From mboxrd@z Thu Jan 1 00:00:00 1970 From: Arnd Bergmann Subject: Re: [PATCH 0/5] usb: phy: samsung: Introducing usb phy driver for samsung SoCs Date: Thu, 2 Aug 2012 11:26:52 +0000 Message-ID: <201208021126.52307.arnd@arndb.de> References: <1343826351-8756-1-git-send-email-p.paneri@samsung.com> <201208011520.02353.arnd@arndb.de> Mime-Version: 1.0 Content-Type: Text/Plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Return-path: Received: from moutng.kundenserver.de ([212.227.126.186]:62848 "EHLO moutng.kundenserver.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752558Ab2HBL06 (ORCPT ); Thu, 2 Aug 2012 07:26:58 -0400 In-Reply-To: Sender: linux-samsung-soc-owner@vger.kernel.org List-Id: linux-samsung-soc@vger.kernel.org To: Praveen Paneri Cc: linux-arm-kernel@lists.infradead.org, linux-usb@vger.kernel.org, l.majewski@samsung.com, kgene.kim@samsung.com, gregkh@linuxfoundation.org, devicetree-discuss@lists.ozlabs.org, broonie@opensource.wolfsonmicro.com, balbi@ti.com, grant.likely@secretlab.ca, kyungmin.park@samsung.com, linux-samsung-soc@vger.kernel.org, thomas.abraham@linaro.org, ben-linux@fluff.org On Thursday 02 August 2012, Praveen Paneri wrote: > Yes! I understand this problem and this is the reason these patches > were sitting in my system for couple of weeks. In a discussion with > Thomas an idea of using the existing regulator framework to > enable/disable numerous PHYs came up. For example the PMU unit > of Exynos4210 has a register set dedicated just to control USBD_PHY, > HDMI_PHY, MIPI_PHY, DAC_PHY and more. If a regulator with > each phy control register as LDO is written, the phy driver becomes a > static consumer and will modify as below. This is roughly what I had in mind, yes. The part I'm not sure about is the subsystem to use. One could obviously express the same logic using the clock or gpio framework, which would work but be conceptually wrong. Some other parts of the PMU functionality are provided through pm-domains rather than regulators and I guess in theory it could all be controlled through pm-domains. Maybe someone else has a better understanding than me what the tradeoffs are here and which subsystem should be used for the PMU. Arnd From mboxrd@z Thu Jan 1 00:00:00 1970 From: arnd@arndb.de (Arnd Bergmann) Date: Thu, 2 Aug 2012 11:26:52 +0000 Subject: [PATCH 0/5] usb: phy: samsung: Introducing usb phy driver for samsung SoCs In-Reply-To: References: <1343826351-8756-1-git-send-email-p.paneri@samsung.com> <201208011520.02353.arnd@arndb.de> Message-ID: <201208021126.52307.arnd@arndb.de> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Thursday 02 August 2012, Praveen Paneri wrote: > Yes! I understand this problem and this is the reason these patches > were sitting in my system for couple of weeks. In a discussion with > Thomas an idea of using the existing regulator framework to > enable/disable numerous PHYs came up. For example the PMU unit > of Exynos4210 has a register set dedicated just to control USBD_PHY, > HDMI_PHY, MIPI_PHY, DAC_PHY and more. If a regulator with > each phy control register as LDO is written, the phy driver becomes a > static consumer and will modify as below. This is roughly what I had in mind, yes. The part I'm not sure about is the subsystem to use. One could obviously express the same logic using the clock or gpio framework, which would work but be conceptually wrong. Some other parts of the PMU functionality are provided through pm-domains rather than regulators and I guess in theory it could all be controlled through pm-domains. Maybe someone else has a better understanding than me what the tradeoffs are here and which subsystem should be used for the PMU. Arnd