From mboxrd@z Thu Jan 1 00:00:00 1970 From: jamie@jamieiles.com (Jamie Iles) Date: Mon, 21 Nov 2011 14:47:46 +0000 Subject: [PATCH] Initial DT support for SIMpad devices. In-Reply-To: <4ECA6118.1050806@scram.de> References: <1321821470-32396-1-git-send-email-jochen@scram.de> <20111121095020.GA7314@totoro> <4ECA6118.1050806@scram.de> Message-ID: <20111121144746.GB25627@totoro> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org Hi Jochen, On Mon, Nov 21, 2011 at 03:32:56PM +0100, Jochen Friedrich wrote: > >>+ localbus { > >>+ compatible = "intel,sa1110-localbus"; > > > >Could this claim compatibility with simple-bus? > > I wasn't sure about this. I took a look in the powerpc DTS files for reference and they used > some kind of -localbus compatible entries. So I took the same approach here. Hmm, I haven't seen any of the ARM platforms doing this, and unless the sa1110 bus has some extra knobs that need tweaking, then simple-bus should be fine. > >>+ uart2: serial at 0x80050000 { > >>+ compatible = "intel,sa1100-uart"; > >>+ reg =<0x80050000 0x24>; > >>+ interrupts =<17>; > >>+ status = "disabled"; > > > >Hmm, I couldn't see status defined in the UART binding or where it was > >used... Is this required? > > status is a global property and it's being used in drivers/of/base.c, of_device_is_available(). > It is used in other dtsi files like e.g. at91sam9g45.dtsi as well to define optional nodes. Ahh, I wasn't aware of that one - thanks! > >>+/ { > >>+ model = "SIEMENS, SIMpad"; > >>+ compatible = "siemens,simpad"; > > > >It may be worth adding the SoC compatible string after the board one for > >completeness. > > Do you mean something like this? > > compatible = "siemens,simpad", "intel,sa1100"; Yup, that looks about right. > >>+ chosen { > >>+ bootargs = "console=ttySA0"; > > > >It is preferred for the bootloader to set these up rather than having > >them statically in the DTS if at all possible. > > Yes, my boot loader does this, but simpad support is not in official U-BOOT yet. > This allows testing with a different boot loader like the hh.org one and a Linux > binary with DTB appended. OK, I guess that's fair enough, though I certainly shouldn't be considered an authoritative source on this! Jamie