From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750964Ab0LFFdd (ORCPT ); Mon, 6 Dec 2010 00:33:33 -0500 Received: from mailout-de.gmx.net ([213.165.64.22]:53076 "HELO mail.gmx.net" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with SMTP id S1750748Ab0LFFdc (ORCPT ); Mon, 6 Dec 2010 00:33:32 -0500 X-Authenticated: #14349625 X-Provags-ID: V01U2FsdGVkX19wi9RQXHRwb5gdJl2DyJc0SciTyp9+NZhnAo8YLQ +FUVv9xc2ihmO0 Subject: Re: Scheduler bug related to rq->skip_clock_update? From: Mike Galbraith To: Yong Zhang Cc: "Bjoern B. Brandenburg" , Peter Zijlstra , Ingo Molnar , Andrea Bastoni , "James H. Anderson" , linux-kernel@vger.kernel.org In-Reply-To: <20101205052819.GA2878@zhy> References: <1290359641.4816.69.camel@maggy.simson.net> <1290442781.16393.22.camel@maggy.simson.net> <20101205052819.GA2878@zhy> Content-Type: text/plain Date: Mon, 06 Dec 2010 06:33:27 +0100 Message-Id: <1291613607.9457.5.camel@marge.simson.net> Mime-Version: 1.0 X-Mailer: Evolution 2.24.1.1 Content-Transfer-Encoding: 7bit X-Y-GMX-Trusted: 0 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Sun, 2010-12-05 at 13:28 +0800, Yong Zhang wrote: > when we init idle task, we doesn't mark it on_rq. > My test show the concern is smoothed by below patch. Close :) The skip_clock_update flag should only be set if rq->curr is on_rq, because it it _that_ clock update during dequeue, and subsequent microscopic vruntime update it causes that we're trying to avoid. I think the below fixes it up properly. Sched: fix skip_clock_update optimization idle_balance() drops/retakes rq->lock, leaving the previous task vulnerable to set_tsk_need_resched(). Clear it after we return from balancing instead, and in setup_thread_stack() as well, so no successfully descheduled or never scheduled task has it set. Need resched confused the skip_clock_update logic, which assumes that the next call to update_rq_clock() will come nearly immediately after being set. Make the optimization robust against the waking a sleeper before it sucessfully deschedules case by checking that the current task has not been dequeued before setting the flag, since it is that useless clock update we're trying to save, and clear in update_rq_clock() to ensure that ONE call may be skipped. Signed-off-by: Mike Galbraith Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Bjoern B. Brandenburg Reported-by: Bjoern B. Brandenburg --- kernel/fork.c | 1 + kernel/sched.c | 6 +++--- 2 files changed, 4 insertions(+), 3 deletions(-) Index: linux-2.6.37.git/kernel/sched.c =================================================================== --- linux-2.6.37.git.orig/kernel/sched.c +++ linux-2.6.37.git/kernel/sched.c @@ -660,6 +660,7 @@ inline void update_rq_clock(struct rq *r sched_irq_time_avg_update(rq, irq_time); } + rq->skip_clock_update = 0; } /* @@ -2138,7 +2139,7 @@ static void check_preempt_curr(struct rq * A queue event has occurred, and we're going to schedule. In * this case, we can save a useless back to back clock update. */ - if (test_tsk_need_resched(rq->curr)) + if (rq->curr->se.on_rq && test_tsk_need_resched(rq->curr)) rq->skip_clock_update = 1; } @@ -3854,7 +3855,6 @@ static void put_prev_task(struct rq *rq, { if (prev->se.on_rq) update_rq_clock(rq); - rq->skip_clock_update = 0; prev->sched_class->put_prev_task(rq, prev); } @@ -3912,7 +3912,6 @@ need_resched_nonpreemptible: hrtick_clear(rq); raw_spin_lock_irq(&rq->lock); - clear_tsk_need_resched(prev); switch_count = &prev->nivcsw; if (prev->state && !(preempt_count() & PREEMPT_ACTIVE)) { @@ -3942,6 +3941,7 @@ need_resched_nonpreemptible: if (unlikely(!rq->nr_running)) idle_balance(cpu, rq); + clear_tsk_need_resched(prev); put_prev_task(rq, prev); next = pick_next_task(rq); Index: linux-2.6.37.git/kernel/fork.c =================================================================== --- linux-2.6.37.git.orig/kernel/fork.c +++ linux-2.6.37.git/kernel/fork.c @@ -275,6 +275,7 @@ static struct task_struct *dup_task_stru setup_thread_stack(tsk, orig); clear_user_return_notifier(tsk); + clear_tsk_need_resched(tsk); stackend = end_of_stack(tsk); *stackend = STACK_END_MAGIC; /* for overflow detection */