Merge branch 'icmp-add-rfc-5837-support'

Ido Schimmel says:

====================
icmp: Add RFC 5837 support

tl;dr
=====

This patchset extends certain ICMP error messages (e.g., "Time
Exceeded") with incoming interface information in accordance with RFC
5837 [1]. This is required for more meaningful traceroute results in
unnumbered networks. Like other ICMP settings, the feature is controlled
via a per-{netns, address family} sysctl. The interface and the
implementation are designed to support more ICMP extensions.

Motivation
==========

Over the years, the kernel was extended with the ability to derive the
source IP of ICMP error messages from the interface that received the
datagram which elicited the ICMP error [2][3][4]. This is especially
important for "Time Exceeded" messages as it allows traceroute users to
trace the actual packet path along the network.

The above scheme does not work in unnumbered networks. In these
networks, only the loopback / VRF interface is assigned a global IP
address while router interfaces are assigned IPv6 link-local addresses.
As such, ICMP error messages are generated with a source IP derived from
the loopback / VRF interface, making it impossible to trace the actual
packet path when parallel links exist between routers.

The problem can be solved by implementing the solution proposed by RFC
4884 [5] and RFC 5837. The former defines an ICMP extension structure
that can be appended to selected ICMP messages and carry extension
objects. The latter defines an extension object called the "Interface
Information Object" (IIO) that can carry interface information (e.g.,
name, index, MTU) about interfaces with certain roles such as the
interface that received the datagram which elicited the ICMP error.

The payload of the datagram that elicited the error (potentially padded
/ trimmed) along with the ICMP extension structure will be queued to the
error queue of the originating socket, thereby allowing traceroute
applications to parse and display the information encoded in the ICMP
extension structure. Example:

# traceroute6 -e 2001:db8:1::3
traceroute to 2001:db8:1::3 (2001:db8:1::3), 30 hops max, 80 byte packets
1 2001:db8:1::2 (2001:db8:1::2) <INC:11,"eth1",mtu=1500> 0.214 ms 0.171 ms 0.162 ms
2 2001:db8:1::3 (2001:db8:1::3) <INC:12,"eth2",mtu=1500> 0.154 ms 0.135 ms 0.127 ms

# traceroute -e 192.0.2.3
traceroute to 192.0.2.3 (192.0.2.3), 30 hops max, 60 byte packets
1 192.0.2.2 (192.0.2.2) <INC:11,"eth1",mtu=1500> 0.191 ms 0.148 ms 0.144 ms
2 192.0.2.3 (192.0.2.3) <INC:12,"eth2",mtu=1500> 0.137 ms 0.122 ms 0.114 ms

Implementation
==============

As previously stated, the feature is controlled via a per-{netns,
address} sysctl. Specifically, a bit mask where each bit controls the
addition of a different ICMP extension to ICMP error messages.
Currently, only a single value is supported, to append the incoming
interface information.

Key points:

1. Global knob vs finer control. I am not aware of users who require
finer control, but it is possible that some users will want to avoid
appending ICMP extensions when the packet is sent out of a specific
interface (e.g., the management interface) or to a specific subnet. This
can be accomplished via a tc-bpf program that trims the ICMP extension
structure. An example program can be found here [6].

2. Split implementation between IPv4 / IPv6. While the implementation is
currently similar, there are some differences between both address
families. In addition, some extensions (e.g., RFC 8883 [7]) are
IPv6-specific. Given the above and given that the implementation is not
very complex, it makes sense to keep both implementations separate.

3. Compatibility with legacy applications. RFC 4884 from 2007 extended
certain ICMP messages with a length field that encodes the length of the
"original datagram" field, so that applications will be able to tell
where the "original datagram" ends and where the ICMP extension
structure starts.

Before the introduction of the IP{,6}_RECVERR_RFC4884 socket options
[8][9] in 2020 it was impossible for applications to know where the ICMP
extension structure starts and to this day some applications assume that
it starts at offset 128, which is the minimum length of the "original
datagram" field as specified by RFC 4884.

Therefore, in order to be compatible with both legacy and modern
applications, the datagram that elicited the ICMP error is trimmed /
padded to 128 bytes before appending the ICMP extension structure.

This behavior is specifically called out by RFC 4884: "Those wishing to
be backward compatible with non-compliant TRACEROUTE implementations
will include exactly 128 octets" [10].

Note that in 128 bytes we should be able to include enough headers for
the originating node to match the ICMP error message with the relevant
socket. For example, the following headers will be present in the
"original datagram" field when a VXLAN encapsulated IPv6 packet elicits
an ICMP error in an IPv6 underlay: IPv6 (40) | UDP (8) | VXLAN (8) | Eth
(14) | IPv6 (40) | UDP (8). Overall, 118 bytes.

If the 128 bytes limit proves to be insufficient for some use case, we
can consider dedicating a new bit in the previously mentioned sysctl to
allow for more bytes to be included in the "original datagram" field.

4. Extensibility. This patchset adds partial support for a single ICMP
extension. However, the interface and the implementation should be able
to support more extensions, if needed. Examples:

* More interface information objects as part of RFC 5837. We should be
able to derive the outgoing interface information and nexthop IP from
the dst entry attached to the packet that elicited the error.

* Node identification object (e.g., hostname / loopback IP) [11].

* Extended Information object which encodes aggregate header limits as
part of RFC 8883.

A previous proposal from Ishaan Gandhi and Ron Bonica is available here
[12].

Testing
=======

The existing traceroute selftest is extended to test that ICMP
extensions are reported correctly when enabled. Both address families
are tested and with different packet sizes in order to make sure that
trimming / padding works correctly. Tested that packets are parsed
correctly by the IP{,6}_RECVERR_RFC4884 socket options using Willem's
selftest [13].

Changelog
=========

Changes since v1 [14]:

* Patches #1-#2: Added a comment about field ordering and review tags.

* Patch #3: Converted "sysctl" to "echo" when testing the return value.
Added a check to skip the test if traceroute version is older
than 2.1.5.

[1] https://datatracker.ietf.org/doc/html/rfc5837
[2] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=1c2fb7f93cb20621772bf304f3dba0849942e5db
[3] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=fac6fce9bdb59837bb89930c3a92f5e0d1482f0b
[4] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=4a8c416602d97a4e2073ed563d4d4c7627de19cf
[5] https://datatracker.ietf.org/doc/html/rfc4884
[6] https://gist.github.com/idosch/5013448cdb5e9e060e6bfdc8b433577c
[7] https://datatracker.ietf.org/doc/html/rfc8883
[8] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=eba75c587e811d3249c8bd50d22bb2266ccd3c0f
[9] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=01370434df85eb76ecb1527a4466013c4aca2436
[10] https://datatracker.ietf.org/doc/html/rfc4884#section-5.3
[11] https://datatracker.ietf.org/doc/html/draft-ietf-intarea-extended-icmp-nodeid-04
[12] https://lore.kernel.org/netdev/[e-mail]/
[13] https://lore.kernel.org/netdev/aPpMItF35gwpgzZx@shredder/
[14] https://lore.kernel.org/netdev/[e-mail]/
====================

Link: https://patch.msgid.link/[e-mail]
Signed-off-by: Jakub Kicinski <[e-mail]>

Documentation/networking/ip-sysctl.rst

···
 
 	Default: 0 (disabled)
 
+icmp_errors_extension_mask - UNSIGNED INTEGER
+	Bitmask of ICMP extensions to append to ICMPv4 error messages
+	("Destination Unreachable", "Time Exceeded" and "Parameter Problem").
+	The original datagram is trimmed / padded to 128 bytes in order to be
+	compatible with applications that do not comply with RFC 4884.
+
+	Possible extensions are:
+
+	==== ==============================================================
+	0x01 Incoming IP interface information according to RFC 5837.
+	     Extension will include the index, IPv4 address (if present),
+	     name and MTU of the IP interface that received the datagram
+	     which elicited the ICMP error.
+	==== ==============================================================
+
+	Default: 0x00 (no extensions)
+
 igmp_max_memberships - INTEGER
 	Change the maximum number of multicast groups we can subscribe to.
 	Default: 20
···
 	- 1 (enabled)
 
 	Default: 0 (disabled)
+
+errors_extension_mask - UNSIGNED INTEGER
+	Bitmask of ICMP extensions to append to ICMPv6 error messages
+	("Destination Unreachable" and "Time Exceeded"). The original datagram
+	is trimmed / padded to 128 bytes in order to be compatible with
+	applications that do not comply with RFC 4884.
+
+	Possible extensions are:
+
+	==== ==============================================================
+	0x01 Incoming IP interface information according to RFC 5837.
+	     Extension will include the index, IPv6 address (if present),
+	     name and MTU of the IP interface that received the datagram
+	     which elicited the ICMP error.
+	==== ==============================================================
+
+	Default: 0x00 (no extensions)
 
 xfrm6_gc_thresh - INTEGER
 	(Obsolete since linux-4.14)

include/linux/icmp.h

···
 			   struct sock_ee_data_rfc4884 *out,
 			   int thlen, int off);
 
+/* RFC 4884 */
+#define ICMP_EXT_ORIG_DGRAM_MIN_LEN	128
+#define ICMP_EXT_VERSION_2		2
+
+/* ICMP Extension Object Classes */
+#define ICMP_EXT_OBJ_CLASS_IIO		2	/* RFC 5837 */
+
+/* Interface Information Object - RFC 5837 */
+enum {
+	ICMP_EXT_CTYPE_IIO_ROLE_IIF,
+};
+
+#define ICMP_EXT_CTYPE_IIO_ROLE(ROLE)	((ROLE) << 6)
+#define ICMP_EXT_CTYPE_IIO_MTU		BIT(0)
+#define ICMP_EXT_CTYPE_IIO_NAME		BIT(1)
+#define ICMP_EXT_CTYPE_IIO_IPADDR	BIT(2)
+#define ICMP_EXT_CTYPE_IIO_IFINDEX	BIT(3)
+
+struct icmp_ext_iio_name_subobj {
+	u8 len;
+	char name[IFNAMSIZ];
+};
+
+enum {
+	/* RFC 5837 - Incoming IP Interface Role */
+	ICMP_ERR_EXT_IIO_IIF,
+	/* Add new constants above. Used by "icmp_errors_extension_mask"
+	 * sysctl.
+	 */
+	ICMP_ERR_EXT_COUNT,
+};
+
 #endif	/* _LINUX_ICMP_H */

include/net/netns/ipv4.h

···
 	u8 sysctl_icmp_echo_ignore_broadcasts;
 	u8 sysctl_icmp_ignore_bogus_error_responses;
 	u8 sysctl_icmp_errors_use_inbound_ifaddr;
+	u8 sysctl_icmp_errors_extension_mask;
 	int sysctl_icmp_ratelimit;
 	int sysctl_icmp_ratemask;
 	int sysctl_icmp_msgs_per_sec;

include/net/netns/ipv6.h

···
 	u8 skip_notify_on_dev_down;
 	u8 fib_notify_on_flag_change;
 	u8 icmpv6_error_anycast_as_unicast;
+	u8 icmpv6_errors_extension_mask;
 };
 
 struct netns_ipv6 {

net/core/dev.c

···
 		strscpy(name, dev->name, IFNAMSIZ);
 	} while (read_seqretry(&netdev_rename_lock, seq));
 }
+EXPORT_IPV6_MOD_GPL(netdev_copy_name);
 
 /**
  *	netdev_get_name - get a netdevice name, knowing its ifindex.

net/ipv4/icmp.c

···
 	return ERR_PTR(err);
 }
 
+struct icmp_ext_iio_addr4_subobj {
+	__be16 afi;
+	__be16 reserved;
+	__be32 addr4;
+};
+
+static unsigned int icmp_ext_iio_len(void)
+{
+	return sizeof(struct icmp_extobj_hdr) +
+		/* ifIndex */
+		sizeof(__be32) +
+		/* Interface Address Sub-Object */
+		sizeof(struct icmp_ext_iio_addr4_subobj) +
+		/* Interface Name Sub-Object. Length must be a multiple of 4
+		 * bytes.
+		 */
+		ALIGN(sizeof(struct icmp_ext_iio_name_subobj), 4) +
+		/* MTU */
+		sizeof(__be32);
+}
+
+static unsigned int icmp_ext_max_len(u8 ext_objs)
+{
+	unsigned int ext_max_len;
+
+	ext_max_len = sizeof(struct icmp_ext_hdr);
+
+	if (ext_objs & BIT(ICMP_ERR_EXT_IIO_IIF))
+		ext_max_len += icmp_ext_iio_len();
+
+	return ext_max_len;
+}
+
+static __be32 icmp_ext_iio_addr4_find(const struct net_device *dev)
+{
+	struct in_device *in_dev;
+	struct in_ifaddr *ifa;
+
+	in_dev = __in_dev_get_rcu(dev);
+	if (!in_dev)
+		return 0;
+
+	/* It is unclear from RFC 5837 which IP address should be chosen, but
+	 * it makes sense to choose a global unicast address.
+	 */
+	in_dev_for_each_ifa_rcu(ifa, in_dev) {
+		if (READ_ONCE(ifa->ifa_flags) & IFA_F_SECONDARY)
+			continue;
+		if (ifa->ifa_scope != RT_SCOPE_UNIVERSE ||
+		    ipv4_is_multicast(ifa->ifa_address))
+			continue;
+		return ifa->ifa_address;
+	}
+
+	return 0;
+}
+
+static void icmp_ext_iio_iif_append(struct net *net, struct sk_buff *skb,
+				    int iif)
+{
+	struct icmp_ext_iio_name_subobj *name_subobj;
+	struct icmp_extobj_hdr *objh;
+	struct net_device *dev;
+	__be32 data;
+
+	if (!iif)
+		return;
+
+	/* Add the fields in the order specified by RFC 5837. */
+	objh = skb_put(skb, sizeof(*objh));
+	objh->class_num = ICMP_EXT_OBJ_CLASS_IIO;
+	objh->class_type = ICMP_EXT_CTYPE_IIO_ROLE(ICMP_EXT_CTYPE_IIO_ROLE_IIF);
+
+	data = htonl(iif);
+	skb_put_data(skb, &data, sizeof(__be32));
+	objh->class_type |= ICMP_EXT_CTYPE_IIO_IFINDEX;
+
+	rcu_read_lock();
+
+	dev = dev_get_by_index_rcu(net, iif);
+	if (!dev)
+		goto out;
+
+	data = icmp_ext_iio_addr4_find(dev);
+	if (data) {
+		struct icmp_ext_iio_addr4_subobj *addr4_subobj;
+
+		addr4_subobj = skb_put_zero(skb, sizeof(*addr4_subobj));
+		addr4_subobj->afi = htons(ICMP_AFI_IP);
+		addr4_subobj->addr4 = data;
+		objh->class_type |= ICMP_EXT_CTYPE_IIO_IPADDR;
+	}
+
+	name_subobj = skb_put_zero(skb, ALIGN(sizeof(*name_subobj), 4));
+	name_subobj->len = ALIGN(sizeof(*name_subobj), 4);
+	netdev_copy_name(dev, name_subobj->name);
+	objh->class_type |= ICMP_EXT_CTYPE_IIO_NAME;
+
+	data = htonl(READ_ONCE(dev->mtu));
+	skb_put_data(skb, &data, sizeof(__be32));
+	objh->class_type |= ICMP_EXT_CTYPE_IIO_MTU;
+
+out:
+	rcu_read_unlock();
+	objh->length = htons(skb_tail_pointer(skb) - (unsigned char *)objh);
+}
+
+static void icmp_ext_objs_append(struct net *net, struct sk_buff *skb,
+				 u8 ext_objs, int iif)
+{
+	if (ext_objs & BIT(ICMP_ERR_EXT_IIO_IIF))
+		icmp_ext_iio_iif_append(net, skb, iif);
+}
+
+static struct sk_buff *
+icmp_ext_append(struct net *net, struct sk_buff *skb_in, struct icmphdr *icmph,
+		unsigned int room, int iif)
+{
+	unsigned int payload_len, ext_max_len, ext_len;
+	struct icmp_ext_hdr *ext_hdr;
+	struct sk_buff *skb;
+	u8 ext_objs;
+	int nhoff;
+
+	switch (icmph->type) {
+	case ICMP_DEST_UNREACH:
+	case ICMP_TIME_EXCEEDED:
+	case ICMP_PARAMETERPROB:
+		break;
+	default:
+		return NULL;
+	}
+
+	ext_objs = READ_ONCE(net->ipv4.sysctl_icmp_errors_extension_mask);
+	if (!ext_objs)
+		return NULL;
+
+	ext_max_len = icmp_ext_max_len(ext_objs);
+	if (ICMP_EXT_ORIG_DGRAM_MIN_LEN + ext_max_len > room)
+		return NULL;
+
+	skb = skb_clone(skb_in, GFP_ATOMIC);
+	if (!skb)
+		return NULL;
+
+	nhoff = skb_network_offset(skb);
+	payload_len = min(skb->len - nhoff, ICMP_EXT_ORIG_DGRAM_MIN_LEN);
+
+	if (!pskb_network_may_pull(skb, payload_len))
+		goto free_skb;
+
+	if (pskb_trim(skb, nhoff + ICMP_EXT_ORIG_DGRAM_MIN_LEN) ||
+	    __skb_put_padto(skb, nhoff + ICMP_EXT_ORIG_DGRAM_MIN_LEN, false))
+		goto free_skb;
+
+	if (pskb_expand_head(skb, 0, ext_max_len, GFP_ATOMIC))
+		goto free_skb;
+
+	ext_hdr = skb_put_zero(skb, sizeof(*ext_hdr));
+	ext_hdr->version = ICMP_EXT_VERSION_2;
+
+	icmp_ext_objs_append(net, skb, ext_objs, iif);
+
+	/* Do not send an empty extension structure. */
+	ext_len = skb_tail_pointer(skb) - (unsigned char *)ext_hdr;
+	if (ext_len == sizeof(*ext_hdr))
+		goto free_skb;
+
+	ext_hdr->checksum = ip_compute_csum(ext_hdr, ext_len);
+	/* The length of the original datagram in 32-bit words (RFC 4884). */
+	icmph->un.reserved[1] = ICMP_EXT_ORIG_DGRAM_MIN_LEN / sizeof(u32);
+
+	return skb;
+
+free_skb:
+	consume_skb(skb);
+	return NULL;
+}
+
 /*
  *	Send an ICMP message in response to a situation
  *
···
 	struct icmp_bxm icmp_param;
 	struct rtable *rt = skb_rtable(skb_in);
 	bool apply_ratelimit = false;
+	struct sk_buff *ext_skb;
 	struct ipcm_cookie ipc;
 	struct flowi4 fl4;
 	__be32 saddr;
···
 	if (room <= (int)sizeof(struct iphdr))
 		goto ende;
 
