fan control on dell inspiron

tux111 · 10-06-2012, 05:53 PM

Dear All,

on a Dell Inspiron the CPU Fan is too loud. It is not running constantly, but it changes from slow to really fast every minute. It's really annoying and hardly possible to work on this machine.

pwmcontrol says "There are no pwm-capable sensor modules installed"

Is it possible to limit the range the fan is using?

- Chris

TobiSGD · 10-06-2012, 07:09 PM

If you have such big changes in fan speed I would think that you power-management is not working correctly.
To help you we need more information:
- Distribution and version.
- model of CPU (can be found with lscpu) and and video chip (can be found with lspci)
- which video drivers are in use
- which power-management modules are loaded (can be found with lsmod | grep cpufreq)

tux111 · 10-07-2012, 03:07 AM

Dear Tobi,

thank you for your quick reply.

Quote:

Originally Posted by TobiSGD

To help you we need more information:
- Distribution and version.

Ubuntu 12.04.1 LTS (Precise Pangolin)

Quote:

- model of CPU (can be found with lscpu) and and video chip (can be found with lspci)

Code:

root@cd:~# lscpu
Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                4
On-line CPU(s) list:   0-3
Thread(s) per core:    1
Core(s) per socket:    4
Socket(s):             1
NUMA node(s):          1
Vendor ID:             GenuineIntel
CPU family:            6
Model:                 23
Stepping:              10
CPU MHz:               1998.000
BogoMIPS:              4987.50
Virtualization:        VT-x
L1d cache:             32K
L1i cache:             32K
L2 cache:              2048K
NUMA node0 CPU(s):     0-3

Quote:

- which video drivers are in use

nvidia 12319264 40 - Live 0xffffffffa01bc000 (P)

Quote:

- which power-management modules are loaded (can be found with lsmod | grep cpufreq)

Code:

root@cd:~# lsmod
Module                  Size  Used by
pci_stub               12622  1 
vboxpci                23200  0 
vboxnetadp             25670  0 
vboxnetflt             23441  1 
vboxdrv               320210  5 vboxpci,vboxnetadp,vboxnetflt
rfcomm                 47604  0 
bnep                   18281  2 
bluetooth             180104  10 rfcomm,bnep
parport_pc             32866  0 
ppdev                  17113  0 
usblp                  18307  0 
nvidia              12319264  40 
snd_hda_codec_hdmi     32474  4 
joydev                 17693  0 
snd_hda_codec_realtek   224173  1 
usbhid                 47199  0 
hid                    99559  1 usbhid
snd_hda_intel          33773  6 
snd_hda_codec         127706  3 snd_hda_codec_hdmi,snd_hda_codec_realtek,snd_hda_intel
snd_hwdep              13668  1 snd_hda_codec
snd_pcm                97188  4 snd_hda_codec_hdmi,snd_hda_intel,snd_hda_codec
snd_seq_midi           13324  0 
snd_rawmidi            30748  1 snd_seq_midi
snd_seq_midi_event     14899  1 snd_seq_midi
snd_seq                61896  2 snd_seq_midi,snd_seq_midi_event
snd_timer              29990  2 snd_pcm,snd_seq
snd_seq_device         14540  3 snd_seq_midi,snd_rawmidi,snd_seq
snd                    78855  21 snd_hda_codec_hdmi,snd_hda_codec_realtek,snd_hda_intel,snd_hda_codec,snd_hwdep,snd_pcm,snd_rawmidi,snd_seq,snd_timer,snd_seq_device
soundcore              15091  1 snd
snd_page_alloc         18529  2 snd_hda_intel,snd_pcm
hwmon_vid              12827  0 
psmouse                97362  0 
serio_raw              13211  0 
mac_hid                13253  0 
dcdbas                 14490  0 
coretemp               13525  0 
lp                     17799  0 
parport                46562  3 parport_pc,ppdev,lp
usb_storage            49198  0 
r8169                  62099  0

The problem is the CPU and not the GPU cooler. The noise is so annoying that I've even cut through the PWM cable! However, the fan was then much louder than before.

