Skip to contentSkip to conversation

net: discard ARP/NDP bcast/null announce (poison)

RFC PATCH Superseded Marc Suñé 2 participants 11 messages Dec 21, 2025 – Dec 22, 2025

Subsystems (4)

v1 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

cover net: discard ARP/NDP bcast/null announce (poison)

The current ARP and NDP implementations accepts announcements with broadcast (mcast, and null) MAC addresses as Sender HW Address (SHA) in ARP or src/target lladdr in NDP, and updates the cache for that neighbour.

Broadcast (and Multicast, see RFC1812, section 3.3.2) and null MAC addresses are reserved addresses and shall never be associated with a unicast or a multicast IPv4/6 address.

ARP/NDP poisioning with a broadcast MAC address, especially when poisoning a Gateway IP, has some undesired implications compared to an ARP/NDP poisioning with a regular MAC. See Note1.

Worth mentioning that if an attacker is able to ARP/NDP poison in a L2 segment, that in itself is probably a bigger security threat (Man-in-middle etc.). See Note2.

However, since these MACs should never be announced, this patch series discards/drops these packets, which prevents broadcast ARP/NDP poisoning vectors.

Comments:

a) This patchset only modifies the behaviour of the neighbouring subsystem when processing network packets. Static entries can still be added with bcast/null MACs.

b) According to RFC1812 multicast MAC addresses should also be rejected:

Quoted 3 lines

A router MUST not believe any ARP reply that claims that the Link Layer address of another host or router is a broadcast or multicast address.

Certain Load Balancers make use of Multicast MAC addresses:

https://support.huawei.com/enterprise/en/doc/EDOC1100213154/d8621162/dynamic-learning-of-arp-entries-with-multicast-mac-addresses
https://learn.microsoft.com/en-us/troubleshoot/windows-server/networking/configure-network-to-support-nlb-operation-mode

But as stated in the Microsoft NLB documentation, it is expected that static entries are be created for it to work:

Quoted 7 lines

To support this configuration, you must configure the network infrastructure to use static ARP entries and MAC address table entries. Network switches cannot learn the NLB multicast MAC address in the course of their usual operations. If you skip the manual configuration step, the network switches may flood NLB traffic to all ports or drop packets. The network may seem to function correctly at first, but problems increase over time.

So I think it's safe to change is_broadcast_ether_addr() to is_multicast_ether_addr() in patches 1 and 3. Alternatively, a per-interface knob to control whether MCAST MAC addresses are learnt, with the default to NOT accept them, could be an option.

c) Scapy: not clear whether acceptable in selftests (used somewhere but not consistently). It is convenient for pkt generation. Patch 5 should be either dropped or squashed (replacing C progs) in patches 2 and 4.

d) In PATCH 1/5 (ARP), I _think_ it's safe to assume that all dev_types with HW addrlen 6 are MAC addresses, but it would need to be double checked by reviewers.

(Notes extracted from the ARP commit. NDP similar. MCAST MACs exhibit a similar behaviour - depends on the exact MCAST MAC.)

Note1:

After a successful broadcast ARP poisioning attack:

1. Unicast packets and refresh ("targeted") ARPs sent to or via the poisioned IP (e.g. the default GW) are flooded by bridges/switches. That is in absence of other security controls.

Hardware swiches generally have rate-limits to prevent/mitigate broadcast storms, since ARPs are usually answered by the CPU. Legit unicast packets could be dropped (perf. degradation).

Most modern NICs implement some form of L2 MAC filtering to early discard irrelevant packets. In contrast to an ARP poisoning attack with any other MAC, both unicast and ARP ("targeted") refresh packets are passed up to the Kernel networking stack (for all hosts in the L2 segment).

2. A single forged ARP packet (e.g. for the Gateway IP) can produce up to N "targeted" (to broadcast) ARPs, where N is the number of hosts in the L2 segment that have an ARP entry for that IP (e.g. GW), and some more traffic, since the real host will answer to targeted refresh ARPs with their (real) reply.

This is a relatively low amount of traffic compared to 1).

3. An attacker could use this form of ARP poisoning to discover all hosts in a L2 segment in a very short period of time with one or few packets.

By poisoning e.g. the default GW (likely multiple times, to avoid races with real gARPs from the GW), all hosts will eventually issue refresh "targeted" ARPs for the GW IP with the broadcast MAC address as destination. These packets will be flooded in the L2 segment, revealing the presence of hosts to the attacker.

For comparison:
* Passive ARP monitoring: also stealthy, but can take a long time or not be possible at all in switches, as most refresh ARPs are targeted.
* ARP req flooding: requires swiping the entire subnet. Noisy and easy to detect.
* ICMP/L4 port scans: similar to the above.

4. In the unlikely case that hosts were to run with /proc/sys/net/ipv4/conf/*/arp_accept=1 (unsafe, and disabled by default), poisoning with the broadcast MAC could be used to create significantly more broadcast traffic (low-volume amplification attack).

An attacker could send M fake gARP with a number of IP addresses, where M is /proc/sys/net/ipv4/neigh/*/gc_thresh3 (1024 by default). This would result in M x R ARPs, where R is the number of hosts in L2 segment with arp_accept=1, and likely other (real) ARP replies coming from the attacked host. This starts to get really relevant when R > 512, which is possible in large LANs but not very common.

Note2:

However, broadcast ARP poisoning might be subtle and difficult to spot. These ARP packets appear on the surface as regular broadcast ARP requests (unless ARP hdr is inspected), traffic continues to flow uninterrupted (unless broadcast rate-limit in switches kick-in) and, the next refresh ARP reply (from the GW) or any (valid) gARP from the GW, will restore the original MAC in the ARP table, making the traffic flow normally again.

Marc Suñé (5):
arp: discard sha bcast/null (bcast ARP poison) selftests/net: add no ARP bcast/null poison test
neigh: discard lladdr bcast/null (bcast poison) selftests/net: add no NDP bcast/null poison test selftests/net: use scapy for no_bcastnull_poison

 net/ipv4/arp.c                                |   9 +
 net/ipv6/ndisc.c                              |  22 ++
 tools/testing/selftests/net/.gitignore        |   2 +
 tools/testing/selftests/net/Makefile          |   1 +
 .../net/arp_ndisc_no_bcastnull_poison.sh      | 334 ++++++++++++++++++
 tools/testing/selftests/net/arp_send.py       |  24 ++
 tools/testing/selftests/net/ndisc_send.py     |  36 ++
 7 files changed, 428 insertions(+)
 create mode 100755 tools/testing/selftests/net/arp_ndisc_no_bcastnull_poison.sh
 create mode 100644 tools/testing/selftests/net/arp_send.py
 create mode 100644 tools/testing/selftests/net/ndisc_send.py
Signature · 1 line
2.47.3
#1 lore raw

RFC PATCH 1/5 arp: discard sha bcast/null (bcast ARP poison)

The current ARP implementation accepts ARP req/replies with the broadcast (mcast, and null) MAC addresses as Sender HW Address (SHA), and updates the ARP cache for that neighbour.

Broadcast (and Multicast, see RFC1812, section 3.3.2) and null MAC addresses are reserved addresses and shall never be associated with a unicast or a multicast IPv4 address.

ARP poisioning with a broadcast MAC address, especially when poisoning a Gateway IP, has some undesired implications compared to an ARP poisioning with a regular MAC. See Note1.

Worth mentioning that if an attacker is able to ARP poison in a L2 segment, that in itself is probably a bigger security threat (Man-in-middle etc.). See Note2.

However, since these MACs should never be announced as SHA, discard/drop ARPs with SHA={bcast, null}, which prevents the broadcast ARP poisoning vector.

Note1:

After a successful broadcast ARP poisioning attack:

1. Unicast packets and refresh ("targeted") ARPs sent to or via the poisioned IP (e.g. the default GW) are flooded by bridges/switches. That is in absence of other security controls.

Hardware swiches generally have rate-limits to prevent/mitigate broadcast storms, since ARPs are usually answered by the CPU. Legit unicast packets could be dropped (perf. degradation).

