fix bug from this MR

fix bug from this MR

stupid-mistake

- Sin and Cos values are obtained from a table with a resolution of
  100.
- dlsim is updated to use this table for phase noise.

This commit contains work done by Eurecom and by Benetel.

FEATURE STATEMENT:
- Introduce linear phase error noise model in Uplink at UE
- Perform common phase error (CPE) estimation and compensation at gNB

SOLUTION:
- A linear phase shift model is introduced in simulation.
- PTRS symbols are used to perform estimation of CPE from DMRS compensated signal
- The estimated values are interpolated in time domain and signal is compensated for the CPE.
- PTRS processing is done in Frequency Domain for each symbol in a slot and
  LLR's are calculated for each symbol accordingly.

IMPLEMENTATION:
* sim.h/channle_sim.c
- Linear Phase Noise Generation model definition.
* nr_ul_channel_estimation.c/nr_ul_estimation.h
- CPE estimation from PTRS and DMRS compensated signal.
* nr_dmrs_rx.c/nr_refsig.h
- Regenerate PTRS symbols at gNB.
* nr_ulsch_demodulation.c
- Removed old PTRS processing code and move to a common PTRS processing function
* defs_gNB.h/init.c
- New PTRS variables definition and initialization
* nr_ulsch_ue.c
- Corrected PTRS parameter to get new PTRS symbols for each OFDM symbol

TESTING
* ulsim.c
- Added Phase noise, Enable PTRS signal and verified the output.

VERIFICATION
- The LLR are rotated back with estimated CPE and no error is observed in scrambling/decoding

fix phy simulators build Pb: move snr_dB global to simulation code, add a channel mode section in configuration for snr_dB ,add a parameter flag in the config module to allow omitting prefix when modifying a parameter (for compatibility of -s option)

add channel model option to rfsimulator and prevent common libraries re-compilation when using both --UE --eNB options in build_oai script

add config options for rfsimulator, set a common lib for l1sim and rfsimulator, avoid simulation sources to be added twice in some execs (tau.c)

For whatever reason most of the files had their permission
changed from 644 to 755, which is not wanted.

not all regression tests done (had to touch some header files)