Code review comment for ~dbungert/curtin:fs_pass_one

I know it's a bit of a pain but I think I would really prefer to avoid emitting passno=1 for tmpfses. One approach would be to have passno == 0 for fstypes marked as "nodev" in /proc/filesystems but maybe that's just silly and you can special case "tmpfs" and "ramfs" (which are the filesystems MAAS supports today that should have passno == 0 by my reading).