GitHub/moto-9609/android_kernel_motorola_exynos9610.git
12 years agostatic keys: Inline the static_key_enabled() function
Jason Baron [Tue, 28 Feb 2012 18:49:01 +0000 (13:49 -0500)]
static keys: Inline the static_key_enabled() function

In the jump label enabled case, calling static_key_enabled()
results in a function call. The function returns the results of
a compare, so it really doesn't need the overhead of a full
function call. Let's make it 'static inline' for both the jump
label enabled and disabled cases.

Signed-off-by: Jason Baron <jbaron@redhat.com>
Cc: a.p.zijlstra@chello.nl
Cc: rostedt@goodmis.org
Cc: mathieu.desnoyers@polymtl.ca
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/201202281849.q1SIn1p2023270@int-mx10.intmail.prod.int.phx2.redhat.com
Signed-off-by: Ingo Molnar <mingo@elte.hu>
12 years agoMerge branch 'perf/jump-labels' into perf/core
Ingo Molnar [Tue, 28 Feb 2012 18:59:45 +0000 (19:59 +0100)]
Merge branch 'perf/jump-labels' into perf/core

Merge reason: After much naming discussion, there seems to be consensus
              now - queue it up for v3.4.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
12 years agoMerge branch 'tip/perf/core' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt...
Ingo Molnar [Mon, 27 Feb 2012 07:44:48 +0000 (08:44 +0100)]
Merge branch 'tip/perf/core' of git://git./linux/kernel/git/rostedt/linux-trace into perf/core

12 years agostatic keys: Introduce 'struct static_key', static_key_true()/false() and static_key_...
Ingo Molnar [Fri, 24 Feb 2012 07:31:31 +0000 (08:31 +0100)]
static keys: Introduce 'struct static_key', static_key_true()/false() and static_key_slow_[inc|dec]()

So here's a boot tested patch on top of Jason's series that does
all the cleanups I talked about and turns jump labels into a
more intuitive to use facility. It should also address the
various misconceptions and confusions that surround jump labels.

Typical usage scenarios:

        #include <linux/static_key.h>

        struct static_key key = STATIC_KEY_INIT_TRUE;

        if (static_key_false(&key))
                do unlikely code
        else
                do likely code

Or:

        if (static_key_true(&key))
                do likely code
        else
                do unlikely code

The static key is modified via:

        static_key_slow_inc(&key);
        ...
        static_key_slow_dec(&key);

The 'slow' prefix makes it abundantly clear that this is an
expensive operation.

I've updated all in-kernel code to use this everywhere. Note
that I (intentionally) have not pushed through the rename
blindly through to the lowest levels: the actual jump-label
patching arch facility should be named like that, so we want to
decouple jump labels from the static-key facility a bit.

On non-jump-label enabled architectures static keys default to
likely()/unlikely() branches.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Acked-by: Jason Baron <jbaron@redhat.com>
Acked-by: Steven Rostedt <rostedt@goodmis.org>
Cc: a.p.zijlstra@chello.nl
Cc: mathieu.desnoyers@efficios.com
Cc: davem@davemloft.net
Cc: ddaney.cavm@gmail.com
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/20120222085809.GA26397@elte.hu
Signed-off-by: Ingo Molnar <mingo@elte.hu>
12 years agostatic keys: Add docs better explaining the whole 'struct static_key' mechanism
Jason Baron [Tue, 21 Feb 2012 20:03:30 +0000 (15:03 -0500)]
static keys: Add docs better explaining the whole 'struct static_key' mechanism

Add better documentation for static keys.

Signed-off-by: Jason Baron <jbaron@redhat.com>
Cc: rostedt@goodmis.org
Cc: mathieu.desnoyers@efficios.com
Cc: davem@davemloft.net
Cc: ddaney.cavm@gmail.com
Cc: a.p.zijlstra@chello.nl
Link: http://lkml.kernel.org/r/52570e566e5f1914f27b67e4eafb5781b8f9f9db.1329851692.git.jbaron@redhat.com
[ Added a 'Summary' section and rewrote it to explain static keys ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
12 years agotracepoint, vfs, sched: Add exec() tracepoint
David Smith [Tue, 7 Feb 2012 16:11:05 +0000 (10:11 -0600)]
tracepoint, vfs, sched: Add exec() tracepoint

Added a minimal exec tracepoint. Exec is an important major event
in the life of a task, like fork(), clone() or exit(), all of
which we already trace.

[ We also do scheduling re-balancing during exec() - so it's useful
  from a scheduler instrumentation POV as well. ]

If you want to watch a task start up, when it gets exec'ed is a good place
to start.  With the addition of this tracepoint, exec's can be monitored
and better picture of general system activity can be obtained. This
tracepoint will also enable better process life tracking, allowing you to
answer questions like "what process keeps starting up binary X?".

This tracepoint can also be useful in ftrace filtering and trigger
conditions: i.e. starting or stopping filtering when exec is called.

Signed-off-by: David Smith <dsmith@redhat.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/4F314D19.7030504@redhat.com
Signed-off-by: Ingo Molnar <mingo@elte.hu>
12 years agoMerge tag 'perf-core-for-mingo' of git://git.kernel.org/pub/scm/linux/kernel/git...
Ingo Molnar [Wed, 22 Feb 2012 09:36:56 +0000 (10:36 +0100)]
Merge tag 'perf-core-for-mingo' of git://git./linux/kernel/git/acme/linux into perf/core

Small fixes, also includes an important fix from Stephane for system
wide monitoring, problem introduced recently in perf/core, in the pid
list patches.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
12 years agojump label: Fix compiler warning
Jason Baron [Tue, 21 Feb 2012 20:02:57 +0000 (15:02 -0500)]
jump label: Fix compiler warning

While cross-compiling on sparc64, I found:

kernel/jump_label.c: In function 'jump_label_update':
kernel/jump_label.c:447:40: warning: cast from pointer to
integer of different size [-Wpointer-to-int-cast]

Fix by casting to 'unsigned long'.

Signed-off-by: Jason Baron <jbaron@redhat.com>
Cc: rostedt@goodmis.org
Cc: mathieu.desnoyers@efficios.com
Cc: davem@davemloft.net
Cc: ddaney.cavm@gmail.com
Cc: a.p.zijlstra@chello.nl
Link: http://lkml.kernel.org/r/08026cbc6df80619cae833ef1ebbbc43efab69ab.1329851692.git.jbaron@redhat.com
Signed-off-by: Ingo Molnar <mingo@elte.hu>
12 years agojump label: Add a WARN() if jump label key count goes negative
Jason Baron [Tue, 21 Feb 2012 20:02:53 +0000 (15:02 -0500)]
jump label: Add a WARN() if jump label key count goes negative

The count on a jump label key should never go negative. Add a
WARN() to check for this condition.

Signed-off-by: Jason Baron <jbaron@redhat.com>
Cc: Gleb Natapov <gleb@redhat.com>
Cc: rostedt@goodmis.org
Cc: mathieu.desnoyers@efficios.com
Cc: davem@davemloft.net
Cc: ddaney.cavm@gmail.com
Cc: a.p.zijlstra@chello.nl
Link: http://lkml.kernel.org/r/3c68556121be4d1920417a3fe367da1ec38246b4.1329851692.git.jbaron@redhat.com
Signed-off-by: Ingo Molnar <mingo@elte.hu>
12 years agoperf tools: fix broken perf record -a mode
Stephane Eranian [Tue, 21 Feb 2012 14:54:25 +0000 (15:54 +0100)]
perf tools: fix broken perf record -a mode

The following commit:
b52956c perf tools: Allow multiple threads or processes in record, stat, top

introduced a bug in the thread_map code which caused perf record -a to
not setup system-wide monitoring properly.

$ taskset -c 1 noploop 1000 &
$ perf record -a -C 1 sleep 10
$ perf report -D | tail -20
cycles stats:
           TOTAL events:       4413
            MMAP events:       4025
            COMM events:        340
          SAMPLE events:         48

Here I was expecting about 10,000 samples and not 48.

In system-wide mode, the PID passed to perf_event_open() must be -1 and
it was 0. That caused the kernel to setup a per-process event on PID:0.
Consequently, the number of samples captured does not correspond to the
requested measurement.

The following one-liner fixes the problem for me with or without -C.

I would also suggest to change the malloc() to something that matches
the struct definition. thread_map->map[] is declared as int map[] and
not pid_t map[]. If map[] can only contain pids, then change the struct
definition.

Acked-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20120221145424.GA6757@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoftrace, perf: Add filter support for function trace event
Jiri Olsa [Wed, 15 Feb 2012 14:51:54 +0000 (15:51 +0100)]
ftrace, perf: Add filter support for function trace event

Adding support to filter function trace event via perf
interface. It is now possible to use filter interface
in the perf tool like:

  perf record -e ftrace:function --filter="(ip == mm_*)" ls

The filter syntax is restricted to the the 'ip' field only,
and following operators are accepted '==' '!=' '||', ending
up with the filter strings like:

  ip == f1[, ]f2 ... || ip != f3[, ]f4 ...

with comma ',' or space ' ' as a function separator. If the
space ' ' is used as a separator, the right side of the
assignment needs to be enclosed in double quotes '"', e.g.:

  perf record -e ftrace:function --filter '(ip == do_execve,sys_*,ext*)' ls
  perf record -e ftrace:function --filter '(ip == "do_execve,sys_*,ext*")' ls
  perf record -e ftrace:function --filter '(ip == "do_execve sys_* ext*")' ls

The '==' operator adds trace filter with same effect as would
be added via set_ftrace_filter file.

The '!=' operator adds trace filter with same effect as would
be added via set_ftrace_notrace file.

The right side of the '!=', '==' operators is list of functions
or regexp. to be added to filter separated by space.

The '||' operator is used for connecting multiple filter definitions
together. It is possible to have more than one '==' and '!='
operators within one filter string.

Link: http://lkml.kernel.org/r/1329317514-8131-8-git-send-email-jolsa@redhat.com
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
12 years agoftrace: Allow to specify filter field type for ftrace events
Jiri Olsa [Wed, 15 Feb 2012 14:51:53 +0000 (15:51 +0100)]
ftrace: Allow to specify filter field type for ftrace events

Adding FILTER_TRACE_FN event field type for function tracepoint
event, so it can be properly recognized within filtering code.

Currently all fields of ftrace subsystem events share the common
field type FILTER_OTHER. Since the function trace fields need
special care within the filtering code we need to recognize it
properly, hence adding the FILTER_TRACE_FN event type.

Adding filter parameter to the FTRACE_ENTRY macro, to specify the
filter field type for the event.

Link: http://lkml.kernel.org/r/1329317514-8131-7-git-send-email-jolsa@redhat.com
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
12 years agoftrace, perf: Add support to use function tracepoint in perf
Jiri Olsa [Wed, 15 Feb 2012 14:51:52 +0000 (15:51 +0100)]
ftrace, perf: Add support to use function tracepoint in perf

Adding perf registration support for the ftrace function event,
so it is now possible to register it via perf interface.

The perf_event struct statically contains ftrace_ops as a handle
for function tracer. The function tracer is registered/unregistered
in open/close actions.

To be efficient, we enable/disable ftrace_ops each time the traced
process is scheduled in/out (via TRACE_REG_PERF_(ADD|DELL) handlers).
This way tracing is enabled only when the process is running.
Intentionally using this way instead of the event's hw state
PERF_HES_STOPPED, which would not disable the ftrace_ops.

It is now possible to use function trace within perf commands
like:

  perf record -e ftrace:function ls
  perf stat -e ftrace:function ls

Allowed only for root.

Link: http://lkml.kernel.org/r/1329317514-8131-6-git-send-email-jolsa@redhat.com
Acked-by: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
12 years agoftrace: Add FTRACE_ENTRY_REG macro to allow event registration
Jiri Olsa [Wed, 15 Feb 2012 14:51:51 +0000 (15:51 +0100)]
ftrace: Add FTRACE_ENTRY_REG macro to allow event registration

Adding FTRACE_ENTRY_REG macro so particular ftrace entries
could specify registration function and thus become accesible
via perf.

This will be used in upcomming patch for function trace.

Link: http://lkml.kernel.org/r/1329317514-8131-5-git-send-email-jolsa@redhat.com
Acked-by: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
12 years agoftrace, perf: Add add/del tracepoint perf registration actions
Jiri Olsa [Wed, 15 Feb 2012 14:51:50 +0000 (15:51 +0100)]
ftrace, perf: Add add/del tracepoint perf registration actions

Adding TRACE_REG_PERF_ADD and TRACE_REG_PERF_DEL to handle
perf event schedule in/out actions.

The add action is invoked for when the perf event is scheduled in,
while the del action is invoked when the event is scheduled out.

Link: http://lkml.kernel.org/r/1329317514-8131-4-git-send-email-jolsa@redhat.com
Acked-by: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
12 years agoftrace, perf: Add open/close tracepoint perf registration actions
Jiri Olsa [Wed, 15 Feb 2012 14:51:49 +0000 (15:51 +0100)]
ftrace, perf: Add open/close tracepoint perf registration actions

Adding TRACE_REG_PERF_OPEN and TRACE_REG_PERF_CLOSE to differentiate
register/unregister from open/close actions.

The register/unregister actions are invoked for the first/last
tracepoint user when opening/closing the event.

The open/close actions are invoked for each tracepoint user when
opening/closing the event.

Link: http://lkml.kernel.org/r/1329317514-8131-3-git-send-email-jolsa@redhat.com
Acked-by: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
12 years agoftrace: Add enable/disable ftrace_ops control interface
Jiri Olsa [Wed, 15 Feb 2012 14:51:48 +0000 (15:51 +0100)]
ftrace: Add enable/disable ftrace_ops control interface

Adding a way to temporarily enable/disable ftrace_ops. The change
follows the same way as 'global' ftrace_ops are done.

Introducing 2 global ftrace_ops - control_ops and ftrace_control_list
which take over all ftrace_ops registered with FTRACE_OPS_FL_CONTROL
flag. In addition new per cpu flag called 'disabled' is also added to
ftrace_ops to provide the control information for each cpu.

When ftrace_ops with FTRACE_OPS_FL_CONTROL is registered, it is
set as disabled for all cpus.

The ftrace_control_list contains all the registered 'control' ftrace_ops.
The control_ops provides function which iterates ftrace_control_list
and does the check for 'disabled' flag on current cpu.

Adding 3 inline functions:
  ftrace_function_local_disable/ftrace_function_local_enable
  - enable/disable the ftrace_ops on current cpu
  ftrace_function_local_disabled
  - get disabled ftrace_ops::disabled value for current cpu

Link: http://lkml.kernel.org/r/1329317514-8131-2-git-send-email-jolsa@redhat.com
Acked-by: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
12 years agotracing: Don't use p->len field to determine output in __print_*() functions
Steven Rostedt [Tue, 21 Feb 2012 01:37:32 +0000 (20:37 -0500)]
tracing: Don't use p->len field to determine output in __print_*() functions

If more than one __print_*() function is used in a tracepoint
(__print_flags(), __print_symbols(), etc), then the temp seq buffer will
not be zero on entry. Using the temp seq buffer's length to know if
data has been printed or not in the current function is incorrect and
may produce incorrect results.

Currently, no in-tree tracepoint causes this bug, but new ones may
be created.

Cc: Andrew Vagin <avagin@openvz.org>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
12 years agotracing: Don't print an extra separator of flags
Andrey Vagin [Sun, 19 Feb 2012 11:16:07 +0000 (14:16 +0300)]
tracing: Don't print an extra separator of flags

If __print_flags() is used after another __print_*() function, the
temp seq_file buffer will not be empty on entry, and the delimiter will
be printed even though there's just one field. We get something like:

|S

instead of just:

S

This is because the length of the temp seq buffer is used to determine
if the delimiter is printed or not. But this algorithm fails when
the seq buffer is not empty on entry, and the delimiter will be printed
because it thinks that a previous field was already printed.

Link: http://lkml.kernel.org/r/1329650167-480655-1-git-send-email-avagin@openvz.org
Signed-off-by: Andrew Vagin <avagin@openvz.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
12 years agoperf tools: Remove duplicated string.h includes
Danny Kukawka [Thu, 16 Feb 2012 13:54:19 +0000 (14:54 +0100)]
perf tools: Remove duplicated string.h includes

tools/perf/util/probe-event.c included 'string.h' twice, remove the
duplicate.

Acked-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Danny Kukawka <dkukawka@suse.de>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Jovi Zhang <bookjovi@gmail.com>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Steven Rostedt <rostedt@goodmis.org>
Link: http://lkml.kernel.org/r/1329400459-31570-1-git-send-email-danny.kukawka@bisect.de
Signed-off-by: Danny Kukawka <danny.kukawka@bisect.de>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: Allow expressions in __print_symbolic() fields
Stefan Hajnoczi [Sun, 4 Sep 2011 15:08:59 +0000 (16:08 +0100)]
perf tools: Allow expressions in __print_symbolic() fields

The __print_symbolic() function takes a sequence of key-value pairs for
pretty-printing a constant.  The new kvm:kvm_exit print fmt uses the
expression:

  __print_symbolic(..., { 0x040 + 1, "DB excp" }, ...)

Currently only atoms are supported and this print fmt fails to parse.
This patch adds support for expressions instead of just atoms so that
0x040 + 1 is parsed successfully.  Also add arg_num_eval() support for
the '+' operator.

Acked-by: Steven Rostedt <rostedt@goodmis.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Steven Rostedt <rostedt@goodmis.org>
Link: http://lkml.kernel.org/r/1315148939-14313-1-git-send-email-stefanha@linux.vnet.ibm.com
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoMerge branch 'tip/perf/core' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt...
Ingo Molnar [Fri, 17 Feb 2012 11:55:07 +0000 (12:55 +0100)]
Merge branch 'tip/perf/core' of git://git./linux/kernel/git/rostedt/linux-trace into perf/core

12 years agoMerge tag 'perf-core-for-mingo' of git://git.kernel.org/pub/scm/linux/kernel/git...
Ingo Molnar [Fri, 17 Feb 2012 08:02:28 +0000 (09:02 +0100)]
Merge tag 'perf-core-for-mingo' of git://git./linux/kernel/git/acme/linux into perf/core

Includes smaller fixes and improvements plus the exclude_{host,guest} feature
test and fallback to handle older kernels.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: Invert the sample_id_all logic
Arnaldo Carvalho de Melo [Tue, 14 Feb 2012 16:18:57 +0000 (14:18 -0200)]
perf tools: Invert the sample_id_all logic

Instead of requiring that users of perf_record_opts set
.sample_id_all_avail to true, just invert the logic, using
.sample_id_all_missing, that doesn't need to be explicitely initialized
since gcc will zero members ommitted in a struct initialization.

