> Yes, you are wrong here, because it was used to restore pointer to > struct space "_index", but "_index" is a system space, and it is never > altered. So pointer to "_index" is always valid, and it makes no sense > to restore it on runtime. Then I am OK with last patch in series.