Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753080Ab3H0IGp (ORCPT ); Tue, 27 Aug 2013 04:06:45 -0400 Received: from mail-ee0-f47.google.com ([74.125.83.47]:62347 "EHLO mail-ee0-f47.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752831Ab3H0IGl (ORCPT ); Tue, 27 Aug 2013 04:06:41 -0400 Date: Tue, 27 Aug 2013 09:06:35 +0100 From: Lee Jones To: Mark Rutland Cc: Sascha Hauer , Linus Walleij , "linux-arm-kernel@lists.infradead.org" , "linux-kernel@vger.kernel.org" , Arnd Bergmann , "devicetree@vger.kernel.org" Subject: Re: [PATCH 09/33] ARM: ux500: Supply the I2C clocks lookup to the DBX500 DT Message-ID: <20130827080635.GC6152@lee--X1> References: <1370521041-32318-10-git-send-email-lee.jones@linaro.org> <20130820093034.GL31036@pengutronix.de> <20130822133730.GB23152@e106331-lin.cambridge.arm.com> <20130822141900.GB17154@lee--X1> <20130822151723.GE23152@e106331-lin.cambridge.arm.com> <20130822154116.GC17154@lee--X1> <20130822211912.GE31036@pengutronix.de> <20130823075607.GD17154@lee--X1> <20130823165539.GD7015@e106331-lin.cambridge.arm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: <20130823165539.GD7015@e106331-lin.cambridge.arm.com> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2258 Lines: 48 On Fri, 23 Aug 2013, Mark Rutland wrote: > On Fri, Aug 23, 2013 at 08:56:07AM +0100, Lee Jones wrote: > > I had a short chat with Rob last night about this. I'm going to loop > > him in to the conversation, as he wrote the binding. > > > > > > When most of the other clocks that we deal with are being requested, > > > > they rely on being index zero: > > > > > > > > drivers/i2c/busses/i2c-nomadik.c: dev->clk = clk_get(&adev->dev, NULL); > > > > > > Look at drivers/clk/clkdev.c, there's some fuzzy matching > > > involved when you pass NULL as connection id. > > > > Yes, I've been looking at that. This is why it works currently. I > > think I need to change all of the drivers to specify which clock they > > want. At the moment that 'fuzzy matching' is what's saving us. If > > anyone were to change our DTS file to match what the binding says, > > then it would cease to work. I'm guessing this is the same for all > > other DTS files too: > > I think if anything, the binding document(s) should be updated to > describe that apb_pclk is referred to by name, and the names of the > other clocks should be described in the specific device bindings. We can > then modify the drivers which grab clock 0 to explicitly grab the first > clock by name, and backwards compatibility should not be broken. > > I don't believe any other OS has implemented the common clock bindings, > and we've never supported the binding as described. Let's correct the > de-facto standard into a standard by decree. I think we need to respect, or at least take into consideration the reason for the original 'de-facto' standard. Other OSes shouldn't be forced to provide a named clock request in order to obtain 'apb_pclk'. If the binding says it should be first, then perhaps we should do just that. It's simply a matter of naming all subsequent clocks related to AMBA devices. -- Lee Jones Linaro ST-Ericsson Landing Team Lead Linaro.org │ Open source software for ARM SoCs Follow Linaro: Facebook | Twitter | Blog -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/