[PATCH] net: macb: add phy-handle support for the macb

This patch adds support for the 'phy-handle' binding which allows for a
system to specifically select a phy which can be attached via any MDIO
bus available in the system.

Signed-off-by: Kedareswara rao Appana <redacted>
---
 Documentation/devicetree/bindings/net/macb.txt |  3 +++
 drivers/net/ethernet/cadence/macb.c            | 11 +++++++++--
 drivers/net/ethernet/cadence/macb.h            |  1 +
 3 files changed, 13 insertions(+), 2 deletions(-)

diff --git a/Documentation/devicetree/bindings/net/macb.txt b/Documentation/devicetree/bindings/net/macb.txt
index b5a42df..3cfff7c 100644
--- a/Documentation/devicetree/bindings/net/macb.txt
+++ b/Documentation/devicetree/bindings/net/macb.txt

@@ -23,6 +23,9 @@ Required properties:
 	Optional elements: 'tx_clk'
 - clocks: Phandles to input clocks.
 
+Optional properties:
+- phy-handle	: See ethernet.txt file in the same directory.
+
 Optional properties for PHY child node:
 - reset-gpios : Should specify the gpio for phy reset
 - magic-packet : If present, indicates that the hardware supports waking

diff --git a/drivers/net/ethernet/cadence/macb.c b/drivers/net/ethernet/cadence/macb.c
index 6b797e3..4ddd45d 100644
--- a/drivers/net/ethernet/cadence/macb.c
+++ b/drivers/net/ethernet/cadence/macb.c

@@ -421,7 +421,7 @@ static int macb_mii_probe(struct net_device *dev)
 static int macb_mii_init(struct macb *bp)
 {
 	struct macb_platform_data *pdata;
-	struct device_node *np;
+	struct device_node *np, *np1;
 	int err = -ENXIO, i;
 
 	/* Enable management port */

@@ -445,7 +445,13 @@ static int macb_mii_init(struct macb *bp)
 	dev_set_drvdata(&bp->dev->dev, bp->mii_bus);
 
 	np = bp->pdev->dev.of_node;
-	if (np) {
+	np1 = of_get_parent(bp->phy_node);
+	if (np1) {
+		of_node_put(np1);
+		err = of_mdiobus_register(bp->mii_bus, np1);
+		if (err)
+			goto err_out_unregister_bus;
+	} else if (np) {
 		/* try dt phy registration */
 		err = of_mdiobus_register(bp->mii_bus, np);
 

@@ -3016,6 +3022,7 @@ static int macb_probe(struct platform_device *pdev)
 	} else {
 		bp->phy_interface = err;
 	}
+	bp->phy_node = of_parse_phandle(np, "phy-handle", 0);
 
 	/* IP specific init */
 	err = init(pdev);

diff --git a/drivers/net/ethernet/cadence/macb.h b/drivers/net/ethernet/cadence/macb.h
index aa3aeec..83d1617 100644
--- a/drivers/net/ethernet/cadence/macb.h
+++ b/drivers/net/ethernet/cadence/macb.h

@@ -855,6 +855,7 @@ struct macb {
 	unsigned int		jumbo_max_len;
 
 	u32			wol;
+	struct device_node	*phy_node;
 };
 
 static inline bool macb_is_gem(struct macb *bp)

-- 
2.1.2

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

From: Kedareswara rao Appana <appana.durga.rao@xilinx.com>
Date: Sat, 13 Aug 2016 15:31:49 +0530

I don't know about this, all OF nodes other than the root have a
parent node which is non-NULL.  This parent node can be anything.

Just blinding assuming that any parent node of the phy_node is
what we are looking for, without any other supplementary checks
at all, seems to be asking for trouble at the very least.

Hi David Miller,

	Thanks for the review...

From: Kedareswara rao Appana <appana.durga.rao@xilinx.com>
Date: Sat, 13 Aug 2016 15:31:49 +0530

quoted

@@ -445,7 +445,13 @@ static int macb_mii_init(struct macb *bp)
 	dev_set_drvdata(&bp->dev->dev, bp->mii_bus);

 	np = bp->pdev->dev.of_node;
-	if (np) {
+	np1 = of_get_parent(bp->phy_node);
+	if (np1) {
+		of_node_put(np1);
+		err = of_mdiobus_register(bp->mii_bus, np1);
+		if (err)
+			goto err_out_unregister_bus;
+	} else if (np) {

I don't know about this, all OF nodes other than the root have a parent node
which is non-NULL.  This parent node can be anything.

Just blinding assuming that any parent node of the phy_node is what we are
looking for, without any other supplementary checks at all, seems to be asking
for trouble at the very least.

Agree with you my intention is if there is a MDIO bus on the device-tree
The MAC driver should create PHY/MDIO devices using of_mdiobus_register().

How about below code...

-       struct device_node *np;
+       struct device_node *np, *np1;
        int err = -ENXIO, i;
 
        /* Enable management port */

@@ -445,7 +445,14 @@ static int macb_mii_init(struct macb *bp)
        dev_set_drvdata(&bp->dev->dev, bp->mii_bus);
 
        np = bp->pdev->dev.of_node;
-       if (np) {
+       np1 = of_get_child_by_name(np, "mdio");
+       if (np1) {
+               of_node_put(np1);
+               err = of_mdiobus_register(bp->mii_bus, np1);
+               if (err)
+                       goto err_out_unregister_bus;
+       } else if (np) {
                /* try dt phy registration */
                err = of_mdiobus_register(bp->mii_bus, np);

If you are ok with the above code please let me know will post it as v2...

Regards,
Kedar.

Agree with you my intention is if there is a MDIO bus on the device-tree
The MAC driver should create PHY/MDIO devices using of_mdiobus_register().

What you suggest is better, and is similar to what other drivers use.

In order to keep backwards compatibility with phy nodes in the MAC
node, you have to do of_mdiobus_register() with the MAC node. However,
please change the binding documentation to say this is deprecated, all
phy nodes should be placed inside an mdio node.

Using some better variable names would also help. mac_np and mdio_np
for example.

    Andrew

Hi Andrew,

	Thanks for the review...

quoted

Agree with you my intention is if there is a MDIO bus on the
device-tree The MAC driver should create PHY/MDIO devices using

of_mdiobus_register().

What you suggest is better, and is similar to what other drivers use.

In order to keep backwards compatibility with phy nodes in the MAC node, you
have to do of_mdiobus_register() with the MAC node. However, please change
the binding documentation to say this is deprecated, all phy nodes should be
placed inside an mdio node.

Using some better variable names would also help. mac_np and mdio_np for
example.

Ok sure will post v2 as you suggested...

Regards,
Kedar.

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html