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;
Date:   Wed, 5 Oct 2022 16:46:47 +0100
From:   Qais Yousef <qais.yousef@arm.com>
To:     David Laight <David.Laight@ACULAB.COM>
Cc:     'John Stultz' <jstultz@google.com>,
        LKML <linux-kernel@vger.kernel.org>,
        Connor O'Brien <connoro@google.com>,
        John Dias <joaodias@google.com>, Rick Yiu <rickyiu@google.com>,
        John Kacur <jkacur@redhat.com>,
        Chris Redpath <chris.redpath@arm.com>,
        Abhijeet Dharmapurikar <adharmap@quicinc.com>,
        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>,
        Thomas Gleixner <tglx@linutronix.de>,
        "kernel-team@android.com" <kernel-team@android.com>,
        "J . Avila" <elavila@google.com>
Subject: Re: [RFC PATCH v3 2/3] sched: Avoid placing RT threads on cores
 handling long softirqs
Message-ID: <20221005154647.vok6unu3vnzxp74d@wubuntu>
References: <20220921012550.3288570-1-jstultz@google.com>
 <20220921012550.3288570-3-jstultz@google.com>
 <20220928125517.ei64pxfucaem55cr@wubuntu>
 <CANDhNCpLp+St4U_fjs+W8HXT-2ff4gpkx0P6vDnrX0CcesiQfA@mail.gmail.com>
 <efbeb4b069384aa39a80d606d14ec833@AcuMS.aculab.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <efbeb4b069384aa39a80d606d14ec833@AcuMS.aculab.com>
Precedence: bulk

On 10/04/22 09:50, David Laight wrote:

[...]

> > That's fair. Digging through the patch history in the Android trees,
> > the first pass was for all softirqs but then restricted to remove
> > known short-running ones.
> > From the bug history and what I can directly reproduce, the net and
> > block softirqs have definitely caused trouble, but I don't see a
> > specific example from TASKLET,  so I'm ok dropping that for now, and
> > should we get specific evidence we can argue for it in a future patch.
> > 
> > So I'll drop TASKLET from the list here. Thanks for the suggestion!
> 
> I've also seen the code that finally frees memory freed under rcu
> take a long time.
> That was a workload sending a lot of UDP/RTP from a raw socket using
> IP_HDRINC - each send allocated a structure (fib?) that was freed from
> the rcu (timer?) softint callback.

I'm assuming this is a network driver that using RCU callback to free some
memory?

> 
> But, actually, one of the biggest causes of RT wakeup latency
> was a normal thread looping without a cond_resched() call.
> In my case some graphics driver doing page flushes of the
> display memory.

Were these drivers that cause these problem in-tree or out-of-tree?


Thanks

--
Qais Yousef