On Mon, Dec 17, 2007 at 02:19:55PM -0600, David Young wrote:
> device_is_active() never used to work the way that it works, now, and
> it is definitely concerned with device PMF states, now, in ways that
> I had to discover by experimenting.  I think that device_is_active()
> deserves to be documented in pmf(9).  Ditto device_has_power().

device_is_active is a function of the device state, just like
device_xname etc. It did exist before and the old behaviour wasn't
documented either.

Joerg