DragonflyBSD added a new acpi_quirk for Notebooks whose DSDTs report a negative 16-bit rate. The quirk can most probably also applied to our apci implementation. Shortly before the DFly commit the affected files were sync'ed from FreeBSD. See http://freshbsd.org/commit/dfbsd/3bb5af08be0b4f2b68e64d7861a103a31a4af4f7 for the commit details.
I don't that notebook, but the patch applies cleanly, compiles on my HEAD notebook and the new kernel works. Link to the raw diff: http://freshbsd.org/commit/dfbsd/3bb5af08be0b4f2b68e64d7861a103a31a4af4f7/diff.txt
Reassigning to -mobile. Would someone with a commit bit please push this into -HEAD? I'm a little swamped atm.
It's hard to verify that the fix is correct unless someone has a busted laptop it can be tested on. Note that the patch adjusts bst[i].rate after it has been added into valid_rate which seems incorrect. (I suspect it should be adjusted before, not after.)
Reassign from non-responding assignee.
This patch is broken, the hardware it addresses is not common, and in two years no one has tested it.