Messages in this thread | | | Date | Mon, 7 Nov 2011 19:17:02 +0200 | From | Gleb Natapov <> | Subject | Re: [PATCHv2 2/9] KVM: Expose a version 2 architectural PMU to a guests |
| |
On Mon, Nov 07, 2011 at 05:45:14PM +0100, Peter Zijlstra wrote: > On Mon, 2011-11-07 at 17:25 +0200, Gleb Natapov wrote: > > > Since the below programming doesn't use perf_event_attr::pinned, yes. > > > > > Yes, that is on todo :). Actually I do want to place all guest perf > > counters into the same event group and make it pinned. But currently perf > > event groups are not very flexible. In our usage scenario we can't have > > one event as a group leader since events are created and destroyed very > > dynamically. What I would like is to have something like meta event that > > will group all other real event. > > Is there a reason to have them grouped if you pin them all anyway? Hmm good question. May be we shouldn't pin then since this will prevent profiling vcpu task on a host while perf is running in a guest, but then we need to group guest created event to get meaningful result. I think I'll pin them for now and will look into grouping them later.
-- Gleb.
| |