From: Michal Hocko <mhocko@suse.cz>
To: Oleg Nesterov <oleg@redhat.com>
Cc: Tejun Heo <tj@kernel.org>,
lizefan@huawei.com, cgroups@vger.kernel.org, hannes@cmpxchg.org,
linux-mm@kvack.org
Subject: Re: [PATCH 3/7] memcg: immigrate charges only when a threadgroup leader is moved
Date: Wed, 20 May 2015 22:22:21 +0200 [thread overview]
Message-ID: <20150520202221.GD14256@dhcp22.suse.cz> (raw)
In-Reply-To: <20150520175302.GA7287@redhat.com>
On Wed 20-05-15 19:53:02, Oleg Nesterov wrote:
> On 05/20, Michal Hocko wrote:
> >
> > So I assume the leader simply waits for its threads to finish and it
> > stays in the sibling list. __unhash_process seems like it does the final
> > cleanup and unlinks the leader from the lists. Which means that
> > mm_update_next_owner never sees !group_leader. Is that correct Oleg?
>
> Yes, yes, the group leader can't go away until the whole thread-group dies.
OK, then we should have a guarantee that mm->owner is always thread
group leader, right?
> But can't we kill mm->owner somehow?
I would be happy about that. But it is not that simple.
> I mean, turn it into something else,
> ideally into "struct mem_cgroup *" although I doubt this is possible.
Sounds like a good idea but... it duplicates the cgroup tracking into
two places and that asks for troubles. On the other hand we are doing
that already because mm->owner might be in a different cgroup than the
current. However, this is an inherent problem because CLONE_VM doesn't
imply CLONE_THREAD. So in the end it doesn't look much worse IMO.
We will loose the "this task is in charge" aspect and that would
be a user space visible change but I am not sure how much it is a
problem. Maybe somebody is (ab)using this to workaround the restriction
that all threads are in the same cgroup.
>From the implementation POV it even looks easier because we just have to
hook to fork (pin the memcg on dup_mm), to attach to change the memcg
and to mmput to unpin the memcg.
I will think about that some more.
> It would be nice to kill mm_update_next_owner()/etc, this looks really
> ugly. We only need it for mem_cgroup_from_task(), and it would be much
> more clean to have mem_cgroup_from_mm(struct mm_struct *mm), imho.
>
> Oleg.
>
--
Michal Hocko
SUSE Labs
--
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>
next prev parent reply other threads:[~2015-05-20 20:22 UTC|newest]
Thread overview: 34+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-05-18 19:49 [PATCHSET cgroup/for-4.2] cgroup: make multi-process migration atomic Tejun Heo
2015-05-18 19:49 ` [PATCH 1/7] cpuset: migrate memory only for threadgroup leaders Tejun Heo
2015-05-18 19:49 ` [PATCH 2/7] memcg: restructure mem_cgroup_can_attach() Tejun Heo
2015-05-19 9:03 ` Michal Hocko
2015-05-18 19:49 ` [PATCH 3/7] memcg: immigrate charges only when a threadgroup leader is moved Tejun Heo
2015-05-19 12:13 ` Michal Hocko
2015-05-19 13:10 ` Michal Hocko
2015-05-19 21:27 ` Tejun Heo
2015-05-20 13:10 ` Michal Hocko
2015-05-20 13:21 ` Michal Hocko
2015-05-20 17:53 ` Oleg Nesterov
2015-05-20 20:22 ` Michal Hocko [this message]
2015-05-21 17:22 ` Johannes Weiner
2015-05-22 9:34 ` Michal Hocko
2015-05-21 19:27 ` Oleg Nesterov
2015-05-22 9:36 ` Michal Hocko
2015-05-22 16:29 ` Oleg Nesterov
2015-05-22 16:57 ` Michal Hocko
2015-05-22 18:30 ` Oleg Nesterov
2015-05-25 16:06 ` Michal Hocko
2015-05-25 17:06 ` Oleg Nesterov
2015-05-26 7:16 ` Michal Hocko
2015-05-22 18:20 ` [PATCH 0/3] memcg: mm_update_next_owner() cleanups Oleg Nesterov
2015-05-22 18:21 ` [PATCH 1/3] memcg: introduce assign_new_owner() Oleg Nesterov
2015-05-22 18:21 ` [PATCH 2/3] memcg: change assign_new_owner() to consider the sub-htreads Oleg Nesterov
2015-05-22 18:21 ` [PATCH 3/3] memcg: change mm_update_next_owner() to search in sub-threads first Oleg Nesterov
2015-05-22 18:22 ` [PATCH 0/3] memcg: mm_update_next_owner() cleanups Oleg Nesterov
2015-05-21 14:12 ` [PATCH 3/7] memcg: immigrate charges only when a threadgroup leader is moved Michal Hocko
2015-05-21 22:09 ` Tejun Heo
2015-05-18 19:49 ` [PATCH 4/7] cgroup, memcg, cpuset: implement cgroup_taskset_for_each_leader() Tejun Heo
2015-05-18 19:49 ` [PATCH 5/7] reorder cgroup_migrate()'s parameters Tejun Heo
2015-05-18 19:49 ` [PATCH 6/7] cgroup: separate out taskset operations from cgroup_migrate() Tejun Heo
2015-05-18 19:49 ` [PATCH 7/7] cgroup: make cgroup_update_dfl_csses() migrate all target processes atomically Tejun Heo
2015-05-19 6:57 ` [PATCHSET cgroup/for-4.2] cgroup: make multi-process migration atomic Zefan Li
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=20150520202221.GD14256@dhcp22.suse.cz \
--to=mhocko@suse.cz \
--cc=cgroups@vger.kernel.org \
--cc=hannes@cmpxchg.org \
--cc=linux-mm@kvack.org \
--cc=lizefan@huawei.com \
--cc=oleg@redhat.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;
as well as URLs for NNTP newsgroup(s).