linux-kernel - Re: [PATCH] sched/fair: Fix non-empty throttled_limbo_list warning in tg_throttle

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-prev] [day] [month] [year] [list]

Message-ID: <561a15f5-3391-2796-6454-b980e0a228bd@gmail.com>
Date: Fri, 24 Oct 2025 14:58:41 +0800
From: Hao Jia <jiahao.kernel@...il.com>
To: K Prateek Nayak <kprateek.nayak@....com>, mingo@...hat.com,
 peterz@...radead.org, mingo@...nel.org, juri.lelli@...hat.com,
 vincent.guittot@...aro.org, dietmar.eggemann@....com, rostedt@...dmis.org,
 bsegall@...gle.com, mgorman@...e.de, vschneid@...hat.com
Cc: linux-kernel@...r.kernel.org, Hao Jia <jiahao1@...iang.com>,
 Aaron Lu <ziqianlu@...edance.com>
Subject: Re: [PATCH] sched/fair: Fix non-empty throttled_limbo_list warning in
 tg_throttle_down()


Hi Prateek,

On 2025/10/24 12:36, K Prateek Nayak wrote:
> Hello Hao,
> 
> On 10/23/2025 5:42 PM, Hao Jia wrote:
>> @@ -5287,7 +5287,9 @@ enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags)
>>   	se->on_rq = 1;
>>   
>>   	if (cfs_rq->nr_queued == 1) {
>> -		check_enqueue_throttle(cfs_rq);
>> +		if (!(flags & ENQUEUE_THROTTLE))
>> +			check_enqueue_throttle(cfs_rq);
>> +
> 
> So my only concern here is:
> 
> check_enqueue_throttle()
>    account_cfs_rq_runtime()
>      __account_cfs_rq_runtime()
>        assign_cfs_rq_runtime()
>          __assign_cfs_rq_runtime()
>            start_cfs_bandwidth() /* Starts the BW timer. */
> 
> If we skip it, we wouldn't know we've run out of bandwidth until the
> hierarchy is picked which would cause additional delay until the
> bandwidth is replenished.
> 
> At the very least, we should pass the enqueue flags to
> check_enqueue_throttle() and only skip the throttle_cfs_rq() part if
> we spot ENQUEUE_THROTTLE.
> 
> Thoughts?
> 

Thanks for your suggestion. This is indeed a potential risk, and it will 
do it in the next version.

Thanks,
Hao