[PATCH net v8 0/3] Handle new Microchip KSZ 9897 Errata

[PATCH net v8 0/3] Handle new Microchip KSZ 9897 Errata

[Enguerrand de Ribaucourt]

These patches implement some suggested workarounds from the Microchip KSZ 9897
Errata [1].

[1] https://ww1.microchip.com/downloads/aemDocuments/documents/UNG/ProductDocuments/Errata/KSZ9897R-Errata-DS80000758.pdf

---
v8:
 - split monitoring function in two to fit into 80 columns
v7: https://lore.kernel.org/netdev/20240621144322.545908-1-enguerrand.de-ribaucourt@savoirfairelinux.com/
 - use dev_crit_once instead of dev_crit_ratelimited
 - add a comment to help users understand the consequences of half-duplex errors
v6: https://lore.kernel.org/netdev/20240614094642.122464-1-enguerrand.de-ribaucourt@savoirfairelinux.com/
 - remove KSZ9897 phy_id workaround (was a configuration issue)
 - use macros for checking link down in monitoring function
 - check if VLAN is enabled before monitoring resources
v5: https://lore.kernel.org/all/20240604092304.314636-1-enguerrand.de-ribaucourt@savoirfairelinux.com/
 - use macros for bitfields
 - rewrap comments
 - check ksz_pread* return values
 - fix spelling mistakes
 - remove KSZ9477 suspend/resume deletion patch
v4: https://lore.kernel.org/all/20240531142430.678198-1-enguerrand.de-ribaucourt@savoirfairelinux.com/
 - Rebase on net/main
 - Add Fixes: tags to the patches
 - reverse x-mas tree order
 - use pseudo phy_id instead of match_phy_device
v3: https://lore.kernel.org/all/20240530102436.226189-1-enguerrand.de-ribaucourt@savoirfairelinux.com/

[PATCH net v8 1/3] net: phy: micrel: add Microchip KSZ 9477 to the device table

[Enguerrand de Ribaucourt]

PHY_ID_KSZ9477 was supported but not added to the device table passed to
MODULE_DEVICE_TABLE.

Fixes: fc3973a1fa09 ("phy: micrel: add Microchip KSZ 9477 Switch PHY support")
Signed-off-by: Enguerrand de Ribaucourt <enguerrand.de-ribaucourt@savoirfairelinux.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
---
v4:
 - rebase on net/main
 - add Fixes tag
v3: https://lore.kernel.org/all/20240530102436.226189-4-enguerrand.de-ribaucourt@savoirfairelinux.com/
---
 drivers/net/phy/micrel.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/net/phy/micrel.c b/drivers/net/phy/micrel.c
index 5aada7cf3da7..ebafedde0ab7 100644
--- a/drivers/net/phy/micrel.c
+++ b/drivers/net/phy/micrel.c
@@ -5607,6 +5607,7 @@ static struct mdio_device_id __maybe_unused micrel_tbl[] = {
 	{ PHY_ID_KSZ8081, MICREL_PHY_ID_MASK },
 	{ PHY_ID_KSZ8873MLL, MICREL_PHY_ID_MASK },
 	{ PHY_ID_KSZ886X, MICREL_PHY_ID_MASK },
+	{ PHY_ID_KSZ9477, MICREL_PHY_ID_MASK },
 	{ PHY_ID_LAN8814, MICREL_PHY_ID_MASK },
 	{ PHY_ID_LAN8804, MICREL_PHY_ID_MASK },
 	{ PHY_ID_LAN8841, MICREL_PHY_ID_MASK },
-- 
2.34.1

[PATCH net v8 2/3] net: dsa: microchip: use collision based back pressure mode

[Enguerrand de Ribaucourt]

Errata DS80000758 states that carrier sense back pressure mode can cause
link down issues in 100BASE-TX half duplex mode. The datasheet also
recommends to always use the collision based back pressure mode.

Fixes: b987e98e50ab ("dsa: add DSA switch driver for Microchip KSZ9477")
Signed-off-by: Enguerrand de Ribaucourt <enguerrand.de-ribaucourt@savoirfairelinux.com>
Reviewed-by: Woojung Huh <Woojung.huh@microchip.com>
Acked-by: Arun Ramadoss <arun.ramadoss@microchip.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
---
v5:
 - define SW_BACK_PRESSURE_COLLISION
v4: https://lore.kernel.org/all/20240531142430.678198-5-enguerrand.de-ribaucourt@savoirfairelinux.com/
 - rebase on net/main
 - add Fixes tag
v3: https://lore.kernel.org/all/20240530102436.226189-5-enguerrand.de-ribaucourt@savoirfairelinux.com/
---
 drivers/net/dsa/microchip/ksz9477.c     | 4 ++++
 drivers/net/dsa/microchip/ksz9477_reg.h | 1 +
 2 files changed, 5 insertions(+)

diff --git a/drivers/net/dsa/microchip/ksz9477.c b/drivers/net/dsa/microchip/ksz9477.c
index f8ad7833f5d9..c2878dd0ad7e 100644
--- a/drivers/net/dsa/microchip/ksz9477.c
+++ b/drivers/net/dsa/microchip/ksz9477.c
@@ -1299,6 +1299,10 @@ int ksz9477_setup(struct dsa_switch *ds)
 	/* Enable REG_SW_MTU__2 reg by setting SW_JUMBO_PACKET */
 	ksz_cfg(dev, REG_SW_MAC_CTRL_1, SW_JUMBO_PACKET, true);
 
+	/* Use collision based back pressure mode. */
+	ksz_cfg(dev, REG_SW_MAC_CTRL_1, SW_BACK_PRESSURE,
+		SW_BACK_PRESSURE_COLLISION);
+
 	/* Now we can configure default MTU value */
 	ret = regmap_update_bits(ksz_regmap_16(dev), REG_SW_MTU__2, REG_SW_MTU_MASK,
 				 VLAN_ETH_FRAME_LEN + ETH_FCS_LEN);
diff --git a/drivers/net/dsa/microchip/ksz9477_reg.h b/drivers/net/dsa/microchip/ksz9477_reg.h
index f3a205ee483f..fb124be8edd3 100644
--- a/drivers/net/dsa/microchip/ksz9477_reg.h
+++ b/drivers/net/dsa/microchip/ksz9477_reg.h
@@ -247,6 +247,7 @@
 #define REG_SW_MAC_CTRL_1		0x0331
 
 #define SW_BACK_PRESSURE		BIT(5)
+#define SW_BACK_PRESSURE_COLLISION	0
 #define FAIR_FLOW_CTRL			BIT(4)
 #define NO_EXC_COLLISION_DROP		BIT(3)
 #define SW_JUMBO_PACKET			BIT(2)
-- 
2.34.1

[PATCH net v8 3/3] net: dsa: microchip: monitor potential faults in half-duplex mode

[Enguerrand de Ribaucourt]

The errata DS80000754 recommends monitoring potential faults in
half-duplex mode for the KSZ9477 family.

half-duplex is not very common so I just added a critical message
when the fault conditions are detected. The switch can be expected
to be unable to communicate anymore in these states and a software
reset of the switch would be required which I did not implement.

Fixes: b987e98e50ab ("dsa: add DSA switch driver for Microchip KSZ9477")
Signed-off-by: Enguerrand de Ribaucourt <enguerrand.de-ribaucourt@savoirfairelinux.com>
---
v8:
 - split the function in two to fit into 80 columns
v7: https://lore.kernel.org/netdev/Znqbe6HDrajK0UVq@shell.armlinux.org.uk/
 - use dev_crit_once instead of dev_crit_ratelimited
   The condition will remain true forever and the routine called every
   5 seconds. There's no need to repeat the message.
 - add explanations in the comment above the warning to help users
   anticipate the consequences of the fault.
v6: https://lore.kernel.org/netdev/20240614094642.122464-4-enguerrand.de-ribaucourt@savoirfairelinux.com/
 - use macros for PORT_INTF_SPEED_MASK check
 - add VLAN condition before checking the resources
v5: https://lore.kernel.org/all/20240604092304.314636-5-enguerrand.de-ribaucourt@savoirfairelinux.com/
 - use macros for bitmasks
 - check for return values on ksz_pread*
v4: https://lore.kernel.org/all/20240531142430.678198-6-enguerrand.de-ribaucourt@savoirfairelinux.com/
 - rebase on net/main
 - add Fixes tag
 - reverse x-mas tree
v3: https://lore.kernel.org/all/20240530102436.226189-6-enguerrand.de-ribaucourt@savoirfairelinux.com/
---
 drivers/net/dsa/microchip/ksz9477.c     | 64 +++++++++++++++++++++++++
 drivers/net/dsa/microchip/ksz9477.h     |  2 +
 drivers/net/dsa/microchip/ksz9477_reg.h | 10 +++-
 drivers/net/dsa/microchip/ksz_common.c  | 11 +++++
 drivers/net/dsa/microchip/ksz_common.h  |  1 +
 5 files changed, 86 insertions(+), 2 deletions(-)

diff --git a/drivers/net/dsa/microchip/ksz9477.c b/drivers/net/dsa/microchip/ksz9477.c
index c2878dd0ad7e..0f29eb66f9c5 100644
--- a/drivers/net/dsa/microchip/ksz9477.c
+++ b/drivers/net/dsa/microchip/ksz9477.c
@@ -429,6 +429,70 @@ void ksz9477_freeze_mib(struct ksz_device *dev, int port, bool freeze)
 	mutex_unlock(&p->mib.cnt_mutex);
 }
 
+static int ksz9477_half_duplex_monitor(struct ksz_device *dev, int port,
+				       u64 tx_late_col)
+{
+	u8 lue_ctrl;
+	u32 pmavbc;
+	u16 pqm;
+	int ret;
+
+	/* Errata DS80000754 recommends monitoring potential faults in
+	 * half-duplex mode. The switch might not be able to communicate anymore
+	 * in these states. If you see this message, please read the
+	 * errata-sheet for more information:
+	 * https://ww1.microchip.com/downloads/aemDocuments/documents
+	 * /UNG/ProductDocuments/Errata/KSZ9477S-Errata-DS80000754.pdf
+	 * To workaround this issue, half-duplex mode should be avoided.
+	 * A software reset could be implemented to recover from this state.
+	 */
+	dev_warn_once(dev->dev,
+		      "Half-duplex detected on port %d, transmission halt may occur\n",
+		      port);
+	if (tx_late_col != 0) {
+		/* Transmission halt with late collisions */
+		dev_crit_once(dev->dev,
+			      "TX late collisions detected, transmission may be halted on port %d\n",
+			      port);
+	}
+	ret = ksz_read8(dev, REG_SW_LUE_CTRL_0, &lue_ctrl);
+	if (ret)
+		return ret;
+	if (lue_ctrl & SW_VLAN_ENABLE) {
+		ret = ksz_pread16(dev, port, REG_PORT_QM_TX_CNT_0__4, &pqm);
+		if (ret)
+			return ret;
+		ret = ksz_read32(dev, REG_PMAVBC, &pmavbc);
+		if (ret)
+			return ret;
+		if ((FIELD_GET(PMAVBC_MASK, pmavbc) <= PMAVBC_MIN) ||
+		    (FIELD_GET(PORT_QM_TX_CNT_M, pqm) >= PORT_QM_TX_CNT_MAX)) {
+			/* Transmission halt with Half-Duplex and VLAN */
+			dev_crit_once(dev->dev,
+				      "resources out of limits, transmission may be halted\n");
+		}
+	}
+
+	return ret;
+}
+
+int ksz9477_errata_monitor(struct ksz_device *dev, int port,
+			   u64 tx_late_col)
+{
+	u8 status;
+	int ret;
+
+	ret = ksz_pread8(dev, port, REG_PORT_STATUS_0, &status);
+	if (ret)
+		return ret;
+	if (!(FIELD_GET(PORT_INTF_SPEED_MASK, status)
+	      == PORT_INTF_SPEED_NONE) &&
+	    !(status & PORT_INTF_FULL_DUPLEX)) {
+		ret = ksz9477_half_duplex_monitor(dev, port, tx_late_col);
+	}
+	return ret;
+}
+
 void ksz9477_port_init_cnt(struct ksz_device *dev, int port)
 {
 	struct ksz_port_mib *mib = &dev->ports[port].mib;
diff --git a/drivers/net/dsa/microchip/ksz9477.h b/drivers/net/dsa/microchip/ksz9477.h
index ce1e656b800b..239a281da10b 100644
--- a/drivers/net/dsa/microchip/ksz9477.h
+++ b/drivers/net/dsa/microchip/ksz9477.h
@@ -36,6 +36,8 @@ int ksz9477_port_mirror_add(struct ksz_device *dev, int port,
 			    bool ingress, struct netlink_ext_ack *extack);
 void ksz9477_port_mirror_del(struct ksz_device *dev, int port,
 			     struct dsa_mall_mirror_tc_entry *mirror);
+int ksz9477_errata_monitor(struct ksz_device *dev, int port,
+			   u64 tx_late_col);
 void ksz9477_get_caps(struct ksz_device *dev, int port,
 		      struct phylink_config *config);
 int ksz9477_fdb_dump(struct ksz_device *dev, int port,
diff --git a/drivers/net/dsa/microchip/ksz9477_reg.h b/drivers/net/dsa/microchip/ksz9477_reg.h
index fb124be8edd3..d5354c600ea1 100644
--- a/drivers/net/dsa/microchip/ksz9477_reg.h
+++ b/drivers/net/dsa/microchip/ksz9477_reg.h
@@ -843,8 +843,8 @@
 
