app_error_fault_handler: id=1, pc=20732, info=0

[DWN]

Hello, In our project the nRF52 resets randomly

we use the NRF52832 with softdevice S332. On the application space we run a FreeRTOS application with 2 tasks.
We want to use Bluetooth stack (central and peripheral role) concurrently.
We use the NUS example in our project to communicate with another central device.
When we run the application we can connect to the NUS and send data. Sometimes the following application error occurs:
app_error_fault_handler: id=1, pc=20732, info=0

Is this an softdevice error?
How can we trace back the cause of the error?

[davege]

Hi DWN,

did you manage to solve this issue?

I'm experiencing exactly the same with a similar setup (nRF52832, SoftDevice S332 and FreeRTOS used to poll events from SoftDevice via 

nrf_sdh_evts_poll()).

In my case this is happening exclusively when debugging. If I set a breakpoint and then try to restart the application, I end up in the forced 

NRF_BREAKPOINT_COND within app_error_weak.

Any idea of what could be the issue is welcome!

Thanks,

D

[Turbo J]

davege said:

In my case this is happening exclusively when debugging. If I set a breakpoint and then try to restart the application, I end up in the forced 

That is a known no-no when running softdevice, as this will fault when it cannot keep the BT LE mandated timings for any reason.

Only known workaround would be SEGGER monitor-mode debugging (which does NOT stop softdevice from running in the background).

[davege]

Hi Turbo,

thanks for the reply.

That's is effectively explaining a lot of things. May I where this forced assertions due to timing constraints from SD is documented? I would be curious to understand more.

Also thanks for pointing to Monitor-mode debugging, I will try to check if I can define a target for this in my project, considering I'm using a pure armgcc Makefile to build the system rather than an IDE solution.

Best,

D