Attention is currently required from: falconia.

pespin has posted comments on this change by falconia. ( https://gerrit.osmocom.org/c/libosmo-netif/+/39280?usp=email )

Change subject: bring twjit into libosmo-netif ......................................................................

Patch Set 5:

(5 comments)

File include/osmocom/netif/twjit.h:

https://gerrit.osmocom.org/c/libosmo-netif/+/39280/comment/bb8990ab_5f15f87a... : PS2, Line 35: /*! buffer depth: starting minimum */

The long-promised document has been written: […]

It probably makes sense to submit this or parts of this document in text format under doc/, so it can be updated whenever something needs to be modified/improved in the future, to keep it in sync as code documentation.

File src/twjit.c:

https://gerrit.osmocom.org/c/libosmo-netif/+/39280/comment/5a2e5e7a_3cd5537b... : PS2, Line 42: /*! Current depth as defined in twjit architecture document */

The master document is here: […]

Done

https://gerrit.osmocom.org/c/libosmo-netif/+/39280/comment/50ab06d1_62f930a8... : PS2, Line 504: rtph = osmo_rtp_get_hdr(msg);

Please see section 2.3.8 of this document: […]

Thanks. After having read your document: The fact that we may like or not like one way to generate RTP transmission (gap vs continous) doesn't mean others will be following that. Hence, for interop it still makes sense to accept the M bit as a trigger for handover. This will help for instance when reading from peers outputting RTP after a gaps which eg. keep an in-sync timestamp+seq+ssrc despite clearly specifying a gap through M bit.

Hence, I still believe this option must be implemented. As mentioned, I'm fine if this can be tweaked with a API & VTY config per twjit.

You also need to update the document, eg. clause 2.3.7 appending text: "or M (Marker) bit set". Also in clause 3.4 you may want to add a new stats "markers_received" (feel free to find a better naming), which should be increased here if the M bit is detected.

See what I'm asking afaiu is not a lot of work, only adding a new API+VTY setting + a few more lines here to check M bit and probably call analytis_init() and increment stats here, plus applying same logic eg. when SSRC changes.

https://gerrit.osmocom.org/c/libosmo-netif/+/39280/comment/ed269504_f092907a... : PS2, Line 544: if (id == INPUT_RESET) {

Please see document sections 2.3.2 and 2.3.3. […]

Done

File src/twjit_vty.c:

https://gerrit.osmocom.org/c/libosmo-netif/+/39280/comment/9695ca35_d8a3518a... : PS2, Line 41: vty_out(vty, "%s%s%s", prefix, name ? : "twjit", VTY_NEWLINE);

Can this point be marked as resolved?

Done