Re: 5.99.42/i386 crash (backtrace + core available)

[David Holland <dholland-current%netbsd.org@localhost>]

On Sun, Jan 09, 2011 at 05:48:47PM +0100, Juergen Hannken-Illjes wrote:
 > Looks like we have
 > 
 >      vnode e15c2170, type nullfs with lower vnode df8295c8, type ffs
 > 
 > so here we lookup the lower vnode, need a new one and hit the upper (nullfs)
 > vnode.  Trying to lock it always has to lock the lower vnode.  This is the
 > background of ad's comment in vclean
 > 
 >      /* XXXAD should not lock vnode under layer */
 >      mutex_exit(&vp->v_interlock);
 >      VOP_LOCK(vp, LK_EXCLUSIVE);
 > 
 > I have no idea how to solve this mess...

Yeah, I saw that when I was looking at revoke and had a feeling we
were about to be in for some fun :(

-- 
David A. Holland
dholland%netbsd.org@localhost