From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755376AbaEHS2y (ORCPT ); Thu, 8 May 2014 14:28:54 -0400 Received: from zene.cmpxchg.org ([85.214.230.12]:60607 "EHLO zene.cmpxchg.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753497AbaEHS2w (ORCPT ); Thu, 8 May 2014 14:28:52 -0400 Date: Thu, 8 May 2014 14:28:48 -0400 From: Johannes Weiner To: Michal Hocko Cc: linux-mm@kvack.org, Hugh Dickins , Tejun Heo , cgroups@vger.kernel.org, linux-kernel@vger.kernel.org Subject: Re: [patch 3/9] mm: memcontrol: retry reclaim for oom-disabled and __GFP_NOFAIL charges Message-ID: <20140508182848.GP19914@cmpxchg.org> References: <1398889543-23671-1-git-send-email-hannes@cmpxchg.org> <1398889543-23671-4-git-send-email-hannes@cmpxchg.org> <20140507144339.GI9489@dhcp22.suse.cz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20140507144339.GI9489@dhcp22.suse.cz> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, May 07, 2014 at 04:43:39PM +0200, Michal Hocko wrote: > On Wed 30-04-14 16:25:37, Johannes Weiner wrote: > > There is no reason why oom-disabled and __GFP_NOFAIL charges should > > try to reclaim only once when every other charge tries several times > > before giving up. Make them all retry the same number of times. > > I guess the idea whas that oom disabled (THP) allocation can fallback to > a smaller allocation. I would suspect this would increase latency for > THP page faults. If it does, we should probably teach THP to use __GFP_NORETRY. On that note, __GFP_NORETRY is currently useless for memcg because it has !__GFP_WAIT semantics... I'll include a fix for that in v2.