nRF52: UARTE lacks pm interface

[tautologyclub]

I noticed that drivers/serial/uart_nrfx_uarte.c doesn't initialize the power management callbacks, but it does for regular UART. However, when inspecting drivers/serial/uart_nrfx_uart.c more closely, you actually catch a bunch of hardcoded references to uart0, which leads me to believe that it's impossible to suspend uart1 under the current API.

Am I missing something here?

[tautologyclub]

So, if this is actually the case (which it seems like), in my mind it should be assigned very high priority because UART_1 is virtually unusable for nRF52 in its current state. I'd be happy to assist with the fix, but would like Nordic to comment first.

[carlescufi]

@tautologyclub I think either @jarz-nordic or @pizi-nordic can comment on this

[tautologyclub]

Looking at the HAL, it sort of looks like this would be very easy by using

__STATIC_INLINE void nrf_uarte_disable(NRF_UARTE_Type * p_reg)

and getting p_reg from

get_uarte_instance(dev)

from drivers/serial//home/benjamin/repos/zephyr/drivers/serial/uart_nrfx_uarte.c

Is this the general idea?

[jakub-uC]

@tautologyclub : UART and UARTE SHIM has been prepared by me at the same time. I had internall discussion about the approach and we decided that UARTE SHIM is multiinsance while UART SHIM is not.
The reason is quite simple: there is no and there will be no Nordic microcontroller with more than 1 UART instance.
And if by any chance this will change we will update the SHIM.

[jakub-uC]

Regarding power management for UARTE we will try to prepare a fix for that shortly.

[tautologyclub]

@jarz-nordic Thanks, that's great. Are we talking days, weeks or months?

[jakub-uC]

@tautologyclub : Once new UART API is merged I hope to have it in next 2-3 weeks.

[tautologyclub]

Any news?

[carlescufi]

@jarz-nordic any update on this?

[carlescufi]

@tautologyclub #13064