[Savage40] New Savage DRM preview

[Felix Kühling]

Am Samstag, den 08.01.2005, 18:59 +0100 schrieb Philippe Villeneuve:
> Hi,
> 
> With the latest instructions I'm finally able to get the DRI running.
> But I've noticed one difference between some info given back.
> 
> bash-2.05b$ dmesg|grep drm
> [drm] Initialized drm 1.0.0 20040925
> [drm] Initialized savage 2.1.0 20050106 on minor 0: S3 Inc. VT8636A
> [ProSavage KN133] AGP4X VGA Controller (TwisterK)
> [drm] Used old pci detect: framebuffer loaded
> 
> and 
> bash-2.05b$ lspci|grep AGP
> 0000:00:01.0 PCI bridge: VIA Technologies, Inc. VT8363/8365 [KT133/KM133 AGP]
> 0000:01:00.0 VGA compatible controller: S3 Inc. VT8636A [ProSavage
> KN133] AGP4X VGA Controller (TwisterK) (rev 01)

AGP4X here is the maximum the chip can do.

> 
> both specify that my twisterK is on a 4x AGP. But 
> 
> bash-2.05b$ glxinfo|grep OpenGL
> OpenGL vendor string: S3 Graphics Inc.
> OpenGL renderer string: Mesa DRI Twister 20050101 AGP 1x x86/MMX+/3DNow!+/SSE
> OpenGL version string: 1.2 Mesa 6.3
> OpenGL extensions:
> 
> mentions a 1x AGP. Should I get any hope that there could be a
> performance gain waiting to be turned on, or is it more of a printf
> problem ;) ?

What glxinfo shows is the AGP speed that is configured in xorg.conf. The
default is 1x. Use the AgpMode option in xorg.conf:

Section "Device"
	...
	Driver		"savage"
	...
        Option          "AgpMode"       "4"
	...
EndSection

See also `man savage`.

My experience with the AgpMode option has always been that it doesn't
have a big effect on performance. Especially for integrated chips (like
the Twister) I wouldn't expect it to make any difference at all. YMMV.

> 
> I wanted also to share a small script I wrote. It covers some
> operations for the compilation of DRI as detailed in the wiki. It is
> my first real script, so I would ask you to be indulgent of my naive
> implementation. Hopefully it will be of use to some of you. I will
> more than welcome any comments on it.

Looks ok. You marked all the critical places with "HACK". A few comments
though.

      * I don't know if the driver works on non-x86 architectures. I've
        never heard of Savages being used on other architectures.
      * I saw some provisions (commented out) for BSD. However, the
        Savage DRM driver currently does not support BSD. Someone with
        BSD and a Savage would have to port it.
      * Recompilation of Xorg can be much faster if you just run "make"
        or "make Everything" instead of "make World". Just "make" will
        break though, if Imakefiles or dependencies changed.
      * Don't blame me if something changes that breaks the scripts. I
        wouldn't entrust the job to a script myself. ;-)

> 
> -philippe

Cheers,
  Felix

-- 
| Felix Kühling <fxkuehl at gmx.de>                     http://fxk.de.vu |
| PGP Fingerprint: 6A3C 9566 5B30 DDED 73C3  B152 151C 5CC1 D888 E595 |