From: Michal Hocko <mhocko@kernel.org>
To: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Cc: guro@fb.com, rientjes@google.com, hannes@cmpxchg.org,
vdavydov.dev@gmail.com, tj@kernel.org, linux-mm@kvack.org,
akpm@linux-foundation.org, torvalds@linux-foundation.org
Subject: Re: [PATCH] mm,oom: Don't call schedule_timeout_killable() with oom_lock held.
Date: Tue, 29 May 2018 19:18:33 +0200 [thread overview]
Message-ID: <20180529171833.GX27180@dhcp22.suse.cz> (raw)
In-Reply-To: <40a5a42f-6812-b4ee-a72e-7f01dc9de464@i-love.sakura.ne.jp>
On Tue 29-05-18 23:33:13, Tetsuo Handa wrote:
> On 2018/05/29 17:16, Michal Hocko wrote:
> > With the full changelog. This can be either folded into the respective
> > patch or applied on top.
> >
> >>From 0bd619e7a68337c97bdaed288e813e96a14ba339 Mon Sep 17 00:00:00 2001
> > From: Michal Hocko <mhocko@suse.com>
> > Date: Tue, 29 May 2018 10:09:33 +0200
> > Subject: [PATCH] mm, memcg, oom: fix pre-mature allocation failures
> >
> > Tetsuo has noticed that "mm, oom: cgroup-aware OOM killer" can lead to a
> > pre-mature allocation failure if the cgroup aware oom killer is enabled
> > and select_victim_memcg doesn't pick up any memcg to kill because there
> > is a memcg already being killed. oc->chosen_memcg will become INFLIGHT_VICTIM
> > and oom_kill_memcg_victim will bail out early. oc->chosen_task will
> > stay NULL, however, and out_of_memory will therefore return false which
> > forces __alloc_pages_may_oom to not set did_some_progress and the page
> > allocator backs out and fails the allocation.
> > U
> > Fix this by checking both chosen_task and chosen_memcg in out_of_memory
> > and return false only when _both_ are NULL.
>
> I don't like this patch. It is not easy to understand and is fragile to
> future changes. Currently the only case !!oc->chosen can become false is that
> there was no eligible tasks when SysRq-f was requested or memcg OOM occurred.
Well, the current contract is not easy unfortunatelly. We have two
different modes of operation. We are either killing whole cgroups or a
task from a cgroup. In any case, the contract says that if we have any
killable entity then at least one of chosen* is set to INFLIGHT_VICTIM.
Other than that one of them has to be !NULL or we have no eligible
killable entity. The return value reflects all these cases.
> /* Found nothing?!?! Either we hang forever, or we panic. */
> if (!oc->chosen && !is_sysrq_oom(oc) && !is_memcg_oom(oc)) {
>
> With this patch applied, what happens if
> mem_cgroup_select_oom_victim(oc) && oom_kill_memcg_victim(oc) forgot to set
> oc->chosen_memcg to NULL and called select_bad_process(oc) and reached
Well, this would be a clear bug and breaking the expected contract. I do
not see such a bug. Do you?
> /* Found nothing?!?! Either we hang forever, or we panic. */
> if (!oc->chosen_task && !is_sysrq_oom(oc) && !is_memcg_oom(oc)) {
>
> but did not trigger panic() because of is_sysrq_oom(oc) || is_memcg_oom(oc)
> and reached the last "!!(oc->chosen_task | oc->chosen_memcg)" line?
> It will by error return "true" when no eligible tasks found...
What are you talking about?
> Don't make return conditions complicated.
> The appropriate fix is to kill "delay" and "goto out;" now! My patch does it!!
I will leave the decision about which fix to take to Roman.
--
Michal Hocko
SUSE Labs
next prev parent reply other threads:[~2018-05-29 17:18 UTC|newest]
Thread overview: 51+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-05-12 14:18 [PATCH] mm,oom: Don't call schedule_timeout_killable() with oom_lock held Tetsuo Handa
2018-05-15 9:16 ` Michal Hocko
2018-05-18 10:14 ` Tetsuo Handa
2018-05-18 12:20 ` Michal Hocko
2018-05-20 15:56 ` Tetsuo Handa
2018-05-22 6:18 ` Michal Hocko
2018-05-23 10:24 ` Tetsuo Handa
2018-05-23 11:57 ` Michal Hocko
2018-05-23 13:45 ` Tetsuo Handa
2018-05-23 14:56 ` Michal Hocko
2018-05-24 10:51 ` Tetsuo Handa
2018-05-24 11:50 ` Michal Hocko
2018-05-25 1:17 ` Tetsuo Handa
2018-05-25 8:31 ` Michal Hocko
2018-05-25 10:57 ` Tetsuo Handa
2018-05-25 11:42 ` Michal Hocko
2018-05-25 11:46 ` Tetsuo Handa
2018-05-28 12:43 ` Michal Hocko
2018-05-28 20:57 ` Tetsuo Handa
2018-05-29 7:17 ` Michal Hocko
2018-05-29 23:07 ` Andrew Morton
2018-05-31 10:10 ` Tetsuo Handa
2018-05-31 10:44 ` Michal Hocko
2018-05-31 15:23 ` Tetsuo Handa
2018-05-31 18:47 ` Michal Hocko
2018-06-01 1:21 ` Tetsuo Handa
2018-06-01 8:04 ` Michal Hocko
2018-06-01 15:28 ` Michal Hocko
2018-06-01 21:11 ` Andrew Morton
2018-06-04 7:04 ` Michal Hocko
2018-06-04 10:41 ` Tetsuo Handa
2018-06-04 11:22 ` Michal Hocko
2018-06-04 11:30 ` Tetsuo Handa
2018-06-06 9:02 ` David Rientjes
2018-06-06 13:37 ` Tetsuo Handa
2018-06-06 18:44 ` David Rientjes
2018-05-29 7:17 ` Michal Hocko
2018-05-29 8:16 ` Michal Hocko
2018-05-29 14:33 ` Tetsuo Handa
2018-05-29 17:18 ` Michal Hocko [this message]
2018-05-29 17:28 ` Michal Hocko
2018-05-31 16:31 ` [PATCH] mm, memcg, oom: fix pre-mature allocation failures kbuild test robot
-- strict thread matches above, loose matches on Subject: below --
2018-03-22 10:51 [PATCH] mm,oom: Don't call schedule_timeout_killable() with oom_lock held Tetsuo Handa
2018-03-22 11:45 ` Michal Hocko
2018-03-22 13:16 ` Tetsuo Handa
2018-01-22 13:46 Tetsuo Handa
2018-01-23 8:38 ` Michal Hocko
2018-01-23 12:07 ` Tetsuo Handa
2018-01-23 12:42 ` Michal Hocko
2018-01-24 13:28 ` Tetsuo Handa
2018-02-13 11:58 ` Tetsuo Handa
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=20180529171833.GX27180@dhcp22.suse.cz \
--to=mhocko@kernel.org \
--cc=akpm@linux-foundation.org \
--cc=guro@fb.com \
--cc=hannes@cmpxchg.org \
--cc=linux-mm@kvack.org \
--cc=penguin-kernel@i-love.sakura.ne.jp \
--cc=rientjes@google.com \
--cc=tj@kernel.org \
--cc=torvalds@linux-foundation.org \
--cc=vdavydov.dev@gmail.com \
/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;
as well as URLs for NNTP newsgroup(s).