Messages in this thread | | | Subject | Re: [RFC][PATCH 07/17] sched: Serialize p->cpus_allowed and ttwu() using p->pi_lock | From | Peter Zijlstra <> | Date | Mon, 03 Jan 2011 12:12:14 +0100 |
| |
On Wed, 2010-12-29 at 22:20 +0800, Yong Zhang wrote: > > - * The caller (fork, wakeup) owns TASK_WAKING, ->cpus_allowed is stable. > > + * The caller (fork, wakeup) owns p->pi_lock, ->cpus_allowed is stable. > > Yes for wakeup, but not true for fork. > I don't see protection in wake_up_new_task(). > Or am I missing something?
Ah, true, wake_up_new_task() holds task_rq_lock() which is sufficient, but yes, we could also make that pi_lock.
| |