* [PATCH net-next v11 1/3] dt-bindings: net: nuvoton: Add schema for Nuvoton MA35 family GMAC
2026-02-05 1:40 [PATCH net-next v11 0/3] Add support for Nuvoton MA35D1 GMAC Joey Lu
@ 2026-02-05 1:40 ` Joey Lu
2026-02-05 1:40 ` [PATCH net-next v11 2/3] arm64: dts: nuvoton: Add Ethernet nodes Joey Lu
2026-02-05 1:40 ` [PATCH net-next v11 3/3] net: stmmac: dwmac-nuvoton: Add dwmac glue for Nuvoton MA35 family Joey Lu
2 siblings, 0 replies; 6+ messages in thread
From: Joey Lu @ 2026-02-05 1:40 UTC (permalink / raw)
To: andrew+netdev, davem, edumazet, kuba, pabeni, robh, krzk+dt,
conor+dt, mcoquelin.stm32, richardcochran
Cc: alexandre.torgue, joabreu, ychuang3, schung, yclu4,
peppe.cavallaro, linux-arm-kernel, netdev, devicetree,
linux-kernel, openbmc, linux-stm32, Joey Lu
Create initial schema for Nuvoton MA35 family Gigabit MAC.
Reviewed-by: Rob Herring (Arm) <robh@kernel.org>
Signed-off-by: Joey Lu <a0987203069@gmail.com>
---
.../bindings/net/nuvoton,ma35d1-dwmac.yaml | 126 ++++++++++++++++++
.../devicetree/bindings/net/snps,dwmac.yaml | 1 +
2 files changed, 127 insertions(+)
create mode 100644 Documentation/devicetree/bindings/net/nuvoton,ma35d1-dwmac.yaml
diff --git a/Documentation/devicetree/bindings/net/nuvoton,ma35d1-dwmac.yaml b/Documentation/devicetree/bindings/net/nuvoton,ma35d1-dwmac.yaml
new file mode 100644
index 000000000000..8eaddfdc937c
--- /dev/null
+++ b/Documentation/devicetree/bindings/net/nuvoton,ma35d1-dwmac.yaml
@@ -0,0 +1,126 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/net/nuvoton,ma35d1-dwmac.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Nuvoton DWMAC glue layer controller
+
+maintainers:
+ - Joey Lu <yclu4@nuvoton.com>
+
+description:
+ Nuvoton 10/100/1000Mbps Gigabit Ethernet MAC Controller is based on
+ Synopsys DesignWare MAC (version 3.73a).
+
+allOf:
+ - $ref: snps,dwmac.yaml#
+
+properties:
+ compatible:
+ items:
+ - enum:
+ - nuvoton,ma35d1-dwmac
+
+ reg:
+ maxItems: 1
+ description:
+ Register range should be one of the GMAC interface.
+
+ interrupts:
+ maxItems: 1
+
+ clocks:
+ items:
+ - description: MAC clock
+ - description: PTP clock
+
+ clock-names:
+ items:
+ - const: stmmaceth
+ - const: ptp_ref
+
+ nuvoton,sys:
+ $ref: /schemas/types.yaml#/definitions/phandle-array
+ items:
+ - items:
+ - description: phandle to access syscon registers.
+ - description: GMAC interface ID.
+ enum:
+ - 0
+ - 1
+ description:
+ A phandle to the syscon with one argument that configures system registers
+ for MA35D1's two GMACs. The argument specifies the GMAC interface ID.
+
+ resets:
+ maxItems: 1
+
+ reset-names:
+ items:
+ - const: stmmaceth
+
+ phy-mode:
+ enum:
+ - rmii
+ - rgmii
+ - rgmii-id
+ - rgmii-txid
+ - rgmii-rxid
+
+ tx-internal-delay-ps:
+ default: 0
+ minimum: 0
+ maximum: 2000
+ description:
+ RGMII TX path delay used only when PHY operates in RGMII mode with
+ internal delay (phy-mode is 'rgmii-id' or 'rgmii-txid') in pico-seconds.
+ Allowed values are from 0 to 2000.
+
+ rx-internal-delay-ps:
+ default: 0
+ minimum: 0
+ maximum: 2000
+ description:
+ RGMII RX path delay used only when PHY operates in RGMII mode with
+ internal delay (phy-mode is 'rgmii-id' or 'rgmii-rxid') in pico-seconds.
+ Allowed values are from 0 to 2000.
