Re: PEBS bug on HSW: "Unexpected number of pebs records 10" (was: Re:[GIT PULL] perf changes for v3.12)

From: Stephane Eranian
Date: Tue Sep 10 2013 - 12:14:22 EST


On Tue, Sep 10, 2013 at 5:28 PM, Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
> On Tue, Sep 10, 2013 at 07:15:19AM -0700, Stephane Eranian wrote:
>> The threshold is where to generate the interrupt. It does not mean
>> where to stop PEBS recording.
>
> It does, since we don't set a reset value. So once a PEBS assist
> happens, that counter stops until we reprogram it in the PMI.
>
I agree with you.

I am curious to see what's in those entries. If they are all zeroes, then
we know only the pointer is bogus. Let me add some instrumentation
to check that.

>> So it is possible that in HSW, we may
>> get into a situation where it takes time to get to the handler to stop
>> the PMU. I don't know how given we use NMI. Well, unless we were
>> already servicing an NMI at the time. But given that we stop the PMU
>> almost immediately in the handler, I don't see how that would
>> possible. The other oddity in HSW is that we clear the NMI on entry
>> to the handler and not at the end. I never gotten an good explanation
>> as to why that was necessary. So maybe it is related...
>
> Even if the PMI was delayed that counter would never do another entry.
> So I really don't see how we can have more than nr_counters PEBS entries
> in the buffer.

I agree.
--
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/