safe/jmp/linux-2.6
14 years agoxfs: Don't flush stale inodes
Dave Chinner [Sat, 2 Jan 2010 02:39:40 +0000 (02:39 +0000)]
xfs: Don't flush stale inodes

Because inodes remain in cache much longer than inode buffers do
under memory pressure, we can get the situation where we have
stale, dirty inodes being reclaimed but the backing storage has
been freed.  Hence we should never, ever flush XFS_ISTALE inodes
to disk as there is no guarantee that the backing buffer is in
cache and still marked stale when the flush occurs.

Signed-off-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years agoxfs: fix timestamp handling in xfs_setattr
Christoph Hellwig [Wed, 23 Dec 2009 16:09:13 +0000 (16:09 +0000)]
xfs: fix timestamp handling in xfs_setattr

We currently have some rather odd code in xfs_setattr for
updating the a/c/mtime timestamps:

 - first we do a non-transaction update if all three are updated
   together
 - second we implicitly update the ctime for various changes
   instead of relying on the ATTR_CTIME flag
 - third we set the timestamps to the current time instead of the
   arguments in the iattr structure in many cases.

This patch makes sure we update it in a consistent way:

 - always transactional
 - ctime is only updated if ATTR_CTIME is set or we do a size
   update, which is a special case
 - always to the times passed in from the caller instead of the
   current time

The only non-size caller of xfs_setattr that doesn't come from
the VFS is updated to set ATTR_CTIME and pass in a valid ctime
value.

Reported-by: Eric Blake <ebb9@byu.net>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years agoxfs: use DECLARE_EVENT_CLASS
Christoph Hellwig [Mon, 21 Dec 2009 14:03:03 +0000 (14:03 +0000)]
xfs: use DECLARE_EVENT_CLASS

Using DECLARE_EVENT_CLASS allows us to to use trace event code
instead of duplicating it in the binary.  This was not available
before 2.6.33 so it had to be done as a separate step once the
prerequisite was merged.

This only requires changes to xfs_trace.h and the results are
rather impressive:

hch@brick:~/work/linux-2.6/obj-kvm$ size fs/xfs/xfs.o*
text    data     bss     dec     hex filename
 607732   41884    3616  653232   9f7b0 fs/xfs/xfs.o
1026732   41884    3808 1072424  105d28 fs/xfs/xfs.o.old

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland...
Linus Torvalds [Fri, 8 Jan 2010 22:05:28 +0000 (14:05 -0800)]
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/addr: Correct CONFIG_IPv6 to CONFIG_IPV6
  mlx4_core: Fix cleanup in __mlx4_init_one() error path
  IB/mlx4: Fix queue overflow check in post_recv
  IB/mlx4: Initialize SRQ scatter entries when creating an SRQ

14 years agoMerge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jwessel...
Linus Torvalds [Fri, 8 Jan 2010 22:04:20 +0000 (14:04 -0800)]
Merge branch 'for_linus' of git://git./linux/kernel/git/jwessel/linux-2.6-kgdb

* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jwessel/linux-2.6-kgdb:
  kgdb: Fix kernel-doc format error in kgdb.h
  blackfin,kgdb: Do not put PC in gdb_regs into retx.
  blackfin,kgdb,probe_kernel: Cleanup probe_kernel_read/write
  maccess,probe_kernel: Allow arch specific override probe_kernel_(read|write)

14 years agoMerge branch 'reiserfs/kill-bkl' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 8 Jan 2010 22:03:55 +0000 (14:03 -0800)]
Merge branch 'reiserfs/kill-bkl' of git://git./linux/kernel/git/frederic/random-tracing

* 'reiserfs/kill-bkl' of git://git.kernel.org/pub/scm/linux/kernel/git/frederic/random-tracing:
  reiserfs: Relax reiserfs_xattr_set_handle() while acquiring xattr locks
  reiserfs: Fix unreachable statement
  reiserfs: Don't call reiserfs_get_acl() with the reiserfs lock
  reiserfs: Relax lock on xattr removing
  reiserfs: Relax the lock before truncating pages
  reiserfs: Fix recursive lock on lchown
  reiserfs: Fix mistake in down_write() conversion

14 years agoMerge branch 'for-linus' of git://oss.sgi.com/xfs/xfs
Linus Torvalds [Fri, 8 Jan 2010 21:57:32 +0000 (13:57 -0800)]
Merge branch 'for-linus' of git://oss.sgi.com/xfs/xfs

* 'for-linus' of git://oss.sgi.com/xfs/xfs:
  xfs: kill some warnings on i386 builds

14 years agoMerge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6
Linus Torvalds [Fri, 8 Jan 2010 21:57:19 +0000 (13:57 -0800)]
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] move fnptr definition inside #ifdef __KERNEL__

14 years agoMerge branch 'for-linus/samsung' of git://git.fluff.org/bjdooks/linux
Linus Torvalds [Fri, 8 Jan 2010 21:56:31 +0000 (13:56 -0800)]
Merge branch 'for-linus/samsung' of git://git.fluff.org/bjdooks/linux

* 'for-linus/samsung' of git://git.fluff.org/bjdooks/linux:
  ARM: S3C64XX: Fix possible clock look in EPLL and MPLL clock chains

14 years agoMerge branch 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 8 Jan 2010 21:55:52 +0000 (13:55 -0800)]
Merge branch 'x86-fixes-for-linus' of git://git./linux/kernel/git/x86/linux-2.6-tip

* 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/x86/linux-2.6-tip:
  x86, irq: Check move_in_progress before freeing the vector mapping
  x86: copy_from_user() should not return -EFAULT
  Revert "x86: Side-step lguest problem by only building cmpxchg8b_emu for pre-Pentium"
  x86/pci: Intel ioh bus num reg accessing fix
  x86: Fix size for ex trampoline with 32bit

14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes...
Linus Torvalds [Fri, 8 Jan 2010 21:55:39 +0000 (13:55 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/jbarnes/pci-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes/pci-2.6:
  PCIe AER: prevent AER injection if hardware masks error reporting
  PCI/PM: Use per-device D3 delays
  PCI: Check the node argument passed to cpumask_of_node
  PCI: AER: fix aer inject result in kernel oops
  PCI: pcie portdrv: style cleanup

14 years agoMerge branch 'bugfixes' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6
Linus Torvalds [Fri, 8 Jan 2010 21:55:14 +0000 (13:55 -0800)]
Merge branch 'bugfixes' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6

* 'bugfixes' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6:
  nfs: fix oops in nfs_rename()
  sunrpc: fix build-time warning
  sunrpc: on successful gss error pipe write, don't return error
  SUNRPC: Fix the return value in gss_import_sec_context()
  SUNRPC: Fix up an error return value in gss_import_sec_context_kerberos()

14 years agoxfs: kill some warnings on i386 builds
Dave Chinner [Thu, 17 Dec 2009 00:20:07 +0000 (00:20 +0000)]
xfs: kill some warnings on i386 builds

Randy Dunlap Reported printk() format-related warnings reported
on i386 builds in his environment.  Dave Chinner provided this
patch to eliminate them.

Signed-off by: Dave Chinner <david@fromorbit.com>
Acked-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
14 years ago[IA64] move fnptr definition inside #ifdef __KERNEL__
Tony Luck [Fri, 8 Jan 2010 18:53:28 +0000 (10:53 -0800)]
[IA64] move fnptr definition inside #ifdef __KERNEL__

Linus pointed out that this definition should not be
exported to user space.

Signed-off-by: Tony Luck <tony.luck@intel.com>
14 years agoMerge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux...
Linus Torvalds [Fri, 8 Jan 2010 17:32:50 +0000 (09:32 -0800)]
Merge branch 'release' of git://git./linux/kernel/git/lenb/linux-acpi-2.6

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6:
  hp-wmi: remove double free caused by merge conflict

14 years agoMerge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6
Linus Torvalds [Fri, 8 Jan 2010 17:32:15 +0000 (09:32 -0800)]
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] __per_cpu_idtrs[] is a memory hog
  [IA64] sanity in #include files.  Move fnptr to types.h
  [IA64] use helpers for rlimits
  [IA64] cpumask_of_node() should handle -1 as a node

14 years ago[IA64] __per_cpu_idtrs[] is a memory hog
Tony Luck [Fri, 8 Jan 2010 00:10:57 +0000 (16:10 -0800)]
[IA64] __per_cpu_idtrs[] is a memory hog

__per_cpu_idtrs is statically allocated ... on CONFIG_NR_CPUS=4096
systems it hogs 16MB of memory. This is way too much for a quite
probably unused facility (only KVM uses dynamic TR registers).

Change to an array of pointers, and allocate entries as needed on
a per cpu basis.  Change the name too as the __per_cpu_ prefix is
confusing (this isn't a classic <linux/percpu.h> type object).

Signed-off-by: Tony Luck <tony.luck@intel.com>
14 years agokgdb: Fix kernel-doc format error in kgdb.h
Randy Dunlap [Thu, 7 Jan 2010 17:58:37 +0000 (11:58 -0600)]
kgdb: Fix kernel-doc format error in kgdb.h

linux-next-20081022//include/linux/kgdb.h:308): duplicate section name 'Description'

and fix typos in that file's kernel-doc comments.

