firefly-linux-kernel-4.4.55.git
15 years agoperf counters: add prctl interface to disable/enable counters
Ingo Molnar [Thu, 11 Dec 2008 13:59:31 +0000 (14:59 +0100)]
perf counters: add prctl interface to disable/enable counters

Add a way for self-monitoring tasks to disable/enable counters summarily,
via a prctl:

PR_TASK_PERF_COUNTERS_DISABLE 31
PR_TASK_PERF_COUNTERS_ENABLE 32

Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoperf counters: implement PERF_COUNT_TASK_CLOCK
Ingo Molnar [Thu, 11 Dec 2008 13:03:20 +0000 (14:03 +0100)]
perf counters: implement PERF_COUNT_TASK_CLOCK

Impact: add new perf-counter type

The 'task clock' counter counts the amount of time a task is executing,
in nanoseconds. It stops ticking when a task is scheduled out either due
to it blocking, sleeping or it being preempted.

This counter type is a Linux kernel based abstraction, it is available
even if the hardware does not support native hardware performance counters.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoperf counters: consolidate hw_perf save/restore APIs
Ingo Molnar [Thu, 11 Dec 2008 12:45:51 +0000 (13:45 +0100)]
perf counters: consolidate hw_perf save/restore APIs

Impact: cleanup

Rename them to better match up the usual IRQ disable/enable APIs:

 hw_perf_disable_all()  => hw_perf_save_disable()
 hw_perf_restore_ctrl() => hw_perf_restore()

Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoperf counters: implement PERF_COUNT_CPU_CLOCK
Ingo Molnar [Thu, 11 Dec 2008 12:21:10 +0000 (13:21 +0100)]
perf counters: implement PERF_COUNT_CPU_CLOCK

Impact: add new perf-counter type

The 'CPU clock' counter counts the amount of CPU clock time that is
elapsing, in nanoseconds. (regardless of how much of it the task is
spending on a CPU executing)

This counter type is a Linux kernel based abstraction, it is available
even if the hardware does not support native hardware performance counters.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoperf counters: hw driver API
Ingo Molnar [Thu, 11 Dec 2008 11:46:46 +0000 (12:46 +0100)]
perf counters: hw driver API

Impact: restructure code, introduce hw_ops driver abstraction

Introduce this abstraction to handle counter details:

 struct hw_perf_counter_ops {
void (*hw_perf_counter_enable) (struct perf_counter *counter);
void (*hw_perf_counter_disable) (struct perf_counter *counter);
void (*hw_perf_counter_read) (struct perf_counter *counter);
 };

This will be useful to support assymetric hw details, and it will also
be useful to implement "software counters". (Counters that count kernel
managed sw events such as pagefaults, context-switches, wall-clock time
or task-local time.)

Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoperf counters: group counter, fixes
Ingo Molnar [Thu, 11 Dec 2008 10:26:29 +0000 (11:26 +0100)]
perf counters: group counter, fixes

Impact: bugfix

Check that a group does not span outside the context of a CPU or a task.

Also, do not allow deep recursive hierarchies.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoperf counters: add support for group counters
Ingo Molnar [Thu, 11 Dec 2008 07:38:42 +0000 (08:38 +0100)]
perf counters: add support for group counters

Impact: add group counters

This patch adds the "counter groups" abstraction.

Groups of counters behave much like normal 'single' counters, with a
few semantic and behavioral extensions on top of that.

A counter group is created by creating a new counter with the open()
syscall's group-leader group_fd file descriptor parameter pointing
to another, already existing counter.

Groups of counters are scheduled in and out in one atomic group, and
they are also roundrobin-scheduled atomically.

Counters that are member of a group can also record events with an
(atomic) extended timestamp that extends to all members of the group,
if the record type is set to PERF_RECORD_GROUP.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoperf counters: restructure the API
Ingo Molnar [Wed, 10 Dec 2008 11:33:23 +0000 (12:33 +0100)]
perf counters: restructure the API

Impact: clean up new API

Thorough cleanup of the new perf counters API, we now get clean separation
of the various concepts:

 - introduce perf_counter_hw_event to separate out the event source details

 - move special type flags into separate attributes: PERF_COUNT_NMI,
   PERF_COUNT_RAW

 - extend the type to u64 and reserve it fully to the architecture in the
   raw type case.

And make use of all these changes in the core and x86 perfcounters code.

Also change the syscall signature to:

  asmlinkage int sys_perf_counter_open(

struct perf_counter_hw_event *hw_event_uptr __user,
pid_t pid,
int cpu,
int group_fd);

( Note that group_fd is unused for now - it's reserved for the counter
  groups abstraction. )

Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoperf counters: expand use of counter->event
Thomas Gleixner [Mon, 8 Dec 2008 18:35:37 +0000 (19:35 +0100)]
perf counters: expand use of counter->event

Impact: change syscall, cleanup

Make use of the new perf_counters event type.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoperf counters: clean up 'raw' type API
Thomas Gleixner [Mon, 8 Dec 2008 18:26:59 +0000 (19:26 +0100)]
perf counters: clean up 'raw' type API

Impact: cleanup

Introduce a separate hw_event type.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoperf counters: protect them against CSTATE transitions
Thomas Gleixner [Tue, 9 Dec 2008 20:43:39 +0000 (21:43 +0100)]
perf counters: protect them against CSTATE transitions

Impact: fix rare lost events problem

There are CPUs whose performance counters misbehave on CSTATE transitions,
so provide a way to just disable/enable them around deep idle methods.

(hw_perf_enable_all() is cheap on x86.)

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoperfcounters: consolidate global-disable codepaths
Ingo Molnar [Tue, 9 Dec 2008 11:23:59 +0000 (12:23 +0100)]
perfcounters: consolidate global-disable codepaths

Impact: cleanup

Simplify global disable handling.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoperfcounters, x86: clean up debug code
Ingo Molnar [Tue, 9 Dec 2008 11:18:18 +0000 (12:18 +0100)]
perfcounters, x86: clean up debug code

Impact: cleanup

Get rid of unused debug code.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoperfcounters, x86: simplify disable/enable of counters
Ingo Molnar [Tue, 9 Dec 2008 10:40:46 +0000 (11:40 +0100)]
perfcounters, x86: simplify disable/enable of counters

Impact: fix spurious missed counter wakeups

In the case of NMI events, close a race window that can occur if an NMI
hits counter code that temporarily disables+enables a counter, and the NMI
leaks into the disabled section.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoperfcounters: select ANON_INODES
Ingo Molnar [Mon, 8 Dec 2008 18:38:33 +0000 (19:38 +0100)]
perfcounters: select ANON_INODES

The perfcounters subsystem depends on CONFIG_ANON_INODES facilities,
so make sure it's selected.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agox86, perfcounters: read out MSR_CORE_PERF_GLOBAL_STATUS with counters disabled
Ingo Molnar [Mon, 8 Dec 2008 13:20:16 +0000 (14:20 +0100)]
x86, perfcounters: read out MSR_CORE_PERF_GLOBAL_STATUS with counters disabled

Impact: make perfcounter NMI and IRQ sequence more robust

Make __smp_perf_counter_interrupt() a bit more conservative: first disable
all counters, then read out the status. Most invocations are because there
are real events, so there's no performance impact.

Code flow gets a bit simpler as well this way.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoperformance counters: x86 support
Ingo Molnar [Wed, 3 Dec 2008 09:39:53 +0000 (10:39 +0100)]
performance counters: x86 support

Implement performance counters for x86 Intel CPUs.

It's simplified right now: the PERFMON CPU feature is assumed,
which is available in Core2 and later Intel CPUs.

The design is flexible to be extended to more CPU types as well.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoperformance counters: documentation
Ingo Molnar [Thu, 4 Dec 2008 19:13:45 +0000 (20:13 +0100)]
performance counters: documentation

Add more documentation about performance counters.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoperformance counters: core code
Thomas Gleixner [Thu, 4 Dec 2008 19:12:29 +0000 (20:12 +0100)]
performance counters: core code

Implement the core kernel bits of Performance Counters subsystem.

The Linux Performance Counter subsystem provides an abstraction of
performance counter hardware capabilities. It provides per task and per
CPU counters, and it provides event capabilities on top of those.

Performance counters are accessed via special file descriptors.
There's one file descriptor per virtual counter used.

The special file descriptor is opened via the perf_counter_open()
system call:

 int
 perf_counter_open(u32 hw_event_type,
                   u32 hw_event_period,
                   u32 record_type,
                   pid_t pid,
                   int cpu);

The syscall returns the new fd. The fd can be used via the normal
VFS system calls: read() can be used to read the counter, fcntl()
can be used to set the blocking mode, etc.

Multiple counters can be kept open at a time, and the counters
can be poll()ed.

See more details in Documentation/perf-counters.txt.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoMerge branches 'x86/signal' and 'x86/irq' into perfcounters/core
Ingo Molnar [Mon, 8 Dec 2008 14:46:30 +0000 (15:46 +0100)]
Merge branches 'x86/signal' and 'x86/irq' into perfcounters/core

Merge these pending x86 tree changes into the perfcounters tree
to avoid conflicts.

15 years agox86: signal: change type of paramter for sys_rt_sigreturn()
Hiroshi Shimamoto [Sat, 6 Dec 2008 01:17:09 +0000 (17:17 -0800)]
x86: signal: change type of paramter for sys_rt_sigreturn()

Impact: cleanup on 32-bit

Peter pointed this parameter can be changed.

Signed-off-by: Hiroshi Shimamoto <h-shimamoto@ct.jp.nec.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agoFix a race condition in FASYNC handling
Jonathan Corbet [Fri, 5 Dec 2008 23:12:48 +0000 (16:12 -0700)]
Fix a race condition in FASYNC handling

Changeset a238b790d5f99c7832f9b73ac8847025815b85f7 (Call fasync()
functions without the BKL) introduced a race which could leave
file->f_flags in a state inconsistent with what the underlying
driver/filesystem believes.  Revert that change, and also fix the same
races in ioctl_fioasync() and ioctl_fionbio().

This is a minimal, short-term fix; the real fix will not involve the
BKL.

Reported-by: Oleg Nesterov <oleg@redhat.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: stable@kernel.org
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoEnforce a minimum SG_IO timeout
Linus Torvalds [Fri, 5 Dec 2008 22:49:18 +0000 (14:49 -0800)]
Enforce a minimum SG_IO timeout

There's no point in having too short SG_IO timeouts, since if the
command does end up timing out, we'll end up through the reset sequence
that is several seconds long in order to abort the command that timed
out.

As a result, shorter timeouts than a few seconds simply do not make
sense, as the recovery would be longer than the timeout itself.

Add a BLK_MIN_SG_TIMEOUT to match the existign BLK_DEFAULT_SG_TIMEOUT.

Suggested-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Acked-by: Tejun Heo <tj@kernel.org>
Acked-by: Jens Axboe <jens.axboe@oracle.com>
Cc: Jeff Garzik <jeff@garzik.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoMerge branch 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus
Linus Torvalds [Fri, 5 Dec 2008 21:42:08 +0000 (13:42 -0800)]
Merge branch 'upstream' of git://ftp.linux-mips.org/upstream-linus

* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
  MIPS: Fix incorrect use of loose in vpe.c

15 years agodrivers/message/i2o/iop.c: cleanup kerneldoc
Qinghuang Feng [Fri, 5 Dec 2008 16:35:07 +0000 (16:35 +0000)]
drivers/message/i2o/iop.c: cleanup kerneldoc

no argument named @msg in i2o_msg_get_wait(), remove it.

Signed-off-by: Qinghuang Feng <qhfeng.kernel@gmail.com>
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoFix incorrect use of loose in i2o_block.c
Nick Andrew [Fri, 5 Dec 2008 16:34:56 +0000 (16:34 +0000)]
Fix incorrect use of loose in i2o_block.c

Fix incorrect use of loose in i2o_block.c

It should be 'lose', not 'loose'.

Signed-off-by: Nick Andrew <nick@nick-andrew.net>
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoFix incorrect use of loose in tty/serial drivers
Nick Andrew [Fri, 5 Dec 2008 16:34:46 +0000 (16:34 +0000)]
Fix incorrect use of loose in tty/serial drivers

[Folded together as one diff from 3]

It should be 'lose', not 'loose'.

Signed-off-by: Nick Andrew <nick@nick-andrew.net>
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoRevert "ACPI: battery: Convert discharge energy rate to current properly"
Linus Torvalds [Fri, 5 Dec 2008 21:30:03 +0000 (13:30 -0800)]
Revert "ACPI: battery: Convert discharge energy rate to current properly"

This reverts commit 558073dd56707864f09d563b64e7c37c021e89d2, along with
the failed try to fix the regression it caused ("ACPI: Fix ACPI battery
regression introduced by commit 558073"), which just made things worse.

Commit aaad077638be1a25871bcae5e43952d6b63abfca (that failed "Fix ACPI
battery regression") got the voltage conversion confused, and fixed the
problem with Rafael's battery monitor apparently just by mistake.

So revert them both, getting us back to the 2.6.27 state in this, and
let's revisit it when people understand what's going on.

Noted-by: Paul Martin <pm@debian.org>
Requested-by: Rafael J. Wysocki <rjw@sisk.pl>
Cc: Len Brown <len.brown@intel.com>
Cc: Alexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoMIPS: Fix incorrect use of loose in vpe.c
Nick Andrew [Fri, 5 Dec 2008 00:36:54 +0000 (11:36 +1100)]
MIPS: Fix incorrect use of loose in vpe.c

It should be 'lose', not 'loose'.

