Re: [PATCH 1/2] perf-events: Add support for supplementary eventregisters v2

[Stephane Eranian]

On Fri, Nov 12, 2010 at 6:33 PM, Peter Zijlstra <a.p.zijlstra@xxxxxxxxx> wrote:
> On Fri, 2010-11-12 at 18:17 +0100, Stephane Eranian wrote:
>> I looked at this patch thinking how could this be reused for LBR_SELECT.
>>
>> I am wondering if storing the extra MSR value in attr.config is really the way
>> to go now as opposed to adding/overloading a field.
>>
>> For OFFCORE_RESPONSE, it makes sense to use attr.config because this is
>> a refinement of the event (a sub-sub event if you want).
>
> Correct, it makes sense for offcore and load-latency, not so much for
> lbr_config.
>
Good.

>> For LBR_SELECT, you also need to pass a value, but there is no specific event
>> associated with it. You can enable LBR with any event you want. Thus,
>> storing the
>> LBR_SELECT value independently would also make sense. And if we have this
>> field for LBR_SELECT then we may as well use it for OFFCORE_REPONSE.
>
> That would assume a single event doesn't contain offcore and lbr, no?
> Currently the extra_reg thing assumes there's only one extra reg encoded
> in the config word.
>
Good point. You want to allow simultaneous LBR + OFFCORE, i.e., to figure
out the path that led to the event (with some skid). So those need to be in
separate fields. I would go with offcore in attr.config, lbr somewhere else.
If you do this then you have the issue of how generic in the content of that
extra field.


>> The alternative would be to consider LBR_SELECT also as a refinement of
>> the event being measured. Though by itself, it wouldn't do anything, it would
>> have to be combine with a PERF_SAMPLE_BRANCH_STACK to attr.sample_type.
>
> A separate field for the lbr_config seems to make most sense, we could
> of course use the top 16 bits for lbr, the next 16 for offcore/ll and
> the bottom 32 for eventsel, but that's mighty crowded.
>
Agreed, not such a good idea. I like the separate lbr_config better.
Could be made
more generic, e.g., branch_stack_config. This is how you've named LBR in the
generic code currently.
--
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/