Skip to contentSkip to conversation

add support for Renesas RZ/N1 ethernet subsystem devices

PATCH Superseded Clément Léger 5 participants 40 messages May 19, 2022 – May 24, 2022

Subsystems (6)

v5 of 9 Superseded Latest is v9 →

lore
Share & tools
lore

Clément Léger

author maintainer
#0 lore raw

v5 cover add support for Renesas RZ/N1 ethernet subsystem devices

The Renesas RZ/N1 SoCs features an ethernet subsystem which contains (most notably) a switch, two GMACs, and a MII converter [1]. This series adds support for the switch and the MII converter.

The MII converter present on this SoC has been represented as a PCS which sit between the MACs and the PHY. This PCS driver is probed from the device-tree since it requires to be configured. Indeed the MII converter also contains the registers that are handling the muxing of ports (Switch, MAC, HSR, RTOS, etc) internally to the SoC.

The switch driver is based on DSA and exposes 4 ports + 1 CPU management port. It include basic bridging support as well as FDB and statistics support.

This series needs commits 14f11da778ff6421 ("soc: renesas: rzn1: Select PM and PM_GENERIC_DOMAINS configs") and ed66b37f916ee23b ("ARM: dts:
r9a06g032: Add missing '#power-domain-cells'") which are available on the renesas-devel tree in order to enable generic power domain on RZ/N1.

-----
Changes in V5:
- MAINTAINERS:
- Add Florian Fainelli Reviewed-by
- Switch:
- Switch Lookup table lock to a mutex instead of a spinlock
- Only handle "ethernet-ports" property for switch ports
- Handle RGMII_ID/RXID/TXID
- Add check for pdata to be non null in remiove
- Add missing of_node_put() for mdio and ports
- Applied Florian Fainelli patch which makes stats description shorter
- Add Kconfig dependency on ARCH_RZN1 to avoid Kconfig "unmet direct dependency"
- PCS:
- Handle RGMII_ID/RXID/TXID
- Use value instead of BIT() for speed/mode
- Tag driver:
- Add Florian Fainelli Reviewed-by

Changes in V4:
- Add ETH_P_DSA_A5PSW in uapi/linux/if_ether.h
- PCS:
- Use devm_pm_runtime_enable() instead of pm_runtime_enable()
- Switch:
- Return -EOPNOTSUPP and set extack when multiple bdridge are created
- Remove error messages in fdb_del if entry does not exists
- Add compatibility with "ethernet-ports" device-tree property
- Tag driver:
- Use ETH_ZLEN as padding len

Changes in V3:
- PCS:
- Fixed reverse christmas tree declaration
- Remove spurious pr_err
- Use pm_runtime functions
- Tag driver:
- Remove packed attribute from the tag struct
- Switch:
- Fix missing spin_unlock in fdb_dump in case of error
- Add static qualifier to dsa_switch_ops
- Add missing documentation for hclk and clk members of struct a5psw
- Changed types of fdb_entry to u16 to discard GCC note on char packed bitfields and add reserved field
- Added Reviewed-by tag from Florian Fainelli

Changes in V2:
- PCS:
- Fix Reverse Christmas tree declaration
- Removed stray newline
- Add PCS remove function and disable clocks in them
- Fix miic_validate function to return correct values
- Split PCS CONV_MODE definition
- Reordered phylink_pcs_ops in definition order
- Remove interface setting in miic_link_up
- Remove useless checks for invalid interface/speed and error prints
- Replace phylink_pcs_to_miic_port macro by a static function
- Add comment in miic_probe about platform_set_drvdata
- Bindings:
- Fix wrong path for mdio.yaml $ref
- Fix yamllint errors
- Tag driver:
- Squashed commit that added tag value with tag driver
- Add BUILD_BUG_ON for tag size
- Split control_data2 in 2 16bits values
- Switch:
- Use .phylink_get_caps instead of .phylink_validate and fill supported_interface correctly
- Use fixed size (ETH_GSTRING_LEN) string for stats and use memcpy
- Remove stats access locking since RTNL lock is used in upper layers
- Check for non C45 addresses in mdio_read/write and return -EOPNOTSUPP
- Add get_eth_mac_stats, get_eth_mac_ctrl_stat, get_rmon_stats
- Fix a few indentation problems
- Remove reset callback from MDIO bus operation
- Add phy/mac/rmon stats
- Add get_rmon_stat to dsa_ops

Clément Léger (13):
net: dsa: allow port_bridge_join() to override extack message
net: dsa: add support for ethtool get_rmon_stats()
net: dsa: add Renesas RZ/N1 switch tag driver
dt-bindings: net: pcs: add bindings for Renesas RZ/N1 MII converter
net: pcs: add Renesas MII converter driver
dt-bindings: net: dsa: add bindings for Renesas RZ/N1 Advanced 5 port
switch
net: dsa: rzn1-a5psw: add Renesas RZ/N1 advanced 5 port switch driver
net: dsa: rzn1-a5psw: add statistics support
net: dsa: rzn1-a5psw: add FDB support
ARM: dts: r9a06g032: describe MII converter
ARM: dts: r9a06g032: describe GMAC2
ARM: dts: r9a06g032: describe switch
MAINTAINERS: add Renesas RZ/N1 switch related driver entry

 .../bindings/net/dsa/renesas,rzn1-a5psw.yaml  |  131 ++
 .../bindings/net/pcs/renesas,rzn1-miic.yaml   |  162 +++
 MAINTAINERS                                   |   11 +
 arch/arm/boot/dts/r9a06g032.dtsi              |   64 +
 drivers/net/dsa/Kconfig                       |    9 +
 drivers/net/dsa/Makefile                      |    1 +
 drivers/net/dsa/rzn1_a5psw.c                  | 1063 +++++++++++++++++
 drivers/net/dsa/rzn1_a5psw.h                  |  259 ++++
 drivers/net/pcs/Kconfig                       |    8 +
 drivers/net/pcs/Makefile                      |    1 +
 drivers/net/pcs/pcs-rzn1-miic.c               |  516 ++++++++
 include/dt-bindings/net/pcs-rzn1-miic.h       |   33 +
 include/linux/pcs-rzn1-miic.h                 |   18 +
 include/net/dsa.h                             |    5 +
 include/uapi/linux/if_ether.h                 |    1 +
 net/dsa/Kconfig                               |    7 +
 net/dsa/Makefile                              |    1 +
 net/dsa/slave.c                               |   18 +-
 net/dsa/tag_rzn1_a5psw.c                      |  113 ++
 19 files changed, 2419 insertions(+), 2 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml
 create mode 100644 Documentation/devicetree/bindings/net/pcs/renesas,rzn1-miic.yaml
 create mode 100644 drivers/net/dsa/rzn1_a5psw.c
 create mode 100644 drivers/net/dsa/rzn1_a5psw.h
 create mode 100644 drivers/net/pcs/pcs-rzn1-miic.c
 create mode 100644 include/dt-bindings/net/pcs-rzn1-miic.h
 create mode 100644 include/linux/pcs-rzn1-miic.h
 create mode 100644 net/dsa/tag_rzn1_a5psw.c
Signature · 1 line
2.36.0
author maintainer
#1 lore raw

PATCH v5 1/13 net: dsa: allow port_bridge_join() to override extack message

Some drivers might report that they are unable to bridge ports by returning -EOPNOTSUPP, but still wants to override extack message. In order to do so, in dsa_slave_changeupper(), if port_bridge_join() returns -EOPNOTSUPP, check if extack message is set and if so, do not override it.

---
 net/dsa/slave.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
net/dsa/slave.c +3 -2
index 801a5d445833..291197859cea 100644
@@ -2460,8 +2460,9 @@ static int dsa_slave_changeupper(struct net_device *dev,
 			if (!err)
 				dsa_bridge_mtu_normalization(dp);
 			if (err == -EOPNOTSUPP) {
-				NL_SET_ERR_MSG_MOD(extack,
-						   "Offloading not supported");
+				if (!extack->_msg)
+					NL_SET_ERR_MSG_MOD(extack,
+							   "Offloading not supported");
 				err = 0;
 			}
 			err = notifier_from_errno(err);
Signature · 1 line
2.36.0
maintainer Reviewed-by on v5 1/13
#2 lore raw

Re: [PATCH net-next v5 01/13] net: dsa: allow port_bridge_join() to override extack message

Quoted 8 lines · Clément Léger

Some drivers might report that they are unable to bridge ports by returning -EOPNOTSUPP, but still wants to override extack message. In order to do so, in dsa_slave_changeupper(), if port_bridge_join() returns -EOPNOTSUPP, check if extack message is set and if so, do not override it.

---

author maintainer
#3 lore raw

PATCH v5 2/13 net: dsa: add support for ethtool get_rmon_stats()

In reply to Clément Léger

Add support to allow dsa drivers to specify the .get_rmon_stats() operation.

---
 include/net/dsa.h |  3 +++
 net/dsa/slave.c   | 13 +++++++++++++
 2 files changed, 16 insertions(+)
include/net/dsa.h +3 -0
index 14f07275852b..64da5ed27fdc 100644
@@ -888,6 +888,9 @@ struct dsa_switch_ops {
 				     struct ethtool_eth_mac_stats *mac_stats);
 	void	(*get_eth_ctrl_stats)(struct dsa_switch *ds, int port,
 				      struct ethtool_eth_ctrl_stats *ctrl_stats);
+	void	(*get_rmon_stats)(struct dsa_switch *ds, int port,
+				  struct ethtool_rmon_stats *rmon_stats,
+				  const struct ethtool_rmon_hist_range **ranges);
 	void	(*get_stats64)(struct dsa_switch *ds, int port,
 				   struct rtnl_link_stats64 *s);
 	void	(*self_test)(struct dsa_switch *ds, int port,
net/dsa/slave.c +13 -0
index 291197859cea..6cfe07ce06d6 100644
@@ -1002,6 +1002,18 @@ dsa_slave_get_eth_ctrl_stats(struct net_device *dev,
 		ds->ops->get_eth_ctrl_stats(ds, dp->index, ctrl_stats);
 }
 
+static void
+dsa_slave_get_rmon_stats(struct net_device *dev,
+			 struct ethtool_rmon_stats *rmon_stats,
+			 const struct ethtool_rmon_hist_range **ranges)
+{
+	struct dsa_port *dp = dsa_slave_to_port(dev);
+	struct dsa_switch *ds = dp->ds;
+
+	if (ds->ops->get_rmon_stats)
+		ds->ops->get_rmon_stats(ds, dp->index, rmon_stats, ranges);
+}
+
 static void dsa_slave_net_selftest(struct net_device *ndev,
 				   struct ethtool_test *etest, u64 *buf)
 {
@@ -2081,6 +2093,7 @@ static const struct ethtool_ops dsa_slave_ethtool_ops = {
 	.get_eth_phy_stats	= dsa_slave_get_eth_phy_stats,
 	.get_eth_mac_stats	= dsa_slave_get_eth_mac_stats,
 	.get_eth_ctrl_stats	= dsa_slave_get_eth_ctrl_stats,
+	.get_rmon_stats		= dsa_slave_get_rmon_stats,
 	.set_wol		= dsa_slave_set_wol,
 	.get_wol		= dsa_slave_get_wol,
 	.set_eee		= dsa_slave_set_eee,
Signature · 1 line
2.36.0
author maintainer
#4 lore raw

PATCH v5 3/13 net: dsa: add Renesas RZ/N1 switch tag driver

In reply to Clément Léger

The switch that is present on the Renesas RZ/N1 SoC uses a specific VLAN value followed by 6 bytes which contains forwarding configuration.

---
 include/net/dsa.h             |   2 +
 include/uapi/linux/if_ether.h |   1 +
 net/dsa/Kconfig               |   7 +++
 net/dsa/Makefile              |   1 +
 net/dsa/tag_rzn1_a5psw.c      | 113 ++++++++++++++++++++++++++++++++++
 5 files changed, 124 insertions(+)
 create mode 100644 net/dsa/tag_rzn1_a5psw.c
include/net/dsa.h +2 -0
index 64da5ed27fdc..33283eeda697 100644
@@ -53,6 +53,7 @@ struct phylink_link_state;
 #define DSA_TAG_PROTO_SJA1110_VALUE		23
 #define DSA_TAG_PROTO_RTL8_4_VALUE		24
 #define DSA_TAG_PROTO_RTL8_4T_VALUE		25
+#define DSA_TAG_PROTO_RZN1_A5PSW_VALUE		26
 
 enum dsa_tag_protocol {
 	DSA_TAG_PROTO_NONE		= DSA_TAG_PROTO_NONE_VALUE,
@@ -81,6 +82,7 @@ enum dsa_tag_protocol {
 	DSA_TAG_PROTO_SJA1110		= DSA_TAG_PROTO_SJA1110_VALUE,
 	DSA_TAG_PROTO_RTL8_4		= DSA_TAG_PROTO_RTL8_4_VALUE,
 	DSA_TAG_PROTO_RTL8_4T		= DSA_TAG_PROTO_RTL8_4T_VALUE,
+	DSA_TAG_PROTO_RZN1_A5PSW	= DSA_TAG_PROTO_RZN1_A5PSW_VALUE,
 };
 
 struct dsa_switch;
include/uapi/linux/if_ether.h +1 -0
index 1d0bccc3fa54..d370165bc621 100644
@@ -116,6 +116,7 @@
 #define ETH_P_QINQ3	0x9300		/* deprecated QinQ VLAN [ 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 ] */
 #define ETH_P_IFE	0xED3E		/* ForCES inter-FE LFB type */
 #define ETH_P_AF_IUCV   0xFBFB		/* IBM af_iucv [ NOT AN OFFICIALLY REGISTERED ID ] */
 
net/dsa/Kconfig +7 -0
index 8cb87b5067ee..63853fff4e2f 100644
@@ -132,6 +132,13 @@ config NET_DSA_TAG_RTL8_4
 	  Say Y or M if you want to enable support for tagging frames for Realtek
 	  switches with 8 byte protocol 4 tags, such as the Realtek RTL8365MB-VC.
 
+config NET_DSA_TAG_RZN1_A5PSW
+	tristate "Tag driver for Renesas RZ/N1 A5PSW switch"
+	help
+	  Say Y or M if you want to enable support for tagging frames for
+	  Renesas RZ/N1 embedded switch that uses an 8 byte tag located after
+	  destination MAC address.
+
 config NET_DSA_TAG_LAN9303
 	tristate "Tag driver for SMSC/Microchip LAN9303 family of switches"
 	help
net/dsa/Makefile +1 -0
index 9f75820e7c98..af28c24ead18 100644
@@ -17,6 +17,7 @@ obj-$(CONFIG_NET_DSA_TAG_OCELOT_8021Q) += tag_ocelot_8021q.o
 obj-$(CONFIG_NET_DSA_TAG_QCA) += tag_qca.o
 obj-$(CONFIG_NET_DSA_TAG_RTL4_A) += tag_rtl4_a.o
 obj-$(CONFIG_NET_DSA_TAG_RTL8_4) += tag_rtl8_4.o
+obj-$(CONFIG_NET_DSA_TAG_RZN1_A5PSW) += tag_rzn1_a5psw.o
 obj-$(CONFIG_NET_DSA_TAG_SJA1105) += tag_sja1105.o
 obj-$(CONFIG_NET_DSA_TAG_TRAILER) += tag_trailer.o
 obj-$(CONFIG_NET_DSA_TAG_XRS700X) += tag_xrs700x.o
net/dsa/tag_rzn1_a5psw.c new file +113 -0

113 lines not on this page · raw message

Signature · 1 line
2.36.0
author maintainer
#5 lore raw

PATCH v5 4/13 dt-bindings: net: pcs: add bindings for Renesas RZ/N1 MII converter

In reply to Clément Léger

This MII converter can be found on the RZ/N1 processor family. The MII converter ports are declared as subnodes which are then referenced by users of the PCS driver such as the switch.

---
 .../bindings/net/pcs/renesas,rzn1-miic.yaml   | 162 ++++++++++++++++++
 include/dt-bindings/net/pcs-rzn1-miic.h       |  33 ++++
 2 files changed, 195 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/net/pcs/renesas,rzn1-miic.yaml
 create mode 100644 include/dt-bindings/net/pcs-rzn1-miic.h
Documentation/devicetree/bindings/net/pcs/renesas,rzn1-miic.yaml new file +162 -0
new file mode 100644
index 000000000000..81a2f7d414a6
@@ -0,0 +1,162 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/net/pcs/renesas,rzn1-miic.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Renesas RZ/N1 MII converter
+
+maintainers:
+  - Clément Léger <[email protected]>
+
+description: |
+  This MII converter is present on the Renesas RZ/N1 SoC family. It is
+  responsible to do MII passthrough or convert it to RMII/RGMII.
+
+properties:
+  '#address-cells':
+    const: 1
+
+  '#size-cells':
+    const: 0
+
+  compatible:
+    items:
+      - enum:
+          - renesas,r9a06g032-miic
+      - const: renesas,rzn1-miic
+
+  reg:
+    maxItems: 1
+
+  clocks:
+    items:
+      - description: MII reference clock
+      - description: RGMII reference clock
+      - description: RMII reference clock
+      - description: AHB clock used for the MII converter register interface
+
+  renesas,miic-switch-portin:
+    description: MII Switch PORTIN configuration. This value should use one of
+      the values defined in dt-bindings/net/pcs-rzn1-miic.h.
+    $ref: /schemas/types.yaml#/definitions/uint32
+    enum: [1, 2]
+
+  power-domains:
+    maxItems: 1
+
+patternProperties:
+  "^mii-conv@[0-5]$":
+    type: object
+    description: MII converter port
+
+    properties:
+      reg:
+        description: MII Converter port number.
+        enum: [1, 2, 3, 4, 5]
+
+      renesas,miic-input:
+        description: Converter input port configuration. This value should use
+          one of the values defined in dt-bindings/net/pcs-rzn1-miic.h.
+        $ref: /schemas/types.yaml#/definitions/uint32
+
+    required:
+      - reg
+      - renesas,miic-input
+
+    additionalProperties: false
+
+    allOf:
+      - if:
+          properties:
+            reg:
+              const: 1
+        then:
+          properties:
+            renesas,miic-input:
+              const: 0
+      - if:
+          properties:
+            reg:
+              const: 2
+        then:
+          properties:
+            renesas,miic-input:
+              enum: [1, 11]
+      - if:
+          properties:
+            reg:
+              const: 3
+        then:
+          properties:
+            renesas,miic-input:
+              enum: [7, 10]
+      - if:
+          properties:
+            reg:
+              const: 4
+        then:
+          properties:
+            renesas,miic-input:
+              enum: [4, 6, 9, 13]
+      - if:
+          properties:
+            reg:
+              const: 5
+        then:
+          properties:
+            renesas,miic-input:
+              enum: [3, 5, 8, 12]
+
+required:
+  - '#address-cells'
+  - '#size-cells'
+  - compatible
+  - reg
+  - clocks
+  - power-domains
+
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/net/pcs-rzn1-miic.h>
+    #include <dt-bindings/clock/r9a06g032-sysctrl.h>
+
+    eth-miic@44030000 {
+      #address-cells = <1>;
+      #size-cells = <0>;
+      compatible = "renesas,r9a06g032-miic", "renesas,rzn1-miic";
+      reg = <0x44030000 0x10000>;
+      clocks = <&sysctrl R9A06G032_CLK_MII_REF>,
+              <&sysctrl R9A06G032_CLK_RGMII_REF>,
+              <&sysctrl R9A06G032_CLK_RMII_REF>,
+              <&sysctrl R9A06G032_HCLK_SWITCH_RG>;
+      renesas,miic-switch-portin = <MIIC_GMAC2_PORT>;
+      power-domains = <&sysctrl>;
+
+      mii_conv1: mii-conv@1 {
+        renesas,miic-input = <MIIC_GMAC1_PORT>;
+        reg = <1>;
+      };
+
+      mii_conv2: mii-conv@2 {
+        renesas,miic-input = <MIIC_SWITCH_PORTD>;
+        reg = <2>;
+      };
+
+      mii_conv3: mii-conv@3 {
+        renesas,miic-input = <MIIC_SWITCH_PORTC>;
+        reg = <3>;
+      };
+
+      mii_conv4: mii-conv@4 {
+        renesas,miic-input = <MIIC_SWITCH_PORTB>;
+        reg = <4>;
+      };
+
+      mii_conv5: mii-conv@5 {
+        renesas,miic-input = <MIIC_SWITCH_PORTA>;
+        reg = <5>;
+      };
+    };
include/dt-bindings/net/pcs-rzn1-miic.h new file +33 -0
new file mode 100644
index 000000000000..784782eaec9e
@@ -0,0 +1,33 @@
+/* SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) */
+/*
+ * Copyright (C) 2022 Schneider-Electric
+ *
+ * Clément Léger <[email protected]>
+ */
+
+#ifndef _DT_BINDINGS_PCS_RZN1_MIIC
+#define _DT_BINDINGS_PCS_RZN1_MIIC
+
+/*
+ * Reefer to the datasheet [1] section 8.2.1, Internal Connection of Ethernet
+ * Ports to check the available combination
+ *
+ * [1] REN_r01uh0750ej0140-rzn1-introduction_MAT_20210228.pdf
+ */
+
+#define MIIC_GMAC1_PORT			0
+#define MIIC_GMAC2_PORT			1
+#define MIIC_RTOS_PORT			2
+#define MIIC_SERCOS_PORTA		3
+#define MIIC_SERCOS_PORTB		4
+#define MIIC_ETHERCAT_PORTA		5
+#define MIIC_ETHERCAT_PORTB		6
+#define MIIC_ETHERCAT_PORTC		7
+#define MIIC_SWITCH_PORTA		8
+#define MIIC_SWITCH_PORTB		9
+#define MIIC_SWITCH_PORTC		10
+#define MIIC_SWITCH_PORTD		11
+#define MIIC_HSR_PORTA			12
+#define MIIC_HSR_PORTB			13
+
+#endif
Signature · 1 line
2.36.0
Reviewed-by on v5 4/13
#6 lore raw

Re: [PATCH net-next v5 04/13] dt-bindings: net: pcs: add bindings for Renesas RZ/N1 MII converter

Hi Clément,

Quoted 7 lines · Clément Léger

This MII converter can be found on the RZ/N1 processor family. The MII converter ports are declared as subnodes which are then referenced by users of the PCS driver such as the switch.

Thanks for your patch!

Quoted 30 lines
Documentation/devicetree/bindings/net/pcs/renesas,rzn1-miic.yaml new file +27 -0
@@ -0,0 +1,162 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/net/pcs/renesas,rzn1-miic.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Renesas RZ/N1 MII converter
+
+maintainers:
+  - Clément Léger <[email protected]>
+
+description: |
+  This MII converter is present on the Renesas RZ/N1 SoC family. It is
+  responsible to do MII passthrough or convert it to RMII/RGMII.
+
+properties:
+  '#address-cells':
+    const: 1
+
+  '#size-cells':
+    const: 0
+
+  compatible:
+    items:
+      - enum:
+          - renesas,r9a06g032-miic
+      - const: renesas,rzn1-miic

Documentation/devicetree/bindings/net/pcs/renesas,rzn1-miic.yaml line 37

Quoted 10 lines
Documentation/devicetree/bindings/net/pcs/renesas,rzn1-miic.yaml new file +10 -0
+
+  reg:
+    maxItems: 1
+
+  clocks:
+    items:
+      - description: MII reference clock
+      - description: RGMII reference clock
+      - description: RMII reference clock
+      - description: AHB clock used for the MII converter register interface

Please add clock-names (and make it required), as there are multiple clocks.

The rest LGTM (from an SoC integration PoV), so

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

author maintainer
#7 lore raw

PATCH v5 5/13 net: pcs: add Renesas MII converter driver

In reply to Clément Léger

Add a PCS driver for the MII converter that is present on the Renesas RZ/N1 SoC. This MII converter is reponsible for converting MII to RMII/RGMII or act as a MII pass-trough. Exposing it as a PCS allows to reuse it in both the switch driver and the stmmac driver. Currently, this driver only allows the PCS to be used by the dual Cortex-A7 subsystem since the register locking system is not used.

---
 drivers/net/pcs/Kconfig         |   8 +
 drivers/net/pcs/Makefile        |   1 +
 drivers/net/pcs/pcs-rzn1-miic.c | 516 ++++++++++++++++++++++++++++++++
 include/linux/pcs-rzn1-miic.h   |  18 ++
 4 files changed, 543 insertions(+)
 create mode 100644 drivers/net/pcs/pcs-rzn1-miic.c
 create mode 100644 include/linux/pcs-rzn1-miic.h
drivers/net/pcs/Kconfig +8 -0
index 22ba7b0b476d..6a500d30ca4c 100644
@@ -18,4 +18,12 @@ config PCS_LYNX
 	  This module provides helpers to phylink for managing the Lynx PCS
 	  which is part of the Layerscape and QorIQ Ethernet SERDES.
 
+config PCS_RZN1_MIIC
+	tristate "Renesas RZ/N1 MII converter"
+	depends on ARCH_RZN1 || COMPILE_TEST
+	help
+	  This module provides a driver for the MII converter that is available
+	  on RZ/N1 SoCs. This PCS converts MII to RMII/RGMII or can be set in
+	  pass-through mode for MII.
+
 endmenu
drivers/net/pcs/Makefile +1 -0
index 0603d469bd57..0ff5388fcdea 100644
@@ -5,3 +5,4 @@ pcs_xpcs-$(CONFIG_PCS_XPCS)	:= pcs-xpcs.o pcs-xpcs-nxp.o
 
 obj-$(CONFIG_PCS_XPCS)		+= pcs_xpcs.o
 obj-$(CONFIG_PCS_LYNX)		+= pcs-lynx.o
+obj-$(CONFIG_PCS_RZN1_MIIC)	+= pcs-rzn1-miic.o
drivers/net/pcs/pcs-rzn1-miic.c new file +516 -0

516 lines not on this page · raw message

include/linux/pcs-rzn1-miic.h new file +18 -0
new file mode 100644
index 000000000000..7a303af57e40
@@ -0,0 +1,18 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (C) 2022 Schneider Electric
+ *
+ * Clément Léger <[email protected]>
+ */
+
+#ifndef __LINUX_PCS_MIIC_H
+#define __LINUX_PCS_MIIC_H
+
+struct phylink;
+struct device_node;
+
+struct phylink_pcs *miic_create(struct device_node *np);
+
+void miic_destroy(struct phylink_pcs *pcs);
+
+#endif /* __LINUX_PCS_MIIC_H */
Signature · 1 line
2.36.0
#8 lore raw

Re: [PATCH net-next v5 05/13] net: pcs: add Renesas MII converter driver

Hi,

Quoted 8 lines · Clément Léger

Add a PCS driver for the MII converter that is present on the Renesas RZ/N1 SoC. This MII converter is reponsible for converting MII to RMII/RGMII or act as a MII pass-trough. Exposing it as a PCS allows to reuse it in both the switch driver and the stmmac driver. Currently, this driver only allows the PCS to be used by the dual Cortex-A7 subsystem since the register locking system is not used.

