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;
References: <20220708211755.73637-1-schspa@gmail.com>
 <20220708172614.14191089@gandalf.local.home>
User-agent: mu4e 1.7.5; emacs 28.1
From:   Schspa Shi <schspa@gmail.com>
To:     Steven Rostedt <rostedt@goodmis.org>
Cc:     mingo@redhat.com, peterz@infradead.org, juri.lelli@redhat.com,
        vincent.guittot@linaro.org, dietmar.eggemann@arm.com,
        bsegall@google.com, mgorman@suse.de, bristot@redhat.com,
        vschneid@redhat.com, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v4 1/2] sched/rt: fix bad task migration for rt tasks
Date:   Sat, 09 Jul 2022 05:32:25 +0800
In-reply-to: <20220708172614.14191089@gandalf.local.home>
Message-ID: <m2v8s7mg78.fsf@gmail.com>
MIME-Version: 1.0
Content-Type: text/plain
Precedence: bulk


Steven Rostedt <rostedt@goodmis.org> writes:

> On Sat,  9 Jul 2022 05:17:54 +0800
> Schspa Shi <schspa@gmail.com> wrote:
>
>> +++ b/kernel/sched/rt.c
>> @@ -1998,11 +1998,14 @@ static struct rq *find_lock_lowest_rq(struct task_struct *task, struct rq *rq)
>>  			 * the mean time, task could have
>>  			 * migrated already or had its affinity changed.
>>  			 * Also make sure that it wasn't scheduled on its rq.
>> +			 * It is possible the task has running for a while,
>
> I don't understand the "running for a while" part. That doesn't make sense.
>

When I say "run for a while" I mean as long as the task has
run capability, we should check the migrate disabled flag again.

> The only way this can happen is that it was scheduled, set
> "migrate_disabled" and then got preempted where it's no longer on the run
> queue.

Yes, it is the only case.

> -- Steve
>
>
>> +			 * And we check task migration disable flag here.
>>  			 */
>>  			if (unlikely(task_rq(task) != rq ||
>>  				     !cpumask_test_cpu(lowest_rq->cpu, &task->cpus_mask) ||
>>  				     task_running(rq, task) ||
>>  				     !rt_task(task) ||
>> +				     is_migration_disabled(task) ||
>>  				     !task_on_rq_queued(task))) {
>>  
>>  				double_unlock_balance(rq, lowest_rq);

-- 
BRs
Schspa Shi