Skip to contentSkip to conversation

net: phy: disable EEE on TI PHYs

PATCH In review Russell King (Oracle) 3 participants 5 messages Nov 13, 2025

Subsystems (3)
lore
Share & tools
lore

Russell King (Oracle)

author reviewer
#0 lore raw

v2 cover net: phy: disable EEE on TI PHYs

Hi,

Towards the end of October, we discussed EEE on TI PHYs which seems to cause problems with the stmmac hardware. This problem was never fully diagnosed, but it was identified that TI PHYs do not support LPI signalling, but report that EEE is supported, and they implement the advertisement registers and that functionality.

This series allows PHY drivers to disable EEE support.

v2:
- integrate Oleksij Rempel's review comments, and merge update into patch 2 to allow EEE on non-1G variants.

 drivers/net/phy/dp83867.c    |  1 +
 drivers/net/phy/dp83869.c    |  1 +
 drivers/net/phy/phy-core.c   |  2 --
 drivers/net/phy/phy_device.c | 32 +++++++++++++++++++++++++++++---
 include/linux/phy.h          |  1 +
 5 files changed, 32 insertions(+), 5 deletions(-)
Signature · 2 lines
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 80Mbps down 10Mbps up. Decent connectivity at last!
#1 lore raw

PATCH v2 1/2 net: phy: allow drivers to disable EEE support via .get_features()

Allow PHY drivers to hook the .get_features() method to disable EEE support. This is useful for TI PHYs, where we have a statement that none of their gigabit products support EEE, yet at least DP83867 reports EEE capabilties and implements EEE negotiation.

---
 drivers/net/phy/phy-core.c   |  2 --
 drivers/net/phy/phy_device.c | 32 +++++++++++++++++++++++++++++---
 include/linux/phy.h          |  1 +
 3 files changed, 30 insertions(+), 5 deletions(-)
drivers/net/phy/phy-core.c +0 -2
index 605ca20ae192..43ccbd3a09f8 100644
@@ -207,8 +207,6 @@ void of_set_phy_eee_broken(struct phy_device *phydev)
 	if (!IS_ENABLED(CONFIG_OF_MDIO) || !node)
 		return;
 
-	linkmode_zero(modes);
-
 	if (of_property_read_bool(node, "eee-broken-100tx"))
 		linkmode_set_bit(ETHTOOL_LINK_MODE_100baseT_Full_BIT, modes);
 	if (of_property_read_bool(node, "eee-broken-1000t"))
drivers/net/phy/phy_device.c +29 -3
index 81984d4ebb7c..b384f99a40f2 100644
@@ -3397,6 +3397,34 @@ struct fwnode_handle *fwnode_get_phy_node(const struct fwnode_handle *fwnode)
 }
 EXPORT_SYMBOL_GPL(fwnode_get_phy_node);
 
+static int phy_get_features(struct phy_device *phydev)
+{
+	int err;
+
+	if (phydev->is_c45)
+		err = genphy_c45_pma_read_abilities(phydev);
+	else
+		err = genphy_read_abilities(phydev);
+
+	return err;
+}
+
+/**
+ * phy_get_features_no_eee - read the PHY features, disabling all EEE
+ * @phydev: phy_device structure to be added to the MDIO bus
+ *
+ * Read the PHY features, and fill the @phydev->eee_disabled_modes to
+ * prevent EEE being used. This is intended to be used for PHY .get_feature
+ * methods where a PHY reports incorrect capabilities.
+ */
+int phy_get_features_no_eee(struct phy_device *phydev)
+{
+	linkmode_fill(phydev->eee_disabled_modes);
+
+	return phy_get_features(phydev);
+}
+EXPORT_SYMBOL_GPL(phy_get_features_no_eee);
+
 /**
  * phy_probe - probe and init a PHY device
  * @dev: device to probe and init
@@ -3442,10 +3470,8 @@ static int phy_probe(struct device *dev)
 	}
 	else if (phydrv->get_features)
 		err = phydrv->get_features(phydev);
-	else if (phydev->is_c45)
-		err = genphy_c45_pma_read_abilities(phydev);
 	else
-		err = genphy_read_abilities(phydev);
+		err = phy_get_features(phydev);
 
 	if (err)
 		goto out;
include/linux/phy.h +1 -0
index bf5457341ca8..2655c0ae6488 100644
@@ -2299,6 +2299,7 @@ void phy_support_sym_pause(struct phy_device *phydev);
 void phy_support_asym_pause(struct phy_device *phydev);
 void phy_support_eee(struct phy_device *phydev);
 void phy_disable_eee(struct phy_device *phydev);
+int phy_get_features_no_eee(struct phy_device *phydev);
 void phy_set_sym_pause(struct phy_device *phydev, bool rx, bool tx,
 		       bool autoneg);
 void phy_set_asym_pause(struct phy_device *phydev, bool rx, bool tx);
Signature · 1 line
2.47.3
#2 lore raw

