GitHub/exynos8895/android_kernel_samsung_universal8895.git
14 years agoperf record: Stop intercepting events, use postprocessing to get build-ids
Arnaldo Carvalho de Melo [Wed, 3 Feb 2010 18:52:05 +0000 (16:52 -0200)]
perf record: Stop intercepting events, use postprocessing to get build-ids

We want to stream events as fast as possible to perf.data, and
also in the future we want to have splice working, when no
interception will be possible.

Using build_id__mark_dso_hit_ops to create the list of DSOs that
back MMAPs we also optimize disk usage in the build-id cache by
only caching DSOs that had hits.

Suggested-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1265223128-11786-6-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf build-id: Move the routine to find DSOs with hits to the lib
Arnaldo Carvalho de Melo [Wed, 3 Feb 2010 18:52:04 +0000 (16:52 -0200)]
perf build-id: Move the routine to find DSOs with hits to the lib

Because 'perf record' will have to find the build-ids in after
we stop recording, so as to reduce even more the impact in the
workload while we do the measurement.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1265223128-11786-5-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf probe: Don't use a perf_session instance just to resolve symbols
Arnaldo Carvalho de Melo [Wed, 3 Feb 2010 18:52:03 +0000 (16:52 -0200)]
perf probe: Don't use a perf_session instance just to resolve symbols

With the recent modifications done to untie the session and
symbol layers, 'perf probe' now can use just the symbols layer.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Acked-by: Masami Hiramatsu <mhiramat@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Masami Hiramatsu <mhiramat@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf symbols: Ditch vdso global variable
Arnaldo Carvalho de Melo [Wed, 3 Feb 2010 18:52:02 +0000 (16:52 -0200)]
perf symbols: Ditch vdso global variable

We can check using strcmp, most DSOs don't start with '[' so the
test is cheap enough and we had to test it there anyway since
when reading perf.data files we weren't calling the routine that
created this global variable and thus weren't setting it as
"loaded", which was causing a bogus:

  Failed to open [vdso], continuing without symbols

Message as the first line of 'perf report'.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1265223128-11786-3-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf symbols: Fixup vsyscall maps
Arnaldo Carvalho de Melo [Wed, 3 Feb 2010 18:52:01 +0000 (16:52 -0200)]
perf symbols: Fixup vsyscall maps

While debugging a problem reported by Pekka Enberg by printing
the IP and all the maps for a thread when we don't find a map
for an IP I noticed that dso__load_sym needs to fixup these
extra maps it creates to hold symbols in different ELF sections
than the main kernel one.

Now we're back showing things like:

[root@doppio linux-2.6-tip]# perf report | grep vsyscall
     0.02%             mutt  [kernel.kallsyms].vsyscall_fn  [.] vread_hpet
     0.01%            named  [kernel.kallsyms].vsyscall_fn  [.] vread_hpet
     0.01%   NetworkManager  [kernel.kallsyms].vsyscall_fn  [.] vread_hpet
     0.01%         gconfd-2  [kernel.kallsyms].vsyscall_0   [.] vgettimeofday
     0.01%  hald-addon-rfki  [kernel.kallsyms].vsyscall_fn  [.] vread_hpet
     0.00%      dbus-daemon  [kernel.kallsyms].vsyscall_fn  [.] vread_hpet
[root@doppio linux-2.6-tip]#

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1265223128-11786-2-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf symbols: Remove perf_session usage in symbols layer
Arnaldo Carvalho de Melo [Wed, 3 Feb 2010 18:52:00 +0000 (16:52 -0200)]
perf symbols: Remove perf_session usage in symbols layer

I noticed while writing the first test in 'perf regtest' that to
just test the symbol handling routines one needs to create a
perf session, that is a layer centered on a perf.data file,
events, etc, so I untied these layers.

This reduces the complexity for the users as the number of
parameters to most of the symbols and session APIs now was
reduced while not adding more state to all the map instances by
only having data that is needed to split the kernel (kallsyms
and ELF symtab sections) maps and do vmlinux relocation on the
main kernel map.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1265223128-11786-1-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf tools: Use O_LARGEFILE to open perf data file
Xiao Guangrong [Wed, 3 Feb 2010 03:53:14 +0000 (11:53 +0800)]
perf tools: Use O_LARGEFILE to open perf data file

Open perf data file with O_LARGEFILE flag since its size is
easily larger that 2G.

For example:

 # rm -rf perf.data
 # ./perf kmem record sleep 300

 [ perf record: Woken up 0 times to write data ]
 [ perf record: Captured and wrote 3142.147 MB perf.data
 (~137282513 samples) ]

 # ll -h perf.data
 -rw------- 1 root root 3.1G .....

Signed-off-by: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
LKML-Reference: <4B68F32A.9040203@cn.fujitsu.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf lock: Clean up various details
Ingo Molnar [Sun, 31 Jan 2010 07:27:58 +0000 (08:27 +0100)]
perf lock: Clean up various details

Fix up a few small stylistic details:

 - use consistent vertical spacing/alignment
 - remove line80 artifacts
 - group some global variables better
 - remove dead code

Plus rename 'prof' to 'report' to make it more in line with other
tools, and remove the line/file keying as we really want to use
IPs like the other tools do.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <1264851813-8413-12-git-send-email-mitake@dcl.info.waseda.ac.jp>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf lock: Introduce new tool "perf lock", for analyzing lock statistics
Hitoshi Mitake [Sat, 30 Jan 2010 11:43:33 +0000 (20:43 +0900)]
perf lock: Introduce new tool "perf lock", for analyzing lock statistics

Adding new subcommand "perf lock" to perf.

I have a lot of remaining ToDos, but for now perf lock can
already provide minimal functionality for analyzing lock
statistics.

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <1264851813-8413-12-git-send-email-mitake@dcl.info.waseda.ac.jp>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf lock: Enhance information of lock trace events
Hitoshi Mitake [Sat, 30 Jan 2010 11:43:32 +0000 (20:43 +0900)]
perf lock: Enhance information of lock trace events

Add wait time and lock identification details.

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <1264851813-8413-11-git-send-email-mitake@dcl.info.waseda.ac.jp>
[ removed the file/line bits as we can do that better via IPs ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf: Add util/include/linuxhash.h to include hash.h of kernel
Hitoshi Mitake [Sat, 30 Jan 2010 11:43:24 +0000 (20:43 +0900)]
perf: Add util/include/linuxhash.h to include hash.h of kernel

linux/hash.h, hash header of kernel, is also useful for perf.

util/include/linuxhash.h includes linux/hash.h, so we can use
hash facilities (e.g. hash_long()) in perf now.

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <1264851813-8413-3-git-send-email-mitake@dcl.info.waseda.ac.jp>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf tools: Add __data_loc support
Hitoshi Mitake [Sat, 30 Jan 2010 11:43:23 +0000 (20:43 +0900)]
perf tools: Add __data_loc support

This patch is required to test the next patch for perf lock.

At 064739bc4b3d7f424b2f25547e6611bcf0132415 ,
support for the modifier "__data_loc" of format is added.

But, when I wanted to parse format of lock_acquired (or some
event else), raw_field_ptr() did not returned correct pointer.

So I modified raw_field_ptr() like this patch. Then
raw_field_ptr() works well.

