linux.git/tools/perf/builtin-stat.c, branch v3.12

perf stat: Set child_pid after perf_evlist__prepare_workload()

2013-10-04T18:16:05+00:00

The commit acf2892270dc ("perf stat: Use perf_evlist__prepare/
start_workload()") converted to use the function but forgot to update
child_pid.  Fix it.

Signed-off-by: Namhyung Kim 
Cc: David Ahern 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Paul Mackerras 
Cc: Peter Zijlstra 
Cc: Stephane Eranian 
Link: http://lkml.kernel.org/r/1380531671-28076-1-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo

perf stat: Flush output after each line in interval mode

2013-08-07T20:35:29+00:00

When interval mode is outputting to a pipe, each measurement should be
flushed individually, so that the reader sees it timely.

With a terminal each line is automatically flushed by stdio, but that is
disabled with non terminal output.

Simply fflush output after each time interval

Signed-off-by: Andi Kleen 
Reviewed-by: Jiri Olsa 
Cc: Jiri Olsa 
Cc: Stephane Eranian 
Link: http://lkml.kernel.org/r/1375490473-1503-5-git-send-email-andi@firstfloor.org
Signed-off-by: Arnaldo Carvalho de Melo

perf stat: Add support for --initial-delay option

2013-08-07T20:35:29+00:00

When measuring workloads the startup phase -- doing page faults, dynamic
linking, opening files -- is often very different from the rest of the
workload.  Especially with smaller kernels and using counter
multiplexing this can give significant measurement errors.

Multiplexing assumes that the workload is mostly the same over longer
periods. But at startup there is typically some spike of activity which
is relatively short.  If many groups are multiplexing the one group
seeing the spike, and which is then scaled up over the time to run all
groups, may see a significant error.

Also in general it's often not useful to measure the startup, because it
is so different from the rest.

One way around this is to use interval mode and discard the first
sample, but this can be awkward because interval mode doesn't support
intervals of less than 100ms, and also a useful interval is not
necessarily the same as a useful startup delay.

This patch adds a new --initial-delay / -D option to skip measuring for
the startup phase. The time can be specified in ms

Here's a simple example:

perf stat -e page-faults bash -c 'for i in $(seq 100000) ; do true ; done'
...
             3,721 page-faults
...

If we just wait 20 ms the number of page faults is 1/3 less:

perf stat -D 20 -e page-faults bash -c 'for i in $(seq 100000) ; do true ; done'
...
             2,823 page-faults
...

So we filtered out most of the startup noise from bash.

Signed-off-by: Andi Kleen 
Reviewed-by: Jiri Olsa 
Cc: Jiri Olsa 
Cc: Stephane Eranian 
Link: http://lkml.kernel.org/r/1375490473-1503-4-git-send-email-andi@firstfloor.org
Signed-off-by: Arnaldo Carvalho de Melo

perf stat: Fix per-socket output bug for uncore events

2013-07-08T21:01:46+00:00

This patch fixes a problem reported by Andi Kleen on perf
stat when measuring uncore events:

 # perf stat --per-socket -e uncore_pcu/event=0x0/ -I1000  -a sleep 2

It would not report counts for the second socket. That was due to a
cpu mapping bug in print_aggr().

This patch also fixes the socket numbering bug for 
events.

Reported-by: Andi Kleen 
Signed-off-by: Stephane Eranian 
Tested-by: Andi Kleen 
Cc: Andi Kleen 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Cc: zheng.z.yan@intel.com
Link: http://lkml.kernel.org/r/20130705170645.GA32519@quad
Signed-off-by: Arnaldo Carvalho de Melo

perf stat: Avoid sending SIGTERM to random processes

2013-07-08T20:36:33+00:00

This patch fixes a problem with perf stat whereby on termination it may
send a SIGTERM signal to random processes on systems with high PID
recycling. I got some actual bug reports on this.

