Hi René, Please try the last SVN trunk - I made a fix on SVN trunk and if ok, I will backport it to 1.5 also. Test it please and let me know.
Regards, Bogdan Rene de Vall wrote: > Hi Bogdan, > > ...sorry fo rmy late reply, we tested with svn branch 1.5 rev. 5768 > and it didn't work. > > It seems that the check in drouting.c > > "if( type==pgwa->type && > (pgwa->port==0 || pgwa->port==msg->rcv.src_port) && > ip_addr_cmp(&pgwa->ip, &msg->rcv.src_ip) )" > > fails because of the fact, that pgwa->port is not 0 but 5060 at this > point, which might be caused by > > "if (port && *port==0) > *port = (is_sips||((*proto)==PROTO_TLS))?SIPS_PORT:SIP_PORT;" > > in resolve.c which is (eg. for me it seems to be) called during > add_dst in routing.c. > > I very much appreciate your efford to fix this, many thanks in advance. > > Best regards, > > René > > > > 2009/6/1 Bogdan-Andrei Iancu <[email protected] > <mailto:[email protected]>> > > Hi Rene, > > I made the fix on the SVN trunk and 1.5 - upload and try again, > > > Regards, > Bogdan > > Rene de Vall wrote: > > Hi Bogdan, > > ...yes, you're right, the gateway "behaves asymmetric", when > it sends out the request it uses just the next port (socket) > which the underlying operating system (IOS, in this case) > offers as source port. > > Your proposed fix sounds goof to me, as it would work for all > types of gateways, no matter if they "behave symmetric" or > "asymmetric". If one needs to diversify a gateway which uses > one IP and different ports, she can always add the port to the > datafilled address in table dr_gateways... > > Thank you for your fast response and best regards, > > René > > > _______________________________________________ Users mailing list [email protected] http://lists.opensips.org/cgi-bin/mailman/listinfo/users
