The following reply was made to PR kern/33861; it has been noted by GNATS.

From: Jed Davis <jld@panix.com>
To: gnats-bugs@NetBSD.org
Cc: yamt@NetBSD.org
Subject: Re: kern/33861
Date: Sat, 1 Jul 2006 05:05:18 -0400

 On Fri, Jun 30, 2006 at 09:40:02AM +0000, YAMAMOTO Takashi wrote:
 >  
 >  i'm not sure if it's a good idea.
 >  isn't it a matter of VOP interface difference rather than carefulness?
 
 Indeed, I see you're right; Solaris doesn't pass those vnodes from the
 fs-independent code, so nfsrename has to do the lookup itself; that it
 happens to be done in more or less the same place in the control flow as
 my addition is somewhat coincidental.
 
 >  coincidently i've written the similar patch.  (attached)
 
 I've tested it, and it works for me.  (Also, FWIW, I've reproduced the
 original problem against an NFS server with ~no other load; there was
 one case where it wasn't happening, but that's clearly not the reason.)
 
 >  next time, let's take gnats "Responsible:" before fixing. :-)
 
 Oops.  Not used to being able to do that.  (-:
 
 --Jed