Signed-off-by: Nick Andrew <nick@nick-andrew.net>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
15 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/bdev
Linus Torvalds [Fri, 5 Dec 2008 05:45:44 +0000 (21:45 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/viro/bdev

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/bdev:
  [PATCH] fix bogus argument of blkdev_put() in pktcdvd
  [PATCH 2/2] documnt FMODE_ constants
  [PATCH 1/2] kill FMODE_NDELAY_NOW
  [PATCH] clean up blkdev_get a little bit
  [PATCH] Fix block dev compat ioctl handling
  [PATCH] kill obsolete temporary comment in swsusp_close()

15 years agoMerge branch 'drm-gem-update' of git://git.kernel.org/pub/scm/linux/kernel/git/airlie...
Linus Torvalds [Fri, 5 Dec 2008 05:44:40 +0000 (21:44 -0800)]
Merge branch 'drm-gem-update' of git://git./linux/kernel/git/airlied/drm-2.6

* 'drm-gem-update' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
  drm/i915: Return error in i915_gem_set_to_gtt_domain if we're not in the GTT.
  drm/i915: Retry execbuffer pinning after clearing the GTT
  drm/i915: Move the execbuffer domain computations together
  drm/i915: Rename object_set_domain to object_set_to_gpu_domain
  drm/i915: Make a single set-to-cpu-domain path and use it wherever needed.
  drm/i915: Make a single set-to-gtt-domain path.
  drm/i915: If interrupted while setting object domains, still emit the flush.
  drm/i915: Move flushing list cleanup from flush request retire to request emit.
  drm/i915: Respect GM965/GM45 bit-17-instead-of-bit-11 option for swizzling.

15 years agoMerge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/galak/powerpc
Linus Torvalds [Fri, 5 Dec 2008 05:44:03 +0000 (21:44 -0800)]
Merge branch 'merge' of git://git./linux/kernel/git/galak/powerpc

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/galak/powerpc:
  powerpc/83xx: Enable FIXED_PHY in mpc834x_itx and mpc83xx defconfigs

15 years agoACPI: Fix ACPI battery regression introduced by commit 558073
Rafael J. Wysocki [Fri, 5 Dec 2008 00:07:51 +0000 (01:07 +0100)]
ACPI: Fix ACPI battery regression introduced by commit 558073

Commit 558073dd56707864f09d563b64e7c37c021e89d2 ("ACPI: battery: Convert
discharge energy rate to current properly") caused the battery subsystem
to report wrong values of the remaining time on battery power and the
time until fully charged on Toshiba Portege R500 (and presumably on
other boxes too).

Fix the issue by correcting the conversion from mW to mA.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoMerge branch 'timers-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 5 Dec 2008 05:40:29 +0000 (21:40 -0800)]
Merge branch 'timers-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'timers-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  time: catch xtime_nsec underflows and fix them
  posix-cpu-timers: fix clock_gettime with CLOCK_PROCESS_CPUTIME_ID

15 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Fri, 5 Dec 2008 05:40:17 +0000 (21:40 -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: Sync FPU state in VIS emulation handler.
  sparc64: Fix VIS emulation bugs
  sparc: asm/bitops.h should define __fls
  sparc64: Fix bug in PTRACE_SETFPREGS64 handling.

15 years agoMerge branch 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 5 Dec 2008 05:40:08 +0000 (21:40 -0800)]
Merge branch 'x86-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  x86: fix early panic with boot option "nosmp"
  x86/oprofile: fix Intel cpu family 6 detection
  oprofile: fix CPU unplug panic in ppro_stop()
  AMD IOMMU: fix possible race while accessing iommu->need_sync
  AMD IOMMU: set device table entry for aliased devices
  AMD IOMMU: struct amd_iommu remove padding on 64 bit
  x86: fix broken flushing in GART nofullflush path
  x86: fix dma_mapping_error for 32bit x86

15 years agoMerge branch 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 5 Dec 2008 05:39:55 +0000 (21:39 -0800)]
Merge branch 'sched-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  sched: don't export sched_mc_power_savings in laptops

15 years agoMerge branch 'core-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 5 Dec 2008 05:39:41 +0000 (21:39 -0800)]
Merge branch 'core-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'core-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  check_hung_task(): unsigned sysctl_hung_task_warnings cannot be less than 0
  documentation: local_ops fix on_each_cpu

15 years agoMerge branch 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus
Linus Torvalds [Fri, 5 Dec 2008 05:39:21 +0000 (21:39 -0800)]
Merge branch 'upstream' of git://ftp.linux-mips.org/upstream-linus

* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
  MIPS: Return ENOSYS from sys32_syscall on 64bit kernels like elsewhere.
  MIPS: 64-bit: vmsplice needs to use the compat wrapper for o32 and N32.
  MIPS: o32: Fix number of arguments to splice(2).
  MIPS: Malta: Consolidate platform device code.
  MIPS: IP22, Fulong, Malta: Update defconfigs.
  MIPS: Malta: Add back RTC support
  MIPS: Fix potential DOS by untrusted user app.

15 years agoMerge branch 'for-linus' of git://oss.sgi.com/xfs/xfs
Linus Torvalds [Fri, 5 Dec 2008 05:38:55 +0000 (21:38 -0800)]
Merge branch 'for-linus' of git://oss.sgi.com/xfs/xfs

* 'for-linus' of git://oss.sgi.com/xfs/xfs:
  [XFS] Fix hang after disallowed rename across directory quota domains

15 years ago[XFS] Fix hang after disallowed rename across directory quota domains
Dave Chinner [Wed, 3 Dec 2008 22:09:34 +0000 (09:09 +1100)]
[XFS] Fix hang after disallowed rename across directory quota domains

When project quota is active and is being used for directory tree
quota control, we disallow rename outside the current directory
tree. This requires a check to be made after all the inodes
involved in the rename are locked. We fail to unlock the inodes
correctly if we disallow the rename when the target is outside the
current directory tree. This results in a hang on the next access
to the inodes involved in failed rename.

Reported-by: Arkadiusz Miskiewicz <arekm@maven.pl>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Tested-by: Arkadiusz Miskiewicz <arekm@maven.pl>
Signed-off-by: Lachlan McIlroy <lachlan@sgi.com>
15 years agopowerpc/83xx: Enable FIXED_PHY in mpc834x_itx and mpc83xx defconfigs
Anton Vorontsov [Thu, 4 Dec 2008 17:52:31 +0000 (20:52 +0300)]
powerpc/83xx: Enable FIXED_PHY in mpc834x_itx and mpc83xx defconfigs

This is needed so that Vitesse 7385 5-port switch could work on
MPC8349E-mITX boards.

Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Acked-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years agoMIPS: Return ENOSYS from sys32_syscall on 64bit kernels like elsewhere.
David Daney [Fri, 31 Oct 2008 18:23:09 +0000 (11:23 -0700)]
MIPS: Return ENOSYS from sys32_syscall on 64bit kernels like elsewhere.

When the o32 errno was changed to ENOSYS, we forgot to update the code
for 64bit kernels.

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
Acked-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
15 years agoMIPS: 64-bit: vmsplice needs to use the compat wrapper for o32 and N32.
Ralf Baechle [Wed, 3 Dec 2008 17:03:27 +0000 (17:03 +0000)]
MIPS: 64-bit: vmsplice needs to use the compat wrapper for o32 and N32.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
15 years agoMIPS: o32: Fix number of arguments to splice(2).
Ralf Baechle [Wed, 3 Dec 2008 16:33:17 +0000 (16:33 +0000)]
MIPS: o32: Fix number of arguments to splice(2).

