Attention is currently required from: wbokslag. laforge has posted comments on this change. ( https://gerrit.osmocom.org/c/osmo-tetra/+/29387 )
Change subject: added support for fill bits and basic link defragmentation (also thanks to SQ5BPF) ......................................................................
Patch Set 3:
(3 comments)
File src/tetra_upper_mac.h:
https://gerrit.osmocom.org/c/osmo-tetra/+/29387/comment/fd4e73ca_93018ea3 PS3, Line 13: in
You're right, it's kind of underdocumented. I'll make sure to annotate the fields. […]
this probably shows how our practices have evolved during the last decade. using bool in C might not have been advisable with comlers used 10+ years ago, but today it clearly is the right thing. please modify, thanks!
File src/tetra_upper_mac.c:
https://gerrit.osmocom.org/c/osmo-tetra/+/29387/comment/7e02e433_691dc283 PS3, Line 44: s
The lower mac allocates a global context structure tetra_cell_data with global pointer tcd. […]
OK, then lets keep the global/static variables for now, but add a TODO comment that this actualY should be in some context
https://gerrit.osmocom.org/c/osmo-tetra/+/29387/comment/68417ba3_a33b8cf0 PS3, Line 50: fragslots[i].msgb = msgb_alloc(FRAGSLOT_MSGB_SIZE, "fragslot");
You're right, I could allocate them on demand, whenever a fragmented resource is encountered on a slot. I'll do that in the next patch set
I'm not saying it must be done that way, was just wondering if there was some specific reason.