Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 14 Mar 2000 02:02:05 +0100 (CET)
From:      Marc van Woerkom <van.woerkom@netcologne.de>
To:        reg@FreeBSD.ORG
Cc:        kuku@gilberto.physik.rwth-aachen.de, cokane@one.net, cracauer@cons.org, multimedia@FreeBSD.ORG
Subject:   Re: GLX Xserver for FreeBSD?
Message-ID:  <200003140102.CAA10193@oranje.my.domain>
In-Reply-To: <20000301085111.F54218@shale.csir.co.za> (message from Jeremy Lea on Wed, 1 Mar 2000 08:51:11 -0800)
References:  <200003011009.LAA37864@gil.physik.rwth-aachen.de> <20000301113408.A79072@cons.org> <20000301104653.A11506@evil.2y.net> <20000301165323.A40061@gil.physik.rwth-aachen.de> <20000301085111.F54218@shale.csir.co.za>

next in thread | previous in thread | raw e-mail | index | archive | help
> > I was about to install /usr/ports/graphics/glx
> > when I found in the end that XF336 which I am using under
> > 4.0-current is 'too new' for that version of GLX.

The version check in the port is just a precaution - the port
is very likely to work with 3.3.6 once you change the number
in the makefile.


> I was playing with Will Andrew's update to the Mesa3 port, which
> compiles, and means that you can use later utah-glx code out the box.

The development effort went into a couple of directions,
with code incarnations showing up on a couple of sites
which makes tracking a bit more complicated than usual:

- Mesa is gone through a few bumps (could be around 3.2 now)

- openprojects.net glx project renamed itself into Utah glx
  to distinguish itself from the SGI implementation of the glx
  protocol (which is the glx from XFree86 version 4 and the reason
  why you can`t use a Utah glx module with XF4)

- XFree86 changed from 3.3.x to 4.x branch

- Next to indirect rendering via glx (where every command goes
  through the X protocol) there is work on the direct rendering
  (DRI) going on - it takes place first on the DRI project
  at SourceForge and then gets patched into the XFree86 main
  branch. So far only the Voodoo driver is in, the Matrox one
  is expected soon. Nothing definitive about nvidia yet, although
  the are believed to do some binary only release using that
  new XF4 loader..

It is clear that the XFree86 4.x version and the Direct Rendering
Infrastructure will be the stuff that has the largest potential.

While 4.0 is labeled as a release, my experiences from
Monday and Tuesday (less stable - got a couple of Linux Netscape 4.72 
crashes and a brand new Mozilla died too on similiar pages, plus the 
lack of certain features like the Shape extension used by Windowmaker) 
let XF 4 feel more like a development version than
a stable release. 

Except maybe for Voodoo people (this stuff has been ported by Doug Rabson to 
FreeBSD), it makes not much sense using now for the other cards. 

Thus it worth the effort to polish the existing stuff for
the XFree86 3.3.6 line - which I continue tomorrow. 

 
> managed to update the glx port to a later snapshot, but haven't been
> able to get around to 1. Getting a good (ie stable) tarball on a fixed
> distribution site,

This can be put on the FreeBSD site once we agree on what
Utah glx snapshot to use.


> 2. Verifying that it actually works,

What do you use?
My own cards are nvidia ones, for Matrox I am sure we find some
tester here or in my list of people who helped testing Matrox in the
past. 


> 3. Polishing up the port.

The glx update should be possible this week. For The Mesa update I have 
to look through my mail archives to understand what happened there.


> This is just more a pointer, that later (and faster) GLX code can work
> with XFree86 3.3.6 and Mesa 3.1. 

You are right. 

The speed increase has been reported for Matrox users primarily,
the nvidia source did see only some bug fixes - the reason is 
that Matrox released real specs and nvidia not.

Regards,
Marc





To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-multimedia" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200003140102.CAA10193>