peer review

Looks good, passes CI so no major complains. Not a big fan that the kernel code starts with a jump to the handler. Would be a lot cleaner to start with that in the kernel code and place the loop last. Just feels like an unnecessary confusion. But not a major issue since still works of course.