Re: MiNT VM on Falcon

[Konrad Kokoszkiewicz <conradus@orient.uw.edu.pl>]

> > needed to fix vm5.spp, because FAlcon IDE = SCSI dev. 16 (mask needs to be
> > changed from $0f to $1f). I am including also other changes i did in mint
> > (not much of them at all, only details).
> 
> why not send it here as unified diffs?

Because the patch is incomplete (i.e. it causes the VM to start on Falcon
IDE, but nothing more), so it can't be considered as a real patch which
improves anything... 

> 
> > 4) if you fill up the ST RAM with programs (so that any next program is
> > loaded to the VM), any program loaded this time crashes. I.e it bombs with
> > strangely low PC value (e.g. $11xxx or something similar). I have no idea,
> > what code is there (it may be NVDI or MiNT or any TSR).
> 
> sounds like a TSR has been swapped out ;-)

It is rather a fault in the MiNT VM kernel...

> Anyway, it sounds like the MiNT 1.14 is not usable at all. Is anybody
> around with MMU knowledges to fix it?

Well, MiNT 1.14 is fully useable as long as you don't use VM ;) I know
that the VM is supposed to be the only difference between 1.12.6 and 1.14,
but I noticed, and it refers at least to my machine, that 1.14 is more
stable than 1.12.6 when used as MultiTOS kernel (i.e. with AES 4.10): I
have a program that causes a lockup when ran on MiNT 1.12.6 + AES 4.10 and
is running fine on 1.14 + the same AES. The proggy does nothing, just does
an usual startup code (Mshrink(), appl_init(), v_opnvwk()) then executes a
infinite assembler loop. I wrote this to check if in the MultiTOS you have
a real preemptive multitasking and was really surprised when MTOS 1.12 got
locked up dead.... :))))

Bye
Draco