ACPI Temperature FAILS after suspend => FAN FAILS

R3DKn16h7 · 2012-12-11 13:45:55

I'm using arch full upgrade in testing and kde-unstable but the problem was present before upgrade to testing) and I currently have a problem with the ACPI Temperature monitor, which leads to a malfunctioning FAN.

After resuming a KDE suspended session I've noticed that the FAN didn't work at all, even if the CPU was heavily occupied. After some check I noticed that the data from the ACPI temperatureindicator was completely abnormal: instead of show the actual temperature, it showed just 0.00 degrees. That's why, I think, my FAN doesn't work after suspend (and after a reboot start to cool like a plane reactor).

Rebooting the problem is solved i.e. the fan start to spin and I get to go to normal temperature (50°C, 60°C).

It's a bit risky b/c some times I've gone up to 100°C and I think is not a good thing for my poor hardware.

Does anyone know what can be the problem and how to solve it? What should I do? Is a KDE bug or an arch bug?

Now I wan't to do some more test. For instance: is there any way to "suspend" the system outside of any desktop environment?

PS: please tell me what information is useful to try to understand the problem.

PPS: running on a Dell Laptop Sudio 17 with Intel CPU, arch x64 full upgraded to testing and KDE unstable (but the problem was present before with no testing and no KDE unstable).

More details:

lm_sensors temperature output doesn't show the same issue.

On a functioning session:

$ acpi -t
Thermal 0: ok, 61.0 degrees C
Thermal 1: ok, 52.0 degrees C
Thermal 2: ok, 58.0 degrees C
$ sensors -u
acpitz-virtual-0
Adapter: Virtual device
temp1:
temp1_input: 58.000
temp1_crit: 100.000
temp2:
temp2_input: 51.000
temp2_crit: 100.000
temp3:
temp3_input: 60.000
temp3_crit: 100.000
coretemp-isa-0000
Adapter: ISA adapter
Core 0:
temp2_input: 57.000
temp2_max: 105.000
temp2_crit: 105.000
temp2_crit_alarm: 0.000
Core 1:
temp3_input: 55.000
temp3_max: 105.000
temp3_crit: 105.000
temp3_crit_alarm: 0.000

After suspend:

$ acpi -t
Thermal 0: ok, 0.0 degrees C
Thermal 1: ok, 0.0 degrees C
Thermal 2: ok, 0.0 degrees C
$ sensors -u
acpitz-virtual-0
Adapter: Virtual device
temp1:
temp1_input: 0.000
temp1_crit: 100.000
temp2:
temp2_input: 0.000
temp2_crit: 100.000
temp3:
temp3_input: 0.000
temp3_crit: 100.000
coretemp-isa-0000
Adapter: ISA adapter
Core 0:
temp2_input: 56.000
temp2_max: 105.000
temp2_crit: 105.000
temp2_crit_alarm: 0.000
Core 1:
temp3_input: 54.000
temp3_max: 105.000
temp3_crit: 105.000
temp3_crit_alarm: 0.000

Last edited by R3DKn16h7 (2012-12-11 13:53:44)

teateawhy · 2012-12-11 17:05:02

Even if i do not know how to debug your issue, i suggest you to use manual fan control. Hopefully it will save your poor hardware until you found a proper solution.
https://wiki.archlinux.org/index.php/Fan_control

cfr · 2012-12-12 01:05:48

sensors *does* show the issue - look at the virtual device readings?

You can suspend outside of the DE if you've configured stuff that way e.g. if you have systemd handle it or whatever.

I doubt this is an issue with either KDE or Arch.

R3DKn16h7 · 2012-12-12 10:44:56

Even if i do not know how to debug your issue, i suggest you to use manual fan control. Hopefully it will save your poor hardware until you found a proper solution.
https://wiki.archlinux.org/index.php/Fan_control

Ty, but until then I simply don't suspend .

cfr wrote:

sensors *does* show the issue - look at the virtual device readings?

You mean sensors are broken and they fail only after a suspended session? Now I try to boot Windows and see if sensors behave the same.

It's lilke sensors doesn't wake up: either is linux that doesn't wake them up or they are broken.

cfr · 2012-12-13 02:18:09

What I meant was that you said that lm_sensors output didn't show the same issue but the output you posted from sensors suggests that it also shows the issue. I wasn't sure why you thought that output didn't demonstrate the same problem.

I don't mean the actual sensors get it wrong - the output seems right for those. But lm_sensors output for acpitz-virtual-0 (again, acpi) shows the same issue as acpi...

R3DKn16h7 · 2012-12-13 07:37:45

cfr wrote:

What I meant was that you said that lm_sensors output didn't show the same issue but the output you posted from sensors suggests that it also shows the issue. I wasn't sure why you thought that output didn't demonstrate the same problem.
I don't mean the actual sensors get it wrong - the output seems right for those. But lm_sensors output for acpitz-virtual-0 (again, acpi) shows the same issue as acpi...

Oh, yes. I said wrong, in fact at first I tought it was a problem of the output of acpid, then I used lm-sensors. What I noticed was that "coretemp-isa-0000" was working fine ad this is an output that I see only using lm_ sensors.

Actually I get 5 different temperatures from variuous points and I've no idea what is the difference between them.

Arch Linux

#1 2012-12-11 13:45:55

ACPI Temperature FAILS after suspend => FAN FAILS

#2 2012-12-11 17:05:02

Re: ACPI Temperature FAILS after suspend => FAN FAILS

#3 2012-12-12 01:05:48

Re: ACPI Temperature FAILS after suspend => FAN FAILS

#4 2012-12-12 10:44:56

Re: ACPI Temperature FAILS after suspend => FAN FAILS

#5 2012-12-13 02:18:09

Re: ACPI Temperature FAILS after suspend => FAN FAILS

#6 2012-12-13 07:37:45

Re: ACPI Temperature FAILS after suspend => FAN FAILS

Board footer