On Tue, Oct 17, 2017 at 07:00:16PM +0200, Caspar Schutijser wrote:
> On Sat, Oct 14, 2017 at 08:17:17PM +0200, Caspar Schutijser wrote:
> > >How-To-Repeat:
> > 1) Turn on the Lenovo ThinkPad P51 and boot OpenBSD.
> > 2) Unplug the power cable.
>
> In the meantime I installed a BIOS update from Lenovo but that did not
> help.
>
> I decided to boot a Linux live environment (Xubuntu 17.04/amd64 with
> Linux version 4.10.0-19-generic) to see what happens on systems with
> another ACPI implementation. When I unplug the power cable while Linux
> is running, I see the following:
>
> [ 45.297388] ACPI Exception: AE_AML_PACKAGE_LIMIT, Index (0x000000005) is
> beyond end of object (length 0x5) (20160930/exoparg2-427)
> [ 45.297435] ACPI Error: Method parse/execution failed
> [\_SB.PCI0.LPCB.EC.BRNS] (Node ffff91242fcfaf50), AE_AML_PACKAGE_LIMIT
> (20160930/psparse-543)
> [ 45.297547] ACPI Error: Method parse/execution failed
> [\_SB.PCI0.LPCB.EC._Q27] (Node ffff91242fcfa8e8), AE_AML_PACKAGE_LIMIT
> (20160930/psparse-543)
>
> That looks similar to what OpenBSD is reporting. Is it reasonable to
> assume that Lenovo made a mistake here with the ACPI table?
Seems like it; there have been quite a few broken ACPI tables and
firmware implementations out there across various vendors/boards.
