Daniel Schepler wrote:
> Having built and installed fglrx-kernel-2.6.15-1-686 using
> module-assistant, when I try to modprobe fglrx, I get the error:
> 
> FATAL: Error inserting fglrx (/lib/modules/2.6.15-1-686/misc/fglrx.ko): No 
> such device
> 
> dmesg shows:
> 
> fglrx: module license 'Proprietary. (C) 2002 - ATI Technologies, Starnberg, 
> GERMANY' taints kernel.
> [fglrx] Maximum main memory to use for locked dma buffers: 432 MBytes.
> [fglrx:firegl_init] *ERROR* Device not found!

This is likely because radeon.ko is already loaded and is therefore in
control of the card. To test if that's the case, unload it with rmmod
then try loading the fglrx module again. To fix this problem
permanently, you'll have to make sure that radeon.ko doesn't get loaded
automatically during system startup: check the configuration of hotplug,
discover and any other hardware detection tools you are using and
disable automatic loading of modules for display devices. Also check
that you aren't loading it explicitly from /etc/modules.

> This is with a laptop card in a machine manufactured by Fujitsu-Siemens.
> The lspci -v output for it is:
> 
> 0000:01:00.0 VGA compatible controller: ATI Technologies Inc RV350 [Mobility 
> Radeon 9600 M10] (prog-if 00 [VGA])
> [...]
> 
> And lspci -n is:
> 
> 0000:01:00.0 0300: 1002:4e52

This is a known problem and it can be fixed by overriding the chip ID in
the X config file's Device section as explained here:

http://www.stanchina.net/~flavio/debian/fglrx-installer.html#chipid

In short, try placing
  ChipID 0x4E50

just below
  Driver "fglrx"

in the X config file.

I will close this bug when ATI finally adds support for chip IDs 4e51
and 4e52. Unfortunately I can't fix it myself, as it is in the
binary-only part of the driver.

> (I can't really try the free r300 driver, either, because of
> xserver-xorg bug #345929 -- and with that too, I have to add my card
> manually to the hardware list for radeon.ko to recognize it.  It seems
> nobody wants to admit my card actually exists.  :( )

That's the real problem. I wonder why ATI doesn't add that chip ID to
the list of supported cards, as it is known to work just fine with the
chip ID override.

-- 
Ciao, Flavio


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Reply via email to