> 
> So, I think pmap_kenter*() is pretty bogus, and should probably be
> nuked.
> 
In favour for what? One of the points with pmap_kenter is that it only
maps kernel pages, and this may simplify the function a lot compared
to pmap_enter. An other point is that it doesn't have to keep track of
pv_table conversions. This simplifies the functions a lot which actually
gives a great speedup. If something uses this function erroneously then
that should be fixed instead.

-- Ragge