Re: [PATCH 2/4] perf: Add exclude_task perf event attribute

[Frederic Weisbecker <fweisbec@gmail.com>]

On Tue, May 25, 2010 at 08:58:08AM +0200, Peter Zijlstra wrote:
> On Tue, 2010-05-25 at 11:43 +1000, Paul Mackerras wrote:
> > On Fri, May 21, 2010 at 04:05:13PM +0200, Frederic Weisbecker wrote:
> > 
> > > Excluding is useful when you want to trace only hard and softirqs.
> > > 
> > > For this we use a new generic perf_exclude_event() (the previous
> > > one beeing turned into perf_exclude_swevent) to which you can pass
> > > the preemption offset to which your events trigger.
> > > 
> > > Computing preempt_count() - offset gives us the preempt_count() of
> > > the context that the event has interrupted, on top of which we
> > > can filter the non-irq contexts.
> > 
> > How does this work for hardware events when we are sampling and
> > getting an interrupt every N events?  It seems like the hardware is
> > still counting all events and interrupting every N events, but we are
> > only recording a sample if the interrupt occurred in the context we
> > want.  In other words the context of the Nth event is considered to be
> > the context for the N-1 events preceding that, which seems a pretty
> > poor approximation.
> > 
> > Also, for hardware events, if we are counting rather than sampling,
> > the exclude_task bit will have no effect.  So perhaps in that case the
> > perf_event_open should fail rather than appear to succeed but give
> > wrong data.
> 
> Right, so for hardware event we'd need to go with those irq_{enter,exit}
> hooks and either fully disable the call, or do as Ingo suggested, read
> the count delta and add that to period_left, so that we'll delay the
> sample (and subtract from ->count, which is I think the trickiest bit as
> it'll generate a non-monotonic ->count).
> 
> So I prefer the disable/enable from irq_enter/exit, however I also
> suspect that that is by far the most expensive option.


Ingo proposed me another trick while discussing other details: having
a per context count instead of a single whole one.

So instead of having event->count, we can have event->task_count/softirq_count
and hardirq_count. Each time we enter irq_enter() (non-nested), we read the count
register and we compute the difference on irq_exit() and add the result on
event->hardirq_count. (similar kind of tricks for task and softirq counts).

So when we want to get the total, we just need to compute the sum, wrt the 
exclude_* options we have.

Now that still requires to keep the samples proxy. And the samples will stay
a bit async as the interrupt period won't be paused when we enter a filtered
context, something that would only be solved with a round of ->stop(). But
as you said, I really suspect this is not viable.