Re: [hatari-devel] It seems I encounter a bug into hatari

[laurent.sallafranque@xxxxxxx]

Hi,

You're right Thomas.
It seems I can abuse these registers in word size (Racer works well to confirm that), but not in long words.

I've written a little program based on Nicolas 's asking which displays the values of the registers $ff8240 and $ff8242 when I first write a long word into them.

I'll send you the result values on both my falcon and Hatari tonight.

Regards
Laurent




----- Mail original -----
De: "Thomas Huth" <th.huth@xxxxxx>
À: hatari-devel@xxxxxxxxxxxxxxxxxxx
Envoyé: Jeudi 27 Mars 2014 16:38:36
Objet: Re: [hatari-devel] It seems I encounter a bug into hatari

On Wed, 26 Mar 2014 19:15:16 +0100
Laurent Sallafranque <laurent.sallafranque@xxxxxxx> wrote:

> Hi,
> 
> While testing racer on my real hardware, I encounter something that 
> works under hatari but not under the falcon (with both old and new cpu)
> 
> The piece of code given above displays the sky (one color per line).
> 
> Under hatari, the sky displays correctly, but under the Falcon, it seems 
> the color palette pointer is wrong and the sky is displayed with bad colors.
> 
> Is it related to the fact I use the ffff8040 + memory address in long mode ?
> 
> At least, Hatari and the real hardware differs here.

I don't think that you can abuse the color registers for storing
arbitrary data since there are likely some bits in the register that
are not implemented in hardware (either tied to zero or random when you
read them back). We likely just lack the masking of the color registers
in Falcon mode in Hatari, that's why it's working by accident there.

 Thomas