OK, that produces the same result:  a happy kernel!

I compared the full 'dmesg' output from the previous build and the
current build using diff:  the only differences were trivial --
return times from initcall functions, and the order in which some
initcalls were performed was different very late in the boot process.


Thanks Yinghai,
Dave W.
--

[ message continues ]

good, let wait for Ingo back to check the two patches too.

YH
--

[ message continues ]