pagezero() in the radix pmap was a plain store loop (bzero), which under
the kernel build flags (-mno-vsx -msoft-float) compiles to byte stores.
dcbz establishes a zeroed cache line directly in the cache without a
read-for-ownership fetch from memory, roughly halving the memory
transactions of page zeroing.

dcbz raises an alignment interrupt on caching-inhibited mappings, and
the kernel does not emulate it, so mmu_radix_zero_page() falls back to
bzero() for any page whose memattr is not the write-back default. The
internal pagezero() callers only touch freshly allocated page-table
pages, which are always write-back.

Only page zeroing is changed. dcbz helps solely because zeroing has no
source to read; for page copying it is a pessimization (a redundant
zeroing pass on top of the mandatory source read), so
mmu_radix_copy_page() is deliberately left as a plain bcopy().

Measured on POWER9 (Raptor Blackbird, DD2.3, bare metal), zeroing a cold
256 MB buffer with 128-byte scalar loops (median of 5, tight variance):

byte stores (what libkern memset compiles to)   8.6 GB/s
doubleword (std) stores                         26.8 GB/s
dcbz                                            34.6 GB/s

Cross-checked on two POWER9 pseries guests; both there hit a ~17 GB/s
memory-bandwidth ceiling that flattens std and dcbz, but neither ever
regresses. Fresh anonymous memory verified to read back byte-exact zero.