Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.37 as permitted sender) client-ip=23.128.96.37;
Sender: Ingo Molnar <mingo.kernel.org@gmail.com>
Date:   Sun, 24 Sep 2023 12:12:35 +0200
From:   Ingo Molnar <mingo@kernel.org>
To:     Josh Don <joshdon@google.com>
Cc:     Ingo Molnar <mingo@redhat.com>,
        Peter Zijlstra <peterz@infradead.org>,
        Juri Lelli <juri.lelli@redhat.com>,
        Vincent Guittot <vincent.guittot@linaro.org>,
        Dietmar Eggemann <dietmar.eggemann@arm.com>,
        Steven Rostedt <rostedt@goodmis.org>,
        Ben Segall <bsegall@google.com>, Mel Gorman <mgorman@suse.de>,
        Daniel Bristot de Oliveira <bristot@redhat.com>,
        Valentin Schneider <vschneid@redhat.com>,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH 2/2] sched: fix warning in bandwidth distribution
Message-ID: <ZRALk+fMpW5895fF@gmail.com>
References: <20230922230535.296350-1-joshdon@google.com>
 <20230922230535.296350-2-joshdon@google.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20230922230535.296350-2-joshdon@google.com>
Precedence: bulk


* Josh Don <joshdon@google.com> wrote:

> We've observed the following warning being hit in
> distribute_cfs_runtime():
> 	SCHED_WARN_ON(cfs_rq->runtime_remaining > 0)
> 
> We have the following race:
> 
> - cpu0: running bandwidth distribution (distribute_cfs_runtime).
>   Inspects the local cfs_rq and makes its runtime_remaining positive.
>   However, we defer unthrottling the local cfs_rq until after
>   considering all remote cfs_rq's.
> - cpu1: starts running bandwidth distribution from the slack timer. When
>   it finds the cfs_rq for cpu 0 on the throttled list, it observers the
>   that the cfs_rq is throttled, yet is not on the CSD list, and has a
>   positive runtime_remaining, thus triggering the warning in
>   distribute_cfs_runtime.
> 
> To fix this, we can rework the local unthrottling logic to put the local
> cfs_rq on a local list, so that any future bandwidth distributions will
> realize that the cfs_rq is about to be unthrottled.
> 
> Signed-off-by: Josh Don <joshdon@google.com>
> ---
>  kernel/sched/fair.c | 36 +++++++++++++++++++++++++-----------
>  1 file changed, 25 insertions(+), 11 deletions(-)
> 
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index 8f4e63fc8900..de002dab28cf 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -5743,13 +5743,13 @@ static void unthrottle_cfs_rq_async(struct cfs_rq *cfs_rq)
>  
>  static bool distribute_cfs_runtime(struct cfs_bandwidth *cfs_b)
>  {
> -	struct cfs_rq *local_unthrottle = NULL;
>  	int this_cpu = smp_processor_id();
>  	u64 runtime, remaining = 1;
>  	bool throttled = false;
> -	struct cfs_rq *cfs_rq;
> +	struct cfs_rq *cfs_rq, *tmp;
>  	struct rq_flags rf;
>  	struct rq *rq;
> +	LIST_HEAD(local_unthrottle);
>  
>  	rcu_read_lock();
>  	list_for_each_entry_rcu(cfs_rq, &cfs_b->throttled_cfs_rq,
> @@ -5784,11 +5784,17 @@ static bool distribute_cfs_runtime(struct cfs_bandwidth *cfs_b)
>  
>  		/* we check whether we're throttled above */
>  		if (cfs_rq->runtime_remaining > 0) {
> -			if (cpu_of(rq) != this_cpu ||
> -			    SCHED_WARN_ON(local_unthrottle))
> +			if (cpu_of(rq) != this_cpu) {
>  				unthrottle_cfs_rq_async(cfs_rq);
> -			else
> -				local_unthrottle = cfs_rq;
> +			} else {
> +				/*
> +				 * We currently only expect to be unthrottling
> +				 * a single cfs_rq locally.
> +				 */
> +				SCHED_WARN_ON(!list_empty(&local_unthrottle));
> +				list_add_tail(&cfs_rq->throttled_csd_list,
> +					      &local_unthrottle);
> +			}
>  		} else {
>  			throttled = true;
>  		}
> @@ -5796,15 +5802,23 @@ static bool distribute_cfs_runtime(struct cfs_bandwidth *cfs_b)
>  next:
>  		rq_unlock_irqrestore(rq, &rf);
>  	}
> -	rcu_read_unlock();
>  
> -	if (local_unthrottle) {
> -		rq = cpu_rq(this_cpu);
> +	list_for_each_entry_safe(cfs_rq, tmp, &local_unthrottle,
> +				 throttled_csd_list) {
> +		struct rq *rq = rq_of(cfs_rq);
> +
>  		rq_lock_irqsave(rq, &rf);
> -		if (cfs_rq_throttled(local_unthrottle))
> -			unthrottle_cfs_rq(local_unthrottle);
> +
> +		list_del_init(&cfs_rq->throttled_csd_list);
> +
> +		if (cfs_rq_throttled(cfs_rq))
> +			unthrottle_cfs_rq(cfs_rq);
> +
>  		rq_unlock_irqrestore(rq, &rf);
>  	}
> +	SCHED_WARN_ON(!list_empty(&local_unthrottle));
> +
> +	rcu_read_unlock();

Thanks, this looks much cleaner.

When the warning hits, we don't have any other side-effects,
such as bad behavior or data corruption, correct?

Under that assumption I've queued your fix in tip:sched/core,
for a v6.7 merge, and not in tip:sched/urgent for a v6.6 merge,
but let me know if I'm reading the code wrong...

Thanks,

	Ingo