Skip to contentSkip to conversation

net: dsa: microchip: Add KSZ8463 switch support

PATCH Superseded [email protected] 5 participants 18 messages Jul 8, 2025 – Jul 9, 2025

Subsystems (5)

v2 of 6 Superseded Latest is v6 →

lore
Share & tools
lore
#0 lore raw

v2 cover net: dsa: microchip: Add KSZ8463 switch support

From: Tristram Ha <[email protected]>

This series of patches is to add KSZ8463 switch support to the KSZ DSA driver.

v2
- Break the KSZ8463 driver code into several patches for easy review
- Replace ntohs with cpu_to_be16

Tristram Ha (6):
dt-bindings: net: dsa: microchip: Add KSZ8463 switch support
net: dsa: microchip: Add KSZ8463 switch support to KSZ DSA driver
net: dsa: microchip: Transform register for use with KSZ8463
net: dsa: microchip: Use different registers for KSZ8463
net: dsa: microchip: Write switch MAC address differently for KSZ8463
net: dsa: microchip: Setup fiber ports for KSZ8463

 .../bindings/net/dsa/microchip,ksz.yaml       |   1 +
 drivers/net/dsa/microchip/ksz8.c              | 201 +++++++++++++++---
 drivers/net/dsa/microchip/ksz8.h              |   4 +
 drivers/net/dsa/microchip/ksz8_reg.h          |  49 +++++
 drivers/net/dsa/microchip/ksz_common.c        | 168 ++++++++++++++-
 drivers/net/dsa/microchip/ksz_common.h        | 104 +++++++--
 drivers/net/dsa/microchip/ksz_dcb.c           |  10 +-
 drivers/net/dsa/microchip/ksz_spi.c           |  14 ++
 include/linux/platform_data/microchip-ksz.h   |   1 +
 9 files changed, 503 insertions(+), 49 deletions(-)
Signature · 1 line
2.34.1
#1 lore raw

PATCH v2 1/6 dt-bindings: net: dsa: microchip: Add KSZ8463 switch support

From: Tristram Ha <[email protected]>

KSZ8463 switch is a 3-port switch based from KSZ8863. Its register access is significantly different from the other KSZ SPI switches.

---
 Documentation/devicetree/bindings/net/dsa/microchip,ksz.yaml | 1 +
 1 file changed, 1 insertion(+)
Documentation/devicetree/bindings/net/dsa/microchip,ksz.yaml +1 -0
index 62ca63e8a26f..eb4607460db7 100644
@@ -18,6 +18,7 @@ properties:
   # required and optional properties.
   compatible:
     enum:
+      - microchip,ksz8463
       - microchip,ksz8765
       - microchip,ksz8794
       - microchip,ksz8795
Signature · 1 line
2.34.1
Acked-by on v2 1/6
#2 lore raw

Re: [PATCH net-next 1/6 v2] dt-bindings: net: dsa: microchip: Add KSZ8463 switch support

Quoted 9 lines · On Mon, Jul 07, 2025 at 08:16:43PM -0700, [email protected] wrote:

From: Tristram Ha <[email protected]>

KSZ8463 switch is a 3-port switch based from KSZ8863. Its register access is significantly different from the other KSZ SPI switches.

---
 Documentation/devicetree/bindings/net/dsa/microchip,ksz.yaml | 1 +
 1 file changed, 1 insertion(+)

Best regards,
Krzysztof

maintainer Reviewed-by on v2 1/6
#3 lore raw

Re: [PATCH net-next 1/6 v2] dt-bindings: net: dsa: microchip: Add KSZ8463 switch support

In reply to [email protected]
Quoted 6 lines · On Mon, Jul 07, 2025 at 08:16:43PM -0700, [email protected] wrote:

From: Tristram Ha <[email protected]>

KSZ8463 switch is a 3-port switch based from KSZ8863. Its register access is significantly different from the other KSZ SPI switches.

Andrew

#4 lore raw

PATCH v2 2/6 net: dsa: microchip: Add KSZ8463 switch support to KSZ DSA driver

In reply to [email protected]

From: Tristram Ha <[email protected]>

KSZ8463 switch is a 3-port switch based from KSZ8863. Its major difference from other KSZ SPI switches is its register access is not a simple continual 8-bit transfer with automatic address increase but uses a byte-enable mechanism specifying 8-bit, 16-bit, or 32-bit access. Its registers are also defined in 16-bit format because it shares a design with a MAC controller using 16-bit access. As a result some common register accesses need to be re-arranged. The 64-bit access used by other switches needs to be broken into 2 32-bit accesses.

This patch adds the basic structure for using KSZ8463. It cannot use the same regmap table for other KSZ switches as it interprets the 16-bit value as little-endian and its SPI commands are different.

KSZ8463's internal PHYs use standard PHY register definitions so there is no need to remap things. However, the hardware has a bug that the high word and low word of the PHY id are swapped. In addition the port registers are arranged differently so KSZ8463 has its own mapping for port registers and PHY registers.

---
 drivers/net/dsa/microchip/ksz8.c            |  87 +++++++++++++++
 drivers/net/dsa/microchip/ksz8.h            |   4 +
 drivers/net/dsa/microchip/ksz8_reg.h        |  49 +++++++++
 drivers/net/dsa/microchip/ksz_common.c      | 114 ++++++++++++++++++++
 drivers/net/dsa/microchip/ksz_common.h      |  33 ++++++
 drivers/net/dsa/microchip/ksz_spi.c         |  14 +++
 include/linux/platform_data/microchip-ksz.h |   1 +
 7 files changed, 302 insertions(+)
drivers/net/dsa/microchip/ksz8.c +87 -0
index be433b4e2b1c..92a720ee1f71 100644
@@ -3,6 +3,7 @@
  * Microchip KSZ8XXX series switch driver
  *
  * It supports the following switches:
+ * - KSZ8463
  * - KSZ8863, KSZ8873 aka KSZ88X3
  * - KSZ8895, KSZ8864 aka KSZ8895 family
  * - KSZ8794, KSZ8795, KSZ8765 aka KSZ87XX
@@ -194,6 +195,7 @@ int ksz8_change_mtu(struct ksz_device *dev, int port, int mtu)
 	case KSZ8794_CHIP_ID:
 	case KSZ8765_CHIP_ID:
 		return ksz8795_change_mtu(dev, frame_size);
+	case KSZ8463_CHIP_ID:
 	case KSZ88X3_CHIP_ID:
 	case KSZ8864_CHIP_ID:
 	case KSZ8895_CHIP_ID:
@@ -1947,6 +1949,91 @@ u32 ksz8_get_port_addr(int port, int offset)
 	return PORT_CTRL_ADDR(port, offset);
 }
 
+u32 ksz8463_get_port_addr(int port, int offset)
+{
+	return offset + 0x18 * port;
+}
+
+static inline u16 ksz8463_get_phy_addr(u16 phy, u16 reg, u16 offset)
+{
+	return offset + reg * 2 + phy * (P2MBCR - P1MBCR);
+}
+
+int ksz8463_r_phy(struct ksz_device *dev, u16 phy, u16 reg, u16 *val)
+{
+	bool processed = true;
+	u16 sw_reg = 0;
+	u16 data;
+	int ret;
+
+	if (phy > 1)
+		return -ENOSPC;
+	switch (reg) {
+	case MII_PHYSID1:
+		sw_reg = ksz8463_get_phy_addr(phy, 0, PHY1IHR);
+		break;
+	case MII_PHYSID2:
+		sw_reg = ksz8463_get_phy_addr(phy, 0, PHY1ILR);
+		break;
+	case MII_BMCR:
+	case MII_BMSR:
+	case MII_ADVERTISE:
+	case MII_LPA:
+		sw_reg = ksz8463_get_phy_addr(phy, reg, P1MBCR);
+		break;
+
+	/* No MMD access. */
+	case MII_MMD_CTRL:
+	case MII_MMD_DATA:
+		break;
+	case MII_TPISTATUS:
+		/* This register holds the PHY interrupt status for simulated
+		 * Micrel KSZ PHY.
+		 */
+		data = 0x0505;
+		break;
+	default:
+		processed = false;
+		break;
+	}
+	if (processed && sw_reg) {
+		ret = ksz_read16(dev, sw_reg, &data);
+		if (ret)
+			return ret;
+		*val = data;
+	}
+
+	return 0;
+}
+
+int ksz8463_w_phy(struct ksz_device *dev, u16 phy, u16 reg, u16 val)
+{
+	u16 sw_reg = 0;
+	int ret;
+
+	if (phy > 1)
+		return -ENOSPC;
+
+	/* No write to fiber port. */
+	if (dev->ports[phy].fiber)
+		return 0;
+	switch (reg) {
+	case MII_BMCR:
+	case MII_ADVERTISE:
+		sw_reg = ksz8463_get_phy_addr(phy, reg, P1MBCR);
+		break;
+	default:
+		break;
+	}
+	if (sw_reg) {
+		ret = ksz_write16(dev, sw_reg, val);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
 int ksz8_switch_init(struct ksz_device *dev)
 {
 	dev->cpu_port = fls(dev->info->cpu_ports) - 1;
drivers/net/dsa/microchip/ksz8.h +4 -0
index e1c79ff97123..0f2cd1474b44 100644
@@ -63,4 +63,8 @@ void ksz8_phylink_mac_link_up(struct phylink_config *config,
 			      bool tx_pause, bool rx_pause);
 int ksz8_all_queues_split(struct ksz_device *dev, int queues);
 
+u32 ksz8463_get_port_addr(int port, int offset);
+int ksz8463_r_phy(struct ksz_device *dev, u16 phy, u16 reg, u16 *val);
+int ksz8463_w_phy(struct ksz_device *dev, u16 phy, u16 reg, u16 val);
+
 #endif
drivers/net/dsa/microchip/ksz8_reg.h +49 -0

55 lines not on this page · raw message

drivers/net/dsa/microchip/ksz_common.c +114 -0

150 lines not on this page · raw message

drivers/net/dsa/microchip/ksz_common.h +33 -0
index a08417df2ca4..3ffca7054983 100644
@@ -222,6 +222,7 @@ struct ksz_device {
 
 /* List of supported models */
 enum ksz_model {
+	KSZ8463,
 	KSZ8563,
 	KSZ8567,
 	KSZ8795,
@@ -761,6 +762,7 @@ static inline bool ksz_is_sgmii_port(struct ksz_device *dev, int port)
 #define REG_CHIP_ID0			0x00
 
 #define SW_FAMILY_ID_M			GENMASK(15, 8)
+#define KSZ84_FAMILY_ID			0x84
 #define KSZ87_FAMILY_ID			0x87
 #define KSZ88_FAMILY_ID			0x88
 #define KSZ8895_FAMILY_ID		0x95
@@ -906,6 +908,9 @@ static inline bool ksz_is_sgmii_port(struct ksz_device *dev, int port)
 #define KSZ_SPI_OP_RD		3
 #define KSZ_SPI_OP_WR		2
 
+#define KSZ8463_SPI_OP_RD	0
+#define KSZ8463_SPI_OP_WR	1
+
 #define swabnot_used(x)		0
 
 #define KSZ_SPI_OP_FLAG_MASK(opcode, swp, regbits, regpad)		\
@@ -939,4 +944,32 @@ static inline bool ksz_is_sgmii_port(struct ksz_device *dev, int port)
 		[KSZ_REGMAP_32] = KSZ_REGMAP_ENTRY(32, swp, (regbits), (regpad), (regalign)), \
 	}
 
+#define KSZ8463_REGMAP_ENTRY(width, swp, regbits, regpad, regalign)	\
+	{								\
+		.name = #width,						\
+		.val_bits = (width),					\
+		.reg_stride = 1,					\
+		.reg_bits = (regbits) + (regalign),			\
+		.pad_bits = (regpad),					\
+		.max_register = BIT(regbits) - 1,			\
+		.cache_type = REGCACHE_NONE,				\
+		.read_flag_mask =					\
+			KSZ_SPI_OP_FLAG_MASK(KSZ8463_SPI_OP_RD, swp,	\
+					     regbits, regpad),		\
+		.write_flag_mask =					\
+			KSZ_SPI_OP_FLAG_MASK(KSZ8463_SPI_OP_WR, swp,	\
+					     regbits, regpad),		\
+		.lock = ksz_regmap_lock,				\
+		.unlock = ksz_regmap_unlock,				\
+		.reg_format_endian = REGMAP_ENDIAN_BIG,			\
+		.val_format_endian = REGMAP_ENDIAN_LITTLE		\
+	}
+
+#define KSZ8463_REGMAP_TABLE(ksz, swp, regbits, regpad, regalign)	\
+	static const struct regmap_config ksz##_regmap_config[] = {	\
+		[KSZ_REGMAP_8] = KSZ8463_REGMAP_ENTRY(8, swp, (regbits), (regpad), (regalign)), \
+		[KSZ_REGMAP_16] = KSZ8463_REGMAP_ENTRY(16, swp, (regbits), (regpad), (regalign)), \
+		[KSZ_REGMAP_32] = KSZ8463_REGMAP_ENTRY(32, swp, (regbits), (regpad), (regalign)), \
+	}
+
 #endif
drivers/net/dsa/microchip/ksz_spi.c +14 -0
index b633d263098c..d7ce2d1cc3f0 100644
@@ -16,6 +16,10 @@
 
 #include "ksz_common.h"
 
+#define KSZ8463_SPI_ADDR_SHIFT			13
+#define KSZ8463_SPI_ADDR_ALIGN			1
+#define KSZ8463_SPI_TURNAROUND_SHIFT		2
+
 #define KSZ8795_SPI_ADDR_SHIFT			12
 #define KSZ8795_SPI_ADDR_ALIGN			3
 #define KSZ8795_SPI_TURNAROUND_SHIFT		1
@@ -28,6 +32,9 @@
 #define KSZ9477_SPI_ADDR_ALIGN			3
 #define KSZ9477_SPI_TURNAROUND_SHIFT		5
 
+KSZ8463_REGMAP_TABLE(ksz8463, 16, KSZ8463_SPI_ADDR_SHIFT,
+		     KSZ8463_SPI_TURNAROUND_SHIFT, KSZ8463_SPI_ADDR_ALIGN);
+
 KSZ_REGMAP_TABLE(ksz8795, 16, KSZ8795_SPI_ADDR_SHIFT,
 		 KSZ8795_SPI_TURNAROUND_SHIFT, KSZ8795_SPI_ADDR_ALIGN);
 
@@ -58,6 +65,8 @@ static int ksz_spi_probe(struct spi_device *spi)
 	dev->chip_id = chip->chip_id;
 	if (chip->chip_id == KSZ88X3_CHIP_ID)
 		regmap_config = ksz8863_regmap_config;
+	else if (chip->chip_id == KSZ8463_CHIP_ID)
+		regmap_config = ksz8463_regmap_config;
 	else if (chip->chip_id == KSZ8795_CHIP_ID ||
 		 chip->chip_id == KSZ8794_CHIP_ID ||
 		 chip->chip_id == KSZ8765_CHIP_ID)
@@ -125,6 +134,10 @@ static void ksz_spi_shutdown(struct spi_device *spi)
 }
 
 static const struct of_device_id ksz_dt_ids[] = {
+	{
+		.compatible = "microchip,ksz8463",
+		.data = &ksz_switch_chips[KSZ8463]
+	},
 	{
 		.compatible = "microchip,ksz8765",
 		.data = &ksz_switch_chips[KSZ8765]
@@ -214,6 +227,7 @@ static const struct of_device_id ksz_dt_ids[] = {
 MODULE_DEVICE_TABLE(of, ksz_dt_ids);
 
 static const struct spi_device_id ksz_spi_ids[] = {
+	{ "ksz8463" },
 	{ "ksz8765" },
 	{ "ksz8794" },
 	{ "ksz8795" },
include/linux/platform_data/microchip-ksz.h +1 -0
index 0e0e8fe6975f..028781ad4059 100644
@@ -23,6 +23,7 @@
 #include <linux/platform_data/dsa.h>
 
 enum ksz_chip_id {
+	KSZ8463_CHIP_ID = 0x8463,
 	KSZ8563_CHIP_ID = 0x8563,
 	KSZ8795_CHIP_ID = 0x8795,
 	KSZ8794_CHIP_ID = 0x8794,
Signature · 1 line
2.34.1
#5 lore raw

PATCH v2 3/6 net: dsa: microchip: Transform register for use with KSZ8463

In reply to [email protected]

From: Tristram Ha <[email protected]>

KSZ8463 uses a byte-enable mechanism to specify 8-bit, 16-bit, and 32-bit access. The register is first shifted right by 2 then left by 4. Extra 4 bits are added. If the access is 8-bit one of the 4 bits is set. If the access is 16-bit two of the 4 bits are set. If the access is 32-bit all 4 bits are set. The SPI command for read or write is then added.

As there are no automatic address increase and continuous SPI transfer the 64-bit access needs to broken into 2 32-bit accesses.

All common access functions in ksz_common.h are updated to include register transformation call so that the correct register access is used for KSZ8463.

KSZ8463/KSZ8795/KSZ8863/KSZ8895 common code for switch access is updated to include the register transformation function if common access functions in ksz_common.h are not used. In addition PORT_CTRL_ADDR is replaced with the get_port_addr helper function as KSZ8463 has different port register arrangement.

---
 drivers/net/dsa/microchip/ksz8.c       | 10 ++--
 drivers/net/dsa/microchip/ksz_common.c |  5 +-
 drivers/net/dsa/microchip/ksz_common.h | 71 +++++++++++++++++++++-----
 3 files changed, 67 insertions(+), 19 deletions(-)
drivers/net/dsa/microchip/ksz8.c +6 -4
index 92a720ee1f71..f0c6049afe51 100644
@@ -36,13 +36,15 @@
 
 static void ksz_cfg(struct ksz_device *dev, u32 addr, u8 bits, bool set)
 {
-	regmap_update_bits(ksz_regmap_8(dev), addr, bits, set ? bits : 0);
+	regmap_update_bits(ksz_regmap_8(dev), reg8(dev, addr), bits,
+			   set ? bits : 0);
 }
 
 static void ksz_port_cfg(struct ksz_device *dev, int port, int offset, u8 bits,
 			 bool set)
 {
-	regmap_update_bits(ksz_regmap_8(dev), PORT_CTRL_ADDR(port, offset),
+	regmap_update_bits(ksz_regmap_8(dev),
+			   reg8(dev, dev->dev_ops->get_port_addr(port, offset)),
 			   bits, set ? bits : 0);
 }
 
@@ -1888,14 +1890,14 @@ int ksz8_setup(struct dsa_switch *ds)
 	ksz_cfg(dev, S_LINK_AGING_CTRL, SW_LINK_AUTO_AGING, true);
 
 	/* Enable aggressive back off algorithm in half duplex mode. */
-	regmap_update_bits(ksz_regmap_8(dev), REG_SW_CTRL_1,
+	regmap_update_bits(ksz_regmap_8(dev), reg8(dev, REG_SW_CTRL_1),
 			   SW_AGGR_BACKOFF, SW_AGGR_BACKOFF);
 
 	/*
 	 * Make sure unicast VLAN boundary is set as default and
 	 * enable no excessive collision drop.
 	 */
-	regmap_update_bits(ksz_regmap_8(dev), REG_SW_CTRL_2,
+	regmap_update_bits(ksz_regmap_8(dev), reg8(dev, REG_SW_CTRL_2),
 			   UNICAST_VLAN_BOUNDARY | NO_EXC_COLLISION_DROP,
 			   UNICAST_VLAN_BOUNDARY | NO_EXC_COLLISION_DROP);
 
drivers/net/dsa/microchip/ksz_common.c +3 -2
index 095e647b3897..5261747b7753 100644
@@ -2991,7 +2991,8 @@ static int ksz_setup(struct dsa_switch *ds)
 
 	ds->num_tx_queues = dev->info->num_tx_queues;
 
-	regmap_update_bits(ksz_regmap_8(dev), regs[S_MULTICAST_CTRL],
+	regmap_update_bits(ksz_regmap_8(dev),
+			   reg8(dev, regs[S_MULTICAST_CTRL]),
 			   MULTICAST_STORM_DISABLE, MULTICAST_STORM_DISABLE);
 
 	ksz_init_mib_timer(dev);
@@ -3051,7 +3052,7 @@ static int ksz_setup(struct dsa_switch *ds)
 		goto out_ptp_clock_unregister;
 
 	/* start switch */
-	regmap_update_bits(ksz_regmap_8(dev), regs[S_START_CTRL],
+	regmap_update_bits(ksz_regmap_8(dev), reg8(dev, regs[S_START_CTRL]),
 			   SW_START, SW_START);
 
 	return 0;
drivers/net/dsa/microchip/ksz_common.h +58 -13

150 lines not on this page · raw message

Signature · 1 line
2.34.1
#6 lore raw

PATCH v2 4/6 net: dsa: microchip: Use different registers for KSZ8463

In reply to [email protected]

From: Tristram Ha <[email protected]>

KSZ8463 does not use same set of registers as KSZ8863 so it is necessary to change some registers when using KSZ8463.

---
 drivers/net/dsa/microchip/ksz8.c       | 78 +++++++++++++++++++-------
 drivers/net/dsa/microchip/ksz_common.c | 35 +++++++++---
 drivers/net/dsa/microchip/ksz_dcb.c    | 10 +++-
 3 files changed, 94 insertions(+), 29 deletions(-)
drivers/net/dsa/microchip/ksz8.c +57 -21
index f0c6049afe51..904db68e11f3 100644
@@ -143,6 +143,11 @@ int ksz8_reset_switch(struct ksz_device *dev)
 			KSZ8863_GLOBAL_SOFTWARE_RESET | KSZ8863_PCS_RESET, true);
 		ksz_cfg(dev, KSZ8863_REG_SW_RESET,
 			KSZ8863_GLOBAL_SOFTWARE_RESET | KSZ8863_PCS_RESET, false);
+	} else if (ksz_is_ksz8463(dev)) {
+		ksz_cfg(dev, KSZ8463_REG_SW_RESET,
+			KSZ8463_GLOBAL_SOFTWARE_RESET, true);
+		ksz_cfg(dev, KSZ8463_REG_SW_RESET,
+			KSZ8463_GLOBAL_SOFTWARE_RESET, false);
 	} else {
 		/* reset switch */
 		ksz_write8(dev, REG_POWER_MANAGEMENT_1,
@@ -231,6 +236,11 @@ static int ksz8_port_queue_split(struct ksz_device *dev, int port, int queues)
 			       WEIGHTED_FAIR_QUEUE_ENABLE);
 		if (ret)
 			return ret;
+	} else if (ksz_is_ksz8463(dev)) {
+		mask_4q = KSZ8873_PORT_4QUEUE_SPLIT_EN;
+		mask_2q = KSZ8873_PORT_2QUEUE_SPLIT_EN;
+		reg_4q = P1CR1;
+		reg_2q = P1CR1 + 1;
 	} else {
 		mask_4q = KSZ8795_PORT_4QUEUE_SPLIT_EN;
 		mask_2q = KSZ8795_PORT_2QUEUE_SPLIT_EN;
@@ -1269,12 +1279,15 @@ int ksz8_w_phy(struct ksz_device *dev, u16 phy, u16 reg, u16 val)
 
 void ksz8_cfg_port_member(struct ksz_device *dev, int port, u8 member)
 {
+	int offset = P_MIRROR_CTRL;
 	u8 data;
 
-	ksz_pread8(dev, port, P_MIRROR_CTRL, &data);
-	data &= ~PORT_VLAN_MEMBERSHIP;
+	if (ksz_is_ksz8463(dev))
+		offset = P1CR2;
+	ksz_pread8(dev, port, offset, &data);
+	data &= ~dev->port_mask;
 	data |= (member & dev->port_mask);
-	ksz_pwrite8(dev, port, P_MIRROR_CTRL, data);
+	ksz_pwrite8(dev, port, offset, data);
 }
 
 void ksz8_flush_dyn_mac_table(struct ksz_device *dev, int port)
@@ -1282,6 +1295,8 @@ void ksz8_flush_dyn_mac_table(struct ksz_device *dev, int port)
 	u8 learn[DSA_MAX_PORTS];
 	int first, index, cnt;
 	const u16 *regs;
+	int reg = S_FLUSH_TABLE_CTRL;
+	int mask = SW_FLUSH_DYN_MAC_TABLE;
 
 	regs = dev->info->regs;
 
@@ -1299,7 +1314,11 @@ void ksz8_flush_dyn_mac_table(struct ksz_device *dev, int port)
 			ksz_pwrite8(dev, index, regs[P_STP_CTRL],
 				    learn[index] | PORT_LEARN_DISABLE);
 	}
-	ksz_cfg(dev, S_FLUSH_TABLE_CTRL, SW_FLUSH_DYN_MAC_TABLE, true);
+	if (ksz_is_ksz8463(dev)) {
+		reg = KSZ8463_FLUSH_TABLE_CTRL;
+		mask = KSZ8463_FLUSH_DYN_MAC_TABLE;
+	}
+	ksz_cfg(dev, reg, mask, true);
 	for (index = first; index < cnt; index++) {
 		if (!(learn[index] & PORT_LEARN_DISABLE))
 			ksz_pwrite8(dev, index, regs[P_STP_CTRL], learn[index]);
@@ -1438,7 +1457,7 @@ int ksz8_fdb_del(struct ksz_device *dev, int port, const unsigned char *addr,
 int ksz8_port_vlan_filtering(struct ksz_device *dev, int port, bool flag,
 			     struct netlink_ext_ack *extack)
 {
-	if (ksz_is_ksz88x3(dev))
+	if (ksz_is_ksz88x3(dev) || ksz_is_ksz8463(dev))
 		return -ENOTSUPP;
 
 	/* Discard packets with VID not enabled on the switch */
@@ -1454,9 +1473,12 @@ int ksz8_port_vlan_filtering(struct ksz_device *dev, int port, bool flag,
 
 static void ksz8_port_enable_pvid(struct ksz_device *dev, int port, bool state)
 {
-	if (ksz_is_ksz88x3(dev)) {
-		ksz_cfg(dev, REG_SW_INSERT_SRC_PVID,
-			0x03 << (4 - 2 * port), state);
+	if (ksz_is_ksz88x3(dev) || ksz_is_ksz8463(dev)) {
+		int reg = REG_SW_INSERT_SRC_PVID;
+
+		if (ksz_is_ksz8463(dev))
+			reg = KSZ8463_REG_SW_CTRL_9;
+		ksz_cfg(dev, reg, 0x03 << (4 - 2 * port), state);
 	} else {
 		ksz_pwrite8(dev, port, REG_PORT_CTRL_12, state ? 0x0f : 0x00);
 	}
@@ -1471,7 +1493,7 @@ int ksz8_port_vlan_add(struct ksz_device *dev, int port,
 	u16 data, new_pvid = 0;
 	u8 fid, member, valid;
 
-	if (ksz_is_ksz88x3(dev))
+	if (ksz_is_ksz88x3(dev) || ksz_is_ksz8463(dev))
 		return -ENOTSUPP;
 
 	/* If a VLAN is added with untagged flag different from the
@@ -1540,7 +1562,7 @@ int ksz8_port_vlan_del(struct ksz_device *dev, int port,
 	u16 data, pvid;
 	u8 fid, member, valid;
 
-	if (ksz_is_ksz88x3(dev))
+	if (ksz_is_ksz88x3(dev) || ksz_is_ksz8463(dev))
 		return -ENOTSUPP;
 
 	ksz_pread16(dev, port, REG_PORT_CTRL_VID, &pvid);
@@ -1570,19 +1592,23 @@ int ksz8_port_mirror_add(struct ksz_device *dev, int port,
 			 struct dsa_mall_mirror_tc_entry *mirror,
 			 bool ingress, struct netlink_ext_ack *extack)
 {
+	int offset = P_MIRROR_CTRL;
+
+	if (ksz_is_ksz8463(dev))
+		offset = P1CR2;
 	if (ingress) {
-		ksz_port_cfg(dev, port, P_MIRROR_CTRL, PORT_MIRROR_RX, true);
+		ksz_port_cfg(dev, port, offset, PORT_MIRROR_RX, true);
 		dev->mirror_rx |= BIT(port);
 	} else {
-		ksz_port_cfg(dev, port, P_MIRROR_CTRL, PORT_MIRROR_TX, true);
+		ksz_port_cfg(dev, port, offset, PORT_MIRROR_TX, true);
 		dev->mirror_tx |= BIT(port);
 	}
 
-	ksz_port_cfg(dev, port, P_MIRROR_CTRL, PORT_MIRROR_SNIFFER, false);
+	ksz_port_cfg(dev, port, offset, PORT_MIRROR_SNIFFER, false);
 
 	/* configure mirror port */
 	if (dev->mirror_rx || dev->mirror_tx)
-		ksz_port_cfg(dev, mirror->to_local_port, P_MIRROR_CTRL,
+		ksz_port_cfg(dev, mirror->to_local_port, offset,
 			     PORT_MIRROR_SNIFFER, true);
 
 	return 0;
@@ -1591,20 +1617,23 @@ int ksz8_port_mirror_add(struct ksz_device *dev, int port,
 void ksz8_port_mirror_del(struct ksz_device *dev, int port,
 			  struct dsa_mall_mirror_tc_entry *mirror)
 {
+	int offset = P_MIRROR_CTRL;
 	u8 data;
 
+	if (ksz_is_ksz8463(dev))
+		offset = P1CR2;
 	if (mirror->ingress) {
-		ksz_port_cfg(dev, port, P_MIRROR_CTRL, PORT_MIRROR_RX, false);
+		ksz_port_cfg(dev, port, offset, PORT_MIRROR_RX, false);
 		dev->mirror_rx &= ~BIT(port);
 	} else {
-		ksz_port_cfg(dev, port, P_MIRROR_CTRL, PORT_MIRROR_TX, false);
+		ksz_port_cfg(dev, port, offset, PORT_MIRROR_TX, false);
 		dev->mirror_tx &= ~BIT(port);
 	}
 
-	ksz_pread8(dev, port, P_MIRROR_CTRL, &data);
+	ksz_pread8(dev, port, offset, &data);
 
 	if (!dev->mirror_rx && !dev->mirror_tx)
-		ksz_port_cfg(dev, mirror->to_local_port, P_MIRROR_CTRL,
+		ksz_port_cfg(dev, mirror->to_local_port, offset,
 			     PORT_MIRROR_SNIFFER, false);
 }
 
@@ -1629,17 +1658,24 @@ void ksz8_port_setup(struct ksz_device *dev, int port, bool cpu_port)
 	const u16 *regs = dev->info->regs;
 	struct dsa_switch *ds = dev->ds;
 	const u32 *masks;
+	int offset;
 	u8 member;
 
 	masks = dev->info->masks;
 
 	/* enable broadcast storm limit */
-	ksz_port_cfg(dev, port, P_BCAST_STORM_CTRL, PORT_BROADCAST_STORM, true);
+	offset = P_BCAST_STORM_CTRL;
+	if (ksz_is_ksz8463(dev))
+		offset = P1CR1;
+	ksz_port_cfg(dev, port, offset, PORT_BROADCAST_STORM, true);
 
 	ksz8_port_queue_split(dev, port, dev->info->num_tx_queues);
 
 	/* replace priority */
-	ksz_port_cfg(dev, port, P_802_1P_CTRL,
+	offset = P_802_1P_CTRL;
+	if (ksz_is_ksz8463(dev))
+		offset = P1CR2;
+	ksz_port_cfg(dev, port, offset,
 		     masks[PORT_802_1P_REMAPPING], false);
 
 	if (cpu_port)
@@ -1905,7 +1941,7 @@ int ksz8_setup(struct dsa_switch *ds)
 
 	ksz_cfg(dev, S_MIRROR_CTRL, SW_MIRROR_RX_TX, false);
 
-	if (!ksz_is_ksz88x3(dev))
+	if (!ksz_is_ksz88x3(dev) && !ksz_is_ksz8463(dev))
 		ksz_cfg(dev, REG_SW_CTRL_19, SW_INS_TAG_ENABLE, true);
 
 	for (i = 0; i < (dev->info->num_vlans / 4); i++)
drivers/net/dsa/microchip/ksz_common.c +28 -7
index 5261747b7753..55108237180c 100644
@@ -2951,6 +2951,7 @@ static int ksz_parse_drive_strength(struct ksz_device *dev);
 static int ksz_setup(struct dsa_switch *ds)
 {
 	struct ksz_device *dev = ds->priv;
+	u16 storm_mask, storm_rate;
 	struct dsa_port *dp;
 	struct ksz_port *p;
 	const u16 *regs;
@@ -2980,10 +2981,15 @@ static int ksz_setup(struct dsa_switch *ds)
 	}
 
 	/* set broadcast storm protection 10% rate */
-	regmap_update_bits(ksz_regmap_16(dev), regs[S_BROADCAST_CTRL],
-			   BROADCAST_STORM_RATE,
-			   (BROADCAST_STORM_VALUE *
-			   BROADCAST_STORM_PROT_RATE) / 100);
+	storm_mask = BROADCAST_STORM_RATE;
+	storm_rate = (BROADCAST_STORM_VALUE * BROADCAST_STORM_PROT_RATE) / 100;
+	if (ksz_is_ksz8463(dev)) {
+		storm_mask = cpu_to_be16(storm_mask);
+		storm_rate = cpu_to_be16(storm_rate);
+	}
+	regmap_update_bits(ksz_regmap_16(dev),
+			   reg16(dev, regs[S_BROADCAST_CTRL]),
+			   storm_mask, storm_rate);
 
 	dev->dev_ops->config_cpu_port(ds);
 
@@ -4222,6 +4228,17 @@ static int ksz_ets_band_to_queue(struct tc_ets_qopt_offload_replace_params *p,
 	return p->bands - 1 - band;
 }
 
+static u8 ksz8463_tc_ctrl(int port, int queue)
+{
+	u8 reg;
+
+	reg = 0xC8 + port * 4;
+	reg += ((3 - queue) / 2) * 2;
+	reg++;
+	reg -= (queue & 1);
+	return reg;
+}
+
 /**
  * ksz88x3_tc_ets_add - Configure ETS (Enhanced Transmission Selection)
  *                      for a port on KSZ88x3 switch
@@ -4257,6 +4274,8 @@ static int ksz88x3_tc_ets_add(struct ksz_device *dev, int port,
 		 * port/queue
 		 */
 		reg = KSZ8873_TXQ_SPLIT_CTRL_REG(port, queue);
+		if (ksz_is_ksz8463(dev))
+			reg = ksz8463_tc_ctrl(port, queue);
 
 		/* Clear WFQ enable bit to select strict priority scheduling */
 		ret = ksz_rmw8(dev, reg, KSZ8873_TXQ_WFQ_ENABLE, 0);
@@ -4292,6 +4311,8 @@ static int ksz88x3_tc_ets_del(struct ksz_device *dev, int port)
 		 * port/queue
 		 */
 		reg = KSZ8873_TXQ_SPLIT_CTRL_REG(port, queue);
+		if (ksz_is_ksz8463(dev))
+			reg = ksz8463_tc_ctrl(port, queue);
 
 		/* Set WFQ enable bit to revert back to default scheduling
 		 * mode
@@ -4439,7 +4460,7 @@ static int ksz_tc_setup_qdisc_ets(struct dsa_switch *ds, int port,
 	struct ksz_device *dev = ds->priv;
 	int ret;
 
-	if (is_ksz8(dev) && !ksz_is_ksz88x3(dev))
+	if (is_ksz8(dev) && !(ksz_is_ksz88x3(dev) || ksz_is_ksz8463(dev)))
 		return -EOPNOTSUPP;
 
 	if (qopt->parent != TC_H_ROOT) {
@@ -4453,13 +4474,13 @@ static int ksz_tc_setup_qdisc_ets(struct dsa_switch *ds, int port,
 		if (ret)
 			return ret;
 
-		if (ksz_is_ksz88x3(dev))
+		if (ksz_is_ksz88x3(dev) || ksz_is_ksz8463(dev))
 			return ksz88x3_tc_ets_add(dev, port,
 						  &qopt->replace_params);
 		else
 			return ksz_tc_ets_add(dev, port, &qopt->replace_params);
 	case TC_ETS_DESTROY:
-		if (ksz_is_ksz88x3(dev))
+		if (ksz_is_ksz88x3(dev) || ksz_is_ksz8463(dev))
 			return ksz88x3_tc_ets_del(dev, port);
 		else
 			return ksz_tc_ets_del(dev, port);
drivers/net/dsa/microchip/ksz_dcb.c +9 -1
index c3b501997ac9..7131c5caac54 100644
@@ -16,10 +16,12 @@
  * Therefore, we define the base offset as 0x00 here to align with that logic.
  */
 #define KSZ8_REG_PORT_1_CTRL_0			0x00
+#define KSZ8463_REG_PORT_1_CTRL_0		0x6C
 #define KSZ8_PORT_DIFFSERV_ENABLE		BIT(6)
 #define KSZ8_PORT_802_1P_ENABLE			BIT(5)
 #define KSZ8_PORT_BASED_PRIO_M			GENMASK(4, 3)
 
+#define KSZ8463_REG_TOS_DSCP_CTRL		0x16
 #define KSZ88X3_REG_TOS_DSCP_CTRL		0x60
 #define KSZ8765_REG_TOS_DSCP_CTRL		0x90
 
@@ -98,6 +100,8 @@ static void ksz_get_default_port_prio_reg(struct ksz_device *dev, int *reg,
 		*reg = KSZ8_REG_PORT_1_CTRL_0;
 		*mask = KSZ8_PORT_BASED_PRIO_M;
 		*shift = __bf_shf(KSZ8_PORT_BASED_PRIO_M);
+		if (ksz_is_ksz8463(dev))
+			*reg = KSZ8463_REG_PORT_1_CTRL_0;
 	} else {
 		*reg = KSZ9477_REG_PORT_MRI_MAC_CTRL;
 		*mask = KSZ9477_PORT_BASED_PRIO_M;
@@ -122,10 +126,12 @@ static void ksz_get_dscp_prio_reg(struct ksz_device *dev, int *reg,
 		*reg = KSZ8765_REG_TOS_DSCP_CTRL;
 		*per_reg = 4;
 		*mask = GENMASK(1, 0);
-	} else if (ksz_is_ksz88x3(dev)) {
+	} else if (ksz_is_ksz88x3(dev) || ksz_is_ksz8463(dev)) {
 		*reg = KSZ88X3_REG_TOS_DSCP_CTRL;
 		*per_reg = 4;
 		*mask = GENMASK(1, 0);
+		if (ksz_is_ksz8463(dev))
+			*reg = KSZ8463_REG_TOS_DSCP_CTRL;
 	} else {
 		*reg = KSZ9477_REG_DIFFSERV_PRIO_MAP;
 		*per_reg = 2;
@@ -151,6 +157,8 @@ static void ksz_get_apptrust_map_and_reg(struct ksz_device *dev,
 		*map = ksz8_apptrust_map_to_bit;
 		*reg = KSZ8_REG_PORT_1_CTRL_0;
 		*mask = KSZ8_PORT_DIFFSERV_ENABLE | KSZ8_PORT_802_1P_ENABLE;
+		if (ksz_is_ksz8463(dev))
+			*reg = KSZ8463_REG_PORT_1_CTRL_0;
 	} else {
 		*map = ksz9477_apptrust_map_to_bit;
 		*reg = KSZ9477_REG_PORT_MRI_PRIO_CTRL;
Signature · 1 line
2.34.1
maintainer
#7 lore raw

Re: [PATCH net-next 4/6 v2] net: dsa: microchip: Use different registers for KSZ8463

Quoted 4 lines · On Mon, 7 Jul 2025 20:16:46 -0700 [email protected] wrote:

From: Tristram Ha <[email protected]>

KSZ8463 does not use same set of registers as KSZ8863 so it is necessary to change some registers when using KSZ8463.

drivers/net/dsa/microchip/ksz_common.c:2987:28: warning: incorrect type in assignment (different base types)
drivers/net/dsa/microchip/ksz_common.c:2987:28: expected unsigned short [assigned] [usertype] storm_mask
drivers/net/dsa/microchip/ksz_common.c:2987:28: got restricted __be16 [usertype]
drivers/net/dsa/microchip/ksz_common.c:2988:28: warning: incorrect type in assignment (different base types)
drivers/net/dsa/microchip/ksz_common.c:2988:28: expected unsigned short [assigned] [usertype] storm_rate
drivers/net/dsa/microchip/ksz_common.c:2988:28: got restricted __be16 [usertype]

Signature · 1 line
pw-bot: cr
#8 lore raw

PATCH v2 5/6 net: dsa: microchip: Write switch MAC address differently for KSZ8463

In reply to [email protected]

From: Tristram Ha <[email protected]>

KSZ8463 uses 16-bit register definitions so it writes differently for 8-bit switch MAC address.

---
 drivers/net/dsa/microchip/ksz_common.c | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)
drivers/net/dsa/microchip/ksz_common.c +10 -1
index 55108237180c..c08e6578a0df 100644
@@ -4823,7 +4823,16 @@ int ksz_switch_macaddr_get(struct dsa_switch *ds, int port,
 
 	/* Program the switch MAC address to hardware */
 	for (i = 0; i < ETH_ALEN; i++) {
-		ret = ksz_write8(dev, regs[REG_SW_MAC_ADDR] + i, addr[i]);
+		if (ksz_is_ksz8463(dev)) {
+			u16 addr16 = ((u16)addr[i] << 8) | addr[i + 1];
+
+			ret = ksz_write16(dev, regs[REG_SW_MAC_ADDR] + i,
+					  addr16);
+			i++;
+		} else {
+			ret = ksz_write8(dev, regs[REG_SW_MAC_ADDR] + i,
+					 addr[i]);
+		}
 		if (ret)
 			goto macaddr_drop;
 	}
Signature · 1 line
2.34.1
#10 lore raw

PATCH v2 6/6 net: dsa: microchip: Setup fiber ports for KSZ8463

In reply to [email protected]

From: Tristram Ha <[email protected]>

The fiber ports in KSZ8463 cannot be detected internally, so it requires specifying that condition in the device tree. Like the one used in Micrel PHY the port link can only be read and there is no write to the PHY. The driver programs registers to operate fiber ports correctly.

The PTP function of the switch is also turned off as it may interfere the normal operation of the MAC.

---
 drivers/net/dsa/microchip/ksz8.c       | 26 ++++++++++++++++++++++++++
 drivers/net/dsa/microchip/ksz_common.c |  3 +++
 2 files changed, 29 insertions(+)
drivers/net/dsa/microchip/ksz8.c +26 -0
index 904db68e11f3..1207879ef80c 100644
@@ -1715,6 +1715,7 @@ void ksz8_config_cpu_port(struct dsa_switch *ds)
 	const u32 *masks;
 	const u16 *regs;
 	u8 remote;
+	u8 fiber_ports = 0;
 	int i;
 
 	masks = dev->info->masks;
@@ -1745,6 +1746,31 @@ void ksz8_config_cpu_port(struct dsa_switch *ds)
 		else
 			ksz_port_cfg(dev, i, regs[P_STP_CTRL],
 				     PORT_FORCE_FLOW_CTRL, false);
+		if (p->fiber)
+			fiber_ports |= (1 << i);
+	}
+	if (ksz_is_ksz8463(dev)) {
+		/* Setup fiber ports. */
+		if (fiber_ports) {
+			regmap_update_bits(ksz_regmap_16(dev),
+					   reg16(dev, KSZ8463_REG_CFG_CTRL),
+					   fiber_ports << PORT_COPPER_MODE_S,
+					   0);
+			regmap_update_bits(ksz_regmap_16(dev),
+					   reg16(dev, KSZ8463_REG_DSP_CTRL_6),
+					   COPPER_RECEIVE_ADJUSTMENT, 0);
+		}
+
+		/* Turn off PTP function as the switch's proprietary way of
+		 * handling timestamp is not supported in current Linux PTP
+		 * stack implementation.
+		 */
+		regmap_update_bits(ksz_regmap_16(dev),
+				   reg16(dev, KSZ8463_PTP_MSG_CONF1),
+				   PTP_ENABLE, 0);
+		regmap_update_bits(ksz_regmap_16(dev),
+				   reg16(dev, KSZ8463_PTP_CLK_CTRL),
+				   PTP_CLK_ENABLE, 0);
 	}
 }
 
drivers/net/dsa/microchip/ksz_common.c +3 -0
index c08e6578a0df..b3153b45ced9 100644
@@ -5441,6 +5441,9 @@ int ksz_switch_register(struct ksz_device *dev)
 						&dev->ports[port_num].interface);
 
 				ksz_parse_rgmii_delay(dev, port_num, port);
+				dev->ports[port_num].fiber =
+					of_property_read_bool(port,
+							      "micrel,fiber-mode");
 			}
 			of_node_put(ports);
 		}
Signature · 1 line
2.34.1
#11 lore raw

