mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Michal Hocko <mhocko@suse.cz>
To: Johannes Weiner <hannes@cmpxchg.org>
Cc: linux-mm@kvack.org, Hugh Dickins <hughd@google.com>,
	Tejun Heo <tj@kernel.org>,
	cgroups@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [patch 7/9] mm: memcontrol: do not acquire page_cgroup lock for kmem pages
Date: Fri, 23 May 2014 15:39:38 +0200	[thread overview]
Message-ID: <20140523133938.GC22135@dhcp22.suse.cz> (raw)
In-Reply-To: <1398889543-23671-8-git-send-email-hannes@cmpxchg.org>

I am adding Vladimir to CC

On Wed 30-04-14 16:25:41, Johannes Weiner wrote:
> Kmem page charging and uncharging is serialized by means of exclusive
> access to the page.  Do not take the page_cgroup lock and don't set
> pc->flags atomically.
> 
> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>

The patch is correct I just have some comments below.
Anyway
Acked-by: Michal Hocko <mhocko@suse.cz>

> ---
>  mm/memcontrol.c | 16 +++-------------
>  1 file changed, 3 insertions(+), 13 deletions(-)
> 
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index c528ae9ac230..d3961fce1d54 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -3535,10 +3535,8 @@ void __memcg_kmem_commit_charge(struct page *page, struct mem_cgroup *memcg,
>  	}
>  

	/*
	 * given page is newly allocated and invisible to everybody but
	 * the caller so there is no need to use page_cgroup lock nor
	 * SetPageCgroupUsed
	 */

would be helpful?

>  	pc = lookup_page_cgroup(page);
> -	lock_page_cgroup(pc);
>  	pc->mem_cgroup = memcg;
> -	SetPageCgroupUsed(pc);
> -	unlock_page_cgroup(pc);
> +	pc->flags = PCG_USED;
>  }
>  
>  void __memcg_kmem_uncharge_pages(struct page *page, int order)
> @@ -3548,19 +3546,11 @@ void __memcg_kmem_uncharge_pages(struct page *page, int order)
>  
>  
>  	pc = lookup_page_cgroup(page);
> -	/*
> -	 * Fast unlocked return. Theoretically might have changed, have to
> -	 * check again after locking.
> -	 */

This comment was there since the code has been merged. Maybe it was true
at the time but after "mm: get rid of __GFP_KMEMCG" it is definitely out
of date.

	/*
	 * the pages is going away and will be freed and nobody can see
	 * it anymore so no need to take page_cgroup lock.
	 */
>  	if (!PageCgroupUsed(pc))
>  		return;
>  
> -	lock_page_cgroup(pc);
> -	if (PageCgroupUsed(pc)) {
> -		memcg = pc->mem_cgroup;
> -		ClearPageCgroupUsed(pc);
> -	}
> -	unlock_page_cgroup(pc);

maybe add
	WARN_ON_ONCE(pc->flags != PCG_USED);

to check for an unexpected flags usage in the kmem path?

> +	memcg = pc->mem_cgroup;
> +	pc->flags = 0;
>  
>  	/*
>  	 * We trust that only if there is a memcg associated with the page, it
> -- 
> 1.9.2
> 

-- 
Michal Hocko
SUSE Labs

  reply	other threads:[~2014-05-23 13:39 UTC|newest]

Thread overview: 35+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-04-30 20:25 [patch 0/9] mm: memcontrol: naturalize charge lifetime Johannes Weiner
2014-04-30 20:25 ` [patch 1/9] mm: memcontrol: fold mem_cgroup_do_charge() Johannes Weiner
2014-04-30 20:25 ` [patch 2/9] mm: memcontrol: rearrange charging fast path Johannes Weiner
2014-05-07 14:33   ` Michal Hocko
2014-05-08 18:22     ` Johannes Weiner
2014-05-12  7:59       ` Michal Hocko
2014-04-30 20:25 ` [patch 3/9] mm: memcontrol: retry reclaim for oom-disabled and __GFP_NOFAIL charges Johannes Weiner
2014-05-07 14:43   ` Michal Hocko
2014-05-08 18:28     ` Johannes Weiner
2014-04-30 20:25 ` [patch 4/9] mm: memcontrol: catch root bypass in move precharge Johannes Weiner
2014-05-07 14:55   ` Michal Hocko
2014-05-08 18:30     ` Johannes Weiner
2014-04-30 20:25 ` [patch 5/9] mm: memcontrol: use root_mem_cgroup res_counter Johannes Weiner
2014-05-07 15:14   ` Michal Hocko
2014-04-30 20:25 ` [patch 6/9] mm: memcontrol: remove ordering between pc->mem_cgroup and PageCgroupUsed Johannes Weiner
2014-05-23 13:20   ` Michal Hocko
2014-05-27 19:45     ` Johannes Weiner
2014-05-28 11:31       ` Michal Hocko
2014-04-30 20:25 ` [patch 7/9] mm: memcontrol: do not acquire page_cgroup lock for kmem pages Johannes Weiner
2014-05-23 13:39   ` Michal Hocko [this message]
2014-05-23 13:40     ` Michal Hocko
2014-05-23 14:29     ` Vladimir Davydov
2014-05-27 19:53     ` Johannes Weiner
2014-05-28 11:33       ` Michal Hocko
2014-04-30 20:25 ` [patch 8/9] mm: memcontrol: rewrite charge API Johannes Weiner
2014-05-23 14:18   ` Michal Hocko
2014-05-23 14:54   ` Michal Hocko
2014-05-23 15:18     ` Michal Hocko
2014-05-27 20:05     ` Johannes Weiner
2014-05-28 11:37       ` Michal Hocko
2014-04-30 20:25 ` [patch 9/9] mm: memcontrol: rewrite uncharge API Johannes Weiner
2014-05-04 14:32   ` Johannes Weiner
2014-05-27  7:43   ` Kamezawa Hiroyuki
2014-05-27 18:59     ` Johannes Weiner
2014-05-02 11:26 ` [patch 0/9] mm: memcontrol: naturalize charge lifetime Michal Hocko

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=20140523133938.GC22135@dhcp22.suse.cz \
    --to=mhocko@suse.cz \
    --cc=cgroups@vger.kernel.org \
    --cc=hannes@cmpxchg.org \
    --cc=hughd@google.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --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

all inboxes | Powered by JetHome®