 #define REG_PORT_STATUS_0		0x0030
 
-#define PORT_INTF_SPEED_M		0x3
-#define PORT_INTF_SPEED_S		3
+#define PORT_INTF_SPEED_MASK		GENMASK(4, 3)
+#define PORT_INTF_SPEED_NONE		GENMASK(1, 0)
 #define PORT_INTF_FULL_DUPLEX		BIT(2)
 #define PORT_TX_FLOW_CTRL		BIT(1)
 #define PORT_RX_FLOW_CTRL		BIT(0)
@@ -1168,6 +1168,11 @@
 #define PORT_RMII_CLK_SEL		BIT(7)
 #define PORT_MII_SEL_EDGE		BIT(5)
 
+#define REG_PMAVBC			0x03AC
+
+#define PMAVBC_MASK			GENMASK(26, 16)
+#define PMAVBC_MIN			0x580
+
 /* 4 - MAC */
 #define REG_PORT_MAC_CTRL_0		0x0400
 
@@ -1495,6 +1500,7 @@
 
 #define PORT_QM_TX_CNT_USED_S		0
 #define PORT_QM_TX_CNT_M		(BIT(11) - 1)
+#define PORT_QM_TX_CNT_MAX		0x200
 
 #define REG_PORT_QM_TX_CNT_1__4		0x0A14
 
diff --git a/drivers/net/dsa/microchip/ksz_common.c b/drivers/net/dsa/microchip/ksz_common.c
index 2818e24e2a51..0433109b42e5 100644
--- a/drivers/net/dsa/microchip/ksz_common.c
+++ b/drivers/net/dsa/microchip/ksz_common.c
@@ -1382,6 +1382,7 @@ const struct ksz_chip_data ksz_switch_chips[] = {
 		.tc_cbs_supported = true,
 		.ops = &ksz9477_dev_ops,
 		.phylink_mac_ops = &ksz9477_phylink_mac_ops,
+		.phy_errata_9477 = true,
 		.mib_names = ksz9477_mib_names,
 		.mib_cnt = ARRAY_SIZE(ksz9477_mib_names),
 		.reg_mib_cnt = MIB_COUNTER_NUM,
@@ -1416,6 +1417,7 @@ const struct ksz_chip_data ksz_switch_chips[] = {
 		.num_ipms = 8,
 		.ops = &ksz9477_dev_ops,
 		.phylink_mac_ops = &ksz9477_phylink_mac_ops,
+		.phy_errata_9477 = true,
 		.mib_names = ksz9477_mib_names,
 		.mib_cnt = ARRAY_SIZE(ksz9477_mib_names),
 		.reg_mib_cnt = MIB_COUNTER_NUM,
@@ -1450,6 +1452,7 @@ const struct ksz_chip_data ksz_switch_chips[] = {
 		.num_ipms = 8,
 		.ops = &ksz9477_dev_ops,
 		.phylink_mac_ops = &ksz9477_phylink_mac_ops,
+		.phy_errata_9477 = true,
 		.mib_names = ksz9477_mib_names,
 		.mib_cnt = ARRAY_SIZE(ksz9477_mib_names),
 		.reg_mib_cnt = MIB_COUNTER_NUM,
@@ -1540,6 +1543,7 @@ const struct ksz_chip_data ksz_switch_chips[] = {
 		.tc_cbs_supported = true,
 		.ops = &ksz9477_dev_ops,
 		.phylink_mac_ops = &ksz9477_phylink_mac_ops,
+		.phy_errata_9477 = true,
 		.mib_names = ksz9477_mib_names,
 		.mib_cnt = ARRAY_SIZE(ksz9477_mib_names),
 		.reg_mib_cnt = MIB_COUNTER_NUM,
@@ -1820,6 +1824,7 @@ void ksz_r_mib_stats64(struct ksz_device *dev, int port)
 	struct rtnl_link_stats64 *stats;
 	struct ksz_stats_raw *raw;
 	struct ksz_port_mib *mib;
+	int ret;
 
 	mib = &dev->ports[port].mib;
 	stats = &mib->stats64;
@@ -1861,6 +1866,12 @@ void ksz_r_mib_stats64(struct ksz_device *dev, int port)
 	pstats->rx_pause_frames = raw->rx_pause;
 
 	spin_unlock(&mib->stats64_lock);
+
+	if (dev->info->phy_errata_9477) {
+		ret = ksz9477_errata_monitor(dev, port, raw->tx_late_col);
+		if (ret)
+			dev_err(dev->dev, "Failed to monitor transmission halt\n");
+	}
 }
 
 void ksz88xx_r_mib_stats64(struct ksz_device *dev, int port)
diff --git a/drivers/net/dsa/microchip/ksz_common.h b/drivers/net/dsa/microchip/ksz_common.h
index c784fd23a993..ee7db46e469d 100644
--- a/drivers/net/dsa/microchip/ksz_common.h
+++ b/drivers/net/dsa/microchip/ksz_common.h
@@ -66,6 +66,7 @@ struct ksz_chip_data {
 	bool tc_cbs_supported;
 	const struct ksz_dev_ops *ops;
 	const struct phylink_mac_ops *phylink_mac_ops;
+	bool phy_errata_9477;
 	bool ksz87xx_eee_link_erratum;
 	const struct ksz_mib_names *mib_names;
 	int mib_cnt;
-- 
2.34.1

Re: [PATCH net v8 0/3] Handle new Microchip KSZ 9897 Errata

[Jakub Kicinski]

On Tue, 25 Jun 2024 16:05:17 +0000 Enguerrand de Ribaucourt wrote:
> v8:
>  - split monitoring function in two to fit into 80 columns
> v7: https://lore.kernel.org/netdev/20240621144322.545908-1-enguerrand.de-ribaucourt@savoirfairelinux.com/

too late, v7 has been applied:

https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net.git/commit/?id=4ae2c67840a0a3c88cd71fc3013f958d60f7e50c

:(
-- 
pw-bot: nap

Re: [PATCH net v8 0/3] Handle new Microchip KSZ 9897 Errata

[Enguerrand de Ribaucourt]

Ok, I can do an extra commit then?

On 26/06/2024 03:47, Jakub Kicinski wrote:
> On Tue, 25 Jun 2024 16:05:17 +0000 Enguerrand de Ribaucourt wrote:
>> v8:
>>   - split monitoring function in two to fit into 80 columns
>> v7: https://lore.kernel.org/netdev/20240621144322.545908-1-enguerrand.de-ribaucourt@savoirfairelinux.com/
> 
> too late, v7 has been applied:
> 
> https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net.git/commit/?id=4ae2c67840a0a3c88cd71fc3013f958d60f7e50c
> 
> :(

-- 
Savoir-faire Linux
Enguerrand de Ribaucourt

Re: [PATCH net v8 0/3] Handle new Microchip KSZ 9897 Errata

[Jakub Kicinski]

On Wed, 26 Jun 2024 09:28:16 +0200 Enguerrand de Ribaucourt wrote:
> Ok, I can do an extra commit then?

Sure, but if it's cosmetic (no functional changes) please wait until
Friday and target net-next.