The syscall code was assuming splice only takes 4 arguments so no stack
arguments were being copied from the userspace stack to the kernel stack.
As the result splice was likely to fail with EINVAL.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
15 years agoMIPS: Malta: Consolidate platform device code.
Ralf Baechle [Mon, 1 Dec 2008 08:16:08 +0000 (08:16 +0000)]
MIPS: Malta: Consolidate platform device code.

After adding the RTC platform device to malta-platform.c malta-mtd.c should
get unified with the rest of the platform device code.

15 years agoMIPS: IP22, Fulong, Malta: Update defconfigs.
Ralf Baechle [Mon, 1 Dec 2008 08:09:10 +0000 (08:09 +0000)]
MIPS: IP22, Fulong, Malta: Update defconfigs.

These haven't seen much attention for too long but particularly important
enable RTC_CLASS and CONFIG_RTC_HCTOSYS so the wall clock time is set on
kernel startup.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
15 years agoMIPS: Malta: Add back RTC support
Tiejun Chen [Tue, 25 Nov 2008 08:33:20 +0000 (16:33 +0800)]
MIPS: Malta: Add back RTC support

With the conversion of MIPS to RTC_LIB the old RTC driver CONFIG_RTC became
unselectable.  Fix by setting up a platform device.  Also enable
RTC_CLASS so system time gets set from RTC on kernel initialization.

[Ralf: Original patch by Tiejun; polished nice and shiny by me]

Signed-off-by: Tiejun Chen <tiejun.chen@windriver.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
15 years agoMIPS: Fix potential DOS by untrusted user app.
Vlad Malov [Tue, 18 Nov 2008 23:05:46 +0000 (15:05 -0800)]
MIPS: Fix potential DOS by untrusted user app.

On a 64 bit kernel if an o32 syscall was made with a syscall number less
than 4000, we would read the function from outside of the bounds of the
syscall table.  This led to non-deterministic behavior including system
crashes.

While we were at it we reworked the 32 bit version as well to use fewer
instructions.  Both 32 and 64 bit versions are use the same code now.

Signed-off-by: Vlad Malov <Vlad.Malov@caviumnetworks.com>
Signed-off-by: David Daney <ddaney@caviumnetworks.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
15 years agosparc64: Sync FPU state in VIS emulation handler.
Hong H. Pham [Thu, 4 Dec 2008 17:12:57 +0000 (09:12 -0800)]
sparc64: Sync FPU state in VIS emulation handler.

Copy the FPU state to the task's thread_info->fpregs for the VIS emulation
functions to access.

Signed-off-by: Hong H. Pham <hong.pham@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agox86: fix early panic with boot option "nosmp"
Andi Kleen [Thu, 4 Dec 2008 12:33:35 +0000 (13:33 +0100)]
x86: fix early panic with boot option "nosmp"

Impact: fix boot crash with numcpus=0 on certain systems

Fix early exception in __get_smp_config with nosmp.

Bail out early when there is no MP table.

Reported-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Tested-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years ago[PATCH] fix bogus argument of blkdev_put() in pktcdvd
Al Viro [Sun, 30 Nov 2008 06:33:57 +0000 (01:33 -0500)]
[PATCH] fix bogus argument of blkdev_put() in pktcdvd

final close of ->bdev should match the initial open, i.e.
get FMODE_READ | FMODE_NDELAY; FMODE_READ|FMODE_WRITE has
been a braino.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
15 years ago[PATCH 2/2] documnt FMODE_ constants
Christoph Hellwig [Wed, 5 Nov 2008 13:58:46 +0000 (14:58 +0100)]
[PATCH 2/2] documnt FMODE_ constants

Make sure all FMODE_ constants are documents, and ensure a coherent
style for the already existing comments.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
15 years ago[PATCH 1/2] kill FMODE_NDELAY_NOW
Christoph Hellwig [Wed, 5 Nov 2008 13:58:42 +0000 (14:58 +0100)]
[PATCH 1/2] kill FMODE_NDELAY_NOW

Update FMODE_NDELAY before each ioctl call so that we can kill the
magic FMODE_NDELAY_NOW.  It would be even better to do this directly
in setfl(), but for that we'd need to have FMODE_NDELAY for all files,
not just block special files.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
15 years ago[PATCH] clean up blkdev_get a little bit
Christoph Hellwig [Wed, 5 Nov 2008 13:54:41 +0000 (14:54 +0100)]
[PATCH] clean up blkdev_get a little bit

The way the bd_claim for the FMODE_EXCL case is implemented is rather
confusing.  Clean it up to the most logical style.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
15 years ago[PATCH] Fix block dev compat ioctl handling
Andreas Schwab [Fri, 31 Oct 2008 21:39:46 +0000 (22:39 +0100)]
[PATCH] Fix block dev compat ioctl handling

Commit 33c2dca4957bd0da3e1af7b96d0758d97e708ef6 (trim file propagation
in block/compat_ioctl.c) removed the handling of some ioctls from
compat_blkdev_driver_ioctl.  That caused them to be rejected as unknown
by the compat layer.

Signed-off-by: Andreas Schwab <schwab@suse.de>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
15 years ago[PATCH] kill obsolete temporary comment in swsusp_close()
Al Viro [Sun, 30 Nov 2008 06:47:12 +0000 (01:47 -0500)]
[PATCH] kill obsolete temporary comment in swsusp_close()

it had been put there to mark the call of blkdev_put() that
needed proper argument propagated to it; later patch in the
same series had done just that.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
15 years agotime: catch xtime_nsec underflows and fix them
john stultz [Tue, 2 Dec 2008 02:34:41 +0000 (18:34 -0800)]
time: catch xtime_nsec underflows and fix them

Impact: fix time warp bug

Alex Shi, along with Yanmin Zhang have been noticing occasional time
inconsistencies recently. Through their great diagnosis, they found that
the xtime_nsec value used in update_wall_time was occasionally going
negative. After looking through the code for awhile, I realized we have
the possibility for an underflow when three conditions are met in
update_wall_time():

  1) We have accumulated a second's worth of nanoseconds, so we
     incremented xtime.tv_sec and appropriately decrement xtime_nsec.
     (This doesn't cause xtime_nsec to go negative, but it can cause it
      to be small).

  2) The remaining offset value is large, but just slightly less then
     cycle_interval.

  3) clocksource_adjust() is speeding up the clock, causing a
     corrective amount (compensating for the increase in the multiplier
     being multiplied against the unaccumulated offset value) to be
     subtracted from xtime_nsec.

This can cause xtime_nsec to underflow.

Unfortunately, since we notify the NTP subsystem via second_overflow()
whenever we accumulate a full second, and this effects the error
accumulation that has already occured, we cannot simply revert the
accumulated second from xtime nor move the second accumulation to after
the clocksource_adjust call without a change in behavior.

This leaves us with (at least) two options:

1) Simply return from clocksource_adjust() without making a change if we
   notice the adjustment would cause xtime_nsec to go negative.

