Ubuntu
xserver-xorg-video-ati package

Xorg freeze, radeon gpu lockup with softreset

Bug #936159 reported by Ernst Persson
44
This bug affects 9 people
Affects Status Importance Assigned to Milestone
X.Org X server
Fix Released
Medium
xserver-xorg-video-ati (Ubuntu)
Fix Released
High
Unassigned

Bug Description

This happens once every week with precise.
When it happens it usually loops until I switch to the terminal and killall -9 compiz.
So obviously compiz is triggering it because it's the only program using 3D.
This happens during normal desktop usage, maybe when new windows are created or similar.

ProblemType: Bug
DistroRelease: Ubuntu 12.04
Package: xorg 1:7.6+10ubuntu1
ProcVersionSignature: Ubuntu 3.2.0-16.25-generic 3.2.6
Uname: Linux 3.2.0-16-generic x86_64
.tmp.unity.support.test.0:

ApportVersion: 1.91-0ubuntu1
Architecture: amd64
CompizPlugins: [core,bailer,detection,composite,opengl,compiztoolbox,decor,snap,move,grid,gnomecompat,place,vpswitch,resize,mousepoll,imgpng,session,regex,wall,animation,expo,fade,ezoom,workarounds,scale,unityshell]
CompositorRunning: compiz
Date: Sun Feb 19 13:16:54 2012
DistUpgraded: Log time: 2012-02-07 10:27:17.880836
DistroCodename: precise
DistroVariant: ubuntu
ExtraDebuggingInterest: Yes, whatever it takes to get this fixed in Ubuntu
GpuHangFrequency: Once a week
GpuHangReproducibility: Seems to happen randomly
GpuHangStarted: Immediately after installing this version of Ubuntu
GraphicsCard:
 Advanced Micro Devices [AMD] nee ATI Barts PRO [ATI Radeon HD 6800 Series] [1002:6739] (prog-if 00 [VGA controller])
   Subsystem: PC Partner Limited Device [174b:174b]
MachineType: System manufacturer System Product Name
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-3.2.0-16-generic root=UUID=f5cb51ea-ed8f-4da3-a77c-a3594554d0a9 ro radeon.pcie_gen2=1 quiet splash vt.handoff=7
SourcePackage: xorg
Symptom: display
Title: Xorg freeze
UpgradeStatus: Upgraded to precise on 2012-02-07 (12 days ago)
dmi.bios.date: 06/08/2011
dmi.bios.vendor: American Megatrends Inc.
dmi.bios.version: 1704
dmi.board.asset.tag: To be filled by O.E.M.
dmi.board.name: P8P67 PRO REV 3.1
dmi.board.vendor: ASUSTeK Computer INC.
dmi.board.version: Rev 1.xx
dmi.chassis.asset.tag: Asset-1234567890
dmi.chassis.type: 3
dmi.chassis.vendor: Chassis Manufacture
dmi.chassis.version: Chassis Version
dmi.modalias: dmi:bvnAmericanMegatrendsInc.:bvr1704:bd06/08/2011:svnSystemmanufacturer:pnSystemProductName:pvrSystemVersion:rvnASUSTeKComputerINC.:rnP8P67PROREV3.1:rvrRev1.xx:cvnChassisManufacture:ct3:cvrChassisVersion:
dmi.product.name: System Product Name
dmi.product.version: System Version
dmi.sys.vendor: System manufacturer
version.compiz: compiz 1:0.9.7.0~bzr2995-0ubuntu4
version.ia32-libs: ia32-libs 20090808ubuntu33
version.libdrm2: libdrm2 2.4.30-1ubuntu1
version.libgl1-mesa-dri: libgl1-mesa-dri 8.0~rc2-0ubuntu5
version.libgl1-mesa-dri-experimental: libgl1-mesa-dri-experimental N/A
version.libgl1-mesa-glx: libgl1-mesa-glx 8.0~rc2-0ubuntu5
version.xserver-xorg-core: xserver-xorg-core 2:1.11.4-0ubuntu3
version.xserver-xorg-input-evdev: xserver-xorg-input-evdev 1:2.6.99.901+git20120126-0ubuntu2
version.xserver-xorg-video-ati: xserver-xorg-video-ati 1:6.14.99~git20111219.aacbd629-0ubuntu2
version.xserver-xorg-video-intel: xserver-xorg-video-intel 2:2.17.0-1ubuntu4
version.xserver-xorg-video-nouveau: xserver-xorg-video-nouveau N/A

Revision history for this message
In , Ernst Persson (ernstp) wrote :

