problem with 0x8C8 line

[gercer]

Hello!
I've have been designing a custom board. The circuit reads ADC channel 1 for long periods, 1 hour at most. When the reading is high enough it will send an alert using ble_nus_data_send. Sometimes the circuit stops at the address 0x8C8

The guy is this post had the same problem solved by comparing the map file but the link provided it does not longer exist.

I was wondering if anyone could provide me that file so I can compare it.

Also, i was hoping someone could explain what are the reasons the code stops there.

Thank you!

[bjorn-spockeli]

Hi Gerardo, 

Which SDK version are you using? There is a known issue in SDK v15.3.0 for ble_nus_data_send, see https://devzone.nordicsemi.com/f/nordic-q-a/51621/hardfault-ble_nus_send_data. 

As stated in the question you linked to, address 0x8C8 is the HardFault handler of the MBR. HardFaults typically occur when the CPU is instructed to write to a protected flash region or branch to an address outside the mapped flash . In the known issue with ble_nus_data_send the issue was that the p_nus->data_handler was a null pointer.

The issue in the case you referred to was that some preprocessor definitions were missing, which caused the incorrect SD APIs being used and this caused a Hardfault. The map file for your application adn the blinky application can be found in the Output\Release\Exe or Debug folder after compiling the SES project. 

Best regards

Bjørn

[gercer]

Thank you for the answer
Im using SDK 15.2.

Should I erase the flash memory often to avoid writing outside the mapped flash?

[bjorn-spockeli]

gercer said:

Should I erase the flash memory often to avoid writing outside the mapped flash?

 No, its usually incorrect or corrupted values that cause the CPU is instructed to write to a protected flash region or branch to an address outside the mapped flash. 

So I would recommend using the HardFault handling library in SDK v15.2.0 to debug the issue. 

Best regards

Bjørn