Skip to contentSkip to conversation

sched/fair: Optimize select_idle_cpu

PATCH New Yang Wei 1 participant 1 message Jun 7, 2021 Stable branch

lore
Share & tools
lore

Yang Wei

author
#0 lore raw

PATCH 1/1 sched/fair: Optimize select_idle_cpu

From: Cheng Jian <[email protected]>

select_idle_cpu() will scan the LLC domain for idle CPUs, it's always expensive. so the next commit :

1ad3aaf3fcd2 ("sched/core: Implement new approach to scale select_idle_cpu()")

introduces a way to limit how many CPUs we scan.

But it consume some CPUs out of 'nr' that are not allowed for the task and thus waste our attempts. The function always return nr_cpumask_bits, and we can't find a CPU which our task is allowed to run.

Cpumask may be too big, similar to select_idle_core(), use per_cpu_ptr 'select_idle_mask' to prevent stack overflow.

---
 kernel/sched/fair.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
kernel/sched/fair.c +4 -3
index 81096dd..37ac76d 100644
@@ -5779,6 +5779,7 @@ static inline int select_idle_smt(struct task_struct *p, struct sched_domain *sd
  */
 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);
 	struct sched_domain *this_sd;
 	u64 avg_cost, avg_idle;
 	u64 time, cost;
@@ -5809,11 +5810,11 @@ static int select_idle_cpu(struct task_struct *p, struct sched_domain *sd, int t
 
 	time = local_clock();
 
-	for_each_cpu_wrap(cpu, sched_domain_span(sd), target) {
+	cpumask_and(cpus, sched_domain_span(sd), &p->cpus_allowed);
+
+	for_each_cpu_wrap(cpu, cpus, target) {
 		if (!--nr)
 			return -1;
-		if (!cpumask_test_cpu(cpu, &p->cpus_allowed))
-			continue;
 		if (idle_cpu(cpu))
 			break;
 	}
Signature · 1 line
1.8.3.1