Skip to contentSkip to conversation

net: phy: dp83867: w/a for fld detect threshold bootstrapping issue

PATCH Applied Grygorii Strashko 2 participants 4 messages Mar 17, 2020 – Mar 24, 2020

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 net: phy: dp83867: w/a for fld detect threshold bootstrapping issue

When the DP83867 PHY is strapped to enable Fast Link Drop (FLD) feature STRAP_STS2.STRAP_ FLD (reg 0x006F bit 10), the Energy Lost Threshold for FLD Energy Lost Mode FLD_THR_CFG.ENERGY_LOST_FLD_THR (reg 0x002e bits 2:0) will be defaulted to 0x2. This may cause the phy link to be unstable. The new DP83867 DM recommends to always restore ENERGY_LOST_FLD_THR to 0x1.

Hence, restore default value of FLD_THR_CFG.ENERGY_LOST_FLD_THR to 0x1 when FLD is enabled by bootstrapping as recommended by DM.

---
 drivers/net/phy/dp83867.c | 21 ++++++++++++++++++++-
 1 file changed, 20 insertions(+), 1 deletion(-)
drivers/net/phy/dp83867.c +20 -1
index 13f7f2d5a2ea..b55e3c0403ed 100644
@@ -30,7 +30,8 @@
 #define DP83867_CTRL		0x1f
 
 /* Extended Registers */
-#define DP83867_CFG4            0x0031
+#define DP83867_FLD_THR_CFG	0x002e
+#define DP83867_CFG4		0x0031
 #define DP83867_CFG4_SGMII_ANEG_MASK (BIT(5) | BIT(6))
 #define DP83867_CFG4_SGMII_ANEG_TIMER_11MS   (3 << 5)
 #define DP83867_CFG4_SGMII_ANEG_TIMER_800US  (2 << 5)
@@ -93,6 +94,7 @@
 #define DP83867_STRAP_STS2_CLK_SKEW_RX_MASK	GENMASK(2, 0)
 #define DP83867_STRAP_STS2_CLK_SKEW_RX_SHIFT	0
 #define DP83867_STRAP_STS2_CLK_SKEW_NONE	BIT(2)
+#define DP83867_STRAP_STS2_STRAP_FLD		BIT(10)
 
 /* PHY CTRL bits */
 #define DP83867_PHYCR_TX_FIFO_DEPTH_SHIFT	14
@@ -145,6 +147,9 @@
 /* CFG4 bits */
 #define DP83867_CFG4_PORT_MIRROR_EN              BIT(0)
 
+/* FLD_THR_CFG */
+#define DP83867_FLD_THR_CFG_ENERGY_LOST_THR_MASK	0x7
+
 enum {
 	DP83867_PORT_MIRROING_KEEP,
 	DP83867_PORT_MIRROING_EN,
@@ -622,6 +627,20 @@ static int dp83867_config_init(struct phy_device *phydev)
 		phy_clear_bits_mmd(phydev, DP83867_DEVADDR, DP83867_CFG4,
 				   BIT(7));
 
+	bs = phy_read_mmd(phydev, DP83867_DEVADDR, DP83867_STRAP_STS2);
+	if (bs & DP83867_STRAP_STS2_STRAP_FLD) {
+		/* When using strap to enable FLD, the ENERGY_LOST_FLD_THR will
+		 * be set to 0x2. This may causes the PHY link to be unstable -
+		 * the default value 0x1 need to be restored.
+		 */
+		ret = phy_modify_mmd(phydev, DP83867_DEVADDR,
+				     DP83867_FLD_THR_CFG,
+				     DP83867_FLD_THR_CFG_ENERGY_LOST_THR_MASK,
+				     0x1);
+		if (ret)
+			return ret;
+	}
+
 	if (phy_interface_is_rgmii(phydev) ||
 	    phydev->interface == PHY_INTERFACE_MODE_SGMII) {
 		val = phy_read(phydev, MII_DP83867_PHYCTRL);
Signature · 1 line
2.17.1
maintainer Applied
#1 lore raw

From: Grygorii Strashko <[email protected]>
Date: Tue, 17 Mar 2020 20:04:54 +0200

Quoted 10 lines

When the DP83867 PHY is strapped to enable Fast Link Drop (FLD) feature STRAP_STS2.STRAP_ FLD (reg 0x006F bit 10), the Energy Lost Threshold for FLD Energy Lost Mode FLD_THR_CFG.ENERGY_LOST_FLD_THR (reg 0x002e bits 2:0) will be defaulted to 0x2. This may cause the phy link to be unstable. The new DP83867 DM recommends to always restore ENERGY_LOST_FLD_THR to 0x1.

Hence, restore default value of FLD_THR_CFG.ENERGY_LOST_FLD_THR to 0x1 when FLD is enabled by bootstrapping as recommended by DM.

Applied, thank you.

Let me know if I should queue this up for -stable.

#2 lore raw
Quoted 9 lines · David Miller

From: Grygorii Strashko <[email protected]>
Date: Tue, 17 Mar 2020 20:04:54 +0200

Quoted 6 lines

When the DP83867 PHY is strapped to enable Fast Link Drop (FLD) feature STRAP_STS2.STRAP_ FLD (reg 0x006F bit 10), the Energy Lost Threshold for FLD Energy Lost Mode FLD_THR_CFG.ENERGY_LOST_FLD_THR (reg 0x002e bits 2:0) will be defaulted to 0x2. This may cause the phy link to be unstable. The new DP83867 DM recommends to always restore ENERGY_LOST_FLD_THR to 0x1.

Quoted 7 lines
Quoted 4 lines

Hence, restore default value of FLD_THR_CFG.ENERGY_LOST_FLD_THR to 0x1 when FLD is enabled by bootstrapping as recommended by DM.

Applied, thank you.

Thank you.

Quoted 2 lines

Let me know if I should queue this up for -stable.

yes, please, as there are real link instability issues were observed without this change.

Signature · 2 lines
Best regards,
grygorii
maintainer
#3 lore raw

From: Grygorii Strashko <[email protected]>
Date: Mon, 23 Mar 2020 22:36:32 +0200

Quoted 6 lines · David Miller
Quoted 2 lines

Let me know if I should queue this up for -stable.

yes, please, as there are real link instability issues were observed without this change.

Ok, done.