Attention is currently required from: pespin. laforge has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-bsc/+/27818 )

Change subject: lchan_fsm: Ignore other SAPIs of RLL_REL_IND for SAPI=0 is received ......................................................................

Patch Set 1:

(2 comments)

File src/osmo-bsc/lchan_fsm.c:

https://gerrit.osmocom.org/c/osmo-bsc/+/27818/comment/0f709e7c_c59ab36f PS1, Line 1207: /* if SAPI=0 is gone, it makes no sense if other SAPIs are still around,

Probably worth adding OS#XYZ here (I recall I saw a ticket a few mins ago?)

can do, but it's in the commitlog as Related.

https://gerrit.osmocom.org/c/osmo-bsc/+/27818/comment/9f2c23ba_a7f2b2f8 PS1, Line 1210: * seem to send a RLL_REL_IND for SAPI=3 if there was already one for SAPI=0 */

this new code added seems to be only applied for TCH according to comment in line 1201. […]

it is _not_ applied for TCH, I guess the rationale is that there may still be speech frmames that we can decode? That also sounds fishy to me (I didn't write that code), but it is a separate issue to resolve.

The pathc has been deployed in production and the SDDCH which previously got stuck in ESTABLISHED did no longer get stuck.