From mboxrd@z Thu Jan 1 00:00:00 1970 From: Charles Keepax Subject: Re: [PATCH v5 1/2] mfd: arizona: Update reset pin to use GPIOD Date: Mon, 26 Mar 2018 09:34:04 +0100 Message-ID: <20180326083404.ewogeas2rdgbfvhh@localhost.localdomain> References: <20180312155201.30108-1-ckeepax@opensource.cirrus.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: Linus Walleij Cc: Lee Jones , Rob Herring , Mark Rutland , "open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS" , "linux-kernel@vger.kernel.org" , patches@opensource.cirrus.com List-Id: devicetree@vger.kernel.org On Sat, Mar 24, 2018 at 03:32:05PM +0100, Linus Walleij wrote: > On Mon, Mar 12, 2018 at 4:52 PM, Charles Keepax > wrote: > > > Now GPIOD has support for both pdata systems and for non-standard DT > > bindings the Arizona reset GPIO can be converted to use it. Worth > > noting gpiod_set_raw_value_cansleep is used to match the behaviour > > of the old GPIOs. This is because the part is fairly widely used and > > it is unknown how many DTs are correctly setting active low through > > device tree, so to avoid breaking any existing users it is best to > > match the previous behaviour. > > > > Signed-off-by: Charles Keepax > > --- > The use of gpiod_set_raw* should be done with caution, but this > is one of the cases where we need to take out the big hammer to > make sure things stay compatible. > > I haven't read why it is like so but I guess because the right flags > in the device tree can not be guaranteed? > Exactly yes there are a lot of these parts out there and since they never required the DT flags to be set correctly before I very much doubt all those DTs are set correctly. Thanks, Charles