>Pardon me?  swapper, pagedaemon, and reaper (kernel-only processes) all
>share vmspace0/kernel_pmap.

Sure, but the mips still maps the kstack at UADDR, for the redzoning.
So the mips locore resume stuff is still doing the dance to invalidate
the TLB entries for the previous guy's kstack -- even if they were
both on vmspace0.  (At least I sincerely hope it is... Michael?)

Thats all I meant by "not fully".


The point was to look at sharing that same vmspace/pmap (and avoid the
pmap_activate() costs) for processes like nfs_iod(), yet give them
back `a vmspace of their own' if/when they return to userland.

That was a noticeable-to-users win on NFS servers, back when SunOS did
it. seems like it'd be the same for us too. yes?