From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932894AbeAXJ2x (ORCPT ); Wed, 24 Jan 2018 04:28:53 -0500 Received: from mx2.suse.de ([195.135.220.15]:41501 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932863AbeAXJ2u (ORCPT ); Wed, 24 Jan 2018 04:28:50 -0500 Date: Wed, 24 Jan 2018 10:28:47 +0100 From: Michal Hocko To: Michel =?iso-8859-1?Q?D=E4nzer?= Cc: Roman Gushchin , Andrey Grodzovsky , linux-kernel@vger.kernel.org, amd-gfx@lists.freedesktop.org, linux-mm@kvack.org, dri-devel@lists.freedesktop.org, Christian.Koenig@amd.com Subject: Re: [RFC] Per file OOM badness Message-ID: <20180124092847.GI1526@dhcp22.suse.cz> References: <1516294072-17841-1-git-send-email-andrey.grodzovsky@amd.com> <20180118170006.GG6584@dhcp22.suse.cz> <20180123152659.GA21817@castle.DHCP.thefacebook.com> <20180123153631.GR1526@dhcp22.suse.cz> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: User-Agent: Mutt/1.9.2 (2017-12-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue 23-01-18 17:39:19, Michel Dänzer wrote: > On 2018-01-23 04:36 PM, Michal Hocko wrote: > > On Tue 23-01-18 15:27:00, Roman Gushchin wrote: > >> On Thu, Jan 18, 2018 at 06:00:06PM +0100, Michal Hocko wrote: > >>> On Thu 18-01-18 11:47:48, Andrey Grodzovsky wrote: > >>>> Hi, this series is a revised version of an RFC sent by Christian König > >>>> a few years ago. The original RFC can be found at > >>>> https://urldefense.proofpoint.com/v2/url?u=https-3A__lists.freedesktop.org_archives_dri-2Ddevel_2015-2DSeptember_089778.html&d=DwIDAw&c=5VD0RTtNlTh3ycd41b3MUw&r=jJYgtDM7QT-W-Fz_d29HYQ&m=R-JIQjy8rqmH5qD581_VYL0Q7cpWSITKOnBCE-3LI8U&s=QZGqKpKuJ2BtioFGSy8_721owcWJ0J6c6d4jywOwN4w& > >>> Here is the origin cover letter text > >>> : I'm currently working on the issue that when device drivers allocate memory on > >>> : behalf of an application the OOM killer usually doesn't knew about that unless > >>> : the application also get this memory mapped into their address space. > >>> : > >>> : This is especially annoying for graphics drivers where a lot of the VRAM > >>> : usually isn't CPU accessible and so doesn't make sense to map into the > >>> : address space of the process using it. > >>> : > >>> : The problem now is that when an application starts to use a lot of VRAM those > >>> : buffers objects sooner or later get swapped out to system memory, but when we > >>> : now run into an out of memory situation the OOM killer obviously doesn't knew > >>> : anything about that memory and so usually kills the wrong process. > >>> : > >>> : The following set of patches tries to address this problem by introducing a per > >>> : file OOM badness score, which device drivers can use to give the OOM killer a > >>> : hint how many resources are bound to a file descriptor so that it can make > >>> : better decisions which process to kill. > >>> : > >>> : So question at every one: What do you think about this approach? > >>> : > >>> : My biggest concern right now is the patches are messing with a core kernel > >>> : structure (adding a field to struct file). Any better idea? I'm considering > >>> : to put a callback into file_ops instead. > >> > >> Hello! > >> > >> I wonder if groupoom (aka cgroup-aware OOM killer) can work for you? > > > > I do not think so. The problem is that the allocating context is not > > identical with the end consumer. > > That's actually not really true. Even in cases where a BO is shared with > a different process, it is still used at least occasionally in the > process which allocated it as well. Otherwise there would be no point in > sharing it between processes. OK, but somebody has to be made responsible. Otherwise you are just killing a process which doesn't really release any memory. > There should be no problem if the memory of a shared BO is accounted for > in each process sharing it. It might be nice to scale each process' > "debt" by 1 / (number of processes sharing it) if possible, but in the > worst case accounting it fully in each process should be fine. So how exactly then helps to kill one of those processes? The memory stays pinned behind or do I still misunderstand? -- Michal Hocko SUSE Labs From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-wm0-f72.google.com (mail-wm0-f72.google.com [74.125.82.72]) by kanga.kvack.org (Postfix) with ESMTP id 303FE800DD for ; Wed, 24 Jan 2018 04:28:50 -0500 (EST) Received: by mail-wm0-f72.google.com with SMTP id v14so1904431wmd.3 for ; Wed, 24 Jan 2018 01:28:50 -0800 (PST) Received: from mx2.suse.de (mx2.suse.de. [195.135.220.15]) by mx.google.com with ESMTPS id j27si1704855wre.25.2018.01.24.01.28.48 for (version=TLS1 cipher=AES128-SHA bits=128/128); Wed, 24 Jan 2018 01:28:49 -0800 (PST) Date: Wed, 24 Jan 2018 10:28:47 +0100 From: Michal Hocko Subject: Re: [RFC] Per file OOM badness Message-ID: <20180124092847.GI1526@dhcp22.suse.cz> References: <1516294072-17841-1-git-send-email-andrey.grodzovsky@amd.com> <20180118170006.GG6584@dhcp22.suse.cz> <20180123152659.GA21817@castle.DHCP.thefacebook.com> <20180123153631.GR1526@dhcp22.suse.cz> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: Sender: owner-linux-mm@kvack.org List-ID: To: Michel =?iso-8859-1?Q?D=E4nzer?= Cc: Roman Gushchin , Andrey Grodzovsky , linux-kernel@vger.kernel.org, amd-gfx@lists.freedesktop.org, linux-mm@kvack.org, dri-devel@lists.freedesktop.org, Christian.Koenig@amd.com On Tue 23-01-18 17:39:19, Michel Danzer wrote: > On 2018-01-23 04:36 PM, Michal Hocko wrote: > > On Tue 23-01-18 15:27:00, Roman Gushchin wrote: > >> On Thu, Jan 18, 2018 at 06:00:06PM +0100, Michal Hocko wrote: > >>> On Thu 18-01-18 11:47:48, Andrey Grodzovsky wrote: > >>>> Hi, this series is a revised version of an RFC sent by Christian Konig > >>>> a few years ago. The original RFC can be found at > >>>> https://urldefense.proofpoint.com/v2/url?u=https-3A__lists.freedesktop.org_archives_dri-2Ddevel_2015-2DSeptember_089778.html&d=DwIDAw&c=5VD0RTtNlTh3ycd41b3MUw&r=jJYgtDM7QT-W-Fz_d29HYQ&m=R-JIQjy8rqmH5qD581_VYL0Q7cpWSITKOnBCE-3LI8U&s=QZGqKpKuJ2BtioFGSy8_721owcWJ0J6c6d4jywOwN4w& > >>> Here is the origin cover letter text > >>> : I'm currently working on the issue that when device drivers allocate memory on > >>> : behalf of an application the OOM killer usually doesn't knew about that unless > >>> : the application also get this memory mapped into their address space. > >>> : > >>> : This is especially annoying for graphics drivers where a lot of the VRAM > >>> : usually isn't CPU accessible and so doesn't make sense to map into the > >>> : address space of the process using it. > >>> : > >>> : The problem now is that when an application starts to use a lot of VRAM those > >>> : buffers objects sooner or later get swapped out to system memory, but when we > >>> : now run into an out of memory situation the OOM killer obviously doesn't knew > >>> : anything about that memory and so usually kills the wrong process. > >>> : > >>> : The following set of patches tries to address this problem by introducing a per > >>> : file OOM badness score, which device drivers can use to give the OOM killer a > >>> : hint how many resources are bound to a file descriptor so that it can make > >>> : better decisions which process to kill. > >>> : > >>> : So question at every one: What do you think about this approach? > >>> : > >>> : My biggest concern right now is the patches are messing with a core kernel > >>> : structure (adding a field to struct file). Any better idea? I'm considering > >>> : to put a callback into file_ops instead. > >> > >> Hello! > >> > >> I wonder if groupoom (aka cgroup-aware OOM killer) can work for you? > > > > I do not think so. The problem is that the allocating context is not > > identical with the end consumer. > > That's actually not really true. Even in cases where a BO is shared with > a different process, it is still used at least occasionally in the > process which allocated it as well. Otherwise there would be no point in > sharing it between processes. OK, but somebody has to be made responsible. Otherwise you are just killing a process which doesn't really release any memory. > There should be no problem if the memory of a shared BO is accounted for > in each process sharing it. It might be nice to scale each process' > "debt" by 1 / (number of processes sharing it) if possible, but in the > worst case accounting it fully in each process should be fine. So how exactly then helps to kill one of those processes? The memory stays pinned behind or do I still misunderstand? -- Michal Hocko SUSE Labs -- 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 From mboxrd@z Thu Jan 1 00:00:00 1970 From: Michal Hocko Subject: Re: [RFC] Per file OOM badness Date: Wed, 24 Jan 2018 10:28:47 +0100 Message-ID: <20180124092847.GI1526@dhcp22.suse.cz> References: <1516294072-17841-1-git-send-email-andrey.grodzovsky@amd.com> <20180118170006.GG6584@dhcp22.suse.cz> <20180123152659.GA21817@castle.DHCP.thefacebook.com> <20180123153631.GR1526@dhcp22.suse.cz> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Return-path: Content-Disposition: inline In-Reply-To: Sender: owner-linux-mm@kvack.org To: Michel =?iso-8859-1?Q?D=E4nzer?= Cc: Roman Gushchin , Andrey Grodzovsky , linux-kernel@vger.kernel.org, amd-gfx@lists.freedesktop.org, linux-mm@kvack.org, dri-devel@lists.freedesktop.org, Christian.Koenig@amd.com List-Id: dri-devel@lists.freedesktop.org On Tue 23-01-18 17:39:19, Michel Dänzer wrote: > On 2018-01-23 04:36 PM, Michal Hocko wrote: > > On Tue 23-01-18 15:27:00, Roman Gushchin wrote: > >> On Thu, Jan 18, 2018 at 06:00:06PM +0100, Michal Hocko wrote: > >>> On Thu 18-01-18 11:47:48, Andrey Grodzovsky wrote: > >>>> Hi, this series is a revised version of an RFC sent by Christian König > >>>> a few years ago. The original RFC can be found at > >>>> https://urldefense.proofpoint.com/v2/url?u=https-3A__lists.freedesktop.org_archives_dri-2Ddevel_2015-2DSeptember_089778.html&d=DwIDAw&c=5VD0RTtNlTh3ycd41b3MUw&r=jJYgtDM7QT-W-Fz_d29HYQ&m=R-JIQjy8rqmH5qD581_VYL0Q7cpWSITKOnBCE-3LI8U&s=QZGqKpKuJ2BtioFGSy8_721owcWJ0J6c6d4jywOwN4w& > >>> Here is the origin cover letter text > >>> : I'm currently working on the issue that when device drivers allocate memory on > >>> : behalf of an application the OOM killer usually doesn't knew about that unless > >>> : the application also get this memory mapped into their address space. > >>> : > >>> : This is especially annoying for graphics drivers where a lot of the VRAM > >>> : usually isn't CPU accessible and so doesn't make sense to map into the > >>> : address space of the process using it. > >>> : > >>> : The problem now is that when an application starts to use a lot of VRAM those > >>> : buffers objects sooner or later get swapped out to system memory, but when we > >>> : now run into an out of memory situation the OOM killer obviously doesn't knew > >>> : anything about that memory and so usually kills the wrong process. > >>> : > >>> : The following set of patches tries to address this problem by introducing a per > >>> : file OOM badness score, which device drivers can use to give the OOM killer a > >>> : hint how many resources are bound to a file descriptor so that it can make > >>> : better decisions which process to kill. > >>> : > >>> : So question at every one: What do you think about this approach? > >>> : > >>> : My biggest concern right now is the patches are messing with a core kernel > >>> : structure (adding a field to struct file). Any better idea? I'm considering > >>> : to put a callback into file_ops instead. > >> > >> Hello! > >> > >> I wonder if groupoom (aka cgroup-aware OOM killer) can work for you? > > > > I do not think so. The problem is that the allocating context is not > > identical with the end consumer. > > That's actually not really true. Even in cases where a BO is shared with > a different process, it is still used at least occasionally in the > process which allocated it as well. Otherwise there would be no point in > sharing it between processes. OK, but somebody has to be made responsible. Otherwise you are just killing a process which doesn't really release any memory. > There should be no problem if the memory of a shared BO is accounted for > in each process sharing it. It might be nice to scale each process' > "debt" by 1 / (number of processes sharing it) if possible, but in the > worst case accounting it fully in each process should be fine. So how exactly then helps to kill one of those processes? The memory stays pinned behind or do I still misunderstand? -- Michal Hocko SUSE Labs -- 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