Re: [hatari-devel] TT emulation crashes when there is no ACSI drive

[Eero Tamminen <oak@xxxxxxxxxxxxxx>]

Hi,

On 8/6/19 9:40 PM, Eero Tamminen wrote:
> On 8/6/19 7:15 PM, Christian Zietz wrote:
> > Eero Tamminen schrieb:
> >
> > > If you look at the cartridge code in my mail, stack pointer is first
> > > moved to a0, and after saving a6 to stack, a0 is moved into a6, then
> > > the extra subroutines are called, after which comes the failing
> > > instruction:
> > >           clr.l   2(a6)
> >
> > But still I don't see how $23000000 could end up in A6.
>
> Oh, you mean that a0 i.e. stack pointer needs to have been corrupted
> before cartridge code gets called.  I.e. TOS has corrupted it somehow,
> or Uwe has some other program in AUTO which Pexec's NF_SCSI.PRG with
> corrupted stack pointer?

No, that's nonsense too, as then the subroutines before crashing
instruction would crash instead.

Because Uwe is using GEMDOS HD, he can't be using any other cartridge
code.

Only other things I can think of are:
* Uwe running some MMU program that maps something else
  to cartridge area.  Is that even possible?
* Something seriously wrong in Uwe's last Hatari build...?


	- Eero

> (Sorry, I'm very rusty on asm, as I haven't written much of it myself
> even back in 90's.)
>
>
> > > The place where bus error happens, is during program loading, not
> > > its execution.  So NF SCSI shouldn't actually be needed to reproduce
> > > this, *if* the crash actually happens at NF_SCSI.PRG loading.
> >
> > OK. I tested with a current Hatari Windows build from antartica.no,
> > Uwe's config file, TOS 3.06, and NF_SCSI.PRG (from Uwe's web page) in
> > the AUTO folder. No crash! (Of course, after loading, NF_SCSI.PRG
> > complains that the Natfeats SCSI feature is not available and terminates
> > itself.)
>
>
>      - Eero