[MTD] NAND: Add Hynix to manufacturer list

Signed-off-by: Nicolas S. Dade <daden@symbol.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Fix it really

tglx declares him self to be the idiot of the day.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] map.h Use the correct macro and fix the resulting compiler warning

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Make map_word_ff ware of the flash buswidth

map_word_ff() was setting the mapword to ~0UL regardless of the
buswidth of the mapped flash chip. The read_map functions are
buswidth aware and therefor the map_word_equal function failed.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] cfi_cmdset_0002: Fix broken status check

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

Merge ... /linux/kernel/git/torvalds/linux-2.6.git

[JFFS2] Whitespace cleanup. Fix missing debug message

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Fix crosscompile

Include sched.h instead of asm/semaphore.h to make it compile everywhere.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Fix cleanup in case of GC-Task not started

Do not call wait_for_completion, when the gc task is not there.
Prevent an oops when the gc thread was not started.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] CFI-0002 - Improve error checking

Check for errors besides infinite loops when writing and erasing.

Signed-off-by: Eric W. Biederman <ebiederman@lnxi.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] CFI flash locking reorg for XIP

This reworks the XIP locking to make sure no lock primitive is ever
called from XIP disabled paths even if in theory they should not
cause any reschedule.  Relying on the current spinlock implementation
is rather fragile and not especially clean from an abstraction pov.
The recent RT work makes it even more obvious.

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Convert thread start semaphore to completion

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] NAND: Honour autoplacement schemes supplied by the caller

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Kill GC thread before cleanup

First kill GC thread, then start clearing the internal structures

Signed-off-by: Artem B. Bityuckiy <dedekind@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Suppress annoying debug messages

Embrace uneeded messages in D1().

Signed-off-by: Artem B. Bityuckiy <dedekind@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Fix NOR only compile

Fix the bug that caouses problems when compiling for NOR.
We read  a newly erased block so we don't need to check ECC.

Define jffs2_is_writebuffered as zero if there is no wbuf.

Signed-off-by: Artem B. Bityuckiy <dedekind@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] CFI DEBUG_LOCK_BITS fixes for Intel NOR flash:

adjust chip-relative offsets to
block address, write to block address + 2 per recent datasheets.

Signed-off-by: Todd Poynor <tpoynor@mvista.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] mtdram: Quick cleanup of the driver:

- Lindent
- Removal of slram/phram functionality
- Removal of most #ifdefs

Signed-off-by: Joern Engel <joern@wohnheim.fh-wedel.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Fix inode allocation race

Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] NAND: Fix the broken dynamic array allocations

Reverting the change from 1.136 to 1.137 (back to static allocation of ecc
arrays) due to stack corruption and ecc errors.

Signed-off-by: Jarkko Lavinen <jarkko.lavinen@nokia.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Fix broken user ABI

Move kernel data where it belongs. Previous change broke user abi.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Fix race in garbage collector

Fix the race problem described here:
http://lists.infradead.org/pipermail/linux-mtd/2005-April/012361.html

Signed-off-by: Artem B. Bityuckiy <dedekind@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] DiskOnChip: Add some comments

Add helpful comment about oobfree so I can't claim two years from
now that I don't remember what I was thinking.

Signed-off-by: Dan Brown <dan_brown@ieee.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] DiskOnChip: Fix compile w/o CONFIG_MTD_PARTITIONS.

Signed-off-by: Dan Brown <dan_brown@ieee.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] DiskOnChip: Prevent problems with existing filesystems

Try not to break existing jffs2 installs, instead break oobfree into
two out-of-order pieces.

Signed-off-by: Dan Brown <dan_brown@ieee.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] NAND: Fix reading of autoplaced OOB when there are multiple free sections.
Signed-off-by: Dan Brown <dan_brown@ieee.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] DiskOnChip: Fix (?) free OOB array info.

I really hope this doesn't break something.

Signed-off-by: Dan Brown <dan_brown@ieee.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Add KERN_DEBUG level to printks

Signed-off-by: Artem B. Bityuckiy <dedekind@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Fix race problems with wbuf.

Signed-off-by: Artem B. Bityuckiy <dedekind@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] NAND: Move the NULL check into the calling function

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] NAND: Fix missing NULL pointer check

Version 1.137 broke nand_read_ecc clients who pass NULL oobsel.  Fixed.

