Attention is currently required from: pespin.
laforge has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-hnbgw/+/36081?usp=email )
Change subject: Introduce concept of per-HNB persistent data structure ......................................................................
Patch Set 3:
(1 comment)
File src/osmo-hnbgw/hnbgw_vty.c:
https://gerrit.osmocom.org/c/osmo-hnbgw/+/36081/comment/0aabdb1d_8b2ff2a6 PS2, Line 825: "hnb UMTS_CELL_ID",
Also not blocking, but I have to say I'm not really liking this concept of "persistent structure" you are adding. I'd probably go for something more similar to what we do with ASPs in libosmo-sccp. The same structure is used to handle everything related to that entity, and if it is dynamically created then a "hnb->dynamic=true" is added to it, so that it is not written back to the config file.
I actually like the approach here. Sure, one can rename structures to improve naming if that's a concern. But fundamentally the persistent ddata like counters is different from the dynamic (connection) part. I see no reason to mix those two.
btw: in libosmo-sigtran the dynamic ASP is also gone after it has disconnected, therefore no persistent counters. So it's not a solution.
Having another syntax to enter the hnb node on the VTY is fine with me, and one can even have multiple different formats to do that. But let's focus on the important part of implementing the persistent counters/statistics first.