Skip to contentSkip to conversation

Add support for imx8mm-venice-gw7901 and DSA switch

PATCH In review Tim Harvey 3 participants 25 messages Jun 30, 2021 – Jul 10, 2021

Subsystems (4)
lore
Share & tools
lore

Tim Harvey

author
#0 lore raw

cover Add support for imx8mm-venice-gw7901 and DSA switch

The GW7901 is an IMX8M Mini based board with a KSZ9897 switch hanging off the IMX8M FEC.

In order to support a fixed-link connection to a U-Boot DSA switch some changes need to be made to the FEC driver.

Additionally, in order to support per-port unique MAC addrs for DSA we must introduce a mechanism to put the master device in promiscuous mode.

All patches are here including adding the dts for the board as well as a dsa driver and necessary patches to fec/dsa so that it is clear how this fits together.

Tim Harvey (9):
board: gateworks: venice: add imx8mm-gw7901 support
net: fec: use device sequence vs index when fetching fec
net: fec: set phy_of_node properly for fixed-link phy
net: add set_promisc function to enable/disable Promiscuous mode
net: fec: add set_promisc function
net: dsa: enable master promisc mode if available and needed
net: add support for KSZ9477/KSZ9897/KSZ9567 GbE switch
arm: dts: imx8mm-venice-gw7901.dts: fix dsa switch configuration
configs: imx8mm_venice_defconfig: add support for gbe switch

 arch/arm/dts/Makefile                         |    1 +
 arch/arm/dts/imx8mm-venice-gw7901-u-boot.dtsi |  118 ++
 arch/arm/dts/imx8mm-venice-gw7901.dts         | 1055 +++++++++++++++++
 board/gateworks/venice/spl.c                  |   55 +-
 configs/imx8mm_venice_defconfig               |    6 +-
 drivers/net/Kconfig                           |    7 +
 drivers/net/Makefile                          |    1 +
 drivers/net/fec_mxc.c                         |   21 +-
 drivers/net/fec_mxc.h                         |    1 +
 drivers/net/ksz9477.c                         |  547 +++++++++
 include/net.h                                 |    2 +
 net/dsa-uclass.c                              |    9 +-
 12 files changed, 1818 insertions(+), 5 deletions(-)
 create mode 100644 arch/arm/dts/imx8mm-venice-gw7901-u-boot.dtsi
 create mode 100644 arch/arm/dts/imx8mm-venice-gw7901.dts
 create mode 100644 drivers/net/ksz9477.c
Signature · 1 line
2.17.1
author
#1 lore raw

PATCH 1/9 board: gateworks: venice: add imx8mm-gw7901 support

The Gateworks GW7901 is an ARM based single board computer (SBC) featuring:
- i.MX8M Mini SoC
- LPDDR4 DRAM
- eMMC FLASH
- SPI FRAM
- Gateworks System Controller (GSC)
- Atmel ATECC Crypto Authentication
- USB 2.0
- Microchip GbE Switch
- Multiple multi-protocol RS232/RS485/RS422 Serial ports
- onboard 802.11ac WiFi / BT
- microSD socket
- miniPCIe socket with PCIe, USB 2.0 and dual SIM sockets
- Wide range DC power input
- 802.3at PoE

To add support for this board:
- add dts from Linux (accepted for v5.14)
- add SPL PMIC config

---
 arch/arm/dts/Makefile                         |    1 +
 arch/arm/dts/imx8mm-venice-gw7901-u-boot.dtsi |  118 ++
 arch/arm/dts/imx8mm-venice-gw7901.dts         | 1020 +++++++++++++++++
 board/gateworks/venice/spl.c                  |   55 +-
 configs/imx8mm_venice_defconfig               |    2 +-
 5 files changed, 1194 insertions(+), 2 deletions(-)
 create mode 100644 arch/arm/dts/imx8mm-venice-gw7901-u-boot.dtsi
 create mode 100644 arch/arm/dts/imx8mm-venice-gw7901.dts
arch/arm/dts/Makefile +1 -0
index 9fb38682e6..b390d6bf1a 100644
@@ -861,6 +861,7 @@ dtb-$(CONFIG_ARCH_IMX8M) += \
 	imx8mm-venice-gw71xx-0x.dtb \
 	imx8mm-venice-gw72xx-0x.dtb \
 	imx8mm-venice-gw73xx-0x.dtb \
+	imx8mm-venice-gw7901.dtb \
 	imx8mm-verdin.dtb \
 	phycore-imx8mm.dtb \
 	imx8mn-ddr4-evk.dtb \
arch/arm/dts/imx8mm-venice-gw7901-u-boot.dtsi new file +118 -0

118 lines not on this page · raw message

arch/arm/dts/imx8mm-venice-gw7901.dts new file +1020 -0

1,020 lines not on this page · raw message

board/gateworks/venice/spl.c +54 -1

74 lines not on this page · raw message

configs/imx8mm_venice_defconfig +1 -1
index daffcc46d5..3961db00cb 100644
@@ -58,7 +58,7 @@ CONFIG_CMD_EXT4_WRITE=y
 # CONFIG_SPL_EFI_PARTITION is not set
 CONFIG_OF_CONTROL=y
 CONFIG_SPL_OF_CONTROL=y
-CONFIG_OF_LIST="imx8mm-venice-gw71xx-0x imx8mm-venice-gw72xx-0x imx8mm-venice-gw73xx-0x"
+CONFIG_OF_LIST="imx8mm-venice-gw71xx-0x imx8mm-venice-gw72xx-0x imx8mm-venice-gw73xx-0x imx8mm-venice-gw7901"
 CONFIG_ENV_IS_IN_MMC=y
 CONFIG_SYS_REDUNDAND_ENVIRONMENT=y
 CONFIG_ENV_VARS_UBOOT_RUNTIME_CONFIG=y
Signature · 1 line
2.17.1
#2 lore raw

PATCH 1/9 board: gateworks: venice: add imx8mm-gw7901 support

Quoted 19 lines

The Gateworks GW7901 is an ARM based single board computer (SBC) featuring:
- i.MX8M Mini SoC
- LPDDR4 DRAM
- eMMC FLASH
- SPI FRAM
- Gateworks System Controller (GSC)
- Atmel ATECC Crypto Authentication
- USB 2.0
- Microchip GbE Switch
- Multiple multi-protocol RS232/RS485/RS422 Serial ports
- onboard 802.11ac WiFi / BT
- microSD socket
- miniPCIe socket with PCIe, USB 2.0 and dual SIM sockets
- Wide range DC power input
- 802.3at PoE
To add support for this board:
- add dts from Linux (accepted for v5.14)
- add SPL PMIC config

Quoted 1 line

Applied to u-boot-imx, master, thanks !

Best regards,
Stefano Babic

Signature · 5 lines
=====================================================================
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-53 Fax: +49-8142-66989-80 Email: [email protected]
=====================================================================
author
#3 lore raw

PATCH 2/9 net: fec: use device sequence vs index when fetching fec

In reply to Tim Harvey

When using uclass_get_device* to get the FEC device we need to use device sequence instead of index into UCLASS_ETH. In systems where for example a I2C based DSA switch exists it will probe before the FEC master and its ports will be registered first and have the first indexes yet the FEC's sequence comes from the device-tree alias.

Take for example the imx8mm-venice-gw7901 board which has an i2c based DSA switch:

u-boot=> net list
eth1 : lan1 00:0d:8d:aa:00:2f
eth2 : lan2 00:0d:8d:aa:00:30
eth3 : lan3 00:0d:8d:aa:00:31
eth4 : lan4 00:0d:8d:aa:00:32
eth0 : ethernet@30be0000 00:0d:8d:aa:00:2e active

Thus in this case uclass_get_device(UCLASS_ETH, 0, &dev) returns lan1 which is wrong but uclass_get_device_seq(UCLASS_ETH, 0, &dev) returns ethernet@30be000 which is correct.

---
 drivers/net/fec_mxc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/fec_mxc.c +1 -1
index 4fd5c01b4a..3abf914833 100644
@@ -146,7 +146,7 @@ static int fec_get_clk_rate(void *udev, int idx)
 	    CONFIG_IS_ENABLED(CLK_CCF)) {
 		dev = udev;
 		if (!dev) {
-			ret = uclass_get_device(UCLASS_ETH, idx, &dev);
+			ret = uclass_get_device_by_seq(UCLASS_ETH, idx, &dev);
 			if (ret < 0) {
 				debug("Can't get FEC udev: %d\n", ret);
 				return ret;
Signature · 1 line
2.17.1
Reviewed-by on v1 2/9
#4 lore raw

Re: [PATCH 2/9] net: fec: use device sequence vs index when fetching fec

Quoted 40 lines · Tim Harvey

When using uclass_get_device* to get the FEC device we need to use device sequence instead of index into UCLASS_ETH. In systems where for example a I2C based DSA switch exists it will probe before the FEC master and its ports will be registered first and have the first indexes yet the FEC's sequence comes from the device-tree alias.

Take for example the imx8mm-venice-gw7901 board which has an i2c based DSA switch:

u-boot=> net list
eth1 : lan1 00:0d:8d:aa:00:2f
eth2 : lan2 00:0d:8d:aa:00:30
eth3 : lan3 00:0d:8d:aa:00:31
eth4 : lan4 00:0d:8d:aa:00:32
eth0 : ethernet@30be0000 00:0d:8d:aa:00:2e active

Thus in this case uclass_get_device(UCLASS_ETH, 0, &dev) returns lan1 which is wrong but uclass_get_device_seq(UCLASS_ETH, 0, &dev) returns ethernet@30be000 which is correct.

---
 drivers/net/fec_mxc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/fec_mxc.c +1 -2
index 4fd5c01b4a..3abf914833 100644
@@ -146,7 +146,7 @@ static int fec_get_clk_rate(void *udev, int idx)
            CONFIG_IS_ENABLED(CLK_CCF)) {
                dev = udev;
                if (!dev) {
-                       ret = uclass_get_device(UCLASS_ETH, idx, &dev);
+                       ret = uclass_get_device_by_seq(UCLASS_ETH, idx, &dev);
                        if (ret < 0) {
                                debug("Can't get FEC udev: %d\n", ret);
                                return ret;
--
Quoted 2 lines

2.17.1

#5 lore raw

PATCH 2/9 net: fec: use device sequence vs index when fetching fec

In reply to Tim Harvey
Quoted 16 lines

When using uclass_get_device* to get the FEC device we need to use device sequence instead of index into UCLASS_ETH. In systems where for example a I2C based DSA switch exists it will probe before the FEC master and its ports will be registered first and have the first indexes yet the FEC's sequence comes from the device-tree alias. Take for example the imx8mm-venice-gw7901 board which has an i2c based DSA switch:
u-boot=> net list
eth1 : lan1 00:0d:8d:aa:00:2f
eth2 : lan2 00:0d:8d:aa:00:30
eth3 : lan3 00:0d:8d:aa:00:31
eth4 : lan4 00:0d:8d:aa:00:32
eth0 : ethernet@30be0000 00:0d:8d:aa:00:2e active
Thus in this case uclass_get_device(UCLASS_ETH, 0, &dev) returns lan1 which is wrong but uclass_get_device_seq(UCLASS_ETH, 0, &dev) returns ethernet@30be000 which is correct.

Quoted 2 lines

Applied to u-boot-imx, master, thanks !

Best regards,
Stefano Babic

Signature · 5 lines
=====================================================================
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-53 Fax: +49-8142-66989-80 Email: [email protected]
=====================================================================
author
#6 lore raw

PATCH 3/9 net: fec: set phy_of_node properly for fixed-link phy

In reply to Tim Harvey

If the FEC is connected to a fixed-link (upstream switch port for example) the phy_of_node should be set to the fixed-link node so that speed and other properties can be found properly.

In addition fix a typo in the debug string.

---
 drivers/net/fec_mxc.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/net/fec_mxc.c +5 -1
index 3abf914833..77680491d0 100644
@@ -1304,7 +1304,11 @@ static int device_get_phy_addr(struct fec_priv *priv, struct udevice *dev)
 	ret = dev_read_phandle_with_args(dev, "phy-handle", NULL, 0, 0,
 					 &phandle_args);
 	if (ret) {
-		debug("Failed to find phy-handle (err = %d\n)", ret);
+		priv->phy_of_node = ofnode_find_subnode(dev_ofnode(dev),
+							"fixed-link");
+		if (ofnode_valid(priv->phy_of_node))
+			return 0;
+		debug("Failed to find phy-handle (err = %d)\n", ret);
 		return ret;
 	}
 
Signature · 1 line
2.17.1
Reviewed-by on v1 3/9
#7 lore raw

Re: [PATCH 3/9] net: fec: set phy_of_node properly for fixed-link phy

Quoted 30 lines · Tim Harvey

If the FEC is connected to a fixed-link (upstream switch port for example) the phy_of_node should be set to the fixed-link node so that speed and other properties can be found properly.

In addition fix a typo in the debug string.

---
 drivers/net/fec_mxc.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/net/fec_mxc.c +5 -2
index 3abf914833..77680491d0 100644
@@ -1304,7 +1304,11 @@ static int device_get_phy_addr(struct fec_priv *priv, struct udevice *dev)
        ret = dev_read_phandle_with_args(dev, "phy-handle", NULL, 0, 0,
                                         &phandle_args);
        if (ret) {
-               debug("Failed to find phy-handle (err = %d\n)", ret);
+               priv->phy_of_node = ofnode_find_subnode(dev_ofnode(dev),
+                                                       "fixed-link");
+               if (ofnode_valid(priv->phy_of_node))
+                       return 0;
+               debug("Failed to find phy-handle (err = %d)\n", ret);
                return ret;
        }
 
--
Quoted 2 lines

2.17.1

#8 lore raw

PATCH 3/9 net: fec: set phy_of_node properly for fixed-link phy

In reply to Tim Harvey
Quoted 6 lines

If the FEC is connected to a fixed-link (upstream switch port for example) the phy_of_node should be set to the fixed-link node so that speed and other properties can be found properly. In addition fix a typo in the debug string.

Applied to u-boot-imx, master, thanks !

Best regards,
Stefano Babic

Signature · 5 lines
=====================================================================
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-53 Fax: +49-8142-66989-80 Email: [email protected]
=====================================================================
author
#9 lore raw

PATCH 4/9 net: add set_promisc function to enable/disable Promiscuous mode

In reply to Tim Harvey

Enabling promiscuous mode can be useful for DSA switches where each port has its own MAC address.

