Skip to contentSkip to conversation

net: phy: Add support for 2.5GBASET PHYs

PATCH In review Maxime Chevallier 4 participants 27 messages Feb 7, 2019 – Feb 20, 2019

Subsystems (5)
lore
Share & tools
lore
#0 lore raw

v2 cover net: phy: Add support for 2.5GBASET PHYs

Hello everyone,

This is the second iteration of the series introducing support for 2.5GBASET and 5GBASET to the network PHY infrastructure.

These 2 modes are described in the 802.3bz specifications, and allow to use 2.5G and 5G speeds on cat5e/cat6 cables.

The required infrastructure code is added for 2.5GBASET and 5GBASET, but only 2.5GBASET support is added to the Marvell10G driver. The reason is because the 5GBASER interface mode used to communicate with the MAC isn't implemented yet, and therefore this can't be tested.

This series has seen some rework since last time, see the full details in the patches changelog. The main highlights are :

- Patch 1 moves the Pause ans Asym_Pause parameters update after calling config_init. This allows us to change the phydev->supported modes in config_init, while still forcing some quirks taken from phydrv->features, to disable unsupported Pause modes. This was proposed by Andrew.

- Patch 2 generalizes the way we mask-out modes we don't want to use when forcing the link speed through DT or ethtool, by walking through the PHY settings table, as proposed by Russell.

- Patch 4 implements automatic setting of TM, FIBRE and Backplane bits from the list of supported linkmodes.

- In Patch 5, we only read abilities from the PMA. Pause parameters aren't built from the genphy_c45_pma_read_abilities, as it was done in the previous iteration of the patch. We also amke use of linkmode_mod_bit to make sure we mask-out unsupported modes.

- In Patch 8, we manually check for the PMA device id to see if we have to use a quirk when reading the 2.5G/5G Extended Abilities

Maxime Chevallier (10):
net: phy: Update PHY linkmodes after config_init
net: phy: Mask-out non-compatible modes when setting the max-speed
net: phy: Move of_set_phy_eee_broken to phy-core.c
net: phy: Automatically fill the generic TP, FIBRE and Backplane modes
net: phy: Extract genphy_c45_pma_read_abilities from marvell10g
net: phy: Add generic support for 2.5GBaseT and 5GBaseT
net: phy: marvell10g: Add support for 2.5GBASET
net: phy: marvell10g: Force reading of 2.5/5G
net: mvpp2: Add 2.5GBaseT support
net: phy: marvell10g: add support for the 88x2110 PHY

 .../net/ethernet/marvell/mvpp2/mvpp2_main.c   |   1 +
 drivers/net/phy/marvell10g.c                  | 142 +++++------
 drivers/net/phy/phy-c45.c                     | 111 ++++++++
 drivers/net/phy/phy-core.c                    |  72 ++++++
 drivers/net/phy/phy_device.c                  | 237 +++++++++---------
 include/linux/linkmode.h                      |   6 +
 include/linux/marvell_phy.h                   |   2 +
 include/linux/phy.h                           |   3 +
 include/uapi/linux/mdio.h                     |  16 ++
 9 files changed, 405 insertions(+), 185 deletions(-)
Signature · 1 line
2.19.2
#1 lore raw

PATCH v2 1/10 net: phy: Update PHY linkmodes after config_init

We want to be able to update a PHY's supported list in the config_init callback, so move the Pause parameters settings from phydrv->features after calling config_init to make sure these parameters aren't overwritten.

---
 drivers/net/phy/phy_device.c | 89 +++++++++++++++++++++++-------------
 1 file changed, 58 insertions(+), 31 deletions(-)
drivers/net/phy/phy_device.c +58 -31
index 891e0178b97f..18a10565efd4 100644
@@ -1059,6 +1059,59 @@ static int phy_poll_reset(struct phy_device *phydev)
 	return 0;
 }
 
+/**
+ * phy_update_linkmodes - Update and sanitize linkmodes and pause parameters
+ * @phydev: The PHY device whose parameters we want to update.
+ *
+ * Description: The list of supported and advertised linkmodes is not
+ * straightforward to maintain, since PHYs and MACs are subject to quirks and
+ * erratas. This function re-builds the list of the supported pause parameters
+ * by taking into account the parameters expressed in the driver's features
+ * list.
+ */
+static void phy_update_linkmodes(struct phy_device *phydev)
+{
+	struct device_driver *drv = phydev->mdio.dev.driver;
+	struct phy_driver *phydrv = to_phy_driver(drv);
+
+	mutex_lock(&phydev->lock);
+
+	/* The Pause Frame bits indicate that the PHY can support passing
+	 * pause frames. During autonegotiation, the PHYs will determine if
+	 * they should allow pause frames to pass.  The MAC driver should then
+	 * use that result to determine whether to enable flow control via
+	 * pause frames.
+	 *
+	 * Normally, PHY drivers should not set the Pause bits, and instead
+	 * allow phylib to do that.  However, there may be some situations
+	 * (e.g. hardware erratum) where the driver wants to set only one
+	 * of these bits.
+	 */
+	if (test_bit(ETHTOOL_LINK_MODE_Pause_BIT, phydrv->features) ||
+	    test_bit(ETHTOOL_LINK_MODE_Asym_Pause_BIT, phydrv->features)) {
+		linkmode_clear_bit(ETHTOOL_LINK_MODE_Pause_BIT,
+				   phydev->supported);
+		linkmode_clear_bit(ETHTOOL_LINK_MODE_Asym_Pause_BIT,
+				   phydev->supported);
+		if (test_bit(ETHTOOL_LINK_MODE_Pause_BIT, phydrv->features))
+			linkmode_set_bit(ETHTOOL_LINK_MODE_Pause_BIT,
+					 phydev->supported);
+		if (test_bit(ETHTOOL_LINK_MODE_Asym_Pause_BIT,
+			     phydrv->features))
+			linkmode_set_bit(ETHTOOL_LINK_MODE_Asym_Pause_BIT,
+					 phydev->supported);
+	} else {
+		linkmode_set_bit(ETHTOOL_LINK_MODE_Pause_BIT,
+				 phydev->supported);
+		linkmode_set_bit(ETHTOOL_LINK_MODE_Asym_Pause_BIT,
+				 phydev->supported);
+	}
+
+	linkmode_copy(phydev->advertising, phydev->supported);
+
+	mutex_unlock(&phydev->lock);
+}
+
 int phy_init_hw(struct phy_device *phydev)
 {
 	int ret = 0;
@@ -1082,6 +1135,11 @@ int phy_init_hw(struct phy_device *phydev)
 	if (phydev->drv->config_init)
 		ret = phydev->drv->config_init(phydev);
 
+	/* Update and sanitize the supported and advertised linkmodes, since
+	 * they might have been changed in config_init
+	 */
+	phy_update_linkmodes(phydev);
+
 	return ret;
 }
 EXPORT_SYMBOL(phy_init_hw);
@@ -2221,37 +2279,6 @@ static int phy_probe(struct device *dev)
 	 */
 	of_set_phy_eee_broken(phydev);
 
-	/* The Pause Frame bits indicate that the PHY can support passing
-	 * pause frames. During autonegotiation, the PHYs will determine if
-	 * they should allow pause frames to pass.  The MAC driver should then
-	 * use that result to determine whether to enable flow control via
-	 * pause frames.
-	 *
-	 * Normally, PHY drivers should not set the Pause bits, and instead
-	 * allow phylib to do that.  However, there may be some situations
-	 * (e.g. hardware erratum) where the driver wants to set only one
-	 * of these bits.
-	 */
-	if (test_bit(ETHTOOL_LINK_MODE_Pause_BIT, phydrv->features) ||
-	    test_bit(ETHTOOL_LINK_MODE_Asym_Pause_BIT, phydrv->features)) {
-		linkmode_clear_bit(ETHTOOL_LINK_MODE_Pause_BIT,
-				   phydev->supported);
-		linkmode_clear_bit(ETHTOOL_LINK_MODE_Asym_Pause_BIT,
-				   phydev->supported);
-		if (test_bit(ETHTOOL_LINK_MODE_Pause_BIT, phydrv->features))
-			linkmode_set_bit(ETHTOOL_LINK_MODE_Pause_BIT,
-					 phydev->supported);
-		if (test_bit(ETHTOOL_LINK_MODE_Asym_Pause_BIT,
-			     phydrv->features))
-			linkmode_set_bit(ETHTOOL_LINK_MODE_Asym_Pause_BIT,
-					 phydev->supported);
-	} else {
-		linkmode_set_bit(ETHTOOL_LINK_MODE_Pause_BIT,
-				 phydev->supported);
-		linkmode_set_bit(ETHTOOL_LINK_MODE_Asym_Pause_BIT,
-				 phydev->supported);
-	}
-
 	/* Set the state to READY by default */
 	phydev->state = PHY_READY;
 
Signature · 1 line
2.19.2
#2 lore raw

Re: [PATCH net-next v2 01/10] net: phy: Update PHY linkmodes after config_init

Hello everyone,

Quoted 4 lines · On Thu, 7 Feb 2019 10:49:30 +0100 Maxime Chevallier…

We want to be able to update a PHY's supported list in the config_init callback, so move the Pause parameters settings from phydrv->features after calling config_init to make sure these parameters aren't overwritten.

I realised that I did not include updating the modes according to the limitations imposed by of_set_phy_supported in this series. If there's a speed limitation set in the DT, it could be overwritten by config_init.

I'll respin the series with this addition, sorry about that.

Maxime

maintainer
#3 lore raw

Re: [PATCH net-next v2 01/10] net: phy: Update PHY linkmodes after config_init

In reply to Maxime Chevallier
Quoted 4 lines · Maxime Chevallier

We want to be able to update a PHY's supported list in the config_init callback, so move the Pause parameters settings from phydrv->features after calling config_init to make sure these parameters aren't overwritten.

Hi Maxime

I have a patch which makes some core changes to support PHY doing runtime feature detection. I would prefer to use them, than this.

Either I or Heiner will post them soon.

Andrew

#4 lore raw

Re: [PATCH net-next v2 01/10] net: phy: Update PHY linkmodes after config_init

Hello Andrew,

Quoted 6 lines · On Thu, 7 Feb 2019 14:48:07 +0100 Andrew Lunn <[email protected]> wrote:
Quoted 4 lines · Maxime Chevallier

We want to be able to update a PHY's supported list in the config_init callback, so move the Pause parameters settings from phydrv->features after calling config_init to make sure these parameters aren't overwritten.

Quoted 6 lines

Hi Maxime

I have a patch which makes some core changes to support PHY doing runtime feature detection. I would prefer to use them, than this.

Either I or Heiner will post them soon.

Sure, no problem, thanks for doing this. I'll be happy to review and test that on my side.

As I said, I lack the big picture view on that part so my approach was pretty naive, I'm glad you can take care of this :)

Thanks,

Maxime

maintainer
#5 lore raw

Re: [PATCH net-next v2 01/10] net: phy: Update PHY linkmodes after config_init

Quoted 13 lines · Maxime Chevallier

Hello Andrew,

Quoted 8 lines · On Thu, 7 Feb 2019 14:48:07 +0100 Andrew Lunn <[email protected]> wrote:
Quoted 4 lines · Maxime Chevallier

We want to be able to update a PHY's supported list in the config_init callback, so move the Pause parameters settings from phydrv->features after calling config_init to make sure these parameters aren't overwritten.

Hi Maxime

Quoted 8 lines
Quoted 4 lines

I have a patch which makes some core changes to support PHY doing runtime feature detection. I would prefer to use them, than this.

Either I or Heiner will post them soon.

Sure, no problem, thanks for doing this. I'll be happy to review and test that on my side.

The second version of the patch set comes in time because I was just about to submit parts of it on Maxime's behalf. We may have some dependencies between adding the get_features callback in phy_driver and this patch set. Maybe the patch set needs to be splitted. I think tomorrow or the day after I can have a closer look at this.

Quoted 3 lines

As I said, I lack the big picture view on that part so my approach was pretty naive, I'm glad you can take care of this :)

Don't be so self-effacing, it's good work :)

Quoted 4 lines

Thanks,

Maxime

Heiner

#6 lore raw

PATCH v2 2/10 net: phy: Mask-out non-compatible modes when setting the max-speed

In reply to Maxime Chevallier

When setting a PHY's max speed using either the max-speed DT property or ethtool, we should mask-out all non-compatible modes according to the settings table, instead of just the 10/100BASET modes.

---
 drivers/net/phy/phy-core.c   | 45 ++++++++++++++++++++++++++++++
 drivers/net/phy/phy_device.c | 53 ------------------------------------
 include/linux/phy.h          |  1 +
 3 files changed, 46 insertions(+), 53 deletions(-)
drivers/net/phy/phy-core.c +45 -0
index 7d6aad287f84..8e54fe396553 100644
@@ -4,6 +4,7 @@
  */
 #include <linux/export.h>
 #include <linux/phy.h>
+#include <linux/of.h>
 
 const char *phy_speed_to_str(int speed)
 {
@@ -338,6 +339,50 @@ size_t phy_speeds(unsigned int *speeds, size_t size,
 	return count;
 }
 
+static int __set_phy_supported(struct phy_device *phydev, u32 max_speed)
+{
+	const struct phy_setting *p;
+	int i;
+
+	for (i = 0, p = settings; i < ARRAY_SIZE(settings); i++, p++) {
+		if (p->speed > max_speed)
+			linkmode_clear_bit(p->bit, phydev->supported);
+		else
+			break;
+	}
+
+	return 0;
+}
+
+int phy_set_max_speed(struct phy_device *phydev, u32 max_speed)
+{
+	int err;
+
+	err = __set_phy_supported(phydev, max_speed);
+	if (err)
+		return err;
+
+	linkmode_copy(phydev->advertising, phydev->supported);
+
+	return 0;
+}
+EXPORT_SYMBOL(phy_set_max_speed);
+
+void of_set_phy_supported(struct phy_device *phydev)
+{
+	struct device_node *node = phydev->mdio.dev.of_node;
+	u32 max_speed;
+
+	if (!IS_ENABLED(CONFIG_OF_MDIO))
+		return;
+
+	if (!node)
+		return;
+
+	if (!of_property_read_u32(node, "max-speed", &max_speed))
+		__set_phy_supported(phydev, max_speed);
+}
+
 /**
  * phy_resolve_aneg_linkmode - resolve the advertisements into phy settings
  * @phydev: The phy_device struct
drivers/net/phy/phy_device.c +0 -53

65 lines not on this page · raw message

include/linux/phy.h +1 -0
index 237dd035858a..cfdd3de38410 100644
@@ -667,6 +667,7 @@ phy_lookup_setting(int speed, int duplex, const unsigned long *mask,
 		   bool exact);
 size_t phy_speeds(unsigned int *speeds, size_t size,
 		  unsigned long *mask);
+void of_set_phy_supported(struct phy_device *phydev);
 
 static inline bool __phy_is_started(struct phy_device *phydev)
 {
Signature · 1 line
2.19.2
maintainer Reviewed-by on v2 2/10
#7 lore raw

Re: [PATCH net-next v2 02/10] net: phy: Mask-out non-compatible modes when setting the max-speed

Quoted 6 lines · Maxime Chevallier

When setting a PHY's max speed using either the max-speed DT property or ethtool, we should mask-out all non-compatible modes according to the settings table, instead of just the 10/100BASET modes.

Andrew

#8 lore raw

PATCH v2 3/10 net: phy: Move of_set_phy_eee_broken to phy-core.c

In reply to Maxime Chevallier

Since of_set_phy_supported was moved to phy-core.c, we can also move of_set_phy_eee_broken to the same location, so that we have all OF functions in the same place.

This patch doesn't intend to introduce any change in behaviour.

---
 drivers/net/phy/phy-core.c   | 27 +++++++++++++++++++++++++++
 drivers/net/phy/phy_device.c | 28 ----------------------------
 include/linux/phy.h          |  1 +
 3 files changed, 28 insertions(+), 28 deletions(-)
drivers/net/phy/phy-core.c +27 -0
index 8e54fe396553..9429b473c791 100644
@@ -383,6 +383,33 @@ void of_set_phy_supported(struct phy_device *phydev)
 		__set_phy_supported(phydev, max_speed);
 }
 
+void of_set_phy_eee_broken(struct phy_device *phydev)
+{
+	struct device_node *node = phydev->mdio.dev.of_node;
+	u32 broken = 0;
+
+	if (!IS_ENABLED(CONFIG_OF_MDIO))
+		return;
+
+	if (!node)
+		return;
+
+	if (of_property_read_bool(node, "eee-broken-100tx"))
+		broken |= MDIO_EEE_100TX;
+	if (of_property_read_bool(node, "eee-broken-1000t"))
+		broken |= MDIO_EEE_1000T;
+	if (of_property_read_bool(node, "eee-broken-10gt"))
+		broken |= MDIO_EEE_10GT;
+	if (of_property_read_bool(node, "eee-broken-1000kx"))
+		broken |= MDIO_EEE_1000KX;
+	if (of_property_read_bool(node, "eee-broken-10gkx4"))
+		broken |= MDIO_EEE_10GKX4;
+	if (of_property_read_bool(node, "eee-broken-10gkr"))
+		broken |= MDIO_EEE_10GKR;
+
+	phydev->eee_broken_modes = broken;
+}
+
 /**
  * phy_resolve_aneg_linkmode - resolve the advertisements into phy settings
  * @phydev: The phy_device struct
drivers/net/phy/phy_device.c +0 -28
index a4ab16992806..a4cbc5a6f09d 100644
@@ -30,7 +30,6 @@
 #include <linux/mdio.h>
 #include <linux/io.h>
 #include <linux/uaccess.h>
-#include <linux/of.h>
 
 MODULE_DESCRIPTION("PHY library");
 MODULE_AUTHOR("Andy Fleming");
@@ -2153,33 +2152,6 @@ bool phy_validate_pause(struct phy_device *phydev,
 }
 EXPORT_SYMBOL(phy_validate_pause);
 
