Merge branch 'qca8k-mdio'

Ansuel Smith says:

====================
Add support for qca8k mdio rw in Ethernet packet

The main reason for this is that we notice some routing problem in the
switch and it seems assisted learning is needed. Considering mdio is
quite slow due to the indirect write using this Ethernet alternative way
seems to be quicker.

The qca8k switch supports a special way to pass mdio read/write request
using specially crafted Ethernet packet.
This works by putting some defined data in the Ethernet header where the
mac source and dst should be placed. The Ethernet type header is set to qca
header and is set to a mdio read/write type.
This is used to communicate to the switch that this is a special packet
and should be parsed differently.

Currently we use Ethernet packet for
- MIB counter
- mdio read/write configuration
- phy read/write for each port

Current implementation of this use completion API to wait for the packet
to be processed by the tagger and has a timeout that fallback to the
legacy mdio way and mutex to enforce one transaction at time.

We now have connect()/disconnect() ops for the tagger. They are used to
allocate priv data in the dsa priv. The header still has to be put in
global include to make it usable by a dsa driver.
They are called when the tag is connect to the dst and the data is freed
using discconect on tagger change.

(if someone wonder why the bind function is put at in the general setup
function it's because tag is set in the cpu port where the notifier is
still not available and we require the notifier to sen the
tag_proto_connect() event.

We now have a tag_proto_connect() for the dsa driver used to put
additional data in the tagger priv (that is actually the dsa priv).
This is called using a switch event DSA_NOTIFIER_TAG_PROTO_CONNECT.
Current use for this is adding handler for the Ethernet packet to keep
the tagger code as dumb as possible.

The tagger priv implement only the handler for the special packet. All the
other stuff is placed in the qca8k_priv and the tagger has to access
it under lock.

We use the new API from Vladimir to track if the master port is
operational or not. We had to track many thing to reach a usable state.
Checking if the port is UP is not enough and tracking a NETDEV_CHANGE is
also not enough since it use also for other task. The correct way was
both track for interface UP and if a qdisc was assigned to the
interface. That tells us the port (and the tagger indirectly) is ready
to accept and process packet.

I tested this with multicpu port and with port6 set as the unique port and
it's sad.
It seems they implemented this feature in a bad way and this is only
supported with cpu port0. When cpu port6 is the unique port, the switch
doesn't send ack packet. With multicpu port, packet ack are not duplicated
and only cpu port0 sends them. This is the same for the MIB counter.
For this reason this feature is enabled only when cpu port0 is enabled and
operational.

v8:
- Reworked to rolling counter for the seq_num
- Reworked the hi/lo cache patch
- Fix multiple missing skb free and mutex lock errors
- Fix some spelling mistake
- Add macro build check for mgmt packet size
- Change some struct naming to make them more descriptive
v7:
- Rebase on net-next changes
- Add bulk patches to speedup this even more
v6:
- Fix some error in ethtool handler caused by rebase/cleanup
v5:
- Adapt to new API fixes
- Fix a wrong logic for noop
- Add additional lock for master_state change
- Limit mdio Ethernet to cpu port0 (switch limitation)
- Add priority to these special packet
- Move mdio cache to qca8k_priv
v4:
- Remove duplicate patch sent by mistake.
v3:
- Include MIB with Ethernet packet.
- Include phy read/write with Ethernet packet.
- Reorganize code with new API.
- Introuce master tracking by Vladimir
v2:
- Address all suggestion from Vladimir.
  Try to generilize this with connect/disconnect function from the
  tagger and tag_proto_connect for the driver.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/dsa/qca8k.h

@@ -11,6 +11,11 @@
 #include <linux/delay.h>
 #include <linux/regmap.h>
 #include <linux/gpio.h>
+#include <linux/dsa/tag_qca.h>
+
+#define QCA8K_ETHERNET_MDIO_PRIORITY			7
+#define QCA8K_ETHERNET_PHY_PRIORITY			6
+#define QCA8K_ETHERNET_TIMEOUT				100
 
 #define QCA8K_NUM_PORTS					7
 #define QCA8K_NUM_CPU_PORTS				2
@@ -63,7 +68,7 @@
 #define QCA8K_REG_MODULE_EN				0x030
 #define   QCA8K_MODULE_EN_MIB				BIT(0)
 #define QCA8K_REG_MIB					0x034
-#define   QCA8K_MIB_FLUSH				BIT(24)
+#define   QCA8K_MIB_FUNC				GENMASK(26, 24)
 #define   QCA8K_MIB_CPU_KEEP				BIT(20)
 #define   QCA8K_MIB_BUSY				BIT(17)
 #define QCA8K_MDIO_MASTER_CTRL				0x3c
@@ -313,6 +318,12 @@ enum qca8k_vlan_cmd {
 	QCA8K_VLAN_READ = 6,
 };
 
+enum qca8k_mid_cmd {
+	QCA8K_MIB_FLUSH = 1,
+	QCA8K_MIB_FLUSH_PORT = 2,
+	QCA8K_MIB_CAST = 3,
+};
+
 struct ar8xxx_port_status {
 	int enabled;
 };
@@ -328,6 +339,22 @@ enum {
 	QCA8K_CPU_PORT6,
 };
 
+struct qca8k_mgmt_eth_data {
+	struct completion rw_done;
+	struct mutex mutex; /* Enforce one mdio read/write at time */
+	bool ack;
+	u32 seq;
+	u32 data[4];
+};
+
+struct qca8k_mib_eth_data {
+	struct completion rw_done;
+	struct mutex mutex; /* Process one command at time */
+	refcount_t port_parsed; /* Counter to track parsed port */
+	u8 req_port;
+	u64 *data; /* pointer to ethtool data */
+};
+
 struct qca8k_ports_config {
 	bool sgmii_rx_clk_falling_edge;
 	bool sgmii_tx_clk_falling_edge;
@@ -336,6 +363,19 @@ struct qca8k_ports_config {
 	u8 rgmii_tx_delay[QCA8K_NUM_CPU_PORTS]; /* 0: CPU port0, 1: CPU port6 */
 };
 
+struct qca8k_mdio_cache {
+/* The 32bit switch registers are accessed indirectly. To achieve this we need
+ * to set the page of the register. Track the last page that was set to reduce
+ * mdio writes
+ */
+	u16 page;
+/* lo and hi can also be cached and from Documentation we can skip one
+ * extra mdio write if lo or hi is didn't change.
+ */
+	u16 lo;
+	u16 hi;
+};
+
 struct qca8k_priv {
 	u8 switch_id;
 	u8 switch_revision;
@@ -353,6 +393,10 @@ struct qca8k_priv {
 	struct dsa_switch_ops ops;
 	struct gpio_desc *reset_gpio;
 	unsigned int port_mtu[QCA8K_NUM_PORTS];
+	struct net_device *mgmt_master; /* Track if mdio/mib Ethernet is available */
+	struct qca8k_mgmt_eth_data mgmt_eth_data;
+	struct qca8k_mib_eth_data mib_eth_data;
+	struct qca8k_mdio_cache mdio_cache;
 };
 
 struct qca8k_mib_desc {

include/linux/dsa/tag_qca.h

@@ -0,0 +1,82 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef __TAG_QCA_H
+#define __TAG_QCA_H
+
+#define QCA_HDR_LEN	2
+#define QCA_HDR_VERSION	0x2
+
+#define QCA_HDR_RECV_VERSION		GENMASK(15, 14)
+#define QCA_HDR_RECV_PRIORITY		GENMASK(13, 11)
+#define QCA_HDR_RECV_TYPE		GENMASK(10, 6)
+#define QCA_HDR_RECV_FRAME_IS_TAGGED	BIT(3)
+#define QCA_HDR_RECV_SOURCE_PORT	GENMASK(2, 0)
+
+/* Packet type for recv */
+#define QCA_HDR_RECV_TYPE_NORMAL	0x0
+#define QCA_HDR_RECV_TYPE_MIB		0x1
+#define QCA_HDR_RECV_TYPE_RW_REG_ACK	0x2
+
+#define QCA_HDR_XMIT_VERSION		GENMASK(15, 14)
+#define QCA_HDR_XMIT_PRIORITY		GENMASK(13, 11)
+#define QCA_HDR_XMIT_CONTROL		GENMASK(10, 8)
+#define QCA_HDR_XMIT_FROM_CPU		BIT(7)
+#define QCA_HDR_XMIT_DP_BIT		GENMASK(6, 0)
+
+/* Packet type for xmit */
+#define QCA_HDR_XMIT_TYPE_NORMAL	0x0
+#define QCA_HDR_XMIT_TYPE_RW_REG	0x1
+
+/* Check code for a valid mgmt packet. Switch will ignore the packet
+ * with this wrong.
+ */
+#define QCA_HDR_MGMT_CHECK_CODE_VAL	0x5
+
+/* Specific define for in-band MDIO read/write with Ethernet packet */
+#define QCA_HDR_MGMT_SEQ_LEN		4 /* 4 byte for the seq */
+#define QCA_HDR_MGMT_COMMAND_LEN	4 /* 4 byte for the command */
+#define QCA_HDR_MGMT_DATA1_LEN		4 /* First 4 byte for the mdio data */
+#define QCA_HDR_MGMT_HEADER_LEN		(QCA_HDR_MGMT_SEQ_LEN + \
+					QCA_HDR_MGMT_COMMAND_LEN + \
+					QCA_HDR_MGMT_DATA1_LEN)
+
+#define QCA_HDR_MGMT_DATA2_LEN		12 /* Other 12 byte for the mdio data */
+#define QCA_HDR_MGMT_PADDING_LEN	34 /* Padding to reach the min Ethernet packet */
+
+#define QCA_HDR_MGMT_PKT_LEN		(QCA_HDR_MGMT_HEADER_LEN + \
+					QCA_HDR_LEN + \
+					QCA_HDR_MGMT_DATA2_LEN + \
+					QCA_HDR_MGMT_PADDING_LEN)
+
+#define QCA_HDR_MGMT_SEQ_NUM		GENMASK(31, 0)  /* 63, 32 */
+#define QCA_HDR_MGMT_CHECK_CODE		GENMASK(31, 29) /* 31, 29 */
+#define QCA_HDR_MGMT_CMD		BIT(28)		/* 28 */
+#define QCA_HDR_MGMT_LENGTH		GENMASK(23, 20) /* 23, 20 */
+#define QCA_HDR_MGMT_ADDR		GENMASK(18, 0)  /* 18, 0 */
+
+/* Special struct emulating a Ethernet header */
+struct qca_mgmt_ethhdr {
+	u32 command;		/* command bit 31:0 */
+	u32 seq;		/* seq 63:32 */
+	u32 mdio_data;		/* first 4byte mdio */
+	__be16 hdr;		/* qca hdr */
+} __packed;
+
+enum mdio_cmd {
+	MDIO_WRITE = 0x0,
+	MDIO_READ
+};
+
+struct mib_ethhdr {
+	u32 data[3];		/* first 3 mib counter */
+	__be16 hdr;		/* qca hdr */
+} __packed;
+
+struct qca_tagger_data {
+	void (*rw_reg_ack_handler)(struct dsa_switch *ds,
+				   struct sk_buff *skb);
+	void (*mib_autocast_handler)(struct dsa_switch *ds,
+				     struct sk_buff *skb);
+};
+
+#endif /* __TAG_QCA_H */

include/net/dsa.h

@@ -278,6 +278,10 @@ struct dsa_port {
 
 	u8			devlink_port_setup:1;
 
+	/* Master state bits, valid only on CPU ports */
+	u8			master_admin_up:1;
+	u8			master_oper_up:1;
+
 	u8			setup:1;
 
 	struct device_node	*dn;
@@ -478,6 +482,12 @@ static inline bool dsa_port_is_unused(struct dsa_port *dp)
 	return dp->type == DSA_PORT_TYPE_UNUSED;
 }
 
+static inline bool dsa_port_master_is_operational(struct dsa_port *dp)
+{
+	return dsa_port_is_cpu(dp) && dp->master_admin_up &&
+	       dp->master_oper_up;
+}
+
 static inline bool dsa_is_unused_port(struct dsa_switch *ds, int p)
 {
 	return dsa_to_port(ds, p)->type == DSA_PORT_TYPE_UNUSED;
@@ -1036,6 +1046,13 @@ struct dsa_switch_ops {
 	int	(*tag_8021q_vlan_add)(struct dsa_switch *ds, int port, u16 vid,
 				      u16 flags);
 	int	(*tag_8021q_vlan_del)(struct dsa_switch *ds, int port, u16 vid);
+
+	/*
+	 * DSA master tracking operations
+	 */
+	void	(*master_state_change)(struct dsa_switch *ds,
+				       const struct net_device *master,
+				       bool operational);
 };
 
 #define DSA_DEVLINK_PARAM_DRIVER(_id, _name, _type, _cmodes)		\

net/dsa/dsa2.c

@@ -15,6 +15,7 @@
 #include <linux/of.h>
 #include <linux/of_net.h>
 #include <net/devlink.h>
+#include <net/sch_generic.h>
 
 #include "dsa_priv.h"
 
@@ -1064,9 +1065,18 @@ static int dsa_tree_setup_master(struct dsa_switch_tree *dst)
 
 	list_for_each_entry(dp, &dst->ports, list) {
 		if (dsa_port_is_cpu(dp)) {
-			err = dsa_master_setup(dp->master, dp);
+			struct net_device *master = dp->master;
+			bool admin_up = (master->flags & IFF_UP) &&
+					!qdisc_tx_is_noop(master);
+
+			err = dsa_master_setup(master, dp);
 			if (err)
 				return err;
+
+			/* Replay master state event */
+			dsa_tree_master_admin_state_change(dst, master, admin_up);
+			dsa_tree_master_oper_state_change(dst, master,
+							  netif_oper_up(master));
 		}
 	}
 
@@ -1081,9 +1091,19 @@ static void dsa_tree_teardown_master(struct dsa_switch_tree *dst)
 
 	rtnl_lock();
 
-	list_for_each_entry(dp, &dst->ports, list)
-		if (dsa_port_is_cpu(dp))
-			dsa_master_teardown(dp->master);
+	list_for_each_entry(dp, &dst->ports, list) {
+		if (dsa_port_is_cpu(dp)) {
+			struct net_device *master = dp->master;
+
+			/* Synthesizing an "admin down" state is sufficient for
+			 * the switches to get a notification if the master is
+			 * currently up and running.
+			 */
+			dsa_tree_master_admin_state_change(dst, master, false);
+
+			dsa_master_teardown(master);
+		}
+	}
 
 	rtnl_unlock();
 }
@@ -1279,6 +1299,52 @@ int dsa_tree_change_tag_proto(struct dsa_switch_tree *dst,
 	return err;
 }
 
+static void dsa_tree_master_state_change(struct dsa_switch_tree *dst,
+					 struct net_device *master)
+{
+	struct dsa_notifier_master_state_info info;
+	struct dsa_port *cpu_dp = master->dsa_ptr;
+
+	info.master = master;
+	info.operational = dsa_port_master_is_operational(cpu_dp);
+
+	dsa_tree_notify(dst, DSA_NOTIFIER_MASTER_STATE_CHANGE, &info);
+}
+
+void dsa_tree_master_admin_state_change(struct dsa_switch_tree *dst,
+					struct net_device *master,
+					bool up)
+{
+	struct dsa_port *cpu_dp = master->dsa_ptr;
+	bool notify = false;
+
+	if ((dsa_port_master_is_operational(cpu_dp)) !=
+	    (up && cpu_dp->master_oper_up))
+		notify = true;
+
+	cpu_dp->master_admin_up = up;
+
+	if (notify)
+		dsa_tree_master_state_change(dst, master);
+}
+
+void dsa_tree_master_oper_state_change(struct dsa_switch_tree *dst,
+				       struct net_device *master,
+				       bool up)
+{
+	struct dsa_port *cpu_dp = master->dsa_ptr;
+	bool notify = false;
+
+	if ((dsa_port_master_is_operational(cpu_dp)) !=
+	    (cpu_dp->master_admin_up && up))
+		notify = true;
+
+	cpu_dp->master_oper_up = up;
+
+	if (notify)
+		dsa_tree_master_state_change(dst, master);
+}
+
 static struct dsa_port *dsa_port_touch(struct dsa_switch *ds, int index)
 {
 	struct dsa_switch_tree *dst = ds->dst;

net/dsa/dsa_priv.h

@@ -40,6 +40,7 @@ enum {
 	DSA_NOTIFIER_TAG_PROTO_DISCONNECT,
 	DSA_NOTIFIER_TAG_8021Q_VLAN_ADD,
 	DSA_NOTIFIER_TAG_8021Q_VLAN_DEL,
+	DSA_NOTIFIER_MASTER_STATE_CHANGE,
 };
 
 /* DSA_NOTIFIER_AGEING_TIME */
@@ -109,6 +110,12 @@ struct dsa_notifier_tag_8021q_vlan_info {
 	u16 vid;
 };
 
+/* DSA_NOTIFIER_MASTER_STATE_CHANGE */
+struct dsa_notifier_master_state_info {
+	const struct net_device *master;
+	bool operational;
+};
+
 struct dsa_switchdev_event_work {
 	struct dsa_switch *ds;
 	int port;
@@ -482,6 +489,12 @@ int dsa_tree_change_tag_proto(struct dsa_switch_tree *dst,
 			      struct net_device *master,
 			      const struct dsa_device_ops *tag_ops,
 			      const struct dsa_device_ops *old_tag_ops);
+void dsa_tree_master_admin_state_change(struct dsa_switch_tree *dst,
+					struct net_device *master,
+					bool up);
+void dsa_tree_master_oper_state_change(struct dsa_switch_tree *dst,
+				       struct net_device *master,
+				       bool up);
 unsigned int dsa_bridge_num_get(const struct net_device *bridge_dev, int max);
 void dsa_bridge_num_put(const struct net_device *bridge_dev,
 			unsigned int bridge_num);

net/dsa/slave.c

@@ -2346,6 +2346,36 @@ static int dsa_slave_netdevice_event(struct notifier_block *nb,
 		err = dsa_port_lag_change(dp, info->lower_state_info);
 		return notifier_from_errno(err);
 	}
+	case NETDEV_CHANGE:
+	case NETDEV_UP: {
+		/* Track state of master port.
+		 * DSA driver may require the master port (and indirectly
+		 * the tagger) to be available for some special operation.
+		 */
+		if (netdev_uses_dsa(dev)) {
+			struct dsa_port *cpu_dp = dev->dsa_ptr;
+			struct dsa_switch_tree *dst = cpu_dp->ds->dst;
+
+			/* Track when the master port is UP */
+			dsa_tree_master_oper_state_change(dst, dev,
+							  netif_oper_up(dev));
+
+			/* Track when the master port is ready and can accept
+			 * packet.
+			 * NETDEV_UP event is not enough to flag a port as ready.
+			 * We also have to wait for linkwatch_do_dev to dev_activate
+			 * and emit a NETDEV_CHANGE event.
+			 * We check if a master port is ready by checking if the dev
+			 * have a qdisc assigned and is not noop.
+			 */
+			dsa_tree_master_admin_state_change(dst, dev,
+							   !qdisc_tx_is_noop(dev));
+
+			return NOTIFY_OK;
+		}
+
+		return NOTIFY_DONE;
+	}
 	case NETDEV_GOING_DOWN: {
 		struct dsa_port *dp, *cpu_dp;
 		struct dsa_switch_tree *dst;
@@ -2357,6 +2387,8 @@ static int dsa_slave_netdevice_event(struct notifier_block *nb,
 		cpu_dp = dev->dsa_ptr;
 		dst = cpu_dp->ds->dst;
 
+		dsa_tree_master_admin_state_change(dst, dev, false);
+
 		list_for_each_entry(dp, &dst->ports, list) {
 			if (!dsa_port_is_user(dp))
 				continue;