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
14 years agoblackfin,kgdb: Do not put PC in gdb_regs into retx.
Sonic Zhang [Thu, 7 Jan 2010 17:58:37 +0000 (11:58 -0600)]
blackfin,kgdb: Do not put PC in gdb_regs into retx.

In blackfin, kgdb is running in delayed exception IRQ5 other than in
exception IRQ3 directly.  Register reti other than retx in pt_regs is
the kgdb return address. So, don't put PC in gdb_regs into retx.

CC: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Sonic Zhang <sonic.adi@gmail.com>
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
14 years agoblackfin,kgdb,probe_kernel: Cleanup probe_kernel_read/write
Jason Wessel [Thu, 7 Jan 2010 17:58:36 +0000 (11:58 -0600)]
blackfin,kgdb,probe_kernel: Cleanup probe_kernel_read/write

Blackfin needs it own arch specific probe_kernel_read() and
probe_kernel_write().

This was moved out of the kgdb code and into the
arch/blackfin/maccess.c, because it is a generic kernel api.

The arch specific kgdb.c for blackfin was cleaned of all functions
which exist in the kgdb core that do the same thing after resolving
the probe_kernel_read() and probe_kernel_write().  This also
eliminated the need for most of the #include's.

CC: Sonic Zhang <sonic.adi@gmail.com>
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
14 years agomaccess,probe_kernel: Allow arch specific override probe_kernel_(read|write)
Jason Wessel [Thu, 7 Jan 2010 17:58:36 +0000 (11:58 -0600)]
maccess,probe_kernel: Allow arch specific override probe_kernel_(read|write)

Some archs such as blackfin, would like to have an arch specific
probe_kernel_read() and probe_kernel_write() implementation which can
fall back to the generic implementation if no special operations are
needed.

CC: Thomas Gleixner <tglx@linutronix.de>
CC: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
14 years agoreiserfs: Relax reiserfs_xattr_set_handle() while acquiring xattr locks
Frederic Weisbecker [Thu, 7 Jan 2010 14:55:31 +0000 (15:55 +0100)]
reiserfs: Relax reiserfs_xattr_set_handle() while acquiring xattr locks

Fix remaining xattr locks acquired in reiserfs_xattr_set_handle()
while we are holding the reiserfs lock to avoid lock inversions.

Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Christian Kujau <lists@nerdbynature.de>
Cc: Alexander Beregalov <a.beregalov@gmail.com>
Cc: Chris Mason <chris.mason@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
14 years agoreiserfs: Fix unreachable statement
Jiri Slaby [Wed, 6 Jan 2010 22:09:50 +0000 (23:09 +0100)]
reiserfs: Fix unreachable statement

Stanse found an unreachable statement in reiserfs_ioctl. There is a
if followed by error assignment and `break' with no braces. Add the
braces so that we don't break every time, but only in error case,
so that REISERFS_IOC_SETVERSION actually works when it returns no
error.

Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Cc: Reiserfs <reiserfs-devel@vger.kernel.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
14 years agoreiserfs: Don't call reiserfs_get_acl() with the reiserfs lock
Frederic Weisbecker [Thu, 7 Jan 2010 11:57:47 +0000 (12:57 +0100)]
reiserfs: Don't call reiserfs_get_acl() with the reiserfs lock

reiserfs_get_acl is usually not called under the reiserfs lock,
as it doesn't need it. But it happens when it is called by
reiserfs_acl_chmod(), which creates a dependency inversion against
the private xattr inodes mutexes for the given inode.

We need to call it without the reiserfs lock, especially since
it's unnecessary.

Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Christian Kujau <lists@nerdbynature.de>
Cc: Alexander Beregalov <a.beregalov@gmail.com>
Cc: Chris Mason <chris.mason@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
14 years agoMerge branch 'drm-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied...
Linus Torvalds [Thu, 7 Jan 2010 04:26:42 +0000 (20:26 -0800)]
Merge branch 'drm-linus' of git://git./linux/kernel/git/airlied/drm-2.6

* 'drm-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
  drm/radeon/kms: rs600: use correct mask for SW interrupt
  gpu/drm/radeon/radeon_irq.c: move a dereference below a NULL test
  drm/radeon/radeon_device.c: move a dereference below a NULL test
  drm/radeon/radeon_fence.c: move a dereference below the NULL test
  drm/radeon/radeon_connectors.c: add a NULL test before dereference
  drm/radeon/kms: fix memory leak
  drm/kms: Fix &&/|| confusion in drm_fb_helper_connector_parse_command_line()
  drm/edid: Fix CVT width/height decode
  drm/edid: Skip empty CVT codepoints
  drm: remove address mask param for drm_pci_alloc()
  drm/radeon/kms: add missing breaks in i2c and ss lookups
  drm/radeon/kms: add primary dac adj values table
  drm/radeon/kms: fallback to default connector table

14 years agoMerge remote branch 'korg/drm-radeon-next' into drm-linus
Dave Airlie [Thu, 7 Jan 2010 04:00:29 +0000 (14:00 +1000)]
Merge remote branch 'korg/drm-radeon-next' into drm-linus

* korg/drm-radeon-next:
  drm/radeon/kms: rs600: use correct mask for SW interrupt
  gpu/drm/radeon/radeon_irq.c: move a dereference below a NULL test
  drm/radeon/radeon_device.c: move a dereference below a NULL test
  drm/radeon/radeon_fence.c: move a dereference below the NULL test
  drm/radeon/radeon_connectors.c: add a NULL test before dereference
  drm/radeon/kms: fix memory leak
  drm/radeon/kms: add missing breaks in i2c and ss lookups
  drm/radeon/kms: add primary dac adj values table
  drm/radeon/kms: fallback to default connector table

14 years agodrm/radeon/kms: rs600: use correct mask for SW interrupt
Luca Tettamanti [Mon, 28 Dec 2009 21:53:05 +0000 (22:53 +0100)]
drm/radeon/kms: rs600: use correct mask for SW interrupt

The mask happens to be the same, but the IH is reading the status, not the
not the control register.

Signed-off-by: Luca Tettamanti <kronos.it@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agogpu/drm/radeon/radeon_irq.c: move a dereference below a NULL test
Darren Jenkins [Wed, 30 Dec 2009 01:16:35 +0000 (12:16 +1100)]
gpu/drm/radeon/radeon_irq.c: move a dereference below a NULL test

If a NULL value is possible, the dereference should only occur after the
NULL test.

Coverity CID: 13338

Signed-off-by: Darren Jenkins <darrenrjenkins@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/radeon_device.c: move a dereference below a NULL test
Darren Jenkins [Wed, 30 Dec 2009 01:18:30 +0000 (12:18 +1100)]
drm/radeon/radeon_device.c: move a dereference below a NULL test

If a NULL value is possible, the dereference should only occur after the
NULL test.

Coverity CID: 13335

Signed-off-by: Darren Jenkins <darrenrjenkins@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/radeon_fence.c: move a dereference below the NULL test
Darren Jenkins [Wed, 30 Dec 2009 01:20:05 +0000 (12:20 +1100)]
drm/radeon/radeon_fence.c: move a dereference below the NULL test

If a NULL value is possible, the dereference should only occur after the
NULL test.

Coverity CID: 13334

Signed-off-by: Darren Jenkins <darrenrjenkins@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/radeon_connectors.c: add a NULL test before dereference
Darren Jenkins [Wed, 30 Dec 2009 01:22:55 +0000 (12:22 +1100)]
drm/radeon/radeon_connectors.c: add a NULL test before dereference

The encoder variable can be NULL in this function so I believe it should
be checked before dereference.

Coverity CID: 13253

[airlied: extremely unlikely to happen]

Signed-off-by: Darren Jenkins <darrenrjenkins@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms: fix memory leak
Jiri Slaby [Wed, 6 Jan 2010 16:39:31 +0000 (17:39 +0100)]
drm/radeon/kms: fix memory leak

Stanse found a memory leak in radeon_master_create. master_priv is not
freed/assigned on all paths. Fix that.

Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agoMerge branch 'drm-core-next' into drm-linus
Dave Airlie [Thu, 7 Jan 2010 03:36:00 +0000 (13:36 +1000)]
Merge branch 'drm-core-next' into drm-linus

* drm-core-next:
  drm/kms: Fix &&/|| confusion in drm_fb_helper_connector_parse_command_line()
  drm/edid: Fix CVT width/height decode
  drm/edid: Skip empty CVT codepoints
  drm: remove address mask param for drm_pci_alloc()

14 years agodrm/kms: Fix &&/|| confusion in drm_fb_helper_connector_parse_command_line()
Roel Kluin [Thu, 31 Dec 2009 12:06:29 +0000 (13:06 +0100)]
drm/kms: Fix &&/|| confusion in drm_fb_helper_connector_parse_command_line()

This always evaluates to true.

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/edid: Fix CVT width/height decode
Adam Jackson [Mon, 4 Jan 2010 22:53:07 +0000 (17:53 -0500)]
drm/edid: Fix CVT width/height decode

Signed-off-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/edid: Skip empty CVT codepoints
Adam Jackson [Mon, 4 Jan 2010 22:53:06 +0000 (17:53 -0500)]
drm/edid: Skip empty CVT codepoints

Signed-off-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm: remove address mask param for drm_pci_alloc()
Zhenyu Wang [Tue, 5 Jan 2010 03:25:05 +0000 (11:25 +0800)]
drm: remove address mask param for drm_pci_alloc()

drm_pci_alloc() has input of address mask for setting pci dma
mask on the device, which should be properly setup by drm driver.
And leave it as a param for drm_pci_alloc() would cause confusion
or mistake would corrupt the correct dma mask setting, as seen on
intel hw which set wrong dma mask for hw status page. So remove
it from drm_pci_alloc() function.

Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agoARM: S3C64XX: Fix possible clock look in EPLL and MPLL clock chains
Ben Dooks [Thu, 7 Jan 2010 02:05:55 +0000 (11:05 +0900)]
ARM: S3C64XX: Fix possible clock look in EPLL and MPLL clock chains

There is a possibility of a loop happening in the PLL output clock
chain on the S3C64XX series. clk_mpll's parent was set to be
clk_mout_mpll, but this is fed from clk_fout_epll (which is also
clk_mpll).

clk_mpll is meant to be the output from the MPLL, and clk_mout_mpll
is a seperate clock derived from the mux of clk_mpll and clk_fin_mpll
and thus should be considered a seperate clock.

Anything using clk_mpll directly really should not be relying on this
being the clock that is eventually routed to a peripheral, so remove the
loop and ensure that the clocks accurately represent the clock chain
in the device.

The clk_mpll is not being used outside of the s3c6400-clock.c code, so
this change should not break anything else.

Do the same for the EPLL.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
14 years agoMerge branch 'drm-intel-next' of git://git.kernel.org/pub/scm/linux/kernel/git/anholt...
Linus Torvalds [Thu, 7 Jan 2010 02:16:17 +0000 (18:16 -0800)]
Merge branch 'drm-intel-next' of git://git./linux/kernel/git/anholt/drm-intel

* 'drm-intel-next' of git://git.kernel.org/pub/scm/linux/kernel/git/anholt/drm-intel: (23 commits)
  drm/i915: remove full registers dump debug
  drm/i915: Add DP dpll limit on ironlake and use existing DPLL search function
  drm/i915: Select the correct BPC for LVDS on Ironlake
  drm/i915: Make the BPC in FDI rx/transcoder be consistent with that in pipeconf on Ironlake
  drm/i915: Enable/disable the dithering for LVDS based on VBT setting
  drm/i915: Permit pinning whilst the device is 'suspended'
  drm/i915: Hold struct mutex whilst pinning power context bo.
  drm/i915: fix unused var
  drm/i915: Storage class should be before const qualifier
  drm/i915: remove render reclock support
  drm/i915: Fix RC6 suspend/resume
  drm/i915: execbuf2 support
  drm/i915: Reload hangcheck timer too for Ironlake
  drm/i915: only enable hotplug for detected outputs
  drm/i915: Track whether cursor needs physical address in intel_device_info
  drm/i915: Implement IS_* macros using static tables
  drm/i915: Move PCI IDs into i915 driver
  drm/i915: Update LVDS connector status when receiving ACPI LID event
  drm/i915: Add MALATA PC-81005 to ACPI LID quirk list
  drm/i915: implement new pm ops for i915
  ...

14 years agoNOMMU: Use copy_*_user_page() in access_process_vm()
Jie Zhang [Wed, 6 Jan 2010 17:23:28 +0000 (17:23 +0000)]
NOMMU: Use copy_*_user_page() in access_process_vm()

The MMU code uses the copy_*_user_page() variants in access_process_vm()
rather than copy_*_user() as the former includes an icache flush.  This
is important when doing things like setting software breakpoints with
gdb.  So switch the NOMMU code over to do the same.

This patch makes the reasonable assumption that copy_from_user_page()
won't fail - which is probably fine, as we've checked the VMA from which
we're copying is usable, and the copy is not allowed to cross VMAs.  The
one case where it might go wrong is if the VMA is a device rather than
RAM, and that device returns an error which - in which case rubbish will
be returned rather than EIO.

Signed-off-by: Jie Zhang <jie.zhang@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: David McCullough <david_mccullough@mcafee.com>
Acked-by: Paul Mundt <lethal@linux-sh.org>
Acked-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoNOMMU: Avoiding duplicate icache flushes of shared maps
Mike Frysinger [Wed, 6 Jan 2010 17:23:23 +0000 (17:23 +0000)]
NOMMU: Avoiding duplicate icache flushes of shared maps

When working with FDPIC, there are many shared mappings of read-only
code regions between applications (the C library, applet packages like
busybox, etc.), but the current do_mmap_pgoff() function will issue an
icache flush whenever a VMA is added to an MM instead of only doing it
when the map is initially created.

The flush can instead be done when a region is first mmapped PROT_EXEC.
Note that we may not rely on the first mapping of a region being
executable - it's possible for it to be PROT_READ only, so we have to
remember whether we've flushed the region or not, and then flush the
entire region when a bit of it is made executable.

However, this also affects the brk area.  That will no longer be
executable.  We can mprotect() it to PROT_EXEC on MPU-mode kernels, but
for NOMMU mode kernels, when it increases the brk allocation, making
sys_brk() flush the extra from the icache should suffice.  The brk area
probably isn't used by NOMMU programs since the brk area can only use up
the leavings from the stack allocation, where the stack allocation is
larger than requested.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoFDPIC: Respect PT_GNU_STACK exec protection markings when creating NOMMU stack
Mike Frysinger [Wed, 6 Jan 2010 17:23:17 +0000 (17:23 +0000)]
FDPIC: Respect PT_GNU_STACK exec protection markings when creating NOMMU stack

The current code will load the stack size and protection markings, but
then only use the markings in the MMU code path.  The NOMMU code path
always passes PROT_EXEC to the mmap() call.  While this doesn't matter
to most people whilst the code is running, it will cause a pointless
icache flush when starting every FDPIC application.  Typically this
icache flush will be of a region on the order of 128KB in size, or may
be the entire icache, depending on the facilities available on the CPU.

In the case where the arch default behaviour seems to be desired
(EXSTACK_DEFAULT), we probe VM_STACK_FLAGS for VM_EXEC to determine
whether we should be setting PROT_EXEC or not.

For arches that support an MPU (Memory Protection Unit - an MMU without
the virtual mapping capability), setting PROT_EXEC or not will make an
important difference.

It should be noted that this change also affects the executability of
the brk region, since ELF-FDPIC has that share with the stack.  However,
this is probably irrelevant as NOMMU programs aren't likely to use the
brk region, preferring instead allocation via mmap().

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge branch 'for-2.6.33' of git://linux-nfs.org/~bfields/linux
Linus Torvalds [Thu, 7 Jan 2010 02:10:15 +0000 (18:10 -0800)]
Merge branch 'for-2.6.33' of git://linux-nfs.org/~bfields/linux

* 'for-2.6.33' of git://linux-nfs.org/~bfields/linux:
  sunrpc: fix peername failed on closed listener
  nfsd: make sure data is on disk before calling ->fsync
  nfsd: fix "insecure" export option

14 years ago[IA64] sanity in #include files. Move fnptr to types.h
Tony Luck [Wed, 6 Jan 2010 23:52:35 +0000 (15:52 -0800)]
[IA64] sanity in #include files.  Move fnptr to types.h

Signed-off-by: Tony Luck <tony.luck@intel.com>
14 years ago[IA64] use helpers for rlimits
Jiri Slaby [Wed, 6 Jan 2010 15:24:30 +0000 (16:24 +0100)]
[IA64] use helpers for rlimits

Make sure compiler won't do weird things with limits. E.g. fetching
them twice may return 2 different values after writable limits are
implemented.

I.e. either use rlimit helpers added in
3e10e716abf3c71bdb5d86b8f507f9e72236c9cd
or ACCESS_ONCE if not applicable.

Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
14 years agonfs: fix oops in nfs_rename()
OGAWA Hirofumi [Wed, 6 Jan 2010 23:48:26 +0000 (18:48 -0500)]
nfs: fix oops in nfs_rename()

Recent change is missing to update "rehash".  With that change, it will
become the cause of adding dentry to hash twice.

This explains the reason of Oops (dereference the freed dentry in
__d_lookup()) on my machine.

Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Reported-by: Marvin <marvin24@gmx.de>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
14 years ago[IA64] cpumask_of_node() should handle -1 as a node
Anton Blanchard [Wed, 6 Jan 2010 04:55:12 +0000 (15:55 +1100)]
[IA64] cpumask_of_node() should handle -1 as a node

pcibus_to_node can return -1 if we cannot determine which node a pci bus
is on. If passed -1, cpumask_of_node will negatively index the lookup array
and pull in random data:

# cat /sys/devices/pci0000:00/0000:00:01.0/local_cpus
00000000,00000003,00000000,00000000
# cat /sys/devices/pci0000:00/0000:00:01.0/local_cpulist
64-65

Change cpumask_of_node to check for -1 and return cpu_all_mask in this
case:

# cat /sys/devices/pci0000:00/0000:00:01.0/local_cpus
ffffffff,ffffffff,ffffffff,ffffffff
# cat /sys/devices/pci0000:00/0000:00:01.0/local_cpulist
0-127

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Tony Luck <tony.luck@intel.com>
14 years agosunrpc: fix peername failed on closed listener
Xiaotian Feng [Thu, 31 Dec 2009 02:52:36 +0000 (10:52 +0800)]
sunrpc: fix peername failed on closed listener

There're some warnings of "nfsd: peername failed (err 107)!"
socket error -107 means Transport endpoint is not connected.
This warning message was outputed by svc_tcp_accept() [net/sunrpc/svcsock.c],
when kernel_getpeername returns -107. This means socket might be CLOSED.

And svc_tcp_accept was called by svc_recv() [net/sunrpc/svc_xprt.c]

        if (test_bit(XPT_LISTENER, &xprt->xpt_flags)) {
        <snip>
                newxpt = xprt->xpt_ops->xpo_accept(xprt);
        <snip>

So this might happen when xprt->xpt_flags has both XPT_LISTENER and XPT_CLOSE.

Let's take a look at commit b0401d72, this commit has moved the close
processing after do recvfrom method, but this commit also introduces this
warnings, if the xpt_flags has both XPT_LISTENER and XPT_CLOSED, we should
close it, not accpet then close.

Signed-off-by: Xiaotian Feng <dfeng@redhat.com>
Cc: J. Bruce Fields <bfields@fieldses.org>
Cc: Neil Brown <neilb@suse.de>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: stable@kernel.org
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
14 years agonfsd: make sure data is on disk before calling ->fsync
Christoph Hellwig [Fri, 25 Dec 2009 16:44:45 +0000 (17:44 +0100)]
nfsd: make sure data is on disk before calling ->fsync

nfsd is not using vfs_fsync, so I missed it when changing the calling
convention during the 2.6.32 window.  This patch fixes it to not only
start the data writeout, but also wait for it to complete before calling
into ->fsync.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: stable@kernel.org
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
14 years agosunrpc: fix build-time warning
Randy Dunlap [Wed, 6 Jan 2010 22:26:27 +0000 (17:26 -0500)]
sunrpc: fix build-time warning

Fix auth_gss printk format warning:

net/sunrpc/auth_gss/auth_gss.c:660: warning: format '%ld' expects type 'long int', but argument 3 has type 'ssize_t'

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Acked-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
14 years agoMerge branches 'misc' and 'mlx4' into for-next
Roland Dreier [Wed, 6 Jan 2010 21:16:47 +0000 (13:16 -0800)]
Merge branches 'misc' and 'mlx4' into for-next

14 years agoIB/addr: Correct CONFIG_IPv6 to CONFIG_IPV6
Robert P. J. Day [Wed, 6 Jan 2010 21:16:30 +0000 (13:16 -0800)]
IB/addr: Correct CONFIG_IPv6 to CONFIG_IPV6

Correct misspelled "CONFIG_IPv6" that was introduced in commit
d14714df ("IB/addr: Fix IPv6 routing lookup").  The config variable
should be all uppercase.

Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
[ This was my fault when I munged the original patch.  - Roland ]

Signed-off-by: Roland Dreier <rolandd@cisco.com>
14 years agomlx4_core: Fix cleanup in __mlx4_init_one() error path
Eli Cohen [Wed, 6 Jan 2010 20:54:39 +0000 (12:54 -0800)]
mlx4_core: Fix cleanup in __mlx4_init_one() error path

If mlx4_init_port_info() fails, cleanup the initialized ports only.

Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
14 years agoIB/mlx4: Fix queue overflow check in post_recv
Or Gerlitz [Wed, 6 Jan 2010 20:51:30 +0000 (12:51 -0800)]
IB/mlx4: Fix queue overflow check in post_recv

In mlx4_ib_post_recv(), we should check the queue for overflow using
recv_cq instead of send_cq (current code looks like a copy-and-paste
mistake).

Signed-off-by: Or Gerlitz <ogerlitz@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
14 years agoIB/mlx4: Initialize SRQ scatter entries when creating an SRQ
Jack Morgenstein [Wed, 6 Jan 2010 20:48:55 +0000 (12:48 -0800)]
IB/mlx4: Initialize SRQ scatter entries when creating an SRQ

As for memfree mthca hardware, ConnectX also requires SRQ WQE scatter
entries to be initialized with the invalid L_Key at SRQ creation time.

Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
14 years agox86, irq: Check move_in_progress before freeing the vector mapping
Suresh Siddha [Wed, 6 Jan 2010 18:56:31 +0000 (10:56 -0800)]
x86, irq: Check move_in_progress before freeing the vector mapping

With the recent irq migration fixes (post 2.6.32), Gary Hade has noticed
"No IRQ handler for vector" messages during the 2.6.33-rc1 kernel boot on IBM
AMD platforms and root caused the issue to this commit:

> commit 23359a88e7eca3c4f402562b102f23014db3c2aa
> Author: Suresh Siddha <suresh.b.siddha@intel.com>
> Date:   Mon Oct 26 14:24:33 2009 -0800
>
>    x86: Remove move_cleanup_count from irq_cfg

As part of this patch, we have removed the move_cleanup_count check
in smp_irq_move_cleanup_interrupt(). With this change, we can run into a
situation where an irq cleanup interrupt on a cpu can cleanup the vector
mappings associated with multiple irqs, of which one of the irq's migration
might be still in progress. As such when that irq hits the old cpu, we get
the "No IRQ handler" messages.

Fix this by checking for the irq_cfg's move_in_progress and if the move
is still in progress delay the vector cleanup to another irq cleanup
interrupt request (which will happen when the irq starts arriving at the
new cpu destination).

Reported-and-tested-by: Gary Hade <garyhade@us.ibm.com>
Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
LKML-Reference: <1262804191.2732.7.camel@sbs-t61.sc.intel.com>
Cc: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
14 years agoMerge branch 'davinci-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/khilman...
Linus Torvalds [Wed, 6 Jan 2010 18:46:27 +0000 (10:46 -0800)]
Merge branch 'davinci-fixes' of git://git./linux/kernel/git/khilman/linux-davinci

* 'davinci-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/khilman/linux-davinci:
  DaVinci: DM365: Add the device_enable for the DaVinci Keyscan
  davinci: enable ARCH_HAS_HOLES_MEMORYMODEL for DaVinci
  davinci: da8xx/omap-l1: mark RTC as a wakeup source
  davinci: cp_intc: provide set_wake function
  Davinci VPFE Capture: Take i2c adapter id through platform data

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6
Linus Torvalds [Wed, 6 Jan 2010 18:45:50 +0000 (10:45 -0800)]
Merge git://git./linux/kernel/git/jejb/scsi-rc-fixes-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6:
  [SCSI] lpfc 8.3.7: Update Driver version to 8.3.7
  [SCSI] lpfc 8.3.7: Fix discovery failures.
  [SCSI] lpfc 8.3.7: Fix SCSI protocol related errors.
  [SCSI] lpfc 8.3.7: Fix hardware/SLI relates issues
  [SCSI] lpfc 8.3.7: Fix NPIV operation errors
  [SCSI] lpfc 8.3.7: Fix FC protocol errors
  [SCSI] stex: fix scan of nonexistent lun
  [SCSI] pmcraid: fix to avoid twice scsi_dma_unmap for a command
  [SCSI] qla2xxx: Update version number to 8.03.01-k9.
  [SCSI] qla2xxx: Added to EEH support.
  [SCSI] qla2xxx: Extend base EEH support in qla2xxx.
  [SCSI] qla2xxx: Fix for a multiqueue bug in CPU affinity mode
  [SCSI] qla2xxx: Get the link data rate explicitly during device resync.
  [SCSI] cxgb3i: Fix a login over vlan issue

14 years agodrm/i915: remove full registers dump debug
Zhenyu Wang [Mon, 28 Dec 2009 05:23:36 +0000 (13:23 +0800)]
drm/i915: remove full registers dump debug

This one reverts 9e3a6d155ed0a7636b926a798dd7221ea107b274.
As reported by http://bugzilla.kernel.org/show_bug.cgi?id=14485,
this dump will cause hang problem on some machine. If something
really needs this kind of full registers dump, that could be done
within intel-gpu-tools.

Cc: Ben Gamari <bgamari.foss@gmail.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
14 years agodrm/i915: Add DP dpll limit on ironlake and use existing DPLL search function
Zhao Yakui [Thu, 31 Dec 2009 08:06:04 +0000 (16:06 +0800)]
drm/i915: Add DP dpll limit on ironlake and use existing DPLL search function

For some clocks, the old Ironlake DPLL calculator wold give m/n/p
combinations that didn't match the spreadsheet of what HW validation
tests.  Instead, use the G4X DPLL calculator, which does a better job
at it.

