linux-kernel - [PATCH] sched_ext: Always call put_prev

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-ID: <20241013173928.20738-1-andrea.righi@linux.dev>
Date: Sun, 13 Oct 2024 19:39:28 +0200
From: Andrea Righi <andrea.righi@...ux.dev>
To: Tejun Heo <tj@...nel.org>,
	David Vernet <void@...ifault.com>,
	Peter Zijlstra <peterz@...radead.org>,
	Ingo Molnar <mingo@...hat.com>,
	Juri Lelli <juri.lelli@...hat.com>,
	Vincent Guittot <vincent.guittot@...aro.org>
Cc: Dietmar Eggemann <dietmar.eggemann@....com>,
	Steven Rostedt <rostedt@...dmis.org>,
	Ben Segall <bsegall@...gle.com>,
	Mel Gorman <mgorman@...e.de>,
	Valentin Schneider <vschneid@...hat.com>,
	linux-kernel@...r.kernel.org
Subject: [PATCH] sched_ext: Always call put_prev_task() with scx enabled

With the consolidation of put_prev_task/set_next_task(), we are now
skipping the sched_ext ops.stopping/running() transitions when the
previous and next tasks are the same, see commit 436f3eed5c69 ("sched:
Combine the last put_prev_task() and the first set_next_task()").

While this optimization makes sense in general, it can negatively impact
performance in some user-space schedulers, that expect to handle such
transitions when tasks exhaust their timeslice (see SCX_OPS_ENQ_LAST).

For example, scx_rustland suffers a significant performance regression
(e.g., gaming benchmarks drop from ~60fps to ~10fps).

To fix this, ensure that put_prev_task()/set_next_task() are never
skipped when the scx scheduling class is enabled, allowing the scx class
to handle such transitions.

This change restores the previous behavior, fixing the performance
regression in scx_rustland.

Link: https://github.com/sched-ext/scx/issues/788
Fixes: 7c65ae81ea86 ("sched_ext: Don't call put_prev_task_scx() before picking the next task")
Signed-off-by: Andrea Righi <andrea.righi@...ux.dev>
---
 kernel/sched/sched.h | 30 +++++++++++++++---------------
 1 file changed, 15 insertions(+), 15 deletions(-)

diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index 6085ef50febf..44d736e49d06 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -2470,21 +2470,6 @@ __put_prev_set_next_dl_server(struct rq *rq,
 	rq->dl_server = NULL;
 }
 
-static inline void put_prev_set_next_task(struct rq *rq,
-					  struct task_struct *prev,
-					  struct task_struct *next)
-{
-	WARN_ON_ONCE(rq->curr != prev);
-
-	__put_prev_set_next_dl_server(rq, prev, next);
-
-	if (next == prev)
-		return;
-
-	prev->sched_class->put_prev_task(rq, prev, next);
-	next->sched_class->set_next_task(rq, next, true);
-}
-
 /*
  * Helper to define a sched_class instance; each one is placed in a separate
  * section which is ordered by the linker script:
@@ -2523,6 +2508,21 @@ DECLARE_STATIC_KEY_FALSE(__scx_switched_all);	/* all fair class tasks on SCX */
 #define scx_switched_all()	false
 #endif /* !CONFIG_SCHED_CLASS_EXT */
 
+static inline void put_prev_set_next_task(struct rq *rq,
+					  struct task_struct *prev,
+					  struct task_struct *next)
+{
+	WARN_ON_ONCE(rq->curr != prev);
+
+	__put_prev_set_next_dl_server(rq, prev, next);
+
+	if (next == prev && !scx_enabled())
+		return;
+
+	prev->sched_class->put_prev_task(rq, prev, next);
+	next->sched_class->set_next_task(rq, next, true);
+}
+
 /*
  * Iterate only active classes. SCX can take over all fair tasks or be
  * completely disabled. If the former, skip fair. If the latter, skip SCX.
-- 
2.47.0