netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH net-next v11 0/5] net: dsa: yt921x: Add support for Motorcomm YT921x
@ 2025-09-22 13:11 David Yang
  2025-09-22 13:11 ` [PATCH net-next v11 1/5] dt-bindings: ethernet-phy: add reverse SGMII phy interface type David Yang
                   ` (4 more replies)
  0 siblings, 5 replies; 19+ messages in thread
From: David Yang @ 2025-09-22 13:11 UTC (permalink / raw)
  To: netdev
  Cc: David Yang, Andrew Lunn, Vladimir Oltean, David S. Miller,
	Eric Dumazet, Jakub Kicinski, Paolo Abeni, Rob Herring,
	Krzysztof Kozlowski, Conor Dooley, Heiner Kallweit, Russell King,
	Simon Horman, devicetree, linux-kernel

Motorcomm YT921x is a series of ethernet switches developed by Shanghai
Motorcomm Electronic Technology, including:

  - YT9215S / YT9215RB / YT9215SC: 5 GbE phys
  - YT9213NB / YT9214NB: 2 GbE phys
  - YT9218N / YT9218MB: 8 GbE phys

and up to 2 serdes interfaces.

This patch adds basic support for a working DSA switch.

v10: https://lore.kernel.org/r/20250919094234.1491638-1-mmyangfl@gmail.com
  - fix warnings related to PHY_INTERFACE_MODE_REVSGMII
v9: https://lore.kernel.org/r/20250913044404.63641-1-mmyangfl@gmail.com
  - add PHY_INTERFACE_MODE_REVSGMII
  - remove mdio_verify()
  - remove uncessary fdb flush opeartions
  - rework mib reading
  - set port pvid by port_set_pvid()
v8: https://lore.kernel.org/r/20250912024620.4032846-1-mmyangfl@gmail.com
  - rework register polling
  - rework mib reading
  - other suggested code style changes
v7: https://lore.kernel.org/r/20250905181728.3169479-1-mmyangfl@gmail.com
  - simplify locking scheme
v6: https://lore.kernel.org/r/20250824005116.2434998-1-mmyangfl@gmail.com
  - handle unforwarded packets in tag driver
  - move register and struct definitions to header file
  - rework register abstraction and implement a driver lock
  - implement *_stats and use a periodic work to fetch MIB
  - remove EEPROM dump
  - remove sysfs attr and other debug leftovers
  - remove ds->user_mii_bus assignment
  - run selftests and fix any errors found
v5: https://lore.kernel.org/r/20250820075420.1601068-1-mmyangfl@gmail.com
  - use enum for reg in dt binding
  - fix phylink_mac_ops in the driver
  - fix coding style
v4: https://lore.kernel.org/r/20250818162445.1317670-1-mmyangfl@gmail.com
  - remove switchid from dt binding
  - remove hsr from tag driver
  - use ratelimited log in tag driver
v3: https://lore.kernel.org/r/20250816052323.360788-1-mmyangfl@gmail.com
  - fix words and warnings in dt binding
  - remove unnecessary dev_warn_ratelimited and u64_from_u32
  - remove lag and mst
  - check for mdio results and fix a unlocked write in conduit_state_change
v2: https://lore.kernel.org/r/20250814065032.3766988-1-mmyangfl@gmail.com
  - fix words in dt binding
  - add support for lag and mst
v1: https://lore.kernel.org/r/20250808173808.273774-1-mmyangfl@gmail.com
  - fix coding style
  - add dt binding
  - add support for fdb, vlan and bridge

David Yang (5):
  dt-bindings: ethernet-phy: add reverse SGMII phy interface type
  net: phy: introduce PHY_INTERFACE_MODE_REVSGMII
  dt-bindings: net: dsa: yt921x: Add Motorcomm YT921x switch support
  net: dsa: tag_yt921x: add support for Motorcomm YT921x tags
  net: dsa: yt921x: Add support for Motorcomm YT921x

 .../bindings/net/dsa/motorcomm,yt921x.yaml    |  169 +
 .../bindings/net/ethernet-controller.yaml     |    1 +
 drivers/net/dsa/Kconfig                       |    7 +
 drivers/net/dsa/Makefile                      |    1 +
 drivers/net/dsa/yt921x.c                      | 2900 +++++++++++++++++
 drivers/net/dsa/yt921x.h                      |  505 +++
 drivers/net/phy/phy-core.c                    |    1 +
 drivers/net/phy/phy_caps.c                    |    1 +
 drivers/net/phy/phylink.c                     |    1 +
 include/linux/phy.h                           |    4 +
 include/net/dsa.h                             |    2 +
 include/uapi/linux/if_ether.h                 |    1 +
 net/dsa/Kconfig                               |    6 +
 net/dsa/Makefile                              |    1 +
 net/dsa/tag_yt921x.c                          |  141 +
 15 files changed, 3741 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/net/dsa/motorcomm,yt921x.yaml
 create mode 100644 drivers/net/dsa/yt921x.c
 create mode 100644 drivers/net/dsa/yt921x.h
 create mode 100644 net/dsa/tag_yt921x.c

-- 
2.51.0


^ permalink raw reply	[flat|nested] 19+ messages in thread

* [PATCH net-next v11 1/5] dt-bindings: ethernet-phy: add reverse SGMII phy interface type
  2025-09-22 13:11 [PATCH net-next v11 0/5] net: dsa: yt921x: Add support for Motorcomm YT921x David Yang
@ 2025-09-22 13:11 ` David Yang
  2025-09-23 13:46   ` Rob Herring (Arm)
  2025-09-22 13:11 ` [PATCH net-next v11 2/5] net: phy: introduce PHY_INTERFACE_MODE_REVSGMII David Yang
                   ` (3 subsequent siblings)
  4 siblings, 1 reply; 19+ messages in thread
From: David Yang @ 2025-09-22 13:11 UTC (permalink / raw)
  To: netdev
  Cc: David Yang, Andrew Lunn, Vladimir Oltean, David S. Miller,
	Eric Dumazet, Jakub Kicinski, Paolo Abeni, Rob Herring,
	Krzysztof Kozlowski, Conor Dooley, Heiner Kallweit, Russell King,
	Simon Horman, devicetree, linux-kernel

The "reverse SGMII" protocol name is a personal invention, derived from
"reverse MII" and "reverse RMII", this means: "behave like an SGMII
PHY".

Signed-off-by: David Yang <mmyangfl@gmail.com>
---
 Documentation/devicetree/bindings/net/ethernet-controller.yaml | 1 +
 1 file changed, 1 insertion(+)

diff --git a/Documentation/devicetree/bindings/net/ethernet-controller.yaml b/Documentation/devicetree/bindings/net/ethernet-controller.yaml
index 2c924d296a8f..8f190fe2208a 100644
--- a/Documentation/devicetree/bindings/net/ethernet-controller.yaml
+++ b/Documentation/devicetree/bindings/net/ethernet-controller.yaml
@@ -42,6 +42,7 @@ properties:
       - mii-lite
       - gmii
       - sgmii
+      - rev-sgmii
       - psgmii
       - qsgmii
       - qusgmii
-- 
2.51.0


^ permalink raw reply related	[flat|nested] 19+ messages in thread

* [PATCH net-next v11 2/5] net: phy: introduce PHY_INTERFACE_MODE_REVSGMII
  2025-09-22 13:11 [PATCH net-next v11 0/5] net: dsa: yt921x: Add support for Motorcomm YT921x David Yang
  2025-09-22 13:11 ` [PATCH net-next v11 1/5] dt-bindings: ethernet-phy: add reverse SGMII phy interface type David Yang
@ 2025-09-22 13:11 ` David Yang
  2025-09-23 14:32   ` Maxime Chevallier
  2025-09-24 17:50   ` Russell King (Oracle)
  2025-09-22 13:11 ` [PATCH net-next v11 3/5] dt-bindings: net: dsa: yt921x: Add Motorcomm YT921x switch support David Yang
                   ` (2 subsequent siblings)
  4 siblings, 2 replies; 19+ messages in thread
From: David Yang @ 2025-09-22 13:11 UTC (permalink / raw)
  To: netdev
  Cc: David Yang, Andrew Lunn, Vladimir Oltean, David S. Miller,
	Eric Dumazet, Jakub Kicinski, Paolo Abeni, Rob Herring,
	Krzysztof Kozlowski, Conor Dooley, Heiner Kallweit, Russell King,
	Simon Horman, devicetree, linux-kernel

The "reverse SGMII" protocol name is a personal invention, derived from
"reverse MII" and "reverse RMII", this means: "behave like an SGMII
PHY".

Signed-off-by: David Yang <mmyangfl@gmail.com>
---
 drivers/net/phy/phy-core.c | 1 +
 drivers/net/phy/phy_caps.c | 1 +
 drivers/net/phy/phylink.c  | 1 +
 include/linux/phy.h        | 4 ++++
 4 files changed, 7 insertions(+)

diff --git a/drivers/net/phy/phy-core.c b/drivers/net/phy/phy-core.c
index 605ca20ae192..074645840cd5 100644
--- a/drivers/net/phy/phy-core.c
+++ b/drivers/net/phy/phy-core.c
@@ -132,6 +132,7 @@ int phy_interface_num_ports(phy_interface_t interface)
 	case PHY_INTERFACE_MODE_TRGMII:
 	case PHY_INTERFACE_MODE_USXGMII:
 	case PHY_INTERFACE_MODE_SGMII:
+	case PHY_INTERFACE_MODE_REVSGMII:
 	case PHY_INTERFACE_MODE_SMII:
 	case PHY_INTERFACE_MODE_1000BASEX:
 	case PHY_INTERFACE_MODE_2500BASEX:
diff --git a/drivers/net/phy/phy_caps.c b/drivers/net/phy/phy_caps.c
index 2cc9ee97e867..9a9a8afc056f 100644
--- a/drivers/net/phy/phy_caps.c
+++ b/drivers/net/phy/phy_caps.c
@@ -299,6 +299,7 @@ unsigned long phy_caps_from_interface(phy_interface_t interface)
 	case PHY_INTERFACE_MODE_PSGMII:
 	case PHY_INTERFACE_MODE_QSGMII:
 	case PHY_INTERFACE_MODE_QUSGMII:
+	case PHY_INTERFACE_MODE_REVSGMII:
 	case PHY_INTERFACE_MODE_SGMII:
 	case PHY_INTERFACE_MODE_GMII:
 		link_caps |= BIT(LINK_CAPA_1000HD) | BIT(LINK_CAPA_1000FD);
diff --git a/drivers/net/phy/phylink.c b/drivers/net/phy/phylink.c
index 1b06805f1bd7..e8e237fb9d35 100644
--- a/drivers/net/phy/phylink.c
+++ b/drivers/net/phy/phylink.c
@@ -255,6 +255,7 @@ static int phylink_interface_max_speed(phy_interface_t interface)
 	case PHY_INTERFACE_MODE_PSGMII:
 	case PHY_INTERFACE_MODE_QSGMII:
 	case PHY_INTERFACE_MODE_QUSGMII:
+	case PHY_INTERFACE_MODE_REVSGMII:
 	case PHY_INTERFACE_MODE_SGMII:
 	case PHY_INTERFACE_MODE_GMII:
 		return SPEED_1000;
diff --git a/include/linux/phy.h b/include/linux/phy.h
index 7da9e19471c9..42d5c1f4d8ad 100644
--- a/include/linux/phy.h
+++ b/include/linux/phy.h
@@ -107,6 +107,7 @@ extern const int phy_basic_ports_array[3];
  * @PHY_INTERFACE_MODE_LAUI: 50 Gigabit Attachment Unit Interface
  * @PHY_INTERFACE_MODE_100GBASEP: 100GBase-P - with Clause 134 FEC
  * @PHY_INTERFACE_MODE_MIILITE: MII-Lite - MII without RXER TXER CRS COL
+ * @PHY_INTERFACE_MODE_REVSGMII: Serial gigabit media-independent interface in PHY role
  * @PHY_INTERFACE_MODE_MAX: Book keeping
  *
  * Describes the interface between the MAC and PHY.
@@ -152,6 +153,7 @@ typedef enum {
 	PHY_INTERFACE_MODE_LAUI,
 	PHY_INTERFACE_MODE_100GBASEP,
 	PHY_INTERFACE_MODE_MIILITE,
+	PHY_INTERFACE_MODE_REVSGMII,
 	PHY_INTERFACE_MODE_MAX,
 } phy_interface_t;
 
@@ -281,6 +283,8 @@ static inline const char *phy_modes(phy_interface_t interface)
 		return "100gbase-p";
 	case PHY_INTERFACE_MODE_MIILITE:
 		return "mii-lite";
+	case PHY_INTERFACE_MODE_REVSGMII:
+		return "rev-sgmii";
 	default:
 		return "unknown";
 	}
-- 
2.51.0


^ permalink raw reply related	[flat|nested] 19+ messages in thread

* [PATCH net-next v11 3/5] dt-bindings: net: dsa: yt921x: Add Motorcomm YT921x switch support
  2025-09-22 13:11 [PATCH net-next v11 0/5] net: dsa: yt921x: Add support for Motorcomm YT921x David Yang
  2025-09-22 13:11 ` [PATCH net-next v11 1/5] dt-bindings: ethernet-phy: add reverse SGMII phy interface type David Yang
  2025-09-22 13:11 ` [PATCH net-next v11 2/5] net: phy: introduce PHY_INTERFACE_MODE_REVSGMII David Yang
@ 2025-09-22 13:11 ` David Yang
  2025-09-22 13:11 ` [PATCH net-next v11 4/5] net: dsa: tag_yt921x: add support for Motorcomm YT921x tags David Yang
  2025-09-22 13:11 ` [PATCH net-next v11 5/5] net: dsa: yt921x: Add support for Motorcomm YT921x David Yang
  4 siblings, 0 replies; 19+ messages in thread
From: David Yang @ 2025-09-22 13:11 UTC (permalink / raw)
  To: netdev
  Cc: David Yang, Andrew Lunn, Vladimir Oltean, David S. Miller,
	Eric Dumazet, Jakub Kicinski, Paolo Abeni, Rob Herring,
	Krzysztof Kozlowski, Conor Dooley, Heiner Kallweit, Russell King,
	Simon Horman, devicetree, linux-kernel, Krzysztof Kozlowski

The Motorcomm YT921x series is a family of Ethernet switches with up to
8 internal GbE PHYs and up to 2 GMACs.

Signed-off-by: David Yang <mmyangfl@gmail.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
---
 .../bindings/net/dsa/motorcomm,yt921x.yaml    | 169 ++++++++++++++++++
 1 file changed, 169 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/net/dsa/motorcomm,yt921x.yaml

diff --git a/Documentation/devicetree/bindings/net/dsa/motorcomm,yt921x.yaml b/Documentation/devicetree/bindings/net/dsa/motorcomm,yt921x.yaml
new file mode 100644
index 000000000000..ff03bff0be4f
--- /dev/null
+++ b/Documentation/devicetree/bindings/net/dsa/motorcomm,yt921x.yaml
@@ -0,0 +1,169 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/net/dsa/motorcomm,yt921x.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Motorcomm YT921x Ethernet switch family
+
+maintainers:
+  - David Yang <mmyangfl@gmail.com>
+
+description: |
+  The Motorcomm YT921x series is a family of Ethernet switches with up to 8
+  internal GbE PHYs and up to 2 GMACs, including:
+
+    - YT9215S / YT9215RB / YT9215SC: 5 GbE PHYs (Port 0-4) + 2 GMACs (Port 8-9)
+    - YT9213NB: 2 GbE PHYs (Port 1/3) + 1 GMAC (Port 9)
+    - YT9214NB: 2 GbE PHYs (Port 1/3) + 2 GMACs (Port 8-9)
+    - YT9218N: 8 GbE PHYs (Port 0-7)
+    - YT9218MB: 8 GbE PHYs (Port 0-7) + 2 GMACs (Port 8-9)
+
+  Any port can be used as the CPU port.
+
+properties:
+  compatible:
+    const: motorcomm,yt9215
+
+  reg:
+    enum: [0x0, 0x1d]
+
+  reset-gpios:
+    maxItems: 1
+
+  mdio:
+    $ref: /schemas/net/mdio.yaml#
+    unevaluatedProperties: false
+    description:
+      Internal MDIO bus for the internal GbE PHYs. PHY 0-7 are used for Port
+      0-7 respectively.
+
+  mdio-external:
+    $ref: /schemas/net/mdio.yaml#
+    unevaluatedProperties: false
+    description:
+      External MDIO bus to access external components. External PHYs for GMACs
+      (Port 8-9) are expected to be connected to the external MDIO bus in
+      vendor's reference design, but that is not a hard limitation from the
+      chip.
+
+required:
+  - compatible
+  - reg
+
+allOf:
+  - $ref: dsa.yaml#/$defs/ethernet-ports
+
+unevaluatedProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/gpio/gpio.h>
+
+    mdio {
+        #address-cells = <1>;
+        #size-cells = <0>;
+
+        switch@1d {
+            compatible = "motorcomm,yt9215";
+            /* default 0x1d, alternate 0x0 */
+            reg = <0x1d>;
+            reset-gpios = <&tlmm 39 GPIO_ACTIVE_LOW>;
+
+            mdio {
+                #address-cells = <1>;
+                #size-cells = <0>;
+
+                sw_phy0: phy@0 {
+                    reg = <0x0>;
+                };
+
+                sw_phy1: phy@1 {
+                    reg = <0x1>;
+                };
+
+                sw_phy2: phy@2 {
+                    reg = <0x2>;
+                };
+
+                sw_phy3: phy@3 {
+                    reg = <0x3>;
+                };
+
+                sw_phy4: phy@4 {
+                    reg = <0x4>;
+                };
+            };
+
+            mdio-external {
+                #address-cells = <1>;
+                #size-cells = <0>;
+
+                phy1: phy@b {
+                    reg = <0xb>;
+                };
+            };
+
+            ethernet-ports {
+                #address-cells = <1>;
+                #size-cells = <0>;
+
+                ethernet-port@0 {
+                    reg = <0>;
+                    label = "lan1";
+                    phy-mode = "internal";
+                    phy-handle = <&sw_phy0>;
+                };
+
+                ethernet-port@1 {
+                    reg = <1>;
+                    label = "lan2";
+                    phy-mode = "internal";
+                    phy-handle = <&sw_phy1>;
+                };
+
+                ethernet-port@2 {
+                    reg = <2>;
+                    label = "lan3";
+                    phy-mode = "internal";
+                    phy-handle = <&sw_phy2>;
+                };
+
+                ethernet-port@3 {
+                    reg = <3>;
+                    label = "lan4";
+                    phy-mode = "internal";
+                    phy-handle = <&sw_phy3>;
+                };
+
+                ethernet-port@4 {
+                    reg = <4>;
+                    label = "lan5";
+                    phy-mode = "internal";
+                    phy-handle = <&sw_phy4>;
+                };
+
+                /* CPU port */
+                ethernet-port@8 {
+                    reg = <8>;
+                    phy-mode = "rev-sgmii";
+                    ethernet = <&eth0>;
+
+                    fixed-link {
+                        speed = <1000>;
+                        full-duplex;
+                        pause;
+                        asym-pause;
+                    };
+                };
+
+                /* if external phy is connected to a MAC */
+                ethernet-port@9 {
+                    reg = <9>;
+                    label = "wan";
+                    phy-mode = "rgmii-id";
+                    phy-handle = <&phy1>;
+                };
+            };
+        };
+    };
-- 
2.51.0


^ permalink raw reply related	[flat|nested] 19+ messages in thread

* [PATCH net-next v11 4/5] net: dsa: tag_yt921x: add support for Motorcomm YT921x tags
  2025-09-22 13:11 [PATCH net-next v11 0/5] net: dsa: yt921x: Add support for Motorcomm YT921x David Yang
                   ` (2 preceding siblings ...)
  2025-09-22 13:11 ` [PATCH net-next v11 3/5] dt-bindings: net: dsa: yt921x: Add Motorcomm YT921x switch support David Yang
@ 2025-09-22 13:11 ` David Yang
  2025-09-22 13:11 ` [PATCH net-next v11 5/5] net: dsa: yt921x: Add support for Motorcomm YT921x David Yang
  4 siblings, 0 replies; 19+ messages in thread
From: David Yang @ 2025-09-22 13:11 UTC (permalink / raw)
  To: netdev
  Cc: David Yang, Andrew Lunn, Vladimir Oltean, David S. Miller,
	Eric Dumazet, Jakub Kicinski, Paolo Abeni, Rob Herring,
	Krzysztof Kozlowski, Conor Dooley, Heiner Kallweit, Russell King,
	Simon Horman, devicetree, linux-kernel

Add support for Motorcomm YT921x tags, which includes a proper
configurable ethertype field (default to 0x9988).

Signed-off-by: David Yang <mmyangfl@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
---
 include/net/dsa.h             |   2 +
 include/uapi/linux/if_ether.h |   1 +
 net/dsa/Kconfig               |   6 ++
 net/dsa/Makefile              |   1 +
 net/dsa/tag_yt921x.c          | 141 ++++++++++++++++++++++++++++++++++
 5 files changed, 151 insertions(+)
 create mode 100644 net/dsa/tag_yt921x.c

diff --git a/include/net/dsa.h b/include/net/dsa.h
index d73ea0880066..67762fdaf3c7 100644
--- a/include/net/dsa.h
+++ b/include/net/dsa.h
@@ -55,6 +55,7 @@ struct tc_action;
 #define DSA_TAG_PROTO_LAN937X_VALUE		27
 #define DSA_TAG_PROTO_VSC73XX_8021Q_VALUE	28
 #define DSA_TAG_PROTO_BRCM_LEGACY_FCS_VALUE	29
+#define DSA_TAG_PROTO_YT921X_VALUE		30
 
 enum dsa_tag_protocol {
 	DSA_TAG_PROTO_NONE		= DSA_TAG_PROTO_NONE_VALUE,
@@ -87,6 +88,7 @@ enum dsa_tag_protocol {
 	DSA_TAG_PROTO_RZN1_A5PSW	= DSA_TAG_PROTO_RZN1_A5PSW_VALUE,
 	DSA_TAG_PROTO_LAN937X		= DSA_TAG_PROTO_LAN937X_VALUE,
 	DSA_TAG_PROTO_VSC73XX_8021Q	= DSA_TAG_PROTO_VSC73XX_8021Q_VALUE,
+	DSA_TAG_PROTO_YT921X		= DSA_TAG_PROTO_YT921X_VALUE,
 };
 
 struct dsa_switch;
diff --git a/include/uapi/linux/if_ether.h b/include/uapi/linux/if_ether.h
index 69e0457eb200..cfd200c87e5e 100644
--- a/include/uapi/linux/if_ether.h
+++ b/include/uapi/linux/if_ether.h
@@ -114,6 +114,7 @@
 #define ETH_P_QINQ1	0x9100		/* deprecated QinQ VLAN [ NOT AN OFFICIALLY REGISTERED ID ] */
 #define ETH_P_QINQ2	0x9200		/* deprecated QinQ VLAN [ NOT AN OFFICIALLY REGISTERED ID ] */
 #define ETH_P_QINQ3	0x9300		/* deprecated QinQ VLAN [ NOT AN OFFICIALLY REGISTERED ID ] */
+#define ETH_P_YT921X	0x9988		/* Motorcomm YT921x DSA [ NOT AN OFFICIALLY REGISTERED ID ] */
 #define ETH_P_EDSA	0xDADA		/* Ethertype DSA [ NOT AN OFFICIALLY REGISTERED ID ] */
 #define ETH_P_DSA_8021Q	0xDADB		/* Fake VLAN Header for DSA [ NOT AN OFFICIALLY REGISTERED ID ] */
 #define ETH_P_DSA_A5PSW	0xE001		/* A5PSW Tag Value [ NOT AN OFFICIALLY REGISTERED ID ] */
diff --git a/net/dsa/Kconfig b/net/dsa/Kconfig
index 869cbe57162f..6b94028b1fcc 100644
--- a/net/dsa/Kconfig
+++ b/net/dsa/Kconfig
@@ -190,4 +190,10 @@ config NET_DSA_TAG_XRS700X
 	  Say Y or M if you want to enable support for tagging frames for
 	  Arrow SpeedChips XRS700x switches that use a single byte tag trailer.
 
