ACPI and kernel patching (was: I fried my Sony Vaio PCG-FXA678)
On Thu, May 01, 2003 at 09:57:31AM -0700, Jaye Inabnit ke6sls wrote:
> Next, and this is what surprised me, is that when I installed the ACPI
> functionality, I seem to have lost the fan support. I am not wording this
> well. The fan *never* turns on now. I think it involves the system reading
> the internal temp incorrectly. I have not yet dug deeply into the ACPI
> module that controls the temp but I think I need to find out how I can
> manually LOWER the temperature settings so that my fan turns on sooner to
> keep the system safely cooled. Before I added ACPI, the system defaulted to
> APM. With APM, my fan stayed on ALL the time, but it was unable to monitor
> my battery or power state.
I recently got a Toshiba Satellite 5100 notebook, which, as far as I
learned, is "legacy free", i.e. it has no traditional BIOS which includes
not having APM. I installed Debian 3.0 (stable), and the fan on this
machine gets on and off (with kernel-image-2.4.18-686).
So, it seems that it is not generally such that the fan is dead with a
plain vanilla kernel.
I've heard, though, that it's at least possible to forcibly turn off the
fan and thereby cause a system overheat with ACPI. Personally, I'd say
that such would be a faulty design or implementation on the hardware
side -- hardware should generally survive even complete hang of the
OS without any permanent damage.
ACPI settings should enable the user to choose whether the system should
be fast and noisy or, alternatively, slower and silent, not to enable the
user to fry the system.
In the meantime, I have applied the toshiba ACPI kernel patch, which gives
me the corresponding /proc entry. In /proc/acpi/toshiba/fan/status, there
are two entries, "running" indicates whether the fan is on, and "force_on"
indicates whether the fan is permanently on due to user request. There is
no "force_off", and rightly so, I'd say.
One question regarding patching the kernel: I tried to apply this patch
(http://memebeam.org/toys/ToshibaAcpiDriver) the "Debian way", by setting
up stuff in /usr/src/kernel-patches etc., but I found this rather tedious
and eventually gave up and just patched the source manually. Are there
any auxiliary tools for generating the scripts in the "apply" and "unpatch"
subdirectores, or something like this?
Greetinx, Jan
--
+- Jan T. Kim -------------------------------------------------------+
| *NEW* email: kim@inb.uni-luebeck.de |
| *NEW* WWW: http://www.inb.uni-luebeck.de/staff/kim.html |
*-----=< hierarchical systems are for files, not for humans >=-----*
Reply to: