From mboxrd@z Thu Jan 1 00:00:00 1970 From: benoitm@perenite.com (Benoit Masson) Date: Mon, 28 Jul 2014 13:17:56 +0200 Subject: [PATCH 2/2] ARM: mvebu: Added dts defintion for Lenovo Iomega ix4-300d NAS In-Reply-To: <6028388.mjACYPxFta@wuerfel> References: <1406418436-61452-1-git-send-email-yahoo@perenite.com> <5442226.1WCY4fjGCy@wuerfel> <78788908-DDB5-46F8-83C1-FAE312D7BFB2@perenite.com> <6028388.mjACYPxFta@wuerfel> Message-ID: <78E68BC6-1D2C-4B75-93BA-5414A358B90D@perenite.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org Le 28 juil. 2014 ? 13:02, Arnd Bergmann a ?crit : > On Monday 28 July 2014 12:50:47 Benoit Masson wrote: >>> >>> It should still list both "marvell,mv78230-i2c" and "marvell,mv64xxx-i2c", >>> as the generic name may not be enough to describe the register layout >>> correctly. >> >> I tend to disagree since the dts include the armada-xp.dtsi, which already declare >> >> "i2c0: i2c at 11000 { >> compatible = "marvell,mv78230-i2c", "marvell,mv64xxx-i2c"; >> reg = <0x11000 0x100>; >> }; >> " >> >> I tested the dts on the device and it does inherits the i2c declaration, did I miss something here ? >> > > Inheriting from armada-xp.dtsi is fine, but if you manually set > "marvell,mv64xxx-i2c", that will override the old setting > and drop the "marvell,mv78230-i2c" part, which is bad because > it may break things later if it turns out that the mv78230 > variant is not completely compatible with the old > mv64xxx variant. Ok then does this mean I can simply remove the "compatible" line to only inherits from armada-xp-dtsi ? "i2c at 11000 { clock-frequency = <400000>; status = "okay"; " Would this be enough ? Benoit > > Arnd