+config NET_DSA_TAG_YT921X
+	tristate "Tag driver for Motorcomm YT921x switches"
+	help
+	  Say Y or M if you want to enable support for tagging frames for
+	  Motorcomm YT921x switches.
+
 endif
diff --git a/net/dsa/Makefile b/net/dsa/Makefile
index 555c07cfeb71..4b011a1d5c87 100644
--- a/net/dsa/Makefile
+++ b/net/dsa/Makefile
@@ -39,6 +39,7 @@ obj-$(CONFIG_NET_DSA_TAG_SJA1105) += tag_sja1105.o
 obj-$(CONFIG_NET_DSA_TAG_TRAILER) += tag_trailer.o
 obj-$(CONFIG_NET_DSA_TAG_VSC73XX_8021Q) += tag_vsc73xx_8021q.o
 obj-$(CONFIG_NET_DSA_TAG_XRS700X) += tag_xrs700x.o
+obj-$(CONFIG_NET_DSA_TAG_YT921X) += tag_yt921x.o
 
 # for tracing framework to find trace.h
 CFLAGS_trace.o := -I$(src)
diff --git a/net/dsa/tag_yt921x.c b/net/dsa/tag_yt921x.c
new file mode 100644
index 000000000000..995da44f0a2a
--- /dev/null
+++ b/net/dsa/tag_yt921x.c
@@ -0,0 +1,141 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Motorcomm YT921x Switch Extended CPU Port Tagging
+ *
+ * Copyright (c) 2025 David Yang <mmyangfl@gmail.com>
+ *
+ * +----+----+-------+-----+----+---------
+ * | DA | SA | TagET | Tag | ET | Payload ...
+ * +----+----+-------+-----+----+---------
+ *   6    6      2      6    2       N
+ *
+ * Tag Ethertype: CPU_TAG_TPID_TPID (default: ETH_P_YT921X = 0x9988)
+ *   * Hardcoded for the moment, but still configurable. Discuss it if there
+ *     are conflicts somewhere and/or you want to change it for some reason.
+ * Tag:
+ *   2: VLAN Tag
+ *   2: Rx Port
+ *     15b: Rx Port Valid
+ *     14b-11b: Rx Port
+ *     10b-0b: Cmd?
+ *   2: Tx Port(s)
+ *     15b: Tx Port(s) Valid
+ *     10b-0b: Tx Port(s) Mask
+ */
+
+#include <linux/etherdevice.h>
+
+#include "tag.h"
+
+#define YT921X_TAG_NAME	"yt921x"
+
+#define YT921X_TAG_LEN	8
+
+#define YT921X_TAG_PORT_EN		BIT(15)
+#define YT921X_TAG_RX_PORT_M		GENMASK(14, 11)
+#define YT921X_TAG_RX_CMD_M		GENMASK(10, 0)
+#define  YT921X_TAG_RX_CMD(x)			FIELD_PREP(YT921X_TAG_RX_CMD_M, (x))
+#define  YT921X_TAG_RX_CMD_FORWARDED		0x80
+#define  YT921X_TAG_RX_CMD_UNK_UCAST		0xb2
+#define  YT921X_TAG_RX_CMD_UNK_MCAST		0xb4
+#define YT921X_TAG_TX_PORTS_M		GENMASK(10, 0)
+#define YT921X_TAG_TX_PORTn(port)	BIT(port)
+
+static struct sk_buff *
+yt921x_tag_xmit(struct sk_buff *skb, struct net_device *netdev)
+{
+	struct dsa_port *dp = dsa_user_to_port(netdev);
+	unsigned int port = dp->index;
+	__be16 *tag;
+	u16 tx;
+
+	skb_push(skb, YT921X_TAG_LEN);
+	dsa_alloc_etype_header(skb, YT921X_TAG_LEN);
+
+	tag = dsa_etype_header_pos_tx(skb);
+
+	tag[0] = htons(ETH_P_YT921X);
+	/* VLAN tag unrelated when TX */
+	tag[1] = 0;
+	tag[2] = 0;
+	tx = YT921X_TAG_PORT_EN | YT921X_TAG_TX_PORTn(port);
+	tag[3] = htons(tx);
+
+	return skb;
+}
+
+static struct sk_buff *
+yt921x_tag_rcv(struct sk_buff *skb, struct net_device *netdev)
+{
+	unsigned int port;
+	__be16 *tag;
+	u16 cmd;
+	u16 rx;
+
+	if (unlikely(!pskb_may_pull(skb, YT921X_TAG_LEN)))
+		return NULL;
+
+	tag = dsa_etype_header_pos_rx(skb);
+
+	if (unlikely(tag[0] != htons(ETH_P_YT921X))) {
+		dev_warn_ratelimited(&netdev->dev,
+				     "Unexpected EtherType 0x%04x\n",
+				     ntohs(tag[0]));
+		return NULL;
+	}
+
+	/* Locate which port this is coming from */
+	rx = ntohs(tag[2]);
+	if (unlikely((rx & YT921X_TAG_PORT_EN) == 0)) {
+		dev_warn_ratelimited(&netdev->dev,
+				     "Unexpected rx tag 0x%04x\n", rx);
+		return NULL;
+	}
+
+	port = FIELD_GET(YT921X_TAG_RX_PORT_M, rx);
+	skb->dev = dsa_conduit_find_user(netdev, 0, port);
+	if (unlikely(!skb->dev)) {
+		dev_warn_ratelimited(&netdev->dev,
+				     "Couldn't decode source port %u\n", port);
+		return NULL;
+	}
+
+	cmd = FIELD_GET(YT921X_TAG_RX_CMD_M, rx);
+	switch (cmd) {
+	case YT921X_TAG_RX_CMD_FORWARDED:
+		/* Already forwarded by hardware */
+		dsa_default_offload_fwd_mark(skb);
+		break;
+	case YT921X_TAG_RX_CMD_UNK_UCAST:
+	case YT921X_TAG_RX_CMD_UNK_MCAST:
+		/* NOTE: hardware doesn't distinguish between TRAP (copy to CPU
+		 * only) and COPY (forward and copy to CPU). In order to perform
+		 * a soft switch, NEVER use COPY action in the switch driver.
+		 */
+		break;
+	default:
+		dev_warn_ratelimited(&netdev->dev,
+				     "Unexpected rx cmd 0x%02x\n", cmd);
+		break;
+	}
+
+	/* Remove YT921x tag and update checksum */
+	skb_pull_rcsum(skb, YT921X_TAG_LEN);
+	dsa_strip_etype_header(skb, YT921X_TAG_LEN);
+
+	return skb;
+}
+
+static const struct dsa_device_ops yt921x_netdev_ops = {
+	.name	= YT921X_TAG_NAME,
+	.proto	= DSA_TAG_PROTO_YT921X,
+	.xmit	= yt921x_tag_xmit,
+	.rcv	= yt921x_tag_rcv,
+	.needed_headroom = YT921X_TAG_LEN,
+};
+
+MODULE_DESCRIPTION("DSA tag driver for Motorcomm YT921x switches");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS_DSA_TAG_DRIVER(DSA_TAG_PROTO_YT921X, YT921X_TAG_NAME);
+
+module_dsa_tag_driver(yt921x_netdev_ops);
-- 
2.51.0


^ permalink raw reply related	[flat|nested] 19+ messages in thread

* [PATCH net-next v11 5/5] net: dsa: yt921x: Add support for Motorcomm YT921x
  2025-09-22 13:11 [PATCH net-next v11 0/5] net: dsa: yt921x: Add support for Motorcomm YT921x David Yang
                   ` (3 preceding siblings ...)
  2025-09-22 13:11 ` [PATCH net-next v11 4/5] net: dsa: tag_yt921x: add support for Motorcomm YT921x tags David Yang
@ 2025-09-22 13:11 ` David Yang
  2025-09-24  0:47   ` Jakub Kicinski
  4 siblings, 1 reply; 19+ messages in thread
From: David Yang @ 2025-09-22 13:11 UTC (permalink / raw)
  To: netdev
  Cc: David Yang, Andrew Lunn, Vladimir Oltean, David S. Miller,
	Eric Dumazet, Jakub Kicinski, Paolo Abeni, Rob Herring,
	Krzysztof Kozlowski, Conor Dooley, Heiner Kallweit, Russell King,
	Simon Horman, devicetree, linux-kernel

Motorcomm YT921x is a series of ethernet switches developed by Shanghai
Motorcomm Electronic Technology, including:

  - YT9215S / YT9215RB / YT9215SC: 5 GbE PHYs
  - YT9213NB / YT9214NB: 2 GbE PHYs
  - YT9218N / YT9218MB: 8 GbE PHYs

and up to 2 GMACs.

Driver verified on a stock wireless router with IPQ5018 + YT9215S.

Signed-off-by: David Yang <mmyangfl@gmail.com>
---
 drivers/net/dsa/Kconfig  |    7 +
 drivers/net/dsa/Makefile |    1 +
 drivers/net/dsa/yt921x.c | 2900 ++++++++++++++++++++++++++++++++++++++
 drivers/net/dsa/yt921x.h |  505 +++++++
 4 files changed, 3413 insertions(+)
 create mode 100644 drivers/net/dsa/yt921x.c
 create mode 100644 drivers/net/dsa/yt921x.h

diff --git a/drivers/net/dsa/Kconfig b/drivers/net/dsa/Kconfig
index 4d9af691b989..7eb301fd987d 100644
--- a/drivers/net/dsa/Kconfig
+++ b/drivers/net/dsa/Kconfig
@@ -154,4 +154,11 @@ config NET_DSA_VITESSE_VSC73XX_PLATFORM
 	  This enables support for the Vitesse VSC7385, VSC7388, VSC7395
 	  and VSC7398 SparX integrated ethernet switches, connected over
 	  a CPU-attached address bus and work in memory-mapped I/O mode.
+
+config NET_DSA_YT921X
+	tristate "Motorcomm YT9215 ethernet switch chip support"
+	select NET_DSA_TAG_YT921X
+	help
+	  This enables support for the Motorcomm YT9215 ethernet switch
+	  chip.
 endmenu
diff --git a/drivers/net/dsa/Makefile b/drivers/net/dsa/Makefile
index 0f8ff4a1a313..16de4ba3fa38 100644
--- a/drivers/net/dsa/Makefile
+++ b/drivers/net/dsa/Makefile
@@ -14,6 +14,7 @@ obj-$(CONFIG_NET_DSA_SMSC_LAN9303_MDIO) += lan9303_mdio.o
 obj-$(CONFIG_NET_DSA_VITESSE_VSC73XX) += vitesse-vsc73xx-core.o
 obj-$(CONFIG_NET_DSA_VITESSE_VSC73XX_PLATFORM) += vitesse-vsc73xx-platform.o
 obj-$(CONFIG_NET_DSA_VITESSE_VSC73XX_SPI) += vitesse-vsc73xx-spi.o
+obj-$(CONFIG_NET_DSA_YT921X) += yt921x.o
 obj-y				+= b53/
 obj-y				+= hirschmann/
 obj-y				+= lantiq/
diff --git a/drivers/net/dsa/yt921x.c b/drivers/net/dsa/yt921x.c
new file mode 100644
index 000000000000..1412e037823f
--- /dev/null
+++ b/drivers/net/dsa/yt921x.c
@@ -0,0 +1,2900 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Driver for Motorcomm YT921x Switch
+ *
+ * Should work on YT9213/YT9214/YT9215/YT9218, but only tested on YT9215+SGMII,
+ * be sure to do your own checks before porting to another chip.
+ *
+ * Copyright (c) 2025 David Yang
+ */
+
+#include <linux/etherdevice.h>
+#include <linux/if_bridge.h>
+#include <linux/if_hsr.h>
+#include <linux/if_vlan.h>
+#include <linux/iopoll.h>
+#include <linux/mdio.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_mdio.h>
+#include <linux/of_net.h>
+
+#include <net/dsa.h>
+
+#include "yt921x.h"
+
+struct yt921x_mib_desc {
+	unsigned int size;
+	unsigned int offset;
+	const char *name;
+	bool unstructured;
+};
+
+#define MIB_DESC(_size, _offset, _name, _unstructured) \
+	{_size, _offset, _name, _unstructured}
+
+/* Must agree with yt921x_mib */
+static const struct yt921x_mib_desc yt921x_mib_descs[] = {
+	MIB_DESC(1, 0x00, "RxBroadcast", false),	/* rx broadcast pkts */
+	MIB_DESC(1, 0x04, "RxPause", false),		/* rx pause pkts */
+	MIB_DESC(1, 0x08, "RxMulticast", false),	/* rx multicast pkts, excluding pause and OAM */
+	MIB_DESC(1, 0x0c, "RxCrcErr", false),		/* rx crc err pkts, len >= 64B */
+
+	MIB_DESC(1, 0x10, "RxAlignErr", false),		/* rx pkts with odd number of bytes */
+	MIB_DESC(1, 0x14, "RxUnderSizeErr", false),	/* rx crc ok pkts, len < 64B */
+	MIB_DESC(1, 0x18, "RxFragErr", false),		/* rx crc err pkts, len < 64B */
+	MIB_DESC(1, 0x1c, "RxPktSz64", false),		/* rx pkts, len == 64B */
+
+	MIB_DESC(1, 0x20, "RxPktSz65To127", false),	/* rx pkts, len >= 65B and <= 127B */
+	MIB_DESC(1, 0x24, "RxPktSz128To255", false),	/* rx pkts, len >= 128B and <= 255B */
+	MIB_DESC(1, 0x28, "RxPktSz256To511", false),	/* rx pkts, len >= 256B and <= 511B */
+	MIB_DESC(1, 0x2c, "RxPktSz512To1023", false),	/* rx pkts, len >= 512B and <= 1023B */
+
+	MIB_DESC(1, 0x30, "RxPktSz1024To1518", false),	/* rx pkts, len >= 1024B and <= 1518B */
+	MIB_DESC(1, 0x34, "RxPktSz1519ToMax", false),	/* rx pkts, len >= 1519B */
+	MIB_DESC(2, 0x38, "RxGoodBytes", false),	/* total bytes of rx ok pkts */
+	/* 0x3c */
+
+	MIB_DESC(2, 0x40, "RxBadBytes", true),		/* total bytes of rx err pkts */
+	/* 0x44 */
+	MIB_DESC(2, 0x48, "RxOverSzErr", false),	/* rx pkts, len > mac frame size */
+	/* 0x4c */
+
+	MIB_DESC(1, 0x50, "RxDropped", false),		/* rx dropped pkts, excluding crc err and pause */
+	MIB_DESC(1, 0x54, "TxBroadcast", false),	/* tx broadcast pkts */
+	MIB_DESC(1, 0x58, "TxPause", false),		/* tx pause pkts */
+	MIB_DESC(1, 0x5c, "TxMulticast", false),	/* tx multicast pkts, excluding pause and OAM */
+
+	MIB_DESC(1, 0x60, "TxUnderSizeErr", false),	/* tx pkts, len < 64B */
+	MIB_DESC(1, 0x64, "TxPktSz64", false),		/* tx pkts, len == 64B */
+	MIB_DESC(1, 0x68, "TxPktSz65To127", false),	/* tx pkts, len >= 65B and <= 127B */
+	MIB_DESC(1, 0x6c, "TxPktSz128To255", false),	/* tx pkts, len >= 128B and <= 255B */
+
+	MIB_DESC(1, 0x70, "TxPktSz256To511", false),	/* tx pkts, len >= 256B and <= 511B */
+	MIB_DESC(1, 0x74, "TxPktSz512To1023", false),	/* tx pkts, len >= 512B and <= 1023B */
+	MIB_DESC(1, 0x78, "TxPktSz1024To1518", false),	/* tx pkts, len >= 1024B and <= 1518B */
+	MIB_DESC(1, 0x7c, "TxPktSz1519ToMax", false),	/* tx pkts, len >= 1519B */
+
+	MIB_DESC(2, 0x80, "TxGoodBytes", false),	/* total bytes of tx ok pkts */
+	/* 0x84 */
+	MIB_DESC(2, 0x88, "TxCollision", false),	/* collisions before 64B */
+	/* 0x8c */
+
+	MIB_DESC(1, 0x90, "TxExcessiveCollistion", false),	/* aborted pkts due to too many colls */
+	MIB_DESC(1, 0x94, "TxMultipleCollision", false),	/* multiple collision for one pkt */
+	MIB_DESC(1, 0x98, "TxSingleCollision", false),	/* one collision for one pkt */
+	MIB_DESC(1, 0x9c, "TxPkt", false),		/* tx ok pkts */
+
+	MIB_DESC(1, 0xa0, "TxDeferred", false),		/* delayed pkts due to defer signal */
+	MIB_DESC(1, 0xa4, "TxLateCollision", false),	/* collisions after 64B */
+	MIB_DESC(1, 0xa8, "RxOAM", true),		/* rx OAM pkts */
+	MIB_DESC(1, 0xac, "TxOAM", true),		/* tx OAM pkts */
+};
+
+struct yt921x_info {
+	const char *name;
+	u16 major;
+	/* Unknown, seems to be plain enumeration */
+	u8 mode;
+	u8 extmode;
+	/* Ports with integral GbE PHYs, not including MCU Port 10 */
+	u16 internal_mask;
+	/* TODO: see comments in yt921x_dsa_phylink_get_caps() */
+	u16 external_mask;
+};
+
+#define YT921X_PORT_MASK_INTn(port)	BIT(port)
+#define YT921X_PORT_MASK_INT0_n(n)	GENMASK((n) - 1, 0)
+#define YT921X_PORT_MASK_EXT0		BIT(8)
+#define YT921X_PORT_MASK_EXT1		BIT(9)
+
+static const struct yt921x_info yt921x_infos[] = {
+	{
+		"YT9215SC", YT9215_MAJOR, 1, 0,
+		YT921X_PORT_MASK_INT0_n(5),
+		YT921X_PORT_MASK_EXT0 | YT921X_PORT_MASK_EXT1,
+	},
+	{
+		"YT9215S", YT9215_MAJOR, 2, 0,
+		YT921X_PORT_MASK_INT0_n(5),
+		YT921X_PORT_MASK_EXT0 | YT921X_PORT_MASK_EXT1,
+	},
+	{
+		"YT9215RB", YT9215_MAJOR, 3, 0,
+		YT921X_PORT_MASK_INT0_n(5),
+		YT921X_PORT_MASK_EXT0 | YT921X_PORT_MASK_EXT1,
+	},
+	{
+		"YT9214NB", YT9215_MAJOR, 3, 2,
+		YT921X_PORT_MASK_INTn(1) | YT921X_PORT_MASK_INTn(3),
+		YT921X_PORT_MASK_EXT0 | YT921X_PORT_MASK_EXT1,
+	},
+	{
+		"YT9213NB", YT9215_MAJOR, 3, 3,
+		YT921X_PORT_MASK_INTn(1) | YT921X_PORT_MASK_INTn(3),
+		YT921X_PORT_MASK_EXT1,
+	},
+	{
+		"YT9218N", YT9218_MAJOR, 0, 0,
+		YT921X_PORT_MASK_INT0_n(8),
+		0,
+	},
+	{
+		"YT9218MB", YT9218_MAJOR, 1, 0,
+		YT921X_PORT_MASK_INT0_n(8),
+		YT921X_PORT_MASK_EXT0 | YT921X_PORT_MASK_EXT1,
+	},
+	{}
+};
+
+#define YT921X_NAME	"yt921x"
+
+#define YT921X_VID_UNWARE	4095
+
+#define YT921X_POLL_SLEEP_US	10000
+#define YT921X_POLL_TIMEOUT_US	100000
+
+/* The interval should be small enough to avoid overflow of 32bit MIBs.
+ *
+ * Until we can read MIBs from stats64 call directly (i.e. sleep
+ * there), we have to poll stats more frequently then it is actually needed.
+ * For overflow protection, normally, 100 sec interval should have been OK.
+ */
+#define YT921X_STATS_INTERVAL_JIFFIES	(3 * HZ)
+
+struct yt921x_reg_mdio {
+	struct mii_bus *bus;
+	int addr;
+	/* SWITCH_ID_1 / SWITCH_ID_0 of the device
+	 *
+	 * This is a way to multiplex multiple devices on the same MII phyaddr
+	 * and should be configurable in DT. However, MDIO core simply doesn't
+	 * allow multiple devices over one reg addr, so this is a fixed value
+	 * for now until a solution is found.
+	 *
+	 * Keep this because we need switchid to form MII regaddrs anyway.
+	 */
+	unsigned char switchid;
+};
+
+/* TODO: SPI/I2C */
+
+#define to_yt921x_priv(_ds) container_of_const(_ds, struct yt921x_priv, ds)
+#define to_device(priv) ((priv)->ds.dev)
+
+static int yt921x_reg_read(struct yt921x_priv *priv, u32 reg, u32 *valp)
+{
+	WARN_ON(!mutex_is_locked(&priv->reg_lock));
+
+	return priv->reg_ops->read(priv->reg_ctx, reg, valp);
+}
+
+static int yt921x_reg_write(struct yt921x_priv *priv, u32 reg, u32 val)
+{
+	WARN_ON(!mutex_is_locked(&priv->reg_lock));
+
+	return priv->reg_ops->write(priv->reg_ctx, reg, val);
+}
+
+static int
+yt921x_reg_wait(struct yt921x_priv *priv, u32 reg, u32 mask, u32 *valp)
+{
+	u32 val;
+	int res;
+	int ret;
+
+	ret = read_poll_timeout(yt921x_reg_read, res,
+				res || (val & mask) == *valp,
+				YT921X_POLL_SLEEP_US, YT921X_POLL_TIMEOUT_US,
+				false, priv, reg, &val);
+	if (ret)
+		return ret;
+	if (res)
+		return res;
+
+	*valp = val;
+	return 0;
+}
+
+static int
+yt921x_reg_update_bits(struct yt921x_priv *priv, u32 reg, u32 mask, u32 val)
+{
+	int res;
+	u32 v;
+	u32 u;
+
+	res = yt921x_reg_read(priv, reg, &v);
+	if (res)
+		return res;
+
+	u = v;
+	u &= ~mask;
+	u |= val;
+	if (u == v)
+		return 0;
+
+	return yt921x_reg_write(priv, reg, u);
+}
+
+static int yt921x_reg_set_bits(struct yt921x_priv *priv, u32 reg, u32 mask)
+{
+	return yt921x_reg_update_bits(priv, reg, 0, mask);
+}
+
+static int yt921x_reg_clear_bits(struct yt921x_priv *priv, u32 reg, u32 mask)
+{
+	return yt921x_reg_update_bits(priv, reg, mask, 0);
+}
+
+static int
+yt921x_reg_toggle_bits(struct yt921x_priv *priv, u32 reg, u32 mask, bool set)
+{
+	return yt921x_reg_update_bits(priv, reg, mask, !set ? 0 : mask);
+}
+
+/* Some registers, like VLANn_CTRL, should always be written in 64-bit, even if
+ * you are to write only the lower / upper 32 bits.
+ *
+ * There is no such restriction for reading, but we still provide 64-bit read
+ * wrappers so that we always handle u64 values.
+ */
+
+static int yt921x_reg64_read(struct yt921x_priv *priv, u32 reg, u64 *valp)
+{
+	u32 lo;
+	u32 hi;
+	int res;
+
+	res = yt921x_reg_read(priv, reg, &lo);
+	if (res)
+		return res;
+	res = yt921x_reg_read(priv, reg + 4, &hi);
+	if (res)
+		return res;
+
+	*valp = ((u64)hi << 32) | lo;
+	return 0;
+}
+
+static int yt921x_reg64_write(struct yt921x_priv *priv, u32 reg, u64 val)
+{
+	int res;
+
+	res = yt921x_reg_write(priv, reg, (u32)val);
+	if (res)
+		return res;
+	return yt921x_reg_write(priv, reg + 4, (u32)(val >> 32));
+}
+
+static int
+yt921x_reg64_update_bits(struct yt921x_priv *priv, u32 reg, u64 mask, u64 val)
+{
+	int res;
+	u64 v;
+	u64 u;
+
+	res = yt921x_reg64_read(priv, reg, &v);
+	if (res)
+		return res;
+
+	u = v;
+	u &= ~mask;
+	u |= val;
+	if (u == v)
+		return 0;
+
+	return yt921x_reg64_write(priv, reg, u);
+}
+
+static int yt921x_reg64_clear_bits(struct yt921x_priv *priv, u32 reg, u64 mask)
+{
+	return yt921x_reg64_update_bits(priv, reg, mask, 0);
+}
+
+static int yt921x_reg_mdio_read(void *context, u32 reg, u32 *valp)
+{
+	struct yt921x_reg_mdio *mdio = context;
+	struct mii_bus *bus = mdio->bus;
+	int addr = mdio->addr;
+	u32 reg_addr;
+	u32 reg_data;
+	u32 val;
+	int res;
+
+	/* Hold the mdio bus lock to avoid (un)locking for 4 times */
+	mutex_lock_nested(&bus->mdio_lock, MDIO_MUTEX_NESTED);
+
+	reg_addr = YT921X_SMI_SWITCHID(mdio->switchid) | YT921X_SMI_ADDR |
+		   YT921X_SMI_READ;
+	res = __mdiobus_write(bus, addr, reg_addr, (u16)(reg >> 16));
+	if (res)
+		goto end;
+	res = __mdiobus_write(bus, addr, reg_addr, (u16)reg);
+	if (res)
+		goto end;
+
+	reg_data = YT921X_SMI_SWITCHID(mdio->switchid) | YT921X_SMI_DATA |
+		   YT921X_SMI_READ;
+	res = __mdiobus_read(bus, addr, reg_data);
+	if (res < 0)
+		goto end;
+	val = (u16)res;
+	res = __mdiobus_read(bus, addr, reg_data);
+	if (res < 0)
+		goto end;
+	val = (val << 16) | (u16)res;
+
+	*valp = val;
+	res = 0;
+
+end:
+	mutex_unlock(&bus->mdio_lock);
+	return res;
+}
+
+static int yt921x_reg_mdio_write(void *context, u32 reg, u32 val)
+{
+	struct yt921x_reg_mdio *mdio = context;
+	struct mii_bus *bus = mdio->bus;
+	int addr = mdio->addr;
+	u32 reg_addr;
+	u32 reg_data;
+	int res;
+
+	mutex_lock_nested(&bus->mdio_lock, MDIO_MUTEX_NESTED);
+
+	reg_addr = YT921X_SMI_SWITCHID(mdio->switchid) | YT921X_SMI_ADDR |
+		   YT921X_SMI_WRITE;
+	res = __mdiobus_write(bus, addr, reg_addr, (u16)(reg >> 16));
+	if (res)
+		goto end;
+	res = __mdiobus_write(bus, addr, reg_addr, (u16)reg);
+	if (res)
+		goto end;
+
+	reg_data = YT921X_SMI_SWITCHID(mdio->switchid) | YT921X_SMI_DATA |
+		   YT921X_SMI_WRITE;
+	res = __mdiobus_write(bus, addr, reg_data, (u16)(val >> 16));
+	if (res)
+		goto end;
+	res = __mdiobus_write(bus, addr, reg_data, (u16)val);
+	if (res)
+		goto end;
+
+	res = 0;
+
+end:
+	mutex_unlock(&bus->mdio_lock);
+	return res;
+}
+
+static const struct yt921x_reg_ops yt921x_reg_ops_mdio = {
+	.read = yt921x_reg_mdio_read,
+	.write = yt921x_reg_mdio_write,
+};
+
+/* TODO: SPI/I2C */
+
+static int yt921x_intif_wait(struct yt921x_priv *priv)
+{
+	u32 val = 0;
+
+	return yt921x_reg_wait(priv, YT921X_INT_MBUS_OP, YT921X_MBUS_OP_START,
+			       &val);
+}
+
+static int
+yt921x_intif_read(struct yt921x_priv *priv, int port, int reg, u16 *valp)
+{
+	struct device *dev = to_device(priv);
+	u32 mask;
+	u32 ctrl;
+	u32 val;
+	int res;
+
+	res = yt921x_intif_wait(priv);
+	if (res)
+		return res;
+
+	mask = YT921X_MBUS_CTRL_PORT_M | YT921X_MBUS_CTRL_REG_M |
+	       YT921X_MBUS_CTRL_OP_M;
+	ctrl = YT921X_MBUS_CTRL_PORT(port) | YT921X_MBUS_CTRL_REG(reg) |
+	       YT921X_MBUS_CTRL_READ;
+	res = yt921x_reg_update_bits(priv, YT921X_INT_MBUS_CTRL, mask, ctrl);
+	if (res)
+		return res;
+	res = yt921x_reg_write(priv, YT921X_INT_MBUS_OP, YT921X_MBUS_OP_START);
+	if (res)
+		return res;
+
+	res = yt921x_intif_wait(priv);
+	if (res)
+		return res;
+	res = yt921x_reg_read(priv, YT921X_INT_MBUS_DIN, &val);
+	if (res)
+		return res;
+
+	if ((u16)val != val)
+		dev_info(dev,
+			 "%s: port %d, reg 0x%x: Expected u16, got 0x%08x\n",
+			 __func__, port, reg, val);
+	*valp = (u16)val;
+	return 0;
+}
+
+static int
+yt921x_intif_write(struct yt921x_priv *priv, int port, int reg, u16 val)
+{
+	u32 mask;
+	u32 ctrl;
+	int res;
+
+	res = yt921x_intif_wait(priv);
+	if (res)
+		return res;
+
+	mask = YT921X_MBUS_CTRL_PORT_M | YT921X_MBUS_CTRL_REG_M |
+	       YT921X_MBUS_CTRL_OP_M;
+	ctrl = YT921X_MBUS_CTRL_PORT(port) | YT921X_MBUS_CTRL_REG(reg) |
+	       YT921X_MBUS_CTRL_WRITE;
+	res = yt921x_reg_update_bits(priv, YT921X_INT_MBUS_CTRL, mask, ctrl);
+	if (res)
+		return res;
+	res = yt921x_reg_write(priv, YT921X_INT_MBUS_DOUT, val);
+	if (res)
+		return res;
+	res = yt921x_reg_write(priv, YT921X_INT_MBUS_OP, YT921X_MBUS_OP_START);
+	if (res)
+		return res;
+
+	return yt921x_intif_wait(priv);
+}
+
+static int yt921x_mbus_int_read(struct mii_bus *mbus, int port, int reg)
+{
+	struct yt921x_priv *priv = mbus->priv;
+	u16 val;
+	int res;
+
+	if (port >= YT921X_PORT_NUM)
+		return U16_MAX;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_intif_read(priv, port, reg, &val);
+	mutex_unlock(&priv->reg_lock);
+
+	if (res)
+		return res;
+	return val;
+}
+
+static int
+yt921x_mbus_int_write(struct mii_bus *mbus, int port, int reg, u16 data)
+{
+	struct yt921x_priv *priv = mbus->priv;
+	int res;
+
+	if (port >= YT921X_PORT_NUM)
+		return -ENODEV;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_intif_write(priv, port, reg, data);
+	mutex_unlock(&priv->reg_lock);
+
+	return res;
+}
+
+static int
+yt921x_mbus_int_init(struct yt921x_priv *priv, struct device_node *mnp)
+{
+	struct device *dev = to_device(priv);
+	struct mii_bus *mbus;
+	int res;
+
+	mbus = devm_mdiobus_alloc(dev);
+	if (!mbus)
+		return -ENOMEM;
+
+	mbus->name = "YT921x internal MDIO bus";
+	snprintf(mbus->id, MII_BUS_ID_SIZE, "%s", dev_name(dev));
+	mbus->priv = priv;
+	mbus->read = yt921x_mbus_int_read;
+	mbus->write = yt921x_mbus_int_write;
+	mbus->parent = dev;
+	mbus->phy_mask = (u32)~GENMASK(YT921X_PORT_NUM - 1, 0);
+
+	res = devm_of_mdiobus_register(dev, mbus, mnp);
+	if (res)
+		return res;
+
+	priv->mbus_int = mbus;
+
+	return 0;
+}
+
+static int yt921x_extif_wait(struct yt921x_priv *priv)
+{
+	u32 val = 0;
+
+	return yt921x_reg_wait(priv, YT921X_EXT_MBUS_OP, YT921X_MBUS_OP_START,
+			       &val);
+}
+
+static int
+yt921x_extif_read(struct yt921x_priv *priv, int port, int reg, u16 *valp)
+{
+	struct device *dev = to_device(priv);
+	u32 mask;
+	u32 ctrl;
+	u32 val;
+	int res;
+
+	res = yt921x_extif_wait(priv);
+	if (res)
+		return res;
+
+	mask = YT921X_MBUS_CTRL_PORT_M | YT921X_MBUS_CTRL_REG_M |
+	       YT921X_MBUS_CTRL_TYPE_M | YT921X_MBUS_CTRL_OP_M;
+	ctrl = YT921X_MBUS_CTRL_PORT(port) | YT921X_MBUS_CTRL_REG(reg) |
+	       YT921X_MBUS_CTRL_TYPE_C22 | YT921X_MBUS_CTRL_READ;
+	res = yt921x_reg_update_bits(priv, YT921X_EXT_MBUS_CTRL, mask, ctrl);
+	if (res)
+		return res;
+	res = yt921x_reg_write(priv, YT921X_EXT_MBUS_OP, YT921X_MBUS_OP_START);
+	if (res)
+		return res;
+
+	res = yt921x_extif_wait(priv);
+	if (res)
+		return res;
+	res = yt921x_reg_read(priv, YT921X_EXT_MBUS_DIN, &val);
+	if (res)
+		return res;
+
+	if ((u16)val != val)
+		dev_info(dev,
+			 "%s: port %d, reg 0x%x: Expected u16, got 0x%08x\n",
+			 __func__, port, reg, val);
+	*valp = (u16)val;
+	return 0;
+}
+
+static int
+yt921x_extif_write(struct yt921x_priv *priv, int port, int reg, u16 val)
+{
+	u32 mask;
+	u32 ctrl;
+	int res;
+
+	res = yt921x_extif_wait(priv);
+	if (res)
+		return res;
+
+	mask = YT921X_MBUS_CTRL_PORT_M | YT921X_MBUS_CTRL_REG_M |
+	       YT921X_MBUS_CTRL_TYPE_M | YT921X_MBUS_CTRL_OP_M;
+	ctrl = YT921X_MBUS_CTRL_PORT(port) | YT921X_MBUS_CTRL_REG(reg) |
+	       YT921X_MBUS_CTRL_TYPE_C22 | YT921X_MBUS_CTRL_WRITE;
+	res = yt921x_reg_update_bits(priv, YT921X_EXT_MBUS_CTRL, mask, ctrl);
+	if (res)
+		return res;
+	res = yt921x_reg_write(priv, YT921X_EXT_MBUS_DOUT, val);
+	if (res)
+		return res;
+	res = yt921x_reg_write(priv, YT921X_EXT_MBUS_OP, YT921X_MBUS_OP_START);
+	if (res)
+		return res;
+
+	return yt921x_extif_wait(priv);
+}
+
+static int yt921x_mbus_ext_read(struct mii_bus *mbus, int port, int reg)
+{
+	struct yt921x_priv *priv = mbus->priv;
+	u16 val;
+	int res;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_extif_read(priv, port, reg, &val);
+	mutex_unlock(&priv->reg_lock);
+
+	if (res)
+		return res;
+	return val;
+}
+
+static int
+yt921x_mbus_ext_write(struct mii_bus *mbus, int port, int reg, u16 data)
+{
+	struct yt921x_priv *priv = mbus->priv;
+	int res;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_extif_write(priv, port, reg, data);
+	mutex_unlock(&priv->reg_lock);
+
+	return res;
+}
+
+static int
+yt921x_mbus_ext_init(struct yt921x_priv *priv, struct device_node *mnp)
+{
+	struct device *dev = to_device(priv);
+	struct mii_bus *mbus;
+	int res;
+
+	mbus = devm_mdiobus_alloc(dev);
+	if (!mbus)
+		return -ENOMEM;
+
+	mbus->name = "YT921x external MDIO bus";
+	snprintf(mbus->id, MII_BUS_ID_SIZE, "%s@ext", dev_name(dev));
+	mbus->priv = priv;
+	/* TODO: c45? */
+	mbus->read = yt921x_mbus_ext_read;
+	mbus->write = yt921x_mbus_ext_write;
+	mbus->parent = dev;
+
+	res = devm_of_mdiobus_register(dev, mbus, mnp);
+	if (res)
+		return res;
+
+	priv->mbus_ext = mbus;
+
+	return 0;
+}
+
+/* Read and handle overflow of 32bit MIBs. MIB buffer must be zeroed before. */
+static int yt921x_read_mib(struct yt921x_priv *priv, int port)
+{
+	struct yt921x_port *pp = &priv->ports[port];
+	struct device *dev = to_device(priv);
+	struct yt921x_mib *mib = &pp->mib;
+	int res = 0;
+
+	/* It's vain to keep consistency of yt921x_port::mib, since we have to
+	 * read them one by one and there is no magic operation to make a
+	 * snapshot of MIB stats.
+	 *
+	 * Our best hope is the consistency of each attribute, which can be done
+	 * without a lock.
+	 */
+	for (size_t i = 0; i < ARRAY_SIZE(yt921x_mib_descs); i++) {
+		const struct yt921x_mib_desc *desc = &yt921x_mib_descs[i];
+		u32 reg = YT921X_MIBn_DATA0(port) + desc->offset;
+		u64 *valp = &((u64 *)mib)[i];
+		u64 val = *valp;
+		u32 val0;
+		u32 val1;
+
+		res = yt921x_reg_read(priv, reg, &val0);
+		if (res)
+			break;
+
+		if (desc->size <= 1) {
+			if (val < (u32)val)
+				/* overflow */
+				val += (u64)U32_MAX + 1;
+			val &= ~U32_MAX;
+			val |= val0;
+		} else {
+			res = yt921x_reg_read(priv, reg + 4, &val1);
+			if (res)
+				break;
+			val = ((u64)val0 << 32) | val1;
+		}
+
+		WRITE_ONCE(*valp, val);
+	}
+
+	if (res) {
+		smp_wmb();
+		dev_err(dev, "Failed to %s port %d: %i\n", "read stats for",
+			port, res);
+		return res;
+	}
+
+	pp->rx_frames = mib->rx_64byte + mib->rx_65_127byte +
+			mib->rx_128_255byte + mib->rx_256_511byte +
+			mib->rx_512_1023byte + mib->rx_1024_1518byte +
+			mib->rx_jumbo;
+	pp->tx_frames = mib->tx_64byte + mib->tx_65_127byte +
+			mib->tx_128_255byte + mib->tx_256_511byte +
+			mib->tx_512_1023byte + mib->tx_1024_1518byte +
+			mib->tx_jumbo;
+
+	/* Flush all writes */
+	smp_wmb();
+	return 0;
+}
+
+static void yt921x_poll_mib(struct work_struct *work)
+{
+	struct yt921x_port *pp = container_of_const(work, struct yt921x_port,
+						    mib_read.work);
+	struct yt921x_priv *priv = (void *)(pp - pp->index) -
+				   offsetof(struct yt921x_priv, ports);
+	unsigned long delay = YT921X_STATS_INTERVAL_JIFFIES;
+	int port = pp->index;
+	int res;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_read_mib(priv, port);
+	mutex_unlock(&priv->reg_lock);
+	if (res)
+		delay *= 4;
+
+	schedule_delayed_work(&pp->mib_read, delay);
+}
+
+static void
+yt921x_dsa_get_strings(struct dsa_switch *ds, int port, u32 stringset,
+		       uint8_t *data)
+{
+	if (stringset != ETH_SS_STATS)
+		return;
+
+	for (size_t i = 0; i < ARRAY_SIZE(yt921x_mib_descs); i++) {
+		const struct yt921x_mib_desc *desc = &yt921x_mib_descs[i];
+
+		if (desc->unstructured)
+			ethtool_puts(&data, desc->name);
+	}
+}
+
+static void
+yt921x_dsa_get_ethtool_stats(struct dsa_switch *ds, int port, uint64_t *data)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	struct yt921x_port *pp = &priv->ports[port];
+	struct yt921x_mib *mib = &pp->mib;
+	size_t j;
+
+	mutex_lock(&priv->reg_lock);
+	yt921x_read_mib(priv, port);
+	mutex_unlock(&priv->reg_lock);
+
+	j = 0;
+	for (size_t i = 0; i < ARRAY_SIZE(yt921x_mib_descs); i++) {
+		const struct yt921x_mib_desc *desc = &yt921x_mib_descs[i];
+
+		if (!desc->unstructured)
+			continue;
+
+		data[j] = ((u64 *)mib)[i];
+		j++;
+	}
+}
+
+static int yt921x_dsa_get_sset_count(struct dsa_switch *ds, int port, int sset)
+{
+	int cnt = 0;
+
+	if (sset != ETH_SS_STATS)
+		return 0;
+
+	for (size_t i = 0; i < ARRAY_SIZE(yt921x_mib_descs); i++) {
+		const struct yt921x_mib_desc *desc = &yt921x_mib_descs[i];
+
+		if (desc->unstructured)
+			cnt++;
+	}
+
+	return cnt;
+}
+
+static void
+yt921x_dsa_get_eth_mac_stats(struct dsa_switch *ds, int port,
+			     struct ethtool_eth_mac_stats *mac_stats)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	struct yt921x_port *pp = &priv->ports[port];
+	struct yt921x_mib *mib = &pp->mib;
+
+	mutex_lock(&priv->reg_lock);
+	yt921x_read_mib(priv, port);
+	mutex_unlock(&priv->reg_lock);
+
+	mac_stats->FramesTransmittedOK = pp->tx_frames;
+	mac_stats->SingleCollisionFrames = mib->tx_single_collisions;
+	mac_stats->MultipleCollisionFrames = mib->tx_multiple_collisions;
+	mac_stats->FramesReceivedOK = pp->rx_frames;
+	mac_stats->FrameCheckSequenceErrors = mib->rx_crc_errors;
+	mac_stats->AlignmentErrors = mib->rx_alignment_errors;
+	mac_stats->OctetsTransmittedOK = mib->tx_good_bytes;
+	mac_stats->FramesWithDeferredXmissions = mib->tx_deferred;
+	mac_stats->LateCollisions = mib->tx_late_collisions;
+	mac_stats->FramesAbortedDueToXSColls = mib->tx_aborted_errors;
+	/* mac_stats->FramesLostDueToIntMACXmitError */
+	/* mac_stats->CarrierSenseErrors */
+	mac_stats->OctetsReceivedOK = mib->rx_good_bytes;
+	/* mac_stats->FramesLostDueToIntMACRcvError */
+	mac_stats->MulticastFramesXmittedOK = mib->tx_multicast;
+	mac_stats->BroadcastFramesXmittedOK = mib->tx_broadcast;
+	/* mac_stats->FramesWithExcessiveDeferral */
+	mac_stats->MulticastFramesReceivedOK = mib->rx_multicast;
+	mac_stats->BroadcastFramesReceivedOK = mib->rx_broadcast;
+	/* mac_stats->InRangeLengthErrors */
+	/* mac_stats->OutOfRangeLengthField */
+	mac_stats->FrameTooLongErrors = mib->rx_oversize_errors;
+}
+
+static void
+yt921x_dsa_get_eth_ctrl_stats(struct dsa_switch *ds, int port,
+			      struct ethtool_eth_ctrl_stats *ctrl_stats)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	struct yt921x_port *pp = &priv->ports[port];
+	struct yt921x_mib *mib = &pp->mib;
+
+	mutex_lock(&priv->reg_lock);
+	yt921x_read_mib(priv, port);
+	mutex_unlock(&priv->reg_lock);
+
+	ctrl_stats->MACControlFramesTransmitted = mib->tx_pause;
+	ctrl_stats->MACControlFramesReceived = mib->rx_pause;
+	/* ctrl_stats->UnsupportedOpcodesReceived */
+}
+
+static const struct ethtool_rmon_hist_range yt921x_rmon_ranges[] = {
+	{ 0, 64 },
+	{ 65, 127 },
+	{ 128, 255 },
+	{ 256, 511 },
+	{ 512, 1023 },
+	{ 1024, 1518 },
+	{ 1519, YT921X_FRAME_SIZE_MAX },
+	{}
+};
+
+static void
+yt921x_dsa_get_rmon_stats(struct dsa_switch *ds, int port,
+			  struct ethtool_rmon_stats *rmon_stats,
+			  const struct ethtool_rmon_hist_range **ranges)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	struct yt921x_port *pp = &priv->ports[port];
+	struct yt921x_mib *mib = &pp->mib;
+
+	mutex_lock(&priv->reg_lock);
+	yt921x_read_mib(priv, port);
+	mutex_unlock(&priv->reg_lock);
+
+	*ranges = yt921x_rmon_ranges;
+
+	rmon_stats->undersize_pkts = mib->rx_undersize_errors;
+	rmon_stats->oversize_pkts = mib->rx_oversize_errors;
+	rmon_stats->fragments = mib->rx_alignment_errors;
+	/* rmon_stats->jabbers */
+
+	rmon_stats->hist[0] = mib->rx_64byte;
+	rmon_stats->hist[1] = mib->rx_65_127byte;
+	rmon_stats->hist[2] = mib->rx_128_255byte;
+	rmon_stats->hist[3] = mib->rx_256_511byte;
+	rmon_stats->hist[4] = mib->rx_512_1023byte;
+	rmon_stats->hist[5] = mib->rx_1024_1518byte;
+	rmon_stats->hist[6] = mib->rx_jumbo;
+
+	rmon_stats->hist_tx[0] = mib->tx_64byte;
+	rmon_stats->hist_tx[1] = mib->tx_65_127byte;
+	rmon_stats->hist_tx[2] = mib->tx_128_255byte;
+	rmon_stats->hist_tx[3] = mib->tx_256_511byte;
+	rmon_stats->hist_tx[4] = mib->tx_512_1023byte;
+	rmon_stats->hist_tx[5] = mib->tx_1024_1518byte;
+	rmon_stats->hist_tx[6] = mib->tx_jumbo;
+}
+
+static void
+yt921x_dsa_get_stats64(struct dsa_switch *ds, int port,
+		       struct rtnl_link_stats64 *stats)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	struct yt921x_port *pp = &priv->ports[port];
+	struct yt921x_mib *mib = &pp->mib;
+
+	stats->rx_length_errors = mib->rx_undersize_errors +
+				  mib->rx_fragment_errors;
+	stats->rx_over_errors = mib->rx_oversize_errors;
+	stats->rx_crc_errors = mib->rx_crc_errors;
+	stats->rx_frame_errors = mib->rx_alignment_errors;
+	/* stats->rx_fifo_errors */
+	/* stats->rx_missed_errors */
+
+	stats->tx_aborted_errors = mib->tx_aborted_errors;
+	/* stats->tx_carrier_errors */
+	stats->tx_fifo_errors = mib->tx_undersize_errors;
+	/* stats->tx_heartbeat_errors */
+	stats->tx_window_errors = mib->tx_late_collisions;
+
+	stats->rx_packets = pp->rx_frames;
+	stats->tx_packets = pp->tx_frames;
+	stats->rx_bytes = mib->rx_good_bytes - ETH_FCS_LEN * stats->rx_packets;
+	stats->tx_bytes = mib->tx_good_bytes - ETH_FCS_LEN * stats->tx_packets;
+	stats->rx_errors = stats->rx_length_errors + stats->rx_over_errors +
+			   stats->rx_crc_errors + stats->rx_frame_errors;
+	stats->tx_errors = stats->tx_aborted_errors + stats->tx_fifo_errors +
+			   stats->tx_window_errors;
+	stats->rx_dropped = mib->rx_dropped;
+	/* stats->tx_dropped */
+	stats->multicast = mib->rx_multicast;
+	stats->collisions = mib->tx_collisions;
+}
+
+static void
+yt921x_dsa_get_pause_stats(struct dsa_switch *ds, int port,
+			   struct ethtool_pause_stats *pause_stats)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	struct yt921x_port *pp = &priv->ports[port];
+	struct yt921x_mib *mib = &pp->mib;
+
+	mutex_lock(&priv->reg_lock);
+	yt921x_read_mib(priv, port);
+	mutex_unlock(&priv->reg_lock);
+
+	pause_stats->tx_pause_frames = mib->tx_pause;
+	pause_stats->rx_pause_frames = mib->rx_pause;
+}
+
+static int
+yt921x_set_eee(struct yt921x_priv *priv, int port, struct ethtool_keee *e)
+{
+	/* Poor datasheet for EEE operations; don't ask if you are confused */
+
+	bool enable = e->eee_enabled;
+	u16 new_mask;
+	int res;
+
+	/* Enable / disable global EEE */
+	new_mask = priv->eee_ports_mask;
+	new_mask &= ~BIT(port);
+	new_mask |= !enable ? 0 : BIT(port);
+
+	if (!!new_mask != !!priv->eee_ports_mask) {
+		res = yt921x_reg_toggle_bits(priv, YT921X_PON_STRAP_FUNC,
+					     YT921X_PON_STRAP_EEE, !!new_mask);
+		if (res)
+			return res;
+		res = yt921x_reg_toggle_bits(priv, YT921X_PON_STRAP_VAL,
+					     YT921X_PON_STRAP_EEE, !!new_mask);
+		if (res)
+			return res;
+	}
+
+	priv->eee_ports_mask = new_mask;
+
+	/* Enable / disable port EEE */
+	res = yt921x_reg_toggle_bits(priv, YT921X_EEE_CTRL,
+				     YT921X_EEE_CTRL_ENn(port), enable);
+	if (res)
+		return res;
+	res = yt921x_reg_toggle_bits(priv, YT921X_EEEn_VAL(port),
+				     YT921X_EEE_VAL_DATA, enable);
+	if (res)
+		return res;
+
+	return 0;
+}
+
+static int
+yt921x_dsa_set_mac_eee(struct dsa_switch *ds, int port, struct ethtool_keee *e)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	int res;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_set_eee(priv, port, e);
+	mutex_unlock(&priv->reg_lock);
+
+	return res;
+}
+
+static int
+yt921x_dsa_port_change_mtu(struct dsa_switch *ds, int port, int new_mtu)
+{
+	/* Only serves as packet filter, since the frame size is always set to
+	 * maximum after reset
+	 */
+
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	struct dsa_port *dp = dsa_to_port(ds, port);
+	int frame_size;
+	int res;
+
+	frame_size = new_mtu + ETH_HLEN + ETH_FCS_LEN;
+	if (dsa_port_is_cpu(dp))
+		frame_size += YT921X_TAG_LEN;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_reg_update_bits(priv, YT921X_MACn_FRAME(port),
+				     YT921X_MAC_FRAME_SIZE_M,
+				     YT921X_MAC_FRAME_SIZE(frame_size));
+	mutex_unlock(&priv->reg_lock);
+
+	return res;
+}
+
+static int yt921x_dsa_port_max_mtu(struct dsa_switch *ds, int port)
+{
+	/* Only called for user ports, exclude tag len here */
+	return YT921X_FRAME_SIZE_MAX - ETH_HLEN - ETH_FCS_LEN - YT921X_TAG_LEN;
+}
+
+static int
+yt921x_mirror_del(struct yt921x_priv *priv, int port, bool ingress)
+{
+	u32 mask;
+
+	if (ingress)
+		mask = YT921X_MIRROR_IGR_PORTn(port);
+	else
+		mask = YT921X_MIRROR_EGR_PORTn(port);
+	return yt921x_reg_clear_bits(priv, YT921X_MIRROR, mask);
+}
+
+static int
+yt921x_mirror_add(struct yt921x_priv *priv, int port, bool ingress,
+		  int to_local_port, struct netlink_ext_ack *extack)
+{
+	u32 srcs;
+	u32 ctrl;
+	u32 val;
+	u32 dst;
+	int res;
+
+	if (ingress)
+		srcs = YT921X_MIRROR_IGR_PORTn(port);
+	else
+		srcs = YT921X_MIRROR_EGR_PORTn(port);
+	dst = YT921X_MIRROR_PORT(to_local_port);
+
+	res = yt921x_reg_read(priv, YT921X_MIRROR, &val);
+	if (res)
+		return res;
+
+	/* other mirror tasks & different dst port -> conflict */
+	if ((val & ~srcs & (YT921X_MIRROR_EGR_PORTS_M |
+			    YT921X_MIRROR_IGR_PORTS_M)) != 0 &&
+	    (val & YT921X_MIRROR_PORT_M) != dst) {
+		NL_SET_ERR_MSG_MOD(extack,
+				   "Sniffer port is already configured, "
+				   "delete existing rules & retry");
+		return -EBUSY;
+	}
+
+	ctrl = val & ~YT921X_MIRROR_PORT_M;
+	ctrl |= srcs;
+	ctrl |= dst;
+
+	if (ctrl == val)
+		return 0;
+
+	return yt921x_reg_write(priv, YT921X_MIRROR, ctrl);
+}
+
+static void
+yt921x_dsa_port_mirror_del(struct dsa_switch *ds, int port,
+			   struct dsa_mall_mirror_tc_entry *mirror)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	struct device *dev = to_device(priv);
+	int res;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_mirror_del(priv, port, mirror->ingress);
+	mutex_unlock(&priv->reg_lock);
+
+	if (res)
+		dev_err(dev, "Failed to %s port %d: %i\n", "unmirror",
+			port, res);
+}
+
+static int
+yt921x_dsa_port_mirror_add(struct dsa_switch *ds, int port,
+			   struct dsa_mall_mirror_tc_entry *mirror,
+			   bool ingress, struct netlink_ext_ack *extack)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	int res;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_mirror_add(priv, port, ingress,
+				mirror->to_local_port, extack);
+	mutex_unlock(&priv->reg_lock);
+
+	return res;
+}
+
+static int yt921x_fdb_wait(struct yt921x_priv *priv, u32 *valp)
+{
+	struct device *dev = to_device(priv);
+	u32 val = YT921X_FDB_RESULT_DONE;
+	int res;
+
+	res = yt921x_reg_wait(priv, YT921X_FDB_RESULT, YT921X_FDB_RESULT_DONE,
+			      &val);
+	if (res) {
+		dev_err(dev, "FDB probably stucked\n");
+		return res;
+	}
+
+	*valp = val;
+	return 0;
+}
+
+static int
+yt921x_fdb_in01(struct yt921x_priv *priv, const unsigned char *addr,
+		u16 vid, u32 ctrl1)
+{
+	u32 ctrl;
+	int res;
+
+	ctrl = (addr[0] << 24) | (addr[1] << 16) | (addr[2] << 8) | addr[3];
+	res = yt921x_reg_write(priv, YT921X_FDB_IN0, ctrl);
+	if (res)
+		return res;
+
+	ctrl = ctrl1 | YT921X_FDB_IO1_FID(vid) | (addr[4] << 8) | addr[5];
+	return yt921x_reg_write(priv, YT921X_FDB_IN1, ctrl);
+}
+
+static int
+yt921x_fdb_has(struct yt921x_priv *priv, const unsigned char *addr, u16 vid,
+	       u16 *indexp)
+{
+	u32 ctrl;
+	u32 val;
+	int res;
+
+	res = yt921x_fdb_in01(priv, addr, vid, 0);
+	if (res)
+		return res;
+
+	ctrl = 0;
+	res = yt921x_reg_write(priv, YT921X_FDB_IN2, ctrl);
+	if (res)
+		return res;
+
+	ctrl = YT921X_FDB_OP_OP_GET_ONE | YT921X_FDB_OP_START;
+	res = yt921x_reg_write(priv, YT921X_FDB_OP, ctrl);
+	if (res)
+		return res;
+
+	res = yt921x_fdb_wait(priv, &val);
+	if (res)
+		return res;
+	if ((val & YT921X_FDB_RESULT_NOTFOUND) != 0) {
+		*indexp = YT921X_FDB_NUM;
+		return 0;
+	}
+
+	*indexp = FIELD_GET(YT921X_FDB_RESULT_INDEX_M, val);
+	return 0;
+}
+
+static int
+yt921x_fdb_read(struct yt921x_priv *priv, unsigned char *addr, u16 *vidp,
+		u16 *ports_maskp, u16 *indexp, u8 *statusp)
+{
+	struct device *dev = to_device(priv);
+	u16 index;
+	u32 data0;
+	u32 data1;
+	u32 data2;
+	u32 val;
+	int res;
+
+	res = yt921x_fdb_wait(priv, &val);
+	if (res)
+		return res;
+	if ((val & YT921X_FDB_RESULT_NOTFOUND) != 0) {
+		*ports_maskp = 0;
+		return 0;
+	}
+	index = FIELD_GET(YT921X_FDB_RESULT_INDEX_M, val);
+
+	res = yt921x_reg_read(priv, YT921X_FDB_OUT1, &data1);
+	if (res)
+		return res;
+	if ((data1 & YT921X_FDB_IO1_STATUS_M) ==
+	    YT921X_FDB_IO1_STATUS_INVALID) {
+		*ports_maskp = 0;
+		return 0;
+	}
+
+	res = yt921x_reg_read(priv, YT921X_FDB_OUT0, &data0);
+	if (res)
+		return res;
+	res = yt921x_reg_read(priv, YT921X_FDB_OUT2, &data2);
+	if (res)
+		return res;
+
+	addr[0] = data0 >> 24;
+	addr[1] = data0 >> 16;
+	addr[2] = data0 >> 8;
+	addr[3] = data0;
+	addr[4] = data1 >> 8;
+	addr[5] = data1;
+	*vidp = FIELD_GET(YT921X_FDB_IO1_FID_M, data1);
+	*indexp = index;
+	*ports_maskp = FIELD_GET(YT921X_FDB_IO2_EGR_PORTS_M, data2);
+	*statusp = FIELD_GET(YT921X_FDB_IO1_STATUS_M, data1);
+
+	dev_dbg(dev,
+		"%s: index 0x%x, mac %02x:%02x:%02x:%02x:%02x:%02x, "
+		"vid %d, ports 0x%x, status %d\n",
+		__func__, *indexp, addr[0], addr[1], addr[2], addr[3],
+		addr[4], addr[5], *vidp, *ports_maskp, *statusp);
+	return 0;
+}
+
+static int
+yt921x_fdb_dump(struct yt921x_priv *priv, u16 ports_mask,
+		dsa_fdb_dump_cb_t *cb, void *data)
+{
+	unsigned char addr[ETH_ALEN];
+	u8 status;
+	u16 pmask;
+	u16 index;
+	u32 ctrl;
+	u16 vid;
+	int res;
+
+	ctrl = YT921X_FDB_OP_INDEX(0) | YT921X_FDB_OP_MODE_INDEX |
+	       YT921X_FDB_OP_OP_GET_ONE | YT921X_FDB_OP_START;
+	res = yt921x_reg_write(priv, YT921X_FDB_OP, ctrl);
+	if (res)
+		return res;
+	res = yt921x_fdb_read(priv, addr, &vid, &pmask, &index, &status);
+	if (res)
+		return res;
+	if ((pmask & ports_mask) != 0 && !is_multicast_ether_addr(addr)) {
+		res = cb(addr, vid,
+			 status == YT921X_FDB_ENTRY_STATUS_STATIC, data);
+		if (res)
+			return res;
+	}
+
+	ctrl = YT921X_FDB_IO2_EGR_PORTS(ports_mask);
+	res = yt921x_reg_write(priv, YT921X_FDB_IN2, ctrl);
+	if (res)
+		return res;
+
+	index = 0;
+	do {
+		ctrl = YT921X_FDB_OP_INDEX(index) | YT921X_FDB_OP_MODE_INDEX |
+		       YT921X_FDB_OP_NEXT_TYPE_UCAST_PORT |
+		       YT921X_FDB_OP_OP_GET_NEXT | YT921X_FDB_OP_START;
+		res = yt921x_reg_write(priv, YT921X_FDB_OP, ctrl);
+		if (res)
+			return res;
+
+		res = yt921x_fdb_read(priv, addr, &vid, &pmask, &index,
+				      &status);
+		if (res)
+			return res;
+		if (!pmask)
+			break;
+
+		if ((pmask & ports_mask) != 0 &&
+		    !is_multicast_ether_addr(addr)) {
+			res = cb(addr, vid,
+				 status == YT921X_FDB_ENTRY_STATUS_STATIC,
+				 data);
+			if (res)
+				return res;
+		}
+
+		/* Never call GET_NEXT with 4095, otherwise it will hang
+		 * forever until a reset!
+		 */
+	} while (index < YT921X_FDB_NUM - 1);
+
+	return 0;
+}
+
+static int
+yt921x_fdb_flush_raw(struct yt921x_priv *priv, u16 ports_mask, u16 vid,
+		     bool flush_static)
+{
+	u32 ctrl;
+	u32 val;
+	int res;
+
+	if (vid < 4096) {
+		ctrl = YT921X_FDB_IO1_FID(vid);
+		res = yt921x_reg_write(priv, YT921X_FDB_IN1, ctrl);
+		if (res)
+			return res;
+	}
+
+	ctrl = YT921X_FDB_IO2_EGR_PORTS(ports_mask);
+	res = yt921x_reg_write(priv, YT921X_FDB_IN2, ctrl);
+	if (res)
+		return res;
+
+	ctrl = YT921X_FDB_OP_OP_FLUSH | YT921X_FDB_OP_START;
+	if (vid >= 4096)
+		ctrl |= YT921X_FDB_OP_FLUSH_PORT;
+	else
+		ctrl |= YT921X_FDB_OP_FLUSH_PORT_VID;
+	if (flush_static)
+		ctrl |= YT921X_FDB_OP_FLUSH_STATIC;
+	res = yt921x_reg_write(priv, YT921X_FDB_OP, ctrl);
+	if (res)
+		return res;
+
+	res = yt921x_fdb_wait(priv, &val);
+	if (res)
+		return res;
+
+	return 0;
+}
+
+static int
+yt921x_fdb_flush_port(struct yt921x_priv *priv, int port, bool flush_static)
+{
+	return yt921x_fdb_flush_raw(priv, BIT(port), 4096, flush_static);
+}
+
+static int
+yt921x_fdb_add_index_in12(struct yt921x_priv *priv, u16 index, u16 ctrl1,
+			  u16 ctrl2)
+{
+	u32 ctrl;
+	u32 val;
+	int res;
+
+	res = yt921x_reg_write(priv, YT921X_FDB_IN1, ctrl1);
+	if (res)
+		return res;
+	res = yt921x_reg_write(priv, YT921X_FDB_IN2, ctrl2);
+	if (res)
+		return res;
+
+	ctrl = YT921X_FDB_OP_INDEX(index) | YT921X_FDB_OP_MODE_INDEX |
+	       YT921X_FDB_OP_OP_ADD | YT921X_FDB_OP_START;
+	res = yt921x_reg_write(priv, YT921X_FDB_OP, ctrl);
+	if (res)
+		return res;
+
+	return yt921x_fdb_wait(priv, &val);
+}
+
+static int
+yt921x_fdb_add(struct yt921x_priv *priv, const unsigned char *addr, u16 vid,
+	       u16 ports_mask)
+{
+	u32 ctrl;
+	u32 val;
+	int res;
+
+	ctrl = YT921X_FDB_IO1_STATUS_STATIC;
+	res = yt921x_fdb_in01(priv, addr, vid, ctrl);
+	if (res)
+		return res;
+
+	ctrl = YT921X_FDB_IO2_EGR_PORTS(ports_mask);
+	res = yt921x_reg_write(priv, YT921X_FDB_IN2, ctrl);
+	if (res)
+		return res;
+
+	ctrl = YT921X_FDB_OP_OP_ADD | YT921X_FDB_OP_START;
+	res = yt921x_reg_write(priv, YT921X_FDB_OP, ctrl);
+	if (res)
+		return res;
+
+	return yt921x_fdb_wait(priv, &val);
+}
+
+static int
+yt921x_fdb_leave(struct yt921x_priv *priv, const unsigned char *addr,
+		 u16 vid, u16 ports_mask)
+{
+	u16 index;
+	u32 ctrl1;
+	u32 ctrl2;
+	u32 ctrl;
+	u32 val2;
+	u32 val;
+	int res;
+
+	/* Check for presence */
+	res = yt921x_fdb_has(priv, addr, vid, &index);
+	if (res)
+		return res;
+	if (index >= YT921X_FDB_NUM)
+		return 0;
+
+	/* Check if action required */
+	res = yt921x_reg_read(priv, YT921X_FDB_OUT2, &val2);
+	if (res)
+		return res;
+
+	ctrl2 = val2 & ~YT921X_FDB_IO2_EGR_PORTS(ports_mask);
+	if (ctrl2 == val2)
+		return 0;
+	if ((ctrl2 & YT921X_FDB_IO2_EGR_PORTS_M) == 0) {
+		ctrl = YT921X_FDB_OP_OP_DEL | YT921X_FDB_OP_START;
+		res = yt921x_reg_write(priv, YT921X_FDB_OP, ctrl);
+		if (res)
+			return res;
+
+		return yt921x_fdb_wait(priv, &val);
+	}
+
+	res = yt921x_reg_read(priv, YT921X_FDB_OUT1, &ctrl1);
+	if (res)
+		return res;
+
+	return yt921x_fdb_add_index_in12(priv, index, ctrl1, ctrl2);
+}
+
+static int
+yt921x_fdb_join(struct yt921x_priv *priv, const unsigned char *addr, u16 vid,
+		u16 ports_mask)
+{
+	u16 index;
+	u32 ctrl1;
+	u32 ctrl2;
+	u32 val1;
+	u32 val2;
+	int res;
+
+	/* Check for presence */
+	res = yt921x_fdb_has(priv, addr, vid, &index);
+	if (res)
+		return res;
+	if (index >= YT921X_FDB_NUM)
+		return yt921x_fdb_add(priv, addr, vid, ports_mask);
+
+	/* Check if action required */
+	res = yt921x_reg_read(priv, YT921X_FDB_OUT1, &val1);
+	if (res)
+		return res;
+	res = yt921x_reg_read(priv, YT921X_FDB_OUT2, &val2);
+	if (res)
+		return res;
+
+	ctrl1 = val1 & ~YT921X_FDB_IO1_STATUS_M;
+	ctrl1 |= YT921X_FDB_IO1_STATUS_STATIC;
+	ctrl2 = val2 | YT921X_FDB_IO2_EGR_PORTS(ports_mask);
+	if (ctrl1 == val1 && ctrl2 == val2)
+		return 0;
+
+	return yt921x_fdb_add_index_in12(priv, index, ctrl1, ctrl2);
+}
+
+static int
+yt921x_dsa_port_fdb_dump(struct dsa_switch *ds, int port,
+			 dsa_fdb_dump_cb_t *cb, void *data)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	int res;
+
+	mutex_lock(&priv->reg_lock);
+	/* Hardware FDB is shared for fdb and mdb, "bridge fdb show"
+	 * only wants to see unicast
+	 */
+	res = yt921x_fdb_dump(priv, BIT(port), cb, data);
+	mutex_unlock(&priv->reg_lock);
+
+	return res;
+}
+
+static void yt921x_dsa_port_fast_age(struct dsa_switch *ds, int port)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	struct device *dev = to_device(priv);
+	int res;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_fdb_flush_port(priv, port, false);
+	mutex_unlock(&priv->reg_lock);
+
+	if (res)
+		dev_err(dev, "Failed to %s port %d: %i\n", "clear FDB for",
+			port, res);
+}
+
+static int
+yt921x_dsa_set_ageing_time(struct dsa_switch *ds, unsigned int msecs)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	u32 ctrl;
+	int res;
+
+	/* AGEING reg is set in 5s step */
+	ctrl = clamp(msecs / 5000, 1, U16_MAX);
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_reg_write(priv, YT921X_AGEING, ctrl);
+	mutex_unlock(&priv->reg_lock);
+
+	return res;
+}
+
+static int
+yt921x_dsa_port_fdb_del(struct dsa_switch *ds, int port,
+			const unsigned char *addr, u16 vid, struct dsa_db db)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	int res;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_fdb_leave(priv, addr, vid, BIT(port));
+	mutex_unlock(&priv->reg_lock);
+
+	return res;
+}
+
+static int
+yt921x_dsa_port_fdb_add(struct dsa_switch *ds, int port,
+			const unsigned char *addr, u16 vid, struct dsa_db db)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	int res;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_fdb_join(priv, addr, vid, BIT(port));
+	mutex_unlock(&priv->reg_lock);
+
+	return res;
+}
+
+static int
+yt921x_dsa_port_mdb_del(struct dsa_switch *ds, int port,
+			const struct switchdev_obj_port_mdb *mdb,
+			struct dsa_db db)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	const unsigned char *addr = mdb->addr;
+	u16 vid = mdb->vid;
+	int res;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_fdb_leave(priv, addr, vid, BIT(port));
+	mutex_unlock(&priv->reg_lock);
+
+	return res;
+}
+
+static int
+yt921x_dsa_port_mdb_add(struct dsa_switch *ds, int port,
+			const struct switchdev_obj_port_mdb *mdb,
+			struct dsa_db db)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	const unsigned char *addr = mdb->addr;
+	u16 vid = mdb->vid;
+	int res;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_fdb_join(priv, addr, vid, BIT(port));
+	mutex_unlock(&priv->reg_lock);
+
+	return res;
+}
+
+static int
+yt921x_port_set_pvid(struct yt921x_priv *priv, int port, u16 vid)
+{
+	u32 mask;
+	u32 ctrl;
+
+	mask = YT921X_PORT_VLAN_CTRL_CVID_M;
+	ctrl = YT921X_PORT_VLAN_CTRL_CVID(vid);
+	return yt921x_reg_update_bits(priv, YT921X_PORTn_VLAN_CTRL(port),
+				      mask, ctrl);
+}
+
+static int
+yt921x_vlan_filtering(struct yt921x_priv *priv, int port, bool vlan_filtering)
+{
+	struct dsa_port *dp = dsa_to_port(&priv->ds, port);
+	struct net_device *bdev;
+	u16 pvid;
+	u32 mask;
+	u32 ctrl;
+	int res;
+
+	bdev = dsa_port_bridge_dev_get(dp);
+
+	if (!bdev || !vlan_filtering)
+		pvid = YT921X_VID_UNWARE;
+	else
+		br_vlan_get_pvid(bdev, &pvid);
+	res = yt921x_port_set_pvid(priv, port, pvid);
+	if (res)
+		return res;
+
+	mask = YT921X_PORT_VLAN_CTRL1_CVLAN_DROP_TAGGED |
+	       YT921X_PORT_VLAN_CTRL1_CVLAN_DROP_UNTAGGED;
+	ctrl = 0;
+	/* Do not drop tagged frames here; let VLAN_IGR_FILTER do it */
+	if (vlan_filtering && !pvid)
+		ctrl |= YT921X_PORT_VLAN_CTRL1_CVLAN_DROP_UNTAGGED;
+	res = yt921x_reg_update_bits(priv, YT921X_PORTn_VLAN_CTRL1(port),
+				     mask, ctrl);
+	if (res)
+		return res;
+
+	res = yt921x_reg_toggle_bits(priv, YT921X_VLAN_IGR_FILTER,
+				     YT921X_VLAN_IGR_FILTER_PORTn(port),
+				     vlan_filtering);
+	if (res)
+		return res;
+
+	/* Turn on / off VLAN awareness */
+	mask = YT921X_PORT_IGR_TPIDn_CTAG_M;
+	if (!vlan_filtering)
+		ctrl = 0;
+	else
+		ctrl = YT921X_PORT_IGR_TPIDn_CTAG(0);
+	res = yt921x_reg_update_bits(priv, YT921X_PORTn_IGR_TPID(port),
+				     mask, ctrl);
+	if (res)
+		return res;
+
+	return 0;
+}
+
+static int
+yt921x_vlan_del(struct yt921x_priv *priv, int port, u16 vid)
+{
+	u64 mask64;
+
+	mask64 = YT921X_VLAN_CTRL_PORTS(port) |
+		 YT921X_VLAN_CTRL_UNTAG_PORTn(port);
+
+	return yt921x_reg64_clear_bits(priv, YT921X_VLANn_CTRL(vid), mask64);
+}
+
+static int
+yt921x_vlan_add(struct yt921x_priv *priv, int port, u16 vid, bool untagged)
+{
+	u64 mask64;
+	u64 ctrl64;
+
+	mask64 = YT921X_VLAN_CTRL_PORTn(port) |
+		 YT921X_VLAN_CTRL_PORTS(priv->cpu_ports_mask);
+	ctrl64 = mask64;
+
+	mask64 |= YT921X_VLAN_CTRL_UNTAG_PORTn(port);
+	if (untagged)
+		ctrl64 |= YT921X_VLAN_CTRL_UNTAG_PORTn(port);
+
+	return yt921x_reg64_update_bits(priv, YT921X_VLANn_CTRL(vid),
+					mask64, ctrl64);
+}
+
+static int
+yt921x_pvid_clear(struct yt921x_priv *priv, int port)
+{
+	struct dsa_port *dp = dsa_to_port(&priv->ds, port);
+	bool vlan_filtering;
+	u32 mask;
+	int res;
+
+	vlan_filtering = dsa_port_is_vlan_filtering(dp);
+
+	res = yt921x_port_set_pvid(priv, port,
+				   vlan_filtering ? 0 : YT921X_VID_UNWARE);
+	if (res)
+		return res;
+
+	if (vlan_filtering) {
+		mask = YT921X_PORT_VLAN_CTRL1_CVLAN_DROP_UNTAGGED;
+		res = yt921x_reg_set_bits(priv, YT921X_PORTn_VLAN_CTRL1(port),
+					  mask);
+		if (res)
+			return res;
+	}
+
+	return 0;
+}
+
+static int
+yt921x_pvid_set(struct yt921x_priv *priv, int port, u16 vid)
+{
+	struct dsa_port *dp = dsa_to_port(&priv->ds, port);
+	bool vlan_filtering;
+	u32 mask;
+	int res;
+
+	vlan_filtering = dsa_port_is_vlan_filtering(dp);
+
+	if (vlan_filtering) {
+		res = yt921x_port_set_pvid(priv, port, vid);
+		if (res)
+			return res;
+	}
+
+	mask = YT921X_PORT_VLAN_CTRL1_CVLAN_DROP_UNTAGGED;
+	res = yt921x_reg_clear_bits(priv, YT921X_PORTn_VLAN_CTRL1(port), mask);
+	if (res)
+		return res;
+
+	return 0;
+}
+
+static int
+yt921x_dsa_port_vlan_filtering(struct dsa_switch *ds, int port,
+			       bool vlan_filtering,
+			       struct netlink_ext_ack *extack)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	int res;
+
+	if (dsa_is_cpu_port(ds, port))
+		return 0;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_vlan_filtering(priv, port, vlan_filtering);
+	mutex_unlock(&priv->reg_lock);
+
+	return res;
+}
+
+static int
+yt921x_dsa_port_vlan_del(struct dsa_switch *ds, int port,
+			 const struct switchdev_obj_port_vlan *vlan)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	u16 vid = vlan->vid;
+	u16 pvid;
+	int res;
+
+	if (dsa_is_cpu_port(ds, port))
+		return 0;
+
+	mutex_lock(&priv->reg_lock);
+	do {
+		struct dsa_port *dp = dsa_to_port(ds, port);
+		struct net_device *bdev;
+
+		res = yt921x_vlan_del(priv, port, vid);
+		if (res)
+			break;
+
+		bdev = dsa_port_bridge_dev_get(dp);
+		if (bdev) {
+			br_vlan_get_pvid(bdev, &pvid);
+			if (pvid == vid)
+				res = yt921x_pvid_clear(priv, port);
+		}
+	} while (0);
+	mutex_unlock(&priv->reg_lock);
+
+	return res;
+}
+
+static int
+yt921x_dsa_port_vlan_add(struct dsa_switch *ds, int port,
+			 const struct switchdev_obj_port_vlan *vlan,
+			 struct netlink_ext_ack *extack)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	u16 vid = vlan->vid;
+	u16 pvid;
+	int res;
+
+	/* CPU port is supposed to be a member of every VLAN; see
+	 * yt921x_vlan_add() and yt921x_port_setup()
+	 */
+	if (dsa_is_cpu_port(ds, port))
+		return 0;
+
+	mutex_lock(&priv->reg_lock);
+	do {
+		struct dsa_port *dp = dsa_to_port(ds, port);
+		struct net_device *bdev;
+
+		res = yt921x_vlan_add(priv, port, vid,
+				      (vlan->flags &
+				       BRIDGE_VLAN_INFO_UNTAGGED) != 0);
+		if (res)
+			break;
+
+		bdev = dsa_port_bridge_dev_get(dp);
+		if (bdev) {
+			if ((vlan->flags & BRIDGE_VLAN_INFO_PVID) != 0) {
+				res = yt921x_pvid_set(priv, port, vid);
+			} else {
+				br_vlan_get_pvid(bdev, &pvid);
+				if (pvid == vid)
+					res = yt921x_pvid_clear(priv, port);
+			}
+		}
+	} while (0);
+	mutex_unlock(&priv->reg_lock);
+
+	return res;
+}
+
+static int yt921x_userport_standalone(struct yt921x_priv *priv, int port)
+{
+	u32 mask;
+	u32 ctrl;
+	int res;
+
+	ctrl = ~priv->cpu_ports_mask;
+	res = yt921x_reg_write(priv, YT921X_PORTn_ISOLATION(port), ctrl);
+	if (res)
+		return res;
+
+	/* Turn off FDB learning to prevent FDB pollution */
+	mask = YT921X_PORT_LEARN_DIS;
+	res = yt921x_reg_set_bits(priv, YT921X_PORTn_LEARN(port), mask);
+	if (res)
+		return res;
+
+	/* Turn off VLAN awareness */
+	mask = YT921X_PORT_IGR_TPIDn_CTAG_M;
+	res = yt921x_reg_clear_bits(priv, YT921X_PORTn_IGR_TPID(port), mask);
+	if (res)
+		return res;
+
+	/* Unrelated since learning is off and all packets are trapped;
+	 * set it anyway
+	 */
+	res = yt921x_port_set_pvid(priv, port, YT921X_VID_UNWARE);
+	if (res)
+		return res;
+
+	return 0;
+}
+
+static int yt921x_userport_bridge(struct yt921x_priv *priv, int port)
+{
+	u32 mask;
+	int res;
+
+	mask = YT921X_PORT_LEARN_DIS;
+	res = yt921x_reg_clear_bits(priv, YT921X_PORTn_LEARN(port), mask);
+	if (res)
+		return res;
+
+	return 0;
+}
+
+static int yt921x_isolate(struct yt921x_priv *priv, int port)
+{
+	u32 mask;
+	int res;
+
+	mask = BIT(port);
+	for (int i = 0; i < YT921X_PORT_NUM; i++) {
+		if ((BIT(i) & priv->cpu_ports_mask) != 0 || i == port)
+			continue;
+
+		res = yt921x_reg_set_bits(priv, YT921X_PORTn_ISOLATION(i),
+					  mask);
+		if (res)
+			return res;
+	}
+
+	return 0;
+}
+
+/* Make sure to include the CPU port in ports_mask, or your bridge will
+ * not have it.
+ */
+static int yt921x_bridge(struct yt921x_priv *priv, u16 ports_mask)
+{
+	unsigned long targets_mask = ports_mask & ~priv->cpu_ports_mask;
+	u32 isolated_mask;
+	u32 ctrl;
+	int port;
+	int res;
+
+	isolated_mask = 0;
+	for_each_set_bit(port, &targets_mask, YT921X_PORT_NUM) {
+		struct yt921x_port *pp = &priv->ports[port];
+
+		if (pp->isolated)
+			isolated_mask |= BIT(port);
+	}
+
+	/* Block from non-cpu bridge ports ... */
+	for_each_set_bit(port, &targets_mask, YT921X_PORT_NUM) {
+		struct yt921x_port *pp = &priv->ports[port];
+
+		/* to non-bridge ports */
+		ctrl = ~ports_mask;
+		/* to isolated ports when isolated */
+		if (pp->isolated)
+			ctrl |= isolated_mask;
+		/* to itself when non-hairpin */
+		if (!pp->hairpin)
+			ctrl |= BIT(port);
+		else
+			ctrl &= ~BIT(port);
+
+		res = yt921x_reg_write(priv, YT921X_PORTn_ISOLATION(port),
+				       ctrl);
+		if (res)
+			return res;
+	}
+
+	return 0;
+}
+
+static int yt921x_bridge_leave(struct yt921x_priv *priv, int port)
+{
+	int res;
+
+	res = yt921x_userport_standalone(priv, port);
+	if (res)
+		return res;
+
+	res = yt921x_isolate(priv, port);
+	if (res)
+		return res;
+
+	return 0;
+}
+
+static int
+yt921x_bridge_join(struct yt921x_priv *priv, int port, u16 ports_mask)
+{
+	int res;
+
+	res = yt921x_userport_bridge(priv, port);
+	if (res)
+		return res;
+
+	res = yt921x_bridge(priv, ports_mask);
+	if (res)
+		return res;
+
+	return 0;
+}
+
+static u32
+dsa_bridge_ports(struct dsa_switch *ds, const struct net_device *bdev)
+{
+	struct dsa_port *dp;
+	u32 mask = 0;
+
+	dsa_switch_for_each_user_port(dp, ds)
+		if (dsa_port_offloads_bridge_dev(dp, bdev))
+			mask |= BIT(dp->index);
+
+	return mask;
+}
+
+static int
+yt921x_bridge_flags(struct yt921x_priv *priv, int port,
+		    struct switchdev_brport_flags flags)
+{
+	struct yt921x_port *pp = &priv->ports[port];
+	bool do_flush;
+	u32 mask;
+	int res;
+
+	if ((flags.mask & BR_LEARNING) != 0) {
+		bool learning = (flags.val & BR_LEARNING) != 0;
+
+		mask = YT921X_PORT_LEARN_DIS;
+		res = yt921x_reg_toggle_bits(priv, YT921X_PORTn_LEARN(port),
+					     mask, !learning);
+		if (res)
+			return res;
+	}
+
+	/* BR_FLOOD, BR_MCAST_FLOOD: see the comment where ACT_UNK_ACTn_TRAP
+	 * is set
+	 */
+
+	/* BR_BCAST_FLOOD: we can filter bcast, but cannot trap them */
+
+	do_flush = false;
+	if ((flags.mask & BR_HAIRPIN_MODE) != 0) {
+		pp->hairpin = (flags.val & BR_HAIRPIN_MODE) != 0;
+		do_flush = true;
+	}
+	if ((flags.mask & BR_ISOLATED) != 0) {
+		pp->isolated = (flags.val & BR_ISOLATED) != 0;
+		do_flush = true;
+	}
+	if (do_flush) {
+		struct dsa_switch *ds = &priv->ds;
+		struct dsa_port *dp = dsa_to_port(ds, port);
+		struct net_device *bdev;
+
+		bdev = dsa_port_bridge_dev_get(dp);
+		if (bdev) {
+			u32 ports_mask;
+
+			ports_mask = dsa_bridge_ports(ds, bdev);
+			ports_mask |= priv->cpu_ports_mask;
+			res = yt921x_bridge(priv, ports_mask);
+			if (res)
+				return res;
+		}
+	}
+
+	return 0;
+}
+
+static int
+yt921x_dsa_port_pre_bridge_flags(struct dsa_switch *ds, int port,
+				 struct switchdev_brport_flags flags,
+				 struct netlink_ext_ack *extack)
+{
+	if ((flags.mask & ~(BR_HAIRPIN_MODE | BR_LEARNING | BR_FLOOD |
+			    BR_MCAST_FLOOD | BR_ISOLATED)) != 0)
+		return -EINVAL;
+	return 0;
+}
+
+static int
+yt921x_dsa_port_bridge_flags(struct dsa_switch *ds, int port,
+			     struct switchdev_brport_flags flags,
+			     struct netlink_ext_ack *extack)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	int res;
+
+	if (dsa_is_cpu_port(ds, port))
+		return 0;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_bridge_flags(priv, port, flags);
+	mutex_unlock(&priv->reg_lock);
+
+	return res;
+}
+
+static void
+yt921x_dsa_port_bridge_leave(struct dsa_switch *ds, int port,
+			     struct dsa_bridge bridge)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	struct device *dev = to_device(priv);
+	int res;
+
+	if (dsa_is_cpu_port(ds, port))
+		return;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_bridge_leave(priv, port);
+	mutex_unlock(&priv->reg_lock);
+
+	if (res)
+		dev_err(dev, "Failed to %s port %d: %i\n", "unbridge",
+			port, res);
+}
+
+static int
+yt921x_dsa_port_bridge_join(struct dsa_switch *ds, int port,
+			    struct dsa_bridge bridge, bool *tx_fwd_offload,
+			    struct netlink_ext_ack *extack)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	u16 ports_mask;
+	int res;
+
+	if (dsa_is_cpu_port(ds, port))
+		return 0;
+
+	ports_mask = dsa_bridge_ports(ds, bridge.dev);
+	ports_mask |= priv->cpu_ports_mask;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_bridge_join(priv, port, ports_mask);
+	mutex_unlock(&priv->reg_lock);
+
+	return res;
+}
+
+static int yt921x_port_down(struct yt921x_priv *priv, int port)
+{
+	u32 mask;
+	int res;
+
+	mask = YT921X_PORT_LINK | YT921X_PORT_RX_MAC_EN | YT921X_PORT_TX_MAC_EN;
+	res = yt921x_reg_clear_bits(priv, YT921X_PORTn_CTRL(port), mask);
+	if (res)
+		return res;
+
+	if (yt921x_port_is_external(port)) {
+		mask = YT921X_SERDES_LINK;
+		res = yt921x_reg_clear_bits(priv, YT921X_SERDESn(port), mask);
+		if (res)
+			return res;
+
+		mask = YT921X_XMII_LINK;
+		res = yt921x_reg_clear_bits(priv, YT921X_XMIIn(port), mask);
+		if (res)
+			return res;
+	}
+
+	return 0;
+}
+
+static int
+yt921x_port_up(struct yt921x_priv *priv, int port, unsigned int mode,
+	       phy_interface_t interface, int speed, int duplex,
+	       bool tx_pause, bool rx_pause)
+{
+	u32 mask;
+	u32 ctrl;
+	int res;
+
+	switch (speed) {
+	case SPEED_10:
+		ctrl = YT921X_PORT_SPEED_10;
+		break;
+	case SPEED_100:
+		ctrl = YT921X_PORT_SPEED_100;
+		break;
+	case SPEED_1000:
+		ctrl = YT921X_PORT_SPEED_1000;
+		break;
+	case SPEED_2500:
+		ctrl = YT921X_PORT_SPEED_2500;
+		break;
+	case SPEED_10000:
+		ctrl = YT921X_PORT_SPEED_10000;
+		break;
+	default:
+		return -EINVAL;
+	}
+	if (duplex == DUPLEX_FULL)
+		ctrl |= YT921X_PORT_DUPLEX_FULL;
+	if (tx_pause)
+		ctrl |= YT921X_PORT_TX_PAUSE;
+	if (rx_pause)
+		ctrl |= YT921X_PORT_RX_PAUSE;
+	ctrl |= YT921X_PORT_RX_MAC_EN | YT921X_PORT_TX_MAC_EN;
+	res = yt921x_reg_write(priv, YT921X_PORTn_CTRL(port), ctrl);
+	if (res)
+		return res;
+
+	if (yt921x_port_is_external(port)) {
+		mask = YT921X_SERDES_SPEED_M;
+		switch (speed) {
+		case SPEED_10:
+			ctrl = YT921X_SERDES_SPEED_10;
+			break;
+		case SPEED_100:
+			ctrl = YT921X_SERDES_SPEED_100;
+			break;
+		case SPEED_1000:
+			ctrl = YT921X_SERDES_SPEED_1000;
+			break;
+		case SPEED_2500:
+			ctrl = YT921X_SERDES_SPEED_2500;
+			break;
+		case SPEED_10000:
+			ctrl = YT921X_SERDES_SPEED_10000;
+			break;
+		default:
+			return -EINVAL;
+		}
+		mask |= YT921X_SERDES_DUPLEX_FULL;
+		if (duplex == DUPLEX_FULL)
+			ctrl |= YT921X_SERDES_DUPLEX_FULL;
+		mask |= YT921X_SERDES_TX_PAUSE;
+		if (tx_pause)
+			ctrl |= YT921X_SERDES_TX_PAUSE;
+		mask |= YT921X_SERDES_RX_PAUSE;
+		if (rx_pause)
+			ctrl |= YT921X_SERDES_RX_PAUSE;
+		mask |= YT921X_SERDES_LINK;
+		ctrl |= YT921X_SERDES_LINK;
+		res = yt921x_reg_update_bits(priv, YT921X_SERDESn(port),
+					     mask, ctrl);
+		if (res)
+			return res;
+
+		mask = YT921X_XMII_LINK;
+		res = yt921x_reg_set_bits(priv, YT921X_XMIIn(port), mask);
+		if (res)
+			return res;
+
+		switch (speed) {
+		case SPEED_10:
+			ctrl = YT921X_MDIO_POLLING_SPEED_10;
+			break;
+		case SPEED_100:
+			ctrl = YT921X_MDIO_POLLING_SPEED_100;
+			break;
+		case SPEED_1000:
+			ctrl = YT921X_MDIO_POLLING_SPEED_1000;
+			break;
+		case SPEED_2500:
+			ctrl = YT921X_MDIO_POLLING_SPEED_2500;
+			break;
+		case SPEED_10000:
+			ctrl = YT921X_MDIO_POLLING_SPEED_10000;
+			break;
+		default:
+			return -EINVAL;
+		}
+		if (duplex == DUPLEX_FULL)
+			ctrl |= YT921X_MDIO_POLLING_DUPLEX_FULL;
+		ctrl |= YT921X_MDIO_POLLING_LINK;
+		res = yt921x_reg_write(priv, YT921X_MDIO_POLLINGn(port), ctrl);
+		if (res)
+			return res;
+	}
+
+	return 0;
+}
+
+static int
+yt921x_port_config(struct yt921x_priv *priv, int port, unsigned int mode,
+		   phy_interface_t interface)
+{
+	struct device *dev = to_device(priv);
+	u32 mask;
+	u32 ctrl;
+	int res;
+
+	if (!yt921x_port_is_external(port)) {
+		if (interface != PHY_INTERFACE_MODE_INTERNAL) {
+			dev_err(dev, "Wrong mode %d on port %d\n",
+				interface, port);
+			return -EINVAL;
+		}
+		return 0;
+	}
+
+	switch (interface) {
+	/* SERDES */
+	case PHY_INTERFACE_MODE_SGMII:
+	case PHY_INTERFACE_MODE_REVSGMII:
+	case PHY_INTERFACE_MODE_100BASEX:
+	case PHY_INTERFACE_MODE_1000BASEX:
+	case PHY_INTERFACE_MODE_2500BASEX:
+		mask = YT921X_SERDES_CTRL_PORTn(port);
+		res = yt921x_reg_set_bits(priv, YT921X_SERDES_CTRL, mask);
+		if (res)
+			return res;
+
+		mask = YT921X_XMII_CTRL_PORTn(port);
+		res = yt921x_reg_clear_bits(priv, YT921X_XMII_CTRL, mask);
+		if (res)
+			return res;
+
+		mask = YT921X_SERDES_MODE_M;
+		switch (interface) {
+		case PHY_INTERFACE_MODE_SGMII:
+			ctrl = YT921X_SERDES_MODE_SGMII;
+			break;
+		case PHY_INTERFACE_MODE_REVSGMII:
+			ctrl = YT921X_SERDES_MODE_REVSGMII;
+			break;
+		case PHY_INTERFACE_MODE_100BASEX:
+			ctrl = YT921X_SERDES_MODE_100BASEX;
+			break;
+		case PHY_INTERFACE_MODE_1000BASEX:
+			ctrl = YT921X_SERDES_MODE_1000BASEX;
+			break;
+		case PHY_INTERFACE_MODE_2500BASEX:
+			ctrl = YT921X_SERDES_MODE_2500BASEX;
+			break;
+		default:
+			return -EINVAL;
+		}
+		res = yt921x_reg_update_bits(priv, YT921X_SERDESn(port),
+					     mask, ctrl);
+		if (res)
+			return res;
+
+		break;
+	/* add XMII support here */
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static void
+yt921x_phylink_mac_link_down(struct phylink_config *config, unsigned int mode,
+			     phy_interface_t interface)
+{
+	struct dsa_port *dp = dsa_phylink_to_port(config);
+	struct yt921x_priv *priv = to_yt921x_priv(dp->ds);
+	int port = dp->index;
+	int res;
+
+	/* No need to sync; port control block is hold until device remove */
+	cancel_delayed_work(&priv->ports[port].mib_read);
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_port_down(priv, port);
+	mutex_unlock(&priv->reg_lock);
+
+	if (res)
+		dev_err(dp->ds->dev, "Failed to %s port %d: %i\n", "bring down",
+			port, res);
+}
+
+static void
+yt921x_phylink_mac_link_up(struct phylink_config *config,
+			   struct phy_device *phydev, unsigned int mode,
+			   phy_interface_t interface, int speed, int duplex,
+			   bool tx_pause, bool rx_pause)
+{
+	struct dsa_port *dp = dsa_phylink_to_port(config);
+	struct yt921x_priv *priv = to_yt921x_priv(dp->ds);
+	int port = dp->index;
+	int res;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_port_up(priv, port, mode, interface, speed, duplex,
+			     tx_pause, rx_pause);
+	mutex_unlock(&priv->reg_lock);
+
+	if (res)
+		dev_err(dp->ds->dev, "Failed to %s port %d: %i\n", "bring up",
+			port, res);
+
+	schedule_delayed_work(&priv->ports[port].mib_read, 0);
+}
+
+static void
+yt921x_phylink_mac_config(struct phylink_config *config, unsigned int mode,
+			  const struct phylink_link_state *state)
+{
+	struct dsa_port *dp = dsa_phylink_to_port(config);
+	struct yt921x_priv *priv = to_yt921x_priv(dp->ds);
+	int port = dp->index;
+	int res;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_port_config(priv, port, mode, state->interface);
+	mutex_unlock(&priv->reg_lock);
+
+	if (res)
+		dev_err(dp->ds->dev, "Failed to %s port %d: %i\n", "config",
+			port, res);
+}
+
+static void
+yt921x_dsa_phylink_get_caps(struct dsa_switch *ds, int port,
+			    struct phylink_config *config)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	const struct yt921x_info *info = priv->info;
+
+	config->mac_capabilities = MAC_ASYM_PAUSE | MAC_SYM_PAUSE |
+				   MAC_10 | MAC_100 | MAC_1000;
+
+	if ((info->internal_mask & BIT(port)) != 0) {
+		/* Port 10 for MCU should probably go here too. But since that
+		 * is untested yet, turn it down for the moment by letting it
+		 * fall to the default branch.
+		 */
+		__set_bit(PHY_INTERFACE_MODE_INTERNAL,
+			  config->supported_interfaces);
+	} else if ((info->external_mask & BIT(port)) != 0) {
+		/* TODO: external ports may support SERDES only, XMII only, or
+		 * SERDES + XMII depending on the chip. However, we can't get
+		 * the accurate config table due to lack of document, thus
+		 * we simply declare SERDES + XMII and rely on the correctness
+		 * of devicetree for now.
+		 */
+
+		/* SERDES */
+		__set_bit(PHY_INTERFACE_MODE_SGMII,
+			  config->supported_interfaces);
+		__set_bit(PHY_INTERFACE_MODE_REVSGMII,
+			  config->supported_interfaces);
+		__set_bit(PHY_INTERFACE_MODE_100BASEX,
+			  config->supported_interfaces);
+		__set_bit(PHY_INTERFACE_MODE_1000BASEX,
+			  config->supported_interfaces);
+		__set_bit(PHY_INTERFACE_MODE_2500BASEX,
+			  config->supported_interfaces);
+		config->mac_capabilities |= MAC_2500FD;
+
+		/* XMII */
+
+		/* Not tested. To add support for XMII:
+		 *   - Add proper interface modes below
+		 *   - Handle them in yt921x_port_config()
+		 */
+	}
+	/* no such port: empty supported_interfaces causes phylink to turn it
+	 * down
+	 */
+}
+
+static int yt921x_port_setup(struct yt921x_priv *priv, int port)
+{
+	struct dsa_switch *ds = &priv->ds;
+	u32 ctrl;
+	int res;
+
+	res = yt921x_userport_standalone(priv, port);
+	if (res)
+		return res;
+
+	if (dsa_is_cpu_port(ds, port)) {
+		/* CPU port egress is supposed to be completely controlled via
+		 * tagging, so set to oneway isolated (drop all packets without
+		 * tag).
+		 */
+		ctrl = ~(u32)0;
+		res = yt921x_reg_write(priv, YT921X_PORTn_ISOLATION(port),
+				       ctrl);
+		if (res)
+			return res;
+
+		/* To simplify FDB "isolation" simulation, we also disable
+		 * learning on the CPU port, and let software identify packets
+		 * towarding CPU (either trapped or a static FDB entry is
+		 * matched, no matter which bridge that entry is for), which is
+		 * already done by yt921x_userport_standalone(). As a result,
+		 * VLAN-awareness becomes unrelated on the CPU port (set to
+		 * VLAN-unaware by the way).
+		 */
+	}
+
+	return 0;
+}
+
+static enum dsa_tag_protocol
+yt921x_dsa_get_tag_protocol(struct dsa_switch *ds, int port,
+			    enum dsa_tag_protocol m)
+{
+	return DSA_TAG_PROTO_YT921X;
+}
+
+static int yt921x_dsa_port_setup(struct dsa_switch *ds, int port)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	int res;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_port_setup(priv, port);
+	mutex_unlock(&priv->reg_lock);
+
+	return res;
+}
+
+static int yt921x_edata_wait(struct yt921x_priv *priv, u32 *valp)
+{
+	u32 val = YT921X_EDATA_DATA_IDLE;
+	int res;
+
+	res = yt921x_reg_wait(priv, YT921X_EDATA_DATA,
+			      YT921X_EDATA_DATA_STATUS_M, &val);
+	if (res)
+		return res;
+
+	*valp = val;
+	return 0;
+}
+
+static int
+yt921x_edata_read_cont(struct yt921x_priv *priv, u8 addr, u8 *valp)
+{
+	u32 ctrl;
+	u32 val;
+	int res;
+
+	ctrl = YT921X_EDATA_CTRL_ADDR(addr) | YT921X_EDATA_CTRL_READ;
+	res = yt921x_reg_write(priv, YT921X_EDATA_CTRL, ctrl);
+	if (res)
+		return res;
+	res = yt921x_edata_wait(priv, &val);
+	if (res)
+		return res;
+
+	*valp = FIELD_GET(YT921X_EDATA_DATA_DATA_M, val);
+	return 0;
+}
+
+static int yt921x_edata_read(struct yt921x_priv *priv, u8 addr, u8 *valp)
+{
+	u32 val;
+	int res;
+
+	res = yt921x_edata_wait(priv, &val);
+	if (res)
+		return res;
+	return yt921x_edata_read_cont(priv, addr, valp);
+}
+
+static int yt921x_chip_detect(struct yt921x_priv *priv)
+{
+	struct device *dev = to_device(priv);
+	const struct yt921x_info *info;
+	u8 extmode;
+	u32 chipid;
+	u32 major;
+	u32 mode;
+	int res;
+
+	res = yt921x_reg_read(priv, YT921X_CHIP_ID, &chipid);
+	if (res)
+		return res;
+
+	major = FIELD_GET(YT921X_CHIP_ID_MAJOR, chipid);
+
+	for (info = yt921x_infos; info->name; info++)
+		if (info->major == major)
+			break;
+	if (!info->name) {
+		dev_err(dev, "Unexpected chipid 0x%x\n", chipid);
+		return -ENODEV;
+	}
+
+	res = yt921x_reg_read(priv, YT921X_CHIP_MODE, &mode);
+	if (res)
+		return res;
+	res = yt921x_edata_read(priv, YT921X_EDATA_EXTMODE, &extmode);
+	if (res)
+		return res;
+
+	for (; info->name; info++)
+		if (info->major == major && info->mode == mode &&
+		    info->extmode == extmode)
+			break;
+	if (!info->name) {
+		dev_err(dev,
+			"Unsupported chipid 0x%x with chipmode 0x%x 0x%x\n",
+			chipid, mode, extmode);
+		return -ENODEV;
+	}
+
+	/* Print chipid here since we are interested in lower 16 bits */
+	dev_info(dev,
+		 "Motorcomm %s ethernet switch, chipid: 0x%x, "
+		 "chipmode: 0x%x 0x%x\n",
+		 info->name, chipid, mode, extmode);
+
+	priv->info = info;
+	return 0;
+}
+
+static int yt921x_chip_reset(struct yt921x_priv *priv)
+{
+	struct device *dev = to_device(priv);
+	u16 eth_p_tag;
+	u32 val;
+	int res;
+
+	res = yt921x_chip_detect(priv);
+	if (res)
+		return res;
+
+	/* Reset */
+	res = yt921x_reg_write(priv, YT921X_RST, YT921X_RST_HW);
+	if (res)
+		return res;
+
+	/* RST_HW is almost same as GPIO hard reset, so we need this delay. */
+	fsleep(YT921X_RST_DELAY_US);
+
+	val = 0;
+	res = yt921x_reg_wait(priv, YT921X_RST, ~0, &val);
+	if (res)
+		return res;
+
+	/* Check for tag EtherType; do it after reset in case you messed it up
+	 * before.
+	 */
+	res = yt921x_reg_read(priv, YT921X_CPU_TAG_TPID, &val);
+	if (res)
+		return res;
+	eth_p_tag = FIELD_GET(YT921X_CPU_TAG_TPID_TPID_M, val);
+	if (eth_p_tag != ETH_P_YT921X) {
+		dev_err(dev, "Tag type 0x%x != 0x%x\n", eth_p_tag,
+			ETH_P_YT921X);
+		/* Despite being possible, we choose not to set CPU_TAG_TPID,
+		 * since there is no way it can be different unless you have the
+		 * wrong chip.
+		 */
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int yt921x_chip_setup(struct yt921x_priv *priv)
+{
+	struct dsa_switch *ds = &priv->ds;
+	unsigned long cpu_ports_mask;
+	u64 ctrl64;
+	u32 ctrl;
+	int port;
+	int res;
+
+	/* Enable DSA */
+	priv->cpu_ports_mask = dsa_cpu_ports(ds);
+
+	ctrl = YT921X_EXT_CPU_PORT_TAG_EN | YT921X_EXT_CPU_PORT_PORT_EN |
+	       YT921X_EXT_CPU_PORT_PORT(__ffs(priv->cpu_ports_mask));
+	res = yt921x_reg_write(priv, YT921X_EXT_CPU_PORT, ctrl);
+	if (res)
+		return res;
+
+	/* Enable and clear MIB */
+	res = yt921x_reg_set_bits(priv, YT921X_FUNC, YT921X_FUNC_MIB);
+	if (res)
+		return res;
+
+	ctrl = YT921X_MIB_CTRL_CLEAN | YT921X_MIB_CTRL_ALL_PORT;
+	res = yt921x_reg_write(priv, YT921X_MIB_CTRL, ctrl);
+	if (res)
+		return res;
+
+	/* Setup software switch */
+	ctrl = YT921X_CPU_COPY_TO_EXT_CPU;
+	res = yt921x_reg_write(priv, YT921X_CPU_COPY, ctrl);
+	if (res)
+		return res;
+
+	ctrl = GENMASK(10, 0);
+	res = yt921x_reg_write(priv, YT921X_FILTER_UNK_UCAST, ctrl);
+	if (res)
+		return res;
+	res = yt921x_reg_write(priv, YT921X_FILTER_UNK_MCAST, ctrl);
+	if (res)
+		return res;
+
+	/* YT921x does not support native DSA port bridging, so we use port
+	 * isolation to emulate it. However, be especially careful that port
+	 * isolation takes _after_ FDB lookups, i.e. if an FDB entry (from
+	 * another bridge) is matched and the destination port (in another
+	 * bridge) is blocked, the packet will be dropped instead of flooding to
+	 * the "bridged" ports, thus we need to trap and handle those packets by
+	 * software.
+	 *
+	 * If there is no more than one bridge, we might be able to drop them
+	 * directly given some conditions are met, but we trap them in all cases
+	 * for now.
+	 */
+	ctrl = 0;
+	for (int i = 0; i < YT921X_PORT_NUM; i++)
+		ctrl |= YT921X_ACT_UNK_ACTn_TRAP(i);
+	/* Except for CPU ports, if any packets are sent via CPU ports without
+	 * tag, they should be dropped.
+	 */
+	cpu_ports_mask = priv->cpu_ports_mask;
+	for_each_set_bit(port, &cpu_ports_mask, YT921X_PORT_NUM) {
+		ctrl &= ~YT921X_ACT_UNK_ACTn_M(port);
+		ctrl |= YT921X_ACT_UNK_ACTn_DROP(port);
+	}
+	res = yt921x_reg_write(priv, YT921X_ACT_UNK_UCAST, ctrl);
+	if (res)
+		return res;
+	res = yt921x_reg_write(priv, YT921X_ACT_UNK_MCAST, ctrl);
+	if (res)
+		return res;
+
+	/* Tagged VID 0 should be treated as untagged, which confuses the
+	 * hardware a lot
+	 */
+	ctrl64 = YT921X_VLAN_CTRL_LEARN_DIS | YT921X_VLAN_CTRL_PORTS_M;
+	res = yt921x_reg64_write(priv, YT921X_VLANn_CTRL(0), ctrl64);
+	if (res)
+		return res;
+
+	/* Miscellaneous */
+	res = yt921x_reg_set_bits(priv, YT921X_SENSOR, YT921X_SENSOR_TEMP);
+	if (res)
+		return res;
+
+	return 0;
+}
+
+static int yt921x_dsa_setup(struct dsa_switch *ds)
+{
+	struct yt921x_priv *priv = to_yt921x_priv(ds);
+	struct device *dev = to_device(priv);
+	struct device_node *np = dev->of_node;
+	struct device_node *child;
+	int res;
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_chip_reset(priv);
+	mutex_unlock(&priv->reg_lock);
+
+	if (res)
+		return res;
+
+	/* Register the internal mdio bus. Nodes for internal ports should have
+	 * proper phy-handle pointing to their PHYs. Not enabling the internal
+	 * bus is possible, though pretty wired, if internal ports are not used.
+	 */
+	child = of_get_child_by_name(np, "mdio");
+	if (child) {
+		res = yt921x_mbus_int_init(priv, child);
+		of_node_put(child);
+		if (res)
+			return res;
+	}
+
+	/* External mdio bus is optional */
+	child = of_get_child_by_name(np, "mdio-external");
+	if (child) {
+		res = yt921x_mbus_ext_init(priv, child);
+		of_node_put(child);
+		if (res)
+			return res;
+
+		dev_err(dev, "Untested external mdio bus\n");
+		return -ENODEV;
+	}
+
+	mutex_lock(&priv->reg_lock);
+	res = yt921x_chip_setup(priv);
+	mutex_unlock(&priv->reg_lock);
+
+	if (res)
+		return res;
+
+	return 0;
+}
+
+static const struct phylink_mac_ops yt921x_phylink_mac_ops = {
+	.mac_link_down	= yt921x_phylink_mac_link_down,
+	.mac_link_up	= yt921x_phylink_mac_link_up,
+	.mac_config	= yt921x_phylink_mac_config,
+};
+
+static const struct dsa_switch_ops yt921x_dsa_switch_ops = {
+	/* mib */
+	.get_strings		= yt921x_dsa_get_strings,
+	.get_ethtool_stats	= yt921x_dsa_get_ethtool_stats,
+	.get_sset_count		= yt921x_dsa_get_sset_count,
+	.get_eth_mac_stats	= yt921x_dsa_get_eth_mac_stats,
+	.get_eth_ctrl_stats	= yt921x_dsa_get_eth_ctrl_stats,
+	.get_rmon_stats		= yt921x_dsa_get_rmon_stats,
+	.get_stats64		= yt921x_dsa_get_stats64,
+	.get_pause_stats	= yt921x_dsa_get_pause_stats,
+	/* eee */
+	.support_eee		= dsa_supports_eee,
+	.set_mac_eee		= yt921x_dsa_set_mac_eee,
+	/* mtu */
+	.port_change_mtu	= yt921x_dsa_port_change_mtu,
+	.port_max_mtu		= yt921x_dsa_port_max_mtu,
+	/* mirror */
+	.port_mirror_del	= yt921x_dsa_port_mirror_del,
+	.port_mirror_add	= yt921x_dsa_port_mirror_add,
+	/* fdb */
+	.port_fdb_dump		= yt921x_dsa_port_fdb_dump,
+	.port_fast_age		= yt921x_dsa_port_fast_age,
+	.set_ageing_time	= yt921x_dsa_set_ageing_time,
+	.port_fdb_del		= yt921x_dsa_port_fdb_del,
+	.port_fdb_add		= yt921x_dsa_port_fdb_add,
+	.port_mdb_del		= yt921x_dsa_port_mdb_del,
+	.port_mdb_add		= yt921x_dsa_port_mdb_add,
+	/* vlan */
+	.port_vlan_filtering	= yt921x_dsa_port_vlan_filtering,
+	.port_vlan_del		= yt921x_dsa_port_vlan_del,
+	.port_vlan_add		= yt921x_dsa_port_vlan_add,
+	/* bridge */
+	.port_pre_bridge_flags	= yt921x_dsa_port_pre_bridge_flags,
+	.port_bridge_flags	= yt921x_dsa_port_bridge_flags,
+	.port_bridge_leave	= yt921x_dsa_port_bridge_leave,
+	.port_bridge_join	= yt921x_dsa_port_bridge_join,
+	/* port */
+	.get_tag_protocol	= yt921x_dsa_get_tag_protocol,
+	.phylink_get_caps	= yt921x_dsa_phylink_get_caps,
+	.port_setup		= yt921x_dsa_port_setup,
+	/* chip */
+	.setup			= yt921x_dsa_setup,
+};
+
+static void yt921x_mdio_shutdown(struct mdio_device *mdiodev)
+{
+	struct yt921x_priv *priv = mdiodev_get_drvdata(mdiodev);
+
+	if (!priv)
+		return;
+
+	dsa_switch_shutdown(&priv->ds);
+}
+
+static void yt921x_mdio_remove(struct mdio_device *mdiodev)
+{
+	struct yt921x_priv *priv = mdiodev_get_drvdata(mdiodev);
+
+	if (!priv)
+		return;
+
+	for (size_t i = ARRAY_SIZE(priv->ports); i-- > 0; ) {
+		struct yt921x_port *pp = &priv->ports[i];
+
+		cancel_delayed_work_sync(&pp->mib_read);
+	}
