openwrt/staging/blogic.git
16 years agoUSB: add missing kerneldoc line for "needs_binding"
Alan Stern [Mon, 28 Jul 2008 14:39:28 +0000 (10:39 -0400)]
USB: add missing kerneldoc line for "needs_binding"

This patch (as1117) adds a kerneldoc line for the "needs_binding"
field in struct usb_interface.  It was accidentally omitted when the
field was added.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: ftdi_sio: add support for Luminance Stellaris Evaluation/Development Kits
Frederik Kriewitz [Wed, 30 Jul 2008 14:53:41 +0000 (16:53 +0200)]
USB: ftdi_sio: add support for Luminance Stellaris Evaluation/Development Kits

The Patch adds support for Luminance Stellaris Evaluation/Development
Kits (FTDI 2232C based).
The PIDs were missing.

Successfully tested with a Stellaris LM3S8962 Evaluation kit.

Signed-off-by: Frederik Kriewitz <frederik@kriewitz.eu>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: ftdi_sio: Add USB Product Id for ELV HS485
André Schenk [Mon, 28 Jul 2008 13:48:46 +0000 (15:48 +0200)]
USB: ftdi_sio: Add USB Product Id for ELV HS485

USB product id registration for the ELV HS485 USB adapter (www.elv.de) to
their home automation bus system. Applies to 2.6.26.

Signed-off-by: Andre Schenk <andre@melior.s.bawue.de>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: fix USB boot crash, ecm_do_notify(), list_add corruption. prev->next should...
David Brownell [Sat, 26 Jul 2008 15:06:24 +0000 (08:06 -0700)]
USB: fix USB boot crash, ecm_do_notify(), list_add corruption. prev->next should be next (ffff88003b8f82f8)

This fixes a BUG() turned up by Ingo via randconfig testing, where
CONFIG_LIST_DEBUG turned up list corruption.  The corruption was
caused by the dummy_hcd (single-machine test harness for gadget and
HCD code) trashing the request queue when driven by the new CDC
composite gadget an I/O pattern that was previously uncommon.
Fix suggested by Alan Stern.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Tested-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agousb: auerswald: remove driver (obsolete)
Wolfgang Mües [Tue, 29 Jul 2008 09:54:43 +0000 (11:54 +0200)]
usb: auerswald: remove driver (obsolete)

This patch removes the auerswald USB driver from the linux kernel
2.6.26.

This driver was included into the kernel mainly to connect to the ISDN
framework. This was done in linux 2.4.x. For 2.6.x, due to the fragile
and moving ISDN support, this connection was never realized, and the
only use of this driver was for device configuration. In the age of DSL,
the demand of ISDN support is getting very low.

Meanwhile, with the advent of libusb, an userspace driver was done for
the device configuration which works fine for linux and mac. (Thanks to
the libusb developers!). The userspace driver is downloadable from the
auerswald web site.

So this driver is obsolete now and has to be removed. Many thanks to all
developers which helped me to bring this driver up and working.

Signed-off-by: Wolfgang Muees <wolfgang@iksw-muees.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Linus Torvalds [Wed, 13 Aug 2008 22:24:35 +0000 (15:24 -0700)]
Merge git://git./linux/kernel/git/herbert/crypto-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  crypto: padlock - fix VIA PadLock instruction usage with irq_ts_save/restore()
  crypto: hash - Add missing top-level functions
  crypto: hash - Fix digest size check for digest type
  crypto: tcrypt - Fix AEAD chunk testing
  crypto: talitos - Add handling for SEC 3.x treatment of link table

16 years agoMerge git://oss.sgi.com:8090/xfs/linux-2.6
Linus Torvalds [Wed, 13 Aug 2008 22:17:49 +0000 (15:17 -0700)]
Merge git://oss.sgi.com:8090/xfs/linux-2.6

* git://oss.sgi.com:8090/xfs/linux-2.6: (45 commits)
  [XFS] Fix use after free in xfs_log_done().
  [XFS] Make xfs_bmap_*_count_leaves void.
  [XFS] Use KM_NOFS for debug trace buffers
  [XFS] use KM_MAYFAIL in xfs_mountfs
  [XFS] refactor xfs_mount_free
  [XFS] don't call xfs_freesb from xfs_unmountfs
  [XFS] xfs_unmountfs should return void
  [XFS] cleanup xfs_mountfs
  [XFS] move root inode IRELE into xfs_unmountfs
  [XFS] stop using file_update_time
  [XFS] optimize xfs_ichgtime
  [XFS] update timestamp in xfs_ialloc manually
  [XFS] remove the sema_t from XFS.
  [XFS] replace dquot flush semaphore with a completion
  [XFS] replace inode flush semaphore with a completion
  [XFS] extend completions to provide XFS object flush requirements
  [XFS] replace the XFS buf iodone semaphore with a completion
  [XFS] clean up stale references to semaphores
  [XFS] use get_unaligned_* helpers
  [XFS] Fix compile failure in xfs_buf_trace()
  ...

16 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/dlm
Linus Torvalds [Wed, 13 Aug 2008 22:16:10 +0000 (15:16 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/teigland/dlm

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/dlm:
  dlm: rename structs
  dlm: add missing kfrees

16 years ago[h8300] move include/asm-h8300 to arch/h8300/include/asm
Linus Torvalds [Wed, 13 Aug 2008 21:26:22 +0000 (14:26 -0700)]
[h8300] move include/asm-h8300 to arch/h8300/include/asm

Done as a script (well, a single "git mv" actually) on request from
Yoshinori Sato as a way to avoid a huge diff.

Requested-by: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agodlm: rename structs
David Teigland [Thu, 31 Jul 2008 14:31:53 +0000 (09:31 -0500)]
dlm: rename structs

Add a dlm_ prefix to the struct names in config.c.  This resolves a
conflict with struct node in particular, when include/linux/node.h
happens to be included.

Reported-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David Teigland <teigland@redhat.com>
16 years agodlm: add missing kfrees
David Teigland [Tue, 29 Jul 2008 20:21:19 +0000 (15:21 -0500)]
dlm: add missing kfrees

A couple of unlikely error conditions were missing a kfree on the error
exit path.

Reported-by: Juha Leppanen <juha_motorsportcom@luukku.com>
Signed-off-by: David Teigland <teigland@redhat.com>
16 years agocrypto: padlock - fix VIA PadLock instruction usage with irq_ts_save/restore()
Suresh Siddha [Wed, 13 Aug 2008 12:02:26 +0000 (22:02 +1000)]
crypto: padlock - fix VIA PadLock instruction usage with irq_ts_save/restore()

Wolfgang Walter reported this oops on his via C3 using padlock for
AES-encryption:

##################################################################

BUG: unable to handle kernel NULL pointer dereference at 000001f0
IP: [<c01028c5>] __switch_to+0x30/0x117
*pde = 00000000
Oops: 0002 [#1] PREEMPT
Modules linked in:

Pid: 2071, comm: sleep Not tainted (2.6.26 #11)
EIP: 0060:[<c01028c5>] EFLAGS: 00010002 CPU: 0
EIP is at __switch_to+0x30/0x117
EAX: 00000000 EBX: c0493300 ECX: dc48dd00 EDX: c0493300
ESI: dc48dd00 EDI: c0493530 EBP: c04cff8c ESP: c04cff7c
 DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 0068
Process sleep (pid: 2071, ti=c04ce000 task=dc48dd00 task.ti=d2fe6000)
Stack: dc48df30 c0493300 00000000 00000000 d2fe7f44 c03b5b43 c04cffc8 00000046
       c0131856 0000005a dc472d3c c0493300 c0493470 d983ae00 00002696 00000000
       c0239f54 00000000 c04c4000 c04cffd8 c01025fe c04f3740 00049800 c04cffe0
Call Trace:
 [<c03b5b43>] ? schedule+0x285/0x2ff
 [<c0131856>] ? pm_qos_requirement+0x3c/0x53
 [<c0239f54>] ? acpi_processor_idle+0x0/0x434
 [<c01025fe>] ? cpu_idle+0x73/0x7f
 [<c03a4dcd>] ? rest_init+0x61/0x63
 =======================

Wolfgang also found out that adding kernel_fpu_begin() and kernel_fpu_end()
around the padlock instructions fix the oops.

