inet: add RCU protection to inet->opt

We lack proper synchronization to manipulate inet->opt ip_options

Problem is ip_make_skb() calls ip_setup_cork() and
ip_setup_cork() possibly makes a copy of ipc->opt (struct ip_options),
without any protection against another thread manipulating inet->opt.

Another thread can change inet->opt pointer and free old one under us.

Use RCU to protect inet->opt (changed to inet->inet_opt).

Instead of handling atomic refcounts, just copy ip_options when
necessary, to avoid cache line dirtying.

We cant insert an rcu_head in struct ip_options since its included in
skb->cb[], so this patch is large because I had to introduce a new
ip_options_rcu structure.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 24 insertions, 11 deletions

diff --git a/net/ipv4/ip_sockglue.c b/net/ipv4/ip_sockglue.c
index 9640900309bb..ab0c9efd1efa 100644
--- a/net/ipv4/ip_sockglue.c
+++ b/net/ipv4/ip_sockglue.c
@@ -451,6 +451,11 @@ out:
 }
 
 
+static void opt_kfree_rcu(struct rcu_head *head)
+{
+	kfree(container_of(head, struct ip_options_rcu, rcu));
+}
+
 /*
  *	Socket option code for IP. This is the end of the line after any
  *	TCP,UDP etc options on an IP socket.
@@ -497,13 +502,16 @@ static int do_ip_setsockopt(struct sock *sk, int level,
 	switch (optname) {
 	case IP_OPTIONS:
 	{
-		struct ip_options *opt = NULL;
+		struct ip_options_rcu *old, *opt = NULL;
+
 		if (optlen > 40)
 			goto e_inval;
 		err = ip_options_get_from_user(sock_net(sk), &opt,
 					       optval, optlen);
 		if (err)
 			break;
+		old = rcu_dereference_protected(inet->inet_opt,
+						sock_owned_by_user(sk));
 		if (inet->is_icsk) {
 			struct inet_connection_sock *icsk = inet_csk(sk);
 #if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
@@ -512,17 +520,18 @@ static int do_ip_setsockopt(struct sock *sk, int level,
 			       (TCPF_LISTEN | TCPF_CLOSE)) &&
 			     inet->inet_daddr != LOOPBACK4_IPV6)) {
 #endif
-				if (inet->opt)
-					icsk->icsk_ext_hdr_len -= inet->opt->optlen;
+				if (old)
+					icsk->icsk_ext_hdr_len -= old->opt.optlen;
 				if (opt)
-					icsk->icsk_ext_hdr_len += opt->optlen;
+					icsk->icsk_ext_hdr_len += opt->opt.optlen;
 				icsk->icsk_sync_mss(sk, icsk->icsk_pmtu_cookie);
 #if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
 			}
 #endif
 		}
-		opt = xchg(&inet->opt, opt);
-		kfree(opt);
+		rcu_assign_pointer(inet->inet_opt, opt);
+		if (old)
+			call_rcu(&old->rcu, opt_kfree_rcu);
 		break;
 	}
 	case IP_PKTINFO:
@@ -1081,12 +1090,16 @@ static int do_ip_getsockopt(struct sock *sk, int level, int optname,
 	case IP_OPTIONS:
 	{
 		unsigned char optbuf[sizeof(struct ip_options)+40];
-		struct ip_options * opt = (struct ip_options *)optbuf;
+		struct ip_options *opt = (struct ip_options *)optbuf;
+		struct ip_options_rcu *inet_opt;
+
+		inet_opt = rcu_dereference_protected(inet->inet_opt,
+						     sock_owned_by_user(sk));
 		opt->optlen = 0;
-		if (inet->opt)
-			memcpy(optbuf, inet->opt,
-			       sizeof(struct ip_options)+
-			       inet->opt->optlen);
+		if (inet_opt)
+			memcpy(optbuf, &inet_opt->opt,
+			       sizeof(struct ip_options) +
+			       inet_opt->opt.optlen);
 		release_sock(sk);
 
 		if (opt->optlen == 0)