Bug 43091 - core2 cpu locked to lowest cpu scaling speed - Lenovo ThinkCentre 8810-91G
Summary: core2 cpu locked to lowest cpu scaling speed - Lenovo ThinkCentre 8810-91G
Status: CLOSED INSUFFICIENT_DATA
Alias: None
Product: ACPI
Classification: Unclassified
Component: Config-Processors (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: Rafael J. Wysocki
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2012-04-11 16:41 UTC by Anton Piatek
Modified: 2016-06-24 17:10 UTC (History)
10 users (show)

See Also:
Kernel Version: 3.4.0-030400rc2
Subsystem:
Regression: No
Bisected commit-id:


Attachments
sensors output (978 bytes, text/plain)
2012-08-13 07:37 UTC, Anton Piatek
Details
dmesg output (67.16 KB, text/plain)
2012-08-13 07:38 UTC, Anton Piatek
Details
acpidump (154.13 KB, text/plain)
2012-08-13 07:40 UTC, Anton Piatek
Details
dmesg with processor.ignore_ppc=1 (62.13 KB, text/plain)
2012-08-13 07:49 UTC, Anton Piatek
Details
grep of cpufreq with processor.ignore_ppc=1 (996 bytes, text/plain)
2012-08-13 07:49 UTC, Anton Piatek
Details

Description Anton Piatek 2012-04-11 16:41:51 UTC
Originally reported against Ubuntu, more information there may be helpful but let me know if you want anything specifically https://bugs.launchpad.net/ubuntu/+source/linux/+bug/962947

It appears that my 1.86GHz Intel Core2 CPU is only running at 700Mhz. Th cpu is actually a 1.86Ghz, but cpufreq only lets me set a maximum of 700mhz. 
The system is a Lenovo ThinkCentre 8810-91G, and the bios is the latest one available. The bios doesn't show any cpu limiting options or restrictions, and shows the cpus as the expected 1.86Mhz (though I think that is just model information rather than the current clock speed). I cannot see any overclocking options in the bios.


anton@smeg:~$uname -a Linux smeg.hursley.ibm.com 3.4.0-030400rc2-generic #201204072235 SMP Sun Apr 8 02:36:11 UTC 2012 x86_64 x86_64 x86_64 GNU/Linux

anton@smeg:~$cpufreq-info cpufrequtils 007: cpufreq-info (C) Dominik Brodowski 2004-2009Report errors and bugs to <email address hidden>, please.
analyzing CPU 0: 
driver: acpi-cpufreq 
CPUs which run at the same hardware frequency: 0 1 
CPUs which need to have their frequency coordinated by software: 0
maximum transition latency: 10.0 us. 
hardware limits: 600 MHz - 700 MHz 
available frequency steps: 700 MHz, 600 MHz 
available cpufreq governors: conservative, ondemand, userspace, powersave, performance
current policy: frequency should be within 600 MHz and 700 MHz. The governor "ondemand" may decide which speed to use within this range. 
current CPU frequency is 600 MHz. 
cpufreq stats: 700 MHz:45.14%, 600 MHz:54.86% (343)

I tried setting processor.ignore_ppc=1 on the kernel boot options, but it made no difference.

I marked this as a regression as it used to work, but as I don't keep an eye on cpu scaling speeds on my desktop I have no idea when it last worked at full speed. I will try some older kernels and see if I can find one which works.
Comment 1 Anton Piatek 2012-04-17 13:54:27 UTC
Reverting back to a Ubuntu 3.0 kernel works fine!

$cat /proc/version_signature 
Ubuntu 3.0.0-17.30-generic 3.0.22



$cpufreq-info 
cpufrequtils 007: cpufreq-info (C) Dominik Brodowski 2004-2009
Report errors and bugs to cpufreq@vger.kernel.org, please.
analyzing CPU 0:
  driver: acpi-cpufreq
  CPUs which run at the same hardware frequency: 0 1
  CPUs which need to have their frequency coordinated by software: 0
  maximum transition latency: 10.0 us.
  hardware limits: 1.60 GHz - 1.87 GHz
  available frequency steps: 1.87 GHz, 1.60 GHz
  available cpufreq governors: conservative, ondemand, userspace, powersave, performance
  current policy: frequency should be within 1.60 GHz and 1.87 GHz.
                  The governor "ondemand" may decide which speed to use
                  within this range.
  current CPU frequency is 1.87 GHz.
  cpufreq stats: 1.87 GHz:75.13%, 1.60 GHz:24.87%  (4778)
analyzing CPU 1:
  driver: acpi-cpufreq
  CPUs which run at the same hardware frequency: 0 1
  CPUs which need to have their frequency coordinated by software: 1
  maximum transition latency: 10.0 us.
  hardware limits: 1.60 GHz - 1.87 GHz
  available frequency steps: 1.87 GHz, 1.60 GHz
  available cpufreq governors: conservative, ondemand, userspace, powersave, performance
  current policy: frequency should be within 1.60 GHz and 1.87 GHz.
                  The governor "ondemand" may decide which speed to use
                  within this range.
  current CPU frequency is 1.60 GHz.
  cpufreq stats: 1.87 GHz:72.79%, 1.60 GHz:27.21%  (4691)
Comment 2 Len Brown 2012-04-24 02:53:28 UTC
3.0 works
3.2 works (according to ubuntu bug report)

3.4-rc2 fails, according to comment #2
How about 3.3?
How about the latest 3.4 (now -rc4)?
Comment 3 Anton Piatek 2012-04-26 07:25:56 UTC
3.2 is odd actually. The following 3.2.0-22 kernel worked
anton@smeg:~$cat /proc/version_signature
Ubuntu 3.2.0-22.35-generic 3.2.14

But this 3.2.0.-23 kernel didn't work.
$cat /proc/version_signature
Ubuntu 3.2.0-23.36-generic 3.2.14

I am wondering if something else rather than the kernel is the cause, or is
part of the cause.

I will try a few more kernels and post here and the ubuntu bug what works and
what doesn't
Comment 4 Anton Piatek 2012-04-26 08:17:16 UTC
It appears that after updating a large number of packages yesterday, all 4 kernels I have installed run at 700mhz today.
3.2.0-22 ubuntu
3.2.0-23 ubuntu
3.0.0    ubuntu
3.4.0    mainline
all run at 700mhz. I am not sure the kernel itself can be the problem here. Is there anything else I should be looking at, or should I just close this bug report as not being a kernel bug?
Comment 5 Leandro Lucarella 2012-05-03 13:25:22 UTC
I'm having the same problem but with a core i5 processor. Also reported a bug first to Ubuntu, you can find all the details there (/proc/cpuinfo, etc):
https://bugs.launchpad.net/bugs/987531

For me it works with the conservative governor but not with the ondemand governor. Should I report another bug?
Comment 6 Len Brown 2012-05-08 01:58:10 UTC
please show the output from

grep . /sys/devices/system/cpu/cpu*/cpufreq/*

in particular, somebody could have scribbled on the max_freq.

Given that all the kernels fail, possibly some user-space package
has gone wrong.
Comment 7 Anton Piatek 2012-05-08 07:01:29 UTC
$sudo grep . /sys/devices/system/cpu/cpu*/cpufreq/*
/sys/devices/system/cpu/cpu0/cpufreq/affected_cpus:0
/sys/devices/system/cpu/cpu0/cpufreq/bios_limit:700000
/sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_cur_freq:700000
/sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq:700000
/sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_min_freq:600000
/sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_transition_latency:10000
/sys/devices/system/cpu/cpu0/cpufreq/related_cpus:0 1
/sys/devices/system/cpu/cpu0/cpufreq/scaling_available_frequencies:700000 600000 
/sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors:conservative ondemand userspace powersave performance 
/sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq:700000
/sys/devices/system/cpu/cpu0/cpufreq/scaling_driver:acpi-cpufreq
/sys/devices/system/cpu/cpu0/cpufreq/scaling_governor:ondemand
/sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq:700000
/sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq:600000
/sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed:<unsupported>
/sys/devices/system/cpu/cpu1/cpufreq/affected_cpus:1
/sys/devices/system/cpu/cpu1/cpufreq/bios_limit:700000
/sys/devices/system/cpu/cpu1/cpufreq/cpuinfo_cur_freq:700000
/sys/devices/system/cpu/cpu1/cpufreq/cpuinfo_max_freq:700000
/sys/devices/system/cpu/cpu1/cpufreq/cpuinfo_min_freq:600000
/sys/devices/system/cpu/cpu1/cpufreq/cpuinfo_transition_latency:10000
/sys/devices/system/cpu/cpu1/cpufreq/related_cpus:0 1
/sys/devices/system/cpu/cpu1/cpufreq/scaling_available_frequencies:700000 600000 
/sys/devices/system/cpu/cpu1/cpufreq/scaling_available_governors:conservative ondemand userspace powersave performance 
/sys/devices/system/cpu/cpu1/cpufreq/scaling_cur_freq:600000
/sys/devices/system/cpu/cpu1/cpufreq/scaling_driver:acpi-cpufreq
/sys/devices/system/cpu/cpu1/cpufreq/scaling_governor:ondemand
/sys/devices/system/cpu/cpu1/cpufreq/scaling_max_freq:700000
/sys/devices/system/cpu/cpu1/cpufreq/scaling_min_freq:600000
/sys/devices/system/cpu/cpu1/cpufreq/scaling_setspeed:<unsupported>

