kmem(9): Create dtrace sdt probes for each kmem cache size.

The names of the probes correspond to the names shown in vmstat -m.
This should make it much easier to track down who's allocating memory
when there's a leak, e.g. by getting a histogram of stack traces for
the matching kmem cache pool:

  # vmstat -m
  Memory resource pool statistics
  Name        Size Requests Fail Releases Pgreq Pgrel Npage Hiwat Minpg Maxpg Idle
  ...
  kmem-00128  256    62242    0        0  3891    0  3891  3891    0  inf    0
  ...
  # dtrace -n 'sdt:kmem:*:kmem-00128 { @[probefunc, stack()] = count() }'
  ^C

When there's no leak, the allocs and frees (probefunc) will be roughly
matched; when there's a leak, the allocs will far outnumber the frees.

(riastradh)