Most modern NICs implement some form of L2 MAC filtering to early discard irrelevant packets. In contrast to an ARP poisoning attack with any other MAC, both unicast and ARP ("targeted") refresh packets are passed up to the Kernel networking stack (for all hosts in the L2 segment).

2. A single forged ARP packet (e.g. for the Gateway IP) can produce up to N "targeted" (to broadcast) ARPs, where N is the number of hosts in the L2 segment that have an ARP entry for that IP (e.g. GW), and some more traffic, since the real host will answer to targeted refresh ARPs with their (real) reply.

This is a relatively low amount of traffic compared to 1).

3. An attacker could use this form of ARP poisoning to discover all hosts in a L2 segment in a very short period of time with one or few packets.

By poisoning e.g. the default GW (likely multiple times, to avoid races with real gARPs from the GW), all hosts will eventually issue refresh "targeted" ARPs for the GW IP with the broadcast MAC address as destination. These packets will be flooded in the L2 segment, revealing the presence of hosts to the attacker.

For comparison:
* Passive ARP monitoring: also stealthy, but can take a long time or not be possible at all in switches, as most refresh ARPs are targeted.
* ARP req flooding: requires swiping the entire subnet. Noisy and easy to detect.
* ICMP/L4 port scans: similar to the above.

4. In the unlikely case that hosts were to run with /proc/sys/net/ipv4/conf/*/arp_accept=1 (unsafe, and disabled by default), poisoning with the broadcast MAC could be used to create significantly more broadcast traffic (low-volume amplification attack).

An attacker could send M fake gARP with a number of IP addresses, where M is /proc/sys/net/ipv4/neigh/*/gc_thresh3 (1024 by default). This would result in M x R ARPs, where R is the number of hosts in L2 segment with arp_accept=1, and likely other (real) ARP replies coming from the attacked host. This starts to get really relevant when R > 512, which is possible in large LANs but not very common.

Note2:

However, broadcast ARP poisoning might be subtle and difficult to spot. These ARP packets appear on the surface as regular broadcast ARP requests (unless ARP hdr is inspected), traffic continues to flow uninterrupted (unless broadcast rate-limit in switches kick-in) and, the next refresh ARP reply (from the GW) or any (valid) gARP from the GW, will restore the original MAC in the ARP table, making the traffic flow normally again.

---
 net/ipv4/arp.c | 9 +++++++++
 1 file changed, 9 insertions(+)
net/ipv4/arp.c +9 -0
index 7f3863daaa40..83b34b89b1be 100644
@@ -799,6 +799,15 @@ static int arp_process(struct net *net, struct sock *sk, struct sk_buff *skb)
 		goto out_free_skb;
 
 /*
+ *	For Ethernet devices, Broadcast/Multicast and zero MAC addresses should
+ *	never be announced and accepted as sender HW address (prevent BCAST MAC
+ *	and NULL ARP poisoning attack).
+ */
+	if (dev->addr_len == ETH_ALEN &&
+	    (is_broadcast_ether_addr(sha) || is_zero_ether_addr(sha)))
+		goto out_free_skb;
+
+ /*
  *     Special case: We must set Frame Relay source Q.922 address
  */
 	if (dev_type == ARPHRD_DLCI)
Signature · 1 line
2.47.3
#5 lore raw

RFC PATCH 2/5 selftests/net: add no ARP bcast/null poison test

In reply to Marc Suñé

Add a selftest to test that ARP bcast/null poisioning checks are never bypassed.

