]> nv-tegra.nvidia Code Review - linux-2.6.git/blobdiff - net/ipv4/ipmr.c
netdev: convert pseudo-devices to netdev_tx_t
[linux-2.6.git] / net / ipv4 / ipmr.c
index 9ccacf57f08bd8ff7a0e6315ae6766f6744fb59c..65d421cf5bc717754353f443a62aa856605f6347 100644 (file)
@@ -1,7 +1,7 @@
 /*
  *     IP multicast routing support for mrouted 3.6/3.8
  *
- *             (c) 1995 Alan Cox, <alan@redhat.com>
+ *             (c) 1995 Alan Cox, <alan@lxorguk.ukuu.org.uk>
  *       Linux Consultancy and Custom Driver Development
  *
  *     This program is free software; you can redistribute it and/or
@@ -9,8 +9,6 @@
  *     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 1.65 2001/10/31 21:55:54 davem Exp $
- *
  *     Fixes:
  *     Michael Chastain        :       Incorrect size of copying.
  *     Alan Cox                :       Added the cache manager code
@@ -31,7 +29,6 @@
 #include <asm/system.h>
 #include <asm/uaccess.h>
 #include <linux/types.h>
-#include <linux/sched.h>
 #include <linux/capability.h>
 #include <linux/errno.h>
 #include <linux/timer.h>
@@ -50,6 +47,7 @@
 #include <linux/mroute.h>
 #include <linux/init.h>
 #include <linux/if_ether.h>
+#include <net/net_namespace.h>
 #include <net/ip.h>
 #include <net/protocol.h>
 #include <linux/skbuff.h>
 #include <linux/netfilter_ipv4.h>
 #include <net/ipip.h>
 #include <net/checksum.h>
+#include <net/netlink.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.
  */
@@ -81,18 +77,9 @@ static DEFINE_RWLOCK(mrt_lock);
  *     Multicast router control variables
  */
 
-static struct vif_device vif_table[MAXVIFS];           /* Devices              */
-static int maxvif;
-
-#define VIF_EXISTS(idx) (vif_table[idx].dev != NULL)
-
-static int mroute_do_assert;                           /* Set in PIM assert    */
-static int mroute_do_pim;
-
-static struct mfc_cache *mfc_cache_array[MFC_LINES];   /* Forwarding cache     */
+#define VIF_EXISTS(_net, _idx) ((_net)->ipv4.vif_table[_idx].dev != NULL)
 
 static struct mfc_cache *mfc_unres_queue;              /* Queue of unresolved entries */
-static atomic_t cache_resolve_queue_len;               /* Size of unresolved   */
 
 /* Special spinlock for queue of unresolved entries */
 static DEFINE_SPINLOCK(mfc_unres_lock);
@@ -105,10 +92,11 @@ static DEFINE_SPINLOCK(mfc_unres_lock);
    In this case data path is free of exclusive locks at all.
  */
 
-static kmem_cache_t *mrt_cachep __read_mostly;
+static struct kmem_cache *mrt_cachep __read_mostly;
 
 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_cache_report(struct net *net,
+                            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);
 
 #ifdef CONFIG_IP_PIMSM_V2
@@ -119,17 +107,48 @@ static struct timer_list ipmr_expire_timer;
 
 /* Service routines creating virtual interfaces: DVMRP tunnels and PIMREG */
 
+static void ipmr_del_tunnel(struct net_device *dev, struct vifctl *v)
+{
+       struct net *net = dev_net(dev);
+
+       dev_close(dev);
+
+       dev = __dev_get_by_name(net, "tunl0");
+       if (dev) {
+               const struct net_device_ops *ops = dev->netdev_ops;
+               struct ifreq ifr;
+               struct ip_tunnel_parm p;
+
+               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 = (__force void __user *)&p;
+
+               if (ops->ndo_do_ioctl) {
+                       mm_segment_t oldfs = get_fs();
+
+                       set_fs(KERNEL_DS);
+                       ops->ndo_do_ioctl(dev, &ifr, SIOCDELTUNNEL);
+                       set_fs(oldfs);
+               }
+       }
+}
+
 static
-struct net_device *ipmr_new_tunnel(struct vifctl *v)
+struct net_device *ipmr_new_tunnel(struct net *net, struct vifctl *v)
 {
        struct net_device  *dev;
 
-       dev = __dev_get_by_name("tunl0");
+       dev = __dev_get_by_name(net, "tunl0");
 
        if (dev) {
+               const struct net_device_ops *ops = dev->netdev_ops;
                int err;
                struct ifreq ifr;
-               mm_segment_t    oldfs;
                struct ip_tunnel_parm p;
                struct in_device  *in_dev;
 
@@ -140,24 +159,33 @@ struct net_device *ipmr_new_tunnel(struct vifctl *v)
                p.iph.ihl = 5;
                p.iph.protocol = IPPROTO_IPIP;
                sprintf(p.name, "dvmrp%d", v->vifc_vifi);
-               ifr.ifr_ifru.ifru_data = (void*)&p;
+               ifr.ifr_ifru.ifru_data = (__force void __user *)&p;
 
-               oldfs = get_fs(); set_fs(KERNEL_DS);
-               err = dev->do_ioctl(dev, &ifr, SIOCADDTUNNEL);
-               set_fs(oldfs);
+               if (ops->ndo_do_ioctl) {
+                       mm_segment_t oldfs = get_fs();
+
+                       set_fs(KERNEL_DS);
+                       err = ops->ndo_do_ioctl(dev, &ifr, SIOCADDTUNNEL);
+                       set_fs(oldfs);
+               } else
+                       err = -EOPNOTSUPP;
 
                dev = NULL;
 
-               if (err == 0 && (dev = __dev_get_by_name(p.name)) != NULL) {
+               if (err == 0 &&
+                   (dev = __dev_get_by_name(net, p.name)) != NULL) {
                        dev->flags |= IFF_MULTICAST;
 
                        in_dev = __in_dev_get_rtnl(dev);
-                       if (in_dev == NULL && (in_dev = inetdev_init(dev)) == NULL)
+                       if (in_dev == NULL)
                                goto failure;
-                       in_dev->cnf.rp_filter = 0;
+
+                       ipv4_devconf_setall(in_dev);
+                       IPV4_DEVCONF(in_dev->cnf, RP_FILTER) = 0;
 
                        if (dev_open(dev))
                                goto failure;
+                       dev_hold(dev);
                }
        }
        return dev;
@@ -173,59 +201,67 @@ failure:
 
 #ifdef CONFIG_IP_PIMSM
 
-static int reg_vif_num = -1;
-
-static int reg_vif_xmit(struct sk_buff *skb, struct net_device *dev)
+static netdev_tx_t reg_vif_xmit(struct sk_buff *skb, struct net_device *dev)
 {
+       struct net *net = dev_net(dev);
+
        read_lock(&mrt_lock);
-       ((struct net_device_stats*)netdev_priv(dev))->tx_bytes += skb->len;
-       ((struct net_device_stats*)netdev_priv(dev))->tx_packets++;
-       ipmr_cache_report(skb, reg_vif_num, IGMPMSG_WHOLEPKT);
+       dev->stats.tx_bytes += skb->len;
+       dev->stats.tx_packets++;
+       ipmr_cache_report(net, skb, net->ipv4.mroute_reg_vif_num,
+                         IGMPMSG_WHOLEPKT);
        read_unlock(&mrt_lock);
        kfree_skb(skb);
-       return 0;
+       return NETDEV_TX_OK;
 }
 
-static struct net_device_stats *reg_vif_get_stats(struct net_device *dev)
-{
-       return (struct net_device_stats*)netdev_priv(dev);
-}
+static const struct net_device_ops reg_vif_netdev_ops = {
+       .ndo_start_xmit = reg_vif_xmit,
+};
 
 static void reg_vif_setup(struct net_device *dev)
 {
        dev->type               = ARPHRD_PIMREG;
        dev->mtu                = ETH_DATA_LEN - sizeof(struct iphdr) - 8;
        dev->flags              = IFF_NOARP;
-       dev->hard_start_xmit    = reg_vif_xmit;
-       dev->get_stats          = reg_vif_get_stats;
+       dev->netdev_ops         = &reg_vif_netdev_ops,
        dev->destructor         = free_netdev;
+       dev->features           |= NETIF_F_NETNS_LOCAL;
 }
 
-static struct net_device *ipmr_reg_vif(void)
+static struct net_device *ipmr_reg_vif(struct net *net)
 {
        struct net_device *dev;
        struct in_device *in_dev;
 
-       dev = alloc_netdev(sizeof(struct net_device_stats), "pimreg",
-                          reg_vif_setup);
+       dev = alloc_netdev(0, "pimreg", reg_vif_setup);
 
        if (dev == NULL)
                return NULL;
 
+       dev_net_set(dev, net);
+
        if (register_netdevice(dev)) {
                free_netdev(dev);
                return NULL;
        }
        dev->iflink = 0;
 
-       if ((in_dev = inetdev_init(dev)) == NULL)
+       rcu_read_lock();
+       if ((in_dev = __in_dev_get_rcu(dev)) == NULL) {
+               rcu_read_unlock();
                goto failure;
+       }
 
-       in_dev->cnf.rp_filter = 0;
+       ipv4_devconf_setall(in_dev);
+       IPV4_DEVCONF(in_dev->cnf, RP_FILTER) = 0;
+       rcu_read_unlock();
 
        if (dev_open(dev))
                goto failure;
 
+       dev_hold(dev);
+
        return dev;
 
 failure:
@@ -240,18 +276,19 @@ failure:
 
 /*
  *     Delete a VIF entry
+ *     @notify: Set to 1, if the caller is a notifier_call
  */
-static int vif_delete(int vifi)
+
+static int vif_delete(struct net *net, int vifi, int notify)
 {
        struct vif_device *v;
        struct net_device *dev;
        struct in_device *in_dev;
 
-       if (vifi < 0 || vifi >= maxvif)
+       if (vifi < 0 || vifi >= net->ipv4.maxvif)
                return -EADDRNOTAVAIL;
 
