perf tools: Fix attributes for '{}' defined event groups

Fixing events attributes for groups defined via '{}'.

Currently 'enable_on_exec' attribute in record command and both
'disabled ' and 'enable_on_exec' attributes in stat command are set
based on the 'group' option. This eliminates proper setup for '{}'
defined groups as they don't set 'group' option.

Making above attributes values based on the 'evsel->leader' as this is
common to both group definition.

Moving perf_evlist__set_leader call within builtin-record ahead
perf_evlist__config_attrs call, because the latter needs possible group
leader links in place.

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.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>
Link: http://lkml.kernel.org/r/1352741644-16809-2-git-send-email-jolsa@redhat.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
This commit is contained in:
Jiri Olsa
2012-11-12 18:34:00 +01:00
committed by Arnaldo Carvalho de Melo
parent 6064803313
commit cac2142557
8 changed files with 20 additions and 28 deletions

View File

@@ -404,8 +404,8 @@ const char *perf_evsel__name(struct perf_evsel *evsel)
return evsel->name ?: "unknown";
}
void perf_evsel__config(struct perf_evsel *evsel, struct perf_record_opts *opts,
struct perf_evsel *first)
void perf_evsel__config(struct perf_evsel *evsel,
struct perf_record_opts *opts)
{
struct perf_event_attr *attr = &evsel->attr;
int track = !evsel->idx; /* only the first counter needs these */
@@ -486,10 +486,8 @@ void perf_evsel__config(struct perf_evsel *evsel, struct perf_record_opts *opts,
attr->mmap = track;
attr->comm = track;
if (perf_target__none(&opts->target) &&
(!opts->group || evsel == first)) {
if (perf_target__none(&opts->target) && (!evsel->leader))
attr->enable_on_exec = 1;
}
}
int perf_evsel__alloc_fd(struct perf_evsel *evsel, int ncpus, int nthreads)