On Tue, 2008-01-29 at 13:14 -0800, Dan Nicholson wrote: > Earlier Victor and I discussed moving the sleep after the check in > spin_lock(). That got lost in the shuffle, so here's a new patch. Applied, thanks. Richard.