Attention is currently required from: laforge, neels.
7 comments:
Patchset:
in the end, after CR, i notice that this is the iu_client. […]
yes I know, but I didn't want to have another major rewrite of code atm.
Fixing the current code to be good enough here.
File src/iu_client.c:
Patch Set #13, Line 842: LOGL_ERROR
(marking: this was ERROR)
yes, but I think error is a little bit harsh, only when a client isn't reachable.
Both cs and ps will use now INFO
Patch Set #13, Line 869: int ranap_iu_page_cs2(const char *imsi, const uint32_t *tmsi, const struct osmo_location_area_id *lai)
it would be great to have an API doc explaining how version 2 behaves differently
I added a comment, but still not describing what's the difference.
Patch Set #13, Line 879: exact
"exactly one" […]
I think this is possible for hnbs and not forbidden.
Patch Set #13, Line 883: if (!osmo_lai_cmp(&entry->rai.lac, lai)) {
(early exit: 'if (osmo_lai_cmp(.. […]
What do you mean?
I think it is ok, to break here. We tx a paging request to a single RNC about this LAC. So breaking this inner llist_for_each is correct.
if it's exactly one, then it seems that this loop could also get an 'if (paged) break;' ?
No, we need to tx a page request to a RNC at most once.
But we don't want to exit, after paging a single RNC.
Patch Set #13, Line 914: int ranap_iu_page_ps2(const char *imsi, const uint32_t *ptmsi, const struct osmo_routing_area_id *rai)
api doc
In the header. I don't see a reason, why to put it into the .c file.
If it's in the header, IDEs can parse it and use it.
To view, visit change 38946. To unsubscribe, or for help writing mail filters, visit settings.