Received-SPF: pass (google.com: domain of linux-kernel+bounces-82168-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Date: Mon, 26 Feb 2024 10:56:27 -0800
From: "Paul E. McKenney" <paulmck@kernel.org>
To: Z qiang <qiang.zhang1211@gmail.com>
Cc: frederic@kernel.org, neeraj.upadhyay@kernel.org, joel@joelfernandes.org,
	rcu@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] rcu-tasks: Fix the comments for
 tasks_rcu_exit_srcu_stall_timer
Message-ID: <ec57f86d-936c-4709-aeda-d9f57a8fd7aa@paulmck-laptop>
Reply-To: paulmck@kernel.org
References: <20240226032439.26359-1-qiang.zhang1211@gmail.com>
 <CALm+0cXvWG2cP2sECzF9JGyUvvp0bBgSXA8NbGf896iMQEcCuA@mail.gmail.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CALm+0cXvWG2cP2sECzF9JGyUvvp0bBgSXA8NbGf896iMQEcCuA@mail.gmail.com>

On Mon, Feb 26, 2024 at 11:28:57AM +0800, Z qiang wrote:
> >
> > The synchronize_srcu() has been removed by commit("rcu-tasks: Eliminate
> > deadlocks involving do_exit() and RCU tasks") in rcu_tasks_postscan.
> > This commit therefore fix the comments of tasks_rcu_exit_srcu_stall_timer.
> >
> > Signed-off-by: Zqiang <qiang.zhang1211@gmail.com>
> > ---
> >  kernel/rcu/tasks.h | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/kernel/rcu/tasks.h b/kernel/rcu/tasks.h
> > index 78d74c81cc24..d5319bbe8c98 100644
> > --- a/kernel/rcu/tasks.h
> > +++ b/kernel/rcu/tasks.h
> > @@ -150,7 +150,7 @@ static struct rcu_tasks rt_name =                                                   \
> >
> >  #ifdef CONFIG_TASKS_RCU
> >
> > -/* Report delay in synchronize_srcu() completion in rcu_tasks_postscan(). */
> > +/* Report delay of scan exiting tasklist in rcu_tasks_postscan(). */
> >  static void tasks_rcu_exit_srcu_stall(struct timer_list *unused);
> >  static DEFINE_TIMER(tasks_rcu_exit_srcu_stall_timer, tasks_rcu_exit_srcu_stall);
> 
> Is this timer not necessary? any thoughts?

We have preemption points in the list traversals, and things like mutex
contention on the do_exit() path could result in extremely long lists,
so I believe we do need the timer.

But what did you have in mind?

							Thanx, Paul

> Thanks
> Zqiang
> 
> 
> >  #endif
> > --
> > 2.17.1
> >