From mboxrd@z Thu Jan 1 00:00:00 1970 From: Arnd Bergmann Subject: Re: [PATCH v2 3/8] arm: add device-tree SoC bindings for Axis Artpec-6 Date: Thu, 11 Feb 2016 15:14:56 +0100 Message-ID: <7242985.WgQduhPVkk@wuerfel> References: <0f096d1ad277cf94032ca274ced0cb35cb5d967c.1455198819.git.larper@axis.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7Bit Return-path: In-Reply-To: <0f096d1ad277cf94032ca274ced0cb35cb5d967c.1455198819.git.larper@axis.com> Sender: linux-kernel-owner@vger.kernel.org To: linux-arm-kernel@lists.infradead.org Cc: Lars Persson , devicetree@vger.kernel.org, linux-clk@vger.kernel.org, mark.rutland@arm.com, pawel.moll@arm.com, ijc+devicetree@hellion.org.uk, Lars Persson , linux-kernel@vger.kernel.org, robh+dt@kernel.org, galak@codeaurora.org List-Id: devicetree@vger.kernel.org On Thursday 11 February 2016 15:04:37 Lars Persson wrote: > + > +The DMA configuration unit muxes dma request lines between single and burst > +alternatives. > + > +Required properties: > +- compatible: axis,artpec6-dmacfg > +- reg: Address and length of DMACFG register. > + > +Example: > + dmacfg { > + compatible = "axis,artpec6-dmacfg"; > + reg = <0xf8000010 4>; > + }; > + This method looks better than before, but putting a single register into DT looks wrong. I can almost guarantee that there are more registers at 0xf8000000-0xf8000fff (or more) that belong to either some system controller or a DMA controller. Whatever it is, please describe it in the DT in a sensible way based on what the hardware does and what other resources it uses (interrupts etc) rather than what today's kernel happens to need. Arnd