On Mon, 2013-09-16 at 16:48 -0700, Olof Johansson wrote: > On Mon, Sep 16, 2013 at 4:47 PM, Benjamin Herrenschmidt > <benh@xxxxxxxxxxxxxxxxxxx> wrote: > > On Mon, 2013-09-16 at 15:48 -0700, Olof Johansson wrote: > >> > A node that has a "reg" property should have the corresponding unit > >> > address. > >> > >> No, absolutely _NOT_ a requirement. Unit address is only required if > >> needed to disambiguate two properties with the same name. > >> > >> If there are no ambiguities, then leaving off the unit address is much > >> preferred. > > > > I disagree :-) > > Well, good thing you've got your own arch to litter the device trees > with unit specifiers in then. :) Right :-) We tend to have multiple memory nodes on server anyway so it's not a big deal. > > Also this would be only true of our find_node_by_path was capable of > > handling it, which it isn't. Thus you end up with generic code looking > > for /memory and finding nothing ... > > Yes, this should be fixed. Right, the whole thing becomes mostly a non-issue once that's fixed. My main objection isn't that ARM doesn't use unit address specifiers. My objection is that the binding documents no unit address :-) It should instead document the unit address with a note indicating that it can be omitted if there is no ambiguity. But first, do we have a volunteer to fix the path parsing code ? Also do we *really* need to keep the path parsing code for fdt ? IE. It would be annoying to have to duplicate that code for before and after expansion... Cheers, Ben. > -Olof > -- > To unsubscribe from this list: send the line "unsubscribe devicetree" in > the body of a message to majordomo@xxxxxxxxxxxxxxx > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html