Re: New CBM-II discoveries

Den Wed, 27 Jun 2018 15:02:03 +0200 skrev Michał Pleban
<lists@michau.name>:
> Hello!
> 
> Mia Magnusson wrote:
> 
> > Btw it would be a good idea to once and for all agreee on a way to
> > read a specific address in the firmware rom to determine if the
> > card has a 6523 or an 8255. (This of course applies to the 8088
> > board. I see no reason for making the Z80 board with a 6523 at all,
> > thus there is no reason for differentiation of any software.
> 
> Good idea. I think a value could be put somewhere around the F000:FFEx
> location to indicate it's a 8255 card (there is obviously a FF value
> in the original BIOS).

(Not that obvious for me :) )
Btw I don't know what a PC BIOS does here. Maybe there is some kind of
"standard" for some identifying strings/data there?

> > Btw I assume the idea from Commodore were that the built in Kernal
> > code for talking with the 8088 card should work with the Z80 card
> > too.
> 
> Yes. The Kernal implements a certain IPC protocol, and as long as the
> card conforms to it, it really doesn't care what is on the other side.
> For what it's worth, you could even create some little hardware to
> place an another computer over there.

Ok.

> > There might be a need for 6509 software to determine if it's a Z80
> > or an 8088 card in the machine though. I wounder how or even if
> > Commodore ment that to work? 
> 
> I am not sure such distinction is needed. The 6509 library implements
> some functions like "print out a character" or "read sector from
> disk". Again it really doesn't care which processor called this
> funcion. And if you really need to do something different for one
> processor, you would simply write a different version of the library?

My main concern is the (rather small) risk of someone starting the 6509
software for the 8088 card while having a Z80 card or vice versa.
Probably not a big problem, but still...

Btw in a theoretic future a recreated CBM-II motherboard could have
hardware for hooking up more than one coprocessor card. That would need
some way to select which coprocessor card that should be enabled. In
that scenario there needs to be some logic to make sure the user
doesn't run software for another card than the active hardware.

> > Does the 6509 software start the coprocessor and
> > first talk to it via the 6526/6523 interface and get some ID thing
> > from the coprocessor to determine if the 6509 software fits the
> > installed card?
> 
> The 6509 sends an IRQ to the coprocessor, and enters an infinite loop
> - it doesn't do anything else. It's up to the coprocessor to call it
> back requesting some I/O function.

Ok. Then the only thing that might go wrong is which (boot) disk that
the coprocessor will read data from.


-- 
(\_/) Copy the bunny to your mails to help
(O.o) him achieve world domination.
(> <) Come join the dark side.
/_|_\ We have cookies.