[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: suspend on laptop overheating



Also sprach Jochen Schulz <jrschulz@well-adjusted.de> (Mon, 21 Aug 2006
19:46:06 +0200):
> Tzafrir Cohen:
> >
> > Aug 21 00:02:40 localhost kernel: ACPI: Critical trip point
> > Aug 21 00:02:40 localhost kernel: Critical temperature reached (73
> > C), shutting down.
> -- snip
> > I would like to change the command it runs to a script that will
> > basically run:
> > 
> >   hibernate || shutdown
> 
> Hm, I only found section 8 at
> <http://www.columbia.edu/~ariel/acpi/acpi_howto.txt>. But his only
> describes how you can change the temperature at which your machine
> shuts down (if it works at all, didn't test it). This is probably
> neither what you want nor what is good for your laptop. But maybe it
> gives you enough words to google for.

I think that's right as it's hardcoded (in opposite to things done by
acpid where this behaviour could be changed in the /e/a/actions/
scripts).

Overheating is often caused by bad airflow inside (dust,..) or plain bad
design. I've also heard of additional RAM or changing to another/bigger
(sometimes getting hotter) HD triggering that.

I don't have a working temperature sensor on my notebook nor is it prone
to overheat. You could keep that trippoint and build a script to
throttle down the CPU if it gets near it - so there's no chance to
barbecue your processor or shutdown the box. Temperatures can be read
out from /proc/acpi/ or maybe sensorsd. For scaling there might be
/proc/acpi/processor//throttling or some so-called cpu-freq governors (I
assume you already use some tool for basic frequency-scaling, e.g. to
enhance batterylife.) There's http://acpi.sourceforge.net and
http://www.thinkwiki.org/wiki/How_to_make_use_of_Dynamic_Frequency_Scaling
providing some information. 

> J.

sl ritch



Reply to: