svn commit: r247086 - head/sys/x86/isa

[Jung-uk Kim]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 2013-05-21 05:18:46 -0400, Bruce Evans wrote:
...
> So it seems that pmtimer has a lot of dead code and reduces to not 
> much more than an inittodr(0) and usually acts like the
> inittodr(!0) in the acpi amd64 resume code.

When I wrote the resume code for amd64, I studied the pmtimer and
realized the same problems, of course.  So, I simply reduced the code
and embedded it in acpi.c (without fixing underlying problems) because
a) I didn't want the dead code in pmtimer, b) acpi is mandatory for
amd64, and c) I didn't want to disturb i386 (or ia64) code paths.

In fact, my evil plan was a) deprecate apm and pmtimer for i386, b)
improve the code in acpi.c, and c) fix "calcru: runtime went
backwards" problem for good.  But I never found free time to do that. :-(

> acpi amd64 resume also seems to be missing the verbose logging of
> the suspension time, but maybe acpi utilities give that.

/etc/rc.{suspend,resume} already do that.  So, yes, the "feature" is
redundant.

Jung-uk Kim
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.20 (FreeBSD)

iQEcBAEBAgAGBQJRm662AAoJECXpabHZMqHO+fgH/iEGdzAXeZuG6F4+fGBk9wCi
nEyz/RoJeQ7OwdR8vVIcs3z9TwZT58QVOipt5Tgh6cr77fRWfTrBhdZEWDSInXcb
wpnFXeXNe0MVNHduuq3QIpmpHxJLHUMLKTdaBJhaVixATcQovYhMZUljq0AxVKDs
VJw346YrdXI3IaAzXbm1OonGNgV48N2ACXCJvOKv1qwtYhx9UMIUkbXoWxGkT6rb
i5hWj3VsVPBG/E4vvKPvZAwmieU7TaUY8/BOjyyWh4kAsRW5N4ZmRY9W0v5EyRid
5qDEfUrW6liVFh+4O4HOL3G0NDY6KCQpoBue91Dp5cOaHl3lzGlBUGQ9/9932nM=
=MldY
-----END PGP SIGNATURE-----