Hi, the ENOSYS comes from a general refusal of libacl ... i'd say.
So where to inspect its source ? (Or dowload it.) -------------------------------------------------------------------- I could not tune the mounted root filesystem: # tunefs.ufs -a enable / tunefs.ufs: POSIX.1e ACLs set tunefs.ufs: /dev/ad0s1: failed to write superblock Rather than looking for a rescue mode, i booted my backup of the virtual "hard disk" as qemu -hda and used the original "disk" as qemu -hdb. After booting i guessed that the slice counterpart of ad0s1 is ad1s1. So i blindly tuned the not mounted device: # tunefs.ufs -a enable /dev/ad1s1 tunefs.ufs: POSIX.1e ACLs set After booting with the tuned "disk" as -hda i now get: # tunefs.ufs -p / tunefs.ufs: POSIX.1e ACLs: (-a) enabled But regrettably only the requested failure case emerges by the (slightly modified) first example from man setfacl: # grep nobody /etc/passwd nobody:x:65534:65534:nobody:/nonexistent:/bin/sh # touch testfile # setfacl -m u:nobody:w testfile setfacl: testfile: Function not implemented So with ACL capable filesystem there is still the ENOSYS error. Most probably from libacl $ ldd $(which setfacl) libacl.so.1 => /lib/i386-kfreebsd-gnu/libacl.so.1 (0x21024000) -------------------------------------------------------------------- Have a nice day :) Thomas
