> I confess again to knowing very little about how the kernel handles
> memory...but....
> 
> Is it not true that at boot certain parts of the Firmware are expected to
> be at 000xxxxx but on Macs remain at FFFxxxxx?  If so, along with setting
> MSR[IP], is there anything stopping us from immediately just copying that
> memory to its new location before using it?

Hmm, just moving the content of memory most likely will not work.  It might
be possible to place stub vectors into the low core that immediately jump to
their high core counter parts to make up for this deficiency in the mac
firmware.  However, regarding this problem it's probably enough to just
simply reprogram the MSR[IP] bit as the system doesn't expect any traps
to the firmware once it expects any traps at all.  There might be a problem
if (note: I don't know enough about the Apple firmware here) the firmware
does expect to do some trap handling, especially interrupt handling, on
its own.

Note that this is by far not the only problem with the Apple firmware.  This
is anything but real OpenFirmware.  Judging from the reports I keep hearing
about it, it's a real PITA.  It doesn't do what a user of OpenFirmware would
expect from it.  Hey, (according to what I hear) it even isn't capable enough
to supply the hardware device tree (the description of the installed hardware
in a machine) to any program running under it, which is the primary reason to
use OpenFirmware to begin with.

Ciao,
Wolfgang
-- 
ws@TooLs.DE     (Wolfgang Solfrank, TooLs GmbH) 	+49-228-985800