10 Jan, 2012
40 commits
-
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
This patch is part of a patch-set which changes the MTD interface
from 'mtd->func()' form to 'mtd_func()' form. We need this because
we want to add common code to to all drivers in the mtd core level,
which is impossible with the current interface when MTD clients
call driver functions like 'read()' or 'write()' directly.At this point we just introduce a new inline wrapper function, but
later some of them are expected to gain more code. E.g., the input
parameters check should be moved to the wrappers rather than be
duplicated at many drivers.This particular patch introduced the 'mtd_erase()' interface. The
following patches add all the other interfaces one by one.Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
We are going to re-work the MTD interface and change 'mtd->write()' to
'mtd_write()', 'mtd->read()' to 'mtd_read()' and so forth for all functions
in the 'struct mtd_info' structure.However, mtdchar.c has its own 'mtd_read()', 'mtd_write()', etc functions
which collide with our changes. This patch renames these functions
to 'mtdchar_read()', 'mtdchar_write()', etc.Additionally, to make the 'mtdchar.c' file look consistent, rename
similarly all the other functions starting with 'mtd_'.Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
We are going to re-work the MTD interface and change 'mtd->write()' to
'mtd_write()', 'mtd->read()' to 'mtd_read()' and so forth for all functions
in the 'struct mtd_info' structure.However, logfs has its own 'mtd_read()', 'mtd_write()', etc functions
which collide with our changes. This patch renames these logfs functions
to 'logfs_mtd_read()', 'logfs_mtd_write()', etc.Additionally, to make the 'fs/logfs/dev_mtd.c' file look consistent, rename
similarly all the other functions starting with 'mtd_'.Cc: Jörn Engel
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
We allocate the "mtd" structure using kzalloc which means we do not have
to initialize unused MTD function pointers to NULL, since it is safe to
assume in Linux that NULL contains all zeroes.Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
This patch fixes the following build failure:
In file included from include/linux/mtd/qinfo.h:4:0,
from include/linux/mtd/pfow.h:7,
from drivers/mtd/lpddr/lpddr_cmds.c:27:
include/linux/mtd/map.h: In function 'inline_map_read':
include/linux/mtd/map.h:409:3: error: implicit declaration of function 'BUILD_BUG_ON' [-Werror=implicit-function-declaration]Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
Reimplement a call to devm_request_mem_region followed by a call to ioremap
or ioremap_nocache by a call to devm_request_and_ioremap.The semantic patch that makes this transformation is as follows:
(http://coccinelle.lip6.fr/)//
@nm@
expression myname;
identifier i;
@@struct platform_driver i = { .driver = { .name = myname } };
@@
expression dev,res,size;
expression nm.myname;
@@-if (!devm_request_mem_region(dev, res->start, size,
- \(res->name\|dev_name(dev)\|myname\))) {
- ...
- return ...;
-}
... when != res->start
(
-devm_ioremap(dev,res->start,size)
+devm_request_and_ioremap(dev,res)
|
-devm_ioremap_nocache(dev,res->start,size)
+devm_request_and_ioremap(dev,res)
)
... when any
when != res->start
//Signed-off-by: Julia Lawall
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
Data allocated with devm_ioremap or devm_ioremap_nocache should not be
freed using iounmap, because doing so causes a dangling pointer, and a
subsequent double free.The semantic match that finds this problem is as follows:
(http://coccinelle.lip6.fr/)//
@r@
expression x;
@@
(
x = devm_ioremap(...)
|
x = devm_ioremap_nocache(...)
)@@
expression r.x;
@@
* iounmap(x)
//Signed-off-by: Julia Lawall
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
Add device tree bindings so that the gpio-nand driver may be
instantiated from the device tree. This also allows the partitions
to be specified in the device tree.v7: - restore runtime device tree/non device tree detection
v6: - convert to mtd_device_parse_register()
v5: - fold dt config helpers into a single gpio_nand_of_get_config()
v4: - get io sync address from gpio-control-nand,io-sync-reg
property rather than a resource
- clarified a few details in the binding
v3: - remove redundant cast and a couple of whitespace/naming
changes
v2: - add CONFIG_OF guards for non-dt platforms
- compatible becomes gpio-control-nand
- clarify some binding detailsSigned-off-by: Jamie Iles
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
Only use the values from the image tag if it is valid. Always create
the CFE, NVRAM and linux partitions, to allow flashing a new image even
if the old is invalid without overwriting CFE or NVRAM.Signed-off-by: Jonas Gorski
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
All CRC32 fields are 32 bit integers, so define them as such to prevent
unnecessary casts if we want to use them.Signed-off-by: Jonas Gorski
Signed-off-by: David Woodhouse -
Instead of referencing the sizes of fixed partitions, use the
precomputed CFE/NVRAM lengths.Signed-off-by: Jonas Gorski
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
The CFE boot loader on BCM63XX platforms assumes itself and the NVRAM
partition to be 64 KiB (or erase block sized, if larger).
Ensure this assumption is also met when creating the partitions to
prevent accidential erasure of CFE or NVRAM.Signed-off-by: Jonas Gorski
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
Recent CFEs do not contain the CFE1CFE1 magic anymore, so check for the
"cfe-v" version marker string instead. As very old CFEs do not have
this string, leave the CFE1CFE1 magic as a fallback for detection.Signed-off-by: Jonas Gorski
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
You didn't mean this to be a bool.
Signed-off-by: Rusty Russell
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
As the MTD api has no use for the number of erase cycles
each block has endured, remove the function which calculated
that value.If one day MTD api finds it usefull for wear levelling
algorithms to have this information, the function should be
put back in place.Signed-off-by: Robert Jarzmik
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
- fix NAND_CMD_READID command for ONFI detect.
