bin/54935: dir corruption not fixed by fsck

[martin%NetBSD.org@localhost]

>Number:         54935
>Category:       bin
>Synopsis:       dir corruption not fixed by fsck
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    bin-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Tue Feb 04 18:50:00 +0000 2020
>Originator:     Martin Husemann
>Release:        NetBSD 9.99.45
>Organization:
The NetBSD Foundation, Inc.
>Environment:
System: NetBSD seven-days-to-the-wolves.aprisoft.de 9.99.45 NetBSD 9.99.45 (GENERIC) #348: Mon Feb 3 08:14:54 CET 2020 martin%seven-days-to-the-wolves.aprisoft.de@localhost:/work/src/sys/arch/amd64/compile/GENERIC amd64
Architecture: x86_64
Machine: amd64
>Description:

I have a machine (running -current as of today) wich apparently has a sligthly
corrupted filesystem.

I booted single user and ran fsck -fy /

That finished and claimed to have fixed the fs.

On reboot I got:
[ 3489.8572502] /: bad dir ino 1054908 at offset 8120: Bad dir (not rounded), reclen=0x803, namlen=116, dirsiz=128 <= reclen=2051 <= maxsize=72, flags=0x5001, entryoffsetinblock=8120, dirblksiz=512

and on next boot:

[  17.5679221] /: bad dir ino 2107008 at offset 8128: NUL in name [raic] i=4, namlen=6
[  27.9025333] panic: /: bad dir ino 1077399 at offset 8112: NUL in name [libdc++.a] i=9, namlen=11

Backup and restore time - but somehow I guess fsck_ffs should have dealt with
it properly.

>How-To-Repeat:
n/a

>Fix:
n/a