Messages in this thread | | | Subject | Re: [PATCH V4 07/14] perf/x86/intel: Support hardware TopDown metrics | From | "Liang, Kan" <> | Date | Mon, 30 Sep 2019 12:17:57 -0400 |
| |
On 9/30/2019 10:53 AM, Peter Zijlstra wrote: > > After that, I think we can simply do something like: > > icl_update_topdown_event(..)
We should call this function in x86_pmu_commit_txn()? In intel_pmu_read_event(), we just simply return, when TXN_READ is set and is_topdown_count().
If so, it looks like we need add a per-cpu variable to store the topdown group events which user tries to read. Then we can update these events in x86_pmu_commit_txn(). The same method as Power does.
Is my understanding correct?
> { > int idx = event->hwc.idx; > > if (is_metric_idx(idx)) > return; > > // must be FIXED_SLOTS
The FIXED_SLOTS may not be in the group.
Thanks, Kan > > /* do teh thing and update SLOTS and METRIC together */ > } > > Hmmm?
| |