linux-stable.git/tools/perf/util/stat-shadow.c, branch v5.3

perf stat: Always separate stalled cycles per insn

2019-07-23T12:03:46+00:00

The "stalled cycles per insn" is appended to "instructions" when the CPU
has this hardware counter directly. We should always make it a separate
line, which also aligns to the output when we hit the "if (total &&
avg)" branch.

Before:

  $ sudo perf stat --all-cpus --field-separator , --log-fd 1 -einstructions,cycles -- sleep 1
  4565048704,,instructions,64114578096,100.00,1.34,insn per cycle,,
  3396325133,,cycles,64146628546,100.00,,

After:

  $ sudo ./tools/perf/perf stat --all-cpus --field-separator , --log-fd 1 -einstructions,cycles -- sleep 1
  6721924,,instructions,24026790339,100.00,0.22,insn per cycle
  ,,,,,0.00,stalled cycles per insn
  30939953,,cycles,24025512526,100.00,,

Signed-off-by: Cong Wang 
Acked-by: Jiri Olsa 
Cc: Andi Kleen 
Link: http://lkml.kernel.org/r/20190517221039.8975-1-xiyou.wangcong@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo

perf tools: Use zfree() where applicable

2019-07-09T13:13:27+00:00

In places where the equivalent was already being done, i.e.:

   free(a);
   a = NULL;

And in placs where struct members are being freed so that if we have
some erroneous reference to its struct, then accesses to freed members
will result in segfaults, which we can detect faster than use after free
to areas that may still have something seemingly valid.

Cc: Adrian Hunter 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Link: https://lkml.kernel.org/n/tip-jatyoofo5boc1bsvoig6bb6i@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo

perf stat: Fix metrics with --no-merge

2019-07-02T01:50:41+00:00

Since Fixes: 8c5421c016a4 ("perf pmu: Display pmu name when printing
unmerged events in stat") using --no-merge adds the PMU name to the
evsel name.

This breaks the metric value lookup because the parser doesn't know
about this.

Remove the extra postfixes for the metric evaluation.

Signed-off-by: Andi Kleen 
Acked-by: Jiri Olsa 
Cc: Agustin Vega-Frias 
Cc: Kan Liang 
Fixes: 8c5421c016a4 ("perf pmu: Display pmu name when printing unmerged events in stat")
Link: http://lkml.kernel.org/r/20190624193711.35241-5-andi@firstfloor.org
Signed-off-by: Arnaldo Carvalho de Melo

perf stat: Make metric event lookup more robust

2019-07-02T01:50:41+00:00

After setting up metric groups through the event parser, the metricgroup
code looks them up again in the event list.

Make sure we only look up events that haven't been used by some other
metric. The data structures currently cannot handle more than one metric
per event. This avoids problems with multiple events partially
overlapping.

Signed-off-by: Andi Kleen 
Acked-by: Jiri Olsa 
Cc: Kan Liang 
Link: http://lkml.kernel.org/r/20190624193711.35241-2-andi@firstfloor.org
Signed-off-by: Arnaldo Carvalho de Melo

perf stat: Support per-die aggregation

2019-06-10T19:19:59+00:00

It is useful to aggregate counts per die. E.g. Uncore becomes die-scope
on Xeon Cascade Lake-AP.

Introduce a new option "--per-die" to support per-die aggregation.

The global id for each core has been changed to socket + die id + core
id. The global id for each die is socket + die id.

Add die information for per-core aggregation. The output of per-core
aggregation will be changed from "S0-C0" to "S0-D0-C0". Any scripts
which rely on the output format of per-core aggregation probably be
broken.

For 'perf stat record/report', there is no die information when
processing the old perf.data. The per-die result will be the same as
per-socket.

Committer notes:

Renamed 'die' variable to 'die_id' to fix the build in some systems:

    CC       /tmp/build/perf/builtin-script.o
  cc1: warnings being treated as errors
  builtin-stat.c: In function 'perf_env__get_die':
  builtin-stat.c:963: error: declaration of 'die' shadows a global declaration
  util/util.h:19: error: shadowed declaration is here
  mv: cannot stat `/tmp/build/perf/.builtin-stat.o.tmp': No such file or directory

Signed-off-by: Kan Liang 
Reviewed-by: Jiri Olsa 
Cc: Andi Kleen 
Cc: Peter Zijlstra 
Link: https://lkml.kernel.org/n/tip-bsnhx7vgsuu6ei307mw60mbj@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo

perf util: Use cached rbtree for rblists

2019-01-25T14:12:10+00:00

At the cost of an extra pointer, we can avoid the O(logN) cost of
finding the first element in the tree (smallest node), which is
something required for any of the strlist or intlist traversals
(XXX_for_each_entry()). There are a number of users in perf of these
(particularly strlists), including probes, and buildid.

Signed-off-by: Davidlohr Bueso 
Tested-by: Arnaldo Carvalho de Melo 
Cc: Jiri Olsa 
Cc: Namhyung Kim 
Link: http://lkml.kernel.org/r/20181206191819.30182-5-dave@stgolabs.net
Signed-off-by: Arnaldo Carvalho de Melo

perf stat: Fix shadow stats for clock events

2018-12-17T17:53:30+00:00

Commit 0aa802a79469 ("perf stat: Get rid of extra clock display
function") introduced scale and unit for clock events. Thus,
perf_stat__update_shadow_stats() now saves scaled values of clock events
in msecs, instead of original nsecs. But while calculating values of
shadow stats we still consider clock event values in nsecs. This results
in a wrong shadow stat values. Ex,

  # ./perf stat -e task-clock,cycles ls
    
              2.60 msec task-clock:u    #    0.877 CPUs utilized
         2,430,564      cycles:u        # 1215282.000 GHz

Fix this by saving original nsec values for clock events in
perf_stat__update_shadow_stats(). After patch:

  # ./perf stat -e task-clock,cycles ls
    
              3.14 msec task-clock:u    #    0.839 CPUs utilized
         3,094,528      cycles:u        #    0.985 GHz

Suggested-by: Jiri Olsa 
Reported-by: Anton Blanchard 
Signed-off-by: Ravi Bangoria 
Reviewed-by: Jiri Olsa 
Cc: Alexander Shishkin 
Cc: Jin Yao 
Cc: Namhyung Kim 
Cc: Thomas Richter 
Cc: yuzhoujian@didichuxing.com
Fixes: 0aa802a79469 ("perf stat: Get rid of extra clock display function")
Link: http://lkml.kernel.org/r/20181116042843.24067-1-ravi.bangoria@linux.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo

perf stat: Use perf_evsel__is_clocki() for clock events

2018-11-22T01:39:57+00:00

We already have function to check if a given event is either
SW_CPU_CLOCK or SW_TASK_CLOCK. Utilize it.

Signed-off-by: Ravi Bangoria 
Acked-by: Jiri Olsa 
Cc: Alexander Shishkin 
Cc: Anton Blanchard 
Cc: Jin Yao 
Cc: Namhyung Kim 
Cc: Thomas Richter 
Cc: yuzhoujian@didichuxing.com
Link: http://lkml.kernel.org/r/20181115095533.16930-1-ravi.bangoria@linux.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo

perf stat: Pass a 'struct perf_stat_config' argument to global print functions

2018-08-30T18:52:23+00:00

Add 'struct perf_stat_config' argument to the global print functions, so
that these functions can be used out of the 'perf stat' command code.

Signed-off-by: Jiri Olsa 
Cc: Alexander Shishkin 
Cc: Andi Kleen 
Cc: David Ahern 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Link: http://lkml.kernel.org/r/20180830063252.23729-20-jolsa@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo

perf stat: Get rid of extra clock display function

2018-07-24T17:54:58+00:00

There's no reason to have separate function to display clock events.
It's only purpose was to convert the nanosecond value into microseconds.
We do that now in generic code, if the unit and scale values are
properly set, which this patch do for clock events.

The output differs in the unit field being displayed in its columns
rather than having it added as a suffix of the event name. Plus the
value is rounded into 2 decimal numbers as for any other event.

Before:

  # perf stat  -e cpu-clock,task-clock -C 0 sleep 3

   Performance counter stats for 'CPU(s) 0':

       3001.123137      cpu-clock (msec)          #    1.000 CPUs utilized
       3001.133250      task-clock (msec)         #    1.000 CPUs utilized

       3.001159813 seconds time elapsed

Now:

  # perf stat  -e cpu-clock,task-clock -C 0 sleep 3

   Performance counter stats for 'CPU(s) 0':

          3,001.05 msec cpu-clock                 #    1.000 CPUs utilized
          3,001.05 msec task-clock                #    1.000 CPUs utilized

       3.001077794 seconds time elapsed

There's a small difference in csv output, as we now output the unit
field, which was empty before. It's in the proper spot, so there's no
compatibility issue.

Before:

  # perf stat  -e cpu-clock,task-clock -C 0 -x, sleep 3
  3001.065177,,cpu-clock,3001064187,100.00,1.000,CPUs utilized
  3001.077085,,task-clock,3001077085,100.00,1.000,CPUs utilized

  # perf stat  -e cpu-clock,task-clock -C 0 -x, sleep 3
  3000.80,msec,cpu-clock,3000799026,100.00,1.000,CPUs utilized
  3000.80,msec,task-clock,3000799550,100.00,1.000,CPUs utilized

Add perf_evsel__is_clock to replace nsec_counter.

Signed-off-by: Jiri Olsa 
Tested-by: Arnaldo Carvalho de Melo 
Cc: Alexander Shishkin 
Cc: Andi Kleen 
Cc: David Ahern 
Cc: Namhyung Kim 
Cc: Peter Zijlstra 
Link: http://lkml.kernel.org/r/20180720110036.32251-2-jolsa@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo