[tip: sched/core] sched: Don't account throttle time for empty groups

[tip-bot2 for Josh Don]

The following commit has been merged into the sched/core branch of tip:

Commit-ID:     eed673d4f701e17b8f887dbcbfd3b4bcc53503ec
Gitweb:        https://git.kernel.org/tip/eed673d4f701e17b8f887dbcbfd3b4bcc53503ec
Author:        Josh Don <joshdon@xxxxxxxxxx>
AuthorDate:    Mon, 12 Jun 2023 16:27:47 -07:00
Committer:     Peter Zijlstra <peterz@xxxxxxxxxxxxx>
CommitterDate: Thu, 15 Jun 2023 13:28:19 +02:00

sched: Don't account throttle time for empty groups

It is easy for a cfs_rq to become throttled even when it has no enqueued
entities (for example, if we have just put_prev()'d the last runnable
task of the cfs_rq, and the cfs_rq is out of quota).

Avoid accounting this time towards total throttle time, since it
otherwise falsely inflates the stats.

Note that the dequeue path is special, since we normally disallow
migrations when a task is in a throttled hierarchy (see
throttled_lb_pair()).

Signed-off-by: Josh Don <joshdon@xxxxxxxxxx>
Signed-off-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
Link: https://lkml.kernel.org/r/20230612232748.3948659-1-joshdon@xxxxxxxxxx
---
 kernel/sched/fair.c | 17 ++++++++++++++---
 1 file changed, 14 insertions(+), 3 deletions(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 6189d1a..b56c86c 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -4873,8 +4873,14 @@ enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags)
 
 	if (cfs_rq->nr_running == 1) {
 		check_enqueue_throttle(cfs_rq);
-		if (!throttled_hierarchy(cfs_rq))
+		if (!throttled_hierarchy(cfs_rq)) {
 			list_add_leaf_cfs_rq(cfs_rq);
+		} else {
+#ifdef CONFIG_CFS_BANDWIDTH
+			if (!cfs_rq->throttled_clock)
+				cfs_rq->throttled_clock = rq_clock(rq_of(cfs_rq));
+#endif
+		}
 	}
 }
 
@@ -5480,7 +5486,9 @@ done:
 	 * throttled-list.  rq->lock protects completion.
 	 */
 	cfs_rq->throttled = 1;
-	cfs_rq->throttled_clock = rq_clock(rq);
+	SCHED_WARN_ON(cfs_rq->throttled_clock);
+	if (cfs_rq->nr_running)
+		cfs_rq->throttled_clock = rq_clock(rq);
 	return true;
 }
 
@@ -5498,7 +5506,10 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
 	update_rq_clock(rq);
 
 	raw_spin_lock(&cfs_b->lock);
-	cfs_b->throttled_time += rq_clock(rq) - cfs_rq->throttled_clock;
+	if (cfs_rq->throttled_clock) {
+		cfs_b->throttled_time += rq_clock(rq) - cfs_rq->throttled_clock;
+		cfs_rq->throttled_clock = 0;
+	}
 	list_del_rcu(&cfs_rq->throttled_list);
 	raw_spin_unlock(&cfs_b->lock);