Suresh wrote:

These padlock instructions though don't use/touch SSE registers, but it behaves
similar to other SSE instructions. For example, it might cause DNA faults
when cr0.ts is set. While this is a spurious DNA trap, it might cause
oops with the recent fpu code changes.

This is the code sequence  that is probably causing this problem:

a) new app is getting exec'd and it is somewhere in between
   start_thread() and flush_old_exec() in the load_xyz_binary()

b) At pont "a", task's fpu state (like TS_USEDFPU, used_math() etc) is
   cleared.

c) Now we get an interrupt/softirq which starts using these encrypt/decrypt
   routines in the network stack. This generates a math fault (as
   cr0.ts is '1') which sets TS_USEDFPU and restores the math that is
   in the task's xstate.

d) Return to exec code path, which does start_thread() which does
   free_thread_xstate() and sets xstate pointer to NULL while
   the TS_USEDFPU is still set.

e) At the next context switch from the new exec'd task to another task,
   we have a scenarios where TS_USEDFPU is set but xstate pointer is null.
   This can cause an oops during unlazy_fpu() in __switch_to()

Now:

1) This should happen with or with out pre-emption. Viro also encountered
   similar problem with out CONFIG_PREEMPT.

2) kernel_fpu_begin() and kernel_fpu_end() will fix this problem, because
   kernel_fpu_begin() will manually do a clts() and won't run in to the
   situation of setting TS_USEDFPU in step "c" above.

3) This was working before the fpu changes, because its a spurious
   math fault  which doesn't corrupt any fpu/sse registers and the task's
   math state was always in an allocated state.

With out the recent lazy fpu allocation changes, while we don't see oops,
there is a possible race still present in older kernels(for example,
while kernel is using kernel_fpu_begin() in some optimized clear/copy
page and an interrupt/softirq happens which uses these padlock
instructions generating DNA fault).

This is the failing scenario that existed even before the lazy fpu allocation
changes:

0. CPU's TS flag is set

1. kernel using FPU in some optimized copy  routine and while doing
kernel_fpu_begin() takes an interrupt just before doing clts()

2. Takes an interrupt and ipsec uses padlock instruction. And we
take a DNA fault as TS flag is still set.

3. We handle the DNA fault and set TS_USEDFPU and clear cr0.ts

4. We complete the padlock routine

5. Go back to step-1, which resumes clts() in kernel_fpu_begin(), finishes
the optimized copy routine and does kernel_fpu_end(). At this point,
we have cr0.ts again set to '1' but the task's TS_USEFPU is stilll
set and not cleared.

6. Now kernel resumes its user operation. And at the next context
switch, kernel sees it has do a FP save as TS_USEDFPU is still set
and then will do a unlazy_fpu() in __switch_to(). unlazy_fpu()
will take a DNA fault, as cr0.ts is '1' and now, because we are
in __switch_to(), math_state_restore() will get confused and will
restore the next task's FP state and will save it in prev tasks's FP state.
Remember, in __switch_to() we are already on the stack of the next task
but take a DNA fault for the prev task.

This causes the fpu leakage.

Fix the padlock instruction usage by calling them inside the
context of new routines irq_ts_save/restore(), which clear/restore cr0.ts
manually in the interrupt context. This will not generate spurious DNA
in the  context of the interrupt which will fix the oops encountered and
the possible FPU leakage issue.

Reported-and-bisected-by: Wolfgang Walter <wolfgang.walter@stwm.de>
Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
16 years agocrypto: hash - Add missing top-level functions
Herbert Xu [Tue, 5 Aug 2008 05:34:30 +0000 (13:34 +0800)]
crypto: hash - Add missing top-level functions

The top-level functions init/update/final were missing for ahash.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
16 years agocrypto: hash - Fix digest size check for digest type
Herbert Xu [Sun, 3 Aug 2008 13:19:43 +0000 (21:19 +0800)]
crypto: hash - Fix digest size check for digest type

The changeset ca786dc738f4f583b57b1bba7a335b5e8233f4b0

crypto: hash - Fixed digest size check

missed one spot for the digest type.  This patch corrects that
error.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
16 years agocrypto: tcrypt - Fix AEAD chunk testing
Herbert Xu [Wed, 30 Jul 2008 08:23:51 +0000 (16:23 +0800)]
crypto: tcrypt - Fix AEAD chunk testing

My changeset 4b22f0ddb6564210c9ded7ba25b2a1007733e784

crypto: tcrpyt - Remove unnecessary kmap/kunmap calls

introduced a typo that broke AEAD chunk testing.  In particular,
axbuf should really be xbuf.

There is also an issue with testing the last segment when encrypting.
The additional part produced by AEAD wasn't tested.  Similarly, on
decryption the additional part of the AEAD input is mistaken for
corruption.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
16 years agocrypto: talitos - Add handling for SEC 3.x treatment of link table
Lee Nipper [Wed, 30 Jul 2008 08:26:57 +0000 (16:26 +0800)]
crypto: talitos - Add handling for SEC 3.x treatment of link table

Later SEC revision requires the link table (used for scatter/gather)
to have an extra entry to account for the total length in descriptor [4],
which contains cipher Input and ICV.
This only applies to decrypt, not encrypt.
Without this change, on 837x, a gather return/length error results
when a decryption uses a link table to gather the fragments.
This is observed by doing a ping with size of 1447 or larger with AES,
or a ping with size 1455 or larger with 3des.

So, add check for SEC compatible "fsl,3.0" for using extra link table entry.

Signed-off-by: Lee Nipper <lee.nipper@freescale.com>
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
16 years ago[XFS] Fix use after free in xfs_log_done().
Lachlan McIlroy [Wed, 13 Aug 2008 06:52:50 +0000 (16:52 +1000)]
[XFS] Fix use after free in xfs_log_done().

The ticket allocation code got reworked in 2.6.26 and we now free tickets
whereas before we used to cache them so the use-after-free went
undetected.

SGI-PV: 985525

SGI-Modid: xfs-linux-melb:xfs-kern:31877a

Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: David Chinner <david@fromorbit.com>
16 years ago[XFS] Make xfs_bmap_*_count_leaves void.
Ruben Porras [Wed, 13 Aug 2008 06:52:25 +0000 (16:52 +1000)]
[XFS] Make xfs_bmap_*_count_leaves void.

xfs_bmap_count_leaves and xfs_bmap_disk_count_leaves always return always
0, make them void.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31844a

Signed-off-by: Ruben Porras <ruben.porras@linworks.de>
Signed-off-by: Donald Douwsma <donaldd@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] Use KM_NOFS for debug trace buffers
Lachlan McIlroy [Wed, 13 Aug 2008 06:51:57 +0000 (16:51 +1000)]
[XFS] Use KM_NOFS for debug trace buffers

Use KM_NOFS to prevent recursion back into the filesystem which can cause
deadlocks.

In the case of xfs_iread() we hold the lock on the inode cluster buffer
while allocating memory for the trace buffers. If we recurse back into XFS
to flush data that may require a transaction to allocate extents which
needs log space. This can deadlock with the xfsaild thread which can't
push the tail of the log because it is trying to get the inode cluster
buffer lock.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31838a

Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: David Chinner <david@fromorbit.com>
16 years ago[XFS] use KM_MAYFAIL in xfs_mountfs
Christoph Hellwig [Wed, 13 Aug 2008 06:51:29 +0000 (16:51 +1000)]
[XFS] use KM_MAYFAIL in xfs_mountfs

Use KM_MAYFAIL for the m_perag allocation, we can deal with the error
easily and blocking forever during mount is not a good idea either.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31837a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] refactor xfs_mount_free
Christoph Hellwig [Wed, 13 Aug 2008 06:50:47 +0000 (16:50 +1000)]
[XFS] refactor xfs_mount_free

xfs_mount_free mostly frees the perag data, which is something that is
duplicated in the mount error path.

Move the XFS_QM_DONE call to the caller and remove the useless
mutex_destroy/spinlock_destroy calls so that we can re-use it for the
mount error path. Also rename it to xfs_free_perag to reflect what it
does.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31836a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] don't call xfs_freesb from xfs_unmountfs
Christoph Hellwig [Wed, 13 Aug 2008 06:50:21 +0000 (16:50 +1000)]
[XFS] don't call xfs_freesb from xfs_unmountfs

xfs_readsb is called before xfs_mount so xfs_freesb should be called after
xfs_unmountfs, too. This means it now happens after a few things during
the of xfs_unmount which all have nothing to do with the superblock.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31835a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] xfs_unmountfs should return void
Christoph Hellwig [Wed, 13 Aug 2008 06:49:57 +0000 (16:49 +1000)]
[XFS] xfs_unmountfs should return void

xfs_unmounts can't and shouldn't return errors so declare it as returning
void.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31833a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] cleanup xfs_mountfs
Christoph Hellwig [Wed, 13 Aug 2008 06:49:32 +0000 (16:49 +1000)]
[XFS] cleanup xfs_mountfs

Remove all the useless flags and code keyed off it in xfs_mountfs.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31831a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] move root inode IRELE into xfs_unmountfs
Christoph Hellwig [Wed, 13 Aug 2008 06:49:04 +0000 (16:49 +1000)]
[XFS] move root inode IRELE into xfs_unmountfs

The root inode is allocated in xfs_mountfs so it should be release in
xfs_unmountfs. For the unmount case that means we do it after the the
xfs_sync(mp, SYNC_WAIT | SYNC_CLOSE) in the forced shutdown case and the
dmapi unmount event. Note that both reference the rip variable which might
be freed by that time in case inode flushing has kicked in, so strictly
speaking this might count as a bug fix

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31830a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] stop using file_update_time
Christoph Hellwig [Wed, 13 Aug 2008 06:48:12 +0000 (16:48 +1000)]
[XFS] stop using file_update_time

xfs_ichtime updates the xfs_inode and Linux inode timestamps just fine, no
need to call file_update_time and then copy the values over to the XFS
inode. The only additional thing in file_update_time are checks not
applicable to the write path.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31829a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: David Chinner <david@fromorbit.com>
16 years ago[XFS] optimize xfs_ichgtime
Christoph Hellwig [Wed, 13 Aug 2008 06:45:13 +0000 (16:45 +1000)]
[XFS] optimize xfs_ichgtime

Port a little optmization from file_update_time to xfs_ichgtime, and only
update the timestamp and mark the inode dirty if the timestamp actually
changes in the timer tick resultion supported by the running kernel.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31827a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] update timestamp in xfs_ialloc manually
Christoph Hellwig [Wed, 13 Aug 2008 06:44:15 +0000 (16:44 +1000)]
[XFS] update timestamp in xfs_ialloc manually

In xfs_ialloc we just want to set all timestamps to the current time. We
don't need to mark the inode dirty like xfs_ichgtime does, and we don't
need nor want the opimizations in xfs_ichgtime that I will introduce in
the next patch.

So just opencode the timestamp update in xfs_ialloc, and remove the new
unused XFS_ICHGTIME_ACC case in xfs_ichgtime.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31825a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] remove the sema_t from XFS.
David Chinner [Wed, 13 Aug 2008 06:42:10 +0000 (16:42 +1000)]
[XFS] remove the sema_t from XFS.

Now that all users of the sema_t are gone from XFS we can finally kill it.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31823a

Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] replace dquot flush semaphore with a completion
David Chinner [Wed, 13 Aug 2008 06:41:43 +0000 (16:41 +1000)]
[XFS] replace dquot flush semaphore with a completion

Use the new completion flush code to implement the dquot flush lock.
Removes one of the final users of semaphores in the XFS code base.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31822a

Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] replace inode flush semaphore with a completion
David Chinner [Wed, 13 Aug 2008 06:41:16 +0000 (16:41 +1000)]
[XFS] replace inode flush semaphore with a completion

Use the new completion flush code to implement the inode flush lock.
Removes one of the final users of semaphores in the XFS code base.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31817a

Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] extend completions to provide XFS object flush requirements
David Chinner [Wed, 13 Aug 2008 06:40:43 +0000 (16:40 +1000)]
[XFS] extend completions to provide XFS object flush requirements

XFS object flushing doesn't quite match existing completion semantics.  It
mixed exclusive access with completion.  That is, we need to mark an object as
being flushed before flushing it to disk, and then block any other attempt to
flush it until the completion occurs.  We do this but adding an extra count to
the completion before we start using them.  However, we still need to
determine if there is a completion in progress, and allow no-blocking attempts
fo completions to decrement the count.

To do this we introduce:

int try_wait_for_completion(struct completion *x)
returns a failure status if done == 0, otherwise decrements done
to zero and returns a "started" status. This is provided
to allow counted completions to begin safely while holding
object locks in inverted order.

int completion_done(struct completion *x)
returns 1 if there is no waiter, 0 if there is a waiter
(i.e. a completion in progress).

This replaces the use of semaphores for providing this exclusion
and completion mechanism.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31816a

Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] replace the XFS buf iodone semaphore with a completion
David Chinner [Wed, 13 Aug 2008 06:36:11 +0000 (16:36 +1000)]
[XFS] replace the XFS buf iodone semaphore with a completion

The xfs_buf_t b_iodonesema is really just a semaphore that wants to be a
completion. Change it to a completion and remove the last user of the
sema_t from XFS.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31815a

Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] clean up stale references to semaphores
David Chinner [Wed, 13 Aug 2008 06:34:31 +0000 (16:34 +1000)]
[XFS] clean up stale references to semaphores

A lot of code has been converted away from semaphores, but there are still
comments that reference semaphore behaviour. The log code is the worst
offender. Update the comments to reflect what the code really does now.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31814a

Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] use get_unaligned_* helpers
Harvey Harrison [Wed, 13 Aug 2008 06:29:21 +0000 (16:29 +1000)]
[XFS] use get_unaligned_* helpers

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31813a

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] Fix compile failure in xfs_buf_trace()
Lachlan McIlroy [Wed, 13 Aug 2008 06:28:40 +0000 (16:28 +1000)]
[XFS] Fix compile failure in xfs_buf_trace()

SGI-PV: 957103

SGI-Modid: xfs-linux-melb:xfs-kern:31804a

Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] Use the same btree_cur union member for alloc and inobt trees.
Christoph Hellwig [Wed, 13 Aug 2008 06:25:27 +0000 (16:25 +1000)]
[XFS] Use the same btree_cur union member for alloc and inobt trees.

The alloc and inobt btree use the same agbp/agno pair in the btree_cur
union. Make them use the same bc_private.a union member so that code for
these two short form btree implementations can be shared.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31788a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Tim Shimmin <tes@sgi.com>
Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] small cleanups in xfs_btree.c
Christoph Hellwig [Wed, 13 Aug 2008 06:23:50 +0000 (16:23 +1000)]
[XFS] small cleanups in xfs_btree.c

Remove unneeded xfs_btree_get_block forward declaration. Move
xfs_btree_firstrec next to xfs_btree_lastrec.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31787a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Tim Shimmin <tes@sgi.com>
Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] sanitize xfs_initialize_vnode
Christoph Hellwig [Wed, 13 Aug 2008 06:23:13 +0000 (16:23 +1000)]
[XFS] sanitize xfs_initialize_vnode

Sanitize setting up the Linux indode.

Setting up the xfs_inode <-> inode link is opencoded in xfs_iget_core now
because that's the only place it needs to be done, xfs_initialize_vnode is
renamed to xfs_setup_inode and loses all superflous paramaters. The check
for I_NEW is removed because it always is true and the di_mode check moves
into xfs_iget_core because it's only needed there.

xfs_set_inodeops and xfs_revalidate_inode are merged into xfs_setup_inode
and the whole things is moved into xfs_iops.c where it belongs.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31782a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] kill bhv_vnode_t
Christoph Hellwig [Wed, 13 Aug 2008 06:22:40 +0000 (16:22 +1000)]
[XFS] kill bhv_vnode_t

All remaining bhv_vnode_t instance are in code that's more or less Linux
specific. (Well, for xfs_acl.c that could be argued, but that code is on
the removal list, too). So just do an s/bhv_vnode_t/struct inode/ over the
whole tree. We can clean up variable naming and some useless helpers
later.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31781a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] remove some easy bhv_vnode_t instances
Christoph Hellwig [Wed, 13 Aug 2008 06:22:09 +0000 (16:22 +1000)]
[XFS] remove some easy bhv_vnode_t instances

In various places we can just move a VFS_I call into the argument list of
called functions/macros instead of having a local bhv_vnode_t.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31776a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] kill xfs_lock_dir_and_entry
Christoph Hellwig [Wed, 13 Aug 2008 06:18:07 +0000 (16:18 +1000)]
[XFS] kill xfs_lock_dir_and_entry

When multiple inodes are locked in XFS it happens in order of the inode
number, with the everything but the first inode trylocked if any of the
previous inodes is in the AIL.

Except for the sorting of the inodes this logic is implemented in
xfs_lock_inodes, but also partially duplicated in xfs_lock_dir_and_entry
in a particularly stupid way adds a lock roundtrip if the inode ordering
is not optimal.

This patch adds a new helper xfs_lock_two_inodes that takes two inodes and
locks them in the most optimal way according to the above locking protocol
and uses it for all places that want to lock two inodes.

The only caller of xfs_lock_inodes is xfs_rename which might lock up to
four inodes.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31772a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Donald Douwsma <donaldd@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] kill INDUCE_IO_ERROR
Christoph Hellwig [Wed, 13 Aug 2008 06:17:37 +0000 (16:17 +1000)]
[XFS] kill INDUCE_IO_ERROR

All the error injection is already enabled through ifdef DEBUG, so kill
the never set second cpp symbol to activate it without the rest of the
debugging infrastructure.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31771a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] implement IHOLD/IRELE directly
Christoph Hellwig [Wed, 13 Aug 2008 06:13:45 +0000 (16:13 +1000)]
[XFS] implement IHOLD/IRELE directly

Now that all direct calls to VN_HOLD/VN_RELE are gone we can implement
IHOLD/IRELE directly.

For the IHOLD case also replace igrab with a direct increment of i_count
because we are guaranteed to already have a live and referenced inode by
the VFS. Also remove the vn_hold statistic because it's been rather
meaningless for some time with most references done by other callers.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31764a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] remove remaining VN_HOLD calls
Christoph Hellwig [Wed, 13 Aug 2008 06:13:09 +0000 (16:13 +1000)]
[XFS] remove remaining VN_HOLD calls

Use IHOLD(ip) instead of VN_HOLD(VFS_I(ip)).

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31765a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] remove spurious VN_HOLD/VN_RELE calls from xfs_acl.c
Christoph Hellwig [Wed, 13 Aug 2008 06:12:37 +0000 (16:12 +1000)]
[XFS] remove spurious VN_HOLD/VN_RELE calls from xfs_acl.c

All the ACL routines are called from inode operations which are guaranteed
to have a referenced inode by the VFS, so there's no need for the ACL code
to grab another temporary one.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31763a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] kill vn_to_inode
Christoph Hellwig [Wed, 13 Aug 2008 06:12:05 +0000 (16:12 +1000)]
[XFS] kill vn_to_inode

bhv_vnode_t is just a typedef for struct inode, so there's
no need for a helper to convert between the two.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31761a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] Remove vn_from_inode()
Christoph Hellwig [Wed, 13 Aug 2008 06:11:26 +0000 (16:11 +1000)]
[XFS] Remove vn_from_inode()

bhv_vnode_t is just a typedef for struct inode, so there's
no need for a helper to convert between the two.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31760a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] remove shouting-indirection macros from xfs_trans.h
Eric Sandeen [Wed, 13 Aug 2008 06:10:52 +0000 (16:10 +1000)]
[XFS] remove shouting-indirection macros from xfs_trans.h

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31758a

Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] convert xfs to use ERR_CAST
Eric Sandeen [Wed, 13 Aug 2008 06:09:25 +0000 (16:09 +1000)]
[XFS] convert xfs to use ERR_CAST

Looks like somehow xfs got missed in the conversion that took place in
e231c2ee64eb1c5cd3c63c31da9dac7d888dcf7f, "Convert ERR_PTR(PTR_ERR(p))
instances to ERR_CAST(p)
<http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit
diff;h=e231c2ee64eb1c5cd3c63c31da9dac7d888dcf7f>"

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31757a

Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Acked-by: David Howells <dhowells@redhat.com>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] remove INT_GET and friends
Eric Sandeen [Wed, 13 Aug 2008 06:07:53 +0000 (16:07 +1000)]
[XFS] remove INT_GET and friends

Thanks to hch's endian work, INT_GET etc are no longer used, and may as
well be removed. INT_SET is still used in the acl code, though.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31756a

Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] Move xfs_attr_rolltrans to xfs_trans_roll
Niv Sardi [Wed, 13 Aug 2008 06:05:49 +0000 (16:05 +1000)]
[XFS] Move xfs_attr_rolltrans to xfs_trans_roll

Move it from the attr code to the transaction code and make
the attr code call the new function.

We rolltrans is really usefull whenever we want to use rolling
transaction, should be generic, it isn't dependent on any part
of the attr code anyway.

We use this excuse to change all the:

if ((error = xfs_attr_rolltrans()))

calls into:

error = xfs_trans_roll();

if (error)

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31729a

Signed-off-by: Niv Sardi <xaiki@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] don't leak m_fsname/m_rtname/m_logname
Christoph Hellwig [Wed, 13 Aug 2008 06:04:05 +0000 (16:04 +1000)]
[XFS] don't leak m_fsname/m_rtname/m_logname

Add a helper to free the m_fsname/m_rtname/m_logname allocations and use
it properly for all mount failure cases. Also switch the allocations for
these to kstrdup while we're at it.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31728a

Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] Move attr log alloc size calculator to another function.
Niv Sardi [Wed, 13 Aug 2008 06:03:35 +0000 (16:03 +1000)]
[XFS] Move attr log alloc size calculator to another function.

We will need that to be able to calculate the size of log we need for a
specific attr (for Create+EA). The local flag is needed so that we can
fail if we run into ENOSPC when trying to alloc blocks.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31727a

Signed-off-by: Niv Sardi <xaiki@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Tim Shimmin <tes@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] Use KM_NOFS for incore inode extent tree allocation V2
David Chinner [Wed, 13 Aug 2008 06:02:51 +0000 (16:02 +1000)]
[XFS] Use KM_NOFS for incore inode extent tree allocation V2

If we allow incore extent tree allocations to recurse into the
filesystem under memory pressure, new delayed allocations through
xfs_iomap_write_delay() can deadlock on themselves if memory
reclaim tries to write back dirty pages from that inode.

It will deadlock in xfs_iomap_write_allocate() trying to take the
ilock we already hold. This can also show up as complex ABBA deadlocks
when multiple threads are triggering memory reclaim when trying to
allocate extents.

The main cause of this is the fact that delayed allocation is not done in
a transaction, so KM_NOFS is not automatically added to the allocations to
prevent this recursion.

Mark all allocations done for the incore inode extent tree as KM_NOFS to
ensure they never recurse back into the filesystem.

Version 2: o KM_NOFS implies KM_SLEEP, so just use KM_NOFS

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31726a

Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] XFS: Kill xfs_vtoi()
David Chinner [Wed, 13 Aug 2008 06:01:45 +0000 (16:01 +1000)]
[XFS] XFS: Kill xfs_vtoi()

xfs_vtoi() is redundant and only unsed in small sections of code.
Replace them with widely used XFS_I() inline and kill xfs_vtoi().

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31725a

Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] Kill shouty XFS_ITOV() macro
David Chinner [Wed, 13 Aug 2008 06:00:45 +0000 (16:00 +1000)]
[XFS] Kill shouty XFS_ITOV() macro

Replace XFS_ITOV() with the new VFS_I() inline.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31724a

Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] kill shouty XFS_ITOV_NULL macro
David Chinner [Wed, 13 Aug 2008 05:47:43 +0000 (15:47 +1000)]
[XFS] kill shouty XFS_ITOV_NULL macro

Replace XFS_ITOV_NULL() with the new VFS_I() inline.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31722a

Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] Avoid directly referencing the VFS inode.
David Chinner [Wed, 13 Aug 2008 05:45:15 +0000 (15:45 +1000)]
[XFS] Avoid directly referencing the VFS inode.

In several places we directly convert from the XFS inode
to the linux (VFS) inode by a simple deference of ip->i_vnode.
We should not do this - a helper function should be used to
extract the VFS inode from the XFS inode.

Introduce the function VFS_I() to extract the VFS inode
from the XFS inode. The name was chosen to match XFS_I() which
is used to extract the XFS inode from the VFS inode.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31720a

Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Niv Sardi <xaiki@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years ago[XFS] Do not access buffers after dropping reference count
Lachlan McIlroy [Wed, 13 Aug 2008 05:42:10 +0000 (15:42 +1000)]
[XFS] Do not access buffers after dropping reference count

We should not access a buffer after dropping it's reference count
otherwise we could race with another thread that releases the final
reference count and frees the buffer causing us to access potentially
unmapped memory. The bug this change fixes only occured on DEBUG XFS since
the offending code was in an ASSERT.

SGI-PV: 984429

SGI-Modid: xfs-linux-melb:xfs-kern:31715a

Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
Signed-off-by: David Chinner <david@fromorbit.com>
16 years ago[XFS] Use the generic bitops rather than implementing them ourselves.
David Chinner [Wed, 13 Aug 2008 05:41:12 +0000 (15:41 +1000)]
[XFS] Use the generic bitops rather than implementing them ourselves.

This keeps xfs_lowbit64 as it was since there aren't good generic helpers
there ... Patch inspired by Andi Kleen.

SGI-PV: 981498

SGI-Modid: xfs-linux-melb:xfs-kern:31472a

Signed-off-by: David Chinner <david@fromorbit.com>
Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Donald Douwsma <donaldd@sgi.com>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
16 years agoLinux 2.6.27-rc3
Linus Torvalds [Wed, 13 Aug 2008 01:55:39 +0000 (18:55 -0700)]
Linux 2.6.27-rc3

16 years agoMerge branch 'for-2.6.27' of git://linux-nfs.org/~bfields/linux
Linus Torvalds [Tue, 12 Aug 2008 23:39:22 +0000 (16:39 -0700)]
Merge branch 'for-2.6.27' of git://linux-nfs.org/~bfields/linux

* 'for-2.6.27' of git://linux-nfs.org/~bfields/linux:
  fs/nfsd/export.c: Adjust error handling code involving auth_domain_put
  MAINTAINERS: mention lockd and sunrpc in nfs entries
  lockd: trivial sparse endian annotations

16 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland...
Linus Torvalds [Tue, 12 Aug 2008 23:38:45 +0000 (16:38 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/roland/infiniband

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  IB/ehca: Discard double CQE for one WR
  IB/ehca: Check idr_find() return value
  IB/ehca: Repoll CQ on invalid opcode
  IB/ehca: Rename goto label in ehca_poll_cq_one()
  IB/ehca: Update qp_state on cached modify_qp()
  IPoIB/cm: Use vmalloc() to allocate rx_rings

16 years agoMerge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6
Linus Torvalds [Tue, 12 Aug 2008 23:07:48 +0000 (16:07 -0700)]
Merge branch 'release' of git://git./linux/kernel/git/aegl/linux-2.6

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6:
  [IA64] use bcd2bin/bin2bcd
  [IA64] Ensure cpu0 can access per-cpu variables in early boot code

16 years agofirmware/memmap: cleanup
Bernhard Walle [Tue, 12 Aug 2008 22:09:14 +0000 (15:09 -0700)]
firmware/memmap: cleanup

Various cleanup the drivers/firmware/memmap (after review by AKPM):

    - fix kdoc to conform to the standard
    - move kdoc from header to implementation files
    - remove superfluous WARN_ON() after kmalloc()
    - WARN_ON(x); if (!x) -> if(!WARN_ON(x))
    - improve some comments

Signed-off-by: Bernhard Walle <bwalle@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoALi M7101 PMU also available on Sun Netra's too
Alexander Clouter [Tue, 12 Aug 2008 22:09:12 +0000 (15:09 -0700)]
ALi M7101 PMU also available on Sun Netra's too

My Sun Netra T1 AC200 has one of these... bit harsh not letting me use it
and all :)

==========
alex@woodchuck:~$ lspci -nn
00:01.0 PCI bridge [0604]: Sun Microsystems Computer Corp. Simba Advanced PCI Bridge [108e:5000] (rev 13)
00:01.1 PCI bridge [0604]: Sun Microsystems Computer Corp. Simba Advanced PCI Bridge [108e:5000] (rev 13)
01:03.0 Non-VGA unclassified device [0000]: ALi Corporation M7101 Power Management Controller [PMU] [10b9:7101]
01:05.1 Ethernet controller [0200]: Sun Microsystems Computer Corp. RIO GEM [108e:1101] (rev 01)
01:05.3 USB Controller [0c03]: Sun Microsystems Computer Corp. RIO USB [108e:1103] (rev 01)
01:07.0 ISA bridge [0601]: ALi Corporation M1533/M1535 PCI to ISA Bridge [Aladdin IV/V/V+] [10b9:1533]
01:0c.0 Bridge [0680]: Sun Microsystems Computer Corp. RIO EBUS [108e:1100] (rev 01)
01:0c.1 Ethernet controller [0200]: Sun Microsystems Computer Corp. RIO GEM [108e:1101] (rev 01)
01:0c.3 USB Controller [0c03]: Sun Microsystems Computer Corp. RIO USB [108e:1103] (rev 01)
01:0d.0 IDE interface [0101]: ALi Corporation M5229 IDE [10b9:5229] (rev c3)
02:08.0 SCSI storage controller [0100]: LSI Logic / Symbios Logic 53C896/897 [1000:000b] (rev 07)
02:08.1 SCSI storage controller [0100]: LSI Logic / Symbios Logic 53C896/897 [1000:000b] (rev 07)
==========

Signed-off-by: Alexander Clouter <alex@digriz.org.uk>
Cc: Wim Van Sebroeck <wim@iguana.be>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoMake ioctl.h compatible with userland
Michael Abbott [Tue, 12 Aug 2008 22:09:11 +0000 (15:09 -0700)]
Make ioctl.h compatible with userland

The attached patch seems to already exist in a number of branches -- it
keeps popping up on Google for me, and is certainly already in Debian --
but is strangely absent from mainstream.

The problem appears to be that the patched file ends up as part of the
target toolchain, but unfortunately the gcc constant folding doesn't
appear to eliminate the __invalid_size_argument_for_IOC value early
enough.  Certainly compiling C++ programs which use _IO...  macros as
constants fails without this patch.

Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agodocsrc: fix getdelays printk formats
Randy Dunlap [Tue, 12 Aug 2008 22:09:10 +0000 (15:09 -0700)]
docsrc: fix getdelays printk formats

Fix printf format type warnings (seen on alpha & ia64):

Documentation/accounting/getdelays.c:206: warning: format '%15llu' expects type 'long long unsigned int', but argument 6 has type '__u64'
Documentation/accounting/getdelays.c:206: warning: format '%15llu' expects type 'long long unsigned int', but argument 7 has type '__u64'
Documentation/accounting/getdelays.c:206: warning: format '%15llu' expects type 'long long unsigned int', but argument 8 has type '__u64'
Documentation/accounting/getdelays.c:206: warning: format '%15llu' expects type 'long long unsigned int', but argument 9 has type '__u64'
Documentation/accounting/getdelays.c:206: warning: format '%15llu' expects type 'long long unsigned int', but argument 12 has type '__u64'
Documentation/accounting/getdelays.c:206: warning: format '%15llu' expects type 'long long unsigned int', but argument 13 has type '__u64'
Documentation/accounting/getdelays.c:206: warning: format '%15llu' expects type 'long long unsigned int', but argument 16 has type '__u64'
Documentation/accounting/getdelays.c:206: warning: format '%15llu' expects type 'long long unsigned int', but argument 17 has type '__u64'
Documentation/accounting/getdelays.c:214: warning: format '%15llu' expects type 'long long unsigned int', but argument 4 has type '__u64'
Documentation/accounting/getdelays.c:214: warning: format '%15llu' expects type 'long long unsigned int', but argument 5 has type '__u64'
Documentation/accounting/getdelays.c:221: warning: format '%llu' expects type 'long long unsigned int', but argument 2 has type '__u64'
Documentation/accounting/getdelays.c:221: warning: format '%llu' expects type 'long long unsigned int', but argument 3 has type '__u64'
Documentation/accounting/getdelays.c:221: warning: format '%llu' expects type 'long long unsigned int', but argument 4 has type '__u64'
Documentation/accounting/getdelays.c:221: warning: format '%llu' expects type 'long long unsigned int', but argument 5 has type '__u64'
Documentation/accounting/getdelays.c:221: warning: format '%llu' expects type 'long long unsigned int', but argument 6 has type '__u64'
Documentation/accounting/getdelays.c:236: warning: 'cmd_type' may be used uninitialized in this function

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agodocsrc: fix crc32hash type
Randy Dunlap [Tue, 12 Aug 2008 22:09:08 +0000 (15:09 -0700)]
docsrc: fix crc32hash type

Fix differing signedness warning:

Documentation/pcmcia/crc32hash.c:29: warning: pointer targets in passing argument 1 of 'crc32' differ in signedness

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Acked-by: Dominik Brodowski <linux@dominikbrodowski.net>
Cc: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agodocsrc: fix ifenslave type
Randy Dunlap [Tue, 12 Aug 2008 22:09:08 +0000 (15:09 -0700)]
docsrc: fix ifenslave type

Documentation/networking/ifenslave.c:1084: warning: pointer targets in assignment differ in signedness

>From include/linux/socket.h:
 * 1003.1g requires sa_family_t and that sa_data is char.

and from SUSv3:
(http://www.opengroup.org/onlinepubs/009695399/basedefs/sys/socket.h.html)

The <sys/socket.h> header shall define the sockaddr structure that includes at least the following members:

sa_family_t  sa_family  Address family.
char         sa_data[]  Socket address (variable-length data).
<end SUSv3>

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agodocsrc: fix procfs example
Randy Dunlap [Tue, 12 Aug 2008 22:09:06 +0000 (15:09 -0700)]
docsrc: fix procfs example

Add MODULE_LICENSE() to DocBook/procfs_example.c since modpost complained
about a missing license there.

Remove tty procfs removal since the creation was deleted long ago
(http://git.kernel.org/?p=linux/kernel/git/tglx/history.git;a=commitdiff;h=5ad9cb65e9b15e5b83e2dd1c10a4bcaccc4ec644).

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Cc: <J.A.K.Mouw@its.tudelft.nl>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agodocsrc: build Documentation/ sources
Randy Dunlap [Tue, 12 Aug 2008 22:09:06 +0000 (15:09 -0700)]
docsrc: build Documentation/ sources

Currently source files in the Documentation/ sub-dir can easily bit-rot
since they are not generally buildable, either because they are hidden in
text files or because there are no Makefile rules for them.  This needs to
be fixed so that the source files remain usable and good examples of code
instead of bad examples.

Add the ability to build source files that are in the Documentation/ dir.
Add to Kconfig as "BUILD_DOCSRC" config symbol.

Use "CONFIG_BUILD_DOCSRC=1 make ..." to build objects from the
Documentation/ sources.  Or enable BUILD_DOCSRC in the *config system.
However, this symbol depends on HEADERS_CHECK since the header files need
to be installed (for userspace builds).

Built (using cross-tools) for x86-64, i386, alpha, ia64, sparc32,
sparc64, powerpc, sh, m68k, & mips.

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agobyteorder: add include/linux/byteorder.h to define endian helpers
Harvey Harrison [Tue, 12 Aug 2008 22:09:05 +0000 (15:09 -0700)]
byteorder: add include/linux/byteorder.h to define endian helpers

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agobyteorder: add a new include/linux/swab.h to define byteswapping functions
Harvey Harrison [Tue, 12 Aug 2008 22:09:04 +0000 (15:09 -0700)]
byteorder: add a new include/linux/swab.h to define byteswapping functions

Collect the implementations from include/linux/byteorder/swab.h, swabb.h
in swab.h

The functionality provided covers:
u16 swab16(u16 val) - return a byteswapped 16 bit value
u32 swab32(u32 val) - return a byteswapped 32 bit value
u64 swab64(u64 val) - return a byteswapped 64 bit value
u32 swahw32(u32 val) - return a wordswapped 32 bit value
u32 swahb32(u32 val) - return a high/low byteswapped 32 bit value

Similar to above, but return swapped value from a naturally-aligned pointer
u16 swab16p(u16 *p)
u32 swab32p(u32 *p)
u64 swab64p(u64 *p)
u32 swahw32p(u32 *p)
u32 swahb32p(u32 *p)

Similar to above, but swap the value in-place (in-situ)
void swab16s(u16 *p)
void swab32s(u32 *p)
void swab64s(u64 *p)
void swahw32s(u32 *p)
void swahb32s(u32 *p)

Arches can override any of these with an optimized version by defining an
inline in their asm/byteorder.h (example given for swab16()):

u16 __arch_swab16() {}
 #define __arch_swab16 __arch_swab16

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agogenirq: switch /proc/irq/*/smp_affinity et al to seqfiles
Alexey Dobriyan [Tue, 12 Aug 2008 22:09:03 +0000 (15:09 -0700)]
genirq: switch /proc/irq/*/smp_affinity et al to seqfiles

Switch /proc/irq/*/smp_affinity , /proc/irq/default_smp_affinity to
seq_files.

cat(1) reads with 1024 chunks by default, with high enough NR_CPUS, there
will be -EINVAL.

As side effect, there are now two less users of the ->read_proc interface.

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Paul Jackson <pj@sgi.com>
Cc: Mike Travis <travis@sgi.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoseq_file: add seq_cpumask(), seq_nodemask()
Alexey Dobriyan [Tue, 12 Aug 2008 22:09:02 +0000 (15:09 -0700)]
seq_file: add seq_cpumask(), seq_nodemask()

Short enough reads from /proc/irq/*/smp_affinity return -EINVAL for no
good reason.

This became noticed with NR_CPUS=4096 patches, when length of printed
representation of cpumask becase 1152, but cat(1) continued to read with
1024-byte chunks.  bitmap_scnprintf() in good faith fills buffer, returns
1023, check returns -EINVAL.

Fix it by switching to seq_file, so handler will just fill buffer and
doesn't care about offsets, length, filling EOF and all this crap.

For that add seq_bitmap(), and wrappers around it -- seq_cpumask() and
seq_nodemask().

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Reviewed-by: Paul Jackson <pj@sgi.com>
Cc: Mike Travis <travis@sgi.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoreiserfs: removed duplicated #include
Huang Weiyi [Tue, 12 Aug 2008 22:09:01 +0000 (15:09 -0700)]
reiserfs: removed duplicated #include

Removed duplicated #include <linux/quotaops.h> in
fs/reiserfs/super.c.

Signed-off-by: Huang Weiyi <weiyi.huang@gmail.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agofs/eventpoll.c: fix sys_epoll_create1() comment
Andrew Morton [Tue, 12 Aug 2008 22:09:01 +0000 (15:09 -0700)]
fs/eventpoll.c: fix sys_epoll_create1() comment

The `size' argument was removed.

Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Ulrich Drepper <drepper@redhat.com>
Cc: Davide Libenzi <davidel@xmailserver.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agodrivers/char/rtc.c: removed duplicated include
Huang Weiyi [Tue, 12 Aug 2008 22:09:00 +0000 (15:09 -0700)]
drivers/char/rtc.c: removed duplicated include

Signed-off-by: Huang Weiyi <weiyi.huang@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agomm/sparse.c: removed duplicated include
Huang Weiyi [Tue, 12 Aug 2008 22:09:00 +0000 (15:09 -0700)]
mm/sparse.c: removed duplicated include

Signed-off-by: Huang Weiyi <weiyi.huang@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agomove kernel-doc comment for might_sleep directly before its defining block
Uwe Kleine-König [Tue, 12 Aug 2008 22:08:59 +0000 (15:08 -0700)]
move kernel-doc comment for might_sleep directly before its defining block

Signed-off-by: Uwe Kleine-König <Uwe.Kleine-Koenig@digi.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agolib/vsprintf.c: wrong conversion function used
Yi Yang [Tue, 12 Aug 2008 22:08:58 +0000 (15:08 -0700)]
lib/vsprintf.c: wrong conversion function used

Fix wrong conversion function used by strict_strtou*

Signed-off-by: Yi Yang <yi.y.yang@intel.com>
Reported-by: Swen Schillig <swen@vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agofbcon: prevent cursor disappearance after switching to 512 character font
Krzysztof Helt [Tue, 12 Aug 2008 22:08:57 +0000 (15:08 -0700)]
fbcon: prevent cursor disappearance after switching to 512 character font

Adjust and honor the vc_scrl_erase_char for 256 and 512 character fonts.

It fixes the issue with disappearing cursor during scrolling
(http://bugzilla.kernel.org/show_bug.cgi?id=11258).  The issue was
reported and tracked by Peter Hanzel.

Signed-off-by: Krzysztof Helt <krzysztof.h1@wp.pl>
Reported-by: Peter Hanzel <hanzelpeter@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoatmel_lcdfb: add board parameter specify framebuffer memory size
Haavard Skinnemoen [Tue, 12 Aug 2008 22:08:57 +0000 (15:08 -0700)]
atmel_lcdfb: add board parameter specify framebuffer memory size

Specify how much physically continuous, DMA capable memory will be
allocated at driver initialization time.  This allow to create framebuffer
device with larger virtual resolution.  Combine with y-panning this can be
used to implement double buffering acceleration method.

Signed-off-by: Stanislaw Gruszka <stf_xl@wp.pl>
Acked-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
Acked-by: Krzysztof Helt <krzysztof.h1@wp.pl>
Cc: Nicolas Ferre <nicolas.ferre@atmel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoatmel_lcdfb: set ypanstep to 1 and enable y-panning on AT91
Haavard Skinnemoen [Tue, 12 Aug 2008 22:08:56 +0000 (15:08 -0700)]
atmel_lcdfb: set ypanstep to 1 and enable y-panning on AT91

Panning in the y-direction can be done by simply changing the DMA base
address.  This code is already in place, but FBIOPAN_DISPLAY will
currently fail because ypanstep is 0.

Set ypanstep to 1 to indicate that we do support y-panning and also set
the necessary acceleration flags on AT91 (AVR32 already have them.)

Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com>
Acked-by: Krzysztof Helt <krzysztof.h1@wp.pl>
Cc: Nicolas Ferre <nicolas.ferre@atmel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agomatrox maven: convert to a new-style i2c driver
Jean Delvare [Tue, 12 Aug 2008 22:08:55 +0000 (15:08 -0700)]
matrox maven: convert to a new-style i2c driver

The legacy i2c model is going away soon, so switch to the new model.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Acked-by: Krzysztof Helt <krzysztof.h1@wp.pl>
Cc: Petr Vandrovec <VANDROVE@vc.cvut.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agomatroxfb: i2c structure templates clean-up
Jean Delvare [Tue, 12 Aug 2008 22:08:54 +0000 (15:08 -0700)]
matroxfb: i2c structure templates clean-up

Clean up the use of structure templates in i2c-matroxfb. In this case
it's more efficient to initialize the few fields we need individually.
This makes i2c-matroxfb.ko 16% smaller on my system.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Acked-by: Krzysztof Helt <krzysztof.h1@wp.pl>
Cc: Petr Vandrovec <VANDROVE@vc.cvut.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agomatrox maven: fix a broken error path
Jean Delvare [Tue, 12 Aug 2008 22:08:53 +0000 (15:08 -0700)]
matrox maven: fix a broken error path

I broke an error path with d03c21ec0be7787ff6b75dcf56c0e96209ccbfbd,
sorry about that.

The machine will crash if the i2c_attach_client() or maven_init_client()
calls fail, although nobody has yet reported this happening.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Acked-by: Krzysztof Helt <krzysztof.h1@wp.pl>
Cc: Petr Vandrovec <VANDROVE@vc.cvut.cz>
Cc: <stable@kernel.org> [2.6.25.x, 2.6.26.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agodo_migrate_pages(): remove unused variable
MinChan Kim [Tue, 12 Aug 2008 22:08:52 +0000 (15:08 -0700)]
do_migrate_pages(): remove unused variable

Signed-off-by: MinChan Kim <minchan.kim@gmail.com>
Acked-by: Christoph Lameter <cl@linux-foundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoGRU: fix preprocessor symbol for sparse
Randy Dunlap [Tue, 12 Aug 2008 22:08:51 +0000 (15:08 -0700)]
GRU: fix preprocessor symbol for sparse

Fix preprocessor symbol so that sparse sees it and does not generate
errors:

  drivers/misc/sgi-gru/grutables.h:286:2: error: "Unsupported architecture"
  drivers/misc/sgi-gru/grutables.h:286:2: error: "Unsupported architecture"
  drivers/misc/sgi-gru/grutables.h:286:2: error: "Unsupported architecture"
  drivers/misc/sgi-gru/grutables.h:286:2: error: "Unsupported architecture"
  drivers/misc/sgi-gru/grutlbpurge.c:185:11: error: undefined identifier 'GRUREGION'
  drivers/misc/sgi-gru/grutables.h:286:2: error: "Unsupported architecture"
  drivers/misc/sgi-gru/grutables.h:286:2: error: "Unsupported architecture"

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Cc: Jack Steiner <steiner@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoradeonfb: fix accel engine hangs
David Miller [Tue, 12 Aug 2008 22:08:51 +0000 (15:08 -0700)]
radeonfb: fix accel engine hangs

Some chips appear to have the 2D engine hang during screen redraw,
typically in a sequence of copyarea operations. This appear to be
solved by adding a flush of the engine destination pixel cache
and waiting for the engine to be idle before issuing the accel
operation. The performance impact seems to be fairly small.

Here is a trace on an RV370 (PCI device ID 0x5b64), it records the
RBBM_STATUS register, then the source x/y, destination x/y, and
width/height used for the copy:

----------------------------------------
radeonfb_prim_copyarea: STATUS[00000140] src[210:70] dst[210:60] wh[a0:10]
radeonfb_prim_copyarea: STATUS[00000140] src[2b8:70] dst[2b8:60] wh[88:10]
radeonfb_prim_copyarea: STATUS[00000140] src[348:70] dst[348:60] wh[40:10]
radeonfb_prim_copyarea: STATUS[80020140] src[390:70] dst[390:60] wh[88:10]
radeonfb_prim_copyarea: STATUS[8002613f] src[40:80] dst[40:70] wh[28:10]
radeonfb_prim_copyarea: STATUS[80026139] src[a8:80] dst[a8:70] wh[38:10]
radeonfb_prim_copyarea: STATUS[80026133] src[e8:80] dst[e8:70] wh[80:10]
radeonfb_prim_copyarea: STATUS[8002612d] src[170:80] dst[170:70] wh[30:10]
radeonfb_prim_copyarea: STATUS[80026127] src[1a8:80] dst[1a8:70] wh[8:10]
radeonfb_prim_copyarea: STATUS[80026121] src[1b8:80] dst[1b8:70] wh[88:10]
radeonfb_prim_copyarea: STATUS[8002611b] src[248:80] dst[248:70] wh[68:10]
----------------------------------------

When things are going fine the copies complete before the next ROP is
even issued, but all of a sudden the 2D unit becomes active (bit 17 in
RBBM_STATUS) and the FIFO retry (bit 13) and FIFO pipeline busy (bit
14) are set as well.  The FIFO begins to backup until it becomes full.

