[Bug 233863] r345425 on PowerMac G5 may require kern.smp.disabled=1 and must set usefdt=1 which causes net interface reorder

[bugzilla-noreply at freebsd.org]

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=233863

--- Comment #7 from Mark Millard <marklmi26-fbsd at yahoo.com> ---
Created attachment 203622
  --> https://bugs.freebsd.org/bugzilla/attachment.cgi?id=203622&action=edit
Investigatory stand/powerpc/ofw/ofwfdt.c and sys/contrib/libfdt/fdt_rw.c
patches

I was asked to make these patches available here.

They are, in part, tied to working around old
PowerMac openfirmware-node-tree oddities, while not
invalidating contexts that follow standards better
They are also tied to FreeBSD-code-assumptions about
relative node positions in the PowerMac trees that
the prior openfirmware->fdt translation did not
preserve.

Be warned that nothing I've done has allowed the
2-socket/1-core-each PowerMac7,2 example to boot to
any visible activity after the "Kernel entry at"
message in usefdt mode. The improvements were
observed on a 2-socket/1-core-each G4 PowerMac3,6
and a 2-socket/2-cores-each G5 PowerMac11,2 .

Other hacks and patches are involved in my
environment but they are not included here.

The 2 patches here are not coded in a form for
acceptable check-in to FreeBSD's code base. As
stands, they are just tools for use in my
investigations. They might form the basis for
official patches if the relevant folks classify
direction of the changes appropriate.

See https://lists.freebsd.org/pipermail/freebsd-ppc/2019-April/010268.html
and any later messages for more context.

-- 
You are receiving this mail because:
You are the assignee for the bug.