Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
[SOLVED] kworkers eating my cpu (again)
View unanswered posts
View posts from last 24 hours

 
Reply to topic    Gentoo Forums Forum Index Kernel & Hardware
View previous topic :: View next topic  
Author Message
Fran
Guru
Guru


Joined: 29 Feb 2004
Posts: 530
Location: Coruña (Spain)

PostPosted: Fri Mar 29, 2013 12:03 pm    Post subject: [SOLVED] kworkers eating my cpu (again) Reply with quote

Sometimes I get a constant 30% use of my puny CPU (ULV SU4100) by one or two kworker threads. It isn't very annoying in the console, but on X it causes everything to stutter like crazy (including the mouse cursor). The computer gets unusable until it passses (sometimes in seconds, sometimes only after a reboot) or it simply freezes to a halt, forcing me to pull the battery.

In the past I already had this problem in this same laptop. I solved it by disabling drm kms polling. But recently (probably with 3.8, I'm not sure) it has appeared again, even with drm polling disabled.

(edit) I'm not alone:
https://bbs.archlinux.org/viewtopic.php?pid=1248190
http://verahill.blogspot.com.es/2013/03/368-slow-mouse-and-keyboard-triggered.html

Also with 3.8, also unrelated to drm kms polling. Does anyone know how to find out which kernel component causes the kworkers to get crazy? I could do a git bisect, but it's a big pain in the ass in a slow processor like the SU4100. I don't see anything weird in /sys/firmware/acpi/interrupts.


Last edited by Fran on Tue Apr 09, 2013 9:30 am; edited 1 time in total
Back to top
View user's profile Send private message
aCOSwt
Bodhisattva
Bodhisattva


Joined: 19 Oct 2007
Posts: 2537
Location: Hilbert space

PostPosted: Fri Mar 29, 2013 12:28 pm    Post subject: Reply with quote

Can you spot on your system any particular interrupt "going mad" ? any particular thread displaying an unusual stime ?
_________________
Back to top
View user's profile Send private message
Fran
Guru
Guru


Joined: 29 Feb 2004
Posts: 530
Location: Coruña (Spain)

PostPosted: Fri Mar 29, 2013 2:14 pm    Post subject: Reply with quote

aCOSwt wrote:
Can you spot on your system any particular interrupt "going mad" ? any particular thread displaying an unusual stime ?

Nope and nope.

(edit) Wait, yeah. i915 starts interrupting like crazy:
Code:
           CPU0       CPU1       
  0:      35433      35563   IO-APIC-edge      timer
  1:        980        917   IO-APIC-edge      i8042
  8:         24         20   IO-APIC-edge      rtc0
  9:         68         72   IO-APIC-fasteoi   acpi
 12:      13887      13928   IO-APIC-edge      i8042
 16:          0          0   IO-APIC-fasteoi   uhci_hcd:usb2, uhci_hcd:usb5
 19:          0          0   IO-APIC-fasteoi   ehci_hcd:usb1, uhci_hcd:usb4
 23:         14         15   IO-APIC-fasteoi   uhci_hcd:usb3, ehci_hcd:usb6
 42:     113836     113519   PCI-MSI-edge      i915
 43:      15801      15858   PCI-MSI-edge      ahci
 44:       5804       5244   PCI-MSI-edge      iwlwifi
 45:        108        113   PCI-MSI-edge      snd_hda_intel
NMI:          0          0   Non-maskable interrupts
LOC:     177735     180565   Local timer interrupts
SPU:          0          0   Spurious interrupts
PMI:          0          0   Performance monitoring interrupts
IWI:          0          0   IRQ work interrupts
RTR:          0          0   APIC ICR read retries
RES:      34498      22338   Rescheduling interrupts
CAL:        104         61   Function call interrupts
TLB:       1189       1143   TLB shootdowns
TRM:      12908      12908   Thermal event interrupts
THR:          0          0   Threshold APIC interrupts
MCE:          0          0   Machine check exceptions
MCP:          2          2   Machine check polls
ERR:          0
MIS:          0

Doesn't seem much higher than the rest, but that's because it had already stopped when I printed it. It was growing at a rate of 1000 interrupts/second instead of the usual 30 per second.

So, again it's caused by the damned i915 driver, like the last time it happened in 2.6 kernels. But this time it's not drm kms polling, unless there is a bug in 3.8 that ignores the setting in /sys/module/drm_kms_helper/parameters/poll
Back to top
View user's profile Send private message
Fran
Guru
Guru


Joined: 29 Feb 2004
Posts: 530
Location: Coruña (Spain)

PostPosted: Sat Mar 30, 2013 9:49 am    Post subject: Reply with quote

Kernel 3.9_rc4 makes X usable again (even when i915 starts interrupting X doesn't stutter) but the i915 interrupt problem persists.
Back to top
View user's profile Send private message
aCOSwt
Bodhisattva
Bodhisattva


Joined: 19 Oct 2007
Posts: 2537
Location: Hilbert space

PostPosted: Sat Mar 30, 2013 10:35 am    Post subject: Reply with quote

Fran wrote:
Code:
           CPU0       CPU1       
TRM:      12908      12908   Thermal event interrupts

should'nt kworkers Heating my cpu be the real title of your thread ? :wink:
_________________
Back to top
View user's profile Send private message
Fran
Guru
Guru


Joined: 29 Feb 2004
Posts: 530
Location: Coruña (Spain)

PostPosted: Sun Mar 31, 2013 9:32 am    Post subject: Reply with quote

aCOSwt wrote:
Fran wrote:
Code:
           CPU0       CPU1       
TRM:      12908      12908   Thermal event interrupts

should'nt kworkers Heating my cpu be the real title of your thread ? :wink:

I actually have no idea what those events mean. Temperature doesn't go above 45º (coretemp).

Oh well. I suppose I'll be happy with 3.9-rc for now. I hate these random kernel bugs. I hope someone with a better CPU and larger hard disk is willing to bisect.
Back to top
View user's profile Send private message
gamanakis
n00b
n00b


Joined: 04 Apr 2013
Posts: 1

PostPosted: Thu Apr 04, 2013 7:47 pm    Post subject: Reply with quote

I did a bisecting and the following seems to be the culprit.
Please someone test and report.
https://bbs.archlinux.org/viewtopic.php?pid=1254285#p1254285
Back to top
View user's profile Send private message
Fran
Guru
Guru


Joined: 29 Feb 2004
Posts: 530
Location: Coruña (Spain)

PostPosted: Tue Apr 09, 2013 9:30 am    Post subject: Reply with quote

gamanakis wrote:
I did a bisecting and the following seems to be the culprit.
Please someone test and report.
https://bbs.archlinux.org/viewtopic.php?pid=1254285#p1254285

I've seen the post in arch taking to the upstream bugs. Marking this as solved.

https://patchwork.kernel.org/patch/2400621/
https://patchwork.kernel.org/patch/2402211/
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Kernel & Hardware All times are GMT
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum