This is a static archive of our old Q&A Site. Please post any new questions and answers at ask.wireshark.org.

Bug in RTP dissector if RTP extension is present

0

It's been a while this bug is present and it's annoying as hell to explain everyone that it's not our code produces malformed packets but whireshark has this bug.

Basically, when RTP extension is present wireshark shows this:

43  17:14:58.142025 10.0.105.172    64.254.226.140  RTP PT=H264, SSRC=0xEDE18064, Seq=5, Time=0, Mark[Malformed Packet]

Here's sample pcap file with malformed RTP packet (#8)

Wireshark version: Version 1.10.2 (SVN Rev 51934 from /trunk-1.10) on Win2008

I suspect that it starts to parse RTP payload without advancing properly for extension header size, and, perhaps, it may have similar bug for CSRC fields (that's usually 0 in p2p calls)

asked 29 Sep '13, 17:38

psp80's gravatar image

psp80
1223
accept rate: 0%

edited 29 Sep '13, 21:11

can you please post a sample capture file (google docs, dropbox, cloudshark)?

And what is your OS and Wireshark version?

(29 Sep '13, 17:42) Kurt Knochner ♦

sample pcap and wireshark version info added.

(29 Sep '13, 17:59) psp80

O.K. looks 'kind of odd'. Please file a bug report at https://bugs.wireshark.org (please attach the sample capture file).

(29 Sep '13, 19:02) Kurt Knochner ♦
(29 Sep '13, 21:29) psp80

One Answer:

0

Here's the bug fix: http://pastie.org/8366148

I attached this bugfix to the bug report. By the way, if you check the pcap file, it uses our extension (it's standard now):

a=extmap:1 urn:3gpp:video-orientation

How do I teach wireshark to understand that one-byte RFC 5285 header with id#1 is 3gpp:video-orientation (as specified in sdp) and properly interpret the data of that header according to 3gpp:video-orientation?

answered 29 Sep '13, 23:15

psp80's gravatar image

psp80
1223
accept rate: 0%