Force softDevice assert

RE: Force softDevice assert

Sigurd — Fri, 17 Apr 2020 13:46:56 GMT

It's the fault-handler used in the sd_softdevice_enable() call that will be used. This is app_error_fault_handler() , and not assert_nrf_callback().

0x00000A60 is the hardfault exception handler inside the MBR. It's normally forwarded to the main application, but will get trapped if breakpoint on hardfault is enabled in SES.
You can disable it like this:

RE: Force softDevice assert

Tilak — Thu, 16 Apr 2020 05:45:56 GMT

Hi,

Thank you for the tip. I tried your code even with 11 second delay after idle_state_handle(); in the example ble_app_blinky_pca10040_s132. I did get a crash but the callstack was blank "unknown function at 0xA60". I did not get the Segger Embedded Studio to stop at DEAD_BEEF. This happened for both Release and Debug builds.

Kindly let me know what am I missing to see the callstack with DEAD_BEEF? Is SoftDevice not seeing my assert_nrf_callback() function in my main.c? If possible, please do share an example.

Thanks,
Tilak

RE: Force softDevice assert

Turbo J — Tue, 14 Apr 2020 19:05:26 GMT

Disabling interrupts long enough should do the trick:

__disable_irq();
nrf_delay_ms(1000);
__enable_irq();

Adjust the time if you use longer advertising or connection interval.

The softdevice uses timers to check whether BTLE timing was met and asserts if not.