[PATCH v2 net-next 0/3] net: stmmac: RK3566

[PATCH v2 net-next 0/3] net: stmmac: RK3566

[Ezequiel Garcia]

Now that RK3568 SoC devicetree upstreaming is happening [1],
here's another round for the RK3566 dwmac. There wasn't any clear
consensus on how to implement the two interfaces present
on RK3568, so I decided to drop that and just submit RK3566 for now.

This has been tested on a Pine64 RK3566 Quartz64 Model B board,
DHCP and iperf are looking good.

For all the people testing, here's Quartz 64 Model B device tree
snippet:

        gmac1: ethernet@fe010000 {
                compatible = "rockchip,rk3566-gmac", "snps,dwmac-4.20a";
                reg = <0x0 0xfe010000 0x0 0x10000>;
                interrupts = <GIC_SPI 32 IRQ_TYPE_LEVEL_HIGH>,
                             <GIC_SPI 29 IRQ_TYPE_LEVEL_HIGH>;
                interrupt-names = "macirq", "eth_wake_irq";
                rockchip,grf = <&grf>;
                clocks = <&cru SCLK_GMAC1>, <&cru SCLK_GMAC1_RX_TX>,
                         <&cru SCLK_GMAC1_RX_TX>, <&cru CLK_MAC1_REFOUT>,
                         <&cru ACLK_GMAC1>, <&cru PCLK_GMAC1>,
                         <&cru SCLK_GMAC1_RX_TX>, <&cru CLK_GMAC1_PTP_REF>;
                clock-names = "stmmaceth", "mac_clk_rx",
                              "mac_clk_tx", "clk_mac_refout",
                              "aclk_mac", "pclk_mac",
                              "clk_mac_speed", "ptp_ref";
                resets = <&cru SRST_A_GMAC1>;
                reset-names = "stmmaceth";

                snps,mixed-burst;
                snps,tso;

                snps,axi-config = <&gmac1_stmmac_axi_setup>;
                snps,mtl-rx-config = <&gmac1_mtl_rx_setup>;
                snps,mtl-tx-config = <&gmac1_mtl_tx_setup>;
                status = "disabled";

                mdio1: mdio {
                        compatible = "snps,dwmac-mdio";
                        #address-cells = <0x1>;
                        #size-cells = <0x0>;
                };

                gmac1_stmmac_axi_setup: stmmac-axi-config {
                        snps,wr_osr_lmt = <4>;
                        snps,rd_osr_lmt = <8>;
                        snps,blen = <0 0 0 0 16 8 4>;
                };

                gmac1_mtl_rx_setup: rx-queues-config {
                        snps,rx-queues-to-use = <1>;
                        queue0 {};
                };

                gmac1_mtl_tx_setup: tx-queues-config {
                        snps,tx-queues-to-use = <1>;
                        queue0 {};
                };
        };

While here, I'm adding a small patch from David Wu, for some
sanity checks for dwmac-rockchip-specific non-NULL ops.

Thanks!

[1] http://lore.kernel.org/r/20210421065921.23917-1-cl@rock-chips.com

David Wu (2):
  net: stmmac: dwmac-rk: Check platform-specific ops
  net: stmmac: Add RK3566 SoC support

Ezequiel Garcia (1):
  net: stmmac: Don't set has_gmac if has_gmac4 is set

 .../bindings/net/rockchip-dwmac.txt           |   1 +
 .../net/ethernet/stmicro/stmmac/dwmac-rk.c    | 126 +++++++++++++++++-
 2 files changed, 124 insertions(+), 3 deletions(-)

-- 
2.30.0


_______________________________________________
Linux-rockchip mailing list
Linux-rockchip@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-rockchip

[PATCH v2 1/3] net: stmmac: Don't set has_gmac if has_gmac4 is set

[Ezequiel Garcia]

Some Rockchip platforms have a GMAC4 core, and therefore
'plat_stmmacenet_data.has_gmac' shouldn't be set if
'plat_stmmacenet_data.has_gmac4' is set.

Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
---
 drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
index 6ef30252bfe0..c432a9592489 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
@@ -1400,7 +1400,11 @@ static int rk_gmac_probe(struct platform_device *pdev)
 	if (IS_ERR(plat_dat))
 		return PTR_ERR(plat_dat);
 
-	plat_dat->has_gmac = true;
+	/* If the stmmac is not already selected as gmac4,
+	 * then make sure we fallback to gmac.
+	 */
+	if (!plat_dat->has_gmac4)
+		plat_dat->has_gmac = true;
 	plat_dat->fix_mac_speed = rk_fix_speed;
 
 	plat_dat->bsp_priv = rk_gmac_setup(pdev, plat_dat, data);
-- 
2.30.0


_______________________________________________
Linux-rockchip mailing list
Linux-rockchip@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-rockchip

[PATCH v2 2/3] net: stmmac: dwmac-rk: Check platform-specific ops

[Ezequiel Garcia]

From: David Wu <david.wu@rock-chips.com>

Add a check for non-null struct rk_gmac_ops for the
configured PHY interface mode, failing if unsupported.

Signed-off-by: David Wu <david.wu@rock-chips.com>
[Ezequiel: Refactor so it fails if unsupported]
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
---
 .../net/ethernet/stmicro/stmmac/dwmac-rk.c    | 31 +++++++++++++++++--
 1 file changed, 29 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
index c432a9592489..d2637d83899e 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
@@ -1294,11 +1294,36 @@ static struct rk_priv_data *rk_gmac_setup(struct platform_device *pdev,
 	return bsp_priv;
 }
 
+static int rk_gmac_check_ops(struct rk_priv_data *bsp_priv)
+{
+	switch (bsp_priv->phy_iface) {
+	case PHY_INTERFACE_MODE_RGMII:
+	case PHY_INTERFACE_MODE_RGMII_ID:
+	case PHY_INTERFACE_MODE_RGMII_RXID:
+	case PHY_INTERFACE_MODE_RGMII_TXID:
+		if (!bsp_priv->ops->set_to_rgmii)
+			return -EINVAL;
+		break;
+	case PHY_INTERFACE_MODE_RMII:
+		if (!bsp_priv->ops->set_to_rmii)
+			return -EINVAL;
+		break;
+	default:
+		dev_err(&bsp_priv->pdev->dev,
+			"unsupported interface %d", bsp_priv->phy_iface);
+	}
+	return 0;
+}
+
 static int rk_gmac_powerup(struct rk_priv_data *bsp_priv)
 {
 	int ret;
 	struct device *dev = &bsp_priv->pdev->dev;
 
+	ret = rk_gmac_check_ops(bsp_priv);
+	if (ret)
+		return ret;
+
 	ret = gmac_clk_enable(bsp_priv, true);
 	if (ret)
 		return ret;
@@ -1369,10 +1394,12 @@ static void rk_fix_speed(void *priv, unsigned int speed)
 	case PHY_INTERFACE_MODE_RGMII_ID:
 	case PHY_INTERFACE_MODE_RGMII_RXID:
 	case PHY_INTERFACE_MODE_RGMII_TXID:
-		bsp_priv->ops->set_rgmii_speed(bsp_priv, speed);
+		if (bsp_priv->ops->set_rgmii_speed)
+			bsp_priv->ops->set_rgmii_speed(bsp_priv, speed);
 		break;
 	case PHY_INTERFACE_MODE_RMII:
-		bsp_priv->ops->set_rmii_speed(bsp_priv, speed);
+		if (bsp_priv->ops->set_rmii_speed)
+			bsp_priv->ops->set_rmii_speed(bsp_priv, speed);
 		break;
 	default:
 		dev_err(dev, "unsupported interface %d", bsp_priv->phy_iface);
-- 
2.30.0


_______________________________________________
Linux-rockchip mailing list
Linux-rockchip@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-rockchip

[PATCH v2 3/3] net: stmmac: Add RK3566 SoC support

[Ezequiel Garcia]

From: David Wu <david.wu@rock-chips.com>

Add constants and callback functions for the dwmac present
on RK3566 SoCs. As can be seen, the base structure
is the same, only registers and the bits in them moved slightly.

The dwmac IP core version v5.1, and so the compatible string
is expected to be:

  compatible = "rockchip,rk3566-gmac", "snps,dwmac-4.20a";

Signed-off-by: David Wu <david.wu@rock-chips.com>
[Ezequiel: Separate rk3566-gmac support]
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
---
 .../bindings/net/rockchip-dwmac.txt           |  1 +
 .../net/ethernet/stmicro/stmmac/dwmac-rk.c    | 89 +++++++++++++++++++
 2 files changed, 90 insertions(+)

diff --git a/Documentation/devicetree/bindings/net/rockchip-dwmac.txt b/Documentation/devicetree/bindings/net/rockchip-dwmac.txt
index 3b71da7e8742..3bd4bbcd6c65 100644
--- a/Documentation/devicetree/bindings/net/rockchip-dwmac.txt
+++ b/Documentation/devicetree/bindings/net/rockchip-dwmac.txt
@@ -12,6 +12,7 @@ Required properties:
    "rockchip,rk3366-gmac": found on RK3366 SoCs
    "rockchip,rk3368-gmac": found on RK3368 SoCs
    "rockchip,rk3399-gmac": found on RK3399 SoCs
+   "rockchip,rk3566-gmac", "snps,dwmac-4.20a": found on RK3566 SoCs
    "rockchip,rv1108-gmac": found on RV1108 SoCs
  - reg: addresses and length of the register sets for the device.
  - interrupts: Should contain the GMAC interrupts.
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
index d2637d83899e..a8850bcd980f 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
@@ -948,6 +948,94 @@ static const struct rk_gmac_ops rk3399_ops = {
 	.set_rmii_speed = rk3399_set_rmii_speed,
 };
 
+#define RK3568_GRF_GMAC0_CON0		0x0380
+#define RK3568_GRF_GMAC0_CON1		0x0384
+#define RK3568_GRF_GMAC1_CON0		0x0388
+#define RK3568_GRF_GMAC1_CON1		0x038c
+
+/* RK3568_GRF_GMAC0_CON1 && RK3568_GRF_GMAC1_CON1 */
+#define RK3568_GMAC_PHY_INTF_SEL_RGMII	\
+		(GRF_BIT(4) | GRF_CLR_BIT(5) | GRF_CLR_BIT(6))
+#define RK3568_GMAC_PHY_INTF_SEL_RMII	\
+		(GRF_CLR_BIT(4) | GRF_CLR_BIT(5) | GRF_BIT(6))
+#define RK3568_GMAC_FLOW_CTRL			GRF_BIT(3)
+#define RK3568_GMAC_FLOW_CTRL_CLR		GRF_CLR_BIT(3)
+#define RK3568_GMAC_RXCLK_DLY_ENABLE		GRF_BIT(1)
+#define RK3568_GMAC_RXCLK_DLY_DISABLE		GRF_CLR_BIT(1)
+#define RK3568_GMAC_TXCLK_DLY_ENABLE		GRF_BIT(0)
+#define RK3568_GMAC_TXCLK_DLY_DISABLE		GRF_CLR_BIT(0)
+
+/* RK3568_GRF_GMAC0_CON0 && RK3568_GRF_GMAC1_CON0 */
+#define RK3568_GMAC_CLK_RX_DL_CFG(val)	HIWORD_UPDATE(val, 0x7F, 8)
+#define RK3568_GMAC_CLK_TX_DL_CFG(val)	HIWORD_UPDATE(val, 0x7F, 0)
+
+static void rk3566_set_to_rgmii(struct rk_priv_data *bsp_priv,
+				int tx_delay, int rx_delay)
+{
+	struct device *dev = &bsp_priv->pdev->dev;
+
+	if (IS_ERR(bsp_priv->grf)) {
+		dev_err(dev, "Missing rockchip,grf property\n");
+		return;
+	}
+
+	regmap_write(bsp_priv->grf, RK3568_GRF_GMAC1_CON1,
+		     RK3568_GMAC_PHY_INTF_SEL_RGMII |
+		     RK3568_GMAC_RXCLK_DLY_ENABLE |
+		     RK3568_GMAC_TXCLK_DLY_ENABLE);
+
+	regmap_write(bsp_priv->grf, RK3568_GRF_GMAC1_CON0,
+		     RK3568_GMAC_CLK_RX_DL_CFG(rx_delay) |
+		     RK3568_GMAC_CLK_TX_DL_CFG(tx_delay));
+}
+
+static void rk3566_set_to_rmii(struct rk_priv_data *bsp_priv)
+{
+	struct device *dev = &bsp_priv->pdev->dev;
+
+	if (IS_ERR(bsp_priv->grf)) {
+		dev_err(dev, "%s: Missing rockchip,grf property\n", __func__);
+		return;
+	}
+
+	regmap_write(bsp_priv->grf, RK3568_GRF_GMAC1_CON1,
+		     RK3568_GMAC_PHY_INTF_SEL_RMII);
+}
+
+static void rk3566_set_gmac_speed(struct rk_priv_data *bsp_priv, int speed)
+{
+	struct device *dev = &bsp_priv->pdev->dev;
+	unsigned long rate;
+	int ret;
+
+	switch (speed) {
+	case 10:
+		rate = 2500000;
+		break;
+	case 100:
+		rate = 25000000;
+		break;
+	case 1000:
+		rate = 125000000;
+		break;
+	default:
+		dev_err(dev, "unknown speed value for GMAC speed=%d", speed);
+		return;
+	}
+
+	ret = clk_set_rate(bsp_priv->clk_mac_speed, rate);
+	if (ret)
+		dev_err(dev, "%s: set clk_mac_speed rate %ld failed %d\n",
+			__func__, rate, ret);
+}
+
+static const struct rk_gmac_ops rk3566_ops = {
+	.set_to_rgmii = rk3566_set_to_rgmii,
+	.set_to_rmii = rk3566_set_to_rmii,
+	.set_rgmii_speed = rk3566_set_gmac_speed,
+	.set_rmii_speed = rk3566_set_gmac_speed,
+};
+
 #define RV1108_GRF_GMAC_CON0		0X0900
 
 /* RV1108_GRF_GMAC_CON0 */
@@ -1512,6 +1600,7 @@ static const struct of_device_id rk_gmac_dwmac_match[] = {
 	{ .compatible = "rockchip,rk3366-gmac", .data = &rk3366_ops },
 	{ .compatible = "rockchip,rk3368-gmac", .data = &rk3368_ops },
 	{ .compatible = "rockchip,rk3399-gmac", .data = &rk3399_ops },
+	{ .compatible = "rockchip,rk3566-gmac", .data = &rk3566_ops },
 	{ .compatible = "rockchip,rv1108-gmac", .data = &rv1108_ops },
 	{ }
 };
-- 
2.30.0


_______________________________________________
Linux-rockchip mailing list
Linux-rockchip@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-rockchip

Re: [PATCH v2 3/3] net: stmmac: Add RK3566 SoC support

[Johan Jonker]

On 4/21/21 10:34 PM, Ezequiel Garcia wrote:
> From: David Wu <david.wu@rock-chips.com>
> 
> Add constants and callback functions for the dwmac present
> on RK3566 SoCs. As can be seen, the base structure
> is the same, only registers and the bits in them moved slightly.
> 
> The dwmac IP core version v5.1, and so the compatible string
> is expected to be:
> 
>   compatible = "rockchip,rk3566-gmac", "snps,dwmac-4.20a";
> 
> Signed-off-by: David Wu <david.wu@rock-chips.com>
> [Ezequiel: Separate rk3566-gmac support]
> Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
> ---
>  .../bindings/net/rockchip-dwmac.txt           |  1 +
>  .../net/ethernet/stmicro/stmmac/dwmac-rk.c    | 89 +++++++++++++++++++
>  2 files changed, 90 insertions(+)
> 
> diff --git a/Documentation/devicetree/bindings/net/rockchip-dwmac.txt b/Documentation/devicetree/bindings/net/rockchip-dwmac.txt
> index 3b71da7e8742..3bd4bbcd6c65 100644
> --- a/Documentation/devicetree/bindings/net/rockchip-dwmac.txt
> +++ b/Documentation/devicetree/bindings/net/rockchip-dwmac.txt
> @@ -12,6 +12,7 @@ Required properties:
>     "rockchip,rk3366-gmac": found on RK3366 SoCs
>     "rockchip,rk3368-gmac": found on RK3368 SoCs
>     "rockchip,rk3399-gmac": found on RK3399 SoCs

> +   "rockchip,rk3566-gmac", "snps,dwmac-4.20a": found on RK3566 SoCs

This is still a text document.
rob+dt has now scripts that check for undocumented compatibility
strings, so first convert rockchip-dwmac.txt to YAML and then add this
in a separated patch.

Johan

>     "rockchip,rv1108-gmac": found on RV1108 SoCs
>   - reg: addresses and length of the register sets for the device.
>   - interrupts: Should contain the GMAC interrupts.
> diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
> index d2637d83899e..a8850bcd980f 100644
> --- a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
> +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c
> @@ -948,6 +948,94 @@ static const struct rk_gmac_ops rk3399_ops = {
>  	.set_rmii_speed = rk3399_set_rmii_speed,
>  };
>  
> +#define RK3568_GRF_GMAC0_CON0		0x0380
> +#define RK3568_GRF_GMAC0_CON1		0x0384
> +#define RK3568_GRF_GMAC1_CON0		0x0388
> +#define RK3568_GRF_GMAC1_CON1		0x038c
> +
> +/* RK3568_GRF_GMAC0_CON1 && RK3568_GRF_GMAC1_CON1 */
> +#define RK3568_GMAC_PHY_INTF_SEL_RGMII	\
> +		(GRF_BIT(4) | GRF_CLR_BIT(5) | GRF_CLR_BIT(6))
> +#define RK3568_GMAC_PHY_INTF_SEL_RMII	\
> +		(GRF_CLR_BIT(4) | GRF_CLR_BIT(5) | GRF_BIT(6))
> +#define RK3568_GMAC_FLOW_CTRL			GRF_BIT(3)
> +#define RK3568_GMAC_FLOW_CTRL_CLR		GRF_CLR_BIT(3)
> +#define RK3568_GMAC_RXCLK_DLY_ENABLE		GRF_BIT(1)
> +#define RK3568_GMAC_RXCLK_DLY_DISABLE		GRF_CLR_BIT(1)
> +#define RK3568_GMAC_TXCLK_DLY_ENABLE		GRF_BIT(0)
> +#define RK3568_GMAC_TXCLK_DLY_DISABLE		GRF_CLR_BIT(0)
> +
> +/* RK3568_GRF_GMAC0_CON0 && RK3568_GRF_GMAC1_CON0 */
> +#define RK3568_GMAC_CLK_RX_DL_CFG(val)	HIWORD_UPDATE(val, 0x7F, 8)
> +#define RK3568_GMAC_CLK_TX_DL_CFG(val)	HIWORD_UPDATE(val, 0x7F, 0)
> +
> +static void rk3566_set_to_rgmii(struct rk_priv_data *bsp_priv,
> +				int tx_delay, int rx_delay)
> +{
> +	struct device *dev = &bsp_priv->pdev->dev;
> +
> +	if (IS_ERR(bsp_priv->grf)) {
> +		dev_err(dev, "Missing rockchip,grf property\n");
> +		return;
> +	}
> +
> +	regmap_write(bsp_priv->grf, RK3568_GRF_GMAC1_CON1,
> +		     RK3568_GMAC_PHY_INTF_SEL_RGMII |
> +		     RK3568_GMAC_RXCLK_DLY_ENABLE |
> +		     RK3568_GMAC_TXCLK_DLY_ENABLE);
> +
> +	regmap_write(bsp_priv->grf, RK3568_GRF_GMAC1_CON0,
> +		     RK3568_GMAC_CLK_RX_DL_CFG(rx_delay) |
> +		     RK3568_GMAC_CLK_TX_DL_CFG(tx_delay));
> +}
> +
> +static void rk3566_set_to_rmii(struct rk_priv_data *bsp_priv)
> +{
> +	struct device *dev = &bsp_priv->pdev->dev;
> +
> +	if (IS_ERR(bsp_priv->grf)) {
> +		dev_err(dev, "%s: Missing rockchip,grf property\n", __func__);
> +		return;
> +	}
> +
> +	regmap_write(bsp_priv->grf, RK3568_GRF_GMAC1_CON1,
> +		     RK3568_GMAC_PHY_INTF_SEL_RMII);
> +}
> +
> +static void rk3566_set_gmac_speed(struct rk_priv_data *bsp_priv, int speed)
> +{
> +	struct device *dev = &bsp_priv->pdev->dev;
> +	unsigned long rate;
> +	int ret;
> +
> +	switch (speed) {
> +	case 10:
> +		rate = 2500000;
> +		break;
> +	case 100:
> +		rate = 25000000;
> +		break;
> +	case 1000:
> +		rate = 125000000;
> +		break;
> +	default:
> +		dev_err(dev, "unknown speed value for GMAC speed=%d", speed);
> +		return;
> +	}
> +
> +	ret = clk_set_rate(bsp_priv->clk_mac_speed, rate);
> +	if (ret)
> +		dev_err(dev, "%s: set clk_mac_speed rate %ld failed %d\n",
> +			__func__, rate, ret);
> +}
> +
> +static const struct rk_gmac_ops rk3566_ops = {
> +	.set_to_rgmii = rk3566_set_to_rgmii,
> +	.set_to_rmii = rk3566_set_to_rmii,
> +	.set_rgmii_speed = rk3566_set_gmac_speed,
> +	.set_rmii_speed = rk3566_set_gmac_speed,
> +};
> +
>  #define RV1108_GRF_GMAC_CON0		0X0900
>  
>  /* RV1108_GRF_GMAC_CON0 */
> @@ -1512,6 +1600,7 @@ static const struct of_device_id rk_gmac_dwmac_match[] = {
>  	{ .compatible = "rockchip,rk3366-gmac", .data = &rk3366_ops },
>  	{ .compatible = "rockchip,rk3368-gmac", .data = &rk3368_ops },
>  	{ .compatible = "rockchip,rk3399-gmac", .data = &rk3399_ops },
> +	{ .compatible = "rockchip,rk3566-gmac", .data = &rk3566_ops },
>  	{ .compatible = "rockchip,rv1108-gmac", .data = &rv1108_ops },
>  	{ }
>  };
> 


_______________________________________________
Linux-rockchip mailing list
Linux-rockchip@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-rockchip

Re: [PATCH v2 3/3] net: stmmac: Add RK3566 SoC support

[Ezequiel Garcia]

On Wed, 2021-04-21 at 22:49 +0200, Johan Jonker wrote:
> On 4/21/21 10:34 PM, Ezequiel Garcia wrote:
> > From: David Wu <david.wu@rock-chips.com>
> > 
> > Add constants and callback functions for the dwmac present
> > on RK3566 SoCs. As can be seen, the base structure
> > is the same, only registers and the bits in them moved slightly.
> > 
> > The dwmac IP core version v5.1, and so the compatible string
> > is expected to be:
> > 
> >   compatible = "rockchip,rk3566-gmac", "snps,dwmac-4.20a";
> > 
> > Signed-off-by: David Wu <david.wu@rock-chips.com>
> > [Ezequiel: Separate rk3566-gmac support]
> > Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
> > ---
> >  .../bindings/net/rockchip-dwmac.txt           |  1 +
> >  .../net/ethernet/stmicro/stmmac/dwmac-rk.c    | 89 +++++++++++++++++++
> >  2 files changed, 90 insertions(+)
> > 
> > diff --git a/Documentation/devicetree/bindings/net/rockchip-dwmac.txt b/Documentation/devicetree/bindings/net/rockchip-dwmac.txt
> > index 3b71da7e8742..3bd4bbcd6c65 100644
> > --- a/Documentation/devicetree/bindings/net/rockchip-dwmac.txt
> > +++ b/Documentation/devicetree/bindings/net/rockchip-dwmac.txt
> > @@ -12,6 +12,7 @@ Required properties:
> >     "rockchip,rk3366-gmac": found on RK3366 SoCs
> >     "rockchip,rk3368-gmac": found on RK3368 SoCs
> >     "rockchip,rk3399-gmac": found on RK3399 SoCs
> 
> > +   "rockchip,rk3566-gmac", "snps,dwmac-4.20a": found on RK3566 SoCs
> 
> This is still a text document.
> rob+dt has now scripts that check for undocumented compatibility
> strings, so first convert rockchip-dwmac.txt to YAML and then add this
> in a separated patch.
> 

Is it a showstopper to convert devicetree bindings to YAML, for driver submission?

Thanks,
Ezequiel


_______________________________________________
Linux-rockchip mailing list
Linux-rockchip@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-rockchip

Re: [PATCH v2 net-next 0/3] net: stmmac: RK3566

[David Wu]

Hi Ezequiel,

在 2021/4/22 上午4:34, Ezequiel Garcia 写道:
> Now that RK3568 SoC devicetree upstreaming is happening [1],
> here's another round for the RK3566 dwmac. There wasn't any clear
> consensus on how to implement the two interfaces present
> on RK3568, so I decided to drop that and just submit RK3566 for now.
> 
> This has been tested on a Pine64 RK3566 Quartz64 Model B board,
> DHCP and iperf are looking good.
> 
> For all the people testing, here's Quartz 64 Model B device tree
> snippet:
> 
>          gmac1: ethernet@fe010000 {
>                  compatible = "rockchip,rk3566-gmac", "snps,dwmac-4.20a";

It is better to use "rockchip,rk3568-gmac" here, "rockchip,rk3566-gmac" 
is not compatible, 3568 has two gmacs, which are compatible with 3566.

If there is no better way, using bus_id from alias is good, it is a 
fixed id, and U-Boot also use the id to write MAC address into kernel DTB.

plat->bus_id = of_alias_get_id(np, "ethernet");

>                  reg = <0x0 0xfe010000 0x0 0x10000>;
>                  interrupts = <GIC_SPI 32 IRQ_TYPE_LEVEL_HIGH>,
>                               <GIC_SPI 29 IRQ_TYPE_LEVEL_HIGH>;
>                  interrupt-names = "macirq", "eth_wake_irq";
>                  rockchip,grf = <&grf>;
>                  clocks = <&cru SCLK_GMAC1>, <&cru SCLK_GMAC1_RX_TX>,
>                           <&cru SCLK_GMAC1_RX_TX>, <&cru CLK_MAC1_REFOUT>,
>                           <&cru ACLK_GMAC1>, <&cru PCLK_GMAC1>,
>                           <&cru SCLK_GMAC1_RX_TX>, <&cru CLK_GMAC1_PTP_REF>;
>                  clock-names = "stmmaceth", "mac_clk_rx",
>                                "mac_clk_tx", "clk_mac_refout",
>                                "aclk_mac", "pclk_mac",
>                                "clk_mac_speed", "ptp_ref";
>                  resets = <&cru SRST_A_GMAC1>;
>                  reset-names = "stmmaceth";
> 
>                  snps,mixed-burst;
>                  snps,tso;
> 
>                  snps,axi-config = <&gmac1_stmmac_axi_setup>;
>                  snps,mtl-rx-config = <&gmac1_mtl_rx_setup>;
>                  snps,mtl-tx-config = <&gmac1_mtl_tx_setup>;
>                  status = "disabled";
> 
>                  mdio1: mdio {
>                          compatible = "snps,dwmac-mdio";
>                          #address-cells = <0x1>;
>                          #size-cells = <0x0>;
>                  };
> 
>                  gmac1_stmmac_axi_setup: stmmac-axi-config {
>                          snps,wr_osr_lmt = <4>;
>                          snps,rd_osr_lmt = <8>;
>                          snps,blen = <0 0 0 0 16 8 4>;
>                  };
> 
>                  gmac1_mtl_rx_setup: rx-queues-config {
>                          snps,rx-queues-to-use = <1>;
>                          queue0 {};
>                  };
> 
>                  gmac1_mtl_tx_setup: tx-queues-config {
>                          snps,tx-queues-to-use = <1>;
>                          queue0 {};
>                  };
>          };
> 
> While here, I'm adding a small patch from David Wu, for some
> sanity checks for dwmac-rockchip-specific non-NULL ops.
> 
> Thanks!
> 
> [1] http://lore.kernel.org/r/20210421065921.23917-1-cl@rock-chips.com
> 
> David Wu (2):
>    net: stmmac: dwmac-rk: Check platform-specific ops
>    net: stmmac: Add RK3566 SoC support
> 
> Ezequiel Garcia (1):
>    net: stmmac: Don't set has_gmac if has_gmac4 is set
> 
>   .../bindings/net/rockchip-dwmac.txt           |   1 +
>   .../net/ethernet/stmicro/stmmac/dwmac-rk.c    | 126 +++++++++++++++++-
>   2 files changed, 124 insertions(+), 3 deletions(-)
> 



_______________________________________________
Linux-rockchip mailing list
Linux-rockchip@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-rockchip

Re: [PATCH v2 net-next 0/3] net: stmmac: RK3566

[Ezequiel Garcia]

Hi David,

Thanks a lot for reviewing.

On Fri, 2021-04-23 at 17:24 +0800, David Wu wrote:
> Hi Ezequiel,
> 
> 在 2021/4/22 上午4:34, Ezequiel Garcia 写道:
> > Now that RK3568 SoC devicetree upstreaming is happening [1],
> > here's another round for the RK3566 dwmac. There wasn't any clear
> > consensus on how to implement the two interfaces present
> > on RK3568, so I decided to drop that and just submit RK3566 for now.
> > 
> > This has been tested on a Pine64 RK3566 Quartz64 Model B board,
> > DHCP and iperf are looking good.
> > 
> > For all the people testing, here's Quartz 64 Model B device tree
> > snippet:
> > 
> >          gmac1: ethernet@fe010000 {
> >                  compatible = "rockchip,rk3566-gmac", "snps,dwmac-4.20a";
> 
> It is better to use "rockchip,rk3568-gmac" here, "rockchip,rk3566-gmac" 
> is not compatible, 3568 has two gmacs, which are compatible with 3566.
> 
> If there is no better way, using bus_id from alias is good, it is a 
> fixed id, and U-Boot also use the id to write MAC address into kernel DTB.
> 
> plat->bus_id = of_alias_get_id(np, "ethernet");
> 

This was discussed, see ChenYu's reply. I think the idea
is considered fragile:

https://lore.kernel.org/netdev/CAGb2v67ZBR=XDFPeXQc429HNu_dbY__-KN50tvBW44fXMs78_w@mail.gmail.com/

However, I agree with you about going back to just "rockchip,rk3568-gmac".
Adding rockchip,rk3566-gmac maybe wasn't a great idea :-)

The most accepted way forward seems Heiko's proposal of hardcoding the mmio
addresses to identify each block, as it's done in the DSI driver:

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/gpu/drm/rockchip/dw-mipi-dsi-rockchip.c#n1170

Would you agree with it?

Thanks!
Ezequiel
 
> >                  reg = <0x0 0xfe010000 0x0 0x10000>;
> >                  interrupts = <GIC_SPI 32 IRQ_TYPE_LEVEL_HIGH>,
> >                               <GIC_SPI 29 IRQ_TYPE_LEVEL_HIGH>;
> >                  interrupt-names = "macirq", "eth_wake_irq";
> >                  rockchip,grf = <&grf>;
> >                  clocks = <&cru SCLK_GMAC1>, <&cru SCLK_GMAC1_RX_TX>,
> >                           <&cru SCLK_GMAC1_RX_TX>, <&cru CLK_MAC1_REFOUT>,
> >                           <&cru ACLK_GMAC1>, <&cru PCLK_GMAC1>,
> >                           <&cru SCLK_GMAC1_RX_TX>, <&cru CLK_GMAC1_PTP_REF>;
> >                  clock-names = "stmmaceth", "mac_clk_rx",
> >                                "mac_clk_tx", "clk_mac_refout",
> >                                "aclk_mac", "pclk_mac",
> >                                "clk_mac_speed", "ptp_ref";
> >                  resets = <&cru SRST_A_GMAC1>;
> >                  reset-names = "stmmaceth";
> > 
> >                  snps,mixed-burst;
> >                  snps,tso;
> > 
> >                  snps,axi-config = <&gmac1_stmmac_axi_setup>;
> >                  snps,mtl-rx-config = <&gmac1_mtl_rx_setup>;
> >                  snps,mtl-tx-config = <&gmac1_mtl_tx_setup>;
> >                  status = "disabled";
> > 
> >                  mdio1: mdio {
> >                          compatible = "snps,dwmac-mdio";
> >                          #address-cells = <0x1>;
> >                          #size-cells = <0x0>;
> >                  };
> > 
> >                  gmac1_stmmac_axi_setup: stmmac-axi-config {
> >                          snps,wr_osr_lmt = <4>;
> >                          snps,rd_osr_lmt = <8>;
> >                          snps,blen = <0 0 0 0 16 8 4>;
> >                  };
> > 
> >                  gmac1_mtl_rx_setup: rx-queues-config {
> >                          snps,rx-queues-to-use = <1>;
> >                          queue0 {};
> >                  };
> > 
> >                  gmac1_mtl_tx_setup: tx-queues-config {
> >                          snps,tx-queues-to-use = <1>;
> >                          queue0 {};
> >                  };
> >          };
> > 
> > While here, I'm adding a small patch from David Wu, for some
> > sanity checks for dwmac-rockchip-specific non-NULL ops.
> > 
> > Thanks!
> > 
> > [1] http://lore.kernel.org/r/20210421065921.23917-1-cl@rock-chips.com
> > 
> > David Wu (2):
> >    net: stmmac: dwmac-rk: Check platform-specific ops
> >    net: stmmac: Add RK3566 SoC support
> > 
> > Ezequiel Garcia (1):
> >    net: stmmac: Don't set has_gmac if has_gmac4 is set
> > 
> >   .../bindings/net/rockchip-dwmac.txt           |   1 +
> >   .../net/ethernet/stmicro/stmmac/dwmac-rk.c    | 126 +++++++++++++++++-
> >   2 files changed, 124 insertions(+), 3 deletions(-)
> > 
> 
> 



_______________________________________________
Linux-rockchip mailing list
Linux-rockchip@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-rockchip

Re: [PATCH v2 net-next 0/3] net: stmmac: RK3566

[David Wu]

Hi Ezequiel,

在 2021/4/23 下午9:57, Ezequiel Garcia 写道:
> Hi David,
> 
> Thanks a lot for reviewing.
> 
> On Fri, 2021-04-23 at 17:24 +0800, David Wu wrote:
>> Hi Ezequiel,
>>
>> 在 2021/4/22 上午4:34, Ezequiel Garcia 写道:
>>> Now that RK3568 SoC devicetree upstreaming is happening [1],
>>> here's another round for the RK3566 dwmac. There wasn't any clear
>>> consensus on how to implement the two interfaces present
>>> on RK3568, so I decided to drop that and just submit RK3566 for now.
>>>
>>> This has been tested on a Pine64 RK3566 Quartz64 Model B board,
>>> DHCP and iperf are looking good.
>>>
>>> For all the people testing, here's Quartz 64 Model B device tree
>>> snippet:
>>>
>>>           gmac1: ethernet@fe010000 {
>>>                   compatible = "rockchip,rk3566-gmac", "snps,dwmac-4.20a";
>>
>> It is better to use "rockchip,rk3568-gmac" here, "rockchip,rk3566-gmac"
>> is not compatible, 3568 has two gmacs, which are compatible with 3566.
>>
>> If there is no better way, using bus_id from alias is good, it is a
>> fixed id, and U-Boot also use the id to write MAC address into kernel DTB.
>>
>> plat->bus_id = of_alias_get_id(np, "ethernet");
>>
> 
> This was discussed, see ChenYu's reply. I think the idea
> is considered fragile:
> 
> https://lore.kernel.org/netdev/CAGb2v67ZBR=XDFPeXQc429HNu_dbY__-KN50tvBW44fXMs78_w@mail.gmail.com/
> 
> However, I agree with you about going back to just "rockchip,rk3568-gmac".
> Adding rockchip,rk3566-gmac maybe wasn't a great idea :-)
> 
> The most accepted way forward seems Heiko's proposal of hardcoding the mmio
> addresses to identify each block, as it's done in the DSI driver:
> 
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/gpu/drm/rockchip/dw-mipi-dsi-rockchip.c#n1170
> 
> Would you agree with it?
> 

Yes, I agree.
I also think the alias should be rarely changed. The DSI mmio addresses 
example here can be used for reference, and it seems that it will be 
stronger than alias.

> Thanks!
> Ezequiel
>   
>>>                   reg = <0x0 0xfe010000 0x0 0x10000>;
>>>                   interrupts = <GIC_SPI 32 IRQ_TYPE_LEVEL_HIGH>,
>>>                                <GIC_SPI 29 IRQ_TYPE_LEVEL_HIGH>;
>>>                   interrupt-names = "macirq", "eth_wake_irq";
>>>                   rockchip,grf = <&grf>;
>>>                   clocks = <&cru SCLK_GMAC1>, <&cru SCLK_GMAC1_RX_TX>,
>>>                            <&cru SCLK_GMAC1_RX_TX>, <&cru CLK_MAC1_REFOUT>,
>>>                            <&cru ACLK_GMAC1>, <&cru PCLK_GMAC1>,
>>>                            <&cru SCLK_GMAC1_RX_TX>, <&cru CLK_GMAC1_PTP_REF>;
>>>                   clock-names = "stmmaceth", "mac_clk_rx",
>>>                                 "mac_clk_tx", "clk_mac_refout",
>>>                                 "aclk_mac", "pclk_mac",
>>>                                 "clk_mac_speed", "ptp_ref";
>>>                   resets = <&cru SRST_A_GMAC1>;
>>>                   reset-names = "stmmaceth";
>>>
>>>                   snps,mixed-burst;
>>>                   snps,tso;
>>>
>>>                   snps,axi-config = <&gmac1_stmmac_axi_setup>;
>>>                   snps,mtl-rx-config = <&gmac1_mtl_rx_setup>;
>>>                   snps,mtl-tx-config = <&gmac1_mtl_tx_setup>;
>>>                   status = "disabled";
>>>
>>>                   mdio1: mdio {
>>>                           compatible = "snps,dwmac-mdio";
>>>                           #address-cells = <0x1>;
>>>                           #size-cells = <0x0>;
>>>                   };
>>>
>>>                   gmac1_stmmac_axi_setup: stmmac-axi-config {
>>>                           snps,wr_osr_lmt = <4>;
>>>                           snps,rd_osr_lmt = <8>;
>>>                           snps,blen = <0 0 0 0 16 8 4>;
>>>                   };
>>>
>>>                   gmac1_mtl_rx_setup: rx-queues-config {
>>>                           snps,rx-queues-to-use = <1>;
>>>                           queue0 {};
>>>                   };
>>>
>>>                   gmac1_mtl_tx_setup: tx-queues-config {
>>>                           snps,tx-queues-to-use = <1>;
>>>                           queue0 {};
>>>                   };
>>>           };
>>>
>>> While here, I'm adding a small patch from David Wu, for some
>>> sanity checks for dwmac-rockchip-specific non-NULL ops.
>>>
>>> Thanks!
>>>
>>> [1] http://lore.kernel.org/r/20210421065921.23917-1-cl@rock-chips.com
>>>
>>> David Wu (2):
>>>     net: stmmac: dwmac-rk: Check platform-specific ops
>>>     net: stmmac: Add RK3566 SoC support
>>>
>>> Ezequiel Garcia (1):
>>>     net: stmmac: Don't set has_gmac if has_gmac4 is set
>>>
>>>    .../bindings/net/rockchip-dwmac.txt           |   1 +
>>>    .../net/ethernet/stmicro/stmmac/dwmac-rk.c    | 126 +++++++++++++++++-
>>>    2 files changed, 124 insertions(+), 3 deletions(-)
>>>
>>
>>
> 
> 
> 
> 
> 



_______________________________________________
Linux-rockchip mailing list
Linux-rockchip@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-rockchip