An update on the SMP work..

After some debugging, the "big lock" and i386-specific SMP code is
starting to stabilize.  Staying up for a "make -j4" of the kernel, or
through an amanda backup of the system, is no longer remarkable,
though there are still a few occasional pesky system-crashing bugs
lurking here and there.

Jason is in the process of merging the largest set of changes into
-current -- namely, a change to how the scheduler and run queues are
locked; once that's in, we'll then work on committing the
machine-independant parts of the "big kernel lock", at which point all
the relevant MI code will be in -current.

					- Bill