A software interrupt ends on its own when viewing on trace...any explanation?

[cpeng]

Hi Devzone:

We are using J-link pro to collect trace data from nRF52833 DK board, and we found a rather strange behavior:

The sw interrupts would end on its own without PE coming back from the call stack to end it.

See ozone screenshot below:

a little explanation of the doodle in the picture: "another main() fills in" is a ozone display artifact. Guess it doesn't know how to handle a interrupt ending pre-maturely so it inserts a second main.

So regarding the problem, shouldn't the sw interrupt (the orange bar) continue extending to the right till PE comes back from the call stack? Why does it terminate pre-maturely?

Thanks.

[cpeng]

Thanks for the tip jmr! I will give it a try if i have a chance.