sock.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.
 *
 *		Generic socket support routines. Memory allocators, socket lock/release
 *		handler for protocols to use and generic option handler.
 *
 *
 * Version:	$Id: sock.c,v 1.102 2000/12/11 23:00:24 davem Exp $
 *
 * Authors:	Ross Biro, <[e-mail]>
 *		Fred N. van Kempen, <[e-mail]>
 *		Florian La Roche, <[e-mail]>
 *		Alan Cox, <[e-mail]>
 *
 * Fixes:
 *		Alan Cox	: 	Numerous verify_area() problems
 *		Alan Cox	:	Connecting on a connecting socket
 *					now returns an error for tcp.
 *		Alan Cox	:	sock->protocol is set correctly.
 *					and is not sometimes left as 0.
 *		Alan Cox	:	connect handles icmp errors on a
 *					connect properly. Unfortunately there
 *					is a restart syscall nasty there. I
 *					can't match BSD without hacking the C
 *					library. Ideas urgently sought!
 *		Alan Cox	:	Disallow bind() to addresses that are
 *					not ours - especially broadcast ones!!
 *		Alan Cox	:	Socket 1024 _IS_ ok for users. (fencepost)
 *		Alan Cox	:	sock_wfree/sock_rfree don't destroy sockets,
 *					instead they leave that for the DESTROY timer.
 *		Alan Cox	:	Clean up error flag in accept
 *		Alan Cox	:	TCP ack handling is buggy, the DESTROY timer
 *					was buggy. Put a remove_sock() in the handler
 *					for memory when we hit 0. Also altered the timer
 *					code. The ACK stuff can wait and needs major 
 *					TCP layer surgery.
 *		Alan Cox	:	Fixed TCP ack bug, removed remove sock
 *					and fixed timer/inet_bh race.
 *		Alan Cox	:	Added zapped flag for TCP
 *		Alan Cox	:	Move kfree_skb into skbuff.c and tidied up surplus code
 *		Alan Cox	:	for new sk_buff allocations wmalloc/rmalloc now call alloc_skb
 *		Alan Cox	:	kfree_s calls now are kfree_skbmem so we can track skb resources
 *		Alan Cox	:	Supports socket option broadcast now as does udp. Packet and raw need fixing.
 *		Alan Cox	:	Added RCVBUF,SNDBUF size setting. It suddenly occurred to me how easy it was so...
 *		Rick Sladkey	:	Relaxed UDP rules for matching packets.
 *		C.E.Hawkins	:	IFF_PROMISC/SIOCGHWADDR support
 *	Pauline Middelink	:	identd support
 *		Alan Cox	:	Fixed connect() taking signals I think.
 *		Alan Cox	:	SO_LINGER supported
 *		Alan Cox	:	Error reporting fixes
 *		Anonymous	:	inet_create tidied up (sk->reuse setting)
 *		Alan Cox	:	inet sockets don't set sk->type!
 *		Alan Cox	:	Split socket option code
 *		Alan Cox	:	Callbacks
 *		Alan Cox	:	Nagle flag for Charles & Johannes stuff
 *		Alex		:	Removed restriction on inet fioctl
 *		Alan Cox	:	Splitting INET from NET core
 *		Alan Cox	:	Fixed bogus SO_TYPE handling in getsockopt()
 *		Adam Caldwell	:	Missing return in SO_DONTROUTE/SO_DEBUG code
 *		Alan Cox	:	Split IP from generic code
 *		Alan Cox	:	New kfree_skbmem()
 *		Alan Cox	:	Make SO_DEBUG superuser only.
 *		Alan Cox	:	Allow anyone to clear SO_DEBUG
 *					(compatibility fix)
 *		Alan Cox	:	Added optimistic memory grabbing for AF_UNIX throughput.
 *		Alan Cox	:	Allocator for a socket is settable.
 *		Alan Cox	:	SO_ERROR includes soft errors.
 *		Alan Cox	:	Allow NULL arguments on some SO_ opts
 *		Alan Cox	: 	Generic socket allocation to make hooks
 *					easier (suggested by Craig Metz).
 *		Michael Pall	:	SO_ERROR returns positive errno again
 *              Steve Whitehouse:       Added default destructor to free
 *                                      protocol private data.
 *              Steve Whitehouse:       Added various other default routines
 *                                      common to several socket families.
 *              Chris Evans     :       Call suser() check last on F_SETOWN
 *		Jay Schulist	:	Added SO_ATTACH_FILTER and SO_DETACH_FILTER.
 *		Andi Kleen	:	Add sock_kmalloc()/sock_kfree_s()
 *		Andi Kleen	:	Fix write_space callback
 *
 * To Fix:
 *
 *
 *		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.
 */

#include <linux/config.h>
#include <linux/errno.h>
#include <linux/types.h>
#include <linux/socket.h>
#include <linux/in.h>
#include <linux/kernel.h>
#include <linux/major.h>
#include <linux/sched.h>
#include <linux/timer.h>
#include <linux/string.h>
#include <linux/sockios.h>
#include <linux/net.h>
#include <linux/fcntl.h>
#include <linux/mm.h>
#include <linux/slab.h>
#include <linux/interrupt.h>
#include <linux/poll.h>
#include <linux/init.h>

#include <asm/uaccess.h>
#include <asm/system.h>

#include <linux/inet.h>
#include <linux/netdevice.h>
#include <net/ip.h>
#include <net/protocol.h>
#include <net/arp.h>
#include <net/route.h>
#include <net/tcp.h>
#include <net/udp.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <net/raw.h>
#include <net/icmp.h>
#include <linux/ipsec.h>

#ifdef CONFIG_FILTER
#include <linux/filter.h>
#endif

#define min(a,b)	((a)<(b)?(a):(b))

/* Run time adjustable parameters. */
__u32 sysctl_wmem_max = SK_WMEM_MAX;
__u32 sysctl_rmem_max = SK_RMEM_MAX;
__u32 sysctl_wmem_default = SK_WMEM_MAX;
__u32 sysctl_rmem_default = SK_RMEM_MAX;

/* Maximal space eaten by iovec or ancilliary data plus some space */
int sysctl_optmem_max = sizeof(unsigned long)*(2*UIO_MAXIOV + 512);

static int sock_set_timeout(long *timeo_p, char *optval, int optlen)
{
	struct timeval tv;

	if (optlen < sizeof(tv))
		return -EINVAL;
	if (copy_from_user(&tv, optval, sizeof(tv)))
		return -EFAULT;

	*timeo_p = MAX_SCHEDULE_TIMEOUT;
	if (tv.tv_sec == 0 && tv.tv_usec == 0)
		return 0;
	if (tv.tv_sec < (MAX_SCHEDULE_TIMEOUT/HZ - 1))
		*timeo_p = tv.tv_sec*HZ + (tv.tv_usec+(1000000/HZ-1))/(1000000/HZ);
	return 0;
}

/*
 *	This is meant for all protocols to use and covers goings on
 *	at the socket level. Everything here is generic.
 */

int sock_setsockopt(struct socket *sock, int level, int optname,
		    char *optval, int optlen)
{
	struct sock *sk=sock->sk;
#ifdef CONFIG_FILTER
	struct sk_filter *filter;
#endif
	int val;
	int valbool;
	int err;
	struct linger ling;
	int ret = 0;
	
	/*
	 *	Options without arguments
	 */

#ifdef SO_DONTLINGER		/* Compatibility item... */
	switch(optname)
	{
		case SO_DONTLINGER:
			sk->linger=0;
			return 0;
	}
#endif	
		
  	if(optlen<sizeof(int))
  		return(-EINVAL);
  	
	err = get_user(val, (int *)optval);
	if (err)
		return err;
	
  	valbool = val?1:0;

	lock_sock(sk);

  	switch(optname) 
  	{
		case SO_DEBUG:	
			if(val && !capable(CAP_NET_ADMIN))
			{
				ret = -EACCES;
			}
			else
				sk->debug=valbool;
			break;
		case SO_REUSEADDR:
			sk->reuse = valbool;
			break;
		case SO_TYPE:
		case SO_ERROR:
			ret = -ENOPROTOOPT;
		  	break;
		case SO_DONTROUTE:
			sk->localroute=valbool;
			break;
		case SO_BROADCAST:
			sk->broadcast=valbool;
			break;
		case SO_SNDBUF:
			/* Don't error on this BSD doesn't and if you think
			   about it this is right. Otherwise apps have to
			   play 'guess the biggest size' games. RCVBUF/SNDBUF
			   are treated in BSD as hints */
			   
			if (val > sysctl_wmem_max)
				val = sysctl_wmem_max;

			sk->userlocks |= SOCK_SNDBUF_LOCK;
			sk->sndbuf = max(val*2,SOCK_MIN_SNDBUF);

			/*
			 *	Wake up sending tasks if we
			 *	upped the value.
			 */
			sk->write_space(sk);
			break;

		case SO_RCVBUF:
			/* Don't error on this BSD doesn't and if you think
			   about it this is right. Otherwise apps have to
			   play 'guess the biggest size' games. RCVBUF/SNDBUF
			   are treated in BSD as hints */
			  
			if (val > sysctl_rmem_max)
				val = sysctl_rmem_max;

			sk->userlocks |= SOCK_RCVBUF_LOCK;
			/* FIXME: is this lower bound the right one? */
			sk->rcvbuf = max(val*2,SOCK_MIN_RCVBUF);
			break;

		case SO_KEEPALIVE:
#ifdef CONFIG_INET
			if (sk->protocol == IPPROTO_TCP)
			{
				tcp_set_keepalive(sk, valbool);
			}
#endif
			sk->keepopen = valbool;
			break;

	 	case SO_OOBINLINE:
			sk->urginline = valbool;
			break;

	 	case SO_NO_CHECK:
			sk->no_check = valbool;
			break;

		case SO_PRIORITY:
			if ((val >= 0 && val <= 6) || capable(CAP_NET_ADMIN)) 
				sk->priority = val;
			else
				ret = -EPERM;
			break;

		case SO_LINGER:
			if(optlen<sizeof(ling)) {
				ret = -EINVAL;	/* 1003.1g */
				break;
			}
			if (copy_from_user(&ling,optval,sizeof(ling))) {
				ret = -EFAULT;
				break;
			}
			if(ling.l_onoff==0) {
				sk->linger=0;
			} else {
#if (BITS_PER_LONG == 32)
				if (ling.l_linger >= MAX_SCHEDULE_TIMEOUT/HZ)
					sk->lingertime=MAX_SCHEDULE_TIMEOUT;
				else
#endif
					sk->lingertime=ling.l_linger*HZ;
				sk->linger=1;
			}
			break;

		case SO_BSDCOMPAT:
			sk->bsdism = valbool;
			break;

		case SO_PASSCRED:
			sock->passcred = valbool;
			break;

		case SO_TIMESTAMP:
			sk->rcvtstamp = valbool;
			break;

		case SO_RCVLOWAT:
			if (val < 0)
				val = INT_MAX;
			sk->rcvlowat = val ? : 1;
			break;

		case SO_RCVTIMEO:
			ret = sock_set_timeout(&sk->rcvtimeo, optval, optlen);
			break;

		case SO_SNDTIMEO:
			ret = sock_set_timeout(&sk->sndtimeo, optval, optlen);
			break;

#ifdef CONFIG_NETDEVICES
		case SO_BINDTODEVICE:
		{
			char devname[IFNAMSIZ]; 

			/* Sorry... */ 
			if (!capable(CAP_NET_RAW)) {
				ret = -EPERM;
				break;
			}

			/* Bind this socket to a particular device like "eth0",
			 * as specified in the passed interface name. If the
			 * name is "" or the option length is zero the socket 
			 * is not bound. 
			 */ 

			if (!valbool) {
				sk->bound_dev_if = 0;
			} else {
				if (optlen > IFNAMSIZ) 
					optlen = IFNAMSIZ; 
				if (copy_from_user(devname, optval, optlen)) {
					ret = -EFAULT;
					break;
				}

				/* Remove any cached route for this socket. */
				sk_dst_reset(sk);

				if (devname[0] == '\0') {
					sk->bound_dev_if = 0;
				} else {
					struct net_device *dev = dev_get_by_name(devname);
					if (!dev) {
						ret = -ENODEV;
						break;
					}
					sk->bound_dev_if = dev->ifindex;
					dev_put(dev);
				}
			}
			break;
		}
#endif


#ifdef CONFIG_FILTER
		case SO_ATTACH_FILTER:
			ret = -EINVAL;
			if (optlen == sizeof(struct sock_fprog)) {
				struct sock_fprog fprog;

				ret = -EFAULT;
				if (copy_from_user(&fprog, optval, sizeof(fprog)))
					break;

				ret = sk_attach_filter(&fprog, sk);
			}
			break;

		case SO_DETACH_FILTER:
			spin_lock_bh(&sk->lock.slock);
			filter = sk->filter;
                        if (filter) {
				sk->filter = NULL;
				spin_unlock_bh(&sk->lock.slock);
				sk_filter_release(sk, filter);
				break;
			}
			spin_unlock_bh(&sk->lock.slock);
			ret = -ENONET;
			break;
#endif
		/* We implement the SO_SNDLOWAT etc to
		   not be settable (1003.1g 5.3) */
		default:
		  	ret = -ENOPROTOOPT;
			break;
  	}
	release_sock(sk);
	return ret;
}


