From: Jon M. Taylor <taylorj@ecs.csus.edu>
  To  : ggi-develop@eskimo.com
  Date: Wed, 29 Jul 1998 13:03:25 -0700 (PDT)

Re: kgicon, etc

On Thu, 30 Jul 1998, Andrew Apted wrote:

> Brian writes:
> 
> >  Umm. Are you implying that it should work with a standard kernel 
> >  off of sunsite, or did you forget to mention the vger patch?
> 
> I'm using vger, but AFAIK the standard kernel has all the necessary
> bits.

	Works for me.

> >  /dev/fb seems broken from a straight sunsite 2.1.111 still
> >  (FBIOPUT_VSCREENINFO fails all the time.)
> 
> That's the "set mode" ioctl, which doesn't work on VGAFB or VESAFB
> (which are fixed modes), nor with kgicon (yet :->).

	fbcon-kgi.c does not handle ANY ioctls right now.  Those ioctls
that work are handled by the fbcon layer itself, not the driver.  The
function to handle the ioctls is just a stub that returns INVALID (or
whatever it is) or all ioctls.  Until recently, almost none of the
existing fbcon drivers handled any ioctls either, so I didn't have
examples to steal^h^h^h^h^hlearn from.  

	Even now, most fbcon drivers still don't handle any ioctls, and
those drivers that do handle diferent ioctls sets and some of the (like
amifb.c, the Amiga driver) seem to handle only Amiga-specific ioctls for
setting the shape of the mouse cursor and things like that. It will not
take too much work to make fbcon-kgi.c more conformant to the fbcon/fbdev
API than any of the other drivers.  Even now, it does a lot of stuf that a
lot of other drivers do not, like supporting all bit depths.

Jon

---
'Cloning and the reprogramming of DNA is the first serious step in 
becoming one with God.'
	- Scientist G. Richard Seed