MIME-Version: 1.0
References: <20230120150246.20797-1-wander@redhat.com> <20230120150246.20797-2-wander@redhat.com>
 <20230123163002.GB6268@redhat.com>
In-Reply-To: <20230123163002.GB6268@redhat.com>
From:   Wander Lairson Costa <wander@redhat.com>
Date:   Mon, 23 Jan 2023 14:24:40 -0300
Message-ID: <CAAq0SUk1vfNDuzGbXNftgW4wq4PC_EzMhpq4E=RBQNkOB3f4YQ@mail.gmail.com>
Subject: Re: [PATCH v2 1/4] sched/task: Add the put_task_struct_atomic_safe function
To:     Oleg Nesterov <oleg@redhat.com>
Cc:     Ingo Molnar <mingo@redhat.com>,
        Peter Zijlstra <peterz@infradead.org>,
        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>,
        "Eric W. Biederman" <ebiederm@xmission.com>,
        Stafford Horne <shorne@gmail.com>,
        Kefeng Wang <wangkefeng.wang@huawei.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
        Andy Lutomirski <luto@kernel.org>,
        "Liam R. Howlett" <Liam.Howlett@oracle.com>,
        Fenghua Yu <fenghua.yu@intel.com>,
        Andrei Vagin <avagin@gmail.com>,
        open list <linux-kernel@vger.kernel.org>,
        Paul McKenney <paulmck@kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Mon, Jan 23, 2023 at 1:30 PM Oleg Nesterov <oleg@redhat.com> wrote:
>
> On 01/20, Wander Lairson Costa wrote:
> >
> > +static inline void put_task_struct_atomic_safe(struct task_struct *task)
> > +{
> > +     if (IS_ENABLED(CONFIG_PREEMPT_RT)) {
> > +             /*
> > +              * Decrement the refcount explicitly to avoid unnecessarily
> > +              * calling call_rcu.
> > +              */
> > +             if (refcount_dec_and_test(&task->usage))
> > +                     /*
> > +                      * under PREEMPT_RT, we can't call put_task_struct
> > +                      * in atomic context because it will indirectly
> > +                      * acquire sleeping locks.
> > +                      */
> > +                     call_rcu(&task->rcu, __delayed_put_task_struct);
>                                   ^^^^^^^^^
> I am not sure the usage of task->rcu is safe...
>
> Suppose that, before __delayed_put_task_struct() is called by RCU, this task
> does the last schedule and calls put_task_struct_rcu_user().
>
> And, can't we simply turn put_task_struct() into something like
>
>         put_task_struct(struct task_struct *t)
>         {
>                 if (refcount_dec_and_test(&t->usage)) {
>                         if (IS_ENABLED(CONFIG_PREEMPT_RT)
>                             && (in_atomic() || irqs_disabled()))
>                                 call_rcu(...);
>                         else
>                                 __put_task_struct(t);
>                 }
>         }
>
> ?

Yeah, that was one approach I thought about. I chose to use an
explicit function because I assumed calling __put_task_struct() from a
non-preemptable context should be the exception, not the rule.
Therefore (if I am correct in my assumption), it would make sense for
only some call sites to pay the overhead price for it. But this is
just a guess, and I have no evidence to support my claim.