ATI radeon KMS driver - gpu lockup

Bug #568605 reported by Michael B. Trausch
94
This bug affects 20 people
Affects Status Importance Assigned to Milestone
linux (Ubuntu)
Expired
High
Unassigned
Lucid
Invalid
High
Unassigned

Bug Description

Using Lucid, up-to-date, kernel 2.6.32-21-generic. I have had this problem since 2.6.32-19-generic, where I was going to report it but updates were available, so I had hoped that the problem would go away. However, it persists.

Using previous kernel 2.6.32-16-generic, this problem did not seem to exist (if it had, it was so rare that I never actually ran into it).

The new kernel now corrupts the display awfully after anywhere from 1 to 8 hours of uptime. I can Control+Alt+F1 to go back to a VT, but then the screen starts to "melt". The system then responds to a Control+Alt+Delete and cleanly reboots.

Sometimes, when I switch back to VT1, I get a console that is extremely corrupt, with what should be column 0, row 0 appearing somewhere in the middle of the display with strange wrapping, or I get an apparently normal console that corrupts itself the next time the kernel attempts to reset the graphics hardware as shown in the log below. The log snippet below was taken under -19-generic, but appears to be the same with -21-generic.

When I encounter this problem, the following shows up in dmesg:

Apr 12 21:01:22 fennel kernel: [11341.070194] [drm:radeon_fence_wait] *ERROR* fence(ffff88009e78a400:0x00048BFE) 510ms timeout going to reset GPU
Apr 12 21:01:22 fennel kernel: [11341.070206] radeon 0000:01:00.0: GPU softreset
Apr 12 21:01:22 fennel kernel: [11341.070212] radeon 0000:01:00.0: R_008010_GRBM_STATUS=0xE00014A4
Apr 12 21:01:22 fennel kernel: [11341.070219] radeon 0000:01:00.0: R_008014_GRBM_STATUS2=0x00300002
Apr 12 21:01:22 fennel kernel: [11341.070225] radeon 0000:01:00.0: R_000E50_SRBM_STATUS=0x200030C0
Apr 12 21:01:23 fennel kernel: [11341.247761] radeon 0000:01:00.0: Wait for MC idle timedout !
Apr 12 21:01:23 fennel kernel: [11341.247767] radeon 0000:01:00.0: R_008020_GRBM_SOFT_RESET=0x00007FEE
Apr 12 21:01:23 fennel kernel: [11341.247825] radeon 0000:01:00.0: R_008020_GRBM_SOFT_RESET=0x00000001
Apr 12 21:01:23 fennel kernel: [11341.247890] radeon 0000:01:00.0: R_000E60_SRBM_SOFT_RESET=0x00000C02
Apr 12 21:01:23 fennel kernel: [11341.296894] radeon 0000:01:00.0: R_008010_GRBM_STATUS=0x00003028
Apr 12 21:01:23 fennel kernel: [11341.296901] radeon 0000:01:00.0: R_008014_GRBM_STATUS2=0x00000002
Apr 12 21:01:23 fennel kernel: [11341.296906] radeon 0000:01:00.0: R_000E50_SRBM_STATUS=0x200000C0
Apr 12 21:01:23 fennel kernel: [11341.299750] [drm:radeon_fence_wait] *ERROR* fence(ffff88009e78a400:0x00048BFE) 740ms timeout
Apr 12 21:01:23 fennel kernel: [11341.299753] [drm:radeon_fence_wait] *ERROR* last signaled fence(0x00048BFE)
Apr 12 21:01:25 fennel kernel: [11343.822667] [drm:radeon_fence_wait] *ERROR* fence(ffff8801088a6880:0x00048BFF) 510ms timeout going to reset GPU