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Acked-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Tom Zanussi <tzanussi@gmail.com>
Cc: Steven Rostedt <srostedt@redhat.com>
LKML-Reference: <1264851813-8413-2-git-send-email-mitake@dcl.info.waseda.ac.jp>
[ v3: fixed minor stylistic detail ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoRevert "perf record: Intercept all events"
Hitoshi Mitake [Sat, 30 Jan 2010 11:55:41 +0000 (20:55 +0900)]
Revert "perf record: Intercept all events"

This reverts commit f5a2c3dce03621b55f84496f58adc2d1a87ca16f.

This patch is required for making "perf lock rec" work.
The commit f5a2c3dce0 changes write_event() of builtin-record.c
. And changed write_event() sometimes doesn't stop with perf
lock rec.

Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <new-submission>
[ that commit also causes perf record to not be Ctrl-C-able,
  and it's concetually wrong to parse the data at record time
  (unconditionally - even when not needed), as we eventually
  want to be able to do zero-copy recording, at least for
  non-archive recordings.  ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf: Ignore perf-archive temp file
John Kacur [Wed, 27 Jan 2010 23:05:54 +0000 (21:05 -0200)]
perf: Ignore perf-archive temp file

Tell git to ignore perf-archive.

Signed-off-by: John Kacur <jkacur@redhat.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <1264633557-17597-6-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agotools/perf/perf.c: Clean up trivial style issues
Thiago Farina [Wed, 27 Jan 2010 23:05:55 +0000 (21:05 -0200)]
tools/perf/perf.c: Clean up trivial style issues

Checked with:
./../scripts/checkpatch.pl --terse --file perf.c

 perf.c: 51: ERROR: open brace '{' following function declarations go on the next line
 perf.c: 73: ERROR: "foo*** bar" should be "foo ***bar"
 perf.c:112: ERROR: space prohibited before that close parenthesis ')'
 perf.c:127: ERROR: space prohibited before that close parenthesis ')'
 perf.c:171: ERROR: "foo** bar" should be "foo **bar"
 perf.c:213: ERROR: "(foo*)" should be "(foo *)"
 perf.c:216: ERROR: "(foo*)" should be "(foo *)"
 perf.c:217: ERROR: space required before that '*' (ctx:OxV)
 perf.c:452: ERROR: do not initialise statics to 0 or NULL
 perf.c:453: ERROR: do not initialise statics to 0 or NULL

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Masami Hiramatsu <mhiramat@redhat.com>
LKML-Reference: <1264633557-17597-7-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoMerge branch 'perf/urgent' into perf/core
Ingo Molnar [Fri, 29 Jan 2010 08:24:57 +0000 (09:24 +0100)]
Merge branch 'perf/urgent' into perf/core

Merge reason: We want to queue up a dependent patch. Also update to
              later -rc's.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf session: Create kernel maps in the constructor
Arnaldo Carvalho de Melo [Wed, 27 Jan 2010 23:05:52 +0000 (21:05 -0200)]
perf session: Create kernel maps in the constructor

Removing one extra step needed in the tools that need this,
fixing a bug in 'perf probe' where this was not being done.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Masami Hiramatsu <mhiramat@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1264633557-17597-4-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf symbols: Split helpers used when creating kernel dso object
Arnaldo Carvalho de Melo [Wed, 27 Jan 2010 23:05:51 +0000 (21:05 -0200)]
perf symbols: Split helpers used when creating kernel dso object

To make it clear and allow for direct usage by, for instance,
regression test suites.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1264633557-17597-3-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf symbols: Factor out dso__load_vmlinux_path()
Arnaldo Carvalho de Melo [Wed, 27 Jan 2010 23:05:50 +0000 (21:05 -0200)]
perf symbols: Factor out dso__load_vmlinux_path()

So that we can call it directly from regression tests, and also
to reduce the size of dso__load_kernel_sym(), making it more
clear.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1264633557-17597-2-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf top: Exit if specified --vmlinux can't be used
Arnaldo Carvalho de Melo [Wed, 27 Jan 2010 23:05:49 +0000 (21:05 -0200)]
perf top: Exit if specified --vmlinux can't be used

As we do lazy loading of symtabs we only will know if the
specified vmlinux file is invalid when we actually have a hit in
kernel space and then try to load it. So if we get kernel hits
and there are _no_ symbols in the DSO backing the kernel map,
bail out.

Reported-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1264633557-17597-1-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf_events: Fix sample_period transfer on inherit
Peter Zijlstra [Fri, 29 Jan 2010 08:04:26 +0000 (09:04 +0100)]
perf_events: Fix sample_period transfer on inherit

One problem with frequency driven counters is that we cannot
predict the rate at which they trigger, therefore we have to
start them at period=1, this causes a ramp up effect. However,
if we fail to propagate the stable state on fork each new child
will have to ramp up again. This can lead to significant
artifacts in sample data.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: eranian@google.com
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <1264752266.4283.2121.camel@laptop>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf_events, x86: Remove spurious counter reset from x86_pmu_enable()
Peter Zijlstra [Wed, 27 Jan 2010 22:07:49 +0000 (23:07 +0100)]
perf_events, x86: Remove spurious counter reset from x86_pmu_enable()

At enable time the counter might still have a ->idx pointing to
a previously occupied location that might now be taken by
another event. Resetting the counter at that location with data
from this event will destroy the other counter's count.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
LKML-Reference: <20100127221122.261477183@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf_events, x86: Implement Intel Westmere support
Peter Zijlstra [Wed, 27 Jan 2010 22:07:48 +0000 (23:07 +0100)]
perf_events, x86: Implement Intel Westmere support

The new Intel documentation includes Westmere arch specific
event maps that are significantly different from the Nehalem
ones. Add support for this generation.

Found the CPUID model numbers on wikipedia.

Also ammend some Nehalem constraints, spotted those when looking
for the differences between Nehalem and Westmere.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Arjan van de Ven <arjan@linux.intel.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Stephane Eranian <eranian@google.com>
LKML-Reference: <20100127221122.151865645@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf_events, x86: Clean up hw_perf_*_all() implementation
Peter Zijlstra [Wed, 27 Jan 2010 22:07:47 +0000 (23:07 +0100)]
perf_events, x86: Clean up hw_perf_*_all() implementation

Put the recursion avoidance code in the generic hook instead of
replicating it in each implementation.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
LKML-Reference: <20100127221122.057507285@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf_events, x86: Fix event constraint masks
Peter Zijlstra [Wed, 27 Jan 2010 22:07:46 +0000 (23:07 +0100)]
perf_events, x86: Fix event constraint masks

Since constraints are specified on the event number, not number
and unit mask shorten the constraint masks so that we'll
actually match something.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
LKML-Reference: <20100127221121.967610372@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf_event: x86: Deduplicate the disable code
Peter Zijlstra [Mon, 25 Jan 2010 14:58:43 +0000 (15:58 +0100)]
perf_event: x86: Deduplicate the disable code

Share the meat of the x86_pmu_disable() code with hw_perf_enable().

Also remove the barrier() from that code, since I could not convince
myself we actually need it.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf, x86: Clean up event constraints code a bit
Ingo Molnar [Wed, 27 Jan 2010 07:39:39 +0000 (08:39 +0100)]
perf, x86: Clean up event constraints code a bit

- Remove stray debug code
 - Improve ugly macros a bit
 - Remove some whitespace damage
 - (Also fix up some accumulated damage in perf_event.h)

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Stephane Eranian <eranian@google.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>

14 years agoperf_event: x86: Optimize x86_pmu_disable()
Peter Zijlstra [Mon, 25 Jan 2010 10:57:25 +0000 (11:57 +0100)]
perf_event: x86: Optimize x86_pmu_disable()

x86_pmu_disable() removes the event from the cpuc->event_list[], however
since an event can only be on that list once, stop looking after we found
it.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf_event: x86: Optimize the fast path a little more
Peter Zijlstra [Fri, 22 Jan 2010 15:40:12 +0000 (16:40 +0100)]
perf_event: x86: Optimize the fast path a little more

Remove num from the fast path and save a few ops.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
LKML-Reference: <20100122155536.056430539@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf_event: x86: Optimize constraint weight computation
Peter Zijlstra [Fri, 22 Jan 2010 15:32:17 +0000 (16:32 +0100)]
perf_event: x86: Optimize constraint weight computation

Add a weight member to the constraint structure and avoid recomputing the
weight at runtime.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
LKML-Reference: <20100122155535.963944926@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf_event: x86: Optimize the constraint searching bits
Peter Zijlstra [Fri, 22 Jan 2010 15:32:17 +0000 (16:32 +0100)]
perf_event: x86: Optimize the constraint searching bits

Instead of copying bitmasks around, pass pointers to the constraint
structure.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
LKML-Reference: <20100122155535.887853503@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agobitops: Provide compile time HWEIGHT{8,16,32,64}
Peter Zijlstra [Fri, 22 Jan 2010 14:59:29 +0000 (15:59 +0100)]
bitops: Provide compile time HWEIGHT{8,16,32,64}

Provide compile time versions of hweight.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
LKML-Reference: <20100122155535.797688466@chello.nl>
[ Remove some whitespace damage while we are at it ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf_event: x86: Reduce some overly long lines with some MACROs
Peter Zijlstra [Fri, 22 Jan 2010 14:38:26 +0000 (15:38 +0100)]
perf_event: x86: Reduce some overly long lines with some MACROs

Introduce INTEL_EVENT_CONSTRAINT and FIXED_EVENT_CONSTRAINT to reduce
some line length and typing work.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
LKML-Reference: <20100122155535.688730371@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf_event: x86: Clean up some of the u64/long bitmask casting
Peter Zijlstra [Fri, 22 Jan 2010 14:25:59 +0000 (15:25 +0100)]
perf_event: x86: Clean up some of the u64/long bitmask casting

We need this to be u64 for direct assigment, but the bitmask functions
all work on unsigned long, leading to cast heaven, solve this by using a
union.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
LKML-Reference: <20100122155535.595961269@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf_event: x86: Fixup constraints typing issue
Peter Zijlstra [Fri, 22 Jan 2010 13:55:22 +0000 (14:55 +0100)]
perf_event: x86: Fixup constraints typing issue

Constraints gets defined an u64 but in long quantities and then cast to
long.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
LKML-Reference: <20100122155535.504916780@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf_event: x86: Allocate the fake_cpuc
Peter Zijlstra [Fri, 22 Jan 2010 13:35:46 +0000 (14:35 +0100)]
perf_event: x86: Allocate the fake_cpuc

GCC was complaining the stack usage was too large, so allocate the
structure.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
LKML-Reference: <20100122155535.411197266@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf_events: Add fast-path to the rescheduling code
Stephane Eranian [Thu, 21 Jan 2010 15:39:01 +0000 (17:39 +0200)]
perf_events: Add fast-path to the rescheduling code

Implement correct fastpath scheduling, i.e., reuse previous assignment.

Signed-off-by: Stephane Eranian <eranian@google.com>
[ split from larger patch]
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <4b588464.1818d00a.4456.383b@mx.google.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf_events, x86: Improve x86 event scheduling
Stephane Eranian [Mon, 18 Jan 2010 08:58:01 +0000 (10:58 +0200)]
perf_events, x86: Improve x86 event scheduling

This patch improves event scheduling by maximizing the use of PMU
registers regardless of the order in which events are created in a group.

The algorithm takes into account the list of counter constraints for each
event. It assigns events to counters from the most constrained, i.e.,
works on only one counter, to the least constrained, i.e., works on any
counter.

Intel Fixed counter events and the BTS special event are also handled via
this algorithm which is designed to be fairly generic.

The patch also updates the validation of an event to use the scheduling
algorithm. This will cause early failure in perf_event_open().

The 2nd version of this patch follows the model used by PPC, by running
the scheduling algorithm and the actual assignment separately. Actual
assignment takes place in hw_perf_enable() whereas scheduling is
implemented in hw_perf_group_sched_in() and x86_pmu_enable().

Signed-off-by: Stephane Eranian <eranian@google.com>
[ fixup whitespace and style nits as well as adding is_x86_event() ]
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <4b5430c6.0f975e0a.1bf9.ffff85fe@mx.google.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agox86/hw-breakpoints: Optimize return code from notifier chain in hw_breakpoint_handler
K.Prasad [Thu, 28 Jan 2010 11:14:15 +0000 (16:44 +0530)]
x86/hw-breakpoints: Optimize return code from notifier chain in hw_breakpoint_handler

Processing of debug exceptions in do_debug() can stop if it
originated from a hw-breakpoint exception by returning NOTIFY_STOP
in most cases.

But for certain cases such as:

a) user-space breakpoints with pending SIGTRAP signal delivery (as
in the case of ptrace induced breakpoints).

b) exceptions due to other causes than breakpoints

We will continue to process the exception by returning NOTIFY_DONE.

Signed-off-by: K.Prasad <prasad@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Roland McGrath <roland@redhat.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Jan Kiszka <jan.kiszka@siemens.com>
LKML-Reference: <20100128111415.GC13935@in.ibm.com>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
14 years agox86/debug: Clear reserved bits of DR6 in do_debug()
K.Prasad [Thu, 28 Jan 2010 11:14:01 +0000 (16:44 +0530)]
x86/debug: Clear reserved bits of DR6 in do_debug()

Clear the reserved bits from the stored copy of debug status
register (DR6).
This will help easy bitwise operations such as quick testing
of a debug event origin.

Signed-off-by: K.Prasad <prasad@linux.vnet.ibm.com>
Cc: Roland McGrath <roland@redhat.com>
Cc: Jan Kiszka <jan.kiszka@siemens.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Ingo Molnar <mingo@elte.hu>
LKML-Reference: <20100128111401.GB13935@in.ibm.com>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
14 years agotracing/kprobe: Cleanup unused return value of tracing functions
Xiao Guangrong [Thu, 28 Jan 2010 01:34:27 +0000 (09:34 +0800)]
tracing/kprobe: Cleanup unused return value of tracing functions

The return values of the kprobe's tracing functions are meaningless,
lets remove these.

