[ltp] Re: Re: ACPI S3 works fine on T41 (not p)

[Volker Braun]

Hi,

On Thu, 26 Aug 2004 10:34:24 +0200, Alexander Gran wrote:
>> Interestingly, there is a LNUX name but this is nowhere used. The dsdt
>> also checks for _REV>=2, but every implementation probably satisfies that.
> 
> Hmm. It is a register/variable (or whatever..) like the ones where IBM stores 
> the windows version. No Idea why it's there..

Afaik _REV is the acpi version, and must be defined by every aml
interpreter. Nothing special here.

Concerning your state=2 vs. 3 suggestion, I read some of the recent lkml
posts. Yet another weirdness brought to you by the acpi commitee :-) But
converting the 2->3 does not help either (in suspend_device()
instead of device_suspend(), kernel 2.6.8.1). Suspend/Resume still works
fine, but power consumption is also unchanged.

The ultrabay is probably also a red herring. The led is on during
windows suspend-to-ram. Moreover, even if I remove the ultrabay device,
power usage stays at the usual 5W in S3/Linux.

I also tried to convert the state=2 to state=4, which is D3cold as far as
I understand it. Suspend/Resume still works somewhat, I
loose access to the ultrabay on resume. Does not improve the power
usage.

-Volker