-	icmp_param.data_len = skb_in->len - icmp_param.offset;
+	ext_skb = icmp_ext_append(net, skb_in, &icmp_param.data.icmph, room,
+				  parm->iif);
+	if (ext_skb)
+		icmp_param.skb = ext_skb;
+
+	icmp_param.data_len = icmp_param.skb->len - icmp_param.offset;
 	if (icmp_param.data_len > room)
 		icmp_param.data_len = room;
 	icmp_param.head_len = sizeof(struct icmphdr);
···
 	trace_icmp_send(skb_in, type, code);
 
 	icmp_push_reply(sk, &icmp_param, &fl4, &ipc, &rt);
+
+	if (ext_skb)
+		consume_skb(ext_skb);
 ende:
 	ip_rt_put(rt);
 out_unlock:
···
 	net->ipv4.sysctl_icmp_ratelimit = 1 * HZ;
 	net->ipv4.sysctl_icmp_ratemask = 0x1818;
 	net->ipv4.sysctl_icmp_errors_use_inbound_ifaddr = 0;
+	net->ipv4.sysctl_icmp_errors_extension_mask = 0;
 	net->ipv4.sysctl_icmp_msgs_per_sec = 1000;
 	net->ipv4.sysctl_icmp_msgs_burst = 50;
 

net/ipv4/sysctl_net_ipv4.c

···
 static int tcp_plb_max_cong_thresh = 256;
 static unsigned int tcp_tw_reuse_delay_max = TCP_PAWS_MSL * MSEC_PER_SEC;
 static int tcp_ecn_mode_max = 2;
+static u32 icmp_errors_extension_mask_all =
+	GENMASK_U8(ICMP_ERR_EXT_COUNT - 1, 0);
 
 /* obsolete */
 static int sysctl_tcp_low_latency __read_mostly;
···
 		.proc_handler	= proc_dou8vec_minmax,
 		.extra1		= SYSCTL_ZERO,
 		.extra2		= SYSCTL_ONE
+	},
+	{
+		.procname	= "icmp_errors_extension_mask",
+		.data		= &init_net.ipv4.sysctl_icmp_errors_extension_mask,
+		.maxlen		= sizeof(u8),
+		.mode		= 0644,
+		.proc_handler	= proc_dou8vec_minmax,
+		.extra1		= SYSCTL_ZERO,
+		.extra2		= &icmp_errors_extension_mask_all,
 	},
 	{
 		.procname	= "icmp_ratelimit",

net/ipv6/af_inet6.c

···
 	net->ipv6.sysctl.icmpv6_echo_ignore_multicast = 0;
 	net->ipv6.sysctl.icmpv6_echo_ignore_anycast = 0;
 	net->ipv6.sysctl.icmpv6_error_anycast_as_unicast = 0;
+	net->ipv6.sysctl.icmpv6_errors_extension_mask = 0;
 
 	/* By default, rate limit error messages.
 	 * Except for pmtu discovery, it would break it.

net/ipv6/icmp.c

···
 	return icmp6_dev(skb)->ifindex;
 }
 
+struct icmp6_ext_iio_addr6_subobj {
+	__be16 afi;
+	__be16 reserved;
+	struct in6_addr addr6;
+};
+
+static unsigned int icmp6_ext_iio_len(void)
+{
+	return sizeof(struct icmp_extobj_hdr) +
+		/* ifIndex */
+		sizeof(__be32) +
+		/* Interface Address Sub-Object */
+		sizeof(struct icmp6_ext_iio_addr6_subobj) +
+		/* Interface Name Sub-Object. Length must be a multiple of 4
+		 * bytes.
+		 */
+		ALIGN(sizeof(struct icmp_ext_iio_name_subobj), 4) +
+		/* MTU */
+		sizeof(__be32);
+}
+
+static unsigned int icmp6_ext_max_len(u8 ext_objs)
+{
+	unsigned int ext_max_len;
+
+	ext_max_len = sizeof(struct icmp_ext_hdr);
+
+	if (ext_objs & BIT(ICMP_ERR_EXT_IIO_IIF))
+		ext_max_len += icmp6_ext_iio_len();
+
+	return ext_max_len;
+}
+
+static struct in6_addr *icmp6_ext_iio_addr6_find(const struct net_device *dev)
+{
+	struct inet6_dev *in6_dev;
+	struct inet6_ifaddr *ifa;
+
+	in6_dev = __in6_dev_get(dev);
+	if (!in6_dev)
+		return NULL;
+
+	/* It is unclear from RFC 5837 which IP address should be chosen, but
+	 * it makes sense to choose a global unicast address.
+	 */
+	list_for_each_entry_rcu(ifa, &in6_dev->addr_list, if_list) {
+		if (ifa->flags & (IFA_F_TENTATIVE | IFA_F_DADFAILED))
+			continue;
+		if (ipv6_addr_type(&ifa->addr) != IPV6_ADDR_UNICAST ||
+		    ipv6_addr_src_scope(&ifa->addr) != IPV6_ADDR_SCOPE_GLOBAL)
+			continue;
+		return &ifa->addr;
+	}
+
+	return NULL;
+}
+
+static void icmp6_ext_iio_iif_append(struct net *net, struct sk_buff *skb,
+				     int iif)
+{
+	struct icmp_ext_iio_name_subobj *name_subobj;
+	struct icmp_extobj_hdr *objh;
+	struct net_device *dev;
+	struct in6_addr *addr6;
+	__be32 data;
+
+	if (!iif)
+		return;
+
+	/* Add the fields in the order specified by RFC 5837. */
+	objh = skb_put(skb, sizeof(*objh));
+	objh->class_num = ICMP_EXT_OBJ_CLASS_IIO;
+	objh->class_type = ICMP_EXT_CTYPE_IIO_ROLE(ICMP_EXT_CTYPE_IIO_ROLE_IIF);
+
+	data = htonl(iif);
+	skb_put_data(skb, &data, sizeof(__be32));
+	objh->class_type |= ICMP_EXT_CTYPE_IIO_IFINDEX;
+
+	rcu_read_lock();
+
+	dev = dev_get_by_index_rcu(net, iif);
+	if (!dev)
+		goto out;
+
+	addr6 = icmp6_ext_iio_addr6_find(dev);
+	if (addr6) {
+		struct icmp6_ext_iio_addr6_subobj *addr6_subobj;
+
+		addr6_subobj = skb_put_zero(skb, sizeof(*addr6_subobj));
+		addr6_subobj->afi = htons(ICMP_AFI_IP6);
+		addr6_subobj->addr6 = *addr6;
+		objh->class_type |= ICMP_EXT_CTYPE_IIO_IPADDR;
+	}
+
+	name_subobj = skb_put_zero(skb, ALIGN(sizeof(*name_subobj), 4));
+	name_subobj->len = ALIGN(sizeof(*name_subobj), 4);
+	netdev_copy_name(dev, name_subobj->name);
+	objh->class_type |= ICMP_EXT_CTYPE_IIO_NAME;
+
+	data = htonl(READ_ONCE(dev->mtu));
+	skb_put_data(skb, &data, sizeof(__be32));
+	objh->class_type |= ICMP_EXT_CTYPE_IIO_MTU;
+
+out:
+	rcu_read_unlock();
+	objh->length = htons(skb_tail_pointer(skb) - (unsigned char *)objh);
+}
+
+static void icmp6_ext_objs_append(struct net *net, struct sk_buff *skb,
+				  u8 ext_objs, int iif)
+{
+	if (ext_objs & BIT(ICMP_ERR_EXT_IIO_IIF))
+		icmp6_ext_iio_iif_append(net, skb, iif);
+}
+
+static struct sk_buff *
+icmp6_ext_append(struct net *net, struct sk_buff *skb_in,
+		 struct icmp6hdr *icmp6h, unsigned int room, int iif)
+{
+	unsigned int payload_len, ext_max_len, ext_len;
+	struct icmp_ext_hdr *ext_hdr;
+	struct sk_buff *skb;
+	u8 ext_objs;
+	int nhoff;
+
+	switch (icmp6h->icmp6_type) {
+	case ICMPV6_DEST_UNREACH:
+	case ICMPV6_TIME_EXCEED:
+		break;
+	default:
+		return NULL;
+	}
+
+	/* Do not overwrite existing extensions. This can happen when we
+	 * receive an ICMPv4 message with extensions from a tunnel and
+	 * translate it to an ICMPv6 message towards an IPv6 host in the
+	 * overlay network.
+	 */
+	if (icmp6h->icmp6_datagram_len)
+		return NULL;
+
+	ext_objs = READ_ONCE(net->ipv6.sysctl.icmpv6_errors_extension_mask);
+	if (!ext_objs)
+		return NULL;
+
+	ext_max_len = icmp6_ext_max_len(ext_objs);
+	if (ICMP_EXT_ORIG_DGRAM_MIN_LEN + ext_max_len > room)
+		return NULL;
+
+	skb = skb_clone(skb_in, GFP_ATOMIC);
+	if (!skb)
+		return NULL;
+
+	nhoff = skb_network_offset(skb);
+	payload_len = min(skb->len - nhoff, ICMP_EXT_ORIG_DGRAM_MIN_LEN);
+
+	if (!pskb_network_may_pull(skb, payload_len))
+		goto free_skb;
+
+	if (pskb_trim(skb, nhoff + ICMP_EXT_ORIG_DGRAM_MIN_LEN) ||
+	    __skb_put_padto(skb, nhoff + ICMP_EXT_ORIG_DGRAM_MIN_LEN, false))
+		goto free_skb;
+
+	if (pskb_expand_head(skb, 0, ext_max_len, GFP_ATOMIC))
+		goto free_skb;
+
+	ext_hdr = skb_put_zero(skb, sizeof(*ext_hdr));
+	ext_hdr->version = ICMP_EXT_VERSION_2;
+
+	icmp6_ext_objs_append(net, skb, ext_objs, iif);
+
+	/* Do not send an empty extension structure. */
+	ext_len = skb_tail_pointer(skb) - (unsigned char *)ext_hdr;
+	if (ext_len == sizeof(*ext_hdr))
+		goto free_skb;
+
+	ext_hdr->checksum = ip_compute_csum(ext_hdr, ext_len);
+	/* The length of the original datagram in 64-bit words (RFC 4884). */
+	icmp6h->icmp6_datagram_len = ICMP_EXT_ORIG_DGRAM_MIN_LEN / sizeof(u64);
+
+	return skb;
+
+free_skb:
+	consume_skb(skb);
+	return NULL;
+}
+
 /*
  *	Send an ICMP message in response to a packet in error
  */
···
 	struct ipv6_pinfo *np;
 	const struct in6_addr *saddr = NULL;
 	bool apply_ratelimit = false;
+	struct sk_buff *ext_skb;
 	struct dst_entry *dst;
+	unsigned int room;
 	struct icmp6hdr tmp_hdr;
 	struct flowi6 fl6;
 	struct icmpv6_msg msg;
···
 	msg.offset = skb_network_offset(skb);
 	msg.type = type;
 
