I thought that would be the issue as well when I dove into this but after a 
couple of hours of log digging and tcpdumps I'm confident the  SPA 3102 is 
sending a Contact field but it is not making it back out of the sipx proxy. 
Maybe its malformed in some subtle (to me) way?

Here's an extract from the packet (as seen by a span tap) as it comes out of 
the SPA headed to sipxProxy. 

Note: I have a patch that's putting quotes around the malformed From: and 
Remote-Party-ID fields and the merged.xml I sent yesterday shows that's working 
as expected (and shouldn't affect the Contact field in any case).

-Eric

Internet Protocol, Src: 192.172.252.21 (192.172.252.21), Dst: 192.172.252.163 
(192.172.252.163)
    Source: 192.172.252.21 (192.172.252.21)
    Destination: 192.172.252.163 (192.172.252.163)
User Datagram Protocol, Src Port: sip-tls (5061), Dst Port: sip (5060)
    Source port: sip-tls (5061)
    Destination port: sip (5060)
Session Initiation Protocol
    Request-Line: INVITE sip:[email protected] SIP/2.0
        Method: INVITE
        Request-URI: sip:[email protected]
            Request-URI User Part: 0
            Request-URI Host Part: d3.foo21.com
        [Resent Packet: False]
    Message Header
        Via: SIP/2.0/UDP 192.172.252.21:5061;branch=z9hG4bK-5c3fd3ea
            Transport: UDP
            Sent-by Address: 192.172.252.21
            Sent-by port: 5061
            Branch: z9hG4bK-5c3fd3ea
        From: SMITH,A <sip:[email protected]>;tag=6e8b9fe8ed9f4d3ao1
            SIP Display info: SMITH,A 
            SIP from address: sip:[email protected]
                SIP from address User Part: 17635191497
                SIP from address Host Part: d3.foo21.com
            SIP tag: 6e8b9fe8ed9f4d3ao1
        To: <sip:[email protected]>
            SIP to address: sip:[email protected]
                SIP to address Host Part: d3.foo21.com
        Remote-Party-ID: SMITH,A 
<sip:[email protected]>;screen=yes;party=calling
        Call-ID: [email protected]
        CSeq: 101 INVITE
            Sequence Number: 101
            Method: INVITE
        Max-Forwards: 70
        Contact: "PSTN FXO1" <sip:[email protected]:5061>
            Contact Binding: "PSTN FXO1" <sip:[email protected]:5061>
                URI: "PSTN FXO1" <sip:[email protected]:5061>
                    SIP Display info: "PSTN FXO1"
                    SIP contact address: sip:[email protected]:5061
        Expires: 240
        User-Agent: Linksys/SPA3102-5.1.10(GW)
        Content-Length: 452
        Allow: ACK, BYE, CANCEL, INFO, INVITE, NOTIFY, OPTIONS, REFER
        Supported: x-sipura, replaces
        Content-Type: application/sdp
    Message Body
        Session Description Protocol
            Session Description Protocol Version (v): 0
            Owner/Creator, Session Id (o): - 138901891 138901891 IN IP4 
192.172.252.21
...

-Eric

On Jan 3, 2010, at 11:30 PM, Scott Lawrence wrote:

> On Sun, 2010-01-03 at 17:19 -0600, Eric Varsanyi wrote:
>> MikeJ of the FreeSwitch project sent me a patch (thank you!) that just
>> ignores the lack of a Contact header when processing the NOTIFY
>> message. This does fix the immediate problem (FreeSwitch doesn't SEGV
>> and I can press 'Reject' as soon as I like and the caller goes to
>> voicemail) but it still seems "wrong" (per the RFC) that the Contact
>> header is missing on the NOTIFY/Ringing... message from sipXproxy.
>> 
>> I'm happy to help (try patches/more testing/whatever), even a shove in
>> the right direction as to where to dig deeper.
> 
> You're correct that it is a bug not to send a Contact in that NOTIFY,
> but you've got the wrong culprit.  The sipXproxy is just forwarding it -
> that message comes from the Linksys and didn't have a Contact there
> either, so you'll have to take it up with them.
> 
> If you'd like, and you can't file an issue with Linksys in some public
> place, you can file an issue in the External project at
> track.sipfoundry.org to record the facts.  You've done a good job of
> documenting the problem (the messages in a sipviewer trace have frame
> numbers - it's easier to identify them that way than copying them, but
> that's a nit... good job).
> 
> 

_______________________________________________
sipx-dev mailing list [email protected]
List Archive: http://list.sipfoundry.org/archive/sipx-dev
Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-dev
sipXecs IP PBX -- http://www.sipfoundry.org/

Reply via email to