+
+required:
+ - clocks
+ - clock-names
+ - nuvoton,sys
+ - resets
+ - reset-names
+
+unevaluatedProperties: false
+
+examples:
+ - |
+ #include <dt-bindings/interrupt-controller/arm-gic.h>
+ #include <dt-bindings/clock/nuvoton,ma35d1-clk.h>
+ #include <dt-bindings/reset/nuvoton,ma35d1-reset.h>
+ ethernet@40120000 {
+ compatible = "nuvoton,ma35d1-dwmac";
+ reg = <0x40120000 0x10000>;
+ interrupts = <GIC_SPI 23 IRQ_TYPE_LEVEL_HIGH>;
+ interrupt-names = "macirq";
+ clocks = <&clk EMAC0_GATE>, <&clk EPLL_DIV8>;
+ clock-names = "stmmaceth", "ptp_ref";
+
+ nuvoton,sys = <&sys 0>;
+ resets = <&sys MA35D1_RESET_GMAC0>;
+ reset-names = "stmmaceth";
+
+ phy-mode = "rgmii-id";
+ phy-handle = <ð_phy0>;
+ mdio {
+ compatible = "snps,dwmac-mdio";
+ #address-cells = <1>;
+ #size-cells = <0>;
+
+ eth_phy0: ethernet-phy@0 {
+ reg = <0>;
+ };
+ };
+ };
diff --git a/Documentation/devicetree/bindings/net/snps,dwmac.yaml b/Documentation/devicetree/bindings/net/snps,dwmac.yaml
index dd3c72e8363e..10cb5e555750 100644
--- a/Documentation/devicetree/bindings/net/snps,dwmac.yaml
+++ b/Documentation/devicetree/bindings/net/snps,dwmac.yaml
@@ -69,6 +69,7 @@ properties:
- ingenic,x2000-mac
- loongson,ls2k-dwmac
- loongson,ls7a-dwmac
+ - nuvoton,ma35d1-dwmac
- nxp,s32g2-dwmac
- qcom,qcs404-ethqos
- qcom,sa8775p-ethqos
--
2.43.0
^ permalink raw reply related [flat|nested] 6+ messages in thread* [PATCH net-next v11 2/3] arm64: dts: nuvoton: Add Ethernet nodes
2026-02-05 1:40 [PATCH net-next v11 0/3] Add support for Nuvoton MA35D1 GMAC Joey Lu
2026-02-05 1:40 ` [PATCH net-next v11 1/3] dt-bindings: net: nuvoton: Add schema for Nuvoton MA35 family GMAC Joey Lu
@ 2026-02-05 1:40 ` Joey Lu
2026-02-05 1:40 ` [PATCH net-next v11 3/3] net: stmmac: dwmac-nuvoton: Add dwmac glue for Nuvoton MA35 family Joey Lu
2 siblings, 0 replies; 6+ messages in thread
From: Joey Lu @ 2026-02-05 1:40 UTC (permalink / raw)
To: andrew+netdev, davem, edumazet, kuba, pabeni, robh, krzk+dt,
conor+dt, mcoquelin.stm32, richardcochran
Cc: alexandre.torgue, joabreu, ychuang3, schung, yclu4,
peppe.cavallaro, linux-arm-kernel, netdev, devicetree,
linux-kernel, openbmc, linux-stm32, Joey Lu
Add GMAC nodes for our MA35D1 development boards:
two RGMII interfaces for SOM board, and one RGMII
and one RMII interface for IoT board.
Signed-off-by: Joey Lu <a0987203069@gmail.com>
---
.../boot/dts/nuvoton/ma35d1-iot-512m.dts | 25 +++++++++++
.../boot/dts/nuvoton/ma35d1-som-256m.dts | 24 ++++++++++
arch/arm64/boot/dts/nuvoton/ma35d1.dtsi | 44 +++++++++++++++++++
3 files changed, 93 insertions(+)
diff --git a/arch/arm64/boot/dts/nuvoton/ma35d1-iot-512m.dts b/arch/arm64/boot/dts/nuvoton/ma35d1-iot-512m.dts
index 9482bec1aa57..78534d613486 100644
--- a/arch/arm64/boot/dts/nuvoton/ma35d1-iot-512m.dts
+++ b/arch/arm64/boot/dts/nuvoton/ma35d1-iot-512m.dts
@@ -18,6 +18,8 @@ aliases {
serial12 = &uart12;
serial13 = &uart13;
serial14 = &uart14;
+ ethernet0 = &gmac0;
+ ethernet1 = &gmac1;
};
chosen {
@@ -126,3 +128,26 @@ &uart14 {
pinctrl-0 = <&pinctrl_uart14>;
status = "okay";
};
+
+&gmac0 {
+ status = "okay";
+ phy-handle = <ð_phy0>;
+};
+
+&mdio0 {
+ eth_phy0: ethernet-phy@0 {
+ reg = <0>;
+ };
+};
+
+&gmac1 {
+ status = "okay";
+ phy-mode = "rmii";
+ phy-handle = <ð_phy1>;
+};
+
+&mdio1 {
+ eth_phy1: ethernet-phy@1 {
+ reg = <1>;
+ };
+};
diff --git a/arch/arm64/boot/dts/nuvoton/ma35d1-som-256m.dts b/arch/arm64/boot/dts/nuvoton/ma35d1-som-256m.dts
index f6f20a17e501..a029b660e8dc 100644
--- a/arch/arm64/boot/dts/nuvoton/ma35d1-som-256m.dts
+++ b/arch/arm64/boot/dts/nuvoton/ma35d1-som-256m.dts
@@ -18,6 +18,8 @@ aliases {
serial12 = &uart12;
serial14 = &uart14;
serial16 = &uart16;
+ ethernet0 = &gmac0;
+ ethernet1 = &gmac1;
};
chosen {
@@ -129,3 +131,25 @@ &uart16 {
pinctrl-0 = <&pinctrl_uart16>;
status = "okay";
};
+
+&gmac0 {
+ status = "okay";
+ phy-handle = <ð_phy0>;
+};
+
+&mdio0 {
+ eth_phy0: ethernet-phy@0 {
+ reg = <0>;
+ };
+};
+
+&gmac1 {
+ status = "okay";
+ phy-handle = <ð_phy1>;
+};
+
+&mdio1 {
+ eth_phy1: ethernet-phy@1 {
+ reg = <1>;
+ };
+};
diff --git a/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi b/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi
index e51b98f5bdce..a39c3640b322 100644
--- a/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi
+++ b/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi
@@ -379,5 +379,49 @@ uart16: serial@40880000 {
clocks = <&clk UART16_GATE>;
status = "disabled";
};
+
+ gmac0: ethernet@40120000 {
+ compatible = "nuvoton,ma35d1-dwmac";
+ reg = <0x0 0x40120000 0x0 0x10000>;
+ interrupts = <GIC_SPI 23 IRQ_TYPE_LEVEL_HIGH>;
+ interrupt-names = "macirq";
+ clocks = <&clk EMAC0_GATE>, <&clk EPLL_DIV8>;
+ clock-names = "stmmaceth", "ptp_ref";
+
+ nuvoton,sys = <&sys 0>;
+ resets = <&sys MA35D1_RESET_GMAC0>;
+ reset-names = "stmmaceth";
+
+ phy-mode = "rgmii-id";
+ status = "disabled";
+
+ mdio0: mdio {
+ compatible = "snps,dwmac-mdio";
+ #address-cells = <1>;
+ #size-cells = <0>;
+ };
+ };
+
+ gmac1: ethernet@40130000 {
+ compatible = "nuvoton,ma35d1-dwmac";
+ reg = <0x0 0x40130000 0x0 0x10000>;
+ interrupts = <GIC_SPI 24 IRQ_TYPE_LEVEL_HIGH>;
+ interrupt-names = "macirq";
+ clocks = <&clk EMAC1_GATE>, <&clk EPLL_DIV8>;
+ clock-names = "stmmaceth", "ptp_ref";
+
+ nuvoton,sys = <&sys 1>;
+ resets = <&sys MA35D1_RESET_GMAC1>;
+ reset-names = "stmmaceth";
+
+ phy-mode = "rgmii-id";
+ status = "disabled";
+
+ mdio1: mdio {
+ compatible = "snps,dwmac-mdio";
+ #address-cells = <1>;
+ #size-cells = <0>;
+ };
+ };
};
};
--
2.43.0
^ permalink raw reply related [flat|nested] 6+ messages in thread* [PATCH net-next v11 3/3] net: stmmac: dwmac-nuvoton: Add dwmac glue for Nuvoton MA35 family
2026-02-05 1:40 [PATCH net-next v11 0/3] Add support for Nuvoton MA35D1 GMAC Joey Lu
2026-02-05 1:40 ` [PATCH net-next v11 1/3] dt-bindings: net: nuvoton: Add schema for Nuvoton MA35 family GMAC Joey Lu
2026-02-05 1:40 ` [PATCH net-next v11 2/3] arm64: dts: nuvoton: Add Ethernet nodes Joey Lu
@ 2026-02-05 1:40 ` Joey Lu
2026-02-05 9:38 ` Russell King (Oracle)
2 siblings, 1 reply; 6+ messages in thread
From: Joey Lu @ 2026-02-05 1:40 UTC (permalink / raw)
To: andrew+netdev, davem, edumazet, kuba, pabeni, robh, krzk+dt,
conor+dt, mcoquelin.stm32, richardcochran
Cc: alexandre.torgue, joabreu, ychuang3, schung, yclu4,
peppe.cavallaro, linux-arm-kernel, netdev, devicetree,
linux-kernel, openbmc, linux-stm32, Joey Lu, Andrew Lunn
Add support for Gigabit Ethernet on Nuvoton MA35 series using dwmac driver.
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Joey Lu <a0987203069@gmail.com>
---
drivers/net/ethernet/stmicro/stmmac/Kconfig | 12 ++
drivers/net/ethernet/stmicro/stmmac/Makefile | 1 +
.../ethernet/stmicro/stmmac/dwmac-nuvoton.c | 173 ++++++++++++++++++
3 files changed, 186 insertions(+)
create mode 100644 drivers/net/ethernet/stmicro/stmmac/dwmac-nuvoton.c
diff --git a/drivers/net/ethernet/stmicro/stmmac/Kconfig b/drivers/net/ethernet/stmicro/stmmac/Kconfig
index 07088d03dbab..861f1c6c14f1 100644
--- a/drivers/net/ethernet/stmicro/stmmac/Kconfig
+++ b/drivers/net/ethernet/stmicro/stmmac/Kconfig
@@ -132,6 +132,18 @@ config DWMAC_MESON
the stmmac device driver. This driver is used for Meson6,
Meson8, Meson8b and GXBB SoCs.
