net: dsa: allow the DSA master to be seen and changed through rtnetlink

Some DSA switches have multiple CPU ports, which can be used to improve
CPU termination throughput, but DSA, through dsa_tree_setup_cpu_ports(),
sets up only the first one, leading to suboptimal use of hardware.

The desire is to not change the default configuration but to permit the
user to create a dynamic mapping between individual user ports and the
CPU port that they are served by, configurable through rtnetlink. It is
also intended to permit load balancing between CPU ports, and in that
case, the foreseen model is for the DSA master to be a bonding interface
whose lowers are the physical DSA masters.

To that end, we create a struct rtnl_link_ops for DSA user ports with
the "dsa" kind. We expose the IFLA_DSA_MASTER link attribute that
contains the ifindex of the newly desired DSA master.

Signed-off-by: Vladimir Oltean <[email protected]>
Signed-off-by: Paolo Abeni <[email protected]>

Author: vladimiroltean

include/net/dsa.h

@@ -559,6 +559,10 @@ static inline bool dsa_is_user_port(struct dsa_switch *ds, int p)
 	list_for_each_entry((_dp), &(_dst)->ports, list) \
 		if (dsa_port_is_user((_dp)))
 
+#define dsa_tree_for_each_user_port_continue_reverse(_dp, _dst) \
+	list_for_each_entry_continue_reverse((_dp), &(_dst)->ports, list) \
+		if (dsa_port_is_user((_dp)))
+
 #define dsa_tree_for_each_cpu_port(_dp, _dst) \
 	list_for_each_entry((_dp), &(_dst)->ports, list) \
 		if (dsa_port_is_cpu((_dp)))
@@ -830,6 +834,10 @@ struct dsa_switch_ops {
 	int	(*connect_tag_protocol)(struct dsa_switch *ds,
 					enum dsa_tag_protocol proto);
 
+	int	(*port_change_master)(struct dsa_switch *ds, int port,
+				      struct net_device *master,
+				      struct netlink_ext_ack *extack);
+
 	/* Optional switch-wide initialization and destruction methods */
 	int	(*setup)(struct dsa_switch *ds);
 	void	(*teardown)(struct dsa_switch *ds);

include/uapi/linux/if_link.h

@@ -1375,4 +1375,14 @@ enum {
 
 #define IFLA_MCTP_MAX (__IFLA_MCTP_MAX - 1)
 
+/* DSA section */
+
+enum {
+	IFLA_DSA_UNSPEC,
+	IFLA_DSA_MASTER,
+	__IFLA_DSA_MAX,
+};
+
+#define IFLA_DSA_MAX	(__IFLA_DSA_MAX - 1)
+
 #endif /* _UAPI_LINUX_IF_LINK_H */

net/dsa/Makefile

@@ -1,7 +1,15 @@
 # SPDX-License-Identifier: GPL-2.0
 # the core
 obj-$(CONFIG_NET_DSA) += dsa_core.o
-dsa_core-y += dsa.o dsa2.o master.o port.o slave.o switch.o tag_8021q.o
+dsa_core-y += \
+	dsa.o \
+	dsa2.o \
+	master.o \
+	netlink.o \
+	port.o \
+	slave.o \
+	switch.o \
+	tag_8021q.o
 
 # tagging formats
 obj-$(CONFIG_NET_DSA_TAG_AR9331) += tag_ar9331.o

net/dsa/dsa.c

@@ -536,8 +536,16 @@ static int __init dsa_init_module(void)
 	dsa_tag_driver_register(&DSA_TAG_DRIVER_NAME(none_ops),
 				THIS_MODULE);
 
+	rc = rtnl_link_register(&dsa_link_ops);
+	if (rc)
+		goto netlink_register_fail;
+
 	return 0;
 
+netlink_register_fail:
+	dsa_tag_driver_unregister(&DSA_TAG_DRIVER_NAME(none_ops));
+	dsa_slave_unregister_notifier();
+	dev_remove_pack(&dsa_pack_type);
 register_notifier_fail:
 	destroy_workqueue(dsa_owq);
 
@@ -547,6 +555,7 @@ module_init(dsa_init_module);
 
 static void __exit dsa_cleanup_module(void)
 {
+	rtnl_link_unregister(&dsa_link_ops);
 	dsa_tag_driver_unregister(&DSA_TAG_DRIVER_NAME(none_ops));
 
 	dsa_slave_unregister_notifier();

net/dsa/dsa2.c

@@ -387,6 +387,20 @@ static struct dsa_port *dsa_tree_find_first_cpu(struct dsa_switch_tree *dst)
 	return NULL;
 }
 
+struct net_device *dsa_tree_find_first_master(struct dsa_switch_tree *dst)
+{
+	struct device_node *ethernet;
+	struct net_device *master;
+	struct dsa_port *cpu_dp;
+
+	cpu_dp = dsa_tree_find_first_cpu(dst);
+	ethernet = of_parse_phandle(cpu_dp->dn, "ethernet", 0);
+	master = of_find_net_device_by_node(ethernet);
+	of_node_put(ethernet);
+
+	return master;
+}
+
 /* Assign the default CPU port (the first one in the tree) to all ports of the
  * fabric which don't already have one as part of their own switch.
  */

net/dsa/dsa_priv.h

@@ -200,6 +200,9 @@ static inline struct net_device *dsa_master_find_slave(struct net_device *dev,
 	return NULL;
 }
 
+/* netlink.c */
+extern struct rtnl_link_ops dsa_link_ops __read_mostly;
+
 /* port.c */
 void dsa_port_set_tag_protocol(struct dsa_port *cpu_dp,
 			       const struct dsa_device_ops *tag_ops);
@@ -292,6 +295,8 @@ void dsa_port_hsr_leave(struct dsa_port *dp, struct net_device *hsr);
 int dsa_port_tag_8021q_vlan_add(struct dsa_port *dp, u16 vid, bool broadcast);
 void dsa_port_tag_8021q_vlan_del(struct dsa_port *dp, u16 vid, bool broadcast);
 void dsa_port_set_host_flood(struct dsa_port *dp, bool uc, bool mc);
+int dsa_port_change_master(struct dsa_port *dp, struct net_device *master,
+			   struct netlink_ext_ack *extack);
 
 /* slave.c */
 extern const struct dsa_device_ops notag_netdev_ops;
@@ -305,8 +310,12 @@ int dsa_slave_suspend(struct net_device *slave_dev);
 int dsa_slave_resume(struct net_device *slave_dev);
 int dsa_slave_register_notifier(void);
 void dsa_slave_unregister_notifier(void);
+void dsa_slave_sync_ha(struct net_device *dev);
+void dsa_slave_unsync_ha(struct net_device *dev);
 void dsa_slave_setup_tagger(struct net_device *slave);
 int dsa_slave_change_mtu(struct net_device *dev, int new_mtu);
+int dsa_slave_change_master(struct net_device *dev, struct net_device *master,
+			    struct netlink_ext_ack *extack);
 int dsa_slave_manage_vlan_filtering(struct net_device *dev,
 				    bool vlan_filtering);
 
@@ -542,6 +551,7 @@ void dsa_lag_map(struct dsa_switch_tree *dst, struct dsa_lag *lag);
 void dsa_lag_unmap(struct dsa_switch_tree *dst, struct dsa_lag *lag);
 struct dsa_lag *dsa_tree_lag_find(struct dsa_switch_tree *dst,
 				  const struct net_device *lag_dev);
+struct net_device *dsa_tree_find_first_master(struct dsa_switch_tree *dst);
 int dsa_tree_notify(struct dsa_switch_tree *dst, unsigned long e, void *v);
 int dsa_broadcast(unsigned long e, void *v);
 int dsa_tree_change_tag_proto(struct dsa_switch_tree *dst,

net/dsa/netlink.c

@@ -0,0 +1,62 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright 2022 NXP
+ */
+#include <linux/netdevice.h>
+#include <net/rtnetlink.h>
+
+#include "dsa_priv.h"
+
+static const struct nla_policy dsa_policy[IFLA_DSA_MAX + 1] = {
+	[IFLA_DSA_MASTER]	= { .type = NLA_U32 },
+};
+
+static int dsa_changelink(struct net_device *dev, struct nlattr *tb[],
+			  struct nlattr *data[],
+			  struct netlink_ext_ack *extack)
+{
+	int err;
+
+	if (!data)
+		return 0;
+
+	if (data[IFLA_DSA_MASTER]) {
+		u32 ifindex = nla_get_u32(data[IFLA_DSA_MASTER]);
+		struct net_device *master;
+
+		master = __dev_get_by_index(dev_net(dev), ifindex);
+		if (!master)
+			return -EINVAL;
+
+		err = dsa_slave_change_master(dev, master, extack);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static size_t dsa_get_size(const struct net_device *dev)
+{
+	return nla_total_size(sizeof(u32)) +	/* IFLA_DSA_MASTER  */
+	       0;
+}
+
+static int dsa_fill_info(struct sk_buff *skb, const struct net_device *dev)
+{
+	struct net_device *master = dsa_slave_to_master(dev);
+
+	if (nla_put_u32(skb, IFLA_DSA_MASTER, master->ifindex))
+		return -EMSGSIZE;
+
+	return 0;
+}
+
+struct rtnl_link_ops dsa_link_ops __read_mostly = {
+	.kind			= "dsa",
+	.priv_size		= sizeof(struct dsa_port),
+	.maxtype		= IFLA_DSA_MAX,
+	.policy			= dsa_policy,
+	.changelink		= dsa_changelink,
+	.get_size		= dsa_get_size,
+	.fill_info		= dsa_fill_info,
+};

net/dsa/port.c

@@ -7,6 +7,7 @@
  */
 
 #include <linux/if_bridge.h>
+#include <linux/netdevice.h>
 #include <linux/notifier.h>
 #include <linux/of_mdio.h>
 #include <linux/of_net.h>
@@ -1374,6 +1375,135 @@ int dsa_port_mrp_del_ring_role(const struct dsa_port *dp,
 	return ds->ops->port_mrp_del_ring_role(ds, dp->index, mrp);
 }
 
+static int dsa_port_assign_master(struct dsa_port *dp,
+				  struct net_device *master,
+				  struct netlink_ext_ack *extack,
+				  bool fail_on_err)
+{
+	struct dsa_switch *ds = dp->ds;
+	int port = dp->index, err;
+
+	err = ds->ops->port_change_master(ds, port, master, extack);
+	if (err && !fail_on_err)
+		dev_err(ds->dev, "port %d failed to assign master %s: %pe\n",
+			port, master->name, ERR_PTR(err));
+
+	if (err && fail_on_err)
+		return err;
+
+	dp->cpu_dp = master->dsa_ptr;
+
+	return 0;
+}
+
+/* Change the dp->cpu_dp affinity for a user port. Note that both cross-chip
+ * notifiers and drivers have implicit assumptions about user-to-CPU-port
+ * mappings, so we unfortunately cannot delay the deletion of the objects
+ * (switchdev, standalone addresses, standalone VLANs) on the old CPU port
+ * until the new CPU port has been set up. So we need to completely tear down
+ * the old CPU port before changing it, and restore it on errors during the
+ * bringup of the new one.
+ */
+int dsa_port_change_master(struct dsa_port *dp, struct net_device *master,
+			   struct netlink_ext_ack *extack)
+{
+	struct net_device *bridge_dev = dsa_port_bridge_dev_get(dp);
+	struct net_device *old_master = dsa_port_to_master(dp);
+	struct net_device *dev = dp->slave;
+	struct dsa_switch *ds = dp->ds;
+	bool vlan_filtering;
+	int err, tmp;
+
+	/* Bridges may hold host FDB, MDB and VLAN objects. These need to be
+	 * migrated, so dynamically unoffload and later reoffload the bridge
+	 * port.
+	 */
+	if (bridge_dev) {
+		dsa_port_pre_bridge_leave(dp, bridge_dev);
+		dsa_port_bridge_leave(dp, bridge_dev);
+	}
+
+	/* The port might still be VLAN filtering even if it's no longer
+	 * under a bridge, either due to ds->vlan_filtering_is_global or
+	 * ds->needs_standalone_vlan_filtering. In turn this means VLANs
+	 * on the CPU port.
+	 */
+	vlan_filtering = dsa_port_is_vlan_filtering(dp);
+	if (vlan_filtering) {
+		err = dsa_slave_manage_vlan_filtering(dev, false);
+		if (err) {
+			NL_SET_ERR_MSG_MOD(extack,
+					   "Failed to remove standalone VLANs");
+			goto rewind_old_bridge;
+		}
+	}
+
+	/* Standalone addresses, and addresses of upper interfaces like
+	 * VLAN, LAG, HSR need to be migrated.
+	 */
+	dsa_slave_unsync_ha(dev);
+
+	err = dsa_port_assign_master(dp, master, extack, true);
+	if (err)
+		goto rewind_old_addrs;
+
+	dsa_slave_sync_ha(dev);
+
+	if (vlan_filtering) {
+		err = dsa_slave_manage_vlan_filtering(dev, true);
+		if (err) {
+			NL_SET_ERR_MSG_MOD(extack,
+					   "Failed to restore standalone VLANs");
+			goto rewind_new_addrs;
+		}
+	}
+
+	if (bridge_dev) {
+		err = dsa_port_bridge_join(dp, bridge_dev, extack);
+		if (err && err == -EOPNOTSUPP) {
+			NL_SET_ERR_MSG_MOD(extack,
+					   "Failed to reoffload bridge");
+			goto rewind_new_vlan;
+		}
+	}
+
+	return 0;
+
+rewind_new_vlan:
+	if (vlan_filtering)
+		dsa_slave_manage_vlan_filtering(dev, false);
+
+rewind_new_addrs:
+	dsa_slave_unsync_ha(dev);
+
+	dsa_port_assign_master(dp, old_master, NULL, false);
+
+/* Restore the objects on the old CPU port */
+rewind_old_addrs:
+	dsa_slave_sync_ha(dev);
+
+	if (vlan_filtering) {
+		tmp = dsa_slave_manage_vlan_filtering(dev, true);
+		if (tmp) {
+			dev_err(ds->dev,
+				"port %d failed to restore standalone VLANs: %pe\n",
+				dp->index, ERR_PTR(tmp));
+		}
+	}
+
+rewind_old_bridge:
+	if (bridge_dev) {
+		tmp = dsa_port_bridge_join(dp, bridge_dev, extack);
+		if (tmp) {
+			dev_err(ds->dev,
+				"port %d failed to rejoin bridge %s: %pe\n",
+				dp->index, bridge_dev->name, ERR_PTR(tmp));
+		}
+	}
+
+	return err;
+}
+
 void dsa_port_set_tag_protocol(struct dsa_port *cpu_dp,
 			       const struct dsa_device_ops *tag_ops)
 {