Messages in this thread | | | Subject | Re: [PATCH 2/5] perf: generic intel uncore support | From | Peter Zijlstra <> | Date | Tue, 03 Apr 2012 00:10:57 +0200 |
| |
On Sun, 2012-04-01 at 11:11 +0800, Yan, Zheng wrote: > >> +static struct intel_uncore_box *uncore_event_to_box(struct perf_event *event) > >> +{ > >> + int phyid = topology_physical_package_id(smp_processor_id()); > > > > Who says that this event has anything to do with the current cpu? > > > Because perf core schedules event on the basis of cpu.
This appears true for all the current callchains, however there's no comment in there making clear you thought about this, nor to remind you future users respect this constraint.
| |