---
 include/net.h | 2 ++
 1 file changed, 2 insertions(+)
include/net.h +2 -0
index b95d6a6f60..cec8c98618 100644
@@ -158,6 +158,7 @@ enum eth_recv_flags {
  *		    ROM on the board. This is how the driver should expose it
  *		    to the network stack. This function should fill in the
  *		    eth_pdata::enetaddr field - optional
+ * set_promisc: Enable or Disable promiscuous mode
  */
 struct eth_ops {
 	int (*start)(struct udevice *dev);
@@ -168,6 +169,7 @@ struct eth_ops {
 	int (*mcast)(struct udevice *dev, const u8 *enetaddr, int join);
 	int (*write_hwaddr)(struct udevice *dev);
 	int (*read_rom_hwaddr)(struct udevice *dev);
+	int (*set_promisc)(struct udevice *dev, bool enable);
 };
 
 #define eth_get_ops(dev) ((struct eth_ops *)(dev)->driver->ops)
Signature · 1 line
2.17.1
Reviewed-by on v1 4/9
#10 lore raw

Re: [PATCH 4/9] net: add set_promisc function to enable/disable Promiscuous mode

Quoted 30 lines · Tim Harvey

Enabling promiscuous mode can be useful for DSA switches where each port has its own MAC address.

---
 include/net.h | 2 ++
 1 file changed, 2 insertions(+)
include/net.h +2 -1
index b95d6a6f60..cec8c98618 100644
@@ -158,6 +158,7 @@ enum eth_recv_flags {
  *                 ROM on the board. This is how the driver should expose it
  *                 to the network stack. This function should fill in the
  *                 eth_pdata::enetaddr field - optional
+ * set_promisc: Enable or Disable promiscuous mode
  */
 struct eth_ops {
        int (*start)(struct udevice *dev);
@@ -168,6 +169,7 @@ struct eth_ops {
        int (*mcast)(struct udevice *dev, const u8 *enetaddr, int join);
        int (*write_hwaddr)(struct udevice *dev);
        int (*read_rom_hwaddr)(struct udevice *dev);
+       int (*set_promisc)(struct udevice *dev, bool enable);
 };
 
 #define eth_get_ops(dev) ((struct eth_ops *)(dev)->driver->ops)
--
Quoted 2 lines

2.17.1

#11 lore raw

PATCH 4/9 net: add set_promisc function to enable/disable Promiscuous mode

In reply to Tim Harvey
Quoted 4 lines

Enabling promiscuous mode can be useful for DSA switches where each port has its own MAC address.

Applied to u-boot-imx, master, thanks !

Best regards,
Stefano Babic

Signature · 5 lines
=====================================================================
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-53 Fax: +49-8142-66989-80 Email: [email protected]
=====================================================================
author
#12 lore raw

PATCH 5/9 net: fec: add set_promisc function

In reply to Tim Harvey

Enabling promiscuous mode is necessary if FEC is the master of a DSA switch driver where each port has their own MAC address.

---
 drivers/net/fec_mxc.c | 13 +++++++++++++
 drivers/net/fec_mxc.h |  1 +
 2 files changed, 14 insertions(+)
drivers/net/fec_mxc.c +13 -0
index 77680491d0..db2cdaf684 100644
@@ -458,6 +458,9 @@ static void fec_reg_setup(struct fec_priv *fec)
 	else if (fec->xcv_type == RMII)
 		rcntrl |= FEC_RCNTRL_RMII;
 
+	if (fec->promisc)
+		rcntrl |= 0x8;
+
 	writel(rcntrl, &fec->eth->r_cntrl);
 }
 
@@ -1278,6 +1281,15 @@ static int fecmxc_read_rom_hwaddr(struct udevice *dev)
 	return fec_get_hwaddr(priv->dev_id, pdata->enetaddr);
 }
 
+static int fecmxc_set_promisc(struct udevice *dev, bool enable)
+{
+	struct fec_priv *priv = dev_get_priv(dev);
+
+	priv->promisc = enable;
+
+	return 0;
+}
+
 static int fecmxc_free_pkt(struct udevice *dev, uchar *packet, int length)
 {
 	if (packet)
@@ -1294,6 +1306,7 @@ static const struct eth_ops fecmxc_ops = {
 	.stop			= fecmxc_halt,
 	.write_hwaddr		= fecmxc_set_hwaddr,
 	.read_rom_hwaddr	= fecmxc_read_rom_hwaddr,
+	.set_promisc		= fecmxc_set_promisc,
 };
 
 static int device_get_phy_addr(struct fec_priv *priv, struct udevice *dev)
drivers/net/fec_mxc.h +1 -0
index 5ccde9193b..62b55ef395 100644
@@ -272,6 +272,7 @@ struct fec_priv {
 	struct clk clk_ref;
 	struct clk clk_ptp;
 	u32 clk_rate;
+	char promisc;
 };
 
 /**
Signature · 1 line
2.17.1
Reviewed-by on v1 5/9
#13 lore raw

Re: [PATCH 5/9] net: fec: add set_promisc function

Quoted 61 lines · Tim Harvey

Enabling promiscuous mode is necessary if FEC is the master of a DSA switch driver where each port has their own MAC address.

---
 drivers/net/fec_mxc.c | 13 +++++++++++++
 drivers/net/fec_mxc.h |  1 +
 2 files changed, 14 insertions(+)
drivers/net/fec_mxc.c +13 -0
index 77680491d0..db2cdaf684 100644
@@ -458,6 +458,9 @@ static void fec_reg_setup(struct fec_priv *fec)
        else if (fec->xcv_type == RMII)
                rcntrl |= FEC_RCNTRL_RMII;
 
+       if (fec->promisc)
+               rcntrl |= 0x8;
+
        writel(rcntrl, &fec->eth->r_cntrl);
 }
 
@@ -1278,6 +1281,15 @@ static int fecmxc_read_rom_hwaddr(struct udevice *dev)
        return fec_get_hwaddr(priv->dev_id, pdata->enetaddr);
 }
 
+static int fecmxc_set_promisc(struct udevice *dev, bool enable)
+{
+       struct fec_priv *priv = dev_get_priv(dev);
+
+       priv->promisc = enable;
+
+       return 0;
+}
+
 static int fecmxc_free_pkt(struct udevice *dev, uchar *packet, int length)
 {
        if (packet)
@@ -1294,6 +1306,7 @@ static const struct eth_ops fecmxc_ops = {
        .stop                   = fecmxc_halt,
        .write_hwaddr           = fecmxc_set_hwaddr,
        .read_rom_hwaddr        = fecmxc_read_rom_hwaddr,
+       .set_promisc            = fecmxc_set_promisc,
 };
 
 static int device_get_phy_addr(struct fec_priv *priv, struct udevice *dev)
drivers/net/fec_mxc.h +1 -1
index 5ccde9193b..62b55ef395 100644
@@ -272,6 +272,7 @@ struct fec_priv {
        struct clk clk_ref;
        struct clk clk_ptp;
        u32 clk_rate;
+       char promisc;
 };
 
 /**
--
Quoted 2 lines

2.17.1

#14 lore raw

PATCH 5/9 net: fec: add set_promisc function

In reply to Tim Harvey
Quoted 4 lines

Enabling promiscuous mode is necessary if FEC is the master of a DSA switch driver where each port has their own MAC address.

Applied to u-boot-imx, master, thanks !

Best regards,
Stefano Babic

Signature · 5 lines
=====================================================================
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-53 Fax: +49-8142-66989-80 Email: [email protected]
=====================================================================
author
#15 lore raw

PATCH 6/9 net: dsa: enable master promisc mode if available and needed

In reply to Tim Harvey

If ports have their own unique MAC addrs and master has a set_promisc function, call it so that packets will be received for ports.

---
 net/dsa-uclass.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
net/dsa-uclass.c +8 -1
index 7ea1cb6949..694664d81b 100644
@@ -277,8 +277,15 @@ static int dsa_port_probe(struct udevice *pdev)
 	 * has a unique MAC address specified in the environment.
 	 */
 	eth_env_get_enetaddr_by_index("eth", dev_seq(pdev), env_enetaddr);
-	if (!is_zero_ethaddr(env_enetaddr))
+	if (!is_zero_ethaddr(env_enetaddr)) {
+		/* individual port mac addrs require master to be promisc */
+		struct eth_ops *eth_ops = eth_get_ops(master);
+
+		if (eth_ops->set_promisc)
+			eth_ops->set_promisc(master, 1);
+
 		return 0;
+	}
 
 	master_pdata = dev_get_plat(master);
 	eth_pdata = dev_get_plat(pdev);
Signature · 1 line
2.17.1
Reviewed-by on v1 6/9
#16 lore raw

Re: [PATCH 6/9] net: dsa: enable master promisc mode if available and needed

Quoted 31 lines · Tim Harvey

If ports have their own unique MAC addrs and master has a set_promisc function, call it so that packets will be received for ports.

---
 net/dsa-uclass.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
net/dsa-uclass.c +8 -2
index 7ea1cb6949..694664d81b 100644
@@ -277,8 +277,15 @@ static int dsa_port_probe(struct udevice *pdev)
         * has a unique MAC address specified in the environment.
         */
        eth_env_get_enetaddr_by_index("eth", dev_seq(pdev), env_enetaddr);
-       if (!is_zero_ethaddr(env_enetaddr))
+       if (!is_zero_ethaddr(env_enetaddr)) {
+               /* individual port mac addrs require master to be promisc */
+               struct eth_ops *eth_ops = eth_get_ops(master);
+
+               if (eth_ops->set_promisc)
+                       eth_ops->set_promisc(master, 1);
+
                return 0;
+       }
 
        master_pdata = dev_get_plat(master);
        eth_pdata = dev_get_plat(pdev);
--
Quoted 2 lines

2.17.1

#17 lore raw

PATCH 6/9 net: dsa: enable master promisc mode if available and needed

In reply to Tim Harvey
Quoted 4 lines

If ports have their own unique MAC addrs and master has a set_promisc function, call it so that packets will be received for ports.

Applied to u-boot-imx, master, thanks !

Best regards,
Stefano Babic

Signature · 5 lines
=====================================================================
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-53 Fax: +49-8142-66989-80 Email: [email protected]
=====================================================================
author
#18 lore raw

PATCH 7/9 net: add support for KSZ9477/KSZ9897/KSZ9567 GbE switch

In reply to Tim Harvey

The Microchip KSZ9477/KSZ9897/KSZ9567 7-Port Gigabit Ethernet Switches support SGMII/RGMII/MII/RMII with register access via SPI, I2C, or MDIO.

This driver currently supports I2C register access but SPI or MDIO register access can be easily added at a later time.

Tagging is not implemented and instead the active port is tracked to avoid needing a tag to store port information.

This was tested with the imx8mm-venice-gw7901 board which has a KSZ9897S switch with an IMX8MM FEC MAC master connected via RGMII_ID.

---
 drivers/net/Kconfig   |   7 +
 drivers/net/Makefile  |   1 +
 drivers/net/ksz9477.c | 547 ++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 555 insertions(+)
 create mode 100644 drivers/net/ksz9477.c
drivers/net/Kconfig +7 -0
index 9fc28b149d..feeea960e2 100644
@@ -402,6 +402,13 @@ config KS8851_MLL_BASEADDR
 endif #DM_ETH
 endif #KS8851_MLL
 
+config KSZ9477
+	bool "Microchip KSZ9477 I2C controller driver"
+	depends on DM_DSA && DM_I2C
+	help
+	  This driver implements a DSA switch driver for the KSZ9477 family
+	  of GbE switches using the I2C interface.
+
 config MVGBE
 	bool "Marvell Orion5x/Kirkwood network interface support"
 	depends on ARCH_KIRKWOOD || ARCH_ORION5X
drivers/net/Makefile +1 -0
index d56baa65b2..03900ffa3a 100644
@@ -39,6 +39,7 @@ obj-$(CONFIG_FTMAC110) += ftmac110.o
 obj-$(CONFIG_FTMAC100) += ftmac100.o
 obj-$(CONFIG_GMAC_ROCKCHIP) += gmac_rockchip.o
 obj-$(CONFIG_KS8851_MLL) += ks8851_mll.o
+obj-$(CONFIG_KSZ9477) += ksz9477.o
 obj-$(CONFIG_LAN91C96) += lan91c96.o
 obj-$(CONFIG_LPC32XX_ETH) += lpc32xx_eth.o
 obj-$(CONFIG_MACB) += macb.o
drivers/net/ksz9477.c new file +547 -0

547 lines not on this page · raw message

Signature · 1 line
2.17.1
Reviewed-by on v1 7/9
#19 lore raw

Re: [PATCH 7/9] net: add support for KSZ9477/KSZ9897/KSZ9567 GbE switch

Quoted 605 lines · Tim Harvey

The Microchip KSZ9477/KSZ9897/KSZ9567 7-Port Gigabit Ethernet Switches support SGMII/RGMII/MII/RMII with register access via SPI, I2C, or MDIO.

This driver currently supports I2C register access but SPI or MDIO register access can be easily added at a later time.

Tagging is not implemented and instead the active port is tracked to avoid needing a tag to store port information.

This was tested with the imx8mm-venice-gw7901 board which has a KSZ9897S switch with an IMX8MM FEC MAC master connected via RGMII_ID.

---
 drivers/net/Kconfig   |   7 +
 drivers/net/Makefile  |   1 +
 drivers/net/ksz9477.c | 547 ++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 555 insertions(+)
 create mode 100644 drivers/net/ksz9477.c
drivers/net/Kconfig +7 -0
index 9fc28b149d..feeea960e2 100644
@@ -402,6 +402,13 @@ config KS8851_MLL_BASEADDR
 endif #DM_ETH
 endif #KS8851_MLL
 
+config KSZ9477
+       bool "Microchip KSZ9477 I2C controller driver"
+       depends on DM_DSA && DM_I2C
+       help
+         This driver implements a DSA switch driver for the KSZ9477 family
+         of GbE switches using the I2C interface.
+
 config MVGBE
        bool "Marvell Orion5x/Kirkwood network interface support"
        depends on ARCH_KIRKWOOD || ARCH_ORION5X
drivers/net/Makefile +1 -0
index d56baa65b2..03900ffa3a 100644
@@ -39,6 +39,7 @@ obj-$(CONFIG_FTMAC110) += ftmac110.o
 obj-$(CONFIG_FTMAC100) += ftmac100.o
 obj-$(CONFIG_GMAC_ROCKCHIP) += gmac_rockchip.o
 obj-$(CONFIG_KS8851_MLL) += ks8851_mll.o
+obj-$(CONFIG_KSZ9477) += ksz9477.o
 obj-$(CONFIG_LAN91C96) += lan91c96.o
 obj-$(CONFIG_LPC32XX_ETH) += lpc32xx_eth.o
 obj-$(CONFIG_MACB) += macb.o
drivers/net/ksz9477.c new file +547 -1
new file mode 100644
index 0000000000..ab32f42127
@@ -0,0 +1,547 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * (C) Copyright 2020
+ * Tim Harvey, Gateworks Corporation
+ */
+
+#include <dm.h>
+#include <dm/device_compat.h>
+#include <dm/device-internal.h>
+#include <dm/lists.h>
+#include <eth_phy.h>
+#include <linux/delay.h>
+#include <miiphy.h>
+#include <i2c.h>
+#include <net/dsa.h>
+
+#include <asm-generic/gpio.h>
+
+/* Global registers */
+
+/* Chip ID */
+#define REG_CHIP_ID0__1                        0x0000
+
+/* Operation control */
+#define REG_SW_OPERATION               0x0300
+#define SW_RESET                       BIT(1)
+#define SW_START                       BIT(0)
+
+/* Port Specific Registers */
+#define PORT_CTRL_ADDR(port, addr) ((addr) | (((port) + 1) << 12))
+
+/* Port Control */
+#define REG_PORT_XMII_CTRL_1           0x0301
+#define PORT_MII_NOT_1GBIT             BIT(6)
+#define PORT_MII_SEL_EDGE              BIT(5)
+#define PORT_RGMII_ID_IG_ENABLE                BIT(4)
+#define PORT_RGMII_ID_EG_ENABLE                BIT(3)
+#define PORT_MII_MAC_MODE              BIT(2)
+#define PORT_MII_SEL_M                 0x3
+#define PORT_RGMII_SEL                 0x0
+#define PORT_RMII_SEL                  0x1
+#define PORT_GMII_SEL                  0x2
+#define PORT_MII_SEL                   0x3
+
+/* Port MSTP State Register */
+#define REG_PORT_MSTP_STATE            0x0b04
+#define PORT_TX_ENABLE                 BIT(2)
+#define PORT_RX_ENABLE                 BIT(1)
+#define PORT_LEARN_DISABLE             BIT(0)
+
+/* MMD */
+#define REG_PORT_PHY_MMD_SETUP         0x011A
+#define PORT_MMD_OP_MODE_M             0x3
+#define PORT_MMD_OP_MODE_S             14
+#define PORT_MMD_OP_INDEX              0
+#define PORT_MMD_OP_DATA_NO_INCR       1
+#define PORT_MMD_OP_DATA_INCR_RW       2
+#define PORT_MMD_OP_DATA_INCR_W                3
+#define PORT_MMD_DEVICE_ID_M           0x1F
+#define MMD_SETUP(mode, dev)           (((u16)(mode) << PORT_MMD_OP_MODE_S) | (dev))
+#define REG_PORT_PHY_MMD_INDEX_DATA    0x011C
+
+struct ksz_dsa_priv {
+       struct udevice *dev;
+       int active_port;
+};
+
+static inline int ksz_read8(struct udevice *dev, u32 reg, u8 *val)
+{
+       int ret = dm_i2c_read(dev, reg, val, 1);
+
+       dev_dbg(dev, "%s 0x%04x<<0x%02x\n", __func__, reg, *val);
+
+       return ret;
+}
+
+static inline int ksz_pread8(struct udevice *dev, int port, int reg, u8 *val)
+{
+       return ksz_read8(dev, PORT_CTRL_ADDR(port, reg), val);
+}
+
+static inline int ksz_write8(struct udevice *dev, u32 reg, u8 val)
+{
+       dev_dbg(dev, "%s 0x%04x>>0x%02x\n", __func__, reg, val);
+       return dm_i2c_write(dev, reg, &val, 1);
+}
+
+static inline int ksz_pwrite8(struct udevice *dev, int port, int reg, u8 val)
+{
+       return ksz_write8(dev, PORT_CTRL_ADDR(port, reg), val);
+}
+
+static inline int ksz_write16(struct udevice *dev, u32 reg, u16 val)
+{
+       u8 buf[2];
+
+       buf[1] = val & 0xff;
+       buf[0] = val >> 8;
+       dev_dbg(dev, "%s 0x%04x>>0x%04x\n", __func__, reg, val);
+
+       return dm_i2c_write(dev, reg, buf, 2);
+}
+
+static inline int ksz_pwrite16(struct udevice *dev, int port, int reg, u16 val)
+{
+       return ksz_write16(dev, PORT_CTRL_ADDR(port, reg), val);
+}
+
+static inline int ksz_read16(struct udevice *dev, u32 reg, u16 *val)
+{
+       u8 buf[2];
+       int ret;
+
+       ret = dm_i2c_read(dev, reg, buf, 2);
+       *val = (buf[0] << 8) | buf[1];
+       dev_dbg(dev, "%s 0x%04x<<0x%04x\n", __func__, reg, *val);
+
+       return ret;
+}
+
+static inline int ksz_pread16(struct udevice *dev, int port, int reg, u16 *val)
+{
+       return ksz_read16(dev, PORT_CTRL_ADDR(port, reg), val);
+}
+
+static inline int ksz_read32(struct udevice *dev, u32 reg, u32 *val)
+{
+       return dm_i2c_read(dev, reg, (u8 *)val, 4);
+}
+
+static inline int ksz_pread32(struct udevice *dev, int port, int reg, u32 *val)
+{
+       return ksz_read32(dev, PORT_CTRL_ADDR(port, reg), val);
+}
+
+static inline int ksz_write32(struct udevice *dev, u32 reg, u32 val)
+{
+       u8 buf[4];
+
+       buf[3] = val & 0xff;
+       buf[2] = (val >> 24) & 0xff;
+       buf[1] = (val >> 16) & 0xff;
+       buf[0] = (val >> 8) & 0xff;
+       dev_dbg(dev, "%s 0x%04x>>0x%04x\n", __func__, reg, val);
+
+       return dm_i2c_write(dev, reg, buf, 4);
+}
+
+static inline int ksz_pwrite32(struct udevice *dev, int port, int reg, u32 val)
+{
+       return ksz_write32(dev, PORT_CTRL_ADDR(port, reg), val);
+}
+
+static __maybe_unused void ksz_port_mmd_read(struct udevice *dev, int port,
+                                            u8 addr, u16 reg, u16 *val)
+{
+       ksz_pwrite16(dev, port, REG_PORT_PHY_MMD_SETUP, MMD_SETUP(PORT_MMD_OP_INDEX, addr));
+       ksz_pwrite16(dev, port, REG_PORT_PHY_MMD_INDEX_DATA, reg);
+       ksz_pwrite16(dev, port, REG_PORT_PHY_MMD_SETUP, MMD_SETUP(PORT_MMD_OP_DATA_NO_INCR, addr));
+       ksz_pread16(dev, port, REG_PORT_PHY_MMD_INDEX_DATA, val);
+       dev_dbg(dev, "%s  P%d 0x%02x:0x%04x<<0x%04x\n", __func__, port + 1, addr, reg, *val);
+}
+
+static void ksz_port_mmd_write(struct udevice *dev, int port, u8 addr, u16 reg, u16 val)
+{
+       dev_dbg(dev, "%s P%d 0x%02x:0x%04x>>0x%04x\n", __func__, port + 1, addr, addr, val);
+       ksz_pwrite16(dev, port, REG_PORT_PHY_MMD_SETUP, MMD_SETUP(PORT_MMD_OP_INDEX, addr));
+       ksz_pwrite16(dev, port, REG_PORT_PHY_MMD_INDEX_DATA, addr);
+       ksz_pwrite16(dev, port, REG_PORT_PHY_MMD_SETUP, MMD_SETUP(PORT_MMD_OP_DATA_NO_INCR, addr));
+       ksz_pwrite16(dev, port, REG_PORT_PHY_MMD_INDEX_DATA, val);
+}
+
+/* Apply PHY settings to address errata listed in KSZ9477, KSZ9897, KSZ9896, KSZ9567
+ * Silicon Errata and Data Sheet Clarification documents
+ */
+static void ksz_phy_errata_setup(struct udevice *dev, int port)
+{
+       dev_dbg(dev, "%s P%d\n", __func__, port + 1);
+
+       /* Register settings are needed to improve PHY receive performance */
+       ksz_port_mmd_write(dev, port, 0x01, 0x6f, 0xdd0b);
+       ksz_port_mmd_write(dev, port, 0x01, 0x8f, 0x6032);
+       ksz_port_mmd_write(dev, port, 0x01, 0x9d, 0x248c);
+       ksz_port_mmd_write(dev, port, 0x01, 0x75, 0x0060);
+       ksz_port_mmd_write(dev, port, 0x01, 0xd3, 0x7777);
+       ksz_port_mmd_write(dev, port, 0x1c, 0x06, 0x3008);
+       ksz_port_mmd_write(dev, port, 0x1c, 0x08, 0x2001);
+
+       /* Transmit waveform amplitude can be improved (1000BASE-T, 100BASE-TX, 10BASE-Te) */
+       ksz_port_mmd_write(dev, port, 0x1c, 0x04, 0x00d0);
+
+       /* Energy Efficient Ethernet (EEE) feature select must be manually disabled */
+       ksz_port_mmd_write(dev, port, 0x07, 0x3c, 0x0000);
+
+       /* Register settings are required to meet data sheet supply current specifications */
+       ksz_port_mmd_write(dev, port, 0x1c, 0x13, 0x6eff);
+       ksz_port_mmd_write(dev, port, 0x1c, 0x14, 0xe6ff);
+       ksz_port_mmd_write(dev, port, 0x1c, 0x15, 0x6eff);
+       ksz_port_mmd_write(dev, port, 0x1c, 0x16, 0xe6ff);
+       ksz_port_mmd_write(dev, port, 0x1c, 0x17, 0x00ff);
+       ksz_port_mmd_write(dev, port, 0x1c, 0x18, 0x43ff);
+       ksz_port_mmd_write(dev, port, 0x1c, 0x19, 0xc3ff);
+       ksz_port_mmd_write(dev, port, 0x1c, 0x1a, 0x6fff);
+       ksz_port_mmd_write(dev, port, 0x1c, 0x1b, 0x07ff);
+       ksz_port_mmd_write(dev, port, 0x1c, 0x1c, 0x0fff);
+       ksz_port_mmd_write(dev, port, 0x1c, 0x1d, 0xe7ff);
+       ksz_port_mmd_write(dev, port, 0x1c, 0x1e, 0xefff);
+       ksz_port_mmd_write(dev, port, 0x1c, 0x20, 0xeeee);
+}
+
+/*
+ * mii bus driver
+ */
+#define KSZ_MDIO_CHILD_DRV_NAME        "ksz_mdio"
+
+struct ksz_mdio_priv {
+       struct ksz_dsa_priv *ksz;
+};
+
+static int dm_ksz_mdio_read(struct udevice *dev, int addr, int devad, int reg)
+{
+       struct ksz_mdio_priv *priv = dev_get_priv(dev);
+       struct ksz_dsa_priv *ksz = priv->ksz;
+       u16 val = 0xffff;
+
+       ksz_pread16(ksz->dev, addr, 0x100 + (reg << 1), &val);
+       dev_dbg(ksz->dev, "%s P%d reg=0x%04x:0x%04x<<0x%04x\n", __func__,
+               addr + 1, reg, 0x100 + (reg << 1), val);
+
+       return val;
+};
+
+static int dm_ksz_mdio_write(struct udevice *dev, int addr, int devad, int reg, u16 val)
+{
+       struct ksz_mdio_priv *priv = dev_get_priv(dev);
+       struct ksz_dsa_priv *ksz = priv->ksz;
+
+       dev_dbg(ksz->dev, "%s P%d reg=0x%04x:%04x>>0x%04x\n",
+               __func__, addr + 1, reg, 0x100 + (reg << 1), val);
+       ksz_pwrite16(ksz->dev, addr, 0x100 + (reg << 1), val);
+
+       return 0;
+}
+
+static const struct mdio_ops ksz_mdio_ops = {
+       .read = dm_ksz_mdio_read,
+       .write = dm_ksz_mdio_write,
+};
+
+static int ksz_mdio_bind(struct udevice *dev)
+{
+       char name[16];
+       static int num_devices;
+
+       dev_dbg(dev, "%s\n", __func__);
+       sprintf(name, "ksz-mdio-%d", num_devices++);
+       device_set_name(dev, name);
+
+       return 0;
+}
+
+static int ksz_mdio_probe(struct udevice *dev)
+{
+       struct ksz_mdio_priv *priv = dev_get_priv(dev);
+
+       dev_dbg(dev, "%s\n", __func__);
+       priv->ksz = dev_get_parent_priv(dev->parent);
+
+       return 0;
+}
+
+static const struct udevice_id ksz_mdio_ids[] = {
+       { .compatible = "microchip,ksz-mdio" },
+       { }
+};
+
+U_BOOT_DRIVER(ksz_mdio) = {
+       .name           = KSZ_MDIO_CHILD_DRV_NAME,
+       .id             = UCLASS_MDIO,
+       .of_match       = ksz_mdio_ids,
+       .bind           = ksz_mdio_bind,
+       .probe          = ksz_mdio_probe,
+       .ops            = &ksz_mdio_ops,
+       .priv_auto      = sizeof(struct ksz_mdio_priv),
+       .plat_auto      = sizeof(struct mdio_perdev_priv),
+};
+
+static int ksz_port_setup(struct udevice *dev, int port,
+                         phy_interface_t interface)
+{
+       struct dsa_pdata *pdata = dev_get_uclass_plat(dev);
+       u8 data8;
+
+       dev_dbg(dev, "%s P%d %s\n", __func__, port + 1,
+               (port == pdata->cpu_port) ? "cpu" : "");
+
+       if (port != pdata->cpu_port) {
+               /* phy port: config errata and leds */
+               ksz_phy_errata_setup(dev, port);
+       } else {
+               /* cpu port: configure MAC interface mode */
+               ksz_pread8(dev, port, REG_PORT_XMII_CTRL_1, &data8);
+               dev_dbg(dev, "%s P%d cpu interface %s\n", __func__, port + 1,
+                       phy_string_for_interface(interface));
+               switch (interface) {
+               case PHY_INTERFACE_MODE_MII:
+                       data8 &= ~PORT_MII_SEL_M;
+                       data8 |= PORT_MII_SEL;
+                       data8 |= PORT_MII_NOT_1GBIT;
+                       break;
+               case PHY_INTERFACE_MODE_RMII:
+                       data8 &= ~PORT_MII_SEL_M;
+                       data8 |= PORT_RMII_SEL;
+                       data8 |= PORT_MII_NOT_1GBIT;
+                       break;
+               case PHY_INTERFACE_MODE_GMII:
+                       data8 &= ~PORT_MII_SEL_M;
+                       data8 |= PORT_GMII_SEL;
+                       data8 &= ~PORT_MII_NOT_1GBIT;
+                       break;
+               default:
+                       data8 &= ~PORT_MII_SEL_M;
+                       data8 |= PORT_RGMII_SEL;
+                       data8 &= ~PORT_MII_NOT_1GBIT;
+                       data8 &= ~PORT_RGMII_ID_IG_ENABLE;
+                       data8 &= ~PORT_RGMII_ID_EG_ENABLE;
+                       if (interface == PHY_INTERFACE_MODE_RGMII_ID ||
+                           interface == PHY_INTERFACE_MODE_RGMII_RXID)
+                               data8 |= PORT_RGMII_ID_IG_ENABLE;
+                       if (interface == PHY_INTERFACE_MODE_RGMII_ID ||
+                           interface == PHY_INTERFACE_MODE_RGMII_TXID)
+                               data8 |= PORT_RGMII_ID_EG_ENABLE;
+                       break;
+               }
+               ksz_write8(dev, PORT_CTRL_ADDR(port, REG_PORT_XMII_CTRL_1), data8);
+       }
+
+       return 0;
+}
+
+static int ksz_port_enable(struct udevice *dev, int port, struct phy_device *phy)
+{
+       struct dsa_pdata *pdata = dev_get_uclass_plat(dev);
+       struct ksz_dsa_priv *priv = dev_get_priv(dev);
+       int supported = PHY_GBIT_FEATURES;
+       u8 data8;
+       int ret;
+
+       dev_dbg(dev, "%s P%d 0x%x %s\n", __func__, port + 1, phy->phy_id,
+               phy_string_for_interface(phy->interface));
+
+       /* setup this port */
+       ret = ksz_port_setup(dev, port, phy->interface);
+       if (ret) {
+               dev_err(dev, "port setup failed: %d\n", ret);
+               return ret;
+       }
+
+       /* enable port forwarding for this port */
+       ksz_pread8(priv->dev, port, REG_PORT_MSTP_STATE, &data8);
+       data8 &= ~(PORT_TX_ENABLE | PORT_RX_ENABLE | PORT_LEARN_DISABLE);
+       data8 |= (PORT_TX_ENABLE | PORT_RX_ENABLE);
+       ksz_pwrite8(priv->dev, port, REG_PORT_MSTP_STATE, data8);
+
+       /* if cpu master we are done */
+       if (port == pdata->cpu_port)
+               return 0;
+
+       /* configure phy */
+       phy->supported &= supported;
+       phy->advertising &= supported;
+       ret = phy_config(phy);
+       if (ret)
+               return ret;
+
+       ret = phy_startup(phy);
+       if (ret)
+               return ret;
+
+       /* start switch */
+       ksz_read8(priv->dev, REG_SW_OPERATION, &data8);
+       data8 |= SW_START;
+       ksz_write8(priv->dev, REG_SW_OPERATION, data8);
+
+       /* keep track of current enabled non-cpu port */
+       priv->active_port = port;
+
+       return 0;
+}
+
+static void ksz_port_disable(struct udevice *dev, int port, struct phy_device *phy)
+{
+       struct dsa_pdata *pdata = dev_get_uclass_plat(dev);
+       struct ksz_dsa_priv *priv = dev_get_priv(dev);
+       u8 data8;
+
+       dev_dbg(dev, "%s P%d 0x%x\n", __func__, port + 1, phy->phy_id);
+
+       /* can't disable CPU port without re-configuring/re-starting switch */
+       if (port == pdata->cpu_port)
+               return;
+
+       /* disable port */
+       ksz_pread8(priv->dev, port, REG_PORT_MSTP_STATE, &data8);
+       data8 &= ~(PORT_TX_ENABLE | PORT_RX_ENABLE | PORT_LEARN_DISABLE);
+       data8 |= PORT_LEARN_DISABLE;
+       ksz_pwrite8(priv->dev, port, REG_PORT_MSTP_STATE, data8);
+
+       /*
+        * we don't call phy_shutdown here to avoid waiting next time we use
+        * the port, but the downside is that remote side will think we're
+        * actively processing traffic although we are not.
+        */
+}
+
+static int ksz_xmit(struct udevice *dev, int port, void *packet, int length)
+{
+       dev_dbg(dev, "%s P%d %d\n", __func__, port + 1, length);
+
+       return 0;
+}
+
+static int ksz_recv(struct udevice *dev, int *port, void *packet, int length)
+{
+       struct ksz_dsa_priv *priv = dev_get_priv(dev);
+
+       dev_dbg(dev, "%s P%d %d\n", __func__, priv->active_port + 1, length);
+       *port = priv->active_port;
+
+       return 0;
+};
+
+static const struct dsa_ops ksz_dsa_ops = {
+       .port_enable = ksz_port_enable,
+       .port_disable = ksz_port_disable,
+       .xmit = ksz_xmit,
+       .rcv = ksz_recv,
+};
+
+static int ksz_probe_mdio(struct udevice *dev)
+{
+       ofnode node, mdios;
+       int ret;
+
+       mdios = dev_read_subnode(dev, "mdios");
+       if (ofnode_valid(mdios)) {
+               ofnode_for_each_subnode(node, mdios) {
+                       const char *name = ofnode_get_name(node);
+                       struct udevice *pdev;
+
+                       ret = device_bind_driver_to_node(dev,
+                                                        KSZ_MDIO_CHILD_DRV_NAME,
+                                                        name, node, &pdev);
+                       if (ret)
+                               dev_err(dev, "failed to probe %s: %d\n", name, ret);
+               }
+       }
+
+       return 0;
+}
+
+/*
+ * I2C driver
+ */
+static int ksz_i2c_probe(struct udevice *dev)
+{
+       struct dsa_pdata *pdata = dev_get_uclass_plat(dev);
+       struct ksz_dsa_priv *priv = dev_get_priv(dev);
+       struct udevice *master = dsa_get_master(dev);
+       int i, ret;
+       u8 data8;
+       u32 id;
+
+       if (!master)
+               return -ENODEV;
+
+       dev_dbg(dev, "%s %s master:%s\n", __func__, dev->name, master->name);
+       dev_set_parent_priv(dev, priv);
+
+       ret = i2c_set_chip_offset_len(dev, 2);
+       if (ret) {
+               printf("i2c_set_chip_offset_len failed: %d\n", ret);
+               return ret;
+       }
+
+       /* default config */
+       priv->dev = dev;
+
+       /* chip level reset */
+       ksz_read8(priv->dev, REG_SW_OPERATION, &data8);
+       data8 |= SW_RESET;
+       ksz_write8(priv->dev, REG_SW_OPERATION, data8);
+
+       /* read chip id */
+       ret = ksz_read32(dev, REG_CHIP_ID0__1, &id);
+       if (ret)
+               return ret;
+       id = __swab32(id);
+       dev_dbg(dev, "%s id=0x%08x\n", __func__, id);
+       switch (id & 0xffffff00) {
+       case 0x00947700:
+               puts("KSZ9477S: ");
+               break;
+       case 0x00956700:
+               puts("KSZ9567R: ");
+               break;
+       case 0x00989700:
+               puts("KSZ9897S: ");
+               break;
+       default:
+               dev_err(dev, "invalid chip id: 0x%08x\n", id);
+               return -EINVAL;
+       }
+
+       /* probe mdio bus */
+       ret = ksz_probe_mdio(dev);
+       if (ret)
+               return ret;
+
+       /* disable ports by default */
+       for (i = 0; i < pdata->num_ports; i++) {
+               ksz_pread8(priv->dev, i, REG_PORT_MSTP_STATE, &data8);
+               data8 &= ~(PORT_TX_ENABLE | PORT_RX_ENABLE | PORT_LEARN_DISABLE);
+               ksz_pwrite8(priv->dev, i, REG_PORT_MSTP_STATE, data8);
+       }
+
+       dsa_set_tagging(dev, 0, 0);
+
+       return 0;
+};
+
+static const struct udevice_id ksz_i2c_ids[] = {
+       { .compatible = "microchip,ksz9897" },
+       { .compatible = "microchip,ksz9477" },
+       { .compatible = "microchip,ksz9567" },
+       { }
+};
+
+U_BOOT_DRIVER(ksz) = {
+       .name           = "ksz-switch",
+       .id             = UCLASS_DSA,
+       .of_match       = ksz_i2c_ids,
+       .probe          = ksz_i2c_probe,
+       .ops            = &ksz_dsa_ops,
+       .priv_auto      = sizeof(struct ksz_dsa_priv),
+};
+
--
Quoted 2 lines

