From cfb6eeb4c860592edd123fdea908d23c6ad1c7dc Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Tue, 14 Nov 2006 19:07:45 -0800 Subject: [PATCH] [TCP]: MD5 Signature Option (RFC2385) support. Based on implementation by Rick Payne. Signed-off-by: YOSHIFUJI Hideaki Signed-off-by: David S. Miller --- CREDITS | 3 + include/linux/tcp.h | 35 +- include/net/request_sock.h | 3 +- include/net/tcp.h | 143 ++++++++ include/net/timewait_sock.h | 3 + net/dccp/ipv4.c | 6 +- net/dccp/ipv6.c | 6 +- net/dccp/minisocks.c | 2 +- net/ipv4/Kconfig | 16 + net/ipv4/tcp.c | 137 ++++++++ net/ipv4/tcp_input.c | 8 + net/ipv4/tcp_ipv4.c | 673 ++++++++++++++++++++++++++++++++++-- net/ipv4/tcp_minisocks.c | 64 +++- net/ipv4/tcp_output.c | 111 +++++- net/ipv6/tcp_ipv6.c | 568 +++++++++++++++++++++++++++++- 15 files changed, 1714 insertions(+), 64 deletions(-) diff --git a/CREDITS b/CREDITS index ccd4f9f4dd71..d0880082c19b 100644 --- a/CREDITS +++ b/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/ diff --git a/include/linux/tcp.h b/include/linux/tcp.h index 2d36f6db3706..0aecfc955591 100644 --- a/include/linux/tcp.h +++ b/include/linux/tcp.h @@ -19,6 +19,7 @@ #include #include +#include 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 @@ -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) diff --git a/include/net/request_sock.h b/include/net/request_sock.h index b5b023e79e5f..e37baaf2080b 100644 --- a/include/net/request_sock.h +++ b/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); }; diff --git a/include/net/tcp.h b/include/net/tcp.h index e1a5d29d0a1f..363960872de0 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -28,6 +28,7 @@ #include #include #include +#include #include #include @@ -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); diff --git a/include/net/timewait_sock.h b/include/net/timewait_sock.h index be293d795e38..d7a306ea560d 100644 --- a/include/net/timewait_sock.h +++ b/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); } diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c index 34d6d197c3b2..35985334daee 100644 --- a/net/dccp/ipv4.c +++ b/net/dccp/ipv4.c @@ -509,7 +509,7 @@ out: 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 @@ no_dccp_socket: 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: diff --git a/net/dccp/ipv6.c b/net/dccp/ipv6.c index fc326173c215..e0a0607862ef 100644 --- a/net/dccp/ipv6.c +++ b/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 @@ no_dccp_socket: 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: diff --git a/net/dccp/minisocks.c b/net/dccp/minisocks.c index 0c49733f5be1..3975048d8094 100644 --- a/net/dccp/minisocks.c +++ b/net/dccp/minisocks.c @@ -246,7 +246,7 @@ listen_overflow: 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; diff --git a/net/ipv4/Kconfig b/net/ipv4/Kconfig index bc298bcc344e..39e0cb763588 100644 --- a/net/ipv4/Kconfig +++ b/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" diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index c05e8edaf544..dadef867a3bb 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -258,6 +258,7 @@ #include #include #include +#include #include #include @@ -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 @@ out: } 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; diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 4a8c96cdec7d..6ab3423674bb 100644 --- a/net/ipv4/tcp_input.c +++ b/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; diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index 0ad0904bf56c..8c8e8112f98d 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -78,6 +78,9 @@ #include #include +#include +#include + int sysctl_tcp_tw_reuse __read_mostly; int sysctl_tcp_low_latency __read_mostly; @@ -89,6 +92,13 @@ static struct socket *tcp_socket; void tcp_v4_send_check(struct sock *sk, int len, struct sk_buff *skb); +#ifdef CONFIG_TCP_MD5SIG +static struct tcp_md5sig_key *tcp_v4_md5_do_lookup(struct sock *sk, __be32 addr); +static int tcp_v4_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, + __be32 saddr, __be32 daddr, struct tcphdr *th, + int protocol, int tcplen); +#endif + struct inet_hashinfo __cacheline_aligned tcp_hashinfo = { .lhash_lock = __RW_LOCK_UNLOCKED(tcp_hashinfo.lhash_lock), .lhash_users = ATOMIC_INIT(0), @@ -526,11 +536,19 @@ int tcp_v4_gso_send_check(struct sk_buff *skb) * Exception: precedence violation. We do not implement it in any case. */ -static void tcp_v4_send_reset(struct sk_buff *skb) +static void tcp_v4_send_reset(struct sock *sk, struct sk_buff *skb) { struct tcphdr *th = skb->h.th; - struct tcphdr rth; + struct { + struct tcphdr th; +#ifdef CONFIG_TCP_MD5SIG + u32 opt[(TCPOLEN_MD5SIG_ALIGNED >> 2)]; +#endif + } rep; struct ip_reply_arg arg; +#ifdef CONFIG_TCP_MD5SIG + struct tcp_md5sig_key *key; +#endif /* Never send a reset in response to a reset. */ if (th->rst) @@ -540,29 +558,50 @@ static void tcp_v4_send_reset(struct sk_buff *skb) return; /* Swap the send and the receive. */ - memset(&rth, 0, sizeof(struct tcphdr)); - rth.dest = th->source; - rth.source = th->dest; - rth.doff = sizeof(struct tcphdr) / 4; - rth.rst = 1; + memset(&rep, 0, sizeof(rep)); + rep.th.dest = th->source; + rep.th.source = th->dest; + rep.th.doff = sizeof(struct tcphdr) / 4; + rep.th.rst = 1; if (th->ack) { - rth.seq = th->ack_seq; + rep.th.seq = th->ack_seq; } else { - rth.ack = 1; - rth.ack_seq = htonl(ntohl(th->seq) + th->syn + th->fin + - skb->len - (th->doff << 2)); + rep.th.ack = 1; + rep.th.ack_seq = htonl(ntohl(th->seq) + th->syn + th->fin + + skb->len - (th->doff << 2)); } memset(&arg, 0, sizeof arg); - arg.iov[0].iov_base = (unsigned char *)&rth; - arg.iov[0].iov_len = sizeof rth; + arg.iov[0].iov_base = (unsigned char *)&rep; + arg.iov[0].iov_len = sizeof(rep.th); + +#ifdef CONFIG_TCP_MD5SIG + key = sk ? tcp_v4_md5_do_lookup(sk, skb->nh.iph->daddr) : NULL; + if (key) { + rep.opt[0] = htonl((TCPOPT_NOP << 24) | + (TCPOPT_NOP << 16) | + (TCPOPT_MD5SIG << 8) | + TCPOLEN_MD5SIG); + /* Update length and the length the header thinks exists */ + arg.iov[0].iov_len += TCPOLEN_MD5SIG_ALIGNED; + rep.th.doff = arg.iov[0].iov_len / 4; + + tcp_v4_do_calc_md5_hash((__u8 *)&rep.opt[1], + key, + skb->nh.iph->daddr, + skb->nh.iph->saddr, + &rep.th, IPPROTO_TCP, + arg.iov[0].iov_len); + } +#endif + arg.csum = csum_tcpudp_nofold(skb->nh.iph->daddr, skb->nh.iph->saddr, /*XXX*/ sizeof(struct tcphdr), IPPROTO_TCP, 0); arg.csumoffset = offsetof(struct tcphdr, check) / 2; - ip_send_reply(tcp_socket->sk, skb, &arg, sizeof rth); + ip_send_reply(tcp_socket->sk, skb, &arg, arg.iov[0].iov_len); TCP_INC_STATS_BH(TCP_MIB_OUTSEGS); TCP_INC_STATS_BH(TCP_MIB_OUTRSTS); @@ -572,15 +611,24 @@ static void tcp_v4_send_reset(struct sk_buff *skb) outside socket context is ugly, certainly. What can I do? */ -static void tcp_v4_send_ack(struct sk_buff *skb, u32 seq, u32 ack, +static void tcp_v4_send_ack(struct tcp_timewait_sock *twsk, + struct sk_buff *skb, u32 seq, u32 ack, u32 win, u32 ts) { struct tcphdr *th = skb->h.th; struct { struct tcphdr th; - u32 tsopt[TCPOLEN_TSTAMP_ALIGNED >> 2]; + u32 opt[(TCPOLEN_TSTAMP_ALIGNED >> 2) +#ifdef CONFIG_TCP_MD5SIG + + (TCPOLEN_MD5SIG_ALIGNED >> 2) +#endif + ]; } rep; struct ip_reply_arg arg; +#ifdef CONFIG_TCP_MD5SIG + struct tcp_md5sig_key *key; + struct tcp_md5sig_key tw_key; +#endif memset(&rep.th, 0, sizeof(struct tcphdr)); memset(&arg, 0, sizeof arg); @@ -588,12 +636,12 @@ static void tcp_v4_send_ack(struct sk_buff *skb, u32 seq, u32 ack, arg.iov[0].iov_base = (unsigned char *)&rep; arg.iov[0].iov_len = sizeof(rep.th); if (ts) { - rep.tsopt[0] = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) | - (TCPOPT_TIMESTAMP << 8) | - TCPOLEN_TIMESTAMP); - rep.tsopt[1] = htonl(tcp_time_stamp); - rep.tsopt[2] = htonl(ts); - arg.iov[0].iov_len = sizeof(rep); + rep.opt[0] = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) | + (TCPOPT_TIMESTAMP << 8) | + TCPOLEN_TIMESTAMP); + rep.opt[1] = htonl(tcp_time_stamp); + rep.opt[2] = htonl(ts); + arg.iov[0].iov_len = TCPOLEN_TSTAMP_ALIGNED; } /* Swap the send and the receive. */ @@ -605,6 +653,44 @@ static void tcp_v4_send_ack(struct sk_buff *skb, u32 seq, u32 ack, rep.th.ack = 1; rep.th.window = htons(win); +#ifdef CONFIG_TCP_MD5SIG + /* + * The SKB holds an imcoming packet, but may not have a valid ->sk + * pointer. This is especially the case when we're dealing with a + * TIME_WAIT ack, because the sk structure is long gone, and only + * the tcp_timewait_sock remains. So the md5 key is stashed in that + * structure, and we use it in preference. I believe that (twsk || + * skb->sk) holds true, but we program defensively. + */ + if (!twsk && skb->sk) { + key = tcp_v4_md5_do_lookup(skb->sk, skb->nh.iph->daddr); + } else if (twsk && twsk->tw_md5_keylen) { + tw_key.key = twsk->tw_md5_key; + tw_key.keylen = twsk->tw_md5_keylen; + key = &tw_key; + } else { + key = NULL; + } + + if (key) { + int offset = (ts) ? 3 : 0; + + rep.opt[offset++] = htonl((TCPOPT_NOP << 24) | + (TCPOPT_NOP << 16) | + (TCPOPT_MD5SIG << 8) | + TCPOLEN_MD5SIG); + arg.iov[0].iov_len += TCPOLEN_MD5SIG_ALIGNED; + rep.th.doff = arg.iov[0].iov_len/4; + + tcp_v4_do_calc_md5_hash((__u8 *)&rep.opt[offset], + key, + skb->nh.iph->daddr, + skb->nh.iph->saddr, + &rep.th, IPPROTO_TCP, + arg.iov[0].iov_len); + } +#endif + arg.csum = csum_tcpudp_nofold(skb->nh.iph->daddr, skb->nh.iph->saddr, /*XXX*/ arg.iov[0].iov_len, IPPROTO_TCP, 0); @@ -618,9 +704,9 @@ static void tcp_v4_send_ack(struct sk_buff *skb, u32 seq, u32 ack, static void tcp_v4_timewait_ack(struct sock *sk, struct sk_buff *skb) { struct inet_timewait_sock *tw = inet_twsk(sk); - const struct tcp_timewait_sock *tcptw = tcp_twsk(sk); + struct tcp_timewait_sock *tcptw = tcp_twsk(sk); - tcp_v4_send_ack(skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt, + tcp_v4_send_ack(tcptw, skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt, tcptw->tw_rcv_wnd >> tw->tw_rcv_wscale, tcptw->tw_ts_recent); inet_twsk_put(tw); @@ -628,7 +714,8 @@ static void tcp_v4_timewait_ack(struct sock *sk, struct sk_buff *skb) static void tcp_v4_reqsk_send_ack(struct sk_buff *skb, struct request_sock *req) { - tcp_v4_send_ack(skb, tcp_rsk(req)->snt_isn + 1, tcp_rsk(req)->rcv_isn + 1, req->rcv_wnd, + tcp_v4_send_ack(NULL, skb, tcp_rsk(req)->snt_isn + 1, + tcp_rsk(req)->rcv_isn + 1, req->rcv_wnd, req->ts_recent); } @@ -714,6 +801,461 @@ static struct ip_options *tcp_v4_save_options(struct sock *sk, return dopt; } +#ifdef CONFIG_TCP_MD5SIG +/* + * RFC2385 MD5 checksumming requires a mapping of + * IP address->MD5 Key. + * We need to maintain these in the sk structure. + */ + +/* Find the Key structure for an address. */ +static struct tcp_md5sig_key *tcp_v4_md5_do_lookup(struct sock *sk, __be32 addr) +{ + struct tcp_sock *tp = tcp_sk(sk); + int i; + + if (!tp->md5sig_info || !tp->md5sig_info->entries4) + return NULL; + for (i = 0; i < tp->md5sig_info->entries4; i++) { + if (tp->md5sig_info->keys4[i].addr == addr) + return (struct tcp_md5sig_key *)&tp->md5sig_info->keys4[i]; + } + return NULL; +} + +struct tcp_md5sig_key *tcp_v4_md5_lookup(struct sock *sk, + struct sock *addr_sk) +{ + return tcp_v4_md5_do_lookup(sk, inet_sk(addr_sk)->daddr); +} + +EXPORT_SYMBOL(tcp_v4_md5_lookup); + +struct tcp_md5sig_key *tcp_v4_reqsk_md5_lookup(struct sock *sk, + struct request_sock *req) +{ + return tcp_v4_md5_do_lookup(sk, inet_rsk(req)->rmt_addr); +} + +/* This can be called on a newly created socket, from other files */ +int tcp_v4_md5_do_add(struct sock *sk, __be32 addr, + u8 *newkey, u8 newkeylen) +{ + /* Add Key to the list */ + struct tcp4_md5sig_key *key; + struct tcp_sock *tp = tcp_sk(sk); + struct tcp4_md5sig_key *keys; + + key = (struct tcp4_md5sig_key *) tcp_v4_md5_do_lookup(sk, addr); + if (key) { + /* Pre-existing entry - just update that one. */ + kfree (key->key); + key->key = newkey; + key->keylen = newkeylen; + } else { + if (!tp->md5sig_info) { + tp->md5sig_info = kzalloc(sizeof(*tp->md5sig_info), GFP_ATOMIC); + if (!tp->md5sig_info) { + kfree(newkey); + return -ENOMEM; + } + } + if (tcp_alloc_md5sig_pool() == NULL) { + kfree(newkey); + return -ENOMEM; + } + if (tp->md5sig_info->alloced4 == tp->md5sig_info->entries4) { + keys = kmalloc((sizeof(struct tcp4_md5sig_key) * + (tp->md5sig_info->entries4 + 1)), GFP_ATOMIC); + if (!keys) { + kfree(newkey); + tcp_free_md5sig_pool(); + return -ENOMEM; + } + + if (tp->md5sig_info->entries4) + memcpy(keys, tp->md5sig_info->keys4, + (sizeof (struct tcp4_md5sig_key) * + tp->md5sig_info->entries4)); + + /* Free old key list, and reference new one */ + if (tp->md5sig_info->keys4) + kfree(tp->md5sig_info->keys4); + tp->md5sig_info->keys4 = keys; + tp->md5sig_info->alloced4++; + } + tp->md5sig_info->entries4++; + tp->md5sig_info->keys4[tp->md5sig_info->entries4 - 1].addr = addr; + tp->md5sig_info->keys4[tp->md5sig_info->entries4 - 1].key = newkey; + tp->md5sig_info->keys4[tp->md5sig_info->entries4 - 1].keylen = newkeylen; + } + return 0; +} + +EXPORT_SYMBOL(tcp_v4_md5_do_add); + +static int tcp_v4_md5_add_func(struct sock *sk, struct sock *addr_sk, + u8 *newkey, u8 newkeylen) +{ + return tcp_v4_md5_do_add(sk, inet_sk(addr_sk)->daddr, + newkey, newkeylen); +} + +int tcp_v4_md5_do_del(struct sock *sk, __be32 addr) +{ + struct tcp_sock *tp = tcp_sk(sk); + int i; + + for (i = 0; i < tp->md5sig_info->entries4; i++) { + if (tp->md5sig_info->keys4[i].addr == addr) { + /* Free the key */ + kfree(tp->md5sig_info->keys4[i].key); + tp->md5sig_info->entries4--; + + if (tp->md5sig_info->entries4 == 0) { + kfree(tp->md5sig_info->keys4); + tp->md5sig_info->keys4 = NULL; + } else { + /* Need to do some manipulation */ + if (tp->md5sig_info->entries4 != i) + memcpy(&tp->md5sig_info->keys4[i], + &tp->md5sig_info->keys4[i+1], + (tp->md5sig_info->entries4 - i) + * sizeof (struct tcp4_md5sig_key)); + } + tcp_free_md5sig_pool(); + return 0; + } + } + return -ENOENT; +} + +EXPORT_SYMBOL(tcp_v4_md5_do_del); + +static void tcp_v4_clear_md5_list (struct sock *sk) +{ + struct tcp_sock *tp = tcp_sk(sk); + + /* Free each key, then the set of key keys, + * the crypto element, and then decrement our + * hold on the last resort crypto. + */ + if (tp->md5sig_info->entries4) { + int i; + for (i = 0; i < tp->md5sig_info->entries4; i++) + kfree(tp->md5sig_info->keys4[i].key); + tp->md5sig_info->entries4 = 0; + tcp_free_md5sig_pool(); + } + if (tp->md5sig_info->keys4) { + kfree(tp->md5sig_info->keys4); + tp->md5sig_info->keys4 = NULL; + tp->md5sig_info->alloced4 = 0; + } +} + +static int tcp_v4_parse_md5_keys (struct sock *sk, char __user *optval, + int optlen) +{ + struct tcp_md5sig cmd; + struct sockaddr_in *sin = (struct sockaddr_in *)&cmd.tcpm_addr; + u8 *newkey; + + if (optlen < sizeof(cmd)) + return -EINVAL; + + if (copy_from_user (&cmd, optval, sizeof(cmd))) + return -EFAULT; + + if (sin->sin_family != AF_INET) + return -EINVAL; + + if (!cmd.tcpm_key || !cmd.tcpm_keylen) { + if (!tcp_sk(sk)->md5sig_info) + return -ENOENT; + return tcp_v4_md5_do_del(sk, sin->sin_addr.s_addr); + } + + if (cmd.tcpm_keylen > TCP_MD5SIG_MAXKEYLEN) + return -EINVAL; + + if (!tcp_sk(sk)->md5sig_info) { + struct tcp_sock *tp = tcp_sk(sk); + struct tcp_md5sig_info *p; + + p = kzalloc(sizeof(struct tcp_md5sig_info), GFP_KERNEL); + if (!p) + return -EINVAL; + + tp->md5sig_info = p; + + } + + newkey = kmalloc(cmd.tcpm_keylen, GFP_KERNEL); + if (!newkey) + return -ENOMEM; + memcpy(newkey, cmd.tcpm_key, cmd.tcpm_keylen); + return tcp_v4_md5_do_add(sk, sin->sin_addr.s_addr, + newkey, cmd.tcpm_keylen); +} + +static int tcp_v4_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, + __be32 saddr, __be32 daddr, + struct tcphdr *th, int protocol, + int tcplen) +{ + struct scatterlist sg[4]; + __u16 data_len; + int block = 0; +#ifdef CONFIG_TCP_MD5SIG_DEBUG + int i; +#endif + __u16 old_checksum; + struct tcp_md5sig_pool *hp; + struct tcp4_pseudohdr *bp; + struct hash_desc *desc; + int err; + unsigned int nbytes = 0; + + /* + * Okay, so RFC2385 is turned on for this connection, + * so we need to generate the MD5 hash for the packet now. + */ + + hp = tcp_get_md5sig_pool(); + if (!hp) + goto clear_hash_noput; + + bp = &hp->md5_blk.ip4; + desc = &hp->md5_desc; + + /* + * 1. the TCP pseudo-header (in the order: source IP address, + * destination IP address, zero-padded protocol number, and + * segment length) + */ + bp->saddr = saddr; + bp->daddr = daddr; + bp->pad = 0; + bp->protocol = protocol; + bp->len = htons(tcplen); + sg_set_buf(&sg[block++], bp, sizeof(*bp)); + nbytes += sizeof(*bp); + +#ifdef CONFIG_TCP_MD5SIG_DEBUG + printk("Calcuating hash for: "); + for (i = 0; i < sizeof (*bp); i++) + printk ("%02x ", (unsigned int)((unsigned char *)bp)[i]); + printk(" "); +#endif + + /* 2. the TCP header, excluding options, and assuming a + * checksum of zero/ + */ + old_checksum = th->check; + th->check = 0; + sg_set_buf(&sg[block++], th, sizeof(struct tcphdr)); + nbytes += sizeof(struct tcphdr); +#ifdef CONFIG_TCP_MD5SIG_DEBUG + for (i = 0; i < sizeof (struct tcphdr); i++) + printk (" %02x", (unsigned int)((unsigned char *)th)[i]); +#endif + /* 3. the TCP segment data (if any) */ + data_len = tcplen - (th->doff << 2); + if (data_len > 0) { + unsigned char *data = (unsigned char *)th + (th->doff << 2); + sg_set_buf(&sg[block++], data, data_len); + nbytes += data_len; + } + + /* 4. an independently-specified key or password, known to both + * TCPs and presumably connection-specific + */ + sg_set_buf(&sg[block++], key->key, key->keylen); + nbytes += key->keylen; + +#ifdef CONFIG_TCP_MD5SIG_DEBUG + printk (" and password: "); + for (i = 0; i < key->keylen; i++) + printk ("%02x ", (unsigned int)key->key[i]); +#endif + + /* Now store the Hash into the packet */ + err = crypto_hash_init(desc); + if (err) + goto clear_hash; + err = crypto_hash_update(desc, sg, nbytes); + if (err) + goto clear_hash; + err = crypto_hash_final(desc, md5_hash); + if (err) + goto clear_hash; + + /* Reset header, and free up the crypto */ + tcp_put_md5sig_pool(); + th->check = old_checksum; + +out: +#ifdef CONFIG_TCP_MD5SIG_DEBUG + printk(" result:"); + for (i = 0; i < 16; i++) + printk (" %02x", (unsigned int)(((u8*)md5_hash)[i])); + printk("\n"); +#endif + return 0; +clear_hash: + tcp_put_md5sig_pool(); +clear_hash_noput: + memset(md5_hash, 0, 16); + goto out; +} + +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) +{ + __be32 saddr, daddr; + + if (sk) { + saddr = inet_sk(sk)->saddr; + daddr = inet_sk(sk)->daddr; + } else { + struct rtable *rt = (struct rtable *)dst; + BUG_ON(!rt); + saddr = rt->rt_src; + daddr = rt->rt_dst; + } + return tcp_v4_do_calc_md5_hash(md5_hash, key, + saddr, daddr, + th, protocol, tcplen); +} + +EXPORT_SYMBOL(tcp_v4_calc_md5_hash); + +static int tcp_v4_inbound_md5_hash (struct sock *sk, struct sk_buff *skb) +{ + /* + * This gets called for each TCP segment that arrives + * so we want to be efficient. + * We have 3 drop cases: + * o No MD5 hash and one expected. + * o MD5 hash and we're not expecting one. + * o MD5 hash and its wrong. + */ + __u8 *hash_location = NULL; + struct tcp_md5sig_key *hash_expected; + struct iphdr *iph = skb->nh.iph; + struct tcphdr *th = skb->h.th; + int length = (th->doff << 2) - sizeof (struct tcphdr); + int genhash; + unsigned char *ptr; + unsigned char newhash[16]; + + hash_expected = tcp_v4_md5_do_lookup(sk, iph->saddr); + + /* + * If the TCP option length is less than the TCP_MD5SIG + * option length, then we can shortcut + */ + if (length < TCPOLEN_MD5SIG) { + if (hash_expected) + return 1; + else + return 0; + } + + /* Okay, we can't shortcut - we have to grub through the options */ + ptr = (unsigned char *)(th + 1); + while (length > 0) { + int opcode = *ptr++; + int opsize; + + switch (opcode) { + case TCPOPT_EOL: + goto done_opts; + case TCPOPT_NOP: + length--; + continue; + default: + opsize = *ptr++; + if (opsize < 2) + goto done_opts; + if (opsize > length) + goto done_opts; + + if (opcode == TCPOPT_MD5SIG) { + hash_location = ptr; + goto done_opts; + } + } + ptr += opsize-2; + length -= opsize; + } +done_opts: + /* We've parsed the options - do we have a hash? */ + if (!hash_expected && !hash_location) + return 0; + + if (hash_expected && !hash_location) { + if (net_ratelimit()) { + printk(KERN_INFO "MD5 Hash NOT expected but found " + "(" NIPQUAD_FMT ", %d)->(" NIPQUAD_FMT ", %d)\n", + NIPQUAD (iph->saddr), ntohs(th->source), + NIPQUAD (iph->daddr), ntohs(th->dest)); + } + return 1; + } + + if (!hash_expected && hash_location) { + if (net_ratelimit()) { + printk(KERN_INFO "MD5 Hash NOT expected but found " + "(" NIPQUAD_FMT ", %d)->(" NIPQUAD_FMT ", %d)\n", + NIPQUAD (iph->saddr), ntohs(th->source), + NIPQUAD (iph->daddr), ntohs(th->dest)); + } + return 1; + } + + /* Okay, so this is hash_expected and hash_location - + * so we need to calculate the checksum. + */ + genhash = tcp_v4_do_calc_md5_hash(newhash, + hash_expected, + iph->saddr, iph->daddr, + th, sk->sk_protocol, + skb->len); + + if (genhash || memcmp(hash_location, newhash, 16) != 0) { + if (net_ratelimit()) { + printk(KERN_INFO "MD5 Hash failed for " + "(" NIPQUAD_FMT ", %d)->(" NIPQUAD_FMT ", %d)%s\n", + NIPQUAD (iph->saddr), ntohs(th->source), + NIPQUAD (iph->daddr), ntohs(th->dest), + genhash ? " tcp_v4_calc_md5_hash failed" : ""); +#ifdef CONFIG_TCP_MD5SIG_DEBUG + do { + int i; + printk("Received: "); + for (i = 0; i < 16; i++) + printk("%02x ", 0xff & (int)hash_location[i]); + printk("\n"); + printk("Calculated: "); + for (i = 0; i < 16; i++) + printk("%02x ", 0xff & (int)newhash[i]); + printk("\n"); + } while(0); +#endif + } + return 1; + } + return 0; +} + +#endif + struct request_sock_ops tcp_request_sock_ops __read_mostly = { .family = PF_INET, .obj_size = sizeof(struct tcp_request_sock), @@ -723,9 +1265,16 @@ struct request_sock_ops tcp_request_sock_ops __read_mostly = { .send_reset = tcp_v4_send_reset, }; +struct tcp_request_sock_ops tcp_request_sock_ipv4_ops = { +#ifdef CONFIG_TCP_MD5SIG + .md5_lookup = tcp_v4_reqsk_md5_lookup, +#endif +}; + static struct timewait_sock_ops tcp_timewait_sock_ops = { .twsk_obj_size = sizeof(struct tcp_timewait_sock), .twsk_unique = tcp_twsk_unique, + .twsk_destructor= tcp_twsk_destructor, }; int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb) @@ -773,6 +1322,10 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb) if (!req) goto drop; +#ifdef CONFIG_TCP_MD5SIG + tcp_rsk(req)->af_specific = &tcp_request_sock_ipv4_ops; +#endif + tcp_clear_options(&tmp_opt); tmp_opt.mss_clamp = 536; tmp_opt.user_mss = tcp_sk(sk)->rx_opt.user_mss; @@ -891,6 +1444,9 @@ struct sock *tcp_v4_syn_recv_sock(struct sock *sk, struct sk_buff *skb, struct inet_sock *newinet; struct tcp_sock *newtp; struct sock *newsk; +#ifdef CONFIG_TCP_MD5SIG + struct tcp_md5sig_key *key; +#endif if (sk_acceptq_is_full(sk)) goto exit_overflow; @@ -925,6 +1481,24 @@ struct sock *tcp_v4_syn_recv_sock(struct sock *sk, struct sk_buff *skb, newtp->advmss = dst_metric(dst, RTAX_ADVMSS); tcp_initialize_rcv_mss(newsk); +#ifdef CONFIG_TCP_MD5SIG + /* Copy over the MD5 key from the original socket */ + if ((key = tcp_v4_md5_do_lookup(sk, newinet->daddr)) != 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) { + memcpy(newkey, key->key, key->keylen); + tcp_v4_md5_do_add(newsk, inet_sk(sk)->daddr, + newkey, key->keylen); + } + } +#endif + __inet_hash(&tcp_hashinfo, newsk, 0); __inet_inherit_port(&tcp_hashinfo, sk, newsk); @@ -1000,10 +1574,24 @@ static int tcp_v4_checksum_init(struct sk_buff *skb) */ int tcp_v4_do_rcv(struct sock *sk, struct sk_buff *skb) { + struct sock *rsk; +#ifdef CONFIG_TCP_MD5SIG + /* + * We really want to reject the packet as early as possible + * if: + * o We're expecting an MD5'd packet and this is no MD5 tcp option + * o There is an MD5 option and we're not expecting one + */ + if (tcp_v4_inbound_md5_hash (sk, skb)) + goto discard; +#endif + if (sk->sk_state == TCP_ESTABLISHED) { /* Fast path */ TCP_CHECK_TIMER(sk); - if (tcp_rcv_established(sk, skb, skb->h.th, skb->len)) + if (tcp_rcv_established(sk, skb, skb->h.th, skb->len)) { + rsk = sk; goto reset; + } TCP_CHECK_TIMER(sk); return 0; } @@ -1017,20 +1605,24 @@ int tcp_v4_do_rcv(struct sock *sk, struct sk_buff *skb) goto discard; if (nsk != sk) { - if (tcp_child_process(sk, nsk, skb)) + if (tcp_child_process(sk, nsk, skb)) { + rsk = nsk; goto reset; + } return 0; } } TCP_CHECK_TIMER(sk); - if (tcp_rcv_state_process(sk, skb, skb->h.th, skb->len)) + if (tcp_rcv_state_process(sk, skb, skb->h.th, skb->len)) { + rsk = sk; goto reset; + } TCP_CHECK_TIMER(sk); return 0; reset: - tcp_v4_send_reset(skb); + tcp_v4_send_reset(rsk, skb); discard: kfree_skb(skb); /* Be careful here. If this function gets more complicated and @@ -1139,7 +1731,7 @@ no_tcp_socket: bad_packet: TCP_INC_STATS_BH(TCP_MIB_INERRS); } else { - tcp_v4_send_reset(skb); + tcp_v4_send_reset(NULL, skb); } discard_it: @@ -1262,6 +1854,15 @@ struct inet_connection_sock_af_ops ipv4_specific = { #endif }; +struct tcp_sock_af_ops tcp_sock_ipv4_specific = { +#ifdef CONFIG_TCP_MD5SIG + .md5_lookup = tcp_v4_md5_lookup, + .calc_md5_hash = tcp_v4_calc_md5_hash, + .md5_add = tcp_v4_md5_add_func, + .md5_parse = tcp_v4_parse_md5_keys, +#endif +}; + /* NOTE: A lot of things set to zero explicitly by call to * sk_alloc() so need not be done here. */ @@ -1301,6 +1902,9 @@ static int tcp_v4_init_sock(struct sock *sk) icsk->icsk_af_ops = &ipv4_specific; icsk->icsk_sync_mss = tcp_sync_mss; +#ifdef CONFIG_TCP_MD5SIG + tp->af_specific = &tcp_sock_ipv4_specific; +#endif sk->sk_sndbuf = sysctl_tcp_wmem[1]; sk->sk_rcvbuf = sysctl_tcp_rmem[1]; @@ -1324,6 +1928,15 @@ int tcp_v4_destroy_sock(struct sock *sk) /* Cleans up our, hopefully empty, out_of_order_queue. */ __skb_queue_purge(&tp->out_of_order_queue); +#ifdef CONFIG_TCP_MD5SIG + /* Clean up the MD5 key list, if any */ + if (tp->md5sig_info) { + tcp_v4_clear_md5_list(sk); + kfree(tp->md5sig_info); + tp->md5sig_info = NULL; + } +#endif + #ifdef CONFIG_NET_DMA /* Cleans up our sk_async_wait_queue */ __skb_queue_purge(&sk->sk_async_wait_queue); diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c index 0163d9826907..ac55d8892cf1 100644 --- a/net/ipv4/tcp_minisocks.c +++ b/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; diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 6a8581ab9a23..32c1a972fa31 100644 --- a/net/ipv4/tcp_output.c +++ b/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; diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 9a88395a7629..663d1d238014 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -66,10 +66,13 @@ #include #include +#include +#include + /* Socket used for sending RSTs and ACKs */ static struct socket *tcp6_socket; -static void tcp_v6_send_reset(struct sk_buff *skb); +static void tcp_v6_send_reset(struct sock *sk, struct sk_buff *skb); static void tcp_v6_reqsk_send_ack(struct sk_buff *skb, struct request_sock *req); static void tcp_v6_send_check(struct sock *sk, int len, struct sk_buff *skb); @@ -78,6 +81,8 @@ static int tcp_v6_do_rcv(struct sock *sk, struct sk_buff *skb); static struct inet_connection_sock_af_ops ipv6_mapped; static struct inet_connection_sock_af_ops ipv6_specific; +static struct tcp_sock_af_ops tcp_sock_ipv6_specific; +static struct tcp_sock_af_ops tcp_sock_ipv6_mapped_specific; static int tcp_v6_get_port(struct sock *sk, unsigned short snum) { @@ -208,6 +213,9 @@ static int tcp_v6_connect(struct sock *sk, struct sockaddr *uaddr, icsk->icsk_af_ops = &ipv6_mapped; sk->sk_backlog_rcv = tcp_v4_do_rcv; +#ifdef CONFIG_TCP_MD5SIG + tp->af_specific = &tcp_sock_ipv6_mapped_specific; +#endif err = tcp_v4_connect(sk, (struct sockaddr *)&sin, sizeof(sin)); @@ -215,6 +223,9 @@ static int tcp_v6_connect(struct sock *sk, struct sockaddr *uaddr, icsk->icsk_ext_hdr_len = exthdrlen; icsk->icsk_af_ops = &ipv6_specific; sk->sk_backlog_rcv = tcp_v6_do_rcv; +#ifdef CONFIG_TCP_MD5SIG + tp->af_specific = &tcp_sock_ipv6_specific; +#endif goto failure; } else { ipv6_addr_set(&np->saddr, 0, 0, htonl(0x0000FFFF), @@ -518,6 +529,396 @@ static void tcp_v6_reqsk_destructor(struct request_sock *req) kfree_skb(inet6_rsk(req)->pktopts); } +#ifdef CONFIG_TCP_MD5SIG +static struct tcp_md5sig_key *tcp_v6_md5_do_lookup(struct sock *sk, + struct in6_addr *addr) +{ + struct tcp_sock *tp = tcp_sk(sk); + int i; + + BUG_ON(tp == NULL); + + if (!tp->md5sig_info || !tp->md5sig_info->entries6) + return NULL; + + for (i = 0; i < tp->md5sig_info->entries6; i++) { + if (ipv6_addr_cmp(&tp->md5sig_info->keys6[i].addr, addr) == 0) + return (struct tcp_md5sig_key *)&tp->md5sig_info->keys6[i]; + } + return NULL; +} + +static struct tcp_md5sig_key *tcp_v6_md5_lookup(struct sock *sk, + struct sock *addr_sk) +{ + return tcp_v6_md5_do_lookup(sk, &inet6_sk(addr_sk)->daddr); +} + +static struct tcp_md5sig_key *tcp_v6_reqsk_md5_lookup(struct sock *sk, + struct request_sock *req) +{ + return tcp_v6_md5_do_lookup(sk, &inet6_rsk(req)->rmt_addr); +} + +static int tcp_v6_md5_do_add(struct sock *sk, struct in6_addr *peer, + char *newkey, u8 newkeylen) +{ + /* Add key to the list */ + struct tcp6_md5sig_key *key; + struct tcp_sock *tp = tcp_sk(sk); + struct tcp6_md5sig_key *keys; + + key = (struct tcp6_md5sig_key*) tcp_v6_md5_do_lookup(sk, peer); + if (key) { + /* modify existing