We've noticed that sometimes "MTD do_write_buffer(): software timeout"
message was printed out when writing to a Fujitsu NOR flash.
It turned out that this was because of a race in the timeout handling
do_write_buffer(). A small timeout of (HZ / 1000) + 1 is used there, and
sometimes if the timer interrupt handling takes more than one or even two
jiffies (which is 1-2 ms with HZ == 1000) and that interrupt happens just
after chip_ready() call, the driver bails out from a ready polling loop
despite the chip has actually become ready while all those interrupts were
handled. To deal with this issue, extra check for chip ready is neccessary on
timeout expiration (and the checks should better be reordered).
As do_write_oneword() uses the same approach, it needs to also be changed.
Signed-off-by: Konstantin Baidarov <kbaidarov@ru.mvista.com>
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Change to the extended cfi table parsing for Intel NOR flash that uses
the info in the extended table to 'walk' the table rather than using
hard coding for various primary extended query table version numbers.
From: Jared Hulbert <jaredeh@gmail.com>
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
While this might be useful for all supported flash types, it is mandatory
for proper JFFS2 support with Sibley flash.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
This updates the Primary Vendor-Specific Extended Query parsing to
version 1.4 in order to get the information about the Configurable
Programming Mode regions implemented in the Sibley flash, as well as
selecting the appropriate write command code.
This flash does not behave like traditional NOR flash when writing data.
While mtdblock should just work, further changes are needed for JFFS2 use.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
This includes improved error handling/reporting plus some other
message cleanups.
Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
I recently picked up my older work to remove unnecessary #includes of
sched.h, starting from a patch by Dave Jones to not include sched.h
from module.h. This reduces the number of indirect includes of sched.h
by ~300. Another ~400 pointless direct includes can be removed after
this disentangling (patch to follow later).
However, quite a few indirect includes need to be fixed up for this.
In order to feed the patches through -mm with as little disturbance as
possible, I've split out the fixes I accumulated up to now (complete for
i386 and x86_64, more archs to follow later) and post them before the real
patch. This way this large part of the patch is kept simple with only
adding #includes, and all hunks are independent of each other. So if any
hunk rejects or gets in the way of other patches, just drop it. My scripts
will pick it up again in the next round.
Signed-off-by: Tim Schmielau <tim@physik3.uni-rostock.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Remove useless udelay(100) after status value already read. Poll
for status OK with reduced udelay if not immediate OK status return.
Fix read and compare of 32-bit status value using 16-bit variable.
Include slab.h since kmalloc/kfree are called.
Signed-off-by: Todd Poynor <tpoynor@mvista.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Modify Amd/Fujitsu CFI NOR flash primary vendor extension table revision
check to recognize version 1.4. Verified the existing driver can
handle version 1.4 chips without additional info from 1.4 extended table.
Move the primary vendor extension table revision check from common file
to the 3 CFI chip driver files, since the data structures and revisions
handled by those data structures are specific to the chip driver.
Modify the error message printed when the revision is unknown to be a
KERN_ERR instead of WARNING since this will cause mtd to ignore the chip.
Signed-off-by: Todd Poynor <tpoynor@mvista.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
ARM is the only known user of this at the moment.
Prevent allyes builds for other archs from failing
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
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>
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>
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>
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>
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>
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>
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>
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>
* 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>
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>
Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.
Let it rip!