Apr 12 21:01:25 fennel kernel: [11343.822679] radeon 0000:01:00.0: GPU softreset
Apr 12 21:01:25 fennel kernel: [11343.822686] radeon 0000:01:00.0: R_008010_GRBM_STATUS=0xA0003028
Apr 12 21:01:25 fennel kernel: [11343.822693] radeon 0000:01:00.0: R_008014_GRBM_STATUS2=0x00000002
Apr 12 21:01:25 fennel kernel: [11343.822698] radeon 0000:01:00.0: R_000E50_SRBM_STATUS=0x200000C0
Apr 12 21:01:25 fennel kernel: [11343.822710] radeon 0000:01:00.0: R_008020_GRBM_SOFT_RESET=0x00007FEE
Apr 12 21:01:25 fennel kernel: [11343.822766] radeon 0000:01:00.0: R_008020_GRBM_SOFT_RESET=0x00000001
Apr 12 21:01:25 fennel kernel: [11343.822831] radeon 0000:01:00.0: R_000E60_SRBM_SOFT_RESET=0x00000402
Apr 12 21:01:25 fennel kernel: [11343.871740] radeon 0000:01:00.0: R_008010_GRBM_STATUS=0x00003028
Apr 12 21:01:25 fennel kernel: [11343.871746] radeon 0000:01:00.0: R_008014_GRBM_STATUS2=0x00000002
Apr 12 21:01:25 fennel kernel: [11343.871752] radeon 0000:01:00.0: R_000E50_SRBM_STATUS=0x200000C0
Apr 12 21:01:25 fennel kernel: [11343.874241] [drm:radeon_fence_wait] *ERROR* fence(ffff8801088a6880:0x00048BFF) 570ms timeout
Apr 12 21:01:25 fennel kernel: [11343.874246] [drm:radeon_fence_wait] *ERROR* last signaled fence(0x00048BFF)
Apr 12 21:01:28 fennel kernel: [11346.390188] [drm:radeon_fence_wait] *ERROR* fence(ffff8801088a6400:0x00048C02) 510ms timeout going to reset GPU
Apr 12 21:01:28 fennel kernel: [11346.390200] radeon 0000:01:00.0: GPU softreset
Apr 12 21:01:28 fennel kernel: [11346.390206] radeon 0000:01:00.0: R_008010_GRBM_STATUS=0xA0003028
Apr 12 21:01:28 fennel kernel: [11346.390213] radeon 0000:01:00.0: R_008014_GRBM_STATUS2=0x00000002
Apr 12 21:01:28 fennel kernel: [11346.390219] radeon 0000:01:00.0: R_000E50_SRBM_STATUS=0x200000C0
Apr 12 21:01:28 fennel kernel: [11346.390230] radeon 0000:01:00.0: R_008020_GRBM_SOFT_RESET=0x00007FEE
Apr 12 21:01:28 fennel kernel: [11346.390287] radeon 0000:01:00.0: R_008020_GRBM_SOFT_RESET=0x00000001
Apr 12 21:01:28 fennel kernel: [11346.390352] radeon 0000:01:00.0: R_000E60_SRBM_SOFT_RESET=0x00000402
Apr 12 21:01:28 fennel kernel: [11346.439261] radeon 0000:01:00.0: R_008010_GRBM_STATUS=0x00003028
Apr 12 21:01:28 fennel kernel: [11346.439267] radeon 0000:01:00.0: R_008014_GRBM_STATUS2=0x00000002
Apr 12 21:01:28 fennel kernel: [11346.439273] radeon 0000:01:00.0: R_000E50_SRBM_STATUS=0x200000C0
Apr 12 21:01:28 fennel kernel: [11346.441532] [drm:radeon_fence_wait] *ERROR* fence(ffff8801088a6400:0x00048C02) 570ms timeout
Apr 12 21:01:28 fennel kernel: [11346.441534] [drm:radeon_fence_wait] *ERROR* last signaled fence(0x00048C02)

All subsequent such blocks in dmesg are identical to the second and third ones shown above.

There does not seem to be any specific behavior that I am doing that is triggering the behavior. At least, not to my knowledge.

I don't really know what else to report here, so please ask for additional information. I will try to apport-collect momentarily, as well.

ProblemType: Bug
DistroRelease: Ubuntu 10.04
Package: linux-image-2.6.32-21-generic 2.6.32-21.32
Regression: Yes
Reproducible: No
ProcVersionSignature: Ubuntu 2.6.32-21.32-generic 2.6.32.11+drm33.2
Uname: Linux 2.6.32-21-generic x86_64
AlsaVersion: Advanced Linux Sound Architecture Driver Version 1.0.21.
Architecture: amd64
ArecordDevices:
 **** List of CAPTURE Hardware Devices ****
 card 0: SB [HDA ATI SB], device 0: STAC92xx Analog [STAC92xx Analog]
   Subdevices: 1/1
   Subdevice #0: subdevice #0
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC0: mbt 1681 F.... pulseaudio
CRDA: Error: [Errno 2] No such file or directory
Card0.Amixer.info:
 Card hw:0 'SB'/'HDA ATI SB at 0xf2400000 irq 16'
   Mixer name : 'IDT 92HD75B3X5'
   Components : 'HDA:111d7603,103c363a,00100202 HDA:11c11040,103c137e,00100200'
   Controls : 20
   Simple ctrls : 12
Card1.Amixer.info:
 Card hw:1 'HDMI'/'HDA ATI HDMI at 0xf2310000 irq 19'
   Mixer name : 'ATI R6xx HDMI'
   Components : 'HDA:1002aa01,00aa0100,00100100'
   Controls : 4
   Simple ctrls : 1
Card1.Amixer.values:
 Simple mixer control 'IEC958',0
   Capabilities: pswitch pswitch-joined penum
   Playback channels: Mono
   Mono: Playback [off]
Date: Thu Apr 22 14:07:47 2010
Frequency: Once a day.
HibernationDevice: RESUME=UUID=036b9a25-3a7c-4e34-93da-98ea8812867b
InstallationMedia: Ubuntu 10.04 "Lucid Lynx" - Beta amd64 (20100318)
MachineType: Hewlett-Packard HP Pavilion dv7 Notebook PC
ProcCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.32-21-generic root=UUID=873a8d7a-e443-400b-890c-c18008f6069c ro quiet splash
ProcEnviron:
 LANG=en_US.utf8
 SHELL=/bin/bash
RelatedPackageVersions: linux-firmware 1.34
RfKill:
 0: phy0: Wireless LAN
  Soft blocked: no
  Hard blocked: no
SourcePackage: linux
dmi.bios.date: 10/10/2009
dmi.bios.vendor: Hewlett-Packard
dmi.bios.version: F.07
dmi.board.name: 363A
dmi.board.vendor: Quanta
dmi.board.version: 33.22
dmi.chassis.type: 10
dmi.chassis.vendor: Quanta
dmi.chassis.version: N/A
dmi.modalias: dmi:bvnHewlett-Packard:bvrF.07:bd10/10/2009:svnHewlett-Packard:pnHPPaviliondv7NotebookPC:pvr039E200000241220000020000:rvnQuanta:rn363A:rvr33.22:cvnQuanta:ct10:cvrN/A:
dmi.product.name: HP Pavilion dv7 Notebook PC
dmi.product.version: 039E200000241220000020000
dmi.sys.vendor: Hewlett-Packard

Revision history for this message
Michael B. Trausch (mtrausch) wrote :
Revision history for this message
Michael B. Trausch (mtrausch) wrote :

Nevermind on the apport-collect statement; I forgot to edit it out since the bug was reported with ubuntu-bug instead of directly through LP.

Revision history for this message
Martin Reiche (grimsrud) wrote :
Bryce Harrington (bryce)
Changed in linux (Ubuntu):
importance: Undecided → High
Revision history for this message
Bryce Harrington (bryce) wrote :

User comment from dupe bug:

This does look like it's similar to my bug 568605, but I used the kernel as the thing to blame, since this problem also affects the standard text-mode console with KMS. I am currently running the packaged upstream to see if the problem persists there, and will try other upstream kernels built using the Ubuntu config files to see if I can replicate this at all upstream. I don't know where the git kernel trees are for Lucid, or I would try to bisect the problem (though note that as unreliably reproducable as this problem appears to be, it may take me longer to bisect than there is time until release day; I'd expect to be able to perform one bisection step every 24–48 hours to try to determine if the problem exists in the kernel or not). Usually anytime I bisect a kernel problem, there are anywhere from 8 to 15 steps involved, so I'd expect anywhere from 16 to 30 days to work through a full bisection properly unless I can find some way to trigger this bug on-demand.

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

If you have any clues on types of activities that might be triggering it (compiz effects, window movement, vt switching, etc.) a common technique is to make a script that does those activities in a loop. Sometimes that can make a bug reproduce a lot faster.

Btw, this bug is titled 'crash' but that's not correct. A crash would give you a stacktrace. You're seeing a GPU lockup.

summary: - ATI radeon KMS driver crashes randomly
+ ATI radeon KMS driver - gpu lockup
tags: added: xorg-needs-kernel-fix
Revision history for this message
Michael B. Trausch (mtrausch) wrote :

I'm using the word "crash" in the sense that a computer system that has ceased to operate correctly is said to have crashed, at least insofar as all the developers I am aware of have used it (including myself). In this case, my meaning should be quite clear: The graphics driver encounters a condition which it cannot rectify and ceases to correctly function with the violent consequence of having to reboot the system without saving any data. I suspect that if the GPU locked up, I would not even be able to see the dmesg output when I switch to the VT, because the GPU would be unable to execute further instructions to draw (and further, if that were the case, it would probably act like my system with the NVIDIA chipset, where the GPU locking up causes the kernel to subsequently lock up, making even the magic sysrq key fail to work, thus requiring a reset or power-cycle to get back up and running).

Software can and does crash without detecting that it crashed.

As an aside, things are so far so good running:

 Linux fennel 2.6.32-02063211-generic #02063211 SMP Mon Apr 12 12:03:56 UTC 2010 x86_64 GNU/Linux

From the mainline packages. I'll comment back in several hours if it hasn't yet crashed.

The only thing that I know is that there is no single activity that occurs before it crashes; sometimes, it will crash just sitting still, without having moved the mouse or switched windows or any of that. Compiz is running on this system (the only aspect that I have tailored of the UI is which fonts I use, not the effects or the extensiveness of them).

Revision history for this message
Jeremy Foshee (jeremyfoshee) wrote :

Hi Michael,

If you could also please test the latest upstream kernel available that would be great. It will allow additional upstream developers to examine the issue. Refer to https://wiki.ubuntu.com/KernelMainlineBuilds . Once you've tested the upstream kernel, please remove the 'needs-upstream-testing' tag. This can be done by clicking on the yellow pencil icon next to the tag located at the bottom of the bug description and deleting the 'needs-upstream-testing' text. Please let us know your results.

Thanks in advance.

    [This is an automated message. Apologies if it has reached you inappropriately; please just reply to this message indicating so.]

tags: added: kj-triage
Revision history for this message
Michael B. Trausch (mtrausch) wrote : Re: [Bug 568605] Re: ATI radeon KMS driver - gpu lockup

On Fri, 2010-04-23 at 00:14 +0000, Jeremy Foshee wrote:
> If you could also please test the latest upstream kernel available
> that would be great. It will allow additional upstream developers to
> examine the issue. Refer to
> https://wiki.ubuntu.com/KernelMainlineBuilds . Once you've tested the
> upstream kernel, please remove the 'needs-upstream-testing' tag. This
> can be done by clicking on the yellow pencil icon next to the tag
> located at the bottom of the bug description and deleting the
> 'needs-upstream-testing' text. Please let us know your results.
>
> Thanks in advance.
>
> [This is an automated message. Apologies if it has reached you
> inappropriately; please just reply to this message indicating so.]

Already running on one of those builds, as commented above (earlier).
So far, so good:

mbt@fennel:~$ uname -a ; uptime
Linux fennel 2.6.32-02063211-generic #02063211 SMP Mon Apr 12 12:03:56 UTC 2010 x86_64 GNU/Linux
 23:33:34 up 8:43, 4 users, load average: 0.19, 0.11, 0.03
mbt@fennel:~$

 --- Mike

--
Even if their crude and anticompetitive business practices don't make
you think about using their software, their use of sweatshops and child
labor should: boycott Microsoft like you would any other amoral child
abuser: http://is.gd/btW8m

Revision history for this message
Daniel Letzeisen (dtl131) wrote :

As I noted in the dupe bug, removing plymouth fixes the problem. Of course, that's not a good solution, but maybe it will provide a clue as to what's happening. If anyone wants to test that out, get mountall package from https://launchpad.net/~dtl131/+archive/mediahacks and then you should be able to remove plymouth packages (except libplymouth2, which is okay to keep).

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

Brian murray suggests his bug 564181 might be related to this; there's upstream discussion about it that might be worth review.

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

Michael, you're welcome to use whatever terminology makes you happy, but understand that the terminology 'freeze' and 'crash' have specific, different meanings with X.org upstream. The practicality of it to you is that each has a different method for analysis/troubleshooting so when you use the wrong terminology you may get the wrong advice and thus end up having lots of your time wasted. But it's your time so use what you feel best.

Revision history for this message
Michael B. Trausch (mtrausch) wrote :

On Fri, 2010-04-23 at 21:38 +0000, Bryce Harrington wrote:
> Michael, you're welcome to use whatever terminology makes you happy,
> but
> understand that the terminology 'freeze' and 'crash' have specific,
> different meanings with X.org upstream. The practicality of it to you
> is that each has a different method for analysis/troubleshooting so
> when
> you use the wrong terminology you may get the wrong advice and thus
> end
> up having lots of your time wasted. But it's your time so use what
> you
> feel best.

Are you suggesting that "crash" means something in the world of X than
it does in the rest of software development? I am confused here.
Doubly confused since the problem appears to be in the kernel and not
X.org.

I've been running a vanilla kernel for two days now and have not
experienced this problem, so it seems to be something in the Ubuntu
kernel causing the problem. I have not removed Plymouth or changed the
Ubuntu system in any way other than using the 2.6.32-02063211-generic
kernel that is packaged and available from the Ubuntu Kernel Team.

 --- Mike

--
Even if their crude and anticompetitive business practices don't make
you think about using their software, their use of sweatshops and child
labor should: boycott Microsoft like you would any other amoral child
abuser: http://is.gd/btW8m

Bryce Harrington (bryce)
tags: added: omit
removed: xorg-needs-kernel-fix
Revision history for this message
Michael B. Trausch (mtrausch) wrote :

I have not experienced this problem running upstream kernels since my last comment on April 23, so it seems to definitely be some issue with the Ubuntu kernel. Tomorrow, I will try to see if I can isolate a means to reliably trigger the problem as an aid to bisection.

Revision history for this message
Victor Pelt (victor-pelt) wrote :

hibernating (using tuxonice) or even resuming from hibernation seems to trigger this for me

Manoj Iyer (manjo)
tags: added: kernel-graphics kernel-reviewed
Revision history for this message
Andy Whitcroft (apw) wrote :

@Michael -- the Lucid kernel actually has the DRM sub-system from v2.6.33 backported into it. Therefore we would need to compare against v2.6.33.y stable releases to see if it is an ubuntu specific patch or just the v2.6.33 DRM which is at fault. Would you be able to test the latest v2.6.33.y kernel from the mainline kernel archive and report back here. Thanks.

Changed in linux (Ubuntu):
status: New → Incomplete
Changed in linux (Ubuntu Lucid):
status: New → Incomplete
Revision history for this message
owens87 (owens87) wrote :

Sounds like the same issue I'm having here with my Radeon Mobility 4500 series. The X window screen never comes back after hibernation, but is perfectly fine with resuming from suspend.

Jul 13 00:58:46 mike-laptop kernel: [ 4601.291348] [drm:radeon_fence_wait] *ERROR* fence(ffff88003700d640:0x00080E12) 7650ms timeout going to reset GPU
Jul 13 00:58:46 mike-laptop kernel: [ 4601.291360] radeon 0000:01:00.0: GPU softreset
Jul 13 00:58:46 mike-laptop kernel: [ 4601.291366] radeon 0000:01:00.0: R_008010_GRBM_STATUS=0xA27034A4
Jul 13 00:58:46 mike-laptop kernel: [ 4601.291373] radeon 0000:01:00.0: R_008014_GRBM_STATUS2=0x00000102
Jul 13 00:58:46 mike-laptop kernel: [ 4601.291379] radeon 0000:01:00.0: R_000E50_SRBM_STATUS=0x200000C0
Jul 13 00:58:46 mike-laptop kernel: [ 4601.291391] radeon 0000:01:00.0: R_008020_GRBM_SOFT_RESET=0x00007FEE
Jul 13 00:58:46 mike-laptop kernel: [ 4601.291449] radeon 0000:01:00.0: R_008020_GRBM_SOFT_RESET=0x00000001
Jul 13 00:58:46 mike-laptop kernel: [ 4601.291513] radeon 0000:01:00.0: R_000E60_SRBM_SOFT_RESET=0x00000402
Jul 13 00:58:46 mike-laptop kernel: [ 4601.315757] radeon 0000:01:00.0: R_008010_GRBM_STATUS=0xFFFFFFFF
Jul 13 00:58:46 mike-laptop kernel: [ 4601.315762] radeon 0000:01:00.0: R_008014_GRBM_STATUS2=0xFFFFFFFF
Jul 13 00:58:46 mike-laptop kernel: [ 4601.315768] radeon 0000:01:00.0: R_000E50_SRBM_STATUS=0xFFFFFFFF
Jul 13 00:58:46 mike-laptop kernel: [ 4601.330146] [drm:radeon_fence_wait] *ERROR* fence(ffff88003700d640:0x00080E12) 7690ms timeout
Jul 13 00:58:46 mike-laptop kernel: [ 4601.330149] [drm:radeon_fence_wait] *ERROR* last signaled fence(0x00080E12)
Jul 13 00:58:47 mike-laptop kernel: [ 4602.541342] [drm:radeon_fence_wait] *ERROR* fence(ffff88003736bc40:0x00080E17) 4150ms timeout going to reset GPU
Jul 13 00:58:47 mike-laptop kernel: [ 4602.541354] radeon 0000:01:00.0: GPU softreset
Jul 13 00:58:47 mike-laptop kernel: [ 4602.541360] radeon 0000:01:00.0: R_008010_GRBM_STATUS=0xA0003028
.....

Revision history for this message
owens87 (owens87) wrote :

2.6.32-23-generic #37-Ubuntu SMP Fri Jun 11 08:03:28 UTC 2010 x86_64 GNU/Linux

Andy Whitcroft (apw)
Changed in linux (Ubuntu):
assignee: nobody → Andy Whitcroft (apw)
Changed in linux (Ubuntu Lucid):
assignee: nobody → Andy Whitcroft (apw)
Revision history for this message
Thomas Antepoth (ta-ubuntu-antepoth) wrote :

I can perfectly reproduce here on a clean Lucid install this while surfing on Firefox onto this URL:

http://www.godmode-trader.de/DU-Pont-Nemours-Aktie/Kennzahlen

After five to ten seconds the screen goes black and the CRT finally enters powersave mode.

Syslog records:

Aug 23 19:26:03 sofa kernel: [ 258.572193] [drm:radeon_fence_wait] *ERROR* fence(ffff880216827300:0x0000265A) 580ms timeout
Aug 23 19:26:03 sofa kernel: [ 258.572199] [drm:radeon_fence_wait] *ERROR* last signaled fence(0x0000265A)
Aug 23 19:26:03 sofa kernel: [ 258.581431] [drm:radeon_fence_wait] *ERROR* fence(ffff8801d7db2500:0x0000265B) 570ms timeout going to reset GPU
Aug 23 19:26:03 sofa kernel: [ 258.581437] radeon 0000:01:00.0: GPU softreset
Aug 23 19:26:03 sofa kernel: [ 258.581440] radeon 0000:01:00.0: R_008010_GRBM_STATUS=0xE57004E0
Aug 23 19:26:03 sofa kernel: [ 258.581444] radeon 0000:01:00.0: R_008014_GRBM_STATUS2=0x00110103
Aug 23 19:26:03 sofa kernel: [ 258.581448] radeon 0000:01:00.0: R_000E50_SRBM_STATUS=0x200020C0
Aug 23 19:26:03 sofa kernel: [ 258.665251] radeon 0000:01:00.0: R_008020_GRBM_SOFT_RESET=0x00007FEE
Aug 23 19:26:03 sofa kernel: [ 258.665305] radeon 0000:01:00.0: R_008020_GRBM_SOFT_RESET=0x00000001
Aug 23 19:26:03 sofa kernel: [ 258.665366] radeon 0000:01:00.0: R_000E60_SRBM_SOFT_RESET=0x00000402
Aug 23 19:26:03 sofa kernel: [ 258.665522] radeon 0000:01:00.0: R_008010_GRBM_STATUS=0x00003030
Aug 23 19:26:03 sofa kernel: [ 258.665524] radeon 0000:01:00.0: R_008014_GRBM_STATUS2=0x00000003
Aug 23 19:26:03 sofa kernel: [ 258.665526] radeon 0000:01:00.0: R_000E50_SRBM_STATUS=0x200000C0
Aug 23 19:26:03 sofa kernel: [ 258.667977] [drm:radeon_fence_wait] *ERROR* fence(ffff8801d7db2500:0x0000265B) 660ms timeout

SYSRQ does not work either.

GPU is a RV610.

Revision history for this message
Thomas Antepoth (ta-ubuntu-antepoth) wrote :

Reproducable also here in this url with the GPU mentioned above when scrolling up and down a bit:

http://www.bilder-hochladen.net/files/fcjc-12-jpg.html

This URL loaded in Firefox resulted in a:

Aug 24 21:35:14 sofa kernel: [91459.900989] [drm:radeon_fence_wait] *ERROR* fence(ffff88009eadfa80:0x0025F614) 20ms timeout
Aug 24 21:35:14 sofa kernel: [91459.900993] [drm:radeon_fence_wait] *ERROR* last signaled fence(0x0025F614)
Aug 24 21:35:14 sofa kernel: [91460.451285] [drm:radeon_fence_wait] *ERROR* fence(ffff8800d050fa40:0x0025F617) 510ms timeout going to reset GPU

and a 950 minute rebuild of the md device here. No SysRQ again and immediately after the black screen all network connections have been dropped.

Revision history for this message
Michael B. Trausch (mtrausch) wrote :

Just a warning to everyone: comment #19 link is pop-uppy and stuff.

Thomas, can you in the future post warnings to that effect when you post on a bug report or in any other sort of forum where it is generally expected that Web links are clean and respectful? That does means disabling ad blocking and pop-up blocking stuff (at least down to the normal level of a off-the-shelf Web browser configuration) before checking the link, which is a downside if you normally run with all that sort of stuff in your browser. You can create a secondary Firefox profile that you could use for that purpose with relative ease. That way when those of us who use "just a browser" we don't have to try to click out of pop-ins and pop-outs and all sorts of other crud. Thanks!

Revision history for this message
David Henningsson (diwic) wrote :

So I'm coming from the duplicate, and just wanted to say, I have been running the maverick backport kernel (2.6.35-generic) from the kernel ppa for a week now and that *seems* to have solved it for me. It happened quite irregular though, so trying to do a git bisect would probably take a few months or so...

Anyway, Andy, if you want me to do something, you know where to find me.

Revision history for this message
Thomas Antepoth (ta-ubuntu-antepoth) wrote :

Thanks David.

Due to your hint using the ppa-Kernel I went off to install this one:

root@sofa:/var/log# uname -a
Linux sofa 2.6.35-18-generic #24~lucid1-Ubuntu SMP Mon Aug 23 04:49:33 UTC 2010 x86_64 GNU/Linux
root@sofa:/var/log#

Overall scrolling performance feels somewhat improved and the first link did not lock up my machine which it did quite reliable before installing the 2.6.35.

The second link provided did also not lock up the machine but while building up the page in the browser there were some interruptions in screen updates (e.g. clock seconds did not advance, mouse pointer froze for several seconds, rendering stopped amidst the picture).

In order to install that kernel I had to set:

root@sofa:/etc/default# grep nomodeset grub
GRUB_CMDLINE_LINUX="nomodeset"
root@sofa:/etc/default#

to avoid a EDID checksum error and to prevent the screen resolution being throttled to 1376x768

Revision history for this message
toad (toad-rubikon) wrote :

Hello David Henningsson,

Did you use the kernel from this ppa: http://www.ubuntuupdates.org/ppas/37 ?

Changed in linux (Ubuntu Lucid):
assignee: Andy Whitcroft (apw) → nobody
Changed in linux (Ubuntu):
assignee: Andy Whitcroft (apw) → nobody
Revision history for this message
Jeremy Foshee (jeremyfoshee) wrote :

This bug report was marked as Incomplete and has not had any updated comments for quite some time. As a result this bug is being closed. Please reopen if this is still an issue in the current Ubuntu development release http://cdimage.ubuntu.com/daily-live/current/ . Also, please be sure to provide any requested information that may have been missing. To reopen the bug, click on the current status under the Status column and change the status back to "New". Thanks.

[This is an automated message. Apologies if it has reached you inappropriately; please just reply to this message indicating so.]

tags: added: kj-expired
Changed in linux (Ubuntu):
status: Incomplete → Expired
Changed in linux (Ubuntu Lucid):
status: Incomplete → Invalid
To post a comment you must log in.
This report contains Public information  
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.