Ubuntu

AGP not detected on Intel 8285P and E7205 chipsets using kernels higher than 2.6.17

Reported by Bernard van der Velden on 2007-01-10
40
Affects Status Importance Assigned to Milestone
linux (Ubuntu)
Undecided
Unassigned
linux-source-2.6.20 (Ubuntu)
Medium
Unassigned

Bug Description

Binary package hint: linux-image-2.6.20-6-generic

When booting Feisty (Kernel 2.6.20-6) my Xorg starts up, but I have no DRI. I use the opensource ati-driver. With kernel 2.6.17 (edgy) everything works fine.
Some highlights from my logs:

/var/log/kern.log:
Jan 10 12:30:45 bernard-desktop kernel: [ 56.584832] [drm] Initialized drm 1.1.0 20060810
Jan 10 12:30:45 bernard-desktop kernel: [ 56.598675] [drm] Initialized radeon 1.25.0 20060524 on minor 0
Jan 10 12:30:46 bernard-desktop kernel: [ 57.482494] [drm:radeon_cp_init] *ERROR* radeon_cp_init called without lock held
Jan 10 12:30:46 bernard-desktop kernel: [ 57.482795] [drm:drm_unlock] *ERROR* Process 4892 using kernel context 0

lsmod | grep agp:
intel_agp 24860 0
agpgart 33480 2 drm,intel_agp

/var/log/Xorg.0.log:
(WW) RADEON(0): [agp] AGP not available
(EE) RADEON(0): [agp] AGP failed to initialize. Disabling the DRI.
(II) RADEON(0): [agp] You may want to make sure the agpgart kernel module

I will attach the full logs to the bug.

======update======

It looks like the modules agpgart/intel_agp don't load my agp:

Kernel 2.6.20-6:
[39.240126] Linux agpgart interface v0.101 (c) Dave Jones
----
[49.257366] [drm] Initialized drm 1.1.0 20060810
[49.276651] ACPI: PCI Interrupt 0000:01:00.0[A] -> GSI 16 (level, low) -> IRQ 16
[49.278751] [drm] Initialized radeon 1.25.0 20060524 on minor 0
[50.191254] [drm:radeon_cp_init] *ERROR* radeon_cp_init called without lock held
[50.191473] [drm:drm_unlock] *ERROR* Process 4928 using kernel context 0

Kernel 2.6.17-10:
[17179584.076000] Linux agpgart interface v0.101 (c) Dave Jones
[17179584.096000] agpgart: Detected an Intel E7205 Chipset.
[17179584.100000] agpgart: AGP aperture is 128M @ 0xf0000000
----
[17179594.472000] [drm] Initialized drm 1.0.1 20051102
[17179594.484000] ACPI: PCI Interrupt 0000:01:00.0[A] -> GSI 16 (level, low) -> IRQ 185
[17179594.484000] [drm] Initialized radeon 1.25.0 20060524 on minor 0
[17179595.580000] agpgart: Found an AGP 3.0 compliant device at 0000:00:00.0.
[17179595.580000] agpgart: Putting AGP V3 device at 0000:00:00.0 into 8x mode
[17179595.580000] agpgart: Putting AGP V3 device at 0000:01:00.0 into 8x mode
[17179595.916000] [drm] Setting GART location based on new memory map
[17179595.916000] [drm] Loading R300 Microcode
[17179595.916000] [drm] writeback test succeeded in 1 usecs

description: updated
description: updated

Same with linux-image-2.6.20-6-generic...

description: updated
description: updated

Looks like an incompatibility between the drm and mesa drivers.

Please provide the output of "lspci -vvn" so I can check this against the PCI feature list in the radeon kernel driver.

Thanks

Changed in linux-source-2.6.20:
assignee: nobody → ben-collins
importance: Undecided → Medium
status: Unconfirmed → Needs Info

I attached the output.

Bernard

Alfredo Matas (amatas) wrote :

Hi, I've the same problem.

 dmesg | grep drm :

[ 54.893021] [drm] Initialized drm 1.1.0 20060810
[ 54.906160] [drm] Initialized radeon 1.25.0 20060524 on minor 0
[ 56.183127] [drm:radeon_cp_init] *ERROR* radeon_cp_init called without lock held
[ 56.183583] [drm:drm_unlock] *ERROR* Process 5691 using kernel context 0

 uname -a

 Linux 2.6.20-11-generic #2 SMP Thu Mar 15 08:03:07 UTC 2007 i686 GNU/Linux

Thanks

Alfredo Matas (amatas) wrote :

The status of this bug is "needs info", what info does the kernel team need?

Can I help?

regards.

Changed in linux-source-2.6.20:
assignee: ben-collins → ubuntu-kernel-team
status: Needs Info → Confirmed
Alfredo Matas (amatas) wrote :

In the bug you mention AGP is loaded correctly, in my situation it looks like AGP isn't detected at all which leads to DRI not loading...

You can check your AGP by running "dmesg | grep agp" in your terminal.

Regards, Bernard

Guere (michiel-4m) wrote :

Hi,

I'm also experiencing this problem. With 2.6.17 (my edgy kernel) AGP/DRI loads and works just fine, but when I use the testing (feisty) kernel (2.6.20-12.19), I'm getting the same 'error'. At first I only saw that xorg couldn't enable the DRI (because of AGP not available); but then I also searched the kernel log and realized it's probably because of this error during startup.

[ 28.704426] [drm] Initialized drm 1.1.0 20060810
[ 28.825819] ACPI: PCI Interrupt 0000:01:00.0[A] -> GSI 16 (level, low) -> IRQ 16
[ 28.826244] [drm] Initialized radeon 1.25.0 20060524 on minor 0
[ 29.062516] NET: Registered protocol family 17
[ 29.951055] [drm:radeon_cp_init] *ERROR* radeon_cp_init called without lock held
[ 29.951111] [drm:drm_unlock] *ERROR* Process 5247 using kernel context 0

My motherboard has an intel 8285P chipset, so even if the problem is related to the chipset, it's not Intel E7205 specific.

Alfredo Matas (amatas) wrote :

[ 22.765396] Linux agpgart interface v0.101 (c) Dave Jones
.... (that's all)....
[ 55.696273] [drm] Initialized drm 1.1.0 20060810
[ 55.724639] ACPI: PCI Interrupt 0000:01:00.0[A] -> GSI 16 (level, low) -> IRQ 16
[ 55.725145] [drm] Initialized radeon 1.25.0 20060524 on minor 0
[ 56.372993] ppdev: user-space parallel port driver
[ 57.173689] [drm:radeon_cp_init] *ERROR* radeon_cp_init called without lock held
[ 57.173747] [drm:drm_unlock] *ERROR* Process 5725 using kernel context 0

I've attached my dmesg.

regards

Alfredo: looks like the same problem indeed...
Guere: could you also attach your full dmesg or dmesg | grep agp so we can be sure your agp also isn't detected? Your lspci -vvn could also be useful...

Regards.

Guere (michiel-4m) wrote :

Of course.

Guere (michiel-4m) wrote :

And here's the lspci.

Guere: also looks the same. Some extra help from the kernel-team or someone else would be handy since I don't know how to pin down the problem better than "AGP not detected on Intel 8285P and E7205 chipsets using kernels higher than 2.6.17".

Alfredo Matas (amatas) wrote :

This bug is still affecting kernel 2.6.20-13

Regards

Edu (martinez-bernal) wrote :

Yesterday I installed feisty beta in an Duron 1Ghz machine with a Radeon 7000. I don't know wich chipset uses, but it loads the via-agp module (not the intel one). DRI doesn't work and I get the same messages.

Guere (michiel-4m) wrote :

I'm also still having this problem. I've tried a few things - including the hints given at http://dri.freedesktop.org/wiki/DriTroubleshooting - but nothing seems to work. I knew the module loading order probably didn't have anything to do with it (since it has worked before), but I wanted to exhaust my options before posting here again. Each and every time X loads, the same radeon_cp_init error pops up.

The weird thing is that we seem to be the only ones experiencing this problem. I've been keeping an eye out on various Ubuntu forums, but haven't seen anyone else reporting this problem yet. And because it's so rare it'll probably not be easy to recreate the problem or track down the cause...

Alfredo Matas (amatas) wrote :

I've just boot my machine with an edgy live cd and I could see that everything that depends on the intel chipset is loaded correctly:

edgy (2.6.17):

lsmod | grep -i intel:

snd_intel8x0 34844 1
snd_ac97_codec 97696 2 snd_emu10k1,snd_intel8x0
snd_pcm 84612 4 snd_emu10k1,snd_intel8x0,snd_pcm_oss,snd_ac97_codec
snd 58372 16 snd_emux_synth,snd_seq_virmidi,snd_seq_oss,snd_seq,snd_emu10k1,snd_rawmidi,snd_intel8x0,snd_pcm_oss,snd_ac97_codec,snd_seq_device,snd_mixer_oss,snd_pcm,snd_hwdep,snd_timer
intel_agp 26012 1
agpgart 34888 2 drm,intel_agp

feisty (2.6.20):

lsmod | grep -i intel
snd_intel8x0 34204 0
snd_ac97_codec 98336 2 snd_emu10k1,snd_intel8x0
snd_pcm 79876 4 snd_emu10k1,snd_intel8x0,snd_ac97_codec,snd_pcm_oss
snd 54020 16 snd_emux_synth,snd_seq_virmidi,snd_seq_oss,snd_seq,snd_emu10k1,snd_rawmidi,snd_intel8x0,snd_ac97_codec,snd_seq_device,snd_pcm_oss,snd_mixer_oss,snd_pcm,snd_hwdep,snd_timer
snd_page_alloc 10888 3 snd_emu10k1,snd_intel8x0,snd_pcm
intel_agp 25116 0
agpgart 35400 2 drm,intel_agp

As you can see, in the feisty lsmod, snd_intel8x0 hasn't detected my chipset i875P, neither the intel_agp. I usually use an emu10k1 sound card, because of that i couldn't see this before.
I think that this bug is near the kernel chipset detection, instead the agpgart module or radeon module.

regards

Z Knezo (burntloser) wrote :

I also believe I have this problem, but with an Intel 875P chipset. Originally thought I couldn't get DRI due to video drivers, but then saw the agp error.

dmesg attached. lspci -vvn to follow.

I never looked at direct rendering with edgy, but I'll try a Live CD later to see it it works.

Z Knezo (burntloser) wrote :
Philipp Stampfli (phildst) wrote :

I'm affected too.
lspci -vvn and dmesg attached if more info is needed just ask

Philipp Stampfli (phildst) wrote :
Z Knezo (burntloser) wrote :

Just to confirm, my AGP is recognized and DRI works fine with a 6.10 Live CD.

Alfredo Matas (amatas) wrote :

I've boot with a 2.6.20.4 stable kernel of kernel.org and it has the same problem.

Guere (michiel-4m) wrote :

Yea, I've also tried building my own kernel (based on the Ubuntu source package, though), but no luck. AGP is still screwed and not found by Xorg. And since I really doubt this problem is going to magically disappear before or when Feisty goes stable, I think we've kinda run out of things to try...

Alfredo Matas (amatas) wrote :

I've removed the following modules and agp works again with the kernel 2.6.20-14:

rm -fr /lib/modules/2.6.20-14-generic/kernel/drivers/edac

can anyone test it?

Regards

Guere (michiel-4m) wrote :

Confirmed. Once I removed the edac modules with rmmod i82875p_edac and rmmod edac_mc, my intel_agp was finally 'bounded' right with agpgart. I'm a bit hesitant to just remove modules, and since any new kernel will have them again, I'm still a bit curious why this these modules are preventing AGP from working correctly. I have an 82875P chipset, so why would that module screw things up? Weird..

But thanks a lot for figuring out this much!

Alfredo Matas (amatas) wrote :

I compiled the kernel with pci debug option and, when I rebooted the computer i got this lines in dmesg:

[ 22.726080] EDAC i82875p: i82875p init one
[ 22.726151] PCI: Found 0000:00:06.0 [8086/257e] 000880 00
[ 22.726155] PCI: Calling quirk c01fb080 for 0000:00:06.0
[ 22.726196] PCI: Calling quirk c0275f80 for 0000:00:06.0
[ 22.726200] PCI: Calling quirk c0275960 for 0000:00:06.0
[ 22.726214] PCI: Unable to reserve mem region #1:1000@fecf0000 for device 0000:00:06.0

I'm not a kernel hacker but I think that EDAC doesn't work fine with some chipsets... and it may cause that some modules fails when they loads.

Strange... When I blacklist the modules e7xxx_edac and edac_mc I get opengl acceleration, only glxinfo reports the direct rendering isn't available:

bernard@bernard-desktop:~$ 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_import_context, GLX_EXT_texture_from_pixmap,
    GLX_EXT_visual_info, GLX_EXT_visual_rating, GLX_MESA_copy_sub_buffer,
    GLX_OML_swap_method, GLX_SGI_make_current_read, GLX_SGI_swap_control,
    GLX_SGIS_multisample, GLX_SGIX_fbconfig, GLX_SGIX_visual_select_group
client glx vendor string: ATI
client glx version string: 1.3
client glx extensions:
    GLX_EXT_visual_info, GLX_EXT_visual_rating, GLX_EXT_import_context,
    GLX_ARB_get_proc_address, GLX_SGI_video_sync, GLX_ARB_multisample,
    GLX_ATI_pixel_format_float, GLX_ATI_render_texture
GLX version: 1.2
GLX extensions:
    GLX_EXT_visual_info, GLX_EXT_visual_rating, GLX_EXT_import_context,
    GLX_ARB_multisample
OpenGL vendor string: Tungsten Graphics, Inc.
OpenGL renderer string: Mesa DRI R300 20060815 AGP 8x x86/MMX/SSE2 TCL
OpenGL version string: 1.3 Mesa 6.5.2
OpenGL extensions:
    GL_ARB_imaging, GL_ARB_multitexture, GL_ARB_texture_border_clamp,
    GL_ARB_texture_cube_map, GL_ARB_texture_env_add,
    GL_ARB_texture_env_combine, GL_ARB_texture_env_dot3,
    GL_ARB_transpose_matrix, GL_EXT_abgr, GL_EXT_blend_color,
    GL_EXT_blend_minmax, GL_EXT_blend_subtract, GL_EXT_texture_env_add,
    GL_EXT_texture_env_combine, GL_EXT_texture_env_dot3,
    GL_EXT_texture_lod_bias

   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 32 0 r y . 8 8 8 8 0 24 8 0 0 0 0 0 0 None
0x24 24 tc 0 32 0 r y . 8 8 8 8 0 24 0 0 0 0 0 0 0 None
0x25 24 tc 0 32 0 r y . 8 8 8 8 0 24 8 16 16 16 16 0 0 Slow
0x26 24 tc 0 32 0 r y . 8 8 8 8 0 24 0 16 16 16 16 0 0 Slow
0x27 24 tc 0 32 0 r . . 8 8 8 8 0 24 8 0 0 0 0 0 0 None
0x28 24 tc 0 32 0 r . . 8 8 8 8 0 24 0 0 0 0 0 0 0 None
0x29 24 tc 0 32 0 r . . 8 8 8 8 0 24 8 16 16 16 16 0 0 Slow
0x2a 24 tc 0 32 0 r . . 8 8 8 8 0 24 0 16 16 16 16 0 0 Slow
0x2b 24 dc 0 32 0 r y . 8 8 8 8 0 24 8 0 0 0 0 0 0 None
0x2c 24 dc 0 32 0 r y . 8 8 8 8 0 24 0 0 0 0 0 0 0 None
0x2d 24 dc 0 32 0 r y . 8 8 8 8 0 24 8 16 16 16 16 0 0 Slow
0x2e 24 dc 0 32 0 r y . 8 8 8 8 0 24 0 16 16 16 16 0 0 Slow
0x2f 24 dc 0 32 0 r . . 8 8 8 8 0 24 8 0 0 0 0 0 0 None
0x30 24 dc 0 32 0 r . . 8 8 8 8 0 24 0 0 0 0 0 0 0 None
0x31 24 dc 0 32 0 r . . 8 8 8 8 0 24 8 16 16 16 16 0 0 Slow
0x32 24 dc 0 32 0 r . . 8 8 8 8 0 24 0 16 16 16 16 0 0 Slow
0x4b 32 tc 1 0 0 c . . 0 0 0 0 0 0 0 0 0 0 0 0 0 None

In short, glx says direct rendering isn't available, but it is working... Tested playing Chromium....

Ben Collins (ben-collins) wrote :

Removing milestone. This wont get fixed for release. Perhaps in an update though.

Confirmed that removing edac related modules and reinserting fglrx fixes agpgart behaviour and restores DRI for me on latest build. Beryl still seems unhappy about something but at least we have a specific place to look for the bug.

Given this affects a fair chunk of the user population I'm a bit underwhelmed that this is not considered a showstopper.

Nicolas Gorguès (gorni-siloh) wrote :

Also confirmed here with kernel 2.6.20 (Feisty Fawn 7.04):
- Shapphire RADEON X800 XL (AGP)
- Chipset Intel 865PE (MB ASUS P4P800)

As a soft workaround, blacklisting the modules has worked perfectly.
=> add the following two lines at the beginning of /etc/modprobe.d/blacklist:
blacklist i82875p_edac
blacklist edac_mc

Now desktop effects (wobbly+cube) work like a charm with the open source RADEON driver.

Sergey Voronin (crazyvoron) wrote :

i have gigabyte motherboard based on intel 875p chipset
and ati radeon 9600xt agp

and i have the same problem
dri works fine on edgy
and agp cannot init on feisty (kernel 2.6.20.x xorg 7.2)

no matter what driver i use fglrx or radeon
blacklisting do not work at all
i have no idea what to do

need any solution
and off course if i can help somehow - i will

ps sorry if my english isn't good enough :)

Cameron Kilgore (ghostfreeman) wrote :

Also confirming this with 2.6.20-16-generic (Feisty) on an Intel 82875P chipset and Radeon 9800XT

walec51 (walec51) wrote :

Had the same problem. Lost 3D acceleration after upgrade from edgy to feisty.

Nicolas Gorguès workaround worked for me.

Mathias Gaunard (loufoque) wrote :

Problem confirmed with gusty, 2.6.22-8-generic

Pekka Panula (grimdin) wrote :

I can also confirm that edac -modules seems to prohibit AGP module to work under Intel 975p chipset & with ATI (AMD) Radeon 9500 (RV200 chipset)
Blacklisting i82875p_edac and edac_mc modules fixes this atm, i got current Feisty and current kernel.

Pekka Panula (grimdin) wrote :

small typo, not 975P but should be 875P chipset

Will we see this fixed before gutsy is released?

Phil Colbourn (philcolbourn) wrote :

I had not noticed any problem with feisty 2.6.20-16 (ie. compiz worked).

Under gutsy 2.6.22-10 compiz doesn't work. I get similar messages in Xorg.0.log:
WW RADEON AGP not available
EE RADEON AGP failed to initialize. Disabling the DRI
II RADEON You may want to make sure the agpgart kernel module is loaded before the radeon kernel module.

So, I rmmod radeon and the modprobe radeon and then logged out and back in. compiz now works.

This is repeatable.

I have a compaq n620c

Can I supply anymore information to help?

DJ_DEF (dj-def) wrote :

I noticed that not always this happens with feisty:
I installed feisty and i tried glxinfo | grep render. It said NO.
So I installed edgy and retried. It said YES.
I reinstalled feisty and retried. It said YES.
I reinstalled again feisty and retried. It said NO.
This is C R A Z Y !!!!!!!!!!
All the installations had been made with same options, same way and always reformatting the HD.

my hardware: Radeon 9600, Intel Corporation E7505/E7205 PCI-to-AGP Bridge on p4g8x motherboard.

DJ_DEF (dj-def) wrote :

news:

Blacklisting drm, radeon, agpgart will not work, I don't know why, or better, it will work but when X will start they will be autoloaded if in xorg.conf radeon driver is set.

I tried to put vesa in xorg.conf, blacklist all the drm, radeon, agpgart, e7xxx_edac, edac_mc, reboot the system and drm, radeon, agpgart too disappeard from lsmod output.

At this point I did:

sudo modprobe agpgart
sudo modprobe intel_agp
sudo modprobe drm
sudo modprobe radeon

Then I changed vesa to radeon in xorg.conf, I hit ctrl+alt+backspace and X restarted.
Nomore radeon_cp_init called without lock held and glxinfo | grep render said YES!

I dit this with a vanilla kernel (2.6.21.5) but I think it should work with ubuntu kernel too.

I think the real problem is that all the necessary modules are not loaded in the correct order and this metod grants the correct order load procedure.

Let me know what you think.

DJ_DEF (dj-def) wrote :

The things above work with ubuntu kernel too but there is a shorter way to let the things work.

I solved the problem simply blacklisting e7xxx_edac and edac_mc, then putting this lines in this exact order in /etc/modules:

agpgart
intel_agp
drm

(intel_agp works fine with my p4g8x but it should be replaced with the right module for your AGP)
When X server will start, it will automatically load the radeon driver (if present in xorg.conf).
This procedure grants the correct order.
When X started if you want, you can modprobe e7xxx_edac and edac_mc, it will not affect dri.
You can also try to load e7xxx_edac and edac_mc after drm in /etc/module after drm, I don't know if it will work, but I think that doing such a thing could be useless.

DJ_DEF (dj-def) wrote :

Summary:

/etc/modprobe.d/blacklist :

e7xxx_edac
edac_mc

/etc/modules for radeon and ati driver:

agpgart
intel_agp
drm

/etc/modules for fglrx driver :
# drm don't have to be here!!!

Same problem with a nvidia nforce3 250 chipset and a radeon 9600 on a new feisty install...

Just after the install, direct rendering was ok, but after all updates (121 updates...), I lost 3D acceleration...

edac modules are not loaded by default on my install...

DJ_DEF (dj-def) wrote :

try to edit /etc/modules and put there in this order:

agpgart
#change this line with your chip agp module#
drm #this line only if you use radeon driver, if you use fglrx this line have not to be here.

You should get back direct rendering: yes

Already test it, but it don't works....

i will attach some logs

DJ_DEF (dj-def) wrote :

Your problem could be the amd64_agp module, try this:

/etc/module

agpgart
nvidia_agp
drm
radeon

and this:

/etc/modprobe.d/blacklist

blacklist amd64_agp

Then if it doesn't work re-post a kernel log and a lsmod output.

DJ_DEF (dj-def) wrote :

there was an error in the comment above, /etc/modules and not /etc/module!

amd64_agp is now blacklisted, after reboot amd64_agp is not loaded (normal), but it didn't correct the problem

DJ_DEF (dj-def) wrote :

try this:

1)
/etc/modules

agpgart
nvidia_agp
drm
radeon

/etc/modprobe.d/blacklist

agpgart
nvidia_agp
amd64_agp
drm
radeon

2)
/etc/modules

agpgart
amd64_agp
drm
radeon

/etc/modprobe.d/blacklist

agpgart
nvidia_agp
amd64_agp
drm
radeon

3)
/etc/modules

agpgart
amd64_agp
nvidia_agp
drm
radeon

/etc/modprobe.d/blacklist

agpgart
nvidia_agp
amd64_agp
drm
radeon

4)
/etc/modules

agpgart
nvidia_agp
amd64_agp
drm
radeon

/etc/modprobe.d/blacklist

agpgart
nvidia_agp
amd64_agp
drm
radeon

Among these one configuration have to work. Then POST a kernel log, we have to verify the presence of [drm:radeon_cp_init] *ERROR* radeon_cp_init called without lock held.

I've tested all configurations, but none of them works....

DJ_DEF (dj-def) wrote :

It seems like both nvidia_agp and amd64_agp cause some problem.
I also see that 8 October you had dri working and there was no "agpgart: Setting up Nforce3 AGP" after agpgart loading.
Maybe you have to blacklist nvidia_agp AND amd64_agp.
Try this:

/etc/modules
#you can remove all the things you added

/etc/blacklist
#here you have to put both:
blacklist nvidia_agp
blacklist amd64_agp

Then post your kernel.log.

DJ_DEF (dj-def) wrote :

I read better, that line is everywhere and 8 October dri was not working... but it was working till 7 October... and agpgart before 8 October was loaded towards the end of the log.
We have to discover the cause. However try to blacklist nvidia_agp and amd64_agp as above.

Blacklist nvidia_agp and amd64_agp didn't work.

Ubuntu was installed on 7 october, and I lost dri after all updates. On the new install without any updates, 3D acceleration works fine, but after all updates there was this problem

kern.log is in attachment

Maybe dpkg.log can be usefull...

I update ubuntu (something like 120 packages...) on october 7, but I didn't reboot before ocotber 8, this is why dri worked fine the 7, and not the 8

DJ_DEF (dj-def) wrote :

maybe we are near to the solution:

now leave nvidia_agp and amd64_agp blacklisted and put this lines in /etc/modules

agpgart
drm

If this change will not work we will have to find THE module that is loaded before agpgart and that causes the problem.

No changes....

kern.log is in attachment

DJ_DEF (dj-def) wrote :

Ok, so restore your /etc/modules and remove the blacklists.
Now try to set agp aperture to 128M in your BIOS.

