[gm45] False GPU lockup render.IPEHR: 0x01820000

[Martin Albicker]

Wenn man den Externen Bildschrim über den VGA stecker anschließt, lassen sich manchmal nicht alle Modis einstellen. Ebenfalls kam es schon vor das anschließend das Bild nicht mehr auf den Laptop zurückzuholen ging.
Erst nach einem Neustart funktionierte es dann wieder.

Bei meiner Hardware handelt es sich um einen Samsung x360

grüße Martin

ProblemType: Crash
DistroRelease: Ubuntu 12.04
Package: xserver-xorg-video-intel 2:2.17.0-1ubuntu4
ProcVersionSignature: Ubuntu 3.2.0-23.36-generic 3.2.14
Uname: Linux 3.2.0-23-generic x86_64
.tmp.unity.support.test.0:

ApportVersion: 2.0.1-0ubuntu2
Architecture: amd64
Chipset: gm45
CompizPlugins: No value set for `/apps/compiz-1/general/screen0/options/active_plugins'
CompositorRunning: compiz
Date: Sun Apr 15 09:29:58 2012
DistUpgraded: Fresh install
DistroCodename: precise
DistroVariant: ubuntu
DkmsStatus:
 virtualbox, 4.1.12, 3.2.0-21-generic, x86_64: installed
 virtualbox, 4.1.12, 3.2.0-22-generic, x86_64: installed
 virtualbox, 4.1.12, 3.2.0-23-generic, x86_64: installed
DuplicateSignature: [gm45] GPU lockup render.IPEHR: 0x01820000 Ubuntu 12.04
ExecutablePath: /usr/share/apport/apport-gpu-error-intel.py
GraphicsCard:
 Intel Corporation Mobile 4 Series Chipset Integrated Graphics Controller [8086:2a42] (rev 07) (prog-if 00 [VGA controller])
   Subsystem: Samsung Electronics Co Ltd Device [144d:c03e]
   Subsystem: Samsung Electronics Co Ltd Device [144d:c03e]
InstallationMedia: Ubuntu 12.04 LTS "Precise Pangolin" - Alpha amd64 (20120321)
InterpreterPath: /usr/bin/python2.7
MachineType: SAMSUNG ELECTRONICS CO., LTD. X360
ProcCmdline: /usr/bin/python /usr/share/apport/apport-gpu-error-intel.py
ProcEnviron:

ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-3.2.0-23-generic root=UUID=816ba687-c7bc-472a-aae3-89afe5cd0d54 ro quiet splash acpi_osi=Linux acpi_backlight=vendor
RelatedPackageVersions:
 xserver-xorg 1:7.6+12ubuntu1
 libdrm2 2.4.32-1ubuntu1
 xserver-xorg-video-intel 2:2.17.0-1ubuntu4
SourcePackage: xserver-xorg-video-intel
Title: [gm45] False GPU lockup render.IPEHR: 0x01820000
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups:

dmi.bios.date: 08/05/2010
dmi.bios.vendor: Phoenix Technologies Ltd.
dmi.bios.version: 10LM.M014.20100805.KSY
dmi.board.name: X360
dmi.board.vendor: SAMSUNG ELECTRONICS CO., LTD.
dmi.board.version: Not Applicable
dmi.chassis.asset.tag: No Asset Tag
dmi.chassis.type: 10
dmi.chassis.vendor: SAMSUNG ELECTRONICS CO., LTD.
dmi.chassis.version: N/A
dmi.modalias: dmi:bvnPhoenixTechnologiesLtd.:bvr10LM.M014.20100805.KSY:bd08/05/2010:svnSAMSUNGELECTRONICSCO.,LTD.:pnX360:pvrNotApplicable:rvnSAMSUNGELECTRONICSCO.,LTD.:rnX360:rvrNotApplicable:cvnSAMSUNGELECTRONICSCO.,LTD.:ct10:cvrN/A:
dmi.product.name: X360
dmi.product.version: Not Applicable
dmi.sys.vendor: SAMSUNG ELECTRONICS CO., LTD.
version.compiz: compiz 1:0.9.7.4-0ubuntu3
version.ia32-libs: ia32-libs N/A
version.libdrm2: libdrm2 2.4.32-1ubuntu1
version.libgl1-mesa-dri: libgl1-mesa-dri 8.0.2-0ubuntu3
version.libgl1-mesa-dri-experimental: libgl1-mesa-dri-experimental N/A
version.libgl1-mesa-glx: libgl1-mesa-glx 8.0.2-0ubuntu3
version.xserver-xorg-core: xserver-xorg-core 2:1.11.4-0ubuntu10
version.xserver-xorg-input-evdev: xserver-xorg-input-evdev 1:2.7.0-0ubuntu1
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 1:0.0.16+git20111201+b5534a1-1build2

[Bryce Harrington (bryce)]

Possibly same as bug #768184 , although error codes differ.

My German is rusty, mind rewriting description in english?

[Martin Albicker (19martin89)]

Hello,

I hope my description in english is good enought, that you can understood it.

if i conect an external monitor, it is sometimes impossible to select all modis (clone, dualview, only laptop, only external monitor).
And after i reconect the monitor, the default resolution of the laptop monitor, dont come back. Sometimes i get a resoltion of 4:3 instead of 16:10, and sometimes the screen persists black. So i musst restart the laptop, that the default resolution come back.

An other Bug that don't associate with this bug is, that the mousepad freezes sometimes. To solve this i musst restart my laptop.

My used Hardware is a Samsung x360 laptop.

[Bryce Harrington (bryce)]

Forwarding this bug from Ubuntu reporter Martin Albicker:
http://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/982251

[Problem]
Black screen lockup, possibly correlated to external monitor hot plug and/or RANDR.

[Original Description]
I hope my description in english is good enought, that you can understood it.

if i conect an external monitor, it is sometimes impossible to select all modis (clone, dualview, only laptop, only external monitor).
And after i reconect the monitor, the default resolution of the laptop monitor, dont come back. Sometimes i get a resoltion of 4:3 instead of 16:10, and sometimes the screen persists black. So i musst restart the laptop, that the default resolution come back.

An other Bug that don't associate with this bug is, that the mousepad freezes sometimes. To solve this i musst restart my laptop.

My used Hardware is a Samsung x360 laptop.

Wenn man den Externen Bildschrim ber den VGA stecker anschliet, lassen sich manchmal nicht alle Modis einstellen. Ebenfalls kam es schon vor das anschlieend das Bild nicht mehr auf den Laptop zurckzuholen ging.
Erst nach einem Neustart funktionierte es dann wieder.

Bei meiner Hardware handelt es sich um einen Samsung x360

gre Martin

ProblemType: Crash
DistroRelease: Ubuntu 12.04
Package: xserver-xorg-video-intel 2:2.17.0-1ubuntu4
ProcVersionSignature: Ubuntu 3.2.0-23.36-generic 3.2.14
Uname: Linux 3.2.0-23-generic x86_64
.tmp.unity.support.test.0:

ApportVersion: 2.0.1-0ubuntu2
Architecture: amd64
Chipset: gm45
CompizPlugins: No value set for `/apps/compiz-1/general/screen0/options/active_plugins'
CompositorRunning: compiz
Date: Sun Apr 15 09:29:58 2012
DistUpgraded: Fresh install
DistroCodename: precise
DistroVariant: ubuntu
DkmsStatus:
 virtualbox, 4.1.12, 3.2.0-21-generic, x86_64: installed
 virtualbox, 4.1.12, 3.2.0-22-generic, x86_64: installed
 virtualbox, 4.1.12, 3.2.0-23-generic, x86_64: installed
DuplicateSignature: [gm45] GPU lockup render.IPEHR: 0x01820000 Ubuntu 12.04
ExecutablePath: /usr/share/apport/apport-gpu-error-intel.py
GraphicsCard:
 Intel Corporation Mobile 4 Series Chipset Integrated Graphics Controller [8086:2a42] (rev 07) (prog-if 00 [VGA controller])
   Subsystem: Samsung Electronics Co Ltd Device [144d:c03e]
   Subsystem: Samsung Electronics Co Ltd Device [144d:c03e]
InstallationMedia: Ubuntu 12.04 LTS "Precise Pangolin" - Alpha amd64 (20120321)
InterpreterPath: /usr/bin/python2.7
MachineType: SAMSUNG ELECTRONICS CO., LTD. X360
ProcCmdline: /usr/bin/python /usr/share/apport/apport-gpu-error-intel.py
ProcEnviron:

ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-3.2.0-23-generic root=UUID=816ba687-c7bc-472a-aae3-89afe5cd0d54 ro quiet splash acpi_osi=Linux acpi_backlight=vendor
RelatedPackageVersions:
 xserver-xorg 1:7.6+12ubuntu1
 libdrm2 2.4.32-1ubuntu1
 xserver-xorg-video-intel 2:2.17.0-1ubuntu4
SourcePackage: xserver-xorg-video-intel
Title: [gm45] False GPU lockup render.IPEHR: 0x01820000
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups:

dmi.bios.date: 08/05/2010
dmi.bios.vendor: Phoenix Technologies Ltd.
dmi.bios.version: 10LM.M014.20100805.KSY
dmi.board.name: X360
dm...

[Bryce Harrington (bryce)]

Created attachment 60206
BootDmesg.txt

[Bryce Harrington (bryce)]

Created attachment 60207
CurrentDmesg.txt

[Bryce Harrington (bryce)]

Created attachment 60208
i915_error_state.txt

[Bryce Harrington (bryce)]

Created attachment 60209
XorgLog.txt

[Chris Wilson (ickle)]

commit 14667a4bde4361b7ac420d68a2e9e9b9b2df5231
Author: Chris Wilson <email address hidden>
Date: Tue Apr 3 17:58:35 2012 +0100

    drm/i915: Finish any pending operations on the framebuffer before disabling

    Similar to the case where we are changing from one framebuffer to
    another, we need to be sure that there are no pending WAIT_FOR_EVENTs on
    the pipe for the current framebuffer before switching. If we disable the
    pipe, and then try to execute a WAIT_FOR_EVENT it will block
    indefinitely and cause a GPU hang.

