neels has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-bsc/+/20924 )
Change subject: bsc_subscr_conn_fsm: use cause code PREEMPTION on lchan loss ...................................................................... Patch Set 2: (1 comment) https://gerrit.osmocom.org/c/osmo-bsc/+/20924/2/src/osmo-bsc/bsc_subscr_conn_fsm.c File src/osmo-bsc/bsc_subscr_conn_fsm.c: https://gerrit.osmocom.org/c/osmo-bsc/+/20924/2/src/osmo-bsc/bsc_subscr_conn_fsm.c@377 PS2, Line 377: gsc > I think the problem is that this code doesn't distinguish on _why_ the lchan > disappeared. […] note the gsm_lchan->release.*cause* items. They are intended to store the individual failure causes to be sent back. maybe gsm_subscriber_connection needs similar fields? So whenever some release reason shows up, store that in gsm_subscriber_connection.clear_cause (to be added), so that gscon_bssmap_clear() can pick them up (and maybe drop the cause argument to gscon_bssmap_clear()). but the real work needed is to figure out which failure should emit which cause and match that up with the implementation. previous osmo-nitb code has been super vague and confused about error codes. i have tried to edge into the right direction there, but so far i have been shying away from doing that properly because it feels like the proper cause codes are not specified well (or at all besides guessing from their names) and that there are too many cases to fix everything at once. so i'm always glad when some customer or test case is clear on expecting a specific cause code so that we can arrange the failure code paths to emit the right sane cause code instead of just "equipment failure" everywhere. i agree it is not helping to change the blanket "equipment failure" to a blanket "preemption", but rather we need fine-grained causes, maybe defaulting to "equipment failure" where we lack better knowledge. -- To view, visit https://gerrit.osmocom.org/c/osmo-bsc/+/20924 To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings Gerrit-Project: osmo-bsc Gerrit-Branch: master Gerrit-Change-Id: I88b47695877a31bc5ede1080027c9c6080eb2090 Gerrit-Change-Number: 20924 Gerrit-PatchSet: 2 Gerrit-Owner: dexter <[email protected]> Gerrit-Reviewer: Jenkins Builder Gerrit-Reviewer: neels <[email protected]> Gerrit-CC: laforge <[email protected]> Gerrit-Comment-Date: Thu, 29 Oct 2020 22:43:11 +0000 Gerrit-HasComments: Yes Gerrit-Has-Labels: No Comment-In-Reply-To: laforge <[email protected]> Gerrit-MessageType: comment
