[Siproxd-users] Via: rport rewriting in siproxd 0.7.1

[Peter Apian-B. <ap...@pa...>]

Hi,

problem:
login with twinkle via siproxd to callcentric.com fails with "network 
failure" on callcentric's side. Running twinkle without siproxd works.

SIP provider's answer:
callcentric's support says siproxd is not following RFC3581 regarding 
the rport parameter. As far as I understand RFC3581 this is correct, 
even if the Via header in question is only the second Via (not the 
topmost) that callcentric's server sees.

data:
The header of the outgoing REQUEST SIP packet with the MD5 hash, leaving 
my gateway, after siproxd processing, looks like:

    Via: SIP/2.0/UDP
    84.56.215.75:5060;branch=z9hG4bKb508ec9eb039c974bd1cccba6117cc33
    Via: SIP/2.0/UDP 192.168.0.11;rport;branch=z9hG4bKcykecwos
    From: <sip:XX...@ca...>;tag=mkqua
    To: <sip:XX...@ca...>
    Call-ID: tmvgrhodrsbtzae@mylocalname
    CSeq: 935 REGISTER
    Contact: <sip:XXX@84.56.215.75>
    Proxy-Authorization: Digest username="XXX", realm="callcentric.com",
    nonce="88a438c17e962decdb3091df2c300169", uri="sip:callcentric.com",
    response="XXXX", algorithm=MD5
    Allow: INVITE
    Allow: ACK
    Allow: BYE
    Allow: CANCEL
    Allow: OPTIONS
    Allow: PRACK
    Allow: REFER
    Allow: NOTIFY
    Allow: SUBSCRIBE
    Allow: INFO
    Allow: MESSAGE
    Max-forwards: 69
    Expires: 1800
    User-agent: Twinkle/1.2
    Content-Length: 0

(with userid details set to XXX). After that there's a 30 second 
silence, followed by a "network failure" packet.

My question is specific about the "rport" parameter in the second Via 
header. RFC3581 says:

"When a server compliant to this specification (which can be a proxy
   or UAS) receives a request, it examines the topmost Via header field
   value.  If this Via header field value contains an "rport" parameter
   with no value, it MUST set the value of the parameter to the source
   port of the request."

Which sounds to me (SIP newbie, UNIX/network veteran) like siproxd 
should add an rport number to the incoming packet sent by twinkle.
Furthermore, shouldn't "use_rport = 3" add a string "rport=" before the 
UDP port number ? At least the RFC seems to give that as example. - ?
Whether callcentric's server should stall at a Via header that doesn't 
matter to it may be another question, but before I nag them I would like 
to be sure that what my siproxd sends them is correct.

any insight much appreciated,
(meaculpa if this is RTFM, I tried to check the mail archives on 
sourceforge)

thanks
Peter

-- 
 pab-opto, Freiburg, Germany, http://www.pab-opto.de
 [see web page to check digital email signature]