The current vector table setup has some shortcomings. First of all
currently the case when the high vectors are inside SDRAM (that is,
SDRAM reaches the end of the address space) is not supported. In this
case we create a secondary page table for the section containing the
vectors which gets overwritten by the general SDRAM secondary page
table entries creation afterwards. On ARMv7 and later the exception
table setup can be improved: Here the vector table address is configurable
in the VBAR register. We can use this register to skip remapping the
vector table.
With this patch we first try to use the VBAR register before doing
something else. Also, when we have to use the high vectors we first
try a request_sdram_region to test if the vector table memory is already
mapped. While at it sprinkle some comments into the code.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The last 64KiB of address space may be used for the vector table at
0xffff0000, so we cannot use it for barebox. The easiest way to archieve
this is to never use the last 64KiB of memory.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
When SDRAM reaches to the end of the address space the comparison
membase + memsize evaluates to 0, so pc - membase < memsize can never
be true. Fix this by substracting membase on both sides of the
comparison.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The name is more suitable for what the function does. Also let the
function return a pointer to the imd data found in the buffer.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Programs compiled for the target need -I $(srctree)/scripts/include/ to
be able to include for example linux/err.h.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
imd_concat_strings() and imd_string_data() are easier to handle when
they take a struct imd_header * instead of a struct imd_entry_string *.
Change this.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
CCM has reset status bits with more detailed information than the
watchdog. Set reset source with higher priority.
Signed-off-by: Alexander Stein <alexander.stein@systec-electronic.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
raw read/write was not possible because we enabled the ECC engine
during driver initialization. To support raw mode we have to disable
the ECC engine dynamically when needed. This has to be done before
the send_page function is called. The places where we have to disable
the ECC engine are not available in the driver, but are buried in the
Nand layer. To make them available we have to implement driver specific
write/read_page functions in which we configure the ECC mode.
This also makes the driver better readable as it makes the detour around
the driver internal data buffer when reading/writing pages unnecessary.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
MTD_DEBUG was only used by the i.MX Nand driver and is now unused.
Remove the mtd specific debug defines.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The relevant informations which command is sent for which page
is already contained in the informations imx_nand_command() prints, so
remove the debug prints in send_cmd and send_addr (which only exist for
v1/v2 controllers, not for v3). Also use dev_dbg to print debug
informations instead of MTD_DEBUG.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Neither CONFIG_MXC_NAND_LOW_LEVEL_ERASE nor CONFIG_MTD_NAND_MXC_FORCE_CE
are defined anywhere, remove the code inside the ifdefs.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The correct return value for a uncorrectable page is -EBADMSG, not -1
(which is -EPERM). Also the max_bitflips returned shall be the bitflips
per ecc step, not per page.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The driver returns wrong values for the ECC correction. For 2k pages
the controller reads and corrects in chunks of 512 bytes. The ECC status
register values are overwritten with each each new chunk read, so after
reading for chunks the .correct callback wil only return the ECC errors
for the last 512 byte chunk. ECC errors in the other three chunks remain
undetected.
Fix this by accumulating the ECC status while reading the chunks and
return the accumulated value in the .correct callback. Also return
-EBADMSG for a bad message and not -1.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
According to commit f6b6f3c7b2bb7d6277801882afdced6f2b10fc17 from
git://git.freescale.com/imx/uboot-imx.git:
Also set Prefetch offset to 15, since it improves
memcpy performance by 35%. Don't enable Incr double
Linefill enable since it adversely affects memcpy
performance by about 32MB/s and reads by 90MB/s. Tested
with 4K to 16MB sized src and dst aligned buffer.
This commit ports those chagnes from U-Boot.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
I.MX6Q Plus parts have r3p2 revision of PL310 so double linefill
errata no longer applies for all of the i.MX6Q SoCs. Change the code to
use PL310's revision inforation to determine if workaround needs to be
applied.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Use constants instead of magic numbers for PL301 registers bits in
imx6_mmu_init()
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Don't wait for more than one second for IP block to finish resetting. If
the block is dead it makes more sence to continue execution in hopes
that the rest of the processor is fine, rather than spin indefinetly
inside of the fec_probe function
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
In v4.2-rc3 the "fsl,imx1-gpt" compatible was replaced with
"fsl,imx21-gpt" so i.MX27 currently ends up without clocksource. Fix
this.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Add a proper check for I/O memory resource allocation failure and
replace dev_request_mem_region with dev_request_mem_resource so it would
be possible to correctly deallocate device's I/O resources when probe
fails.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Call clk_enable on mmdc_ch0_axi_podf in order to properly increase
reference counters for all of the nodes in this particular clock
path. Otherwise it becomes possible for peripherals, located on other
branches stemming from "periph", to shut down the whole clock tree (up
to "pll2_bus") when they try to manage their own local clocks.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Make sure that Barebox specific .dtsi files are included after .dtsi
files imported from Linux kernel. This way those local .dtsi files can
reference phandles defined in Linux kernel files.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>