system/1823: fsck segfaults in single user mode

[wfaulk@beaglebros.com]

>Number:         1823
>Category:       system
>Synopsis:       fsck segfaults in single user mode
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    bugs
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Mon May 14 19:30:02 MDT 2001
>Last-Modified:
>Originator:     Bitt Faulk
>Organization:
net
>Release:        2.8
>Environment:
	
	System      : OpenBSD 2.8
	Architecture: OpenBSD.i386
	Machine     : i386
>Description:
	If one of my filesystems needs to be checked on bootup, the fsck
	*always* segfaults with signal 11.  Upon entering single user mode
	(because it wasn't able to fsck all the filesystems it needed to),
	I still get the same segfault every time I run fsck.  If I allow
	it to boot anyway (the offending filesystem is just data), and
	run fsck against it in multiuser mode, it works fine.  The offending
	filesystem is a 56754678kB ffs filesystem.  I have 7 other, much
	smaller, ffs filesystems on the same machine that do not exhibit
	this problem.  Also, the offending filesystem is the only one on the
	machine that is attached to a PCI IDE card (a Promise UltraATA/100).
	The others are on the on-motherboard IDE controllers.  But I've found
	no problems once it's been fsck'd properly and mounted.  I've got
	a copy of a ktrace done on the fsck in single user mode (where it
	failed, obviously), and one in multiuser mode (where it worked).  I
	don't know how to include them here, so let me know if and how you
	want them.
>How-To-Repeat:
	Unfortunately, I'm sure that it's specific to my system in some way,
	but, unmount that particular filesystem uncleanly and boot.  Then
	run fsck.  It always fails while in single user mode.  It has never
	failed in multiuser mode.
>Fix:
	No idea.

>Audit-Trail:
>Unformatted: