Robert Jarzmik [Thu, 22 Mar 2012 20:00:50 +0000 (21:00 +0100)]
mtd: docg3 fix inbound calculations
The last erase block was not accessible, as the out of bound
check was incorrectly rejecting the last block.
The read/write/erase offset checks were forbidding the usage of the
last block, because of the calculation which was considering the
byte after the last instead of the last byte.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Wolfram Sang [Wed, 21 Mar 2012 18:29:17 +0000 (19:29 +0100)]
mtd: nand: gpmi: fix function annotations
A lot of functions have been marked __devinit, but they shouldn't, because they
are needed for bbt_scan. While I believe the whole MX23 handling should be done
entirely different, I am missing the resources to fix it. So, let's have at least
the annotations correct.
Signed-off-by: Wolfram Sang <w.sang@pengutronix.de> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Shmulik Ladkani [Thu, 9 Feb 2012 13:36:29 +0000 (15:36 +0200)]
mtd: unify initialization of erase_info->fail_addr
Initialization of 'erase_info->fail_addr' to MTD_FAIL_ADDR_UNKNOWN prior
erase operation is duplicated accross several MTD drivers, and also taken
care of by some MTD users as well.
Harmonize it: initialize 'fail_addr' within 'mtd_erase()' interface.
Matthieu CASTET [Mon, 19 Mar 2012 14:35:25 +0000 (15:35 +0100)]
mtd: support ONFI multi lun NAND
With onfi a flash is organized into one or more logical units (LUNs).
A logical unit (LUN) is the minimum unit that can independently execute
commands and report status.
Mtd does not exploit LUN, so make it see a big single flash where size is
lun_size * number_of_lun.
Without this patch MT29F8G08ADBDAH4 size is 512MiB instead of 1GiB.
Vipin Kumar [Wed, 14 Mar 2012 06:17:18 +0000 (11:47 +0530)]
mtd: nand/fsmc: Add DMA support
The fsmc_nand driver uses cpu to read/write onto the device. This is inefficient
because of two reasons
- the cpu gets locked on AHB bus while reading from NAND
- the cpu is unnecessarily used when dma can do the job
This patch adds the support for accessing the device through DMA
Vipin Kumar [Wed, 14 Mar 2012 06:17:17 +0000 (11:47 +0530)]
mtd: nand/fsmc: Access the NAND device word by word whenever possible
The default way of accessing nand device is using the nand width. This means
that 8bit devices are using u8 * and 16bit devices are accessed using u16 *.
This results in a non-optimal performance since the FSMC is designed to
translate the normal word accesses into device width based accesses. This patch
implements read_buf and write_buf callbacks using word by word accesses.
Vipin Kumar [Wed, 14 Mar 2012 06:17:16 +0000 (11:47 +0530)]
mtd: nand/fsmc: Use dev_err to report error scenario
fsmc controller takes time to calculate the bch8 codes and the error offsets.
The calculate logic checks for completion upto a timeout. This patch adds a
error print when this timer expires and the ecc or error offsets are not yet
calculated.
Vipin Kumar [Wed, 14 Mar 2012 06:17:15 +0000 (11:47 +0530)]
mtd: nand/fsmc: Use devm routines
fsmc_nand driver currently uses normal kzalloc, request_mem etc routines. This
patch replaces these routines with devm_kzalloc and devm_request_mem_region etc.
Consequently, the error and driver removal scenarios are curtailed.
Vipin Kumar [Wed, 14 Mar 2012 06:17:14 +0000 (11:47 +0530)]
mtd: nand/fsmc: Modify fsmc driver to accept nand timing parameters via platform
FSMC controllers provide registers to program the required timing values for
attached NAND device. The timing values used until now are relaxed and should
work for all devices.
Although, for read/write performance improvements, the fsmc nand driver should
accept nand timings as a platform data and program the timing parameters into
fsmc registers accordingly.
This patch implements this modification. Additionally, it programs the default
timing parameters if these are not passed via platform data.
Vipin Kumar [Wed, 14 Mar 2012 06:17:12 +0000 (11:47 +0530)]
mtd: nand/fsmc: Initialize the badblockbits to 7
Ideally, the block should have 0xff written on the bad block position. Any value
other than 0xff implies a bad block. In practical situations, there can be
bit flips in the oob area as well which means that a block with 0x7f being read
at bad block position may imply a bad block but it is infact only a bit flip in
the bad block byte.
To resolve this problem, the block is marked as good if number of high bits is
greater than or equal to badblockbits (initialized to 7)
Vipin Kumar [Wed, 14 Mar 2012 06:17:11 +0000 (11:47 +0530)]
mtd: nand/fsmc: Flip the bit only if the error index is < 4096
ECC can correct up to 8 bits in 512 bytes data + 13 bytes ecc. This means that
the algorithm can correct a max of 8 bits in 4200 bits ie the error indices can
be from 0 to 4199. Of these 0 to 4095 are for data and 4096 to 4199 for ecc.
The driver flips the bit only if the index is <= 4096. This is a bug since the
data bits are only from 0 to 4095.
Vipin Kumar [Wed, 14 Mar 2012 06:17:10 +0000 (11:47 +0530)]
mtd: nand/fsmc: Read only 512 + 13 bytes for 8bit NAND devices
The ECC logic of FSMC works on 512 bytes data + 13 bytes ECC to generate error
indices of up to 8 incorrect bits. The FSMC driver reads 14 instead of 13 oob
bytes to accommodate for 16 bit device as well.
Unfortunately, the internal ecc state machine gets corrupted for 8 bit devices
reading 512 + 14 bytes of data resulting in error indices not getting reported.
Fix this by reading 14 bytes only for 16 bit devices
Vipin Kumar [Wed, 14 Mar 2012 06:17:07 +0000 (11:47 +0530)]
mtd: nand/fsmc: Pass partition information through platform data
This patch reimplements the passing of partition information through platform
data. This was unintentionally deleted in commit 0d04eda1430e9a796214bee644b7e05d99cfe613
"mtd: fsmc_nand.c: use mtd_device_parse_register"
Artem: fix gcc warning about passin 0 instead of NULL.
Hervé Fache [Tue, 13 Mar 2012 15:07:53 +0000 (16:07 +0100)]
mtd: phram: dot not crash when built-in and passing boot param
This patch is based on Ville Herva's similar patch to block2mtd.
Trying to pass a parameter through the kernel command line when built-in would
crash the kernel, as phram_setup() was called so early that kmalloc() was not
functional yet.
This patch only saves the parameter string at the early boot stage, and parses
it later when init_phram() is called. The same happens in both module and
built-in cases.
With this patch, I can boot with a statically-compiled phram, and mount a
ext2 root fs from physical RAM, without the need for a initrd.
This has been tested in built-in and module cases, with and without a
parameter string.
Huang Shijie [Mon, 12 Mar 2012 02:22:18 +0000 (10:22 +0800)]
mtd: gpmi: fix the wrong DMA command.
The last DMA command of ECC read page is used to disable the BCH module.
But the original code missed to set the pio[2] which is used to set the
GPMI_HW_GPMI_ECCCTRL register. fix it now.
Mike Dunn [Sun, 11 Mar 2012 21:21:11 +0000 (14:21 -0700)]
mtd: flash drivers set ecc strength
Flash device drivers initialize 'ecc_strength' in struct mtd_info, which is the
maximum number of bit errors that can be corrected in one writesize region.
Drivers using the nand interface intitialize 'strength' in struct nand_ecc_ctrl,
which is the maximum number of bit errors that can be corrected in one ecc step.
Nand infrastructure code translates this to 'ecc_strength'.
Also for nand drivers, the nand infrastructure code sets ecc.strength for ecc
modes NAND_ECC_SOFT, NAND_ECC_SOFT_BCH, and NAND_ECC_NONE. It is set in the
driver for all other modes.
Signed-off-by: Mike Dunn <mikedunn@newsguy.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Mike Dunn [Sun, 11 Mar 2012 21:21:10 +0000 (14:21 -0700)]
mtd: add ecc_strength fields to mtd structs
This adds 'ecc_strength' to struct mtd_info. This stores the maximum number of
bit errors that can be corrected in one writesize region.
For consistency with the nand code, 'strength' is similiarly added to struct
nand_ecc_ctrl. This stores the maximum number of bit errors that can be
corrected in one ecc step.
Signed-off-by: Mike Dunn <mikedunn@newsguy.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Robert Jarzmik [Sun, 11 Mar 2012 10:28:45 +0000 (11:28 +0100)]
mtd: docg3 reduce read alignment burden
The read function was so far requiring the reads to be aligned on page
boundaries, and be page length multiples in size. Relieve these
constraints to ease the userspace ubifs programs runs, which read ubifs
headers of 64 bytes.
