> Mac OS might not be setting this temperature, > so I don't know why OpenBSD should set it
Indeed, I belive MacOS doesn't set it, hence the interest in developing things like G4FanControl. If I'm not mistaken, OpenBSD manages the temperature manually towards a hardcoded state in smu(4), though I don't know if MacOS does the same. > The fans are quiet Yes indeed this diff makes them spin earlier, that is the whole idea of G4FanControl for example. I personally find it extremely unobstrusive - fans at lower speed are barely noticeable but kicking them earlier helps not needing them to ramp up like a rocket out of the sudden. > To reduce wear on this old PowerBook, I run apmd(8) -L I can't imagine. I daily drive this powerbook and having the cpu at half speed all the time would really make it much more difficult. I do run apm -L every now and then since for some workloads, although light, apmd tends to keep the cpu high. Right now, I get the following stats with this diff with a light workload: hw.sensors.adt0.temp0=44.00 degC (Remote) hw.sensors.adt0.temp1=41.00 degC (Internal) hw.sensors.adt0.temp2=45.00 degC (Remote) hw.sensors.adt0.fan0=2620 RPM hw.sensors.adt0.fan1=2816 RPM The fans are barely audible in this powerbook 5,9, cpu is at max speed. I would argue it's likely for these machines to have a high load given their limited resources, and thus worth keeping cool and healthy as much as possible.
