Re: [PATCH] perf, x86: Add Silvermont (22nm Atom) support

[Peter Zijlstra <peterz@infradead.org>]

On Thu, Jul 18, 2013 at 04:27:31PM +0800, Yan, Zheng wrote:
> On 07/18/2013 04:23 PM, Peter Zijlstra wrote:
> > On Thu, Jul 18, 2013 at 01:36:07PM +0800, Yan, Zheng wrote:
> >> +static struct event_constraint intel_slm_event_constraints[] __read_mostly =
> >> +{
> >> +	FIXED_EVENT_CONSTRAINT(0x00c0, 0), /* INST_RETIRED.ANY */
> >> +	FIXED_EVENT_CONSTRAINT(0x003c, 1), /* CPU_CLK_UNHALTED.CORE */
> >> +	FIXED_EVENT_CONSTRAINT(0x013c, 2), /* CPU_CLK_UNHALTED.REF */
> >> +	FIXED_EVENT_CONSTRAINT(0x0300, 2), /* pseudo CPU_CLK_UNHALTED.REF */
> > 
> > So the normal event 0x13c and the fixed counter 2 are normally _not_ the
> > same. Are they for slm? Are you sure?
> > 
> 
> yes, I'm sure. see page 15-15 of http://www.intel.com/content/dam/www/public/us/en/documents/manuals/64-ia-32-architectures-optimization-manual.pdf

OK, then put in a comment how slm is 'special' and you might want to fix
intel_pmu_init():

	if (x86_pmu.event_constraints) {
		/*
		 * event on fixed counter2 (REF_CYCLES) only works on this
		 * counter, so do not extend mask to generic counters
		 */
		for_each_event_constraint(c, x86_pmu.event_constraints) {
			if (c->cmask != FIXED_EVENT_FLAGS
			    || c->idxmsk64 == INTEL_PMC_MSK_FIXED_REF_CYCLES) {
				continue;
			}

			c->idxmsk64 |= (1ULL << x86_pmu.num_counters) - 1;
			c->weight += x86_pmu.num_counters;
		}
	}

Since that explicitly skips the fixed counter 2 and doesn't extend its
constraint to include all other counters.