[ltp] disk-protect patch in linux kernel 2.6.28 => hdapsd works out of the box

[Henrique de Moraes Holschuh]

On Mon, 29 Dec 2008, Chris Penn wrote:
> I have just noticed that battery time tends to get diminished with
> hdapsd compared to say windows.  I am curious on any improvements that
> are being made.

There are none, Windows also has to poll the firmware at 50Hz or more,
just like Linux does (or you get too much latency and won't notice a
fall is imminent with enough lead time to unload the heads).

Windows with the Lenovo drivers does better power management than
Linux can right now, regardless of HDAPS.  Maybe that's what you're
observing?

There is, of course, a chance that the HDAPS firmware was MUCH
improved in the latest models from Lenovo and can now do the decision
making by itself and just report "park the heads!" interrupts, like
Apple and HP do (but do notice the Apple and HP APS systems are much
less adaptative than the one Lenovo provides).

Another possibility is that the windows HDAPS utility notices you have
an HD with APS sensors built-in, and switches to a monitor-the-HD
passive mode instead of a monitor-the-accelerometers active mode
unless you tell it to do something the built-in HD-based APS cannot
do.

Either way, if Windows is really doing better, you need to find
someone that is available for clean-room reverse-engineering work in
either the Windows drivers or the EC binary firmware, and get person
to send an email to LKML documenting what we'd need to do.

-- 
  "One disk to rule them all, One disk to find them. One disk to bring
  them all and in the darkness grind them. In the Land of Redmond
  where the shadows lie." -- The Silicon Valley Tarot
  Henrique Holschuh