-       v = &vif_table[vifi];
+       v = &net->ipv4.vif_table[vifi];
 
        write_lock_bh(&mrt_lock);
        dev = v->dev;
@@ -263,17 +300,17 @@ static int vif_delete(int vifi)
        }
 
 #ifdef CONFIG_IP_PIMSM
-       if (vifi == reg_vif_num)
-               reg_vif_num = -1;
+       if (vifi == net->ipv4.mroute_reg_vif_num)
+               net->ipv4.mroute_reg_vif_num = -1;
 #endif
 
-       if (vifi+1 == maxvif) {
+       if (vifi+1 == net->ipv4.maxvif) {
                int tmp;
                for (tmp=vifi-1; tmp>=0; tmp--) {
-                       if (VIF_EXISTS(tmp))
+                       if (VIF_EXISTS(net, tmp))
                                break;
                }
-               maxvif = tmp+1;
+               net->ipv4.maxvif = tmp+1;
        }
 
        write_unlock_bh(&mrt_lock);
@@ -281,17 +318,23 @@ static int vif_delete(int vifi)
        dev_set_allmulti(dev, -1);
 
        if ((in_dev = __in_dev_get_rtnl(dev)) != NULL) {
-               in_dev->cnf.mc_forwarding--;
+               IPV4_DEVCONF(in_dev->cnf, MC_FORWARDING)--;
                ip_rt_multicast_event(in_dev);
        }
 
-       if (v->flags&(VIFF_TUNNEL|VIFF_REGISTER))
+       if (v->flags&(VIFF_TUNNEL|VIFF_REGISTER) && !notify)
                unregister_netdevice(dev);
 
        dev_put(dev);
        return 0;
 }
 
+static inline void ipmr_cache_free(struct mfc_cache *c)
+{
+       release_net(mfc_net(c));
+       kmem_cache_free(mrt_cachep, c);
+}
+
 /* Destroy an unresolved cache entry, killing queued skbs
    and reporting error to netlink readers.
  */
@@ -300,11 +343,12 @@ static void ipmr_destroy_unres(struct mfc_cache *c)
 {
        struct sk_buff *skb;
        struct nlmsgerr *e;
+       struct net *net = mfc_net(c);
 
-       atomic_dec(&cache_resolve_queue_len);
+       atomic_dec(&net->ipv4.cache_resolve_queue_len);
 
-       while((skb=skb_dequeue(&c->mfc_un.unres.unresolved))) {
-               if (skb->nh.iph->version == 0) {
+       while ((skb = skb_dequeue(&c->mfc_un.unres.unresolved))) {
+               if (ip_hdr(skb)->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));
@@ -312,12 +356,13 @@ static void ipmr_destroy_unres(struct mfc_cache *c)
                        e = NLMSG_DATA(nlh);
                        e->error = -ETIMEDOUT;
                        memset(&e->msg, 0, sizeof(e->msg));
-                       netlink_unicast(rtnl, skb, NETLINK_CB(skb).dst_pid, MSG_DONTWAIT);
+
+                       rtnl_unicast(skb, net, NETLINK_CB(skb).pid);
                } else
                        kfree_skb(skb);
        }
 
-       kmem_cache_free(mrt_cachep, c);
+       ipmr_cache_free(c);
 }
 
 
@@ -334,7 +379,7 @@ static void ipmr_expire_process(unsigned long dummy)
                return;
        }
 
-       if (atomic_read(&cache_resolve_queue_len) == 0)
+       if (mfc_unres_queue == NULL)
                goto out;
 
        now = jiffies;
@@ -355,7 +400,7 @@ static void ipmr_expire_process(unsigned long dummy)
                ipmr_destroy_unres(c);
        }
 
-       if (atomic_read(&cache_resolve_queue_len))
+       if (mfc_unres_queue != NULL)
                mod_timer(&ipmr_expire_timer, jiffies + expires);
 
 out:
@@ -367,13 +412,15 @@ out:
 static void ipmr_update_thresholds(struct mfc_cache *cache, unsigned char *ttls)
 {
        int vifi;
+       struct net *net = mfc_net(cache);
 
        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) {
+       for (vifi = 0; vifi < net->ipv4.maxvif; vifi++) {
+               if (VIF_EXISTS(net, 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;
@@ -383,15 +430,16 @@ static void ipmr_update_thresholds(struct mfc_cache *cache, unsigned char *ttls)
        }
 }
 
-static int vif_add(struct vifctl *vifc, int mrtsock)
+static int vif_add(struct net *net, struct vifctl *vifc, int mrtsock)
 {
        int vifi = vifc->vifc_vifi;
-       struct vif_device *v = &vif_table[vifi];
+       struct vif_device *v = &net->ipv4.vif_table[vifi];
        struct net_device *dev;
        struct in_device *in_dev;
+       int err;
 
        /* Is vif busy ? */
-       if (VIF_EXISTS(vifi))
+       if (VIF_EXISTS(net, vifi))
                return -EADDRINUSE;
 
        switch (vifc->vifc_flags) {
@@ -401,23 +449,39 @@ static int vif_add(struct vifctl *vifc, int mrtsock)
                 * Special Purpose VIF in PIM
                 * All the packets will be sent to the daemon
                 */
-               if (reg_vif_num >= 0)
+               if (net->ipv4.mroute_reg_vif_num >= 0)
                        return -EADDRINUSE;
-               dev = ipmr_reg_vif();
+               dev = ipmr_reg_vif(net);
                if (!dev)
                        return -ENOBUFS;
+               err = dev_set_allmulti(dev, 1);
+               if (err) {
+                       unregister_netdevice(dev);
+                       dev_put(dev);
+                       return err;
+               }
                break;
 #endif
-       case VIFF_TUNNEL:       
-               dev = ipmr_new_tunnel(vifc);
+       case VIFF_TUNNEL:
+               dev = ipmr_new_tunnel(net, vifc);
                if (!dev)
                        return -ENOBUFS;
+               err = dev_set_allmulti(dev, 1);
+               if (err) {
+                       ipmr_del_tunnel(dev, vifc);
+                       dev_put(dev);
+                       return err;
+               }
                break;
        case 0:
-               dev = ip_dev_find(vifc->vifc_lcl_addr.s_addr);
+               dev = ip_dev_find(net, vifc->vifc_lcl_addr.s_addr);
                if (!dev)
                        return -EADDRNOTAVAIL;
-               dev_put(dev);
+               err = dev_set_allmulti(dev, 1);
+               if (err) {
+                       dev_put(dev);
+                       return err;
+               }
                break;
        default:
                return -EINVAL;
@@ -425,20 +489,19 @@ static int vif_add(struct vifctl *vifc, int mrtsock)
 
        if ((in_dev = __in_dev_get_rtnl(dev)) == NULL)
                return -EADDRNOTAVAIL;
-       in_dev->cnf.mc_forwarding++;
-       dev_set_allmulti(dev, +1);
+       IPV4_DEVCONF(in_dev->cnf, MC_FORWARDING)++;
        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;
+       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->threshold = vifc->vifc_threshold;
        v->bytes_in = 0;
        v->bytes_out = 0;
        v->pkt_in = 0;
@@ -449,24 +512,25 @@ static int vif_add(struct vifctl *vifc, int mrtsock)
 
        /* And finish update writing critical data */
        write_lock_bh(&mrt_lock);
-       dev_hold(dev);
-       v->dev=dev;
+       v->dev = dev;
 #ifdef CONFIG_IP_PIMSM
        if (v->flags&VIFF_REGISTER)
-               reg_vif_num = vifi;
+               net->ipv4.mroute_reg_vif_num = vifi;
 #endif
-       if (vifi+1 > maxvif)
-               maxvif = vifi+1;
+       if (vifi+1 > net->ipv4.maxvif)
+               net->ipv4.maxvif = vifi+1;
        write_unlock_bh(&mrt_lock);
        return 0;
 }
 
-static struct mfc_cache *ipmr_cache_find(__u32 origin, __u32 mcastgrp)
+static struct mfc_cache *ipmr_cache_find(struct net *net,
+                                        __be32 origin,
+                                        __be32 mcastgrp)
 {
-       int line=MFC_HASH(mcastgrp,origin);
+       int line = MFC_HASH(mcastgrp, origin);
        struct mfc_cache *c;
 
-       for (c=mfc_cache_array[line]; c; c = c->next) {
+       for (c = net->ipv4.mfc_cache_array[line]; c; c = c->next) {
                if (c->mfc_origin==origin && c->mfc_mcastgrp==mcastgrp)
                        break;
        }
@@ -476,31 +540,31 @@ static struct mfc_cache *ipmr_cache_find(__u32 origin, __u32 mcastgrp)
 /*
  *     Allocate a multicast cache entry
  */
-static struct mfc_cache *ipmr_cache_alloc(void)
+static struct mfc_cache *ipmr_cache_alloc(struct net *net)
 {
-       struct mfc_cache *c=kmem_cache_alloc(mrt_cachep, GFP_KERNEL);
-       if(c==NULL)
+       struct mfc_cache *c = kmem_cache_zalloc(mrt_cachep, GFP_KERNEL);
+       if (c == NULL)
                return NULL;
-       memset(c, 0, sizeof(*c));
        c->mfc_un.res.minvif = MAXVIFS;
+       mfc_net_set(c, net);
        return c;
 }
 
-static struct mfc_cache *ipmr_cache_alloc_unres(void)
+static struct mfc_cache *ipmr_cache_alloc_unres(struct net *net)
 {
-       struct mfc_cache *c=kmem_cache_alloc(mrt_cachep, GFP_ATOMIC);
-       if(c==NULL)
+       struct mfc_cache *c = kmem_cache_zalloc(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;
+       mfc_net_set(c, net);
        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;
@@ -510,13 +574,13 @@ static void ipmr_cache_resolve(struct mfc_cache *uc, struct mfc_cache *c)
         *      Play the pending entries through our router
         */
 
-       while((skb=__skb_dequeue(&uc->mfc_un.unres.unresolved))) {
-               if (skb->nh.iph->version == 0) {
-                       int err;
+       while ((skb = __skb_dequeue(&uc->mfc_un.unres.unresolved))) {
+               if (ip_hdr(skb)->version == 0) {
                        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;
+                               nlh->nlmsg_len = (skb_tail_pointer(skb) -
+                                                 (u8 *)nlh);
                        } else {
                                nlh->nlmsg_type = NLMSG_ERROR;
                                nlh->nlmsg_len = NLMSG_LENGTH(sizeof(struct nlmsgerr));
@@ -525,7 +589,8 @@ static void ipmr_cache_resolve(struct mfc_cache *uc, struct mfc_cache *c)
                                e->error = -EMSGSIZE;
                                memset(&e->msg, 0, sizeof(e->msg));
                        }
-                       err = netlink_unicast(rtnl, skb, NETLINK_CB(skb).dst_pid, MSG_DONTWAIT);
+
+                       rtnl_unicast(skb, mfc_net(c), NETLINK_CB(skb).pid);
                } else
                        ip_mr_forward(skb, c, 0);
        }
@@ -537,11 +602,12 @@ static void ipmr_cache_resolve(struct mfc_cache *uc, struct mfc_cache *c)
  *
  *     Called under mrt_lock.
  */
-static int ipmr_cache_report(struct sk_buff *pkt, vifi_t vifi, int assert)
+
+static int ipmr_cache_report(struct net *net,
+                            struct sk_buff *pkt, vifi_t vifi, int assert)
 {
        struct sk_buff *skb;
-       int ihl = pkt->nh.iph->ihl<<2;
+       const int ihl = ip_hdrlen(pkt);
        struct igmphdr *igmp;
        struct igmpmsg *msg;
        int ret;
@@ -553,7 +619,7 @@ static int ipmr_cache_report(struct sk_buff *pkt, vifi_t vifi, int assert)
 #endif
                skb = alloc_skb(128, GFP_ATOMIC);
 
-       if(!skb)
+       if (!skb)
                return -ENOBUFS;
 
 #ifdef CONFIG_IP_PIMSM
@@ -563,42 +629,46 @@ static int ipmr_cache_report(struct sk_buff *pkt, vifi_t vifi, int assert)
                   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));
+               skb_push(skb, sizeof(struct iphdr));
+               skb_reset_network_header(skb);
+               skb_reset_transport_header(skb);
+               msg = (struct igmpmsg *)skb_network_header(skb);
+               memcpy(msg, skb_network_header(pkt), 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 
+               msg->im_vif = net->ipv4.mroute_reg_vif_num;
+               ip_hdr(skb)->ihl = sizeof(struct iphdr) >> 2;
+               ip_hdr(skb)->tot_len = htons(ntohs(ip_hdr(pkt)->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;
+       skb->network_header = skb->tail;
+       skb_put(skb, ihl);
+       skb_copy_to_linear_data(skb, pkt->data, ihl);
+       ip_hdr(skb)->protocol = 0;                      /* Flag to the kernel this is a route add */
+       msg = (struct igmpmsg *)skb_network_header(skb);
        msg->im_vif = vifi;
-       skb->dst = dst_clone(pkt->dst);
+       skb_dst_set(skb, dst_clone(skb_dst(pkt)));
 
        /*
         *      Add our header
         */
 
-       igmp=(struct igmphdr *)skb_put(skb,sizeof(struct igmphdr));
+       igmp=(struct igmphdr *)skb_put(skb, sizeof(struct igmphdr));
        igmp->type      =
        msg->im_msgtype = assert;
        igmp->code      =       0;
-       skb->nh.iph->tot_len=htons(skb->len);                   /* Fix the length */
-       skb->h.raw = skb->nh.raw;
-        }
+       ip_hdr(skb)->tot_len = htons(skb->len);                 /* Fix the length */
+       skb->transport_header = skb->network_header;
+       }
 
-       if (mroute_socket == NULL) {
+       if (net->ipv4.mroute_sk == NULL) {
                kfree_skb(skb);
                return -EINVAL;
        }
@@ -606,7 +676,8 @@ static int ipmr_cache_report(struct sk_buff *pkt, vifi_t vifi, int assert)
        /*
         *      Deliver to mrouted
         */
-       if ((ret=sock_queue_rcv_skb(mroute_socket,skb))<0) {
+       ret = sock_queue_rcv_skb(net->ipv4.mroute_sk, skb);
+       if (ret < 0) {
                if (net_ratelimit())
                        printk(KERN_WARNING "mroute: pending queue full, dropping entries.\n");
                kfree_skb(skb);
@@ -618,17 +689,19 @@ static int ipmr_cache_report(struct sk_buff *pkt, vifi_t vifi, int assert)
 /*
  *     Queue a packet for resolution. It gets locked cache entry!
  */
+
 static int
-ipmr_cache_unresolved(vifi_t vifi, struct sk_buff *skb)
+ipmr_cache_unresolved(struct net *net, vifi_t vifi, struct sk_buff *skb)
 {
        int err;
        struct mfc_cache *c;
+       const struct iphdr *iph = ip_hdr(skb);
 
        spin_lock_bh(&mfc_unres_lock);
        for (c=mfc_unres_queue; c; c=c->next) {
-               if (c->mfc_mcastgrp == skb->nh.iph->daddr &&
-                   c->mfc_origin == skb->nh.iph->saddr)
+               if (net_eq(mfc_net(c), net) &&
+                   c->mfc_mcastgrp == iph->daddr &&
+                   c->mfc_origin == iph->saddr)
                        break;
        }
 
@@ -637,8 +710,8 @@ ipmr_cache_unresolved(vifi_t vifi, struct sk_buff *skb)
                 *      Create a new entry if allowable
                 */
 
-               if (atomic_read(&cache_resolve_queue_len)>=10 ||
-                   (c=ipmr_cache_alloc_unres())==NULL) {
+               if (atomic_read(&net->ipv4.cache_resolve_queue_len) >= 10 ||
+                   (c = ipmr_cache_alloc_unres(net)) == NULL) {
                        spin_unlock_bh(&mfc_unres_lock);
 
                        kfree_skb(skb);
@@ -648,25 +721,26 @@ ipmr_cache_unresolved(vifi_t vifi, struct sk_buff *skb)
                /*
                 *      Fill in the new cache entry
                 */
-               c->mfc_parent=-1;
-               c->mfc_origin=skb->nh.iph->saddr;
-               c->mfc_mcastgrp=skb->nh.iph->daddr;
+               c->mfc_parent   = -1;
+               c->mfc_origin   = iph->saddr;
+               c->mfc_mcastgrp = iph->daddr;
 
                /*
                 *      Reflect first query at mrouted.
                 */
-               if ((err = ipmr_cache_report(skb, vifi, IGMPMSG_NOCACHE))<0) {
-                       /* If the report failed throw the cache entry 
+               err = ipmr_cache_report(net, skb, vifi, IGMPMSG_NOCACHE);
+               if (err < 0) {
+                       /* If the report failed throw the cache entry
                           out - Brad Parker
                         */
                        spin_unlock_bh(&mfc_unres_lock);
 
-                       kmem_cache_free(mrt_cachep, c);
+                       ipmr_cache_free(c);
                        kfree_skb(skb);
                        return err;
                }
 
-               atomic_inc(&cache_resolve_queue_len);
+               atomic_inc(&net->ipv4.cache_resolve_queue_len);
                c->next = mfc_unres_queue;
                mfc_unres_queue = c;
 
@@ -680,7 +754,7 @@ ipmr_cache_unresolved(vifi_t vifi, struct sk_buff *skb)
                kfree_skb(skb);
                err = -ENOBUFS;
        } else {
-               skb_queue_tail(&c->mfc_un.unres.unresolved,skb);
+               skb_queue_tail(&c->mfc_un.unres.unresolved, skb);
                err = 0;
        }
 
@@ -692,35 +766,37 @@ ipmr_cache_unresolved(vifi_t vifi, struct sk_buff *skb)
  *     MFC cache manipulation by user space mroute daemon
  */
 
-static int ipmr_mfc_delete(struct mfcctl *mfc)
+static int ipmr_mfc_delete(struct net *net, struct mfcctl *mfc)
 {
        int line;
        struct mfc_cache *c, **cp;
 
-       line=MFC_HASH(mfc->mfcc_mcastgrp.s_addr, mfc->mfcc_origin.s_addr);
+       line = MFC_HASH(mfc->mfcc_mcastgrp.s_addr, mfc->mfcc_origin.s_addr);
 
-       for (cp=&mfc_cache_array[line]; (c=*cp) != NULL; cp = &c->next) {
+       for (cp = &net->ipv4.mfc_cache_array[line];
+            (c = *cp) != NULL; cp = &c->next) {
                if (c->mfc_origin == mfc->mfcc_origin.s_addr &&
                    c->mfc_mcastgrp == mfc->mfcc_mcastgrp.s_addr) {
                        write_lock_bh(&mrt_lock);
                        *cp = c->next;
                        write_unlock_bh(&mrt_lock);
 
-                       kmem_cache_free(mrt_cachep, c);
+                       ipmr_cache_free(c);
                        return 0;
                }
        }
        return -ENOENT;
 }
 
-static int ipmr_mfc_add(struct mfcctl *mfc, int mrtsock)
+static int ipmr_mfc_add(struct net *net, struct mfcctl *mfc, int mrtsock)
 {
        int line;
        struct mfc_cache *uc, *c, **cp;
 
-       line=MFC_HASH(mfc->mfcc_mcastgrp.s_addr, mfc->mfcc_origin.s_addr);
+       line = MFC_HASH(mfc->mfcc_mcastgrp.s_addr, mfc->mfcc_origin.s_addr);
 
-       for (cp=&mfc_cache_array[line]; (c=*cp) != NULL; cp = &c->next) {
+       for (cp = &net->ipv4.mfc_cache_array[line];
+            (c = *cp) != NULL; cp = &c->next) {
                if (c->mfc_origin == mfc->mfcc_origin.s_addr &&
                    c->mfc_mcastgrp == mfc->mfcc_mcastgrp.s_addr)
                        break;
@@ -736,23 +812,23 @@ static int ipmr_mfc_add(struct mfcctl *mfc, int mrtsock)
                return 0;
        }
 
-       if(!MULTICAST(mfc->mfcc_mcastgrp.s_addr))
+       if (!ipv4_is_multicast(mfc->mfcc_mcastgrp.s_addr))
                return -EINVAL;
 
-       c=ipmr_cache_alloc();
-       if (c==NULL)
+       c = ipmr_cache_alloc(net);
+       if (c == NULL)
                return -ENOMEM;
 
-       c->mfc_origin=mfc->mfcc_origin.s_addr;
-       c->mfc_mcastgrp=mfc->mfcc_mcastgrp.s_addr;
-       c->mfc_parent=mfc->mfcc_parent;
+       c->mfc_origin = mfc->mfcc_origin.s_addr;
+       c->mfc_mcastgrp = mfc->mfcc_mcastgrp.s_addr;
+       c->mfc_parent = mfc->mfcc_parent;
        ipmr_update_thresholds(c, mfc->mfcc_ttls);
        if (!mrtsock)
                c->mfc_flags |= MFC_STATIC;
 
        write_lock_bh(&mrt_lock);
-       c->next = mfc_cache_array[line];
-       mfc_cache_array[line] = c;
+       c->next = net->ipv4.mfc_cache_array[line];
+       net->ipv4.mfc_cache_array[line] = c;
        write_unlock_bh(&mrt_lock);
 
        /*
@@ -762,19 +838,21 @@ static int ipmr_mfc_add(struct mfcctl *mfc, int mrtsock)
        spin_lock_bh(&mfc_unres_lock);
        for (cp = &mfc_unres_queue; (uc=*cp) != NULL;
             cp = &uc->next) {
-               if (uc->mfc_origin == c->mfc_origin &&
+               if (net_eq(mfc_net(uc), net) &&
+                   uc->mfc_origin == c->mfc_origin &&
                    uc->mfc_mcastgrp == c->mfc_mcastgrp) {
                        *cp = uc->next;
-                       if (atomic_dec_and_test(&cache_resolve_queue_len))
-                               del_timer(&ipmr_expire_timer);
+                       atomic_dec(&net->ipv4.cache_resolve_queue_len);
                        break;
                }
        }
+       if (mfc_unres_queue == NULL)
+               del_timer(&ipmr_expire_timer);
        spin_unlock_bh(&mfc_unres_lock);
 
        if (uc) {
                ipmr_cache_resolve(uc, c);
-               kmem_cache_free(mrt_cachep, uc);
+               ipmr_cache_free(uc);
        }
        return 0;
 }
@@ -782,26 +860,26 @@ static int ipmr_mfc_add(struct mfcctl *mfc, int mrtsock)
 /*
  *     Close the multicast socket, and clear the vif tables etc
  */
-static void mroute_clean_tables(struct sock *sk)
+
+static void mroute_clean_tables(struct net *net)
 {
        int i;
-               
+
        /*
         *      Shut down all active vif entries
         */
-       for(i=0; i<maxvif; i++) {
-               if (!(vif_table[i].flags&VIFF_STATIC))
-                       vif_delete(i);
+       for (i = 0; i < net->ipv4.maxvif; i++) {
+               if (!(net->ipv4.vif_table[i].flags&VIFF_STATIC))
+                       vif_delete(net, i, 0);
        }
 
        /*
         *      Wipe the cache
         */
-       for (i=0;i<MFC_LINES;i++) {
+       for (i=0; i<MFC_LINES; i++) {
                struct mfc_cache *c, **cp;
 
-               cp = &mfc_cache_array[i];
+               cp = &net->ipv4.mfc_cache_array[i];
                while ((c = *cp) != NULL) {
                        if (c->mfc_flags&MFC_STATIC) {
                                cp = &c->next;
@@ -811,22 +889,23 @@ static void mroute_clean_tables(struct sock *sk)
                        *cp = c->next;
                        write_unlock_bh(&mrt_lock);
 
-                       kmem_cache_free(mrt_cachep, c);
+                       ipmr_cache_free(c);
                }
        }
 
-       if (atomic_read(&cache_resolve_queue_len) != 0) {
-               struct mfc_cache *c;
+       if (atomic_read(&net->ipv4.cache_resolve_queue_len) != 0) {
+               struct mfc_cache *c, **cp;
 
                spin_lock_bh(&mfc_unres_lock);
-               while (mfc_unres_queue != NULL) {
-                       c = mfc_unres_queue;
-                       mfc_unres_queue = c->next;
-                       spin_unlock_bh(&mfc_unres_lock);
+               cp = &mfc_unres_queue;
+               while ((c = *cp) != NULL) {
+                       if (!net_eq(mfc_net(c), net)) {
+                               cp = &c->next;
+                               continue;
+                       }
+                       *cp = c->next;
 
                        ipmr_destroy_unres(c);
-
-                       spin_lock_bh(&mfc_unres_lock);
                }
                spin_unlock_bh(&mfc_unres_lock);
        }
@@ -834,15 +913,17 @@ static void mroute_clean_tables(struct sock *sk)
 
 static void mrtsock_destruct(struct sock *sk)
 {
+       struct net *net = sock_net(sk);
+
        rtnl_lock();
-       if (sk == mroute_socket) {
-               ipv4_devconf.mc_forwarding--;
+       if (sk == net->ipv4.mroute_sk) {
+               IPV4_DEVCONF_ALL(net, MC_FORWARDING)--;
 
                write_lock_bh(&mrt_lock);
-               mroute_socket=NULL;
+               net->ipv4.mroute_sk = NULL;
                write_unlock_bh(&mrt_lock);
 
-               mroute_clean_tables(sk);
+               mroute_clean_tables(net);
        }
        rtnl_unlock();
 }
@@ -853,139 +934,131 @@ static void mrtsock_destruct(struct sock *sk)
  *     that's how BSD mrouted happens to think. Maybe one day with a proper
  *     MOSPF/PIM router set up we can clean this up.
  */
-int ip_mroute_setsockopt(struct sock *sk,int optname,char __user *optval,int optlen)
+
+int ip_mroute_setsockopt(struct sock *sk, int optname, char __user *optval, int optlen)
 {
        int ret;
        struct vifctl vif;
        struct mfcctl mfc;
-       
-       if(optname!=MRT_INIT)
-       {
-               if(sk!=mroute_socket && !capable(CAP_NET_ADMIN))
+       struct net *net = sock_net(sk);
+
+       if (optname != MRT_INIT) {
+               if (sk != net->ipv4.mroute_sk && !capable(CAP_NET_ADMIN))
                        return -EACCES;
        }
 
-       switch(optname)
-       {
-               case MRT_INIT:
-                       if (sk->sk_type != SOCK_RAW ||
-                           inet_sk(sk)->num != IPPROTO_IGMP)
-                               return -EOPNOTSUPP;
-                       if(optlen!=sizeof(int))
-                               return -ENOPROTOOPT;
-
-                       rtnl_lock();
-                       if (mroute_socket) {
-                               rtnl_unlock();
-                               return -EADDRINUSE;
-                       }
-
-                       ret = ip_ra_control(sk, 1, mrtsock_destruct);
-                       if (ret == 0) {
-                               write_lock_bh(&mrt_lock);
-                               mroute_socket=sk;
-                               write_unlock_bh(&mrt_lock);
+       switch (optname) {
+       case MRT_INIT:
+               if (sk->sk_type != SOCK_RAW ||
+                   inet_sk(sk)->num != IPPROTO_IGMP)
+                       return -EOPNOTSUPP;
+               if (optlen != sizeof(int))
+                       return -ENOPROTOOPT;
 
-                               ipv4_devconf.mc_forwarding++;
-                       }
+               rtnl_lock();
+               if (net->ipv4.mroute_sk) {
                        rtnl_unlock();
-                       return ret;
-               case MRT_DONE:
-                       if (sk!=mroute_socket)
-                               return -EACCES;
-                       return ip_ra_control(sk, 0, NULL);
-               case MRT_ADD_VIF:
-               case MRT_DEL_VIF:
-                       if(optlen!=sizeof(vif))
-                               return -EINVAL;
-                       if (copy_from_user(&vif,optval,sizeof(vif)))
-                               return -EFAULT; 
-                       if(vif.vifc_vifi >= MAXVIFS)
-                               return -ENFILE;
-                       rtnl_lock();
-                       if (optname==MRT_ADD_VIF) {
-                               ret = vif_add(&vif, sk==mroute_socket);
-                       } else {
-                               ret = vif_delete(vif.vifc_vifi);
-                       }
-                       rtnl_unlock();
-                       return ret;
+                       return -EADDRINUSE;
+               }
+
+               ret = ip_ra_control(sk, 1, mrtsock_destruct);
+               if (ret == 0) {
+                       write_lock_bh(&mrt_lock);
+                       net->ipv4.mroute_sk = sk;
+                       write_unlock_bh(&mrt_lock);
+
+                       IPV4_DEVCONF_ALL(net, MC_FORWARDING)++;
+               }
+               rtnl_unlock();
+               return ret;
+       case MRT_DONE:
+               if (sk != net->ipv4.mroute_sk)
+                       return -EACCES;
+               return ip_ra_control(sk, 0, NULL);
+       case MRT_ADD_VIF:
+       case MRT_DEL_VIF:
+               if (optlen != sizeof(vif))
+                       return -EINVAL;
+               if (copy_from_user(&vif, optval, sizeof(vif)))
+                       return -EFAULT;
+               if (vif.vifc_vifi >= MAXVIFS)
+                       return -ENFILE;
+               rtnl_lock();
+               if (optname == MRT_ADD_VIF) {
+                       ret = vif_add(net, &vif, sk == net->ipv4.mroute_sk);
+               } else {
+                       ret = vif_delete(net, vif.vifc_vifi, 0);
+               }
+               rtnl_unlock();
+               return ret;
 
                /*
                 *      Manipulate the forwarding caches. These live
                 *      in a sort of kernel/user symbiosis.
                 */
-               case MRT_ADD_MFC:
-               case MRT_DEL_MFC:
-                       if(optlen!=sizeof(mfc))
-                               return -EINVAL;
-                       if (copy_from_user(&mfc,optval, sizeof(mfc)))
-                               return -EFAULT;
-                       rtnl_lock();
-                       if (optname==MRT_DEL_MFC)
-                               ret = ipmr_mfc_delete(&mfc);
-                       else
-                               ret = ipmr_mfc_add(&mfc, sk==mroute_socket);
-                       rtnl_unlock();
-                       return ret;
+       case MRT_ADD_MFC:
+       case MRT_DEL_MFC:
+               if (optlen != sizeof(mfc))
+                       return -EINVAL;
+               if (copy_from_user(&mfc, optval, sizeof(mfc)))
+                       return -EFAULT;
+               rtnl_lock();
+               if (optname == MRT_DEL_MFC)
+                       ret = ipmr_mfc_delete(net, &mfc);
+               else
+                       ret = ipmr_mfc_add(net, &mfc, sk == net->ipv4.mroute_sk);
+               rtnl_unlock();
+               return ret;
                /*
                 *      Control PIM assert.
                 */
-               case MRT_ASSERT:
-               {
-                       int v;
-                       if(get_user(v,(int __user *)optval))
-                               return -EFAULT;
-                       mroute_do_assert=(v)?1:0;
-                       return 0;
-               }
+       case MRT_ASSERT:
+       {
+               int v;
+               if (get_user(v,(int __user *)optval))
+                       return -EFAULT;
+               net->ipv4.mroute_do_assert = (v) ? 1 : 0;
+               return 0;
+       }
 #ifdef CONFIG_IP_PIMSM
-               case MRT_PIM:
-               {
-                       int v, ret;
-                       if(get_user(v,(int __user *)optval))
-                               return -EFAULT;
-                       v = (v)?1:0;
-                       rtnl_lock();
-                       ret = 0;
-                       if (v != mroute_do_pim) {
-                               mroute_do_pim = v;
-                               mroute_do_assert = v;
-#ifdef CONFIG_IP_PIMSM_V2
-                               if (mroute_do_pim)
-                                       ret = inet_add_protocol(&pim_protocol,
-                                                               IPPROTO_PIM);
-                               else
-                                       ret = inet_del_protocol(&pim_protocol,
-                                                               IPPROTO_PIM);
-                               if (ret < 0)
-                                       ret = -EAGAIN;
-#endif
-                       }
-                       rtnl_unlock();
-                       return ret;
+       case MRT_PIM:
+       {
+               int v;
+
+               if (get_user(v,(int __user *)optval))
+                       return -EFAULT;
+               v = (v) ? 1 : 0;
+
+               rtnl_lock();
+               ret = 0;
+               if (v != net->ipv4.mroute_do_pim) {
+                       net->ipv4.mroute_do_pim = v;
+                       net->ipv4.mroute_do_assert = v;
                }
+               rtnl_unlock();
+               return ret;
+       }
 #endif
-               /*
-                *      Spurious command, or MRT_VERSION which you cannot
-                *      set.
-                */
-               default:
-                       return -ENOPROTOOPT;
+       /*
+        *      Spurious command, or MRT_VERSION which you cannot
+        *      set.
+        */
+       default:
+               return -ENOPROTOOPT;
        }
 }
 
 /*
  *     Getsock opt support for the multicast routing system.
  */
-int ip_mroute_getsockopt(struct sock *sk,int optname,char __user *optval,int __user *optlen)
+
+int ip_mroute_getsockopt(struct sock *sk, int optname, char __user *optval, int __user *optlen)
 {
        int olr;
        int val;
+       struct net *net = sock_net(sk);
 
-       if(optname!=MRT_VERSION && 
+       if (optname != MRT_VERSION &&
 #ifdef CONFIG_IP_PIMSM
           optname!=MRT_PIM &&
 #endif
@@ -998,18 +1071,18 @@ int ip_mroute_getsockopt(struct sock *sk,int optname,char __user *optval,int __u
        olr = min_t(unsigned int, olr, sizeof(int));
        if (olr < 0)
                return -EINVAL;
-               
-       if(put_user(olr,optlen))
+
+       if (put_user(olr, optlen))
                return -EFAULT;
-       if(optname==MRT_VERSION)
-               val=0x0305;
+       if (optname == MRT_VERSION)
+               val = 0x0305;
 #ifdef CONFIG_IP_PIMSM
-       else if(optname==MRT_PIM)
-               val=mroute_do_pim;
+       else if (optname == MRT_PIM)
+               val = net->ipv4.mroute_do_pim;
 #endif
        else
-               val=mroute_do_assert;
-       if(copy_to_user(optval,&val,olr))
+               val = net->ipv4.mroute_do_assert;
+       if (copy_to_user(optval, &val, olr))
                return -EFAULT;
        return 0;
 }
@@ -1017,76 +1090,82 @@ int ip_mroute_getsockopt(struct sock *sk,int optname,char __user *optval,int __u
 /*
  *     The IP multicast ioctl support routines.
  */
+
 int ipmr_ioctl(struct sock *sk, int cmd, void __user *arg)
 {
        struct sioc_sg_req sr;
        struct sioc_vif_req vr;
        struct vif_device *vif;
        struct mfc_cache *c;
-       
-       switch(cmd)
-       {
-               case SIOCGETVIFCNT:
-                       if (copy_from_user(&vr,arg,sizeof(vr)))
-                               return -EFAULT; 
-                       if(vr.vifi>=maxvif)
-                               return -EINVAL;
-                       read_lock(&mrt_lock);
-                       vif=&vif_table[vr.vifi];
-                       if(VIF_EXISTS(vr.vifi)) {
-                               vr.icount=vif->pkt_in;
-                               vr.ocount=vif->pkt_out;
-                               vr.ibytes=vif->bytes_in;
-                               vr.obytes=vif->bytes_out;
-                               read_unlock(&mrt_lock);
-
-                               if (copy_to_user(arg,&vr,sizeof(vr)))
-                                       return -EFAULT;
-                               return 0;
-                       }
+       struct net *net = sock_net(sk);
+
+       switch (cmd) {
+       case SIOCGETVIFCNT:
+               if (copy_from_user(&vr, arg, sizeof(vr)))
+                       return -EFAULT;
+               if (vr.vifi >= net->ipv4.maxvif)
+                       return -EINVAL;
+               read_lock(&mrt_lock);
+               vif = &net->ipv4.vif_table[vr.vifi];
+               if (VIF_EXISTS(net, vr.vifi)) {
+                       vr.icount = vif->pkt_in;
+                       vr.ocount = vif->pkt_out;
+                       vr.ibytes = vif->bytes_in;
+                       vr.obytes = vif->bytes_out;
                        read_unlock(&mrt_lock);
-                       return -EADDRNOTAVAIL;
-               case SIOCGETSGCNT:
-                       if (copy_from_user(&sr,arg,sizeof(sr)))
-                               return -EFAULT;
 
-                       read_lock(&mrt_lock);
-                       c = ipmr_cache_find(sr.src.s_addr, sr.grp.s_addr);
-                       if (c) {
-                               sr.pktcnt = c->mfc_un.res.pkt;
-                               sr.bytecnt = c->mfc_un.res.bytes;
-                               sr.wrong_if = c->mfc_un.res.wrong_if;
-                               read_unlock(&mrt_lock);
-
-                               if (copy_to_user(arg,&sr,sizeof(sr)))
-                                       return -EFAULT;
-                               return 0;
-                       }
+                       if (copy_to_user(arg, &vr, sizeof(vr)))
+                               return -EFAULT;
+                       return 0;
+               }
+               read_unlock(&mrt_lock);
+               return -EADDRNOTAVAIL;
+       case SIOCGETSGCNT:
+               if (copy_from_user(&sr, arg, sizeof(sr)))
+                       return -EFAULT;
+
+               read_lock(&mrt_lock);
+               c = ipmr_cache_find(net, sr.src.s_addr, sr.grp.s_addr);
+               if (c) {
+                       sr.pktcnt = c->mfc_un.res.pkt;
+                       sr.bytecnt = c->mfc_un.res.bytes;
+                       sr.wrong_if = c->mfc_un.res.wrong_if;
                        read_unlock(&mrt_lock);
-                       return -EADDRNOTAVAIL;
-               default:
-                       return -ENOIOCTLCMD;
+
+                       if (copy_to_user(arg, &sr, sizeof(sr)))
+                               return -EFAULT;
+                       return 0;
+               }
+               read_unlock(&mrt_lock);
+               return -EADDRNOTAVAIL;
+       default:
+               return -ENOIOCTLCMD;
        }
 }
 
 
 static int ipmr_device_event(struct notifier_block *this, unsigned long event, void *ptr)
 {
+       struct net_device *dev = ptr;
+       struct net *net = dev_net(dev);
        struct vif_device *v;
        int ct;
+
+       if (!net_eq(dev_net(dev), net))
+               return NOTIFY_DONE;
+
        if (event != NETDEV_UNREGISTER)
                return NOTIFY_DONE;
-       v=&vif_table[0];
-       for(ct=0;ct<maxvif;ct++,v++) {
-               if (v->dev==ptr)
-                       vif_delete(ct);
+       v = &net->ipv4.vif_table[0];
+       for (ct = 0; ct < net->ipv4.maxvif; ct++, v++) {
+               if (v->dev == dev)
+                       vif_delete(net, ct, 1);
        }
        return NOTIFY_DONE;
 }
 
 
-static struct notifier_block ip_mr_notifier={
+static struct notifier_block ip_mr_notifier = {
        .notifier_call = ipmr_device_event,
 };
 
@@ -1095,25 +1174,29 @@ static struct notifier_block ip_mr_notifier={
  *     This avoids tunnel drivers and other mess and gives us the speed so
  *     important for multicast video.
  */
-static void ip_encap(struct sk_buff *skb, u32 saddr, u32 daddr)
+
+static void ip_encap(struct sk_buff *skb, __be32 saddr, __be32 daddr)
 {
-       struct iphdr *iph = (struct iphdr *)skb_push(skb,sizeof(struct iphdr));
+       struct iphdr *iph;
+       struct iphdr *old_iph = ip_hdr(skb);
+
+       skb_push(skb, sizeof(struct iphdr));
+       skb->transport_header = skb->network_header;
+       skb_reset_network_header(skb);
+       iph = ip_hdr(skb);
 
        iph->version    =       4;
-       iph->tos        =       skb->nh.iph->tos;
-       iph->ttl        =       skb->nh.iph->ttl;
+       iph->tos        =       old_iph->tos;
+       iph->ttl        =       old_iph->ttl;
        iph->frag_off   =       0;
        iph->daddr      =       daddr;
        iph->saddr      =       saddr;
        iph->protocol   =       IPPROTO_IPIP;
        iph->ihl        =       5;
        iph->tot_len    =       htons(skb->len);
-       ip_select_ident(iph, skb->dst, NULL);
+       ip_select_ident(iph, skb_dst(skb), NULL);
        ip_send_check(iph);
 
-       skb->h.ipiph = skb->nh.iph;
-       skb->nh.iph = iph;
        memset(&(IPCB(skb)->opt), 0, sizeof(IPCB(skb)->opt));
        nf_reset(skb);
 }
@@ -1122,7 +1205,7 @@ static inline int ipmr_forward_finish(struct sk_buff *skb)
 {
        struct ip_options * opt = &(IPCB(skb)->opt);
 
-       IP_INC_STATS_BH(IPSTATS_MIB_OUTFORWDATAGRAMS);
+       IP_INC_STATS_BH(dev_net(skb_dst(skb)->dev), IPSTATS_MIB_OUTFORWDATAGRAMS);
 
        if (unlikely(opt->optlen))
                ip_forward_options(skb);
@@ -1136,8 +1219,9 @@ static inline int ipmr_forward_finish(struct sk_buff *skb)
 
 static void ipmr_queue_xmit(struct sk_buff *skb, struct mfc_cache *c, int vifi)
 {
-       struct iphdr *iph = skb->nh.iph;
-       struct vif_device *vif = &vif_table[vifi];
+       struct net *net = mfc_net(c);
+       const struct iphdr *iph = ip_hdr(skb);
+       struct vif_device *vif = &net->ipv4.vif_table[vifi];
        struct net_device *dev;
        struct rtable *rt;
        int    encap = 0;
@@ -1148,12 +1232,11 @@ static void ipmr_queue_xmit(struct sk_buff *skb, struct mfc_cache *c, int vifi)
 #ifdef CONFIG_IP_PIMSM
        if (vif->flags & VIFF_REGISTER) {
                vif->pkt_out++;
-               vif->bytes_out+=skb->len;
-               ((struct net_device_stats*)netdev_priv(vif->dev))->tx_bytes += skb->len;
-               ((struct net_device_stats*)netdev_priv(vif->dev))->tx_packets++;
-               ipmr_cache_report(skb, vifi, IGMPMSG_WHOLEPKT);
-               kfree_skb(skb);
-               return;
+               vif->bytes_out += skb->len;
+               vif->dev->stats.tx_bytes += skb->len;
+               vif->dev->stats.tx_packets++;
+               ipmr_cache_report(net, skb, vifi, IGMPMSG_WHOLEPKT);
+               goto out_free;
        }
 #endif
 
@@ -1164,7 +1247,7 @@ static void ipmr_queue_xmit(struct sk_buff *skb, struct mfc_cache *c, int vifi)
                                                .saddr = vif->local,
                                                .tos = RT_TOS(iph->tos) } },
                                    .proto = IPPROTO_IPIP };
-               if (ip_route_output_key(&rt, &fl))
+               if (ip_route_output_key(net, &rt, &fl))
                        goto out_free;
                encap = sizeof(struct iphdr);
        } else {
@@ -1173,7 +1256,7 @@ static void ipmr_queue_xmit(struct sk_buff *skb, struct mfc_cache *c, int vifi)
                                              { .daddr = iph->daddr,
                                                .tos = RT_TOS(iph->tos) } },
                                    .proto = IPPROTO_IPIP };
-               if (ip_route_output_key(&rt, &fl))
+               if (ip_route_output_key(net, &rt, &fl))
                        goto out_free;
        }
 
@@ -1185,7 +1268,7 @@ static void ipmr_queue_xmit(struct sk_buff *skb, struct mfc_cache *c, int vifi)
                   to blackhole.
                 */
 
-               IP_INC_STATS_BH(IPSTATS_MIB_FRAGFAILS);
+               IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_FRAGFAILS);
                ip_rt_put(rt);
                goto out_free;
        }
@@ -1193,25 +1276,24 @@ static void ipmr_queue_xmit(struct sk_buff *skb, struct mfc_cache *c, int vifi)
        encap += LL_RESERVED_SPACE(dev) + rt->u.dst.header_len;
 
        if (skb_cow(skb, encap)) {
-               ip_rt_put(rt);
+               ip_rt_put(rt);
                goto out_free;
        }
 
        vif->pkt_out++;
-       vif->bytes_out+=skb->len;
+       vif->bytes_out += skb->len;
 
-       dst_release(skb->dst);
-       skb->dst = &rt->u.dst;
-       iph = skb->nh.iph;
-       ip_decrease_ttl(iph);
+       skb_dst_drop(skb);
+       skb_dst_set(skb, &rt->u.dst);
+       ip_decrease_ttl(ip_hdr(skb));
 
        /* FIXME: forward and output firewalls used to be called here.
         * What do we do with netfilter? -- RR */
        if (vif->flags & VIFF_TUNNEL) {
                ip_encap(skb, vif->local, vif->remote);
                /* FIXME: extra output firewall step used to be here. --RR */
-               ((struct ip_tunnel *)netdev_priv(vif->dev))->stat.tx_packets++;
-               ((struct ip_tunnel *)netdev_priv(vif->dev))->stat.tx_bytes+=skb->len;
+               vif->dev->stats.tx_packets++;
+               vif->dev->stats.tx_bytes += skb->len;
        }
 
        IPCB(skb)->flags |= IPSKB_FORWARDED;
@@ -1227,7 +1309,7 @@ static void ipmr_queue_xmit(struct sk_buff *skb, struct mfc_cache *c, int vifi)
         * not mrouter) cannot join to more than one interface - it will
         * result in receiving multiple packets.
         */
-       NF_HOOK(PF_INET, NF_IP_FORWARD, skb, skb->dev, dev, 
+       NF_HOOK(PF_INET, NF_INET_FORWARD, skb, skb->dev, dev,
                ipmr_forward_finish);
        return;
 
@@ -1238,9 +1320,10 @@ out_free:
 
 static int ipmr_find_vif(struct net_device *dev)
 {
+       struct net *net = dev_net(dev);
        int ct;
-       for (ct=maxvif-1; ct>=0; ct--) {
-               if (vif_table[ct].dev == dev)
+       for (ct = net->ipv4.maxvif-1; ct >= 0; ct--) {
+               if (net->ipv4.vif_table[ct].dev == dev)
                        break;
        }
        return ct;
@@ -1252,6 +1335,7 @@ static int ip_mr_forward(struct sk_buff *skb, struct mfc_cache *cache, int local
 {
        int psend = -1;
        int vif, ct;
+       struct net *net = mfc_net(cache);
 
        vif = cache->mfc_parent;
        cache->mfc_un.res.pkt++;
@@ -1260,10 +1344,10 @@ static int ip_mr_forward(struct sk_buff *skb, struct mfc_cache *cache, int local
        /*
         * Wrong interface: drop packet and (maybe) send PIM assert.
         */
-       if (vif_table[vif].dev != skb->dev) {
+       if (net->ipv4.vif_table[vif].dev != skb->dev) {
                int true_vifi;
 
-               if (((struct rtable*)skb->dst)->fl.iif == 0) {
+               if (skb_rtable(skb)->fl.iif == 0) {
                        /* It is our own packet, looped back.
                           Very complicated situation...
 
@@ -1281,35 +1365,36 @@ static int ip_mr_forward(struct sk_buff *skb, struct mfc_cache *cache, int local
                cache->mfc_un.res.wrong_if++;
                true_vifi = ipmr_find_vif(skb->dev);
 
-               if (true_vifi >= 0 && mroute_do_assert &&
+               if (true_vifi >= 0 && net->ipv4.mroute_do_assert &&
                    /* pimsm uses asserts, when switching from RPT to SPT,
                       so that we cannot check that packet arrived on an oif.
                       It is bad, but otherwise we would need to move pretty
                       large chunk of pimd to kernel. Ough... --ANK
                     */
-                   (mroute_do_pim || cache->mfc_un.res.ttls[true_vifi] < 255) &&
-                   time_after(jiffies, 
+                   (net->ipv4.mroute_do_pim ||
+                    cache->mfc_un.res.ttls[true_vifi] < 255) &&
+                   time_after(jiffies,
                               cache->mfc_un.res.last_assert + MFC_ASSERT_THRESH)) {
                        cache->mfc_un.res.last_assert = jiffies;
-                       ipmr_cache_report(skb, true_vifi, IGMPMSG_WRONGVIF);
+                       ipmr_cache_report(net, skb, true_vifi, IGMPMSG_WRONGVIF);
                }
                goto dont_forward;
        }
 
-       vif_table[vif].pkt_in++;
-       vif_table[vif].bytes_in+=skb->len;
+       net->ipv4.vif_table[vif].pkt_in++;
+       net->ipv4.vif_table[vif].bytes_in += skb->len;
 
        /*
         *      Forward the frame
         */
        for (ct = cache->mfc_un.res.maxvif-1; ct >= cache->mfc_un.res.minvif; ct--) {
-               if (skb->nh.iph->ttl > cache->mfc_un.res.ttls[ct]) {
+               if (ip_hdr(skb)->ttl > cache->mfc_un.res.ttls[ct]) {
                        if (psend != -1) {
                                struct sk_buff *skb2 = skb_clone(skb, GFP_ATOMIC);
                                if (skb2)
                                        ipmr_queue_xmit(skb2, cache, psend);
                        }
-                       psend=ct;
+                       psend = ct;
                }
        }
        if (psend != -1) {
@@ -1337,7 +1422,8 @@ dont_forward:
 int ip_mr_input(struct sk_buff *skb)
 {
        struct mfc_cache *cache;
-       int local = ((struct rtable*)skb->dst)->rt_flags&RTCF_LOCAL;
+       struct net *net = dev_net(skb->dev);
+       int local = skb_rtable(skb)->rt_flags & RTCF_LOCAL;
 
