Re: [PATCH v2 3/3] sched/tp: Add new tracepoint to track compute energy computation

[Lukasz Luba]

On 2/21/23 12:08, Qais Yousef wrote:
> On 02/20/23 11:22, Lukasz Luba wrote:
> > Hi Qais,
> >
> > On 2/5/23 22:43, Qais Yousef wrote:
> > > It was useful to track feec() placement decision and debug the spare
> > > capacity and optimization issues vs uclamp_max.
> > >
> > > Signed-off-by: Qais Yousef <qyousef@xxxxxxxxxxx>
> > > ---
> > >    include/trace/events/sched.h | 4 ++++
> > >    kernel/sched/core.c          | 1 +
> > >    kernel/sched/fair.c          | 7 ++++++-
> > >    3 files changed, 11 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/include/trace/events/sched.h b/include/trace/events/sched.h
> > > index fbb99a61f714..20cc884f72ff 100644
> > > --- a/include/trace/events/sched.h
> > > +++ b/include/trace/events/sched.h
> > > @@ -735,6 +735,10 @@ DECLARE_TRACE(sched_update_nr_running_tp,
> > >    	TP_PROTO(struct rq *rq, int change),
> > >    	TP_ARGS(rq, change));
> > > +DECLARE_TRACE(sched_compute_energy_tp,
> > > +	TP_PROTO(struct task_struct *p, int dst_cpu, unsigned long energy),
> > > +	TP_ARGS(p, dst_cpu, energy));
> > > +
> > >    #endif /* _TRACE_SCHED_H */
> > >    /* This part must be outside protection */
> > > diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> > > index 4580fe3e1d0c..5f6dde9b892b 100644
> > > --- a/kernel/sched/core.c
> > > +++ b/kernel/sched/core.c
> > > @@ -110,6 +110,7 @@ EXPORT_TRACEPOINT_SYMBOL_GPL(sched_overutilized_tp);
> > >    EXPORT_TRACEPOINT_SYMBOL_GPL(sched_util_est_cfs_tp);
> > >    EXPORT_TRACEPOINT_SYMBOL_GPL(sched_util_est_se_tp);
> > >    EXPORT_TRACEPOINT_SYMBOL_GPL(sched_update_nr_running_tp);
> > > +EXPORT_TRACEPOINT_SYMBOL_GPL(sched_compute_energy_tp);
> > >    DEFINE_PER_CPU_SHARED_ALIGNED(struct rq, runqueues);
> > > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> > > index a8c3d92ff3f6..801e903c4307 100644
> > > --- a/kernel/sched/fair.c
> > > +++ b/kernel/sched/fair.c
> > > @@ -7295,11 +7295,16 @@ compute_energy(struct energy_env *eenv, struct perf_domain *pd,
> > >    {
> > >    	unsigned long max_util = eenv_pd_max_util(eenv, pd_cpus, p, dst_cpu);
> > >    	unsigned long busy_time = eenv->pd_busy_time;
> > > +	unsigned long energy;
> > >    	if (dst_cpu >= 0)
> > >    		busy_time = min(eenv->pd_cap, busy_time + eenv->task_busy_time);
> > > -	return em_cpu_energy(pd->em_pd, max_util, busy_time, eenv->cpu_cap);
> > > +	energy = em_cpu_energy(pd->em_pd, max_util, busy_time, eenv->cpu_cap);
> > > +
> > > +	trace_sched_compute_energy_tp(p, dst_cpu, energy);
> >
> > Could we also dump the max_util and busy_time?
>
> We certainly can :)

Great, thanks!