From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754368Ab1KWMfY (ORCPT ); Wed, 23 Nov 2011 07:35:24 -0500 Received: from cantor2.suse.de ([195.135.220.15]:53849 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751115Ab1KWMfX (ORCPT ); Wed, 23 Nov 2011 07:35:23 -0500 Subject: Re: [patch 6/7] sched: use rq->avg_event to resurrect nohz ratelimiting From: Mike Galbraith To: Peter Zijlstra Cc: Suresh Siddha , linux-kernel , Ingo Molnar , Paul Turner , Norbert Preining In-Reply-To: <1321971893.6855.24.camel@marge.simson.net> References: <1321350377.1421.55.camel@twins> <1321406062.16760.60.camel@sbsiddha-desk.sc.intel.com> <1321435455.5072.64.camel@marge.simson.net> <1321468646.11680.2.camel@sbsiddha-desk.sc.intel.com> <1321495153.5100.7.camel@marge.simson.net> <1321544313.6308.25.camel@marge.simson.net> <1321545376.2495.1.camel@laptop> <1321547917.6308.48.camel@marge.simson.net> <1321551381.15339.21.camel@sbsiddha-desk.sc.intel.com> <1321629267.7080.13.camel@marge.simson.net> <1321629441.7080.15.camel@marge.simson.net> <1321630552.2197.15.camel@twins> <1321971445.6855.14.camel@marge.simson.net> <1321971893.6855.24.camel@marge.simson.net> Content-Type: text/plain; charset="UTF-8" Date: Wed, 23 Nov 2011 13:35:19 +0100 Message-ID: <1322051719.7041.15.camel@marge.simson.net> Mime-Version: 1.0 X-Mailer: Evolution 2.32.1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Added CC of the person who reported that the last nohz ratelimit patch caused increased power consumption. Norbert: can you please test this patch to make sure it's not going to cause you the same trouble the last one did? I see no ill effects with this patch, but then I didn't with the last either. TIA, -Mike On Tue, 2011-11-22 at 15:24 +0100, Mike Galbraith wrote: > Entering nohz at high frequency eats cycles, ratelimit it > just as we do idle_balance(), and for the same reason. > > Signed-off-by: Mike Galbraith > > --- > include/linux/sched.h | 2 ++ > kernel/sched/core.c | 7 +++++++ > kernel/time/tick-sched.c | 2 +- > 3 files changed, 10 insertions(+), 1 deletion(-) > > Index: linux-3.0-tip/include/linux/sched.h > =================================================================== > --- linux-3.0-tip.orig/include/linux/sched.h > +++ linux-3.0-tip/include/linux/sched.h > @@ -1988,8 +1988,10 @@ static inline void idle_task_exit(void) > > #if defined(CONFIG_NO_HZ) && defined(CONFIG_SMP) > extern void wake_up_idle_cpu(int cpu); > +extern int sched_needs_cpu(int cpu); > #else > static inline void wake_up_idle_cpu(int cpu) { } > +static inline void sched_needs_cpu(int cpu) { } > #endif > > extern unsigned int sysctl_sched_latency; > Index: linux-3.0-tip/kernel/sched/core.c > =================================================================== > --- linux-3.0-tip.orig/kernel/sched/core.c > +++ linux-3.0-tip/kernel/sched/core.c > @@ -587,6 +587,13 @@ static inline bool got_nohz_idle_kick(vo > return idle_cpu(smp_processor_id()) && this_rq()->nohz_balance_kick; > } > > +#ifdef CONFIG_SMP > +int sched_needs_cpu(int cpu) > +{ > + return cpu_rq(cpu)->avg_event < sysctl_sched_migration_cost; > +} > +#endif > + > #else /* CONFIG_NO_HZ */ > > static inline bool got_nohz_idle_kick(void) > Index: linux-3.0-tip/kernel/time/tick-sched.c > =================================================================== > --- linux-3.0-tip.orig/kernel/time/tick-sched.c > +++ linux-3.0-tip/kernel/time/tick-sched.c > @@ -352,7 +352,7 @@ void tick_nohz_stop_sched_tick(int inidl > } while (read_seqretry(&xtime_lock, seq)); > > if (rcu_needs_cpu(cpu) || printk_needs_cpu(cpu) || > - arch_needs_cpu(cpu)) { > + arch_needs_cpu(cpu) || sched_needs_cpu(cpu)) { > next_jiffies = last_jiffies + 1; > delta_jiffies = 1; > } else { > >