Chris

TobiSGD · 10-07-2012, 04:19 AM

Please don't modify the hardware if you don't know exactly what you are doing. With cutting through the PWM wire you have taken away any possibility to control the fan from the mainboard.
Also, this is not a hardware problem, the problem is that you don't have any power management running on that machine.
Try to load the module acpi-cpufreq with

Code:

modprobe -v acpi-cpufreq

, then post the output of

Code:

lsmod |grep cpufreq

cascade9 · 10-07-2012, 04:26 AM

Quote:

Originally Posted by TobiSGD

- model of CPU (can be found with lscpu)

Do you actually like this method? I find the lscpu output to be annoying if you actually want to know the CPU model. Lshw is better IMO (even though you will get a huge readout, at least it tells you the CPU model)

Quote:

Originally Posted by tux111

The noise is so annoying that I've even cut through the PWM cable! However, the fan was then much louder than before.

You'll have to either repair the wire you've cut, or replace the fan, or fan/heatsink.

TobiSGD · 10-07-2012, 04:30 AM

Quote:

Originally Posted by cascade9

Do you actually like this method? I find the lscpu output to be annoying if you actually want to know the CPU model. Lshw is better IMO (even though you will get a huge readout, at least it tells you the CPU model)

lshw is not installed by default on most distros. Actually I thought lscpu will show me the CPU type, may be I mixed that up, but it took me about a minute to see that this CPU is a mobile Core 2 Quad 9000. But you are right, a better way would have been to use

Code:

cat /proc/cpuinfo | grep name

Works always and without installing extra software.

cascade9 · 10-07-2012, 04:42 AM

Fair point about lshw not being installed by defualt on many distros. I dont think that adding lshw is going to make any real difference, and should be easy, but I'll try to remeber your 'cat /proc/cpuinfo | grep name' command for next time this sort of thread pops up.

Is it really a mobile Core 2 Quad 9000? They should have 6MB (2 x 3MB) L2 cache, not 2MB.

http://ark.intel.com/products/40480

tux111 · 10-07-2012, 05:21 AM

Quote:

Originally Posted by TobiSGD

Please don't modify the hardware if you don't know exactly what you are doing. With cutting through the PWM wire you have taken away any possibility to control the fan from the mainboard.

I've replaced the fan in the meantime.

Quote:

Also, this is not a hardware problem, the problem is that you don't have any power management running on that machine.
Try to load the module acpi-cpufreq with

Code:

modprobe -v acpi-cpufreq

, then post the output of

Code:

lsmod |grep cpufreq

In fact, there's no output, the module isn't loaded.

Code:

