Bug 338930

glxinf and fglrxinfo both correctly see the card:

15:55 rankin-p35a~/Pictures/screenshots> fglrxinfo
display: :0.0  screen: 0
OpenGL vendor string: ATI Technologies Inc.
OpenGL renderer string: ATI MOBILITY RADEON 9600/9700 Series
OpenGL version string: 2.0.6958 Release

The biggest problem is when sax2 selects the "ATI Rv350 NP" card it says there is no 3D. When you attempt to change any sax2 setting, sax2 provides a dialog that says 3D is not available, which is not the case. See the error:

http://www.3111skyline.com/download/screenshot/sax2ati3d_error.png

The 9600/9700 card does 3D just fine, but I suspect that the initial sax2 choice of the wrong card without 3D is causing many additional problems.

if you don't mind call

   init 3
   sax2

   ==> click on change configuration
   ==> could you select 3D y/n
   ==> exit sax2

send the file /var/log/SaX.log

thanks

*** Bug 338953 has been marked as a duplicate of this bug. ***

This card is registered as 3D capable with radeon driver in our database.

Created attachment 182105 [details]
Requested SaX.log

You can *not* select 3D in sax2. The selection is greyed out. Currently sax2 hardlocks the machine when started from runlevel 3 so the SaX.log may be a partial. This occurs if the console is reached by init 3 from kwin or by passing "runlevel 3" as a grub boot parameter. I can start sax from yast in kwin.

Here is something bizarre.

If I delete the link: /usr/lib/libIndirectGL.so.1 by moving it,

[root rankin-p35a/home/david] # l /usr/lib/libInd*
-rwxr-xr-x 1 root root 440676 2007-09-21 20:34 /usr/lib/libIndirectGL.so.1.2*
lrwxrwxrwx 1 root root     20 2007-11-05 09:35 /usr/lib/libIndirectGL.so.1.sav -> libIndirectGL.so.1.2*

Then, I can start sax2 from within kwin or compiz and the card is then shown correctly. See:

http://www.rbpllc.com/download/screenshots/sax2/sax2-ATI9600-9700.jpg

However, when the /usr/lib/libIndirectGL.so.1 link is there, sax2 freezes from the console and I believe also causes the card to be configured as the ATI Rv350 NP.

Way beyond me!

I suggest to uninstall the fglrx driver.

Uninstalling the fglrx driver will leave the laptop very crippled. But we may be on to something. Here is what is currently installed:

[root rankin-p35a/home/david] # rpm -qa | grep fglrx
ati-fglrxG01-kmp-default-8.42.3_2.6.22.9_0.4-1.1
x11-video-fglrxG01-8.42.3-2.1

it appears the x11-video rpm is not matched to the ati driver after 8.42.3-2.xx. Any way to fix this?

You installed the fglrx driver but the RPMs are still installed ?

No? Below is exactly how I built and installed the driver, all fglrx rpms were removed. With 10.3, this laptop has always had problems with graphics for some reason. 10.2 worked fine. The only other clue I could add was that during the original install of 10.3, the screen froze, and the install bombed, but Yast was able to pick back up almost at the point where the screen froze. (rescue was borked) So I just had to continue the install the best I could. Is there anything related to the initial 10.3 install that could still be causing this problem?

Would I get any benefit from booting from the new kde-LIVE cd and trying rescue from there?  Anyway, here is exactly how I installed the fglrx driver:

------------------------------------------
Building and Installing the New ATI Driver
------------------------------------------

Prepare the kernel

   # su
   password
   # cd /usr/src/linux
   # make mrproper
   # make cloneconfig
   # make modules_prepare
   # make clean

Remove old drivers

   # rpm -e $(rpm -qa | grep fglrx)

Generate new driver

   # sh ati-driver-installer-8.42.3-x86.x86_64.run --listpkg
   # sh ati-driver-installer-8.42.3-x86.x86_64.run --buildpkg
SuSE/SUSE103-IA32

Check that the new driver rpms were successfully created in
/usr/src/packages/RPMS/i386

   # l /usr/src/packages/RPMS/i386/
	-rw-r--r-- 1 root root 11073729 2007-10-22 19:45 fglrx_7_1_0_SUSE102-8.40.4-1.i386.rpm
	-rw-r--r-- 1 root root 15365859 2007-10-24 20:31
fglrx_7_1_0_SUSE103-8.42.3-1.i386.rpm