While running a 3.2.0-23.36-generic ubuntu kernel.

I am leaning towards some userspace package being the cause, as I have tried several kernels and all produce the same error after now. I have absolutely no idea what package, but hopefully my Ubuntu bug can track that down. Should I close this? Or is there anything else to try before assuming it is all userspace
Comment 8 Anton Piatek 2012-05-14 14:12:30 UTC
The plot thickens...
I turned of the computer for a few minutes this morning (it is normally always on), and when it came back up the CPU now shows a full speed range. Either it is something that a cold start reset in the bios, or perhaps something related to thermal protection trying to lock the max speed.
I had actually rebooted the box a few minutes before powering it off, and it was still running with a scaled back cpu speed, so it really doesn't look like it is related to the version of my kernel nor any packages.
Comment 9 Len Brown 2012-06-05 02:46:40 UTC
re: /sys/devices/system/cpu/cpu0/cpufreq/bios_limit:700000

I think this means that the BIOS is telling your system
not to exceed 700 Mhz.

Are there any messages in dmesg when this happens,
say about thermal throttling due to over-heating.
(please attach full dmesg upon failure)

What do you see when you poke at the temperature
of the system?

Please show the output from 
grep . /sys/class/thermal/thermal*/*

and see if the temperature "temp" relates to success and failure.

Also, when you boot with processor.ignore_ppc=1, please attach
the output from

grep . /sys/devices/system/cpu/cpu0/cpufreq/*

finally, 
Please attach the output from acpidump
(collected in any configuration)
Comment 10 Anton Piatek 2012-08-13 07:37:42 UTC
Created attachment 77501 [details]
sensors output
Comment 11 Anton Piatek 2012-08-13 07:38:01 UTC
Created attachment 77511 [details]
dmesg output
Comment 12 Anton Piatek 2012-08-13 07:40:54 UTC
Created attachment 77521 [details]
acpidump
Comment 13 Anton Piatek 2012-08-13 07:41:00 UTC
Have just rebooted and got limited cpu. 

I cannot find the thermal sys entries you wanted, would they have been renamed?
$ls /sys/class/thermal/
cooling_device0  cooling_device1

$ls /sys/class/thermal/cooling_device*/ -l
/sys/class/thermal/cooling_device0/:
total 0
-rw-r--r-- 1 root root 4096 Aug 13 08:38 cur_state
lrwxrwxrwx 1 root root    0 Aug 13 08:38 device -> ../../../LNXSYSTM:00/LNXCPU:00
-r--r--r-- 1 root root 4096 Aug 13 08:38 max_state
drwxr-xr-x 2 root root    0 Aug 13 08:38 power
lrwxrwxrwx 1 root root    0 Aug 13 08:30 subsystem -> ../../../../class/thermal
-r--r--r-- 1 root root 4096 Aug 13 08:38 type
-rw-r--r-- 1 root root 4096 Aug 13 08:30 uevent

/sys/class/thermal/cooling_device1/:
total 0
-rw-r--r-- 1 root root 4096 Aug 13 08:38 cur_state
lrwxrwxrwx 1 root root    0 Aug 13 08:38 device -> ../../../LNXSYSTM:00/LNXCPU:01
-r--r--r-- 1 root root 4096 Aug 13 08:38 max_state
drwxr-xr-x 2 root root    0 Aug 13 08:38 power
lrwxrwxrwx 1 root root    0 Aug 13 08:30 subsystem -> ../../../../class/thermal
-r--r--r-- 1 root root 4096 Aug 13 08:38 type
-rw-r--r-- 1 root root 4096 Aug 13 08:30 ueven

Will reboot with the processor.ignore_ppc=1 and gather that data too
Comment 14 Anton Piatek 2012-08-13 07:49:26 UTC
Created attachment 77531 [details]
dmesg with processor.ignore_ppc=1
Comment 15 Anton Piatek 2012-08-13 07:49:56 UTC
Created attachment 77541 [details]
grep of cpufreq with processor.ignore_ppc=1
Comment 16 Anton Piatek 2012-08-13 07:52:45 UTC
For the record, my kernel at this time is 3.2.0-29-generic Ubuntu
$ cat /proc/version_signature 
Ubuntu 3.2.0-29.46-generic 3.2.24

I'm not sure if I should be changing the state of this bug now that I have collected data.

I can't prove the temperature is the cause, but turning off the box for 10 minutes and then rebooting it appears to be a workaround to get back to a full speed cpu.
Comment 17 Zhang Rui 2012-11-28 13:24:16 UTC
so processor.ignore_ppc does not help in your case, right?
Comment 18 Anton Piatek 2012-12-04 11:08:52 UTC
correct
Comment 19 Yill Din 2012-12-04 11:17:32 UTC
Just in case it helps, I had a similar problem with my Notebook (Toshiba Satellite Z830) and I "fixed" it by changing some BIOS option. Here is an article I wrote about it:
http://www.llucax.com.ar/blog/blog/post/-31ba9e5e

You might want to try messing around with the BIOS settings...
Comment 20 Aaron Lu 2013-04-02 06:33:04 UTC
Hi Anton,

Does justincase's suggestion help for you?
Comment 21 Lan Tianyu 2013-04-18 00:58:14 UTC
ping ...
Comment 22 Anton Piatek 2013-04-18 07:37:44 UTC
No, there are no BIOS options which appear to help.

The best I can figure out is something related to a thermal cutout, as rebooting often sees the issue continue, but powering off for a few minutes and letting the cpu cool seems to have the cpu come up at full speed.

It could even be a thermal limit built in to the bios, rather than the kernel, but that is pure speculation and I have no way of testing that hypothesis.
Comment 23 Zhang Rui 2013-05-13 02:10:37 UTC
(In reply to comment #5)
> I'm having the same problem but with a core i5 processor. Also reported a bug
> first to Ubuntu, you can find all the details there (/proc/cpuinfo, etc):
> https://bugs.launchpad.net/bugs/987531
> 
> For me it works with the conservative governor but not with the ondemand
> governor. Should I report another bug?

please file a new bug report if the problem still exists in the latest kernel.
Comment 24 Zhang Rui 2013-05-13 02:12:45 UTC
Anton,
First, please find a kernel that the problem does not exist. or else I'll remove the regression flag.
Second, please attach the output of " grep . /sys/devices/system/cpu/cpu*/cpufreq/*" in the latest kernel, say, 3.10-rc1, both with and without processor.ignore_ppc.
Comment 25 Rafael J. Wysocki 2013-06-04 01:16:12 UTC
There's no evidence that this is a regression.
Comment 26 Anton Piatek 2013-11-18 08:46:14 UTC
This hansn't gone away. Finding a kernel which pre-dates this bug is almost impossible as I have seen it on a 3.0 kernel, I am now running 3.11... I do not have time to revert the system to a 2.6 kernel just to see if this reverts itself.

