Re: [PATCH 0/1] (Was uprobes/perf: pre-filtering)

[Oleg Nesterov <oleg@redhat.com>]

Hi,

On 02/07, Namhyung Kim wrote:
>
> On Wed, 6 Feb 2013 20:42:18 +0100, Oleg Nesterov wrote:
> >
> > And why opts->target.system_wide is only set by OPT_BOOLEAN("all-cpus") ?
> > I meant, why I can't do "perf record -e whatever -C0" to create a "global"
> > counter on CPU_0? This doesn't work because __cmd_record() sees !.system_wide
> > and assumes we need perf_event__synthesize_thread_map() which silently fail.
> >
> > So I am sending a single patch to fix the problem which complicated my
> > testing. It is trivial but I am not sure it correct, please review.
>
> Yes, it's not clear how it handles above (-C0) case.  I think it should
> be treated as a system_wide mode like --all-cpus (-a).  So we could set
> ->system_wide to true if -C is given and/or test perf_target__has_cpu()
> for perf_event__synthesize_thread_map() or both.

Yes, thanks... but to be honest I do not understand opts->systemwide as
well.

OK, both 'perf record ... sleep 1' and 'perf record ... -a sleep 1' attach
the counter(s) to the child process, but opts->systemwide differs. In the
latter case run_command() does perf_event__synthesize_threads(). Not sure
this is right.

Btw, I am just curious. You can override the target with "-p" and run the
command, but it seems that it is not possible to create a global counter
and run the command. Not that important, but could be handy.

Oleg.