        /* Packet is looped back after forward, it should not be
           forwarded second time, but still can be delivered locally.
@@ -1349,7 +1435,7 @@ int ip_mr_input(struct sk_buff *skb)
                    if (IPCB(skb)->opt.router_alert) {
                            if (ip_call_ra_chain(skb))
                                    return 0;
-                   } else if (skb->nh.iph->protocol == IPPROTO_IGMP){
+                   } else if (ip_hdr(skb)->protocol == IPPROTO_IGMP){
                            /* IGMPv1 (and broken IGMPv2 implementations sort of
                               Cisco IOS <= 11.2(8)) do not put router alert
                               option to IGMP packets destined to routable
@@ -1357,9 +1443,9 @@ int ip_mr_input(struct sk_buff *skb)
                               that we can forward NO IGMP messages.
                             */
                            read_lock(&mrt_lock);
-                           if (mroute_socket) {
+                           if (net->ipv4.mroute_sk) {
                                    nf_reset(skb);
-                                   raw_rcv(mroute_socket, skb);
+                                   raw_rcv(net->ipv4.mroute_sk, skb);
                                    read_unlock(&mrt_lock);
                                    return 0;
                            }
@@ -1368,12 +1454,12 @@ int ip_mr_input(struct sk_buff *skb)
        }
 
        read_lock(&mrt_lock);
-       cache = ipmr_cache_find(skb->nh.iph->saddr, skb->nh.iph->daddr);
+       cache = ipmr_cache_find(net, ip_hdr(skb)->saddr, ip_hdr(skb)->daddr);
 
        /*
         *      No usable cache entry
         */
-       if (cache==NULL) {
+       if (cache == NULL) {
                int vif;
 
                if (local) {
@@ -1388,7 +1474,7 @@ int ip_mr_input(struct sk_buff *skb)
 
                vif = ipmr_find_vif(skb->dev);
                if (vif >= 0) {
-                       int err = ipmr_cache_unresolved(vif, skb);
+                       int err = ipmr_cache_unresolved(net, vif, skb);
                        read_unlock(&mrt_lock);
 
                        return err;
@@ -1414,66 +1500,76 @@ dont_forward:
        return 0;
 }
 
-#ifdef CONFIG_IP_PIMSM_V1
-/*
- * Handle IGMP messages of PIMv1
- */
-
-int pim_rcv_v1(struct sk_buff * skb)
+#ifdef CONFIG_IP_PIMSM
+static int __pim_rcv(struct sk_buff *skb, unsigned int pimlen)
 {
-       struct igmphdr *pim;
-       struct iphdr   *encap;
-       struct net_device  *reg_dev = NULL;
-
-       if (!pskb_may_pull(skb, sizeof(*pim) + sizeof(*encap))) 
-               goto drop;
+       struct net_device *reg_dev = NULL;
+       struct iphdr *encap;
+       struct net *net = dev_net(skb->dev);
 
-       pim = (struct igmphdr*)skb->h.raw;
-
-        if (!mroute_do_pim ||
-           skb->len < sizeof(*pim) + sizeof(*encap) ||
-           pim->group != PIM_V1_VERSION || pim->code != PIM_V1_REGISTER) 
-               goto drop;
-
-       encap = (struct iphdr*)(skb->h.raw + sizeof(struct igmphdr));
+       encap = (struct iphdr *)(skb_transport_header(skb) + pimlen);
        /*
           Check that:
           a. packet is really destinted to a multicast group
           b. packet is not a NULL-REGISTER
           c. packet is not truncated
         */
-       if (!MULTICAST(encap->daddr) ||
+       if (!ipv4_is_multicast(encap->daddr) ||
            encap->tot_len == 0 ||
-           ntohs(encap->tot_len) + sizeof(*pim) > skb->len) 
-               goto drop;
+           ntohs(encap->tot_len) + pimlen > skb->len)
+               return 1;
 
        read_lock(&mrt_lock);
-       if (reg_vif_num >= 0)
-               reg_dev = vif_table[reg_vif_num].dev;
+       if (net->ipv4.mroute_reg_vif_num >= 0)
+               reg_dev = net->ipv4.vif_table[net->ipv4.mroute_reg_vif_num].dev;
        if (reg_dev)
                dev_hold(reg_dev);
        read_unlock(&mrt_lock);
 
-       if (reg_dev == NULL) 
-               goto drop;
+       if (reg_dev == NULL)
+               return 1;
 
-       skb->mac.raw = skb->nh.raw;
+       skb->mac_header = skb->network_header;
        skb_pull(skb, (u8*)encap - skb->data);
-       skb->nh.iph = (struct iphdr *)skb->data;
+       skb_reset_network_header(skb);
        skb->dev = reg_dev;
        skb->protocol = htons(ETH_P_IP);
        skb->ip_summed = 0;
        skb->pkt_type = PACKET_HOST;
-       dst_release(skb->dst);
-       skb->dst = NULL;
-       ((struct net_device_stats*)netdev_priv(reg_dev))->rx_bytes += skb->len;
-       ((struct net_device_stats*)netdev_priv(reg_dev))->rx_packets++;
+       skb_dst_drop(skb);
+       reg_dev->stats.rx_bytes += skb->len;
+       reg_dev->stats.rx_packets++;
        nf_reset(skb);
        netif_rx(skb);
        dev_put(reg_dev);
+
        return 0;
- drop:
-       kfree_skb(skb);
+}
+#endif
+
+#ifdef CONFIG_IP_PIMSM_V1
+/*
+ * Handle IGMP messages of PIMv1
+ */
+
+int pim_rcv_v1(struct sk_buff * skb)
+{
+       struct igmphdr *pim;
+       struct net *net = dev_net(skb->dev);
+
+       if (!pskb_may_pull(skb, sizeof(*pim) + sizeof(struct iphdr)))
+               goto drop;
+
+       pim = igmp_hdr(skb);
+
+       if (!net->ipv4.mroute_do_pim ||
+           pim->group != PIM_V1_VERSION || pim->code != PIM_V1_REGISTER)
+               goto drop;
+
+       if (__pim_rcv(skb, sizeof(*pim))) {
+drop:
+               kfree_skb(skb);
+       }
        return 0;
 }
 #endif
@@ -1482,53 +1578,21 @@ int pim_rcv_v1(struct sk_buff * skb)
 static int pim_rcv(struct sk_buff * skb)
 {
        struct pimreghdr *pim;
-       struct iphdr   *encap;
-       struct net_device  *reg_dev = NULL;
 
-       if (!pskb_may_pull(skb, sizeof(*pim) + sizeof(*encap))) 
+       if (!pskb_may_pull(skb, sizeof(*pim) + sizeof(struct iphdr)))
                goto drop;
 
-       pim = (struct pimreghdr*)skb->h.raw;
-        if (pim->type != ((PIM_VERSION<<4)|(PIM_REGISTER)) ||
+       pim = (struct pimreghdr *)skb_transport_header(skb);
+       if (pim->type != ((PIM_VERSION<<4)|(PIM_REGISTER)) ||
            (pim->flags&PIM_NULL_REGISTER) ||
-           (ip_compute_csum((void *)pim, sizeof(*pim)) != 0 && 
-            (u16)csum_fold(skb_checksum(skb, 0, skb->len, 0)))) 
+           (ip_compute_csum((void *)pim, sizeof(*pim)) != 0 &&
+            csum_fold(skb_checksum(skb, 0, skb->len, 0))))
                goto drop;
 
-       /* check if the inner packet is destined to mcast group */
-       encap = (struct iphdr*)(skb->h.raw + sizeof(struct pimreghdr));
-       if (!MULTICAST(encap->daddr) ||
-           encap->tot_len == 0 ||
-           ntohs(encap->tot_len) + sizeof(*pim) > skb->len) 
-               goto drop;
-
-       read_lock(&mrt_lock);
-       if (reg_vif_num >= 0)
-               reg_dev = vif_table[reg_vif_num].dev;
-       if (reg_dev)
-               dev_hold(reg_dev);
-       read_unlock(&mrt_lock);
-
-       if (reg_dev == NULL) 
-               goto drop;
-
-       skb->mac.raw = skb->nh.raw;
-       skb_pull(skb, (u8*)encap - skb->data);
-       skb->nh.iph = (struct iphdr *)skb->data;
-       skb->dev = reg_dev;
-       skb->protocol = htons(ETH_P_IP);
-       skb->ip_summed = 0;
-       skb->pkt_type = PACKET_HOST;
-       dst_release(skb->dst);
-       ((struct net_device_stats*)netdev_priv(reg_dev))->rx_bytes += skb->len;
-       ((struct net_device_stats*)netdev_priv(reg_dev))->rx_packets++;
-       skb->dst = NULL;
-       nf_reset(skb);
-       netif_rx(skb);
-       dev_put(reg_dev);
-       return 0;
- drop:
-       kfree_skb(skb);
+       if (__pim_rcv(skb, sizeof(*pim))) {
+drop:
+               kfree_skb(skb);
+       }
        return 0;
 }
 #endif
@@ -1538,46 +1602,50 @@ ipmr_fill_mroute(struct sk_buff *skb, struct mfc_cache *c, struct rtmsg *rtm)
 {
        int ct;
        struct rtnexthop *nhp;
-       struct net_device *dev = vif_table[c->mfc_parent].dev;
-       u8 *b = skb->tail;
+       struct net *net = mfc_net(c);
+       struct net_device *dev = net->ipv4.vif_table[c->mfc_parent].dev;
+       u8 *b = skb_tail_pointer(skb);
        struct rtattr *mp_head;
 
        if (dev)
                RTA_PUT(skb, RTA_IIF, 4, &dev->ifindex);
 