Go to Runlevel 3

   # init 3

Install the new rpms created in /usr/src/packages/RPMS/i386

   # rpm -Uvh /usr/src/packages/RPMS/i386/fglrx*.rpm

After the install is finished, you need to launch the aticonfig utility

   # ldconfig
   # aticonfig --initial --input=/etc/X11/xorg.conf

Configure the new driver with sax2

   # sax2 -r -m 0=fglrx

Check other important Settings in the xorg.conf

	For best and most stable working fglrx driver set the following options
to Section "Device" in the xorg.conf:

   Option      "BlockSignalsOnLock" "on"
   Option      "KernelModuleParm" "locked-userpages=0"
   Option      "OpenGLOverlay" "off"
   Option      "UseFastTLS" "2"
   Option      "UseInternalAGPGART" "no"
   Option      "VideoOverlay" "on"
   Option      "mtrr" "off"
   Option      "no_accel" "no"
   Option      "no_dri" "no"
   Option      "EnablePrivateBackZ" "no"
   Option      "backingstore" "true"
   Option      "XAANoOffscreenPixmaps" "true"

        Add the addition xorg.conf settings:

Section "ServerFlags"
  Option       "AIGLX" "True"
EndSection

Section "Extensions"
  Option       "DAMAGE" "true"
  Option "Composite" "true"
EndSection


All OK - reboot
-----------------------------------------

You can't enable the 3D button in SaX2 if you don't use the official 8.42.3 RPMs for openSUSE 10.3, since only these RPMs are prepared for SaX2. Please read

  http://www.suse.de/~sndirsch/ati-installer-HOWTO.html

for installation instructions via YaST.

OK, next test and strange result. I downloaded the new kde-LIVE cd and booted from it. Sax2 configured the card again incorrectly as "ATI Rv350 NP". This time however I *could* select 3D on that card. However it was still the incorrect card.

The Live cd also allowed me to "log out" without the system hanging in hardlock. However attempting to *reboot* hardlocks the machine at the end of the shutdown cycle. I could *shutdown* cleanly though.

Is there anyway I can use the Live cd to just reinstall the base graphics system on my laptop without losing my laptop install? I want to avoid reformatting at all costs. I don't have another 20hrs. to redo all of the configuration, etc.

Would a copy of the SaX.log from the cd boot be of any help? If so, I'll be glad to send one (after I load the madwifi drivers on the live cd)

Rv350 NP is your graphics chip, not the card. This chip is marked as 3D capable when using the radeon driver. Do not enable 3D support if it crashes your machine.

Guys,

    It's NOT the driver. The same hardlock occurs with a reboot after booting from the openSUSE-10.3-GM-KDE-Live-i386.iso. After openSuSE caught up ati driver release and had the 8.42.3.2 driver out, I did then install them from the www2.ati.com repository.

    What is this "WONT FIX" garbage. Look at the fglrx driver info, that is the openSuSE release. So what? Are we just supposed to live with a box that you cannot log out from or reboot from without hardlocking the system (with both the new kde_LIVE cd and the ati repository drivers). None of these issues were present in 10.2.

    Do you even know what it is you "WONT FIX"?  Do we have a handle on it? If so, would you mind telling me what it is? This laptop worked fine with 10.2, now in 10.3 you can't log out or reboot? What gives? What else do I need to check to make sure the 10.3 install freeze without a functioning rescue hasn't screwed up my box?

    I'm sorry, but "WONT FIX" sure smells like a "COP OUT" on this one. There is a bug somewhere and I'm doing my best to help you try to fix it. Tell me what it is that you have made a conscious decision to "NOT FIX" and at least then I'll know what it is, maybe I can continue to try and fix it! Thanks.

The WONTFIX was about the disabled 3D button in SaX2 when the fglrx driver has been installed manually and SaX2 has been started with "sax2 -r -m 0=fglrx". 

I suggest to not enable 3D support using the radeonold driver and test again if this prevents the freeze at the logout. Run "sax2 -r -m 0=radeonold" for configuration. This would be a beginning.

Could you try this and check if the freeze at logout and reboot no longer happens?

Ok, sorry for getting us sidetracked. I appologize, I think that you are correct that part (but not all) of the problem was with driver. Here is what I have done. 