Just like the newly introduced .exclude_{guest,host} feature test.

Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ab772uzk78cwybihf0vt7kxw@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: Handle kernels that don't support attr.exclude_{guest,host}
Arnaldo Carvalho de Melo [Tue, 14 Feb 2012 16:05:30 +0000 (14:05 -0200)]
perf tools: Handle kernels that don't support attr.exclude_{guest,host}

Just fall back to resetting those fields, if set, warning the user that
that feature is not available.

If guest samples appear they will just be discarded because no struct
machine will be found and thus the event will be accounted as not
handled and dropped, see 0c09571.

Reported-by: Namhyung Kim <namhyung@gmail.com>
Tested-by: Joerg Roedel <joerg.roedel@amd.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Joerg Roedel <joerg.roedel@amd.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-vuwxig36mzprl5n7nzvnxxsh@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: cleanup initialization of attr->size
Stephane Eranian [Thu, 9 Feb 2012 15:12:38 +0000 (16:12 +0100)]
perf tools: cleanup initialization of attr->size

The perf_event_attr size needs to be initialized in all cases because it
captures the ABI version.

This patch moves the initialization of the field from the
perf_event_open() syscall stub to its proper location in the
event_attr_init().

Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20120209151238.GA10272@quad
Signed-off-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: Factor out feature op to process header sections
Robert Richter [Fri, 10 Feb 2012 14:41:56 +0000 (15:41 +0100)]
perf tools: Factor out feature op to process header sections

There is individual code for each feature to process header sections.

Adding a function pointer .process to struct feature_ops for keeping the
implementation in separate functions. Code to process header sections is
now a generic function.

Cc: Ingo Molnar <mingo@elte.hu>
Link: http://lkml.kernel.org/r/1328884916-5901-2-git-send-email-robert.richter@amd.com
Signed-off-by: Robert Richter <robert.richter@amd.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: Moving code in header.c
Robert Richter [Fri, 10 Feb 2012 14:41:55 +0000 (15:41 +0100)]
perf tools: Moving code in header.c

Needed for later changes. No modified functionality.

Cc: Ingo Molnar <mingo@elte.hu>
Link: http://lkml.kernel.org/r/1328884916-5901-1-git-send-email-robert.richter@amd.com
Signed-off-by: Robert Richter <robert.richter@amd.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: Add bitmap_or function into bitmap object
Jiri Olsa [Fri, 27 Jan 2012 14:34:23 +0000 (15:34 +0100)]
perf tools: Add bitmap_or function into bitmap object

Adding implementation os bitmap_or function to the bitmap object. It is
stolen from the kernel lib/bitmap.o object.

It is used in upcomming patches.

Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1327674868-10486-5-git-send-email-jolsa@redhat.com
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: Add sysfs mountpoint interface
Jiri Olsa [Fri, 27 Jan 2012 14:34:22 +0000 (15:34 +0100)]
perf tools: Add sysfs mountpoint interface

Adding sysfs object to provide sysfs mount information in the same way
as debugfs object does.

The object provides following function:
  sysfs_find_mountpoint

which returns the sysfs mount mount.

Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1327674868-10486-4-git-send-email-jolsa@redhat.com
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: Remove unused functions from debugfs object
Jiri Olsa [Fri, 27 Jan 2012 14:34:21 +0000 (15:34 +0100)]
perf tools: Remove unused functions from debugfs object

Following debugfs object functions are not referenced
within the code:

  int debugfs_valid_entry(const char *path);
  int debugfs_umount(void);
  int debugfs_write(const char *entry, const char *value);
  int debugfs_read(const char *entry, char *buffer, size_t size);
  void debugfs_force_cleanup(void);
  int debugfs_make_path(const char *element, char *buffer, int size);

Removing them.

Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1327674868-10486-3-git-send-email-jolsa@redhat.com
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: Get rid of ctype.h in symbol.c
Namhyung Kim [Fri, 10 Feb 2012 01:10:17 +0000 (10:10 +0900)]
perf tools: Get rid of ctype.h in symbol.c

The ctype.h in symbol.c was needed because of isupper(). However we now
have it in util.h, it can be changed to use our implementation.

Cc: Ingo Molnar <mingo@elte.hu>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1328836217-9118-3-git-send-email-namhyung.kim@lge.com
Signed-off-by: Namhyung Kim <namhyung.kim@lge.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: ctype.c only wants util.h
Namhyung Kim [Fri, 10 Feb 2012 01:10:16 +0000 (10:10 +0900)]
perf tools: ctype.c only wants util.h

The implementation of sane ctype macros only depends on symbols in
util.h not cache.h.

Cc: Ingo Molnar <mingo@elte.hu>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1328836217-9118-2-git-send-email-namhyung.kim@lge.com
Signed-off-by: Namhyung Kim <namhyung.kim@lge.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: Implement islower/isupper macro into util.h
Namhyung Kim [Fri, 10 Feb 2012 01:10:15 +0000 (10:10 +0900)]
perf tools: Implement islower/isupper macro into util.h

The util.h header provides various ctype macros but lacks those two.

Add them.

Cc: Ingo Molnar <mingo@elte.hu>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1328836217-9118-1-git-send-email-namhyung.kim@lge.com
Signed-off-by: Namhyung Kim <namhyung.kim@lge.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: Change perf_guest default back to false
Joerg Roedel [Fri, 10 Feb 2012 17:05:05 +0000 (18:05 +0100)]
perf tools: Change perf_guest default back to false

Setting perf_guest to true by default makes no sense because the perf
subcommands can not setup guest symbol information and thus not process
and guest samples. The only exception is perf-kvm which changes the
perf_guest value on its own.  So change the default for perf_guest back
to false.

Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1328893505-4115-3-git-send-email-joerg.roedel@amd.com
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf top: Don't process samples with no valid machine object
Joerg Roedel [Fri, 10 Feb 2012 17:05:04 +0000 (18:05 +0100)]
perf top: Don't process samples with no valid machine object

The perf sample processing code relies on a valid machine object. Make
sure that this path is only entered when such a object exists.

A counter for samples where no machine object exits is also introduced
to give the user a message about these samples.

Reported-by: David Ahern <dsahern@gmail.com>
Reported-by: Jason Wang <jasowang@redhat.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1328893505-4115-2-git-send-email-joerg.roedel@amd.com
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: Allow multiple threads or processes in record, stat, top
David Ahern [Wed, 8 Feb 2012 16:32:52 +0000 (09:32 -0700)]
perf tools: Allow multiple threads or processes in record, stat, top

Allow a user to collect events for multiple threads or processes
using a comma separated list.

e.g., collect data on a VM and its vhost thread:
  perf top -p 21483,21485
  perf stat -p 21483,21485 -ddd
  perf record -p 21483,21485

or monitoring vcpu threads
  perf top -t 21488,21489
  perf stat -t 21488,21489 -ddd
  perf record -t 21488,21489

Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1328718772-16688-1-git-send-email-dsahern@gmail.com
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: Fix out of tree compiles
David Ahern [Mon, 13 Feb 2012 14:57:37 +0000 (07:57 -0700)]
perf tools: Fix out of tree compiles

For latest tip/perf/core tree Compiles are failing on:

GEN common-cmds.h
make: *** No rule to make target `../../arch/x86/lib/memset_64.S', needed by `builtin-annotate.o'.  Stop.

Resolve by adding memset.* to the tar file.

Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1329145057-26302-1-git-send-email-dsahern@gmail.com
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: Fix build dependency of perf python extension
Namhyung Kim [Sun, 12 Feb 2012 10:45:24 +0000 (19:45 +0900)]
perf tools: Fix build dependency of perf python extension

The perf python extention (perf.so) file lacks its dependencies in the
Makefile so that it cannot be refreshed if one of source files it depends
is changed. Fix it by putting them in a separate file and processing it in
both of Makefile and setup.py.

Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1329043524-12470-1-git-send-email-namhyung@gmail.com
Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agotracing/trivial: Use kcalloc instead of kzalloc to allocate array
Thomas Meyer [Tue, 29 Nov 2011 21:08:00 +0000 (22:08 +0100)]
tracing/trivial: Use kcalloc instead of kzalloc to allocate array

The advantage of kcalloc is, that will prevent integer overflows which could
result from the multiplication of number of elements and size and it is also
a bit nicer to read.

