From: Rainer Buchty (buchty_at_cs.tum.edu)
Date: 2003-08-23 21:04:23
> Won't an ACIA NMI get lost in this case? Assume this order: > > 1. CIA NMI triggered > 2. CIA NMI gets handled, but still stays active > 3. ACIA NMI occurs, since /NMI is already active, nothing happens > 4. CIA NMI released. /NMI still active (from ACIA), nothing happens Well, then after releasing the CIA NMI the ACIA IRQ status has to be checked to see if something did happen meanwhile and then jump to a "userland" copy if the NMI handler? If I understood your first mail correctly, all you wanted to prevent is the occurance of an ACIA NMI during access of the ACIA control register happening outside the NMI routine. To Spiros: Yeah, I indeed mixed them up (as a quick look into the 6502 datasheet would have shown me first place): IRQ is level-triggered (i.e. level stays, IRQ will be re-entered after RTI) where NMI is falling-edge triggered (i.e. staying level prevents further occurances). So it's not you who should get asleep but rather me not writing mails so early after getting up :) Rainer Message was sent through the cbm-hackers mailing list
Archive generated by hypermail pre-2.1.8.