriting our objects */ + rv515_vga_render_disable(rdev); +} + +/* SI MC address space is 40 bits */ +static void si_vram_location(struct radeon_device *rdev, + struct radeon_mc *mc, u64 base) +{ + mc->vram_start = base; + if (mc->mc_vram_size > (0xFFFFFFFFFFULL - base + 1)) { + dev_warn(rdev->dev, "limiting VRAM to PCI aperture size\n"); + mc->real_vram_size = mc->aper_size; + mc->mc_vram_size = mc->aper_size; + } + mc->vram_end = mc->vram_start + mc->mc_vram_size - 1; + dev_info(rdev->dev, "VRAM: %lluM 0x%016llX - 0x%016llX (%lluM used)\n", + mc->mc_vram_size >> 20, mc->vram_start, + mc->vram_end, mc->real_vram_size >> 20); +} + +static void si_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc) +{ + u64 size_af, size_bf; + + size_af = ((0xFFFFFFFFFFULL - mc->vram_end) + mc->gtt_base_align) & ~mc->gtt_base_align; + size_bf = mc->vram_start & ~mc->gtt_base_align; + if (size_bf > size_af) { + if (mc->gtt_size > size_bf) { + dev_warn(rdev->dev, "limiting GTT\n"); + mc->gtt_size = size_bf; + } + mc->gtt_start = (mc->vram_start & ~mc->gtt_base_align) - mc->gtt_size; + } else { + if (mc->gtt_size > size_af) { + dev_warn(rdev->dev, "limiting GTT\n"); + mc->gtt_size = size_af; + } + mc->gtt_start = (mc->vram_end + 1 + mc->gtt_base_align) & ~mc->gtt_base_align; + } + mc->gtt_end = mc->gtt_start + mc->gtt_size - 1; + dev_info(rdev->dev, "GTT: %lluM 0x%016llX - 0x%016llX\n", + mc->gtt_size >> 20, mc->gtt_start, mc->gtt_end); +} + +static void si_vram_gtt_location(struct radeon_device *rdev, + struct radeon_mc *mc) +{ + if (mc->mc_vram_size > 0xFFC0000000ULL) { + /* leave room for at least 1024M GTT */ + dev_warn(rdev->dev, "limiting VRAM\n"); + mc->real_vram_size = 0xFFC0000000ULL; + mc->mc_vram_size = 0xFFC0000000ULL; + } + si_vram_location(rdev, &rdev->mc, 0); + rdev->mc.gtt_base_align = 0; + si_gtt_location(rdev, mc); +} + +static int si_mc_init(struct radeon_device *rdev) +{ + u32 tmp; + int chansize, numchan; + + /* Get VRAM informations */ + rdev->mc.vram_is_ddr = true; + tmp = RREG32(MC_ARB_RAMCFG); + if (tmp & CHANSIZE_OVERRIDE) { + chansize = 16; + } else if (tmp & CHANSIZE_MASK) { + chansize = 64; + } else { + chansize = 32; + } + tmp = RREG32(MC_SHARED_CHMAP); + switch ((tmp & NOOFCHAN_MASK) >> NOOFCHAN_SHIFT) { + case 0: + default: + numchan = 1; + break; + case 1: + numchan = 2; + break; + case 2: + numchan = 4; + break; + case 3: + numchan = 8; + break; + case 4: + numchan = 3; + break; + case 5: + numchan = 6; + break; + case 6: + numchan = 10; + break; + case 7: + numchan = 12; + break; + case 8: + numchan = 16; + break; + } + rdev->mc.vram_width = numchan * chansize; + /* Could aper size report 0 ? */ + rdev->mc.aper_base = pci_resource_start(rdev->pdev, 0); + rdev->mc.aper_size = pci_resource_len(rdev->pdev, 0); + /* size in MB on si */ + rdev->mc.mc_vram_size = RREG32(CONFIG_MEMSIZE) * 1024 * 1024; + rdev->mc.real_vram_size = RREG32(CONFIG_MEMSIZE) * 1024 * 1024; + rdev->mc.visible_vram_size = rdev->mc.aper_size; + si_vram_gtt_location(rdev, &rdev->mc); + radeon_update_bandwidth_info(rdev); + + return 0; +} + +/* + * GART + */ +void si_pcie_gart_tlb_flush(struct radeon_device *rdev) +{ + /* flush hdp cache */ + WREG32(HDP_MEM_COHERENCY_FLUSH_CNTL, 0x1); + + /* bits 0-15 are the VM contexts0-15 */ + WREG32(VM_INVALIDATE_REQUEST, 1); +} + +int si_pcie_gart_enable(struct radeon_device *rdev) +{ + int r, i; + + if (rdev->gart.robj == NULL) { + dev_err(rdev->dev, "No VRAM object for PCIE GART.\n"); + return -EINVAL; + } + r = radeon_gart_table_vram_pin(rdev); + if (r) + return r; + radeon_gart_restore(rdev); + /* Setup TLB control */ + WREG32(MC_VM_MX_L1_TLB_CNTL, + (0xA << 7) | + ENABLE_L1_TLB | + SYSTEM_ACCESS_MODE_NOT_IN_SYS | + ENABLE_ADVANCED_DRIVER_MODEL | + SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU); + /* Setup L2 cache */ + WREG32(VM_L2_CNTL, ENABLE_L2_CACHE | + ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE | + ENABLE_L2_PDE0_CACHE_LRU_UPDATE_BY_WRITE | + EFFECTIVE_L2_QUEUE_SIZE(7) | + CONTEXT1_IDENTITY_ACCESS_MODE(1)); + WREG32(VM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS | INVALIDATE_L2_CACHE); + WREG32(VM_L2_CNTL3, L2_CACHE_BIGK_ASSOCIATIVITY | + L2_CACHE_BIGK_FRAGMENT_SIZE(0)); + /* setup context0 */ + WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR, rdev->mc.gtt_start >> 12); + WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR, rdev->mc.gtt_end >> 12); + WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR, rdev->gart.table_addr >> 12); + WREG32(VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR, + (u32)(rdev->dummy_page.addr >> 12)); + WREG32(VM_CONTEXT0_CNTL2, 0); + WREG32(VM_CONTEXT0_CNTL, (ENABLE_CONTEXT | PAGE_TABLE_DEPTH(0) | + RANGE_PROTECTION_FAULT_ENABLE_DEFAULT)); + + WREG32(0x15D4, 0); + WREG32(0x15D8, 0); + WREG32(0x15DC, 0); + + /* empty context1-15 */ + /* FIXME start with 4G, once using 2 level pt switch to full + * vm size space + */ + /* set vm size, must be a multiple of 4 */ + WREG32(VM_CONTEXT1_PAGE_TABLE_START_ADDR, 0); + WREG32(VM_CONTEXT1_PAGE_TABLE_END_ADDR, rdev->vm_manager.max_pfn); + for (i = 1; i < 16; i++) { + if (i < 8) + WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2), + rdev->gart.table_addr >> 12); + else + WREG32(VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((i - 8) << 2), + rdev->gart.table_addr >> 12); + } + + /* enable context1-15 */ + WREG32(VM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR, + (u32)(rdev->dummy_page.addr >> 12)); + WREG32(VM_CONTEXT1_CNTL2, 0); + WREG32(VM_CONTEXT1_CNTL, ENABLE_CONTEXT | PAGE_TABLE_DEPTH(0) | + RANGE_PROTECTION_FAULT_ENABLE_DEFAULT); + + si_pcie_gart_tlb_flush(rdev); + DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n", + (unsigned)(rdev->mc.gtt_size >> 20), + (unsigned long long)rdev->gart.table_addr); + rdev->gart.ready = true; + return 0; +} + +void si_pcie_gart_disable(struct radeon_device *rdev) +{ + /* Disable all tables */ + WREG32(VM_CONTEXT0_CNTL, 0); + WREG32(VM_CONTEXT1_CNTL, 0); + /* Setup TLB control */ + WREG32(MC_VM_MX_L1_TLB_CNTL, SYSTEM_ACCESS_MODE_NOT_IN_SYS | + SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU); + /* Setup L2 cache */ + WREG32(VM_L2_CNTL, ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE | + ENABLE_L2_PDE0_CACHE_LRU_UPDATE_BY_WRITE | + EFFECTIVE_L2_QUEUE_SIZE(7) | + CONTEXT1_IDENTITY_ACCESS_MODE(1)); + WREG32(VM_L2_CNTL2, 0); + WREG32(VM_L2_CNTL3, L2_CACHE_BIGK_ASSOCIATIVITY | + L2_CACHE_BIGK_FRAGMENT_SIZE(0)); + radeon_gart_table_vram_unpin(rdev); +} + +void si_pcie_gart_fini(struct radeon_device *rdev) +{ + si_pcie_gart_disable(rdev); + radeon_gart_table_vram_free(rdev); + radeon_gart_fini(rdev); +} + +/* vm parser */ +static bool si_vm_reg_valid(u32 reg) +{ + /* context regs are fine */ + if (reg >= 0x28000) + return true; + + /* check config regs */ + switch (reg) { + case GRBM_GFX_INDEX: + case CP_STRMOUT_CNTL: + case VGT_VTX_VECT_EJECT_REG: + case VGT_CACHE_INVALIDATION: + case VGT_ESGS_RING_SIZE: + case VGT_GSVS_RING_SIZE: + case VGT_GS_VERTEX_REUSE: + case VGT_PRIMITIVE_TYPE: + case VGT_INDEX_TYPE: + case VGT_NUM_INDICES: + case VGT_NUM_INSTANCES: + case VGT_TF_RING_SIZE: + case VGT_HS_OFFCHIP_PARAM: + case VGT_TF_MEMORY_BASE: + case PA_CL_ENHANCE: + case PA_SU_LINE_STIPPLE_VALUE: + case PA_SC_LINE_STIPPLE_STATE: + case PA_SC_ENHANCE: + case SQC_CACHES: + case SPI_STATIC_THREAD_MGMT_1: + case SPI_STATIC_THREAD_MGMT_2: + case SPI_STATIC_THREAD_MGMT_3: + case SPI_PS_MAX_WAVE_ID: + case SPI_CONFIG_CNTL: + case SPI_CONFIG_CNTL_1: + case TA_CNTL_AUX: + return true; + default: + DRM_ERROR("Invalid register 0x%x in CS\n", reg); + return false; + } +} + +static int si_vm_packet3_ce_check(struct radeon_device *rdev, + u32 *ib, struct radeon_cs_packet *pkt) +{ + switch (pkt->opcode) { + case PACKET3_NOP: + case PACKET3_SET_BASE: + case PACKET3_SET_CE_DE_COUNTERS: + case PACKET3_LOAD_CONST_RAM: + case PACKET3_WRITE_CONST_RAM: + case PACKET3_WRITE_CONST_RAM_OFFSET: + case PACKET3_DUMP_CONST_RAM: + case PACKET3_INCREMENT_CE_COUNTER: + case PACKET3_WAIT_ON_DE_COUNTER: + case PACKET3_CE_WRITE: + break; + default: + DRM_ERROR("Invalid CE packet3: 0x%x\n", pkt->opcode); + return -EINVAL; + } + return 0; +} + +static int si_vm_packet3_gfx_check(struct radeon_device *rdev, + u32 *ib, struct radeon_cs_packet *pkt) +{ + u32 idx = pkt->idx + 1; + u32 idx_value = ib[idx]; + u32 start_reg, end_reg, reg, i; + + switch (pkt->opcode) { + case PACKET3_NOP: + case PACKET3_SET_BASE: + case PACKET3_CLEAR_STATE: + case PACKET3_INDEX_BUFFER_SIZE: + case PACKET3_DISPATCH_DIRECT: + case PACKET3_DISPATCH_INDIRECT: + case PACKET3_ALLOC_GDS: + case PACKET3_WRITE_GDS_RAM: + case PACKET3_ATOMIC_GDS: + case PACKET3_ATOMIC: + case PACKET3_OCCLUSION_QUERY: + case PACKET3_SET_PREDICATION: + case PACKET3_COND_EXEC: + case PACKET3_PRED_EXEC: + case PACKET3_DRAW_INDIRECT: + case PACKET3_DRAW_INDEX_INDIRECT: + case PACKET3_INDEX_BASE: + case PACKET3_DRAW_INDEX_2: + case PACKET3_CONTEXT_CONTROL: + case PACKET3_INDEX_TYPE: + case PACKET3_DRAW_INDIRECT_MULTI: + case PACKET3_DRAW_INDEX_AUTO: + case PACKET3_DRAW_INDEX_IMMD: + case PACKET3_NUM_INSTANCES: + case PACKET3_DRAW_INDEX_MULTI_AUTO: + case PACKET3_STRMOUT_BUFFER_UPDATE: + case PACKET3_DRAW_INDEX_OFFSET_2: + case PACKET3_DRAW_INDEX_MULTI_ELEMENT: + case PACKET3_DRAW_INDEX_INDIRECT_MULTI: + case PACKET3_MPEG_INDEX: + case PACKET3_WAIT_REG_MEM: + case PACKET3_MEM_WRITE: + case PACKET3_PFP_SYNC_ME: + case PACKET3_SURFACE_SYNC: + case PACKET3_EVENT_WRITE: + case PACKET3_EVENT_WRITE_EOP: + case PACKET3_EVENT_WRITE_EOS: + case PACKET3_SET_CONTEXT_REG: + case PACKET3_SET_CONTEXT_REG_INDIRECT: + case PACKET3_SET_SH_REG: + case PACKET3_SET_SH_REG_OFFSET: + case PACKET3_INCREMENT_DE_COUNTER: + case PACKET3_WAIT_ON_CE_COUNTER: + case PACKET3_WAIT_ON_AVAIL_BUFFER: + case PACKET3_ME_WRITE: + break; + case PACKET3_COPY_DATA: + if ((idx_value & 0xf00) == 0) { + reg = ib[idx + 3] * 4; + if (!si_vm_reg_valid(reg)) + return -EINVAL; + } + break; + case PACKET3_WRITE_DATA: + if ((idx_value & 0xf00) == 0) { + start_reg = ib[idx + 1] * 4; + if (idx_value & 0x10000) { + if (!si_vm_reg_valid(start_reg)) + return -EINVAL; + } else { + for (i = 0; i < (pkt->count - 2); i++) { + reg = start_reg + (4 * i); + if (!si_vm_reg_valid(reg)) + return -EINVAL; + } + } + } + break; + case PACKET3_COND_WRITE: + if (idx_value & 0x100) { + reg = ib[idx + 5] * 4; + if (!si_vm_reg_valid(reg)) + return -EINVAL; + } + break; + case PACKET3_COPY_DW: + if (idx_value & 0x2) { + reg = ib[idx + 3] * 4; + if (!si_vm_reg_valid(reg)) + return -EINVAL; + } + break; + case PACKET3_SET_CONFIG_REG: + start_reg = (idx_value << 2) + PACKET3_SET_CONFIG_REG_START; + end_reg = 4 * pkt->count + start_reg - 4; + if ((start_reg < PACKET3_SET_CONFIG_REG_START) || + (start_reg >= PACKET3_SET_CONFIG_REG_END) || + (end_reg >= PACKET3_SET_CONFIG_REG_END)) { + DRM_ERROR("bad PACKET3_SET_CONFIG_REG\n"); + return -EINVAL; + } + for (i = 0; i < pkt->count; i++) { + reg = start_reg + (4 * i); + if (!si_vm_reg_valid(reg)) + return -EINVAL; + } + break; + default: + DRM_ERROR("Invalid GFX packet3: 0x%x\n", pkt->opcode); + return -EINVAL; + } + return 0; +} + +static int si_vm_packet3_compute_check(struct radeon_device *rdev, + u32 *ib, struct radeon_cs_packet *pkt) +{ + u32 idx = pkt->idx + 1; + u32 idx_value = ib[idx]; + u32 start_reg, reg, i; + + switch (pkt->opcode) { + case PACKET3_NOP: + case PACKET3_SET_BASE: + case PACKET3_CLEAR_STATE: + case PACKET3_DISPATCH_DIRECT: + case PACKET3_DISPATCH_INDIRECT: + case PACKET3_ALLOC_GDS: + case PACKET3_WRITE_GDS_RAM: + case PACKET3_ATOMIC_GDS: + case PACKET3_ATOMIC: + case PACKET3_OCCLUSION_QUERY: + case PACKET3_SET_PREDICATION: + case PACKET3_COND_EXEC: + case PACKET3_PRED_EXEC: + case PACKET3_CONTEXT_CONTROL: + case PACKET3_STRMOUT_BUFFER_UPDATE: + case PACKET3_WAIT_REG_MEM: + case PACKET3_MEM_WRITE: + case PACKET3_PFP_SYNC_ME: + case PACKET3_SURFACE_SYNC: + case PACKET3_EVENT_WRITE: + case PACKET3_EVENT_WRITE_EOP: + case PACKET3_EVENT_WRITE_EOS: + case PACKET3_SET_CONTEXT_REG: + case PACKET3_SET_CONTEXT_REG_INDIRECT: + case PACKET3_SET_SH_REG: + case PACKET3_SET_SH_REG_OFFSET: + case PACKET3_INCREMENT_DE_COUNTER: + case PACKET3_WAIT_ON_CE_COUNTER: + case PACKET3_WAIT_ON_AVAIL_BUFFER: + case PACKET3_ME_WRITE: + break; + case PACKET3_COPY_DATA: + if ((idx_value & 0xf00) == 0) { + reg = ib[idx + 3] * 4; + if (!si_vm_reg_valid(reg)) + return -EINVAL; + } + break; + case PACKET3_WRITE_DATA: + if ((idx_value & 0xf00) == 0) { + start_reg = ib[idx + 1] * 4; + if (idx_value & 0x10000) { + if (!si_vm_reg_valid(start_reg)) + return -EINVAL; + } else { + for (i = 0; i < (pkt->count - 2); i++) { + reg = start_reg + (4 * i); + if (!si_vm_reg_valid(reg)) + return -EINVAL; + } + } + } + break; + case PACKET3_COND_WRITE: + if (idx_value & 0x100) { + reg = ib[idx + 5] * 4; + if (!si_vm_reg_valid(reg)) + return -EINVAL; + } + break; + case PACKET3_COPY_DW: + if (idx_value & 0x2) { + reg = ib[idx + 3] * 4; + if (!si_vm_reg_valid(reg)) + return -EINVAL; + } + break; + default: + DRM_ERROR("Invalid Compute packet3: 0x%x\n", pkt->opcode); + return -EINVAL; + } + return 0; +} + +int si_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib) +{ + int ret = 0; + u32 idx = 0; + struct radeon_cs_packet pkt; + + do { + pkt.idx = idx; + pkt.type = CP_PACKET_GET_TYPE(ib->ptr[idx]); + pkt.count = CP_PACKET_GET_COUNT(ib->ptr[idx]); + pkt.one_reg_wr = 0; + switch (pkt.type) { + case PACKET_TYPE0: + dev_err(rdev->dev, "Packet0 not allowed!\n"); + ret = -EINVAL; + break; + case PACKET_TYPE2: + idx += 1; + break; + case PACKET_TYPE3: + pkt.opcode = CP_PACKET3_GET_OPCODE(ib->ptr[idx]); + if (ib->is_const_ib) + ret = si_vm_packet3_ce_check(rdev, ib->ptr, &pkt); + else { + switch (ib->fence->ring) { + case RADEON_RING_TYPE_GFX_INDEX: + ret = si_vm_packet3_gfx_check(rdev, ib->ptr, &pkt); + break; + case CAYMAN_RING_TYPE_CP1_INDEX: + case CAYMAN_RING_TYPE_CP2_INDEX: + ret = si_vm_packet3_compute_check(rdev, ib->ptr, &pkt); + break; + default: + dev_err(rdev->dev, "Non-PM4 ring %d !\n", ib->fence->ring); + ret = -EINVAL; + break; + } + } + idx += pkt.count + 2; + break; + default: + dev_err(rdev->dev, "Unknown packet type %d !\n", pkt.type); + ret = -EINVAL; + break; + } + if (ret) + break; + } while (idx < ib->length_dw); + + return ret; +} + +/* + * vm + */ +int si_vm_init(struct radeon_device *rdev) +{ + /* number of VMs */ + rdev->vm_manager.nvm = 16; + /* base offset of vram pages */ + rdev->vm_manager.vram_base_offset = 0; + + return 0; +} + +void si_vm_fini(struct radeon_device *rdev) +{ +} + +int si_vm_bind(struct radeon_device *rdev, struct radeon_vm *vm, int id) +{ + if (id < 8) + WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (id << 2), vm->pt_gpu_addr >> 12); + else + WREG32(VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((id - 8) << 2), + vm->pt_gpu_addr >> 12); + /* flush hdp cache */ + WREG32(HDP_MEM_COHERENCY_FLUSH_CNTL, 0x1); + /* bits 0-15 are the VM contexts0-15 */ + WREG32(VM_INVALIDATE_REQUEST, 1 << id); + return 0; +} + +void si_vm_unbind(struct radeon_device *rdev, struct radeon_vm *vm) +{ + if (vm->id < 8) + WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2), 0); + else + WREG32(VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm->id - 8) << 2), 0); + /* flush hdp cache */ + WREG32(HDP_MEM_COHERENCY_FLUSH_CNTL, 0x1); + /* bits 0-15 are the VM contexts0-15 */ + WREG32(VM_INVALIDATE_REQUEST, 1 << vm->id); +} + +void si_vm_tlb_flush(struct radeon_device *rdev, struct radeon_vm *vm) +{ + if (vm->id == -1) + return; + + /* flush hdp cache */ + WREG32(HDP_MEM_COHERENCY_FLUSH_CNTL, 0x1); + /* bits 0-15 are the VM contexts0-15 */ + WREG32(VM_INVALIDATE_REQUEST, 1 << vm->id); +} + +/* + * RLC + */ +void si_rlc_fini(struct radeon_device *rdev) +{ + int r; + + /* save restore block */ + if (rdev->rlc.save_restore_obj) { + r = radeon_bo_reserve(rdev->rlc.save_restore_obj, false); + if (unlikely(r != 0)) + dev_warn(rdev->dev, "(%d) reserve RLC sr bo failed\n", r); + radeon_bo_unpin(rdev->rlc.save_restore_obj); + radeon_bo_unreserve(rdev->rlc.save_restore_obj); + + radeon_bo_unref(&rdev->rlc.save_restore_obj); + rdev->rlc.save_restore_obj = NULL; + } + + /* clear state block */ + if (rdev->rlc.clear_state_obj) { + r = radeon_bo_reserve(rdev->rlc.clear_state_obj, false); + if (unlikely(r != 0)) + dev_warn(rdev->dev, "(%d) reserve RLC c bo failed\n", r); + radeon_bo_unpin(rdev->rlc.clear_state_obj); + radeon_bo_unreserve(rdev->rlc.clear_state_obj); + + radeon_bo_unref(&rdev->rlc.clear_state_obj); + rdev->rlc.clear_state_obj = NULL; + } +} + +int si_rlc_init(struct radeon_device *rdev) +{ + int r; + + /* save restore block */ + if (rdev->rlc.save_restore_obj == NULL) { + r = radeon_bo_create(rdev, RADEON_GPU_PAGE_SIZE, PAGE_SIZE, true, + RADEON_GEM_DOMAIN_VRAM, &rdev->rlc.save_restore_obj); + if (r) { + dev_warn(rdev->dev, "(%d) create RLC sr bo failed\n", r); + return r; + } + } + + r = radeon_bo_reserve(rdev->rlc.save_restore_obj, false); + if (unlikely(r != 0)) { + si_rlc_fini(rdev); + return r; + } + r = radeon_bo_pin(rdev->rlc.save_restore_obj, RADEON_GEM_DOMAIN_VRAM, + &rdev->rlc.save_restore_gpu_addr); + radeon_bo_unreserve(rdev->rlc.save_restore_obj); + if (r) { + dev_warn(rdev->dev, "(%d) pin RLC sr bo failed\n", r); + si_rlc_fini(rdev); + return r; + } + + /* clear state block */ + if (rdev->rlc.clear_state_obj == NULL) { + r = radeon_bo_create(rdev, RADEON_GPU_PAGE_SIZE, PAGE_SIZE, true, + RADEON_GEM_DOMAIN_VRAM, &rdev->rlc.clear_state_obj); + if (r) { + dev_warn(rdev->dev, "(%d) create RLC c bo failed\n", r); + si_rlc_fini(rdev); + return r; + } + } + r = radeon_bo_reserve(rdev->rlc.clear_state_obj, false); + if (unlikely(r != 0)) { + si_rlc_fini(rdev); + return r; + } + r = radeon_bo_pin(rdev->rlc.clear_state_obj, RADEON_GEM_DOMAIN_VRAM, + &rdev->rlc.clear_state_gpu_addr); + radeon_bo_unreserve(rdev->rlc.clear_state_obj); + if (r) { + dev_warn(rdev->dev, "(%d) pin RLC c bo failed\n", r); + si_rlc_fini(rdev); + return r; + } + + return 0; +} + +static void si_rlc_stop(struct radeon_device *rdev) +{ + WREG32(RLC_CNTL, 0); +} + +static void si_rlc_start(struct radeon_device *rdev) +{ + WREG32(RLC_CNTL, RLC_ENABLE); +} + +static int si_rlc_resume(struct radeon_device *rdev) +{ + u32 i; + const __be32 *fw_data; + + if (!rdev->rlc_fw) + return -EINVAL; + + si_rlc_stop(rdev); + + WREG32(RLC_RL_BASE, 0); + WREG32(RLC_RL_SIZE, 0); + WREG32(RLC_LB_CNTL, 0); + WREG32(RLC_LB_CNTR_MAX, 0xffffffff); + WREG32(RLC_LB_CNTR_INIT, 0); + + WREG32(RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8); + WREG32(RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8); + + WREG32(RLC_MC_CNTL, 0); + WREG32(RLC_UCODE_CNTL, 0); + + fw_data = (const __be32 *)rdev->rlc_fw->data; + for (i = 0; i < SI_RLC_UCODE_SIZE; i++) { + WREG32(RLC_UCODE_ADDR, i); + WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++)); + } + WREG32(RLC_UCODE_ADDR, 0); + + si_rlc_start(rdev); + + return 0; +} + +static void si_enable_interrupts(struct radeon_device *rdev) +{ + u32 ih_cntl = RREG32(IH_CNTL); + u32 ih_rb_cntl = RREG32(IH_RB_CNTL); + + ih_cntl |= ENABLE_INTR; + ih_rb_cntl |= IH_RB_ENABLE; + WREG32(IH_CNTL, ih_cntl); + WREG32(IH_RB_CNTL, ih_rb_cntl); + rdev->ih.enabled = true; +} + +static void si_disable_interrupts(struct radeon_device *rdev) +{ + u32 ih_rb_cntl = RREG32(IH_RB_CNTL); + u32 ih_cntl = RREG32(IH_CNTL); + + ih_rb_cntl &= ~IH_RB_ENABLE; + ih_cntl &= ~ENABLE_INTR; + WREG32(IH_RB_CNTL, ih_rb_cntl); + WREG32(IH_CNTL, ih_cntl); + /* set rptr, wptr to 0 */ + WREG32(IH_RB_RPTR, 0); + WREG32(IH_RB_WPTR, 0); + rdev->ih.enabled = false; + rdev->ih.wptr = 0; + rdev->ih.rptr = 0; +} + +static void si_disable_interrupt_state(struct radeon_device *rdev) +{ + u32 tmp; + + WREG32(CP_INT_CNTL_RING0, CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE); + WREG32(CP_INT_CNTL_RING1, 0); + WREG32(CP_INT_CNTL_RING2, 0); + WREG32(GRBM_INT_CNTL, 0); + WREG32(INT_MASK + EVERGREEN_CRTC0_REGISTER_OFFSET, 0); + WREG32(INT_MASK + EVERGREEN_CRTC1_REGISTER_OFFSET, 0); + if (rdev->num_crtc >= 4) { + WREG32(INT_MASK + EVERGREEN_CRTC2_REGISTER_OFFSET, 0); + WREG32(INT_MASK + EVERGREEN_CRTC3_REGISTER_OFFSET, 0); + } + if (rdev->num_crtc >= 6) { + WREG32(INT_MASK + EVERGREEN_CRTC4_REGISTER_OFFSET, 0); + WREG32(INT_MASK + EVERGREEN_CRTC5_REGISTER_OFFSET, 0); + } + + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, 0); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, 0); + if (rdev->num_crtc >= 4) { + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET, 0); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET, 0); + } + if (rdev->num_crtc >= 6) { + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET, 0); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, 0); + } + + WREG32(DACA_AUTODETECT_INT_CONTROL, 0); + + tmp = RREG32(DC_HPD1_INT_CONTROL) & DC_HPDx_INT_POLARITY; + WREG32(DC_HPD1_INT_CONTROL, tmp); + tmp = RREG32(DC_HPD2_INT_CONTROL) & DC_HPDx_INT_POLARITY; + WREG32(DC_HPD2_INT_CONTROL, tmp); + tmp = RREG32(DC_HPD3_INT_CONTROL) & DC_HPDx_INT_POLARITY; + WREG32(DC_HPD3_INT_CONTROL, tmp); + tmp = RREG32(DC_HPD4_INT_CONTROL) & DC_HPDx_INT_POLARITY; + WREG32(DC_HPD4_INT_CONTROL, tmp); + tmp = RREG32(DC_HPD5_INT_CONTROL) & DC_HPDx_INT_POLARITY; + WREG32(DC_HPD5_INT_CONTROL, tmp); + tmp = RREG32(DC_HPD6_INT_CONTROL) & DC_HPDx_INT_POLARITY; + WREG32(DC_HPD6_INT_CONTROL, tmp); + +} + +static int si_irq_init(struct radeon_device *rdev) +{ + int ret = 0; + int rb_bufsz; + u32 interrupt_cntl, ih_cntl, ih_rb_cntl; + + /* allocate ring */ + ret = r600_ih_ring_alloc(rdev); + if (ret) + return ret; + + /* disable irqs */ + si_disable_interrupts(rdev); + + /* init rlc */ + ret = si_rlc_resume(rdev); + if (ret) { + r600_ih_ring_fini(rdev); + return ret; + } + + /* setup interrupt control */ + /* set dummy read address to ring address */ + WREG32(INTERRUPT_CNTL2, rdev->ih.gpu_addr >> 8); + interrupt_cntl = RREG32(INTERRUPT_CNTL); + /* IH_DUMMY_RD_OVERRIDE=0 - dummy read disabled with msi, enabled without msi + * IH_DUMMY_RD_OVERRIDE=1 - dummy read controlled by IH_DUMMY_RD_EN + */ + interrupt_cntl &= ~IH_DUMMY_RD_OVERRIDE; + /* IH_REQ_NONSNOOP_EN=1 if ring is in non-cacheable memory, e.g., vram */ + interrupt_cntl &= ~IH_REQ_NONSNOOP_EN; + WREG32(INTERRUPT_CNTL, interrupt_cntl); + + WREG32(IH_RB_BASE, rdev->ih.gpu_addr >> 8); + rb_bufsz = drm_order(rdev->ih.ring_size / 4); + + ih_rb_cntl = (IH_WPTR_OVERFLOW_ENABLE | + IH_WPTR_OVERFLOW_CLEAR | + (rb_bufsz << 1)); + + if (rdev->wb.enabled) + ih_rb_cntl |= IH_WPTR_WRITEBACK_ENABLE; + + /* set the writeback address whether it's enabled or not */ + WREG32(IH_RB_WPTR_ADDR_LO, (rdev->wb.gpu_addr + R600_WB_IH_WPTR_OFFSET) & 0xFFFFFFFC); + WREG32(IH_RB_WPTR_ADDR_HI, upper_32_bits(rdev->wb.gpu_addr + R600_WB_IH_WPTR_OFFSET) & 0xFF); + + WREG32(IH_RB_CNTL, ih_rb_cntl); + + /* set rptr, wptr to 0 */ + WREG32(IH_RB_RPTR, 0); + WREG32(IH_RB_WPTR, 0); + + /* Default settings for IH_CNTL (disabled at first) */ + ih_cntl = MC_WRREQ_CREDIT(0x10) | MC_WR_CLEAN_CNT(0x10) | MC_VMID(0); + /* RPTR_REARM only works if msi's are enabled */ + if (rdev->msi_enabled) + ih_cntl |= RPTR_REARM; + WREG32(IH_CNTL, ih_cntl); + + /* force the active interrupt state to all disabled */ + si_disable_interrupt_state(rdev); + + /* enable irqs */ + si_enable_interrupts(rdev); + + return ret; +} + +int si_irq_set(struct radeon_device *rdev) +{ + u32 cp_int_cntl = CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE; + u32 cp_int_cntl1 = 0, cp_int_cntl2 = 0; + u32 crtc1 = 0, crtc2 = 0, crtc3 = 0, crtc4 = 0, crtc5 = 0, crtc6 = 0; + u32 hpd1, hpd2, hpd3, hpd4, hpd5, hpd6; + u32 grbm_int_cntl = 0; + u32 grph1 = 0, grph2 = 0, grph3 = 0, grph4 = 0, grph5 = 0, grph6 = 0; + + if (!rdev->irq.installed) { + WARN(1, "Can't enable IRQ/MSI because no handler is installed\n"); + return -EINVAL; + } + /* don't enable anything if the ih is disabled */ + if (!rdev->ih.enabled) { + si_disable_interrupts(rdev); + /* force the active interrupt state to all disabled */ + si_disable_interrupt_state(rdev); + return 0; + } + + hpd1 = RREG32(DC_HPD1_INT_CONTROL) & ~DC_HPDx_INT_EN; + hpd2 = RREG32(DC_HPD2_INT_CONTROL) & ~DC_HPDx_INT_EN; + hpd3 = RREG32(DC_HPD3_INT_CONTROL) & ~DC_HPDx_INT_EN; + hpd4 = RREG32(DC_HPD4_INT_CONTROL) & ~DC_HPDx_INT_EN; + hpd5 = RREG32(DC_HPD5_INT_CONTROL) & ~DC_HPDx_INT_EN; + hpd6 = RREG32(DC_HPD6_INT_CONTROL) & ~DC_HPDx_INT_EN; + + /* enable CP interrupts on all rings */ + if (rdev->irq.sw_int[RADEON_RING_TYPE_GFX_INDEX]) { + DRM_DEBUG("si_irq_set: sw int gfx\n"); + cp_int_cntl |= TIME_STAMP_INT_ENABLE; + } + if (rdev->irq.sw_int[CAYMAN_RING_TYPE_CP1_INDEX]) { + DRM_DEBUG("si_irq_set: sw int cp1\n"); + cp_int_cntl1 |= TIME_STAMP_INT_ENABLE; + } + if (rdev->irq.sw_int[CAYMAN_RING_TYPE_CP2_INDEX]) { + DRM_DEBUG("si_irq_set: sw int cp2\n"); + cp_int_cntl2 |= TIME_STAMP_INT_ENABLE; + } + if (rdev->irq.crtc_vblank_int[0] || + rdev->irq.pflip[0]) { + DRM_DEBUG("si_irq_set: vblank 0\n"); + crtc1 |= VBLANK_INT_MASK; + } + if (rdev->irq.crtc_vblank_int[1] || + rdev->irq.pflip[1]) { + DRM_DEBUG("si_irq_set: vblank 1\n"); + crtc2 |= VBLANK_INT_MASK; + } + if (rdev->irq.crtc_vblank_int[2] || + rdev->irq.pflip[2]) { + DRM_DEBUG("si_irq_set: vblank 2\n"); + crtc3 |= VBLANK_INT_MASK; + } + if (rdev->irq.crtc_vblank_int[3] || + rdev->irq.pflip[3]) { + DRM_DEBUG("si_irq_set: vblank 3\n"); + crtc4 |= VBLANK_INT_MASK; + } + if (rdev->irq.crtc_vblank_int[4] || + rdev->irq.pflip[4]) { + DRM_DEBUG("si_irq_set: vblank 4\n"); + crtc5 |= VBLANK_INT_MASK; + } + if (rdev->irq.crtc_vblank_int[5] || + rdev->irq.pflip[5]) { + DRM_DEBUG("si_irq_set: vblank 5\n"); + crtc6 |= VBLANK_INT_MASK; + } + if (rdev->irq.hpd[0]) { + DRM_DEBUG("si_irq_set: hpd 1\n"); + hpd1 |= DC_HPDx_INT_EN; + } + if (rdev->irq.hpd[1]) { + DRM_DEBUG("si_irq_set: hpd 2\n"); + hpd2 |= DC_HPDx_INT_EN; + } + if (rdev->irq.hpd[2]) { + DRM_DEBUG("si_irq_set: hpd 3\n"); + hpd3 |= DC_HPDx_INT_EN; + } + if (rdev->irq.hpd[3]) { + DRM_DEBUG("si_irq_set: hpd 4\n"); + hpd4 |= DC_HPDx_INT_EN; + } + if (rdev->irq.hpd[4]) { + DRM_DEBUG("si_irq_set: hpd 5\n"); + hpd5 |= DC_HPDx_INT_EN; + } + if (rdev->irq.hpd[5]) { + DRM_DEBUG("si_irq_set: hpd 6\n"); + hpd6 |= DC_HPDx_INT_EN; + } + if (rdev->irq.gui_idle) { + DRM_DEBUG("gui idle\n"); + grbm_int_cntl |= GUI_IDLE_INT_ENABLE; + } + + WREG32(CP_INT_CNTL_RING0, cp_int_cntl); + WREG32(CP_INT_CNTL_RING1, cp_int_cntl1); + WREG32(CP_INT_CNTL_RING2, cp_int_cntl2); + + WREG32(GRBM_INT_CNTL, grbm_int_cntl); + + WREG32(INT_MASK + EVERGREEN_CRTC0_REGISTER_OFFSET, crtc1); + WREG32(INT_MASK + EVERGREEN_CRTC1_REGISTER_OFFSET, crtc2); + if (rdev->num_crtc >= 4) { + WREG32(INT_MASK + EVERGREEN_CRTC2_REGISTER_OFFSET, crtc3); + WREG32(INT_MASK + EVERGREEN_CRTC3_REGISTER_OFFSET, crtc4); + } + if (rdev->num_crtc >= 6) { + WREG32(INT_MASK + EVERGREEN_CRTC4_REGISTER_OFFSET, crtc5); + WREG32(INT_MASK + EVERGREEN_CRTC5_REGISTER_OFFSET, crtc6); + } + + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, grph1); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, grph2); + if (rdev->num_crtc >= 4) { + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC2_REGISTER_OFFSET, grph3); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC3_REGISTER_OFFSET, grph4); + } + if (rdev->num_crtc >= 6) { + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC4_REGISTER_OFFSET, grph5); + WREG32(GRPH_INT_CONTROL + EVERGREEN_CRTC5_REGISTER_OFFSET, grph6); + } + + WREG32(DC_HPD1_INT_CONTROL, hpd1); + WREG32(DC_HPD2_INT_CONTROL, hpd2); + WREG32(DC_HPD3_INT_CONTROL, hpd3); + WREG32(DC_HPD4_INT_CONTROL, hpd4); + WREG32(DC_HPD5_INT_CONTROL, hpd5); + WREG32(DC_HPD6_INT_CONTROL, hpd6); + + return 0; +} + +static inline void si_irq_ack(struct radeon_device *rdev) +{ + u32 tmp; + + rdev->irq.stat_regs.evergreen.disp_int = RREG32(DISP_INTERRUPT_STATUS); + rdev->irq.stat_regs.evergreen.disp_int_cont = RREG32(DISP_INTERRUPT_STATUS_CONTINUE); + rdev->irq.stat_regs.evergreen.disp_int_cont2 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE2); + rdev->irq.stat_regs.evergreen.disp_int_cont3 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE3); + rdev->irq.stat_regs.evergreen.disp_int_cont4 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE4); + rdev->irq.stat_regs.evergreen.disp_int_cont5 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE5); + rdev->irq.stat_regs.evergreen.d1grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET); + rdev->irq.stat_regs.evergreen.d2grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET); + if (rdev->num_crtc >= 4) { + rdev->irq.stat_regs.evergreen.d3grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET); + rdev->irq.stat_regs.evergreen.d4grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET); + } + if (rdev->num_crtc >= 6) { + rdev->irq.stat_regs.evergreen.d5grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET); + rdev->irq.stat_regs.evergreen.d6grph_int = RREG32(GRPH_INT_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET); + } + + if (rdev->irq.stat_regs.evergreen.d1grph_int & GRPH_PFLIP_INT_OCCURRED) + WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR); + if (rdev->irq.stat_regs.evergreen.d2grph_int & GRPH_PFLIP_INT_OCCURRED) + WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR); + if (rdev->irq.stat_regs.evergreen.disp_int & LB_D1_VBLANK_INTERRUPT) + WREG32(VBLANK_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET, VBLANK_ACK); + if (rdev->irq.stat_regs.evergreen.disp_int & LB_D1_VLINE_INTERRUPT) + WREG32(VLINE_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET, VLINE_ACK); + if (rdev->irq.stat_regs.evergreen.disp_int_cont & LB_D2_VBLANK_INTERRUPT) + WREG32(VBLANK_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET, VBLANK_ACK); + if (rdev->irq.stat_regs.evergreen.disp_int_cont & LB_D2_VLINE_INTERRUPT) + WREG32(VLINE_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET, VLINE_ACK); + + if (rdev->num_crtc >= 4) { + if (rdev->irq.stat_regs.evergreen.d3grph_int & GRPH_PFLIP_INT_OCCURRED) + WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR); + if (rdev->irq.stat_regs.evergreen.d4grph_int & GRPH_PFLIP_INT_OCCURRED) + WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR); + if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & LB_D3_VBLANK_INTERRUPT) + WREG32(VBLANK_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET, VBLANK_ACK); + if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & LB_D3_VLINE_INTERRUPT) + WREG32(VLINE_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET, VLINE_ACK); + if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & LB_D4_VBLANK_INTERRUPT) + WREG32(VBLANK_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET, VBLANK_ACK); + if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & LB_D4_VLINE_INTERRUPT) + WREG32(VLINE_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET, VLINE_ACK); + } + + if (rdev->num_crtc >= 6) { + if (rdev->irq.stat_regs.evergreen.d5grph_int & GRPH_PFLIP_INT_OCCURRED) + WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR); + if (rdev->irq.stat_regs.evergreen.d6grph_int & GRPH_PFLIP_INT_OCCURRED) + WREG32(GRPH_INT_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET, GRPH_PFLIP_INT_CLEAR); + if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & LB_D5_VBLANK_INTERRUPT) + WREG32(VBLANK_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET, VBLANK_ACK); + if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & LB_D5_VLINE_INTERRUPT) + WREG32(VLINE_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET, VLINE_ACK); + if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & LB_D6_VBLANK_INTERRUPT) + WREG32(VBLANK_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET, VBLANK_ACK); + if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & LB_D6_VLINE_INTERRUPT) + WREG32(VLINE_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET, VLINE_ACK); + } + + if (rdev->irq.stat_regs.evergreen.disp_int & DC_HPD1_INTERRUPT) { + tmp = RREG32(DC_HPD1_INT_CONTROL); + tmp |= DC_HPDx_INT_ACK; + WREG32(DC_HPD1_INT_CONTROL, tmp); + } + if (rdev->irq.stat_regs.evergreen.disp_int_cont & DC_HPD2_INTERRUPT) { + tmp = RREG32(DC_HPD2_INT_CONTROL); + tmp |= DC_HPDx_INT_ACK; + WREG32(DC_HPD2_INT_CONTROL, tmp); + } + if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & DC_HPD3_INTERRUPT) { + tmp = RREG32(DC_HPD3_INT_CONTROL); + tmp |= DC_HPDx_INT_ACK; + WREG32(DC_HPD3_INT_CONTROL, tmp); + } + if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & DC_HPD4_INTERRUPT) { + tmp = RREG32(DC_HPD4_INT_CONTROL); + tmp |= DC_HPDx_INT_ACK; + WREG32(DC_HPD4_INT_CONTROL, tmp); + } + if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & DC_HPD5_INTERRUPT) { + tmp = RREG32(DC_HPD5_INT_CONTROL); + tmp |= DC_HPDx_INT_ACK; + WREG32(DC_HPD5_INT_CONTROL, tmp); + } + if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & DC_HPD6_INTERRUPT) { + tmp = RREG32(DC_HPD5_INT_CONTROL); + tmp |= DC_HPDx_INT_ACK; + WREG32(DC_HPD6_INT_CONTROL, tmp); + } +} + +static void si_irq_disable(struct radeon_device *rdev) +{ + si_disable_interrupts(rdev); + /* Wait and acknowledge irq */ + mdelay(1); + si_irq_ack(rdev); + si_disable_interrupt_state(rdev); +} + +static void si_irq_suspend(struct radeon_device *rdev) +{ + si_irq_disable(rdev); + si_rlc_stop(rdev); +} + +static void si_irq_fini(struct radeon_device *rdev) +{ + si_irq_suspend(rdev); + r600_ih_ring_fini(rdev); +} + +static inline u32 si_get_ih_wptr(struct radeon_device *rdev) +{ + u32 wptr, tmp; + + if (rdev->wb.enabled) + wptr = le32_to_cpu(rdev->wb.wb[R600_WB_IH_WPTR_OFFSET/4]); + else + wptr = RREG32(IH_RB_WPTR); + + if (wptr & RB_OVERFLOW) { + /* When a ring buffer overflow happen start parsing interrupt + * from the last not overwritten vector (wptr + 16). Hopefully + * this should allow us to catchup. + */ + dev_warn(rdev->dev, "IH ring buffer overflow (0x%08X, %d, %d)\n", + wptr, rdev->ih.rptr, (wptr + 16) + rdev->ih.ptr_mask); + rdev->ih.rptr = (wptr + 16) & rdev->ih.ptr_mask; + tmp = RREG32(IH_RB_CNTL); + tmp |= IH_WPTR_OVERFLOW_CLEAR; + WREG32(IH_RB_CNTL, tmp); + } + return (wptr & rdev->ih.ptr_mask); +} + +/* SI IV Ring + * Each IV ring entry is 128 bits: + * [7:0] - interrupt source id + * [31:8] - reserved + * [59:32] - interrupt source data + * [63:60] - reserved + * [71:64] - RINGID + * [79:72] - VMID + * [127:80] - reserved + */ +int si_irq_process(struct radeon_device *rdev) +{ + u32 wptr; + u32 rptr; + u32 src_id, src_data, ring_id; + u32 ring_index; + unsigned long flags; + bool queue_hotplug = false; + + if (!rdev->ih.enabled || rdev->shutdown) + return IRQ_NONE; + + wptr = si_get_ih_wptr(rdev); + rptr = rdev->ih.rptr; + DRM_DEBUG("si_irq_process start: rptr %d, wptr %d\n", rptr, wptr); + + spin_lock_irqsave(&rdev->ih.lock, flags); + if (rptr == wptr) { + spin_unlock_irqrestore(&rdev->ih.lock, flags); + return IRQ_NONE; + } +restart_ih: + /* Order reading of wptr vs. reading of IH ring data */ + rmb(); + + /* display interrupts */ + si_irq_ack(rdev); + + rdev->ih.wptr = wptr; + while (rptr != wptr) { + /* wptr/rptr are in bytes! */ + ring_index = rptr / 4; + src_id = le32_to_cpu(rdev->ih.ring[ring_index]) & 0xff; + src_data = le32_to_cpu(rdev->ih.ring[ring_index + 1]) & 0xfffffff; + ring_id = le32_to_cpu(rdev->ih.ring[ring_index + 2]) & 0xff; + + switch (src_id) { + case 1: /* D1 vblank/vline */ + switch (src_data) { + case 0: /* D1 vblank */ + if (rdev->irq.stat_regs.evergreen.disp_int & LB_D1_VBLANK_INTERRUPT) { + if (rdev->irq.crtc_vblank_int[0]) { + drm_handle_vblank(rdev->ddev, 0); + rdev->pm.vblank_sync = true; + wake_up(&rdev->irq.vblank_queue); + } + if (rdev->irq.pflip[0]) + radeon_crtc_handle_flip(rdev, 0); + rdev->irq.stat_regs.evergreen.disp_int &= ~LB_D1_VBLANK_INTERRUPT; + DRM_DEBUG("IH: D1 vblank\n"); + } + break; + case 1: /* D1 vline */ + if (rdev->irq.stat_regs.evergreen.disp_int & LB_D1_VLINE_INTERRUPT) { + rdev->irq.stat_regs.evergreen.disp_int &= ~LB_D1_VLINE_INTERRUPT; + DRM_DEBUG("IH: D1 vline\n"); + } + break; + default: + DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data); + break; + } + break; + case 2: /* D2 vblank/vline */ + switch (src_data) { + case 0: /* D2 vblank */ + if (rdev->irq.stat_regs.evergreen.disp_int_cont & LB_D2_VBLANK_INTERRUPT) { + if (rdev->irq.crtc_vblank_int[1]) { + drm_handle_vblank(rdev->ddev, 1); + rdev->pm.vblank_sync = true; + wake_up(&rdev->irq.vblank_queue); + } + if (rdev->irq.pflip[1]) + radeon_crtc_handle_flip(rdev, 1); + rdev->irq.stat_regs.evergreen.disp_int_cont &= ~LB_D2_VBLANK_INTERRUPT; + DRM_DEBUG("IH: D2 vblank\n"); + } + break; + case 1: /* D2 vline */ + if (rdev->irq.stat_regs.evergreen.disp_int_cont & LB_D2_VLINE_INTERRUPT) { + rdev->irq.stat_regs.evergreen.disp_int_cont &= ~LB_D2_VLINE_INTERRUPT; + DRM_DEBUG("IH: D2 vline\n"); + } + break; + default: + DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data); + break; + } + break; + case 3: /* D3 vblank/vline */ + switch (src_data) { + case 0: /* D3 vblank */ + if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & LB_D3_VBLANK_INTERRUPT) { + if (rdev->irq.crtc_vblank_int[2]) { + drm_handle_vblank(rdev->ddev, 2); + rdev->pm.vblank_sync = true; + wake_up(&rdev->irq.vblank_queue); + } + if (rdev->irq.pflip[2]) + radeon_crtc_handle_flip(rdev, 2); + rdev->irq.stat_regs.evergreen.disp_int_cont2 &= ~LB_D3_VBLANK_INTERRUPT; + DRM_DEBUG("IH: D3 vblank\n"); + } + break; + case 1: /* D3 vline */ + if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & LB_D3_VLINE_INTERRUPT) { + rdev->irq.stat_regs.evergreen.disp_int_cont2 &= ~LB_D3_VLINE_INTERRUPT; + DRM_DEBUG("IH: D3 vline\n"); + } + break; + default: + DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data); + break; + } + break; + case 4: /* D4 vblank/vline */ + switch (src_data) { + case 0: /* D4 vblank */ + if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & LB_D4_VBLANK_INTERRUPT) { + if (rdev->irq.crtc_vblank_int[3]) { + drm_handle_vblank(rdev->ddev, 3); + rdev->pm.vblank_sync = true; + wake_up(&rdev->irq.vblank_queue); + } + if (rdev->irq.pflip[3]) + radeon_crtc_handle_flip(rdev, 3); + rdev->irq.stat_regs.evergreen.disp_int_cont3 &= ~LB_D4_VBLANK_INTERRUPT; + DRM_DEBUG("IH: D4 vblank\n"); + } + break; + case 1: /* D4 vline */ + if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & LB_D4_VLINE_INTERRUPT) { + rdev->irq.stat_regs.evergreen.disp_int_cont3 &= ~LB_D4_VLINE_INTERRUPT; + DRM_DEBUG("IH: D4 vline\n"); + } + break; + default: + DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data); + break; + } + break; + case 5: /* D5 vblank/vline */ + switch (src_data) { + case 0: /* D5 vblank */ + if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & LB_D5_VBLANK_INTERRUPT) { + if (rdev->irq.crtc_vblank_int[4]) { + drm_handle_vblank(rdev->ddev, 4); + rdev->pm.vblank_sync = true; + wake_up(&rdev->irq.vblank_queue); + } + if (rdev->irq.pflip[4]) + radeon_crtc_handle_flip(rdev, 4); + rdev->irq.stat_regs.evergreen.disp_int_cont4 &= ~LB_D5_VBLANK_INTERRUPT; + DRM_DEBUG("IH: D5 vblank\n"); + } + break; + case 1: /* D5 vline */ + if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & LB_D5_VLINE_INTERRUPT) { + rdev->irq.stat_regs.evergreen.disp_int_cont4 &= ~LB_D5_VLINE_INTERRUPT; + DRM_DEBUG("IH: D5 vline\n"); + } + break; + default: + DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data); + break; + } + break; + case 6: /* D6 vblank/vline */ + switch (src_data) { + case 0: /* D6 vblank */ + if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & LB_D6_VBLANK_INTERRUPT) { + if (rdev->irq.crtc_vblank_int[5]) { + drm_handle_vblank(rdev->ddev, 5); + rdev->pm.vblank_sync = true; + wake_up(&rdev->irq.vblank_queue); + } + if (rdev->irq.pflip[5]) + radeon_crtc_handle_flip(rdev, 5); + rdev->irq.stat_regs.evergreen.disp_int_cont5 &= ~LB_D6_VBLANK_INTERRUPT; + DRM_DEBUG("IH: D6 vblank\n"); + } + break; + case 1: /* D6 vline */ + if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & LB_D6_VLINE_INTERRUPT) { + rdev->irq.stat_regs.evergreen.disp_int_cont5 &= ~LB_D6_VLINE_INTERRUPT; + DRM_DEBUG("IH: D6 vline\n"); + } + break; + default: + DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data); + break; + } + break; + case 42: /* HPD hotplug */ + switch (src_data) { + case 0: + if (rdev->irq.stat_regs.evergreen.disp_int & DC_HPD1_INTERRUPT) { + rdev->irq.stat_regs.evergreen.disp_int &= ~DC_HPD1_INTERRUPT; + queue_hotplug = true; + DRM_DEBUG("IH: HPD1\n"); + } + break; + case 1: + if (rdev->irq.stat_regs.evergreen.disp_int_cont & DC_HPD2_INTERRUPT) { + rdev->irq.stat_regs.evergreen.disp_int_cont &= ~DC_HPD2_INTERRUPT; + queue_hotplug = true; + DRM_DEBUG("IH: HPD2\n"); + } + break; + case 2: + if (rdev->irq.stat_regs.evergreen.disp_int_cont2 & DC_HPD3_INTERRUPT) { + rdev->irq.stat_regs.evergreen.disp_int_cont2 &= ~DC_HPD3_INTERRUPT; + queue_hotplug = true; + DRM_DEBUG("IH: HPD3\n"); + } + break; + case 3: + if (rdev->irq.stat_regs.evergreen.disp_int_cont3 & DC_HPD4_INTERRUPT) { + rdev->irq.stat_regs.evergreen.disp_int_cont3 &= ~DC_HPD4_INTERRUPT; + queue_hotplug = true; + DRM_DEBUG("IH: HPD4\n"); + } + break; + case 4: + if (rdev->irq.stat_regs.evergreen.disp_int_cont4 & DC_HPD5_INTERRUPT) { + rdev->irq.stat_regs.evergreen.disp_int_cont4 &= ~DC_HPD5_INTERRUPT; + queue_hotplug = true; + DRM_DEBUG("IH: HPD5\n"); + } + break; + case 5: + if (rdev->irq.stat_regs.evergreen.disp_int_cont5 & DC_HPD6_INTERRUPT) { + rdev->irq.stat_regs.evergreen.disp_int_cont5 &= ~DC_HPD6_INTERRUPT; + queue_hotplug = true; + DRM_DEBUG("IH: HPD6\n"); + } + break; + default: + DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data); + break; + } + break; + case 176: /* RINGID0 CP_INT */ + radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX); + break; + case 177: /* RINGID1 CP_INT */ + radeon_fence_process(rdev, CAYMAN_RING_TYPE_CP1_INDEX); + break; + case 178: /* RINGID2 CP_INT */ + radeon_fence_process(rdev, CAYMAN_RING_TYPE_CP2_INDEX); + break; + case 181: /* CP EOP event */ + DRM_DEBUG("IH: CP EOP\n"); + switch (ring_id) { + case 0: + radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX); + break; + case 1: + radeon_fence_process(rdev, CAYMAN_RING_TYPE_CP1_INDEX); + break; + case 2: + radeon_fence_process(rdev, CAYMAN_RING_TYPE_CP2_INDEX); + break; + } + break; + case 233: /* GUI IDLE */ + DRM_DEBUG("IH: GUI idle\n"); + rdev->pm.gui_idle = true; + wake_up(&rdev->irq.idle_queue); + break; + default: + DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data); + break; + } + + /* wptr/rptr are in bytes! */ + rptr += 16; + rptr &= rdev->ih.ptr_mask; + } + /* make sure wptr hasn't changed while processing */ + wptr = si_get_ih_wptr(rdev); + if (wptr != rdev->ih.wptr) + goto restart_ih; + if (queue_hotplug) + schedule_work(&rdev->hotplug_work); + rdev->ih.rptr = rptr; + WREG32(IH_RB_RPTR, rdev->ih.rptr); + spin_unlock_irqrestore(&rdev->ih.lock, flags); + return IRQ_HANDLED; +} + +/* + * startup/shutdown callbacks + */ +static int si_startup(struct radeon_device *rdev) +{ + struct radeon_ring *ring; + int r; + + if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw || + !rdev->rlc_fw || !rdev->mc_fw) { + r = si_init_microcode(rdev); + if (r) { + DRM_ERROR("Failed to load firmware!\n"); + return r; + } + } + + r = si_mc_load_microcode(rdev); + if (r) { + DRM_ERROR("Failed to load MC firmware!\n"); + return r; + } + + r = r600_vram_scratch_init(rdev); + if (r) + return r; + + si_mc_program(rdev); + r = si_pcie_gart_enable(rdev); + if (r) + return r; + si_gpu_init(rdev); + +#if 0 + r = evergreen_blit_init(rdev); + if (r) { + r600_blit_fini(rdev); + rdev->asic->copy = NULL; + dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r); + } +#endif + /* allocate rlc buffers */ + r = si_rlc_init(rdev); + if (r) { + DRM_ERROR("Failed to init rlc BOs!\n"); + return r; + } + + /* allocate wb buffer */ + r = radeon_wb_init(rdev); + if (r) + return r; + + r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r); + return r; + } + + r = radeon_fence_driver_start_ring(rdev, CAYMAN_RING_TYPE_CP1_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r); + return r; + } + + r = radeon_fence_driver_start_ring(rdev, CAYMAN_RING_TYPE_CP2_INDEX); + if (r) { + dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r); + return r; + } + + /* Enable IRQ */ + r = si_irq_init(rdev); + if (r) { + DRM_ERROR("radeon: IH init failed (%d).\n", r); + radeon_irq_kms_fini(rdev); + return r; + } + si_irq_set(rdev); + + ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; + r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET, + CP_RB0_RPTR, CP_RB0_WPTR, + 0, 0xfffff, RADEON_CP_PACKET2); + if (r) + return r; + + ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]; + r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP1_RPTR_OFFSET, + CP_RB1_RPTR, CP_RB1_WPTR, + 0, 0xfffff, RADEON_CP_PACKET2); + if (r) + return r; + + ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]; + r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP2_RPTR_OFFSET, + CP_RB2_RPTR, CP_RB2_WPTR, + 0, 0xfffff, RADEON_CP_PACKET2); + if (r) + return r; + + r = si_cp_load_microcode(rdev); + if (r) + return r; + r = si_cp_resume(rdev); + if (r) + return r; + + r = radeon_ib_pool_start(rdev); + if (r) + return r; + + r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]); + if (r) { + DRM_ERROR("radeon: failed testing IB (%d) on CP ring 0\n", r); + rdev->accel_working = false; + return r; + } + + r = radeon_ib_test(rdev, CAYMAN_RING_TYPE_CP1_INDEX, &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]); + if (r) { + DRM_ERROR("radeon: failed testing IB (%d) on CP ring 1\n", r); + rdev->accel_working = false; + return r; + } + + r = radeon_ib_test(rdev, CAYMAN_RING_TYPE_CP2_INDEX, &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]); + if (r) { + DRM_ERROR("radeon: failed testing IB (%d) on CP ring 2\n", r); + rdev->accel_working = false; + return r; + } + + r = radeon_vm_manager_start(rdev); + if (r) + return r; + + return 0; +} + +int si_resume(struct radeon_device *rdev) +{ + int r; + + /* Do not reset GPU before posting, on rv770 hw unlike on r500 hw, + * posting will perform necessary task to bring back GPU into good + * shape. + */ + /* post card */ + atom_asic_init(rdev->mode_info.atom_context); + + rdev->accel_working = true; + r = si_startup(rdev); + if (r) { + DRM_ERROR("si startup failed on resume\n"); + rdev->accel_working = false; + return r; + } + + return r; + +} + +int si_suspend(struct radeon_device *rdev) +{ + /* FIXME: we should wait for ring to be empty */ + radeon_ib_pool_suspend(rdev); + radeon_vm_manager_suspend(rdev); +#if 0 + r600_blit_suspend(rdev); +#endif + si_cp_enable(rdev, false); + rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false; + rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX].ready = false; + rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX].ready = false; + si_irq_suspend(rdev); + radeon_wb_disable(rdev); + si_pcie_gart_disable(rdev); + return 0; +} + +/* Plan is to move initialization in that function and use + * helper function so that radeon_device_init pretty much + * do nothing more than calling asic specific function. This + * should also allow to remove a bunch of callback function + * like vram_info. + */ +int si_init(struct radeon_device *rdev) +{ + struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; + int r; + + /* This don't do much */ + r = radeon_gem_init(rdev); + if (r) + return r; + /* Read BIOS */ + if (!radeon_get_bios(rdev)) { + if (ASIC_IS_AVIVO(rdev)) + return -EINVAL; + } + /* Must be an ATOMBIOS */ + if (!rdev->is_atom_bios) { + dev_err(rdev->dev, "Expecting atombios for cayman GPU\n"); + return -EINVAL; + } + r = radeon_atombios_init(rdev); + if (r) + return r; + + /* Post card if necessary */ + if (!radeon_card_posted(rdev)) { + if (!rdev->bios) { + dev_err(rdev->dev, "Card not posted and no BIOS - ignoring\n"); + return -EINVAL; + } + DRM_INFO("GPU not posted. posting now...\n"); + atom_asic_init(rdev->mode_info.atom_context); + } + /* Initialize scratch registers */ + si_scratch_init(rdev); + /* Initialize surface registers */ + radeon_surface_init(rdev); + /* Initialize clocks */ + radeon_get_clock_info(rdev->ddev); + + /* Fence driver */ + r = radeon_fence_driver_init(rdev); + if (r) + return r; + + /* initialize memory controller */ + r = si_mc_init(rdev); + if (r) + return r; + /* Memory manager */ + r = radeon_bo_init(rdev); + if (r) + return r; + + r = radeon_irq_kms_init(rdev); + if (r) + return r; + + ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]; + ring->ring_obj = NULL; + r600_ring_init(rdev, ring, 1024 * 1024); + + ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX]; + ring->ring_obj = NULL; + r600_ring_init(rdev, ring, 1024 * 1024); + + ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX]; + ring->ring_obj = NULL; + r600_ring_init(rdev, ring, 1024 * 1024); + + rdev->ih.ring_obj = NULL; + r600_ih_ring_init(rdev, 64 * 1024); + + r = r600_pcie_gart_init(rdev); + if (r) + return r; + + r = radeon_ib_pool_init(rdev); + rdev->accel_working = true; + if (r) { + dev_err(rdev->dev, "IB initialization failed (%d).\n", r); + rdev->accel_working = false; + } + r = radeon_vm_manager_init(rdev); + if (r) { + dev_err(rdev->dev, "vm manager initialization failed (%d).\n", r); + } + + r = si_startup(rdev); + if (r) { + dev_err(rdev->dev, "disabling GPU acceleration\n"); + si_cp_fini(rdev); + si_irq_fini(rdev); + si_rlc_fini(rdev); + radeon_wb_fini(rdev); + r100_ib_fini(rdev); + radeon_vm_manager_fini(rdev); + radeon_irq_kms_fini(rdev); + si_pcie_gart_fini(rdev); + rdev->accel_working = false; + } + + /* Don't start up if the MC ucode is missing. + * The default clocks and voltages before the MC ucode + * is loaded are not suffient for advanced operations. + */ + if (!rdev->mc_fw) { + DRM_ERROR("radeon: MC ucode required for NI+.\n"); + return -EINVAL; + } + + return 0; +} + +void si_fini(struct radeon_device *rdev) +{ +#if 0 + r600_blit_fini(rdev); +#endif + si_cp_fini(rdev); + si_irq_fini(rdev); + si_rlc_fini(rdev); + radeon_wb_fini(rdev); + radeon_vm_manager_fini(rdev); + r100_ib_fini(rdev); + radeon_irq_kms_fini(rdev); + si_pcie_gart_fini(rdev); + r600_vram_scratch_fini(rdev); + radeon_gem_fini(rdev); + radeon_semaphore_driver_fini(rdev); + radeon_fence_driver_fini(rdev); + radeon_bo_fini(rdev); + radeon_atombios_fini(rdev); + kfree(rdev->bios); + rdev->bios = NULL; +} + diff --git a/drivers/gpu/drm/radeon/si_blit_shaders.c b/drivers/gpu/drm/radeon/si_blit_shaders.c new file mode 100644 index 0000000..ec415e7 --- /dev/null +++ b/drivers/gpu/drm/radeon/si_blit_shaders.c @@ -0,0 +1,253 @@ +/* + * Copyright 2011 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * Alex Deucher + */ + +#include +#include +#include + +const u32 si_default_state[] = +{ + 0xc0066900, + 0x00000000, + 0x00000060, /* DB_RENDER_CONTROL */ + 0x00000000, /* DB_COUNT_CONTROL */ + 0x00000000, /* DB_DEPTH_VIEW */ + 0x0000002a, /* DB_RENDER_OVERRIDE */ + 0x00000000, /* DB_RENDER_OVERRIDE2 */ + 0x00000000, /* DB_HTILE_DATA_BASE */ + + 0xc0046900, + 0x00000008, + 0x00000000, /* DB_DEPTH_BOUNDS_MIN */ + 0x00000000, /* DB_DEPTH_BOUNDS_MAX */ + 0x00000000, /* DB_STENCIL_CLEAR */ + 0x00000000, /* DB_DEPTH_CLEAR */ + + 0xc0036900, + 0x0000000f, + 0x00000000, /* DB_DEPTH_INFO */ + 0x00000000, /* DB_Z_INFO */ + 0x00000000, /* DB_STENCIL_INFO */ + + 0xc0016900, + 0x00000080, + 0x00000000, /* PA_SC_WINDOW_OFFSET */ + + 0xc00d6900, + 0x00000083, + 0x0000ffff, /* PA_SC_CLIPRECT_RULE */ + 0x00000000, /* PA_SC_CLIPRECT_0_TL */ + 0x20002000, /* PA_SC_CLIPRECT_0_BR */ + 0x00000000, + 0x20002000, + 0x00000000, + 0x20002000, + 0x00000000, + 0x20002000, + 0xaaaaaaaa, /* PA_SC_EDGERULE */ + 0x00000000, /* PA_SU_HARDWARE_SCREEN_OFFSET */ + 0x0000000f, /* CB_TARGET_MASK */ + 0x0000000f, /* CB_SHADER_MASK */ + + 0xc0226900, + 0x00000094, + 0x80000000, /* PA_SC_VPORT_SCISSOR_0_TL */ + 0x20002000, /* PA_SC_VPORT_SCISSOR_0_BR */ + 0x80000000, + 0x20002000, + 0x80000000, + 0x20002000, + 0x80000000, + 0x20002000, + 0x80000000, + 0x20002000, + 0x80000000, + 0x20002000, + 0x80000000, + 0x20002000, + 0x80000000, + 0x20002000, + 0x80000000, + 0x20002000, + 0x80000000, + 0x20002000, + 0x80000000, + 0x20002000, + 0x80000000, + 0x20002000, + 0x80000000, + 0x20002000, + 0x80000000, + 0x20002000, + 0x80000000, + 0x20002000, + 0x80000000, + 0x20002000, + 0x00000000, /* PA_SC_VPORT_ZMIN_0 */ + 0x3f800000, /* PA_SC_VPORT_ZMAX_0 */ + + 0xc0026900, + 0x000000d9, + 0x00000000, /* CP_RINGID */ + 0x00000000, /* CP_VMID */ + + 0xc0046900, + 0x00000100, + 0xffffffff, /* VGT_MAX_VTX_INDX */ + 0x00000000, /* VGT_MIN_VTX_INDX */ + 0x00000000, /* VGT_INDX_OFFSET */ + 0x00000000, /* VGT_MULTI_PRIM_IB_RESET_INDX */ + + 0xc0046900, + 0x00000105, + 0x00000000, /* CB_BLEND_RED */ + 0x00000000, /* CB_BLEND_GREEN */ + 0x00000000, /* CB_BLEND_BLUE */ + 0x00000000, /* CB_BLEND_ALPHA */ + + 0xc0016900, + 0x000001e0, + 0x00000000, /* CB_BLEND0_CONTROL */ + + 0xc00e6900, + 0x00000200, + 0x00000000, /* DB_DEPTH_CONTROL */ + 0x00000000, /* DB_EQAA */ + 0x00cc0010, /* CB_COLOR_CONTROL */ + 0x00000210, /* DB_SHADER_CONTROL */ + 0x00010000, /* PA_CL_CLIP_CNTL */ + 0x00000004, /* PA_SU_SC_MODE_CNTL */ + 0x00000100, /* PA_CL_VTE_CNTL */ + 0x00000000, /* PA_CL_VS_OUT_CNTL */ + 0x00000000, /* PA_CL_NANINF_CNTL */ + 0x00000000, /* PA_SU_LINE_STIPPLE_CNTL */ + 0x00000000, /* PA_SU_LINE_STIPPLE_SCALE */ + 0x00000000, /* PA_SU_PRIM_FILTER_CNTL */ + 0x00000000, /* */ + 0x00000000, /* */ + + 0xc0116900, + 0x00000280, + 0x00000000, /* PA_SU_POINT_SIZE */ + 0x00000000, /* PA_SU_POINT_MINMAX */ + 0x00000008, /* PA_SU_LINE_CNTL */ + 0x00000000, /* PA_SC_LINE_STIPPLE */ + 0x00000000, /* VGT_OUTPUT_PATH_CNTL */ + 0x00000000, /* VGT_HOS_CNTL */ + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, /* VGT_GS_MODE */ + + 0xc0026900, + 0x00000292, + 0x00000000, /* PA_SC_MODE_CNTL_0 */ + 0x00000000, /* PA_SC_MODE_CNTL_1 */ + + 0xc0016900, + 0x000002a1, + 0x00000000, /* VGT_PRIMITIVEID_EN */ + + 0xc0016900, + 0x000002a5, + 0x00000000, /* VGT_MULTI_PRIM_IB_RESET_EN */ + + 0xc0026900, + 0x000002a8, + 0x00000000, /* VGT_INSTANCE_STEP_RATE_0 */ + 0x00000000, + + 0xc0026900, + 0x000002ad, + 0x00000000, /* VGT_REUSE_OFF */ + 0x00000000, + + 0xc0016900, + 0x000002d5, + 0x00000000, /* VGT_SHADER_STAGES_EN */ + + 0xc0016900, + 0x000002dc, + 0x0000aa00, /* DB_ALPHA_TO_MASK */ + + 0xc0066900, + 0x000002de, + 0x00000000, /* PA_SU_POLY_OFFSET_DB_FMT_CNTL */ + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + + 0xc0026900, + 0x000002e5, + 0x00000000, /* VGT_STRMOUT_CONFIG */ + 0x00000000, + + 0xc01b6900, + 0x000002f5, + 0x76543210, /* PA_SC_CENTROID_PRIORITY_0 */ + 0xfedcba98, /* PA_SC_CENTROID_PRIORITY_1 */ + 0x00000000, /* PA_SC_LINE_CNTL */ + 0x00000000, /* PA_SC_AA_CONFIG */ + 0x00000005, /* PA_SU_VTX_CNTL */ + 0x3f800000, /* PA_CL_GB_VERT_CLIP_ADJ */ + 0x3f800000, /* PA_CL_GB_VERT_DISC_ADJ */ + 0x3f800000, /* PA_CL_GB_HORZ_CLIP_ADJ */ + 0x3f800000, /* PA_CL_GB_HORZ_DISC_ADJ */ + 0x00000000, /* PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0 */ + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0x00000000, + 0xffffffff, /* PA_SC_AA_MASK_X0Y0_X1Y0 */ + 0xffffffff, + + 0xc0026900, + 0x00000316, + 0x0000000e, /* VGT_VERTEX_REUSE_BLOCK_CNTL */ + 0x00000010, /* */ +}; + +const u32 si_default_size = ARRAY_SIZE(si_default_state); diff --git a/drivers/gpu/drm/radeon/si_blit_shaders.h b/drivers/gpu/drm/radeon/si_blit_shaders.h new file mode 100644 index 0000000..c739e51 --- /dev/null +++ b/drivers/gpu/drm/radeon/si_blit_shaders.h @@ -0,0 +1,32 @@ +/* + * Copyright 2011 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + */ + +#ifndef SI_BLIT_SHADERS_H +#define SI_BLIT_SHADERS_H + +extern const u32 si_default_state[]; + +extern const u32 si_default_size; + +#endif diff --git a/drivers/gpu/drm/radeon/si_reg.h b/drivers/gpu/drm/radeon/si_reg.h new file mode 100644 index 0000000..eda938a --- /dev/null +++ b/drivers/gpu/drm/radeon/si_reg.h @@ -0,0 +1,33 @@ +/* + * Copyright 2010 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Alex Deucher + */ +#ifndef __SI_REG_H__ +#define __SI_REG_H__ + +/* SI */ +#define SI_DC_GPIO_HPD_MASK 0x65b0 +#define SI_DC_GPIO_HPD_A 0x65b4 +#define SI_DC_GPIO_HPD_EN 0x65b8 +#define SI_DC_GPIO_HPD_Y 0x65bc + +#endif diff --git a/drivers/gpu/drm/radeon/sid.h b/drivers/gpu/drm/radeon/sid.h new file mode 100644 index 0000000..2c2bc63 --- /dev/null +++ b/drivers/gpu/drm/radeon/sid.h @@ -0,0 +1,887 @@ +/* + * Copyright 2011 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Alex Deucher + */ +#ifndef SI_H +#define SI_H + +#define CG_MULT_THERMAL_STATUS 0x714 +#define ASIC_MAX_TEMP(x) ((x) << 0) +#define ASIC_MAX_TEMP_MASK 0x000001ff +#define ASIC_MAX_TEMP_SHIFT 0 +#define CTF_TEMP(x) ((x) << 9) +#define CTF_TEMP_MASK 0x0003fe00 +#define CTF_TEMP_SHIFT 9 + +#define SI_MAX_SH_GPRS 256 +#define SI_MAX_TEMP_GPRS 16 +#define SI_MAX_SH_THREADS 256 +#define SI_MAX_SH_STACK_ENTRIES 4096 +#define SI_MAX_FRC_EOV_CNT 16384 +#define SI_MAX_BACKENDS 8 +#define SI_MAX_BACKENDS_MASK 0xFF +#define SI_MAX_BACKENDS_PER_SE_MASK 0x0F +#define SI_MAX_SIMDS 12 +#define SI_MAX_SIMDS_MASK 0x0FFF +#define SI_MAX_SIMDS_PER_SE_MASK 0x00FF +#define SI_MAX_PIPES 8 +#define SI_MAX_PIPES_MASK 0xFF +#define SI_MAX_PIPES_PER_SIMD_MASK 0x3F +#define SI_MAX_LDS_NUM 0xFFFF +#define SI_MAX_TCC 16 +#define SI_MAX_TCC_MASK 0xFFFF + +#define VGA_HDP_CONTROL 0x328 +#define VGA_MEMORY_DISABLE (1 << 4) + +#define DMIF_ADDR_CONFIG 0xBD4 + +#define SRBM_STATUS 0xE50 + +#define CC_SYS_RB_BACKEND_DISABLE 0xe80 +#define GC_USER_SYS_RB_BACKEND_DISABLE 0xe84 + +#define VM_L2_CNTL 0x1400 +#define ENABLE_L2_CACHE (1 << 0) +#define ENABLE_L2_FRAGMENT_PROCESSING (1 << 1) +#define L2_CACHE_PTE_ENDIAN_SWAP_MODE(x) ((x) << 2) +#define L2_CACHE_PDE_ENDIAN_SWAP_MODE(x) ((x) << 4) +#define ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE (1 << 9) +#define ENABLE_L2_PDE0_CACHE_LRU_UPDATE_BY_WRITE (1 << 10) +#define EFFECTIVE_L2_QUEUE_SIZE(x) (((x) & 7) << 15) +#define CONTEXT1_IDENTITY_ACCESS_MODE(x) (((x) & 3) << 19) +#define VM_L2_CNTL2 0x1404 +#define INVALIDATE_ALL_L1_TLBS (1 << 0) +#define INVALIDATE_L2_CACHE (1 << 1) +#define INVALIDATE_CACHE_MODE(x) ((x) << 26) +#define INVALIDATE_PTE_AND_PDE_CACHES 0 +#define INVALIDATE_ONLY_PTE_CACHES 1 +#define INVALIDATE_ONLY_PDE_CACHES 2 +#define VM_L2_CNTL3 0x1408 +#define BANK_SELECT(x) ((x) << 0) +#define L2_CACHE_UPDATE_MODE(x) ((x) << 6) +#define L2_CACHE_BIGK_FRAGMENT_SIZE(x) ((x) << 15) +#define L2_CACHE_BIGK_ASSOCIATIVITY (1 << 20) +#define VM_L2_STATUS 0x140C +#define L2_BUSY (1 << 0) +#define VM_CONTEXT0_CNTL 0x1410 +#define ENABLE_CONTEXT (1 << 0) +#define PAGE_TABLE_DEPTH(x) (((x) & 3) << 1) +#define RANGE_PROTECTION_FAULT_ENABLE_DEFAULT (1 << 4) +#define VM_CONTEXT1_CNTL 0x1414 +#define VM_CONTEXT0_CNTL2 0x1430 +#define VM_CONTEXT1_CNTL2 0x1434 +#define VM_CONTEXT8_PAGE_TABLE_BASE_ADDR 0x1438 +#define VM_CONTEXT9_PAGE_TABLE_BASE_ADDR 0x143c +#define VM_CONTEXT10_PAGE_TABLE_BASE_ADDR 0x1440 +#define VM_CONTEXT11_PAGE_TABLE_BASE_ADDR 0x1444 +#define VM_CONTEXT12_PAGE_TABLE_BASE_ADDR 0x1448 +#define VM_CONTEXT13_PAGE_TABLE_BASE_ADDR 0x144c +#define VM_CONTEXT14_PAGE_TABLE_BASE_ADDR 0x1450 +#define VM_CONTEXT15_PAGE_TABLE_BASE_ADDR 0x1454 + +#define VM_INVALIDATE_REQUEST 0x1478 +#define VM_INVALIDATE_RESPONSE 0x147c + +#define VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR 0x1518 +#define VM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR 0x151c + +#define VM_CONTEXT0_PAGE_TABLE_BASE_ADDR 0x153c +#define VM_CONTEXT1_PAGE_TABLE_BASE_ADDR 0x1540 +#define VM_CONTEXT2_PAGE_TABLE_BASE_ADDR 0x1544 +#define VM_CONTEXT3_PAGE_TABLE_BASE_ADDR 0x1548 +#define VM_CONTEXT4_PAGE_TABLE_BASE_ADDR 0x154c +#define VM_CONTEXT5_PAGE_TABLE_BASE_ADDR 0x1550 +#define VM_CONTEXT6_PAGE_TABLE_BASE_ADDR 0x1554 +#define VM_CONTEXT7_PAGE_TABLE_BASE_ADDR 0x1558 +#define VM_CONTEXT0_PAGE_TABLE_START_ADDR 0x155c +#define VM_CONTEXT1_PAGE_TABLE_START_ADDR 0x1560 + +#define VM_CONTEXT0_PAGE_TABLE_END_ADDR 0x157C +#define VM_CONTEXT1_PAGE_TABLE_END_ADDR 0x1580 + +#define MC_SHARED_CHMAP 0x2004 +#define NOOFCHAN_SHIFT 12 +#define NOOFCHAN_MASK 0x0000f000 +#define MC_SHARED_CHREMAP 0x2008 + +#define MC_VM_FB_LOCATION 0x2024 +#define MC_VM_AGP_TOP 0x2028 +#define MC_VM_AGP_BOT 0x202C +#define MC_VM_AGP_BASE 0x2030 +#define MC_VM_SYSTEM_APERTURE_LOW_ADDR 0x2034 +#define MC_VM_SYSTEM_APERTURE_HIGH_ADDR 0x2038 +#define MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR 0x203C + +#define MC_VM_MX_L1_TLB_CNTL 0x2064 +#define ENABLE_L1_TLB (1 << 0) +#define ENABLE_L1_FRAGMENT_PROCESSING (1 << 1) +#define SYSTEM_ACCESS_MODE_PA_ONLY (0 << 3) +#define SYSTEM_ACCESS_MODE_USE_SYS_MAP (1 << 3) +#define SYSTEM_ACCESS_MODE_IN_SYS (2 << 3) +#define SYSTEM_ACCESS_MODE_NOT_IN_SYS (3 << 3) +#define SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU (0 << 5) +#define ENABLE_ADVANCED_DRIVER_MODEL (1 << 6) + +#define MC_SHARED_BLACKOUT_CNTL 0x20ac + +#define MC_ARB_RAMCFG 0x2760 +#define NOOFBANK_SHIFT 0 +#define NOOFBANK_MASK 0x00000003 +#define NOOFRANK_SHIFT 2 +#define NOOFRANK_MASK 0x00000004 +#define NOOFROWS_SHIFT 3 +#define NOOFROWS_MASK 0x00000038 +#define NOOFCOLS_SHIFT 6 +#define NOOFCOLS_MASK 0x000000C0 +#define CHANSIZE_SHIFT 8 +#define CHANSIZE_MASK 0x00000100 +#define CHANSIZE_OVERRIDE (1 << 11) +#define NOOFGROUPS_SHIFT 12 +#define NOOFGROUPS_MASK 0x00001000 + +#define MC_SEQ_TRAIN_WAKEUP_CNTL 0x2808 +#define TRAIN_DONE_D0 (1 << 30) +#define TRAIN_DONE_D1 (1 << 31) + +#define MC_SEQ_SUP_CNTL 0x28c8 +#define RUN_MASK (1 << 0) +#define MC_SEQ_SUP_PGM 0x28cc + +#define MC_IO_PAD_CNTL_D0 0x29d0 +#define MEM_FALL_OUT_CMD (1 << 8) + +#define MC_SEQ_IO_DEBUG_INDEX 0x2a44 +#define MC_SEQ_IO_DEBUG_DATA 0x2a48 + +#define HDP_HOST_PATH_CNTL 0x2C00 +#define HDP_NONSURFACE_BASE 0x2C04 +#define HDP_NONSURFACE_INFO 0x2C08 +#define HDP_NONSURFACE_SIZE 0x2C0C + +#define HDP_ADDR_CONFIG 0x2F48 +#define HDP_MISC_CNTL 0x2F4C +#define HDP_FLUSH_INVALIDATE_CACHE (1 << 0) + +#define IH_RB_CNTL 0x3e00 +# define IH_RB_ENABLE (1 << 0) +# define IH_IB_SIZE(x) ((x) << 1) /* log2 */ +# define IH_RB_FULL_DRAIN_ENABLE (1 << 6) +# define IH_WPTR_WRITEBACK_ENABLE (1 << 8) +# define IH_WPTR_WRITEBACK_TIMER(x) ((x) << 9) /* log2 */ +# define IH_WPTR_OVERFLOW_ENABLE (1 << 16) +# define IH_WPTR_OVERFLOW_CLEAR (1 << 31) +#define IH_RB_BASE 0x3e04 +#define IH_RB_RPTR 0x3e08 +#define IH_RB_WPTR 0x3e0c +# define RB_OVERFLOW (1 << 0) +# define WPTR_OFFSET_MASK 0x3fffc +#define IH_RB_WPTR_ADDR_HI 0x3e10 +#define IH_RB_WPTR_ADDR_LO 0x3e14 +#define IH_CNTL 0x3e18 +# define ENABLE_INTR (1 << 0) +# define IH_MC_SWAP(x) ((x) << 1) +# define IH_MC_SWAP_NONE 0 +# define IH_MC_SWAP_16BIT 1 +# define IH_MC_SWAP_32BIT 2 +# define IH_MC_SWAP_64BIT 3 +# define RPTR_REARM (1 << 4) +# define MC_WRREQ_CREDIT(x) ((x) << 15) +# define MC_WR_CLEAN_CNT(x) ((x) << 20) +# define MC_VMID(x) ((x) << 25) + +#define CONFIG_MEMSIZE 0x5428 + +#define INTERRUPT_CNTL 0x5468 +# define IH_DUMMY_RD_OVERRIDE (1 << 0) +# define IH_DUMMY_RD_EN (1 << 1) +# define IH_REQ_NONSNOOP_EN (1 << 3) +# define GEN_IH_INT_EN (1 << 8) +#define INTERRUPT_CNTL2 0x546c + +#define HDP_MEM_COHERENCY_FLUSH_CNTL 0x5480 + +#define BIF_FB_EN 0x5490 +#define FB_READ_EN (1 << 0) +#define FB_WRITE_EN (1 << 1) + +#define HDP_REG_COHERENCY_FLUSH_CNTL 0x54A0 + +#define DC_LB_MEMORY_SPLIT 0x6b0c +#define DC_LB_MEMORY_CONFIG(x) ((x) << 20) + +#define PRIORITY_A_CNT 0x6b18 +#define PRIORITY_MARK_MASK 0x7fff +#define PRIORITY_OFF (1 << 16) +#define PRIORITY_ALWAYS_ON (1 << 20) +#define PRIORITY_B_CNT 0x6b1c + +#define DPG_PIPE_ARBITRATION_CONTROL3 0x6cc8 +# define LATENCY_WATERMARK_MASK(x) ((x) << 16) +#define DPG_PIPE_LATENCY_CONTROL 0x6ccc +# define LATENCY_LOW_WATERMARK(x) ((x) << 0) +# define LATENCY_HIGH_WATERMARK(x) ((x) << 16) + +/* 0x6bb8, 0x77b8, 0x103b8, 0x10fb8, 0x11bb8, 0x127b8 */ +#define VLINE_STATUS 0x6bb8 +# define VLINE_OCCURRED (1 << 0) +# define VLINE_ACK (1 << 4) +# define VLINE_STAT (1 << 12) +# define VLINE_INTERRUPT (1 << 16) +# define VLINE_INTERRUPT_TYPE (1 << 17) +/* 0x6bbc, 0x77bc, 0x103bc, 0x10fbc, 0x11bbc, 0x127bc */ +#define VBLANK_STATUS 0x6bbc +# define VBLANK_OCCURRED (1 << 0) +# define VBLANK_ACK (1 << 4) +# define VBLANK_STAT (1 << 12) +# define VBLANK_INTERRUPT (1 << 16) +# define VBLANK_INTERRUPT_TYPE (1 << 17) + +/* 0x6b40, 0x7740, 0x10340, 0x10f40, 0x11b40, 0x12740 */ +#define INT_MASK 0x6b40 +# define VBLANK_INT_MASK (1 << 0) +# define VLINE_INT_MASK (1 << 4) + +#define DISP_INTERRUPT_STATUS 0x60f4 +# define LB_D1_VLINE_INTERRUPT (1 << 2) +# define LB_D1_VBLANK_INTERRUPT (1 << 3) +# define DC_HPD1_INTERRUPT (1 << 17) +# define DC_HPD1_RX_INTERRUPT (1 << 18) +# define DACA_AUTODETECT_INTERRUPT (1 << 22) +# define DACB_AUTODETECT_INTERRUPT (1 << 23) +# define DC_I2C_SW_DONE_INTERRUPT (1 << 24) +# define DC_I2C_HW_DONE_INTERRUPT (1 << 25) +#define DISP_INTERRUPT_STATUS_CONTINUE 0x60f8 +# define LB_D2_VLINE_INTERRUPT (1 << 2) +# define LB_D2_VBLANK_INTERRUPT (1 << 3) +# define DC_HPD2_INTERRUPT (1 << 17) +# define DC_HPD2_RX_INTERRUPT (1 << 18) +# define DISP_TIMER_INTERRUPT (1 << 24) +#define DISP_INTERRUPT_STATUS_CONTINUE2 0x60fc +# define LB_D3_VLINE_INTERRUPT (1 << 2) +# define LB_D3_VBLANK_INTERRUPT (1 << 3) +# define DC_HPD3_INTERRUPT (1 << 17) +# define DC_HPD3_RX_INTERRUPT (1 << 18) +#define DISP_INTERRUPT_STATUS_CONTINUE3 0x6100 +# define LB_D4_VLINE_INTERRUPT (1 << 2) +# define LB_D4_VBLANK_INTERRUPT (1 << 3) +# define DC_HPD4_INTERRUPT (1 << 17) +# define DC_HPD4_RX_INTERRUPT (1 << 18) +#define DISP_INTERRUPT_STATUS_CONTINUE4 0x614c +# define LB_D5_VLINE_INTERRUPT (1 << 2) +# define LB_D5_VBLANK_INTERRUPT (1 << 3) +# define DC_HPD5_INTERRUPT (1 << 17) +# define DC_HPD5_RX_INTERRUPT (1 << 18) +#define DISP_INTERRUPT_STATUS_CONTINUE5 0x6150 +# define LB_D6_VLINE_INTERRUPT (1 << 2) +# define LB_D6_VBLANK_INTERRUPT (1 << 3) +# define DC_HPD6_INTERRUPT (1 << 17) +# define DC_HPD6_RX_INTERRUPT (1 << 18) + +/* 0x6858, 0x7458, 0x10058, 0x10c58, 0x11858, 0x12458 */ +#define GRPH_INT_STATUS 0x6858 +# define GRPH_PFLIP_INT_OCCURRED (1 << 0) +# define GRPH_PFLIP_INT_CLEAR (1 << 8) +/* 0x685c, 0x745c, 0x1005c, 0x10c5c, 0x1185c, 0x1245c */ +#define GRPH_INT_CONTROL 0x685c +# define GRPH_PFLIP_INT_MASK (1 << 0) +# define GRPH_PFLIP_INT_TYPE (1 << 8) + +#define DACA_AUTODETECT_INT_CONTROL 0x66c8 + +#define DC_HPD1_INT_STATUS 0x601c +#define DC_HPD2_INT_STATUS 0x6028 +#define DC_HPD3_INT_STATUS 0x6034 +#define DC_HPD4_INT_STATUS 0x6040 +#define DC_HPD5_INT_STATUS 0x604c +#define DC_HPD6_INT_STATUS 0x6058 +# define DC_HPDx_INT_STATUS (1 << 0) +# define DC_HPDx_SENSE (1 << 1) +# define DC_HPDx_RX_INT_STATUS (1 << 8) + +#define DC_HPD1_INT_CONTROL 0x6020 +#define DC_HPD2_INT_CONTROL 0x602c +#define DC_HPD3_INT_CONTROL 0x6038 +#define DC_HPD4_INT_CONTROL 0x6044 +#define DC_HPD5_INT_CONTROL 0x6050 +#define DC_HPD6_INT_CONTROL 0x605c +# define DC_HPDx_INT_ACK (1 << 0) +# define DC_HPDx_INT_POLARITY (1 << 8) +# define DC_HPDx_INT_EN (1 << 16) +# define DC_HPDx_RX_INT_ACK (1 << 20) +# define DC_HPDx_RX_INT_EN (1 << 24) + +#define DC_HPD1_CONTROL 0x6024 +#define DC_HPD2_CONTROL 0x6030 +#define DC_HPD3_CONTROL 0x603c +#define DC_HPD4_CONTROL 0x6048 +#define DC_HPD5_CONTROL 0x6054 +#define DC_HPD6_CONTROL 0x6060 +# define DC_HPDx_CONNECTION_TIMER(x) ((x) << 0) +# define DC_HPDx_RX_INT_TIMER(x) ((x) << 16) +# define DC_HPDx_EN (1 << 28) + +/* 0x6e98, 0x7a98, 0x10698, 0x11298, 0x11e98, 0x12a98 */ +#define CRTC_STATUS_FRAME_COUNT 0x6e98 + +#define GRBM_CNTL 0x8000 +#define GRBM_READ_TIMEOUT(x) ((x) << 0) + +#define GRBM_STATUS2 0x8008 +#define RLC_RQ_PENDING (1 << 0) +#define RLC_BUSY (1 << 8) +#define TC_BUSY (1 << 9) + +#define GRBM_STATUS 0x8010 +#define CMDFIFO_AVAIL_MASK 0x0000000F +#define RING2_RQ_PENDING (1 << 4) +#define SRBM_RQ_PENDING (1 << 5) +#define RING1_RQ_PENDING (1 << 6) +#define CF_RQ_PENDING (1 << 7) +#define PF_RQ_PENDING (1 << 8) +#define GDS_DMA_RQ_PENDING (1 << 9) +#define GRBM_EE_BUSY (1 << 10) +#define DB_CLEAN (1 << 12) +#define CB_CLEAN (1 << 13) +#define TA_BUSY (1 << 14) +#define GDS_BUSY (1 << 15) +#define VGT_BUSY (1 << 17) +#define IA_BUSY_NO_DMA (1 << 18) +#define IA_BUSY (1 << 19) +#define SX_BUSY (1 << 20) +#define SPI_BUSY (1 << 22) +#define BCI_BUSY (1 << 23) +#define SC_BUSY (1 << 24) +#define PA_BUSY (1 << 25) +#define DB_BUSY (1 << 26) +#define CP_COHERENCY_BUSY (1 << 28) +#define CP_BUSY (1 << 29) +#define CB_BUSY (1 << 30) +#define GUI_ACTIVE (1 << 31) +#define GRBM_STATUS_SE0 0x8014 +#define GRBM_STATUS_SE1 0x8018 +#define SE_DB_CLEAN (1 << 1) +#define SE_CB_CLEAN (1 << 2) +#define SE_BCI_BUSY (1 << 22) +#define SE_VGT_BUSY (1 << 23) +#define SE_PA_BUSY (1 << 24) +#define SE_TA_BUSY (1 << 25) +#define SE_SX_BUSY (1 << 26) +#define SE_SPI_BUSY (1 << 27) +#define SE_SC_BUSY (1 << 29) +#define SE_DB_BUSY (1 << 30) +#define SE_CB_BUSY (1 << 31) + +#define GRBM_SOFT_RESET 0x8020 +#define SOFT_RESET_CP (1 << 0) +#define SOFT_RESET_CB (1 << 1) +#define SOFT_RESET_RLC (1 << 2) +#define SOFT_RESET_DB (1 << 3) +#define SOFT_RESET_GDS (1 << 4) +#define SOFT_RESET_PA (1 << 5) +#define SOFT_RESET_SC (1 << 6) +#define SOFT_RESET_BCI (1 << 7) +#define SOFT_RESET_SPI (1 << 8) +#define SOFT_RESET_SX (1 << 10) +#define SOFT_RESET_TC (1 << 11) +#define SOFT_RESET_TA (1 << 12) +#define SOFT_RESET_VGT (1 << 14) +#define SOFT_RESET_IA (1 << 15) + +#define GRBM_GFX_INDEX 0x802C + +#define GRBM_INT_CNTL 0x8060 +# define RDERR_INT_ENABLE (1 << 0) +# define GUI_IDLE_INT_ENABLE (1 << 19) + +#define CP_STRMOUT_CNTL 0x84FC +#define SCRATCH_REG0 0x8500 +#define SCRATCH_REG1 0x8504 +#define SCRATCH_REG2 0x8508 +#define SCRATCH_REG3 0x850C +#define SCRATCH_REG4 0x8510 +#define SCRATCH_REG5 0x8514 +#define SCRATCH_REG6 0x8518 +#define SCRATCH_REG7 0x851C + +#define SCRATCH_UMSK 0x8540 +#define SCRATCH_ADDR 0x8544 + +#define CP_SEM_WAIT_TIMER 0x85BC + +#define CP_SEM_INCOMPLETE_TIMER_CNTL 0x85C8 + +#define CP_ME_CNTL 0x86D8 +#define CP_CE_HALT (1 << 24) +#define CP_PFP_HALT (1 << 26) +#define CP_ME_HALT (1 << 28) + +#define CP_COHER_CNTL2 0x85E8 + +#define CP_RB2_RPTR 0x86f8 +#define CP_RB1_RPTR 0x86fc +#define CP_RB0_RPTR 0x8700 +#define CP_RB_WPTR_DELAY 0x8704 + +#define CP_QUEUE_THRESHOLDS 0x8760 +#define ROQ_IB1_START(x) ((x) << 0) +#define ROQ_IB2_START(x) ((x) << 8) +#define CP_MEQ_THRESHOLDS 0x8764 +#define MEQ1_START(x) ((x) << 0) +#define MEQ2_START(x) ((x) << 8) + +#define CP_PERFMON_CNTL 0x87FC + +#define VGT_VTX_VECT_EJECT_REG 0x88B0 + +#define VGT_CACHE_INVALIDATION 0x88C4 +#define CACHE_INVALIDATION(x) ((x) << 0) +#define VC_ONLY 0 +#define TC_ONLY 1 +#define VC_AND_TC 2 +#define AUTO_INVLD_EN(x) ((x) << 6) +#define NO_AUTO 0 +#define ES_AUTO 1 +#define GS_AUTO 2 +#define ES_AND_GS_AUTO 3 +#define VGT_ESGS_RING_SIZE 0x88C8 +#define VGT_GSVS_RING_SIZE 0x88CC + +#define VGT_GS_VERTEX_REUSE 0x88D4 + +#define VGT_PRIMITIVE_TYPE 0x8958 +#define VGT_INDEX_TYPE 0x895C + +#define VGT_NUM_INDICES 0x8970 +#define VGT_NUM_INSTANCES 0x8974 + +#define VGT_TF_RING_SIZE 0x8988 + +#define VGT_HS_OFFCHIP_PARAM 0x89B0 + +#define VGT_TF_MEMORY_BASE 0x89B8 + +#define CC_GC_SHADER_ARRAY_CONFIG 0x89bc +#define GC_USER_SHADER_ARRAY_CONFIG 0x89c0 + +#define PA_CL_ENHANCE 0x8A14 +#define CLIP_VTX_REORDER_ENA (1 << 0) +#define NUM_CLIP_SEQ(x) ((x) << 1) + +#define PA_SU_LINE_STIPPLE_VALUE 0x8A60 + +#define PA_SC_LINE_STIPPLE_STATE 0x8B10 + +#define PA_SC_FORCE_EOV_MAX_CNTS 0x8B24 +#define FORCE_EOV_MAX_CLK_CNT(x) ((x) << 0) +#define FORCE_EOV_MAX_REZ_CNT(x) ((x) << 16) + +#define PA_SC_FIFO_SIZE 0x8BCC +#define SC_FRONTEND_PRIM_FIFO_SIZE(x) ((x) << 0) +#define SC_BACKEND_PRIM_FIFO_SIZE(x) ((x) << 6) +#define SC_HIZ_TILE_FIFO_SIZE(x) ((x) << 15) +#define SC_EARLYZ_TILE_FIFO_SIZE(x) ((x) << 23) + +#define PA_SC_ENHANCE 0x8BF0 + +#define SQ_CONFIG 0x8C00 + +#define SQC_CACHES 0x8C08 + +#define SX_DEBUG_1 0x9060 + +#define SPI_STATIC_THREAD_MGMT_1 0x90E0 +#define SPI_STATIC_THREAD_MGMT_2 0x90E4 +#define SPI_STATIC_THREAD_MGMT_3 0x90E8 +#define SPI_PS_MAX_WAVE_ID 0x90EC + +#define SPI_CONFIG_CNTL 0x9100 + +#define SPI_CONFIG_CNTL_1 0x913C +#define VTX_DONE_DELAY(x) ((x) << 0) +#define INTERP_ONE_PRIM_PER_ROW (1 << 4) + +#define CGTS_TCC_DISABLE 0x9148 +#define CGTS_USER_TCC_DISABLE 0x914C +#define TCC_DISABLE_MASK 0xFFFF0000 +#define TCC_DISABLE_SHIFT 16 + +#define TA_CNTL_AUX 0x9508 + +#define CC_RB_BACKEND_DISABLE 0x98F4 +#define BACKEND_DISABLE(x) ((x) << 16) +#define GB_ADDR_CONFIG 0x98F8 +#define NUM_PIPES(x) ((x) << 0) +#define NUM_PIPES_MASK 0x00000007 +#define NUM_PIPES_SHIFT 0 +#define PIPE_INTERLEAVE_SIZE(x) ((x) << 4) +#define PIPE_INTERLEAVE_SIZE_MASK 0x00000070 +#define PIPE_INTERLEAVE_SIZE_SHIFT 4 +#define NUM_SHADER_ENGINES(x) ((x) << 12) +#define NUM_SHADER_ENGINES_MASK 0x00003000 +#define NUM_SHADER_ENGINES_SHIFT 12 +#define SHADER_ENGINE_TILE_SIZE(x) ((x) << 16) +#define SHADER_ENGINE_TILE_SIZE_MASK 0x00070000 +#define SHADER_ENGINE_TILE_SIZE_SHIFT 16 +#define NUM_GPUS(x) ((x) << 20) +#define NUM_GPUS_MASK 0x00700000 +#define NUM_GPUS_SHIFT 20 +#define MULTI_GPU_TILE_SIZE(x) ((x) << 24) +#define MULTI_GPU_TILE_SIZE_MASK 0x03000000 +#define MULTI_GPU_TILE_SIZE_SHIFT 24 +#define ROW_SIZE(x) ((x) << 28) +#define ROW_SIZE_MASK 0x30000000 +#define ROW_SIZE_SHIFT 28 + +#define GB_TILE_MODE0 0x9910 +# define MICRO_TILE_MODE(x) ((x) << 0) +# define ADDR_SURF_DISPLAY_MICRO_TILING 0 +# define ADDR_SURF_THIN_MICRO_TILING 1 +# define ADDR_SURF_DEPTH_MICRO_TILING 2 +# define ARRAY_MODE(x) ((x) << 2) +# define ARRAY_LINEAR_GENERAL 0 +# define ARRAY_LINEAR_ALIGNED 1 +# define ARRAY_1D_TILED_THIN1 2 +# define ARRAY_2D_TILED_THIN1 4 +# define PIPE_CONFIG(x) ((x) << 6) +# define ADDR_SURF_P2 0 +# define ADDR_SURF_P4_8x16 4 +# define ADDR_SURF_P4_16x16 5 +# define ADDR_SURF_P4_16x32 6 +# define ADDR_SURF_P4_32x32 7 +# define ADDR_SURF_P8_16x16_8x16 8 +# define ADDR_SURF_P8_16x32_8x16 9 +# define ADDR_SURF_P8_32x32_8x16 10 +# define ADDR_SURF_P8_16x32_16x16 11 +# define ADDR_SURF_P8_32x32_16x16 12 +# define ADDR_SURF_P8_32x32_16x32 13 +# define ADDR_SURF_P8_32x64_32x32 14 +# define TILE_SPLIT(x) ((x) << 11) +# define ADDR_SURF_TILE_SPLIT_64B 0 +# define ADDR_SURF_TILE_SPLIT_128B 1 +# define ADDR_SURF_TILE_SPLIT_256B 2 +# define ADDR_SURF_TILE_SPLIT_512B 3 +# define ADDR_SURF_TILE_SPLIT_1KB 4 +# define ADDR_SURF_TILE_SPLIT_2KB 5 +# define ADDR_SURF_TILE_SPLIT_4KB 6 +# define BANK_WIDTH(x) ((x) << 14) +# define ADDR_SURF_BANK_WIDTH_1 0 +# define ADDR_SURF_BANK_WIDTH_2 1 +# define ADDR_SURF_BANK_WIDTH_4 2 +# define ADDR_SURF_BANK_WIDTH_8 3 +# define BANK_HEIGHT(x) ((x) << 16) +# define ADDR_SURF_BANK_HEIGHT_1 0 +# define ADDR_SURF_BANK_HEIGHT_2 1 +# define ADDR_SURF_BANK_HEIGHT_4 2 +# define ADDR_SURF_BANK_HEIGHT_8 3 +# define MACRO_TILE_ASPECT(x) ((x) << 18) +# define ADDR_SURF_MACRO_ASPECT_1 0 +# define ADDR_SURF_MACRO_ASPECT_2 1 +# define ADDR_SURF_MACRO_ASPECT_4 2 +# define ADDR_SURF_MACRO_ASPECT_8 3 +# define NUM_BANKS(x) ((x) << 20) +# define ADDR_SURF_2_BANK 0 +# define ADDR_SURF_4_BANK 1 +# define ADDR_SURF_8_BANK 2 +# define ADDR_SURF_16_BANK 3 + +#define CB_PERFCOUNTER0_SELECT0 0x9a20 +#define CB_PERFCOUNTER0_SELECT1 0x9a24 +#define CB_PERFCOUNTER1_SELECT0 0x9a28 +#define CB_PERFCOUNTER1_SELECT1 0x9a2c +#define CB_PERFCOUNTER2_SELECT0 0x9a30 +#define CB_PERFCOUNTER2_SELECT1 0x9a34 +#define CB_PERFCOUNTER3_SELECT0 0x9a38 +#define CB_PERFCOUNTER3_SELECT1 0x9a3c + +#define GC_USER_RB_BACKEND_DISABLE 0x9B7C +#define BACKEND_DISABLE_MASK 0x00FF0000 +#define BACKEND_DISABLE_SHIFT 16 + +#define TCP_CHAN_STEER_LO 0xac0c +#define TCP_CHAN_STEER_HI 0xac10 + +#define CP_RB0_BASE 0xC100 +#define CP_RB0_CNTL 0xC104 +#define RB_BUFSZ(x) ((x) << 0) +#define RB_BLKSZ(x) ((x) << 8) +#define BUF_SWAP_32BIT (2 << 16) +#define RB_NO_UPDATE (1 << 27) +#define RB_RPTR_WR_ENA (1 << 31) + +#define CP_RB0_RPTR_ADDR 0xC10C +#define CP_RB0_RPTR_ADDR_HI 0xC110 +#define CP_RB0_WPTR 0xC114 + +#define CP_PFP_UCODE_ADDR 0xC150 +#define CP_PFP_UCODE_DATA 0xC154 +#define CP_ME_RAM_RADDR 0xC158 +#define CP_ME_RAM_WADDR 0xC15C +#define CP_ME_RAM_DATA 0xC160 + +#define CP_CE_UCODE_ADDR 0xC168 +#define CP_CE_UCODE_DATA 0xC16C + +#define CP_RB1_BASE 0xC180 +#define CP_RB1_CNTL 0xC184 +#define CP_RB1_RPTR_ADDR 0xC188 +#define CP_RB1_RPTR_ADDR_HI 0xC18C +#define CP_RB1_WPTR 0xC190 +#define CP_RB2_BASE 0xC194 +#define CP_RB2_CNTL 0xC198 +#define CP_RB2_RPTR_ADDR 0xC19C +#define CP_RB2_RPTR_ADDR_HI 0xC1A0 +#define CP_RB2_WPTR 0xC1A4 +#define CP_INT_CNTL_RING0 0xC1A8 +#define CP_INT_CNTL_RING1 0xC1AC +#define CP_INT_CNTL_RING2 0xC1B0 +# define CNTX_BUSY_INT_ENABLE (1 << 19) +# define CNTX_EMPTY_INT_ENABLE (1 << 20) +# define WAIT_MEM_SEM_INT_ENABLE (1 << 21) +# define TIME_STAMP_INT_ENABLE (1 << 26) +# define CP_RINGID2_INT_ENABLE (1 << 29) +# define CP_RINGID1_INT_ENABLE (1 << 30) +# define CP_RINGID0_INT_ENABLE (1 << 31) +#define CP_INT_STATUS_RING0 0xC1B4 +#define CP_INT_STATUS_RING1 0xC1B8 +#define CP_INT_STATUS_RING2 0xC1BC +# define WAIT_MEM_SEM_INT_STAT (1 << 21) +# define TIME_STAMP_INT_STAT (1 << 26) +# define CP_RINGID2_INT_STAT (1 << 29) +# define CP_RINGID1_INT_STAT (1 << 30) +# define CP_RINGID0_INT_STAT (1 << 31) + +#define CP_DEBUG 0xC1FC + +#define RLC_CNTL 0xC300 +# define RLC_ENABLE (1 << 0) +#define RLC_RL_BASE 0xC304 +#define RLC_RL_SIZE 0xC308 +#define RLC_LB_CNTL 0xC30C +#define RLC_SAVE_AND_RESTORE_BASE 0xC310 +#define RLC_LB_CNTR_MAX 0xC314 +#define RLC_LB_CNTR_INIT 0xC318 + +#define RLC_CLEAR_STATE_RESTORE_BASE 0xC320 + +#define RLC_UCODE_ADDR 0xC32C +#define RLC_UCODE_DATA 0xC330 + +#define RLC_MC_CNTL 0xC344 +#define RLC_UCODE_CNTL 0xC348 + +#define VGT_EVENT_INITIATOR 0x28a90 +# define SAMPLE_STREAMOUTSTATS1 (1 << 0) +# define SAMPLE_STREAMOUTSTATS2 (2 << 0) +# define SAMPLE_STREAMOUTSTATS3 (3 << 0) +# define CACHE_FLUSH_TS (4 << 0) +# define CACHE_FLUSH (6 << 0) +# define CS_PARTIAL_FLUSH (7 << 0) +# define VGT_STREAMOUT_RESET (10 << 0) +# define END_OF_PIPE_INCR_DE (11 << 0) +# define END_OF_PIPE_IB_END (12 << 0) +# define RST_PIX_CNT (13 << 0) +# define VS_PARTIAL_FLUSH (15 << 0) +# define PS_PARTIAL_FLUSH (16 << 0) +# define CACHE_FLUSH_AND_INV_TS_EVENT (20 << 0) +# define ZPASS_DONE (21 << 0) +# define CACHE_FLUSH_AND_INV_EVENT (22 << 0) +# define PERFCOUNTER_START (23 << 0) +# define PERFCOUNTER_STOP (24 << 0) +# define PIPELINESTAT_START (25 << 0) +# define PIPELINESTAT_STOP (26 << 0) +# define PERFCOUNTER_SAMPLE (27 << 0) +# define SAMPLE_PIPELINESTAT (30 << 0) +# define SAMPLE_STREAMOUTSTATS (32 << 0) +# define RESET_VTX_CNT (33 << 0) +# define VGT_FLUSH (36 << 0) +# define BOTTOM_OF_PIPE_TS (40 << 0) +# define DB_CACHE_FLUSH_AND_INV (42 << 0) +# define FLUSH_AND_INV_DB_DATA_TS (43 << 0) +# define FLUSH_AND_INV_DB_META (44 << 0) +# define FLUSH_AND_INV_CB_DATA_TS (45 << 0) +# define FLUSH_AND_INV_CB_META (46 << 0) +# define CS_DONE (47 << 0) +# define PS_DONE (48 << 0) +# define FLUSH_AND_INV_CB_PIXEL_DATA (49 << 0) +# define THREAD_TRACE_START (51 << 0) +# define THREAD_TRACE_STOP (52 << 0) +# define THREAD_TRACE_FLUSH (54 << 0) +# define THREAD_TRACE_FINISH (55 << 0) + +/* + * PM4 + */ +#define PACKET_TYPE0 0 +#define PACKET_TYPE1 1 +#define PACKET_TYPE2 2 +#define PACKET_TYPE3 3 + +#define CP_PACKET_GET_TYPE(h) (((h) >> 30) & 3) +#define CP_PACKET_GET_COUNT(h) (((h) >> 16) & 0x3FFF) +#define CP_PACKET0_GET_REG(h) (((h) & 0xFFFF) << 2) +#define CP_PACKET3_GET_OPCODE(h) (((h) >> 8) & 0xFF) +#define PACKET0(reg, n) ((PACKET_TYPE0 << 30) | \ + (((reg) >> 2) & 0xFFFF) | \ + ((n) & 0x3FFF) << 16) +#define CP_PACKET2 0x80000000 +#define PACKET2_PAD_SHIFT 0 +#define PACKET2_PAD_MASK (0x3fffffff << 0) + +#define PACKET2(v) (CP_PACKET2 | REG_SET(PACKET2_PAD, (v))) + +#define PACKET3(op, n) ((PACKET_TYPE3 << 30) | \ + (((op) & 0xFF) << 8) | \ + ((n) & 0x3FFF) << 16) + +#define PACKET3_COMPUTE(op, n) (PACKET3(op, n) | 1 << 1) + +/* Packet 3 types */ +#define PACKET3_NOP 0x10 +#define PACKET3_SET_BASE 0x11 +#define PACKET3_BASE_INDEX(x) ((x) << 0) +#define GDS_PARTITION_BASE 2 +#define CE_PARTITION_BASE 3 +#define PACKET3_CLEAR_STATE 0x12 +#define PACKET3_INDEX_BUFFER_SIZE 0x13 +#define PACKET3_DISPATCH_DIRECT 0x15 +#define PACKET3_DISPATCH_INDIRECT 0x16 +#define PACKET3_ALLOC_GDS 0x1B +#define PACKET3_WRITE_GDS_RAM 0x1C +#define PACKET3_ATOMIC_GDS 0x1D +#define PACKET3_ATOMIC 0x1E +#define PACKET3_OCCLUSION_QUERY 0x1F +#define PACKET3_SET_PREDICATION 0x20 +#define PACKET3_REG_RMW 0x21 +#define PACKET3_COND_EXEC 0x22 +#define PACKET3_PRED_EXEC 0x23 +#define PACKET3_DRAW_INDIRECT 0x24 +#define PACKET3_DRAW_INDEX_INDIRECT 0x25 +#define PACKET3_INDEX_BASE 0x26 +#define PACKET3_DRAW_INDEX_2 0x27 +#define PACKET3_CONTEXT_CONTROL 0x28 +#define PACKET3_INDEX_TYPE 0x2A +#define PACKET3_DRAW_INDIRECT_MULTI 0x2C +#define PACKET3_DRAW_INDEX_AUTO 0x2D +#define PACKET3_DRAW_INDEX_IMMD 0x2E +#define PACKET3_NUM_INSTANCES 0x2F +#define PACKET3_DRAW_INDEX_MULTI_AUTO 0x30 +#define PACKET3_INDIRECT_BUFFER_CONST 0x31 +#define PACKET3_INDIRECT_BUFFER 0x32 +#define PACKET3_STRMOUT_BUFFER_UPDATE 0x34 +#define PACKET3_DRAW_INDEX_OFFSET_2 0x35 +#define PACKET3_DRAW_INDEX_MULTI_ELEMENT 0x36 +#define PACKET3_WRITE_DATA 0x37 +#define PACKET3_DRAW_INDEX_INDIRECT_MULTI 0x38 +#define PACKET3_MEM_SEMAPHORE 0x39 +#define PACKET3_MPEG_INDEX 0x3A +#define PACKET3_COPY_DW 0x3B +#define PACKET3_WAIT_REG_MEM 0x3C +#define PACKET3_MEM_WRITE 0x3D +#define PACKET3_COPY_DATA 0x40 +#define PACKET3_PFP_SYNC_ME 0x42 +#define PACKET3_SURFACE_SYNC 0x43 +# define PACKET3_DEST_BASE_0_ENA (1 << 0) +# define PACKET3_DEST_BASE_1_ENA (1 << 1) +# define PACKET3_CB0_DEST_BASE_ENA (1 << 6) +# define PACKET3_CB1_DEST_BASE_ENA (1 << 7) +# define PACKET3_CB2_DEST_BASE_ENA (1 << 8) +# define PACKET3_CB3_DEST_BASE_ENA (1 << 9) +# define PACKET3_CB4_DEST_BASE_ENA (1 << 10) +# define PACKET3_CB5_DEST_BASE_ENA (1 << 11) +# define PACKET3_CB6_DEST_BASE_ENA (1 << 12) +# define PACKET3_CB7_DEST_BASE_ENA (1 << 13) +# define PACKET3_DB_DEST_BASE_ENA (1 << 14) +# define PACKET3_DEST_BASE_2_ENA (1 << 19) +# define PACKET3_DEST_BASE_3_ENA (1 << 21) +# define PACKET3_TCL1_ACTION_ENA (1 << 22) +# define PACKET3_TC_ACTION_ENA (1 << 23) +# define PACKET3_CB_ACTION_ENA (1 << 25) +# define PACKET3_DB_ACTION_ENA (1 << 26) +# define PACKET3_SH_KCACHE_ACTION_ENA (1 << 27) +# define PACKET3_SH_ICACHE_ACTION_ENA (1 << 29) +#define PACKET3_ME_INITIALIZE 0x44 +#define PACKET3_ME_INITIALIZE_DEVICE_ID(x) ((x) << 16) +#define PACKET3_COND_WRITE 0x45 +#define PACKET3_EVENT_WRITE 0x46 +#define EVENT_TYPE(x) ((x) << 0) +#define EVENT_INDEX(x) ((x) << 8) + /* 0 - any non-TS event + * 1 - ZPASS_DONE + * 2 - SAMPLE_PIPELINESTAT + * 3 - SAMPLE_STREAMOUTSTAT* + * 4 - *S_PARTIAL_FLUSH + * 5 - EOP events + * 6 - EOS events + * 7 - CACHE_FLUSH, CACHE_FLUSH_AND_INV_EVENT + */ +#define INV_L2 (1 << 20) + /* INV TC L2 cache when EVENT_INDEX = 7 */ +#define PACKET3_EVENT_WRITE_EOP 0x47 +#define DATA_SEL(x) ((x) << 29) + /* 0 - discard + * 1 - send low 32bit data + * 2 - send 64bit data + * 3 - send 64bit counter value + */ +#define INT_SEL(x) ((x) << 24) + /* 0 - none + * 1 - interrupt only (DATA_SEL = 0) + * 2 - interrupt when data write is confirmed + */ +#define PACKET3_EVENT_WRITE_EOS 0x48 +#define PACKET3_PREAMBLE_CNTL 0x4A +# define PACKET3_PREAMBLE_BEGIN_CLEAR_STATE (2 << 28) +# define PACKET3_PREAMBLE_END_CLEAR_STATE (3 << 28) +#define PACKET3_ONE_REG_WRITE 0x57 +#define PACKET3_LOAD_CONFIG_REG 0x5F +#define PACKET3_LOAD_CONTEXT_REG 0x60 +#define PACKET3_LOAD_SH_REG 0x61 +#define PACKET3_SET_CONFIG_REG 0x68 +#define PACKET3_SET_CONFIG_REG_START 0x00008000 +#define PACKET3_SET_CONFIG_REG_END 0x0000b000 +#define PACKET3_SET_CONTEXT_REG 0x69 +#define PACKET3_SET_CONTEXT_REG_START 0x00028000 +#define PACKET3_SET_CONTEXT_REG_END 0x00029000 +#define PACKET3_SET_CONTEXT_REG_INDIRECT 0x73 +#define PACKET3_SET_RESOURCE_INDIRECT 0x74 +#define PACKET3_SET_SH_REG 0x76 +#define PACKET3_SET_SH_REG_START 0x0000b000 +#define PACKET3_SET_SH_REG_END 0x0000c000 +#define PACKET3_SET_SH_REG_OFFSET 0x77 +#define PACKET3_ME_WRITE 0x7A +#define PACKET3_SCRATCH_RAM_WRITE 0x7D +#define PACKET3_SCRATCH_RAM_READ 0x7E +#define PACKET3_CE_WRITE 0x7F +#define PACKET3_LOAD_CONST_RAM 0x80 +#define PACKET3_WRITE_CONST_RAM 0x81 +#define PACKET3_WRITE_CONST_RAM_OFFSET 0x82 +#define PACKET3_DUMP_CONST_RAM 0x83 +#define PACKET3_INCREMENT_CE_COUNTER 0x84 +#define PACKET3_INCREMENT_DE_COUNTER 0x85 +#define PACKET3_WAIT_ON_CE_COUNTER 0x86 +#define PACKET3_WAIT_ON_DE_COUNTER 0x87 +#define PACKET3_WAIT_ON_DE_COUNTER_DIFF 0x88 +#define PACKET3_SET_CE_DE_COUNTERS 0x89 +#define PACKET3_WAIT_ON_AVAIL_BUFFER 0x8A + +#endif diff --git a/drivers/gpu/drm/savage/savage_bci.c b/drivers/gpu/drm/savage/savage_bci.c index cb1ee4e..2a25888 100644 --- a/drivers/gpu/drm/savage/savage_bci.c +++ b/drivers/gpu/drm/savage/savage_bci.c @@ -547,6 +547,8 @@ int savage_driver_load(struct drm_device *dev, unsigned long chipset) dev_priv->chipset = (enum savage_family)chipset; + pci_set_master(dev->pdev); + return 0; } diff --git a/drivers/gpu/drm/savage/savage_drv.c b/drivers/gpu/drm/savage/savage_drv.c index 5468d1c..89afe0b 100644 --- a/drivers/gpu/drm/savage/savage_drv.c +++ b/drivers/gpu/drm/savage/savage_drv.c @@ -35,6 +35,17 @@ static struct pci_device_id pciidlist[] = { savage_PCI_IDS }; +static const struct file_operations savage_driver_fops = { + .owner = THIS_MODULE, + .open = drm_open, + .release = drm_release, + .unlocked_ioctl = drm_ioctl, + .mmap = drm_mmap, + .poll = drm_poll, + .fasync = drm_fasync, + .llseek = noop_llseek, +}; + static struct drm_driver driver = { .driver_features = DRIVER_USE_AGP | DRIVER_USE_MTRR | DRIVER_HAVE_DMA | DRIVER_PCI_DMA, @@ -46,17 +57,7 @@ static struct drm_driver driver = { .reclaim_buffers = savage_reclaim_buffers, .ioctls = savage_ioctls, .dma_ioctl = savage_bci_buffers, - .fops = { - .owner = THIS_MODULE, - .open = drm_open, - .release = drm_release, - .unlocked_ioctl = drm_ioctl, - .mmap = drm_mmap, - .poll = drm_poll, - .fasync = drm_fasync, - .llseek = noop_llseek, - }, - + .fops = &savage_driver_fops, .name = DRIVER_NAME, .desc = DRIVER_DESC, .date = DRIVER_DATE, diff --git a/drivers/gpu/drm/savage/savage_state.c b/drivers/gpu/drm/savage/savage_state.c index 8a3e315..b6d8608 100644 --- a/drivers/gpu/drm/savage/savage_state.c +++ b/drivers/gpu/drm/savage/savage_state.c @@ -988,7 +988,7 @@ int savage_bci_cmdbuf(struct drm_device *dev, void *data, struct drm_file *file_ * for locking on FreeBSD. */ if (cmdbuf->size) { - kcmd_addr = kmalloc(cmdbuf->size * 8, GFP_KERNEL); + kcmd_addr = kmalloc_array(cmdbuf->size, 8, GFP_KERNEL); if (kcmd_addr == NULL) return -ENOMEM; @@ -1015,8 +1015,8 @@ int savage_bci_cmdbuf(struct drm_device *dev, void *data, struct drm_file *file_ cmdbuf->vb_addr = kvb_addr; } if (cmdbuf->nbox) { - kbox_addr = kmalloc(cmdbuf->nbox * sizeof(struct drm_clip_rect), - GFP_KERNEL); + kbox_addr = kmalloc_array(cmdbuf->nbox, sizeof(struct drm_clip_rect), + GFP_KERNEL); if (kbox_addr == NULL) { ret = -ENOMEM; goto done; @@ -1057,7 +1057,8 @@ int savage_bci_cmdbuf(struct drm_device *dev, void *data, struct drm_file *file_ DRM_ERROR("indexed drawing command extends " "beyond end of command buffer\n"); DMA_FLUSH(); - return -EINVAL; + ret = -EINVAL; + goto done; } /* fall through */ case SAVAGE_CMD_DMA_PRIM: @@ -1076,7 +1077,7 @@ int savage_bci_cmdbuf(struct drm_device *dev, void *data, struct drm_file *file_ cmdbuf->vb_stride, cmdbuf->nbox, cmdbuf->box_addr); if (ret != 0) - return ret; + goto done; first_draw_cmd = NULL; } } diff --git a/drivers/gpu/drm/sis/sis_drv.c b/drivers/gpu/drm/sis/sis_drv.c index a9c5716..dd14cd1 100644 --- a/drivers/gpu/drm/sis/sis_drv.c +++ b/drivers/gpu/drm/sis/sis_drv.c @@ -40,51 +40,78 @@ static struct pci_device_id pciidlist[] = { static int sis_driver_load(struct drm_device *dev, unsigned long chipset) { drm_sis_private_t *dev_priv; - int ret; + + pci_set_master(dev->pdev); dev_priv = kzalloc(sizeof(drm_sis_private_t), GFP_KERNEL); if (dev_priv == NULL) return -ENOMEM; + idr_init(&dev_priv->object_idr); dev->dev_private = (void *)dev_priv; dev_priv->chipset = chipset; - ret = drm_sman_init(&dev_priv->sman, 2, 12, 8); - if (ret) - kfree(dev_priv); - return ret; + return 0; } static int sis_driver_unload(struct drm_device *dev) { drm_sis_private_t *dev_priv = dev->dev_private; - drm_sman_takedown(&dev_priv->sman); + idr_remove_all(&dev_priv->object_idr); + idr_destroy(&dev_priv->object_idr); + kfree(dev_priv); return 0; } +static const struct file_operations sis_driver_fops = { + .owner = THIS_MODULE, + .open = drm_open, + .release = drm_release, + .unlocked_ioctl = drm_ioctl, + .mmap = drm_mmap, + .poll = drm_poll, + .fasync = drm_fasync, + .llseek = noop_llseek, +}; + +static int sis_driver_open(struct drm_device *dev, struct drm_file *file) +{ + struct sis_file_private *file_priv; + + DRM_DEBUG_DRIVER("\n"); + file_priv = kmalloc(sizeof(*file_priv), GFP_KERNEL); + if (!file_priv) + return -ENOMEM; + + file->driver_priv = file_priv; + + INIT_LIST_HEAD(&file_priv->obj_list); + + return 0; +} + +void sis_driver_postclose(struct drm_device *dev, struct drm_file *file) +{ + struct sis_file_private *file_priv = file->driver_priv; + + kfree(file_priv); +} + static struct drm_driver driver = { .driver_features = DRIVER_USE_AGP | DRIVER_USE_MTRR, .load = sis_driver_load, .unload = sis_driver_unload, + .open = sis_driver_open, + .postclose = sis_driver_postclose, .dma_quiescent = sis_idle, .reclaim_buffers = NULL, .reclaim_buffers_idlelocked = sis_reclaim_buffers_locked, .lastclose = sis_lastclose, .ioctls = sis_ioctls, - .fops = { - .owner = THIS_MODULE, - .open = drm_open, - .release = drm_release, - .unlocked_ioctl = drm_ioctl, - .mmap = drm_mmap, - .poll = drm_poll, - .fasync = drm_fasync, - .llseek = noop_llseek, - }, - + .fops = &sis_driver_fops, .name = DRIVER_NAME, .desc = DRIVER_DESC, .date = DRIVER_DATE, diff --git a/drivers/gpu/drm/sis/sis_drv.h b/drivers/gpu/drm/sis/sis_drv.h index 194303c..573758b 100644 --- a/drivers/gpu/drm/sis/sis_drv.h +++ b/drivers/gpu/drm/sis/sis_drv.h @@ -44,7 +44,7 @@ enum sis_family { SIS_CHIP_315 = 1, }; -#include "drm_sman.h" +#include "drm_mm.h" #define SIS_BASE (dev_priv->mmio) @@ -54,12 +54,15 @@ enum sis_family { typedef struct drm_sis_private { drm_local_map_t *mmio; unsigned int idle_fault; - struct drm_sman sman; unsigned int chipset; int vram_initialized; int agp_initialized; unsigned long vram_offset; unsigned long agp_offset; + struct drm_mm vram_mm; + struct drm_mm agp_mm; + /** Mapping of userspace keys to mm objects */ + struct idr object_idr; } drm_sis_private_t; extern int sis_idle(struct drm_device *dev); diff --git a/drivers/gpu/drm/sis/sis_mm.c b/drivers/gpu/drm/sis/sis_mm.c index 7fe2b63..dd4a316 100644 --- a/drivers/gpu/drm/sis/sis_mm.c +++ b/drivers/gpu/drm/sis/sis_mm.c @@ -41,40 +41,18 @@ #define AGP_TYPE 1 +struct sis_memblock { + struct drm_mm_node mm_node; + struct sis_memreq req; + struct list_head owner_list; +}; + #if defined(CONFIG_FB_SIS) || defined(CONFIG_FB_SIS_MODULE) /* fb management via fb device */ #define SIS_MM_ALIGN_SHIFT 0 #define SIS_MM_ALIGN_MASK 0 -static void *sis_sman_mm_allocate(void *private, unsigned long size, - unsigned alignment) -{ - struct sis_memreq req; - - req.size = size; - sis_malloc(&req); - if (req.size == 0) - return NULL; - else - return (void *)(unsigned long)~req.offset; -} - -static void sis_sman_mm_free(void *private, void *ref) -{ - sis_free(~((unsigned long)ref)); -} - -static void sis_sman_mm_destroy(void *private) -{ - ; -} - -static unsigned long sis_sman_mm_offset(void *private, void *ref) -{ - return ~((unsigned long)ref); -} - #else /* CONFIG_FB_SIS[_MODULE] */ #define SIS_MM_ALIGN_SHIFT 4 @@ -86,30 +64,11 @@ static int sis_fb_init(struct drm_device *dev, void *data, struct drm_file *file { drm_sis_private_t *dev_priv = dev->dev_private; drm_sis_fb_t *fb = data; - int ret; mutex_lock(&dev->struct_mutex); -#if