Netfilter fixes for net

Hi David,

The following patchset contains Netfilter fixes for your net tree:

1) Skip ip_sabotage_in() for packet making into the VRF driver, otherwise packets are dropped, from David Ahern.

2) Clang compilation warning uncovering typo in the nft_validate_register_store() call from nft_osf, from Stefan Agner.

3) Double sizeof netlink message length calculations in ctnetlink, from zhong jiang.

4) Missing rb_erase() on batch full in rbtree garbage collector, from Taehee Yoo.

5) Calm down compilation warning in nf_hook(), from Florian Westphal.

6) Missing check for non-null sk in xt_socket before validating netns procedence, from Flavio Leitner.

You can pull these changes from:

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

Thanks.

----------------------------------------------------------------

The following changes since commit 56ce3c5a50f4d8cc95361b1ec7f152006c6320d8:

smc: generic netlink family should be __ro_after_init (2018-09-20 07:49:55 -0700)

are available in the git repository at:

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

for you to fetch changes up to 40e4f26e6a14fc1496eabb8b0004a547303114e6:

netfilter: xt_socket: check sk before checking for netns. (2018-09-28 14:47:41 +0200)

----------------------------------------------------------------
David Ahern (1):
netfilter: bridge: Don't sabotage nf_hook calls from an l3mdev

Flavio Leitner (1):
netfilter: xt_socket: check sk before checking for netns.

Florian Westphal (1):
netfilter: avoid erronous array bounds warning

Stefan Agner (1):
netfilter: nft_osf: use enum nft_data_types for nft_validate_register_store

Taehee Yoo (1):
netfilter: nft_set_rbtree: add missing rb_erase() in GC routine

zhong jiang (1):
netfilter: conntrack: get rid of double sizeof

 include/linux/netfilter.h              |  2 ++
 net/bridge/br_netfilter_hooks.c        |  3 ++-
 net/netfilter/nf_conntrack_proto_tcp.c |  4 ++--
 net/netfilter/nft_osf.c                |  2 +-
 net/netfilter/nft_set_rbtree.c         | 28 ++++++++++++++--------------
 net/netfilter/xt_socket.c              |  4 ++--
 6 files changed, 23 insertions(+), 20 deletions(-)

From: David Ahern <[email protected]>

For starters, the bridge netfilter code registers operations that are invoked any time nh_hook is called. Specifically, ip_sabotage_in watches for nested calls for NF_INET_PRE_ROUTING when a bridge is in the stack.

Packet wise, the bridge netfilter hook runs first. br_nf_pre_routing allocates nf_bridge, sets in_prerouting to 1 and calls NF_HOOK for NF_INET_PRE_ROUTING. It's finish function, br_nf_pre_routing_finish, then resets in_prerouting flag to 0 and the packet continues up the stack. The packet eventually makes it to the VRF driver and it invokes nf_hook for NF_INET_PRE_ROUTING in case any rules have been added against the vrf device.

Because of the registered operations the call to nf_hook causes ip_sabotage_in to be invoked. That function sees the nf_bridge on the skb and that in_prerouting is not set. Thinking it is an invalid nested call it steals (drops) the packet.

Update ip_sabotage_in to recognize that the bridge or one of its upper devices (e.g., vlan) can be enslaved to a VRF (L3 master device) and allow the packet to go through the nf_hook a second time.

---
 net/bridge/br_netfilter_hooks.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

index 6e0dc6bcd32a..37278dc280eb 100644
@@ -835,7 +835,8 @@ static unsigned int ip_sabotage_in(void *priv,
 				   struct sk_buff *skb,
 				   const struct nf_hook_state *state)
 {
-	if (skb->nf_bridge && !skb->nf_bridge->in_prerouting) {
+	if (skb->nf_bridge && !skb->nf_bridge->in_prerouting &&
+	    !netif_is_l3_master(skb->dev)) {
 		state->okfn(state->net, state->sk, skb);
 		return NF_STOLEN;
 	}

2.11.0

From: Stefan Agner <[email protected]>

The function nft_validate_register_store requires a struct of type struct nft_data_types. NFTA_DATA_VALUE is of type enum nft_verdict_attributes. Pass the correct enum type.

