Received-SPF: pass (google.com: domain of linux-kernel+bounces-126933-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Precedence: bulk
MIME-Version: 1.0
References: <20240328062757.29803-1-mingyang.cui@horizon.ai> <073e5bf0-99c4-4dc5-8894-5442e2d53a34@126.com>
In-Reply-To: <073e5bf0-99c4-4dc5-8894-5442e2d53a34@126.com>
From: John Stultz <jstultz@google.com>
Date: Mon, 1 Apr 2024 11:09:04 -0700
Message-ID: <CANDhNCoBh778bVOSdEtNTDQdw3rt6-zjnLFiiyYR6PZUrhOfew@mail.gmail.com>
Subject: Re: [PATCH] sched/fair: Fix forked task check in vruntime_normalized
To: Honglei Wang <jameshongleiwang@126.com>
Cc: "mingyang.cui" <mingyang.cui@horizon.ai>, mingo@redhat.com, peterz@infradead.org, 
	juri.lelli@redhat.com, vincent.guittot@linaro.org, dietmar.eggemann@arm.com, 
	rostedt@goodmis.org, bsegall@google.com, mgorman@suse.de, bristot@redhat.com, 
	linux-kernel@vger.kernel.org, kernel-team@android.com, stable@vger.kernel.org, 
	tkjos@google.com, pjt@google.com, quentin.perret@arm.com, 
	Patrick.Bellasi@arm.com, Chris.Redpath@arm.com, Morten.Rasmussen@arm.com, 
	joaodias@google.com
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Mon, Apr 1, 2024 at 8:20=E2=80=AFAM Honglei Wang <jameshongleiwang@126.c=
om> wrote:
> On 2024/3/28 14:27, mingyang.cui wrote:
> > When rt_mutex_setprio changes a task's scheduling class to RT,
> > sometimes the task's vruntime is not updated correctly upon
> > return to the fair class.
> > Specifically, the following is being observed:
> > - task has just been created and running for a short time
> > - task sleep while still in the fair class
> > - task is boosted to RT via rt_mutex_setprio, which changes
> >    the task to RT and calls check_class_changed.
> > - check_class_changed leads to detach_task_cfs_rq, at which point
> >    the vruntime_normalized check sees that the task's sum_exec_runtime
> >    is zero, which results in skipping the subtraction of the
> >    rq's min_vruntime from the task's vruntime
>
> Did you do the test on the latest tree? vruntime_normalized was removed
> by e8f331bcc2 (sched/smp: Use lag to simplify cross-runqueue placement).

Indeed (I was looking at an older tree last week and missed it was
removed as well). Though something like this change probably will be
needed for the -stable trees?

thanks
-john