COMMODORE ROM PROJECT

From: Craig Taylor <ctalkobt_at_gmail.com>
Date: Fri, 3 Feb 2012 08:03:58 -0500
Message-ID: <CAE6OHdWE1YMN2s0KzQidb83jMkGU-2VysmkKJ68MW6Mg8EQtHw@mail.gmail.com>
This topic reminds me of a basic extended that Trsnsactor published
extensions for.  The User could load into memory his custom set of
extensions.

If direct kernel access is limited - eg, jump table, few other direct
common access routines -- then modules could register themselves for a
custom kernel.  Stock Roms could be generated to always boot an "unit.Kern"
file enumerating the modules.

In this convrrsation, basic "essential" and boot from disk would be loaded.

On Friday, February 3, 2012, Gábor Lénárt <lgb@lgb.hu> wrote:
> Hi,
>
> On Fri, Feb 03, 2012 at 08:32:36AM +0100, Baltissen, GJPAA (Ruud) wrote:
>> Hallo Gábor,
>>
>>
>> > I guess it's a bit overkill to have a compilable (and working!)
>> > source of the kernal
>>
>> If it is a lot of work to generate it, yes. But I have a disassembler
that, with a bit of tweaking, produces readable and in most cases,
relocatable sources.
>
> Wow, ok then :) I am curious how it can be done, since it's not an easy
> task: kernal has entry pointes at fixed addresses used by various
programs,
> also the not so official ones (which should not be used, but maybe some
> softwares use them), so if you modify the source these addresses will be
not
> the same anymore.
>
> How can you do that, to have an assembly source which will produce as much
> "fixed points code" as possible with checking that no "overrun" of a code
> part if someone want to push "too much" before the next "should be fixed"
> address? It's OK that kernal has entries like CHRIN, etc, but I think
there
> are much more "odd" usage of the kernal with using not so official
routines
> to call (hmm so "private" functions designed to be used by the kernal
> itself).
>
> Anyway, my answer is still valid I guess, it's even easier with the source
> of course :) If someone does not need tape routines, there is some place
to
> implement other stuff there (however even DTV's kernal has bugs, not every
> vectors etc are patched which should so some programs crashes because of
it
> - it's documented for some degree at the links I've mentioned in my
previous
> mail).
>
>
>       Message was sent through the cbm-hackers mailing list
>


       Message was sent through the cbm-hackers mailing list
Received on 2012-02-03 14:00:08

Archive generated by hypermail 2.2.0.