-static void of_set_phy_eee_broken(struct phy_device *phydev)
-{
-	struct device_node *node = phydev->mdio.dev.of_node;
-	u32 broken = 0;
-
-	if (!IS_ENABLED(CONFIG_OF_MDIO))
-		return;
-
-	if (!node)
-		return;
-
-	if (of_property_read_bool(node, "eee-broken-100tx"))
-		broken |= MDIO_EEE_100TX;
-	if (of_property_read_bool(node, "eee-broken-1000t"))
-		broken |= MDIO_EEE_1000T;
-	if (of_property_read_bool(node, "eee-broken-10gt"))
-		broken |= MDIO_EEE_10GT;
-	if (of_property_read_bool(node, "eee-broken-1000kx"))
-		broken |= MDIO_EEE_1000KX;
-	if (of_property_read_bool(node, "eee-broken-10gkx4"))
-		broken |= MDIO_EEE_10GKX4;
-	if (of_property_read_bool(node, "eee-broken-10gkr"))
-		broken |= MDIO_EEE_10GKR;
-
-	phydev->eee_broken_modes = broken;
-}
-
 static bool phy_drv_supports_irq(struct phy_driver *phydrv)
 {
 	return phydrv->config_intr && phydrv->ack_interrupt;
include/linux/phy.h +1 -0
index cfdd3de38410..017118061ecf 100644
@@ -668,6 +668,7 @@ phy_lookup_setting(int speed, int duplex, const unsigned long *mask,
 size_t phy_speeds(unsigned int *speeds, size_t size,
 		  unsigned long *mask);
 void of_set_phy_supported(struct phy_device *phydev);
+void of_set_phy_eee_broken(struct phy_device *phydev);
 
 static inline bool __phy_is_started(struct phy_device *phydev)
 {
Signature · 1 line
2.19.2
maintainer Reviewed-by on v2 3/10
#9 lore raw

Re: [PATCH net-next v2 03/10] net: phy: Move of_set_phy_eee_broken to phy-core.c

Quoted 7 lines · Maxime Chevallier

Since of_set_phy_supported was moved to phy-core.c, we can also move of_set_phy_eee_broken to the same location, so that we have all OF functions in the same place.

This patch doesn't intend to introduce any change in behaviour.

Andrew

#10 lore raw

PATCH v2 4/10 net: phy: Automatically fill the generic TP, FIBRE and Backplane modes

In reply to Maxime Chevallier

PHY advertised and supported linkmodes contain both specific modes such as 1000BASET Half/Full and generic ones such as TP that represent a class of modes.

Since some modes such as Fibre, TP or Backplane match a wide range of specific modes, we can automatically set these bits if one of the specific modes it corresponds to is present in the list.

The 'TP' bit is set whenever there's a BaseT linkmode in phydev->supported.

The 'FIBRE' bit is set for BaseL, BaseS and BaseE linkmodes.

Finally, the 'Backplane' is set whenever a BaseK mode is supported.

---
 drivers/net/phy/phy_device.c | 67 +++++++++++++++++++++++++++++++++++-
 include/linux/linkmode.h     |  6 ++++
 2 files changed, 72 insertions(+), 1 deletion(-)
drivers/net/phy/phy_device.c +66 -1
index a4cbc5a6f09d..942cfa7548c4 100644
@@ -1066,15 +1066,80 @@ static int phy_poll_reset(struct phy_device *phydev)
  * straightforward to maintain, since PHYs and MACs are subject to quirks and
  * erratas. This function re-builds the list of the supported pause parameters
  * by taking into account the parameters expressed in the driver's features
- * list.
+ * list. It also sets the generic bits indicating Twisted Pair, Fibre and
+ * Backaplane link modes support based on the detailed list that can be built
+ * from the PHY's ability list.
  */
 static void phy_update_linkmodes(struct phy_device *phydev)
 {
+	__ETHTOOL_DECLARE_LINK_MODE_MASK(phy_baset_features) = { 0, };
+	__ETHTOOL_DECLARE_LINK_MODE_MASK(phy_fibre_features) = { 0, };
+	__ETHTOOL_DECLARE_LINK_MODE_MASK(phy_backplane_features) = { 0, };
 	struct device_driver *drv = phydev->mdio.dev.driver;
 	struct phy_driver *phydrv = to_phy_driver(drv);
 
+	const int phy_baset_features_array[] = {
+		ETHTOOL_LINK_MODE_10baseT_Half_BIT,
+		ETHTOOL_LINK_MODE_10baseT_Full_BIT,
+		ETHTOOL_LINK_MODE_100baseT_Half_BIT,
+		ETHTOOL_LINK_MODE_100baseT_Full_BIT,
+		ETHTOOL_LINK_MODE_1000baseT_Half_BIT,
+		ETHTOOL_LINK_MODE_1000baseT_Full_BIT,
+		ETHTOOL_LINK_MODE_10000baseT_Full_BIT,
+	};
+
+	const int phy_fibre_features_array[] = {
+		ETHTOOL_LINK_MODE_40000baseSR4_Full_BIT,
+		ETHTOOL_LINK_MODE_40000baseLR4_Full_BIT,
+		ETHTOOL_LINK_MODE_56000baseSR4_Full_BIT,
+		ETHTOOL_LINK_MODE_56000baseLR4_Full_BIT,
+		ETHTOOL_LINK_MODE_25000baseSR_Full_BIT,
+		ETHTOOL_LINK_MODE_100000baseSR4_Full_BIT,
+		ETHTOOL_LINK_MODE_100000baseLR4_ER4_Full_BIT,
+		ETHTOOL_LINK_MODE_50000baseSR2_Full_BIT,
+		ETHTOOL_LINK_MODE_10000baseSR_Full_BIT,
+		ETHTOOL_LINK_MODE_10000baseLR_Full_BIT,
+		ETHTOOL_LINK_MODE_10000baseLRM_Full_BIT,
+		ETHTOOL_LINK_MODE_10000baseER_Full_BIT,
+	};
+
+	const int phy_backplane_features_array[] = {
+		ETHTOOL_LINK_MODE_1000baseKX_Full_BIT,
+		ETHTOOL_LINK_MODE_10000baseKX4_Full_BIT,
+		ETHTOOL_LINK_MODE_10000baseKR_Full_BIT,
+		ETHTOOL_LINK_MODE_20000baseKR2_Full_BIT,
+		ETHTOOL_LINK_MODE_40000baseKR4_Full_BIT,
+		ETHTOOL_LINK_MODE_56000baseKR4_Full_BIT,
+		ETHTOOL_LINK_MODE_25000baseKR_Full_BIT,
+		ETHTOOL_LINK_MODE_50000baseKR2_Full_BIT,
+		ETHTOOL_LINK_MODE_100000baseKR4_Full_BIT,
+	};
+
+	linkmode_set_bit_array(phy_baset_features_array,
+			       ARRAY_SIZE(phy_baset_features_array),
+			       phy_baset_features);
+
+	linkmode_set_bit_array(phy_fibre_features_array,
+			       ARRAY_SIZE(phy_fibre_features_array),
+			       phy_fibre_features);
+
+	linkmode_set_bit_array(phy_backplane_features_array,
+			       ARRAY_SIZE(phy_backplane_features_array),
+			       phy_backplane_features);
+
 	mutex_lock(&phydev->lock);
 
+	if (linkmode_intersects(phydev->supported, phy_baset_features))
+		linkmode_set_bit(ETHTOOL_LINK_MODE_TP_BIT, phydev->supported);
+
+	if (linkmode_intersects(phydev->supported, phy_fibre_features))
+		linkmode_set_bit(ETHTOOL_LINK_MODE_FIBRE_BIT,
+				 phydev->supported);
+
+	if (linkmode_intersects(phydev->supported, phy_backplane_features))
+		linkmode_set_bit(ETHTOOL_LINK_MODE_Backplane_BIT,
+				 phydev->supported);
+
 	/* The Pause Frame bits indicate that the PHY can support passing
 	 * pause frames. During autonegotiation, the PHYs will determine if
 	 * they should allow pause frames to pass.  The MAC driver should then
include/linux/linkmode.h +6 -0
index a99c58866860..49ab6415e1e0 100644
@@ -82,4 +82,10 @@ static inline int linkmode_equal(const unsigned long *src1,
 	return bitmap_equal(src1, src2, __ETHTOOL_LINK_MODE_MASK_NBITS);
 }
 
+static inline bool linkmode_intersects(const unsigned long *src1,
+				       const unsigned long *src2)
+{
+	return bitmap_intersects(src1, src2, __ETHTOOL_LINK_MODE_MASK_NBITS);
+}
+
 #endif /* __LINKMODE_H */
Signature · 1 line
2.19.2
maintainer
#11 lore raw

Re: [PATCH net-next v2 04/10] net: phy: Automatically fill the generic TP, FIBRE and Backplane modes

Quoted 8 lines · Maxime Chevallier

PHY advertised and supported linkmodes contain both specific modes such as 1000BASET Half/Full and generic ones such as TP that represent a class of modes.

Since some modes such as Fibre, TP or Backplane match a wide range of specific modes, we can automatically set these bits if one of the specific modes it corresponds to is present in the list.

Quoted 6 lines

The 'TP' bit is set whenever there's a BaseT linkmode in phydev->supported.

The 'FIBRE' bit is set for BaseL, BaseS and BaseE linkmodes.

Finally, the 'Backplane' is set whenever a BaseK mode is supported.

Hi Maxime

Interesting idea.

But what exactly are we supposed to be representing here? That PHY can do these modes, or that the port exists on the device? The marvell10g can do fibre, but do all boards have an SFP/SFF, or do some only have an RJ-45 for TP? Are there boards without TP and just SFP/SFF?

Is there documentation in ethtool which gives a clue as to what is expected?

