ipmr.c

Linux内核源代码 为压缩文件 是<<Linux内核>>一书中的源代码

/*
 *	IP multicast routing support for mrouted 3.6/3.8
 *
 *		(c) 1995 Alan Cox, <[e-mail]>
 *	  Linux Consultancy and Custom Driver Development
 *
 *	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.
 *
 *	Version: $Id: ipmr.c,v 1.55 2000/11/28 13:13:27 davem Exp $
 *
 *	Fixes:
 *	Michael Chastain	:	Incorrect size of copying.
 *	Alan Cox		:	Added the cache manager code
 *	Alan Cox		:	Fixed the clone/copy bug and device race.
 *	Mike McLagan		:	Routing by source
 *	Malcolm Beattie		:	Buffer handling fixes.
 *	Alexey Kuznetsov	:	Double buffer free and other fixes.
 *	SVR Anand		:	Fixed several multicast bugs and problems.
 *	Alexey Kuznetsov	:	Status, optimisations and more.
 *	Brad Parker		:	Better behaviour on mrouted upcall
 *					overflow.
 *      Carlos Picoto           :       PIMv1 Support
 *	Pavlin Ivanov Radoslavov:	PIMv2 Registers must checksum only PIM header
 *					Relax this requrement to work with older peers.
 *
 */

#include <linux/config.h>
#include <asm/system.h>
#include <asm/uaccess.h>
#include <linux/types.h>
#include <linux/sched.h>
#include <linux/errno.h>
#include <linux/timer.h>
#include <linux/mm.h>
#include <linux/kernel.h>
#include <linux/fcntl.h>
#include <linux/stat.h>
#include <linux/socket.h>
#include <linux/in.h>
#include <linux/inet.h>
#include <linux/netdevice.h>
#include <linux/inetdevice.h>
#include <linux/igmp.h>
#include <linux/proc_fs.h>
#include <linux/mroute.h>
#include <linux/init.h>
#include <net/ip.h>
#include <net/protocol.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <net/icmp.h>
#include <net/udp.h>
#include <net/raw.h>
#include <linux/notifier.h>
#include <linux/if_arp.h>
#include <linux/netfilter_ipv4.h>
#include <net/ipip.h>
#include <net/checksum.h>

#if defined(CONFIG_IP_PIMSM_V1) || defined(CONFIG_IP_PIMSM_V2)
#define CONFIG_IP_PIMSM	1
#endif

static struct sock *mroute_socket;


/* Big lock, protecting vif table, mrt cache and mroute socket state.
   Note that the changes are semaphored via rtnl_lock.
 */

static rwlock_t mrt_lock = RW_LOCK_UNLOCKED;

/*
 *	Multicast router control variables
 */

static struct vif_device vif_table[MAXVIFS];		/* Devices 		*/
static int maxvif;

#define VIF_EXISTS(idx) (vif_table[idx].dev != NULL)

int mroute_do_assert = 0;				/* Set in PIM assert	*/
int mroute_do_pim = 0;

static struct mfc_cache *mfc_cache_array[MFC_LINES];	/* Forwarding cache	*/

static struct mfc_cache *mfc_unres_queue;		/* Queue of unresolved entries */
atomic_t cache_resolve_queue_len;			/* Size of unresolved	*/

/* Special spinlock for queue of unresolved entries */
static spinlock_t mfc_unres_lock = SPIN_LOCK_UNLOCKED;

/* We return to original Alan's scheme. Hash table of resolved
   entries is changed only in process context and protected
   with weak lock mrt_lock. Queue of unresolved entries is protected
   with strong spinlock mfc_unres_lock.

   In this case data path is free of exclusive locks at all.
 */

kmem_cache_t *mrt_cachep;

static int ip_mr_forward(struct sk_buff *skb, struct mfc_cache *cache, int local);
static int ipmr_cache_report(struct sk_buff *pkt, vifi_t vifi, int assert);
static int ipmr_fill_mroute(struct sk_buff *skb, struct mfc_cache *c, struct rtmsg *rtm);

extern struct inet_protocol pim_protocol;

static struct timer_list ipmr_expire_timer;

/* Service routines creating virtual interfaces: DVMRP tunnels and PIMREG */

static
struct net_device *ipmr_new_tunnel(struct vifctl *v)
{
	struct net_device  *dev;

	dev = __dev_get_by_name("tunl0");

	if (dev) {
		int err;
		struct ifreq ifr;
		mm_segment_t	oldfs;
		struct ip_tunnel_parm p;
		struct in_device  *in_dev;

		memset(&p, 0, sizeof(p));
		p.iph.daddr = v->vifc_rmt_addr.s_addr;
		p.iph.saddr = v->vifc_lcl_addr.s_addr;
		p.iph.version = 4;
		p.iph.ihl = 5;
		p.iph.protocol = IPPROTO_IPIP;
		sprintf(p.name, "dvmrp%d", v->vifc_vifi);
		ifr.ifr_ifru.ifru_data = (void*)&p;

		oldfs = get_fs(); set_fs(KERNEL_DS);
		err = dev->do_ioctl(dev, &ifr, SIOCADDTUNNEL);
		set_fs(oldfs);

		dev = NULL;

		if (err == 0 && (dev = __dev_get_by_name(p.name)) != NULL) {
			dev->flags |= IFF_MULTICAST;

			in_dev = __in_dev_get(dev);
			if (in_dev == NULL && (in_dev = inetdev_init(dev)) == NULL)
				goto failure;
			in_dev->cnf.rp_filter = 0;

			if (dev_open(dev))
				goto failure;
		}
	}
	return dev;

failure:
	unregister_netdevice(dev);
	return NULL;
}

#ifdef CONFIG_IP_PIMSM

static int reg_vif_num = -1;

static int reg_vif_xmit(struct sk_buff *skb, struct net_device *dev)
{
	read_lock(&mrt_lock);
	((struct net_device_stats*)dev->priv)->tx_bytes += skb->len;
	((struct net_device_stats*)dev->priv)->tx_packets++;
	ipmr_cache_report(skb, reg_vif_num, IGMPMSG_WHOLEPKT);
	read_unlock(&mrt_lock);
	kfree_skb(skb);
	return 0;
}

static struct net_device_stats *reg_vif_get_stats(struct net_device *dev)
{
	return (struct net_device_stats*)dev->priv;
}

static
struct net_device *ipmr_reg_vif(struct vifctl *v)
{
	struct net_device  *dev;
	struct in_device *in_dev;
	int size;

	size = sizeof(*dev) + sizeof(struct net_device_stats);
	dev = kmalloc(size, GFP_KERNEL);
	if (!dev)
		return NULL;

	memset(dev, 0, size);

	dev->priv = dev + 1;

	strcpy(dev->name, "pimreg");

	dev->type		= ARPHRD_PIMREG;
	dev->mtu		= 1500 - sizeof(struct iphdr) - 8;
	dev->flags		= IFF_NOARP;
	dev->hard_start_xmit	= reg_vif_xmit;
	dev->get_stats		= reg_vif_get_stats;
	dev->features		|= NETIF_F_DYNALLOC;

	if (register_netdevice(dev)) {
		kfree(dev);
		return NULL;
	}
	dev->iflink = 0;

	if ((in_dev = inetdev_init(dev)) == NULL)
		goto failure;

	in_dev->cnf.rp_filter = 0;

	if (dev_open(dev))
		goto failure;

	return dev;

failure:
	unregister_netdevice(dev);
	return NULL;
}
#endif

/*
 *	Delete a VIF entry
 */
 
static int vif_delete(int vifi)
{
	struct vif_device *v;
	struct net_device *dev;
	struct in_device *in_dev;

	if (vifi < 0 || vifi >= maxvif)
		return -EADDRNOTAVAIL;

	v = &vif_table[vifi];

	write_lock_bh(&mrt_lock);
	dev = v->dev;
	v->dev = NULL;

	if (!dev) {
		write_unlock_bh(&mrt_lock);
		return -EADDRNOTAVAIL;
	}

#ifdef CONFIG_IP_PIMSM
	if (vifi == reg_vif_num)
		reg_vif_num = -1;
#endif

	if (vifi+1 == maxvif) {
		int tmp;
		for (tmp=vifi-1; tmp>=0; tmp--) {
			if (VIF_EXISTS(tmp))
				break;
		}
		maxvif = tmp+1;
	}

	write_unlock_bh(&mrt_lock);

	dev_set_allmulti(dev, -1);

	if ((in_dev = __in_dev_get(dev)) != NULL) {
		in_dev->cnf.mc_forwarding--;
		ip_rt_multicast_event(in_dev);
	}

	if (v->flags&(VIFF_TUNNEL|VIFF_REGISTER))
		unregister_netdevice(dev);

	dev_put(dev);
	return 0;
}

/* Destroy an unresolved cache entry, killing queued skbs
   and reporting error to netlink readers.
 */

static void ipmr_destroy_unres(struct mfc_cache *c)
{
	struct sk_buff *skb;

	atomic_dec(&cache_resolve_queue_len);

	while((skb=skb_dequeue(&c->mfc_un.unres.unresolved))) {
#ifdef CONFIG_RTNETLINK
		if (skb->nh.iph->version == 0) {
			struct nlmsghdr *nlh = (struct nlmsghdr *)skb_pull(skb, sizeof(struct iphdr));
			nlh->nlmsg_type = NLMSG_ERROR;
			nlh->nlmsg_len = NLMSG_LENGTH(sizeof(struct nlmsgerr));
			skb_trim(skb, nlh->nlmsg_len);
			((struct nlmsgerr*)NLMSG_DATA(nlh))->error = -ETIMEDOUT;
			netlink_unicast(rtnl, skb, NETLINK_CB(skb).dst_pid, MSG_DONTWAIT);
		} else
#endif
			kfree_skb(skb);
	}

	kmem_cache_free(mrt_cachep, c);
}


/* Single timer process for all the unresolved queue. */

void ipmr_expire_process(unsigned long dummy)
{
	unsigned long now;
	unsigned long expires;
	struct mfc_cache *c, **cp;

	if (!spin_trylock(&mfc_unres_lock)) {
		mod_timer(&ipmr_expire_timer, jiffies+HZ/10);
		return;
	}

	if (atomic_read(&cache_resolve_queue_len) == 0)
		goto out;

	now = jiffies;
	expires = 10*HZ;
	cp = &mfc_unres_queue;

	while ((c=*cp) != NULL) {
		long interval = c->mfc_un.unres.expires - now;

		if (interval > 0) {
			if (interval < expires)
				expires = interval;
			cp = &c->next;
			continue;
		}

		*cp = c->next;

		ipmr_destroy_unres(c);
	}

	if (atomic_read(&cache_resolve_queue_len))
		mod_timer(&ipmr_expire_timer, jiffies + expires);

out:
	spin_unlock(&mfc_unres_lock);
}

/* Fill oifs list. It is called under write locked mrt_lock. */

static void ipmr_update_threshoulds(struct mfc_cache *cache, unsigned char *ttls)
{
	int vifi;

	cache->mfc_un.res.minvif = MAXVIFS;
	cache->mfc_un.res.maxvif = 0;
	memset(cache->mfc_un.res.ttls, 255, MAXVIFS);

	for (vifi=0; vifi<maxvif; vifi++) {
		if (VIF_EXISTS(vifi) && ttls[vifi] && ttls[vifi] < 255) {
			cache->mfc_un.res.ttls[vifi] = ttls[vifi];
			if (cache->mfc_un.res.minvif > vifi)
				cache->mfc_un.res.minvif = vifi;
			if (cache->mfc_un.res.maxvif <= vifi)
				cache->mfc_un.res.maxvif = vifi + 1;
		}
	}
}

static int vif_add(struct vifctl *vifc, int mrtsock)
{
	int vifi = vifc->vifc_vifi;
	struct vif_device *v = &vif_table[vifi];
	struct net_device *dev;
	struct in_device *in_dev;

	/* Is vif busy ? */
	if (VIF_EXISTS(vifi))
		return -EADDRINUSE;

	switch (vifc->vifc_flags) {
#ifdef CONFIG_IP_PIMSM
	case VIFF_REGISTER:
		/*
		 * Special Purpose VIF in PIM
		 * All the packets will be sent to the daemon
		 */
		if (reg_vif_num >= 0)
			return -EADDRINUSE;
		dev = ipmr_reg_vif(vifc);
		if (!dev)
			return -ENOBUFS;
		break;
#endif
	case VIFF_TUNNEL:	
		dev = ipmr_new_tunnel(vifc);
		if (!dev)
			return -ENOBUFS;
		break;
	case 0:
		dev=ip_dev_find(vifc->vifc_lcl_addr.s_addr);
		if (!dev)
			return -EADDRNOTAVAIL;
		__dev_put(dev);
		break;
	default:
		return -EINVAL;
	}

	if ((in_dev = __in_dev_get(dev)) == NULL)
		return -EADDRNOTAVAIL;
	in_dev->cnf.mc_forwarding++;
	dev_set_allmulti(dev, +1);
	ip_rt_multicast_event(in_dev);

	/*
	 *	Fill in the VIF structures
	 */
	v->rate_limit=vifc->vifc_rate_limit;
	v->local=vifc->vifc_lcl_addr.s_addr;
	v->remote=vifc->vifc_rmt_addr.s_addr;
	v->flags=vifc->vifc_flags;
	if (!mrtsock)
		v->flags |= VIFF_STATIC;
	v->threshold=vifc->vifc_threshold;
	v->bytes_in = 0;
	v->bytes_out = 0;
	v->pkt_in = 0;
	v->pkt_out = 0;
	v->link = dev->ifindex;
	if (v->flags&(VIFF_TUNNEL|VIFF_REGISTER))
		v->link = dev->iflink;

	/* And finish update writing critical data */
	write_lock_bh(&mrt_lock);
	dev_hold(dev);
	v->dev=dev;
#ifdef CONFIG_IP_PIMSM
	if (v->flags&VIFF_REGISTER)
		reg_vif_num = vifi;
#endif
	if (vifi+1 > maxvif)
		maxvif = vifi+1;
	write_unlock_bh(&mrt_lock);
	return 0;
}

static struct mfc_cache *ipmr_cache_find(__u32 origin, __u32 mcastgrp)
{
	int line=MFC_HASH(mcastgrp,origin);
	struct mfc_cache *c;

	for (c=mfc_cache_array[line]; c; c = c->next) {
		if (c->mfc_origin==origin && c->mfc_mcastgrp==mcastgrp)
			break;
	}
	return c;
}

/*
 *	Allocate a multicast cache entry
 */
static struct mfc_cache *ipmr_cache_alloc(void)
{
	struct mfc_cache *c=kmem_cache_alloc(mrt_cachep, GFP_KERNEL);
	if(c==NULL)
		return NULL;
	memset(c, 0, sizeof(*c));
	c->mfc_un.res.minvif = MAXVIFS;
	return c;
}

static struct mfc_cache *ipmr_cache_alloc_unres(void)
{
	struct mfc_cache *c=kmem_cache_alloc(mrt_cachep, GFP_ATOMIC);
	if(c==NULL)
		return NULL;
	memset(c, 0, sizeof(*c));
	skb_queue_head_init(&c->mfc_un.unres.unresolved);
	c->mfc_un.unres.expires = jiffies + 10*HZ;
	return c;
}

/*
 *	A cache entry has gone into a resolved state from queued
 */
 
static void ipmr_cache_resolve(struct mfc_cache *uc, struct mfc_cache *c)
{
	struct sk_buff *skb;

	/*
	 *	Play the pending entries through our router
	 */

	while((skb=__skb_dequeue(&uc->mfc_un.unres.unresolved))) {
#ifdef CONFIG_RTNETLINK
		if (skb->nh.iph->version == 0) {
			int err;
			struct nlmsghdr *nlh = (struct nlmsghdr *)skb_pull(skb, sizeof(struct iphdr));

			if (ipmr_fill_mroute(skb, c, NLMSG_DATA(nlh)) > 0) {
				nlh->nlmsg_len = skb->tail - (u8*)nlh;
			} else {
				nlh->nlmsg_type = NLMSG_ERROR;
				nlh->nlmsg_len = NLMSG_LENGTH(sizeof(struct nlmsgerr));
				skb_trim(skb, nlh->nlmsg_len);
				((struct nlmsgerr*)NLMSG_DATA(nlh))->error = -EMSGSIZE;
			}
			err = netlink_unicast(rtnl, skb, NETLINK_CB(skb).dst_pid, MSG_DONTWAIT);
		} else
#endif
			ip_mr_forward(skb, c, 0);
	}
}

/*
 *	Bounce a cache query up to mrouted. We could use netlink for this but mrouted
 *	expects the following bizarre scheme.
 *
 *	Called under mrt_lock.
 */
 
static int ipmr_cache_report(struct sk_buff *pkt, vifi_t vifi, int assert)
{
	struct sk_buff *skb;
	int ihl = pkt->nh.iph->ihl<<2;
	struct igmphdr *igmp;
	struct igmpmsg *msg;
	int ret;

#ifdef CONFIG_IP_PIMSM
	if (assert == IGMPMSG_WHOLEPKT)
		skb = skb_realloc_headroom(pkt, sizeof(struct iphdr));
	else
#endif
		skb = alloc_skb(128, GFP_ATOMIC);

	if(!skb)
		return -ENOBUFS;

#ifdef CONFIG_IP_PIMSM
	if (assert == IGMPMSG_WHOLEPKT) {
		/* Ugly, but we have no choice with this interface.
		   Duplicate old header, fix ihl, length etc.
		   And all this only to mangle msg->im_msgtype and
		   to set msg->im_mbz to "mbz" :-)
		 */
		msg = (struct igmpmsg*)skb_push(skb, sizeof(struct iphdr));
		skb->nh.raw = skb->h.raw = (u8*)msg;
		memcpy(msg, pkt->nh.raw, sizeof(struct iphdr));
		msg->im_msgtype = IGMPMSG_WHOLEPKT;
		msg->im_mbz = 0;
 		msg->im_vif = reg_vif_num;
		skb->nh.iph->ihl = sizeof(struct iphdr) >> 2;
		skb->nh.iph->tot_len = htons(ntohs(pkt->nh.iph->tot_len) + sizeof(struct iphdr));
	} else 
#endif
	{	
		
	/*
	 *	Copy the IP header
	 */

	skb->nh.iph = (struct iphdr *)skb_put(skb, ihl);
	memcpy(skb->data,pkt->data,ihl);
	skb->nh.iph->protocol = 0;			/* Flag to the kernel this is a route add */
	msg = (struct igmpmsg*)skb->nh.iph;
	msg->im_vif = vifi;
	skb->dst = dst_clone(pkt->dst);

	/*
	 *	Add our header
	 */

	igmp=(struct igmphdr *)skb_put(skb,sizeof(struct igmphdr));
	igmp->type	=