From: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
To: David Rientjes <rientjes@google.com>
Cc: kosaki.motohiro@jp.fujitsu.com, "Figo.zhang" <figo1802@gmail.com>,
lkml <linux-kernel@vger.kernel.org>,
"linux-mm@kvack.org" <linux-mm@kvack.org>,
Andrew Morton <akpm@osdl.org>,
Linus Torvalds <torvalds@linux-foundation.org>
Subject: Re: [PATCH v2]mm/oom-kill: direct hardware access processes should get bonus
Date: Sun, 14 Nov 2010 14:07:10 +0900 (JST) [thread overview]
Message-ID: <20101112104140.DFFF.A69D9226@jp.fujitsu.com> (raw)
In-Reply-To: <alpine.DEB.2.00.1011091307240.7730@chino.kir.corp.google.com>
> > the victim should not directly access hardware devices like Xorg server,
> > because the hardware could be left in an unpredictable state, although
> > user-application can set /proc/pid/oom_score_adj to protect it. so i think
> > those processes should get 3% bonus for protection.
> >
>
> The logic here is wrong: if killing these tasks can leave hardware in an
> unpredictable state (and that state is presumably harmful), then they
> should be completely immune from oom killing since you're still leaving
> them exposed here to be killed.
>
> So the question that needs to be answered is: why do these threads deserve
> to use 3% more memory (not >4%) than others without getting killed? If
> there was some evidence that these threads have a certain quantity of
> memory they require as a fundamental attribute of CAP_SYS_RAWIO, then I
> have no objection, but that's going to be expressed in a memory quantity
> not a percentage as you have here.
3% is choosed by you :-/
> The CAP_SYS_ADMIN heuristic has a background: it is used in the oom killer
> because we have used the same 3% in __vm_enough_memory() for a long time
> and we want consistency amongst the heuristics. Adding additional bonuses
> with arbitrary values like 3% of memory for things like CAP_SYS_RAWIO
> makes the heuristic less predictable and moves us back toward the old
> heuristic which was almost entirely arbitrary.
That's bogus. __vm_enough_memory() does track virtual adress space. oom-killer
doesn't. It's unrelated.
> Now before KOSAKI-san comes out and says the old heuristic considered
> CAP_SYS_RAWIO and the new one does not so it _must_ be a regression: the
> old heuristic also divided the badness score by 4 for that capability as a
> completely arbitrary value (just like 3% is here). Other traits like
> runtime and nice levels were also removed from the heuristic. What needs
> to be shown is that CAP_SYS_RAWIO requires additional memory just to run
> or we should neglect to free 3% of memory, which could be gigabytes,
> because it has this trait.
Old background is very simple and cleaner.
CAP_SYS_RESOURCE mean the process has a privilege of using more resource.
then, oom-killer gave it additonal bonus.
CAP_SYS_RAWIO mean the process has a direct hardware access privilege
(eg X.org, RDB). and then, killing it might makes system crash.
In another story, somebody doubt 4x bonus is good or not. but 3% has
the same problem.
next prev parent reply other threads:[~2010-11-14 5:07 UTC|newest]
Thread overview: 45+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-11-02 1:43 [PATCH]oom-kill: direct hardware access processes should get bonus Figo.zhang
2010-11-02 3:10 ` David Rientjes
2010-11-02 14:24 ` Figo.zhang
2010-11-02 19:34 ` David Rientjes
2010-11-03 23:43 ` [PATCH v2]oom-kill: CAP_SYS_RESOURCE " Figo.zhang
2010-11-03 23:47 ` David Rientjes
[not found] ` <AANLkTimjfmLzr_9+Sf4gk0xGkFjffQ1VcCnwmCXA88R8@mail.gmail.com>
2010-11-04 1:38 ` Figo.zhang
2010-11-04 1:50 ` David Rientjes
2010-11-04 2:12 ` Figo.zhang
2010-11-04 2:54 ` David Rientjes
2010-11-04 4:42 ` Figo.zhang
2010-11-04 5:08 ` David Rientjes
2010-11-09 11:01 ` [PATCH " KOSAKI Motohiro
2010-11-09 12:24 ` Alan Cox
2010-11-09 21:06 ` David Rientjes
2010-11-09 21:25 ` David Rientjes
2010-11-10 14:38 ` Figo.zhang
2010-11-10 20:50 ` David Rientjes
2010-11-09 10:41 ` [PATCH]oom-kill: direct hardware access processes " KOSAKI Motohiro
2010-11-09 12:24 ` [PATCH v2]mm/oom-kill: " Figo.zhang
2010-11-09 21:16 ` David Rientjes
2010-11-10 14:48 ` Figo.zhang
2010-11-14 5:07 ` KOSAKI Motohiro [this message]
2010-11-14 21:29 ` David Rientjes
2010-11-15 1:24 ` KOSAKI Motohiro
2010-11-15 10:03 ` David Rientjes
2010-11-23 7:16 ` KOSAKI Motohiro
2010-11-28 1:36 ` David Rientjes
2010-11-30 13:00 ` KOSAKI Motohiro
2010-11-30 20:05 ` David Rientjes
2010-11-10 15:14 ` [PATCH v3]mm/oom-kill: " Figo.zhang
2010-11-10 15:24 ` Figo.zhang
2010-11-10 21:00 ` David Rientjes
2010-11-14 5:21 ` KOSAKI Motohiro
2010-11-14 21:33 ` David Rientjes
2010-11-15 3:26 ` [PATCH] Revert oom rewrite series Figo.zhang
2010-11-15 10:14 ` David Rientjes
2010-11-15 10:57 ` Alan Cox
2010-11-15 20:54 ` David Rientjes
2010-11-23 7:16 ` KOSAKI Motohiro
2011-01-04 7:51 ` [PATCH v3]mm/oom-kill: direct hardware access processes should get bonus Figo.zhang
2011-01-04 8:28 ` KAMEZAWA Hiroyuki
2011-01-04 8:56 ` Figo.zhang
2011-01-06 0:55 ` KAMEZAWA Hiroyuki
2011-01-05 3:32 ` David Rientjes
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=20101112104140.DFFF.A69D9226@jp.fujitsu.com \
--to=kosaki.motohiro@jp.fujitsu.com \
--cc=akpm@osdl.org \
--cc=figo1802@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=rientjes@google.com \
--cc=torvalds@linux-foundation.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;
as well as URLs for NNTP newsgroup(s).