From: Glauber Costa <glommer@parallels.com>
To: Michal Hocko <mhocko@suse.cz>
Cc: David Rientjes <rientjes@google.com>, <linux-mm@kvack.org>,
<cgroups@vger.kernel.org>, Mel Gorman <mgorman@suse.de>,
Tejun Heo <tj@kernel.org>,
Andrew Morton <akpm@linux-foundation.org>,
Johannes Weiner <hannes@cmpxchg.org>,
KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>,
Christoph Lameter <cl@linux.com>,
Pekka Enberg <penberg@kernel.org>, <devel@openvz.org>,
<linux-kernel@vger.kernel.org>,
Pekka Enberg <penberg@cs.helsinki.fi>
Subject: Re: [PATCH v5 06/14] memcg: kmem controller infrastructure
Date: Mon, 22 Oct 2012 16:52:41 +0400 [thread overview]
Message-ID: <50854199.9020804@parallels.com> (raw)
In-Reply-To: <20121022125124.GA8344@dhcp22.suse.cz>
On 10/22/2012 04:51 PM, Michal Hocko wrote:
> [Sorry for the late reply]
>
> On Mon 22-10-12 16:34:15, Glauber Costa wrote:
>> On 10/20/2012 12:34 AM, David Rientjes wrote:
>>> On Fri, 19 Oct 2012, Glauber Costa wrote:
>>>
>>>>>>> What about gfp & __GFP_FS?
>>>>>>>
>>>>>>
>>>>>> Do you intend to prevent or allow OOM under that flag? I personally
>>>>>> think that anything that accepts to be OOM-killed should have GFP_WAIT
>>>>>> set, so that ought to be enough.
>>>>>>
>>>>>
>>>>> The oom killer in the page allocator cannot trigger without __GFP_FS
>>>>> because direct reclaim has little chance of being very successful and
>>>>> thus we end up needlessly killing processes, and that tends to happen
>>>>> quite a bit if we dont check for it. Seems like this would also happen
>>>>> with memcg if mem_cgroup_reclaim() has a large probability of failing?
>>>>>
>>>>
>>>> I can indeed see tests for GFP_FS in some key locations in mm/ before
>>>> calling the OOM Killer.
>>>>
>>>> Should I test for GFP_IO as well?
>>>
>>> It's not really necessary, if __GFP_IO isn't set then it wouldn't make
>>> sense for __GFP_FS to be set.
>>>
>>>> If the idea is preventing OOM to
>>>> trigger for allocations that can write their pages back, how would you
>>>> feel about the following test:
>>>>
>>>> may_oom = (gfp & GFP_KERNEL) && !(gfp & __GFP_NORETRY) ?
>>>>
>>>
>>> I would simply copy the logic from the page allocator and only trigger oom
>>> for __GFP_FS and !__GFP_NORETRY.
>>>
>>
>> That seems reasonable to me. Michal ?
>
> Yes it makes sense to be consistent with the global case. While we are
> at it, do we need to consider PF_DUMPCORE resp. !__GFP_NOFAIL?
>
at least from kmem, GFP_NOFAIL will not reach this codepath. We will
ditch it to the root in memcontrol.h
next prev parent reply other threads:[~2012-10-22 12:53 UTC|newest]
Thread overview: 70+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-10-16 10:16 [PATCH v5 00/14] kmem controller for memcg Glauber Costa
2012-10-16 10:16 ` [PATCH v5 01/14] memcg: Make it possible to use the stock for more than one page Glauber Costa
2012-10-17 22:11 ` Andrew Morton
2012-10-18 16:54 ` Glauber Costa
2012-10-16 10:16 ` [PATCH v5 02/14] memcg: Reclaim when more than one page needed Glauber Costa
2012-10-17 21:46 ` David Rientjes
2012-10-16 10:16 ` [PATCH v5 03/14] memcg: change defines to an enum Glauber Costa
2012-10-17 21:50 ` David Rientjes
2012-10-16 10:16 ` [PATCH v5 04/14] kmem accounting basic infrastructure Glauber Costa
2012-10-16 12:14 ` Michal Hocko
2012-10-17 22:08 ` David Rientjes
2012-10-18 17:01 ` Glauber Costa
2012-10-18 19:47 ` Tejun Heo
2012-10-18 19:37 ` Tejun Heo
2012-10-17 22:12 ` Andrew Morton
2012-10-18 17:03 ` Glauber Costa
2012-10-16 10:16 ` [PATCH v5 05/14] Add a __GFP_KMEMCG flag Glauber Costa
2012-10-16 12:15 ` Michal Hocko
2012-10-17 22:09 ` David Rientjes
2012-10-16 10:16 ` [PATCH v5 06/14] memcg: kmem controller infrastructure Glauber Costa
2012-10-17 6:40 ` Kamezawa Hiroyuki
2012-10-17 22:12 ` Andrew Morton
2012-10-18 9:16 ` Glauber Costa
2012-10-18 22:06 ` David Rientjes
2012-10-19 9:10 ` Glauber Costa
2012-10-19 9:31 ` David Rientjes
2012-10-19 10:00 ` Glauber Costa
2012-10-17 22:37 ` David Rientjes
2012-10-18 9:23 ` Glauber Costa
2012-10-18 21:59 ` David Rientjes
2012-10-19 10:08 ` Glauber Costa
2012-10-19 20:34 ` David Rientjes
2012-10-22 12:34 ` Glauber Costa
2012-10-22 12:51 ` Michal Hocko
2012-10-22 12:52 ` Glauber Costa [this message]
2012-10-16 10:16 ` [PATCH v5 07/14] mm: Allocate kernel pages to the right memcg Glauber Costa
2012-10-16 15:31 ` Christoph Lameter
2012-10-16 18:55 ` Glauber Costa
2012-10-17 22:12 ` Andrew Morton
2012-10-18 9:24 ` Glauber Costa
2012-10-18 20:44 ` Andrew Morton
2012-10-18 11:53 ` Glauber Costa
2012-10-17 22:43 ` David Rientjes
2012-10-16 10:16 ` [PATCH v5 08/14] res_counter: return amount of charges after res_counter_uncharge Glauber Costa
2012-10-17 23:23 ` David Rientjes
2012-10-16 10:16 ` [PATCH v5 09/14] memcg: kmem accounting lifecycle management Glauber Costa
2012-10-17 23:28 ` David Rientjes
2012-10-18 6:14 ` Michal Hocko
2012-10-18 9:42 ` Glauber Costa
2012-10-16 10:16 ` [PATCH v5 10/14] memcg: use static branches when code not in use Glauber Costa
2012-10-16 10:16 ` [PATCH v5 11/14] memcg: allow a memcg with kmem charges to be destructed Glauber Costa
2012-10-17 22:12 ` Andrew Morton
2012-10-18 9:33 ` Glauber Costa
2012-10-16 10:16 ` [PATCH v5 12/14] execute the whole memcg freeing in free_worker Glauber Costa
2012-10-17 6:56 ` Kamezawa Hiroyuki
2012-10-16 10:16 ` [PATCH v5 13/14] protect architectures where THREAD_SIZE >= PAGE_SIZE against fork bombs Glauber Costa
2012-10-17 22:12 ` Andrew Morton
2012-10-18 9:37 ` Glauber Costa
2012-10-16 10:16 ` [PATCH v5 14/14] Add documentation about the kmem controller Glauber Costa
2012-10-16 12:23 ` Michal Hocko
2012-10-16 18:25 ` Christoph Lameter
2012-10-16 18:55 ` Aristeu Rozanski
2012-10-16 19:02 ` Glauber Costa
2012-10-16 19:30 ` Christoph Lameter
2012-10-17 22:12 ` Andrew Morton
2012-10-18 9:38 ` Glauber Costa
2012-10-17 22:11 ` [PATCH v5 00/14] kmem controller for memcg Andrew Morton
2012-10-18 16:51 ` Glauber Costa
2012-10-18 19:21 ` Andrew Morton
2012-10-19 9:55 ` 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=50854199.9020804@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=hannes@cmpxchg.org \
--cc=kamezawa.hiroyu@jp.fujitsu.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mgorman@suse.de \
--cc=mhocko@suse.cz \
--cc=penberg@cs.helsinki.fi \
--cc=penberg@kernel.org \
--cc=rientjes@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: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
Powered by JetHome