int sock_getsockopt(struct socket *sock, int level, int optname,
		    char *optval, int *optlen)
{
	struct sock *sk = sock->sk;
	
	union
	{
  		int val;
  		struct linger ling;
		struct timeval tm;
	} v;
	
	int lv=sizeof(int),len;
  	
  	if(get_user(len,optlen))
  		return -EFAULT;

  	switch(optname) 
  	{
		case SO_DEBUG:		
			v.val = sk->debug;
			break;
		
		case SO_DONTROUTE:
			v.val = sk->localroute;
			break;
		
		case SO_BROADCAST:
			v.val= sk->broadcast;
			break;

		case SO_SNDBUF:
			v.val=sk->sndbuf;
			break;
		
		case SO_RCVBUF:
			v.val =sk->rcvbuf;
			break;

		case SO_REUSEADDR:
			v.val = sk->reuse;
			break;

		case SO_KEEPALIVE:
			v.val = sk->keepopen;
			break;

		case SO_TYPE:
			v.val = sk->type;		  		
			break;

		case SO_ERROR:
			v.val = -sock_error(sk);
			if(v.val==0)
				v.val=xchg(&sk->err_soft,0);
			break;

		case SO_OOBINLINE:
			v.val = sk->urginline;
			break;
	
		case SO_NO_CHECK:
			v.val = sk->no_check;
			break;

		case SO_PRIORITY:
			v.val = sk->priority;
			break;
		
		case SO_LINGER:	
			lv=sizeof(v.ling);
			v.ling.l_onoff=sk->linger;
 			v.ling.l_linger=sk->lingertime/HZ;
			break;
					
		case SO_BSDCOMPAT:
			v.val = sk->bsdism;
			break;

		case SO_TIMESTAMP:
			v.val = sk->rcvtstamp;
			break;

		case SO_RCVTIMEO:
			lv=sizeof(struct timeval);
			if (sk->rcvtimeo == MAX_SCHEDULE_TIMEOUT) {
				v.tm.tv_sec = 0;
				v.tm.tv_usec = 0;
			} else {
				v.tm.tv_sec = sk->rcvtimeo/HZ;
				v.tm.tv_usec = ((sk->rcvtimeo%HZ)*1000)/HZ;
			}
			break;

		case SO_SNDTIMEO:
			lv=sizeof(struct timeval);
			if (sk->sndtimeo == MAX_SCHEDULE_TIMEOUT) {
				v.tm.tv_sec = 0;
				v.tm.tv_usec = 0;
			} else {
				v.tm.tv_sec = sk->sndtimeo/HZ;
				v.tm.tv_usec = ((sk->sndtimeo%HZ)*1000)/HZ;
			}
			break;

		case SO_RCVLOWAT:
			v.val = sk->rcvlowat;
			break;

		case SO_SNDLOWAT:
			v.val=1;
			break; 

		case SO_PASSCRED:
			v.val = sock->passcred;
			break;

		case SO_PEERCRED:
			lv=sizeof(sk->peercred);
			len=min(len, lv);
			if(copy_to_user((void*)optval, &sk->peercred, len))
				return -EFAULT;
			goto lenout;

		case SO_PEERNAME:
		{
			char address[128];

			if (sock->ops->getname(sock, (struct sockaddr *)address, &lv, 2))
				return -ENOTCONN;
			if (lv < len)
				return -EINVAL;
			if(copy_to_user((void*)optval, address, len))
				return -EFAULT;
			goto lenout;
		}

		default:
			return(-ENOPROTOOPT);
	}
	len=min(len,lv);
	if(copy_to_user(optval,&v,len))
		return -EFAULT;
lenout:
  	if(put_user(len, optlen))
  		return -EFAULT;
  	return 0;
}

static kmem_cache_t *sk_cachep;

/*
 *	All socket objects are allocated here. This is for future
 *	usage.
 */
 
struct sock *sk_alloc(int family, int priority, int zero_it)
{
	struct sock *sk = kmem_cache_alloc(sk_cachep, priority);

	if(sk && zero_it) {
		memset(sk, 0, sizeof(struct sock));
		sk->family = family;
		sock_lock_init(sk);