Salve Marco,
Hmm... that reminds me of the problem I had with a MacBook (port-i386/37305) in which the second core wouldn't start due to timing problems associated with USB legacy emulation stuff (timing calibration was botched due to some SMIs being generated by a "USB legacy" feature). But that was with NetBSD 4.0...
This is different. Here, not only the CPU would not start, but it does not get initialized neither. It is as if it was non-existant.
I think there is two different stages in an AP init. First one is to boot it, set it up in the correct mode (i386 or x64), init the VM system and park it into an idle lwp ("init"). Then the processor is integrated into the scheduler operation and activated ("running"). Correct me if I am wrong.
Buon finesettimana, VincentPS : I'm planning to buy the new MacBookPro due at mid-October. I guess NetBSD-current works correctly on MacBook, doesn't it?