From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756521Ab3H2SCf (ORCPT ); Thu, 29 Aug 2013 14:02:35 -0400 Received: from mail-ea0-f172.google.com ([209.85.215.172]:45997 "EHLO mail-ea0-f172.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756491Ab3H2SCd (ORCPT ); Thu, 29 Aug 2013 14:02:33 -0400 Message-ID: <521F8CB3.5060407@gmail.com> Date: Thu, 29 Aug 2013 20:02:27 +0200 From: Sebastian Hesselbarth User-Agent: Mozilla/5.0 (X11; Linux i686; rv:17.0) Gecko/17.0 Thunderbird/17.0 To: Mark Brown CC: Russell King - ARM Linux , Thomas Petazzoni , Mark Rutland , Jean-Francois Moine , Jason Cooper , Pawel Moll , devicetree@vger.kernel.org, Stephen Warren , linux-kernel@vger.kernel.org, Rob Herring , Gregory CLEMENT , linux-arm-kernel@lists.infradead.org, Ian Campbell , liam.r.girdwood@intel.com Subject: Re: [PATCH 1/2] ARM: Dove: Add the audio devices in DT References: <20130828113459.48ecbb34@armhf> <521DCD33.2070008@gmail.com> <20130828121943.1c8327ca@skate> <521DD057.4040208@gmail.com> <20130829161217.GK10783@sirena.org.uk> <20130829163358.GS6617@n2100.arm.linux.org.uk> <20130829171210.GM10783@sirena.org.uk> In-Reply-To: <20130829171210.GM10783@sirena.org.uk> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 08/29/13 19:12, Mark Brown wrote: > On Thu, Aug 29, 2013 at 05:33:58PM +0100, Russell King - ARM Linux wrote: >> On Thu, Aug 29, 2013 at 05:12:17PM +0100, Mark Brown wrote: >>> On Wed, Aug 28, 2013 at 12:26:31PM +0200, Sebastian Hesselbarth wrote: > >>>> Also, we'll need to distinguish between the different audio controllers >>>> on a single SoC, i.e. i2s0 and i2s1. I suggest checking the (phys) reg >>>> base passed. > >>> Why is this required - ideally this would have been mentioned in some of >>> the previous reviews... > >> I've mentioned the differences between the blocks to you repeatedly in >> our massive thread, including that some contain the block with different > > You have described some additional features which will require > additional driver support. I would expect that the device tree bindings > for these features would be added as the features are added and the DTS > files updated, for example by listing additional compatible strings if > that was the binding update, as is the normal practice. Obviously any > hardware which is not compatible with the current binding should not be > being registered using the current binding. > > It is not clear from the above comment by Sebastian if he is referring > to the same set of hardware differences or something new - doing things > based on device address is highly unusual, it sounds like something to > do with the integration into the SoC rather than to do with the IP. > Mark, it is referring the same differences Russell already mentioned. But I already came to the conclusion, that we don't need the information in the binding. For example, if you use that controller on Dove and you hook it up for SPDIF-in (which it hasn't), than I consider this a DT bug. No need to double-check that in the driver. From that p-o-v, please just let the current binding as is. Thomas Petazzoni mentioned earlier, that the _usual_ procedure to name the compatibles is to pick the SoC that the IP appeared in first. But I am also fine with "marvell,mvebu-audio" and adding compatibles for dove or kirkwood _if_ we will ever need them. Please, just stop fighting over this again - it is not getting anything any further. Sebastian From mboxrd@z Thu Jan 1 00:00:00 1970 From: sebastian.hesselbarth@gmail.com (Sebastian Hesselbarth) Date: Thu, 29 Aug 2013 20:02:27 +0200 Subject: [PATCH 1/2] ARM: Dove: Add the audio devices in DT In-Reply-To: <20130829171210.GM10783@sirena.org.uk> References: <20130828113459.48ecbb34@armhf> <521DCD33.2070008@gmail.com> <20130828121943.1c8327ca@skate> <521DD057.4040208@gmail.com> <20130829161217.GK10783@sirena.org.uk> <20130829163358.GS6617@n2100.arm.linux.org.uk> <20130829171210.GM10783@sirena.org.uk> Message-ID: <521F8CB3.5060407@gmail.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On 08/29/13 19:12, Mark Brown wrote: > On Thu, Aug 29, 2013 at 05:33:58PM +0100, Russell King - ARM Linux wrote: >> On Thu, Aug 29, 2013 at 05:12:17PM +0100, Mark Brown wrote: >>> On Wed, Aug 28, 2013 at 12:26:31PM +0200, Sebastian Hesselbarth wrote: > >>>> Also, we'll need to distinguish between the different audio controllers >>>> on a single SoC, i.e. i2s0 and i2s1. I suggest checking the (phys) reg >>>> base passed. > >>> Why is this required - ideally this would have been mentioned in some of >>> the previous reviews... > >> I've mentioned the differences between the blocks to you repeatedly in >> our massive thread, including that some contain the block with different > > You have described some additional features which will require > additional driver support. I would expect that the device tree bindings > for these features would be added as the features are added and the DTS > files updated, for example by listing additional compatible strings if > that was the binding update, as is the normal practice. Obviously any > hardware which is not compatible with the current binding should not be > being registered using the current binding. > > It is not clear from the above comment by Sebastian if he is referring > to the same set of hardware differences or something new - doing things > based on device address is highly unusual, it sounds like something to > do with the integration into the SoC rather than to do with the IP. > Mark, it is referring the same differences Russell already mentioned. But I already came to the conclusion, that we don't need the information in the binding. For example, if you use that controller on Dove and you hook it up for SPDIF-in (which it hasn't), than I consider this a DT bug. No need to double-check that in the driver. From that p-o-v, please just let the current binding as is. Thomas Petazzoni mentioned earlier, that the _usual_ procedure to name the compatibles is to pick the SoC that the IP appeared in first. But I am also fine with "marvell,mvebu-audio" and adding compatibles for dove or kirkwood _if_ we will ever need them. Please, just stop fighting over this again - it is not getting anything any further. Sebastian