Signed-off-by: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Acked-by: Masami Hiramatsu <mhiramat@redhat.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Jason Baron <jbaron@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
LKML-Reference: <4B60E9A3.2040505@cn.fujitsu.com>
[fweisbec@gmail: whitespace fixes, drop useless void returns in end
of functions]
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
14 years agoperf: Factorize trace events raw sample buffer operations
Xiao Guangrong [Thu, 28 Jan 2010 01:32:29 +0000 (09:32 +0800)]
perf: Factorize trace events raw sample buffer operations

Introduce ftrace_perf_buf_prepare() and ftrace_perf_buf_submit() to
gather the common code that operates on raw events sampling buffer.
This cleans up redundant code between regular trace events, syscall
events and kprobe events.

Changelog v1->v2:
- Rename function name as per Masami and Frederic's suggestion
- Add __kprobes for ftrace_perf_buf_prepare() and make
  ftrace_perf_buf_submit() inline as per Masami's suggestion
- Export ftrace_perf_buf_prepare since modules will use it

Signed-off-by: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Acked-by: Masami Hiramatsu <mhiramat@redhat.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Jason Baron <jbaron@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
LKML-Reference: <4B60E92D.9000808@cn.fujitsu.com>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
14 years agoperf: Fix inconsistency between IP and callchain sampling
Anton Blanchard [Mon, 18 Jan 2010 05:47:07 +0000 (16:47 +1100)]
perf: Fix inconsistency between IP and callchain sampling

When running perf across all cpus with backtracing (-a -g), sometimes we
get samples without associated backtraces:

    23.44%         init  [kernel]                     [k] restore
    11.46%         init                       eeba0c  [k] 0x00000000eeba0c
     6.77%      swapper  [kernel]                     [k] .perf_ctx_adjust_freq
     5.73%         init  [kernel]                     [k] .__trace_hcall_entry
     4.69%         perf  libc-2.9.so                  [.] 0x0000000006bb8c
                       |
                       |--11.11%-- 0xfffa941bbbc

It turns out the backtrace code has a check for the idle task and the IP
sampling does not. This creates problems when profiling an interrupt
heavy workload (in my case 10Gbit ethernet) since we get no backtraces
for interrupts received while idle (ie most of the workload).

Right now x86 and sh check that current is not NULL, which should never
happen so remove that too.

Idle task's exclusion must be performed from the core code, on top
of perf_event_attr:exclude_idle.

Signed-off-by: Anton Blanchard <anton@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mundt <lethal@linux-sh.org>
LKML-Reference: <20100118054707.GT12666@kryten>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
14 years agohw_breakpoints: Release the bp slot if arch_validate_hwbkpt_settings() fails.
Mahesh Salgaonkar [Thu, 21 Jan 2010 12:55:16 +0000 (18:25 +0530)]
hw_breakpoints: Release the bp slot if arch_validate_hwbkpt_settings() fails.

On a given architecture, when hardware breakpoint registration fails
due to un-supported access type (read/write/execute), we lose the bp
slot since register_perf_hw_breakpoint() does not release the bp slot
on failure.
Hence, any subsequent hardware breakpoint registration starts failing
with 'no space left on device' error.

This patch introduces error handling in register_perf_hw_breakpoint()
function and releases bp slot on error.

Signed-off-by: Mahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Cc: K. Prasad <prasad@linux.vnet.ibm.com>
Cc: Maneesh Soni <maneesh@in.ibm.com>
LKML-Reference: <20100121125516.GA32521@in.ibm.com>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
14 years agoperf trace: Add -i option for choosing input file
Hitoshi Mitake [Fri, 22 Jan 2010 13:45:29 +0000 (22:45 +0900)]
perf trace: Add -i option for choosing input file

perf trace lacks -i option for choosing input file.
This patch adds it to perf trace.

Signed-off-by: Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <1264167929-6741-1-git-send-email-mitake@dcl.info.waseda.ac.jp>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf symbols: Use the right variable to check for kallsyms in the cache
Arnaldo Carvalho de Melo [Fri, 22 Jan 2010 16:35:02 +0000 (14:35 -0200)]
perf symbols: Use the right variable to check for kallsyms in the cache

Probably this wasn't noticed when testing this on my parisc
machine because I must have copied manually to its cache the
vmlinux file used in the x86_64 machine, now that I tried
looking on a x86-32 machine with a fresh cache, kernel symbols
weren't being resolved even with the right kallsyms copy on its
cache, duh.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1264178102-4203-2-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf symbols: Fix inverted logic for showing kallsyms as the source of symbols
Arnaldo Carvalho de Melo [Fri, 22 Jan 2010 16:35:01 +0000 (14:35 -0200)]
perf symbols: Fix inverted logic for showing kallsyms as the source of symbols

Only if we parsed /proc/kallsyms (or a copy found in the buildid
cache) we should set the dso long name to "[kernel.kallsyms]".

Reported-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1264178102-4203-1-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf top: Handle PERF_RECORD_{FORK,EXIT} events
Arnaldo Carvalho de Melo [Thu, 21 Jan 2010 15:04:44 +0000 (13:04 -0200)]
perf top: Handle PERF_RECORD_{FORK,EXIT} events

As noticed by Mike, symbols in new tasks were not being
processed as we weren't processing these events.

Reported-by: Mike Galbraith <efault@gmx.de>
Tested-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1264086284-1431-2-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf top: Fix sample counting
Arnaldo Carvalho de Melo [Thu, 21 Jan 2010 15:04:43 +0000 (13:04 -0200)]
perf top: Fix sample counting

Broken since "5b2bb75 perf top: Support userspace symbols too".

Reported-by: Mike Galbraith <efault@gmx.de>
Tested-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frédéric Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1264086284-1431-1-git-send-email-acme@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf: Ignore perf.data.old
Amerigo Wang [Mon, 25 Jan 2010 05:07:30 +0000 (00:07 -0500)]
perf: Ignore perf.data.old

Tell git to ignore this file.

Signed-off-by: WANG Cong <amwang@redhat.com>
LKML-Reference: <20100125051052.3999.28082.sendpatchset@localhost.localdomain>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf report: Fix segmentation fault when running with '-g none'
Yong Wang [Fri, 22 Jan 2010 01:47:50 +0000 (09:47 +0800)]
perf report: Fix segmentation fault when running with '-g none'

Segmentation fault occurs when running perf report with '-g
none'.

Reported-by: Austin Zhang <austin.zhang@intel.com>
Signed-off-by: Yong Wang <yong.y.wang@intel.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <20100122014750.GA4111@ywang-moblin2.bj.intel.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agoperf: Reimplement frequency driven sampling
Peter Zijlstra [Tue, 26 Jan 2010 17:50:16 +0000 (18:50 +0100)]
perf: Reimplement frequency driven sampling

There was a bug in the old period code that caused intel_pmu_enable_all()
or native_write_msr_safe() to show up quite high in the profiles.

In staring at that code it made my head hurt, so I rewrote it in a
hopefully simpler fashion. Its now fully symetric between tick and
overflow driven adjustments and uses less data to boot.

The only complication is that it basically wants to do a u128 division.
The code approximates that in a rather simple truncate until it fits
fashion, taking care to balance the terms while truncating.

This version does not generate that sampling artefact.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Cc: <stable@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agofnctl: f_modown should call write_lock_irqsave/restore
Greg Kroah-Hartman [Tue, 26 Jan 2010 23:04:02 +0000 (15:04 -0800)]
fnctl: f_modown should call write_lock_irqsave/restore

Commit 703625118069f9f8960d356676662d3db5a9d116 exposed that f_modown()
should call write_lock_irqsave instead of just write_lock_irq so that
because a caller could have a spinlock held and it would not be good to
renable interrupts.

Cc: Eric W. Biederman <ebiederm@xmission.com>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Tavis Ormandy <taviso@google.com>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4
Linus Torvalds [Tue, 26 Jan 2010 03:05:06 +0000 (19:05 -0800)]
Merge branch 'for_linus' of git://git./linux/kernel/git/tytso/ext4

* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
  ext4: Drop EXT4_GET_BLOCKS_UPDATE_RESERVE_SPACE flag
  ext4: Fix quota accounting error with fallocate
  ext4: Handle -EDQUOT error on write

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/wim/linux-2.6-watchdog
Linus Torvalds [Tue, 26 Jan 2010 03:03:58 +0000 (19:03 -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] sbc_fitpc2_wdt: fix I/O space access technique.
  [WATCHDOG] ixp2000: Fix build failure caused by missing include

14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6
Linus Torvalds [Tue, 26 Jan 2010 03:03:45 +0000 (19:03 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/tiwai/sound-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6:
  ASoC: fix a memory-leak in wm8903
  ALSA: hda - add possibility to choose speakers configuration for 4930g
  ALSA: hda - Fix HP T5735 automute
  ALSA: hda - Turn on EAPD only if available for Realtek codecs
  ALSA: hda - Fix parsing pin node 0x21 on ALC259

14 years agoMerge branch 'kvm-updates/2.6.33' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Linus Torvalds [Tue, 26 Jan 2010 03:02:31 +0000 (19:02 -0800)]
Merge branch 'kvm-updates/2.6.33' of git://git./virt/kvm/kvm

* 'kvm-updates/2.6.33' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  KVM: x86: Fix leak of free lapic date in kvm_arch_vcpu_init()
  KVM: x86: Fix probable memory leak of vcpu->arch.mce_banks
  KVM: S390: fix potential array overrun in intercept handling
  KVM: fix spurious interrupt with irqfd
  eventfd - allow atomic read and waitqueue remove
  KVM: MMU: bail out pagewalk on kvm_read_guest error
  KVM: properly check max PIC pin in irq route setup
  KVM: only allow one gsi per fd
  KVM: x86: Fix host_mapping_level()
  KVM: powerpc: Show timing option only on embedded
  KVM: Fix race between APIC TMR and IRR

14 years agoMerge branch 'linux-next' of git://git.infradead.org/ubi-2.6
Linus Torvalds [Tue, 26 Jan 2010 03:02:06 +0000 (19:02 -0800)]
Merge branch 'linux-next' of git://git.infradead.org/ubi-2.6

* 'linux-next' of git://git.infradead.org/ubi-2.6:
  UBI: fix memory leak in update path
  UBI: add more checks to chdev open
  UBI: initialise update marker

14 years agoMerge branch 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelv...
Linus Torvalds [Tue, 26 Jan 2010 03:00:56 +0000 (19:00 -0800)]
Merge branch 'hwmon-for-linus' of git://git./linux/kernel/git/jdelvare/staging

* 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvare/staging:
  hwmon: (fschmd) Fix a memleak on multiple opens of /dev/watchdog
  hwmon: (asus_atk0110) Do not fail if MBIF is missing
  hwmon: (amc6821) Double unlock bug
  hwmon: (smsc47m1) Fix section mismatch

14 years agoMerge branch 'drm-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied...
Linus Torvalds [Tue, 26 Jan 2010 02:59:47 +0000 (18:59 -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: (95 commits)
  drm/radeon/kms: preface warning printk with driver name
  drm/radeon/kms: drop unnecessary printks.
  drm: fix regression in fb blank handling
  drm/radeon/kms: make hibernate work on IGPs
  drm/vmwgfx: Optimize memory footprint for DMA buffers.
  drm/ttm: Allow system memory as a busy placement.
  drm/ttm: Fix race condition in ttm_bo_delayed_delete (v3, final)
  drm/nv50: prevent switching off SOR when in use for DVI-over-DP
  drm/nv50: fail auxch transaction if reply count not what we expect
  drm/nouveau: fix failure path if userspace specifies no valid memtypes
  drm/nouveau: report LVDS as disconnected if lid closed
  drm/radeon/kms: fix legacy get_engine/memory clock
  drm/radeon/kms/atom: atom parser fixes
  drm/radeon/kms: clean up atombios pll code
  drm/radeon/kms: clean up pll struct
  drm/radeon/kms/atom: fix crtc lock ordering
  drm/radeon: r6xx/r7xx possible security issue, system ram access
  drm/radeon/kms: r600/r700 don't test ib if ib initialization fails
  drm/radeon/kms: Forbid creation of framebuffer with no valid GEM object
  drm/radeon/kms: r600 handle irq vector ring overflow
  ...

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Tue, 26 Jan 2010 02:57:25 +0000 (18:57 -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 IRQ ->set_affinity() methods.
  sparc: cpumask_of_node() should handle -1 as a node
  sparc64: Update defconfig.
  sparc: Add missing SW perf fault events.
  sparc64: Fully support both performance counters.
  sparc64: Add perf callchain support.
  sparc: convert to arch_gettimeoffset()
  sparc: leds_resource.end assigned to itself in clock_board_probe()
  sparc32: Fix page_to_phys().
  sparc: Simplify param.h by simply including <asm-generic/param.h>
  sparc32: Update defconfig.
  SPARC: use helpers for rlimits
  sparc: copy_from_user() should not return -EFAULT

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Tue, 26 Jan 2010 02:57:07 +0000 (18:57 -0800)]
Merge git://git./linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (42 commits)
  virtio_net: Make delayed refill more reliable
  sfc: Use fixed-size buffers for MCDI NVRAM requests
  sfc: Add workspace for GMAC bug workaround to MCDI MAC_STATS buffer
  tcp_probe: avoid modulus operation and wrap fix
  qlge: Only free resources if they were allocated
  netns xfrm: deal with dst entries in netns
  sky2: revert config space change
  vlan: fix vlan_skb_recv()
  netns xfrm: fix "ip xfrm state|policy count" misreport
  sky2: Enable/disable WOL per hardware device
  net: Fix IPv6 GSO type checks in Intel ethernet drivers
  igb/igbvf: cleanup exception handling in tx_map_adv
  MAINTAINERS: Add Intel igbvf maintainer
  e1000/e1000e: don't use small hardware rx buffers
  fmvj18x_cs: add new id (Panasonic lan & modem card)
  be2net: swap only first 2 fields of mcc_wrb
  Please add support for Microsoft MN-120 PCMCIA network card
  be2net: fix bug in rx page posting
  wimax/i2400m: Add support for more i6x50 SKUs
  e1000e: enhance frame fragment detection
  ...