The semantic patch that makes this change is available
in https://lkml.org/lkml/2011/11/25/107

Link: http://lkml.kernel.org/r/1322600880.1534.347.camel@localhost.localdomain
Signed-off-by: Thomas Meyer <thomas@m3y3r.de>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
12 years agoprintk/tracing: Add console output tracing
Johannes Berg [Thu, 24 Nov 2011 19:03:08 +0000 (20:03 +0100)]
printk/tracing: Add console output tracing

Add a printk.console trace point to record any printk
messages into the trace, regardless of the current
console loglevel. This can help correlate (existing)
printk debugging with other tracing.

Link: http://lkml.kernel.org/r/1322161388.5366.54.camel@jlt3.sipsolutions.net
Acked-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Ingo Molnar <mingo@redhat.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
12 years agoftrace: Append wakeup_rt description of ftrace doc
Geunsik Lim [Wed, 8 Feb 2012 10:05:37 +0000 (19:05 +0900)]
ftrace: Append wakeup_rt description of ftrace doc

Append and update the description about wakeup/wakeup_rt usage.

Link: http://lkml.kernel.org/r/1328695537-15081-2-git-send-email-geunsik.lim@gmail.com
Cc: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Geunsik Lim <geunsik.lim@samsung.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
12 years agoftrace: sched_switch plugin is deprecated
Geunsik Lim [Wed, 8 Feb 2012 10:05:36 +0000 (19:05 +0900)]
ftrace: sched_switch plugin is deprecated

Actually, sched_switch function tracer is merged into wakeup/wakeup_rt
Update 'mini-HOWTO' for ftrace(Kernel function tracer).
If we want to trace "sched:sched_switch" to trace sched_switch func,
We may utilize event option.(e.g: trace-cmd list -e | grep sched)
This patch is based on Linux-3.3.rc2-SMP-PREEMPT

Link: http://lkml.kernel.org/r/1328695537-15081-1-git-send-email-geunsik.lim@gmail.com
Cc: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Geunsik Lim <geunsik.lim@samsung.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
12 years agocpuidle/tracing: Denote the tracepoints as being in rcu_idle_exit() section
Steven Rostedt [Tue, 7 Feb 2012 14:46:01 +0000 (09:46 -0500)]
cpuidle/tracing: Denote the tracepoints as being in rcu_idle_exit() section

As the tracepoints in the cpuidle code are called when rcu_idle_exit() is in
effect, the _rcuidle() version must be used, otherwise the rcu_read_lock()s
that protect the tracepoint will not be honored.

Cc: Len Brown <len.brown@intel.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
12 years agox86/tracing: Denote the power and cpuidle tracepoints as _rcuidle()
Steven Rostedt [Tue, 7 Feb 2012 14:40:30 +0000 (09:40 -0500)]
x86/tracing: Denote the power and cpuidle tracepoints as _rcuidle()

The power and cpuidle tracepoints are called within a rcu_idle_exit()
section, and must be denoted with the _rcuidle() version of the tracepoint.

Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
12 years agotracing/rcu: Add trace_##name##__rcuidle() static tracepoint for inside rcu_idle_exit...
Steven Rostedt [Tue, 7 Feb 2012 14:32:43 +0000 (09:32 -0500)]
tracing/rcu: Add trace_##name##__rcuidle() static tracepoint for inside rcu_idle_exit() sections

Added is a new static inline function that lets *any* tracepoint be used
inside a rcu_idle_exit() section. And this also solves the problem where
the same tracepoint may be used inside a rcu_idle_exit() section as well
as outside of one.

I added a new tracepoint function with a "_rcuidle" extension. All
tracepoints can be used with either the normal "trace_foobar()"
function, or the "trace_foobar_rcuidle()" function when inside a
rcu_idle_exit() section.

All tracepoints defined by TRACE_EVENT() or any of the derivatives
will have a "_rcuidle()" function also defined. When a tracepoint is
used within an rcu_idle_exit() section, the "_rcuidle()" version must
be used. This denotes that the tracepoint is within rcu_idle_exit()
and it allows the rcu read locks within the tracepoint to still
be valid, as this version takes us out of rcu_idle_exit().

Another nice aspect about this patch is that "static inline"s are not
compiled into text when not used. So only the tracepoints that actually
use the _rcuidle() version will have them defined in the actual text
that is booted.

Link: http://lkml.kernel.org/r/1328563113.2200.39.camel@gandalf.stny.rr.com>
Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
12 years agox86: Fix to decode grouped AVX with VEX pp bits
Masami Hiramatsu [Fri, 10 Feb 2012 05:33:40 +0000 (14:33 +0900)]
x86: Fix to decode grouped AVX with VEX pp bits

Fix to decode grouped AVX with VEX pp bits which should be
handled as same as last-prefixes. This fixes below warnings
in posttest with CONFIG_CRYPTO_SHA1_SSSE3=y.

 Warning: arch/x86/tools/test_get_len found difference at <sha1_transform_avx>:ffffffff810d5fc0
 Warning: ffffffff810d6069: c5 f9 73 de 04        vpsrldq $0x4,%xmm6,%xmm0
 Warning: objdump says 5 bytes, but insn_get_length() says 4
 ...

With this change, test_get_len can decode it correctly.

 $ arch/x86/tools/test_get_len -v -y
 ffffffff810d6069:       c5 f9 73 de 04          vpsrldq $0x4,%xmm6,%xmm0
 Succeed: decoded and checked 1 instructions

Reported-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: yrl.pp-manager.tt@hitachi.com
Link: http://lkml.kernel.org/r/20120210053340.30429.73410.stgit@localhost.localdomain
Signed-off-by: Ingo Molnar <mingo@elte.hu>
12 years agowatchdog: Fix code/comments mismatches
Fernando Luis Vázquez Cao [Thu, 9 Feb 2012 22:42:22 +0000 (17:42 -0500)]
watchdog: Fix code/comments mismatches

Reflect the change in the soft and hard lockup thresholds and
their relation to the frequency of the hrtimer and NMI events in
the code comments. While at it, remove references to files that
do not exist anymore.

Signed-off-by: Fernando Luis Vazquez Cao <fernando@oss.ntt.co.jp>
Signed-off-by: Don Zickus <dzickus@redhat.com>
Link: http://lkml.kernel.org/r/1328827342-6253-3-git-send-email-dzickus@redhat.com
Signed-off-by: Ingo Molnar <mingo@elte.hu>
12 years agowatchdog: Update Kconfig entries
Fernando Luis Vázquez Cao [Thu, 9 Feb 2012 22:42:21 +0000 (17:42 -0500)]
watchdog: Update Kconfig entries

The soft and hard lockup thresholds have changed so the
corresponding Kconfig entries need to be updated accordingly.
Add a reference to  watchdog_thresh while at it.

Signed-off-by: Fernando Luis Vazquez Cao <fernando@oss.ntt.co.jp>
Signed-off-by: Don Zickus <dzickus@redhat.com>
Link: http://lkml.kernel.org/r/1328827342-6253-2-git-send-email-dzickus@redhat.com
Signed-off-by: Ingo Molnar <mingo@elte.hu>
12 years agowatchdog: Update documentation
Fernando Luis Vázquez Cao [Thu, 9 Feb 2012 22:42:20 +0000 (17:42 -0500)]
watchdog: Update documentation

The soft and hard lockup detectors are now built on top of the
hrtimer and perf subsystems. Update the documentation
accordingly.

Signed-off-by: Fernando Luis Vazquez Cao<fernando@oss.ntt.co.jp>
Acked-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Don Zickus <dzickus@redhat.com>
Link: http://lkml.kernel.org/r/1328827342-6253-1-git-send-email-dzickus@redhat.com
Signed-off-by: Ingo Molnar <mingo@elte.hu>
12 years agoperf record: No build id option fails
David Ahern [Mon, 6 Feb 2012 22:27:52 +0000 (15:27 -0700)]
perf record: No build id option fails

A recent refactoring of perf-record introduced the following:

perf record -a -B
Couldn't generating buildids. Use --no-buildid to profile anyway.
sleep: Terminated

I believe the triple negative was meant to be only a double negative.
:-) While I'm there, fixed the grammar on the error message.

Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1328567272-13190-1-git-send-email-dsahern@gmail.com
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: fix endianness detection in perf.data
Stephane Eranian [Thu, 2 Feb 2012 12:54:44 +0000 (13:54 +0100)]
perf tools: fix endianness detection in perf.data

The current version of perf detects whether or not the perf.data file is
written in a different endianness using the attr_size field in the
header of the file. This field represents sizeof(struct perf_event_attr)
as known to perf record. If the sizes do not match, then perf tries the
byte-swapped version. If they match, then the tool assumes a different
endianness.

The issue with the approach is that it assumes the size of
perf_event_attr always has to match between perf record and perf report.
However, the kernel perf_event ABI is extensible.  New fields can be
added to struct perf_event_attr. Consequently, it is not possible to use
attr_size to detect endianness.

This patch takes another approach by using the magic number written at
the beginning of the perf.data file to detect endianness. The magic
number is an eight-byte signature.  It's primary purpose is to identify
(signature) a perf.data file. But it could also be used to encode the
endianness.

The patch introduces a new value for this signature. The key difference
is that the signature is written differently in the file depending on
the endianness. Thus, by comparing the signature from the file with the
tool's own signature it is possible to detect endianness. The new
signature is "PERFILE2".

