MXS specific devices have some common infrastructure in the kernel
known as STMP devices. We have the same in barebox, but with a
mxs_ prefix instead of a stmp_ prefix. As some STMP devices are
also found on i.MX6 move the common infrastructure out of MXS
specific files and use the stmp_ prefix.
This is done in preparation for i.MX6 NAND support.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This updates the NAND stuff to Linux-3.11-rc1. It is synchronized
as best as we can get:
- locks removed
- The splitting in different files we had to better support different
features has been dropped. Instead this is now done mostly with the
use of __maybe_unused
Some barebox adjustments are forward ported, like:
- Allow partial page writes
- Optionally allow to erase bad blocks
- check for all_ff before writing a page
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This introduces the ecc stength fields in the structures and fills
them in, but leaves them unused right now.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
There is the same function name in the Kernel but with different
semantics. Rename to avoid naming conflicts when we update the
mtd support from the kernel.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This is based on Linux:
commit e2414f4c20bd4dc62186fbfd7bdec50bce6d2ead
Author: Brian Norris <computersforpeace@gmail.com>
Date: Mon Feb 6 13:44:00 2012 -0800
Currently, the flash-based BBT implementation writes bad block data only
to its flash-based table and not to the OOB marker area. Then, as new bad
blocks are marked over time, the OOB markers become incomplete and the
flash-based table becomes the only source of current bad block
information. This becomes an obvious problem when, for example:
* code accessing the flash cannot read the flash-based BBT format
* BBT is corrupted and the flash must be rescanned for bad
blocks; we want to remember bad blocks that were marked from Linux
So to keep the bad block markers in sync with the flash-based BBT, this
patch changes the default so that we write bad block markers to the proper
OOB area on each block in addition to flash-based BBT. Comments are
updated, expanded, and/or relocated as necessary.
The new flash-based BBT procedure for marking bad blocks:
(1) erase the affected block, to allow OOB marker to be written cleanly
(2) update in-memory BBT
(3) write bad block marker to OOB area of affected block
(4) update flash-based BBT
Note that we retain the first error encountered in (3) or (4), finish the
procedures, and dump the error in the end.
This should handle power cuts gracefully enough. (1) and (2) are mostly
harmless (note that (1) will not erase an already-recognized bad block).
The OOB and BBT may be "out of sync" if we experience power loss bewteen
(3) and (4), but we can reasonably expect that on next boot, subsequent
I/O operations will discover that the block should be marked bad again,
thus re-syncing the OOB and BBT.
Note that this is a change from the previous default flash-based BBT
behavior.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Chipsize didn't take number of LUNs into account. Sync chipsize calculation
to kernel commit 63795755
Tested with MT29F8G16ADBDAH4 on OMAP4
Signed-off-by: Jan Weitzel <j.weitzel@phytec.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Our erase command used to align to eraseblocks when necessary.
This worked well until recently when the m25p80, mtd_dataflash
and cfi flash were added / converted to mtd. This patch aligns
the input to the erase fileoperation to eraseblock boundaries.
Also tested with non uniform flashes with multiple eraseregions.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Since the info is device specific and not driver specific, attach
the callback to the device. This makes it possible to have a info
callback for a device which does not have a driver attached.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
If a device does not have bad blocks loop over the eraseblocks
in the driver instead of the core. This allows the mtd_dataflash
driver to erase blocks instead of pages to gain more speed during
erasing. Also the mtd_dataflash driver modifies the erase_info
struct which causes the outer loop in the core to never end.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Many cfi chips support 16 and 8 bit modes. Most important
difference is use of so called "Q15/A-1" pin. In 16bit mode this
pin is used for data IO. In 8bit mode, it is an address input
which add one more least significant bit (LSB). In this case
we should shift all adresses by one:
For example 0xaa << 1 = 0x154
Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
this fix the problems introduced when detecting non ONFI flashes in
commit 4c2bdc8728
"nand_base: detect more ONFI flash"
Signed-off-by: Eric Bénard <eric@eukrea.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
actually ops.ooboffs is not defaulted so when its value gets
added to chip->oob_poi in nand_fill_oob or nand_transfer_oob
the respective memcpy is using a wrong address.
With this patch, both md -s /dev/nandraw0 and cp xyz /dev/nandraw0.sb
are working fine on an i.MX28 target (instead of crashing the board).
Signed-off-by: Eric Bénard <eric@eukrea.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
if the flash has a known type, the ONFI detection won't occur
and thus we may not detect the right parameters.
