From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from service87.mimecast.com ([91.220.42.44]:47928 "EHLO service87.mimecast.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754699Ab3GVNBL convert rfc822-to-8bit (ORCPT ); Mon, 22 Jul 2013 09:01:11 -0400 Message-ID: <51ED2D27.2030904@arm.com> Date: Mon, 22 Jul 2013 14:01:27 +0100 From: Sudeep KarkadaNagesha MIME-Version: 1.0 Subject: Re: [PATCH 0/2] PM / OPP: updates to enable sharing OPPs info References: <1367406679-21603-1-git-send-email-Sudeep.KarkadaNagesha@arm.com> <519B45C3.8050700@arm.com> <20130720050943.5EB853E169C@localhost> In-Reply-To: <20130720050943.5EB853E169C@localhost> Content-Type: text/plain; charset=WINDOWS-1252 Content-Transfer-Encoding: 8BIT Sender: devicetree-owner@vger.kernel.org To: Grant Likely Cc: Sudeep KarkadaNagesha , "rob.herring@calxeda.com" , devicetree@vger.kernel.org, "linux-kernel@vger.kernel.org" , "linux-pm@vger.kernel.org" , Rob Landley , "Rafael J. Wysocki" , Shawn Guo , "linux-doc@vger.kernel.org" , mturquette@linaro.org List-ID: (sorry with new DT mailing list address this time) On 20/07/13 06:09, Grant Likely wrote: > On Tue, 21 May 2013 11:00:35 +0100, Sudeep KarkadaNagesha wrote: >> Hi Rob, Grant, >> >> On 01/05/13 12:11, Sudeep KarkadaNagesha wrote: >>> From: Sudeep KarkadaNagesha >>> >>> These are couple of updates to existing PM/OPP library to support >>> sharing of OPPs between different device nodes. >>> >>> Currently all the cpu nodes are parsed until the OPPs are found. This >>> is essential to support cpuhotplug without having to replicate OPP >>> information in all the cpu nodes. >>> >>> However in systems with multiple cpu power domain, its better to have >>> OPP entry for each cpu. To avoid replication, phandle can be specified >>> to the node which contains the full OPP information. >>> >> Is proposed option of phandle for OPP acceptable to avoid replication ? >> Any suggestions to proceed on this ? This is needed to support CPU >> hotplug on big LITTLE system where current methods like parsing all the >> nodes or just CPU0 node will not work. > > Looks fine to me. > Hi Grant, Thanks for the response. However I had a thought after seeing recent patch series by Mike[1] Since the OPPs are usually associated with clocks, and multiple devices sharing clocks will point to same clock node in DT, clk node is more logical place to specify the OPPs. IMO this will be good alignment for the consolidation effort by Mike. One issue with this approach is backward compatibility(using old DT) Regards, Sudeep [1] http://lists.infradead.org/pipermail/linux-arm-kernel/2013-July/182231.html