Merge branch 'net-metrics-consolidate'

David Ahern says:

====================
net: Consolidate metrics handling for ipv4 and ipv6

As part of the IPv6 fib info refactoring, the intent was to make metrics
handling for ipv6 identical to ipv4. One oversight in ip6_dst_destroy
led to confusion and a couple of incomplete attempts at finding and
fixing the resulting memory leak which was ultimately resolved by
ce7ea4af0838 ("ipv6: fix memory leak on dst->_metrics").

Refactor metrics hanlding make the code really identical for v4 and v6,
and add a few test cases.
====================

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

5 files changed, 50 insertions, 73 deletions

diff --git a/net/ipv4/fib_semantics.c b/net/ipv4/fib_semantics.c
index bee8db979195..f8c7ec8171a8 100644
--- a/net/ipv4/fib_semantics.c
+++ b/net/ipv4/fib_semantics.c
@@ -208,7 +208,6 @@ static void rt_fibinfo_free_cpus(struct rtable __rcu * __percpu *rtp)
 static void free_fib_info_rcu(struct rcu_head *head)
 {
 	struct fib_info *fi = container_of(head, struct fib_info, rcu);
-	struct dst_metrics *m;
 
 	change_nexthops(fi) {
 		if (nexthop_nh->nh_dev)
@@ -219,9 +218,8 @@ static void free_fib_info_rcu(struct rcu_head *head)
 		rt_fibinfo_free(&nexthop_nh->nh_rth_input);
 	} endfor_nexthops(fi);
 
-	m = fi->fib_metrics;
-	if (m != &dst_default_metrics && refcount_dec_and_test(&m->refcnt))
-		kfree(m);
+	ip_fib_metrics_put(fi->fib_metrics);
+
 	kfree(fi);
 }
 
@@ -1020,13 +1018,6 @@ static bool fib_valid_prefsrc(struct fib_config *cfg, __be32 fib_prefsrc)
 	return true;
 }
 
-static int
-fib_convert_metrics(struct fib_info *fi, const struct fib_config *cfg)
-{
-	return ip_metrics_convert(fi->fib_net, cfg->fc_mx, cfg->fc_mx_len,
-				  fi->fib_metrics->metrics);
-}
-
 struct fib_info *fib_create_info(struct fib_config *cfg,
 				 struct netlink_ext_ack *extack)
 {
@@ -1084,16 +1075,14 @@ struct fib_info *fib_create_info(struct fib_config *cfg,
 	fi = kzalloc(sizeof(*fi)+nhs*sizeof(struct fib_nh), GFP_KERNEL);
 	if (!fi)
 		goto failure;
-	if (cfg->fc_mx) {
-		fi->fib_metrics = kzalloc(sizeof(*fi->fib_metrics), GFP_KERNEL);
-		if (unlikely(!fi->fib_metrics)) {
-			kfree(fi);
-			return ERR_PTR(err);
-		}
-		refcount_set(&fi->fib_metrics->refcnt, 1);
-	} else {
-		fi->fib_metrics = (struct dst_metrics *)&dst_default_metrics;
+	fi->fib_metrics = ip_fib_metrics_init(fi->fib_net, cfg->fc_mx,
+					      cfg->fc_mx_len);
+	if (unlikely(IS_ERR(fi->fib_metrics))) {
+		err = PTR_ERR(fi->fib_metrics);
+		kfree(fi);
+		return ERR_PTR(err);
 	}
+
 	fib_info_cnt++;
 	fi->fib_net = net;
 	fi->fib_protocol = cfg->fc_protocol;
@@ -1112,10 +1101,6 @@ struct fib_info *fib_create_info(struct fib_config *cfg,
 			goto failure;
 	} endfor_nexthops(fi)
 
-	err = fib_convert_metrics(fi, cfg);
-	if (err)
-		goto failure;
-
 	if (cfg->fc_mp) {
 #ifdef CONFIG_IP_ROUTE_MULTIPATH
 		err = fib_get_nhs(fi, cfg->fc_mp, cfg->fc_mp_len, cfg, extack);
diff --git a/net/ipv4/metrics.c b/net/ipv4/metrics.c
index 04311f7067e2..6d218f5a2e71 100644
--- a/net/ipv4/metrics.c
+++ b/net/ipv4/metrics.c
@@ -5,8 +5,8 @@
 #include <net/net_namespace.h>
 #include <net/tcp.h>
 
-int ip_metrics_convert(struct net *net, struct nlattr *fc_mx, int fc_mx_len,
-		       u32 *metrics)
+static int ip_metrics_convert(struct net *net, struct nlattr *fc_mx,
+			      int fc_mx_len, u32 *metrics)
 {
 	bool ecn_ca = false;
 	struct nlattr *nla;
@@ -52,4 +52,28 @@ int ip_metrics_convert(struct net *net, struct nlattr *fc_mx, int fc_mx_len,
 
 	return 0;
 }
-EXPORT_SYMBOL_GPL(ip_metrics_convert);
+
+struct dst_metrics *ip_fib_metrics_init(struct net *net, struct nlattr *fc_mx,
+					int fc_mx_len)
+{
+	struct dst_metrics *fib_metrics;
+	int err;
+
+	if (!fc_mx)
+		return (struct dst_metrics *)&dst_default_metrics;
+
+	fib_metrics = kzalloc(sizeof(*fib_metrics), GFP_KERNEL);
+	if (unlikely(!fib_metrics))
+		return ERR_PTR(-ENOMEM);
+
+	err = ip_metrics_convert(net, fc_mx, fc_mx_len, fib_metrics->metrics);
+	if (!err) {
+		refcount_set(&fib_metrics->refcnt, 1);
+	} else {
+		kfree(fib_metrics);
+		fib_metrics = ERR_PTR(err);
+	}
+
+	return fib_metrics;
+}
+EXPORT_SYMBOL_GPL(ip_fib_metrics_init);
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 048919713f4e..f71d2395c428 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -1476,12 +1476,9 @@ void rt_del_uncached_list(struct rtable *rt)
 
 static void ipv4_dst_destroy(struct dst_entry *dst)
 {
-	struct dst_metrics *p = (struct dst_metrics *)DST_METRICS_PTR(dst);
 	struct rtable *rt = (struct rtable *)dst;
 
-	if (p != &dst_default_metrics && refcount_dec_and_test(&p->refcnt))
-		kfree(p);
-
+	ip_dst_metrics_put(dst);
 	rt_del_uncached_list(rt);
 }
 
@@ -1528,11 +1525,8 @@ static void rt_set_nexthop(struct rtable *rt, __be32 daddr,
 			rt->rt_gateway = nh->nh_gw;
 			rt->rt_uses_gateway = 1;
 		}
-		dst_init_metrics(&rt->dst, fi->fib_metrics->metrics, true);
-		if (fi->fib_metrics != &dst_default_metrics) {
-			rt->dst._metrics |= DST_METRICS_REFCOUNTED;
-			refcount_inc(&fi->fib_metrics->refcnt);
-		}
+		ip_dst_init_metrics(&rt->dst, fi->fib_metrics);
+
 #ifdef CONFIG_IP_ROUTE_CLASSID
 		rt->dst.tclassid = nh->nh_tclassid;
 #endif
diff --git a/net/ipv6/ip6_fib.c b/net/ipv6/ip6_fib.c
index 5516f55e214b..cf709eadc932 100644
--- a/net/ipv6/ip6_fib.c
+++ b/net/ipv6/ip6_fib.c
@@ -29,6 +29,7 @@
 #include <linux/list.h>
 #include <linux/slab.h>
 
+#include <net/ip.h>
 #include <net/ipv6.h>
 #include <net/ndisc.h>
 #include <net/addrconf.h>
@@ -160,8 +161,6 @@ struct fib6_info *fib6_info_alloc(gfp_t gfp_flags)
 	}
 
 	INIT_LIST_HEAD(&f6i->fib6_siblings);
-	f6i->fib6_metrics = (struct dst_metrics *)&dst_default_metrics;
-
 	atomic_inc(&f6i->fib6_ref);
 
 	return f6i;
@@ -171,7 +170,6 @@ void fib6_info_destroy_rcu(struct rcu_head *head)
 {
 	struct fib6_info *f6i = container_of(head, struct fib6_info, rcu);
 	struct rt6_exception_bucket *bucket;
-	struct dst_metrics *m;
 
