On Wed, Feb 06, 2019 at 05:25:31PM +0000, Valentin Schneider wrote:
> Hi,
>
> On 06/02/2019 16:04, Peter Zijlstra wrote:
> [...]
> >> @@ -9561,6 +9573,14 @@ static void nohz_balancer_kick(struct rq *rq)
> >
> > sd = rcu_dereference(rq->sd);
> > if (sd) {
> > if ((rq->cfs.h_nr_running >= 1) &&
> > check_cpu_capacity(rq, sd)) {
> > flags = NOHZ_KICK_MASK;
> > goto unlock;
> >> }
> >> }
> >>
> >> + sd = rcu_dereference(per_cpu(sd_asym_cpucapacity, cpu));
> >> + if (sd) {
> >> + if (check_misfit_status(rq, sd)) {
> >> + flags = NOHZ_KICK_MASK;
> >> + goto unlock;
> >> + }
> >> + }
> >
> > So while the exact @sd to use for check_cpu_capacity() likely doesn't
> > matter; this is a 'implicit' test for actually having asym_capacity.
> >
>
> I did feel compelled to use the "right" @sd to have a coherent
> imbalance_pct being used - on big.LITTLE, the @sd above this hunk would be
> MC (117 imbalance_pct) whereas the sd_asym_cpucapacity one would be
> DIE (125 imbalance_pct).
Fair enough.
> > @@ -9606,6 +9614,10 @@ static void nohz_balancer_kick(struct rq
> >
> > sd = rcu_dereference(per_cpu(sd_asym_packing, cpu));
> > if (sd) {
> > + /*
> > + * When ASYM_PACKING; see if there's a more preferred CPU going
> > + * idle; in which case, kick the ILB to move tasks around.
> > + */
>
> s/going idle/currently idle/, no?
Yah.. D'0h..
> > for_each_cpu_and(i, sched_domain_span(sd), nohz.idle_cpus_mask)
> > {
> > if (sched_asym_prefer(i, cpu)) {
> > flags = NOHZ_KICK_MASK;
> >