defined(CONFIG_FB_SIS) || defined(CONFIG_FB_SIS_MODULE) - { - struct drm_sman_mm sman_mm; - sman_mm.private = (void *)0xFFFFFFFF; - sman_mm.allocate = sis_sman_mm_allocate; - sman_mm.free = sis_sman_mm_free; - sman_mm.destroy = sis_sman_mm_destroy; - sman_mm.offset = sis_sman_mm_offset; - ret = - drm_sman_set_manager(&dev_priv->sman, VIDEO_TYPE, &sman_mm); - } -#else - ret = drm_sman_set_range(&dev_priv->sman, VIDEO_TYPE, 0, - fb->size >> SIS_MM_ALIGN_SHIFT); -#endif - - if (ret) { - DRM_ERROR("VRAM memory manager initialisation error\n"); - mutex_unlock(&dev->struct_mutex); - return ret; - } + /* Unconditionally init the drm_mm, even though we don't use it when the + * fb sis driver is available - make cleanup easier. */ + drm_mm_init(&dev_priv->vram_mm, 0, fb->size >> SIS_MM_ALIGN_SHIFT); dev_priv->vram_initialized = 1; dev_priv->vram_offset = fb->offset; @@ -120,13 +79,15 @@ static int sis_fb_init(struct drm_device *dev, void *data, struct drm_file *file return 0; } -static int sis_drm_alloc(struct drm_device *dev, struct drm_file *file_priv, +static int sis_drm_alloc(struct drm_device *dev, struct drm_file *file, void *data, int pool) { drm_sis_private_t *dev_priv = dev->dev_private; drm_sis_mem_t *mem = data; - int retval = 0; - struct drm_memblock_item *item; + int retval = 0, user_key; + struct sis_memblock *item; + struct sis_file_private *file_priv = file->driver_priv; + unsigned long offset; mutex_lock(&dev->struct_mutex); @@ -138,25 +99,68 @@ static int sis_drm_alloc(struct drm_device *dev, struct drm_file *file_priv, return -EINVAL; } - mem->size = (mem->size + SIS_MM_ALIGN_MASK) >> SIS_MM_ALIGN_SHIFT; - item = drm_sman_alloc(&dev_priv->sman, pool, mem->size, 0, - (unsigned long)file_priv); + item = kzalloc(sizeof(*item), GFP_KERNEL); + if (!item) { + retval = -ENOMEM; + goto fail_alloc; + } - mutex_unlock(&dev->struct_mutex); - if (item) { - mem->offset = ((pool == 0) ? - dev_priv->vram_offset : dev_priv->agp_offset) + - (item->mm-> - offset(item->mm, item->mm_info) << SIS_MM_ALIGN_SHIFT); - mem->free = item->user_hash.key; - mem->size = mem->size << SIS_MM_ALIGN_SHIFT; + mem->size = (mem->size + SIS_MM_ALIGN_MASK) >> SIS_MM_ALIGN_SHIFT; + if (pool == AGP_TYPE) { + retval = drm_mm_insert_node(&dev_priv->agp_mm, + &item->mm_node, + mem->size, 0); + offset = item->mm_node.start; } else { - mem->offset = 0; - mem->size = 0; - mem->free = 0; +#if defined(CONFIG_FB_SIS) || defined(CONFIG_FB_SIS_MODULE) + item->req.size = mem->size; + sis_malloc(&item->req); + if (item->req.size == 0) + retval = -ENOMEM; + offset = item->req.offset; +#else + retval = drm_mm_insert_node(&dev_priv->vram_mm, + &item->mm_node, + mem->size, 0); + offset = item->mm_node.start; +#endif + } + if (retval) + goto fail_alloc; + +again: + if (idr_pre_get(&dev_priv->object_idr, GFP_KERNEL) == 0) { retval = -ENOMEM; + goto fail_idr; } + retval = idr_get_new_above(&dev_priv->object_idr, item, 1, &user_key); + if (retval == -EAGAIN) + goto again; + if (retval) + goto fail_idr; + + list_add(&item->owner_list, &file_priv->obj_list); + mutex_unlock(&dev->struct_mutex); + + mem->offset = ((pool == 0) ? + dev_priv->vram_offset : dev_priv->agp_offset) + + (offset << SIS_MM_ALIGN_SHIFT); + mem->free = user_key; + mem->size = mem->size << SIS_MM_ALIGN_SHIFT; + + return 0; + +fail_idr: + drm_mm_remove_node(&item->mm_node); +fail_alloc: + kfree(item); + mutex_unlock(&dev->struct_mutex); + + mem->offset = 0; + mem->size = 0; + mem->free = 0; + DRM_DEBUG("alloc %d, size = %d, offset = %d\n", pool, mem->size, mem->offset); @@ -167,14 +171,28 @@ static int sis_drm_free(struct drm_device *dev, void *data, struct drm_file *fil { drm_sis_private_t *dev_priv = dev->dev_private; drm_sis_mem_t *mem = data; - int ret; + struct sis_memblock *obj; mutex_lock(&dev->struct_mutex); - ret = drm_sman_free_key(&dev_priv->sman, mem->free); + obj = idr_find(&dev_priv->object_idr, mem->free); + if (obj == NULL) { + mutex_unlock(&dev->struct_mutex); + return -EINVAL; + } + + idr_remove(&dev_priv->object_idr, mem->free); + list_del(&obj->owner_list); + if (drm_mm_node_allocated(&obj->mm_node)) + drm_mm_remove_node(&obj->mm_node); +#if defined(CONFIG_FB_SIS) || defined(CONFIG_FB_SIS_MODULE) + else + sis_free(obj->req.offset); +#endif + kfree(obj); mutex_unlock(&dev->struct_mutex); DRM_DEBUG("free = 0x%lx\n", mem->free); - return ret; + return 0; } static int sis_fb_alloc(struct drm_device *dev, void *data, @@ -188,18 +206,10 @@ static int sis_ioctl_agp_init(struct drm_device *dev, void *data, { drm_sis_private_t *dev_priv = dev->dev_private; drm_sis_agp_t *agp = data; - int ret; dev_priv = dev->dev_private; mutex_lock(&dev->struct_mutex); - ret = drm_sman_set_range(&dev_priv->sman, AGP_TYPE, 0, - agp->size >> SIS_MM_ALIGN_SHIFT); - - if (ret) { - DRM_ERROR("AGP memory manager initialisation error\n"); - mutex_unlock(&dev->struct_mutex); - return ret; - } + drm_mm_init(&dev_priv->agp_mm, 0, agp->size >> SIS_MM_ALIGN_SHIFT); dev_priv->agp_initialized = 1; dev_priv->agp_offset = agp->offset; @@ -293,20 +303,26 @@ void sis_lastclose(struct drm_device *dev) return; mutex_lock(&dev->struct_mutex); - drm_sman_cleanup(&dev_priv->sman); - dev_priv->vram_initialized = 0; - dev_priv->agp_initialized = 0; + if (dev_priv->vram_initialized) { + drm_mm_takedown(&dev_priv->vram_mm); + dev_priv->vram_initialized = 0; + } + if (dev_priv->agp_initialized) { + drm_mm_takedown(&dev_priv->agp_mm); + dev_priv->agp_initialized = 0; + } dev_priv->mmio = NULL; mutex_unlock(&dev->struct_mutex); } void sis_reclaim_buffers_locked(struct drm_device *dev, - struct drm_file *file_priv) + struct drm_file *file) { - drm_sis_private_t *dev_priv = dev->dev_private; + struct sis_file_private *file_priv = file->driver_priv; + struct sis_memblock *entry, *next; mutex_lock(&dev->struct_mutex); - if (drm_sman_owner_clean(&dev_priv->sman, (unsigned long)file_priv)) { + if (list_empty(&file_priv->obj_list)) { mutex_unlock(&dev->struct_mutex); return; } @@ -314,7 +330,18 @@ void sis_reclaim_buffers_locked(struct drm_device *dev, if (dev->driver->dma_quiescent) dev->driver->dma_quiescent(dev); - drm_sman_owner_cleanup(&dev_priv->sman, (unsigned long)file_priv); + + list_for_each_entry_safe(entry, next, &file_priv->obj_list, + owner_list) { + list_del(&entry->owner_list); + if (drm_mm_node_allocated(&entry->mm_node)) + drm_mm_remove_node(&entry->mm_node); +#if defined(CONFIG_FB_SIS) || defined(CONFIG_FB_SIS_MODULE) + else + sis_free(entry->req.offset); +#endif + kfree(entry); + } mutex_unlock(&dev->struct_mutex); return; } diff --git a/drivers/gpu/drm/tdfx/tdfx_drv.c b/drivers/gpu/drm/tdfx/tdfx_drv.c index cda2991..1613c78 100644 --- a/drivers/gpu/drm/tdfx/tdfx_drv.c +++ b/drivers/gpu/drm/tdfx/tdfx_drv.c @@ -41,20 +41,21 @@ static struct pci_device_id pciidlist[] = { tdfx_PCI_IDS }; +static const struct file_operations tdfx_driver_fops = { + .owner = THIS_MODULE, + .open = drm_open, + .release = drm_release, + .unlocked_ioctl = drm_ioctl, + .mmap = drm_mmap, + .poll = drm_poll, + .fasync = drm_fasync, + .llseek = noop_llseek, +}; + static struct drm_driver driver = { .driver_features = DRIVER_USE_MTRR, .reclaim_buffers = drm_core_reclaim_buffers, - .fops = { - .owner = THIS_MODULE, - .open = drm_open, - .release = drm_release, - .unlocked_ioctl = drm_ioctl, - .mmap = drm_mmap, - .poll = drm_poll, - .fasync = drm_fasync, - .llseek = noop_llseek, - }, - + .fops = &tdfx_driver_fops, .name = DRIVER_NAME, .desc = DRIVER_DESC, .date = DRIVER_DATE, diff --git a/drivers/gpu/drm/ttm/Makefile b/drivers/gpu/drm/ttm/Makefile index f3cf6f0..b2b33dd 100644 --- a/drivers/gpu/drm/ttm/Makefile +++ b/drivers/gpu/drm/ttm/Makefile @@ -7,4 +7,8 @@ ttm-y := ttm_agp_backend.o ttm_memory.o ttm_tt.o ttm_bo.o \ ttm_object.o ttm_lock.o ttm_execbuf_util.o ttm_page_alloc.o \ ttm_bo_manager.o +ifeq ($(CONFIG_SWIOTLB),y) +ttm-y += ttm_page_alloc_dma.o +endif + obj-$(CONFIG_DRM_TTM) += ttm.o diff --git a/drivers/gpu/drm/ttm/ttm_agp_backend.c b/drivers/gpu/drm/ttm/ttm_agp_backend.c index 1c4a72f..4a87282 100644 --- a/drivers/gpu/drm/ttm/ttm_agp_backend.c +++ b/drivers/gpu/drm/ttm/ttm_agp_backend.c @@ -29,8 +29,11 @@ * Keith Packard. */ +#define pr_fmt(fmt) "[TTM] " fmt + #include "ttm/ttm_module.h" #include "ttm/ttm_bo_driver.h" +#include "ttm/ttm_page_alloc.h" #ifdef TTM_HAS_AGP #include "ttm/ttm_placement.h" #include @@ -40,100 +43,77 @@ #include struct ttm_agp_backend { - struct ttm_backend backend; + struct ttm_tt ttm; struct agp_memory *mem; struct agp_bridge_data *bridge; }; -static int ttm_agp_populate(struct ttm_backend *backend, - unsigned long num_pages, struct page **pages, - struct page *dummy_read_page, - dma_addr_t *dma_addrs) +static int ttm_agp_bind(struct ttm_tt *ttm, struct ttm_mem_reg *bo_mem) { - struct ttm_agp_backend *agp_be = - container_of(backend, struct ttm_agp_backend, backend); - struct page **cur_page, **last_page = pages + num_pages; + struct ttm_agp_backend *agp_be = container_of(ttm, struct ttm_agp_backend, ttm); + struct drm_mm_node *node = bo_mem->mm_node; struct agp_memory *mem; + int ret, cached = (bo_mem->placement & TTM_PL_FLAG_CACHED); + unsigned i; - mem = agp_allocate_memory(agp_be->bridge, num_pages, AGP_USER_MEMORY); + mem = agp_allocate_memory(agp_be->bridge, ttm->num_pages, AGP_USER_MEMORY); if (unlikely(mem == NULL)) return -ENOMEM; mem->page_count = 0; - for (cur_page = pages; cur_page < last_page; ++cur_page) { - struct page *page = *cur_page; + for (i = 0; i < ttm->num_pages; i++) { + struct page *page = ttm->pages[i]; + if (!page) - page = dummy_read_page; + page = ttm->dummy_read_page; mem->pages[mem->page_count++] = page; } agp_be->mem = mem; - return 0; -} - -static int ttm_agp_bind(struct ttm_backend *backend, struct ttm_mem_reg *bo_mem) -{ - struct ttm_agp_backend *agp_be = - container_of(backend, struct ttm_agp_backend, backend); - struct drm_mm_node *node = bo_mem->mm_node; - struct agp_memory *mem = agp_be->mem; - int cached = (bo_mem->placement & TTM_PL_FLAG_CACHED); - int ret; mem->is_flushed = 1; mem->type = (cached) ? AGP_USER_CACHED_MEMORY : AGP_USER_MEMORY; ret = agp_bind_memory(mem, node->start); if (ret) - printk(KERN_ERR TTM_PFX "AGP Bind memory failed.\n"); + pr_err("AGP Bind memory failed\n"); return ret; } -static int ttm_agp_unbind(struct ttm_backend *backend) +static int ttm_agp_unbind(struct ttm_tt *ttm) { - struct ttm_agp_backend *agp_be = - container_of(backend, struct ttm_agp_backend, backend); - - if (agp_be->mem->is_bound) - return agp_unbind_memory(agp_be->mem); - else - return 0; -} + struct ttm_agp_backend *agp_be = container_of(ttm, struct ttm_agp_backend, ttm); -static void ttm_agp_clear(struct ttm_backend *backend) -{ - struct ttm_agp_backend *agp_be = - container_of(backend, struct ttm_agp_backend, backend); - struct agp_memory *mem = agp_be->mem; - - if (mem) { - ttm_agp_unbind(backend); - agp_free_memory(mem); + if (agp_be->mem) { + if (agp_be->mem->is_bound) + return agp_unbind_memory(agp_be->mem); + agp_free_memory(agp_be->mem); + agp_be->mem = NULL; } - agp_be->mem = NULL; + return 0; } -static void ttm_agp_destroy(struct ttm_backend *backend) +static void ttm_agp_destroy(struct ttm_tt *ttm) { - struct ttm_agp_backend *agp_be = - container_of(backend, struct ttm_agp_backend, backend); + struct ttm_agp_backend *agp_be = container_of(ttm, struct ttm_agp_backend, ttm); if (agp_be->mem) - ttm_agp_clear(backend); + ttm_agp_unbind(ttm); + ttm_tt_fini(ttm); kfree(agp_be); } static struct ttm_backend_func ttm_agp_func = { - .populate = ttm_agp_populate, - .clear = ttm_agp_clear, .bind = ttm_agp_bind, .unbind = ttm_agp_unbind, .destroy = ttm_agp_destroy, }; -struct ttm_backend *ttm_agp_backend_init(struct ttm_bo_device *bdev, - struct agp_bridge_data *bridge) +struct ttm_tt *ttm_agp_tt_create(struct ttm_bo_device *bdev, + struct agp_bridge_data *bridge, + unsigned long size, uint32_t page_flags, + struct page *dummy_read_page) { struct ttm_agp_backend *agp_be; @@ -143,10 +123,29 @@ struct ttm_backend *ttm_agp_backend_init(struct ttm_bo_device *bdev, agp_be->mem = NULL; agp_be->bridge = bridge; - agp_be->backend.func = &ttm_agp_func; - agp_be->backend.bdev = bdev; - return &agp_be->backend; + agp_be->ttm.func = &ttm_agp_func; + + if (ttm_tt_init(&agp_be->ttm, bdev, size, page_flags, dummy_read_page)) { + return NULL; + } + + return &agp_be->ttm; +} +EXPORT_SYMBOL(ttm_agp_tt_create); + +int ttm_agp_tt_populate(struct ttm_tt *ttm) +{ + if (ttm->state != tt_unpopulated) + return 0; + + return ttm_pool_populate(ttm); +} +EXPORT_SYMBOL(ttm_agp_tt_populate); + +void ttm_agp_tt_unpopulate(struct ttm_tt *ttm) +{ + ttm_pool_unpopulate(ttm); } -EXPORT_SYMBOL(ttm_agp_backend_init); +EXPORT_SYMBOL(ttm_agp_tt_unpopulate); #endif diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index 0d27bff..8b73ae8 100644 --- a/drivers/gpu/drm/ttm/ttm_bo.c +++ b/drivers/gpu/drm/ttm/ttm_bo.c @@ -28,6 +28,8 @@ * Authors: Thomas Hellstrom */ +#define pr_fmt(fmt) "[TTM] " fmt + #include "ttm/ttm_module.h" #include "ttm/ttm_bo_driver.h" #include "ttm/ttm_placement.h" @@ -68,15 +70,13 @@ static void ttm_mem_type_debug(struct ttm_bo_device *bdev, int mem_type) { struct ttm_mem_type_manager *man = &bdev->man[mem_type]; - printk(KERN_ERR TTM_PFX " has_type: %d\n", man->has_type); - printk(KERN_ERR TTM_PFX " use_type: %d\n", man->use_type); - printk(KERN_ERR TTM_PFX " flags: 0x%08X\n", man->flags); - printk(KERN_ERR TTM_PFX " gpu_offset: 0x%08lX\n", man->gpu_offset); - printk(KERN_ERR TTM_PFX " size: %llu\n", man->size); - printk(KERN_ERR TTM_PFX " available_caching: 0x%08X\n", - man->available_caching); - printk(KERN_ERR TTM_PFX " default_caching: 0x%08X\n", - man->default_caching); + pr_err(" has_type: %d\n", man->has_type); + pr_err(" use_type: %d\n", man->use_type); + pr_err(" flags: 0x%08X\n", man->flags); + pr_err(" gpu_offset: 0x%08lX\n", man->gpu_offset); + pr_err(" size: %llu\n", man->size); + pr_err(" available_caching: 0x%08X\n", man->available_caching); + pr_err(" default_caching: 0x%08X\n", man->default_caching); if (mem_type != TTM_PL_SYSTEM) (*man->func->debug)(man, TTM_PFX); } @@ -86,16 +86,16 @@ static void ttm_bo_mem_space_debug(struct ttm_buffer_object *bo, { int i, ret, mem_type; - printk(KERN_ERR TTM_PFX "No space for %p (%lu pages, %luK, %luM)\n", - bo, bo->mem.num_pages, bo->mem.size >> 10, - bo->mem.size >> 20); + pr_err("No space for %p (%lu pages, %luK, %luM)\n", + bo, bo->mem.num_pages, bo->mem.size >> 10, + bo->mem.size >> 20); for (i = 0; i < placement->num_placement; i++) { ret = ttm_mem_type_from_flags(placement->placement[i], &mem_type); if (ret) return; - printk(KERN_ERR TTM_PFX " placement[%d]=0x%08X (%d)\n", - i, placement->placement[i], mem_type); + pr_err(" placement[%d]=0x%08X (%d)\n", + i, placement->placement[i], mem_type); ttm_mem_type_debug(bo->bdev, mem_type); } } @@ -137,6 +137,7 @@ static void ttm_bo_release_list(struct kref *list_kref) struct ttm_buffer_object *bo = container_of(list_kref, struct ttm_buffer_object, list_kref); struct ttm_bo_device *bdev = bo->bdev; + size_t acc_size = bo->acc_size; BUG_ON(atomic_read(&bo->list_kref.refcount)); BUG_ON(atomic_read(&bo->kref.refcount)); @@ -152,9 +153,9 @@ static void ttm_bo_release_list(struct kref *list_kref) if (bo->destroy) bo->destroy(bo); else { - ttm_mem_global_free(bdev->glob->mem_glob, bo->acc_size); kfree(bo); } + ttm_mem_global_free(bdev->glob->mem_glob, acc_size); } int ttm_bo_wait_unreserved(struct ttm_buffer_object *bo, bool interruptible) @@ -337,29 +338,13 @@ static int ttm_bo_add_ttm(struct ttm_buffer_object *bo, bool zero_alloc) if (zero_alloc) page_flags |= TTM_PAGE_FLAG_ZERO_ALLOC; case ttm_bo_type_kernel: - bo->ttm = ttm_tt_create(bdev, bo->num_pages << PAGE_SHIFT, - page_flags, glob->dummy_read_page); + bo->ttm = bdev->driver->ttm_tt_create(bdev, bo->num_pages << PAGE_SHIFT, + page_flags, glob->dummy_read_page); if (unlikely(bo->ttm == NULL)) ret = -ENOMEM; break; - case ttm_bo_type_user: - bo->ttm = ttm_tt_create(bdev, bo->num_pages << PAGE_SHIFT, - page_flags | TTM_PAGE_FLAG_USER, - glob->dummy_read_page); - if (unlikely(bo->ttm == NULL)) { - ret = -ENOMEM; - break; - } - - ret = ttm_tt_set_user(bo->ttm, current, - bo->buffer_start, bo->num_pages); - if (unlikely(ret != 0)) { - ttm_tt_destroy(bo->ttm); - bo->ttm = NULL; - } - break; default: - printk(KERN_ERR TTM_PFX "Illegal buffer object type\n"); + pr_err("Illegal buffer object type\n"); ret = -EINVAL; break; } @@ -431,14 +416,23 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo, else ret = ttm_bo_move_memcpy(bo, evict, no_wait_reserve, no_wait_gpu, mem); - if (ret) + if (ret) { + if (bdev->driver->move_notify) { + struct ttm_mem_reg tmp_mem = *mem; + *mem = bo->mem; + bo->mem = tmp_mem; + bdev->driver->move_notify(bo, mem); + bo->mem = *mem; + } + goto out_err; + } moved: if (bo->evicted) { ret = bdev->driver->invalidate_caches(bdev, bo->mem.placement); if (ret) - printk(KERN_ERR TTM_PFX "Can not flush read caches\n"); + pr_err("Can not flush read caches\n"); bo->evicted = false; } @@ -472,6 +466,9 @@ out_err: static void ttm_bo_cleanup_memtype_use(struct ttm_buffer_object *bo) { + if (bo->bdev->driver->move_notify) + bo->bdev->driver->move_notify(bo, NULL); + if (bo->ttm) { ttm_tt_unbind(bo->ttm); ttm_tt_destroy(bo->ttm); @@ -737,9 +734,7 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible, if (unlikely(ret != 0)) { if (ret != -ERESTARTSYS) { - printk(KERN_ERR TTM_PFX - "Failed to expire sync object before " - "buffer eviction.\n"); + pr_err("Failed to expire sync object before buffer eviction\n"); } goto out; } @@ -760,9 +755,8 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible, no_wait_reserve, no_wait_gpu); if (ret) { if (ret != -ERESTARTSYS) { - printk(KERN_ERR TTM_PFX - "Failed to find memory space for " - "buffer 0x%p eviction.\n", bo); + pr_err("Failed to find memory space for buffer 0x%p eviction\n", + bo); ttm_bo_mem_space_debug(bo, &placement); } goto out; @@ -772,7 +766,7 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible, no_wait_reserve, no_wait_gpu); if (ret) { if (ret != -ERESTARTSYS) - printk(KERN_ERR TTM_PFX "Buffer eviction failed\n"); + pr_err("Buffer eviction failed\n"); ttm_bo_mem_put(bo, &evict_mem); goto out; } @@ -913,16 +907,12 @@ static uint32_t ttm_bo_select_caching(struct ttm_mem_type_manager *man, } static bool ttm_bo_mt_compatible(struct ttm_mem_type_manager *man, - bool disallow_fixed, uint32_t mem_type, uint32_t proposed_placement, uint32_t *masked_placement) { uint32_t cur_flags = ttm_bo_type_flags(mem_type); - if ((man->flags & TTM_MEMTYPE_FLAG_FIXED) && disallow_fixed) - return false; - if ((cur_flags & proposed_placement & TTM_PL_MASK_MEM) == 0) return false; @@ -967,7 +957,6 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo, man = &bdev->man[mem_type]; type_ok = ttm_bo_mt_compatible(man, - bo->type == ttm_bo_type_user, mem_type, placement->placement[i], &cur_flags); @@ -1015,7 +1004,6 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo, if (!man->has_type) continue; if (!ttm_bo_mt_compatible(man, - bo->type == ttm_bo_type_user, mem_type, placement->busy_placement[i], &cur_flags)) @@ -1185,15 +1173,27 @@ int ttm_bo_init(struct ttm_bo_device *bdev, { int ret = 0; unsigned long num_pages; + struct ttm_mem_global *mem_glob = bdev->glob->mem_glob; + + ret = ttm_mem_global_alloc(mem_glob, acc_size, false, false); + if (ret) { + pr_err("Out of kernel memory\n"); + if (destroy) + (*destroy)(bo); + else + kfree(bo); + return -ENOMEM; + } size += buffer_start & ~PAGE_MASK; num_pages = (size + PAGE_SIZE - 1) >> PAGE_SHIFT; if (num_pages == 0) { - printk(KERN_ERR TTM_PFX "Illegal buffer object size.\n"); + pr_err("Illegal buffer object size\n"); if (destroy) (*destroy)(bo); else kfree(bo); + ttm_mem_global_free(mem_glob, acc_size); return -EINVAL; } bo->destroy = destroy; @@ -1255,14 +1255,34 @@ out_err: } EXPORT_SYMBOL(ttm_bo_init); -static inline size_t ttm_bo_size(struct ttm_bo_global *glob, - unsigned long num_pages) +size_t ttm_bo_acc_size(struct ttm_bo_device *bdev, + unsigned long bo_size, + unsigned struct_size) { - size_t page_array_size = (num_pages * sizeof(void *) + PAGE_SIZE - 1) & - PAGE_MASK; + unsigned npages = (PAGE_ALIGN(bo_size)) >> PAGE_SHIFT; + size_t size = 0; - return glob->ttm_bo_size + 2 * page_array_size; + size += ttm_round_pot(struct_size); + size += PAGE_ALIGN(npages * sizeof(void *)); + size += ttm_round_pot(sizeof(struct ttm_tt)); + return size; } +EXPORT_SYMBOL(ttm_bo_acc_size); + +size_t ttm_bo_dma_acc_size(struct ttm_bo_device *bdev, + unsigned long bo_size, + unsigned struct_size) +{ + unsigned npages = (PAGE_ALIGN(bo_size)) >> PAGE_SHIFT; + size_t size = 0; + + size += ttm_round_pot(struct_size); + size += PAGE_ALIGN(npages * sizeof(void *)); + size += PAGE_ALIGN(npages * sizeof(dma_addr_t)); + size += ttm_round_pot(sizeof(struct ttm_dma_tt)); + return size; +} +EXPORT_SYMBOL(ttm_bo_dma_acc_size); int ttm_bo_create(struct ttm_bo_device *bdev, unsigned long size, @@ -1275,22 +1295,14 @@ int ttm_bo_create(struct ttm_bo_device *bdev, struct ttm_buffer_object **p_bo) { struct ttm_buffer_object *bo; - struct ttm_mem_global *mem_glob = bdev->glob->mem_glob; + size_t acc_size; int ret; - size_t acc_size = - ttm_bo_size(bdev->glob, (size + PAGE_SIZE - 1) >> PAGE_SHIFT); - ret = ttm_mem_global_alloc(mem_glob, acc_size, false, false); - if (unlikely(ret != 0)) - return ret; - bo = kzalloc(sizeof(*bo), GFP_KERNEL); - - if (unlikely(bo == NULL)) { - ttm_mem_global_free(mem_glob, acc_size); + if (unlikely(bo == NULL)) return -ENOMEM; - } + acc_size = ttm_bo_acc_size(bdev, size, sizeof(struct ttm_buffer_object)); ret = ttm_bo_init(bdev, bo, size, type, placement, page_alignment, buffer_start, interruptible, persistent_swap_storage, acc_size, NULL); @@ -1320,8 +1332,7 @@ static int ttm_bo_force_list_clean(struct ttm_bo_device *bdev, if (allow_errors) { return ret; } else { - printk(KERN_ERR TTM_PFX - "Cleanup eviction failed\n"); + pr_err("Cleanup eviction failed\n"); } } spin_lock(&glob->lru_lock); @@ -1336,14 +1347,14 @@ int ttm_bo_clean_mm(struct ttm_bo_device *bdev, unsigned mem_type) int ret = -EINVAL; if (mem_type >= TTM_NUM_MEM_TYPES) { - printk(KERN_ERR TTM_PFX "Illegal memory type %d\n", mem_type); + pr_err("Illegal memory type %d\n", mem_type); return ret; } man = &bdev->man[mem_type]; if (!man->has_type) { - printk(KERN_ERR TTM_PFX "Trying to take down uninitialized " - "memory manager type %u\n", mem_type); + pr_err("Trying to take down uninitialized memory manager type %u\n", + mem_type); return ret; } @@ -1366,16 +1377,12 @@ int ttm_bo_evict_mm(struct ttm_bo_device *bdev, unsigned mem_type) struct ttm_mem_type_manager *man = &bdev->man[mem_type]; if (mem_type == 0 || mem_type >= TTM_NUM_MEM_TYPES) { - printk(KERN_ERR TTM_PFX - "Illegal memory manager memory type %u.\n", - mem_type); + pr_err("Illegal memory manager memory type %u\n", mem_type); return -EINVAL; } if (!man->has_type) { - printk(KERN_ERR TTM_PFX - "Memory type %u has not been initialized.\n", - mem_type); + pr_err("Memory type %u has not been initialized\n", mem_type); return 0; } @@ -1460,18 +1467,10 @@ int ttm_bo_global_init(struct drm_global_reference *ref) ttm_mem_init_shrink(&glob->shrink, ttm_bo_swapout); ret = ttm_mem_register_shrink(glob->mem_glob, &glob->shrink); if (unlikely(ret != 0)) { - printk(KERN_ERR TTM_PFX - "Could not register buffer object swapout.\n"); + pr_err("Could not register buffer object swapout\n"); goto out_no_shrink; } - glob->ttm_bo_extra_size = - ttm_round_pot(sizeof(struct ttm_tt)) + - ttm_round_pot(sizeof(struct ttm_backend)); - - glob->ttm_bo_size = glob->ttm_bo_extra_size + - ttm_round_pot(sizeof(struct ttm_buffer_object)); - atomic_set(&glob->bo_count, 0); ret = kobject_init_and_add( @@ -1501,9 +1500,8 @@ int ttm_bo_device_release(struct ttm_bo_device *bdev) man->use_type = false; if ((i != TTM_PL_SYSTEM) && ttm_bo_clean_mm(bdev, i)) { ret = -EBUSY; - printk(KERN_ERR TTM_PFX - "DRM memory manager type %d " - "is not clean.\n", i); + pr_err("DRM memory manager type %d is not clean\n", + i); } man->has_type = false; } diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c index 082fcae..f8187ea 100644 --- a/drivers/gpu/drm/ttm/ttm_bo_util.c +++ b/drivers/gpu/drm/ttm/ttm_bo_util.c @@ -244,7 +244,7 @@ static int ttm_copy_io_ttm_page(struct ttm_tt *ttm, void *src, unsigned long page, pgprot_t prot) { - struct page *d = ttm_tt_get_page(ttm, page); + struct page *d = ttm->pages[page]; void *dst; if (!d) @@ -281,7 +281,7 @@ static int ttm_copy_ttm_io_page(struct ttm_tt *ttm, void *dst, unsigned long page, pgprot_t prot) { - struct page *s = ttm_tt_get_page(ttm, page); + struct page *s = ttm->pages[page]; void *src; if (!s) @@ -342,6 +342,12 @@ int ttm_bo_move_memcpy(struct ttm_buffer_object *bo, if (old_iomap == NULL && ttm == NULL) goto out2; + if (ttm->state == tt_unpopulated) { + ret = ttm->bdev->driver->ttm_tt_populate(ttm); + if (ret) + goto out1; + } + add = 0; dir = 1; @@ -439,6 +445,7 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo, kref_init(&fbo->list_kref); kref_init(&fbo->kref); fbo->destroy = &ttm_transfered_destroy; + fbo->acc_size = 0; *new_obj = fbo; return 0; @@ -502,10 +509,16 @@ static int ttm_bo_kmap_ttm(struct ttm_buffer_object *bo, { struct ttm_mem_reg *mem = &bo->mem; pgprot_t prot; struct ttm_tt *ttm = bo->ttm; - struct page *d; - int i; + int ret; BUG_ON(!ttm); + + if (ttm->state == tt_unpopulated) { + ret = ttm->bdev->driver->ttm_tt_populate(ttm); + if (ret) + return ret; + } + if (num_pages == 1 && (mem->placement & TTM_PL_FLAG_CACHED)) { /* * We're mapping a single page, and the desired @@ -513,18 +526,9 @@ static int ttm_bo_kmap_ttm(struct ttm_buffer_object *bo, */ map->bo_kmap_type = ttm_bo_map_kmap; - map->page = ttm_tt_get_page(ttm, start_page); + map->page = ttm->pages[start_page]; map->virtual = kmap(map->page); } else { - /* - * Populate the part we're mapping; - */ - for (i = start_page; i < start_page + num_pages; ++i) { - d = ttm_tt_get_page(ttm, i); - if (!d) - return -ENOMEM; - } - /* * We need to use vmap to get the desired page protection * or to make the buffer object look contiguous. diff --git a/drivers/gpu/drm/ttm/ttm_bo_vm.c b/drivers/gpu/drm/ttm/ttm_bo_vm.c index 221b924..a877813 100644 --- a/drivers/gpu/drm/ttm/ttm_bo_vm.c +++ b/drivers/gpu/drm/ttm/ttm_bo_vm.c @@ -28,6 +28,8 @@ * Authors: Thomas Hellstrom */ +#define pr_fmt(fmt) "[TTM] " fmt + #include #include #include @@ -174,18 +176,23 @@ static int ttm_bo_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf) vma->vm_page_prot = (bo->mem.placement & TTM_PL_FLAG_CACHED) ? vm_get_page_prot(vma->vm_flags) : ttm_io_prot(bo->mem.placement, vma->vm_page_prot); + + /* Allocate all page at once, most common usage */ + if (ttm->bdev->driver->ttm_tt_populate(ttm)) { + retval = VM_FAULT_OOM; + goto out_io_unlock; + } } /* * Speculatively prefault a number of pages. Only error on * first page. */ - for (i = 0; i < TTM_BO_VM_NUM_PREFAULT; ++i) { if (bo->mem.bus.is_iomem) pfn = ((bo->mem.bus.base + bo->mem.bus.offset) >> PAGE_SHIFT) + page_offset; else { - page = ttm_tt_get_page(ttm, page_offset); + page = ttm->pages[page_offset]; if (unlikely(!page && i == 0)) { retval = VM_FAULT_OOM; goto out_io_unlock; @@ -257,8 +264,7 @@ int ttm_bo_mmap(struct file *filp, struct vm_area_struct *vma, read_unlock(&bdev->vm_lock); if (unlikely(bo == NULL)) { - printk(KERN_ERR TTM_PFX - "Could not find buffer object to map.\n"); + pr_err("Could not find buffer object to map\n"); return -EINVAL; } diff --git a/drivers/gpu/drm/ttm/ttm_memory.c b/drivers/gpu/drm/ttm/ttm_memory.c index e70ddd8..23d2ecb 100644 --- a/drivers/gpu/drm/ttm/ttm_memory.c +++ b/drivers/gpu/drm/ttm/ttm_memory.c @@ -25,6 +25,8 @@ * **************************************************************************/ +#define pr_fmt(fmt) "[TTM] " fmt + #include "ttm/ttm_memory.h" #include "ttm/ttm_module.h" #include "ttm/ttm_page_alloc.h" @@ -74,9 +76,8 @@ static void ttm_mem_zone_kobj_release(struct kobject *kobj) struct ttm_mem_zone *zone = container_of(kobj, struct ttm_mem_zone, kobj); - printk(KERN_INFO TTM_PFX - "Zone %7s: Used memory at exit: %llu kiB.\n", - zone->name, (unsigned long long) zone->used_mem >> 10); + pr_info("Zone %7s: Used memory at exit: %llu kiB\n", + zone->name, (unsigned long long)zone->used_mem >> 10); kfree(zone); } @@ -390,11 +391,11 @@ int ttm_mem_global_init(struct ttm_mem_global *glob) #endif for (i = 0; i < glob->num_zones; ++i) { zone = glob->zones[i]; - printk(KERN_INFO TTM_PFX - "Zone %7s: Available graphics memory: %llu kiB.\n", - zone->name, (unsigned long long) zone->max_mem >> 10); + pr_info("Zone %7s: Available graphics memory: %llu kiB\n", + zone->name, (unsigned long long)zone->max_mem >> 10); } ttm_page_alloc_init(glob, glob->zone_kernel->max_mem/(2*PAGE_SIZE)); + ttm_dma_page_alloc_init(glob, glob->zone_kernel->max_mem/(2*PAGE_SIZE)); return 0; out_no_zone: ttm_mem_global_release(glob); @@ -409,6 +410,7 @@ void ttm_mem_global_release(struct ttm_mem_global *glob) /* let the page allocator first stop the shrink work. */ ttm_page_alloc_fini(); + ttm_dma_page_alloc_fini(); flush_workqueue(glob->swap_queue); destroy_workqueue(glob->swap_queue); diff --git a/drivers/gpu/drm/ttm/ttm_object.c b/drivers/gpu/drm/ttm/ttm_object.c index 93577f2..68daca4 100644 --- a/drivers/gpu/drm/ttm/ttm_object.c +++ b/drivers/gpu/drm/ttm/ttm_object.c @@ -49,6 +49,8 @@ * for fast lookup of ref objects given a base object. */ +#define pr_fmt(fmt) "[TTM] " fmt + #include "ttm/ttm_object.h" #include "ttm/ttm_module.h" #include @@ -232,8 +234,7 @@ struct ttm_base_object *ttm_base_object_lookup(struct ttm_object_file *tfile, return NULL; if (tfile != base->tfile && !base->shareable) { - printk(KERN_ERR TTM_PFX - "Attempted access of non-shareable object.\n"); + pr_err("Attempted access of non-shareable object\n"); ttm_base_object_unref(&base); return NULL; } diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c index 9e4313e..578207e 100644 --- a/drivers/gpu/drm/ttm/ttm_page_alloc.c +++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c @@ -30,6 +30,9 @@ * - Use page->lru to keep a free list * - doesn't track currently in use pages */ + +#define pr_fmt(fmt) "[TTM] " fmt + #include #include #include @@ -167,18 +170,13 @@ static ssize_t ttm_pool_store(struct kobject *kobj, m->options.small = val; else if (attr == &ttm_page_pool_alloc_size) { if (val > NUM_PAGES_TO_ALLOC*8) { - printk(KERN_ERR TTM_PFX - "Setting allocation size to %lu " - "is not allowed. Recommended size is " - "%lu\n", + pr_err("Setting allocation size to %lu is not allowed. Recommended size is %lu\n", NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 7), NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 10)); return size; } else if (val > NUM_PAGES_TO_ALLOC) { - printk(KERN_WARNING TTM_PFX - "Setting allocation size to " - "larger than %lu is not recommended.\n", - NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 10)); + pr_warn("Setting allocation size to larger than %lu is not recommended\n", + NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 10)); } m->options.alloc_size = val; } @@ -279,8 +277,7 @@ static void ttm_pages_put(struct page *pages[], unsigned npages) { unsigned i; if (set_pages_array_wb(pages, npages)) - printk(KERN_ERR TTM_PFX "Failed to set %d pages to wb!\n", - npages); + pr_err("Failed to set %d pages to wb!\n", npages); for (i = 0; i < npages; ++i) __free_page(pages[i]); } @@ -315,8 +312,7 @@ static int ttm_page_pool_free(struct ttm_page_pool *pool, unsigned nr_free) pages_to_free = kmalloc(npages_to_free * sizeof(struct page *), GFP_KERNEL); if (!pages_to_free) { - printk(KERN_ERR TTM_PFX - "Failed to allocate memory for pool free operation.\n"); + pr_err("Failed to allocate memory for pool free operation\n"); return 0; } @@ -438,16 +434,12 @@ static int ttm_set_pages_caching(struct page **pages, case tt_uncached: r = set_pages_array_uc(pages, cpages); if (r) - printk(KERN_ERR TTM_PFX - "Failed to set %d pages to uc!\n", - cpages); + pr_err("Failed to set %d pages to uc!\n", cpages); break; case tt_wc: r = set_pages_array_wc(pages, cpages); if (r) - printk(KERN_ERR TTM_PFX - "Failed to set %d pages to wc!\n", - cpages); + pr_err("Failed to set %d pages to wc!\n", cpages); break; default: break; @@ -492,8 +484,7 @@ static int ttm_alloc_new_pages(struct list_head *pages, gfp_t gfp_flags, caching_array = kmalloc(max_cpages*sizeof(struct page *), GFP_KERNEL); if (!caching_array) { - printk(KERN_ERR TTM_PFX - "Unable to allocate table for new pages."); + pr_err("Unable to allocate table for new pages\n"); return -ENOMEM; } @@ -501,7 +492,7 @@ static int ttm_alloc_new_pages(struct list_head *pages, gfp_t gfp_flags, p = alloc_page(gfp_flags); if (!p) { - printk(KERN_ERR TTM_PFX "Unable to get page %u.\n", i); + pr_err("Unable to get page %u\n", i); /* store already allocated pages in the pool after * setting the caching state */ @@ -599,8 +590,7 @@ static void ttm_page_pool_fill_locked(struct ttm_page_pool *pool, ++pool->nrefills; pool->npages += alloc_size; } else { - printk(KERN_ERR TTM_PFX - "Failed to fill pool (%p).", pool); + pr_err("Failed to fill pool (%p)\n", pool); /* If we have any pages left put them to the pool. */ list_for_each_entry(p, &pool->list, lru) { ++cpages; @@ -619,8 +609,10 @@ static void ttm_page_pool_fill_locked(struct ttm_page_pool *pool, * @return count of pages still required to fulfill the request. */ static unsigned ttm_page_pool_get_pages(struct ttm_page_pool *pool, - struct list_head *pages, int ttm_flags, - enum ttm_caching_state cstate, unsigned count) + struct list_head *pages, + int ttm_flags, + enum ttm_caching_state cstate, + unsigned count) { unsigned long irq_flags; struct list_head *p; @@ -660,17 +652,63 @@ out: return count; } +/* Put all pages in pages list to correct pool to wait for reuse */ +static void ttm_put_pages(struct page **pages, unsigned npages, int flags, + enum ttm_caching_state cstate) +{ + unsigned long irq_flags; + struct ttm_page_pool *pool = ttm_get_pool(flags, cstate); + unsigned i; + + if (pool == NULL) { + /* No pool for this memory type so free the pages */ + for (i = 0; i < npages; i++) { + if (pages[i]) { + if (page_count(pages[i]) != 1) + pr_err("Erroneous page count. Leaking pages.\n"); + __free_page(pages[i]); + pages[i] = NULL; + } + } + return; + } + + spin_lock_irqsave(&pool->lock, irq_flags); + for (i = 0; i < npages; i++) { + if (pages[i]) { + if (page_count(pages[i]) != 1) + pr_err("Erroneous page count. Leaking pages.\n"); + list_add_tail(&pages[i]->lru, &pool->list); + pages[i] = NULL; + pool->npages++; + } + } + /* Check that we don't go over the pool limit */ + npages = 0; + if (pool->npages > _manager->options.max_size) { + npages = pool->npages - _manager->options.max_size; + /* free at least NUM_PAGES_TO_ALLOC number of pages + * to reduce calls to set_memory_wb */ + if (npages < NUM_PAGES_TO_ALLOC) + npages = NUM_PAGES_TO_ALLOC; + } + spin_unlock_irqrestore(&pool->lock, irq_flags); + if (npages) + ttm_page_pool_free(pool, npages); +} + /* * On success pages list will hold count number of correctly * cached pages. */ -int ttm_get_pages(struct list_head *pages, int flags, - enum ttm_caching_state cstate, unsigned count, - dma_addr_t *dma_address) +static int ttm_get_pages(struct page **pages, unsigned npages, int flags, + enum ttm_caching_state cstate) { struct ttm_page_pool *pool = ttm_get_pool(flags, cstate); + struct list_head plist; struct page *p = NULL; gfp_t gfp_flags = GFP_USER; + unsigned count; int r; /* set zero flag for page allocation if required */ @@ -684,30 +722,33 @@ int ttm_get_pages(struct list_head *pages, int flags, else gfp_flags |= GFP_HIGHUSER; - for (r = 0; r < count; ++r) { + for (r = 0; r < npages; ++r) { p = alloc_page(gfp_flags); if (!p) { - printk(KERN_ERR TTM_PFX - "Unable to allocate page."); + pr_err("Unable to allocate page\n"); return -ENOMEM; } - list_add(&p->lru, pages); + pages[r] = p; } return 0; } - /* combine zero flag to pool flags */ gfp_flags |= pool->gfp_flags; /* First we take pages from the pool */ - count = ttm_page_pool_get_pages(pool, pages, flags, cstate, count); + INIT_LIST_HEAD(&plist); + npages = ttm_page_pool_get_pages(pool, &plist, flags, cstate, npages); + count = 0; + list_for_each_entry(p, &plist, lru) { + pages[count++] = p; + } /* clear the pages coming from the pool if requested */ if (flags & TTM_PAGE_FLAG_ZERO_ALLOC) { - list_for_each_entry(p, pages, lru) { + list_for_each_entry(p, &plist, lru) { if (PageHighMem(p)) clear_highpage(p); else @@ -716,67 +757,27 @@ int ttm_get_pages(struct list_head *pages, int flags, } /* If pool didn't have enough pages allocate new one. */ - if (count > 0) { + if (npages > 0) { /* ttm_alloc_new_pages doesn't reference pool so we can run * multiple requests in parallel. **/ - r = ttm_alloc_new_pages(pages, gfp_flags, flags, cstate, count); + INIT_LIST_HEAD(&plist); + r = ttm_alloc_new_pages(&plist, gfp_flags, flags, cstate, npages); + list_for_each_entry(p, &plist, lru) { + pages[count++] = p; + } if (r) { /* If there is any pages in the list put them back to * the pool. */ - printk(KERN_ERR TTM_PFX - "Failed to allocate extra pages " - "for large request."); - ttm_put_pages(pages, 0, flags, cstate, NULL); + pr_err("Failed to allocate extra pages for large request\n"); + ttm_put_pages(pages, count, flags, cstate); return r; } } - return 0; } -/* Put all pages in pages list to correct pool to wait for reuse */ -void ttm_put_pages(struct list_head *pages, unsigned page_count, int flags, - enum ttm_caching_state cstate, dma_addr_t *dma_address) -{ - unsigned long irq_flags; - struct ttm_page_pool *pool = ttm_get_pool(flags, cstate); - struct page *p, *tmp; - - if (pool == NULL) { - /* No pool for this memory type so free the pages */ - - list_for_each_entry_safe(p, tmp, pages, lru) { - __free_page(p); - } - /* Make the pages list empty */ - INIT_LIST_HEAD(pages); - return; - } - if (page_count == 0) { - list_for_each_entry_safe(p, tmp, pages, lru) { - ++page_count; - } - } - - spin_lock_irqsave(&pool->lock, irq_flags); - list_splice_init(pages, &pool->list); - pool->npages += page_count; - /* Check that we don't go over the pool limit */ - page_count = 0; - if (pool->npages > _manager->options.max_size) { - page_count = pool->npages - _manager->options.max_size; - /* free at least NUM_PAGES_TO_ALLOC number of pages - * to reduce calls to set_memory_wb */ - if (page_count < NUM_PAGES_TO_ALLOC) - page_count = NUM_PAGES_TO_ALLOC; - } - spin_unlock_irqrestore(&pool->lock, irq_flags); - if (page_count) - ttm_page_pool_free(pool, page_count); -} - static void ttm_page_pool_init_locked(struct ttm_page_pool *pool, int flags, char *name) { @@ -794,7 +795,7 @@ int ttm_page_alloc_init(struct ttm_mem_global *glob, unsigned max_pages) WARN_ON(_manager); - printk(KERN_INFO TTM_PFX "Initializing pool allocator.\n"); + pr_info("Initializing pool allocator\n"); _manager = kzalloc(sizeof(*_manager), GFP_KERNEL); @@ -829,7 +830,7 @@ void ttm_page_alloc_fini(void) { int i; - printk(KERN_INFO TTM_PFX "Finalizing pool allocator.\n"); + pr_info("Finalizing pool allocator\n"); ttm_pool_mm_shrink_fini(_manager); for (i = 0; i < NUM_POOLS; ++i) @@ -839,6 +840,62 @@ void ttm_page_alloc_fini(void) _manager = NULL; } +int ttm_pool_populate(struct ttm_tt *ttm) +{ + struct ttm_mem_global *mem_glob = ttm->glob->mem_glob; + unsigned i; + int ret; + + if (ttm->state != tt_unpopulated) + return 0; + + for (i = 0; i < ttm->num_pages; ++i) { + ret = ttm_get_pages(&ttm->pages[i], 1, + ttm->page_flags, + ttm->caching_state); + if (ret != 0) { + ttm_pool_unpopulate(ttm); + return -ENOMEM; + } + + ret = ttm_mem_global_alloc_page(mem_glob, ttm->pages[i], + false, false); + if (unlikely(ret != 0)) { + ttm_pool_unpopulate(ttm); + return -ENOMEM; + } + } + + if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { + ret = ttm_tt_swapin(ttm); + if (unlikely(ret != 0)) { + ttm_pool_unpopulate(ttm); + return ret; + } + } + + ttm->state = tt_unbound; + return 0; +} +EXPORT_SYMBOL(ttm_pool_populate); + +void ttm_pool_unpopulate(struct ttm_tt *ttm) +{ + unsigned i; + + for (i = 0; i < ttm->num_pages; ++i) { + if (ttm->pages[i]) { + ttm_mem_global_free_page(ttm->glob->mem_glob, + ttm->pages[i]); + ttm_put_pages(&ttm->pages[i], 1, + ttm->page_flags, + ttm->caching_state); + } + } + ttm->state = tt_unpopulated; +} +EXPORT_SYMBOL(ttm_pool_unpopulate); + int ttm_page_alloc_debugfs(struct seq_file *m, void *data) { struct ttm_page_pool *p; diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c new file mode 100644 index 0000000..4f9e548 --- /dev/null +++ b/drivers/gpu/drm/ttm/ttm_page_alloc_dma.c @@ -0,0 +1,1134 @@ +/* + * Copyright 2011 (c) Oracle Corp. + + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sub license, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Author: Konrad Rzeszutek Wilk + */ + +/* + * A simple DMA pool losely based on dmapool.c. It has certain advantages + * over the DMA pools: + * - Pool collects resently freed pages for reuse (and hooks up to + * the shrinker). + * - Tracks currently in use pages + * - Tracks whether the page is UC, WB or cached (and reverts to WB + * when freed). + */ + +#define pr_fmt(fmt) "[TTM] " fmt + +#include +#include +#include /* for seq_printf */ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include "ttm/ttm_bo_driver.h" +#include "ttm/ttm_page_alloc.h" +#ifdef TTM_HAS_AGP +#include +#endif + +#define NUM_PAGES_TO_ALLOC (PAGE_SIZE/sizeof(struct page *)) +#define SMALL_ALLOCATION 4 +#define FREE_ALL_PAGES (~0U) +/* times are in msecs */ +#define IS_UNDEFINED (0) +#define IS_WC (1<<1) +#define IS_UC (1<<2) +#define IS_CACHED (1<<3) +#define IS_DMA32 (1<<4) + +enum pool_type { + POOL_IS_UNDEFINED, + POOL_IS_WC = IS_WC, + POOL_IS_UC = IS_UC, + POOL_IS_CACHED = IS_CACHED, + POOL_IS_WC_DMA32 = IS_WC | IS_DMA32, + POOL_IS_UC_DMA32 = IS_UC | IS_DMA32, + POOL_IS_CACHED_DMA32 = IS_CACHED | IS_DMA32, +}; +/* + * The pool structure. There are usually six pools: + * - generic (not restricted to DMA32): + * - write combined, uncached, cached. + * - dma32 (up to 2^32 - so up 4GB): + * - write combined, uncached, cached. + * for each 'struct device'. The 'cached' is for pages that are actively used. + * The other ones can be shrunk by the shrinker API if neccessary. + * @pools: The 'struct device->dma_pools' link. + * @type: Type of the pool + * @lock: Protects the inuse_list and free_list from concurrnet access. Must be + * used with irqsave/irqrestore variants because pool allocator maybe called + * from delayed work. + * @inuse_list: Pool of pages that are in use. The order is very important and + * it is in the order that the TTM pages that are put back are in. + * @free_list: Pool of pages that are free to be used. No order requirements. + * @dev: The device that is associated with these pools. + * @size: Size used during DMA allocation. + * @npages_free: Count of available pages for re-use. + * @npages_in_use: Count of pages that are in use. + * @nfrees: Stats when pool is shrinking. + * @nrefills: Stats when the pool is grown. + * @gfp_flags: Flags to pass for alloc_page. + * @name: Name of the pool. + * @dev_name: Name derieved from dev - similar to how dev_info works. + * Used during shutdown as the dev_info during release is unavailable. + */ +struct dma_pool { + struct list_head pools; /* The 'struct device->dma_pools link */ + enum pool_type type; + spinlock_t lock; + struct list_head inuse_list; + struct list_head free_list; + struct device *dev; + unsigned size; + unsigned npages_free; + unsigned npages_in_use; + unsigned long nfrees; /* Stats when shrunk. */ + unsigned long nrefills; /* Stats when grown. */ + gfp_t gfp_flags; + char name[13]; /* "cached dma32" */ + char dev_name[64]; /* Constructed from dev */ +}; + +/* + * The accounting page keeping track of the allocated page along with + * the DMA address. + * @page_list: The link to the 'page_list' in 'struct dma_pool'. + * @vaddr: The virtual address of the page + * @dma: The bus address of the page. If the page is not allocated + * via the DMA API, it will be -1. + */ +struct dma_page { + struct list_head page_list; + void *vaddr; + struct page *p; + dma_addr_t dma; +}; + +/* + * Limits for the pool. They are handled without locks because only place where + * they may change is in sysfs store. They won't have immediate effect anyway + * so forcing serialization to access them is pointless. + */ + +struct ttm_pool_opts { + unsigned alloc_size; + unsigned max_size; + unsigned small; +}; + +/* + * Contains the list of all of the 'struct device' and their corresponding + * DMA pools. Guarded by _mutex->lock. + * @pools: The link to 'struct ttm_pool_manager->pools' + * @dev: The 'struct device' associated with the 'pool' + * @pool: The 'struct dma_pool' associated with the 'dev' + */ +struct device_pools { + struct list_head pools; + struct device *dev; + struct dma_pool *pool; +}; + +/* + * struct ttm_pool_manager - Holds memory pools for fast allocation + * + * @lock: Lock used when adding/removing from pools + * @pools: List of 'struct device' and 'struct dma_pool' tuples. + * @options: Limits for the pool. + * @npools: Total amount of pools in existence. + * @shrinker: The structure used by [un|]register_shrinker + */ +struct ttm_pool_manager { + struct mutex lock; + struct list_head pools; + struct ttm_pool_opts options; + unsigned npools; + struct shrinker mm_shrink; + struct kobject kobj; +}; + +static struct ttm_pool_manager *_manager; + +static struct attribute ttm_page_pool_max = { + .name = "pool_max_size", + .mode = S_IRUGO | S_IWUSR +}; +static struct attribute ttm_page_pool_small = { + .name = "pool_small_allocation", + .mode = S_IRUGO | S_IWUSR +}; +static struct attribute ttm_page_pool_alloc_size = { + .name = "pool_allocation_size", + .mode = S_IRUGO | S_IWUSR +}; + +static struct attribute *ttm_pool_attrs[] = { + &ttm_page_pool_max, + &ttm_page_pool_small, + &ttm_page_pool_alloc_size, + NULL +}; + +static void ttm_pool_kobj_release(struct kobject *kobj) +{ + struct ttm_pool_manager *m = + container_of(kobj, struct ttm_pool_manager, kobj); + kfree(m); +} + +static ssize_t ttm_pool_store(struct kobject *kobj, struct attribute *attr, + const char *buffer, size_t size) +{ + struct ttm_pool_manager *m = + container_of(kobj, struct ttm_pool_manager, kobj); + int chars; + unsigned val; + chars = sscanf(buffer, "%u", &val); + if (chars == 0) + return size; + + /* Convert kb to number of pages */ + val = val / (PAGE_SIZE >> 10); + + if (attr == &ttm_page_pool_max) + m->options.max_size = val; + else if (attr == &ttm_page_pool_small) + m->options.small = val; + else if (attr == &ttm_page_pool_alloc_size) { + if (val > NUM_PAGES_TO_ALLOC*8) { + pr_err("Setting allocation size to %lu is not allowed. Recommended size is %lu\n", + NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 7), + NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 10)); + return size; + } else if (val > NUM_PAGES_TO_ALLOC) { + pr_warn("Setting allocation size to larger than %lu is not recommended\n", + NUM_PAGES_TO_ALLOC*(PAGE_SIZE >> 10)); + } + m->options.alloc_size = val; + } + + return size; +} + +static ssize_t ttm_pool_show(struct kobject *kobj, struct attribute *attr, + char *buffer) +{ + struct ttm_pool_manager *m = + container_of(kobj, struct ttm_pool_manager, kobj); + unsigned val = 0; + + if (attr == &ttm_page_pool_max) + val = m->options.max_size; + else if (attr == &ttm_page_pool_small) + val = m->options.small; + else if (attr == &ttm_page_pool_alloc_size) + val = m->options.alloc_size; + + val = val * (PAGE_SIZE >> 10); + + return snprintf(buffer, PAGE_SIZE, "%u\n", val); +} + +static const struct sysfs_ops ttm_pool_sysfs_ops = { + .show = &ttm_pool_show, + .store = &ttm_pool_store, +}; + +static struct kobj_type ttm_pool_kobj_type = { + .release = &ttm_pool_kobj_release, + .sysfs_ops = &ttm_pool_sysfs_ops, + .default_attrs = ttm_pool_attrs, +}; + +#ifndef CONFIG_X86 +static int set_pages_array_wb(struct page **pages, int addrinarray) +{ +#ifdef TTM_HAS_AGP + int i; + + for (i = 0; i < addrinarray; i++) + unmap_page_from_agp(pages[i]); +#endif + return 0; +} + +static int set_pages_array_wc(struct page **pages, int addrinarray) +{ +#ifdef TTM_HAS_AGP + int i; + + for (i = 0; i < addrinarray; i++) + map_page_into_agp(pages[i]); +#endif + return 0; +} + +static int set_pages_array_uc(struct page **pages, int addrinarray) +{ +#ifdef TTM_HAS_AGP + int i; + + for (i = 0; i < addrinarray; i++) + map_page_into_agp(pages[i]); +#endif + return 0; +} +#endif /* for !CONFIG_X86 */ + +static int ttm_set_pages_caching(struct dma_pool *pool, + struct page **pages, unsigned cpages) +{ + int r = 0; + /* Set page caching */ + if (pool->type & IS_UC) { + r = set_pages_array_uc(pages, cpages); + if (r) + pr_err("%s: Failed to set %d pages to uc!