Mouse often hangs after upgrade to 35-2, 35-4 kernels

[Jonathan Thomas]

After the upgrade from 35-1 to 35-2 and 35-4, the mouse often hangs momemtarily when I move it across the screen. I also see that kslowd000, kslowd001 and kslowd002 are showing up in top as taking cpu time in accordance with the mouse freezes. It seems that the faster I move the mouse, the more frequent the hangs are, and these kslowd00* processes are, if not the highest consumer of CPU, are directly behind Xorg's process.

ProblemType: Bug
DistroRelease: Ubuntu 10.10
Package: linux-image-2.6.35-4-generic 2.6.35-4.5
Regression: Yes
Reproducible: Yes
ProcVersionSignature: Ubuntu 2.6.35-4.5-generic 2.6.35-rc3
Uname: Linux 2.6.35-4-generic x86_64
AlsaVersion: Advanced Linux Sound Architecture Driver Version 1.0.23.
AplayDevices:
 **** List of PLAYBACK Hardware Devices ****
 card 0: Intel [HDA Intel], device 0: VT1708S Analog [VT1708S Analog]
   Subdevices: 1/2
   Subdevice #0: subdevice #0
   Subdevice #1: subdevice #1
Architecture: amd64
ArecordDevices:
 **** List of CAPTURE Hardware Devices ****
 card 0: Intel [HDA Intel], device 0: VT1708S Analog [VT1708S Analog]
   Subdevices: 2/2
   Subdevice #0: subdevice #0
   Subdevice #1: subdevice #1
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC0: jonathan 1567 F.... kmix
 /dev/snd/pcmC0D0p: jonathan 1453 F...m knotify4
 /dev/snd/timer: jonathan 1453 f.... knotify4
CRDA: Error: [Errno 2] No such file or directory
Card0.Amixer.info:
 Card hw:0 'Intel'/'HDA Intel at 0xfe9f4000 irq 45'
   Mixer name : 'VIA VT1708S'
   Components : 'HDA:11060397,10431111,00100000'
   Controls : 21
   Simple ctrls : 14
Date: Thu Jun 17 22:42:58 2010
HibernationDevice: RESUME=UUID=9d0658a5-9c45-4297-9c1a-5d37f23f0eed
InstallationMedia: Kubuntu 10.04 "Lucid Lynx" - Alpha amd64 (20100416)
MachineType: ASUSTeK Computer Inc. K60IJ
ProcCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.35-4-generic root=UUID=ca1ae227-2c4d-4613-9de4-a7b29a4eb014 ro quiet splash
ProcEnviron:
 LANGUAGE=en
 PATH=(custom, no user)
 LANG=en_US.UTF-8
 SHELL=/bin/bash
RelatedPackageVersions: linux-firmware 1.36
RfKill:
 0: phy0: Wireless LAN
  Soft blocked: no
  Hard blocked: no
SourcePackage: linux
dmi.bios.date: 08/26/2009
dmi.bios.vendor: American Megatrends Inc.
dmi.bios.version: 203
dmi.board.asset.tag: ATN12345678901234567
dmi.board.name: K60IJ
dmi.board.vendor: ASUSTeK Computer Inc.
dmi.board.version: 1.0
dmi.chassis.asset.tag: ATN12345678901234567
dmi.chassis.type: 10
dmi.chassis.vendor: ASUSTeK Computer Inc.
dmi.chassis.version: 1.0
dmi.modalias: dmi:bvnAmericanMegatrendsInc.:bvr203:bd08/26/2009:svnASUSTeKComputerInc.:pnK60IJ:pvr1.0:rvnASUSTeKComputerInc.:rnK60IJ:rvr1.0:cvnASUSTeKComputerInc.:ct10:cvr1.0:
dmi.product.name: K60IJ
dmi.product.version: 1.0
dmi.sys.vendor: ASUSTeK Computer Inc.

[Jonathan Thomas (echidnaman)]

Correction, the last kernel version that did not have this bug was -1, and I am experiencing this with both -2 and -4. (Using -1 right now to keep my sanity ;)

[Jeremy Foshee (jeremyfoshee)]

Hi Jonathan,

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.]

[Jonathan Thomas (echidnaman)]

Still happens with -5 (just released)

[Joel Stanley (shenki)]

This appears to be the issue described in this lkml thread:

 http://www.gossamer-threads.com/lists/linux/kernel/1236723

I have seen it with 2.6.35-{2,5} kernels with Lucid and also Maverick userspace. Happens on every boot, 4 kslowd threads. My hardware is a Thinkpad X200 with Intel GM45 graphics.

[Jonathan Thomas (echidnaman)]

-7 seems to have fixed this. No more hangs, and kslowd* are not taking up any CPU time at all.

