Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753460AbdCBTO0 (ORCPT ); Thu, 2 Mar 2017 14:14:26 -0500 Received: from mx2.suse.de ([195.135.220.15]:48156 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752200AbdCBTNm (ORCPT ); Thu, 2 Mar 2017 14:13:42 -0500 Date: Thu, 2 Mar 2017 16:47:44 +0100 From: Michal Hocko To: Brian Foster Cc: Christoph Hellwig , Tetsuo Handa , Xiong Zhou , linux-xfs@vger.kernel.org, linux-mm@kvack.org, linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org Subject: Re: mm allocation failure and hang when running xfstests generic/269 on xfs Message-ID: <20170302154744.GN1404@dhcp22.suse.cz> References: <20170302124909.GE1404@dhcp22.suse.cz> <20170302130009.GC3213@bfoster.bfoster> <20170302132755.GG1404@dhcp22.suse.cz> <20170302134157.GD3213@bfoster.bfoster> <20170302135001.GI1404@dhcp22.suse.cz> <20170302142315.GE3213@bfoster.bfoster> <20170302143441.GL1404@dhcp22.suse.cz> <20170302145131.GF3213@bfoster.bfoster> <20170302151411.GM1404@dhcp22.suse.cz> <20170302153002.GG3213@bfoster.bfoster> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20170302153002.GG3213@bfoster.bfoster> User-Agent: Mutt/1.5.23 (2014-03-12) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1051 Lines: 27 On Thu 02-03-17 10:30:02, Brian Foster wrote: > On Thu, Mar 02, 2017 at 04:14:11PM +0100, Michal Hocko wrote: [...] > > I am not objecting to adding fatal_signal_pending as well I just thought > > that from the logic POV breaking after reaching the minimum size is just > > the right thing to do. We can optimize further by checking > > fatal_signal_pending and reducing retries when we know it doesn't make > > much sense but that should be done on top as an optimization IMHO. > > > > I don't think of it as an optimization to not invoke calls (a > non-deterministic number of times) that we know are going to fail, but the point is that vmalloc failure modes are an implementation detail which might change in the future. The fix should be really independent on the current implementation that is why I think the fatal_signal_pending is just an optimization. > ultimately I don't care too much how it's framed or if it's done in > separate patches or whatnot. As long as they are posted at the same > time. ;) Done -- Michal Hocko SUSE Labs