From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stephen Warren Subject: thermal-zones DT node bound by name rather than compatible property Date: Thu, 06 Mar 2014 10:57:01 -0700 Message-ID: <5318B6ED.3060104@wwwdotorg.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: Sender: linux-pm-owner@vger.kernel.org To: Eduardo Valentin Cc: Wei Ni , Zhang Rui , "devicetree@vger.kernel.org" , "linux-pm@vger.kernel.org" , "linux-kernel@vger.kernel.org" , Rob Herring , Pawel Moll , Mark Rutland , Ian Campbell , Kumar Gala List-Id: devicetree@vger.kernel.org Commit 4e5e4705bf69 "thermal: introduce device tree parser" introduced the text below into Documentation/devicetree/bindings/thermal/thermal.txt: > * The thermal-zones node > > The "thermal-zones" node is a container for all thermal zone nodes. It shall > contain only sub-nodes describing thermal zones as in the section > "Thermal zone nodes". The "thermal-zones" node appears under "/". This implies that software must find the thermal-zones node by node name. Node names aren't supposed to be significant in DT. Rather, software is supposed to bind to a node by searching for all nodes with a particular value in the compatible property. While there are some legacy counter-examples such as /aliases, /chosen, and /cpus, I don't think we should propagate any more of these in new bindings. Can this mistake in the binding definition be rectified, or is it too late?