Re: port-amd64/58344: Machine hangs at end of shutdown -r

[RVP <rvp%SDF.ORG@localhost>]

On Mon, 17 Jun 2024, RVP wrote:

> > ACPI: RSDP 0x00000000000FA240 000024 (v02 ACPIAM)
> > ACPI: XSDT 0x00000000DFEA0100 00008C (v01 SMCI            20180419 MSFT 00000097)
> > ACPI: FACP 0x00000000DFEA0290 0000F4 (v03 041918 FACP1644 20180419 MSFT 00000097)
> > Firmware Warning (ACPI): 32/64X length mismatch in FADT/Gpe0Block: 64/32 (20190405/tbfadt-642)
> > ACPI: DSDT 0x00000000DFEA0640 005A69 (v01 1CD11  1CD11000 00000000 INTL 20051117)
> > ACPI: FACS 0x00000000DFEB2000 000040
> > [...]
> >        instead of rebooting. It will then require a power cycle,
> >        which is most awkward when I happen to not be on site.
>
> Can you try the other way of inducing the triple-fault?
> Change the `#if 0' to `#if 1' and move it after x86_reset():
>
> https://github.com/NetBSD/src/blob/netbsd-9/sys/arch/amd64/amd64/machdep.c#L1964

Since your FADT doesn't look like it's being fully parsed, a more effective
workaround, I think, would be to just not call acpi_reset() in x86_reset():

https://github.com/NetBSD/src/blob/netbsd-9/sys/arch/x86/x86/x86_machdep.c#L1005

-RVP