Artem: squashed a later fix from Robert Jarzmik into this patch.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Bastian Hecht [Sun, 18 Mar 2012 14:13:20 +0000 (15:13 +0100)]
mtd: sh_flctl: Add power management with QoS request
Adds power management code with fine granularity. Every flash control
command is enclosed by runtime_put()/get()s. To make sure that no
overhead is generated by too frequent power state switches, a quality of
service request is issued.
The first 3 arguments of 'mtd_device_parse_register()' are pointers,
but many callers pass '0' instead of 'NULL'. Fix this globally. Thanks
to coccinelle for making it easy to do with the following semantic patch:
Paul Parsons [Wed, 7 Mar 2012 14:14:31 +0000 (14:14 +0000)]
mtd: maps: sa1100-flash: Add reference counter to set_vpp()
This patch is part of a set which fixes unnecessary flash erase and write errors
resulting from the MTD CFI driver turning off vpp while an erase is in progress.
This patch allows sa1100_set_vpp() calls to be nested by adding a reference
counter.
Signed-off-by: Paul Parsons <lost.distance@yahoo.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Paul Parsons [Wed, 7 Mar 2012 14:13:41 +0000 (14:13 +0000)]
mtd: maps: pcmciamtd: Add reference counter to set_vpp()
This patch is part of a set which fixes unnecessary flash erase and write errors
resulting from the MTD CFI driver turning off vpp while an erase is in progress.
This patch allows pcmciamtd_set_vpp() calls to be nested by adding a reference
counter.
Signed-off-by: Paul Parsons <lost.distance@yahoo.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Paul Parsons [Wed, 7 Mar 2012 14:12:50 +0000 (14:12 +0000)]
mtd: maps: l440gx: Add reference counter to set_vpp()
This patch is part of a set which fixes unnecessary flash erase and write errors
resulting from the MTD CFI driver turning off vpp while an erase is in progress.
This patch allows l440gx_set_vpp() calls to be nested by adding a reference
counter.
Signed-off-by: Paul Parsons <lost.distance@yahoo.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Paul Parsons [Wed, 7 Mar 2012 14:12:08 +0000 (14:12 +0000)]
mtd: maps: physmap: Add reference counter to set_vpp()
This patch is part of a set which fixes unnecessary flash erase and write errors
resulting from the MTD CFI driver turning off vpp while an erase is in progress.
This patch allows physmap_set_vpp() calls to be nested by adding a reference
counter.
omap1_set_vpp() already used a reference counter. Since it is called from
physmap_set_vpp(), omap1_set_vpp() can now be simplified.
simtec_nor_vpp() already disabled hard interrupts. Since it is called from
physmap_set_vpp(), simtec_nor_vpp() can now be simplified.
Signed-off-by: Paul Parsons <lost.distance@yahoo.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Paul Parsons [Wed, 7 Mar 2012 14:11:16 +0000 (14:11 +0000)]
mtd: chips: cfi_cmdset_0002: Match ENABLE_VPP()/DISABLE_VPP() calls
This patch is part of a set which fixes unnecessary flash erase and write errors
resulting from the MTD CFI driver turning off vpp while an erase is in progress.
This patch ensures that only those flash operations which call ENABLE_VPP() can
then call DISABLE_VPP(). Other operations should never call DISABLE_VPP().
Signed-off-by: Paul Parsons <lost.distance@yahoo.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Paul Parsons [Wed, 7 Mar 2012 14:10:33 +0000 (14:10 +0000)]
mtd: chips: cfi_cmdset_0001: Match ENABLE_VPP()/DISABLE_VPP() calls
This patch is part of a set which fixes unnecessary flash erase and write errors
resulting from the MTD CFI driver turning off vpp while an erase is in progress.
This patch ensures that only those flash operations which call ENABLE_VPP() can
then call DISABLE_VPP(). Other operations should never call DISABLE_VPP().
Signed-off-by: Paul Parsons <lost.distance@yahoo.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Since change_bit() requires a (unsigned int *) as second arg,
the correct definition of err_idx[] array declared as
local variable of fsmc_correct_data() is the following:
Shiraz Hashim [Wed, 7 Mar 2012 11:30:52 +0000 (17:00 +0530)]
mtd: fsmc: Move ALE, CLE defines to their respective platform
Address Latch Enable (ALE) and Command Latch Enable (CLE) defines are
platform specific and were wrongly put in driver specific fsmc.h file.
Move such defines to their respective platform.
Also instead of relying on fsmc driver, pass ALE, CLE offsets explicitly
from individual platform.
Bhavna Yadav [Wed, 7 Mar 2012 11:30:50 +0000 (17:00 +0530)]
mtd: fsmc_nand: ECC1 & ECC4 layout separated for different page sizes
ECC1 & ECC4 layout for NAND of different pages sizes for e.g. 512bytes,
2KiB, 4KiB and 8KiB are separated. Previously there existed one ECC4
layout for 2KiB & 4KiB page size due to which oob test module available
in drivers/mtd/nand/test was failing.
A newly erased page contains ff in data as well as spare area. While reading an
erased page, the read out ecc from spare area does not match the ecc generated
by fsmc ecc hardware accelerator. This is because ecc of data ff ff is not ff
ff. This leads to errors when file system erases and reads back the pages to
ensure consistency.
This patch adds a software workaround to ensure that the ecc check is not
performed for erased pages. This problem is solved by checking the number of
bits (in 512 byte data + 13 byte ecc) which are 0. If these number of bits are
less than 8, the page is considered erased and correction algorithm is not tried
on that page
Mike Dunn [Sat, 3 Mar 2012 21:13:06 +0000 (13:13 -0800)]
mtd: fix partition wrapper functions
This patch reverts a change that may have been mistakenly included with the set
of patches that introduced the new mtd api entry functions. Or perhaps I am
mistaken :)
The problem is in the partition wrapper functions, where the calls to the driver
methods were replaced with calls to the new mtd api functions. This causes the
api function to be called a second time, further down the call stack. This is
not only unnecessary and redundant - because the sanity checking code and (more
restrictive) bounds checks for the partition were done in the first call - but
is potentially problematic and confusing.
For example, the call stack for a call to mtd_read() on a partitioned device
currently looks like this:
mtd_read() gets struct mtd_info for the partition
|
+-> part_read() via the pointer assigned when the partition was created
|
+->mtd_read() this time gets struct mtd_info for the master
|
+->xyz_driver_read() via the pointer asigned by the driver
It seems that this can cause a variety of problems. For example, if you want to
add code to the api function that tests a value in mtd_info that is relevant
only to the partition. Or (in my case) you want the driver to return a value
that may be different from that returned by the mtd api function.
This patch eliminates the second call to the mtd api function. It was tested on
the docg4 nand driver with a subset of the api functions, but I inspected the
rest and don't see any problems.
Signed-off-by: Mike Dunn <mikedunn@newsguy.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Bastian Hecht [Thu, 1 Mar 2012 09:48:35 +0000 (10:48 +0100)]
mtd: sh_flctl: Expand FLCMNCR register bit field
Add support for a new hardware generation. The meaning of some bits
of the FLCMNCR register changed, so some new defines are added
parallel to the existing ones to keep backward compatibility.
The defines allow to choose an appropriate clocking scheme.
Huang Shijie [Fri, 17 Feb 2012 03:22:37 +0000 (11:22 +0800)]
mtd: change the location of the ONFI detected log
Some strange nand chip(such as Hynix H27UBG8T2A) can pass the `ONFI` signature
check. So the log can be printed out even it is not an ONFI nand indeed.
Change this log to the end of the function. Print out the log only when we
really detect an ONFI nand.
Signed-off-by: Huang Shijie <b32955@freescale.com> Acked-by: Florian Fainelli <ffainelli@freebox.fr> Acked-by: Brian Norris <computersforpeace@gmail.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Huang Shijie [Thu, 16 Feb 2012 06:17:33 +0000 (14:17 +0800)]
mxs-dma : rewrite the last parameter of mxs_dma_prep_slave_sg()
[1] Background :
The GPMI does ECC read page operation with a DMA chain consist of three DMA
Command Structures. The middle one of the chain is used to enable the BCH,
and read out the NAND page.
The WAIT4END(wait for command end) is a comunication signal between
the GPMI and MXS-DMA.
