From mboxrd@z Thu Jan 1 00:00:00 1970 From: Arnd Bergmann Subject: Re: [PATCH] of: support an enumerated-bus compatible value Date: Tue, 24 Jul 2012 20:05:50 +0000 Message-ID: <201207242005.50982.arnd@arndb.de> References: <4FF1C567.4060809@wwwdotorg.org> <201207241848.53308.arnd@arndb.de> <500EF7C5.6060406@wwwdotorg.org> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <500EF7C5.6060406-3lzwWm7+Weoh9ZMKESR00Q@public.gmane.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: devicetree-discuss-bounces+gldd-devicetree-discuss=m.gmane.org-uLR06cmDAlY/bJ5BZ2RsiQ@public.gmane.org Sender: "devicetree-discuss" To: Stephen Warren Cc: devicetree-discuss-uLR06cmDAlY/bJ5BZ2RsiQ@public.gmane.org, Mark Brown , Rob Herring List-Id: devicetree@vger.kernel.org On Tuesday 24 July 2012, Stephen Warren wrote: > > One would think that, but the of_address handling code actually treats > > empty ranges the same as missing ranges, in violation of the spec, > > and as a workaround to deal with some powermac machines that required > > this. > > > > I'd rather fix the code to deal with this correctly. > > OK, so what is correctly then? > > Adding an explicit enumerated-bus compatible value/node-type/binding > seems like "correctly" to me, but it seemed like others didn't agree. I'm talking about the issue that you cannot currently have a bus without a "ranges" property and #size-cells=<0> and expect Linux to turn the devices on it into platform devices, because the of_translate_address function fails on this. Arnd