This would work, but I'm concerned that if a large adjustment was needed
(due to the error being large), it may be possible to get stuck with an
ever increasing error that becomes too large to correct (since it may
always force xtime_nsec negative). This may just be paranoia on my part.

2) Catch xtime_nsec if it is negative, then add back the amount its
   negative to both xtime_nsec and the error.

This second method is consistent with how we've handled earlier rounding
issues, and also has the benefit that the error being added is always in
the oposite direction also always equal or smaller then the correction
being applied. So the risk of a corner case where things get out of
control is lessened.

This patch fixes bug 11970, as tested by Yanmin Zhang
http://bugzilla.kernel.org/show_bug.cgi?id=11970

Reported-by: alex.shi@intel.com
Signed-off-by: John Stultz <johnstul@us.ibm.com>
Acked-by: "Zhang, Yanmin" <yanmin_zhang@linux.intel.com>
Tested-by: "Zhang, Yanmin" <yanmin_zhang@linux.intel.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agosparc64: Fix VIS emulation bugs
Joseph Myers [Thu, 4 Dec 2008 03:36:05 +0000 (19:36 -0800)]
sparc64: Fix VIS emulation bugs

This patch fixes some bugs in VIS emulation that cause the GCC test
failure

FAIL: gcc.target/sparc/pdist-3.c execution test

for both 32-bit and 64-bit testing on hardware lacking these
instructions.  The emulation code for the pdist instruction uses
RS1(insn) for both source registers rs1 and rs2, which is obviously
wrong and leads to the instruction doing nothing (the observed
problem), and further inspection of the code shows that RS1 uses a
shift of 24 and RD a shift of 25, which clearly cannot both be right;
examining SPARC documentation indicates the correct shift for RS1 is
14.

This patch fixes the bug if single-stepping over the affected
instruction in the debugger, but not if the testcase is run
standalone.  For that, Wind River has another patch I hope they will
send as a followup to this patch submission.

Signed-off-by: Joseph Myers <joseph@codesourcery.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agodrm/i915: Return error in i915_gem_set_to_gtt_domain if we're not in the GTT.
Eric Anholt [Wed, 26 Nov 2008 21:58:13 +0000 (13:58 -0800)]
drm/i915: Return error in i915_gem_set_to_gtt_domain if we're not in the GTT.

It's only for flushing caches appropriately for GTT access, not for actually
getting it there.  Prevents potential smashing of cpu read/write domains on
unbound objects.

Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
15 years agodrm/i915: Retry execbuffer pinning after clearing the GTT
Keith Packard [Fri, 21 Nov 2008 07:30:27 +0000 (23:30 -0800)]
drm/i915: Retry execbuffer pinning after clearing the GTT

If we fail to pin all of the buffers in an execbuffer request, go through
and clear the GTT and try again to see if its just a matter of fragmentation

Signed-off-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
15 years agodrm/i915: Move the execbuffer domain computations together
Keith Packard [Fri, 21 Nov 2008 07:23:03 +0000 (23:23 -0800)]
drm/i915: Move the execbuffer domain computations together

This eliminates the dev_set_domain function and just in-lines it
where its used, with the goal of moving the manipulation and use of
invalidate_domains and flush_domains closer together. This also
avoids calling add_request unless some domain has been flushed.

Signed-off-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
15 years agodrm/i915: Rename object_set_domain to object_set_to_gpu_domain
Keith Packard [Fri, 21 Nov 2008 07:11:08 +0000 (23:11 -0800)]
drm/i915: Rename object_set_domain to object_set_to_gpu_domain

Now that the CPU and GTT domain operations are isolated to their own
functions, the previously general-purpose set_domain function is now used
only to set GPU domains. It also has no failure cases, which is important as
this eliminates any possible interruption of the computation of new object
domains and subsequent emmission of the flushing instructions into the ring.

Signed-off-by: Keith Packard <keithp@keithp.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
15 years agodrm/i915: Make a single set-to-cpu-domain path and use it wherever needed.
Eric Anholt [Fri, 14 Nov 2008 21:35:19 +0000 (13:35 -0800)]
drm/i915: Make a single set-to-cpu-domain path and use it wherever needed.

This fixes several domain management bugs, including potential lack of cache
invalidation for pread, potential failure to wait for set_domain(CPU, 0),
and more, along with producing more intelligible code.

Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
15 years agodrm/i915: Make a single set-to-gtt-domain path.
Eric Anholt [Mon, 10 Nov 2008 18:53:25 +0000 (10:53 -0800)]
drm/i915: Make a single set-to-gtt-domain path.

This fixes failure to flush caches in the relocation update path, and
failure to wait in the set_domain ioctl, each of which could lead to incorrect
rendering.

Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
15 years agodrm/i915: If interrupted while setting object domains, still emit the flush.
Eric Anholt [Sat, 15 Nov 2008 00:27:47 +0000 (16:27 -0800)]
drm/i915: If interrupted while setting object domains, still emit the flush.

Otherwise, we would leave the objects in an inconsistent state, such as
write_domain == 0 but on the flushing list.

Signed-off-by: Dave Airlie <airlied@redhat.com>
15 years agodrm/i915: Move flushing list cleanup from flush request retire to request emit.
Eric Anholt [Fri, 7 Nov 2008 00:00:31 +0000 (16:00 -0800)]
drm/i915: Move flushing list cleanup from flush request retire to request emit.

obj_priv->write_domain is "write domain if the GPU went idle now", not
"write domain at this moment."  By postponing the clear, we confused the
concept, required more storage, and potentially emitted more flushes than
are required.

Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
15 years agodrm/i915: Respect GM965/GM45 bit-17-instead-of-bit-11 option for swizzling.
Eric Anholt [Tue, 25 Nov 2008 22:02:05 +0000 (14:02 -0800)]
drm/i915: Respect GM965/GM45 bit-17-instead-of-bit-11 option for swizzling.

This fixes readpixels and buffer corruption when swapped out and in by
disabling tiling on them.

Now that we know that the bit 17 mode isn't just a mistake of older chipsets,
we'll need to work on a clever fix so that we can get the performance of
tiling on these chipsets, but that will require intrusive changes targeted
at the next kernel release, not this one.

Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
15 years agoMN10300: Introduce barriers to replace removed volatiles in gdbstub 16550 driver
David Howells [Wed, 3 Dec 2008 16:33:14 +0000 (16:33 +0000)]
MN10300: Introduce barriers to replace removed volatiles in gdbstub 16550 driver

Introduce into the MN10300 gdbstub 16550 driver a couple of barrier() calls to
replace the removed volatility of the input/output index variables for the Rx
ring buffer.  A previous patch added them into the on-chip serial port driver.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoMAINTAINERS: Add security subsystem maintainer
James Morris [Wed, 3 Dec 2008 16:19:45 +0000 (03:19 +1100)]
MAINTAINERS: Add security subsystem maintainer

Add myself as overall maintainer of the security subsystem (generally,
components under the top-level security directory).  This addresses
the lack of an official maintainer for the increasing number of
security projects being incorporated into the kernel.

