Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20221117005418.3499691-1-joshdon@google.com> <Y3d+1a9AEnWaxFwq@hirez.programming.kicks-ass.net>
 <CABk29NtSmXVCvkdpymeam7AYmXhZy2JLYLPFTdKpk5g6AN1-zg@mail.gmail.com> <Y3xnUhjSb56ex9XX@ziqianlu-desk2>
In-Reply-To: <Y3xnUhjSb56ex9XX@ziqianlu-desk2>
From:   Josh Don <joshdon@google.com>
Date:   Tue, 22 Nov 2022 11:41:04 -0800
Message-ID: <CABk29Ntdztkv4jT87vFgTtC99d49iUu9CcHPruh9MxDazaAW7A@mail.gmail.com>
Subject: Re: [PATCH v3] sched: async unthrottling for cfs bandwidth
To:     Aaron Lu <aaron.lu@intel.com>,
        Peter Zijlstra <peterz@infradead.org>
Cc:     Ingo Molnar <mingo@redhat.com>, 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, Tejun Heo <tj@kernel.org>,
        =?UTF-8?Q?Michal_Koutn=C3=BD?= <mkoutny@suse.com>,
        Christian Brauner <brauner@kernel.org>,
        Zefan Li <lizefan.x@bytedance.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

> > +        */
> > +       if (local_unthrottle) {
> > +               rq = cpu_rq(this_cpu);
> > +               rq_lock_irqsave(rq, &rf);
>
> Should we add:
>                 if (cfs_rq_throttled(local_unthrottle))
>
> before calling into unthrottle_cfs_rq_async(local_unthrottle) to avoid a
> potential WARN?
>
> As for whether the local cfs_rq can be unthrottled now after rq lock is
> re-acquired, I suppose it can be. e.g. another user sets a new quota to
> this task group during the window of rq lock gets dropped in the above
> loop and re-acquired here IIUC.
>
> > +               unthrottle_cfs_rq_async(local_unthrottle);
> > +               rq_unlock_irqrestore(rq, &rf);
> > +       }
> > +
> >         return throttled;
> >  }

Yes, we should add that check due to the case you described with a
user concurrently configuring bandwidth. And as long as we're doing
that, we might as well make this unthrottle_cfs_rq() instead and snip
the comment. Peter, would you mind adding that delta?