[v1,4/5] net: hpe: Add GXP UMAC Driver

Message ID 20230721212044.59666-5-nick.hawkins@hpe.com
State New
Headers
Series ARM: Add GXP UMAC Support |

Commit Message

Hawkins, Nick July 21, 2023, 9:20 p.m. UTC
  From: Nick Hawkins <nick.hawkins@hpe.com>

The GXP contains two Ethernet MACs that can be connected externally
to several physical devices. From an external interface perspective
the BMC provides two SERDES interface connections capable of either
SGMII or 1000Base-X operation. The BMC also provides a RMII interface
for sideband connections to external Ethernet controllers.

The primary MAC (umac0) can be mapped to either SGMII/1000-BaseX
SERDES interface.  The secondary MAC (umac1) can be mapped to only
the second SGMII/1000-Base X Serdes interface or it can be mapped for
RMII sideband.

The MDIO(mdio0) interface from the primary MAC (umac0) is used for
external PHY status and configuration. The MDIO(mdio1) interface from
the secondary MAC (umac1) is routed to the SGMII/100Base-X IP blocks
on the two SERDES interface connections.

Signed-off-by: Nick Hawkins <nick.hawkins@hpe.com>
---
 drivers/net/ethernet/hpe/Kconfig    |  15 +
 drivers/net/ethernet/hpe/Makefile   |   1 +
 drivers/net/ethernet/hpe/gxp-umac.c | 911 ++++++++++++++++++++++++++++
 drivers/net/ethernet/hpe/gxp-umac.h |  89 +++
 4 files changed, 1016 insertions(+)
 create mode 100644 drivers/net/ethernet/hpe/gxp-umac.c
 create mode 100644 drivers/net/ethernet/hpe/gxp-umac.h
  

Comments

Christophe JAILLET July 21, 2023, 10 p.m. UTC | #1
Le 21/07/2023 à 23:20, nick.hawkins-ZPxbGqLxI0U@public.gmane.org a écrit :
> From: Nick Hawkins <nick.hawkins-ZPxbGqLxI0U@public.gmane.org>
> 
> The GXP contains two Ethernet MACs that can be connected externally
> to several physical devices. From an external interface perspective
> the BMC provides two SERDES interface connections capable of either
> SGMII or 1000Base-X operation. The BMC also provides a RMII interface
> for sideband connections to external Ethernet controllers.
> 
> The primary MAC (umac0) can be mapped to either SGMII/1000-BaseX
> SERDES interface.  The secondary MAC (umac1) can be mapped to only
> the second SGMII/1000-Base X Serdes interface or it can be mapped for
> RMII sideband.
> 
> The MDIO(mdio0) interface from the primary MAC (umac0) is used for
> external PHY status and configuration. The MDIO(mdio1) interface from
> the secondary MAC (umac1) is routed to the SGMII/100Base-X IP blocks
> on the two SERDES interface connections.
> 
> Signed-off-by: Nick Hawkins <nick.hawkins-ZPxbGqLxI0U@public.gmane.org>
> ---

[...]

> +static int umac_poll(struct napi_struct *napi, int budget)
> +{
> +	struct umac_priv *umac = container_of(napi, struct umac_priv, napi);
> +	struct net_device *ndev = umac->ndev;
> +	unsigned int value;
> +	int rx_done = 0;

No need to init.

> +
> +	umac_tx_done(ndev);
> +
> +	rx_done = umac_rx(ndev, budget);
> +
> +	if (rx_done < budget) {
> +		napi_complete_done(napi, rx_done);
> +		/* clear rx interrupt */
> +		value = readl(umac->base + UMAC_INTERRUPT);
> +		value &= ~UMAC_RX_INT;
> +		writel(value, umac->base + UMAC_INTERRUPT);
> +
> +		/* enable interrupt */
> +		umac_irq_enable(umac);
> +	}
> +
> +	return rx_done;
> +}

[...]

> +static int umac_setup_phy(struct net_device *ndev)
> +{
> +	struct umac_priv *umac = netdev_priv(ndev);
> +	struct platform_device *pdev = umac->pdev;
> +	struct device_node *phy_handle;
> +	phy_interface_t interface;
> +	struct device_node *eth_ports_np;
> +	struct device_node *port_np;
> +	int ret;
> +	int err;

Could there be only one ret ou err variable?

> +	int i;
> +
> +	/* Get child node ethernet-ports. */
> +	eth_ports_np = of_get_child_by_name(pdev->dev.of_node, "ethernet-ports");
> +	if (!eth_ports_np) {
> +		dev_err(&pdev->dev, "No ethernet-ports child node found!\n");
> +		return -ENODEV;
> +	}
> +
> +	for (i = 0; i < NUMBER_OF_PORTS; i++) {
> +		/* Get port@i of node ethernet-ports */
> +		port_np = gxp_umac_get_eth_child_node(eth_ports_np, i);
> +		if (!port_np)
> +			break;
> +
> +		if (i == INTERNAL_PORT) {
> +			phy_handle = of_parse_phandle(port_np, "phy-handle", 0);
> +			if (phy_handle) {
> +				umac->int_phy_dev = of_phy_find_device(phy_handle);
> +				if (!umac->int_phy_dev)
> +					return -ENODEV;
> +
> +				umac_int_phy_init(umac);
> +			} else {
> +				return dev_err_probe(&pdev->dev, PTR_ERR(phy_handle),
> +						     "Failed to map phy-handle for port %d", i);
> +			}
> +		}
> +
> +		if (i == EXTERNAL_PORT) {
> +			phy_handle = of_parse_phandle(port_np, "phy-handle", 0);
> +			if (phy_handle) {
> +				/* register the phy board fixup */
> +				ret = phy_register_fixup_for_uid(0x01410dd1, 0xffffffff,
> +								 umac_phy_fixup);
> +				if (ret)
> +					dev_err(&pdev->dev, "cannot register phy board fixup\n");
> +
> +				err = of_get_phy_mode(phy_handle, &interface);
> +				if (err)
> +					interface = PHY_INTERFACE_MODE_NA;
> +
> +				umac->phy_dev = of_phy_connect(ndev, phy_handle,
> +							       &umac_adjust_link,
> +							       0, interface);
> +
> +				if (!umac->phy_dev)
> +					return -ENODEV;
> +
> +				/* If the specified phy-handle has a fixed-link declaration, use the
> +				 * fixed-link properties to set the configuration for the PHY
> +				 */
> +				if (of_phy_is_fixed_link(phy_handle)) {
> +					struct device_node *fixed_link_node =
> +						of_get_child_by_name(phy_handle,
> +								     "fixed-link");
> +
> +					if (of_property_read_u32(fixed_link_node, "speed",
> +								 &umac->phy_dev->speed)) {
> +						netdev_err(ndev, "Invalid fixed-link specified.\n");
> +						return -EINVAL;
> +					}
> +					umac->phy_dev->duplex =
> +						of_property_read_bool(fixed_link_node,
> +								      "full-duplex");
> +					umac->phy_dev->pause =
> +						of_property_read_bool(fixed_link_node,
> +								      "pause");
> +					umac->phy_dev->asym_pause =
> +						of_property_read_bool(fixed_link_node,
> +								      "asym-pause");
> +					umac->phy_dev->autoneg = AUTONEG_DISABLE;
> +					__clear_bit(ETHTOOL_LINK_MODE_Autoneg_BIT,
> +						    umac->phy_dev->advertising);
> +				}
> +			} else {
> +				return dev_err_probe(&pdev->dev, PTR_ERR(phy_handle),
> +						     "Failed to map phy-handle for port %d", i);
> +			}
> +		}
> +	}
> +
> +	return 0;
> +}
> +
> +static int umac_probe(struct platform_device *pdev)
> +{
> +	struct umac_priv *umac;
> +	struct net_device *ndev;
> +	struct resource *res;
> +	int ret = 0;
> +
> +	ndev = alloc_etherdev(sizeof(*umac));

devm_alloc_etherdev()?
This saves a lot of free_netdev() below.

> +	if (!ndev)
> +		return -ENOMEM;
> +
> +	SET_NETDEV_DEV(ndev, &pdev->dev);
> +
> +	umac = netdev_priv(ndev);
> +	umac->pdev = pdev;
> +	umac->ndev = ndev;
> +
> +	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> +	if (!res) {
> +		netdev_err(ndev, "failed to get I/O memory\n");
> +		free_netdev(ndev);
> +		return -ENXIO;
> +	}
> +
> +	umac->base = devm_ioremap_resource(&pdev->dev, res);
> +	if (!umac->base) {
> +		netdev_err(ndev, "failed to remap I/O memory\n");
> +		free_netdev(ndev);
> +		return -EBUSY;
> +	}
> +
> +	ndev->irq = platform_get_irq(pdev, 0);
> +	if (ndev->irq < 0) {
> +		netdev_err(ndev, "failed to get irq\n");
> +		free_netdev(ndev);
> +		return -ENXIO;
> +	}
> +
> +	platform_set_drvdata(pdev, ndev);
> +
> +	ndev->netdev_ops = &umac_netdev_ops;
> +	ndev->ethtool_ops = &umac_ethtool_ops;
> +
> +	umac_init_mac_address(ndev);
> +	umac_channel_disable(umac);
> +	ret = umac_setup_phy(ndev);
> +	if (ret != 0) {
> +		netdev_err(ndev, "failed to setup phy ret=%d\n", ret);

free_netdev() missing. (should alloc_etherdev() be left as-is)

> +		return -ENODEV;

Does it makes sense to return ret as-is?

> +	}
> +
> +	umac->use_ncsi = false;
> +	if (of_get_property(pdev->dev.of_node, "use-ncsi", NULL)) {
> +		if (!IS_ENABLED(CONFIG_NET_NCSI)) {
> +			netdev_err(ndev, "NCSI stack not enabled\n");
> +			free_netdev(ndev);
> +			return 0;

So register_netdev() below could not be called.
In such a case unregister_netdev() would be unbalanced in the remove 
function. Using devm_register_netdev() below would avoid it.

> +		}
> +
> +		dev_info(&pdev->dev, "Using NCSI interface\n");
> +		umac->use_ncsi = true;
> +		umac->ncsidev = ncsi_register_dev(ndev, umac_ncsi_handler);
> +		if (!umac->ncsidev) {
> +			free_netdev(ndev);
> +			return -ENODEV;
> +		}
> +	}
> +
> +	netif_napi_add(ndev, &umac->napi, umac_poll);
> +	ret = register_netdev(ndev);

devm_register_netdev()

> +	if (ret != 0) {
> +		netdev_err(ndev, "failed to register UMAC ret=%d\n", ret);
> +		netif_napi_del(&umac->napi);

(devm_)free_netdev() already call netif_napi_del().

> +		free_netdev(ndev);
> +		return -ENODEV;

Does it makes sense to return ret as-is?

> +	}
> +
> +	return ret;
> +}
> +
> +static int umac_remove(struct platform_device *pdev)
> +{
> +	struct net_device *ndev = platform_get_drvdata(pdev);
> +	struct umac_priv *umac = netdev_priv(ndev);
> +
> +	unregister_netdev(ndev);
> +	iounmap(umac->base);

umac->base is a managed resource.

> +	free_netdev(ndev);
> +	return 0;
> +}
> +
> +static const struct of_device_id umac_of_matches[] = {
> +	{ .compatible = "hpe, gxp-umac", },
> +	{},
> +};
> +MODULE_DEVICE_TABLE(of, umac_of_matches);
> +
> +static struct platform_driver umac_driver = {
> +	.driver	= {
> +		.name    = "gxp-umac",
> +		.of_match_table = of_match_ptr(umac_of_matches),
> +	},
> +	.probe   = umac_probe,
> +	.remove  = umac_remove,

With the above proposed changes, the remove function could be removed.

> +};

[...]
  
Andrew Lunn July 21, 2023, 10:29 p.m. UTC | #2
> +#define PHY_88E1514_COPPER_CONTROL_REG		0
> +#define PHY_88E1514_PAGE_ADDRESS		22
> +
> +#define PHY_88E1514_GENERAL_CONTROL_REG1	20

This looks wrong. A MAC driver should never access the PHY directly.

> +
> +#define DRV_MODULE_NAME		"gxp-umac"
> +#define DRV_MODULE_VERSION	"0.1"

Versions are pointless. Please remove.

> +
> +#define NUMBER_OF_PORTS 2
> +#define EXTERNAL_PORT 1
> +#define INTERNAL_PORT 0
> +
> +struct umac_priv {
> +	void __iomem *base;
> +	int irq;
> +	struct platform_device *pdev;
> +	struct umac_tx_descs *tx_descs;
> +	struct umac_rx_descs *rx_descs;
> +	dma_addr_t tx_descs_dma_addr;
> +	dma_addr_t rx_descs_dma_addr;
> +	unsigned int tx_cur;
> +	unsigned int tx_done;
> +	unsigned int rx_cur;
> +	struct napi_struct napi;
> +	struct net_device *ndev;
> +	struct phy_device *phy_dev;
> +	struct phy_device *int_phy_dev;
> +	struct ncsi_dev *ncsidev;
> +	bool use_ncsi;
> +};
> +
> +static void umac_get_drvinfo(struct net_device *ndev,
> +			     struct ethtool_drvinfo *info)
> +{
> +	strscpy(info->driver, DRV_MODULE_NAME, sizeof(info->driver));
> +	strscpy(info->version, DRV_MODULE_VERSION, sizeof(info->version));

If you leave the version empty, you get the kernel version, and i
think hash. That is actually meaningful.

> +static int umac_get_link_ksettings(struct net_device *ndev,
> +				   struct ethtool_link_ksettings *cmd)
> +{
> +	phy_ethtool_ksettings_get(ndev->phydev, cmd);
> +	return 0;

return whatever phy_ethtool_ksettings_get() returns.

phy_ethtool_get_link_ksettings() is better here. You then don't even
need a function, you can reference it directly.

> +}
> +
> +static int umac_set_link_ksettings(struct net_device *ndev,
> +				   const struct ethtool_link_ksettings *cmd)
> +{
> +	return phy_ethtool_ksettings_set(ndev->phydev, cmd);

phy_ethtool_set_link_ksettings().

> +static int umac_nway_reset(struct net_device *ndev)
> +{
> +	return genphy_restart_aneg(ndev->phydev);

You locking is broken here. Use phy_ethtool_nway_reset()

> +static u32 umac_get_link(struct net_device *ndev)
> +{
> +	int err;
> +
> +	err = genphy_update_link(ndev->phydev);
> +	if (err)
> +		return ethtool_op_get_link(ndev);
> +
> +	return ndev->phydev->link;
> +}

You have something really wrong if you are doing this.

> +static int umac_ioctl(struct net_device *ndev, struct ifreq *ifr, int cmd)
> +{
> +	if (!netif_running(ndev))
> +		return -EINVAL;
> +
> +	if (!ndev->phydev)
> +		return -ENODEV;
> +
> +	return phy_mii_ioctl(ndev->phydev, ifr, cmd);
> +}

> +static int umac_int_phy_init(struct umac_priv *umac)
> +{
> +	struct phy_device *phy_dev = umac->int_phy_dev;
> +	unsigned int value;
> +
> +	value = phy_read(phy_dev, 0);
> +	if (value & 0x4000)
> +		pr_info("Internal PHY loopback is enabled - clearing\n");

The MAC driver never access the PHY directly.

What is putting it into loopback? The bootloader?

> +
> +	value &= ~0x4000; /* disable loopback */
> +	phy_write(phy_dev, 0, value);
> +
> +	value = phy_read(phy_dev, 0);
> +	value |= 0x1000; /* set aneg enable */
> +	value |= 0x8000; /* SW reset */
> +	phy_write(phy_dev, 0, value);
> +
> +	do {
> +		value = phy_read(phy_dev, 0);
> +	} while (value & 0x8000);

phy_start() will do this for you.

> +static int umac_phy_fixup(struct phy_device *phy_dev)
> +{
> +	unsigned int value;
> +
> +	/* set phy mode to SGMII to copper */
> +	/* set page to 18 by writing 18 to register 22 */
> +	phy_write(phy_dev, PHY_88E1514_PAGE_ADDRESS, 18);
> +	value = phy_read(phy_dev, PHY_88E1514_GENERAL_CONTROL_REG1);
> +	value &= ~0x07;
> +	value |= 0x01;
> +	phy_write(phy_dev, PHY_88E1514_GENERAL_CONTROL_REG1, value);
> +
> +	/* perform mode reset by setting bit 15 in general_control_reg1 */
> +	phy_write(phy_dev, PHY_88E1514_GENERAL_CONTROL_REG1, value | 0x8000);
> +
> +	do {
> +		value = phy_read(phy_dev, PHY_88E1514_GENERAL_CONTROL_REG1);
> +	} while (value & 0x8000);
> +
> +	/* after setting the mode, must perform a SW reset */
> +	phy_write(phy_dev, PHY_88E1514_PAGE_ADDRESS, 0); /* set page to 0 */
> +
> +	value = phy_read(phy_dev, PHY_88E1514_COPPER_CONTROL_REG);
> +	value |= 0x8000;
> +	phy_write(phy_dev, PHY_88E1514_COPPER_CONTROL_REG, value);
> +
> +	do {
> +		value = phy_read(phy_dev, PHY_88E1514_COPPER_CONTROL_REG);
> +	} while (value & 0x8000);

Please extend the PHY driver to do this. You can pass SGMII as the
interface, and have the PHY driver act on it.

