Nr pdcch updates

See merge request oai/openairinterface5g!514

  -- should prevent VM-creation crashes during test stages
Signed-off-by: Raphael Defosseux <raphael.defosseux@eurecom.fr>
(cherry picked from commit eeffdfa2)
Signed-off-by: Raphael Defosseux <raphael.defosseux@eurecom.fr>

issue401-nr-pucch-dmrs-format_2

See merge request oai/openairinterface5g!510

Conflicts:
	openair1/PHY/MODULATION/slot_fep_nr.c
	openair1/PHY/NR_UE_ESTIMATION/nr_dl_channel_estimation.c
	targets/RT/USER/nr-ue.c

This partially reverts commit 449da9ac.

Signed-off-by: HOAI SON DANG <sondh8@viettel.com.vn>

Merge branch 'nr_pdcch_updates' of https://gitlab.eurecom.fr/oai/openairinterface5g into nr_pdcch_updates

Nr cfo estimation

See merge request oai/openairinterface5g!499

Develop nr fixes

See merge request oai/openairinterface5g!505

reverting some changes in the UE that were introduced by nr_polar_encoder_optimizations to make UE work again on real-time testbench. Now at least the PDCCH can be detected and used for time tracking. However, the DCI is not yet decoded (this will be fixed on branch nr_pdcch_fixes). The reverted changes are most notably passing the correct slot number to slot_fep in phy_procedures_nrUE. This will have to be changed as well in the future when we pass to slot-based processing at the UE.

This reverts commit 61e36839.

Resolve "NR PDSCH: Deinterleaving and rate-matching are in wrong order"

See merge request oai/openairinterface5g!502

On the gNB side tha data are first rate-matched and then interleaved.
So on the UE side, the received LLRs have to be deinterleaved first before rate-matching.

This significantly improved the performance, especially for MCS values > 3.