Ozone debugger causes hardfault

[BvdP4Celsius]

I have a rather big application on a nRF52840 under Zephyr 2.4.0 that needs debugging. When starting the Ozone debugger, I can load and run the program as expected. I can see the output in the terminal, all seems to be ok. But when I set a breakpoint, sometimes a hardfault is generated immediately after setting the breakpoint, even before it ever hits it. The fault is not catched by the debugger, but only visible in the terminal window. Sometimes, the breakpoint seems to be ignored totally. I can see from the terminal window het the position of the breakpoint was passed (some log info) but Ozone did not stop. Sometimes the debugger stops several lines after the breakpoint.
Any suggestion where to start looking for the cause of this problem?

[Sigurd]

Hi,
If you are halting the CPU using breakpoints, while the SoftDevice is active, then it will assert, since it needs the CPU to meet it's real-time requirements, and it will detect that it was not able to process interrupts and events in time if the CPU is halted, and it will then assert.

[BvdP4Celsius]

Hi Sigurd,

Is your answer relevant for Zephyr OS also? You mentioned 'Softdevice', however under Zephyr I don't use the softdevice (right?)

[Sigurd]

BvdP4Celsius said:

Is your answer relevant for Zephyr OS also? You mentioned 'Softdevice', however under Zephyr I don't use the softdevice (right?)

Yes, in nRF Connect SDK's case, it will be the SoftDevice Controller. If you pause the CPU, so that BLE real-time requirements are not meet, you wil get a MPSL assert and then a hardfault in the log. Example: