Attention is currently required from: laforge, pespin, daniel.

arehbein has posted comments on this change. ( https://gerrit.osmocom.org/c/libosmo-netif/+/33652 )

Change subject: ipa: Add segmentation callback ......................................................................

Patch Set 5:

(3 comments)

File include/osmocom/netif/ipa.h:

https://gerrit.osmocom.org/c/libosmo-netif/+/33652/comment/6a49898d_bb42ab28 PS4, Line 27: #define msgb_ipa_proto(__x) OSMO_MSGB_IPA_CB(__x)->proto

osmo_msgb_ipa_cb_proto […]

I put `osmo_ipa_msgb_cb*`, because it's in `ipa.h` and not in `msgb.h` and because to me it seems less confusing to stick to filename prefixes.

File src/ipa.c:

https://gerrit.osmocom.org/c/libosmo-netif/+/33652/comment/80f13ef4_74d76a85 PS2, Line 378: MSG_CB_IPA_PROTO_EXT_IS_SET_OFFSET

We wouldn't need the IS_SET flag anyway, I'd expect proto_ext to only have a meaning if proto is IPA […]

Concerning the idea of a struct, I deliberately did not go for a struct in order to avoid violating the strict aliasing rule, since `msg->cb` is an array of `unsigned long`. Seems like we don't care/looks like we use `-fno-strict-aliasing` everywhere anyways, so I'll keep that in mind. At least now I know where we stand on this

https://gerrit.osmocom.org/c/libosmo-netif/+/33652/comment/639e4653_5142065a PS2, Line 432: /* Below: Helper functions for addition of send_ipa functionality in later commit */

Iirc the idea was to introduce ipa_check_pull_headers() AND use it in the segmentation_cb(). […]

Done