[FEATURE] New Modbus implementation for NuttX

[raiden00pl]

Is your feature request related to a problem? Please describe.

After struggling with the FreeModBus implementation available in NuttX I decided to write my own implementation of Modbus stack.

The FreeModBus in my opinion is very poor quality and NuttX deserve something better.

Problems I want to solve:

• [BUG] Modbus RTU slave response time is too high #3037
• follow NuttX coding standard
• clean API for callback registration
• support for Modbus TCP
• Modbus TLS someday
• support for many Modbus instances on the same firmware
• support for hardware-accelerated Modbus handling available in some USART peripherals
• support for not standard interfaces (Modbus over BLE or other weird things)

Tools:

• modbus slave example common for Modbus RTU and Modbus TCP in examples
• modbus master client common for Modbus RTU and Modbus TCP in system

Demos:

• Modbus RTU with simulator (master and slave talk over simulated UART)
• Modbus TCP with qemu (master and slave talk over host network)

[cederom]

I will be using ModBus/RS485/Master in a few days too, let me know if you need support / testing and if you found a better solution @raiden00pl :-)

Particularly we may want to have two modbus instances running in a master-slave proxy with read/write intercept capabilities :-)

I was playing with ESP-IDF esp-modbus that was using FreeModbus and it was kinda pita :D

[raiden00pl]

@cederom Of course, the more people involved in testing, the better. I already have it somehow working, but so far I've been doing all my testing on the host (qemu and SIM). The library passes the tests I previously wrote for the freemodbus implementation. I haven't tested it on real hw yet