Attention is currently required from: fixeria, jolly, laforge, pespin.
1 comment:
File src/twjit.c:
Patch Set #2, Line 504: rtph = osmo_rtp_get_hdr(msg);
I wouldn't mind getting rid of ortp and switching to something else. […]
Regarding the question of what to do in OsmoBTS, I really prefer to tackle that question *after* the present addition of twrtp to libosmo-netif is merged, not before. The present concern of extending libosmo-netif should be independent of what we ultimately decide to do in OsmoBTS.
Now back to the original thread question of adding support for M bit in twjit: I am now ready to work on this task, and I have an additional question to @pespin@sysmocom.de. My current idea is to add the optional check of M bit to `check_input_for_subbuf()` function - I hope you'll agree it seems like the best place. The question is: should this M bit check go before or after the currently implemented `if (ts_delta < 0)` check? Consider this scenario: a packet comes in M bit set, but the timestamp is in the past relative to `head_ts`. Should the past timestamp take precedence (discard the packet as too old), or should the M bit take precedence and invoke reset in HUNT or handover in FLOWING? You might wish to consider scenarios in which marker and non-marker packets get reordered by the IP network, if such reordering is possible in those worlds where you say M bit handling is needed.
From my PoV, I treat this M bit handling as a customer-requested feature, and you are the customer-like requestor - hence I need you to tell me exactly how you want it to work.
To view, visit change 39280. To unsubscribe, or for help writing mail filters, visit settings.