X-Git-Url: https://err.no/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=net%2Fipv4%2Ftcp_ipv4.c;h=29adc668ad51d87555330d653a9315c27fc9d26b;hb=ed88098e25d77bef3b2ad8c9d8e2ebf454d9ccbf;hp=cd601a866c2f5856e3f8b69ffc3fd542c465e34a;hpb=38e80121bd7d0c493072442ac7eddcba165a07a8;p=linux-2.6

diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index cd601a866c..29adc668ad 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -5,8 +5,6 @@
  *
  *		Implementation of the Transmission Control Protocol(TCP).
  *
- * Version:	$Id: tcp_ipv4.c,v 1.240 2002/02/01 22:01:04 davem Exp $
- *
  *		IPv4 specific functions
  *
  *
@@ -85,18 +83,19 @@
 int sysctl_tcp_tw_reuse __read_mostly;
 int sysctl_tcp_low_latency __read_mostly;
 
-/* Check TCP sequence numbers in ICMP packets. */
-#define ICMP_MIN_LENGTH 8
-
-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,
-				   unsigned int tcplen);
+				   struct tcphdr *th, unsigned int tcplen);
+#else
+static inline
+struct tcp_md5sig_key *tcp_v4_md5_do_lookup(struct sock *sk, __be32 addr)
+{
+	return NULL;
+}
 #endif
 
 struct inet_hashinfo __cacheline_aligned tcp_hashinfo = {
@@ -176,7 +175,7 @@ int tcp_v4_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)
 			       inet->sport, usin->sin_port, sk, 1);
 	if (tmp < 0) {
 		if (tmp == -ENETUNREACH)
-			IP_INC_STATS_BH(IPSTATS_MIB_OUTNOROUTES);
+			IP_INC_STATS_BH(sock_net(sk), IPSTATS_MIB_OUTNOROUTES);
 		return tmp;
 	}
 
@@ -344,16 +343,17 @@ void tcp_v4_err(struct sk_buff *skb, u32 info)
 	struct sock *sk;
 	__u32 seq;
 	int err;
+	struct net *net = dev_net(skb->dev);
 
 	if (skb->len < (iph->ihl << 2) + 8) {
-		ICMP_INC_STATS_BH(ICMP_MIB_INERRORS);
+		ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS);
 		return;
 	}
 
-	sk = inet_lookup(dev_net(skb->dev), &tcp_hashinfo, iph->daddr, th->dest,
+	sk = inet_lookup(net, &tcp_hashinfo, iph->daddr, th->dest,
 			iph->saddr, th->source, inet_iif(skb));
 	if (!sk) {
-		ICMP_INC_STATS_BH(ICMP_MIB_INERRORS);
+		ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS);
 		return;
 	}
 	if (sk->sk_state == TCP_TIME_WAIT) {
@@ -366,7 +366,7 @@ void tcp_v4_err(struct sk_buff *skb, u32 info)
 	 * servers this needs to be solved differently.
 	 */
 	if (sock_owned_by_user(sk))
-		NET_INC_STATS_BH(LINUX_MIB_LOCKDROPPEDICMPS);
+		NET_INC_STATS_BH(net, LINUX_MIB_LOCKDROPPEDICMPS);
 
 	if (sk->sk_state == TCP_CLOSE)
 		goto out;
@@ -375,7 +375,7 @@ void tcp_v4_err(struct sk_buff *skb, u32 info)
 	seq = ntohl(th->seq);
 	if (sk->sk_state != TCP_LISTEN &&
 	    !between(seq, tp->snd_una, tp->snd_nxt)) {
-		NET_INC_STATS_BH(LINUX_MIB_OUTOFWINDOWICMPS);
+		NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS);
 		goto out;
 	}
 
@@ -422,7 +422,7 @@ void tcp_v4_err(struct sk_buff *skb, u32 info)
 		BUG_TRAP(!req->sk);
 
 		if (seq != tcp_rsk(req)->snt_isn) {
-			NET_INC_STATS_BH(LINUX_MIB_OUTOFWINDOWICMPS);
+			NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS);
 			goto out;
 		}
 
@@ -544,6 +544,7 @@ static void tcp_v4_send_reset(struct sock *sk, struct sk_buff *skb)
 #ifdef CONFIG_TCP_MD5SIG
 	struct tcp_md5sig_key *key;
 #endif
+	struct net *net;
 
 	/* Never send a reset in response to a reset. */
 	if (th->rst)
@@ -586,8 +587,7 @@ static void tcp_v4_send_reset(struct sock *sk, struct sk_buff *skb)
 					key,
 					ip_hdr(skb)->daddr,
 					ip_hdr(skb)->saddr,
-					&rep.th, IPPROTO_TCP,
-					arg.iov[0].iov_len);
+					&rep.th, arg.iov[0].iov_len);
 	}
 #endif
 	arg.csum = csum_tcpudp_nofold(ip_hdr(skb)->daddr,
@@ -595,20 +595,21 @@ static void tcp_v4_send_reset(struct sock *sk, struct sk_buff *skb)
 				      sizeof(struct tcphdr), IPPROTO_TCP, 0);
 	arg.csumoffset = offsetof(struct tcphdr, check) / 2;
 
-	ip_send_reply(dev_net(skb->dst->dev)->ipv4.tcp_sock, skb,
+	net = dev_net(skb->dst->dev);
+	ip_send_reply(net->ipv4.tcp_sock, skb,
 		      &arg, arg.iov[0].iov_len);
 
-	TCP_INC_STATS_BH(TCP_MIB_OUTSEGS);
-	TCP_INC_STATS_BH(TCP_MIB_OUTRSTS);
+	TCP_INC_STATS_BH(net, TCP_MIB_OUTSEGS);
+	TCP_INC_STATS_BH(net, TCP_MIB_OUTRSTS);
 }
 
 /* The code following below sending ACKs in SYN-RECV and TIME-WAIT states
    outside socket context is ugly, certainly. What can I do?
  */
 
-static void tcp_v4_send_ack(struct tcp_timewait_sock *twsk,
-			    struct sk_buff *skb, u32 seq, u32 ack,
-			    u32 win, u32 ts)
+static void tcp_v4_send_ack(struct sk_buff *skb, u32 seq, u32 ack,
+			    u32 win, u32 ts, int oif,
+			    struct tcp_md5sig_key *key)
 {
 	struct tcphdr *th = tcp_hdr(skb);
 	struct {
@@ -620,10 +621,7 @@ static void tcp_v4_send_ack(struct tcp_timewait_sock *twsk,
 			];
 	} rep;
 	struct ip_reply_arg arg;
-#ifdef CONFIG_TCP_MD5SIG
-	struct tcp_md5sig_key *key;
-	struct tcp_md5sig_key tw_key;
-#endif
+	struct net *net = dev_net(skb->dev);
 
 	memset(&rep.th, 0, sizeof(struct tcphdr));
 	memset(&arg, 0, sizeof(arg));
@@ -649,23 +647,6 @@ static void tcp_v4_send_ack(struct tcp_timewait_sock *twsk,
 	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, ip_hdr(skb)->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;
 
@@ -680,21 +661,20 @@ static void tcp_v4_send_ack(struct tcp_timewait_sock *twsk,
 					key,
 					ip_hdr(skb)->daddr,
 					ip_hdr(skb)->saddr,
-					&rep.th, IPPROTO_TCP,
-					arg.iov[0].iov_len);
+					&rep.th, arg.iov[0].iov_len);
 	}
 #endif
 	arg.csum = csum_tcpudp_nofold(ip_hdr(skb)->daddr,
 				      ip_hdr(skb)->saddr, /* XXX */
 				      arg.iov[0].iov_len, IPPROTO_TCP, 0);
 	arg.csumoffset = offsetof(struct tcphdr, check) / 2;
-	if (twsk)
-		arg.bound_dev_if = twsk->tw_sk.tw_bound_dev_if;
+	if (oif)
+		arg.bound_dev_if = oif;
 
-	ip_send_reply(dev_net(skb->dev)->ipv4.tcp_sock, skb,
+	ip_send_reply(net->ipv4.tcp_sock, skb,
 		      &arg, arg.iov[0].iov_len);
 
-	TCP_INC_STATS_BH(TCP_MIB_OUTSEGS);
+	TCP_INC_STATS_BH(net, TCP_MIB_OUTSEGS);
 }
 
 static void tcp_v4_timewait_ack(struct sock *sk, struct sk_buff *skb)
@@ -702,9 +682,12 @@ static void tcp_v4_timewait_ack(struct sock *sk, struct sk_buff *skb)
 	struct inet_timewait_sock *tw = inet_twsk(sk);
 	struct tcp_timewait_sock *tcptw = tcp_twsk(sk);
 
-	tcp_v4_send_ack(tcptw, skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt,
+	tcp_v4_send_ack(skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt,
 			tcptw->tw_rcv_wnd >> tw->tw_rcv_wscale,
-			tcptw->tw_ts_recent);
+			tcptw->tw_ts_recent,
+			tw->tw_bound_dev_if,
+			tcp_twsk_md5_key(tcptw)
+			);
 
 	inet_twsk_put(tw);
 }
@@ -712,9 +695,11 @@ 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(NULL, skb, tcp_rsk(req)->snt_isn + 1,
+	tcp_v4_send_ack(skb, tcp_rsk(req)->snt_isn + 1,
 			tcp_rsk(req)->rcv_isn + 1, req->rcv_wnd,
-			req->ts_recent);
+			req->ts_recent,
+			0,
+			tcp_v4_md5_do_lookup(skb->sk, ip_hdr(skb)->daddr));
 }
 
 /*
@@ -1006,18 +991,12 @@ static int tcp_v4_parse_md5_keys(struct sock *sk, char __user *optval,
 
 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,
+				   struct tcphdr *th,
 				   unsigned int tcplen)
 {
-	struct scatterlist sg[4];
-	__u16 data_len;
-	int block = 0;
-	__sum16 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,
@@ -1029,63 +1008,25 @@ static int tcp_v4_do_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key,
 		goto clear_hash_noput;
 
 	bp = &hp->md5_blk.ip4;
-	desc = &hp->md5_desc;
 
 	/*
-	 * 1. the TCP pseudo-header (in the order: source IP address,
+	 * 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->protocol = IPPROTO_TCP;
 	bp->len = htons(tcplen);
 
-	sg_init_table(sg, 4);
-
-	sg_set_buf(&sg[block++], bp, sizeof(*bp));
-	nbytes += sizeof(*bp);
-
-	/* 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);
-
-	/* 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;
-
-	sg_mark_end(&sg[block - 1]);
-
-	/* 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);
+	err = tcp_calc_md5_hash(md5_hash, key, sizeof(*bp),
+				th, tcplen, hp);
 	if (err)
 		goto clear_hash;
 
-	/* Reset header, and free up the crypto */
+	/* Free up the crypto pool */
 	tcp_put_md5sig_pool();
-	th->check = old_checksum;
-
 out:
 	return 0;
 clear_hash:
@@ -1099,7 +1040,7 @@ 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,
+			 struct tcphdr *th,
 			 unsigned int tcplen)
 {
 	__be32 saddr, daddr;
@@ -1115,7 +1056,7 @@ int tcp_v4_calc_md5_hash(char *md5_hash, struct tcp_md5sig_key *key,
 	}
 	return tcp_v4_do_calc_md5_hash(md5_hash, key,
 				       saddr, daddr,
-				       th, protocol, tcplen);
+				       th, tcplen);
 }
 
 EXPORT_SYMBOL(tcp_v4_calc_md5_hash);
@@ -1134,52 +1075,12 @@ static int tcp_v4_inbound_md5_hash(struct sock *sk, struct sk_buff *skb)
 	struct tcp_md5sig_key *hash_expected;
 	const struct iphdr *iph = ip_hdr(skb);
 	struct tcphdr *th = tcp_hdr(skb);
-	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);
+	hash_location = tcp_parse_md5sig_option(th);
 
-	/*
-	 * 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;
@@ -1206,8 +1107,7 @@ done_opts:
 	genhash = tcp_v4_do_calc_md5_hash(newhash,
 					  hash_expected,
 					  iph->saddr, iph->daddr,
-					  th, sk->sk_protocol,
-					  skb->len);
+					  th, skb->len);
 
 	if (genhash || memcmp(hash_location, newhash, 16) != 0) {
 		if (net_ratelimit()) {
@@ -1285,7 +1185,7 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
 	if (sk_acceptq_is_full(sk) && inet_csk_reqsk_queue_young(sk) > 1)
 		goto drop;
 
-	req = reqsk_alloc(&tcp_request_sock_ops);
+	req = inet_reqsk_alloc(&tcp_request_sock_ops);
 	if (!req)
 		goto drop;
 
@@ -1351,7 +1251,7 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
 			if (get_seconds() < peer->tcp_ts_stamp + TCP_PAWS_MSL &&
 			    (s32)(peer->tcp_ts - req->ts_recent) >
 							TCP_PAWS_WINDOW) {
-				NET_INC_STATS_BH(LINUX_MIB_PAWSPASSIVEREJECTED);
+				NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_PAWSPASSIVEREJECTED);
 				goto drop_and_release;
 			}
 		}
@@ -1465,9 +1365,9 @@ struct sock *tcp_v4_syn_recv_sock(struct sock *sk, struct sk_buff *skb,
 	return newsk;
 
 exit_overflow:
-	NET_INC_STATS_BH(LINUX_MIB_LISTENOVERFLOWS);
+	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENOVERFLOWS);
 exit:
-	NET_INC_STATS_BH(LINUX_MIB_LISTENDROPS);
+	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS);
 	dst_release(dst);
 	return NULL;
 }
@@ -1594,7 +1494,7 @@ discard:
 	return 0;
 
 csum_err:
-	TCP_INC_STATS_BH(TCP_MIB_INERRS);
+	TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_INERRS);
 	goto discard;
 }
 
@@ -1608,12 +1508,13 @@ int tcp_v4_rcv(struct sk_buff *skb)
 	struct tcphdr *th;
 	struct sock *sk;
 	int ret;
+	struct net *net = dev_net(skb->dev);
 
 	if (skb->pkt_type != PACKET_HOST)
 		goto discard_it;
 
 	/* Count it even if it's bad */
-	TCP_INC_STATS_BH(TCP_MIB_INSEGS);
+	TCP_INC_STATS_BH(net, TCP_MIB_INSEGS);
 
 	if (!pskb_may_pull(skb, sizeof(struct tcphdr)))
 		goto discard_it;
@@ -1642,7 +1543,7 @@ int tcp_v4_rcv(struct sk_buff *skb)
 	TCP_SKB_CB(skb)->flags	 = iph->tos;
 	TCP_SKB_CB(skb)->sacked	 = 0;
 
-	sk = __inet_lookup(dev_net(skb->dev), &tcp_hashinfo, iph->saddr,
+	sk = __inet_lookup(net, &tcp_hashinfo, iph->saddr,
 			th->source, iph->daddr, th->dest, inet_iif(skb));
 	if (!sk)
 		goto no_tcp_socket;
@@ -1689,7 +1590,7 @@ no_tcp_socket:
 
 	if (skb->len < (th->doff << 2) || tcp_checksum_complete(skb)) {
 bad_packet:
-		TCP_INC_STATS_BH(TCP_MIB_INERRS);
+		TCP_INC_STATS_BH(net, TCP_MIB_INERRS);
 	} else {
 		tcp_v4_send_reset(NULL, skb);
 	}
@@ -1710,7 +1611,7 @@ do_time_wait:
 	}
 
 	if (skb->len < (th->doff << 2) || tcp_checksum_complete(skb)) {
-		TCP_INC_STATS_BH(TCP_MIB_INERRS);
+		TCP_INC_STATS_BH(net, TCP_MIB_INERRS);
 		inet_twsk_put(inet_twsk(sk));
 		goto discard_it;
 	}
@@ -1875,7 +1776,7 @@ static int tcp_v4_init_sock(struct sock *sk)
 	return 0;
 }
 
-int tcp_v4_destroy_sock(struct sock *sk)
+void tcp_v4_destroy_sock(struct sock *sk)
 {
 	struct tcp_sock *tp = tcp_sk(sk);
 
@@ -1918,17 +1819,7 @@ int tcp_v4_destroy_sock(struct sock *sk)
 		sk->sk_sndmsg_page = NULL;
 	}
 
-	if (tp->defer_tcp_accept.request) {
-		reqsk_free(tp->defer_tcp_accept.request);
-		sock_put(tp->defer_tcp_accept.listen_sk);
-		sock_put(sk);
-		tp->defer_tcp_accept.listen_sk = NULL;
-		tp->defer_tcp_accept.request = NULL;
-	}
-
 	atomic_dec(&tcp_sockets_allocated);
-
-	return 0;
 }
 
 EXPORT_SYMBOL(tcp_v4_destroy_sock);
@@ -2303,7 +2194,7 @@ static void get_tcp4_sock(struct sock *sk, struct seq_file *f, int i, int *len)
 	}
 
 	seq_printf(f, "%4d: %08X:%04X %08X:%04X %02X %08X:%08X %02X:%08lX "
-			"%08X %5d %8d %lu %d %p %u %u %u %u %d%n",
+			"%08X %5d %8d %lu %d %p %lu %lu %u %u %d%n",
 		i, src, srcp, dest, destp, sk->sk_state,
 		tp->write_seq - tp->snd_una,
 		sk->sk_state == TCP_LISTEN ? sk->sk_ack_backlog :
@@ -2315,8 +2206,8 @@ static void get_tcp4_sock(struct sock *sk, struct seq_file *f, int i, int *len)
 		icsk->icsk_probes_out,
 		sock_i_ino(sk),
 		atomic_read(&sk->sk_refcnt), sk,
-		icsk->icsk_rto,
-		icsk->icsk_ack.ato,
+		jiffies_to_clock_t(icsk->icsk_rto),
+		jiffies_to_clock_t(icsk->icsk_ack.ato),
 		(icsk->icsk_ack.quick << 1) | icsk->icsk_ack.pingpong,
 		tp->snd_cwnd,
 		tp->snd_ssthresh >= 0xFFFF ? -1 : tp->snd_ssthresh,