[PATCH RFC net-next 10/18] net/ipv6: Make fib6_nh optional at the end of fib6_info

[dsahern@kernel.org]

From: David Ahern <dsahern@gmail.com>

Move fib6_nh to the end of fib6_info and make an array of
size 0. Pass a flag to fib6_info_alloc indicating if the
allocation needs to add space for a fib6_nh.

The current code path always has a fib6_nh allocated; with
nexthop objects they will not.

Signed-off-by: David Ahern <dsahern@gmail.com>
---
 include/net/ip6_fib.h       |   8 +--
 include/net/ip6_route.h     |  10 ++-
 include/trace/events/fib6.h |  15 ++--
 net/core/filter.c           |   6 +-
 net/ipv6/addrconf.c         |   2 +-
 net/ipv6/ip6_fib.c          |  15 ++--
 net/ipv6/ndisc.c            |  13 ++--
 net/ipv6/route.c            | 165 ++++++++++++++++++++++----------------------
 8 files changed, 124 insertions(+), 110 deletions(-)

diff --git a/include/net/ip6_fib.h b/include/net/ip6_fib.h
index 2a1fae1247a9..9526eef711d5 100644
--- a/include/net/ip6_fib.h
+++ b/include/net/ip6_fib.h
@@ -170,8 +170,8 @@ struct fib6_info {
 					dst_host:1,
 					unused:3;
 
-	struct fib6_nh			fib6_nh;
 	struct rcu_head			rcu;
+	struct fib6_nh			fib6_nh[0];
 };
 
 struct rt6_info {
@@ -274,7 +274,7 @@ static inline void ip6_rt_put(struct rt6_info *rt)
 	dst_release(&rt->dst);
 }
 
-struct fib6_info *fib6_info_alloc(gfp_t gfp_flags);
+struct fib6_info *fib6_info_alloc(gfp_t gfp_flags, bool with_fib6_nh);
 void fib6_info_destroy_rcu(struct rcu_head *head);
 
 static inline void fib6_info_hold(struct fib6_info *f6i)
@@ -426,13 +426,13 @@ static inline void fib6_nh_release(struct fib6_nh *fib6_nh)
 
 static inline struct net_device *fib6_info_nh_dev(const struct fib6_info *f6i)
 {
-	return f6i->fib6_nh.nh_dev;
+	return f6i->fib6_nh->nh_dev;
 }
 
 static inline
 struct lwtunnel_state *fib6_info_nh_lwt(const struct fib6_info *f6i)
 {
-	return f6i->fib6_nh.nh_lwtstate;
+	return f6i->fib6_nh->nh_lwtstate;
 }
 
 void inet6_rt_notify(int event, struct fib6_info *rt, struct nl_info *info,
diff --git a/include/net/ip6_route.h b/include/net/ip6_route.h
index 7b9c82de11cc..b1ca637acb2a 100644
--- a/include/net/ip6_route.h
+++ b/include/net/ip6_route.h
@@ -274,9 +274,13 @@ static inline struct in6_addr *rt6_nexthop(struct rt6_info *rt,
 
 static inline bool rt6_duplicate_nexthop(struct fib6_info *a, struct fib6_info *b)
 {
-	return a->fib6_nh.nh_dev == b->fib6_nh.nh_dev &&
-	       ipv6_addr_equal(&a->fib6_nh.nh_gw, &b->fib6_nh.nh_gw) &&
-	       !lwtunnel_cmp_encap(a->fib6_nh.nh_lwtstate, b->fib6_nh.nh_lwtstate);
+// TO-DO:
+	//if (a->nh || b->nh)
+	//	return nexthop_cmp(a->nh, b->nh);
+
+	return a->fib6_nh->nh_dev == b->fib6_nh->nh_dev &&
+	       ipv6_addr_equal(&a->fib6_nh->nh_gw, &b->fib6_nh->nh_gw) &&
+	       !lwtunnel_cmp_encap(a->fib6_nh->nh_lwtstate, b->fib6_nh->nh_lwtstate);
 }
 
 static inline unsigned int ip6_dst_mtu_forward(const struct dst_entry *dst)
diff --git a/include/trace/events/fib6.h b/include/trace/events/fib6.h
index b088b54d699c..037df3d2be0b 100644
--- a/include/trace/events/fib6.h
+++ b/include/trace/events/fib6.h
@@ -12,7 +12,7 @@
 
 TRACE_EVENT(fib6_table_lookup,
 
-	TP_PROTO(const struct net *net, const struct fib6_info *f6i,
+	TP_PROTO(const struct net *net, struct fib6_info *f6i,
 		 struct fib6_table *table, const struct flowi6 *flp),
 
 	TP_ARGS(net, f6i, table, flp),
@@ -36,6 +36,7 @@ TRACE_EVENT(fib6_table_lookup,
 	),
 
 	TP_fast_assign(
+		struct fib6_nh *fib6_nh = f6i->fib6_nh;
 		struct in6_addr *in6;
 