So we use the intel_g4x_find_best_pll to calculate the DPLL for CRT/HDMI/LVDS
on ironlake. At the same time to consider the dpll setting for display port, we
add the display port DPLL limit on ironlake, which will directly use the
function of intel_find_pll_ironlake_dp to get the corresponding dpll setting.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
14 years agodrm/i915: Select the correct BPC for LVDS on Ironlake
Zhao Yakui [Mon, 4 Jan 2010 08:29:32 +0000 (16:29 +0800)]
drm/i915: Select the correct BPC for LVDS on Ironlake

Select the correct BPC for LVDS on Ironlake. If it is 18-bit LVDS panel,
the BPC will be 6. When it is 24-bit LVDS panel, the BPC will 8.
At the same time the BPC will be 8 when the output device is CRT/HDMI/DP.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
14 years agodrm/i915: Make the BPC in FDI rx/transcoder be consistent with that in pipeconf on...
Zhao Yakui [Mon, 4 Jan 2010 08:29:31 +0000 (16:29 +0800)]
drm/i915: Make the BPC in FDI rx/transcoder be consistent with that in pipeconf on Ironlake

Make the BPC in FDI rx/transcoder be consistent with that in pipeconf on Ironlake.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
14 years agodrm/i915: Enable/disable the dithering for LVDS based on VBT setting
Zhao Yakui [Mon, 4 Jan 2010 08:29:30 +0000 (16:29 +0800)]
drm/i915: Enable/disable the dithering for LVDS based on VBT setting

Enable/disable the dithering for LVDS based on VBT setting. On the 965/g4x
platform the dithering flag is defined in LVDS register. And on the ironlake
the dithering flag is defined in pipeconf register.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
14 years agodrm/i915: Permit pinning whilst the device is 'suspended'
Chris Wilson [Mon, 4 Jan 2010 18:57:57 +0000 (18:57 +0000)]
drm/i915: Permit pinning whilst the device is 'suspended'

As pinning (allocating and binding GTT memory) does not actually invoke
GPU commands, it is safe, and indeed is attempted, during resumption
from suspension:

  [drm:intel_init_clock_gating] *ERROR* failed to pin power context: -16

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reported-by: Hugh Dickins <hugh.dickins@tiscali.co.uk>
Cc: stable@kernel.org
Signed-off-by: Eric Anholt <eric@anholt.net>
14 years agodrm/i915: Hold struct mutex whilst pinning power context bo.
Chris Wilson [Mon, 4 Jan 2010 18:57:56 +0000 (18:57 +0000)]
drm/i915: Hold struct mutex whilst pinning power context bo.

Hugh found an error path where we were attempting to unref a bo without
holding the struct mutex:

  [drm:intel_init_clock_gating] *ERROR* failed to pin power context: -16
  ------------[ cut here ]------------
  WARNING: at drivers/gpu/drm/drm_gem.c:438 drm_gem_object_free+0x20/0x5e()
  Hardware name: ESPRIMO Mobile V5505
  Modules linked in: snd_pcm_oss snd_mixer_oss snd_seq snd_seq_device
  Pid: 3793, comm: s2ram Not tainted 2.6.33-rc2 #4
  Call Trace:
   [<7815298e>] warn_slowpath_common+0x59/0x6b
   [<781529b3>] warn_slowpath_null+0x13/0x18
   [<78317c1a>] ? drm_gem_object_free+0x20/0x5e
   [<78317c1a>] drm_gem_object_free+0x20/0x5e
   [<78317bfa>] ? drm_gem_object_free+0x0/0x5e
   [<7829df11>] kref_put+0x38/0x45
   [<7833a5f0>] intel_init_clock_gating+0x232/0x271
   [<78317bfa>] ? drm_gem_object_free+0x0/0x5e
   [<7832c307>] i915_restore_state+0x21a/0x2b3
   [<7832379d>] i915_resume+0x3c/0xbb
   [<78174fe5>] ? trace_hardirqs_on_caller+0xfc/0x123
   [<7831c756>] ? drm_class_resume+0x0/0x3e
   [<7831c78d>] drm_class_resume+0x37/0x3e
   [<78351e0a>] legacy_resume+0x1e/0x51
   [<78351ece>] device_resume+0x91/0xab
   [<7831c756>] ? drm_class_resume+0x0/0x3e
   [<78352226>] dpm_resume+0x58/0x10f
   [<783522fb>] dpm_resume_end+0x1e/0x2c
   [<78180f80>] suspend_devices_and_enter+0x61/0x84
   [<78180ff8>] enter_state+0x55/0x83
   [<7818091c>] state_store+0x94/0xaa
   [<7829d09e>] kobj_attr_store+0x1e/0x23
   [<782098e0>] sysfs_write_file+0x66/0x99
   [<781cd2f0>] vfs_write+0x8a/0x108
   [<781cd408>] sys_write+0x3c/0x63
   [<78125c10>] sysenter_do_call+0x12/0x36
  ---[ end trace a343537f29950fda ]---

It is in fact slightly more insiduous that first appears since we are
attempting to not just free the object without the lock, but are trying
to do the whole bo manipulation without holding the lock.

Reported-by: Hugh Dickins <hugh.dickins@tiscali.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org
Signed-off-by: Eric Anholt <eric@anholt.net>
14 years agodrm/i915: fix unused var
Andrew Morton [Tue, 17 Nov 2009 22:08:52 +0000 (14:08 -0800)]
drm/i915: fix unused var

drivers/gpu/drm/i915/i915_dma.c: In function 'i915_driver_load':
drivers/gpu/drm/i915/i915_dma.c:1114: warning: 'll_base' may be used uninitialized in this function

Partly this is because gcc isn't smart enough.  But `ll_base' does get used
uninitialised in the DRM_DEBUG() call.

Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: Eric Anholt <eric@anholt.net>
Cc: Dave Airlie <airlied@linux.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Eric Anholt <eric@anholt.net>
14 years agodrm/i915: Storage class should be before const qualifier
Tobias Klauser [Wed, 23 Dec 2009 13:14:34 +0000 (14:14 +0100)]
drm/i915: Storage class should be before const qualifier

The C99 specification states in section 6.11.5:

The placement of a storage-class specifier other than at the beginning
of the declaration specifiers in a declaration is an obsolescent
feature.

Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Signed-off-by: Eric Anholt <eric@anholt.net>
14 years agodrm/i915: remove render reclock support
Jesse Barnes [Thu, 17 Dec 2009 19:11:13 +0000 (11:11 -0800)]
drm/i915: remove render reclock support

This code generally fails to adjust the render clock, and when it does,
it conflicts with some other register settings and can cause problems.

So remove this code altogether.  I'm reworking it now to do the right
thing, but the only bit it will share is the VBT check for whether
reclocking is supported, so I'm leaving that bit.

Reverts most of 652c393a3368af84359da37c45afc35a91144960 ("add dynamic
clock frequency control"), though for many the regressions showed up
in the later 181a5336d6cc836f05507410d66988c483ad0154 ("Fix render
reclock availability detection").

Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Eric Anholt <eric@anholt.net>
14 years agodrm/i915: Fix RC6 suspend/resume
Andrew Lutomirski [Mon, 21 Dec 2009 15:10:22 +0000 (10:10 -0500)]
drm/i915: Fix RC6 suspend/resume

We restored RC6 twice on resume, even with modesetting off.  Instead,
only restore it once and skip RC6 initialization entirely in non-KMS mode.

Signed-off-by: Andy Lutomirski <luto@mit.edu>
Tested-by: Jeff Chua <jeff.chua.linux@gmail.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
14 years agodrm/i915: execbuf2 support
Jesse Barnes [Fri, 18 Dec 2009 03:05:42 +0000 (22:05 -0500)]
drm/i915: execbuf2 support

This patch adds a new execbuf ioctl, execbuf2, for use by clients that
want to control fence register allocation more finely.  The buffer
passed in to the new ioctl includes a new relocation type to indicate
whether a given object needs a fence register assigned for the command
buffer in question.

Compatibility with the existing execbuf ioctl is implemented in terms
of the new code, preserving the assumption that fence registers are
required for pre-965 rendering commands.

Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
[ickle: Remove pre-emptive clear_fence_reg()]
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Kristian Høgsberg <krh@bitplanet.net>
[anholt: Removed dmesg spam]
Signed-off-by: Eric Anholt <eric@anholt.net>
14 years agoDaVinci: DM365: Add the device_enable for the DaVinci Keyscan
Miguel Aguilar [Fri, 13 Nov 2009 19:44:23 +0000 (13:44 -0600)]
DaVinci: DM365: Add the device_enable for the DaVinci Keyscan

Adds the device_enable function to the DaVinci Keyscan platform data
to setup the PINMUX configuration.

It also removes #ifdef from the DM365 EVM board in order to load it
properly as a module.