Signed-off-by: Dan Brown <dan_brown@ieee.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Fix OTP for top-parameter devices

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Reset file position when switching OTP mode

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] NAND: Fix oob available calculation

Use oobfree to calculate the number of oob bytes available for fs usage

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Add reboot notifier to Intel NOR flash driver

to make sure the flash is in array mode whenever we're about to
reboot. This is especially useful to allow "soft" reboot to work
which consists of branching back into the bootloader.

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Avoid compile warnings for Intel CFI flash without OTP support.

Signed-off-by: Todd Poynor <tpoynor@mvista.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] cfi_cmdset_0001: Skip delay if Instant Block Locking is set

Skip jiffy delay after each block lock/unlock for Intel CFI flash
with the "Instant Individual Block Locking" feature bit set.

Signed-off-by: Todd Poynor <tpoynor@mvista.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] DiskOnChip: Scan the entire device for Media Headers.

Add a new module param, show_firmware_partition.

Signed-off-by: Dan Brown <dan_brown@ieee.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] NAND: Use arrays of needed size instead of constant-sized.

Signed-off-by: Artem B. Bityuckiy <dedekind@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] sharpsl-flash: Correct error paths

Signed-off-by: Richard Purdie <rpurdie@rpsys.net>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Add support for more SharpSL machines and fix missing mapping init

Signed-off-by: Richard Purdie <rpurdie@rpsys.net>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Prevent deadlock when flushing write buffer

Signed-off-by: Artem B. Bityuckiy <dedekind@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Forbid to free inode_cache objects if its nlink isn't zero.

Signed-off-by: Artem B. Bityuckiy <dedekind@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Improve garbage collector block selection

Make sure the erase_pending_wbuf_list's blocks are taken into account
when picking the block to GC.

Signed-off-by: Artem B. Bityuckiy <dedekind@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] plat-ram: Make it usable on non ARM platforms

Use memset instead of ARM only memzero function

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] block2mtd: Fix incompatible pointer type

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] cfi_cmdset_0001: Fix compiler warnings

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] NAND nandsim: Use NAND_SKIP_BBT option

Use the new NAND_SKIP_BBT option instead of defining a fake scan_bbt
handler.

Signed-off-by: Artem B. Bityuckiy <dedekind@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] ixp2000: Remove port setting code

Setting the slowport to 8-bit mode is something that ought to be
done in the IXP2000 generic code, not in the MTD map driver.  See
the description for ARM patch 2493/1 for an explanation.

http://www.arm.linux.org.uk/developer/patches/viewpatch.php?id=2493/1

Now that 2493/1 has been accepted and will be upstream soon, this
doesn't need to be done in the map driver anymore.

Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Kernel Janitor fixes.

Signed-off-by: Domen Puncer <domen@coderock.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Use function to manipulate superblock dirty flag

Use the corresponding function to mark Superblock dirty instead
of doing it directly.

Signed-off-by: Artem B. Bityuckiy <dedekind@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Fix typo in Kconfig

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] NAND s3c2410: Simplify command handling

Updated with tglx's suggestion to simply the command invocation by
simply changing the address of the IO write area

Signed-off-by: Ben Dooks <ben@simtec.co.uk>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] cfi_cmdset_0001: Fix state after sync

oldstate has to be reset to FL_READY after sync completion.

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] cfi_cmdset_0001: Fix the buggy status check.

The change makes the code endianess aware and replaces the bogus
nested loop to or the status flags together.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Add the reverse operation of cfi_build_cmd()

This is necessary to fix the broken status check in cfi_cmdset_0001

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] plat-ram: removed extraneous debugging code

removed define of DEBUG
removed extraneous debugging code

Signed-off-by: Ben Dooks <ben@simtec.co.uk>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] NAND: Fixed unused loop variable

Signed-off-by: Ben Dooks <ben@simtec.co.uk>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Fixed signed 1bit bitfield

Signed-off-by: Ben Dooks <ben@simtec.co.uk>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] phram: Allow short reads.

Jffs2 apparently needs this.
Accept newline at the end of input.

Signed-off-by: Joern Engel <joern@wohnheim.fh-wedel.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] block2mtd: Remove copyright. Fix offset calculation

- Remove Gareth from the Copyrights (at his own request)
- Fix the "fscking embarrassment"

Signed-off-by: Joern Engel <joern@wohnheim.fh-wedel.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Remove Elan-104NC

Remove support for the Arcom Elan-104NC since it's no longer being maintained.

Signed-off-by: David Vrabel <dvrabel@arcom.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Add symlink caching support.

Signed-off-by: Artem B. Bityuckiy <dedekind@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] NAND: Use cond_resched instead of msleep

Replace msleep by cond_resched. On machines with HZ=100 (e.g. ARM)
msleep slows down the operation by factor 10

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Fix NOR specific scan BUG

Fix fairly sad NOR-specific bug - during FS building ic->scan_dents
isn't zero, but jffs2_mark_node_obsolete() migt be called it tries to
finde the ic corresponding to ref - this requires ic->scan_dents = 0.

Signed-off-by: Artem B. Bityuckiy <dedekind@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Prevent ino cache removal for inodes in use

Don't remove inocache for inodes which are in read_inode() or
clear_inode() until they're done.

Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Replace all the Au1x mapping drivers with a simplified single driver

This driver does not have as many options but it's easier to maintain.
And, it turns out AMD never shipped boards with different flash densities.

Signed-off-by: Pete Popov <ppopov@pacbell.net>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Use after free, found by the Coverity tool

Signed-off-by: Alexander Nyberg <alexn@dsv.su.se>
Signed-off-by: Joern Engel <joern@wohnheim.fh-wedel.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] NAND: Check command timeout

Check timeout while we wait for the command to finish. No worry about a
false result. This prevents deadlocking when detecting an unknown number
of chips and is useful for removable media too.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] DiskOnChip: Wait for the command to finish.

Do not use the ready function here, as it might hang for ever.
The result will show, whether the chip is there or not

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Sparse fixes

Fix sparse errors due to lack of address-space markers
Updated header comments
Small re-format of initialiser

Signed-off-by: Ben Dooks <ben@simtec.co.uk>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Fix unregister_mtd_user() public function documentation.

Signed-off-by: Artem B. Bityuckiy <dedekind@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Update BAST driver configuration

update the BAST driver config (which already supports the vr1000) to
be selected only if the vr1000 has been configured

Signed-off-by: Ben Dooks <ben@simtec.co.uk>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Make OTP actually work.

The OTP code is rather broken without this.

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Fix node lookup

Look the ref->next_phys field instead of ->next_in_ino to determine
if the block has more then one node.

Signed-off-by: Artem B. Bityuckiy <dedekind@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] NAND: Fix bad block table scan for small page devices

Scan 1st and 2nd pages of SP devices for BB marker by default.
Fix more then one page scanning in create_bbt.c.

Signed-off-by: Artem B. Bityuckiy <dedekind@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Quiet unused variable warning

Signed-off-by: Nioclas Pitre <nico@cam.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] NAND: Early Manufacturer ID lookup

Move manufacturer ID search to display correct ID in case of buswidth
mismatch.

Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Add SST 39VF1601 (MPF+) ID

Signed-off-by: Ben Dooks <ben@simtec.co.uk>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Fixup probing logic for single 16bit devices

The change to the generic probe to look for the
smallest width of chip first is causing some problems
on boards with a single 16bit device.

The problem seems to be the jedec_match() is truncating
the device-id read from the table to match against the
one read from the hardware, causing a match against the
partial id of some chips with 16bit IDs (such as the
SST39LF160)

This fixes things for my own board, but something may
need to be done if the same problem is exhibited for
chips with an 8bit ID

Signed-off-by: Ben Dooks <ben@simtec.co.uk>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] NAND: Read only OOB bytes during bad block scan

When scanning NAND for bad blocks, don't read the whole page, read
only needed OOB bytes instead. Also check the return code of the
nand_read_raw() function. Correctly free the this->bbt array in
case of failure. Tested with Large page NAND.

Fix debugging message.

Signed-off-by: Artem B. Bityuckiy <dedekind@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] NAND: Allow operation without bad block table

Small bugfix. Sometimes it may be handy not to have bbt.
So, this->bbt might be NULL.

