detecting hibernation capability

[Richard Hughes]

On Tue, 2005-11-29 at 17:40 -0500, David Zeuthen wrote:
> On Tue, 2005-11-29 at 22:11 +0000, Richard Hughes wrote:
> > (bool) power_management.can_suspend
> > (bool) power_management.can_hibernate
> > (bool) power_management.can_standby
> 
> Not sure we really want to make a distinction between suspend/standby
> (not many ppl know the difference between ACPI S1 and S3).

Sure, drop the standby.

> > 
> > I think we need to expose these keys (rather than just returning invalid
> > from the scripts) so GUI programs can hide the options that do not
> > apply.
> 
> So take a look at what will be Fedora Core 5. The vendor supplied kernel
> is configured on x86 to support ACPI and given the laptop has those
> capabilities we'll always set
> 
>  .can_suspend
>  .can_hibernate
> 
> to TRUE. However this will only work on about 80% of the systems
> actually deployed because of bugs (the laptop simply will refuse to
> either standby or resume). 

Yes agreed, we should assume it will work.

> Now, it turns out that we can't even white list this, I thought we could
> - at least this was what I was told here
> 
>   https://www.redhat.com/archives/fedora-maintainers/2005-March/msg00067.html

Hmm. I still like the blacklist idea, as some laptops are so broken its
unlikely hibernate would ever work... Still, with these keys you can
just match and merge with and fdi file.

> So in other words I seriously doubt how this is useful? Sure, it may be
> sorta useful to grey out stuff on some platforms; i'd say go ahead and
> add this if you have a patch.

For people running HAL on non-vendor kernels, or for vendor kernels that
do not support hibernation this makes sense. I'll try to cook something
up tonight.

Richard.