entry - just update that one */ + kfree(key->key); + key->key = newkey; + key->keylen = newkeylen; + } else { + /* reallocate new list if current one is full. */ + if (!tp->md5sig_info) { + tp->md5sig_info = kzalloc(sizeof(*tp->md5sig_info), GFP_ATOMIC); + if (!tp->md5sig_info) { + kfree(newkey); + return -ENOMEM; + } + } + tcp_alloc_md5sig_pool(); + if (tp->md5sig_info->alloced6 == tp->md5sig_info->entries6) { + keys = kmalloc((sizeof (tp->md5sig_info->keys6[0]) * + (tp->md5sig_info->entries6 + 1)), GFP_ATOMIC); + + if (!keys) { + tcp_free_md5sig_pool(); + kfree(newkey); + return -ENOMEM; + } + + if (tp->md5sig_info->entries6) + memmove(keys, tp->md5sig_info->keys6, + (sizeof (tp->md5sig_info->keys6[0]) * + tp->md5sig_info->entries6)); + + kfree(tp->md5sig_info->keys6); + tp->md5sig_info->keys6 = keys; + tp->md5sig_info->alloced6++; + } + + ipv6_addr_copy(&tp->md5sig_info->keys6[tp->md5sig_info->entries6].addr, + peer); + tp->md5sig_info->keys6[tp->md5sig_info->entries6].key = newkey; + tp->md5sig_info->keys6[tp->md5sig_info->entries6].keylen = newkeylen; + + tp->md5sig_info->entries6++; + } + return 0; +} + +static int tcp_v6_md5_add_func(struct sock *sk, struct sock *addr_sk, + u8 *newkey, __u8 newkeylen) +{ + return tcp_v6_md5_do_add(sk, &inet6_sk(addr_sk)->daddr, + newkey, newkeylen); +} + +static int tcp_v6_md5_do_del(struct sock *sk, struct in6_addr *peer) +{ + struct tcp_sock *tp = tcp_sk(sk); + int i; + + for (i = 0; i < tp->md5sig_info->entries6; i++) { + if (ipv6_addr_cmp(&tp->md5sig_info->keys6[i].addr, peer) == 0) { + /* Free the key */ + kfree(tp->md5sig_info->keys6[i].key); + tp->md5sig_info->entries6--; + + if (tp->md5sig_info->entries6 == 0) { + kfree(tp->md5sig_info->keys6); + tp->md5sig_info->keys6 = NULL; + + tcp_free_md5sig_pool(); + + return 0; + } else { + /* shrink the database */ + if (tp->md5sig_info->entries6 != i) + memmove(&tp->md5sig_info->keys6[i], + &tp->md5sig_info->keys6[i+1], + (tp->md5sig_info->entries6 - i) + * sizeof (tp->md5sig_info->keys6[0])); + } + } + } + return -ENOENT; +} + +static void tcp_v6_clear_md5_list (struct sock *sk) +{ + struct tcp_sock *tp = tcp_sk(sk); + int i; + + if (tp->md5sig_info->entries6) { + for (i = 0; i < tp->md5sig_info->entries6; i++) + kfree(tp->md5sig_info->keys6[i].key); + tp->md5sig_info->entries6 = 0; + tcp_free_md5sig_pool(); + } + + kfree(tp->md5sig_info->keys6); + tp->md5sig_info->keys6 = NULL; + tp->md5sig_info->alloced6 = 0; + + if (tp->md5sig_info->entries4) { + for (i = 0; i < tp->md5sig_info->entries4; i++) + kfree(tp->md5sig_info->keys4[i].key); + tp->md5sig_info->entries4 = 0; + tcp_free_md5sig_pool(); + } + + kfree(tp->md5sig_info->keys4); + tp->md5sig_info->keys4 = NULL; + tp->md5sig_info->alloced4 = 0; +} + +static int tcp_v6_parse_md5_keys (struct sock *sk, char __user *optval, + int optlen) +{ + struct tcp_md5sig cmd; + struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)&cmd.tcpm_addr; + u8 *newkey; + + if (optlen < sizeof(cmd)) + return -EINVAL; + + if (copy_from_user(&cmd, optval, sizeof(cmd))) + return -EFAULT; + + if (sin6->sin6_family != AF_INET6) + return -EINVAL; + + if (!cmd.tcpm_keylen) { + if (!tcp_sk(sk)->md5sig_info) + return -ENOENT; + if (ipv6_addr_type(&sin6->sin6_addr) & IPV6_ADDR_MAPPED) + return tcp_v4_md5_do_del(sk, sin6->sin6_addr.s6_addr32[3]); + return tcp_v6_md5_do_del(sk, &sin6->sin6_addr); + } + + if (cmd.tcpm_keylen > TCP_MD5SIG_MAXKEYLEN) + return -EINVAL; + + if (!tcp_sk(sk)->md5sig_info) { + struct tcp_sock *tp = tcp_sk(sk); + struct tcp_md5sig_info *p; + + p = kzalloc(sizeof(struct tcp_md5sig_info), GFP_KERNEL); + if (!p) + return -ENOMEM; + + tp->md5sig_info = p; + } + + newkey = kmalloc(cmd.tcpm_keylen, GFP_KERNEL); + if (!newkey) + return -ENOMEM; + memcpy(newkey, cmd.tcpm_key, cmd.tcpm_keylen); + if (ipv6_addr_type(&sin6->sin6_addr) & IPV6_ADDR_MAPPED) { + return tcp_v4_md5_do_add(sk, sin6->sin6_addr.s6_addr32[3], + newkey, cmd.tcpm_keylen); + } + return tcp_v6_md5_do_add(sk, &sin6->sin6_addr, newkey, cmd.tcpm_keylen); +} + +static int tcp_v6_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key, + struct in6_addr *saddr, + struct in6_addr *daddr, + struct tcphdr *th, int protocol, + int tcplen) +{ + struct scatterlist sg[4]; + __u16 data_len; + int block = 0; + __u16 cksum; + struct tcp_md5sig_pool *hp; + struct tcp6_pseudohdr *bp; + struct hash_desc *desc; + int err; + unsigned int nbytes = 0; + + hp = tcp_get_md5sig_pool(); + if (!hp) { + printk(KERN_WARNING "%s(): hash pool not found...\n", __FUNCTION__); + goto clear_hash_noput; + } + bp = &hp->md5_blk.ip6; + desc = &hp->md5_desc; + + /* 1. TCP pseudo-header (RFC2460) */ + ipv6_addr_copy(&bp->saddr, saddr); + ipv6_addr_copy(&bp->daddr, daddr); + bp->len = htonl(tcplen); + bp->protocol = htonl(protocol); + + sg_set_buf(&sg[block++], bp, sizeof(*bp)); + nbytes += sizeof(*bp); + + /* 2. TCP header, excluding options */ + cksum = th->check; + th->check = 0; + sg_set_buf(&sg[block++], th, sizeof(*th)); + nbytes += sizeof(*th); + + /* 3. TCP segment data (if any) */ + data_len = tcplen - (th->doff << 2); + if (data_len > 0) { + u8 *data = (u8 *)th + (th->doff << 2); + sg_set_buf(&sg[block++], data, data_len); + nbytes += data_len; + } + + /* 4. shared key */ + sg_set_buf(&sg[block++], key->key, key->keylen); + nbytes += key->keylen; + + /* Now store the hash into the packet */ + err = crypto_hash_init(desc); + if (err) { + printk(KERN_WARNING "%s(): hash_init failed\n", __FUNCTION__); + goto clear_hash; + } + err = crypto_hash_update(desc, sg, nbytes); + if (err) { + printk(KERN_WARNING "%s(): hash_update failed\n", __FUNCTION__); + goto clear_hash; + } + err = crypto_hash_final(desc, md5_hash); + if (err) { + printk(KERN_WARNING "%s(): hash_final failed\n", __FUNCTION__); + goto clear_hash; + } + + /* Reset header, and free up the crypto */ + tcp_put_md5sig_pool(); + th->check = cksum; +out: + return 0; +clear_hash: + tcp_put_md5sig_pool(); +clear_hash_noput: + memset(md5_hash, 0, 16); + goto out; +} + +static int tcp_v6_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) +{ + struct in6_addr *saddr, *daddr; + + if (sk) { + saddr = &inet6_sk(sk)->saddr; + daddr = &inet6_sk(sk)->daddr; + } else { + saddr = &inet6_rsk(req)->loc_addr; + daddr = &inet6_rsk(req)->rmt_addr; + } + return tcp_v6_do_calc_md5_hash(md5_hash, key, + saddr, daddr, + th, protocol, tcplen); +} + +static int tcp_v6_inbound_md5_hash (struct sock *sk, struct sk_buff *skb) +{ + __u8 *hash_location = NULL; + struct tcp_md5sig_key *hash_expected; + struct ipv6hdr *ip6h = skb->nh.ipv6h; + struct tcphdr *th = skb->h.th; + int length = (th->doff << 2) - sizeof (*th); + int genhash; + u8 *ptr; + u8 newhash[16]; + + hash_expected = tcp_v6_md5_do_lookup(sk, &ip6h->saddr); + + /* If the TCP option is too short, we can short cut */ + if (length < TCPOLEN_MD5SIG) + return hash_expected ? 1 : 0; + + /* parse options */ + ptr = (u8*)(th + 1); + while (length > 0) { + int opcode = *ptr++; + int opsize; + + switch(opcode) { + case TCPOPT_EOL: + goto done_opts; + case TCPOPT_NOP: + length--; + continue; + default: + opsize = *ptr++; + if (opsize < 2 || opsize > length) + goto done_opts; + if (opcode == TCPOPT_MD5SIG) { + hash_location = ptr; + goto done_opts; + } + } + ptr += opsize - 2; + length -= opsize; + } + +done_opts: + /* do we have a hash as expected? */ + if (!hash_expected) { + if (!hash_location) + return 0; + if (net_ratelimit()) { + printk(KERN_INFO "MD5 Hash NOT expected but found " + "(" NIP6_FMT ", %u)->" + "(" NIP6_FMT ", %u)\n", + NIP6(ip6h->saddr), ntohs(th->source), + NIP6(ip6h->daddr), ntohs(th->dest)); + } + return 1; + } + + if (!hash_location) { + if (net_ratelimit()) { + printk(KERN_INFO "MD5 Hash expected but NOT found " + "(" NIP6_FMT ", %u)->" + "(" NIP6_FMT ", %u)\n", + NIP6(ip6h->saddr), ntohs(th->source), + NIP6(ip6h->daddr), ntohs(th->dest)); + } + return 1; + } + + /* check the signature */ + genhash = tcp_v6_do_calc_md5_hash(newhash, + hash_expected, + &ip6h->saddr, &ip6h->daddr, + th, sk->sk_protocol, + skb->len); + if (genhash || memcmp(hash_location, newhash, 16) != 0) { + if (net_ratelimit()) { + printk(KERN_INFO "MD5 Hash %s for " + "(" NIP6_FMT ", %u)->" + "(" NIP6_FMT ", %u)\n", + genhash ? "failed" : "mismatch", + NIP6(ip6h->saddr), ntohs(th->source), + NIP6(ip6h->daddr), ntohs(th->dest)); + } + return 1; + } + return 0; +} +#endif + static struct request_sock_ops tcp6_request_sock_ops __read_mostly = { .family = AF_INET6, .obj_size = sizeof(struct tcp6_request_sock), @@ -527,9 +928,16 @@ static struct request_sock_ops tcp6_request_sock_ops __read_mostly = { .send_reset = tcp_v6_send_reset }; +struct tcp_request_sock_ops tcp_request_sock_ipv6_ops = { +#ifdef CONFIG_TCP_MD5SIG + .md5_lookup = tcp_v6_reqsk_md5_lookup, +#endif +}; + static struct timewait_sock_ops tcp6_timewait_sock_ops = { .twsk_obj_size = sizeof(struct tcp6_timewait_sock), .twsk_unique = tcp_twsk_unique, + .twsk_destructor= tcp_twsk_destructor, }; static void tcp_v6_send_check(struct sock *sk, int len, struct sk_buff *skb) @@ -566,11 +974,15 @@ static int tcp_v6_gso_send_check(struct sk_buff *skb) return 0; } -static void tcp_v6_send_reset(struct sk_buff *skb) +static void tcp_v6_send_reset(struct sock *sk, struct sk_buff *skb) { struct tcphdr *th = skb->h.th, *t1; struct sk_buff *buff; struct flowi fl; + int tot_len = sizeof(*th); +#ifdef CONFIG_TCP_MD5SIG + struct tcp_md5sig_key *key; +#endif if (th->rst) return; @@ -578,25 +990,35 @@ static void tcp_v6_send_reset(struct sk_buff *skb) if (!ipv6_unicast_destination(skb)) return; +#ifdef CONFIG_TCP_MD5SIG + if (sk) + key = tcp_v6_md5_do_lookup(sk, &skb->nh.ipv6h->daddr); + else + key = NULL; + + if (key) + tot_len += TCPOLEN_MD5SIG_ALIGNED; +#endif + /* * We need to grab some memory, and put together an RST, * and then put it into the queue to be sent. */ - buff = alloc_skb(MAX_HEADER + sizeof(struct ipv6hdr) + sizeof(struct tcphdr), + buff = alloc_skb(MAX_HEADER + sizeof(struct ipv6hdr) + tot_len, GFP_ATOMIC); if (buff == NULL) return; - skb_reserve(buff, MAX_HEADER + sizeof(struct ipv6hdr) + sizeof(struct tcphdr)); + skb_reserve(buff, MAX_HEADER + sizeof(struct ipv6hdr) + tot_len); - t1 = (struct tcphdr *) skb_push(buff,sizeof(struct tcphdr)); + t1 = (struct tcphdr *) skb_push(buff, tot_len); /* Swap the send and the receive. */ memset(t1, 0, sizeof(*t1)); t1->dest = th->source; t1->source = th->dest; - t1->doff = sizeof(*t1)/4; + t1->doff = tot_len / 4; t1->rst = 1; if(th->ack) { @@ -607,6 +1029,22 @@ static void tcp_v6_send_reset(struct sk_buff *skb) + skb->len - (th->doff<<2)); } +#ifdef CONFIG_TCP_MD5SIG + if (key) { + u32 *opt = (u32*)(t1 + 1); + opt[0] = htonl((TCPOPT_NOP << 24) | + (TCPOPT_NOP << 16) | + (TCPOPT_MD5SIG << 8) | + TCPOLEN_MD5SIG); + tcp_v6_do_calc_md5_hash((__u8*)&opt[1], + key, + &skb->nh.ipv6h->daddr, + &skb->nh.ipv6h->saddr, + t1, IPPROTO_TCP, + tot_len); + } +#endif + buff->csum = csum_partial((char *)t1, sizeof(*t1), 0); memset(&fl, 0, sizeof(fl)); @@ -637,15 +1075,37 @@ static void tcp_v6_send_reset(struct sk_buff *skb) kfree_skb(buff); } -static void tcp_v6_send_ack(struct sk_buff *skb, u32 seq, u32 ack, u32 win, u32 ts) +static void tcp_v6_send_ack(struct tcp_timewait_sock *tw, + struct sk_buff *skb, u32 seq, u32 ack, u32 win, u32 ts) { struct tcphdr *th = skb->h.th, *t1; struct sk_buff *buff; struct flowi fl; int tot_len = sizeof(struct tcphdr); + u32 *topt; +#ifdef CONFIG_TCP_MD5SIG + struct tcp_md5sig_key *key; + struct tcp_md5sig_key tw_key; +#endif + +#ifdef CONFIG_TCP_MD5SIG + if (!tw && skb->sk) { + key = tcp_v6_md5_do_lookup(skb->sk, &skb->nh.ipv6h->daddr); + } else if (tw && tw->tw_md5_keylen) { + tw_key.key = tw->tw_md5_key; + tw_key.keylen = tw->tw_md5_keylen; + key = &tw_key; + } else { + key = NULL; + } +#endif if (ts) tot_len += TCPOLEN_TSTAMP_ALIGNED; +#ifdef CONFIG_TCP_MD5SIG + if (key) + tot_len += TCPOLEN_MD5SIG_ALIGNED; +#endif buff = alloc_skb(MAX_HEADER + sizeof(struct ipv6hdr) + tot_len, GFP_ATOMIC); @@ -665,15 +1125,29 @@ static void tcp_v6_send_ack(struct sk_buff *skb, u32 seq, u32 ack, u32 win, u32 t1->ack_seq = htonl(ack); t1->ack = 1; t1->window = htons(win); + + topt = (u32*)(t1 + 1); if (ts) { - u32 *ptr = (u32*)(t1 + 1); - *ptr++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) | - (TCPOPT_TIMESTAMP << 8) | TCPOLEN_TIMESTAMP); - *ptr++ = htonl(tcp_time_stamp); - *ptr = htonl(ts); + *topt++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) | + (TCPOPT_TIMESTAMP << 8) | TCPOLEN_TIMESTAMP); + *topt++ = htonl(tcp_time_stamp); + *topt = htonl(ts); } +#ifdef CONFIG_TCP_MD5SIG + if (key) { + *topt++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) | + (TCPOPT_MD5SIG << 8) | TCPOLEN_MD5SIG); + tcp_v6_do_calc_md5_hash((__u8 *)topt, + key, + &skb->nh.ipv6h->daddr, + &skb->nh.ipv6h->saddr, + t1, IPPROTO_TCP, + tot_len); + } +#endif + buff->csum = csum_partial((char *)t1, tot_len, 0); memset(&fl, 0, sizeof(fl)); @@ -704,9 +1178,9 @@ static void tcp_v6_send_ack(struct sk_buff *skb, u32 seq, u32 ack, u32 win, u32 static void tcp_v6_timewait_ack(struct sock *sk, struct sk_buff *skb) { struct inet_timewait_sock *tw = inet_twsk(sk); - const struct tcp_timewait_sock *tcptw = tcp_twsk(sk); + struct tcp_timewait_sock *tcptw = tcp_twsk(sk); - tcp_v6_send_ack(skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt, + tcp_v6_send_ack(tcptw, skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt, tcptw->tw_rcv_wnd >> tw->tw_rcv_wscale, tcptw->tw_ts_recent); @@ -715,7 +1189,7 @@ static void tcp_v6_timewait_ack(struct sock *sk, struct sk_buff *skb) static void tcp_v6_reqsk_send_ack(struct sk_buff *skb, struct request_sock *req) { - tcp_v6_send_ack(skb, tcp_rsk(req)->snt_isn + 1, tcp_rsk(req)->rcv_isn + 1, req->rcv_wnd, req->ts_recent); + tcp_v6_send_ack(NULL, skb, tcp_rsk(req)->snt_isn + 1, tcp_rsk(req)->rcv_isn + 1, req->rcv_wnd, req->ts_recent); } @@ -786,6 +1260,10 @@ static int tcp_v6_conn_request(struct sock *sk, struct sk_buff *skb) if (req == NULL) goto drop; +#ifdef CONFIG_TCP_MD5SIG + tcp_rsk(req)->af_specific = &tcp_request_sock_ipv6_ops; +#endif + tcp_clear_options(&tmp_opt); tmp_opt.mss_clamp = IPV6_MIN_MTU - sizeof(struct tcphdr) - sizeof(struct ipv6hdr); tmp_opt.user_mss = tp->rx_opt.user_mss; @@ -844,6 +1322,9 @@ static struct sock * tcp_v6_syn_recv_sock(struct sock *sk, struct sk_buff *skb, struct tcp_sock *newtp; struct sock *newsk; struct ipv6_txoptions *opt; +#ifdef CONFIG_TCP_MD5SIG + struct tcp_md5sig_key *key; +#endif if (skb->protocol == htons(ETH_P_IP)) { /* @@ -874,6 +1355,10 @@ static struct sock * tcp_v6_syn_recv_sock(struct sock *sk, struct sk_buff *skb, inet_csk(newsk)->icsk_af_ops = &ipv6_mapped; newsk->sk_backlog_rcv = tcp_v4_do_rcv; +#ifdef CONFIG_TCP_MD5SIG + newtp->af_specific = &tcp_sock_ipv6_mapped_specific; +#endif + newnp->pktoptions = NULL; newnp->opt = NULL; newnp->mcast_oif = inet6_iif(skb); @@ -1008,6 +1493,23 @@ static struct sock * tcp_v6_syn_recv_sock(struct sock *sk, struct sk_buff *skb, newinet->daddr = newinet->saddr = newinet->rcv_saddr = LOOPBACK4_IPV6; +#ifdef CONFIG_TCP_MD5SIG + /* Copy over the MD5 key from the original socket */ + if ((key = tcp_v6_md5_do_lookup(sk, &newnp->daddr)) != 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) { + memcpy(newkey, key->key, key->keylen); + tcp_v6_md5_do_add(newsk, &inet6_sk(sk)->daddr, + newkey, key->keylen); + } + } +#endif + __inet6_hash(&tcp_hashinfo, newsk); inet_inherit_port(&tcp_hashinfo, sk, newsk); @@ -1067,6 +1569,11 @@ static int tcp_v6_do_rcv(struct sock *sk, struct sk_buff *skb) if (skb->protocol == htons(ETH_P_IP)) return tcp_v4_do_rcv(sk, skb); +#ifdef CONFIG_TCP_MD5SIG + if (tcp_v6_inbound_md5_hash (sk, skb)) + goto discard; +#endif + if (sk_filter(sk, skb)) goto discard; @@ -1132,7 +1639,7 @@ static int tcp_v6_do_rcv(struct sock *sk, struct sk_buff *skb) return 0; reset: - tcp_v6_send_reset(skb); + tcp_v6_send_reset(sk, skb); discard: if (opt_skb) __kfree_skb(opt_skb); @@ -1257,7 +1764,7 @@ no_tcp_socket: bad_packet: TCP_INC_STATS_BH(TCP_MIB_INERRS); } else { - tcp_v6_send_reset(skb); + tcp_v6_send_reset(NULL, skb); } discard_it: @@ -1336,6 +1843,15 @@ static struct inet_connection_sock_af_ops ipv6_specific = { #endif }; +static struct tcp_sock_af_ops tcp_sock_ipv6_specific = { +#ifdef CONFIG_TCP_MD5SIG + .md5_lookup = tcp_v6_md5_lookup, + .calc_md5_hash = tcp_v6_calc_md5_hash, + .md5_add = tcp_v6_md5_add_func, + .md5_parse = tcp_v6_parse_md5_keys, +#endif +}; + /* * TCP over IPv4 via INET6 API */ @@ -1358,6 +1874,15 @@ static struct inet_connection_sock_af_ops ipv6_mapped = { #endif }; +static struct tcp_sock_af_ops tcp_sock_ipv6_mapped_specific = { +#ifdef CONFIG_TCP_MD5SIG + .md5_lookup = tcp_v4_md5_lookup, + .calc_md5_hash = tcp_v4_calc_md5_hash, + .md5_add = tcp_v6_md5_add_func, + .md5_parse = tcp_v6_parse_md5_keys, +#endif +}; + /* NOTE: A lot of things set to zero explicitly by call to * sk_alloc() so need not be done here. */ @@ -1397,6 +1922,10 @@ static int tcp_v6_init_sock(struct sock *sk) sk->sk_write_space = sk_stream_write_space; sock_set_flag(sk, SOCK_USE_WRITE_QUEUE); +#ifdef CONFIG_TCP_MD5SIG + tp->af_specific = &tcp_sock_ipv6_specific; +#endif + sk->sk_sndbuf = sysctl_tcp_wmem[1]; sk->sk_rcvbuf = sysctl_tcp_rmem[1]; @@ -1407,6 +1936,11 @@ static int tcp_v6_init_sock(struct sock *sk) static int tcp_v6_destroy_sock(struct sock *sk) { +#ifdef CONFIG_TCP_MD5SIG + /* Clean up the MD5 key list */ + if (tcp_sk(sk)->md5sig_info) + tcp_v6_clear_md5_list(sk); +#endif tcp_v4_destroy_sock(sk); return inet6_destroy_sock(sk); }