[TCP]: MD5 Signature Option (RFC2385) support.

Based on implementation by Rick Payne.
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

CREDITS

@@ -2598,6 +2598,9 @@ S: Ucitelska 1576
 S: Prague 8
 S: 182 00 Czech Republic
 
+N: Rick Payne
+D: RFC2385 Support for TCP
+
 N: Barak A. Pearlmutter
 E: bap@cs.unm.edu
 W: http://www.cs.unm.edu/~bap/

include/linux/tcp.h

@@ -19,6 +19,7 @@
 
 #include <linux/types.h>
 #include <asm/byteorder.h>
+#include <linux/socket.h>
 
 struct tcphdr {
 	__be16	source;
@@ -94,6 +95,7 @@ enum {
 #define TCP_INFO		11	/* Information about this connection. */
 #define TCP_QUICKACK		12	/* Block/reenable quick acks */
 #define TCP_CONGESTION		13	/* Congestion control algorithm */
+#define TCP_MD5SIG		14	/* TCP MD5 Signature (RFC2385) */
 
 #define TCPI_OPT_TIMESTAMPS	1
 #define TCPI_OPT_SACK		2
@@ -157,6 +159,17 @@ struct tcp_info
 	__u32	tcpi_total_retrans;
 };
 
+/* for TCP_MD5SIG socket option */
+#define TCP_MD5SIG_MAXKEYLEN	80
+
+struct tcp_md5sig {
+	struct __kernel_sockaddr_storage tcpm_addr;	/* address associated */
+	__u16	__tcpm_pad1;				/* zero */
+	__u16	tcpm_keylen;				/* key length */
+	__u32	__tcpm_pad2;				/* zero */
+	__u8	tcpm_key[TCP_MD5SIG_MAXKEYLEN];		/* key (binary) */
+};
+
 #ifdef __KERNEL__
 
 #include <linux/skbuff.h>
@@ -197,9 +210,13 @@ struct tcp_options_received {
 };
 
 struct tcp_request_sock {
-	struct inet_request_sock req;
-	__u32			 rcv_isn;
-	__u32			 snt_isn;
+	struct inet_request_sock 	req;
+#ifdef CONFIG_TCP_MD5SIG
+	/* Only used by TCP MD5 Signature so far. */
+	struct tcp_request_sock_ops	*af_specific;
+#endif
+	__u32			 	rcv_isn;
+	__u32			 	snt_isn;
 };
 
 static inline struct tcp_request_sock *tcp_rsk(const struct request_sock *req)
@@ -363,6 +380,14 @@ struct tcp_sock {
 		__u32		  probe_seq_start;
 		__u32		  probe_seq_end;
 	} mtu_probe;
+
+#ifdef CONFIG_TCP_MD5SIG
+/* TCP AF-Specific parts; only used by MD5 Signature support so far */
+	struct tcp_sock_af_ops	*af_specific;
+
+/* TCP MD5 Signagure Option information */
+	struct tcp_md5sig_info	*md5sig_info;
+#endif
 };
 
 static inline struct tcp_sock *tcp_sk(const struct sock *sk)
@@ -377,6 +402,10 @@ struct tcp_timewait_sock {
 	__u32			  tw_rcv_wnd;
 	__u32			  tw_ts_recent;
 	long			  tw_ts_recent_stamp;
+#ifdef CONFIG_TCP_MD5SIG
+	__u16			  tw_md5_keylen;
+	__u8			  tw_md5_key[TCP_MD5SIG_MAXKEYLEN];
+#endif
 };
 
 static inline struct tcp_timewait_sock *tcp_twsk(const struct sock *sk)

include/net/request_sock.h

@@ -35,7 +35,8 @@ struct request_sock_ops {
 				       struct dst_entry *dst);
 	void		(*send_ack)(struct sk_buff *skb,
 				    struct request_sock *req);
-	void		(*send_reset)(struct sk_buff *skb);
+	void		(*send_reset)(struct sock *sk,
+				      struct sk_buff *skb);
 	void		(*destructor)(struct request_sock *req);
 };
 

include/net/tcp.h

@@ -28,6 +28,7 @@
 #include <linux/percpu.h>
 #include <linux/skbuff.h>
 #include <linux/dmaengine.h>
+#include <linux/crypto.h>
 
 #include <net/inet_connection_sock.h>
 #include <net/inet_timewait_sock.h>
@@ -161,6 +162,7 @@ extern void tcp_time_wait(struct sock *sk, int state, int timeo);
 #define TCPOPT_SACK_PERM        4       /* SACK Permitted */
 #define TCPOPT_SACK             5       /* SACK Block */
 #define TCPOPT_TIMESTAMP	8	/* Better RTT estimations/PAWS */
+#define TCPOPT_MD5SIG		19	/* MD5 Signature (RFC2385) */
 
 /*
  *     TCP option lengths
@@ -170,6 +172,7 @@ extern void tcp_time_wait(struct sock *sk, int state, int timeo);
 #define TCPOLEN_WINDOW         3
 #define TCPOLEN_SACK_PERM      2
 #define TCPOLEN_TIMESTAMP      10
+#define TCPOLEN_MD5SIG         18
 
 /* But this is what stacks really send out. */
 #define TCPOLEN_TSTAMP_ALIGNED		12
@@ -178,6 +181,7 @@ extern void tcp_time_wait(struct sock *sk, int state, int timeo);
 #define TCPOLEN_SACK_BASE		2
 #define TCPOLEN_SACK_BASE_ALIGNED	4
 #define TCPOLEN_SACK_PERBLOCK		8
+#define TCPOLEN_MD5SIG_ALIGNED		20
 
 /* Flags in tp->nonagle */
 #define TCP_NAGLE_OFF		1	/* Nagle's algo is disabled */
@@ -299,6 +303,8 @@ extern void			tcp_cleanup_rbuf(struct sock *sk, int copied);
 extern int			tcp_twsk_unique(struct sock *sk,
 						struct sock *sktw, void *twp);
 
+extern void			tcp_twsk_destructor(struct sock *sk);
+
 static inline void tcp_dec_quickack_mode(struct sock *sk,
 					 const unsigned int pkts)
 {
@@ -1064,6 +1070,114 @@ static inline void clear_all_retrans_hints(struct tcp_sock *tp){
 	tp->fastpath_skb_hint = NULL;
 }
 
+/* MD5 Signature */
+struct crypto_hash;
+
+/* - key database */
+struct tcp_md5sig_key {
+	u8			*key;
+	u8			keylen;
+};
+
+struct tcp4_md5sig_key {
+	u8			*key;
+	u16			keylen;
+	__be32			addr;
+};
+
+struct tcp6_md5sig_key {
+	u8			*key;
+	u16			keylen;
+#if 0
+	u32			scope_id;	/* XXX */
+#endif
+	struct in6_addr		addr;
+};
+
+/* - sock block */
+struct tcp_md5sig_info {
+	struct tcp4_md5sig_key	*keys4;
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+	struct tcp6_md5sig_key	*keys6;
+	u32			entries6;
+	u32			alloced6;
+#endif
+	u32			entries4;
+	u32			alloced4;
+};
+
+/* - pseudo header */
+struct tcp4_pseudohdr {
+	__be32		saddr;
+	__be32		daddr;
+	__u8		pad;
+	__u8		protocol;
+	__be16		len;
+};
+
+struct tcp6_pseudohdr {
+	struct in6_addr	saddr;
+	struct in6_addr daddr;
+	__be32		len;
+	__be32		protocol;	/* including padding */
+};
+
+union tcp_md5sum_block {
+	struct tcp4_pseudohdr ip4;
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+	struct tcp6_pseudohdr ip6;
+#endif
+};
+
+/* - pool: digest algorithm, hash description and scratch buffer */
+struct tcp_md5sig_pool {
+	struct hash_desc	md5_desc;
+	union tcp_md5sum_block	md5_blk;
+};
+
+#define TCP_MD5SIG_MAXKEYS	(~(u32)0)	/* really?! */
+
+/* - functions */
+extern int			tcp_v4_calc_md5_hash(char *md5_hash,
+						     struct tcp_md5sig_key *key,
+						     struct sock *sk,
+						     struct dst_entry *dst,
+						     struct request_sock *req,
+						     struct tcphdr *th,
+						     int protocol, int tcplen);
+extern struct tcp_md5sig_key	*tcp_v4_md5_lookup(struct sock *sk,
+						   struct sock *addr_sk);
+
+extern int			tcp_v4_md5_do_add(struct sock *sk,
+						  __be32 addr,
+						  u8 *newkey,
+						  u8 newkeylen);
+
+extern int			tcp_v4_md5_do_del(struct sock *sk,
+						  u32 addr);
+
+extern struct tcp_md5sig_pool	**tcp_alloc_md5sig_pool(void);
+extern void			tcp_free_md5sig_pool(void);
+
+extern struct tcp_md5sig_pool	*__tcp_get_md5sig_pool(int cpu);
+extern void			__tcp_put_md5sig_pool(void);
+
+static inline
+struct tcp_md5sig_pool		*tcp_get_md5sig_pool(void)
+{
+	int cpu = get_cpu();
+	struct tcp_md5sig_pool *ret = __tcp_get_md5sig_pool(cpu);
+	if (!ret)
+		put_cpu();
+	return ret;
+}
+
+static inline void		tcp_put_md5sig_pool(void)
+{
+	__tcp_put_md5sig_pool();
+	put_cpu();
+}
+
 /* /proc */
 enum tcp_seq_states {
 	TCP_SEQ_STATE_LISTENING,
@@ -1103,6 +1217,35 @@ extern int  tcp4_proc_init(void);
 extern void tcp4_proc_exit(void);
 #endif
 
+/* TCP af-specific functions */
+struct tcp_sock_af_ops {
+#ifdef CONFIG_TCP_MD5SIG
+	struct tcp_md5sig_key	*(*md5_lookup) (struct sock *sk,
+						struct sock *addr_sk);
+	int			(*calc_md5_hash) (char *location,
+						  struct tcp_md5sig_key *md5,
+						  struct sock *sk,
+						  struct dst_entry *dst,
+						  struct request_sock *req,
+						  struct tcphdr *th,
+						  int protocol, int len);
+	int			(*md5_add) (struct sock *sk,
+					    struct sock *addr_sk,
+					    u8 *newkey,
+					    u8 len);
+	int			(*md5_parse) (struct sock *sk,
+					      char __user *optval,
+					      int optlen);
+#endif
+};
+
+struct tcp_request_sock_ops {
+#ifdef CONFIG_TCP_MD5SIG
+	struct tcp_md5sig_key	*(*md5_lookup) (struct sock *sk,
+						struct request_sock *req);
+#endif
+};
+
 extern void tcp_v4_init(struct net_proto_family *ops);
 extern void tcp_init(void);
 

include/net/timewait_sock.h

@@ -31,6 +31,9 @@ static inline int twsk_unique(struct sock *sk, struct sock *sktw, void *twp)
 
 static inline void twsk_destructor(struct sock *sk)
 {
+	BUG_ON(sk == NULL);
+	BUG_ON(sk->sk_prot == NULL);
+	BUG_ON(sk->sk_prot->twsk_prot == NULL);
 	if (sk->sk_prot->twsk_prot->twsk_destructor != NULL)
 		sk->sk_prot->twsk_prot->twsk_destructor(sk);
 }

net/dccp/ipv4.c

@@ -509,7 +509,7 @@ static int dccp_v4_send_response(struct sock *sk, struct request_sock *req,
 	return err;
 }
 
-static void dccp_v4_ctl_send_reset(struct sk_buff *rxskb)
+static void dccp_v4_ctl_send_reset(struct sock *sk, struct sk_buff *rxskb)
 {
 	int err;
 	struct dccp_hdr *rxdh = dccp_hdr(rxskb), *dh;
@@ -724,7 +724,7 @@ int dccp_v4_do_rcv(struct sock *sk, struct sk_buff *skb)
 	return 0;
 
 reset:
-	dccp_v4_ctl_send_reset(skb);
+	dccp_v4_ctl_send_reset(sk, skb);
 discard:
 	kfree_skb(skb);
 	return 0;
@@ -913,7 +913,7 @@ static int dccp_v4_rcv(struct sk_buff *skb)
 	if (dh->dccph_type != DCCP_PKT_RESET) {
 		DCCP_SKB_CB(skb)->dccpd_reset_code =
 					DCCP_RESET_CODE_NO_CONNECTION;
-		dccp_v4_ctl_send_reset(skb);
+		dccp_v4_ctl_send_reset(sk, skb);
 	}
 
 discard_it:

net/dccp/ipv6.c

@@ -310,7 +310,7 @@ static void dccp_v6_reqsk_destructor(struct request_sock *req)
 		kfree_skb(inet6_rsk(req)->pktopts);
 }
 
