- remove restriction on OFW kernel allocations being 4M

- remove restriction on OFW kernel allocations being 4M
- shuffle memory range following kernel to the beginning of phys_avail
- have the direct area use 256MB pages where possible
- remove dead code from the end of pmap_bootstrap
- have pmap_alloc_contig_pages check all memory ranges in phys_avail before
  giving up

- informal benchmarking indicates a ~5% speedup on buildworld