Re: [PATCH V2 1/2] sched/core: Rename and move schedutil_cpu_util() to core.c

[Lukasz Luba]

On 10/23/20 11:54 AM, Viresh Kumar wrote:
> On 23-10-20, 12:34, Peter Zijlstra wrote:
> > On Fri, Oct 23, 2020 at 03:50:20PM +0530, Viresh Kumar wrote:
> > > diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> > > index d2003a7d5ab5..369ff54d11d4 100644
> > > --- a/kernel/sched/core.c
> > > +++ b/kernel/sched/core.c
> > > @@ -5117,6 +5117,119 @@ struct task_struct *idle_task(int cpu)
> > >   	return cpu_rq(cpu)->idle;
> > >   }
> > >   
> > > +/*
> > > + * This function computes an effective utilization for the given CPU, to be
> > > + * used for frequency selection given the linear relation: f = u * f_max.
> > > + *
> > > + * The scheduler tracks the following metrics:
> > > + *
> > > + *   cpu_util_{cfs,rt,dl,irq}()
> > > + *   cpu_bw_dl()
> > > + *
> > > + * Where the cfs,rt and dl util numbers are tracked with the same metric and
> > > + * synchronized windows and are thus directly comparable.
> > > + *
> > > + * The cfs,rt,dl utilization are the running times measured with rq->clock_task
> > > + * which excludes things like IRQ and steal-time. These latter are then accrued
> > > + * in the irq utilization.
> > > + *
> > > + * The DL bandwidth number otoh is not a measured metric but a value computed
> > > + * based on the task model parameters and gives the minimal utilization
> > > + * required to meet deadlines.
> > > + */
> > > +unsigned long effective_cpu_util(int cpu, unsigned long util_cfs,
> > > +				 unsigned long max, enum cpu_util_type type,
> > > +				 struct task_struct *p)
> > > +{
> > 	...
> > > +}
> > > +
> > > +unsigned long sched_cpu_util(int cpu, enum cpu_util_type type,
> > > +			     unsigned long max)
> > > +{
> > > +	return effective_cpu_util(cpu, cpu_util_cfs(cpu_rq(cpu)), max, type,
> > > +				  NULL);
> > > +}
> >
> > Shouldn't all that be: #ifdef CONFIG_SMP ?
>
> I didn't realize that these matrices are only available in case of SMP
> and that's why schedutil isn't available for !SMP. I wonder what we
> should be doing in cpufreq_cooling now ? Make it depend on SMP ? Or
> calculate load the traditional way (the stuff I just removed) for !SMP
> case ?

IMO the !SMP can leave with the old design, so keeping two
implementations under #ifdef CONFIG_SMP is fair I would say in this
case.

There are popular platforms !SMP (BeagleBone, RPi1, RPiZero) but I
haven't heard anyone was using IPA on them.

Regards,
Lukasz

> :)