Thanks
Andrew

#12 lore raw

Re: [PATCH net-next v2 04/10] net: phy: Automatically fill the generic TP, FIBRE and Backplane modes

Hello Andrew,

Quoted 18 lines · On Thu, 7 Feb 2019 15:09:39 +0100 Andrew Lunn <[email protected]> wrote:
Quoted 8 lines · Maxime Chevallier

PHY advertised and supported linkmodes contain both specific modes such as 1000BASET Half/Full and generic ones such as TP that represent a class of modes.

Since some modes such as Fibre, TP or Backplane match a wide range of specific modes, we can automatically set these bits if one of the specific modes it corresponds to is present in the list.

Quoted 6 lines

The 'TP' bit is set whenever there's a BaseT linkmode in phydev->supported.

The 'FIBRE' bit is set for BaseL, BaseS and BaseE linkmodes.

Finally, the 'Backplane' is set whenever a BaseK mode is supported.

Hi Maxime

Quoted 7 lines

Interesting idea.

But what exactly are we supposed to be representing here? That PHY can do these modes, or that the port exists on the device? The marvell10g can do fibre, but do all boards have an SFP/SFF, or do some only have an RJ-45 for TP? Are there boards without TP and just SFP/SFF?

My understanding is that this would represent what the PHY is capable of, at least when set in the supported field, but I agree that this doesn't reflect what's on the device.

I extrapolated this logic from the ability detection in marvell10g, that would set these bits according to the abilities reported by the PHY. This was done based on the PHY register values, instead of the linkmodes in the 'supported' field.

Quoted 2 lines

Is there documentation in ethtool which gives a clue as to what is expected?

Not really, at least to my knowledge. I think this would be used by "ethtool -s ethX port tp|fibre|aui|etc". Maybe this could be useful if we decide to implement port selection with ethtool on the MCBin, but I'm getting a bit ahead of myself.

Thanks,

Maxime

#13 lore raw

PATCH v2 5/10 net: phy: Extract genphy_c45_pma_read_abilities from marvell10g

In reply to Maxime Chevallier

Marvell 10G PHY driver has a generic way of initializing the supported link modes by reading the PHY's C45 PMA abilities. This can be made generic, since these registers are part of the 802.3 specifications.

This commit extracts the config_init link_mode initialization code from marvell10g and uses it to introduce the genphy_c45_pma_read_abilities function.

Only PMA modes are read, it's still up to the caller to set the Pause parameters.

---
V1 -> V2: The Pause parameters setting stays in marvell10g. Make use of
linkmode_mod_bit. Renamed the function from genphy_c45_read_abilities to genphy_c45_pma_read_abilities, as advised by Andrew.

 drivers/net/phy/marvell10g.c | 78 ++++--------------------------------
 drivers/net/phy/phy-c45.c    | 74 ++++++++++++++++++++++++++++++++++
 include/linux/phy.h          |  1 +
 3 files changed, 83 insertions(+), 70 deletions(-)
drivers/net/phy/marvell10g.c +8 -70
index 296a537cdfcb..07df87b81369 100644
@@ -234,8 +234,7 @@ static int mv3310_resume(struct phy_device *phydev)
 
 static int mv3310_config_init(struct phy_device *phydev)
 {
-	__ETHTOOL_DECLARE_LINK_MODE_MASK(supported) = { 0, };
-	int val;
+	int ret, val;
 
 	/* Check that the PHY interface type is compatible */
 	if (phydev->interface != PHY_INTERFACE_MODE_SGMII &&
@@ -244,8 +243,8 @@ static int mv3310_config_init(struct phy_device *phydev)
 	    phydev->interface != PHY_INTERFACE_MODE_10GKR)
 		return -ENODEV;
 
-	__set_bit(ETHTOOL_LINK_MODE_Pause_BIT, supported);
-	__set_bit(ETHTOOL_LINK_MODE_Asym_Pause_BIT, supported);
+	__set_bit(ETHTOOL_LINK_MODE_Pause_BIT, phydev->supported);
+	__set_bit(ETHTOOL_LINK_MODE_Asym_Pause_BIT, phydev->supported);
 
 	if (phydev->c45_ids.devices_in_package & MDIO_DEVS_AN) {
 		val = phy_read_mmd(phydev, MDIO_MMD_AN, MDIO_STAT1);
@@ -253,74 +252,13 @@ static int mv3310_config_init(struct phy_device *phydev)
 			return val;
 
 		if (val & MDIO_AN_STAT1_ABLE)
-			__set_bit(ETHTOOL_LINK_MODE_Autoneg_BIT, supported);
+			__set_bit(ETHTOOL_LINK_MODE_Autoneg_BIT,
+				  phydev->supported);
 	}
 
-	val = phy_read_mmd(phydev, MDIO_MMD_PMAPMD, MDIO_STAT2);
-	if (val < 0)
-		return val;
-
-	/* Ethtool does not support the WAN mode bits */
-	if (val & (MDIO_PMA_STAT2_10GBSR | MDIO_PMA_STAT2_10GBLR |
-		   MDIO_PMA_STAT2_10GBER | MDIO_PMA_STAT2_10GBLX4 |
-		   MDIO_PMA_STAT2_10GBSW | MDIO_PMA_STAT2_10GBLW |
-		   MDIO_PMA_STAT2_10GBEW))
-		__set_bit(ETHTOOL_LINK_MODE_FIBRE_BIT, supported);
-	if (val & MDIO_PMA_STAT2_10GBSR)
-		__set_bit(ETHTOOL_LINK_MODE_10000baseSR_Full_BIT, supported);
-	if (val & MDIO_PMA_STAT2_10GBLR)
-		__set_bit(ETHTOOL_LINK_MODE_10000baseLR_Full_BIT, supported);
-	if (val & MDIO_PMA_STAT2_10GBER)
-		__set_bit(ETHTOOL_LINK_MODE_10000baseER_Full_BIT, supported);
-
-	if (val & MDIO_PMA_STAT2_EXTABLE) {
-		val = phy_read_mmd(phydev, MDIO_MMD_PMAPMD, MDIO_PMA_EXTABLE);
-		if (val < 0)
-			return val;
-
-		if (val & (MDIO_PMA_EXTABLE_10GBT | MDIO_PMA_EXTABLE_1000BT |
-			   MDIO_PMA_EXTABLE_100BTX | MDIO_PMA_EXTABLE_10BT))
-			__set_bit(ETHTOOL_LINK_MODE_TP_BIT, supported);
-		if (val & MDIO_PMA_EXTABLE_10GBLRM)
-			__set_bit(ETHTOOL_LINK_MODE_FIBRE_BIT, supported);
-		if (val & (MDIO_PMA_EXTABLE_10GBKX4 | MDIO_PMA_EXTABLE_10GBKR |
-			   MDIO_PMA_EXTABLE_1000BKX))
-			__set_bit(ETHTOOL_LINK_MODE_Backplane_BIT, supported);
-		if (val & MDIO_PMA_EXTABLE_10GBLRM)
-			__set_bit(ETHTOOL_LINK_MODE_10000baseLRM_Full_BIT,
-				  supported);
-		if (val & MDIO_PMA_EXTABLE_10GBT)
-			__set_bit(ETHTOOL_LINK_MODE_10000baseT_Full_BIT,
-				  supported);
-		if (val & MDIO_PMA_EXTABLE_10GBKX4)
-			__set_bit(ETHTOOL_LINK_MODE_10000baseKX4_Full_BIT,
-				  supported);
-		if (val & MDIO_PMA_EXTABLE_10GBKR)
-			__set_bit(ETHTOOL_LINK_MODE_10000baseKR_Full_BIT,
-				  supported);
-		if (val & MDIO_PMA_EXTABLE_1000BT)
-			__set_bit(ETHTOOL_LINK_MODE_1000baseT_Full_BIT,
-				  supported);
-		if (val & MDIO_PMA_EXTABLE_1000BKX)
-			__set_bit(ETHTOOL_LINK_MODE_1000baseKX_Full_BIT,
-				  supported);
-		if (val & MDIO_PMA_EXTABLE_100BTX) {
-			__set_bit(ETHTOOL_LINK_MODE_100baseT_Full_BIT,
-				  supported);
-			__set_bit(ETHTOOL_LINK_MODE_100baseT_Half_BIT,
-				  supported);
-		}
-		if (val & MDIO_PMA_EXTABLE_10BT) {
-			__set_bit(ETHTOOL_LINK_MODE_10baseT_Full_BIT,
-				  supported);
-			__set_bit(ETHTOOL_LINK_MODE_10baseT_Half_BIT,
-				  supported);
-		}
-	}
-
-	linkmode_copy(phydev->supported, supported);
-	linkmode_and(phydev->advertising, phydev->advertising,
-		     phydev->supported);
+	ret = genphy_c45_pma_read_abilities(phydev);
+	if (ret)
+		return ret;
 
 	return 0;
 }
drivers/net/phy/phy-c45.c +74 -0

80 lines not on this page · raw message

include/linux/phy.h +1 -0
index 017118061ecf..863d0e00db89 100644
@@ -1102,6 +1102,7 @@ int genphy_c45_read_pma(struct phy_device *phydev);
 int genphy_c45_pma_setup_forced(struct phy_device *phydev);
 int genphy_c45_an_disable_aneg(struct phy_device *phydev);
 int genphy_c45_read_mdix(struct phy_device *phydev);
+int genphy_c45_pma_read_abilities(struct phy_device *phydev);
 
 /* The gen10g_* functions are the old Clause 45 stub */
 int gen10g_config_aneg(struct phy_device *phydev);
Signature · 1 line
2.19.2
#14 lore raw

PATCH v2 6/10 net: phy: Add generic support for 2.5GBaseT and 5GBaseT

In reply to Maxime Chevallier

The 802.3bz specification, based on previous by the NBASET alliance, defines the 2.5GBaseT and 5GBaseT link modes for ethernet traffic on cat5e, cat6 and cat7 cables.

