From mboxrd@z Thu Jan 1 00:00:00 1970 From: Viresh Kumar Subject: Re: [PATCH 01/16] PM / OPP: Add 'supply-names' binding Date: Thu, 8 Oct 2015 14:57:09 +0530 Message-ID: <20151008092709.GC18898@linux> References: <2b87b162eabd1570ae2311e1ef8655acda72f678.1441972771.git.viresh.kumar@linaro.org> <55F72C97.2030306@kernel.org> <20150915024747.GI32551@linux> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from mail-pa0-f44.google.com ([209.85.220.44]:36531 "EHLO mail-pa0-f44.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755296AbbJHJ1O (ORCPT ); Thu, 8 Oct 2015 05:27:14 -0400 Received: by pablk4 with SMTP id lk4so49813346pab.3 for ; Thu, 08 Oct 2015 02:27:14 -0700 (PDT) Content-Disposition: inline In-Reply-To: <20150915024747.GI32551@linux> Sender: linux-pm-owner@vger.kernel.org List-Id: linux-pm@vger.kernel.org To: Rob Herring Cc: Rafael Wysocki , nm@ti.com, sboyd@codeaurora.org, linaro-kernel@lists.linaro.org, linux-pm@vger.kernel.org, rob.herring@linaro.org, lee.jones@linaro.org, Mark Brown , devicetree@vger.kernel.org, Ian Campbell , Kumar Gala , open list , Mark Rutland , Pawel Moll , "Rafael J. Wysocki" Rob, On 15-09-15, 08:17, Viresh Kumar wrote: > On 14-09-15, 15:22, Rob Herring wrote: > > What if we have a 2nd device and supply rail? For example, what if the > > L2$ has a separate rail from the cores but is linked to the OPPs. > > Right, so that is the case with the Mediatek SoC as well, AFAIR. How > do we plan to treat L2 devices? For example, in the mediatek cpufreq > driver, they change L2's supplies together with CPUs. > > One way to get that done, with such very closely bound devices is to > add two regulators: > > cpu-supply = ...; > l2-supply = ...; > > And then a property in OPP table: > > supply-name = "cpu", "l2"; > > And maybe fix the order in which the supplies which be updated, based > on the order in which entries are present in the above property. > > Any other way you suggest for doing that ? > > > Remind me of when do we have multiple regulators for a cpu? > > I haven't seen that yet, but its more like what I explained above. > i.e. one for the CPU and other one for the L2 cache. > > But, these bindings do apply for other devices as well, where it > should be very much possible. Not sure if you still have any objections to this patch? -- viresh