Merge git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf

Pablo Neira Ayuso says:

====================
Netfilter fixes for net

The following patchset contains Netfilter fixes for net:

1) Fix warnings suspicious rcu usage when handling base chain
   statistics, from Taehee Yoo.

2) Refetch pointer to tcp header from nf_ct_sack_adjust() since
   skb_make_writable() may reallocate data area, reported by Google
   folks patch from Florian.

3) Incorrect netlink nest end after previous cancellation from error
   path in ipset, from Pan Bian.

4) Use dst_hold_safe() from nf_xfrm_me_harder(), from Florian.

5) Use rb_link_node_rcu() for rcu-protected rbtree node in
   nf_conncount, from Taehee Yoo.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

include/linux/netfilter/nfnetlink.h

@@ -62,18 +62,6 @@ static inline bool lockdep_nfnl_is_held(__u8 subsys_id)
 }
 #endif /* CONFIG_PROVE_LOCKING */
 
-/*
- * nfnl_dereference - fetch RCU pointer when updates are prevented by subsys mutex
- *
- * @p: The pointer to read, prior to dereferencing
- * @ss: The nfnetlink subsystem ID
- *
- * Return the value of the specified RCU-protected pointer, but omit
- * the READ_ONCE(), because caller holds the NFNL subsystem mutex.
- */
-#define nfnl_dereference(p, ss)					\
-	rcu_dereference_protected(p, lockdep_nfnl_is_held(ss))
-
 #define MODULE_ALIAS_NFNL_SUBSYS(subsys) \
 	MODULE_ALIAS("nfnetlink-subsys-" __stringify(subsys))
 

net/netfilter/ipset/ip_set_list_set.c

@@ -531,8 +531,8 @@ list_set_list(const struct ip_set *set,
 		ret = -EMSGSIZE;
 	} else {
 		cb->args[IPSET_CB_ARG0] = i;
+		ipset_nest_end(skb, atd);
 	}
-	ipset_nest_end(skb, atd);
 out:
 	rcu_read_unlock();
 	return ret;

net/netfilter/nf_conncount.c

@@ -427,7 +427,7 @@ insert_tree(struct net *net,
 	count = 1;
 	rbconn->list.count = count;
 
-	rb_link_node(&rbconn->node, parent, rbnode);
+	rb_link_node_rcu(&rbconn->node, parent, rbnode);
 	rb_insert_color(&rbconn->node, root);
 out_unlock:
 	spin_unlock_bh(&nf_conncount_locks[hash % CONNCOUNT_LOCK_SLOTS]);

net/netfilter/nf_conntrack_seqadj.c

@@ -115,12 +115,12 @@ static void nf_ct_sack_block_adjust(struct sk_buff *skb,
 /* TCP SACK sequence number adjustment */
 static unsigned int nf_ct_sack_adjust(struct sk_buff *skb,
 				      unsigned int protoff,
-				      struct tcphdr *tcph,
 				      struct nf_conn *ct,
 				      enum ip_conntrack_info ctinfo)
 {
-	unsigned int dir, optoff, optend;
+	struct tcphdr *tcph = (void *)skb->data + protoff;
 	struct nf_conn_seqadj *seqadj = nfct_seqadj(ct);
+	unsigned int dir, optoff, optend;
 
 	optoff = protoff + sizeof(struct tcphdr);
 	optend = protoff + tcph->doff * 4;
@@ -128,6 +128,7 @@ static unsigned int nf_ct_sack_adjust(struct sk_buff *skb,
 	if (!skb_make_writable(skb, optend))
 		return 0;
 
+	tcph = (void *)skb->data + protoff;
 	dir = CTINFO2DIR(ctinfo);
 
 	while (optoff < optend) {
@@ -207,7 +208,7 @@ int nf_ct_seq_adjust(struct sk_buff *skb,
 		 ntohl(newack));
 	tcph->ack_seq = newack;
 
