Hi Tim,

As I suspected, it is empty branch is triggered by a bogus SIP flow. The ACK generated by the caller device is bogus as the TO uri is different than the TO in INVITE. Also the caller device does use branch at all in VIA and forces openser to do pre-RFC3261 transaction matching with involves the TO uri. As the log show, the ACK does not match any transaction: DEBUG: t_lookup_request: start searching: hash=11223, isACK=1 DEBUG: proceeding to pre-RFC3261 transaction matching DEBUG: t_lookup_request: no transaction found

So, the ACK is forwarded as stateless request - for this case, there is complicated alg for computing the branch param based on the branch params from the previous VIA headers. But the previous VIA is bogus (no branch at all) and openser finds itself in the impossibility to compute a branch for its VIA.

regards, bogdan

Tim Madorma wrote:

Hi Bogdan,

Attached is the log and the openser.cfg file.

thanks, Tim

On 5/30/07, Tim Madorma tmadorma@gmail.com wrote:

...

Hi Bogdan,

I have attached an html file that shows the call trace (note that the OpenSER is 65.185.232.62:5061in the flow). I will send the log file and openser.cfg in a separate email since it would put me over the 40K limit.

thanks, Tim

On 5/30/07, Bogdan-Andrei Iancu bogdan@voice-system.ro wrote:

...

Hi Tim,

the VIA branch param must not be 0, indeed. Could you provide an

network

...

trace for the entire call and the debug log for level 9?

regards, bogdan

Tim Madorma wrote:

...

Hi,

In testing with OpenSER 1.2 I noticed that the branch parameter

of the

...

...

Via in the ACK to 200 OK has a value of zero. This does not seem right. From 16.6 of RFC-3261:

    The proxy MUST insert a Via header field value into the copy
    before the existing Via header field values.  The 

construction

...

...

    of this value follows the same guidelines of Section 

8.1.1.7.

...

...

    This implies that the proxy will compute its own branch
    parameter, which will be globally unique for that branch, 

and

...

...

    contain the requisite magic cookie.

Please advise.

thanks, Tim

---

Users mailing list Users@openser.org http://openser.org/cgi-bin/mailman/listinfo/users