the latest pool pull-ups seem to have tripped a bug that makes
GENERIC.MP fail...  

[by hand]

...
202 TLB shootdown pool entries; page size 4096

kernel: page fault trap, code=0 at uvm_km_kmemalloc + 0xf, cmpl $0xc02fce60, 0(%eax)

uvm_km_kmemalloc (0, c02c8580, 1000, 1, ca)
uvm_km_alloc_poolpage (0, c02c8580, 0, bfc0000, ca)
pool_page_alloc (1000, 0, 22, c054ef9, c02fcdb0)
pool_prime (c02fcd40, ca, 0, c02b00c, c032d010)
pmap_init (c02ee900, c02ec900, c032ffa8,c013b85d, c0b7d000)
uvm_init (c032d010, 32d000, 334000, c00007ff, c049)
main (0, 0, 0, 0)

	Sean.