Signed-off-by: James Morris <jmorris@namei.org>
Acked-by: Chris Wright <chrisw@sous-sol.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agoMerge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block
Linus Torvalds [Thu, 4 Dec 2008 00:45:56 +0000 (16:45 -0800)]
Merge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block

* 'for-linus' of git://git.kernel.dk/linux-2.6-block:
  block: fix setting of max_segment_size and seg_boundary mask
  block: internal dequeue shouldn't start timer
  block: set disk->node_id before it's being used
  When block layer fails to map iov, it calls bio_unmap_user to undo

15 years agoMerge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc
Linus Torvalds [Thu, 4 Dec 2008 00:41:15 +0000 (16:41 -0800)]
Merge branch 'merge' of git://git./linux/kernel/git/paulus/powerpc

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc:
  powerpc/83xx: Fix MCU support merge issue in mpc8349emitx.dts
  powerpc: Fix dma_map_sg() cache flushing on non coherent platforms

15 years agoMerge branch 'for-2.6.28' of git://linux-nfs.org/~bfields/linux
Linus Torvalds [Thu, 4 Dec 2008 00:40:37 +0000 (16:40 -0800)]
Merge branch 'for-2.6.28' of git://linux-nfs.org/~bfields/linux

* 'for-2.6.28' of git://linux-nfs.org/~bfields/linux:
  NLM: client-side nlm_lookup_host() should avoid matching on srcaddr
  nfsd: use of unitialized list head on error exit in nfs4recover.c
  Add a reference to sunrpc in svc_addsock
  nfsd: clean up grace period on early exit

15 years agoiTCO_wdt: fix typo when setting TCO_EN bit
Linus Torvalds [Thu, 4 Dec 2008 00:20:19 +0000 (16:20 -0800)]
iTCO_wdt: fix typo when setting TCO_EN bit

The code used '&= 0x00002000' when it tried to set the TCO_EN bit, which
obviously didn't set that bit at all, but instead just reset all the
other bits in the SMI_EN register.

This bug seemingly caused various random behavior, with Frans Pop
reporting that X.org just silently hung at startup and Rafael Wysocki
reports the fan spinning with full speed.

See
http://lkml.org/lkml/2008/12/3/178
http://bugzilla.kernel.org/show_bug.cgi?id=12162

The problem seems to have been triggered by "[WATCHDOG] iTCO_wdt :
problem with rebooting on new ICH9 based motherboards" (commit
7cd5b08be3c489df11b559fef210b81133764ad4), but the bogus code existed
before that too (in the "supermicro_old_pre_stop()" function), it just
apparently never showed up due to different logic.

In that commit the broken code got moved around and now gets executed
much more.

Reported-by: Rafael J. Wysocki <rjw@sisk.pl>
Tested-by: Frans Pop <elendil@planet.nl>
Cc: Wim Van Sebroeck <wim@iguana.be>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
15 years agosparc: asm/bitops.h should define __fls
Rusty Russell [Thu, 4 Dec 2008 00:04:52 +0000 (16:04 -0800)]
sparc: asm/bitops.h should define __fls

bitops_64.h includes the generic one; pretty sure 32 should too.

(Found by using __fls in generic code and breaking sparc defconfig build:
 thanks Stephen and linux-next!)

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoMerge branch 'oprofile-for-tip' of git://git.kernel.org/pub/scm/linux/kernel/git...
Ingo Molnar [Wed, 3 Dec 2008 17:52:46 +0000 (18:52 +0100)]
Merge branch 'oprofile-for-tip' of git://git./linux/kernel/git/rric/oprofile into x86/urgent

15 years agox86/oprofile: fix Intel cpu family 6 detection
William Cohen [Sun, 30 Nov 2008 20:39:10 +0000 (15:39 -0500)]
x86/oprofile: fix Intel cpu family 6 detection

Alan Jenkins wrote:
> This is on an EeePC 701, /proc/cpuinfo as attached.
>
> Is this expected?  Will the next release work?
>
> Thanks, Alan
>
> # opcontrol --setup --no-vmlinux
> cpu_type 'unset' is not valid
> you should upgrade oprofile or force the use of timer mode
>
> # opcontrol -v
> opcontrol: oprofile 0.9.4 compiled on Nov 29 2008 22:44:10
>
> # cat /dev/oprofile/cpu_type
> i386/p6
> # uname -r
> 2.6.28-rc6eeepc

Hi Alan,

Looking at the kernel driver code for oprofile it can return the "i386/p6" for
the cpu_type. However, looking at the user-space oprofile code there isn't the
matching entry in libop/op_cpu_type.c or the events/unit_mask files in
events/i386 directory.

The Intel AP-485 says this is a "Intel Pentium M processor model D". Seems like
the oprofile kernel driver should be identifying the processor as "i386/p6_mobile"

The driver identification code doesn't look quite right in nmi_init.c

http://git.kernel.org/?p=linux/kernel/git/sfr/linux-next.git;a=blob;f=arch/x86/oprofile/nmi_int.c;h=022cd41ea9b4106e5884277096e80e9088a7c7a9;hb=HEAD

has:

409         case 10 ... 13:
410                 *cpu_type = "i386/p6";
411                 break;

Referring to the Intel AP-485:
case 10 and 11 should produce "i386/piii"
case 13 should produce "i386/p6_mobile"

I didn't see anything for case 12.

Something like the attached patch. I don't have a celeron machine to verify that
changes in this area of the kernel fix thing.

-Will

Signed-off-by: William Cohen <wcohen@redhat.com>
Tested-by: Alan Jenkins <alan-jenkins@tuffmail.co.uk>
Acked-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Robert Richter <robert.richter@amd.com>
15 years agopowerpc/83xx: Fix MCU support merge issue in mpc8349emitx.dts
Anton Vorontsov [Thu, 27 Nov 2008 17:36:45 +0000 (20:36 +0300)]
powerpc/83xx: Fix MCU support merge issue in mpc8349emitx.dts

Just found the merge issue in 442746989d92afc125040e0f29b33602ad94da99
("powerpc/83xx: Add support for MCU microcontroller in .dts files"):
the commit adds the MCU controller node into the DMA node, which is
wrong because the MCU sits on the I2C bus. Fix this by moving the MCU
node into the I2C controller node.

The original patch[1] was OK though. ;-)

Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years agooprofile: fix CPU unplug panic in ppro_stop()
Eric Dumazet [Tue, 2 Dec 2008 06:21:21 +0000 (07:21 +0100)]
oprofile: fix CPU unplug panic in ppro_stop()

If oprofile statically compiled in kernel, a cpu unplug triggers
a panic in ppro_stop(), because a NULL pointer is dereferenced.

Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: Robert Richter <robert.richter@amd.com>
15 years agoblock: fix setting of max_segment_size and seg_boundary mask
Milan Broz [Wed, 3 Dec 2008 11:55:08 +0000 (12:55 +0100)]
block: fix setting of max_segment_size and seg_boundary mask

Fix setting of max_segment_size and seg_boundary mask for stacked md/dm
devices.

When stacking devices (LVM over MD over SCSI) some of the request queue
parameters are not set up correctly in some cases by default, namely
max_segment_size and and seg_boundary mask.

