linux-kernel - Re: [PATCH] [media] media-entity: only call dev_dbg

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <f3f83e8f-41e3-3567-8ec6-c4e693e7297e@collabora.com>
Date:   Fri, 7 Apr 2017 11:36:29 -0300
From:   Helen Koike <helen.koike@...labora.com>
To:     Sakari Ailus <sakari.ailus@....fi>
Cc:     linux-media@...r.kernel.org, mchehab@...nel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] [media] media-entity: only call dev_dbg_obj if mdev is
 not NULL

Hi Sakari,

On 2017-04-07 04:40 AM, Sakari Ailus wrote:
> Hi Helen,
>
> On Thu, Apr 06, 2017 at 04:32:00PM -0300, Helen Koike wrote:
>> Fix kernel Oops NULL pointer deference
>> Call dev_dbg_obj only after checking if gobj->mdev is not NULL
>>
>> Signed-off-by: Helen Koike <helen.koike@...labora.com>
>> ---
>>  drivers/media/media-entity.c | 4 ++--
>>  1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/media/media-entity.c b/drivers/media/media-entity.c
>> index 5640ca2..bc44193 100644
>> --- a/drivers/media/media-entity.c
>> +++ b/drivers/media/media-entity.c
>> @@ -199,12 +199,12 @@ void media_gobj_create(struct media_device *mdev,
>>
>>  void media_gobj_destroy(struct media_gobj *gobj)
>>  {
>> -	dev_dbg_obj(__func__, gobj);
>> -
>>  	/* Do nothing if the object is not linked. */
>>  	if (gobj->mdev == NULL)
>>  		return;
>>
>> +	dev_dbg_obj(__func__, gobj);
>> +
>>  	gobj->mdev->topology_version++;
>>
>>  	/* Remove the object from mdev list */
>
> Where is media_gobj_destroy() called with an object with NULL mdev?
>
> I do not object to the change, but would like to know because I don't think
> it's supposed to happen.


This happens when media_device_unregister(mdev) is called before 
unregistering the subdevices v4l2_device_unregister_subdev(sd) (which 
should be possible).

v4l2_device_unregister_subdev(sd) ends up calling v4l2_device_release() 
that calls media_device_unregister_entity() again (previously called by 
media_device_unregister(mdev)

Helen

>
> There are issues though, until the patches fixing object referencing are
> finished and merged. Unfortunately I haven't been able to work on those
> recently, will pick them up again soon...
>