kilian@seas.gwu.edu (Jens Kilian) (01/23/91)
In article <1991Jan21.115306.17458@ifi.uio.no> jkr@ifi.uio.no (Johan Kristian Rosenvold) writes: > >My revision of the developer system doesn't say anything about this. Are the >chip register adresses defined as a negative offset from the top of the addressable >range or as the absolute address 00 FF FA 01 ? > A pity that Atari didn't document this. This is exactly the same kind of brain-damage that causes the 640 K memory limit on you-know-who PCs. The whole confusion probably stems from Atari's use of 24-bit addresses in their meager documentation. Did they never expect to use a 68020 in future models ? My advice is to use 16-bit addresses for hardware registers (if you REALLY must use the registers, that is ...), and cursed be those who don't ! Jens Kilian -- \/// Internet: kilian@seas.gwu.edu |Snail: 4715 MacArthur Blvd. \\\/ /// UUCP: ...uunet!seas.gwu.edu!kilian| Washington, DC 20007 \\\ \\\ [These addresses will change (to kilian@cc.gatech.edu) soon] /// /\\\ "Wos deht ich-en so gern haage, awwer er spihrt doch nix ..." ///\
micro@imada.dk (Klaus Pedersen) (01/23/91)
jkr@ifi.uio.no (Johan Kristian Rosenvold) writes: > Are the chip register adresses defined as a negative offset from the top of > the addressable range or as the absolute address 00 FF FA 01 ? If think that the addresses is defined to be in the top of the memorymap eg. 0xFFFF_FA01, and that just happens to be in the reach for a negative short absolute addressing. But it should be a simple problem for the onchip MMU, to map the 24 bit addressing used by some programmers/compilers/assemblers to the full 32 bit on the TT. - Klaus >-- >K. Rosenvsold, jkr@ifi.uio.no / ...!{uunet,mcvax,sunic}!ifi.uio.no!jkr >Short signatures R cute.