[Freeswitch-users] sdp header rewrite

[Gabriel Kuri]

I ran into this posting which is similar, although not exactly the same,
as the failure mode I'm experiencing.

    http://bugs.digium.com/view.php?id=11483

following what this other person tried to temporarily fix the issue, I
changed the name of the rtpmap on the linksys spa from G729a to G729 and
it works - FS no longer transmits an audio port of 0 in the sdp headers
when inbound-proxy-media and late-negotiation are enabled.

correct sdp header excerpt on a call ...

v=0.
o=01Nextone 2341985734634606731 5798373005113647141 IN IP4 YY.YY.YY.YY.
s=sip call.
c=IN IP4 YY.YY.YY.YY.
t=0 0.
m=audio 25454 RTP/AVP 18 101.
a=rtpmap:18 G729/8000.
a=fmtp:18 annexb=no.
a=rtpmap:101 telephone-event/8000.
a=fmtp:101 0-15.
a=ptime:20.

So is this an underlying issue with the linksys spa units or FS?

Gabe


Gabriel Kuri wrote:
> I'm having an issue with the linksys spa devices when enabling inbound
> proxy media mode (inbound-proxy-media=true) and late negotiation
> (inbound-late-negotiation=true) in the sofia profile. The spa
> immediately sends a BYE when the call is answered by the called party.
> For whatever reason, it works fine between two linksys devices directly
> connected to FS, but when the call goes out to the PSTN via the SIP
> provider, the spa isn't happy and sends a BYE.
>
> After comparing the raw SIP packets on the wire (tcpdump) and between
> enabling/disabling proxy-media mode and late negotiation, the only
> difference I notice is the port in the m= line of the SDP header.
>
> According to the freeswitch log, the rtp port would be rewritten to
> 28044 in the sdp header of the SIP packet sent to the spa device. But on
> the wire, the port is rewritten to 0, which I'm guessing is why the spa
> isn't happy and sending a BYE.
>
> Here's the excerpt from the freeswitch log showing FS rewriting the port
> to 28044 for the packet going to the spa device.
>
>
> [DEBUG] sofia_glue.c:1003 sofia_glue_tech_patch_sdp()
> sofia/internal/<phone_number_removed>@mydomain.net Patched SDP
> ---
> v=0
> o=01Nextone 3587 27824 IN IP4 XX.XX.XX.XX
> s=sip call
> c=IN IP4 XX.XX.XX.XX
> t=0 0
> m=audio 24174 RTP/AVP 18 101
> a=rtpmap:18 G729/8000
> a=fmtp:18 annexb=no
> a=rtpmap:101 telephone-event/8000
> a=fmtp:101 0-15
> a=ptime:20
>
> +++
> v=0
> o=01Nextone 3587 27824 IN IP4 XX.XX.XX.XX
> s=sip call
> c=IN IP4 YY.YY.YY.YY
> t=0 0
> m=audio 28044 RTP/AVP 18 101
> a=rtpmap:18 G729/8000
> a=fmtp:18 annexb=no
> a=rtpmap:101 telephone-event/8000
> a=fmtp:101 0-15
> a=ptime:20
>
>
> However, the packet on the wire reveals FS rewriting the port to 0  ...
>
> v=0.
> o=01Nextone 7852943629956191733 8120394851828294756 IN IP4 YY.YY.YY.YY.
> s=sip call.
> c=IN IP4 YY.YY.YY.YY.
> t=0 0.
> m=audio 0 RTP/AVP 96 101.
> a=rtpmap:96 G729/8000.
> a=fmtp:96 annexb=no.
> a=rtpmap:101 telephone-event/8000.
> a=fmtp:101 0-15.
>
> Is this a bug or is there some other problem?
>
> Thanks for the help,
> Gabe
>
>
>
>
>
>
>