Signed-off-by: Artem B. Bityuckiy <dedekind@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] NAND: Skip bad block table scan on request

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Use a single config option for write buffer support

This patch replaces the current CONFIG_JFFS2_FS_NAND, CONFIG_JFFS2_FS_NOR_ECC
and CONFIG_JFFS2_FS_DATAFLASH with a single configuration option -
CONFIG_JFFS2_FS_WRITEBUFFER.

The only functional change of this patch is that the slower div/mod
calculations for SECTOR_ADDR(), PAGE_DIV() and PAGE_MOD() are now always
used when CONFIG_JFFS2_FS_WRITEBUFFER is enabled.

Signed-off-by: Andrew Victor <andrew@sanpeople.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Add support for JFFS2-on-Dataflash devices.

For Dataflash, can_mark_obsolete = false and the NAND write buffering
code (wbuf.c) is used.

Since the DataFlash chip will automatically erase pages when writing,
the cleanmarkers are not needed - so cleanmarker_oob = false and
cleanmarker_size = 0

DataFlash page-sizes are not a power of two (they're multiples of 528
bytes).  The SECTOR_ADDR macro (added in the previous core patch) is
replaced with a (slower) div/mod version if CONFIG_JFFS2_FS_DATAFLASH is
selected.

Signed-off-by: Andrew Victor <andrew@sanpeople.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Core changes required to support JFFS2-on-Dataflash devices.

DataFlash page-sizes are not a power of two (they're multiples of 528
bytes).  There are a few places in JFFS2 code where sector_size is used
as a bitmask.  A new macro (SECTOR_ADDR) was defined to calculate these
sector addresses. For non-DataFlash devices, the original (faster)
bitmask operation is still used.

In scan.c, the EMPTY_SCAN_SIZE was a constant of 1024.
Since this could be larger than the sector size of the DataFlash, this
is now basically set to MIN(sector_size, 1024).

Addition of a jffs2_is_writebuffered() macro.

Signed-off-by: Andrew Victor <andrew@sanpeople.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Unabuse file-f_mode for OTP purpose

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] User interface to Protection Registers

This is implemented using a ioctl to switch the MTD char device into
one of the different OTP "modes", at which point read/write/seek can
operate on the selected OTP area.  Also some extra ioctls to query
for size and lock protection segments or groups.  Some example user
space utilities are provided.

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Support for protection register support on Intel FLASH chips

This enables support for reading, writing and locking so called
"Protection Registers" present on some flash chips.
A subset of them are pre-programmed at the factory with a
unique set of values. The rest is user-programmable.

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Prevent deadlock during write buffer recovery

Prevent deadlock when checking erased block for
space allocation during wbuf recovery.

Signed-off-by: Estelle Hammache <estelle.hammache@st.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Add OTP basisc

add structure definition for OTP region info

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] amd_flash: Fix chip ID clash

* Removed table entry for AM29BDS643D, since device ID clashes with AM29DL640G
and both chips support CFI.

Signed-off-by: Jonas Holmberg <jonas.holmberg@axis.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Remove NAND dependencies for NOR FLASH

make NAND code work on NOR flash again

Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] DiskOnChip: big endian fix for NFTL devices

Make NFTL devices work on big endian machines.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] DiskOnChip code cleanup

Remove commented ugliness

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] DiskOnChip use CONFIG_ options instead of random symbols

Using the CONFIG_ options from KConfig seems to work better :8

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Code cleanup

Code beautification and block filing correction for optimization.

Signed-off-by: Estelle Hammache <estelle.hammache@st.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[MTD] Fix MTD device probing

Try larger numbers of chips before smaller
numbers of chips across the bus width.

This means we'll avoid misdetecting a 2 x16 array as 1 x32 if the
high 16-bits happen to read as zeros in the QRY area.

Signed-off-by: Russell King <linux@arm.linux.org.uk>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Fix refile of blocks due to write failure.

avoid segfault when nextblock was refiled because of a write failure
- avoid filing blocks on the clean list when they have wasted
space

Signed-off-by: Estelle Hammache <estelle.hammache@st.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

[JFFS2] Fix block refiling

- block refiling when writing directly to flash a buffer
which is bigger than wbuf
- retry cases for flushing wbuf

Signed-off-by: Estelle Hammache <estelle.hammache@st.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>