Re: [hatari-devel] Some findings

[Thomas Huth <th.huth@xxxxxxxxx>]

Am Sat, 01 Sep 2018 09:34:00 +0200
schrieb Thorsten Otto <admin@xxxxxxxxxxx>:

> On Samstag, 1. September 2018 09:03:09 CEST Thomas Huth wrote:
> > Did you move the mouse or press any keys during that phase?  
> 
> I pressed keys, yes, but i can't remember whether i moved the mouse.
> 
> >I think that's kind of expected. Hatari only supports the serial port
> >that is wired to the MFP so far,   
> 
> I've have set debug device in mint to 1, that should be AUX, the
> serial port driven by MFP, and present in all Ataris. There was no
> error message or hang, just nothing written to the file (a regular
> file on the host, not some device).

The serial port of the MFP is not connected in the Falcon, so I'd be
surprised if the Falcon BIOS and MiNT would try to use it.

> >Known issue,  
> 
> Yes, expected that. Just a bit strange that it worked with EmuTOS,
> which also has very rudimentary support for serials, if at all. Maybe
> it would be an option to allow dummy accesses to the SCC registers if
> EmuTOS is detected, to avoid crashes in MiNT later.

We've discussed that earlier this year. Dummy access might have other
disadvantages instead (e.g. software might hang because it polls for a
bit that does not behave correctly).

> But MiNT should
> also be fixed to not access them, instead of blindly assuming that
> they are present only because the _MCH type is TT.

.... or simply finally implement SCC emulation in Hatari, which would
certainly be the best solution ;-)

 Thomas