 		__entry->tb_id = table->tb6_id;
@@ -62,20 +63,20 @@ TRACE_EVENT(fib6_table_lookup,
 			__entry->dport = 0;
 		}
 
-		if (f6i->fib6_nh.nh_dev) {
-			__assign_str(name, f6i->fib6_nh.nh_dev);
+		if (fib6_nh && fib6_nh->nh_dev) {
+			__assign_str(name, fib6_nh->nh_dev);
 		} else {
 			__assign_str(name, "-");
 		}
-		if (f6i == net->ipv6.fib6_null_entry) {
+
+		if (!fib6_nh) {
 			struct in6_addr in6_zero = {};
 
 			in6 = (struct in6_addr *)__entry->gw;
 			*in6 = in6_zero;
-
-		} else if (f6i) {
+		} else {
 			in6 = (struct in6_addr *)__entry->gw;
-			*in6 = f6i->fib6_nh.nh_gw;
+			*in6 = fib6_nh->nh_gw;
 		}
 	),
 
diff --git a/net/core/filter.c b/net/core/filter.c
index 0ba4c477415d..bc979edf06ca 100644
--- a/net/core/filter.c
+++ b/net/core/filter.c
@@ -4428,13 +4428,13 @@ static int bpf_ipv6_fib_lookup(struct net *net, struct bpf_fib_lookup *params,
 			return BPF_FIB_LKUP_RET_FRAG_NEEDED;
 	}
 
-	if (f6i->fib6_nh.nh_lwtstate)
+	if (f6i->fib6_nh->nh_lwtstate)
 		return BPF_FIB_LKUP_RET_UNSUPP_LWT;
 
 	if (f6i->fib6_flags & RTF_GATEWAY)
-		*dst = f6i->fib6_nh.nh_gw;
+		*dst = f6i->fib6_nh->nh_gw;
 
-	dev = f6i->fib6_nh.nh_dev;
+	dev = f6i->fib6_nh->nh_dev;
 	params->rt_metric = f6i->fib6_metric;
 
 	/* xdp and cls_bpf programs are run in RCU-bh so rcu_read_lock_bh is
diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
index d51a8c0b3372..da5102bff2a9 100644
--- a/net/ipv6/addrconf.c
+++ b/net/ipv6/addrconf.c
@@ -2366,7 +2366,7 @@ static struct fib6_info *addrconf_get_prefix_route(const struct in6_addr *pfx,
 		goto out;
 
 	for_each_fib6_node_rt_rcu(fn) {
-		if (rt->fib6_nh.nh_dev->ifindex != dev->ifindex)
+		if (rt->fib6_nh->nh_dev->ifindex != dev->ifindex)
 			continue;
 		if ((rt->fib6_flags & flags) != flags)
 			continue;
diff --git a/net/ipv6/ip6_fib.c b/net/ipv6/ip6_fib.c
index c1c23427a81e..5b0ca5b3710d 100644
--- a/net/ipv6/ip6_fib.c
+++ b/net/ipv6/ip6_fib.c
@@ -145,11 +145,15 @@ static __be32 addr_bit_set(const void *token, int fn_bit)
 	       addr[fn_bit >> 5];
 }
 
-struct fib6_info *fib6_info_alloc(gfp_t gfp_flags)
+struct fib6_info *fib6_info_alloc(gfp_t gfp_flags, bool with_fib6_nh)
 {
 	struct fib6_info *f6i;
+	size_t sz = sizeof(*f6i);
 
-	f6i = kzalloc(sizeof(*f6i), gfp_flags);
+	if (with_fib6_nh)
+		sz += sizeof(struct fib6_nh);
+
+	f6i = kzalloc(sz, gfp_flags);
 	if (!f6i)
 		return NULL;
 
@@ -198,7 +202,7 @@ void fib6_info_destroy_rcu(struct rcu_head *head)
 		}
 	}
 
-	fib6_nh_release(&f6i->fib6_nh);
+	fib6_nh_release(f6i->fib6_nh);
 
 	m = f6i->fib6_metrics;
 	if (m != &dst_default_metrics && refcount_dec_and_test(&m->refcnt))
@@ -2247,6 +2251,7 @@ void fib6_gc_cleanup(void)
 static int ipv6_route_seq_show(struct seq_file *seq, void *v)
 {
 	struct fib6_info *rt = v;
+	struct fib6_nh *fib6_nh = rt->fib6_nh;
 	struct ipv6_route_iter *iter = seq->private;
 	const struct net_device *dev;
 
@@ -2258,11 +2263,11 @@ static int ipv6_route_seq_show(struct seq_file *seq, void *v)
 	seq_puts(seq, "00000000000000000000000000000000 00 ");
 #endif
 	if (rt->fib6_flags & RTF_GATEWAY)
-		seq_printf(seq, "%pi6", &rt->fib6_nh.nh_gw);
+		seq_printf(seq, "%pi6", &fib6_nh->nh_gw);
 	else
 		seq_puts(seq, "00000000000000000000000000000000");
 
-	dev = rt->fib6_nh.nh_dev;
+	dev = fib6_nh->nh_dev;
 	seq_printf(seq, " %08x %08x %08x %08x %8s\n",
 		   rt->fib6_metric, atomic_read(&rt->fib6_ref), 0,
 		   rt->fib6_flags, dev ? dev->name : "");
diff --git a/net/ipv6/ndisc.c b/net/ipv6/ndisc.c
index 0ec273997d1d..4bc47b9db35b 100644
--- a/net/ipv6/ndisc.c
+++ b/net/ipv6/ndisc.c
@@ -1156,6 +1156,7 @@ static void ndisc_router_discovery(struct sk_buff *skb)
 	struct neighbour *neigh = NULL;
 	struct inet6_dev *in6_dev;
 	struct fib6_info *rt = NULL;
+	struct fib6_nh *fib6_nh;
 	struct net *net;
 	int lifetime;
 	struct ndisc_options ndopts;
@@ -1276,9 +1277,9 @@ static void ndisc_router_discovery(struct sk_buff *skb)
 	rt = rt6_get_dflt_router(net, &ipv6_hdr(skb)->saddr, skb->dev);
 
 	if (rt) {
-		neigh = ip6_neigh_lookup(&rt->fib6_nh.nh_gw,
-					 rt->fib6_nh.nh_dev, NULL,
-					  &ipv6_hdr(skb)->saddr);
+		fib6_nh = rt->fib6_nh;
+		neigh = ip6_neigh_lookup(&fib6_nh->nh_gw, fib6_nh->nh_dev, NULL,
+					 &ipv6_hdr(skb)->saddr);
 		if (!neigh) {
 			ND_PRINTK(0, err,
 				  "RA: %s got default router without neighbour\n",
@@ -1306,9 +1307,9 @@ static void ndisc_router_discovery(struct sk_buff *skb)
 			return;
 		}
 
-		neigh = ip6_neigh_lookup(&rt->fib6_nh.nh_gw,
-					 rt->fib6_nh.nh_dev, NULL,
-					  &ipv6_hdr(skb)->saddr);
+		fib6_nh = rt->fib6_nh;
+		neigh = ip6_neigh_lookup(&fib6_nh->nh_gw, fib6_nh->nh_dev, NULL,
+					 &ipv6_hdr(skb)->saddr);
 		if (!neigh) {
 			ND_PRINTK(0, err,
 				  "RA: %s got default router without neighbour\n",
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index aa44cd5b3217..5792f57fdb91 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -439,14 +439,14 @@ struct fib6_info *fib6_multipath_select(const struct net *net,
 	if (!fl6->mp_hash)
 		fl6->mp_hash = rt6_multipath_hash(net, fl6, skb, NULL);
 
-	if (fl6->mp_hash <= atomic_read(&match->fib6_nh.nh_upper_bound))
+	if (fl6->mp_hash <= atomic_read(&match->fib6_nh->nh_upper_bound))
 		return match;
 
 	list_for_each_entry_safe(sibling, next_sibling, &match->fib6_siblings,
 				 fib6_siblings) {
 		int nh_upper_bound;
 
-		nh_upper_bound = atomic_read(&sibling->fib6_nh.nh_upper_bound);
+		nh_upper_bound = atomic_read(&sibling->fib6_nh->nh_upper_bound);
 		if (fl6->mp_hash > nh_upper_bound)
 			continue;
 		if (rt6_score_route(sibling, oif, strict) < 0)
@@ -471,13 +471,13 @@ static inline struct fib6_info *rt6_device_match(struct net *net,
 	struct fib6_info *sprt;
 
 	if (!oif && ipv6_addr_any(saddr) &&
-	    !(rt->fib6_nh.nh_flags & RTNH_F_DEAD))
+	    !(rt->fib6_nh->nh_flags & RTNH_F_DEAD))
 		return rt;
 
 	for (sprt = rt; sprt; sprt = rcu_dereference(sprt->fib6_next)) {
-		const struct net_device *dev = sprt->fib6_nh.nh_dev;
+		const struct net_device *dev = sprt->fib6_nh->nh_dev;
 
-		if (sprt->fib6_nh.nh_flags & RTNH_F_DEAD)
+		if (sprt->fib6_nh->nh_flags & RTNH_F_DEAD)
 			continue;
 
 		if (oif) {
@@ -493,7 +493,7 @@ static inline struct fib6_info *rt6_device_match(struct net *net,
 	if (oif && flags & RT6_LOOKUP_F_IFACE)
 		return net->ipv6.fib6_null_entry;
 
-	return rt->fib6_nh.nh_flags & RTNH_F_DEAD ? net->ipv6.fib6_null_entry : rt;
+	return rt->fib6_nh->nh_flags & RTNH_F_DEAD ? net->ipv6.fib6_null_entry : rt;
 }
 
 #ifdef CONFIG_IPV6_ROUTER_PREF
@@ -533,8 +533,8 @@ static void rt6_probe(struct fib6_info *rt)
 	if (!rt || !(rt->fib6_flags & RTF_GATEWAY))
 		return;
 
-	nh_gw = &rt->fib6_nh.nh_gw;
-	dev = rt->fib6_nh.nh_dev;
+	nh_gw = &rt->fib6_nh->nh_gw;
+	dev = rt->fib6_nh->nh_dev;
 	rcu_read_lock_bh();
 	neigh = __ipv6_neigh_lookup_noref(dev, nh_gw);
 	if (neigh) {
@@ -580,7 +580,7 @@ static inline void rt6_probe(struct fib6_info *rt)
  */
 static inline int rt6_check_dev(struct fib6_info *rt, int oif)
 {
-	const struct net_device *dev = rt->fib6_nh.nh_dev;
+	const struct net_device *dev = rt->fib6_nh->nh_dev;
 
 	if (!oif || dev->ifindex == oif)
 		return 2;
@@ -597,8 +597,8 @@ static inline enum rt6_nud_state rt6_check_neigh(struct fib6_info *rt)
 		return RT6_NUD_SUCCEED;
 
 	rcu_read_lock_bh();
-	neigh = __ipv6_neigh_lookup_noref(rt->fib6_nh.nh_dev,
-					  &rt->fib6_nh.nh_gw);
+	neigh = __ipv6_neigh_lookup_noref(rt->fib6_nh->nh_dev,
+					  &rt->fib6_nh->nh_gw);
 	if (neigh) {
 		read_lock(&neigh->lock);
 		if (neigh->nud_state & NUD_VALID)
@@ -638,6 +638,7 @@ static int rt6_score_route(struct fib6_info *rt, int oif, int strict)
 }
 