> +static int umac_init_hw(struct net_device *ndev)
> +{
> +	if (umac->use_ncsi) {
> +		/* set correct tx clock */
> +		value &= UMAC_CFG_TX_CLK_EN;
> +		value &= ~UMAC_CFG_GTX_CLK_EN;
> +		value &= ~UMAC_CFG_GIGABIT_MODE; /* RMII mode */
> +		value |= UMAC_CFG_FULL_DUPLEX; /* full duplex */
> +	} else {
> +		if (ndev->phydev->duplex)
> +			value |= UMAC_CFG_FULL_DUPLEX;
> +		else
> +			value &= ~UMAC_CFG_FULL_DUPLEX;
> +
> +		if (ndev->phydev->speed == SPEED_1000) {

The MAC driver should only access phydev members inside the
adjust_link callback. Outside of that, these members can in
inconsistent.

> +static int umac_open(struct net_device *ndev)
> +{

...

> +	netdev_info(ndev, "%s is OPENED\n", ndev->name);

Please don't spam the kernel log.

> +static int umac_init_mac_address(struct net_device *ndev)
> +{
> +	struct umac_priv *umac = netdev_priv(ndev);
> +	struct platform_device *pdev = umac->pdev;
> +	char addr[ETH_ALEN];
> +	int err;
> +
> +	err = of_get_mac_address(pdev->dev.of_node, addr);
> +	if (err)
> +		netdev_err(ndev, "Failed to get address from device-tree: %d\n",
> +			   err);
> +
> +	if (is_valid_ether_addr(addr)) {
> +		dev_addr_set(ndev, addr);
> +		netdev_info(ndev,
> +			    "Read MAC address %pM from DTB\n", ndev->dev_addr);

netdev_dbg()

> +static void umac_adjust_link(struct net_device *ndev)
> +{
> +	struct umac_priv *umac = netdev_priv(ndev);
> +	int value;
> +
> +	if (ndev->phydev->link) {
> +		/* disable both clock */
> +		value = readl(umac->base + UMAC_CONFIG_STATUS);
> +		value &= 0xfffff9ff;
> +		writel(value, umac->base + UMAC_CONFIG_STATUS);
> +		udelay(2);
> +
> +		if (ndev->phydev->duplex)
> +			value |= UMAC_CFG_FULL_DUPLEX;
> +		else
> +			value &= ~UMAC_CFG_FULL_DUPLEX;
> +
> +		switch (ndev->phydev->speed) {
> +		case SPEED_1000:
> +			value &= ~UMAC_CFG_TX_CLK_EN;
> +			value |= UMAC_CFG_GTX_CLK_EN;
> +			value |= UMAC_CFG_GIGABIT_MODE;
> +			break;
> +		case SPEED_100:
> +			value |= UMAC_CFG_TX_CLK_EN;
> +			value &= ~UMAC_CFG_GTX_CLK_EN;
> +			value &= ~UMAC_CFG_GIGABIT_MODE;
> +			break;
> +		}
> +		/* update duplex and gigabit_mode to umac */
> +		writel(value, umac->base + UMAC_CONFIG_STATUS);
> +		udelay(2);
> +
> +		netif_carrier_on(ndev);

phylib will do this for you.

> +static int umac_setup_phy(struct net_device *ndev)
> +{
> +	struct umac_priv *umac = netdev_priv(ndev);
> +	struct platform_device *pdev = umac->pdev;
> +	struct device_node *phy_handle;
> +	phy_interface_t interface;
> +	struct device_node *eth_ports_np;
> +	struct device_node *port_np;
> +	int ret;
> +	int err;
> +	int i;
> +
> +	/* Get child node ethernet-ports. */
> +	eth_ports_np = of_get_child_by_name(pdev->dev.of_node, "ethernet-ports");
> +	if (!eth_ports_np) {
> +		dev_err(&pdev->dev, "No ethernet-ports child node found!\n");
> +		return -ENODEV;
> +	}
> +
> +	for (i = 0; i < NUMBER_OF_PORTS; i++) {
> +		/* Get port@i of node ethernet-ports */
> +		port_np = gxp_umac_get_eth_child_node(eth_ports_np, i);
> +		if (!port_np)
> +			break;
> +
> +		if (i == INTERNAL_PORT) {
> +			phy_handle = of_parse_phandle(port_np, "phy-handle", 0);
> +			if (phy_handle) {
> +				umac->int_phy_dev = of_phy_find_device(phy_handle);
> +				if (!umac->int_phy_dev)
> +					return -ENODEV;
> +
> +				umac_int_phy_init(umac);
> +			} else {
> +				return dev_err_probe(&pdev->dev, PTR_ERR(phy_handle),
> +						     "Failed to map phy-handle for port %d", i);
> +			}
> +		}
> +
> +		if (i == EXTERNAL_PORT) {
> +			phy_handle = of_parse_phandle(port_np, "phy-handle", 0);
> +			if (phy_handle) {
> +				/* register the phy board fixup */
> +				ret = phy_register_fixup_for_uid(0x01410dd1, 0xffffffff,
> +								 umac_phy_fixup);
> +				if (ret)
> +					dev_err(&pdev->dev, "cannot register phy board fixup\n");
> +
> +				err = of_get_phy_mode(phy_handle, &interface);
> +				if (err)
> +					interface = PHY_INTERFACE_MODE_NA;
> +
> +				umac->phy_dev = of_phy_connect(ndev, phy_handle,
> +							       &umac_adjust_link,
> +							       0, interface);
> +
> +				if (!umac->phy_dev)
> +					return -ENODEV;

It looks like you MAC does not support 10Mbps. At some point you need
to remove the two link modes using phy_remove_link_mode().

> +
> +				/* If the specified phy-handle has a fixed-link declaration, use the
> +				 * fixed-link properties to set the configuration for the PHY

This is wrong. Look at other MAC drivers using fixed-link.

     Andrew
  
Simon Horman July 24, 2023, 3:01 p.m. UTC | #3
On Fri, Jul 21, 2023 at 04:20:43PM -0500, nick.hawkins@hpe.com wrote:

...

Hi Nick,

> diff --git a/drivers/net/ethernet/hpe/gxp-umac.c b/drivers/net/ethernet/hpe/gxp-umac.c

...

> +static void umac_set_mac_address(struct net_device *ndev, void *p_addr)
> +{
> +	struct umac_priv *umac = netdev_priv(ndev);
> +	char *addr = (char *)p_addr;
> +	unsigned int value;
> +
> +	/* update address to register */
> +	value = addr[0] << 8 | addr[1];
> +	writel(value, umac->base + UMAC_MAC_ADDR_HI);
> +	value = addr[2] << 8 | addr[3];
> +	writel(value, umac->base + UMAC_MAC_ADDR_MID);
> +	value = addr[4] << 8 | addr[5];
> +	writel(value, umac->base + UMAC_MAC_ADDR_LO);
> +}
> +
> +static int umac_eth_mac_addr(struct net_device *ndev, void *p)
> +{
> +	int ret;
> +	struct sockaddr *addr = p;

Please use reverse xmas tree - longest like to shortest - for
new Networking code. Likewise in some other places in this patch/series.

This tool can be helpful in this regard.
https://github.com/ecree-solarflare/xmastree

...

> +static int umac_init_hw(struct net_device *ndev)
> +{
> +	struct umac_priv *umac = netdev_priv(ndev);
> +	unsigned int value;
> +
> +	/* initialize tx and rx rings to first entry */
> +	writel(0, umac->base + UMAC_RING_PTR);
> +
> +	/* clear the missed bit */
> +	writel(0, umac->base + UMAC_CLEAR_STATUS);
> +
> +	/* disable checksum generation */
> +	writel(0, umac->base + UMAC_CKSUM_CONFIG);
> +
> +	/* write the ring size register */
> +	value = ((UMAC_RING_SIZE_256 << UMAC_TX_RING_SIZE_SHIFT) &
> +			UMAC_TX_RING_SIZE_MASK) |
> +		((UMAC_RING_SIZE_256 << UMAC_RX_RING_SIZE_SHIFT) &
> +			UMAC_RX_RING_SIZE_MASK);
> +	writel(value, umac->base + UMAC_RING_SIZE);
> +
> +	/* write rx ring base address */
> +	writel(cpu_to_le32(umac->rx_descs_dma_addr),
> +	       umac->base + UMAC_RX_RING_ADDR);

It is my understanding that writel will convert the value
from host byte order to little endien. If so then pre-converting value
seems incorrect. Perhaps this should be:

	writel(umac->rx_descs_dma_addr, umac->base + UMAC_RX_RING_ADDR);

Flagged by Sparse.

> +
> +	/* write tx ring base address */
> +	writel(cpu_to_le32(umac->tx_descs_dma_addr),
> +	       umac->base + UMAC_TX_RING_ADDR);

Ditto.

> +
> +	/* write burst size */
> +	writel(0x22, umac->base + UMAC_DMA_CONFIG);
> +
> +	umac_channel_disable(umac);
> +
> +	/* disable clocks and gigabit mode (leave channels disabled) */
> +	value = readl(umac->base + UMAC_CONFIG_STATUS);
> +	value &= 0xfffff9ff;
> +	writel(value, umac->base + UMAC_CONFIG_STATUS);
> +	udelay(2);
> +
> +	if (umac->use_ncsi) {
> +		/* set correct tx clock */
> +		value &= UMAC_CFG_TX_CLK_EN;
> +		value &= ~UMAC_CFG_GTX_CLK_EN;
> +		value &= ~UMAC_CFG_GIGABIT_MODE; /* RMII mode */
> +		value |= UMAC_CFG_FULL_DUPLEX; /* full duplex */
> +	} else {
> +		if (ndev->phydev->duplex)
> +			value |= UMAC_CFG_FULL_DUPLEX;
> +		else
> +			value &= ~UMAC_CFG_FULL_DUPLEX;
> +
> +		if (ndev->phydev->speed == SPEED_1000) {
> +			value &= ~UMAC_CFG_TX_CLK_EN;
> +			value |= UMAC_CFG_GTX_CLK_EN;
> +			value |= UMAC_CFG_GIGABIT_MODE;
> +		} else {
> +			value |= UMAC_CFG_TX_CLK_EN;
> +			value &= ~UMAC_CFG_GTX_CLK_EN;
> +			value &= ~UMAC_CFG_GIGABIT_MODE;
> +		}
> +	}
> +	writel(value, umac->base + UMAC_CONFIG_STATUS);
> +	udelay(2);
> +
> +	umac_channel_enable(umac);
> +
> +	return 0;
> +}

...

> diff --git a/drivers/net/ethernet/hpe/gxp-umac.h b/drivers/net/ethernet/hpe/gxp-umac.h

...

> +struct umac_rx_desc_entry {
> +	u32  dmaaddress;   // Start address for DMA operationg

1. operationg -> operating
2. It appears that this field is used to hold __le32 values.
   Perhaps it's type should be __le32.
   As is, Sparse complains about this.

> +	u16  status;       // Packet tx status and ownership flag
> +	u16  count;        // Number of bytes received
> +	u16  checksum;     // On-the-fly packet checksum
> +	u16  control;      // Checksum-in-time flag
> +	u32  reserved;
> +} __aligned(16);
> +
> +struct umac_rx_descs {
> +	struct umac_rx_desc_entry entrylist[UMAC_MAX_RX_DESC_ENTRIES];
> +	u8 framelist[UMAC_MAX_RX_DESC_ENTRIES][UMAC_MAX_RX_FRAME_SIZE];
> +} __packed;
> +
> +struct umac_tx_desc_entry {
> +	u32  dmaaddress;   // Start address for DMA operationg

Ditto.

> +	u16  status;       // Packet rx status, type, and ownership flag
> +	u16  count;        // Number of bytes received
> +	u32  cksumoffset;  // Specifies where to place packet checksum
> +	u32  reserved;
> +} __aligned(16);
> +
> +struct umac_tx_descs {
> +	struct umac_tx_desc_entry entrylist[UMAC_MAX_TX_DESC_ENTRIES];
> +	u8 framelist[UMAC_MAX_TX_DESC_ENTRIES][UMAC_MAX_TX_FRAME_SIZE];
> +} __packed;
> +
> +#endif
> -- 
> 2.17.1
> 
>
  

Patch

diff --git a/drivers/net/ethernet/hpe/Kconfig b/drivers/net/ethernet/hpe/Kconfig
index 461aa15ace34..cc269558b707 100644
--- a/drivers/net/ethernet/hpe/Kconfig
+++ b/drivers/net/ethernet/hpe/Kconfig
@@ -14,6 +14,21 @@  config NET_VENDOR_HPE
 
 if NET_VENDOR_HPE
 
+config GXP_UMAC
+	tristate "GXP UMAC support"
+	depends on ARCH_HPE
+	select CRC32
+	select MII
+	select PHYLIB
+	select GXP_UMAC_MDIO
+	help
+	  Say y here to support the GXP UMACs interface. The
+	  primary MAC (umac0) can be mapped to either
+	  SGMII/1000-BaseX SERDES interface. The secondary MAC
+	  (umac1) can be mapped to only the second
+	  SGMII/1000-Base X Serdes interface or it can be
+	  mapped for RMII sideband.
+
 config GXP_UMAC_MDIO
 	tristate "GXP UMAC mdio support"
 	depends on ARCH_HPE
diff --git a/drivers/net/ethernet/hpe/Makefile b/drivers/net/ethernet/hpe/Makefile
index e84c8786ba04..5c2c3bcde532 100644
--- a/drivers/net/ethernet/hpe/Makefile
+++ b/drivers/net/ethernet/hpe/Makefile
@@ -1 +1,2 @@ 
 obj-$(CONFIG_GXP_UMAC_MDIO) += gxp-umac-mdio.o
+obj-$(CONFIG_GXP_UMAC) += gxp-umac.o
diff --git a/drivers/net/ethernet/hpe/gxp-umac.c b/drivers/net/ethernet/hpe/gxp-umac.c
new file mode 100644
index 000000000000..a4fe3eb9f68f
--- /dev/null
+++ b/drivers/net/ethernet/hpe/gxp-umac.c
@@ -0,0 +1,911 @@ 
+// SPDX-License-Identifier: GPL-2.0-only
+/* Copyright (C) 2023 Hewlett-Packard Enterprise Development Company, L.P. */
+
+#include <linux/dma-mapping.h>
+#include <linux/etherdevice.h>
+#include <linux/ethtool.h>
+#include <linux/iopoll.h>
+#include <linux/module.h>
+#include <net/ncsi.h>
+#include <linux/of_device.h>
+#include <linux/of_mdio.h>
+#include <linux/of_net.h>
+#include <linux/phy.h>
+#include "gxp-umac.h"
+
+#define PHY_88E1514_COPPER_CONTROL_REG		0
+#define PHY_88E1514_PAGE_ADDRESS		22
+
+#define PHY_88E1514_GENERAL_CONTROL_REG1	20
+
+#define DRV_MODULE_NAME		"gxp-umac"
+#define DRV_MODULE_VERSION	"0.1"
+
+#define NUMBER_OF_PORTS 2
+#define EXTERNAL_PORT 1
+#define INTERNAL_PORT 0
+
+struct umac_priv {
+	void __iomem *base;
+	int irq;
+	struct platform_device *pdev;
+	struct umac_tx_descs *tx_descs;
+	struct umac_rx_descs *rx_descs;
+	dma_addr_t tx_descs_dma_addr;
+	dma_addr_t rx_descs_dma_addr;
+	unsigned int tx_cur;
+	unsigned int tx_done;
+	unsigned int rx_cur;
+	struct napi_struct napi;
+	struct net_device *ndev;
+	struct phy_device *phy_dev;
+	struct phy_device *int_phy_dev;
+	struct ncsi_dev *ncsidev;
+	bool use_ncsi;
+};
+
+static void umac_get_drvinfo(struct net_device *ndev,
+			     struct ethtool_drvinfo *info)
+{
+	strscpy(info->driver, DRV_MODULE_NAME, sizeof(info->driver));
+	strscpy(info->version, DRV_MODULE_VERSION, sizeof(info->version));
+}
+
+static int umac_get_link_ksettings(struct net_device *ndev,
+				   struct ethtool_link_ksettings *cmd)
+{
+	phy_ethtool_ksettings_get(ndev->phydev, cmd);
+	return 0;
+}
+
+static int umac_set_link_ksettings(struct net_device *ndev,
+				   const struct ethtool_link_ksettings *cmd)
+{
+	return phy_ethtool_ksettings_set(ndev->phydev, cmd);
+}
+
+static int umac_nway_reset(struct net_device *ndev)
+{
+	return genphy_restart_aneg(ndev->phydev);
+}
+
+static u32 umac_get_link(struct net_device *ndev)
+{
+	int err;
+
+	err = genphy_update_link(ndev->phydev);
+	if (err)
+		return ethtool_op_get_link(ndev);
+
+	return ndev->phydev->link;
+}
+
+static struct net_device_stats *umac_get_stats(struct net_device *ndev)
+{
+	return &ndev->stats;
+}
+
+static int umac_ioctl(struct net_device *ndev, struct ifreq *ifr, int cmd)
+{
+	if (!netif_running(ndev))
+		return -EINVAL;
+
+	if (!ndev->phydev)
+		return -ENODEV;
+
+	return phy_mii_ioctl(ndev->phydev, ifr, cmd);
+}
+
+static void umac_set_mac_address(struct net_device *ndev, void *p_addr)
+{
+	struct umac_priv *umac = netdev_priv(ndev);
+	char *addr = (char *)p_addr;
+	unsigned int value;
+
+	/* update address to register */
+	value = addr[0] << 8 | addr[1];
+	writel(value, umac->base + UMAC_MAC_ADDR_HI);
+	value = addr[2] << 8 | addr[3];
+	writel(value, umac->base + UMAC_MAC_ADDR_MID);
+	value = addr[4] << 8 | addr[5];
+	writel(value, umac->base + UMAC_MAC_ADDR_LO);
+}
+
+static int umac_eth_mac_addr(struct net_device *ndev, void *p)
+{
+	int ret;
+	struct sockaddr *addr = p;
+
+	ret = eth_prepare_mac_addr_change(ndev, p);
+	if (ret < 0)
+		return ret;
+
+	eth_commit_mac_addr_change(ndev, p);
+	umac_set_mac_address(ndev, addr->sa_data);
+
+	return 0;
+}
+
+static void umac_channel_enable(struct umac_priv *umac)
+{
+	unsigned int value;
+
+	value = readl(umac->base + UMAC_CONFIG_STATUS);
+	value |= UMAC_CFG_TXEN | UMAC_CFG_RXEN;
+	writel(value, umac->base + UMAC_CONFIG_STATUS);
+
+	/* start processing by writing the ring prompt register */
+	writel(0, umac->base + UMAC_RING_PROMPT);
+}
+
+static void umac_channel_disable(struct umac_priv *umac)
+{
+	writel(0, umac->base + UMAC_CONFIG_STATUS);
+}
+
+static int umac_init_ring_discriptor(struct net_device *ndev)
+{
+	struct umac_priv *umac = netdev_priv(ndev);
+	struct platform_device *pdev = umac->pdev;
+
+	struct umac_tx_desc_entry *ptxdesc;
+	struct umac_rx_desc_entry *prxdesc;
+
+	unsigned int i;
+
+	if (dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(32))) {
+		netdev_err(ndev, "No suitable DMA available\n");
+		return -ENOMEM;
+	}
+
+	umac->tx_descs = dma_alloc_coherent(&pdev->dev,
+					    sizeof(struct umac_tx_descs),
+					&umac->tx_descs_dma_addr, GFP_KERNEL);
+	if (!umac->tx_descs)
+		return -ENOMEM;
+
+	umac->rx_descs = dma_alloc_coherent(&pdev->dev,
+					    sizeof(struct umac_rx_descs),
+					&umac->rx_descs_dma_addr, GFP_KERNEL);
+	if (!umac->rx_descs) {
+		dma_free_coherent(&pdev->dev, sizeof(struct umac_tx_descs),
+				  umac->tx_descs,
+					umac->tx_descs_dma_addr);
+		return -ENOMEM;
+	}
+
+	for (i = 0; i < UMAC_MAX_TX_DESC_ENTRIES; i++) {
+		ptxdesc = &umac->tx_descs->entrylist[i];
+		ptxdesc->dmaaddress = cpu_to_le32(umac->tx_descs_dma_addr +
+						offsetof(struct umac_tx_descs,
+							 framelist[i][0]));
+	}
+
+	for (i = 0; i < UMAC_MAX_RX_DESC_ENTRIES; i++) {
+		prxdesc = &umac->rx_descs->entrylist[i];
+		prxdesc->dmaaddress = cpu_to_le32(umac->rx_descs_dma_addr +
+						offsetof(struct umac_rx_descs,
+							 framelist[i][0]));
+		prxdesc->status = UMAC_RING_ENTRY_HW_OWN;
+		prxdesc->count = UMAC_MAX_RX_FRAME_SIZE;
+	}
+
+	umac->tx_cur = 0;
+	umac->tx_done = 0;
+	umac->rx_cur = 0;
+
+	return 0;
+}
+
+static int umac_int_phy_init(struct umac_priv *umac)
+{
+	struct phy_device *phy_dev = umac->int_phy_dev;
+	unsigned int value;
+
+	value = phy_read(phy_dev, 0);
+	if (value & 0x4000)
+		pr_info("Internal PHY loopback is enabled - clearing\n");
+
+	value &= ~0x4000; /* disable loopback */
+	phy_write(phy_dev, 0, value);
+
+	value = phy_read(phy_dev, 0);
+	value |= 0x1000; /* set aneg enable */
+	value |= 0x8000; /* SW reset */
+	phy_write(phy_dev, 0, value);
+
+	do {
+		value = phy_read(phy_dev, 0);
+	} while (value & 0x8000);
+
+	return 0;
+}
+
+static int umac_phy_fixup(struct phy_device *phy_dev)
+{
+	unsigned int value;
+
+	/* set phy mode to SGMII to copper */
+	/* set page to 18 by writing 18 to register 22 */
+	phy_write(phy_dev, PHY_88E1514_PAGE_ADDRESS, 18);
+	value = phy_read(phy_dev, PHY_88E1514_GENERAL_CONTROL_REG1);
+	value &= ~0x07;
+	value |= 0x01;
+	phy_write(phy_dev, PHY_88E1514_GENERAL_CONTROL_REG1, value);
+
+	/* perform mode reset by setting bit 15 in general_control_reg1 */
+	phy_write(phy_dev, PHY_88E1514_GENERAL_CONTROL_REG1, value | 0x8000);
+
+	do {
+		value = phy_read(phy_dev, PHY_88E1514_GENERAL_CONTROL_REG1);
+	} while (value & 0x8000);
+
+	/* after setting the mode, must perform a SW reset */
+	phy_write(phy_dev, PHY_88E1514_PAGE_ADDRESS, 0); /* set page to 0 */
+
+	value = phy_read(phy_dev, PHY_88E1514_COPPER_CONTROL_REG);
+	value |= 0x8000;
+	phy_write(phy_dev, PHY_88E1514_COPPER_CONTROL_REG, value);
+
+	do {
+		value = phy_read(phy_dev, PHY_88E1514_COPPER_CONTROL_REG);
+	} while (value & 0x8000);
+
+	return 0;
+}
+
+static int umac_init_hw(struct net_device *ndev)
+{
+	struct umac_priv *umac = netdev_priv(ndev);
+	unsigned int value;
+
+	/* initialize tx and rx rings to first entry */
+	writel(0, umac->base + UMAC_RING_PTR);
+
+	/* clear the missed bit */
+	writel(0, umac->base + UMAC_CLEAR_STATUS);
+
+	/* disable checksum generation */
+	writel(0, umac->base + UMAC_CKSUM_CONFIG);
+
+	/* write the ring size register */
+	value = ((UMAC_RING_SIZE_256 << UMAC_TX_RING_SIZE_SHIFT) &
+			UMAC_TX_RING_SIZE_MASK) |
+		((UMAC_RING_SIZE_256 << UMAC_RX_RING_SIZE_SHIFT) &
+			UMAC_RX_RING_SIZE_MASK);
+	writel(value, umac->base + UMAC_RING_SIZE);
+
+	/* write rx ring base address */
+	writel(cpu_to_le32(umac->rx_descs_dma_addr),
+	       umac->base + UMAC_RX_RING_ADDR);
+
+	/* write tx ring base address */
+	writel(cpu_to_le32(umac->tx_descs_dma_addr),
+	       umac->base + UMAC_TX_RING_ADDR);
+
+	/* write burst size */
+	writel(0x22, umac->base + UMAC_DMA_CONFIG);
+
+	umac_channel_disable(umac);
+
+	/* disable clocks and gigabit mode (leave channels disabled) */
+	value = readl(umac->base + UMAC_CONFIG_STATUS);
+	value &= 0xfffff9ff;
+	writel(value, umac->base + UMAC_CONFIG_STATUS);
+	udelay(2);
+
+	if (umac->use_ncsi) {
+		/* set correct tx clock */
+		value &= UMAC_CFG_TX_CLK_EN;
+		value &= ~UMAC_CFG_GTX_CLK_EN;
+		value &= ~UMAC_CFG_GIGABIT_MODE; /* RMII mode */
+		value |= UMAC_CFG_FULL_DUPLEX; /* full duplex */
+	} else {
+		if (ndev->phydev->duplex)
+			value |= UMAC_CFG_FULL_DUPLEX;
+		else
+			value &= ~UMAC_CFG_FULL_DUPLEX;
+
+		if (ndev->phydev->speed == SPEED_1000) {
+			value &= ~UMAC_CFG_TX_CLK_EN;
+			value |= UMAC_CFG_GTX_CLK_EN;
+			value |= UMAC_CFG_GIGABIT_MODE;
+		} else {
+			value |= UMAC_CFG_TX_CLK_EN;
+			value &= ~UMAC_CFG_GTX_CLK_EN;
+			value &= ~UMAC_CFG_GIGABIT_MODE;
+		}
+	}
+	writel(value, umac->base + UMAC_CONFIG_STATUS);
+	udelay(2);
+
+	umac_channel_enable(umac);
+
+	return 0;
+}
+
+static int umac_start_xmit(struct sk_buff *skb, struct net_device *ndev)
+{
+	struct umac_priv *umac = netdev_priv(ndev);
+	struct umac_tx_desc_entry *ptxdesc;
+	u8 *pframe;
+	unsigned int length;
+
+	ptxdesc = &umac->tx_descs->entrylist[umac->tx_cur];
+	pframe = umac->tx_descs->framelist[umac->tx_cur];
+
+	length = skb->len;
+	if (length > 1514) {
+		netdev_err(ndev, "send data %d bytes > 1514, clamp it to 1514\n",
+			   skb->len);
+		length = 1514;
+	}
+
+	memset(pframe, 0, UMAC_MAX_FRAME_SIZE);
+	memcpy(pframe, skb->data, length);
+
+	if (length < ETH_ZLEN)
+		length = ETH_ZLEN; /* minimum tx byte */
+
+	ptxdesc->count = length;
+	ptxdesc->status = UMAC_RING_ENTRY_HW_OWN;
+	ptxdesc->cksumoffset = 0; /* disable checksum generation */
+
+	umac->tx_cur++;
+	if (umac->tx_cur >= UMAC_MAX_TX_DESC_ENTRIES)
+		umac->tx_cur = 0;
+
+	/* if current tx ring buffer is full, stop the queue */
+	ptxdesc = &umac->tx_descs->entrylist[umac->tx_cur];
+	if (ptxdesc->status & UMAC_RING_ENTRY_HW_OWN)
+		netif_stop_queue(ndev);
+
+	/* start processing by writing the ring prompt register */
+	writel(0, umac->base + UMAC_RING_PROMPT);
+	dev_kfree_skb(skb);
+
+	return NETDEV_TX_OK;
+}
+
+static int umac_rx(struct net_device *ndev, int budget)
+{
+	struct umac_priv *umac = netdev_priv(ndev);
+
+	struct umac_rx_desc_entry *prxdesc;
+	struct sk_buff *skb;
+
+	unsigned int rxlength;
+	int rxpktcount = 0;
+	u8 *pframe;
+	u8 *skb_buf;
+
+	prxdesc = &umac->rx_descs->entrylist[umac->rx_cur];
+	pframe = umac->rx_descs->framelist[umac->rx_cur];
+
+	while (!(prxdesc->status & UMAC_RING_ENTRY_HW_OWN)) {
+		rxlength = prxdesc->count;
+		skb = netdev_alloc_skb(ndev, rxlength);
+		if (!skb) {
+			/* run out of memory */
+			ndev->stats.rx_dropped++;
+			return rxpktcount;
+		}
+
+		/* make 16 bytes aligned for 14 bytes ethernet header */
+		skb_buf = skb_put(skb, rxlength);
+		memcpy(skb_buf, pframe, rxlength);
+
+		skb->protocol = eth_type_trans(skb, ndev);
+		netif_receive_skb(skb);
+		rxpktcount++;
+
+		prxdesc->status = UMAC_RING_ENTRY_HW_OWN;
+		prxdesc->count = UMAC_MAX_FRAME_SIZE;
+
+		ndev->stats.rx_packets++;
+		ndev->stats.rx_bytes += rxlength;
+
+		/* move to next buffer */
+		umac->rx_cur++;
+		if (umac->rx_cur >= UMAC_MAX_RX_DESC_ENTRIES)
+			umac->rx_cur = 0;
+
+		if (rxpktcount >= budget)
+			break;
+
+		prxdesc = &umac->rx_descs->entrylist[umac->rx_cur];
+		pframe = umac->rx_descs->framelist[umac->rx_cur];
+	}
+	/* start processing by writing the ring prompt register */
+	writel(0, umac->base + UMAC_RING_PROMPT);
+
+	return rxpktcount;
+}
+
+static void umac_tx_done(struct net_device *ndev)
+{
+	struct umac_priv *umac = netdev_priv(ndev);
+
+	unsigned int txptr;
+	unsigned int value;
+	struct umac_tx_desc_entry *ptxdesc;
+
+	value = readl(umac->base + UMAC_RING_PTR);
+	txptr = (value & UMAC_TX_RING_PTR_MASK) >> UMAC_TX_RING_PTR_SHIFT;
+
+	ptxdesc = &umac->tx_descs->entrylist[umac->tx_done];
+
+	while (!(ptxdesc->status & UMAC_RING_ENTRY_HW_OWN)) {
+		if (umac->tx_done == txptr)
+			break;
+
+		ndev->stats.tx_packets++;
+		ndev->stats.tx_bytes += ptxdesc->count;
+
+		umac->tx_done++;
+		if (umac->tx_done >= UMAC_MAX_TX_DESC_ENTRIES)
+			umac->tx_done = 0;
+		ptxdesc = &umac->tx_descs->entrylist[umac->tx_done];
+	}
+
+	/* clear tx interrupt */
+	value = readl(umac->base + UMAC_INTERRUPT);
+	value &= ~UMAC_TX_INT;
+	writel(value, umac->base + UMAC_INTERRUPT);
+
+	if (netif_queue_stopped(ndev))
+		netif_wake_queue(ndev);
+}
+
+static void umac_irq_enable(struct umac_priv *umac)
+{
+	unsigned int value;
+
+	/* enable interrupt */
+	value = readl(umac->base + UMAC_INTERRUPT);
+	value |= (UMAC_RX_INTEN | UMAC_TX_INTEN);
+	writel(value, umac->base + UMAC_INTERRUPT);
+}
+
+static void umac_irq_disable(struct umac_priv *umac)
+{
+	unsigned int value;
+
+	/* clear and disable interrupt */
+	value = readl(umac->base + UMAC_INTERRUPT);
+	value |= (UMAC_RX_INT | UMAC_TX_INT);
+	value &= ~(UMAC_RX_INTEN | UMAC_TX_INTEN);
+	writel(value, umac->base + UMAC_INTERRUPT);
+}
+
+static irqreturn_t umac_interrupt(int irq, void *p_ndev)
+{
+	struct net_device *ndev = (struct net_device *)p_ndev;
+	struct umac_priv *umac = netdev_priv(ndev);
+
+	if (umac->use_ncsi || netif_running(ndev)) {
+		umac_irq_disable(umac);
+		napi_schedule(&umac->napi);
+	}
+
+	return IRQ_HANDLED;
+}
+
+static int umac_poll(struct napi_struct *napi, int budget)
+{
+	struct umac_priv *umac = container_of(napi, struct umac_priv, napi);
+	struct net_device *ndev = umac->ndev;
+	unsigned int value;