[2] The current DMA code sets the WAIT4END bit at the last one, such as:
[3] But in the new GPMI version (used in MX50/MX60), the WAIT4END bit should
be set not only at the last DMA Command Structure,
but also at the middle one, such as:
+-----+ +-----+ +-----+
| cmd | ------------> | cmd | ------------------> | cmd |
+-----+ +-----+ +-----+
^ ^
| |
| |
set WAIT4END here too set WAIT4END here
If we do not set WAIT4END, the BCH maybe stalls in "ECC reading page" state.
In the next ECC write page operation, a DMA-timeout occurs.
This has been catched in the MX6Q board.
[4] In order to fix the bug, rewrite the last parameter of mxs_dma_prep_slave_sg(),
and use the dma_ctrl_flags:
---------------------------------------------------------
DMA_PREP_INTERRUPT : append a new DMA Command Structrue.
DMA_CTRL_ACK : set the WAIT4END bit for this DMA Command Structure.
---------------------------------------------------------
[5] changes to the relative drivers:
<1> For mxs-mmc driver, just use the new flags, do not change any logic.
<2> For gpmi-nand driver, and use the new flags to set the DMA
chain, especially for ecc read page.
Huang Shijie [Thu, 16 Feb 2012 06:17:32 +0000 (14:17 +0800)]
mxs-dma : move the mxs dma.h to a more common place
Move the header to a more common place.
The mxs dma engine is not only used in mx23/mx28, but also used
in mx50/mx6q. It will also be used in the future chips.
Rename it to mxs-dma.h, and create a new folder include/linux/fsl/ to
store the Freescale's header files.
change mxs-dma driver, mxs-mmc driver, gpmi-nand driver, mxs-saif driver
to the new header file.
Acked-by: Shawn Guo <shawn.guo@linaro.org> Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com> Signed-off-by: Huang Shijie <b32955@freescale.com> Acked-by: Vinod Koul <vinod.koul@linux.intel.com> Acked-by: Chris Ball <cjb@laptop.org> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Russell King [Wed, 15 Feb 2012 11:48:34 +0000 (11:48 +0000)]
mtd: fix 'Flash device refused suspend due to active operation' message
While debugging on SA11x0, the following message was observed:
"Flash device refused suspend due to active operation (state 20)"
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
In commit "c797533 mtd: abstract last MTD partition parser argument" the
third argument of "mtd_device_parse_register()" changed from start address
of the MTD device to a pointer to a struct.
The "ixp4xx_flash_probe()" function was not converted properly, causing
an oops during boot.
This patch fixes the problem by filling the needed information into a
"struct mtd_part_parser_data" and passing it to
"mtd_device_parse_register()".
Signed-off-by: Marc Kleine-Budde <mkl@blackshift.org> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: stable@kernel.org [3.2+] Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
This patch changes all the OTP functions like 'mtd_get_fact_prot_info()' and
makes them return zero immediately if the input 'len' parameter is 0. This is
not really needed currently, but most of the other functions do this, and it is
just consistent to do the same in the OTP functions.
This patch also moves the OTP functions from the header file to mtdcore.c
because they become a bit too big for being inlined.
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
mtd: move zero length verification to MTD API functions
In many places in drivers we verify for the zero length, but this is very
inconsistent across drivers. This is obviously the right thing to do, though.
This patch moves the check to the MTD API functions instead and removes a lot
of duplication.
Some MTD drivers return -EINVAL if the 'phys' parameter is not NULL, trying to
convey that they cannot return the physical address. However, this is not very
logical because they still can return the virtual address ('virt'). But some
drivers (lpddr) just ignore the 'phys' parameter instead, which is a more
logical thing to do.
Let's harmonize this and:
1. Always initialize 'virt' and 'phys' to 'NULL' in 'mtd_point()'.
2. Do not return an error if the physical address cannot be found.
So as a result, all drivers will set 'phys' to 'NULL' if it is not supported.
None of the 'mtd_point()' users use 'phys' anyway, so this should not break
anything. I guess we could also just delete this parameter later.
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The MTD API function now zero the 'retlen' parameter before calling
the driver's method — do not do this again in drivers. This removes
duplicated '*retlen = 0' assignent from the following methods:
Many drivers check whether the partition is R/O and return -EROFS if yes.
Let's stop having duplicated checks and move them to the API functions
instead.
And again a bit of noise - deleted few too sparse newlines, sorry.
