[PATCH] phy: micrel.c: Support ksz9031 energy-detect power-down mode

[PATCH] phy: micrel.c: Support ksz9031 energy-detect power-down mode

[Mike Looijmans]

Set bit 0 in register 1C.23 to enable the EDPD feature of the
KSZ9031 PHY. This reduces power consumption when the link is
down. To use this, set "enable-edpd" in the devicetree.

Signed-off-by: Mike Looijmans <mike.looijmans@topic.nl>
---
 .../devicetree/bindings/net/micrel-ksz90x1.txt        |  6 ++++++
 drivers/net/phy/micrel.c                              | 19 +++++++++++++++++++
 2 files changed, 25 insertions(+)

diff --git a/Documentation/devicetree/bindings/net/micrel-ksz90x1.txt b/Documentation/devicetree/bindings/net/micrel-ksz90x1.txt
index f9c32ad..fbec6f8 100644
--- a/Documentation/devicetree/bindings/net/micrel-ksz90x1.txt
+++ b/Documentation/devicetree/bindings/net/micrel-ksz90x1.txt
@@ -56,6 +56,11 @@ KSZ9031:
       - txd2-skew-ps : Skew control of TX data 2 pad
       - txd3-skew-ps : Skew control of TX data 3 pad
 
+    Boolean:
+
+      - enable-edpd : Not related to timing. Specify this property to enable
+                      energy-detect power-down mode in the PHY.
+
 Examples:
 
 	mdio {
@@ -64,6 +69,7 @@ Examples:
 			rxdv-skew-ps = <0>;
 			txc-skew-ps = <3000>;
 			txen-skew-ps = <0>;
+			enable-edpd;
 			reg = <0>;
 		};
 	};
diff --git a/drivers/net/phy/micrel.c b/drivers/net/phy/micrel.c
index 4516c8a..0a365b4 100644
--- a/drivers/net/phy/micrel.c
+++ b/drivers/net/phy/micrel.c
@@ -409,6 +409,10 @@ static int ksz9021_config_init(struct phy_device *phydev)
 #define MII_KSZ9031RN_TX_DATA_PAD_SKEW	6
 #define MII_KSZ9031RN_CLK_PAD_SKEW	8
 
+/* MMD Address 0x1C */
+#define MII_KSZ9031RN_EDPD		0x23
+#define MII_KSZ9031RN_EDPD_ENABLE	BIT(0)
+
 static int ksz9031_extended_write(struct phy_device *phydev,
 				  u8 mode, u32 dev_addr, u32 regnum, u16 val)
 {
@@ -480,6 +484,18 @@ static int ksz9031_center_flp_timing(struct phy_device *phydev)
 	return genphy_restart_aneg(phydev);
 }
 
+/* Enable energy-detect power-down mode */
+static int ksz9031_enable_edpd(struct phy_device *phydev)
+{
+	int reg;
+
+	reg = ksz9031_extended_read(phydev, OP_DATA, 0x1C, MII_KSZ9031RN_EDPD);
+	if (reg < 0)
+		return reg;
+	return ksz9031_extended_write(phydev, OP_DATA, 0x1C, MII_KSZ9031RN_EDPD,
+				      reg | MII_KSZ9031RN_EDPD_ENABLE);
+}
+
 static int ksz9031_config_init(struct phy_device *phydev)
 {
 	const struct device *dev = &phydev->mdio.dev;
@@ -522,6 +538,9 @@ static int ksz9031_config_init(struct phy_device *phydev)
 		ksz9031_of_load_skew_values(phydev, of_node,
 				MII_KSZ9031RN_TX_DATA_PAD_SKEW, 4,
 				tx_data_skews, 4);
+
+		if (of_property_read_bool(of_node, "enable-edpd"))
+			ksz9031_enable_edpd(phydev);
 	}
 
 	return ksz9031_center_flp_timing(phydev);
-- 
1.9.1

Re: [PATCH] phy: micrel.c: Support ksz9031 energy-detect power-down mode

[Andrew Lunn]

> +    Boolean:
> +
> +      - enable-edpd : Not related to timing. Specify this property to enable
> +                      energy-detect power-down mode in the PHY.

The Broadcom PHYs unconditionally enable this feature.

The SMSC driver has an option: smsc,disable-energy-detect because it
seems to have a rare hardware bug when it is enabled.

And you would like to make it off by default.

Could we avoid having three drivers doing three different things?
Could you have it unconditionally on, and see if anybody reports it
being broken?

      Andrew
--
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

Re: [PATCH] phy: micrel.c: Support ksz9031 energy-detect power-down mode

[Mike Looijmans]

On 03-10-16 16:35, Andrew Lunn wrote:
>> +    Boolean:
>> +
>> +      - enable-edpd : Not related to timing. Specify this property to enable
>> +                      energy-detect power-down mode in the PHY.
>
> The Broadcom PHYs unconditionally enable this feature.
>
> The SMSC driver has an option: smsc,disable-energy-detect because it
> seems to have a rare hardware bug when it is enabled.
>
> And you would like to make it off by default.
>
> Could we avoid having three drivers doing three different things?
> Could you have it unconditionally on, and see if anybody reports it
> being broken?

Good suggestion I think. It's a feature found on many PHYs, apparently 
without any ill effects, so just enabling unconditionally simplifies 
things. And it's good for the environment...

I'll post a v2 patch (which won't need devicetree changes then).

-- 
Mike Looijmans