On Mon, Feb 7, 2011 at 3:26 PM, Guenter Roeck <guenter.roeck@xxxxxxxxxxxx> wrote: > Hi Andrew, > > On Mon, 2011-02-07 at 14:53 -0500, Andrew Lutomirski wrote: > [ ... ] >> > >> > Anyway, please try again. I updated the code to fix the pwmX_enable bug. >> >> Mostly works. >> >> If I set pwm2_enable to 1, then pwm2 controls the fan. >> >> If I try to set pwm2_enable to 5, I get: >> >> # echo 5 >pwm2_enable >> bash: echo: write error: Invalid argument >> > For now that is on purpose. Problem is that if SmartFanIV mode was not > initialized by the BIOS, it probably won't work. There are lots of > parameters which have to be configured. So one can not just let users > enable it without forcing them to configure all the other parameters .. > which is currently not possible with the sysfs ABI. I guess what you'd want is a "set lots of parameters all at once" interface. > > So this is really a one-way street for now. > > Nuvoton's position on this matter is that fan control should be left to > the board vendor to prevent system damage, ie that users should not try > to configure it at all other than through the BIOS. They might have a > point. Fair enough. I actually plan to let BIOS configure everything, since it seems to work nicely. But as long as there is a driver that offers fan control, might as well have it work properly. (I care a lot more about the sensors than the fan control.) Should there be a "read-only" module option? --Andy _______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
