Attention is currently required from: lynxis lazus, neels.
pespin has posted comments on this change by pespin. ( https://gerrit.osmocom.org/c/osmo-hnbgw/+/40162?usp=email )
Change subject: rua: Handle event TX_DIRECT_TRANSFER in disconnected state discarding msg ......................................................................
Patch Set 1:
(3 comments)
File src/osmo-hnbgw/context_map_rua.c:
https://gerrit.osmocom.org/c/osmo-hnbgw/+/40162/comment/7d2df55d_62c2ae82?us... : PS1, Line 285: condition
We have data to be sent out on RUA, but RUA is already disconnected. […]
There can be several reasons, but the fact that we are in the state I'm explaining here still holds. eg. if there's a RUA conn going down we'd be sending an SCCP Release over the other side.
It's a race condition in the sense that ending up in this path happens given a specific sequence and timing of events, ie. CN sending some DL traffic to HNBGW in the meantime that RUA side went disconnected and Release (SCCP or Iu) was sent to CN.
https://gerrit.osmocom.org/c/osmo-hnbgw/+/40162/comment/b8b41525_361cca7a?us... : PS1, Line 290: msgb_l2len(ranap_msg), osmo_hexdump(msgb_l2(ranap_msg), msgb_l2len(ranap_msg)));
I think we should not print entire messages' hexdump on NOTICE. […]
I can move the hexdump to DEBUG. Since it's going to be debug I think it's fine printing the whole thing.
https://gerrit.osmocom.org/c/osmo-hnbgw/+/40162/comment/48abe63d_27b7b698?us... : PS1, Line 295: all
it's no longer all events
I'll change it to "Ignore events", but I think it's obvious anyway.