libGL error: Calling driver entry point failed

[duffster]

I'm having some sort of problem getting hardware acceleration.
Can anyone help?

Below is what I think is the necessary information, but obviously I can provide anything that would be required. I am tempted to see if fglrx would work, but I fear I'll just end up spending hours trying to figure out what else I've broken if I give it a try and it doesn't work.

I'm running (uptodate) Debian testing with kernel 2.6.38.8 with a:
01:00.0 VGA compatible controller: ATI Technologies Inc RV710 [Radeon HD 4350]

I suspect my libGL.so.1 is the problem as it looks very outdated, but I can't work out what I need to do to bring it back up to date.

#ls -l /usr/lib/libGL*
-rw-r--r-- 1 root root 405290 Mar 15 07:55 /usr/lib/libGLEW.a
lrwxrwxrwx 1 root root 16 Mar 15 07:55 /usr/lib/libGLEW.so -> libGLEW.so.1.5.8
lrwxrwxrwx 1 root root 16 Mar 15 07:55 /usr/lib/libGLEW.so.1.5 -> libGLEW.so.1.5.8
-rw-r--r-- 1 root root 347376 Mar 15 07:55 /usr/lib/libGLEW.so.1.5.8
lrwxrwxrwx 1 root root 20 Jul 4 15:04 /usr/lib/libGL.so.1 -> libGL.so.1.2
-rw-r--r-- 1 root root 564300 Mar 28 2009 /usr/lib/libGL.so.1.2
-rw-r--r-- 1 root root 394924 Mar 28 2009 /usr/lib/libGL.so.1.old

Here is the appropriate section from the command:
LD_DEBUG=files LIBGL_DEBUG=verbose glxinfo

15872: transferring control: glxinfo
15872:
libGL: XF86DRIGetClientDriverName: 4.3.0 r600 (screen 0)
15872: opening file=/usr/lib/i386-linux-gnu/libGL.so.1 [0]; direct_opencount=1
15872:
libGL: OpenDriver: trying /usr/lib/i386-linux-gnu/dri/tls/r600_dri.so
15872:
15872: file=/usr/lib/i386-linux-gnu/dri/tls/r600_dri.so [0]; needed by /usr/lib/i386-linux-gnu/libGL.so.1 [0]
libGL: OpenDriver: trying /usr/lib/i386-linux-gnu/dri/r600_dri.so
15872:
15872: file=/usr/lib/i386-linux-gnu/dri/r600_dri.so [0]; needed by /usr/lib/i386-linux-gnu/libGL.so.1 [0]
15872: file=/usr/lib/i386-linux-gnu/dri/r600_dri.so [0]; generating link map
15872: dynamic: 0xb71ec028 base: 0xb6e98000 size: 0x003a7604
15872: entry: 0xb6ea94e0 phdr: 0xb6e98034 phnum: 6
15872:
15872:
15872: file=libexpat.so.1 [0]; needed by /usr/lib/i386-linux-gnu/dri/r600_dri.so [0]
15872: file=libexpat.so.1 [0]; generating link map
15872: dynamic: 0xb6e6ee80 base: 0xb6e49000 size: 0x00025fb4
15872: entry: 0xb6e4b160 phdr: 0xb6e49034 phnum: 6
15872:
15872:
15872: calling init: /usr/lib/libexpat.so.1
15872:
15872:
15872: calling init: /usr/lib/i386-linux-gnu/dri/r600_dri.so
15872:
15872: opening file=/usr/lib/i386-linux-gnu/dri/r600_dri.so [0]; direct_opencount=1
15872:
15872:
15872: closing file=/usr/lib/i386-linux-gnu/libGL.so.1; direct_opencount=0
drmOpenDevice: node name is /dev/dri/card0
drmOpenDevice: open result is 4, (OK)
drmOpenByBusid: Searching for BusID pci:0000:01:00.0
drmOpenDevice: node name is /dev/dri/card0
drmOpenDevice: open result is 4, (OK)
drmOpenByBusid: drmOpenMinor returns 4
drmOpenByBusid: Interface 1.4 failed, trying 1.1
drmOpenByBusid: drmGetBusid reports pci:0000:01:00.0
libGL error: Calling driver entry point failedlibGL error: reverting to software direct rendering
15872:
15872: calling fini: /usr/lib/i386-linux-gnu/dri/r600_dri.so [0]

Thanks in advance!

[duffster]

no-one got any suggestions?

I've tried everything I can think of, but I'm getting nowhere.

[whizse]

My first guess would be fallout from the multiarch transition ongoing in Debian at the moment.

(BTW why are you poking around /usr/lib/libGL* when the driver loaded comes from /usr/lib/i386-linux-gnu/ ?)

[whizse]

This is what I was referring to:

http://blog.mraw.org/2011/06/18/mesa_a_disturbance_in_the_Force/

Not sure how much of that is already fixed. It's usually a better idea to run unstable than deal with all these problems.

[duffster]

That certainly looks like it is my problem.

I think the reason I was looking in the wrong place is that I did a partial upgrade (intending to not upgrade any X stuff) and ran into problems pointing at /usr/lib/libGL* and then did a full testing upgrade once I realised something was broken, but clearly I did it just at the wrong moment and have hit the problem you mentioned.

Thanks very much for the pointer - you've saved me a lot of head-scratching.