> > Wow: > > efi0 at bios0: UEFI 2.0 > > that is ancient. I also found > > https://docs.oracle.com/cd/E26502_01/html/E28978/hardw.html > > so clearly the UEFI BIOS has bugs. Using UEFI instead of the legacy > BIOS on a machine that old may not be the wisest choice. But I think > we can just avoid using UEFI in the kernel in this case. > > Diff below should fix it. > > ok? > > > Index: arch/amd64/amd64/efi_machdep.c > =================================================================== > RCS file: /cvs/src/sys/arch/amd64/amd64/efi_machdep.c,v > retrieving revision 1.1 > diff -u -p -r1.1 efi_machdep.c > --- arch/amd64/amd64/efi_machdep.c 16 Oct 2022 15:03:39 -0000 1.1 > +++ arch/amd64/amd64/efi_machdep.c 19 Oct 2022 17:18:19 -0000 > @@ -112,6 +112,10 @@ efi_attach(struct device *parent, struct > printf(".%d", minor % 10); > printf("\n"); > > + /* Early implementations can be buggy. */ > + if (major < 2 || (major == 2 && minor < 10)) > + return; > + > if ((bios_efiinfo->flags & BEI_64BIT) == 0) > return; > > Index: arch/arm64/dev/efi_machdep.c > =================================================================== > RCS file: /cvs/src/sys/arch/arm64/dev/efi_machdep.c,v > retrieving revision 1.2 > diff -u -p -r1.2 efi_machdep.c > --- arch/arm64/dev/efi_machdep.c 12 Oct 2022 13:39:50 -0000 1.2 > +++ arch/arm64/dev/efi_machdep.c 19 Oct 2022 17:18:19 -0000 > @@ -118,6 +118,10 @@ efi_attach(struct device *parent, struct > printf(".%d", minor % 10); > printf("\n"); > > + /* Early implementations can be buggy. */ > + if (major < 2 || (major == 2 && minor < 10)) > + return; > + > efi_map_runtime(sc); > > /* >
Heh, yeah. :-) I figured I'd give it a go with UEFI on this old beast and it worked without that kind of issue until Monday of this week. I'll see how I can apply this diff if possible since I can't boot at all, unless it will be included in an upcoming snapshot. Thanks again. -Claudio
