[lm-sensors] dme177737: pwm not working

Juergen Bausa Juergen.Bausa at web.de
Sat Feb 26 15:32:18 CET 2011

Von: "Guenter Roeck" <guenter.roeck at ericsson.com>
Gesendet: 23.02.2011 23:23:58

>On Wed, 2011-02-23 at 17:12 -0500, Juergen Bausa wrote:
>> >Do you see any related output with dmesg ?
>> nothing:
>> jba at lisa:~$ dmesg | grep dme1737
>> jba at lisa:~$ dmesg | grep pwm
>> jba at lisa:~$ 
>> Juergen
>Should have noticed earlier: You other email shows 
>jba at lisa:/sys/class/hwmon/hwmon1$ ll
>insgesamt 0
>lrwxrwxrwx 1 root root    0 23. Feb 21:01 device -> ../../../ATK0110:00
>which indicates the loaded module is atk0110, not dme1737. Which afaik
>means that the device is detected as ASUS ACPI device. It loads the
>asus_atk0110 driver which does not support pwm attributes.
>Not sure what to do about that, especially since this suggests that
>ACPI claims access to the dme1737 registers, and parallel access may
>cause trouble.

Thanks for the tip, that helped a lot. Searching the web I found some cases, where the 
asus_atk0110 prevented other modules from working correctly. Mostly it was suggested to
add "acpi_enforce_resources=lax" to the kernel parameters. Allthough some people say this 
may be dangerous (others said it is not) I tried it and now the pwms are there again!

However, next problem was, that after each boot, hwmon1 and hwmon2 (dme1737 and 
asus_atk0110) may change. As a result, I was not able to write fixed adresses for the pwm in
the configuration file for the fan-controller. After blacklisting the asus_atk0110 this problem was 
also solved. As I dont need any of the outputs of the asus_atk0110 I dont think this is problem.
Most outputs are also given by the dme1737.

Could anyone please comment on this? Could it be dangerous to add "acpi_enforce_resources=lax"
to the kernel parameters? If so, what could happen? And, is there any reason to not blacklist the 
asus_atk0110 module?



Schon gehört? WEB.DE hat einen genialen Phishing-Filter in die
Toolbar eingebaut! http://produkte.web.de/go/toolbar

More information about the lm-sensors mailing list