sched/deadline: Fix stale throttling on de-/boosted tasks
authorLucas Stach <l.stach@pengutronix.de>
Mon, 31 Aug 2020 11:07:19 +0000 (13:07 +0200)
committerPeter Zijlstra <peterz@infradead.org>
Fri, 25 Sep 2020 12:23:24 +0000 (14:23 +0200)
commit46fcc4b00c3cca8adb9b7c9afdd499f64e427135
tree7bf0a9f37a399724d5042a823edadfe095bf8102
parent8e0e0eda6a13e242239799263cc354796c05434a
sched/deadline: Fix stale throttling on de-/boosted tasks

When a boosted task gets throttled, what normally happens is that it's
immediately enqueued again with ENQUEUE_REPLENISH, which replenishes the
runtime and clears the dl_throttled flag. There is a special case however:
if the throttling happened on sched-out and the task has been deboosted in
the meantime, the replenish is skipped as the task will return to its
normal scheduling class. This leaves the task with the dl_throttled flag
set.

Now if the task gets boosted up to the deadline scheduling class again
while it is sleeping, it's still in the throttled state. The normal wakeup
however will enqueue the task with ENQUEUE_REPLENISH not set, so we don't
actually place it on the rq. Thus we end up with a task that is runnable,
but not actually on the rq and neither a immediate replenishment happens,
nor is the replenishment timer set up, so the task is stuck in
forever-throttled limbo.

Clear the dl_throttled flag before dropping back to the normal scheduling
class to fix this issue.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Juri Lelli <juri.lelli@redhat.com>
Link: https://lkml.kernel.org/r/20200831110719.2126930-1-l.stach@pengutronix.de
kernel/sched/deadline.c