Backward compatiblity with existing perf.data file is ensured.

Tested-by: David Ahern <dsahern@gmail.com>
Acked-by: David Ahern <dsahern@gmail.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Anshuman Khandual <khandual@linux.vnet.ibm.com>
Cc: Arun Sharma <asharma@fb.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Lin Ming <ming.m.lin@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Roberto Agostino Vitillo <ravitillo@lbl.gov>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Vince Weaver <vweaver1@eecs.utk.edu>
Link: http://lkml.kernel.org/r/1328187288-24395-15-git-send-email-eranian@google.com
Signed-off-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agox86/sched/perf/AMD: Set sched_clock_stable
Borislav Petkov [Tue, 7 Feb 2012 12:08:52 +0000 (13:08 +0100)]
x86/sched/perf/AMD: Set sched_clock_stable

Stephane Eranian reported that doing a scheduler latency
measurements with perf on AMD doesn't work out as expected due
to the fact that the sched_clock() granularity is too coarse,
i.e. done in jiffies due to the sched_clock_stable not set,
which, if set, would mean that we get to use the TSC as sample
source which would give us much higher precision.

However, there's no reason not to set sched_clock_stable on AMD
because all families from F10h and upwards do have an invariant
TSC and have the CPUID flag to prove (CPUID_8000_0007_EDX[8]).

Make it so, #1.

Signed-off-by: Borislav Petkov <bp@alien8.de>
Cc: Borislav Petkov <bp@amd64.org>
Cc: Venki Pallipadi <venki@google.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Arnaldo Carvalho de Melo <acme@infradead.org>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Andreas Herrmann <andreas.herrmann3@amd.com>
Link: http://lkml.kernel.org/r/20120206132546.GA30854@quad
[ Should any non-standard system break the TSC, we should
  mark them so explicitly, in their platform init handler, or
  in a DMI quirk. ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
12 years agoMerge tag 'perf-core-for-mingo' of git://git.kernel.org/pub/scm/linux/kernel/git...
Ingo Molnar [Tue, 7 Feb 2012 09:00:06 +0000 (10:00 +0100)]
Merge tag 'perf-core-for-mingo' of git://git./linux/kernel/git/acme/linux into perf/core

perf/core fixes and improvements.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
12 years agoMerge tag 'v3.3-rc2' into perf/core
Ingo Molnar [Tue, 7 Feb 2012 08:59:29 +0000 (09:59 +0100)]
Merge tag 'v3.3-rc2' into perf/core

Linux 3.3-rc2

Pick up the latest fixes.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
12 years agoperf stat: Align scaled output of cpu-clock
Namhyung Kim [Mon, 6 Feb 2012 07:44:45 +0000 (16:44 +0900)]
perf stat: Align scaled output of cpu-clock

The output of cpu-clock event is controlled in nsec_printout(),
but its alignment was broken:

 Performance counter stats for 'sleep 1':

         6,038,774 instructions              #    0.00  insns per cycle
               180 faults                    #    0.007 K/sec                   [99.95%]
         1,282,201 branches                  #    0.053 M/sec                   [99.84%]
      24126.221811 cpu-clock                 [99.62%]
      24121.689540 task-clock                #   24.098 CPUs utilized           [99.52%]

       1.001001017 seconds time elapsed

This patch fixes this:

 Performance counter stats for 'sleep 1':

        13,540,843 instructions              #    0.00  insns per cycle
               180 faults                    #    0.007 K/sec                   [99.94%]
         2,875,386 branches                  #    0.119 M/sec                   [99.82%]
      24144.221137 cpu-clock                                                    [99.61%]
      24133.515366 task-clock                #   24.109 CPUs utilized           [99.52%]

       1.001020946 seconds time elapsed

Cc: Ingo Molnar <mingo@elte.hu>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1328514285-26232-2-git-send-email-namhyung.kim@lge.com
Signed-off-by: Namhyung Kim <namhyung.kim@lge.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf stat: Adjust print unit
Namhyung Kim [Mon, 6 Feb 2012 07:44:44 +0000 (16:44 +0900)]
perf stat: Adjust print unit

The default 'M/sec' unit is not useful if the result is small enough.

Adjust it dynamically according to the value.

Cc: Ingo Molnar <mingo@elte.hu>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1328514285-26232-1-git-send-email-namhyung.kim@lge.com
Signed-off-by: Namhyung Kim <namhyung.kim@lge.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf doc: Allow producing documentation in a specified output directory
Franck Bui-Huu [Mon, 6 Feb 2012 15:17:23 +0000 (16:17 +0100)]
perf doc: Allow producing documentation in a specified output directory

Currently we can put the object files in a different directory by using
'O=' comand line argument.

However the generated documentation files don't honor this directive,

This patch fixes that. It's been tested for man target but the others
seems currently broken so no tests have been done on them so far.

Link: http://lkml.kernel.org/r/1328541443-18003-1-git-send-email-fbuihuu@gmail.com
Signed-off-by: Franck Bui-Huu <fbuihuu@gmail.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tool: Fix perf stack to non executable on x86_64
Jiri Olsa [Wed, 1 Feb 2012 12:54:08 +0000 (13:54 +0100)]
perf tool: Fix perf stack to non executable on x86_64

By adding following objects:
  bench/mem-memset-x86-64-asm.o
  bench/mem-memcpy-x86-64-asm.o
the x86_64 perf binary ended up with executable stack.

The reason was that above objects are assembler sourced and are missing the
GNU-stack note section. In such case the linker assumes that the final binary
should not be restricted at all and mark the stack as RWX.

Adding section ".note.GNU-stack" definition to mentioned objects, with all
flags disabled, thus omiting those objects from linker stack flags decision.

Reported-at: https://bugzilla.redhat.com/show_bug.cgi?id=783570
Reported-by: Clark Williams <williams@redhat.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1328100848-5630-1-git-send-email-jolsa@redhat.com
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
[ committer note: Remaining bits after what was already added to perf/urgent ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoMerge branch 'perf/urgent' into perf/core
Arnaldo Carvalho de Melo [Mon, 6 Feb 2012 21:10:52 +0000 (19:10 -0200)]
Merge branch 'perf/urgent' into perf/core

So that we can get the perf bench exec stack fixes and then apply the
remaining fix for the files added after what is in perf/urgent.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf evsel: Fix an issue where perf report fails to show the proper percentage
Naveen N. Rao [Fri, 3 Feb 2012 17:01:13 +0000 (22:31 +0530)]
perf evsel: Fix an issue where perf report fails to show the proper percentage

This patch fixes an issue where perf report shows nan% for certain
perf.data files. The below is from a report for a do_fork probe:

   -nan%           sshd  [kernel.kallsyms]  [k] do_fork
   -nan%    packagekitd  [kernel.kallsyms]  [k] do_fork
   -nan%    dbus-daemon  [kernel.kallsyms]  [k] do_fork
   -nan%           bash  [kernel.kallsyms]  [k] do_fork

A git bisect shows commit f3bda2c as the cause. However, looking back
through the git history, I saw commit 640c03c which seems to have
removed the required initialization for perf_sample->period. The problem
only started showing after commit f3bda2c. The below patch re-introduces
the initialization and it fixes the problem for me.

With the below patch, for the same perf.data:

  73.08%             bash  [kernel.kallsyms]  [k] do_fork
   8.97%      11-dhclient  [kernel.kallsyms]  [k] do_fork
   6.41%             sshd  [kernel.kallsyms]  [k] do_fork
   3.85%        20-chrony  [kernel.kallsyms]  [k] do_fork
   2.56%         sendmail  [kernel.kallsyms]  [k] do_fork

This patch applies over current linux-tip commit 9949284.

Problem introduced in:

$ git describe 640c03c
v2.6.37-rc3-83-g640c03c

Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: stable@kernel.org
Link: http://lkml.kernel.org/r/20120203170113.5190.25558.stgit@localhost6.localdomain6
Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: Fix prefix matching for kernel maps
Jiri Olsa [Sun, 5 Feb 2012 17:11:05 +0000 (18:11 +0100)]
perf tools: Fix prefix matching for kernel maps

In some perf ancient versions we used '[kernel.kallsyms._text]' as the
name for the kernel map.

This got changed with commit:
  perf: 'perf kvm' tool for monitoring guest performance from host
  commit a1645ce12adb6c9cc9e19d7695466204e3f017fe
  Author: Zhang, Yanmin <yanmin_zhang@linux.intel.com>

and we started to use following name '[kernel.kallsyms]_text'.

This name change is important for the report code dealing with ancient
perf data. When processing the kernel map event, we need to recognize
the old naming (dont match the last ']') and initialize the kernel map
correctly.

The subsequent call to maps__set_kallsyms_ref_reloc_sym deals with the
superfluous ']' to get correct symbol name.

Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1328461865-6127-1-git-send-email-jolsa@redhat.com
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: Fix perf stack to non executable on x86_64
Jiri Olsa [Mon, 6 Feb 2012 20:54:06 +0000 (18:54 -0200)]
perf tools: Fix perf stack to non executable on x86_64

By adding following objects:
  bench/mem-memcpy-x86-64-asm.o
the x86_64 perf binary ended up with executable stack.

The reason was that above object are assembler sourced and is missing the
GNU-stack note section. In such case the linker assumes that the final binary
should not be restricted at all and mark the stack as RWX.

Adding section ".note.GNU-stack" definition to mentioned object, with all
flags disabled, thus omiting this object from linker stack flags decision.

Problem introduced in:

  $ git describe ea7872b
  v2.6.37-rc2-19-gea7872b

Reported-at: https://bugzilla.redhat.com/show_bug.cgi?id=783570
Reported-by: Clark Williams <williams@redhat.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: stable@kernel.org
Link: http://lkml.kernel.org/r/1328100848-5630-1-git-send-email-jolsa@redhat.com
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
[ committer note: Backported fix to perf/urgent (3.3-rc2+) ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agotracing/softirq: Move __raise_softirq_irqoff() out of header
Steven Rostedt [Thu, 26 Jan 2012 01:18:55 +0000 (20:18 -0500)]
tracing/softirq: Move __raise_softirq_irqoff() out of header

The __raise_softirq_irqoff() contains a tracepoint. As tracepoints in headers
can cause issues, and not to mention, bloats the kernel when they are
in a static inline, it is best to move the function that contains the
tracepoint out of the header and into softirq.c.

Link: http://lkml.kernel.org/r/20120118120711.GB14863@elte.hu
Suggested-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
12 years agoftrace: Change filter/notrace set functions to return exit code
Jiri Olsa [Mon, 2 Jan 2012 09:04:14 +0000 (10:04 +0100)]
ftrace: Change filter/notrace set functions to return exit code

Currently the ftrace_set_filter and ftrace_set_notrace functions
do not return any return code. So there's no way for ftrace_ops
user to tell wether the filter was correctly applied.

The set_ftrace_filter interface returns error in case the filter
did not match:

  # echo krava > set_ftrace_filter
  bash: echo: write error: Invalid argument

Changing both ftrace_set_filter and ftrace_set_notrace functions
to return zero if the filter was applied correctly or -E* values
in case of error.

Link: http://lkml.kernel.org/r/1325495060-6402-2-git-send-email-jolsa@redhat.com
Acked-by: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
12 years agoperf: Remove deprecated WARN_ON_ONCE()
Stephane Eranian [Thu, 2 Feb 2012 11:04:01 +0000 (12:04 +0100)]
perf: Remove deprecated WARN_ON_ONCE()

With the new throttling/unthrottling code introduced with
commit:

  e050e3f0a71b ("perf: Fix broken interrupt rate throttling")

we occasionally hit two WARN_ON_ONCE() checks in:

  - intel_pmu_pebs_enable()
  - intel_pmu_lbr_enable()
  - x86_pmu_start()

The assertions are no longer problematic. There is a valid
path where they can trigger but it is harmless.

The assertion can be triggered with:

  $ perf record -e instructions:pp ....

Leading to paths:

  intel_pmu_pebs_enable
  intel_pmu_enable_event
  x86_perf_event_set_period
  x86_pmu_start
  perf_adjust_freq_unthr_context
  perf_event_task_tick
  scheduler_tick

And:

  intel_pmu_lbr_enable
  intel_pmu_enable_event
  x86_perf_event_set_period
  x86_pmu_start
  perf_adjust_freq_unthr_context.
  perf_event_task_tick
  scheduler_tick

cpuc->enabled is always on because when we get to
perf_adjust_freq_unthr_context() the PMU is not totally
disabled. Furthermore when we need to adjust a period,
we only stop the event we need to change and not the
entire PMU. Thus, when we re-enable, cpuc->enabled is
already set. Note that when we stop the event, both
pebs and lbr are stopped if necessary (and possible).

Signed-off-by: Stephane Eranian <eranian@google.com>
Cc: peterz@infradead.org
Link: http://lkml.kernel.org/r/20120202110401.GA30911@quad
Signed-off-by: Ingo Molnar <mingo@elte.hu>
12 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph...
Linus Torvalds [Thu, 2 Feb 2012 23:47:33 +0000 (15:47 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/sage/ceph-client

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client:
  rbd: fix safety of rbd_put_client()
  rbd: fix a memory leak in rbd_get_client()
  ceph: create a new session lock to avoid lock inversion
  ceph: fix length validation in parse_reply_info()
  ceph: initialize client debugfs outside of monc->mutex
  ceph: change "ceph.layout" xattr to be "ceph.file.layout"

12 years agoinclude/linux/lp8727.h: Remove executable bit
Josh Triplett [Thu, 2 Feb 2012 23:29:08 +0000 (15:29 -0800)]
include/linux/lp8727.h: Remove executable bit

Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agorbd: fix safety of rbd_put_client()
Alex Elder [Sun, 29 Jan 2012 19:57:43 +0000 (13:57 -0600)]
rbd: fix safety of rbd_put_client()

The rbd_client structure uses a kref to arrange for cleaning up and
freeing an instance when its last reference is dropped.  The cleanup
routine is rbd_client_release(), and one of the things it does is
delete the rbd_client from rbd_client_list.  It acquires node_lock
to do so, but the way it is done is still not safe.

The problem is that when attempting to reuse an existing rbd_client,
the structure found might already be in the process of getting
destroyed and cleaned up.

Here's the scenario, with "CLIENT" representing an existing
rbd_client that's involved in the race:

 Thread on CPU A                | Thread on CPU B
 ---------------                | ---------------
 rbd_put_client(CLIENT)         | rbd_get_client()
   kref_put()                   |   (acquires node_lock)
     kref->refcount becomes 0   |   __rbd_client_find() returns CLIENT
     calls rbd_client_release() |   kref_get(&CLIENT->kref);
                                |   (releases node_lock)
       (acquires node_lock)     |
       deletes CLIENT from list | ...and starts using CLIENT...
       (releases node_lock)     |
       and frees CLIENT         | <-- but CLIENT gets freed here

Fix this by having rbd_put_client() acquire node_lock.  The result
could still be improved, but at least it avoids this problem.

Signed-off-by: Alex Elder <elder@dreamhost.com>
Signed-off-by: Sage Weil <sage@newdream.net>
12 years agoFix race in process_vm_rw_core
Christopher Yeoh [Thu, 2 Feb 2012 01:04:09 +0000 (11:34 +1030)]
Fix race in process_vm_rw_core

This fixes the race in process_vm_core found by Oleg (see

  http://article.gmane.org/gmane.linux.kernel/1235667/

for details).

This has been updated since I last sent it as the creation of the new
mm_access() function did almost exactly the same thing as parts of the
previous version of this patch did.

In order to use mm_access() even when /proc isn't enabled, we move it to
kernel/fork.c where other related process mm access functions already
are.

Signed-off-by: Chris Yeoh <yeohc@au1.ibm.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agorbd: fix a memory leak in rbd_get_client()
Alex Elder [Tue, 24 Jan 2012 16:08:36 +0000 (10:08 -0600)]
rbd: fix a memory leak in rbd_get_client()

If an existing rbd client is found to be suitable for use in
rbd_get_client(), the rbd_options structure is not being
freed as it should.  Fix that.

Signed-off-by: Alex Elder <elder@dreamhost.com>
Signed-off-by: Sage Weil <sage@newdream.net>
12 years agoceph: create a new session lock to avoid lock inversion
Alex Elder [Fri, 13 Jan 2012 01:48:10 +0000 (17:48 -0800)]
ceph: create a new session lock to avoid lock inversion

Lockdep was reporting a possible circular lock dependency in
dentry_lease_is_valid().  That function needs to sample the
session's s_cap_gen and and s_cap_ttl fields coherently, but needs
to do so while holding a dentry lock.  The s_cap_lock field was
being used to protect the two fields, but that can't be taken while
holding a lock on a dentry within the session.

In most cases, the s_cap_gen and s_cap_ttl fields only get operated
on separately.  But in three cases they need to be updated together.
Implement a new lock to protect the spots updating both fields
atomically is required.

Signed-off-by: Alex Elder <elder@dreamhost.com>
Reviewed-by: Sage Weil <sage@newdream.net>
12 years agoceph: fix length validation in parse_reply_info()
Xi Wang [Sun, 15 Jan 2012 03:20:59 +0000 (22:20 -0500)]
ceph: fix length validation in parse_reply_info()

"len" is read from network and thus needs validation.  Otherwise, given
a bogus "len" value, p+len could be an out-of-bounds pointer, which is
used in further parsing.

Signed-off-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Sage Weil <sage@newdream.net>
12 years agoceph: initialize client debugfs outside of monc->mutex
Sage Weil [Sat, 14 Jan 2012 06:22:03 +0000 (22:22 -0800)]
ceph: initialize client debugfs outside of monc->mutex

Initializing debufs under monc->mutex introduces a lock dependency for
sb->s_type->i_mutex_key, which (combined with several other dependencies)
leads to an annoying lockdep warning.  There's no particular reason to do
the debugfs setup under this lock, so move it out.

It used to be the case that our first monmap could come from the OSD; that
is no longer the case with recent servers, so we will reliably set up the
client entry during the initial authentication.

We don't have to worry about racing with debugfs teardown by
ceph_debugfs_client_cleanup() because ceph_destroy_client() calls
ceph_msgr_flush() first, which will wait for the message dispatch work
to complete (and the debugfs init to complete).

Fixes: #1940
Signed-off-by: Sage Weil <sage@newdream.net>
12 years agoceph: change "ceph.layout" xattr to be "ceph.file.layout"
Alex Elder [Thu, 12 Jan 2012 01:41:01 +0000 (17:41 -0800)]
ceph: change "ceph.layout" xattr to be "ceph.file.layout"

The virtual extended attribute named "ceph.layout" is meaningful
only for regular files.  Change its name to be "ceph.file.layout" to
more directly reflect that in the ceph xattr namespace.  Preserve
the old "ceph.layout" name for the time being (until we decide it's
safe to get rid of it entirely).

Add a missing initializer for "readonly" in the terminating entry.

Signed-off-by: Alex Elder <elder@dreamhost.com>
Reviewed-by: Sage Weil <sage@newdream.net>
12 years agoperf record: Make feature initialization generic
Robert Richter [Thu, 15 Dec 2011 16:32:40 +0000 (17:32 +0100)]
perf record: Make feature initialization generic

Loop over all features to enable it instead of explicitly enabling every
single feature. Reducing duplicate code and making it more robust to
later changes e.g. when adding more features.

Cc: Ingo Molnar <mingo@elte.hu>
Link: http://lkml.kernel.org/r/1323966762-8574-3-git-send-email-robert.richter@amd.com
Signed-off-by: Robert Richter <robert.richter@amd.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf probe: Rename target_module to target
Srikar Dronamraju [Thu, 2 Feb 2012 14:20:40 +0000 (19:50 +0530)]
perf probe: Rename target_module to target

This is a precursor patch that modifies names that refer to
kernel/module to also refer to user space names.

Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Anton Arapov <anton@redhat.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Jim Keniston <jkenisto@linux.vnet.ibm.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Linux-mm <linux-mm@kvack.org>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Roland McGrath <roland@hack.frob.com>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20120202142040.5967.64156.sendpatchset@srdronam.in.ibm.com
Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoperf tools: Remove distclean from Makefile help output
John Kacur [Wed, 1 Feb 2012 22:16:31 +0000 (23:16 +0100)]
perf tools: Remove distclean from Makefile help output

distclean as an alias for clean was removed from the perf Makefile by
commit a3d1ee10d1bf4520af3d44c1aa6cd46956ec4fd7

However, that commit neglected to remove it from the help output of
the perf Makefile, which could result in a user trying the following.

$ cd tools/perf/
$ make help | grep distclean
  distclean - alias to clean
$ make distclean
make: *** No rule to make target `distclean'.  Stop.

This patch removes it from the Makefile help output.

Link: http://lkml.kernel.org/r/1328134591-19851-1-git-send-email-jkacur@redhat.com
Signed-off-by: John Kacur <jkacur@redhat.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agoMerge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux
Linus Torvalds [Thu, 2 Feb 2012 19:19:03 +0000 (11:19 -0800)]
Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux:
  drm/radeon/kms/blit: fix blit copy for very large buffers
  drm/radeon/kms: fix TRAVIS panel setup
  drm/radeon: fix use after free in ATRM bios reading code.
  drm/radeon/kms: Fix device tree linkage of DP i2c buses too
  drm/radeon: Set DESKTOP_HEIGHT register to the framebuffer (not mode) height.
  drm/radeon/kms: disable output polling when suspended
  drm/nv50/pm: signedness bug in nv50_pm_clocks_pre()
  drm/nouveau/gem: fix fence_sync race / oops
  drm/nouveau: fix typo on mxmdcb option
  drm/nouveau/mxm: pretend to succeed, even if we can't shadow the MXM-SIS
  drm/nouveau/disp: check that panel power gpio is enabled at init time

12 years agoMerge branch 'next' of git://git.monstr.eu/linux-2.6-microblaze
Linus Torvalds [Thu, 2 Feb 2012 19:18:18 +0000 (11:18 -0800)]
Merge branch 'next' of git://git.monstr.eu/linux-2.6-microblaze

* 'next' of git://git.monstr.eu/linux-2.6-microblaze:
  Revert "microblaze: Add topology init"

12 years agoMerge branches 'core-urgent-for-linus', 'perf-urgent-for-linus', 'sched-urgent-for...
Linus Torvalds [Thu, 2 Feb 2012 19:11:13 +0000 (11:11 -0800)]
Merge branches 'core-urgent-for-linus', 'perf-urgent-for-linus', 'sched-urgent-for-linus' and 'x86-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

* 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  bugs, x86: Fix printk levels for panic, softlockups and stack dumps

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf top: Fix number of samples displayed
  perf tools: Fix strlen() bug in perf_event__synthesize_event_type()
  perf tools: Fix broken build by defining _GNU_SOURCE in Makefile
  x86/dumpstack: Remove unneeded check in dump_trace()
  perf: Fix broken interrupt rate throttling

* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched/rt: Fix task stack corruption under __ARCH_WANT_INTERRUPTS_ON_CTXSW
  sched: Fix ancient race in do_exit()
  sched/nohz: Fix nohz cpu idle load balancing state with cpu hotplug
  sched/s390: Fix compile error in sched/core.c
  sched: Fix rq->nr_uninterruptible update race

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/reboot: Remove VersaLogic Menlow reboot quirk
  x86/reboot: Skip DMI checks if reboot set by user
  x86: Properly parenthesize cmpxchg() macro arguments

12 years agodrm/radeon/kms/blit: fix blit copy for very large buffers
Ilija Hadzic [Thu, 2 Feb 2012 15:26:24 +0000 (10:26 -0500)]
drm/radeon/kms/blit: fix blit copy for very large buffers

Evergreen and NI blit copy was broken if the buffer maps to a rectangle
whose one dimension is 16384 (max dimension allowed by these chips).
In the mainline kernel, the problem is exposed only when buffers are
very large (1G), but it's still a problem. The problem could be exposed
for smaller buffers if anyone modifies the algorithm for rectangle
construction in r600_blit_create_rect() (the reason why someone would
modify that algorithm is to tune the performance of buffer moves).

The root cause was in i2f() function which only operated on range between
0 and 16383. Fix this by extending the range of i2f() function to 0 to
32767.

While at it improve the function so that the range can be easily
extended in the future (if it becomes necessary), cleanup lines
over 80 characters, and replace in-line comments with one strategic
comment that explains the crux of the function.

Credits to michel@daenzer.net for pointing out the root cause of
the bug.

v2: Fix I2F_MAX_INPUT constant definition goof and warn only once
    if input argument is out of range. Edit the comment a little
    bit to avoid some linguistic confusion and make it look better
    in general.

Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Michel Dänzer <michel@daenzer.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agodrm/radeon/kms: fix TRAVIS panel setup
Alex Deucher [Thu, 2 Feb 2012 15:18:00 +0000 (10:18 -0500)]
drm/radeon/kms: fix TRAVIS panel setup

Different versions of the DP to LVDS bridge chip
need different panel mode settings depending on
the chip version used.

Fixes:
https://bugs.freedesktop.org/show_bug.cgi?id=41569

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agodrm/radeon: fix use after free in ATRM bios reading code.
Dave Airlie [Thu, 2 Feb 2012 15:25:16 +0000 (15:25 +0000)]
drm/radeon: fix use after free in ATRM bios reading code.

Fixes:
https://bugs.freedesktop.org/show_bug.cgi?id=45503

Reported-and-Debugged-by: mlambda@gmail.com
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agoMerge tag 'firewire-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394...
Linus Torvalds [Thu, 2 Feb 2012 02:30:15 +0000 (18:30 -0800)]
Merge tag 'firewire-fixes' of git://git./linux/kernel/git/ieee1394/linux1394

firewire fixes post v3.3-rc1

Add workarounds table entries for hardware bugs in
  - FireWire part of Sound Blaster Audigy cards,
  - Ricoh PCIe 1394 controllers.
Without these, several protocols, e.g. AV/C, do not work on the
Audigy, and the Ricoh PCIe controllers wouldn't work at all.
This does not concern the older Ricoh PCI controllers.

* tag 'firewire-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394:
  firewire: ohci: disable MSI on Ricoh controllers
  firewire: ohci: add reset packet quirk for SB Audigy

12 years agostaging: fix go7007-usb license
Randy Dunlap [Thu, 2 Feb 2012 03:04:02 +0000 (19:04 -0800)]
staging: fix go7007-usb license

Add MODULE_LICENSE() as per the license in the comment at the top of the
file for this source module to fix build warning:

  WARNING: modpost: missing MODULE_LICENSE() in drivers/staging/media/go7007/go7007-usb.o
  see include/linux/module.h for more information

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Cc: Ross Cohen <rcohen@snurgle.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agodocbook: fix fatal errors in device-drivers docbook and add DMA Management section
Randy Dunlap [Thu, 2 Feb 2012 02:15:49 +0000 (18:15 -0800)]
docbook: fix fatal errors in device-drivers docbook and add DMA Management section

Fix 2 fatal errors in the device-drivers docbook.
Also add some missing files from drivers/base/; since several
of these are DMA-related, add a section for DMA Management.

docproc: drivers/base/sys.c: No such file or directory
docproc: drivers/tty/serial/8250.c: No such file or directory

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoMerge branch 'for-linus2' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris...
Linus Torvalds [Thu, 2 Feb 2012 00:00:38 +0000 (16:00 -0800)]
Merge branch 'for-linus2' of git://git./linux/kernel/git/jmorris/linux-security

* 'for-linus2' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security:
  lib: Fix 32-bit sparc udiv_qrnnd() definition in mpilib's longlong.h
  lib: Fix multiple definitions of clz_tab
  lib/digsig: checks for NULL return value
  lib/mpi: added missing NULL check
  lib/mpi: added comment on divide by 0 case
  lib/mpi: check for possible zero length
  lib/digsig: pkcs_1_v1_5_decode_emsa cleanup
  lib/digsig: additional sanity checks against badly formated key payload
  lib/mpi: removed unused functions
  lib/mpi: checks for zero divisor length
  lib/mpi: return error code on dividing by zero
  lib/mpi: replaced MPI_NULL with normal NULL
  lib/mpi: added missing NULL check

12 years agolib: Fix 32-bit sparc udiv_qrnnd() definition in mpilib's longlong.h
David Miller [Wed, 1 Feb 2012 22:17:55 +0000 (00:17 +0200)]
lib: Fix 32-bit sparc udiv_qrnnd() definition in mpilib's longlong.h

This copy of longlong.h is extremely dated and results in compile
errors on sparc32 when MPILIB is enabled, copy over the more uptodate
implementation from arch/sparc/math/sfp-util_32.h

Reported-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: James Morris <jmorris@namei.org>
12 years agolib: Fix multiple definitions of clz_tab
David Miller [Wed, 1 Feb 2012 22:17:54 +0000 (00:17 +0200)]
lib: Fix multiple definitions of clz_tab

Both sparc 32-bit's software divide assembler and MPILIB provide
clz_tab[] with identical contents.

Break it out into a seperate object file and select it when
SPARC32 or MPILIB is set.

Reported-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: James Morris <jmorris@namei.org>
12 years agoMerge tag 'rdma-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland...
Linus Torvalds [Wed, 1 Feb 2012 23:18:39 +0000 (15:18 -0800)]
Merge tag 'rdma-for-linus' of git://git./linux/kernel/git/roland/infiniband

InfiniBand/RDMA fixes for 3.3:

 - Fix a crash due to a regression (uninitialized refcnt) introduced in
   3.2 with XRC support.
 - Close race in how ucma reports events when connect fails.
 - Process vendor-specific MADs in mlx4 so that eg FDR-10 data rate works.
 - Fix regression in qib caused by over-aggressive PCIe tuning.
 - Other small fixes for hardware drivers (ipath, nes, qib).

* tag 'rdma-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  RDMA/nes: Copyright update
  IB/mlx4: pass SMP vendor-specific attribute MADs to firmware
  RDMA/nes: Fix fast memory registration opcode
  RDMA/nes: Fix fast memory registration length
  RDMA/ucma: Discard all events for new connections until accepted
  IB/qib: Roll back PCIe tuning change
  IB/qib: Use GFP_ATOMIC when locks are held
  RDMA/nes: Add missing rcu_read_unlock() in nes_addr_resolve_neigh()
  RDMA/nes: Fix for sending MPA reject frame
  IB/ipath: Calling PTR_ERR() on right variable in create_file()
  RDMA/core: Fix kernel panic by always initializing qp->usecnt

12 years agoproc: make sure mem_open() doesn't pin the target's memory
Oleg Nesterov [Tue, 31 Jan 2012 16:15:11 +0000 (17:15 +0100)]
proc: make sure mem_open() doesn't pin the target's memory

Once /proc/pid/mem is opened, the memory can't be released until
mem_release() even if its owner exits.

Change mem_open() to do atomic_inc(mm_count) + mmput(), this only
pins mm_struct. Change mem_rw() to do atomic_inc_not_zero(mm_count)
before access_remote_vm(), this verifies that this mm is still alive.

I am not sure what should mem_rw() return if atomic_inc_not_zero()
fails. With this patch it returns zero to match the "mm == NULL" case,
may be it should return -EINVAL like it did before e268337d.

Perhaps it makes sense to add the additional fatal_signal_pending()
check into the main loop, to ensure we do not hold this memory if
the target task was oom-killed.

Cc: stable@kernel.org
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoproc: unify mem_read() and mem_write()
Oleg Nesterov [Tue, 31 Jan 2012 16:14:54 +0000 (17:14 +0100)]
proc: unify mem_read() and mem_write()

No functional changes, cleanup and preparation.

mem_read() and mem_write() are very similar. Move this code into the
new common helper, mem_rw(), which takes the additional "int write"
argument.

Cc: stable@kernel.org
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoproc: mem_release() should check mm != NULL
Oleg Nesterov [Tue, 31 Jan 2012 16:14:38 +0000 (17:14 +0100)]
proc: mem_release() should check mm != NULL

mem_release() can hit mm == NULL, add the necessary check.

Cc: stable@kernel.org
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoMAINTAINERS: Greg's suse email address is dead
Greg KH [Wed, 1 Feb 2012 04:02:00 +0000 (20:02 -0800)]
MAINTAINERS: Greg's suse email address is dead

My email address has changed, the suse.de one is now dead, so update all
of my MAINTAINER entries with the correct one so that patches don't get
lost.

Also change the status of some of my entries as I'm supposed to be doing
this stuff now for real.

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agomtd: fix merge conflict resolution breakage
Artem Bityutskiy [Wed, 1 Feb 2012 19:10:24 +0000 (11:10 -0800)]
mtd: fix merge conflict resolution breakage

This patch fixes merge conflict resolution breakage introduced by merge
d3712b9dfcf4 ("Merge tag 'for-linus' of git://github.com/prasad-joshi/logfs_upstream").

The commit changed 'mtd_can_have_bb()' function and made it always
return zero, which is incorrect.  Instead, we need it to return whether
the underlying flash device can have bad eraseblocks or not.  UBI needs
this information because it affects how it handles the underlying flash.
E.g., if the underlying flash is NOR, it cannot have bad blocks and any
write or erase error is fatal, and all we can do is to switch to R/O
mode.  We do not need to reserve a pool of good eraseblocks for bad
eraseblocks handling, and so on.

This patch also removes 'mtd_can_have_bb()' invocations from Logfs to
ensure correct Logfs behavior.

I've tested that with this patch UBI works on top of NOR and NAND
flashes emulated by mtdram and nandsim correspondingly.

This patch is based on patch from Linus Torvalds.

Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Acked-by: Jörn Engel <joern@logfs.org>
Acked-by: Prasad Joshi <prasadjoshi.linux@gmail.com>
Acked-by: Brian Norris <computersforpeace@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoperf evlist: Make splice_list_tail method public
Jiri Olsa [Fri, 27 Jan 2012 14:34:20 +0000 (15:34 +0100)]
perf evlist: Make splice_list_tail method public

Making perf_evlist__splice_list_tail globaly accessible.

It is used in the upcomming paches.

Cc: Ingo Molnar <mingo@elte.hu>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Link: http://lkml.kernel.org/r/1327674868-10486-2-git-send-email-jolsa@redhat.com
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
12 years agodrm/radeon/kms: Fix device tree linkage of DP i2c buses too
Jean Delvare [Tue, 31 Jan 2012 08:55:21 +0000 (09:55 +0100)]
drm/radeon/kms: Fix device tree linkage of DP i2c buses too

Properly set the parent device of DP i2c buses before registering them
too.

Signed-off-by: Jean Delvare <jdelvare@suse.de>
Reviewed-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agodrm/radeon: Set DESKTOP_HEIGHT register to the framebuffer (not mode) height.
Michel Dänzer [Wed, 1 Feb 2012 11:09:55 +0000 (12:09 +0100)]
drm/radeon: Set DESKTOP_HEIGHT register to the framebuffer (not mode) height.

The value of this register is transferred to the V_COUNTER register at the
beginning of vertical blank. V_COUNTER is the reference for VLINE waits and
goes from VIEWPORT_Y_START to VIEWPORT_Y_START+VIEWPORT_HEIGHT during scanout,
so if VIEWPORT_Y_START is not 0, V_COUNTER actually went backwards at the
beginning of vertical blank, and VLINE waits excluding the whole scanout area
could never finish (possibly only if VIEWPORT_Y_START is larger than the length
of vertical blank in scanlines). Setting DESKTOP_HEIGHT to the framebuffer
height should prevent this for any kind of VLINE wait.

Fixes https://bugs.freedesktop.org/show_bug.cgi?id=45329 .

CC: stable@vger.kernel.org
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agodrm/radeon/kms: disable output polling when suspended
Seth Forshee [Wed, 1 Feb 2012 01:06:25 +0000 (19:06 -0600)]
drm/radeon/kms: disable output polling when suspended

Polling the outputs when the device is suspended can result in erroneous
status updates. Disable output polling during suspend to prevent this
from happening.

Signed-off-by: Seth Forshee <seth.forshee@canonical.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>