14 years agoMerge branch 'omap-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Tue, 26 Jan 2010 02:56:12 +0000 (18:56 -0800)]
Merge branch 'omap-fixes-for-linus' of git://git./linux/kernel/git/tmlind/linux-omap-2.6

* 'omap-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap-2.6: (25 commits)
  OMAP2/3: DMTIMER: Clear pending interrupts when stopping a timer
  PM debug: Fix warning when no CONFIG_DEBUG_FS
  OMAP3: PM: DSS PM_WKEN to refill DMA
  OMAP: timekeeping: time should not stop during suspend
  OMAP3: PM: Force write last pad config register into save area
  OMAP: omap3_pm_get_suspend_state() error ignored in pwrdm_suspend_get()
  OMAP3: PM: Enable wake-up from McBSP2, 3 and 4 modules
  OMAP3: PM debug: fix build error when !CONFIG_DEBUG_FS
  OMAP3: PM: Removing redundant and potentially dangerous PRCM configration
  OMAP3: Fixed ARM aux ctrl register save/restore
  OMAP3: CPUidle: Fixed timer resolution
  OMAP3: PM: Remove duplicate code blocks
  OMAP3: PM: Disable interrupt controller AUTOIDLE before WFI
  OMAP3: PM: Enable system control module autoidle
  OMAP3: PM: Ack pending interrupts before entering suspend
  omap: Enable GPMC clock in gpmc_init
  OMAP1 clock: fix for "BUG: spinlock lockup on CPU#0"
  OMAP4: clocks: Fix the clksel_rate struct DPLL divs
  OMAP4: PRCM: Fix the base address for CHIRONSS reg defines
  OMAP: dma_chan[lch_head].flag & OMAP_DMA_ACTIVE tested twice in omap_dma_unlink_lch()
  ...

14 years agovirtio_net: Make delayed refill more reliable
Herbert Xu [Mon, 25 Jan 2010 23:51:01 +0000 (15:51 -0800)]
virtio_net: Make delayed refill more reliable

I have seen RX stalls on a machine that experienced a suspected
OOM.  After the stall, the RX buffer is empty on the guest side
and there are exactly 16 entries available on the host side.  As
the number of entries is less than that required by a maximal
skb, the host cannot proceed.

The guest did not have a refill job scheduled.

My diagnosis is that an OOM had occured, with the delayed refill
job scheduled.  The job was able to allocate at least one skb, but
not enough to overcome the minimum required by the host to proceed.

As the refill job would only reschedule itself if it failed completely
to allocate any skbs, this would lead to an RX stall.

The following patch removes this stall possibility by always
rescheduling the refill job until the ring is totally refilled.

Testing has shown that the RX stall no longer occurs whereas
previously it would occur within a day.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Acked-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agosfc: Use fixed-size buffers for MCDI NVRAM requests
Ben Hutchings [Mon, 25 Jan 2010 23:49:59 +0000 (15:49 -0800)]
sfc: Use fixed-size buffers for MCDI NVRAM requests

The low-level MCDI code always uses 32-bit MMIO operations, and
callers must pad input and output buffers to multiples of 4 bytes.
The MCDI NVRAM functions are not doing this.  Also, their buffers are
declared as variable-length arrays with no explicit maximum length.

Switch to a fixed buffer size based on the chunk size used by the
MTD driver (which is a multiple of 4).

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agosfc: Add workspace for GMAC bug workaround to MCDI MAC_STATS buffer
Guido Barzini [Mon, 25 Jan 2010 23:49:19 +0000 (15:49 -0800)]
sfc: Add workspace for GMAC bug workaround to MCDI MAC_STATS buffer

Due to a hardware bug in the SFC9000 family, the firmware must
transfer raw GMAC statistics to host memory before aggregating them
into the cooked (speed-independent) MAC statistics.  Extend the stats
buffer to support this.

The length of the buffer is explicit in the MAC_STATS command, so this
change is backward-compatible on both sides.

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agotcp_probe: avoid modulus operation and wrap fix
Stephen Hemminger [Mon, 25 Jan 2010 23:47:50 +0000 (15:47 -0800)]
tcp_probe: avoid modulus operation and wrap fix

By rounding up the buffer size to power of 2, several expensive
modulus operations can be avoided.  This patch also solves a bug where
the gap need when ring gets full was not being accounted for.

Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agoqlge: Only free resources if they were allocated
Breno Leitao [Mon, 25 Jan 2010 23:46:58 +0000 (15:46 -0800)]
qlge: Only free resources if they were allocated

Currently qlge tries to release regions even if they were not allocated.
This causes messages like the following in the kernel log

Trying to free nonexistent resource <00000000006af400-00000000006af4ff>
Trying to free nonexistent resource <00003c04ff9f4000-00003c04ff9f7fff>
Trying to free nonexistent resource <00003c04ffc00000-00003c04ffcfffff>

This patch fixes the goto logic in order to not release the resources
if they were not allocated.

Signed-off-by: Breno Leitao <leitao@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[WATCHDOG] sbc_fitpc2_wdt: fix I/O space access technique.
Denis Turischev [Thu, 21 Jan 2010 14:10:07 +0000 (16:10 +0200)]
[WATCHDOG] sbc_fitpc2_wdt: fix I/O space access technique.

The mdelay function was used between I/O access commands, that causes peak
in CPU usage. Fix it by substitution mdelay to msleep.

Expand usage on fitPC2 compatible boards according to DMI identification.

Signed-off-by: Denis Turischev <denis@compulab.co.il>
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>
14 years agoMerge branch 'fix/hda' into for-linus
Takashi Iwai [Mon, 25 Jan 2010 16:00:01 +0000 (17:00 +0100)]
Merge branch 'fix/hda' into for-linus

14 years agoASoC: fix a memory-leak in wm8903
Guennadi Liakhovetski [Fri, 22 Jan 2010 17:00:03 +0000 (18:00 +0100)]
ASoC: fix a memory-leak in wm8903

Remember to free the temporary register-cache.

Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: Liam Girdwood <lrg@slimlogic.co.uk>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Cc: stable@kernel.org
14 years agoKVM: x86: Fix leak of free lapic date in kvm_arch_vcpu_init()
Wei Yongjun [Fri, 22 Jan 2010 06:21:29 +0000 (14:21 +0800)]
KVM: x86: Fix leak of free lapic date in kvm_arch_vcpu_init()

In function kvm_arch_vcpu_init(), if the memory malloc for
vcpu->arch.mce_banks is fail, it does not free the memory
of lapic date. This patch fixed it.

Cc: stable@kernel.org
Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
14 years agoKVM: x86: Fix probable memory leak of vcpu->arch.mce_banks
Wei Yongjun [Fri, 22 Jan 2010 06:18:47 +0000 (14:18 +0800)]
KVM: x86: Fix probable memory leak of vcpu->arch.mce_banks

vcpu->arch.mce_banks is malloc in kvm_arch_vcpu_init(), but
never free in any place, this may cause memory leak. So this
patch fixed to free it in kvm_arch_vcpu_uninit().

Cc: stable@kernel.org
Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
14 years agoKVM: S390: fix potential array overrun in intercept handling
Christian Borntraeger [Thu, 21 Jan 2010 11:19:07 +0000 (12:19 +0100)]
KVM: S390: fix potential array overrun in intercept handling

kvm_handle_sie_intercept uses a jump table to get the intercept handler
for a SIE intercept. Static code analysis revealed a potential problem:
the intercept_funcs jump table was defined to contain (0x48 >> 2) entries,
but we only checked for code > 0x48 which would cause an off-by-one
array overflow if code == 0x48.

Use the compiler and ARRAY_SIZE to automatically set the limits.

