Re: [PATCH] perf: Fix intel shared extra msr allocation

[Stephane Eranian]

On Tue, Jun 5, 2012 at 12:14 PM, Peter Zijlstra <a.p.zijlstra@xxxxxxxxx> wrote:
> On Fri, 2012-06-01 at 22:11 +0800, Yan, Zheng wrote:
>>
>> The real problem is from __intel_shared_reg_put_constraints(). it set
>> reg->alloc to 0
>
> ok
>
>> and decreases fake_cpuc->shared_regs->regs[reg->idx]'s
>> reference count.
>
> So? That's fake state, who cares what its left in?
>
Right, that's what I am trying to elucidate. Who cares about it
beyond validate_group()? The fake_cpuc is reallocated for each
validate group. And this includes the fake shared_regs.

>> Later when deleting the event, Âput_constraints() will find
>> reg->alloc is 0 and it won't decrease the shared msr's reference count.
>
> OK, so the only problem is us setting reg->alloc to 0?

I agree with the first part of the patch in intel_try_alt_er(), we
should not touch
the actual event struct. But I am still unclear about the reg->alloc part.
--
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/