Re: [PATCH] sound/usb: fix to release stream resources from media_snd_device_delete()

From: Shuah Khan
Date: Sat Mar 19 2016 - 09:31:35 EST


On 03/19/2016 04:39 AM, Mauro Carvalho Chehab wrote:
> Em Fri, 18 Mar 2016 23:57:08 -0300
> Mauro Carvalho Chehab <mchehab@xxxxxxxxxxxxxxx> escreveu:
>
>> Em Fri, 18 Mar 2016 20:50:31 -0600
>> Shuah Khan <shuahkh@xxxxxxxxxxxxxxx> escreveu:
>>
>>> Fix to release stream resources from media_snd_device_delete() before
>>> media device is unregistered. Without this change, stream resource free
>>> is attempted after the media device is unregistered which would result
>>> in use-after-free errors.
>>>
>>> Signed-off-by: Shuah Khan <shuahkh@xxxxxxxxxxxxxxx>
>>> ---
>>>
>>> - Ran bind/unbind loop (1000 iteration) test on snd-usb-audio
>>> while running mc_nextgen_test loop (1000 iterations) in parallel.
>>> - Ran bind/unbind and rmmod/modprobe tests on both drivers. Also
>>> generated graphs when after bind/unbind, rmmod/modprobe. Graphs
>>> look good.
>>> - Note: Please apply the following patch to fix memory leak:
>>> sound/usb: Fix memory leak in media_snd_stream_delete() during unbind
>>> https://lkml.org/lkml/2016/3/16/1050
>>>
>>> sound/usb/media.c | 7 +++++++
>>> 1 file changed, 7 insertions(+)
>>>
>>> diff --git a/sound/usb/media.c b/sound/usb/media.c
>>> index de4a815..e35af88 100644
>>> --- a/sound/usb/media.c
>>> +++ b/sound/usb/media.c
>>> @@ -301,6 +301,13 @@ int media_snd_device_create(struct snd_usb_audio *chip,
>>> void media_snd_device_delete(struct snd_usb_audio *chip)
>>> {
>>> struct media_device *mdev = chip->media_dev;
>>> + struct snd_usb_stream *stream;
>>> +
>>> + /* release resources */
>>> + list_for_each_entry(stream, &chip->pcm_list, list) {
>>> + media_snd_stream_delete(&stream->substream[0]);
>>> + media_snd_stream_delete(&stream->substream[1]);
>>
>> I'll look on it better tomorrow, but it sounds weird to hardcode
>> substream[0] and [1] here... are you sure that this is valid for
>> *all* devices supported by snd-usb-audio?
>
> After looking at pcm.c and finding this:
>
> static void snd_usb_audio_stream_free(struct snd_usb_stream *stream)
> {
> free_substream(&stream->substream[0]);
> free_substream(&stream->substream[1]);
> list_del(&stream->list);
> kfree(stream);
> }
>
> It seems that assuming that substream is always an array with size 2
> is right.
>
> I'll do some tests with it today with your patch.
>

Right. snd-usb-audio uses this in several places like the one above you found.

thanks,
-- Shuah


--
Shuah Khan
Sr. Linux Kernel Developer
Open Source Innovation Group
Samsung Research America (Silicon Valley)
shuahkh@xxxxxxxxxxxxxxx | (970) 217-8978