Attention is currently required from: pespin.
neels has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-mgw/+/35176?usp=email )
Change subject: mgcp_network: Allow rx of IuUP in loopback mode with set rem IP address and unset rem port ......................................................................
Patch Set 2:
(2 comments)
File src/libosmo-mgcp/mgcp_network.c:
https://gerrit.osmocom.org/c/osmo-mgw/+/35176/comment/036f8de3_d0de7ad6 PS2, Line 866: HACK: for IuUP, we want to reply with an IuUP Initialization ACK upon the first RTP : * messag
Ah, after some more investigation, this seems to only be transmitting the exact same packet back.
This code was only active in LOOPBACK mode; since patch [1] we no longer send CRCX in LOOPBACK from osmo-msc, as we used to do for the IuUP hack. So this code is definitely not needed for IuUP init.
Also the hack itself no longer exists in osmo-mgw, since your patch [2] =) Now, instead, we have proper IuUP protocol support in osmo-mgw.
But you may be right about needing to allow IuUP Init to pass thru here. In rx_rtp(), we first do check_rtp_origin(), and then forward RTP to mgcp_iuup.c via dispatch_rtp_cb() ... I just ran a test on a nano3G and indeed the Initialization gets no ACK. I'll post this for now and test a bit further.
Does our test suite lack some important tests...?
[1] osmo-msc a8181539671f79016c968d25bb214c784ecbdc3c [2] osmo-mgw bb3ccdea1ac4d9467b2701328d6f10c4c1c8fa9a
https://gerrit.osmocom.org/c/osmo-mgw/+/35176/comment/05587916_43d7a09f PS2, Line 876: return 0; it's like this, right?
Before this patch, line 913 below disallows rx of any RTP when the MGCP configured port == 0.
This new condition punches a hole in that rejection; but only when in LOOPBACK mode.
osmo-msc does not use LOOPBACK mode anymore since march this year. Still this patch can be relevant for using an osmo-msc older than that...