On Mon, Mar 05, 2018 at 05:40:13PM +1100, David Gibson wrote: > On Fri, Feb 23, 2018 at 09:35:15AM +0100, Geert Uytterhoeven wrote: > > Hi all, > > > > Using "fragment@N" as node name for overlay fragments violates unit > > address rules: > > > > <stdout>: Warning (unit_address_vs_reg): Node /fragment@0 has a unit name, but no reg property > > Hm, regardless of this patch, that's really a bug in the checks not to > handle overlays as a different case. > > The quick fix would be to just suppress that check if an __overlay__ > subnode exists. THe more ambitious approach would be to actually > parse things that look like overlays into their components and apply > checks separately to each fragment (except for ones that don't make > sense except on a fully resolved tree). I had plans for that, but am > unlikely to get time to work on it any time soon. Fwiw, I've now made such a change to the checks. However, I think you can still get a similar problem in the generated __local_fixups__ node for similar reasons. That one's a bit harder to check. > > > This patch series (for dtc) fixes: > > 1. generation of node names when using overlay sugar syntax, > > 2. all documentation and tests. > > > > If this is accepted, I'll send similar patches for Linux. > > > > Note that I've been using "fragment-N" in hand-written DT overlays for > > years, so this is known to work fine with Linux. > > Existing tools expecting the @ would be my only concern with this. > You've check Linux is ok, and I've checked libfdt is ok too. I'm not > really confident if that's enough to call it everything we care about. > -- David Gibson | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ | _way_ _around_! http://www.ozlabs.org/~dgibson
Attachment:
signature.asc
Description: PGP signature