-	len = skb->len - msg.offset;
-	len = min_t(unsigned int, len, IPV6_MIN_MTU - sizeof(struct ipv6hdr) - sizeof(struct icmp6hdr));
+	room = IPV6_MIN_MTU - sizeof(struct ipv6hdr) - sizeof(struct icmp6hdr);
+	ext_skb = icmp6_ext_append(net, skb, &tmp_hdr, room, parm->iif);
+	if (ext_skb)
+		msg.skb = ext_skb;
+
+	len = msg.skb->len - msg.offset;
+	len = min_t(unsigned int, len, room);
 	if (len < 0) {
 		net_dbg_ratelimited("icmp: len problem [%pI6c > %pI6c]\n",
 				    &hdr->saddr, &hdr->daddr);
···
 	}
 
 out_dst_release:
+	if (ext_skb)
+		consume_skb(ext_skb);
 	dst_release(dst);
 out_unlock:
 	icmpv6_xmit_unlock(sk);
···
 EXPORT_SYMBOL(icmpv6_err_convert);
 
 #ifdef CONFIG_SYSCTL
+
+static u32 icmpv6_errors_extension_mask_all =
+	GENMASK_U8(ICMP_ERR_EXT_COUNT - 1, 0);
+
 static struct ctl_table ipv6_icmp_table_template[] = {
 	{
 		.procname	= "ratelimit",
···
 		.extra1		= SYSCTL_ZERO,
 		.extra2		= SYSCTL_ONE,
 	},
+	{
+		.procname	= "errors_extension_mask",
+		.data		= &init_net.ipv6.sysctl.icmpv6_errors_extension_mask,
+		.maxlen		= sizeof(u8),
+		.mode		= 0644,
+		.proc_handler	= proc_dou8vec_minmax,
+		.extra1		= SYSCTL_ZERO,
+		.extra2		= &icmpv6_errors_extension_mask_all,
+	},
 };
 
 struct ctl_table * __net_init ipv6_icmp_sysctl_init(struct net *net)
···
 		table[3].data = &net->ipv6.sysctl.icmpv6_echo_ignore_anycast;
 		table[4].data = &net->ipv6.sysctl.icmpv6_ratemask_ptr;
 		table[5].data = &net->ipv6.sysctl.icmpv6_error_anycast_as_unicast;
+		table[6].data = &net->ipv6.sysctl.icmpv6_errors_extension_mask;
 	}
 	return table;
 }

tools/testing/selftests/net/traceroute.sh

···
 	return $rc
 }
 
+__check_traceroute_version()
+{
+	local cmd=$1; shift
+	local req_ver=$1; shift
+	local ver
+
+	req_ver=$(echo "$req_ver" | sed 's/\.//g')
+	ver=$($cmd -V 2>&1 | grep -Eo '[0-9]+.[0-9]+.[0-9]+' | sed 's/\.//g')
+	if [[ $ver -lt $req_ver ]]; then
+		return 1
+	else
+		return 0
+	fi
+}
+
+check_traceroute6_version()
+{
+	local req_ver=$1; shift
+
+	__check_traceroute_version traceroute6 "$req_ver"
+}
+
+check_traceroute_version()
+{
+	local req_ver=$1; shift
+
+	__check_traceroute_version traceroute "$req_ver"
+}
+
 ################################################################################
 # create namespaces and interconnects
 
···
 	ip netns exec ${ns} ip -6 ro add unreachable default metric 8192
 
 	ip netns exec ${ns} sysctl -qw net.ipv4.ip_forward=1
+	ip netns exec ${ns} sysctl -qw net.ipv4.icmp_ratelimit=0
+	ip netns exec ${ns} sysctl -qw net.ipv6.icmp.ratelimit=0
 	ip netns exec ${ns} sysctl -qw net.ipv6.conf.all.keep_addr_on_down=1
 	ip netns exec ${ns} sysctl -qw net.ipv6.conf.all.forwarding=1
 	ip netns exec ${ns} sysctl -qw net.ipv6.conf.default.forwarding=1
···
 }
 
