From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail143.messagelabs.com (mail143.messagelabs.com [216.82.254.35]) by kanga.kvack.org (Postfix) with ESMTP id B8E9C6B00EC for ; Thu, 21 Jul 2011 20:26:10 -0400 (EDT) Received: from m2.gw.fujitsu.co.jp (unknown [10.0.50.72]) by fgwmail6.fujitsu.co.jp (Postfix) with ESMTP id 8919C3EE0B6 for ; Fri, 22 Jul 2011 09:26:07 +0900 (JST) Received: from smail (m2 [127.0.0.1]) by outgoing.m2.gw.fujitsu.co.jp (Postfix) with ESMTP id 458BD2E68C4 for ; Fri, 22 Jul 2011 09:26:07 +0900 (JST) Received: from s2.gw.fujitsu.co.jp (s2.gw.fujitsu.co.jp [10.0.50.92]) by m2.gw.fujitsu.co.jp (Postfix) with ESMTP id 9D1BC2E68CA for ; Fri, 22 Jul 2011 09:26:06 +0900 (JST) Received: from s2.gw.fujitsu.co.jp (localhost.localdomain [127.0.0.1]) by s2.gw.fujitsu.co.jp (Postfix) with ESMTP id 90BD61DB803C for ; Fri, 22 Jul 2011 09:26:06 +0900 (JST) Received: from ml13.s.css.fujitsu.com (ml13.s.css.fujitsu.com [10.240.81.133]) by s2.gw.fujitsu.co.jp (Postfix) with ESMTP id 5A9B41DB8040 for ; Fri, 22 Jul 2011 09:26:06 +0900 (JST) Date: Fri, 22 Jul 2011 09:18:58 +0900 From: KAMEZAWA Hiroyuki Subject: Re: [PATCH 4/4] memcg: prevent from reclaiming if there are per-cpu cached charges Message-Id: <20110722091858.f4b78e14.kamezawa.hiroyu@jp.fujitsu.com> In-Reply-To: <20110722085652.759aded2.kamezawa.hiroyu@jp.fujitsu.com> References: <0ed59a22cc84037d6e42b258981c75e3a6063899.1311241300.git.mhocko@suse.cz> <20110721195411.f4fa9f91.kamezawa.hiroyu@jp.fujitsu.com> <20110721123012.GD27855@tiehlicka.suse.cz> <20110722085652.759aded2.kamezawa.hiroyu@jp.fujitsu.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-mm@kvack.org List-ID: To: KAMEZAWA Hiroyuki Cc: Michal Hocko , linux-mm@kvack.org, Balbir Singh , Daisuke Nishimura , linux-kernel@vger.kernel.org On Fri, 22 Jul 2011 08:56:52 +0900 KAMEZAWA Hiroyuki wrote: > On Thu, 21 Jul 2011 14:30:12 +0200 > Michal Hocko wrote: > Please wait until "background reclaim" stuff. I don't stop it and it will > make this cpu-caching stuff better because we can drain before hitting > limit. > > If you cannot wait.... > > One idea is to have a threshold to call async "drain". For example, > > threshould = limit_of_memory - nr_online_cpu() * (BATCH_SIZE + 1) > > if (usage > threshould) > drain_all_stock_async(). > > Then, situation will be much better. > Of course, frequency of this call can be controlled by event counter. Thanks, -Kame -- 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/ . Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/ Don't email: email@kvack.org