Hi, thanks. I'll send your info to the team for review. Phil Peter Oehry wrote: >Hello Philip > >I updated the w83627hf driver with some code to support the QFan options >of the w83627thf chip. I think, if we have a Chip, that is able to >control the fans, it's better to let the chip do this work rather than >software. >So far my updated driver can control nearly all qfan settings through >some new sysfs nodes. It does the work perfect for me. I was only able >to test the settings for fan2 as the MB (ASUS P4P800) so far as I now >has only the cpu-fan-pwm connected to the chip. > >I don't know if there is another driver in the lm_sensors, that can >already do this. The problem of getting this things into the lm_sensors >project might be de discussion of how shoud the new nodes in the sysfs >be named an how do they behave. I have now idea how many chips the >ability of qfan have and how different they are to controll. > >So far I created the following nodes: >They correspond direct with the datasheet of the w83627thf > >qfan[1-3]_mode Mode are definde as follows: > 0 = manual Mode > 1 = thermal cruise > 2 = fan speed cruise > 3 = reserved no function > offset of 4 controls duty-Bit (if set fan does > not turn off) >qfan[1-3]_start When at Thermal Cruise mode, fan_pwm will > increase from 0 to this register value to prived > a minimum value to turn on the fan. >qfan[1-3]_stop minmal value for fan_pwm (bevore it turns of) >qfan[1-3]_stop_time defines the time from stop value to 0 in 0.1s >qfan[1-3]_target target temperature in ?C >qfan[1-3]_tolerance +/- tolerance for target temperature >qfan_step_down determines the speed of fan_pwm decreasing the > speed, when temperature is below > target-tolerance (Unit is 1.6 seconds) >qfan_step_up determines the speed of fan_pwm increasing the > speed, when temperature is above > target+tolerance (Unit is 1.6 seconds) > >the good thing about all these changes is that all other functions >(except that it's not possible to change the pwm_value during cruise >mode) remains the same. So any front-end application can use all the old >functions as bevore. > >Let me know if you're interested in my work. I can send you the driver >as it is or make a diff against my source. It might need som cosmetic >work bevore I can give you the driver. I thought I'll do this only if >you're interested. I can also help in furhter development or testing. By >the way I worked with the driver I got from the latest Kernel patch >(2.6.5-rc2). I am stil using a 2.6.1 Kernel patched with some drivers. > > > >
