[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20151027135154.GA21699@krava.brq.redhat.com>
Date: Tue, 27 Oct 2015 14:51:54 +0100
From: Jiri Olsa <jolsa@...hat.com>
To: Arnaldo Carvalho de Melo <acme@...nel.org>
Cc: Jiri Olsa <jolsa@...nel.org>, lkml <linux-kernel@...r.kernel.org>,
David Ahern <dsahern@...il.com>,
Ingo Molnar <mingo@...nel.org>,
Namhyung Kim <namhyung@...nel.org>,
Peter Zijlstra <a.p.zijlstra@...llo.nl>,
"Liang, Kan" <kan.liang@...el.com>
Subject: Re: [PATCH 11/52] perf tools: Add stat config event
On Tue, Oct 27, 2015 at 10:44:40AM -0300, Arnaldo Carvalho de Melo wrote:
> Em Tue, Oct 27, 2015 at 02:30:43PM +0100, Jiri Olsa escreveu:
> > On Tue, Oct 27, 2015 at 10:16:05AM -0300, Arnaldo Carvalho de Melo wrote:
> > > Em Sun, Oct 25, 2015 at 03:51:27PM +0100, Jiri Olsa escreveu:
> > > > Adding stat config event to pass/store stat config data,
> > > > so report tools (report/script) know how to interpret
> > > > stat data.
> > > >
> > > > The config data are stored in 'tag|value' way to allow
> > > > easy extension and backward compatibility.
> > >
> > > I wonder if this couldn't be renamed 'PERF_RECORD_CONFIG' and just go
> > > on using one of those 2^64-1 tags for the 'stat config' needs, but then
> > > this is just a matter of changing the name of this event, which won't
> > > break anything when done.
> >
> > we use this event to update 'struct perf_stat_config',
> > so it's kind of stat specific
>
> What makes it specific? Isn't this just tag/value? If it was generic all
> it would take would be for us to register a range of values (or several)
> to be handled by the stat config reading routine.
>
> > > Other stuff we may want to have stored like this include sysctl, sysfs
> > > values, kernel command line options used, etc.
> >
> > I think we could easily add record specific event for this
> > once it's needed.. there's plenty of free numbers in user
> > events area ;-)
>
> Shhh, don't ask PeterZ about it ;-)
>
> But seriously, what makes this specific? Do you envision the stat code
> needing 2^64-1 tags?
nope ;-)
it's just the read function perf_event__read_stat_config takes the
'struct perf_stat_config *' and updates it based on the event data.
we would need to change the code to be more generic.. but as you said,
it's future ;-)
jirka
--
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/
Powered by blists - more mailing lists