> Btw, is it possible for the kernel to initialise VGA cards if the
> firmware neglects to do it? My test would have been more reliable if I
> could have seen what was happening...

It is in theory possible, however such code needs to be written, and (as
always) is subject to availability of documentation and/or sample source.

I've found S3 to be pretty cool about sending me old data books and CD's
for their Trio and ViRGE series chips. Many of those come up after reset
in a PCI-friendly state, but once the VGA BIOS runs, the card locks into
icky legacy-compatible mode.

Unfortunately, I've discovered that my application (multi-headed machine
for info kiosks and data entry counters) will require plenty of wscons
work even after we get multiple vga* drivers attaching, so I've had to
scrap it for this year. I'd still like to do it though, but I only care
about text mode, so you may want more than what I might eventually come
up with.

XFree 4 apparently is doing a lot of research into this area so they can
do multi-monitor desktops; I don't know how far they've got however.

Todd Whitesel
toddpw @ best.com