---
 tools/testing/selftests/net/.gitignore        |   1 +
 tools/testing/selftests/net/Makefile          |   2 +
 .../selftests/net/arp_no_bcastnull_poision.sh | 159 ++++++++++++++++++
 tools/testing/selftests/net/arp_send.c        | 138 +++++++++++++++
 4 files changed, 300 insertions(+)
 create mode 100755 tools/testing/selftests/net/arp_no_bcastnull_poision.sh
 create mode 100644 tools/testing/selftests/net/arp_send.c
tools/testing/selftests/net/.gitignore +1 -0
index 6930fe926c58..fd08ceeab07c 100644
@@ -1,4 +1,5 @@
 # SPDX-License-Identifier: GPL-2.0-only
+arp_send
 bind_bhash
 bind_timewait
 bind_wildcard
tools/testing/selftests/net/Makefile +2 -0
index b66ba04f19d9..8308f0067547 100644
@@ -9,6 +9,7 @@ CFLAGS += -I../
 TEST_PROGS := \
 	altnames.sh \
 	amt.sh \
+	arp_no_bcastnull_poision.sh \
 	arp_ndisc_evict_nocarrier.sh \
 	arp_ndisc_untracked_subnets.sh \
 	bareudp.sh \
@@ -163,6 +164,7 @@ TEST_GEN_FILES := \
 # end of TEST_GEN_FILES
 
 TEST_GEN_PROGS := \
+	arp_send \
 	bind_timewait \
 	bind_wildcard \
 	epoll_busy_poll \
tools/testing/selftests/net/arp_no_bcastnull_poision.sh new file +159 -0

159 lines not on this page · raw message

tools/testing/selftests/net/arp_send.c new file +138 -0

138 lines not on this page · raw message

Signature · 1 line
2.47.3
#6 lore raw

RFC PATCH 3/5 neigh: discard lladdr bcast/null (bcast poison)

In reply to Marc Suñé

The current NDP implementation accepts NDP NS/NS with the broadcast (mcast, and null) MAC addresses as src/dst lladdr, and updates the neighbour cache for that host.

Broadcast (and Multicast, see RFC1812, section 3.3.2) and null MAC addresses are reserved addresses and shall never be associated with a unicast or a multicast IPv6 address.

NDP poisioning with a broadcast MAC address, especially when poisoning a Gateway IP, has some undesired implications compared to an NDP poisioning with a regular MAC. (see ARP bcast poison commit for more details).

Worth mentioning that if an attacker is able to NDP poison in a L2 segment, that in itself is probably a bigger security threat (Man-in-middle etc.).

However, since these MACs should never be announced as SHA, discard/drop NDP with lladdr={bcast, null}, which prevents the broadcast NDP poisoning vector.

---
 net/ipv6/ndisc.c | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)
net/ipv6/ndisc.c +22 -0
index 59d17b6f06bf..980768a79092 100644
@@ -830,6 +830,17 @@ static enum skb_drop_reason ndisc_recv_ns(struct sk_buff *skb)
 			return reason;
 		}
 
+		/*
+		 * Broadcast/Multicast and zero MAC addresses should
+		 * never be announced and accepted as llsrc address (prevent
+		 * NDP BCAST MAC poisoning attack).
+		 */
+		if (dev->addr_len == ETH_ALEN &&
+		    (is_broadcast_ether_addr(lladdr) ||
+		     is_zero_ether_addr(lladdr))) {
+			return reason;
+		}
+
 		/* RFC2461 7.1.1:
 		 *	If the IP source address is the unspecified address,
 		 *	there MUST NOT be source link-layer address option
@@ -1033,6 +1044,17 @@ static enum skb_drop_reason ndisc_recv_na(struct sk_buff *skb)
 			net_dbg_ratelimited("NA: invalid link-layer address length\n");
 			return reason;
 		}
+
+		/*
+		 * Broadcast/Multicast and zero MAC addresses should
+		 * never be announced and accepted as llsrc address (prevent
+		 * NDP BCAST MAC poisoning attack).
+		 */
+		if (dev->addr_len == ETH_ALEN &&
+		    (is_broadcast_ether_addr(lladdr) ||
+		     is_zero_ether_addr(lladdr))) {
+			return reason;
+		}
 	}
 	ifp = ipv6_get_ifaddr(dev_net(dev), &msg->target, dev, 1);
 	if (ifp) {
Signature · 1 line
2.47.3
#7 lore raw

RFC PATCH 4/5 selftests/net: add no NDP bcast/null poison test

In reply to Marc Suñé

Add a selftest to test that NDP bcast/null poisioning checks are never bypassed.

---
 tools/testing/selftests/net/.gitignore        |   1 +
 tools/testing/selftests/net/Makefile          |   1 +
 .../net/arp_ndisc_no_bcastnull_poison.sh      | 334 ++++++++++++++++++
 .../selftests/net/arp_no_bcastnull_poision.sh | 159 ---------
 tools/testing/selftests/net/ndisc_send.c      | 198 +++++++++++
 5 files changed, 534 insertions(+), 159 deletions(-)
 create mode 100755 tools/testing/selftests/net/arp_ndisc_no_bcastnull_poison.sh
 delete mode 100755 tools/testing/selftests/net/arp_no_bcastnull_poision.sh
 create mode 100644 tools/testing/selftests/net/ndisc_send.c
tools/testing/selftests/net/.gitignore +1 -0
index fd08ceeab07c..5a82300a22a9 100644
@@ -18,6 +18,7 @@ ipv6_flowlabel_mgr
 ipv6_fragmentation
 log.txt
 msg_zerocopy
+ndisc_send
 netlink-dumps
 nettest
 proc_net_pktgen
tools/testing/selftests/net/Makefile +1 -0
index 8308f0067547..a5bd845e1c58 100644
@@ -169,6 +169,7 @@ TEST_GEN_PROGS := \
 	bind_wildcard \
 	epoll_busy_poll \
 	ipv6_fragmentation \
+	ndisc_send \
 	proc_net_pktgen \
 	reuseaddr_conflict \
 	reuseport_bpf \
tools/testing/selftests/net/arp_ndisc_no_bcastnull_poison.sh new file +334 -0

334 lines not on this page · raw message

tools/testing/selftests/net/arp_no_bcastnull_poision.sh deleted +0 -159

159 lines not on this page · raw message

tools/testing/selftests/net/ndisc_send.c new file +198 -0

198 lines not on this page · raw message

Signature · 1 line
2.47.3
#8 lore raw

RFC PATCH 5/5 selftests/net: use scapy for no_bcastnull_poison

In reply to Marc Suñé

Use Scapy to generate ARP/ND packets for ARP/ND no bcast/NULL MAC poisoning.

---
 tools/testing/selftests/net/Makefile          |   2 -
 .../net/arp_ndisc_no_bcastnull_poison.sh      |  12 +-
 tools/testing/selftests/net/arp_send.c        | 138 ------------
 tools/testing/selftests/net/arp_send.py       |  24 +++
 tools/testing/selftests/net/ndisc_send.c      | 198 ------------------
 tools/testing/selftests/net/ndisc_send.py     |  36 ++++
 6 files changed, 66 insertions(+), 344 deletions(-)
 delete mode 100644 tools/testing/selftests/net/arp_send.c
 create mode 100644 tools/testing/selftests/net/arp_send.py
 delete mode 100644 tools/testing/selftests/net/ndisc_send.c
 create mode 100644 tools/testing/selftests/net/ndisc_send.py
tools/testing/selftests/net/Makefile +0 -2
index a5bd845e1c58..6fe0b962cd05 100644
@@ -164,12 +164,10 @@ TEST_GEN_FILES := \
 # end of TEST_GEN_FILES
 
 TEST_GEN_PROGS := \
-	arp_send \
 	bind_timewait \
 	bind_wildcard \
 	epoll_busy_poll \
 	ipv6_fragmentation \
-	ndisc_send \
 	proc_net_pktgen \
 	reuseaddr_conflict \
 	reuseport_bpf \
tools/testing/selftests/net/arp_ndisc_no_bcastnull_poison.sh +6 -6
index dd0bdd0e3f37..8dc79da27a2e 100755
@@ -75,7 +75,7 @@ run_no_arp_poisoning() {
 	ip netns exec ${PEER_NS} ping -c 1 ${V4_ADDR0} >/dev/null 2>&1
 
