Add a small perl script to help mount BSD slices in Linux when
customizing a generic image by doing the following:

- Parse a BSD disklabel on a block device
- Map BSD partitions to their Linux device names
- Mount BSD root partition, parse its /etc/fstab to find
  which partitions in the slice should be mounted and where
- Mount remaining BSD partitions in the correct order

Users still need to manually unmount the BSD partitions, but this should
at least help reduce the confusion due to the disparity in partition
naming schemes.  Even if you aren't confused by the way Linux maps BSD
partitions, it's annoying to figure out the mapping by hand each time.