Re: [RFC PATCH 5/6] sched/fair: Select an energy-efficient CPU on task wake-up

From: Quentin Perret
Date: Sat Mar 24 2018 - 22:00:07 EST


On Wednesday 21 Mar 2018 at 15:35:18 (+0000), Patrick Bellasi wrote:
> On 20-Mar 09:43, Dietmar Eggemann wrote:
> > From: Quentin Perret <quentin.perret@xxxxxxx>
>
> [...]
>
> > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> > index 76bd46502486..65a1bead0773 100644
> > --- a/kernel/sched/fair.c
> > +++ b/kernel/sched/fair.c
> > @@ -6513,6 +6513,60 @@ static unsigned long compute_energy(struct task_struct *p, int dst_cpu)
> > return energy;
> > }
> >
> > +static bool task_fits(struct task_struct *p, int cpu)
> > +{
> > + unsigned long next_util = cpu_util_next(cpu, p, cpu);
> > +
> > + return util_fits_capacity(next_util, capacity_orig_of(cpu));
> ^^^^^^^^^^^^^^^^^^^^^
>
> Since here we are at scheduling CFS tasks, should we not better use
> capacity_of() to account for RT/IRQ pressure ?

Yes, definitely. I change this in v2.

>
> > +}
> > +
> > +static int find_energy_efficient_cpu(struct sched_domain *sd,
> > + struct task_struct *p, int prev_cpu)
> > +{
> > + unsigned long cur_energy, prev_energy, best_energy;
> > + int cpu, best_cpu = prev_cpu;
> > +
> > + if (!task_util(p))
>
> We are still waking up a task... what if the task was previously
> running on a big CPU which is now idle?
>
> I understand that from a _relative_ energy_diff standpoint there is
> not much to do for a 0 utilization task. However, for those tasks we
> can still try to return the most energy efficient CPU among the ones
> in their cpus_allowed mask.
>
> It should be a relatively low overhead (maybe contained in a fallback
> most_energy_efficient_cpu() kind of function) which allows, for
> example on ARM big.LITTLE systems, to consolidate those tasks on
> LITTLE CPUs instead for example keep running them on a big CPU.

Hmmmm so the difficult thing about a task with 0 util is that you don't
know if this is really a small task, or a big task with a very long
period. The only useful thing you know for sure about the task is where
it ran last time, so I guess that makes sense to use that information
rather than make assumptions. There is no perfect solution using the
util_avg of the task.

Now, UTIL_EST is changing the game here. If we use it for task placement
(which I think is the right thing to do), this issue should be a lot
easier to solve. What do you think ?

Thanks,
Quentin