> Uh, "sort of".  You only need a kernel stack *per process* because we always 
> store state on the state when we switch.  We can arrange (cf. later versions 
> of Mach) to not do so in many cases, and instead have a kernel stack per CPU 
> most of the time.  This would be especially beneficial for threads, as would 
> it remove 8k of per-LWP overhead.

for mach's "stack handoff" tricks, you need to switch address spaces and
stacks separately.  i guess it's actually harder with switch-by-TSS.

YAMAMOTO Takashi