What happens next is the radeon_fifo_wait() times out, and we access
the chip illegally leading to a bus error which usually wedges the
box.  None of this makes it to the console screen, of course :-)
radeon_fifo_wait() should be modified to reset the accelerator when
this timeout happens instead of programming the chip anyways.

----------------------------------------
radeonfb: FIFO Timeout !
ERROR(0): Cheetah error trap taken afsr[0010080005000000] afar[000007f900800e40] TL1(0)
ERROR(0): TPC[595114] TNPC[595118] O7[459788] TSTATE[11009601]
ERROR(0): TPC<radeonfb_copyarea+0xfc/0x248>
ERROR(0): M_SYND(0),  E_SYND(0), Privileged
ERROR(0): Highest priority error (0000080000000000) "Bus error response from system bus"
ERROR(0): D-cache idx[0] tag[0000000000000000] utag[0000000000000000] stag[0000000000000000]
ERROR(0): D-cache data0[0000000000000000] data1[0000000000000000] data2[0000000000000000] data3[0000000000000000]
ERROR(0): I-cache idx[0] tag[0000000000000000] utag[0000000000000000] stag[0000000000000000] u[0000000000000000] l[00\

ERROR(0): I-cache INSN0[0000000000000000] INSN1[0000000000000000] INSN2[0000000000000000] INSN3[0000000000000000]
ERROR(0): I-cache INSN4[0000000000000000] INSN5[0000000000000000] INSN6[0000000000000000] INSN7[0000000000000000]
ERROR(0): E-cache idx[800e40] tag[000000000e049f4c]
ERROR(0): E-cache data0[fffff8127d300180] data1[00000000004b5384] data2[0000000000000000] data3[0000000000000000]
Ker:xnel panic - not syncing: Irrecoverable deferred error trap.
----------------------------------------

Another quirk is that these copyarea calls will not happen until the
first drivers/char/vt.c:redraw_screen() occurs.  This will only happen
if you 1) VC switch or 2) run "consolechars" or 3) unblank the screen.

This seems to happen because until a redraw_screen() the screen scrolling
method used by fbcon is not finalized yet.  I've seen this with other fb
drivers too.

So if all you do is boot straight into X you will never see this bug on
the relevant chips.

Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: <stable@kernel.org> [2.6.25.x, 2.6.26.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoallocate structures for reservation tracking in hugetlbfs outside of spinlocks v2
Andy Whitcroft [Tue, 12 Aug 2008 22:08:49 +0000 (15:08 -0700)]
allocate structures for reservation tracking in hugetlbfs outside of spinlocks v2

[Andrew this should replace the previous version which did not check
the returns from the region prepare for errors.  This has been tested by
us and Gerald and it looks good.

Bah, while reviewing the locking based on your previous email I spotted
that we need to check the return from the vma_needs_reservation call for
allocation errors.  Here is an updated patch to correct this.  This passes
testing here.]

Signed-off-by: Andy Whitcroft <apw@shadowen.org>
Tested-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agohugetlbfs: allocate structures for reservation tracking outside of spinlocks
Andy Whitcroft [Tue, 12 Aug 2008 22:08:47 +0000 (15:08 -0700)]
hugetlbfs: allocate structures for reservation tracking outside of spinlocks

In the normal case, hugetlbfs reserves hugepages at map time so that the
pages exist for future faults.  A struct file_region is used to track when
reservations have been consumed and where.  These file_regions are
allocated as necessary with kmalloc() which can sleep with the
mm->page_table_lock held.  This is wrong and triggers may-sleep warning
when PREEMPT is enabled.

Updates to the underlying file_region are done in two phases.  The first
phase prepares the region for the change, allocating any necessary memory,
without actually making the change.  The second phase actually commits the
change.  This patch makes use of this by checking the reservations before
the page_table_lock is taken; triggering any necessary allocations.  This
may then be safely repeated within the locks without any allocations being
required.

Credit to Mel Gorman for diagnosing this failure and initial versions of
the patch.

Signed-off-by: Andy Whitcroft <apw@shadowen.org>
Tested-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoisight_firmware: fix a leak and double kfree()
Parag Warudkar [Tue, 12 Aug 2008 22:08:46 +0000 (15:08 -0700)]
isight_firmware: fix a leak and double kfree()

Signed-off-by: Parag Warudkar <parag.warudkar@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agocpuidle: make sysfs attributes sysdev class attributes
Rabin Vincent [Tue, 12 Aug 2008 22:08:45 +0000 (15:08 -0700)]
cpuidle: make sysfs attributes sysdev class attributes

These attributes are really sysdev class attributes.  The incorrect
definition leads to an oops because of recent changes which make sysdev
attributes use a different prototype.

Based on Andi's f718cd4add5aea9d379faff92f162571e356cc5f ("sched: make
scheduler sysfs attributes sysdev class devices")

Reported-by: Eric Sesterhenn <snakebyte@gmx.de>
Signed-off-by: Rabin Vincent <rabin@rab.in>
Acked-by: Andi Kleen <ak@linux.intel.com>
Cc: "Li, Shaohua" <shaohua.li@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agortc-isl1208: fix double removal of a sysfs entry
Alessandro Zummo [Tue, 12 Aug 2008 22:08:44 +0000 (15:08 -0700)]
rtc-isl1208: fix double removal of a sysfs entry

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Cc: Herbert Valerio Riedel <hvr@gnu.org>
Cc: Hartley Sweeten <hartleys@visionengravers.com>
Cc: David Brownell <david-b@pacbell.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agoh8300: fix section mismatches
Yoshinori Sato [Tue, 12 Aug 2008 22:08:43 +0000 (15:08 -0700)]
h8300: fix section mismatches

WARNING: vmlinux.o(.text+0x2fdf): Section mismatch in reference from the variable .LM3 to the variable .init.text:___alloc_bootmem
The function .LM3() references
the variable __init ___alloc_bootmem.
This is often because .LM3 lacks a __init
annotation or the annotation of ___alloc_bootmem is wrong.

WARNING: vmlinux.o(.text+0x2ff5): Section mismatch in reference from the variable .LM4 to the variable .init.text:___alloc_bootmem
The function .LM4() references
the variable __init ___alloc_bootmem.
This is often because .LM4 lacks a __init
annotation or the annotation of ___alloc_bootmem is wrong.

WARNING: vmlinux.o(.text+0x300b): Section mismatch in reference from the variable .LM5 to the variable .init.text:___alloc_bootmem
The function .LM5() references
the variable __init ___alloc_bootmem.
This is often because .LM5 lacks a __init
annotation or the annotation of ___alloc_bootmem is wrong.

WARNING: vmlinux.o(.text+0x304b): Section mismatch in reference from the variable .LM10 to the variable .init.text:_free_area_init
The function .LM10() references
the variable __init _free_area_init.
This is often because .LM10 lacks a __init
annotation or the annotation of _free_area_init is wrong.

WARNING: vmlinux.o(.text+0x30a3): Section mismatch in reference from the variable .LM17 to the variable .init.text:_free_all_bootmem
The function .LM17() references
the variable __init _free_all_bootmem.
This is often because .LM17 lacks a __init
annotation or the annotation of _free_all_bootmem is wrong.

Signed-off-by: Yoshinori Sato <ysato@users.sourceforge.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agorevert "rtc: cdev lock_kernel() pushdown"
David Brownell [Tue, 12 Aug 2008 22:08:41 +0000 (15:08 -0700)]
revert "rtc: cdev lock_kernel() pushdown"

Revert commit 51a776fa7a7997e726d4a478eda0854c6f9143bd ("rtc: cdev
lock_kernel() pushdown").  The RTC framework does not need BKL
protection.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Alessandro Zummo <alessandro.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
16 years agomemcg: fix oops in mem_cgroup_shrink_usage
Hugh Dickins [Tue, 12 Aug 2008 22:08:41 +0000 (15:08 -0700)]
memcg: fix oops in mem_cgroup_shrink_usage

Got an oops in mem_cgroup_shrink_usage() when testing loop over tmpfs:
yes, of course, loop0 has no mm: other entry points check but this didn't.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>