These mode integrate with the already defined C45 MDIO PMA/PMD registers set that added 10G support, by defining some previously reserved bits, and adding a new register (2.5G/5G Extended abilities).

This commit adds the required definitions in include/uapi/linux/mdio.h to support these modes, and detect when a link-partner advertises them.

It also adds support for these mode in the generic C45 PHY infrastructure.

---
V1 -> V2: Merged in code for reading 2.5G/5G abilities.

 drivers/net/phy/phy-c45.c    | 37 ++++++++++++++++++++++++++++++++++++
 drivers/net/phy/phy_device.c |  2 ++
 include/uapi/linux/mdio.h    | 16 ++++++++++++++++
 3 files changed, 55 insertions(+)
drivers/net/phy/phy-c45.c +37 -0
index 32e72b170ea9..92897cb26557 100644
@@ -47,6 +47,16 @@ int genphy_c45_pma_setup_forced(struct phy_device *phydev)
 		/* Assume 1000base-T */
 		ctrl2 |= MDIO_PMA_CTRL2_1000BT;
 		break;
+	case SPEED_2500:
+		ctrl1 |= MDIO_CTRL1_SPEED2_5G;
+		/* Assume 2.5Gbase-T */
+		ctrl2 |= MDIO_PMA_CTRL2_2_5GBT;
+		break;
+	case SPEED_5000:
+		ctrl1 |= MDIO_CTRL1_SPEED5G;
+		/* Assume 5Gbase-T */
+		ctrl2 |= MDIO_PMA_CTRL2_5GBT;
+		break;
 	case SPEED_10000:
 		ctrl1 |= MDIO_CTRL1_SPEED10G;
 		/* Assume 10Gbase-T */
@@ -179,6 +189,12 @@ int genphy_c45_read_lpa(struct phy_device *phydev)
 	if (val < 0)
 		return val;
 
+	if (val & MDIO_AN_10GBT_STAT_LP2_5G)
+		linkmode_set_bit(ETHTOOL_LINK_MODE_2500baseT_Full_BIT,
+				 phydev->lp_advertising);
+	if (val & MDIO_AN_10GBT_STAT_LP5G)
+		linkmode_set_bit(ETHTOOL_LINK_MODE_5000baseT_Full_BIT,
+				 phydev->lp_advertising);
 	if (val & MDIO_AN_10GBT_STAT_LP10G)
 		linkmode_set_bit(ETHTOOL_LINK_MODE_10000baseT_Full_BIT,
 				 phydev->lp_advertising);
@@ -209,6 +225,12 @@ int genphy_c45_read_pma(struct phy_device *phydev)
 	case MDIO_PMA_CTRL1_SPEED1000:
 		phydev->speed = SPEED_1000;
 		break;
+	case MDIO_CTRL1_SPEED2_5G:
+		phydev->speed = SPEED_2500;
+		break;
+	case MDIO_CTRL1_SPEED5G:
+		phydev->speed = SPEED_5000;
+		break;
 	case MDIO_CTRL1_SPEED10G:
 		phydev->speed = SPEED_10000;
 		break;
@@ -324,6 +346,21 @@ int genphy_c45_pma_read_abilities(struct phy_device *phydev)
 		linkmode_mod_bit(ETHTOOL_LINK_MODE_10baseT_Half_BIT,
 				 phydev->supported,
 				 val & MDIO_PMA_EXTABLE_10BT);
+
+		if (val & MDIO_PMA_EXTABLE_NBT) {
+			val = phy_read_mmd(phydev, MDIO_MMD_PMAPMD,
+					   MDIO_PMA_NG_EXTABLE);
+			if (val < 0)
+				return val;
+
+			linkmode_mod_bit(ETHTOOL_LINK_MODE_2500baseT_Full_BIT,
+					 phydev->supported,
+					 val & MDIO_PMA_NG_EXTABLE_2_5GBT);
+
+			linkmode_mod_bit(ETHTOOL_LINK_MODE_5000baseT_Full_BIT,
+					 phydev->supported,
+					 val & MDIO_PMA_NG_EXTABLE_5GBT);
+		}
 	}
 
 	return 0;
drivers/net/phy/phy_device.c +2 -0
index 942cfa7548c4..13ac4bdb9b69 100644
@@ -1085,6 +1085,8 @@ static void phy_update_linkmodes(struct phy_device *phydev)
 		ETHTOOL_LINK_MODE_100baseT_Full_BIT,
 		ETHTOOL_LINK_MODE_1000baseT_Half_BIT,
 		ETHTOOL_LINK_MODE_1000baseT_Full_BIT,
+		ETHTOOL_LINK_MODE_2500baseT_Full_BIT,
+		ETHTOOL_LINK_MODE_5000baseT_Full_BIT,
 		ETHTOOL_LINK_MODE_10000baseT_Full_BIT,
 	};
 
include/uapi/linux/mdio.h +16 -0
index d435b00d64ad..546509898867 100644
@@ -45,6 +45,7 @@
 #define MDIO_AN_ADVERTISE	16	/* AN advertising (base page) */
 #define MDIO_AN_LPA		19	/* AN LP abilities (base page) */
 #define MDIO_PCS_EEE_ABLE	20	/* EEE Capability register */
+#define MDIO_PMA_NG_EXTABLE	21	/* 2.5G/5G PMA/PMD extended ability */
 #define MDIO_PCS_EEE_WK_ERR	22	/* EEE wake error counter */
 #define MDIO_PHYXS_LNSTAT	24	/* PHY XGXS lane state */
 #define MDIO_AN_EEE_ADV		60	/* EEE advertisement */
@@ -92,6 +93,10 @@
 #define MDIO_CTRL1_SPEED10G		(MDIO_CTRL1_SPEEDSELEXT | 0x00)
 /* 10PASS-TS/2BASE-TL */
 #define MDIO_CTRL1_SPEED10P2B		(MDIO_CTRL1_SPEEDSELEXT | 0x04)
+/* 2.5 Gb/s */
+#define MDIO_CTRL1_SPEED2_5G		(MDIO_CTRL1_SPEEDSELEXT | 0x18)
+/* 5 Gb/s */
+#define MDIO_CTRL1_SPEED5G		(MDIO_CTRL1_SPEEDSELEXT | 0x1c)
 
 /* Status register 1. */
 #define MDIO_STAT1_LPOWERABLE		0x0002	/* Low-power ability */
@@ -142,6 +147,8 @@
 #define MDIO_PMA_CTRL2_1000BKX		0x000d	/* 1000BASE-KX type */
 #define MDIO_PMA_CTRL2_100BTX		0x000e	/* 100BASE-TX type */
 #define MDIO_PMA_CTRL2_10BT		0x000f	/* 10BASE-T type */
+#define MDIO_PMA_CTRL2_2_5GBT		0x0030  /* 2.5GBaseT type */
+#define MDIO_PMA_CTRL2_5GBT		0x0031  /* 5GBaseT type */
 #define MDIO_PCS_CTRL2_TYPE		0x0003	/* PCS type selection */
 #define MDIO_PCS_CTRL2_10GBR		0x0000	/* 10GBASE-R type */
 #define MDIO_PCS_CTRL2_10GBX		0x0001	/* 10GBASE-X type */
@@ -195,6 +202,7 @@
 #define MDIO_PMA_EXTABLE_1000BKX	0x0040	/* 1000BASE-KX ability */
 #define MDIO_PMA_EXTABLE_100BTX		0x0080	/* 100BASE-TX ability */
 #define MDIO_PMA_EXTABLE_10BT		0x0100	/* 10BASE-T ability */
+#define MDIO_PMA_EXTABLE_NBT		0x4000  /* 2.5/5GBASE-T ability */
 
 /* PHY XGXS lane state register. */
 #define MDIO_PHYXS_LNSTAT_SYNC0		0x0001
@@ -231,9 +239,13 @@
 #define MDIO_PCS_10GBRT_STAT2_BER	0x3f00
 
 /* AN 10GBASE-T control register. */
+#define MDIO_AN_10GBT_CTRL_ADV2_5G	0x0080	/* Advertise 2.5GBASE-T */
+#define MDIO_AN_10GBT_CTRL_ADV5G	0x0100	/* Advertise 5GBASE-T */
 #define MDIO_AN_10GBT_CTRL_ADV10G	0x1000	/* Advertise 10GBASE-T */
 
 /* AN 10GBASE-T status register. */
+#define MDIO_AN_10GBT_STAT_LP2_5G	0x0020  /* LP is 2.5GBT capable */
+#define MDIO_AN_10GBT_STAT_LP5G		0x0040  /* LP is 5GBT capable */
 #define MDIO_AN_10GBT_STAT_LPTRR	0x0200	/* LP training reset req. */
 #define MDIO_AN_10GBT_STAT_LPLTABLE	0x0400	/* LP loop timing ability */
 #define MDIO_AN_10GBT_STAT_LP10G	0x0800	/* LP is 10GBT capable */
@@ -262,6 +274,10 @@
 #define MDIO_EEE_10GKX4		0x0020	/* 10G KX4 EEE cap */
 #define MDIO_EEE_10GKR		0x0040	/* 10G KR EEE cap */
 
+/* 2.5G/5G Extended abilities register. */
+#define MDIO_PMA_NG_EXTABLE_2_5GBT	0x0001	/* 2.5GBASET ability */
+#define MDIO_PMA_NG_EXTABLE_5GBT	0x0002	/* 5GBASET ability */
+
 /* LASI RX_ALARM control/status registers. */
 #define MDIO_PMA_LASI_RX_PHYXSLFLT	0x0001	/* PHY XS RX local fault */
 #define MDIO_PMA_LASI_RX_PCSLFLT	0x0008	/* PCS RX local fault */
Signature · 1 line
2.19.2
#15 lore raw

PATCH v2 7/10 net: phy: marvell10g: Add support for 2.5GBASET

In reply to Maxime Chevallier

The Marvell Alaska family of PHYs supports 2.5GBaseT and 5GBaseT modes, as defined in the 802.3bz specification.

When the link partner requests a 2.5GBASET link, the PHY will reconfigure it's MII interface to 2500BASEX.

At 5G, the PHY will reconfigure it's interface to 5GBASE-R, but this mode isn't supported by any MAC for now.

This was tested with :
- The 88X3310, which is on the MacchiatoBin
- The 88E2010, an Alaska PHY that has no fiber interfaces, and is limited to 5G maximum speed.

---
V1 -> V2: Use a #define for the 88X3310 PHY ID, since it's reused in
various places in the code. Rebased on Heiner Kallweit's patch introducing the phy_modify_mmd accessor.

 drivers/net/phy/marvell10g.c | 30 ++++++++++++++++++++++--------
 include/linux/marvell_phy.h  |  1 +
 2 files changed, 23 insertions(+), 8 deletions(-)
drivers/net/phy/marvell10g.c +22 -8
index 07df87b81369..581b4b6e31e9 100644
@@ -238,6 +238,7 @@ static int mv3310_config_init(struct phy_device *phydev)
 
 	/* Check that the PHY interface type is compatible */
 	if (phydev->interface != PHY_INTERFACE_MODE_SGMII &&
+	    phydev->interface != PHY_INTERFACE_MODE_2500BASEX &&
 	    phydev->interface != PHY_INTERFACE_MODE_XAUI &&
 	    phydev->interface != PHY_INTERFACE_MODE_RXAUI &&
 	    phydev->interface != PHY_INTERFACE_MODE_10GKR)
@@ -307,8 +308,18 @@ static int mv3310_config_aneg(struct phy_device *phydev)
 	else
 		reg = 0;
 
+	if (linkmode_test_bit(ETHTOOL_LINK_MODE_2500baseT_Full_BIT,
+			      phydev->advertising))
+		reg |= MDIO_AN_10GBT_CTRL_ADV2_5G;
+	if (linkmode_test_bit(ETHTOOL_LINK_MODE_5000baseT_Full_BIT,
+			      phydev->advertising))
+		reg |= MDIO_AN_10GBT_CTRL_ADV5G;
+
 	ret = phy_modify_mmd(phydev, MDIO_MMD_AN, MDIO_AN_10GBT_CTRL,
-			     MDIO_AN_10GBT_CTRL_ADV10G, reg);
+			     MDIO_AN_10GBT_CTRL_ADV10G |
+			     MDIO_AN_10GBT_CTRL_ADV5G |
+			     MDIO_AN_10GBT_CTRL_ADV2_5G, reg);
+
 	if (ret < 0)
 		return ret;
 	if (ret > 0)
@@ -337,17 +348,20 @@ static int mv3310_aneg_done(struct phy_device *phydev)
 static void mv3310_update_interface(struct phy_device *phydev)
 {
 	if ((phydev->interface == PHY_INTERFACE_MODE_SGMII ||
+	     phydev->interface == PHY_INTERFACE_MODE_2500BASEX ||
 	     phydev->interface == PHY_INTERFACE_MODE_10GKR) && phydev->link) {
 		/* The PHY automatically switches its serdes interface (and
-		 * active PHYXS instance) between Cisco SGMII and 10GBase-KR
-		 * modes according to the speed.  Florian suggests setting
-		 * phydev->interface to communicate this to the MAC. Only do
-		 * this if we are already in either SGMII or 10GBase-KR mode.
+		 * active PHYXS instance) between Cisco SGMII, 10GBase-KR and
+		 * 2500BaseX modes according to the speed.  Florian suggests
+		 * setting phydev->interface to communicate this to the MAC.
+		 * Only do this if we are already in one of the above modes.
 		 */
 		if (phydev->speed == SPEED_10000)
 			phydev->interface = PHY_INTERFACE_MODE_10GKR;
+		else if (phydev->speed == SPEED_2500)
+			phydev->interface = PHY_INTERFACE_MODE_2500BASEX;
 		else if (phydev->speed >= SPEED_10 &&
-			 phydev->speed < SPEED_10000)
+			 phydev->speed < SPEED_2500)
 			phydev->interface = PHY_INTERFACE_MODE_SGMII;
 	}
 }
@@ -450,7 +464,7 @@ static int mv3310_read_status(struct phy_device *phydev)
 
 static struct phy_driver mv3310_drivers[] = {
 	{
-		.phy_id		= 0x002b09aa,
+		.phy_id		= MARVELL_PHY_ID_88X3310,
 		.phy_id_mask	= MARVELL_PHY_ID_MASK,
 		.name		= "mv88x3310",
 		.features	= PHY_10GBIT_FEATURES,
@@ -468,7 +482,7 @@ static struct phy_driver mv3310_drivers[] = {
 module_phy_driver(mv3310_drivers);
 
 static struct mdio_device_id __maybe_unused mv3310_tbl[] = {
-	{ 0x002b09aa, MARVELL_PHY_ID_MASK },
+	{ MARVELL_PHY_ID_88X3310, MARVELL_PHY_ID_MASK },
 	{ },
 };
 MODULE_DEVICE_TABLE(mdio, mv3310_tbl);
include/linux/marvell_phy.h +1 -0
index 1eb6f244588d..5851d68d828a 100644
@@ -20,6 +20,7 @@
 #define MARVELL_PHY_ID_88E1540		0x01410eb0
 #define MARVELL_PHY_ID_88E1545		0x01410ea0
 #define MARVELL_PHY_ID_88E3016		0x01410e60
+#define MARVELL_PHY_ID_88X3310		0x002b09aa
 
 /* The MV88e6390 Ethernet switch contains embedded PHYs. These PHYs do
  * not have a model ID. So the switch driver traps reads to the ID2
Signature · 1 line
2.19.2
maintainer
#16 lore raw

Re: [PATCH net-next v2 07/10] net: phy: marvell10g: Add support for 2.5GBASET

drivers/net/phy/marvell10g.c static int mv3310_aneg_done(struct phy_device *phydev) · line 366

Quoted 9 lines
drivers/net/phy/marvell10g.c +3 -1
 		if (phydev->speed == SPEED_10000)
 			phydev->interface = PHY_INTERFACE_MODE_10GKR;
+		else if (phydev->speed == SPEED_2500)
+			phydev->interface = PHY_INTERFACE_MODE_2500BASEX;
 		else if (phydev->speed >= SPEED_10 &&
-			 phydev->speed < SPEED_10000)
+			 phydev->speed < SPEED_2500)
 			phydev->interface = PHY_INTERFACE_MODE_SGMII;
 	}

Maybe swap to a switch statement?

drivers/net/phy/marvell10g.c static int mv3310_read_status(struct phy_device *phydev) · line 467

Quoted 4 lines

static struct phy_driver mv3310_drivers[] = {

drivers/net/phy/marvell10g.c +1 -1
 	{
-		.phy_id		= 0x002b09aa,
+		.phy_id		= MARVELL_PHY_ID_88X3310,

What does the datasheet say about the lower nibble? Often it is a silicon revision field, so you don't match on it. But 0xa is a rather odd revision.

Thanks
Andrew

#17 lore raw

Re: [PATCH net-next v2 07/10] net: phy: marvell10g: Add support for 2.5GBASET

In reply to Maxime Chevallier
Quoted 11 lines · Maxime Chevallier

The Marvell Alaska family of PHYs supports 2.5GBaseT and 5GBaseT modes, as defined in the 802.3bz specification.

When the link partner requests a 2.5GBASET link, the PHY will reconfigure it's MII interface to 2500BASEX.

At 5G, the PHY will reconfigure it's interface to 5GBASE-R, but this mode isn't supported by any MAC for now.

This was tested with :
- The 88X3310, which is on the MacchiatoBin

Hi Maxime,

Looking deeper at this, I think we actually need an additional patch at the beginning of your series.

The default AN advertisement in 7.32 is 0x1181 - which includes the 2.5G and 5G modes. We need to clear these bits, so that when the 10G mode disabled via ethtool, we do not switch to 2.5G or 5G speed (both of which are not currently reported as supported.) Such a patch needs backporting to stable kernels.

Quoted 105 lines

- The 88E2010, an Alaska PHY that has no fiber interfaces, and is limited to 5G maximum speed.

---
V1 -> V2: Use a #define for the 88X3310 PHY ID, since it's reused in
various places in the code. Rebased on Heiner Kallweit's patch introducing the phy_modify_mmd accessor.

 drivers/net/phy/marvell10g.c | 30 ++++++++++++++++++++++--------
 include/linux/marvell_phy.h  |  1 +
 2 files changed, 23 insertions(+), 8 deletions(-)
drivers/net/phy/marvell10g.c +22 -8
index 07df87b81369..581b4b6e31e9 100644
@@ -238,6 +238,7 @@ static int mv3310_config_init(struct phy_device *phydev)
 
 	/* Check that the PHY interface type is compatible */
 	if (phydev->interface != PHY_INTERFACE_MODE_SGMII &&
+	    phydev->interface != PHY_INTERFACE_MODE_2500BASEX &&
 	    phydev->interface != PHY_INTERFACE_MODE_XAUI &&
 	    phydev->interface != PHY_INTERFACE_MODE_RXAUI &&
 	    phydev->interface != PHY_INTERFACE_MODE_10GKR)
@@ -307,8 +308,18 @@ static int mv3310_config_aneg(struct phy_device *phydev)
 	else
 		reg = 0;
 
+	if (linkmode_test_bit(ETHTOOL_LINK_MODE_2500baseT_Full_BIT,
+			      phydev->advertising))
+		reg |= MDIO_AN_10GBT_CTRL_ADV2_5G;
+	if (linkmode_test_bit(ETHTOOL_LINK_MODE_5000baseT_Full_BIT,
+			      phydev->advertising))
+		reg |= MDIO_AN_10GBT_CTRL_ADV5G;
+
 	ret = phy_modify_mmd(phydev, MDIO_MMD_AN, MDIO_AN_10GBT_CTRL,
-			     MDIO_AN_10GBT_CTRL_ADV10G, reg);
+			     MDIO_AN_10GBT_CTRL_ADV10G |
+			     MDIO_AN_10GBT_CTRL_ADV5G |
+			     MDIO_AN_10GBT_CTRL_ADV2_5G, reg);
+
 	if (ret < 0)
 		return ret;
 	if (ret > 0)
@@ -337,17 +348,20 @@ static int mv3310_aneg_done(struct phy_device *phydev)
 static void mv3310_update_interface(struct phy_device *phydev)
 {
 	if ((phydev->interface == PHY_INTERFACE_MODE_SGMII ||
+	     phydev->interface == PHY_INTERFACE_MODE_2500BASEX ||
 	     phydev->interface == PHY_INTERFACE_MODE_10GKR) && phydev->link) {
 		/* The PHY automatically switches its serdes interface (and
-		 * active PHYXS instance) between Cisco SGMII and 10GBase-KR
-		 * modes according to the speed.  Florian suggests setting
-		 * phydev->interface to communicate this to the MAC. Only do
-		 * this if we are already in either SGMII or 10GBase-KR mode.
+		 * active PHYXS instance) between Cisco SGMII, 10GBase-KR and
+		 * 2500BaseX modes according to the speed.  Florian suggests
+		 * setting phydev->interface to communicate this to the MAC.
+		 * Only do this if we are already in one of the above modes.
 		 */
 		if (phydev->speed == SPEED_10000)
 			phydev->interface = PHY_INTERFACE_MODE_10GKR;
+		else if (phydev->speed == SPEED_2500)
+			phydev->interface = PHY_INTERFACE_MODE_2500BASEX;
 		else if (phydev->speed >= SPEED_10 &&
-			 phydev->speed < SPEED_10000)
+			 phydev->speed < SPEED_2500)
 			phydev->interface = PHY_INTERFACE_MODE_SGMII;
 	}
 }
@@ -450,7 +464,7 @@ static int mv3310_read_status(struct phy_device *phydev)
 
 static struct phy_driver mv3310_drivers[] = {
 	{
-		.phy_id		= 0x002b09aa,
+		.phy_id		= MARVELL_PHY_ID_88X3310,
 		.phy_id_mask	= MARVELL_PHY_ID_MASK,
 		.name		= "mv88x3310",
 		.features	= PHY_10GBIT_FEATURES,
@@ -468,7 +482,7 @@ static struct phy_driver mv3310_drivers[] = {
 module_phy_driver(mv3310_drivers);
 
 static struct mdio_device_id __maybe_unused mv3310_tbl[] = {
-	{ 0x002b09aa, MARVELL_PHY_ID_MASK },
+	{ MARVELL_PHY_ID_88X3310, MARVELL_PHY_ID_MASK },
 	{ },
 };
 MODULE_DEVICE_TABLE(mdio, mv3310_tbl);
include/linux/marvell_phy.h +1 -0
index 1eb6f244588d..5851d68d828a 100644
@@ -20,6 +20,7 @@
 #define MARVELL_PHY_ID_88E1540		0x01410eb0
 #define MARVELL_PHY_ID_88E1545		0x01410ea0
 #define MARVELL_PHY_ID_88E3016		0x01410e60
+#define MARVELL_PHY_ID_88X3310		0x002b09aa
 
 /* The MV88e6390 Ethernet switch contains embedded PHYs. These PHYs do
  * not have a model ID. So the switch driver traps reads to the ID2
Signature · 1 line
2.19.2
Signature · 3 lines
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line in suburbia: sync at 12.1Mbps down 622kbps up
According to speedtest.net: 11.9Mbps down 500kbps up
#18 lore raw

Re: [PATCH net-next v2 07/10] net: phy: marvell10g: Add support for 2.5GBASET

Hello Russell,

Quoted 15 lines · On Thu, 7 Feb 2019 23:48:24 +0000 Russell King - ARM Linux admin…
Quoted 11 lines · Maxime Chevallier

The Marvell Alaska family of PHYs supports 2.5GBaseT and 5GBaseT modes, as defined in the 802.3bz specification.

When the link partner requests a 2.5GBASET link, the PHY will reconfigure it's MII interface to 2500BASEX.

At 5G, the PHY will reconfigure it's interface to 5GBASE-R, but this mode isn't supported by any MAC for now.

This was tested with :
- The 88X3310, which is on the MacchiatoBin

Hi Maxime,

Quoted 8 lines

Looking deeper at this, I think we actually need an additional patch at the beginning of your series.

The default AN advertisement in 7.32 is 0x1181 - which includes the 2.5G and 5G modes. We need to clear these bits, so that when the 10G mode disabled via ethtool, we do not switch to 2.5G or 5G speed (both of which are not currently reported as supported.) Such a patch needs backporting to stable kernels.

Good catch. The issue seems fixed by Andrew's patch :
3de97f3c6308 ("net: phy: marvell10g: use genphy_c45_an_config_aneg")

However, the fix should indeed be backported to the -stable trees, I've been able to repdocude this on 4.20. I'll take care of sending a patch to -net for that.

Thanks for reporting this,

Maxime

#19 lore raw

PATCH v2 8/10 net: phy: marvell10g: Force reading of 2.5/5G

In reply to Maxime Chevallier

As per 802.3bz, if bit 14 of (1.11) "PMA Extended Abilities" indicates whether or not we should read register (1.21) "2.52/5G PMA Extended Abilities", which contains information on the support of 2.5GBASET and 5GBASET.

After testing on several variants of PHYS of this family, it appears that bit 14 in (1.11) isn't always set when it should be.

PHYs 88X3310 (on MacchiatoBin) and 88E2010 do support 2.5G and 5GBASET, but don't have 1.11.14 set. Their register 1.21 is filled with the correct values, indicating 2.5G and 5G support.

PHYs 88E2110 do have their 1.11.14 bit set, as it should.

---
V1 -> V2: Use the PMA device id to identify which devices need this
quirk, based on Andrew's idea.

 drivers/net/phy/marvell10g.c | 31 +++++++++++++++++++++++++++++++
 1 file changed, 31 insertions(+)
drivers/net/phy/marvell10g.c +31 -0
index 581b4b6e31e9..2593db608af7 100644
@@ -232,6 +232,23 @@ static int mv3310_resume(struct phy_device *phydev)
 	return mv3310_hwmon_config(phydev, true);
 }
 
+/* Some PHYs in the Alaska family such as the 88X3310 and the 88E2010
+ * don't set bit 14 in PMA Extended Abilities (1.11), although they do
+ * support 2.5GBASET and 5GBASET. For these models, we can still read their
+ * 2.5G/5G extended abilities register (1.21). We detect these models based on
+ * the PMA device identifier, with a mask matching models known to have this
+ * issue
+ */
+static bool mv3310_has_pma_ngbaset_quirk(struct phy_device *phydev)
+{
+	if (!(phydev->c45_ids.devices_in_package & MDIO_DEVS_PMAPMD))
+		return false;
+
+	/* Only some revisions of the 88X3310 family PMA seem to be impacted */
+	return (phydev->c45_ids.device_ids[MDIO_MMD_PMAPMD] & 0xfffffffe) ==
+	       (MARVELL_PHY_ID_88X3310 & 0xfffffffe);
+}
+
 static int mv3310_config_init(struct phy_device *phydev)
 {
 	int ret, val;
@@ -261,6 +278,20 @@ static int mv3310_config_init(struct phy_device *phydev)
 	if (ret)
 		return ret;
 
+	if (mv3310_has_pma_ngbaset_quirk(phydev)) {
+		val = phy_read_mmd(phydev, MDIO_MMD_PMAPMD,
+				   MDIO_PMA_NG_EXTABLE);
+		if (val < 0)
+			return val;
+
+		if (val & MDIO_PMA_NG_EXTABLE_2_5GBT)
+			__set_bit(ETHTOOL_LINK_MODE_2500baseT_Full_BIT,
+				  phydev->supported);
+		if (val & MDIO_PMA_NG_EXTABLE_5GBT)
+			__set_bit(ETHTOOL_LINK_MODE_5000baseT_Full_BIT,
+				  phydev->supported);
+	}
+
 	return 0;
 }
 
Signature · 1 line
2.19.2
#20 lore raw

PATCH v2 9/10 net: mvpp2: Add 2.5GBaseT support

In reply to Maxime Chevallier

The PPv2 controller is able to support 2.5G speeds, allowing to use 2.5GBASET in conjunction with PHYs that use 2500BASEX as their MII interface when using this mode.

