From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S964980AbbESO7s (ORCPT <rfc822;w@1wt.eu>);
	Tue, 19 May 2015 10:59:48 -0400
Received: from mail-pd0-f169.google.com ([209.85.192.169]:33516 "EHLO
	mail-pd0-f169.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S964947AbbESO7p (ORCPT
	<rfc822;linux-kernel@vger.kernel.org>);
	Tue, 19 May 2015 10:59:45 -0400
Date: Tue, 19 May 2015 23:58:22 +0900
From: Namhyung Kim <namhyung@kernel.org>
To: Jiri Olsa <jolsa@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>, Ingo Molnar <mingo@kernel.org>,
        Peter Zijlstra <a.p.zijlstra@chello.nl>,
        LKML <linux-kernel@vger.kernel.org>, David Ahern <dsahern@gmail.com>,
        Adrian Hunter <adrian.hunter@intel.com>,
        Andi Kleen <andi@firstfloor.org>,
        Frederic Weisbecker <fweisbec@gmail.com>,
        Stephane Eranian <eranian@google.com>
Subject: Re: [PATCH 34/40] perf report: Parallelize perf report using
 multi-thread
Message-ID: <20150519145822.GB29162@danjae.kornet>
References: <1431909055-21442-1-git-send-email-namhyung@kernel.org>
 <1431909055-21442-35-git-send-email-namhyung@kernel.org>
 <20150519101230.GC22060@krava.redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <20150519101230.GC22060@krava.redhat.com>
User-Agent: Mutt/1.5.23 (2014-03-12)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On Tue, May 19, 2015 at 12:12:30PM +0200, Jiri Olsa wrote:
> On Mon, May 18, 2015 at 09:30:49AM +0900, Namhyung Kim wrote:
> 
> SNIP
> 
> > +
> > +int perf_session__process_events_mt(struct perf_session *session, void *arg)
> > +{
> > +	struct perf_data_file *file = session->file;
> > +	struct perf_evlist *evlist = session->evlist;
> > +	struct perf_evsel *evsel;
> > +	u64 nr_entries = 0;
> > +	struct perf_tool *tool = session->tool;
> > +	struct perf_session *mt_sessions = NULL;
> > +	struct perf_tool_mt *mt_tools = NULL;
> > +	struct perf_session *ms;
> > +	struct perf_tool_mt *mt;
> > +	pthread_t *th_id;
> > +	int err, i, k;
> > +	int nr_index = session->header.nr_index;
> > +	u64 size = perf_data_file__size(file);
> > +
> 
> SNIP
> 
> > +
> > +	for (i = 1; i < nr_index; i++) {
> > +		mt = &mt_tools[i];
> > +
> > +		evlist__for_each(evlist, evsel) {
> > +			struct hists *hists = evsel__hists(evsel);
> > +
> > +			if (perf_evsel__is_dummy_tracking(evsel))
> > +				continue;
> > +
> > +			hists__mt_resort(hists, &mt->hists[evsel->idx]);
> > +
> > +			/* Non-group events are considered as leader */
> > +			if (symbol_conf.event_group &&
> > +			    !perf_evsel__is_group_leader(evsel)) {
> > +				struct hists *leader_hists;
> > +
> > +				leader_hists = evsel__hists(evsel->leader);
> > +				hists__match(leader_hists, hists);
> > +				hists__link(leader_hists, hists);
> 
> hum, could you please describe/comment on why is this needed?

This is because it skips report__collapse_hists() when multi-thread
processing is enabled.  For multi-thread, it needs to collect
per-thread hists into a global hists.  This step is very similar to
hists__collapse_resort() and called regardless of sort keys, so it
skips to call the report__collapse_hists().

But I think it should be called after collapsing all per-thread hists.
I'll move it out of the loop.

Thanks,
Namhyung