Re: [PATCH 10/10] perf tests: Use lower sample_freq in sw clockevent period test

[Arnaldo Carvalho de Melo]

Em Tue, Nov 12, 2013 at 05:40:39PM +0900, Namhyung Kim escreveu:
> Hi Adrian,
> 
> On Tue, 12 Nov 2013 09:07:36 +0200, Adrian Hunter wrote:
> > On 11/11/13 22:22, Arnaldo Carvalho de Melo wrote:
> >> From: Arnaldo Carvalho de Melo <acme@xxxxxxxxxx>
> >>
> >> We were using it at 10 kHz, which doesn't work in machines where somehow
> >> the max freq was auto reduced by the kernel:
> >>
> >> [root@ssdandy ~]# perf test 19
> >> 19: Test software clock events have valid period values    : FAILED!
> >> [root@ssdandy ~]# perf test -v 19
> >> 19: Test software clock events have valid period values    :
> >> --- start ---
> >> Couldn't open evlist: Invalid argument
> >> ---- end ----
> >> Test software clock events have valid period values: FAILED!
> >> [root@ssdandy ~]#
> >>
> >> [root@ssdandy ~]# cat /proc/sys/kernel/perf_event_max_sample_rate
> >> 7000
> >>
> >> Reducing it to 500 Hz should be good enough for this test and also
> >> shouldn't affect what it is testing.
> >>
> >> But warn the user if it fails, informing the knob and the freq tried.
> >
> > Doesn't work for me:
> >
> > ./perf test -v 19
> > 19: Test software clock events have valid period values    :
> > --- start ---
> > mmap size 528384B
> > mmap size 528384B
> > All (0) samples have period value of 1!
> > ---- end ----
> > Test software clock events have valid period values: FAILED!
> >
> > But this fixes it:
> >
> > diff --git a/tools/perf/tests/sw-clock.c b/tools/perf/tests/sw-clock.c
> > index 93a7139..6664a7c 100644
> > --- a/tools/perf/tests/sw-clock.c
> > +++ b/tools/perf/tests/sw-clock.c
> > @@ -9,7 +9,7 @@
> >  #include "util/cpumap.h"
> >  #include "util/thread_map.h"
> >  
> > -#define NR_LOOPS  1000000
> > +#define NR_LOOPS  10000000
> 
> Or else, why not using the max_sample_rate as the freq value?

Sure, that probably is better, I was just trying the lazy way, as doing
what you suggest entails reworking what Jiri did, since the routine that
reads that max_sample_rate is not exported, needs some massaging, etc.

Using a low enough max_sample_rate and warning the user that anyway
would be better knowing that the sample rate lowered so dramatically
looked quicker/lazy enough ;-)

- Arnaldo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/