> It does, over time. I can see the sensor files moving to be class > attributes instead of device attributes. But we need to do that so we > don't break all of the userspace code at the same time :) Yes, fair enough :). That's basically the point Mark made too, and I can understand, the userspace interface hasn't exactly been stable for lm-sensors 2.6. I really like the hwmon direction overall though, and hopefully the ipmi sensors driver will be an ideal example of what the hwmon patch is all about. I keep encountering little hurdles to jump over though on the way.. Thanks, Yani