> Currently, if ft_node_add() is passed NULL it will allocate an > entry for > it and return a non-NULL phandle. This patch makes it simply pass the > NULL through. Hrm, is this a good thing? Shouldn't you just BUG_ON() instead? Segher