Messages in this thread | | | Date | Fri, 27 Mar 2015 08:28:43 -0700 | Subject | Re: [PATCH 1/1] perf/x86: filter branches for PEBS event | From | Stephane Eranian <> |
| |
On Fri, Mar 27, 2015 at 5:10 AM, Peter Zijlstra <peterz@infradead.org> wrote: > On Thu, Mar 26, 2015 at 12:40:14PM -0700, Stephane Eranian wrote: >> You are addressing one of the problems of this routine. But I think >> there is a more serious issue which is not addressed here. The >> intel_shared_regs_constraints() assumes that the associated event is >> necessarily unconstrained: >> >> __intel_shared_reg_get_constraints() >> { >> struct event_constraint *c = &emptyconstraint; >> ... >> } > > emptyconstraint != unconstrained. > > Note how that function only returns emptyconstraint if its rejecting the > event, otherwise it returns NULL such that we continue calling > x86_get_event_constraint(). > Ah, yes, I had forgotten about that. Then everything is fine.
>> This is true for offcore_response, but for LBR this may not always be the case. >> I may want to use LBR on the L1D_PEND_MISS event and it would need to >> be on counter 2. > >> But I believe that the current code could place it on counter 0 simply >> because you return if shared_reg_get_constraint() is successful, but >> it looks only at the LBR constraint not the event constraint. I think >> in the presence of LBR, you always need to call share_get_reg() and >> x86_get_event_constraint(). > > Which, I think it does.
Indeed.
| |