kern/124744: [acpi] [patch] incorrect _BST result validation for Tosh Satellite P20

[Andriy Gapon]

I think that the patch is good and I have slightly extended it:
http://people.freebsd.org/~avg/acpi-bat.diff

Part of the problem is that the same definitions are used for interpreting status
returned by _BST and for maintaining internal driver status.
I have added a comment that explains this duality, retired now unused
ACPI_BATT_STAT_MAX, added code for cleaning extended/undefined bits in _BST status
and added a warning about charging+discharging bits being set at the same time.

I'd appreciate reviews and testing.
Thanks!

-- 
Andriy Gapon