 /* called with rc_read_lock held */
+// TO-DO: if (!f6i->nh)
 static inline bool fib6_ignore_linkdown(const struct fib6_info *f6i)
 {
 	const struct net_device *dev = fib6_info_nh_dev(f6i);
@@ -659,11 +660,11 @@ static struct fib6_info *find_match(struct fib6_info *rt, int oif, int strict,
 	int m;
 	bool match_do_rr = false;
 
-	if (rt->fib6_nh.nh_flags & RTNH_F_DEAD)
+	if (rt->fib6_nh->nh_flags & RTNH_F_DEAD)
 		goto out;
 
 	if (fib6_ignore_linkdown(rt) &&
-	    rt->fib6_nh.nh_flags & RTNH_F_LINKDOWN &&
+	    rt->fib6_nh->nh_flags & RTNH_F_LINKDOWN &&
 	    !(strict & RT6_LOOKUP_F_IGNORE_LINKSTATE))
 		goto out;
 
@@ -868,7 +869,7 @@ int rt6_route_rcv(struct net_device *dev, u8 *opt, int len,
 /* called with rcu_lock held */
 static struct net_device *ip6_rt_get_dev_rcu(struct fib6_info *rt)
 {
-	struct net_device *dev = rt->fib6_nh.nh_dev;
+	struct net_device *dev = rt->fib6_nh->nh_dev;
 
 	if (rt->fib6_flags & (RTF_LOCAL | RTF_ANYCAST)) {
 		/* for copies of local routes, dst->dev needs to be the
@@ -964,8 +965,8 @@ static void ip6_rt_init_dst(struct rt6_info *rt, struct fib6_info *ort)
 		rt->dst.input = ip6_forward;
 	}
 
-	if (ort->fib6_nh.nh_lwtstate) {
-		rt->dst.lwtstate = lwtstate_get(ort->fib6_nh.nh_lwtstate);
+	if (ort->fib6_nh->nh_lwtstate) {
+		rt->dst.lwtstate = lwtstate_get(ort->fib6_nh->nh_lwtstate);
 		lwtunnel_set_redirect(&rt->dst);
 	}
 
@@ -989,14 +990,14 @@ static void ip6_rt_copy_init(struct rt6_info *rt, struct fib6_info *ort)
 
 	rt->rt6i_dst = ort->fib6_dst;
 	rt->rt6i_idev = dev ? in6_dev_get(dev) : NULL;
-	rt->rt6i_gateway = ort->fib6_nh.nh_gw;
+	rt->rt6i_gateway = ort->fib6_nh->nh_gw;
 	rt->rt6i_flags = ort->fib6_flags;
 	rt6_set_from(rt, ort);
 #ifdef CONFIG_IPV6_SUBTREES
 	rt->rt6i_src = ort->fib6_src;
 #endif
 	rt->rt6i_prefsrc = ort->fib6_prefsrc;
-	rt->dst.lwtstate = lwtstate_get(ort->fib6_nh.nh_lwtstate);
+	rt->dst.lwtstate = lwtstate_get(ort->fib6_nh->nh_lwtstate);
 }
 
 static struct fib6_node* fib6_backtrack(struct fib6_node *fn,
@@ -1038,7 +1039,7 @@ static bool ip6_hold_safe(struct net *net, struct rt6_info **prt,
 static struct rt6_info *ip6_create_rt_rcu(struct fib6_info *rt)
 {
 	unsigned short flags = fib6_info_dst_flags(rt);
-	struct net_device *dev = rt->fib6_nh.nh_dev;
+	struct net_device *dev = rt->fib6_nh->nh_dev;
 	struct rt6_info *nrt;
 
 	if (!fib6_info_hold_safe(rt))
@@ -1409,7 +1410,7 @@ static unsigned int fib6_mtu(const struct fib6_info *rt)
 
 	mtu = min_t(unsigned int, mtu, IP6_MAX_MTU);
 
-	return mtu - lwtunnel_headroom(rt->fib6_nh.nh_lwtstate, mtu);
+	return mtu - lwtunnel_headroom(rt->fib6_nh->nh_lwtstate, mtu);
 }
 
 static int rt6_insert_exception(struct rt6_info *nrt,
@@ -2453,7 +2454,7 @@ static struct rt6_info *__ip6_route_redirect(struct net *net,
 	fn = fib6_node_lookup(&table->tb6_root, &fl6->daddr, &fl6->saddr);
 restart:
 	for_each_fib6_node_rt_rcu(fn) {
-		if (rt->fib6_nh.nh_flags & RTNH_F_DEAD)
+		if (rt->fib6_nh->nh_flags & RTNH_F_DEAD)
 			continue;
 		if (fib6_check_expired(rt))
 			continue;
@@ -2461,14 +2462,14 @@ static struct rt6_info *__ip6_route_redirect(struct net *net,
 			break;
 		if (!(rt->fib6_flags & RTF_GATEWAY))
 			continue;
-		if (fl6->flowi6_oif != rt->fib6_nh.nh_dev->ifindex)
+		if (fl6->flowi6_oif != rt->fib6_nh->nh_dev->ifindex)
 			continue;
 		/* rt_cache's gateway might be different from its 'parent'
 		 * in the case of an ip redirect.
 		 * So we keep searching in the exception table if the gateway
 		 * is different.
 		 */
-		if (!ipv6_addr_equal(&rdfl->gateway, &rt->fib6_nh.nh_gw)) {
+		if (!ipv6_addr_equal(&rdfl->gateway, &rt->fib6_nh->nh_gw)) {
 			rt_cache = rt6_find_cached_rt(rt,
 						      &fl6->daddr,
 						      &fl6->saddr);
@@ -3004,8 +3005,7 @@ int fib6_nh_init(struct net *net, struct fib6_nh *fib6_nh,
 			}
 		}
 		cfg->fc_flags = RTF_REJECT | RTF_NONEXTHOP;
-		err = 0;
-		goto out;
+		goto set_dev;
 	}
 
 	if (cfg->fc_flags & RTF_GATEWAY) {
@@ -3036,7 +3036,9 @@ int fib6_nh_init(struct net *net, struct fib6_nh *fib6_nh,
 	    !netif_carrier_ok(dev))
 		fib6_nh->nh_flags |= RTNH_F_LINKDOWN;
 
+set_dev:
 	fib6_nh->nh_dev = dev;
+	err = 0;
 
 out:
 	if (idev)
@@ -3108,7 +3110,7 @@ static struct fib6_info *ip6_route_info_create(struct fib6_config *cfg,
 		goto out;
 
 	err = -ENOMEM;
-	rt = fib6_info_alloc(gfp_flags);
+	rt = fib6_info_alloc(gfp_flags, true);
 	if (!rt)
 		goto out;
 
@@ -3142,7 +3144,7 @@ static struct fib6_info *ip6_route_info_create(struct fib6_config *cfg,
 	ipv6_addr_prefix(&rt->fib6_src.addr, &cfg->fc_src, cfg->fc_src_len);
 	rt->fib6_src.plen = cfg->fc_src_len;
 #endif
-	err = fib6_nh_init(net, &rt->fib6_nh, cfg, extack);
+	err = fib6_nh_init(net, rt->fib6_nh, cfg, extack);
 	if (err)
 		goto out;
 
@@ -3318,11 +3320,11 @@ static int ip6_route_del(struct fib6_config *cfg,
 				continue;
 			}
 			if (cfg->fc_ifindex &&
-			    (!rt->fib6_nh.nh_dev ||
-			     rt->fib6_nh.nh_dev->ifindex != cfg->fc_ifindex))
+			    (!rt->fib6_nh->nh_dev ||
+			     rt->fib6_nh->nh_dev->ifindex != cfg->fc_ifindex))
 				continue;
 			if (cfg->fc_flags & RTF_GATEWAY &&
-			    !ipv6_addr_equal(&cfg->fc_gateway, &rt->fib6_nh.nh_gw))
+			    !ipv6_addr_equal(&cfg->fc_gateway, &rt->fib6_nh->nh_gw))
 				continue;
 			if (cfg->fc_metric && cfg->fc_metric != rt->fib6_metric)
 				continue;
@@ -3493,11 +3495,11 @@ static struct fib6_info *rt6_get_route_info(struct net *net,
 		goto out;
 
 	for_each_fib6_node_rt_rcu(fn) {
-		if (rt->fib6_nh.nh_dev->ifindex != ifindex)
+		if (rt->fib6_nh->nh_dev->ifindex != ifindex)
 			continue;
 		if ((rt->fib6_flags & (RTF_ROUTEINFO|RTF_GATEWAY)) != (RTF_ROUTEINFO|RTF_GATEWAY))
 			continue;
-		if (!ipv6_addr_equal(&rt->fib6_nh.nh_gw, gwaddr))
+		if (!ipv6_addr_equal(&rt->fib6_nh->nh_gw, gwaddr))
 			continue;
 		if (!fib6_info_hold_safe(rt))
 			continue;
@@ -3555,9 +3557,9 @@ struct fib6_info *rt6_get_dflt_router(struct net *net,
 
 	rcu_read_lock();
 	for_each_fib6_node_rt_rcu(&table->tb6_root) {
-		if (dev == rt->fib6_nh.nh_dev &&
+		if (dev == rt->fib6_nh->nh_dev &&
 		    ((rt->fib6_flags & (RTF_ADDRCONF | RTF_DEFAULT)) == (RTF_ADDRCONF | RTF_DEFAULT)) &&
-		    ipv6_addr_equal(&rt->fib6_nh.nh_gw, addr))
+		    ipv6_addr_equal(&rt->fib6_nh->nh_gw, addr))
 			break;
 	}
 	if (rt && !fib6_info_hold_safe(rt))
@@ -3763,7 +3765,7 @@ struct fib6_info *addrconf_f6i_alloc(struct net *net,
 	struct net_device *dev = idev->dev;
 	struct fib6_info *f6i;
 
-	f6i = fib6_info_alloc(gfp_flags);
+	f6i = fib6_info_alloc(gfp_flags, true);
 	if (!f6i)
 		return ERR_PTR(-ENOMEM);
 
@@ -3779,9 +3781,9 @@ struct fib6_info *addrconf_f6i_alloc(struct net *net,
 		f6i->fib6_flags |= RTF_LOCAL;
 	}
 