-	res = nf_ct_sack_adjust(skb, protoff, tcph, ct, ctinfo);
+	res = nf_ct_sack_adjust(skb, protoff, ct, ctinfo);
 out:
 	spin_unlock_bh(&ct->lock);
 

net/netfilter/nf_nat_core.c

@@ -117,7 +117,8 @@ int nf_xfrm_me_harder(struct net *net, struct sk_buff *skb, unsigned int family)
 	dst = skb_dst(skb);
 	if (dst->xfrm)
 		dst = ((struct xfrm_dst *)dst)->route;
-	dst_hold(dst);
+	if (!dst_hold_safe(dst))
+		return -EHOSTUNREACH;
 
 	if (sk && !net_eq(net, sock_net(sk)))
 		sk = NULL;

net/netfilter/nf_tables_api.c

@@ -1216,7 +1216,8 @@ static int nf_tables_fill_chain_info(struct sk_buff *skb, struct net *net,
 		if (nla_put_string(skb, NFTA_CHAIN_TYPE, basechain->type->name))
 			goto nla_put_failure;
 
-		if (basechain->stats && nft_dump_stats(skb, basechain->stats))
+		if (rcu_access_pointer(basechain->stats) &&
+		    nft_dump_stats(skb, rcu_dereference(basechain->stats)))
 			goto nla_put_failure;
 	}
 
@@ -1392,7 +1393,8 @@ static struct nft_stats __percpu *nft_stats_alloc(const struct nlattr *attr)
 	return newstats;
 }
 
-static void nft_chain_stats_replace(struct nft_base_chain *chain,
+static void nft_chain_stats_replace(struct net *net,
+				    struct nft_base_chain *chain,
 				    struct nft_stats __percpu *newstats)
 {
 	struct nft_stats __percpu *oldstats;
@@ -1400,8 +1402,9 @@ static void nft_chain_stats_replace(struct nft_base_chain *chain,
 	if (newstats == NULL)
 		return;
 
-	if (chain->stats) {
-		oldstats = nfnl_dereference(chain->stats, NFNL_SUBSYS_NFTABLES);
+	if (rcu_access_pointer(chain->stats)) {
+		oldstats = rcu_dereference_protected(chain->stats,
+					lockdep_commit_lock_is_held(net));
 		rcu_assign_pointer(chain->stats, newstats);
 		synchronize_rcu();
 		free_percpu(oldstats);
@@ -1439,9 +1442,10 @@ static void nf_tables_chain_destroy(struct nft_ctx *ctx)
 		struct nft_base_chain *basechain = nft_base_chain(chain);
 
 		module_put(basechain->type->owner);
-		free_percpu(basechain->stats);
-		if (basechain->stats)
+		if (rcu_access_pointer(basechain->stats)) {
 			static_branch_dec(&nft_counters_enabled);
+			free_percpu(rcu_dereference_raw(basechain->stats));
+		}
 		kfree(chain->name);
 		kfree(basechain);
 	} else {
@@ -1590,7 +1594,7 @@ static int nf_tables_addchain(struct nft_ctx *ctx, u8 family, u8 genmask,
 				kfree(basechain);
 				return PTR_ERR(stats);
 			}
-			basechain->stats = stats;
+			rcu_assign_pointer(basechain->stats, stats);
 			static_branch_inc(&nft_counters_enabled);
 		}
 
@@ -6180,7 +6184,8 @@ static void nft_chain_commit_update(struct nft_trans *trans)
 		return;
 
 	basechain = nft_base_chain(trans->ctx.chain);
-	nft_chain_stats_replace(basechain, nft_trans_chain_stats(trans));
+	nft_chain_stats_replace(trans->ctx.net, basechain,
+				nft_trans_chain_stats(trans));
 
 	switch (nft_trans_chain_policy(trans)) {
 	case NF_DROP:

net/netfilter/nf_tables_core.c

@@ -101,7 +101,7 @@ static noinline void nft_update_chain_stats(const struct nft_chain *chain,
 	struct nft_stats *stats;
 
 	base_chain = nft_base_chain(chain);
-	if (!base_chain->stats)
+	if (!rcu_access_pointer(base_chain->stats))
 		return;
 
 	local_bh_disable();