Peter, this should apply on top of your 11 vg_read() patches posted 1-12-2009. I've been thinking about ways to clean up vg_might_exist() and this is one potential route but has a downside. It makes the code simpler and more readable but does have one shortfall. Normally when a VG does not exist we print an error message to this effect. This patch removes the message. One approach is to add another flag (ugh - we just removed one) to this effect. Might not be possible this go around but perhaps this is close enough to spur some thoughts of how we could clean this up.