On Thu, Nov 24, 2016 at 04:19:09PM +0000, Mark Rutland wrote:
> > static void
> > event_sched_out(struct perf_event *event,
> > struct perf_cpu_context *cpuctx,
> > - struct perf_event_context *ctx)
> > + struct perf_event_context *ctx,
> > + bool log_overhead)
>
> Boolean parameter are always confusing. Why not pass the flags directly? That
> way we can pass *which* overhead to log, and make the callsites easier to
> understand.
>
> > event->tstamp_stopped = tstamp;
> > - event->pmu->del(event, 0);
> > + event->pmu->del(event, log_overhead ? PERF_EF_LOG : 0);
>
> ... which we could pass on here.
>
> > @@ -1835,20 +1835,21 @@ event_sched_out(struct perf_event *event,
> > static void
> > group_sched_out(struct perf_event *group_event,
> > struct perf_cpu_context *cpuctx,
> > - struct perf_event_context *ctx)
> > + struct perf_event_context *ctx,
> > + bool log_overhead)
>
> Likewise.
>
> > @@ -1872,7 +1873,7 @@ __perf_remove_from_context(struct perf_event *event,
> > {
> > unsigned long flags = (unsigned long)info;
> >
> > - event_sched_out(event, cpuctx, ctx);
> > + event_sched_out(event, cpuctx, ctx, false);
> > if (flags & DETACH_GROUP)
> > perf_group_detach(event);
> > list_del_event(event, ctx);
> > @@ -1918,9 +1919,9 @@ static void __perf_event_disable(struct perf_event
> > *event,
> > update_cgrp_time_from_event(event);
> > update_group_times(event);
> > if (event == event->group_leader)
> > - group_sched_out(event, cpuctx, ctx);
> > + group_sched_out(event, cpuctx, ctx, true);
> > else
> > - event_sched_out(event, cpuctx, ctx);
> > + event_sched_out(event, cpuctx, ctx, true);
>
> Why does this differ from __perf_remove_from_context()?
So I'm not a great fan of sprinkling all this through the core code at
all.
