net: core: Make the FIB notification chain generic

The FIB notification chain is currently soley used by IPv4 code.
However, we're going to introduce IPv6 FIB offload support, which
requires these notification as well.

As explained in commit c3852ef7f2 ("ipv4: fib: Replay events when
registering FIB notifier"), upon registration to the chain, the callee
receives a full dump of the FIB tables and rules by traversing all the
net namespaces. The integrity of the dump is ensured by a per-namespace
sequence counter that is incremented whenever a change to the tables or
rules occurs.

In order to allow more address families to use the chain, each family is
expected to register its fib_notifier_ops in its pernet init. These
operations allow the common code to read the family's sequence counter
as well as dump its tables and rules in the given net namespace.

Additionally, a 'family' parameter is added to sent notifications, so
that listeners could distinguish between the different families.

Implement the common code that allows listeners to register to the chain
and for address families to register their fib_notifier_ops. Subsequent
patches will implement these operations in IPv6.

In the future, ipmr and ip6mr will be extended to provide these
notifications as well.

Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c

@@ -53,6 +53,7 @@
 #include <net/addrconf.h>
 #include <net/ndisc.h>
 #include <net/ipv6.h>
+#include <net/fib_notifier.h>
 
 #include "spectrum.h"
 #include "core.h"

drivers/net/ethernet/rocker/rocker_main.c

@@ -34,6 +34,7 @@
 #include <net/netevent.h>
 #include <net/arp.h>
 #include <net/fib_rules.h>
+#include <net/fib_notifier.h>
 #include <linux/io-64-nonatomic-lo-hi.h>
 #include <generated/utsrelease.h>
 

include/net/fib_notifier.h

@@ -0,0 +1,44 @@
+#ifndef __NET_FIB_NOTIFIER_H
+#define __NET_FIB_NOTIFIER_H
+
+#include <linux/types.h>
+#include <linux/notifier.h>
+#include <net/net_namespace.h>
+
+struct fib_notifier_info {
+	struct net *net;
+	int family;
+};
+
+enum fib_event_type {
+	FIB_EVENT_ENTRY_REPLACE,
+	FIB_EVENT_ENTRY_APPEND,
+	FIB_EVENT_ENTRY_ADD,
+	FIB_EVENT_ENTRY_DEL,
+	FIB_EVENT_RULE_ADD,
+	FIB_EVENT_RULE_DEL,
+	FIB_EVENT_NH_ADD,
+	FIB_EVENT_NH_DEL,
+};
+
+struct fib_notifier_ops {
+	int family;
+	struct list_head list;
+	unsigned int (*fib_seq_read)(struct net *net);
+	int (*fib_dump)(struct net *net, struct notifier_block *nb);
+	struct rcu_head rcu;
+};
+
+int call_fib_notifier(struct notifier_block *nb, struct net *net,
+		      enum fib_event_type event_type,
+		      struct fib_notifier_info *info);
+int call_fib_notifiers(struct net *net, enum fib_event_type event_type,
+		       struct fib_notifier_info *info);
+int register_fib_notifier(struct notifier_block *nb,
+			  void (*cb)(struct notifier_block *nb));
+int unregister_fib_notifier(struct notifier_block *nb);
+struct fib_notifier_ops *
+fib_notifier_ops_register(const struct fib_notifier_ops *tmpl, struct net *net);
+void fib_notifier_ops_unregister(struct fib_notifier_ops *ops);
+
+#endif

include/net/ip_fib.h

@@ -19,6 +19,7 @@
 #include <net/flow.h>
 #include <linux/seq_file.h>
 #include <linux/rcupdate.h>
+#include <net/fib_notifier.h>
 #include <net/fib_rules.h>
 #include <net/inetpeer.h>
 #include <linux/percpu.h>
@@ -188,10 +189,6 @@ __be32 fib_info_update_nh_saddr(struct net *net, struct fib_nh *nh);
 #define FIB_RES_PREFSRC(net, res)	((res).fi->fib_prefsrc ? : \
 					 FIB_RES_SADDR(net, res))
 
-struct fib_notifier_info {
-	struct net *net;
-};
-
 struct fib_entry_notifier_info {
 	struct fib_notifier_info info; /* must be first */
 	u32 dst;
@@ -212,25 +209,14 @@ struct fib_nh_notifier_info {
 	struct fib_nh *fib_nh;
 };
 
-enum fib_event_type {
+int call_fib4_notifier(struct notifier_block *nb, struct net *net,
-	FIB_EVENT_ENTRY_REPLACE,
+		       enum fib_event_type event_type,
-	FIB_EVENT_ENTRY_APPEND,
-	FIB_EVENT_ENTRY_ADD,
-	FIB_EVENT_ENTRY_DEL,
-	FIB_EVENT_RULE_ADD,
-	FIB_EVENT_RULE_DEL,
-	FIB_EVENT_NH_ADD,
-	FIB_EVENT_NH_DEL,
-};
-
-int register_fib_notifier(struct notifier_block *nb,
-			  void (*cb)(struct notifier_block *nb));
-int unregister_fib_notifier(struct notifier_block *nb);
-int call_fib_notifier(struct notifier_block *nb, struct net *net,
-		      enum fib_event_type event_type,
-		      struct fib_notifier_info *info);
-int call_fib_notifiers(struct net *net, enum fib_event_type event_type,
 		       struct fib_notifier_info *info);
+int call_fib4_notifiers(struct net *net, enum fib_event_type event_type,
+			struct fib_notifier_info *info);
+
+int __net_init fib4_notifier_init(struct net *net);
+void __net_exit fib4_notifier_exit(struct net *net);
 
 void fib_notify(struct net *net, struct notifier_block *nb);
 #ifdef CONFIG_IP_MULTIPLE_TABLES

include/net/net_namespace.h

@@ -88,6 +88,7 @@ struct net {
 	/* core fib_rules */
 	struct list_head	rules_ops;
 
+	struct list_head	fib_notifier_ops;  /* protected by net_mutex */
 
 	struct net_device       *loopback_dev;          /* The loopback */
 	struct netns_core	core;

include/net/netns/ipv4.h

@@ -159,6 +159,7 @@ struct netns_ipv4 {
 	int sysctl_fib_multipath_hash_policy;
 #endif
 
+	struct fib_notifier_ops	*notifier_ops;
 	unsigned int	fib_seq;	/* protected by rtnl_mutex */
 
 	atomic_t	rt_genid;

net/core/Makefile

@@ -9,7 +9,8 @@ obj-$(CONFIG_SYSCTL) += sysctl_net_core.o
 
 obj-y		     += dev.o ethtool.o dev_addr_lists.o dst.o netevent.o \
 			neighbour.o rtnetlink.o utils.o link_watch.o filter.o \
-			sock_diag.o dev_ioctl.o tso.o sock_reuseport.o
+			sock_diag.o dev_ioctl.o tso.o sock_reuseport.o \
+			fib_notifier.o
 
