RC6 enabled causes severe graphics corruption

[Daniel T Chen]

Running current 12.04 (as of Sun Feb 19 01:15:24 EST 2012) and experiencing severe graphics corruption with RC6 enabled. Appending "i915.i915_enable_rc6=0" to the kernel command line upon reboot results in a usable Unity session.

Processor Information
 Socket Designation: CPU
 Type: Central Processor
 Family: Core i7
 Manufacturer: Intel(R) Corporation
 ID: A7 06 02 00 FF FB EB BF
 Signature: Type 0, Family 6, Model 42, Stepping 7

ProblemType: Bug
DistroRelease: Ubuntu 12.04
Package: linux-image-3.2.0-17-generic 3.2.0-17.26
ProcVersionSignature: Ubuntu 3.2.0-17.26-generic 3.2.6
Uname: Linux 3.2.0-17-generic x86_64
AlsaVersion: Advanced Linux Sound Architecture Driver Version 1.0.24.
ApportVersion: 1.91-0ubuntu1
Architecture: amd64
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC0: crimsun 2157 F.... pulseaudio
 /dev/snd/pcmC0D0p: crimsun 2157 F...m pulseaudio
CRDA: Error: [Errno 2] No such file or directory
Card0.Amixer.info:
 Card hw:0 'PCH'/'HDA Intel PCH at 0xc0800000 irq 55'
   Mixer name : 'Intel CougarPoint HDMI'
   Components : 'HDA:10ec0269,144dc0b3,00100202 HDA:80862805,80860101,00100000'
   Controls : 24
   Simple ctrls : 11
Date: Sun Feb 19 01:13:40 2012
EcryptfsInUse: Yes
HibernationDevice: RESUME=UUID=e98aa071-f1fc-44b7-afcc-4fbc8e1e1564
InstallationMedia: Ubuntu 12.04 LTS "Precise Pangolin" - Alpha amd64 (20120129)
MachineType: SAMSUNG ELECTRONICS CO., LTD. 700Z3A/700Z4A/700Z5A/700Z5B
ProcEnviron:
 PATH=(custom, no user)
 LANG=en_US.UTF-8
 SHELL=/bin/bash
ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-3.2.0-17-generic root=/dev/mapper/hostname-root ro quiet splash i915.i915_enable_rc6=0 vt.handoff=7
RelatedPackageVersions:
 linux-restricted-modules-3.2.0-17-generic N/A
 linux-backports-modules-3.2.0-17-generic N/A
 linux-firmware 1.69
RfKill:
 0: phy0: Wireless LAN
  Soft blocked: no
  Hard blocked: no
SourcePackage: linux
StagingDrivers: mei
UpgradeStatus: No upgrade log present (probably fresh install)
dmi.bios.date: 09/06/2011
dmi.bios.vendor: Phoenix Technologies Ltd.
dmi.bios.version: 03FD
dmi.board.asset.tag: Base Board Asset Tag
dmi.board.name: 700Z3A/700Z4A/700Z5A/700Z5B
dmi.board.vendor: SAMSUNG ELECTRONICS CO., LTD.
dmi.board.version: FAB1
dmi.chassis.asset.tag: Asset Tag
dmi.chassis.type: 9
dmi.chassis.vendor: SAMSUNG ELECTRONICS CO., LTD.
dmi.chassis.version: 0.1
dmi.modalias: dmi:bvnPhoenixTechnologiesLtd.:bvr03FD:bd09/06/2011:svnSAMSUNGELECTRONICSCO.,LTD.:pn700Z3A/700Z4A/700Z5A/700Z5B:pvr0.1:rvnSAMSUNGELECTRONICSCO.,LTD.:rn700Z3A/700Z4A/700Z5A/700Z5B:rvrFAB1:cvnSAMSUNGELECTRONICSCO.,LTD.:ct9:cvr0.1:
dmi.product.name: 700Z3A/700Z4A/700Z5A/700Z5B
dmi.product.version: 0.1
dmi.sys.vendor: SAMSUNG ELECTRONICS CO., LTD.

[Robert Hooker (sarvatt)]

Can you describe the corruption please? Can you take a picture of it?

[Daniel T Chen (crimsun)]

Most of the artifacts are portions of application windows that do not
refresh properly, e.g., they remain across all virtual desktops. With
slightly lesser frequency are triangular textures that consume a quarter to
a third of the upper or lower horizontal edge of the virtual desktop. With
very limited frequency are flashes or flickers of the entire current
virtual desktop when switching to another virtual desktop using the
keyboard shortcuts.

I'll attempt to snap a photo.

[Jorge G (geojorg)]

I did some testing for the last 3 days with Precise Kernel and freezes are finally out of the picture but there are very rare glitches in Chrome and Firefox, I can't reproduce them in other parts of the desktop which is really weird, because i could reproduce them in Oneiric with rc6 manually active on different apps.

This graphics curruption only appears in Web Browsers, so i don't know if is related to RC6, I post screenshot.

On the other hand my laptop is using 25% less energy and is not getting hot.

I would keep testing RC6 since the problems are really rare and less frequent.

[Leann Ogasawara (leannogasawara)]

Assigning myself to this bug so that we keep track of it. This is a direct result of us having enabled RC6 by default in the 3.2.0-17.26 kernel.

[noname2 (noname2-deactivatedaccount)]

are u using framebuffer compression or lvds downclock or any other special aprameters for the intel driver/kernel?

[Robert Hooker (sarvatt)]

Jorge, would you mind please filing a new bug with ubuntu-bug linux that we can duplicate to this one just so we can have your specific machine's hardware information for working out an eventual quirk for them?

[Jorge G (geojorg)]

Ok i just created a new bug report here https://bugs.launchpad.net/ubuntu/+source/linux/+bug/938833 with the information you need.

[Jorge G (geojorg)]

#7 I was using drm.vblankoffdelay=1 i just remove it from the etc/default/grub, update grub and start testing with no kernel parameters to see if I can reproduce the bug that seem to only appear in web browsers.

[Björn (lajvaren)]

Having the same problem i think, though its really minor. 3 or four times since i upgraded to 12.04 the day before yesterday and i think its just in firefox (though i've been using that most of the time so that can be the reason). One time it was just the X:s that you close the tabs with that were bad. It seems to be confined to a certain "texture", for example one time it was in the main browser area but followed the page up when i scrolled down and dissapeared when i scrolled back up.

[Robert Hooker (sarvatt)]

Björn: Does it go away with i915.i915_enable_rc6=0 added to your kernel command line? Can you also please file a new bug with ubuntu-bug linux and duplicate it to this one if it does go away with that? We really need the unique information from every machine with problems, at this point it looks like we're going to have to quirk RC6 off by default on samsung machines in the short term until a fix comes around. Jorge G: Thanks for that!

[Robert Hooker (sarvatt)]

Jorge, Daniel or Björn, apologies for throwing more testing at you, but would you mind giving this kernel a try to see if it fares any better for you on your laptops?

http://kernel.ubuntu.com/~sarvatt/rc6/

[Daniel T Chen (crimsun)]

This kernel seems much more usable with the default setting, i.e., *no*
disabling of RC6 specified on the kernel command line. I've tested with
Unity, Unity 2d, and cinnamon; no noticeable graphical glitches have been
visible.

[Robert Hooker (sarvatt)]

Daniel: Thank you so much for testing it out! Jorge and Björn: would you mind trying the new kernel out? It's going to be released as -18 soon and it'd be much appreciated, we noticed a problem with 3.2.0-17 that should account for your problems and be fixed in the new one but want to be sure.

[Jorge G (geojorg)]

I just install it, would test and give all the feedback you need.

[Leann Ogasawara (leannogasawara)]

Hi All,

We just uploaded a new ubuntu kernel, 3.2.0-17.27. It contains the fix Robet mentioned in comment #13. We believe this should resolve the issue you are seeing here. If you could please test and let us know your results, it would be greatly appreciated. Thanks.

https://launchpad.net/ubuntu/+source/linux/3.2.0-17.27

[Jorge G (geojorg)]

OK i have tested kernel from comment #13 all day and could not reproduce the graphical glitches neither in Chrome or Firefox or any other app in Unity 3D. I guess the new patch work as expected.

[a (tuomasjjrasanen)]

3.2.0-17.27 seems to fix rc6-related graphics problems also on my machine:

  Lenovo X220
  Version: 8DET54WW (1.24 )
  00:02.0 VGA compatible controller [0300]: Intel Corporation 2nd Generation Core Processor Family Integrated Graphics Controller [8086:0126] (rev 09)

With previous kernel, 3.2.0-17.26, I had occasional graphic glitches. Glitches appeared as unrefreshed screen areas (displaying content from windows below the top one). Now they are all gone. Power levels have remained at the same level (avg=6.67W, std=0.11) in idle mode.

Well done!

[Leann Ogasawara (leannogasawara)]

Thanks everyone for all of the testing and feedback. The current RC6 patches we are carrying sound most promising. For now I'm going to mark this bug Fix Released. We intend to carry these patches for the 12.04 release assuming we do not discover any other show stopping regressions with the RC6 patches. If you do notice any further graphics corruption issues due to RC6, please re-open this bug by setting the status back to New. Thanks.

[Björn (lajvaren)]

 Sorry for not answering, just wanted to let you know its all working as it should now.