Attention is currently required from: falconia, jolly.

fixeria has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-bts/+/35132?usp=email )

Change subject: Transmit invalid AMR speech blocks instead of dummy FACCH ......................................................................

Patch Set 1:

(5 comments)

File src/osmo-bts-trx/sched_lchan_tchf.c:

https://gerrit.osmocom.org/c/osmo-bts/+/35132/comment/28b4f113_24e5f2f4 PS1, Line 550: FIXME: someone who knows AMR needs to look at this problem : * and decide what is the correct BTS Tx behavior for frame : * gaps in TCH/AFS. See OS#6049. This block can now be removed?

https://gerrit.osmocom.org/c/osmo-bts/+/35132/comment/a3bd575a_53d5a890 PS1, Line 571: goto send_burst Similarly to how it's done below for `gsm0503_tch_fr_encode()`, let's check the returned value. Old version of `gsm0503_tch_afs_encode()` preceding your libosmocore patch (I82ce2adf995a4b42d1f378c5819f88d773b9104a) would return -1 and leave the burst buffer unpopulated. In this case we should fall-back to generating dummy FACCH.

File src/osmo-bts-trx/sched_lchan_tchh.c:

https://gerrit.osmocom.org/c/osmo-bts/+/35132/comment/97317e92_2e7d5d18 PS1, Line 453: TCH/HS Mention TCH/AHS here? Or add a separate bullet point, like you did in `tx_tchf_fn()`?

https://gerrit.osmocom.org/c/osmo-bts/+/35132/comment/0c4be259_75680d62 PS1, Line 462: FIXME: someone who knows AMR needs to look at this problem : * and decide what is the correct BTS Tx behavior for frame : * gaps in TCH/AHS. See OS#6049. Likewise, can be removed.

https://gerrit.osmocom.org/c/osmo-bts/+/35132/comment/f91355a8_c1dcc480 PS1, Line 474: gsm0503_tch_ahs_encode Same here, we need to check returned value.