mbox series

[v1,0/4] add dsa switch support for ar9331

Message ID 20191014061549.3669-1-o.rempel@pengutronix.de
Headers show
Series add dsa switch support for ar9331 | expand

Message

Oleksij Rempel Oct. 14, 2019, 6:15 a.m. UTC
This patch series provides dsa switch support for Atheros ar9331 WiSoC.
As side effect ag71xx needed to be ported to phylink to make the switch
driver (as well phylink based) work properly.

Oleksij Rempel (4):
  net: ag71xx: port to phylink
  dt-bindings: net: dsa: qca,ar9331 switch documentation
  MIPS: ath79: ar9331: add ar9331-switch node
  net: dsa: add support for Atheros AR9331 build-in switch

 .../devicetree/bindings/net/dsa/ar9331.txt    | 155 ++++
 arch/mips/boot/dts/qca/ar9331.dtsi            | 128 ++-
 arch/mips/boot/dts/qca/ar9331_dpt_module.dts  |  13 +
 drivers/net/dsa/Kconfig                       |   2 +
 drivers/net/dsa/Makefile                      |   1 +
 drivers/net/dsa/qca/Kconfig                   |  11 +
 drivers/net/dsa/qca/Makefile                  |   2 +
 drivers/net/dsa/qca/ar9331.c                  | 822 ++++++++++++++++++
 drivers/net/ethernet/atheros/Kconfig          |   2 +-
 drivers/net/ethernet/atheros/ag71xx.c         | 144 +--
 include/net/dsa.h                             |   2 +
 net/dsa/Kconfig                               |   6 +
 net/dsa/Makefile                              |   1 +
 net/dsa/tag_ar9331.c                          |  97 +++
 14 files changed, 1324 insertions(+), 62 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/net/dsa/ar9331.txt
 create mode 100644 drivers/net/dsa/qca/Kconfig
 create mode 100644 drivers/net/dsa/qca/Makefile
 create mode 100644 drivers/net/dsa/qca/ar9331.c
 create mode 100644 net/dsa/tag_ar9331.c

Comments

Andrew Lunn Oct. 16, 2019, 12:12 p.m. UTC | #1
On Mon, Oct 14, 2019 at 08:15:46AM +0200, Oleksij Rempel wrote:
> The port to phylink was done as close as possible to initial
> functionality.
> Theoretically this HW can support flow control, practically seems to be not
> enough to just enable it. So, more work should be done.
> 
> Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>

Hi Oleksij

Please include Russell King in Cc: in future.

> -static void ag71xx_phy_link_adjust(struct net_device *ndev)
> +static void ag71xx_mac_validate(struct phylink_config *config,
> +			    unsigned long *supported,
> +			    struct phylink_link_state *state)
>  {
> -	struct ag71xx *ag = netdev_priv(ndev);
> +	__ETHTOOL_DECLARE_LINK_MODE_MASK(mask) = { 0, };
> +
> +	if (state->interface != PHY_INTERFACE_MODE_NA &&
> +	    state->interface != PHY_INTERFACE_MODE_GMII &&
> +	    state->interface != PHY_INTERFACE_MODE_MII) {
> +		bitmap_zero(supported, __ETHTOOL_LINK_MODE_MASK_NBITS);
> +		return;
> +	}
> +
> +	phylink_set(mask, MII);
> +
> +	/* flow control is not supported */
>  
> -	ag71xx_link_adjust(ag, true);
> +	phylink_set(mask, 10baseT_Half);
> +	phylink_set(mask, 10baseT_Full);
> +	phylink_set(mask, 100baseT_Half);
> +	phylink_set(mask, 100baseT_Full);
> +
> +	phylink_set(mask, 1000baseT_Full);
> +	phylink_set(mask, 1000baseX_Full);

Can the MAC/PHY dynamically switch between MII and GMII? Maybe you
should only add 1G support when interface is GMII?

> @@ -1239,6 +1255,13 @@ static int ag71xx_open(struct net_device *ndev)
>  	unsigned int max_frame_len;
>  	int ret;
>  
> +	ret = phylink_of_phy_connect(ag->phylink, ag->pdev->dev.of_node, 0);
> +	if (ret) {
> +		netif_info(ag, link, ndev, "phylink_of_phy_connect filed with err: %i\n",
> +			   ret);

netif_info seems wrong. _err()?

	   Andrew
Oleksij Rempel Oct. 16, 2019, 12:24 p.m. UTC | #2
On Wed, Oct 16, 2019 at 02:12:16PM +0200, Andrew Lunn wrote:
> On Mon, Oct 14, 2019 at 08:15:46AM +0200, Oleksij Rempel wrote:
> > The port to phylink was done as close as possible to initial
> > functionality.
> > Theoretically this HW can support flow control, practically seems to be not
> > enough to just enable it. So, more work should be done.
> > 
> > Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
> 
> Hi Oleksij
> 
> Please include Russell King in Cc: in future.

He was included in To:. Do you mean, I need to move him from To to Cc?

> > -static void ag71xx_phy_link_adjust(struct net_device *ndev)
> > +static void ag71xx_mac_validate(struct phylink_config *config,
> > +			    unsigned long *supported,
> > +			    struct phylink_link_state *state)
> >  {
> > -	struct ag71xx *ag = netdev_priv(ndev);
> > +	__ETHTOOL_DECLARE_LINK_MODE_MASK(mask) = { 0, };
> > +
> > +	if (state->interface != PHY_INTERFACE_MODE_NA &&
> > +	    state->interface != PHY_INTERFACE_MODE_GMII &&
> > +	    state->interface != PHY_INTERFACE_MODE_MII) {
> > +		bitmap_zero(supported, __ETHTOOL_LINK_MODE_MASK_NBITS);
> > +		return;
> > +	}
> > +
> > +	phylink_set(mask, MII);
> > +
> > +	/* flow control is not supported */
> >  
> > -	ag71xx_link_adjust(ag, true);
> > +	phylink_set(mask, 10baseT_Half);
> > +	phylink_set(mask, 10baseT_Full);
> > +	phylink_set(mask, 100baseT_Half);
> > +	phylink_set(mask, 100baseT_Full);
> > +
> > +	phylink_set(mask, 1000baseT_Full);
> > +	phylink_set(mask, 1000baseX_Full);
> 
> Can the MAC/PHY dynamically switch between MII and GMII? Maybe you
> should only add 1G support when interface is GMII?

OK, good point.

> > @@ -1239,6 +1255,13 @@ static int ag71xx_open(struct net_device *ndev)
> >  	unsigned int max_frame_len;
> >  	int ret;
> >  
> > +	ret = phylink_of_phy_connect(ag->phylink, ag->pdev->dev.of_node, 0);
> > +	if (ret) {
> > +		netif_info(ag, link, ndev, "phylink_of_phy_connect filed with err: %i\n",
> > +			   ret);
> 
> netif_info seems wrong. _err()?

Yes, will fix it.

Regards,
Oleksij.
Andrew Lunn Oct. 16, 2019, 1 p.m. UTC | #3
> +++ b/drivers/net/dsa/qca/ar9331.c
> @@ -0,0 +1,822 @@
> +// SPDX-License-Identifier: GPL-2.0-only

I think C files should use /*  */, and header files //, for SPDX.

> +// Copyright (c) 2019 Pengutronix, Oleksij Rempel <kernel@pengutronix.de>
> +
> +#include <linux/bitfield.h>
> +#include <linux/module.h>
> +#include <linux/of_irq.h>
> +#include <linux/of_mdio.h>
> +#include <linux/regmap.h>
> +#include <linux/reset.h>
> +#include <net/dsa.h>
> +
> +#define AR9331_SW_NAME				"ar9331_switch"
> +#define AR9331_SW_PORTS				6
> +
> +/* dummy reg to change page */
> +#define AR9331_SW_REG_PAGE			BIT(18)
> +
> +/* Global Interrupt */
> +#define AR9331_SW_REG_GINT			0x10
> +#define AR9331_SW_REG_GINT_MASK			0x14
> +#define AR9331_SW_GINT_PHY_INT			BIT(2)
> +
> +#define AR9331_SW_REG_FLOOD_MASK		0x2c
> +#define AR9331_SW_FLOOD_MASK_BROAD_TO_CPU	BIT(26)
> +
> +#define AR9331_SW_REG_GLOBAL_CTRL		0x30
> +#define AR9331_SW_GLOBAL_CTRL_MFS_M		GENMASK(13, 0)
> +
> +#define AR9331_SW_REG_MDIO_CTRL			0x98
> +#define AR9331_SW_MDIO_CTRL_BUSY		BIT(31)
> +#define AR9331_SW_MDIO_CTRL_MASTER_EN		BIT(30)
> +#define AR9331_SW_MDIO_CTRL_CMD_READ		BIT(27)
> +#define AR9331_SW_MDIO_CTRL_PHY_ADDR_M		GENMASK(25, 21)
> +#define AR9331_SW_MDIO_CTRL_REG_ADDR_M		GENMASK(20, 16)
> +#define AR9331_SW_MDIO_CTRL_DATA_M		GENMASK(16, 0)
> +
> +#define AR9331_SW_REG_PORT_STATUS(_port)	(0x100 + (_port) * 0x100)
> +
> +/* FLOW_LINK_EN - enable mac flow control config auto-neg with phy.
> + * If not set, mac can be config by software.
> + */
> +#define AR9331_SW_PORT_STATUS_FLOW_LINK_EN	BIT(12)
> +
> +/* LINK_EN - If set, MAC is configured from PHY link status.
> + * If not set, MAC should be configured by software.
> + */
> +#define AR9331_SW_PORT_STATUS_LINK_EN		BIT(9)
> +#define AR9331_SW_PORT_STATUS_DUPLEX_MODE	BIT(6)
> +#define AR9331_SW_PORT_STATUS_RX_FLOW_EN	BIT(5)
> +#define AR9331_SW_PORT_STATUS_TX_FLOW_EN	BIT(4)
> +#define AR9331_SW_PORT_STATUS_RXMAC		BIT(3)
> +#define AR9331_SW_PORT_STATUS_TXMAC		BIT(2)
> +#define AR9331_SW_PORT_STATUS_SPEED_M		GENMASK(1, 0)
> +#define AR9331_SW_PORT_STATUS_SPEED_1000	2
> +#define AR9331_SW_PORT_STATUS_SPEED_100		1
> +#define AR9331_SW_PORT_STATUS_SPEED_10		0
> +
> +#define AR9331_SW_PORT_STATUS_MAC_MASK \
> +	(AR9331_SW_PORT_STATUS_TXMAC | AR9331_SW_PORT_STATUS_RXMAC)
> +
> +#define AR9331_SW_PORT_STATUS_LINK_MASK \
> +	(AR9331_SW_PORT_STATUS_LINK_EN | AR9331_SW_PORT_STATUS_FLOW_LINK_EN | \
> +	 AR9331_SW_PORT_STATUS_DUPLEX_MODE | \
> +	 AR9331_SW_PORT_STATUS_RX_FLOW_EN | AR9331_SW_PORT_STATUS_TX_FLOW_EN | \
> +	 AR9331_SW_PORT_STATUS_SPEED_M)
> +
> +/* Phy bypass mode
> + * ------------------------------------------------------------------------
> + * Bit:   | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |10 |11 |12 |13 |14 |15 |
> + *
> + * real   | start |   OP  | PhyAddr           |  Reg Addr         |  TA   |
> + * atheros| start |   OP  | 2'b00 |PhyAdd[2:0]|  Reg Addr[4:0]    |  TA   |
> + *
> + *
> + * Bit:   |16 |17 |18 |19 |20 |21 |22 |23 |24 |25 |26 |27 |28 |29 |30 |31 |
> + * real   |  Data                                                         |
> + * atheros|  Data                                                         |
> + *
> + * ------------------------------------------------------------------------
> + * Page address mode
> + * ------------------------------------------------------------------------
> + * Bit:   | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |10 |11 |12 |13 |14 |15 |
> + * real   | start |   OP  | PhyAddr           |  Reg Addr         |  TA   |
> + * atheros| start |   OP  | 2'b11 |                          8'b0 |  TA   |
> + *
> + * Bit:   |16 |17 |18 |19 |20 |21 |22 |23 |24 |25 |26 |27 |28 |29 |30 |31 |
> + * real   |  Data                                                         |
> + * atheros|                       | Page [9:0]                            |
> + */
> +/* In case of Page Address mode, Bit[18:9] of 32 bit register address should be
> + * written to bits[9:0] of mdio data register.
> + */
> +#define AR9331_SW_ADDR_PAGE			GENMASK(18, 9)
> +
> +/* ------------------------------------------------------------------------
> + * Normal register access mode
> + * ------------------------------------------------------------------------
> + * Bit:   | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |10 |11 |12 |13 |14 |15 |
> + * real   | start |   OP  | PhyAddr           |  Reg Addr         |  TA   |
> + * atheros| start |   OP  | 2'b10 |  low_addr[7:0]                |  TA   |
> + *
> + * Bit:   |16 |17 |18 |19 |20 |21 |22 |23 |24 |25 |26 |27 |28 |29 |30 |31 |
> + * real   |  Data                                                         |
> + * atheros|  Data                                                         |
> + * ------------------------------------------------------------------------
> + */
> +#define AR9331_SW_LOW_ADDR_PHY			GENMASK(8, 6)
> +#define AR9331_SW_LOW_ADDR_REG			GENMASK(5, 1)
> +
> +#define AR9331_SW_MDIO_PHY_MODE_M		GENMASK(4, 3)
> +#define AR9331_SW_MDIO_PHY_MODE_PAGE		3
> +#define AR9331_SW_MDIO_PHY_MODE_REG		2
> +#define AR9331_SW_MDIO_PHY_MODE_BYPASS		0
> +#define AR9331_SW_MDIO_PHY_ADDR_M		GENMASK(2, 0)
> +
> +/* Empirical determined values */
> +#define AR9331_SW_MDIO_POLL_SLEEP_US		1
> +#define AR9331_SW_MDIO_POLL_TIMEOUT_US		20
> +
> +struct ar9331_sw_priv {
> +	struct device *dev;
> +	struct dsa_switch *ds;
> +	struct dsa_switch_ops ops;
> +	struct irq_domain *irqdomain;
> +	struct mii_bus *mbus; /* mdio master */
> +	struct mii_bus *sbus; /* mdio slave */
> +	struct regmap *regmap;
> +	struct reset_control *sw_reset;
> +};
> +
> +/* Warning: switch reset will reset last AR9331_SW_MDIO_PHY_MODE_PAGE request
> + * If some kind of optimization is used, the request should be repeated.
> + */
> +static int ar9331_sw_reset(struct ar9331_sw_priv *priv)
> +{
> +	int ret;
> +
> +	ret = reset_control_assert(priv->sw_reset);
> +	if (ret)
> +		goto error;
> +
> +	/* AR9331 doc do not provide any information about proper reset
> +	 * sequence. The AR8136 (the closes switch to the AR9331) doc says:
> +	 * reset duration should be greater than 10ms. So, let's use this value
> +	 * for now.
> +	 */
> +	usleep_range(10000, 15000);
> +	ret = reset_control_deassert(priv->sw_reset);
> +	if (ret)
> +		goto error;

Any comments in the documentation about needing to wait for the reset
to complete?


> +
> +	return 0;
> +error:
> +	dev_err_ratelimited(priv->dev, "%s: %i\n", __func__, ret);
> +	return ret;
> +}
> +
> +static int ar9331_sw_mbus_write(struct mii_bus *mbus, int port, int regnum,
> +				u16 data)
> +{
> +	struct ar9331_sw_priv *priv = mbus->priv;
> +	struct regmap *regmap = priv->regmap;
> +	u32 val;
> +	int ret;
> +
> +	ret = regmap_write(regmap, AR9331_SW_REG_MDIO_CTRL,
> +			   AR9331_SW_MDIO_CTRL_BUSY |
> +			   AR9331_SW_MDIO_CTRL_MASTER_EN |
> +			   FIELD_PREP(AR9331_SW_MDIO_CTRL_PHY_ADDR_M, port) |
> +			   FIELD_PREP(AR9331_SW_MDIO_CTRL_REG_ADDR_M, regnum) |
> +			   FIELD_PREP(AR9331_SW_MDIO_CTRL_DATA_M, data));
> +	if (ret)
> +		goto error;
> +
> +	ret = regmap_read_poll_timeout(regmap, AR9331_SW_REG_MDIO_CTRL, val,
> +				       !(val & AR9331_SW_MDIO_CTRL_BUSY),
> +				       AR9331_SW_MDIO_POLL_SLEEP_US,
> +				       AR9331_SW_MDIO_POLL_TIMEOUT_US);
> +	if (ret)
> +		goto error;
> +
> +	return 0;
> +error:
> +	dev_err_ratelimited(priv->dev, "PHY write error: %i\n", ret);
> +	return ret;
> +}
> +
> +static int ar9331_sw_mbus_read(struct mii_bus *mbus, int port, int regnum)
> +{
> +	struct ar9331_sw_priv *priv = mbus->priv;
> +	struct regmap *regmap = priv->regmap;
> +	u32 val;
> +	int ret;
> +
> +	ret = regmap_write(regmap, AR9331_SW_REG_MDIO_CTRL,
> +			   AR9331_SW_MDIO_CTRL_BUSY |
> +			   AR9331_SW_MDIO_CTRL_MASTER_EN |
> +			   AR9331_SW_MDIO_CTRL_CMD_READ |
> +			   FIELD_PREP(AR9331_SW_MDIO_CTRL_PHY_ADDR_M, port) |
> +			   FIELD_PREP(AR9331_SW_MDIO_CTRL_REG_ADDR_M, regnum));
> +	if (ret)
> +		goto error;
> +
> +	ret = regmap_read_poll_timeout(regmap, AR9331_SW_REG_MDIO_CTRL, val,
> +				       !(val & AR9331_SW_MDIO_CTRL_BUSY),
> +				       AR9331_SW_MDIO_POLL_SLEEP_US,
> +				       AR9331_SW_MDIO_POLL_TIMEOUT_US);
> +	if (ret)
> +		goto error;
> +
> +	ret = regmap_read(regmap, AR9331_SW_REG_MDIO_CTRL, &val);
> +	if (ret)
> +		goto error;
> +
> +	return FIELD_GET(AR9331_SW_MDIO_CTRL_DATA_M, val);
> +
> +error:
> +	dev_err_ratelimited(priv->dev, "PHY read error: %i\n", ret);
> +	return ret;
> +}
> +
> +static int ar9331_sw_mbus_init(struct ar9331_sw_priv *priv)
> +{
> +	struct device *dev = priv->dev;
> +	static struct mii_bus *mbus;
> +	struct device_node *np, *mnp;
> +	int ret;
> +
> +	np = dev->of_node;
> +
> +	mbus = devm_mdiobus_alloc(dev);
> +	if (!mbus)
> +		return -ENOMEM;
> +
> +	mbus->name = np->full_name;
> +	snprintf(mbus->id, MII_BUS_ID_SIZE, "%pOF", np);
> +
> +	mbus->read = ar9331_sw_mbus_read;
> +	mbus->write = ar9331_sw_mbus_write;
> +	mbus->priv = priv;
> +	mbus->parent = dev;
> +
> +	mnp = of_get_child_by_name(np, "mdio");

You should check if mnp is NULL. You want it to mandatory. The current
code will pass NULL to of_mdiobus_register(), which is legal, and it
will look one level higher for the PHYs.


> +	ret = of_mdiobus_register(mbus, mnp);
> +	of_node_put(mnp);
> +	if (ret)
> +		return ret;
> +
> +	priv->mbus = mbus;
> +
> +	return 0;
> +}
> +
> +static int ar9331_sw_setup(struct dsa_switch *ds)
> +{
> +	struct ar9331_sw_priv *priv = (struct ar9331_sw_priv *)ds->priv;
> +	struct regmap *regmap = priv->regmap;
> +	int ret;
> +
> +	ret = ar9331_sw_reset(priv);
> +	if (ret)
> +		return ret;
> +
> +	/* Reset will set proper defaults. CPU - Port0 will be enabled and
> +	 * configured. All other ports (ports 1 - 5) are disabled
> +	 */

Nice, some hardware engineer thought about that.

> +	ret = ar9331_sw_mbus_init(priv);
> +	if (ret)
> +		return ret;
> +
> +	/* Do not drop broadcast frames */
> +	ret = regmap_write_bits(regmap, AR9331_SW_REG_FLOOD_MASK,
> +				AR9331_SW_FLOOD_MASK_BROAD_TO_CPU,
> +				AR9331_SW_FLOOD_MASK_BROAD_TO_CPU);
> +	if (ret)
> +		goto error;
> +
> +	/* Sync max frame size with value used in
> +	 * drivers/net/ethernet/atheros/ag71xx.c for ar9330 SoC.
> +	 * TODO: In both drivers this value seems to be not real maximal size
> +	 * The switch is able to configure 0x3fff and ethernet controller
> +	 * 0xffff. Are there any better way to sync this values?
> +	 */
> +	ret = regmap_write_bits(regmap, AR9331_SW_REG_GLOBAL_CTRL,
> +				AR9331_SW_GLOBAL_CTRL_MFS_M,
> +				FIELD_PREP(AR9331_SW_GLOBAL_CTRL_MFS_M, 1540));
> +	if (ret)
> +		goto error;

Jumbo is not so easy. I would avoid the plain number 1540. There
should be a #define. Also, you might want to allow space for a VLAN
header? Does enabbling jumbo have a performance impact? If not, you
can configure the switch to its maximum size.

> +
> +	return 0;
> +error:
> +	dev_err_ratelimited(priv->dev, "%s: %i\n", __func__, ret);
> +	return ret;
> +}
> +
> +static int ar9331_sw_port_enable(struct dsa_switch *ds, int port,
> +				 struct phy_device *phy)
> +{
> +	struct ar9331_sw_priv *priv = (struct ar9331_sw_priv *)ds->priv;
> +	struct regmap *regmap = priv->regmap;
> +	int ret;
> +
> +	/* nothing to enable. Just set link to initial state */
> +	ret = regmap_write(regmap, AR9331_SW_REG_PORT_STATUS(port), 0);
> +	if (ret)
> +		dev_err_ratelimited(priv->dev, "%s: %i\n", __func__, ret);
> +
> +	return ret;
> +}
> +
> +static void ar9331_sw_port_disable(struct dsa_switch *ds, int port)
> +{
> +	struct ar9331_sw_priv *priv = (struct ar9331_sw_priv *)ds->priv;
> +	struct regmap *regmap = priv->regmap;
> +	int ret;
> +
> +	ret = regmap_write(regmap, AR9331_SW_REG_PORT_STATUS(port), 0);
> +	if (ret)
> +		dev_err_ratelimited(priv->dev, "%s: %i\n", __func__, ret);
> +}

port_enable() and port_disable() look the same?

> +
> +static enum dsa_tag_protocol ar9331_sw_get_tag_protocol(struct dsa_switch *ds,
> +							int port)
> +{
> +	return DSA_TAG_PROTO_AR9331;
> +}
> +
> +static void ar9331_sw_phylink_validate(struct dsa_switch *ds, int port,
> +				       unsigned long *supported,
> +				       struct phylink_link_state *state)
> +{
> +	__ETHTOOL_DECLARE_LINK_MODE_MASK(mask) = { 0, };
> +
> +	switch (port) {
> +	case 0:
> +		if (state->interface != PHY_INTERFACE_MODE_GMII)
> +			goto unsupported;
> +
> +		phylink_set(mask, 1000baseT_Full);
> +		phylink_set(mask, 1000baseT_Half);
> +		break;
> +	case 1:
> +	case 2:
> +	case 3:
> +	case 4:
> +	case 5:
> +		if (state->interface != PHY_INTERFACE_MODE_INTERNAL)
> +			goto unsupported;
> +		break;
> +	default:
> +		bitmap_zero(supported, __ETHTOOL_LINK_MODE_MASK_NBITS);
> +		dev_err(ds->dev, "Unsupported port: %i\n", port);
> +		return;
> +	}
> +
> +	phylink_set_port_modes(mask);
> +	phylink_set(mask, Pause);
> +	phylink_set(mask, Asym_Pause);
> +
> +	phylink_set(mask, 10baseT_Half);
> +	phylink_set(mask, 10baseT_Full);
> +	phylink_set(mask, 100baseT_Half);
> +	phylink_set(mask, 100baseT_Full);

So the CPU port is 1G capable. All the other ports are only Fast Ethernet?

> +
> +	bitmap_and(supported, supported, mask,
> +		   __ETHTOOL_LINK_MODE_MASK_NBITS);
> +	bitmap_and(state->advertising, state->advertising, mask,
> +		   __ETHTOOL_LINK_MODE_MASK_NBITS);
> +
> +	return;
> +
> +unsupported:
> +	bitmap_zero(supported, __ETHTOOL_LINK_MODE_MASK_NBITS);
> +	dev_err(ds->dev, "Unsupported interface: %d, port: %d\n",
> +		state->interface, port);
> +}
> +
> +static void ar9331_sw_phylink_mac_config(struct dsa_switch *ds, int port,
> +					 unsigned int mode,
> +					 const struct phylink_link_state *state)
> +{
> +	struct ar9331_sw_priv *priv = (struct ar9331_sw_priv *)ds->priv;
> +	struct regmap *regmap = priv->regmap;
> +	int ret;
> +	u32 val;
> +
> +	switch (state->speed) {
> +	case SPEED_1000:
> +		val = AR9331_SW_PORT_STATUS_SPEED_1000;
> +		break;
> +	case SPEED_100:
> +		val = AR9331_SW_PORT_STATUS_SPEED_100;
> +		break;
> +	case SPEED_10:
> +		val = AR9331_SW_PORT_STATUS_SPEED_10;
> +		break;
> +	default:
> +		return;
> +	}
> +
> +	if (state->duplex)
> +		val |= AR9331_SW_PORT_STATUS_DUPLEX_MODE;
> +
> +	if (state->pause & MLO_PAUSE_TX)
> +		val |= AR9331_SW_PORT_STATUS_TX_FLOW_EN;
> +
> +	if (state->pause & MLO_PAUSE_RX)
> +		val |= AR9331_SW_PORT_STATUS_RX_FLOW_EN;
> +
> +	ret = regmap_update_bits(regmap, AR9331_SW_REG_PORT_STATUS(port),
> +				 AR9331_SW_PORT_STATUS_LINK_MASK, val);
> +	if (ret)
> +		dev_err_ratelimited(priv->dev, "%s: %i\n", __func__, ret);
> +}
> +
> +static void ar9331_sw_phylink_mac_link_down(struct dsa_switch *ds, int port,
> +					    unsigned int mode,
> +					    phy_interface_t interface)
> +{
> +	struct ar9331_sw_priv *priv = (struct ar9331_sw_priv *)ds->priv;
> +	struct regmap *regmap = priv->regmap;
> +	int ret;
> +
> +	ret = regmap_update_bits(regmap, AR9331_SW_REG_PORT_STATUS(port),
> +				 AR9331_SW_PORT_STATUS_MAC_MASK, 0);
> +	if (ret)
> +		dev_err_ratelimited(priv->dev, "%s: %i\n", __func__, ret);
> +}
> +
> +static void ar9331_sw_phylink_mac_link_up(struct dsa_switch *ds, int port,
> +					  unsigned int mode,
> +					  phy_interface_t interface,
> +					  struct phy_device *phydev)
> +{
> +	struct ar9331_sw_priv *priv = (struct ar9331_sw_priv *)ds->priv;
> +	struct regmap *regmap = priv->regmap;
> +	int ret;
> +
> +	ret = regmap_update_bits(regmap, AR9331_SW_REG_PORT_STATUS(port),
> +				 AR9331_SW_PORT_STATUS_MAC_MASK,
> +				 AR9331_SW_PORT_STATUS_MAC_MASK);
> +	if (ret)
> +		dev_err_ratelimited(priv->dev, "%s: %i\n", __func__, ret);
> +}
> +
> +static const struct dsa_switch_ops ar9331_sw_ops = {
> +	.get_tag_protocol = ar9331_sw_get_tag_protocol,
> +	.setup = ar9331_sw_setup,
> +	.port_enable = ar9331_sw_port_enable,
> +	.port_disable = ar9331_sw_port_disable,
> +	.phylink_validate	= ar9331_sw_phylink_validate,
> +	.phylink_mac_config	= ar9331_sw_phylink_mac_config,
> +	.phylink_mac_link_down	= ar9331_sw_phylink_mac_link_down,
> +	.phylink_mac_link_up	= ar9331_sw_phylink_mac_link_up,
> +};
> +
> +static irqreturn_t ar9331_sw_irq(int irq, void *data)
> +{
> +	struct ar9331_sw_priv *priv = data;
> +	struct regmap *regmap = priv->regmap;
> +	u32 stat;
> +	int ret;
> +
> +	ret = regmap_read(regmap, AR9331_SW_REG_GINT, &stat);
> +	if (ret) {
> +		dev_err(priv->dev, "can't read interrupt status\n");
> +		return IRQ_NONE;
> +	}
> +
> +	if (!stat)
> +		return IRQ_NONE;
> +
> +	if (stat & AR9331_SW_GINT_PHY_INT) {
> +		int child_irq;
> +
> +		child_irq = irq_find_mapping(priv->irqdomain, 0);
> +		handle_nested_irq(child_irq);
> +	}
> +
> +	ret = regmap_write(regmap, AR9331_SW_REG_GINT, stat);
> +	if (ret) {
> +		dev_err(priv->dev, "can't write interrupt status\n");
> +		return IRQ_NONE;
> +	}
> +
> +	return IRQ_HANDLED;
> +}
> +
> +static void ar9331_sw_mask_irq(struct irq_data *d)
> +{
> +	struct ar9331_sw_priv *priv = irq_data_get_irq_chip_data(d);
> +	struct regmap *regmap = priv->regmap;
> +	int ret;
> +
> +	ret = regmap_update_bits(regmap, AR9331_SW_REG_GINT_MASK,
> +				 AR9331_SW_GINT_PHY_INT, 0);
> +	if (ret)
> +		dev_err(priv->dev, "could not mask IRQ\n");
> +}
> +
> +static void ar9331_sw_unmask_irq(struct irq_data *d)
> +{
> +	struct ar9331_sw_priv *priv = irq_data_get_irq_chip_data(d);
> +	struct regmap *regmap = priv->regmap;
> +	int ret;
> +
> +	ret = regmap_update_bits(regmap, AR9331_SW_REG_GINT_MASK,
> +				 AR9331_SW_GINT_PHY_INT,
> +				 AR9331_SW_GINT_PHY_INT);
> +	if (ret)
> +		dev_err(priv->dev, "could not unmask IRQ\n");
> +}
> +
> +static struct irq_chip ar9331_sw_irq_chip = {
> +	.name = AR9331_SW_NAME,
> +	.irq_mask = ar9331_sw_mask_irq,
> +	.irq_unmask = ar9331_sw_unmask_irq,
> +};
> +
> +static int ar9331_sw_irq_map(struct irq_domain *domain, unsigned int irq,
> +			     irq_hw_number_t hwirq)
> +{
> +	irq_set_chip_data(irq, domain->host_data);
> +	irq_set_chip_and_handler(irq, &ar9331_sw_irq_chip, handle_simple_irq);
> +	irq_set_nested_thread(irq, 1);
> +	irq_set_noprobe(irq);
> +
> +	return 0;
> +}
> +
> +static void ar9331_sw_irq_unmap(struct irq_domain *d, unsigned int irq)
> +{
> +	irq_set_nested_thread(irq, 0);
> +	irq_set_chip_and_handler(irq, NULL, NULL);
> +	irq_set_chip_data(irq, NULL);
> +}
> +
> +static const struct irq_domain_ops ar9331_sw_irqdomain_ops = {
> +	.map = ar9331_sw_irq_map,
> +	.unmap = ar9331_sw_irq_unmap,
> +	.xlate = irq_domain_xlate_onecell,
> +};
> +
> +static int ar9331_sw_irq_init(struct ar9331_sw_priv *priv)
> +{
> +	struct device_node *np = priv->dev->of_node;
> +	struct device *dev = priv->dev;
> +	int ret, irq;
> +
> +	irq = of_irq_get(np, 0);
> +	if (irq <= 0) {
> +		dev_err(dev, "failed to get parent IRQ\n");
> +		return irq ? irq : -EINVAL;
> +	}
> +
> +	ret = devm_request_threaded_irq(dev, irq, NULL, ar9331_sw_irq,
> +					IRQF_ONESHOT, AR9331_SW_NAME, priv);
> +	if (ret) {
> +		dev_err(dev, "unable to request irq: %d\n", ret);
> +		return ret;
> +	}
> +
> +	priv->irqdomain = irq_domain_add_linear(np, 1, &ar9331_sw_irqdomain_ops,
> +						priv);
> +	if (!priv->irqdomain) {
> +		dev_err(dev, "failed to create IRQ domain\n");
> +		return -EINVAL;
> +	}
> +
> +	irq_set_parent(irq_create_mapping(priv->irqdomain, 0), irq);
> +
> +	return 0;
> +}
> +
> +static int __ar9331_mdio_write(struct mii_bus *sbus, u8 mode, u16 reg, u16 val)
> +{
> +	u8 r, p;
> +
> +	p = FIELD_PREP(AR9331_SW_MDIO_PHY_MODE_M, mode) |
> +		FIELD_GET(AR9331_SW_LOW_ADDR_PHY, reg);
> +	r = FIELD_GET(AR9331_SW_LOW_ADDR_REG, reg);
> +
> +	return sbus->write(sbus, p, r, val);

Why not use the mdiobus_write() and mdiobus_read()?

> +static int ar9331_sw_probe(struct mdio_device *mdiodev)
> +{
> +	struct ar9331_sw_priv *priv;
> +	int ret;
> +
> +	/* allocate the private data struct so that we can probe the switches
> +	 * ID register
> +	 */

I don't see the code actually getting the ID register?


> +	priv = devm_kzalloc(&mdiodev->dev, sizeof(*priv), GFP_KERNEL);
> +	if (!priv)
> +		return -ENOMEM;
> +
> +	priv->regmap = devm_regmap_init(&mdiodev->dev, &ar9331_sw_bus, priv,
> +					&ar9331_mdio_regmap_config);
> +	if (IS_ERR(priv->regmap)) {
> +		ret = PTR_ERR(priv->regmap);
> +		dev_err(&mdiodev->dev, "regmap init failed: %d\n", ret);
> +		return ret;
> +	}
> +
> +	priv->sw_reset = devm_reset_control_get(&mdiodev->dev, "switch");
> +	if (IS_ERR(priv->sw_reset)) {
> +		dev_err(&mdiodev->dev, "missing switch reset\n");
> +		return PTR_ERR(priv->sw_reset);
> +	}
> +
> +	priv->sbus = mdiodev->bus;
> +	priv->dev = &mdiodev->dev;
> +
> +	ret = ar9331_sw_irq_init(priv);
> +	if (ret)
> +		return ret;
> +
> +	priv->ds = dsa_switch_alloc(&mdiodev->dev, AR9331_SW_PORTS);
> +	if (!priv->ds)
> +		return -ENOMEM;
> +
> +	priv->ds->priv = priv;
> +	priv->ops = ar9331_sw_ops;
> +	priv->ds->ops = &priv->ops;
> +	dev_set_drvdata(&mdiodev->dev, priv);
> +
> +	return dsa_register_switch(priv->ds);
> +}
> +
> +static void ar9331_sw_remove(struct mdio_device *mdiodev)
> +{
> +	struct ar9331_sw_priv *priv = dev_get_drvdata(&mdiodev->dev);
> +
> +	mdiobus_unregister(priv->mbus);
> +	dsa_unregister_switch(priv->ds);
> +
> +	reset_control_assert(priv->sw_reset);
> +}
> +
> +static const struct of_device_id ar9331_sw_of_match[] = {
> +	{ .compatible = "qca,ar9331-switch" },
> +	{ },
> +};
> +
> +static struct mdio_driver ar9331_sw_mdio_driver = {
> +	.probe = ar9331_sw_probe,
> +	.remove = ar9331_sw_remove,
> +	.mdiodrv.driver = {
> +		.name = AR9331_SW_NAME,
> +		.of_match_table = ar9331_sw_of_match,
> +	},
> +};
> +
> +mdio_module_driver(ar9331_sw_mdio_driver);
> +
> +MODULE_AUTHOR("Oleksij Rempel <kernel@pengutronix.de>");
> +MODULE_DESCRIPTION("Driver for Atheros AR9331 switch");
> +MODULE_LICENSE("GPL v2");
> diff --git a/include/net/dsa.h b/include/net/dsa.h
> index 541fb514e31d..89a334e68d42 100644
> --- a/include/net/dsa.h
> +++ b/include/net/dsa.h
> @@ -42,6 +42,7 @@ struct phylink_link_state;
>  #define DSA_TAG_PROTO_8021Q_VALUE		12
>  #define DSA_TAG_PROTO_SJA1105_VALUE		13
>  #define DSA_TAG_PROTO_KSZ8795_VALUE		14
> +#define DSA_TAG_PROTO_AR9331_VALUE		15
>  
>  enum dsa_tag_protocol {
>  	DSA_TAG_PROTO_NONE		= DSA_TAG_PROTO_NONE_VALUE,
> @@ -59,6 +60,7 @@ enum dsa_tag_protocol {
>  	DSA_TAG_PROTO_8021Q		= DSA_TAG_PROTO_8021Q_VALUE,
>  	DSA_TAG_PROTO_SJA1105		= DSA_TAG_PROTO_SJA1105_VALUE,
>  	DSA_TAG_PROTO_KSZ8795		= DSA_TAG_PROTO_KSZ8795_VALUE,
> +	DSA_TAG_PROTO_AR9331		= DSA_TAG_PROTO_AR9331_VALUE,
>  };
>  
>  struct packet_type;
> diff --git a/net/dsa/Kconfig b/net/dsa/Kconfig
> index 29e2bd5cc5af..6e015309a7fe 100644
> --- a/net/dsa/Kconfig
> +++ b/net/dsa/Kconfig
> @@ -107,4 +107,10 @@ config NET_DSA_TAG_TRAILER
>  	  Say Y or M if you want to enable support for tagging frames at
>  	  with a trailed. e.g. Marvell 88E6060.
>  
> +config NET_DSA_TAG_AR9331
> +	tristate "Tag driver for Atheros AR9331 SoC with build-in switch"
> +	help
> +	  Say Y or M if you want to enable support for tagging frames for
> +	  the Atheros AR9331 SoC with build-in switch.
> +

These are somewhat sorted, based on the tristate string. So this
should go before NET_DSA_TAG_BRCM_COMMON.


>  endif
> diff --git a/net/dsa/Makefile b/net/dsa/Makefile
> index 2c6d286f0511..67caebf602be 100644
> --- a/net/dsa/Makefile
> +++ b/net/dsa/Makefile
> @@ -15,3 +15,4 @@ obj-$(CONFIG_NET_DSA_TAG_MTK) += tag_mtk.o
>  obj-$(CONFIG_NET_DSA_TAG_QCA) += tag_qca.o
>  obj-$(CONFIG_NET_DSA_TAG_SJA1105) += tag_sja1105.o
>  obj-$(CONFIG_NET_DSA_TAG_TRAILER) += tag_trailer.o
> +obj-$(CONFIG_NET_DSA_TAG_AR9331) += tag_ar9331.o

Please keep with the sorting.

> diff --git a/net/dsa/tag_ar9331.c b/net/dsa/tag_ar9331.c
> new file mode 100644
> index 000000000000..b32a8d3d48b9
> --- /dev/null
> +++ b/net/dsa/tag_ar9331.c
> @@ -0,0 +1,97 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * Copyright (c) 2019 Pengutronix, Oleksij Rempel <kernel@pengutronix.de>
> + */
> +
> +
> +#include <linux/bitfield.h>
> +#include <linux/etherdevice.h>
> +
> +#include "dsa_priv.h"
> +
> +#define AR9331_HDR_LEN			2
> +#define AR9331_HDR_VERSION		1
> +
> +#define AR9331_HDR_VERSION_MASK		GENMASK(15, 14)
> +#define AR9331_HDR_PRIORITY_MASK	GENMASK(13, 12)
> +#define AR9331_HDR_TYPE_MASK		GENMASK(10, 8)
> +#define AR9331_HDR_BROADCAST		BIT(7)
> +#define AR9331_HDR_FROM_CPU		BIT(6)
> +/* AR9331_HDR_RESERVED - not used or may be version filed.

field

> + * According to the AR8216 doc it should 0b10. On AR9331 it is 0b11 on RX path
> + * and should be set to 0b11 to make it work.
> + */
> +#define AR9331_HDR_RESERVED_MASK	GENMASK(5, 4)
> +#define AR9331_HDR_PORT_NUM_MASK	GENMASK(3, 0)
> +
> +static struct sk_buff *ar9331_tag_xmit(struct sk_buff *skb,
> +				       struct net_device *dev)
> +{
> +	struct dsa_port *dp = dsa_slave_to_port(dev);
> +	__le16 *phdr;
> +	u16 hdr;
> +
> +	if (skb_cow_head(skb, 0) < 0)
> +		return NULL;
> +
> +	phdr = skb_push(skb, AR9331_HDR_LEN);
> +
> +	hdr = FIELD_PREP(AR9331_HDR_VERSION_MASK, AR9331_HDR_VERSION);
> +	hdr |= AR9331_HDR_FROM_CPU | dp->index;
> +	/* 0b10 for AR8216 and 0b11 for AR9331 */
> +	hdr |= AR9331_HDR_RESERVED_MASK;
> +
> +	phdr[0] = cpu_to_le16(hdr);
> +
> +	return skb;
> +}
> +
> +static struct sk_buff *ar9331_tag_rcv(struct sk_buff *skb,
> +				      struct net_device *ndev,
> +				      struct packet_type *pt)
> +{
> +	u8 ver, port;
> +	u16 hdr;
> +
> +	if (unlikely(!pskb_may_pull(skb, AR9331_HDR_LEN)))
> +		return NULL;
> +
> +	hdr = le16_to_cpu(*(__le16 *)skb_mac_header(skb));
> +
> +	ver = FIELD_GET(AR9331_HDR_VERSION_MASK, hdr);
> +	if (unlikely(ver != AR9331_HDR_VERSION)) {
> +		netdev_warn(ndev, "%s:%i wrong header version 0x%2x\n",
> +			    __func__, __LINE__, hdr);

This would should probably be rate limited.

> +		return NULL;
> +	}
> +
> +	if (unlikely(hdr & AR9331_HDR_FROM_CPU)) {
> +		netdev_warn(ndev, "%s:%i packet should not be from cpu 0x%2x\n",
> +			    __func__, __LINE__, hdr);

This as well.

     Andrew
Andrew Lunn Oct. 16, 2019, 1:32 p.m. UTC | #4
On Wed, Oct 16, 2019 at 02:24:01PM +0200, Oleksij Rempel wrote:
> On Wed, Oct 16, 2019 at 02:12:16PM +0200, Andrew Lunn wrote:
> > On Mon, Oct 14, 2019 at 08:15:46AM +0200, Oleksij Rempel wrote:
> > > The port to phylink was done as close as possible to initial
> > > functionality.
> > > Theoretically this HW can support flow control, practically seems to be not
> > > enough to just enable it. So, more work should be done.
> > > 
> > > Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
> > 
> > Hi Oleksij
> > 
> > Please include Russell King in Cc: in future.
> 
> He was included in To:. Do you mean, I need to move him from To to Cc?

Ah, sorry. Missed him among all the other To:

I don't know if there are any strict rules, but i tend to use To: for
the maintainer you expect to merge the patch, and Cc: for everybody
else, and the lists.

    Andrew
Oleksij Rempel Oct. 16, 2019, 2:07 p.m. UTC | #5
On Wed, Oct 16, 2019 at 03:00:57PM +0200, Andrew Lunn wrote:
> > +++ b/drivers/net/dsa/qca/ar9331.c
> > @@ -0,0 +1,822 @@
> > +// SPDX-License-Identifier: GPL-2.0-only
> 
> I think C files should use /*  */, and header files //, for SPDX.

OK. In fact, there are both variants present in Kernel.
I'll do what ever format the maintainer think is acceptable.

> > +// Copyright (c) 2019 Pengutronix, Oleksij Rempel <kernel@pengutronix.de>
> > +
> > +#include <linux/bitfield.h>
> > +#include <linux/module.h>
> > +#include <linux/of_irq.h>
> > +#include <linux/of_mdio.h>
> > +#include <linux/regmap.h>
> > +#include <linux/reset.h>
> > +#include <net/dsa.h>
> > +
> > +#define AR9331_SW_NAME				"ar9331_switch"
> > +#define AR9331_SW_PORTS				6
> > +
> > +/* dummy reg to change page */
> > +#define AR9331_SW_REG_PAGE			BIT(18)
> > +
> > +/* Global Interrupt */
> > +#define AR9331_SW_REG_GINT			0x10
> > +#define AR9331_SW_REG_GINT_MASK			0x14
> > +#define AR9331_SW_GINT_PHY_INT			BIT(2)
> > +
> > +#define AR9331_SW_REG_FLOOD_MASK		0x2c
> > +#define AR9331_SW_FLOOD_MASK_BROAD_TO_CPU	BIT(26)
> > +
> > +#define AR9331_SW_REG_GLOBAL_CTRL		0x30
> > +#define AR9331_SW_GLOBAL_CTRL_MFS_M		GENMASK(13, 0)
> > +
> > +#define AR9331_SW_REG_MDIO_CTRL			0x98
> > +#define AR9331_SW_MDIO_CTRL_BUSY		BIT(31)
> > +#define AR9331_SW_MDIO_CTRL_MASTER_EN		BIT(30)
> > +#define AR9331_SW_MDIO_CTRL_CMD_READ		BIT(27)
> > +#define AR9331_SW_MDIO_CTRL_PHY_ADDR_M		GENMASK(25, 21)
> > +#define AR9331_SW_MDIO_CTRL_REG_ADDR_M		GENMASK(20, 16)
> > +#define AR9331_SW_MDIO_CTRL_DATA_M		GENMASK(16, 0)
> > +
> > +#define AR9331_SW_REG_PORT_STATUS(_port)	(0x100 + (_port) * 0x100)
> > +
> > +/* FLOW_LINK_EN - enable mac flow control config auto-neg with phy.
> > + * If not set, mac can be config by software.
> > + */
> > +#define AR9331_SW_PORT_STATUS_FLOW_LINK_EN	BIT(12)
> > +
> > +/* LINK_EN - If set, MAC is configured from PHY link status.
> > + * If not set, MAC should be configured by software.
> > + */
> > +#define AR9331_SW_PORT_STATUS_LINK_EN		BIT(9)
> > +#define AR9331_SW_PORT_STATUS_DUPLEX_MODE	BIT(6)
> > +#define AR9331_SW_PORT_STATUS_RX_FLOW_EN	BIT(5)
> > +#define AR9331_SW_PORT_STATUS_TX_FLOW_EN	BIT(4)
> > +#define AR9331_SW_PORT_STATUS_RXMAC		BIT(3)
> > +#define AR9331_SW_PORT_STATUS_TXMAC		BIT(2)
> > +#define AR9331_SW_PORT_STATUS_SPEED_M		GENMASK(1, 0)
> > +#define AR9331_SW_PORT_STATUS_SPEED_1000	2
> > +#define AR9331_SW_PORT_STATUS_SPEED_100		1
> > +#define AR9331_SW_PORT_STATUS_SPEED_10		0
> > +
> > +#define AR9331_SW_PORT_STATUS_MAC_MASK \
> > +	(AR9331_SW_PORT_STATUS_TXMAC | AR9331_SW_PORT_STATUS_RXMAC)
> > +
> > +#define AR9331_SW_PORT_STATUS_LINK_MASK \
> > +	(AR9331_SW_PORT_STATUS_LINK_EN | AR9331_SW_PORT_STATUS_FLOW_LINK_EN | \
> > +	 AR9331_SW_PORT_STATUS_DUPLEX_MODE | \
> > +	 AR9331_SW_PORT_STATUS_RX_FLOW_EN | AR9331_SW_PORT_STATUS_TX_FLOW_EN | \
> > +	 AR9331_SW_PORT_STATUS_SPEED_M)
> > +
> > +/* Phy bypass mode
> > + * ------------------------------------------------------------------------
> > + * Bit:   | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |10 |11 |12 |13 |14 |15 |
> > + *
> > + * real   | start |   OP  | PhyAddr           |  Reg Addr         |  TA   |
> > + * atheros| start |   OP  | 2'b00 |PhyAdd[2:0]|  Reg Addr[4:0]    |  TA   |
> > + *
> > + *
> > + * Bit:   |16 |17 |18 |19 |20 |21 |22 |23 |24 |25 |26 |27 |28 |29 |30 |31 |
> > + * real   |  Data                                                         |
> > + * atheros|  Data                                                         |
> > + *
> > + * ------------------------------------------------------------------------
> > + * Page address mode
> > + * ------------------------------------------------------------------------
> > + * Bit:   | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |10 |11 |12 |13 |14 |15 |
> > + * real   | start |   OP  | PhyAddr           |  Reg Addr         |  TA   |
> > + * atheros| start |   OP  | 2'b11 |                          8'b0 |  TA   |
> > + *
> > + * Bit:   |16 |17 |18 |19 |20 |21 |22 |23 |24 |25 |26 |27 |28 |29 |30 |31 |
> > + * real   |  Data                                                         |
> > + * atheros|                       | Page [9:0]                            |
> > + */
> > +/* In case of Page Address mode, Bit[18:9] of 32 bit register address should be
> > + * written to bits[9:0] of mdio data register.
> > + */
> > +#define AR9331_SW_ADDR_PAGE			GENMASK(18, 9)
> > +
> > +/* ------------------------------------------------------------------------
> > + * Normal register access mode
> > + * ------------------------------------------------------------------------
> > + * Bit:   | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |10 |11 |12 |13 |14 |15 |
> > + * real   | start |   OP  | PhyAddr           |  Reg Addr         |  TA   |
> > + * atheros| start |   OP  | 2'b10 |  low_addr[7:0]                |  TA   |
> > + *
> > + * Bit:   |16 |17 |18 |19 |20 |21 |22 |23 |24 |25 |26 |27 |28 |29 |30 |31 |
> > + * real   |  Data                                                         |
> > + * atheros|  Data                                                         |
> > + * ------------------------------------------------------------------------
> > + */
> > +#define AR9331_SW_LOW_ADDR_PHY			GENMASK(8, 6)
> > +#define AR9331_SW_LOW_ADDR_REG			GENMASK(5, 1)
> > +
> > +#define AR9331_SW_MDIO_PHY_MODE_M		GENMASK(4, 3)
> > +#define AR9331_SW_MDIO_PHY_MODE_PAGE		3
> > +#define AR9331_SW_MDIO_PHY_MODE_REG		2
> > +#define AR9331_SW_MDIO_PHY_MODE_BYPASS		0
> > +#define AR9331_SW_MDIO_PHY_ADDR_M		GENMASK(2, 0)
> > +
> > +/* Empirical determined values */
> > +#define AR9331_SW_MDIO_POLL_SLEEP_US		1
> > +#define AR9331_SW_MDIO_POLL_TIMEOUT_US		20
> > +
> > +struct ar9331_sw_priv {
> > +	struct device *dev;
> > +	struct dsa_switch *ds;
> > +	struct dsa_switch_ops ops;
> > +	struct irq_domain *irqdomain;
> > +	struct mii_bus *mbus; /* mdio master */
> > +	struct mii_bus *sbus; /* mdio slave */
> > +	struct regmap *regmap;
> > +	struct reset_control *sw_reset;
> > +};
> > +
> > +/* Warning: switch reset will reset last AR9331_SW_MDIO_PHY_MODE_PAGE request
> > + * If some kind of optimization is used, the request should be repeated.
> > + */
> > +static int ar9331_sw_reset(struct ar9331_sw_priv *priv)
> > +{
> > +	int ret;
> > +
> > +	ret = reset_control_assert(priv->sw_reset);
> > +	if (ret)
> > +		goto error;
> > +
> > +	/* AR9331 doc do not provide any information about proper reset
> > +	 * sequence. The AR8136 (the closes switch to the AR9331) doc says:
> > +	 * reset duration should be greater than 10ms. So, let's use this value
> > +	 * for now.
> > +	 */
> > +	usleep_range(10000, 15000);
> > +	ret = reset_control_deassert(priv->sw_reset);
> > +	if (ret)
> > +		goto error;
> 
> Any comments in the documentation about needing to wait for the reset
> to complete?

No. AR8136 has EEPROM_INT and EEPROM_ERR_INT for EEPROM load status,
which can be used as wait sequence after reset. But this is not present on
AR9331. As long as we support only AR9331 in this driver, looks like there is
no need to wait after reset.

Hm.. I think I'll add this comment to the code, just in case there is a
AR8136 switch out there on a system which can run modern linux kernel.

> 
> > +
> > +	return 0;
> > +error:
> > +	dev_err_ratelimited(priv->dev, "%s: %i\n", __func__, ret);
> > +	return ret;
> > +}
> > +
> > +static int ar9331_sw_mbus_write(struct mii_bus *mbus, int port, int regnum,
> > +				u16 data)
> > +{
> > +	struct ar9331_sw_priv *priv = mbus->priv;
> > +	struct regmap *regmap = priv->regmap;
> > +	u32 val;
> > +	int ret;
> > +
> > +	ret = regmap_write(regmap, AR9331_SW_REG_MDIO_CTRL,
> > +			   AR9331_SW_MDIO_CTRL_BUSY |
> > +			   AR9331_SW_MDIO_CTRL_MASTER_EN |
> > +			   FIELD_PREP(AR9331_SW_MDIO_CTRL_PHY_ADDR_M, port) |
> > +			   FIELD_PREP(AR9331_SW_MDIO_CTRL_REG_ADDR_M, regnum) |
> > +			   FIELD_PREP(AR9331_SW_MDIO_CTRL_DATA_M, data));
> > +	if (ret)
> > +		goto error;
> > +
> > +	ret = regmap_read_poll_timeout(regmap, AR9331_SW_REG_MDIO_CTRL, val,
> > +				       !(val & AR9331_SW_MDIO_CTRL_BUSY),
> > +				       AR9331_SW_MDIO_POLL_SLEEP_US,
> > +				       AR9331_SW_MDIO_POLL_TIMEOUT_US);
> > +	if (ret)
> > +		goto error;
> > +
> > +	return 0;
> > +error:
> > +	dev_err_ratelimited(priv->dev, "PHY write error: %i\n", ret);
> > +	return ret;
> > +}
> > +
> > +static int ar9331_sw_mbus_read(struct mii_bus *mbus, int port, int regnum)
> > +{
> > +	struct ar9331_sw_priv *priv = mbus->priv;
> > +	struct regmap *regmap = priv->regmap;
> > +	u32 val;
> > +	int ret;
> > +
> > +	ret = regmap_write(regmap, AR9331_SW_REG_MDIO_CTRL,
> > +			   AR9331_SW_MDIO_CTRL_BUSY |
> > +			   AR9331_SW_MDIO_CTRL_MASTER_EN |
> > +			   AR9331_SW_MDIO_CTRL_CMD_READ |
> > +			   FIELD_PREP(AR9331_SW_MDIO_CTRL_PHY_ADDR_M, port) |
> > +			   FIELD_PREP(AR9331_SW_MDIO_CTRL_REG_ADDR_M, regnum));
> > +	if (ret)
> > +		goto error;
> > +
> > +	ret = regmap_read_poll_timeout(regmap, AR9331_SW_REG_MDIO_CTRL, val,
> > +				       !(val & AR9331_SW_MDIO_CTRL_BUSY),
> > +				       AR9331_SW_MDIO_POLL_SLEEP_US,
> > +				       AR9331_SW_MDIO_POLL_TIMEOUT_US);
> > +	if (ret)
> > +		goto error;
> > +
> > +	ret = regmap_read(regmap, AR9331_SW_REG_MDIO_CTRL, &val);
> > +	if (ret)
> > +		goto error;
> > +
> > +	return FIELD_GET(AR9331_SW_MDIO_CTRL_DATA_M, val);
> > +
> > +error:
> > +	dev_err_ratelimited(priv->dev, "PHY read error: %i\n", ret);
> > +	return ret;
> > +}
> > +
> > +static int ar9331_sw_mbus_init(struct ar9331_sw_priv *priv)
> > +{
> > +	struct device *dev = priv->dev;
> > +	static struct mii_bus *mbus;
> > +	struct device_node *np, *mnp;
> > +	int ret;
> > +
> > +	np = dev->of_node;
> > +
> > +	mbus = devm_mdiobus_alloc(dev);
> > +	if (!mbus)
> > +		return -ENOMEM;
> > +
> > +	mbus->name = np->full_name;
> > +	snprintf(mbus->id, MII_BUS_ID_SIZE, "%pOF", np);
> > +
> > +	mbus->read = ar9331_sw_mbus_read;
> > +	mbus->write = ar9331_sw_mbus_write;
> > +	mbus->priv = priv;
> > +	mbus->parent = dev;
> > +
> > +	mnp = of_get_child_by_name(np, "mdio");
> 
> You should check if mnp is NULL. You want it to mandatory. The current
> code will pass NULL to of_mdiobus_register(), which is legal, and it
> will look one level higher for the PHYs.

ok

> 
> > +	ret = of_mdiobus_register(mbus, mnp);
> > +	of_node_put(mnp);
> > +	if (ret)
> > +		return ret;
> > +
> > +	priv->mbus = mbus;
> > +
> > +	return 0;
> > +}
> > +
> > +static int ar9331_sw_setup(struct dsa_switch *ds)
> > +{
> > +	struct ar9331_sw_priv *priv = (struct ar9331_sw_priv *)ds->priv;
> > +	struct regmap *regmap = priv->regmap;
> > +	int ret;
> > +
> > +	ret = ar9331_sw_reset(priv);
> > +	if (ret)
> > +		return ret;
> > +
> > +	/* Reset will set proper defaults. CPU - Port0 will be enabled and
> > +	 * configured. All other ports (ports 1 - 5) are disabled
> > +	 */
> 
> Nice, some hardware engineer thought about that.

:) is it not common case?

> > +	ret = ar9331_sw_mbus_init(priv);
> > +	if (ret)
> > +		return ret;
> > +
> > +	/* Do not drop broadcast frames */
> > +	ret = regmap_write_bits(regmap, AR9331_SW_REG_FLOOD_MASK,
> > +				AR9331_SW_FLOOD_MASK_BROAD_TO_CPU,
> > +				AR9331_SW_FLOOD_MASK_BROAD_TO_CPU);
> > +	if (ret)
> > +		goto error;
> > +
> > +	/* Sync max frame size with value used in
> > +	 * drivers/net/ethernet/atheros/ag71xx.c for ar9330 SoC.
> > +	 * TODO: In both drivers this value seems to be not real maximal size
> > +	 * The switch is able to configure 0x3fff and ethernet controller
> > +	 * 0xffff. Are there any better way to sync this values?
> > +	 */
> > +	ret = regmap_write_bits(regmap, AR9331_SW_REG_GLOBAL_CTRL,
> > +				AR9331_SW_GLOBAL_CTRL_MFS_M,
> > +				FIELD_PREP(AR9331_SW_GLOBAL_CTRL_MFS_M, 1540));
> > +	if (ret)
> > +		goto error;
> 
> Jumbo is not so easy. I would avoid the plain number 1540. There
> should be a #define.

ok

> Also, you might want to allow space for a VLAN header?

yes.

> Does enabbling jumbo have a performance impact? If not, you
> can configure the switch to its maximum size.

ok.

> > +
> > +	return 0;
> > +error:
> > +	dev_err_ratelimited(priv->dev, "%s: %i\n", __func__, ret);
> > +	return ret;
> > +}
> > +
> > +static int ar9331_sw_port_enable(struct dsa_switch *ds, int port,
> > +				 struct phy_device *phy)
> > +{
> > +	struct ar9331_sw_priv *priv = (struct ar9331_sw_priv *)ds->priv;
> > +	struct regmap *regmap = priv->regmap;
> > +	int ret;
> > +
> > +	/* nothing to enable. Just set link to initial state */
> > +	ret = regmap_write(regmap, AR9331_SW_REG_PORT_STATUS(port), 0);
> > +	if (ret)
> > +		dev_err_ratelimited(priv->dev, "%s: %i\n", __func__, ret);
> > +
> > +	return ret;
> > +}
> > +
> > +static void ar9331_sw_port_disable(struct dsa_switch *ds, int port)
> > +{
> > +	struct ar9331_sw_priv *priv = (struct ar9331_sw_priv *)ds->priv;
> > +	struct regmap *regmap = priv->regmap;
> > +	int ret;
> > +
> > +	ret = regmap_write(regmap, AR9331_SW_REG_PORT_STATUS(port), 0);
> > +	if (ret)
> > +		dev_err_ratelimited(priv->dev, "%s: %i\n", __func__, ret);
> > +}
> 
> port_enable() and port_disable() look the same?

yes. there only two bits in port register: tx and rx mac enable. So, I
was not able to decide where to set this bits - In link_up/down or
port_up/down. I decided to make sure, port up/down should have always
predictable configuration: tx/rx mac is off and automatic
negation is disabled.

What is the better way?

> > +
> > +static enum dsa_tag_protocol ar9331_sw_get_tag_protocol(struct dsa_switch *ds,
> > +							int port)
> > +{
> > +	return DSA_TAG_PROTO_AR9331;
> > +}
> > +
> > +static void ar9331_sw_phylink_validate(struct dsa_switch *ds, int port,
> > +				       unsigned long *supported,
> > +				       struct phylink_link_state *state)
> > +{
> > +	__ETHTOOL_DECLARE_LINK_MODE_MASK(mask) = { 0, };
> > +
> > +	switch (port) {
> > +	case 0:
> > +		if (state->interface != PHY_INTERFACE_MODE_GMII)
> > +			goto unsupported;
> > +
> > +		phylink_set(mask, 1000baseT_Full);
> > +		phylink_set(mask, 1000baseT_Half);
> > +		break;
> > +	case 1:
> > +	case 2:
> > +	case 3:
> > +	case 4:
> > +	case 5:
> > +		if (state->interface != PHY_INTERFACE_MODE_INTERNAL)
> > +			goto unsupported;
> > +		break;
> > +	default:
> > +		bitmap_zero(supported, __ETHTOOL_LINK_MODE_MASK_NBITS);
> > +		dev_err(ds->dev, "Unsupported port: %i\n", port);
> > +		return;
> > +	}
> > +
> > +	phylink_set_port_modes(mask);
> > +	phylink_set(mask, Pause);
> > +	phylink_set(mask, Asym_Pause);
> > +
> > +	phylink_set(mask, 10baseT_Half);
> > +	phylink_set(mask, 10baseT_Full);
> > +	phylink_set(mask, 100baseT_Half);
> > +	phylink_set(mask, 100baseT_Full);
> 
> So the CPU port is 1G capable. All the other ports are only Fast Ethernet?

yes. 

> > +
> > +	bitmap_and(supported, supported, mask,
> > +		   __ETHTOOL_LINK_MODE_MASK_NBITS);
> > +	bitmap_and(state->advertising, state->advertising, mask,
> > +		   __ETHTOOL_LINK_MODE_MASK_NBITS);
> > +
> > +	return;
> > +
> > +unsupported:
> > +	bitmap_zero(supported, __ETHTOOL_LINK_MODE_MASK_NBITS);
> > +	dev_err(ds->dev, "Unsupported interface: %d, port: %d\n",
> > +		state->interface, port);
> > +}
> > +
> > +static void ar9331_sw_phylink_mac_config(struct dsa_switch *ds, int port,
> > +					 unsigned int mode,
> > +					 const struct phylink_link_state *state)
> > +{
> > +	struct ar9331_sw_priv *priv = (struct ar9331_sw_priv *)ds->priv;
> > +	struct regmap *regmap = priv->regmap;
> > +	int ret;
> > +	u32 val;
> > +
> > +	switch (state->speed) {
> > +	case SPEED_1000:
> > +		val = AR9331_SW_PORT_STATUS_SPEED_1000;
> > +		break;
> > +	case SPEED_100:
> > +		val = AR9331_SW_PORT_STATUS_SPEED_100;
> > +		break;
> > +	case SPEED_10:
> > +		val = AR9331_SW_PORT_STATUS_SPEED_10;
> > +		break;
> > +	default:
> > +		return;
> > +	}
> > +
> > +	if (state->duplex)
> > +		val |= AR9331_SW_PORT_STATUS_DUPLEX_MODE;
> > +
> > +	if (state->pause & MLO_PAUSE_TX)
> > +		val |= AR9331_SW_PORT_STATUS_TX_FLOW_EN;
> > +
> > +	if (state->pause & MLO_PAUSE_RX)
> > +		val |= AR9331_SW_PORT_STATUS_RX_FLOW_EN;
> > +
> > +	ret = regmap_update_bits(regmap, AR9331_SW_REG_PORT_STATUS(port),
> > +				 AR9331_SW_PORT_STATUS_LINK_MASK, val);
> > +	if (ret)
> > +		dev_err_ratelimited(priv->dev, "%s: %i\n", __func__, ret);
> > +}
> > +
> > +static void ar9331_sw_phylink_mac_link_down(struct dsa_switch *ds, int port,
> > +					    unsigned int mode,
> > +					    phy_interface_t interface)
> > +{
> > +	struct ar9331_sw_priv *priv = (struct ar9331_sw_priv *)ds->priv;
> > +	struct regmap *regmap = priv->regmap;
> > +	int ret;
> > +
> > +	ret = regmap_update_bits(regmap, AR9331_SW_REG_PORT_STATUS(port),
> > +				 AR9331_SW_PORT_STATUS_MAC_MASK, 0);
> > +	if (ret)
> > +		dev_err_ratelimited(priv->dev, "%s: %i\n", __func__, ret);
> > +}
> > +
> > +static void ar9331_sw_phylink_mac_link_up(struct dsa_switch *ds, int port,
> > +					  unsigned int mode,
> > +					  phy_interface_t interface,
> > +					  struct phy_device *phydev)
> > +{
> > +	struct ar9331_sw_priv *priv = (struct ar9331_sw_priv *)ds->priv;
> > +	struct regmap *regmap = priv->regmap;
> > +	int ret;
> > +
> > +	ret = regmap_update_bits(regmap, AR9331_SW_REG_PORT_STATUS(port),
> > +				 AR9331_SW_PORT_STATUS_MAC_MASK,
> > +				 AR9331_SW_PORT_STATUS_MAC_MASK);
> > +	if (ret)
> > +		dev_err_ratelimited(priv->dev, "%s: %i\n", __func__, ret);
> > +}
> > +
> > +static const struct dsa_switch_ops ar9331_sw_ops = {
> > +	.get_tag_protocol = ar9331_sw_get_tag_protocol,
> > +	.setup = ar9331_sw_setup,
> > +	.port_enable = ar9331_sw_port_enable,
> > +	.port_disable = ar9331_sw_port_disable,
> > +	.phylink_validate	= ar9331_sw_phylink_validate,
> > +	.phylink_mac_config	= ar9331_sw_phylink_mac_config,
> > +	.phylink_mac_link_down	= ar9331_sw_phylink_mac_link_down,
> > +	.phylink_mac_link_up	= ar9331_sw_phylink_mac_link_up,
> > +};
> > +
> > +static irqreturn_t ar9331_sw_irq(int irq, void *data)
> > +{
> > +	struct ar9331_sw_priv *priv = data;
> > +	struct regmap *regmap = priv->regmap;
> > +	u32 stat;
> > +	int ret;
> > +
> > +	ret = regmap_read(regmap, AR9331_SW_REG_GINT, &stat);
> > +	if (ret) {
> > +		dev_err(priv->dev, "can't read interrupt status\n");
> > +		return IRQ_NONE;
> > +	}
> > +
> > +	if (!stat)
> > +		return IRQ_NONE;
> > +
> > +	if (stat & AR9331_SW_GINT_PHY_INT) {
> > +		int child_irq;
> > +
> > +		child_irq = irq_find_mapping(priv->irqdomain, 0);
> > +		handle_nested_irq(child_irq);
> > +	}
> > +
> > +	ret = regmap_write(regmap, AR9331_SW_REG_GINT, stat);
> > +	if (ret) {
> > +		dev_err(priv->dev, "can't write interrupt status\n");
> > +		return IRQ_NONE;
> > +	}
> > +
> > +	return IRQ_HANDLED;
> > +}
> > +
> > +static void ar9331_sw_mask_irq(struct irq_data *d)
> > +{
> > +	struct ar9331_sw_priv *priv = irq_data_get_irq_chip_data(d);
> > +	struct regmap *regmap = priv->regmap;
> > +	int ret;
> > +
> > +	ret = regmap_update_bits(regmap, AR9331_SW_REG_GINT_MASK,
> > +				 AR9331_SW_GINT_PHY_INT, 0);
> > +	if (ret)
> > +		dev_err(priv->dev, "could not mask IRQ\n");
> > +}
> > +
> > +static void ar9331_sw_unmask_irq(struct irq_data *d)
> > +{
> > +	struct ar9331_sw_priv *priv = irq_data_get_irq_chip_data(d);
> > +	struct regmap *regmap = priv->regmap;
> > +	int ret;
> > +
> > +	ret = regmap_update_bits(regmap, AR9331_SW_REG_GINT_MASK,
> > +				 AR9331_SW_GINT_PHY_INT,
> > +				 AR9331_SW_GINT_PHY_INT);
> > +	if (ret)
> > +		dev_err(priv->dev, "could not unmask IRQ\n");
> > +}
> > +
> > +static struct irq_chip ar9331_sw_irq_chip = {
> > +	.name = AR9331_SW_NAME,
> > +	.irq_mask = ar9331_sw_mask_irq,
> > +	.irq_unmask = ar9331_sw_unmask_irq,
> > +};
> > +
> > +static int ar9331_sw_irq_map(struct irq_domain *domain, unsigned int irq,
> > +			     irq_hw_number_t hwirq)
> > +{
> > +	irq_set_chip_data(irq, domain->host_data);
> > +	irq_set_chip_and_handler(irq, &ar9331_sw_irq_chip, handle_simple_irq);
> > +	irq_set_nested_thread(irq, 1);
> > +	irq_set_noprobe(irq);
> > +
> > +	return 0;
> > +}
> > +
> > +static void ar9331_sw_irq_unmap(struct irq_domain *d, unsigned int irq)
> > +{
> > +	irq_set_nested_thread(irq, 0);
> > +	irq_set_chip_and_handler(irq, NULL, NULL);
> > +	irq_set_chip_data(irq, NULL);
> > +}
> > +
> > +static const struct irq_domain_ops ar9331_sw_irqdomain_ops = {
> > +	.map = ar9331_sw_irq_map,
> > +	.unmap = ar9331_sw_irq_unmap,
> > +	.xlate = irq_domain_xlate_onecell,
> > +};
> > +
> > +static int ar9331_sw_irq_init(struct ar9331_sw_priv *priv)
> > +{
> > +	struct device_node *np = priv->dev->of_node;
> > +	struct device *dev = priv->dev;
> > +	int ret, irq;
> > +
> > +	irq = of_irq_get(np, 0);
> > +	if (irq <= 0) {
> > +		dev_err(dev, "failed to get parent IRQ\n");
> > +		return irq ? irq : -EINVAL;
> > +	}
> > +
> > +	ret = devm_request_threaded_irq(dev, irq, NULL, ar9331_sw_irq,
> > +					IRQF_ONESHOT, AR9331_SW_NAME, priv);
> > +	if (ret) {
> > +		dev_err(dev, "unable to request irq: %d\n", ret);
> > +		return ret;
> > +	}
> > +
> > +	priv->irqdomain = irq_domain_add_linear(np, 1, &ar9331_sw_irqdomain_ops,
> > +						priv);
> > +	if (!priv->irqdomain) {
> > +		dev_err(dev, "failed to create IRQ domain\n");
> > +		return -EINVAL;
> > +	}
> > +
> > +	irq_set_parent(irq_create_mapping(priv->irqdomain, 0), irq);
> > +
> > +	return 0;
> > +}
> > +
> > +static int __ar9331_mdio_write(struct mii_bus *sbus, u8 mode, u16 reg, u16 val)
> > +{
> > +	u8 r, p;
> > +
> > +	p = FIELD_PREP(AR9331_SW_MDIO_PHY_MODE_M, mode) |
> > +		FIELD_GET(AR9331_SW_LOW_ADDR_PHY, reg);
> > +	r = FIELD_GET(AR9331_SW_LOW_ADDR_REG, reg);
> > +
> > +	return sbus->write(sbus, p, r, val);
> 
> Why not use the mdiobus_write() and mdiobus_read()?

no special reason. I'll fix.

> > +static int ar9331_sw_probe(struct mdio_device *mdiodev)
> > +{
> > +	struct ar9331_sw_priv *priv;
> > +	int ret;
> > +
> > +	/* allocate the private data struct so that we can probe the switches
> > +	 * ID register
> > +	 */
> 
> I don't see the code actually getting the ID register?

old artifact, i'll remove it. ID bits are actually not documented.


