Re: Drop WARN on AMD lack of perfctrs

From: Peter Zijlstra
Date: Fri May 17 2013 - 05:06:51 EST

On Thu, May 16, 2013 at 11:34:20PM +0200, Borislav Petkov wrote:
> On Thu, May 16, 2013 at 10:55:58PM +0200, Robert Richter wrote:
> > iirc the reason was the different msr range that is switched on fam15h
> > with a different counter to counter msr offset of 2 instead of 1. The
> > code relies on the assumption that the msrs exist on that cpu. Thus
> > the warning if not. Also note that code may have changed in 3.10 in
> > that area.
> Stupid question: why is check_hw_exists() *after* the vendor-specific
> counter detection code in init_hw_perf_events() even though it is
> supposed to check whether hardware is emulated?

Mostly so that check_hw_exists() doesn't need to know about the vendor
specifics like where the MSRs live, how many there are etc..

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at