Merge branch 'dsa-add-fabric-notifier'

Vivien Didelot says:

====================
net: dsa: add fabric notifier

When a switch fabric is composed of multiple switch chips, these chips
must be programmed accordingly when an event occurred on one of them.

Examples of such event include hardware bridging: when a Linux bridge
spans interconnected chips, they must be programmed to allow external
ports to ingress frames on their internal ports.

Another example is cross-chip hardware VLANs. Switch chips in-between
interconnected bridge ports must also configure a given VLAN to allow
packets to pass through them.

In order to support that, this patchset introduces a non-intrusive
notifier mechanism. It adds a notifier head in every DSA switch tree
(the said fabric), and a notifier block in every DSA switch chip.

When an even occurs, it is chained to all notifiers of the fabric.
Switch chips can react accordingly if they are cross-chip capable.

On a dynamic debug enabled system, bridging a port in a multi-chip
fabric will print something like this (ZII Rev B board):

    # brctl addif br0 lan3
    mv88e6085 0.1:00: crosschip DSA port 1.0 bridged to br0
    mv88e6085 0.4:00: crosschip DSA port 1.0 bridged to br0
    # brctl delif br0 lan3
    mv88e6085 0.1:00: crosschip DSA port 1.0 unbridged from br0
    mv88e6085 0.4:00: crosschip DSA port 1.0 unbridged from br0

Currently only bridging events are added. A patchset introducing support
for cross-chip hardware bridging configuration in mv88e6xxx will follow
right after. Then events for switchdev operations are next on the line.

We should note that non-switchdev events do not support rolling-back
switch-wide operations. We'll have to work on closer integration with
switchdev for that, like introducing new attributes or objects, to
benefit from the prepare and commit phases.
====================

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

include/net/dsa.h

@@ -13,6 +13,7 @@
 
 #include <linux/if_ether.h>
 #include <linux/list.h>
+#include <linux/notifier.h>
 #include <linux/timer.h>
 #include <linux/workqueue.h>
 #include <linux/of.h>
@@ -92,6 +93,9 @@ struct packet_type;
 struct dsa_switch_tree {
 	struct list_head	list;
 
+	/* Notifier chain for switch-wide events */
+	struct raw_notifier_head	nh;
+
 	/* Tree identifier */
 	u32 tree;
 
@@ -182,6 +186,9 @@ struct dsa_switch {
 	struct dsa_switch_tree	*dst;
 	int			index;
 
+	/* Listener for switch fabric events */
+	struct notifier_block	nb;
+
 	/*
 	 * Give the switch driver somewhere to hang its private data
 	 * structure.
@@ -261,6 +268,16 @@ struct switchdev_obj_port_fdb;
 struct switchdev_obj_port_mdb;
 struct switchdev_obj_port_vlan;
 
+#define DSA_NOTIFIER_BRIDGE_JOIN		1
+#define DSA_NOTIFIER_BRIDGE_LEAVE		2
+
+/* DSA_NOTIFIER_BRIDGE_* */
+struct dsa_notifier_bridge_info {
+	struct net_device *br;
+	int sw_index;
+	int port;
+};
+
 struct dsa_switch_ops {
 	/*
 	 * Probing and setup.

net/dsa/Makefile

@@ -1,6 +1,7 @@
 # the core
 obj-$(CONFIG_NET_DSA) += dsa_core.o
 dsa_core-y += dsa.o slave.o dsa2.o
+dsa_core-y += dsa.o slave.o dsa2.o switch.o
 
 # tagging formats
 dsa_core-$(CONFIG_NET_DSA_TAG_BRCM) += tag_brcm.o

net/dsa/dsa.c

@@ -275,6 +275,10 @@ static int dsa_switch_setup_one(struct dsa_switch *ds, struct device *parent)
 	if (ret < 0)
 		return ret;
 
+	ret = dsa_switch_register_notifier(ds);
+	if (ret)
+		return ret;
+
 	if (ops->set_addr) {
 		ret = ops->set_addr(ds, dst->master_netdev->dev_addr);
 		if (ret < 0)
@@ -400,6 +404,8 @@ static void dsa_switch_destroy(struct dsa_switch *ds)
 
 	if (ds->slave_mii_bus && ds->ops->phy_read)
 		mdiobus_unregister(ds->slave_mii_bus);
+
+	dsa_switch_unregister_notifier(ds);
 }
 
 #ifdef CONFIG_PM_SLEEP
@@ -903,10 +909,6 @@ static struct packet_type dsa_pack_type __read_mostly = {
 	.func	= dsa_switch_rcv,
 };
 
-static struct notifier_block dsa_netdevice_nb __read_mostly = {
-	.notifier_call	= dsa_slave_netdevice_event,
-};
-
 #ifdef CONFIG_PM_SLEEP
 static int dsa_suspend(struct device *d)
 {
@@ -964,7 +966,9 @@ static int __init dsa_init_module(void)
 {
 	int rc;
 
-	register_netdevice_notifier(&dsa_netdevice_nb);
+	rc = dsa_slave_register_notifier();
+	if (rc)
+		return rc;
 
 	rc = platform_driver_register(&dsa_driver);
 	if (rc)
@@ -978,7 +982,7 @@ module_init(dsa_init_module);
 
 static void __exit dsa_cleanup_module(void)
 {
-	unregister_netdevice_notifier(&dsa_netdevice_nb);
+	dsa_slave_unregister_notifier();
 	dev_remove_pack(&dsa_pack_type);
 	platform_driver_unregister(&dsa_driver);
 }

net/dsa/dsa2.c

@@ -294,6 +294,10 @@ static int dsa_ds_apply(struct dsa_switch_tree *dst, struct dsa_switch *ds)
 	if (err < 0)
 		return err;
 
+	err = dsa_switch_register_notifier(ds);
+	if (err)
+		return err;
+
 	if (ds->ops->set_addr) {
 		err = ds->ops->set_addr(ds, dst->master_netdev->dev_addr);
 		if (err < 0)
@@ -364,6 +368,8 @@ static void dsa_ds_unapply(struct dsa_switch_tree *dst, struct dsa_switch *ds)
 
 	if (ds->slave_mii_bus && ds->ops->phy_read)
 		mdiobus_unregister(ds->slave_mii_bus);
+
+	dsa_switch_unregister_notifier(ds);
 }
 
 static int dsa_dst_apply(struct dsa_switch_tree *dst)

net/dsa/dsa_priv.h

@@ -63,8 +63,12 @@ int dsa_slave_create(struct dsa_switch *ds, struct device *parent,
 void dsa_slave_destroy(struct net_device *slave_dev);
 int dsa_slave_suspend(struct net_device *slave_dev);
 int dsa_slave_resume(struct net_device *slave_dev);
-int dsa_slave_netdevice_event(struct notifier_block *unused,
-			      unsigned long event, void *ptr);
+int dsa_slave_register_notifier(void);
+void dsa_slave_unregister_notifier(void);
+
+/* switch.c */
+int dsa_switch_register_notifier(struct dsa_switch *ds);
+void dsa_switch_unregister_notifier(struct dsa_switch *ds);
 
 /* tag_dsa.c */
 extern const struct dsa_device_ops dsa_netdev_ops;

net/dsa/slave.c

@@ -27,6 +27,17 @@
 
 static bool dsa_slave_dev_check(struct net_device *dev);
 
+static int dsa_slave_notify(struct net_device *dev, unsigned long e, void *v)
+{
+	struct dsa_slave_priv *p = netdev_priv(dev);
+	struct raw_notifier_head *nh = &p->dp->ds->dst->nh;
+	int err;
+
+	err = raw_notifier_call_chain(nh, e, v);
+
+	return notifier_to_errno(err);
+}
+
 /* slave mii_bus handling ***************************************************/
 static int dsa_slave_phy_read(struct mii_bus *bus, int addr, int reg)
 {
@@ -74,9 +85,12 @@ static inline bool dsa_port_is_bridged(struct dsa_port *dp)
 	return !!dp->bridge_dev;
 }
 
-static void dsa_port_set_stp_state(struct dsa_switch *ds, int port, u8 state)
+static void dsa_slave_set_state(struct net_device *dev, u8 state)
 {
-	struct dsa_port *dp = &ds->ports[port];
+	struct dsa_slave_priv *p = netdev_priv(dev);
+	struct dsa_port *dp = p->dp;
+	struct dsa_switch *ds = dp->ds;
+	int port = dp->index;
 
 	if (ds->ops->port_stp_state_set)
 		ds->ops->port_stp_state_set(ds, port, state);
@@ -133,7 +147,7 @@ static int dsa_slave_open(struct net_device *dev)
 			goto clear_promisc;
 	}
 
-	dsa_port_set_stp_state(ds, p->dp->index, stp_state);
+	dsa_slave_set_state(dev, stp_state);
 
 	if (p->phy)
 		phy_start(p->phy);
@@ -175,7 +189,7 @@ static int dsa_slave_close(struct net_device *dev)
 	if (ds->ops->port_disable)
 		ds->ops->port_disable(ds, p->dp->index, p->phy);
 
