🚀 go-pugleaf

In article <109194@sun.Eng.Sun.COM> cmcmanis@sun.UUCP (Chuck McManis) writes:
>In article <17184@louie.udel.EDU> new@udel.EDU (Darren New) writes:

[deleted stuff about Amiga memory map questions and answers]

>space. Of course due to the way the 68000 resets, ROM has to be located
>at $FFFFFF, but can extend "down" toward 0 an arbitrary distance, 256K
>in the case of the current Amigas, probably 512K in a future Amiga. (

Well, not quite. The 68000 loads its initial PC and Stack Pointer from the
first 8 bytes in the address space ($000000 to $000007). This normally
requires one of the following:

	1. ROM is permanently located starting at 0. (Unless special
	   provisions are made to accommodate it, this usually makes it 
	   real ugly to try to redirect exception vectors.)
	2. ROM responds to an address range starting at 0 until some action 
	   causes it to assume its usual address. I've seen this action 
	   consist of anything from an explicit write to a port address to 
	   an automatic toggle that happens when you branch to an address 
	   in the normal ROM address range.
	3. ROM located at both 0 and its normal address until a similar
	   action described in 2.
	4. The base address (or entry point) of the ROM is permanently 
	   gated onto the data bus when the first longword in memory is 
	   accessed, effectively replacing the first 4 bytes of memory with 
	   a 4 byte ROM.

I'm sure there are other methods of handling this, but these are the major ones
I've encountered. Which one does the Amiga use? Well... I knew once, but I
forgot! (Shows you how important it is once you know it works, eh?)

The location of ROM at $FFFFFF and "down" is usually to avoid fragmenting
the available address space unnecessarily, or to simplify chip select decoding.

>
>--Chuck McManis
>uucp: {anywhere}!sun!cmcmanis   BIX: cmcmanis  ARPAnet: cmcmanis@sun.com
>These opinions are my own and no one elses, but you knew that didn't you.
>"A most excellent barbarian ... Genghis Kahn!"

(I guess I got a bit longwinded picking a very small nit, didn't I? Sorry.)

-Phil
-- 
------------------------------------------------------------------------------
Phil Staub        
Tektronix, Inc., Vancouver, Washington 98668
phils@tekigm2.MEN.TEK.COM