Skip to contentSkip to conversation

ipv4: introduce new IP_MTU_DISCOVER mode IP_PMTUDISC_INTERFACE

PATCH In review Hannes Frederic Sowa 2 participants 3 messages Nov 5, 2013 – Nov 6, 2013

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ipv4: introduce new IP_MTU_DISCOVER mode IP_PMTUDISC_INTERFACE

Sockets marked with IP_PMTUDISC_INTERFACE won't do path mtu discovery, their sockets won't accept and install new path mtu information and they will always use the interface mtu for outgoing packets. It is guaranteed that the packet is not fragmented locally. But we won't set the DF-Flag on the outgoing frames.

Florian Weimer had the idea to use this flag to ensure DNS servers are never generating outgoing fragments. They may well be fragmented on the path, but the server never stores or usees path mtu values, which could well be forged in an attack.

(The root of the problem with path MTU discovery is that there is no reliable way to authenticate ICMP Fragmentation Needed But DF Set messages because they are sent from intermediate routers with their source addresses, and the IMCP payload will not always contain sufficient information to identify a flow.)

Recent research in the DNS community showed that it is possible to implement an attack where DNS cache poisoning is feasible by spoofing fragments. This work was done by Amir Herzberg and Haya Shulman:
<https://sites.google.com/site/hayashulman/files/fragmentation-poisoning.pdf>

This issue was previously discussed among the DNS community, e.g.
<http://www.ietf.org/mail-archive/web/dnsext/current/msg01204.html>,
without leading to fixes.

This patch depends on the patch "ipv4: fix DO and PROBE pmtu mode regarding local fragmentation with UFO/CORK" for the enforcement of the non-fragmentable checks. If other users than ip_append_page/data should use this semantic too, we have to add a new flag to IPCB(skb)->flags to suppress local fragmentation and check for this in ip_finish_output.

Many thanks to Florian Weimer for the idea and feedback while implementing this patch.

---
 include/net/route.h     | 16 ++++++++++++----
 include/uapi/linux/in.h |  5 +++++
 net/dccp/ipv4.c         |  1 +
 net/ipv4/ip_output.c    |  8 ++++----
 net/ipv4/ip_sockglue.c  |  2 +-
 net/ipv4/route.c        |  4 ++++
 net/ipv4/tcp_ipv4.c     |  1 +
 7 files changed, 28 insertions(+), 9 deletions(-)
include/net/route.h +12 -4
index dd4ae00..f68c167 100644
@@ -313,12 +313,20 @@ static inline int ip4_dst_hoplimit(const struct dst_entry *dst)
 	return hoplimit;
 }
 
-static inline int ip_skb_dst_mtu(struct sk_buff *skb)
+static inline bool ip_sk_accept_pmtu(const struct sock *sk)
 {
-	struct inet_sock *inet = skb->sk ? inet_sk(skb->sk) : NULL;
+	return inet_sk(sk)->pmtudisc != IP_PMTUDISC_INTERFACE;
+}
 
-	return (inet && inet->pmtudisc == IP_PMTUDISC_PROBE) ?
-	       skb_dst(skb)->dev->mtu : dst_mtu(skb_dst(skb));
+static inline bool ip_sk_use_pmtu(const struct sock *sk)
+{
+	return inet_sk(sk)->pmtudisc < IP_PMTUDISC_PROBE;
+}
+
+static inline int ip_skb_dst_mtu(const struct sk_buff *skb)
+{
+	return (!skb->sk || ip_sk_use_pmtu(skb->sk)) ?
+	       dst_mtu(skb_dst(skb)) : skb_dst(skb)->dev->mtu;
 }
 
 #endif	/* _ROUTE_H */
include/uapi/linux/in.h +5 -0
index f9e8e49..393c5de 100644
@@ -115,6 +115,11 @@ struct in_addr {
 #define IP_PMTUDISC_WANT		1	/* Use per route hints	*/
 #define IP_PMTUDISC_DO			2	/* Always DF		*/
 #define IP_PMTUDISC_PROBE		3       /* Ignore dst pmtu      */
+/* Always use interface mtu (ignores dst pmtu) but don't set DF flag.
+ * Also incoming ICMP frag_needed notifications will be ignored on
+ * this socket to prevent accepting spoofed ones.
+ */
+#define IP_PMTUDISC_INTERFACE		4
 
 #define IP_MULTICAST_IF			32
 #define IP_MULTICAST_TTL 		33
net/dccp/ipv4.c +1 -0
index 720c362..d9f65fc 100644
@@ -174,6 +174,7 @@ static inline void dccp_do_pmtu_discovery(struct sock *sk,
 	mtu = dst_mtu(dst);
 
 	if (inet->pmtudisc != IP_PMTUDISC_DONT &&
+	    ip_sk_accept_pmtu(sk) &&
 	    inet_csk(sk)->icsk_pmtu_cookie > mtu) {
 		dccp_sync_mss(sk, mtu);
 
net/ipv4/ip_output.c +4 -4
index 51be64e..9124027 100644
@@ -1037,7 +1037,6 @@ error:
 static int ip_setup_cork(struct sock *sk, struct inet_cork *cork,
 			 struct ipcm_cookie *ipc, struct rtable **rtp)
 {
-	struct inet_sock *inet = inet_sk(sk);
 	struct ip_options_rcu *opt;
 	struct rtable *rt;
 
@@ -1063,8 +1062,8 @@ static int ip_setup_cork(struct sock *sk, struct inet_cork *cork,
 	 * We steal reference to this route, caller should not release it
 	 */
 	*rtp = NULL;
-	cork->fragsize = inet->pmtudisc == IP_PMTUDISC_PROBE ?
-			 rt->dst.dev->mtu : dst_mtu(&rt->dst);
+	cork->fragsize = ip_sk_use_pmtu(sk) ?
+			 dst_mtu(&rt->dst) : rt->dst.dev->mtu;
 	cork->dst = &rt->dst;
 	cork->length = 0;
 	cork->ttl = ipc->ttl;
@@ -1315,7 +1314,8 @@ struct sk_buff *__ip_make_skb(struct sock *sk,
 	/* DF bit is set when we want to see DF on outgoing frames.
 	 * If local_df is set too, we still allow to fragment this frame
 	 * locally. */
-	if (inet->pmtudisc >= IP_PMTUDISC_DO ||
+	if (inet->pmtudisc == IP_PMTUDISC_DO ||
+	    inet->pmtudisc == IP_PMTUDISC_PROBE ||
 	    (skb->len <= dst_mtu(&rt->dst) &&
 	     ip_dont_fragment(sk, &rt->dst)))
 		df = htons(IP_DF);
net/ipv4/ip_sockglue.c +1 -1
index 0626f2c..3f85826 100644
@@ -627,7 +627,7 @@ static int do_ip_setsockopt(struct sock *sk, int level,
 		inet->nodefrag = val ? 1 : 0;
 		break;
 	case IP_MTU_DISCOVER:
-		if (val < IP_PMTUDISC_DONT || val > IP_PMTUDISC_PROBE)
+		if (val < IP_PMTUDISC_DONT || val > IP_PMTUDISC_INTERFACE)
 			goto e_inval;
 		inet->pmtudisc = val;
 		break;
net/ipv4/route.c +4 -0
index d2d3253..f428935 100644
@@ -1036,6 +1036,10 @@ void ipv4_sk_update_pmtu(struct sk_buff *skb, struct sock *sk, u32 mtu)
 	bool new = false;
 
 	bh_lock_sock(sk);
+
+	if (!ip_sk_accept_pmtu(sk))
+		goto out;
+
 	rt = (struct rtable *) __sk_dst_get(sk);
 
 	if (sock_owned_by_user(sk) || !rt) {
net/ipv4/tcp_ipv4.c +1 -0
index 300ab2c..14bba8a 100644
@@ -288,6 +288,7 @@ static void tcp_v4_mtu_reduced(struct sock *sk)
 	mtu = dst_mtu(dst);
 
 	if (inet->pmtudisc != IP_PMTUDISC_DONT &&
+	    ip_sk_accept_pmtu(sk) &&
 	    inet_csk(sk)->icsk_pmtu_cookie > mtu) {
 		tcp_sync_mss(sk, mtu);
 
Signature · 1 line
1.8.3.1
maintainer
#1 lore raw

From: Hannes Frederic Sowa <[email protected]>
Date: Tue, 5 Nov 2013 02:24:17 +0100

Quoted 32 lines

Sockets marked with IP_PMTUDISC_INTERFACE won't do path mtu discovery, their sockets won't accept and install new path mtu information and they will always use the interface mtu for outgoing packets. It is guaranteed that the packet is not fragmented locally. But we won't set the DF-Flag on the outgoing frames.

