[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20240712174313.GW27299@noisy.programming.kicks-ass.net>
Date: Fri, 12 Jul 2024 19:43:13 +0200
From: Peter Zijlstra <peterz@...radead.org>
To: Valentin Schneider <vschneid@...hat.com>
Cc: linux-kernel@...r.kernel.org, rcu@...r.kernel.org,
Ingo Molnar <mingo@...hat.com>, 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>,
Phil Auld <pauld@...hat.com>, Clark Williams <williams@...hat.com>,
Tomas Glozar <tglozar@...hat.com>,
"Paul E. McKenney" <paulmck@...nel.org>,
Frederic Weisbecker <frederic@...nel.org>,
Neeraj Upadhyay <neeraj.upadhyay@...nel.org>,
Joel Fernandes <joel@...lfernandes.org>,
Josh Triplett <josh@...htriplett.org>,
Boqun Feng <boqun.feng@...il.com>,
Mathieu Desnoyers <mathieu.desnoyers@...icios.com>,
Lai Jiangshan <jiangshanlai@...il.com>,
Zqiang <qiang.zhang1211@...il.com>,
Alexander Gordeev <agordeev@...ux.ibm.com>,
Catalin Marinas <catalin.marinas@....com>,
Arnd Bergmann <arnd@...db.de>, Guo Ren <guoren@...nel.org>,
Palmer Dabbelt <palmer@...osinc.com>,
Andrew Morton <akpm@...ux-foundation.org>,
Oleg Nesterov <oleg@...hat.com>, Jens Axboe <axboe@...nel.dk>
Subject: Re: [RFC PATCH v3 10/10] sched/fair: Throttle CFS tasks on return to
userspace
On Thu, Jul 11, 2024 at 03:00:04PM +0200, Valentin Schneider wrote:
> +static void throttle_one_task(struct cfs_rq *cfs_rq, struct task_struct *p)
> {
> + long task_delta, idle_task_delta;
> + struct sched_entity *se = &p->se;
> +
> + list_add(&p->throttle_node, &cfs_rq->throttled_limbo_list);
>
> + task_delta = 1;
> + idle_task_delta = cfs_rq_is_idle(cfs_rq) ? 1 : 0;
> +
> + for_each_sched_entity(se) {
> + cfs_rq = cfs_rq_of(se);
> +
> + if (!se->on_rq)
> + return;
> +
> + dequeue_entity(cfs_rq, se, DEQUEUE_SLEEP);
> + cfs_rq->h_nr_running -= task_delta;
> + cfs_rq->idle_h_nr_running -= idle_task_delta;
> +
> + if (cfs_rq->load.weight) {
> + /* Avoid re-evaluating load for this entity: */
> + se = parent_entity(se);
> + break;
> + }
> + }
> +
> + for_each_sched_entity(se) {
> + cfs_rq = cfs_rq_of(se);
> + /* throttled entity or throttle-on-deactivate */
> + if (!se->on_rq)
> + goto throttle_done;
> +
> + update_load_avg(cfs_rq, se, 0);
> + se_update_runnable(se);
> + cfs_rq->h_nr_running -= task_delta;
> + cfs_rq->h_nr_running -= idle_task_delta;
> + }
> +
> +throttle_done:
> + /* At this point se is NULL and we are at root level*/
> + sub_nr_running(rq_of(cfs_rq), 1);
> }
I know you're just moving code around, but we should look if we can
share code between this and dequeue_task_fair().
I have patches around this in that eevdf series I should send out again,
I'll try and have a stab.
> -static void task_throttle_cancel_irq_work_fn(struct irq_work *work)
> +static void throttle_cfs_rq_work(struct callback_head *work)
> {
> - /* Write me */
> + struct task_struct *p = container_of(work, struct task_struct, sched_throttle_work);
> + struct sched_entity *se;
> + struct rq *rq;
> + struct cfs_rq *cfs_rq;
> +
> + WARN_ON_ONCE(p != current);
> + p->sched_throttle_work.next = &p->sched_throttle_work;
> + /*
> + * If task is exiting, then there won't be a return to userspace, so we
> + * don't have to bother with any of this.
> + */
> + if ((p->flags & PF_EXITING))
> + return;
> +
> + CLASS(task_rq_lock, rq_guard)(p);
> + rq = rq_guard.rq;
The other way to write this is:
scoped_guard (task_rq_lock, p) {
struct rq *rq = scope.rq;
> + se = &p->se;
> + cfs_rq = cfs_rq_of(se);
> +
> + /*
> + * If not in limbo, then either replenish has happened or this task got
> + * migrated out of the throttled cfs_rq, move along
> + */
> + if (!cfs_rq->throttle_count)
> + return;
> +
> + update_rq_clock(rq);
> +
> + throttle_one_task(cfs_rq, p);
> +
> + resched_curr(rq);
}
> }
Powered by blists - more mailing lists