2.17.1

#20 lore raw

PATCH 7/9 net: add support for KSZ9477/KSZ9897/KSZ9567 GbE switch

In reply to Tim Harvey
Quoted 10 lines

The Microchip KSZ9477/KSZ9897/KSZ9567 7-Port Gigabit Ethernet Switches support SGMII/RGMII/MII/RMII with register access via SPI, I2C, or MDIO. This driver currently supports I2C register access but SPI or MDIO register access can be easily added at a later time.
Tagging is not implemented and instead the active port is tracked to avoid needing a tag to store port information.
This was tested with the imx8mm-venice-gw7901 board which has a KSZ9897S switch with an IMX8MM FEC MAC master connected via RGMII_ID.

Applied to u-boot-imx, master, thanks !

Best regards,
Stefano Babic

Signature · 5 lines
=====================================================================
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-53 Fax: +49-8142-66989-80 Email: [email protected]
=====================================================================
author
#21 lore raw

PATCH 8/9 arm: dts: imx8mm-venice-gw7901.dts: fix dsa switch configuration

In reply to Tim Harvey

Fix the dsa switch config:
- remove the unnecessary phy-mode from the switch itself
- added the necessary fixed-link node to the non-cpu ports required for U-Boot DSA

---
 arch/arm/dts/imx8mm-venice-gw7901.dts | 37 ++++++++++++++++++++++++++-
 1 file changed, 36 insertions(+), 1 deletion(-)
arch/arm/dts/imx8mm-venice-gw7901.dts +36 -1
index 0216facb2a..124e1e4e70 100644
@@ -577,7 +577,6 @@
 		pinctrl-0 = <&pinctrl_ksz>;
 		interrupt-parent = <&gpio4>;
 		interrupts = <18 IRQ_TYPE_EDGE_FALLING>;
-		phy-mode = "rgmii-id";
 
 		ports {
 			#address-cells = <1>;
@@ -587,24 +586,32 @@
 				reg = <0>;
 				label = "lan1";
 				local-mac-address = [00 00 00 00 00 00];
+				phy-handle = <&sw_phy0>;
+				phy-mode = "internal";
 			};
 
 			lan2: port@1 {
 				reg = <1>;
 				label = "lan2";
 				local-mac-address = [00 00 00 00 00 00];
+				phy-handle = <&sw_phy1>;
+				phy-mode = "internal";
 			};
 
 			lan3: port@2 {
 				reg = <2>;
 				label = "lan3";
 				local-mac-address = [00 00 00 00 00 00];
+				phy-handle = <&sw_phy2>;
+				phy-mode = "internal";
 			};
 
 			lan4: port@3 {
 				reg = <3>;
 				label = "lan4";
 				local-mac-address = [00 00 00 00 00 00];
+				phy-handle = <&sw_phy3>;
+				phy-mode = "internal";
 			};
 
 			port@5 {
@@ -619,6 +626,34 @@
 				};
 			};
 		};
+
+		mdios {
+			#address-cells = <1>;
+			#size-cells = <0>;
+
+			mdio@0 {
+				reg = <0>;
+				compatible = "microchip,ksz-mdio";
+				#address-cells = <1>;
+				#size-cells = <0>;
+
+				sw_phy0: ethernet-phy@0 {
+					reg = <0x0>;
+				};
+
+				sw_phy1: ethernet-phy@1 {
+					reg = <0x1>;
+				};
+
+				sw_phy2: ethernet-phy@2 {
+					reg = <0x2>;
+				};
+
+				sw_phy3: ethernet-phy@3 {
+					reg = <0x3>;
+				};
+			};
+		};
 	};
 
 	crypto@60 {
Signature · 1 line
2.17.1
#22 lore raw

PATCH 8/9 arm: dts: imx8mm-venice-gw7901.dts: fix dsa switch configuration

Quoted 5 lines

Fix the dsa switch config:
- remove the unnecessary phy-mode from the switch itself
- added the necessary fixed-link node to the non-cpu ports required for U-Boot DSA

Applied to u-boot-imx, master, thanks !

Best regards,
Stefano Babic

Signature · 5 lines
=====================================================================
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-53 Fax: +49-8142-66989-80 Email: [email protected]
=====================================================================
author
#23 lore raw

PATCH 9/9 configs: imx8mm_venice_defconfig: add support for gbe switch

In reply to Tim Harvey

The imx8mm-venice-gw7901 board has an I2C connected KSZ9897S GbE switch with an IMX8MM FEC MAC master connected via RGMII_ID.

---
 configs/imx8mm_venice_defconfig | 4 ++++
 1 file changed, 4 insertions(+)
configs/imx8mm_venice_defconfig +4 -0
index 3961db00cb..49affccec4 100644
@@ -87,9 +87,13 @@ CONFIG_SPL_MMC_HS400_SUPPORT=y
 CONFIG_FSL_USDHC=y
 CONFIG_PHYLIB=y
 CONFIG_PHY_TI_DP83867=y
+CONFIG_PHY_FIXED=y
 CONFIG_DM_ETH=y
+CONFIG_DM_MDIO=y
+CONFIG_DM_DSA=y
 CONFIG_PHY_GIGE=y
 CONFIG_FEC_MXC=y
+CONFIG_KSZ9477=y
 CONFIG_MII=y
 CONFIG_PINCTRL=y
 CONFIG_SPL_PINCTRL=y
Signature · 1 line
2.17.1
#24 lore raw

PATCH 9/9 configs: imx8mm_venice_defconfig: add support for gbe switch

Quoted 3 lines

The imx8mm-venice-gw7901 board has an I2C connected KSZ9897S GbE switch with an IMX8MM FEC MAC master connected via RGMII_ID.

Applied to u-boot-imx, master, thanks !

Best regards,
Stefano Babic

Signature · 5 lines
=====================================================================
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-53 Fax: +49-8142-66989-80 Email: [email protected]
=====================================================================