> I disagree (about the "bad" part - it was a cut and paste for sure). > adm1021 was at the limit for number of chips in SENSORS_INSMOD_x. True. > Making a new driver is a prerequisite for having better detection > in sensors-detect and recommending the correct driver, > without the user having to edit sensors.conf. > The new driver matches the xeon datasheet - exactly one sensor. Not > two. But we still need to add support to sensors-detect. Since I had been adding extra checks to the MAX1617 (and LM84) detection, I'd expect Xeon sensors not to be misdetected as these, so adding a basic check for xeontemp (without the extra knowledge of the value stored in the "unused regs") with confidence 2 should be sufficent for now. -- Jean Delvare http://www.ensicaen.ismra.fr/~delvare/
