Re: lockmgr: thread <..> unlocking unheld lock

From: Kris Kennaway <kris_at_obsecurity.org>
Date: Tue, 4 Oct 2005 15:57:42 -0400
On Tue, Oct 04, 2005 at 01:57:40PM -0400, Andrew Gallatin wrote:
> 
> I'm getting tons of these warnings on a new AMD64x2 box when doing
> heavy disk activity (buildworld or portsnap):
> 
> lockmgr: thread 0xffffff00162a1be0 unlocking unheld lock
> KDB: stack backtrace:
> lockmgr() at lockmgr+0x6ee
> VOP_UNLOCK_APV() at VOP_UNLOCK_APV+0x49
> vput() at vput+0x81
> lookup() at lookup+0xae2
> namei() at namei+0x34d
> vn_open_cred() at vn_open_cred+0x208
> kern_open() at kern_open+0xfc
> open() at open+0x25
> syscall() at syscall+0x642
> Xfast_syscall() at Xfast_syscall+0xa8
> --- syscall (5, FreeBSD ELF64, open), rip = 0x8006e2f4c, rsp = 0x7fffffffe638, rbp = 0x4 ---
> 
> This is with -current as of ~9am EDT today.   I've seen 19 in the
> course of a j4 buildworld and a portsnap fetch and extract.
> 
> Kris reported the same message, but with a different stack here:
> http://lists.freebsd.org/pipermail/freebsd-current/2005-September/055142.html

What filesystems are you using?  Any non-standard kernel options?
I've not seen any more of these on nullfs, nfs or ufs in my
environment (mine were nullfs-related).

Kris

Received on Tue Oct 04 2005 - 17:57:51 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:44 UTC