From: Jon M. Taylor <taylorj@ecs.csus.edu>
  To  : ggi-develop@eskimo.com
  Date: Sun, 9 Aug 1998 15:33:51 -0700 (PDT)

Re: kgicon problem

On Sun, 9 Aug 1998, Brian Julin wrote:

> On Sun, 9 Aug 1998, Dirk Lattermann wrote:
> > However, when I insmodded kgicon.o (using the insert script), the
> > system froze, showing vertical grey bars with some coloured dots in the
> > upper half. 

	Same here with the Trio64 driver.  The grey bars+colored dots 
thing lasted of a couple of seconds and then it all went black.

> > Did I make a mistake or is vga-kgicon, or the kernel,
> > still buggy? I had the dali vga driver running, some time earlier!

	My driver worked fine until I upgraded to 2.1.115.

> The 4bit modes with the std vga driver currently lock the machine.
> Same symptoms you describe.

	I thought the machine had locked, but as I was reaching for the
big red button shutdown finally kicked in.  I didn't have any fs
corruption, but I also only got the first two lines of the oops trace in
my syslog too.  Very strange.

> See my post a couple of days ago on "vga kgicon problem summary" for 
> how to get it working with 1bit modes.  

	640x400x1 == 32K == the FB memory size that the vgacon driver
uses.  Coincidence?

> Note though that there is a bug that causes a kernel oops in
> the fbcon renderer for 1bit modes; probably the same one that
> happened to people in 8-bit modes when they started lynx.  It
> happened to me in emacs tho.
> 
> Is this bug fixed under 8-bit modes now?  If so, can anyone
> venture how to fix 1-bit modes similarly?

	Nope.  Debugging this stuff is hard because it gets called so
much.  If you put any printk()s in those renderers you'll fill your hard
drive up with your syslog if you aren't careful.

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