set_ecc_config function only set ECC_CORRECTION register, so
move register setting to probe function.
Since controller only support 15bit and 8bit ecc correction,
updated nand ecc layout information.
Signed-off-by: Chuanxiao Dong <chuanxiao.dong@intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Hi David,
I sent 4 patches using my intel email account. If there is any
problem about the format of these patches, I will resend them after
I arrived at home by using my gmail account, and I will keep on
using gmail account to send patches.
Thanks.
>From 242e3bf5e17f54b1df8cf285154a7c7a61ff62e9 Mon Sep 17 00:00:00 2001
From: Chuanxiao Dong <chuanxiao.dong@intel.com>
Date: Fri, 6 Aug 2010 15:29:41 +0800
Subject: [PATCH 1/4] mtd: denali: Remove device_info_tag structure.
Most of the variables in this structure are useless, so just
remove this structure and relevant codes.
Signed-off-by: Chuanxiao Dong <chuanxiao.dong@intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
This patch fixes
drivers/mtd/nand/atmel_nand.c:372: warning: 'part_probes' defined but not used
which is issued when CONFIG_MTD_PARTITIONS is defined but
CONFIG_MTD_CMDLINE_PARTS not.
Signed-off-by: Andreas Bießmann <biessmann@corscience.de>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
This patch corrects a problem where gen_nand driver assumed there can be only
one chip and ignored the pdata->chip.nr_chips value.
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
In denali driver, wTotalBlocks variable is useless, so just
remove them in get_toshiba_nand_para and get_hynix_nand_para
first. Other wTotalBlocks in denali.c and other variables in
dev_info struct are also useless, will remove them later.
Also add a parameter in get_hynix_nand_para to prepare to
remove dev_info struct. This parameter comes by reading NAND
ID directly. Reading register will report wrong value for
some NAND chips, e.g. some Micron MT29F32G08QAA chips.
So update NAND_CMD_READID method as well.
Signed-off-by: Chuanxiao Dong <chuanxiao.dong@intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
rename these functions' name and remove additional declarations
in header file
Signed-off-by: Chuanxiao Dong <chuanxiao.dong@intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Hi,
I have changed the outlook mail cliet to be linux mutt client and use my
personal gmail to submit patches.
Here are 5 new patches to fix nand/denali check patch errors. The other
4 patches will be sent out after this mail.
Thanks for your review.
>From d125ad3f57bbf517131dccad6b5933edf8c2632a Mon Sep 17 00:00:00 2001
From: Chuanxiao Dong <chuanxiao.dong@intel.com>
Date: Tue, 3 Aug 2010 15:54:48 +0800
Subject: [PATCH 1/5] mtd: denali.c: clean up all whitespaces in code indent
Signed-off-by: Chuanxiao Dong <chuanxiao.dong@intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The IRQSTAT register is a W1C register used by the interrupt handler and
may have its BUSY bit changed. This makes it somewhat unreliable for the
polling devready function. So switch it over to use the BUSY bit in the
STAT register that always reflects the current state of the hardware.
This fixes driver hangs seen when the NAND flash is under heavy system
load (like I/O benchmarks).
Signed-off-by: Barry Song <barry.song@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Rather than forcing the platform resources to declare the desired page
size, simply use the existing information passed down to us by the higher
layers. This way we work out of the box with all flash chips that the
kernel knows about.
Signed-off-by: Barry Song <barry.song@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
When resetting the ECC registers/counters, the bit will automatically
clear itself once the reset has actually finished. So make sure we
wait for that to occur before doing anything else rather than assuming
everything is peachy and proceeding with stale ECC values.
Signed-off-by: Barry Song <barry.song@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
We used the platform rd_dly field when we meant to use the wr_dly field.
Signed-off-by: Barry Song <barry.song@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The NAND base may send some controls which are neither CLE nor ALE, so
we need to explicitly check both instead of assuming things are always
one or the other. Otherwise, we sometimes send out illegal addresses
to the NAND device.
Signed-off-by: Barry Song <barry.song@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Our write_buf/read_buf funcs always do ECC in HW ECC mode. That is not
needed for raw funcs. In fact, write_buf/read_buf should be a pure func
for data input/output while chip->ecc.hwctl controls ECC. Unfortunately,
we can't separate ECC from normal data input/output in our NFC, so our DMA
write_buf/read_buf entries are coupled with ECC operations closely.
Thus we need to provide dedicated read_page_raw/write_page_raw funcs where
we do non-DMA transactions so as to avoid automatic ECC.
Signed-off-by: Barry Song <barry.song@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The low level NAND driver doesn't care about filesystems, so punt the
local comment indicating otherwise.
Signed-off-by: Barry Song <barry.song@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The mtd layer is responsible to call mtd->{suspend,resume}. Doing it
again in the driver is wrong and results in a warning:
nand_resume called for a chip which is not in suspended state
at resume time. Removing the calls from the resume and suspend
functions makes them empty allowing them to be deleted completely.
Signed-off-by: Eric Bénard <eric@eukrea.com>
Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The variable 'syn' was being used uninitialized. Also
fixed incorrect use of syn[] vs s[].
Tested on powerpc board with 64MiB DOC2000.
Signed-off-by: Mark Ware <mware@elphinstone.net>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
This is an update that depends on the previous patches I sent.
We can now write to all the appropriate BB marker locations (i.e.
pages 1 AND 2, bytes 1 AND 6) with nand_default_block_markbad() if
necessary, according to the flags marked in chip->options.
Note that I removed the line:
ofs += mtd->oobsize;
Unless I am wrong, this line was completely unnecessary in the
first place.
Signed-off-by: Brian Norris <norris@broadcom.com>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Fix all checkpatch.pl complaints.
Artem: tweaked a little and fix tab indentations, so now this is not
only about checkpatch, but also about making indentations look
sane.
Signed-off-by: Chuanxiao Dong <chuanxiao.dong@intel.com>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
On TI's DA830/OMAP-L137, DA850/OMAP-L138 and DM365, after setting the
4BITECC_ADD_CALC_START bit in the NAND Flash control register to 1 and
before waiting for the NAND Flash status register to be equal to 1, 2 or
3, we have to wait till the ECC HW goes to correction state. Without this
wait, ECC correction calculations will not be proper.
This has been tested on DA830/OMAP-L137, DA850/OMAP-L138, DM355 and DM365
EVMs.
Signed-off-by: Sudhakar Rajashekhara <sudhakar.raj@ti.com>
Acked-by: Sneha Narnakaje <nsnehaprabha@ti.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
This is a revision to PATCH 2/2 that I sent. Link:
http://lists.infradead.org/pipermail/linux-mtd/2010-July/030911.html
Added new flag for scanning of both bytes 1 and 6 of the OOB for
a BB marker (instead of simply one or the other).
The "check_pattern" and "check_short_pattern" functions were updated
to include support for scanning the two different locations in the OOB.
In order to handle increases in variety of necessary scanning patterns,
I implemented dynamic memory allocation of nand_bbt_descr structs
in new function 'nand_create_default_bbt_descr()'. This replaces
some increasingly-unwieldy, statically-declared descriptors. It can
replace several more (e.g. "flashbased" structs). However, I do not
test the flashbased options personally.
How this was tested:
I referenced 30+ data sheets (covering 100+ parts), and I tested a
selection of 10 different chips to varying degrees. Particularly, I
tested the creation of bad-block descriptors and basic BB scanning on
three parts:
ST NAND04GW3B2D, 2K page
ST NAND128W3A, 512B page
Samsung K9F1G08U0A, 2K page
To test these, I wrote some fake bad block markers to the flash (in OOB
bytes 1, 6, and elsewhere) to see if the scanning routine would detect
them properly. However, this method was somewhat limited because the
driver I am using has some bugs in its OOB write functionality.
Signed-off-by: Brian Norris <norris@broadcom.com>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Some level of support for various scanning locations was already built in,
but this required clean-up. First, BB marker location cannot be determined
_only_ by the page size. Instead, I implemented some heuristic detection
based on data sheets from various manufacturers (all found in
nand_base.c:nand_get_flash_type()).
Second, once these options were identified, they were not handled properly
by nand_bbt.c:nand_default_bbt(). I updated the static nand_bbt_desc structs
to reflect the need for more combinations of detection. The memory allocation
here probably needs to be done dynamically in the very near future (see next
patches).
Signed-off-by: Brian Norris <norris@broadcom.com>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Current driver prevents command-line partitions from being parsed when built-in
partitions are defined in s3c2410_nand_set object, but it is not desirable in some
cases. This patch tries to parse commad-line partitions prior to the built-in.
Signed-off-by: Conke Hu <conke@maxwit.com>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
NAND_BB_LAST_PAGE used to be in nand.h, but it pertained to bad block
management and so belongs next to NAND_BBT_SCAN2NDPAGE in bbm.h. Also,
its previous flag value (0x00000400) conflicted with NAND_BBT_SCANALLPAGES
so I changed its value to 0x00008000. All uses of the name were modified to
provide consistency with other "NAND_BBT_*" flags.
Signed-off-by: Brian Norris <norris@broadcom.com>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Included the basic size info for NAND chips with ID of 0xAD or
0xD7. The first can be found in Hynix HY27SF161G2M, while the
second can be found in Micron MT29F64G08 and the Samsung K9LBG08U0D
(among others). Also, some 64 Gbit (or larger) chips identify as
0xD7 because they contain multiple smaller 32 Gbit chips. I
assume it's safe to classify these under the 32 Gbit listing.
Signed-off-by: Brian Norris <norris@broadcom.com>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
This fixes:
drivers/mtd/nand/mxc_nand.c: In function 'mxcnd_resume':
drivers/mtd/nand/mxc_nand.c:901: warning: unused variable 'host'
Removing this variable was missed in 9c14b153e6.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Look for dependency checks for "FOO" when inside of an "if FOO" block and remove them.
Signed-off-by: Kyle Spaans <kspaans@uwaterloo.ca>
Reviewed-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The OOB handling in the mxc_nand driver is broken for v1 type
controllers (i.MX27/31) with 512 byte page size. This perhaps
did not show up because ubi does not use OOB.
Update the driver to always read/write a whole page even if
only OOB is requested. With this patch the driver passes the
mtd_oobtest on i.MX27 with 512 byte page size. Also tested
with 2048 byte page size and on i.MX35 (v2 type controller)
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Fixes build errors in drivers caused by the OF device_node
pointer being moved into struct device
Signed-off-by: Anatolij Gustschin <agust@denx.de>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Move MTD_NAND_ECC and MTD_NAND_ECC_SMC above NAND memuconfig, to unbreak
display in xconfig. This shouldn't change any dependencies.
Signed-off-by: Maxim Levitsky <maximlevitsky@gmail.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
This turns out to be the reason for DMA timeouts on resume,
if card was inserted while system was suspended
Signed-off-by: Maxim Levitsky <maximlevitsky@gmail.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
* Don't call complete on dma completion
* do a INIT_COMPLETE before using it each time
* Report DMA read error via ecc 'correct'
I finally managed to make my system do suspend to ram propertly, and I see that
if card was inserted during suspend (while system was off), I get dma timeouts
on resume. Simple card reinsert solves the issue.
This patch solves a crash that would happen otherwise
Signed-off-by: Maxim Levitsky <maximlevitsky@gmail.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Otherwise, if it fires right away, it might access
uninitialized spinlock
Signed-off-by: Maxim Levitsky <maximlevitsky@gmail.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
This reverts commit 66803762 ("mtd: mxc_nand: add RESET command support").
Support for NAND_CMD_RESET was added separately in commit d4840180
("mtd: mxc_nand: set NFC registers after reset"), causing a build error:
drivers/mtd/nand/mxc_nand.c: In function 'mxc_nand_command':
drivers/mtd/nand/mxc_nand.c:689: error: duplicate case value
drivers/mtd/nand/mxc_nand.c:606: error: previously used here
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Implicit slab.h inclusion via percpu.h is about to go away. Make sure
gfp.h or slab.h is included as necessary.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
This NAND flash part advertises 0xD1 as an identifier but is still a working
128MBytes x 8bits 3.3V NAND part.
Signed-off-by: Florian Fainelli <ffainelli@freebox.fr>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
This is a slightly modified version of a patch submitted last year by
Reuben Dowle <reuben.dowle@navico.com>. His original comments follow:
This patch adds support for some MLC NAND flashes that place the BB
marker in the LAST page of the bad block rather than the FIRST page used
for SLC NAND and other types of MLC nand.
Lifted from Samsung datasheet for K9LG8G08U0A (1Gbyte MLC NAND):
"
Identifying Initial Invalid Block(s)
All device locations are erased(FFh) except locations where the initial
invalid block(s) information is written prior to shipping. The initial
invalid block(s) status is defined by the 1st byte in the spare area.
Samsung makes sure that the last page of every initial invalid block has
non-FFh data at the column address of 2,048.
...
"
As far as I can tell, this is the same for all Samsung MLC nand, and in
fact the samsung bsp for the processor used in our project (s3c6410)
actually contained a hack similar to this patch but less portable to
enable use of their NAND parts. I discovered this problem when trying to
use a Micron NAND which does not used this layout - I wish samsung would
put their stuff in main-line to avoid this type of problem.
Currently this patch causes all MLC nand with manufacturer codes from
Samsung and ST(Numonyx) to use this alternative location, since these
are the manufactures that I know of that use this layout.
Signed-off-by: Kevin Cernekee <cernekee@gmail.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
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. For instance:
Samsung K9GAG08U0M (5-byte sequence): ec d5 14 b6 74
4th byte, bits 1:0 encode the page size: 0=1KiB, 1=2KiB, 2=4KiB, 3=8KiB
4th byte, bits 5:4 encode the block size: 0=64KiB, 1=128KiB, ...
4th byte, bit 6 encodes the OOB size: 0=8B/512B, 1=16B/512B
Samsung K9GAG08U0D (6-byte sequence): ec d5 94 29 34 41
4th byte, bits 1:0 encode the page size: 0=2KiB, 1=4KiB, 3=8KiB, 4=rsvd
4th byte, bits 7;5:4 encode the block size: 0=128KiB, 1=256KiB, ...
4th byte, bits 6;3:2 encode the OOB size: 1=128B/page, 2=218B/page
This patch uses the new 6-byte scheme if the following conditions are
all true:
1) The ID code wraps around after exactly 6 bytes
2) Manufacturer is Samsung
3) 6th byte is zero
The patch also extends the maximum OOB size from 128B to 256B.
Signed-off-by: Kevin Cernekee <cernekee@gmail.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