 	# Poison with a valid MAC to ensure injection is working
-	./arp_send ${veth0_ifindex} ${BCAST_MAC} ${VALID_MAC} ${op} \
+	python3 ./arp_send.py "veth0" ${BCAST_MAC} ${VALID_MAC} ${op} \
 		${V4_ADDR0} ${VALID_MAC} ${V4_ADDR0} ${VALID_MAC}
 
 	neigh=$(ip netns exec ${PEER_NS} ip neigh show ${V4_ADDR0} | \
@@ -88,7 +88,7 @@ run_no_arp_poisoning() {
 	fi
 
 	# Poison with tmac
-	./arp_send ${veth0_ifindex} ${l2_dmac} ${VALID_MAC} ${op} \
+	python3 ./arp_send.py "veth0" ${BCAST_MAC} ${VALID_MAC} ${op} \
 		${V4_ADDR0} ${tmac} ${V4_ADDR0} ${tmac}
 
 	neigh=$(ip netns exec ${PEER_NS} ip neigh show ${V4_ADDR0} | \
@@ -119,8 +119,8 @@ run_no_ndp_poisoning() {
 	ip netns exec ${PEER_NS} ping -c 1 ${V6_ADDR0} >/dev/null 2>&1
 
 	# Poison with a valid MAC to ensure injection is working
-	./ndisc_send ${veth0_ifindex} ${l2_dmac} ${VALID_MAC} ${dst_ip} \
-		${V6_ADDR0} ${tip} ${op} ${VALID_MAC}
+	python3 ./ndisc_send.py "veth0" ${l2_dmac} ${VALID_MAC} \
+		${dst_ip} ${V6_ADDR0} ${tip} ${op} ${VALID_MAC}
 	neigh=$(ip netns exec ${PEER_NS} ip neigh show ${V6_ADDR0} | \
 		grep ${VALID_MAC})
 	if [ "${neigh}" == "" ]; then
@@ -131,8 +131,8 @@ run_no_ndp_poisoning() {
 	fi
 
 	# Poison with tmac
-	./ndisc_send ${veth0_ifindex} ${l2_dmac} ${VALID_MAC} ${dst_ip} \
-		${V6_ADDR0} ${tip} ${op} ${tmac}
+	python3 ./ndisc_send.py "veth0" ${l2_dmac} ${VALID_MAC} \
+		${dst_ip} ${V6_ADDR0} ${tip} ${op} ${tmac}
 	neigh=$(ip netns exec ${PEER_NS} ip neigh show ${V6_ADDR0} | \
 		grep ${tmac})
 	if [ "${neigh}" != "" ]; then
tools/testing/selftests/net/arp_send.c deleted +0 -138

138 lines not on this page · raw message

tools/testing/selftests/net/arp_send.py new file +24 -0
new file mode 100644
index 000000000000..1161dfd60b27
@@ -0,0 +1,24 @@
+#!/bin/bash
+# SPDX-License-Identifier: GPL-2.0
+
+import sys
+from scapy.all import *
+
+if len(sys.argv) != 9:
+    print(f"Usage: {sys.argv[0]} <iface> <mac_dst> <mac_src> <op_code> <target-ip> <target-hwaddr> <sender-ip> <sender-hwaddr>\n");
+
+iface = sys.argv[1]
+mac_dst = sys.argv[2]
+mac_src = sys.argv[3]
+op = int(sys.argv[4])
+tip = sys.argv[5]
+tha = sys.argv[6]
+sip = sys.argv[5]
+sha = sys.argv[6]
+
+pkt = (
+    Ether(dst=mac_dst, src=mac_src) /
+    ARP(op=op, psrc=sip, hwsrc=sha, pdst=tip, hwdst=tha)
+)
+
+sendp(pkt, iface=iface, verbose=False)
tools/testing/selftests/net/ndisc_send.c deleted +0 -198

198 lines not on this page · raw message

tools/testing/selftests/net/ndisc_send.py new file +36 -0
new file mode 100644
index 000000000000..7b1a1c057862
@@ -0,0 +1,36 @@
+#!/bin/bash
+# SPDX-License-Identifier: GPL-2.0
+
+import sys
+from scapy.all import *
+
+if len(sys.argv) != 9:
+    print(f"Usage: {sys.argv[0]} <iface> <mac_dst> <mac_src> <ip_dst> <ip_src> <target_ip> <op> <lladr>")
+
+iface = sys.argv[1]
+mac_dst = sys.argv[2]
+mac_src = sys.argv[3]
+ip_dst = sys.argv[4]
+ip_src = sys.argv[5]
+tip = sys.argv[6]
+op = int(sys.argv[7])
+lladdr = sys.argv[8]
+
+NDP_NA=136
+
+if op == NDP_NA:
+    pkt = (
+        Ether(dst=mac_dst, src=mac_src) /
+        IPv6(src=ip_src, dst=ip_dst, hlim=255) /
+        ICMPv6ND_NA(R=0, S=0, O=1, tgt=tip) /
+        ICMPv6NDOptDstLLAddr(lladdr=lladdr)
+    )
+else:
+    pkt = (
+        Ether(dst=mac_dst, src=mac_src) /
+        IPv6(src=ip_src, dst=ip_dst, hlim=255) /
+        ICMPv6ND_NS(tgt=tip) /
+        ICMPv6NDOptSrcLLAddr(lladdr=lladdr)
+    )
+
+sendp(pkt, iface=iface, verbose=False)
Signature · 1 line
2.47.3
#9 lore raw

Re: [PATCH RFC net 5/5] selftests/net: use scapy for no_bcastnull_poison

Quoted 2 lines · Marc Suñé

Use Scapy to generate ARP/ND packets for ARP/ND no bcast/NULL MAC poisoning.

What's wrong with current implementation in arp_send.c and ndisc_send.c?

#2 lore raw

Re: [PATCH RFC net 1/5] arp: discard sha bcast/null (bcast ARP poison)

In reply to Marc Suñé

net/ipv4/arp.c static int arp_process(struct net *net, struct sock *sk, struct sk_buff *skb) · line 806

Quoted 7 lines · Marc Suñé

/*
+ * For Ethernet devices, Broadcast/Multicast and zero MAC addresses should
+ * never be announced and accepted as sender HW address (prevent BCAST MAC
+ * and NULL ARP poisoning attack).
+ */

net/ipv4/arp.c +1 -0
+	if (dev->addr_len == ETH_ALEN &&

dev_type == ARPHRD_ETHER ?

net/ipv4/arp.c static int arp_process(struct net *net, struct sock *sk, struct sk_buff *skb) · line 807

Quoted 1 line
net/ipv4/arp.c +1 -0
+	    (is_broadcast_ether_addr(sha) || is_zero_ether_addr(sha)))

RFC says that neither broadcast, nor multicast must be believed. You check for broadcast only. The better check would be:

!is_unicast_ether_addr(sha)

Quoted 6 lines
net/ipv4/arp.c +3 -0
+		goto out_free_skb;
+
+ /*
   *     Special case: We must set Frame Relay source Q.922 address
   */
  	if (dev_type == ARPHRD_DLCI)
author
#3 lore raw

Re: [PATCH RFC net 1/5] arp: discard sha bcast/null (bcast ARP poison)

Missatge de Vadim Fedorenko <[email protected]> del dia dl., 22 de des. 2025 a les 10:47:

Quoted 11 lines · Marc Suñé
Quoted 7 lines

/*

+ *   For Ethernet devices, Broadcast/Multicast and zero MAC addresses should
+ *   never be announced and accepted as sender HW address (prevent BCAST MAC
+ *   and NULL ARP poisoning attack).

+ */
+ if (dev->addr_len == ETH_ALEN &&

dev_type == ARPHRD_ETHER ?

This is discussed in the cover letter, comments section d). I would think more dev_types than that need to check this, at least:

+       case ARPHRD_ETHER:
+       case ARPHRD_EETHER:
+       case ARPHRD_FDDI:
+       case ARPHRD_IEEE802:
+       case ARPHRD_IEEE80211:

but as said, I _think_ it's sufficient to check for HW addrlen == ETH_ALEN.

Quoted 8 lines
Quoted 3 lines

+ (is_broadcast_ether_addr(sha) || is_zero_ether_addr(sha)))

RFC says that neither broadcast, nor multicast must be believed. You check for broadcast only. The better check would be:

!is_unicast_ether_addr(sha)

This is discussed in the cover letter, comments section b). In short, some NLBs announce MCAST MAC addresses.

Mind the context there, but I think it's safe. This is applicable to ARP and NDP, so I would suggest to follow up there.

Btw, the is_zero_ether_addr(addr) check is still needed. is_unicast_ether_addr(addr) is implemented as !is_multicast_ether_addr(addr), and the NULL mac (00:00:00:00:00:00) doesn't have the "MCAST bit" set to 1.

Quoted 8 lines
Quoted 7 lines
net/ipv4/arp.c +3 -0
+             goto out_free_skb;
+
+ /*
   *     Special case: We must set Frame Relay source Q.922 address
   */
      if (dev_type == ARPHRD_DLCI)
author
#10 lore raw

Re: [PATCH RFC net 5/5] selftests/net: use scapy for no_bcastnull_poison

In reply to Vadim Fedorenko

Missatge de Vadim Fedorenko <[email protected]> del dia dl., 22 de des. 2025 a les 10:34:

Quoted 6 lines · Marc Suñé
Quoted 2 lines

Use Scapy to generate ARP/ND packets for ARP/ND no bcast/NULL MAC poisoning.

What's wrong with current implementation in arp_send.c and ndisc_send.c?

Nothing. The Scapy version of the tooling is shorter, that's all.

This is discussed in the cover letter, comments section c). Patch 5/5 can be dropped if Scapy is not acceptable, else squash them with patches 2 (arp_send.c) and 4 (ndisc_send.c).

#4 lore raw

Re: [PATCH RFC net 1/5] arp: discard sha bcast/null (bcast ARP poison)

In reply to Marc Sune
Quoted 23 lines · Marc Sune

Missatge de Vadim Fedorenko <[email protected]> del dia dl., 22 de des. 2025 a les 10:47:

Quoted 11 lines · Marc Suñé
Quoted 7 lines
   /*
+ *   For Ethernet devices, Broadcast/Multicast and zero MAC addresses should
+ *   never be announced and accepted as sender HW address (prevent BCAST MAC
+ *   and NULL ARP poisoning attack).

+ */
+ if (dev->addr_len == ETH_ALEN &&

dev_type == ARPHRD_ETHER ?

This is discussed in the cover letter, comments section d). I would think more dev_types than that need to check this, at least:

(quoted diff) +5 -0
+       case ARPHRD_ETHER:
+       case ARPHRD_EETHER:
+       case ARPHRD_FDDI:
+       case ARPHRD_IEEE802:
+       case ARPHRD_IEEE80211:
Quoted 1 line

but as said, I _think_ it's sufficient to check for HW addrlen == ETH_ALEN.

ARPHRD_EETHER and ARPHRD_IEEE80211 are not really used in the kernel. For other 3 we already have such case a bit earlier in arp_process(), it's fine to be aligned with the existing code.

Quoted 13 lines
Quoted 9 lines
Quoted 4 lines

+ (is_broadcast_ether_addr(sha) || is_zero_ether_addr(sha)))

RFC says that neither broadcast, nor multicast must be believed. You check for broadcast only. The better check would be:

!is_unicast_ether_addr(sha)

This is discussed in the cover letter, comments section b). In short, some NLBs announce MCAST MAC addresses.

Quoted 7 lines

Mind the context there, but I think it's safe. This is applicable to ARP and NDP, so I would suggest to follow up there.

Btw, the is_zero_ether_addr(addr) check is still needed. is_unicast_ether_addr(addr) is implemented as !is_multicast_ether_addr(addr), and the NULL mac (00:00:00:00:00:00) doesn't have the "MCAST bit" set to 1.

Yeah, "!is_valid_ether_addr(sha)" is better in this case.

Quoted 9 lines
Quoted 9 lines
Quoted 8 lines
net/ipv4/arp.c +3 -0
+             goto out_free_skb;
+
+ /*
    *     Special case: We must set Frame Relay source Q.922 address
    */
       if (dev_type == ARPHRD_DLCI)