tcp_ipv4.c

Linux内核源代码 为压缩文件 是<<Linux内核>>一书中的源代码

/*
 * INET		An implementation of the TCP/IP protocol suite for the LINUX
 *		operating system.  INET is implemented using the  BSD Socket
 *		interface as the means of communication with the user level.
 *
 *		Implementation of the Transmission Control Protocol(TCP).
 *
 * Version:	$Id: tcp_ipv4.c,v 1.222 2000/12/08 17:15:53 davem Exp $
 *
 *		IPv4 specific functions
 *
 *
 *		code split from:
 *		linux/ipv4/tcp.c
 *		linux/ipv4/tcp_input.c
 *		linux/ipv4/tcp_output.c
 *
 *		See tcp.c for author information
 *
 *	This program is free software; you can redistribute it and/or
 *      modify it under the terms of the GNU General Public License
 *      as published by the Free Software Foundation; either version
 *      2 of the License, or (at your option) any later version.
 */

/*
 * Changes:
 *		David S. Miller	:	New socket lookup architecture.
 *					This code is dedicated to John Dyson.
 *		David S. Miller :	Change semantics of established hash,
 *					half is devoted to TIME_WAIT sockets
 *					and the rest go in the other half.
 *		Andi Kleen :		Add support for syncookies and fixed
 *					some bugs: ip options weren't passed to
 *					the TCP layer, missed a check for an ACK bit.
 *		Andi Kleen :		Implemented fast path mtu discovery.
 *	     				Fixed many serious bugs in the
 *					open_request handling and moved
 *					most of it into the af independent code.
 *					Added tail drop and some other bugfixes.
 *					Added new listen sematics.
 *		Mike McLagan	:	Routing by source
 *	Juan Jose Ciarlante:		ip_dynaddr bits
 *		Andi Kleen:		various fixes.
 *	Vitaly E. Lavrov	:	Transparent proxy revived after year coma.
 *	Andi Kleen		:	Fix new listen.
 *	Andi Kleen		:	Fix accept error reporting.
 */

#include <linux/config.h>
#include <linux/types.h>
#include <linux/fcntl.h>
#include <linux/random.h>
#include <linux/cache.h>
#include <linux/init.h>

#include <net/icmp.h>
#include <net/tcp.h>
#include <net/ipv6.h>
#include <net/inet_common.h>

#include <linux/inet.h>
#include <linux/stddef.h>
#include <linux/ipsec.h>

extern int sysctl_ip_dynaddr;

/* Check TCP sequence numbers in ICMP packets. */
#define ICMP_MIN_LENGTH 8

/* Socket used for sending RSTs */ 	
static struct inode tcp_inode;
static struct socket *tcp_socket=&tcp_inode.u.socket_i;

void tcp_v4_send_check(struct sock *sk, struct tcphdr *th, int len, 
		       struct sk_buff *skb);

/*
 * ALL members must be initialised to prevent gcc-2.7.2.3 miscompilation
 */
struct tcp_hashinfo __cacheline_aligned tcp_hashinfo = {
	__tcp_ehash:          NULL,
	__tcp_bhash:          NULL,
	__tcp_bhash_size:     0,
	__tcp_ehash_size:     0,
	__tcp_listening_hash: { NULL, },
	__tcp_lhash_lock:     RW_LOCK_UNLOCKED,
	__tcp_lhash_users:    ATOMIC_INIT(0),
	__tcp_lhash_wait:
	  __WAIT_QUEUE_HEAD_INITIALIZER(tcp_hashinfo.__tcp_lhash_wait),
	__tcp_portalloc_lock: SPIN_LOCK_UNLOCKED
};

/*
 * This array holds the first and last local port number.
 * For high-usage systems, use sysctl to change this to
 * 32768-61000
 */
int sysctl_local_port_range[2] = { 1024, 4999 };
int tcp_port_rover = (1024 - 1);

static __inline__ int tcp_hashfn(__u32 laddr, __u16 lport,
				 __u32 faddr, __u16 fport)
{
	int h = ((laddr ^ lport) ^ (faddr ^ fport));
	h ^= h>>16;
	h ^= h>>8;
	return h & (tcp_ehash_size - 1);
}

static __inline__ int tcp_sk_hashfn(struct sock *sk)
{
	__u32 laddr = sk->rcv_saddr;
	__u16 lport = sk->num;
	__u32 faddr = sk->daddr;
	__u16 fport = sk->dport;

	return tcp_hashfn(laddr, lport, faddr, fport);
}

/* Allocate and initialize a new TCP local port bind bucket.
 * The bindhash mutex for snum's hash chain must be held here.
 */
struct tcp_bind_bucket *tcp_bucket_create(struct tcp_bind_hashbucket *head,
					  unsigned short snum)
{
	struct tcp_bind_bucket *tb;

	tb = kmem_cache_alloc(tcp_bucket_cachep, SLAB_ATOMIC);
	if(tb != NULL) {
		tb->port = snum;
		tb->fastreuse = 0;
		tb->owners = NULL;
		if((tb->next = head->chain) != NULL)
			tb->next->pprev = &tb->next;
		head->chain = tb;
		tb->pprev = &head->chain;
	}
	return tb;
}

