GPIO event shall increment counter without CPU

RE: GPIO event shall increment counter without CPU

Håkon Alseth — Fri, 01 Aug 2025 14:12:43 GMT

Hi,

[quote user="Artnia"]

First of all it doesn't build:

- /src/modules/network/network.c:16:30: error: 'CONFIG_MQTT_SAMPLE_NETWORK_LOG_LEVEL' undeclared here (not in a function); did you mean 'CONFIG_MQTT_SAMPLE_ERROR_LOG_LEVEL'?

- /src/modules/network/network.c:110:17: error: 'CONFIG_MQTT_SAMPLE_NETWORK_THREAD_STACK_SIZE' undeclared here (not in a function); did you mean 'CONFIG_MQTT_SAMPLE_SAMPLER_THREAD_STACK_SIZE'?

[/quote]

It is hard to say what has happened here, have you done any changes to the project?

First, try to delete your build folder and regenerate the build.

The API has been changed for nrfx_gpiote_init function. Instead of:

nrfx_gpiote_init(NULL, 0);

You need to provide the instance:

nrfx_gpiote_t const gpiote_inst = NRFX_GPIOTE_INSTANCE(GPIOTE_INST_IDX);
status = nrfx_gpiote_init(&gpiote_inst, NRFX_GPIOTE_DEFAULT_CONFIG_IRQ_PRIORITY);

Where 0 is for secure, and 1 is for non-secure:

https://docs.nordicsemi.com/bundle/ps_nrf9160/page/gpiote.html#d558e286

Kind regards,

Håkon

RE: GPIO event shall increment counter without CPU

Artnia — Thu, 31 Jul 2025 08:14:43 GMT

1. As a base for a minimal sample I use /home/art/ncs/v3.0.2/nrf/samples/net/mqtt for nrf9160dk/nrf9160/ns.

Usually is says on boot:

*** Booting nRF Connect SDK v2.8.0-a2386bfc8401 ***
*** Using Zephyr OS v3.7.99-0bc3393fb112 ***
[00:00:00.395,111] <inf> network: Bringing network interface up and connecting to the network

- /src/modules/network/network.c:110:17: error: 'CONFIG_MQTT_SAMPLE_NETWORK_THREAD_STACK_SIZE' undeclared here (not in a function); did you mean 'CONFIG_MQTT_SAMPLE_SAMPLER_THREAD_STACK_SIZE'?

When I replace both with a hard coded value it builds.
Inside the trigger task I add line 14+15

static void trigger_task(void)
{
#if CONFIG_DK_LIBRARY
	int err = dk_buttons_init(button_handler);

	if (err) {
		LOG_ERR("dk_buttons_init, error: %d", err);
		SEND_FATAL_ERROR();
		return;
	}
#endif /* CONFIG_DK_LIBRARY */


	nrfx_gpiote_init(NULL, 0);
	LOG_INF("GPIOTE initialized");

	while (true) {
		message_send();
		k_sleep(K_SECONDS(CONFIG_MQTT_SAMPLE_TRIGGER_TIMEOUT_SECONDS));
	}
}

and the behaviour is that nothing is printed.

2. If I do the same for hello_world it seems to work:

int main(void)
{
	printf("Hello World! %s\n", CONFIG_BOARD_TARGET);

	nrfx_gpiote_init(NULL, 0);

	printf("GPIOTE initialized successfully.\n");

	return 0;
}

Output:

All pins have been configured as non-secure
Booting TF-M v2.1.1-ncs4-2
[Sec Thread] Secure image initializing!
*** Booting nRF Connect SDK v3.0.2-89ba1294ac9b ***
*** Using Zephyr OS v4.0.99-f791c49f492c ***
Hello World! nrf9160dk@0.14.0/nrf9160/ns
GPIOTE initialized successfully.

3. According to https://docs.nordicsemi.com/bundle/ps_nrf9160/page/gpiote.html#d558e286
which one (GPIOTE0 or GPIOTE1) should I be using in my application if ns/TFM is enabled? And does ns/TFM also use one of them?
https://docs.nordicsemi.com/bundle/ps_nrf9160/page/dppi.html is also related

RE: GPIO event shall increment counter without CPU

Håkon Alseth — Thu, 31 Jul 2025 06:01:06 GMT

Hi,

Could you try to place your logic in a minimal sample, like zephyr/samples/hello_world or similar, then see if the gpiote/ppi/timer logic runs as expected?

[quote user="Artnia"]When I try to build path/to/ncs/modules/hal/nordic/nrfx/samples/src/nrfx_gppi without any modification but just the build config:[/quote]

Inside that folder there are two samples. You could look at both to see which one is closest to your use-case.

Kind regards,

Håkon

RE: GPIO event shall increment counter without CPU

Artnia — Wed, 30 Jul 2025 17:42:34 GMT

Hi,

1. For the build without ns I get:

/ncs/v3.0.2/nrf/lib/at_monitor/at_monitor.c:11:10: fatal error: nrf_modem_at.h: No such file or directory

2. When I build for ns and comment the function button_hw_counter_init() which i posted above, it boots without error. But obviously I need that function.

If I shrink the function to

void button_hw_counter_init(void)
{
	nrfx_gpiote_init(NULL, 0);
}

the behaviour is the same. So it already crashed here. Is GPIOTE already used by TFM and hence not available for me?

I build upton the pp_modules example.

3. When I try to build path/to/ncs/modules/hal/nordic/nrfx/samples/src/nrfx_gppi without any modification but just the build config:

I get:

CMake Error at CMakeLists.txt:11 (GET_DEVICE_CONFIG_FILES):
  GET_DEVICE_CONFIG_FILES Macro invoked with incorrect arguments for macro
  named: GET_DEVICE_CONFIG_FILES

while

cmake_minimum_required(VERSION 3.20.0)

if(DEFINED CUSTOM_BOILERPLATE)
    include(${CUSTOM_BOILERPLATE})
    return()
endif()

set(COMMON_PATH ${CMAKE_CURRENT_LIST_DIR}/common)
include(${COMMON_PATH}/common.cmake)

GET_DEVICE_CONFIG_FILES(${BOARD} boards)

find_package(Zephyr REQUIRED HINTS $ENV{ZEPHYR_BASE})
project(one_to_one)
target_sources(app PRIVATE main.c)
target_include_directories(app PRIVATE common)

If i fix the error by

# GET_DEVICE_CONFIG_FILES(${BOARD} boards)

I can build for nrf9160dk/nrf9160 but not for nrf9160dk/nrf9160/ns:

/ncs/v3.0.2/modules/hal/nordic/nrfx/drivers/include/nrfx_gpiote.h:183:57: error: 'nrfx_gpiote_0_irq_handler' undeclared (first use in this function); did you mean 'nrfx_timer_0_irq_handler'?

/ncs/v3.0.2/modules/hal/nordic/nrfx/drivers/include/nrfx_gpiote.h:76:33: error: 'NRFX_GPIOTE0_INST_IDX' undeclared (first use in this function); did you mean 'NRFX_TIMER0_INST_IDX'?

RE: GPIO event shall increment counter without CPU

Håkon Alseth — Mon, 28 Jul 2025 14:45:13 GMT

Hi,

Here's a sample that uses gpiote + timer + dppi to measure a edge:

Best way to measure time between falling/rising edges (NCS, Zephyr, NRF52840)

There are also samples in the nrfx folder which could be of interest for you:

path/to/ncs/modules/hal/nordic/nrfx/samples/src/nrfx_gppi

Since you are seeing a fault condition, I would recommend that you run this sample isolated for testing purposes, for board nrf9160dk/nrf9160 (note the lack of /ns), and see if this fault still occurs. If the fault still occurs, please share the output.

Kind regards,

Håkon