DJ_DEF (dj-def) wrote :

If aperture size changing doesn't work you may consider to use the previous kernel version (you should still have an entry in your bootloader).

I will test apeture size, but for the previous kernel version, I already tested it with no success :\

AGP aperture size was already to 128M

When passing option "BusType" in xorg.conf to PCI mode, dri is now working and drm start without errors. I don't know why it was working fine in AGP mode before doing all updates...

DJ_DEF (dj-def) wrote :

so... why now is this in kernel log? -> agpgart: aperture base > 4G

before there was: AGP aperture is 128M @ 0xf0000000

Can you post an "lspci -n" command output and your /etc/X11/xorg.conf?

Strange....

I will test with a slower AGP apeture size in the bios...

DJ_DEF (dj-def) wrote :

I found a possible solution:

http://wiki.archlinux.org/index.php/ATI_Radeon_&_Kernel_2.6 (see hardware conflicts section)
http://forums.gentoo.org/viewtopic-p-3954435.html
http://bugzilla.kernel.org/show_bug.cgi?id=6350

At the moment it seems that downgrading bios or patching the kernel is the only solution.

Lowering AGP aperture size in the BIOS doesn't change anything...

I will look at those links!

Thanks for your help!

Ok I've tested to boot on windows and then warm boot on linux, and I get dri working, so I need to downgrade the bios (this is the problem mentionned in the link you gave: http://bugzilla.kernel.org/show_bug.cgi?id=6350)

DJ_DEF (dj-def) wrote :

yes, you need that or a kernel patch... I hope that soon kernel developers will fix this problem

Pekka Panula (grimdin) wrote :

I upgraded to Gutsy Gibbon (RC-version) and still same problem, but, fix is same as with Feisty Fawn, eg.

/etc/modprobe.d/blacklist :
blacklist i82875p_edac
blacklist edac_mc

After that AGP-modules, and ATI drivers, etc. loads ok and direct rendering ok.

smacfarl (smacfarl) wrote :

So I have an ATI Radeon 7000 card. No 3d acceleration at all until I applied the blacklist and module fix.

I am currently running Gutsy, and actually upgraded because I wasn't getting OpenGL support in feisty.

Since this card is RV100 it's supposed to be 100% supported out of the box. What gives?

Also now that I can get direct render to indicate yes from glxinfo, I have a lot of other random problems. So it looks like stuff may still need a little tweaking.

You can read about my travails here. http://ubuntuforums.org/showthread.php?t=570382

I am user smacfarl in the linked thread. The helpful person on the ubuntu forums sent me to this for to regale you all with my woes.

What can I do to help get this problem fixed?

dobratzp (peter-dobratz) wrote :

Just wiped my Dell Dimension 8300 of Ubuntu 7.10 and did a clean install of Ubuntu 8.04 and this bug still persists. To resolve the problem, I did the previously mentioned solution of adding lines to /etc/modprobe.d/blacklist:

blacklist i82875p_edac
blacklist edac_mc

After adding this and restarting, /var/log/Xorg.0.log does not list any errors.

J Hertel (herteljt) wrote :

I have a Radeon 9600 and can confirm that with Ubuntu 8.04

adding lines to /etc/modprobe.d/blacklist:

blacklist i82875p_edac
blacklist edac_mc

Fixes the issue. Thank you for this thread, I was beginning drift back to the dark side just to play fps :)

Richard (maclan) wrote :

I have a Radeon 9800XP and a Dell Dimension 8300 that I have been trying to get compiz in Hardy 8.04 to work with. I finally got it to work by blacklisting the edac modules.

adding lines to /etc/modprobe.d/blacklist:

blacklist i82875p_edac
blacklist edac_mc

Thank you so much guys. I hope other people are able to get some help from this.

Beginning with the Hardy Heron 8.04 development cycle, all open Ubuntu kernel bugs need to be reported against the "linux" kernel package. We are automatically migrating this bug to the new "linux" package. However, development has already began for the upcoming Intrepid Ibex 8.10 release. It would be helpful if you could test the upcoming release and verify if this is still an issue - http://www.ubuntu.com/testing . If the issue still exists, please update this report by changing the Status of the "linux" task from "Incomplete" to "New". We appreciate your patience and understanding as we make this transition. Thanks!

