Re: Fan control on ibook2

[Cedric Pradalier <cedric.pradalier@inrialpes.fr>]

Hi,

	I just tried your patch. It seems to work, at least for acces to the
chip, even with a joint driver. Some side effects : 

	First it sets a very high load on the CPU. I wonder if the task is
correctly scheduled. I'll have a look at that. It's as if the task
monitoring task was running without yielding the processor...

	Second, when rmmoding the module, it sets thermal limits to such a
low value (44), that fan becomes crazy, running at full speed.
(Measured temps are 53 and 56 °C).

	Third, when starting I got the following message in syslog. The
lowering max temp from 44 to 50 is funny !

kernel: therm_adt7467: no version for "struct_module" found: kernel
tainted.
kernel: adt746x: Thermostat bus: 1, address: 0x2c, limit_adjust: 0,
fan_speed: -1
kernel: adt746x: ADT7467 initializing
kernel: adm313x: Lowering max temperatures from 44, 44, 44 to 50, 50, 70
(°C)
kernel: adt746x: Manual setting of speed not supported on adm103x
kernel: adt746x: Setting speed to: automatic for CPU fan.
kernel: adt746x: Manual setting of speed not supported on adm103x
last message repeated 16 times
last message repeated 31 times
last message repeated 31 times
last message repeated 31 times
last message repeated 31 times
last message repeated 31 times
last message repeated 15 times
kernel: adm103x: Putting max temperatures back from 50, 50, 70, to 44,
44, 44, (°C)


Before I RTFM, I am not sure of the meaning of thermal limits : is it
the temperature at which fan are started ? 

Thanks

On Wed, 17 Mar 2004 11:24:02 +0100, 
Colin Leroy gracefully wrote:
>> >You are welcome to consolidate these into a single driver allowing
>> >userland access to those values & finer thermal control :) A bit
>> >like Colin Leroy did for the recent models ...
>>
>> I'll try to make some steps in this direction ASAP.
>> Thanks for the informations.
>
>
>For kernel 2.6, you can try this patch and maybe use it as a base. It
>adds support for ADM103x to the therm_adt746x driver. However it is
>completely untested, not even compiled, and lacks features such as
>manual control and fan speed measurement.
>If it doesn't work, any patch would be welcomed :)
>
>the chips docs can be found at
>http://www.analog.com/UploadedFiles/Data_Sheets/878926113ADM1030_a.pdf
>http://www.analog.com/UploadedFiles/Data_Sheets/27250527ADM1031_a.pdf
>
>In this patch i made a few assumptions about sensor locations. ADM1030
>has two (local and remote), i considered local=cpu and remote=gpu;
>ADM1031 has three (local, rem1, rem2), i considered rem1=cpu, rem2=gpu,
>as it is the case for iBook G4 and 2003 AlBooks.
>
>HTH,


-- 
Cedric

"[Of course] I'm French! Why do think I have this outrageous 
accent, you silly king-a?!"  Monty Python and the Holy Grail