From mboxrd@z Thu Jan 1 00:00:00 1970 From: jgunthorpe@obsidianresearch.com (Jason Gunthorpe) Date: Fri, 21 Mar 2014 10:49:43 -0600 Subject: [PATCH v4 6/9] ARM: shmobile: Add PCIe device tree nodes for R8A7790 In-Reply-To: References: <1395397968-6242-1-git-send-email-phil.edworthy@renesas.com> <532C5ADE.4030708@cogentembedded.com> <6044873.2S2Xlhp0hC@wuerfel> <532C6BF7.1050404@cogentembedded.com> Message-ID: <20140321164943.GB3374@obsidianresearch.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Fri, Mar 21, 2014 at 04:31:24PM +0000, Phil.Edworthy at renesas.com wrote: > > OK, I have found where it does that but the recommended value > > seems to be "pciex" for PCI-Express case. > > Ok, but why then do all the other PCIe hosts use "pci"? Because that is what the PCI bus binding standard says to do, and new drivers have been pushed toward conforming to that document. I'm not sure if "pciex" was ever standardized, I haven't see a spec that reflects that at least. Commit 14e2abb732e485ee57d9d5b2cb8884652238e5c1 introduced support for the "pciex" word, but it looks driven by proprietary choices by IBM.. In any event there isn't a single 'device_type = "pciex";' pre-exsting in the kernel bindings today, so please don't add one :) Regards, Jason