The BIOS on this VGA card shouldn't be there...

[per]

One of my VGA cards are a bit weird. Instead of having the BIOS at segment D000 where most other VGA cards got it, it's placed at segment C000.

The weird thing is that it successfully mannages to take over int 10h and int 6Dh, even though the XT's BIOS only scans for extensions from offset C800 to F400. I can't see any way how it is able to do it, it can't really do an IRQ without hooking an interrupt first, and it might be a little unsafe for the card to take controll over the system bus address and data lines (if possible at all). Maybe it's DMA... I don't know...

Another fact is that the VGA-copyright date/manufacter shows up on the monitor before the memory test, but that might have been set up by the graphics processor itself.

 

[Chuck(G)]

C000 is special-cased for video BIOS ahead of all other BIOS expansion. It's where the IBM EGA put its code, after all.

It just makes sense. You want your adapters to be able to display messages, but how can you do that if the display adapter's not initialized?

 

[Marrr]

One of my VGA cards are a bit weird. Instead of having the BIOS at segment D000 where most other VGA cards got it
[...]

There must be some mistake there...
Most (or even all) EGA/VGA/SVGA cards have BIOS at segment C000, typically 32 KB from C0000 to C7FFF.

 

[per]

There must be some mistake there...
Most (or even all) EGA/VGA/SVGA cards have BIOS at segment C000, typically 32 KB from C0000 to C7FFF.

Sorry, my bad.

I also figured about why it is loaded. Instead of searching for the modules between C000 and C780 when the other modules (C800 to F580) are being searched for, it happens in the end of the video-check routine.

 

[Chuck(G)]

Sorry, my bad.

I also figured about why it is loaded. Instead of searching for the modules between C000 and C780 when the other modules (C800 to F580) are being searched for, it happens in the end of the video-check routine.

What I said.