This fixes a warning seen with Clang:
net/netfilter/nft_osf.c:52:8: warning: implicit conversion from
enumeration type 'enum nft_data_attributes' to different enumeration type 'enum nft_data_types' [-Wenum-conversion]
NFTA_DATA_VALUE, NFT_OSF_MAXGENRELEN); ^~~~~~~~~~~~~~~

---
 net/netfilter/nft_osf.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 5af74b37f423..a35fb59ace73 100644
@@ -49,7 +49,7 @@ static int nft_osf_init(const struct nft_ctx *ctx,
 
 	priv->dreg = nft_parse_register(tb[NFTA_OSF_DREG]);
 	err = nft_validate_register_store(ctx, priv->dreg, NULL,
-					  NFTA_DATA_VALUE, NFT_OSF_MAXGENRELEN);
+					  NFT_DATA_VALUE, NFT_OSF_MAXGENRELEN);
 	if (err < 0)
 		return err;
 

2.11.0

From: zhong jiang <[email protected]>

sizeof(sizeof()) is quite strange and does not seem to be what is wanted here.

The issue is detected with the help of Coccinelle.

---
 net/netfilter/nf_conntrack_proto_tcp.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

index b4bdf9eda7b7..247b89784a6f 100644
@@ -1213,8 +1213,8 @@ static const struct nla_policy tcp_nla_policy[CTA_PROTOINFO_TCP_MAX+1] = {
 #define TCP_NLATTR_SIZE	( \
 	NLA_ALIGN(NLA_HDRLEN + 1) + \
 	NLA_ALIGN(NLA_HDRLEN + 1) + \
-	NLA_ALIGN(NLA_HDRLEN + sizeof(sizeof(struct nf_ct_tcp_flags))) + \
-	NLA_ALIGN(NLA_HDRLEN + sizeof(sizeof(struct nf_ct_tcp_flags))))
+	NLA_ALIGN(NLA_HDRLEN + sizeof(struct nf_ct_tcp_flags)) + \
+	NLA_ALIGN(NLA_HDRLEN + sizeof(struct nf_ct_tcp_flags)))
 
 static int nlattr_to_tcp(struct nlattr *cda[], struct nf_conn *ct)
 {

2.11.0

From: Taehee Yoo <[email protected]>

The nft_set_gc_batch_check() checks whether gc buffer is full. If gc buffer is full, gc buffer is released by
the nft_set_gc_batch_complete() internally.
In case of rbtree, the rb_erase() should be called before calling the nft_set_gc_batch_complete(). therefore the rb_erase() should be called before calling the nft_set_gc_batch_check() too.

test commands:

   table ip filter {
	   set set1 {
		   type ipv4_addr; flags interval, timeout;
		   gc-interval 10s;
		   timeout 1s;
		   elements = {
			   1-2,
			   3-4,
			   5-6,
			   ...
			   10000-10001,
		   }
	   }

}
%nft -f test.nft

splat looks like:

[  430.273885] kasan: GPF could be caused by NULL-ptr deref or user memory access
[  430.282158] general protection fault: 0000 [#1] SMP DEBUG_PAGEALLOC KASAN PTI
[  430.283116] CPU: 1 PID: 190 Comm: kworker/1:2 Tainted: G    B             4.18.0+ #7
[  430.283116] Workqueue: events_power_efficient nft_rbtree_gc [nf_tables_set]
[  430.313559] RIP: 0010:rb_next+0x81/0x130
[  430.313559] Code: 08 49 bd 00 00 00 00 00 fc ff df 48 bb 00 00 00 00 00 fc ff df 48 85 c0 75 05 eb 58 48 89 d4
[  430.313559] RSP: 0018:ffff88010cdb7680 EFLAGS: 00010207
[  430.313559] RAX: 0000000000b84854 RBX: dffffc0000000000 RCX: ffffffff83f01973
[  430.313559] RDX: 000000000017090c RSI: 0000000000000008 RDI: 0000000000b84864
[  430.313559] RBP: ffff8801060d4588 R08: fffffbfff09bc349 R09: fffffbfff09bc349
[  430.313559] R10: 0000000000000001 R11: fffffbfff09bc348 R12: ffff880100f081a8
[  430.313559] R13: dffffc0000000000 R14: ffff880100ff8688 R15: dffffc0000000000
[  430.313559] FS:  0000000000000000(0000) GS:ffff88011b400000(0000) knlGS:0000000000000000
[  430.313559] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  430.313559] CR2: 0000000001551008 CR3: 000000005dc16000 CR4: 00000000001006e0
[  430.313559] Call Trace:
[  430.313559]  nft_rbtree_gc+0x112/0x5c0 [nf_tables_set]
[  430.313559]  process_one_work+0xc13/0x1ec0
[  430.313559]  ? _raw_spin_unlock_irq+0x29/0x40
[  430.313559]  ? pwq_dec_nr_in_flight+0x3c0/0x3c0
[  430.313559]  ? set_load_weight+0x270/0x270
[  430.313559]  ? __switch_to_asm+0x34/0x70
[  430.313559]  ? __switch_to_asm+0x40/0x70
[  430.313559]  ? __switch_to_asm+0x34/0x70
[  430.313559]  ? __switch_to_asm+0x34/0x70
[  430.313559]  ? __switch_to_asm+0x40/0x70
[  430.313559]  ? __switch_to_asm+0x34/0x70
[  430.313559]  ? __switch_to_asm+0x40/0x70
[  430.313559]  ? __switch_to_asm+0x34/0x70
[  430.313559]  ? __switch_to_asm+0x34/0x70
[  430.313559]  ? __switch_to_asm+0x40/0x70
[  430.313559]  ? __switch_to_asm+0x34/0x70
[  430.313559]  ? __schedule+0x6d3/0x1f50
[  430.313559]  ? find_held_lock+0x39/0x1c0
[  430.313559]  ? __sched_text_start+0x8/0x8
[  430.313559]  ? cyc2ns_read_end+0x10/0x10
[  430.313559]  ? save_trace+0x300/0x300
[  430.313559]  ? sched_clock_local+0xd4/0x140
[  430.313559]  ? find_held_lock+0x39/0x1c0
[  430.313559]  ? worker_thread+0x353/0x1120
[  430.313559]  ? worker_thread+0x353/0x1120
[  430.313559]  ? lock_contended+0xe70/0xe70
[  430.313559]  ? __lock_acquire+0x4500/0x4500
[  430.535635]  ? do_raw_spin_unlock+0xa5/0x330
[  430.535635]  ? do_raw_spin_trylock+0x101/0x1a0
[  430.535635]  ? do_raw_spin_lock+0x1f0/0x1f0
[  430.535635]  ? _raw_spin_lock_irq+0x10/0x70
[  430.535635]  worker_thread+0x15d/0x1120

[ ... ]

---
 net/netfilter/nft_set_rbtree.c | 28 ++++++++++++++--------------
 1 file changed, 14 insertions(+), 14 deletions(-)

index 55e2d9215c0d..0e5ec126f6ad 100644
@@ -355,12 +355,11 @@ static void nft_rbtree_walk(const struct nft_ctx *ctx,
 
 static void nft_rbtree_gc(struct work_struct *work)
 {
+	struct nft_rbtree_elem *rbe, *rbe_end = NULL, *rbe_prev = NULL;
 	struct nft_set_gc_batch *gcb = NULL;
-	struct rb_node *node, *prev = NULL;
-	struct nft_rbtree_elem *rbe;
 	struct nft_rbtree *priv;
+	struct rb_node *node;
 	struct nft_set *set;
-	int i;
 
 	priv = container_of(work, struct nft_rbtree, gc_work.work);
 	set  = nft_set_container_of(priv);

@@ -371,7 +370,7 @@ static void nft_rbtree_gc(struct work_struct *work)
 		rbe = rb_entry(node, struct nft_rbtree_elem, node);
 
 		if (nft_rbtree_interval_end(rbe)) {
-			prev = node;
+			rbe_end = rbe;
 			continue;
 		}
 		if (!nft_set_elem_expired(&rbe->ext))

@@ -379,29 +378,30 @@ static void nft_rbtree_gc(struct work_struct *work)
 		if (nft_set_elem_mark_busy(&rbe->ext))
 			continue;
 
+		if (rbe_prev) {
+			rb_erase(&rbe_prev->node, &priv->root);
+			rbe_prev = NULL;
+		}
 		gcb = nft_set_gc_batch_check(set, gcb, GFP_ATOMIC);
 		if (!gcb)
 			break;
 
 		atomic_dec(&set->nelems);
 		nft_set_gc_batch_add(gcb, rbe);
+		rbe_prev = rbe;
 
-		if (prev) {
-			rbe = rb_entry(prev, struct nft_rbtree_elem, node);
+		if (rbe_end) {
 			atomic_dec(&set->nelems);
-			nft_set_gc_batch_add(gcb, rbe);
-			prev = NULL;
+			nft_set_gc_batch_add(gcb, rbe_end);
+			rb_erase(&rbe_end->node, &priv->root);
+			rbe_end = NULL;
 		}
 		node = rb_next(node);
 		if (!node)
 			break;
 	}
-	if (gcb) {
-		for (i = 0; i < gcb->head.cnt; i++) {
-			rbe = gcb->elems[i];
-			rb_erase(&rbe->node, &priv->root);
-		}
-	}
+	if (rbe_prev)
+		rb_erase(&rbe_prev->node, &priv->root);
 	write_seqcount_end(&priv->count);
 	write_unlock_bh(&priv->lock);
 

2.11.0

From: Florian Westphal <[email protected]>

Unfortunately some versions of gcc emit following warning:

  $ make net/xfrm/xfrm_output.o
  linux/compiler.h:252:20: warning: array subscript is above array bounds [-Warray-bounds]
  hook_head = rcu_dereference(net->nf.hooks_arp[hook]);
                            ^~~~~~~~~~~~~~~~~~~~~
xfrm_output_resume passes skb_dst(skb)->ops->family as its 'pf' arg so compiler
can't know that we'll never access hooks_arp[].
(NFPROTO_IPV4 or NFPROTO_IPV6 are only possible cases).

Avoid this by adding an explicit WARN_ON_ONCE() check.

This patch has no effect if the family is a compile-time constant as gcc will remove the switch() construct entirely.

---
 include/linux/netfilter.h | 2 ++
 1 file changed, 2 insertions(+)

index 07efffd0c759..bbe99d2b28b4 100644
@@ -215,6 +215,8 @@ static inline int nf_hook(u_int8_t pf, unsigned int hook, struct net *net,
 		break;
 	case NFPROTO_ARP:
 #ifdef CONFIG_NETFILTER_FAMILY_ARP
+		if (WARN_ON_ONCE(hook >= ARRAY_SIZE(net->nf.hooks_arp)))
+			break;
 		hook_head = rcu_dereference(net->nf.hooks_arp[hook]);
 #endif
 		break;

2.11.0

From: Flavio Leitner <[email protected]>

Only check for the network namespace if the socket is available.

---
 net/netfilter/xt_socket.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

index 0472f3472842..ada144e5645b 100644
@@ -56,7 +56,7 @@ socket_match(const struct sk_buff *skb, struct xt_action_param *par,
 	struct sk_buff *pskb = (struct sk_buff *)skb;
 	struct sock *sk = skb->sk;
 
-	if (!net_eq(xt_net(par), sock_net(sk)))
+	if (sk && !net_eq(xt_net(par), sock_net(sk)))
 		sk = NULL;
 
 	if (!sk)

@@ -117,7 +117,7 @@ socket_mt6_v1_v2_v3(const struct sk_buff *skb, struct xt_action_param *par)
 	struct sk_buff *pskb = (struct sk_buff *)skb;
 	struct sock *sk = skb->sk;
 
-	if (!net_eq(xt_net(par), sock_net(sk)))
+	if (sk && !net_eq(xt_net(par), sock_net(sk)))
 		sk = NULL;
 
 	if (!sk)

2.11.0

From: Pablo Neira Ayuso <[email protected]>
Date: Tue, 2 Oct 2018 00:37:39 +0200

...

Pulled, thanks.