Re: [hatari-devel] Reeking rubber analysis

[Laurent Sallafranque <laurent.sallafranque@xxxxxxx>]

Ah, ok
I'll have to test them on my falcon one of these days.

But at least, I think the ST palette may need the $fff mask.
I've thrown a little look at video.c and it seems that the st palettes 
are masked ($777 for stf palette and $fff for ste palette).
So I think it may be the same for the Falcon ST palette (but I can be 
wrong here).

Your advice ?

Regards

Laurent


Le 09/09/2013 21:54, Thomas Huth a écrit :
> Am Mon, 09 Sep 2013 21:26:21 +0200
> schrieb Laurent Sallafranque <laurent.sallafranque@xxxxxxx>:
>
> > Hi,
> >
> > I wanted to do the same patch tonight, but as you've done it, it's
> > perfect. I'm happy this game works under hatari now.
> >
> > By the way, while having a look at the videl code, I saw that many
> > registers should be set to 0 on certain bits (like for the falcon
> > palette).
> >
> > For example : in the StColorRegsWrite, a $fff mask should be applied
> > to the colors
> >
> > The same for the registers that don't use all the bits.
> >
> > Do you agree with that ?
> Well, it depends on how the register is implemented in hardware. Unused
> bits can either be tied to 0, or to 1 or not be implemented at all - in
> that case, the result depends on the traffic that has been sent via the
> bus before. We've seen that with some registers on the ST already,
> especially the ST shift mode register if I remember clearly.
>
> So for emulating unused bits properly, you would need to test them on a
> real Falcon to see whether they are 0 or 1 or "random".
>
>   Thomas