Index: [thread] [date] [subject] [author]
  From: Jon M. Taylor <taylorj@ecs.csus.edu>
  To  : ggi-develop@eskimo.com
  Date: Thu, 2 Jul 1998 14:42:17 -0700 (PDT)

Re: KGI Drivers: dali->degas transition

On Thu, 2 Jul 1998, David Waite wrote:

> -----Original Message-----
> From: Rodolphe Ortalo <ortalo@laas.fr>
> 
> >So, what's the status of the "Degas KGI drivers writer guide" ?
> >What about this fbcon<->KGI and its use of dali drivers ?
> >What should I do to have my Laguna board working again ?!!!
> 
> Thats my question too. Will it (fbcon) use dali or degas drivers? I've heard
> both from people. 

	Modified Dali.  I am cleaning up the Dali KGI driver stuff by
cutting out all the unneeded includes (serialmouse.h, etc) and renaming
some stuff from ggi* to kgi* to make it all neat and tidy and consistent. 
Autoconfig is gone for now because the required utilities are no longer
present - this should be added back, but using the PNP facilities like the
sound card drivers do.  The driver module is called kgicon.o.

	The moved include directories will necessitate some changes in the
drivers and makefiles, but it should all be mindless so either I will
grind through it myself for all the drivers or write a perl script to make
the necessary changes.  The only tricky point may be the removal of the
input- and console-related elements of some structs if initializers count
on their presence.  But it needs to be done.  KGI and GGI includes and
function naming need to be completely separated and right now they are
not.

> I don't really care if the driver format changes, as long
> as they are removable (so I can write drivers without rebooting whenever
> they don't work).

	They are already removable in theory, but the current Dali drivers
have some problems cleaning up after themselves (e.g. the Trio64V+ driver
leaves the card in MMIO mode).  Also I had to disable the I/O port
allocation code in the Trio64V+ driver as it complained that the existing
fbcon VGA+ driver had already allocated the VGA ports (which was correct
of course).  

	The result of this is that once I insert the Trio64V+ driver and
migrate the consoles to it, I cannot migrate them back to the old VGA+
driver and get my display back.  Technically, the VGA+ driver and the
Trio64V+ driver are trying to drive the same piece of hardware
simultaneously, which should be a no-no.  The inital display hardware
should always be controlled by a compiled-in driver. 

Jon 

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

Index: [thread] [date] [subject] [author]