From: "Jeff" <fyberoptic1979@...>
Sep 6, 2007
How hard do you suppose it would be to make the first 16KB of address
spac=
e swappable, like the middle two "slots" can do? I haven't gone
over the =
circuits hardly at all so I don't know really what does what
at the moment=
. But if it wasn't too elaborate, and one could take
advantage of some sp=
are port or something like that, it might be
worth trying. Even if I had =
to hard-wire the page of ram it went to,
that'd be okay too. And I know t=
here's some undocumented pins on the
CPU still.
In FyOS, I've written in =
compatibility for almost all of CP/M's BDOS
function calls except file ope=
rations. Eventually I want to get that
too. But the point is, most CP/M =
apps naturally expect to run from
0x0100. They call 0x0005 for BDOS funct=
ions. These two aspects
alone aren't TOO bad; I mean, with a little work,=
one could relocate
a program to run at 0x4000 instead (which I've been do=
ing to one for
testing purposes)
But obviously there's much more than sim=
ple relocation most times,
since various apps can have hard-coded memory l=
ocations and routines
that do things in places that'd require digging real=
ly deep in the
code, and truly learning what every aspect of the code is d=
oing. A
very time consuming and outright confusing chore at times.
So, o=
bviously, I opt to take the easy way out if possible! I'm not
calling any=
firmware functions at all, so I don't need anything from
the first 16KB o=
f address space once FyOS is running. If I could
just swap ram there, it =
would be a much more functional system. So
I'm mostly wondering if anyone=
has examined the hardware enough to
determine if this is very possible?
=