On 9/1/07, Pavel Machek <pavel at ucw.cz> wrote: > > diff --git a/Documentation/hwmon/sysfs-interface b/Documentation/hwmon/sysfs-interface > > index b3a9e1b..da546ce 100644 > > --- a/Documentation/hwmon/sysfs-interface > > +++ b/Documentation/hwmon/sysfs-interface > > @@ -304,6 +304,21 @@ curr[1-*]_input Current input value > > Unit: milliampere > > RO > > > > +********* > > +* Power * > > +********* > > + > > +power[1-*]_input Current power use > > + Unit: milliWatt > > + RO > > + > > +power[1-*]_max_input Historical maximum power use > > + Unit: milliWatt > > + RO > > + > > +power[1-*]_min_input Historical minimum power use > > + Unit: milliWatt > > + RO > > Should we add power?_10sec_avg_input? IIRC thinkpads export that, > too. Indeed, ThinkPads report both instantaneous power/current and a rolling average (exponentially decaying, I think) over the last ~10 seconds. ACPI provides only the rolling average, and the out-of-tree tp_smapi driver provides both. More generally, linux/power_supply.h defines these attributes: POWER_SUPPLY_PROP_VOLTAGE_NOW, POWER_SUPPLY_PROP_VOLTAGE_AVG, POWER_SUPPLY_PROP_CURRENT_NOW, POWER_SUPPLY_PROP_CURRENT_AVG, It would be nice if the power meter interface uses the same conventions as the power supply interface, since the former is essentially a subset of the latter. Userspace app that read power meters via sysfs should work for power supplies too. Shem
