On Fri, Apr 13, 2018 at 12:04:26PM +0100, Patrick Bellasi wrote: > On 13-Apr 12:22, Peter Zijlstra wrote: > > On Fri, Apr 13, 2018 at 10:26:48AM +0200, Peter Zijlstra wrote: > > > On Mon, Apr 09, 2018 at 05:56:09PM +0100, Patrick Bellasi wrote: > > > > +static inline void uclamp_cpu_get(struct task_struct *p, int cpu, int > > > > clamp_id) > > > > +{ > > > > + struct uclamp_cpu *uc_cpu = &cpu_rq(cpu)->uclamp[clamp_id]; > > > > > > > +static inline void uclamp_cpu_put(struct task_struct *p, int cpu, int > > > > clamp_id) > > > > +{ > > > > + struct uclamp_cpu *uc_cpu = &cpu_rq(cpu)->uclamp[clamp_id]; > > > > > > That all seems daft, since you already have rq at the call site. > > > > > > > +static inline void uclamp_task_update(struct rq *rq, struct > > > > task_struct *p) > > > > +{ > > > > + int cpu = cpu_of(rq); > > > > + int clamp_id; > > > > + > > > > + /* The idle task does not affect CPU's clamps */ > > > > + if (unlikely(p->sched_class == &idle_sched_class)) > > > > + return; > > > > + /* DEADLINE tasks do not affect CPU's clamps */ > > > > + if (unlikely(p->sched_class == &dl_sched_class)) > > > > + return; > > > > > > This is wrong; it misses the stop_sched_class. > > > > > > And since you're looking at sched_class anyway, maybe put a marker in > > > there: > > > > > > if (!p->sched_class->has_clamping) > > > return; > > > > Alternatively, we could simply add uclamp_task_{en,de}queue() into > > {en,de}queue_task_{fair,rt}(). > > I like better your first proposal, I think it makes sense to factor > out in core code used by both RT and FAIR the same way. > > Do you have a strong preference?
The second is probably faster as it avoids the load+branch; then again, without LTO you'll get an actual call in return. Dunno...