if_awg: Support new emac bindings

[kevans]

Summary

Highlights of the new bindings:

• ahb clock is specified as 'stmmaceth'
• The PHY to be used is now specified as phy-handle
• We must now check the parent of the node phy-handle points to in order to discover if we're using internal PHY.
• The ephy clk/reset will be specified on the PHY node, not the emac node.

[manu]

Works for me :

Tested on OrangePi-One (H3 with internal phy)
awg0: <Allwinner Gigabit Ethernet> mem 0x1c30000-0x1c3ffff irq 21 on simplebus0
simplebus0: no default resources for rid = 1, type = 3
awg0: PHY type: mii, conf mode: reg
awg0: Using internal PHY
awg0: EMAC clock: 0x00168000
awg0: AHB frequency 300000000 Hz, MDC div: 0x3

Tested On NanoPi M1 Plus (H3 with external phy) :
awg0: <Allwinner Gigabit Ethernet> mem 0x1c30000-0x1c3ffff irq 21 on simplebus0
simplebus0: no default resources for rid = 1, type = 3
awg0: clk_ephy not found
awg0: rst_ephy not found
awg0: PHY type: rgmii, conf mode: reg
awg0: No internal PHY
awg0: EMAC clock: 0x00050006
awg0: AHB frequency 300000000 Hz, MDC div: 0x3

Just a little nit, the printf that print that clk/rst for ephy isn't found when it's not enabled.
Also should we attach is phy_node == 0 ? That would mean that either internal or external phy was found no ?

[manu]

All test were done using linux 4.15-rc6 dts with u-boot from ports. Booted from sdcard or netbooted.

[manu]

And Tested on Pine64-LTS (with dts from sunxi/for-next https://git.kernel.org/pub/scm/linux/kernel/git/sunxi/linux.git/tree/arch/arm64/boot/dts/allwinner?h=sunxi/for-next)

[kevans]

In D13777#288714, @manu wrote:

Works for me :

Dang =)

Just a little nit, the printf that print that clk/rst for ephy isn't found when it's not enabled.

I intend to remove these, unless you think they'd be helpful in debugging other stuff- then I could fix them.

Also should we attach is phy_node == 0 ? That would mean that either internal or external phy was found no ?

IIRC, that would break our legacy compatibility. The phy-handle is a new binding thing (again, IIRC), so we'll need to maintain it for some time like this.

[kevans]

• Removed a bunch of device_printf's that were added in frustration while trying to figure out why it wasn't working on my OrangePi One.
• Made awg fail to attach if we have a phy-handle that we failed to resolve to a node.

[kevans]

@manu - any further objections? I'll add some "Tested on" lines to the commit message prior to commit.