[PATCH 0/6] netfilter/ipvs fixes for net

[PATCH 0/6] netfilter/ipvs fixes for net

[Pablo Neira Ayuso]

Hi David,

The following patchset contains fixes for netfilter/ipvs. This round of
fixes is larger than usual at this stage, specifically because of the
nf_tables bridge reject fixes that I would like to see in 3.18. The
patches are:

1) Fix a null-pointer dereference that may occur when logging
   errors. This problem was introduced by 4a4739d56b0 ("ipvs: Pull
   out crosses_local_route_boundary logic") in v3.17-rc5.

2) Update hook mask in nft_reject_bridge so we can also filter out
   packets from there. This fixes 36d2af5 ("netfilter: nf_tables: allow
   to filter from prerouting and postrouting"), which needs this chunk
   to work.

3) Two patches to refactor common code to forge the IPv4 and IPv6
   reject packets from the bridge. These are required by the nf_tables
   reject bridge fix.

4) Fix nft_reject_bridge by avoiding the use of the IP stack to reject
   packets from the bridge. The idea is to forge the reject packets and
   inject them to the original port via br_deliver() which is now
   exported for that purpose.

5) Restrict nft_reject_bridge to bridge prerouting and input hooks.
   the original skbuff may cloned after prerouting when the bridge stack
   needs to flood it to several bridge ports, it is too late to reject
   the traffic.

You can pull these changes from:

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

Thanks!

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

The following changes since commit 7965ee93719921ea5978f331da653dfa2d7b99f5:

  netfilter: nft_compat: fix wrong target lookup in nft_target_select_ops() (2014-10-27 22:17:46 +0100)

are available in the git repository at:

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

for you to fetch changes up to 127917c29a432c3b798e014a1714e9c1af0f87fe:

  netfilter: nft_reject_bridge: restrict reject to prerouting and input (2014-10-31 12:50:09 +0100)

----------------------------------------------------------------
Alex Gartrell (1):
      ipvs: Avoid null-pointer deref in debug code

Pablo Neira Ayuso (5):
      netfilter: nf_tables_bridge: update hook_mask to allow {pre,post}routing
      netfilter: nf_reject_ipv4: split nf_send_reset() in smaller functions
      netfilter: nf_reject_ipv6: split nf_send_reset6() in smaller functions
      netfilter: nft_reject_bridge: don't use IP stack to reject traffic
      netfilter: nft_reject_bridge: restrict reject to prerouting and input

 include/net/netfilter/ipv4/nf_reject.h   |   10 +
 include/net/netfilter/ipv6/nf_reject.h   |   10 +
 net/bridge/br_forward.c		  |    1 +
 net/bridge/netfilter/nf_tables_bridge.c  |    6 +-
 net/bridge/netfilter/nft_reject_bridge.c |  296 ++++++++++++++++++++++++++++--
 net/ipv4/netfilter/nf_reject_ipv4.c	  |   88 ++++++---
 net/ipv6/netfilter/nf_reject_ipv6.c	  |  175 +++++++++++-------
 net/netfilter/ipvs/ip_vs_xmit.c	  |    4 +-
 8 files changed, 483 insertions(+), 107 deletions(-)

[PATCH 1/6] ipvs: Avoid null-pointer deref in debug code

[Pablo Neira Ayuso]

From: Alex Gartrell <agartrell@fb.com>

Use daddr instead of reaching into dest.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alex Gartrell <agartrell@fb.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
---
 net/netfilter/ipvs/ip_vs_xmit.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/net/netfilter/ipvs/ip_vs_xmit.c b/net/netfilter/ipvs/ip_vs_xmit.c
index 91f17c1..437a366 100644
--- a/net/netfilter/ipvs/ip_vs_xmit.c
+++ b/net/netfilter/ipvs/ip_vs_xmit.c
@@ -316,7 +316,7 @@ __ip_vs_get_out_rt(int skb_af, struct sk_buff *skb, struct ip_vs_dest *dest,
 	if (unlikely(crosses_local_route_boundary(skb_af, skb, rt_mode,
 						  local))) {
 		IP_VS_DBG_RL("We are crossing local and non-local addresses"
-			     " daddr=%pI4\n", &dest->addr.ip);
+			     " daddr=%pI4\n", &daddr);
 		goto err_put;
 	}
 
@@ -458,7 +458,7 @@ __ip_vs_get_out_rt_v6(int skb_af, struct sk_buff *skb, struct ip_vs_dest *dest,
 	if (unlikely(crosses_local_route_boundary(skb_af, skb, rt_mode,
 						  local))) {
 		IP_VS_DBG_RL("We are crossing local and non-local addresses"
-			     " daddr=%pI6\n", &dest->addr.in6);
+			     " daddr=%pI6\n", daddr);
 		goto err_put;
 	}
 
-- 
1.7.10.4

[PATCH 2/6] netfilter: nf_tables_bridge: update hook_mask to allow {pre,post}routing

[Pablo Neira Ayuso]