Looks much better now, thanks. Only one thing I've spotted is:

drivers/net/pcs/pcs-rzn1-miic.c line 268

Quoted 7 lines
drivers/net/pcs/pcs-rzn1-miic.c new file +7 -0
+static int miic_validate(struct phylink_pcs *pcs, unsigned long *supported,
+			 const struct phylink_link_state *state)
+{
+	if (state->interface == PHY_INTERFACE_MODE_RGMII ||
+	    state->interface == PHY_INTERFACE_MODE_RGMII_ID ||
+	    state->interface == PHY_INTERFACE_MODE_RGMII_TXID ||
+	    state->interface == PHY_INTERFACE_MODE_RGMII_RXID ||

The above could use:

	if (phy_interface_mode_is_rgmii(state->interface) ||

Also, as a request to unbind this driver would be disasterous to users, I think you should set ".suppress_bind_attrs = true" to prevent the sysfs bind/unbind facility being available. This doesn't completely solve the problem.

Thanks.

Signature · 2 lines
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!
author maintainer
#9 lore raw

Re: [PATCH net-next v5 05/13] net: pcs: add Renesas MII converter driver

Quoted 24 lines · Russell King (Oracle)

Hi,

Quoted 8 lines · Clément Léger

Add a PCS driver for the MII converter that is present on the Renesas RZ/N1 SoC. This MII converter is reponsible for converting MII to RMII/RGMII or act as a MII pass-trough. Exposing it as a PCS allows to reuse it in both the switch driver and the stmmac driver. Currently, this driver only allows the PCS to be used by the dual Cortex-A7 subsystem since the register locking system is not used.

Looks much better now, thanks. Only one thing I've spotted is:

drivers/net/pcs/pcs-rzn1-miic.c line 268

Quoted 7 lines
drivers/net/pcs/pcs-rzn1-miic.c new file +7 -0
+static int miic_validate(struct phylink_pcs *pcs, unsigned long *supported,
+			 const struct phylink_link_state *state)
+{
+	if (state->interface == PHY_INTERFACE_MODE_RGMII ||
+	    state->interface == PHY_INTERFACE_MODE_RGMII_ID ||
+	    state->interface == PHY_INTERFACE_MODE_RGMII_TXID ||
+	    state->interface == PHY_INTERFACE_MODE_RGMII_RXID ||  

The above could use:

Quoted 1 line
	if (phy_interface_mode_is_rgmii(state->interface) ||

Thanks, I did found the one to set the bit for phylink part but not this one.

Quoted 5 lines

Also, as a request to unbind this driver would be disasterous to users, I think you should set ".suppress_bind_attrs = true" to prevent the sysfs bind/unbind facility being available. This doesn't completely solve the problem.

Acked. What should I do to make it more robust ? Should I use a refcount per pdev and check that in the remove() callback to avoid removing the pdev if used ?

Thanks,

Signature · 3 lines
Clément Léger,
Embedded Linux and Kernel engineer at Bootlin
https://bootlin.com
maintainer
#10 lore raw

Re: [PATCH net-next v5 05/13] net: pcs: add Renesas MII converter driver

Quoted 8 lines · Clément Léger
Quoted 4 lines

Also, as a request to unbind this driver would be disasterous to users, I think you should set ".suppress_bind_attrs = true" to prevent the sysfs bind/unbind facility being available. This doesn't completely solve the problem.

Acked. What should I do to make it more robust ? Should I use a refcount per pdev and check that in the remove() callback to avoid removing the pdev if used ?

I wonder, if you call device_link_add(ds->dev, miic->dev, DL_FLAG_AUTOREMOVE_CONSUMER), wouldn't that be enough to auto-unbind the DSA driver when the MII converter driver unbinds?

author maintainer
#11 lore raw

Re: [PATCH net-next v5 05/13] net: pcs: add Renesas MII converter driver

Quoted 6 lines · Vladimir Oltean
Quoted 5 lines · Clément Léger
Quoted 4 lines

Also, as a request to unbind this driver would be disasterous to users, I think you should set ".suppress_bind_attrs = true" to prevent the sysfs bind/unbind facility being available. This doesn't completely solve the problem.

Quoted 7 lines
Quoted 3 lines

Acked. What should I do to make it more robust ? Should I use a refcount per pdev and check that in the remove() callback to avoid removing the pdev if used ?

I wonder, if you call device_link_add(ds->dev, miic->dev, DL_FLAG_AUTOREMOVE_CONSUMER), wouldn't that be enough to auto-unbind the DSA driver when the MII converter driver unbinds?

I looiked at that a bit and I'm not sure how to achieve that cleanly. If I need to create this link, then I need to do it once for the dsa switch device. However, currently, the way I get the references to the MII converter are via the pcs-handle properties which are for each port.

So, I'm not sure creating the link multiple times in miic_create() would be ok and also, I'm not sure how to create the link once without adding a specific property which points on the MII converter node and use that to create the link by adding miic_device_add_link() for instance.

Do you have any preference ?

Thanks,

Signature · 3 lines
Clément Léger,
Embedded Linux and Kernel engineer at Bootlin
https://bootlin.com
maintainer
#12 lore raw

Re: [PATCH net-next v5 05/13] net: pcs: add Renesas MII converter driver

Quoted 22 lines · Clément Léger
Quoted 6 lines · Vladimir Oltean
Quoted 5 lines · Clément Léger
Quoted 4 lines

Also, as a request to unbind this driver would be disasterous to users, I think you should set ".suppress_bind_attrs = true" to prevent the sysfs bind/unbind facility being available. This doesn't completely solve the problem.

Quoted 7 lines
Quoted 3 lines

Acked. What should I do to make it more robust ? Should I use a refcount per pdev and check that in the remove() callback to avoid removing the pdev if used ?

I wonder, if you call device_link_add(ds->dev, miic->dev, DL_FLAG_AUTOREMOVE_CONSUMER), wouldn't that be enough to auto-unbind the DSA driver when the MII converter driver unbinds?

I looiked at that a bit and I'm not sure how to achieve that cleanly. If I need to create this link, then I need to do it once for the dsa switch device. However, currently, the way I get the references to the MII converter are via the pcs-handle properties which are for each port.

Quoted 6 lines

So, I'm not sure creating the link multiple times in miic_create() would be ok and also, I'm not sure how to create the link once without adding a specific property which points on the MII converter node and use that to create the link by adding miic_device_add_link() for instance.

Do you have any preference ?

The simplest (although not the most elegant) way would probably be to pass the ds->dev as a second argument to miic_create(), and call device_link_add() multiple times, letting all but the first call fail, and ignoring the resulting NULL return code. Maybe others have a better idea.

author maintainer
#13 lore raw

Re: [PATCH net-next v5 05/13] net: pcs: add Renesas MII converter driver

Quoted 28 lines · Vladimir Oltean
Quoted 27 lines · Clément Léger
Quoted 6 lines · Vladimir Oltean
Quoted 5 lines · Clément Léger
Quoted 4 lines

Also, as a request to unbind this driver would be disasterous to users, I think you should set ".suppress_bind_attrs = true" to prevent the sysfs bind/unbind facility being available. This doesn't completely solve the problem.

Quoted 7 lines
Quoted 3 lines

Acked. What should I do to make it more robust ? Should I use a refcount per pdev and check that in the remove() callback to avoid removing the pdev if used ?

I wonder, if you call device_link_add(ds->dev, miic->dev, DL_FLAG_AUTOREMOVE_CONSUMER), wouldn't that be enough to auto-unbind the DSA driver when the MII converter driver unbinds?

I looiked at that a bit and I'm not sure how to achieve that cleanly. If I need to create this link, then I need to do it once for the dsa switch device. However, currently, the way I get the references to the MII converter are via the pcs-handle properties which are for each port.

So, I'm not sure creating the link multiple times in miic_create() would be ok and also, I'm not sure how to create the link once without adding a specific property which points on the MII converter node and use that to create the link by adding miic_device_add_link() for instance.

Quoted 6 lines
Quoted 1 line

Do you have any preference ?

The simplest (although not the most elegant) way would probably be to pass the ds->dev as a second argument to miic_create(), and call device_link_add() multiple times, letting all but the first call fail, and ignoring the resulting NULL return code. Maybe others have a better idea.

That's indeed what I started to do although it's nasty to say the least... Moreover, the device_link_del() calls in miic_destroy() would have to be carefully made after all miic ports have been destroyed. Not sure this is going to be cleaner ! I'll try to think about it a bit more.

Thanks,

Signature · 3 lines
Clément Léger,
Embedded Linux and Kernel engineer at Bootlin
https://bootlin.com
maintainer
#14 lore raw

Re: [PATCH net-next v5 05/13] net: pcs: add Renesas MII converter driver

Quoted 38 lines · Clément Léger
Quoted 28 lines · Vladimir Oltean
Quoted 27 lines · Clément Léger
Quoted 6 lines · Vladimir Oltean
Quoted 5 lines · Clément Léger
Quoted 4 lines

Also, as a request to unbind this driver would be disasterous to users, I think you should set ".suppress_bind_attrs = true" to prevent the sysfs bind/unbind facility being available. This doesn't completely solve the problem.

Quoted 7 lines
Quoted 3 lines

Acked. What should I do to make it more robust ? Should I use a refcount per pdev and check that in the remove() callback to avoid removing the pdev if used ?

I wonder, if you call device_link_add(ds->dev, miic->dev, DL_FLAG_AUTOREMOVE_CONSUMER), wouldn't that be enough to auto-unbind the DSA driver when the MII converter driver unbinds?

I looiked at that a bit and I'm not sure how to achieve that cleanly. If I need to create this link, then I need to do it once for the dsa switch device. However, currently, the way I get the references to the MII converter are via the pcs-handle properties which are for each port.

So, I'm not sure creating the link multiple times in miic_create() would be ok and also, I'm not sure how to create the link once without adding a specific property which points on the MII converter node and use that to create the link by adding miic_device_add_link() for instance.

Quoted 6 lines
Quoted 1 line

Do you have any preference ?

The simplest (although not the most elegant) way would probably be to pass the ds->dev as a second argument to miic_create(), and call device_link_add() multiple times, letting all but the first call fail, and ignoring the resulting NULL return code. Maybe others have a better idea.

Quoted 5 lines

That's indeed what I started to do although it's nasty to say the least... Moreover, the device_link_del() calls in miic_destroy() would have to be carefully made after all miic ports have been destroyed. Not sure this is going to be cleaner ! I'll try to think about it a bit more.

Wait... the whole idea with AUTOREMOVE_CONSUMER is that you _don't_ remove the device link.. you let it sit there such that the device core knows there are other consumers it needs to remove when this driver unbinds from the device.

author maintainer
#15 lore raw

Re: [PATCH net-next v5 05/13] net: pcs: add Renesas MII converter driver

Quoted 45 lines · Vladimir Oltean
Quoted 38 lines · Clément Léger
Quoted 28 lines · Vladimir Oltean
Quoted 27 lines · Clément Léger
Quoted 6 lines · Vladimir Oltean
Quoted 5 lines · Clément Léger
Quoted 4 lines

Also, as a request to unbind this driver would be disasterous to users, I think you should set ".suppress_bind_attrs = true" to prevent the sysfs bind/unbind facility being available. This doesn't completely solve the problem.

Quoted 7 lines
Quoted 3 lines

Acked. What should I do to make it more robust ? Should I use a refcount per pdev and check that in the remove() callback to avoid removing the pdev if used ?

I wonder, if you call device_link_add(ds->dev, miic->dev, DL_FLAG_AUTOREMOVE_CONSUMER), wouldn't that be enough to auto-unbind the DSA driver when the MII converter driver unbinds?

I looiked at that a bit and I'm not sure how to achieve that cleanly. If I need to create this link, then I need to do it once for the dsa switch device. However, currently, the way I get the references to the MII converter are via the pcs-handle properties which are for each port.

So, I'm not sure creating the link multiple times in miic_create() would be ok and also, I'm not sure how to create the link once without adding a specific property which points on the MII converter node and use that to create the link by adding miic_device_add_link() for instance.

Quoted 6 lines
Quoted 1 line

Do you have any preference ?

The simplest (although not the most elegant) way would probably be to pass the ds->dev as a second argument to miic_create(), and call device_link_add() multiple times, letting all but the first call fail, and ignoring the resulting NULL return code. Maybe others have a better idea.

Quoted 5 lines

That's indeed what I started to do although it's nasty to say the least... Moreover, the device_link_del() calls in miic_destroy() would have to be carefully made after all miic ports have been destroyed. Not sure this is going to be cleaner ! I'll try to think about it a bit more.

Quoted 4 lines

Wait... the whole idea with AUTOREMOVE_CONSUMER is that you _don't_ remove the device link.. you let it sit there such that the device core knows there are other consumers it needs to remove when this driver unbinds from the device.

Hum ok, got it, so let's try this solution.

Signature · 3 lines
Clément Léger,
Embedded Linux and Kernel engineer at Bootlin
https://bootlin.com
author maintainer
#16 lore raw

PATCH v5 6/13 dt-bindings: net: dsa: add bindings for Renesas RZ/N1 Advanced 5 port switch

In reply to Clément Léger

Add bindings for Renesas RZ/N1 Advanced 5 port switch. This switch is present on Renesas RZ/N1 SoC and was probably provided by MoreThanIP. This company does not exists anymore and has been bought by Synopsys. Since this IP can't be find anymore in the Synospsy portfolio, lets use Renesas as the vendor compatible for this IP.

---
 .../bindings/net/dsa/renesas,rzn1-a5psw.yaml  | 131 ++++++++++++++++++
 1 file changed, 131 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml
Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml new file +131 -0
new file mode 100644
index 000000000000..e3ab1cf1575d
@@ -0,0 +1,131 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/net/dsa/renesas,rzn1-a5psw.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Renesas RZ/N1 Advanced 5 ports ethernet switch
+
+maintainers:
+  - Clément Léger <[email protected]>
+
+description: |
+  The advanced 5 ports switch is present on the Renesas RZ/N1 SoC family and
+  handles 4 ports + 1 CPU management port.
+
+allOf:
+  - $ref: dsa.yaml#
+
+properties:
+  compatible:
+    items:
+      - enum:
+          - renesas,r9a06g032-a5psw
+      - const: renesas,rzn1-a5psw
+
+  reg:
+    maxItems: 1
+
+  mdio:
+    $ref: /schemas/net/mdio.yaml#
+    unevaluatedProperties: false
+
+  clocks:
+    items:
+      - description: AHB clock used for the switch register interface
+      - description: Switch system clock
+
+  clock-names:
+    items:
+      - const: hclk
+      - const: clk
+
+  ethernet-ports:
+    type: object
+    properties:
+      '#address-cells':
+        const: 1
+      '#size-cells':
+        const: 0
+
+    patternProperties:
+      "^(ethernet-)?port@[0-4]$":
+        type: object
+        description: Ethernet switch ports
+
+        properties:
+          pcs-handle:
+            description:
+              phandle pointing to a PCS sub-node compatible with
+              renesas,rzn1-miic.yaml#
+            $ref: /schemas/types.yaml#/definitions/phandle
+
+unevaluatedProperties: false
+
+required:
+  - compatible
+  - reg
+  - clocks
+  - clock-names
+
+examples:
+  - |
+    #include <dt-bindings/gpio/gpio.h>
+    #include <dt-bindings/clock/r9a06g032-sysctrl.h>
+
+    switch@44050000 {
+        compatible = "renesas,r9a06g032-a5psw", "renesas,rzn1-a5psw";
+        reg = <0x44050000 0x10000>;
+        clocks = <&sysctrl R9A06G032_HCLK_SWITCH>, <&sysctrl R9A06G032_CLK_SWITCH>;
+        clock-names = "hclk", "clk";
+        pinctrl-names = "default";
+        pinctrl-0 = <&pins_mdio1>, <&pins_eth3>, <&pins_eth4>;
+
+        dsa,member = <0 0>;
+
+        ethernet-ports {
+            #address-cells = <1>;
+            #size-cells = <0>;
+
+            port@0 {
+                reg = <0>;
+                label = "lan0";
+                phy-handle = <&switch0phy3>;
+                pcs-handle = <&mii_conv4>;
+            };
+
+            port@1 {
+                reg = <1>;
+                label = "lan1";
+                phy-handle = <&switch0phy1>;
+                pcs-handle = <&mii_conv3>;
+            };
+
+            port@4 {
+                reg = <4>;
+                ethernet = <&gmac1>;
+                label = "cpu";
+                fixed-link {
+                  speed = <1000>;
+                  full-duplex;
+                };
+            };
+        };
+
+        mdio {
+            #address-cells = <1>;
+            #size-cells = <0>;
+
+            reset-gpios = <&gpio0a 2 GPIO_ACTIVE_HIGH>;
+            reset-delay-us = <15>;
+            clock-frequency = <2500000>;
+
+            switch0phy1: ethernet-phy@1{
+                reg = <1>;
+            };
+
+            switch0phy3: ethernet-phy@3{
+                reg = <3>;
+            };
+        };
+    };
Signature · 1 line
2.36.0
Reviewed-by on v5 6/13
#17 lore raw

Re: [PATCH net-next v5 06/13] dt-bindings: net: dsa: add bindings for Renesas RZ/N1 Advanced 5 port switch

Hi Clément,

Quoted 7 lines · Clément Léger

Add bindings for Renesas RZ/N1 Advanced 5 port switch. This switch is present on Renesas RZ/N1 SoC and was probably provided by MoreThanIP. This company does not exists anymore and has been bought by Synopsys. Since this IP can't be find anymore in the Synospsy portfolio, lets use Renesas as the vendor compatible for this IP.

Thanks for your patch!

Quoted 34 lines
Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml new file +31 -0
@@ -0,0 +1,131 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/net/dsa/renesas,rzn1-a5psw.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Renesas RZ/N1 Advanced 5 ports ethernet switch
+
+maintainers:
+  - Clément Léger <[email protected]>
+
+description: |
+  The advanced 5 ports switch is present on the Renesas RZ/N1 SoC family and
+  handles 4 ports + 1 CPU management port.
+
+allOf:
+  - $ref: dsa.yaml#
+
+properties:
+  compatible:
+    items:
+      - enum:
+          - renesas,r9a06g032-a5psw
+      - const: renesas,rzn1-a5psw
+
+  reg:
+    maxItems: 1
+
+  mdio:
+    $ref: /schemas/net/mdio.yaml#
+    unevaluatedProperties: false

Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml line 41

Quoted 10 lines
Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml new file +10 -0
+
+  clocks:
+    items:
+      - description: AHB clock used for the switch register interface
+      - description: Switch system clock
+
+  clock-names:
+    items:
+      - const: hclk
+      - const: clk

(Good, "clock-names" is present ;-)

Missing "power-domains" property.

Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml line 82

Quoted 12 lines
Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml new file +12 -0
+examples:
+  - |
+    #include <dt-bindings/gpio/gpio.h>
+    #include <dt-bindings/clock/r9a06g032-sysctrl.h>
+
+    switch@44050000 {
+        compatible = "renesas,r9a06g032-a5psw", "renesas,rzn1-a5psw";
+        reg = <0x44050000 0x10000>;
+        clocks = <&sysctrl R9A06G032_HCLK_SWITCH>, <&sysctrl R9A06G032_CLK_SWITCH>;
+        clock-names = "hclk", "clk";
+        pinctrl-names = "default";
+        pinctrl-0 = <&pins_mdio1>, <&pins_eth3>, <&pins_eth4>;

Usually we don't list pinctrl-* properties in examples.

The rest LGTM (from an SoC integration PoV), so with the above fixed

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

author maintainer
#18 lore raw

Re: [PATCH net-next v5 06/13] dt-bindings: net: dsa: add bindings for Renesas RZ/N1 Advanced 5 port switch

Quoted 60 lines · Geert Uytterhoeven

Hi Clément,

Quoted 7 lines · Clément Léger

Add bindings for Renesas RZ/N1 Advanced 5 port switch. This switch is present on Renesas RZ/N1 SoC and was probably provided by MoreThanIP. This company does not exists anymore and has been bought by Synopsys. Since this IP can't be find anymore in the Synospsy portfolio, lets use Renesas as the vendor compatible for this IP.

Thanks for your patch!

Quoted 34 lines
Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml new file +31 -0
@@ -0,0 +1,131 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/net/dsa/renesas,rzn1-a5psw.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Renesas RZ/N1 Advanced 5 ports ethernet switch
+
+maintainers:
+  - Clément Léger <[email protected]>
+
+description: |
+  The advanced 5 ports switch is present on the Renesas RZ/N1 SoC family and
+  handles 4 ports + 1 CPU management port.
+
+allOf:
+  - $ref: dsa.yaml#
+
+properties:
+  compatible:
+    items:
+      - enum:
+          - renesas,r9a06g032-a5psw
+      - const: renesas,rzn1-a5psw
+
+  reg:
+    maxItems: 1
+
+  mdio:
+    $ref: /schemas/net/mdio.yaml#
+    unevaluatedProperties: false

Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml line 41

Quoted 10 lines
Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml new file +10 -0
+
+  clocks:
+    items:
+      - description: AHB clock used for the switch register interface
+      - description: Switch system clock
+
+  clock-names:
+    items:
+      - const: hclk
+      - const: clk  

(Good, "clock-names" is present ;-)

Quoted 2 lines

Missing "power-domains" property.

I do not use pm_runtime* in the switch driver. I should probably do that right ?

Quoted 6 lines
Quoted 6 lines
Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml new file +6 -0
+examples:
+  - |
+    #include <dt-bindings/gpio/gpio.h>
+    #include <dt-bindings/clock/r9a06g032-sysctrl.h>
+
+    switch@44050000 {
Quoted 9 lines

Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml line 82

Quoted 6 lines
Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml new file +6 -0
+        compatible = "renesas,r9a06g032-a5psw", "renesas,rzn1-a5psw";
+        reg = <0x44050000 0x10000>;
+        clocks = <&sysctrl R9A06G032_HCLK_SWITCH>, <&sysctrl R9A06G032_CLK_SWITCH>;
+        clock-names = "hclk", "clk";
+        pinctrl-names = "default";
+        pinctrl-0 = <&pins_mdio1>, <&pins_eth3>, <&pins_eth4>;  

Usually we don't list pinctrl-* properties in examples.

Acked, I'll remove that.

Quoted 13 lines

The rest LGTM (from an SoC integration PoV), so with the above fixed

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

Signature · 3 lines
Clément Léger,
Embedded Linux and Kernel engineer at Bootlin
https://bootlin.com
#19 lore raw

Re: [PATCH net-next v5 06/13] dt-bindings: net: dsa: add bindings for Renesas RZ/N1 Advanced 5 port switch

Hi Clément,

Quoted 9 lines · Clément Léger
Quoted 7 lines · Geert Uytterhoeven
Quoted 6 lines · Clément Léger

Add bindings for Renesas RZ/N1 Advanced 5 port switch. This switch is present on Renesas RZ/N1 SoC and was probably provided by MoreThanIP. This company does not exists anymore and has been bought by Synopsys. Since this IP can't be find anymore in the Synospsy portfolio, lets use Renesas as the vendor compatible for this IP.

Quoted 8 lines
Quoted 5 lines

Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml

Quoted 3 lines
Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml new file +0 -0

Missing "power-domains" property.

I do not use pm_runtime* in the switch driver. I should probably do that right ?

For now you don't have to. But I think it is a good idea, and it helps if the IP block is ever reused in an SoC with real power areas.

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

author maintainer
#20 lore raw

Re: [PATCH net-next v5 06/13] dt-bindings: net: dsa: add bindings for Renesas RZ/N1 Advanced 5 port switch

Quoted 18 lines · Geert Uytterhoeven

Hi Clément,

Quoted 15 lines · Clément Léger
Quoted 12 lines · Geert Uytterhoeven

Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml

Quoted 9 lines · Clément Léger

Add bindings for Renesas RZ/N1 Advanced 5 port switch. This switch is present on Renesas RZ/N1 SoC and was probably provided by MoreThanIP. This company does not exists anymore and has been bought by Synopsys. Since this IP can't be find anymore in the Synospsy portfolio, lets use Renesas as the vendor compatible for this IP.

Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml new file +0 -0

Missing "power-domains" property.

Quoted 5 lines
Quoted 2 lines

I do not use pm_runtime* in the switch driver. I should probably do that right ?

For now you don't have to. But I think it is a good idea, and it helps if the IP block is ever reused in an SoC with real power areas.

Ok, sounds good, I'll probably also set that as required to be compatible with these potentials modifications without breaking the bindings.

Thanks,

Quoted 11 lines

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

Signature · 3 lines
Clément Léger,
Embedded Linux and Kernel engineer at Bootlin
https://bootlin.com
maintainer Reviewed-by on v5 6/13
#21 lore raw

Re: [PATCH net-next v5 06/13] dt-bindings: net: dsa: add bindings for Renesas RZ/N1 Advanced 5 port switch

In reply to Clément Léger
Quoted 6 lines · Clément Léger

Add bindings for Renesas RZ/N1 Advanced 5 port switch. This switch is present on Renesas RZ/N1 SoC and was probably provided by MoreThanIP. This company does not exists anymore and has been bought by Synopsys. Since this IP can't be find anymore in the Synospsy portfolio, lets use Renesas as the vendor compatible for this IP.

Quoted 6 lines
---
 .../bindings/net/dsa/renesas,rzn1-a5psw.yaml  | 131 ++++++++++++++++++
 1 file changed, 131 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml
author maintainer
#22 lore raw

PATCH v5 7/13 net: dsa: rzn1-a5psw: add Renesas RZ/N1 advanced 5 port switch driver

In reply to Clément Léger

Add Renesas RZ/N1 advanced 5 port switch driver. This switch handles 5 ports including 1 CPU management port. A MDIO bus is also exposed by this switch and allows to communicate with PHYs connected to the ports. Each switch port (except for the CPU management ports) is connected to the MII converter.

This driver includes basic bridging support, more support will be added later (vlan, etc).

---
 drivers/net/dsa/Kconfig      |   9 +
 drivers/net/dsa/Makefile     |   1 +
 drivers/net/dsa/rzn1_a5psw.c | 720 +++++++++++++++++++++++++++++++++++
 drivers/net/dsa/rzn1_a5psw.h | 198 ++++++++++
 4 files changed, 928 insertions(+)
 create mode 100644 drivers/net/dsa/rzn1_a5psw.c
 create mode 100644 drivers/net/dsa/rzn1_a5psw.h
drivers/net/dsa/Kconfig +9 -0
index 37a3dabdce31..66ff3a4a3a98 100644
@@ -70,6 +70,15 @@ config NET_DSA_QCA8K
 
 source "drivers/net/dsa/realtek/Kconfig"
 
+config NET_DSA_RZN1_A5PSW
+	tristate "Renesas RZ/N1 A5PSW Ethernet switch support"
+	depends on ARCH_RZN1
+	select NET_DSA_TAG_RZN1_A5PSW
+	select PCS_RZN1_MIIC
+	help
+	  This driver supports the A5PSW switch, which is embedded in Renesas
+	  RZ/N1 SoC.
+
 config NET_DSA_SMSC_LAN9303
 	tristate
 	depends on VLAN_8021Q || VLAN_8021Q=n
drivers/net/dsa/Makefile +1 -0
index e73838c12256..b32907afa702 100644
@@ -9,6 +9,7 @@ obj-$(CONFIG_NET_DSA_LANTIQ_GSWIP) += lantiq_gswip.o
 obj-$(CONFIG_NET_DSA_MT7530)	+= mt7530.o
 obj-$(CONFIG_NET_DSA_MV88E6060) += mv88e6060.o
 obj-$(CONFIG_NET_DSA_QCA8K)	+= qca8k.o
+obj-$(CONFIG_NET_DSA_RZN1_A5PSW) += rzn1_a5psw.o
 obj-$(CONFIG_NET_DSA_SMSC_LAN9303) += lan9303-core.o
 obj-$(CONFIG_NET_DSA_SMSC_LAN9303_I2C) += lan9303_i2c.o
 obj-$(CONFIG_NET_DSA_SMSC_LAN9303_MDIO) += lan9303_mdio.o
drivers/net/dsa/rzn1_a5psw.c new file +720 -0

720 lines not on this page · raw message

drivers/net/dsa/rzn1_a5psw.h new file +198 -0

198 lines not on this page · raw message

Signature · 1 line
2.36.0
maintainer
#23 lore raw

Re: [PATCH net-next v5 07/13] net: dsa: rzn1-a5psw: add Renesas RZ/N1 advanced 5 port switch driver

Quoted 9 lines · Clément Léger
drivers/net/dsa/rzn1_a5psw.c new file +9 -0
+static int a5psw_pcs_get(struct a5psw *a5psw)
+{
+	struct device_node *ports, *port, *pcs_node;
+	struct phylink_pcs *pcs;
+	int ret;
+	u32 reg;
+
+	ports = of_get_child_by_name(a5psw->dev->of_node, "ethernet-ports");
+	if (!ports)

drivers/net/dsa/rzn1_a5psw.c line 561

Quoted 10 lines
drivers/net/dsa/rzn1_a5psw.c new file +10 -0
+		return -EINVAL;
+
+	for_each_available_child_of_node(ports, port) {
+		pcs_node = of_parse_phandle(port, "pcs-handle", 0);
+		if (!pcs_node)
+			continue;
+
+		if (of_property_read_u32(port, "reg", &reg)) {
+			ret = -EINVAL;
+			goto free_pcs;

I think when you exit the for_each_available_child_of_node() loop you need to manually call of_node_put(port).

Quoted 27 lines
drivers/net/dsa/rzn1_a5psw.c new file +27 -0
+		}
+
+		if (reg >= ARRAY_SIZE(a5psw->pcs)) {
+			ret = -ENODEV;
+			goto free_pcs;
+		}
+
+		pcs = miic_create(pcs_node);
+		if (IS_ERR(pcs)) {
+			dev_err(a5psw->dev, "Failed to create PCS for port %d\n",
+				reg);
+			ret = PTR_ERR(pcs);
+			goto free_pcs;
+		}
+
+		a5psw->pcs[reg] = pcs;
+	}
+	of_node_put(ports);
+
+	return 0;
+
+free_pcs:
+	a5psw_pcs_free(a5psw);
+	of_node_put(ports);
+
+	return ret;
+}
author maintainer
#24 lore raw

Re: [PATCH net-next v5 07/13] net: dsa: rzn1-a5psw: add Renesas RZ/N1 advanced 5 port switch driver

Quoted 2 lines · Vladimir Oltean

I think when you exit the for_each_available_child_of_node() loop you need to manually call of_node_put(port).

Yes you are right on that point. I'll fix that.

Signature · 3 lines
Clément Léger,
Embedded Linux and Kernel engineer at Bootlin
https://bootlin.com
author maintainer
#25 lore raw

PATCH v5 8/13 net: dsa: rzn1-a5psw: add statistics support

In reply to Clément Léger

Add statistics support to the rzn1-a5psw driver by implementing the following dsa_switch_ops callbacks:
- get_sset_count()
- get_strings()
- get_ethtool_stats()
- get_eth_mac_stats()
- get_eth_ctrl_stats()
- get_rmon_stats()

---
 drivers/net/dsa/rzn1_a5psw.c | 174 +++++++++++++++++++++++++++++++++++
 drivers/net/dsa/rzn1_a5psw.h |  46 ++++++++-
 2 files changed, 219 insertions(+), 1 deletion(-)
drivers/net/dsa/rzn1_a5psw.c +174 -0
index 3c5e905e3a0f..3ef68bf1c8c6 100644
@@ -17,6 +17,57 @@
 
 #include "rzn1_a5psw.h"
 
+struct a5psw_stats {
+	u16 offset;
+	const char name[ETH_GSTRING_LEN];
+};
+
+#define STAT_DESC(_offset) {	\
+	.offset = A5PSW_##_offset,	\
+	.name = __stringify(_offset),	\
+}
+
+static const struct a5psw_stats a5psw_stats[] = {
+	STAT_DESC(aFramesTransmittedOK),
+	STAT_DESC(aFramesReceivedOK),
+	STAT_DESC(aFrameCheckSequenceErrors),
+	STAT_DESC(aAlignmentErrors),
+	STAT_DESC(aOctetsTransmittedOK),
+	STAT_DESC(aOctetsReceivedOK),
+	STAT_DESC(aTxPAUSEMACCtrlFrames),
+	STAT_DESC(aRxPAUSEMACCtrlFrames),
+	STAT_DESC(ifInErrors),
+	STAT_DESC(ifOutErrors),
+	STAT_DESC(ifInUcastPkts),
+	STAT_DESC(ifInMulticastPkts),
+	STAT_DESC(ifInBroadcastPkts),
+	STAT_DESC(ifOutDiscards),
+	STAT_DESC(ifOutUcastPkts),
+	STAT_DESC(ifOutMulticastPkts),
+	STAT_DESC(ifOutBroadcastPkts),
+	STAT_DESC(etherStatsDropEvents),
+	STAT_DESC(etherStatsOctets),
+	STAT_DESC(etherStatsPkts),
+	STAT_DESC(etherStatsUndersizePkts),
+	STAT_DESC(etherStatsOversizePkts),
+	STAT_DESC(etherStatsPkts64Octets),
+	STAT_DESC(etherStatsPkts65to127Octets),
+	STAT_DESC(etherStatsPkts128to255Octets),
+	STAT_DESC(etherStatsPkts256to511Octets),
+	STAT_DESC(etherStatsPkts1024to1518Octets),
+	STAT_DESC(etherStatsPkts1519toXOctets),
+	STAT_DESC(etherStatsJabbers),
+	STAT_DESC(etherStatsFragments),
+	STAT_DESC(VLANReceived),
+	STAT_DESC(VLANTransmitted),
+	STAT_DESC(aDeferred),
+	STAT_DESC(aMultipleCollisions),
+	STAT_DESC(aSingleCollisions),
+	STAT_DESC(aLateCollisions),
+	STAT_DESC(aExcessiveCollisions),
+	STAT_DESC(aCarrierSenseErrors),
+};
+
 static void a5psw_reg_writel(struct a5psw *a5psw, int offset, u32 value)
 {
 	writel(value, a5psw->base + offset);
@@ -324,6 +375,123 @@ static void a5psw_port_fast_age(struct dsa_switch *ds, int port)
 	a5psw_port_fdb_flush(a5psw, port);
 }
 
+static u64 a5psw_read_stat(struct a5psw *a5psw, u32 offset, int port)
+{
+	u32 reg_lo, reg_hi;
+
+	reg_lo = a5psw_reg_readl(a5psw, offset + A5PSW_PORT_OFFSET(port));
+	/* A5PSW_STATS_HIWORD is latched on stat read */
+	reg_hi = a5psw_reg_readl(a5psw, A5PSW_STATS_HIWORD);
+
+	return ((u64)reg_hi << 32) | reg_lo;
+}
+
+static void a5psw_get_strings(struct dsa_switch *ds, int port, u32 stringset,
+			      uint8_t *data)
+{
+	unsigned int u;
+
+	if (stringset != ETH_SS_STATS)
+		return;
+
+	for (u = 0; u < ARRAY_SIZE(a5psw_stats); u++) {
+		memcpy(data + u * ETH_GSTRING_LEN, a5psw_stats[u].name,
+		       ETH_GSTRING_LEN);
+	}
+}
+
+static void a5psw_get_ethtool_stats(struct dsa_switch *ds, int port,
+				    uint64_t *data)
+{
+	struct a5psw *a5psw = ds->priv;
+	unsigned int u;
+
+	for (u = 0; u < ARRAY_SIZE(a5psw_stats); u++)
+		data[u] = a5psw_read_stat(a5psw, a5psw_stats[u].offset, port);
+}
+
+static int a5psw_get_sset_count(struct dsa_switch *ds, int port, int sset)
+{
+	if (sset != ETH_SS_STATS)
+		return 0;
+
+	return ARRAY_SIZE(a5psw_stats);
+}
+
+static void a5psw_get_eth_mac_stats(struct dsa_switch *ds, int port,
+				    struct ethtool_eth_mac_stats *mac_stats)
+{
+	struct a5psw *a5psw = ds->priv;
+
+#define RD(name) a5psw_read_stat(a5psw, A5PSW_##name, port)
+	mac_stats->FramesTransmittedOK = RD(aFramesTransmittedOK);
+	mac_stats->SingleCollisionFrames = RD(aSingleCollisions);
+	mac_stats->MultipleCollisionFrames = RD(aMultipleCollisions);
+	mac_stats->FramesReceivedOK = RD(aFramesReceivedOK);
+	mac_stats->FrameCheckSequenceErrors = RD(aFrameCheckSequenceErrors);
+	mac_stats->AlignmentErrors = RD(aAlignmentErrors);
+	mac_stats->OctetsTransmittedOK = RD(aOctetsTransmittedOK);
+	mac_stats->FramesWithDeferredXmissions = RD(aDeferred);
+	mac_stats->LateCollisions = RD(aLateCollisions);
+	mac_stats->FramesAbortedDueToXSColls = RD(aExcessiveCollisions);
+	mac_stats->FramesLostDueToIntMACXmitError = RD(ifOutErrors);
+	mac_stats->CarrierSenseErrors = RD(aCarrierSenseErrors);
+	mac_stats->OctetsReceivedOK = RD(aOctetsReceivedOK);
+	mac_stats->FramesLostDueToIntMACRcvError = RD(ifInErrors);
+	mac_stats->MulticastFramesXmittedOK = RD(ifOutMulticastPkts);
+	mac_stats->BroadcastFramesXmittedOK = RD(ifOutBroadcastPkts);
+	mac_stats->FramesWithExcessiveDeferral = RD(aDeferred);
+	mac_stats->MulticastFramesReceivedOK = RD(ifInMulticastPkts);
+	mac_stats->BroadcastFramesReceivedOK = RD(ifInBroadcastPkts);
+#undef RD
+}
+
+static const struct ethtool_rmon_hist_range a5psw_rmon_ranges[] = {
+	{ 0, 64 },
+	{ 65, 127 },
+	{ 128, 255 },
+	{ 256, 511 },
+	{ 512, 1023 },
+	{ 1024, 1518 },
+	{ 1519, A5PSW_MAX_MTU },
+	{}
+};
+
+static void a5psw_get_rmon_stats(struct dsa_switch *ds, int port,
+				 struct ethtool_rmon_stats *rmon_stats,
+				 const struct ethtool_rmon_hist_range **ranges)
+{
+	struct a5psw *a5psw = ds->priv;
+
+#define RD(name) a5psw_read_stat(a5psw, A5PSW_##name, port)
+	rmon_stats->undersize_pkts = RD(etherStatsUndersizePkts);
+	rmon_stats->oversize_pkts = RD(etherStatsOversizePkts);
+	rmon_stats->fragments = RD(etherStatsFragments);
+	rmon_stats->jabbers = RD(etherStatsJabbers);
+	rmon_stats->hist[0] = RD(etherStatsPkts64Octets);
+	rmon_stats->hist[1] = RD(etherStatsPkts65to127Octets);
+	rmon_stats->hist[2] = RD(etherStatsPkts128to255Octets);
+	rmon_stats->hist[3] = RD(etherStatsPkts256to511Octets);
+	rmon_stats->hist[4] = RD(etherStatsPkts512to1023Octets);
+	rmon_stats->hist[5] = RD(etherStatsPkts1024to1518Octets);
+	rmon_stats->hist[6] = RD(etherStatsPkts1519toXOctets);
+#undef RD
+
+	*ranges = a5psw_rmon_ranges;
+}
+
+static void a5psw_get_eth_ctrl_stats(struct dsa_switch *ds, int port,
+				     struct ethtool_eth_ctrl_stats *ctrl_stats)
+{
+	struct a5psw *a5psw = ds->priv;
+	u64 stat;
+
+	stat = a5psw_read_stat(a5psw, A5PSW_aTxPAUSEMACCtrlFrames, port);
+	ctrl_stats->MACControlFramesTransmitted = stat;
+	stat = a5psw_read_stat(a5psw, A5PSW_aRxPAUSEMACCtrlFrames, port);
+	ctrl_stats->MACControlFramesReceived = stat;
+}
+
 static int a5psw_setup(struct dsa_switch *ds)
 {
 	struct a5psw *a5psw = ds->priv;
@@ -412,6 +580,12 @@ static const struct dsa_switch_ops a5psw_switch_ops = {
 	.phylink_mac_link_up = a5psw_phylink_mac_link_up,
 	.port_change_mtu = a5psw_port_change_mtu,
 	.port_max_mtu = a5psw_port_max_mtu,
+	.get_sset_count = a5psw_get_sset_count,
+	.get_strings = a5psw_get_strings,
+	.get_ethtool_stats = a5psw_get_ethtool_stats,
+	.get_eth_mac_stats = a5psw_get_eth_mac_stats,
+	.get_eth_ctrl_stats = a5psw_get_eth_ctrl_stats,
+	.get_rmon_stats = a5psw_get_rmon_stats,
 	.set_ageing_time = a5psw_set_ageing_time,
 	.port_bridge_join = a5psw_port_bridge_join,
 	.port_bridge_leave = a5psw_port_bridge_leave,
drivers/net/dsa/rzn1_a5psw.h +45 -1

58 lines not on this page · raw message

Signature · 1 line
2.36.0
maintainer
#26 lore raw

Re: [PATCH net-next v5 08/13] net: dsa: rzn1-a5psw: add statistics support

Quoted 9 lines · Clément Léger
drivers/net/dsa/rzn1_a5psw.h +0 -1
index 306cab55db24..649165d37fde 100644
@@ -146,7 +146,50 @@
 
 #define A5PSW_STATS_HIWORD		0x900
 
-#define A5PSW_DUMMY_WORKAROUND		0x5000

drivers/net/dsa/rzn1_a5psw.h line 158

Quoted 10 lines
drivers/net/dsa/rzn1_a5psw.h +10 -0
+/* Stats */
+#define A5PSW_aFramesTransmittedOK		0x868
+#define A5PSW_aFramesReceivedOK			0x86C
+#define A5PSW_aFrameCheckSequenceErrors		0x870
+#define A5PSW_aAlignmentErrors			0x874
+#define A5PSW_aOctetsTransmittedOK		0x878
+#define A5PSW_aOctetsReceivedOK			0x87C
+#define A5PSW_aTxPAUSEMACCtrlFrames		0x880
+#define A5PSW_aRxPAUSEMACCtrlFrames		0x884
+	/* If */

It would probably be good to keep the same alignment for all comments here.

Quoted 34 lines
drivers/net/dsa/rzn1_a5psw.h +34 -0
+#define A5PSW_ifInErrors			0x888
+#define A5PSW_ifOutErrors			0x88C
+#define A5PSW_ifInUcastPkts			0x890
+#define A5PSW_ifInMulticastPkts			0x894
+#define A5PSW_ifInBroadcastPkts			0x898
+#define A5PSW_ifOutDiscards			0x89C
+#define A5PSW_ifOutUcastPkts			0x8A0
+#define A5PSW_ifOutMulticastPkts		0x8A4
+#define A5PSW_ifOutBroadcastPkts		0x8A8
+	/* Ether */
+#define A5PSW_etherStatsDropEvents		0x8AC
+#define A5PSW_etherStatsOctets			0x8B0
+#define A5PSW_etherStatsPkts			0x8B4
+#define A5PSW_etherStatsUndersizePkts		0x8B8
+#define A5PSW_etherStatsOversizePkts		0x8BC
+#define A5PSW_etherStatsPkts64Octets		0x8C0
+#define A5PSW_etherStatsPkts65to127Octets	0x8C4
+#define A5PSW_etherStatsPkts128to255Octets	0x8C8
+#define A5PSW_etherStatsPkts256to511Octets	0x8CC
+#define A5PSW_etherStatsPkts512to1023Octets	0x8D0
+#define A5PSW_etherStatsPkts1024to1518Octets	0x8D4
+#define A5PSW_etherStatsPkts1519toXOctets	0x8D8
+#define A5PSW_etherStatsJabbers			0x8DC
+#define A5PSW_etherStatsFragments		0x8E0
+
+#define A5PSW_VLANReceived			0x8E8
+#define A5PSW_VLANTransmitted			0x8EC
+
+#define A5PSW_aDeferred				0x910
+#define A5PSW_aMultipleCollisions		0x914
+#define A5PSW_aSingleCollisions			0x918
+#define A5PSW_aLateCollisions			0x91C
+#define A5PSW_aExcessiveCollisions		0x920
+#define A5PSW_aCarrierSenseErrors		0x924
author maintainer
#27 lore raw

PATCH v5 9/13 net: dsa: rzn1-a5psw: add FDB support

In reply to Clément Léger

This commits add forwarding database support to the driver. It implements fdb_add(), fdb_del() and fdb_dump().

---
 drivers/net/dsa/rzn1_a5psw.c | 175 ++++++++++++++++++++++++++++++++++-
 drivers/net/dsa/rzn1_a5psw.h |  19 +++-
 2 files changed, 190 insertions(+), 4 deletions(-)
drivers/net/dsa/rzn1_a5psw.c +172 -3
index 3ef68bf1c8c6..125491d24c68 100644
@@ -150,9 +150,9 @@ static void a5psw_port_fdb_flush(struct a5psw *a5psw, int port)
 {
 	u32 ctrl = A5PSW_LK_ADDR_CTRL_DELETE_PORT | BIT(port);
 
-	spin_lock(&a5psw->lk_lock);
+	mutex_lock(&a5psw->lk_lock);
 	a5psw_lk_execute_ctrl(a5psw, &ctrl);
-	spin_unlock(&a5psw->lk_lock);
+	mutex_unlock(&a5psw->lk_lock);
 }
 
 static void a5psw_port_authorize_set(struct a5psw *a5psw, int port,
@@ -375,6 +375,172 @@ static void a5psw_port_fast_age(struct dsa_switch *ds, int port)
 	a5psw_port_fdb_flush(a5psw, port);
 }
 
+static int a5psw_lk_execute_lookup(struct a5psw *a5psw, union lk_data *lk_data,
+				   u16 *entry)
+{
+	u32 ctrl;
+	int ret;
+
+	a5psw_reg_writel(a5psw, A5PSW_LK_DATA_LO, lk_data->lo);
+	a5psw_reg_writel(a5psw, A5PSW_LK_DATA_HI, lk_data->hi);
+
+	ctrl = A5PSW_LK_ADDR_CTRL_LOOKUP;
+	ret = a5psw_lk_execute_ctrl(a5psw, &ctrl);
+	if (ret)
+		return ret;
+
+	*entry = ctrl & A5PSW_LK_ADDR_CTRL_ADDRESS;
+
+	return 0;
+}
+
+static int a5psw_port_fdb_add(struct dsa_switch *ds, int port,
+			      const unsigned char *addr, u16 vid,
+			      struct dsa_db db)
+{
+	struct a5psw *a5psw = ds->priv;
+	union lk_data lk_data = {0};
+	bool inc_learncount = false;
+	int ret = 0;
+	u16 entry;
+	u32 reg;
+
+	ether_addr_copy(lk_data.entry.mac, addr);
+	lk_data.entry.port_mask = BIT(port);
+
+	mutex_lock(&a5psw->lk_lock);
+
+	/* Set the value to be written in the lookup table */
+	ret = a5psw_lk_execute_lookup(a5psw, &lk_data, &entry);
+	if (ret)
+		goto lk_unlock;
+
+	lk_data.hi = a5psw_reg_readl(a5psw, A5PSW_LK_DATA_HI);
+	if (!lk_data.entry.valid) {
+		inc_learncount = true;
+		/* port_mask set to 0x1f when entry is not valid, clear it */
+		lk_data.entry.port_mask = 0;
+		lk_data.entry.prio = 0;
+	}
+
+	lk_data.entry.port_mask |= BIT(port);
+	lk_data.entry.is_static = 1;
+	lk_data.entry.valid = 1;
+
+	a5psw_reg_writel(a5psw, A5PSW_LK_DATA_HI, lk_data.hi);
+
+	reg = A5PSW_LK_ADDR_CTRL_WRITE | entry;
+	ret = a5psw_lk_execute_ctrl(a5psw, &reg);
+	if (ret)
+		goto lk_unlock;
+
+	if (inc_learncount) {
+		reg = A5PSW_LK_LEARNCOUNT_MODE_INC;
+		a5psw_reg_writel(a5psw, A5PSW_LK_LEARNCOUNT, reg);
+	}
+
+lk_unlock:
+	mutex_unlock(&a5psw->lk_lock);
+
+	return ret;
+}
+
+static int a5psw_port_fdb_del(struct dsa_switch *ds, int port,
+			      const unsigned char *addr, u16 vid,
+			      struct dsa_db db)
+{
+	struct a5psw *a5psw = ds->priv;
+	union lk_data lk_data = {0};
+	bool clear = false;
+	u16 entry;
+	u32 reg;
+	int ret;
+
+	ether_addr_copy(lk_data.entry.mac, addr);
+
+	mutex_lock(&a5psw->lk_lock);
+
+	ret = a5psw_lk_execute_lookup(a5psw, &lk_data, &entry);
+	if (ret)
+		goto lk_unlock;
+
+	lk_data.hi = a5psw_reg_readl(a5psw, A5PSW_LK_DATA_HI);
+
+	/* Our hardware does not associate any VID to the FDB entries so this
+	 * means that if two entries were added for the same mac but for
+	 * different VID, then, on the deletion of the first one, we would also
+	 * delete the second one. Since there is unfortunately nothing we can do
+	 * about that, do not return an error...
+	 */
+	if (!lk_data.entry.valid)
+		goto lk_unlock;
+
+	lk_data.entry.port_mask &= ~BIT(port);
+	/* If there is no more port in the mask, clear the entry */
+	if (lk_data.entry.port_mask == 0)
+		clear = true;
+
+	a5psw_reg_writel(a5psw, A5PSW_LK_DATA_HI, lk_data.hi);
+
+	reg = entry;
+	if (clear)
+		reg |= A5PSW_LK_ADDR_CTRL_CLEAR;
+	else
+		reg |= A5PSW_LK_ADDR_CTRL_WRITE;
+
+	ret = a5psw_lk_execute_ctrl(a5psw, &reg);
+	if (ret)
+		goto lk_unlock;
+
+	/* Decrement LEARNCOUNT */
+	if (clear) {
+		reg = A5PSW_LK_LEARNCOUNT_MODE_DEC;
+		a5psw_reg_writel(a5psw, A5PSW_LK_LEARNCOUNT, reg);
+	}
+
+lk_unlock:
+	mutex_unlock(&a5psw->lk_lock);
+
+	return ret;
+}
+
+static int a5psw_port_fdb_dump(struct dsa_switch *ds, int port,
+			       dsa_fdb_dump_cb_t *cb, void *data)
+{
+	struct a5psw *a5psw = ds->priv;
+	union lk_data lk_data;
+	int i = 0, ret;
+	u32 reg;
+
+	for (i = 0; i < A5PSW_TABLE_ENTRIES; i++) {
+		reg = A5PSW_LK_ADDR_CTRL_READ | A5PSW_LK_ADDR_CTRL_WAIT | i;
+		mutex_lock(&a5psw->lk_lock);
+
+		ret = a5psw_lk_execute_ctrl(a5psw, &reg);
+		if (ret) {
+			mutex_unlock(&a5psw->lk_lock);
+			return ret;
+		}
+
+		lk_data.hi = a5psw_reg_readl(a5psw, A5PSW_LK_DATA_HI);
+		/* If entry is not valid or does not contain the port, skip */
+		if (!lk_data.entry.valid ||
+		    !(lk_data.entry.port_mask & BIT(port))) {
+			mutex_unlock(&a5psw->lk_lock);
+			continue;
+		}
+
+		lk_data.lo = a5psw_reg_readl(a5psw, A5PSW_LK_DATA_LO);
+		mutex_unlock(&a5psw->lk_lock);
+
+		ret = cb(lk_data.entry.mac, 0, lk_data.entry.is_static, data);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
 static u64 a5psw_read_stat(struct a5psw *a5psw, u32 offset, int port)
 {
 	u32 reg_lo, reg_hi;
@@ -591,6 +757,9 @@ static const struct dsa_switch_ops a5psw_switch_ops = {
 	.port_bridge_leave = a5psw_port_bridge_leave,
 	.port_stp_state_set = a5psw_port_stp_state_set,
 	.port_fast_age = a5psw_port_fast_age,
+	.port_fdb_add = a5psw_port_fdb_add,
+	.port_fdb_del = a5psw_port_fdb_del,
+	.port_fdb_dump = a5psw_port_fdb_dump,
 };
 
 static int a5psw_mdio_wait_busy(struct a5psw *a5psw)
@@ -774,7 +943,7 @@ static int a5psw_probe(struct platform_device *pdev)
 		return -ENOMEM;
 
 	a5psw->dev = dev;
-	spin_lock_init(&a5psw->lk_lock);
+	mutex_init(&a5psw->lk_lock);
 	spin_lock_init(&a5psw->reg_lock);
 	a5psw->base = devm_platform_ioremap_resource(pdev, 0);
 	if (!a5psw->base)
drivers/net/dsa/rzn1_a5psw.h +18 -1
index 649165d37fde..11fcea45b255 100644
@@ -211,6 +211,23 @@
 #define A5PSW_CTRL_TIMEOUT		1000
 #define A5PSW_TABLE_ENTRIES		8192
 
+struct fdb_entry {
+	u8 mac[ETH_ALEN];
+	u16 valid:1;
+	u16 is_static:1;
+	u16 prio:3;
+	u16 port_mask:5;
+	u16 reserved:6;
+} __packed;
+
+union lk_data {
+	struct {
+		u32 lo;
+		u32 hi;
+	};
+	struct fdb_entry entry;
+};
+
 /**
  * struct a5psw - switch struct
  * @base: Base address of the switch
@@ -235,7 +252,7 @@ struct a5psw {
 	struct mii_bus	*mii_bus;
 	struct phylink_pcs *pcs[A5PSW_PORTS_NUM - 1];
 	struct dsa_switch ds;
-	spinlock_t lk_lock;
+	struct mutex lk_lock;
 	spinlock_t reg_lock;
 	u32 bridged_ports;
 	struct net_device *br_dev;
Signature · 1 line
2.36.0
maintainer
#28 lore raw

Re: [PATCH net-next v5 09/13] net: dsa: rzn1-a5psw: add FDB support

drivers/net/dsa/rzn1_a5psw.c static void a5psw_port_fast_age(struct dsa_switch *ds, int port) · line 517

Quoted 11 lines · Clément Léger
drivers/net/dsa/rzn1_a5psw.c +11 -0
+static int a5psw_port_fdb_dump(struct dsa_switch *ds, int port,
+			       dsa_fdb_dump_cb_t *cb, void *data)
+{
+	struct a5psw *a5psw = ds->priv;
+	union lk_data lk_data;
+	int i = 0, ret;
+	u32 reg;
+
+	for (i = 0; i < A5PSW_TABLE_ENTRIES; i++) {
+		reg = A5PSW_LK_ADDR_CTRL_READ | A5PSW_LK_ADDR_CTRL_WAIT | i;
+		mutex_lock(&a5psw->lk_lock);

It might be more efficient to lock the lookup table only once, outside the for loop, rather than 8192 times (plus the fact that when you run plain "bridge fdb show", this gets repeated for each switch user port, which is a nuisance of its own).

Quoted 25 lines
drivers/net/dsa/rzn1_a5psw.c +25 -0
+
+		ret = a5psw_lk_execute_ctrl(a5psw, &reg);
+		if (ret) {
+			mutex_unlock(&a5psw->lk_lock);
+			return ret;
+		}
+
+		lk_data.hi = a5psw_reg_readl(a5psw, A5PSW_LK_DATA_HI);
+		/* If entry is not valid or does not contain the port, skip */
+		if (!lk_data.entry.valid ||
+		    !(lk_data.entry.port_mask & BIT(port))) {
+			mutex_unlock(&a5psw->lk_lock);
+			continue;
+		}
+
+		lk_data.lo = a5psw_reg_readl(a5psw, A5PSW_LK_DATA_LO);
+		mutex_unlock(&a5psw->lk_lock);
+
+		ret = cb(lk_data.entry.mac, 0, lk_data.entry.is_static, data);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
author maintainer
#29 lore raw

PATCH v5 10/13 ARM: dts: r9a06g032: describe MII converter

In reply to Clément Léger

Add the MII converter node which describes the MII converter that is present on the RZ/N1 SoC.

---
 arch/arm/boot/dts/r9a06g032.dtsi | 38 ++++++++++++++++++++++++++++++++
 1 file changed, 38 insertions(+)
arch/arm/boot/dts/r9a06g032.dtsi +38 -0
index 636a6ab31c58..0051fe9f44fd 100644
@@ -200,6 +200,44 @@ nand_controller: nand-controller@40102000 {
 			status = "disabled";
 		};
 
+		eth_miic: eth-miic@44030000 {
+			compatible = "renesas,r9a06g032-miic", "renesas,rzn1-miic";
+			#address-cells = <1>;
+			#size-cells = <0>;
+			reg = <0x44030000 0x10000>;
+			clocks = <&sysctrl R9A06G032_CLK_MII_REF>,
+				 <&sysctrl R9A06G032_CLK_RGMII_REF>,
+				 <&sysctrl R9A06G032_CLK_RMII_REF>,
+				 <&sysctrl R9A06G032_HCLK_SWITCH_RG>;
+			power-domains = <&sysctrl>;
+			status = "disabled";
+
+			mii_conv1: mii-conv@1 {
+				reg = <1>;
+				status = "disabled";
+			};
+
+			mii_conv2: mii-conv@2 {
+				reg = <2>;
+				status = "disabled";
+			};
+
+			mii_conv3: mii-conv@3 {
+				reg = <3>;
+				status = "disabled";
+			};
+
+			mii_conv4: mii-conv@4 {
+				reg = <4>;
+				status = "disabled";
+			};
+
+			mii_conv5: mii-conv@5 {
+				reg = <5>;
+				status = "disabled";
+			};
+		};
+
 		gic: interrupt-controller@44101000 {
 			compatible = "arm,gic-400", "arm,cortex-a7-gic";
 			interrupt-controller;
Signature · 1 line
2.36.0
author maintainer
#30 lore raw

PATCH v5 11/13 ARM: dts: r9a06g032: describe GMAC2

In reply to Clément Léger

RZ/N1 SoC includes two MAC named GMACx that are compatible with the "snps,dwmac" driver. GMAC1 is connected directly to the MII converter port 1. GMAC2 however can be used as the MAC for the switch CPU management port or can be muxed to be connected directly to the MII converter port 2. This commit add description for the GMAC2 which will be used by the switch description.

---
 arch/arm/boot/dts/r9a06g032.dtsi | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)
arch/arm/boot/dts/r9a06g032.dtsi +17 -0
index 0051fe9f44fd..31c4b2e2950a 100644
@@ -200,6 +200,23 @@ nand_controller: nand-controller@40102000 {
 			status = "disabled";
 		};
 
+		gmac2: ethernet@44002000 {
+			compatible = "snps,dwmac";
+			reg = <0x44002000 0x2000>;
+			interrupt-parent = <&gic>;
+			interrupts = <GIC_SPI 37 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 39 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 38 IRQ_TYPE_LEVEL_HIGH>;
+			interrupt-names = "macirq", "eth_wake_irq", "eth_lpi";
+			clock-names = "stmmaceth";
+			clocks = <&sysctrl R9A06G032_HCLK_GMAC1>;
+			snps,multicast-filter-bins = <256>;
+			snps,perfect-filter-entries = <128>;
+			tx-fifo-depth = <2048>;
+			rx-fifo-depth = <4096>;
+			status = "disabled";
+		};
+
 		eth_miic: eth-miic@44030000 {
 			compatible = "renesas,r9a06g032-miic", "renesas,rzn1-miic";
 			#address-cells = <1>;
Signature · 1 line
2.36.0
#31 lore raw

Re: [PATCH net-next v5 11/13] ARM: dts: r9a06g032: describe GMAC2

Hi Clément

Quoted 8 lines · Clément Léger

RZ/N1 SoC includes two MAC named GMACx that are compatible with the "snps,dwmac" driver. GMAC1 is connected directly to the MII converter port 1. GMAC2 however can be used as the MAC for the switch CPU management port or can be muxed to be connected directly to the MII converter port 2. This commit add description for the GMAC2 which will be used by the switch description.

Thanks for your patch!

arch/arm/boot/dts/r9a06g032.dtsi nand_controller: nand-controller@40102000 { · line 204

Quoted 8 lines
arch/arm/boot/dts/r9a06g032.dtsi +2 -0
@@ -200,6 +200,23 @@ nand_controller: nand-controller@40102000 {
                        status = "disabled";
                };
 
+               gmac2: ethernet@44002000 {
+                       compatible = "snps,dwmac";

Does this need an SoC-specific compatible value?

arch/arm/boot/dts/r9a06g032.dtsi nand_controller: nand-controller@40102000 { · line 212

Quoted 8 lines
arch/arm/boot/dts/r9a06g032.dtsi +8 -0
+                       reg = <0x44002000 0x2000>;
+                       interrupt-parent = <&gic>;
+                       interrupts = <GIC_SPI 37 IRQ_TYPE_LEVEL_HIGH>,
+                                    <GIC_SPI 39 IRQ_TYPE_LEVEL_HIGH>,
+                                    <GIC_SPI 38 IRQ_TYPE_LEVEL_HIGH>;
+                       interrupt-names = "macirq", "eth_wake_irq", "eth_lpi";
+                       clock-names = "stmmaceth";
+                       clocks = <&sysctrl R9A06G032_HCLK_GMAC1>;

Missing "power-domains", also in the DT bindings.
The driver already uses Runtime PM.

arch/arm/boot/dts/r9a06g032.dtsi nand_controller: nand-controller@40102000 { · line 222

Quoted 10 lines
arch/arm/boot/dts/r9a06g032.dtsi +7 -0
+                       snps,multicast-filter-bins = <256>;
+                       snps,perfect-filter-entries = <128>;
+                       tx-fifo-depth = <2048>;
+                       rx-fifo-depth = <4096>;
+                       status = "disabled";
+               };
+
                eth_miic: eth-miic@44030000 {
                        compatible = "renesas,r9a06g032-miic", "renesas,rzn1-miic";
                        #address-cells = <1>;

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

author maintainer
#32 lore raw

Re: [PATCH net-next v5 11/13] ARM: dts: r9a06g032: describe GMAC2

Quoted 20 lines · Geert Uytterhoeven

Hi Clément

Quoted 8 lines · Clément Léger

RZ/N1 SoC includes two MAC named GMACx that are compatible with the "snps,dwmac" driver. GMAC1 is connected directly to the MII converter port 1. GMAC2 however can be used as the MAC for the switch CPU management port or can be muxed to be connected directly to the MII converter port 2. This commit add description for the GMAC2 which will be used by the switch description.

Thanks for your patch!

Quoted 6 lines
arch/arm/boot/dts/r9a06g032.dtsi +0 -0
@@ -200,6 +200,23 @@ nand_controller: nand-controller@40102000 {
                        status = "disabled";
                };
 
Quoted 4 lines

arch/arm/boot/dts/r9a06g032.dtsi nand_controller: nand-controller@40102000 { · line 204

Quoted 2 lines
arch/arm/boot/dts/r9a06g032.dtsi +2 -0
+               gmac2: ethernet@44002000 {
+                       compatible = "snps,dwmac";  

Does this need an SoC-specific compatible value?

Indeed, it might be useful to introduce a specific SoC compatible since in a near future, there might be some specific support for that gmac. Here is an overview of the gmac connection on the SoC:

                                          ┌─────────┐   ┌──────────┐
                                          │         │   │          │
                                          │  GMAC2  │   │  GMAC1   │
                                          │         │   │          │
                                          └───┬─────┘   └─────┬────┘
                                              │               │
                                              │               │
                                              │               │
                                         ┌────▼──────┐        │
                                         │           │        │
            ┌────────────────────────────┤  SWITCH   │        │
            │                            │           │        │
            │          ┌─────────────────┴─┬────┬────┘        │
            │          │            ┌──────┘    │             │
            │          │            │           │             │
       ┌────▼──────────▼────────────▼───────────▼─────────────▼────┐
       │                      MII Converter                        │
       │                                                           │
       │                                                           │
       │ port 1      port 2       port 3      port 4       port 5  │
       └───────────────────────────────────────────────────────────┘

As you can see, the GMAC1 is directly connected to MIIC converter and thus will need a "pcs-handle" property to point on the MII converter port whereas the GMAC2 is directly connected to the switch in GMII.

Is "renesas,r9a06g032-gmac2", "renesas,rzn1-switch-gmac2" looks ok for you for this one ?

Thanks,

Clément

Quoted 14 lines

arch/arm/boot/dts/r9a06g032.dtsi nand_controller: nand-controller@40102000 { · line 212

Quoted 13 lines
arch/arm/boot/dts/r9a06g032.dtsi +6 -0
+                       reg = <0x44002000 0x2000>;
+                       interrupt-parent = <&gic>;
+                       interrupts = <GIC_SPI 37
 IRQ_TYPE_LEVEL_HIGH>,
+                                    <GIC_SPI 39
 IRQ_TYPE_LEVEL_HIGH>,
+                                    <GIC_SPI 38
 IRQ_TYPE_LEVEL_HIGH>;
+                       interrupt-names = "macirq", "eth_wake_irq",

"eth_lpi";

arch/arm/boot/dts/r9a06g032.dtsi +2 -0
+                       clock-names = "stmmaceth";
+                       clocks = <&sysctrl R9A06G032_HCLK_GMAC1>;  
Quoted 2 lines

Missing "power-domains", also in the DT bindings.
The driver already uses Runtime PM.

Ok,I'll add that to the DT and modify the DWMAC bindings.

Quoted 23 lines
Quoted 11 lines
arch/arm/boot/dts/r9a06g032.dtsi +7 -0
+                       snps,multicast-filter-bins = <256>;
+                       snps,perfect-filter-entries = <128>;
+                       tx-fifo-depth = <2048>;
+                       rx-fifo-depth = <4096>;
+                       status = "disabled";
+               };
+
                eth_miic: eth-miic@44030000 {
                        compatible = "renesas,r9a06g032-miic",

"renesas,rzn1-miic"; #address-cells = <1>;

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds

Signature · 3 lines
Clément Léger,
Embedded Linux and Kernel engineer at Bootlin
https://bootlin.com
#33 lore raw

Re: [PATCH net-next v5 11/13] ARM: dts: r9a06g032: describe GMAC2

Hi Clément,

Quoted 49 lines · Clément Léger
Quoted 16 lines · Geert Uytterhoeven
Quoted 15 lines · Clément Léger

RZ/N1 SoC includes two MAC named GMACx that are compatible with the "snps,dwmac" driver. GMAC1 is connected directly to the MII converter port 1. GMAC2 however can be used as the MAC for the switch CPU management port or can be muxed to be connected directly to the MII converter port 2. This commit add description for the GMAC2 which will be used by the switch description.

arch/arm/boot/dts/r9a06g032.dtsi +0 -0
@@ -200,6 +200,23 @@ nand_controller: nand-controller@40102000 {
                        status = "disabled";
                };
 
Quoted 4 lines

arch/arm/boot/dts/r9a06g032.dtsi nand_controller: nand-controller@40102000 { · line 204

Quoted 2 lines
arch/arm/boot/dts/r9a06g032.dtsi +2 -0
+               gmac2: ethernet@44002000 {
+                       compatible = "snps,dwmac";

Does this need an SoC-specific compatible value?