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
mtd: do not duplicate length and offset checks in drivers
We already verify that offset and length are within the MTD device size
in the MTD API functions. Let's remove the duplicated checks in drivers.
This patch only affects the following API's:
The 'mtd_unpoint()' API function should be able to return an error code because
it may fail if you specify incorrect offset. This patch changes this MTD API
function and amends all the drivers correspondingly.
Also return '-EOPNOTSUPP' from 'mtd_unpoint()' when the '->unpoint()' method is
undefined. We do not really need this currently, but this just makes
sense to be consistent with 'mtd_point()'.
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Brian Norris [Mon, 6 Feb 2012 21:44:00 +0000 (13:44 -0800)]
mtd: nand: write BBM to OOB even with flash-based BBT
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:
* bootloader 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. If your system cannot support writing bad block markers to OOB,
use the new NAND_BBT_NO_OOB_BBM option (in combination with
NAND_BBT_USE_FLASH and NAND_BBT_NO_OOB).
Signed-off-by: Brian Norris <computersforpeace@gmail.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
We do not need to invoke 'mtd_can_have_bb()' before invoking
'mtd_block_isbad()' because the latter already handles the case when the MTD
device does not support bad blocks.
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
mtd: mtdconcat: return -EOPNOTSUPP if block_markbad is undefined
The main 'mtd_block_markbad()' function returns -EOPNOTSUPP if the
'->block_markbad' method is undefined, and mtdconcat should do the same.
Fix this by simply removing the 'mtd_can_have_bb()' because it is not
really necessary. It could be treated as an optimization, but this function is
expected to be used so rarely that it does not matter.
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The writebufsize concept was introduce by commit
"0e4ca7e mtd: add writebufsize field to mtd_info struct" and it represents
the maximum amount of data the device writes to the media at a time. This is
an important parameter for UBIFS which is used during recovery and which
basically defines how big a corruption caused by a power cut can be.
Set writebufsize to the flash page size because it is the maximum amount of
data it writes at a time.
The writebufsize concept was introduce by commit
"0e4ca7e mtd: add writebufsize field to mtd_info struct" and it represents
the maximum amount of data the device writes to the media at a time. This is
an important parameter for UBIFS which is used during recovery and which
basically defines how big a corruption caused by a power cut can be.
Set writebufsize to the flash page size because it is the maximum amount of
data it writes at a time.
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Acked-by: Stefan Roese <sr@denx.de> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The writebufsize concept was introduce by commit
"0e4ca7e mtd: add writebufsize field to mtd_info struct" and it represents
the maximum amount of data the device writes to the media at a time. This is
an important parameter for UBIFS which is used during recovery and which
basically defines how big a corruption caused by a power cut can be.
Set writebufsize to 4 because this drivers writes at max 4 bytes at a time.
The writebufsize concept was introduce by commit
"0e4ca7e mtd: add writebufsize field to mtd_info struct" and it represents
the maximum amount of data the device writes to the media at a time. This is
an important parameter for UBIFS which is used during recovery and which
basically defines how big a corruption caused by a power cut can be.
Set it to be equivalent to mtd->writesize because this is the maximum amount
of data the driver writes at a time.
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Acked-by: Robert Jarzmik <robert.jarzmik@free.fr> Cc: stable@kernel.org [3.2+] Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The writebufsize concept was introduce by commit
"0e4ca7e mtd: add writebufsize field to mtd_info struct" and it represents
the maximum amount of data the device writes to the media at a time. This is
an important parameter for UBIFS which is used during recovery and which
basically defines how big a corruption caused by a power cut can be.
Set it to be equivalent to mtd->writesize because this is the maximum amount
of data the driver writes at a time.
The writebufsize concept was introduce by commit
"0e4ca7e mtd: add writebufsize field to mtd_info struct" and it represents
the maximum amount of data the device writes to the media at a time. This is
an important parameter for UBIFS which is used during recovery and which
basically defines how big a corruption caused by a power cut can be.
Set it to be equivalent to mtd->writesize because this is the maximum amount
of data the driver writes at a time.
The writebufsize concept was introduce by commit
"0e4ca7e mtd: add writebufsize field to mtd_info struct" and it represents
the maximum amount of data the device writes to the media at a time. This is
an important parameter for UBIFS which is used during recovery and which
basically defines how big a corruption caused by a power cut can be.
Set it to be equivalent to mtd->writesize because this is the maximum amount
of data the driver writes at a time.
The writebufsize concept was introduce by commit
"0e4ca7e mtd: add writebufsize field to mtd_info struct" and it represents
the maximum amount of data the device writes to the media at a time. This is
an important parameter for UBIFS which is used during recovery and which
basically defines how big a corruption caused by a power cut can be.
However, we forgot to set this parameter for block2mtd. Set it to PAGE_SIZE
because this is actually the amount of data we write at a time.
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Acked-by: Joern Engel <joern@lazybastard.org> Cc: stable@kernel.org [2.6.38+] Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Dan Carpenter [Tue, 31 Jan 2012 08:54:06 +0000 (11:54 +0300)]
mtd: docg4: fix printk() warnings
Gcc complains here:
drivers/mtd/nand/docg4.c: In function ‘probe_docg4’:
drivers/mtd/nand/docg4.c:1277:4: warning: format ‘%x’ expects argument of type ‘unsigned int’, but argument 3 has type ‘resource_size_t’ [-Wformat]
drivers/mtd/nand/docg4.c:1277:4: warning: format ‘%x’ expects argument of type ‘unsigned int’, but argument 4 has type ‘resource_size_t’ [-Wformat]
We have a standard way of printing these using a format string
extension.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Acked-by: Randy Dunlap <rdunlap@xenotime.net> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Wolfram Sang [Tue, 31 Jan 2012 12:10:43 +0000 (13:10 +0100)]
mtd: nand: gpmi: use correct member for checking NAND_BBT_USE_FLASH
This has been moved from .options to .bbt_options meanwhile. So, it
currently checks for something totally different (NAND_OWN_BUFFERS) and
decides according to that.
Artem Bityutskiy: the options were moved in a40f734 mtd: nand: consolidate redundant flash-based BBT flags
Artem Bityutskiy: CCing -stable
Signed-off-by: Wolfram Sang <w.sang@pengutronix.de> Acked-by: Huang Shijie <b32955@freescale.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Cc: stable@kernel.org [3.2+] Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The reason is that we are re-working the MTD API and from now on it is
an error to use MTD function pointers directly - we have a corresponding
API call for every pointer. By adding a leading "_" we achieve the following:
1. Make sure we convert every direct pointer users
2. A leading "_" suggests that this interface is internal and it becomes
less likely that people will use them directly
3. Make sure all the out-of-tree modules stop compiling and the owners
spot the big API change and amend them.
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Brian Norris [Fri, 3 Feb 2012 00:21:09 +0000 (16:21 -0800)]
jffs2: update to new MTD interface
There were a few instances of the old MTD interface remaining for JFFS2. We
fix one error that shows up (only when CONFIG_JFFS2_FS_WRITEBUFFER is not
defined) like this:
fs/jffs2/read.c: In function 'jffs2_read_dnode':
fs/jffs2/read.c:36:8: error: 'struct mtd_info' has no member named 'read'
fs/jffs2/read.c:112:8: error: 'struct mtd_info' has no member named 'read'
...
We also simply remove two macros that are not in use, were not updated to
the new MTD interface, and don't even utilize the old interface properly.
(That means they weren't used since commit 8593fbc6, year 2006; almost 6
years ago, for those who don't want to do the math)
Reported-by: Randy Dunlap <rdunlap@xenotime.net> Signed-off-by: Brian Norris <computersforpeace@gmail.com> Acked-by: Randy Dunlap <rdunlap@xenotime.net> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Mike Dunn [Wed, 4 Jan 2012 00:05:44 +0000 (16:05 -0800)]
mtd: nand: add support for diskonchip G4 nand flash device
This patch adds a driver for the M-Sys / Sandisk diskonchip G4 nand flash found
in various smartphones and PDAs, among them the Palm Treo680, HTC Prophet and
Wizard, Toshiba Portege G900, Asus P526, and O2 XDA Zinc. It was tested on the
Treo 680, but should work generically.
Since v3, this patch adds power management functions, a scan of the factory bad
block table during initialization, several fixes, and more extensive testing.
Also, the platform data header file, which only contained partitioning
information, was removed. Command-line partitioning can be used, at least until
an mtd parser is written for the saftl format with which these chips are
shipped.
Signed-off-by: Mike Dunn <mikedunn@newsguy.com> Reviewed-by: Robert Jarzmik <robert.jarzmik@free.fr> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>