Page 4 of 4
Re: Mini-challenge: detecting activity on a port
Posted: Tue Apr 07, 2020 8:32 am
by Chromatix
The /NMI sampling mentioned will be in the context of the edge detection logic. If it was high this cycle but not last cycle, set the flip-flop so we remember to take the interrupt at the next instruction boundary; reset that flipflop as part of the NMI interrupt entry.
Re: Mini-challenge: detecting activity on a port
Posted: Tue Apr 07, 2020 9:42 am
by BigEd
The general question of interrupt timing got a good airing in the thread
A taken branch delays interrupt handling by one instruction
but I think the question of how WAI resumed might be an interesting one to measure. I'm sure there's documentation, but as we know well, the reality could possibly be different: and a single cycle pulse on IRQ is not a normal thing to do.