[teachop (teachop)]

Perhaps not fixed. This is happening with the Maverick beta 2.6.35-19-generic on a Dell E6510 with i7 and nvidia. The process showing in top is kslowd001.

[Andrew Hutchings (linuxjedi)]

Same on my Dell E6510 with i5 and nouveau. Suspect the second patch here will work but I've not tried yet:
http://bugs.freedesktop.org/show_bug.cgi?id=29536

[Vincenzo Ampolo (vincenzo-ampolo)]

I'm experiencing exactly the same problem on a Sony Vaio VGN-SR21M with an ATI RADEON graphic card

01:00.0 VGA compatible controller: ATI Technologies Inc Mobility Radeon HD 3400 Series

my kernel is Linux earth 2.6.35-22-generic #33-Ubuntu SMP Sun Sep 19 20:32:27 UTC 2010 x86_64 GNU/Linux and i'm on maverick updated just now installed 1 hour ago from the maverik beta.

System load is almost 5.10 with nothing running and in top appear the kslowd000 process. The same problem appears in tty1 so it's NOT related to graphic card.

[Vincenzo Ampolo (vincenzo-ampolo)]

A quick workaround which worked for me is to install the kernel from:

http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.36-rc5-maverick/

kslowd is kworker in that version and it works much more better, no more hangs.

[Andy Whitcroft (apw)]

@Vincenzo -- the issue which was triggering these deamons to eat cpu was resolved for maverick (for this bug). If you are seeing similar issue on -22 we need you to open a new bug as although the symptoms are similar the bug is very likely not. Thanks!

[Jonathan Thomas (echidnaman)]

Re-closing as my issue was fixed.

[Alexandros Vellis (avel)]

I experienced the same issue with kslowd, with 2.6.35-22-generic in a Thinkpad R50e (Intel graphics). Using the 2.6.36-rc5 kernel the system works smoothly again.

[burli (mb-embedit)]

Two days until Maverick release and the bug is still there. My laptop (Belinea o.book 13011, Intel 4500HD) is unusable while the kslowd tasks are running

[zzeroo (zzeroo)]

Vincenzo's Quick n Durty Tip works fine for me. What's different with that kernel?

http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.36-rc5-maverick/

[T3$T (anonymat-fr)]

I have the same problem with a T400 and Intel graphic card. sometimes my laptop is unusable because of this problem

[Søren Holm (sgh)]

These are potential fixes (might be more) since 2,6.35

930a9e2 drm: Use a nondestructive mode for output detect when polling (v2)
7b334fc drm: Use a nondestructive mode for output detect when polling
551402a drm: Fix regression in disable polling e58f637
e58f637 drm/kms: Add a module parameter to disable polling
9a919c4 drm: fix fallouts from slow-work -> wq conversion
991ea75 drm: use workqueue instead of slow-work

I will try to apply some of them, unless someone knows which changes fix the issue.

[Nicholas J Kreucher (kreucher)]

Hmm, I've tested 4 kernels and the only one that doesn't have this behavior for me (so far) is vanilla 2.6.34.

2.6.36-rc7+ (vanilla) -- broken (kslowd renamed to kworker? seems even worse)
2.6.35-22 (Ubuntu) -- broken
2.6.35 (vanilla) -- broken
2.6.34 (vanilla) -- works? (so far)

So unless something was fixed in -rc5, and broken again by -rc7+ (85a3318), might I be seeing a different bug? Also, it seems intermittent, so I am not 100% convinced even 2.6.34 works at this point.

Using a ThinkPad T400 type 6474 in 32-bit mode (w/i915 graphics).

[juliobahar (yahalla-julio)]

I must confirm that kernel 2.6.36-rc5 worked well for me on my toshiba satellite a55 running ubuntu 10.10, 32-bit, and the notoriously supported 855gm intel graphics.

Yes, No more jerkiness nor stuttering. I'm happy now

I've also tried the -rc7 but not as smooth at the -rc5

[zzeroo (zzeroo)]

You must disable polling to remove the odd behavior, either on the fly
via:

# sudo -i && echo N >/sys/module/drm_kms_helper/parameters/poll

or for a permanent solution.

# sudo -i
# echo "options drm_kms_helper poll=N">/etc/modprobe.d/local.conf
# reboot

Am Sonntag, den 03.10.2010, 16:48 +0000 schrieb Alexandros Vellis:
> I experienced the same issue with kslowd, with 2.6.35-22-generic in a
> Thinkpad R50e (Intel graphics). Using the 2.6.36-rc5 kernel the system
> works smoothly again.
>

[Maor (maor-gm)]

I've tested few kernel and all kernel post 2.6.35 are no good, 2.6.36-rc7 even worse it cause a complete lockup of the OS for few few seconds.