And a few patches to the DDX to shrink the race.

[Bryce Harrington (bryce)]

Martin Albicker - I've forwarded this bug upstream to https://bugs.freedesktop.org/show_bug.cgi?id=48838 - please subscribe yourself to this bug, in case they need further information or wish you to test something. Thanks ahead of time!

[Bryce Harrington (bryce)]

Which DDX patches?

[Bryce Harrington (bryce)]

@kernel team, upstream suggests this patch:

commit 14667a4bde4361b7ac420d68a2e9e9b9b2df5231
Author: Chris Wilson <email address hidden>
Date: Tue Apr 3 17:58:35 2012 +0100

    drm/i915: Finish any pending operations on the framebuffer before disabling

    Similar to the case where we are changing from one framebuffer to
    another, we need to be sure that there are no pending WAIT_FOR_EVENTs on
    the pipe for the current framebuffer before switching. If we disable the
    pipe, and then try to execute a WAIT_FOR_EVENT it will block
    indefinitely and cause a GPU hang.

[Bryce Harrington (bryce)]

Same patch was flagged for bug #982415 too

[Chris Wilson (ickle)]

In this particular case, just the kernel patch should be sufficient.

In the general case, the ddx needs:

commit b817200371bfe16f44b879a793cf4a75ad17bc5c
Author: Chris Wilson <email address hidden>
Date: Tue Apr 17 17:54:58 2012 +0100

    Don't issue a scanline wait while VT switched

and

commit cc20c45aa0ca15720510668d6918bf3c99104626
Author: Chris Wilson <email address hidden>
Date: Fri Mar 30 22:51:21 2012 +0100

    sna: Minimise the risk of hotplug hangs by checking fb before vsync

(which is just papering over the bugs).

[Bryce Harrington (bryce)]

Thanks; we don't ship/support sna but will look into the other ddx patch.

[Bryce Harrington (bryce)]

Upstream suggests a DDX patch that would be worth including as well.

"""
In this particular case, just the kernel patch should be sufficient.

In the general case, the ddx needs:

commit b817200371bfe16f44b879a793cf4a75ad17bc5c
Author: Chris Wilson <email address hidden>
Date: Tue Apr 17 17:54:58 2012 +0100

    Don't issue a scanline wait while VT switched

and

commit cc20c45aa0ca15720510668d6918bf3c99104626
Author: Chris Wilson <email address hidden>
Date: Fri Mar 30 22:51:21 2012 +0100

    sna: Minimise the risk of hotplug hangs by checking fb before vsync

(which is just papering over the bugs).
"""

I think we don't need the sna patch since we don't ship with that on yet.

[Leann Ogasawara (leannogasawara)]

Hi Martin,

I've build a test kernel with the patch Bryce noted from comment #5. Could you please test and let us know your results? Thanks in advance.

http://people.canonical.com/~ogasawara/lp982251/amd64/

[Chris Wilson (ickle)]

And the last patch landed in drm-intel-next-queued:

commit 0f91128d88bbb8b0a8e7bb93df2c40680871d45a
Author: Chris Wilson <email address hidden>
Date: Tue Apr 17 10:05:38 2012 +0100

    drm/i915: Wait for all pending operations to the fb before disabling the pip

    During modeset we have to disable the pipe to reconfigure its timings
    and maybe its size. Userspace may have queued up command buffers that
    depend upon the pipe running in a certain configuration and so the
    commands may become confused across the modeset. At the moment, we use a
    less than satisfactory kick-scanline-waits should the GPU hang during
    the modeset. It should be more reliable to wait for the pending
    operations to complete first, even though we still have a window for
    userspace to submit a broken command buffer during the modeset.

    Signed-off-by: Chris Wilson <email address hidden>
    Signed-off-by: Daniel Vetter <email address hidden>

[Sergio Callegari (callegar)]

This is biting me too...
As I could not find this bug by title, I reported it on the intel drivers bug tracker on freedesktop and Chris Wilson's answer (in closing the bug) pointed me here. He indicated that 3.4 kernels already have the fix.

Can anyone indicate if this is getting backported to precise kernel?
It is important for having a machine in shape for giving presentations.

[Timo Aaltonen (tjaalton)]

I'll send the missing upstream kernel commit to the stable maintainer, precise will get it in the end.

[Timo Aaltonen (tjaalton)]

closing it for quantal, as it should be closed there.

[Timo Aaltonen (tjaalton)]

included in 3.2.0-31.50