ACPI refcount increasing

[Jung-uk Kim]

On Monday 26 March 2012 01:42 am, Poul-Henning Kamp wrote:
> I tried -current as of approx one day ago and was greeted with a
> kernel printf flooding the screen with something about a ACPI(?)
> mutex(?) refcount increasing.
>
> I were unable to divine any identifying info from the messages
> because the screen scrolled too fast and the message was longer
> than the width of the screen.
>
> I were unable to get a core-dump and had to reset the laptop
> (Lenovo T400s)

I am well aware of the issue and identified the regression was 
introduced in these changes:

http://svnweb.freebsd.org/base/head/sys/contrib/dev/acpica/components/namespace/nspredef.c?r1=231844&r2=233250&view=patch
http://svnweb.freebsd.org/base/head/sys/contrib/dev/acpica/components/namespace/nsrepair.c?r1=231844&r2=233250&view=patch
http://svnweb.freebsd.org/base/head/sys/contrib/dev/acpica/include/aclocal.h?r1=229989&r2=233250&view=patch
http://svnweb.freebsd.org/base/head/sys/contrib/dev/acpica/include/acnamesp.h?r1=229989&r2=233250&view=patch

I notified the upstream maintainers and they are working very hard to 
correct the issue ATM.  I'll update ACPICA as soon as they find a fix 
but please revert the changes for now if you are in hurry.

Sorry for the breakage,

Jung-uk Kim