Signed-off-by: Miguel Aguilar <miguel.aguilar@ridgerun.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
14 years agodavinci: enable ARCH_HAS_HOLES_MEMORYMODEL for DaVinci
Sekhar Nori [Thu, 19 Nov 2009 10:04:01 +0000 (15:34 +0530)]
davinci: enable ARCH_HAS_HOLES_MEMORYMODEL for DaVinci

All DaVinci platforms include a DSP or co-processor for
audio/video acceleration.

While creating memory for the DSP/co-processor, system
integrator can end up creating a hole in the memory map
of the sort:

<kernel memory> <hole (memory for DSP)> <kernel memory>

This sort of configuration needs ARCH_HAS_HOLES_MEMORYMODEL
enabled. See further details see this discussion on ARM
linux mailing list:
http://www.mail-archive.com/linux-omap@vger.kernel.org/msg15262.html

The patch is boot tested on OMAP-L138, DM6446 and DM355 EVMs

Signed-off-by: Sekhar Nori <nsekhar@ti.com>
CC: Sriramakrishnan <srk@ti.com>
CC: Khasim Syed Mohammed <khasim@ti.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
14 years agodavinci: da8xx/omap-l1: mark RTC as a wakeup source
Sekhar Nori [Mon, 16 Nov 2009 11:51:31 +0000 (17:21 +0530)]
davinci: da8xx/omap-l1: mark RTC as a wakeup source

On da850, RTC alarm is a wakeup source from deep sleep.
Mark it as a wakeup source after the rtc platform device
is registered.

Without this patch, the rtc-omap driver suspends the RTC
during the suspend sequence and hence it cannot wakeup the
SoC.

Signed-off-by: Sekhar Nori <nsekhar@ti.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
14 years agodavinci: cp_intc: provide set_wake function
Sekhar Nori [Mon, 16 Nov 2009 11:51:30 +0000 (17:21 +0530)]
davinci: cp_intc: provide set_wake function

There is nothing special to be done for interrupts
which can wakeup the device from sleep on CP-INTC,
but not having a set_wake implemented prevents use
of common drivers which expect this function to be
implemented for all wakeup interrupt sources.

This patch fixes the issue encountered when using the
omap-rtc driver on DA850. On DA850 the RTC alarm
interrupt is used to wake up the SoC from deep sleep
mode. Without this patch, the disable_irq_wake throws
an unbalanced wake disable warning while resuming
because the previous enable call fails for lack of
set_wake implementation.

Signed-off-by: Sekhar Nori <nsekhar@ti.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
14 years agoDavinci VPFE Capture: Take i2c adapter id through platform data
Vaibhav Hiremath [Tue, 13 Oct 2009 15:08:54 +0000 (15:08 +0000)]
Davinci VPFE Capture: Take i2c adapter id through platform data

The I2C adapter ID is actually depends on Board and may vary, Davinci
uses id=1, but in case of AM3517 id=3.

So modified respective davinci board files.

Signed-off-by: Vaibhav Hiremath <hvaibhav@ti.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
14 years agoMerge branch 'for-linus' of git://git.open-osd.org/linux-open-osd
Linus Torvalds [Wed, 6 Jan 2010 09:41:07 +0000 (01:41 -0800)]
Merge branch 'for-linus' of git://git.open-osd.org/linux-open-osd

* 'for-linus' of git://git.open-osd.org/linux-open-osd:
  exofs: simple_write_end does not mark_inode_dirty
  exofs: fix pnfs_osd re-definitions in pre-pnfs trees

14 years agoMerge master.kernel.org:/home/rmk/linux-2.6-arm
Linus Torvalds [Wed, 6 Jan 2010 09:16:26 +0000 (01:16 -0800)]
Merge master.kernel.org:/home/rmk/linux-2.6-arm

* master.kernel.org:/home/rmk/linux-2.6-arm: (22 commits)
  ARM: 5865/1: nuc900 ethernet driver needs mii
  ARM: 5864/1: Implement arch_reset() in NUC900
  ARM: 5863/1: fix bugs of clock source of NUC900
  ARM: 5858/1: Remove unused vma_vm_flags macro from v7wbi_flush_user_tlb_range
  imx/mx3: depend on USB_ULPI for otg_ulpi_create
  ARM: MX3: make CPU revision number detection work on all boards
  mx25: pdk: add platform code for FEC support
  mx25: add support for FEC on i.MX25
  mx25: s/NO_PAD_CTL/NO_PAD_CTRL/
  mx31moboard: fix usbh device names
  mx3: add support for the mt9v022 camera sensor to pcm037 platform
  mx27: mxt_td60: Remove not used UART pins
  [ARM] pxa/poodle: fix incorrect 'gpio_card_detect' of MMC
  [ARM] pxa/zylonite: simplify reduntant gpio settings on mmc slot
  [ARM] pxa/ttc_dkb: remove duplicate macro definition
  [ARM] pxa/zeus: provide power-source information when APM is enabled
  [ARM] pxa/zeus: relax memory timings on Zeus ethernet ports
  [ARM] pxa/zeus: make internal zeus_get_pcb_info static
  [ARM] pxa/littleton: select CPU_PXA300 and CPU_PXA310
  [ARM] pxa/littleton: add UART3 GPIO config
  ...

14 years agomodules: Skip empty sections when exporting section notes
Ben Hutchings [Sat, 19 Dec 2009 14:43:01 +0000 (14:43 +0000)]
modules: Skip empty sections when exporting section notes

Commit 35dead4 "modules: don't export section names of empty sections
via sysfs" changed the set of sections that have attributes, but did
not change the iteration over these attributes in add_notes_attrs().
This can lead to add_notes_attrs() creating attributes with the wrong
names or with null name pointers.

Introduce a sect_empty() function and use it in both add_sect_attrs()
and add_notes_attrs().

Reported-by: Martin Michlmayr <tbm@cyrius.com>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Tested-by: Martin Michlmayr <tbm@cyrius.com>
Cc: stable@kernel.org
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agodrm/radeon/kms: add missing breaks in i2c and ss lookups
Alex Deucher [Mon, 28 Dec 2009 18:45:23 +0000 (13:45 -0500)]
drm/radeon/kms: add missing breaks in i2c and ss lookups

Should fix fdo bug 25741

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms: add primary dac adj values table
Alex Deucher [Tue, 29 Dec 2009 17:09:17 +0000 (12:09 -0500)]
drm/radeon/kms: add primary dac adj values table

Look up primary dac adj values from the table if
there is no bios or bios dac table to reference.

The lookup table may need to be adjusted for certain
families.

Should fix kernel bug 14945.

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms: fallback to default connector table
Alex Deucher [Tue, 5 Jan 2010 00:12:02 +0000 (19:12 -0500)]
drm/radeon/kms: fallback to default connector table

if necessary for combios

Some early combios radeon cards don't have a connector
table or dac table in the bios, if they do not, fallback
to the default tables.

Should fix kernel bug 14963.

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agoLinux 2.6.33-rc3 v2.6.33-rc3
Linus Torvalds [Wed, 6 Jan 2010 00:02:46 +0000 (16:02 -0800)]
Linux 2.6.33-rc3

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Wed, 6 Jan 2010 00:01:58 +0000 (16:01 -0800)]
Merge git://git./linux/kernel/git/davem/sparc-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6:
  sparc64: Fix Niagara2 perf event handling.
  sparc64: Fix NMI programming when perf events are active.
  bbc_envctrl: Clean up properly if kthread_run() fails.

14 years agoRevert "x86: Side-step lguest problem by only building cmpxchg8b_emu for pre-Pentium"
Rusty Russell [Tue, 5 Jan 2010 02:18:49 +0000 (12:48 +1030)]
Revert "x86: Side-step lguest problem by only building cmpxchg8b_emu for pre-Pentium"

This reverts commit ae1b22f6e46c03cede7cea234d0bf2253b4261cf.

As Linus said in 982d007a6ee: "There was something really messy about
cmpxchg8b and clone CPU's, so if you enable it on other CPUs later, do it
carefully."

This breaks lguest for those configs, but we can fix that by emulating
if we have to.

Fixes: http://bugzilla.kernel.org/show_bug.cgi?id=14884
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: stable@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jlbec...
Linus Torvalds [Wed, 6 Jan 2010 00:01:04 +0000 (16:01 -0800)]
Merge branch 'upstream-linus' of git://git./linux/kernel/git/jlbec/ocfs2

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jlbec/ocfs2:
  ocfs2: Handle O_DIRECT when writing to a refcounted cluster.

14 years agoMerge branch 'for-2.6.33' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie...
Linus Torvalds [Tue, 5 Jan 2010 23:59:56 +0000 (15:59 -0800)]
Merge branch 'for-2.6.33' of git://git./linux/kernel/git/broonie/sound-2.6

* 'for-2.6.33' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound-2.6:
  ASoC: fixup oops in generic AC97 codec glue
  ASoC: fix params_rate() macro use in several codecs
  ASoC: fsi-ak4642: Remove ak4642_add_i2c_device

14 years agox86: copy_from_user() should not return -EFAULT
Heiko Carstens [Tue, 5 Jan 2010 13:19:11 +0000 (14:19 +0100)]
x86: copy_from_user() should not return -EFAULT

Callers of copy_from_user() expect it to return the number of bytes
it could not copy. In no case it is supposed to return -EFAULT.

In case of a detected buffer overflow just return the requested
length. In addition one could think of a memset that would clear
the size of the target object.

[ hpa: code is not in .32 so not needed for -stable ]

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Acked-by: Arjan van de Ven <arjan@linux.intel.com>
LKML-Reference: <20100105131911.GC5480@osiris.boeblingen.de.ibm.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
14 years agoARM: 5865/1: nuc900 ethernet driver needs mii
Li Jie [Thu, 31 Dec 2009 15:03:16 +0000 (16:03 +0100)]
ARM: 5865/1: nuc900 ethernet driver needs mii

nuc900 ethernet driver uses mii_xx_xx serials api, so mii module should be selected.

Signed-off-by: lijie <eltshanli@gmail.com>
Acked-by: Wan ZongShun <mcuos.com@gmail.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
14 years agoARM: 5864/1: Implement arch_reset() in NUC900
Li Jie [Thu, 31 Dec 2009 14:50:02 +0000 (15:50 +0100)]
ARM: 5864/1: Implement arch_reset() in NUC900

Implement arch_reset(), reboot from shell become possible.

Signed-off-by: lijie <eltshanli@gmail.com>
Acked-by: Wan ZongShun <mcuos.com@gmail.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
14 years agoARM: 5863/1: fix bugs of clock source of NUC900
Li Jie [Thu, 31 Dec 2009 14:57:53 +0000 (15:57 +0100)]
ARM: 5863/1: fix bugs of clock source of NUC900

This patch fix following bugs:

1. typo error, CLOCK_EVT_MODE_PERIODIC -> CLOCK_EVT_FEAT_PERIODIC
2. TCSR register of timer1 missed PRESCALE
3. timer1 should be enabled before register it to clock source.

Signed-off-by: lijie <eltshanli@gmail.com>
Acked-by: Wan ZongShun <mcuos.com@gmail.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
14 years agoARM: 5858/1: Remove unused vma_vm_flags macro from v7wbi_flush_user_tlb_range
Bahadir Balban [Fri, 25 Dec 2009 13:25:48 +0000 (14:25 +0100)]
ARM: 5858/1: Remove unused vma_vm_flags macro from v7wbi_flush_user_tlb_range

Signed-off-by: Bahadir Balban <bbalban@b-labs.co.uk>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
14 years agosparc64: Fix Niagara2 perf event handling.
David S. Miller [Tue, 5 Jan 2010 07:16:03 +0000 (23:16 -0800)]
sparc64: Fix Niagara2 perf event handling.

For chips like Niagara2 that have true overflow indications
in the %pcr (which we don't actually need and don't use)
the interrupt signal persists until the overflow bits are
cleared by an explicit %pcr write.

Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agoexofs: simple_write_end does not mark_inode_dirty
Boaz Harrosh [Sun, 27 Dec 2009 15:01:42 +0000 (17:01 +0200)]
exofs: simple_write_end does not mark_inode_dirty

exofs uses simple_write_end() for it's .write_end handler. But
it is not enough because simple_write_end() does not call
mark_inode_dirty() when it extends i_size. So even if we do
call mark_inode_dirty at beginning of write out, with a very
long IO and a saturated system we might get the .write_inode()
called while still extend-writing to file and miss out on the last
i_size updates.

So override .write_end, call simple_write_end(), and afterwords if
i_size was changed call mark_inode_dirty().

It stands to logic that since simple_write_end() was the one extending
i_size it should also call mark_inode_dirty(). But it looks like all
users of simple_write_end() are memory-bound pseudo filesystems, who
could careless about mark_inode_dirty(). I might submit a
warning-comment patch to simple_write_end() in future.

CC: Stable <stable@kernel.org>
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
14 years agoexofs: fix pnfs_osd re-definitions in pre-pnfs trees
Boaz Harrosh [Mon, 21 Dec 2009 14:36:23 +0000 (16:36 +0200)]
exofs: fix pnfs_osd re-definitions in pre-pnfs trees

Some on disk exofs constants and types are defined in the pnfs_osd_xdr.h
file. Since we needed these types before the pnfs-objects code was
accepted to mainline we duplicated the minimal needed definitions into
an exofs local header. The definitions where conditionally included
depending on !CONFIG_PNFS defined. So if PNFS was present in the tree
definitions are taken from there and if not they are defined locally.

That was all good but, the CONFIG_PNFS is planed to be included upstream
before the pnfs-objects is also included. (The first pnfs batch might be
pnfs-files only)

So condition exofs local definitions on the absence of pnfs_osd_xdr.h
inclusion (__PNFS_OSD_XDR_H__ not defined). User code must make sure
that in future pnfs_osd_xdr.h will be included before fs/exofs/pnfs.h,
which happens to be so in current code.

Once pnfs-objects hits mainline, exofs's local header will be removed.

Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
14 years agoreiserfs: Relax lock on xattr removing
Frederic Weisbecker [Tue, 5 Jan 2010 01:14:30 +0000 (02:14 +0100)]
reiserfs: Relax lock on xattr removing

When we remove an xattr, we call lookup_and_delete_xattr()
that takes some private xattr inodes mutexes. But we hold
the reiserfs lock at this time, which leads to dependency
inversions.

We can safely call lookup_and_delete_xattr() without the
reiserfs lock, where xattr inodes lookups only need the
xattr inodes mutexes.

Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Christian Kujau <lists@nerdbynature.de>
Cc: Alexander Beregalov <a.beregalov@gmail.com>
Cc: Chris Mason <chris.mason@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
14 years agoreiserfs: Relax the lock before truncating pages
Frederic Weisbecker [Mon, 4 Jan 2010 23:15:38 +0000 (00:15 +0100)]
reiserfs: Relax the lock before truncating pages

While truncating a file, reiserfs_setattr() calls inode_setattr()
that will truncate the mapping for the given inode, but for that
it needs the pages locks.

In order to release these, the owners need the reiserfs lock to
complete their jobs. But they can't, as we don't release it before
calling inode_setattr().

We need to do that to fix the following softlockups:

INFO: task flush-8:0:2149 blocked for more than 120 seconds.
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
flush-8:0     D f51af998     0  2149      2 0x00000000
 f51af9ac 00000092 00000002 f51af998 c2803304 00000000 c1894ad0 010f3000
 f51af9cc c1462604 c189ef80 f51af974 c1710304 f715b450 f715b5ec c2807c40
 00000000 0005bb00 c2803320 c102c55b c1710304 c2807c50 c2803304 00000246
Call Trace:
 [<c1462604>] ? schedule+0x434/0xb20
 [<c102c55b>] ? resched_task+0x4b/0x70
 [<c106fa22>] ? mark_held_locks+0x62/0x80
 [<c146414d>] ? mutex_lock_nested+0x1fd/0x350
 [<c14640b9>] mutex_lock_nested+0x169/0x350
 [<c1178cde>] ? reiserfs_write_lock+0x2e/0x40
 [<c1178cde>] reiserfs_write_lock+0x2e/0x40
 [<c11719a2>] do_journal_end+0xc2/0xe70
 [<c1172912>] journal_end+0xb2/0x120
 [<c11686b3>] ? pathrelse+0x33/0xb0
 [<c11729e4>] reiserfs_end_persistent_transaction+0x64/0x70
 [<c1153caa>] reiserfs_get_block+0x12ba/0x15f0
 [<c106fa22>] ? mark_held_locks+0x62/0x80
 [<c1154b24>] reiserfs_writepage+0xa74/0xe80
 [<c1465a27>] ? _raw_spin_unlock_irq+0x27/0x50
 [<c11f3d25>] ? radix_tree_gang_lookup_tag_slot+0x95/0xc0
 [<c10b5377>] ? find_get_pages_tag+0x127/0x1a0
 [<c106fa22>] ? mark_held_locks+0x62/0x80
 [<c106fcd4>] ? trace_hardirqs_on_caller+0x124/0x170
 [<c10bc1e0>] __writepage+0x10/0x40
 [<c10bc9ab>] write_cache_pages+0x16b/0x320
 [<c10bc1d0>] ? __writepage+0x0/0x40
 [<c10bcb88>] generic_writepages+0x28/0x40
 [<c10bcbd5>] do_writepages+0x35/0x40
 [<c11059f7>] writeback_single_inode+0xc7/0x330
 [<c11067b2>] writeback_inodes_wb+0x2c2/0x490
 [<c1106a86>] wb_writeback+0x106/0x1b0
 [<c1106cf6>] wb_do_writeback+0x106/0x1e0
 [<c1106c18>] ? wb_do_writeback+0x28/0x1e0
 [<c1106e0a>] bdi_writeback_task+0x3a/0xb0
 [<c10cbb13>] bdi_start_fn+0x63/0xc0
 [<c10cbab0>] ? bdi_start_fn+0x0/0xc0
 [<c105d1f4>] kthread+0x74/0x80
 [<c105d180>] ? kthread+0x0/0x80
 [<c100327a>] kernel_thread_helper+0x6/0x10
3 locks held by flush-8:0/2149:
 #0:  (&type->s_umount_key#30){+++++.}, at: [<c110676f>] writeback_inodes_wb+0x27f/0x490
 #1:  (&journal->j_mutex){+.+...}, at: [<c117199a>] do_journal_end+0xba/0xe70
 #2:  (&REISERFS_SB(s)->lock){+.+.+.}, at: [<c1178cde>] reiserfs_write_lock+0x2e/0x40
INFO: task fstest:3813 blocked for more than 120 seconds.
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
fstest        D 00000002     0  3813   3812 0x00000000
 f5103c94 00000082 f5103c40 00000002 f5ad5450 00000007 f5103c28 011f3000
 00000006 f5ad5450 c10bb005 00000480 c1710304 f5ad5450 f5ad55ec c2907c40
 00000001 f5ad5450 f5103c74 00000046 00000002 f5ad5450 00000007 f5103c6c
Call Trace:
 [<c10bb005>] ? free_hot_cold_page+0x1d5/0x280
 [<c1462d64>] io_schedule+0x74/0xc0
 [<c10b5a45>] sync_page+0x35/0x60
 [<c146325a>] __wait_on_bit_lock+0x4a/0x90
 [<c10b5a10>] ? sync_page+0x0/0x60
 [<c10b59e5>] __lock_page+0x85/0x90
 [<c105d660>] ? wake_bit_function+0x0/0x60
 [<c10bf654>] truncate_inode_pages_range+0x1e4/0x2d0
 [<c10bf75f>] truncate_inode_pages+0x1f/0x30
 [<c10bf7cf>] truncate_pagecache+0x5f/0xa0
 [<c10bf86a>] vmtruncate+0x5a/0x70
 [<c10fdb7d>] inode_setattr+0x5d/0x190
 [<c1150117>] reiserfs_setattr+0x1f7/0x2f0
 [<c1464569>] ? down_write+0x49/0x70
 [<c10fde01>] notify_change+0x151/0x330
 [<c10e6f3d>] do_truncate+0x6d/0xa0
 [<c10f4ce2>] do_filp_open+0x9a2/0xcf0
 [<c1465aec>] ? _raw_spin_unlock+0x2c/0x50
 [<c10fec50>] ? alloc_fd+0xe0/0x100
 [<c10e602d>] do_sys_open+0x6d/0x130
 [<c1002cfb>] ? sysenter_exit+0xf/0x16
 [<c10e615e>] sys_open+0x2e/0x40
 [<c1002ccc>] sysenter_do_call+0x12/0x32
3 locks held by fstest/3813:
 #0:  (&sb->s_type->i_mutex_key#4){+.+.+.}, at: [<c10e6f33>] do_truncate+0x63/0xa0
 #1:  (&sb->s_type->i_alloc_sem_key#3){+.+.+.}, at: [<c10fdf07>] notify_change+0x257/0x330
 #2:  (&REISERFS_SB(s)->lock){+.+.+.}, at: [<c1178c8e>] reiserfs_write_lock_once+0x2e/0x50

Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Christian Kujau <lists@nerdbynature.de>
Cc: Alexander Beregalov <a.beregalov@gmail.com>
Cc: Chris Mason <chris.mason@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
14 years agoreiserfs: Fix recursive lock on lchown
Frederic Weisbecker [Mon, 4 Jan 2010 21:04:01 +0000 (22:04 +0100)]
reiserfs: Fix recursive lock on lchown

On chown, reiserfs will call reiserfs_setattr() to change the owner
of the given inode, but it may also recursively call
reiserfs_setattr() to propagate the owner change to the private xattr
files for this inode.

Hence, the reiserfs lock may be acquired twice which is not wanted
as reiserfs_setattr() calls journal_begin() that is going to try to
relax the lock in order to safely acquire the journal mutex.

Using reiserfs_write_lock_once() from reiserfs_setattr() solves
the problem.

This fixes the following warning, that precedes a lockdep report.

WARNING: at fs/reiserfs/lock.c:95 reiserfs_lock_check_recursive+0x3f/0x50()
Hardware name: MS-7418
Unwanted recursive reiserfs lock!
Pid: 4189, comm: fsstress Not tainted 2.6.33-rc2-tip-atom+ #195
Call Trace:
 [<c1178bff>] ? reiserfs_lock_check_recursive+0x3f/0x50
 [<c1178bff>] ? reiserfs_lock_check_recursive+0x3f/0x50
 [<c103f7ac>] warn_slowpath_common+0x6c/0xc0
 [<c1178bff>] ? reiserfs_lock_check_recursive+0x3f/0x50
 [<c103f84b>] warn_slowpath_fmt+0x2b/0x30
 [<c1178bff>] reiserfs_lock_check_recursive+0x3f/0x50
 [<c1172ae3>] do_journal_begin_r+0x83/0x350
 [<c1172f2d>] journal_begin+0x7d/0x140
 [<c106509a>] ? in_group_p+0x2a/0x30
 [<c10fda71>] ? inode_change_ok+0x91/0x140
 [<c115007d>] reiserfs_setattr+0x15d/0x2e0
 [<c10f9bf3>] ? dput+0xe3/0x140
 [<c1465adc>] ? _raw_spin_unlock+0x2c/0x50
 [<c117831d>] chown_one_xattr+0xd/0x10
 [<c11780a3>] reiserfs_for_each_xattr+0x113/0x2c0
 [<c1178310>] ? chown_one_xattr+0x0/0x10
 [<c14641e9>] ? mutex_lock_nested+0x2a9/0x350
 [<c117826f>] reiserfs_chown_xattrs+0x1f/0x60
 [<c106509a>] ? in_group_p+0x2a/0x30
 [<c10fda71>] ? inode_change_ok+0x91/0x140
 [<c1150046>] reiserfs_setattr+0x126/0x2e0
 [<c1177c20>] ? reiserfs_getxattr+0x0/0x90
 [<c11b0d57>] ? cap_inode_need_killpriv+0x37/0x50
 [<c10fde01>] notify_change+0x151/0x330
 [<c10e659f>] chown_common+0x6f/0x90
 [<c10e67bd>] sys_lchown+0x6d/0x80
 [<c1002ccc>] sysenter_do_call+0x12/0x32
---[ end trace 7c2b77224c1442fc ]---

Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Christian Kujau <lists@nerdbynature.de>
Cc: Alexander Beregalov <a.beregalov@gmail.com>
Cc: Chris Mason <chris.mason@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
14 years agoRevert "x86: Side-step lguest problem by only building cmpxchg8b_emu for pre-Pentium"
Rusty Russell [Tue, 5 Jan 2010 02:18:49 +0000 (12:48 +1030)]
Revert "x86: Side-step lguest problem by only building cmpxchg8b_emu for pre-Pentium"

This reverts commit ae1b22f6e46c03cede7cea234d0bf2253b4261cf.

As Linus said in 982d007a6ee: "There was something really messy about
cmpxchg8b and clone CPU's, so if you enable it on other CPUs later, do it
carefully."

This breaks lguest for those configs, but we can fix that by emulating
if we have to.

Fixes: http://bugzilla.kernel.org/show_bug.cgi?id=14884
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
LKML-Reference: <201001051248.49700.rusty@rustcorp.com.au>
Cc: stable@kernel.org
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
14 years agoPCIe AER: prevent AER injection if hardware masks error reporting
Youquan,Song [Thu, 17 Dec 2009 13:22:48 +0000 (08:22 -0500)]
PCIe AER: prevent AER injection if hardware masks error reporting

The Correcteable/Uncorrectable Error Mask Registers are used by PCIe AER
driver which will controls the reporting of individual errors to PCIe RC
via PCIe error messages.

If hardware masks special error reporting to RC, the aer_inject driver
should not inject aer error.

Acked-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Youquan, Song <youquan.song@intel.com>
Acked-by: Ying, Huang <ying.huang@intel.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
14 years agoPCI/PM: Use per-device D3 delays
Rafael J. Wysocki [Thu, 31 Dec 2009 11:15:54 +0000 (12:15 +0100)]
PCI/PM: Use per-device D3 delays

It turns out that some PCI devices require extra delays when changing
power state from D3 to D0 (and the other way around).  Although this
is against the PCI specification, we can handle it quite easily by
allowing drivers to define arbitrary D3 delays for devices known to
require extra time for switching power states.

Introduce additional field d3_delay in struct pci_dev and use it to
store the value of the device's D0->D3 delay, in miliseconds.  Make
the PCI PM core code use the per-device d3_delay unless
pci_pm_d3_delay is greater (in which case the latter is used).
[This also allows the driver to specify d3_delay shorter than the
 10 ms required by the PCI standard if the device is known to be able
 to handle that.]

Make the sky2 driver set d3_delay to 150 for devices handled by it.

Fixes http://bugzilla.kernel.org/show_bug.cgi?id=14730 which is a
listed regression from 2.6.30.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>