cvs commit: src/etc/rc.d apm apmd archdep

[Doug Barton]

On Mon, 5 May 2003, Peter Wemm wrote:

> Mike Makonnen wrote:
> > On Mon, 5 May 2003 20:06:13 -0400 (EDT)
> > Robert Watson <rwatson at FreeBSD.org> wrote:
> >
> > >
> > > > Umm, I dont think so.  apm does not exist on amd64, cannot exist and
> > > > never will exist.  Please back this out.
> > >
> > > Well, ACPI does provide emulation for some of the apm(4) control ioctls,
> > > such that apm(8) actually usefully reports on the power condition on ACPI
> > > boxes.  I'd imagine the functionality provided by apmd(8) would be useful
> > > still with ACPI, but I don't know if it either directly speaks ACPI or
> > > whether the ACPI apm emulation is sufficient to support apmd.  If it does,
> > > then in theory the apm startup scripts could be useful on ACPI-supporting
> > > platforms, including ia64 and AMD64.  If that's the case, we might
> > > seriously think about renaming apmd to powerd or the like.
> > >
> >
> > Should this still be backed out ?
> > FWIW, everything that I ported over to rc.d *was* in rc.amd64.
>
> Robert is incorrect.  Only the i386 implementation of acpica provides
> the APM API emulation.  The include files do not exist on the other platforms.
>
> ACPI provides its own ioctl() interface (in addition to the sysctl stuff),
> which is cross platform.
>
> Yes, it should still be backed out.  And so should the ibcs2_compat thing.
>
> rc.i386 was repocopied to rc.amd64 without changes.  Thats where this mess
> started from.

I share some responsibility for this too. I noticed the problem with the
rc.amd64 stuff, but I was too busy fixing my own mistake with rc.sendmail
to send a head's up to the -rc list, and when Mike jumped on what he
thought was the fix for this, I decided to hope it was the right thing to
do and let more experienced people speak on the topic.

Doug

-- 

    This .signature sanitized for your protection