2.6.36-rc7 - not good
2.6.36.-rc5 - not good (although better than rc7)
2.6.35-22 - no good
2.6.34 - good
2.6.32 (lucid kernel) - good

The only option to work with newer kernels is to disable polling.

Using a ThinkPad T400 in 32-bit PAE mode (w/i915 graphics).

[Nicholas J Kreucher (kreucher)]

Following up on my #18 post, I can confirm 2.6.34 still works fine after 24+ hours. This supports maor's findings as well.

I don't see /sys/module/drm_kms_helper/parameters at all in 2.6.34... so was polling added in 2.6.35?

If not that, someone have time to git bisect 2.6.34 and 2.6.35 ? :)

[Gijutsu (gijutsu)]

Nick Bowler (nbowler [at] elliptictech) has already done a git bisect and concluded that reverting commit fbf81762e385d ("drm/kms: disable/enable poll around switcheroo on/off") solves this issue.
Reference: http://www.gossamer-threads.com/lists/linux/kernel/1238374#1239762

I can also confirm that the issue is still present in the latest kernel (2.6.35-22-generic) for Ubuntu 10.10. Therefore this bug should be changed from "Fix Released" to get the attention it needs.

[Jonathan Thomas (echidnaman)]

See comment #12. Other issues should be tracked elsewhere.

[DrPaulaner (dr-paulaner)]

same in here, using 35.22 kernel, ubuntu 10.10 last updates

[Vincenzo Ampolo (vincenzo-ampolo)]

Ok Since Jonathan Thomas thinks this is a fixed issue, let's post our problems to this bug report instead:

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/662946

[baracus2k (baracus2k)]

I have the same issue with a Dell Inspiron 1150 (Intel 852/855 gfx) running ubuntu 10.10 on 35-22 kernel. I can't get the patch to apply, I always get hunk failed errors. I tried the above option with the drm parameters but there is no parameters folder and creating the local.conf with the options caused ubuntu to give an error at boot saying option 'poll' not found.

Can anyone tell me what exactly I'm doing wrong? Using kernel rc5 fixes it for me but my wifi doesn't work so I need to get this patch working somehow under kernel 35-22.

[Vincenzo Ampolo (vincenzo-ampolo)]

@baracus2k: This bug is marked "Fix Released" so any other comment will not get the needed attenction.
Please post in https://bugs.launchpad.net/ubuntu/+source/linux/+bug/662946 in which there is a new bug report for the same issue.

[Mihai (udev)]

This bug also affects me, on a T400 with Intel graphics, 35.22 x64 kernel, ubuntu 10.10

[therealv (therealv)]

samsung x120 laptop with intel HD4500 graphics has the same "kslowd" problem with maverick ubuntu all updates

[Carlos (hirs)]

Same here with an asus ul30a, intel HD4500 graphics card as well. It usually happens waking up from suspend

[Tim Utschig (tim-tetro)]

Same issue on a Dell OptiPlex 755 under 2.6.3[56]*

I installed 2.6.37-rc2 from: http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.37-rc2-maverick/
Seems slightly better with kworker but still noticeable and annoying.

I set poll=N for drm_kms_helper and now I don't notice the problem anymore.

[Arangel Angov (arangel)]

Issue persists on 2.6.35-24 and 2.6.36-02063602.

[zzeroo (zzeroo)]

Selbst in Vanilla-Kernel 2.6.38 Brauch-rc2 präsentieren. Und Umfrage = N für drm_kms_helper hat keine Wirkung. Ich meine nicht gelöst.

[zzeroo (zzeroo)]

Even in custom vanilla kernel 2.6.38-rc2 present. And poll=N for drm_kms_helper has no effect. I mean not solved.

[Achim (ach1m)]

I also have noticed that setting
# echo n > /sys/module/drm_kms_helper/parameters/poll
has no effect.

I also checked that the parameter was set right.
cat /sys/module/drm_kms_helper/parameters/poll
N

I still see this message in my logs
kernel: [ 833.140022] [drm:intel_crt_detect], CRT not detected via hotplug
I would expect it to disappear, if the parameter was set right.

I am reporting this from Natty

linux:
  Installed: (none)
  Candidate: 2.6.38.8.22
  Version table:
     2.6.38.8.22 0
        500 http://archive.ubuntu.com/ubuntu/ natty/main amd64 Packages

[Achim (ach1m)]

As it seems this is a bug in upowerd I have filled a new report:
https://bugs.launchpad.net/ubuntu/+source/upower/+bug/758593