Indeed, it might be useful to introduce a specific SoC compatible since in a near future, there might be some specific support for that gmac. Here is an overview of the gmac connection on the SoC:

                                          ┌─────────┐   ┌──────────┐
                                          │         │   │          │
                                          │  GMAC2  │   │  GMAC1   │
                                          │         │   │          │
                                          └───┬─────┘   └─────┬────┘
                                              │               │
                                              │               │
                                              │               │
                                         ┌────▼──────┐        │
                                         │           │        │
            ┌────────────────────────────┤  SWITCH   │        │
            │                            │           │        │
            │          ┌─────────────────┴─┬────┬────┘        │
            │          │            ┌──────┘    │             │
            │          │            │           │             │
       ┌────▼──────────▼────────────▼───────────▼─────────────▼────┐
       │                      MII Converter                        │
       │                                                           │
       │                                                           │
       │ port 1      port 2       port 3      port 4       port 5  │
       └───────────────────────────────────────────────────────────┘
Quoted 6 lines

As you can see, the GMAC1 is directly connected to MIIC converter and thus will need a "pcs-handle" property to point on the MII converter port whereas the GMAC2 is directly connected to the switch in GMII.

Is "renesas,r9a06g032-gmac2", "renesas,rzn1-switch-gmac2" looks ok for you for this one ?

Why "switch" in the family-specific value, but not in the SoC-specific value?

Are GMAC1 and GMAC2 really different, or are they identical, and is the only difference in the wiring, which can be detected at run-time using this "pcs-handle" property? If they're identical, they should use the same compatible value.

Thanks!

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

author maintainer
#34 lore raw

Re: [PATCH net-next v5 11/13] ARM: dts: r9a06g032: describe GMAC2

Quoted 57 lines · Geert Uytterhoeven

Hi Clément,

Quoted 53 lines · Clément Léger
Quoted 16 lines · Geert Uytterhoeven
Quoted 15 lines · Clément Léger

RZ/N1 SoC includes two MAC named GMACx that are compatible with the "snps,dwmac" driver. GMAC1 is connected directly to the MII converter port 1. GMAC2 however can be used as the MAC for the switch CPU management port or can be muxed to be connected directly to the MII converter port 2. This commit add description for the GMAC2 which will be used by the switch description.

arch/arm/boot/dts/r9a06g032.dtsi +0 -0
@@ -200,6 +200,23 @@ nand_controller: nand-controller@40102000 {
                        status = "disabled";
                };
 
Quoted 4 lines

arch/arm/boot/dts/r9a06g032.dtsi nand_controller: nand-controller@40102000 { · line 204

Quoted 2 lines
arch/arm/boot/dts/r9a06g032.dtsi +2 -0
+               gmac2: ethernet@44002000 {
+                       compatible = "snps,dwmac";  

Does this need an SoC-specific compatible value?

Indeed, it might be useful to introduce a specific SoC compatible since in a near future, there might be some specific support for that gmac. Here is an overview of the gmac connection on the SoC:

                                          ┌─────────┐   ┌──────────┐
                                          │         │   │          │
                                          │  GMAC2  │   │  GMAC1   │
                                          │         │   │          │
                                          └───┬─────┘   └─────┬────┘
                                              │               │
                                              │               │
                                              │               │
                                         ┌────▼──────┐        │
                                         │           │        │
            ┌────────────────────────────┤  SWITCH   │        │
            │                            │           │        │
            │          ┌─────────────────┴─┬────┬────┘        │
            │          │            ┌──────┘    │             │
            │          │            │           │             │
       ┌────▼──────────▼────────────▼───────────▼─────────────▼────┐
       │                      MII Converter                        │
       │                                                           │
       │                                                           │
       │ port 1      port 2       port 3      port 4       port 5  │
       └───────────────────────────────────────────────────────────┘

As you can see, the GMAC1 is directly connected to MIIC converter and thus will need a "pcs-handle" property to point on the MII converter port whereas the GMAC2 is directly connected to the switch in GMII.

Quoted 5 lines
Quoted 2 lines

Is "renesas,r9a06g032-gmac2", "renesas,rzn1-switch-gmac2" looks ok for you for this one ?

Why "switch" in the family-specific value, but not in the SoC-specific value?

That's a typo, switch should be removed.

Quoted 5 lines

Are GMAC1 and GMAC2 really different, or are they identical, and is the only difference in the wiring, which can be detected at run-time using this "pcs-handle" property? If they're identical, they should use the same compatible value.

They are actually identical except the requirement for a "pcs-handle" for gmac1. I thought about using different compatible to enforce this by making it "required" with the "renesas,r9a06g032-gmac1" compatible but not the "renesas,r9a06g032-gmac2" one. If it's ok for you to let it optional and use a single compatible, I'm ok with that !

Thanks !

Quoted 13 lines

Thanks!

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

Signature · 3 lines
Clément Léger,
Embedded Linux and Kernel engineer at Bootlin
https://bootlin.com
#35 lore raw

Re: [PATCH net-next v5 11/13] ARM: dts: r9a06g032: describe GMAC2

Hi Clément,

Quoted 57 lines · Clément Léger
Quoted 55 lines · Geert Uytterhoeven
Quoted 53 lines · Clément Léger
Quoted 16 lines · Geert Uytterhoeven
Quoted 15 lines · Clément Léger

RZ/N1 SoC includes two MAC named GMACx that are compatible with the "snps,dwmac" driver. GMAC1 is connected directly to the MII converter port 1. GMAC2 however can be used as the MAC for the switch CPU management port or can be muxed to be connected directly to the MII converter port 2. This commit add description for the GMAC2 which will be used by the switch description.

arch/arm/boot/dts/r9a06g032.dtsi +0 -0
@@ -200,6 +200,23 @@ nand_controller: nand-controller@40102000 {
                        status = "disabled";
                };
 
Quoted 4 lines

arch/arm/boot/dts/r9a06g032.dtsi nand_controller: nand-controller@40102000 { · line 204

Quoted 2 lines
arch/arm/boot/dts/r9a06g032.dtsi +2 -0
+               gmac2: ethernet@44002000 {
+                       compatible = "snps,dwmac";

Does this need an SoC-specific compatible value?

Indeed, it might be useful to introduce a specific SoC compatible since in a near future, there might be some specific support for that gmac. Here is an overview of the gmac connection on the SoC:

                                          ┌─────────┐   ┌──────────┐
                                          │         │   │          │
                                          │  GMAC2  │   │  GMAC1   │
                                          │         │   │          │
                                          └───┬─────┘   └─────┬────┘
                                              │               │
                                              │               │
                                              │               │
                                         ┌────▼──────┐        │
                                         │           │        │
            ┌────────────────────────────┤  SWITCH   │        │
            │                            │           │        │
            │          ┌─────────────────┴─┬────┬────┘        │
            │          │            ┌──────┘    │             │
            │          │            │           │             │
       ┌────▼──────────▼────────────▼───────────▼─────────────▼────┐
       │                      MII Converter                        │
       │                                                           │
       │                                                           │
       │ port 1      port 2       port 3      port 4       port 5  │
       └───────────────────────────────────────────────────────────┘

As you can see, the GMAC1 is directly connected to MIIC converter and thus will need a "pcs-handle" property to point on the MII converter port whereas the GMAC2 is directly connected to the switch in GMII.

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

Is "renesas,r9a06g032-gmac2", "renesas,rzn1-switch-gmac2" looks ok for you for this one ?

Why "switch" in the family-specific value, but not in the SoC-specific value?

That's a typo, switch should be removed.

OK.

Quoted 10 lines
Quoted 4 lines

Are GMAC1 and GMAC2 really different, or are they identical, and is the only difference in the wiring, which can be detected at run-time using this "pcs-handle" property? If they're identical, they should use the same compatible value.

They are actually identical except the requirement for a "pcs-handle" for gmac1. I thought about using different compatible to enforce this by making it "required" with the "renesas,r9a06g032-gmac1" compatible but not the "renesas,r9a06g032-gmac2" one. If it's ok for you to let it optional and use a single compatible, I'm ok with that !

OK to make it optional. Thanks!

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

author maintainer
#36 lore raw

PATCH v5 12/13 ARM: dts: r9a06g032: describe switch

In reply to Clément Léger

Add description of the switch that is present on the RZ/N1 SoC.

---
 arch/arm/boot/dts/r9a06g032.dtsi | 9 +++++++++
 1 file changed, 9 insertions(+)
arch/arm/boot/dts/r9a06g032.dtsi +9 -0
index 31c4b2e2950a..20d3dce632ce 100644
@@ -255,6 +255,15 @@ mii_conv5: mii-conv@5 {
 			};
 		};
 
+		switch: switch@44050000 {
+			compatible = "renesas,r9a06g032-a5psw", "renesas,rzn1-a5psw";
+			reg = <0x44050000 0x10000>;
+			clocks = <&sysctrl R9A06G032_HCLK_SWITCH>,
+				 <&sysctrl R9A06G032_CLK_SWITCH>;
+			clock-names = "hclk", "clk";
+			status = "disabled";
+		};
+
 		gic: interrupt-controller@44101000 {
 			compatible = "arm,gic-400", "arm,cortex-a7-gic";
 			interrupt-controller;
Signature · 1 line
2.36.0
maintainer
#37 lore raw

Re: [PATCH net-next v5 12/13] ARM: dts: r9a06g032: describe switch

Quoted 12 lines · Clément Léger

Add description of the switch that is present on the RZ/N1 SoC.

---
 arch/arm/boot/dts/r9a06g032.dtsi | 9 +++++++++
 1 file changed, 9 insertions(+)
arch/arm/boot/dts/r9a06g032.dtsi +0 -0
index 31c4b2e2950a..20d3dce632ce 100644

arch/arm/boot/dts/r9a06g032.dtsi mii_conv5: mii-conv@5 { · line 264

Quoted 10 lines
arch/arm/boot/dts/r9a06g032.dtsi +7 -0
@@ -255,6 +255,15 @@ mii_conv5: mii-conv@5 {
 			};
 		};
 
+		switch: switch@44050000 {
+			compatible = "renesas,r9a06g032-a5psw", "renesas,rzn1-a5psw";
+			reg = <0x44050000 0x10000>;
+			clocks = <&sysctrl R9A06G032_HCLK_SWITCH>,
+				 <&sysctrl R9A06G032_CLK_SWITCH>;
+			clock-names = "hclk", "clk";
+			status = "disabled";

Does the switch port count depend on anything? If it doesn't, maybe you could add the "ethernet-ports" node and all the ports here, with status = "disabled", so that board files don't need to spell them out each time? I'm also thinking you could define the fixed-link and phy-mode = "internal" property of the CPU port with this occasion. That surely isn't a per-board thing.

Quoted 8 lines
arch/arm/boot/dts/r9a06g032.dtsi +2 -0
+		};
+
 		gic: interrupt-controller@44101000 {
 			compatible = "arm,gic-400", "arm,cortex-a7-gic";
 			interrupt-controller;
Signature · 1 line
2.36.0
author maintainer
#38 lore raw

Re: [PATCH net-next v5 12/13] ARM: dts: r9a06g032: describe switch

Quoted 3 lines · Vladimir Oltean

Does the switch port count depend on anything? If it doesn't, maybe you could add the "ethernet-ports" node and all the ports here, with status = "disabled", so that board files don't need to spell them out each time?

Port count does not depends on anything, it's always fixed so indeed, it would be a good idea to provide all the ports as disabled.

Quoted 3 lines

I'm also thinking you could define the fixed-link and phy-mode = "internal" property of the CPU port with this occasion. That surely isn't a per-board thing.

Totally.

Thanks,

Signature · 3 lines
Clément Léger,
Embedded Linux and Kernel engineer at Bootlin
https://bootlin.com
author maintainer
#39 lore raw

PATCH v5 13/13 MAINTAINERS: add Renesas RZ/N1 switch related driver entry

In reply to Clément Léger

After contributing the drivers, volunteer for maintenance and add myself as the maintainer for Renesas RZ/N1 switch related drivers.

---
 MAINTAINERS | 11 +++++++++++
 1 file changed, 11 insertions(+)
MAINTAINERS +11 -0
index 69b597aa4bc7..9f03e4fba1a7 100644
@@ -16889,6 +16889,17 @@ S:	Supported
 F:	Documentation/devicetree/bindings/iio/adc/renesas,rzg2l-adc.yaml
 F:	drivers/iio/adc/rzg2l_adc.c
 
+RENESAS RZ/N1 A5PSW SWITCH DRIVER
+M:	Clément Léger <[email protected]>
+S:	Maintained
+F:	Documentation/devicetree/bindings/net/dsa/renesas,rzn1-a5psw.yaml
+F:	Documentation/devicetree/bindings/net/pcs/renesas,rzn1-miic.yaml
+F:	drivers/net/dsa/rzn1_a5psw*
+F:	drivers/net/pcs/pcs-rzn1-miic.c
+F:	include/dt-bindings/net/pcs-rzn1-miic.h
+F:	include/linux/pcs-rzn1-miic.h
+F:	net/dsa/tag_rzn1_a5psw.c
+
 RENESAS R-CAR GEN3 & RZ/N1 NAND CONTROLLER DRIVER
 M:	Miquel Raynal <[email protected]>
 L:	[email protected]
Signature · 1 line
2.36.0