Fixes: 36d2af5 ("netfilter: nf_tables: allow to filter from prerouting and postrouting")
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
 net/bridge/netfilter/nf_tables_bridge.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/net/bridge/netfilter/nf_tables_bridge.c b/net/bridge/netfilter/nf_tables_bridge.c
index da17a5e..074c557 100644
--- a/net/bridge/netfilter/nf_tables_bridge.c
+++ b/net/bridge/netfilter/nf_tables_bridge.c
@@ -75,9 +75,11 @@ static const struct nf_chain_type filter_bridge = {
 	.type		= NFT_CHAIN_T_DEFAULT,
 	.family		= NFPROTO_BRIDGE,
 	.owner		= THIS_MODULE,
-	.hook_mask	= (1 << NF_BR_LOCAL_IN) |
+	.hook_mask	= (1 << NF_BR_PRE_ROUTING) |
+			  (1 << NF_BR_LOCAL_IN) |
 			  (1 << NF_BR_FORWARD) |
-			  (1 << NF_BR_LOCAL_OUT),
+			  (1 << NF_BR_LOCAL_OUT) |
+			  (1 << NF_BR_POST_ROUTING),
 };
 
 static int __init nf_tables_bridge_init(void)
-- 
1.7.10.4

[PATCH 3/6] netfilter: nf_reject_ipv4: split nf_send_reset() in smaller functions

[Pablo Neira Ayuso]

That can be reused by the reject bridge expression to build the reject
packet. The new functions are:

* nf_reject_ip_tcphdr_get(): to sanitize and to obtain the TCP header.
* nf_reject_iphdr_put(): to build the IPv4 header.
* nf_reject_ip_tcphdr_put(): to build the TCP header.

Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
 include/net/netfilter/ipv4/nf_reject.h |   10 ++++
 net/ipv4/netfilter/nf_reject_ipv4.c    |   88 ++++++++++++++++++++++----------
 2 files changed, 72 insertions(+), 26 deletions(-)

diff --git a/include/net/netfilter/ipv4/nf_reject.h b/include/net/netfilter/ipv4/nf_reject.h
index e842719..03e928a 100644
--- a/include/net/netfilter/ipv4/nf_reject.h
+++ b/include/net/netfilter/ipv4/nf_reject.h
@@ -1,6 +1,8 @@
 #ifndef _IPV4_NF_REJECT_H
 #define _IPV4_NF_REJECT_H
 
+#include <linux/skbuff.h>
+#include <net/ip.h>
 #include <net/icmp.h>
 
 static inline void nf_send_unreach(struct sk_buff *skb_in, int code)
@@ -10,4 +12,12 @@ static inline void nf_send_unreach(struct sk_buff *skb_in, int code)
 
 void nf_send_reset(struct sk_buff *oldskb, int hook);
 
+const struct tcphdr *nf_reject_ip_tcphdr_get(struct sk_buff *oldskb,
+					     struct tcphdr *_oth, int hook);
+struct iphdr *nf_reject_iphdr_put(struct sk_buff *nskb,
+				  const struct sk_buff *oldskb,
+				  __be16 protocol, int ttl);
+void nf_reject_ip_tcphdr_put(struct sk_buff *nskb, const struct sk_buff *oldskb,
+			     const struct tcphdr *oth);
+
 #endif /* _IPV4_NF_REJECT_H */
diff --git a/net/ipv4/netfilter/nf_reject_ipv4.c b/net/ipv4/netfilter/nf_reject_ipv4.c
index 92b303d..1baaa83 100644
--- a/net/ipv4/netfilter/nf_reject_ipv4.c
+++ b/net/ipv4/netfilter/nf_reject_ipv4.c
@@ -12,43 +12,39 @@
 #include <net/route.h>
 #include <net/dst.h>
 #include <linux/netfilter_ipv4.h>
+#include <net/netfilter/ipv4/nf_reject.h>
 
-/* Send RST reply */
-void nf_send_reset(struct sk_buff *oldskb, int hook)
+const struct tcphdr *nf_reject_ip_tcphdr_get(struct sk_buff *oldskb,
+					     struct tcphdr *_oth, int hook)
 {
-	struct sk_buff *nskb;
-	const struct iphdr *oiph;
-	struct iphdr *niph;
 	const struct tcphdr *oth;
-	struct tcphdr _otcph, *tcph;
 
 	/* IP header checks: fragment. */
 	if (ip_hdr(oldskb)->frag_off & htons(IP_OFFSET))
-		return;
+		return NULL;
 
 	oth = skb_header_pointer(oldskb, ip_hdrlen(oldskb),
-				 sizeof(_otcph), &_otcph);
+				 sizeof(struct tcphdr), _oth);
 	if (oth == NULL)
-		return;
+		return NULL;
 
 	/* No RST for RST. */
 	if (oth->rst)
-		return;
-
-	if (skb_rtable(oldskb)->rt_flags & (RTCF_BROADCAST | RTCF_MULTICAST))
-		return;
+		return NULL;
 
 	/* Check checksum */
 	if (nf_ip_checksum(oldskb, hook, ip_hdrlen(oldskb), IPPROTO_TCP))
-		return;
-	oiph = ip_hdr(oldskb);
+		return NULL;
 
-	nskb = alloc_skb(sizeof(struct iphdr) + sizeof(struct tcphdr) +
-			 LL_MAX_HEADER, GFP_ATOMIC);
-	if (!nskb)
-		return;
+	return oth;
+}
+EXPORT_SYMBOL_GPL(nf_reject_ip_tcphdr_get);
 
-	skb_reserve(nskb, LL_MAX_HEADER);
+struct iphdr *nf_reject_iphdr_put(struct sk_buff *nskb,
+				  const struct sk_buff *oldskb,
+				  __be16 protocol, int ttl)
+{
+	struct iphdr *niph, *oiph = ip_hdr(oldskb);
 
 	skb_reset_network_header(nskb);
 	niph = (struct iphdr *)skb_put(nskb, sizeof(struct iphdr));
@@ -57,10 +53,23 @@ void nf_send_reset(struct sk_buff *oldskb, int hook)
 	niph->tos	= 0;
 	niph->id	= 0;
 	niph->frag_off	= htons(IP_DF);
-	niph->protocol	= IPPROTO_TCP;
+	niph->protocol	= protocol;
 	niph->check	= 0;
 	niph->saddr	= oiph->daddr;
 	niph->daddr	= oiph->saddr;
+	niph->ttl	= ttl;
+
+	nskb->protocol = htons(ETH_P_IP);
+
+	return niph;
+}
+EXPORT_SYMBOL_GPL(nf_reject_iphdr_put);
+
+void nf_reject_ip_tcphdr_put(struct sk_buff *nskb, const struct sk_buff *oldskb,
+			  const struct tcphdr *oth)
+{
+	struct iphdr *niph = ip_hdr(nskb);
+	struct tcphdr *tcph;
 
 	skb_reset_transport_header(nskb);
 	tcph = (struct tcphdr *)skb_put(nskb, sizeof(struct tcphdr));
@@ -69,9 +78,9 @@ void nf_send_reset(struct sk_buff *oldskb, int hook)
 	tcph->dest	= oth->source;
 	tcph->doff	= sizeof(struct tcphdr) / 4;
 
-	if (oth->ack)
+	if (oth->ack) {
 		tcph->seq = oth->ack_seq;
-	else {
+	} else {
 		tcph->ack_seq = htonl(ntohl(oth->seq) + oth->syn + oth->fin +
 				      oldskb->len - ip_hdrlen(oldskb) -
 				      (oth->doff << 2));
@@ -84,16 +93,43 @@ void nf_send_reset(struct sk_buff *oldskb, int hook)
 	nskb->ip_summed = CHECKSUM_PARTIAL;
 	nskb->csum_start = (unsigned char *)tcph - nskb->head;
 	nskb->csum_offset = offsetof(struct tcphdr, check);
+}
+EXPORT_SYMBOL_GPL(nf_reject_ip_tcphdr_put);
+
+/* Send RST reply */
+void nf_send_reset(struct sk_buff *oldskb, int hook)
+{
+	struct sk_buff *nskb;
+	const struct iphdr *oiph;
+	struct iphdr *niph;
+	const struct tcphdr *oth;
+	struct tcphdr _oth;
+
+	oth = nf_reject_ip_tcphdr_get(oldskb, &_oth, hook);
+	if (!oth)
+		return;
+
+	if (skb_rtable(oldskb)->rt_flags & (RTCF_BROADCAST | RTCF_MULTICAST))
+		return;
+
+	oiph = ip_hdr(oldskb);
+
+	nskb = alloc_skb(sizeof(struct iphdr) + sizeof(struct tcphdr) +
+			 LL_MAX_HEADER, GFP_ATOMIC);
+	if (!nskb)
+		return;
 
 	/* ip_route_me_harder expects skb->dst to be set */
 	skb_dst_set_noref(nskb, skb_dst(oldskb));
 
-	nskb->protocol = htons(ETH_P_IP);
+	skb_reserve(nskb, LL_MAX_HEADER);
+	niph = 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(nskb, RTN_UNSPEC))
 		goto free_nskb;
 
-	niph->ttl	= ip4_dst_hoplimit(skb_dst(nskb));
-
 	/* "Never happens" */
 	if (nskb->len > dst_mtu(skb_dst(nskb)))
 		goto free_nskb;
-- 
1.7.10.4

[PATCH 4/6] netfilter: nf_reject_ipv6: split nf_send_reset6() in smaller functions

[Pablo Neira Ayuso]

That can be reused by the reject bridge expression to build the reject
packet. The new functions are:

* nf_reject_ip6_tcphdr_get(): to sanitize and to obtain the TCP header.
* nf_reject_ip6hdr_put(): to build the IPv6 header.
* nf_reject_ip6_tcphdr_put(): to build the TCP header.

Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
 include/net/netfilter/ipv6/nf_reject.h |   10 ++
 net/ipv6/netfilter/nf_reject_ipv6.c    |  175 ++++++++++++++++++++------------
 2 files changed, 119 insertions(+), 66 deletions(-)

diff --git a/include/net/netfilter/ipv6/nf_reject.h b/include/net/netfilter/ipv6/nf_reject.h
index 48e1881..23216d4 100644
--- a/include/net/netfilter/ipv6/nf_reject.h
+++ b/include/net/netfilter/ipv6/nf_reject.h
@@ -15,4 +15,14 @@ nf_send_unreach6(struct net *net, struct sk_buff *skb_in, unsigned char code,
 
 void nf_send_reset6(struct net *net, struct sk_buff *oldskb, int hook);
 
+const struct tcphdr *nf_reject_ip6_tcphdr_get(struct sk_buff *oldskb,
+					      struct tcphdr *otcph,
+					      unsigned int *otcplen, int hook);
+struct ipv6hdr *nf_reject_ip6hdr_put(struct sk_buff *nskb,
+				     const struct sk_buff *oldskb,
+				     __be16 protocol, int hoplimit);
+void nf_reject_ip6_tcphdr_put(struct sk_buff *nskb,
+			      const struct sk_buff *oldskb,
+			      const struct tcphdr *oth, unsigned int otcplen);
+
 #endif /* _IPV6_NF_REJECT_H */
diff --git a/net/ipv6/netfilter/nf_reject_ipv6.c b/net/ipv6/netfilter/nf_reject_ipv6.c
index 20d9def..015eb8a 100644
--- a/net/ipv6/netfilter/nf_reject_ipv6.c
+++ b/net/ipv6/netfilter/nf_reject_ipv6.c
@@ -12,116 +12,102 @@
 #include <net/ip6_fib.h>
 #include <net/ip6_checksum.h>
 #include <linux/netfilter_ipv6.h>
+#include <net/netfilter/ipv6/nf_reject.h>
 