The Ubuntu Kernel Team is planning to move to the 2.6.27 kernel for the upcoming Intrepid Ibex 8.10 release. As a result, the kernel team would appreciate it if you could please test this newer 2.6.27 Ubuntu kernel. There are one of two ways you should be able to test:

1) If you are comfortable installing packages on your own, the linux-image-2.6.27-* package is currently available for you to install and test.

--or--

2) The upcoming Alpha5 for Intrepid Ibex 8.10 will contain this newer 2.6.27 Ubuntu kernel. Alpha5 is set to be released Thursday Sept 4. Please watch http://www.ubuntu.com/testing for Alpha5 to be announced. You should then be able to test via a LiveCD.

Please let us know immediately if this newer 2.6.27 kernel resolves the bug reported here or if the issue remains. More importantly, please open a new bug report for each new bug/regression introduced by the 2.6.27 kernel and tag the bug report with 'linux-2.6.27'. Also, please specifically note if the issue does or does not appear in the 2.6.26 kernel. Thanks again, we really appreicate your help and feedback.

Downloaded the Live CD beta last night.

Burned it. The Live CD does not identify my graphics card. After the boot
bar I get a blank screen, with my lcd monitor claiming "unknown signal and
then "going to sleep". Pressing alt-f2 and looking at the xorg.conf shows a
"configured graphics card" rather than the "ati" driver in my 8.04
xorg.conf. Selecting safe mode on the livecd has no effect.

What can I do to help resolve this,
Shawn MacFarland

On Thu, Aug 28, 2008 at 7:33 PM, Leann Ogasawara <email address hidden> wrote:

> The Ubuntu Kernel Team is planning to move to the 2.6.27 kernel for the
> upcoming Intrepid Ibex 8.10 release. As a result, the kernel team would
> appreciate it if you could please test this newer 2.6.27 Ubuntu kernel.
> There are one of two ways you should be able to test:
>
> 1) If you are comfortable installing packages on your own, the linux-
> image-2.6.27-* package is currently available for you to install and
> test.
>
> --or--
>
> 2) The upcoming Alpha5 for Intrepid Ibex 8.10 will contain this newer
> 2.6.27 Ubuntu kernel. Alpha5 is set to be released Thursday Sept 4.
> Please watch http://www.ubuntu.com/testing for Alpha5 to be announced.
> You should then be able to test via a LiveCD.
>
> Please let us know immediately if this newer 2.6.27 kernel resolves the
> bug reported here or if the issue remains. More importantly, please
> open a new bug report for each new bug/regression introduced by the
> 2.6.27 kernel and tag the bug report with 'linux-2.6.27'. Also, please
> specifically note if the issue does or does not appear in the 2.6.26
> kernel. Thanks again, we really appreicate your help and feedback.
>
> ** Tags added: cft-2.6.27
>
> --
> AGP not detected on Intel 8285P and E7205 chipsets using kernels higher
> than 2.6.17
> https://bugs.launchpad.net/bugs/78684
> You received this bug notification because you are a direct subscriber
> of the bug.
>

Intrepid Ibex 8.10 was released at the end of October. Assuming that you are still experiencing issues, if you can please attach an updated dmesg output from the 2.6.27-7 Intrepid kernel that would be great. Thanks.

Per a decision made by the Ubuntu Kernel Team, bugs will longer be assigned to the ubuntu-kernel-team in Launchpad as part of the bug triage process. The ubuntu-kernel-team is being unassigned from this bug report. Refer to https://wiki.ubuntu.com/KernelTeamBugPolicies for more information. Thanks.

dobratzp (peter-dobratz) wrote :

I just installed Ubuntu 8.10 on a machine that was experiencing this problem with prior releases of Ubuntu. Without adding any lines to the blacklist file, there are no errors in the Xorg log file for me. The visual effects appear to be working properly. I'm using a Radeon 9800 Pro (Radeon R350) in a Dell Dimension 8300.

Thanks dobratzp. Can anyone else confirm?

With no other feedback, I'll go ahead and mark this Fix Released based on https://bugs.edge.launchpad.net/ubuntu/+source/linux-source-2.6.20/+bug/78684/comments/94

Changed in linux (Ubuntu):
status: Incomplete → Fix Released
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Duplicates of this bug

Other bug subscribers

Remote bug watches

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