Re: [PATCH] perf report: Fix --task and --stat with pipe input

From: Arnaldo Carvalho de Melo <acme@kernel.org>
To: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>,
	Ingo Molnar <mingo@kernel.org>,
	Peter Zijlstra <peterz@infradead.org>,
	LKML <linux-kernel@vger.kernel.org>,
	Andi Kleen <ak@linux.intel.com>, Ian Rogers <irogers@google.com>,
	stable@kernel.org
Subject: Re: [PATCH] perf report: Fix --task and --stat with pipe input
Date: Fri, 2 Jul 2021 09:46:08 -0300	[thread overview]
Message-ID: <YN8KkP759PDZY6dQ@kernel.org> (raw)
In-Reply-To: <YN7skX0uRz5SrB0w@krava>

Em Fri, Jul 02, 2021 at 12:38:09PM +0200, Jiri Olsa escreveu:
> On Tue, Jun 29, 2021 at 09:30:58PM -0700, Namhyung Kim wrote:
> > Current perf report fails to process a pipe input when --task
> > or --stat option is used.  This is because they reset all the
> > tool callbacks and failed to find a matching event for a sample.
> > 
> > When pipe input is used, the event info is passed via ATTR records
> > so it needs to handle that operation.  Otherwise the following error
> > occurred.  Note, -14 (= -EFAULT) comes from evlist__parse_sample():
> > 
> >   # perf record -a -o- sleep 1 | perf report -i- --stat
> >   Can't parse sample, err = -14
> >   0x271044 [0x38]: failed to process type: 9
> >   Error:
> >   failed to process sample
> > 
> > Signed-off-by: Namhyung Kim <namhyung@kernel.org>
> 
> Acked-by: Jiri Olsa <jolsa@redhat.com>

Added this for the the benefit of stable@kernel.org:

Fixes: a4a4d0a7a2b20f78 ("perf report: Add --stats option to display quick data statistics")

> thanks,
> jirka
> 
> > ---
> >  tools/perf/builtin-report.c | 6 ++++++
> >  1 file changed, 6 insertions(+)
> > 
> > diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
> > index bc5c393021dc..8639bbe0969d 100644
> > --- a/tools/perf/builtin-report.c
> > +++ b/tools/perf/builtin-report.c
> > @@ -729,9 +729,14 @@ static int count_sample_event(struct perf_tool *tool __maybe_unused,
> >  	return 0;
> >  }
> >  
> > +static int process_attr(struct perf_tool *tool __maybe_unused,
> > +			union perf_event *event,
> > +			struct evlist **pevlist);
> > +
> >  static void stats_setup(struct report *rep)
> >  {
> >  	memset(&rep->tool, 0, sizeof(rep->tool));
> > +	rep->tool.attr = process_attr;
> >  	rep->tool.sample = count_sample_event;
> >  	rep->tool.no_warn = true;
> >  }
> > @@ -753,6 +758,7 @@ static void tasks_setup(struct report *rep)
> >  		rep->tool.mmap = perf_event__process_mmap;
> >  		rep->tool.mmap2 = perf_event__process_mmap2;
> >  	}
> > +	rep->tool.attr = process_attr;
> >  	rep->tool.comm = perf_event__process_comm;
> >  	rep->tool.exit = perf_event__process_exit;
> >  	rep->tool.fork = perf_event__process_fork;
> > -- 
> > 2.32.0.93.g670b81a890-goog
> > 
> 

-- 

- Arnaldo