S112 crash at address 0x18808

[Gaétan]

Hi,

On a dice projet based on NRF52810. Sometime when I plug the battery charger the firmware crash at address 0x18808 and the call stack is empty !

This address is in softdevice , and I can't determine the cause !

SDK 16.0

Softdevice version: s112_nrf52_7.0.1_softdevice.hex

Please can you explain what is the function at 0x18808 address ?

Regards,

Gaétan

[Gaétan]

Hi,

The batterie charger is :

I have found a current injection on a pin "LED" connected directly to nrf52810.

When bootloader software is present -> sometime  fix at adresse 0x18808

When bootloader software is not in flash -> sometime  firmwre restart

Now I make multiple tests with nRF52810 pin forced to ouput whith high current configuration...

Regards,

Gaétan

EDIT:

Now, without charger...

I allways have the same fix at address 0x18808 when I connected to NUS !

Last log is : <info> app: Data len is set to 0xB6(182)

NrfToolBox: 

With s112_nrf52_7.3.0_softdevice.hex address is 0x188D4

EDIT 2:

When I put the firmware compiled for nRF52810 directly in a PCA10040 (NnRF52832) all is OK no fix but in a nRF52810 allways fix when etablish a NUS connection. No problem when using DFU service. 

[Hung Bui]

Hi Gaetan, 

Could you please confirm that even without the charger you still have the issue when you are trying to connect with NUS ? 

Could you check if you have any log on the device ? Any assertion ? Could you send the full log ?
If you test with unmodified NUS example do you have the same issue  ? 

[Hung Bui]

Hi Gaetan,

I have just tested your BL+Settings+App+SD.hex file on my nRF52810 dev kit and it worked fine. 
You can see the screen recording here: 

This is quite strange. Do you have any other hardware board to test ? 
Could you record the screen when you test ? What exactly you meant by "connecting to NUS" 
If you test with the buttonless example (no NUS) do you have any trouble ? 

[Gaétan]

Hi, 
I'm back from holiday...

I have make some tests with 

Gaétan said:

Firmware: BL+Settings+App+SD.zip

and the crash occur only if SES is in debug mode !

Thanks for confirmation:

Hung Bui said:

The error at 0x18808 and 0x188D4 pointing us to the hardfault handler in the softdevices. 

Please can you explain why softdevice don't fire the function app_error_fault_handler() when hard fault is catched ?

Thanks

[Hung Bui]

Hi Gaetan, 
When you are running in debug mode did you stop at any breakpoint and then continue ? 

Have you built your application so that it's friendly for debugging ? (build with no optimization )
The app_error_fault_handler() only handle assertion that thrown by the softdevice or the SDK, not the hardfault. It's not possible to return to the application from the softdevice when you have a hardfault. 

Please try to narrow down the issue by comparing to our DFU Buttonless example. Please try to test if you can reproduce the issue with the example. 

[Gaétan]

Hung Bui said:

When you are running in debug mode did you stop at any breakpoint and then continue ? 

Never. But I download firmware in flash memory. and after I use "attach to debbuger" menu for connect Jlink RTT to debug terminal in SES.

I continue to investigate... and send feedback.

[Hung Bui]

You can use the RTT Viewer application to see the log without having to "attach to debugger" in SES.