 ################################################################################
+# traceroute6 with ICMP extensions test
+#
+# Verify that in this scenario
+#
+# ----                          ----                          ----
+# |H1|--------------------------|R1|--------------------------|H2|
+# ----            N1            ----            N2            ----
+#
+# ICMP extensions are correctly reported. The loopback interfaces on all the
+# nodes are assigned global addresses and the interfaces connecting the nodes
+# are assigned IPv6 link-local addresses.
+
+cleanup_traceroute6_ext()
+{
+	cleanup_all_ns
+}
+
+setup_traceroute6_ext()
+{
+	# Start clean
+	cleanup_traceroute6_ext
+
+	setup_ns h1 r1 h2
+	create_ns "$h1"
+	create_ns "$r1"
+	create_ns "$h2"
+
+	# Setup N1
+	connect_ns "$h1" eth1 - fe80::1/64 "$r1" eth1 - fe80::2/64
+	# Setup N2
+	connect_ns "$r1" eth2 - fe80::3/64 "$h2" eth2 - fe80::4/64
+
+	# Setup H1
+	ip -n "$h1" address add 2001:db8:1::1/128 dev lo
+	ip -n "$h1" route add ::/0 nexthop via fe80::2 dev eth1
+
+	# Setup R1
+	ip -n "$r1" address add 2001:db8:1::2/128 dev lo
+	ip -n "$r1" route add 2001:db8:1::1/128 nexthop via fe80::1 dev eth1
+	ip -n "$r1" route add 2001:db8:1::3/128 nexthop via fe80::4 dev eth2
+
+	# Setup H2
+	ip -n "$h2" address add 2001:db8:1::3/128 dev lo
+	ip -n "$h2" route add ::/0 nexthop via fe80::3 dev eth2
+
+	# Prime the network
+	ip netns exec "$h1" ping6 -c5 2001:db8:1::3 >/dev/null 2>&1
+}
+
+traceroute6_ext_iio_iif_test()
+{
+	local r1_ifindex h2_ifindex
+	local pkt_len=$1; shift
+
+	# Test that incoming interface info is not appended by default.
+	run_cmd "$h1" "traceroute6 -e 2001:db8:1::3 $pkt_len | grep INC"
+	check_fail $? "Incoming interface info appended by default when should not"
+
+	# Test that the extension is appended when enabled.
+	run_cmd "$r1" "bash -c \"echo 0x01 > /proc/sys/net/ipv6/icmp/errors_extension_mask\""
+	check_err $? "Failed to enable incoming interface info extension on R1"
+
+	run_cmd "$h1" "traceroute6 -e 2001:db8:1::3 $pkt_len | grep INC"
+	check_err $? "Incoming interface info not appended after enable"
+
+	# Test that the extension is not appended when disabled.
+	run_cmd "$r1" "bash -c \"echo 0x00 > /proc/sys/net/ipv6/icmp/errors_extension_mask\""
+	check_err $? "Failed to disable incoming interface info extension on R1"
+
+	run_cmd "$h1" "traceroute6 -e 2001:db8:1::3 $pkt_len | grep INC"
+	check_fail $? "Incoming interface info appended after disable"
+
+	# Test that the extension is sent correctly from both R1 and H2.
+	run_cmd "$r1" "sysctl -w net.ipv6.icmp.errors_extension_mask=0x01"
+	r1_ifindex=$(ip -n "$r1" -j link show dev eth1 | jq '.[]["ifindex"]')
+	run_cmd "$h1" "traceroute6 -e 2001:db8:1::3 $pkt_len | grep '<INC:$r1_ifindex,\"eth1\",mtu=1500>'"
+	check_err $? "Wrong incoming interface info reported from R1"
+
+	run_cmd "$h2" "sysctl -w net.ipv6.icmp.errors_extension_mask=0x01"
+	h2_ifindex=$(ip -n "$h2" -j link show dev eth2 | jq '.[]["ifindex"]')
+	run_cmd "$h1" "traceroute6 -e 2001:db8:1::3 $pkt_len | grep '<INC:$h2_ifindex,\"eth2\",mtu=1500>'"
+	check_err $? "Wrong incoming interface info reported from H2"
+
+	# Add a global address on the incoming interface of R1 and check that
+	# it is reported.
+	run_cmd "$r1" "ip address add 2001:db8:100::1/64 dev eth1 nodad"
+	run_cmd "$h1" "traceroute6 -e 2001:db8:1::3 $pkt_len | grep '<INC:$r1_ifindex,2001:db8:100::1,\"eth1\",mtu=1500>'"
+	check_err $? "Wrong incoming interface info reported from R1 after address addition"
+	run_cmd "$r1" "ip address del 2001:db8:100::1/64 dev eth1"
+
+	# Change name and MTU and make sure the result is still correct.
+	run_cmd "$r1" "ip link set dev eth1 name eth1tag mtu 1501"
+	run_cmd "$h1" "traceroute6 -e 2001:db8:1::3 $pkt_len | grep '<INC:$r1_ifindex,\"eth1tag\",mtu=1501>'"
+	check_err $? "Wrong incoming interface info reported from R1 after name and MTU change"
+	run_cmd "$r1" "ip link set dev eth1tag name eth1 mtu 1500"
+
+	run_cmd "$r1" "sysctl -w net.ipv6.icmp.errors_extension_mask=0x00"
+	run_cmd "$h2" "sysctl -w net.ipv6.icmp.errors_extension_mask=0x00"
+}
+
+run_traceroute6_ext()
+{
+	# Need at least version 2.1.5 for RFC 5837 support.
+	if ! check_traceroute6_version 2.1.5; then
+		log_test_skip "traceroute6 too old, missing ICMP extensions support"
+		return
+	fi
+
+	setup_traceroute6_ext
+
+	RET=0
+
+	## General ICMP extensions tests
+
+	# Test that ICMP extensions are disabled by default.
+	run_cmd "$h1" "sysctl net.ipv6.icmp.errors_extension_mask | grep \"= 0$\""
+	check_err $? "ICMP extensions are not disabled by default"
+
+	# Test that unsupported values are rejected. Do not use "sysctl" as
+	# older versions do not return an error code upon failure.
+	run_cmd "$h1" "bash -c \"echo 0x80 > /proc/sys/net/ipv6/icmp/errors_extension_mask\""
+	check_fail $? "Unsupported sysctl value was not rejected"
+
+	## Extension-specific tests
+
+	# Incoming interface info test. Test with various packet sizes,
+	# including the default one.
+	traceroute6_ext_iio_iif_test
+	traceroute6_ext_iio_iif_test 127
+	traceroute6_ext_iio_iif_test 128
+	traceroute6_ext_iio_iif_test 129
+
+	log_test "IPv6 traceroute with ICMP extensions"
+
+	cleanup_traceroute6_ext
+}
+
+################################################################################
 # traceroute test
 #
 # Verify that traceroute from H1 to H2 shows 1.0.3.1 and 1.0.1.1 when
···
 }
 
 ################################################################################
+# traceroute with ICMP extensions test
+#
+# Verify that in this scenario
+#
+# ----                          ----                          ----
+# |H1|--------------------------|R1|--------------------------|H2|
+# ----            N1            ----            N2            ----
+#
+# ICMP extensions are correctly reported. The loopback interfaces on all the
+# nodes are assigned global addresses and the interfaces connecting the nodes
+# are assigned IPv6 link-local addresses.
+
+cleanup_traceroute_ext()
+{
+	cleanup_all_ns
+}
+
+setup_traceroute_ext()
+{
+	# Start clean
+	cleanup_traceroute_ext
+
+	setup_ns h1 r1 h2
+	create_ns "$h1"
+	create_ns "$r1"
+	create_ns "$h2"
+
+	# Setup N1
+	connect_ns "$h1" eth1 - fe80::1/64 "$r1" eth1 - fe80::2/64
+	# Setup N2
+	connect_ns "$r1" eth2 - fe80::3/64 "$h2" eth2 - fe80::4/64
+
+	# Setup H1
+	ip -n "$h1" address add 192.0.2.1/32 dev lo
+	ip -n "$h1" route add 0.0.0.0/0 nexthop via inet6 fe80::2 dev eth1
+
+	# Setup R1
+	ip -n "$r1" address add 192.0.2.2/32 dev lo
+	ip -n "$r1" route add 192.0.2.1/32 nexthop via inet6 fe80::1 dev eth1
+	ip -n "$r1" route add 192.0.2.3/32 nexthop via inet6 fe80::4 dev eth2
+
+	# Setup H2
+	ip -n "$h2" address add 192.0.2.3/32 dev lo
+	ip -n "$h2" route add 0.0.0.0/0 nexthop via inet6 fe80::3 dev eth2
+
+	# Prime the network
+	ip netns exec "$h1" ping -c5 192.0.2.3 >/dev/null 2>&1
+}
+
+traceroute_ext_iio_iif_test()
+{
+	local r1_ifindex h2_ifindex
+	local pkt_len=$1; shift
+
+	# Test that incoming interface info is not appended by default.
+	run_cmd "$h1" "traceroute -e 192.0.2.3 $pkt_len | grep INC"
+	check_fail $? "Incoming interface info appended by default when should not"
+
+	# Test that the extension is appended when enabled.
+	run_cmd "$r1" "bash -c \"echo 0x01 > /proc/sys/net/ipv4/icmp_errors_extension_mask\""
+	check_err $? "Failed to enable incoming interface info extension on R1"
+
+	run_cmd "$h1" "traceroute -e 192.0.2.3 $pkt_len | grep INC"
+	check_err $? "Incoming interface info not appended after enable"
+
+	# Test that the extension is not appended when disabled.
+	run_cmd "$r1" "bash -c \"echo 0x00 > /proc/sys/net/ipv4/icmp_errors_extension_mask\""
+	check_err $? "Failed to disable incoming interface info extension on R1"
+
+	run_cmd "$h1" "traceroute -e 192.0.2.3 $pkt_len | grep INC"
+	check_fail $? "Incoming interface info appended after disable"
+
+	# Test that the extension is sent correctly from both R1 and H2.
+	run_cmd "$r1" "sysctl -w net.ipv4.icmp_errors_extension_mask=0x01"
+	r1_ifindex=$(ip -n "$r1" -j link show dev eth1 | jq '.[]["ifindex"]')
+	run_cmd "$h1" "traceroute -e 192.0.2.3 $pkt_len | grep '<INC:$r1_ifindex,\"eth1\",mtu=1500>'"
+	check_err $? "Wrong incoming interface info reported from R1"
+
+	run_cmd "$h2" "sysctl -w net.ipv4.icmp_errors_extension_mask=0x01"
+	h2_ifindex=$(ip -n "$h2" -j link show dev eth2 | jq '.[]["ifindex"]')
+	run_cmd "$h1" "traceroute -e 192.0.2.3 $pkt_len | grep '<INC:$h2_ifindex,\"eth2\",mtu=1500>'"
+	check_err $? "Wrong incoming interface info reported from H2"
+
+	# Add a global address on the incoming interface of R1 and check that
+	# it is reported.
+	run_cmd "$r1" "ip address add 198.51.100.1/24 dev eth1"
+	run_cmd "$h1" "traceroute -e 192.0.2.3 $pkt_len | grep '<INC:$r1_ifindex,198.51.100.1,\"eth1\",mtu=1500>'"
+	check_err $? "Wrong incoming interface info reported from R1 after address addition"
+	run_cmd "$r1" "ip address del 198.51.100.1/24 dev eth1"
+
+	# Change name and MTU and make sure the result is still correct.
+	# Re-add the route towards H1 since it was deleted when we removed the
+	# last IPv4 address from eth1 on R1.
+	run_cmd "$r1" "ip route add 192.0.2.1/32 nexthop via inet6 fe80::1 dev eth1"
+	run_cmd "$r1" "ip link set dev eth1 name eth1tag mtu 1501"
+	run_cmd "$h1" "traceroute -e 192.0.2.3 $pkt_len | grep '<INC:$r1_ifindex,\"eth1tag\",mtu=1501>'"
+	check_err $? "Wrong incoming interface info reported from R1 after name and MTU change"
+	run_cmd "$r1" "ip link set dev eth1tag name eth1 mtu 1500"
+
+	run_cmd "$r1" "sysctl -w net.ipv4.icmp_errors_extension_mask=0x00"
+	run_cmd "$h2" "sysctl -w net.ipv4.icmp_errors_extension_mask=0x00"
+}
+
+run_traceroute_ext()
+{
+	# Need at least version 2.1.5 for RFC 5837 support.
+	if ! check_traceroute_version 2.1.5; then
+		log_test_skip "traceroute too old, missing ICMP extensions support"
+		return
+	fi
+
+	setup_traceroute_ext
+
+	RET=0
+
+	## General ICMP extensions tests
+
+	# Test that ICMP extensions are disabled by default.
+	run_cmd "$h1" "sysctl net.ipv4.icmp_errors_extension_mask | grep \"= 0$\""
+	check_err $? "ICMP extensions are not disabled by default"
+
+	# Test that unsupported values are rejected. Do not use "sysctl" as
+	# older versions do not return an error code upon failure.
+	run_cmd "$h1" "bash -c \"echo 0x80 > /proc/sys/net/ipv4/icmp_errors_extension_mask\""
+	check_fail $? "Unsupported sysctl value was not rejected"
+
+	## Extension-specific tests
+
+	# Incoming interface info test. Test with various packet sizes,
+	# including the default one.
+	traceroute_ext_iio_iif_test
+	traceroute_ext_iio_iif_test 127
+	traceroute_ext_iio_iif_test 128
+	traceroute_ext_iio_iif_test 129
+
+	log_test "IPv4 traceroute with ICMP extensions"
+
+	cleanup_traceroute_ext
+}
+
+################################################################################
 # Run tests
 
 run_tests()
 {
 	run_traceroute6
 	run_traceroute6_vrf
+	run_traceroute6_ext
 	run_traceroute
 	run_traceroute_vrf
+	run_traceroute_ext
 }
 
 ################################################################################
···
 
 require_command traceroute6
 require_command traceroute
+require_command jq
 
 run_tests