Attention is currently required from: dexter, laforge, pespin.
neels has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-mgw/+/31214?usp=email )
Change subject: invent fmtp 'gsm-hr-format', for RFC5993 vs 3GPP TS 101.318 ......................................................................
Patch Set 15: Code-Review-1
(2 comments)
Patchset:
PS15: my conclusion is that this patch needs more work, explained below
File src/libosmo-mgcp/mgcp_network.c:
https://gerrit.osmocom.org/c/osmo-mgw/+/31214/comment/5d099e5b_773d446b PS13, Line 692: explicitly specified the target format will be the opposite of the input : * format.
this makes no sense to me. […]
I took a look and this is what i find:
Before this patch, we have a VTY option that would enable the HR conversion for an entire trunk. This does not indicate which format should appear on which side, it always converts the format that arrives to the other format. This works when one entire side of the MGW consistently uses one HR format, while the other side consistently uses the other format. This does not work when any one BTS has a different format, or when any CN call leg happens to already send the BTS format.
A good solution is the fmtp that this patch added, because with it we can indicate on each side of each endpoint individually which format is used. If we forward a HR packet, convert only when the packet has the wrong format. The presence of the fmtp alone is not sufficient.
Other idea, relatively simple: in the very first HR packet arriving on a conn, look which format is used on that side. From then on use that format to forward future HR to that side.