From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752977AbbJQRJp (ORCPT ); Sat, 17 Oct 2015 13:09:45 -0400 Received: from seldrel01.sonyericsson.com ([37.139.156.2]:6062 "EHLO seldrel01.sonyericsson.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751697AbbJQRJn (ORCPT ); Sat, 17 Oct 2015 13:09:43 -0400 Date: Sat, 17 Oct 2015 10:09:38 -0700 From: Bjorn Andersson To: Ulf Hansson CC: Rob Herring , Pawel Moll , Mark Rutland , Ian Campbell , Kumar Gala , "devicetree@vger.kernel.org" , "linux-kernel@vger.kernel.org" , linux-mmc Subject: Re: [PATCH v2] mmc: core: Allow specifying current consumption Message-ID: <20151017170938.GP24668@usrtlx11787.corpusers.net> References: <1435725525-4367-1-git-send-email-bjorn.andersson@sonymobile.com> <1444698032-9051-1-git-send-email-bjorn.andersson@sonymobile.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.22 (2013-10-16) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri 16 Oct 05:55 PDT 2015, Ulf Hansson wrote: > On 13 October 2015 at 03:00, Bjorn Andersson > wrote: > > This allows us to specify expected current consumption of the vmmc and > > vqmmc regulators. This is needed to bring the supplying regulators out > > of their low-power-mode while accessing the mmc. > > This indeed makes sense, still I need to think a bit more on this. > Thanks. We can currently work around this by using "regulator-system-load" on the vmmc regulator, but a proper fix would be nice. > For example, can we allow these regulators to enter low power again at > some times? If so, when and what should that current value be. > The regulator framework recalculates the total load on a regulator upon enable/disable, but currently it doesn't depend on the state of the regulator. I talked to Mark about this a while back and he was positive to us correcting this, so disabled regulators wouldn't count towards the total load of a regulator. (So that still is on my todo) The second option would be to do what is done in the Qualcomm codeaurora kernels, where the regulators are never disabled and there's a smaller current specified for the suspend case. This would require bigger changes to the mmc core I presume, but the naming of the dt properties support such future adaption. > Moreover, wouldn't vmmc|vqmmc-active-current be depending what eMMC/SD > card that is attached? > I would say so, I do however not know how to acquire or encode that information. Any suggestions here are welcome. Regards, Bjorn