Re: [PATCH 0/5] perf events finer grained context instrumentation / context exclusion

[Frederic Weisbecker <fweisbec@gmail.com>]

On Thu, Jun 10, 2010 at 08:26:18AM +0200, Ingo Molnar wrote:
> 
> * Frederic Weisbecker <fweisbec@gmail.com> wrote:
> 
> > Here is the new version of per context exclusion, based on hooks on 
> > irq_enter/irq_exit. I haven't observed slowdowns but I haven't actually 
> > measured the impact.
> 
> One thing that would be nice to see in this discussion is a comparison of 
> before/after perf stat --repeat runs.
> 
> Something like:
> 
>   perf stat --repeat ./hackbench 5
> 
> Done with full stat, and then also done with hardirqs/softirqs excluded. (i.e. 
> task context stats only)



Right, so I just tried each perf stat default events with :t and it hung up ;-)

(Not severely, I can kill perf stat with ^Z, but still there is something I need
to fix).



> 
> I.e. does the feature really give us the expected statistical stability in 
> results? Does it really exclude hardirq/softirq workloads, etc.?


But yeah, before posting these patches I gave that a try with the instruction
counter and it didn't change much against the usual results, it's about the same
variations.

I just know the exclusion works by using perf record -g / perf report, as the
callchains are truly reliable against the exclusion rules, and since
counting and samples are treated the same in this scheme (we just deactivate
/reactivate, there is no post blocking or fixup).

I just don't know where the entropy comes from. May be once I'll have the hang
fixed I'll be able to test with all the default stat events and see a better
overview of progress.