From mboxrd@z Thu Jan 1 00:00:00 1970 From: soren.brinkmann@xilinx.com (=?utf-8?B?U8O2cmVu?= Brinkmann) Date: Mon, 7 Apr 2014 08:39:34 -0700 Subject: [PATCH v2 2/5] ARM: zynq: dt: Convert to preprocessor includes In-Reply-To: <53429754.8080105@topic.nl> References: <1396653256-28397-1-git-send-email-soren.brinkmann@xilinx.com> <1396653256-28397-3-git-send-email-soren.brinkmann@xilinx.com> <53423E75.3020008@monstr.eu> <53429754.8080105@topic.nl> Message-ID: <8db0b480-2c27-4f97-8712-e16fd39d4159@CH1EHSMHS038.ehs.local> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Mon, 2014-04-07 at 02:17PM +0200, Mike Looijmans wrote: > On 04/07/2014 07:58 AM, Michal Simek wrote: > >Hi Soren, > > > >On 04/05/2014 01:14 AM, Soren Brinkmann wrote: > >>Convert all Zynq DT files to the dtc preprocessor include syntax. > >>This allows to include header files in the devicetrees like other > >>SoC-types already do. > >> > >>Inspired-by: Steffen Trumtrar > >>(http://www.spinics.net/lists/arm-kernel/msg319832.html) > >> > >>Signed-off-by: Soren Brinkmann > > > >These 4 patches needs more wider discussion if this is helpful or > >not. Currently I can't see any value in it because everything > >is just generated and fixed. I think I had the same discussion > >with Laurent some weeks ago regarding this. > > I would be kinda neutral here. I'd consider it helpful, it improves > readability (regardless of whether they are generated or hand > crafted). That's convenient for things like interrupt sensitivity, I > can't remember whether 4 is level or edge type. On the other hand, > the clock indices are just as much magic numbers as the memory > addresses. If I suspect an error in that area, I'd start by lokking > in /sys/kernel/debug/clk but wouldn't start in the DT. I agree. Readability is better this way. And whether our generator spits out magic numbers of these symbolic names should not be a big difference, is it? S?ren