Attention is currently required from: laforge, pespin. neels has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-hnbgw/+/28816 )
Change subject: add ps_rab_ass FSM to map GTP via UPF ......................................................................
Patch Set 2:
(3 comments)
File include/osmocom/hnbgw/hnbgw.h:
https://gerrit.osmocom.org/c/osmo-hnbgw/+/28816/comment/d54c4fe7_c1f5b210 PS2, Line 193: /* Return true when the user configured GTP mapping to be enabled, by configuring a PFCP link to a UPF.
I we should not allow running osmo-hnbgw without proxying GTP through it (in detail, through its ass […]
i disagree, osmo-hnbgw has so far always run without proxying/mapping GTP. we are adding a new feature to use a UPF, and should still allow osmo-hnbgw to run the same way it always did. Firstly to allow users to keep their current installations working, secondly it also makes sense in lab or non-complex sites to just pass GTP through without a UPF.
The same is true for RTP proxying, next time i test 3G with a lab setup i will be annoyed that the hnbgw *requires* an MGW, and will rather submit a patch to make MGW optional and allow running as it did before RTP proxying
File include/osmocom/hnbgw/hnbgw.h:
https://gerrit.osmocom.org/c/osmo-hnbgw/+/28816/comment/1f75cf19_ef2877b5 PS1, Line 193: static inline bool hnb_gw_is_gtp_mapping_enabled(const struct hnb_gw *gw)
can you document this function? what does gtp_mapping mean exactly?
"mapping" == "proxying" ... should we decide on one of those terms?
File src/osmo-hnbgw/hnbgw.c:
https://gerrit.osmocom.org/c/osmo-hnbgw/+/28816/comment/43e9c42d_e3e95728 PS1, Line 230: LOGP(DHNBAP, LOGL_INFO, "deallocating UE context: id 0x%x, imsi %s, tmsi 0x%x\n",
AFAICT this was not resolved.
sorry i was distracted by a visitor and lost track
-
neels