Re: [PATCH] perf, x86: Make cycles:p working on SNB

[Stephane Eranian]

On Thu, May 24, 2012 at 10:59 AM, Namhyung Kim <namhyung.kim@xxxxxxx> wrote:
> Hi, Stephane
>
> On Thu, 24 May 2012 09:41:45 +0200, Stephane Eranian wrote:
>> On Thu, May 24, 2012 at 9:27 AM, Peter Zijlstra <a.p.zijlstra@xxxxxxxxx> wrote:
>>> On Thu, 2012-05-24 at 12:02 +0900, Namhyung Kim wrote:
>>>
>>>> --- a/arch/x86/kernel/cpu/perf_event_intel.c
>>>> +++ b/arch/x86/kernel/cpu/perf_event_intel.c
>>>> @@ -1329,6 +1329,12 @@ static int intel_pmu_hw_config(struct perf_event *event)
>>>> Â Â Â Â Â Â Â Â*/
>>>> Â Â Â Â Â Â Â u64 alt_config = X86_CONFIG(.event=0xc0, .inv=1, .cmask=16);
>>>>
>>>> + Â Â Â Â Â Â /*
>>>> + Â Â Â Â Â Â Â* SNB introduced INST_RETIRED.PREC_DIST for this purpose.
>>>> + Â Â Â Â Â Â Â*/
>>>> + Â Â Â Â Â Â if (x86_pmu.pebs_constraints == intel_snb_pebs_event_constraints)
>>>> + Â Â Â Â Â Â Â Â Â Â alt_config = X86_CONFIG(.event=0xc0, .umask=0x01,
>>>> + Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â .inv=1, .cmask=16);
>>>>
>>>> Â Â Â Â Â Â Â alt_config |= (event->hw.config & ~X86_RAW_EVENT_MASK);
>>>> Â Â Â Â Â Â Â event->hw.config = alt_config;
>>>
>>> That's rather ugly.. but that's okay, I've actually got the patch for
>>> this still laying around, it needs a bit of an update though.
>>>
>> You cannot simply use PREC_DIST. This umask has some severe
>> restriction. When you measure it, NO other event on the the entire PMU
>> can be measured at the same time. It needs exclusive mode on SNB.
>>
>
> Yeah, I read something like above on the SDM. But just got confused with
> this:
>
> $ ./perf stat -e cycles:p,instructions,cache-references,cache-misses noploop 1
>
Passing :p in counting mode is useless, it does not do anything.
The :p  suffix is only meaningful is sampling mode where it enables PEBS.

> ÂPerformance counter stats for 'noploop 1':
>
>   3,741,658,837 cycles          Â#  Â0.000 GHz
>   3,618,983,116 instructions       Â#  Â0.97 Âinsns per cycle
> Â Â Â Â Â Â51,126 cache-references
>       7,357 cache-misses       Â#  14.390 % of all cache refs
>
> Â Â Â 1.000692634 seconds time elapsed
>
>
> Thanks,
> Namhyung
--
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/