ACPI power-down error on Toshiba M40 laptop

[Wayne Chapeskie]

I have a Toshiba M40-SF3 laptop that has been running various FreeBSD
versions going back to 5.4.  It has been running with ACPI disable,
since ACPI support has been gradually getting better, but is not
quite working yet.  The machine's ACPI BIOS was blacklisted until
6.2; 6.2 tried to enable ACPI by default, but the hard disk was not
recognized. After the major ACPI update earlier this year,
7-CURRENT snapshots would mostly work with ACPI turned on.
The machine normally runs 6.2, with ACPI turned off; there is a spare
partition for testing 7-CURRENT snapshots.

With ACPI enabled under the latest 7-CURRENT snapshot, the system boots,
devices are recognized, and the system runs without noticable
problems.  From either single-user or multi-user, shutting the system
down does not work properly:
    - shutdown -r: works properly
    - shutdown -h: halts the system normally; pressing a key to reboot
      works; the power button does not work--it is necessary to remove
      the battery to shut the machine off
    - shutdown -p: shuts down the operating system, but does not turn
      the power off; the power button does not work--it is necessary
      to remove the battery to shut the machine off

In the shutdown -p case, verbose logging shows an acpi error message
near the point where power would normally be turned off:

    Uptime: 7m28s
    acpi_lid0: wake_prep disabled wake for \_SB_.LID0 (S5)
       ... more wake_prep disable messages ...
    acpi: bad write to port 0x080 (32), val 0x55
    Powering system off using ACPI

DSDT tables, ASL output, and verbose dmesg output can be found at
http://www.spinnaker.com/~waynec/tech/toshiba-m40/acpi.html

I have only recently been able to configure this machine with enough
extra space to run a proper testing/debugging partition for -CURRENT,
and so have not submitted PR's for this, and a couple of other
problems with hardware support on this machine.  I'm not quite sure
what the proper procedure is for including large supplementary
information such as DSDT's and verbose dmesg output in a PR.

Thanks,

-- 
Wayne Chapeskie