-static void dccp_v6_ctl_send_reset(struct sk_buff *rxskb)
+static void dccp_v6_ctl_send_reset(struct sock *sk, struct sk_buff *rxskb)
 {
 	struct dccp_hdr *rxdh = dccp_hdr(rxskb), *dh;
 	const u32 dccp_hdr_reset_len = sizeof(struct dccp_hdr) +
@@ -805,7 +805,7 @@ static int dccp_v6_do_rcv(struct sock *sk, struct sk_buff *skb)
 	return 0;
 
 reset:
-	dccp_v6_ctl_send_reset(skb);
+	dccp_v6_ctl_send_reset(sk, skb);
 discard:
 	if (opt_skb != NULL)
 		__kfree_skb(opt_skb);
@@ -902,7 +902,7 @@ static int dccp_v6_rcv(struct sk_buff **pskb)
 	if (dh->dccph_type != DCCP_PKT_RESET) {
 		DCCP_SKB_CB(skb)->dccpd_reset_code =
 					DCCP_RESET_CODE_NO_CONNECTION;
-		dccp_v6_ctl_send_reset(skb);
+		dccp_v6_ctl_send_reset(sk, skb);
 	}
 
 discard_it:

net/dccp/minisocks.c

@@ -246,7 +246,7 @@ struct sock *dccp_check_req(struct sock *sk, struct sk_buff *skb,
 	DCCP_SKB_CB(skb)->dccpd_reset_code = DCCP_RESET_CODE_TOO_BUSY;
 drop:
 	if (dccp_hdr(skb)->dccph_type != DCCP_PKT_RESET)
-		req->rsk_ops->send_reset(skb);
+		req->rsk_ops->send_reset(sk, skb);
 
 	inet_csk_reqsk_queue_drop(sk, req, prev);
 	goto out;

net/ipv4/Kconfig

@@ -618,5 +618,21 @@ config DEFAULT_TCP_CONG
 	default "reno" if DEFAULT_RENO
 	default "cubic"
 
+config TCP_MD5SIG
+	bool "TCP: MD5 Signature Option support (RFC2385) (EXPERIMENTAL)"
+	depends on EXPERIMENTAL
+	select CRYPTO
+	select CRYPTO_MD5
+	---help---
+	  RFC2385 specifices a method of giving MD5 protection to TCP sessions.
+	  Its main (only?) use is to protect BGP sessions between core routers
+	  on the Internet.
+
+	  If unsure, say N.
+
+config TCP_MD5SIG_DEBUG
+	bool "TCP: MD5 Signature Option debugging"
+	depends on TCP_MD5SIG
+
 source "net/ipv4/ipvs/Kconfig"
 

net/ipv4/tcp.c

@@ -258,6 +258,7 @@
 #include <linux/bootmem.h>
 #include <linux/cache.h>
 #include <linux/err.h>
+#include <linux/crypto.h>
 
 #include <net/icmp.h>
 #include <net/tcp.h>
@@ -1942,6 +1943,13 @@ static int do_tcp_setsockopt(struct sock *sk, int level,
 		}
 		break;
 
+#ifdef CONFIG_TCP_MD5SIG
+	case TCP_MD5SIG:
+		/* Read the IP->Key mappings from userspace */
+		err = tp->af_specific->md5_parse(sk, optval, optlen);
+		break;
+#endif
+
 	default:
 		err = -ENOPROTOOPT;
 		break;
@@ -2231,6 +2239,135 @@ struct sk_buff *tcp_tso_segment(struct sk_buff *skb, int features)
 }
 EXPORT_SYMBOL(tcp_tso_segment);
 
+#ifdef CONFIG_TCP_MD5SIG
+static unsigned long tcp_md5sig_users;
+static struct tcp_md5sig_pool **tcp_md5sig_pool;
+static DEFINE_SPINLOCK(tcp_md5sig_pool_lock);
+
+static void __tcp_free_md5sig_pool(struct tcp_md5sig_pool **pool)
+{
+	int cpu;
+	for_each_possible_cpu(cpu) {
+		struct tcp_md5sig_pool *p = *per_cpu_ptr(pool, cpu);
+		if (p) {
+			if (p->md5_desc.tfm)
+				crypto_free_hash(p->md5_desc.tfm);
+			kfree(p);
+			p = NULL;
+		}
+	}
+	free_percpu(pool);
+}
+
+void tcp_free_md5sig_pool(void)
+{
+	struct tcp_md5sig_pool **pool = NULL;
+
+	spin_lock(&tcp_md5sig_pool_lock);
+	if (--tcp_md5sig_users == 0) {
+		pool = tcp_md5sig_pool;
+		tcp_md5sig_pool = NULL;
+	}
+	spin_unlock(&tcp_md5sig_pool_lock);
+	if (pool)
+		__tcp_free_md5sig_pool(pool);
+}
+
+EXPORT_SYMBOL(tcp_free_md5sig_pool);
+
+struct tcp_md5sig_pool **__tcp_alloc_md5sig_pool(void)
+{
+	int cpu;
+	struct tcp_md5sig_pool **pool;
+
+	pool = alloc_percpu(struct tcp_md5sig_pool *);
+	if (!pool)
+		return NULL;
+
+	for_each_possible_cpu(cpu) {
+		struct tcp_md5sig_pool *p;
+		struct crypto_hash *hash;
+
+		p = kzalloc(sizeof(*p), GFP_KERNEL);
+		if (!p)
+			goto out_free;
+		*per_cpu_ptr(pool, cpu) = p;
+
+		hash = crypto_alloc_hash("md5", 0, CRYPTO_ALG_ASYNC);
+		if (!hash || IS_ERR(hash))
+			goto out_free;
+
+		p->md5_desc.tfm = hash;
+	}
+	return pool;
+out_free:
+	__tcp_free_md5sig_pool(pool);
+	return NULL;
+}
+
+struct tcp_md5sig_pool **tcp_alloc_md5sig_pool(void)
+{
+	struct tcp_md5sig_pool **pool;
+	int alloc = 0;
+
+retry:
+	spin_lock(&tcp_md5sig_pool_lock);
+	pool = tcp_md5sig_pool;
+	if (tcp_md5sig_users++ == 0) {
+		alloc = 1;
+		spin_unlock(&tcp_md5sig_pool_lock);
+	} else if (!pool) {
+		tcp_md5sig_users--;
+		spin_unlock(&tcp_md5sig_pool_lock);
+		cpu_relax();
+		goto retry;
+	} else
+		spin_unlock(&tcp_md5sig_pool_lock);
+
+	if (alloc) {
+		/* we cannot hold spinlock here because this may sleep. */
+		struct tcp_md5sig_pool **p = __tcp_alloc_md5sig_pool();
+		spin_lock(&tcp_md5sig_pool_lock);
+		if (!p) {
+			tcp_md5sig_users--;
+			spin_unlock(&tcp_md5sig_pool_lock);
+			return NULL;
+		}
+		pool = tcp_md5sig_pool;
+		if (pool) {
+			/* oops, it has already been assigned. */
+			spin_unlock(&tcp_md5sig_pool_lock);
+			__tcp_free_md5sig_pool(p);
+		} else {
+			tcp_md5sig_pool = pool = p;
+			spin_unlock(&tcp_md5sig_pool_lock);
+		}
+	}
+	return pool;
+}
+
+EXPORT_SYMBOL(tcp_alloc_md5sig_pool);
+
+struct tcp_md5sig_pool *__tcp_get_md5sig_pool(int cpu)
+{
+	struct tcp_md5sig_pool **p;
+	spin_lock(&tcp_md5sig_pool_lock);
+	p = tcp_md5sig_pool;
+	if (p)
+		tcp_md5sig_users++;
+	spin_unlock(&tcp_md5sig_pool_lock);
+	return (p ? *per_cpu_ptr(p, cpu) : NULL);
+}
+
+EXPORT_SYMBOL(__tcp_get_md5sig_pool);
+
+void __tcp_put_md5sig_pool(void) {
+	__tcp_free_md5sig_pool(tcp_md5sig_pool);
+}
+
+EXPORT_SYMBOL(__tcp_put_md5sig_pool);
+#endif
+
 extern void __skb_cb_too_small_for_tcp(int, int);
 extern struct tcp_congestion_ops tcp_reno;
 

net/ipv4/tcp_input.c

@@ -2677,6 +2677,14 @@ void tcp_parse_options(struct sk_buff *skb, struct tcp_options_received *opt_rx,
 					   opt_rx->sack_ok) {
 						TCP_SKB_CB(skb)->sacked = (ptr - 2) - (unsigned char *)th;
 					}
+#ifdef CONFIG_TCP_MD5SIG
+				case TCPOPT_MD5SIG:
+					/*
+					 * The MD5 Hash has already been
+					 * checked (see tcp_v{4,6}_do_rcv()).
+					 */
+					break;
+#endif
 	  			};
 	  			ptr+=opsize-2;
 	  			length-=opsize;

net/ipv4/tcp_minisocks.c

@@ -306,6 +306,28 @@ void tcp_time_wait(struct sock *sk, int state, int timeo)
 			tw->tw_ipv6only = np->ipv6only;
 		}
 #endif
+
+#ifdef CONFIG_TCP_MD5SIG
+		/*
+		 * The timewait bucket does not have the key DB from the
+		 * sock structure. We just make a quick copy of the
+		 * md5 key being used (if indeed we are using one)
+		 * so the timewait ack generating code has the key.
+		 */
+		do {
+			struct tcp_md5sig_key *key;
+			memset(tcptw->tw_md5_key, 0, sizeof(tcptw->tw_md5_key));
+			tcptw->tw_md5_keylen = 0;
+			key = tp->af_specific->md5_lookup(sk, sk);
+			if (key != NULL) {
+				memcpy(&tcptw->tw_md5_key, key->key, key->keylen);
+				tcptw->tw_md5_keylen = key->keylen;
+				if (tcp_alloc_md5sig_pool() == NULL)
+					BUG();
+			}
+		} while(0);
+#endif
+
 		/* Linkage updates. */
 		__inet_twsk_hashdance(tw, sk, &tcp_hashinfo);
 
@@ -337,6 +359,17 @@ void tcp_time_wait(struct sock *sk, int state, int timeo)
 	tcp_done(sk);
 }
 
+void tcp_twsk_destructor(struct sock *sk)
+{
+	struct tcp_timewait_sock *twsk = tcp_twsk(sk);
+#ifdef CONFIG_TCP_MD5SIG
+	if (twsk->tw_md5_keylen)
+		tcp_put_md5sig_pool();
+#endif
+}
+
+EXPORT_SYMBOL_GPL(tcp_twsk_destructor);
+
 /* This is not only more efficient than what we used to do, it eliminates
  * a lot of code duplication between IPv4/IPv6 SYN recv processing. -DaveM
  *
@@ -435,6 +468,11 @@ struct sock *tcp_create_openreq_child(struct sock *sk, struct request_sock *req,
 			newtp->rx_opt.ts_recent_stamp = 0;
 			newtp->tcp_header_len = sizeof(struct tcphdr);
 		}
+#ifdef CONFIG_TCP_MD5SIG
+		newtp->md5sig_info = NULL;	/*XXX*/
+		if (newtp->af_specific->md5_lookup(sk, newsk))
+			newtp->tcp_header_len += TCPOLEN_MD5SIG_ALIGNED;
+#endif
 		if (skb->len >= TCP_MIN_RCVMSS+newtp->tcp_header_len)
 			newicsk->icsk_ack.last_seg_size = skb->len - newtp->tcp_header_len;
 		newtp->rx_opt.mss_clamp = req->mss;
@@ -617,6 +655,30 @@ struct sock *tcp_check_req(struct sock *sk,struct sk_buff *skb,
 								 req, NULL);
 		if (child == NULL)
 			goto listen_overflow;
+#ifdef CONFIG_TCP_MD5SIG
+		else {
+			/* Copy over the MD5 key from the original socket */
+			struct tcp_md5sig_key *key;
+			struct tcp_sock *tp = tcp_sk(sk);
+			key = tp->af_specific->md5_lookup(sk, child);
+			if (key != NULL) {
+				/*
+				 * We're using one, so create a matching key on the
+				 * newsk structure. If we fail to get memory then we
+				 * end up not copying the key across. Shucks.
+				 */
+				char *newkey = kmalloc(key->keylen, GFP_ATOMIC);
+				if (newkey) {
+					if (!tcp_alloc_md5sig_pool())
+						BUG();
+					memcpy(newkey, key->key, key->keylen);
+					tp->af_specific->md5_add(child, child,
+								 newkey,
+								 key->keylen);
+				}
+			}
+		}
+#endif
 
 		inet_csk_reqsk_queue_unlink(sk, req, prev);
 		inet_csk_reqsk_queue_removed(sk, req);
@@ -633,7 +695,7 @@ struct sock *tcp_check_req(struct sock *sk,struct sk_buff *skb,
 	embryonic_reset:
 		NET_INC_STATS_BH(LINUX_MIB_EMBRYONICRSTS);
 		if (!(flg & TCP_FLAG_RST))
-			req->rsk_ops->send_reset(skb);
+			req->rsk_ops->send_reset(sk, skb);
 
 		inet_csk_reqsk_queue_drop(sk, req, prev);
 		return NULL;