anton@smeg:~$  grep . /sys/devices/system/cpu/cpu*/cpufreq/*
/sys/devices/system/cpu/cpu0/cpufreq/affected_cpus:0
/sys/devices/system/cpu/cpu0/cpufreq/bios_limit:700000
grep: /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_cur_freq: Permission denied
/sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq:700000
/sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_min_freq:600000
/sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_transition_latency:10000
/sys/devices/system/cpu/cpu0/cpufreq/freqdomain_cpus:0 1
/sys/devices/system/cpu/cpu0/cpufreq/related_cpus:0
/sys/devices/system/cpu/cpu0/cpufreq/scaling_available_frequencies:700000 600000 
/sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors:conservative ondemand userspace powersave performance 
/sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq:600000
/sys/devices/system/cpu/cpu0/cpufreq/scaling_driver:acpi-cpufreq
/sys/devices/system/cpu/cpu0/cpufreq/scaling_governor:ondemand
/sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq:700000
/sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq:600000
/sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed:<unsupported>
grep: /sys/devices/system/cpu/cpu0/cpufreq/stats: Is a directory
/sys/devices/system/cpu/cpu1/cpufreq/affected_cpus:1
/sys/devices/system/cpu/cpu1/cpufreq/bios_limit:700000
grep: /sys/devices/system/cpu/cpu1/cpufreq/cpuinfo_cur_freq: Permission denied
/sys/devices/system/cpu/cpu1/cpufreq/cpuinfo_max_freq:700000
/sys/devices/system/cpu/cpu1/cpufreq/cpuinfo_min_freq:600000
/sys/devices/system/cpu/cpu1/cpufreq/cpuinfo_transition_latency:10000
/sys/devices/system/cpu/cpu1/cpufreq/freqdomain_cpus:0 1
/sys/devices/system/cpu/cpu1/cpufreq/related_cpus:1
/sys/devices/system/cpu/cpu1/cpufreq/scaling_available_frequencies:700000 600000 
/sys/devices/system/cpu/cpu1/cpufreq/scaling_available_governors:conservative ondemand userspace powersave performance 
/sys/devices/system/cpu/cpu1/cpufreq/scaling_cur_freq:600000
/sys/devices/system/cpu/cpu1/cpufreq/scaling_driver:acpi-cpufreq
/sys/devices/system/cpu/cpu1/cpufreq/scaling_governor:ondemand
/sys/devices/system/cpu/cpu1/cpufreq/scaling_max_freq:700000
/sys/devices/system/cpu/cpu1/cpufreq/scaling_min_freq:600000
/sys/devices/system/cpu/cpu1/cpufreq/scaling_setspeed:<unsupported>
grep: /sys/devices/system/cpu/cpu1/cpufreq/stats: Is a directory
anton@smeg:~$ uname -a
Linux smeg 3.11.0-11-generic #17-Ubuntu SMP Tue Oct 1 19:42:04 UTC 2013 x86_64 x86_64 x86_64 GNU/Linux

This is again an Ubuntu kernel, I will try some more pure kernels at some point, but so far nothing has got any useful information out of my system other than the problem definitely exists...

There must be some real diagnostics I can try instead of this messing about with the same greps or cpufreq-info every time someone else has a look at this bug...
Comment 27 Lan Tianyu 2013-12-11 14:52:43 UTC
Hi Anton:
        Please check speedstep in your Bios. If it existed and was enabled, disable it and test again.
Comment 28 Anton Piatek 2013-12-11 15:58:07 UTC
Hi Lan, 
There was an option for "GV1/GV3 P type power management" which once disabled seems to disable cpufreq from really doing much:
$ cpufreq-info 
cpufrequtils 008: cpufreq-info (C) Dominik Brodowski 2004-2009
Report errors and bugs to cpufreq@vger.kernel.org, please.
analyzing CPU 0:
  no or unknown cpufreq driver is active on this CPU
  maximum transition latency: 4294.55 ms.
analyzing CPU 1:
  no or unknown cpufreq driver is active on this CPU
  maximum transition latency: 4294.55 ms.

$ ls /sys/devices/system/cpu/cpu*/cpufreq*
ls: cannot access /sys/devices/system/cpu/cpu*/cpufreq*:

I can't tell what frequency the system is running at, but presumably is no longer throttled
Comment 29 Zhang Rui 2014-06-03 02:08:07 UTC
As I said, please attach the output of " grep . /sys/devices/system/cpu/cpu*/cpufreq/*" in the latest vanilla kernel you're using, both with and without boot option processor.ignore_ppc.
Comment 30 Zhang Rui 2014-10-23 07:53:42 UTC
bug closed as there is no response from the bug reporter.
please feel free to re-open it once you can provide the information required in comment #29.
Comment 31 Anton Piatek 2014-10-23 08:01:31 UTC
I no longer have the hardware so can't do anything on this

On 23 October 2014 08:53,  <bugzilla-daemon@bugzilla.kernel.org> wrote:
> https://bugzilla.kernel.org/show_bug.cgi?id=43091
>
> Zhang Rui <rui.zhang@intel.com> changed:
>
>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>              Status|NEEDINFO                    |CLOSED
>          Resolution|---                         |INSUFFICIENT_DATA
>
> --- Comment #30 from Zhang Rui <rui.zhang@intel.com> ---
> bug closed as there is no response from the bug reporter.
> please feel free to re-open it once you can provide the information required
> in
> comment #29.
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.
> You reported the bug.

Note You need to log in before you can comment on or make changes to this bug.