-void nf_send_reset6(struct net *net, struct sk_buff *oldskb, int hook)
+const struct tcphdr *nf_reject_ip6_tcphdr_get(struct sk_buff *oldskb,
+					      struct tcphdr *otcph,
+					      unsigned int *otcplen, int hook)
 {
-	struct sk_buff *nskb;
-	struct tcphdr otcph, *tcph;
-	unsigned int otcplen, hh_len;
-	int tcphoff, needs_ack;
 	const struct ipv6hdr *oip6h = ipv6_hdr(oldskb);
-	struct ipv6hdr *ip6h;
-#define DEFAULT_TOS_VALUE	0x0U
-	const __u8 tclass = DEFAULT_TOS_VALUE;
-	struct dst_entry *dst = NULL;
 	u8 proto;
 	__be16 frag_off;
-	struct flowi6 fl6;
-
-	if ((!(ipv6_addr_type(&oip6h->saddr) & IPV6_ADDR_UNICAST)) ||
-	    (!(ipv6_addr_type(&oip6h->daddr) & IPV6_ADDR_UNICAST))) {
-		pr_debug("addr is not unicast.\n");
-		return;
-	}
+	int tcphoff;
 
 	proto = oip6h->nexthdr;
-	tcphoff = ipv6_skip_exthdr(oldskb, ((u8*)(oip6h+1) - oldskb->data), &proto, &frag_off);
+	tcphoff = ipv6_skip_exthdr(oldskb, ((u8*)(oip6h+1) - oldskb->data),
+				   &proto, &frag_off);
 
 	if ((tcphoff < 0) || (tcphoff > oldskb->len)) {
 		pr_debug("Cannot get TCP header.\n");
-		return;
+		return NULL;
 	}
 
-	otcplen = oldskb->len - tcphoff;
+	*otcplen = oldskb->len - tcphoff;
 
 	/* IP header checks: fragment, too short. */
-	if (proto != IPPROTO_TCP || otcplen < sizeof(struct tcphdr)) {
-		pr_debug("proto(%d) != IPPROTO_TCP, "
-			 "or too short. otcplen = %d\n",
-			 proto, otcplen);
-		return;
+	if (proto != IPPROTO_TCP || *otcplen < sizeof(struct tcphdr)) {
+		pr_debug("proto(%d) != IPPROTO_TCP or too short (len = %d)\n",
+			 proto, *otcplen);
+		return NULL;
 	}
 
-	if (skb_copy_bits(oldskb, tcphoff, &otcph, sizeof(struct tcphdr)))
-		BUG();
+	otcph = skb_header_pointer(oldskb, tcphoff, sizeof(struct tcphdr),
+				   otcph);
+	if (otcph == NULL)
+		return NULL;
 
 	/* No RST for RST. */
-	if (otcph.rst) {
+	if (otcph->rst) {
 		pr_debug("RST is set\n");
-		return;
+		return NULL;
 	}
 
 	/* Check checksum. */
 	if (nf_ip6_checksum(oldskb, hook, tcphoff, IPPROTO_TCP)) {
 		pr_debug("TCP checksum is invalid\n");
-		return;
-	}
-
-	memset(&fl6, 0, sizeof(fl6));
-	fl6.flowi6_proto = IPPROTO_TCP;
-	fl6.saddr = oip6h->daddr;
-	fl6.daddr = oip6h->saddr;
-	fl6.fl6_sport = otcph.dest;
-	fl6.fl6_dport = otcph.source;
-	security_skb_classify_flow(oldskb, flowi6_to_flowi(&fl6));
-	dst = ip6_route_output(net, NULL, &fl6);
-	if (dst == NULL || dst->error) {
-		dst_release(dst);
-		return;
-	}
-	dst = xfrm_lookup(net, dst, flowi6_to_flowi(&fl6), NULL, 0);
-	if (IS_ERR(dst))
-		return;
-
-	hh_len = (dst->dev->hard_header_len + 15)&~15;
-	nskb = alloc_skb(hh_len + 15 + dst->header_len + sizeof(struct ipv6hdr)
-			 + sizeof(struct tcphdr) + dst->trailer_len,
-			 GFP_ATOMIC);
-
-	if (!nskb) {
-		net_dbg_ratelimited("cannot alloc skb\n");
-		dst_release(dst);
-		return;
+		return NULL;
 	}
 
-	skb_dst_set(nskb, dst);
+	return otcph;
+}
+EXPORT_SYMBOL_GPL(nf_reject_ip6_tcphdr_get);
 
-	skb_reserve(nskb, hh_len + dst->header_len);
+struct ipv6hdr *nf_reject_ip6hdr_put(struct sk_buff *nskb,
+				     const struct sk_buff *oldskb,
+				     __be16 protocol, int hoplimit)
+{
+	struct ipv6hdr *ip6h;
+	const struct ipv6hdr *oip6h = ipv6_hdr(oldskb);
+#define DEFAULT_TOS_VALUE	0x0U
+	const __u8 tclass = DEFAULT_TOS_VALUE;
 
 	skb_put(nskb, sizeof(struct ipv6hdr));
 	skb_reset_network_header(nskb);
 	ip6h = ipv6_hdr(nskb);
 	ip6_flow_hdr(ip6h, tclass, 0);
-	ip6h->hop_limit = ip6_dst_hoplimit(dst);
-	ip6h->nexthdr = IPPROTO_TCP;
+	ip6h->hop_limit = hoplimit;
+	ip6h->nexthdr = protocol;
 	ip6h->saddr = oip6h->daddr;
 	ip6h->daddr = oip6h->saddr;
 
