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, 20 insertions, 14 deletions

diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 310454c..d60732f 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -154,6 +154,7 @@ int tcp_v4_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)
 	struct flowi4 fl4;
 	struct rtable *rt;
 	int err;
+	struct ip_options_rcu *inet_opt;
 
 	if (addr_len < sizeof(struct sockaddr_in))
 		return -EINVAL;
@@ -162,10 +163,12 @@ int tcp_v4_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)
 		return -EAFNOSUPPORT;
 
 	nexthop = daddr = usin->sin_addr.s_addr;
-	if (inet->opt && inet->opt->srr) {
+	inet_opt = rcu_dereference_protected(inet->inet_opt,
+					     sock_owned_by_user(sk));
+	if (inet_opt && inet_opt->opt.srr) {
 		if (!daddr)
 			return -EINVAL;
-		nexthop = inet->opt->faddr;
+		nexthop = inet_opt->opt.faddr;
 	}
 
 	orig_sport = inet->inet_sport;
@@ -186,7 +189,7 @@ int tcp_v4_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)
 		return -ENETUNREACH;
 	}
 
-	if (!inet->opt || !inet->opt->srr)
+	if (!inet_opt || !inet_opt->opt.srr)
 		daddr = rt->rt_dst;
 
 	if (!inet->inet_saddr)
@@ -222,8 +225,8 @@ int tcp_v4_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)
 	inet->inet_daddr = daddr;
 
 	inet_csk(sk)->icsk_ext_hdr_len = 0;
-	if (inet->opt)
-		inet_csk(sk)->icsk_ext_hdr_len = inet->opt->optlen;
+	if (inet_opt)
+		inet_csk(sk)->icsk_ext_hdr_len = inet_opt->opt.optlen;
 
 	tp->rx_opt.mss_clamp = TCP_MSS_DEFAULT;
 
@@ -820,17 +823,18 @@ static void syn_flood_warning(const struct sk_buff *skb)
 /*
  * Save and compile IPv4 options into the request_sock if needed.
  */
-static struct ip_options *tcp_v4_save_options(struct sock *sk,
-					      struct sk_buff *skb)
+static struct ip_options_rcu *tcp_v4_save_options(struct sock *sk,
+						  struct sk_buff *skb)
 {
-	struct ip_options *opt = &(IPCB(skb)->opt);
-	struct ip_options *dopt = NULL;
+	const struct ip_options *opt = &(IPCB(skb)->opt);
+	struct ip_options_rcu *dopt = NULL;
 
 	if (opt && opt->optlen) {
-		int opt_size = optlength(opt);
+		int opt_size = sizeof(*dopt) + opt->optlen;
+
 		dopt = kmalloc(opt_size, GFP_ATOMIC);
 		if (dopt) {
-			if (ip_options_echo(dopt, skb)) {
+			if (ip_options_echo(&dopt->opt, skb)) {
 				kfree(dopt);
 				dopt = NULL;
 			}
@@ -1411,6 +1415,7 @@ struct sock *tcp_v4_syn_recv_sock(struct sock *sk, struct sk_buff *skb,
 #ifdef CONFIG_TCP_MD5SIG
 	struct tcp_md5sig_key *key;
 #endif
+	struct ip_options_rcu *inet_opt;
 
 	if (sk_acceptq_is_full(sk))
 		goto exit_overflow;
@@ -1431,13 +1436,14 @@ struct sock *tcp_v4_syn_recv_sock(struct sock *sk, struct sk_buff *skb,
 	newinet->inet_daddr   = ireq->rmt_addr;
 	newinet->inet_rcv_saddr = ireq->loc_addr;
 	newinet->inet_saddr	      = ireq->loc_addr;
-	newinet->opt	      = ireq->opt;
+	inet_opt	      = ireq->opt;
+	rcu_assign_pointer(newinet->inet_opt, inet_opt);
 	ireq->opt	      = NULL;
 	newinet->mc_index     = inet_iif(skb);
 	newinet->mc_ttl	      = ip_hdr(skb)->ttl;
 	inet_csk(newsk)->icsk_ext_hdr_len = 0;
-	if (newinet->opt)
-		inet_csk(newsk)->icsk_ext_hdr_len = newinet->opt->optlen;
+	if (inet_opt)
+		inet_csk(newsk)->icsk_ext_hdr_len = inet_opt->opt.optlen;
 	newinet->inet_id = newtp->write_seq ^ jiffies;
 
 	tcp_mtup_init(newsk);