\n", + pool->dev_name, cpages); + } + if (pool->type & IS_WC) { + r = set_pages_array_wc(pages, cpages); + if (r) + pr_err("%s: Failed to set %d pages to wc!\n", + pool->dev_name, cpages); + } + return r; +} + +static void __ttm_dma_free_page(struct dma_pool *pool, struct dma_page *d_page) +{ + dma_addr_t dma = d_page->dma; + dma_free_coherent(pool->dev, pool->size, d_page->vaddr, dma); + + kfree(d_page); + d_page = NULL; +} +static struct dma_page *__ttm_dma_alloc_page(struct dma_pool *pool) +{ + struct dma_page *d_page; + + d_page = kmalloc(sizeof(struct dma_page), GFP_KERNEL); + if (!d_page) + return NULL; + + d_page->vaddr = dma_alloc_coherent(pool->dev, pool->size, + &d_page->dma, + pool->gfp_flags); + if (d_page->vaddr) + d_page->p = virt_to_page(d_page->vaddr); + else { + kfree(d_page); + d_page = NULL; + } + return d_page; +} +static enum pool_type ttm_to_type(int flags, enum ttm_caching_state cstate) +{ + enum pool_type type = IS_UNDEFINED; + + if (flags & TTM_PAGE_FLAG_DMA32) + type |= IS_DMA32; + if (cstate == tt_cached) + type |= IS_CACHED; + else if (cstate == tt_uncached) + type |= IS_UC; + else + type |= IS_WC; + + return type; +} + +static void ttm_pool_update_free_locked(struct dma_pool *pool, + unsigned freed_pages) +{ + pool->npages_free -= freed_pages; + pool->nfrees += freed_pages; + +} + +/* set memory back to wb and free the pages. */ +static void ttm_dma_pages_put(struct dma_pool *pool, struct list_head *d_pages, + struct page *pages[], unsigned npages) +{ + struct dma_page *d_page, *tmp; + + /* Don't set WB on WB page pool. */ + if (npages && !(pool->type & IS_CACHED) && + set_pages_array_wb(pages, npages)) + pr_err("%s: Failed to set %d pages to wb!\n", + pool->dev_name, npages); + + list_for_each_entry_safe(d_page, tmp, d_pages, page_list) { + list_del(&d_page->page_list); + __ttm_dma_free_page(pool, d_page); + } +} + +static void ttm_dma_page_put(struct dma_pool *pool, struct dma_page *d_page) +{ + /* Don't set WB on WB page pool. */ + if (!(pool->type & IS_CACHED) && set_pages_array_wb(&d_page->p, 1)) + pr_err("%s: Failed to set %d pages to wb!\n", + pool->dev_name, 1); + + list_del(&d_page->page_list); + __ttm_dma_free_page(pool, d_page); +} + +/* + * Free pages from pool. + * + * To prevent hogging the ttm_swap process we only free NUM_PAGES_TO_ALLOC + * number of pages in one go. + * + * @pool: to free the pages from + * @nr_free: If set to true will free all pages in pool + **/ +static unsigned ttm_dma_page_pool_free(struct dma_pool *pool, unsigned nr_free) +{ + unsigned long irq_flags; + struct dma_page *dma_p, *tmp; + struct page **pages_to_free; + struct list_head d_pages; + unsigned freed_pages = 0, + npages_to_free = nr_free; + + if (NUM_PAGES_TO_ALLOC < nr_free) + npages_to_free = NUM_PAGES_TO_ALLOC; +#if 0 + if (nr_free > 1) { + pr_debug("%s: (%s:%d) Attempting to free %d (%d) pages\n", + pool->dev_name, pool->name, current->pid, + npages_to_free, nr_free); + } +#endif + pages_to_free = kmalloc(npages_to_free * sizeof(struct page *), + GFP_KERNEL); + + if (!pages_to_free) { + pr_err("%s: Failed to allocate memory for pool free operation\n", + pool->dev_name); + return 0; + } + INIT_LIST_HEAD(&d_pages); +restart: + spin_lock_irqsave(&pool->lock, irq_flags); + + /* We picking the oldest ones off the list */ + list_for_each_entry_safe_reverse(dma_p, tmp, &pool->free_list, + page_list) { + if (freed_pages >= npages_to_free) + break; + + /* Move the dma_page from one list to another. */ + list_move(&dma_p->page_list, &d_pages); + + pages_to_free[freed_pages++] = dma_p->p; + /* We can only remove NUM_PAGES_TO_ALLOC at a time. */ + if (freed_pages >= NUM_PAGES_TO_ALLOC) { + + ttm_pool_update_free_locked(pool, freed_pages); + /** + * Because changing page caching is costly + * we unlock the pool to prevent stalling. + */ + spin_unlock_irqrestore(&pool->lock, irq_flags); + + ttm_dma_pages_put(pool, &d_pages, pages_to_free, + freed_pages); + + INIT_LIST_HEAD(&d_pages); + + if (likely(nr_free != FREE_ALL_PAGES)) + nr_free -= freed_pages; + + if (NUM_PAGES_TO_ALLOC >= nr_free) + npages_to_free = nr_free; + else + npages_to_free = NUM_PAGES_TO_ALLOC; + + freed_pages = 0; + + /* free all so restart the processing */ + if (nr_free) + goto restart; + + /* Not allowed to fall through or break because + * following context is inside spinlock while we are + * outside here. + */ + goto out; + + } + } + + /* remove range of pages from the pool */ + if (freed_pages) { + ttm_pool_update_free_locked(pool, freed_pages); + nr_free -= freed_pages; + } + + spin_unlock_irqrestore(&pool->lock, irq_flags); + + if (freed_pages) + ttm_dma_pages_put(pool, &d_pages, pages_to_free, freed_pages); +out: + kfree(pages_to_free); + return nr_free; +} + +static void ttm_dma_free_pool(struct device *dev, enum pool_type type) +{ + struct device_pools *p; + struct dma_pool *pool; + + if (!dev) + return; + + mutex_lock(&_manager->lock); + list_for_each_entry_reverse(p, &_manager->pools, pools) { + if (p->dev != dev) + continue; + pool = p->pool; + if (pool->type != type) + continue; + + list_del(&p->pools); + kfree(p); + _manager->npools--; + break; + } + list_for_each_entry_reverse(pool, &dev->dma_pools, pools) { + if (pool->type != type) + continue; + /* Takes a spinlock.. */ + ttm_dma_page_pool_free(pool, FREE_ALL_PAGES); + WARN_ON(((pool->npages_in_use + pool->npages_free) != 0)); + /* This code path is called after _all_ references to the + * struct device has been dropped - so nobody should be + * touching it. In case somebody is trying to _add_ we are + * guarded by the mutex. */ + list_del(&pool->pools); + kfree(pool); + break; + } + mutex_unlock(&_manager->lock); +} + +/* + * On free-ing of the 'struct device' this deconstructor is run. + * Albeit the pool might have already been freed earlier. + */ +static void ttm_dma_pool_release(struct device *dev, void *res) +{ + struct dma_pool *pool = *(struct dma_pool **)res; + + if (pool) + ttm_dma_free_pool(dev, pool->type); +} + +static int ttm_dma_pool_match(struct device *dev, void *res, void *match_data) +{ + return *(struct dma_pool **)res == match_data; +} + +static struct dma_pool *ttm_dma_pool_init(struct device *dev, gfp_t flags, + enum pool_type type) +{ + char *n[] = {"wc", "uc", "cached", " dma32", "unknown",}; + enum pool_type t[] = {IS_WC, IS_UC, IS_CACHED, IS_DMA32, IS_UNDEFINED}; + struct device_pools *sec_pool = NULL; + struct dma_pool *pool = NULL, **ptr; + unsigned i; + int ret = -ENODEV; + char *p; + + if (!dev) + return NULL; + + ptr = devres_alloc(ttm_dma_pool_release, sizeof(*ptr), GFP_KERNEL); + if (!ptr) + return NULL; + + ret = -ENOMEM; + + pool = kmalloc_node(sizeof(struct dma_pool), GFP_KERNEL, + dev_to_node(dev)); + if (!pool) + goto err_mem; + + sec_pool = kmalloc_node(sizeof(struct device_pools), GFP_KERNEL, + dev_to_node(dev)); + if (!sec_pool) + goto err_mem; + + INIT_LIST_HEAD(&sec_pool->pools); + sec_pool->dev = dev; + sec_pool->pool = pool; + + INIT_LIST_HEAD(&pool->free_list); + INIT_LIST_HEAD(&pool->inuse_list); + INIT_LIST_HEAD(&pool->pools); + spin_lock_init(&pool->lock); + pool->dev = dev; + pool->npages_free = pool->npages_in_use = 0; + pool->nfrees = 0; + pool->gfp_flags = flags; + pool->size = PAGE_SIZE; + pool->type = type; + pool->nrefills = 0; + p = pool->name; + for (i = 0; i < 5; i++) { + if (type & t[i]) { + p += snprintf(p, sizeof(pool->name) - (p - pool->name), + "%s", n[i]); + } + } + *p = 0; + /* We copy the name for pr_ calls b/c when dma_pool_destroy is called + * - the kobj->name has already been deallocated.*/ + snprintf(pool->dev_name, sizeof(pool->dev_name), "%s %s", + dev_driver_string(dev), dev_name(dev)); + mutex_lock(&_manager->lock); + /* You can get the dma_pool from either the global: */ + list_add(&sec_pool->pools, &_manager->pools); + _manager->npools++; + /* or from 'struct device': */ + list_add(&pool->pools, &dev->dma_pools); + mutex_unlock(&_manager->lock); + + *ptr = pool; + devres_add(dev, ptr); + + return pool; +err_mem: + devres_free(ptr); + kfree(sec_pool); + kfree(pool); + return ERR_PTR(ret); +} + +static struct dma_pool *ttm_dma_find_pool(struct device *dev, + enum pool_type type) +{ + struct dma_pool *pool, *tmp, *found = NULL; + + if (type == IS_UNDEFINED) + return found; + + /* NB: We iterate on the 'struct dev' which has no spinlock, but + * it does have a kref which we have taken. The kref is taken during + * graphic driver loading - in the drm_pci_init it calls either + * pci_dev_get or pci_register_driver which both end up taking a kref + * on 'struct device'. + * + * On teardown, the graphic drivers end up quiescing the TTM (put_pages) + * and calls the dev_res deconstructors: ttm_dma_pool_release. The nice + * thing is at that point of time there are no pages associated with the + * driver so this function will not be called. + */ + list_for_each_entry_safe(pool, tmp, &dev->dma_pools, pools) { + if (pool->type != type) + continue; + found = pool; + break; + } + return found; +} + +/* + * Free pages the pages that failed to change the caching state. If there + * are pages that have changed their caching state already put them to the + * pool. + */ +static void ttm_dma_handle_caching_state_failure(struct dma_pool *pool, + struct list_head *d_pages, + struct page **failed_pages, + unsigned cpages) +{ + struct dma_page *d_page, *tmp; + struct page *p; + unsigned i = 0; + + p = failed_pages[0]; + if (!p) + return; + /* Find the failed page. */ + list_for_each_entry_safe(d_page, tmp, d_pages, page_list) { + if (d_page->p != p) + continue; + /* .. and then progress over the full list. */ + list_del(&d_page->page_list); + __ttm_dma_free_page(pool, d_page); + if (++i < cpages) + p = failed_pages[i]; + else + break; + } + +} + +/* + * Allocate 'count' pages, and put 'need' number of them on the + * 'pages' and as well on the 'dma_address' starting at 'dma_offset' offset. + * The full list of pages should also be on 'd_pages'. + * We return zero for success, and negative numbers as errors. + */ +static int ttm_dma_pool_alloc_new_pages(struct dma_pool *pool, + struct list_head *d_pages, + unsigned count) +{ + struct page **caching_array; + struct dma_page *dma_p; + struct page *p; + int r = 0; + unsigned i, cpages; + unsigned max_cpages = min(count, + (unsigned)(PAGE_SIZE/sizeof(struct page *))); + + /* allocate array for page caching change */ + caching_array = kmalloc(max_cpages*sizeof(struct page *), GFP_KERNEL); + + if (!caching_array) { + pr_err("%s: Unable to allocate table for new pages\n", + pool->dev_name); + return -ENOMEM; + } + + if (count > 1) { + pr_debug("%s: (%s:%d) Getting %d pages\n", + pool->dev_name, pool->name, current->pid, count); + } + + for (i = 0, cpages = 0; i < count; ++i) { + dma_p = __ttm_dma_alloc_page(pool); + if (!dma_p) { + pr_err("%s: Unable to get page %u\n", + pool->dev_name, i); + + /* store already allocated pages in the pool after + * setting the caching state */ + if (cpages) { + r = ttm_set_pages_caching(pool, caching_array, + cpages); + if (r) + ttm_dma_handle_caching_state_failure( + pool, d_pages, caching_array, + cpages); + } + r = -ENOMEM; + goto out; + } + p = dma_p->p; +#ifdef CONFIG_HIGHMEM + /* gfp flags of highmem page should never be dma32 so we + * we should be fine in such case + */ + if (!PageHighMem(p)) +#endif + { + caching_array[cpages++] = p; + if (cpages == max_cpages) { + /* Note: Cannot hold the spinlock */ + r = ttm_set_pages_caching(pool, caching_array, + cpages); + if (r) { + ttm_dma_handle_caching_state_failure( + pool, d_pages, caching_array, + cpages); + goto out; + } + cpages = 0; + } + } + list_add(&dma_p->page_list, d_pages); + } + + if (cpages) { + r = ttm_set_pages_caching(pool, caching_array, cpages); + if (r) + ttm_dma_handle_caching_state_failure(pool, d_pages, + caching_array, cpages); + } +out: + kfree(caching_array); + return r; +} + +/* + * @return count of pages still required to fulfill the request. + */ +static int ttm_dma_page_pool_fill_locked(struct dma_pool *pool, + unsigned long *irq_flags) +{ + unsigned count = _manager->options.small; + int r = pool->npages_free; + + if (count > pool->npages_free) { + struct list_head d_pages; + + INIT_LIST_HEAD(&d_pages); + + spin_unlock_irqrestore(&pool->lock, *irq_flags); + + /* Returns how many more are neccessary to fulfill the + * request. */ + r = ttm_dma_pool_alloc_new_pages(pool, &d_pages, count); + + spin_lock_irqsave(&pool->lock, *irq_flags); + if (!r) { + /* Add the fresh to the end.. */ + list_splice(&d_pages, &pool->free_list); + ++pool->nrefills; + pool->npages_free += count; + r = count; + } else { + struct dma_page *d_page; + unsigned cpages = 0; + + pr_err("%s: Failed to fill %s pool (r:%d)!\n", + pool->dev_name, pool->name, r); + + list_for_each_entry(d_page, &d_pages, page_list) { + cpages++; + } + list_splice_tail(&d_pages, &pool->free_list); + pool->npages_free += cpages; + r = cpages; + } + } + return r; +} + +/* + * @return count of pages still required to fulfill the request. + * The populate list is actually a stack (not that is matters as TTM + * allocates one page at a time. + */ +static int ttm_dma_pool_get_pages(struct dma_pool *pool, + struct ttm_dma_tt *ttm_dma, + unsigned index) +{ + struct dma_page *d_page; + struct ttm_tt *ttm = &ttm_dma->ttm; + unsigned long irq_flags; + int count, r = -ENOMEM; + + spin_lock_irqsave(&pool->lock, irq_flags); + count = ttm_dma_page_pool_fill_locked(pool, &irq_flags); + if (count) { + d_page = list_first_entry(&pool->free_list, struct dma_page, page_list); + ttm->pages[index] = d_page->p; + ttm_dma->dma_address[index] = d_page->dma; + list_move_tail(&d_page->page_list, &ttm_dma->pages_list); + r = 0; + pool->npages_in_use += 1; + pool->npages_free -= 1; + } + spin_unlock_irqrestore(&pool->lock, irq_flags); + return r; +} + +/* + * On success pages list will hold count number of correctly + * cached pages. On failure will hold the negative return value (-ENOMEM, etc). + */ +int ttm_dma_populate(struct ttm_dma_tt *ttm_dma, struct device *dev) +{ + struct ttm_tt *ttm = &ttm_dma->ttm; + struct ttm_mem_global *mem_glob = ttm->glob->mem_glob; + struct dma_pool *pool; + enum pool_type type; + unsigned i; + gfp_t gfp_flags; + int ret; + + if (ttm->state != tt_unpopulated) + return 0; + + type = ttm_to_type(ttm->page_flags, ttm->caching_state); + if (ttm->page_flags & TTM_PAGE_FLAG_DMA32) + gfp_flags = GFP_USER | GFP_DMA32; + else + gfp_flags = GFP_HIGHUSER; + if (ttm->page_flags & TTM_PAGE_FLAG_ZERO_ALLOC) + gfp_flags |= __GFP_ZERO; + + pool = ttm_dma_find_pool(dev, type); + if (!pool) { + pool = ttm_dma_pool_init(dev, gfp_flags, type); + if (IS_ERR_OR_NULL(pool)) { + return -ENOMEM; + } + } + + INIT_LIST_HEAD(&ttm_dma->pages_list); + for (i = 0; i < ttm->num_pages; ++i) { + ret = ttm_dma_pool_get_pages(pool, ttm_dma, i); + if (ret != 0) { + ttm_dma_unpopulate(ttm_dma, dev); + return -ENOMEM; + } + + ret = ttm_mem_global_alloc_page(mem_glob, ttm->pages[i], + false, false); + if (unlikely(ret != 0)) { + ttm_dma_unpopulate(ttm_dma, dev); + return -ENOMEM; + } + } + + if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { + ret = ttm_tt_swapin(ttm); + if (unlikely(ret != 0)) { + ttm_dma_unpopulate(ttm_dma, dev); + return ret; + } + } + + ttm->state = tt_unbound; + return 0; +} +EXPORT_SYMBOL_GPL(ttm_dma_populate); + +/* Get good estimation how many pages are free in pools */ +static int ttm_dma_pool_get_num_unused_pages(void) +{ + struct device_pools *p; + unsigned total = 0; + + mutex_lock(&_manager->lock); + list_for_each_entry(p, &_manager->pools, pools) + total += p->pool->npages_free; + mutex_unlock(&_manager->lock); + return total; +} + +/* Put all pages in pages list to correct pool to wait for reuse */ +void ttm_dma_unpopulate(struct ttm_dma_tt *ttm_dma, struct device *dev) +{ + struct ttm_tt *ttm = &ttm_dma->ttm; + struct dma_pool *pool; + struct dma_page *d_page, *next; + enum pool_type type; + bool is_cached = false; + unsigned count = 0, i, npages = 0; + unsigned long irq_flags; + + type = ttm_to_type(ttm->page_flags, ttm->caching_state); + pool = ttm_dma_find_pool(dev, type); + if (!pool) + return; + + is_cached = (ttm_dma_find_pool(pool->dev, + ttm_to_type(ttm->page_flags, tt_cached)) == pool); + + /* make sure pages array match list and count number of pages */ + list_for_each_entry(d_page, &ttm_dma->pages_list, page_list) { + ttm->pages[count] = d_page->p; + count++; + } + + spin_lock_irqsave(&pool->lock, irq_flags); + pool->npages_in_use -= count; + if (is_cached) { + pool->nfrees += count; + } else { + pool->npages_free += count; + list_splice(&ttm_dma->pages_list, &pool->free_list); + npages = count; + if (pool->npages_free > _manager->options.max_size) { + npages = pool->npages_free - _manager->options.max_size; + /* free at least NUM_PAGES_TO_ALLOC number of pages + * to reduce calls to set_memory_wb */ + if (npages < NUM_PAGES_TO_ALLOC) + npages = NUM_PAGES_TO_ALLOC; + } + } + spin_unlock_irqrestore(&pool->lock, irq_flags); + + if (is_cached) { + list_for_each_entry_safe(d_page, next, &ttm_dma->pages_list, page_list) { + ttm_mem_global_free_page(ttm->glob->mem_glob, + d_page->p); + ttm_dma_page_put(pool, d_page); + } + } else { + for (i = 0; i < count; i++) { + ttm_mem_global_free_page(ttm->glob->mem_glob, + ttm->pages[i]); + } + } + + INIT_LIST_HEAD(&ttm_dma->pages_list); + for (i = 0; i < ttm->num_pages; i++) { + ttm->pages[i] = NULL; + ttm_dma->dma_address[i] = 0; + } + + /* shrink pool if necessary (only on !is_cached pools)*/ + if (npages) + ttm_dma_page_pool_free(pool, npages); + ttm->state = tt_unpopulated; +} +EXPORT_SYMBOL_GPL(ttm_dma_unpopulate); + +/** + * Callback for mm to request pool to reduce number of page held. + */ +static int ttm_dma_pool_mm_shrink(struct shrinker *shrink, + struct shrink_control *sc) +{ + static atomic_t start_pool = ATOMIC_INIT(0); + unsigned idx = 0; + unsigned pool_offset = atomic_add_return(1, &start_pool); + unsigned shrink_pages = sc->nr_to_scan; + struct device_pools *p; + + if (list_empty(&_manager->pools)) + return 0; + + mutex_lock(&_manager->lock); + pool_offset = pool_offset % _manager->npools; + list_for_each_entry(p, &_manager->pools, pools) { + unsigned nr_free; + + if (!p->dev) + continue; + if (shrink_pages == 0) + break; + /* Do it in round-robin fashion. */ + if (++idx < pool_offset) + continue; + nr_free = shrink_pages; + shrink_pages = ttm_dma_page_pool_free(p->pool, nr_free); + pr_debug("%s: (%s:%d) Asked to shrink %d, have %d more to go\n", + p->pool->dev_name, p->pool->name, current->pid, + nr_free, shrink_pages); + } + mutex_unlock(&_manager->lock); + /* return estimated number of unused pages in pool */ + return ttm_dma_pool_get_num_unused_pages(); +} + +static void ttm_dma_pool_mm_shrink_init(struct ttm_pool_manager *manager) +{ + manager->mm_shrink.shrink = &ttm_dma_pool_mm_shrink; + manager->mm_shrink.seeks = 1; + register_shrinker(&manager->mm_shrink); +} + +static void ttm_dma_pool_mm_shrink_fini(struct ttm_pool_manager *manager) +{ + unregister_shrinker(&manager->mm_shrink); +} + +int ttm_dma_page_alloc_init(struct ttm_mem_global *glob, unsigned max_pages) +{ + int ret = -ENOMEM; + + WARN_ON(_manager); + + pr_info("Initializing DMA pool allocator\n"); + + _manager = kzalloc(sizeof(*_manager), GFP_KERNEL); + if (!_manager) + goto err_manager; + + mutex_init(&_manager->lock); + INIT_LIST_HEAD(&_manager->pools); + + _manager->options.max_size = max_pages; + _manager->options.small = SMALL_ALLOCATION; + _manager->options.alloc_size = NUM_PAGES_TO_ALLOC; + + /* This takes care of auto-freeing the _manager */ + ret = kobject_init_and_add(&_manager->kobj, &ttm_pool_kobj_type, + &glob->kobj, "dma_pool"); + if (unlikely(ret != 0)) { + kobject_put(&_manager->kobj); + goto err; + } + ttm_dma_pool_mm_shrink_init(_manager); + return 0; +err_manager: + kfree(_manager); + _manager = NULL; +err: + return ret; +} + +void ttm_dma_page_alloc_fini(void) +{ + struct device_pools *p, *t; + + pr_info("Finalizing DMA pool allocator\n"); + ttm_dma_pool_mm_shrink_fini(_manager); + + list_for_each_entry_safe_reverse(p, t, &_manager->pools, pools) { + dev_dbg(p->dev, "(%s:%d) Freeing.\n", p->pool->name, + current->pid); + WARN_ON(devres_destroy(p->dev, ttm_dma_pool_release, + ttm_dma_pool_match, p->pool)); + ttm_dma_free_pool(p->dev, p->pool->type); + } + kobject_put(&_manager->kobj); + _manager = NULL; +} + +int ttm_dma_page_alloc_debugfs(struct seq_file *m, void *data) +{ + struct device_pools *p; + struct dma_pool *pool = NULL; + char *h[] = {"pool", "refills", "pages freed", "inuse", "available", + "name", "virt", "busaddr"}; + + if (!_manager) { + seq_printf(m, "No pool allocator running.\n"); + return 0; + } + seq_printf(m, "%13s %12s %13s %8s %8s %8s\n", + h[0], h[1], h[2], h[3], h[4], h[5]); + mutex_lock(&_manager->lock); + list_for_each_entry(p, &_manager->pools, pools) { + struct device *dev = p->dev; + if (!dev) + continue; + pool = p->pool; + seq_printf(m, "%13s %12ld %13ld %8d %8d %8s\n", + pool->name, pool->nrefills, + pool->nfrees, pool->npages_in_use, + pool->npages_free, + pool->dev_name); + } + mutex_unlock(&_manager->lock); + return 0; +} +EXPORT_SYMBOL_GPL(ttm_dma_page_alloc_debugfs); diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c index f9cc548..fa09daf 100644 --- a/drivers/gpu/drm/ttm/ttm_tt.c +++ b/drivers/gpu/drm/ttm/ttm_tt.c @@ -28,6 +28,8 @@ * Authors: Thomas Hellstrom */ +#define pr_fmt(fmt) "[TTM] " fmt + #include #include #include @@ -43,139 +45,20 @@ #include "ttm/ttm_placement.h" #include "ttm/ttm_page_alloc.h" -static int ttm_tt_swapin(struct ttm_tt *ttm); - /** * Allocates storage for pointers to the pages that back the ttm. */ static void ttm_tt_alloc_page_directory(struct ttm_tt *ttm) { - ttm->pages = drm_calloc_large(ttm->num_pages, sizeof(*ttm->pages)); - ttm->dma_address = drm_calloc_large(ttm->num_pages, - sizeof(*ttm->dma_address)); -} - -static void ttm_tt_free_page_directory(struct ttm_tt *ttm) -{ - drm_free_large(ttm->pages); - ttm->pages = NULL; - drm_free_large(ttm->dma_address); - ttm->dma_address = NULL; -} - -static void ttm_tt_free_user_pages(struct ttm_tt *ttm) -{ - int write; - int dirty; - struct page *page; - int i; - struct ttm_backend *be = ttm->be; - - BUG_ON(!(ttm->page_flags & TTM_PAGE_FLAG_USER)); - write = ((ttm->page_flags & TTM_PAGE_FLAG_WRITE) != 0); - dirty = ((ttm->page_flags & TTM_PAGE_FLAG_USER_DIRTY) != 0); - - if (be) - be->func->clear(be); - - for (i = 0; i < ttm->num_pages; ++i) { - page = ttm->pages[i]; - if (page == NULL) - continue; - - if (page == ttm->dummy_read_page) { - BUG_ON(write); - continue; - } - - if (write && dirty && !PageReserved(page)) - set_page_dirty_lock(page); - - ttm->pages[i] = NULL; - ttm_mem_global_free(ttm->glob->mem_glob, PAGE_SIZE); - put_page(page); - } - ttm->state = tt_unpopulated; - ttm->first_himem_page = ttm->num_pages; - ttm->last_lomem_page = -1; + ttm->pages = drm_calloc_large(ttm->num_pages, sizeof(void*)); } -static struct page *__ttm_tt_get_page(struct ttm_tt *ttm, int index) +static void ttm_dma_tt_alloc_page_directory(struct ttm_dma_tt *ttm) { - struct page *p; - struct list_head h; - struct ttm_mem_global *mem_glob = ttm->glob->mem_glob; - int ret; - - while (NULL == (p = ttm->pages[index])) { - - INIT_LIST_HEAD(&h); - - ret = ttm_get_pages(&h, ttm->page_flags, ttm->caching_state, 1, - &ttm->dma_address[index]); - - if (ret != 0) - return NULL; - - p = list_first_entry(&h, struct page, lru); - - ret = ttm_mem_global_alloc_page(mem_glob, p, false, false); - if (unlikely(ret != 0)) - goto out_err; - - if (PageHighMem(p)) - ttm->pages[--ttm->first_himem_page] = p; - else - ttm->pages[++ttm->last_lomem_page] = p; - } - return p; -out_err: - put_page(p); - return NULL; -} - -struct page *ttm_tt_get_page(struct ttm_tt *ttm, int index) -{ - int ret; - - if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { - ret = ttm_tt_swapin(ttm); - if (unlikely(ret != 0)) - return NULL; - } - return __ttm_tt_get_page(ttm, index); -} - -int ttm_tt_populate(struct ttm_tt *ttm) -{ - struct page *page; - unsigned long i; - struct ttm_backend *be; - int ret; - - if (ttm->state != tt_unpopulated) - return 0; - - if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { - ret = ttm_tt_swapin(ttm); - if (unlikely(ret != 0)) - return ret; - } - - be = ttm->be; - - for (i = 0; i < ttm->num_pages; ++i) { - page = __ttm_tt_get_page(ttm, i); - if (!page) - return -ENOMEM; - } - - be->func->populate(be, ttm->num_pages, ttm->pages, - ttm->dummy_read_page, ttm->dma_address); - ttm->state = tt_unbound; - return 0; + ttm->ttm.pages = drm_calloc_large(ttm->ttm.num_pages, sizeof(void*)); + ttm->dma_address = drm_calloc_large(ttm->ttm.num_pages, + sizeof(*ttm->dma_address)); } -EXPORT_SYMBOL(ttm_tt_populate); #ifdef CONFIG_X86 static inline int ttm_tt_set_page_caching(struct page *p, @@ -278,153 +161,100 @@ int ttm_tt_set_placement_caching(struct ttm_tt *ttm, uint32_t placement) } EXPORT_SYMBOL(ttm_tt_set_placement_caching); -static void ttm_tt_free_alloced_pages(struct ttm_tt *ttm) -{ - int i; - unsigned count = 0; - struct list_head h; - struct page *cur_page; - struct ttm_backend *be = ttm->be; - - INIT_LIST_HEAD(&h); - - if (be) - be->func->clear(be); - for (i = 0; i < ttm->num_pages; ++i) { - - cur_page = ttm->pages[i]; - ttm->pages[i] = NULL; - if (cur_page) { - if (page_count(cur_page) != 1) - printk(KERN_ERR TTM_PFX - "Erroneous page count. " - "Leaking pages.\n"); - ttm_mem_global_free_page(ttm->glob->mem_glob, - cur_page); - list_add(&cur_page->lru, &h); - count++; - } - } - ttm_put_pages(&h, count, ttm->page_flags, ttm->caching_state, - ttm->dma_address); - ttm->state = tt_unpopulated; - ttm->first_himem_page = ttm->num_pages; - ttm->last_lomem_page = -1; -} - void ttm_tt_destroy(struct ttm_tt *ttm) { - struct ttm_backend *be; - if (unlikely(ttm == NULL)) return; - be = ttm->be; - if (likely(be != NULL)) { - be->func->destroy(be); - ttm->be = NULL; + if (ttm->state == tt_bound) { + ttm_tt_unbind(ttm); } if (likely(ttm->pages != NULL)) { - if (ttm->page_flags & TTM_PAGE_FLAG_USER) - ttm_tt_free_user_pages(ttm); - else - ttm_tt_free_alloced_pages(ttm); - - ttm_tt_free_page_directory(ttm); + ttm->bdev->driver->ttm_tt_unpopulate(ttm); } if (!(ttm->page_flags & TTM_PAGE_FLAG_PERSISTENT_SWAP) && ttm->swap_storage) fput(ttm->swap_storage); - kfree(ttm); + ttm->swap_storage = NULL; + ttm->func->destroy(ttm); } -int ttm_tt_set_user(struct ttm_tt *ttm, - struct task_struct *tsk, - unsigned long start, unsigned long num_pages) +int ttm_tt_init(struct ttm_tt *ttm, struct ttm_bo_device *bdev, + unsigned long size, uint32_t page_flags, + struct page *dummy_read_page) { - struct mm_struct *mm = tsk->mm; - int ret; - int write = (ttm->page_flags & TTM_PAGE_FLAG_WRITE) != 0; - struct ttm_mem_global *mem_glob = ttm->glob->mem_glob; - - BUG_ON(num_pages != ttm->num_pages); - BUG_ON((ttm->page_flags & TTM_PAGE_FLAG_USER) == 0); - - /** - * Account user pages as lowmem pages for now. - */ - - ret = ttm_mem_global_alloc(mem_glob, num_pages * PAGE_SIZE, - false, false); - if (unlikely(ret != 0)) - return ret; - - down_read(&mm->mmap_sem); - ret = get_user_pages(tsk, mm, start, num_pages, - write, 0, ttm->pages, NULL); - up_read(&mm->mmap_sem); + ttm->bdev = bdev; + ttm->glob = bdev->glob; + ttm->num_pages = (size + PAGE_SIZE - 1) >> PAGE_SHIFT; + ttm->caching_state = tt_cached; + ttm->page_flags = page_flags; + ttm->dummy_read_page = dummy_read_page; + ttm->state = tt_unpopulated; + ttm->swap_storage = NULL; - if (ret != num_pages && write) { - ttm_tt_free_user_pages(ttm); - ttm_mem_global_free(mem_glob, num_pages * PAGE_SIZE); + ttm_tt_alloc_page_directory(ttm); + if (!ttm->pages) { + ttm_tt_destroy(ttm); + pr_err("Failed allocating page table\n"); return -ENOMEM; } - - ttm->tsk = tsk; - ttm->start = start; - ttm->state = tt_unbound; - return 0; } +EXPORT_SYMBOL(ttm_tt_init); -struct ttm_tt *ttm_tt_create(struct ttm_bo_device *bdev, unsigned long size, - uint32_t page_flags, struct page *dummy_read_page) +void ttm_tt_fini(struct ttm_tt *ttm) { - struct ttm_bo_driver *bo_driver = bdev->driver; - struct ttm_tt *ttm; - - if (!bo_driver) - return NULL; + drm_free_large(ttm->pages); + ttm->pages = NULL; +} +EXPORT_SYMBOL(ttm_tt_fini); - ttm = kzalloc(sizeof(*ttm), GFP_KERNEL); - if (!ttm) - return NULL; +int ttm_dma_tt_init(struct ttm_dma_tt *ttm_dma, struct ttm_bo_device *bdev, + unsigned long size, uint32_t page_flags, + struct page *dummy_read_page) +{ + struct ttm_tt *ttm = &ttm_dma->ttm; + ttm->bdev = bdev; ttm->glob = bdev->glob; ttm->num_pages = (size + PAGE_SIZE - 1) >> PAGE_SHIFT; - ttm->first_himem_page = ttm->num_pages; - ttm->last_lomem_page = -1; ttm->caching_state = tt_cached; ttm->page_flags = page_flags; - ttm->dummy_read_page = dummy_read_page; + ttm->state = tt_unpopulated; + ttm->swap_storage = NULL; - ttm_tt_alloc_page_directory(ttm); - if (!ttm->pages) { - ttm_tt_destroy(ttm); - printk(KERN_ERR TTM_PFX "Failed allocating page table\n"); - return NULL; - } - ttm->be = bo_driver->create_ttm_backend_entry(bdev); - if (!ttm->be) { + INIT_LIST_HEAD(&ttm_dma->pages_list); + ttm_dma_tt_alloc_page_directory(ttm_dma); + if (!ttm->pages || !ttm_dma->dma_address) { ttm_tt_destroy(ttm); - printk(KERN_ERR TTM_PFX "Failed creating ttm backend entry\n"); - return NULL; + pr_err("Failed allocating page table\n"); + return -ENOMEM; } - ttm->state = tt_unpopulated; - return ttm; + return 0; } +EXPORT_SYMBOL(ttm_dma_tt_init); + +void ttm_dma_tt_fini(struct ttm_dma_tt *ttm_dma) +{ + struct ttm_tt *ttm = &ttm_dma->ttm; + + drm_free_large(ttm->pages); + ttm->pages = NULL; + drm_free_large(ttm_dma->dma_address); + ttm_dma->dma_address = NULL; +} +EXPORT_SYMBOL(ttm_dma_tt_fini); void ttm_tt_unbind(struct ttm_tt *ttm) { int ret; - struct ttm_backend *be = ttm->be; if (ttm->state == tt_bound) { - ret = be->func->unbind(be); + ret = ttm->func->unbind(ttm); BUG_ON(ret); ttm->state = tt_unbound; } @@ -433,7 +263,6 @@ void ttm_tt_unbind(struct ttm_tt *ttm) int ttm_tt_bind(struct ttm_tt *ttm, struct ttm_mem_reg *bo_mem) { int ret = 0; - struct ttm_backend *be; if (!ttm) return -EINVAL; @@ -441,25 +270,21 @@ int ttm_tt_bind(struct ttm_tt *ttm, struct ttm_mem_reg *bo_mem) if (ttm->state == tt_bound) return 0; - be = ttm->be; - - ret = ttm_tt_populate(ttm); + ret = ttm->bdev->driver->ttm_tt_populate(ttm); if (ret) return ret; - ret = be->func->bind(be, bo_mem); + ret = ttm->func->bind(ttm, bo_mem); if (unlikely(ret != 0)) return ret; ttm->state = tt_bound; - if (ttm->page_flags & TTM_PAGE_FLAG_USER) - ttm->page_flags |= TTM_PAGE_FLAG_USER_DIRTY; return 0; } EXPORT_SYMBOL(ttm_tt_bind); -static int ttm_tt_swapin(struct ttm_tt *ttm) +int ttm_tt_swapin(struct ttm_tt *ttm) { struct address_space *swap_space; struct file *swap_storage; @@ -470,16 +295,6 @@ static int ttm_tt_swapin(struct ttm_tt *ttm) int i; int ret = -ENOMEM; - if (ttm->page_flags & TTM_PAGE_FLAG_USER) { - ret = ttm_tt_set_user(ttm, ttm->tsk, ttm->start, - ttm->num_pages); - if (unlikely(ret != 0)) - return ret; - - ttm->page_flags &= ~TTM_PAGE_FLAG_SWAPPED; - return 0; - } - swap_storage = ttm->swap_storage; BUG_ON(swap_storage == NULL); @@ -491,16 +306,16 @@ static int ttm_tt_swapin(struct ttm_tt *ttm) ret = PTR_ERR(from_page); goto out_err; } - to_page = __ttm_tt_get_page(ttm, i); + to_page = ttm->pages[i]; if (unlikely(to_page == NULL)) goto out_err; preempt_disable(); - from_virtual = kmap_atomic(from_page, KM_USER0); - to_virtual = kmap_atomic(to_page, KM_USER1); + from_virtual = kmap_atomic(from_page); + to_virtual = kmap_atomic(to_page); memcpy(to_virtual, from_virtual, PAGE_SIZE); - kunmap_atomic(to_virtual, KM_USER1); - kunmap_atomic(from_virtual, KM_USER0); + kunmap_atomic(to_virtual); + kunmap_atomic(from_virtual); preempt_enable(); page_cache_release(from_page); } @@ -512,7 +327,6 @@ static int ttm_tt_swapin(struct ttm_tt *ttm) return 0; out_err: - ttm_tt_free_alloced_pages(ttm); return ret; } @@ -530,24 +344,12 @@ int ttm_tt_swapout(struct ttm_tt *ttm, struct file *persistent_swap_storage) BUG_ON(ttm->state != tt_unbound && ttm->state != tt_unpopulated); BUG_ON(ttm->caching_state != tt_cached); - /* - * For user buffers, just unpin the pages, as there should be - * vma references. - */ - - if (ttm->page_flags & TTM_PAGE_FLAG_USER) { - ttm_tt_free_user_pages(ttm); - ttm->page_flags |= TTM_PAGE_FLAG_SWAPPED; - ttm->swap_storage = NULL; - return 0; - } - if (!persistent_swap_storage) { swap_storage = shmem_file_setup("ttm swap", ttm->num_pages << PAGE_SHIFT, 0); if (unlikely(IS_ERR(swap_storage))) { - printk(KERN_ERR "Failed allocating swap storage.\n"); + pr_err("Failed allocating swap storage\n"); return PTR_ERR(swap_storage); } } else @@ -565,18 +367,18 @@ int ttm_tt_swapout(struct ttm_tt *ttm, struct file *persistent_swap_storage) goto out_err; } preempt_disable(); - from_virtual = kmap_atomic(from_page, KM_USER0); - to_virtual = kmap_atomic(to_page, KM_USER1); + from_virtual = kmap_atomic(from_page); + to_virtual = kmap_atomic(to_page); memcpy(to_virtual, from_virtual, PAGE_SIZE); - kunmap_atomic(to_virtual, KM_USER1); - kunmap_atomic(from_virtual, KM_USER0); + kunmap_atomic(to_virtual); + kunmap_atomic(from_virtual); preempt_enable(); set_page_dirty(to_page); mark_page_accessed(to_page); page_cache_release(to_page); } - ttm_tt_free_alloced_pages(ttm); + ttm->bdev->driver->ttm_tt_unpopulate(ttm); ttm->swap_storage = swap_storage; ttm->page_flags |= TTM_PAGE_FLAG_SWAPPED; if (persistent_swap_storage) diff --git a/drivers/gpu/drm/udl/Kconfig b/drivers/gpu/drm/udl/Kconfig new file mode 100644 index 0000000..0b5e096 --- /dev/null +++ b/drivers/gpu/drm/udl/Kconfig @@ -0,0 +1,12 @@ +config DRM_UDL + tristate "DisplayLink" + depends on DRM && EXPERIMENTAL + select DRM_USB + select FB_SYS_FILLRECT + select FB_SYS_COPYAREA + select FB_SYS_IMAGEBLIT + select FB_DEFERRED_IO + select DRM_KMS_HELPER + help + This is a KMS driver for the USB displaylink video adapters. + Say M/Y to add support for these devices via drm/kms interfaces. diff --git a/drivers/gpu/drm/udl/Makefile b/drivers/gpu/drm/udl/Makefile new file mode 100644 index 0000000..05c7481 --- /dev/null +++ b/drivers/gpu/drm/udl/Makefile @@ -0,0 +1,6 @@ + +ccflags-y := -Iinclude/drm + +udl-y := udl_drv.o udl_modeset.o udl_connector.o udl_encoder.o udl_main.o udl_fb.o udl_transfer.o udl_gem.o + +obj-$(CONFIG_DRM_UDL) := udl.o diff --git a/drivers/gpu/drm/udl/udl_connector.c b/drivers/gpu/drm/udl/udl_connector.c new file mode 100644 index 0000000..3234224 --- /dev/null +++ b/drivers/gpu/drm/udl/udl_connector.c @@ -0,0 +1,161 @@ +/* + * Copyright (C) 2012 Red Hat + * based in parts on udlfb.c: + * Copyright (C) 2009 Roberto De Ioris + * Copyright (C) 2009 Jaya Kumar + * Copyright (C) 2009 Bernie Thompson + * + * This file is subject to the terms and conditions of the GNU General Public + * License v2. See the file COPYING in the main directory of this archive for + * more details. + */ + +#include "drmP.h" +#include "drm_crtc.h" +#include "drm_edid.h" +#include "drm_crtc_helper.h" +#include "udl_drv.h" + +/* dummy connector to just get EDID, + all UDL appear to have a DVI-D */ + +static u8 *udl_get_edid(struct udl_device *udl) +{ + u8 *block; + char *rbuf; + int ret, i; + + block = kmalloc(EDID_LENGTH, GFP_KERNEL); + if (block == NULL) + return NULL; + + rbuf = kmalloc(2, GFP_KERNEL); + if (rbuf == NULL) + goto error; + + for (i = 0; i < EDID_LENGTH; i++) { + ret = usb_control_msg(udl->ddev->usbdev, + usb_rcvctrlpipe(udl->ddev->usbdev, 0), (0x02), + (0x80 | (0x02 << 5)), i << 8, 0xA1, rbuf, 2, + HZ); + if (ret < 1) { + DRM_ERROR("Read EDID byte %d failed err %x\n", i, ret); + goto error; + } + block[i] = rbuf[1]; + } + + kfree(rbuf); + return block; + +error: + kfree(block); + kfree(rbuf); + return NULL; +} + +static int udl_get_modes(struct drm_connector *connector) +{ + struct udl_device *udl = connector->dev->dev_private; + struct edid *edid; + int ret; + + edid = (struct edid *)udl_get_edid(udl); + + connector->display_info.raw_edid = (char *)edid; + + /* + * We only read the main block, but if the monitor reports extension + * blocks then the drm edid code expects them to be present, so patch + * the extension count to 0. + */ + edid->checksum += edid->extensions; + edid->extensions = 0; + + drm_mode_connector_update_edid_property(connector, edid); + ret = drm_add_edid_modes(connector, edid); + connector->display_info.raw_edid = NULL; + kfree(edid); + return ret; +} + +static int udl_mode_valid(struct drm_connector *connector, + struct drm_display_mode *mode) +{ + struct udl_device *udl = connector->dev->dev_private; + if (!udl->sku_pixel_limit) + return 0; + + if (mode->vdisplay * mode->hdisplay > udl->sku_pixel_limit) + return MODE_VIRTUAL_Y; + + return 0; +} + +static enum drm_connector_status +udl_detect(struct drm_connector *connector, bool force) +{ + if (drm_device_is_unplugged(connector->dev)) + return connector_status_disconnected; + return connector_status_connected; +} + +struct drm_encoder *udl_best_single_encoder(struct drm_connector *connector) +{ + int enc_id = connector->encoder_ids[0]; + struct drm_mode_object *obj; + struct drm_encoder *encoder; + + obj = drm_mode_object_find(connector->dev, enc_id, DRM_MODE_OBJECT_ENCODER); + if (!obj) + return NULL; + encoder = obj_to_encoder(obj); + return encoder; +} + +int udl_connector_set_property(struct drm_connector *connector, struct drm_property *property, + uint64_t val) +{ + return 0; +} + +static void udl_connector_destroy(struct drm_connector *connector) +{ + drm_sysfs_connector_remove(connector); + drm_connector_cleanup(connector); + kfree(connector); +} + +struct drm_connector_helper_funcs udl_connector_helper_funcs = { + .get_modes = udl_get_modes, + .mode_valid = udl_mode_valid, + .best_encoder = udl_best_single_encoder, +}; + +struct drm_connector_funcs udl_connector_funcs = { + .dpms = drm_helper_connector_dpms, + .detect = udl_detect, + .fill_modes = drm_helper_probe_single_connector_modes, + .destroy = udl_connector_destroy, + .set_property = udl_connector_set_property, +}; + +int udl_connector_init(struct drm_device *dev, struct drm_encoder *encoder) +{ + struct drm_connector *connector; + + connector = kzalloc(sizeof(struct drm_connector), GFP_KERNEL); + if (!connector) + return -ENOMEM; + + drm_connector_init(dev, connector, &udl_connector_funcs, DRM_MODE_CONNECTOR_DVII); + drm_connector_helper_add(connector, &udl_connector_helper_funcs); + + drm_sysfs_connector_add(connector); + drm_mode_connector_attach_encoder(connector, encoder); + + drm_connector_attach_property(connector, + dev->mode_config.dirty_info_property, + 1); + return 0; +} diff --git a/drivers/gpu/drm/udl/udl_drv.c b/drivers/gpu/drm/udl/udl_drv.c new file mode 100644 index 0000000..08eff0d --- /dev/null +++ b/drivers/gpu/drm/udl/udl_drv.c @@ -0,0 +1,112 @@ +/* + * Copyright (C) 2012 Red Hat + * + * This file is subject to the terms and conditions of the GNU General Public + * License v2. See the file COPYING in the main directory of this archive for + * more details. + */ + +#include +#include "drm_usb.h" +#include "drm_crtc_helper.h" +#include "udl_drv.h" + +static struct drm_driver driver; + +/* + * There are many DisplayLink-based graphics products, all with unique PIDs. + * So we match on DisplayLink's VID + Vendor-Defined Interface Class (0xff) + * We also require a match on SubClass (0x00) and Protocol (0x00), + * which is compatible with all known USB 2.0 era graphics chips and firmware, + * but allows DisplayLink to increment those for any future incompatible chips + */ +static struct usb_device_id id_table[] = { + {.idVendor = 0x17e9, .bInterfaceClass = 0xff, + .bInterfaceSubClass = 0x00, + .bInterfaceProtocol = 0x00, + .match_flags = USB_DEVICE_ID_MATCH_VENDOR | + USB_DEVICE_ID_MATCH_INT_CLASS | + USB_DEVICE_ID_MATCH_INT_SUBCLASS | + USB_DEVICE_ID_MATCH_INT_PROTOCOL,}, + {}, +}; +MODULE_DEVICE_TABLE(usb, id_table); + +MODULE_LICENSE("GPL"); + +static int udl_usb_probe(struct usb_interface *interface, + const struct usb_device_id *id) +{ + return drm_get_usb_dev(interface, id, &driver); +} + +static void udl_usb_disconnect(struct usb_interface *interface) +{ + struct drm_device *dev = usb_get_intfdata(interface); + + drm_kms_helper_poll_disable(dev); + drm_connector_unplug_all(dev); + udl_fbdev_unplug(dev); + udl_drop_usb(dev); + drm_unplug_dev(dev); +} + +static struct vm_operations_struct udl_gem_vm_ops = { + .fault = udl_gem_fault, + .open = drm_gem_vm_open, + .close = drm_gem_vm_close, +}; + +static const struct file_operations udl_driver_fops = { + .owner = THIS_MODULE, + .open = drm_open, + .mmap = udl_drm_gem_mmap, + .poll = drm_poll, + .read = drm_read, + .unlocked_ioctl = drm_ioctl, + .release = drm_release, + .fasync = drm_fasync, + .llseek = noop_llseek, +}; + +static struct drm_driver driver = { + .driver_features = DRIVER_MODESET | DRIVER_GEM, + .load = udl_driver_load, + .unload = udl_driver_unload, + + /* gem hooks */ + .gem_init_object = udl_gem_init_object, + .gem_free_object = udl_gem_free_object, + .gem_vm_ops = &udl_gem_vm_ops, + + .dumb_create = udl_dumb_create, + .dumb_map_offset = udl_gem_mmap, + .dumb_destroy = udl_dumb_destroy, + .fops = &udl_driver_fops, + .name = DRIVER_NAME, + .desc = DRIVER_DESC, + .date = DRIVER_DATE, + .major = DRIVER_MAJOR, + .minor = DRIVER_MINOR, + .patchlevel = DRIVER_PATCHLEVEL, +}; + +static struct usb_driver udl_driver = { + .name = "udl", + .probe = udl_usb_probe, + .disconnect = udl_usb_disconnect, + .id_table = id_table, +}; + +static int __init udl_init(void) +{ + return drm_usb_init(&driver, &udl_driver); +} + +static void __exit udl_exit(void) +{ + drm_usb_exit(&driver, &udl_driver); +} + +module_init(udl_init); +module_exit(udl_exit); diff --git a/drivers/gpu/drm/udl/udl_drv.h b/drivers/gpu/drm/udl/udl_drv.h new file mode 100644 index 0000000..e760575 --- /dev/null +++ b/drivers/gpu/drm/udl/udl_drv.h @@ -0,0 +1,142 @@ +/* + * Copyright (C) 2012 Red Hat + * + * based in parts on udlfb.c: + * Copyright (C) 2009 Roberto De Ioris + * Copyright (C) 2009 Jaya Kumar + * Copyright (C) 2009 Bernie Thompson + * + * This file is subject to the terms and conditions of the GNU General Public + * License v2. See the file COPYING in the main directory of this archive for + * more details. + */ + +#ifndef UDL_DRV_H +#define UDL_DRV_H + +#include + +#define DRIVER_NAME "udl" +#define DRIVER_DESC "DisplayLink" +#define DRIVER_DATE "20120220" + +#define DRIVER_MAJOR 0 +#define DRIVER_MINOR 0 +#define DRIVER_PATCHLEVEL 1 + +struct udl_device; + +struct urb_node { + struct list_head entry; + struct udl_device *dev; + struct delayed_work release_urb_work; + struct urb *urb; +}; + +struct urb_list { + struct list_head list; + spinlock_t lock; + struct semaphore limit_sem; + int available; + int count; + size_t size; +}; + +struct udl_fbdev; + +struct udl_device { + struct device *dev; + struct drm_device *ddev; + + int sku_pixel_limit; + + struct urb_list urbs; + atomic_t lost_pixels; /* 1 = a