 obj-y += net-sysfs.o
 obj-$(CONFIG_PROC_FS) += net-procfs.o

net/core/fib_notifier.c

@@ -0,0 +1,164 @@
+#include <linux/rtnetlink.h>
+#include <linux/notifier.h>
+#include <linux/rcupdate.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <net/net_namespace.h>
+#include <net/fib_notifier.h>
+
+static ATOMIC_NOTIFIER_HEAD(fib_chain);
+
+int call_fib_notifier(struct notifier_block *nb, struct net *net,
+		      enum fib_event_type event_type,
+		      struct fib_notifier_info *info)
+{
+	info->net = net;
+	return nb->notifier_call(nb, event_type, info);
+}
+EXPORT_SYMBOL(call_fib_notifier);
+
+int call_fib_notifiers(struct net *net, enum fib_event_type event_type,
+		       struct fib_notifier_info *info)
+{
+	info->net = net;
+	return atomic_notifier_call_chain(&fib_chain, event_type, info);
+}
+EXPORT_SYMBOL(call_fib_notifiers);
+
+static unsigned int fib_seq_sum(void)
+{
+	struct fib_notifier_ops *ops;
+	unsigned int fib_seq = 0;
+	struct net *net;
+
+	rtnl_lock();
+	for_each_net(net) {
+		list_for_each_entry(ops, &net->fib_notifier_ops, list)
+			fib_seq += ops->fib_seq_read(net);
+	}
+	rtnl_unlock();
+
+	return fib_seq;
+}
+
+static int fib_net_dump(struct net *net, struct notifier_block *nb)
+{
+	struct fib_notifier_ops *ops;
+
+	list_for_each_entry_rcu(ops, &net->fib_notifier_ops, list) {
+		int err = ops->fib_dump(net, nb);
+
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static bool fib_dump_is_consistent(struct notifier_block *nb,
+				   void (*cb)(struct notifier_block *nb),
+				   unsigned int fib_seq)
+{
+	atomic_notifier_chain_register(&fib_chain, nb);
+	if (fib_seq == fib_seq_sum())
+		return true;
+	atomic_notifier_chain_unregister(&fib_chain, nb);
+	if (cb)
+		cb(nb);
+	return false;
+}
+
+#define FIB_DUMP_MAX_RETRIES 5
+int register_fib_notifier(struct notifier_block *nb,
+			  void (*cb)(struct notifier_block *nb))
+{
+	int retries = 0;
+	int err;
+
+	do {
+		unsigned int fib_seq = fib_seq_sum();
+		struct net *net;
+
+		rcu_read_lock();
+		for_each_net_rcu(net) {
+			err = fib_net_dump(net, nb);
+			if (err)
+				goto err_fib_net_dump;
+		}
+		rcu_read_unlock();
+
+		if (fib_dump_is_consistent(nb, cb, fib_seq))
+			return 0;
+	} while (++retries < FIB_DUMP_MAX_RETRIES);
+
+	return -EBUSY;
+
+err_fib_net_dump:
+	rcu_read_unlock();
+	return err;
+}
+EXPORT_SYMBOL(register_fib_notifier);
+
+int unregister_fib_notifier(struct notifier_block *nb)
+{
+	return atomic_notifier_chain_unregister(&fib_chain, nb);
+}
+EXPORT_SYMBOL(unregister_fib_notifier);
+
+static int __fib_notifier_ops_register(struct fib_notifier_ops *ops,
+				       struct net *net)
+{
+	struct fib_notifier_ops *o;
+
+	list_for_each_entry(o, &net->fib_notifier_ops, list)
+		if (ops->family == o->family)
+			return -EEXIST;
+	list_add_tail_rcu(&ops->list, &net->fib_notifier_ops);
+	return 0;
+}
+
+struct fib_notifier_ops *
+fib_notifier_ops_register(const struct fib_notifier_ops *tmpl, struct net *net)
+{
+	struct fib_notifier_ops *ops;
+	int err;
+
+	ops = kmemdup(tmpl, sizeof(*ops), GFP_KERNEL);
+	if (!ops)
+		return ERR_PTR(-ENOMEM);
+
+	err = __fib_notifier_ops_register(ops, net);
+	if (err)
+		goto err_register;
+
+	return ops;
+
+err_register:
+	kfree(ops);
+	return ERR_PTR(err);
+}
+EXPORT_SYMBOL(fib_notifier_ops_register);
+
+void fib_notifier_ops_unregister(struct fib_notifier_ops *ops)
+{
+	list_del_rcu(&ops->list);
+	kfree_rcu(ops, rcu);
+}
+EXPORT_SYMBOL(fib_notifier_ops_unregister);
+
+static int __net_init fib_notifier_net_init(struct net *net)
+{
+	INIT_LIST_HEAD(&net->fib_notifier_ops);
+	return 0;
+}
+
+static struct pernet_operations fib_notifier_net_ops = {
+	.init = fib_notifier_net_init,
+};
+
+static int __init fib_notifier_init(void)
+{
+	return register_pernet_subsys(&fib_notifier_net_ops);
+}
+
+subsys_initcall(fib_notifier_init);

net/ipv4/fib_frontend.c

@@ -1247,22 +1247,28 @@ static int __net_init ip_fib_net_init(struct net *net)
 	int err;
 	size_t size = sizeof(struct hlist_head) * FIB_TABLE_HASHSZ;
 
-	net->ipv4.fib_seq = 0;
+	err = fib4_notifier_init(net);
+	if (err)
+		return err;
 
 	/* Avoid false sharing : Use at least a full cache line */
 	size = max_t(size_t, size, L1_CACHE_BYTES);
 
 	net->ipv4.fib_table_hash = kzalloc(size, GFP_KERNEL);
-	if (!net->ipv4.fib_table_hash)
+	if (!net->ipv4.fib_table_hash) {
-		return -ENOMEM;
+		err = -ENOMEM;
+		goto err_table_hash_alloc;
+	}
 
 	err = fib4_rules_init(net);
 	if (err < 0)
-		goto fail;
+		goto err_rules_init;
 	return 0;
 
-fail:
+err_rules_init:
 	kfree(net->ipv4.fib_table_hash);
+err_table_hash_alloc:
+	fib4_notifier_exit(net);
 	return err;
 }
 
@@ -1292,6 +1298,7 @@ static void ip_fib_net_exit(struct net *net)
 #endif
 	rtnl_unlock();
 	kfree(net->ipv4.fib_table_hash);
+	fib4_notifier_exit(net);
 }
 
 static int __net_init fib_net_init(struct net *net)

net/ipv4/fib_notifier.c

@@ -1,86 +1,66 @@
 #include <linux/rtnetlink.h>
 #include <linux/notifier.h>
-#include <linux/rcupdate.h>
+#include <linux/socket.h>
 #include <linux/kernel.h>
 #include <net/net_namespace.h>
+#include <net/fib_notifier.h>
 #include <net/netns/ipv4.h>
 #include <net/ip_fib.h>
 
-static ATOMIC_NOTIFIER_HEAD(fib_chain);
+int call_fib4_notifier(struct notifier_block *nb, struct net *net,
-
+		       enum fib_event_type event_type,
-int call_fib_notifier(struct notifier_block *nb, struct net *net,
-		      enum fib_event_type event_type,
-		      struct fib_notifier_info *info)
-{
-	info->net = net;
-	return nb->notifier_call(nb, event_type, info);
-}
-
-int call_fib_notifiers(struct net *net, enum fib_event_type event_type,
 		       struct fib_notifier_info *info)
 {
+	info->family = AF_INET;
+	return call_fib_notifier(nb, net, event_type, info);
+}
+
+int call_fib4_notifiers(struct net *net, enum fib_event_type event_type,
+			struct fib_notifier_info *info)
+{
+	ASSERT_RTNL();
+
+	info->family = AF_INET;
 	net->ipv4.fib_seq++;
-	info->net = net;
+	return call_fib_notifiers(net, event_type, info);
-	return atomic_notifier_call_chain(&fib_chain, event_type, info);
 }
 
-static unsigned int fib_seq_sum(void)
+static unsigned int fib4_seq_read(struct net *net)
 {
-	unsigned int fib_seq = 0;
+	ASSERT_RTNL();
-	struct net *net;
 
-	rtnl_lock();
+	return net->ipv4.fib_seq;
-	for_each_net(net)
-		fib_seq += net->ipv4.fib_seq;
-	rtnl_unlock();
-
-	return fib_seq;
 }
 
-static bool fib_dump_is_consistent(struct notifier_block *nb,
+static int fib4_dump(struct net *net, struct notifier_block *nb)
-				   void (*cb)(struct notifier_block *nb),
-				   unsigned int fib_seq)
 {
-	atomic_notifier_chain_register(&fib_chain, nb);
+	fib_rules_notify(net, nb);
-	if (fib_seq == fib_seq_sum())
+	fib_notify(net, nb);
-		return true;
+
-	atomic_notifier_chain_unregister(&fib_chain, nb);
+	return 0;
-	if (cb)
-		cb(nb);
-	return false;
 }
 
-#define FIB_DUMP_MAX_RETRIES 5
+static const struct fib_notifier_ops fib4_notifier_ops_template = {
-int register_fib_notifier(struct notifier_block *nb,
+	.family		= AF_INET,
-			  void (*cb)(struct notifier_block *nb))
+	.fib_seq_read	= fib4_seq_read,
+	.fib_dump	= fib4_dump,
+};
+
+int __net_init fib4_notifier_init(struct net *net)
 {
-	int retries = 0;
+	struct fib_notifier_ops *ops;
 
-	do {
+	net->ipv4.fib_seq = 0;
-		unsigned int fib_seq = fib_seq_sum();
-		struct net *net;
 
-		/* Mutex semantics guarantee that every change done to
+	ops = fib_notifier_ops_register(&fib4_notifier_ops_template, net);
-		 * FIB tries before we read the change sequence counter
+	if (IS_ERR(ops))
-		 * is now visible to us.
+		return PTR_ERR(ops);
-		 */
+	net->ipv4.notifier_ops = ops;
-		rcu_read_lock();
-		for_each_net_rcu(net) {
-			fib_rules_notify(net, nb);
-			fib_notify(net, nb);
-		}
-		rcu_read_unlock();
 
-		if (fib_dump_is_consistent(nb, cb, fib_seq))
+	return 0;
-			return 0;
-	} while (++retries < FIB_DUMP_MAX_RETRIES);
-
-	return -EBUSY;
 }
-EXPORT_SYMBOL(register_fib_notifier);
 
-int unregister_fib_notifier(struct notifier_block *nb)
+void __net_exit fib4_notifier_exit(struct net *net)
 {
-	return atomic_notifier_chain_unregister(&fib_chain, nb);
+	fib_notifier_ops_unregister(net->ipv4.notifier_ops);
 }
-EXPORT_SYMBOL(unregister_fib_notifier);

net/ipv4/fib_rules.c

@@ -32,6 +32,7 @@
 #include <net/tcp.h>
 #include <net/ip_fib.h>
 #include <net/fib_rules.h>
+#include <net/fib_notifier.h>
 
 struct fib4_rule {
 	struct fib_rule		common;
@@ -193,7 +194,7 @@ static int call_fib_rule_notifier(struct notifier_block *nb, struct net *net,
 		.rule = rule,
 	};
 
-	return call_fib_notifier(nb, net, event_type, &info.info);
+	return call_fib4_notifier(nb, net, event_type, &info.info);
 }
 
 static int call_fib_rule_notifiers(struct net *net,
@@ -204,7 +205,7 @@ static int call_fib_rule_notifiers(struct net *net,
 		.rule = rule,
 	};
 
-	return call_fib_notifiers(net, event_type, &info.info);
+	return call_fib4_notifiers(net, event_type, &info.info);
 }
 
 /* Called with rcu_read_lock() */

net/ipv4/fib_semantics.c

@@ -44,6 +44,7 @@
 #include <net/netlink.h>
 #include <net/nexthop.h>
 #include <net/lwtunnel.h>
+#include <net/fib_notifier.h>
 
 #include "fib_lookup.h"
 
@@ -1451,14 +1452,14 @@ static int call_fib_nh_notifiers(struct fib_nh *fib_nh,
 		if (IN_DEV_IGNORE_ROUTES_WITH_LINKDOWN(in_dev) &&
 		    fib_nh->nh_flags & RTNH_F_LINKDOWN)
 			break;
-		return call_fib_notifiers(dev_net(fib_nh->nh_dev), event_type,
+		return call_fib4_notifiers(dev_net(fib_nh->nh_dev), event_type,
-					  &info.info);
+					   &info.info);
 	case FIB_EVENT_NH_DEL:
 		if ((in_dev && IN_DEV_IGNORE_ROUTES_WITH_LINKDOWN(in_dev) &&
 		     fib_nh->nh_flags & RTNH_F_LINKDOWN) ||
 		    (fib_nh->nh_flags & RTNH_F_DEAD))
-			return call_fib_notifiers(dev_net(fib_nh->nh_dev),
+			return call_fib4_notifiers(dev_net(fib_nh->nh_dev),
-						  event_type, &info.info);
+						   event_type, &info.info);
 	default:
 		break;
 	}

net/ipv4/fib_trie.c

@@ -81,6 +81,7 @@
 #include <net/tcp.h>
 #include <net/sock.h>
 #include <net/ip_fib.h>
+#include <net/fib_notifier.h>
 #include <trace/events/fib.h>
 #include "fib_lookup.h"
 
@@ -97,7 +98,7 @@ static int call_fib_entry_notifier(struct notifier_block *nb, struct net *net,
 		.type = type,
 		.tb_id = tb_id,
 	};
-	return call_fib_notifier(nb, net, event_type, &info.info);
+	return call_fib4_notifier(nb, net, event_type, &info.info);
 }
 
 static int call_fib_entry_notifiers(struct net *net,
@@ -113,7 +114,7 @@ static int call_fib_entry_notifiers(struct net *net,
 		.type = type,
 		.tb_id = tb_id,
 	};
-	return call_fib_notifiers(net, event_type, &info.info);
+	return call_fib4_notifiers(net, event_type, &info.info);
 }
 
 #define MAX_STAT_DEPTH 32