xf86-video-intel 2.9.99.902-1 breaks stuff

eldragon · 2009-12-21 02:31:56

dmesg gets spammed with

...
[drm:i915_gem_execbuffer] *ERROR* Execbuf while wedged
[drm:i915_gem_execbuffer] *ERROR* Execbuf while wedged
[drm:i915_gem_execbuffer] *ERROR* Execbuf while wedged
[drm:i915_gem_execbuffer] *ERROR* Execbuf while wedged
[drm:i915_gem_execbuffer] *ERROR* Execbuf while wedged
[drm:i915_gem_execbuffer] *ERROR* Execbuf while wedged
[drm:i915_gem_execbuffer] *ERROR* Execbuf while wedged
...

by breaks stuff i mean, only sysrq to reboot the system

any clues?

combuster · 2009-12-21 08:35:19

New requirements of 2.10 compared to 2.9
----------------------------------------
* Linux kernel with kernel-modesetting (KMS)
All user-modesetting code has now been removed from the driver.
We highly recommend kernel version 2.6.32 or later.
* Libdrm >= 2.4.16

Do you have kms enabled? I'm going to try to compile it now (from git) and report back later if any problems appear...

Ok, tested it with 2.6.32.2 and 2.6.31-rc1 and works ok for now. Does this kernel panic appear random, right after you login or after specific tasks (gaming, watching videos or something else)?

Kernels mentioned above are custom with kms enabled and i915 driver is compiled into the kernel image.

Last edited by combuster (2009-12-21 08:48:53)

chpln · 2009-12-21 09:50:42

Similar issues for me, though the symptoms varied. I experienced X failing to start (repeated errors similar to that of eldragon printed in stderr if I recall correctly), X locking (but still being able to switch to a tty and recover), and X locking and needing to use Sysrq to reboot.

% lspci | grep VGA
00:02.0 VGA compatible controller: Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller (rev 03)

kernel26-2.6.32.1-1 with KMS enabled.

I've since reverted back to xf86-video-intel-2.9.1-1.

combuster · 2009-12-21 10:03:45

I have an uptime of 1.5h on 2.6.33-rc1 and xf86-video-intel 2.9.99.902 and no problem so far. Played urban terror, several video clips (with xvmc output), and it's still rocksolid.

lspci | grep VGA
00:02.0 VGA compatible controller: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 0c)

It could be a chipset specific issue...

eldragon · 2009-12-21 11:55:24

what chpln said exactly.

same hardware.

Sertse · 2009-12-21 12:18:50

I get X freezing, requiring a reboot. KMS enabled, kernel from testing.

00:02.0 VGA compatible controller: Intel Corporation 82915G/GV/910GL Integrated Graphics Controller (rev 0e)

patroclo7 · 2009-12-21 12:45:09

Kernel from testing, KMS enabled, sysrq is the only solution anytime I try to start the X server.

00:02.0 VGA compatible controller: Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller (rev 03)
00:02.1 Display controller: Intel Corporation Mobile 945GM/GMS/GME, 943/940GML Express Integrated Graphics Controller (rev 03)

jowilkin · 2009-12-21 14:44:13

I think I'm having this issue as well.

jowilkin@copper ~ $ lspci | grep 'Intel.*Graphics'
00:02.0 VGA compatible controller: Intel Corporation Mobile 4 Series Chipset Integrated Graphics Controller (rev 07)
00:02.1 Display controller: Intel Corporation Mobile 4 Series Chipset Integrated Graphics Controller (rev 07)

I get a complete freeze, only a hard reboot works.

.:B:. · 2009-12-21 15:34:54

[stijn@hermes downloads]$ lspci |grep Graphics
00:02.0 VGA compatible controller: Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller (rev 03)
00:02.1 Display controller: Intel Corporation Mobile 945GM/GMS/GME, 943/940GML Express Integrated Graphics Controller (rev 03)
[stijn@hermes downloads]$

Works fine here.

[stijn@hermes ~]$ dmesg |grep drm
[drm] Initialized drm 1.1.0 20060810
[drm] LVDS-8: set mode 1366x768 d
[drm] fb0: inteldrmfb frame buffer device
[drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0

No graphical corruption either. KMS is enabled.

[stijn@hermes ~]$ zgrep -i kms /proc/config.gz 
# CONFIG_DRM_I915_KMS is not set

Hmm . I should set that in my kernel instead of using the initcpio image .

Edit: I'm using my own 2.6.31 kernel.

ilpianista · 2009-12-21 18:24:40

same problem as eldragon, KMS enabled (anyway isn't enabled by default?)

00:02.0 VGA compatible controller: Intel Corporation Mobile 915GM/GMS/910GML Express Graphics Controller (rev 03)

kernel26 2.6.32.2-1
xf86-video-intel 2.9.99.902-1

[drm] Initialized drm 1.1.0 20060810                                                                                                     
i915 0000:00:02.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16                                                                            
i915 0000:00:02.0: setting latency timer to 64                                                                                           
[drm] set up 7M of stolen space                                                                                                          
[drm] TV-14: set mode NTSC 480i 0                                                                                                        
render error detected, EIR: 0x00000010                                                                                                   
page table error                                                                                                                         
  PGTBL_ER: 0x00000100                                                                                                                   
[drm:i915_handle_error] *ERROR* EIR stuck: 0x00000010, masking                                                                           
render error detected, EIR: 0x00000010                                                                                                   
page table error                                                                                                                         
  PGTBL_ER: 0x00000100                                                                                                                   
[drm] DAC-6: set mode 1280x1024 2d                                                                                                       
[drm] LVDS-8: set mode 1280x768 2e                                                                                                       
Console: switching to colour frame buffer device 160x48                                                                                  
fb0: inteldrmfb frame buffer device                                                                                                      
registered panic notifier                                                                                                                
[drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0
[...]
[drm:i915_gem_execbuffer] *ERROR* Execbuf while wedged
[drm:i915_gem_execbuffer] *ERROR* Execbuf while wedged
[...]

jowilkin · 2009-12-21 21:01:40

What is a known good config on these intel cards? I have a new laptop I'm trying to get running. I tried rolling back to xf86-video-intel 2.9.1-1 and kernel26 2.6.32.1-1, but I'm still having issues.

.:B:. · 2009-12-21 21:18:41

I am having lockups it seems with the xf86-video-intel from testing. I rolled back to the one in extra for the time being, I had two lockups within a few hours.

Skoll · 2009-12-21 22:52:49

X freezed two times on my laptop and I decided to replace libdrm and xf86-video-intel with libdrm-newest and xf86-intel-newest from AUR. Everything seems stable even with KMS enabled.

Edit: VGA compatible controller: Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller (rev 03)

Last edited by Skoll (2009-12-21 22:57:02)

JGC · 2009-12-21 23:26:19

The libdrm-newest and xf86-video-intel-newest packages from AUR are actually the same versions as we have, so that shouldn't make a difference. I think the new driver utilizes KMS code that is broken in 2.6.32, causing the GPU lockups. I also reverted my system to the stable version for now.

On the other side, there's 845G hardware that locks up instantly with the stable driver, but works fine with the new driver.

eldragon · 2009-12-22 11:37:46

JGC wrote:

The libdrm-newest and xf86-video-intel-newest packages from AUR are actually the same versions as we have, so that shouldn't make a difference. I think the new driver utilizes KMS code that is broken in 2.6.32, causing the GPU lockups. I also reverted my system to the stable version for now.
On the other side, there's 845G hardware that locks up instantly with the stable driver, but works fine with the new driver.

im using 2.6.33-rc1 with 2.9.99 and it works ok so far... maybe i can get suspend to work with this toaster now

JGC · 2009-12-22 14:26:29

Try booting with i915.powersave=0. According to some developers the problems are caused by the new powermanagement features in the 2.6.32 kernel. I think some of that has been reverted in 2.6.33.

Renan Birck · 2009-12-22 15:16:41

I also am having similar problems. X crashes often, especially when loading lots of tabs in Firefox.

As I am using testing, I will downgrade to the driver in [extra].

jowilkin · 2009-12-22 18:14:41

I rolled back to the driver in extra and still had problems. I rolled my kernel back to 2.6.31.6 and now seem to have a stable system.

firecat53 · 2009-12-22 18:35:05

I have

00:02.0 VGA compatible controller: Intel Corporation 82845G/GL[Brookdale-G]/GE Chipset Integrated Graphics Device (rev 01)

and with 2.6.32 kernel from testing and the 2.9.99.902-3 xf86-video-intel-newest, I get lockups with the following errors:

EE intel(0): Failed to submit batch buff, expecting rendering corruption or even a frozen display: Input/Output error

This is actually an improvement over previous versions, as you can actually ctrl-alt-backspace to kill X and go back to the console. If you try to restart X it just gives a black screen. Previous versions hard-locked everything except moving the mouse.

