From mboxrd@z Thu Jan 1 00:00:00 1970 From: Arnd Bergmann Subject: Re: [RFC PATCH 0/5] omap: add ocp2scp as a misc driver Date: Thu, 31 May 2012 09:22:18 +0000 Message-ID: <201205310922.18550.arnd@arndb.de> References: <1338375394-16268-1-git-send-email-kishon@ti.com> <201205301843.51568.arnd@arndb.de> Mime-Version: 1.0 Content-Type: Text/Plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org To: "ABRAHAM, KISHON VIJAY" Cc: paul@pwsan.com, tony@atomide.com, linux@arm.linux.org.uk, b-cousson@ti.com, gregkh@linuxfoundation.org, linux-omap@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, balbi@ti.com List-Id: linux-omap@vger.kernel.org On Thursday 31 May 2012, ABRAHAM, KISHON VIJAY wrote: > > > > Just put the devices you actually want into the device tree instead and > > have them automatically created. > > But shouldn't device tree support and non-device tree support co-exist > till the non-device tree is completely knocked out of the kernel? Even Well, it's a new driver, so there is no regression in supporting it only on DT-enabled boards. OMAP5 is DT-only anyway and OMAP4 only supports two boards that should both be working with the generic DT code already. > then shouldn't there be a separate driver to extract the child nodes > of ocp2scp and create a device (using of_platform_populate)? > As far as I've seen, the devices are automatically created (by virtue > of of_platform_populate() in mach-omap2/board-generic.c in omap case) > only for the devices that is connected to the system bus. For other > cases, the parent should be responsible for creating the child > devices. I would mark the multiplexed device compatible with "simple-bus", which results in the child devices automatically getting added. Arnd From mboxrd@z Thu Jan 1 00:00:00 1970 From: arnd@arndb.de (Arnd Bergmann) Date: Thu, 31 May 2012 09:22:18 +0000 Subject: [RFC PATCH 0/5] omap: add ocp2scp as a misc driver In-Reply-To: References: <1338375394-16268-1-git-send-email-kishon@ti.com> <201205301843.51568.arnd@arndb.de> Message-ID: <201205310922.18550.arnd@arndb.de> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Thursday 31 May 2012, ABRAHAM, KISHON VIJAY wrote: > > > > Just put the devices you actually want into the device tree instead and > > have them automatically created. > > But shouldn't device tree support and non-device tree support co-exist > till the non-device tree is completely knocked out of the kernel? Even Well, it's a new driver, so there is no regression in supporting it only on DT-enabled boards. OMAP5 is DT-only anyway and OMAP4 only supports two boards that should both be working with the generic DT code already. > then shouldn't there be a separate driver to extract the child nodes > of ocp2scp and create a device (using of_platform_populate)? > As far as I've seen, the devices are automatically created (by virtue > of of_platform_populate() in mach-omap2/board-generic.c in omap case) > only for the devices that is connected to the system bus. For other > cases, the parent should be responsible for creating the child > devices. I would mark the multiplexed device compatible with "simple-bus", which results in the child devices automatically getting added. Arnd