Maximum number of pin interrupts

[Eric Wiebols]

Hi,

In this post  Maximum number of interrupt enabled pins? Hung Bui states that there can only be "maximum 4 pin events, each for one GPIO pin (so max 4 pins at a time)". This is (apperently) not true since I have 8 pins set up with individual interrupt callbacks and that works just fine. This makes sense to me as there are 8 GPIOTE->CONFIG[] registers.

Consequently, adding a 9th input causes strange behaviour.But when the 9th input is configured via a call to  gpio_pin_interrupt_configure() this function does not return an error. Shouldn't this be considered a bug in the library function ?

\Eric

[Håkon Alseth]

Hi,

 

In this post  Maximum number of interrupt enabled pins? Hung Bui states that there can only be "maximum 4 pin events, each for one GPIO pin (so max 4 pins at a time)".

This is true for GPIOTE IN channels for the nRF51-series devices.

For nRF52-series devices, it is 8 GPIOTE IN channels:

https://infocenter.nordicsemi.com/topic/struct_nrf52/struct/nrf52.html?cp=5

 

That being mentioned, you can still use level triggered wakeup sources, ie. GPIOTE PORT event. This can be set on all GPIOs if you want.

To implement a software-emulated edge-triggered detection based on GPIOTE PORT event, you can use "sense-edge-mask" in DeviceTree:

 How to use sense-edge-mask? 

 

Kind regards,

Håkon

[Eric Wiebols]

Hi Håkon,

Thank you for the feedback. I will look into using the "sense-edge-mask".  However, I still think gpio_pin_interrupt_configure() should return an error if you try to configure more interrupts than are available in the chip. This would have saved me a lot of time trying to figure out what was going on.

\Eric

[Håkon Alseth]

Hi Eric,

 

I ran a test, where I added this loop to the zephyr/samples/basic/button:

	struct gpio_dt_spec test = {
		.port = button.port,
		.pin = 0,
		.dt_flags = button.dt_flags,
	};
	for (int i = 0; i < 10; i++) {
		test.pin = i;
		ret = gpio_pin_interrupt_configure_dt(&test,
							GPIO_INT_EDGE_TO_ACTIVE);
		if (ret != 0) {
			printk("Error %d: failed to configure interrupt on %s pin %d\n",
				ret, test.port->name, test.pin);
			return 0;
		} else {
			printk("Success: %d\n", i);
		}
	}

 

This now shows:

*** Booting nRF Connect SDK v2.5.0 ***
Success: 0
Success: 1
Success: 2
Success: 3
Success: 4
Success: 5
Success: 6
Success: 7
Error -12: failed to configure interrupt on gpio@842500 pin 8

 

However, If I add the below mentioned "sense-edge-mask", I can configure more than 8:

*** Booting nRF Connect SDK v2.5.0 ***
Success: 0
Success: 1
Success: 2
Success: 3
Success: 4
Success: 5
Success: 6
Success: 7
Success: 8
Success: 9
Set up button at gpio@842500 pin 6
Set up LED at gpio@842500 pin 2
Press the button

 

Are you certain that the "sense-edge-mask" is not configured in your overlay?

 

Kind regards,

Håkon

[Eric Wiebols]

This is my configuration:

&gpio0 {

    status = "okay";

};

i.e no "sense-edge-mask = < 0x... >;" definition.

My (inherited) code uses  gpio_pin_interrupt_configure(). Maybe gpio_pin_interrupt_configure_dt() works differently ?

[Håkon Alseth]

Hi,

 

It wraps around the function that you're using:

https://github.com/nrfconnect/sdk-zephyr/blob/v3.4.99-ncs1-1/include/zephyr/drivers/gpio.h#L931

 

If you call the above with a GPIO_INT_EDGE* input, it will become GPIOTE allocated channel (unless it is masked in sense-edge-mask).

A GPIO_INT_LEVEL* input will become a GPIOTE PORT event, where you can have as many as you have GPIOs available.

 

Kind regards,

Håkon