Re: [Freemint-discuss] Disassembly of TOS 1.00 with Ghidra

[Vincent R. <vin...@fr...>]

On 16/01/2025 at 18:11, Thorsten Otto via Freemint-discuss wrote:
> BIOS of 1.00 can now be compiled to identical code.

Well done 🙂

Now you're only missing TOS 1.02.
As I said earlier, Ghidra *can* be used to match many labels from one ROM to 
another. And some editions of Atari ST Intern (at least my French one) has 
its BIOS disassembly, named "Mega ST TOS". That could help.

Due to Line-F hacks, that automatic matching doesn't work for AES/Desktop 
code. But I guess that with your recent work on processor patch, that could 
work better.

> I wonder whether they already had access to the sources? (the book is
> from 1988,
I guess yes.
Or at least a disassembly with labels.

Note that, at that time, the BDOS/VDI/AES/Desktop was still copyrighted by 
Digital Research (later open-sourced as GPL by Lineo). Maybe Atari just 
shared their own parts with those people? Namely, the BIOS? This was also 
easier, because it was obviously entirely manually written in assembly language.
Or maybe, is it possible that such commented disassembly was available with 
official Atari development kits? I have no idea.

> but there is also an earlier edition from 1986).
Do you mean this one?
https://www.synacktiv.com/ressources/Atari-ST-Internals.pdf

At the end, there is the disassembly of an early TOS 1.0 with a build date 
of 20/06/1985.

Note that in this early version, the ROM is located at address 0x5000 
instead of standard 0xfc0000. Pretty weird, definitely not a stock ST.

And at page 297 of the PDF, we can see it's the ancient "Mushroom" version, 
with mushroom clouds instead of bombs.

> And another small gimmick: the data of the bomb image differs by 1 pixel 
> compared to later versions.

Funny.

-- 
Vincent Rivière