---
 drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c | 1 +
 1 file changed, 1 insertion(+)
drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c +1 -0
index 5c0c26fd8363..6f36aed5bd0a 100644
@@ -4410,6 +4410,7 @@ static void mvpp2_phylink_validate(struct net_device *dev,
 	case PHY_INTERFACE_MODE_2500BASEX:
 		phylink_set(mask, 1000baseT_Full);
 		phylink_set(mask, 1000baseX_Full);
+		phylink_set(mask, 2500baseT_Full);
 		phylink_set(mask, 2500baseX_Full);
 		break;
 	default:
Signature · 1 line
2.19.2
#22 lore raw

PATCH v2 10/10 net: phy: marvell10g: add support for the 88x2110 PHY

In reply to Maxime Chevallier

This patch adds support for the 88x2110 PHY, which is similar to the already supported 88x3310 PHY without the SFP interface.

It supports 10/100/1000BASET along with 2.5GBASET, 5GBASET and 10GBASET, with the same interface modes that are used by the 3310.

This PHY don't have the same issue as the 88x3310 regarding 2.5/5G abilities, and correctly follows the 802.3bz standard to list the supported abilities.

---
V1 -> V2: Use a #define for the PHY ID, since we also do that for the
3310. Removed the mv2110_config_init implementation since we now manually check for the PMA device id when using the quirk.

 drivers/net/phy/marvell10g.c | 13 +++++++++++++
 include/linux/marvell_phy.h  |  1 +
 2 files changed, 14 insertions(+)
drivers/net/phy/marvell10g.c +13 -0
index 2593db608af7..fa4847662c81 100644
@@ -508,12 +508,25 @@ static struct phy_driver mv3310_drivers[] = {
 		.aneg_done	= mv3310_aneg_done,
 		.read_status	= mv3310_read_status,
 	},
+	{
+		.phy_id		= MARVELL_PHY_ID_88E2110,
+		.phy_id_mask	= MARVELL_PHY_ID_MASK,
+		.name		= "mv88x2110",
+		.features	= PHY_10GBIT_FEATURES,
+		.probe		= mv3310_probe,
+		.soft_reset	= gen10g_no_soft_reset,
+		.config_init	= mv3310_config_init,
+		.config_aneg	= mv3310_config_aneg,
+		.aneg_done	= mv3310_aneg_done,
+		.read_status	= mv3310_read_status,
+	},
 };
 
 module_phy_driver(mv3310_drivers);
 
 static struct mdio_device_id __maybe_unused mv3310_tbl[] = {
 	{ MARVELL_PHY_ID_88X3310, MARVELL_PHY_ID_MASK },
+	{ MARVELL_PHY_ID_88E2110, MARVELL_PHY_ID_MASK },
 	{ },
 };
 MODULE_DEVICE_TABLE(mdio, mv3310_tbl);
include/linux/marvell_phy.h +1 -0
index 5851d68d828a..8596a861940e 100644
@@ -21,6 +21,7 @@
 #define MARVELL_PHY_ID_88E1545		0x01410ea0
 #define MARVELL_PHY_ID_88E3016		0x01410e60
 #define MARVELL_PHY_ID_88X3310		0x002b09aa
+#define MARVELL_PHY_ID_88E2110		0x002b09b8
 
 /* The MV88e6390 Ethernet switch contains embedded PHYs. These PHYs do
  * not have a model ID. So the switch driver traps reads to the ID2
Signature · 1 line
2.19.2
maintainer
#23 lore raw
In reply to Maxime Chevallier
Quoted 49 lines · Maxime Chevallier

Hello everyone,

This is the second iteration of the series introducing support for 2.5GBASET and 5GBASET to the network PHY infrastructure.

These 2 modes are described in the 802.3bz specifications, and allow to use 2.5G and 5G speeds on cat5e/cat6 cables.

The required infrastructure code is added for 2.5GBASET and 5GBASET, but only 2.5GBASET support is added to the Marvell10G driver. The reason is because the 5GBASER interface mode used to communicate with the MAC isn't implemented yet, and therefore this can't be tested.

This series has seen some rework since last time, see the full details in the patches changelog. The main highlights are :

- Patch 1 moves the Pause ans Asym_Pause parameters update after calling config_init. This allows us to change the phydev->supported modes in config_init, while still forcing some quirks taken from phydrv->features, to disable unsupported Pause modes. This was proposed by Andrew.

- Patch 2 generalizes the way we mask-out modes we don't want to use when forcing the link speed through DT or ethtool, by walking through the PHY settings table, as proposed by Russell.

- Patch 4 implements automatic setting of TM, FIBRE and Backplane bits from the list of supported linkmodes.

- In Patch 5, we only read abilities from the PMA. Pause parameters aren't built from the genphy_c45_pma_read_abilities, as it was done in the previous iteration of the patch. We also amke use of linkmode_mod_bit to make sure we mask-out unsupported modes.

- In Patch 8, we manually check for the PMA device id to see if we have to use a quirk when reading the 2.5G/5G Extended Abilities

Maxime Chevallier (10):
net: phy: Update PHY linkmodes after config_init
net: phy: Mask-out non-compatible modes when setting the max-speed
net: phy: Move of_set_phy_eee_broken to phy-core.c
net: phy: Automatically fill the generic TP, FIBRE and Backplane modes
net: phy: Extract genphy_c45_pma_read_abilities from marvell10g
net: phy: Add generic support for 2.5GBaseT and 5GBaseT
net: phy: marvell10g: Add support for 2.5GBASET
net: phy: marvell10g: Force reading of 2.5/5G
net: mvpp2: Add 2.5GBaseT support
net: phy: marvell10g: add support for the 88x2110 PHY

Quoted 11 lines
 .../net/ethernet/marvell/mvpp2/mvpp2_main.c   |   1 +
 drivers/net/phy/marvell10g.c                  | 142 +++++------
 drivers/net/phy/phy-c45.c                     | 111 ++++++++
 drivers/net/phy/phy-core.c                    |  72 ++++++
 drivers/net/phy/phy_device.c                  | 237 +++++++++---------
 include/linux/linkmode.h                      |   6 +
 include/linux/marvell_phy.h                   |   2 +
 include/linux/phy.h                           |   3 +
 include/uapi/linux/mdio.h                     |  16 ++
 9 files changed, 405 insertions(+), 185 deletions(-)

Hi Maxime,

Andrew and me are working on Aquantia PHY support and he handed over to me a patch series which includes parts of the first version of your series. Having said that I'm especially interested in your patches 5 and 6. Because your series is somewhat bigger and there are a few review comments, preparing the next round may take time.

I'd propose that you extract generic patches being submission-ready and split the patch series into two. I think the following patches would be candidates for the first series: 2, 3, 5, 6 (provided they have no dependency on the other patches) Based on that both of us can go on with our work.

Andrew, what do you think?

Heiner

maintainer
#24 lore raw
Quoted 7 lines

I'd propose that you extract generic patches being submission-ready and split the patch series into two. I think the following patches would be candidates for the first series: 2, 3, 5, 6 (provided they have no dependency on the other patches) Based on that both of us can go on with our work.

Andrew, what do you think?

Yes, that would help.

Heiner, can you also submit the .get_features patches soon. That is another important piece of the puzzle for both drivers.

Thanks
Andrew

maintainer
#25 lore raw
Quoted 6 lines · Andrew Lunn
Quoted 6 lines

I'd propose that you extract generic patches being submission-ready and split the patch series into two. I think the following patches would be candidates for the first series: 2, 3, 5, 6 (provided they have no dependency on the other patches) Based on that both of us can go on with our work.

Quoted 7 lines
Quoted 1 line

Andrew, what do you think?

Yes, that would help.

Heiner, can you also submit the .get_features patches soon. That is another important piece of the puzzle for both drivers.

Just submitted it few hours ago.
https://patchwork.ozlabs.org/patch/1039237/

Quoted 3 lines

Thanks
Andrew

Heiner

#26 lore raw
In reply to Heiner Kallweit

Hello Heiner,

Quoted 8 lines

Hi Maxime,

Andrew and me are working on Aquantia PHY support and he handed over to me a patch series which includes parts of the first version of your series. Having said that I'm especially interested in your patches 5 and 6. Because your series is somewhat bigger and there are a few review comments, preparing the next round may take time.

Quoted 5 lines

I'd propose that you extract generic patches being submission-ready and split the patch series into two. I think the following patches would be candidates for the first series: 2, 3, 5, 6 (provided they have no dependency on the other patches) Based on that both of us can go on with our work.

Sure, I'll sent that shortly. Thanks for the help,

Maxime