From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mika Westerberg Subject: Re: [PATCH] i2c: designware: use enable on resume instead initialization Date: Mon, 15 Jun 2015 12:29:03 +0300 Message-ID: <20150615092903.GG1478@lahna.fi.intel.com> References: <1433785828-4100-1-git-send-email-lucas.de.marchi@gmail.com> <20150610075555.GO1478@lahna.fi.intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org To: Lucas De Marchi Cc: linux-i2c@vger.kernel.org, lkml , Wolfram Sang , Jarkko Nikula , Fabio Mello , Lucas De Marchi List-Id: linux-i2c@vger.kernel.org On Fri, Jun 12, 2015 at 07:45:00PM -0300, Lucas De Marchi wrote: > Hi Mika, > > On Wed, Jun 10, 2015 at 4:55 AM, Mika Westerberg > wrote: > > On Mon, Jun 08, 2015 at 02:50:28PM -0300, lucas.de.marchi@gmail.com wrote: > >> @@ -320,7 +320,7 @@ static int dw_i2c_resume(struct device *dev) > >> clk_prepare_enable(i_dev->clk); > >> > >> if (!i_dev->pm_runtime_disabled) > >> - i2c_dw_init(i_dev); > >> + i2c_dw_enable(i_dev); > > > > This will not work if the device power gets removed (for example being > > put to D3cold) as it looses context. > > Do you mean we should keep the i2c_dw_init() here? Yes, that's safer if the controller power gets removed.