DRI and the radeon.o kernel module -- Joy

[Stefan Bruda]

Hi again.

At 01:50 -0400 on 2003-4-11 Bill Fink wrote:
 >
 > On Thu Apr 10 2003, Stefan Bruda wrote:
 > 
 > > I have recently installed XFree86 4.3.0 on my Tibook (the one with
 > > Radeon M7 LW (AGP) in it).  The installation was, by the way,
 > > painless, maybe except the rather large build time.  Anyway, the
 > > problem is now that DRI refuses to start
 > 
 > I'll tell you what I did to get DRI working with my ATI Rage 128 card
 > (I suspect the Radeon should be similar).

Joy, well, sort of: DRI is now up and running, the FPS jumped about
three times, yay.  Thanks for the solution.

However, there is a problem in the sense that I am no longer use AGP.
The same /var/log/XFree86.0.log now says:

(II) RADEON(0): [drm] created "radeon" driver at busid "PCI:0:16:0"
(II) RADEON(0): [drm] added 8192 byte SAREA at 0xe3013000
(II) RADEON(0): [drm] mapped SAREA 0xe3013000 to 0x30019000
(II) RADEON(0): [drm] framebuffer handle = 0xb8000000
(II) RADEON(0): [drm] added 1 reserved context for kernel
(II) RADEON(0): [agp] Mode 0x07000207 [AGP 0x106b/0x002d; Card 0x1002/0x4c57]
(II) RADEON(0): [agp] 8192 kB allocated with handle 0xe3016000
(II) RADEON(0): [agp] ring handle = 0x00000000
(EE) RADEON(0): [agp] Could not map ring
(WW) RADEON(0): [agp] AGP failed to initialize -- falling back to PCI mode.
(WW) RADEON(0): [agp] If this is an AGP card, you may want to make sure the agpgart
kernel module is loaded before the radeon kernel module.

and then goes on and does the right thing, but on PCI instead of AGP.
I doubt the solution is the one sugggested above, both agpgart and
radeon are compiled into the kernel instead of being modules.  I.e., I
have into my kernel config

CONFIG_AGP=y
CONFIG_AGP_UNINORTH=y
CONFIG_DRM_NEW=y
CONFIG_DRM=y
CONFIG_DRM_RADEON=y

I also tried CONFIG_AGP=m and CONFIG_DRM_RADEON=m and loaded them
explicitly before starting X--just for the heck of it, I did not
expect (and did not have) any success.  Am I missing anything here?
The relevant part of XF86Config-4 looks like this:

Section "Device"
	Identifier "Radeon"
	Driver "radeon"
	BusID "PCI:0:16:0"

        Option "UseFBDev"
        Option "AGPMode" "4"
EndSection

Does the order in which modules are loaded (in Section "Module" of
XF86Config-4) matter?

Thanks again,
Stefan

-- 
If it was so, it might be; and if it were so, it would be; but as
it isn't, it ain't. That's logic.
    --Lewis Carroll, Through the Looking-Glass