Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755833Ab1EXJkl (ORCPT ); Tue, 24 May 2011 05:40:41 -0400 Received: from fgwmail6.fujitsu.co.jp ([192.51.44.36]:55229 "EHLO fgwmail6.fujitsu.co.jp" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755521Ab1EXJkk (ORCPT ); Tue, 24 May 2011 05:40:40 -0400 X-SecurityPolicyCheck-FJ: OK by FujitsuOutboundMailChecker v1.3.1 Message-ID: <4DDB7D0F.3060204@jp.fujitsu.com> Date: Tue, 24 May 2011 18:40:31 +0900 From: KOSAKI Motohiro User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; ja; rv:1.9.2.17) Gecko/20110414 Lightning/1.0b2 Thunderbird/3.1.10 MIME-Version: 1.0 To: mgorman@suse.de CC: minchan.kim@gmail.com, abarry@cray.com, akpm@linux-foundation.org, linux-mm@kvack.org, riel@redhat.com, hannes@cmpxchg.org, linux-kernel@vger.kernel.org Subject: Re: Unending loop in __alloc_pages_slowpath following OOM-kill; rfc: patch. References: <4DCDA347.9080207@cray.com> <4DD2991B.5040707@cray.com> <20110520164924.GB2386@barrios-desktop> <4DDB3A1E.6090206@jp.fujitsu.com> <20110524083008.GA5279@suse.de> <4DDB6DF6.2050700@jp.fujitsu.com> <20110524084915.GC5279@suse.de> <4DDB74F7.9020109@jp.fujitsu.com> <20110524091611.GD5279@suse.de> In-Reply-To: <20110524091611.GD5279@suse.de> Content-Type: text/plain; charset=ISO-8859-15 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1463 Lines: 36 (2011/05/24 18:16), Mel Gorman wrote: > On Tue, May 24, 2011 at 06:05:59PM +0900, KOSAKI Motohiro wrote: >>>>> Why? >>>> >>>> Otherwise, we don't have good PCP dropping trigger. Big machine might have >>>> big pcp cache. >>>> >>> >>> Big machines also have a large cost for sending IPIs. >> >> Yes. But it's only matter if IPIs are frequently happen. >> But, drain_all_pages() is NOT only IPI source. some vmscan function (e.g. >> try_to_umap) makes a lot of IPIs. >> >> Then, it's _relatively_ not costly. I have a question. Do you compare which >> operation and drain_all_pages()? IOW, your "costly" mean which scenario suspect? >> > > I am concerned that if the machine gets into trouble and we are failing > to reclaim that sending more IPIs is not going to help any. There is no > evidence at the moment that sending extra IPIs here will help anything. In old days, we always call drain_all_pages() if did_some_progress!=0. But current kernel only call it when get_page_from_freelist() fail. So, wait_iff_congested() may help but no guarantee to help us. If you still strongly worry about IPI cost, I'm concern to move drain_all_pages() to more unfrequently point. but to ignore pcp makes less sense, IMHO. -- 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/