Can anyone tell all the potential reasons for stack corruption? I'm debuging a problem where sometimes there is no stack trace when putting a breakpoint on the first line of the app_error_fault_handler. I'm causing a deliberate crash by putting an APP_ERROR_HANDLER(0) statement in the first line of the main function. Our custom app_error_fault_handler looks like this:
void app_error_fault_handler(uint32_t id, uint32_t pc, uint32_t info)
{
if (m_err_info == NULL)
{
m_err_info = (error_info_t *)info;
nrf_sdh_disable_request();
ret_code_t err_code = batt_recording_store_err_info(m_err_info);
}
// disable wake-on-gpio
for (int i = 0; i < NUMBER_OF_PINS; ++i)
{
if (i != 21) // exclude the reset pin
nrf_gpio_cfg_default(i);
}
// disable wake-on-nfc
//NRF_NFCT->TASKS_DISABLE = 1;
__sd_nvic_irq_disable();
while (true)
{
nrf_power_system_off();
}
}
If I comment out the endless loop at the end of the fault handler I get a stack trace, but if I leave it in I don't. Can't really see how it could be a buffer overflow since the APP_ERROR_FAULT_HANDLER is the first statement in main. The optimization is O0.
Can't make sense of this.
