Laptop ACPI question

[Kevin Oberman]

> From: Peter Schuller <peter.schuller at infidyne.com>
> Date: Sun, 2 May 2004 23:41:47 +0200
> 
> > Actually, ACPI will greatly improve battery life soon, but not yet. The
> > bits and pieces are being fed into CURRENT and I suspect that SpeedStep
> > support will be coming soon.
> 
> Cool! I always wondered what was "missing", since I get about 50% longer 
> battery life in Windows than in FreeBSD on my laptop (T40p).
> 
> I read somewhere that someone mentioned something about PCI power save 
> modes... would this be part of SpeedStep? Will SpeedStep support bring power 
> consumption to the levels one gets in Windows, or is there additional power 
> saving features on modern laptops that need to be supported?
> 
> (I'm asking even though it's not relevant to the original question, since you 
> seem to know a lot about this stuff, and I haven't found much on this 
> googling around.)

I am learning,  but I am WAY behind the experts. I just have more time
to answer questions. Hopefully the answers are sometimes correct. ;-)

PCI power modes are not SpeedStep but are important in power management.
They are related to ACPI, but not really a part of it, as I read the
spec. They are a part of the PCI spec and ACPI needs to provide control
information for these for proper PCI power control. This is very
important for things like suspend and resume.

SpeedStep, as I understand it, is adjustment of the CPU clock speed and
is VERY directly tied to battery life. It is what systems supporting
SpeedStep adjust when moving from AC to battery modes. E.g. my T30, by
default, runs at 1.8 GHz on AC and 1.2 GHz on battery (with my BIOS
configuration). SpeedStep is a marketing name and I am unwilling to say
that it is limited to the CPU speed. It may include some of the other
power management tools like deep sleep modes.

Support of CPU frequency is something that is actively being developed
and will almost certainly be in 5.3. PCI power mode support is also in
active development and was committed to CURRENT a few weeks ago, but at
least partly backed out due to problems on a few laptops. (It worked
GREAT on mine, though). I hope it is returned soon as it make my sound
work after a suspend.

The other item of importance to laptop users is turning off (not just
blanking) the display. It appears that this is on the verge of
commitment. I ran an early version of it and I could turn off the
display on suspend for the first time without the old APM stuff.

So power management is getting closer all the time and I am pretty
confident (for someone who is NOT writing the code) that it will all be
ready when V5 is declared STABLE.
-- 
R. Kevin Oberman, Network Engineer
Energy Sciences Network (ESnet)
Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab)
E-mail: oberman at es.net			Phone: +1 510 486-8634