[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAG6tG3wrHKTJBOu69XiV6BAPfvmK-QqJkBVpzkmkrLdLV5jRwQ@mail.gmail.com>
Date: Thu, 25 Apr 2013 11:05:53 -0400
From: Robert Love <rlove@...gle.com>
To: Shankar Brahadeeswaran <shankoo77@...il.com>
Cc: Dan Carpenter <dan.carpenter@...cle.com>,
LKML <linux-kernel@...r.kernel.org>,
Bjorn Bringert <bringert@...gle.com>,
Al Viro <viro@...iv.linux.org.uk>, devel@...verdev.osuosl.org,
Hugh Dickins <hughd@...gle.com>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Anjana V Kumar <anjanavk12@...il.com>
Subject: Re: [BUG] staging: android: ashmem: Deadlock during ashmem_shrink
On Thu, Apr 25, 2013 at 9:54 AM, Shankar Brahadeeswaran
<shankoo77@...il.com> wrote:
> Also, there are other places in the code where ashmem_mutex is held and memory
> allocation functions are called, ex:- range_alloc, calls kmem_cache_zalloc
>
> Since ashmem_shrink holds the ashmem_mutex, any where from ashmem driver
> if a memory allocation function is called with the ashmem_mutex held
> && if there is a low memory condition that leads to shrinkers being called
> we'll hit the deadlock.
The usual way this is solved is by checking the gfp_mask in the
shrinker code and bailing out (like we do now) for certain masks. So
e.g. the kmem_cache_zalloc in range_alloc is fixed by changing the
mask to GFP_FS.
> I'm trying to see if the ashmem_shrink should really hold the ashmem_mutex,
> but looks like its necessary.
Yes, it needs to hold ashmem_mutex.
There's no reason we have to run ashmem_shrink, though. See attached (untested).
Robert
Download attachment "ashmem-lock-fix-rlove-2.patch" of type "application/octet-stream" (739 bytes)
Powered by blists - more mailing lists