Hi @haugoug,
you were an hour too early with the merge ;)
I was working on fixing a major problem over the weekend: single stepping and the ID mapping of signals sent back to GDB were completely wrong - probably to the lousy documentation of the DBG_CAUSE register.
I have not tested these changes with interrupts or anything but ordinary breakpoints but it is pretty clear that this is a huge improvement. One could even say that debugging worked only by accident in the past :)
Hi @haugoug, you were an hour too early with the merge ;) I was working on fixing a major problem over the weekend: single stepping and the ID mapping of signals sent back to GDB were completely wrong - probably to the lousy documentation of the
DBG_CAUSE
register.I have not tested these changes with interrupts or anything but ordinary breakpoints but it is pretty clear that this is a huge improvement. One could even say that debugging worked only by accident in the past :)