From: Brian Julin <bri@forcade.calyx.net>
  To  : ggi-develop@eskimo.com
  Date: Mon, 3 Aug 1998 08:55:47 -0400 (EDT)

Re: More progress with kgicon

On Sun, 2 Aug 1998, Jon M. Taylor wrote:

> Jonas writes:
> 
> >>On Sun, Aug 02, 1998 at 01:21:27AM +1000, Andrew Apted wrote:
> > Oh, sorry I thought everything was supposed to set up the MMIO
> > information probably (except maybe the stock VGA driver).

> 	Also, in fbcon-kgi.c in kgi_init(), you commented out the call to
> kgim_check_mode(), thus passing the default kgi_mode set by your chipset
> driver through unmodified.  Problem is, that kgi_mode is set by the
> chipset driver at chipset init time, and the chipset doesn't know anything
> about the other driver subsections.  

What he said.  I hacked it by importing graphics_setup_mmio from
the accel driver (someone missed renaming it to accel_setup_mmio.)
And actually got my first legible output from kgicon yesterday,
but it selected an 8bpp rendering function set despite both kgi and fbcon
registering the console as 4bpp.

The MMIO base can be different between two modes for tiny reasons 
internal to the driver.  If it is necessary to pass the base mmio 
address to fbcon rather than wrap fbcon around a native kgi scroller,
(the only reason I could see is that the scroller's not implemented,
or to make /dev/fb compatible for graphics access rather than using
kgi ioctls to get a /dev/graph) this must be after the mode is 
actually set, or at minimum checked.

Also note that mode.fb although defined as a mem_vaddr is actually
given a mem_paddr by the vga driver and lots of others.

Finally, don't expect banked modes to map through to /dev/fb*.  fbcon
kernels may lack an equivalent to fast_remap_page IINM; haven't looked tho.

The vgafb is doing something to the font plane locations or
memory address translation circuitry such that setfont doesn't work; 
neither before or after kgicon is inserted/removed.  It's not
that the PIO_SETFONT ioctl is broken because I've poked every
location in the device MMIO window without hitting the font tables.
Haven't figured this out what yet but will keep trying after work...
maybe with more new bugfixes?? :)

(P.S. right now using 2.1.113 sunsite kernel.)

--
Brian S. Julin