-	dsa_port_set_stp_state(ds, p->dp->index, BR_STATE_DISABLED);
+	dsa_slave_set_state(dev, BR_STATE_DISABLED);
 
 	return 0;
 }
@@ -382,7 +396,7 @@ static int dsa_slave_stp_state_set(struct net_device *dev,
 	if (switchdev_trans_ph_prepare(trans))
 		return ds->ops->port_stp_state_set ? 0 : -EOPNOTSUPP;
 
-	dsa_port_set_stp_state(ds, p->dp->index, attr->u.stp_state);
+	dsa_slave_set_state(dev, attr->u.stp_state);
 
 	return 0;
 }
@@ -559,32 +573,51 @@ static int dsa_slave_bridge_port_join(struct net_device *dev,
 				      struct net_device *br)
 {
 	struct dsa_slave_priv *p = netdev_priv(dev);
-	struct dsa_switch *ds = p->dp->ds;
-	int ret = -EOPNOTSUPP;
+	struct dsa_notifier_bridge_info info = {
+		.sw_index = p->dp->ds->index,
+		.port = p->dp->index,
+		.br = br,
+	};
+	int err;
 
+	/* Here the port is already bridged. Reflect the current configuration
+	 * so that drivers can program their chips accordingly.
+	 */
 	p->dp->bridge_dev = br;
 
-	if (ds->ops->port_bridge_join)
-		ret = ds->ops->port_bridge_join(ds, p->dp->index, br);
+	err = dsa_slave_notify(dev, DSA_NOTIFIER_BRIDGE_JOIN, &info);
 
-	return ret == -EOPNOTSUPP ? 0 : ret;
+	/* The bridging is rolled back on error */
+	if (err)
+		p->dp->bridge_dev = NULL;
+
+	return err;
 }
 
 static void dsa_slave_bridge_port_leave(struct net_device *dev,
 					struct net_device *br)
 {
 	struct dsa_slave_priv *p = netdev_priv(dev);
-	struct dsa_switch *ds = p->dp->ds;
+	struct dsa_notifier_bridge_info info = {
+		.sw_index = p->dp->ds->index,
+		.port = p->dp->index,
+		.br = br,
+	};
+	int err;
 
+	/* Here the port is already unbridged. Reflect the current configuration
+	 * so that drivers can program their chips accordingly.
+	 */
 	p->dp->bridge_dev = NULL;
 
-	if (ds->ops->port_bridge_leave)
-		ds->ops->port_bridge_leave(ds, p->dp->index, br);
+	err = dsa_slave_notify(dev, DSA_NOTIFIER_BRIDGE_LEAVE, &info);
+	if (err)
+		netdev_err(dev, "failed to notify DSA_NOTIFIER_BRIDGE_LEAVE\n");
 
 	/* Port left the bridge, put in BR_STATE_DISABLED by the bridge layer,
 	 * so allow it to be in BR_STATE_FORWARDING to be kept functional
 	 */
-	dsa_port_set_stp_state(ds, p->dp->index, BR_STATE_FORWARDING);
+	dsa_slave_set_state(dev, BR_STATE_FORWARDING);
 }
 
 static int dsa_slave_port_attr_get(struct net_device *dev,
@@ -1491,46 +1524,52 @@ static bool dsa_slave_dev_check(struct net_device *dev)
 	return dev->netdev_ops == &dsa_slave_netdev_ops;
 }
 
-static int dsa_slave_port_upper_event(struct net_device *dev,
-				      unsigned long event, void *ptr)
+static int dsa_slave_changeupper(struct net_device *dev,
+				 struct netdev_notifier_changeupper_info *info)
 {
-	struct netdev_notifier_changeupper_info *info = ptr;
-	struct net_device *upper = info->upper_dev;
-	int err = 0;
+	int err = NOTIFY_DONE;
 
-	switch (event) {
-	case NETDEV_CHANGEUPPER:
-		if (netif_is_bridge_master(upper)) {
-			if (info->linking)
-				err = dsa_slave_bridge_port_join(dev, upper);
-			else
-				dsa_slave_bridge_port_leave(dev, upper);
+	if (netif_is_bridge_master(info->upper_dev)) {
+		if (info->linking) {
+			err = dsa_slave_bridge_port_join(dev, info->upper_dev);
+			err = notifier_from_errno(err);
+		} else {
+			dsa_slave_bridge_port_leave(dev, info->upper_dev);
+			err = NOTIFY_OK;
 		}
-
-		break;
 	}
 
