Attention is currently required from: pespin.

View Change

2 comments:

• File include/osmocom/trau/trau_rtp.h:

  • Patch Set #1, Line 27: uint8_t rtp_extensions;

    missing entry in TODO-RELEASE file stating ABI change.

    Will add in the next patch iteration.

• File src/trau/trau_rtp_conv.c:

  • Patch Set #1, Line 291: crc_bfi = true;

    why is all the code path logic changing here? now if it's considered a bgi, then all the computation […]

    With TW-TS-001 it is now possible to emit BFI-with-data in RTP, just like in TRAU-UL frames. In the case of EFR, if the CRC is good but C12 is set (BTS emitted BFI marker, rather than transmission errors on Abis), we emit BFI-with-data in RTP: the whole point of TW-TS-001, aside from preserving TAF bit. Now the question is: if the TRAU-UL frame came in with CRC errors, should we also emit BFI-with-data, or emit a No_Data frame?

    TW-TS-001 allows either option: see Annex C in the spec. However, when I wrote TW-TS-002 spec for HR codec, I specified (section 6.2.3) that bad CRC-3 in the TRAU frame shall result in No_Data output. Hence thinking about it now, I'll make the EFR case consistent.

    I also just noticed that my code is currently plain-broken here in that I don't actually set the BFI bit in the TEH octet - will fix.

To view, visit change 37156. To unsubscribe, or for help writing mail filters, visit settings.