A hint for my Core 2 Duo MP bug

[Vincent <10.50%free.fr@localhost>]

Hi there,
I am back with my bug (the second CPU of my Core 2 Duo CPU won't start on
-current).

I've tried today a 4.0.0_PATCH amd64 kernel. Although the userland start crashes
(why? I cannot get to the end of the init sequence), I clearly have a message
"CPU 1 started", showing that the second core actually runs on 4.0.

So, my question is: what is it, related to low level starting of cores, that has
changed between 4.0 and -current?

Thanks,
Vincent