Re: [hatari-devel] Sparrow TOS 2.07 support

[Thomas Huth <th.huth@xxxxxxxxx>]

Am Sun, 27 May 2018 11:56:14 +0200
schrieb Christian Zietz <czietz@xxxxxxx>:

> Thomas Huth schrieb:
> 
> > Anyway, it does not boot in Falcon mode. I've run it with "--trace
> > cpu_disasm", and it seems like TOS jumps back to the beginning quite
> > soon:
> > 
> > 00E037CE 2078 0004                MOVEA.L $0004 [00e00030],A0
> > 00E037D2 4ed0                     JMP (A0)
> > 
> > No clue why it is doing that yet. If somebody wants to have a look:
> > I've attached the trace.  
> 
> I have also played around with Miro's TOS 2.07 dump. There is an
> unhandled bus error accessing $FFFF8E0D, a hardware register that does
> not exist on the Falcon. According the documentation on the Sparrow
> that I have access to, this is the SCU, like on the MSTE/TT.

You're right - I initially missed that one ... I've added now a TOS
patch for this location to simply skip that access.

> All subsequent bus errors are handled. Thus, after patching out this
> one access I got TOS 2.07 running on Hatari to the point that the
> familiar keyclick sound is heard, when a key is pressed. However,
> there's no screen display; probably because the video hardware of the
> Sparrow is completely different from the one of the Falcon.

The Sparrow TOS apparently really expects a video hardware that is more
close to the STE than Falcon. It does not set up most of the Falcon
registers and simply only writes a value to $ff8260 to set the
resolution. I've added a hack to the Videl emulation, and now it boots
to the desktop. Looks pretty close to TOS 2.06, indeed.

By the way, I don't have much experience with real Videl programming,
but does anybody know whether changing $ff8260 also influences the
values in $ff82a4 - $ff82aa on real hardware?

 Thomas