linux-kernel - Re: [PATCH 3/3] perf, tool: Add new event group management

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: <20120322124049.GE21163@infradead.org>
Date:	Thu, 22 Mar 2012 09:40:49 -0300
From:	Arnaldo Carvalho de Melo <acme@...hat.com>
To:	Jiri Olsa <jolsa@...hat.com>
Cc:	Peter Zijlstra <a.p.zijlstra@...llo.nl>, mingo@...e.hu,
	paulus@...ba.org, cjashfor@...ux.vnet.ibm.com, fweisbec@...il.com,
	linux-kernel@...r.kernel.org, Thomas Gleixner <tglx@...utronix.de>
Subject: Re: [PATCH 3/3] perf, tool: Add new event group management

Em Wed, Mar 21, 2012 at 11:27:24PM +0100, Jiri Olsa escreveu:
> On Wed, Mar 21, 2012 at 12:11:19PM -0300, Arnaldo Carvalho de Melo wrote:
> > Em Wed, Mar 21, 2012 at 11:42:14AM +0100, Jiri Olsa escreveu:
> > > On Tue, Mar 20, 2012 at 07:44:41PM +0100, Peter Zijlstra wrote:
> > > > If so, we need some option like:
> > > > 
> > > >   perf record -e "{cpu-clock,cache-misses,cache-references}:1"
> > > > 
> > > > to mean, only sample on cpu-clock but use PERF_SAMPLE_READ and
> > > > PERF_FORMAT_GROUP to read all siblings on every cpu-clock sample.
> > > > 
> > > > Now the disadvantage is that {} needs quotes on bash, the advantage is
> > > > that its completely natural on how to construct groups, without weird
> > > > --group/--group parsed. Also it provides a place for group modifiers.
> > > 
> > > how about '=group' keyword followed by ':' modifier
> > >   perf record -e "cpu-clock,cache-misses,cache-references=group:1"
> > > 
> > > or '=$groupname' and use the group name in output like
> > >   perf stat -e "task-clock,context-switches=krava:1" ls
> > > 
> > >      Performance counter stats for 'ls':
> > > 
> > >           1.003695 task-clock/krava          #    0.083 CPUs utilized          
> > >                  0 context-switches/krava    #    0.000 K/sec                  
> > > 
> > 
> > Humm, interesting, perhaps putting the group name before the =event_list
> > will be more intuitive?
> > 
> > I.e.:
> > 
> >   perf stat -e "krava=task-clock,context-switches:1" ls
> 
> yep, I put it at the end because of the ':' modifier stuff.
> We have it for each event, and it'd be in conflict having
> it at the end..
> 
> maybe we could put it like:
>    perf stat -e "krava:1=task-clock,context-switches" ls
> 
> or put some other separator instead of ':' ... but ':' is the best ;)

That thought crossed my mind as well, indeed it seems more intuitive:
everything before the '=' defines the group wide attributes (name,
modifiers, whatevever else we need later) after it, the evlist, where
each evsel can have its own attributes.

- Arnaldo
--
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/