Re: [perfmon2] [PATCH 1/2] perf_events: add support for per-cpu per-cgroup monitoring (v5)

[Stephane E. <er...@go...>]

On Thu, Nov 25, 2010 at 12:20 PM, Peter Zijlstra <pe...@in...> wrote:
>
> On Thu, 2010-11-18 at 12:40 +0200, Stephane Eranian wrote:
> > @@ -919,6 +945,10 @@ static inline void perf_event_task_sched_in(struct task_struct *task)
> >  static inline
> >  void perf_event_task_sched_out(struct task_struct *task, struct task_struct *next)
> >  {
> > +#ifdef CONFIG_CGROUPS
> > +       atomic_t *cgroup_events = &__get_cpu_var(perf_cgroup_events);
> > +       COND_STMT(cgroup_events, perf_cgroup_switch(task, next));
> > +#endif
> >         COND_STMT(&perf_task_events, __perf_event_task_sched_out(task, next));
> >  }
>
> I don't think that'll actually work, the jump label stuff needs a static
> address.
>

I did not know that.

> Why not simply: s/perf_task_events/perf_sched_events/ and increment it
> for cgroup events as well?
>

But you would need to demultiplex. that's not because perf_sched_events is
set that you want BOTH perf_cgroup_switch() AND perf_event_task_sched_out().
Cgroup mode is per-cpu only.