X.org crashes sometimes. [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!

Hi gordev,

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.]

Seems no error on 2.6.34-020634-generic #020634 SMP Mon May 17 20:34:55 UTC 2010 i686 GNU/Linux kernel.

I have the same problem.
wouter@wouter-desktop:~$ uname -a
Linux wouter-desktop 2.6.32-22-generic #35-Ubuntu SMP Tue Jun 1 14:17:36 UTC 2010 i686 GNU/Linux

For what it's worth - here are the messages from my syslog just as it happened:

Jun 3 17:10:11 wouter-desktop kernel: [ 2093.042961] ACPI Error: Illegal I/O port address/length above 64K: 0x00400020/4 (20090903/hwvalid-154)
Jun 3 17:10:11 wouter-desktop kernel: [ 2093.042983] ACPI Exception: AE_LIMIT, Returned by Handler for [SystemIO] (20090903/evregion-424)
Jun 3 17:10:11 wouter-desktop kernel: [ 2093.042999] ACPI Error (psparse-0537): Method parse/execution failed [\_GPE._L09] (Node f7010d98), AE_LIMIT
Jun 3 17:10:11 wouter-desktop kernel: [ 2093.043064] ACPI Exception: AE_LIMIT, while evaluating GPE method [_L09] (20090903/evgpe-568)
Jun 3 17:10:11 wouter-desktop kernel: [ 2093.055943] ACPI Error: Illegal I/O port address/length above 64K: 0x00400020/4 (20090903/hwvalid-154)
Jun 3 17:10:11 wouter-desktop kernel: [ 2093.055960] ACPI Exception: AE_LIMIT, Returned by Handler for [SystemIO] (20090903/evregion-424)
Jun 3 17:10:11 wouter-desktop kernel: [ 2093.055976] ACPI Error (psparse-0537): Method parse/execution failed [\_GPE._L09] (Node f7010d98), AE_LIMIT
Jun 3 17:10:11 wouter-desktop kernel: [ 2093.056490] ACPI Exception: AE_LIMIT, while evaluating GPE method [_L09] (20090903/evgpe-568)
Jun 3 17:10:19 wouter-desktop kernel: [ 2101.048904] ACPI Error: Illegal I/O port address/length above 64K: 0x00400020/4 (20090903/hwvalid-154)
Jun 3 17:10:19 wouter-desktop kernel: [ 2101.048926] ACPI Exception: AE_LIMIT, Returned by Handler for [SystemIO] (20090903/evregion-424)
Jun 3 17:10:19 wouter-desktop kernel: [ 2101.048941] ACPI Error (psparse-0537): Method parse/execution failed [\_GPE._L09] (Node f7010d98), AE_LIMIT
Jun 3 17:10:19 wouter-desktop kernel: [ 2101.049005] ACPI Exception: AE_LIMIT, while evaluating GPE method [_L09] (20090903/evgpe-568)
Jun 3 17:10:20 wouter-desktop kernel: [ 2102.558143] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!

I am seeing this as well with 2.6.32-22-generic

This one is associated with a crash ( X server for display :0 terminated unexpectedly):

Jun 3 08:17:01 ns1 CRON[3291]: (root) CMD ( cd / && run-parts --report /etc/cron.hourly)
Jun 3 08:25:02 ns1 CRON[3348]: (root) CMD (command -v debian-sa1 > /dev/null && debian-sa1 1 1)
Jun 3 08:31:48 ns1 kernel: [ 4773.852874] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!
Jun 3 08:31:51 ns1 kdm[771]: X server for display :0 terminated unexpectedly
Jun 3 08:31:56 ns1 acpid: client connected from 3406[0:0]
Jun 3 08:31:56 ns1 acpid: 1 client rule loaded
Jun 3 08:31:56 ns1 kernel: [ 4782.416110] [drm:drm_mode_getfb] *ERROR* invalid framebuffer id
Jun 3 08:31:58 ns1 kdm_greet[3408]: Cannot load /usr/share/kde4/apps/kdm/faces/.default.face: No such file or directory
Jun 3 08:32:09 ns1 anacron[3565]: Anacron 2.3 started on 2010-06-03
Jun 3 08:32:09 ns1 anacron[3565]: Normal exit (0 jobs run)

These did not crash me:

Jun 3 11:09:01 ns1 CRON[11207]: (root) CMD ( [ -x /usr/lib/php5/maxlifetime ] && [ -d /var/lib/php5 ] && find /var/lib/php5/ -type f -cmin +$(/usr/lib/php5/maxlifetime) -print0 | xargs -n 200 -r -0 rm)
Jun 3 11:12:46 ns1 kernel: [14432.189115] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!
Jun 3 11:12:51 ns1 kernel: [14436.823590] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!
Jun 3 11:12:51 ns1 kernel: [14437.105722] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!
Jun 3 11:15:01 ns1 CRON[11263]: (root) CMD (command -v debian-sa1 > /dev/null && debian-sa1 1 1)
Jun 3 11:15:33 ns1 kernel: [14598.907186] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!
Jun 3 11:15:39 ns1 kernel: [14605.553271] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!
Jun 3 11:17:01 ns1 CRON[11272]: (root) CMD ( cd / && run-parts --report /etc/cron.hourly)
Jun 3 11:25:01 ns1 CRON[11374]: (root) CMD (command -v debian-sa1 > /dev/null && debian-sa1 1 1)

This also happens on Kubuntu, not using firefox. It can happen with different applications showing graphics, such as Akregator, Konqueror, etc. I found that you can actually predict when X is going to crash and bring you straight to login screen, because characters within the window start to be replaced by senseless symbols; or images within the frame of the window turn into thick coloured horizontal stripes. Whenever I see that, I immediately close the window in question, and go check the System log, which invariably shows such messages as:

2010-06-04 14:58:08 EGO kernel [20138.652499] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!
2010-06-04 14:58:08 EGO kernel [20139.291065] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!
2010-06-04 14:58:09 EGO kernel [20139.797174] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!

I found that this prevents crashes, and interruption to normal workflow. Still, it's merely a temporary workaround that doesn't solve the issue.

The crash strikes again:

Jun 8 07:55:16 ns1 anacron[858]: Normal exit (1 job run)
Jun 8 08:05:01 ns1 CRON[3278]: (root) CMD (command -v debian-sa1 > /dev/null && debian-sa1 1 1)
Jun 8 08:06:42 ns1 kernel: [ 2716.581275] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!
Jun 8 08:06:43 ns1 kdm[755]: X server for display :0 terminated unexpectedly
Jun 8 08:06:48 ns1 acpid: client connected from 3321[0:0]
Jun 8 08:06:48 ns1 acpid: 1 client rule loaded
Jun 8 08:06:48 ns1 kernel: [ 2722.716857] [drm:drm_mode_getfb] *ERROR* invalid framebuffer id
Jun 8 08:06:50 ns1 kdm_greet[3323]: Cannot load /usr/share/kde4/apps/kdm/faces/.default.face: No such file or directory
Jun 8 08:06:59 ns1 anacron[3480]: Anacron 2.3 started on 2010-06-08
Jun 8 08:06:59 ns1 anacron[3480]: Normal exit (0 jobs run)
Jun 8 08:09:01 ns1 CRON[4069]: (root) CMD ( [ -x /usr/lib/php5/maxlifetime ] && [ -d /var/lib/php5 ] && find /var/lib/php5/ -type f -cmin +$(/usr/lib/php5/maxlifetime) -print0 | xargs -n 200 -r -0 rm)

And like lelamal, I see glyph issues as well. I most likely will revert to 8.03, because I cannot afford to keep crashing.

For me this crash seems to correlate when I switch between desktops/applications. For example, I was using command line apt-get install - started the download, clicked on the firefox app/desktop - and crash..... And when I think back - it seems always to be when I switch applications - instead of getting to the selected app - I crash and must login again. Had 3 crashes this AM already.

I can reproduce the "kernel: [ 6229.694385] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!" entry in the syslog.

I am tailing the syslog and when I switch from viewing the tail in a console to view Firefox using the tabs at the bottom, then this error appears in the tail most every time. Switching between other applications does not generate this error. It does not always crash though, but still always appears before the crash:

Jun 18 08:15:16 ns1 kernel: [ 5867.293755] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!
Jun 18 08:15:18 ns1 kdm[737]: X server for display :0 terminated unexpectedly
Jun 18 08:15:23 ns1 acpid: client 834[0:0] has disconnected
Jun 18 08:15:23 ns1 acpid: client connected from 3393[0:0]
Jun 18 08:15:23 ns1 acpid: 1 client rule loaded
Jun 18 08:15:24 ns1 kernel: [ 5874.786774] [drm:drm_mode_getfb] *ERROR* invalid framebuffer id
Jun 18 08:15:25 ns1 kdm_greet[3395]: Cannot load /usr/share/kde4/apps/kdm/faces/.default.face: No such file or directory
Jun 18 08:15:33 ns1 anacron[3554]: Anacron 2.3 started on 2010-06-18
Jun 18 08:15:33 ns1 anacron[3554]: Normal exit (0 jobs run)
Jun 18 08:17:01 ns1 CRON[4183]: (root) CMD ( cd / && run-parts --report /etc/cron.hourly)

@John Gilmore: I'm not entirely convinced this bug is a duplicate of bug #550850. The latter seems to be specific to Compiz, as the title itself entails, and users reportedly refer to Compiz, including you, while the present bug report also includes Kubuntu, which relies on Kwin as a window manager.

As I understand from the original report and various comments, including yours, in bug #550850 it's just Compiz to exit, but X is not crashing as reported in this bug. You report that the "window manager goes away", leaving you with undecorated windows, and that you're still able to log out and then back in to temporarily solve the issue. Also, you add that when it happens, you "can usually type into a terminal and/or a web browser, so there's lots that could be done". Conversely, this bug reports X crashing and taking users straight to the login screen (imagine hitting Ctrl+Alt+Backspace in the old days, to get the effect), and when that happens there's not much we can actually do apart from logging back in.

I'm not entirely convinced either. But I find it suspicious that various processes are crashing with the same kernel error message, coming from the same DRM driver in the kernel. Whichever process gets the error is the one that crashes, so when it's compiz, you have windows but no window manager; when it's X, the whole window system dies.

The fact that a variety of programs are crashing -- all talking to one driver -- leads me to suspect that the bug is in the driver, not in the programs. But only time and further investigation will tell.

I used to crash multiple times daily. Now I do not. I believe the system updates may have fixed this problem for me..... I still see a few "Failed to parse relocation -12!", but not nearly as many and again - not with a crash.

I have also the same problem.
I have a ATI Radeon 9000 mobility
Linux mtst010 2.6.35-6-generic #9~lucid1-Ubuntu SMP Wed Jun 30 11:34:29 UTC 2010 i686 GNU/Linux
X.Org X Server 1.7.6
KDE 4.4.5

kdm.log says:
drmRadeonCmdBuffer: -12. Kernel failed to parse or rejected command stream. See dmesg for more info.

dmesg:
[ 2128.536437] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!

syslog:
Jul 8 09:55:33 mtst010 anacron[866]: Job `cron.weekly' terminated
Jul 8 09:55:33 mtst010 anacron[866]: Normal exit (2 jobs run)
Jul 8 10:06:03 mtst010 kernel: [ 2128.536437] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!
Jul 8 10:06:04 mtst010 NetworkManager: <info> (eth1): device state change: 8 -> 3 (reason 38)
Jul 8 10:06:04 mtst010 NetworkManager: <info> (eth1): deactivating device (reason: 38).
Jul 8 10:06:04 mtst010 kdm[736]: X server for display :0 terminated unexpectedly

This happens very often. System is unuseable. But I can't say how to reproduce. But I think it happens when I change the focus to an other window or maximise a window

@madbiologist
I am sorry, but where exactly was the fix committed? What package is this concerning? I am still experiencing this problem. Where can I get said fix?

Still a problem for me in Maverick alpha 3

@madbiologist: no fix was committed, the issue still affects users - setting it back to "triaged".

It seems that option is not available to me, marking it as "confirmed".

I have tried using latest kernel 2.6.35-17-generic
from sudo add-apt-repository ppa:kernel-ppa/pre-proposed

but still get this problem

[drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!

its easily reproduced on my machine by changing between windows after boot up
and results in the windows borders disappearing. As a way to keep working I have added
an icon to my menu that runs "metacity --replace", which then puts the window decoration
back in place and I can continue (usually without further crashes)

I am more than willing to try a new kernel patch for this is one is made available.

Can someone explain if this is a problem with the video driver in kernel or the way metacity is using the graphics.
Regards,
Luke

This is not a duclicate of the compiz bug (https://bugs.launchpad.net/ubuntu/+source/linux/+bug/550850).
I have the same problem on KDE4 (Kubuntu Lucid) and I don't have compiz installed.

I think its related to desktop effects. I have disabled them and the error message disappears and no crash anymore. but no effects too :(

I've just had the same thing on Maverick rc, under KDE with desktop effects; it's been running fine all day, but I just seemed to have triggered it repeatedly by zooming out in Google maps running in Chromium-browser.

Radeon HD4350/RV710 on the open driver.

Dave

Two things; one I don't get the same problem running on Gnome without desktop effects rather than KDE with desktop effects; haven't tried the two in between.

-12 is -ENOMEM and I've traced it back a little:

Oct 4 13:11:27 major kernel: [ 795.810631] [drm:ttm_bo_validate] *ERROR* ttm_bo_validate: ttm_bo_move_buffer gave ENOMEM
Oct 4 13:11:27 major kernel: [ 795.810634] [drm:radeon_bo_list_validate] *ERROR* radeon_bo_list_validate: ttm_bo_validate gave NOMEM
Oct 4 13:11:27 major kernel: [ 795.810636] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!

Hi - I'm new to this, so apologies if I'm being daft. I get this same result when trying to start up Vegastrike using a an ATI 9 series card, thus:

. . . blah, blah, blah

Loading a starsystem
Loading Star System Crucible/Cephid_17
 Next To: Crucible/17-ar
 Next To: Crucible/Cardell
 Next To: Crucible/Cephid_17
 Next To: Crucible/Enyo
 Next To: Crucible/Cephid_17
 Next To: Crucible/OldzieyLoading active missions True
drmRadeonCmdBuffer: -12. Kernel failed to parse or rejected command stream. See dmesg for more info.
AL lib: ALc.c:1253: exit() 1 device(s) and 1 context(s) NOT deleted
AL lib: alSource.c:2291: alcDestroyContext(): 32 Source(s) NOT deleted
AL lib: alBuffer.c:1097: exit() 16 Buffer(s) NOT deleted

dmesg:

after a fair bit of blah blah blah,

[ 6.744269] fb0: radeondrmfb frame buffer device
[ 6.744274] registered panic notifier
[ 6.744284] [drm] Initialized radeon 2.0.0 20080528 for 0000:01:00.0 on minor 0
[ 6.851022] vga16fb: initializing
[ 6.851027] vga16fb: mapped to 0xc00a0000
[ 6.851033] vga16fb: not registering due to another framebuffer present
[ 7.690852] alloc irq_desc for 17 on node -1
[ 7.690856] alloc kstat_irqs on node -1
[ 7.690866] Intel ICH 0000:00:1f.5: PCI INT B -> GSI 17 (level, low) -> IRQ 17
[ 7.690916] Intel ICH 0000:00:1f.5: setting latency timer to 64
[ 7.714135] Console: switching to colour frame buffer device 128x48
[ 8.028044] intel8x0_measure_ac97_clock: measured 54423 usecs (2622 samples)
[ 8.028049] intel8x0: clocking to 48000
[ 8.784948] type=1505 audit(1286356674.706:5): operation="profile_load" pid=644 name="/usr/share/gdm/guest-session/Xsession"
[ 8.789271] type=1505 audit(1286356674.710:6): operation="profile_replace" pid=647 name="/sbin/dhclient3"
[ 8.789844] type=1505 audit(1286356674.710:7): operation="profile_replace" pid=647 name="/usr/lib/NetworkManager/nm-dhcp-client.action"
[ 8.790161] type=1505 audit(1286356674.710:8): operation="profile_replace" pid=647 name="/usr/lib/connman/scripts/dhclient-script"
[ 9.882622] type=1505 audit(1286356675.802:9): operation="profile_load" pid=649 name="/usr/bin/evince"
[ 9.890537] type=1505 audit(1286356675.810:10): operation="profile_load" pid=649 name="/usr/bin/evince-previewer"
[ 9.895433] type=1505 audit(1286356675.814:11): operation="profile_load" pid=649 name="/usr/bin/evince-thumbnailer"
[ 10.006663] type=1505 audit(1286356675.926:12): operation="profile_load" pid=654 name="/usr/lib/cups/backend/cups-pdf"
[ 10.007321] type=1505 audit(1286356675.926:13): operation="profile_load" pid=654 name="/usr/sbin/cupsd"
[ 10.024698] type=1505 audit(1286356675.946:14): operation="profile_load" pid=655 name="/usr/sbin/tcpdump"
[ 10.816940] eth0: link up, 100Mbps, full-duplex, lpa 0x45E1
[ 16.650169] i2c i2c-0: sendbytes: NAK bailout.
[ 21.204025] eth0: no IPv6 routers present
[ 5373.646201] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!
[ 5527.210871] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!
[ 6928.308771] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!
[11409.283978] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12...

Got it again in Maverik :(
When i was on 2.6.34 kernel everything was ok. But now it started again.

Has anyone tested this with the closed source driver?
I will test it when i have time. Without Desktop Effects I have no crashes

I think it is related to this post:
http://blog.martin-graesslin.com/blog/2010/09/demystifying-opengl-desktop-effects/

He said closed source driver should work fine

I have got the same problem on ubuntu natty since the last days after recent updates

I get this errors in latest natty update too. I don't know if we have to investigate further.
It usually happens when compiz breaks (it removes the Window decorations) and looks like it try to redraw the windows from outside the drawing area.
My graphic card is an ATI Radeon HD 4870. Hope it helps

I've been having this problem for almost a year, since 10.04 and including 10.10. Since the bug has yet to be fixed, does anyone have a workaround? I use my Ubuntu computer for work and it crashes on me almost daily.

turn off desktop effects. I have had no crashes since i turn it off

I hadn't had this for a while; just triggered it in KDE - started assault cube.

after a stream of Failed to parse relocation -12!'s in the log there is a

[drm:radeon_gem_object_create] *ERROR* Failed to allocate GEM object (-883785728, 1, 4096, -12)

This is with last weeks daily DRM-next tree build; 2.6.38-996-generic #201103041138.

Running in kwin with effects. I just started assaultcube and it did that, only the starting logo came up for ac; it's dropped me back into my kde session but there are lots of random redraw errors and flashing etc now.

Dave

Still an issue , with screen corruption, was watching a youtube video

I add my info as attachments

gordev, this bug was reported a while ago and there hasn't been any activity in it recently. We were wondering if this is still an issue? Can you try with the latest development release of Ubuntu? ISO CD images are available from http://cdimage.ubuntu.com/releases/ .

If it remains an issue, could you run the following command in the development release from a Terminal (Applications->Accessories->Terminal). It will automatically gather and attach updated debug information to this report.

apport-collect -p linux <replace-with-bug-number>

Also, if you could 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 . Please do not test the kernel in the daily folder, but the one all the way at the bottom. 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. As well, please comment on which kernel version specifically you tested.

If this bug is fixed in the mainline kernel, please add the following tag 'kernel-fixed-upstream'.

If the mainline kernel does not fix this bug, please add the tag: 'kernel-bug-exists-upstream'.

If you are unable to test the mainline kernel, for example it will not boot, please add the tag: 'kernel-unable-to-test-upstream', and comment as to why specifically you were unable to test it.

Please let us know your results. Thanks in advance.