samples: nrfx_prs: Fix sample frequency and baudrate

[joerchan]

Fix sample frequency and baudrate.
The nrfx drivers are now using raw integere values instead of nrf definitions.

Add handling of SCK initial value based on the SPI mode used.
If mode is using CPOL then SCK needs to be set.

Fixes: #71154

[anangl]

This is incorrect. The field here is still of the enumerated type, so this line should stay untouched.

The UARTE initialization fails due to a different reason - the RXSTARTED event appears to be set because its register is at the same location as the SPIM event STARTED. So we'll need something like:

	/* Make sure all SPIM events are cleared. */
	nrfy_spim_int_init(spim.p_reg, 0xFFFFFFFF, 0, false);

in the if (spim_initialized) block right above and a similar clean-up in switch_to_spim().

[joerchan]

You're right. thanks.

I tested with your workarounds an it works.

However to me it seems like this should be part of nrfx uninit function.
Clearing the event registers seems like an important step. For example if these functions are used in this scenario, or between images in a bootchain, like mcuboot -> tf-m -> application.

It can be added here as a temporary workaround until a new release of NRFX of course.

[joerchan]

@anangl Added workarounds.

I think the proper fix is to make sure that events get cleared in the nrfx_<peripheral>_uninit functions.
That seems important for a lot of cases, especially boot-loader chained boot (including with TF-M).

[anangl]

@anangl Added workarounds.

Did you forget to push the branch?

I think the proper fix is to make sure that events get cleared in the nrfx__uninit functions. That seems important for a lot of cases, especially boot-loader chained boot (including with TF-M).

I completely agree, this should be ultimately addressed in nrfx.

[joerchan]

Did you forget to push the branch?

Sorry, pushed now.