render op failed. Need screen refresh */ + + struct udl_fbdev *fbdev; + char mode_buf[1024]; + uint32_t mode_buf_len; + atomic_t bytes_rendered; /* raw pixel-bytes driver asked to render */ + atomic_t bytes_identical; /* saved effort with backbuffer comparison */ + atomic_t bytes_sent; /* to usb, after compression including overhead */ + atomic_t cpu_kcycles_used; /* transpired during pixel processing */ +}; + +struct udl_gem_object { + struct drm_gem_object base; + struct page **pages; + void *vmapping; +}; + +#define to_udl_bo(x) container_of(x, struct udl_gem_object, base) + +struct udl_framebuffer { + struct drm_framebuffer base; + struct udl_gem_object *obj; + bool active_16; /* active on the 16-bit channel */ +}; + +#define to_udl_fb(x) container_of(x, struct udl_framebuffer, base) + +/* modeset */ +int udl_modeset_init(struct drm_device *dev); +void udl_modeset_cleanup(struct drm_device *dev); +int udl_connector_init(struct drm_device *dev, struct drm_encoder *encoder); + +struct drm_encoder *udl_encoder_init(struct drm_device *dev); + +struct urb *udl_get_urb(struct drm_device *dev); + +int udl_submit_urb(struct drm_device *dev, struct urb *urb, size_t len); +void udl_urb_completion(struct urb *urb); + +int udl_driver_load(struct drm_device *dev, unsigned long flags); +int udl_driver_unload(struct drm_device *dev); + +int udl_fbdev_init(struct drm_device *dev); +void udl_fbdev_cleanup(struct drm_device *dev); +void udl_fbdev_unplug(struct drm_device *dev); +struct drm_framebuffer * +udl_fb_user_fb_create(struct drm_device *dev, + struct drm_file *file, + struct drm_mode_fb_cmd2 *mode_cmd); + +int udl_render_hline(struct drm_device *dev, int bpp, struct urb **urb_ptr, + const char *front, char **urb_buf_ptr, + u32 byte_offset, u32 device_byte_offset, u32 byte_width, + int *ident_ptr, int *sent_ptr); + +int udl_dumb_create(struct drm_file *file_priv, + struct drm_device *dev, + struct drm_mode_create_dumb *args); +int udl_gem_mmap(struct drm_file *file_priv, struct drm_device *dev, + uint32_t handle, uint64_t *offset); +int udl_dumb_destroy(struct drm_file *file_priv, struct drm_device *dev, + uint32_t handle); + +int udl_gem_init_object(struct drm_gem_object *obj); +void udl_gem_free_object(struct drm_gem_object *gem_obj); +struct udl_gem_object *udl_gem_alloc_object(struct drm_device *dev, + size_t size); + +int udl_gem_vmap(struct udl_gem_object *obj); +void udl_gem_vunmap(struct udl_gem_object *obj); +int udl_drm_gem_mmap(struct file *filp, struct vm_area_struct *vma); +int udl_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf); + +int udl_handle_damage(struct udl_framebuffer *fb, int x, int y, + int width, int height); + +int udl_drop_usb(struct drm_device *dev); + +#define CMD_WRITE_RAW8 "\xAF\x60" /**< 8 bit raw write command. */ +#define CMD_WRITE_RL8 "\xAF\x61" /**< 8 bit run length command. */ +#define CMD_WRITE_COPY8 "\xAF\x62" /**< 8 bit copy command. */ +#define CMD_WRITE_RLX8 "\xAF\x63" /**< 8 bit extended run length command. */ + +#define CMD_WRITE_RAW16 "\xAF\x68" /**< 16 bit raw write command. */ +#define CMD_WRITE_RL16 "\xAF\x69" /**< 16 bit run length command. */ +#define CMD_WRITE_COPY16 "\xAF\x6A" /**< 16 bit copy command. */ +#define CMD_WRITE_RLX16 "\xAF\x6B" /**< 16 bit extended run length command. */ + +#endif diff --git a/drivers/gpu/drm/udl/udl_encoder.c b/drivers/gpu/drm/udl/udl_encoder.c new file mode 100644 index 0000000..56e75f0 --- /dev/null +++ b/drivers/gpu/drm/udl/udl_encoder.c @@ -0,0 +1,80 @@ +/* + * Copyright (C) 2012 Red Hat + * based in parts on udlfb.c: + * Copyright (C) 2009 Roberto De Ioris + * Copyright (C) 2009 Jaya Kumar + * Copyright (C) 2009 Bernie Thompson + * + * This file is subject to the terms and conditions of the GNU General Public + * License v2. See the file COPYING in the main directory of this archive for + * more details. + */ + +#include "drmP.h" +#include "drm_crtc.h" +#include "drm_crtc_helper.h" +#include "udl_drv.h" + +/* dummy encoder */ +void udl_enc_destroy(struct drm_encoder *encoder) +{ + drm_encoder_cleanup(encoder); + kfree(encoder); +} + +static void udl_encoder_disable(struct drm_encoder *encoder) +{ +} + +static bool udl_mode_fixup(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + return true; +} + +static void udl_encoder_prepare(struct drm_encoder *encoder) +{ +} + +static void udl_encoder_commit(struct drm_encoder *encoder) +{ +} + +static void udl_encoder_mode_set(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ +} + +static void +udl_encoder_dpms(struct drm_encoder *encoder, int mode) +{ +} + +static const struct drm_encoder_helper_funcs udl_helper_funcs = { + .dpms = udl_encoder_dpms, + .mode_fixup = udl_mode_fixup, + .prepare = udl_encoder_prepare, + .mode_set = udl_encoder_mode_set, + .commit = udl_encoder_commit, + .disable = udl_encoder_disable, +}; + +static const struct drm_encoder_funcs udl_enc_funcs = { + .destroy = udl_enc_destroy, +}; + +struct drm_encoder *udl_encoder_init(struct drm_device *dev) +{ + struct drm_encoder *encoder; + + encoder = kzalloc(sizeof(struct drm_encoder), GFP_KERNEL); + if (!encoder) + return NULL; + + drm_encoder_init(dev, encoder, &udl_enc_funcs, DRM_MODE_ENCODER_TMDS); + drm_encoder_helper_add(encoder, &udl_helper_funcs); + encoder->possible_crtcs = 1; + return encoder; +} diff --git a/drivers/gpu/drm/udl/udl_fb.c b/drivers/gpu/drm/udl/udl_fb.c new file mode 100644 index 0000000..b9282cf --- /dev/null +++ b/drivers/gpu/drm/udl/udl_fb.c @@ -0,0 +1,613 @@ +/* + * Copyright (C) 2012 Red Hat + * + * based in parts on udlfb.c: + * Copyright (C) 2009 Roberto De Ioris + * Copyright (C) 2009 Jaya Kumar + * Copyright (C) 2009 Bernie Thompson + * + * This file is subject to the terms and conditions of the GNU General Public + * License v2. See the file COPYING in the main directory of this archive for + * more details. + */ +#include +#include +#include + +#include "drmP.h" +#include "drm.h" +#include "drm_crtc.h" +#include "drm_crtc_helper.h" +#include "udl_drv.h" + +#include "drm_fb_helper.h" + +#define DL_DEFIO_WRITE_DELAY 5 /* fb_deferred_io.delay in jiffies */ + +static int fb_defio = 1; /* Optionally enable experimental fb_defio mmap support */ +static int fb_bpp = 16; + +module_param(fb_bpp, int, S_IWUSR | S_IRUSR | S_IWGRP | S_IRGRP); +module_param(fb_defio, int, S_IWUSR | S_IRUSR | S_IWGRP | S_IRGRP); + +struct udl_fbdev { + struct drm_fb_helper helper; + struct udl_framebuffer ufb; + struct list_head fbdev_list; + int fb_count; +}; + +#define DL_ALIGN_UP(x, a) ALIGN(x, a) +#define DL_ALIGN_DOWN(x, a) ALIGN(x-(a-1), a) + +/** Read the red component (0..255) of a 32 bpp colour. */ +#define DLO_RGB_GETRED(col) (uint8_t)((col) & 0xFF) + +/** Read the green component (0..255) of a 32 bpp colour. */ +#define DLO_RGB_GETGRN(col) (uint8_t)(((col) >> 8) & 0xFF) + +/** Read the blue component (0..255) of a 32 bpp colour. */ +#define DLO_RGB_GETBLU(col) (uint8_t)(((col) >> 16) & 0xFF) + +/** Return red/green component of a 16 bpp colour number. */ +#define DLO_RG16(red, grn) (uint8_t)((((red) & 0xF8) | ((grn) >> 5)) & 0xFF) + +/** Return green/blue component of a 16 bpp colour number. */ +#define DLO_GB16(grn, blu) (uint8_t)(((((grn) & 0x1C) << 3) | ((blu) >> 3)) & 0xFF) + +/** Return 8 bpp colour number from red, green and blue components. */ +#define DLO_RGB8(red, grn, blu) ((((red) << 5) | (((grn) & 3) << 3) | ((blu) & 7)) & 0xFF) + +#if 0 +static uint8_t rgb8(uint32_t col) +{ + uint8_t red = DLO_RGB_GETRED(col); + uint8_t grn = DLO_RGB_GETGRN(col); + uint8_t blu = DLO_RGB_GETBLU(col); + + return DLO_RGB8(red, grn, blu); +} + +static uint16_t rgb16(uint32_t col) +{ + uint8_t red = DLO_RGB_GETRED(col); + uint8_t grn = DLO_RGB_GETGRN(col); + uint8_t blu = DLO_RGB_GETBLU(col); + + return (DLO_RG16(red, grn) << 8) + DLO_GB16(grn, blu); +} +#endif + +/* + * NOTE: fb_defio.c is holding info->fbdefio.mutex + * Touching ANY framebuffer memory that triggers a page fault + * in fb_defio will cause a deadlock, when it also tries to + * grab the same mutex. + */ +static void udlfb_dpy_deferred_io(struct fb_info *info, + struct list_head *pagelist) +{ + struct page *cur; + struct fb_deferred_io *fbdefio = info->fbdefio; + struct udl_fbdev *ufbdev = info->par; + struct drm_device *dev = ufbdev->ufb.base.dev; + struct udl_device *udl = dev->dev_private; + struct urb *urb; + char *cmd; + cycles_t start_cycles, end_cycles; + int bytes_sent = 0; + int bytes_identical = 0; + int bytes_rendered = 0; + + if (!fb_defio) + return; + + start_cycles = get_cycles(); + + urb = udl_get_urb(dev); + if (!urb) + return; + + cmd = urb->transfer_buffer; + + /* walk the written page list and render each to device */ + list_for_each_entry(cur, &fbdefio->pagelist, lru) { + + if (udl_render_hline(dev, (ufbdev->ufb.base.bits_per_pixel / 8), + &urb, (char *) info->fix.smem_start, + &cmd, cur->index << PAGE_SHIFT, + cur->index << PAGE_SHIFT, + PAGE_SIZE, &bytes_identical, &bytes_sent)) + goto error; + bytes_rendered += PAGE_SIZE; + } + + if (cmd > (char *) urb->transfer_buffer) { + /* Send partial buffer remaining before exiting */ + int len = cmd - (char *) urb->transfer_buffer; + udl_submit_urb(dev, urb, len); + bytes_sent += len; + } else + udl_urb_completion(urb); + +error: + atomic_add(bytes_sent, &udl->bytes_sent); + atomic_add(bytes_identical, &udl->bytes_identical); + atomic_add(bytes_rendered, &udl->bytes_rendered); + end_cycles = get_cycles(); + atomic_add(((unsigned int) ((end_cycles - start_cycles) + >> 10)), /* Kcycles */ + &udl->cpu_kcycles_used); +} + +int udl_handle_damage(struct udl_framebuffer *fb, int x, int y, + int width, int height) +{ + struct drm_device *dev = fb->base.dev; + struct udl_device *udl = dev->dev_private; + int i, ret; + char *cmd; + cycles_t start_cycles, end_cycles; + int bytes_sent = 0; + int bytes_identical = 0; + struct urb *urb; + int aligned_x; + int bpp = (fb->base.bits_per_pixel / 8); + + if (!fb->active_16) + return 0; + + if (!fb->obj->vmapping) + udl_gem_vmap(fb->obj); + + start_cycles = get_cycles(); + + aligned_x = DL_ALIGN_DOWN(x, sizeof(unsigned long)); + width = DL_ALIGN_UP(width + (x-aligned_x), sizeof(unsigned long)); + x = aligned_x; + + if ((width <= 0) || + (x + width > fb->base.width) || + (y + height > fb->base.height)) + return -EINVAL; + + urb = udl_get_urb(dev); + if (!urb) + return 0; + cmd = urb->transfer_buffer; + + for (i = y; i < y + height ; i++) { + const int line_offset = fb->base.pitches[0] * i; + const int byte_offset = line_offset + (x * bpp); + const int dev_byte_offset = (fb->base.width * bpp * i) + (x * bpp); + if (udl_render_hline(dev, bpp, &urb, + (char *) fb->obj->vmapping, + &cmd, byte_offset, dev_byte_offset, + width * bpp, + &bytes_identical, &bytes_sent)) + goto error; + } + + if (cmd > (char *) urb->transfer_buffer) { + /* Send partial buffer remaining before exiting */ + int len = cmd - (char *) urb->transfer_buffer; + ret = udl_submit_urb(dev, urb, len); + bytes_sent += len; + } else + udl_urb_completion(urb); + +error: + atomic_add(bytes_sent, &udl->bytes_sent); + atomic_add(bytes_identical, &udl->bytes_identical); + atomic_add(width*height*bpp, &udl->bytes_rendered); + end_cycles = get_cycles(); + atomic_add(((unsigned int) ((end_cycles - start_cycles) + >> 10)), /* Kcycles */ + &udl->cpu_kcycles_used); + + return 0; +} + +static int udl_fb_mmap(struct fb_info *info, struct vm_area_struct *vma) +{ + unsigned long start = vma->vm_start; + unsigned long size = vma->vm_end - vma->vm_start; + unsigned long offset = vma->vm_pgoff << PAGE_SHIFT; + unsigned long page, pos; + + if (offset + size > info->fix.smem_len) + return -EINVAL; + + pos = (unsigned long)info->fix.smem_start + offset; + + pr_notice("mmap() framebuffer addr:%lu size:%lu\n", + pos, size); + + while (size > 0) { + page = vmalloc_to_pfn((void *)pos); + if (remap_pfn_range(vma, start, page, PAGE_SIZE, PAGE_SHARED)) + return -EAGAIN; + + start += PAGE_SIZE; + pos += PAGE_SIZE; + if (size > PAGE_SIZE) + size -= PAGE_SIZE; + else + size = 0; + } + + vma->vm_flags |= VM_RESERVED; /* avoid to swap out this VMA */ + return 0; +} + +static void udl_fb_fillrect(struct fb_info *info, const struct fb_fillrect *rect) +{ + struct udl_fbdev *ufbdev = info->par; + + sys_fillrect(info, rect); + + udl_handle_damage(&ufbdev->ufb, rect->dx, rect->dy, rect->width, + rect->height); +} + +static void udl_fb_copyarea(struct fb_info *info, const struct fb_copyarea *region) +{ + struct udl_fbdev *ufbdev = info->par; + + sys_copyarea(info, region); + + udl_handle_damage(&ufbdev->ufb, region->dx, region->dy, region->width, + region->height); +} + +static void udl_fb_imageblit(struct fb_info *info, const struct fb_image *image) +{ + struct udl_fbdev *ufbdev = info->par; + + sys_imageblit(info, image); + + udl_handle_damage(&ufbdev->ufb, image->dx, image->dy, image->width, + image->height); +} + +/* + * It's common for several clients to have framebuffer open simultaneously. + * e.g. both fbcon and X. Makes things interesting. + * Assumes caller is holding info->lock (for open and release at least) + */ +static int udl_fb_open(struct fb_info *info, int user) +{ + struct udl_fbdev *ufbdev = info->par; + struct drm_device *dev = ufbdev->ufb.base.dev; + struct udl_device *udl = dev->dev_private; + + /* If the USB device is gone, we don't accept new opens */ + if (drm_device_is_unplugged(udl->ddev)) + return -ENODEV; + + ufbdev->fb_count++; + + if (fb_defio && (info->fbdefio == NULL)) { + /* enable defio at last moment if not disabled by client */ + + struct fb_deferred_io *fbdefio; + + fbdefio = kmalloc(sizeof(struct fb_deferred_io), GFP_KERNEL); + + if (fbdefio) { + fbdefio->delay = DL_DEFIO_WRITE_DELAY; + fbdefio->deferred_io = udlfb_dpy_deferred_io; + } + + info->fbdefio = fbdefio; + fb_deferred_io_init(info); + } + + pr_notice("open /dev/fb%d user=%d fb_info=%p count=%d\n", + info->node, user, info, ufbdev->fb_count); + + return 0; +} + + +/* + * Assumes caller is holding info->lock mutex (for open and release at least) + */ +static int udl_fb_release(struct fb_info *info, int user) +{ + struct udl_fbdev *ufbdev = info->par; + + ufbdev->fb_count--; + + if ((ufbdev->fb_count == 0) && (info->fbdefio)) { + fb_deferred_io_cleanup(info); + kfree(info->fbdefio); + info->fbdefio = NULL; + info->fbops->fb_mmap = udl_fb_mmap; + } + + pr_warn("released /dev/fb%d user=%d count=%d\n", + info->node, user, ufbdev->fb_count); + + return 0; +} + +static struct fb_ops udlfb_ops = { + .owner = THIS_MODULE, + .fb_check_var = drm_fb_helper_check_var, + .fb_set_par = drm_fb_helper_set_par, + .fb_fillrect = udl_fb_fillrect, + .fb_copyarea = udl_fb_copyarea, + .fb_imageblit = udl_fb_imageblit, + .fb_pan_display = drm_fb_helper_pan_display, + .fb_blank = drm_fb_helper_blank, + .fb_setcmap = drm_fb_helper_setcmap, + .fb_debug_enter = drm_fb_helper_debug_enter, + .fb_debug_leave = drm_fb_helper_debug_leave, + .fb_mmap = udl_fb_mmap, + .fb_open = udl_fb_open, + .fb_release = udl_fb_release, +}; + +void udl_crtc_fb_gamma_set(struct drm_crtc *crtc, u16 red, u16 green, + u16 blue, int regno) +{ +} + +void udl_crtc_fb_gamma_get(struct drm_crtc *crtc, u16 *red, u16 *green, + u16 *blue, int regno) +{ + *red = 0; + *green = 0; + *blue = 0; +} + +static int udl_user_framebuffer_dirty(struct drm_framebuffer *fb, + struct drm_file *file, + unsigned flags, unsigned color, + struct drm_clip_rect *clips, + unsigned num_clips) +{ + struct udl_framebuffer *ufb = to_udl_fb(fb); + int i; + + if (!ufb->active_16) + return 0; + + for (i = 0; i < num_clips; i++) { + udl_handle_damage(ufb, clips[i].x1, clips[i].y1, + clips[i].x2 - clips[i].x1, + clips[i].y2 - clips[i].y1); + } + return 0; +} + +static void udl_user_framebuffer_destroy(struct drm_framebuffer *fb) +{ + struct udl_framebuffer *ufb = to_udl_fb(fb); + + if (ufb->obj) + drm_gem_object_unreference_unlocked(&ufb->obj->base); + + drm_framebuffer_cleanup(fb); + kfree(ufb); +} + +static const struct drm_framebuffer_funcs udlfb_funcs = { + .destroy = udl_user_framebuffer_destroy, + .dirty = udl_user_framebuffer_dirty, + .create_handle = NULL, +}; + + +static int +udl_framebuffer_init(struct drm_device *dev, + struct udl_framebuffer *ufb, + struct drm_mode_fb_cmd2 *mode_cmd, + struct udl_gem_object *obj) +{ + int ret; + + ufb->obj = obj; + ret = drm_framebuffer_init(dev, &ufb->base, &udlfb_funcs); + drm_helper_mode_fill_fb_struct(&ufb->base, mode_cmd); + return ret; +} + + +static int udlfb_create(struct udl_fbdev *ufbdev, + struct drm_fb_helper_surface_size *sizes) +{ + struct drm_device *dev = ufbdev->helper.dev; + struct fb_info *info; + struct device *device = &dev->usbdev->dev; + struct drm_framebuffer *fb; + struct drm_mode_fb_cmd2 mode_cmd; + struct udl_gem_object *obj; + uint32_t size; + int ret = 0; + + if (sizes->surface_bpp == 24) + sizes->surface_bpp = 32; + + mode_cmd.width = sizes->surface_width; + mode_cmd.height = sizes->surface_height; + mode_cmd.pitches[0] = mode_cmd.width * ((sizes->surface_bpp + 7) / 8); + + mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp, + sizes->surface_depth); + + size = mode_cmd.pitches[0] * mode_cmd.height; + size = ALIGN(size, PAGE_SIZE); + + obj = udl_gem_alloc_object(dev, size); + if (!obj) + goto out; + + ret = udl_gem_vmap(obj); + if (ret) { + DRM_ERROR("failed to vmap fb\n"); + goto out_gfree; + } + + info = framebuffer_alloc(0, device); + if (!info) { + ret = -ENOMEM; + goto out_gfree; + } + info->par = ufbdev; + + ret = udl_framebuffer_init(dev, &ufbdev->ufb, &mode_cmd, obj); + if (ret) + goto out_gfree; + + fb = &ufbdev->ufb.base; + + ufbdev->helper.fb = fb; + ufbdev->helper.fbdev = info; + + strcpy(info->fix.id, "udldrmfb"); + + info->screen_base = ufbdev->ufb.obj->vmapping; + info->fix.smem_len = size; + info->fix.smem_start = (unsigned long)ufbdev->ufb.obj->vmapping; + + info->flags = FBINFO_DEFAULT | FBINFO_CAN_FORCE_OUTPUT; + info->fbops = &udlfb_ops; + drm_fb_helper_fill_fix(info, fb->pitches[0], fb->depth); + drm_fb_helper_fill_var(info, &ufbdev->helper, sizes->fb_width, sizes->fb_height); + + ret = fb_alloc_cmap(&info->cmap, 256, 0); + if (ret) { + ret = -ENOMEM; + goto out_gfree; + } + + + DRM_DEBUG_KMS("allocated %dx%d vmal %p\n", + fb->width, fb->height, + ufbdev->ufb.obj->vmapping); + + return ret; +out_gfree: + drm_gem_object_unreference(&ufbdev->ufb.obj->base); +out: + return ret; +} + +static int udl_fb_find_or_create_single(struct drm_fb_helper *helper, + struct drm_fb_helper_surface_size *sizes) +{ + struct udl_fbdev *ufbdev = (struct udl_fbdev *)helper; + int new_fb = 0; + int ret; + + if (!helper->fb) { + ret = udlfb_create(ufbdev, sizes); + if (ret) + return ret; + + new_fb = 1; + } + return new_fb; +} + +static struct drm_fb_helper_funcs udl_fb_helper_funcs = { + .gamma_set = udl_crtc_fb_gamma_set, + .gamma_get = udl_crtc_fb_gamma_get, + .fb_probe = udl_fb_find_or_create_single, +}; + +static void udl_fbdev_destroy(struct drm_device *dev, + struct udl_fbdev *ufbdev) +{ + struct fb_info *info; + if (ufbdev->helper.fbdev) { + info = ufbdev->helper.fbdev; + unregister_framebuffer(info); + if (info->cmap.len) + fb_dealloc_cmap(&info->cmap); + framebuffer_release(info); + } + drm_fb_helper_fini(&ufbdev->helper); + drm_framebuffer_cleanup(&ufbdev->ufb.base); + drm_gem_object_unreference_unlocked(&ufbdev->ufb.obj->base); +} + +int udl_fbdev_init(struct drm_device *dev) +{ + struct udl_device *udl = dev->dev_private; + int bpp_sel = fb_bpp; + struct udl_fbdev *ufbdev; + int ret; + + ufbdev = kzalloc(sizeof(struct udl_fbdev), GFP_KERNEL); + if (!ufbdev) + return -ENOMEM; + + udl->fbdev = ufbdev; + ufbdev->helper.funcs = &udl_fb_helper_funcs; + + ret = drm_fb_helper_init(dev, &ufbdev->helper, + 1, 1); + if (ret) { + kfree(ufbdev); + return ret; + + } + + drm_fb_helper_single_add_all_connectors(&ufbdev->helper); + drm_fb_helper_initial_config(&ufbdev->helper, bpp_sel); + return 0; +} + +void udl_fbdev_cleanup(struct drm_device *dev) +{ + struct udl_device *udl = dev->dev_private; + if (!udl->fbdev) + return; + + udl_fbdev_destroy(dev, udl->fbdev); + kfree(udl->fbdev); + udl->fbdev = NULL; +} + +void udl_fbdev_unplug(struct drm_device *dev) +{ + struct udl_device *udl = dev->dev_private; + struct udl_fbdev *ufbdev; + if (!udl->fbdev) + return; + + ufbdev = udl->fbdev; + if (ufbdev->helper.fbdev) { + struct fb_info *info; + info = ufbdev->helper.fbdev; + unlink_framebuffer(info); + } +} + +struct drm_framebuffer * +udl_fb_user_fb_create(struct drm_device *dev, + struct drm_file *file, + struct drm_mode_fb_cmd2 *mode_cmd) +{ + struct drm_gem_object *obj; + struct udl_framebuffer *ufb; + int ret; + + obj = drm_gem_object_lookup(dev, file, mode_cmd->handles[0]); + if (obj == NULL) + return ERR_PTR(-ENOENT); + + ufb = kzalloc(sizeof(*ufb), GFP_KERNEL); + if (ufb == NULL) + return ERR_PTR(-ENOMEM); + + ret = udl_framebuffer_init(dev, ufb, mode_cmd, to_udl_bo(obj)); + if (ret) { + kfree(ufb); + return ERR_PTR(-EINVAL); + } + return &ufb->base; +} diff --git a/drivers/gpu/drm/udl/udl_gem.c b/drivers/gpu/drm/udl/udl_gem.c new file mode 100644 index 0000000..92f19ef --- /dev/null +++ b/drivers/gpu/drm/udl/udl_gem.c @@ -0,0 +1,241 @@ +/* + * Copyright (C) 2012 Red Hat + * + * This file is subject to the terms and conditions of the GNU General Public + * License v2. See the file COPYING in the main directory of this archive for + * more details. + */ + +#include "drmP.h" +#include "udl_drv.h" +#include + +struct udl_gem_object *udl_gem_alloc_object(struct drm_device *dev, + size_t size) +{ + struct udl_gem_object *obj; + + obj = kzalloc(sizeof(*obj), GFP_KERNEL); + if (obj == NULL) + return NULL; + + if (drm_gem_object_init(dev, &obj->base, size) != 0) { + kfree(obj); + return NULL; + } + + return obj; +} + +static int +udl_gem_create(struct drm_file *file, + struct drm_device *dev, + uint64_t size, + uint32_t *handle_p) +{ + struct udl_gem_object *obj; + int ret; + u32 handle; + + size = roundup(size, PAGE_SIZE); + + obj = udl_gem_alloc_object(dev, size); + if (obj == NULL) + return -ENOMEM; + + ret = drm_gem_handle_create(file, &obj->base, &handle); + if (ret) { + drm_gem_object_release(&obj->base); + kfree(obj); + return ret; + } + + drm_gem_object_unreference(&obj->base); + *handle_p = handle; + return 0; +} + +int udl_dumb_create(struct drm_file *file, + struct drm_device *dev, + struct drm_mode_create_dumb *args) +{ + args->pitch = args->width * ((args->bpp + 1) / 8); + args->size = args->pitch * args->height; + return udl_gem_create(file, dev, + args->size, &args->handle); +} + +int udl_dumb_destroy(struct drm_file *file, struct drm_device *dev, + uint32_t handle) +{ + return drm_gem_handle_delete(file, handle); +} + +int udl_drm_gem_mmap(struct file *filp, struct vm_area_struct *vma) +{ + int ret; + + ret = drm_gem_mmap(filp, vma); + if (ret) + return ret; + + vma->vm_flags &= ~VM_PFNMAP; + vma->vm_flags |= VM_MIXEDMAP; + + return ret; +} + +int udl_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf) +{ + struct udl_gem_object *obj = to_udl_bo(vma->vm_private_data); + struct page *page; + unsigned int page_offset; + int ret = 0; + + page_offset = ((unsigned long)vmf->virtual_address - vma->vm_start) >> + PAGE_SHIFT; + + if (!obj->pages) + return VM_FAULT_SIGBUS; + + page = obj->pages[page_offset]; + ret = vm_insert_page(vma, (unsigned long)vmf->virtual_address, page); + switch (ret) { + case -EAGAIN: + set_need_resched(); + case 0: + case -ERESTARTSYS: + return VM_FAULT_NOPAGE; + case -ENOMEM: + return VM_FAULT_OOM; + default: + return VM_FAULT_SIGBUS; + } +} + +int udl_gem_init_object(struct drm_gem_object *obj) +{ + BUG(); + + return 0; +} + +static int udl_gem_get_pages(struct udl_gem_object *obj, gfp_t gfpmask) +{ + int page_count, i; + struct page *page; + struct inode *inode; + struct address_space *mapping; + + if (obj->pages) + return 0; + + page_count = obj->base.size / PAGE_SIZE; + BUG_ON(obj->pages != NULL); + obj->pages = drm_malloc_ab(page_count, sizeof(struct page *)); + if (obj->pages == NULL) + return -ENOMEM; + + inode = obj->base.filp->f_path.dentry->d_inode; + mapping = inode->i_mapping; + gfpmask |= mapping_gfp_mask(mapping); + + for (i = 0; i < page_count; i++) { + page = shmem_read_mapping_page_gfp(mapping, i, gfpmask); + if (IS_ERR(page)) + goto err_pages; + obj->pages[i] = page; + } + + return 0; +err_pages: + while (i--) + page_cache_release(obj->pages[i]); + drm_free_large(obj->pages); + obj->pages = NULL; + return PTR_ERR(page); +} + +static void udl_gem_put_pages(struct udl_gem_object *obj) +{ + int page_count = obj->base.size / PAGE_SIZE; + int i; + + for (i = 0; i < page_count; i++) + page_cache_release(obj->pages[i]); + + drm_free_large(obj->pages); + obj->pages = NULL; +} + +int udl_gem_vmap(struct udl_gem_object *obj) +{ + int page_count = obj->base.size / PAGE_SIZE; + int ret; + + ret = udl_gem_get_pages(obj, GFP_KERNEL); + if (ret) + return ret; + + obj->vmapping = vmap(obj->pages, page_count, 0, PAGE_KERNEL); + if (!obj->vmapping) + return -ENOMEM; + return 0; +} + +void udl_gem_vunmap(struct udl_gem_object *obj) +{ + if (obj->vmapping) + vunmap(obj->vmapping); + + udl_gem_put_pages(obj); +} + +void udl_gem_free_object(struct drm_gem_object *gem_obj) +{ + struct udl_gem_object *obj = to_udl_bo(gem_obj); + + if (obj->vmapping) + udl_gem_vunmap(obj); + + if (obj->pages) + udl_gem_put_pages(obj); + + if (gem_obj->map_list.map) + drm_gem_free_mmap_offset(gem_obj); +} + +/* the dumb interface doesn't work with the GEM straight MMAP + interface, it expects to do MMAP on the drm fd, like normal */ +int udl_gem_mmap(struct drm_file *file, struct drm_device *dev, + uint32_t handle, uint64_t *offset) +{ + struct udl_gem_object *gobj; + struct drm_gem_object *obj; + int ret = 0; + + mutex_lock(&dev->struct_mutex); + obj = drm_gem_object_lookup(dev, file, handle); + if (obj == NULL) { + ret = -ENOENT; + goto unlock; + } + gobj = to_udl_bo(obj); + + ret = udl_gem_get_pages(gobj, GFP_KERNEL); + if (ret) + return ret; + if (!gobj->base.map_list.map) { + ret = drm_gem_create_mmap_offset(obj); + if (ret) + goto out; + } + + *offset = (u64)gobj->base.map_list.hash.key << PAGE_SHIFT; + +out: + drm_gem_object_unreference(&gobj->base); +unlock: + mutex_unlock(&dev->struct_mutex); + return ret; +} diff --git a/drivers/gpu/drm/udl/udl_main.c b/drivers/gpu/drm/udl/udl_main.c new file mode 100644 index 0000000..a8d5f09 --- /dev/null +++ b/drivers/gpu/drm/udl/udl_main.c @@ -0,0 +1,338 @@ +/* + * Copyright (C) 2012 Red Hat + * + * based in parts on udlfb.c: + * Copyright (C) 2009 Roberto De Ioris + * Copyright (C) 2009 Jaya Kumar + * Copyright (C) 2009 Bernie Thompson + * + * This file is subject to the terms and conditions of the GNU General Public + * License v2. See the file COPYING in the main directory of this archive for + * more details. + */ +#include "drmP.h" +#include "udl_drv.h" + +/* -BULK_SIZE as per usb-skeleton. Can we get full page and avoid overhead? */ +#define BULK_SIZE 512 + +#define MAX_TRANSFER (PAGE_SIZE*16 - BULK_SIZE) +#define WRITES_IN_FLIGHT (4) +#define MAX_VENDOR_DESCRIPTOR_SIZE 256 + +#define GET_URB_TIMEOUT HZ +#define FREE_URB_TIMEOUT (HZ*2) + +static int udl_parse_vendor_descriptor(struct drm_device *dev, + struct usb_device *usbdev) +{ + struct udl_device *udl = dev->dev_private; + char *desc; + char *buf; + char *desc_end; + + u8 total_len = 0; + + buf = kzalloc(MAX_VENDOR_DESCRIPTOR_SIZE, GFP_KERNEL); + if (!buf) + return false; + desc = buf; + + total_len = usb_get_descriptor(usbdev, 0x5f, /* vendor specific */ + 0, desc, MAX_VENDOR_DESCRIPTOR_SIZE); + if (total_len > 5) { + DRM_INFO("vendor descriptor length:%x data:%02x %02x %02x %02x" \ + "%02x %02x %02x %02x %02x %02x %02x\n", + total_len, desc[0], + desc[1], desc[2], desc[3], desc[4], desc[5], desc[6], + desc[7], desc[8], desc[9], desc[10]); + + if ((desc[0] != total_len) || /* descriptor length */ + (desc[1] != 0x5f) || /* vendor descriptor type */ + (desc[2] != 0x01) || /* version (2 bytes) */ + (desc[3] != 0x00) || + (desc[4] != total_len - 2)) /* length after type */ + goto unrecognized; + + desc_end = desc + total_len; + desc += 5; /* the fixed header we've already parsed */ + + while (desc < desc_end) { + u8 length; + u16 key; + + key = *((u16 *) desc); + desc += sizeof(u16); + length = *desc; + desc++; + + switch (key) { + case 0x0200: { /* max_area */ + u32 max_area; + max_area = le32_to_cpu(*((u32 *)desc)); + DRM_DEBUG("DL chip limited to %d pixel modes\n", + max_area); + udl->sku_pixel_limit = max_area; + break; + } + default: + break; + } + desc += length; + } + } + + goto success; + +unrecognized: + /* allow udlfb to load for now even if firmware unrecognized */ + DRM_ERROR("Unrecognized vendor firmware descriptor\n"); + +success: + kfree(buf); + return true; +} + +static void udl_release_urb_work(struct work_struct *work) +{ + struct urb_node *unode = container_of(work, struct urb_node, + release_urb_work.work); + + up(&unode->dev->urbs.limit_sem); +} + +void udl_urb_completion(struct urb *urb) +{ + struct urb_node *unode = urb->context; + struct udl_device *udl = unode->dev; + unsigned long flags; + + /* sync/async unlink faults aren't errors */ + if (urb->status) { + if (!(urb->status == -ENOENT || + urb->status == -ECONNRESET || + urb->status == -ESHUTDOWN)) { + DRM_ERROR("%s - nonzero write bulk status received: %d\n", + __func__, urb->status); + atomic_set(&udl->lost_pixels, 1); + } + } + + urb->transfer_buffer_length = udl->urbs.size; /* reset to actual */ + + spin_lock_irqsave(&udl->urbs.lock, flags); + list_add_tail(&unode->entry, &udl->urbs.list); + udl->urbs.available++; + spin_unlock_irqrestore(&udl->urbs.lock, flags); + +#if 0 + /* + * When using fb_defio, we deadlock if up() is called + * while another is waiting. So queue to another process. + */ + if (fb_defio) + schedule_delayed_work(&unode->release_urb_work, 0); + else +#endif + up(&udl->urbs.limit_sem); +} + +static void udl_free_urb_list(struct drm_device *dev) +{ + struct udl_device *udl = dev->dev_private; + int count = udl->urbs.count; + struct list_head *node; + struct urb_node *unode; + struct urb *urb; + int ret; + unsigned long flags; + + DRM_DEBUG("Waiting for completes and freeing all render urbs\n"); + + /* keep waiting and freeing, until we've got 'em all */ + while (count--) { + + /* Getting interrupted means a leak, but ok at shutdown*/ + ret = down_interruptible(&udl->urbs.limit_sem); + if (ret) + break; + + spin_lock_irqsave(&udl->urbs.lock, flags); + + node = udl->urbs.list.next; /* have reserved one with sem */ + list_del_init(node); + + spin_unlock_irqrestore(&udl->urbs.lock, flags); + + unode = list_entry(node, struct urb_node, entry); + urb = unode->urb; + + /* Free each separately allocated piece */ + usb_free_coherent(urb->dev, udl->urbs.size, + urb->transfer_buffer, urb->transfer_dma); + usb_free_urb(urb); + kfree(node); + } + udl->urbs.count = 0; +} + +static int udl_alloc_urb_list(struct drm_device *dev, int count, size_t size) +{ + struct udl_device *udl = dev->dev_private; + int i = 0; + struct urb *urb; + struct urb_node *unode; + char *buf; + + spin_lock_init(&udl->urbs.lock); + + udl->urbs.size = size; + INIT_LIST_HEAD(&udl->urbs.list); + + while (i < count) { + unode = kzalloc(sizeof(struct urb_node), GFP_KERNEL); + if (!unode) + break; + unode->dev = udl; + + INIT_DELAYED_WORK(&unode->release_urb_work, + udl_release_urb_work); + + urb = usb_alloc_urb(0, GFP_KERNEL); + if (!urb) { + kfree(unode); + break; + } + unode->urb = urb; + + buf = usb_alloc_coherent(udl->ddev->usbdev, MAX_TRANSFER, GFP_KERNEL, + &urb->transfer_dma); + if (!buf) { + kfree(unode); + usb_free_urb(urb); + break; + } + + /* urb->transfer_buffer_length set to actual before submit */ + usb_fill_bulk_urb(urb, udl->ddev->usbdev, usb_sndbulkpipe(udl->ddev->usbdev, 1), + buf, size, udl_urb_completion, unode); + urb->transfer_flags |= URB_NO_TRANSFER_DMA_MAP; + + list_add_tail(&unode->entry, &udl->urbs.list); + + i++; + } + + sema_init(&udl->urbs.limit_sem, i); + udl->urbs.count = i; + udl->urbs.available = i; + + DRM_DEBUG("allocated %d %d byte urbs\n", i, (int) size); + + return i; +} + +struct urb *udl_get_urb(struct drm_device *dev) +{ + struct udl_device *udl = dev->dev_private; + int ret = 0; + struct list_head *entry; + struct urb_node *unode; + struct urb *urb = NULL; + unsigned long flags; + + /* Wait for an in-flight buffer to complete and get re-queued */ + ret = down_timeout(&udl->urbs.limit_sem, GET_URB_TIMEOUT); + if (ret) { + atomic_set(&udl->lost_pixels, 1); + DRM_INFO("wait for urb interrupted: %x available: %d\n", + ret, udl->urbs.available); + goto error; + } + + spin_lock_irqsave(&udl->urbs.lock, flags); + + BUG_ON(list_empty(&udl->urbs.list)); /* reserved one with limit_sem */ + entry = udl->urbs.list.next; + list_del_init(entry); + udl->urbs.available--; + + spin_unlock_irqrestore(&udl->urbs.lock, flags); + + unode = list_entry(entry, struct urb_node, entry); + urb = unode->urb; + +error: + return urb; +} + +int udl_submit_urb(struct drm_device *dev, struct urb *urb, size_t len) +{ + struct udl_device *udl = dev->dev_private; + int ret; + + BUG_ON(len > udl->urbs.size); + + urb->transfer_buffer_length = len; /* set to actual payload len */ + ret = usb_submit_urb(urb, GFP_ATOMIC); + if (ret) { + udl_urb_completion(urb); /* because no one else will */ + atomic_set(&udl->lost_pixels, 1); + DRM_ERROR("usb_submit_urb error %x\n", ret); + } + return ret; +} + +int udl_driver_load(struct drm_device *dev, unsigned long flags) +{ + struct udl_device *udl; + int ret; + + DRM_DEBUG("\n"); + udl = kzalloc(sizeof(struct udl_device), GFP_KERNEL); + if (!udl) + return -ENOMEM; + + udl->ddev = dev; + dev->dev_private = udl; + + if (!udl_parse_vendor_descriptor(dev, dev->usbdev)) { + DRM_ERROR("firmware not recognized. Assume incompatible device\n"); + goto err; + } + + if (!udl_alloc_urb_list(dev, WRITES_IN_FLIGHT, MAX_TRANSFER)) { + ret = -ENOMEM; + DRM_ERROR("udl_alloc_urb_list failed\n"); + goto err; + } + + DRM_DEBUG("\n"); + ret = udl_modeset_init(dev); + + ret = udl_fbdev_init(dev); + return 0; +err: + kfree(udl); + DRM_ERROR("%d\n", ret); + return ret; +} + +int udl_drop_usb(struct drm_device *dev) +{ + udl_free_urb_list(dev); + return 0; +} + +int udl_driver_unload(struct drm_device *dev) +{ + struct udl_device *udl = dev->dev_private; + + if (udl->urbs.count) + udl_free_urb_list(dev); + + udl_fbdev_cleanup(dev); + udl_modeset_cleanup(dev); + kfree(udl); + return 0; +} diff --git a/drivers/gpu/drm/udl/udl_modeset.c b/drivers/gpu/drm/udl/udl_modeset.c new file mode 100644 index 0000000..b3ecb3d --- /dev/null +++ b/drivers/gpu/drm/udl/udl_modeset.c @@ -0,0 +1,414 @@ +/* + * Copyright (C) 2012 Red Hat + * + * based in parts on udlfb.c: + * Copyright (C) 2009 Roberto De Ioris + * Copyright (C) 2009 Jaya Kumar + * Copyright (C) 2009 Bernie Thompson + + * This file is subject to the terms and conditions of the GNU General Public + * License v2. See the file COPYING in the main directory of this archive for + * more details. + */ + +#include "drmP.h" +#include "drm_crtc.h" +#include "drm_crtc_helper.h" +#include "udl_drv.h" + +/* + * All DisplayLink bulk operations start with 0xAF, followed by specific code + * All operations are written to buffers which then later get sent to device + */ +static char *udl_set_register(char *buf, u8 reg, u8 val) +{ + *buf++ = 0xAF; + *buf++ = 0x20; + *buf++ = reg; + *buf++ = val; + return buf; +} + +static char *udl_vidreg_lock(char *buf) +{ + return udl_set_register(buf, 0xFF, 0x00); +} + +static char *udl_vidreg_unlock(char *buf) +{ + return udl_set_register(buf, 0xFF, 0xFF); +} + +/* + * On/Off for driving the DisplayLink framebuffer to the display + * 0x00 H and V sync on + * 0x01 H and V sync off (screen blank but powered) + * 0x07 DPMS powerdown (requires modeset to come back) + */ +static char *udl_enable_hvsync(char *buf, bool enable) +{ + if (enable) + return udl_set_register(buf, 0x1F, 0x00); + else + return udl_set_register(buf, 0x1F, 0x07); +} + +static char *udl_set_color_depth(char *buf, u8 selection) +{ + return udl_set_register(buf, 0x00, selection); +} + +static char *udl_set_base16bpp(char *wrptr, u32 base) +{ + /* the base pointer is 16 bits wide, 0x20 is hi byte. */ + wrptr = udl_set_register(wrptr, 0x20, base >> 16); + wrptr = udl_set_register(wrptr, 0x21, base >> 8); + return udl_set_register(wrptr, 0x22, base); +} + +/* + * DisplayLink HW has separate 16bpp and 8bpp framebuffers. + * In 24bpp modes, the low 323 RGB bits go in the 8bpp framebuffer + */ +static char *udl_set_base8bpp(char *wrptr, u32 base) +{ + wrptr = udl_set_register(wrptr, 0x26, base >> 16); + wrptr = udl_set_register(wrptr, 0x27, base >> 8); + return udl_set_register(wrptr, 0x28, base); +} + +static char *udl_set_register_16(char *wrptr, u8 reg, u16 value) +{ + wrptr = udl_set_register(wrptr, reg, value >> 8); + return udl_set_register(wrptr, reg+1, value); +} + +/* + * This is kind of weird because the controller takes some + * register values in a different byte order than other registers. + */ +static char *udl_set_register_16be(char *wrptr, u8 reg, u16 value) +{ + wrptr = udl_set_register(wrptr, reg, value); + return udl_set_register(wrptr, reg+1, value >> 8); +} + +/* + * LFSR is linear feedback shift register. The reason we have this is + * because the display controller needs to minimize the clock depth of + * various counters used in the display path. So this code reverses the + * provided value into the lfsr16 value by counting backwards to get + * the value that needs to be set in the hardware comparator to get the + * same actual count. This makes sense once you read above a couple of + * times and think about it from a hardware perspective. + */ +static u16 udl_lfsr16(u16 actual_count) +{ + u32 lv = 0xFFFF; /* This is the lfsr value that the hw starts with */ + + while (actual_count--) { + lv = ((lv << 1) | + (((lv >> 15) ^ (lv >> 4) ^ (lv >> 2) ^ (lv >> 1)) & 1)) + & 0xFFFF; + } + + return (u16) lv; +} + +/* + * This does LFSR conversion on the value that is to be written. + * See LFSR explanation above for more detail. + */ +static char *udl_set_register_lfsr16(char *wrptr, u8 reg, u16 value) +{ + return udl_set_register_16(wrptr, reg, udl_lfsr16(value)); +} + +/* + * This takes a standard fbdev screeninfo struct and all of its monitor mode + * details and converts them into the DisplayLink equivalent register commands. + ERR(vreg(dev, 0x00, (color_depth == 16) ? 0 : 1)); + ERR(vreg_lfsr16(dev, 0x01, xDisplayStart)); + ERR(vreg_lfsr16(dev, 0x03, xDisplayEnd)); + ERR(vreg_lfsr16(dev, 0x05, yDisplayStart)); + ERR(vreg_lfsr16(dev, 0x07, yDisplayEnd)); + ERR(vreg_lfsr16(dev, 0x09, xEndCount)); + ERR(vreg_lfsr16(dev, 0x0B, hSyncStart)); + ERR(vreg_lfsr16(dev, 0x0D, hSyncEnd)); + ERR(vreg_big_endian(dev, 0x0F, hPixels)); + ERR(vreg_lfsr16(dev, 0x11, yEndCount)); + ERR(vreg_lfsr16(dev, 0x13, vSyncStart)); + ERR(vreg_lfsr16(dev, 0x15, vSyncEnd)); + ERR(vreg_big_endian(dev, 0x17, vPixels)); + ERR(vreg_little_endian(dev, 0x1B, pixelClock5KHz)); + + ERR(vreg(dev, 0x1F, 0)); + + ERR(vbuf(dev, WRITE_VIDREG_UNLOCK, DSIZEOF(WRITE_VIDREG_UNLOCK))); + */ +static char *udl_set_vid_cmds(char *wrptr, struct drm_display_mode *mode) +{ + u16 xds, yds; + u16 xde, yde; + u16 yec; + + /* x display start */ + xds = mode->crtc_htotal - mode->crtc_hsync_start; + wrptr = udl_set_register_lfsr16(wrptr, 0x01, xds); + /* x display end */ + xde = xds + mode->crtc_hdisplay; + wrptr = udl_set_register_lfsr16(wrptr, 0x03, xde); + + /* y display start */ + yds = mode->crtc_vtotal - mode->crtc_vsync_start; + wrptr = udl_set_register_lfsr16(wrptr, 0x05, yds); + /* y display end */ + yde = yds + mode->crtc_vdisplay; + wrptr = udl_set_register_lfsr16(wrptr, 0x07, yde); + + /* x end count is active + blanking - 1 */ + wrptr = udl_set_register_lfsr16(wrptr, 0x09, + mode->crtc_htotal - 1); + + /* libdlo hardcodes hsync start to 1 */ + wrptr = udl_set_register_lfsr16(wrptr, 0x0B, 1); + + /* hsync end is width of sync pulse + 1 */ + wrptr = udl_set_register_lfsr16(wrptr, 0x0D, + mode->crtc_hsync_end - mode->crtc_hsync_start + 1); + + /* hpixels is active pixels */ + wrptr = udl_set_register_16(wrptr, 0x0F, mode->hdisplay); + + /* yendcount is vertical active + vertical blanking */ + yec = mode->crtc_vtotal; + wrptr = udl_set_register_lfsr16(wrptr, 0x11, yec); + + /* libdlo hardcodes vsync start to 0 */ + wrptr = udl_set_register_lfsr16(wrptr, 0x13, 0); + + /* vsync end is width of vsync pulse */ + wrptr = udl_set_register_lfsr16(wrptr, 0x15, mode->crtc_vsync_end - mode->crtc_vsync_start); + + /* vpixels is active pixels */ + wrptr = udl_set_register_16(wrptr, 0x17, mode->crtc_vdisplay); + + wrptr = udl_set_register_16be(wrptr, 0x1B, + mode->clock / 5); + + return wrptr; +} + +static int udl_crtc_write_mode_to_hw(struct drm_crtc *crtc) +{ + struct drm_device *dev = crtc->dev; + struct udl_device *udl = dev->dev_private; + struct urb *urb; + char *buf; + int retval; + + urb = udl_get_urb(dev); + if (!urb) + return -ENOMEM; + + buf = (char *)urb->transfer_buffer; + + memcpy(buf, udl->mode_buf, udl->mode_buf_len); + retval = udl_submit_urb(dev, urb, udl->mode_buf_len); + DRM_INFO("write mode info %d\n", udl->mode_buf_len); + return retval; +} + + +static void udl_crtc_dpms(struct drm_crtc *crtc, int mode) +{ + struct drm_device *dev = crtc->dev; + struct udl_device *udl = dev->dev_private; + int retval; + + if (mode == DRM_MODE_DPMS_OFF) { + char *buf; + struct urb *urb; + urb = udl_get_urb(dev); + if (!urb) + return; + + buf = (char *)urb->transfer_buffer; + buf = udl_vidreg_lock(buf); + buf = udl_enable_hvsync(buf, false); + buf = udl_vidreg_unlock(buf); + + retval = udl_submit_urb(dev, urb, buf - (char *) + urb->transfer_buffer); + } else { + if (udl->mode_buf_len == 0) { + DRM_ERROR("Trying to enable DPMS with no mode\n"); + return; + } + udl_crtc_write_mode_to_hw(crtc); + } + +} + +static bool udl_crtc_mode_fixup(struct drm_crtc *crtc, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) + +{ + return true; +} + +#if 0 +static int +udl_pipe_set_base_atomic(struct drm_crtc *crtc, struct drm_framebuffer *fb, + int x, int y, enum mode_set_atomic state) +{ + return 0; +} + +static int +udl_pipe_set_base(struct drm_crtc *crtc, int x, int y, + struct drm_framebuffer *old_fb) +{ + return 0; +} +#endif + +static int udl_crtc_mode_set(struct drm_crtc *crtc, + struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode, + int x, int y, + struct drm_framebuffer *old_fb) + +{ + struct drm_device *dev = crtc->dev; + struct udl_framebuffer *ufb = to_udl_fb(crtc->fb); + struct udl_device *udl = dev->dev_private; + char *buf; + char *wrptr; + int color_depth = 0; + + buf = (char *)udl->mode_buf; + + /* for now we just clip 24 -> 16 - if we fix that fix this */ + /*if (crtc->fb->bits_per_pixel != 16) + color_depth = 1; */ + + /* This first section has to do with setting the base address on the + * controller * associated with the display. There are 2 base + * pointers, currently, we only * use the 16 bpp segment. + */ + wrptr = udl_vidreg_lock(buf); + wrptr = udl_set_color_depth(wrptr, color_depth); + /* set base for 16bpp segment to 0 */ + wrptr = udl_set_base16bpp(wrptr, 0); + /* set base for 8bpp segment to end of fb */ + wrptr = udl_set_base8bpp(wrptr, 2 * mode->vdisplay * mode->hdisplay); + + wrptr = udl_set_vid_cmds(wrptr, adjusted_mode); + wrptr = udl_enable_hvsync(wrptr, true); + wrptr = udl_vidreg_unlock(wrptr); + + ufb->active_16 = true; + if (old_fb) { + struct udl_framebuffer *uold_fb = to_udl_fb(old_fb); + uold_fb->active_16 = false; + } + udl->mode_buf_len = wrptr - buf; + + /* damage all of it */ + udl_handle_damage(ufb, 0, 0, ufb->base.width, ufb->base.height); + return 0; +} + + +static void udl_crtc_disable(struct drm_crtc *crtc) +{ + + +} + +static void udl_crtc_destroy(struct drm_crtc *crtc) +{ + drm_crtc_cleanup(crtc); + kfree(crtc); +} + +static void udl_load_lut(struct drm_crtc *crtc) +{ +} + +static void udl_crtc_prepare(struct drm_crtc *crtc) +{ +} + +static void udl_crtc_commit(struct drm_crtc *crtc) +{ + udl_crtc_dpms(crtc, DRM_MODE_DPMS_ON); +} + +static struct drm_crtc_helper_funcs udl_helper_funcs = { + .dpms = udl_crtc_dpms, + .mode_fixup = udl_crtc_mode_fixup, + .mode_set = udl_crtc_mode_set, + .prepare = udl_crtc_prepare, + .commit = udl_crtc_commit, + .disable = udl_crtc_disable, + .load_lut = udl_load_lut, +}; + +static const struct drm_crtc_funcs udl_crtc_funcs = { + .set_config = drm_crtc_helper_set_config, + .destroy = udl_crtc_destroy, +}; + +int udl_crtc_init(struct drm_device *dev) +{ + struct drm_crtc *crtc; + + crtc = kzalloc(sizeof(struct drm_crtc) + sizeof(struct drm_connector *), GFP_KERNEL); + if (crtc == NULL) + return -ENOMEM; + + drm_crtc_init(dev, crtc, &udl_crtc_funcs); + drm_crtc_helper_add(crtc, &udl_helper_funcs); + + return 0; +} + +static const struct drm_mode_config_funcs udl_mode_funcs = { + .fb_create = udl_fb_user_fb_create, + .output_poll_changed = NULL, +}; + +int udl_modeset_init(struct drm_device *dev) +{ + struct drm_encoder *encoder; + drm_mode_config_init(dev); + + dev->mode_config.min_width = 640; + dev->mode_config.min_height = 480; + + dev->mode_config.max_width = 2048; + dev->mode_config.max_height = 2048; + + dev->mode_config.prefer_shadow = 0; + dev->mode_config.preferred_depth = 24; + + dev->mode_config.funcs = (void *)&udl_mode_funcs; + + drm_mode_create_dirty_info_property(dev); + + udl_crtc_init(dev); + + encoder = udl_encoder_init(dev); + + udl_connector_init(dev, encoder); + + return 0; +} + +void udl_modeset_cleanup(struct drm_device *dev) +{ + drm_mode_config_cleanup(dev); +} diff --git a/drivers/gpu/drm/udl/udl_transfer.c b/drivers/gpu/drm/udl/udl_transfer.c new file mode 100644 index 0000000..fc11344 --- /dev/null +++ b/drivers/gpu/drm/udl/udl_transfer.c @@ -0,0 +1,254 @@ +/* + * Copyright (C) 2012 Red Hat + * based in parts on udlfb.c: + * Copyright (C) 2009 Roberto De Ioris + * Copyright (C) 2009 Jaya Kumar + * Copyright (C) 2009 Bernie Thompson + * + * This file is subject to the terms and conditions of the GNU General Public + * License v2. See the file COPYING in the main directory of this archive for + * more details. + */ + +#include +#include +#include +#include + +#include "drmP.h" +#include "udl_drv.h" + +#define MAX_CMD_PIXELS 255 + +#define RLX_HEADER_BYTES 7 +#define MIN_RLX_PIX_BYTES 4 +#define MIN_RLX_CMD_BYTES (RLX_HEADER_BYTES + MIN_RLX_PIX_BYTES) + +#define RLE_HEADER_BYTES 6 +#define MIN_RLE_PIX_BYTES 3 +#define MIN_RLE_CMD_BYTES (RLE_HEADER_BYTES + MIN_RLE_PIX_BYTES) + +#define RAW_HEADER_BYTES 6 +#define MIN_RAW_PIX_BYTES 2 +#define MIN_RAW_CMD_BYTES (RAW_HEADER_BYTES + MIN_RAW_PIX_BYTES) + +/* + * Trims identical data from front and back of line + * Sets new front buffer address and width + * And returns byte count of identical pixels + * Assumes CPU natural alignment (unsigned long) + * for back and front buffer ptrs and width + */ +#if 0 +static int udl_trim_hline(const u8 *bback, const u8 **bfront, int *width_bytes) +{ + int j, k; + const unsigned long *back = (const unsigned long *) bback; + const unsigned long *front = (const unsigned long *) *bfront; + const int width = *width_bytes / sizeof(unsigned long); + int identical = width; + int start = width; + int end = width; + + prefetch((void *) front); + prefetch((void *) back); + + for (j = 0; j < width; j++) { + if (back[j] != front[j]) { + start = j; + break; + } + } + + for (k = width - 1; k > j; k--) { + if (back[k] != front[k]) { + end = k+1; + break; + } + } + + identical = start + (width - end); + *bfront = (u8 *) &front[start]; + *width_bytes = (end - start) * sizeof(unsigned long); + + return identical * sizeof(unsigned long); +} +#endif + +static inline u16 pixel32_to_be16p(const uint8_t *pixel) +{ + uint32_t pix = *(uint32_t *)pixel; + u16 retval; + + retval = (((pix >> 3) & 0x001f) | + ((pix >> 5) & 0x07e0) | + ((pix >> 8) & 0xf800)); + return retval; +} + +/* + * Render a command stream for an encoded horizontal line segment of pixels. + * + * A command buffer holds several commands. + * It always begins with a fresh command header + * (the protocol doesn't require this, but we enforce it to allow + * multiple buffers to be potentially encoded and sent in parallel). + * A single command encodes one contiguous horizontal line of pixels + * + * The function relies on the client to do all allocation, so that + * rendering can be done directly to output buffers (e.g. USB URBs). + * The function fills the supplied command buffer, providing information + * on where it left off, so the client may call in again with additional + * buffers if the line will take several buffers to complete. + * + * A single command can transmit a maximum of 256 pixels, + * regardless of the compression ratio (protocol design limit). + * To the hardware, 0 for a size byte means 256 + * + * Rather than 256 pixel commands which are either rl or raw encoded, + * the rlx command simply assumes alternating raw and rl spans within one cmd. + * This has a slightly larger header overhead, but produces more even results. + * It also processes all data (read and write) in a single pass. + * Performance benchmarks of common cases show it having just slightly better + * compression than 256 pixel raw or rle commands, with similar CPU consumpion. + * But for very rl friendly data, will compress not quite as well. + */ +static void udl_compress_hline16( + const u8 **pixel_start_ptr, + const u8 *const pixel_end, + uint32_t *device_address_ptr, + uint8_t **command_buffer_ptr, + const uint8_t *const cmd_buffer_end, int bpp) +{ + const u8 *pixel = *pixel_start_ptr; + uint32_t dev_addr = *device_address_ptr; + uint8_t *cmd = *command_buffer_ptr; + + while ((pixel_end > pixel) && + (cmd_buffer_end - MIN_RLX_CMD_BYTES > cmd)) { + uint8_t *raw_pixels_count_byte = 0; + uint8_t *cmd_pixels_count_byte = 0; + const u8 *raw_pixel_start = 0; + const u8 *cmd_pixel_start, *cmd_pixel_end = 0; + + prefetchw((void *) cmd); /* pull in one cache line at least */ + + *cmd++ = 0xaf; + *cmd++ = 0x6b; + *cmd++ = (uint8_t) ((dev_addr >> 16) & 0xFF); + *cmd++ = (uint8_t) ((dev_addr >> 8) & 0xFF); + *cmd++ = (uint8_t) ((dev_addr) & 0xFF); + + cmd_pixels_count_byte = cmd++; /* we'll know this later */ + cmd_pixel_start = pixel; + + raw_pixels_count_byte = cmd++; /* we'll know this later */ + raw_pixel_start = pixel; + + cmd_pixel_end = pixel + (min(MAX_CMD_PIXELS + 1, + min((int)(pixel_end - pixel) / bpp, + (int)(cmd_buffer_end - cmd) / 2))) * bpp; + + prefetch_range((void *) pixel, (cmd_pixel_end - pixel) * bpp); + + while (pixel < cmd_pixel_end) { + const u8 * const repeating_pixel = pixel; + + if (bpp == 2) + *(uint16_t *)cmd = cpu_to_be16p((uint16_t *)pixel); + else if (bpp == 4) + *(uint16_t *)cmd = cpu_to_be16(pixel32_to_be16p(pixel)); + + cmd += 2; + pixel += bpp; + + if (unlikely((pixel < cmd_pixel_end) && + (!memcmp(pixel, repeating_pixel, bpp)))) { + /* go back and fill in raw pixel count */ + *raw_pixels_count_byte = (((repeating_pixel - + raw_pixel_start) / bpp) + 1) & 0xFF; + + while ((pixel < cmd_pixel_end) + && (!memcmp(pixel, repeating_pixel, bpp))) { + pixel += bpp; + } + + /* immediately after raw data is repeat byte */ + *cmd++ = (((pixel - repeating_pixel) / bpp) - 1) & 0xFF; + + /* Then start another raw pixel span */ + raw_pixel_start = pixel; + raw_pixels_count_byte = cmd++; + } + } + + if (pixel > raw_pixel_start) { + /* finalize last RAW span */ + *raw_pixels_count_byte = ((pixel-raw_pixel_start) / bpp) & 0xFF; + } + + *cmd_pixels_count_byte = ((pixel - cmd_pixel_start) / bpp) & 0xFF; + dev_addr += ((pixel - cmd_pixel_start) / bpp) * 2; + } + + if (cmd_buffer_end <= MIN_RLX_CMD_BYTES + cmd) { + /* Fill leftover bytes with no-ops */ + if (cmd_buffer_end > cmd) + memset(cmd, 0xAF, cmd_buffer_end - cmd); + cmd = (uint8_t *) cmd_buffer_end; + } + + *command_buffer_ptr = cmd; + *pixel_start_ptr = pixel; + *device_address_ptr = dev_addr; + + return; +} + +/* + * There are 3 copies of every pixel: The front buffer that the fbdev + * client renders to, the actual framebuffer across the USB bus in hardware + * (that we can only write to, slowly, and can never read), and (optionally) + * our shadow copy that tracks what's been sent to that hardware buffer. + */ +int udl_render_hline(struct drm_device *dev, int bpp, struct urb **urb_ptr, + const char *front, char **urb_buf_ptr, + u32 byte_offset, u32 device_byte_offset, + u32 byte_width, + int *ident_ptr, int *sent_ptr) +{ + const u8 *line_start, *line_end, *next_pixel; + u32 base16 = 0 + (device_byte_offset / bpp) * 2; + struct urb *urb = *urb_ptr; + u8 *cmd = *urb_buf_ptr; + u8 *cmd_end = (u8 *) urb->transfer_buffer + urb->transfer_buffer_length; + + line_start = (u8 *) (front + byte_offset); + next_pixel = line_start; + line_end = next_pixel + byte_width; + + while (next_pixel < line_end) { + + udl_compress_hline16(&next_pixel, + line_end, &base16, + (u8 **) &cmd, (u8 *) cmd_end, bpp); + + if (cmd >= cmd_end) { + int len = cmd - (u8 *) urb->transfer_buffer; + if (udl_submit_urb(dev, urb, len)) + return 1; /* lost pixels is set */ + *sent_ptr += len; + urb = udl_get_urb(dev); + if (!urb) + return 1; /* lost_pixels is set */ + *urb_ptr = urb; + cmd = urb->transfer_buffer; + cmd_end = &cmd[urb->transfer_buffer_length]; + } + } + + *urb_buf_ptr = cmd; + + return 0; +} + diff --git a/drivers/gpu/drm/via/via_drv.c b/drivers/gpu/drm/via/via_drv.c index a83e86d..02661f3 100644 --- a/drivers/gpu/drm/via/via_drv.c +++ b/drivers/gpu/drm/via/via_drv.c @@ -30,16 +30,52 @@ #include "drm_pciids.h" +static int via_driver_open(struct drm_device *dev, struct drm_file *file) +{ + struct via_file_private *file_priv; + + DRM_DEBUG_DRIVER("\n"); + file_priv = kmalloc(sizeof(*file_priv), GFP_KERNEL); + if (!file_priv) + return -ENOMEM; + + file->driver_priv = file_priv; + + INIT_LIST_HEAD(&file_priv->obj_list); + + return 0; +} + +void via_driver_postclose(struct drm_device *dev, struct drm_file *file) +{ + struct via_file_private *file_priv = file->driver_priv; + + kfree(file_priv); +} + static struct pci_device_id pciidlist[] = { viadrv_PCI_IDS }; +static const struct file_operations via_driver_fops = { + .owner = THIS_MODULE, + .open = drm_open, + .release = drm_release, + .unlocked_ioctl = drm_ioctl, + .mmap = drm_mmap, + .poll = drm_poll, + .fasync = drm_fasync, + .llseek = noop_llseek, +}; + static struct drm_driver driver = { .driver_features = DRIVER_USE_AGP | DRIVER_USE_MTRR | DRIVER_HAVE_IRQ | DRIVER_IRQ_SHARED, .load = via_driver_load, .unload = via_driver_unload, + .open = via_driver_open, + .postclose = via_driver_postclose, .context_dtor = via_final_context, .get_vblank_counter = via_get_vblank_counter, .enable_vblank = via_enable_vblank, @@ -54,17 +90,7 @@ static struct drm_driver driver = { .reclaim_buffers_idlelocked = via_reclaim_buffers_locked, .lastclose = via_lastclose, .ioctls = via_ioctls, - .fops = { - .owner = THIS_MODULE, - .open = drm_open, - .release = drm_release, - .unlocked_ioctl = drm_ioctl, - .mmap = drm_mmap, - .poll = drm_poll, - .fasync = drm_fasync, - .llseek = noop_llseek, - }, - + .fops = &via_driver_fops, .name = DRIVER_NAME, .desc = DRIVER_DESC, .date = DRIVER_DATE, diff --git a/drivers/gpu/drm/via/via_drv.h b/drivers/gpu/drm/via/via_drv.h index 9cf87d9..88edacc 100644 --- a/drivers/gpu/drm/via/via_drv.h +++ b/drivers/gpu/drm/via/via_drv.h @@ -24,7 +24,7 @@ #ifndef _VIA_DRV_H_ #define _VIA_DRV_H_ -#include "drm_sman.h" +#include "drm_mm.h" #define DRIVER_AUTHOR "Various" #define DRIVER_NAME "via" @@ -88,9 +88,12 @@ typedef struct drm_via_private { uint32_t irq_pending_mask; int *irq_map; unsigned int idle_fault; - struct drm_sman sman; int vram_initialized; + struct drm_mm vram_mm; int agp_initialized; + struct drm_mm agp_mm; + /** Mapping of userspace keys to mm objects */ + struct idr object_idr; unsigned long vram_offset; unsigned long agp_offset; drm_via_blitq_t blit_queues[VIA_NUM_BLIT_ENGINES]; diff --git a/drivers/gpu/drm/via/via_map.c b/drivers/gpu/drm/via/via_map.c index 6cca9a7..c126182 100644 --- a/drivers/gpu/drm/via/via_map.c +++ b/drivers/gpu/drm/via/via_map.c @@ -100,19 +100,15 @@ int via_driver_load(struct drm_device *dev, unsigned long chipset) if (dev_priv == NULL) return -ENOMEM; + idr_init(&dev_priv->object_idr); dev->dev_private = (void *)dev_priv; dev_priv->chipset = chipset; - ret = drm_sman_init(&dev_priv->sman, 2, 12, 8); - if (ret) { - kfree(dev_priv); - return ret; - } + pci_set_master(dev->pdev); ret = drm_vblank_init(dev, 1); if (ret) { - drm_sman_takedown(&dev_priv->sman); kfree(dev_priv); return ret; } @@ -124,7 +120,8 @@ int via_driver_unload(struct drm_device *dev) { drm_via_private_t *dev_priv = dev->dev_private; - drm_sman_takedown(&dev_priv->sman); + idr_remove_all(&dev_priv->object_idr); + idr_destroy(&dev_priv->object_idr); kfree(dev_priv); diff --git a/drivers/gpu/drm/via/via_mm.c b/drivers/gpu/drm/via/via_mm.c index 6cc2dad..a3574d0 100644 --- a/drivers/gpu/drm/via/via_mm.c +++ b/drivers/gpu/drm/via/via_mm.c @@ -28,26 +28,22 @@ #include "drmP.h" #include "via_drm.h" #include "via_drv.h" -#include "drm_sman.h" #define VIA_MM_ALIGN_SHIFT 4 #define VIA_MM_ALIGN_MASK ((1 << VIA_MM_ALIGN_SHIFT) - 1) +struct via_memblock { + struct drm_mm_node mm_node; + struct list_head owner_list; +}; + int via_agp_init(struct drm_device *dev, void *data, struct drm_file *file_priv) { drm_via_agp_t *agp = data; drm_via_private_t *dev_priv = (drm_via_private_t *) dev->dev_private; - int ret; mutex_lock(&dev->struct_mutex); - ret = drm_sman_set_range(&dev_priv->sman, VIA_MEM_AGP, 0, - agp->size >> VIA_MM_ALIGN_SHIFT); - - if (ret) { - DRM_ERROR("AGP memory manager initialisation error\n"); - mutex_unlock(&dev->struct_mutex); - return ret; - } + drm_mm_init(&dev_priv->agp_mm, 0, agp->size >> VIA_MM_ALIGN_SHIFT); dev_priv->agp_initialized = 1; dev_priv->agp_offset = agp->offset; @@ -61,17 +57,9 @@ int via_fb_init(struct drm_device *dev, void *data, struct drm_file *file_priv) { drm_via_fb_t *fb = data; drm_via_private_t *dev_priv = (drm_via_private_t *) dev->dev_private; - int ret; mutex_lock(&dev->struct_mutex); - ret = drm_sman_set_range(&dev_priv->sman, VIA_MEM_VIDEO, 0, - fb->size >> VIA_MM_ALIGN_SHIFT); - - if (ret) { - DRM_ERROR("VRAM memory manager initialisation error\n"); - mutex_unlock(&dev->struct_mutex); - return ret; - } + drm_mm_init(&dev_priv->vram_mm, 0, fb->size >> VIA_MM_ALIGN_SHIFT); dev_priv->vram_initialized = 1; dev_priv->vram_offset = fb->offset; @@ -108,19 +96,25 @@ void via_lastclose(struct drm_device *dev) return; mutex_lock(&dev->struct_mutex); - drm_sman_cleanup(&dev_priv->sman); - dev_priv->vram_initialized = 0; - dev_priv->agp_initialized = 0; + if (dev_priv->vram_initialized) { + drm_mm_takedown(&dev_priv->vram_mm); + dev_priv->vram_initialized = 0; + } + if (dev_priv->agp_initialized) { + drm_mm_takedown(&dev_priv->agp_mm); + dev_priv->agp_initialized = 0; + } mutex_unlock(&dev->struct_mutex); } int via_mem_alloc(struct drm_device *dev, void *data, - struct drm_file *file_priv) + struct drm_file *file) { drm_via_mem_t *mem = data; - int retval = 0; - struct drm_memblock_item *item; + int retval = 0, user_key; + struct via_memblock *item; drm_via_private_t *dev_priv = (drm_via_private_t *) dev->dev_private; + struct via_file_private *file_priv = file->driver_priv; unsigned long tmpSize; if (mem->type > VIA_MEM_AGP) { @@ -136,24 +130,57 @@ int via_mem_alloc(struct drm_device *dev, void *data, return -EINVAL; } + item = kzalloc(sizeof(*item), GFP_KERNEL); + if (!item) { + retval = -ENOMEM; + goto fail_alloc; + } + tmpSize = (mem->size + VIA_MM_ALIGN_MASK) >> VIA_MM_ALIGN_SHIFT; - item = drm_sman_alloc(&dev_priv->sman, mem->type, tmpSize, 0, - (unsigned long)file_priv); - mutex_unlock(&dev->struct_mutex); - if (item) { - mem->offset = ((mem->type == VIA_MEM_VIDEO) ? - dev_priv->vram_offset : dev_priv->agp_offset) + - (item->mm-> - offset(item->mm, item->mm_info) << VIA_MM_ALIGN_SHIFT); - mem->index = item->user_hash.key; - } else { - mem->offset = 0; - mem->size = 0; - mem->index = 0; - DRM_DEBUG("Video memory allocation failed\n"); + if (mem->type == VIA_MEM_AGP) + retval = drm_mm_insert_node(&dev_priv->agp_mm, + &item->mm_node, + tmpSize, 0); + else + retval = drm_mm_insert_node(&dev_priv->vram_mm, + &item->mm_node, + tmpSize, 0); + if (retval) + goto fail_alloc; + +again: + if (idr_pre_get(&dev_priv->object_idr, GFP_KERNEL) == 0) { retval = -ENOMEM; + goto fail_idr; } + retval = idr_get_new_above(&dev_priv->object_idr, item, 1, &user_key); + if (retval == -EAGAIN) + goto again; + if (retval) + goto fail_idr; + + list_add(&item->owner_list, &file_priv->obj_list); + mutex_unlock(&dev->struct_mutex); + + mem->offset = ((mem->type == VIA_MEM_VIDEO) ? + dev_priv->vram_offset : dev_priv->agp_offset) + + ((item->mm_node.start) << VIA_MM_ALIGN_SHIFT); + mem->index = user_key; + + return 0; + +fail_idr: + drm_mm_remove_node(&item->mm_node); +fail_alloc: + kfree(item); + mutex_unlock(&dev->struct_mutex); + + mem->offset = 0; + mem->size = 0; + mem->index = 0; + DRM_DEBUG("Video memory allocation failed\n"); + return retval; } @@ -161,24 +188,35 @@ int via_mem_free(struct drm_device *dev, void *data, struct drm_file *file_priv) { drm_via_private_t *dev_priv = dev->dev_private; drm_via_mem_t *mem = data; - int ret; + struct via_memblock *obj; mutex_lock(&dev->struct_mutex); - ret = drm_sman_free_key(&dev_priv->sman, mem->index); + obj = idr_find(&dev_priv->object_idr, mem->index); + if (obj == NULL) { + mutex_unlock(&dev->struct_mutex); + return -EINVAL; + } + + idr_remove(&dev_priv->object_idr, mem->index); + list_del(&obj->owner_list); + drm_mm_remove_node(&obj->mm_node); + kfree(obj); mutex_unlock(&dev->struct_mutex); + DRM_DEBUG("free = 0x%lx\n", mem->index); - return ret; + return 0; } void via_reclaim_buffers_locked(struct drm_device *dev, - struct drm_file *file_priv) + struct drm_file *file) { - drm_via_private_t *dev_priv = dev->dev_private; + struct via_file_private *file_priv = file->driver_priv; + struct via_memblock *entry, *next; mutex_lock(&dev->struct_mutex); - if (drm_sman_owner_clean(&dev_priv->sman, (unsigned long)file_priv)) { + if (list_empty(&file_priv->obj_list)) { mutex_unlock(&dev->struct_mutex); return; } @@ -186,7 +224,12 @@ void via_reclaim_buffers_locked(struct drm_device *dev, if (dev->driver->dma_quiescent) dev->driver->dma_quiescent(dev); - drm_sman_owner_cleanup(&dev_priv->sman, (unsigned long)file_priv); + list_for_each_entry_safe(entry, next, &file_priv->obj_list, + owner_list) { + list_del(&entry->owner_list); + drm_mm_remove_node(&entry->mm_node); + kfree(entry); + } mutex_unlock(&dev->struct_mutex); return; } diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c index 5a72ed9..1e2c0fb 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c @@ -28,6 +28,7 @@ #include "vmwgfx_drv.h" #include "ttm/ttm_bo_driver.h" #include "ttm/ttm_placement.h" +#include "ttm/ttm_page_alloc.h" static uint32_t vram_placement_flags = TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED; @@ -139,85 +140,63 @@ struct ttm_placement vmw_srf_placement = { .busy_placement = gmr_vram_placement_flags }; -struct vmw_ttm_backend { - struct ttm_backend backend; - struct page **pages; - unsigned long num_pages; +struct vmw_ttm_tt { + struct ttm_tt ttm; struct vmw_private *dev_priv; int gmr_id; }; -static int vmw_ttm_populate(struct ttm_backend *backend, - unsigned long num_pages, struct page **pages, - struct page *dummy_read_page, - dma_addr_t *dma_addrs) +static int vmw_ttm_bind(struct ttm_tt *ttm, struct ttm_mem_reg *bo_mem) { - struct vmw_ttm_backend *vmw_be = - container_of(backend, struct vmw_ttm_backend, backend); - - vmw_be->pages = pages; - vmw_be->num_pages = num_pages; - - return 0; -} - -static int vmw_ttm_bind(struct ttm_backend *backend, struct ttm_mem_reg *bo_mem) -{ - struct vmw_ttm_backend *vmw_be = - container_of(backend, struct vmw_ttm_backend, backend); + struct vmw_ttm_tt *vmw_be = container_of(ttm, struct vmw_ttm_tt, ttm); vmw_be->gmr_id = bo_mem->start; - return vmw_gmr_bind(vmw_be->dev_priv, vmw_be->pages, - vmw_be->num_pages, vmw_be->gmr_id); + return vmw_gmr_bind(vmw_be->dev_priv, ttm->pages, + ttm->num_pages, vmw_be->gmr_id); } -static int vmw_ttm_unbind(struct ttm_backend *backend) +static int vmw_ttm_unbind(struct ttm_tt *ttm) { - struct vmw_ttm_backend *vmw_be = - container_of(backend, struct vmw_ttm_backend, backend); + struct vmw_ttm_tt *vmw_be = container_of(ttm, struct vmw_ttm_tt, ttm); vmw_gmr_unbind(vmw_be->dev_priv, vmw_be->gmr_id); return 0; } -static void vmw_ttm_clear(struct ttm_backend *backend) +static void vmw_ttm_destroy(struct ttm_tt *ttm) { - struct vmw_ttm_backend *vmw_be = - container_of(backend, struct vmw_ttm_backend, backend); - - vmw_be->pages = NULL; - vmw_be->num_pages = 0; -} - -static void vmw_ttm_destroy(struct ttm_backend *backend) -{ - struct vmw_ttm_backend *vmw_be = - container_of(backend, struct vmw_ttm_backend, backend); + struct vmw_ttm_tt *vmw_be = container_of(ttm, struct vmw_ttm_tt, ttm); + ttm_tt_fini(ttm); kfree(vmw_be); } static struct ttm_backend_func vmw_ttm_func = { - .populate = vmw_ttm_populate, - .clear = vmw_ttm_clear, .bind = vmw_ttm_bind, .unbind = vmw_ttm_unbind, .destroy = vmw_ttm_destroy, }; -struct ttm_backend *vmw_ttm_backend_init(struct ttm_bo_device *bdev) +struct ttm_tt *vmw_ttm_tt_create(struct ttm_bo_device *bdev, + unsigned long size, uint32_t page_flags, + struct page *dummy_read_page) { - struct vmw_ttm_backend *vmw_be; + struct vmw_ttm_tt *vmw_be; vmw_be = kmalloc(sizeof(*vmw_be), GFP_KERNEL); if (!vmw_be) return NULL; - vmw_be->backend.func = &vmw_ttm_func; + vmw_be->ttm.func = &vmw_ttm_func; vmw_be->dev_priv = container_of(bdev, struct vmw_private, bdev); - return &vmw_be->backend; + if (ttm_tt_init(&vmw_be->ttm, bdev, size, page_flags, dummy_read_page)) { + kfree(vmw_be); + return NULL; + } + + return &vmw_be->ttm; } int vmw_invalidate_caches(struct ttm_bo_device *bdev, uint32_t flags) @@ -357,7 +336,9 @@ static int vmw_sync_obj_wait(void *sync_obj, void *sync_arg, } struct ttm_bo_driver vmw_bo_driver = { - .create_ttm_backend_entry = vmw_ttm_backend_init, + .ttm_tt_create = &vmw_ttm_tt_create, + .ttm_tt_populate = &ttm_pool_populate, + .ttm_tt_unpopulate = &ttm_pool_unpopulate, .invalidate_caches = vmw_invalidate_caches, .init_mem_type = vmw_init_mem_type, .evict_flags = vmw_evict_flags, diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c index b639536..db50604 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c @@ -38,6 +38,10 @@ #define VMWGFX_CHIP_SVGAII 0 #define VMW_FB_RESERVATION 0 +#define VMW_MIN_INITIAL_WIDTH 800 +#define VMW_MIN_INITIAL_HEIGHT 600 + + /** * Fully encoded drm commands. Might move to vmw_drm.h */ @@ -388,6 +392,41 @@ void vmw_3d_resource_dec(struct vmw_private *dev_priv, BUG_ON(n3d < 0); } +/** + * Sets the initial_[width|height] fields on the given vmw_private. + * + * It does so by reading SVGA_REG_[WIDTH|HEIGHT] regs and then + * clamping the value to fb_max_[width|height] fields and the + * VMW_MIN_INITIAL_[WIDTH|HEIGHT]. + * If the values appear to be invalid, set them to + * VMW_MIN_INITIAL_[WIDTH|HEIGHT]. + */ +static void vmw_get_initial_size(struct vmw_private *dev_priv) +{ + uint32_t width; + uint32_t height; + + width = vmw_read(dev_priv, SVGA_REG_WIDTH); + height = vmw_read(dev_priv, SVGA_REG_HEIGHT); + + width = max_t(uint32_t, width, VMW_MIN_INITIAL_WIDTH); + height = max_t(uint32_t, height, VMW_MIN_INITIAL_HEIGHT); + + if (width > dev_priv->fb_max_width || + height > dev_priv->fb_max_height) { + + /* + * This is a host error and shouldn't occur. + */ + + width = VMW_MIN_INITIAL_WIDTH; + height = VMW_MIN_INITIAL_HEIGHT; + } + + dev_priv->initial_width = width; + dev_priv->initial_height = height; +} + static int vmw_driver_load(struct drm_device *dev, unsigned long chipset) { struct vmw_private *dev_priv; @@ -401,6 +440,8 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset) } memset(dev_priv, 0, sizeof(*dev_priv)); + pci_set_master(dev->pdev); + dev_priv->dev = dev; dev_priv->vmw_chipset = chipset; dev_priv->last_read_seqno = (uint32_t) -100; @@ -431,7 +472,7 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset) svga_id = vmw_read(dev_priv, SVGA_REG_ID); if (svga_id != SVGA_ID_2) { ret = -ENOSYS; - DRM_ERROR("Unsuported SVGA ID 0x%x\n", svga_id); + DRM_ERROR("Unsupported SVGA ID 0x%x\n", svga_id); mutex_unlock(&dev_priv->hw_mutex); goto out_err0; } @@ -442,6 +483,9 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset) dev_priv->mmio_size = vmw_read(dev_priv, SVGA_REG_MEM_SIZE); dev_priv->fb_max_width = vmw_read(dev_priv, SVGA_REG_MAX_WIDTH); dev_priv->fb_max_height = vmw_read(dev_priv, SVGA_REG_MAX_HEIGHT); + + vmw_get_initial_size(dev_priv); + if (dev_priv->capabilities & SVGA_CAP_GMR) { dev_priv->max_gmr_descriptors = vmw_read(dev_priv, @@ -689,6 +733,15 @@ static int vmw_driver_unload(struct drm_device *dev) return 0; } +static void vmw_preclose(struct drm_device *dev, + struct drm_file *file_priv) +{ + struct vmw_fpriv *vmw_fp = vmw_fpriv(file_priv); + struct vmw_private *dev_priv = vmw_priv(dev); + + vmw_event_fence_fpriv_gone(dev_priv->fman, &vmw_fp->fence_events); +} + static void vmw_postclose(struct drm_device *dev, struct drm_file *file_priv) { @@ -711,6 +764,7 @@ static int vmw_driver_open(struct drm_device *dev, struct drm_file *file_priv) if (unlikely(vmw_fp == NULL)) return ret; + INIT_LIST_HEAD(&vmw_fp->fence_events); vmw_fp->tfile = ttm_object_file_init(dev_priv->tdev, 10); if (unlikely(vmw_fp->tfile == NULL)) goto out_no_tfile; @@ -1070,6 +1124,21 @@ static const struct dev_pm_ops vmw_pm_ops = { .resume = vmw_pm_resume, }; +static const struct file_operations vmwgfx_driver_fops = { + .owner = THIS_MODULE, + .open = drm_open, + .release = drm_release, + .unlocked_ioctl = vmw_unlocked_ioctl, + .mmap = vmw_mmap, + .poll = vmw_fops_poll, + .read = vmw_fops_read, + .fasync = drm_fasync, +#if defined(CONFIG_COMPAT) + .compat_ioctl = drm_compat_ioctl, +#endif + .llseek = noop_llseek, +}; + static struct drm_driver driver = { .driver_features = DRIVER_HAVE_IRQ | DRIVER_IRQ_SHARED | DRIVER_MODESET, @@ -1093,26 +1162,14 @@ static struct drm_driver driver = { .master_set = vmw_master_set, .master_drop = vmw_master_drop, .open = vmw_driver_open, + .preclose = vmw_preclose, .postclose = vmw_postclose, .dumb_create = vmw_dumb_create, .dumb_map_offset = vmw_dumb_map_offset, .dumb_destroy = vmw_dumb_destroy, - .fops = { - .owner = THIS_MODULE, - .open = drm_open, - .release = drm_release, - .unlocked_ioctl = vmw_unlocked_ioctl, - .mmap = vmw_mmap, - .poll = vmw_fops_poll, - .read = vmw_fops_read, - .fasync = drm_fasync, -#if defined(CONFIG_COMPAT) - .compat_ioctl = drm_compat_ioctl, -#endif - .llseek = noop_llseek, - }, + .fops = &vmwgfx_driver_fops, .name = VMWGFX_DRIVER_NAME, .desc = VMWGFX_DRIVER_DESC, .date = VMWGFX_DRIVER_DATE, diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h index 0e3fa7d..29c984f 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h @@ -40,9 +40,9 @@ #include "ttm/ttm_module.h" #include "vmwgfx_fence.h" -#define VMWGFX_DRIVER_DATE "20111025" +#define VMWGFX_DRIVER_DATE "20120209" #define VMWGFX_DRIVER_MAJOR 2 -#define VMWGFX_DRIVER_MINOR 3 +#define VMWGFX_DRIVER_MINOR 4 #define VMWGFX_DRIVER_PATCHLEVEL 0 #define VMWGFX_FILE_PAGE_OFFSET 0x00100000 #define VMWGFX_FIFO_STATIC_SIZE (1024*1024) @@ -62,6 +62,7 @@ struct vmw_fpriv { struct drm_master *locked_master; struct ttm_object_file *tfile; + struct list_head fence_events; }; struct vmw_dma_buffer { @@ -202,6 +203,8 @@ struct vmw_private { uint32_t mmio_size; uint32_t fb_max_width; uint32_t fb_max_height; + uint32_t initial_width; + uint32_t initial_height; __le32 __iomem *mmio_virt; int mmio_mtrr; uint32_t capabilities; @@ -533,7 +536,8 @@ extern int vmw_execbuf_process(struct drm_file *file_priv, uint32_t command_size, uint64_t throttle_us, struct drm_vmw_fence_rep __user - *user_fence_rep); + *user_fence_rep, + struct vmw_fence_obj **out_fence); extern void vmw_execbuf_release_pinned_bo(struct vmw_private *dev_priv, diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c index 40932fb..4acced4 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c @@ -1109,10 +1109,11 @@ int vmw_execbuf_process(struct drm_file *file_priv, void *kernel_commands, uint32_t command_size, uint64_t throttle_us, - struct drm_vmw_fence_rep __user *user_fence_rep) + struct drm_vmw_fence_rep __user *user_fence_rep, + struct vmw_fence_obj **out_fence) { struct vmw_sw_context *sw_context = &dev_priv->ctx; - struct vmw_fence_obj *fence; + struct vmw_fence_obj *fence = NULL; uint32_t handle; void *cmd; int ret; @@ -1208,8 +1209,13 @@ int vmw_execbuf_process(struct drm_file *file_priv, vmw_execbuf_copy_fence_user(dev_priv, vmw_fpriv(file_priv), ret, user_fence_rep, fence, handle); - if (likely(fence != NULL)) + /* Don't unreference when handing fence out */ + if (unlikely(out_fence != NULL)) { + *out_fence = fence; + fence = NULL; + } else if (likely(fence != NULL)) { vmw_fence_obj_unreference(&fence); + } mutex_unlock(&dev_priv->cmdbuf_mutex); return 0; @@ -1362,7 +1368,8 @@ int vmw_execbuf_ioctl(struct drm_device *dev, void *data, ret = vmw_execbuf_process(file_priv, dev_priv, (void __user *)(unsigned long)arg->commands, NULL, arg->command_size, arg->throttle_us, - (void __user *)(unsigned long)arg->fence_rep); + (void __user *)(unsigned long)arg->fence_rep, + NULL); if (unlikely(ret != 0)) goto out_unlock; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c index 34e51a1..3c447bf 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c @@ -414,10 +414,6 @@ int vmw_fb_init(struct vmw_private *vmw_priv) unsigned fb_bpp, fb_depth, fb_offset, fb_pitch, fb_size; int ret; - /* XXX These shouldn't be hardcoded. */ - initial_width = 800; - initial_height = 600; - fb_bpp = 32; fb_depth = 24; @@ -425,8 +421,8 @@ int vmw_fb_init(struct vmw_private *vmw_priv) fb_width = min(vmw_priv->fb_max_width, (unsigned)2048); fb_height = min(vmw_priv->fb_max_height, (unsigned)2048); - initial_width = min(fb_width, initial_width); - initial_height = min(fb_height, initial_height); + initial_width = min(vmw_priv->initial_width, fb_width); + initial_height = min(vmw_priv->initial_height, fb_height); fb_pitch = fb_width * fb_bpp / 8; fb_size = fb_pitch * fb_height; @@ -515,19 +511,7 @@ int vmw_fb_init(struct vmw_private *vmw_priv) info->var.xres = initial_width; info->var.yres = initial_height; -#if 0 - info->pixmap.size = 64*1024; - info->pixmap.buf_align = 8; - info->pixmap.access_align = 32; - info->pixmap.flags = FB_PIXMAP_SYSTEM; - info->pixmap.scan_align = 1; -#else - info->pixmap.size = 0; - info->pixmap.buf_align = 8; - info->pixmap.access_align = 32; - info->pixmap.flags = FB_PIXMAP_SYSTEM; - info->pixmap.scan_align = 1; -#endif + /* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */ info->apertures = alloc_apertures(1); if (!info->apertures) { diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c index 15fb260..7e07433 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c @@ -69,12 +69,13 @@ struct vmw_user_fence { * be assigned the current time tv_usec val when the fence signals. */ struct vmw_event_fence_action { - struct drm_pending_event e; struct vmw_fence_action action; + struct list_head fpriv_head; + + struct drm_pending_event *event; struct vmw_fence_obj *fence; struct drm_device *dev; - struct kref kref; - uint32_t size; + uint32_t *tv_sec; uint32_t *tv_usec; }; @@ -784,46 +785,40 @@ int vmw_fence_obj_unref_ioctl(struct drm_device *dev, void *data, } /** - * vmw_event_fence_action_destroy - * - * @kref: The struct kref embedded in a struct vmw_event_fence_action. - * - * The vmw_event_fence_action destructor that may be called either after - * the fence action cleanup, or when the event is delivered. - * It frees both the vmw_event_fence_action struct and the actual - * event structure copied to user-space. - */ -static void vmw_event_fence_action_destroy(struct kref *kref) -{ - struct vmw_event_fence_action *eaction = - container_of(kref, struct vmw_event_fence_action, kref); - struct ttm_mem_global *mem_glob = - vmw_mem_glob(vmw_priv(eaction->dev)); - uint32_t size = eaction->size; - - kfree(eaction->e.event); - kfree(eaction); - ttm_mem_global_free(mem_glob, size); -} - - -/** - * vmw_event_fence_action_delivered + * vmw_event_fence_fpriv_gone - Remove references to struct drm_file objects * - * @e: The struct drm_pending_event embedded in a struct - * vmw_event_fence_action. + * @fman: Pointer to a struct vmw_fence_manager + * @event_list: Pointer to linked list of struct vmw_event_fence_action objects + * with pointers to a struct drm_file object about to be closed. * - * The struct drm_pending_event destructor that is called by drm - * once the event is delivered. Since we don't know whether this function - * will be called before or after the fence action destructor, we - * free a refcount and destroy if it becomes zero. + * This function removes all pending fence events with references to a + * specific struct drm_file object about to be closed. The caller is required + * to pass a list of all struct vmw_event_fence_action objects with such + * events attached. This function is typically called before the + * struct drm_file object's event management is taken down. */ -static void vmw_event_fence_action_delivered(struct drm_pending_event *e) +void vmw_event_fence_fpriv_gone(struct vmw_fence_manager *fman, + struct list_head *event_list) { - struct vmw_event_fence_action *eaction = - container_of(e, struct vmw_event_fence_action, e); + struct vmw_event_fence_action *eaction; + struct drm_pending_event *event; + unsigned long irq_flags; - kref_put(&eaction->kref, vmw_event_fence_action_destroy); + while (1) { + spin_lock_irqsave(&fman->lock, irq_flags); + if (list_empty(event_list)) + goto out_unlock; + eaction = list_first_entry(event_list, + struct vmw_event_fence_action, + fpriv_head); + list_del_init(&eaction->fpriv_head); + event = eaction->event; + eaction->event = NULL; + spin_unlock_irqrestore(&fman->lock, irq_flags); + event->destroy(event); + } +out_unlock: + spin_unlock_irqrestore(&fman->lock, irq_flags); } @@ -836,18 +831,21 @@ static void vmw_event_fence_action_delivered(struct drm_pending_event *e) * This function is called when the seqno of the fence where @action is * attached has passed. It queues the event on the submitter's event list. * This function is always called from atomic context, and may be called - * from irq context. It ups a refcount reflecting that we now have two - * destructors. + * from irq context. */ static void vmw_event_fence_action_seq_passed(struct vmw_fence_action *action) { struct vmw_event_fence_action *eaction = container_of(action, struct vmw_event_fence_action, action); struct drm_device *dev = eaction->dev; - struct drm_file *file_priv = eaction->e.file_priv; + struct drm_pending_event *event = eaction->event; + struct drm_file *file_priv; unsigned long irq_flags; - kref_get(&eaction->kref); + if (unlikely(event == NULL)) + return; + + file_priv = event->file_priv; spin_lock_irqsave(&dev->event_lock, irq_flags); if (likely(eaction->tv_sec != NULL)) { @@ -858,7 +856,9 @@ static void vmw_event_fence_action_seq_passed(struct vmw_fence_action *action) *eaction->tv_usec = tv.tv_usec; } - list_add_tail(&eaction->e.link, &file_priv->event_list); + list_del_init(&eaction->fpriv_head); + list_add_tail(&eaction->event->link, &file_priv->event_list); + eaction->event = NULL; wake_up_all(&file_priv->event_wait); spin_unlock_irqrestore(&dev->event_lock, irq_flags); } @@ -876,9 +876,15 @@ static void vmw_event_fence_action_cleanup(struct vmw_fence_action *action) { struct vmw_event_fence_action *eaction = container_of(action, struct vmw_event_fence_action, action); + struct vmw_fence_manager *fman = eaction->fence->fman; + unsigned long irq_flags; + + spin_lock_irqsave(&fman->lock, irq_flags); + list_del(&eaction->fpriv_head); + spin_unlock_irqrestore(&fman->lock, irq_flags); vmw_fence_obj_unreference(&eaction->fence); - kref_put(&eaction->kref, vmw_event_fence_action_destroy); + kfree(eaction); } @@ -946,39 +952,23 @@ void vmw_fence_obj_add_action(struct vmw_fence_obj *fence, * an error code, the caller needs to free that object. */ -int vmw_event_fence_action_create(struct drm_file *file_priv, - struct vmw_fence_obj *fence, - struct drm_event *event, - uint32_t *tv_sec, - uint32_t *tv_usec, - bool interruptible) +int vmw_event_fence_action_queue(struct drm_file *file_priv, + struct vmw_fence_obj *fence, + struct drm_pending_event *event, + uint32_t *tv_sec, + uint32_t *tv_usec, + bool interruptible) { struct vmw_event_fence_action *eaction; - struct ttm_mem_global *mem_glob = - vmw_mem_glob(fence->fman->dev_priv); struct vmw_fence_manager *fman = fence->fman; - uint32_t size = fman->event_fence_action_size + - ttm_round_pot(event->length); - int ret; - - /* - * Account for internal structure size as well as the - * event size itself. - */ - - ret = ttm_mem_global_alloc(mem_glob, size, false, interruptible); - if (unlikely(ret != 0)) - return ret; + struct vmw_fpriv *vmw_fp = vmw_fpriv(file_priv); + unsigned long irq_flags; eaction = kzalloc(sizeof(*eaction), GFP_KERNEL); - if (unlikely(eaction == NULL)) { - ttm_mem_global_free(mem_glob, size); + if (unlikely(eaction == NULL)) return -ENOMEM; - } - eaction->e.event = event; - eaction->e.file_priv = file_priv; - eaction->e.destroy = vmw_event_fence_action_delivered; + eaction->event = event; eaction->action.seq_passed = vmw_event_fence_action_seq_passed; eaction->action.cleanup = vmw_event_fence_action_cleanup; @@ -986,16 +976,89 @@ int vmw_event_fence_action_create(struct drm_file *file_priv, eaction->fence = vmw_fence_obj_reference(fence); eaction->dev = fman->dev_priv->dev; - eaction->size = size; eaction->tv_sec = tv_sec; eaction->tv_usec = tv_usec; - kref_init(&eaction->kref); + spin_lock_irqsave(&fman->lock, irq_flags); + list_add_tail(&eaction->fpriv_head, &vmw_fp->fence_events); + spin_unlock_irqrestore(&fman->lock, irq_flags); + vmw_fence_obj_add_action(fence, &eaction->action); return 0; } +struct vmw_event_fence_pending { + struct drm_pending_event base; + struct drm_vmw_event_fence event; +}; + +int vmw_event_fence_action_create(struct drm_file *file_priv, + struct vmw_fence_obj *fence, + uint32_t flags, + uint64_t user_data, + bool interruptible) +{ + struct vmw_event_fence_pending *event; + struct drm_device *dev = fence->fman->dev_priv->dev; + unsigned long irq_flags; + int ret; + + spin_lock_irqsave(&dev->event_lock, irq_flags); + + ret = (file_priv->event_space < sizeof(event->event)) ? -EBUSY : 0; + if (likely(ret == 0)) + file_priv->event_space -= sizeof(event->event); + + spin_unlock_irqrestore(&dev->event_lock, irq_flags); + + if (unlikely(ret != 0)) { + DRM_ERROR("Failed to allocate event space for this file.\n"); + goto out_no_space; + } + + + event = kzalloc(sizeof(*event), GFP_KERNEL); + if (unlikely(event == NULL)) { + DRM_ERROR("Failed to allocate an event.\n"); + ret = -ENOMEM; + goto out_no_event; + } + + event->event.base.type = DRM_VMW_EVENT_FENCE_SIGNALED; + event->event.base.length = sizeof(*event); + event->event.user_data = user_data; + + event->base.event = &event->event.base; + event->base.file_priv = file_priv; + event->base.destroy = (void (*) (struct drm_pending_event *)) kfree; + + + if (flags & DRM_VMW_FE_FLAG_REQ_TIME) + ret = vmw_event_fence_action_queue(file_priv, fence, + &event->base, + &event->event.tv_sec, + &event->event.tv_usec, + interruptible); + else + ret = vmw_event_fence_action_queue(file_priv, fence, + &event->base, + NULL, + NULL, + interruptible); + if (ret != 0) + goto out_no_queue; + +out_no_queue: + event->base.destroy(&event->base); +out_no_event: + spin_lock_irqsave(&dev->event_lock, irq_flags); + file_priv->event_space += sizeof(*event); + spin_unlock_irqrestore(&dev->event_lock, irq_flags); +out_no_space: + return ret; +} + int vmw_fence_event_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { @@ -1008,8 +1071,6 @@ int vmw_fence_event_ioctl(struct drm_device *dev, void *data, (struct drm_vmw_fence_rep __user *)(unsigned long) arg->fence_rep; uint32_t handle; - unsigned long irq_flags; - struct drm_vmw_event_fence *event; int ret; /* @@ -1062,59 +1123,28 @@ int vmw_fence_event_ioctl(struct drm_device *dev, void *data, BUG_ON(fence == NULL); - spin_lock_irqsave(&dev->event_lock, irq_flags); - - ret = (file_priv->event_space < sizeof(*event)) ? -EBUSY : 0; - if (likely(ret == 0)) - file_priv->event_space -= sizeof(*event); - - spin_unlock_irqrestore(&dev->event_lock, irq_flags); - - if (unlikely(ret != 0)) { - DRM_ERROR("Failed to allocate event space for this file.\n"); - goto out_no_event_space; - } - - event = kzalloc(sizeof(*event), GFP_KERNEL); - if (unlikely(event == NULL)) { - DRM_ERROR("Failed to allocate an event.\n"); - goto out_no_event; - } - - event->base.type = DRM_VMW_EVENT_FENCE_SIGNALED; - event->base.length = sizeof(*event); - event->user_data = arg->user_data; - if (arg->flags & DRM_VMW_FE_FLAG_REQ_TIME) ret = vmw_event_fence_action_create(file_priv, fence, - &event->base, - &event->tv_sec, - &event->tv_usec, + arg->flags, + arg->user_data, true); else ret = vmw_event_fence_action_create(file_priv, fence, - &event->base, - NULL, - NULL, + arg->flags, + arg->user_data, true); if (unlikely(ret != 0)) { if (ret != -ERESTARTSYS) DRM_ERROR("Failed to attach event to fence.\n"); - goto out_no_attach; + goto out_no_create; } vmw_execbuf_copy_fence_user(dev_priv, vmw_fp, 0, user_fence_rep, fence, handle); vmw_fence_obj_unreference(&fence); return 0; -out_no_attach: - kfree(event); -out_no_event: - spin_lock_irqsave(&dev->event_lock, irq_flags); - file_priv->event_space += sizeof(*event); - spin_unlock_irqrestore(&dev->event_lock, irq_flags); -out_no_event_space: +out_no_create: if (user_fence_rep != NULL) ttm_ref_object_base_unref(vmw_fpriv(file_priv)->tfile, handle, TTM_REF_USAGE); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h index 0854a20..faf2e78 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h @@ -109,5 +109,12 @@ extern int vmw_fence_obj_unref_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); extern int vmw_fence_event_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); - +extern void vmw_event_fence_fpriv_gone(struct vmw_fence_manager *fman, + struct list_head *event_list); +extern int vmw_event_fence_action_queue(struct drm_file *filee_priv, + struct vmw_fence_obj *fence, + struct drm_pending_event *event, + uint32_t *tv_sec, + uint32_t *tv_usec, + bool interruptible); #endif /* _VMWGFX_FENCE_H_ */ diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_gmr.c b/drivers/gpu/drm/vmwgfx/vmwgfx_gmr.c index c41226a..21ee782 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_gmr.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_gmr.c @@ -136,10 +136,10 @@ static int vmw_gmr_build_descriptors(struct list_head *desc_pages, if (likely(page_virtual != NULL)) { desc_virtual->ppn = page_to_pfn(page); - kunmap_atomic(page_virtual, KM_USER0); + kunmap_atomic(page_virtual); } - page_virtual = kmap_atomic(page, KM_USER0); + page_virtual = kmap_atomic(page); desc_virtual = page_virtual - 1; prev_pfn = ~(0UL); @@ -169,7 +169,7 @@ static int vmw_gmr_build_descriptors(struct list_head *desc_pages, } if (likely(page_virtual != NULL)) - kunmap_atomic(page_virtual, KM_USER0); + kunmap_atomic(page_virtual); return 0; out_err: diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c index 7c88f1f..00fb5aa 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c @@ -422,7 +422,8 @@ static int do_surface_dirty_sou(struct vmw_private *dev_priv, struct vmw_framebuffer *framebuffer, unsigned flags, unsigned color, struct drm_clip_rect *clips, - unsigned num_clips, int inc) + unsigned num_clips, int inc, + struct vmw_fence_obj **out_fence) { struct vmw_display_unit *units[VMWGFX_NUM_DISPLAY_UNITS]; struct drm_clip_rect *clips_ptr; @@ -542,12 +543,15 @@ static int do_surface_dirty_sou(struct vmw_private *dev_priv, if (num == 0) continue; + /* only return the last fence */ + if (out_fence && *out_fence) + vmw_fence_obj_unreference(out_fence); /* recalculate package length */ fifo_size = sizeof(*cmd) + sizeof(SVGASignedRect) * num; cmd->header.size = cpu_to_le32(fifo_size - sizeof(cmd->header)); ret = vmw_execbuf_process(file_priv, dev_priv, NULL, cmd, - fifo_size, 0, NULL); + fifo_size, 0, NULL, out_fence); if (unlikely(ret != 0)) break; @@ -598,7 +602,7 @@ int vmw_framebuffer_surface_dirty(struct drm_framebuffer *framebuffer, ret = do_surface_dirty_sou(dev_priv, file_priv, &vfbs->base, flags, color, - clips, num_clips, inc); + clips, num_clips, inc, NULL); ttm_read_unlock(&vmaster->lock); return 0; @@ -690,7 +694,7 @@ static int vmw_kms_new_framebuffer_surface(struct vmw_private *dev_priv, /* XXX get the first 3 from the surface info */ vfbs->base.base.bits_per_pixel = mode_cmd->bpp; - vfbs->base.base.pitch = mode_cmd->pitch; + vfbs->base.base.pitches[0] = mode_cmd->pitch; vfbs->base.base.depth = mode_cmd->depth; vfbs->base.base.width = mode_cmd->width; vfbs->base.base.height = mode_cmd->height; @@ -804,12 +808,12 @@ static int do_dmabuf_define_gmrfb(struct drm_file *file_priv, cmd->body.format.bitsPerPixel = framebuffer->base.bits_per_pixel; cmd->body.format.colorDepth = depth; cmd->body.format.reserved = 0; - cmd->body.bytesPerLine = framebuffer->base.pitch; + cmd->body.bytesPerLine = framebuffer->base.pitches[0]; cmd->body.ptr.gmrId = framebuffer->user_handle; cmd->body.ptr.offset = 0; ret = vmw_execbuf_process(file_priv, dev_priv, NULL, cmd, - fifo_size, 0, NULL); + fifo_size, 0, NULL, NULL); kfree(cmd); @@ -821,7 +825,8 @@ static int do_dmabuf_dirty_sou(struct drm_file *file_priv, struct vmw_framebuffer *framebuffer, unsigned flags, unsigned color, struct drm_clip_rect *clips, - unsigned num_clips, int increment) + unsigned num_clips, int increment, + struct vmw_fence_obj **out_fence) { struct vmw_display_unit *units[VMWGFX_NUM_DISPLAY_UNITS]; struct drm_clip_rect *clips_ptr; @@ -894,9 +899,13 @@ static int do_dmabuf_dirty_sou(struct drm_file *file_priv, if (hit_num == 0) continue; + /* only return the last fence */ + if (out_fence && *out_fence) + vmw_fence_obj_unreference(out_fence); + fifo_size = sizeof(*blits) * hit_num; ret = vmw_execbuf_process(file_priv, dev_priv, NULL, blits, - fifo_size, 0, NULL); + fifo_size, 0, NULL, out_fence); if (unlikely(ret != 0)) break; @@ -942,7 +951,7 @@ int vmw_framebuffer_dmabuf_dirty(struct drm_framebuffer *framebuffer, } else { ret = do_dmabuf_dirty_sou(file_priv, dev_priv, &vfbd->base, flags, color, - clips, num_clips, increment); + clips, num_clips, increment, NULL); } ttm_read_unlock(&vmaster->lock); @@ -1056,7 +1065,7 @@ static int vmw_kms_new_framebuffer_dmabuf(struct vmw_private *dev_priv, } vfbd->base.base.bits_per_pixel = mode_cmd->bpp; - vfbd->base.base.pitch = mode_cmd->pitch; + vfbd->base.base.pitches[0] = mode_cmd->pitch; vfbd->base.base.depth = mode_cmd->depth; vfbd->base.base.width = mode_cmd->width; vfbd->base.base.height = mode_cmd->height; @@ -1085,7 +1094,7 @@ out_err1: static struct drm_framebuffer *vmw_kms_fb_create(struct drm_device *dev, struct drm_file *file_priv, - struct drm_mode_fb_cmd *mode_cmd) + struct drm_mode_fb_cmd2 *mode_cmd2) { struct vmw_private *dev_priv = vmw_priv(dev); struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile; @@ -1093,8 +1102,16 @@ static struct drm_framebuffer *vmw_kms_fb_create(struct drm_device *dev, struct vmw_surface *surface = NULL; struct vmw_dma_buffer *bo = NULL; struct ttm_base_object *user_obj; + struct drm_mode_fb_cmd mode_cmd; int ret; + mode_cmd.width = mode_cmd2->width; + mode_cmd.height = mode_cmd2->height; + mode_cmd.pitch = mode_cmd2->pitches[0]; + mode_cmd.handle = mode_cmd2->handles[0]; + drm_fb_get_bpp_depth(mode_cmd2->pixel_format, &mode_cmd.depth, + &mode_cmd.bpp); + /** * This code should be conditioned on Screen Objects not being used. * If screen objects are used, we can allocate a GMR to hold the @@ -1102,8 +1119,8 @@ static struct drm_framebuffer *vmw_kms_fb_create(struct drm_device *dev, */ if (!vmw_kms_validate_mode_vram(dev_priv, - mode_cmd->pitch, - mode_cmd->height)) { + mode_cmd.pitch, + mode_cmd.height)) { DRM_ERROR("VRAM size is too small for requested mode.