the same code present in the normal HAL-SITL scheduler was not copied across when this method was created.
Without it, if a non-main thread called delay we will attempt to do things like update the sitl_model (ie. SIM_Aircraft) object. We have no protections in place (nor should we have) for this happening when the main thread is in the middle of doing the SITL updates, so corruption is definitely going to happen at some stage!
this is the loop which ensures the amount of data sent to the mavlink client (usually Python) is limited - if we don't do this then we lose vast amounts of data when running at large speedups.
By attempting to process inbound data we may realise that the TCP connection has been dropped, and in that case we will start to listen for another connection.
This allows you to terminate the sim_vehicle.py MAVProxy and have it automagically restart (when running under GDB). This is very useful for testing MAVProxy patches with SITL; it's a different workflow to opening an output and connecting a new version of MAVProxy to that outout.
Removes busywait for simulation state packet, dramatically reducing CPU
usage. The dominant wait time in AP_Periph is 1024 microseconds as this
is the default value of HAL_PERIPH_LOOP_DELAY_US, so a 1ms wait is
unlikely to be a problem.
Add code to reflect USB ACM parity setting to the passthrough port alongside existing support for ACM baud rate changes. Some use cases for serial passthrough require specific parity settings.
For example, even parity is used and required by the USART protocol used in the STM32 system bootloader. This enhancement allows the use of standard flash programming tools such as STM32CubeProgrammer to flash connected STM based peripherals such as Receivers and Telemetry radios via serial passthrough. Some examples of such peripherals include the FrSky R9 receivers as well as various other STM based LoRa modules used by the mLRS project.
Code remains to check for use of these options and give an error which
contains the correct option to use, due to the non-intuitive mapping
between --uartX and --serialN.
A future version will remove that code too.
