Bug #4745
open
Related function where the address is added to o=: mgcp_write_response_sdp()
I think we need to clarify what is correct before jumping to any implementation
dexter, neels - what is your take on this?
I think you are right. The o= field should indeed contain the IP address of the party that started the session. The osmo-mgcp-client seems to be correct already. The MGW is the only party that would need fixing. It should be simple to fix since its just meta data. We are not using the contents of the o= field to get the RTP ip address, that info is in the c= field.
No the question is: Who begins the session. Is it the party that sends the CRCX without SDP or is it the MGW that assigns the session ID?
RFC 4566, capter 5.2 says: "For privacy reasons, it is sometimes desirable to obfuscate the username and IP address of the session originator. If this is a concern, an arbitrary <username> and private <unicast-address> MAY be chosen to populate the "o=" field, provided that these are selected in a manner that does not affect the global uniqueness of the field." This would mean that if we return something like 255.255.255.255 in the field we are still spec compliant since the session id still ensures uniqueness.
RFC 3435 "3.4 Encoding of the Session Description (SDP)":
* The address SHOULD be the same address as provided in the
connection information ("c=") field.
Also available in: Atom
PDF