[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20161206112107.GK3124@twins.programming.kicks-ass.net>
Date: Tue, 6 Dec 2016 12:21:07 +0100
From: Peter Zijlstra <peterz@...radead.org>
To: kan.liang@...el.com
Cc: mingo@...hat.com, acme@...nel.org, linux-kernel@...r.kernel.org,
alexander.shishkin@...ux.intel.com, tglx@...utronix.de,
namhyung@...nel.org, jolsa@...nel.org, adrian.hunter@...el.com,
wangnan0@...wei.com, mark.rutland@....com, andi@...stfloor.org
Subject: Re: [PATCH V2 02/13] perf/core: output overhead when sched out from
context
On Fri, Dec 02, 2016 at 04:19:10PM -0500, kan.liang@...el.com wrote:
> From: Kan Liang <kan.liang@...el.com>
>
> Outputing every overhead when it happens is very costly. The accumulated
> time is more meaningful. So the overhead information should be outputted
> at the very end.
>
> The overhead information is outputted when task is scheduling out or the
> event is going to be disabled.
Why? The context switch stuff can happen very frequently. Doesn't seem
like a good place to add overhead.
> The arch specific overhead is outputted in event pmu delete, when Flag
> PERF_EF_LOG is set.
Powered by blists - more mailing lists