svn commit: r328013 - head/sbin/fsck_ffs

[Mark Johnston]

On Sat, Mar 10, 2018 at 05:01:40PM -0500, David Bright wrote:
> With regard to the fsck_ffs behavior being a regression because formerly the FS would be mounted successfully:
> 
> That was not my experience. What I observed was that the “fsck -y” would give the “please re-run” message, exit with 0 status so the boot would continue, the subsequent mount would fail because the filesystem was not clean, and *then* the boot would stop and drop to single user.

I think my problem is specific to SU without journaling. The UFS code
allows one to mount an unclean filesystem in that configuration since SU
guarantees that on-disk metadata is consistent. A background fsck takes
care of leaked inodes and data blocks.

(FWIW, I'm not using journaling only because makefs(8) doesn't support
the creation of SU+J filesystems.)

/dev/gpt/rootfs: FREE BLK COUNT(S) WRONG IN SUPERBLK (SALVAGED)
/dev/gpt/rootfs: SUMMARY INFORMATION BAD (SALVAGED)
/dev/gpt/rootfs: BLK(S) MISSING IN BIT MAPS (SALVAGED)
/dev/gpt/rootfs: 32664 files, 495447 used, 813272 free (176 frags, 203274 blocks, 0.0% fragmentation)

***** PLEASE RERUN FSCK *****
WARNING: /: reload pending error: blocks 192 files 3
Unknown error 16; help!
ERROR: ABORTING BOOT (sending SIGTERM to parent)!
Mar 10 12:47:50 init: /bin/sh on /etc/rc terminated abnormally, going to single user mode
Enter full pathname of shell or RETURN for /bin/sh: 
# mount
/dev/gpt/rootfs on / (ufs, local, read-only)
devfs on /dev (devfs, local, multilabel)
# mount -u -o rw /
WARNING: / was not properly dismounted
# echo $?
0
# mount
/dev/gpt/rootfs on / (ufs, local, soft-updates)
devfs on /dev (devfs, local, multilabel)