> My guess it that is an agp(4) problem because this driver is known to
> be buggy. If you are an experienced C programmer you could compare
> the i830 support in FreeBSD and NetBSD. Maybe you'll find something
> there which our driver gets wrong.

this sounds like something out of a bad sg-1 episode:
    
			       DANIEL JACKSON
    
    There are two transcripts here. One written in the language of The
    Ancients, the other written in an obscure G'ould dialect. If we can
    determine the difference between them, then we should be able to get
    the hyperdrive working again.
    
				JACK O'NEIL
    
    Daniel, thanks, very helpful. Any other suggestions? Anyone?
    
/mtr