prlw1@newn.cam.ac.uk said:
> pcdisplay0      at isa?                 # CGA, MDA, EGA, HGA 
> [...]
> With such a kernel, I only get 1 console, and ctl-alt-fn doesn't do anything.

That's right. The pcdisplay driver is a minimal last-resort thing.

> No X (as opposed to just X)

It should be possible to run X. Perhaps it tries to use a specific vt
which doesn't exist?

> On a side note, I get a rectangular steady cursor which follows the booting
> text on the bottom line, but I also have a flashing underline cursor which
> stays put at 0,2 from bottom left hand corner.

I'd guess this is a bug in the SOFTCURSOR code which went in recently.
Mostly harmless...

Anyway - we know now that there is not another driver stepping on
your ATI card but something in the vga driver it doesn't like.
If we get it narrowed down, we could eg add an ATI specific quirk.
Can you try to comment out some suspicious parts in vga.c?
I'd start with the monochrome detection - you could force
the "mono" variable in vga_common_probe() and the vh->vh_mono flag
in vga_init() to zero.
Besides this, you could replace vga_copyrows() by pcdisplay_copyrows()
in vga_emulops.
If this doesn't help, we'd have to kick out the parts dealing
with "dispoffset", but hopefully this is not necessary.

best regards
Matthias