>Number:         8051
>Category:       kern
>Synopsis:       nullfs panic in readdir
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    kern-bug-people (Kernel Bug People)
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Fri Jul 23 17:20:01 1999
>Last-Modified:
>Originator:     Konrad Schroder
>Organization:
>Release:        22 July 1999
>Environment:
System: NetBSD amarillo 1.4G NetBSD 1.4G (AMARILLO) #0: Thu Jul 22 15:17:09 PDT 1999 perseant@amarillo:/usr/src/sys/arch/i386/compile/AMARILLO i386

>Description:
	Nullfs crashes during the first readdir, when mounted from an NFS
	filesystem.

	The problem seems to be related to the locking scheme of the layered
	fs when the lower vnode does not export a lock; and its relation to
	the fact that NFS uses the genfs_nolock style of locking.
>How-To-Repeat:
	amarillo# mount -t nfs stout:/host/stout4 /host/stout4
	amarillo# mount -t null /host/stout4 /mnt
	amarillo# ls /mnt
>Fix:
	Unknown.

	(Is there some reason why NFS can't use a real vnode locking scheme?
	I realize it wouldn't be able to lock the file data, but shouldn't it
	at least be able to lock the in-kernel vnode?)
>Audit-Trail:
>Unformatted: