[PATCH v3 0/4] net: stmmac: Add Toshiba Visconti SoCs glue driver

[PATCH v3 0/4] net: stmmac: Add Toshiba Visconti SoCs glue driver

[Nobuhiro Iwamatsu]

Hi,

This series is the ethernet driver for Toshiba's ARM SoC, Visconti[0].
This provides DT binding documentation, device driver, MAINTAINER files,
and updates to DT files.

Best regards,
  Nobuhiro

[0]:
https://toshiba.semicon-storage.com/ap-en/semiconductor/product/image-recognition-processors-visconti.htmli

Updates:

  dt-bindings: net: Add DT bindings for Toshiba Visconti TMPV7700 SoC
    v2 -> v3: Change to the correct dwmac version.
              Remove description of reg property.
              Update examples, drop dma-coherent, add snps,tso, fix phy-mode.
    v1 -> v2: No update.

  net: stmmac: Add Toshiba Visconti SoCs glue driver
    v2 -> v3: Remove code that is no longer needed by using compatible string.
              Drop def_bool y in Kconfig
    v1 -> v2: Use reverse christmas tree ordering for local variable declarations.

  MAINTAINERS: Add entries for Toshiba Visconti ethernet controller
    v2 -> v3: No update.
    v1 -> v2: No update.

  arm: dts: visconti: Add DT support for Toshiba Visconti5 ethernet controller
    v2 -> v3: Add "snps,dwmac-4.20a" as compatible string.
              Add snps,tso.
    v1 -> v2: No update.

Nobuhiro Iwamatsu (4):
  dt-bindings: net: Add DT bindings for Toshiba Visconti TMPV7700 SoC
  net: stmmac: Add Toshiba Visconti SoCs glue driver
  MAINTAINERS: Add entries for Toshiba Visconti ethernet controller
  arm: dts: visconti: Add DT support for Toshiba Visconti5 ethernet
    controller

 .../bindings/net/toshiba,visconti-dwmac.yaml  |  85 ++++++
 MAINTAINERS                                   |   2 +
 .../boot/dts/toshiba/tmpv7708-rm-mbrc.dts     |  18 ++
 arch/arm64/boot/dts/toshiba/tmpv7708.dtsi     |  25 ++
 drivers/net/ethernet/stmicro/stmmac/Kconfig   |   8 +
 drivers/net/ethernet/stmicro/stmmac/Makefile  |   1 +
 .../ethernet/stmicro/stmmac/dwmac-visconti.c  | 285 ++++++++++++++++++
 7 files changed, 424 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/net/toshiba,visconti-dwmac.yaml
 create mode 100644 drivers/net/ethernet/stmicro/stmmac/dwmac-visconti.c

-- 
2.30.0.rc2

[PATCH 1/4] dt-bindings: net: Add DT bindings for Toshiba Visconti TMPV7700 SoC

[Nobuhiro Iwamatsu]

Add device tree bindings for ethernet controller of Toshiba Visconti
TMPV7700 SoC series.

Signed-off-by: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp>
---
 .../bindings/net/toshiba,visconti-dwmac.yaml  | 85 +++++++++++++++++++
 1 file changed, 85 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/net/toshiba,visconti-dwmac.yaml

diff --git a/Documentation/devicetree/bindings/net/toshiba,visconti-dwmac.yaml b/Documentation/devicetree/bindings/net/toshiba,visconti-dwmac.yaml
new file mode 100644
index 000000000000..59724d18e6f3
--- /dev/null
+++ b/Documentation/devicetree/bindings/net/toshiba,visconti-dwmac.yaml
@@ -0,0 +1,85 @@
+# SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: "http://devicetree.org/schemas/net/toshiba,visconti-dwmac.yaml#"
+$schema: "http://devicetree.org/meta-schemas/core.yaml#"
+
+title: Toshiba Visconti DWMAC Ethernet controller
+
+maintainers:
+  - Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp>
+
+select:
+  properties:
+    compatible:
+      contains:
+        enum:
+          - toshiba,visconti-dwmac
+  required:
+    - compatible
+
+allOf:
+  - $ref: "snps,dwmac.yaml#"
+
+properties:
+  compatible:
+    oneOf:
+      - items:
+          - enum:
+              - toshiba,visconti-dwmac
+          - const: snps,dwmac-4.20a
+
+  reg:
+    maxItems: 1
+
+  clocks:
+    items:
+      - description: main clock
+      - description: PHY reference clock
+
+  clock-names:
+    items:
+      - const: stmmaceth
+      - const: phy_ref_clk
+
+required:
+  - compatible
+  - reg
+  - clocks
+  - clock-names
+
+unevaluatedProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/interrupt-controller/arm-gic.h>
+
+    soc {
+        #address-cells = <2>;
+        #size-cells = <2>;
+
+        piether: ethernet@28000000 {
+            compatible = "toshiba,visconti-dwmac", "snps,dwmac-4.20a";
+            reg = <0 0x28000000 0 0x10000>;
+            interrupts = <GIC_SPI 156 IRQ_TYPE_LEVEL_HIGH>;
+            interrupt-names = "macirq";
+            clocks = <&clk300mhz>, <&clk125mhz>;
+            clock-names = "stmmaceth", "phy_ref_clk";
+            snps,txpbl = <4>;
+            snps,rxpbl = <4>;
+            snps,tso;
+            phy-mode = "rgmii-id";
+            phy-handle = <&phy0>;
+
+            mdio0 {
+                #address-cells = <0x1>;
+                #size-cells = <0x0>;
+                compatible = "snps,dwmac-mdio";
+
+                phy0: ethernet-phy@1 {
+                    device_type = "ethernet-phy";
+                    reg = <0x1>;
+                };
+            };
+        };
+    };
-- 
2.30.0.rc2

[PATCH 2/4] net: stmmac: Add Toshiba Visconti SoCs glue driver

[Nobuhiro Iwamatsu]

Add dwmac-visconti to the stmmac driver in Toshiba Visconti ARM SoCs.
This patch contains only the basic function of the device. There is no
clock control, PM, etc. yet. These will be added in the future.

Signed-off-by: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp>
---
 drivers/net/ethernet/stmicro/stmmac/Kconfig   |   8 +
 drivers/net/ethernet/stmicro/stmmac/Makefile  |   1 +
 .../ethernet/stmicro/stmmac/dwmac-visconti.c  | 285 ++++++++++++++++++
 3 files changed, 294 insertions(+)
 create mode 100644 drivers/net/ethernet/stmicro/stmmac/dwmac-visconti.c

diff --git a/drivers/net/ethernet/stmicro/stmmac/Kconfig b/drivers/net/ethernet/stmicro/stmmac/Kconfig
index 53f14c5a9e02..55ba67a550b9 100644
--- a/drivers/net/ethernet/stmicro/stmmac/Kconfig
+++ b/drivers/net/ethernet/stmicro/stmmac/Kconfig
@@ -219,6 +219,14 @@ config DWMAC_INTEL_PLAT
 	  This selects the Intel platform specific glue layer support for
 	  the stmmac device driver. This driver is used for the Intel Keem Bay
 	  SoC.
+
+config DWMAC_VISCONTI
+	bool "Toshiba Visconti DWMAC support"
+	def_bool y
+	depends on OF && COMMON_CLK && (ARCH_VISCONTI || COMPILE_TEST)
+	help
+	  Support for ethernet controller on Visconti SoCs.
+
 endif
 
 config DWMAC_INTEL
diff --git a/drivers/net/ethernet/stmicro/stmmac/Makefile b/drivers/net/ethernet/stmicro/stmmac/Makefile
index 24e6145d4eae..366740ab9c5a 100644
--- a/drivers/net/ethernet/stmicro/stmmac/Makefile
+++ b/drivers/net/ethernet/stmicro/stmmac/Makefile
@@ -29,6 +29,7 @@ obj-$(CONFIG_DWMAC_DWC_QOS_ETH)	+= dwmac-dwc-qos-eth.o
 obj-$(CONFIG_DWMAC_INTEL_PLAT)	+= dwmac-intel-plat.o
 obj-$(CONFIG_DWMAC_GENERIC)	+= dwmac-generic.o
 obj-$(CONFIG_DWMAC_IMX8)	+= dwmac-imx.o
+obj-$(CONFIG_DWMAC_VISCONTI)	+= dwmac-visconti.o
 stmmac-platform-objs:= stmmac_platform.o
 dwmac-altr-socfpga-objs := altr_tse_pcs.o dwmac-socfpga.o
 
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-visconti.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-visconti.c
new file mode 100644
index 000000000000..b7a0c57dfbfb
--- /dev/null
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-visconti.c
@@ -0,0 +1,285 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Toshiba Visconti Ethernet Support
+ *
+ * (C) Copyright 2020 TOSHIBA CORPORATION
+ * (C) Copyright 2020 Toshiba Electronic Devices & Storage Corporation
+ */
+
+#include <linux/module.h>
+#include <linux/of_device.h>
+#include <linux/of_net.h>
+#include <linux/stmmac.h>
+
+#include "stmmac_platform.h"
+#include "dwmac4.h"
+
+#define REG_ETHER_CONTROL	0x52D4
+#define ETHER_ETH_CONTROL_RESET BIT(17)
+
+#define REG_ETHER_CLOCK_SEL	0x52D0
+#define ETHER_CLK_SEL_TX_CLK_EN BIT(0)
+#define ETHER_CLK_SEL_RX_CLK_EN BIT(1)
+#define ETHER_CLK_SEL_RMII_CLK_EN BIT(2)
+#define ETHER_CLK_SEL_RMII_CLK_RST BIT(3)
+#define ETHER_CLK_SEL_DIV_SEL_2 BIT(4)
+#define ETHER_CLK_SEL_DIV_SEL_20 BIT(0)
+#define ETHER_CLK_SEL_FREQ_SEL_125M	(BIT(9) | BIT(8))
+#define ETHER_CLK_SEL_FREQ_SEL_50M	BIT(9)
+#define ETHER_CLK_SEL_FREQ_SEL_25M	BIT(8)
+#define ETHER_CLK_SEL_FREQ_SEL_2P5M	BIT(0)
+#define ETHER_CLK_SEL_TX_CLK_EXT_SEL_IN BIT(0)
+#define ETHER_CLK_SEL_TX_CLK_EXT_SEL_TXC BIT(10)
+#define ETHER_CLK_SEL_TX_CLK_EXT_SEL_DIV BIT(11)
+#define ETHER_CLK_SEL_RX_CLK_EXT_SEL_IN  BIT(0)
+#define ETHER_CLK_SEL_RX_CLK_EXT_SEL_RXC BIT(12)
+#define ETHER_CLK_SEL_RX_CLK_EXT_SEL_DIV BIT(13)
+#define ETHER_CLK_SEL_TX_CLK_O_TX_I	 BIT(0)
+#define ETHER_CLK_SEL_TX_CLK_O_RMII_I	 BIT(14)
+#define ETHER_CLK_SEL_TX_O_E_N_IN	 BIT(15)
+#define ETHER_CLK_SEL_RMII_CLK_SEL_IN	 BIT(0)
+#define ETHER_CLK_SEL_RMII_CLK_SEL_RX_C	 BIT(16)
+
+#define ETHER_CLK_SEL_RX_TX_CLK_EN (ETHER_CLK_SEL_RX_CLK_EN | ETHER_CLK_SEL_TX_CLK_EN)
+
+#define ETHER_CONFIG_INTF_MII 0
+#define ETHER_CONFIG_INTF_RGMII BIT(0)
+#define ETHER_CONFIG_INTF_RMII BIT(2)
+
+struct visconti_eth {
+	void __iomem *reg;
+	u32 phy_intf_sel;
+	struct clk *phy_ref_clk;
+	spinlock_t lock; /* lock to protect register update */
+};
+
+static void visconti_eth_fix_mac_speed(void *priv, unsigned int speed)
+{
+	struct visconti_eth *dwmac = priv;
+	unsigned int val, clk_sel_val;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dwmac->lock, flags);
+
+	/* adjust link */
+	val = readl(dwmac->reg + MAC_CTRL_REG);
+	val &= ~(GMAC_CONFIG_PS | GMAC_CONFIG_FES);
+
+	switch (speed) {
+	case SPEED_1000:
+		if (dwmac->phy_intf_sel == ETHER_CONFIG_INTF_RGMII)
+			clk_sel_val = ETHER_CLK_SEL_FREQ_SEL_125M;
+		break;
+	case SPEED_100:
+		if (dwmac->phy_intf_sel == ETHER_CONFIG_INTF_RGMII)
+			clk_sel_val = ETHER_CLK_SEL_FREQ_SEL_25M;
+		if (dwmac->phy_intf_sel == ETHER_CONFIG_INTF_RMII)
+			clk_sel_val = ETHER_CLK_SEL_DIV_SEL_2;
+		val |= GMAC_CONFIG_PS | GMAC_CONFIG_FES;
+		break;
+	case SPEED_10:
+		if (dwmac->phy_intf_sel == ETHER_CONFIG_INTF_RGMII)
+			clk_sel_val = ETHER_CLK_SEL_FREQ_SEL_2P5M;
+		if (dwmac->phy_intf_sel == ETHER_CONFIG_INTF_RMII)
+			clk_sel_val = ETHER_CLK_SEL_DIV_SEL_20;
+		val |= GMAC_CONFIG_PS;
+		break;
+	default:
+		/* No bit control */
+		break;
+	}
+
+	writel(val, dwmac->reg + MAC_CTRL_REG);
+
+	/* Stop internal clock */
+	val = readl(dwmac->reg + REG_ETHER_CLOCK_SEL);
+	val &= ~(ETHER_CLK_SEL_RMII_CLK_EN | ETHER_CLK_SEL_RX_TX_CLK_EN);
+	val |= ETHER_CLK_SEL_TX_O_E_N_IN;
+	writel(val, dwmac->reg + REG_ETHER_CLOCK_SEL);
+
+	switch (dwmac->phy_intf_sel) {
+	case ETHER_CONFIG_INTF_RGMII:
+		val = clk_sel_val | ETHER_CLK_SEL_RX_CLK_EXT_SEL_RXC;
+		break;
+	case ETHER_CONFIG_INTF_RMII:
+		val = clk_sel_val | ETHER_CLK_SEL_RX_CLK_EXT_SEL_DIV |
+			ETHER_CLK_SEL_TX_CLK_EXT_SEL_TXC | ETHER_CLK_SEL_TX_O_E_N_IN |
+			ETHER_CLK_SEL_RMII_CLK_SEL_RX_C;
+		break;
+	case ETHER_CONFIG_INTF_MII:
+	default:
+		val = clk_sel_val | ETHER_CLK_SEL_RX_CLK_EXT_SEL_RXC |
+			ETHER_CLK_SEL_TX_CLK_EXT_SEL_DIV | ETHER_CLK_SEL_TX_O_E_N_IN |
+			ETHER_CLK_SEL_RMII_CLK_EN;
+		break;
+	}
+
+	/* Start clock */
+	writel(val, dwmac->reg + REG_ETHER_CLOCK_SEL);
+	val |= ETHER_CLK_SEL_RX_TX_CLK_EN;
+	writel(val, dwmac->reg + REG_ETHER_CLOCK_SEL);
+
+	val &= ~ETHER_CLK_SEL_TX_O_E_N_IN;
+	writel(val, dwmac->reg + REG_ETHER_CLOCK_SEL);
+
+	spin_unlock_irqrestore(&dwmac->lock, flags);
+}
+
+static int visconti_eth_init_hw(struct platform_device *pdev, struct plat_stmmacenet_data *plat_dat)
+{
+	struct visconti_eth *dwmac = plat_dat->bsp_priv;
+	unsigned int reg_val, clk_sel_val;
+
+	switch (plat_dat->phy_interface) {
+	case PHY_INTERFACE_MODE_RGMII:
+	case PHY_INTERFACE_MODE_RGMII_ID:
+	case PHY_INTERFACE_MODE_RGMII_RXID:
+	case PHY_INTERFACE_MODE_RGMII_TXID:
+		dwmac->phy_intf_sel = ETHER_CONFIG_INTF_RGMII;
+		break;
+	case PHY_INTERFACE_MODE_MII:
+		dwmac->phy_intf_sel = ETHER_CONFIG_INTF_MII;
+		break;
+	case PHY_INTERFACE_MODE_RMII:
+		dwmac->phy_intf_sel = ETHER_CONFIG_INTF_RMII;
+		break;
+	default:
+		dev_err(&pdev->dev, "Unsupported phy-mode (%d)\n", plat_dat->phy_interface);
+		return -EOPNOTSUPP;
+	}
+
+	reg_val = dwmac->phy_intf_sel;
+	writel(reg_val, dwmac->reg + REG_ETHER_CONTROL);
+
+	/* Enable TX/RX clock */
+	clk_sel_val = ETHER_CLK_SEL_FREQ_SEL_125M;
+	writel(clk_sel_val, dwmac->reg + REG_ETHER_CLOCK_SEL);
+
+	writel((clk_sel_val | ETHER_CLK_SEL_RMII_CLK_EN | ETHER_CLK_SEL_RX_TX_CLK_EN),
+	       dwmac->reg + REG_ETHER_CLOCK_SEL);
+
+	/* release internal-reset */
+	reg_val |= ETHER_ETH_CONTROL_RESET;
+	writel(reg_val, dwmac->reg + REG_ETHER_CONTROL);
+
+	return 0;
+}
+
+static int visconti_eth_clock_probe(struct platform_device *pdev,
+				    struct plat_stmmacenet_data *plat_dat)
+{
+	struct visconti_eth *dwmac = plat_dat->bsp_priv;
+	int err;
+
+	dwmac->phy_ref_clk = devm_clk_get(&pdev->dev, "phy_ref_clk");
+	if (IS_ERR(dwmac->phy_ref_clk)) {
+		dev_err(&pdev->dev, "phy_ref_clk clock not found.\n");
+		return PTR_ERR(dwmac->phy_ref_clk);
+	}
+
+	err = clk_prepare_enable(dwmac->phy_ref_clk);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to enable phy_ref clock: %d\n", err);
+		return err;
+	}
+
+	return 0;
+}
+
+static int visconti_eth_clock_remove(struct platform_device *pdev)
+{
+	struct visconti_eth *dwmac = get_stmmac_bsp_priv(&pdev->dev);
+	struct net_device *ndev = platform_get_drvdata(pdev);
+	struct stmmac_priv *priv = netdev_priv(ndev);
+
+	clk_disable_unprepare(dwmac->phy_ref_clk);
+	clk_disable_unprepare(priv->plat->stmmac_clk);
+
+	return 0;
+}
+
+static int visconti_eth_dwmac_probe(struct platform_device *pdev)
+{
+	struct plat_stmmacenet_data *plat_dat;
+	struct stmmac_resources stmmac_res;
+	struct visconti_eth *dwmac;
+	int ret;
+
+	ret = stmmac_get_platform_resources(pdev, &stmmac_res);
+	if (ret)
+		return ret;
+
+	plat_dat = stmmac_probe_config_dt(pdev, &stmmac_res.mac);
+	if (IS_ERR(plat_dat))
+		return PTR_ERR(plat_dat);
+
+	dwmac = devm_kzalloc(&pdev->dev, sizeof(*dwmac), GFP_KERNEL);
+	if (!dwmac) {
+		ret = -ENOMEM;
+		goto remove_config;
+	}
+
+	dwmac->reg = stmmac_res.addr;
+	plat_dat->bsp_priv = dwmac;
+	plat_dat->fix_mac_speed = visconti_eth_fix_mac_speed;
+
+	ret = visconti_eth_clock_probe(pdev, plat_dat);
+	if (ret)
+		goto remove_config;
+
+	visconti_eth_init_hw(pdev, plat_dat);
+
+	plat_dat->dma_cfg->aal = 1;
+
+	ret = stmmac_dvr_probe(&pdev->dev, plat_dat, &stmmac_res);
+	if (ret)
+		goto remove;
+
+	return ret;
+
+remove:
+	visconti_eth_clock_remove(pdev);
+remove_config:
+	stmmac_remove_config_dt(pdev, plat_dat);
+
+	return ret;
+}
+
+static int visconti_eth_dwmac_remove(struct platform_device *pdev)
+{
+	struct net_device *ndev = platform_get_drvdata(pdev);
+	struct stmmac_priv *priv = netdev_priv(ndev);
+	int err;
+
+	err = stmmac_pltfr_remove(pdev);
+	if (err < 0)
+		dev_err(&pdev->dev, "failed to remove platform: %d\n", err);
+
+	err = visconti_eth_clock_remove(pdev);
+	if (err < 0)
+		dev_err(&pdev->dev, "failed to remove clock: %d\n", err);
+
+	stmmac_remove_config_dt(pdev, priv->plat);
+
+	return err;
+}
+
+static const struct of_device_id visconti_eth_dwmac_match[] = {
+	{ .compatible = "toshiba,visconti-dwmac" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, visconti_eth_dwmac_match);
+
+static struct platform_driver visconti_eth_dwmac_driver = {
+	.probe  = visconti_eth_dwmac_probe,
+	.remove = visconti_eth_dwmac_remove,
+	.driver = {
+		.name           = "visconti-eth-dwmac",
+		.of_match_table = visconti_eth_dwmac_match,
+	},
+};
+module_platform_driver(visconti_eth_dwmac_driver);
+
+MODULE_AUTHOR("Toshiba");
+MODULE_DESCRIPTION("Toshiba Visconti Ethernet DWMAC glue driver");
+MODULE_AUTHOR("Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp");
+MODULE_LICENSE("GPL v2");
-- 
2.30.0.rc2

Re: [PATCH 2/4] net: stmmac: Add Toshiba Visconti SoCs glue driver

[Leon Romanovsky]

On Mon, Feb 15, 2021 at 02:06:53PM +0900, Nobuhiro Iwamatsu wrote:
> Add dwmac-visconti to the stmmac driver in Toshiba Visconti ARM SoCs.
> This patch contains only the basic function of the device. There is no
> clock control, PM, etc. yet. These will be added in the future.
>
> Signed-off-by: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp>
> ---
>  drivers/net/ethernet/stmicro/stmmac/Kconfig   |   8 +
>  drivers/net/ethernet/stmicro/stmmac/Makefile  |   1 +
>  .../ethernet/stmicro/stmmac/dwmac-visconti.c  | 285 ++++++++++++++++++
>  3 files changed, 294 insertions(+)
>  create mode 100644 drivers/net/ethernet/stmicro/stmmac/dwmac-visconti.c
>
> diff --git a/drivers/net/ethernet/stmicro/stmmac/Kconfig b/drivers/net/ethernet/stmicro/stmmac/Kconfig
> index 53f14c5a9e02..55ba67a550b9 100644
> --- a/drivers/net/ethernet/stmicro/stmmac/Kconfig
> +++ b/drivers/net/ethernet/stmicro/stmmac/Kconfig
> @@ -219,6 +219,14 @@ config DWMAC_INTEL_PLAT
>  	  This selects the Intel platform specific glue layer support for
>  	  the stmmac device driver. This driver is used for the Intel Keem Bay
>  	  SoC.
> +
> +config DWMAC_VISCONTI
> +	bool "Toshiba Visconti DWMAC support"
> +	def_bool y

I asked it before, but never received an answer.
Why did you use "def_bool y" and not "default y"? Isn't it supposed to be
"depends on STMMAC_ETH"? And probably it shouldn't be set as a default as "y".

Thanks

Re: [PATCH 2/4] net: stmmac: Add Toshiba Visconti SoCs glue driver

[Nobuhiro Iwamatsu]

Hi,

Thanks for your review.

On Mon, Feb 15, 2021 at 08:07:21AM +0200, Leon Romanovsky wrote:
> On Mon, Feb 15, 2021 at 02:06:53PM +0900, Nobuhiro Iwamatsu wrote:
> > Add dwmac-visconti to the stmmac driver in Toshiba Visconti ARM SoCs.
> > This patch contains only the basic function of the device. There is no
> > clock control, PM, etc. yet. These will be added in the future.
> >
> > Signed-off-by: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp>
> > ---
> >  drivers/net/ethernet/stmicro/stmmac/Kconfig   |   8 +
> >  drivers/net/ethernet/stmicro/stmmac/Makefile  |   1 +
> >  .../ethernet/stmicro/stmmac/dwmac-visconti.c  | 285 ++++++++++++++++++
> >  3 files changed, 294 insertions(+)
> >  create mode 100644 drivers/net/ethernet/stmicro/stmmac/dwmac-visconti.c
> >
> > diff --git a/drivers/net/ethernet/stmicro/stmmac/Kconfig b/drivers/net/ethernet/stmicro/stmmac/Kconfig
> > index 53f14c5a9e02..55ba67a550b9 100644
> > --- a/drivers/net/ethernet/stmicro/stmmac/Kconfig
> > +++ b/drivers/net/ethernet/stmicro/stmmac/Kconfig
> > @@ -219,6 +219,14 @@ config DWMAC_INTEL_PLAT
> >  	  This selects the Intel platform specific glue layer support for
> >  	  the stmmac device driver. This driver is used for the Intel Keem Bay
> >  	  SoC.
> > +
> > +config DWMAC_VISCONTI
> > +	bool "Toshiba Visconti DWMAC support"
> > +	def_bool y
> 

Sorry, I sent the wrong patchset that didn't fix this point out.

> I asked it before, but never received an answer.

I have received your point out and have sent an email with the content
to remove this line. But it may not have arrived yet...

> Why did you use "def_bool y" and not "default y"? Isn't it supposed to be
> "depends on STMMAC_ETH"? And probably it shouldn't be set as a default as "y".
> 

The reason why "def_bool y" was set is that the wrong fix was left when
debugging. Also, I don't think it is necessary to set "default y".
This is also incorrect because it says "bool" Toshiba Visconti DWMAC
support "". I change it to trustate in the new patch.

And this driver is enabled when STMMAC_PLATFORM was Y. And STMMAC_PLATFORM
depends on STMMAC_ETH.
So I understand that STMMAC_ETH does not need to be dependents. Is this
understanding wrong?

> Thanks
> 

Best regards,
  Nobuhiro

Re: [PATCH 2/4] net: stmmac: Add Toshiba Visconti SoCs glue driver

[Leon Romanovsky]

On Mon, Feb 15, 2021 at 04:28:09PM +0900, Nobuhiro Iwamatsu wrote:
> Hi,
>
> Thanks for your review.
>
> On Mon, Feb 15, 2021 at 08:07:21AM +0200, Leon Romanovsky wrote:
> > On Mon, Feb 15, 2021 at 02:06:53PM +0900, Nobuhiro Iwamatsu wrote:
> > > Add dwmac-visconti to the stmmac driver in Toshiba Visconti ARM SoCs.
> > > This patch contains only the basic function of the device. There is no
> > > clock control, PM, etc. yet. These will be added in the future.
> > >
> > > Signed-off-by: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp>
> > > ---
> > >  drivers/net/ethernet/stmicro/stmmac/Kconfig   |   8 +
> > >  drivers/net/ethernet/stmicro/stmmac/Makefile  |   1 +
> > >  .../ethernet/stmicro/stmmac/dwmac-visconti.c  | 285 ++++++++++++++++++
> > >  3 files changed, 294 insertions(+)
> > >  create mode 100644 drivers/net/ethernet/stmicro/stmmac/dwmac-visconti.c
> > >
> > > diff --git a/drivers/net/ethernet/stmicro/stmmac/Kconfig b/drivers/net/ethernet/stmicro/stmmac/Kconfig
> > > index 53f14c5a9e02..55ba67a550b9 100644
> > > --- a/drivers/net/ethernet/stmicro/stmmac/Kconfig
> > > +++ b/drivers/net/ethernet/stmicro/stmmac/Kconfig
> > > @@ -219,6 +219,14 @@ config DWMAC_INTEL_PLAT
> > >  	  This selects the Intel platform specific glue layer support for
> > >  	  the stmmac device driver. This driver is used for the Intel Keem Bay
> > >  	  SoC.
> > > +
> > > +config DWMAC_VISCONTI
> > > +	bool "Toshiba Visconti DWMAC support"
> > > +	def_bool y
> >
>
> Sorry, I sent the wrong patchset that didn't fix this point out.
>
> > I asked it before, but never received an answer.
>
> I have received your point out and have sent an email with the content
> to remove this line. But it may not have arrived yet...
>
> > Why did you use "def_bool y" and not "default y"? Isn't it supposed to be
> > "depends on STMMAC_ETH"? And probably it shouldn't be set as a default as "y".
> >
>
> The reason why "def_bool y" was set is that the wrong fix was left when
> debugging. Also, I don't think it is necessary to set "default y".
> This is also incorrect because it says "bool" Toshiba Visconti DWMAC
> support "". I change it to trustate in the new patch.
>
> And this driver is enabled when STMMAC_PLATFORM was Y. And STMMAC_PLATFORM
> depends on STMMAC_ETH.
> So I understand that STMMAC_ETH does not need to be dependents. Is this
> understanding wrong?

This is correct understanding, just need to clean other entries in that
Kconfig that depends on STMMAC_ETH.

Thanks

>
> > Thanks
> >
>
> Best regards,
>   Nobuhiro

Re: [PATCH 2/4] net: stmmac: Add Toshiba Visconti SoCs glue driver

[Arnd Bergmann]

On Mon, Feb 15, 2021 at 10:23 AM Leon Romanovsky <leon@kernel.org> wrote:
> On Mon, Feb 15, 2021 at 04:28:09PM +0900, Nobuhiro Iwamatsu wrote:
> >
> > Sorry, I sent the wrong patchset that didn't fix this point out.
> >
> > > I asked it before, but never received an answer.
> >
> > I have received your point out and have sent an email with the content
> > to remove this line. But it may not have arrived yet...
> >
> > > Why did you use "def_bool y" and not "default y"? Isn't it supposed to be
> > > "depends on STMMAC_ETH"? And probably it shouldn't be set as a default as "y".
> > >
> >
> > The reason why "def_bool y" was set is that the wrong fix was left when
> > debugging. Also, I don't think it is necessary to set "default y".
> > This is also incorrect because it says "bool" Toshiba Visconti DWMAC
> > support "". I change it to trustate in the new patch.
> >
> > And this driver is enabled when STMMAC_PLATFORM was Y. And STMMAC_PLATFORM
> > depends on STMMAC_ETH.
> > So I understand that STMMAC_ETH does not need to be dependents. Is this
> > understanding wrong?
>
> This is correct understanding, just need to clean other entries in that
> Kconfig that depends on STMMAC_ETH.

'tristate' with no default sounds right. I see that some platforms have a
default according to the platform, which also makes sense but isn't
required. What I would suggest though is a dependency on the platform,
to make it easier to disable the front-end based on which platforms
are enabled. This would end up as

config DWMAC_VISCONTI
        tristate "Toshiba Visconti DWMAC support"
        depends on ARCH_VISCONTI || COMPILE_TEST
        depends on OF && COMMON_CLK # only add this line if it's
required for compilation
        default ARCH_VISCONTI

      Arnd

Re: [PATCH 2/4] net: stmmac: Add Toshiba Visconti SoCs glue driver

[Nobuhiro Iwamatsu]

Hi,

On Mon, Feb 15, 2021 at 01:19:18PM +0100, Arnd Bergmann wrote:
> On Mon, Feb 15, 2021 at 10:23 AM Leon Romanovsky <leon@kernel.org> wrote:
> > On Mon, Feb 15, 2021 at 04:28:09PM +0900, Nobuhiro Iwamatsu wrote:
> > >
> > > Sorry, I sent the wrong patchset that didn't fix this point out.
> > >
> > > > I asked it before, but never received an answer.
> > >
> > > I have received your point out and have sent an email with the content
> > > to remove this line. But it may not have arrived yet...
> > >
> > > > Why did you use "def_bool y" and not "default y"? Isn't it supposed to be
> > > > "depends on STMMAC_ETH"? And probably it shouldn't be set as a default as "y".
> > > >
> > >
> > > The reason why "def_bool y" was set is that the wrong fix was left when
> > > debugging. Also, I don't think it is necessary to set "default y".
> > > This is also incorrect because it says "bool" Toshiba Visconti DWMAC
> > > support "". I change it to trustate in the new patch.
> > >
> > > And this driver is enabled when STMMAC_PLATFORM was Y. And STMMAC_PLATFORM
> > > depends on STMMAC_ETH.
> > > So I understand that STMMAC_ETH does not need to be dependents. Is this
> > > understanding wrong?
> >
> > This is correct understanding, just need to clean other entries in that
> > Kconfig that depends on STMMAC_ETH.
> 
> 'tristate' with no default sounds right. I see that some platforms have a
> default according to the platform, which also makes sense but isn't
> required. What I would suggest though is a dependency on the platform,
> to make it easier to disable the front-end based on which platforms
> are enabled. This would end up as
> 
> config DWMAC_VISCONTI
>         tristate "Toshiba Visconti DWMAC support"
>         depends on ARCH_VISCONTI || COMPILE_TEST
>         depends on OF && COMMON_CLK # only add this line if it's
> required for compilation
>         default ARCH_VISCONTI
>

The fix at hand is the same as your suggestion.
Thank you for your comment.

>       Arnd
> 

Best regards,
  Nobuhiro

Re: [PATCH 2/4] net: stmmac: Add Toshiba Visconti SoCs glue driver

[Nobuhiro Iwamatsu]

Hi,

On Mon, Feb 15, 2021 at 11:22:33AM +0200, Leon Romanovsky wrote:
> On Mon, Feb 15, 2021 at 04:28:09PM +0900, Nobuhiro Iwamatsu wrote:
> >
> > I have received your point out and have sent an email with the content
> > to remove this line. But it may not have arrived yet...
> >
> > > Why did you use "def_bool y" and not "default y"? Isn't it supposed to be
> > > "depends on STMMAC_ETH"? And probably it shouldn't be set as a default as "y".
> > >
> >
> > The reason why "def_bool y" was set is that the wrong fix was left when
> > debugging. Also, I don't think it is necessary to set "default y".
> > This is also incorrect because it says "bool" Toshiba Visconti DWMAC
> > support "". I change it to trustate in the new patch.
> >
> > And this driver is enabled when STMMAC_PLATFORM was Y. And STMMAC_PLATFORM
> > depends on STMMAC_ETH.
> > So I understand that STMMAC_ETH does not need to be dependents. Is this
> > understanding wrong?
> 
> This is correct understanding, just need to clean other entries in that
> Kconfig that depends on STMMAC_ETH.

OK, thanks.

Best regards,
  Nobuhiro

[PATCH 3/4] MAINTAINERS: Add entries for Toshiba Visconti ethernet controller

[Nobuhiro Iwamatsu]

Add entries for Toshiba Visconti ethernet controller binding and driver.

Signed-off-by: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp>
---
 MAINTAINERS | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/MAINTAINERS b/MAINTAINERS
index cbf4b94f89d4..6be4bdaabf32 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -2641,8 +2641,10 @@ L:	linux-arm-kernel@lists.infradead.org (moderated for non-subscribers)
 S:	Supported
 T:	git git://git.kernel.org/pub/scm/linux/kernel/git/iwamatsu/linux-visconti.git
 F:	Documentation/devicetree/bindings/arm/toshiba.yaml
+F:	Documentation/devicetree/bindings/net/toshiba,visconti-dwmac.yaml
 F:	Documentation/devicetree/bindings/pinctrl/toshiba,tmpv7700-pinctrl.yaml
 F:	arch/arm64/boot/dts/toshiba/
+F:	drivers/net/ethernet/stmicro/stmmac/dwmac-visconti.c
 F:	drivers/pinctrl/visconti/
 N:	visconti
 
-- 
2.30.0.rc2

[PATCH 4/4] arm: dts: visconti: Add DT support for Toshiba Visconti5 ethernet controller

[Nobuhiro Iwamatsu]

Add the ethernet controller node in Toshiba Visconti5 SoC-specific DT file.
And enable this node in TMPV7708 RM main board's board-specific DT file.

Signed-off-by: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp>
---
 .../boot/dts/toshiba/tmpv7708-rm-mbrc.dts     | 18 +++++++++++++
 arch/arm64/boot/dts/toshiba/tmpv7708.dtsi     | 25 +++++++++++++++++++
 2 files changed, 43 insertions(+)

diff --git a/arch/arm64/boot/dts/toshiba/tmpv7708-rm-mbrc.dts b/arch/arm64/boot/dts/toshiba/tmpv7708-rm-mbrc.dts
index ed0bf7f13f54..48fa8776e36f 100644
--- a/arch/arm64/boot/dts/toshiba/tmpv7708-rm-mbrc.dts
+++ b/arch/arm64/boot/dts/toshiba/tmpv7708-rm-mbrc.dts
@@ -41,3 +41,21 @@ &uart1 {
 	clocks = <&uart_clk>;
 	clock-names = "apb_pclk";
 };
+
+&piether {
+	status = "okay";
+	phy-handle = <&phy0>;
+	phy-mode = "rgmii-id";
+	clocks = <&clk300mhz>, <&clk125mhz>;
+	clock-names = "stmmaceth", "phy_ref_clk";
+
+	mdio0 {
+		#address-cells = <1>;
+		#size-cells = <0>;
+		compatible = "snps,dwmac-mdio";
+		phy0: ethernet-phy@1 {
+			device_type = "ethernet-phy";
+			reg = <0x1>;
+		};
+	};
+};
diff --git a/arch/arm64/boot/dts/toshiba/tmpv7708.dtsi b/arch/arm64/boot/dts/toshiba/tmpv7708.dtsi
index 242f25f4e12a..3366786699fc 100644
--- a/arch/arm64/boot/dts/toshiba/tmpv7708.dtsi
+++ b/arch/arm64/boot/dts/toshiba/tmpv7708.dtsi
@@ -134,6 +134,20 @@ uart_clk: uart-clk {
 		#clock-cells = <0>;
 	};
 
+	clk125mhz: clk125mhz {
+		compatible = "fixed-clock";
+		clock-frequency = <125000000>;
+		#clock-cells = <0>;
+		clock-output-names = "clk125mhz";
+	};
+
+	clk300mhz: clk300mhz {
+		compatible = "fixed-clock";
+		clock-frequency = <300000000>;
+		#clock-cells = <0>;
+		clock-output-names = "clk300mhz";
+	};
+
 	soc {
 		#address-cells = <2>;
 		#size-cells = <2>;
@@ -384,6 +398,17 @@ spi6: spi@28146000 {
 			#size-cells = <0>;
 			status = "disabled";
 		};
+
+		piether: ethernet@28000000 {
+			compatible = "toshiba,visconti-dwmac", "snps,dwmac-4.20a";
+			reg = <0 0x28000000 0 0x10000>;
+			interrupts = <GIC_SPI 156 IRQ_TYPE_LEVEL_HIGH>;
+			interrupt-names = "macirq";
+			snps,txpbl = <4>;
+			snps,rxpbl = <4>;
+			snps,tso;
+			status = "disabled";
+		};
 	};
 };
 
-- 
2.30.0.rc2

[PATCH 0/4] net: stmmac: Add Toshiba Visconti SoCs glue driver

[Nobuhiro Iwamatsu]

Hi,

This series is the ethernet driver for Toshiba's ARM SoC, Visconti[0].
This provides DT binding documentation, device driver, MAINTAINER files, and updates to DT files.

Best regards,
  Nobuhiro

[0]: https://toshiba.semicon-storage.com/ap-en/semiconductor/product/image-recognition-processors-visconti.htmli

Nobuhiro Iwamatsu (4):
  dt-bindings: net: Add DT bindings for Toshiba Visconti TMPV7700 SoC
  net: stmmac: Add Toshiba Visconti SoCs glue driver
  MAINTAINERS: Add entries for Toshiba Visconti ethernet controller
  arm: dts: visconti: Add DT support for Toshiba Visconti5 ethernet
    controller

 .../bindings/net/toshiba,visconti-dwmac.yaml  |  87 ++++++
 MAINTAINERS                                   |   2 +
 .../boot/dts/toshiba/tmpv7708-rm-mbrc.dts     |  18 ++
 arch/arm64/boot/dts/toshiba/tmpv7708.dtsi     |  24 ++
 drivers/net/ethernet/stmicro/stmmac/Kconfig   |   8 +
 drivers/net/ethernet/stmicro/stmmac/Makefile  |   1 +
 .../ethernet/stmicro/stmmac/dwmac-visconti.c  | 292 ++++++++++++++++++
 7 files changed, 432 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/net/toshiba,visconti-dwmac.yaml
 create mode 100644 drivers/net/ethernet/stmicro/stmmac/dwmac-visconti.c

-- 
2.27.0

[PATCH 2/4] net: stmmac: Add Toshiba Visconti SoCs glue driver

[Nobuhiro Iwamatsu]

Add dwmac-visconti to the stmmac driver in Toshiba Visconti ARM SoCs.
This patch contains only the basic function of the device. There is no
clock control, PM, etc. yet. These will be added in the future.

Signed-off-by: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp>
---
 drivers/net/ethernet/stmicro/stmmac/Kconfig   |   8 +
 drivers/net/ethernet/stmicro/stmmac/Makefile  |   1 +
 .../ethernet/stmicro/stmmac/dwmac-visconti.c  | 292 ++++++++++++++++++
 3 files changed, 301 insertions(+)
 create mode 100644 drivers/net/ethernet/stmicro/stmmac/dwmac-visconti.c

diff --git a/drivers/net/ethernet/stmicro/stmmac/Kconfig b/drivers/net/ethernet/stmicro/stmmac/Kconfig
index 53f14c5a9e02..55ba67a550b9 100644
--- a/drivers/net/ethernet/stmicro/stmmac/Kconfig
+++ b/drivers/net/ethernet/stmicro/stmmac/Kconfig
@@ -219,6 +219,14 @@ config DWMAC_INTEL_PLAT
 	  This selects the Intel platform specific glue layer support for
 	  the stmmac device driver. This driver is used for the Intel Keem Bay
 	  SoC.
+
+config DWMAC_VISCONTI
+	bool "Toshiba Visconti DWMAC support"
+	def_bool y
+	depends on OF && COMMON_CLK && (ARCH_VISCONTI || COMPILE_TEST)
+	help
+	  Support for ethernet controller on Visconti SoCs.
+
 endif
 
 config DWMAC_INTEL
diff --git a/drivers/net/ethernet/stmicro/stmmac/Makefile b/drivers/net/ethernet/stmicro/stmmac/Makefile
index 24e6145d4eae..366740ab9c5a 100644
--- a/drivers/net/ethernet/stmicro/stmmac/Makefile
+++ b/drivers/net/ethernet/stmicro/stmmac/Makefile
@@ -29,6 +29,7 @@ obj-$(CONFIG_DWMAC_DWC_QOS_ETH)	+= dwmac-dwc-qos-eth.o
 obj-$(CONFIG_DWMAC_INTEL_PLAT)	+= dwmac-intel-plat.o
 obj-$(CONFIG_DWMAC_GENERIC)	+= dwmac-generic.o
 obj-$(CONFIG_DWMAC_IMX8)	+= dwmac-imx.o
+obj-$(CONFIG_DWMAC_VISCONTI)	+= dwmac-visconti.o
 stmmac-platform-objs:= stmmac_platform.o
 dwmac-altr-socfpga-objs := altr_tse_pcs.o dwmac-socfpga.o
 
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-visconti.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-visconti.c
new file mode 100644
index 000000000000..604b0a513fbb
--- /dev/null
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-visconti.c
@@ -0,0 +1,292 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Toshiba Visconti Ethernet Support
+ *
+ * (C) Copyright 2020 TOSHIBA CORPORATION
+ * (C) Copyright 2020 Toshiba Electronic Devices & Storage Corporation
+ */
+
+#include <linux/module.h>
+#include <linux/of_device.h>
+#include <linux/of_net.h>
+#include <linux/stmmac.h>
+
+#include "stmmac_platform.h"
+#include "dwmac4.h"
+
+#define REG_ETHER_CONTROL	0x52D4
+#define ETHER_ETH_CONTROL_RESET BIT(17)
+
+#define REG_ETHER_CLOCK_SEL	0x52D0
+#define ETHER_CLK_SEL_TX_CLK_EN BIT(0)
+#define ETHER_CLK_SEL_RX_CLK_EN BIT(1)
+#define ETHER_CLK_SEL_RMII_CLK_EN BIT(2)
+#define ETHER_CLK_SEL_RMII_CLK_RST BIT(3)
+#define ETHER_CLK_SEL_DIV_SEL_2 BIT(4)
+#define ETHER_CLK_SEL_DIV_SEL_20 BIT(0)
+#define ETHER_CLK_SEL_FREQ_SEL_125M	(BIT(9) | BIT(8))
+#define ETHER_CLK_SEL_FREQ_SEL_50M	BIT(9)
+#define ETHER_CLK_SEL_FREQ_SEL_25M	BIT(8)
+#define ETHER_CLK_SEL_FREQ_SEL_2P5M	BIT(0)
+#define ETHER_CLK_SEL_TX_CLK_EXT_SEL_IN BIT(0)
+#define ETHER_CLK_SEL_TX_CLK_EXT_SEL_TXC BIT(10)
+#define ETHER_CLK_SEL_TX_CLK_EXT_SEL_DIV BIT(11)
+#define ETHER_CLK_SEL_RX_CLK_EXT_SEL_IN  BIT(0)
+#define ETHER_CLK_SEL_RX_CLK_EXT_SEL_RXC BIT(12)
+#define ETHER_CLK_SEL_RX_CLK_EXT_SEL_DIV BIT(13)
+#define ETHER_CLK_SEL_TX_CLK_O_TX_I	 BIT(0)
+#define ETHER_CLK_SEL_TX_CLK_O_RMII_I	 BIT(14)
+#define ETHER_CLK_SEL_TX_O_E_N_IN	 BIT(15)
+#define ETHER_CLK_SEL_RMII_CLK_SEL_IN	 BIT(0)
+#define ETHER_CLK_SEL_RMII_CLK_SEL_RX_C	 BIT(16)
+
+#define ETHER_CLK_SEL_RX_TX_CLK_EN (ETHER_CLK_SEL_RX_CLK_EN | ETHER_CLK_SEL_TX_CLK_EN)
+
+#define ETHER_CONFIG_INTF_MII 0
+#define ETHER_CONFIG_INTF_RGMII BIT(0)
+#define ETHER_CONFIG_INTF_RMII BIT(2)
+
+struct visconti_eth {
+	void __iomem *reg;
+	u32 phy_intf_sel;
+	struct clk *phy_ref_clk;
+	spinlock_t lock; /* lock to protect register update */
+};
+
+static void visconti_eth_fix_mac_speed(void *priv, unsigned int speed)
+{
+	struct visconti_eth *dwmac = priv;
+	unsigned int val, clk_sel_val;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dwmac->lock, flags);
+
+	/* adjust link */
+	val = readl(dwmac->reg + MAC_CTRL_REG);
+	val &= ~(GMAC_CONFIG_PS | GMAC_CONFIG_FES);
+
+	switch (speed) {
+	case SPEED_1000:
+		if (dwmac->phy_intf_sel == ETHER_CONFIG_INTF_RGMII)
+			clk_sel_val = ETHER_CLK_SEL_FREQ_SEL_125M;
+		break;
+	case SPEED_100:
+		if (dwmac->phy_intf_sel == ETHER_CONFIG_INTF_RGMII)
+			clk_sel_val = ETHER_CLK_SEL_FREQ_SEL_25M;
+		if (dwmac->phy_intf_sel == ETHER_CONFIG_INTF_RMII)
+			clk_sel_val = ETHER_CLK_SEL_DIV_SEL_2;
+		val |= GMAC_CONFIG_PS | GMAC_CONFIG_FES;
+		break;
+	case SPEED_10:
+		if (dwmac->phy_intf_sel == ETHER_CONFIG_INTF_RGMII)
+			clk_sel_val = ETHER_CLK_SEL_FREQ_SEL_2P5M;
+		if (dwmac->phy_intf_sel == ETHER_CONFIG_INTF_RMII)
+			clk_sel_val = ETHER_CLK_SEL_DIV_SEL_20;
+		val |= GMAC_CONFIG_PS;
+		break;
+	default:
+		/* No bit control */
+		break;
+	}
+
+	writel(val, dwmac->reg + MAC_CTRL_REG);
+
+	/* Stop internal clock */
+	val = readl(dwmac->reg + REG_ETHER_CLOCK_SEL);
+	val &= ~(ETHER_CLK_SEL_RMII_CLK_EN | ETHER_CLK_SEL_RX_TX_CLK_EN);
+	val |= ETHER_CLK_SEL_TX_O_E_N_IN;
+	writel(val, dwmac->reg + REG_ETHER_CLOCK_SEL);
+
+	switch (dwmac->phy_intf_sel) {
+	case ETHER_CONFIG_INTF_RGMII:
+		val = clk_sel_val | ETHER_CLK_SEL_RX_CLK_EXT_SEL_RXC;
+		break;
+	case ETHER_CONFIG_INTF_RMII:
+		val = clk_sel_val | ETHER_CLK_SEL_RX_CLK_EXT_SEL_DIV |
+			ETHER_CLK_SEL_TX_CLK_EXT_SEL_TXC | ETHER_CLK_SEL_TX_O_E_N_IN |
+			ETHER_CLK_SEL_RMII_CLK_SEL_RX_C;
+		break;
+	case ETHER_CONFIG_INTF_MII:
+	default:
+		val = clk_sel_val | ETHER_CLK_SEL_RX_CLK_EXT_SEL_RXC |
+			ETHER_CLK_SEL_TX_CLK_EXT_SEL_DIV | ETHER_CLK_SEL_TX_O_E_N_IN |
+			ETHER_CLK_SEL_RMII_CLK_EN;
+		break;
+	}
+
+	/* Start clock */
+	writel(val, dwmac->reg + REG_ETHER_CLOCK_SEL);
+	val |= ETHER_CLK_SEL_RX_TX_CLK_EN;
+	writel(val, dwmac->reg + REG_ETHER_CLOCK_SEL);
+
+	val &= ~ETHER_CLK_SEL_TX_O_E_N_IN;
+	writel(val, dwmac->reg + REG_ETHER_CLOCK_SEL);
+
+	spin_unlock_irqrestore(&dwmac->lock, flags);
+}
+
+static int visconti_eth_init_hw(struct platform_device *pdev, struct plat_stmmacenet_data *plat_dat)
+{
+	struct visconti_eth *dwmac;
+	unsigned int reg_val, clk_sel_val;
+
+	dwmac = plat_dat->bsp_priv;
+
+	switch (plat_dat->phy_interface) {
+	case PHY_INTERFACE_MODE_RGMII:
+	case PHY_INTERFACE_MODE_RGMII_ID:
+	case PHY_INTERFACE_MODE_RGMII_RXID:
+	case PHY_INTERFACE_MODE_RGMII_TXID:
+		dwmac->phy_intf_sel = ETHER_CONFIG_INTF_RGMII;
+		break;
+	case PHY_INTERFACE_MODE_MII:
+		dwmac->phy_intf_sel = ETHER_CONFIG_INTF_MII;
+		break;
+	case PHY_INTERFACE_MODE_RMII:
+		dwmac->phy_intf_sel = ETHER_CONFIG_INTF_RMII;
+		break;
+	default:
+		dev_err(&pdev->dev, "Unsupported phy-mode (%d)\n", plat_dat->phy_interface);
+		return -EOPNOTSUPP;
+	}
+
+	reg_val = dwmac->phy_intf_sel;
+	writel(reg_val, dwmac->reg + REG_ETHER_CONTROL);
+
+	/* Enable TX/RX clock */
+	clk_sel_val = ETHER_CLK_SEL_FREQ_SEL_125M;
+	writel(clk_sel_val, dwmac->reg + REG_ETHER_CLOCK_SEL);
+
+	writel((clk_sel_val | ETHER_CLK_SEL_RMII_CLK_EN | ETHER_CLK_SEL_RX_TX_CLK_EN),
+	       dwmac->reg + REG_ETHER_CLOCK_SEL);
+
+	/* release internal-reset */
+	reg_val |= ETHER_ETH_CONTROL_RESET;
+	writel(reg_val, dwmac->reg + REG_ETHER_CONTROL);
+
+	return 0;
+}
+
+static int visconti_eth_clock_probe(struct platform_device *pdev,
+				    struct plat_stmmacenet_data *plat_dat)
+{
+	int err;
+	struct visconti_eth *dwmac;
+
+	dwmac = plat_dat->bsp_priv;
+
+	dwmac->phy_ref_clk = devm_clk_get(&pdev->dev, "phy_ref_clk");
+	if (IS_ERR(dwmac->phy_ref_clk)) {
+		dev_err(&pdev->dev, "phy_ref_clk clock not found.\n");
+		return PTR_ERR(dwmac->phy_ref_clk);
+	}
+
+	err = clk_prepare_enable(dwmac->phy_ref_clk);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to enable phy_ref clock: %d\n", err);
+		return err;
+	}
+
+	return 0;
+}
+
+static int visconti_eth_clock_remove(struct platform_device *pdev)
+{
+	struct net_device *ndev = platform_get_drvdata(pdev);
+	struct stmmac_priv *priv = netdev_priv(ndev);
+	struct visconti_eth *dwmac = get_stmmac_bsp_priv(&pdev->dev);
+
+	clk_disable_unprepare(dwmac->phy_ref_clk);
+	clk_disable_unprepare(priv->plat->stmmac_clk);
+
+	return 0;
+}
+
+static int visconti_eth_dwmac_probe(struct platform_device *pdev)
+{
+	struct plat_stmmacenet_data *plat_dat;
+	struct stmmac_resources stmmac_res;
+	struct visconti_eth *dwmac;
+	int ret;
+
+	ret = stmmac_get_platform_resources(pdev, &stmmac_res);
+	if (ret)
+		return ret;
+
+	plat_dat = stmmac_probe_config_dt(pdev, &stmmac_res.mac);
+	if (IS_ERR(plat_dat))
+		return PTR_ERR(plat_dat);
+
+	dwmac = devm_kzalloc(&pdev->dev, sizeof(*dwmac), GFP_KERNEL);
+	if (!dwmac) {
+		ret = -ENOMEM;
+		goto remove_config;
+	}
+
+	dwmac->reg = stmmac_res.addr;
+	plat_dat->bsp_priv = dwmac;
+	plat_dat->fix_mac_speed = visconti_eth_fix_mac_speed;
+
+	ret = visconti_eth_clock_probe(pdev, plat_dat);
+	if (ret)
+		goto remove_config;
+
+	visconti_eth_init_hw(pdev, plat_dat);
+
+	plat_dat->has_gmac4 = 1;
+	plat_dat->dma_cfg->aal = 1;
+	plat_dat->tso_en = 1;
+	plat_dat->pmt = 1;
+
+	ret = stmmac_dvr_probe(&pdev->dev, plat_dat, &stmmac_res);
+	if (ret)
+		goto remove;
+
+	return ret;
+
+remove:
+	visconti_eth_clock_remove(pdev);
+remove_config:
+	stmmac_remove_config_dt(pdev, plat_dat);
+
+	return ret;
+}
+
+static int visconti_eth_dwmac_remove(struct platform_device *pdev)
+{
+	struct net_device *ndev = platform_get_drvdata(pdev);
+	struct stmmac_priv *priv = netdev_priv(ndev);
+	int err;
+
+	err = stmmac_pltfr_remove(pdev);
+	if (err < 0)
+		dev_err(&pdev->dev, "failed to remove platform: %d\n", err);
+
+	err = visconti_eth_clock_remove(pdev);
+	if (err < 0)
+		dev_err(&pdev->dev, "failed to remove clock: %d\n", err);
+
+	stmmac_remove_config_dt(pdev, priv->plat);
+
+	return err;
+}
+
+static const struct of_device_id visconti_eth_dwmac_match[] = {
+	{ .compatible = "toshiba,visconti-dwmac" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, visconti_eth_dwmac_match);
+
+static struct platform_driver visconti_eth_dwmac_driver = {
+	.probe  = visconti_eth_dwmac_probe,
+	.remove = visconti_eth_dwmac_remove,
+	.driver = {
+		.name           = "visconti-eth-dwmac",
+		.of_match_table = visconti_eth_dwmac_match,
+	},
+};
+module_platform_driver(visconti_eth_dwmac_driver);
+
+MODULE_AUTHOR("Toshiba");
+MODULE_DESCRIPTION("Toshiba Visconti Ethernet DWMAC glue driver");
+MODULE_AUTHOR("Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp");
+MODULE_LICENSE("GPL v2");
-- 
2.27.0

Re: [PATCH 2/4] net: stmmac: Add Toshiba Visconti SoCs glue driver

[David Miller]

From: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp>
Date: Thu, 11 Feb 2021 01:29:52 +0900

> +static int visconti_eth_init_hw(struct platform_device *pdev, struct plat_stmmacenet_data *plat_dat)
> +{
> +	struct visconti_eth *dwmac;
> +	unsigned int reg_val, clk_sel_val;

Please use reverse christmas tree ordering for local variable declarations.

> +static int visconti_eth_clock_probe(struct platform_device *pdev,
> +				    struct plat_stmmacenet_data *plat_dat)
> +{
> +	int err;
> +	struct visconti_eth *dwmac;

Likewise.

> +static int visconti_eth_clock_remove(struct platform_device *pdev)
> +{
> +	struct net_device *ndev = platform_get_drvdata(pdev);
> +	struct stmmac_priv *priv = netdev_priv(ndev);
> +	struct visconti_eth *dwmac = get_stmmac_bsp_priv(&pdev->dev);
Likewise.

Thanks.

Re: [PATCH 2/4] net: stmmac: Add Toshiba Visconti SoCs glue driver

[Nobuhiro Iwamatsu]

Hi,

Thanks for your comment.

On Thu, Feb 11, 2021 at 02:13:07PM -0800, David Miller wrote:
> From: Nobuhiro Iwamatsu <nobuhiro1.iwamatsu@toshiba.co.jp>
> Date: Thu, 11 Feb 2021 01:29:52 +0900
> 
> > +static int visconti_eth_init_hw(struct platform_device *pdev, struct plat_stmmacenet_data *plat_dat)
> > +{
> > +	struct visconti_eth *dwmac;
> > +	unsigned int reg_val, clk_sel_val;
> 
> Please use reverse christmas tree ordering for local variable declarations.
> 
> > +static int visconti_eth_clock_probe(struct platform_device *pdev,
> > +				    struct plat_stmmacenet_data *plat_dat)
> > +{
> > +	int err;
> > +	struct visconti_eth *dwmac;
> 
> Likewise.
> 
> > +static int visconti_eth_clock_remove(struct platform_device *pdev)
> > +{
> > +	struct net_device *ndev = platform_get_drvdata(pdev);
> > +	struct stmmac_priv *priv = netdev_priv(ndev);
> > +	struct visconti_eth *dwmac = get_stmmac_bsp_priv(&pdev->dev);
> Likewise.
> 

OK, I will fix these.

Best regards,
  Nobuhiro