You are not logged in.

#1 2012-11-20 11:02:17

werto
Member
Registered: 2012-11-20
Posts: 4

[SOLVED] Fan and therm issue.

Hi, that's my first post in this forum and this is like my first week using Arch linux (which I'm loving).
First of all I'm sorry for my spagehtti english.

From when I'm using Arch there's an issue I didn't found solution: my fan is turning loudly all the time and the cpu temperature is something like 16/18C° higher than normally.

As title says i'm using a laptop with an Indel Core i5-2450M Cpu with an integrated Intel HD3000 graphic card.

Actually this is my pstree:

systemd─┬─2*[dbus-daemon]
        ├─dbus-launch
        ├─dhcpcd
        ├─login───bash───startx───xinit─┬─X
        │                               └─awesome─┬─conky───4*[{conky}]
        │                                         └─sh───sleep
        ├─2*[python2]
        ├─systemd-journal
        ├─systemd-logind
        ├─systemd-udevd
        ├─uzbl-core───6*[{uzbl-core}]
        ├─wpa_actiond
        ├─wpa_supplicant
        ├─xterm───bash───irssi───{irssi}
        ├─2*[xterm───bash]
        └─xterm───bash─┬─mutt
                       └─pstree

Usually the CPU therm was almost 52C°, since i'm using Arch it's almost 68C°. Now that i'm writing is 69,5C° and the fan is turning loudly all the time. I know that if I start using some more cpu the temperature will start growing to 72C° and more.
I also ear that the CPU adapt its turning to the temperature. Infact now it grows in speed a little and then go back to the speed it had before.

About the GPU I configured xorg /etc/X11/xorg.conf.d/20-intel.conf that way, I tried also sna, but nothing changes:

#/etc/X11/xorg.conf.d/20-intel.conf

Section "Device"
        Identifier      "Intel Graphics"
        Driver          "intel"
        Option          "AccelMethod"   "sna"
        #Option         "AccelMethod"   "uxa"
        Option          "TearFree"      "true"
EndSection

My cpupower governor is set "powerspace" and the actual frequency is set to 800MHz. The maximum is 2.50GHz.

The CPU usage is actually 2%.

Here's a portion of inxi:

CPU~Dual core Intel Core i5-2450M CPU (-HT-MCP-) clocked at 800.000 Mhz Kernel~3.6.6-1-ARCH x86_64 Up~19 min Mem~197.5/3867.0MB Procs~105 Client~Shell inxi~1.8.22  

This is my actual lsmod: http://nopaste.me/paste/112657216950ab6012206a5

I've also checked if there's something useful in the bios, but actually the only interesting options i've found are Intel Virtualization and Intel SpeedStep which are both active.

Does someone knows what's up?
Thank you :)

Last edited by werto (2013-01-31 10:37:03)

Offline

#2 2012-11-21 22:50:01

@op
Member
Registered: 2011-04-05
Posts: 49

Re: [SOLVED] Fan and therm issue.

Hi werto,
it seems that you're also affected by the current linux kernel power regression issues.
There are some possible fixes listed in the linked thread, like a patched kernel. As I'm affected too, I simply did a downgrade to kernel 3.5 and I'm running fine for now.

If you don't have kernel 3.5.x in your package cache, you can use the package downgrade form the AUR. Downgrade at least the linux and linux-headers package.
If you're having trouble downgrading, let me know.

Last edited by @op (2012-11-21 22:51:34)

Offline

#3 2012-11-22 09:54:24

werto
Member
Registered: 2012-11-20
Posts: 4

Re: [SOLVED] Fan and therm issue.

Hi @op,
thank you so much for your kind answer.

Yesterday I discovered that that happens most of time when the power feeder is plugged.
When I use it only with the battery it doesn't happen and the cpu's temperature goes slowly to 42 C°, which is awesome.

Unfortunately I don't know what happened yesterday night,
I was watching a movie and suddenly the laptop shutdown istantly.
The feeder was plugged in but the issue wasn't happening because I first powerup it without plug, then a minute after the boot I plugged the feeder in, as it worked as a solution to the issue (yes I know I could check the battery/plug policy, but it was late and I wanted to relax).

When the laptop shutdown I suddenly controlled if it was hot but it was totally fresh as breeze.
I've seen that the light of the power feeder, when it's not connected to the laptop is bright, but when I connect it to the laptop it goes weak and also if I disconnect it from the laptop it continues to be weak. I have to plug our from the laptop and from the electricity, than wait 5 seconds and replug to the electricity to turn it bright (until I don't plug it to the laptop);
it seems to me that there was a short circuit.

The laptop can't turn on in any way, with or without battery, with or without power feeder.

Fortunately the laptop is still in warrantee;
the guy from the assistance said that he thinks that happened for a voltage issue in my home electronics, but I don't think so.

Tomorrow they'll come taking it to the lab because fortunately it's in warrantee, so I can't do your suggestions; also that should take almost 10 days.

I don't think that happened because of the issue, this sound not so possible to me; I used laptop-utils, cpupower and acpid, all with default configurations and everything seemed okey in the last our before that happened.

Now I have to backup because of the stupid (my opinion) assistance policy.
Whish I was not so ot.
II'll let you know all the news I have o/

Offline

#4 2012-11-22 17:37:55

MadTux
Member
Registered: 2009-09-20
Posts: 553

Re: [SOLVED] Fan and therm issue.

werto wrote:

Yesterday I discovered that that happens most of time when the power feeder is plugged.
When I use it only with the battery it doesn't happen and the cpu's temperature goes slowly to 42 C°, which is awesome.
/

Charging the battery produces additonal heat. Additionally, the system may use a different CPU frequency governor when on battery.

werto wrote:

Unfortunately I don't know what happened yesterday night,
I was watching a movie and suddenly the laptop shutdown istantly.
/

Maybe the laptop was overheating? If the temperature reaches a certain point, most machines have a self-protection mechanism that prevents hardware damage and does an emergency shutdown.

Offline

#5 2012-11-23 09:31:53

werto
Member
Registered: 2012-11-20
Posts: 4

Re: [SOLVED] Fan and therm issue.

MadTux wrote:

Charging the battery produces additonal heat. Additionally, the system may use a different CPU frequency governor when on battery.

Yes, I've seen that. But I saw also that if I setup the CPU freq same as on battery when on power feeder not so much changed; I whish I wrote it correct/clear smile
As probably @op said it should be the linux kernel power regression issue.

MadTux wrote:

Maybe the laptop was overheating? If the temperature reaches a certain point, most machines have a self-protection mechanism that prevents hardware damage and does an emergency shutdown.

Yes, usually laptop shutdown because of overheating, but I think that's not the case, because:

werto wrote:

When the laptop shutdown I suddenly controlled if it was hot but it was totally fresh as breeze.

Thank you for you kind answer MadTux.

Offline

#6 2013-01-31 10:36:19

werto
Member
Registered: 2012-11-20
Posts: 4

Re: [SOLVED] Fan and therm issue.

Sorry for the long delay,
"fortunately" as op said it was the linux kernel regression issue.
So I only had to downgrade the kernel, following the guide op linked me.
The problem is not still solved in 3.6.11-1.
Thank you so much for you help.

ps: I'm actually at the 3.5.3-1. If you have to downgrade and you use VirtualBox, don't forget to downgrade it also using the A.R.M. and each other software still used to work under the 3.6 kernel.

Offline

Board footer

Powered by FluxBB