Test development using CLI/python/RTT

[Tim Adu]

Hi,

I'd like to start developing automated tests for my custom NRF52840 board.

Currently I use the CLI library (with RTT backend) for communicating with the board.

Any ideas for how I could communicate with my board via a python script?

Thanks

Tim

[Jørgen Holmefjord]

Hi Tim,

If you want to continue using CLI over RTT for your communication with the board, you can look into our pynrfjprog library. This enables you to read and write RTT channels from a python script.

Best regards,
Jørgen

[Tim Adu]

Hi Jørgen,

I would like to avoid the middleman (CLI) and just use pynrfjprog directly.

1. I plan to use the pynrfjprog read and write functions directly as you suggested. However, on the firmware side, I'm considering one potential issue: my logger uses the RTT backend. Would I have to disable  the logger?
2. As for implmentation, would it make sense to use a repeating timer to call SEGGER_RTTRead() and have a task (I'm using FreeRTOS) decode the command, since there is no interrupt to tell when data is received in the RTT buffer?

Tim

[Jørgen Holmefjord]

1. RTT supports mulitple channels. If you want to keep logging, use another channel than 0 for the test communications.
2. Yes, this makes sense to me.

[Tim Adu]

What does it take to use a different channel? I just read this post: https://devzone.nordicsemi.com/nordic/nordic-blog/b/blog/posts/multiple-channels-in-rtt

Does it paint the full picture?

[Tim Adu]

Okay, so I tried using Channel 1 (both in my python script and firmware), Everytime I run the script, I hit a breakpoint in the code. Specifically, this breakpoint (in cmsis_gcc.h) is hit everytime I send a string from the script:

__attribute__((always_inline)) __STATIC_INLINE void __SEV(void)
{
  __ASM volatile ("sev");
}

And the program halts. Is this expected?