Florian Weimer had the idea to use this flag to ensure DNS servers are never generating outgoing fragments. They may well be fragmented on the path, but the server never stores or usees path mtu values, which could well be forged in an attack.

(The root of the problem with path MTU discovery is that there is no reliable way to authenticate ICMP Fragmentation Needed But DF Set messages because they are sent from intermediate routers with their source addresses, and the IMCP payload will not always contain sufficient information to identify a flow.)

Recent research in the DNS community showed that it is possible to implement an attack where DNS cache poisoning is feasible by spoofing fragments. This work was done by Amir Herzberg and Haya Shulman:
<https://sites.google.com/site/hayashulman/files/fragmentation-poisoning.pdf>

This issue was previously discussed among the DNS community, e.g.
<http://www.ietf.org/mail-archive/web/dnsext/current/msg01204.html>,
without leading to fixes.

This patch depends on the patch "ipv4: fix DO and PROBE pmtu mode regarding local fragmentation with UFO/CORK" for the enforcement of the non-fragmentable checks. If other users than ip_append_page/data should use this semantic too, we have to add a new flag to IPCB(skb)->flags to suppress local fragmentation and check for this in ip_finish_output.

Quoted 6 lines

Many thanks to Florian Weimer for the idea and feedback while implementing this patch.

Ok I changed my mind and decided to apply this, thanks.

BTW, what about IPV6? Even if ipv6 doesn't need this facility, for whatever reason, these IP_PMTUDISC_* values are shared with ipv6 (via the IPV6_MTU_DISCOVER socket option) so we should at least make sure that we do something reasonable if the new value happens to get passed in.

#2 lore raw
Quoted 42 lines · David Miller

From: Hannes Frederic Sowa <[email protected]>
Date: Tue, 5 Nov 2013 02:24:17 +0100

Quoted 32 lines

Sockets marked with IP_PMTUDISC_INTERFACE won't do path mtu discovery, their sockets won't accept and install new path mtu information and they will always use the interface mtu for outgoing packets. It is guaranteed that the packet is not fragmented locally. But we won't set the DF-Flag on the outgoing frames.

Florian Weimer had the idea to use this flag to ensure DNS servers are never generating outgoing fragments. They may well be fragmented on the path, but the server never stores or usees path mtu values, which could well be forged in an attack.

(The root of the problem with path MTU discovery is that there is no reliable way to authenticate ICMP Fragmentation Needed But DF Set messages because they are sent from intermediate routers with their source addresses, and the IMCP payload will not always contain sufficient information to identify a flow.)

Recent research in the DNS community showed that it is possible to implement an attack where DNS cache poisoning is feasible by spoofing fragments. This work was done by Amir Herzberg and Haya Shulman:
<https://sites.google.com/site/hayashulman/files/fragmentation-poisoning.pdf>

This issue was previously discussed among the DNS community, e.g.
<http://www.ietf.org/mail-archive/web/dnsext/current/msg01204.html>,
without leading to fixes.

This patch depends on the patch "ipv4: fix DO and PROBE pmtu mode regarding local fragmentation with UFO/CORK" for the enforcement of the non-fragmentable checks. If other users than ip_append_page/data should use this semantic too, we have to add a new flag to IPCB(skb)->flags to suppress local fragmentation and check for this in ip_finish_output.

Quoted 6 lines

Many thanks to Florian Weimer for the idea and feedback while implementing this patch.

Quoted 7 lines

Ok I changed my mind and decided to apply this, thanks.

BTW, what about IPV6? Even if ipv6 doesn't need this facility, for whatever reason, these IP_PMTUDISC_* values are shared with ipv6 (via the IPV6_MTU_DISCOVER socket option) so we should at least make sure that we do something reasonable if the new value happens to get passed in.

Of course I will look after IPv6. ;)

IPPROTO_IP setsockopt level does change the value of inet_sk->pmtudisc. The IPPROTO_IPV6 would change the pmtudisc in ipv6_pinfo. So we currently don't accept this option on IPv6 sockets (-EINVAL) and when applied with IPPROTO_IP level it will change the sockets behaviour for IPv4 mapped sockets only. So nothing will break with this change.

It should be easy to do and I plan to send the changes this week. pmtudisc in ipv6_pinfo has only 2 bits so I will have to increase the socket size a bit.

Thank you,

Hannes