perf: Fix running time accounting

A recent fix to the shadow timestamp inadvertly broke the running time
accounting.

We must not update the running timestamp if we fail to schedule the
event, the event will not have ran. This can (and did) result in
negative total runtime because the stopped timestamp was before the
running timestamp (we 'started' but never stopped the event -- because
it never really started we didn't have to stop it either).

Reported-and-Tested-by: Vince Weaver <vincent.weaver@maine.edu>
Fixes: 72f669c008 ("perf: Update shadow timestamp before add event")
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: stable@vger.kernel.org # 4.1
Cc: Shaohua Li <shli@fb.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
This commit is contained in:
Peter Zijlstra 2015-07-27 10:35:07 +02:00 committed by Thomas Gleixner
parent cbfe8fa6cd
commit 00a2916f7f

View File

@ -1868,8 +1868,6 @@ event_sched_in(struct perf_event *event,
perf_pmu_disable(event->pmu); perf_pmu_disable(event->pmu);
event->tstamp_running += tstamp - event->tstamp_stopped;
perf_set_shadow_time(event, ctx, tstamp); perf_set_shadow_time(event, ctx, tstamp);
perf_log_itrace_start(event); perf_log_itrace_start(event);
@ -1881,6 +1879,8 @@ event_sched_in(struct perf_event *event,
goto out; goto out;
} }
event->tstamp_running += tstamp - event->tstamp_stopped;
if (!is_software_event(event)) if (!is_software_event(event))
cpuctx->active_oncpu++; cpuctx->active_oncpu++;
if (!ctx->nr_active++) if (!ctx->nr_active++)