I've been running NT as my primary OS as well as administrating it since the
3.5 days (circa 1995), and have yet to see a BSOD caused (even indirectly)
by a defective video driver. I have seen sub-optimal performance with
SKU-derived video drivers (particularly on Win2K and XP where OpenGL is
wrapped to use Direct3D until after installing the appropriate vendor
drivers). I have seen drivers that refused to get out of 640x480. I have
seen drivers that refused to get into a video mode / refresh rate I know my
system supported.

But I haven't seen a BSOD related to faulty video drivers.

Anyway, the only thing that moved into the kernel between NT 3.51 and NT 4.0
is the GDI and a goodly portion of the win32 sub-system (into win32k.sys);
the video drivers were always in kernel space from NT 3.1 days. And yes, if
you cross the ring 0 / ring 3 boundary less often, performance goes up. This
will be true of XFree86 once it provides a shortcut to ditch tcp/ip pipes to
communicate with local graphics hardware from local X clients, like SGI
does.

Andrew

----- Original Message -----
From: "Charles Shannon Hendrix" <shannon@widomaker.com>
To: "NetBSD-current Discussion List" <current-users@NetBSD.ORG>
Sent: Thursday, July 19, 2001 4:11 PM
Subject: Re: Why not track our xsrc with X11R6.6 from X.org?

[snip]
> > > We've also seen how stupid it is to put graphics drivers in the
> > kernel.
> >
> > We have? Where? Certainly not in the Unix workstation world we haven't
> > (at least not in any of the ones who survived!).
>
> Windows NT since 4.0 put the drivers in the kernel. It sucks, and it
> does definitely impact stability. They claimed you cannot get good
> graphics speed if the graphics code is in userland.
[snip]