On Fri, Dec 2, 2022 at 9:31 AM Andrew Lunn <andrew@xxxxxxx> wrote: > > > > > > port@5 { > > > > > reg = <5>; > > > > > label = "cpu"; > > > > > ethernet = <&fec>; > > > > > + phy-mode = "rgmii-id"; > > > > > + > > > > > + fixed-link { > > > > > + speed = <1000>; > > > > > + full-duplex; > > > > > + }; > > > > > }; > > > > > > This part is needed to make a warning go away. Does the SoC network interface > > > have phy-mode = "rgmii"; ? > > > > No, it looks like this: > > > > &fec { > > pinctrl-names = "default"; > > pinctrl-0 = <&pinctrl_enet>; > > phy-mode = "rgmii-id"; > > > Is something here wrong? > > It suggests both ends should insert RGMII delays. So you will end up > with double delays. Have one end say plain "rgmii" and the other > "rgmii-id". I would probably go with the FEC being "rgmii". > > Andrew Andrew, That makes sense - I will change the fec node to rgmii. Upon further testing I find there is something else wrong with this patch however that I don't yet understand. Without it the switch works fine (due to RGMII delay being configured via boot firmware) but I do get the warning you had mentioned due to the phy-mode/phy-handle props missing: mv88e6085 2188000.ethernet-1:00: switch 0x1760 detected: Marvell 88E6176, revision 1 mv88e6085 2188000.ethernet-1:00: OF node /soc/bus@2100000/ethernet@2188000/mdio/switch@0/ports/port@5 of CPU port 5 lacks the required "phy-mode" property mv88e6085 2188000.ethernet-1:00: OF node /soc/bus@2100000/ethernet@2188000/mdio/switch@0/ports/port@5 of CPU port 5 lacks the required "phy-handle", "fixed-link" or "managed" properties mv88e6085 2188000.ethernet-1:00: Skipping phylink registration for CPU port 5 mv88e6085 2188000.ethernet-1:00 lan4 (uninitialized): PHY [mv88e6xxx-1:00] driver [Generic PHY] (irq=POLL) mv88e6085 2188000.ethernet-1:00 lan3 (uninitialized): PHY [mv88e6xxx-1:01] driver [Generic PHY] (irq=POLL) mv88e6085 2188000.ethernet-1:00 lan2 (uninitialized): PHY [mv88e6xxx-1:02] driver [Generic PHY] (irq=POLL) mv88e6085 2188000.ethernet-1:00 lan1 (uninitialized): PHY [mv88e6xxx-1:03] driver [Generic PHY] (irq=POLL) When I add the phy-mode/phy-handle props with this patch I get the following failure: mv88e6085 2188000.ethernet-1:00: switch 0x1760 detected: Marvell 88E6176, revision 1 mv88e6085 2188000.ethernet-1:00: switch 0x1760 detected: Marvell 88E6176, revision 1 mv88e6085 2188000.ethernet-1:00: configuring for fixed/rgmii-id link mode mv88e6085 2188000.ethernet-1:00: p5: delay RXCLK yes, TXCLK yes mv88e6085 2188000.ethernet-1:00: p5: delay RXCLK yes, TXCLK yes mv88e6085 2188000.ethernet-1:00: Link is Up - 1Gbps/Full - flow control off mv88e6085 2188000.ethernet-1:00 lan4 (uninitialized): validation of internal with support 00000000,00000000,000062ff and advertisement 00000000,00000000,000062ff failed: -EINVAL mv88e6085 2188000.ethernet-1:00 lan4 (uninitialized): failed to connect to PHY: -EINVAL mv88e6085 2188000.ethernet-1:00 lan4 (uninitialized): error -22 setting up PHY for tree 0, switch 0, port 0 mv88e6085 2188000.ethernet-1:00 lan3 (uninitialized): validation of internal with support 00000000,00000000,000062ff and advertisement 00000000,00000000,000062ff failed: -EINVAL mv88e6085 2188000.ethernet-1:00 lan3 (uninitialized): failed to connect to PHY: -EINVAL mv88e6085 2188000.ethernet-1:00 lan3 (uninitialized): error -22 setting up PHY for tree 0, switch 0, port 1 mv88e6085 2188000.ethernet-1:00 lan2 (uninitialized): validation of internal with support 00000000,00000000,000062ff and advertisement 00000000,00000000,000062ff failed: -EINVAL mv88e6085 2188000.ethernet-1:00 lan2 (uninitialized): failed to connect to PHY: -EINVAL mv88e6085 2188000.ethernet-1:00 lan2 (uninitialized): error -22 setting up PHY for tree 0, switch 0, port 2 mv88e6085 2188000.ethernet-1:00 lan1 (uninitialized): validation of internal with support 00000000,00000000,000062ff and advertisement 00000000,00000000,000062ff failed: -EINVAL mv88e6085 2188000.ethernet-1:00 lan1 (uninitialized): failed to connect to PHY: -EINVAL mv88e6085 2188000.ethernet-1:00 lan1 (uninitialized): error -22 setting up PHY for tree 0, switch 0, port 3 I've run into this message before and had a hard time understanding the issue from the message - it seems to indicate the phy status matches advertisement but that its an invalid mode? Thanks, Tim