that is pretty nice 😀
I must be blind
but where do you write the new
BIOSMEM_SEG,BIOSMEM_MODESET_CTL
in the int10_modes.cpp ?
Or doesn't it need to be written anymore ?
Water flows down the stream
How to ask questions the smart way!
I've tried S3, ET4000 and Cirrus BIOS on HW and neither modify data at 488h/489h in a modeset.
1+1=10
wonder why we removed that one bit then in modeset....
That line has been present since the first public source release. I doubt that our private CVS server is still up.
Water flows down the stream
How to ask questions the smart way!
According to the interrupt list this bit is cleared on modeswitch. Now if we knew what system/VGA they derived this info from... possibly original IBM VGA. But who owned one of these anyway 😉
1+1=10
well Would it hurt to leave there ?
Water flows down the stream
How to ask questions the smart way!
A practical result of not resetting that data when setting the mode is that the selected number of lines (200/350/400) will persist unless specifically changed. Whether that is seen as a good or bad thing is subjective...
The s3 (86csomething) and elpin both modify 0x40:0x49 (same way as dosbox).
The other bytes are most likely unmodified for the reason ripsaw mentioned
(those special int10 functions are used to change the values).
Oh yeah we're talking about 0x40:0x89... sorry...
0x40:0x89 is not changed on the cards I tried.
1+1=10
Your diff is fine wrt those changes, but i didn't check the rest yet (scanline compatibility).
488 is also changed by INT10/BL=30h AX=1200h, that part is missing in my patch.
1+1=10
