Merge tag 'nf-24-09-26' of git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf

Pablo Neira Ayuso says:

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

v2: with kdoc fixes per Paolo Abeni.

The following patchset contains Netfilter fixes for net:

Patch #1 and #2 handle an esoteric scenario: Given two tasks sending UDP
packets to one another, two packets of the same flow in each direction
handled by different CPUs that result in two conntrack objects in NEW
state, where reply packet loses race. Then, patch #3 adds a testcase for
this scenario. Series from Florian Westphal.

1) NAT engine can falsely detect a port collision if it happens to pick
   up a reply packet as NEW rather than ESTABLISHED. Add extra code to
   detect this and suppress port reallocation in this case.

2) To complete the clash resolution in the reply direction, extend conntrack
   logic to detect clashing conntrack in the reply direction to existing entry.

3) Adds a test case.

Then, an assorted list of fixes follow:

4) Add a selftest for tproxy, from Antonio Ojea.

5) Guard ctnetlink_*_size() functions under
   #if defined(CONFIG_NETFILTER_NETLINK_GLUE_CT) || defined(CONFIG_NF_CONNTRACK_EVENTS)
   From Andy Shevchenko.

6) Use -m socket --transparent in iptables tproxy documentation.
   From XIE Zhibang.

7) Call kfree_rcu() when releasing flowtable hooks to address race with
   netlink dump path, from Phil Sutter.

8) Fix compilation warning in nf_reject with CONFIG_BRIDGE_NETFILTER=n.
   From Simon Horman.

9) Guard ctnetlink_label_size() under CONFIG_NF_CONNTRACK_EVENTS which
   is its only user, to address a compilation warning. From Simon Horman.

10) Use rcu-protected list iteration over basechain hooks from netlink
    dump path.

11) Fix memcg for nf_tables, use GFP_KERNEL_ACCOUNT is not complete.

12) Remove old nfqueue conntrack clash resolution. Instead trying to
    use same destination address consistently which requires double DNAT,
    use the existing clash resolution which allows clashing packets
    go through with different destination. Antonio Ojea originally
    reported an issue from the postrouting chain, I proposed a fix:
    https://lore.kernel.org/netfilter-devel/ZuwSwAqKgCB2a51-@calendula/T/
    which he reported it did not work for him.

13) Adds a selftest for patch 12.

14) Fixes ipvs.sh selftest.

netfilter pull request 24-09-26

* tag 'nf-24-09-26' of git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf:
  selftests: netfilter: Avoid hanging ipvs.sh
  kselftest: add test for nfqueue induced conntrack race
  netfilter: nfnetlink_queue: remove old clash resolution logic
  netfilter: nf_tables: missing objects with no memcg accounting
  netfilter: nf_tables: use rcu chain hook list iterator from netlink dump path
  netfilter: ctnetlink: compile ctnetlink_label_size with CONFIG_NF_CONNTRACK_EVENTS
  netfilter: nf_reject: Fix build warning when CONFIG_BRIDGE_NETFILTER=n
  netfilter: nf_tables: Keep deleted flowtable hooks until after RCU
  docs: tproxy: ignore non-transparent sockets in iptables
  netfilter: ctnetlink: Guard possible unused functions
  selftests: netfilter: nft_tproxy.sh: add tcp tests
  selftests: netfilter: add reverse-clash resolution test case
  netfilter: conntrack: add clash resolution for reverse collisions
  netfilter: nf_nat: don't try nat source port reallocation for reverse dir clash
====================

Link: https://patch.msgid.link/20240926110717.102194-1-pablo@netfilter.org
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

12 files changed, 197 insertions, 125 deletions

diff --git a/net/ipv4/netfilter/nf_reject_ipv4.c b/net/ipv4/netfilter/nf_reject_ipv4.c
index 04504b2b51df..87fd945a0d27 100644
--- a/net/ipv4/netfilter/nf_reject_ipv4.c
+++ b/net/ipv4/netfilter/nf_reject_ipv4.c
@@ -239,9 +239,8 @@ static int nf_reject_fill_skb_dst(struct sk_buff *skb_in)
 void nf_send_reset(struct net *net, struct sock *sk, struct sk_buff *oldskb,
 		   int hook)
 {
-	struct sk_buff *nskb;
-	struct iphdr *niph;
 	const struct tcphdr *oth;
+	struct sk_buff *nskb;
 	struct tcphdr _oth;
 