Re: [PATCH net-next 6/6 v2] net: dsa: microchip: Setup fiber ports for KSZ8463

Hi Tristram,

Quoted 29 lines · On Mon, 7 Jul 2025 20:16:48 -0700 <[email protected]> wrote:

From: Tristram Ha <[email protected]>

The fiber ports in KSZ8463 cannot be detected internally, so it requires specifying that condition in the device tree. Like the one used in Micrel PHY the port link can only be read and there is no write to the PHY. The driver programs registers to operate fiber ports correctly.

The PTP function of the switch is also turned off as it may interfere the normal operation of the MAC.

---
 drivers/net/dsa/microchip/ksz8.c       | 26 ++++++++++++++++++++++++++
 drivers/net/dsa/microchip/ksz_common.c |  3 +++
 2 files changed, 29 insertions(+)
drivers/net/dsa/microchip/ksz8.c +1 -0
index 904db68e11f3..1207879ef80c 100644
@@ -1715,6 +1715,7 @@ void ksz8_config_cpu_port(struct dsa_switch *ds)
 	const u32 *masks;
 	const u16 *regs;
 	u8 remote;
+	u8 fiber_ports = 0;
 	int i;
 
 	masks = dev->info->masks;

drivers/net/dsa/microchip/ksz8.c void ksz8_config_cpu_port(struct dsa_switch *ds) · line 1753

Quoted 8 lines
drivers/net/dsa/microchip/ksz8.c +5 -0
@@ -1745,6 +1746,31 @@ void ksz8_config_cpu_port(struct dsa_switch *ds)
 		else
 			ksz_port_cfg(dev, i, regs[P_STP_CTRL],
 				     PORT_FORCE_FLOW_CTRL, false);
+		if (p->fiber)
+			fiber_ports |= (1 << i);
+	}
+	if (ksz_is_ksz8463(dev)) {
+		/* Setup fiber ports. */

What does fiber port mean ? Is it 100BaseFX ? As this configuration is done only for the CPU port (it seems), looks like this mode is planned to be used as the MAC to MAC mode on the DSA conduit. So, instead of using this property maybe you should implement that as handling the "100base-x" phy-mode ?

Quoted 28 lines
drivers/net/dsa/microchip/ksz8.c +20 -0
+		if (fiber_ports) {
+			regmap_update_bits(ksz_regmap_16(dev),
+					   reg16(dev, KSZ8463_REG_CFG_CTRL),
+					   fiber_ports << PORT_COPPER_MODE_S,
+					   0);
+			regmap_update_bits(ksz_regmap_16(dev),
+					   reg16(dev, KSZ8463_REG_DSP_CTRL_6),
+					   COPPER_RECEIVE_ADJUSTMENT, 0);
+		}
+
+		/* Turn off PTP function as the switch's proprietary way of
+		 * handling timestamp is not supported in current Linux PTP
+		 * stack implementation.
+		 */
+		regmap_update_bits(ksz_regmap_16(dev),
+				   reg16(dev, KSZ8463_PTP_MSG_CONF1),
+				   PTP_ENABLE, 0);
+		regmap_update_bits(ksz_regmap_16(dev),
+				   reg16(dev, KSZ8463_PTP_CLK_CTRL),
+				   PTP_CLK_ENABLE, 0);
 	}
 }
 
drivers/net/dsa/microchip/ksz_common.c +0 -0
index c08e6578a0df..b3153b45ced9 100644

drivers/net/dsa/microchip/ksz_common.c int ksz_switch_register(struct ksz_device *dev) · line 5446

Quoted 6 lines
drivers/net/dsa/microchip/ksz_common.c +3 -0
@@ -5441,6 +5441,9 @@ int ksz_switch_register(struct ksz_device *dev)
 						&dev->ports[port_num].interface);
 
 				ksz_parse_rgmii_delay(dev, port_num, port);
+				dev->ports[port_num].fiber =
+					of_property_read_bool(port,
+							      "micrel,fiber-mode");

Shouldn't this be described in the binding ?

drivers/net/dsa/microchip/ksz_common.c int ksz_switch_register(struct ksz_device *dev) · line 5449

Quoted 3 lines
drivers/net/dsa/microchip/ksz_common.c +0 -0
 			}
 			of_node_put(ports);
 		}

Maxime

#12 lore raw

RE: [PATCH net-next 6/6 v2] net: dsa: microchip: Setup fiber ports for KSZ8463

Quoted 80 lines

Hi Tristram,

Quoted 29 lines · On Mon, 7 Jul 2025 20:16:48 -0700 <[email protected]> wrote:

From: Tristram Ha <[email protected]>

The fiber ports in KSZ8463 cannot be detected internally, so it requires specifying that condition in the device tree. Like the one used in Micrel PHY the port link can only be read and there is no write to the PHY. The driver programs registers to operate fiber ports correctly.

The PTP function of the switch is also turned off as it may interfere the normal operation of the MAC.

---
 drivers/net/dsa/microchip/ksz8.c       | 26 ++++++++++++++++++++++++++
 drivers/net/dsa/microchip/ksz_common.c |  3 +++
 2 files changed, 29 insertions(+)
drivers/net/dsa/microchip/ksz8.c +1 -0
index 904db68e11f3..1207879ef80c 100644
@@ -1715,6 +1715,7 @@ void ksz8_config_cpu_port(struct dsa_switch *ds)
      const u32 *masks;
      const u16 *regs;
      u8 remote;
+     u8 fiber_ports = 0;
      int i;
 
      masks = dev->info->masks;

drivers/net/dsa/microchip/ksz8.c void ksz8_config_cpu_port(struct dsa_switch *ds) · line 1753

Quoted 8 lines
drivers/net/dsa/microchip/ksz8.c +5 -0
@@ -1745,6 +1746,31 @@ void ksz8_config_cpu_port(struct dsa_switch *ds)
              else
                      ksz_port_cfg(dev, i, regs[P_STP_CTRL],
                                   PORT_FORCE_FLOW_CTRL, false);
+             if (p->fiber)
+                     fiber_ports |= (1 << i);
+     }
+     if (ksz_is_ksz8463(dev)) {
+             /* Setup fiber ports. */

What does fiber port mean ? Is it 100BaseFX ? As this configuration is done only for the CPU port (it seems), looks like this mode is planned to be used as the MAC to MAC mode on the DSA conduit. So, instead of using this property maybe you should implement that as handling the "100base-x" phy-mode ?

Review

Quoted 24 lines
drivers/net/dsa/microchip/ksz8.c +20 -0
+             if (fiber_ports) {
+                     regmap_update_bits(ksz_regmap_16(dev),
+                                        reg16(dev, KSZ8463_REG_CFG_CTRL),
+                                        fiber_ports << PORT_COPPER_MODE_S,
+                                        0);
+                     regmap_update_bits(ksz_regmap_16(dev),
+                                        reg16(dev, KSZ8463_REG_DSP_CTRL_6),
+                                        COPPER_RECEIVE_ADJUSTMENT, 0);
+             }
+
+             /* Turn off PTP function as the switch's proprietary way of
+              * handling timestamp is not supported in current Linux PTP
+              * stack implementation.
+              */
+             regmap_update_bits(ksz_regmap_16(dev),
+                                reg16(dev, KSZ8463_PTP_MSG_CONF1),
+                                PTP_ENABLE, 0);
+             regmap_update_bits(ksz_regmap_16(dev),
+                                reg16(dev, KSZ8463_PTP_CLK_CTRL),
+                                PTP_CLK_ENABLE, 0);
      }
 }
 
(quoted diff) +0 -0

b/drivers/net/dsa/microchip/ksz_common.c

Quoted 6 lines

index c08e6578a0df..b3153b45ced9 100644

drivers/net/dsa/microchip/ksz_common.c +0 -0
@@ -5441,6 +5441,9 @@ int ksz_switch_register(struct ksz_device *dev)
                                              &dev->ports[port_num].interface);
 
Quoted 10 lines

drivers/net/dsa/microchip/ksz_common.c int ksz_switch_register(struct ksz_device *dev) · line 5446

Quoted 4 lines
drivers/net/dsa/microchip/ksz_common.c +3 -0
                              ksz_parse_rgmii_delay(dev, port_num, port);
+                             dev->ports[port_num].fiber =
+                                     of_property_read_bool(port,
+                                                           "micrel,fiber-mode");

Shouldn't this be described in the binding ?

