From: Glauber Costa <glommer@parallels.com> To: David Rientjes <rientjes@google.com> Cc: <cgroups@vger.kernel.org>, <linux-mm@kvack.org>, Andrew Morton <akpm@linux-foundation.org>, <linux-kernel@vger.kernel.org>, "Frederic Weisbecker" <fweisbec@gmail.com>, Pekka Enberg <penberg@kernel.org>, "Michal Hocko" <mhocko@suse.cz>, Johannes Weiner <hannes@cmpxchg.org>, "Christoph Lameter" <cl@linux.com>, <devel@openvz.org>, <kamezawa.hiroyu@jp.fujitsu.com>, Tejun Heo <tj@kernel.org>, Suleiman Souhlal <suleiman@google.com> Subject: Re: [PATCH 02/11] memcg: Reclaim when more than one page needed. Date: Tue, 26 Jun 2012 13:23:40 +0400 [thread overview] Message-ID: <4FE97F9C.5030907@parallels.com> (raw) In-Reply-To: <alpine.DEB.2.00.1206260214310.16020@chino.kir.corp.google.com> On 06/26/2012 01:17 PM, David Rientjes wrote: > On Tue, 26 Jun 2012, Glauber Costa wrote: > >>> Nope, have you checked the output of /sys/kernel/slab/.../order when >>> running slub? On my workstation 127 out of 316 caches have order-2 or >>> higher by default. >>> >> >> Well, this is still on the side of my argument, since this is still a majority >> of them being low ordered. > > Ok, so what happens if I pass slub_min_order=2 on the command line? We > never retry? Well, indeed. The function has many other RETRY points, but I believe we'd reach none of them without triggering oom first. >> The code here does not necessarily have to retry - >> if I understand it correctly - we just retry for very small allocations >> because that is where our likelihood of succeeding is. >> > > Well, the comment for NR_PAGES_TO_RETRY says > > /* > * We need a number that is small enough to be likely to have been > * reclaimed even under pressure, but not too big to trigger unnecessary > * retries > */ > > and mmzone.h says > > /* > * PAGE_ALLOC_COSTLY_ORDER is the order at which allocations are deemed > * costly to service. That is between allocation orders which should > * coalesce naturally under reasonable reclaim pressure and those which > * will not. > */ > #define PAGE_ALLOC_COSTLY_ORDER 3 > > so I'm trying to reconcile which one is correct. > I am not myself against reverting back to costly order. The check we have here, of course, is stricter than costly order, so there is nothing to reconcile. Now if we really need a stricter check or not, is the subject of discussion.
WARNING: multiple messages have this Message-ID (diff)
From: Glauber Costa <glommer@parallels.com> To: David Rientjes <rientjes@google.com> Cc: cgroups@vger.kernel.org, linux-mm@kvack.org, Andrew Morton <akpm@linux-foundation.org>, linux-kernel@vger.kernel.org, Frederic Weisbecker <fweisbec@gmail.com>, Pekka Enberg <penberg@kernel.org>, Michal Hocko <mhocko@suse.cz>, Johannes Weiner <hannes@cmpxchg.org>, Christoph Lameter <cl@linux.com>, devel@openvz.org, kamezawa.hiroyu@jp.fujitsu.com, Tejun Heo <tj@kernel.org>, Suleiman Souhlal <suleiman@google.com> Subject: Re: [PATCH 02/11] memcg: Reclaim when more than one page needed. Date: Tue, 26 Jun 2012 13:23:40 +0400 [thread overview] Message-ID: <4FE97F9C.5030907@parallels.com> (raw) In-Reply-To: <alpine.DEB.2.00.1206260214310.16020@chino.kir.corp.google.com> On 06/26/2012 01:17 PM, David Rientjes wrote: > On Tue, 26 Jun 2012, Glauber Costa wrote: > >>> Nope, have you checked the output of /sys/kernel/slab/.../order when >>> running slub? On my workstation 127 out of 316 caches have order-2 or >>> higher by default. >>> >> >> Well, this is still on the side of my argument, since this is still a majority >> of them being low ordered. > > Ok, so what happens if I pass slub_min_order=2 on the command line? We > never retry? Well, indeed. The function has many other RETRY points, but I believe we'd reach none of them without triggering oom first. >> The code here does not necessarily have to retry - >> if I understand it correctly - we just retry for very small allocations >> because that is where our likelihood of succeeding is. >> > > Well, the comment for NR_PAGES_TO_RETRY says > > /* > * We need a number that is small enough to be likely to have been > * reclaimed even under pressure, but not too big to trigger unnecessary > * retries > */ > > and mmzone.h says > > /* > * PAGE_ALLOC_COSTLY_ORDER is the order at which allocations are deemed > * costly to service. That is between allocation orders which should > * coalesce naturally under reasonable reclaim pressure and those which > * will not. > */ > #define PAGE_ALLOC_COSTLY_ORDER 3 > > so I'm trying to reconcile which one is correct. > I am not myself against reverting back to costly order. The check we have here, of course, is stricter than costly order, so there is nothing to reconcile. Now if we really need a stricter check or not, is the subject of discussion. -- 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: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2012-06-26 9:26 UTC|newest] Thread overview: 238+ messages / expand[flat|nested] mbox.gz Atom feed top 2012-06-25 14:15 [PATCH 00/11] kmem controller for memcg: stripped down version Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-25 14:15 ` [PATCH 01/11] memcg: Make it possible to use the stock for more than one page Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-25 17:44 ` Tejun Heo 2012-06-25 17:44 ` Tejun Heo 2012-06-25 17:44 ` Tejun Heo 2012-06-25 22:29 ` Glauber Costa 2012-06-25 22:29 ` Glauber Costa 2012-06-25 22:33 ` Tejun Heo 2012-06-25 22:33 ` Tejun Heo 2012-06-26 4:01 ` David Rientjes 2012-06-26 4:01 ` David Rientjes 2012-06-25 14:15 ` [PATCH 02/11] memcg: Reclaim when more than one page needed Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-25 23:33 ` Suleiman Souhlal 2012-06-25 23:33 ` Suleiman Souhlal 2012-06-26 8:39 ` Glauber Costa 2012-06-26 8:39 ` Glauber Costa 2012-06-27 16:16 ` Suleiman Souhlal 2012-06-27 16:16 ` Suleiman Souhlal 2012-06-27 16:16 ` Suleiman Souhlal 2012-06-26 4:09 ` David Rientjes 2012-06-26 4:09 ` David Rientjes 2012-06-26 7:12 ` Glauber Costa 2012-06-26 7:12 ` Glauber Costa 2012-06-26 7:12 ` Glauber Costa 2012-06-26 8:54 ` David Rientjes 2012-06-26 8:54 ` David Rientjes 2012-06-26 8:54 ` David Rientjes 2012-06-26 9:08 ` Glauber Costa 2012-06-26 9:08 ` Glauber Costa 2012-06-26 9:17 ` David Rientjes 2012-06-26 9:17 ` David Rientjes 2012-06-26 9:23 ` Glauber Costa [this message] 2012-06-26 9:23 ` Glauber Costa 2012-06-27 10:03 ` Glauber Costa 2012-06-27 10:03 ` Glauber Costa 2012-06-27 19:48 ` David Rientjes 2012-06-27 19:48 ` David Rientjes 2012-06-27 19:48 ` David Rientjes 2012-06-27 20:47 ` Glauber Costa 2012-06-27 20:47 ` Glauber Costa 2012-06-25 14:15 ` [PATCH 03/11] memcg: change defines to an enum Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-26 4:11 ` David Rientjes 2012-06-26 4:11 ` David Rientjes 2012-06-26 8:28 ` Glauber Costa 2012-06-26 8:28 ` Glauber Costa 2012-06-26 9:01 ` David Rientjes 2012-06-26 9:01 ` David Rientjes 2012-06-26 9:01 ` David Rientjes 2012-06-25 14:15 ` [PATCH 04/11] kmem slab accounting basic infrastructure Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-26 4:22 ` David Rientjes 2012-06-26 4:22 ` David Rientjes 2012-06-26 7:09 ` Glauber Costa 2012-06-26 7:09 ` Glauber Costa 2012-06-25 14:15 ` [PATCH 05/11] Add a __GFP_KMEMCG flag Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-26 4:25 ` David Rientjes 2012-06-26 4:25 ` David Rientjes 2012-06-26 7:08 ` Glauber Costa 2012-06-26 7:08 ` Glauber Costa 2012-06-26 9:03 ` David Rientjes 2012-06-26 9:03 ` David Rientjes 2012-06-25 14:15 ` [PATCH 06/11] memcg: kmem controller infrastructure Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-25 18:06 ` Tejun Heo 2012-06-25 18:06 ` Tejun Heo 2012-06-25 18:06 ` Tejun Heo 2012-06-25 22:28 ` Glauber Costa 2012-06-25 22:28 ` Glauber Costa 2012-06-25 23:17 ` Andrew Morton 2012-06-25 23:17 ` Andrew Morton 2012-06-25 23:17 ` Andrew Morton 2012-06-26 14:40 ` Glauber Costa 2012-06-26 14:40 ` Glauber Costa 2012-06-26 15:01 ` Glauber Costa 2012-06-26 15:01 ` Glauber Costa 2012-06-26 15:01 ` Glauber Costa 2012-06-26 18:01 ` Andrew Morton 2012-06-26 18:01 ` Andrew Morton 2012-06-26 18:01 ` Andrew Morton 2012-06-26 18:08 ` Tejun Heo 2012-06-26 18:08 ` Tejun Heo 2012-06-26 18:08 ` Tejun Heo 2012-06-26 18:14 ` Glauber Costa 2012-06-26 18:14 ` Glauber Costa 2012-06-26 19:20 ` Andrew Morton 2012-06-26 19:20 ` Andrew Morton 2012-06-26 15:29 ` Glauber Costa 2012-06-26 15:29 ` Glauber Costa 2012-06-26 15:29 ` Glauber Costa 2012-06-26 9:12 ` David Rientjes 2012-06-26 9:12 ` David Rientjes 2012-06-26 9:12 ` David Rientjes 2012-06-26 9:17 ` Glauber Costa 2012-06-26 9:17 ` Glauber Costa 2012-06-26 9:17 ` Glauber Costa 2012-06-27 4:01 ` David Rientjes 2012-06-27 4:01 ` David Rientjes 2012-06-27 9:33 ` Glauber Costa 2012-06-27 9:33 ` Glauber Costa 2012-06-27 9:33 ` Glauber Costa 2012-06-27 19:46 ` David Rientjes 2012-06-27 19:46 ` David Rientjes 2012-06-25 14:15 ` [PATCH 07/11] mm: Allocate kernel pages to the right memcg Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-25 18:07 ` Tejun Heo 2012-06-25 18:07 ` Tejun Heo 2012-06-25 22:27 ` Glauber Costa 2012-06-25 22:27 ` Glauber Costa 2012-06-25 14:15 ` [PATCH 08/11] memcg: disable kmem code when not in use Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-26 5:51 ` Kamezawa Hiroyuki 2012-06-26 5:51 ` Kamezawa Hiroyuki 2012-06-25 14:15 ` [PATCH 09/11] memcg: propagate kmem limiting information to children Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-25 18:29 ` Tejun Heo 2012-06-25 18:29 ` Tejun Heo 2012-06-25 22:36 ` Glauber Costa 2012-06-25 22:36 ` Glauber Costa 2012-06-25 22:36 ` Glauber Costa 2012-06-25 22:49 ` Tejun Heo 2012-06-25 22:49 ` Tejun Heo 2012-06-25 22:49 ` Tejun Heo 2012-06-25 23:21 ` Andrew Morton 2012-06-25 23:21 ` Andrew Morton 2012-06-26 5:23 ` David Rientjes 2012-06-26 5:23 ` David Rientjes 2012-06-25 23:23 ` Andrew Morton 2012-06-25 23:23 ` Andrew Morton 2012-06-26 5:24 ` David Rientjes 2012-06-26 5:24 ` David Rientjes 2012-06-26 5:31 ` Andrew Morton 2012-06-26 5:31 ` Andrew Morton 2012-06-26 5:31 ` Andrew Morton 2012-06-26 7:23 ` Glauber Costa 2012-06-26 7:23 ` Glauber Costa 2012-06-26 7:23 ` Glauber Costa 2012-06-25 14:15 ` [PATCH 10/11] memcg: allow a memcg with kmem charges to be destructed Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-25 18:34 ` Tejun Heo 2012-06-25 18:34 ` Tejun Heo 2012-06-25 22:25 ` Glauber Costa 2012-06-25 22:25 ` Glauber Costa 2012-06-25 22:25 ` Glauber Costa 2012-06-26 5:59 ` Kamezawa Hiroyuki 2012-06-26 5:59 ` Kamezawa Hiroyuki 2012-06-26 7:21 ` Glauber Costa 2012-06-26 7:21 ` Glauber Costa 2012-06-26 7:21 ` Glauber Costa 2012-06-25 14:15 ` [PATCH 11/11] protect architectures where THREAD_SIZE >= PAGE_SIZE against fork bombs Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-25 14:15 ` Glauber Costa 2012-06-25 16:55 ` Frederic Weisbecker 2012-06-25 16:55 ` Frederic Weisbecker 2012-06-25 18:38 ` Tejun Heo 2012-06-25 18:38 ` Tejun Heo 2012-06-25 18:38 ` Tejun Heo 2012-06-25 20:57 ` Frederic Weisbecker 2012-06-25 20:57 ` Frederic Weisbecker 2012-06-25 20:57 ` Frederic Weisbecker 2012-06-26 12:48 ` Glauber Costa 2012-06-26 12:48 ` Glauber Costa 2012-06-26 12:48 ` Glauber Costa 2012-06-26 13:38 ` Frederic Weisbecker 2012-06-26 13:38 ` Frederic Weisbecker 2012-06-26 13:38 ` Frederic Weisbecker 2012-06-26 13:37 ` Glauber Costa 2012-06-26 13:37 ` Glauber Costa 2012-06-26 13:37 ` Glauber Costa 2012-06-26 13:44 ` Frederic Weisbecker 2012-06-26 13:44 ` Frederic Weisbecker 2012-06-26 13:44 ` Frederic Weisbecker 2012-06-26 4:57 ` David Rientjes 2012-06-26 4:57 ` David Rientjes 2012-06-26 5:35 ` Kamezawa Hiroyuki 2012-06-26 5:35 ` Kamezawa Hiroyuki 2012-06-26 5:35 ` Kamezawa Hiroyuki 2012-06-26 7:23 ` Glauber Costa 2012-06-26 7:23 ` Glauber Costa 2012-06-26 7:23 ` Glauber Costa 2012-06-26 8:45 ` David Rientjes 2012-06-26 8:45 ` David Rientjes 2012-06-26 8:45 ` David Rientjes 2012-06-26 8:44 ` Glauber Costa 2012-06-26 8:44 ` Glauber Costa 2012-06-26 8:44 ` Glauber Costa 2012-06-26 9:05 ` David Rientjes 2012-06-26 9:05 ` David Rientjes 2012-06-25 23:27 ` [PATCH 00/11] kmem controller for memcg: stripped down version Andrew Morton 2012-06-25 23:27 ` Andrew Morton 2012-06-26 7:17 ` Glauber Costa 2012-06-26 7:17 ` Glauber Costa 2012-06-26 21:55 ` Andrew Morton 2012-06-26 21:55 ` Andrew Morton 2012-06-26 21:55 ` Andrew Morton 2012-06-27 1:08 ` David Rientjes 2012-06-27 1:08 ` David Rientjes 2012-06-27 8:39 ` Glauber Costa 2012-06-27 8:39 ` Glauber Costa 2012-06-27 8:39 ` Glauber Costa 2012-06-27 9:29 ` Fork bomb limitation in memcg WAS: " Glauber Costa 2012-06-27 9:29 ` Glauber Costa 2012-06-27 9:29 ` Glauber Costa 2012-06-27 12:29 ` Frederic Weisbecker 2012-06-27 12:29 ` Frederic Weisbecker 2012-06-27 12:28 ` Glauber Costa 2012-06-27 12:28 ` Glauber Costa 2012-06-27 12:35 ` Frederic Weisbecker 2012-06-27 12:35 ` Frederic Weisbecker 2012-06-27 19:38 ` David Rientjes 2012-06-27 19:38 ` David Rientjes 2012-06-27 19:38 ` David Rientjes 2012-06-28 9:01 ` Glauber Costa 2012-06-28 9:01 ` Glauber Costa 2012-06-28 22:25 ` Andrew Morton 2012-06-28 22:25 ` Andrew Morton 2012-07-03 11:38 ` Glauber Costa 2012-07-03 11:38 ` Glauber Costa 2012-07-03 11:38 ` Glauber Costa 2012-07-12 15:40 ` Frederic Weisbecker 2012-07-12 15:40 ` Frederic Weisbecker 2012-08-07 13:59 ` Glauber Costa 2012-08-07 13:59 ` Glauber Costa 2012-08-07 13:59 ` Glauber Costa 2012-08-08 14:15 ` Glauber Costa 2012-08-08 14:15 ` Glauber Costa 2012-08-08 14:15 ` Glauber Costa
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=4FE97F9C.5030907@parallels.com \ --to=glommer@parallels.com \ --cc=akpm@linux-foundation.org \ --cc=cgroups@vger.kernel.org \ --cc=cl@linux.com \ --cc=devel@openvz.org \ --cc=fweisbec@gmail.com \ --cc=hannes@cmpxchg.org \ --cc=kamezawa.hiroyu@jp.fujitsu.com \ --cc=linux-kernel@vger.kernel.org \ --cc=linux-mm@kvack.org \ --cc=mhocko@suse.cz \ --cc=penberg@kernel.org \ --cc=rientjes@google.com \ --cc=suleiman@google.com \ --cc=tj@kernel.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.