There is race between the SIGCHLD and sig_atexit() handlers.  This patch
addresses this problem by clearing child_pid in the SIGCHLD handler.

Signed-off-by: Stephane Eranian 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Link: http://lkml.kernel.org/r/20130604154426.GA2928@quad
Signed-off-by: Arnaldo Carvalho de Melo

perf stat: Add per-core aggregation

2013-03-25T19:13:26+00:00

This patch adds the --per-core option to perf stat.

This option is used to aggregate system-wide counts
on a per physical core basis. On processors with
hyperthreading, this means counts of all HT threads
running on a physical core are aggregated.

This mode is useful to find imblance between physical
cores running an uniform workload. Cores are identified
by socket: S0-C1, means physical core 1 on socket 0. Note
that cores are identified using their physical core id,
thus their numbering may not be continuous.

Per core aggregation can be combined with interval printing:

 # perf stat -a --per-core -I 1000 -e cycles sleep 1000
 #           time core         cpus             counts events
      1.000090030 S0-C0           1          4,765,747 cycles
      1.000090030 S0-C1           1          5,580,647 cycles
      1.000090030 S0-C2           1            221,181 cycles
      1.000090030 S0-C3           1            266,092 cycles

Signed-off-by: Stephane Eranian 
Cc: Andi Kleen 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Link: http://lkml.kernel.org/r/1360846649-6411-4-git-send-email-eranian@google.com
[ committer note: Remove parts already applied on 86ee6e1 to keep bisectability ]
Signed-off-by: Arnaldo Carvalho de Melo

perf stat: Rename --aggr-socket to --per-socket

2013-03-25T19:09:24+00:00

To make it more obvious what this option does as suggested by Andi on
LKML.

Signed-off-by: Stephane Eranian 
Cc: Andi Kleen 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Link: http://lkml.kernel.org/r/1360846649-6411-3-git-send-email-eranian@google.com
Signed-off-by: Arnaldo Carvalho de Melo

perf stat: Refactor aggregation code

2013-03-25T18:29:53+00:00

Refactor aggregation code by introducing a single aggr_mode variable and an
enum for aggregation.

Also refactor cpumap code having to do with cpu to socket mappings. All in
preparation for extended modes, such as cpu -> core.

Also fix socket aggregation and ensure that sockets are printed in increasing
order.

Signed-off-by: Stephane Eranian 
Cc: Andi Kleen 
Cc: Ingo Molnar 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Link: http://lkml.kernel.org/r/1360846649-6411-2-git-send-email-eranian@google.com
[ committer note: Fixup conflicts with a7e191c "--repeat forever" and
  acf2892 "Use perf_evlist__prepare/start_workload()" ]
Signed-off-by: Arnaldo Carvalho de Melo

perf stat: Introduce evlist methods to allocate/free the stats

2013-03-18T14:34:09+00:00

Reducing the noise in the main logic.

Cc: David Ahern 
Cc: Frederic Weisbecker 
Cc: Jiri Olsa 
Cc: Mike Galbraith 
Cc: Namhyung Kim 
Cc: Paul Mackerras 
Cc: Peter Zijlstra 
Cc: Stephane Eranian 
Link: http://lkml.kernel.org/n/tip-o219lnci04hlilxi6711wtcr@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo

perf stat: Introduce --repeat forever

2013-03-15T17:01:26+00:00

The following patch causes 'perf stat --repeat 0' to be interpreted as
'forever', displaying the stats for every run.

We act as if a single run was asked, and reset the stats in each
iteration. In this mode SIGINT is passed to perf to be able to stop the
loop with Ctrl+C.

Signed-off-by: Frederik Deweerdt 
Cc: Ingo Molnar 
Cc: Paul Mackerras 
Cc: Peter Zijlstra 
Link: http://lkml.kernel.org/r/20130301180227.GA24385@ks398093.ip-192-95-24.net
Signed-off-by: Arnaldo Carvalho de Melo