[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-kernel
Subject:    Re: SVGA kernel chipset drivers.
From:       Albert Cahalan <albert () ccs ! neu ! edu>
Date:       1996-05-31 17:22:47
[Download RAW message or body]

From: Ian main <imain@vcc.bc.ca>
> On Thu, 30 May 1996, root wrote:
> 
> > I was wondering if there are any plans incorporate any of the major video 
> > card chipsets into low level kernel driver support. SVGALIB at current 
> > has a very limited number of chipsets that it supports not to mention I 
> > would think a low level kernel driver would get better performence out of 
> > the video hardware. I for one would like to be able to use the features 
> > of my Matrox MGA Millenium in linux :)
> 
> 	Call me crazy..  I had a talk with someone a while back about 
> this.  We thought it'd be a good idea to have a single driver implemented 
> somehow.. one that X could use, and any other app that chose to.  This 
> would have a few advantages.  It would be more stable (ie no wierd text 
> distortion etc. like you get with SVGALIB sometimes) because the one 
> driver could have complete control over the video card. Also... Xfree 
> supports a lot more hardware that svgalib does.  If some of the code for 
> the X drivers could be used in this new implementation, it would be much 
> better than SVGALIB is now.  

A project called GGI is doing exactly that.  Want to help support
your video card or work on the library?  You can join the GGI
mailing list linux-ggi@eskimo.com by sending mail to listserv@eskimo.com
or to the maintainer, irish@eskimo.com.  Note that he will be gone
for a few weeks, so be patient.

> 	Now I am not the one to say where this should be implemented.. 
> whether it be in the kernel, or just a standard library..  I dunno, but 
> it makes a lot of sense to me.  

GGI is using a multi-part approach. All hardware IO and such is done
in a kernel driver.  The kernel provides acceleration features when
the hardware has them, but does _not_ provide an emulation library.
The userspace library uses acceleration if available, otherwise it
just uses a mmapped framebuffer.

> Having a single _smart_ driver for the video hardware would make 
> graphical applications much more attainable, stable, standardized, 
> supported, etc.

Telnet in and 'kill -9' your X server.  Now go to the console and...
Oh, there is no more console. :-(

Together with the SAK patch, GGI fixes this completely.

----

OK now, everyone interested in this _please_ join the GGI mailing list.
This thread has been here before, several times.

Steffen Seeger (misc hacker, maintainer)
   http://www.tu-chemnitz.de/~sse/scrdrv
Andreas Beck (memory, libs)
   http://www-public.rz.uni-duesseldorf.de/~becka/doc/scrdrv.html


[prev in list] [next in list] [prev in thread] [next in thread]