The extra race-with-truncate-then-retry logic around
ext3_get_block_handle(), which was inherited from ext2, becomes unecessary
for ext3, since we have already obtained the ei->truncate_sem in
ext3_get_block_handle() before calling ext3_alloc_branch().  The
ei->truncate_sem is already there to block concurrent truncate and block
allocation on the same inode.  So the inode's indirect addressing tree
won't be changed after we grab that semaphore.

We could, after get the semaphore, re-verify the branch is up-to-date or
not.  If it has been changed, then get the updated branch.  If we still
need block allocation, we will have a safe version of the branch to work
with in the ext3_find_goal()/ext3_splice_branch().

The code becomes more readable after remove those retry logic.  The patch
also clean up some gotos in ext3_get_block_handle() to make it more
readable.
Signed-off-by: NMingming Cao <cmm@us.ibm.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

msp3400 update: Fix and enable "simpler" mode, some other minor fixes.
Signed-off-by: NGerd Knorr <kraxel@bytesex.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

update cx22702 fe driver, add support for using the dvb pll lib, enable
cx22702 support in cx88-dvb.
Signed-off-by: NGerd Knorr <kraxel@bytesex.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Fixup error path, without that one the driver kills the machine by oopsing
in the IRQ handler in case the frontend initialization fails.
Signed-off-by: NGerd Knorr <kraxel@bytesex.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Since we only access reiserfs_key ->u.k_offset_v2 guts in four helper
functions, we are free to sanitize those, as long as

- layout of the structure is unchanged (it's on-disk object)

- behaviour of these helpers is same as before.

Patch kills the mess with endianness-dependent bitfields and replaces them
with a single __le64.  Helpers are switched to straightforward shift/and/or.

Benefits:

- exact same definitions for little- and big-endian architectures; no ifdefs
  in sight.

- generate the same code on little-endian and improved on big-endian.

- doesn't rely on lousy bitfields handling in gcc codegenerator.

- happens to be standard C (unsigned long long is not a valid type for a
  bitfield; it's a gccism and not well-implemented one, at that).
Signed-off-by: NAl Viro <viro@parcelfarce.linux.theplanet.co.uk>
Cc: <reiserfs-dev@namesys.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

comp_short_keys() massaged into sane form, which kills the last place where
pointer to in_core_key (or any object containing such) would be cast to or
from something else.  At that point we are free to change layout of
in_core_key - nothing depends on it anymore.

So we drop the mess with union in there and simply use (unconditional) __u64
k_offset and __u8 k_type instead; places using in_core_key switched to those.
That gives _far_ better code than current mess - on all platforms.
Signed-off-by: NAl Viro <viro@parcelfarce.linux.theplanet.co.uk>
Cc: <reiserfs-dev@namesys.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

fixes for a couple of bugs exposed by the above: le32_to_cpu() used on 16bit
value and missing conversion in comparison of host- and little-endian values.
Signed-off-by: NAl Viro <viro@parcelfarce.linux.theplanet.co.uk>
Cc: <reiserfs-dev@namesys.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

little-endian objects annotated as such; again, obviously no changes of
resulting code, we only replace __u16 with __le16, etc.  in relevant places.
Signed-off-by: NAl Viro <viro@parcelfarce.linux.theplanet.co.uk>
Cc: <reiserfs-dev@namesys.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

struct reiserfs_key cloned; (currently) identical struct in_core_key added.
Places that expect host-endian data in reiserfs_key switched to in_core_key.
Basically, we get annotation of reiserfs_key users and keep the resulting tree
obviously equivalent to original.
Signed-off-by: NAl Viro <viro@parcelfarce.linux.theplanet.co.uk>
Cc: <reiserfs-dev@namesys.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Bump autofs4 version so we know what's going on.
Signed-off-by: NIan Kent <raven@themaw.net>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

For tree mount maps, a call to chdir or chroot, to a directory above the
moint point directories at a certain time during the expire results in the
expire incorrectly thinking the tree is not busy.  This patch adds a check
to see if the filesystem above the tree mount points is busy and also locks
the filesystem during the tree mount expire to prevent the race.
Signed-off-by: NIan Kent <raven@themaw.net>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

It's possible for an event wait request to arive before the event
requestor.  If this happens the daemon never gets notified and autofs
hangs.
Signed-off-by: NIan Kent <raven@themaw.net>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

This patch fixes the leak of sb->s_fs_info in both the HFS and HFS+
modules.  In addition to this, it fixes an oops happening when trying to
mount a non-hfsplus filesystem using hfsplus.  This patch is from Roman
Zippel, based off patches sent by myself.
Signed-off-by: NColin Leroy <colin@colino.net>
Signed-off-by: NRoman Zippel <zippel@linux-m68k.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

This patch optimizes io_submit_one to call aio_run_iocb() directly if
ctx->run_list is empty.  When the list is empty, the operation of adding to
the list, then call to __aio_run_iocbs() is unnecessary because these
operations are done in one atomic step.  ctx->run_list always has only one
element in this case.  This optimization speeds up industry standard db
transaction processing benchmark by 0.2%.
Signed-off-by: NKen Chen <kenneth.w.chen@intel.com>
Cc: Benjamin LaHaise <bcrl@kvack.org>
Cc: Suparna Bhattacharya <suparna@in.ibm.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Clean up code that was previously used for debug purpose.  Remove aio_run,
aio_wakeups, iocb->ki_queued and iocb->ki_kicked.  Also clean up unused
variable count in __aio_run_iocbs() and debug code in read_events().
Signed-off-by: NKen Chen <kenneth.w.chen@intel.com>
Cc: Benjamin LaHaise <bcrl@kvack.org>
Cc: Suparna Bhattacharya <suparna@in.ibm.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Since the tail pointer in aio_ring structure never wrap ring size more than
once, so a simple compare is sufficient to wrap the index around.  This avoid
a more expensive mod operation.
Signed-off-by: NKen Chen <kenneth.w.chen@intel.com>
Cc: Benjamin LaHaise <bcrl@kvack.org>
Cc: Suparna Bhattacharya <suparna@in.ibm.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

This patch removes superfluous kiocb member initialization in the AIO
allocation and deallocation path.  For example, in really_put_req(),
right before kiocb is returned to slab, 5 variables are reset to NULL.
The same variables will be initialized at the kiocb allocation time,
so why bother reset them knowing that they will be set to valid data
at alloc time?  Another example: ki_retry is initialized in __aio_get_req,
but is initialized again in io_submit_one.
Signed-off-by: NKen Chen <kenneth.w.chen@intel.com>
Cc: Benjamin LaHaise <bcrl@kvack.org>
Cc: Suparna Bhattacharya <suparna@in.ibm.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

I've noticed that, starting from linux-2.6.12-rc1, in the top Makefile the
"cmd_tags" variable has been changed in a way incompatible with *emacs
ctags.  Since the "--extra" option exists only in "exuberant ctags", it
should be included in the CTAGSF shell variable.

Cc: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

This patch reduces the stack usage of the function serial_event() in
serial_cs from 2212 to 228.  I used a patched version of gcc 3.4.3 on i386
with -fno-unit-at-a-time disabled.

This patch is only compile tested.
Acked-by: NRandy Dunlap <rddunlap@osdl.org>
Signed-off-by: NYum Rayan <yum.rayan@gmail.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Fix order of #include lines in mthca_memfree.c
Signed-off-by: NRoland Dreier <roland@topspin.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Convert most of the current code that uses _NSIG directly to instead use
valid_signal().  This avoids gcc -W warnings and off-by-one errors.
Signed-off-by: NJesper Juhl <juhl-lkml@dif.dk>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

This patch adds a new function valid_signal() that tests if its argument is
a valid signal number.

The reasons for adding this new function are:

- some code currently testing _NSIG directly has off-by-one errors.
  Using this function instead avoids such errors.

- some code currently tests unsigned signal numbers for <0 which is
  pointless and generates warnings when building with gcc -W.  Using this
  function instead avoids such warnings.

I considered various places to add this function but eventually settled on
include/linux/signal.h as the most logical place for it.  If there's some
reason this is a bad choice then please let me know (hints as to a better
location are then welcome of course).
Signed-off-by: NJesper Juhl <juhl-lkml@dif.dk>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

This patch adds the maintainers entry.
Signed-off-by: NKylene Hall <kjhall@us.ibm.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Signed-off-by: NStephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Signed-off-by: NAndreas Schwab <schwab@suse.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Correct an issue with the IPMI message layer taking a lock and calling
lower layer driver.  If an error occrues at the lower layer the lock can be
taken again causing a deadlock.  The lock is released before calling the
lower layer.
Signed-off-by: NDavid Griego <dgriego@mvista.com>
Signed-off-by: NCorey Minyard <minyard@acm.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Enable interrupts for a BT interface.  There is a specific register that
needs to be set up to enable interrupts that also must be modified to clear
the irq.

Also, don't reset the BMC on a BT interface.  That's probably not a good
idea as the BMC may be performing other important functions and a reset
should only be a last resort.  Also, that register is also used to
enable/disable interrupts to the BT; modifying it may screw up the
interrupts.
Signed-off-by: NCorey Minyard <minyard@acm.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

If there is an unexpected close, still allow the watchdog interface to be
re-opened on the IPMI watchdog.
Signed-off-by: NCorey Minyard <minyard@acm.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

If the ACPI register bit width is zero (an invalid value) assume it is the
default spacing.  This avoids some coredumps on invalid data and makes some
systems work that have broken ACPI data.
Signed-off-by: NCorey Minyard <minyard@acm.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Ignore the bottom bit of the base address from the DMI data.  It is
supposed to be set to 1 if it is I/O space.  Few systems do this, but this
enables the ones that do set it to work properly.
Signed-off-by: NCorey Minyard <minyard@acm.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Documentation/dontdiff is a little messy.  Here is a patch to sort the
content of that file in alphabetical
Signed-off-by: NNguyen Anh Quynh <aquynh@gmail.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

This fixes segmentation fault when specifying bad journal device via
a mount option.

Don't pass a zero pointer to bdevname() if filp_open() returns error.
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Add missing button codes for the Leadtek Winfast remote controls.
Signed-off-by: NNicolas Boichat <nicolas@boichat.ch>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

The variable attributes "packed" and "align" when used with struct, should
have the following order:

struct ... {...} __attribute__((packed)) var;

This patch fixes few instances where the variable and attributes are placed
the other way around and had no effect.
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

There were still a few comments left refering to verify_area, and two
functions, verify_area_skas & verify_area_tt that just wrap corresponding
access_ok_skas & access_ok_tt functions, just like verify_area does for
access_ok - deprecate those.

There was also a few places that still used verify_area in commented-out
code, fix those up to use access_ok.

After applying this one there should not be anything left but finally
removing verify_area completely, which will happen after a kernel release
or two.
Signed-off-by: NJesper Juhl <juhl-lkml@dif.dk>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Discussing with Matthew Wilcox some of his outstanding patches lead me to
this patch (among others).

The preamble in struct sigevent can be expressed independently of the
architecture.

Also use __ARCH_SI_PREAMBLE_SIZE on ia64.
Signed-off-by: NStephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

I recently realized that the in-kernel copy of hangcheck-timer was quite
stale.  Here's the latest.  It adds support for s390, ppc64, and ia64 too.
Signed-off-by: NJoel Becker <joel.becker@oracle.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

When allocating a new VC with vgacon_init(), the font is shared across all
the VGA consoles.  However, the font mask was always set to the default
value of zero in visual_init(), even if we were using 512 character fonts
at the time.

Moreover, code in vgacon.c:vga_do_font_op() didn't reset the mask if the
console driver thinks it's already in 512 character mode.  This means that
to *fix* it, you'd actually have to take the console out of 512 character
mode and then set it back.

The attached sets vc_hi_font_mask in vgacon_init() for any new consoles
opened if the vgacon driver is already in 512 character mode, solving this.

This bug goes back to 2.4.18 at least, probably earlier.
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Allow rewriting of a file and extending a file upto the end of the
allocated block on a full filesystem.

From: Chris Mason <mason@suse.com>
Signed-off-by: NJan Kara <jack@suse.cz>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

This patches adds the "nbds_max" parameter to the nbd kernel module, which
limits the number of nbds allocated.  Previously, always all 128 entries
were allocated unconditionally, which used to waste resources and
needlessly flood the hotplug system with events.  (Defaults to 16 now.)
Signed-off-by: NLars Marowsky-Bree <lmb@suse.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>