Re: Fan control on ibook2
I've done my adm103x module. Thanks a lot to Colin and Frank. Without
them it would have been much longer. I first tried to integrate the
driver in the adt7....c module, according to Colin's patches. But I
believe the two chips behaviour does not fit well in a common module.
On the ibook, the fan must be activated very sparsely, and the adm chip
as an integrated behavior (histeresis, fan motor kickoff, and
proportional control which is really sufficient). The only thing to
parameterize is at which temperature the fan starts and the slope of the
So, on the contrary of adt module, I didn't daemonized the module, but I
only provide a sysfs interface to look at measured temperatures and
current parameters, and also to set them.
For instance :
T:51°C S:56°C R:10°C
T:48°C S:76°C R:10°C
#echo "56 10" > /sys/device/temperature/info0
make the fan starts at 56 degrees and accelerate progressively till max
at 66 degrees (56 + 10), on the sensor 0. After boot, my ibook is
parameterized at 72 degrees and a slope of 5 !
I don't think manual control of the fan speed is needed, and anyway, the
motor torque seems not sufficient to have any reactivity in some PID
Now the question is : should this be integrated (or proposed) in the
kernel ? What should I do to distribute it ? I'm waiting instructions !
On Wed, 17 Mar 2004 14:45:00 +0100,
Colin Leroy gracefully wrote:
>>>No, it just seems to :) You have a 1.0 load average, that's it ? this
>>>is due to TASK_UNINTERRUPTIBLE, which is needed according to
>Benjamin;>>however this doesn't eat the cpu power :)
>>Hmmm. Isn't it a problem in load accounting ?
>>I'll try this.
>Here's a third patch, which factorizes things a bit more and isn't
>reversed as the second one was.
>>>Did you look at the values in /sys/devices/temperatures ? Do they
>>Seems so. 53 and 56 °C
"[Of course] I'm French! Why do think I have this outrageous
accent, you silly king-a?!" Monty Python and the Holy Grail