Attention is currently required from: falconia.

View Change

2 comments:

• File src/osmo-bts-trx/sched_lchan_tchf.c:

  • Patch Set #1, Line 500:

    what is the correct BTS Tx behavior for frame
    		 * gaps in TCH/AFS
    

    I guess we could indicate a bad frame via AMR's in-band signalling? Just an idea.
    IIRC, this is what ip.access nanoBTS does in the absence of RTP frames.
    See also: https://osmocom.org/issues/4823.

  • Patch Set #1, Line 505: gsm0503_tch_fr_encode

    I suggest using a switch here:

    ```
    switch (tch_mode) {
    case GSM48_CMODE_SPEECH_V1:
    case GSM48_CMODE_SPEECH_EFR:
        rc = gsm0503_tch_fr_encode(*bursts_p, NULL, 0, 1);
        if (rc == 0)
            break;
        /* fall-through */
    default:
        gsm0503_tch_fr_encode(*bursts_p, dummy, sizeof(dummy), 1);
        chan_state->dl_facch_bursts = 8;
    }
    ```

    The reason why I am suggesting this, specifically the fall-through if `rc != 0`, is that the user may be running an older version of libosmocore, predating your patch `Iade3310e16b906efb6892d28f474a0d15204e861`. In this case `gsm0503_tch_fr_encode()` would return -1 and we will fall-back to FACCH rather than sending garbage.

To view, visit change 33069. To unsubscribe, or for help writing mail filters, visit settings.