All of lore.kernel.org
 help / color / mirror / Atom feed
From: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
To: Michal Hocko <mhocko@suse.cz>
Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	Johannes Weiner <hannes@cmpxchg.org>,
	Ying Han <yinghan@google.com>, Tejun Heo <htejun@gmail.com>,
	Glauber Costa <glommer@parallels.com>
Subject: Re: [RFC 1/5] memcg: synchronize per-zone iterator access by a spinlock
Date: Wed, 14 Nov 2012 09:03:40 +0900	[thread overview]
Message-ID: <50A2DFDC.90402@jp.fujitsu.com> (raw)
In-Reply-To: <1352820639-13521-2-git-send-email-mhocko@suse.cz>

(2012/11/14 0:30), Michal Hocko wrote:
> per-zone per-priority iterator is aimed at coordinating concurrent
> reclaimers on the same hierarchy (or the global reclaim when all
> groups are reclaimed) so that all groups get reclaimed evenly as
> much as possible. iter->position holds the last css->id visited
> and iter->generation signals the completed tree walk (when it is
> incremented).
> Concurrent reclaimers are supposed to provide a reclaim cookie which
> holds the reclaim priority and the last generation they saw. If cookie's
> generation doesn't match the iterator's view then other concurrent
> reclaimer already did the job and the tree walk is done for that
> priority.
> 
> This scheme works nicely in most cases but it is not raceless. Two
> racing reclaimers can see the same iter->position and so bang on the
> same group. iter->generation increment is not serialized as well so a
> reclaimer can see an updated iter->position with and old generation so
> the iteration might be restarted from the root of the hierarchy.
> 
> The simplest way to fix this issue is to synchronise access to the
> iterator by a lock. This implementation uses per-zone per-priority
> spinlock which linearizes only directly racing reclaimers which use
> reclaim cookies so the effect of the new locking should be really
> minimal.
> 
> I have to note that I haven't seen this as a real issue so far. The
> primary motivation for the change is different. The following patch
> will change the way how the iterator is implemented and css->id
> iteration will be replaced cgroup generic iteration which requires
> storing mem_cgroup pointer into iterator and that requires reference
> counting and so concurrent access will be a problem.
> 
> Signed-off-by: Michal Hocko <mhocko@suse.cz>

Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>


--
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>

WARNING: multiple messages have this Message-ID (diff)
From: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
To: Michal Hocko <mhocko@suse.cz>
Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	Johannes Weiner <hannes@cmpxchg.org>,
	Ying Han <yinghan@google.com>, Tejun Heo <htejun@gmail.com>,
	Glauber Costa <glommer@parallels.com>
Subject: Re: [RFC 1/5] memcg: synchronize per-zone iterator access by a spinlock
Date: Wed, 14 Nov 2012 09:03:40 +0900	[thread overview]
Message-ID: <50A2DFDC.90402@jp.fujitsu.com> (raw)
In-Reply-To: <1352820639-13521-2-git-send-email-mhocko@suse.cz>

(2012/11/14 0:30), Michal Hocko wrote:
> per-zone per-priority iterator is aimed at coordinating concurrent
> reclaimers on the same hierarchy (or the global reclaim when all
> groups are reclaimed) so that all groups get reclaimed evenly as
> much as possible. iter->position holds the last css->id visited
> and iter->generation signals the completed tree walk (when it is
> incremented).
> Concurrent reclaimers are supposed to provide a reclaim cookie which
> holds the reclaim priority and the last generation they saw. If cookie's
> generation doesn't match the iterator's view then other concurrent
> reclaimer already did the job and the tree walk is done for that
> priority.
> 
> This scheme works nicely in most cases but it is not raceless. Two
> racing reclaimers can see the same iter->position and so bang on the
> same group. iter->generation increment is not serialized as well so a
> reclaimer can see an updated iter->position with and old generation so
> the iteration might be restarted from the root of the hierarchy.
> 
> The simplest way to fix this issue is to synchronise access to the
> iterator by a lock. This implementation uses per-zone per-priority
> spinlock which linearizes only directly racing reclaimers which use
> reclaim cookies so the effect of the new locking should be really
> minimal.
> 
> I have to note that I haven't seen this as a real issue so far. The
> primary motivation for the change is different. The following patch
> will change the way how the iterator is implemented and css->id
> iteration will be replaced cgroup generic iteration which requires
> storing mem_cgroup pointer into iterator and that requires reference
> counting and so concurrent access will be a problem.
> 
> Signed-off-by: Michal Hocko <mhocko@suse.cz>

Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>



  reply	other threads:[~2012-11-14  0:03 UTC|newest]

Thread overview: 62+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-11-13 15:30 [RFC] rework mem_cgroup iterator Michal Hocko
2012-11-13 15:30 ` Michal Hocko
2012-11-13 15:30 ` [RFC 1/5] memcg: synchronize per-zone iterator access by a spinlock Michal Hocko
2012-11-13 15:30   ` Michal Hocko
2012-11-14  0:03   ` Kamezawa Hiroyuki [this message]
2012-11-14  0:03     ` Kamezawa Hiroyuki
2012-11-13 15:30 ` [RFC 2/5] memcg: rework mem_cgroup_iter to use cgroup iterators Michal Hocko
2012-11-13 15:30   ` Michal Hocko
2012-11-13 16:14   ` Tejun Heo
2012-11-13 16:14     ` Tejun Heo
2012-11-14  8:51     ` Michal Hocko
2012-11-14  8:51       ` Michal Hocko
2012-11-14 18:52       ` Tejun Heo
2012-11-14 18:52         ` Tejun Heo
2012-11-15  9:51         ` Michal Hocko
2012-11-15  9:51           ` Michal Hocko
2012-11-15 14:47           ` Tejun Heo
2012-11-15 14:47             ` Tejun Heo
2012-11-15 15:12             ` Michal Hocko
2012-11-15 15:12               ` Michal Hocko
2012-11-15 15:31               ` Tejun Heo
2012-11-15 15:31                 ` Tejun Heo
2012-11-15 16:15                 ` Michal Hocko
2012-11-15 16:15                   ` Michal Hocko
2012-11-14  0:20   ` Kamezawa Hiroyuki
2012-11-14  0:20     ` Kamezawa Hiroyuki
2012-11-14 10:10     ` Michal Hocko
2012-11-14 10:10       ` Michal Hocko
2012-11-15  4:12       ` Kamezawa Hiroyuki
2012-11-15  4:12         ` Kamezawa Hiroyuki
2012-11-15  9:52         ` Michal Hocko
2012-11-15  9:52           ` Michal Hocko
2012-11-19 14:05       ` Michal Hocko
2012-11-19 14:05         ` Michal Hocko
2012-11-19 15:11   ` Michal Hocko
2012-11-19 15:11     ` Michal Hocko
2012-11-13 15:30 ` [RFC 3/5] memcg: simplify mem_cgroup_iter Michal Hocko
2012-11-13 15:30   ` Michal Hocko
2012-11-13 15:30 ` [RFC 4/5] memcg: clean up mem_cgroup_iter Michal Hocko
2012-11-13 15:30   ` Michal Hocko
2012-11-13 15:30 ` [RFC 5/5] cgroup: remove css_get_next Michal Hocko
2012-11-13 15:30   ` Michal Hocko
2012-11-14  0:13 ` [RFC] rework mem_cgroup iterator Kamezawa Hiroyuki
2012-11-14  0:13   ` Kamezawa Hiroyuki
2012-11-14  1:55 ` Li Zefan
2012-11-14  1:55   ` Li Zefan
2012-11-14  8:36   ` Michal Hocko
2012-11-14  8:36     ` Michal Hocko
2012-11-14 18:30     ` Tejun Heo
2012-11-14 18:30       ` Tejun Heo
2012-11-15  2:12   ` Kamezawa Hiroyuki
2012-11-15  2:12     ` Kamezawa Hiroyuki
2012-11-14 16:17 ` Glauber Costa
2012-11-14 16:17   ` Glauber Costa
2012-11-14  8:40   ` Michal Hocko
2012-11-14  8:40     ` Michal Hocko
2012-11-14 18:41   ` Tejun Heo
2012-11-14 18:41     ` Tejun Heo
2012-11-15  2:44     ` Glauber Costa
2012-11-15  2:44       ` Glauber Costa
2012-11-14 18:46       ` Tejun Heo
2012-11-14 18:46         ` Tejun Heo

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=50A2DFDC.90402@jp.fujitsu.com \
    --to=kamezawa.hiroyu@jp.fujitsu.com \
    --cc=glommer@parallels.com \
    --cc=hannes@cmpxchg.org \
    --cc=htejun@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mhocko@suse.cz \
    --cc=yinghan@google.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 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.