\n"); return ERR_PTR(-ENOMEM); } @@ -1117,15 +1134,19 @@ static struct drm_framebuffer *vmw_kms_fb_create(struct drm_device *dev, * command stream using user-space handles. */ - user_obj = ttm_base_object_lookup(tfile, mode_cmd->handle); + user_obj = ttm_base_object_lookup(tfile, mode_cmd.handle); if (unlikely(user_obj == NULL)) { DRM_ERROR("Could not locate requested kms frame buffer.\n"); return ERR_PTR(-ENOENT); } + /** + * End conditioned code. + */ + /* returns either a dmabuf or surface */ ret = vmw_user_lookup_handle(dev_priv, tfile, - mode_cmd->handle, + mode_cmd.handle, &surface, &bo); if (ret) goto err_out; @@ -1133,10 +1154,10 @@ static struct drm_framebuffer *vmw_kms_fb_create(struct drm_device *dev, /* Create the new framebuffer depending one what we got back */ if (bo) ret = vmw_kms_new_framebuffer_dmabuf(dev_priv, bo, &vfb, - mode_cmd); + &mode_cmd); else if (surface) ret = vmw_kms_new_framebuffer_surface(dev_priv, file_priv, - surface, &vfb, mode_cmd); + surface, &vfb, &mode_cmd); else BUG(); @@ -1284,7 +1305,7 @@ int vmw_kms_present(struct vmw_private *dev_priv, fifo_size = sizeof(*cmd) + sizeof(SVGASignedRect) * num; cmd->header.size = cpu_to_le32(fifo_size - sizeof(cmd->header)); ret = vmw_execbuf_process(file_priv, dev_priv, NULL, cmd, - fifo_size, 0, NULL); + fifo_size, 0, NULL, NULL); if (unlikely(ret != 0)) break; @@ -1344,7 +1365,7 @@ int vmw_kms_readback(struct vmw_private *dev_priv, cmd->body.format.bitsPerPixel = vfb->base.bits_per_pixel; cmd->body.format.colorDepth = vfb->base.depth; cmd->body.format.reserved = 0; - cmd->body.bytesPerLine = vfb->base.pitch; + cmd->body.bytesPerLine = vfb->base.pitches[0]; cmd->body.ptr.gmrId = vfb->user_handle; cmd->body.ptr.offset = 0; @@ -1397,7 +1418,7 @@ int vmw_kms_readback(struct vmw_private *dev_priv, fifo_size = sizeof(*cmd) + sizeof(*blits) * blits_pos; ret = vmw_execbuf_process(file_priv, dev_priv, NULL, cmd, fifo_size, - 0, user_fence_rep); + 0, user_fence_rep, NULL); kfree(cmd); @@ -1660,6 +1681,74 @@ int vmw_du_update_layout(struct vmw_private *dev_priv, unsigned num, return 0; } +int vmw_du_page_flip(struct drm_crtc *crtc, + struct drm_framebuffer *fb, + struct drm_pending_vblank_event *event) +{ + struct vmw_private *dev_priv = vmw_priv(crtc->dev); + struct drm_framebuffer *old_fb = crtc->fb; + struct vmw_framebuffer *vfb = vmw_framebuffer_to_vfb(fb); + struct drm_file *file_priv ; + struct vmw_fence_obj *fence = NULL; + struct drm_clip_rect clips; + int ret; + + if (event == NULL) + return -EINVAL; + + /* require ScreenObject support for page flipping */ + if (!dev_priv->sou_priv) + return -ENOSYS; + + file_priv = event->base.file_priv; + if (!vmw_kms_screen_object_flippable(dev_priv, crtc)) + return -EINVAL; + + crtc->fb = fb; + + /* do a full screen dirty update */ + clips.x1 = clips.y1 = 0; + clips.x2 = fb->width; + clips.y2 = fb->height; + + if (vfb->dmabuf) + ret = do_dmabuf_dirty_sou(file_priv, dev_priv, vfb, + 0, 0, &clips, 1, 1, &fence); + else + ret = do_surface_dirty_sou(dev_priv, file_priv, vfb, + 0, 0, &clips, 1, 1, &fence); + + + if (ret != 0) + goto out_no_fence; + if (!fence) { + ret = -EINVAL; + goto out_no_fence; + } + + ret = vmw_event_fence_action_queue(file_priv, fence, + &event->base, + &event->event.tv_sec, + &event->event.tv_usec, + true); + + /* + * No need to hold on to this now. The only cleanup + * we need to do if we fail is unref the fence. + */ + vmw_fence_obj_unreference(&fence); + + if (vmw_crtc_to_du(crtc)->is_implicit) + vmw_kms_screen_object_update_implicit_fb(dev_priv, crtc); + + return ret; + +out_no_fence: + crtc->fb = old_fb; + return ret; +} + + void vmw_du_crtc_save(struct drm_crtc *crtc) { } diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h index e1cb855..8184bc5 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h @@ -29,6 +29,7 @@ #define VMWGFX_KMS_H_ #include "drmP.h" +#include "drm_crtc_helper.h" #include "vmwgfx_drv.h" #define VMWGFX_NUM_DISPLAY_UNITS 8 @@ -120,6 +121,9 @@ struct vmw_display_unit { * Shared display unit functions - vmwgfx_kms.c */ void vmw_display_unit_cleanup(struct vmw_display_unit *du); +int vmw_du_page_flip(struct drm_crtc *crtc, + struct drm_framebuffer *fb, + struct drm_pending_vblank_event *event); void vmw_du_crtc_save(struct drm_crtc *crtc); void vmw_du_crtc_restore(struct drm_crtc *crtc); void vmw_du_crtc_gamma_set(struct drm_crtc *crtc, @@ -153,5 +157,10 @@ int vmw_kms_init_screen_object_display(struct vmw_private *dev_priv); int vmw_kms_close_screen_object_display(struct vmw_private *dev_priv); int vmw_kms_sou_update_layout(struct vmw_private *dev_priv, unsigned num, struct drm_vmw_rect *rects); +bool vmw_kms_screen_object_flippable(struct vmw_private *dev_priv, + struct drm_crtc *crtc); +void vmw_kms_screen_object_update_implicit_fb(struct vmw_private *dev_priv, + struct drm_crtc *crtc); + #endif diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c index 8f8dbd4..070fb23 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c @@ -95,7 +95,7 @@ static int vmw_ldu_commit_list(struct vmw_private *dev_priv) return 0; fb = entry->base.crtc.fb; - return vmw_kms_write_svga(dev_priv, w, h, fb->pitch, + return vmw_kms_write_svga(dev_priv, w, h, fb->pitches[0], fb->bits_per_pixel, fb->depth); } @@ -103,7 +103,7 @@ static int vmw_ldu_commit_list(struct vmw_private *dev_priv) entry = list_entry(lds->active.next, typeof(*entry), active); fb = entry->base.crtc.fb; - vmw_kms_write_svga(dev_priv, fb->width, fb->height, fb->pitch, + vmw_kms_write_svga(dev_priv, fb->width, fb->height, fb->pitches[0], fb->bits_per_pixel, fb->depth); } @@ -354,8 +354,8 @@ static int vmw_ldu_init(struct vmw_private *dev_priv, unsigned unit) INIT_LIST_HEAD(&ldu->active); ldu->base.pref_active = (unit == 0); - ldu->base.pref_width = 800; - ldu->base.pref_height = 600; + ldu->base.pref_width = dev_priv->initial_width; + ldu->base.pref_height = dev_priv->initial_height; ldu->base.pref_mode = NULL; ldu->base.is_implicit = true; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c index 0795d17..059b32c 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c @@ -1540,29 +1540,10 @@ out_bad_surface: /** * Buffer management. */ - -static size_t vmw_dmabuf_acc_size(struct ttm_bo_global *glob, - unsigned long num_pages) -{ - static size_t bo_user_size = ~0; - - size_t page_array_size = - (num_pages * sizeof(void *) + PAGE_SIZE - 1) & PAGE_MASK; - - if (unlikely(bo_user_size == ~0)) { - bo_user_size = glob->ttm_bo_extra_size + - ttm_round_pot(sizeof(struct vmw_dma_buffer)); - } - - return bo_user_size + page_array_size; -} - void vmw_dmabuf_bo_free(struct ttm_buffer_object *bo) { struct vmw_dma_buffer *vmw_bo = vmw_dma_buffer(bo); - struct ttm_bo_global *glob = bo->glob; - ttm_mem_global_free(glob->mem_glob, bo->acc_size); kfree(vmw_bo); } @@ -1573,24 +1554,12 @@ int vmw_dmabuf_init(struct vmw_private *dev_priv, void (*bo_free) (struct ttm_buffer_object *bo)) { struct ttm_bo_device *bdev = &dev_priv->bdev; - struct ttm_mem_global *mem_glob = bdev->glob->mem_glob; size_t acc_size; int ret; BUG_ON(!bo_free); - acc_size = - vmw_dmabuf_acc_size(bdev->glob, - (size + PAGE_SIZE - 1) >> PAGE_SHIFT); - - ret = ttm_mem_global_alloc(mem_glob, acc_size, false, false); - if (unlikely(ret != 0)) { - /* we must free the bo here as - * ttm_buffer_object_init does so as well */ - bo_free(&vmw_bo->base); - return ret; - } - + acc_size = ttm_bo_acc_size(bdev, size, sizeof(struct vmw_dma_buffer)); memset(vmw_bo, 0, sizeof(*vmw_bo)); INIT_LIST_HEAD(&vmw_bo->validate_list); @@ -1605,9 +1574,7 @@ int vmw_dmabuf_init(struct vmw_private *dev_priv, static void vmw_user_dmabuf_destroy(struct ttm_buffer_object *bo) { struct vmw_user_dma_buffer *vmw_user_bo = vmw_user_dma_buffer(bo); - struct ttm_bo_global *glob = bo->glob; - ttm_mem_global_free(glob->mem_glob, bo->acc_size); kfree(vmw_user_bo); } diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c b/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c index 4defdcf..6deaf2f 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c @@ -394,6 +394,7 @@ static struct drm_crtc_funcs vmw_screen_object_crtc_funcs = { .gamma_set = vmw_du_crtc_gamma_set, .destroy = vmw_sou_crtc_destroy, .set_config = vmw_sou_crtc_set_config, + .page_flip = vmw_du_page_flip, }; /* @@ -448,8 +449,8 @@ static int vmw_sou_init(struct vmw_private *dev_priv, unsigned unit) sou->active_implicit = false; sou->base.pref_active = (unit == 0); - sou->base.pref_width = 800; - sou->base.pref_height = 600; + sou->base.pref_width = dev_priv->initial_width; + sou->base.pref_height = dev_priv->initial_height; sou->base.pref_mode = NULL; sou->base.is_implicit = true; @@ -535,3 +536,36 @@ int vmw_kms_close_screen_object_display(struct vmw_private *dev_priv) return 0; } + +/** + * Returns if this unit can be page flipped. + * Must be called with the mode_config mutex held. + */ +bool vmw_kms_screen_object_flippable(struct vmw_private *dev_priv, + struct drm_crtc *crtc) +{ + struct vmw_screen_object_unit *sou = vmw_crtc_to_sou(crtc); + + if (!sou->base.is_implicit) + return true; + + if (dev_priv->sou_priv->num_implicit != 1) + return false; + + return true; +} + +/** + * Update the implicit fb to the current fb of this crtc. + * Must be called with the mode_config mutex held. + */ +void vmw_kms_screen_object_update_implicit_fb(struct vmw_private *dev_priv, + struct drm_crtc *crtc) +{ + struct vmw_screen_object_unit *sou = vmw_crtc_to_sou(crtc); + + BUG_ON(!sou->base.is_implicit); + + dev_priv->sou_priv->implicit_fb = + vmw_framebuffer_to_vfb(sou->base.crtc.fb); +} diff --git a/include/drm/Kbuild b/include/drm/Kbuild index 3a60ac8..1e38a19 100644 --- a/include/drm/Kbuild +++ b/include/drm/Kbuild @@ -1,6 +1,8 @@ header-y += drm.h +header-y += drm_fourcc.h header-y += drm_mode.h header-y += drm_sarea.h +header-y += exynos_drm.h header-y += i810_drm.h header-y += i915_drm.h header-y += mga_drm.h diff --git a/include/drm/drm.h b/include/drm/drm.h index 4be33b4..64ff02d 100644 --- a/include/drm/drm.h +++ b/include/drm/drm.h @@ -617,6 +617,17 @@ struct drm_get_cap { __u64 value; }; +#define DRM_CLOEXEC O_CLOEXEC +struct drm_prime_handle { + __u32 handle; + + /** Flags.. only applicable for handle->fd */ + __u32 flags; + + /** Returned dmabuf file descriptor */ + __s32 fd; +}; + #include "drm_mode.h" #define DRM_IOCTL_BASE 'd' @@ -673,7 +684,8 @@ struct drm_get_cap { #define DRM_IOCTL_UNLOCK DRM_IOW( 0x2b, struct drm_lock) #define DRM_IOCTL_FINISH DRM_IOW( 0x2c, struct drm_lock) -#define DRM_IOCTL_GEM_PRIME_OPEN DRM_IOWR(0x2e, struct drm_gem_open) +#define DRM_IOCTL_PRIME_HANDLE_TO_FD DRM_IOWR(0x2d, struct drm_prime_handle) +#define DRM_IOCTL_PRIME_FD_TO_HANDLE DRM_IOWR(0x2e, struct drm_prime_handle) #define DRM_IOCTL_AGP_ACQUIRE DRM_IO( 0x30) #define DRM_IOCTL_AGP_RELEASE DRM_IO( 0x31) @@ -714,6 +726,10 @@ struct drm_get_cap { #define DRM_IOCTL_MODE_CREATE_DUMB DRM_IOWR(0xB2, struct drm_mode_create_dumb) #define DRM_IOCTL_MODE_MAP_DUMB DRM_IOWR(0xB3, struct drm_mode_map_dumb) #define DRM_IOCTL_MODE_DESTROY_DUMB DRM_IOWR(0xB4, struct drm_mode_destroy_dumb) +#define DRM_IOCTL_MODE_GETPLANERESOURCES DRM_IOWR(0xB5, struct drm_mode_get_plane_res) +#define DRM_IOCTL_MODE_GETPLANE DRM_IOWR(0xB6, struct drm_mode_get_plane) +#define DRM_IOCTL_MODE_SETPLANE DRM_IOWR(0xB7, struct drm_mode_set_plane) +#define DRM_IOCTL_MODE_ADDFB2 DRM_IOWR(0xB8, struct drm_mode_fb_cmd2) /** * Device specific ioctls should only be in their respective headers @@ -757,6 +773,8 @@ struct drm_event_vblank { #define DRM_CAP_DUMB_BUFFER 0x1 #define DRM_CAP_VBLANK_HIGH_CRTC 0x2 +#define DRM_CAP_DUMB_PREFERRED_DEPTH 0x3 +#define DRM_CAP_DUMB_PREFER_SHADOW 0x4 /* typedef area */ #ifndef __KERNEL__ diff --git a/include/drm/drmP.h b/include/drm/drmP.h index bf4b2dc..dd73104 100644 --- a/include/drm/drmP.h +++ b/include/drm/drmP.h @@ -91,6 +91,7 @@ struct drm_device; #define DRM_UT_CORE 0x01 #define DRM_UT_DRIVER 0x02 #define DRM_UT_KMS 0x04 +#define DRM_UT_PRIME 0x08 /* * Three debug levels are defined. * drm_core, drm_driver, drm_kms @@ -150,6 +151,7 @@ int drm_err(const char *func, const char *format, ...); #define DRIVER_IRQ_VBL2 0x800 #define DRIVER_GEM 0x1000 #define DRIVER_MODESET 0x2000 +#define DRIVER_PRIME 0x4000 #define DRIVER_BUS_PCI 0x1 #define DRIVER_BUS_PLATFORM 0x2 @@ -215,6 +217,11 @@ int drm_err(const char *func, const char *format, ...); drm_ut_debug_printk(DRM_UT_KMS, DRM_NAME, \ __func__, fmt, ##args); \ } while (0) +#define DRM_DEBUG_PRIME(fmt, args...) \ + do { \ + drm_ut_debug_printk(DRM_UT_PRIME, DRM_NAME, \ + __func__, fmt, ##args); \ + } while (0) #define DRM_LOG(fmt, args...) \ do { \ drm_ut_debug_printk(DRM_UT_CORE, NULL, \ @@ -238,6 +245,7 @@ int drm_err(const char *func, const char *format, ...); #else #define DRM_DEBUG_DRIVER(fmt, args...) do { } while (0) #define DRM_DEBUG_KMS(fmt, args...) do { } while (0) +#define DRM_DEBUG_PRIME(fmt, args...) do { } while (0) #define DRM_DEBUG(fmt, arg...) do { } while (0) #define DRM_LOG(fmt, arg...) do { } while (0) #define DRM_LOG_KMS(fmt, args...) do { } while (0) @@ -410,6 +418,12 @@ struct drm_pending_event { void (*destroy)(struct drm_pending_event *event); }; +/* initial implementaton using a linked list - todo hashtab */ +struct drm_prime_file_private { + struct list_head head; + struct mutex lock; +}; + /** File private data */ struct drm_file { int authenticated; @@ -437,6 +451,8 @@ struct drm_file { wait_queue_head_t event_wait; struct list_head event_list; int event_space; + + struct drm_prime_file_private prime; }; /** Wait queue */ @@ -652,6 +668,12 @@ struct drm_gem_object { uint32_t pending_write_domain; void *driver_private; + + /* dma buf exported from this GEM object */ + struct dma_buf *export_dma_buf; + + /* dma buf attachment backing this object */ + struct dma_buf_attachment *import_attach; }; #include "drm_crtc.h" @@ -820,7 +842,7 @@ struct drm_driver { * Specifically, the timestamp in @vblank_time should correspond as * closely as possible to the time when the first video scanline of * the video frame after the end of VBLANK will start scanning out, - * the time immmediately after end of the VBLANK interval. If the + * the time immediately after end of the VBLANK interval. If the * @crtc is currently inside VBLANK, this will be a time in the future. * If the @crtc is currently scanning out a frame, this will be the * past start time of the current scanout. This is meant to adhere @@ -890,6 +912,20 @@ struct drm_driver { int (*gem_open_object) (struct drm_gem_object *, struct drm_file *); void (*gem_close_object) (struct drm_gem_object *, struct drm_file *); + /* prime: */ + /* export handle -> fd (see drm_gem_prime_handle_to_fd() helper) */ + int (*prime_handle_to_fd)(struct drm_device *dev, struct drm_file *file_priv, + uint32_t handle, uint32_t flags, int *prime_fd); + /* import fd -> handle (see drm_gem_prime_fd_to_handle() helper) */ + int (*prime_fd_to_handle)(struct drm_device *dev, struct drm_file *file_priv, + int prime_fd, uint32_t *handle); + /* export GEM -> dmabuf */ + struct dma_buf * (*gem_prime_export)(struct drm_device *dev, + struct drm_gem_object *obj, int flags); + /* import dmabuf -> GEM */ + struct drm_gem_object * (*gem_prime_import)(struct drm_device *dev, + struct dma_buf *dma_buf); + /* vga arb irq handler */ void (*vgaarb_irq)(struct drm_device *dev, bool state); @@ -918,7 +954,7 @@ struct drm_driver { int dev_priv_size; struct drm_ioctl_desc *ioctls; int num_ioctls; - struct file_operations fops; + const struct file_operations *fops; union { struct pci_driver *pci; struct platform_device *platform_device; @@ -1170,6 +1206,8 @@ struct drm_device { struct idr object_name_idr; /*@} */ int switch_power_state; + + atomic_t unplugged; /* device has been unplugged or gone away */ }; #define DRM_SWITCH_POWER_ON 0 @@ -1235,6 +1273,19 @@ static inline int drm_mtrr_del(int handle, unsigned long offset, } #endif +static inline void drm_device_set_unplugged(struct drm_device *dev) +{ + smp_wmb(); + atomic_set(&dev->unplugged, 1); +} + +static inline int drm_device_is_unplugged(struct drm_device *dev) +{ + int ret = atomic_read(&dev->unplugged); + smp_rmb(); + return ret; +} + /******************************************************************/ /** \name Internal function definitions */ /*@{*/ @@ -1264,11 +1315,6 @@ extern unsigned int drm_poll(struct file *filp, struct poll_table_struct *wait); /* Memory management support (drm_memory.h) */ #include "drm_memory.h" -extern void drm_mem_init(void); -extern int drm_mem_info(char *buf, char **start, off_t offset, - int request, int *eof, void *data); -extern void *drm_realloc(void *oldpt, size_t oldsize, size_t size, int area); - extern void drm_free_agp(DRM_AGP_MEM * handle, int pages); extern int drm_bind_agp(DRM_AGP_MEM * handle, unsigned int start); extern DRM_AGP_MEM *drm_agp_bind_pages(struct drm_device *dev, @@ -1383,12 +1429,8 @@ extern void drm_core_reclaim_buffers(struct drm_device *dev, /* IRQ support (drm_irq.h) */ extern int drm_control(struct drm_device *dev, void *data, struct drm_file *file_priv); -extern irqreturn_t drm_irq_handler(DRM_IRQ_ARGS); extern int drm_irq_install(struct drm_device *dev); extern int drm_irq_uninstall(struct drm_device *dev); -extern void drm_driver_irq_preinstall(struct drm_device *dev); -extern void drm_driver_irq_postinstall(struct drm_device *dev); -extern void drm_driver_irq_uninstall(struct drm_device *dev); extern int drm_vblank_init(struct drm_device *dev, int num_crtcs); extern int drm_wait_vblank(struct drm_device *dev, void *data, @@ -1464,6 +1506,7 @@ extern void drm_master_put(struct drm_master **master); extern void drm_put_dev(struct drm_device *dev); extern int drm_put_minor(struct drm_minor **minor); +extern void drm_unplug_dev(struct drm_device *dev); extern unsigned int drm_debug; extern unsigned int drm_vblank_offdelay; @@ -1502,6 +1545,32 @@ extern int drm_vblank_info(struct seq_file *m, void *data); extern int drm_clients_info(struct seq_file *m, void* data); extern int drm_gem_name_info(struct seq_file *m, void *data); + +extern int drm_gem_prime_handle_to_fd(struct drm_device *dev, + struct drm_file *file_priv, uint32_t handle, uint32_t flags, + int *prime_fd); +extern int drm_gem_prime_fd_to_handle(struct drm_device *dev, + struct drm_file *file_priv, int prime_fd, uint32_t *handle); + +extern int drm_prime_handle_to_fd_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv); +extern int drm_prime_fd_to_handle_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv); + +extern struct sg_table *drm_prime_pages_to_sg(struct page **pages, int nr_pages); +extern void drm_prime_gem_destroy(struct drm_gem_object *obj, struct sg_table *sg); + + +void drm_prime_init_file_private(struct drm_prime_file_private *prime_fpriv); +void drm_prime_destroy_file_private(struct drm_prime_file_private *prime_fpriv); +int drm_prime_add_imported_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf, uint32_t handle); +int drm_prime_lookup_imported_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf, uint32_t *handle); +void drm_prime_remove_imported_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf); + +int drm_prime_add_dma_buf(struct drm_device *dev, struct drm_gem_object *obj); +int drm_prime_lookup_obj(struct drm_device *dev, struct dma_buf *buf, + struct drm_gem_object **obj); + #if DRM_DEBUG_CODE extern int drm_vma_info(struct seq_file *m, void *data); #endif @@ -1697,5 +1766,13 @@ extern void drm_platform_exit(struct drm_driver *driver, struct platform_device extern int drm_get_platform_dev(struct platform_device *pdev, struct drm_driver *driver); +/* returns true if currently okay to sleep */ +static __inline__ bool drm_can_sleep(void) +{ + if (in_atomic() || in_dbg_master() || irqs_disabled()) + return false; + return true; +} + #endif /* __KERNEL__ */ #endif diff --git a/include/drm/drm_crtc.h b/include/drm/drm_crtc.h index 8020798..e250eda 100644 --- a/include/drm/drm_crtc.h +++ b/include/drm/drm_crtc.h @@ -29,9 +29,10 @@ #include #include #include - #include +#include + struct drm_device; struct drm_mode_set; struct drm_framebuffer; @@ -44,6 +45,7 @@ struct drm_framebuffer; #define DRM_MODE_OBJECT_PROPERTY 0xb0b0b0b0 #define DRM_MODE_OBJECT_FB 0xfbfbfbfb #define DRM_MODE_OBJECT_BLOB 0xbbbbbbbb +#define DRM_MODE_OBJECT_PLANE 0xeeeeeeee struct drm_mode_object { uint32_t id; @@ -118,9 +120,8 @@ struct drm_display_mode { char name[DRM_DISPLAY_MODE_LEN]; - int connector_count; enum drm_mode_status status; - int type; + unsigned int type; /* Proposed mode values */ int clock; /* in kHz */ @@ -238,13 +239,15 @@ struct drm_framebuffer { struct list_head head; struct drm_mode_object base; const struct drm_framebuffer_funcs *funcs; - unsigned int pitch; + unsigned int pitches[4]; + unsigned int offsets[4]; unsigned int width; unsigned int height; /* depth can be 15 or 16 */ unsigned int depth; int bits_per_pixel; int flags; + uint32_t pixel_format; /* fourcc format */ struct list_head filp_head; /* if you are using the helper */ void *helper_private; @@ -254,7 +257,7 @@ struct drm_property_blob { struct drm_mode_object base; struct list_head head; unsigned int length; - void *data; + unsigned char data[]; }; struct drm_property_enum { @@ -278,6 +281,7 @@ struct drm_crtc; struct drm_connector; struct drm_encoder; struct drm_pending_vblank_event; +struct drm_plane; /** * drm_crtc_funcs - control CRTCs for a given device @@ -341,10 +345,21 @@ struct drm_crtc_funcs { /** * drm_crtc - central CRTC control structure + * @dev: parent DRM device + * @head: list management + * @base: base KMS object for ID tracking etc. * @enabled: is this CRTC enabled? + * @mode: current mode timings + * @hwmode: mode timings as programmed to hw regs * @x: x position on screen * @y: y position on screen * @funcs: CRTC control functions + * @gamma_size: size of gamma ramp + * @gamma_store: gamma ramp values + * @framedur_ns: precise frame timing + * @framedur_ns: precise line timing + * @pixeldur_ns: precise pixel timing + * @helper_private: mid-layer private data * * Each CRTC may have one or more connectors associated with it. This structure * allows the CRTC to be controlled. @@ -423,6 +438,13 @@ struct drm_connector_funcs { void (*force)(struct drm_connector *connector); }; +/** + * drm_encoder_funcs - encoder controls + * @reset: reset state (e.g. at init or resume time) + * @destroy: cleanup and free associated data + * + * Encoders sit between CRTCs and connectors. + */ struct drm_encoder_funcs { void (*reset)(struct drm_encoder *encoder); void (*destroy)(struct drm_encoder *encoder); @@ -431,10 +453,22 @@ struct drm_encoder_funcs { #define DRM_CONNECTOR_MAX_UMODES 16 #define DRM_CONNECTOR_MAX_PROPERTY 16 #define DRM_CONNECTOR_LEN 32 -#define DRM_CONNECTOR_MAX_ENCODER 2 +#define DRM_CONNECTOR_MAX_ENCODER 3 /** * drm_encoder - central DRM encoder structure + * @dev: parent DRM device + * @head: list management + * @base: base KMS object + * @encoder_type: one of the %DRM_MODE_ENCODER_ types in drm_mode.h + * @possible_crtcs: bitmask of potential CRTC bindings + * @possible_clones: bitmask of potential sibling encoders for cloning + * @crtc: currently bound CRTC + * @funcs: control functions + * @helper_private: mid-layer private data + * + * CRTCs drive pixels to encoders, which convert them into signals + * appropriate for a given connector or set of connectors. */ struct drm_encoder { struct drm_device *dev; @@ -470,14 +504,37 @@ enum drm_connector_force { /** * drm_connector - central DRM connector control structure - * @crtc: CRTC this connector is currently connected to, NULL if none + * @dev: parent DRM device + * @kdev: kernel device for sysfs attributes + * @attr: sysfs attributes + * @head: list management + * @base: base KMS object + * @connector_type: one of the %DRM_MODE_CONNECTOR_ types from drm_mode.h + * @connector_type_id: index into connector type enum * @interlace_allowed: can this connector handle interlaced modes? * @doublescan_allowed: can this connector handle doublescan? - * @available_modes: modes available on this connector (from get_modes() + user) - * @initial_x: initial x position for this connector - * @initial_y: initial y position for this connector - * @status: connector connected? + * @modes: modes available on this connector (from fill_modes() + user) + * @status: one of the drm_connector_status enums (connected, not, or unknown) + * @probed_modes: list of modes derived directly from the display + * @display_info: information about attached display (e.g. from EDID) * @funcs: connector control functions + * @user_modes: user added mode list + * @edid_blob_ptr: DRM property containing EDID if present + * @property_ids: property tracking for this connector + * @property_values: value pointers or data for properties + * @polled: a %DRM_CONNECTOR_POLL_ value for core driven polling + * @dpms: current dpms state + * @helper_private: mid-layer private data + * @force: a %DRM_FORCE_ state for forced mode sets + * @encoder_ids: valid encoders for this connector + * @encoder: encoder driving this connector, if any + * @eld: EDID-like data, if present + * @dvi_dual: dual link DVI, if found + * @max_tmds_clock: max clock rate, if found + * @latency_present: AV delay info from ELD, if found + * @video_latency: video latency info from ELD, if found + * @audio_latency: audio latency info from ELD, if found + * @null_edid_counter: track sinks that give us all zeros for the EDID * * Each connector may be connected to one or more CRTCs, or may be clonable by * another connector if they can share a CRTC. Each connector also has a specific @@ -498,7 +555,6 @@ struct drm_connector { bool doublescan_allowed; struct list_head modes; /* list of modes on this connector */ - int initial_x, initial_y; enum drm_connector_status status; /* these are modes added by probing with DDC or the BIOS */ @@ -522,7 +578,6 @@ struct drm_connector { /* forced on connector */ enum drm_connector_force force; uint32_t encoder_ids[DRM_CONNECTOR_MAX_ENCODER]; - uint32_t force_encoder_id; struct drm_encoder *encoder; /* currently active encoder */ /* EDID bits */ @@ -536,7 +591,71 @@ struct drm_connector { }; /** - * struct drm_mode_set + * drm_plane_funcs - driver plane control functions + * @update_plane: update the plane configuration + * @disable_plane: shut down the plane + * @destroy: clean up plane resources + */ +struct drm_plane_funcs { + int (*update_plane)(struct drm_plane *plane, + struct drm_crtc *crtc, struct drm_framebuffer *fb, + int crtc_x, int crtc_y, + unsigned int crtc_w, unsigned int crtc_h, + uint32_t src_x, uint32_t src_y, + uint32_t src_w, uint32_t src_h); + int (*disable_plane)(struct drm_plane *plane); + void (*destroy)(struct drm_plane *plane); +}; + +/** + * drm_plane - central DRM plane control structure + * @dev: DRM device this plane belongs to + * @head: for list management + * @base: base mode object + * @possible_crtcs: pipes this plane can be bound to + * @format_types: array of formats supported by this plane + * @format_count: number of formats supported + * @crtc: currently bound CRTC + * @fb: currently bound fb + * @gamma_size: size of gamma table + * @gamma_store: gamma correction table + * @enabled: enabled flag + * @funcs: helper functions + * @helper_private: storage for drver layer + */ +struct drm_plane { + struct drm_device *dev; + struct list_head head; + + struct drm_mode_object base; + + uint32_t possible_crtcs; + uint32_t *format_types; + uint32_t format_count; + + struct drm_crtc *crtc; + struct drm_framebuffer *fb; + + /* CRTC gamma size for reporting to userspace */ + uint32_t gamma_size; + uint16_t *gamma_store; + + bool enabled; + + const struct drm_plane_funcs *funcs; + void *helper_private; +}; + +/** + * drm_mode_set - new values for a CRTC config change + * @head: list management + * @fb: framebuffer to use for new config + * @crtc: CRTC whose configuration we're about to change + * @mode: mode timings to use + * @x: position of this CRTC relative to @fb + * @y: position of this CRTC relative to @fb + * @connectors: array of connectors to drive with this CRTC if possible + * @num_connectors: size of @connectors array * * Represents a single crtc the connectors that it drives with what mode * and from which framebuffer it scans out from. @@ -558,13 +677,33 @@ struct drm_mode_set { }; /** - * struct drm_mode_config_funcs - configure CRTCs for a given screen layout + * struct drm_mode_config_funcs - basic driver provided mode setting functions + * @fb_create: create a new framebuffer object + * @output_poll_changed: function to handle output configuration changes + * + * Some global (i.e. not per-CRTC, connector, etc) mode setting functions that + * involve drivers. */ struct drm_mode_config_funcs { - struct drm_framebuffer *(*fb_create)(struct drm_device *dev, struct drm_file *file_priv, struct drm_mode_fb_cmd *mode_cmd); + struct drm_framebuffer *(*fb_create)(struct drm_device *dev, + struct drm_file *file_priv, + struct drm_mode_fb_cmd2 *mode_cmd); void (*output_poll_changed)(struct drm_device *dev); }; +/** + * drm_mode_group - group of mode setting resources for potential sub-grouping + * @num_crtcs: CRTC count + * @num_encoders: encoder count + * @num_connectors: connector count + * @id_list: list of KMS object IDs in this group + * + * Currently this simply tracks the global mode setting state. But in the + * future it could allow groups of objects to be set aside into independent + * control groups for use by different user level processes (e.g. two X servers + * running simultaneously on different heads, each with their own mode + * configuration and freedom of mode setting). + */ struct drm_mode_group { uint32_t num_crtcs; uint32_t num_encoders; @@ -576,7 +715,30 @@ struct drm_mode_group { /** * drm_mode_config - Mode configuration control structure + * @mutex: mutex protecting KMS related lists and structures + * @idr_mutex: mutex for KMS ID allocation and management + * @crtc_idr: main KMS ID tracking object + * @num_fb: number of fbs available + * @fb_list: list of framebuffers available + * @num_connector: number of connectors on this device + * @connector_list: list of connector objects + * @num_encoder: number of encoders on this device + * @encoder_list: list of encoder objects + * @num_crtc: number of CRTCs on this device + * @crtc_list: list of CRTC objects + * @min_width: minimum pixel width on this device + * @min_height: minimum pixel height on this device + * @max_width: maximum pixel width on this device + * @max_height: maximum pixel height on this device + * @funcs: core driver provided mode setting functions + * @fb_base: base address of the framebuffer + * @poll_enabled: track polling status for this device + * @output_poll_work: delayed work for polling in process context + * @*_property: core property tracking * + * Core mode resource tracking structure. All CRTC, encoders, and connectors + * enumerated by the driver are added here, as are global properties. Some + * global restrictions are also here, e.g. dimension restrictions. */ struct drm_mode_config { struct mutex mutex; /* protects configuration (mode lists etc.) */ @@ -589,6 +751,8 @@ struct drm_mode_config { struct list_head connector_list; int num_encoder; struct list_head encoder_list; + int num_plane; + struct list_head plane_list; int num_crtc; struct list_head crtc_list; @@ -632,6 +796,9 @@ struct drm_mode_config { struct drm_property *scaling_mode_property; struct drm_property *dithering_mode_property; struct drm_property *dirty_info_property; + + /* dumb ioctl parameters */ + uint32_t preferred_depth, prefer_shadow; }; #define obj_to_crtc(x) container_of(x, struct drm_crtc, base) @@ -641,24 +808,39 @@ struct drm_mode_config { #define obj_to_fb(x) container_of(x, struct drm_framebuffer, base) #define obj_to_property(x) container_of(x, struct drm_property, base) #define obj_to_blob(x) container_of(x, struct drm_property_blob, base) +#define obj_to_plane(x) container_of(x, struct drm_plane, base) +struct drm_prop_enum_list { + int type; + char *name; +}; -extern void drm_crtc_init(struct drm_device *dev, - struct drm_crtc *crtc, - const struct drm_crtc_funcs *funcs); +extern int drm_crtc_init(struct drm_device *dev, + struct drm_crtc *crtc, + const struct drm_crtc_funcs *funcs); extern void drm_crtc_cleanup(struct drm_crtc *crtc); -extern void drm_connector_init(struct drm_device *dev, - struct drm_connector *connector, - const struct drm_connector_funcs *funcs, - int connector_type); +extern int drm_connector_init(struct drm_device *dev, + struct drm_connector *connector, + const struct drm_connector_funcs *funcs, + int connector_type); extern void drm_connector_cleanup(struct drm_connector *connector); - -extern void drm_encoder_init(struct drm_device *dev, - struct drm_encoder *encoder, - const struct drm_encoder_funcs *funcs, - int encoder_type); +/* helper to unplug all connectors from sysfs for device */ +extern void drm_connector_unplug_all(struct drm_device *dev); + +extern int drm_encoder_init(struct drm_device *dev, + struct drm_encoder *encoder, + const struct drm_encoder_funcs *funcs, + int encoder_type); + +extern int drm_plane_init(struct drm_device *dev, + struct drm_plane *plane, + unsigned long possible_crtcs, + const struct drm_plane_funcs *funcs, + const uint32_t *formats, uint32_t format_count, + bool priv); +extern void drm_plane_cleanup(struct drm_plane *plane); extern void drm_encoder_cleanup(struct drm_encoder *encoder); @@ -675,6 +857,7 @@ extern struct edid *drm_get_edid(struct drm_connector *connector, extern int drm_add_edid_modes(struct drm_connector *connector, struct edid *edid); extern void drm_mode_probed_add(struct drm_connector *connector, struct drm_display_mode *mode); extern void drm_mode_remove(struct drm_connector *connector, struct drm_display_mode *mode); +extern void drm_mode_copy(struct drm_display_mode *dst, const struct drm_display_mode *src); extern struct drm_display_mode *drm_mode_duplicate(struct drm_device *dev, const struct drm_display_mode *mode); extern void drm_mode_debug_printmodeline(struct drm_display_mode *mode); @@ -689,7 +872,7 @@ extern int drm_mode_height(struct drm_display_mode *mode); /* for us by fb module */ extern int drm_mode_attachmode_crtc(struct drm_device *dev, struct drm_crtc *crtc, - struct drm_display_mode *mode); + const struct drm_display_mode *mode); extern int drm_mode_detachmode_crtc(struct drm_device *dev, struct drm_display_mode *mode); extern struct drm_display_mode *drm_mode_create(struct drm_device *dev); @@ -731,6 +914,13 @@ extern int drm_connector_attach_property(struct drm_connector *connector, struct drm_property *property, uint64_t init_val); extern struct drm_property *drm_property_create(struct drm_device *dev, int flags, const char *name, int num_values); +extern struct drm_property *drm_property_create_enum(struct drm_device *dev, int flags, + const char *name, + const struct drm_prop_enum_list *props, + int num_values); +struct drm_property *drm_property_create_range(struct drm_device *dev, int flags, + const char *name, + uint64_t min, uint64_t max); extern void drm_property_destroy(struct drm_device *dev, struct drm_property *property); extern int drm_property_add_enum(struct drm_property *property, int index, uint64_t value, const char *name); @@ -746,24 +936,32 @@ extern int drm_mode_connector_attach_encoder(struct drm_connector *connector, struct drm_encoder *encoder); extern void drm_mode_connector_detach_encoder(struct drm_connector *connector, struct drm_encoder *encoder); -extern bool drm_mode_crtc_set_gamma_size(struct drm_crtc *crtc, +extern int drm_mode_crtc_set_gamma_size(struct drm_crtc *crtc, int gamma_size); extern struct drm_mode_object *drm_mode_object_find(struct drm_device *dev, uint32_t id, uint32_t type); /* IOCTLs */ extern int drm_mode_getresources(struct drm_device *dev, void *data, struct drm_file *file_priv); - +extern int drm_mode_getplane_res(struct drm_device *dev, void *data, + struct drm_file *file_priv); extern int drm_mode_getcrtc(struct drm_device *dev, void *data, struct drm_file *file_priv); extern int drm_mode_getconnector(struct drm_device *dev, void *data, struct drm_file *file_priv); extern int drm_mode_setcrtc(struct drm_device *dev, void *data, struct drm_file *file_priv); +extern int drm_mode_getplane(struct drm_device *dev, + void *data, struct drm_file *file_priv); +extern int drm_mode_setplane(struct drm_device *dev, + void *data, struct drm_file *file_priv); extern int drm_mode_cursor_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); extern int drm_mode_addfb(struct drm_device *dev, void *data, struct drm_file *file_priv); +extern int drm_mode_addfb2(struct drm_device *dev, + void *data, struct drm_file *file_priv); +extern uint32_t drm_mode_legacy_fb_format(uint32_t bpp, uint32_t depth); extern int drm_mode_rmfb(struct drm_device *dev, void *data, struct drm_file *file_priv); extern int drm_mode_getfb(struct drm_device *dev, @@ -814,6 +1012,7 @@ extern int drm_add_modes_noedid(struct drm_connector *connector, int hdisplay, int vdisplay); extern int drm_edid_header_is_valid(const u8 *raw_edid); +extern bool drm_edid_block_valid(u8 *raw_edid); extern bool drm_edid_is_valid(struct edid *edid); struct drm_display_mode *drm_mode_find_dmt(struct drm_device *dev, int hsize, int vsize, int fresh); @@ -824,4 +1023,7 @@ extern int drm_mode_mmap_dumb_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); extern int drm_mode_destroy_dumb_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); + +extern void drm_fb_get_bpp_depth(uint32_t format, unsigned int *depth, + int *bpp); #endif /* __DRM_CRTC_H__ */ diff --git a/include/drm/drm_crtc_helper.h b/include/drm/drm_crtc_helper.h index 73b0712..37515d1 100644 --- a/include/drm/drm_crtc_helper.h +++ b/include/drm/drm_crtc_helper.h @@ -117,7 +117,7 @@ extern bool drm_helper_encoder_in_use(struct drm_encoder *encoder); extern void drm_helper_connector_dpms(struct drm_connector *connector, int mode); extern int drm_helper_mode_fill_fb_struct(struct drm_framebuffer *fb, - struct drm_mode_fb_cmd *mode_cmd); + struct drm_mode_fb_cmd2 *mode_cmd); static inline void drm_crtc_helper_add(struct drm_crtc *crtc, const struct drm_crtc_helper_funcs *funcs) @@ -144,4 +144,7 @@ extern void drm_helper_hpd_irq_event(struct drm_device *dev); extern void drm_kms_helper_poll_disable(struct drm_device *dev); extern void drm_kms_helper_poll_enable(struct drm_device *dev); + +extern int drm_format_num_planes(uint32_t format); + #endif diff --git a/include/drm/drm_edid.h b/include/drm/drm_edid.h index 74ce916..bcb9a66 100644 --- a/include/drm/drm_edid.h +++ b/include/drm/drm_edid.h @@ -238,5 +238,6 @@ int drm_av_sync_delay(struct drm_connector *connector, struct drm_display_mode *mode); struct drm_connector *drm_select_eld(struct drm_encoder *encoder, struct drm_display_mode *mode); +int drm_load_edid_firmware(struct drm_connector *connector); #endif /* __DRM_EDID_H__ */ diff --git a/include/drm/drm_fb_helper.h b/include/drm/drm_fb_helper.h index 6e3076a..5120b01 100644 --- a/include/drm/drm_fb_helper.h +++ b/include/drm/drm_fb_helper.h @@ -35,7 +35,6 @@ struct drm_fb_helper; #include struct drm_fb_helper_crtc { - uint32_t crtc_id; struct drm_mode_set mode_set; struct drm_display_mode *desired_mode; }; @@ -74,7 +73,6 @@ struct drm_fb_helper { int connector_count; struct drm_fb_helper_connector **connector_info; struct drm_fb_helper_funcs *funcs; - int conn_limit; struct fb_info *fbdev; u32 pseudo_palette[17]; struct list_head kernel_fb_list; diff --git a/include/drm/drm_fourcc.h b/include/drm/drm_fourcc.h new file mode 100644 index 0000000..bdf0152 --- /dev/null +++ b/include/drm/drm_fourcc.h @@ -0,0 +1,137 @@ +/* + * Copyright 2011 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * VA LINUX SYSTEMS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ + +#ifndef DRM_FOURCC_H +#define DRM_FOURCC_H + +#include + +#define fourcc_code(a, b, c, d) ((__u32)(a) | ((__u32)(b) << 8) | \ + ((__u32)(c) << 16) | ((__u32)(d) << 24)) + +#define DRM_FORMAT_BIG_ENDIAN (1<<31) /* format is big endian instead of little endian */ + +/* color index */ +#define DRM_FORMAT_C8 fourcc_code('C', '8', ' ', ' ') /* [7:0] C */ + +/* 8 bpp RGB */ +#define DRM_FORMAT_RGB332 fourcc_code('R', 'G', 'B', '8') /* [7:0] R:G:B 3:3:2 */ +#define DRM_FORMAT_BGR233 fourcc_code('B', 'G', 'R', '8') /* [7:0] B:G:R 2:3:3 */ + +/* 16 bpp RGB */ +#define DRM_FORMAT_XRGB4444 fourcc_code('X', 'R', '1', '2') /* [15:0] x:R:G:B 4:4:4:4 little endian */ +#define DRM_FORMAT_XBGR4444 fourcc_code('X', 'B', '1', '2') /* [15:0] x:B:G:R 4:4:4:4 little endian */ +#define DRM_FORMAT_RGBX4444 fourcc_code('R', 'X', '1', '2') /* [15:0] R:G:B:x 4:4:4:4 little endian */ +#define DRM_FORMAT_BGRX4444 fourcc_code('B', 'X', '1', '2') /* [15:0] B:G:R:x 4:4:4:4 little endian */ + +#define DRM_FORMAT_ARGB4444 fourcc_code('A', 'R', '1', '2') /* [15:0] A:R:G:B 4:4:4:4 little endian */ +#define DRM_FORMAT_ABGR4444 fourcc_code('A', 'B', '1', '2') /* [15:0] A:B:G:R 4:4:4:4 little endian */ +#define DRM_FORMAT_RGBA4444 fourcc_code('R', 'A', '1', '2') /* [15:0] R:G:B:A 4:4:4:4 little endian */ +#define DRM_FORMAT_BGRA4444 fourcc_code('B', 'A', '1', '2') /* [15:0] B:G:R:A 4:4:4:4 little endian */ + +#define DRM_FORMAT_XRGB1555 fourcc_code('X', 'R', '1', '5') /* [15:0] x:R:G:B 1:5:5:5 little endian */ +#define DRM_FORMAT_XBGR1555 fourcc_code('X', 'B', '1', '5') /* [15:0] x:B:G:R 1:5:5:5 little endian */ +#define DRM_FORMAT_RGBX5551 fourcc_code('R', 'X', '1', '5') /* [15:0] R:G:B:x 5:5:5:1 little endian */ +#define DRM_FORMAT_BGRX5551 fourcc_code('B', 'X', '1', '5') /* [15:0] B:G:R:x 5:5:5:1 little endian */ + +#define DRM_FORMAT_ARGB1555 fourcc_code('A', 'R', '1', '5') /* [15:0] A:R:G:B 1:5:5:5 little endian */ +#define DRM_FORMAT_ABGR1555 fourcc_code('A', 'B', '1', '5') /* [15:0] A:B:G:R 1:5:5:5 little endian */ +#define DRM_FORMAT_RGBA5551 fourcc_code('R', 'A', '1', '5') /* [15:0] R:G:B:A 5:5:5:1 little endian */ +#define DRM_FORMAT_BGRA5551 fourcc_code('B', 'A', '1', '5') /* [15:0] B:G:R:A 5:5:5:1 little endian */ + +#define DRM_FORMAT_RGB565 fourcc_code('R', 'G', '1', '6') /* [15:0] R:G:B 5:6:5 little endian */ +#define DRM_FORMAT_BGR565 fourcc_code('B', 'G', '1', '6') /* [15:0] B:G:R 5:6:5 little endian */ + +/* 24 bpp RGB */ +#define DRM_FORMAT_RGB888 fourcc_code('R', 'G', '2', '4') /* [23:0] R:G:B little endian */ +#define DRM_FORMAT_BGR888 fourcc_code('B', 'G', '2', '4') /* [23:0] B:G:R little endian */ + +/* 32 bpp RGB */ +#define DRM_FORMAT_XRGB8888 fourcc_code('X', 'R', '2', '4') /* [31:0] x:R:G:B 8:8:8:8 little endian */ +#define DRM_FORMAT_XBGR8888 fourcc_code('X', 'B', '2', '4') /* [31:0] x:B:G:R 8:8:8:8 little endian */ +#define DRM_FORMAT_RGBX8888 fourcc_code('R', 'X', '2', '4') /* [31:0] R:G:B:x 8:8:8:8 little endian */ +#define DRM_FORMAT_BGRX8888 fourcc_code('B', 'X', '2', '4') /* [31:0] B:G:R:x 8:8:8:8 little endian */ + +#define DRM_FORMAT_ARGB8888 fourcc_code('A', 'R', '2', '4') /* [31:0] A:R:G:B 8:8:8:8 little endian */ +#define DRM_FORMAT_ABGR8888 fourcc_code('A', 'B', '2', '4') /* [31:0] A:B:G:R 8:8:8:8 little endian */ +#define DRM_FORMAT_RGBA8888 fourcc_code('R', 'A', '2', '4') /* [31:0] R:G:B:A 8:8:8:8 little endian */ +#define DRM_FORMAT_BGRA8888 fourcc_code('B', 'A', '2', '4') /* [31:0] B:G:R:A 8:8:8:8 little endian */ + +#define DRM_FORMAT_XRGB2101010 fourcc_code('X', 'R', '3', '0') /* [31:0] x:R:G:B 2:10:10:10 little endian */ +#define DRM_FORMAT_XBGR2101010 fourcc_code('X', 'B', '3', '0') /* [31:0] x:B:G:R 2:10:10:10 little endian */ +#define DRM_FORMAT_RGBX1010102 fourcc_code('R', 'X', '3', '0') /* [31:0] R:G:B:x 10:10:10:2 little endian */ +#define DRM_FORMAT_BGRX1010102 fourcc_code('B', 'X', '3', '0') /* [31:0] B:G:R:x 10:10:10:2 little endian */ + +#define DRM_FORMAT_ARGB2101010 fourcc_code('A', 'R', '3', '0') /* [31:0] A:R:G:B 2:10:10:10 little endian */ +#define DRM_FORMAT_ABGR2101010 fourcc_code('A', 'B', '3', '0') /* [31:0] A:B:G:R 2:10:10:10 little endian */ +#define DRM_FORMAT_RGBA1010102 fourcc_code('R', 'A', '3', '0') /* [31:0] R:G:B:A 10:10:10:2 little endian */ +#define DRM_FORMAT_BGRA1010102 fourcc_code('B', 'A', '3', '0') /* [31:0] B:G:R:A 10:10:10:2 little endian */ + +/* packed YCbCr */ +#define DRM_FORMAT_YUYV fourcc_code('Y', 'U', 'Y', 'V') /* [31:0] Cr0:Y1:Cb0:Y0 8:8:8:8 little endian */ +#define DRM_FORMAT_YVYU fourcc_code('Y', 'V', 'Y', 'U') /* [31:0] Cb0:Y1:Cr0:Y0 8:8:8:8 little endian */ +#define DRM_FORMAT_UYVY fourcc_code('U', 'Y', 'V', 'Y') /* [31:0] Y1:Cr0:Y0:Cb0 8:8:8:8 little endian */ +#define DRM_FORMAT_VYUY fourcc_code('V', 'Y', 'U', 'Y') /* [31:0] Y1:Cb0:Y0:Cr0 8:8:8:8 little endian */ + +#define DRM_FORMAT_AYUV fourcc_code('A', 'Y', 'U', 'V') /* [31:0] A:Y:Cb:Cr 8:8:8:8 little endian */ + +/* + * 2 plane YCbCr + * index 0 = Y plane, [7:0] Y + * index 1 = Cr:Cb plane, [15:0] Cr:Cb little endian + * or + * index 1 = Cb:Cr plane, [15:0] Cb:Cr little endian + */ +#define DRM_FORMAT_NV12 fourcc_code('N', 'V', '1', '2') /* 2x2 subsampled Cr:Cb plane */ +#define DRM_FORMAT_NV21 fourcc_code('N', 'V', '2', '1') /* 2x2 subsampled Cb:Cr plane */ +#define DRM_FORMAT_NV16 fourcc_code('N', 'V', '1', '6') /* 2x1 subsampled Cr:Cb plane */ +#define DRM_FORMAT_NV61 fourcc_code('N', 'V', '6', '1') /* 2x1 subsampled Cb:Cr plane */ + +/* 2 non contiguous plane YCbCr */ +#define DRM_FORMAT_NV12M fourcc_code('N', 