Cc: stable@kernel.org
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
14 years agoKVM: fix spurious interrupt with irqfd
Michael S. Tsirkin [Wed, 13 Jan 2010 17:12:30 +0000 (19:12 +0200)]
KVM: fix spurious interrupt with irqfd

kvm didn't clear irqfd counter on deassign, as a result we could get a
spurious interrupt when irqfd is assigned back. this leads to poor
performance and, in theory, guest crash.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
14 years agoeventfd - allow atomic read and waitqueue remove
Davide Libenzi [Wed, 13 Jan 2010 17:34:36 +0000 (09:34 -0800)]
eventfd - allow atomic read and waitqueue remove

KVM needs a wait to atomically remove themselves from the eventfd ->poll()
wait queue head, in order to handle correctly their IRQfd deassign
operation.

This patch introduces such API, plus a way to read an eventfd from its
context.

Signed-off-by: Davide Libenzi <davidel@xmailserver.org>
Signed-off-by: Avi Kivity <avi@redhat.com>
14 years agoKVM: MMU: bail out pagewalk on kvm_read_guest error
Marcelo Tosatti [Thu, 14 Jan 2010 19:41:27 +0000 (17:41 -0200)]
KVM: MMU: bail out pagewalk on kvm_read_guest error

Exit the guest pagetable walk loop if reading gpte failed. Otherwise its
possible to enter an endless loop processing the previous present pte.

Cc: stable@kernel.org
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
14 years agoKVM: properly check max PIC pin in irq route setup
Marcelo Tosatti [Tue, 12 Jan 2010 18:42:09 +0000 (16:42 -0200)]
KVM: properly check max PIC pin in irq route setup

Otherwise memory beyond irq_states[16] might be accessed.

Noticed by Juan Quintela.

Cc: stable@kernel.org
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Acked-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
14 years agoKVM: only allow one gsi per fd
Michael S. Tsirkin [Wed, 13 Jan 2010 16:58:09 +0000 (18:58 +0200)]
KVM: only allow one gsi per fd

Looks like repeatedly binding same fd to multiple gsi's with irqfd can
use up a ton of kernel memory for irqfd structures.

A simple fix is to allow each fd to only trigger one gsi: triggering a
storm of interrupts in guest is likely useless anyway, and we can do it
by binding a single gsi to many interrupts if we really want to.

Cc: stable@kernel.org
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Acked-by: Gregory Haskins <ghaskins@novell.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
14 years agoKVM: x86: Fix host_mapping_level()
Sheng Yang [Tue, 5 Jan 2010 11:02:28 +0000 (19:02 +0800)]
KVM: x86: Fix host_mapping_level()

When found a error hva, should not return PAGE_SIZE but the level...

Also clean up the coding style of the following loop.

Cc: stable@kernel.org
Signed-off-by: Sheng Yang <sheng@linux.intel.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
14 years agoKVM: powerpc: Show timing option only on embedded
Alexander Graf [Sun, 20 Dec 2009 21:24:06 +0000 (22:24 +0100)]
KVM: powerpc: Show timing option only on embedded

Embedded PowerPC KVM has an exit timing implementation to track and evaluate
how much time was spent in which exit path.

For Book3S, we don't implement it. So let's not expose it as a config option
either.

Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Avi Kivity <avi@redhat.com>
14 years agoKVM: Fix race between APIC TMR and IRR
Avi Kivity [Tue, 29 Dec 2009 10:42:16 +0000 (12:42 +0200)]
KVM: Fix race between APIC TMR and IRR

When we queue an interrupt to the local apic, we set the IRR before the TMR.
The vcpu can pick up the IRR and inject the interrupt before setting the TMR,
and perhaps even EOI it, causing incorrect behaviour.

The race is really insignificant since it can only occur on the first
interrupt (usually following interrupts will not change TMR), but it's better
closed than open.

Fixed by reordering setting the TMR vs IRR.

Cc: stable@kernel.org
Signed-off-by: Avi Kivity <avi@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
14 years agohwmon: (fschmd) Fix a memleak on multiple opens of /dev/watchdog
Hans de Goede [Mon, 25 Jan 2010 14:00:50 +0000 (15:00 +0100)]
hwmon: (fschmd) Fix a memleak on multiple opens of /dev/watchdog

When /dev/watchdog gets opened a second time we return -EBUSY, but
we already have got a kref then, so we end up leaking our data struct.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Cc: stable@kernel.org
14 years agohwmon: (asus_atk0110) Do not fail if MBIF is missing
Luca Tettamanti [Mon, 25 Jan 2010 14:00:49 +0000 (15:00 +0100)]
hwmon: (asus_atk0110) Do not fail if MBIF is missing

MBIF (motherboard identification) is only used to print the name of
the board, it's not essential for the driver; do not fail if it's
missing. Based on Juan's patch.

Signed-off-by: Luca Tettamanti <kronos.it@gmail.com>
Acked-by: Juan RP <xtraeme@gmail.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
14 years agohwmon: (amc6821) Double unlock bug
Dan Carpenter [Mon, 25 Jan 2010 14:00:49 +0000 (15:00 +0100)]
hwmon: (amc6821) Double unlock bug

The mutex gets unlocked after we goto EXIT.

Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
14 years agohwmon: (smsc47m1) Fix section mismatch
Jeff Mahoney [Mon, 25 Jan 2010 14:00:48 +0000 (15:00 +0100)]
hwmon: (smsc47m1) Fix section mismatch

smsc47m1_restore is called from sm_smsc47m1_exit, which is an __exit
function, so it can't be __init.

Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
14 years agoALSA: hda - add possibility to choose speakers configuration for 4930g
Łukasz Wojniłowicz [Sun, 24 Jan 2010 13:12:37 +0000 (14:12 +0100)]
ALSA: hda - add possibility to choose speakers configuration for 4930g

Now one can choose speaker configuration in e.g. PulseAudio mixer

Signed-off-by: Łukasz Wojniłowicz <lukasz.wojnilowicz@gmail.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
14 years agonetns xfrm: deal with dst entries in netns
Alexey Dobriyan [Mon, 25 Jan 2010 06:47:53 +0000 (22:47 -0800)]
netns xfrm: deal with dst entries in netns

GC is non-existent in netns, so after you hit GC threshold, no new
dst entries will be created until someone triggers cleanup in init_net.

Make xfrm4_dst_ops and xfrm6_dst_ops per-netns.
This is not done in a generic way, because it woule waste
(AF_MAX - 2) * sizeof(struct dst_ops) bytes per-netns.

Reorder GC threshold initialization so it'd be done before registering
XFRM policies.

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agosky2: revert config space change
stephen hemminger [Sun, 24 Jan 2010 18:46:06 +0000 (18:46 +0000)]
sky2: revert config space change

Obviously, this register had some other impact that is causing
the regression.  Either it is masking some other access or needs
to be reset in some path.

Either, way it is best to just revert the change for 2.6.33

This reverts commit 166a0fd4c788ec7f10ca8194ec6d526afa12db75.

Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agodrm/radeon/kms: preface warning printk with driver name
Dave Airlie [Mon, 25 Jan 2010 06:13:55 +0000 (16:13 +1000)]
drm/radeon/kms: preface warning printk with driver name

This just adds a little more info to the warning for old -ati/mesa
userspaces.

Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms: drop unnecessary printks.
Dave Airlie [Mon, 25 Jan 2010 06:13:12 +0000 (16:13 +1000)]
drm/radeon/kms: drop unnecessary printks.

These printks aren't required anymore.

Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm: fix regression in fb blank handling
Zhenyu Wang [Mon, 18 Jan 2010 08:47:04 +0000 (16:47 +0800)]
drm: fix regression in fb blank handling

commit 731b5a15a3b1474a41c2ca29b4c32b0f21bc852e
Author: James Simmons <jsimmons@infradead.org>
Date:   Thu Oct 29 20:39:07 2009 +0000

    drm/kms: properly handle fbdev blanking

uses DRM_MODE_DPMS_ON for FB_BLANK_NORMAL, but DRM_MODE_DPMS_ON
is actually for turning output on instead of blank.

This makes fb blank broken on my T61, it put LVDS on but leave
pipe disabled which made screen totally white or caused some
'burning' effect.

[airlied: James objects to this but at this point in 2.6.33,
I can't see a patch that will fix this properly like he wants coming
in time and otherwise this is a regression - proper fix for 2.6.34
hopefully.]

Cc: James Simmons <jsimmons@infradead.org>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms: make hibernate work on IGPs
Dave Airlie [Mon, 25 Jan 2010 03:08:08 +0000 (13:08 +1000)]
drm/radeon/kms: make hibernate work on IGPs

This is the least invasive fix without migrating the radeon driver
to pm_ops from what I can see. We just always migrate VRAM objects
on IGPs for now and we can fix it up later to migrate depending
on STR vs STD.

Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/vmwgfx: Optimize memory footprint for DMA buffers.
Thomas Hellstrom [Sat, 16 Jan 2010 15:05:05 +0000 (16:05 +0100)]
drm/vmwgfx: Optimize memory footprint for DMA buffers.

Use VRAM whenever there is free space for DMA buffers,
but use system GMR memory if using VRAM would cause an eviction.

This significantly reduces the guest system memory usage for
VMs with a large amount of VRAM allocated.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/ttm: Allow system memory as a busy placement.
Thomas Hellstrom [Sat, 16 Jan 2010 15:05:04 +0000 (16:05 +0100)]
drm/ttm: Allow system memory as a busy placement.

This is needed to fix a vmwgfx memory usage bug.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agoMerge remote branch 'korg/drm-radeon-next' into drm-linus
Dave Airlie [Mon, 25 Jan 2010 06:04:21 +0000 (16:04 +1000)]
Merge remote branch 'korg/drm-radeon-next' into drm-linus

* korg/drm-radeon-next:
  drm/radeon/kms: fix legacy get_engine/memory clock
  drm/radeon/kms/atom: atom parser fixes
  drm/radeon/kms: clean up atombios pll code
  drm/radeon/kms: clean up pll struct
  drm/radeon/kms/atom: fix crtc lock ordering
  drm/radeon: r6xx/r7xx possible security issue, system ram access
  drm/radeon/kms: r600/r700 don't test ib if ib initialization fails
  drm/radeon/kms: Forbid creation of framebuffer with no valid GEM object
  drm/radeon/kms: r600 handle irq vector ring overflow
  drm/radeon/kms: r600/r700 don't process IRQ if not initialized
  drm/radeon/kms: r600/r700 disable irq at suspend
  drm/radeon/kms/r4xx: cleanup atom path
  drm/radeon/kms: fix atombios_crtc_set_base
  drm/radeon/kms/atom: upstream parser updates
  drm/radeon/kms/atom: fix some parser bugs
  drm/radeon/kms: fix hardcoded mmio size in register functions
  drm/radeon/kms/r100: fix bug in CS parser
  drm/radeon/kms/r200: fix bug in CS parser
  drm/radeon/kms/r200: fix bug in CS parser

14 years agoMerge remote branch 'nouveau/for-airlied' of ../drm-nouveau-next into drm-linus
Dave Airlie [Mon, 25 Jan 2010 06:04:11 +0000 (16:04 +1000)]
Merge remote branch 'nouveau/for-airlied' of ../drm-nouveau-next into drm-linus

* 'nouveau/for-airlied' of ../drm-nouveau-next:
  drm/nv50: prevent switching off SOR when in use for DVI-over-DP
  drm/nv50: fail auxch transaction if reply count not what we expect
  drm/nouveau: fix failure path if userspace specifies no valid memtypes
  drm/nouveau: report LVDS as disconnected if lid closed
  drm/nv50: prevent accidently turning off encoders we're actually using
  drm/nv50: fix alignment of per-channel fifo cache
  drm/nouveau: Evict buffers in VRAM before freeing sgdma
  drm/nouveau: Acknowledge DMA_VTX_PROTECTION PGRAPH interrupts
  drm/nouveau: fix thinko in nv04_instmem.c
  drm/nouveau: fix a race condition in nouveau_dma_wait()

14 years agovlan: fix vlan_skb_recv()
Eric Dumazet [Mon, 25 Jan 2010 03:52:24 +0000 (19:52 -0800)]
vlan: fix vlan_skb_recv()

Bruno Prémont found commit 9793241fe92f7d930
(vlan: Precise RX stats accounting) added a regression for non
hw accelerated vlans.

[   26.390576] BUG: unable to handle kernel NULL pointer dereference at (null)
[   26.396369] IP: [<df856b89>] vlan_skb_recv+0x89/0x280 [8021q]

vlan_dev_info() was used with original device, instead of
skb->dev. Also spotted by Américo Wang.

Reported-By: Bruno Prémont <bonbons@linux-vserver.org>
Tested-By: Bruno Prémont <bonbons@linux-vserver.org>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agodrm/ttm: Fix race condition in ttm_bo_delayed_delete (v3, final)
Luca Barbieri [Wed, 20 Jan 2010 19:01:30 +0000 (20:01 +0100)]
drm/ttm: Fix race condition in ttm_bo_delayed_delete (v3, final)

Resending this with Thomas Hellstrom's signoff for merging into 2.6.33

ttm_bo_delayed_delete has a race condition, because after we do:
kref_put(&nentry->list_kref, ttm_bo_release_list);

we are not holding the list lock and not holding any reference to
objects, and thus every bo in the list can be removed and freed at
this point.

However, we then use the next pointer we stored, which is not guaranteed
to be valid.

This was apparently the cause of some Nouveau oopses I experienced.

This patch rewrites the function so that it keeps the reference to nentry
until nentry itself is freed and we already got a reference to nentry->next.

v2 updated by me according to Thomas Hellstrom's feedback.
v3 proposed by Thomas Hellstrom. Commit comment updated by me.

Both updates fixed minor efficiency/style issues only and all three versions
should be correct.

Signed-off-by: Luca Barbieri <luca@luca-barbieri.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/nv50: prevent switching off SOR when in use for DVI-over-DP
Ben Skeggs [Fri, 22 Jan 2010 00:57:01 +0000 (10:57 +1000)]
drm/nv50: prevent switching off SOR when in use for DVI-over-DP

Another hack because of us exposing each encoder block's function as
an encoder rather than exposing a single encoder that deals with them
all.

A proper fix will come, it's just rather invasive so this hack will
do until then.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>