From mboxrd@z Thu Jan 1 00:00:00 1970 From: Romain Perier Subject: Re: [PATCH v1 0/4] net: stmmac: dwmac-rk: Fix phy regulator issues Date: Tue, 20 Jan 2015 07:58:43 +0100 Message-ID: References: <1421690889-11901-1-git-send-email-romain.perier@gmail.com> <1446006.BXnWRvK5p3@phil> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: <1446006.BXnWRvK5p3@phil> Sender: linux-kernel-owner@vger.kernel.org To: =?UTF-8?Q?Heiko_St=C3=BCbner?= Cc: David Miller , peppe.cavallaro@st.com, netdev , Linux Kernel Mailing List , "open list:ARM/Rockchip SoC..." , "linux-arm-kernel@lists.infradead.org" , devicetree , roger.chen@rock-chips.com List-Id: devicetree@vger.kernel.org Hi all, 2015-01-19 21:19 GMT+01:00 Heiko St=C3=BCbner : > Hi Romain > > Am Montag, 19. Januar 2015, 18:08:05 schrieb Romain Perier: >> This series fixes few issues in dwmac-rk: >> >> 1. Voltage settings was hardcoded into the driver for the phy regula= tor. >> The driver now uses the default voltage settings found in the dev= icetree, >> which are applied throught the regulator framework. >> 2. The regulator name used to power on or power off the phy was put = in the >> devicetree variable "phy_regulator", which is not standard and added= a lot >> of code for nothing. The driver now uses the devicetree property >> "phy-supply" and the corresponding functions to manipulate this regu= lator. >> >> The corresponding devicetree files are also updated. As, dwmac-rk wa= s >> recently pushed in the development tree, I don't care about devicetr= ee >> backward compatibility issues. > > This last sentence is slightly misleading :-) . > Yes, I meant that I don't need to care about it, as you explain it well= below. Sorry for my misleading sentence ;) . I will fix it in my second serie. > The actual fact is, that these new bindings for the rk3288 gmac have = not been > released with any official kernel release yet ... i.e. the will be re= leased with > 3.20 in whatever form, so we don't _need_ to care about keeping compa= tibility > still for the next 2.5 weeks or so. > > @Dave: it would be good if this series (when fixed) could still go in= to the > 3.20 material so we don't get stuck with the non-standard regulator p= roperty. > > > As we'll probably need a v2 due to at the issue in patch3, could you = also > switch places of patch1 and 2, which would keep bisecatbility (i.e. r= egulator > property before removing the voltage setting from the driver). This sentence about re-ordering patches is for Dave or for me ? Thanks, Romain