Re: [PATCH 3/4] perf, x86: Use ALTERNATIVE() to check for X86_FEATURE_PERFCTR_CORE

[Andi Kleen]

On Tue, Apr 19, 2011 at 12:39:27PM +0200, Robert Richter wrote:
> On 18.04.11 16:00:57, Andi Kleen wrote:
> > Robert Richter <robert.richter@xxxxxxx> writes:
> > 
> > > Using ALTERNATIVE() when checking for X86_FEATURE_PERFCTR_CORE avoids
> > > an extra pointer chase and data cache hit.
> > 
> > Is that really a performance critical path?
> > 
> > Seems more like unnecessary obfuscation to me.
> 
> We hotest path is in perf_pmu_disable(), which happens at least with
> every task switch when calling the event scheduler.

Yes but that's already a slow path isn't it? It better is, because
the MSR accesses alone are incredibly expensive. I guess your test
and jump isn't even on the radar after that ...

-Andi
--
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/