Re: ACPI S1 state

[David Young <dyoung%pobox.com@localhost>]

On Wed, Jun 11, 2008 at 06:08:55AM -0400, Jared D. McNeill wrote:
> On Wed, 11 Jun 2008, David Young wrote:
> >If I do not put the devices to sleep, then the system hangs instead
> >of waking up.  It seems to have something to do with interrupts from
> >PCI devices.
> 
> Cute, but we can't run pmf handlers in this case as it introduces too 
> much latency, and is redundant since hardware state is preserved on S1 
> anyway. Is there a way we can suspend all interrupt handlers before 
> sleeping?

Let me unpack some of the power-saving issues that I think are involved
here:

1 I want to conserve as much battery as possible when the machine
  is idle.

2 I don't want to shut down the machine when it is idle.

3 S1 is the only sleep state that the BIOS on my laptop supports.
  (S4 *may* work, but I'm not holding my breath.)

4 Slow as waking all of the devices may be, it only takes about
  one second for the system to become interactive.  It only
  takes a few seconds longer for the HDD to wake.

5 If there is a strict wake latency requirement for S1, I am not aware
  of it.  As best I can tell, it need only wake faster than S2 and S3,
  and neither of those is an option.

6 Compliance with ACPI guidelines/standards is important, but when
  you get down to it, users want something that can meet both their
  power-saving and latency requirements.

Dave

-- 
David Young             OJC Technologies
dyoung%ojctech.com@localhost      Urbana, IL * (217) 278-3933 ext 24