icmp: support rfc 4884

Add setsockopt SOL_IP/IP_RECVERR_4884 to return the offset to an
extension struct if present.

ICMP messages may include an extension structure after the original
datagram. RFC 4884 standardized this behavior. It stores the offset
in words to the extension header in u8 icmphdr.un.reserved[1].

The field is valid only for ICMP types destination unreachable, time
exceeded and parameter problem, if length is at least 128 bytes and
entire packet does not exceed 576 bytes.

Return the offset to the start of the extension struct when reading an
ICMP error from the error queue, if it matches the above constraints.

Do not return the raw u8 field. Return the offset from the start of
the user buffer, in bytes. The kernel does not return the network and
transport headers, so subtract those.

Also validate the headers. Return the offset regardless of validation,
as an invalid extension must still not be misinterpreted as part of
the original datagram. Note that !invalid does not imply valid. If
the extension version does not match, no validation can take place,
for instance.

For backward compatibility, make this optional, set by setsockopt
SOL_IP/IP_RECVERR_RFC4884. For API example and feature test, see
github.com/wdebruij/kerneltools/blob/master/tests/recv_icmp_v2.c

For forward compatibility, reserve only setsockopt value 1, leaving
other bits for additional icmp extensions.

Changes
v1->v2:
- convert word offset to byte offset from start of user buffer
- return in ee_data as u8 may be insufficient
- define extension struct and object header structs
- return len only if constraints met
- if returning len, also validate

Signed-off-by: Willem de Bruijn <[e-mail]>
Signed-off-by: David S. Miller <[e-mail]>

include/linux/icmp.h

···
 
 #include <linux/skbuff.h>
 #include <uapi/linux/icmp.h>
+#include <uapi/linux/errqueue.h>
 
 static inline struct icmphdr *icmp_hdr(const struct sk_buff *skb)
 {
···
 
 	return false;
 }
+
+void ip_icmp_error_rfc4884(const struct sk_buff *skb,
+			   struct sock_ee_data_rfc4884 *out);
 
 #endif	/* _LINUX_ICMP_H */

include/net/inet_sock.h

···
 				mc_all:1,
 				nodefrag:1;
 	__u8			bind_address_no_port:1,
+				recverr_rfc4884:1,
 				defer_connect:1; /* Indicates that fastopen_connect is set
 						  * and cookie exists so we defer connect
 						  * until first data frame is written

include/uapi/linux/errqueue.h

···
 #include <linux/types.h>
 #include <linux/time_types.h>
 
+/* RFC 4884: return offset to extension struct + validation */
+struct sock_ee_data_rfc4884 {
+	__u16	len;
+	__u8	flags;
+	__u8	reserved;
+};
+
 struct sock_extended_err {
 	__u32	ee_errno;	
 	__u8	ee_origin;
···
 	__u8	ee_code;
 	__u8	ee_pad;
 	__u32   ee_info;
-	__u32   ee_data;
+	union	{
+		__u32   ee_data;
+		struct sock_ee_data_rfc4884 ee_rfc4884;
+	};
 };
 
 #define SO_EE_ORIGIN_NONE	0
···
 
 #define SO_EE_CODE_TXTIME_INVALID_PARAM	1
 #define SO_EE_CODE_TXTIME_MISSED	2
+
+#define SO_EE_RFC4884_FLAG_INVALID	1
 
 /**
  *	struct scm_timestamping - timestamps exposed through cmsg

include/uapi/linux/icmp.h

···
 #define _UAPI_LINUX_ICMP_H
 
 #include <linux/types.h>
+#include <asm/byteorder.h>
 
 #define ICMP_ECHOREPLY		0	/* Echo Reply			*/
 #define ICMP_DEST_UNREACH	3	/* Destination Unreachable	*/
···
 	__u32		data;
 };
 
+/* RFC 4884 extension struct: one per message */
+struct icmp_ext_hdr {
+#if defined(__LITTLE_ENDIAN_BITFIELD)
+	__u8		reserved1:4,
+			version:4;
+#elif defined(__BIG_ENDIAN_BITFIELD)
+	__u8		version:4,
+			reserved1:4;
+#else
+#error	"Please fix <asm/byteorder.h>"
+#endif
+	__u8		reserved2;
+	__sum16		checksum;
+};
+
+/* RFC 4884 extension object header: one for each object */
+struct icmp_extobj_hdr {
+	__be16		length;
+	__u8		class_num;
+	__u8		class_type;
+};
 
 #endif /* _UAPI_LINUX_ICMP_H */

include/uapi/linux/in.h

···
 #define IP_CHECKSUM	23
 #define IP_BIND_ADDRESS_NO_PORT	24
 #define IP_RECVFRAGSIZE	25
+#define IP_RECVERR_RFC4884	26
 
 /* IP_MTU_DISCOVER values */
 #define IP_PMTUDISC_DONT		0	/* Never send DF frames */

net/ipv4/icmp.c

···
 	goto drop;
 }
 
+static bool ip_icmp_error_rfc4884_validate(const struct sk_buff *skb, int off)
+{
+	struct icmp_extobj_hdr *objh, _objh;
+	struct icmp_ext_hdr *exth, _exth;
+	u16 olen;
+
+	exth = skb_header_pointer(skb, off, sizeof(_exth), &_exth);
+	if (!exth)
+		return false;
+	if (exth->version != 2)
+		return true;
+
+	if (exth->checksum &&
+	    csum_fold(skb_checksum(skb, off, skb->len - off, 0)))
+		return false;
+
+	off += sizeof(_exth);
+	while (off < skb->len) {
+		objh = skb_header_pointer(skb, off, sizeof(_objh), &_objh);
+		if (!objh)
+			return false;
+
+		olen = ntohs(objh->length);
+		if (olen < sizeof(_objh))
+			return false;
+
+		off += olen;
+		if (off > skb->len)
+			return false;
+	}
+
+	return true;
+}
+
+void ip_icmp_error_rfc4884(const struct sk_buff *skb,
+			   struct sock_ee_data_rfc4884 *out)
+{
+	int hlen, off;
+
+	switch (icmp_hdr(skb)->type) {
+	case ICMP_DEST_UNREACH:
+	case ICMP_TIME_EXCEEDED:
+	case ICMP_PARAMETERPROB:
+		break;
+	default:
+		return;
+	}
+
+	/* outer headers up to inner iph. skb->data is at inner payload */
+	hlen = -skb_transport_offset(skb) - sizeof(struct icmphdr);
+
+	/* per rfc 791: maximum packet length of 576 bytes */
+	if (hlen + skb->len > 576)
+		return;
+
+	/* per rfc 4884: minimal datagram length of 128 bytes */
+	off = icmp_hdr(skb)->un.reserved[1] * sizeof(u32);
+	if (off < 128)
+		return;
+
+	/* kernel has stripped headers: return payload offset in bytes */
+	off -= hlen;
+	if (off + sizeof(struct icmp_ext_hdr) > skb->len)
+		return;
+
+	out->len = off;
+
+	if (!ip_icmp_error_rfc4884_validate(skb, off))
+		out->flags |= SO_EE_RFC4884_FLAG_INVALID;
+}
+
 int icmp_err(struct sk_buff *skb, u32 info)
 {
 	struct iphdr *iph = (struct iphdr *)skb->data;

net/ipv4/ip_sockglue.c

···
 	serr->port = port;
 
 	if (skb_pull(skb, payload - skb->data)) {
+		if (inet_sk(sk)->recverr_rfc4884)
+			ip_icmp_error_rfc4884(skb, &serr->ee.ee_rfc4884);
+
 		skb_reset_transport_header(skb);
 		if (sock_queue_err_skb(sk, skb) == 0)
 			return;
···
 	case IP_RECVORIGDSTADDR:
 	case IP_CHECKSUM:
 	case IP_RECVFRAGSIZE:
+	case IP_RECVERR_RFC4884:
 		if (optlen >= sizeof(int)) {
 			if (get_user(val, (int __user *) optval))
 				return -EFAULT;
···
 		inet->recverr = !!val;
 		if (!val)
 			skb_queue_purge(&sk->sk_error_queue);
+		break;
+	case IP_RECVERR_RFC4884:
+		if (val < 0 || val > 1)
+			goto e_inval;
+		inet->recverr_rfc4884 = !!val;
 		break;
 	case IP_MULTICAST_TTL:
 		if (sk->sk_type == SOCK_STREAM)
···
 	}
 	case IP_RECVERR:
 		val = inet->recverr;
+		break;
+	case IP_RECVERR_RFC4884:
+		val = inet->recverr_rfc4884;
 		break;
 	case IP_MULTICAST_TTL:
 		val = inet->mc_ttl;