perf annotate: Add comment about annotated_src->nr_histograms
authorArnaldo Carvalho de Melo <acme@redhat.com>
Fri, 25 May 2018 14:27:38 +0000 (11:27 -0300)
committerArnaldo Carvalho de Melo <acme@redhat.com>
Mon, 4 Jun 2018 13:28:52 +0000 (10:28 -0300)
commit9132d3d92d8953ee79690408a9a24d938be22cd8
tree385f92a51cfd839aeaeb04e3c2ce15b4b6de017e
parent9fd5578a3cdd6cb59603b04a9282199d08f85607
perf annotate: Add comment about annotated_src->nr_histograms

When we have multiple groups in an evlist, say:

  $ perf stat -e '{cycles,instructions},{cache-references,cache-misses}' sleep 1

   Performance counter stats for 'sleep 1':

           343,134      cycles:u
           249,292      instructions:u            #    0.73  insn per cycle
            15,556      cache-references:u
             8,925      cache-misses:u            #   57.373 % of all cache refs

       1.000957550 seconds time elapsed

  $

Then the perf_evsel instances for the two group leaders ("cycles" and
"cache-references") will have evsel->nr_members set to 2, while all the
evsel->evlist->nr_entries will be set to 4, so we can't use
evsel->evlist->nr_entries everywhere, as event groups need to be taken
into account.

But this probably requires us to audit at least the forced-group code,
where we want all of the events to be in a "group", to see them all in
the screen, one column for each, even knowing that they were not
necessarily scheduled to count at the same time by the kernel perf
subsystem.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Wang Nan <wangnan0@huawei.com>
Link: https://lkml.kernel.org/n/tip-2g0vwqnc49wl4ttjk8dvpgcc@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
tools/perf/util/annotate.h