By testing both namd and pagesize, as done in the kernel, we
can detect ONFI flash with know IDs.
As an example on an i.MX53 board :
- without the patch :
NAND device: Manufacturer ID: 0x2c, Chip ID: 0xd3
(Micron NAND 1GiB 3,3V 8-bit), page size: 4096, OOB size: 128
- with the patch :
ONFI flash detected ... ONFI param page 0 valid
NAND device: Manufacturer ID: 0x2c, Chip ID: 0xd3
(Micron MT29F8G08ABACAWP), page size: 4096, OOB size: 224
in the first case the OOB size is wrong.
Signed-off-by: Eric Bénard <eric@eukrea.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
With 2k NAND flashes the data layout in memory is not what is
written on the flash device. This leads to the result that the
factory provided bad block markers are not recognized correctly.
To preserve the factory bad block information the i.MX NAND driver
will not scan for the bad blocks itself when there is no flash based
bbt available, because the mtd layer would do so based on wrong
information. Instead, a new command is introduced which allows to
manually create a flash bbt based on the correct information.
As this command is tightly coupled to mtd and the i.MX NAND driver
the command is placed under drivers/mtd/nand/ instead of commands/
where a command normally belongs to.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The information which kind of bbt (if any) is used is hidden somewhere
in the NAND layer. Expose it to a device parameter to make it detectable
and visible during runtime.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This patch fixes following warning:
drivers/mtd/nand/nand_write.c: In function 'nand_do_write_ops':
drivers/mtd/nand/nand_write.c:272: warning: 'ret' may be used uninitialized in this function
Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
While erasing bad blocks is a potentially dangerous operation
it is sometimes needed during development or when some foreign
code has touched the flash.
This patch adds a device parameter 'erasebad' to allow erasing
bad blocks. Since this is not wanted during production this is
behind a Kconfig option.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This allows us to have some NAND specific stuff during registration,
like for example adding NAND specific device parameters.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Since recently we check for the page being written for not being
empty and do not actually write it when it is. This fails for a
freshly created flash bad block table when all blocks are good.
In this case the bbt code will try to write an empty page, but
with the BBT marker in OOB. This page never gets written, so the
BBT code will not find a bad block table during next start up and
writes it again. Fix this by checking if we want to write OOB data,
if we do, then write the page, even if the data is empty.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Do not write pages which only contain 0xff. UBI expects pages which
seem empty to be writable. This got lost with:
| commit 3139c3e9a6
| Author: Sascha Hauer <s.hauer@pengutronix.de>
| Date: Thu Nov 29 11:16:40 2012 +0100
|
| mtd core: call driver write function with complete buffer
|
| mtd->write is supposed to loop around pages internally, no need
| to do this in mtd_write. This fixes a huge write performance drop
| with the m25p80 driver when it was converted to a mtd driver recently.
| Since mtd->writesize is 1 for this driver mtd_write ended up doing
| single byte writes on the flash.
Introduce mtd_all_ff as a global function since UBI currently has its own
implementation.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
CFI Flash is currently handled outside the mtd layer which makes it
a special case. Integrate it into mtd so that we get rid of this
special status.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Instead of dereferencing struct mtd_info members directly use
the wrapper functions which have an additional check if the
callback exists if it is optional, like mark_bad or is_bad.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The Kernel has mtd_read, mtd_write, mtd_erase and mtd_block_markbad.
Add these functions to barebox aswell to make future mtd synchronizations
with the kernel easier.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
These are currently named mtd_read/write/erase. Functions with
the same name exist as global functions in the kernel. Rename
them to mtd_op_* to avoid name clashes with future mtd updates.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Previously, during a multi-page write, chip->oob_poi would not be
reinitialized.
See commit f722013ee9fd24623df31dec9a91a6d02c3e2f2f in the kernel
for an analogous change.
Signed-off-by: Jan Luebbe <jlu@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Got this compile warning if MTD_WRITE support isn't set.
drivers/mtd/core.c: In function 'mtd_ioctl':
drivers/mtd/core.c:117:26: warning: unused variable 'ei' [-Wunused-variable]
Add a #ifdef to declaration of variable 'ei' to solve it.
Signed-off-by: Alexander Aring <alex.aring@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
- remove mach/silicon.h and include omap?-silicon.h directly
- include mach/omap?-clock.h directly where needed
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
To make that, we need to shift mtd_erase before mtd_ioctl.
ubi-utils need that, especially ubiformat.
Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
We have two versions in the tree. Use the newer one, and put it into the
mtd directory while we are at it.
Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
It is not the chip, but this controller which cannot do subpage writes.
So, make sure we add the flag at the proper place, so it doesn't get
overwritten by flash detection anymore.
Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
dev_add_child is a very unsafe function. If called multiple times
it allows setting the same device to different parents thus corrupting
the siblings list. This happens regularly since:
| commit c2e568d19c
| Author: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
| Date: Sat Nov 3 16:11:05 2012 +0100
|
| bus: add bus device
|
| automatically add it as parent of any bus device if none already specified
|
| we have now a nice output per bus
If for example a FATfs is mounted this nice output per bus often ends with:
> `---- fat0
> `---- 0
> `---- 0x86f0000087020031-0x86f000410df27124: /dev/<NULL>
> `---- sram00
> `---- 0x00000000-0xffffffffffffffff: /dev/<NULL>
> `---- 0x00000000-0xffffffffffffffff: /dev/<NULL>
> unable to handle NULL pointer dereference at address 0x0000000c
> pc : [<87f08a20>] lr : [<87f08a04>]
> sp : 86eff8c0 ip : 87f3fbde fp : ffffffff
> r10: ffffffff r9 : 00000000 r8 : 00000003
> r7 : 86f075b8 r6 : 00000002 r5 : ffffffec r4 : 86f07544
> r3 : 00000000 r2 : 43f900b4 r1 : 00000020 r0 : 00000005
> Flags: Nzcv IRQs off FIQs off Mode SVC_32
> [<87f08a20>] (do_devinfo_subtree+0x90/0x130) from [<87f08a90>] (do_devinfo_subtree+0x100/0x130)
>
> [<87f3e070>] (unwind_backtrace+0x0/0x90) from [<87f28514>] (panic+0x28/0x3c)
> [<87f28514>] (panic+0x28/0x3c) from [<87f3e4b8>] (do_exception+0x10/0x14)
> [<87f3e4b8>] (do_exception+0x10/0x14) from [<87f3e544>] (do_data_abort+0x2c/0x38)
> [<87f3e544>] (do_data_abort+0x2c/0x38) from [<87f3e268>] (data_abort+0x48/0x60)
This patch fixes this by adding a device to its parents children list in
register_device so that dev_add_child is no longer needed. This function
is removed from the tree. Now callers of register_device have to clearly
set the parent *before* registering a device.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Reported-by: Jan Lübbe <jlu@pengutronix.de>
Barebox not contain symbol MTD_NAND, so remove all references to it.
Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This patch provides a global cleanup barebox Kconfig files. This includes
replacing spaces to tabs, formatting in accordance format, removing
extraneous lines and spaces. No functional changes.
Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
mtd->write is supposed to loop around pages internally, no need
to do this in mtd_write. This fixes a huge write performance drop
with the m25p80 driver when it was converted to a mtd driver recently.
Since mtd->writesize is 1 for this driver mtd_write ended up doing
single byte writes on the flash.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The nand layer handles partial page writes just fine. If the start or
end of the data is not page aligned, the nand layer will copy the data
to a temporary page buffer.
Remove the check which disallows partial page writes since this is what
we want to do on barebox when for example an image is written to nand
which is not padded to page size.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
It could happen (1 out of 100 times) that NAND did not start up
correctly after warm rebooting, so barebox could not find its
environment or DMA timed out due to a stalled BCH. When resetting BCH
together with GPMI, the issue could not be observed anymore. We probably
need the consistent state already before sending commands to NAND.
Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
As in the kernel we revert as this was supposed to work but does not yet
this may need more work on the smc to be able to use it
So for now revert it
This reverts commit 809f0f6327.
Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
the timeout is 100ms not 100ns
on sam9x5 the SoC is fast enough so we can not see it
but not on sam9n12
Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
so we now create the cdev via mtd
This will also simplify sync with linux
to avoid the m25p8000 or m25p00 the cdev is still named name m25p and the
drivers m25p80
Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Fixed possible crashs if nand-bb device is removed
and the list entry was not deleted from the list.
Signed-off-by: Juergen Kilb <J.Kilb@phytec.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
this allow to do not provide block_isbad at mtd driver level
as example spi flash
Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
on some system as the spi_transfer is not set to 0 if the rx_buf or tx_buf
are not used they are not set to NULL
Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Got this when compiling sandbox on a 64-bit system:
drivers/mtd/ubi/cdev.c: In function ‘ubi_volume_cdev_read’:
drivers/mtd/ubi/cdev.c:26:2: warning: format ‘%d’ expects argument of type ‘int’, but argument 3 has type ‘size_t’ [-Wformat]
Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This patch is from linux 3.7-rc1 and adapt to Barebox
Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
as we do not need to probe them and they have no driver or bus attached
Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
On i.MX we enable all necessary clocks during startup of the clock
controller driver, so we do not need the register hacking in the drivers
anymore.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
toread is unitialized. We have to use count instead.
| commit 992c291e95
| Author: Sascha Hauer <s.hauer@pengutronix.de>
| Date: Sat Sep 15 16:54:47 2012 +0200
|
| mtd mtdraw: Fix partial page read
|
| When reading parts of a page we have to limit the maximum bytes copied
| to the remaining bytes of a page.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The FSF address has changed in the past. Instead of updating it
each time the address changes, just drop it completely treewide.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
When reading parts of a page we have to limit the maximum bytes copied
to the remaining bytes of a page.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Acked-by: Robert Jarzmik <robert.jarzmik@free.fr>
Some of the newer MLC devices have a 6-byte ID sequence in which
several field definitions differ from older chips in a manner that is
not backward compatible.
This method is already used in the Linux Kernel.
Signed-off-by: Marcus Folkesson <marcus.folkesson@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
For the size_t type the format specifier %zu or %zx must be used.
See Documentation/printk-formats.txt in the kernel for details.
Signed-off-by: Jan Luebbe <jlu@pengutronix.de>
Acked-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The mtdoob and mtdraw device don't clean up correctly.
Added a private data element to hold allocated memory.
Fix remove of mtdoob and mtdraw device.
Signed-off-by: Alexander Aring <alex.aring@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
With pbl support enabled most boards need a pbl-y for their lowlevel
stuff.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Acked-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
The existing nand_boot functions all do the same, so move it to
a common place. To be flexible enough for future boards the real
image size is used instead of hardcoded 256k.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Acked-by: Juergen Beisert <jbe@pengutronix.de>
Acked-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
This introduces a new NAND_BUSWIDTH_AUTO flag which can be used
to automatically detect the nand buswidth. The id is always read
in 8bit mode. An additional callback is needed to switch the nand
controller into 16bit mode.
This currently depends on a safe read_byte (always) and read_buf
(for onfi-only flashes) callback. It has been tested on OMAP, but
is not something that generally works. For this reason the existence
of the set_buswidth callback is used to determine whether we are
able to do autodetection or not.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Use the prefetch engine to improve NAND performance. The howto
is derived from the Kernel. Unlike the kernel we do not make
the access mode configurable.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
- dev_ready is supposed to return whether the device is ready or
not, not to poll until the device is ready.
- dev_ready should return true for ready and false for not ready
- waitpin polarity is not needed (at least the kernel does not have it)
- wait_mon_mask must be 32bit.
The code was unused since no board specified a wait pin, so no breakage
included. This also removes the now unused timeout variable from
platformdata.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The 'off_t cur_ofs' variable was missed during the 64 bit conversion.
For the MEMGETBADBLOCK ioctl, a pointer to a loff_t is needed.
Also adjust the debug format strings.
Signed-off-by: Jan Luebbe <jlu@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Since this commit we interpret the argument to the bad block ioctls
as a pointer to a 64bit number:
|commit e71c343668
|Author: Sascha Hauer <s.hauer@pengutronix.de>
|Date: Fri Oct 14 11:57:55 2011 +0200
|
| mtd: fix arguments to bad block ioctls
|
| In the Kernel the mtd ioctls expect a pointer to the offset, whereas
| barebox interprets the pointer itself as an offset. Since we want
| to add 64bit support for file sizes a pointer may not be sufficient,
| so align with the kernel and convert it to a pointer to the offset.
|
| Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This missed some places, fix them aswell.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
- add CMD_PARAM and read_param to get the ONFI structure
- fix OOB size for flash with 224 OOB on i.MX51/3
- add the same ecc layout as the one in the kernel for
4k page flashs
Tested on an i.MX53.
Signed-off-by: Eric Bénard <eric@eukrea.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
the code is taken from linux & u-boot implementations
Validated on an i.MX53 which gives the following log :
ONFI flash detected ... ONFI param page 0 valid
NAND device: Manufacturer ID: 0x2c, Chip ID: 0x38 (Micron MT29F8G08ABABAWP), page size: 4096, OOB size: 224
Signed-off-by: Eric Bénard <eric@eukrea.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>