#228 VDP Pirates shouldn't work on the TMS99xx VDP's

[Manuel Bilderbeek]

The VDP Pirates demo does not run properly on (at
least) one real Toshiba HX-10, while it does run on
openMSX's HX-10. Attached is a ROM and a screenshot of
the result on the real HX-10 (it blinks). See also
http://www.msx.org/VDP-Pirates-Demo.newspost3752.html

[Manuel Bilderbeek]

Screenshot on real HX-10

[Manuel Bilderbeek]

The demo

[Manuel Bilderbeek]

Logged In: YES
user_id=78178
Originator: YES

The same happens on my real Yashica YC-64. I couldn't run the demo on real MSX2's either: it hung with "VDP test" on the screen. I'm not sure if this is caused by my Flash device or another problem in this demo.

The VDP test works fine on the MSX1 machines.

[Manuel Bilderbeek]

Logged In: YES
user_id=78178
Originator: YES

I reproduced this problem also on my own Toshiba HX-10, exactly like the screenshot.

The demo works fine on a Sony HB-201P (which should also have a TMS VDP).

[Manuel Bilderbeek]

<hap> the cause is the extra mask on pattern table on 99xx
<hap> that doesn't happen on 91xx
<hap> (or v9938)
<hap> pattern mask becomes (p_mask & 0x3800) | (c_mask & 0x7ff)
<hap> instead of (p_mask & 0x3800) | 0x7ff
<hap> it is the feature mentioned here http://www.msx.org/forum/development/msx-development/new-hybrid-mode-found

[Manuel Bilderbeek]

Or, as MAME puts it:

jul 13 2014 19:48:46 <hap> .// on 91xx family, the colour table mask doesn't affect the pattern table mask
jul 13 2014 19:48:46 <hap> m_patternmask = ( (m_Regs[4] & 3) << 8 ) | ( m_99 ? (m_colourmask & 0xff) : 0xff );</hap></hap>

[Manuel Bilderbeek]

or https://www.youtube.com/watch?v=XJljSJqzDR0

or how meisei puts it:

            ct_mask=ct|0x3f;
            if (is_99xx) pg_mask=0x3800|ct_mask; /* additional pg mirroring on 99xx VDPs */

[Manuel Bilderbeek]

Implemented by Wouter and committed by me in commit 8651f152158e0828ab93486c948eea3c4a82de76

Thanks Wouter and hap!