From mboxrd@z Thu Jan 1 00:00:00 1970 From: hdegoede@redhat.com (Hans de Goede) Date: Tue, 11 Feb 2014 16:54:28 +0100 Subject: ohci-/ehci-platform: Change compatible string to ?hci-platform In-Reply-To: <1635726.gc3dEs0zOi@wuerfel> References: <1635726.gc3dEs0zOi@wuerfel> Message-ID: <52FA47B4.50507@redhat.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org Hi, On 02/11/2014 04:43 PM, Arnd Bergmann wrote: > On Tuesday 11 February 2014 10:27:12 Alan Stern wrote: >> >> It might even be a good idea to change the "xhci-platform" string to >> match, it that doesn't cause too much trouble. > > The original xhci binding was contributed by Al Cooper, but I don't > see any dts files using it. I agree that xhci-generic is a better > name than xhci-platform, and I think we should make that the recommended > string. If Al or someone thinks the xhci-generic string might already > be used in production devices, we should however allow both names > in the binding and in the driver. I can live with generic as pre/post fix. Looking at what seems to be the common trend I believe it should be a pre-fix and nost a post-fix though. The common way to build a compatible string seems to be: vendor,soc_model-function. We have a few deviations from this in the usb bindings, but ie the interrupt-controller bindings use this structure for the compatible strings everywhere. So since generic has no vendor, we fill just the soc_model-function part leading to: generic-ohci generic-ehci And possible also: generic-uhci generic-xhci So I'm going to respin my 2 fixup patches to move from ?hci-platform to generic-?hci. Regards, Hans