linux-kernel - Re: [PATCH 3/4] perf: teach perf inject to merge sched_stat_* and sched

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20120827221423.GE16230@one.firstfloor.org>
Date:	Tue, 28 Aug 2012 00:14:23 +0200
From:	Andi Kleen <andi@...stfloor.org>
To:	David Ahern <dsahern@...il.com>
Cc:	Andi Kleen <andi@...stfloor.org>, Andrew Vagin <avagin@...nvz.org>,
	linux-kernel@...r.kernel.org,
	Peter Zijlstra <a.p.zijlstra@...llo.nl>,
	Paul Mackerras <paulus@...ba.org>,
	Ingo Molnar <mingo@...hat.com>,
	Arnaldo Carvalho de Melo <acme@...stprotocols.net>
Subject: Re: [PATCH 3/4] perf: teach perf inject to merge sched_stat_* and sched_switch events (v2)

On Mon, Aug 27, 2012 at 03:56:23PM -0600, David Ahern wrote:
> On 8/27/12 2:51 PM, Andi Kleen wrote:
> >Andrew Vagin <avagin@...nvz.org> writes:
> >>
> >>My code saves sched_switch event for each process and when it meets
> >>stat_iowait, it reports the sched_switch event, because this event
> >>contains a correct callchain. By another words it replaces all
> >>stat_iowait events on proper sched_switch events.
> >
> >Is there any way to generalize this to arbitary events? I found myself
> >writing scripts to compute lengths between various events.  But I
> 
> You mean delta-time between events? I have toyed around with patches to 
> perf-script for this -- dt between same event on a cpu, dt between 
> consecutive events on a cpu, dt between events for a thread, etc.

I did the same. But I then realized perf script is the wrong approach. 
perf script is a deadend as far as perf is concerned.
You want a filter that generates another perf.data that you then can 
view in the browser. So perf inject is the better model.

-Andi
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/