-	return notifier_from_errno(err);
+	return err;
 }
 
-static int dsa_slave_port_event(struct net_device *dev, unsigned long event,
-				void *ptr)
+static int dsa_slave_netdevice_event(struct notifier_block *nb,
+				     unsigned long event, void *ptr)
 {
-	switch (event) {
-	case NETDEV_CHANGEUPPER:
-		return dsa_slave_port_upper_event(dev, event, ptr);
-	}
+	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
+
+	if (dev->netdev_ops != &dsa_slave_netdev_ops)
+		return NOTIFY_DONE;
+
+	if (event == NETDEV_CHANGEUPPER)
+		return dsa_slave_changeupper(dev, ptr);
 
 	return NOTIFY_DONE;
 }
 
-int dsa_slave_netdevice_event(struct notifier_block *unused,
-			      unsigned long event, void *ptr)
+static struct notifier_block dsa_slave_nb __read_mostly = {
+	.notifier_call	= dsa_slave_netdevice_event,
+};
+
+int dsa_slave_register_notifier(void)
 {
-	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
+	return register_netdevice_notifier(&dsa_slave_nb);
+}
 
-	if (dsa_slave_dev_check(dev))
-		return dsa_slave_port_event(dev, event, ptr);
+void dsa_slave_unregister_notifier(void)
+{
+	int err;
 
-	return NOTIFY_DONE;
+	err = unregister_netdevice_notifier(&dsa_slave_nb);
+	if (err)
+		pr_err("DSA: failed to unregister slave notifier (%d)\n", err);
 }

net/dsa/switch.c

@@ -0,0 +1,85 @@
+/*
+ * Handling of a single switch chip, part of a switch fabric
+ *
+ * Copyright (c) 2017 Vivien Didelot <vivien.didelot@savoirfairelinux.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <linux/netdevice.h>
+#include <linux/notifier.h>
+#include <net/dsa.h>
+
+static int dsa_switch_bridge_join(struct dsa_switch *ds,
+				  struct dsa_notifier_bridge_info *info)
+{
+	if (ds->index == info->sw_index && ds->ops->port_bridge_join)
+		return ds->ops->port_bridge_join(ds, info->port, info->br);
+
+	if (ds->index != info->sw_index)
+		dev_dbg(ds->dev, "crosschip DSA port %d.%d bridged to %s\n",
+			info->sw_index, info->port, netdev_name(info->br));
+
+	return 0;
+}
+
+static int dsa_switch_bridge_leave(struct dsa_switch *ds,
+				   struct dsa_notifier_bridge_info *info)
+{
+	if (ds->index == info->sw_index && ds->ops->port_bridge_leave)
+		ds->ops->port_bridge_leave(ds, info->port, info->br);
+
+	if (ds->index != info->sw_index)
+		dev_dbg(ds->dev, "crosschip DSA port %d.%d unbridged from %s\n",
+			info->sw_index, info->port, netdev_name(info->br));
+
+	return 0;
+}
+
+static int dsa_switch_event(struct notifier_block *nb,
+			    unsigned long event, void *info)
+{
+	struct dsa_switch *ds = container_of(nb, struct dsa_switch, nb);
+	int err;
+
+	switch (event) {
+	case DSA_NOTIFIER_BRIDGE_JOIN:
+		err = dsa_switch_bridge_join(ds, info);
+		break;
+	case DSA_NOTIFIER_BRIDGE_LEAVE:
+		err = dsa_switch_bridge_leave(ds, info);
+		break;
+	default:
+		err = -EOPNOTSUPP;
+		break;
+	}
+
+	/* Non-switchdev operations cannot be rolled back. If a DSA driver
+	 * returns an error during the chained call, switch chips may be in an
+	 * inconsistent state.
+	 */
+	if (err)
+		dev_dbg(ds->dev, "breaking chain for DSA event %lu (%d)\n",
+			event, err);
+
+	return notifier_from_errno(err);
+}
+
+int dsa_switch_register_notifier(struct dsa_switch *ds)
+{
+	ds->nb.notifier_call = dsa_switch_event;
+
+	return raw_notifier_chain_register(&ds->dst->nh, &ds->nb);
+}
+
+void dsa_switch_unregister_notifier(struct dsa_switch *ds)
+{
+	int err;
+
+	err = raw_notifier_chain_unregister(&ds->dst->nh, &ds->nb);
+	if (err)
+		dev_err(ds->dev, "failed to unregister notifier (%d)\n", err);
+}