Created attachment 56286
GPU locksups log

I'm getting gpu lockups every now and then. More likely during with heavy OpenGL usage, but Compiz is enough sometimes. I can switch to console and kill the application and usually get a functioning system back.
Setup:
Intel Core i5 2500k
Radeon 6850
Ubuntu 11.10
Kernel: Ubuntu 3.2.0-7.13-generic 3.2.0-rc7
3D stack from xorg-edgers ppa

Revision history for this message
Ernst Persson (ernstp) wrote :
Bug Watch Updater (bug-watch-updater)
Changed in xorg-server:
importance: Unknown → Medium
status: Unknown → Confirmed
Revision history for this message
Ernst Persson (ernstp) wrote :

Happened again today:

[ 1673.084392] radeon 0000:01:00.0: GPU lockup CP stall for more than 10004msec
[ 1673.084394] GPU lockup (waiting for 0x0001838A last fence id 0x00018387)
[ 1673.085465] radeon 0000:01:00.0: GPU softreset
[ 1673.085466] radeon 0000:01:00.0: GRBM_STATUS=0xB2701828
[ 1673.085467] radeon 0000:01:00.0: GRBM_STATUS_SE0=0x08000003
[ 1673.085468] radeon 0000:01:00.0: GRBM_STATUS_SE1=0x1C000003
[ 1673.085469] radeon 0000:01:00.0: SRBM_STATUS=0x200000C0
[ 1673.085478] radeon 0000:01:00.0: GRBM_SOFT_RESET=0x00007F6B
[ 1673.085579] radeon 0000:01:00.0: GRBM_STATUS=0x00003828
[ 1673.085580] radeon 0000:01:00.0: GRBM_STATUS_SE0=0x00000007
[ 1673.085581] radeon 0000:01:00.0: GRBM_STATUS_SE1=0x00000007
[ 1673.085582] radeon 0000:01:00.0: SRBM_STATUS=0x200000C0
[ 1673.086576] radeon 0000:01:00.0: GPU reset succeed
[ 1673.107797] [drm] PCIE GART of 512M enabled (table at 0x0000000000040000).
[ 1673.107884] radeon 0000:01:00.0: WB enabled
[ 1673.124003] [drm] ring test succeeded in 3 usecs
[ 1673.124012] [drm] ib test succeeded in 3 usecs

Revision history for this message
In , Torsten Krah (tkrah) wrote :

Created attachment 57325
GPU lookup - kernel 3.2.6; Ubuntu Oneiric

Got some GPU lookup today too (hope its the same thing and the correct bug report to add to) - happend the first time.
No openGL involved afaik can tell.
System recovered without i need to kill something, but mouse cursor is borked - so i needed to restart X anyway.

Revision history for this message
In , Torsten Krah (tkrah) wrote :

Addon: Restart X did not help, mouse still garbage - had to reboot. Any other option i can try next time?

Revision history for this message
In , Ernst Persson (ernstp) wrote :

So you don't run a compositing windowmanager?
killall -9 compiz get's me back to a fully working desktop.

Bryce Harrington (bryce)
affects: xorg (Ubuntu) → xserver-xorg-video-ati (Ubuntu)
Revision history for this message
Ernst Persson (ernstp) wrote :

This is sometimes triggered by a "fast user switch". But not every time, and most of the time by other things.
I've added the xorg-edgers ppa now to see if it still happens.
I guess the bug is in userspace, it sends a bad command stream or something?

Revision history for this message
In , Ernst Persson (ernstp) wrote :

Created attachment 57602
More lockups

More lockups

Bryce Harrington (bryce)
Changed in xserver-xorg-video-ati (Ubuntu):
status: New → Triaged
importance: Undecided → Medium
importance: Medium → High
Revision history for this message
In , Ernst Persson (ernstp) wrote :

This might be related to reclocking the GPU because when I tried dynpm it happened more often and it often happens when waking the computer up from DPMS.

Revision history for this message
In , Ernst Persson (ernstp) wrote :

I can now reproduce this consistenly I think:
Install Ubuntu Precise
Add xorg-edgers ppa
Create a 2:nd user
Log in as user 1
Switch to user 2
Switch to user 1

Then when you have unlocked user 1's screen with your password and it redraws the desktop I get a GPU reset loop.

Revision history for this message
Michael Tsang (miklcct) wrote :

I'm also suffering this on my Debian wheezy system. The symptom is display frozen for 10 seconds, a flicker than returns to normal.

