sched: Fix schedule_tail() to disable preemption
authorOleg Nesterov <oleg@redhat.com>
Wed, 8 Oct 2014 19:36:44 +0000 (21:36 +0200)
committerIngo Molnar <mingo@kernel.org>
Tue, 28 Oct 2014 09:47:54 +0000 (10:47 +0100)
finish_task_switch() enables preemption, so post_schedule(rq) can be
called on the wrong (and even dead) CPU. Afaics, nothing really bad
can happen, but in this case we can wrongly clear rq->post_schedule
on that CPU. And this simply looks wrong in any case.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Kirill Tkhai <tkhai@yandex.ru>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/20141008193644.GA32055@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
kernel/sched/core.c

index cde848149dd6de87a75f417f45d07996143bf367..b4935600cd85b7aded970f57e5fcf1e09bcc5462 100644 (file)
@@ -2309,15 +2309,14 @@ static inline void post_schedule(struct rq *rq)
 asmlinkage __visible void schedule_tail(struct task_struct *prev)
        __releases(rq->lock)
 {
-       struct rq *rq = this_rq();
+       struct rq *rq;
 
+       /* finish_task_switch() drops rq->lock and enables preemtion */
+       preempt_disable();
+       rq = this_rq();
        finish_task_switch(rq, prev);
-
-       /*
-        * FIXME: do we need to worry about rq being invalidated by the
-        * task_switch?
-        */
        post_schedule(rq);
+       preempt_enable();
 
        if (current->set_child_tid)
                put_user(task_pid_vnr(current), current->set_child_tid);