On Thu, Mar 04, 2004 at 10:42:17PM +1100, matthew green wrote:
>    
>     (1) CPU cross-calls
>     (2) TLB invalidation
>     (3) lazy FPU context switching
>     (4) getting the non-boot CPUs to enter the scheduler and run stuff
>     ... and probably more once we have initial versions of all the above.
>    
>    dennis and I have versions of (1) and (2)
> 
> 
> commitable?  for code #ifdef MULTIPROCESSOR, don't feel shy about
> adding it not 100% cooked :)

I'll see about committing what we've got so far this weekend.
it's really more like 10% cooked.  but as long as it doesn't break
UP kernels then any progress is good.

-Chuck