bin/44204: resize_ffs(8) problems shrinking 64K blocksize ffsv1 < 1cyl

[riz%netbsd.org@localhost]

>Number:         44204
>Category:       bin
>Synopsis:       resize_ffs(8) problems shrinking 64K blocksize ffsv1 < 1cyl
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    bin-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Wed Dec 08 07:40:00 +0000 2010
>Originator:     Jeff Rizzo <riz%NetBSD.org@localhost>
>Release:        NetBSD 5.99.41
>Organization:
        
>Environment:
        
        
System: NetBSD netbsd6 5.99.41 NetBSD 5.99.41 (GENERIC) #0: Sun Dec 5 19:35:40 
PST 2010 riz@netbsd6:/home/riz/obj/sys/arch/amd64/compile/GENERIC amd64
Architecture: x86_64
Machine: amd64
>Description:
        When shrinking a file system with 64K blocksize less than a whole
cylinder group, the file system is marked clean but is actually not, as
evidenced by a run through fsck_ffs. Dropping a whole cg or more seems
to work OK.
>How-To-Repeat:

- Create a file system with 64K blocks
- use resize_ffs to shrink it less than one whole cg
- run fsck_ffs -f on it

netbsd6:riz  /usr/src/sbin/resize_ffs> newfs -O1 -s 65536 -f 8192 -s 131072 -F 
/tmp/fs
/tmp/fs: 64.0MB (131072 sectors) block size 65536, fragment size 8192
        using 3 cylinder groups of 21.38MB, 342 blks, 512 inodes.
super-block backups (for fsck_ffs -b #) at:
128, 43904, 87680,
netbsd6:riz  /usr/src/sbin/resize_ffs> resize_ffs -y -s 98304 /tmp/fs
netbsd6:riz  /usr/src/sbin/resize_ffs> fsck_ffs -F -f /tmp/fs
** /tmp/fs
** File system is already clean
** Last Mounted on 
** Phase 1 - Check Blocks and Sizes
** Phase 2 - Check Pathnames
** Phase 3 - Check Connectivity
** Phase 4 - Check Reference Counts
** Phase 5 - Check Cyl groups
FREE BLK COUNT(S) WRONG IN SUPERBLK
SALVAGE? [yn] n

SUMMARY INFORMATION BAD
SALVAGE? [yn] n

BLK(S) MISSING IN BIT MAPS
SALVAGE? [yn] n

1 files, 1 used, 6046 free (14 frags, 754 blocks, 0.2% fragmentation)

***** UNRESOLVED INCONSISTENCIES REMAIN *****


>Fix:
        None given.  There is some code to deal with partial cgs;  
check there first.

>Unformatted: