Re: [hatari-devel] Extended VDI screen in Falcon mode

[Thomas Huth <th.huth@xxxxxx>]

Am Mon, 4 Feb 2013 14:32:56 +0200
schrieb Eero Tamminen <oak@xxxxxxxxxxxxxx>:

> Hi,
> 
> On maanantai 04 helmikuu 2013, Vincent Rivière wrote:
> > On 04/02/2013 00:31, Roger Burrows wrote:
> > > EmuTOS allocates memory for screen RAM at the top of Falcon
> > > memory.  It reserves 300K which is the largest amount it needs
> > > for any legal screen size. I suggest that Hatari limits its
> > > extended screen resolutions accordingly.
> > 
> > The problem is similar on ST, where the allocated screen memory is
> > always 32 kB. Not enough for extended resolutions.
> > 
> > IIRC, Thomas said that that he used a very clever trick to allow
> > extended resolutions. Since the screen is always at the very end of
> > the ST-RAM, he just artificially extends the ST-RAM to fit the
> > extended resolution requirements. Very simple and efficient.
> 
> Thomas, where in the code that happens?   I see only Hatari's internal
> ST screen conversion buffers to use max. VDI screen size, and
> VDI/Line-A variable changing.

IIRC it's the other way round, Hatari is not extending the ST memory,
it is decreasing the memtop system variable so that TOS does not touch
the upper range of the memory anymore (except for screen accesses of
course).

Have a look at stMemory.c and search for "memtop" there.


> > Hatari should use the same same strategy for Falcon extended video
> > modes. The situation is similar, anyway.

I guess that memtop hack does somehow not work with TOS 4.0x anymore.
Maybe TOS places the screen at other memory locations, too, not only at
the top of the RAM?

 Thomas