From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755825Ab1ERVro (ORCPT ); Wed, 18 May 2011 17:47:44 -0400 Received: from mail-iy0-f174.google.com ([209.85.210.174]:54603 "EHLO mail-iy0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754743Ab1ERVrm convert rfc822-to-8bit (ORCPT ); Wed, 18 May 2011 17:47:42 -0400 DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:cc:content-type :content-transfer-encoding; b=Yrk8O9tw2HmhCbXVMjzxcLaxTaJ4vszsglIl7nzMkc8zKGylvwKH50fuBa4joOIpH5 +tVfh8UWu0N+SOProeNezyh1/sOEVESwZR9Gf7JkYZ8oShi+YrRAOR/mDGFchRj5I6+E 3MQKvjUK1FeWUswRDqMcHOSNGS6gWZPhcknio= MIME-Version: 1.0 In-Reply-To: <20110518155615.GJ15292@game.jcrosoft.org> References: <201105181047.17309.arnd@arndb.de> <20110518155615.GJ15292@game.jcrosoft.org> Date: Wed, 18 May 2011 22:47:41 +0100 X-Google-Sender-Auth: 1yO3wjtpxSKf7eVXkVhOPQXl84U Message-ID: Subject: Re: [RFC] ARM Subarchitecture group maintainership From: Catalin Marinas To: Jean-Christophe PLAGNIOL-VILLARD Cc: Arnd Bergmann , "linux-arm-kernel@lists.infradead.org" , Linus Torvalds , Thomas Gleixner , Russell King , "linux-kernel@vger.kernel.org" , Nicolas Pitre Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8BIT Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wednesday, 18 May 2011, Jean-Christophe PLAGNIOL-VILLARD wrote: > On 10:47 Wed 18 May     , Arnd Bergmann wrote: >> This is the draft plan for maintaining the ARM subarchitectures in a common >> tree, as a way to help coordinate the upstream merging of the >> arch/arm/{plat,mach}-* changes into Linus' tree. > > How do you plan to manage with already sub arch maintainers? > > In my ming such tree could be good to organise cross arch drivers > but for real arch specific the current workflow is good > > If we add a new stage it will be difficult to follow for people > > I'd like to have scenario example where such workflow will give a significant > advantage compare to the current workflow. And please real example. I'll let Arnd and the other members of this group come up with real examples. In the meantime I'll give my view on this. The aim is not to hide existing sub-architecture patches behind another git tree so that people won't notice the conflicts. It's not just a workflow change, the Acked-by mechanism may have worked as well, though some people may skip it (even unintentionally). There is a lot of code duplication between various sub-architectures under arch/arm/ and this group, together with the sub-arch maintainers, will try to consolidate this, move common code out into drivers/ so that it can be easily shared, start moving platforms to FDT etc. (they could give a roadmap but it's early stages now). This needs to be a coordinated effort and the group will ensure that the general direction of code reduction is followed by all the current and new sub-arch maintainers. A lot of platform ports have been done just by copying existing code without any effort in exploiting the commonality. This could be spotted much easier with a dedicated group of people and guide the sub-arch maintainers in the write direction. It may be even quicker for them to get code merged into drivers/ where applicable. I expect the activity of this group to be reduced in the future, once the consolidation work is complete (but that's not a trivial task). At that point maybe one or two people would just keep an eye on what gets pushed into mainline in sub-arch directories and sub-arch maintainers would just work as before, sending pull requests to Linus directly. As the incentive for existing sub-arch maintainers to agree with this strategy - given Linus' recent comments on the ARM sub-arch trees, it is possible that he won't merge such trees pushed directly to him if they don't show a consistent direction towards code consolidation. I think this group and tree would actually make upstream pushing quicker. -- Catalin From mboxrd@z Thu Jan 1 00:00:00 1970 From: catalin.marinas@arm.com (Catalin Marinas) Date: Wed, 18 May 2011 22:47:41 +0100 Subject: [RFC] ARM Subarchitecture group maintainership In-Reply-To: <20110518155615.GJ15292@game.jcrosoft.org> References: <201105181047.17309.arnd@arndb.de> <20110518155615.GJ15292@game.jcrosoft.org> Message-ID: To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Wednesday, 18 May 2011, Jean-Christophe PLAGNIOL-VILLARD wrote: > On 10:47 Wed 18 May ? ? , Arnd Bergmann wrote: >> This is the draft plan for maintaining the ARM subarchitectures in a common >> tree, as a way to help coordinate the upstream merging of the >> arch/arm/{plat,mach}-* changes into Linus' tree. > > How do you plan to manage with already sub arch maintainers? > > In my ming such tree could be good to organise cross arch drivers > but for real arch specific the current workflow is good > > If we add a new stage it will be difficult to follow for people > > I'd like to have scenario example where such workflow will give a significant > advantage compare to the current workflow. And please real example. I'll let Arnd and the other members of this group come up with real examples. In the meantime I'll give my view on this. The aim is not to hide existing sub-architecture patches behind another git tree so that people won't notice the conflicts. It's not just a workflow change, the Acked-by mechanism may have worked as well, though some people may skip it (even unintentionally). There is a lot of code duplication between various sub-architectures under arch/arm/ and this group, together with the sub-arch maintainers, will try to consolidate this, move common code out into drivers/ so that it can be easily shared, start moving platforms to FDT etc. (they could give a roadmap but it's early stages now). This needs to be a coordinated effort and the group will ensure that the general direction of code reduction is followed by all the current and new sub-arch maintainers. A lot of platform ports have been done just by copying existing code without any effort in exploiting the commonality. This could be spotted much easier with a dedicated group of people and guide the sub-arch maintainers in the write direction. It may be even quicker for them to get code merged into drivers/ where applicable. I expect the activity of this group to be reduced in the future, once the consolidation work is complete (but that's not a trivial task). At that point maybe one or two people would just keep an eye on what gets pushed into mainline in sub-arch directories and sub-arch maintainers would just work as before, sending pull requests to Linus directly. As the incentive for existing sub-arch maintainers to agree with this strategy - given Linus' recent comments on the ARM sub-arch trees, it is possible that he won't merge such trees pushed directly to him if they don't show a consistent direction towards code consolidation. I think this group and tree would actually make upstream pushing quicker. -- Catalin