From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jason Gunthorpe Subject: Re: [PATCH v2 03/18] PCI: designware: Configuration space should be specified in 'reg' Date: Thu, 29 May 2014 10:30:37 -0600 Message-ID: <20140529163037.GC2552@obsidianresearch.com> References: <1401345500-20188-1-git-send-email-kishon@ti.com> <1401345500-20188-4-git-send-email-kishon@ti.com> <98E18225-3C63-4106-901A-A5D3DEC268C8@codeaurora.org> <20140529151840.GD1677@bart.dudau.co.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: Content-Disposition: inline In-Reply-To: Sender: linux-pci-owner@vger.kernel.org To: Kumar Gala Cc: Liviu Dudau , Kishon Vijay Abraham I , devicetree , linux-doc@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-omap@vger.kernel.org, linux-pci@vger.kernel.org, linux-kernel@vger.kernel.org, Marek Vasut , Arnd Bergmann , tony@atomide.com, Mohit Kumar , Jingoo Han , Bjorn Helgaas List-Id: devicetree@vger.kernel.org On Thu, May 29, 2014 at 11:03:36AM -0500, Kumar Gala wrote: > Just because the kernel doesn=E2=80=99t handle this is NO reason to c= hange > the way the DT works. The OF specs do not specify how to process a config type ranges entry, and we all mutually agreed that the only sane interpretation for such a thing would be to describe an ECAM memory space so generic code could potentially make use of it. Since designware is not ECAM it should not use config ranges. This has come up multiple times now, and the above is the consensus. Jason