If you create MD device over SCSI, these attributes are zeroed.

Problem become when there is over this mapping next device-mapper mapping
- queue attributes are set in DM this way:

request_queue   max_segment_size  seg_boundary_mask
SCSI                65536             0xffffffff
MD RAID1                0                      0
LVM                 65536                 -1 (64bit)

Unfortunately bio_add_page (resp.  bio_phys_segments) calculates number of
physical segments according to these parameters.

During the generic_make_request() is segment cout recalculated and can
increase bio->bi_phys_segments count over the allowed limit.  (After
bio_clone() in stack operation.)

Thi is specially problem in CCISS driver, where it produce OOPS here

    BUG_ON(creq->nr_phys_segments > MAXSGENTRIES);

(MAXSEGENTRIES is 31 by default.)

Sometimes even this command is enough to cause oops:

  dd iflag=direct if=/dev/<vg>/<lv> of=/dev/null bs=128000 count=10

This command generates bios with 250 sectors, allocated in 32 4k-pages
(last page uses only 1024 bytes).

For LVM layer, it allocates bio with 31 segments (still OK for CCISS),
unfortunatelly on lower layer it is recalculated to 32 segments and this
violates CCISS restriction and triggers BUG_ON().

The patch tries to fix it by:

 * initializing attributes above in queue request constructor
   blk_queue_make_request()

 * make sure that blk_queue_stack_limits() inherits setting

 (DM uses its own function to set the limits because it
 blk_queue_stack_limits() was introduced later.  It should probably switch
 to use generic stack limit function too.)

 * sets the default seg_boundary value in one place (blkdev.h)

 * use this mask as default in DM (instead of -1, which differs in 64bit)

Bugs related to this:
https://bugzilla.redhat.com/show_bug.cgi?id=471639
http://bugzilla.kernel.org/show_bug.cgi?id=8672

Signed-off-by: Milan Broz <mbroz@redhat.com>
Reviewed-by: Alasdair G Kergon <agk@redhat.com>
Cc: Neil Brown <neilb@suse.de>
Cc: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Cc: Tejun Heo <htejun@gmail.com>
Cc: Mike Miller <mike.miller@hp.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
15 years agoMerge branch 'iommu-fixes-2.6.28' of git://git.kernel.org/pub/scm/linux/kernel/git...
Ingo Molnar [Wed, 3 Dec 2008 11:54:45 +0000 (12:54 +0100)]
Merge branch 'iommu-fixes-2.6.28' of git://git./linux/kernel/git/joro/linux-2.6-iommu into x86/urgent

15 years agoblock: internal dequeue shouldn't start timer
Tejun Heo [Wed, 3 Dec 2008 11:41:26 +0000 (12:41 +0100)]
block: internal dequeue shouldn't start timer

blkdev_dequeue_request() and elv_dequeue_request() are equivalent and
both start the timeout timer.  Barrier code dequeues the original
barrier request but doesn't passes the request itself to lower level
driver, only broken down proxy requests; however, as the original
barrier code goes through the same dequeue path and timeout timer is
started on it.  If barrier sequence takes long enough, this timer
expires but the low level driver has no idea about this request and
oops follows.

Timeout timer shouldn't have been started on the original barrier
request as it never goes through actual IO.  This patch unexports
elv_dequeue_request(), which has no external user anyway, and makes it
operate on elevator proper w/o adding the timer and make
blkdev_dequeue_request() call elv_dequeue_request() and add timer.
Internal users which don't pass the request to driver - barrier code
and end_that_request_last() - are converted to use
elv_dequeue_request().

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Mike Anderson <andmike@linux.vnet.ibm.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
15 years agoblock: set disk->node_id before it's being used
Cheng Renquan [Thu, 20 Nov 2008 07:37:37 +0000 (08:37 +0100)]
block: set disk->node_id before it's being used

disk->node_id will be refered in allocating in disk_expand_part_tbl, so we
should set it before disk->node_id is refered.

Signed-off-by: Cheng Renquan <crquan@gmail.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
15 years agoWhen block layer fails to map iov, it calls bio_unmap_user to undo
Petr Vandrovec [Wed, 19 Nov 2008 10:12:14 +0000 (11:12 +0100)]
When block layer fails to map iov, it calls bio_unmap_user to undo
mapping.  Which is good if pages were mapped - but if they were provided
by someone else and just copied then bad things happen - pages are
released once here, and once by caller, leading to user triggerable BUG
at include/linux/mm.h:246.

Signed-off-by: Petr Vandrovec <petr@vandrovec.name>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
15 years agoAMD IOMMU: fix possible race while accessing iommu->need_sync
Joerg Roedel [Wed, 3 Dec 2008 11:19:27 +0000 (12:19 +0100)]
AMD IOMMU: fix possible race while accessing iommu->need_sync

The access to the iommu->need_sync member needs to be protected by the
iommu->lock. Otherwise this is a possible race condition. Fix it with
this patch.

Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
15 years agoAMD IOMMU: set device table entry for aliased devices
Joerg Roedel [Tue, 25 Nov 2008 11:56:12 +0000 (12:56 +0100)]
AMD IOMMU: set device table entry for aliased devices

In some rare cases a request can arrive an IOMMU with its originial
requestor id even it is aliased. Handle this by setting the device table
entry to the same protection domain for the original and the aliased
requestor id.

Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
15 years agoAMD IOMMU: struct amd_iommu remove padding on 64 bit
Richard Kennedy [Mon, 24 Nov 2008 13:53:24 +0000 (13:53 +0000)]
AMD IOMMU: struct amd_iommu remove padding on 64 bit

Remove 16 bytes of padding from struct amd_iommu on 64bit builds
reducing its size to 120 bytes, allowing it to span one fewer
cachelines.

Signed-off-by: Richard Kennedy <richard@rsk.demon.co.uk>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
15 years agocheck_hung_task(): unsigned sysctl_hung_task_warnings cannot be less than 0
Roel Kluin [Tue, 2 Dec 2008 21:55:38 +0000 (22:55 +0100)]
check_hung_task(): unsigned sysctl_hung_task_warnings cannot be less than 0

Impact: fix warnings-limit cutoff check for debug feature

unsigned sysctl_hung_task_warnings cannot be less than 0

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agox86: fix broken flushing in GART nofullflush path
Joerg Roedel [Tue, 2 Dec 2008 19:16:03 +0000 (20:16 +0100)]
x86: fix broken flushing in GART nofullflush path

Impact: remove stale IOTLB entries

In the non-default nofullflush case the GART is only flushed when
next_bit wraps around. But it can happen that an unmap operation unmaps
memory which is behind the current next_bit location. If these addresses
are reused it may result in stale GART IO/TLB entries. Fix this by
setting the GART next_bit always behind an unmapped location.

Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
15 years agosparc64: Fix bug in PTRACE_SETFPREGS64 handling.
Chris Torek [Wed, 3 Dec 2008 08:47:28 +0000 (00:47 -0800)]
sparc64: Fix bug in PTRACE_SETFPREGS64 handling.

