linux-kernel - Re: [PATCH v2] sched: async unthrottling for cfs bandwidth

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

Message-ID: <20221116095740.GA29859@blackbody.suse.cz>
Date:   Wed, 16 Nov 2022 10:57:40 +0100
From:   Michal Koutný <mkoutny@...e.com>
To:     Josh Don <joshdon@...gle.com>
Cc:     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>,
        Daniel Bristot de Oliveira <bristot@...hat.com>,
        Valentin Schneider <vschneid@...hat.com>,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2] sched: async unthrottling for cfs bandwidth

On Tue, Nov 15, 2022 at 07:01:31PM -0800, Josh Don <joshdon@...gle.com> wrote:
> After more thought, I realized that we can't reuse the throttled_list
> list_head, since that would potentially break the lockless traversal
> of a concurrent list_for_each_entry_rcu() (ie. if we removed the
> element from the throttled list and then added it to the CSD list).

I see, the concurrent RCU traversal is a valid point for the two heads.

What does it mean for SCHED_WARN_ON in __unthrottle_cfs_rq_async()?

IIUC, if the concurrency of cfs_b->throttled_cfs_rq list is
expected (hence I'm not sure about the SCHED_WARN_ON), then it may
happen that __unthrottle_cfs_rq_async is called on cfs_rq that's already
on rq->cfsb_csd_list (there's still rq lock but it's only help inside
cfs_b->throttled_cfs_rq iteration).

Thanks,
Michal

Download attachment "signature.asc" of type "application/pgp-signature" (229 bytes)