-	f6i->fib6_nh.nh_gw = *addr;
+	f6i->fib6_nh->nh_gw = *addr;
 	dev_hold(dev);
-	f6i->fib6_nh.nh_dev = dev;
+	f6i->fib6_nh->nh_dev = dev;
 	f6i->fib6_dst.addr = *addr;
 	f6i->fib6_dst.plen = 128;
 	tb_id = l3mdev_fib_table(idev->dev) ? : RT6_TABLE_LOCAL;
@@ -3803,7 +3805,7 @@ static int fib6_remove_prefsrc(struct fib6_info *rt, void *arg)
 	struct net *net = ((struct arg_dev_net_ip *)arg)->net;
 	struct in6_addr *addr = ((struct arg_dev_net_ip *)arg)->addr;
 
-	if (((void *)rt->fib6_nh.nh_dev == dev || !dev) &&
+	if (((void *)rt->fib6_nh->nh_dev == dev || !dev) &&
 	    rt != net->ipv6.fib6_null_entry &&
 	    ipv6_addr_equal(addr, &rt->fib6_prefsrc.addr)) {
 		spin_lock_bh(&rt6_exception_lock);
@@ -3835,7 +3837,7 @@ static int fib6_clean_tohost(struct fib6_info *rt, void *arg)
 	struct in6_addr *gateway = (struct in6_addr *)arg;
 
 	if (((rt->fib6_flags & RTF_RA_ROUTER) == RTF_RA_ROUTER) &&
-	    ipv6_addr_equal(gateway, &rt->fib6_nh.nh_gw)) {
+	    ipv6_addr_equal(gateway, &rt->fib6_nh->nh_gw)) {
 		return -1;
 	}
 
@@ -3883,8 +3885,8 @@ static struct fib6_info *rt6_multipath_first_sibling(const struct fib6_info *rt)
 
 static bool rt6_is_dead(const struct fib6_info *rt)
 {
-	if (rt->fib6_nh.nh_flags & RTNH_F_DEAD ||
-	    (rt->fib6_nh.nh_flags & RTNH_F_LINKDOWN &&
+	if (rt->fib6_nh->nh_flags & RTNH_F_DEAD ||
+	    (rt->fib6_nh->nh_flags & RTNH_F_LINKDOWN &&
 	     fib6_ignore_linkdown(rt)))
 		return true;
 
@@ -3897,11 +3899,11 @@ static int rt6_multipath_total_weight(const struct fib6_info *rt)
 	int total = 0;
 
 	if (!rt6_is_dead(rt))
-		total += rt->fib6_nh.nh_weight;
+		total += rt->fib6_nh->nh_weight;
 
 	list_for_each_entry(iter, &rt->fib6_siblings, fib6_siblings) {
 		if (!rt6_is_dead(iter))
-			total += iter->fib6_nh.nh_weight;
+			total += iter->fib6_nh->nh_weight;
 	}
 
 	return total;
@@ -3912,11 +3914,11 @@ static void rt6_upper_bound_set(struct fib6_info *rt, int *weight, int total)
 	int upper_bound = -1;
 
 	if (!rt6_is_dead(rt)) {
-		*weight += rt->fib6_nh.nh_weight;
+		*weight += rt->fib6_nh->nh_weight;
 		upper_bound = DIV_ROUND_CLOSEST_ULL((u64) (*weight) << 31,
 						    total) - 1;
 	}
-	atomic_set(&rt->fib6_nh.nh_upper_bound, upper_bound);
+	atomic_set(&rt->fib6_nh->nh_upper_bound, upper_bound);
 }
 
 static void rt6_multipath_upper_bound_set(struct fib6_info *rt, int total)
@@ -3959,8 +3961,9 @@ static int fib6_ifup(struct fib6_info *rt, void *p_arg)
 	const struct arg_netdev_event *arg = p_arg;
 	struct net *net = dev_net(arg->dev);
 
-	if (rt != net->ipv6.fib6_null_entry && rt->fib6_nh.nh_dev == arg->dev) {
-		rt->fib6_nh.nh_flags &= ~arg->nh_flags;
+	if (rt != net->ipv6.fib6_null_entry &&
+	    rt->fib6_nh->nh_dev == arg->dev) {
+		rt->fib6_nh->nh_flags &= ~arg->nh_flags;
 		fib6_update_sernum_upto_root(net, rt);
 		rt6_multipath_rebalance(rt);
 	}
@@ -3988,10 +3991,10 @@ static bool rt6_multipath_uses_dev(const struct fib6_info *rt,
 {
 	struct fib6_info *iter;
 
-	if (rt->fib6_nh.nh_dev == dev)
+	if (rt->fib6_nh->nh_dev == dev)
 		return true;
 	list_for_each_entry(iter, &rt->fib6_siblings, fib6_siblings)
-		if (iter->fib6_nh.nh_dev == dev)
+		if (iter->fib6_nh->nh_dev == dev)
 			return true;
 
 	return false;
@@ -4012,12 +4015,12 @@ static unsigned int rt6_multipath_dead_count(const struct fib6_info *rt,
 	struct fib6_info *iter;
 	unsigned int dead = 0;
 
-	if (rt->fib6_nh.nh_dev == down_dev ||
-	    rt->fib6_nh.nh_flags & RTNH_F_DEAD)
+	if (rt->fib6_nh->nh_dev == down_dev ||
+	    rt->fib6_nh->nh_flags & RTNH_F_DEAD)
 		dead++;
 	list_for_each_entry(iter, &rt->fib6_siblings, fib6_siblings)
-		if (iter->fib6_nh.nh_dev == down_dev ||
-		    iter->fib6_nh.nh_flags & RTNH_F_DEAD)
+		if (iter->fib6_nh->nh_dev == down_dev ||
+		    iter->fib6_nh->nh_flags & RTNH_F_DEAD)
 			dead++;
 
 	return dead;
@@ -4029,11 +4032,11 @@ static void rt6_multipath_nh_flags_set(struct fib6_info *rt,
 {
 	struct fib6_info *iter;
 
-	if (rt->fib6_nh.nh_dev == dev)
-		rt->fib6_nh.nh_flags |= nh_flags;
+	if (rt->fib6_nh->nh_dev == dev)
+		rt->fib6_nh->nh_flags |= nh_flags;
 	list_for_each_entry(iter, &rt->fib6_siblings, fib6_siblings)
-		if (iter->fib6_nh.nh_dev == dev)
-			iter->fib6_nh.nh_flags |= nh_flags;
+		if (iter->fib6_nh->nh_dev == dev)
+			iter->fib6_nh->nh_flags |= nh_flags;
 }
 
 /* called with write lock held for table with rt */
@@ -4048,12 +4051,12 @@ static int fib6_ifdown(struct fib6_info *rt, void *p_arg)
 
 	switch (arg->event) {
 	case NETDEV_UNREGISTER:
-		return rt->fib6_nh.nh_dev == dev ? -1 : 0;
+		return rt->fib6_nh->nh_dev == dev ? -1 : 0;
 	case NETDEV_DOWN:
 		if (rt->should_flush)
 			return -1;
 		if (!rt->fib6_nsiblings)
-			return rt->fib6_nh.nh_dev == dev ? -1 : 0;
+			return rt->fib6_nh->nh_dev == dev ? -1 : 0;
 		if (rt6_multipath_uses_dev(rt, dev)) {
 			unsigned int count;
 
@@ -4069,10 +4072,10 @@ static int fib6_ifdown(struct fib6_info *rt, void *p_arg)
 		}
 		return -2;
 	case NETDEV_CHANGE:
-		if (rt->fib6_nh.nh_dev != dev ||
+		if (rt->fib6_nh->nh_dev != dev ||
 		    rt->fib6_flags & (RTF_LOCAL | RTF_ANYCAST))
 			break;
-		rt->fib6_nh.nh_flags |= RTNH_F_LINKDOWN;
+		rt->fib6_nh->nh_flags |= RTNH_F_LINKDOWN;
 		rt6_multipath_rebalance(rt);
 		break;
 	}
@@ -4124,7 +4127,7 @@ static int rt6_mtu_change_route(struct fib6_info *rt, void *p_arg)
 	   Since RFC 1981 doesn't include administrative MTU increase
 	   update PMTU increase is a MUST. (i.e. jumbo frame)
 	 */
-	if (rt->fib6_nh.nh_dev == arg->dev &&
+	if (rt->fib6_nh->nh_dev == arg->dev &&
 	    !fib6_metric_locked(rt, RTAX_MTU)) {
 		u32 mtu = rt->fib6_pmtu;
 
@@ -4426,7 +4429,7 @@ static int ip6_route_multipath_add(struct fib6_config *cfg,
 			goto cleanup;
 		}
 
-		rt->fib6_nh.nh_weight = rtnh->rtnh_hops + 1;
+		rt->fib6_nh->nh_weight = rtnh->rtnh_hops + 1;
 
 		err = ip6_route_info_append(info->nl_net, &rt6_nh_list,
 					    rt, &r_cfg);
@@ -4589,7 +4592,7 @@ static size_t rt6_nlmsg_size(struct fib6_info *rt)
 		nexthop_len = nla_total_size(0)	 /* RTA_MULTIPATH */
 			    + NLA_ALIGN(sizeof(struct rtnexthop))
 			    + nla_total_size(16) /* RTA_GATEWAY */
-			    + lwtunnel_get_encap_size(rt->fib6_nh.nh_lwtstate);
+			    + lwtunnel_get_encap_size(rt->fib6_nh->nh_lwtstate);
 
 		nexthop_len *= rt->fib6_nsiblings;
 	}
@@ -4607,17 +4610,17 @@ static size_t rt6_nlmsg_size(struct fib6_info *rt)
 	       + nla_total_size(sizeof(struct rta_cacheinfo))
 	       + nla_total_size(TCP_CA_NAME_MAX) /* RTAX_CC_ALGO */
 	       + nla_total_size(1) /* RTA_PREF */
-	       + lwtunnel_get_encap_size(rt->fib6_nh.nh_lwtstate)
+	       + lwtunnel_get_encap_size(rt->fib6_nh->nh_lwtstate)
 	       + nexthop_len;
 }
 
 static int rt6_nexthop_info(struct sk_buff *skb, struct fib6_info *rt,
 			    unsigned int *flags, bool skip_oif)
 {
-	if (rt->fib6_nh.nh_flags & RTNH_F_DEAD)
+	if (rt->fib6_nh->nh_flags & RTNH_F_DEAD)
 		*flags |= RTNH_F_DEAD;
 
-	if (rt->fib6_nh.nh_flags & RTNH_F_LINKDOWN) {
+	if (rt->fib6_nh->nh_flags & RTNH_F_LINKDOWN) {
 		*flags |= RTNH_F_LINKDOWN;
 
 		rcu_read_lock();
@@ -4627,21 +4630,21 @@ static int rt6_nexthop_info(struct sk_buff *skb, struct fib6_info *rt,
 	}
 
 	if (rt->fib6_flags & RTF_GATEWAY) {
-		if (nla_put_in6_addr(skb, RTA_GATEWAY, &rt->fib6_nh.nh_gw) < 0)
+		if (nla_put_in6_addr(skb, RTA_GATEWAY, &rt->fib6_nh->nh_gw) < 0)
 			goto nla_put_failure;
 	}
 
-	*flags |= (rt->fib6_nh.nh_flags & RTNH_F_ONLINK);
-	if (rt->fib6_nh.nh_flags & RTNH_F_OFFLOAD)
+	*flags |= (rt->fib6_nh->nh_flags & RTNH_F_ONLINK);
+	if (rt->fib6_nh->nh_flags & RTNH_F_OFFLOAD)
 		*flags |= RTNH_F_OFFLOAD;
 
 	/* not needed for multipath encoding b/c it has a rtnexthop struct */
-	if (!skip_oif && rt->fib6_nh.nh_dev &&
-	    nla_put_u32(skb, RTA_OIF, rt->fib6_nh.nh_dev->ifindex))
+	if (!skip_oif && rt->fib6_nh->nh_dev &&
+	    nla_put_u32(skb, RTA_OIF, rt->fib6_nh->nh_dev->ifindex))
 		goto nla_put_failure;
 
-	if (rt->fib6_nh.nh_lwtstate &&
-	    lwtunnel_fill_encap(skb, rt->fib6_nh.nh_lwtstate) < 0)
+	if (rt->fib6_nh->nh_lwtstate &&
+	    lwtunnel_fill_encap(skb, rt->fib6_nh->nh_lwtstate) < 0)
 		goto nla_put_failure;
 
 	return 0;
@@ -4653,7 +4656,7 @@ static int rt6_nexthop_info(struct sk_buff *skb, struct fib6_info *rt,
 /* add multipath next hop */
 static int rt6_add_nexthop(struct sk_buff *skb, struct fib6_info *rt)
 {
-	const struct net_device *dev = rt->fib6_nh.nh_dev;
+	const struct net_device *dev = rt->fib6_nh->nh_dev;
 	struct rtnexthop *rtnh;
 	unsigned int flags = 0;
 
@@ -4661,7 +4664,7 @@ static int rt6_add_nexthop(struct sk_buff *skb, struct fib6_info *rt)
 	if (!rtnh)
 		goto nla_put_failure;
 
-	rtnh->rtnh_hops = rt->fib6_nh.nh_weight - 1;
+	rtnh->rtnh_hops = rt->fib6_nh->nh_weight - 1;
 	rtnh->rtnh_ifindex = dev ? dev->ifindex : 0;
 
 	if (rt6_nexthop_info(skb, rt, &flags, true) < 0)
@@ -5017,7 +5020,7 @@ static int ip6_route_dev_notify(struct notifier_block *this,
 		return NOTIFY_OK;
 
 	if (event == NETDEV_REGISTER) {
-		net->ipv6.fib6_null_entry->fib6_nh.nh_dev = dev;
+		net->ipv6.fib6_null_entry->fib6_nh->nh_dev = dev;
 		net->ipv6.ip6_null_entry->dst.dev = dev;
 		net->ipv6.ip6_null_entry->rt6i_idev = in6_dev_get(dev);
 #ifdef CONFIG_IPV6_MULTIPLE_TABLES
@@ -5194,11 +5197,11 @@ static int __net_init ip6_route_net_init(struct net *net)
 	if (dst_entries_init(&net->ipv6.ip6_dst_ops) < 0)
 		goto out_ip6_dst_ops;
 
-	net->ipv6.fib6_null_entry = kmemdup(&fib6_null_entry_template,
-					    sizeof(*net->ipv6.fib6_null_entry),
-					    GFP_KERNEL);
+	net->ipv6.fib6_null_entry = fib6_info_alloc(GFP_KERNEL, true);
 	if (!net->ipv6.fib6_null_entry)
 		goto out_ip6_dst_entries;
+	memcpy(net->ipv6.fib6_null_entry, &fib6_null_entry_template,
+	       sizeof(*net->ipv6.fib6_null_entry));
 
 	net->ipv6.ip6_null_entry = kmemdup(&ip6_null_entry_template,
 					   sizeof(*net->ipv6.ip6_null_entry),
@@ -5334,7 +5337,7 @@ void __init ip6_route_init_special_entries(void)
 	/* Registering of the loopback is done before this portion of code,
 	 * the loopback reference in rt6_info will not be taken, do it
 	 * manually for init_net */
-	init_net.ipv6.fib6_null_entry->fib6_nh.nh_dev = init_net.loopback_dev;
+	init_net.ipv6.fib6_null_entry->fib6_nh->nh_dev = init_net.loopback_dev;
 	init_net.ipv6.ip6_null_entry->dst.dev = init_net.loopback_dev;
 	init_net.ipv6.ip6_null_entry->rt6i_idev = in6_dev_get(init_net.loopback_dev);
   #ifdef CONFIG_IPV6_MULTIPLE_TABLES
-- 
2.11.0