'M', '1', '2') /* 2x2 subsampled Cr:Cb plane */ +#define DRM_FORMAT_NV12MT fourcc_code('T', 'M', '1', '2') /* 2x2 subsampled Cr:Cb plane 64x32 macroblocks */ + +/* + * 3 plane YCbCr + * index 0: Y plane, [7:0] Y + * index 1: Cb plane, [7:0] Cb + * index 2: Cr plane, [7:0] Cr + * or + * index 1: Cr plane, [7:0] Cr + * index 2: Cb plane, [7:0] Cb + */ +#define DRM_FORMAT_YUV410 fourcc_code('Y', 'U', 'V', '9') /* 4x4 subsampled Cb (1) and Cr (2) planes */ +#define DRM_FORMAT_YVU410 fourcc_code('Y', 'V', 'U', '9') /* 4x4 subsampled Cr (1) and Cb (2) planes */ +#define DRM_FORMAT_YUV411 fourcc_code('Y', 'U', '1', '1') /* 4x1 subsampled Cb (1) and Cr (2) planes */ +#define DRM_FORMAT_YVU411 fourcc_code('Y', 'V', '1', '1') /* 4x1 subsampled Cr (1) and Cb (2) planes */ +#define DRM_FORMAT_YUV420 fourcc_code('Y', 'U', '1', '2') /* 2x2 subsampled Cb (1) and Cr (2) planes */ +#define DRM_FORMAT_YVU420 fourcc_code('Y', 'V', '1', '2') /* 2x2 subsampled Cr (1) and Cb (2) planes */ +#define DRM_FORMAT_YUV422 fourcc_code('Y', 'U', '1', '6') /* 2x1 subsampled Cb (1) and Cr (2) planes */ +#define DRM_FORMAT_YVU422 fourcc_code('Y', 'V', '1', '6') /* 2x1 subsampled Cr (1) and Cb (2) planes */ +#define DRM_FORMAT_YUV444 fourcc_code('Y', 'U', '2', '4') /* non-subsampled Cb (1) and Cr (2) planes */ +#define DRM_FORMAT_YVU444 fourcc_code('Y', 'V', '2', '4') /* non-subsampled Cr (1) and Cb (2) planes */ + +/* 3 non contiguous plane YCbCr */ +#define DRM_FORMAT_YUV420M fourcc_code('Y', 'M', '1', '2') /* 2x2 subsampled Cb (1) and Cr (2) planes */ + +#endif /* DRM_FOURCC_H */ diff --git a/include/drm/drm_mode.h b/include/drm/drm_mode.h index 7639f18..9242310 100644 --- a/include/drm/drm_mode.h +++ b/include/drm/drm_mode.h @@ -27,6 +27,8 @@ #ifndef _DRM_MODE_H #define _DRM_MODE_H +#include + #define DRM_DISPLAY_INFO_LEN 32 #define DRM_CONNECTOR_NAME_LEN 32 #define DRM_DISPLAY_MODE_LEN 32 @@ -120,11 +122,48 @@ struct drm_mode_crtc { struct drm_mode_modeinfo mode; }; -#define DRM_MODE_ENCODER_NONE 0 -#define DRM_MODE_ENCODER_DAC 1 -#define DRM_MODE_ENCODER_TMDS 2 -#define DRM_MODE_ENCODER_LVDS 3 -#define DRM_MODE_ENCODER_TVDAC 4 +#define DRM_MODE_PRESENT_TOP_FIELD (1<<0) +#define DRM_MODE_PRESENT_BOTTOM_FIELD (1<<1) + +/* Planes blend with or override other bits on the CRTC */ +struct drm_mode_set_plane { + __u32 plane_id; + __u32 crtc_id; + __u32 fb_id; /* fb object contains surface format type */ + __u32 flags; /* see above flags */ + + /* Signed dest location allows it to be partially off screen */ + __s32 crtc_x, crtc_y; + __u32 crtc_w, crtc_h; + + /* Source values are 16.16 fixed point */ + __u32 src_x, src_y; + __u32 src_h, src_w; +}; + +struct drm_mode_get_plane { + __u32 plane_id; + + __u32 crtc_id; + __u32 fb_id; + + __u32 possible_crtcs; + __u32 gamma_size; + + __u32 count_format_types; + __u64 format_type_ptr; +}; + +struct drm_mode_get_plane_res { + __u64 plane_id_ptr; + __u32 count_planes; +}; + +#define DRM_MODE_ENCODER_NONE 0 +#define DRM_MODE_ENCODER_DAC 1 +#define DRM_MODE_ENCODER_TMDS 2 +#define DRM_MODE_ENCODER_LVDS 3 +#define DRM_MODE_ENCODER_TVDAC 4 #define DRM_MODE_ENCODER_VIRTUAL 5 struct drm_mode_get_encoder { @@ -231,6 +270,33 @@ struct drm_mode_fb_cmd { __u32 handle; }; +#define DRM_MODE_FB_INTERLACED (1<<0) /* for interlaced framebuffers */ + +struct drm_mode_fb_cmd2 { + __u32 fb_id; + __u32 width, height; + __u32 pixel_format; /* fourcc code from drm_fourcc.h */ + __u32 flags; /* see above flags */ + + /* + * In case of planar formats, this ioctl allows up to 4 + * buffer objects with offets and pitches per plane. + * The pitch and offset order is dictated by the fourcc, + * e.g. NV12 (http://fourcc.org/yuv.php#NV12) is described as: + * + * YUV 4:2:0 image with a plane of 8 bit Y samples + * followed by an interleaved U/V plane containing + * 8 bit 2x2 subsampled colour difference samples. + * + * So it would consist of Y as offset[0] and UV as + * offeset[1]. Note that offset[0] will generally + * be 0. + */ + __u32 handles[4]; + __u32 pitches[4]; /* pitch for each plane */ + __u32 offsets[4]; /* offset of each plane */ +}; + #define DRM_MODE_FB_DIRTY_ANNOTATE_COPY 0x01 #define DRM_MODE_FB_DIRTY_ANNOTATE_FILL 0x02 #define DRM_MODE_FB_DIRTY_FLAGS 0x03 diff --git a/include/drm/drm_pciids.h b/include/drm/drm_pciids.h index def807c..7207a99 100644 --- a/include/drm/drm_pciids.h +++ b/include/drm/drm_pciids.h @@ -204,11 +204,57 @@ {0x1002, 0x6778, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CAICOS|RADEON_NEW_MEMMAP}, \ {0x1002, 0x6779, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CAICOS|RADEON_NEW_MEMMAP}, \ {0x1002, 0x677B, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_CAICOS|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6780, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6784, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6788, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x678A, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6790, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6791, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6792, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6798, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6799, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x679A, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x679B, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x679E, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x679F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6800, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6801, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6802, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6806, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6808, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6809, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6810, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6811, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6816, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6817, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6818, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6819, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6820, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6821, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6823, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6824, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6825, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6826, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6827, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6828, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6829, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x682B, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x682D, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x682F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6830, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6831, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6837, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6838, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x6839, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x683B, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x683D, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x683F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|RADEON_NEW_MEMMAP}, \ {0x1002, 0x6840, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TURKS|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ {0x1002, 0x6841, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TURKS|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ {0x1002, 0x6842, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TURKS|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ {0x1002, 0x6843, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TURKS|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP}, \ {0x1002, 0x6849, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TURKS|RADEON_NEW_MEMMAP}, \ + {0x1002, 0x684C, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN|RADEON_NEW_MEMMAP}, \ {0x1002, 0x6850, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TURKS|RADEON_NEW_MEMMAP}, \ {0x1002, 0x6858, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TURKS|RADEON_NEW_MEMMAP}, \ {0x1002, 0x6859, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TURKS|RADEON_NEW_MEMMAP}, \ @@ -516,6 +562,30 @@ {0x1002, 0x9808, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PALM|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ {0x1002, 0x9809, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PALM|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ {0x1002, 0x980A, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PALM|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x9900, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x9901, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x9903, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x9904, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x9905, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x9906, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x9907, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x9908, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x9909, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x990A, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x990F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x9910, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x9913, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x9917, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x9918, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x9919, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x9990, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x9991, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x9992, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x9993, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x9994, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x99A0, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x99A2, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_IS_MOBILITY|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ + {0x1002, 0x99A4, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ARUBA|RADEON_NEW_MEMMAP|RADEON_IS_IGP}, \ {0, 0, 0} #define r128_PCI_IDS \ diff --git a/include/drm/drm_sman.h b/include/drm/drm_sman.h deleted file mode 100644 index 08ecf83..0000000 --- a/include/drm/drm_sman.h +++ /dev/null @@ -1,176 +0,0 @@ -/************************************************************************** - * - * Copyright 2006 Tungsten Graphics, Inc., Bismarck, ND., USA. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, - * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR - * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE - * USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * - **************************************************************************/ -/* - * Simple memory MANager interface that keeps track on allocate regions on a - * per "owner" basis. All regions associated with an "owner" can be released - * with a simple call. Typically if the "owner" exists. The owner is any - * "unsigned long" identifier. Can typically be a pointer to a file private - * struct or a context identifier. - * - * Authors: - * Thomas Hellström - */ - -#ifndef DRM_SMAN_H -#define DRM_SMAN_H - -#include "drmP.h" -#include "drm_hashtab.h" - -/* - * A class that is an abstration of a simple memory allocator. - * The sman implementation provides a default such allocator - * using the drm_mm.c implementation. But the user can replace it. - * See the SiS implementation, which may use the SiS FB kernel module - * for memory management. - */ - -struct drm_sman_mm { - /* private info. If allocated, needs to be destroyed by the destroy - function */ - void *private; - - /* Allocate a memory block with given size and alignment. - Return an opaque reference to the memory block */ - - void *(*allocate) (void *private, unsigned long size, - unsigned alignment); - - /* Free a memory block. "ref" is the opaque reference that we got from - the "alloc" function */ - - void (*free) (void *private, void *ref); - - /* Free all resources associated with this allocator */ - - void (*destroy) (void *private); - - /* Return a memory offset from the opaque reference returned from the - "alloc" function */ - - unsigned long (*offset) (void *private, void *ref); -}; - -struct drm_memblock_item { - struct list_head owner_list; - struct drm_hash_item user_hash; - void *mm_info; - struct drm_sman_mm *mm; - struct drm_sman *sman; -}; - -struct drm_sman { - struct drm_sman_mm *mm; - int num_managers; - struct drm_open_hash owner_hash_tab; - struct drm_open_hash user_hash_tab; - struct list_head owner_items; -}; - -/* - * Take down a memory manager. This function should only be called after a - * successful init and after a call to drm_sman_cleanup. - */ - -extern void drm_sman_takedown(struct drm_sman * sman); - -/* - * Allocate structures for a manager. - * num_managers are the number of memory pools to manage. (VRAM, AGP, ....) - * user_order is the log2 of the number of buckets in the user hash table. - * set this to approximately log2 of the max number of memory regions - * that will be allocated for _all_ pools together. - * owner_order is the log2 of the number of buckets in the owner hash table. - * set this to approximately log2 of - * the number of client file connections that will - * be using the manager. - * - */ - -extern int drm_sman_init(struct drm_sman * sman, unsigned int num_managers, - unsigned int user_order, unsigned int owner_order); - -/* - * Initialize a drm_mm.c allocator. Should be called only once for each - * manager unless a customized allogator is used. - */ - -extern int drm_sman_set_range(struct drm_sman * sman, unsigned int manager, - unsigned long start, unsigned long size); - -/* - * Initialize a customized allocator for one of the managers. - * (See the SiS module). The object pointed to by "allocator" is copied, - * so it can be destroyed after this call. - */ - -extern int drm_sman_set_manager(struct drm_sman * sman, unsigned int mananger, - struct drm_sman_mm * allocator); - -/* - * Allocate a memory block. Aligment is not implemented yet. - */ - -extern struct drm_memblock_item *drm_sman_alloc(struct drm_sman * sman, - unsigned int manager, - unsigned long size, - unsigned alignment, - unsigned long owner); -/* - * Free a memory block identified by its user hash key. - */ - -extern int drm_sman_free_key(struct drm_sman * sman, unsigned int key); - -/* - * returns 1 iff there are no stale memory blocks associated with this owner. - * Typically called to determine if we need to idle the hardware and call - * drm_sman_owner_cleanup. If there are no stale memory blocks, it removes all - * resources associated with owner. - */ - -extern int drm_sman_owner_clean(struct drm_sman * sman, unsigned long owner); - -/* - * Frees all stale memory blocks associated with this owner. Note that this - * requires that the hardware is finished with all blocks, so the graphics engine - * should be idled before this call is made. This function also frees - * any resources associated with "owner" and should be called when owner - * is not going to be referenced anymore. - */ - -extern void drm_sman_owner_cleanup(struct drm_sman * sman, unsigned long owner); - -/* - * Frees all stale memory blocks associated with the memory manager. - * See idling above. - */ - -extern void drm_sman_cleanup(struct drm_sman * sman); - -#endif diff --git a/include/drm/exynos_drm.h b/include/drm/exynos_drm.h index 1205043..e478de4 100644 --- a/include/drm/exynos_drm.h +++ b/include/drm/exynos_drm.h @@ -74,9 +74,38 @@ struct drm_exynos_gem_mmap { uint64_t mapped; }; +/** + * A structure for user connection request of virtual display. + * + * @connection: indicate whether doing connetion or not by user. + * @extensions: if this value is 1 then the vidi driver would need additional + * 128bytes edid data. + * @edid: the edid data pointer from user side. + */ +struct drm_exynos_vidi_connection { + unsigned int connection; + unsigned int extensions; + uint64_t edid; +}; + +struct drm_exynos_plane_set_zpos { + __u32 plane_id; + __s32 zpos; +}; + +/* memory type definitions. */ +enum e_drm_exynos_gem_mem_type { + /* Physically Non-Continuous memory. */ + EXYNOS_BO_NONCONTIG = 1 << 0, + EXYNOS_BO_MASK = EXYNOS_BO_NONCONTIG +}; + #define DRM_EXYNOS_GEM_CREATE 0x00 #define DRM_EXYNOS_GEM_MAP_OFFSET 0x01 #define DRM_EXYNOS_GEM_MMAP 0x02 +/* Reserved 0x03 ~ 0x05 for exynos specific gem ioctl */ +#define DRM_EXYNOS_PLANE_SET_ZPOS 0x06 +#define DRM_EXYNOS_VIDI_CONNECTION 0x07 #define DRM_IOCTL_EXYNOS_GEM_CREATE DRM_IOWR(DRM_COMMAND_BASE + \ DRM_EXYNOS_GEM_CREATE, struct drm_exynos_gem_create) @@ -87,19 +116,70 @@ struct drm_exynos_gem_mmap { #define DRM_IOCTL_EXYNOS_GEM_MMAP DRM_IOWR(DRM_COMMAND_BASE + \ DRM_EXYNOS_GEM_MMAP, struct drm_exynos_gem_mmap) +#define DRM_IOCTL_EXYNOS_PLANE_SET_ZPOS DRM_IOWR(DRM_COMMAND_BASE + \ + DRM_EXYNOS_PLANE_SET_ZPOS, struct drm_exynos_plane_set_zpos) + +#define DRM_IOCTL_EXYNOS_VIDI_CONNECTION DRM_IOWR(DRM_COMMAND_BASE + \ + DRM_EXYNOS_VIDI_CONNECTION, struct drm_exynos_vidi_connection) + +#ifdef __KERNEL__ + /** - * Platform Specific Structure for DRM based FIMD. + * A structure for lcd panel information. * * @timing: default video mode for initializing + * @width_mm: physical size of lcd width. + * @height_mm: physical size of lcd height. + */ +struct exynos_drm_panel_info { + struct fb_videomode timing; + u32 width_mm; + u32 height_mm; +}; + +/** + * Platform Specific Structure for DRM based FIMD. + * + * @panel: default panel info for initializing * @default_win: default window layer number to be used for UI. * @bpp: default bit per pixel. */ struct exynos_drm_fimd_pdata { - struct fb_videomode timing; + struct exynos_drm_panel_info panel; u32 vidcon0; u32 vidcon1; unsigned int default_win; unsigned int bpp; }; -#endif +/** + * Platform Specific Structure for DRM based HDMI. + * + * @hdmi_dev: device point to specific hdmi driver. + * @mixer_dev: device point to specific mixer driver. + * + * this structure is used for common hdmi driver and each device object + * would be used to access specific device driver(hdmi or mixer driver) + */ +struct exynos_drm_common_hdmi_pd { + struct device *hdmi_dev; + struct device *mixer_dev; +}; + +/** + * Platform Specific Structure for DRM based HDMI core. + * + * @timing: default video mode for initializing + * @default_win: default window layer number to be used for UI. + * @bpp: default bit per pixel. + * @is_v13: set if hdmi version 13 is. + */ +struct exynos_drm_hdmi_pdata { + struct fb_videomode timing; + unsigned int default_win; + unsigned int bpp; + unsigned int is_v13:1; +}; + +#endif /* __KERNEL__ */ +#endif /* _EXYNOS_DRM_H_ */ diff --git a/include/drm/gma_drm.h b/include/drm/gma_drm.h new file mode 100644 index 0000000..884613e --- /dev/null +++ b/include/drm/gma_drm.h @@ -0,0 +1,91 @@ +/************************************************************************** + * Copyright (c) 2007-2011, Intel Corporation. + * All Rights Reserved. + * Copyright (c) 2008, Tungsten Graphics Inc. Cedar Park, TX., USA. + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. + * + **************************************************************************/ + +#ifndef _PSB_DRM_H_ +#define _PSB_DRM_H_ + +/* + * Manage the LUT for an output + */ +struct drm_psb_dpst_lut_arg { + uint8_t lut[256]; + int output_id; +}; + +/* + * Validate modes + */ +struct drm_psb_mode_operation_arg { + u32 obj_id; + u16 operation; + struct drm_mode_modeinfo mode; + u64 data; +}; + +/* + * Query the stolen memory for smarter management of + * memory by the server + */ +struct drm_psb_stolen_memory_arg { + u32 base; + u32 size; +}; + +struct drm_psb_get_pipe_from_crtc_id_arg { + /** ID of CRTC being requested **/ + u32 crtc_id; + /** pipe of requested CRTC **/ + u32 pipe; +}; + +struct drm_psb_gem_create { + __u64 size; + __u32 handle; + __u32 flags; +#define GMA_GEM_CREATE_STOLEN 1 /* Stolen memory can be used */ +}; + +struct drm_psb_gem_mmap { + __u32 handle; + __u32 pad; + /** + * Fake offset to use for subsequent mmap call + * + * This is a fixed-size type for 32/64 compatibility. + */ + __u64 offset; +}; + +/* Controlling the kernel modesetting buffers */ + +#define DRM_GMA_GEM_CREATE 0x00 /* Create a GEM object */ +#define DRM_GMA_GEM_MMAP 0x01 /* Map GEM memory */ +#define DRM_GMA_STOLEN_MEMORY 0x02 /* Report stolen memory */ +#define DRM_GMA_2D_OP 0x03 /* Will be merged later */ +#define DRM_GMA_GAMMA 0x04 /* Set gamma table */ +#define DRM_GMA_ADB 0x05 /* Get backlight */ +#define DRM_GMA_DPST_BL 0x06 /* Set backlight */ +#define DRM_GMA_MODE_OPERATION 0x07 /* Mode validation/DC set */ +#define PSB_MODE_OPERATION_MODE_VALID 0x01 +#define DRM_GMA_GET_PIPE_FROM_CRTC_ID 0x08 /* CRTC to physical pipe# */ + + +#endif diff --git a/include/drm/i915_drm.h b/include/drm/i915_drm.h index 28c0d11..da929bb 100644 --- a/include/drm/i915_drm.h +++ b/include/drm/i915_drm.h @@ -198,6 +198,8 @@ typedef struct _drm_i915_sarea { #define DRM_I915_OVERLAY_PUT_IMAGE 0x27 #define DRM_I915_OVERLAY_ATTRS 0x28 #define DRM_I915_GEM_EXECBUFFER2 0x29 +#define DRM_I915_GET_SPRITE_COLORKEY 0x2a +#define DRM_I915_SET_SPRITE_COLORKEY 0x2b #define DRM_IOCTL_I915_INIT DRM_IOW( DRM_COMMAND_BASE + DRM_I915_INIT, drm_i915_init_t) #define DRM_IOCTL_I915_FLUSH DRM_IO ( DRM_COMMAND_BASE + DRM_I915_FLUSH) @@ -239,6 +241,8 @@ typedef struct _drm_i915_sarea { #define DRM_IOCTL_I915_GEM_MADVISE DRM_IOWR(DRM_COMMAND_BASE + DRM_I915_GEM_MADVISE, struct drm_i915_gem_madvise) #define DRM_IOCTL_I915_OVERLAY_PUT_IMAGE DRM_IOW(DRM_COMMAND_BASE + DRM_I915_OVERLAY_PUT_IMAGE, struct drm_intel_overlay_put_image) #define DRM_IOCTL_I915_OVERLAY_ATTRS DRM_IOWR(DRM_COMMAND_BASE + DRM_I915_OVERLAY_ATTRS, struct drm_intel_overlay_attrs) +#define DRM_IOCTL_I915_SET_SPRITE_COLORKEY DRM_IOWR(DRM_COMMAND_BASE + DRM_I915_SET_SPRITE_COLORKEY, struct drm_intel_sprite_colorkey) +#define DRM_IOCTL_I915_GET_SPRITE_COLORKEY DRM_IOWR(DRM_COMMAND_BASE + DRM_I915_SET_SPRITE_COLORKEY, struct drm_intel_sprite_colorkey) /* Allow drivers to submit batchbuffers directly to hardware, relying * on the security mechanisms provided by hardware. @@ -291,6 +295,8 @@ typedef struct drm_i915_irq_wait { #define I915_PARAM_HAS_COHERENT_RINGS 13 #define I915_PARAM_HAS_EXEC_CONSTANTS 14 #define I915_PARAM_HAS_RELAXED_DELTA 15 +#define I915_PARAM_HAS_GEN7_SOL_RESET 16 +#define I915_PARAM_HAS_LLC 17 typedef struct drm_i915_getparam { int param; @@ -653,6 +659,9 @@ struct drm_i915_gem_execbuffer2 { __u64 rsvd2; }; +/** Resets the SO write offset registers for transform feedback on gen7. */ +#define I915_EXEC_GEN7_SOL_RESET (1<<8) + struct drm_i915_gem_pin { /** Handle of the buffer to be pinned. */ __u32 handle; @@ -844,4 +853,36 @@ struct drm_intel_overlay_attrs { __u32 gamma5; }; +/* + * Intel sprite handling + * + * Color keying works with a min/mask/max tuple. Both source and destination + * color keying is allowed. + * + * Source keying: + * Sprite pixels within the min & max values, masked against the color channels + * specified in the mask field, will be transparent. All other pixels will + * be displayed on top of the primary plane. For RGB surfaces, only the min + * and mask fields will be used; ranged compares are not allowed. + * + * Destination keying: + * Primary plane pixels that match the min value, masked against the color + * channels specified in the mask field, will be replaced by corresponding + * pixels from the sprite plane. + * + * Note that source & destination keying are exclusive; only one can be + * active on a given plane. + */ + +#define I915_SET_COLORKEY_NONE (1<<0) /* disable color key matching */ +#define I915_SET_COLORKEY_DESTINATION (1<<1) +#define I915_SET_COLORKEY_SOURCE (1<<2) +struct drm_intel_sprite_colorkey { + __u32 plane_id; + __u32 min_value; + __u32 channel_mask; + __u32 max_value; + __u32 flags; +}; + #endif /* _I915_DRM_H_ */ diff --git a/include/drm/intel-gtt.h b/include/drm/intel-gtt.h index b174620..923afb5 100644 --- a/include/drm/intel-gtt.h +++ b/include/drm/intel-gtt.h @@ -15,6 +15,10 @@ const struct intel_gtt { unsigned int needs_dmar : 1; /* Whether we idle the gpu before mapping/unmapping */ unsigned int do_idle_maps : 1; + /* Share the scratch page dma with ppgtts. */ + dma_addr_t scratch_page_dma; + /* for ppgtt PDE access */ + u32 __iomem *gtt; } *intel_gtt_get(void); void intel_gtt_chipset_flush(void); @@ -40,4 +44,8 @@ void intel_gtt_insert_pages(unsigned int first_entry, unsigned int num_entries, /* flag for GFDT type */ #define AGP_USER_CACHED_MEMORY_GFDT (1 << 3) +#ifdef CONFIG_INTEL_IOMMU +extern int intel_iommu_gfx_mapped; +#endif + #endif diff --git a/include/drm/radeon_drm.h b/include/drm/radeon_drm.h index be94be6..7c491b4 100644 --- a/include/drm/radeon_drm.h +++ b/include/drm/radeon_drm.h @@ -509,6 +509,7 @@ typedef struct { #define DRM_RADEON_GEM_SET_TILING 0x28 #define DRM_RADEON_GEM_GET_TILING 0x29 #define DRM_RADEON_GEM_BUSY 0x2a +#define DRM_RADEON_GEM_VA 0x2b #define DRM_IOCTL_RADEON_CP_INIT DRM_IOW( DRM_COMMAND_BASE + DRM_RADEON_CP_INIT, drm_radeon_init_t) #define DRM_IOCTL_RADEON_CP_START DRM_IO( DRM_COMMAND_BASE + DRM_RADEON_CP_START) @@ -550,6 +551,7 @@ typedef struct { #define DRM_IOCTL_RADEON_GEM_SET_TILING DRM_IOWR(DRM_COMMAND_BASE + DRM_RADEON_GEM_SET_TILING, struct drm_radeon_gem_set_tiling) #define DRM_IOCTL_RADEON_GEM_GET_TILING DRM_IOWR(DRM_COMMAND_BASE + DRM_RADEON_GEM_GET_TILING, struct drm_radeon_gem_get_tiling) #define DRM_IOCTL_RADEON_GEM_BUSY DRM_IOWR(DRM_COMMAND_BASE + DRM_RADEON_GEM_BUSY, struct drm_radeon_gem_busy) +#define DRM_IOCTL_RADEON_GEM_VA DRM_IOWR(DRM_COMMAND_BASE + DRM_RADEON_GEM_VA, struct drm_radeon_gem_va) typedef struct drm_radeon_init { enum { @@ -802,13 +804,23 @@ struct drm_radeon_gem_create { uint32_t flags; }; -#define RADEON_TILING_MACRO 0x1 -#define RADEON_TILING_MICRO 0x2 -#define RADEON_TILING_SWAP_16BIT 0x4 -#define RADEON_TILING_SWAP_32BIT 0x8 -#define RADEON_TILING_SURFACE 0x10 /* this object requires a surface - * when mapped - i.e. front buffer */ -#define RADEON_TILING_MICRO_SQUARE 0x20 +#define RADEON_TILING_MACRO 0x1 +#define RADEON_TILING_MICRO 0x2 +#define RADEON_TILING_SWAP_16BIT 0x4 +#define RADEON_TILING_SWAP_32BIT 0x8 +/* this object requires a surface when mapped - i.e. front buffer */ +#define RADEON_TILING_SURFACE 0x10 +#define RADEON_TILING_MICRO_SQUARE 0x20 +#define RADEON_TILING_EG_BANKW_SHIFT 8 +#define RADEON_TILING_EG_BANKW_MASK 0xf +#define RADEON_TILING_EG_BANKH_SHIFT 12 +#define RADEON_TILING_EG_BANKH_MASK 0xf +#define RADEON_TILING_EG_MACRO_TILE_ASPECT_SHIFT 16 +#define RADEON_TILING_EG_MACRO_TILE_ASPECT_MASK 0xf +#define RADEON_TILING_EG_TILE_SPLIT_SHIFT 24 +#define RADEON_TILING_EG_TILE_SPLIT_MASK 0xf +#define RADEON_TILING_EG_STENCIL_TILE_SPLIT_SHIFT 28 +#define RADEON_TILING_EG_STENCIL_TILE_SPLIT_MASK 0xf struct drm_radeon_gem_set_tiling { uint32_t handle; @@ -872,12 +884,40 @@ struct drm_radeon_gem_pwrite { uint64_t data_ptr; }; +#define RADEON_VA_MAP 1 +#define RADEON_VA_UNMAP 2 + +#define RADEON_VA_RESULT_OK 0 +#define RADEON_VA_RESULT_ERROR 1 +#define RADEON_VA_RESULT_VA_EXIST 2 + +#define RADEON_VM_PAGE_VALID (1 << 0) +#define RADEON_VM_PAGE_READABLE (1 << 1) +#define RADEON_VM_PAGE_WRITEABLE (1 << 2) +#define RADEON_VM_PAGE_SYSTEM (1 << 3) +#define RADEON_VM_PAGE_SNOOPED (1 << 4) + +struct drm_radeon_gem_va { + uint32_t handle; + uint32_t operation; + uint32_t vm_id; + uint32_t flags; + uint64_t offset; +}; + #define RADEON_CHUNK_ID_RELOCS 0x01 #define RADEON_CHUNK_ID_IB 0x02 #define RADEON_CHUNK_ID_FLAGS 0x03 +#define RADEON_CHUNK_ID_CONST_IB 0x04 /* The first dword of RADEON_CHUNK_ID_FLAGS is a uint32 of these flags: */ #define RADEON_CS_KEEP_TILING_FLAGS 0x01 +#define RADEON_CS_USE_VM 0x02 +/* The second dword of RADEON_CHUNK_ID_FLAGS is a uint32 that sets the ring type */ +#define RADEON_CS_RING_GFX 0 +#define RADEON_CS_RING_COMPUTE 1 +/* The third dword of RADEON_CHUNK_ID_FLAGS is a sint32 that sets the priority */ +/* 0 = normal, + = higher priority, - = lower priority */ struct drm_radeon_cs_chunk { uint32_t chunk_id; @@ -885,6 +925,9 @@ struct drm_radeon_cs_chunk { uint64_t chunk_data; }; +/* drm_radeon_cs_reloc.flags */ +#define RADEON_RELOC_DONT_SYNC 0x01 + struct drm_radeon_cs_reloc { uint32_t handle; uint32_t read_domains; @@ -916,6 +959,12 @@ struct drm_radeon_cs { #define RADEON_INFO_NUM_TILE_PIPES 0x0b /* tile pipes for r600+ */ #define RADEON_INFO_FUSION_GART_WORKING 0x0c /* fusion writes to GTT were broken before this */ #define RADEON_INFO_BACKEND_MAP 0x0d /* pipe to backend map, needed by mesa */ +/* virtual address start, va < start are reserved by the kernel */ +#define RADEON_INFO_VA_START 0x0e +/* maximum size of ib using the virtual memory cs */ +#define RADEON_INFO_IB_VM_MAX_SIZE 0x0f +/* max pipes - needed for compute shaders */ +#define RADEON_INFO_MAX_PIPES 0x10 struct drm_radeon_info { uint32_t request; diff --git a/include/drm/sis_drm.h b/include/drm/sis_drm.h index 30f7b38..035b804 100644 --- a/include/drm/sis_drm.h +++ b/include/drm/sis_drm.h @@ -64,4 +64,8 @@ typedef struct { unsigned int offset, size; } drm_sis_fb_t; +struct sis_file_private { + struct list_head obj_list; +}; + #endif /* __SIS_DRM_H__ */ diff --git a/include/drm/ttm/ttm_bo_api.h b/include/drm/ttm/ttm_bo_api.h index 42e3469..974c8f8 100644 --- a/include/drm/ttm/ttm_bo_api.h +++ b/include/drm/ttm/ttm_bo_api.h @@ -122,17 +122,12 @@ struct ttm_mem_reg { * be mmapped by user space. Each of these bos occupy a slot in the * device address space, that can be used for normal vm operations. * - * @ttm_bo_type_user: These are user-space memory areas that are made - * available to the GPU by mapping the buffer pages into the GPU aperture - * space. These buffers cannot be mmaped from the device address space. - * * @ttm_bo_type_kernel: These buffers are like ttm_bo_type_device buffers, * but they cannot be accessed from user-space. For kernel-only use. */ enum ttm_bo_type { ttm_bo_type_device, - ttm_bo_type_user, ttm_bo_type_kernel }; @@ -434,9 +429,9 @@ extern void ttm_bo_unlock_delayed_workqueue(struct ttm_bo_device *bdev, * -EBUSY if the buffer is busy and no_wait is true. * -ERESTARTSYS if interrupted by a signal. */ - extern int ttm_bo_synccpu_write_grab(struct ttm_buffer_object *bo, bool no_wait); + /** * ttm_bo_synccpu_write_release: * @@ -447,6 +442,22 @@ ttm_bo_synccpu_write_grab(struct ttm_buffer_object *bo, bool no_wait); extern void ttm_bo_synccpu_write_release(struct ttm_buffer_object *bo); /** + * ttm_bo_acc_size + * + * @bdev: Pointer to a ttm_bo_device struct. + * @bo_size: size of the buffer object in byte. + * @struct_size: size of the structure holding buffer object datas + * + * Returns size to account for a buffer object + */ +size_t ttm_bo_acc_size(struct ttm_bo_device *bdev, + unsigned long bo_size, + unsigned struct_size); +size_t ttm_bo_dma_acc_size(struct ttm_bo_device *bdev, + unsigned long bo_size, + unsigned struct_size); + +/** * ttm_bo_init * * @bdev: Pointer to a ttm_bo_device struct. @@ -493,6 +504,7 @@ extern int ttm_bo_init(struct ttm_bo_device *bdev, struct file *persistent_swap_storage, size_t acc_size, void (*destroy) (struct ttm_buffer_object *)); + /** * ttm_bo_synccpu_object_init * diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h index 94eb143..d43e892 100644 --- a/include/drm/ttm/ttm_bo_driver.h +++ b/include/drm/ttm/ttm_bo_driver.h @@ -43,36 +43,9 @@ struct ttm_backend; struct ttm_backend_func { /** - * struct ttm_backend_func member populate - * - * @backend: Pointer to a struct ttm_backend. - * @num_pages: Number of pages to populate. - * @pages: Array of pointers to ttm pages. - * @dummy_read_page: Page to be used instead of NULL pages in the - * array @pages. - * @dma_addrs: Array of DMA (bus) address of the ttm pages. - * - * Populate the backend with ttm pages. Depending on the backend, - * it may or may not copy the @pages array. - */ - int (*populate) (struct ttm_backend *backend, - unsigned long num_pages, struct page **pages, - struct page *dummy_read_page, - dma_addr_t *dma_addrs); - /** - * struct ttm_backend_func member clear - * - * @backend: Pointer to a struct ttm_backend. - * - * This is an "unpopulate" function. Release all resources - * allocated with populate. - */ - void (*clear) (struct ttm_backend *backend); - - /** * struct ttm_backend_func member bind * - * @backend: Pointer to a struct ttm_backend. + * @ttm: Pointer to a struct ttm_tt. * @bo_mem: Pointer to a struct ttm_mem_reg describing the * memory type and location for binding. * @@ -80,46 +53,29 @@ struct ttm_backend_func { * indicated by @bo_mem. This function should be able to handle * differences between aperture and system page sizes. */ - int (*bind) (struct ttm_backend *backend, struct ttm_mem_reg *bo_mem); + int (*bind) (struct ttm_tt *ttm, struct ttm_mem_reg *bo_mem); /** * struct ttm_backend_func member unbind * - * @backend: Pointer to a struct ttm_backend. + * @ttm: Pointer to a struct ttm_tt. * * Unbind previously bound backend pages. This function should be * able to handle differences between aperture and system page sizes. */ - int (*unbind) (struct ttm_backend *backend); + int (*unbind) (struct ttm_tt *ttm); /** * struct ttm_backend_func member destroy * - * @backend: Pointer to a struct ttm_backend. + * @ttm: Pointer to a struct ttm_tt. * - * Destroy the backend. + * Destroy the backend. This will be call back from ttm_tt_destroy so + * don't call ttm_tt_destroy from the callback or infinite loop. */ - void (*destroy) (struct ttm_backend *backend); -}; - -/** - * struct ttm_backend - * - * @bdev: Pointer to a struct ttm_bo_device. - * @flags: For driver use. - * @func: Pointer to a struct ttm_backend_func that describes - * the backend methods. - * - */ - -struct ttm_backend { - struct ttm_bo_device *bdev; - uint32_t flags; - struct ttm_backend_func *func; + void (*destroy) (struct ttm_tt *ttm); }; -#define TTM_PAGE_FLAG_USER (1 << 1) -#define TTM_PAGE_FLAG_USER_DIRTY (1 << 2) #define TTM_PAGE_FLAG_WRITE (1 << 3) #define TTM_PAGE_FLAG_SWAPPED (1 << 4) #define TTM_PAGE_FLAG_PERSISTENT_SWAP (1 << 5) @@ -135,23 +91,18 @@ enum ttm_caching_state { /** * struct ttm_tt * + * @bdev: Pointer to a struct ttm_bo_device. + * @func: Pointer to a struct ttm_backend_func that describes + * the backend methods. * @dummy_read_page: Page to map where the ttm_tt page array contains a NULL * pointer. * @pages: Array of pages backing the data. - * @first_himem_page: Himem pages are put last in the page array, which - * enables us to run caching attribute changes on only the first part - * of the page array containing lomem pages. This is the index of the - * first himem page. - * @last_lomem_page: Index of the last lomem page in the page array. * @num_pages: Number of pages in the page array. * @bdev: Pointer to the current struct ttm_bo_device. * @be: Pointer to the ttm backend. - * @tsk: The task for user ttm. - * @start: virtual address for user ttm. * @swap_storage: Pointer to shmem struct file for swap storage. * @caching_state: The current caching state of the pages. * @state: The current binding state of the pages. - * @dma_address: The DMA (bus) addresses of the pages (if TTM_PAGE_FLAG_DMA32) * * This is a structure holding the pages, caching- and aperture binding * status for a buffer object that isn't backed by fixed (VRAM / AGP) @@ -159,16 +110,14 @@ enum ttm_caching_state { */ struct ttm_tt { + struct ttm_bo_device *bdev; + struct ttm_backend_func *func; struct page *dummy_read_page; struct page **pages; - long first_himem_page; - long last_lomem_page; uint32_t page_flags; unsigned long num_pages; struct ttm_bo_global *glob; struct ttm_backend *be; - struct task_struct *tsk; - unsigned long start; struct file *swap_storage; enum ttm_caching_state caching_state; enum { @@ -176,7 +125,23 @@ struct ttm_tt { tt_unbound, tt_unpopulated, } state; +}; + +/** + * struct ttm_dma_tt + * + * @ttm: Base ttm_tt struct. + * @dma_address: The DMA (bus) addresses of the pages + * @pages_list: used by some page allocation backend + * + * This is a structure holding the pages, caching- and aperture binding + * status for a buffer object that isn't backed by fixed (VRAM / AGP) + * memory. + */ +struct ttm_dma_tt { + struct ttm_tt ttm; dma_addr_t *dma_address; + struct list_head pages_list; }; #define TTM_MEMTYPE_FLAG_FIXED (1 << 0) /* Fixed (on-card) PCI memory */ @@ -351,15 +316,42 @@ struct ttm_mem_type_manager { struct ttm_bo_driver { /** - * struct ttm_bo_driver member create_ttm_backend_entry + * ttm_tt_create * - * @bdev: The buffer object device. + * @bdev: pointer to a struct ttm_bo_device: + * @size: Size of the data needed backing. + * @page_flags: Page flags as identified by TTM_PAGE_FLAG_XX flags. + * @dummy_read_page: See struct ttm_bo_device. * - * Create a driver specific struct ttm_backend. + * Create a struct ttm_tt to back data with system memory pages. + * No pages are actually allocated. + * Returns: + * NULL: Out of memory. */ + struct ttm_tt *(*ttm_tt_create)(struct ttm_bo_device *bdev, + unsigned long size, + uint32_t page_flags, + struct page *dummy_read_page); - struct ttm_backend *(*create_ttm_backend_entry) - (struct ttm_bo_device *bdev); + /** + * ttm_tt_populate + * + * @ttm: The struct ttm_tt to contain the backing pages. + * + * Allocate all backing pages + * Returns: + * -ENOMEM: Out of memory. + */ + int (*ttm_tt_populate)(struct ttm_tt *ttm); + + /** + * ttm_tt_unpopulate + * + * @ttm: The struct ttm_tt to contain the backing pages. + * + * Free all backing page + */ + void (*ttm_tt_unpopulate)(struct ttm_tt *ttm); /** * struct ttm_bo_driver member invalidate_caches @@ -477,9 +469,6 @@ struct ttm_bo_global_ref { * @dummy_read_page: Pointer to a dummy page used for mapping requests * of unpopulated pages. * @shrink: A shrink callback object used for buffer object swap. - * @ttm_bo_extra_size: Extra size (sizeof(struct ttm_buffer_object) excluded) - * used by a buffer object. This is excluding page arrays and backing pages. - * @ttm_bo_size: This is @ttm_bo_extra_size + sizeof(struct ttm_buffer_object). * @device_list_mutex: Mutex protecting the device list. * This mutex is held while traversing the device list for pm options. * @lru_lock: Spinlock protecting the bo subsystem lru lists. @@ -497,8 +486,6 @@ struct ttm_bo_global { struct ttm_mem_global *mem_glob; struct page *dummy_read_page; struct ttm_mem_shrink shrink; - size_t ttm_bo_extra_size; - size_t ttm_bo_size; struct mutex device_list_mutex; spinlock_t lru_lock; @@ -600,8 +587,9 @@ ttm_flag_masked(uint32_t *old, uint32_t new, uint32_t mask) } /** - * ttm_tt_create + * ttm_tt_init * + * @ttm: The struct ttm_tt. * @bdev: pointer to a struct ttm_bo_device: * @size: Size of the data needed backing. * @page_flags: Page flags as identified by TTM_PAGE_FLAG_XX flags. @@ -612,28 +600,22 @@ ttm_flag_masked(uint32_t *old, uint32_t new, uint32_t mask) * Returns: * NULL: Out of memory. */ -extern struct ttm_tt *ttm_tt_create(struct ttm_bo_device *bdev, - unsigned long size, - uint32_t page_flags, - struct page *dummy_read_page); +extern int ttm_tt_init(struct ttm_tt *ttm, struct ttm_bo_device *bdev, + unsigned long size, uint32_t page_flags, + struct page *dummy_read_page); +extern int ttm_dma_tt_init(struct ttm_dma_tt *ttm_dma, struct ttm_bo_device *bdev, + unsigned long size, uint32_t page_flags, + struct page *dummy_read_page); /** - * ttm_tt_set_user: + * ttm_tt_fini * - * @ttm: The struct ttm_tt to populate. - * @tsk: A struct task_struct for which @start is a valid user-space address. - * @start: A valid user-space address. - * @num_pages: Size in pages of the user memory area. + * @ttm: the ttm_tt structure. * - * Populate a struct ttm_tt with a user-space memory area after first pinning - * the pages backing it. - * Returns: - * !0: Error. + * Free memory of ttm_tt structure */ - -extern int ttm_tt_set_user(struct ttm_tt *ttm, - struct task_struct *tsk, - unsigned long start, unsigned long num_pages); +extern void ttm_tt_fini(struct ttm_tt *ttm); +extern void ttm_dma_tt_fini(struct ttm_dma_tt *ttm_dma); /** * ttm_ttm_bind: @@ -646,20 +628,11 @@ extern int ttm_tt_set_user(struct ttm_tt *ttm, extern int ttm_tt_bind(struct ttm_tt *ttm, struct ttm_mem_reg *bo_mem); /** - * ttm_tt_populate: - * - * @ttm: The struct ttm_tt to contain the backing pages. - * - * Add backing pages to all of @ttm - */ -extern int ttm_tt_populate(struct ttm_tt *ttm); - -/** * ttm_ttm_destroy: * * @ttm: The struct ttm_tt. * - * Unbind, unpopulate and destroy a struct ttm_tt. + * Unbind, unpopulate and destroy common struct ttm_tt. */ extern void ttm_tt_destroy(struct ttm_tt *ttm); @@ -673,19 +646,13 @@ extern void ttm_tt_destroy(struct ttm_tt *ttm); extern void ttm_tt_unbind(struct ttm_tt *ttm); /** - * ttm_ttm_destroy: + * ttm_tt_swapin: * * @ttm: The struct ttm_tt. - * @index: Index of the desired page. - * - * Return a pointer to the struct page backing @ttm at page - * index @index. If the page is unpopulated, one will be allocated to - * populate that index. * - * Returns: - * NULL on OOM. + * Swap in a previously swap out ttm_tt. */ -extern struct page *ttm_tt_get_page(struct ttm_tt *ttm, int index); +extern int ttm_tt_swapin(struct ttm_tt *ttm); /** * ttm_tt_cache_flush: @@ -1046,17 +1013,25 @@ extern const struct ttm_mem_type_manager_func ttm_bo_manager_func; #include /** - * ttm_agp_backend_init + * ttm_agp_tt_create * * @bdev: Pointer to a struct ttm_bo_device. * @bridge: The agp bridge this device is sitting on. + * @size: Size of the data needed backing. + * @page_flags: Page flags as identified by TTM_PAGE_FLAG_XX flags. + * @dummy_read_page: See struct ttm_bo_device. + * * * Create a TTM backend that uses the indicated AGP bridge as an aperture * for TT memory. This function uses the linux agpgart interface to * bind and unbind memory backing a ttm_tt. */ -extern struct ttm_backend *ttm_agp_backend_init(struct ttm_bo_device *bdev, - struct agp_bridge_data *bridge); +extern struct ttm_tt *ttm_agp_tt_create(struct ttm_bo_device *bdev, + struct agp_bridge_data *bridge, + unsigned long size, uint32_t page_flags, + struct page *dummy_read_page); +int ttm_agp_tt_populate(struct ttm_tt *ttm); +void ttm_agp_tt_unpopulate(struct ttm_tt *ttm); #endif #endif diff --git a/include/drm/ttm/ttm_memory.h b/include/drm/ttm/ttm_memory.h index 26c1f78..d6d1da4 100644 --- a/include/drm/ttm/ttm_memory.h +++ b/include/drm/ttm/ttm_memory.h @@ -30,6 +30,7 @@ #include #include +#include #include #include #include diff --git a/include/drm/ttm/ttm_page_alloc.h b/include/drm/ttm/ttm_page_alloc.h index 129de12..5fe2740 100644 --- a/include/drm/ttm/ttm_page_alloc.h +++ b/include/drm/ttm/ttm_page_alloc.h @@ -30,45 +30,70 @@ #include "ttm_memory.h" /** - * Get count number of pages from pool to pages list. + * Initialize pool allocator. + */ +int ttm_page_alloc_init(struct ttm_mem_global *glob, unsigned max_pages); +/** + * Free pool allocator. + */ +void ttm_page_alloc_fini(void); + +/** + * ttm_pool_populate: + * + * @ttm: The struct ttm_tt to contain the backing pages. * - * @pages: head of empty linked list where pages are filled. - * @flags: ttm flags for page allocation. - * @cstate: ttm caching state for the page. - * @count: number of pages to allocate. - * @dma_address: The DMA (bus) address of pages (if TTM_PAGE_FLAG_DMA32 set). + * Add backing pages to all of @ttm */ -int ttm_get_pages(struct list_head *pages, - int flags, - enum ttm_caching_state cstate, - unsigned count, - dma_addr_t *dma_address); +extern int ttm_pool_populate(struct ttm_tt *ttm); + /** - * Put linked list of pages to pool. + * ttm_pool_unpopulate: + * + * @ttm: The struct ttm_tt which to free backing pages. * - * @pages: list of pages to free. - * @page_count: number of pages in the list. Zero can be passed for unknown - * count. - * @flags: ttm flags for page allocation. - * @cstate: ttm caching state. - * @dma_address: The DMA (bus) address of pages (if TTM_PAGE_FLAG_DMA32 set). + * Free all pages of @ttm */ -void ttm_put_pages(struct list_head *pages, - unsigned page_count, - int flags, - enum ttm_caching_state cstate, - dma_addr_t *dma_address); +extern void ttm_pool_unpopulate(struct ttm_tt *ttm); + +/** + * Output the state of pools to debugfs file + */ +extern int ttm_page_alloc_debugfs(struct seq_file *m, void *data); + + +#ifdef CONFIG_SWIOTLB /** * Initialize pool allocator. */ -int ttm_page_alloc_init(struct ttm_mem_global *glob, unsigned max_pages); +int ttm_dma_page_alloc_init(struct ttm_mem_global *glob, unsigned max_pages); + /** * Free pool allocator. */ -void ttm_page_alloc_fini(void); +void ttm_dma_page_alloc_fini(void); /** * Output the state of pools to debugfs file */ -extern int ttm_page_alloc_debugfs(struct seq_file *m, void *data); +extern int ttm_dma_page_alloc_debugfs(struct seq_file *m, void *data); + +extern int ttm_dma_populate(struct ttm_dma_tt *ttm_dma, struct device *dev); +extern void ttm_dma_unpopulate(struct ttm_dma_tt *ttm_dma, struct device *dev); + +#else +static inline int ttm_dma_page_alloc_init(struct ttm_mem_global *glob, + unsigned max_pages) +{ + return -ENODEV; +} + +static inline void ttm_dma_page_alloc_fini(void) { return; } + +static inline int ttm_dma_page_alloc_debugfs(struct seq_file *m, void *data) +{ + return 0; +} +#endif + #endif diff --git a/include/drm/via_drm.h b/include/drm/via_drm.h index fd11a5b..79b3b6e 100644 --- a/include/drm/via_drm.h +++ b/include/drm/via_drm.h @@ -274,4 +274,8 @@ typedef struct drm_via_dmablit { drm_via_blitsync_t sync; } drm_via_dmablit_t; +struct via_file_private { + struct list_head obj_list; +}; + #endif /* _VIA_DRM_H_ */