+config DWMAC_NUVOTON
+ tristate "Nuvoton MA35 dwmac support"
+ default ARCH_MA35
+ depends on OF && (ARCH_MA35 || COMPILE_TEST)
+ select MFD_SYSCON
+ help
+ Support for Ethernet controller on Nuvoton MA35 series SoC.
+
+ This selects the Nuvoton MA35 series SoC glue layer support
+ for the stmmac device driver. The nuvoton-dwmac driver is
+ used for MA35 series SoCs.
+
config DWMAC_QCOM_ETHQOS
tristate "Qualcomm ETHQOS support"
default ARCH_QCOM
diff --git a/drivers/net/ethernet/stmicro/stmmac/Makefile b/drivers/net/ethernet/stmicro/stmmac/Makefile
index c9263987ef8d..4ade030b634f 100644
--- a/drivers/net/ethernet/stmicro/stmmac/Makefile
+++ b/drivers/net/ethernet/stmicro/stmmac/Makefile
@@ -20,6 +20,7 @@ obj-$(CONFIG_DWMAC_IPQ806X) += dwmac-ipq806x.o
obj-$(CONFIG_DWMAC_LPC18XX) += dwmac-lpc18xx.o
obj-$(CONFIG_DWMAC_MEDIATEK) += dwmac-mediatek.o
obj-$(CONFIG_DWMAC_MESON) += dwmac-meson.o dwmac-meson8b.o
+obj-$(CONFIG_DWMAC_NUVOTON) += dwmac-nuvoton.o
obj-$(CONFIG_DWMAC_QCOM_ETHQOS) += dwmac-qcom-ethqos.o
obj-$(CONFIG_DWMAC_RENESAS_GBETH) += dwmac-renesas-gbeth.o
obj-$(CONFIG_DWMAC_ROCKCHIP) += dwmac-rk.o
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-nuvoton.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-nuvoton.c
new file mode 100644
index 000000000000..dd233e0a32ee
--- /dev/null
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-nuvoton.c
@@ -0,0 +1,173 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Nuvoton DWMAC specific glue layer
+ *
+ * Copyright (C) 2025 Nuvoton Technology Corp.
+ *
+ * Author: Joey Lu <a0987203069@gmail.com>
+ */
+
+#include <linux/mfd/syscon.h>
+#include <linux/of_device.h>
+#include <linux/of_net.h>
+#include <linux/platform_device.h>
+#include <linux/regmap.h>
+#include <linux/stmmac.h>
+
+#include "stmmac.h"
+#include "stmmac_platform.h"
+
+#define NVT_REG_SYS_GMAC0MISCR 0x108
+#define NVT_REG_SYS_GMAC1MISCR 0x10C
+
+#define NVT_MISCR_RMII BIT(0)
+
+/* Two thousand picoseconds are evenly mapped to a 4-bit field,
+ * resulting in each step being 2000/15 picoseconds.
+ */
+#define NVT_PATH_DELAY_STEP 134
+#define NVT_TX_DELAY_MASK GENMASK(19, 16)
+#define NVT_RX_DELAY_MASK GENMASK(23, 20)
+
+struct nvt_priv_data {
+ struct platform_device *pdev;
+ struct regmap *regmap;
+};
+
+static struct nvt_priv_data *
+nvt_gmac_setup(struct platform_device *pdev, struct plat_stmmacenet_data *plat)
+{
+ struct device *dev = &pdev->dev;
+ struct nvt_priv_data *bsp_priv;
+ phy_interface_t phy_mode;
+ u32 macid, arg, reg;
+ u32 tx_delay_step;
+ u32 rx_delay_step;
+ u32 miscr;
+
+ bsp_priv = devm_kzalloc(dev, sizeof(*bsp_priv), GFP_KERNEL);
+ if (!bsp_priv)
+ return ERR_PTR(-ENOMEM);
+
+ bsp_priv->regmap =
+ syscon_regmap_lookup_by_phandle_args(dev->of_node, "nuvoton,sys", 1, &macid);
+ if (IS_ERR(bsp_priv->regmap))
+ return ERR_PTR(dev_err_probe(dev, PTR_ERR(bsp_priv->regmap),
+ "Failed to get sys register\n"));
+ if (macid > 1) {
+ dev_err(dev, "Invalid sys arguments\n");
+ return ERR_PTR(-EINVAL);
+ }
+
+ if (of_property_read_u32(dev->of_node, "tx-internal-delay-ps", &arg)) {
+ tx_delay_step = 0;
+ } else {
+ if (arg <= 2000) {
+ tx_delay_step = (arg == 2000) ? 0xf : (arg / NVT_PATH_DELAY_STEP);
+ dev_dbg(dev, "Set Tx path delay to 0x%x\n", tx_delay_step);
+ } else {
+ dev_err(dev, "Invalid Tx path delay argument.\n");
+ return ERR_PTR(-EINVAL);
+ }
+ }
+ if (of_property_read_u32(dev->of_node, "rx-internal-delay-ps", &arg)) {
+ rx_delay_step = 0;
+ } else {
+ if (arg <= 2000) {
+ rx_delay_step = (arg == 2000) ? 0xf : (arg / NVT_PATH_DELAY_STEP);
+ dev_dbg(dev, "Set Rx path delay to 0x%x\n", rx_delay_step);
+ } else {
+ dev_err(dev, "Invalid Rx path delay argument.\n");
+ return ERR_PTR(-EINVAL);
+ }
+ }
+
+ miscr = (macid == 0) ? NVT_REG_SYS_GMAC0MISCR : NVT_REG_SYS_GMAC1MISCR;
+ regmap_read(bsp_priv->regmap, miscr, ®);
+ reg &= ~(NVT_TX_DELAY_MASK | NVT_RX_DELAY_MASK);
+
+ if (of_get_phy_mode(pdev->dev.of_node, &phy_mode)) {
+ dev_err(dev, "missing phy mode property\n");
+ return ERR_PTR(-EINVAL);
+ }
+
+ switch (phy_mode) {
+ case PHY_INTERFACE_MODE_RGMII:
+ case PHY_INTERFACE_MODE_RGMII_ID:
+ case PHY_INTERFACE_MODE_RGMII_RXID:
+ case PHY_INTERFACE_MODE_RGMII_TXID:
+ reg &= ~NVT_MISCR_RMII;
+ break;
+ case PHY_INTERFACE_MODE_RMII:
+ reg |= NVT_MISCR_RMII;
+ break;
+ default:
+ dev_err(dev, "Unsupported phy-mode (%d)\n", phy_mode);
+ return ERR_PTR(-EINVAL);
+ }
+
+ if (!(reg & NVT_MISCR_RMII)) {
+ reg |= FIELD_PREP(NVT_TX_DELAY_MASK, tx_delay_step);
+ reg |= FIELD_PREP(NVT_RX_DELAY_MASK, rx_delay_step);
+ }
+
+ regmap_write(bsp_priv->regmap, miscr, reg);
+
+ bsp_priv->pdev = pdev;
+
+ return bsp_priv;
+}
+
+static int nvt_gmac_probe(struct platform_device *pdev)
+{
+ struct plat_stmmacenet_data *plat_dat;
+ struct stmmac_resources stmmac_res;
+ struct nvt_priv_data *priv_data;
+ int ret;
+
+ ret = stmmac_get_platform_resources(pdev, &stmmac_res);
+ if (ret)
+ return ret;
+
+ plat_dat = devm_stmmac_probe_config_dt(pdev, stmmac_res.mac);
+ if (IS_ERR(plat_dat))
+ return PTR_ERR(plat_dat);
+
+ /* Nuvoton DWMAC configs */
+ plat_dat->core_type = DWMAC_CORE_GMAC;
+ plat_dat->tx_fifo_size = 2048;
+ plat_dat->rx_fifo_size = 4096;
+ plat_dat->multicast_filter_bins = 0;
+ plat_dat->unicast_filter_entries = 8;
+
+ priv_data = nvt_gmac_setup(pdev, plat_dat);
+ if (IS_ERR(priv_data))
+ return PTR_ERR(priv_data);
+
+ ret = stmmac_pltfr_probe(pdev, plat_dat, &stmmac_res);
+ if (ret)
+ return ret;
+
+ return 0;
+}
+
+static const struct of_device_id nvt_dwmac_match[] = {
+ { .compatible = "nuvoton,ma35d1-dwmac"},
+ { }
+};
+MODULE_DEVICE_TABLE(of, nvt_dwmac_match);
+
+static struct platform_driver nvt_dwmac_driver = {
+ .probe = nvt_gmac_probe,
+ .remove = stmmac_pltfr_remove,
+ .driver = {
+ .name = "nuvoton-dwmac",
+ .pm = &stmmac_pltfr_pm_ops,
+ .of_match_table = nvt_dwmac_match,
+ },
+};
+module_platform_driver(nvt_dwmac_driver);
+
+MODULE_AUTHOR("Joey Lu <a0987203069@gmail.com>");
+MODULE_DESCRIPTION("Nuvoton DWMAC specific glue layer");
+MODULE_LICENSE("GPL");
--
2.43.0
^ permalink raw reply related [flat|nested] 6+ messages in thread* Re: [PATCH net-next v11 3/3] net: stmmac: dwmac-nuvoton: Add dwmac glue for Nuvoton MA35 family
2026-02-05 1:40 ` [PATCH net-next v11 3/3] net: stmmac: dwmac-nuvoton: Add dwmac glue for Nuvoton MA35 family Joey Lu
@ 2026-02-05 9:38 ` Russell King (Oracle)
2026-02-06 9:53 ` Joey Lu
0 siblings, 1 reply; 6+ messages in thread
From: Russell King (Oracle) @ 2026-02-05 9:38 UTC (permalink / raw)
To: Joey Lu
Cc: andrew+netdev, davem, edumazet, kuba, pabeni, robh, krzk+dt,
conor+dt, mcoquelin.stm32, richardcochran, alexandre.torgue,
joabreu, ychuang3, schung, yclu4, peppe.cavallaro,
linux-arm-kernel, netdev, devicetree, linux-kernel, openbmc,
linux-stm32, Andrew Lunn
Hi,
On Thu, Feb 05, 2026 at 09:40:05AM +0800, Joey Lu wrote:
> +
> +struct nvt_priv_data {
> + struct platform_device *pdev;
This looks to me like it's write-only, does it serve a useful purpose?
> + struct regmap *regmap;
This doesn't seem to be used outside of nvt_gmac_setup().
> +};
Given the above two comments, do you actually need struct nvt_priv_data ?
> +
> +static struct nvt_priv_data *
> +nvt_gmac_setup(struct platform_device *pdev, struct plat_stmmacenet_data *plat)
> +{
> + struct device *dev = &pdev->dev;
> + struct nvt_priv_data *bsp_priv;
> + phy_interface_t phy_mode;
> + u32 macid, arg, reg;
> + u32 tx_delay_step;
> + u32 rx_delay_step;
> + u32 miscr;
> +
> + bsp_priv = devm_kzalloc(dev, sizeof(*bsp_priv), GFP_KERNEL);
> + if (!bsp_priv)
> + return ERR_PTR(-ENOMEM);
> +
> + bsp_priv->regmap =
> + syscon_regmap_lookup_by_phandle_args(dev->of_node, "nuvoton,sys", 1, &macid);
> + if (IS_ERR(bsp_priv->regmap))
> + return ERR_PTR(dev_err_probe(dev, PTR_ERR(bsp_priv->regmap),
> + "Failed to get sys register\n"));
> + if (macid > 1) {
> + dev_err(dev, "Invalid sys arguments\n");
> + return ERR_PTR(-EINVAL);
> + }
> +
> + if (of_property_read_u32(dev->of_node, "tx-internal-delay-ps", &arg)) {
> + tx_delay_step = 0;
> + } else {
> + if (arg <= 2000) {
> + tx_delay_step = (arg == 2000) ? 0xf : (arg / NVT_PATH_DELAY_STEP);
> + dev_dbg(dev, "Set Tx path delay to 0x%x\n", tx_delay_step);
> + } else {
> + dev_err(dev, "Invalid Tx path delay argument.\n");
> + return ERR_PTR(-EINVAL);
> + }
> + }
> + if (of_property_read_u32(dev->of_node, "rx-internal-delay-ps", &arg)) {
> + rx_delay_step = 0;
> + } else {
> + if (arg <= 2000) {
> + rx_delay_step = (arg == 2000) ? 0xf : (arg / NVT_PATH_DELAY_STEP);
> + dev_dbg(dev, "Set Rx path delay to 0x%x\n", rx_delay_step);
> + } else {
> + dev_err(dev, "Invalid Rx path delay argument.\n");
> + return ERR_PTR(-EINVAL);
> + }
> + }
Each of these could be moved into a separate function:
static int nvt_gmac_get_delay(struct device *dev, const char *property)
{
u32 arg;
if (of_property_read_u32(dev->of_node, property, &arg))
return 0;
if (arg > 2000) {
dev_err(dev, "Invalid %s argument.\n", property);
return -EINVAL;
}
if (arg == 2000)
return 15;
return arg / NVT_PATH_DELAY_STEP;
}
then:
int ret;
ret = nvt_gmac_get_delay(dev, "tx-internal-delay-ps");
if (ret < 0)
return ERR_PTR(ret);
tx_delay = ret;
ret = nvt_gmac_get_delay(dev, "rx-internal-delay-ps");
if (ret < 0)
return ERR_PTR(ret);
rx_delay = ret;
> +
> + miscr = (macid == 0) ? NVT_REG_SYS_GMAC0MISCR : NVT_REG_SYS_GMAC1MISCR;
> + regmap_read(bsp_priv->regmap, miscr, ®);
> + reg &= ~(NVT_TX_DELAY_MASK | NVT_RX_DELAY_MASK);
> +
> + if (of_get_phy_mode(pdev->dev.of_node, &phy_mode)) {
> + dev_err(dev, "missing phy mode property\n");
> + return ERR_PTR(-EINVAL);
> + }
> +
> + switch (phy_mode) {
> + case PHY_INTERFACE_MODE_RGMII:
> + case PHY_INTERFACE_MODE_RGMII_ID:
> + case PHY_INTERFACE_MODE_RGMII_RXID:
> + case PHY_INTERFACE_MODE_RGMII_TXID:
> + reg &= ~NVT_MISCR_RMII;
> + break;
> + case PHY_INTERFACE_MODE_RMII:
> + reg |= NVT_MISCR_RMII;
> + break;
> + default:
> + dev_err(dev, "Unsupported phy-mode (%d)\n", phy_mode);
> + return ERR_PTR(-EINVAL);
> + }
> +
> + if (!(reg & NVT_MISCR_RMII)) {
> + reg |= FIELD_PREP(NVT_TX_DELAY_MASK, tx_delay_step);
> + reg |= FIELD_PREP(NVT_RX_DELAY_MASK, rx_delay_step);
You can move this inside the switch above under the RGMII case. Theses
delays are, after all, only for RGMII.
> + }
> +
> + regmap_write(bsp_priv->regmap, miscr, reg);
Consider:
regmap_update_bits(bsp_priv->regmap, miscr,
NVT_TX_DELAY_MASK | NVT_RX_DELAY_MASK |
NVT_MISCR_RMII, reg);
> + plat_dat = devm_stmmac_probe_config_dt(pdev, stmmac_res.mac);
> + if (IS_ERR(plat_dat))
> + return PTR_ERR(plat_dat);
> +
> + /* Nuvoton DWMAC configs */
> + plat_dat->core_type = DWMAC_CORE_GMAC;
Is the hardware not compatible with any of the compatible types that
devm_stmmac_probe_config_dt() will automatically set this for you?
Which version of the core do you have?
> + plat_dat->tx_fifo_size = 2048;
> + plat_dat->rx_fifo_size = 4096;
There are tx-fifo-depth / rx-fifo-depth properties that can be used to
describe these in DT.
> + plat_dat->multicast_filter_bins = 0;
> + plat_dat->unicast_filter_entries = 8;
If this core is v3.50, v3.70 or v3.72, then there are
snps,multicast-filter-bins and snps,perfect-filter-entries which
can be used to describe both of these.
Thanks.
--
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 80Mbps down 10Mbps up. Decent connectivity at last!
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [PATCH net-next v11 3/3] net: stmmac: dwmac-nuvoton: Add dwmac glue for Nuvoton MA35 family
2026-02-05 9:38 ` Russell King (Oracle)
@ 2026-02-06 9:53 ` Joey Lu
0 siblings, 0 replies; 6+ messages in thread
From: Joey Lu @ 2026-02-06 9:53 UTC (permalink / raw)
To: Russell King (Oracle)
Cc: andrew+netdev, davem, edumazet, kuba, pabeni, robh, krzk+dt,
conor+dt, mcoquelin.stm32, richardcochran, alexandre.torgue,
joabreu, ychuang3, schung, yclu4, peppe.cavallaro,
linux-arm-kernel, netdev, devicetree, linux-kernel, openbmc,
linux-stm32, Andrew Lunn
On 2/5/2026 5:38 PM, Russell King (Oracle) wrote:
> Hi,
>
> On Thu, Feb 05, 2026 at 09:40:05AM +0800, Joey Lu wrote:
>> +
>> +struct nvt_priv_data {
>> + struct platform_device *pdev;
> This looks to me like it's write-only, does it serve a useful purpose?
>
>> + struct regmap *regmap;
> This doesn't seem to be used outside of nvt_gmac_setup().
>
>> +};
> Given the above two comments, do you actually need struct nvt_priv_data ?
You are right. I'll drop it in the next revision.
>
>> +
>> +static struct nvt_priv_data *
>> +nvt_gmac_setup(struct platform_device *pdev, struct plat_stmmacenet_data *plat)
>> +{
>> + struct device *dev = &pdev->dev;
>> + struct nvt_priv_data *bsp_priv;
>> + phy_interface_t phy_mode;
>> + u32 macid, arg, reg;
>> + u32 tx_delay_step;
>> + u32 rx_delay_step;
>> + u32 miscr;
>> +
>> + bsp_priv = devm_kzalloc(dev, sizeof(*bsp_priv), GFP_KERNEL);
>> + if (!bsp_priv)
>> + return ERR_PTR(-ENOMEM);
>> +
>> + bsp_priv->regmap =
>> + syscon_regmap_lookup_by_phandle_args(dev->of_node, "nuvoton,sys", 1, &macid);
>> + if (IS_ERR(bsp_priv->regmap))
>> + return ERR_PTR(dev_err_probe(dev, PTR_ERR(bsp_priv->regmap),
>> + "Failed to get sys register\n"));
>> + if (macid > 1) {
>> + dev_err(dev, "Invalid sys arguments\n");
>> + return ERR_PTR(-EINVAL);
>> + }
>> +
>> + if (of_property_read_u32(dev->of_node, "tx-internal-delay-ps", &arg)) {
>> + tx_delay_step = 0;
>> + } else {
>> + if (arg <= 2000) {
>> + tx_delay_step = (arg == 2000) ? 0xf : (arg / NVT_PATH_DELAY_STEP);
>> + dev_dbg(dev, "Set Tx path delay to 0x%x\n", tx_delay_step);
>> + } else {
>> + dev_err(dev, "Invalid Tx path delay argument.\n");
>> + return ERR_PTR(-EINVAL);
>> + }
>> + }
>> + if (of_property_read_u32(dev->of_node, "rx-internal-delay-ps", &arg)) {
>> + rx_delay_step = 0;
>> + } else {
>> + if (arg <= 2000) {
