soc: samd5x: xosc32 initialization - 8 seconds start-up time

[wag-thom]

The commit 58a54e3 (@nandojve) changed in soc_samd5x.c in the osc32k_init() function the register configuration for the external 32K crystal.

zephyr/soc/atmel/sam0/common/soc_samd5x.c

Line 27 in e40fa89

| OSC32KCTRL_XOSC32K_STARTUP(6) | OSC32KCTRL_XOSC32K_CGM_XT;

Previously, the start-up value was set to 7 - which was a reserved according to the datasheet.
Now it is set to 6 - which translates to a start-up time of 8 seconds. I don't think this is a sensible value for most users.

I would suggest to either:

• change the value to 0 (Which will result in the same behavior as before the change - I get the same start-up time with 0 or 7)
• or let the user set the value with some Kconfig option

[nandojve]

Hi @wag-thom ,

We encourage community to send patches. Feel free to submit what fits better to you.
For me, both solutions are Ok.

[nandojve]

Hi @wag-thom ,
Hi @dkalowsk ,

It was given to you green light to @wag-thom that we will receive one of those solutions that was mentioned in the issue. You should keep this as an issue that will be automatically closed when you or someone send a patch to the issue. He can get more information about how submit a Pull Request in the contribute-guidelines. In special see adding-links.

Why are you looking the issue into conversations?

[henrikbrixandersen]

It was not clear to the release engineering meeting yesterday if this was being submitted as a bug or as a discussion starter. Since the original issue did not even remotely follow our bug reporting template, we decided to move it to a discussion instead.

[wag-thom]

Hi @nandojve,
Unfortunately, I am not able to contribute at the moment. I try to get permission, so I can contribute in the future.

Regarding the issue, maybe as a start setting it to 0 would be enough - then it is the same as it was before and that did not seem like a problem until now...

If there is a need to make it configurable, that can still be added later.

[nandojve]

Recap origins, see #73961