+
+	dsa_unregister_switch(&priv->ds);
+
+	mutex_destroy(&priv->reg_lock);
+}
+
+static int yt921x_mdio_probe(struct mdio_device *mdiodev)
+{
+	struct device *dev = &mdiodev->dev;
+	struct yt921x_reg_mdio *mdio;
+	struct yt921x_priv *priv;
+	struct dsa_switch *ds;
+
+	priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	mdio = devm_kzalloc(dev, sizeof(*mdio), GFP_KERNEL);
+	if (!mdio)
+		return -ENOMEM;
+
+	mdio->bus = mdiodev->bus;
+	mdio->addr = mdiodev->addr;
+	mdio->switchid = 0;
+
+	mutex_init(&priv->reg_lock);
+
+	priv->reg_ops = &yt921x_reg_ops_mdio;
+	priv->reg_ctx = mdio;
+
+	for (size_t i = 0; i < ARRAY_SIZE(priv->ports); i++) {
+		struct yt921x_port *pp = &priv->ports[i];
+
+		pp->index = i;
+		INIT_DELAYED_WORK(&pp->mib_read, yt921x_poll_mib);
+	}
+
+	ds = &priv->ds;
+	ds->dev = dev;
+	ds->assisted_learning_on_cpu_port = true;
+	ds->priv = priv;
+	ds->ops = &yt921x_dsa_switch_ops;
+	ds->phylink_mac_ops = &yt921x_phylink_mac_ops;
+	ds->num_ports = YT921X_PORT_NUM;
+
+	mdiodev_set_drvdata(mdiodev, priv);
+
+	return dsa_register_switch(ds);
+}
+
+static const struct of_device_id yt921x_of_match[] = {
+	{ .compatible = "motorcomm,yt9215" },
+	{}
+};
+MODULE_DEVICE_TABLE(of, yt921x_of_match);
+
+static struct mdio_driver yt921x_mdio_driver = {
+	.probe = yt921x_mdio_probe,
+	.remove = yt921x_mdio_remove,
+	.shutdown = yt921x_mdio_shutdown,
+	.mdiodrv.driver = {
+		.name = YT921X_NAME,
+		.of_match_table = yt921x_of_match,
+	},
+};
+
+mdio_module_driver(yt921x_mdio_driver);
+
+MODULE_AUTHOR("David Yang <mmyangfl@gmail.com>");
+MODULE_DESCRIPTION("Driver for Motorcomm YT921x Switch");
+MODULE_LICENSE("GPL");
diff --git a/drivers/net/dsa/yt921x.h b/drivers/net/dsa/yt921x.h
new file mode 100644
index 000000000000..a269b313d2a8
--- /dev/null
+++ b/drivers/net/dsa/yt921x.h
@@ -0,0 +1,505 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (c) 2025 David Yang
+ */
+
+#ifndef __YT921X_H
+#define __YT921X_H
+
+#include <net/dsa.h>
+
+#define YT921X_SMI_SWITCHID_M		GENMASK(3, 2)
+#define  YT921X_SMI_SWITCHID(x)			FIELD_PREP(YT921X_SMI_SWITCHID_M, (x))
+#define YT921X_SMI_AD			BIT(1)
+#define  YT921X_SMI_ADDR			0
+#define  YT921X_SMI_DATA			YT921X_SMI_AD
+#define YT921X_SMI_RW			BIT(0)
+#define  YT921X_SMI_WRITE			0
+#define  YT921X_SMI_READ			YT921X_SMI_RW
+
+#define YT921X_SWITCHID_NUM		4
+
+#define YT921X_RST			0x80000
+#define  YT921X_RST_HW				BIT(31)
+#define  YT921X_RST_SW				BIT(1)
+#define YT921X_FUNC			0x80004
+#define  YT921X_FUNC_MIB			BIT(1)
+#define YT921X_CHIP_ID			0x80008
+#define  YT921X_CHIP_ID_MAJOR			GENMASK(31, 16)
+#define YT921X_EXT_CPU_PORT		0x8000c
+#define  YT921X_EXT_CPU_PORT_TAG_EN		BIT(15)
+#define  YT921X_EXT_CPU_PORT_PORT_EN		BIT(14)
+#define  YT921X_EXT_CPU_PORT_PORT_M		GENMASK(3, 0)
+#define   YT921X_EXT_CPU_PORT_PORT(x)			FIELD_PREP(YT921X_EXT_CPU_PORT_PORT_M, (x))
+#define YT921X_CPU_TAG_TPID		0x80010
+#define  YT921X_CPU_TAG_TPID_TPID_M		GENMASK(15, 0)
+/* Same as ETH_P_YT921X, but this represents the true HW default, while the
+ * former is a local convention chosen by us.
+ */
+#define   YT921X_CPU_TAG_TPID_TPID_DEFAULT		0x9988
+#define YT921X_PVID_SEL			0x80014
+#define  YT921X_PVID_SEL_SVID_PORTn(port)	BIT(port)
+#define YT921X_SERDES_CTRL		0x80028
+#define  YT921X_SERDES_CTRL_PORTn_TEST(port)	BIT((port) - 3)
+#define  YT921X_SERDES_CTRL_PORTn(port)		BIT((port) - 8)
+#define YT921X_IO_LEVEL			0x80030
+#define  YT9215_IO_LEVEL_NORMAL_M		GENMASK(5, 4)
+#define   YT9215_IO_LEVEL_NORMAL(x)			FIELD_PREP(YT9215_IO_LEVEL_NORMAL_M, (x))
+#define   YT9215_IO_LEVEL_NORMAL_3V3			YT9215_IO_LEVEL_NORMAL(0)
+#define   YT9215_IO_LEVEL_NORMAL_1V8			YT9215_IO_LEVEL_NORMAL(3)
+#define  YT9215_IO_LEVEL_RGMII1_M		GENMASK(3, 2)
+#define   YT9215_IO_LEVEL_RGMII1(x)			FIELD_PREP(YT9215_IO_LEVEL_RGMII1_M, (x))
+#define   YT9215_IO_LEVEL_RGMII1_3V3			YT9215_IO_LEVEL_RGMII1(0)
+#define   YT9215_IO_LEVEL_RGMII1_2V5			YT9215_IO_LEVEL_RGMII1(1)
+#define   YT9215_IO_LEVEL_RGMII1_1V8			YT9215_IO_LEVEL_RGMII1(2)
+#define  YT9215_IO_LEVEL_RGMII0_M		GENMASK(1, 0)
+#define   YT9215_IO_LEVEL_RGMII0(x)			FIELD_PREP(YT9215_IO_LEVEL_RGMII0_M, (x))
+#define   YT9215_IO_LEVEL_RGMII0_3V3			YT9215_IO_LEVEL_RGMII0(0)
+#define   YT9215_IO_LEVEL_RGMII0_2V5			YT9215_IO_LEVEL_RGMII0(1)
+#define   YT9215_IO_LEVEL_RGMII0_1V8			YT9215_IO_LEVEL_RGMII0(2)
+#define  YT9218_IO_LEVEL_RGMII1_M		GENMASK(5, 4)
+#define   YT9218_IO_LEVEL_RGMII1(x)			FIELD_PREP(YT9218_IO_LEVEL_RGMII1_M, (x))
+#define   YT9218_IO_LEVEL_RGMII1_3V3			YT9218_IO_LEVEL_RGMII1(0)
+#define   YT9218_IO_LEVEL_RGMII1_2V5			YT9218_IO_LEVEL_RGMII1(1)
+#define   YT9218_IO_LEVEL_RGMII1_1V8			YT9218_IO_LEVEL_RGMII1(2)
+#define  YT9218_IO_LEVEL_RGMII0_M		GENMASK(3, 2)
+#define   YT9218_IO_LEVEL_RGMII0(x)			FIELD_PREP(YT9218_IO_LEVEL_RGMII0_M, (x))
+#define   YT9218_IO_LEVEL_RGMII0_3V3			YT9218_IO_LEVEL_RGMII0(0)
+#define   YT9218_IO_LEVEL_RGMII0_2V5			YT9218_IO_LEVEL_RGMII0(1)
+#define   YT9218_IO_LEVEL_RGMII0_1V8			YT9218_IO_LEVEL_RGMII0(2)
+#define  YT9218_IO_LEVEL_NORMAL_M		GENMASK(1, 0)
+#define   YT9218_IO_LEVEL_NORMAL(x)			FIELD_PREP(YT9218_IO_LEVEL_NORMAL_M, (x))
+#define   YT9218_IO_LEVEL_NORMAL_3V3			YT9218_IO_LEVEL_NORMAL(0)
+#define   YT9218_IO_LEVEL_NORMAL_1V8			YT9218_IO_LEVEL_NORMAL(3)
+#define YT921X_MAC_ADDR_HI2		0x80080
+#define YT921X_MAC_ADDR_LO4		0x80084
+#define YT921X_SERDESn(port)		(0x8008c + 4 * ((port) - 8))
+#define  YT921X_SERDES_MODE_M			GENMASK(9, 7)
+#define   YT921X_SERDES_MODE(x)				FIELD_PREP(YT921X_SERDES_MODE_M, (x))
+#define   YT921X_SERDES_MODE_SGMII			YT921X_SERDES_MODE(0)
+#define   YT921X_SERDES_MODE_REVSGMII			YT921X_SERDES_MODE(1)
+#define   YT921X_SERDES_MODE_1000BASEX			YT921X_SERDES_MODE(2)
+#define   YT921X_SERDES_MODE_100BASEX			YT921X_SERDES_MODE(3)
+#define   YT921X_SERDES_MODE_2500BASEX			YT921X_SERDES_MODE(4)
+#define   YT921X_SERDES_MODE_BASEX			YT921X_SERDES_MODE(5)
+#define  YT921X_SERDES_RX_PAUSE			BIT(6)
+#define  YT921X_SERDES_TX_PAUSE			BIT(5)
+#define  YT921X_SERDES_LINK			BIT(4)  /* force link */
+#define  YT921X_SERDES_DUPLEX_FULL		BIT(3)
+#define  YT921X_SERDES_SPEED_M			GENMASK(2, 0)
+#define   YT921X_SERDES_SPEED(x)			FIELD_PREP(YT921X_SERDES_SPEED_M, (x))
+#define   YT921X_SERDES_SPEED_10			YT921X_SERDES_SPEED(0)
+#define   YT921X_SERDES_SPEED_100			YT921X_SERDES_SPEED(1)
+#define   YT921X_SERDES_SPEED_1000			YT921X_SERDES_SPEED(2)
+#define   YT921X_SERDES_SPEED_10000			YT921X_SERDES_SPEED(3)
+#define   YT921X_SERDES_SPEED_2500			YT921X_SERDES_SPEED(4)
+#define YT921X_PORTn_CTRL(port)		(0x80100 + 4 * (port))
+#define  YT921X_PORT_CTRL_PAUSE_AN		BIT(10)
+#define YT921X_PORTn_STATUS(port)	(0x80200 + 4 * (port))
+#define  YT921X_PORT_LINK			BIT(9)  /* CTRL: auto negotiation */
+#define  YT921X_PORT_HALF_PAUSE			BIT(8)  /* Half-duplex back pressure mode */
+#define  YT921X_PORT_DUPLEX_FULL		BIT(7)
+#define  YT921X_PORT_RX_PAUSE			BIT(6)
+#define  YT921X_PORT_TX_PAUSE			BIT(5)
+#define  YT921X_PORT_RX_MAC_EN			BIT(4)
+#define  YT921X_PORT_TX_MAC_EN			BIT(3)
+#define  YT921X_PORT_SPEED_M			GENMASK(2, 0)
+#define   YT921X_PORT_SPEED(x)				FIELD_PREP(YT921X_PORT_SPEED_M, (x))
+#define   YT921X_PORT_SPEED_10				YT921X_PORT_SPEED(0)
+#define   YT921X_PORT_SPEED_100				YT921X_PORT_SPEED(1)
+#define   YT921X_PORT_SPEED_1000			YT921X_PORT_SPEED(2)
+#define   YT921X_PORT_SPEED_10000			YT921X_PORT_SPEED(3)
+#define   YT921X_PORT_SPEED_2500			YT921X_PORT_SPEED(4)
+#define YT921X_PON_STRAP_FUNC		0x80320
+#define YT921X_PON_STRAP_VAL		0x80324
+#define YT921X_PON_STRAP_CAP		0x80328
+#define  YT921X_PON_STRAP_EEE			BIT(16)
+#define  YT921X_PON_STRAP_LOOP_DETECT		BIT(7)
+#define YT921X_MDIO_POLLINGn(port)	(0x80364 + 4 * ((port) - 8))
+#define  YT921X_MDIO_POLLING_DUPLEX_FULL	BIT(4)
+#define  YT921X_MDIO_POLLING_LINK		BIT(3)
+#define  YT921X_MDIO_POLLING_SPEED_M		GENMASK(2, 0)
+#define   YT921X_MDIO_POLLING_SPEED(x)			FIELD_PREP(YT921X_MDIO_POLLING_SPEED_M, (x))
+#define   YT921X_MDIO_POLLING_SPEED_10			YT921X_MDIO_POLLING_SPEED(0)
+#define   YT921X_MDIO_POLLING_SPEED_100			YT921X_MDIO_POLLING_SPEED(1)
+#define   YT921X_MDIO_POLLING_SPEED_1000		YT921X_MDIO_POLLING_SPEED(2)
+#define   YT921X_MDIO_POLLING_SPEED_10000		YT921X_MDIO_POLLING_SPEED(3)
+#define   YT921X_MDIO_POLLING_SPEED_2500		YT921X_MDIO_POLLING_SPEED(4)
+#define YT921X_SENSOR			0x8036c
+#define  YT921X_SENSOR_TEMP			BIT(18)
+#define YT921X_TEMP			0x80374
+#define YT921X_CHIP_MODE		0x80388
+#define  YT921X_CHIP_MODE_MODE			GENMASK(1, 0)
+#define YT921X_XMII_CTRL		0x80394
+#define  YT921X_XMII_CTRL_PORTn(port)		BIT(9 - (port))  /* Yes, it's reversed */
+#define YT921X_XMIIn(port)		(0x80400 + 8 * ((port) - 8))
+#define  YT921X_XMII_MODE_M			GENMASK(31, 29)
+#define   YT921X_XMII_MODE(x)				FIELD_PREP(YT921X_XMII_MODE_M, (x))
+#define   YT921X_XMII_MODE_MII				YT921X_XMII_MODE(0)
+#define   YT921X_XMII_MODE_REVMII			YT921X_XMII_MODE(1)
+#define   YT921X_XMII_MODE_RMII				YT921X_XMII_MODE(2)
+#define   YT921X_XMII_MODE_REVRMII			YT921X_XMII_MODE(3)
+#define   YT921X_XMII_MODE_RGMII			YT921X_XMII_MODE(4)
+#define   YT921X_XMII_MODE_DISABLE			YT921X_XMII_MODE(5)
+#define  YT921X_XMII_LINK			BIT(19)  /* force link */
+#define  YT921X_XMII_EN				BIT(18)
+#define  YT921X_XMII_SOFT_RST			BIT(17)
+#define  YT921X_XMII_RGMII_TX_DELAY_150PS_M	GENMASK(16, 13)
+#define   YT921X_XMII_RGMII_TX_DELAY_150PS(x)		FIELD_PREP(YT921X_XMII_RGMII_TX_DELAY_150PS_M, (x))
+#define  YT921X_XMII_TX_CLK_IN			BIT(11)
+#define  YT921X_XMII_RX_CLK_IN			BIT(10)
+#define  YT921X_XMII_RGMII_TX_DELAY_2NS		BIT(8)
+#define  YT921X_XMII_RGMII_TX_CLK_OUT		BIT(7)
+#define  YT921X_XMII_RGMII_RX_DELAY_150PS_M	GENMASK(6, 3)
+#define   YT921X_XMII_RGMII_RX_DELAY_150PS(x)		FIELD_PREP(YT921X_XMII_RGMII_RX_DELAY_150PS_M, (x))
+#define  YT921X_XMII_RMII_PHY_TX_CLK_OUT	BIT(2)
+#define  YT921X_XMII_REVMII_TX_CLK_OUT		BIT(1)
+#define  YT921X_XMII_REVMII_RX_CLK_OUT		BIT(0)
+
+#define YT921X_MACn_FRAME(port)		(0x81008 + 0x1000 * (port))
+#define  YT921X_MAC_FRAME_SIZE_M		GENMASK(21, 8)
+#define   YT921X_MAC_FRAME_SIZE(x)			FIELD_PREP(YT921X_MAC_FRAME_SIZE_M, (x))
+
+#define YT921X_EEEn_VAL(port)		(0xa0000 + 0x40 * (port))
+#define  YT921X_EEE_VAL_DATA			BIT(1)
+
+#define YT921X_EEE_CTRL			0xb0000
+#define  YT921X_EEE_CTRL_ENn(port)		BIT(port)
+
+#define YT921X_MIB_CTRL			0xc0004
+#define  YT921X_MIB_CTRL_CLEAN			BIT(30)
+#define  YT921X_MIB_CTRL_PORT_M			GENMASK(6, 3)
+#define   YT921X_MIB_CTRL_PORT(x)			FIELD_PREP(YT921X_MIB_CTRL_PORT_M, (x))
+#define  YT921X_MIB_CTRL_ONE_PORT		BIT(1)
+#define  YT921X_MIB_CTRL_ALL_PORT		BIT(0)
+#define YT921X_MIBn_DATA0(port)		(0xc0100 + 0x100 * (port))
+#define YT921X_MIBn_DATAm(port, x)	(YT921X_MIBn_DATA0(port) + 4 * (x))
+
+#define YT921X_EDATA_CTRL		0xe0000
+#define  YT921X_EDATA_CTRL_ADDR_M		GENMASK(15, 8)
+#define   YT921X_EDATA_CTRL_ADDR(x)			FIELD_PREP(YT921X_EDATA_CTRL_ADDR_M, (x))
+#define  YT921X_EDATA_CTRL_OP_M			GENMASK(3, 0)
+#define   YT921X_EDATA_CTRL_OP(x)			FIELD_PREP(YT921X_EDATA_CTRL_OP_M, (x))
+#define   YT921X_EDATA_CTRL_READ			YT921X_EDATA_CTRL_OP(5)
+#define YT921X_EDATA_DATA		0xe0004
+#define  YT921X_EDATA_DATA_DATA_M			GENMASK(31, 24)
+#define  YT921X_EDATA_DATA_STATUS_M		GENMASK(3, 0)
+#define   YT921X_EDATA_DATA_STATUS(x)			FIELD_PREP(YT921X_EDATA_DATA_STATUS_M, (x))
+#define   YT921X_EDATA_DATA_IDLE			YT921X_EDATA_DATA_STATUS(3)
+
+#define YT921X_EXT_MBUS_OP		0x6a000
+#define YT921X_INT_MBUS_OP		0xf0000
+#define  YT921X_MBUS_OP_START			BIT(0)
+#define YT921X_EXT_MBUS_CTRL		0x6a004
+#define YT921X_INT_MBUS_CTRL		0xf0004
+#define  YT921X_MBUS_CTRL_PORT_M		GENMASK(25, 21)
+#define   YT921X_MBUS_CTRL_PORT(x)			FIELD_PREP(YT921X_MBUS_CTRL_PORT_M, (x))
+#define  YT921X_MBUS_CTRL_REG_M			GENMASK(20, 16)
+#define   YT921X_MBUS_CTRL_REG(x)			FIELD_PREP(YT921X_MBUS_CTRL_REG_M, (x))
+#define  YT921X_MBUS_CTRL_TYPE_M		GENMASK(11, 8)  /* wild guess */
+#define   YT921X_MBUS_CTRL_TYPE(x)			FIELD_PREP(YT921X_MBUS_CTRL_TYPE_M, (x))
+#define   YT921X_MBUS_CTRL_TYPE_C22			YT921X_MBUS_CTRL_TYPE(4)
+#define  YT921X_MBUS_CTRL_OP_M			GENMASK(3, 2)  /* wild guess */
+#define   YT921X_MBUS_CTRL_OP(x)			FIELD_PREP(YT921X_MBUS_CTRL_OP_M, (x))
+#define   YT921X_MBUS_CTRL_WRITE			YT921X_MBUS_CTRL_OP(1)
+#define   YT921X_MBUS_CTRL_READ				YT921X_MBUS_CTRL_OP(2)
+#define YT921X_EXT_MBUS_DOUT		0x6a008
+#define YT921X_INT_MBUS_DOUT		0xf0008
+#define YT921X_EXT_MBUS_DIN		0x6a00c
+#define YT921X_INT_MBUS_DIN		0xf000c
+
+#define YT921X_PORTn_EGR(port)		(0x100000 + 4 * (port))
+#define  YT921X_PORT_EGR_TPID_CTAG_M		GENMASK(5, 4)
+#define   YT921X_PORT_EGR_TPID_CTAG(x)			FIELD_PREP(YT921X_PORT_EGR_TPID_CTAG_M, (x))
+#define  YT921X_PORT_EGR_TPID_STAG_M		GENMASK(3, 2)
+#define   YT921X_PORT_EGR_TPID_STAG(x)			FIELD_PREP(YT921X_PORT_EGR_TPID_STAG_M, (x))
+#define YT921X_TPID_EGRn(x)		(0x100300 + 4 * (x))	/* [0, 3] */
+#define  YT921X_TPID_EGR_TPID_M			GENMASK(15, 0)
+
+#define YT921X_VLAN_IGR_FILTER		0x180280
+#define  YT921X_VLAN_IGR_FILTER_PORTn_BYPASS_IGMP(port)	BIT((port) + 11)
+#define  YT921X_VLAN_IGR_FILTER_PORTn(port)	BIT(port)
+#define YT921X_PORTn_ISOLATION(port)	(0x180294 + 4 * (port))
+#define  YT921X_PORT_ISOLATION_BLOCKn(port)	BIT(port)
+#define YT921X_PORTn_LEARN(port)	(0x1803d0 + 4 * (port))
+#define  YT921X_PORT_LEARN_VID_LEARN_MULTI_EN	BIT(22)
+#define  YT921X_PORT_LEARN_VID_LEARN_MODE	BIT(21)
+#define  YT921X_PORT_LEARN_VID_LEARN_EN		BIT(20)
+#define  YT921X_PORT_LEARN_SUSPEND_COPY_EN	BIT(19)
+#define  YT921X_PORT_LEARN_SUSPEND_DROP_EN	BIT(18)
+#define  YT921X_PORT_LEARN_DIS			BIT(17)
+#define  YT921X_PORT_LEARN_LIMIT_EN		BIT(16)
+#define  YT921X_PORT_LEARN_LIMIT_M		GENMASK(15, 8)
+#define   YT921X_PORT_LEARN_LIMIT(x)			FIELD_PREP(YT921X_PORT_LEARN_LIMIT_M, (x))
+#define  YT921X_PORT_LEARN_DROP_ON_EXCEEDED	BIT(2)
+#define  YT921X_PORT_LEARN_MODE_M		GENMASK(1, 0)
+#define   YT921X_PORT_LEARN_MODE(x)			FIELD_PREP(YT921X_PORT_LEARN_MODE_M, (x))
+#define   YT921X_PORT_LEARN_MODE_AUTO			YT921X_PORT_LEARN_MODE(0)
+#define   YT921X_PORT_LEARN_MODE_AUTO_AND_COPY		YT921X_PORT_LEARN_MODE(1)
+#define   YT921X_PORT_LEARN_MODE_CPU_CONTROL		YT921X_PORT_LEARN_MODE(2)
+#define YT921X_AGEING			0x180440
+#define  YT921X_AGEING_INTERVAL_M		GENMASK(15, 0)
+#define YT921X_FDB_IN0			0x180454
+#define YT921X_FDB_IN1			0x180458
+#define YT921X_FDB_IN2			0x18045c
+#define YT921X_FDB_OP			0x180460
+#define  YT921X_FDB_OP_INDEX_M			GENMASK(22, 11)
+#define   YT921X_FDB_OP_INDEX(x)			FIELD_PREP(YT921X_FDB_OP_INDEX_M, (x))
+#define  YT921X_FDB_OP_MODE_INDEX		BIT(10)  /* mac+fid / index */
+#define  YT921X_FDB_OP_FLUSH_MCAST		BIT(9)  /* ucast / mcast */
+#define  YT921X_FDB_OP_FLUSH_M			GENMASK(8, 7)
+#define   YT921X_FDB_OP_FLUSH(x)			FIELD_PREP(YT921X_FDB_OP_FLUSH_M, (x))
+#define   YT921X_FDB_OP_FLUSH_ALL			YT921X_FDB_OP_FLUSH(0)
+#define   YT921X_FDB_OP_FLUSH_PORT			YT921X_FDB_OP_FLUSH(1)
+#define   YT921X_FDB_OP_FLUSH_PORT_VID			YT921X_FDB_OP_FLUSH(2)
+#define   YT921X_FDB_OP_FLUSH_VID			YT921X_FDB_OP_FLUSH(3)
+#define  YT921X_FDB_OP_FLUSH_STATIC		BIT(6)
+#define  YT921X_FDB_OP_NEXT_TYPE_M		GENMASK(5, 4)
+#define   YT921X_FDB_OP_NEXT_TYPE(x)			FIELD_PREP(YT921X_FDB_OP_NEXT_TYPE_M, (x))
+#define   YT921X_FDB_OP_NEXT_TYPE_UCAST_PORT		YT921X_FDB_OP_NEXT_TYPE(0)
+#define   YT921X_FDB_OP_NEXT_TYPE_UCAST_VID		YT921X_FDB_OP_NEXT_TYPE(1)
+#define   YT921X_FDB_OP_NEXT_TYPE_UCAST			YT921X_FDB_OP_NEXT_TYPE(2)
+#define   YT921X_FDB_OP_NEXT_TYPE_MCAST			YT921X_FDB_OP_NEXT_TYPE(3)
+#define  YT921X_FDB_OP_OP_M			GENMASK(3, 1)
+#define   YT921X_FDB_OP_OP(x)				FIELD_PREP(YT921X_FDB_OP_OP_M, (x))
+#define   YT921X_FDB_OP_OP_ADD				YT921X_FDB_OP_OP(0)
+#define   YT921X_FDB_OP_OP_DEL				YT921X_FDB_OP_OP(1)
+#define   YT921X_FDB_OP_OP_GET_ONE			YT921X_FDB_OP_OP(2)
+#define   YT921X_FDB_OP_OP_GET_NEXT			YT921X_FDB_OP_OP(3)
+#define   YT921X_FDB_OP_OP_FLUSH			YT921X_FDB_OP_OP(4)
+#define  YT921X_FDB_OP_START			BIT(0)
+#define YT921X_FDB_RESULT		0x180464
+#define  YT921X_FDB_RESULT_DONE			BIT(15)
+#define  YT921X_FDB_RESULT_NOTFOUND		BIT(14)
+#define  YT921X_FDB_RESULT_OVERWRITED		BIT(13)
+#define  YT921X_FDB_RESULT_INDEX_M		GENMASK(11, 0)
+#define   YT921X_FDB_RESULT_INDEX(x)			FIELD_PREP(YT921X_FDB_RESULT_INDEX_M, (x))
+#define YT921X_FDB_OUT0			0x1804b0
+#define  YT921X_FDB_IO0_ADDR_HI4_M		GENMASK(31, 0)
+#define YT921X_FDB_OUT1			0x1804b4
+#define  YT921X_FDB_IO1_EGR_INT_PRI_EN		BIT(31)
+#define  YT921X_FDB_IO1_STATUS_M		GENMASK(30, 28)
+#define   YT921X_FDB_IO1_STATUS(x)			FIELD_PREP(YT921X_FDB_IO1_STATUS_M, (x))
+#define   YT921X_FDB_IO1_STATUS_INVALID			YT921X_FDB_IO1_STATUS(0)
+#define   YT921X_FDB_IO1_STATUS_MIN_TIME		YT921X_FDB_IO1_STATUS(1)
+#define   YT921X_FDB_IO1_STATUS_MOVE_AGING_MAX_TIME	YT921X_FDB_IO1_STATUS(3)
+#define   YT921X_FDB_IO1_STATUS_MAX_TIME		YT921X_FDB_IO1_STATUS(5)
+#define   YT921X_FDB_IO1_STATUS_PENDING			YT921X_FDB_IO1_STATUS(6)
+#define   YT921X_FDB_IO1_STATUS_STATIC			YT921X_FDB_IO1_STATUS(7)
+#define  YT921X_FDB_IO1_FID_M			GENMASK(27, 16)  /* filtering ID (VID) */
+#define   YT921X_FDB_IO1_FID(x)				FIELD_PREP(YT921X_FDB_IO1_FID_M, (x))
+#define  YT921X_FDB_IO1_ADDR_LO2_M		GENMASK(15, 0)
+#define YT921X_FDB_OUT2			0x1804b8
+#define  YT921X_FDB_IO2_MOVE_AGING_STATUS_M	GENMASK(31, 30)
+#define  YT921X_FDB_IO2_IGR_DROP		BIT(29)
+#define  YT921X_FDB_IO2_EGR_PORTS_M		GENMASK(28, 18)
+#define   YT921X_FDB_IO2_EGR_PORTS(x)			FIELD_PREP(YT921X_FDB_IO2_EGR_PORTS_M, (x))
+#define  YT921X_FDB_IO2_EGR_DROP		BIT(17)
+#define  YT921X_FDB_IO2_COPY_TO_CPU		BIT(16)
+#define  YT921X_FDB_IO2_IGR_INT_PRI_EN		BIT(15)
+#define  YT921X_FDB_IO2_INT_PRI_M		GENMASK(14, 12)
+#define   YT921X_FDB_IO2_INT_PRI(x)			FIELD_PREP(YT921X_FDB_IO2_INT_PRI_M, (x))
+#define  YT921X_FDB_IO2_NEW_VID_M		GENMASK(11, 0)
+#define   YT921X_FDB_IO2_NEW_VID(x)			FIELD_PREP(YT921X_FDB_IO2_NEW_VID_M, (x))
+#define YT921X_FILTER_UNK_UCAST		0x180508
+#define YT921X_FILTER_UNK_MCAST		0x18050c
+#define YT921X_FILTER_MCAST		0x180510
+#define YT921X_FILTER_BCAST		0x180514
+#define  YT921X_FILTER_PORTS_M			GENMASK(10, 0)
+#define   YT921X_FILTER_PORTS(x)			FIELD_PREP(YT921X_FILTER_PORTS_M, (x))
+#define  YT921X_FILTER_PORTn(port)		BIT(port)
+#define YT921X_VLAN_EGR_FILTER		0x180598
+#define  YT921X_VLAN_EGR_FILTER_PORTn(port)	BIT(port)
+#define YT921X_CPU_COPY			0x180690
+#define  YT921X_CPU_COPY_FORCE_INT_PORT		BIT(2)
+#define  YT921X_CPU_COPY_TO_INT_CPU		BIT(1)
+#define  YT921X_CPU_COPY_TO_EXT_CPU		BIT(0)
+#define YT921X_ACT_UNK_UCAST		0x180734
+#define YT921X_ACT_UNK_MCAST		0x180738
+#define  YT921X_ACT_UNK_MCAST_BYPASS_DROP_RMA	BIT(23)
+#define  YT921X_ACT_UNK_MCAST_BYPASS_DROP_IGMP	BIT(22)
+#define  YT921X_ACT_UNK_ACTn_M(port)		GENMASK(2 * (port) + 1, 2 * (port))
+#define   YT921X_ACT_UNK_ACTn(port, x)			((x) << (2 * (port)))
+#define   YT921X_ACT_UNK_ACTn_FORWARD(port)		YT921X_ACT_UNK_ACTn(port, 0)  /* flood */
+#define   YT921X_ACT_UNK_ACTn_TRAP(port)		YT921X_ACT_UNK_ACTn(port, 1)  /* steer to CPU */
+#define   YT921X_ACT_UNK_ACTn_DROP(port)		YT921X_ACT_UNK_ACTn(port, 2)  /* discard */
+/* NEVER use this action; see comments in the tag driver */
+#define   YT921X_ACT_UNK_ACTn_COPY(port)		YT921X_ACT_UNK_ACTn(port, 3)  /* flood and copy */
+#define YT921X_FDB_HW_FLUSH		0x180958
+#define  YT921X_FDB_HW_FLUSH_ON_LINKDOWN	BIT(0)
+
+#define YT921X_VLANn_CTRL(vlan)		(0x188000 + 8 * (vlan))
+#define  YT921X_VLAN_CTRL_UNTAG_PORTS_M		GENMASK(50, 40)
+#define   YT921X_VLAN_CTRL_UNTAG_PORTS(x)		FIELD_PREP(YT921X_VLAN_CTRL_UNTAG_PORTS_M, (x))
+#define  YT921X_VLAN_CTRL_UNTAG_PORTn(port)	BIT((port) + 40)
+#define  YT921X_VLAN_CTRL_STP_ID_M		GENMASK(39, 36)
+#define   YT921X_VLAN_CTRL_STP_ID(x)			FIELD_PREP(YT921X_VLAN_CTRL_STP_ID_M, (x))
+#define  YT921X_VLAN_CTRL_SVLAN_EN		BIT(35)
+#define  YT921X_VLAN_CTRL_FID_M			GENMASK(34, 23)
+#define   YT921X_VLAN_CTRL_FID(x)			FIELD_PREP(YT921X_VLAN_CTRL_FID_M, (x))
+#define  YT921X_VLAN_CTRL_LEARN_DIS		BIT(22)
+#define  YT921X_VLAN_CTRL_INT_PRI_EN		BIT(21)
+#define  YT921X_VLAN_CTRL_INT_PRI_M		GENMASK(20, 18)
+#define  YT921X_VLAN_CTRL_PORTS_M		GENMASK(17, 7)
+#define   YT921X_VLAN_CTRL_PORTS(x)			FIELD_PREP(YT921X_VLAN_CTRL_PORTS_M, (x))
+#define  YT921X_VLAN_CTRL_PORTn(port)		BIT((port) + 7)
+#define  YT921X_VLAN_CTRL_BYPASS_1X_AC		BIT(6)
+#define  YT921X_VLAN_CTRL_METER_EN		BIT(5)
+#define  YT921X_VLAN_CTRL_METER_ID_M		GENMASK(4, 0)
+
+#define YT921X_TPID_IGRn(x)		(0x210000 + 4 * (x))	/* [0, 3] */
+#define  YT921X_TPID_IGR_TPID_M			GENMASK(15, 0)
+#define YT921X_PORTn_IGR_TPID(port)	(0x210010 + 4 * (port))
+#define  YT921X_PORT_IGR_TPIDn_STAG_M		GENMASK(7, 4)
+#define  YT921X_PORT_IGR_TPIDn_STAG(x)		BIT((x) + 4)
+#define  YT921X_PORT_IGR_TPIDn_CTAG_M		GENMASK(3, 0)
+#define  YT921X_PORT_IGR_TPIDn_CTAG(x)		BIT(x)
+
+#define YT921X_PORTn_VLAN_CTRL(port)	(0x230010 + 4 * (port))
+#define  YT921X_PORT_VLAN_CTRL_SVLAN_PRI_EN	BIT(31)
+#define  YT921X_PORT_VLAN_CTRL_CVLAN_PRI_EN	BIT(30)
+#define  YT921X_PORT_VLAN_CTRL_SVID_M		GENMASK(29, 18)
+#define   YT921X_PORT_VLAN_CTRL_SVID(x)			FIELD_PREP(YT921X_PORT_VLAN_CTRL_SVID_M, (x))
+#define  YT921X_PORT_VLAN_CTRL_CVID_M		GENMASK(17, 6)
+#define   YT921X_PORT_VLAN_CTRL_CVID(x)			FIELD_PREP(YT921X_PORT_VLAN_CTRL_CVID_M, (x))
+#define  YT921X_PORT_VLAN_CTRL_SVLAN_PRI_M	GENMASK(5, 3)
+#define  YT921X_PORT_VLAN_CTRL_CVLAN_PRI_M	GENMASK(2, 0)
+#define YT921X_PORTn_VLAN_CTRL1(port)	(0x230080 + 4 * (port))
+#define  YT921X_PORT_VLAN_CTRL1_VLAN_RANGE_EN	BIT(8)
+#define  YT921X_PORT_VLAN_CTRL1_VLAN_RANGE_PROFILE_ID_M	GENMASK(7, 4)
+#define  YT921X_PORT_VLAN_CTRL1_SVLAN_DROP_TAGGED	BIT(3)
+#define  YT921X_PORT_VLAN_CTRL1_SVLAN_DROP_UNTAGGED	BIT(2)
+#define  YT921X_PORT_VLAN_CTRL1_CVLAN_DROP_TAGGED	BIT(1)
+#define  YT921X_PORT_VLAN_CTRL1_CVLAN_DROP_UNTAGGED	BIT(0)
+
+#define YT921X_MIRROR			0x300300
+#define  YT921X_MIRROR_IGR_PORTS_M		GENMASK(26, 16)
+#define   YT921X_MIRROR_IGR_PORTS(x)			FIELD_PREP(YT921X_MIRROR_IGR_PORTS_M, (x))
+#define  YT921X_MIRROR_IGR_PORTn(port)		BIT((port) + 16)
+#define  YT921X_MIRROR_EGR_PORTS_M		GENMASK(14, 4)
+#define   YT921X_MIRROR_EGR_PORTS(x)			FIELD_PREP(YT921X_MIRROR_EGR_PORTS_M, (x))
+#define  YT921X_MIRROR_EGR_PORTn(port)		BIT((port) + 4)
+#define  YT921X_MIRROR_PORT_M			GENMASK(3, 0)
+#define   YT921X_MIRROR_PORT(x)				FIELD_PREP(YT921X_MIRROR_PORT_M, (x))
+
+#define YT921X_EDATA_EXTMODE	0xfb
+#define YT921X_EDATA_LEN	0x100
+
+#define YT921X_FDB_NUM	4096
+
+enum yt921x_fdb_entry_status {
+	YT921X_FDB_ENTRY_STATUS_INVALID = 0,
+	YT921X_FDB_ENTRY_STATUS_MIN_TIME = 1,
+	YT921X_FDB_ENTRY_STATUS_MOVE_AGING_MAX_TIME = 3,
+	YT921X_FDB_ENTRY_STATUS_MAX_TIME = 5,
+	YT921X_FDB_ENTRY_STATUS_PENDING = 6,
+	YT921X_FDB_ENTRY_STATUS_STATIC = 7,
+};
+
+#define YT9215_MAJOR	0x9002
+#define YT9218_MAJOR	0x9001
+
+/* required for a hard reset */
+#define YT921X_RST_DELAY_US	10000
+
+#define YT921X_FRAME_SIZE_MAX	0x2400  /* 9216 */
+
+#define YT921X_TAG_LEN	8
+
+/* 8 internal + 2 external + 1 mcu */
+#define YT921X_PORT_NUM			11
+
+#define yt921x_port_is_internal(port) ((port) < 8)
+#define yt921x_port_is_external(port) (8 <= (port) && (port) < 9)
+
+struct yt921x_mib {
+	u64 rx_broadcast;
+	u64 rx_pause;
+	u64 rx_multicast;
+	u64 rx_crc_errors;
+
+	u64 rx_alignment_errors;
+	u64 rx_undersize_errors;
+	u64 rx_fragment_errors;
+	u64 rx_64byte;
+
+	u64 rx_65_127byte;
+	u64 rx_128_255byte;
+	u64 rx_256_511byte;
+	u64 rx_512_1023byte;
+
+	u64 rx_1024_1518byte;
+	u64 rx_jumbo;
+	u64 rx_good_bytes;
+
+	u64 rx_bad_bytes;
+	u64 rx_oversize_errors;
+
+	u64 rx_dropped;
+	u64 tx_broadcast;
+	u64 tx_pause;
+	u64 tx_multicast;
+
+	u64 tx_undersize_errors;
+	u64 tx_64byte;
+	u64 tx_65_127byte;
+	u64 tx_128_255byte;
+
+	u64 tx_256_511byte;
+	u64 tx_512_1023byte;
+	u64 tx_1024_1518byte;
+	u64 tx_jumbo;
+
+	u64 tx_good_bytes;
+	u64 tx_collisions;
+
+	u64 tx_aborted_errors;
+	u64 tx_multiple_collisions;
+	u64 tx_single_collisions;
+	u64 tx_good;
+
+	u64 tx_deferred;
+	u64 tx_late_collisions;
+	u64 rx_oam;
+	u64 tx_oam;
+};
+
+struct yt921x_port {
+	unsigned char index;
+
+	bool hairpin;
+	bool isolated;
+
+	struct delayed_work mib_read;
+	struct yt921x_mib mib;
+	u64 rx_frames;
+	u64 tx_frames;
+};
+
+struct yt921x_reg_ops {
+	int (*read)(void *context, u32 reg, u32 *valp);
+	int (*write)(void *context, u32 reg, u32 val);
+};
+
+struct yt921x_priv {
+	struct dsa_switch ds;
+
+	const struct yt921x_info *info;
+	/* cache of dsa_cpu_ports(ds) */
+	u16 cpu_ports_mask;
+
+	/* protect the access to the switch registers */
+	struct mutex reg_lock;
+	const struct yt921x_reg_ops *reg_ops;
+	void *reg_ctx;
+
+	/* mdio master bus */
+	struct mii_bus *mbus_int;
+	struct mii_bus *mbus_ext;
+
+	struct yt921x_port ports[YT921X_PORT_NUM];
+
+	u16 eee_ports_mask;
+};
+
+#endif
-- 
2.51.0