>> + rx_delay_step = (arg == 2000) ? 0xf : (arg / NVT_PATH_DELAY_STEP);
>> + dev_dbg(dev, "Set Rx path delay to 0x%x\n", rx_delay_step);
>> + } else {
>> + dev_err(dev, "Invalid Rx path delay argument.\n");
>> + return ERR_PTR(-EINVAL);
>> + }
>> + }
> Each of these could be moved into a separate function:
>
> static int nvt_gmac_get_delay(struct device *dev, const char *property)
> {
> u32 arg;
>
> if (of_property_read_u32(dev->of_node, property, &arg))
> return 0;
>
> if (arg > 2000) {
> dev_err(dev, "Invalid %s argument.\n", property);
> return -EINVAL;
> }
>
> if (arg == 2000)
> return 15;
>
> return arg / NVT_PATH_DELAY_STEP;
> }
>
> then:
> int ret;
>
> ret = nvt_gmac_get_delay(dev, "tx-internal-delay-ps");
> if (ret < 0)
> return ERR_PTR(ret);
>
> tx_delay = ret;
>
> ret = nvt_gmac_get_delay(dev, "rx-internal-delay-ps");
> if (ret < 0)
> return ERR_PTR(ret);
>
> rx_delay = ret;
I'll update the code according to your suggestions.
>> +
>> + miscr = (macid == 0) ? NVT_REG_SYS_GMAC0MISCR : NVT_REG_SYS_GMAC1MISCR;
>> + regmap_read(bsp_priv->regmap, miscr, ®);
>> + reg &= ~(NVT_TX_DELAY_MASK | NVT_RX_DELAY_MASK);
>> +
>> + if (of_get_phy_mode(pdev->dev.of_node, &phy_mode)) {
>> + dev_err(dev, "missing phy mode property\n");
>> + return ERR_PTR(-EINVAL);
>> + }
>> +
>> + switch (phy_mode) {
>> + case PHY_INTERFACE_MODE_RGMII:
>> + case PHY_INTERFACE_MODE_RGMII_ID:
>> + case PHY_INTERFACE_MODE_RGMII_RXID:
>> + case PHY_INTERFACE_MODE_RGMII_TXID:
>> + reg &= ~NVT_MISCR_RMII;
>> + break;
>> + case PHY_INTERFACE_MODE_RMII:
>> + reg |= NVT_MISCR_RMII;
>> + break;
>> + default:
>> + dev_err(dev, "Unsupported phy-mode (%d)\n", phy_mode);
>> + return ERR_PTR(-EINVAL);
>> + }
>> +
>> + if (!(reg & NVT_MISCR_RMII)) {
>> + reg |= FIELD_PREP(NVT_TX_DELAY_MASK, tx_delay_step);
>> + reg |= FIELD_PREP(NVT_RX_DELAY_MASK, rx_delay_step);
> You can move this inside the switch above under the RGMII case. Theses
> delays are, after all, only for RGMII.
Got it. I'll move them into the RGMII case.
>> + }
>> +
>> + regmap_write(bsp_priv->regmap, miscr, reg);
> Consider:
>
> regmap_update_bits(bsp_priv->regmap, miscr,
> NVT_TX_DELAY_MASK | NVT_RX_DELAY_MASK |
> NVT_MISCR_RMII, reg);
>
>> + plat_dat = devm_stmmac_probe_config_dt(pdev, stmmac_res.mac);
>> + if (IS_ERR(plat_dat))
>> + return PTR_ERR(plat_dat);
>> +
>> + /* Nuvoton DWMAC configs */
>> + plat_dat->core_type = DWMAC_CORE_GMAC;
> Is the hardware not compatible with any of the compatible types that
> devm_stmmac_probe_config_dt() will automatically set this for you?
> Which version of the core do you have?
>
>> + plat_dat->tx_fifo_size = 2048;
>> + plat_dat->rx_fifo_size = 4096;
> There are tx-fifo-depth / rx-fifo-depth properties that can be used to
> describe these in DT.
>
>> + plat_dat->multicast_filter_bins = 0;
>> + plat_dat->unicast_filter_entries = 8;
> If this core is v3.50, v3.70 or v3.72, then there are
> snps,multicast-filter-bins and snps,perfect-filter-entries which
> can be used to describe both of these.
>
> Thanks.
Thanks for the feedback.
This GMAC is based on v3.73a. While this specific revision isn’t
explicitly documented in the current DT binding YAML, the relevant FIFO
sizing and filter capabilities match the behavior introduced in earlier
v3.70+ cores.
Given that, I agree it makes sense to describe these parameters using
the existing DT properties.
I will update the DT and driver accordingly in the next revision.
Joey
>
^ permalink raw reply [flat|nested] 6+ messages in thread