Troubleshooting steps performed:

(1) init 3
(2) rpm -e $(rpm -qa | grep fglrx)
(3)  # su
   password
   # cd /usr/src/linux
   # make mrproper
   # make cloneconfig
   # make modules_prepare
   # make clean

(4) sax2  (configure w/o 3D and default radeon driver)

See: http://www.rbpllc.com/download/screenshots/sax2/sax2-No3D.jpg

(5) reboot

I can now log out without the screen freezing and then log back in. I have not yet tried to reboot, but I will and I will report back with a separate comment. 

However, other parts of the graphics system are now broken. glxgears no longer works:

15:25 rankin-p35a~> glxgears
glxgears: error while loading shared libraries: libGL.so.1: cannot open shared object file: No such file or directory

Hmm.., where did libGL.so.1 go?

[root rankin-p35a/home/david] # find /usr/lib -name libGL*
/usr/lib/libGLU.so.1
/usr/lib/libGLU.so.1.3.070001
/usr/lib/libGLcore.so.1
/usr/lib/libGLcore.so.1.0
/usr/lib/libGL.so
/usr/lib/xorg/modules/extensions/libGLcore.so
/usr/lib/libGLU.so

Is there a needed link that has been trashed? Also, libGL.so is now a broken link:

[root rankin-p35a/home/david] # l /usr/lib/libGL.so
lrwxrwxrwx 1 root root 10 2007-10-12 03:00 /usr/lib/libGL.so -> libGL.so.1

Let me know what else you want me to try. If it would help and If you want access to this laptop, I can set my router to make that so. Just let me know.

I really do appreciate you taking the time to reopen this issue. I will do everything I can to help resolve this.

Thanks!

Please update the Mesa package to install libGl.so.1 again. I suspect it vanished by installing/uninstalling the fglrx driver manually. But that's just a guess. I will wait patiently for the feedback if the non-3D radeon driver now also fix the reboot freeze.

Ok,

The system still freezes on reboot. It gets to the point where it unloads the file systems, the flashes the "Rebooting, Please Stand By ..." or whatever it says there, and then it hardlocks. Thankfully the freeze doesn't occur until after the file systems are unmounted, but the reboot problem is still there. 

Is there anything I can send? The logs are silent as far as I can tell, but I will send whatever you would like.

Thanks.

Here is a crude performance comparison between the ATI drivers and the stock radeon driver:

            Stock Radeon        ATI Driver
glxgears:   308 FPS             2700 FPS

Is there any way to get more performance out of the stock drivers?

Could you please open a seperate bugreport for the reboot freeze issue. For me this looks completely unrelated to the X configuration.

Back to 3D performance. Could you just add the line ' Load "dri" ' to Section "Module" of your /etc/X11/xorg.conf and restart your session?
Better fps values? Still no freezes during logout? 

Done,

Section "Module"
  Load         "dbe"
  Load         "type1"
  Load         "freetype"
  Load         "extmod"
  Load         "glx"
  Load         "dri"
EndSection

    dri makes no difference with performance. The same 308 FPS are received after restarting the server. However, I can still log in and log out without a freeze as long as I am using the radeon driver. Graphics performance sucks, but you can log in and out just fine.

    The ati driver problem looks more and more like a memory leak or a piece of the driver that gets left in memory when an X session in terminated. To answer your question, I dropped to runlevel 3, reconfigured with the original radeon driver, loaded dri and was able to log in and log out fine. The bizarre part is that after I reconfigured to use the ati driver in runlevel 3 and went to start x (startx), the system immediately hardlocked displaying the same symptoms that originally started this Bug.

    There shouldn't be anything left in memory after stopping the x server with the /opt/kde3/share/config/kdm/kdmrc config set with:

# Core config for 1st local display
[X-:0-Core]
TerminateServer=true

    Shutting down the system and clearing the memory immediately allowed me to go back into x with the ati driver and start compiz without any problem. Of course if I drop to runlevel 3 and then try and go back in, the system will hang again -- every time.

    I had installed the ati driver prior to conducting the test you wanted. Before installing the ati driver, I saved the xorg.conf from the stock radeon driver for that purpose. To configure for the test, I simply created a couple of scripts that would configure the system to the original state and then another script to set the system back in the state to use the ati driver with compiz.

    To set the system to its original state I used:

#!/bin/bash
#
# Unlink and Restore libGL.so.1.2
#
unlink /usr/lib/libGL.so.1.2
mv /usr/lib/libGL.so.1.2.sav /usr/lib/libGL.so.1.2
#
# Restore /usr/lib/libIndirectGL.so.1
#
mv /usr/lib/libIndirectGL.so.1.sav /usr/lib/libIndirectGL.so.1
#
# Set the proper xorg.conf
#
cp /home/david/linux/ati-compiz/xorg.conf.radeon /etc/X11/xorg.conf


    To configure the system for the ati driver I do:

#!/bin/bash
#
# Move Old libGL.so.1.2 and Link New Lib
#
mv /usr/lib/libGL.so.1.2 /usr/lib/libGL.so.1.2.sav
ln -s /usr/X11R6/lib/libGL.so.1.2 /usr/lib/libGL.so.1.2
#
# Remove /usr/lib/libIndirectGL.so.1
#
mv /usr/lib/libIndirectGL.so.1 /usr/lib/libIndirectGL.so.1.sav
#
# Set the proper xorg.conf
#
cp /home/david/linux/ati-compiz/xorg.conf.compiz /etc/X11/xorg.conf

    Crazy? Yes, but this is what it takes.

    I have also done additional documenting of the library problems. See Bug 338947, for more details on the libraries that must be changed to get compiz to work with this system. Hopefully some of this additional information will pin-point exactly what the problem is and why the libraries and drivers don't like this Toshiba laptop...

Do you guys have any way of getting AMD/ATI in the loop on this problem? Should I also post some of this on their site? Do they have anyone that looks at linux issues seriously that might be of help to get involved?

I would like to concentrate on radeon 3D for now. No need to involve ATI here.
Please attach /var/log/Xorg.0.log with DRI enabled. Also attach the output of
'ldd /usr/bin/glxinfo | grep GL.so' and 'LIBGL_DEBUG=1 glxinfo'.

[root Rankin-P35a/home/david] # ldd /usr/bin/glxinfo | grep GL.so
        libGL.so.1 => /usr/lib/libGL.so.1 (0xb7e40000)

[root Rankin-P35a/home/david] # LIBGL_DEBUG=1 glxinfo
name of display: :0.0
display: :0  screen: 0
direct rendering: No
server glx vendor string: SGI
server glx version string: 1.2
server glx extensions:
    GLX_ARB_multisample, GLX_EXT_visual_info, GLX_EXT_visual_rating,
    GLX_EXT_import_context, GLX_EXT_texture_from_pixmap, GLX_OML_swap_method,
    GLX_SGI_make_current_read, GLX_SGIS_multisample, GLX_SGIX_hyperpipe,
    GLX_SGIX_swap_barrier, GLX_SGIX_fbconfig, GLX_MESA_copy_sub_buffer
client glx vendor string: SGI
client glx version string: 1.4
client glx extensions:
    GLX_ARB_get_proc_address, GLX_ARB_multisample, GLX_EXT_import_context,
    GLX_EXT_visual_info, GLX_EXT_visual_rating, GLX_MESA_allocate_memory,
    GLX_MESA_copy_sub_buffer, GLX_MESA_swap_control,
    GLX_MESA_swap_frame_usage, GLX_OML_swap_method, GLX_OML_sync_control,
    GLX_SGI_make_current_read, GLX_SGI_swap_control, GLX_SGI_video_sync,
    GLX_SGIS_multisample, GLX_SGIX_fbconfig, GLX_SGIX_pbuffer,
    GLX_SGIX_visual_select_group, GLX_EXT_texture_from_pixmap
GLX extensions:
    GLX_ARB_get_proc_address, GLX_ARB_multisample, GLX_EXT_import_context,
    GLX_EXT_visual_info, GLX_EXT_visual_rating, GLX_MESA_copy_sub_buffer,
    GLX_OML_swap_method, GLX_SGI_make_current_read, GLX_SGIS_multisample,
    GLX_SGIX_fbconfig, GLX_EXT_texture_from_pixmap
OpenGL vendor string: Mesa project: www.mesa3d.org
OpenGL renderer string: Mesa GLX Indirect
OpenGL version string: 1.4 (2.1 Mesa 7.0.1)
OpenGL extensions:
    GL_ARB_depth_texture, GL_ARB_draw_buffers, GL_ARB_fragment_program,
    GL_ARB_imaging, GL_ARB_multisample, GL_ARB_multitexture,
    GL_ARB_occlusion_query, GL_ARB_point_parameters, GL_ARB_point_sprite,
    GL_ARB_shadow, GL_ARB_shadow_ambient, GL_ARB_texture_border_clamp,
    GL_ARB_texture_compression, GL_ARB_texture_cube_map,
    GL_ARB_texture_env_add, GL_ARB_texture_env_combine,
    GL_ARB_texture_env_crossbar, GL_ARB_texture_env_dot3,
    GL_ARB_texture_mirrored_repeat, GL_ARB_texture_non_power_of_two,
    GL_ARB_texture_rectangle, GL_ARB_transpose_matrix, GL_ARB_vertex_program,
    GL_ARB_window_pos, GL_EXT_abgr, GL_EXT_bgra, GL_EXT_blend_color,
    GL_EXT_blend_equation_separate, GL_EXT_blend_func_separate,
    GL_EXT_blend_logic_op, GL_EXT_blend_minmax, GL_EXT_blend_subtract,
    GL_EXT_clip_volume_hint, GL_EXT_copy_texture, GL_EXT_draw_range_elements,
    GL_EXT_fog_coord, GL_EXT_framebuffer_object, GL_EXT_multi_draw_arrays,
    GL_EXT_packed_pixels, GL_EXT_paletted_texture, GL_EXT_point_parameters,
    GL_EXT_polygon_offset, GL_EXT_rescale_normal, GL_EXT_secondary_color,
    GL_EXT_separate_specular_color, GL_EXT_shadow_funcs,
    GL_EXT_shared_texture_palette, GL_EXT_stencil_wrap, GL_EXT_subtexture,
    GL_EXT_texture, GL_EXT_texture3D, GL_EXT_texture_edge_clamp,
    GL_EXT_texture_env_add, GL_EXT_texture_env_combine,
    GL_EXT_texture_env_dot3, GL_EXT_texture_lod_bias,
    GL_EXT_texture_mirror_clamp, GL_EXT_texture_object,
    GL_EXT_texture_rectangle, GL_EXT_vertex_array, GL_APPLE_packed_pixels,
    GL_ATI_draw_buffers, GL_ATI_texture_env_combine3,
    GL_ATI_texture_mirror_once, GL_ATIX_texture_env_combine3,
    GL_IBM_texture_mirrored_repeat, GL_INGR_blend_func_separate,
    GL_MESA_pack_invert, GL_MESA_ycbcr_texture, GL_NV_blend_square,
    GL_NV_fragment_program, GL_NV_light_max_exponent, GL_NV_point_sprite,
    GL_NV_texgen_reflection, GL_NV_texture_rectangle, GL_NV_vertex_program,
    GL_NV_vertex_program1_1, GL_SGI_color_matrix, GL_SGI_color_table,
    GL_SGIS_generate_mipmap, GL_SGIS_texture_border_clamp,
    GL_SGIS_texture_edge_clamp, GL_SGIS_texture_lod, GL_SGIX_depth_texture,
    GL_SGIX_shadow, GL_SGIX_shadow_ambient, GL_SUN_multi_draw_arrays
glu version: 1.3
glu extensions:
    GLU_EXT_nurbs_tessellator, GLU_EXT_object_space_tess

   visual  x  bf lv rg d st colorbuffer ax dp st accumbuffer  ms  cav
 id dep cl sp sz l  ci b ro  r  g  b  a bf th cl  r  g  b  a ns b eat
----------------------------------------------------------------------
0x23 24 tc  0 24  0 r  y  .  8  8  8  0  0 16  0  0  0  0  0  0 0 None
0x24 24 tc  0 24  0 r  y  .  8  8  8  0  0 16  8 16 16 16  0  0 0 None
0x25 24 tc  0 32  0 r  y  .  8  8  8  8  0 16  8 16 16 16 16  0 0 None
0x26 24 tc  0 32  0 r  .  .  8  8  8  8  0 16  8 16 16 16 16  0 0 None
0x27 24 dc  0 24  0 r  y  .  8  8  8  0  0 16  0  0  0  0  0  0 0 None
0x28 24 dc  0 24  0 r  y  .  8  8  8  0  0 16  8 16 16 16  0  0 0 None
0x29 24 dc  0 32  0 r  y  .  8  8  8  8  0 16  8 16 16 16 16  0 0 None
0x2a 24 dc  0 32  0 r  .  .  8  8  8  8  0 16  8 16 16 16 16  0 0 None

> [root Rankin-P35a/home/david] # ldd /usr/bin/glxinfo | grep GL.so
>         libGL.so.1 => /usr/lib/libGL.so.1 (0xb7e40000)
Looks good. Seems to be the Mesa and not the ATI one.

> [root Rankin-P35a/home/david] # LIBGL_DEBUG=1 glxinfo
> name of display: :0.0
> display: :0  screen: 0
> direct rendering: No
Bad. I would expect more debug here. Could you try again by running

  LIBGL_DEBUG=1 glxinfo 2>&1| grep -B100 "^direct rendering"

/var/log/Xorg.0.log is still missing.

> /var/log/Xorg.0.log is still missing.
Just found this information in Bug #338947, comment #13.

(EE) RADEON(0): [dri] RADEONDRIGetVersion failed to open the DRM
[dri] Disabling DRI.
[...]
(WW) RADEON(0): Direct rendering disabled

Could you run "rmmod radeon; modprobe radeon" and add the lines in /var/log/Xorg.0.log, which appear there at the time you run these commands?

I think I know what's the issue here. The fglrx kernel module is still loaded. Then the radeon kernel module cannot be loaded. Try to unload the fglrx module with "rmmod fglrx". If this doesn't work, you need to reboot.

BTW, this is what happens if you try to investigate two different exclusive drivers at the same time. It always ends up in a mess and makes things more difficult for the reporter and of course the poor guy, who needs to investigate the issue and has no access to the system.

Sorry, looks like the Xorg.0.log got posted to Bug 338947#c13. If you need another copy, just let me know. Here is the additional output:

[root Rankin-P35a/home/david/linux/ati-compiz] # rmmod radeon; modprobe radeon
ERROR: Module radeon does not exist in /proc/modules
WARNING: Error inserting drm (/lib/modules/2.6.22.9-0.4-default/kernel/drivers/char/drm/drm.ko): Cannot allocate memory
FATAL: Error inserting radeon (/lib/modules/2.6.22.9-0.4-default/kernel/drivers/char/drm/radeon.ko): Unknown symbol in module, or unknown parameter (see dmesg)


    There was no output to Xorg.0.log: 

[root Rankin-P35a/home/david] # tailf /var/log/Xorg.0.log
SynapticsCtrl called.
(II) Mouse[1]: ps2EnableDataReporting: succeeded
Synaptics DeviceOn called
(--) Mouse[3] auto-dev sets device to /dev/input/event3 (waited 0 msec)
(**) Option "Device" "/dev/input/event3"
(--) Mouse[3] touchpad found
(II) 3rd Button detected: disabling emulate3Button
SetClientVersion: 0 9
SetGrabKeysState - disabled
SetGrabKeysState - enabled

Damn your good,

[root Rankin-P35a/home/david/linux/ati-compiz] # modprobe -l | grep fglrx
/lib/modules/2.6.22.9-0.4-default/extra/fglrx.ko
[root Rankin-P35a/home/david/linux/ati-compiz] # rmmod fglrx

**now lets double-check to see if it was removed

[root Rankin-P35a/home/david/linux/ati-compiz] # modprobe -l | grep fglrx
/lib/modules/2.6.22.9-0.4-default/extra/fglrx.ko

(Rebooting) I guess fglrx is really stuck in memory. SHouldn't all this be wiped from memory with TerminateServer=true in kdmrc?

Thanks for your help! 

> (Rebooting) I guess fglrx is really stuck in memory. SHouldn't all this be
> wiped from memory with TerminateServer=true in kdmrc?
Not really. kernel modules are not unloaded, even if they are no longer in use.

Please reboot your machine and try 3D enabled radeon driver again.

Ok, progress:

    I blacklisted fglrx to make sure it wasn't giving us any trouble. Confirmed the right modules are loaded:

[root rankin-p35a/home/david] # lsmod | grep radeon
radeon                116228  1
drm                    79636  2 radeon
radeonfb              101168  0
fb_ddc                  6656  1 radeonfb
i2c_algo_bit            9988  1 radeonfb
i2c_core               27520  4 radeonfb,fb_ddc,i2c_algo_bit,i2c_piix4

[root rankin-p35a/home/david] # lsmod | grep fglrx
[root rankin-p35a/home/david] #

    When I checked with Sax2, 3D was already enabled, see:

http://www.rbpllc.com/download/screenshots/sax2/sax3-3D.jpg

    Framerates are better:

[root rankin-p35a/home/david] # glxgears
9085 frames in 5.0 seconds = 1816.883 FPS
9099 frames in 5.0 seconds = 1819.645 FPS
9156 frames in 5.0 seconds = 1831.155 FPS
9133 frames in 5.0 seconds = 1826.408 FPS

    Your other requests look different:

[root rankin-p35a/home/david] # ldd /usr/bin/glxinfo | grep GL.so
        libGL.so.1 => /usr/lib/libGL.so.1 (0xb7e49000)

    Here is a new one:

[root rankin-p35a/home/david] # LIBGL_DEBUG=1 glxinfo 2>&1| grep -B100 "^direct rendering"
libGL error: dlopen /usr/lib/dri/updates/r300_dri.so failed (/usr/lib/dri/updates/r300_dri.so: cannot open shared object file: No such file or directory)
libGL error:
Can't open configuration file /root/.drirc: No such file or directory.
name of display: :0.0
display: :0  screen: 0
direct rendering: Yes

(same results as user)

Xorg.0.log will follow as an attachment (posted to the right bug)
    

Created attachment 182499 [details]
Xorg log with 3D enabled, better frame rates

Great. So you finally managed it to enable radeon 3D support. Still freezes during logout with this setup?

No Joy.

    The system hangs miserably on log out with the radeon 3D driver enabled. Here is what occurred:

(1) /etc/inittab was set to default to runlevel 3. So to conduct your test, I used log out to drop to runlevel 3. (a quick vi of /etc/inittab to reset default runlevel to 5 to prepare for test 2 below, in case of a lockup) Then an init 5 to restart kdm. The screen hung with the black screen of death. The mouse pointer was visible and would still move. ctrl+alt+backspace did not work. However (this is a first), pressing the power button did gracefully shut the system down;

(2) reboot with default runlevel 5. Log in with login screen without problems. Confirmed 3D active with glxgears and confirmed that proper radeon driver was loaded and no fglrx with lsmod. All OK. (see below). Same exact freeze on log out as in (1) above. Black screen of death, mouse moves, can't kill X, powerbutton gracefully shut down. --- UUGHH....

    What else can I send?

----------- Confirmation Details -------------

[root rankin-p35a/home/david] # glxgears
9048 frames in 5.0 seconds = 1809.596 FPS
9018 frames in 5.0 seconds = 1803.434 FPS
9086 frames in 5.0 seconds = 1817.050 FPS

[root rankin-p35a/home/david] # lsmod | grep radeon
radeon                116228  1
drm                    79636  2 radeon
radeonfb              101168  0
fb_ddc                  6656  1 radeonfb
i2c_algo_bit            9988  1 radeonfb
i2c_core               27520  4 radeonfb,fb_ddc,i2c_algo_bit,i2c_piix4
[root rankin-p35a/home/david] # lsmod | grep fglrx

Ok. At least the problem is understood now. Unfortunately we don't have the ressources to fix the issues of the reverse engineered r300 DRI driver, so I need to set this to LATER (in the hope that it will be fixed in a later
X.Org/Mesa release).

For the fglrx freezes during logout please open a seperate bugreport, which I then can reassign to ATI. Please also try 8.40 release of this driver, which is known to be more stable that 8.41/8.42 for some graphics cards.

It is possible that you need to do without 3D for now.

OK,

   Thanks for following this to the bitter end. Now all of our effort has not been in vain because we now have specifically identified and diagnosed the problem and it can be confirmed and set to be fixed for a future release. And we now know that all people who have ati Mobility Radeon cards in their laptops are just stuck with the bug until X.org/Mesa or ATI can fix a problem with the radeon driver or, open source it and have it fixed for them in less that 24 hours.

    I will open a fglrx bug so it can go to ati, but I can tell you that much investigation has been done there and:

(1) the freeze occurs with the 8.40, 8.41 and 8.42 drivers;
(2) the 8.41 driver is so borked you can't even get compiz to run to test with it; and 
(3) the XGL/AIGLX change that occurred between the 8.41 and 8.42 drivers makes testing with anything prior to the 8.42 driver somewhat wasted effort.

    But we'll sort through that in a new story.

    Thanks Again!