soc: kinetis: full pinctrl enablement for Kinetis boards

[danieldegrasse]

This PR enables pinctrl for all drivers used by NXP Kinetis SOCs, and disables pinmux drivers for in tree Kinetis boards. The in tree pinmux driver is still present, but is no longer enabled at the SOC level. All in tree Kinetis boards should now be converted for pinctrl support.

Dependent on #42238 for pinctrl support in the MCUX LPUART driver.

[danieldegrasse]

@hakehuang Could you run CI on this PR, specifically targeting Kinetis boards? I want to make sure all the pinmux settings are correct.

[hakehuang]

@danieldegrasse CI is running will feedback tomorrow.

[henrikbrixandersen]

Looks good, thank you. I know it's still in draft, but I took the liberty to review it anyway :)

I noticed that in some drivers you conditionally use pinctrl, others you always use it regardless if it is enabled or not. I take it the conditional ones are due to the same drivers being in use in other SoC families (e.g. i.MX RT)?

Anyways, for drivers that unconditionally use the pinctrl API, I think the Kconfig symbols needs to depend on CONFIG_PINCTRL?

[hakehuang]

@danieldegrasse on frdm_k64f I meet two driver test error.

1. tests/driver/adc. failure log is
   adc_failure_log.txt
2. tests/driver/spi/spi_loopback. failure log is
   spi_failure_log.txt

for spi case, you need short the sin and sout which are: ptd2 - ptd3

the two above cases are OK in mainline weekly testing which is zephyr-v3.0.0-838-gc050d1f2ec.

[danieldegrasse]

@henrikbrixandersen Thanks for the feedback. That is why some drivers have conditional pinctrl enablement- once full iMX.RT support is merged, that should be dropped.

As far as the Kconfig symbols, I think that rather than depending on CONFIG_PINCTRL the drivers should select CONFIG_PINCTRL. This way, a new user to the project doesn't need to debug Kconfig symbols to understand why a driver isn't being included.

[danieldegrasse]

@hakehuang I reproduced and fixed the ADC test failure, but I cannot reproduce the SPI failure. I tried testing with and without the mcux-dspi-dma.conf overlay applied, and in both cases the test passed when ptd2 and ptd3 were shorted. However, the pinctrl driver did have an error regarding the setting of the slew rate (SRE field), where it incorrectly set SRE=1 when a fast slew rate was selected, so I've fixed that.

[hakehuang]

@hakehuang I reproduced and fixed the ADC test failure, but I cannot reproduce the SPI failure. I tried testing with and without the mcux-dspi-dma.conf overlay applied, and in both cases the test passed when ptd2 and ptd3 were shorted. However, the pinctrl driver did have an error regarding the setting of the slew rate (SRE field), where it incorrectly set SRE=1 when a fast slew rate was selected, so I've fixed that.

I will re-run CI based your updates, and feedback

[hakehuang]

I will re-run CI based your updates, and feedback

@danieldegrasse All CI testing PASS. except frdm_kl25z

[danieldegrasse]

@hakehuang I have updated the LPSCI UART peripheral driver for the KL25Z to support pinctrl, which should fix the CI failure. Can you retry CI on that board?

[hakehuang]

have updated the LPSCI UART peripheral driver for the KL25Z to support pinctrl, which should fix the CI failure. Can you retry CI on that board?

running. @danieldegrasse all PASS.

[gmarull]

LGTM, great work