linux-kernel - Re: [PATCH] venus: vdec: fix decoded data size

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

Date:   Mon, 17 Sep 2018 16:32:54 +0200
From:   Hans Verkuil <hverkuil@...all.nl>
To:     Stanimir Varbanov <stanimir.varbanov@...aro.org>,
        Nicolas Dufresne <nicolas@...fresne.ca>,
        Vikash Garodia <vgarodia@...eaurora.org>
Cc:     linux-media@...r.kernel.org, linux-arm-msm@...r.kernel.org,
        linux-kernel@...r.kernel.org, acourbot@...omium.org
Subject: Re: [PATCH] venus: vdec: fix decoded data size

On 09/17/2018 04:30 PM, Stanimir Varbanov wrote:
> Hi Hans,
> 
> On 09/17/2018 01:00 PM, Hans Verkuil wrote:
>> On 07/18/2018 04:37 PM, Stanimir Varbanov wrote:
>>> Hi,
>>>
>>> On 07/18/2018 04:26 PM, Nicolas Dufresne wrote:
>>>> Le mercredi 18 juillet 2018 à 14:31 +0300, Stanimir Varbanov a écrit :
>>>>> Hi Vikash,
>>>>>
>>>>> On 07/02/2018 10:44 AM, Vikash Garodia wrote:
>>>>>> Exisiting code returns the max of the decoded
>>>>>> size and buffer size. It turns out that buffer
>>>>>> size is always greater due to hardware alignment
>>>>>> requirement. As a result, payload size given to
>>>>>> client is incorrect. This change ensures that
>>>>>> the bytesused is assigned to actual payload size.
>>>>>>
>>>>>> Change-Id: Ie6f3429c0cb23f682544748d181fa4fa63ca2e28
>>>>>> Signed-off-by: Vikash Garodia <vgarodia@...eaurora.org>
>>>>>> ---
>>>>>>  drivers/media/platform/qcom/venus/vdec.c | 2 +-
>>>>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>>>>
>>>>>> diff --git a/drivers/media/platform/qcom/venus/vdec.c
>>>>>> b/drivers/media/platform/qcom/venus/vdec.c
>>>>>> index d079aeb..ada1d2f 100644
>>>>>> --- a/drivers/media/platform/qcom/venus/vdec.c
>>>>>> +++ b/drivers/media/platform/qcom/venus/vdec.c
>>>>>> @@ -890,7 +890,7 @@ static void vdec_buf_done(struct venus_inst
>>>>>> *inst, unsigned int buf_type,
>>>>>>  
>>>>>>  		vb = &vbuf->vb2_buf;
>>>>>>  		vb->planes[0].bytesused =
>>>>>> -			max_t(unsigned int, opb_sz, bytesused);
>>>>>> +			min_t(unsigned int, opb_sz, bytesused);
>>>>>
>>>>> Most probably my intension was to avoid bytesused == 0, but that is
>>>>> allowed from v4l2 driver -> userspace direction
>>>>
>>>> It remains bad practice since it was used by decoders to indicate the
>>>> last buffer. Some userspace (some GStreamer versions) will stop working
>>>> if you start returning 0.
>>>
>>> I think it is legal v4l2 driver to return bytesused = 0 when userspace
>>> issues streamoff on both queues before EOS, no? Simply because the
>>> capture buffers are empty.
>>>
>>
>> Going through some of the older pending patches I found this one:
>>
>> So is this patch right or wrong?
> 
> I'm not sure either, let's not applying it for now (if Nicolas is right
> this will break gstreamer plugin).
> 

OK, I marked this as Rejected. If you change your mind it can be reposted :-)

Regards,

	Hans