On Mon, Mar 19, 2018 at 04:23:11PM +0000, Alayev Michael wrote: > Hello Andrew, > > You deserve a big thank you for your solution to our device-tree - the linux > bootup log looks a lot better and It made a big progress for us. > Though we still have some issues... > > 1. Attached are 2 log files that are the result of your suggested dts: > 1st is your device-tree as is : results in kernel panic. Its probably > caused by the "link" parameter.
The link parameters look correct. It is a pointer to the switches port in the DSA link. But i spotted a real problem. See bellow. > 2. The switch's product number should be 0x0a1 but instead its 0xa10 (is it > just a high-low byte thing?) This is correct. The product number is contains the revision. You have revision 0. > 3. The stand-alone phy (gem1) is not detected properly. It should be device > id mv88e1510. > switch0: switch@1c: { > compatible = "marvell,mv88e6190"; > reg = <0x1c>; > > ports { > #address-cells = <1>; > #size-cells = <0>; > > port@0 { > reg = <0>; > label = "cpu"; > fixed-link { > speed = <1000>; > full-duplex; > }; There is a missing property here: ethernet = <&gem0>; > port@1 { > reg = <1>; > label = "lan0"; > }; > port@2 { > reg = <2>; > label = "lan1"; > }; > ... > switch0port10: port@10 { > reg = <10>; > label = "dsa"; > link = <&switch1port10>; > phy-mode = "sgmii"; > fixed-link { > speed = <1000>; > full-duplex; > }; > }; > }; > }; > > > gem1: { > phy-handle = <phy0>; and this is missing an & phy-handle = <&phy0>; Andrew