- add NAND_CMD_PARAM command to read the ONFI parameter page.Signed-off-by: Shengzhou Liu
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
There was a bug for fmr initialization, which lead to fmr was always 0x100
in fsl_elbc_chip_init() and caused FCM command timeout before calling
fsl_elbc_chip_init_tail(), now we initialize CWTO to maximum timeout value
and not relying on the setting of bootloader.Signed-off-by: Shengzhou Liu
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
If we use the Nand flash chip whose number of pages in a block is greater
than 64(for large page), we must treat the low bit of FBAR as being the
high bit of the page address due to the limitation of FCM, it simply uses
the low 6-bits (for large page) of the combined block/page address as the
FPAR component, rather than considering the actual block size.Signed-off-by: Liu Shuo
Signed-off-by: Jerry Huang
Signed-off-by: Tang Yuantian
Signed-off-by: Li Yang
Acked-by: Scott Wood
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
The array of unsigned long pointed by oops_page_used is allocated
by vmalloc which requires the size to be in bytes.BITS_PER_LONG is equal to 32.
If we want to allocate memory for 32 pages with one bit per page then
32 / BITS_PER_LONG is equal to 1 byte that is 8 bits.
To fix it we need to multiply the result by sizeof(unsigned long) equal to 4.Cc: stable@kernel.org
Signed-off-by: Roman Tereshonkov
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
after 250df6ed274d767da844a5d9f05720b804240197
(fs: protect inode->i_state with inode->i_lock), insert_inode_locked()
no longer returns the inode with I_NEW set on failure. However,
the error handler still calls unlock_new_inode() on failure,
which does a WARN_ON if I_NEW is not set, so any failure spews
a lot of warnings.We can just drop the unlock_new_inode() if insert_inode_locked()
fails here.Signed-off-by: Eric Sandeen
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
bcm963xx-flash does nothing meaningful anymore.
Signed-off-by: Jonas Gorski
Acked-by: Florian Fainelli
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
Signed-off-by: Jonas Gorski
Acked-by: Florian Fainelli
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
Arch setup code might want to use their own partition parsers, but still
use the generic physmap flash driver.Signed-off-by: Jonas Gorski
Acked-by: Florian Fainelli
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
Recent BCM63XX devices support a variety of flash types (parallel, SPI,
NAND) and share the partition layout. To prevent code duplication make
the CFE partition parsing code a stand alone mtd parser to allow SPI or
NAND flash drivers to use it.Signed-off-by: Jonas Gorski
Acked-by: Florian Fainelli
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
Replace raw printk's with their pr_XXX equivalent and unify broken up
strings so they become grepable.Also replace the PFX definition with a pr_fmt().
Signed-off-by: Jonas Gorski
Acked-by: Florian Fainelli
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
Signed-off-by: Jonas Gorski
Acked-by: Florian Fainelli
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
Signed-off-by: Jonas Gorski
Acked-by: Florian Fainelli
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
Signed-off-by: Heiko Schocher
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
When we do a non-full-page write, the length be set to FBCR should
not be 'elbc_fcm_ctrl->index', it should be 'elbc_fcm_ctrl->index -
elbc_fcm_ctrl->column'.Signed-off-by: Liu Shuo
Signed-off-by: Li Yang
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
On both of large-page chip and small-page chip, we always should use
'elbc_fcm_ctrl->oob' to set the FPAR_LP_MS/FPAR_SP_MS bit of FPAR, don't
use a overflowed 'column' to set it.Signed-off-by: Liu Shuo
Signed-off-by: Li Yang
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
This patch takes into account checkpatch, sparse and ECC
comments.Signed-off-by: Robert Jarzmik
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
Use block_isbad to check and skip the bad blocks reading.
This will allow to get rid of the read errors if bad blocks
are present initially.Cc: stable@kernel.org
Signed-off-by: Roman Tereshonkov
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
stresstest needs at least two eraseblocks. Bail out gracefully if that
condition is not met. Fixes the following 'division by zero' OOPS:[ 619.100000] mtd_stresstest: MTD device size 131072, eraseblock size 131072, page size 2048, count of eraseblocks 1, pages per eraseblock 64, OOB size 64
[ 619.120000] mtd_stresstest: scanning for bad eraseblocks
[ 619.120000] mtd_stresstest: scanned 1 eraseblocks, 0 are bad
[ 619.130000] mtd_stresstest: doing operations
[ 619.130000] mtd_stresstest: 0 operations done
[ 619.140000] Division by zero in kernel.
...caused by
/* Read or write up 2 eraseblocks at a time - hence 'ebcnt - 1' */
eb %= (ebcnt - 1);Cc: stable@kernel.org
Signed-off-by: Wolfram Sang
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
This patch converts the drivers in drivers/mtd/* to use the
module_platform_driver() macro which makes the code smaller and a bit
simpler.Signed-off-by: Axel Lin
Acked by: Haojian Zhuang
Signed-off-by: Artem BityutskiySigned-off-by: David Woodhouse
-
Add documentation for MSystems disk-on-chip docg3 chips
sysfs entries, which enable and disable protection areas,
giving or disabling access to the chip's memory.Signed-off-by: Robert Jarzmik
Signed-off-by: David Woodhouse -
If doc_probe_device() returned an ERR_PTR, then we accidentally saved
that to docg3_floors[floor] = mtd; which gets derefenced in the error
handling when we call doc_release_device().I've reworked the error handling to take care of that and hopefully
make it a little simpler.Signed-off-by: Dan Carpenter
Acked-by: Robert Jarzmik
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse
