[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20071213092338.8b10944c.kamezawa.hiroyu@jp.fujitsu.com>
Date: Thu, 13 Dec 2007 09:23:38 +0900
From: KAMEZAWA Hiroyuki <kamezawa.hiroyu@...fujitsu.com>
To: Lee Schermerhorn <Lee.Schermerhorn@...com>
Cc: Mel Gorman <mel@....ul.ie>, akpm@...ux-foundation.org,
linux-kernel@...r.kernel.org, linux-mm@...ck.org,
rientjes@...gle.com, clameter@....com
Subject: Re: [PATCH 0/6] Use two zonelists per node instead of multiple
zonelists v11r2
On Wed, 12 Dec 2007 16:32:51 -0500
Lee Schermerhorn <Lee.Schermerhorn@...com> wrote:
> Just this afternoon, I hit a null pointer deref in
> __mem_cgroup_remove_list() [called from mem_cgroup_uncharge() if I can
> trust the stack trace] attempting to unmap a page for migration. I'm
> just starting to investigate this.
>
> I'll replace the series I have [~V10] with V11r2 and continue testing in
> anticipation of the day that we can get this into -mm.
>
Hi, Lee-san.
Could you know what is the caller of page migration ?
system call ? hot removal ? or some new thing ?
Note: 2.6.24-rc4-mm1's cgroup/migration logic.
In 2.6.24-rc4-mm1, in page migration, mem_cgroup_prepare_migration() increments
page_cgroup's refcnt before calling try_to_unmap(). This extra refcnt guarantees
the page_cgroup's refcnt will not drop to 0 in sequence of
unmap_and_move() -> try_to_unmap() -> page_remove_rmap() -> mem_cgroup_unchage().
Thanks,
-Kame
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists