From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail144.messagelabs.com (mail144.messagelabs.com [216.82.254.51]) by kanga.kvack.org (Postfix) with SMTP id 1C2D56B01C1 for ; Fri, 28 May 2010 02:38:29 -0400 (EDT) Received: from m1.gw.fujitsu.co.jp ([10.0.50.71]) by fgwmail5.fujitsu.co.jp (Fujitsu Gateway) with ESMTP id o4S6cPRN011918 for (envelope-from kosaki.motohiro@jp.fujitsu.com); Fri, 28 May 2010 15:38:25 +0900 Received: from smail (m1 [127.0.0.1]) by outgoing.m1.gw.fujitsu.co.jp (Postfix) with ESMTP id 7545645DE4F for ; Fri, 28 May 2010 15:38:25 +0900 (JST) Received: from s1.gw.fujitsu.co.jp (s1.gw.fujitsu.co.jp [10.0.50.91]) by m1.gw.fujitsu.co.jp (Postfix) with ESMTP id 5573E45DE4D for ; Fri, 28 May 2010 15:38:25 +0900 (JST) Received: from s1.gw.fujitsu.co.jp (localhost.localdomain [127.0.0.1]) by s1.gw.fujitsu.co.jp (Postfix) with ESMTP id 4248DE08001 for ; Fri, 28 May 2010 15:38:25 +0900 (JST) Received: from m107.s.css.fujitsu.com (m107.s.css.fujitsu.com [10.249.87.107]) by s1.gw.fujitsu.co.jp (Postfix) with ESMTP id EB70D1DB803C for ; Fri, 28 May 2010 15:38:24 +0900 (JST) From: KOSAKI Motohiro Subject: Re: [RFC] oom-kill: give the dying task a higher priority In-Reply-To: <20100528062701.GA3519@balbir.in.ibm.com> References: <20100528134133.7E24.A69D9226@jp.fujitsu.com> <20100528062701.GA3519@balbir.in.ibm.com> Message-Id: <20100528153410.7E30.A69D9226@jp.fujitsu.com> MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit Date: Fri, 28 May 2010 15:38:24 +0900 (JST) Sender: owner-linux-mm@kvack.org To: balbir@linux.vnet.ibm.com Cc: kosaki.motohiro@jp.fujitsu.com, "Luis Claudio R. Goncalves" , Oleg Nesterov , linux-kernel@vger.kernel.org, linux-mm@kvack.org, Thomas Gleixner , Peter Zijlstra , David Rientjes , Mel Gorman , williams@redhat.com List-ID: > * KOSAKI Motohiro [2010-05-28 13:46:53]: > > > > * Luis Claudio R. Goncalves [2010-05-28 00:51:47]: > > > > > > > @@ -382,6 +382,8 @@ static void dump_header(struct task_struct *p, gfp_t gfp_mask, int order, > > > > */ > > > > static void __oom_kill_task(struct task_struct *p, int verbose) > > > > { > > > > + struct sched_param param; > > > > + > > > > if (is_global_init(p)) { > > > > WARN_ON(1); > > > > printk(KERN_WARNING "tried to kill init!\n"); > > > > @@ -413,8 +415,9 @@ static void __oom_kill_task(struct task_struct *p, int verbose) > > > > */ > > > > p->rt.time_slice = HZ; > > > > set_tsk_thread_flag(p, TIF_MEMDIE); > > > > - > > > > force_sig(SIGKILL, p); > > > > + param.sched_priority = MAX_RT_PRIO-1; > > > > + sched_setscheduler_nocheck(p, SCHED_FIFO, ¶m); > > > > } > > > > > > > > > > I would like to understand the visible benefits of this patch. Have > > > you seen an OOM kill tasked really get bogged down. Should this task > > > really be competing with other important tasks for run time? > > > > What you mean important? Until OOM victim task exit completely, the system have no memory. > > all of important task can't do anything. > > > > In almost kernel subsystems, automatically priority boost is really bad idea because > > it may break RT task's deterministic behavior. but OOM is one of exception. The deterministic > > was alread broken by memory starvation. > > > > I am still not convinced, specially if we are running under mem > cgroup. Even setting SCHED_FIFO does not help, you could have other > things like cpusets that might restrict the CPUs you can run on, or > any other policy and we could end up contending anyway with other > SCHED_FIFO tasks. Ah, right you are. I had missed mem-cgroup. But I think memcgroup also don't need following two boost. Can we get rid of it? p->rt.time_slice = HZ; set_tsk_thread_flag(p, TIF_MEMDIE); I mean we need distinguish global oom and memcg oom, perhapls. > > That's the reason I acked it. > > If we could show faster recovery from OOM or anything else, I would be > more convinced. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: email@kvack.org