Quoted 3 lines
drivers/net/dsa/microchip/ksz_common.c +0 -0
                      }
                      of_node_put(ports);
              }

The "micrel,fiber-mode" is described in Documentation/devicetree/ bindings/net/micrel.txt.

Some old KSZ88XX switches have option of using fiber in a port running 100base-fx. Typically they have a register indicating that configuration and the driver just treats the port as having a PHY and reads the link status and speed as normal except there is no write to those PHY related registers. KSZ8463 does not have that option so the driver needs to be told.

#13 lore raw

Re: [PATCH net-next 6/6 v2] net: dsa: microchip: Setup fiber ports for KSZ8463

Quoted 101 lines · On Tue, 8 Jul 2025 19:45:44 +0000 <[email protected]> wrote:
Quoted 80 lines

Hi Tristram,

Quoted 29 lines · On Mon, 7 Jul 2025 20:16:48 -0700 <[email protected]> wrote:

From: Tristram Ha <[email protected]>

The fiber ports in KSZ8463 cannot be detected internally, so it requires specifying that condition in the device tree. Like the one used in Micrel PHY the port link can only be read and there is no write to the PHY. The driver programs registers to operate fiber ports correctly.

The PTP function of the switch is also turned off as it may interfere the normal operation of the MAC.

---
 drivers/net/dsa/microchip/ksz8.c       | 26 ++++++++++++++++++++++++++
 drivers/net/dsa/microchip/ksz_common.c |  3 +++
 2 files changed, 29 insertions(+)
drivers/net/dsa/microchip/ksz8.c +1 -0
index 904db68e11f3..1207879ef80c 100644
@@ -1715,6 +1715,7 @@ void ksz8_config_cpu_port(struct dsa_switch *ds)
      const u32 *masks;
      const u16 *regs;
      u8 remote;
+     u8 fiber_ports = 0;
      int i;
 
      masks = dev->info->masks;

drivers/net/dsa/microchip/ksz8.c void ksz8_config_cpu_port(struct dsa_switch *ds) · line 1753

Quoted 8 lines
drivers/net/dsa/microchip/ksz8.c +5 -0
@@ -1745,6 +1746,31 @@ void ksz8_config_cpu_port(struct dsa_switch *ds)
              else
                      ksz_port_cfg(dev, i, regs[P_STP_CTRL],
                                   PORT_FORCE_FLOW_CTRL, false);
+             if (p->fiber)
+                     fiber_ports |= (1 << i);
+     }
+     if (ksz_is_ksz8463(dev)) {
+             /* Setup fiber ports. */  

What does fiber port mean ? Is it 100BaseFX ? As this configuration is done only for the CPU port (it seems), looks like this mode is planned to be used as the MAC to MAC mode on the DSA conduit. So, instead of using this property maybe you should implement that as handling the "100base-x" phy-mode ?

Review

Quoted 24 lines
drivers/net/dsa/microchip/ksz8.c +20 -0
+             if (fiber_ports) {
+                     regmap_update_bits(ksz_regmap_16(dev),
+                                        reg16(dev, KSZ8463_REG_CFG_CTRL),
+                                        fiber_ports << PORT_COPPER_MODE_S,
+                                        0);
+                     regmap_update_bits(ksz_regmap_16(dev),
+                                        reg16(dev, KSZ8463_REG_DSP_CTRL_6),
+                                        COPPER_RECEIVE_ADJUSTMENT, 0);
+             }
+
+             /* Turn off PTP function as the switch's proprietary way of
+              * handling timestamp is not supported in current Linux PTP
+              * stack implementation.
+              */
+             regmap_update_bits(ksz_regmap_16(dev),
+                                reg16(dev, KSZ8463_PTP_MSG_CONF1),
+                                PTP_ENABLE, 0);
+             regmap_update_bits(ksz_regmap_16(dev),
+                                reg16(dev, KSZ8463_PTP_CLK_CTRL),
+                                PTP_CLK_ENABLE, 0);
      }
 }
 
(quoted diff) +0 -0

b/drivers/net/dsa/microchip/ksz_common.c

Quoted 6 lines

index c08e6578a0df..b3153b45ced9 100644

drivers/net/dsa/microchip/ksz_common.c +0 -0
@@ -5441,6 +5441,9 @@ int ksz_switch_register(struct ksz_device *dev)
                                              &dev->ports[port_num].interface);
 
Quoted 10 lines

drivers/net/dsa/microchip/ksz_common.c int ksz_switch_register(struct ksz_device *dev) · line 5446

Quoted 4 lines
drivers/net/dsa/microchip/ksz_common.c +3 -0
                              ksz_parse_rgmii_delay(dev, port_num, port);
+                             dev->ports[port_num].fiber =
+                                     of_property_read_bool(port,
+                                                           "micrel,fiber-mode");  

Shouldn't this be described in the binding ?

Quoted 3 lines
drivers/net/dsa/microchip/ksz_common.c +0 -0
                      }
                      of_node_put(ports);
              }  

The "micrel,fiber-mode" is described in Documentation/devicetree/ bindings/net/micrel.txt.

Some old KSZ88XX switches have option of using fiber in a port running 100base-fx. Typically they have a register indicating that configuration and the driver just treats the port as having a PHY and reads the link status and speed as normal except there is no write to those PHY related registers. KSZ8463 does not have that option so the driver needs to be told.

#14 lore raw

Re: [PATCH net-next 6/6 v2] net: dsa: microchip: Setup fiber ports for KSZ8463

In reply to [email protected]

Hi Tristram

Quoted 87 lines · On Tue, 8 Jul 2025 19:45:44 +0000 <[email protected]> wrote:
Quoted 87 lines

Hi Tristram,

Quoted 29 lines · On Mon, 7 Jul 2025 20:16:48 -0700 <[email protected]> wrote:

From: Tristram Ha <[email protected]>

The fiber ports in KSZ8463 cannot be detected internally, so it requires specifying that condition in the device tree. Like the one used in Micrel PHY the port link can only be read and there is no write to the PHY. The driver programs registers to operate fiber ports correctly.

The PTP function of the switch is also turned off as it may interfere the normal operation of the MAC.

---
 drivers/net/dsa/microchip/ksz8.c       | 26 ++++++++++++++++++++++++++
 drivers/net/dsa/microchip/ksz_common.c |  3 +++
 2 files changed, 29 insertions(+)
drivers/net/dsa/microchip/ksz8.c +1 -0
index 904db68e11f3..1207879ef80c 100644
@@ -1715,6 +1715,7 @@ void ksz8_config_cpu_port(struct dsa_switch *ds)
      const u32 *masks;
      const u16 *regs;
      u8 remote;
+     u8 fiber_ports = 0;
      int i;
 
      masks = dev->info->masks;

drivers/net/dsa/microchip/ksz8.c void ksz8_config_cpu_port(struct dsa_switch *ds) · line 1753

Quoted 8 lines
drivers/net/dsa/microchip/ksz8.c +5 -0
@@ -1745,6 +1746,31 @@ void ksz8_config_cpu_port(struct dsa_switch *ds)
              else
                      ksz_port_cfg(dev, i, regs[P_STP_CTRL],
                                   PORT_FORCE_FLOW_CTRL, false);
+             if (p->fiber)
+                     fiber_ports |= (1 << i);
+     }
+     if (ksz_is_ksz8463(dev)) {
+             /* Setup fiber ports. */  

What does fiber port mean ? Is it 100BaseFX ? As this configuration is done only for the CPU port (it seems), looks like this mode is planned to be used as the MAC to MAC mode on the DSA conduit. So, instead of using this property maybe you should implement that as handling the "100base-x" phy-mode ?

Review

Quoted 24 lines
drivers/net/dsa/microchip/ksz8.c +20 -0
+             if (fiber_ports) {
+                     regmap_update_bits(ksz_regmap_16(dev),
+                                        reg16(dev, KSZ8463_REG_CFG_CTRL),
+                                        fiber_ports << PORT_COPPER_MODE_S,
+                                        0);
+                     regmap_update_bits(ksz_regmap_16(dev),
+                                        reg16(dev, KSZ8463_REG_DSP_CTRL_6),
+                                        COPPER_RECEIVE_ADJUSTMENT, 0);
+             }
+
+             /* Turn off PTP function as the switch's proprietary way of
+              * handling timestamp is not supported in current Linux PTP
+              * stack implementation.
+              */
+             regmap_update_bits(ksz_regmap_16(dev),
+                                reg16(dev, KSZ8463_PTP_MSG_CONF1),
+                                PTP_ENABLE, 0);
+             regmap_update_bits(ksz_regmap_16(dev),
+                                reg16(dev, KSZ8463_PTP_CLK_CTRL),
+                                PTP_CLK_ENABLE, 0);
      }
 }
 
(quoted diff) +0 -0

b/drivers/net/dsa/microchip/ksz_common.c

drivers/net/dsa/microchip/ksz_common.c int ksz_switch_register(struct ksz_device *dev) · line 5446

Quoted 10 lines

index c08e6578a0df..b3153b45ced9 100644

drivers/net/dsa/microchip/ksz_common.c +3 -0
@@ -5441,6 +5441,9 @@ int ksz_switch_register(struct ksz_device *dev)
                                              &dev->ports[port_num].interface);
 
                              ksz_parse_rgmii_delay(dev, port_num, port);
+                             dev->ports[port_num].fiber =
+                                     of_property_read_bool(port,
+                                                           "micrel,fiber-mode");  

Shouldn't this be described in the binding ?

Quoted 6 lines
Quoted 3 lines

drivers/net/dsa/microchip/ksz_common.c int ksz_switch_register(struct ksz_device *dev) · line 5449

Quoted 3 lines
drivers/net/dsa/microchip/ksz_common.c +0 -0
                      }
                      of_node_put(ports);
              }  

The "micrel,fiber-mode" is described in Documentation/devicetree/ bindings/net/micrel.txt.

Yes but that's for PHYs right ? Yours is under the DSA "ports" node.

Quoted 7 lines

Some old KSZ88XX switches have option of using fiber in a port running 100base-fx. Typically they have a register indicating that configuration and the driver just treats the port as having a PHY and reads the link status and speed as normal except there is no write to those PHY related registers. KSZ8463 does not have that option so the driver needs to be told.

That's what I understood from your comments indeed, what thew me off guard is that all ports's fiber mode is configured in the config_cpu_port() callback.

I'd like to one day be able to deprecate these micrel,fiber-mode/ti,fiber-mode properties in favor of the ports API that's being worked on, but I guess we can roll with it for now.

Maxime

#15 lore raw

RE: [PATCH net-next 6/6 v2] net: dsa: microchip: Setup fiber ports for KSZ8463

Quoted 110 lines

Hi Tristram

Quoted 87 lines · On Tue, 8 Jul 2025 19:45:44 +0000 <[email protected]> wrote:
Quoted 87 lines

Hi Tristram,

Quoted 29 lines · On Mon, 7 Jul 2025 20:16:48 -0700 <[email protected]> wrote:

From: Tristram Ha <[email protected]>

The fiber ports in KSZ8463 cannot be detected internally, so it requires specifying that condition in the device tree. Like the one used in Micrel PHY the port link can only be read and there is no write to the PHY. The driver programs registers to operate fiber ports correctly.

The PTP function of the switch is also turned off as it may interfere the normal operation of the MAC.

---
 drivers/net/dsa/microchip/ksz8.c       | 26 ++++++++++++++++++++++++++
 drivers/net/dsa/microchip/ksz_common.c |  3 +++
 2 files changed, 29 insertions(+)
drivers/net/dsa/microchip/ksz8.c +1 -0
index 904db68e11f3..1207879ef80c 100644
@@ -1715,6 +1715,7 @@ void ksz8_config_cpu_port(struct dsa_switch *ds)
      const u32 *masks;
      const u16 *regs;
      u8 remote;
+     u8 fiber_ports = 0;
      int i;
 
      masks = dev->info->masks;

drivers/net/dsa/microchip/ksz8.c void ksz8_config_cpu_port(struct dsa_switch *ds) · line 1753

Quoted 8 lines
drivers/net/dsa/microchip/ksz8.c +5 -0
@@ -1745,6 +1746,31 @@ void ksz8_config_cpu_port(struct dsa_switch *ds)
              else
                      ksz_port_cfg(dev, i, regs[P_STP_CTRL],
                                   PORT_FORCE_FLOW_CTRL, false);
+             if (p->fiber)
+                     fiber_ports |= (1 << i);
+     }
+     if (ksz_is_ksz8463(dev)) {
+             /* Setup fiber ports. */

What does fiber port mean ? Is it 100BaseFX ? As this configuration is done only for the CPU port (it seems), looks like this mode is planned to be used as the MAC to MAC mode on the DSA conduit. So, instead of using this property maybe you should implement that as handling the "100base-x" phy-mode ?

Review

Quoted 24 lines
drivers/net/dsa/microchip/ksz8.c +20 -0
+             if (fiber_ports) {
+                     regmap_update_bits(ksz_regmap_16(dev),
+                                        reg16(dev, KSZ8463_REG_CFG_CTRL),
+                                        fiber_ports << PORT_COPPER_MODE_S,
+                                        0);
+                     regmap_update_bits(ksz_regmap_16(dev),
+                                        reg16(dev, KSZ8463_REG_DSP_CTRL_6),
+                                        COPPER_RECEIVE_ADJUSTMENT, 0);
+             }
+
+             /* Turn off PTP function as the switch's proprietary way of
+              * handling timestamp is not supported in current Linux PTP
+              * stack implementation.
+              */
+             regmap_update_bits(ksz_regmap_16(dev),
+                                reg16(dev, KSZ8463_PTP_MSG_CONF1),
+                                PTP_ENABLE, 0);
+             regmap_update_bits(ksz_regmap_16(dev),
+                                reg16(dev, KSZ8463_PTP_CLK_CTRL),
+                                PTP_CLK_ENABLE, 0);
      }
 }
 
(quoted diff) +0 -0

b/drivers/net/dsa/microchip/ksz_common.c

drivers/net/dsa/microchip/ksz_common.c int ksz_switch_register(struct ksz_device *dev) · line 5446

Quoted 10 lines

index c08e6578a0df..b3153b45ced9 100644

drivers/net/dsa/microchip/ksz_common.c +3 -0
@@ -5441,6 +5441,9 @@ int ksz_switch_register(struct ksz_device *dev)
                                              &dev->ports[port_num].interface);
 
                              ksz_parse_rgmii_delay(dev, port_num, port);
+                             dev->ports[port_num].fiber =
+                                     of_property_read_bool(port,
+                                                           "micrel,fiber-mode");

Shouldn't this be described in the binding ?

Quoted 6 lines
Quoted 3 lines

drivers/net/dsa/microchip/ksz_common.c int ksz_switch_register(struct ksz_device *dev) · line 5449

Quoted 3 lines
drivers/net/dsa/microchip/ksz_common.c +0 -0
                      }
                      of_node_put(ports);
              }

The "micrel,fiber-mode" is described in Documentation/devicetree/ bindings/net/micrel.txt.

Yes but that's for PHYs right ? Yours is under the DSA "ports" node.

Quoted 7 lines

Some old KSZ88XX switches have option of using fiber in a port running 100base-fx. Typically they have a register indicating that configuration and the driver just treats the port as having a PHY and reads the link status and speed as normal except there is no write to those PHY related registers. KSZ8463 does not have that option so the driver needs to be told.

Quoted 7 lines

That's what I understood from your comments indeed, what thew me off guard is that all ports's fiber mode is configured in the config_cpu_port() callback.

I'd like to one day be able to deprecate these micrel,fiber-mode/ti,fiber-mode properties in favor of the ports API that's being worked on, but I guess we can roll with it for now.

Is that required to declare the parameter in the device tree document? It is just a bit of hassle to create another one just for KSZ8463. The other old switches likely do not need this. And the new KSZ9477 and LAN937X switches use the SFP case logic.

maintainer
#16 lore raw

Re: [PATCH net-next 6/6 v2] net: dsa: microchip: Setup fiber ports for KSZ8463

In reply to [email protected]
Quoted 9 lines · On Mon, Jul 07, 2025 at 08:16:48PM -0700, [email protected] wrote:

From: Tristram Ha <[email protected]>

The fiber ports in KSZ8463 cannot be detected internally, so it requires specifying that condition in the device tree. Like the one used in Micrel PHY the port link can only be read and there is no write to the PHY. The driver programs registers to operate fiber ports correctly.

The PTP function of the switch is also turned off as it may interfere the normal operation of the MAC.

Is this PTP problem anything to do with fibre?

It seems like this should be a patch of its own, unless it does have something to do with fibre.

Andrew

#17 lore raw

RE: [PATCH net-next 6/6 v2] net: dsa: microchip: Setup fiber ports for KSZ8463

Quoted 5 lines
Quoted 5 lines

The fiber ports in KSZ8463 cannot be detected internally, so it requires specifying that condition in the device tree. Like the one used in Micrel PHY the port link can only be read and there is no write to the PHY. The driver programs registers to operate fiber ports correctly.

Quoted 7 lines
Quoted 2 lines

The PTP function of the switch is also turned off as it may interfere the normal operation of the MAC.

Is this PTP problem anything to do with fibre?

It seems like this should be a patch of its own, unless it does have something to do with fibre.

This is kind of the initial setup operation for KSZ8463. I can break it up into 2 patches.