Re: thermal_zone trip_point_0_temp 200°C

[Zhang Rui <rui.zhang@intel.com>]

On 一, 2012-06-04 at 10:41 -0400, Mark wrote:
> Hi Rui,
> 
> Thanks for writing;
> 
> On 06/03/2012 10:56 PM, Zhang Rui wrote:
> > Usually, the critical trip point value is a hard coded number provided
> > by the BIOS.

This is the ASL code for critical trip point in the ACPI table,

            Name (DCRT, 0x127C)

            Method (_CRT, 0, NotSerialized)
            {
                Return (DCRT)
            }

_CRT, the control method which OS evaluates to get the thermal critical
trip point, returns DCRT, which is a hard coded value 0x127C. And this
equals 473.2 K, or 200C.

So I do not see ACPI thermal does anything wrong here.

> 200° though, possibly the kernel should set a default upper limit, unless there's
> a rational reason

you can use module parameter thermal.crt= to override the critical trip
point.
But I'm not sure if kernel should set a default upper limit or not.
Maybe we need another entry for this laptop in thermal_dmi_table.

> > About fan control, it seems that there is no ACPI FAN on this machine,
> > so the fan may be controlled either by firmware or by some platform
> > specific driver.
> > To make a double check, it would be great if you can refer to
> > http://www.lesswatts.org/projects/acpi/utilities.php
> > to get the acpidump output of this machine.
> >
> > thanks,
> > rui
> 
> I'm tending to a similar conclusion, it looks as though that's generally the case
> with many Aspires; there is a script for poking the hardware registers
> http://code.google.com/p/aceracpi/source/browse/trunk/acer_ec/acer_ec.pl referred
> to at http://www.linuxquestions.org/questions/showthread.php?p=4657451 though it
> sounds a somewhat risky/random business without the manufacturers' specifications
> at the ready
> 
> there are 2-3 symbols in the ACPI that may be relevant, FANG, FANW, possibly
> FANU; I attach the relevant files
> 
FANG/FANW/FANU can be used for fan control?
I do not know what these mean as they are not ACPI pre-defined control
method.
But if all the Aspires machines follow the same rule, then maybe we need
a kernel Acer platform driver that handles this.

thanks,
rui