[fglrx] DRI not working with fglrx

Bug #199359 reported by Милош
12
Affects Status Importance Assigned to Milestone
linux-restricted-modules-2.6.24 (Ubuntu)
Invalid
Medium
Unassigned

Bug Description

Binary package hint: xorg-driver-fglrx

I'm running Hardy Alpha 5 with the latest updates and I cannot get DRI working with fglrx. I tried using the official drivers through Restricted manager, also installed envyng, with the latest ATI drivers (8.3) with no luck.

misos@misos-laptop:~$ glxinfo
name of display: :1.0
display: :1 screen: 0
direct rendering: No (If you want to find out why, try setting LIBGL_DEBUG=verbose)
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
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_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 version: 1.2
GLX extensions:
    GLX_ARB_get_proc_address, GLX_ARB_multisample, GLX_EXT_import_context,
    GLX_EXT_visual_info, GLX_EXT_visual_rating, 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: 1.2 (2.1 Mesa 7.0.3)
OpenGL extensions:
    GL_ARB_depth_texture, GL_ARB_fragment_program, GL_ARB_imaging,
    GL_ARB_multitexture, GL_ARB_point_parameters, GL_ARB_point_sprite,
    GL_ARB_shadow, GL_ARB_shadow_ambient, GL_ARB_texture_border_clamp,
    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_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_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_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_object, GL_EXT_texture_rectangle,
    GL_EXT_vertex_array, GL_APPLE_packed_pixels, 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_point_sprite, GL_NV_texgen_reflection, GL_NV_texture_rectangle,
    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

   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
----------------------------------------------------------------------
0x2c 24 tc 0 32 0 r y . 8 8 8 8 0 24 8 0 0 0 0 0 0 None
0x2d 24 tc 0 32 0 r . . 8 8 8 8 0 0 0 0 0 0 0 0 0 None
0x2e 32 tc 0 32 0 r y . 8 8 8 8 0 24 8 0 0 0 0 0 0 Ncon
0x2f 32 tc 0 32 0 r . . 8 8 8 8 0 0 0 0 0 0 0 0 0 Ncon
misos@misos-laptop:~$ fglrxinfo
display: :1.0 screen: 0
OpenGL vendor string: Mesa project: www.mesa3d.org
OpenGL renderer string: Mesa GLX Indirect
OpenGL version string: 1.2 (2.1 Mesa 7.0.3)

Revision history for this message
Милош (misos) wrote :
Revision history for this message
Милош (misos) wrote :
Revision history for this message
Timo Aaltonen (tjaalton) wrote :

Try commenting out the Extensions section.

Changed in linux-restricted-modules-2.6.24:
status: New → Incomplete
Revision history for this message
Милош (misos) wrote :

Just tried it, no change.

Милош (misos)
Changed in linux-restricted-modules-2.6.24:
status: Incomplete → New
Revision history for this message
Saivann Carignan (oxmosys) wrote :

Thanks for your bug report. Can you try with the latest hardy alpha 6 release? In my case, it started to work with hardy alpha 6 once the bug in jockey got fixed.

Changed in linux-restricted-modules-2.6.24:
assignee: nobody → saivann
status: New → Incomplete
Revision history for this message
Милош (misos) wrote :

Now X does not start at all, I had to restore the original VESA configuration.

Revision history for this message
Saivann Carignan (oxmosys) wrote :

Did you do a fresh install of the hardy alpha 6 release and installed restricted drivers before X stops to work?

Revision history for this message
Милош (misos) wrote :

No, I upgraded the system, removed restricted drivers, rebooted, installed them again, and rebooted again. Actualy, X fails to load even if I used the open source radeonhd drivers.

I think I forgot to say, I have ATI Radeon X1400 Mobility.

Revision history for this message
Saivann Carignan (oxmosys) wrote :

I believe that since this bug relates to X config files, you should start with a clean installation to confirm if the problem still exist. If you can reproduce it after installing a brand new hardy alpha 6, we can set the status to confirmed. Otherwise, this bug has many chances to persist because of the config files you have from older alpha. Is it possible for you to do that?

Revision history for this message
lod (altoas) wrote :

isn't it unconvinient to force users to reinstall upon new release?? what's the use of dist-upgrade then?

Revision history for this message
Милош (misos) wrote :

Is there another way to do the same? Is there a list of config files that X uses that I just can delete?

Revision history for this message
Saivann Carignan (oxmosys) wrote :

Yes it's inconvenient, I agree, however it's sometime normal inside a development release. We don't speak of a update between Gutsy - Hardy (which should not have this kinds of problems) but about a update between two development releases of the same system. I can't confirm this bug if it has not been tested at least with the latest hardy release since this released fixed the problem for a lot of people and since installing from the alpha is the only way I can consider to determine if this bug is fixed in that case or not. Sorry, I know that this is not great. I don't have a sufficient knowledge of all aspects of this bug so if I confirm this bug while it has not been tested with latest alpha ISO release (and with config files from latest ubiquity/jockey/Xorg/displayconfig-gtk), we have great chances to confirm a bug that is in fact fixed. Maybe you can just install hardy in a temporary partition/HD and see what's happening? Thanks for your contribution so far.

Revision history for this message
Милош (misos) wrote :

I did some testing, and came up with this:
1. If I boot normally, the login window does not show up at all, and the system reboots itself or lock up, with a black screen, and the cursor stuck at the top left corner, not blinking. The same thing happens with every driver (ati, radeonhd etc) driver.
2. If I go to recovery mode, and Resume right away, the system boots. So far I have tested radeonhd only. From the boot messages I saw, the only difference was that during the recovery mode, a message Saving VESA state showed up.

Any clue?

I'll install on a separate partition this weekend, since I havent downloaded the entire iso image.

Revision history for this message
Saivann Carignan (oxmosys) wrote :

Thank you very much for your testing, I will wait to have your comment once you tried with a fresh default install of hardy alpha 6.

Revision history for this message
lod (altoas) wrote :

isn't it better to try with latest daily live cd, since it would have the latest kernel and xorg, ati drivers?

Revision history for this message
Saivann Carignan (oxmosys) wrote :

lod : Daily build or the alpha 6, both should have this bug fixed.

Revision history for this message
lod (altoas) wrote :

installed fresh ubuntu from daily build and i have now some 3d but still there are some black diagonal triangles. but all seams to work fine. but i'm wondering will I have same issues when the final release comes and must I every time reinstall the hole system?

Revision history for this message
lod (altoas) wrote :

 the 3D apps are blinking..

Revision history for this message
Saivann Carignan (oxmosys) wrote :

lod : Thanks for your confirmation on this! If Милош confirms the same, this bug can be set as fixed.

No, with a final release, you should not have the obligation to reinstall your system, but it's pretty common during the development of a release (alpha releases exist for that purpose), you should keep in mind that Hardy alpha/beta releases are not for production use. If the final release ships some bad pre-configured things, updates will fix it without the needs of reinstalling your system.

Concerning strange 3D behavior you have, you can post a separate bug since it describe different issues with the ATI drivers (good package name is linux-restricted-modules-2.6.24).

Thanks for your comments, your participation with bug reports is really appreciated.

Revision history for this message
Jordan (jordanu) wrote :

Милош:

I am not a developer, and I am not claiming that this is a replacement for re-installing, but try:

sudo dpkg-reconfigure xserver-xorg -phigh

Then you will need to re-enable fglrx from System -> Administration -> Hardware Drivers

Then confirm if that works for you ( it did for me with the same problem ).

Revision history for this message
Bryce Harrington (bryce) wrote :

Setting to fix committed pending response from Милош.

Regarding xorg.conf, we deliberately do not alter the user's xorg.conf during upgrades, both because we can't be 100% certain that our updates may break things, and because the user probably has their xorg.conf set up the way they do for a reason.

If you can narrow down a specific xorg.conf setting that breaks on upgrade from Gutsy to Hardy, please do report that as a bug. But like Saïvann said, upgrades within a development release sometimes get broken. (Make sure to attach both your Gutsy and Hardy xorg.conf's).

Changed in linux-restricted-modules-2.6.24:
importance: Undecided → Medium
status: Incomplete → Fix Committed
Revision history for this message
Saivann Carignan (oxmosys) wrote :

Set back to incomplete, fix committed is for bugs that have patch ready to upload. In this case, we still don't know if this is a genuine bug or a duplicate of a already fixed bug.

Changed in linux-restricted-modules-2.6.24:
assignee: saivann → nobody
status: Fix Committed → Incomplete
Revision history for this message
Милош (misos) wrote :

I'm sorry I havent testest the lates live cd on a clean install, but will probably do it this week. However, I think I have located the problem.

misos@misos-laptop:~$ echo $DISPLAY
:1.0
misos@misos-laptop:~$ sudo aticonfig --query-monitor
  Connected monitors: none
  Enabled monitors: none
misos@misos-laptop:~$ export DISPLAY=:0.0
misos@misos-laptop:~$ sudo aticonfig --query-monitor
  Connected monitors: lvds
  Enabled monitors: lvds
misos@misos-laptop:~$ sudo aticonfig --enable-monitor=lvds
Using /etc/X11/xorg.conf
Saved back-up to /etc/X11/xorg.conf.fglrx-11
misos@misos-laptop:~$ fglrxinfo
display: :0.0 screen: 0
OpenGL vendor string: ATI Technologies Inc.
OpenGL renderer string: ATI Mobility Radeon X1400
OpenGL version string: 2.1.7412 Release

One problem though, the graphics remain slugish, as with the indirect driver.

Revision history for this message
PeterPall (peterpall) wrote :

I think I have the solution:

Ubuntu still tries to emulate 3D features by starting an XGL session even if the new fglrx driver knows how to do it by itselves - and will disable the 3D features that way.

You can disable the now unnecessary XGL by just creating a file with the name "disable" in the folder ".config/xserver-xgl" in your home directory

Just issue the following commands:

cd
mkdir -p .config/xserver-xgl/
touch disable

This should do it.
At least it works *fine* for me.

Revision history for this message
PeterPall (peterpall) wrote :

As a second idea You could try the following:
In /usr//lib there are libGL libraries that seem to be generic software rendering ones to me even if I have no real clue about these things.
In /usr/lib/fglrx there are files that should be hardware-specific 3d libraries. Don't know if Ubuntu knows that they are the right ones for Your setup.

Perhaps copying them to /usr/lib and renaming them to libGL.so, libGL.so.1 and libGL.so.1.2 will help solving the problem, too. I'm afraid I did this somehow during trying to find a solution to my 3D problems.

Revision history for this message
PeterPall (peterpall) wrote :

Found a solution to this Bug - even if it needs still manual shell work by the user to solve it.
I am therefore rather confident that this means we actually have a bug that was introduced by introducing AIGLX into fglrx.

Changed in linux-restricted-modules-2.6.24:
status: Incomplete → Confirmed
Revision history for this message
PeterPall (peterpall) wrote :

Tried it again.
Seems like disabling xgl did fox the problem. Copying the files in /usr/lib broke compiz, so I recommend not to play with them.
Would have been a rather monstrous Idea, anyway.

Revision history for this message
Милош (misos) wrote :

I can confirm, disabling xgl fixes the problem.

One problem with the libs, I have a broken symlink:
ls -l /usr/lib/fglrx/libGL.so.1.xlibmesa
lrwxrwxrwx 1 root root 12 2008-04-11 12:20 /usr/lib/fglrx/libGL.so.1.xlibmesa -> libGL.so.1.2

Revision history for this message
Saivann Carignan (oxmosys) wrote :

PeterPall or Милош : Did you try the latest Hardy release at least? Perhaps that this is already fixed here and if it is not, your comments on this would be very helpful since the RC approach and we don't want this issue to persist. I doubt that your issue is still valid with latest ISO since I have a ATI cards and everything works correctly but if this issue is still valid, we would need you to confirm it from the latest Hardy beta ISO release.

Revision history for this message
Shawn McMahon (smcmahon) wrote :

Started with an up-to-date Gutsy a couple of days ago. dist-upgrade to Hardy. (I have full disk encryption so I couldn't use the CD.)

Same problem. I tried PeterPall's fix (disabling Xgl) and it worked, all is good now.

ATI Radeon Mobility M10.

Revision history for this message
Saivann Carignan (oxmosys) wrote :

Than that means that the bug is still confirmed.

Revision history for this message
arno_b (arno.b) wrote :

> Милош wrote on 2008-04-12:
>
>One problem with the libs, I have a broken symlink:
>ls -l /usr/lib/fglrx/libGL.so.1.xlibmesa
>lrwxrwxrwx 1 root root 12 2008-04-11 12:20 /usr/lib/fglrx/libGL.so.1.xlibmesa -> libGL.so.1.2

I have this problem too, the link is broken.

Revision history for this message
youss44 (chahibi-deactivatedaccount) wrote :

Try to enable the restricted driver in "Hardware Drivers", remove the "fglrx" line in /etc/modprobe.d/blacklist-restricted , restart, and test again.

Revision history for this message
Alberto Milone (albertomilone) wrote :

/usr/lib/fglrx/libGL.so.1.xlibmesa is the result of a diversion of /usr/lib/libGL.so.1 performed by xorg-driver-fglrx. It shouldn't be a symlink.

I'm closing this bugreport since XGL was the cause of the problem instead of xorg-driver-fglrx.

Changed in linux-restricted-modules-2.6.24:
status: Confirmed → Invalid
Revision history for this message
gajanan (gajanan-khandake-hotmail) wrote :

Post #24 by PeterPall on 2008-04-12 worked for me.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.