From mboxrd@z Thu Jan 1 00:00:00 1970 From: Maxime Coquelin Subject: Re: [PATCH v4 6/6] ARM: dts: STiH407: Add B2120 board support Date: Tue, 20 May 2014 09:53:57 +0200 Message-ID: <537B0A15.409@st.com> References: <1394614210-15698-1-git-send-email-maxime.coquelin@st.com> <1394614210-15698-7-git-send-email-maxime.coquelin@st.com> <20140520072048.GC20874@lee--X1> <20140520074353.GD20874@lee--X1> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; Format="flowed" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20140520074353.GD20874@lee--X1> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: "linux-arm-kernel" Errors-To: linux-arm-kernel-bounces+linux-arm-kernel=m.gmane.org@lists.infradead.org To: Lee Jones , Olof Johansson Cc: Mark Rutland , "devicetree@vger.kernel.org" , Russell King , "kernel@stlinux.com" , Pawel Moll , Ian Campbell , Linus Walleij , Srinivas Kandagatla , "linux-doc@vger.kernel.org" , "linux-kernel@vger.kernel.org" , Stuart Menefy , Rob Herring , Rob Landley , Kumar Gala , Giuseppe Cavallaro , "linux-arm-kernel@lists.infradead.org" List-Id: devicetree@vger.kernel.org On 05/20/2014 09:43 AM, Lee Jones wrote: >>>>> + soc { >>>>> + sbc_serial0: serial@9530000 { >>>>> + status = "okay"; >>>>> + }; >>>> >>>> You might want to consider reference-based syntax here instead, so you >>>> don't have to mimic the hierarchy. That'd be (at the root level of the >>>> file, below this secion: >>>> >>>> &sbc_serial0: { >>>> status = "okay"; >>>> }; >>> >>> I'm personally not keen on this scheme. It's sometimes helpful to know >>> the hierarchy and I don't think it's a large overhead to format the >>> subordinate DTS files in this way. >>> >>> Please consider not enforcing this. >> >> Definitely not enforcing it, and I didn't use to like it either but it >> has some real upsides. >> >> In particular, it saves a lot of grief when you're changing something >> like the unit-id of a node in .dtsi and forget to do the same update >> in the dts. > > I'm not entirely sure what a unit-id is, but I can see that there > would be benefits to using the referenced-based syntax as you call > it. If any of those benefits hold true here I won't push back, but I > would personally like to see us default to the hierarchical scheme. +1, I would prefer to keep the hierarchical scheme.