On 07 Feb 2003 13:52:16 -0500 "Perry E. Metzger" <perry@piermont.com> wrote:

> 
> "Steven M. Bellovin" <smb@research.att.com> writes:
> > I know there's talk of the "right" way to talk to the acpi drivers, but 
> > even when something like that is finished, I think we'll still need 
> > something that provides the same API as /dev/apm, just to preserve all 
> > the tools that know how to use it.  (I'd love to try acpi on my laptop, 
> > but if I can't even check the battery status it's a non-starter.)
> 
> There isn't really that much infrastructure that uses APM, and it
> appears ACPI needs a lot more infrastructure than APM provides. In
> particular, ACPI requires that the OS handle things like powering down
> idle devices, so ultimately our user tools will need to permit
> registration of power management policies. This goes far beyond the
> tiny amount of code in apmd etc.
> 
> Perry

I think we're in one of those 'negative sweet spots' right now, where apm/apmd
provides some behaviours laptop users want (battery monitoring, power
management options) and ACPI provides other features just as interesting
(indicated for some cardbus device recognition problems on some laptops, fan
management, longer-term architecture) and you can't have both at the same
time.

YMMV. I think because of the cardbus thing on my Dell L400, I'll forgoe the
apmd power management. However I'm a bit worried that since I did ACPI I've
seen the return of phantom halts which in FreeBSD were put down to CPU
overheating.(eg during build.sh)

cheers, thanks for confirming status

	-George