| From | Valentin Schneider <> | Subject | Re: [PATCH v4 19/19] sched: Comment affine_move_task() | Date | Thu, 29 Oct 2020 16:27:45 +0000 |
| |
On 23/10/20 11:12, Peter Zijlstra wrote: > + * (1) In the cases covered above. There is one more where the completion is > + * signaled within affine_move_task() itself: when a subsequent affinity request > + * cancels the need for an active migration. Consider: > + * > + * Initial conditions: P0->cpus_mask = [0, 1] > + * > + * P0@CPU0 P1 P2 > + * > + * migrate_disable(); > + * <preempted> > + * set_cpus_allowed_ptr(P0, [1]); > + * <blocks> > + * set_cpus_allowed_ptr(P0, [0, 1]); > + * <signal completion> > + * <awakes> > + * > + * Note that the above is safe vs a concurrent migrate_enable(), as any > + * pending affinity completion is preceded an uninstallion of > + * p->migration_pending done with p->pi_lock held.
I too must have been thinking too much about ponies lately.
diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 0a3f9fd3b061..d8c85f180b09 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -2147,7 +2147,7 @@ void do_set_cpus_allowed(struct task_struct *p, const struct cpumask *new_mask) * <awakes> * * Note that the above is safe vs a concurrent migrate_enable(), as any - * pending affinity completion is preceded an uninstallion of + * pending affinity completion is preceded by an uninstallation of * p->migration_pending done with p->pi_lock held. */ static int affine_move_task(struct rq *rq, struct task_struct *p, struct rq_flags *rf,
|