[FFmpeg-devel] [PATCH] libavformat/rtsp: fix rtsp multicasts

[Wolfgang Haupt]

On 20.04.20 10:08, Marton Balint wrote:
>
>
> On Mon, 20 Apr 2020, Wolfgang Haupt wrote:
>
>> On 19.04.20 23:43, Marton Balint wrote:
>>>
>>>
>>> On Sun, 19 Apr 2020, Wolfgang Haupt wrote:
>>>
>>>> On 19.04.20 14:53, Marton Balint wrote:
>>>>>
>>>>>
>>>>> On Sun, 19 Apr 2020, Wolfgang Haupt wrote:
>>>>>
>>>>>> ping
>>>>>>
>>>>>> On 03.04.20 08:42, Wolfgang Haupt wrote:
>>>>>>> Hey,
>>>>>>>
>>>>>>> is someone up to review this patch?
>>>>>>>
>>>>>>> It's an attempt to fix rtsp streams that use udp multicasts as 
>>>>>>> the underlying
>>>>>>> transmission protocol.
>>>>>>> The idea was taken from live555 as the said stream worked in VLC.
>>>>>>>
>>>>>>> It still applies cleanly on current master.
>>>>>>>
>>>>>>>
>>>>>>> Best Regards,
>>>>>>> Wolfgang
>>>>>>>
>>>>>>> On 18.10.19 18:59, Wolfgang Haupt wrote:
>>>>>>>> If an rtsp server offers a udp multicast
>>>>>>>> address as response of a DESCRIBE command
>>>>>>>> the rtsp client is expected to issue
>>>>>>>> SETUP with "Transport: RTP/AVP/UDP;multicast".
>>>>>>>> Some rtsp servers bail out otherwise.
>>>>>>>> ---
>>>>>>>>   libavformat/rtsp.c | 3 +++
>>>>>>>>   1 file changed, 3 insertions(+)
>>>>>>>>
>>>>>>>> diff --git a/libavformat/rtsp.c b/libavformat/rtsp.c
>>>>>>>> index 859defa592..3f0cbfc98b 100644
>>>>>>>> --- a/libavformat/rtsp.c
>>>>>>>> +++ b/libavformat/rtsp.c
>>>>>>>> @@ -1913,6 +1913,9 @@ redirect:
>>>>>>>>                   && (rt->rtsp_flags & RTSP_FLAG_PREFER_TCP))
>>>>>>>>               lower_transport = RTSP_LOWER_TRANSPORT_TCP;
>>>>>>>>   +        if (ff_is_multicast_address((struct 
>>>>>>>> sockaddr*)&rt->rtsp_streams[rt->nb_rtsp_streams-1]->sdp_ip))
>>>>>>>> +            lower_transport = RTSP_LOWER_TRANSPORT_UDP_MULTICAST;
>>>>>>>> +
>>>>>
>>>>> Shouldn't this take into account lower_transport_mask? E.g. it 
>>>>> should only prefer multicast if it is allowed in the mask?
>>>>
>>>> Can you explain what the lower_transport_mask is about?
>>>> I cannot think of any case where you wouldn't want use udp 
>>>> multicast when the rtsp server gives you a mutlicast-ip
>>>> in response to a DESCRIBE request.
>>>
>>> I am not familiar with the code, but it looks like it is used to 
>>> control the allowed (tried) transport methods. The rtsp protocol has 
>>> an rtsp_transport option, that is the source of the mask:
>>> https://www.ffmpeg.org/ffmpeg-all.html#rtsp
>>> If UDP multicast transport is disallowed because it is not in the 
>>> mask then you should not set it as far as I see.
>>
>> Mhm, yeah that's actually the problem I tried to fix.
>> Reading the docs again, I'm not sure if I missed it before, but the 
>> docs are stating:
>> Multiple lower transport protocols may be specified, in that case 
>> they are tried one at a time (if the setup of one fails, the next one 
>> is tried). For the muxer, only the ‘tcp’ and ‘udp’ options are 
>> supported.
>
> Are you fixing the muxer or the demuxer? So are you reading via 
> multicast or generating multicast?
I'm trying to fix the demuxer, so I'm reading a multicast udp stream 
from an IP camera.
>
>> However I still cannot manage to get udp + udp_multicast into the 
>> lower_transport_mask on the commandline, so I'm still not sure how/if 
>> it's possible to really set multiple protocols.
>
> Why not? Isn't it just specifying -rtsp_transport +udp+udp_multicast ?
I tried various separators like blank, comma, colon - now I gave your 
version a try, but it still fails.

./ffprobe -rtsp_transport +udp+udp_multicast -i rtsp://172.17.1.131/stream1m

[rtsp @ 0x55be51d7c600] Nonmatching transport in server reply
rtsp://172.17.1.131/stream1m: Invalid data found when processing input

While writing this reply, I did some more debugging:
Your version of the commandline successfully changes the value of the 
lower_transport_mask but it still just fails to probe - I'll dig into 
why it's not working as expected and come back with
my conclusions.

>
>> Therefore my patch always adds "udp_multicast" as lower_transport no 
>> matter what the lower_transport_mask is as soon as the DESCRIBE 
>> response returns a multicast IP, because
>> clients don't know when to set udp vs. udp_multicast as you can't 
>> know which stream you get.
>
> That is why lower_transport is a mask, a combination of the possible 
> values. And it looks like by default everything _is_ allowed.
My stream always fails until I specify exactly: ./ffprobe 
-rtsp_transport udp_multicast -i rtsp://172.17.1.131/stream1m
>
> Regards,
> Marton
>
>>
>>>
>>> Regards,
>>> Marton
>>>
>>>>
>>>>>
>>>>> Thanks,
>>>>> Marton
>>>>>
>>>>>>>>           err = ff_rtsp_make_setup_request(s, host, port, 
>>>>>>>> lower_transport,
>>>>>>>> rt->server_type == RTSP_SERVER_REAL ?
>>>>>>>> real_challenge : NULL);
>>>>>>>
>>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> ffmpeg-devel mailing list
>>>>>> ffmpeg-devel at ffmpeg.org
>>>>>> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>>>>>>
>>>>>> To unsubscribe, visit link above, or email
>>>>>> ffmpeg-devel-request at ffmpeg.org with subject "unsubscribe".
>>>>> _______________________________________________
>>>>> ffmpeg-devel mailing list
>>>>> ffmpeg-devel at ffmpeg.org
>>>>> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>>>>>
>>>>> To unsubscribe, visit link above, or email
>>>>> ffmpeg-devel-request at ffmpeg.org with subject "unsubscribe".
>>>>
>>>>
>>>> _______________________________________________
>>>> ffmpeg-devel mailing list
>>>> ffmpeg-devel at ffmpeg.org
>>>> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>>>>
>>>> To unsubscribe, visit link above, or email
>>>> ffmpeg-devel-request at ffmpeg.org with subject "unsubscribe".
>>> _______________________________________________
>>> ffmpeg-devel mailing list
>>> ffmpeg-devel at ffmpeg.org
>>> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>>>
>>> To unsubscribe, visit link above, or email
>>> ffmpeg-devel-request at ffmpeg.org with subject "unsubscribe".
>>
>>
>> _______________________________________________
>> ffmpeg-devel mailing list
>> ffmpeg-devel at ffmpeg.org
>> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>>
>> To unsubscribe, visit link above, or email
>> ffmpeg-devel-request at ffmpeg.org with subject "unsubscribe".
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request at ffmpeg.org with subject "unsubscribe".