Attention is currently required from: pespin.
fixeria has posted comments on this change by pespin. ( https://gerrit.osmocom.org/c/erlang/osmo-s1gw/+/37923?usp=email )
Change subject: erab_fsm: Abort and reply call with error if unable to establish PFCP sess ......................................................................
Patch Set 1:
(2 comments)
File src/erab_fsm.erl:
https://gerrit.osmocom.org/c/erlang/osmo-s1gw/+/37923/comment/b294973f_390ac... : PS1, Line 176: next_state, ?FUNCTION_NAME, S, %% loop transition to enable state_timeout
Yes, that sounded weird to me at the time when you said so. […]
Ok, I see you have also changed the loop transition to `keep_state` in the `pfcp_peer` (Ic17c1c7210ab4ae0ddc840a5ffd13b7cdda243c9). I gave it a try and indeed it's working as expected: the peer re-tries association.
This is weird, because I spent quite some time debugging and figuring out why state timeout does not work, e.g. here: https://gerrit.osmocom.org/c/erlang/osmo-s1gw/+/37265. Maybe it was a bug of older Erlang/OTP version, or this problem occurs only under some conditions.
I would keep it as-is for now, until we know more about this problem. Strictly speaking, changing from `next_state` to `keep_state` is not needed to get things done in this patch. You're just doing this by the way since you're touching the related code path, but this is clearly not a must in this case.
https://gerrit.osmocom.org/c/erlang/osmo-s1gw/+/37923/comment/763ea469_5e49a... : PS1, Line 179: {stop_and_reply, (cosmetic) Also, please move the case body to the next line for consistency with the existing code (like done above).
``` {error, Reason} -> {stop_and_reply, ... ```