^ permalink raw reply related	[flat|nested] 19+ messages in thread

* Re: [PATCH net-next v11 1/5] dt-bindings: ethernet-phy: add reverse SGMII phy interface type
  2025-09-22 13:11 ` [PATCH net-next v11 1/5] dt-bindings: ethernet-phy: add reverse SGMII phy interface type David Yang
@ 2025-09-23 13:46   ` Rob Herring (Arm)
  0 siblings, 0 replies; 19+ messages in thread
From: Rob Herring (Arm) @ 2025-09-23 13:46 UTC (permalink / raw)
  To: David Yang
  Cc: linux-kernel, Paolo Abeni, Krzysztof Kozlowski, Heiner Kallweit,
	Simon Horman, netdev, Jakub Kicinski, Vladimir Oltean,
	Conor Dooley, Andrew Lunn, Russell King, Eric Dumazet, devicetree,
	David S. Miller


On Mon, 22 Sep 2025 21:11:39 +0800, David Yang wrote:
> The "reverse SGMII" protocol name is a personal invention, derived from
> "reverse MII" and "reverse RMII", this means: "behave like an SGMII
> PHY".
> 
> Signed-off-by: David Yang <mmyangfl@gmail.com>
> ---
>  Documentation/devicetree/bindings/net/ethernet-controller.yaml | 1 +
>  1 file changed, 1 insertion(+)
> 

Acked-by: Rob Herring (Arm) <robh@kernel.org>


^ permalink raw reply	[flat|nested] 19+ messages in thread

* Re: [PATCH net-next v11 2/5] net: phy: introduce PHY_INTERFACE_MODE_REVSGMII
  2025-09-22 13:11 ` [PATCH net-next v11 2/5] net: phy: introduce PHY_INTERFACE_MODE_REVSGMII David Yang
@ 2025-09-23 14:32   ` Maxime Chevallier
  2025-09-24 17:50   ` Russell King (Oracle)
  1 sibling, 0 replies; 19+ messages in thread
From: Maxime Chevallier @ 2025-09-23 14:32 UTC (permalink / raw)
  To: David Yang, netdev
  Cc: Andrew Lunn, Vladimir Oltean, David S. Miller, Eric Dumazet,
	Jakub Kicinski, Paolo Abeni, Rob Herring, Krzysztof Kozlowski,
	Conor Dooley, Heiner Kallweit, Russell King, Simon Horman,
	devicetree, linux-kernel

On 22/09/2025 18:41, David Yang wrote:
> The "reverse SGMII" protocol name is a personal invention, derived from
> "reverse MII" and "reverse RMII", this means: "behave like an SGMII
> PHY".

Wasn't it Russell who suggested that name ? :D

> Signed-off-by: David Yang <mmyangfl@gmail.com>

Reviewed-by: Maxime Chevallier <maxime.chevallier@bootlin.com>

Maxime


^ permalink raw reply	[flat|nested] 19+ messages in thread

* Re: [PATCH net-next v11 5/5] net: dsa: yt921x: Add support for Motorcomm YT921x
  2025-09-22 13:11 ` [PATCH net-next v11 5/5] net: dsa: yt921x: Add support for Motorcomm YT921x David Yang
@ 2025-09-24  0:47   ` Jakub Kicinski
  2025-09-24 12:33     ` Yangfl
  0 siblings, 1 reply; 19+ messages in thread
From: Jakub Kicinski @ 2025-09-24  0:47 UTC (permalink / raw)
  To: David Yang
  Cc: netdev, Andrew Lunn, Vladimir Oltean, David S. Miller,
	Eric Dumazet, Paolo Abeni, Rob Herring, Krzysztof Kozlowski,
	Conor Dooley, Heiner Kallweit, Russell King, Simon Horman,
	devicetree, linux-kernel

On Mon, 22 Sep 2025 21:11:43 +0800 David Yang wrote:
> +	MIB_DESC(1, 0x00, "RxBroadcast", false),	/* rx broadcast pkts */
> +	MIB_DESC(1, 0x04, "RxPause", false),		/* rx pause pkts */
> +	MIB_DESC(1, 0x08, "RxMulticast", false),	/* rx multicast pkts, excluding pause and OAM */
> +	MIB_DESC(1, 0x0c, "RxCrcErr", false),		/* rx crc err pkts, len >= 64B */

Keeping the string names for every stat, and the boolean seems like 
an overkill when there is grand total of 3 stats that set "true".
The comments for IEEE stats can also go, there's no extra information
here.

> +static void yt921x_mdio_remove(struct mdio_device *mdiodev)
> +{

> +		cancel_delayed_work_sync(&pp->mib_read);
> +	}
> +
> +	dsa_unregister_switch(&priv->ds);

The work canceling looks racy, the port can come up in between
cancel_work and dsa_unregister ? disable_delayed_work.. will likely 
do the job.
-- 
pw-bot: cr

^ permalink raw reply	[flat|nested] 19+ messages in thread

* Re: [PATCH net-next v11 5/5] net: dsa: yt921x: Add support for Motorcomm YT921x
  2025-09-24  0:47   ` Jakub Kicinski
@ 2025-09-24 12:33     ` Yangfl
  2025-09-25  0:35       ` Jakub Kicinski
  0 siblings, 1 reply; 19+ messages in thread
From: Yangfl @ 2025-09-24 12:33 UTC (permalink / raw)
  To: Jakub Kicinski
  Cc: netdev, Andrew Lunn, Vladimir Oltean, David S. Miller,
	Eric Dumazet, Paolo Abeni, Rob Herring, Krzysztof Kozlowski,
	Conor Dooley, Heiner Kallweit, Russell King, Simon Horman,
	devicetree, linux-kernel

On Wed, Sep 24, 2025 at 8:47 AM Jakub Kicinski <kuba@kernel.org> wrote:
...
>
> > +static void yt921x_mdio_remove(struct mdio_device *mdiodev)
> > +{
>
> > +             cancel_delayed_work_sync(&pp->mib_read);
> > +     }
> > +
> > +     dsa_unregister_switch(&priv->ds);
>
> The work canceling looks racy, the port can come up in between
> cancel_work and dsa_unregister ? disable_delayed_work.. will likely
> do the job.

Are you sure about this? There are many others who use
cancel_delayed_work_sync in their teardown methods (for example
ar9331_sw_remove). If that is true, they should be fixed too.

^ permalink raw reply	[flat|nested] 19+ messages in thread

* Re: [PATCH net-next v11 2/5] net: phy: introduce PHY_INTERFACE_MODE_REVSGMII
  2025-09-22 13:11 ` [PATCH net-next v11 2/5] net: phy: introduce PHY_INTERFACE_MODE_REVSGMII David Yang
  2025-09-23 14:32   ` Maxime Chevallier
@ 2025-09-24 17:50   ` Russell King (Oracle)
  2025-09-24 18:41     ` Andrew Lunn
  1 sibling, 1 reply; 19+ messages in thread
From: Russell King (Oracle) @ 2025-09-24 17:50 UTC (permalink / raw)
  To: David Yang, Andrew Lunn
  Cc: netdev, Vladimir Oltean, David S. Miller, Eric Dumazet,
	Jakub Kicinski, Paolo Abeni, Rob Herring, Krzysztof Kozlowski,
	Conor Dooley, Heiner Kallweit, Simon Horman, devicetree,
	linux-kernel

On Mon, Sep 22, 2025 at 09:11:40PM +0800, David Yang wrote:
> The "reverse SGMII" protocol name is a personal invention, derived from
> "reverse MII" and "reverse RMII", this means: "behave like an SGMII
> PHY".

Sorry to mess you around, but... I've been getting further with stmmac's
PCS stuff (I've started again with it) and I've come to realise that the
stmmac driver is full of worms here.

I think we need to have a bigger discussion here.

Today, we have:

- PHY_INTERFACE_MODE_REVMII
- PHY_INTERFACE_MODE_REVRMII

which both complement their _MII and _RMII definitions. So, it seems
entirely sensible to also introduce REVSGMII to complement SGMII.

However, stmmac hardware supports "reverse" mode for more than just
SGMII, also RGMII and SMII. The driver doesn't support SMII, and is
actually buggy - despite having the DT configuration knobs (which
are used), the hardware is never actually configured to operate in
"reverse" mode (it never has the GMAC_CONFIG_TC bit set to allow the
core to, in effect, act as a PHY.) That said, the core does have
it's SGMII rate adapter switched from using the incoming inband word
to using the MAC configuration.

So, while I thought this would be useful for stmmac, given that all
the platforms we have today aren't actually using "reverse SGMII"
mode, I don't think this will be useful there.

If we go round the route of adding REVSGMII, we are also opening
the path to also having REVSMII, and all four REVRGMII* as well.
Is this a good idea?

Would it be better to have phy_interface_t + reverse-mode flag
and accept REVMII and REVRMII as a pecularity? That's probably
going to be a very painful change.

Andrew, any views?

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

^ permalink raw reply	[flat|nested] 19+ messages in thread

* Re: [PATCH net-next v11 2/5] net: phy: introduce PHY_INTERFACE_MODE_REVSGMII
  2025-09-24 17:50   ` Russell King (Oracle)
@ 2025-09-24 18:41     ` Andrew Lunn
  2025-09-24 19:18       ` Russell King (Oracle)
  0 siblings, 1 reply; 19+ messages in thread
From: Andrew Lunn @ 2025-09-24 18:41 UTC (permalink / raw)
  To: Russell King (Oracle)
  Cc: David Yang, netdev, Vladimir Oltean, David S. Miller,
	Eric Dumazet, Jakub Kicinski, Paolo Abeni, Rob Herring,
	Krzysztof Kozlowski, Conor Dooley, Heiner Kallweit, Simon Horman,
	devicetree, linux-kernel

On Wed, Sep 24, 2025 at 06:50:19PM +0100, Russell King (Oracle) wrote:
> On Mon, Sep 22, 2025 at 09:11:40PM +0800, David Yang wrote:
> > The "reverse SGMII" protocol name is a personal invention, derived from
> > "reverse MII" and "reverse RMII", this means: "behave like an SGMII
> > PHY".
> 
> Sorry to mess you around, but... I've been getting further with stmmac's
> PCS stuff (I've started again with it) and I've come to realise that the
> stmmac driver is full of worms here.
> 
> I think we need to have a bigger discussion here.
> 
> Today, we have:
> 
> - PHY_INTERFACE_MODE_REVMII
> - PHY_INTERFACE_MODE_REVRMII
> 
> which both complement their _MII and _RMII definitions. So, it seems
> entirely sensible to also introduce REVSGMII to complement SGMII.

Maybe we need to think about, what does REVfoo actually mean?

Is it simply about, who provides the clock? For MII, the 'PHY'
provides the clock to the 'MAC;. So does REVMII simply mean a MAC in
REVMII mode provides the clock? Is more needed? As far as i know MII
does not have any inband signalling.

For RMII, it appears each side can provide the clock, or consume the
clock, or a 3rd party can provide the clock. It is a hardware design
choice. So does REVRMII actual mean anything? Again there is no inband
signalling.

GMII the transmit clock is provided by the transmitter, the receive
clock by the receiver. It is symmetrical. REV has no meaning here?

In theory, {R}GMII does have inband signalling, but it is pretty much
never used. REV for GMII could thus indicate what role the device is
playing in this in-band signalling?

For any SERDES based links likes like SGMII, 1000Base-X and above,
clocking is part of the SERDES, so symmetrical. There clearly is
inband signalling, mostly, when it is not broken because of
overclocked SGMII. But we have never needed to specify what role each
end needs to play.

> However, stmmac hardware supports "reverse" mode for more than just
> SGMII, also RGMII and SMII.

How does the databook describe reverse SGMII? How does it differ from
SGMII?

	Andrew

^ permalink raw reply	[flat|nested] 19+ messages in thread

* Re: [PATCH net-next v11 2/5] net: phy: introduce PHY_INTERFACE_MODE_REVSGMII
  2025-09-24 18:41     ` Andrew Lunn
@ 2025-09-24 19:18       ` Russell King (Oracle)
  2025-09-26  6:30         ` Yangfl
  0 siblings, 1 reply; 19+ messages in thread
From: Russell King (Oracle) @ 2025-09-24 19:18 UTC (permalink / raw)
  To: Andrew Lunn
  Cc: David Yang, netdev, Vladimir Oltean, David S. Miller,
	Eric Dumazet, Jakub Kicinski, Paolo Abeni, Rob Herring,
	Krzysztof Kozlowski, Conor Dooley, Heiner Kallweit, Simon Horman,
	devicetree, linux-kernel

On Wed, Sep 24, 2025 at 08:41:06PM +0200, Andrew Lunn wrote:
> In theory, {R}GMII does have inband signalling, but it is pretty much
> never used. REV for GMII could thus indicate what role the device is
> playing in this in-band signalling?

For RGMII, as you say, the in-band signalling is pretty much never used.
The stmmac code as it stands today does have support for using it, but
the code has been broken for longer than six years:

1. the longest historical breakage, it's conditional on the hardware
   reporting that it has a PCS integrated into the design, but a PCS
   won't be integrated into the design for RGMII-only cases.

2. even if (1) was fixed, that would result in the driver manipulating
   the netif carrier state from interrupt context, always beating
   phylink's resolve worker, meaning that mac_link_(down|up) never get
   called. This results in no traffic flow and a non-functional
   interface.

So, maybe we should just ignore the RGMII in-band signalling until
someone pops up with a hard and fast requirement for it.

> For any SERDES based links likes like SGMII, 1000Base-X and above,
> clocking is part of the SERDES, so symmetrical. There clearly is
> inband signalling, mostly, when it is not broken because of
> overclocked SGMII. But we have never needed to specify what role each
> end needs to play.

100base-X is intentionally symmetric, and designed for:

	MAC----PCS---- some kind of link ----PCS----MAC

where "some kind of link" is fibre or copper. There is no reverse mode
possible there, because "reverse" is just the same as "normal".

For SGMII though, it's a different matter. The PHY-like end transmits
the link configuration. The MAC-like end receives the link
configuration and configures itself to it - and never sends a link
configuration back.

So, the formats of the in-band tx_config_reg[15:0] are different
depending on the role each end is in.

In order for a SGMII link with in-band signalling to work, one end
has to assume the MAC-like role and the other a PHY-like role.

PHY_INTERFACE_MODE_SGMII generally means that the MAC is acting in a
MAC-like role. However, stmmac had the intention (but broken) idea
that setting the DT snps,ps-speed property would configure it into a
PHY-like role. It almost does... but instead of setting the "transmit
configuration" (TC) bit, someone typo'd and instead set the "transmit
enable" (TE) bit. So no one has actually had their stmmac-based
device operating in a PHY-like role, even if they _thought_ it was!

> > However, stmmac hardware supports "reverse" mode for more than just
> > SGMII, also RGMII and SMII.
> 
> How does the databook describe reverse SGMII? How does it differ from
> SGMII?

It doesn't describe "reverse SGMII". Instead, it describes:

1. The TC bit in the MAC configuration register, which makes the block
   transmit the speed and duplex from the MAC configuration register
   over RGMII, SGMII or SMII links (only, not 1000base-X.)

2. The SGMIIRAL bit in the PCS control register, which switches where
   the SGMII rate adapter layer takes its speed configuration from -
   either the incoming in-band tx_config_reg[15:0] word, or from the
   MAC configuration register. It is explicitly stated for this bit
   that it is for back-to-back MAC links, and as it's specific to
   SGMII, that means a back-to-back SGMII MAC link.

Set both these bits while the MAC is configured for SGMII mode, and
you have a stmmac MAC which immitates a SGMII PHY as far as the
in-band tx_config_reg[15:0] word is concerned.

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

^ permalink raw reply	[flat|nested] 19+ messages in thread

* Re: [PATCH net-next v11 5/5] net: dsa: yt921x: Add support for Motorcomm YT921x
  2025-09-24 12:33     ` Yangfl
@ 2025-09-25  0:35       ` Jakub Kicinski
  0 siblings, 0 replies; 19+ messages in thread
From: Jakub Kicinski @ 2025-09-25  0:35 UTC (permalink / raw)
  To: Yangfl
  Cc: netdev, Andrew Lunn, Vladimir Oltean, David S. Miller,
	Eric Dumazet, Paolo Abeni, Rob Herring, Krzysztof Kozlowski,
	Conor Dooley, Heiner Kallweit, Russell King, Simon Horman,
	devicetree, linux-kernel

On Wed, 24 Sep 2025 20:33:57 +0800 Yangfl wrote:
> On Wed, Sep 24, 2025 at 8:47 AM Jakub Kicinski <kuba@kernel.org> wrote:
> > > +             cancel_delayed_work_sync(&pp->mib_read);
> > > +     }
> > > +
> > > +     dsa_unregister_switch(&priv->ds);  
> >
> > The work canceling looks racy, the port can come up in between
> > cancel_work and dsa_unregister ? disable_delayed_work.. will likely
> > do the job.  
> 
> Are you sure about this? There are many others who use
> cancel_delayed_work_sync in their teardown methods (for example
> ar9331_sw_remove). If that is true, they should be fixed too.

Not at all! I'll gladly accept an explanation of why the code is
correct. "Someone else is doing it too" is not an explanation.

^ permalink raw reply	[flat|nested] 19+ messages in thread

* Re: [PATCH net-next v11 2/5] net: phy: introduce PHY_INTERFACE_MODE_REVSGMII
  2025-09-24 19:18       ` Russell King (Oracle)
@ 2025-09-26  6:30         ` Yangfl
  2025-09-26  8:45           ` Russell King (Oracle)
  2025-09-26 16:09           ` Andrew Lunn
  0 siblings, 2 replies; 19+ messages in thread
From: Yangfl @ 2025-09-26  6:30 UTC (permalink / raw)
  To: Russell King (Oracle)
  Cc: Andrew Lunn, netdev, Vladimir Oltean, David S. Miller,
	Eric Dumazet, Jakub Kicinski, Paolo Abeni, Rob Herring,
	Krzysztof Kozlowski, Conor Dooley, Heiner Kallweit, Simon Horman,
	devicetree, linux-kernel

On Thu, Sep 25, 2025 at 3:18 AM Russell King (Oracle)
<linux@armlinux.org.uk> wrote:
>
> On Wed, Sep 24, 2025 at 08:41:06PM +0200, Andrew Lunn wrote:
> > In theory, {R}GMII does have inband signalling, but it is pretty much
> > never used. REV for GMII could thus indicate what role the device is
> > playing in this in-band signalling?
>
> For RGMII, as you say, the in-band signalling is pretty much never used.
> The stmmac code as it stands today does have support for using it, but
> the code has been broken for longer than six years:
>
> 1. the longest historical breakage, it's conditional on the hardware
>    reporting that it has a PCS integrated into the design, but a PCS
>    won't be integrated into the design for RGMII-only cases.
>
> 2. even if (1) was fixed, that would result in the driver manipulating
>    the netif carrier state from interrupt context, always beating
>    phylink's resolve worker, meaning that mac_link_(down|up) never get
>    called. This results in no traffic flow and a non-functional
>    interface.
>
> So, maybe we should just ignore the RGMII in-band signalling until
> someone pops up with a hard and fast requirement for it.
>
> > For any SERDES based links likes like SGMII, 1000Base-X and above,
> > clocking is part of the SERDES, so symmetrical. There clearly is
> > inband signalling, mostly, when it is not broken because of
> > overclocked SGMII. But we have never needed to specify what role each
> > end needs to play.
>
> 100base-X is intentionally symmetric, and designed for:
>
>         MAC----PCS---- some kind of link ----PCS----MAC
>
> where "some kind of link" is fibre or copper. There is no reverse mode
> possible there, because "reverse" is just the same as "normal".
>
> For SGMII though, it's a different matter. The PHY-like end transmits
> the link configuration. The MAC-like end receives the link
> configuration and configures itself to it - and never sends a link
> configuration back.
>
> So, the formats of the in-band tx_config_reg[15:0] are different
> depending on the role each end is in.
>
> In order for a SGMII link with in-band signalling to work, one end
> has to assume the MAC-like role and the other a PHY-like role.
>
> PHY_INTERFACE_MODE_SGMII generally means that the MAC is acting in a
> MAC-like role. However, stmmac had the intention (but broken) idea
> that setting the DT snps,ps-speed property would configure it into a
> PHY-like role. It almost does... but instead of setting the "transmit
> configuration" (TC) bit, someone typo'd and instead set the "transmit
> enable" (TE) bit. So no one has actually had their stmmac-based
> device operating in a PHY-like role, even if they _thought_ it was!
>
> > > However, stmmac hardware supports "reverse" mode for more than just
> > > SGMII, also RGMII and SMII.
> >
> > How does the databook describe reverse SGMII? How does it differ from
> > SGMII?
>
> It doesn't describe "reverse SGMII". Instead, it describes:
>
> 1. The TC bit in the MAC configuration register, which makes the block
>    transmit the speed and duplex from the MAC configuration register
>    over RGMII, SGMII or SMII links (only, not 1000base-X.)
>
> 2. The SGMIIRAL bit in the PCS control register, which switches where
>    the SGMII rate adapter layer takes its speed configuration from -
>    either the incoming in-band tx_config_reg[15:0] word, or from the
>    MAC configuration register. It is explicitly stated for this bit
>    that it is for back-to-back MAC links, and as it's specific to
>    SGMII, that means a back-to-back SGMII MAC link.
>
> Set both these bits while the MAC is configured for SGMII mode, and
> you have a stmmac MAC which immitates a SGMII PHY as far as the
> in-band tx_config_reg[15:0] word is concerned.
>
> --
> RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
> FTTP is here! 80Mbps down 10Mbps up. Decent connectivity at last!

So any conclusion? Should I go on with REV*MII, or wait for (or write
it myself) reverse-mode flag?

^ permalink raw reply	[flat|nested] 19+ messages in thread

* Re: [PATCH net-next v11 2/5] net: phy: introduce PHY_INTERFACE_MODE_REVSGMII
  2025-09-26  6:30         ` Yangfl
@ 2025-09-26  8:45           ` Russell King (Oracle)
  2025-09-26 16:09           ` Andrew Lunn
  1 sibling, 0 replies; 19+ messages in thread
From: Russell King (Oracle) @ 2025-09-26  8:45 UTC (permalink / raw)
  To: Yangfl
  Cc: Andrew Lunn, netdev, Vladimir Oltean, David S. Miller,
	Eric Dumazet, Jakub Kicinski, Paolo Abeni, Rob Herring,
	Krzysztof Kozlowski, Conor Dooley, Heiner Kallweit, Simon Horman,
	devicetree, linux-kernel

On Fri, Sep 26, 2025 at 02:30:25PM +0800, Yangfl wrote:
> On Thu, Sep 25, 2025 at 3:18 AM Russell King (Oracle)
> <linux@armlinux.org.uk> wrote:
> >
> > On Wed, Sep 24, 2025 at 08:41:06PM +0200, Andrew Lunn wrote:
> > > In theory, {R}GMII does have inband signalling, but it is pretty much
> > > never used. REV for GMII could thus indicate what role the device is
> > > playing in this in-band signalling?
> >
> > For RGMII, as you say, the in-band signalling is pretty much never used.
> > The stmmac code as it stands today does have support for using it, but
> > the code has been broken for longer than six years:
> >
> > 1. the longest historical breakage, it's conditional on the hardware
> >    reporting that it has a PCS integrated into the design, but a PCS
> >    won't be integrated into the design for RGMII-only cases.
> >
> > 2. even if (1) was fixed, that would result in the driver manipulating
> >    the netif carrier state from interrupt context, always beating
> >    phylink's resolve worker, meaning that mac_link_(down|up) never get
> >    called. This results in no traffic flow and a non-functional
> >    interface.
> >
> > So, maybe we should just ignore the RGMII in-band signalling until
> > someone pops up with a hard and fast requirement for it.
> >
> > > For any SERDES based links likes like SGMII, 1000Base-X and above,
> > > clocking is part of the SERDES, so symmetrical. There clearly is
> > > inband signalling, mostly, when it is not broken because of
> > > overclocked SGMII. But we have never needed to specify what role each
> > > end needs to play.
> >
> > 100base-X is intentionally symmetric, and designed for:
> >
> >         MAC----PCS---- some kind of link ----PCS----MAC
> >
> > where "some kind of link" is fibre or copper. There is no reverse mode
> > possible there, because "reverse" is just the same as "normal".
> >
> > For SGMII though, it's a different matter. The PHY-like end transmits
> > the link configuration. The MAC-like end receives the link
> > configuration and configures itself to it - and never sends a link
> > configuration back.
> >
> > So, the formats of the in-band tx_config_reg[15:0] are different
> > depending on the role each end is in.
> >
> > In order for a SGMII link with in-band signalling to work, one end
> > has to assume the MAC-like role and the other a PHY-like role.
> >
> > PHY_INTERFACE_MODE_SGMII generally means that the MAC is acting in a
> > MAC-like role. However, stmmac had the intention (but broken) idea
> > that setting the DT snps,ps-speed property would configure it into a
> > PHY-like role. It almost does... but instead of setting the "transmit
> > configuration" (TC) bit, someone typo'd and instead set the "transmit
> > enable" (TE) bit. So no one has actually had their stmmac-based
> > device operating in a PHY-like role, even if they _thought_ it was!
> >
> > > > However, stmmac hardware supports "reverse" mode for more than just
> > > > SGMII, also RGMII and SMII.
> > >
> > > How does the databook describe reverse SGMII? How does it differ from
> > > SGMII?
> >
> > It doesn't describe "reverse SGMII". Instead, it describes:
> >
> > 1. The TC bit in the MAC configuration register, which makes the block
> >    transmit the speed and duplex from the MAC configuration register
> >    over RGMII, SGMII or SMII links (only, not 1000base-X.)
> >
> > 2. The SGMIIRAL bit in the PCS control register, which switches where
> >    the SGMII rate adapter layer takes its speed configuration from -
> >    either the incoming in-band tx_config_reg[15:0] word, or from the
> >    MAC configuration register. It is explicitly stated for this bit
> >    that it is for back-to-back MAC links, and as it's specific to
> >    SGMII, that means a back-to-back SGMII MAC link.
> >
> > Set both these bits while the MAC is configured for SGMII mode, and
> > you have a stmmac MAC which immitates a SGMII PHY as far as the
> > in-band tx_config_reg[15:0] word is concerned.
> 
> So any conclusion? Should I go on with REV*MII, or wait for (or write
> it myself) reverse-mode flag?

Clearly not as there's been no discussion beyond my response to Andrew.
I don't know what to suggest, as whatever decision we make here, we
will have to live with the consequences of it for a very long time.

I suspect no one really knows the answer, so given the lack of
engagement on the issue, my suggestion would be to just press ahead
with your current approach.

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

^ permalink raw reply	[flat|nested] 19+ messages in thread

* Re: [PATCH net-next v11 2/5] net: phy: introduce PHY_INTERFACE_MODE_REVSGMII
  2025-09-26  6:30         ` Yangfl
  2025-09-26  8:45           ` Russell King (Oracle)
@ 2025-09-26 16:09           ` Andrew Lunn
  2025-09-26 16:28             ` Yangfl
  1 sibling, 1 reply; 19+ messages in thread
From: Andrew Lunn @ 2025-09-26 16:09 UTC (permalink / raw)
  To: Yangfl
  Cc: Russell King (Oracle), netdev, Vladimir Oltean, David S. Miller,
	Eric Dumazet, Jakub Kicinski, Paolo Abeni, Rob Herring,
	Krzysztof Kozlowski, Conor Dooley, Heiner Kallweit, Simon Horman,
	devicetree, linux-kernel

> > > How does the databook describe reverse SGMII? How does it differ from
> > > SGMII?
> >
> > It doesn't describe "reverse SGMII". Instead, it describes:
> >
> > 1. The TC bit in the MAC configuration register, which makes the block
> >    transmit the speed and duplex from the MAC configuration register
> >    over RGMII, SGMII or SMII links (only, not 1000base-X.)
> >
> > 2. The SGMIIRAL bit in the PCS control register, which switches where
> >    the SGMII rate adapter layer takes its speed configuration from -
> >    either the incoming in-band tx_config_reg[15:0] word, or from the
> >    MAC configuration register. It is explicitly stated for this bit
> >    that it is for back-to-back MAC links, and as it's specific to
> >    SGMII, that means a back-to-back SGMII MAC link.
> >
> > Set both these bits while the MAC is configured for SGMII mode, and
> > you have a stmmac MAC which immitates a SGMII PHY as far as the
> > in-band tx_config_reg[15:0] word is concerned.
> 
> So any conclusion? Should I go on with REV*MII, or wait for (or write
> it myself) reverse-mode flag?

Sorry, i'm missing some context here.

Why do you actually need REVSGMII, or at least the concept?

REVMII is used when you connect one MAC to another. You need to
indicate one ends needs to play the PHY role. This is generally when
you connect a host MAC to an Ethernet switch, and you want the switch
to play the PHY role.

Now consider SGMII, when connecting a host MAC to a switch. Why would
you even use SGMII, 1000BaseX is the more logical choice. You don't
want the link to run at 100Mbps, or 10Mbps. The link between the host
and the switch should run as fast as possible. And 1000BaseX is
symmetrical, you don't need a REV concept.

Also, in these cases, stmmmac is on the host, not the switch, so it
will have the host role, leaving the switch to play 'PHY'. I'm not
sure you could even embedded stmmac in a switch, where it might want
to play 'PHY', because stmmac is software driven, where as a switch is
all hardware.

So the hardware supports reverse SGMII, but it is not clear to me why
you would want to use it.

	Andrew


^ permalink raw reply	[flat|nested] 19+ messages in thread

* Re: [PATCH net-next v11 2/5] net: phy: introduce PHY_INTERFACE_MODE_REVSGMII
  2025-09-26 16:09           ` Andrew Lunn
@ 2025-09-26 16:28             ` Yangfl
  2025-09-26 18:06               ` Russell King (Oracle)
  0 siblings, 1 reply; 19+ messages in thread
From: Yangfl @ 2025-09-26 16:28 UTC (permalink / raw)
  To: Andrew Lunn
  Cc: Russell King (Oracle), netdev, Vladimir Oltean, David S. Miller,
	Eric Dumazet, Jakub Kicinski, Paolo Abeni, Rob Herring,
	Krzysztof Kozlowski, Conor Dooley, Heiner Kallweit, Simon Horman,
	devicetree, linux-kernel

On Sat, Sep 27, 2025 at 12:09 AM Andrew Lunn <andrew@lunn.ch> wrote:
>
> > > > How does the databook describe reverse SGMII? How does it differ from
> > > > SGMII?
> > >
> > > It doesn't describe "reverse SGMII". Instead, it describes:
> > >
> > > 1. The TC bit in the MAC configuration register, which makes the block
> > >    transmit the speed and duplex from the MAC configuration register
> > >    over RGMII, SGMII or SMII links (only, not 1000base-X.)
> > >
> > > 2. The SGMIIRAL bit in the PCS control register, which switches where
> > >    the SGMII rate adapter layer takes its speed configuration from -
> > >    either the incoming in-band tx_config_reg[15:0] word, or from the
> > >    MAC configuration register. It is explicitly stated for this bit
> > >    that it is for back-to-back MAC links, and as it's specific to
> > >    SGMII, that means a back-to-back SGMII MAC link.
> > >
> > > Set both these bits while the MAC is configured for SGMII mode, and
> > > you have a stmmac MAC which immitates a SGMII PHY as far as the
> > > in-band tx_config_reg[15:0] word is concerned.
> >
> > So any conclusion? Should I go on with REV*MII, or wait for (or write
> > it myself) reverse-mode flag?
>
> Sorry, i'm missing some context here.
>
> Why do you actually need REVSGMII, or at least the concept?
>
> REVMII is used when you connect one MAC to another. You need to
> indicate one ends needs to play the PHY role. This is generally when
> you connect a host MAC to an Ethernet switch, and you want the switch
> to play the PHY role.
>
> Now consider SGMII, when connecting a host MAC to a switch. Why would
> you even use SGMII, 1000BaseX is the more logical choice. You don't
> want the link to run at 100Mbps, or 10Mbps. The link between the host
> and the switch should run as fast as possible. And 1000BaseX is
> symmetrical, you don't need a REV concept.
>
> Also, in these cases, stmmmac is on the host, not the switch, so it
> will have the host role, leaving the switch to play 'PHY'. I'm not
> sure you could even embedded stmmac in a switch, where it might want
> to play 'PHY', because stmmac is software driven, where as a switch is
> all hardware.
>
> So the hardware supports reverse SGMII, but it is not clear to me why
> you would want to use it.
>
>         Andrew
>

Cause I couldn't make 1000BaseX work with qca-ssdk, so I can only
confirm and test REVSGMII mode on my device.

^ permalink raw reply	[flat|nested] 19+ messages in thread

* Re: [PATCH net-next v11 2/5] net: phy: introduce PHY_INTERFACE_MODE_REVSGMII
  2025-09-26 16:28             ` Yangfl
@ 2025-09-26 18:06               ` Russell King (Oracle)
  0 siblings, 0 replies; 19+ messages in thread
From: Russell King (Oracle) @ 2025-09-26 18:06 UTC (permalink / raw)
  To: Yangfl
  Cc: Andrew Lunn, netdev, Vladimir Oltean, David S. Miller,
	Eric Dumazet, Jakub Kicinski, Paolo Abeni, Rob Herring,
	Krzysztof Kozlowski, Conor Dooley, Heiner Kallweit, Simon Horman,
	devicetree, linux-kernel

On Sat, Sep 27, 2025 at 12:28:05AM +0800, Yangfl wrote:
> On Sat, Sep 27, 2025 at 12:09 AM Andrew Lunn <andrew@lunn.ch> wrote:
> >
> > > > > How does the databook describe reverse SGMII? How does it differ from
> > > > > SGMII?
> > > >
> > > > It doesn't describe "reverse SGMII". Instead, it describes:
> > > >
> > > > 1. The TC bit in the MAC configuration register, which makes the block
> > > >    transmit the speed and duplex from the MAC configuration register
> > > >    over RGMII, SGMII or SMII links (only, not 1000base-X.)
> > > >
> > > > 2. The SGMIIRAL bit in the PCS control register, which switches where
> > > >    the SGMII rate adapter layer takes its speed configuration from -
> > > >    either the incoming in-band tx_config_reg[15:0] word, or from the
> > > >    MAC configuration register. It is explicitly stated for this bit
> > > >    that it is for back-to-back MAC links, and as it's specific to
> > > >    SGMII, that means a back-to-back SGMII MAC link.
> > > >
> > > > Set both these bits while the MAC is configured for SGMII mode, and
> > > > you have a stmmac MAC which immitates a SGMII PHY as far as the
> > > > in-band tx_config_reg[15:0] word is concerned.
> > >
> > > So any conclusion? Should I go on with REV*MII, or wait for (or write
> > > it myself) reverse-mode flag?
> >
> > Sorry, i'm missing some context here.
> >
> > Why do you actually need REVSGMII, or at least the concept?
> >
> > REVMII is used when you connect one MAC to another. You need to
> > indicate one ends needs to play the PHY role. This is generally when
> > you connect a host MAC to an Ethernet switch, and you want the switch
> > to play the PHY role.
> >
> > Now consider SGMII, when connecting a host MAC to a switch. Why would
> > you even use SGMII, 1000BaseX is the more logical choice. You don't
> > want the link to run at 100Mbps, or 10Mbps. The link between the host
> > and the switch should run as fast as possible. And 1000BaseX is
> > symmetrical, you don't need a REV concept.
> >
> > Also, in these cases, stmmmac is on the host, not the switch, so it
> > will have the host role, leaving the switch to play 'PHY'. I'm not
> > sure you could even embedded stmmac in a switch, where it might want
> > to play 'PHY', because stmmac is software driven, where as a switch is
> > all hardware.
> >
> > So the hardware supports reverse SGMII, but it is not clear to me why
> > you would want to use it.
> >
> >         Andrew
> >
> 
> Cause I couldn't make 1000BaseX work with qca-ssdk, so I can only
> confirm and test REVSGMII mode on my device.

I think it would help if you could show what you tried for 1000base-X
in terms of dts fragments for both ends.

Marvell DSA switches support 1000base-X, but it defaults to link-down
and without AN, so expecting in-band to work with it doesn't result
in a working link, but using fixed-link on both ends does.

Maybe qca-ssdk needs that as well? Is that what you tried?

It could also be a buggy MAC driver that doesn't disable in-band AN
for 1000base-X in fixed link mode.

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

^ permalink raw reply	[flat|nested] 19+ messages in thread

end of thread, other threads:[~2025-09-26 18:06 UTC | newest]

Thread overview: 19+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-09-22 13:11 [PATCH net-next v11 0/5] net: dsa: yt921x: Add support for Motorcomm YT921x David Yang
2025-09-22 13:11 ` [PATCH net-next v11 1/5] dt-bindings: ethernet-phy: add reverse SGMII phy interface type David Yang
2025-09-23 13:46   ` Rob Herring (Arm)
2025-09-22 13:11 ` [PATCH net-next v11 2/5] net: phy: introduce PHY_INTERFACE_MODE_REVSGMII David Yang
2025-09-23 14:32   ` Maxime Chevallier
2025-09-24 17:50   ` Russell King (Oracle)
2025-09-24 18:41     ` Andrew Lunn
2025-09-24 19:18       ` Russell King (Oracle)
2025-09-26  6:30         ` Yangfl
2025-09-26  8:45           ` Russell King (Oracle)
2025-09-26 16:09           ` Andrew Lunn
2025-09-26 16:28             ` Yangfl
2025-09-26 18:06               ` Russell King (Oracle)
2025-09-22 13:11 ` [PATCH net-next v11 3/5] dt-bindings: net: dsa: yt921x: Add Motorcomm YT921x switch support David Yang
2025-09-22 13:11 ` [PATCH net-next v11 4/5] net: dsa: tag_yt921x: add support for Motorcomm YT921x tags David Yang
2025-09-22 13:11 ` [PATCH net-next v11 5/5] net: dsa: yt921x: Add support for Motorcomm YT921x David Yang
2025-09-24  0:47   ` Jakub Kicinski
2025-09-24 12:33     ` Yangfl
2025-09-25  0:35       ` Jakub Kicinski

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).