On Sun, Nov 13, 2016 at 11:50:27AM +0100, Juergen Gross wrote: > On 12/11/16 16:10, Wei Liu wrote: > > On Fri, Nov 11, 2016 at 09:00:00AM +0100, Juergen Gross wrote: > >> Instead of calling add_change_node() at places where write_node() is > >> called, do that inside write_node(). > >> > >> Signed-off-by: Juergen Gross <jgr...@suse.com> > > > > > > There seems to be a subtle change in behaviour -- previously in > > create_node, there is no add_chnage_node called. Now it has. > > No change. add_change_node() has been called after calling create_node() >
I don't think that's true for all cases. For example, setup_structure calls create_node but doesn't call add_change_node after that. But I have convinced myself that it would be fine to call add_change_node in write_node because that seems rather logical -- writing a node is certainly changing it. Wei. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xen.org https://lists.xen.org/xen-devel