+	nskb->protocol = htons(ETH_P_IPV6);
+
+	return ip6h;
+}
+EXPORT_SYMBOL_GPL(nf_reject_ip6hdr_put);
+
+void nf_reject_ip6_tcphdr_put(struct sk_buff *nskb,
+			      const struct sk_buff *oldskb,
+			      const struct tcphdr *oth, unsigned int otcplen)
+{
+	struct tcphdr *tcph;
+	int needs_ack;
+
 	skb_reset_transport_header(nskb);
 	tcph = (struct tcphdr *)skb_put(nskb, sizeof(struct tcphdr));
 	/* Truncate to length (no data) */
 	tcph->doff = sizeof(struct tcphdr)/4;
-	tcph->source = otcph.dest;
-	tcph->dest = otcph.source;
+	tcph->source = oth->dest;
+	tcph->dest = oth->source;
 
-	if (otcph.ack) {
+	if (oth->ack) {
 		needs_ack = 0;
-		tcph->seq = otcph.ack_seq;
+		tcph->seq = oth->ack_seq;
 		tcph->ack_seq = 0;
 	} else {
 		needs_ack = 1;
-		tcph->ack_seq = htonl(ntohl(otcph.seq) + otcph.syn + otcph.fin
-				      + otcplen - (otcph.doff<<2));
+		tcph->ack_seq = htonl(ntohl(oth->seq) + oth->syn + oth->fin +
+				      otcplen - (oth->doff<<2));
 		tcph->seq = 0;
 	}
 
@@ -139,6 +125,63 @@ void nf_send_reset6(struct net *net, struct sk_buff *oldskb, int hook)
 				      sizeof(struct tcphdr), IPPROTO_TCP,
 				      csum_partial(tcph,
 						   sizeof(struct tcphdr), 0));
+}
+EXPORT_SYMBOL_GPL(nf_reject_ip6_tcphdr_put);
+
+void nf_send_reset6(struct net *net, struct sk_buff *oldskb, int hook)
+{
+	struct sk_buff *nskb;
+	struct tcphdr _otcph;
+	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;
+
+	if ((!(ipv6_addr_type(&oip6h->saddr) & IPV6_ADDR_UNICAST)) ||
+	    (!(ipv6_addr_type(&oip6h->daddr) & IPV6_ADDR_UNICAST))) {
+		pr_debug("addr is not unicast.\n");
+		return;
+	}
+
+	otcph = nf_reject_ip6_tcphdr_get(oldskb, &_otcph, &otcplen, hook);
+	if (!otcph)
+		return;
+
+	memset(&fl6, 0, sizeof(fl6));
+	fl6.flowi6_proto = IPPROTO_TCP;
+	fl6.saddr = oip6h->daddr;
+	fl6.daddr = oip6h->saddr;
+	fl6.fl6_sport = otcph->dest;
+	fl6.fl6_dport = otcph->source;
+	security_skb_classify_flow(oldskb, flowi6_to_flowi(&fl6));
+	dst = ip6_route_output(net, NULL, &fl6);
+	if (dst == NULL || dst->error) {
+		dst_release(dst);
+		return;
+	}
+	dst = xfrm_lookup(net, dst, flowi6_to_flowi(&fl6), NULL, 0);
+	if (IS_ERR(dst))
+		return;
+
+	hh_len = (dst->dev->hard_header_len + 15)&~15;
+	nskb = alloc_skb(hh_len + 15 + dst->header_len + sizeof(struct ipv6hdr)
+			 + sizeof(struct tcphdr) + dst->trailer_len,
+			 GFP_ATOMIC);
+
+	if (!nskb) {
+		net_dbg_ratelimited("cannot alloc skb\n");
+		dst_release(dst);
+		return;
+	}
+
+	skb_dst_set(nskb, dst);
+
+	skb_reserve(nskb, hh_len + dst->header_len);
+	ip6h = 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);
 
-- 
1.7.10.4

[PATCH 5/6] netfilter: nft_reject_bridge: don't use IP stack to reject traffic

[Pablo Neira Ayuso]

If the packet is received via the bridge stack, this cannot reject
packets from the IP stack.

This adds functions to build the reject packet and send it from the
bridge stack. Comments and assumptions on this patch:

1) Validate the IPv4 and IPv6 headers before further processing,
   given that the packet comes from the bridge stack, we cannot assume
   they are clean. Truncated packets are dropped, we follow similar
   approach in the existing iptables match/target extensions that need
   to inspect layer 4 headers that is not available. This also includes
   packets that are directed to multicast and broadcast ethernet
   addresses.

2) br_deliver() is exported to inject the reject packet via
   bridge localout -> postrouting. So the approach is similar to what
   we already do in the iptables reject target. The reject packet is
   sent to the bridge port from which we have received the original
   packet.

3) The reject packet is forged based on the original packet. The TTL
   is set based on sysctl_ip_default_ttl for IPv4 and per-net
   ipv6.devconf_all hoplimit for IPv6.

Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
 net/bridge/br_forward.c                  |    1 +
 net/bridge/netfilter/nft_reject_bridge.c |  263 ++++++++++++++++++++++++++++--
 2 files changed, 254 insertions(+), 10 deletions(-)

diff --git a/net/bridge/br_forward.c b/net/bridge/br_forward.c
index 992ec49..44cb786 100644
--- a/net/bridge/br_forward.c
+++ b/net/bridge/br_forward.c
@@ -112,6 +112,7 @@ void br_deliver(const struct net_bridge_port *to, struct sk_buff *skb)
 
 	kfree_skb(skb);
 }
+EXPORT_SYMBOL_GPL(br_deliver);
 
 /* called with rcu_read_lock */
 void br_forward(const struct net_bridge_port *to, struct sk_buff *skb, struct sk_buff *skb0)
diff --git a/net/bridge/netfilter/nft_reject_bridge.c b/net/bridge/netfilter/nft_reject_bridge.c
index a764795..31b27e1 100644
--- a/net/bridge/netfilter/nft_reject_bridge.c
+++ b/net/bridge/netfilter/nft_reject_bridge.c
@@ -16,6 +16,237 @@
 #include <net/netfilter/nft_reject.h>
 #include <net/netfilter/ipv4/nf_reject.h>
 #include <net/netfilter/ipv6/nf_reject.h>
+#include <linux/ip.h>
+#include <net/ip.h>
+#include "../br_private.h"
+
+static void nft_reject_br_push_etherhdr(struct sk_buff *oldskb,
+					struct sk_buff *nskb)
+{
+	struct ethhdr *eth;
+
+	eth = (struct ethhdr *)skb_push(nskb, ETH_HLEN);
+	skb_reset_mac_header(nskb);
+	ether_addr_copy(eth->h_source, eth_hdr(oldskb)->h_dest);
+	ether_addr_copy(eth->h_dest, eth_hdr(oldskb)->h_source);
+	eth->h_proto = eth_hdr(oldskb)->h_proto;
+	skb_pull(nskb, ETH_HLEN);
+}
+
+static int nft_reject_iphdr_validate(struct sk_buff *oldskb)
+{
+	struct iphdr *iph;
+	u32 len;
+
+	if (!pskb_may_pull(oldskb, sizeof(struct iphdr)))
+		return 0;
+
+	iph = ip_hdr(oldskb);
+	if (iph->ihl < 5 || iph->version != 4)
+		return 0;
+
+	len = ntohs(iph->tot_len);
+	if (oldskb->len < len)
+		return 0;
+	else if (len < (iph->ihl*4))
+		return 0;
+
+	if (!pskb_may_pull(oldskb, iph->ihl*4))
+		return 0;
+
+	return 1;
+}
+
+static void nft_reject_br_send_v4_tcp_reset(struct sk_buff *oldskb, int hook)
+{
+	struct sk_buff *nskb;
+	struct iphdr *niph;
+	const struct tcphdr *oth;
+	struct tcphdr _oth;
+
+	if (!nft_reject_iphdr_validate(oldskb))
+		return;
+
+	oth = nf_reject_ip_tcphdr_get(oldskb, &_oth, hook);
+	if (!oth)
+		return;
+
+	nskb = alloc_skb(sizeof(struct iphdr) + sizeof(struct tcphdr) +
+			 LL_MAX_HEADER, GFP_ATOMIC);
+	if (!nskb)
+		return;
+
+	skb_reserve(nskb, LL_MAX_HEADER);
+	niph = nf_reject_iphdr_put(nskb, oldskb, IPPROTO_TCP,
+				   sysctl_ip_default_ttl);
+	nf_reject_ip_tcphdr_put(nskb, oldskb, oth);
+	niph->ttl	= sysctl_ip_default_ttl;
+	niph->tot_len	= htons(nskb->len);
+	ip_send_check(niph);
+
+	nft_reject_br_push_etherhdr(oldskb, nskb);
+
+	br_deliver(br_port_get_rcu(oldskb->dev), nskb);
+}
+
+static void nft_reject_br_send_v4_unreach(struct sk_buff *oldskb, int hook,
+					  u8 code)
+{
+	struct sk_buff *nskb;
+	struct iphdr *niph;
+	struct icmphdr *icmph;
+	unsigned int len;
+	void *payload;
+	__wsum csum;
+
+	if (!nft_reject_iphdr_validate(oldskb))
+		return;
+
+	/* IP header checks: fragment. */
+	if (ip_hdr(oldskb)->frag_off & htons(IP_OFFSET))
+		return;
+
+	/* RFC says return as much as we can without exceeding 576 bytes. */
+	len = min_t(unsigned int, 536, oldskb->len);
+
+	if (!pskb_may_pull(oldskb, len))
+		return;
+
+	if (nf_ip_checksum(oldskb, hook, ip_hdrlen(oldskb), 0))
+		return;
+
+	nskb = alloc_skb(sizeof(struct iphdr) + sizeof(struct icmphdr) +
+			 LL_MAX_HEADER + len, GFP_ATOMIC);
+	if (!nskb)
+		return;
+
+	skb_reserve(nskb, LL_MAX_HEADER);
+	niph = nf_reject_iphdr_put(nskb, oldskb, IPPROTO_ICMP,
+				   sysctl_ip_default_ttl);
+
+	skb_reset_transport_header(nskb);
+	icmph = (struct icmphdr *)skb_put(nskb, sizeof(struct icmphdr));
+	memset(icmph, 0, sizeof(*icmph));
+	icmph->type     = ICMP_DEST_UNREACH;
+	icmph->code	= code;
+
+	payload = skb_put(nskb, len);
+	memcpy(payload, skb_network_header(oldskb), len);
+
+	csum = csum_partial((void *)icmph, len + sizeof(struct icmphdr), 0);
+	icmph->checksum = csum_fold(csum);
+
+	niph->tot_len	= htons(nskb->len);
+	ip_send_check(niph);
+
+	nft_reject_br_push_etherhdr(oldskb, nskb);
+
+	br_deliver(br_port_get_rcu(oldskb->dev), nskb);
+}
+
+static int nft_reject_ip6hdr_validate(struct sk_buff *oldskb)
+{
+	struct ipv6hdr *hdr;
+	u32 pkt_len;
+
+	if (!pskb_may_pull(oldskb, sizeof(struct ipv6hdr)))
+		return 0;
+
+	hdr = ipv6_hdr(oldskb);
+	if (hdr->version != 6)
+		return 0;
+
+	pkt_len = ntohs(hdr->payload_len);
+	if (pkt_len + sizeof(struct ipv6hdr) > oldskb->len)
+		return 0;
+
+	return 1;
+}
+
+static void nft_reject_br_send_v6_tcp_reset(struct net *net,
+					    struct sk_buff *oldskb, int hook)
+{
+	struct sk_buff *nskb;
+	const struct tcphdr *oth;
+	struct tcphdr _oth;
+	unsigned int otcplen;
+	struct ipv6hdr *nip6h;
+
+	if (!nft_reject_ip6hdr_validate(oldskb))
+		return;
+
+	oth = nf_reject_ip6_tcphdr_get(oldskb, &_oth, &otcplen, hook);
+	if (!oth)
+		return;
+
+	nskb = alloc_skb(sizeof(struct ipv6hdr) + sizeof(struct tcphdr) +
+			 LL_MAX_HEADER, GFP_ATOMIC);
+	if (!nskb)
+		return;
+
+	skb_reserve(nskb, LL_MAX_HEADER);
+	nip6h = nf_reject_ip6hdr_put(nskb, oldskb, IPPROTO_TCP,
+				     net->ipv6.devconf_all->hop_limit);
+	nf_reject_ip6_tcphdr_put(nskb, oldskb, oth, otcplen);
+	nip6h->payload_len = htons(nskb->len - sizeof(struct ipv6hdr));
+
+	nft_reject_br_push_etherhdr(oldskb, nskb);
+
+	br_deliver(br_port_get_rcu(oldskb->dev), nskb);
+}
+
+static void nft_reject_br_send_v6_unreach(struct net *net,
+					  struct sk_buff *oldskb, int hook,
+					  u8 code)
+{
+	struct sk_buff *nskb;
+	struct ipv6hdr *nip6h;
+	struct icmp6hdr *icmp6h;
+	unsigned int len;
+	void *payload;
+
+	if (!nft_reject_ip6hdr_validate(oldskb))
+		return;
+
+	/* Include "As much of invoking packet as possible without the ICMPv6
+	 * packet exceeding the minimum IPv6 MTU" in the ICMP payload.
+	 */
+	len = min_t(unsigned int, 1220, oldskb->len);
+
+	if (!pskb_may_pull(oldskb, len))
+		return;
+
+	nskb = alloc_skb(sizeof(struct iphdr) + sizeof(struct icmp6hdr) +
+			 LL_MAX_HEADER + len, GFP_ATOMIC);
+	if (!nskb)
+		return;
+
+	skb_reserve(nskb, LL_MAX_HEADER);
+	nip6h = nf_reject_ip6hdr_put(nskb, oldskb, IPPROTO_ICMPV6,
+				     net->ipv6.devconf_all->hop_limit);
+
+	skb_reset_transport_header(nskb);
+	icmp6h = (struct icmp6hdr *)skb_put(nskb, sizeof(struct icmp6hdr));
+	memset(icmp6h, 0, sizeof(*icmp6h));
+	icmp6h->icmp6_type = ICMPV6_DEST_UNREACH;
+	icmp6h->icmp6_code = code;
+
+	payload = skb_put(nskb, len);
+	memcpy(payload, skb_network_header(oldskb), len);
+	nip6h->payload_len = htons(nskb->len - sizeof(struct ipv6hdr));
+
+	icmp6h->icmp6_cksum =
+		csum_ipv6_magic(&nip6h->saddr, &nip6h->daddr,
+				nskb->len - sizeof(struct ipv6hdr),
+				IPPROTO_ICMPV6,
+				csum_partial(icmp6h,
+					     nskb->len - sizeof(struct ipv6hdr),
+					     0));
+
+	nft_reject_br_push_etherhdr(oldskb, nskb);
+
+	br_deliver(br_port_get_rcu(oldskb->dev), nskb);
+}
 
