[hatari-devel] BITS demo : thread from atari forum
Laurent Sallafranque
laurent.sallafranque at free.fr
Tue Jul 12 23:28:34 CEST 2011
I've got a 520 STF and a Falcon.
I can try the test on both computers, but I'll have to connect them first.
(some tests after the release is OK for you ?)
Regards
Laurent
Le 12/07/2011 23:14, Nicolas Pomarède a écrit :
> Le 12/07/2011 20:00, Anders Eriksson a écrit :
>> On Tue, 12 Jul 2011, Nicolas Pomarède wrote:
>>
>>> Either way, it would be great if an STE's owner could dump the memory
>>> region from $ff8200 to $ff827f (Anders maybe ? ;-) )
>>
>> Hi!
>>
>> 48 bytes did buserror (from offset $0f) so I left that area blank. The
>> remaning 80 bytes are in the file. See the source in case I did
>> something wrong.
>>
>> http://ae.dhs.nu/tmp/regdump.zip
>
> Thanks also for this test.
>
> From what I see, Hatari STE's mode is OK :
> - hatari has bus error between ff8210 and ff823f (Thomas wrote a
> program to check those bus errors area)
> - hatari sets unused byte (even addresses) between ff8200 and ff820f
> to $ff
> - sync byte at $ff820a has unused bits 2-7 set to 1 on STE (which is
> different from $ff8260 on STE), OK on Hatari
>
> - unused bytes $ff820b returns 0 on STE, but $ff on Hatari
>
> - colors reg between $ff8240 and $ff8260 sets unused bits 12-15 to 0
> on STE, this is OK on hatari (on STF, those bits seem to be randomly
> set to 1)
>
> - res byte $ff8260 has unused bits 2-7 set to 0 on STE.
>
> - scroll bytes $ff8264/65 have nothing special
>
> - all other unused bytes between $ff8260 and $ff827f are set to 0 and
> not to $ff (as the unused bytes between $ff8200 and $ff820f) ; hatari
> is not correct, it returns $ff.
>
>
> The cases were Hatari returns $ff instead of $00 for unused bytes is
> not good, but it should be harmless regarding emulation results.
>
> I will add a 'todo' about this for the next Hatari version.
>
> It would need to be confirmed on real hardware, but as Mega STE, TT
> and Falcon were released after the STE, the same rules might apply for
> bytes that return 0 instead of $ff on STF.
>
>
> If someone here on this list can run Anders regdump.tos program under
> STF / TT / Falcon or any non STE machines, it could interesting to
> compare the results.
>
>
>
> Nicolas
> _______________________________________________
> hatari-devel mailing list
> hatari-devel at lists.berlios.de
> https://lists.berlios.de/mailman/listinfo/hatari-devel
More information about the hatari-devel
mailing list