[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <4F223F11.4050307@suse.com>
Date: Fri, 27 Jan 2012 11:37:13 +0530
From: Suresh Jayaraman <sjayaraman@...e.com>
To: Dirk Gouders <gouders@...bocholt.fh-gelsenkirchen.de>
CC: Vivek Goyal <vgoyal@...hat.com>, Tejun Heo <tj@...nel.org>,
LKML <linux-kernel@...r.kernel.org>, Jens Axboe <axboe@...nel.dk>
Subject: Re: Slab corruption in floppy driver module
On 01/27/2012 03:18 AM, Dirk Gouders wrote:
> Vivek Goyal <vgoyal@...hat.com> writes:
>
>> On Thu, Jan 26, 2012 at 10:05:32AM -0800, Tejun Heo wrote:
>>> Hello,
>>>
>>> On Thu, Jan 26, 2012 at 10:04:20AM -0500, Vivek Goyal wrote:
>>>> out_put_disk:
>>>> while (dr--) {
>>>> del_timer_sync(&motor_off_timer[dr]);
>>>> - if (disks[dr]->queue)
>>>> + if (disks[dr]->queue) {
>>>> blk_cleanup_queue(disks[dr]->queue);
>>>> + /*
>>>> + * The request queue reference we took at device
>>>> + * creation time has been put by above
>>>> + * blk_cleanup_queue(). We have not called add_disk()
>>>> + * yet and due to failure calling put_disk(). Put disk
>>>> + * will try to put a reference to disk->queue which is
>>>> + * taken in add_disk(). As we have not taken that
>>>> + * extra reference, putting extra reference down
>>>> + * will try to access already freed queue. Clear
>>>> + * disk->queue before calling put_disk().>
>>>> + */
>>>> + disks[dr]->queue = NULL;
>>>
>>> Yeah, this looks correct to me. It might be better to tone down the
>>> comment a bit tho. Wouldn't it be sufficient to say put_disk() isn't
>>> paired with add_disk() and will put one extra time?
>>
>> Sure. Toned down the comment as suggested. Here is the new patch.
>>
>> floppy: Cleanup disk->queue before caling put_disk() if add_disk() was never called
>>
>> add_disk() takes gendisk reference on request queue. If driver failed during
>> initialization and never called add_disk() then that extra reference is not
>> taken. That reference is put in put_disk(). floppy driver allocates the
>> disk, allocates queue, sets disk->queue and then relizes that floppy
>> controller is not present. It tries to tear down everything and tries to
>> put a reference down in put_disk() which was never taken.
>>
>> In such error cases cleanup disk->queue before calling put_disk() so that
>> we never try to put down a reference which was never taken in first place.
>>
>> Reported-by: Suresh Jayaraman <sjayaraman@...e.com>
>> Tested-by: Dirk Gouders <gouders@...bocholt.fh-gelsenkirchen.de>
>> Signed-off-by: Vivek Goyal <vgoyal@...hat.com>
>> ---
>> drivers/block/floppy.c | 8 +++++++-
>> 1 file changed, 7 insertions(+), 1 deletion(-)
>>
>> Index: linux-2.6/drivers/block/floppy.c
>> ===================================================================
>> --- linux-2.6.orig/drivers/block/floppy.c 2012-01-15 09:49:14.000000000 -0500
>> +++ linux-2.6/drivers/block/floppy.c 2012-01-26 14:35:14.662374464 -0500
>> @@ -4368,8 +4368,14 @@ out_unreg_blkdev:
>> out_put_disk:
>> while (dr--) {
>> del_timer_sync(&motor_off_timer[dr]);
>> - if (disks[dr]->queue)
>> + if (disks[dr]->queue) {
>> blk_cleanup_queue(disks[dr]->queue);
>> + /*
>> + * put_disk() is not paired with add_disk() and
>> + * will put queue reference one extra time. fix it.
>> + */
>> + disks[dr]->queue = NULL;
>> + }
>> put_disk(disks[dr]);
>> }
>> return err;
>
>
> Probably a rare and uncommon one but it seems that the reloading case on
> a machine that has a floppy controller is a different problem. To be
> sure I tested the patch on a machine that has a floppy controller and
> when unloading and reloading the floppy module the log messages that I
> attached to a mail earlier in this thread are still generated.
>
Yeah, this seems like a different problem. Could you please try enabling
CONFIG_DEBUG_PAGEALLOC and see whether is it pointing to the problem
code while loading/unloading the module?
Suresh
--
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