Path: utzoo!attcan!uunet!cbmvax!valentin
From: valentin@cbmvax.commodore.com (Valentin Pepelea)
Newsgroups: comp.sys.atari.st
Subject: Re: TOS 1.6 and the 68030
Message-ID: <9299@cbmvax.commodore.com>
Date: 11 Jan 90 20:15:20 GMT
References: <1149@crash.cts.com>
Reply-To: valentin@cbmvax.commodore.com (Valentin Pepelea)
Organization: Commodore, West Chester, PA
Lines: 58

In article <1149@crash.cts.com> bradm@pro-grouch.cts.com (Brad Martin) writes:
>
>I have heard rumors (God I hate rumors) that TOS 1.6 would allow the STe to
>work with an 68030 processor (and perhaps an 1040ST or Mega if outfited with
>1.6).  I would like to know if any of the Atari reps here on the net could
>comment on this.  I have a friend who works for a company who makes 68030
>boards for 'other' computers who could have an '030 board out for the ST way
>before another well publisized project (Hi Dave :-) ).  If this is not true, I
>will have to convince them that they want to write the software to make it
>compatable

Well, I too work for a company that makes 68030 boards, though not likely to
ever produce ST boards.  :-)  Now I don't know how TOS is written but I can
tell you what the difference in the processors are.

All user-mode instructions, as opposed to supervisor-mode, are fully upwardly
compatible from the 68000 up to the 68040. Except for the MOVE <ea>,SR
instruction of the 68000 which is now executable only in supervisor mode on
68010+ processors.

This however can be easely fixed through software. Executing this instruction
in user mode on 68010+ processors will cause a priviledge violation exception,
and an exception software handler can substitue the right code there.

Another difference between the 68000 and 68010+ processors is the exception
stack frame created during an exception, such as interrupts. The differences
are small and well documented, but they have to be taken into account anyway.
If Atari provides source code to its OS, then fixing this should be no problem,
particularly on 68030+'s. You see, the 68030 has a built-in MMU, thus allowing
you to load a fake image of ROM in RAM and then translating all memory accesses
to the ROM area to the RAM area.

If source code to the OS is not provided, you're screwed, unless the OS already
takes account of the different exception stack frames created on the different
processors. You can probably disassemble the exception processing code yourself,
since the exception processing routines have their pointers starting at
location 8.

Of course, software which depends on busy-waiting timing loops will be
completely out of wack. So even if you get your act together, that does not
mean that third party software out there does too.

(how would you like to have you ST running at 40mhz?).

Don't hold your breath. Even if you get a processor running at 40MHz, the rest
of the hardware will not. You would have to provide your own 32-bit memory board
to go along the 68030, and it will not be accessible to the blitter or any other
device at high speeds.

By the way, if your friend's company would like to make a board for this 'other'
computer I work on, you know who to have him contact.  :-)

Valentin
-- 
The Goddess of democracy? "The tyrants     Name:    Valentin Pepelea
may distroy a statue,  but they cannot     Phone:   (215) 431-9327
kill a god."                               UseNet:  cbmvax!valentin@uunet.uu.net
             - Ancient Chinese Proverb     Claimer: I not Commodore spokesman be