The PC would fall behind whenever an IRQ was called because the pipeline
was reloaded (+8 to PC), however that was never actually done by any code
Now, the PC is always incremented when the pipeline is reloaded
I believe this to be necessary in order to get hardware interrupts
working.
thumb.gba test 108 fails but I'm committing anyways (despite the
regression) because this is kind of rebase/merge hell and I have
something that at least sort of works rn