[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CANDhNCrN6o=_uKN4xK2uFfsKBQuhq7Rz82jgBss4LSWgK-OpqQ@mail.gmail.com>
Date: Tue, 22 Aug 2023 12:24:56 -0700
From: John Stultz <jstultz@...gle.com>
To: Waiman Long <longman@...hat.com>
Cc: LKML <linux-kernel@...r.kernel.org>,
Joel Fernandes <joelaf@...gle.com>,
Qais Yousef <qyousef@...gle.com>,
Ingo Molnar <mingo@...hat.com>,
Peter Zijlstra <peterz@...radead.org>,
Juri Lelli <juri.lelli@...hat.com>,
Vincent Guittot <vincent.guittot@...aro.org>,
Dietmar Eggemann <dietmar.eggemann@....com>,
Valentin Schneider <vschneid@...hat.com>,
Steven Rostedt <rostedt@...dmis.org>,
Ben Segall <bsegall@...gle.com>,
Zimuzo Ezeozue <zezeozue@...gle.com>,
Youssef Esmat <youssefesmat@...gle.com>,
Mel Gorman <mgorman@...e.de>,
Daniel Bristot de Oliveira <bristot@...hat.com>,
Will Deacon <will@...nel.org>,
Boqun Feng <boqun.feng@...il.com>,
"Paul E . McKenney" <paulmck@...nel.org>, kernel-team@...roid.com
Subject: Re: [PATCH v5 02/19] locking/mutex: Removes wakeups from under mutex::wait_lock
On Tue, Aug 22, 2023 at 12:11 PM Waiman Long <longman@...hat.com> wrote:
> On 8/19/23 02:08, John Stultz wrote:
> > @@ -676,6 +677,10 @@ __mutex_lock_common(struct mutex *lock, unsigned int state, unsigned int subclas
> > }
> >
> > raw_spin_unlock(&lock->wait_lock);
> > + /* Make sure we do wakeups before calling schedule */
> > + wake_up_q(&wake_q);
> > + wake_q_init(&wake_q);
> > +
>
> The wake_q may have task to wake up only in the case of ww_mutex which
> is a minority in the kernel. IOW, wake_up_q() which is a function call
> will do nothing in most cases. From an optimization point of view, it is
> better to do a "!wake_q_empty(&wake_q)" check before calling wake_up_q().
Thanks for the suggestion! Updated for the next version!
> > @@ -946,9 +953,11 @@ static noinline void __sched __mutex_unlock_slowpath(struct mutex *lock, unsigne
> > if (owner & MUTEX_FLAG_HANDOFF)
> > __mutex_handoff(lock, next);
> >
> > + preempt_disable();
> > raw_spin_unlock(&lock->wait_lock);
> >
> > wake_up_q(&wake_q);
> > + preempt_enable();
> > }
>
> I think it looks better to put the preempt_disable() right before
> raw_spin_lock() for proper nesting.
Agreed.
Thanks so much for the review and feedback! I really appreciate it!
-john
Powered by blists - more mailing lists