Re: [PATCH RFC] sched/fair: fix sudden expiration of cfq quota in put_prev_task()

[Peter Zijlstra]

On Fri, Apr 03, 2015 at 03:51:17PM +0300, Konstantin Khlebnikov wrote:
> On 03.04.2015 15:41, Konstantin Khlebnikov wrote:
> >Pick_next_task_fair() must be sure that here is at least one runnable
> >task before calling put_prev_task(), but put_prev_task() can expire
> >last remains of cfs quota and throttle all currently runnable tasks.
> >As a result pick_next_task_fair() cannot find next task and crashes.
> 
> Kernel crash looks like this:
> 
> <1>[50288.719491] BUG: unable to handle kernel NULL pointer dereference at 0000000000000038
> <1>[50288.719538] IP: [<ffffffff81097b8c>] set_next_entity+0x1c/0x80

> <4>[50288.720388] Call Trace:
> <4>[50288.720402]  [<ffffffff8109ead8>] pick_next_task_fair+0x88/0x5d0
> <4>[50288.720479]  [<ffffffff816b5b99>] __schedule+0xf9/0x7d0

Which set_next_entity() is that? There are 3 in pick_next_task_fair().

I have a vague suspicion its in the 'simple' code, please verify.

The thinking is that if it was the 'complex' pick_next_entity()
returning NULL we'd have exploded elsewhere, the cfs_rq iteration
would've wandered off into random memory and most likely exploded on
cfs_rq->curr.

Which too would suggest the check_cfs_rq_runtime() thing works just
fine, it send us to the simple code.

> >This patch leaves 1 in ->runtime_remaining when current assignation
> >expires and tries to refill it right after that. In the worst case
> >task will be scheduled once and throttled at the end of slice.

Which is a strange approach. If pick_next_task_fair() is borken, we
should fix that, no?

In any case, it appears to me that: 606dba2e2894 ("sched: Push
put_prev_task() into pick_next_task()") inverted the ->nr_running and
put_prev_task() statements.

If the above set_next_entity() is indeed the simple one, does the below
cure things?

---
 kernel/sched/fair.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index fdae26eb7218..df72d61138a8 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -5176,12 +5176,11 @@ pick_next_task_fair(struct rq *rq, struct task_struct *prev)
 simple:
 	cfs_rq = &rq->cfs;
 #endif
+	put_prev_task(rq, prev);
 
 	if (!cfs_rq->nr_running)
 		goto idle;
 
-	put_prev_task(rq, prev);
-
 	do {
 		se = pick_next_entity(cfs_rq, NULL);
 		set_next_entity(cfs_rq, se);
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/