Re: [PATCH] fs-writeback: fix NULL pointer dereference in, __mark_inode_dirty

[Jens Axboe <axboe@kernel.dk>]

On 2011-03-15 13:11, Anton Altaparmakov wrote:
> Hi,
> 
> On 15 Mar 2011, at 10:17, Torsten Hilbrich wrote:
>> On 10.03.2011 11:39, Torsten Hilbrich wrote:
>>> I ran into the same problem and successfully applied your fix.
>>>
>>> I was able to reproduce this panic and bisected it to the following commit:
>>>
>>> commit aaead25b954879e1a708ff2f3602f494c18d20b5
>>> Author: Christoph Hellwig <hch@lst.de>
>>> Date:   Mon Oct 4 14:25:33 2010 +0200
>>>
>>>    writeback: always use sb->s_bdi for writeback purposes
>>>
>>> The steps to reproduce it on my test system (T60p with Intel Core Duo) were.
>>
>> Added Christoph to CC. I also open a bug report at https://bugzilla.kernel.org/show_bug.cgi?id=31112
> 
> This is already being handled.  It is the same as other bug reports,
> i.e. the fact that sb->s_bdi is made NULL on device removal and if it
> happens at the wrong time you then get a NULL pointer dereference.
> 
> Jens Axboe just only yesterday posted an initial patch for this.  Can
> you please test it and report back if it does indeed cure the problem?
> 
> The patch can be found here for example:
> 
> 	https://lkml.org/lkml/2011/3/14/25

Yes, any testing of that patch would be greatly appreciated.

-- 
Jens Axboe