Stuck in GPIO interrupt during init

[Chuppster]

While attempting to run the following code from a MESH example in IAR, the program gets stuck.  It only seems to be calling button_event_handler continuously, as if the interrupt is never clearing.

static void initialize(void)
{
    __LOG_INIT(LOG_SRC_APP | LOG_SRC_ACCESS, LOG_LEVEL_INFO, LOG_CALLBACK_DEFAULT);
    __LOG(LOG_SRC_APP, LOG_LEVEL_INFO, "----- BLE Mesh Light Switch Provisioner Demo -----\n");

    hal_leds_init();
#if BUTTON_BOARD
    ERROR_CHECK(hal_buttons_init(button_event_handler));
#endif

    /* Mesh Init */
    nrf_clock_lf_cfg_t lfc_cfg = DEV_BOARD_LF_CLK_CFG;
    ERROR_CHECK(mesh_softdevice_init(lfc_cfg));
    mesh_init();
}

Pretty much everything is bone stock in this code base, and I have logging disabled.  Does anyone have any leads on what I can check to debug this?  I've tried checking the registers in the NVIC but everything looks okay.  It only finishes initialization when I push button 1 on the board.  This is my Main() where initialize() gets called from.

int main(void)
{
    
    initialize();
    NRF_LOG_INFO("Initialization complete");
    execution_start(start);

    for (;;)
    {
        (void)sd_app_evt_wait();
    }
}

[Hung Bui]

Hi, 

Which hardware board did you use for testing ? 

Which button did you have the issue ? Have you check the level of the button ? 

Have you tried to compile using Segger Embedded Studio ? 

 

[Chuppster]

1. PCA10056

2. It's button 1.  How can I check the level of the button?

3.I'm trying to replicate the MESH light_switch provisioner example.  Runs just fine in SES, but we are trying to get away from it.

[Hung Bui]

I mean have you checked what the actual voltage level of the pin connect to the button. 

You can try to test without  starting the mesh stack, just check what triggers GPIOTE_IRQHandler(). Note that the PORT event only triggered on the rising edge of the DETECT signal. I'm suspecting there is one of the pin staying float and trigger lots of event ? 

[Chuppster]

The pin on the micro is normally at 1.5v until the switch is closed, in which case it goes to 0v.  I'm guessing it should be at 3.3v, so it is indeed floating?

[Chuppster]

I'm pretty sure this has to do with the fact that we are using the 20 pin trace debug port.  I'm going to try and figure out how to switch the buttons over to the alternate.

[Hung Bui]

 Button 1 using pin P0.11 which is also TRACE DATA 2. I guess you find the root cause.