On Mon, 1 Sep 2025 11:47:49 -0500, Paul Edwards <[email protected]> wrote:

>Basically with everything written in terms of C90,
>everything is everything. Pretty much any API
>you throw at it - even the MVS API (SVC 120 to
>get memory etc), I convert it back to a call to
>malloc() and we're back to where we started -
>anything, anywhere. I haven't done that work yet
>though. Other than proof of concept that I can
>override an SVC in userspace with cooperative
>programming (read: just PDPCLIB).

Oh - and so that means you'll have an ASCII version of
(a mini) MVS too. Potentially MVS hasn't been emphatically
proven conceptually. Specifically I need to show that from
C I can have an assembler callback function for use in a
RDJFCB exit or whatever that will get me back to C code,
and in the right AMODE, and I can control it from C code.

But it's a pretty sure bet that it's a "simple matter of
programming". There's no magical feature that can't
be replicated with a bit of supporting assembler. But
I haven't looked into all the internals of mvssupa.asm -
maybe every single register is accounted for and I
can't possibly do something from C. But I don't believe
there are any missing concepts left for me to complete
what I want to do.

Note that the technique for overriding an SVC (which I
thought was impossible), only came about when I was
trying to do the same thing on the Atari - I thought it was
impossible to override a 68000 trap from usermode. So
I could do the Atari, which had a C interface to the OS,
but Atari was impossible.

No. It wasn't impossible. And that opened up MVS SVCs.
It was someone in the Atari forum who pointed out how
I could do it within my listed constraints - I didn't invent
it myself. And the whole PDOS-generic design was
invented in an Amiga forum. And a lot was achieved trying
to force the XT to work too. And MVS was a big influence
the whole time too. And nothing was done in a vacuum.

BFN. Paul.

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to