> > +	priv = devm_kzalloc(&mdiodev->dev, sizeof(*priv), GFP_KERNEL);
> > +	if (!priv)
> > +		return -ENOMEM;
> > +
> > +	priv->regmap = devm_regmap_init(&mdiodev->dev, &ar9331_sw_bus, priv,
> > +					&ar9331_mdio_regmap_config);
> > +	if (IS_ERR(priv->regmap)) {
> > +		ret = PTR_ERR(priv->regmap);
> > +		dev_err(&mdiodev->dev, "regmap init failed: %d\n", ret);
> > +		return ret;
> > +	}
> > +
> > +	priv->sw_reset = devm_reset_control_get(&mdiodev->dev, "switch");
> > +	if (IS_ERR(priv->sw_reset)) {
> > +		dev_err(&mdiodev->dev, "missing switch reset\n");
> > +		return PTR_ERR(priv->sw_reset);
> > +	}
> > +
> > +	priv->sbus = mdiodev->bus;
> > +	priv->dev = &mdiodev->dev;
> > +
> > +	ret = ar9331_sw_irq_init(priv);
> > +	if (ret)
> > +		return ret;
> > +
> > +	priv->ds = dsa_switch_alloc(&mdiodev->dev, AR9331_SW_PORTS);
> > +	if (!priv->ds)
> > +		return -ENOMEM;
> > +
> > +	priv->ds->priv = priv;
> > +	priv->ops = ar9331_sw_ops;
> > +	priv->ds->ops = &priv->ops;
> > +	dev_set_drvdata(&mdiodev->dev, priv);
> > +
> > +	return dsa_register_switch(priv->ds);
> > +}
> > +
> > +static void ar9331_sw_remove(struct mdio_device *mdiodev)
> > +{
> > +	struct ar9331_sw_priv *priv = dev_get_drvdata(&mdiodev->dev);
> > +
> > +	mdiobus_unregister(priv->mbus);
> > +	dsa_unregister_switch(priv->ds);
> > +
> > +	reset_control_assert(priv->sw_reset);
> > +}
> > +
> > +static const struct of_device_id ar9331_sw_of_match[] = {
> > +	{ .compatible = "qca,ar9331-switch" },
> > +	{ },
> > +};
> > +
> > +static struct mdio_driver ar9331_sw_mdio_driver = {
> > +	.probe = ar9331_sw_probe,
> > +	.remove = ar9331_sw_remove,
> > +	.mdiodrv.driver = {
> > +		.name = AR9331_SW_NAME,
> > +		.of_match_table = ar9331_sw_of_match,
> > +	},
> > +};
> > +
> > +mdio_module_driver(ar9331_sw_mdio_driver);
> > +
> > +MODULE_AUTHOR("Oleksij Rempel <kernel@pengutronix.de>");
> > +MODULE_DESCRIPTION("Driver for Atheros AR9331 switch");
> > +MODULE_LICENSE("GPL v2");
> > diff --git a/include/net/dsa.h b/include/net/dsa.h
> > index 541fb514e31d..89a334e68d42 100644
> > --- a/include/net/dsa.h
> > +++ b/include/net/dsa.h
> > @@ -42,6 +42,7 @@ struct phylink_link_state;
> >  #define DSA_TAG_PROTO_8021Q_VALUE		12
> >  #define DSA_TAG_PROTO_SJA1105_VALUE		13
> >  #define DSA_TAG_PROTO_KSZ8795_VALUE		14
> > +#define DSA_TAG_PROTO_AR9331_VALUE		15
> >  
> >  enum dsa_tag_protocol {
> >  	DSA_TAG_PROTO_NONE		= DSA_TAG_PROTO_NONE_VALUE,
> > @@ -59,6 +60,7 @@ enum dsa_tag_protocol {
> >  	DSA_TAG_PROTO_8021Q		= DSA_TAG_PROTO_8021Q_VALUE,
> >  	DSA_TAG_PROTO_SJA1105		= DSA_TAG_PROTO_SJA1105_VALUE,
> >  	DSA_TAG_PROTO_KSZ8795		= DSA_TAG_PROTO_KSZ8795_VALUE,
> > +	DSA_TAG_PROTO_AR9331		= DSA_TAG_PROTO_AR9331_VALUE,
> >  };
> >  
> >  struct packet_type;
> > diff --git a/net/dsa/Kconfig b/net/dsa/Kconfig
> > index 29e2bd5cc5af..6e015309a7fe 100644
> > --- a/net/dsa/Kconfig
> > +++ b/net/dsa/Kconfig
> > @@ -107,4 +107,10 @@ config NET_DSA_TAG_TRAILER
> >  	  Say Y or M if you want to enable support for tagging frames at
> >  	  with a trailed. e.g. Marvell 88E6060.
> >  
> > +config NET_DSA_TAG_AR9331
> > +	tristate "Tag driver for Atheros AR9331 SoC with build-in switch"
> > +	help
> > +	  Say Y or M if you want to enable support for tagging frames for
> > +	  the Atheros AR9331 SoC with build-in switch.
> > +
> 
> These are somewhat sorted, based on the tristate string. So this
> should go before NET_DSA_TAG_BRCM_COMMON.

ok

> 
> >  endif
> > diff --git a/net/dsa/Makefile b/net/dsa/Makefile
> > index 2c6d286f0511..67caebf602be 100644
> > --- a/net/dsa/Makefile
> > +++ b/net/dsa/Makefile
> > @@ -15,3 +15,4 @@ obj-$(CONFIG_NET_DSA_TAG_MTK) += tag_mtk.o
> >  obj-$(CONFIG_NET_DSA_TAG_QCA) += tag_qca.o
> >  obj-$(CONFIG_NET_DSA_TAG_SJA1105) += tag_sja1105.o
> >  obj-$(CONFIG_NET_DSA_TAG_TRAILER) += tag_trailer.o
> > +obj-$(CONFIG_NET_DSA_TAG_AR9331) += tag_ar9331.o
> 
> Please keep with the sorting.

ok

> > diff --git a/net/dsa/tag_ar9331.c b/net/dsa/tag_ar9331.c
> > new file mode 100644
> > index 000000000000..b32a8d3d48b9
> > --- /dev/null
> > +++ b/net/dsa/tag_ar9331.c
> > @@ -0,0 +1,97 @@
> > +// SPDX-License-Identifier: GPL-2.0
> > +/*
> > + * Copyright (c) 2019 Pengutronix, Oleksij Rempel <kernel@pengutronix.de>
> > + */
> > +
> > +
> > +#include <linux/bitfield.h>
> > +#include <linux/etherdevice.h>
> > +
> > +#include "dsa_priv.h"
> > +
> > +#define AR9331_HDR_LEN			2
> > +#define AR9331_HDR_VERSION		1
> > +
> > +#define AR9331_HDR_VERSION_MASK		GENMASK(15, 14)
> > +#define AR9331_HDR_PRIORITY_MASK	GENMASK(13, 12)
> > +#define AR9331_HDR_TYPE_MASK		GENMASK(10, 8)
> > +#define AR9331_HDR_BROADCAST		BIT(7)
> > +#define AR9331_HDR_FROM_CPU		BIT(6)
> > +/* AR9331_HDR_RESERVED - not used or may be version filed.
> 
> field

thx. ok

> > + * According to the AR8216 doc it should 0b10. On AR9331 it is 0b11 on RX path
> > + * and should be set to 0b11 to make it work.
> > + */
> > +#define AR9331_HDR_RESERVED_MASK	GENMASK(5, 4)
> > +#define AR9331_HDR_PORT_NUM_MASK	GENMASK(3, 0)
> > +
> > +static struct sk_buff *ar9331_tag_xmit(struct sk_buff *skb,
> > +				       struct net_device *dev)
> > +{
> > +	struct dsa_port *dp = dsa_slave_to_port(dev);
> > +	__le16 *phdr;
> > +	u16 hdr;
> > +
> > +	if (skb_cow_head(skb, 0) < 0)
> > +		return NULL;
> > +
> > +	phdr = skb_push(skb, AR9331_HDR_LEN);
> > +
> > +	hdr = FIELD_PREP(AR9331_HDR_VERSION_MASK, AR9331_HDR_VERSION);
> > +	hdr |= AR9331_HDR_FROM_CPU | dp->index;
> > +	/* 0b10 for AR8216 and 0b11 for AR9331 */
> > +	hdr |= AR9331_HDR_RESERVED_MASK;
> > +
> > +	phdr[0] = cpu_to_le16(hdr);
> > +
> > +	return skb;
> > +}
> > +
> > +static struct sk_buff *ar9331_tag_rcv(struct sk_buff *skb,
> > +				      struct net_device *ndev,
> > +				      struct packet_type *pt)
> > +{
> > +	u8 ver, port;
> > +	u16 hdr;
> > +
> > +	if (unlikely(!pskb_may_pull(skb, AR9331_HDR_LEN)))
> > +		return NULL;
> > +
> > +	hdr = le16_to_cpu(*(__le16 *)skb_mac_header(skb));
> > +
> > +	ver = FIELD_GET(AR9331_HDR_VERSION_MASK, hdr);
> > +	if (unlikely(ver != AR9331_HDR_VERSION)) {
> > +		netdev_warn(ndev, "%s:%i wrong header version 0x%2x\n",
> > +			    __func__, __LINE__, hdr);
> 
> This would should probably be rate limited.

good point. ok

> > +		return NULL;
> > +	}
> > +
> > +	if (unlikely(hdr & AR9331_HDR_FROM_CPU)) {
> > +		netdev_warn(ndev, "%s:%i packet should not be from cpu 0x%2x\n",
> > +			    __func__, __LINE__, hdr);
> 
> This as well.

ok.

>      Andrew
> 

Regards,
Oleksij
Fabio Estevam Oct. 16, 2019, 4:20 p.m. UTC | #6
On Wed, Oct 16, 2019 at 1:01 PM Andrew Lunn <andrew@lunn.ch> wrote:

> I think C files should use /*  */, and header files //, for SPDX.

Not really.

From Documentation/process/license-rules.rst:

"C source: // SPDX-License-Identifier: <SPDX License Expression>
 C header: /* SPDX-License-Identifier: <SPDX License Expression> */ "
David Miller Oct. 16, 2019, 5:45 p.m. UTC | #7
From: Andrew Lunn <andrew@lunn.ch>
Date: Wed, 16 Oct 2019 15:32:15 +0200

> I don't know if there are any strict rules, but i tend to use To: for
> the maintainer you expect to merge the patch, and Cc: for everybody
> else, and the lists.

This is a good way to handle things.
Florian Fainelli Oct. 17, 2019, 6:35 p.m. UTC | #8
On 10/13/2019 11:15 PM, Oleksij Rempel wrote:
> Provide basic support for Atheros AR9331 build-in switch. So far it
> works as port multiplexer without any hardware offloading support.

I glanced through the functional parts of the code, and it looks pretty
straight forward, since there is no offloading done so far, do you plan
on adding bridge offload eventually if nothing more?

When you submit v2, I would suggest splitting the tagger code from the
switch driver code, just to make them easier to review.

Thanks!
Oleksij Rempel Oct. 18, 2019, 4:52 a.m. UTC | #9
On Thu, Oct 17, 2019 at 11:35:48AM -0700, Florian Fainelli wrote:
> 
> 
> On 10/13/2019 11:15 PM, Oleksij Rempel wrote:
> > Provide basic support for Atheros AR9331 build-in switch. So far it
> > works as port multiplexer without any hardware offloading support.
> 
> I glanced through the functional parts of the code, and it looks pretty
> straight forward, since there is no offloading done so far, do you plan
> on adding bridge offload eventually if nothing more?

Currently not. There are following reasons:
- I do it for the Freifunk project. It is currently not clear, what
  functionality has higher priority.
- there are not many ar9331 based devices with enough RAM to run any
  thing modern. There is even less devices using more then one switch
  port.
- IPv6 support is important for this project, but old Atheros switches have some
  known issues with IPv6 packages in hardware bridge mode. So, this
  functionality will need more testing.

> When you submit v2, I would suggest splitting the tagger code from the
> switch driver code, just to make them easier to review.

ok.

Regards,
Oleksij