Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751866AbbDHLNZ (ORCPT ); Wed, 8 Apr 2015 07:13:25 -0400 Received: from e23smtp08.au.ibm.com ([202.81.31.141]:51063 "EHLO e23smtp08.au.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751229AbbDHLNX (ORCPT ); Wed, 8 Apr 2015 07:13:23 -0400 Date: Wed, 8 Apr 2015 16:42:16 +0530 From: Srikar Dronamraju To: Jason Low Cc: Preeti U Murthy , Morten Rasmussen , "peterz@infradead.org" , "mingo@kernel.org" , Daniel Lezcano , "riel@redhat.com" , "vincent.guittot@linaro.org" , "pjt@google.com" , "benh@kernel.crashing.org" , "efault@gmx.de" , "linux-kernel@vger.kernel.org" , "iamjoonsoo.kim@lge.com" , "svaidy@linux.vnet.ibm.com" , "tim.c.chen@linux.intel.com" Subject: Re: sched: Improve load balancing in the presence of idle CPUs Message-ID: <20150408111216.GA24645@linux.vnet.ibm.com> Reply-To: Srikar Dronamraju References: <1427741729.5694.24.camel@j-VirtualBox> <551A5CCE.70008@linux.vnet.ibm.com> <1427828056.2492.24.camel@j-VirtualBox> <551B9514.80701@linux.vnet.ibm.com> <20150401170418.GX18994@e105550-lin.cambridge.arm.com> <1427954347.2556.43.camel@j-VirtualBox> <551FB5F5.5050906@linux.vnet.ibm.com> <1428449300.2556.79.camel@j-VirtualBox> <1428451666.2556.84.camel@j-VirtualBox> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline In-Reply-To: <1428451666.2556.84.camel@j-VirtualBox> User-Agent: Mutt/1.5.21 (2010-09-15) X-TM-AS-MML: disable X-Content-Scanned: Fidelis XPS MAILER x-cbid: 15040811-0029-0000-0000-0000015CAAAA Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 3310 Lines: 110 * Jason Low [2015-04-07 17:07:46]: > On Tue, 2015-04-07 at 16:28 -0700, Jason Low wrote: > > > Okay, so perhaps we can also try continuing nohz load balancing if we > > find that there are overloaded CPUs in the system. > > Something like the following. > > --- > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c > index fdae26e..d636bf7 100644 > --- a/kernel/sched/fair.c > +++ b/kernel/sched/fair.c > @@ -7620,6 +7620,16 @@ out: > } > > #ifdef CONFIG_NO_HZ_COMMON > +static inline bool nohz_kick_needed(struct rq *rq); > + > +static inline void pass_nohz_balance(struct rq *this_rq, int this_cpu) > +{ > + clear_bit(NOHZ_BALANCE_KICK, nohz_flags(this_cpu)); > + nohz.next_balance = jiffies; Why are we updating nohz.next_balance here? > + if (nohz_kick_needed(this_rq)) > + nohz_balancer_kick(); > +} > + > /* > * In CONFIG_NO_HZ_COMMON case, the idle balance kickee will do the > * rebalancing for all the cpus for whom scheduler ticks are stopped. > @@ -7631,8 +7641,10 @@ static void nohz_idle_balance(struct rq *this_rq, enum cpu_idle_type idle) > int balance_cpu; > > if (idle != CPU_IDLE || Would it make sense to add need_resched here like http://mid.gmane.org/1427442750-8112-1-git-send-email-wanpeng.li@linux.intel.com > - !test_bit(NOHZ_BALANCE_KICK, nohz_flags(this_cpu))) > - goto end; > + !test_bit(NOHZ_BALANCE_KICK, nohz_flags(this_cpu))) { > + pass_nohz_balance(this_rq, this_cpu); > + return; > + } > > for_each_cpu(balance_cpu, nohz.idle_cpus_mask) { > if (balance_cpu == this_cpu || !idle_cpu(balance_cpu)) > @@ -7687,7 +7700,7 @@ static inline bool nohz_kick_needed(struct rq *rq) > int nr_busy, cpu = rq->cpu; > bool kick = false; > > - if (unlikely(rq->idle_balance)) > + if (unlikely(idle_cpu(cpu))) > return false; The only other place that we use idle_balance is run_rebalance_domains(). Would it make sense to just use idle_cpu() in run_rebalance_domains() and remove rq->idle_balance? > > /* > @@ -7707,7 +7720,7 @@ static inline bool nohz_kick_needed(struct rq *rq) > if (time_before(now, nohz.next_balance)) > return false; > > - if (rq->nr_running >= 2) > + if (rq->nr_running >= 2 || rq->rd->overload) > return true; > > rcu_read_lock(); > @@ -7757,16 +7770,14 @@ static void run_rebalance_domains(struct softirq_action *h) > enum cpu_idle_type idle = this_rq->idle_balance ? > CPU_IDLE : CPU_NOT_IDLE; > > + rebalance_domains(this_rq, idle); > + > /* > * If this cpu has a pending nohz_balance_kick, then do the > * balancing on behalf of the other idle cpus whose ticks are > - * stopped. Do nohz_idle_balance *before* rebalance_domains to > - * give the idle cpus a chance to load balance. Else we may > - * load balance only within the local sched_domain hierarchy > - * and abort nohz_idle_balance altogether if we pull some load. > + * stopped. > */ > nohz_idle_balance(this_rq, idle); > - rebalance_domains(this_rq, idle); > } > > /* > > -- Thanks and Regards Srikar Dronamraju -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/