Re: [PATCH 1/3] ARM: dts: imx51: ZII: Add missing phy-mode

[Vladimir Oltean <vladimir.oltean@nxp.com>]

On Mon, Apr 10, 2023 at 02:35:28PM +0200, Andrew Lunn wrote:
> I tested vf610-zii-devel-rev-c.dtb:
> 
> [    2.307416] mv88e6085 mdio_mux-0.1:00: configuring for fixed/rev-rmii link mode
> [    2.314459] mv88e6085 mdio_mux-0.1:00: configuring for fixed/xaui link mode
> [    2.320588] mv88e6085 mdio_mux-0.1:00: Link is Up - 100Mbps/Full - flow control off
> [    2.327722] mv88e6085 mdio_mux-0.2:00: configuring for fixed/xaui link mode
> [    2.334729] mv88e6085 mdio_mux-0.2:00: Link is Up - 10Gbps/Full - flow control off
> [    2.343263] mv88e6085 mdio_mux-0.1:00: Link is Up - 10Gbps/Full - flow control off
> [    2.396110] mv88e6085 mdio_mux-0.1:00 lan1 (uninitialized): PHY [!mdio-mux!mdio@1!switch@0!mdio:01] driver [Marvell 88E6390 Fa)
> [    2.498137] mv88e6085 mdio_mux-0.1:00 lan2 (uninitialized): PHY [!mdio-mux!mdio@1!switch@0!mdio:02] driver [Marvell 88E6390 Fa)
> [    2.566028] mv88e6085 mdio_mux-0.1:00 lan3 (uninitialized): PHY [!mdio-mux!mdio@1!switch@0!mdio:03] driver [Marvell 88E6390 Fa)
> [    2.663995] mv88e6085 mdio_mux-0.1:00 lan4 (uninitialized): PHY [!mdio-mux!mdio@1!switch@0!mdio:04] driver [Marvell 88E6390 Fa)
> [    2.746064] mv88e6085 mdio_mux-0.2:00 lan5 (uninitialized): PHY [!mdio-mux!mdio@2!switch@0!mdio:01] driver [Marvell 88E6390 Fa)
> [    2.834000] mv88e6085 mdio_mux-0.2:00 lan6 (uninitialized): PHY [!mdio-mux!mdio@2!switch@0!mdio:02] driver [Marvell 88E6390 Fa)
> [    2.933998] mv88e6085 mdio_mux-0.2:00 lan7 (uninitialized): PHY [!mdio-mux!mdio@2!switch@0!mdio:03] driver [Marvell 88E6390 Fa)
> [    3.016031] mv88e6085 mdio_mux-0.2:00 lan8 (uninitialized): PHY [!mdio-mux!mdio@2!switch@0!mdio:04] driver [Marvell 88E6390 Fa)
> [    3.033976] mv88e6085 mdio_mux-0.2:00 sff2 (uninitialized): switched to inband/2500base-x link mode
> 
> So we can see link mode rev-rmii and there are no errors.
> 
> Testing a board using mii, not rmii is going to be a bit harder. I
> don't have one set up right now.

hmmm... why does this work?

would you mind adding this small debug print and booting again?

diff --git a/drivers/net/phy/phylink.c b/drivers/net/phy/phylink.c
index a4111f1be375..0b9754d4db80 100644
--- a/drivers/net/phy/phylink.c
+++ b/drivers/net/phy/phylink.c
@@ -695,6 +695,10 @@ static int phylink_validate(struct phylink *pl, unsigned long *supported,
 {
 	const unsigned long *interfaces = pl->config->supported_interfaces;
 
+	phylink_err(pl, "%s: supported_interfaces=%*pbl, interface %s\n",
+		    __func__, (int)PHY_INTERFACE_MODE_MAX,
+		    interfaces, phy_modes(state->interface));
+
 	if (!phy_interface_empty(interfaces)) {
 		if (state->interface == PHY_INTERFACE_MODE_NA)
 			return phylink_validate_mask(pl, supported, state,

I would've thought this check below would fail:

		if (!test_bit(state->interface, interfaces))
			return -EINVAL;