On 05/02/2001 07:09:37 PM CST Lawrence Gold wrote:
>
>On Tue, May 01, 2001 at 01:38:18AM +0200, Christian Groessler wrote:
>>
>> Since this a 8bpp display I'm a bit confused as to why this is
>> necessary.
>>
>> So I'm not suggesting to use this modification, I'm curious as to why
>> it is needed on my machine. Maybe something else is misconfigured.
>
>That's really odd. I would normally suspect some sore of implicit
>little-endianness in the core, but as you point out, this is an 8bpp
>display.
OK, I think it has something to do with the bit depth.
I'm using XFree 4.0.3 with a ATI R128 card, testing mostly with the
games pacman and mslug.
The DGA mode selected is DGA 2.
When I said, it's an 8bpp display I did this because of the message
from xmame "XF86-DGA2 running at: 8bpp"
Also, the 8-to-8bit blit-ops are used, as I changed one of them to fix
the display.
I didn't follow the recent DGA discussion very carefully, but I think
DGA2 is able to change the display bit depth "on the fly", which DGA1
cannot do.
I had in my XF86Config a "DefaultDepth 16" line, so I the default
display bit depth should be 16. The bit depth switching of DGA2 (or
xmame) is not complete, because the display somehow remembers it's old
16bit-ness and wants the (byte-sized) display data apparently as 16bit
values (which breaks on ppc due to endianness issues).
When I set "DefaultDepth 8" in my XF86Config, it works perfectly.
Also when I change mame to use DGA1 only.
When I set "DefaultDepth 24" in XF86Config, the garbledness is even
greater, it appears that then not every 16bit, but every 32bit value
should be swapped.
I cannot decide now whether the bug is in xmame or XFree. Does anyone
know some other DGA2 progs which do similar things like xmame to check
out and empirically decide where the bug resides? :-)
regards,
chris
_______________________________________________
Xmame mailing list
[EMAIL PROTECTED]
http://toybox.twisted.org.uk/mailman/listinfo/xmame
