Re: [PATCH 4/4] sched/fair: Merge select_idle_core/cpu()

[Mel Gorman]

On Wed, Jan 27, 2021 at 11:43:22AM +0100, Vincent Guittot wrote:
> > @@ -6149,18 +6161,31 @@ static int select_idle_cpu(struct task_struct *p, struct sched_domain *sd, int t
> >         }
> >
> >         for_each_cpu_wrap(cpu, cpus, target) {
> > -               if (!--nr)
> > -                       return -1;
> > -               if (available_idle_cpu(cpu) || sched_idle_cpu(cpu))
> > -                       break;
> > +               if (smt) {
> > +                       i = select_idle_core(p, cpu, cpus, &idle_cpu);
> > +                       if ((unsigned int)i < nr_cpumask_bits)
> > +                               return i;
> > +
> > +               } else {
> > +                       if (!--nr)
> > +                               return -1;
> > +                       i = __select_idle_cpu(cpu);
> 
> you should use idle_cpu directly instead of this intermediate i variable
> 
> +                       idle_cpu = __select_idle_cpu(cpu);
> +                       if ((unsigned int)idle_cpu < nr_cpumask_bits)
> +                               break;
> 
> Apart ths small comment above, the patch looks good to me and I
> haven't any performance regression anymore
> 

It's matching the code sequence in the SMT block. If we are going to make
that change, then go the full way with this?

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 52a650aa2108..01e40e36c386 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -6129,7 +6129,7 @@ static inline int select_idle_core(struct task_struct *p, int core, struct cpuma
 static int select_idle_cpu(struct task_struct *p, struct sched_domain *sd, int target)
 {
 	struct cpumask *cpus = this_cpu_cpumask_var_ptr(select_idle_mask);
-	int i, cpu, idle_cpu = -1, nr = INT_MAX;
+	int cpu, idle_cpu = -1, nr = INT_MAX;
 	bool smt = test_idle_cores(target, false);
 	int this = smp_processor_id();
 	struct sched_domain *this_sd;
@@ -6162,18 +6162,16 @@ static int select_idle_cpu(struct task_struct *p, struct sched_domain *sd, int t
 
 	for_each_cpu_wrap(cpu, cpus, target) {
 		if (smt) {
-			i = select_idle_core(p, cpu, cpus, &idle_cpu);
-			if ((unsigned int)i < nr_cpumask_bits)
-				return i;
+			idle_cpu = select_idle_core(p, cpu, cpus, &idle_cpu);
+			if ((unsigned int)idle_cpu < nr_cpumask_bits)
+				return idle_cpu;
 
 		} else {
 			if (!--nr)
 				return -1;
-			i = __select_idle_cpu(cpu);
-			if ((unsigned int)i < nr_cpumask_bits) {
-				idle_cpu = i;
+			idle_cpu = __select_idle_cpu(cpu);
+			if ((unsigned int)idle_cpu < nr_cpumask_bits)
 				break;
-			}
 		}
 	}
 

-- 
Mel Gorman
SUSE Labs