net/ipv4/tcp_output.c

@@ -270,7 +270,7 @@ static u16 tcp_select_window(struct sock *sk)
 }
 
 static void tcp_build_and_update_options(__be32 *ptr, struct tcp_sock *tp,
-					 __u32 tstamp)
+					 __u32 tstamp, __u8 **md5_hash)
 {
 	if (tp->rx_opt.tstamp_ok) {
 		*ptr++ = htonl((TCPOPT_NOP << 24) |
@@ -298,16 +298,29 @@ static void tcp_build_and_update_options(__be32 *ptr, struct tcp_sock *tp,
 			tp->rx_opt.eff_sacks--;
 		}
 	}
+#ifdef CONFIG_TCP_MD5SIG
+	if (md5_hash) {
+		*ptr++ = htonl((TCPOPT_NOP << 24) |
+			       (TCPOPT_NOP << 16) |
+			       (TCPOPT_MD5SIG << 8) |
+			       TCPOLEN_MD5SIG);
+		*md5_hash = (__u8 *)ptr;
+	}
+#endif
 }
 
 /* Construct a tcp options header for a SYN or SYN_ACK packet.
  * If this is every changed make sure to change the definition of
  * MAX_SYN_SIZE to match the new maximum number of options that you
  * can generate.
+ *
+ * Note - that with the RFC2385 TCP option, we make room for the
+ * 16 byte MD5 hash. This will be filled in later, so the pointer for the
+ * location to be filled is passed back up.
  */
 static void tcp_syn_build_options(__be32 *ptr, int mss, int ts, int sack,
 				  int offer_wscale, int wscale, __u32 tstamp,
-				  __u32 ts_recent)
+				  __u32 ts_recent, __u8 **md5_hash)
 {
 	/* We always get an MSS option.
 	 * The option bytes which will be seen in normal data
@@ -346,6 +359,20 @@ static void tcp_syn_build_options(__be32 *ptr, int mss, int ts, int sack,
 			       (TCPOPT_WINDOW << 16) |
 			       (TCPOLEN_WINDOW << 8) |
 			       (wscale));
+#ifdef CONFIG_TCP_MD5SIG
+	/*
+	 * If MD5 is enabled, then we set the option, and include the size
+	 * (always 18). The actual MD5 hash is added just before the
+	 * packet is sent.
+	 */
+	if (md5_hash) {
+		*ptr++ = htonl((TCPOPT_NOP << 24) |
+			       (TCPOPT_NOP << 16) |
+			       (TCPOPT_MD5SIG << 8) |
+			       TCPOLEN_MD5SIG);
+		*md5_hash = (__u8 *) ptr;
+	}
+#endif
 }
 
 /* This routine actually transmits TCP packets queued in by
@@ -366,6 +393,10 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it,
 	struct tcp_sock *tp;
 	struct tcp_skb_cb *tcb;
 	int tcp_header_size;
+#ifdef CONFIG_TCP_MD5SIG
+	struct tcp_md5sig_key *md5;
+	__u8 *md5_hash_location;
+#endif
 	struct tcphdr *th;
 	int sysctl_flags;
 	int err;
@@ -424,6 +455,16 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it,
 	if (tcp_packets_in_flight(tp) == 0)
 		tcp_ca_event(sk, CA_EVENT_TX_START);
 
+#ifdef CONFIG_TCP_MD5SIG
+	/*
+	 * Are we doing MD5 on this segment? If so - make
+	 * room for it.
+	 */
+	md5 = tp->af_specific->md5_lookup(sk, sk);
+	if (md5)
+		tcp_header_size += TCPOLEN_MD5SIG_ALIGNED;
+#endif
+
 	th = (struct tcphdr *) skb_push(skb, tcp_header_size);
 	skb->h.th = th;
 
@@ -460,13 +501,34 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it,
 				      (sysctl_flags & SYSCTL_FLAG_WSCALE),
 				      tp->rx_opt.rcv_wscale,
 				      tcb->when,
-				      tp->rx_opt.ts_recent);
+				      tp->rx_opt.ts_recent,
+
+#ifdef CONFIG_TCP_MD5SIG
+				      md5 ? &md5_hash_location :
+#endif
+				      NULL);
 	} else {
 		tcp_build_and_update_options((__be32 *)(th + 1),
-					     tp, tcb->when);
+					     tp, tcb->when,
+#ifdef CONFIG_TCP_MD5SIG
+					     md5 ? &md5_hash_location :
+#endif
+					     NULL);
 		TCP_ECN_send(sk, tp, skb, tcp_header_size);
 	}
 