From: Chris Torek <chris.torek@windriver.com>

>The SPARC64 kernel code for PTRACE_SETFPREGS64 appears to be an exact copy
>of that for PTRACE_GETFPREGS64.  This means that gdbserver and native
>64-bit GDB cannot set floating-point registers.

It looks like a simple typo.

Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agopowerpc: Fix dma_map_sg() cache flushing on non coherent platforms
Benjamin Herrenschmidt [Sun, 30 Nov 2008 18:53:40 +0000 (18:53 +0000)]
powerpc: Fix dma_map_sg() cache flushing on non coherent platforms

On PowerPC 4xx or other non cache-coherent platforms, we lost the
appropriate cache flushing in dma_map_sg() when merging the 32 and
64-bit DMA code (commit 4fc665b88a79a45bae8bbf3a05563c27c7337c3d,
"powerpc: Merge 32 and 64-bit dma code").  This restores it.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Becky Bruce <beckyb@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
15 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/wim/linux-2.6-watchdog
Linus Torvalds [Tue, 2 Dec 2008 23:58:20 +0000 (15:58 -0800)]
Merge git://git./linux/kernel/git/wim/linux-2.6-watchdog

* git://git.kernel.org/pub/scm/linux/kernel/git/wim/linux-2.6-watchdog:
  [WATCHDOG] hpwdt: Fix kdump when using hpwdt
  [WATCHDOG] hpwdt: set the mapped BIOS address space as executable
  [WATCHDOG] iTCO_wdt: add PCI ID's for ICH9 & ICH10 chipsets
  [WATCHDOG] iTCO_wdt : correct status clearing
  [WATCHDOG] iTCO_wdt : problem with rebooting on new ICH9 based motherboards
  [WATCHDOG] fix mtx1_wdt compilation failure

15 years agoMerge branch 'linux-next' of git://git.infradead.org/ubifs-2.6
Linus Torvalds [Tue, 2 Dec 2008 23:56:55 +0000 (15:56 -0800)]
Merge branch 'linux-next' of git://git.infradead.org/ubifs-2.6

* 'linux-next' of git://git.infradead.org/ubifs-2.6:
  UBIFS: pre-allocate bulk-read buffer
  UBIFS: do not allocate too much
  UBIFS: do not print scary memory allocation warnings
  UBIFS: allow for gaps when dirtying the LPT
  UBIFS: fix compilation warnings
  MAINTAINERS: change UBI/UBIFS git tree URLs
  UBIFS: endian handling fixes and annotations
  UBIFS: remove printk

15 years agoMerge branch 'kvm-updates/2.6.28' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Tue, 2 Dec 2008 23:56:17 +0000 (15:56 -0800)]
Merge branch 'kvm-updates/2.6.28' of git://git./linux/kernel/git/avi/kvm

* 'kvm-updates/2.6.28' of git://git.kernel.org/pub/scm/linux/kernel/git/avi/kvm:
  KVM: MMU: avoid creation of unreachable pages in the shadow
  KVM: ppc: stop leaking host memory on VM exit
  KVM: MMU: fix sync of ptes addressed at owner pagetable
  KVM: ia64: Fix: Use correct calling convention for PAL_VPS_RESUME_HANDLER
  KVM: ia64: Fix incorrect kbuild CFLAGS override
  KVM: VMX: Fix interrupt loss during race with NMI
  KVM: s390: Fix problem state handling in guest sigp handler

15 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Tue, 2 Dec 2008 23:55:43 +0000 (15:55 -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 offset calculation in compute_size()
  rtc: rtc-starfire fixes

15 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Tue, 2 Dec 2008 23:55:05 +0000 (15:55 -0800)]
Merge git://git./linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (30 commits)
  MAINTAINERS: add netdev to ATM
  ATM: horizon, fix hrz_probe fail path
  pppol2tp: Add missing sock_put() in pppol2tp_release()
  net: Fix soft lockups/OOM issues w/ unix garbage collector
  macvlan: don't broadcast PAUSE frames to macvlan devices
  Phonet: fix oops in phonet_address_del() on non-Phonet device
  netfilter: ctnetlink: fix GFP_KERNEL allocation under spinlock
  sungem: Fix PCS_MIICTRL register write in gem_init_phy().
  net: make skb_truesize_bug() call WARN()
  net: hp-plus uses eip_poll
  net/wireless/reg.c: fix bad WARN_ON in if statement
  ath5k: disable beacon filter when station is not associated
  ath5k: fix Security issue in DebugFS part of ath5k
  ath9k: correct expected max RX buffer size
  ath9k: Fix SW-IOMMU bounce buffer starvation
  mac80211 : Fix setting ad-hoc mode and non-ibss channel
  iwlagn: fix DMA sync
  phylib: Add Vitesse VSC8221 SGMII PHY
  rose: zero length frame filtering in af_rose.c
  bridge: netfilter: fix update_pmtu crash with GRE
  ...

15 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux...
Linus Torvalds [Tue, 2 Dec 2008 23:53:41 +0000 (15:53 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/geert/linux-m68k

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k:
  m68k: Update defconfigs for 2.6.28-rc7
  macfb: Do not overflow fb_fix_screeninfo.id

15 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/bart/ide-2.6
Linus Torvalds [Tue, 2 Dec 2008 23:53:10 +0000 (15:53 -0800)]
Merge git://git./linux/kernel/git/bart/ide-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/bart/ide-2.6:
  alim15x3: fix sparse warning
  ide: remove dead code from drive_is_ready()
  ide: fix build for DEBUG_PM
  ide: respect current DMA setting during resume
  ide: add SAMSUNG SP0822N with firmware WA100-10 to ivb_list[]
  amd74xx: workaround unreliable AltStatus register for nVidia controllers
  ide: fix the ide_release_lock imbalance

15 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6
Linus Torvalds [Tue, 2 Dec 2008 23:52:28 +0000 (15:52 -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] stex: switch to block timeout
  [SCSI] make scsi_eh_try_stu use block timeout
  [SCSI] megaraid_sas: switch to block timeout
  [SCSI] ibmvscsi: switch to block timeout
  [SCSI] aacraid: switch to block timeout
  [SCSI] zfcp: prevent double decrement on host_busy while being busy
  [SCSI] zfcp: fix deadlock between wq triggered port scan and ERP
  [SCSI] zfcp: eliminate race between validation and locking
  [SCSI] zfcp: verify for correct rport state before scanning for SCSI devs
  [SCSI] zfcp: returning an ERR_PTR where a NULL value is expected
  [SCSI] zfcp: Fix opening of wka ports
  [SCSI] zfcp: fix remote port status check
  [SCSI] fc_transport: fix old bug on bitflag definitions
  [SCSI] Fix hang in starved list processing

15 years agoMN10300: Fix application of kernel module relocations
Mark Salter [Tue, 2 Dec 2008 14:38:09 +0000 (14:38 +0000)]
MN10300: Fix application of kernel module relocations

This fixes the MN10300 kernel module linking to match the toolchain.  RELA
relocs don't use the value at the location being relocated.  This has been
working because the tools always leave the value at the target location
cleared.

Signed-off-by: Mark Salter <msalter@redhat.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>