> No really, what could possibly set hw.setperf besides sysctl > (which I do not call) and apmd (which is not running)? And > where does the number 5 come from? > > Or, what obvious triviality have I overlooked? > Is hw.setperf meaningless when apm/acpi is disabled?
The kernel is manipulating this. I've already told the guilty parties that they are wrong in doing so.