For whatever reason we need N_TA_offset to be 0 for the basic simulator
to function properly.

This is more a hack than a proper solution. More analysis is needed
to understand what happens and what should be really done.

* join pthread_FH (ru_thread) once all condition variables have been set
* join RU FHTX only if it has been started (check as for pthread_create())
* join PRACH thread in the monolithic case
* send broadcast on condition variable cond_eNBs, as there can be multiple
  waits on it
* integrate stop_ru(*ru) into kill_RU_proc(*ru)
* Correct memory freeing

LOG_F replaced by LOG_DUMPMSG. Add a generic function to dump a buffer or message. Go on with LOG_M implementation improvement. Remove redundant options to enable/disable logs.

USRP addressess and clock sources can be configured in the configuration file with the RU section with comma separated IP addresses in 'sdr_addrs' and with strings "internal"/"external"/"gpsdo" in 'clock_src'.

TODO: Currently the driver picks the two first RX and TX channels when nb_tx and nb_rx are both set to 2. If a developers wants to force usage of two devices, 'set_rx_subdev_spec' and 'set_tx_subdev_spec' can be used to indicate to the driver so that antennas from separate devices will be used. To rephrase: until 4 antennas are supported by OAI, testing with two devices, so that antennas from both devices are used, requires manual configuration of which subdevices' antennas are used.

Conflicts:
	targets/ARCH/USRP/USERSPACE/LIB/usrp_lib.cpp

log.c simplification, fix usefull functionality which never worked, move all log code from openair2 to common

Remove compilation warnings, also fix telnet server compilation failure and NB-IoT shared lib compilation failure introduced when reviewing L1 sources  organization

We have to not call that function here because
it blocks everything if number of proc > 4.