+#ifdef CONFIG_TCP_MD5SIG
+	/* Calculate the MD5 hash, as we have all we need now */
+	if (md5) {
+		tp->af_specific->calc_md5_hash(md5_hash_location,
+					       md5,
+					       sk, NULL, NULL,
+					       skb->h.th,
+					       sk->sk_protocol,
+					       skb->len);
+	}
+#endif
+
 	icsk->icsk_af_ops->send_check(sk, skb->len, skb);
 
 	if (likely(tcb->flags & TCPCB_FLAG_ACK))
@@ -840,6 +902,11 @@ unsigned int tcp_current_mss(struct sock *sk, int large_allowed)
 		mss_now -= (TCPOLEN_SACK_BASE_ALIGNED +
 			    (tp->rx_opt.eff_sacks * TCPOLEN_SACK_PERBLOCK));
 
+#ifdef CONFIG_TCP_MD5SIG
+	if (tp->af_specific->md5_lookup(sk, sk))
+		mss_now -= TCPOLEN_MD5SIG_ALIGNED;
+#endif
+
 	xmit_size_goal = mss_now;
 
 	if (doing_tso) {
@@ -2033,6 +2100,10 @@ struct sk_buff * tcp_make_synack(struct sock *sk, struct dst_entry *dst,
 	struct tcphdr *th;
 	int tcp_header_size;
 	struct sk_buff *skb;
+#ifdef CONFIG_TCP_MD5SIG
+	struct tcp_md5sig_key *md5;
+	__u8 *md5_hash_location;
+#endif
 
 	skb = sock_wmalloc(sk, MAX_TCP_HEADER + 15, 1, GFP_ATOMIC);
 	if (skb == NULL)
@@ -2048,6 +2119,13 @@ struct sk_buff * tcp_make_synack(struct sock *sk, struct dst_entry *dst,
 			   (ireq->wscale_ok ? TCPOLEN_WSCALE_ALIGNED : 0) +
 			   /* SACK_PERM is in the place of NOP NOP of TS */
 			   ((ireq->sack_ok && !ireq->tstamp_ok) ? TCPOLEN_SACKPERM_ALIGNED : 0));
+
+#ifdef CONFIG_TCP_MD5SIG
+	/* Are we doing MD5 on this segment? If so - make room for it */
+	md5 = tcp_rsk(req)->af_specific->md5_lookup(sk, req);
+	if (md5)
+		tcp_header_size += TCPOLEN_MD5SIG_ALIGNED;
+#endif
 	skb->h.th = th = (struct tcphdr *) skb_push(skb, tcp_header_size);
 
 	memset(th, 0, sizeof(struct tcphdr));
@@ -2085,11 +2163,29 @@ struct sk_buff * tcp_make_synack(struct sock *sk, struct dst_entry *dst,
 	tcp_syn_build_options((__be32 *)(th + 1), dst_metric(dst, RTAX_ADVMSS), ireq->tstamp_ok,
 			      ireq->sack_ok, ireq->wscale_ok, ireq->rcv_wscale,
 			      TCP_SKB_CB(skb)->when,
-			      req->ts_recent);
+			      req->ts_recent,
+			      (
+#ifdef CONFIG_TCP_MD5SIG
+			       md5 ? &md5_hash_location :
+#endif
+			       NULL)
+			      );
 
 	skb->csum = 0;
 	th->doff = (tcp_header_size >> 2);
 	TCP_INC_STATS(TCP_MIB_OUTSEGS);
+
+#ifdef CONFIG_TCP_MD5SIG
+	/* Okay, we have all we need - do the md5 hash if needed */
+	if (md5) {
+		tp->af_specific->calc_md5_hash(md5_hash_location,
+					       md5,
+					       NULL, dst, req,
+					       skb->h.th, sk->sk_protocol,
+					       skb->len);
+	}
+#endif
+
 	return skb;
 }
 
@@ -2108,6 +2204,11 @@ static void tcp_connect_init(struct sock *sk)
 	tp->tcp_header_len = sizeof(struct tcphdr) +
 		(sysctl_tcp_timestamps ? TCPOLEN_TSTAMP_ALIGNED : 0);
 
+#ifdef CONFIG_TCP_MD5SIG
+	if (tp->af_specific->md5_lookup(sk, sk) != NULL)
+		tp->tcp_header_len += TCPOLEN_MD5SIG_ALIGNED;
+#endif
+
 	/* If user gave his TCP_MAXSEG, record it to clamp */
 	if (tp->rx_opt.user_mss)
 		tp->rx_opt.mss_clamp = tp->rx_opt.user_mss;