sched/fair: Do not announce throttled next buddy in dequeue_task_fair()
authorKonstantin Khlebnikov <khlebnikov@yandex-team.ru>
Thu, 16 Jun 2016 12:57:15 +0000 (15:57 +0300)
committerIngo Molnar <mingo@kernel.org>
Fri, 24 Jun 2016 06:26:45 +0000 (08:26 +0200)
Hierarchy could be already throttled at this point. Throttled next
buddy could trigger a NULL pointer dereference in pick_next_task_fair().

Signed-off-by: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Ben Segall <bsegall@google.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/146608183552.21905.15924473394414832071.stgit@buzz
Signed-off-by: Ingo Molnar <mingo@kernel.org>
kernel/sched/fair.c

index 8c5d8c0c88273d7f5e03f1aeb4fbd3e97a2201f7..bdcbeea90c950523b5276cfd1d8464ceef2d4fd8 100644 (file)
@@ -4537,15 +4537,14 @@ static void dequeue_task_fair(struct rq *rq, struct task_struct *p, int flags)
 
                /* Don't dequeue parent if it has other entities besides us */
                if (cfs_rq->load.weight) {
+                       /* Avoid re-evaluating load for this entity: */
+                       se = parent_entity(se);
                        /*
                         * Bias pick_next to pick a task from this cfs_rq, as
                         * p is sleeping when it is within its sched_slice.
                         */
-                       if (task_sleep && parent_entity(se))
-                               set_next_buddy(parent_entity(se));
-
-                       /* avoid re-evaluating load for this entity */
-                       se = parent_entity(se);
+                       if (task_sleep && se && !throttled_hierarchy(cfs_rq))
+                               set_next_buddy(se);
                        break;
                }
                flags |= DEQUEUE_SLEEP;