Date:   Fri, 19 Nov 2021 08:33:17 +0800
From:   Leo Yan <leo.yan@linaro.org>
To:     Mathieu Poirier <mathieu.poirier@linaro.org>
Cc:     Arnaldo Carvalho de Melo <acme@kernel.org>,
        Namhyung Kim <namhyung@kernel.org>,
        Suzuki K Poulose <suzuki.poulose@arm.com>,
        Mike Leach <mike.leach@linaro.org>,
        John Garry <john.garry@huawei.com>,
        Will Deacon <will@kernel.org>, Ingo Molnar <mingo@redhat.com>,
        Mark Rutland <mark.rutland@arm.com>,
        Alexander Shishkin <alexander.shishkin@linux.intel.com>,
        Jiri Olsa <jolsa@redhat.com>,
        linux-arm-kernel@lists.infradead.org, coresight@lists.linaro.org,
        linux-perf-users@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v1] perf cs-etm: Pass -1 as pid value for
 machine__set_current_tid()
Message-ID: <20211119003317.GD69886@leoy-ThinkPad-X240s>
References: <20211113143540.53957-1-leo.yan@linaro.org>
 <20211118171412.GB2530497@p14s>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20211118171412.GB2530497@p14s>
Precedence: bulk

Hi Mathieu,

On Thu, Nov 18, 2021 at 10:14:12AM -0700, Mathieu Poirier wrote:
> Good morning Leo,
> 
> On Sat, Nov 13, 2021 at 10:35:40PM +0800, Leo Yan wrote:
> > Currently, cs-etm passes the tid value for both tid and pid parameters
> > when calling machine__set_current_tid(), this can lead to confusion for
> > thread handling.  E.g. we arbitrarily pass the same value for pid and
> > tid, perf tool will be misled to consider it is a main thread (see
> > thread__main_thread()).
> > 
> > On the other hand, Perf tool only can retrieve tid from Arm CoreSight
> > context packet, and we have no chance to know pid (it maps to kernel's
> > task_struct::tgid) from hardware tracing data.  For this reason, this
> > patch passes -1 as pid for function machine__set_current_tid().
> > 
> > Signed-off-by: Leo Yan <leo.yan@linaro.org>
> > ---
> >  tools/perf/util/cs-etm.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/tools/perf/util/cs-etm.c b/tools/perf/util/cs-etm.c
> > index f323adb1af85..eed1a5930072 100644
> > --- a/tools/perf/util/cs-etm.c
> > +++ b/tools/perf/util/cs-etm.c
> > @@ -1118,7 +1118,7 @@ int cs_etm__etmq_set_tid(struct cs_etm_queue *etmq,
> >  	if (cs_etm__get_cpu(trace_chan_id, &cpu) < 0)
> >  		return err;
> >  
> > -	err = machine__set_current_tid(etm->machine, cpu, tid, tid);
> > +	err = machine__set_current_tid(etm->machine, cpu, -1, tid);
> 
> I remember wondering about what to do with the pid parameter when I wrote this
> patch... 
> 
> Do you have a before-and-after snapshot you can add to the changelog?

I tried to capture log but I didn't observe the difference introduced
by this patch, this might because I didn't per-process mode for
multi-threading case.  I will try more case for this.

> I also think it will require a "Fixes" tag.  In your next revision please CC James
> since you guys are working in that area nowadays.

Will do.  And will Cc James and German in next spin.

Thanks for review and suggestion.

Leo