root@cd:~# modprobe -v acpi-cpufreq
root@cd:~# lsmod |grep cpufreq
root@cd:~# lsmod
Module                  Size  Used by
pci_stub               12622  1 
vboxpci                23200  0 
vboxnetadp             25670  0 
vboxnetflt             23441  1 
vboxdrv               320210  5 vboxpci,vboxnetadp,vboxnetflt
rfcomm                 47604  0 
bnep                   18281  2 
bluetooth             180104  10 rfcomm,bnep
parport_pc             32866  0 
ppdev                  17113  0 
usblp                  18307  0 
nvidia              12319264  50 
snd_hda_codec_hdmi     32474  4 
joydev                 17693  0 
snd_hda_codec_realtek   224173  1 
usbhid                 47199  0 
hid                    99559  1 usbhid
snd_hda_intel          33773  6 
snd_hda_codec         127706  3 snd_hda_codec_hdmi,snd_hda_codec_realtek,snd_hda_intel
snd_hwdep              13668  1 snd_hda_codec
snd_pcm                97188  4 snd_hda_codec_hdmi,snd_hda_intel,snd_hda_codec
snd_seq_midi           13324  0 
snd_rawmidi            30748  1 snd_seq_midi
snd_seq_midi_event     14899  1 snd_seq_midi
snd_seq                61896  2 snd_seq_midi,snd_seq_midi_event
snd_timer              29990  2 snd_pcm,snd_seq
snd_seq_device         14540  3 snd_seq_midi,snd_rawmidi,snd_seq
snd                    78855  21 snd_hda_codec_hdmi,snd_hda_codec_realtek,snd_hda_intel,snd_hda_codec,snd_hwdep,snd_pcm,snd_rawmidi,snd_seq,snd_timer,snd_seq_device
soundcore              15091  1 snd
snd_page_alloc         18529  2 snd_hda_intel,snd_pcm
hwmon_vid              12827  0 
psmouse                97362  0 
serio_raw              13211  0 
mac_hid                13253  0 
dcdbas                 14490  0 
coretemp               13525  0 
lp                     17799  0 
parport                46562  3 parport_pc,ppdev,lp
usb_storage            49198  0 
r8169                  62099  0

The module seems to be available:

Code:

root@cd:/lib/modules/3.2.0-31-generic/kernel/drivers/cpufreq# ll
insgesamt 40
-rw-r--r-- 1 root root 11136 Sep  7 18:59 p4-clockmod.ko
-rw-r--r-- 1 root root 14240 Sep  7 18:59 pcc-cpufreq.ko
-rw-r--r-- 1 root root 10416 Sep  7 18:59 speedstep-lib.ko

Nothing is logged:

Code:

root@cd:/var/log# grep cpufreq *
root@cd:/var/log#

tux111 · 10-07-2012, 05:25 AM

Quote:

Originally Posted by cascade9

Is it really a mobile Core 2 Quad 9000? They should have 6MB (2 x 3MB) L2 cache, not 2MB.

http://ark.intel.com/products/40480

It's a Q8300 with 4 MiB. This is lshw output:

Code:

     *-cpu
          Beschreibung: CPU
          Produkt: Intel(R) Core(TM)2 Quad CPU    Q8300  @ 2.50GHz
          Hersteller: Intel Corp.
          Physische ID: 4
          Bus-Informationen: cpu@0
          Version: Intel(R) Core(TM)2 Quad CPU    Q8300  @ 2.50GHz
          Steckplatz: Socket 775
          Größe: 2498MHz
          Kapazität: 4GHz
          Breite: 64 bits
          Uhr: 333MHz
          Fähigkeiten: x86-64 fpu fpu_exception wp vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx constant_tsc arch_perfmon pebs bts rep_good nopl aperfmperf pni dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm sse4_1 xsave lahf_lm dtherm tpr_shadow vnmi flexpriority cpufreq
          Konfiguration: cores=4 enabledcores=4 threads=4
        *-cache:0
             Beschreibung: L1 Cache
             Physische ID: a
             Steckplatz: Internal Cache
             Größe: 32KiB
             Kapazität: 32KiB
             Fähigkeiten: synchronous internal write-back
        *-cache:1
             Beschreibung: L2 Cache
             Physische ID: b
             Steckplatz: External Cache
             Größe: 4MiB
             Kapazität: 4MiB
             Fähigkeiten: synchronous external write-back

Thank you for all the replies so far. I'm looking forward to getting ahead with this issue.

TobiSGD · 10-07-2012, 07:27 AM

OK, something weird is going on here:
1. The output of lscpu and lshw show different CPUs.
2. If lshw is correct Dell thought it would be a good idea to put a desktop CPU with 95W thermal design power into your laptop.
3. If the acpi-cpufreq module doesn't load modprobe should throw out an error message, especially with the -v option.
4. The OP states that the module is available, although it isn't on the output of the ls command that should show it.

tux111 · 10-07-2012, 09:00 AM

Quote:

Originally Posted by TobiSGD

OK, something weird is going on here:

That's interesting.

Quote:

1. The output of lscpu and lshw show different CPUs.

The posted output is true. I've just verified this.

Quote:

2. If lshw is correct Dell thought it would be a good idea to put a desktop CPU with 95W thermal design power into your laptop.

It's not a laptop. It's a desktop.

Quote:

3. If the acpi-cpufreq module doesn't load modprobe should throw out an error message, especially with the -v option.

Is there any way to narrow this issue down? Are there any similar commands to find evidence for a rootkit? If this behavior is caused by a rootkit at all. rkhunter and chkrootkit haven't found anything.

Code:

root@cd:/lib/modules/3.2.0-31-generic/kernel# modprobe -v nichtda
FATAL: Module nichtda not found.
root@cd:/lib/modules/3.2.0-31-generic/kernel# modprobe -v acpi-cpufreq
root@cd:/lib/modules/3.2.0-31-generic/kernel# lsmod|grep cpufreq
root@cd:/lib/modules/3.2.0-31-generic/kernel# find -iname *cpufreq*
./drivers/cpufreq
./drivers/cpufreq/pcc-cpufreq.ko

Quote:

4. The OP states that the module is available, although it isn't on the output of the ls command that should show it.

I was wrong. I thought p4-clockmod.ko pcc-cpufreq.ko and speedstep-lib.ko in the cpufreq folder would make up this module.

TobiSGD · 10-07-2012, 10:02 AM

Quote:

Originally Posted by tux111

It's not a laptop. It's a desktop.

My fault, sorry, Dell seems to have a laptop series with the same name also, so I mixed that up.

Quote:

Are there any similar commands to find evidence for a rootkit? If this behavior is caused by a rootkit at all. rkhunter and chkrootkit haven't found anything.

Why do you think that a rootkit is involved here?

I would try to run your system from a live-CD and look if the power savings work there (looking for the module also). If so I would think something is wrong with your installation. Is this by any chance a self compiled kernel?

tux111 · 10-07-2012, 11:25 AM

Quote:

Originally Posted by TobiSGD

Why do you think that a rootkit is involved here?

Could have been.

Quote:

I would try to run your system from a live-CD and look if the power savings work there (looking for the module also).

It seems to be a specialty of Ubuntu. The 12.04 Live-CD is also giving no output when loading the acpi-cpufreq module with verbose-switch. It isn't listed with lsmod.

Quote:

If so I would think something is wrong with your installation. Is this by any chance a self compiled kernel?

No, it isn't.

Do you have any other ideas howto fix this?

TobiSGD · 10-07-2012, 12:40 PM

Please post the output of

Code:

grep CPUFREQ /boot/config
cpufreq-info

I don't know if the second command is installed by default on Ubuntu, it may be possible that you have to install cpufrequtils.

tux111 · 10-07-2012, 12:47 PM

Quote:

Originally Posted by TobiSGD

Please post the output of

Code:

grep CPUFREQ /boot/config
cpufreq-info

Code:

root@cd:~# grep CPUFREQ /boot/config-3.2.0-31-generic 
CONFIG_X86_PCC_CPUFREQ=m
CONFIG_X86_ACPI_CPUFREQ=y

Code:

root@cd:~# cpufreq-info
cpufrequtils 007: cpufreq-info (C) Dominik Brodowski 2004-2009
Bitte melden Sie Fehler an cpufreq@vger.kernel.org.
analysiere CPU 0:
  Treiber: acpi-cpufreq
  Folgende CPUs laufen mit der gleichen Hardware-Taktfrequenz: 0
  Die Taktfrequenz folgender CPUs werden per Software koordiniert: 0
  Maximale Dauer eines Taktfrequenzwechsels: 10.0 us.
  Hardwarebedingte Grenzen der Taktfrequenz: 2.00 GHz - 2.50 GHz
  mögliche Taktfrequenzen: 2.50 GHz, 2.00 GHz
  mögliche Regler: conservative, ondemand, userspace, powersave, performance
  momentane Taktik: die Frequenz soll innerhalb 2.00 GHz und 2.50 GHz.
                    liegen. Der Regler "ondemand" kann frei entscheiden,
                    welche Taktfrequenz innerhalb dieser Grenze verwendet wird.
  momentane Taktfrequenz ist 2.00 GHz  (verifiziert durch Nachfrage bei der Hardware).
  Statistik:2.50 GHz:5,76%, 2.00 GHz:94,24%  (5651)
analysiere CPU 1:
  Treiber: acpi-cpufreq
  Folgende CPUs laufen mit der gleichen Hardware-Taktfrequenz: 1
  Die Taktfrequenz folgender CPUs werden per Software koordiniert: 1
  Maximale Dauer eines Taktfrequenzwechsels: 10.0 us.
  Hardwarebedingte Grenzen der Taktfrequenz: 2.00 GHz - 2.50 GHz
  mögliche Taktfrequenzen: 2.50 GHz, 2.00 GHz
  mögliche Regler: conservative, ondemand, userspace, powersave, performance
  momentane Taktik: die Frequenz soll innerhalb 2.00 GHz und 2.50 GHz.
                    liegen. Der Regler "ondemand" kann frei entscheiden,
                    welche Taktfrequenz innerhalb dieser Grenze verwendet wird.
  momentane Taktfrequenz ist 2.00 GHz  (verifiziert durch Nachfrage bei der Hardware).
  Statistik:2.50 GHz:5,15%, 2.00 GHz:94,85%  (4967)
analysiere CPU 2:
  Treiber: acpi-cpufreq
  Folgende CPUs laufen mit der gleichen Hardware-Taktfrequenz: 2
  Die Taktfrequenz folgender CPUs werden per Software koordiniert: 2
  Maximale Dauer eines Taktfrequenzwechsels: 10.0 us.
  Hardwarebedingte Grenzen der Taktfrequenz: 2.00 GHz - 2.50 GHz
  mögliche Taktfrequenzen: 2.50 GHz, 2.00 GHz
  mögliche Regler: conservative, ondemand, userspace, powersave, performance
  momentane Taktik: die Frequenz soll innerhalb 2.00 GHz und 2.50 GHz.
                    liegen. Der Regler "ondemand" kann frei entscheiden,
                    welche Taktfrequenz innerhalb dieser Grenze verwendet wird.
  momentane Taktfrequenz ist 2.00 GHz  (verifiziert durch Nachfrage bei der Hardware).
  Statistik:2.50 GHz:3,57%, 2.00 GHz:96,43%  (2945)
analysiere CPU 3:
  Treiber: acpi-cpufreq
  Folgende CPUs laufen mit der gleichen Hardware-Taktfrequenz: 3
  Die Taktfrequenz folgender CPUs werden per Software koordiniert: 3
  Maximale Dauer eines Taktfrequenzwechsels: 10.0 us.
  Hardwarebedingte Grenzen der Taktfrequenz: 2.00 GHz - 2.50 GHz
  mögliche Taktfrequenzen: 2.50 GHz, 2.00 GHz
  mögliche Regler: conservative, ondemand, userspace, powersave, performance
  momentane Taktik: die Frequenz soll innerhalb 2.00 GHz und 2.50 GHz.
                    liegen. Der Regler "ondemand" kann frei entscheiden,
                    welche Taktfrequenz innerhalb dieser Grenze verwendet wird.
  momentane Taktfrequenz ist 2.00 GHz  (verifiziert durch Nachfrage bei der Hardware).
  Statistik:2.50 GHz:4,98%, 2.00 GHz:95,02%  (4423)

Is it possible to stop this howling cpu-fan by modifying the CPUFREQ kernel option? It's really, really annoying.