Scott

anrxc · 2009-12-22 19:48:11

This problem (looks like multiple problems actually) is known for a while now, for example:
https://bugs.freedesktop.org/show_bug.cgi?id=25475
https://bugs.freedesktop.org/show_bug.cgi?id=25406
https://bugs.freedesktop.org/show_bug.cgi?id=25598

Some patches are already available, unfortunately for the kernel... keep an eye on those discussions.

Rednaxela · 2009-12-28 20:37:00

After my first reboot after a long while of constant uptime I started having this/these issue(s). Been seeing the same kind of errors everyone has been reporting above. I've found it to affect:
- 2.6.32 kernel with 2.9.99.902 drivers
- 2.6.32 kernel with 2.9.1 drivers
- 2.6.31 kernel with 2.9.99.902 drivers
So stuck on non-testing for both kernel26 and xf86-video-intel in order to have a system that is remotely usable. Does anyone have any of the above combinations working at all?

chpln · 2009-12-28 22:23:36

Rednaxela wrote:

- 2.6.32 kernel with 2.9.99.902 drivers
- 2.6.32 kernel with 2.9.1 drivers
- 2.6.31 kernel with 2.9.99.902 drivers

kernel26 2.6.32 with xf86-video-intel 2.9.1 are without issue here. I only experience problems with 2.9.99.902. Though there seems to be others reporting the same issue with 2.6.32: here.

eldragon · 2009-12-29 03:33:14

mesa 7.7 and its child packages + 2.6.32 work for me.. just to let people know

Onwards · 2009-12-30 10:27:10

JGC wrote:

Try booting with i915.powersave=0. According to some developers the problems are caused by the new powermanagement features in the 2.6.32 kernel. I think some of that has been reverted in 2.6.33.

That solves my X freezes. Switching back and forth b/w vts and X causes problems though.
I get these lines in my X-log:

Fatal server error:
Failed to submit batchbuffer: Input/output error

(II) Macintosh mouse button emulation: Close
(II) UnloadModule: "evdev"
(II) AT Translated Set 2 keyboard: Close
(II) UnloadModule: "evdev"
(II) USB Optical Mouse: Close
(II) UnloadModule: "evdev"
(II) Sleep Button: Close
(II) UnloadModule: "evdev"
(II) Power Button: Close
(II) UnloadModule: "evdev"

nightmaresiege · 2010-01-01 02:53:39

This issue makes my MSI Wind essentially unusable. It almost seems that the freezing is completely random. I'm having a hard time reproducing it. Using powersave=0 has not fixed the problem for me.

Last edited by nightmaresiege (2010-01-01 02:54:13)

Arch Linux

#1 2009-12-21 02:31:56

xf86-video-intel 2.9.99.902-1 breaks stuff

#2 2009-12-21 08:35:19

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#3 2009-12-21 09:50:42

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#4 2009-12-21 10:03:45

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#5 2009-12-21 11:55:24

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#6 2009-12-21 12:18:50

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#7 2009-12-21 12:45:09

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#8 2009-12-21 14:44:13

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#9 2009-12-21 15:34:54

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#10 2009-12-21 18:24:40

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#11 2009-12-21 21:01:40

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#12 2009-12-21 21:18:41

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#13 2009-12-21 22:52:49

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#14 2009-12-21 23:26:19

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#15 2009-12-22 11:37:46

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#16 2009-12-22 14:26:29

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#17 2009-12-22 15:16:41

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#18 2009-12-22 18:14:41

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#19 2009-12-22 18:35:05

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#20 2009-12-22 19:48:11

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#21 2009-12-28 20:37:00

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#22 2009-12-28 22:23:36

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#23 2009-12-29 03:33:14

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#24 2009-12-30 10:27:10

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

#25 2010-01-01 02:53:39

Re: xf86-video-intel 2.9.99.902-1 breaks stuff

Board footer