There was a bug. Uplink MAC PDU were corrupted. It turns out
that at the end of the processing of an LDPC segment, the decoded
data is copied to the output buffer, but! with a wrong size, bigger
than it should. The bug I saw was with three segments: 0, 1, and 2.
Segment 1 was finished first, then 0, then 2. When copying segment
0, the beginning of segment 1 was overwritten with the end
of segment 1 (which contains non-data bytes but CRC bytes then 0s),
leading to a corrupted MAC PDU.

Let's copy only the decoded data bytes instead.

MR890 : moving the OFDM phase rotation from RU to DU  
Moves the recently introduced phase rotations from RU to DU. 
This is important for setups that do not use the monolithic mode (like Benetel RRU).

MR944 : fix pdsch low prb 
Fix for DLSCH decoding for low TBS and for ULSCH. 
In the UE, DLSCH decoding stats are printed periodically for every 64 frames. 
DLSCH decoding is skipped if UE receives retransmission for successfully decoded CW.
 
MR946 : S1ap bugfix rab setup  
Fix a bug in S1AP

MR951 : CI NSA FR1 logging and reporting update  
Enhancement of gNB html + log report showing (RAPROC PUSCH check, dlsch and ulsch statistics)

M954 : Mac fixes wk45 2
Hotfix: preserves "low" delay (20-30ms) and correct HARQ retransmissions by disabling timing advance (TA)
Reduces log level, compiler warning, correct TA condition

MR936: nr_fdd_if_fix
- Add support for 5G NR FDD bands and introduce configurable IF frequencies (esp. on nrUE) with configurable UL-DL-frequency offset. 
 
MR940: dongzhanyi-zte-develop2  
- Solved the problem that UE UL LDPC code can only process data within 8 code blocks (the UE UL LDPC code could only process data within 8 code blocks. If there were more than 8 code blocks, an error would occur)

MR943: NR_UE_MAC_scheduler
- Dynamic calculation of slot for PUSCH transmission based on slot offset k2. PUSCH slot is no longer hard-coded to 8.
- Support for multiple UL allocations in a frame by supporting a ul_config_request per UL slot.
- Support for multiple PDUs per slot
- Fixes for a few UE crashes/assertion failures due to invalid DCI received
- UE scheduler code clean-up

- wrap around 1024 (frame number is 0-1023)
- check for equality of frames with high numbers will trigger every 10ms

This reverts commit 199ce5d953ed90e052a69f520b30b7253b163250.

If UE receives retransmission for already decoded PDSCH, it skips decoding and sets ACK directly.
And adding PDSCH decoding stats and printing then every 64 frames for CI purpose

The BG was set to its default for rounds > 0. This is now fixed by adding it to harq structure.
Also a fix for wrong I_lbrm set in dlsch decoding.

Integration 2020 wk45 (2) comprising:

MR937:Nr dl sched fixes  
-restore old phytest scheduler behavior
-correct TA scheduling
-correct SSB
-shorten interface of `nr_fill_nfapi_dl_pdu()`
-Track RB allocation on symbol basis (bitmap)
-add a `remove_nr_ue_list()`, could be used instead of MR !934

MR910:5G SA NAS messages 
- 5G SA nass messages and a simple unitary test code
 
MR927:Nr rrc configuration fr2
- introduces some bug fixes for FR2 that might also have an impact on FR1.

Signed-off-by: Raphael Defosseux <raphael.defosseux@eurecom.fr>

 Conflicts:
	openair2/LAYER2/NR_MAC_gNB/gNB_scheduler_bch.c