boards: support Arduino Portenta C33

[facchinm]

As per title, this PR is a WIP for Arduino Portenta C33 support.
The porting status is quite advanced but it's still missing a couple of very important features.
Based on RA6M5 microcontroller, onboard peripherals that need to be supported are:

• RTC
• ADC
• GPIO
• PWM
• I2C ~ some exposed I2C ports are based on SCI, not IIC peripherals
• SPI ~ some exposed SPI ports are based on SCI, not SPI peripherals
• UART
• CAN
• PMIC ~ same as boards: stm: arduino_portenta_h7: support onboard ethernet + pf1550 driver #76542
• BLE
• WiFi ~ drivers: wifi: Add WiFi drivers for ESP hosted firmware. #87685
• Ethernet
• USB HS (device)
• USB FS (host) ~ drivers: usb: uhc: initial support for UHC driver on Renesas RA family #82730
• Flash over QSPI
• SDCARD

https://builds.zephyrproject.io/zephyr/pr/85337/docs/boards/arduino/portenta_c33/doc/index.html

[beriberikix]

Is there a reason to use esp_host over esp_at, which is already in-tree? Unfortunately that's also missing SPI support.

[nordicjm]

Suggested change

CONFIG_USE_DT_CODE_PARTITION=y

[facchinm]

This needs to be set to preserve the default functionality (with dfu bootloader). Is there another way to achieve the same result?

[nordicjm]

if this is required then needs to go in Kconfig and the board Kconfig select it

[nordicjm]

have a common .c file with all of the code functions then use a board hook to run a function which runs them

[facchinm]

I removed the USB init file (which was a dirty solution anyway), will apply #85294 whrn it gets merged

[facchinm]

Is there a reason to use esp_host over esp_at, which is already in-tree? Unfortunately that's also missing SPI support.

When we had to decide which firmware to flash on the network companion, esp-at for ESP32C3 was brand new and very unstable, so we took the other direction 🥲

[pillo79]

Review and CI comments addressed. @nordicjm please revisit.

[nordicjm]

Suggested change

	zephyr_library()
	zephyr_library_sources_ifdef(CONFIG_NETWORKING board.c)
	if(CONFIG_NETWORKING)
	zephyr_library()
	zephyr_library_sources(board.c)
	endif()

otherwise will have an empty library warning when NETWORKING is disabled

[pillo79]

Thanks for pointing this out. Having a board.c conditional on NETWORKING looked strange as well, so I renamed the file to eth_clock.c and dropped the library step.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[pillo79]

@nordicjm please revisit, thanks! 🙇

[pillo79]

@nordicjm thanks for the review, concerns should have been addressed. [...]

EDIT: Note that I had to restore the SYS_INIT entry as in the original PR, because neither early nor late board hooks are acceptable to set up the Ethernet clock; it needs to be exactly between PWM clock init and Eth device init.

@nordicjm As already mentioned last week, I need precise ordering between the above devices:

• early board hook is executed before the clock subsys is even initialized,
• late board hook is after the Ethernet device is probed unsuccessfully due to the missing clock.

I see no way to achieve this without a SYS_INIT... 🙁 but really open to suggestions!

[nordicjm]

@nordicjm thanks for the review, concerns should have been addressed. [...]
EDIT: Note that I had to restore the SYS_INIT entry as in the original PR, because neither early nor late board hooks are acceptable to set up the Ethernet clock; it needs to be exactly between PWM clock init and Eth device init.

@nordicjm As already mentioned last week, I need precise ordering between the above devices:

* `early` board hook is executed before the clock subsys is even initialized,

* `late` board hook is after the Ethernet device is probed unsuccessfully due to the missing clock.

I see no way to achieve this without a SYS_INIT... 🙁 but really open to suggestions!

FYI @nashif