/* Caller must disable local BH processing. */
static __inline__ void __tcp_inherit_port(struct sock *sk, struct sock *child)
{
	struct tcp_bind_hashbucket *head = &tcp_bhash[tcp_bhashfn(child->num)];
	struct tcp_bind_bucket *tb;

	spin_lock(&head->lock);
	tb = (struct tcp_bind_bucket *)sk->prev;
	if ((child->bind_next = tb->owners) != NULL)
		tb->owners->bind_pprev = &child->bind_next;
	tb->owners = child;
	child->bind_pprev = &tb->owners;
	child->prev = (struct sock *) tb;
	spin_unlock(&head->lock);
}

__inline__ void tcp_inherit_port(struct sock *sk, struct sock *child)
{
	local_bh_disable();
	__tcp_inherit_port(sk, child);
	local_bh_enable();
}

/* Obtain a reference to a local port for the given sock,
 * if snum is zero it means select any available local port.
 */
static int tcp_v4_get_port(struct sock *sk, unsigned short snum)
{
	struct tcp_bind_hashbucket *head;
	struct tcp_bind_bucket *tb;
	int ret;

	local_bh_disable();
	if (snum == 0) {
		int low = sysctl_local_port_range[0];
		int high = sysctl_local_port_range[1];
		int remaining = (high - low) + 1;
		int rover;

		spin_lock(&tcp_portalloc_lock);
		rover = tcp_port_rover;
		do {	rover++;
			if ((rover < low) || (rover > high))
				rover = low;
			head = &tcp_bhash[tcp_bhashfn(rover)];
			spin_lock(&head->lock);
			for (tb = head->chain; tb; tb = tb->next)
				if (tb->port == rover)
					goto next;
			break;
		next:
			spin_unlock(&head->lock);
		} while (--remaining > 0);
		tcp_port_rover = rover;
		spin_unlock(&tcp_portalloc_lock);

		/* Exhausted local port range during search? */
		ret = 1;
		if (remaining <= 0)
			goto fail;

		/* OK, here is the one we will use.  HEAD is
		 * non-NULL and we hold it's mutex.
		 */
		snum = rover;
		tb = NULL;
	} else {
		head = &tcp_bhash[tcp_bhashfn(snum)];
		spin_lock(&head->lock);
		for (tb = head->chain; tb != NULL; tb = tb->next)
			if (tb->port == snum)
				break;
	}
	if (tb != NULL && tb->owners != NULL) {
		if (tb->fastreuse != 0 && sk->reuse != 0 && sk->state != TCP_LISTEN) {
			goto success;
		} else {
			struct sock *sk2 = tb->owners;
			int sk_reuse = sk->reuse;

			for( ; sk2 != NULL; sk2 = sk2->bind_next) {
				if (sk != sk2 &&
				    sk->bound_dev_if == sk2->bound_dev_if) {
					if (!sk_reuse	||
					    !sk2->reuse	||
					    sk2->state == TCP_LISTEN) {
						if (!sk2->rcv_saddr	||
						    !sk->rcv_saddr	||
						    (sk2->rcv_saddr == sk->rcv_saddr))
							break;
					}
				}
			}
			/* If we found a conflict, fail. */
			ret = 1;
			if (sk2 != NULL)
				goto fail_unlock;
		}
	}
	ret = 1;
	if (tb == NULL &&
	    (tb = tcp_bucket_create(head, snum)) == NULL)
			goto fail_unlock;
	if (tb->owners == NULL) {
		if (sk->reuse && sk->state != TCP_LISTEN)
			tb->fastreuse = 1;
		else
			tb->fastreuse = 0;
	} else if (tb->fastreuse &&
		   ((sk->reuse == 0) || (sk->state == TCP_LISTEN)))
		tb->fastreuse = 0;
success:
	sk->num = snum;
	if (sk->prev == NULL) {
		if ((sk->bind_next = tb->owners) != NULL)
			tb->owners->bind_pprev = &sk->bind_next;
		tb->owners = sk;
		sk->bind_pprev = &tb->owners;
		sk->prev = (struct sock *) tb;
	} else {
		BUG_TRAP(sk->prev == (struct sock *) tb);
	}
	ret = 0;

fail_unlock:
	spin_unlock(&head->lock);
fail:
	local_bh_enable();
	return ret;
}

/* Get rid of any references to a local port held by the
 * given sock.
 */
__inline__ void __tcp_put_port(struct sock *sk)
{
	struct tcp_bind_hashbucket *head = &tcp_bhash[tcp_bhashfn(sk->num)];
	struct tcp_bind_bucket *tb;

	spin_lock(&head->lock);
	tb = (struct tcp_bind_bucket *) sk->prev;
	if (sk->bind_next)
		sk->bind_next->bind_pprev = sk->bind_pprev;
	*(sk->bind_pprev) = sk->bind_next;
	sk->prev = NULL;
	sk->num = 0;
	if (tb->owners == NULL) {
		if (tb->next)
			tb->next->pprev = tb->pprev;
		*(tb->pprev) = tb->next;
		kmem_cache_free(tcp_bucket_cachep, tb);
	}
	spin_unlock(&head->lock);
}

void tcp_put_port(struct sock *sk)
{
	local_bh_disable();
	__tcp_put_port(sk);
	local_bh_enable();
}

/* This lock without WQ_FLAG_EXCLUSIVE is good on UP and it can be very bad on SMP.
 * Look, when several writers sleep and reader wakes them up, all but one
 * immediately hit write lock and grab all the cpus. Exclusive sleep solves
 * this, _but_ remember, it adds useless work on UP machines (wake up each
 * exclusive lock release). It should be ifdefed really.
 */

void tcp_listen_wlock(void)
{
	write_lock(&tcp_lhash_lock);

	if (atomic_read(&tcp_lhash_users)) {
		DECLARE_WAITQUEUE(wait, current);

		add_wait_queue_exclusive(&tcp_lhash_wait, &wait);
		for (;;) {
			set_current_state(TASK_UNINTERRUPTIBLE);
			if (atomic_read(&tcp_lhash_users) == 0)
				break;
			write_unlock_bh(&tcp_lhash_lock);
			schedule();
			write_lock_bh(&tcp_lhash_lock);
		}

		__set_current_state(TASK_RUNNING);
		remove_wait_queue(&tcp_lhash_wait, &wait);
	}
}

static __inline__ void __tcp_v4_hash(struct sock *sk)
{
	struct sock **skp;
	rwlock_t *lock;

	BUG_TRAP(sk->pprev==NULL);
	if(sk->state == TCP_LISTEN) {
		skp = &tcp_listening_hash[tcp_sk_listen_hashfn(sk)];
		lock = &tcp_lhash_lock;
		tcp_listen_wlock();
	} else {
		skp = &tcp_ehash[(sk->hashent = tcp_sk_hashfn(sk))].chain;
		lock = &tcp_ehash[sk->hashent].lock;
		write_lock(lock);
	}
	if((sk->next = *skp) != NULL)
		(*skp)->pprev = &sk->next;
	*skp = sk;
	sk->pprev = skp;
	sock_prot_inc_use(sk->prot);
	write_unlock(lock);
	if (sk->state == TCP_LISTEN)
		wake_up(&tcp_lhash_wait);
}

static void tcp_v4_hash(struct sock *sk)
{
	if (sk->state != TCP_CLOSE) {
		local_bh_disable();
		__tcp_v4_hash(sk);
		local_bh_enable();
	}
}

void tcp_unhash(struct sock *sk)
{
	rwlock_t *lock;

	if (sk->state == TCP_LISTEN) {
		local_bh_disable();
		tcp_listen_wlock();
		lock = &tcp_lhash_lock;
	} else {
		struct tcp_ehash_bucket *head = &tcp_ehash[sk->hashent];
		lock = &head->lock;
		write_lock_bh(&head->lock);
	}

	if(sk->pprev) {
		if(sk->next)
			sk->next->pprev = sk->pprev;
		*sk->pprev = sk->next;
		sk->pprev = NULL;
		sock_prot_dec_use(sk->prot);
	}
	write_unlock_bh(lock);
	if (sk->state == TCP_LISTEN)
		wake_up(&tcp_lhash_wait);
}

/* Don't inline this cruft.  Here are some nice properties to
 * exploit here.  The BSD API does not allow a listening TCP
 * to specify the remote port nor the remote address for the
 * connection.  So always assume those are both wildcarded
 * during the search since they can never be otherwise.
 */
static struct sock *__tcp_v4_lookup_listener(struct sock *sk, u32 daddr, unsigned short hnum, int dif)
{
	struct sock *result = NULL;
	int score, hiscore;

	hiscore=0;
	for(; sk; sk = sk->next) {
		if(sk->num == hnum) {
			__u32 rcv_saddr = sk->rcv_saddr;

			score = 1;
			if(rcv_saddr) {
				if (rcv_saddr != daddr)
					continue;
				score++;
			}
			if (sk->bound_dev_if) {
				if (sk->bound_dev_if != dif)
					continue;
				score++;
			}
			if (score == 3)
				return sk;
			if (score > hiscore) {
				hiscore = score;
				result = sk;
			}
		}
	}
	return result;
}

/* Optimize the common listener case. */
__inline__ struct sock *tcp_v4_lookup_listener(u32 daddr, unsigned short hnum, int dif)
{
	struct sock *sk;

	read_lock(&tcp_lhash_lock);
	sk = tcp_listening_hash[tcp_lhashfn(hnum)];
	if (sk) {
		if (sk->num == hnum &&
		    sk->next == NULL &&
		    (!sk->rcv_saddr || sk->rcv_saddr == daddr) &&
		    !sk->bound_dev_if)
			goto sherry_cache;
		sk = __tcp_v4_lookup_listener(sk, daddr, hnum, dif);
	}
	if (sk) {
sherry_cache:
		sock_hold(sk);
	}
	read_unlock(&tcp_lhash_lock);
	return sk;
}

/* Sockets in TCP_CLOSE state are _always_ taken out of the hash, so
 * we need not check it for TCP lookups anymore, thanks Alexey. -DaveM
 *
 * Local BH must be disabled here.
 */

static inline struct sock *__tcp_v4_lookup_established(u32 saddr, u16 sport,
						       u32 daddr, u16 hnum, int dif)
{
	struct tcp_ehash_bucket *head;
	TCP_V4_ADDR_COOKIE(acookie, saddr, daddr)
	__u32 ports = TCP_COMBINED_PORTS(sport, hnum);
	struct sock *sk;
	int hash;

	/* Optimize here for direct hit, only listening connections can
	 * have wildcards anyways.
	 */
	hash = tcp_hashfn(daddr, hnum, saddr, sport);
	head = &tcp_ehash[hash];
	read_lock(&head->lock);
	for(sk = head->chain; sk; sk = sk->next) {
		if(TCP_IPV4_MATCH(sk, acookie, saddr, daddr, ports, dif))
			goto hit; /* You sunk my battleship! */
	}

	/* Must check for a TIME_WAIT'er before going to listener hash. */
	for(sk = (head + tcp_ehash_size)->chain; sk; sk = sk->next)
		if(TCP_IPV4_MATCH(sk, acookie, saddr, daddr, ports, dif))
			goto hit;
	read_unlock(&head->lock);

	return NULL;

hit:
	sock_hold(sk);
	read_unlock(&head->lock);
	return sk;
}

static inline struct sock *__tcp_v4_lookup(u32 saddr, u16 sport,
					   u32 daddr, u16 hnum, int dif)
{
	struct sock *sk;

	sk = __tcp_v4_lookup_established(saddr, sport, daddr, hnum, dif);

	if (sk)
		return sk;
		
	return tcp_v4_lookup_listener(daddr, hnum, dif);
}

__inline__ struct sock *tcp_v4_lookup(u32 saddr, u16 sport, u32 daddr, u16 dport, int dif)
{
	struct sock *sk;

	local_bh_disable();
	sk = __tcp_v4_lookup(saddr, sport, daddr, ntohs(dport), dif);
	local_bh_enable();

	return sk;
}

static inline __u32 tcp_v4_init_sequence(struct sock *sk, struct sk_buff *skb)
{
	return secure_tcp_sequence_number(skb->nh.iph->daddr,
					  skb->nh.iph->saddr,
					  skb->h.th->dest,
					  skb->h.th->source);
}

static int tcp_v4_check_established(struct sock *sk)
{
	u32 daddr = sk->rcv_saddr;
	u32 saddr = sk->daddr;
	int dif = sk->bound_dev_if;
	TCP_V4_ADDR_COOKIE(acookie, saddr, daddr)
	__u32 ports = TCP_COMBINED_PORTS(sk->dport, sk->num);
	int hash = tcp_hashfn(daddr, sk->num, saddr, sk->dport);
	struct tcp_ehash_bucket *head = &tcp_ehash[hash];
	struct sock *sk2, **skp;
	struct tcp_tw_bucket *tw;

	write_lock_bh(&head->lock);

	/* Check TIME-WAIT sockets first. */
	for(skp = &(head + tcp_ehash_size)->chain; (sk2=*skp) != NULL;
	    skp = &sk2->next) {
		tw = (struct tcp_tw_bucket*)sk2;

		if(TCP_IPV4_MATCH(sk2, acookie, saddr, daddr, ports, dif)) {
			struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);

			/* With PAWS, it is safe from the viewpoint
			   of data integrity. Even without PAWS it
			   is safe provided sequence spaces do not
			   overlap i.e. at data rates <= 80Mbit/sec.

			   Actually, the idea is close to VJ's one,
			   only timestamp cache is held not per host,
			   but per port pair and TW bucket is used