PATCH v2 2/2 net: phy: TI PHYs use phy_get_features_no_eee()

In reply to Russell King (Oracle)

As TI Gigabit PHYs do not support EEE, use the newly introduced phy_get_features_no_eee() to read the features but mark EEE as disabled.

---
 drivers/net/phy/dp83867.c | 1 +
 drivers/net/phy/dp83869.c | 1 +
 2 files changed, 2 insertions(+)
drivers/net/phy/dp83867.c +1 -0
index 36a0c1b7f59c..da055ff861be 100644
@@ -1124,6 +1124,7 @@ static struct phy_driver dp83867_driver[] = {
 		/* PHY_GBIT_FEATURES */
 
 		.probe          = dp83867_probe,
+		.get_features	= phy_get_features_no_eee,
 		.config_init	= dp83867_config_init,
 		.soft_reset	= dp83867_phy_reset,
 
drivers/net/phy/dp83869.c +1 -0
index 1f381d7b13ff..4400654b0f72 100644
@@ -906,6 +906,7 @@ static int dp83869_phy_reset(struct phy_device *phydev)
 	PHY_ID_MATCH_MODEL(_id),				\
 	.name		= (_name),				\
 	.probe          = dp83869_probe,			\
+	.get_features	= phy_get_features_no_eee,		\
 	.config_init	= dp83869_config_init,			\
 	.soft_reset	= dp83869_phy_reset,			\
 	.config_intr	= dp83869_config_intr,			\
Signature · 1 line
2.47.3
maintainer
#3 lore raw
In reply to Russell King (Oracle)
Quoted 14 lines · Russell King (Oracle)

Hi,

Towards the end of October, we discussed EEE on TI PHYs which seems to cause problems with the stmmac hardware. This problem was never fully diagnosed, but it was identified that TI PHYs do not support LPI signalling, but report that EEE is supported, and they implement the advertisement registers and that functionality.

This series allows PHY drivers to disable EEE support.

v2:
- integrate Oleksij Rempel's review comments, and merge update into patch 2 to allow EEE on non-1G variants.

Quoted 7 lines
 drivers/net/phy/dp83867.c    |  1 +
 drivers/net/phy/dp83869.c    |  1 +
 drivers/net/phy/phy-core.c   |  2 --
 drivers/net/phy/phy_device.c | 32 +++++++++++++++++++++++++++++---
 include/linux/phy.h          |  1 +
 5 files changed, 32 insertions(+), 5 deletions(-)

Alternatively the PHY driver could call phy_disable_eee() in its config_init. Then we wouldn't have to touch core code.

author reviewer
#4 lore raw
Quoted 23 lines · Heiner Kallweit
Quoted 14 lines · Russell King (Oracle)

Hi,

Towards the end of October, we discussed EEE on TI PHYs which seems to cause problems with the stmmac hardware. This problem was never fully diagnosed, but it was identified that TI PHYs do not support LPI signalling, but report that EEE is supported, and they implement the advertisement registers and that functionality.

This series allows PHY drivers to disable EEE support.

v2:
- integrate Oleksij Rempel's review comments, and merge update into patch 2 to allow EEE on non-1G variants.

Quoted 7 lines
 drivers/net/phy/dp83867.c    |  1 +
 drivers/net/phy/dp83869.c    |  1 +
 drivers/net/phy/phy-core.c   |  2 --
 drivers/net/phy/phy_device.c | 32 +++++++++++++++++++++++++++++---
 include/linux/phy.h          |  1 +
 5 files changed, 32 insertions(+), 5 deletions(-)
Quoted 2 lines

Alternatively the PHY driver could call phy_disable_eee() in its config_init. Then we wouldn't have to touch core code.

If that's what you intend, then please update the kerneldoc for phy_disable_eee() so that it doesn't state that it's for MAC use (because that implies it's not for PHY drivers.)

Signature · 2 lines
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 80Mbps down 10Mbps up. Decent connectivity at last!