STM32: configure wake-up pins with GPIOs from devicetree for exiting Poweroff

[ajarmouni-st]

This PR implements end-to-end configuration of GPIO pins as sources for STM32 PWR wake-up pins behind the scenes exclusively from devicetree information for all series using a newly introduced public stm32_pwr_wkup_pin_cfg_gpio() function inside gpio_stm32.c driver during initial GPIO pin configuration.

For this, we add a devicetree binding for the STM32 power contoller (PWR) peripheral which manages all device power supplies and power modes transitions, including the control of wake-up pins. We note that this initial PWR binding is focused on wake-up pins.

We also introduce a STM32 specific sample application that demonstrates how GPIO pins can be used to power on the system after a Poweroff, just like a reset pin.
The user only needs to add the STM32_GPIO_WKUP custom GPIO flag to the GPIO pin in "gpios" property of a DT device, such as a button or a sensor.
The sample powers off the system after a delay then the user can power it on again with the user button that has the STM32_GPIO_WKUP flag in DT.
Will only work on STM32 boards that support Zephyr POWEROFF (U5, WL, WB, & L4 SoC series), & also happen to have a user button whose GPIO pin is associated with a wake-up pin (see PWR node in SoC dtsi file).

For more information about how wake-up pins are managed & configured on a given STM32 SoC, please refer to the "Power control (PWR)" section in the Reference Manual. Information about GPIOs associated with wake-up pins can be found in the "pinouts and pin description" section in the datasheet.

[erwango]

@nandojve ^^

[ajarmouni-st]

Additional context: this PR introduced a STM32 custom GPIO flag, STM32_GPIO_WKUP, to be used to signal to GPIO driver that a GPIO pin should be setup as a source for wake-up pins, instead of using the wakeup-source property, because using a flag is more generic, & allows for use cases, beyond just GPIO Keys, where a DT device could have multiple GPIO pins in its gpiosproperty, & only one of those GPIO pins is supposed to be used as wake-up source.

[ajarmouni-st]

Hi @bjarki-trackunit, I would appreciate your feedback, thanks!

[ajarmouni-st]

Rebase

[erwango]

Moving assignee from @galak to @henrikbrixandersen as the owner of the next impacted subsystem (GPIO)

[erwango]

Otherwise LGTM in the general lines.

[zafersn]

Hi, @ajarmouni-st , I am not quite sure to discuss an issue here about your implementation. I just tested wake up pin sample you implemented in your repo: https://github.com/ajarmouni-st/zephyr_dev/tree/dev_stm32_wkup_pins_gpios on the stm32u575zq dev board.

1. the user button couldnt wake up the system.
   

Before powering the system off, current consumption: ~ 12mA

After powering the system off, current consumption: 14uA

I was expecting less than 500 nA of consumption in shutdown mode:

So, I was wondering if I was following up on the correct repo and steps? And any ideas?

[ajarmouni-st]

Hi @zafersn, was it a nucleo_u575zi_q board? I just tested it on my side & the user button is working.

/ {
	aliases {
		wkup-src = &user_button;
	};
};

&user_button {
	gpios = <&gpioc 13 (GPIO_ACTIVE_HIGH | STM32_GPIO_WKUP)>;
};

&pwr {
	status = "okay";
};

Just to verify quickly that the system actually goes into shutdown mode, LED 1 will be turned off.

Will try to verify current consumption on nucleo_u575zi_q.

[zafersn]

Hi @zafersn, was it a nucleo_u575zi_q board? I just tested it on my side & the user button is working.

/ {
	aliases {
		wkup-src = &user_button;
	};
};

&user_button {
	gpios = <&gpioc 13 (GPIO_ACTIVE_HIGH | ST_STM32_GPIO_WKUP)>;
};

&pwr {
	status = "okay";
};

Just to verify quickly that the system actually goes into shutdown mode, LED 1 will be turned off.

Will try to verify current consumption on nucleo_u575zi_q.

Thanks for quick reply @ajarmouni-st :

I see the differences now that you have (GPIO_ACTIVE_HIGH | ST_STM32_GPIO_WKUP) flag configs. I tried with this configs and now button works. and also I have new current consumtion of 5.5uA but it is still not even close to 160nA shutdown mode consumption. :

[ajarmouni-st]

also I have new current consumtion of 5.5uA but it is still not even close to 160nA shutdown mode consumption.

@zafersn I got 1 uA but I don't know if it's the real consu. or it's bcz my ammeter cannot go lower.

In any case, feel free to open an issue regarding current consumption in shutdown mode on STM32U575 & we can continue the discussion there.

[zafersn]

also I have new current consumtion of 5.5uA but it is still not even close to 160nA shutdown mode consumption.

@zafersn I got 1 uA but I don't know if it's the real consu. or it's bcz my ammeter cannot go lower.

In any case, feel free to open an issue regarding current consumption in shutdown mode on STM32U575 & we can continue the discussion there.

Ok thanks for this information.

BTW, I also tried to get reset cause information.

hwinfo_get_reset_cause(&cause);

cause return with 0 after waking up from LL_PWR_SHUTDOWN_MODE
Is there any possibility to get the correct cause info?

[ajarmouni-st]

BTW, I also tried to get reset cause information.

hwinfo_get_reset_cause(&cause);

cause return with 0 after waking up from LL_PWR_SHUTDOWN_MODE Is there any possibility to get the correct cause info?

@zafersn I had it in the PR previously but decided to remove it cz I couldn't get consistent results on all boards, see https://github.com/zephyrproject-rtos/zephyr/compare/4235e22f6c855f3c58386a04812f57ce084618a0..601e1ded1dc99273ab9ce6c4b253fdfd4363279a