hal: Branch 'master'

[David Zeuthen]

On Fri, 2007-03-30 at 19:37 -0400, David Zeuthen wrote:
> On Fri, 2007-03-30 at 23:39 +0200, Danny Kukawka wrote:
> > Hi,
> > 
> > Maybe someone could verify this, but on my IBM X41 and 2.6.21-rc4 I don't get 
> > any event for sleep via the input layer (cat /dev/input/event6 and xev show 
> > nothing for sleep). So this change would only lead to get no events for 
> > sleep.
> 
> Hardly something to do with this patch - the only change here is that we
> emit the event on the root computer device object... fwiw, none of this
> works for me on my T41; acpi-ibm refuses to load on my Fedora kernel.
> Ugh.

With an updated kernel, I got ibm-acpi to work and I can confirm that
events emitted from ibm-acpi, including sleep, are properly propagated
on HAL's computer object and that at least g-p-m handles that correctly.

However, I think something fishy is going on; sometimes ibm-acpi don't
emit events at all... but that's most probably a problem with
ibm.acpi.ko and/or the other things in the kernel.

     David