 static void nft_reject_bridge_eval(const struct nft_expr *expr,
 				 struct nft_data data[NFT_REG_MAX + 1],
@@ -23,35 +254,46 @@ static void nft_reject_bridge_eval(const struct nft_expr *expr,
 {
 	struct nft_reject *priv = nft_expr_priv(expr);
 	struct net *net = dev_net((pkt->in != NULL) ? pkt->in : pkt->out);
+	const unsigned char *dest = eth_hdr(pkt->skb)->h_dest;
+
+	if (is_broadcast_ether_addr(dest) ||
+	    is_multicast_ether_addr(dest))
+		goto out;
 
 	switch (eth_hdr(pkt->skb)->h_proto) {
 	case htons(ETH_P_IP):
 		switch (priv->type) {
 		case NFT_REJECT_ICMP_UNREACH:
-			nf_send_unreach(pkt->skb, priv->icmp_code);
+			nft_reject_br_send_v4_unreach(pkt->skb,
+						      pkt->ops->hooknum,
+						      priv->icmp_code);
 			break;
 		case NFT_REJECT_TCP_RST:
-			nf_send_reset(pkt->skb, pkt->ops->hooknum);
+			nft_reject_br_send_v4_tcp_reset(pkt->skb,
+							pkt->ops->hooknum);
 			break;
 		case NFT_REJECT_ICMPX_UNREACH:
-			nf_send_unreach(pkt->skb,
-					nft_reject_icmp_code(priv->icmp_code));
+			nft_reject_br_send_v4_unreach(pkt->skb,
+						      pkt->ops->hooknum,
+						      nft_reject_icmp_code(priv->icmp_code));
 			break;
 		}
 		break;
 	case htons(ETH_P_IPV6):
 		switch (priv->type) {
 		case NFT_REJECT_ICMP_UNREACH:
-			nf_send_unreach6(net, pkt->skb, priv->icmp_code,
-					 pkt->ops->hooknum);
+			nft_reject_br_send_v6_unreach(net, pkt->skb,
+						      pkt->ops->hooknum,
+						      priv->icmp_code);
 			break;
 		case NFT_REJECT_TCP_RST:
-			nf_send_reset6(net, pkt->skb, pkt->ops->hooknum);
+			nft_reject_br_send_v6_tcp_reset(net, pkt->skb,
+							pkt->ops->hooknum);
 			break;
 		case NFT_REJECT_ICMPX_UNREACH:
-			nf_send_unreach6(net, pkt->skb,
-					 nft_reject_icmpv6_code(priv->icmp_code),
-					 pkt->ops->hooknum);
+			nft_reject_br_send_v6_unreach(net, pkt->skb,
+						      pkt->ops->hooknum,
+						      nft_reject_icmpv6_code(priv->icmp_code));
 			break;
 		}
 		break;
@@ -59,6 +301,7 @@ static void nft_reject_bridge_eval(const struct nft_expr *expr,
 		/* No explicit way to reject this protocol, drop it. */
 		break;
 	}
+out:
 	data[NFT_REG_VERDICT].verdict = NF_DROP;
 }
 
-- 
1.7.10.4

[PATCH 6/6] netfilter: nft_reject_bridge: restrict reject to prerouting and input

[Pablo Neira Ayuso]

Restrict the reject expression to the prerouting and input bridge
hooks. If we allow this to be used from forward or any other later
bridge hook, if the frame is flooded to several ports, we'll end up
sending several reject packets, one per cloned packet.

Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
 net/bridge/netfilter/nft_reject_bridge.c |   33 +++++++++++++++++++++++++++++-
 1 file changed, 32 insertions(+), 1 deletion(-)

diff --git a/net/bridge/netfilter/nft_reject_bridge.c b/net/bridge/netfilter/nft_reject_bridge.c
index 31b27e1..654c901 100644
--- a/net/bridge/netfilter/nft_reject_bridge.c
+++ b/net/bridge/netfilter/nft_reject_bridge.c
@@ -18,6 +18,7 @@
 #include <net/netfilter/ipv6/nf_reject.h>
 #include <linux/ip.h>
 #include <net/ip.h>
+#include <linux/netfilter_bridge.h>
 #include "../br_private.h"
 
 static void nft_reject_br_push_etherhdr(struct sk_buff *oldskb,
@@ -305,12 +306,34 @@ out:
 	data[NFT_REG_VERDICT].verdict = NF_DROP;
 }
 
+static int nft_reject_bridge_validate_hooks(const struct nft_chain *chain)
+{
+	struct nft_base_chain *basechain;
+
+	if (chain->flags & NFT_BASE_CHAIN) {
+		basechain = nft_base_chain(chain);
+
+		switch (basechain->ops[0].hooknum) {
+		case NF_BR_PRE_ROUTING:
+		case NF_BR_LOCAL_IN:
+			break;
+		default:
+			return -EOPNOTSUPP;
+		}
+	}
+	return 0;
+}
+
 static int nft_reject_bridge_init(const struct nft_ctx *ctx,
 				  const struct nft_expr *expr,
 				  const struct nlattr * const tb[])
 {
 	struct nft_reject *priv = nft_expr_priv(expr);
-	int icmp_code;
+	int icmp_code, err;
+
+	err = nft_reject_bridge_validate_hooks(ctx->chain);
+	if (err < 0)
+		return err;
 
 	if (tb[NFTA_REJECT_TYPE] == NULL)
 		return -EINVAL;
@@ -359,6 +382,13 @@ nla_put_failure:
 	return -1;
 }
 
+static int nft_reject_bridge_validate(const struct nft_ctx *ctx,
+				      const struct nft_expr *expr,
+				      const struct nft_data **data)
+{
+	return nft_reject_bridge_validate_hooks(ctx->chain);
+}
+
 static struct nft_expr_type nft_reject_bridge_type;
 static const struct nft_expr_ops nft_reject_bridge_ops = {
 	.type		= &nft_reject_bridge_type,
@@ -366,6 +396,7 @@ static const struct nft_expr_ops nft_reject_bridge_ops = {
 	.eval		= nft_reject_bridge_eval,
 	.init		= nft_reject_bridge_init,
 	.dump		= nft_reject_bridge_dump,
+	.validate	= nft_reject_bridge_validate,
 };
 
 static struct nft_expr_type nft_reject_bridge_type __read_mostly = {
-- 
1.7.10.4

Re: [PATCH 0/6] netfilter/ipvs fixes for net

[David Miller]

From: Pablo Neira Ayuso <pablo@netfilter.org>
Date: Fri, 31 Oct 2014 13:18:26 +0100

> The following patchset contains fixes for netfilter/ipvs. This round of
> fixes is larger than usual at this stage, specifically because of the
> nf_tables bridge reject fixes that I would like to see in 3.18. The
> patches are:
 ...
> You can pull these changes from:
> 
>   git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf.git

Pulled, thanks Pablo.