 	oth = nf_reject_ip_tcphdr_get(oldskb, &_oth, hook);
@@ -266,14 +265,12 @@ void nf_send_reset(struct net *net, struct sock *sk, struct sk_buff *oldskb,
 	nskb->mark = IP4_REPLY_MARK(net, oldskb->mark);
 
 	skb_reserve(nskb, LL_MAX_HEADER);
-	niph = nf_reject_iphdr_put(nskb, oldskb, IPPROTO_TCP,
-				   ip4_dst_hoplimit(skb_dst(nskb)));
+	nf_reject_iphdr_put(nskb, oldskb, IPPROTO_TCP,
+			    ip4_dst_hoplimit(skb_dst(nskb)));
 	nf_reject_ip_tcphdr_put(nskb, oldskb, oth);
 	if (ip_route_me_harder(net, sk, nskb, RTN_UNSPEC))
 		goto free_nskb;
 
-	niph = ip_hdr(nskb);
-
 	/* "Never happens" */
 	if (nskb->len > dst_mtu(skb_dst(nskb)))
 		goto free_nskb;
@@ -290,6 +287,7 @@ void nf_send_reset(struct net *net, struct sock *sk, struct sk_buff *oldskb,
 	 */
 	if (nf_bridge_info_exists(oldskb)) {
 		struct ethhdr *oeth = eth_hdr(oldskb);
+		struct iphdr *niph = ip_hdr(nskb);
 		struct net_device *br_indev;
 
 		br_indev = nf_bridge_get_physindev(oldskb, net);
diff --git a/net/ipv6/netfilter/nf_reject_ipv6.c b/net/ipv6/netfilter/nf_reject_ipv6.c
index b9457473c176..7db0437140bf 100644
--- a/net/ipv6/netfilter/nf_reject_ipv6.c
+++ b/net/ipv6/netfilter/nf_reject_ipv6.c
@@ -273,7 +273,6 @@ void nf_send_reset6(struct net *net, struct sock *sk, struct sk_buff *oldskb,
 	const struct tcphdr *otcph;
 	unsigned int otcplen, hh_len;
 	const struct ipv6hdr *oip6h = ipv6_hdr(oldskb);
-	struct ipv6hdr *ip6h;
 	struct dst_entry *dst = NULL;
 	struct flowi6 fl6;
 
@@ -329,8 +328,7 @@ void nf_send_reset6(struct net *net, struct sock *sk, struct sk_buff *oldskb,
 	nskb->mark = fl6.flowi6_mark;
 
 	skb_reserve(nskb, hh_len + dst->header_len);
-	ip6h = nf_reject_ip6hdr_put(nskb, oldskb, IPPROTO_TCP,
-				    ip6_dst_hoplimit(dst));
+	nf_reject_ip6hdr_put(nskb, oldskb, IPPROTO_TCP, ip6_dst_hoplimit(dst));
 	nf_reject_ip6_tcphdr_put(nskb, oldskb, otcph, otcplen);
 
 	nf_ct_attach(nskb, oldskb);
@@ -345,6 +343,7 @@ void nf_send_reset6(struct net *net, struct sock *sk, struct sk_buff *oldskb,
 	 */
 	if (nf_bridge_info_exists(oldskb)) {
 		struct ethhdr *oeth = eth_hdr(oldskb);
+		struct ipv6hdr *ip6h = ipv6_hdr(nskb);
 		struct net_device *br_indev;
 
 		br_indev = nf_bridge_get_physindev(oldskb, net);
diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c
index d3cb53b008f5..9db3e2b0b1c3 100644
--- a/net/netfilter/nf_conntrack_core.c
+++ b/net/netfilter/nf_conntrack_core.c
@@ -988,6 +988,56 @@ static void __nf_conntrack_insert_prepare(struct nf_conn *ct)
 		tstamp->start = ktime_get_real_ns();
 }
 
+/**
+ * nf_ct_match_reverse - check if ct1 and ct2 refer to identical flow
+ * @ct1: conntrack in hash table to check against
+ * @ct2: merge candidate
+ *
+ * returns true if ct1 and ct2 happen to refer to the same flow, but
+ * in opposing directions, i.e.
+ * ct1: a:b -> c:d
+ * ct2: c:d -> a:b
+ * for both directions.  If so, @ct2 should not have been created
+ * as the skb should have been picked up as ESTABLISHED flow.
+ * But ct1 was not yet committed to hash table before skb that created
+ * ct2 had arrived.
+ *
+ * Note we don't compare netns because ct entries in different net
+ * namespace cannot clash to begin with.
+ *
+ * @return: true if ct1 and ct2 are identical when swapping origin/reply.
+ */
+static bool
+nf_ct_match_reverse(const struct nf_conn *ct1, const struct nf_conn *ct2)
+{
+	u16 id1, id2;
+
+	if (!nf_ct_tuple_equal(&ct1->tuplehash[IP_CT_DIR_ORIGINAL].tuple,
+			       &ct2->tuplehash[IP_CT_DIR_REPLY].tuple))
+		return false;
+
+	if (!nf_ct_tuple_equal(&ct1->tuplehash[IP_CT_DIR_REPLY].tuple,
+			       &ct2->tuplehash[IP_CT_DIR_ORIGINAL].tuple))
+		return false;
+
+	id1 = nf_ct_zone_id(nf_ct_zone(ct1), IP_CT_DIR_ORIGINAL);
+	id2 = nf_ct_zone_id(nf_ct_zone(ct2), IP_CT_DIR_REPLY);
+	if (id1 != id2)
+		return false;
+
+	id1 = nf_ct_zone_id(nf_ct_zone(ct1), IP_CT_DIR_REPLY);
+	id2 = nf_ct_zone_id(nf_ct_zone(ct2), IP_CT_DIR_ORIGINAL);
+
+	return id1 == id2;
+}
+
+static int nf_ct_can_merge(const struct nf_conn *ct,
+			   const struct nf_conn *loser_ct)
+{
+	return nf_ct_match(ct, loser_ct) ||
+	       nf_ct_match_reverse(ct, loser_ct);
+}
+
 /* caller must hold locks to prevent concurrent changes */
 static int __nf_ct_resolve_clash(struct sk_buff *skb,
 				 struct nf_conntrack_tuple_hash *h)
@@ -999,11 +1049,7 @@ static int __nf_ct_resolve_clash(struct sk_buff *skb,
 
 	loser_ct = nf_ct_get(skb, &ctinfo);
 
-	if (nf_ct_is_dying(ct))
-		return NF_DROP;
-
-	if (((ct->status & IPS_NAT_DONE_MASK) == 0) ||
-	    nf_ct_match(ct, loser_ct)) {
+	if (nf_ct_can_merge(ct, loser_ct)) {
 		struct net *net = nf_ct_net(ct);
 
 		nf_conntrack_get(&ct->ct_general);
@@ -2151,80 +2197,6 @@ static void nf_conntrack_attach(struct sk_buff *nskb, const struct sk_buff *skb)
 	nf_conntrack_get(skb_nfct(nskb));
 }
 
-static int __nf_conntrack_update(struct net *net, struct sk_buff *skb,
-				 struct nf_conn *ct,
-				 enum ip_conntrack_info ctinfo)
-{
-	const struct nf_nat_hook *nat_hook;
-	struct nf_conntrack_tuple_hash *h;
-	struct nf_conntrack_tuple tuple;
-	unsigned int status;
-	int dataoff;
-	u16 l3num;
-	u8 l4num;
-
-	l3num = nf_ct_l3num(ct);
-
-	dataoff = get_l4proto(skb, skb_network_offset(skb), l3num, &l4num);
-	if (dataoff <= 0)
-		return NF_DROP;
-
-	if (!nf_ct_get_tuple(skb, skb_network_offset(skb), dataoff, l3num,
-			     l4num, net, &tuple))
-		return NF_DROP;
-
-	if (ct->status & IPS_SRC_NAT) {
-		memcpy(tuple.src.u3.all,
-		       ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple.src.u3.all,
-		       sizeof(tuple.src.u3.all));
-		tuple.src.u.all =
-			ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple.src.u.all;
-	}
-
-	if (ct->status & IPS_DST_NAT) {
-		memcpy(tuple.dst.u3.all,
-		       ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple.dst.u3.all,
-		       sizeof(tuple.dst.u3.all));
-		tuple.dst.u.all =
-			ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple.dst.u.all;
-	}
-
-	h = nf_conntrack_find_get(net, nf_ct_zone(ct), &tuple);
-	if (!h)
-		return NF_ACCEPT;
-
-	/* Store status bits of the conntrack that is clashing to re-do NAT
-	 * mangling according to what it has been done already to this packet.
-	 */
-	status = ct->status;
-
-	nf_ct_put(ct);
-	ct = nf_ct_tuplehash_to_ctrack(h);
-	nf_ct_set(skb, ct, ctinfo);
-
-	nat_hook = rcu_dereference(nf_nat_hook);
-	if (!nat_hook)
-		return NF_ACCEPT;
-
-	if (status & IPS_SRC_NAT) {
-		unsigned int verdict = nat_hook->manip_pkt(skb, ct,
-							   NF_NAT_MANIP_SRC,
-							   IP_CT_DIR_ORIGINAL);
-		if (verdict != NF_ACCEPT)
-			return verdict;
-	}
-
-	if (status & IPS_DST_NAT) {
-		unsigned int verdict = nat_hook->manip_pkt(skb, ct,
-							   NF_NAT_MANIP_DST,
-							   IP_CT_DIR_ORIGINAL);
-		if (verdict != NF_ACCEPT)
-			return verdict;
-	}
-
-	return NF_ACCEPT;
-}
-
 /* This packet is coming from userspace via nf_queue, complete the packet
  * processing after the helper invocation in nf_confirm().
  */
@@ -2288,17 +2260,6 @@ static int nf_conntrack_update(struct net *net, struct sk_buff *skb)
 	if (!ct)
 		return NF_ACCEPT;
 
-	if (!nf_ct_is_confirmed(ct)) {
-		int ret = __nf_conntrack_update(net, skb, ct, ctinfo);
-
-		if (ret != NF_ACCEPT)
-			return ret;
-
-		ct = nf_ct_get(skb, &ctinfo);
-		if (!ct)
-			return NF_ACCEPT;
-	}
-
 	return nf_confirm_cthelper(skb, ct, ctinfo);
 }
 
diff --git a/net/netfilter/nf_conntrack_netlink.c b/net/netfilter/nf_conntrack_netlink.c
index 123e2e933e9b..6a1239433830 100644
--- a/net/netfilter/nf_conntrack_netlink.c
+++ b/net/netfilter/nf_conntrack_netlink.c
@@ -382,7 +382,7 @@ nla_put_failure:
 #define ctnetlink_dump_secctx(a, b) (0)
 #endif
 
-#ifdef CONFIG_NF_CONNTRACK_LABELS
+#ifdef CONFIG_NF_CONNTRACK_EVENTS
 static inline int ctnetlink_label_size(const struct nf_conn *ct)
 {
 	struct nf_conn_labels *labels = nf_ct_labels_find(ct);
@@ -391,6 +391,7 @@ static inline int ctnetlink_label_size(const struct nf_conn *ct)
 		return 0;
 	return nla_total_size(sizeof(labels->bits));
 }
+#endif
 
 static int
 ctnetlink_dump_labels(struct sk_buff *skb, const struct nf_conn *ct)
@@ -411,10 +412,6 @@ ctnetlink_dump_labels(struct sk_buff *skb, const struct nf_conn *ct)
 
 	return 0;
 }
-#else
-#define ctnetlink_dump_labels(a, b) (0)
-#define ctnetlink_label_size(a)	(0)
-#endif
 
 #define master_tuple(ct) &(ct->master->tuplehash[IP_CT_DIR_ORIGINAL].tuple)
 
@@ -652,7 +649,6 @@ static size_t ctnetlink_proto_size(const struct nf_conn *ct)
 
 	return len + len4;
 }
-#endif
 
 static inline size_t ctnetlink_acct_size(const struct nf_conn *ct)
 {
@@ -690,6 +686,7 @@ static inline size_t ctnetlink_timestamp_size(const struct nf_conn *ct)
 	return 0;
 #endif
 }
+#endif
 
 #ifdef CONFIG_NF_CONNTRACK_EVENTS
 static size_t ctnetlink_nlmsg_size(const struct nf_conn *ct)
diff --git a/net/netfilter/nf_nat_core.c b/net/netfilter/nf_nat_core.c
index 6d8da6dddf99..4085c436e306 100644
--- a/net/netfilter/nf_nat_core.c
+++ b/net/netfilter/nf_nat_core.c
@@ -183,7 +183,35 @@ hash_by_src(const struct net *net,
 	return reciprocal_scale(hash, nf_nat_htable_size);
 }
 
-/* Is this tuple already taken? (not by us) */
+/**
+ * nf_nat_used_tuple - check if proposed nat tuple clashes with existing entry
+ * @tuple: proposed NAT binding
+ * @ignored_conntrack: our (unconfirmed) conntrack entry
+ *
+ * A conntrack entry can be inserted to the connection tracking table
+ * if there is no existing entry with an identical tuple in either direction.
+ *
+ * Example:
+ * INITIATOR -> NAT/PAT -> RESPONDER
+ *
+ * INITIATOR passes through NAT/PAT ("us") and SNAT is done (saddr rewrite).
+ * Then, later, NAT/PAT itself also connects to RESPONDER.
+ *
+ * This will not work if the SNAT done earlier has same IP:PORT source pair.
+ *
+ * Conntrack table has:
+ * ORIGINAL: $IP_INITIATOR:$SPORT -> $IP_RESPONDER:$DPORT
+ * REPLY:    $IP_RESPONDER:$DPORT -> $IP_NAT:$SPORT
+ *
+ * and new locally originating connection wants:
+ * ORIGINAL: $IP_NAT:$SPORT -> $IP_RESPONDER:$DPORT
+ * REPLY:    $IP_RESPONDER:$DPORT -> $IP_NAT:$SPORT
+ *
+ * ... which would mean incoming packets cannot be distinguished between
+ * the existing and the newly added entry (identical IP_CT_DIR_REPLY tuple).
+ *
+ * @return: true if the proposed NAT mapping collides with an existing entry.
+ */
 static int
 nf_nat_used_tuple(const struct nf_conntrack_tuple *tuple,
 		  const struct nf_conn *ignored_conntrack)
@@ -200,6 +228,94 @@ nf_nat_used_tuple(const struct nf_conntrack_tuple *tuple,
 	return nf_conntrack_tuple_taken(&reply, ignored_conntrack);
 }
 
+static bool nf_nat_allow_clash(const struct nf_conn *ct)
+{
+	return nf_ct_l4proto_find(nf_ct_protonum(ct))->allow_clash;
+}
+
+/**
+ * nf_nat_used_tuple_new - check if to-be-inserted conntrack collides with existing entry
+ * @tuple: proposed NAT binding
+ * @ignored_ct: our (unconfirmed) conntrack entry
+ *
+ * Same as nf_nat_used_tuple, but also check for rare clash in reverse
+ * direction. Should be called only when @tuple has not been altered, i.e.
+ * @ignored_conntrack will not be subject to NAT.
+ *
+ * @return: true if the proposed NAT mapping collides with existing entry.
+ */
+static noinline bool
+nf_nat_used_tuple_new(const struct nf_conntrack_tuple *tuple,
+		      const struct nf_conn *ignored_ct)
+{
+	static const unsigned long uses_nat = IPS_NAT_MASK | IPS_SEQ_ADJUST_BIT;
+	const struct nf_conntrack_tuple_hash *thash;
+	const struct nf_conntrack_zone *zone;
+	struct nf_conn *ct;
+	bool taken = true;
+	struct net *net;
+
+	if (!nf_nat_used_tuple(tuple, ignored_ct))
+		return false;
+
+	if (!nf_nat_allow_clash(ignored_ct))
+		return true;
+
+	/* Initial choice clashes with existing conntrack.
+	 * Check for (rare) reverse collision.
+	 *
+	 * This can happen when new packets are received in both directions
+	 * at the exact same time on different CPUs.
+	 *
+	 * Without SMP, first packet creates new conntrack entry and second
+	 * packet is resolved as established reply packet.
+	 *
+	 * With parallel processing, both packets could be picked up as
+	 * new and both get their own ct entry allocated.
+	 *
+	 * If ignored_conntrack and colliding ct are not subject to NAT then
+	 * pretend the tuple is available and let later clash resolution
+	 * handle this at insertion time.
+	 *
+	 * Without it, the 'reply' packet has its source port rewritten
+	 * by nat engine.
+	 */
+	if (READ_ONCE(ignored_ct->status) & uses_nat)
+		return true;
+
+	net = nf_ct_net(ignored_ct);
+	zone = nf_ct_zone(ignored_ct);
+
+	thash = nf_conntrack_find_get(net, zone, tuple);
+	if (unlikely(!thash)) /* clashing entry went away */
+		return false;
+
+	ct = nf_ct_tuplehash_to_ctrack(thash);
+
+	/* NB: IP_CT_DIR_ORIGINAL should be impossible because
+	 * nf_nat_used_tuple() handles origin collisions.
+	 *
+	 * Handle remote chance other CPU confirmed its ct right after.
+	 */
+	if (thash->tuple.dst.dir != IP_CT_DIR_REPLY)
+		goto out;
+
+	/* clashing connection subject to NAT? Retry with new tuple. */
+	if (READ_ONCE(ct->status) & uses_nat)
+		goto out;
+
+	if (nf_ct_tuple_equal(&ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple,
+			      &ignored_ct->tuplehash[IP_CT_DIR_REPLY].tuple) &&
+	    nf_ct_tuple_equal(&ct->tuplehash[IP_CT_DIR_REPLY].tuple,
+			      &ignored_ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple)) {
+		taken = false;
+		goto out;
+	}
+out:
+	nf_ct_put(ct);
+	return taken;
+}
+
 static bool nf_nat_may_kill(struct nf_conn *ct, unsigned long flags)
 {
 	static const unsigned long flags_refuse = IPS_FIXED_TIMEOUT |
@@ -611,7 +727,7 @@ get_unique_tuple(struct nf_conntrack_tuple *tuple,
 	    !(range->flags & NF_NAT_RANGE_PROTO_RANDOM_ALL)) {
 		/* try the original tuple first */
 		if (nf_in_range(orig_tuple, range)) {
-			if (!nf_nat_used_tuple(orig_tuple, ct)) {
+			if (!nf_nat_used_tuple_new(orig_tuple, ct)) {
 				*tuple = *orig_tuple;
 				return;
 			}
@@ -1208,7 +1324,6 @@ static const struct nf_nat_hook nat_hook = {
 #ifdef CONFIG_XFRM
 	.decode_session		= __nf_nat_decode_session,
 #endif
-	.manip_pkt		= nf_nat_manip_pkt,
 	.remove_nat_bysrc	= nf_nat_cleanup_conntrack,
 };
 
diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c
index 57259b5f3ef5..a24fe62650a7 100644
--- a/net/netfilter/nf_tables_api.c
+++ b/net/netfilter/nf_tables_api.c
@@ -1849,7 +1849,7 @@ static int nft_dump_basechain_hook(struct sk_buff *skb, int family,
 		if (!hook_list)
 			hook_list = &basechain->hook_list;
 
-		list_for_each_entry(hook, hook_list, list) {
+		list_for_each_entry_rcu(hook, hook_list, list) {
 			if (!first)
 				first = hook;
 
@@ -6684,7 +6684,7 @@ static int nft_setelem_catchall_insert(const struct net *net,
 		}
 	}
 
-	catchall = kmalloc(sizeof(*catchall), GFP_KERNEL);
+	catchall = kmalloc(sizeof(*catchall), GFP_KERNEL_ACCOUNT);
 	if (!catchall)
 		return -ENOMEM;
 
@@ -9207,7 +9207,7 @@ static void nf_tables_flowtable_destroy(struct nft_flowtable *flowtable)
 		flowtable->data.type->setup(&flowtable->data, hook->ops.dev,
 					    FLOW_BLOCK_UNBIND);
 		list_del_rcu(&hook->list);
-		kfree(hook);
+		kfree_rcu(hook, rcu);
 	}
 	kfree(flowtable->name);
 	module_put(flowtable->data.type->owner);
diff --git a/net/netfilter/nft_compat.c b/net/netfilter/nft_compat.c
index 52cdfee17f73..7ca4f0d21fe2 100644
--- a/net/netfilter/nft_compat.c
+++ b/net/netfilter/nft_compat.c
@@ -535,7 +535,7 @@ nft_match_large_init(const struct nft_ctx *ctx, const struct nft_expr *expr,
 	struct xt_match *m = expr->ops->data;
 	int ret;
 
-	priv->info = kmalloc(XT_ALIGN(m->matchsize), GFP_KERNEL);
+	priv->info = kmalloc(XT_ALIGN(m->matchsize), GFP_KERNEL_ACCOUNT);
 	if (!priv->info)
 		return -ENOMEM;
 
@@ -808,7 +808,7 @@ nft_match_select_ops(const struct nft_ctx *ctx,
 		goto err;
 	}
 
-	ops = kzalloc(sizeof(struct nft_expr_ops), GFP_KERNEL);
+	ops = kzalloc(sizeof(struct nft_expr_ops), GFP_KERNEL_ACCOUNT);
 	if (!ops) {
 		err = -ENOMEM;
 		goto err;
@@ -898,7 +898,7 @@ nft_target_select_ops(const struct nft_ctx *ctx,
 		goto err;
 	}
 
-	ops = kzalloc(sizeof(struct nft_expr_ops), GFP_KERNEL);
+	ops = kzalloc(sizeof(struct nft_expr_ops), GFP_KERNEL_ACCOUNT);
 	if (!ops) {
 		err = -ENOMEM;
 		goto err;
diff --git a/net/netfilter/nft_log.c b/net/netfilter/nft_log.c
index 5defe6e4fd98..e35588137995 100644
--- a/net/netfilter/nft_log.c
+++ b/net/netfilter/nft_log.c
@@ -163,7 +163,7 @@ static int nft_log_init(const struct nft_ctx *ctx,
 
 	nla = tb[NFTA_LOG_PREFIX];
 	if (nla != NULL) {
-		priv->prefix = kmalloc(nla_len(nla) + 1, GFP_KERNEL);
+		priv->prefix = kmalloc(nla_len(nla) + 1, GFP_KERNEL_ACCOUNT);
 		if (priv->prefix == NULL)
 			return -ENOMEM;
 		nla_strscpy(priv->prefix, nla, nla_len(nla) + 1);
diff --git a/net/netfilter/nft_meta.c b/net/netfilter/nft_meta.c
index 8c8eb14d647b..05cd1e6e6a2f 100644
--- a/net/netfilter/nft_meta.c
+++ b/net/netfilter/nft_meta.c
@@ -952,7 +952,7 @@ static int nft_secmark_obj_init(const struct nft_ctx *ctx,
 	if (tb[NFTA_SECMARK_CTX] == NULL)
 		return -EINVAL;
 
-	priv->ctx = nla_strdup(tb[NFTA_SECMARK_CTX], GFP_KERNEL);
+	priv->ctx = nla_strdup(tb[NFTA_SECMARK_CTX], GFP_KERNEL_ACCOUNT);
 	if (!priv->ctx)
 		return -ENOMEM;
 
diff --git a/net/netfilter/nft_numgen.c b/net/netfilter/nft_numgen.c
index 7d29db7c2ac0..bd058babfc82 100644
--- a/net/netfilter/nft_numgen.c
+++ b/net/netfilter/nft_numgen.c
@@ -66,7 +66,7 @@ static int nft_ng_inc_init(const struct nft_ctx *ctx,
 	if (priv->offset + priv->modulus - 1 < priv->offset)
 		return -EOVERFLOW;
 
-	priv->counter = kmalloc(sizeof(*priv->counter), GFP_KERNEL);
+	priv->counter = kmalloc(sizeof(*priv->counter), GFP_KERNEL_ACCOUNT);
 	if (!priv->counter)
 		return -ENOMEM;
 
diff --git a/net/netfilter/nft_set_pipapo.c b/net/netfilter/nft_set_pipapo.c
index eb4c4a4ac7ac..7be342b495f5 100644
--- a/net/netfilter/nft_set_pipapo.c
+++ b/net/netfilter/nft_set_pipapo.c
@@ -663,7 +663,7 @@ static int pipapo_realloc_mt(struct nft_pipapo_field *f,
 	    check_add_overflow(rules, extra, &rules_alloc))
 		return -EOVERFLOW;
 
-	new_mt = kvmalloc_array(rules_alloc, sizeof(*new_mt), GFP_KERNEL);
+	new_mt = kvmalloc_array(rules_alloc, sizeof(*new_mt), GFP_KERNEL_ACCOUNT);
 	if (!new_mt)
 		return -ENOMEM;
 
@@ -936,7 +936,7 @@ static void pipapo_lt_bits_adjust(struct nft_pipapo_field *f)
 		return;
 	}
 
-	new_lt = kvzalloc(lt_size + NFT_PIPAPO_ALIGN_HEADROOM, GFP_KERNEL);
+	new_lt = kvzalloc(lt_size + NFT_PIPAPO_ALIGN_HEADROOM, GFP_KERNEL_ACCOUNT);
 	if (!new_lt)
 		return;
 
@@ -1212,7 +1212,7 @@ static int pipapo_realloc_scratch(struct nft_pipapo_match *clone,
 		scratch = kzalloc_node(struct_size(scratch, map,
 						   bsize_max * 2) +
 				       NFT_PIPAPO_ALIGN_HEADROOM,
-				       GFP_KERNEL, cpu_to_node(i));
+				       GFP_KERNEL_ACCOUNT, cpu_to_node(i));
 		if (!scratch) {
 			/* On failure, there's no need to undo previous
 			 * allocations: this means that some scratch maps have
@@ -1427,7 +1427,7 @@ static struct nft_pipapo_match *pipapo_clone(struct nft_pipapo_match *old)
 	struct nft_pipapo_match *new;
 	int i;
 
-	new = kmalloc(struct_size(new, f, old->field_count), GFP_KERNEL);
+	new = kmalloc(struct_size(new, f, old->field_count), GFP_KERNEL_ACCOUNT);
 	if (!new)
 		return NULL;
 
@@ -1457,7 +1457,7 @@ static struct nft_pipapo_match *pipapo_clone(struct nft_pipapo_match *old)
 		new_lt = kvzalloc(src->groups * NFT_PIPAPO_BUCKETS(src->bb) *
 				  src->bsize * sizeof(*dst->lt) +
 				  NFT_PIPAPO_ALIGN_HEADROOM,
-				  GFP_KERNEL);
+				  GFP_KERNEL_ACCOUNT);
 		if (!new_lt)
 			goto out_lt;
 
@@ -1470,7 +1470,8 @@ static struct nft_pipapo_match *pipapo_clone(struct nft_pipapo_match *old)
 
 		if (src->rules > 0) {
 			dst->mt = kvmalloc_array(src->rules_alloc,
-						 sizeof(*src->mt), GFP_KERNEL);
+						 sizeof(*src->mt),
+						 GFP_KERNEL_ACCOUNT);
 			if (!dst->mt)
 				goto out_mt;
 
diff --git a/net/netfilter/nft_tunnel.c b/net/netfilter/nft_tunnel.c
index 60a76e6e348e..5c6ed68cc6e0 100644
--- a/net/netfilter/nft_tunnel.c
+++ b/net/netfilter/nft_tunnel.c
@@ -509,13 +509,14 @@ static int nft_tunnel_obj_init(const struct nft_ctx *ctx,
 			return err;
 	}
 
-	md = metadata_dst_alloc(priv->opts.len, METADATA_IP_TUNNEL, GFP_KERNEL);
+	md = metadata_dst_alloc(priv->opts.len, METADATA_IP_TUNNEL,
+				GFP_KERNEL_ACCOUNT);
 	if (!md)
 		return -ENOMEM;
 
 	memcpy(&md->u.tun_info, &info, sizeof(info));
 #ifdef CONFIG_DST_CACHE
-	err = dst_cache_init(&md->u.tun_info.dst_cache, GFP_KERNEL);
+	err = dst_cache_init(&md->u.tun_info.dst_cache, GFP_KERNEL_ACCOUNT);
 	if (err < 0) {
 		metadata_dst_free(md);
 		return err;