[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Zw7MpoDcKmX9mGlK@slm.duckdns.org>
Date: Tue, 15 Oct 2024 10:12:22 -1000
From: Tejun Heo <tj@...nel.org>
To: Zhang Qiao <zhangqiao22@...wei.com>
Cc: David Vernet <void@...ifault.com>, Ingo Molnar <mingo@...hat.com>,
Peter Zijlstra <peterz@...radead.org>,
Juri Lelli <juri.lelli@...hat.com>,
Vincent Guittot <vincent.guittot@...aro.org>,
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: Re: [PATCH] sched_ext: Fix missing unlock in consume_dispatch_q()
Hello,
On Tue, Oct 15, 2024 at 10:29:17PM +0800, Zhang Qiao wrote:
> When the function consume_dispatch_q() returns true, the dsq lock may
> remains held and is not unlocked.
>
> Signed-off-by: Zhang Qiao <zhangqiao22@...wei.com>
> ---
> kernel/sched/ext.c | 16 +++++++++-------
> 1 file changed, 9 insertions(+), 7 deletions(-)
>
> diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c
> index 410a4df8a121..4d80aa3de00e 100644
> --- a/kernel/sched/ext.c
> +++ b/kernel/sched/ext.c
> @@ -2377,7 +2377,8 @@ static inline bool consume_remote_task(struct rq *this_rq, struct task_struct *p
> static bool consume_dispatch_q(struct rq *rq, struct scx_dispatch_q *dsq)
> {
> struct task_struct *p;
> -retry:
> + bool ret = false;
> +
> /*
> * The caller can't expect to successfully consume a task if the task's
> * addition to @dsq isn't guaranteed to be visible somehow. Test
> @@ -2394,19 +2395,20 @@ static bool consume_dispatch_q(struct rq *rq, struct scx_dispatch_q *dsq)
> if (rq == task_rq) {
> task_unlink_from_dsq(p, dsq);
> move_local_task_to_local_dsq(p, 0, dsq, rq);
> - raw_spin_unlock(&dsq->lock);
> - return true;
> + ret = true;
> + break;
> }
>
> if (task_can_run_on_remote_rq(p, rq, false)) {
> - if (likely(consume_remote_task(rq, p, dsq, task_rq)))
> - return true;
> - goto retry;
> + if (likely(consume_remote_task(rq, p, dsq, task_rq))) {
> + ret = true;
> + break;
> + }
> }
> }
>
> raw_spin_unlock(&dsq->lock);
> - return false;
> + return ret;
Has this change been tested at all? There's quite a bit of lock dancing
happening in the remote consumption path because the task needs to get
hot-migrated to the local CPU - consume_remote_task() calls
unlink_dsq_and_lock_src_rq() which drops the DSQ lock among other things.
Thanks.
--
tejun
Powered by blists - more mailing lists