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;
Sender: Tejun Heo <htejun@gmail.com>
Date:   Tue, 1 Nov 2022 12:38:23 -1000
From:   Tejun Heo <tj@kernel.org>
To:     Josh Don <joshdon@google.com>
Cc:     Peter Zijlstra <peterz@infradead.org>,
        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,
        Joel Fernandes <joel@joelfernandes.org>,
        Michal =?iso-8859-1?Q?Koutn=FD?= <mkoutny@suse.com>,
        Christian Brauner <brauner@kernel.org>,
        Zefan Li <lizefan.x@bytedance.com>
Subject: Re: [PATCH v2] sched: async unthrottling for cfs bandwidth
Message-ID: <Y2Gf3zxJqxRnkVyf@slm.duckdns.org>
References: <Y2BDFNpkSawKnE9S@slm.duckdns.org>
 <CABk29Nta-RJpTcybgOk9u4DH=1mwQFZsOxFuQ-UpCcTwzjzAuA@mail.gmail.com>
 <Y2Bf+CeQ8x2jKQ3S@slm.duckdns.org>
 <CABk29Nvqv-T1JuAq2cf9=AwRu=y1+YOR4xS2qnVo6+XpWd2UNQ@mail.gmail.com>
 <Y2B6V1PPuCcTXGp6@slm.duckdns.org>
 <CABk29Ns1VWEVRYENud4CW3JQPrcr79i_F2PBTANqt3t-LaYCfQ@mail.gmail.com>
 <Y2FwVX42LIKXSTz3@slm.duckdns.org>
 <CABk29Nua8ZsDfhY+x+VfYDkbkjfXLXTZ5JMVR9uiBygraxDM+g@mail.gmail.com>
 <Y2GUg8CiI68ZBznr@slm.duckdns.org>
 <CABk29Nvj8nRyD0HGo+gZ4CEr0kOJSsUbJnSNFs62D66EDTMGog@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CABk29Nvj8nRyD0HGo+gZ4CEr0kOJSsUbJnSNFs62D66EDTMGog@mail.gmail.com>
Precedence: bulk

Hello,

(cc'ing Michal, Christian and Li for context)

On Tue, Nov 01, 2022 at 02:59:56PM -0700, Josh Don wrote:
> > If most of the problems were with cpu bw control, fixing that should do for
> > the time being. Otherwise, we'll have to think about finishing kernfs
> > locking granularity improvements and doing something similar to cgroup
> > locking too.
> 
> Oh we've easily hit stalls measured in multiple seconds. We
> extensively use cpu.idle to group batch tasks. One of the memory
> bandwidth mitigations implemented in userspace is cpu jailing, which
> can end up pushing lots and lots of these batch threads onto a small
> number of cpus. 5ms min gran * 200 threads is already one second :)

Ah, I see.

> We're in the process of transitioning to using bw instead for this
> instead in order to maintain parallelism. Fixing bw is definitely
> going to be useful, but I'm afraid we'll still likely have some issues
> from low throughput for non-bw reasons (some of which we can't
> directly control, since arbitrary jobs can spin up and configure their
> hierarchy/threads in antagonistic ways, in effect pushing out the
> latency of some of their threads).

Yeah, thanks for the explanation. Making the lock more granular is tedious
but definitely doable. I don't think I can work on it in the near future but
will keep it on mind. If anyone's interested in attacking it, please be my
guest.

Thanks.

-- 
tejun