From: Jon M. Taylor <taylorj@gaia.ecs.csus.edu>
  To  : ggi-develop@eskimo.com
  Date: Fri, 6 Aug 1999 17:26:07 -0700 (PDT)

Re: Accelleration: A summary

On Fri, 6 Aug 1999, James Simmons wrote:

> 
> > 	Absolutely.  As soon as I can get some time to hack KGIcon some more,
> > one of the first items on my list of things to add/port is the PP buffer
> > stuff (well, after the essential management facilities, of course).  It will
> > be nice to see accels scream on KGI drivers.
> 
>  It would be nice to add PP buffer to KGIcon but I was told that mixing
> accel handling with /dev/fb would not be accepted 

	KGIcon is distributed as modules and does not require a kernel
patch, so no acceptance is required.  And basically anything non-fbdev
related will be pushed out into a dynamic /proc-based tree, so we won't 
need to get Linus to assign us major and minor device numbers or anything 
like that.

> but handling accels with
> a second driver (/dev/gfx) would be. 

	James, I appreciate very much all the work you are doing and all
the crap you are putting up with working with the kernel developers. 
However, I cannot _depend_ on /dev/gfx getting past Linus, nice though it
would be if/when it happens.  Fast command FIFOs are 100% necessary to
have a fast direct rendering model and the PP buffers are here now with
code (KGI/Dali) and an example driver implementation (MGA 1064 driver).  I
have to use these now.  

	When /dev/gfx arrives on the scene it will probably be easy to
hook it to the PP buffers.  I imagine that the SGI /dev/gfx uses something
similar internally, after all - didn't we get the idea from Linas Vepstas
way back when, and IIRC he has worked on IRIX.  Hm, SGI has donated a lot
of IRIX code to Linux recently - why can't they donate their /dev/gfx code
too? 

> Even this will be tough to get pass
> linus but with enough press from other kernel developers which back me it
> would get in. 

	Good luck dude.

Jon

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