From: Kamezawa Hiroyuki <kamezawa.hiroyu-+CUm20s59erQFUHtdCDX3A@public.gmane.org>
To: Li Zefan <lizefan-hv44wF8Li93QT0dZR+AlfA@public.gmane.org>
Cc: Andrew Morton
<akpm-de/tnXTf+JLsfHDXvbKv3WD2FQJk+8+b@public.gmane.org>,
Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>,
Glauber Costa <glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org>,
Johannes Weiner <hannes-druUgvl0LCNAfugRpC6u6w@public.gmane.org>,
LKML <linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org>,
Cgroups <cgroups-u79uwXL29TY76Z2rM5mHXA@public.gmane.org>,
linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public.gmane.org
Subject: Re: [PATCH 04/12] memcg, kmem: fix reference count handling on the error path
Date: Tue, 09 Apr 2013 11:51:39 +0900 [thread overview]
Message-ID: <5163823B.4020109@jp.fujitsu.com> (raw)
In-Reply-To: <516264D6.3090100-hv44wF8Li93QT0dZR+AlfA@public.gmane.org>
(2013/04/08 15:33), Li Zefan wrote:
> From: Michal Hocko <mhocko-AlSwsSmVLrQ@public.gmane.org>
>
> mem_cgroup_css_online calls mem_cgroup_put if memcg_init_kmem
> fails. This is not correct because only memcg_propagate_kmem takes an
> additional reference while mem_cgroup_sockets_init is allowed to fail as
> well (although no current implementation fails) but it doesn't take any
> reference. This all suggests that it should be memcg_propagate_kmem that
> should clean up after itself so this patch moves mem_cgroup_put over
> there.
>
> Unfortunately this is not that easy (as pointed out by Li Zefan) because
> memcg_kmem_mark_dead marks the group dead (KMEM_ACCOUNTED_DEAD) if it
> is marked active (KMEM_ACCOUNTED_ACTIVE) which is the case even if
> memcg_propagate_kmem fails so the additional reference is dropped in
> that case in kmem_cgroup_destroy which means that the reference would be
> dropped two times.
>
> The easiest way then would be to simply remove mem_cgrroup_put from
> mem_cgroup_css_online and rely on kmem_cgroup_destroy doing the right
> thing.
>
> Signed-off-by: Li Zefan <lizefan-hv44wF8Li93QT0dZR+AlfA@public.gmane.org>
> Signed-off-by: Michal Hocko <mhocko-AlSwsSmVLrQ@public.gmane.org>
> Cc: <stable-u79uwXL29TY76Z2rM5mHXA@public.gmane.org> # 3.8.x
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu-+CUm20s59erQFUHtdCDX3A@public.gmane.org>
WARNING: multiple messages have this Message-ID (diff)
From: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
To: Li Zefan <lizefan@huawei.com>
Cc: Andrew Morton <akpm@linux-foundation.org>,
Tejun Heo <tj@kernel.org>, Glauber Costa <glommer@parallels.com>,
Johannes Weiner <hannes@cmpxchg.org>,
LKML <linux-kernel@vger.kernel.org>,
Cgroups <cgroups@vger.kernel.org>,
linux-mm@kvack.org
Subject: Re: [PATCH 04/12] memcg, kmem: fix reference count handling on the error path
Date: Tue, 09 Apr 2013 11:51:39 +0900 [thread overview]
Message-ID: <5163823B.4020109@jp.fujitsu.com> (raw)
In-Reply-To: <516264D6.3090100@huawei.com>
(2013/04/08 15:33), Li Zefan wrote:
> From: Michal Hocko <mhocko@suse.cz>
>
> mem_cgroup_css_online calls mem_cgroup_put if memcg_init_kmem
> fails. This is not correct because only memcg_propagate_kmem takes an
> additional reference while mem_cgroup_sockets_init is allowed to fail as
> well (although no current implementation fails) but it doesn't take any
> reference. This all suggests that it should be memcg_propagate_kmem that
> should clean up after itself so this patch moves mem_cgroup_put over
> there.
>
> Unfortunately this is not that easy (as pointed out by Li Zefan) because
> memcg_kmem_mark_dead marks the group dead (KMEM_ACCOUNTED_DEAD) if it
> is marked active (KMEM_ACCOUNTED_ACTIVE) which is the case even if
> memcg_propagate_kmem fails so the additional reference is dropped in
> that case in kmem_cgroup_destroy which means that the reference would be
> dropped two times.
>
> The easiest way then would be to simply remove mem_cgrroup_put from
> mem_cgroup_css_online and rely on kmem_cgroup_destroy doing the right
> thing.
>
> Signed-off-by: Li Zefan <lizefan@huawei.com>
> Signed-off-by: Michal Hocko <mhocko@suse.cz>
> Cc: <stable@vger.kernel.org> # 3.8.x
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: Li Zefan <lizefan@huawei.com>
Cc: Andrew Morton <akpm@linux-foundation.org>,
Tejun Heo <tj@kernel.org>, Glauber Costa <glommer@parallels.com>,
Johannes Weiner <hannes@cmpxchg.org>,
LKML <linux-kernel@vger.kernel.org>,
Cgroups <cgroups@vger.kernel.org>,
linux-mm@kvack.org
Subject: Re: [PATCH 04/12] memcg, kmem: fix reference count handling on the error path
Date: Tue, 09 Apr 2013 11:51:39 +0900 [thread overview]
Message-ID: <5163823B.4020109@jp.fujitsu.com> (raw)
In-Reply-To: <516264D6.3090100@huawei.com>
(2013/04/08 15:33), Li Zefan wrote:
> From: Michal Hocko <mhocko@suse.cz>
>
> mem_cgroup_css_online calls mem_cgroup_put if memcg_init_kmem
> fails. This is not correct because only memcg_propagate_kmem takes an
> additional reference while mem_cgroup_sockets_init is allowed to fail as
> well (although no current implementation fails) but it doesn't take any
> reference. This all suggests that it should be memcg_propagate_kmem that
> should clean up after itself so this patch moves mem_cgroup_put over
> there.
>
> Unfortunately this is not that easy (as pointed out by Li Zefan) because
> memcg_kmem_mark_dead marks the group dead (KMEM_ACCOUNTED_DEAD) if it
> is marked active (KMEM_ACCOUNTED_ACTIVE) which is the case even if
> memcg_propagate_kmem fails so the additional reference is dropped in
> that case in kmem_cgroup_destroy which means that the reference would be
> dropped two times.
>
> The easiest way then would be to simply remove mem_cgrroup_put from
> mem_cgroup_css_online and rely on kmem_cgroup_destroy doing the right
> thing.
>
> Signed-off-by: Li Zefan <lizefan@huawei.com>
> Signed-off-by: Michal Hocko <mhocko@suse.cz>
> Cc: <stable@vger.kernel.org> # 3.8.x
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
next prev parent reply other threads:[~2013-04-09 2:51 UTC|newest]
Thread overview: 81+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-04-08 6:32 [PATCH 0/12][V2] memcg: make memcg's life cycle the same as cgroup Li Zefan
2013-04-08 6:32 ` Li Zefan
2013-04-08 6:32 ` Li Zefan
2013-04-08 6:34 ` [PATCH 05/12] memcg: use css_get() in sock_update_memcg() Li Zefan
2013-04-08 6:34 ` Li Zefan
2013-04-08 6:34 ` Li Zefan
2013-04-08 6:34 ` [PATCH 06/12] memcg: don't use mem_cgroup_get() when creating a kmemcg cache Li Zefan
2013-04-08 6:34 ` Li Zefan
2013-04-08 6:34 ` Li Zefan
2013-04-09 2:53 ` Kamezawa Hiroyuki
2013-04-09 2:53 ` Kamezawa Hiroyuki
2013-04-08 6:34 ` [PATCH 07/12] memcg: use css_get/put when charging/uncharging kmem Li Zefan
2013-04-08 6:34 ` Li Zefan
2013-04-08 6:34 ` Li Zefan
2013-04-08 14:14 ` Michal Hocko
2013-04-08 14:14 ` Michal Hocko
2013-04-09 2:55 ` Kamezawa Hiroyuki
2013-04-09 2:55 ` Kamezawa Hiroyuki
2013-04-08 6:34 ` [PATCH 08/12] memcg: use css_get/put for swap memcg Li Zefan
2013-04-08 6:34 ` Li Zefan
2013-04-08 6:34 ` Li Zefan
2013-04-08 6:36 ` [PATCH 11/12] memcg: kill memcg refcnt Li Zefan
2013-04-08 6:36 ` Li Zefan
2013-04-08 6:36 ` Li Zefan
[not found] ` <5162648B.9070802-hv44wF8Li93QT0dZR+AlfA@public.gmane.org>
2013-04-08 6:33 ` [PATCH 01/12] memcg: take reference before releasing rcu_read_lock Li Zefan
2013-04-08 6:33 ` Li Zefan
2013-04-08 6:33 ` Li Zefan
2013-04-08 6:33 ` [PATCH 02/12] memcg: avoid accessing memcg after releasing reference Li Zefan
2013-04-08 6:33 ` Li Zefan
2013-04-08 6:33 ` Li Zefan
2013-04-08 6:33 ` [PATCH 03/12] Revert "memcg: avoid dangling reference count in creation failure." Li Zefan
2013-04-08 6:33 ` Li Zefan
2013-04-08 6:33 ` Li Zefan
2013-04-09 2:50 ` Kamezawa Hiroyuki
2013-04-09 2:50 ` Kamezawa Hiroyuki
2013-04-08 6:33 ` [PATCH 04/12] memcg, kmem: fix reference count handling on the error path Li Zefan
2013-04-08 6:33 ` Li Zefan
2013-04-08 6:33 ` Li Zefan
[not found] ` <516264D6.3090100-hv44wF8Li93QT0dZR+AlfA@public.gmane.org>
2013-04-09 2:51 ` Kamezawa Hiroyuki [this message]
2013-04-09 2:51 ` Kamezawa Hiroyuki
2013-04-09 2:51 ` Kamezawa Hiroyuki
2013-04-08 6:35 ` [PATCH 09/12] cgroup: make sure parent won't be destroyed before its children Li Zefan
2013-04-08 6:35 ` Li Zefan
2013-04-08 6:35 ` Li Zefan
[not found] ` <51626516.3000603-hv44wF8Li93QT0dZR+AlfA@public.gmane.org>
2013-04-08 15:36 ` Tejun Heo
2013-04-08 15:36 ` Tejun Heo
2013-04-08 15:36 ` Tejun Heo
2013-04-08 6:35 ` [PATCH 10/12] memcg: don't need to get a reference to the parent Li Zefan
2013-04-08 6:35 ` Li Zefan
2013-04-08 6:35 ` Li Zefan
2013-04-08 6:36 ` [PATCH 12/12] memcg: don't need to free memcg via RCU or workqueue Li Zefan
2013-04-08 6:36 ` Li Zefan
2013-04-08 6:36 ` Li Zefan
2013-04-08 14:15 ` Michal Hocko
2013-04-08 14:15 ` Michal Hocko
[not found] ` <51626570.8000400-hv44wF8Li93QT0dZR+AlfA@public.gmane.org>
2013-04-09 2:57 ` Kamezawa Hiroyuki
2013-04-09 2:57 ` Kamezawa Hiroyuki
2013-04-09 2:57 ` Kamezawa Hiroyuki
2013-04-08 6:36 ` [PATCH 13/12] memcg: don't need memcg->memcg_name Li Zefan
2013-04-08 6:36 ` Li Zefan
2013-04-08 6:36 ` Li Zefan
2013-04-08 14:25 ` Michal Hocko
2013-04-08 14:25 ` Michal Hocko
[not found] ` <20130408142503.GH17178-2MMpYkNvuYDjFM9bn6wA6Q@public.gmane.org>
2013-04-09 1:28 ` Li Zefan
2013-04-09 1:28 ` Li Zefan
2013-04-09 1:28 ` Li Zefan
2013-04-09 3:10 ` Kamezawa Hiroyuki
2013-04-09 3:10 ` Kamezawa Hiroyuki
[not found] ` <5163868B.3020905-+CUm20s59erQFUHtdCDX3A@public.gmane.org>
2013-04-09 3:18 ` Li Zefan
2013-04-09 3:18 ` Li Zefan
2013-04-09 3:18 ` Li Zefan
2013-04-09 3:46 ` Kamezawa Hiroyuki
2013-04-09 3:46 ` Kamezawa Hiroyuki
[not found] ` <51638F2B.3000800-+CUm20s59erQFUHtdCDX3A@public.gmane.org>
2013-04-09 7:55 ` Glauber Costa
2013-04-09 7:55 ` Glauber Costa
2013-04-09 7:55 ` Glauber Costa
[not found] ` <5163887D.1040809-hv44wF8Li93QT0dZR+AlfA@public.gmane.org>
2013-04-09 3:59 ` Tejun Heo
2013-04-09 3:59 ` Tejun Heo
2013-04-09 3:59 ` Tejun Heo
2013-04-09 7:53 ` Glauber Costa
2013-04-09 7:53 ` 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=5163823B.4020109@jp.fujitsu.com \
--to=kamezawa.hiroyu-+cum20s59erqfuhtdcdx3a@public.gmane.org \
--cc=akpm-de/tnXTf+JLsfHDXvbKv3WD2FQJk+8+b@public.gmane.org \
--cc=cgroups-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
--cc=glommer-bzQdu9zFT3WakBO8gow8eQ@public.gmane.org \
--cc=hannes-druUgvl0LCNAfugRpC6u6w@public.gmane.org \
--cc=linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
--cc=linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public.gmane.org \
--cc=lizefan-hv44wF8Li93QT0dZR+AlfA@public.gmane.org \
--cc=tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.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 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.