+	int rx_done = 0;
+
+	umac_tx_done(ndev);
+
+	rx_done = umac_rx(ndev, budget);
+
+	if (rx_done < budget) {
+		napi_complete_done(napi, rx_done);
+		/* clear rx interrupt */
+		value = readl(umac->base + UMAC_INTERRUPT);
+		value &= ~UMAC_RX_INT;
+		writel(value, umac->base + UMAC_INTERRUPT);
+
+		/* enable interrupt */
+		umac_irq_enable(umac);
+	}
+
+	return rx_done;
+}
+
+static int umac_open(struct net_device *ndev)
+{
+	struct umac_priv *umac = netdev_priv(ndev);
+	int err;
+
+	if (request_irq(ndev->irq, umac_interrupt, 0x0, ndev->name, ndev)) {
+		netdev_err(ndev, "failed to register irq\n");
+		return -EAGAIN;
+	}
+
+	umac_init_ring_discriptor(ndev);
+	umac_init_hw(ndev);
+
+	if (umac->use_ncsi)
+		netif_carrier_on(ndev);
+	else
+		phy_start(ndev->phydev);
+
+	napi_enable(&umac->napi);
+	netif_start_queue(ndev);
+	umac_irq_enable(umac);
+
+	if (umac->use_ncsi) {
+		err = ncsi_start_dev(umac->ncsidev);
+		if (err) {
+			netdev_err(ndev, "failed to start ncsi\n");
+			free_irq(ndev->irq, ndev);
+			return err;
+		}
+	}
+
+	netdev_info(ndev, "%s is OPENED\n", ndev->name);
+	return 0;
+}
+
+static int umac_stop(struct net_device *ndev)
+{
+	struct umac_priv *umac = netdev_priv(ndev);
+	struct platform_device *pdev = umac->pdev;
+
+	dma_free_coherent(&pdev->dev, sizeof(struct umac_tx_descs),
+			  umac->tx_descs, umac->tx_descs_dma_addr);
+	dma_free_coherent(&pdev->dev, sizeof(struct umac_rx_descs),
+			  umac->rx_descs, umac->rx_descs_dma_addr);
+	netif_stop_queue(ndev);
+
+	if (umac->use_ncsi)
+		ncsi_stop_dev(umac->ncsidev);
+	else
+		phy_stop(ndev->phydev);
+	umac_irq_disable(umac);
+	umac_channel_disable(umac);
+	napi_disable(&umac->napi);
+
+	free_irq(ndev->irq, ndev);
+
+	return 0;
+}
+
+static const struct ethtool_ops umac_ethtool_ops = {
+	.get_ts_info	= ethtool_op_get_ts_info,
+	.get_link_ksettings	= umac_get_link_ksettings,
+	.set_link_ksettings	= umac_set_link_ksettings,
+	.get_drvinfo	= umac_get_drvinfo,
+	.nway_reset	= umac_nway_reset,
+	.get_link	= umac_get_link,
+};
+
+static const struct net_device_ops umac_netdev_ops = {
+	.ndo_open		= umac_open,
+	.ndo_stop		= umac_stop,
+	.ndo_start_xmit		= umac_start_xmit,
+	.ndo_get_stats		= umac_get_stats,
+	.ndo_do_ioctl		= umac_ioctl,
+	.ndo_validate_addr	= eth_validate_addr,
+	.ndo_set_mac_address	= umac_eth_mac_addr,
+};
+
+static int umac_init_mac_address(struct net_device *ndev)
+{
+	struct umac_priv *umac = netdev_priv(ndev);
+	struct platform_device *pdev = umac->pdev;
+	char addr[ETH_ALEN];
+	int err;
+
+	err = of_get_mac_address(pdev->dev.of_node, addr);
+	if (err)
+		netdev_err(ndev, "Failed to get address from device-tree: %d\n",
+			   err);
+
+	if (is_valid_ether_addr(addr)) {
+		dev_addr_set(ndev, addr);
+		netdev_info(ndev,
+			    "Read MAC address %pM from DTB\n", ndev->dev_addr);
+	} else {
+		eth_hw_addr_random(ndev);
+		netdev_info(ndev, "Generated random MAC address %pM\n",
+			    ndev->dev_addr);
+	}
+
+	dev_addr_set(ndev, addr);
+	umac_set_mac_address(ndev, addr);
+
+	return 0;
+}
+
+static void umac_ncsi_handler(struct ncsi_dev *ncsidev)
+{
+	if (unlikely(ncsidev->state != ncsi_dev_state_functional))
+		return;
+
+	netdev_info(ncsidev->dev, "NCSI interface %s\n",
+		    ncsidev->link_up ? "up" : "down");
+}
+
+static void umac_adjust_link(struct net_device *ndev)
+{
+	struct umac_priv *umac = netdev_priv(ndev);
+	int value;
+
+	if (ndev->phydev->link) {
+		/* disable both clock */
+		value = readl(umac->base + UMAC_CONFIG_STATUS);
+		value &= 0xfffff9ff;
+		writel(value, umac->base + UMAC_CONFIG_STATUS);
+		udelay(2);
+
+		if (ndev->phydev->duplex)
+			value |= UMAC_CFG_FULL_DUPLEX;
+		else
+			value &= ~UMAC_CFG_FULL_DUPLEX;
+
+		switch (ndev->phydev->speed) {
+		case SPEED_1000:
+			value &= ~UMAC_CFG_TX_CLK_EN;
+			value |= UMAC_CFG_GTX_CLK_EN;
+			value |= UMAC_CFG_GIGABIT_MODE;
+			break;
+		case SPEED_100:
+			value |= UMAC_CFG_TX_CLK_EN;
+			value &= ~UMAC_CFG_GTX_CLK_EN;
+			value &= ~UMAC_CFG_GIGABIT_MODE;
+			break;
+		}
+		/* update duplex and gigabit_mode to umac */
+		writel(value, umac->base + UMAC_CONFIG_STATUS);
+		udelay(2);
+
+		netif_carrier_on(ndev);
+	} else {
+		/* disable both clock */
+		value = readl(umac->base + UMAC_CONFIG_STATUS);
+		value &= 0xfffff9ff;
+		writel(value, umac->base + UMAC_CONFIG_STATUS);
+		udelay(2);
+
+		value &= ~UMAC_CFG_FULL_DUPLEX;
+		value &= ~UMAC_CFG_GTX_CLK_EN;
+		value &= ~UMAC_CFG_GIGABIT_MODE;
+		value |= UMAC_CFG_TX_CLK_EN;
+		writel(value, umac->base + UMAC_CONFIG_STATUS);
+		udelay(2);
+
+		netif_carrier_off(ndev);
+	}
+}
+
+static struct device_node *gxp_umac_get_eth_child_node(struct device_node *ether_np, int id)
+{
+	struct device_node *port_np;
+	int port_id;
+
+	for_each_child_of_node(ether_np, port_np) {
+		/* It is not a 'port' node, continue. */
+		if (strcmp(port_np->name, "port"))
+			continue;
+		if (of_property_read_u32(port_np, "reg", &port_id) < 0)
+			continue;
+
+		if (port_id == id)
+			return port_np;
+	}
+
+	/* Not found! */
+	return NULL;
+}
+
+static int umac_setup_phy(struct net_device *ndev)
+{
+	struct umac_priv *umac = netdev_priv(ndev);
+	struct platform_device *pdev = umac->pdev;
+	struct device_node *phy_handle;
+	phy_interface_t interface;
+	struct device_node *eth_ports_np;
+	struct device_node *port_np;
+	int ret;
+	int err;
+	int i;
+
+	/* Get child node ethernet-ports. */
+	eth_ports_np = of_get_child_by_name(pdev->dev.of_node, "ethernet-ports");
+	if (!eth_ports_np) {
+		dev_err(&pdev->dev, "No ethernet-ports child node found!\n");
+		return -ENODEV;
+	}
+
+	for (i = 0; i < NUMBER_OF_PORTS; i++) {
+		/* Get port@i of node ethernet-ports */
+		port_np = gxp_umac_get_eth_child_node(eth_ports_np, i);
+		if (!port_np)
+			break;
+
+		if (i == INTERNAL_PORT) {
+			phy_handle = of_parse_phandle(port_np, "phy-handle", 0);
+			if (phy_handle) {
+				umac->int_phy_dev = of_phy_find_device(phy_handle);
+				if (!umac->int_phy_dev)
+					return -ENODEV;
+
+				umac_int_phy_init(umac);
+			} else {
+				return dev_err_probe(&pdev->dev, PTR_ERR(phy_handle),
+						     "Failed to map phy-handle for port %d", i);
+			}
+		}
+
+		if (i == EXTERNAL_PORT) {
+			phy_handle = of_parse_phandle(port_np, "phy-handle", 0);
+			if (phy_handle) {
+				/* register the phy board fixup */
+				ret = phy_register_fixup_for_uid(0x01410dd1, 0xffffffff,
+								 umac_phy_fixup);
+				if (ret)
+					dev_err(&pdev->dev, "cannot register phy board fixup\n");
+
+				err = of_get_phy_mode(phy_handle, &interface);
+				if (err)
+					interface = PHY_INTERFACE_MODE_NA;
+
+				umac->phy_dev = of_phy_connect(ndev, phy_handle,
+							       &umac_adjust_link,
+							       0, interface);
+
+				if (!umac->phy_dev)
+					return -ENODEV;
+
+				/* If the specified phy-handle has a fixed-link declaration, use the
+				 * fixed-link properties to set the configuration for the PHY
+				 */
+				if (of_phy_is_fixed_link(phy_handle)) {
+					struct device_node *fixed_link_node =
+						of_get_child_by_name(phy_handle,
+								     "fixed-link");
+
+					if (of_property_read_u32(fixed_link_node, "speed",
+								 &umac->phy_dev->speed)) {
+						netdev_err(ndev, "Invalid fixed-link specified.\n");
+						return -EINVAL;
+					}
+					umac->phy_dev->duplex =
+						of_property_read_bool(fixed_link_node,
+								      "full-duplex");
+					umac->phy_dev->pause =
+						of_property_read_bool(fixed_link_node,
+								      "pause");
+					umac->phy_dev->asym_pause =
+						of_property_read_bool(fixed_link_node,
+								      "asym-pause");
+					umac->phy_dev->autoneg = AUTONEG_DISABLE;
+					__clear_bit(ETHTOOL_LINK_MODE_Autoneg_BIT,
+						    umac->phy_dev->advertising);
+				}
+			} else {
+				return dev_err_probe(&pdev->dev, PTR_ERR(phy_handle),
+						     "Failed to map phy-handle for port %d", i);
+			}
+		}
+	}
+
+	return 0;
+}
+
+static int umac_probe(struct platform_device *pdev)
+{
+	struct umac_priv *umac;
+	struct net_device *ndev;
+	struct resource *res;
+	int ret = 0;
+
+	ndev = alloc_etherdev(sizeof(*umac));
+	if (!ndev)
+		return -ENOMEM;
+
+	SET_NETDEV_DEV(ndev, &pdev->dev);
+
+	umac = netdev_priv(ndev);
+	umac->pdev = pdev;
+	umac->ndev = ndev;
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!res) {
+		netdev_err(ndev, "failed to get I/O memory\n");
+		free_netdev(ndev);
+		return -ENXIO;
+	}
+
+	umac->base = devm_ioremap_resource(&pdev->dev, res);
+	if (!umac->base) {
+		netdev_err(ndev, "failed to remap I/O memory\n");
+		free_netdev(ndev);
+		return -EBUSY;
+	}
+
+	ndev->irq = platform_get_irq(pdev, 0);
+	if (ndev->irq < 0) {
+		netdev_err(ndev, "failed to get irq\n");
+		free_netdev(ndev);
+		return -ENXIO;
+	}
+
+	platform_set_drvdata(pdev, ndev);
+
+	ndev->netdev_ops = &umac_netdev_ops;
+	ndev->ethtool_ops = &umac_ethtool_ops;
+
+	umac_init_mac_address(ndev);
+	umac_channel_disable(umac);
+	ret = umac_setup_phy(ndev);
+	if (ret != 0) {
+		netdev_err(ndev, "failed to setup phy ret=%d\n", ret);
+		return -ENODEV;
+	}
+
+	umac->use_ncsi = false;
+	if (of_get_property(pdev->dev.of_node, "use-ncsi", NULL)) {
+		if (!IS_ENABLED(CONFIG_NET_NCSI)) {
+			netdev_err(ndev, "NCSI stack not enabled\n");
+			free_netdev(ndev);
+			return 0;
+		}
+
+		dev_info(&pdev->dev, "Using NCSI interface\n");
+		umac->use_ncsi = true;
+		umac->ncsidev = ncsi_register_dev(ndev, umac_ncsi_handler);
+		if (!umac->ncsidev) {
+			free_netdev(ndev);
+			return -ENODEV;
+		}
+	}
+
+	netif_napi_add(ndev, &umac->napi, umac_poll);
+	ret = register_netdev(ndev);
+	if (ret != 0) {
+		netdev_err(ndev, "failed to register UMAC ret=%d\n", ret);
+		netif_napi_del(&umac->napi);
+		free_netdev(ndev);
+		return -ENODEV;
+	}
+
+	return ret;
+}
+
+static int umac_remove(struct platform_device *pdev)
+{
+	struct net_device *ndev = platform_get_drvdata(pdev);
+	struct umac_priv *umac = netdev_priv(ndev);
+
+	unregister_netdev(ndev);
+	iounmap(umac->base);
+	free_netdev(ndev);
+	return 0;
+}
+
+static const struct of_device_id umac_of_matches[] = {
+	{ .compatible = "hpe, gxp-umac", },
+	{},
+};
+MODULE_DEVICE_TABLE(of, umac_of_matches);
+
+static struct platform_driver umac_driver = {
+	.driver	= {
+		.name    = "gxp-umac",
+		.of_match_table = of_match_ptr(umac_of_matches),
+	},
+	.probe   = umac_probe,
+	.remove  = umac_remove,
+};
+
+module_platform_driver(umac_driver);
+
+MODULE_AUTHOR("Nick Hawkins <nick.hawkins@hpe.com");
+MODULE_DESCRIPTION("HPE GXP UMAC driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/net/ethernet/hpe/gxp-umac.h b/drivers/net/ethernet/hpe/gxp-umac.h
new file mode 100644
index 000000000000..2d3c777dd5b5
--- /dev/null
+++ b/drivers/net/ethernet/hpe/gxp-umac.h
@@ -0,0 +1,89 @@ 
+/* SPDX-License-Identifier: GPL-2.0-only */
+/* Copyright (C) 2023 Hewlett-Packard Development Company, L.P. */
+
+#ifndef _UMAC_H_
+#define _UMAC_H_
+
+#define UMAC_CONFIG_STATUS	0x00        /* R/W Configuration and Status
+					     * Register I
+					     */
+#define UMAC_CFG_TXEN		0x00001000  // transmit enable
+#define UMAC_CFG_RXEN		0x00000800  // receive enable
+#define UMAC_CFG_GTX_CLK_EN	0x00000400  // gigabit clock enable
+#define UMAC_CFG_TX_CLK_EN	0x00000200  // 10/100 clock enable
+#define UMAC_CFG_GIGABIT_MODE	0x00000004  // MAC gigabit mode enable
+#define UMAC_CFG_FULL_DUPLEX	0x00000001  // enable ignoring of collisions
+#define UMAC_RING_PTR		0x04        //  R/W Ring Pointer Register
+#define UMAC_TX_RING_PTR_MASK	0x7FFF0000  // transmit ring entry pointer
+#define UMAC_TX_RING_PTR_SHIFT	16
+#define UMAC_RX_RING_PTR_MASK	0x00007FFF  // receive ring entry pointer
+#define UMAC_RX_RING_PTR_SHIFT	0
+
+#define UMAC_CLEAR_STATUS	0x0C	//  W   Clear Status Register
+#define UMAC_RING_PROMPT	0x08	//  W	Ring Prompt Register
+#define UMAC_CKSUM_CONFIG	0x10	//  R/W	Checksum Config Register
+#define UMAC_RING_SIZE		0x14	//  R/W	Ring Size Register
+#define UMAC_TX_RING_SIZE_MASK	0xFF000000
+#define UMAC_TX_RING_SIZE_SHIFT	24
+#define UMAC_RX_RING_SIZE_MASK	0x00FF0000
+#define UMAC_RX_RING_SIZE_SHIFT	16
+#define UMAC_LAST_RX_PKT_SIZE	0x0000FFFF
+#define UMAC_RING_SIZE_256	0x3F
+
+#define UMAC_MAC_ADDR_HI	0x18	//  R/W	MAC Address[47:32] Register
+#define UMAC_MAC_ADDR_MID	0x1C	//  R/W	MAC Address[31:16] Register
+#define UMAC_MAC_ADDR_LO	0x20	//  R/W	MAC Address[15:0] Register
+#define UMAC_INTERRUPT		0x30	//  R/W	MAC Interrupt Configuration
+					// and Status Register
+#define UMAC_RX_INTEN		0x00000008
+#define UMAC_RX_INT		0x00000004
+#define UMAC_TX_INTEN		0x00000002
+#define UMAC_TX_INT		0x00000001
+
+#define UMAC_RX_RING_ADDR	0x4C	//  R/W	Rx Ring Base Address Register
+#define UMAC_TX_RING_ADDR	0x50	//  R/W	Tx Ring Base Address Register
+#define UMAC_DMA_CONFIG		0x54	//  R/W	DMA Config Register
+
+#define UMAC_MAX_TX_DESC_ENTRIES	0x100	/* 256,number of ring buffer
+						 *  entries supported
+						 */
+#define UMAC_MAX_RX_DESC_ENTRIES	0x100
+#define UMAC_MAX_TX_FRAME_SIZE		0x600
+#define UMAC_MAX_RX_FRAME_SIZE		0x600
+
+// ring status masks
+#define UMAC_RING_ENTRY_HW_OWN		0x8000
+
+// maximum ethernet frame size
+#define UMAC_MIN_FRAME_SIZE		60    // excludes preable, sfd, and fcs
+#define UMAC_MAX_PAYLOAD_SIZE		1500
+#define UMAC_MAX_FRAME_SIZE		1514  // excludes preable, sfd, and fcs
+
+struct umac_rx_desc_entry {
+	u32  dmaaddress;   // Start address for DMA operationg
+	u16  status;       // Packet tx status and ownership flag
+	u16  count;        // Number of bytes received
+	u16  checksum;     // On-the-fly packet checksum
+	u16  control;      // Checksum-in-time flag
+	u32  reserved;
+} __aligned(16);
+
+struct umac_rx_descs {
+	struct umac_rx_desc_entry entrylist[UMAC_MAX_RX_DESC_ENTRIES];
+	u8 framelist[UMAC_MAX_RX_DESC_ENTRIES][UMAC_MAX_RX_FRAME_SIZE];
+} __packed;
+
+struct umac_tx_desc_entry {
+	u32  dmaaddress;   // Start address for DMA operationg
+	u16  status;       // Packet rx status, type, and ownership flag
+	u16  count;        // Number of bytes received
+	u32  cksumoffset;  // Specifies where to place packet checksum
+	u32  reserved;
+} __aligned(16);
+
+struct umac_tx_descs {
+	struct umac_tx_desc_entry entrylist[UMAC_MAX_TX_DESC_ENTRIES];
+	u8 framelist[UMAC_MAX_TX_DESC_ENTRIES][UMAC_MAX_TX_FRAME_SIZE];
+} __packed;
+
+#endif