On 16. 11. 18 12:50, Martin Lund wrote: > Hi Naga, > > I've been working on running up the latest kernel (v4.20-rc2) on our > custom Xilinx hw board so that I can test the v12 version of your > Arasan nand driver. > > I've managed to get the driver successfully up and running and ready > for testing with a Micron MT29F64G08AFAAAWP device. However, setting > it up I've found a few inaccuracies in the documentation of the device > tree bindings. > > This is the device configuration that ended up working for me with > linux v4.20-rc2: > > nfc: nand@ff100000 { > compatible = "xlnx,zynqmp-nand", "arasan,nfc-v3p10"; > reg = <0x0 0xff100000 0x0 0x1000>; > clock-names = "clk_sys", "clk_flash"; > clocks = <&clk200>, <&clk100>; > interrupt-parent = <&gic>; > interrupts = <0 14 4>; > #address-cells = <1>; > #size-cells = <0>; > > nand@0 { > reg = <0>; > nand-ecc-mode = "hw"; > }; > }; > > Compared with the example you will notice that "clock-name" should be > "clock-names". reg was missing a "0x0". clock-names and even that names - you are right it is not correct and should be fixed. Missing 0x0 in reg doesn't matter because it depends on address/size cells. > I think it is helpful to provide a real-world working example, so you > might also consider changing the example "clocks" configuration to > clk200/clk100 since there is no clk_misc among the clock sources of > any of the xilinx zynqmp board device tree configurations. Real example is the best normally just c&p from existing dts is the way to go. But in connection to clocks it doesn't matter what exactly should be there and I don't think there is any consistency in that. Hopefully this will be removed by yaml conversion. Thanks, Michal ______________________________________________________ Linux MTD discussion mailing list http://lists.infradead.org/mailman/listinfo/linux-mtd/