From mboxrd@z Thu Jan 1 00:00:00 1970 From: Felipe Balbi Subject: Re: Query: Regulator framework in EHCI driver Date: Thu, 5 Nov 2009 11:41:34 +0200 Message-ID: <20091105094134.GB3045@nokia.com> References: <20091103153046.GA30320@nokia.com> <19F8576C6E063C45BE387C64729E73940436F9375B@dbde02.ent.ti.com> <20091104133922.GD25044@sirena.org.uk> <5A47E75E594F054BAF48C5E4FC4B92AB030A67DE6A@dbde02.ent.ti.com> <20091104144126.GA30644@rakim.wolfsonmicro.main> <5A47E75E594F054BAF48C5E4FC4B92AB030A67DE7F@dbde02.ent.ti.com> <20091104144811.GB30644@rakim.wolfsonmicro.main> <5A47E75E594F054BAF48C5E4FC4B92AB030A67DE8A@dbde02.ent.ti.com> <19F8576C6E063C45BE387C64729E73940436F9384E@dbde02.ent.ti.com> <20091105093641.GA21779@rakim.wolfsonmicro.main> Reply-To: felipe.balbi@nokia.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from smtp.nokia.com ([192.100.105.134]:47841 "EHLO mgw-mx09.nokia.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752083AbZKEJmu (ORCPT ); Thu, 5 Nov 2009 04:42:50 -0500 Content-Disposition: inline In-Reply-To: <20091105093641.GA21779@rakim.wolfsonmicro.main> Sender: linux-omap-owner@vger.kernel.org List-Id: linux-omap@vger.kernel.org To: ext Mark Brown Cc: "Gupta, Ajay Kumar" , "Gadiyar, Anand" , "Balbi Felipe (Nokia-D/Helsinki)" , "linux-omap@vger.kernel.org" , "Aggarwal, Anuj" Hi, On Thu, Nov 05, 2009 at 10:36:41AM +0100, ext Mark Brown wrote: > On Thu, Nov 05, 2009 at 08:54:02AM +0530, Gupta, Ajay Kumar wrote: > > > I think we can add a check for supply name in EHCI driver, if its > valid then we will call regulator_get()/enable() but if it's NULL then > don't do anything. > > > > So the boards which don't use any regulator they can pass a NULL to > supply name. I don't know if I'm the only one, but Anand's mails are coming in over 80 chars lines. > No, this would be a substantial misuse of the regulator API. The supply > name should not be being passed through as platform data, the driver > should request a fixed name (usually the name of the relevant physical > supply to the chip) and let the API map this onto the actual supply for > the system. I agree with Mark here. The best solution would be to provide a fixed voltage regulator for the other boards. -- balbi