Here's the related dmesg output:

Mar 24 11:36:06 server kernel: [ 8901.252040] radeon 0000:01:05.0: GPU lockup CP stall for more than 10000msec
Mar 24 11:36:06 server kernel: [ 8901.252046] GPU lockup (waiting for 0x000170CC last fence id 0x000170CB)
Mar 24 11:36:06 server kernel: [ 8901.253150] radeon 0000:01:05.0: GPU softreset
Mar 24 11:36:06 server kernel: [ 8901.253153] radeon 0000:01:05.0: R_008010_GRBM_STATUS=0xA27334AC
Mar 24 11:36:06 server kernel: [ 8901.253156] radeon 0000:01:05.0: R_008014_GRBM_STATUS2=0x00000103
Mar 24 11:36:06 server kernel: [ 8901.253159] radeon 0000:01:05.0: R_000E50_SRBM_STATUS=0x20000040
Mar 24 11:36:06 server kernel: [ 8901.253167] radeon 0000:01:05.0: R_008020_GRBM_SOFT_RESET=0x00007FEE
Mar 24 11:36:06 server kernel: [ 8901.268052] radeon 0000:01:05.0: R_008020_GRBM_SOFT_RESET=0x00000001
Mar 24 11:36:06 server kernel: [ 8901.283921] radeon 0000:01:05.0: R_008010_GRBM_STATUS=0xA0003030
Mar 24 11:36:06 server kernel: [ 8901.283924] radeon 0000:01:05.0: R_008014_GRBM_STATUS2=0x00000003
Mar 24 11:36:06 server kernel: [ 8901.283928] radeon 0000:01:05.0: R_000E50_SRBM_STATUS=0x20008040
Mar 24 11:36:06 server kernel: [ 8901.284930] radeon 0000:01:05.0: GPU reset succeed
Mar 24 11:36:06 server kernel: [ 8901.306127] [drm] PCIE GART of 512M enabled (table at 0x00000000C0040000).
Mar 24 11:36:06 server kernel: [ 8901.306188] radeon 0000:01:05.0: WB enabled
Mar 24 11:36:06 server kernel: [ 8901.338269] [drm] ring test succeeded in 0 usecs
Mar 24 11:36:06 server kernel: [ 8901.338279] [drm] ib test succeeded in 1 usecs

I don't have any /etc/X11/xorg.conf

Revision history for this message
In , Ernst Persson (ernstp) wrote :
Download full text (5.5 KiB)

Now this happened to me during login! :-(
[ 188.237505] radeon 0000:01:00.0: GPU lockup CP stall for more than 10008msec
[ 188.237511] GPU lockup (waiting for 0x00000E47 last fence id 0x00000E46)
[ 188.238681] radeon 0000:01:00.0: GPU softreset
[ 188.238684] radeon 0000:01:00.0: GRBM_STATUS=0xB2703828
[ 188.238687] radeon 0000:01:00.0: GRBM_STATUS_SE0=0x1C000007
[ 188.238689] radeon 0000:01:00.0: GRBM_STATUS_SE1=0x08000007
[ 188.238692] radeon 0000:01:00.0: SRBM_STATUS=0x200000C0
[ 188.238704] radeon 0000:01:00.0: GRBM_SOFT_RESET=0x00007F6B
[ 188.238806] radeon 0000:01:00.0: GRBM_STATUS=0x00003828
[ 188.238809] radeon 0000:01:00.0: GRBM_STATUS_SE0=0x00000007
[ 188.238811] radeon 0000:01:00.0: GRBM_STATUS_SE1=0x00000007
[ 188.238814] radeon 0000:01:00.0: SRBM_STATUS=0x200000C0
[ 188.239811] radeon 0000:01:00.0: GPU reset succeed
[ 188.262214] [drm] PCIE GART of 512M enabled (table at 0x0000000000040000).
[ 188.262326] radeon 0000:01:00.0: WB enabled
[ 188.278447] [drm] ring test succeeded in 2 usecs
[ 188.278456] [drm] ib test succeeded in 2 usecs
[ 199.622493] radeon 0000:01:00.0: GPU lockup CP stall for more than 10020msec
[ 199.622498] GPU lockup (waiting for 0x00000E63 last fence id 0x00000E60)
[ 199.623667] radeon 0000:01:00.0: GPU softreset
[ 199.623670] radeon 0000:01:00.0: GRBM_STATUS=0xB2701828
[ 199.623673] radeon 0000:01:00.0: GRBM_STATUS_SE0=0x1C000003
[ 199.623675] radeon 0000:01:00.0: GRBM_STATUS_SE1=0x08000003
[ 199.623678] radeon 0000:01:00.0: SRBM_STATUS=0x20000AC0
[ 199.623689] radeon 0000:01:00.0: GRBM_SOFT_RESET=0x00007F6B
[ 199.623792] radeon 0000:01:00.0: GRBM_STATUS=0x00003828
[ 199.623794] radeon 0000:01:00.0: GRBM_STATUS_SE0=0x00000007
[ 199.623797] radeon 0000:01:00.0: GRBM_STATUS_SE1=0x00000007
[ 199.623800] radeon 0000:01:00.0: SRBM_STATUS=0x200000C0
[ 199.624796] radeon 0000:01:00.0: GPU reset succeed
[ 199.647176] [drm] PCIE GART of 512M enabled (table at 0x0000000000040000).
[ 199.647279] radeon 0000:01:00.0: WB enabled
[ 199.901376] [drm] ring test succeeded in 3 usecs
[ 199.901393] [drm] ib test succeeded in 3 usecs
[ 210.619789] radeon 0000:01:00.0: GPU lockup CP stall for more than 10004msec
[ 210.619794] GPU lockup (waiting for 0x00000F19 last fence id 0x00000F18)
[ 210.620964] radeon 0000:01:00.0: GPU softreset
[ 210.620967] radeon 0000:01:00.0: GRBM_STATUS=0xB2701828
[ 210.620970] radeon 0000:01:00.0: GRBM_STATUS_SE0=0x08000003
[ 210.620973] radeon 0000:01:00.0: GRBM_STATUS_SE1=0x1C000003
[ 210.620975] radeon 0000:01:00.0: SRBM_STATUS=0x200000C0
[ 210.620987] radeon 0000:01:00.0: GRBM_SOFT_RESET=0x00007F6B
[ 210.621089] radeon 0000:01:00.0: GRBM_STATUS=0x00003828
[ 210.621092] radeon 0000:01:00.0: GRBM_STATUS_SE0=0x00000007
[ 210.621094] radeon 0000:01:00.0: GRBM_STATUS_SE1=0x00000007
[ 210.621097] radeon 0000:01:00.0: SRBM_STATUS=0x200000C0
[ 210.622094] radeon 0000:01:00.0: GPU reset succeed
[ 210.644468] [drm] PCIE GART of 512M enabled (table at 0x0000000000040000).
[ 210.644581] radeon 0000:01:00.0: WB enabled
[ 210.660703] [drm] ring test succeeded in 2 usecs
[ 210.660712] [drm] ib test succeede...

Read more...

Revision history for this message
In , Ernst Persson (ernstp) wrote :

(In reply to comment #6)
> I can now reproduce this consistenly I think:
> Install Ubuntu Precise
> Add xorg-edgers ppa
> Create a 2:nd user
> Log in as user 1
> Switch to user 2
> Switch to user 1
>
> Then when you have unlocked user 1's screen with your password and it redraws
> the desktop I get a GPU reset loop.

Still easily reproducible with 3.4-rc2 vanilla and xorg-edgers. No config options, power-profile forced to high.

Revision history for this message
Xavier Tardivel (the-frog45) wrote :

I am affected with a stock Precise install. It seems to happen more often when I am invoking the Workspace switcher or starting an application, and for what it worth I also got a 6850 GPU.

Revision history for this message
Ernst Persson (ernstp) wrote :

You can use Unity 2D as a workaround, doesn't trigger this even with compositing.

Revision history for this message
In , Ernst Persson (ernstp) wrote :

I can no longer reproduce this with latest Xorg driver/mesa/drmlib! It's fixed in userspace somehow, not kernel.

Revision history for this message
Ernst Persson (ernstp) wrote :

I can no longer reproduce this with latest xorg-edgers! Xorg-edgers hadn't been updated in a while and then there was a big push around 2nd or 4th May, and that solved it. I did not update the kernel when this stopped happening, so the fix seems to be in userspace. No idea what to backport...

Bug Watch Updater (bug-watch-updater)
Changed in xorg-server:
status: Confirmed → Fix Released
Revision history for this message
Oibaf (oibaf) wrote :

Is this still an issue?

Changed in xserver-xorg-video-ati (Ubuntu):
status: Triaged → Incomplete
Revision history for this message
Xavier Tardivel (the-frog45) wrote :

I had no lock up since about a year.

Revision history for this message
Oibaf (oibaf) wrote :

Looks fixed then.

Changed in xserver-xorg-video-ati (Ubuntu):
status: Incomplete → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.
Subscribing...

Other bug subscribers

Remote bug watches

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