 	WARN_ON(f6i->fib6_node);
 
@@ -203,9 +201,7 @@ void fib6_info_destroy_rcu(struct rcu_head *head)
 	if (f6i->fib6_nh.nh_dev)
 		dev_put(f6i->fib6_nh.nh_dev);
 
-	m = f6i->fib6_metrics;
-	if (m != &dst_default_metrics && refcount_dec_and_test(&m->refcnt))
-		kfree(m);
+	ip_fib_metrics_put(f6i->fib6_metrics);
 
 	kfree(f6i);
 }
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index 3adf107b42d2..6c1d817151ca 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -364,14 +364,11 @@ EXPORT_SYMBOL(ip6_dst_alloc);
 
 static void ip6_dst_destroy(struct dst_entry *dst)
 {
-	struct dst_metrics *p = (struct dst_metrics *)DST_METRICS_PTR(dst);
 	struct rt6_info *rt = (struct rt6_info *)dst;
 	struct fib6_info *from;
 	struct inet6_dev *idev;
 
-	if (p != &dst_default_metrics && refcount_dec_and_test(&p->refcnt))
-		kfree(p);
-
+	ip_dst_metrics_put(dst);
 	rt6_uncached_list_del(rt);
 
 	idev = rt->rt6i_idev;
@@ -978,11 +975,7 @@ static void rt6_set_from(struct rt6_info *rt, struct fib6_info *from)
 {
 	rt->rt6i_flags &= ~RTF_EXPIRES;
 	rcu_assign_pointer(rt->from, from);
-	dst_init_metrics(&rt->dst, from->fib6_metrics->metrics, true);
-	if (from->fib6_metrics != &dst_default_metrics) {
-		rt->dst._metrics |= DST_METRICS_REFCOUNTED;
-		refcount_inc(&from->fib6_metrics->refcnt);
-	}
+	ip_dst_init_metrics(&rt->dst, from->fib6_metrics);
 }
 
 /* Caller must already hold reference to @ort */
@@ -2705,24 +2698,6 @@ out:
 	return entries > rt_max_size;
 }
 
-static int ip6_convert_metrics(struct net *net, struct fib6_info *rt,
-			       struct fib6_config *cfg)
-{
-	struct dst_metrics *p;
-
-	if (!cfg->fc_mx)
-		return 0;
-
-	p = kzalloc(sizeof(*rt->fib6_metrics), GFP_KERNEL);
-	if (unlikely(!p))
-		return -ENOMEM;
-
-	refcount_set(&p->refcnt, 1);
-	rt->fib6_metrics = p;
-
-	return ip_metrics_convert(net, cfg->fc_mx, cfg->fc_mx_len, p->metrics);
-}
-
 static struct rt6_info *ip6_nh_lookup_table(struct net *net,
 					    struct fib6_config *cfg,
 					    const struct in6_addr *gw_addr,
@@ -2998,13 +2973,15 @@ static struct fib6_info *ip6_route_info_create(struct fib6_config *cfg,
 	if (!rt)
 		goto out;
 
+	rt->fib6_metrics = ip_fib_metrics_init(net, cfg->fc_mx, cfg->fc_mx_len);
+	if (IS_ERR(rt->fib6_metrics)) {
+		err = PTR_ERR(rt->fib6_metrics);
+		goto out;
+	}
+
 	if (cfg->fc_flags & RTF_ADDRCONF)
 		rt->dst_nocount = true;
 
-	err = ip6_convert_metrics(net, rt, cfg);
-	if (err < 0)
-		goto out;
-
 	if (cfg->fc_flags & RTF_EXPIRES)
 		fib6_set_expires(rt, jiffies +
 				clock_t_to_jiffies(cfg->fc_expires));
@@ -3727,6 +3704,7 @@ struct fib6_info *addrconf_f6i_alloc(struct net *net,
 	if (!f6i)
 		return ERR_PTR(-ENOMEM);
 
+	f6i->fib6_metrics = ip_fib_metrics_init(net, NULL, 0);
 	f6i->dst_nocount = true;
 	f6i->dst_host = true;
 	f6i->fib6_protocol = RTPROT_KERNEL;