-       mp_head = (struct rtattr*)skb_put(skb, RTA_LENGTH(0));
+       mp_head = (struct rtattr *)skb_put(skb, RTA_LENGTH(0));
 
        for (ct = c->mfc_un.res.minvif; ct < c->mfc_un.res.maxvif; ct++) {
                if (c->mfc_un.res.ttls[ct] < 255) {
                        if (skb_tailroom(skb) < RTA_ALIGN(RTA_ALIGN(sizeof(*nhp)) + 4))
                                goto rtattr_failure;
-                       nhp = (struct rtnexthop*)skb_put(skb, RTA_ALIGN(sizeof(*nhp)));
+                       nhp = (struct rtnexthop *)skb_put(skb, RTA_ALIGN(sizeof(*nhp)));
                        nhp->rtnh_flags = 0;
                        nhp->rtnh_hops = c->mfc_un.res.ttls[ct];
-                       nhp->rtnh_ifindex = vif_table[ct].dev->ifindex;
+                       nhp->rtnh_ifindex = net->ipv4.vif_table[ct].dev->ifindex;
                        nhp->rtnh_len = sizeof(*nhp);
                }
        }
        mp_head->rta_type = RTA_MULTIPATH;
-       mp_head->rta_len = skb->tail - (u8*)mp_head;
+       mp_head->rta_len = skb_tail_pointer(skb) - (u8 *)mp_head;
        rtm->rtm_type = RTN_MULTICAST;
        return 1;
 
 rtattr_failure:
-       skb_trim(skb, b - skb->data);
+       nlmsg_trim(skb, b);
        return -EMSGSIZE;
 }
 
-int ipmr_get_route(struct sk_buff *skb, struct rtmsg *rtm, int nowait)
+int ipmr_get_route(struct net *net,
+                  struct sk_buff *skb, struct rtmsg *rtm, int nowait)
 {
        int err;
        struct mfc_cache *cache;
-       struct rtable *rt = (struct rtable*)skb->dst;
+       struct rtable *rt = skb_rtable(skb);
 
        read_lock(&mrt_lock);
-       cache = ipmr_cache_find(rt->rt_src, rt->rt_dst);
+       cache = ipmr_cache_find(net, rt->rt_src, rt->rt_dst);
 
-       if (cache==NULL) {
+       if (cache == NULL) {
+               struct sk_buff *skb2;
+               struct iphdr *iph;
                struct net_device *dev;
                int vif;
 
@@ -1591,12 +1659,20 @@ int ipmr_get_route(struct sk_buff *skb, struct rtmsg *rtm, int nowait)
                        read_unlock(&mrt_lock);
                        return -ENODEV;
                }
-               skb->nh.raw = skb_push(skb, sizeof(struct iphdr));
-               skb->nh.iph->ihl = sizeof(struct iphdr)>>2;
-               skb->nh.iph->saddr = rt->rt_src;
-               skb->nh.iph->daddr = rt->rt_dst;
-               skb->nh.iph->version = 0;
-               err = ipmr_cache_unresolved(vif, skb);
+               skb2 = skb_clone(skb, GFP_ATOMIC);
+               if (!skb2) {
+                       read_unlock(&mrt_lock);
+                       return -ENOMEM;
+               }
+
+               skb_push(skb2, sizeof(struct iphdr));
+               skb_reset_network_header(skb2);
+               iph = ip_hdr(skb2);
+               iph->ihl = sizeof(struct iphdr) >> 2;
+               iph->saddr = rt->rt_src;
+               iph->daddr = rt->rt_dst;
+               iph->version = 0;
+               err = ipmr_cache_unresolved(net, vif, skb2);
                read_unlock(&mrt_lock);
                return err;
        }
@@ -1608,58 +1684,67 @@ int ipmr_get_route(struct sk_buff *skb, struct rtmsg *rtm, int nowait)
        return err;
 }
 
-#ifdef CONFIG_PROC_FS  
+#ifdef CONFIG_PROC_FS
 /*
  *     The /proc interfaces to multicast routing /proc/ip_mr_cache /proc/ip_mr_vif
  */
 struct ipmr_vif_iter {
+       struct seq_net_private p;
        int ct;
 };
 
-static struct vif_device *ipmr_vif_seq_idx(struct ipmr_vif_iter *iter,
+static struct vif_device *ipmr_vif_seq_idx(struct net *net,
+                                          struct ipmr_vif_iter *iter,
                                           loff_t pos)
 {
-       for (iter->ct = 0; iter->ct < maxvif; ++iter->ct) {
-               if(!VIF_EXISTS(iter->ct))
+       for (iter->ct = 0; iter->ct < net->ipv4.maxvif; ++iter->ct) {
+               if (!VIF_EXISTS(net, iter->ct))
                        continue;
-               if (pos-- == 0) 
-                       return &vif_table[iter->ct];
+               if (pos-- == 0)
+                       return &net->ipv4.vif_table[iter->ct];
        }
        return NULL;
 }
 
 static void *ipmr_vif_seq_start(struct seq_file *seq, loff_t *pos)
+       __acquires(mrt_lock)
 {
+       struct net *net = seq_file_net(seq);
+
        read_lock(&mrt_lock);
-       return *pos ? ipmr_vif_seq_idx(seq->private, *pos - 1) 
+       return *pos ? ipmr_vif_seq_idx(net, seq->private, *pos - 1)
                : SEQ_START_TOKEN;
 }
 
 static void *ipmr_vif_seq_next(struct seq_file *seq, void *v, loff_t *pos)
 {
        struct ipmr_vif_iter *iter = seq->private;
+       struct net *net = seq_file_net(seq);
 
        ++*pos;
        if (v == SEQ_START_TOKEN)
-               return ipmr_vif_seq_idx(iter, 0);
-       
-       while (++iter->ct < maxvif) {
-               if(!VIF_EXISTS(iter->ct))
+               return ipmr_vif_seq_idx(net, iter, 0);
+
+       while (++iter->ct < net->ipv4.maxvif) {
+               if (!VIF_EXISTS(net, iter->ct))
                        continue;
-               return &vif_table[iter->ct];
+               return &net->ipv4.vif_table[iter->ct];
        }
        return NULL;
 }
 
 static void ipmr_vif_seq_stop(struct seq_file *seq, void *v)
+       __releases(mrt_lock)
 {
        read_unlock(&mrt_lock);
 }
 
 static int ipmr_vif_seq_show(struct seq_file *seq, void *v)
 {
+       struct net *net = seq_file_net(seq);
+
        if (v == SEQ_START_TOKEN) {
-               seq_puts(seq, 
+               seq_puts(seq,
                         "Interface      BytesIn  PktsIn  BytesOut PktsOut Flags Local    Remote\n");
        } else {
                const struct vif_device *vif = v;
@@ -1667,15 +1752,15 @@ static int ipmr_vif_seq_show(struct seq_file *seq, void *v)
 
                seq_printf(seq,
                           "%2Zd %-10s %8ld %7ld  %8ld %7ld %05X %08X %08X\n",
-                          vif - vif_table,
-                          name, vif->bytes_in, vif->pkt_in, 
+                          vif - net->ipv4.vif_table,
+                          name, vif->bytes_in, vif->pkt_in,
                           vif->bytes_out, vif->pkt_out,
                           vif->flags, vif->local, vif->remote);
        }
        return 0;
 }
 
-static struct seq_operations ipmr_vif_seq_ops = {
+static const struct seq_operations ipmr_vif_seq_ops = {
        .start = ipmr_vif_seq_start,
        .next  = ipmr_vif_seq_next,
        .stop  = ipmr_vif_seq_stop,
@@ -1684,58 +1769,44 @@ static struct seq_operations ipmr_vif_seq_ops = {
 
 static int ipmr_vif_open(struct inode *inode, struct file *file)
 {
-       struct seq_file *seq;
-       int rc = -ENOMEM;
-       struct ipmr_vif_iter *s = kmalloc(sizeof(*s), GFP_KERNEL);
-       
-       if (!s)
-               goto out;
-
-       rc = seq_open(file, &ipmr_vif_seq_ops);
-       if (rc)
-               goto out_kfree;
-
-       s->ct = 0;
-       seq = file->private_data;
-       seq->private = s;
-out:
-       return rc;
-out_kfree:
-       kfree(s);
-       goto out;
-
+       return seq_open_net(inode, file, &ipmr_vif_seq_ops,
+                           sizeof(struct ipmr_vif_iter));
 }
 
-static struct file_operations ipmr_vif_fops = {
+static const struct file_operations ipmr_vif_fops = {
        .owner   = THIS_MODULE,
        .open    = ipmr_vif_open,
        .read    = seq_read,
        .llseek  = seq_lseek,
-       .release = seq_release_private,
+       .release = seq_release_net,
 };
 
 struct ipmr_mfc_iter {
+       struct seq_net_private p;
        struct mfc_cache **cache;
        int ct;
 };
 
 
-static struct mfc_cache *ipmr_mfc_seq_idx(struct ipmr_mfc_iter *it, loff_t pos)
+static struct mfc_cache *ipmr_mfc_seq_idx(struct net *net,
+                                         struct ipmr_mfc_iter *it, loff_t pos)
 {
        struct mfc_cache *mfc;
 
-       it->cache = mfc_cache_array;
+       it->cache = net->ipv4.mfc_cache_array;
        read_lock(&mrt_lock);
-       for (it->ct = 0; it->ct < MFC_LINES; it->ct++) 
-               for(mfc = mfc_cache_array[it->ct]; mfc; mfc = mfc->next) 
-                       if (pos-- == 0) 
+       for (it->ct = 0; it->ct < MFC_LINES; it->ct++)
+               for (mfc = net->ipv4.mfc_cache_array[it->ct];
+                    mfc; mfc = mfc->next)
+                       if (pos-- == 0)
                                return mfc;
        read_unlock(&mrt_lock);
 
        it->cache = &mfc_unres_queue;
        spin_lock_bh(&mfc_unres_lock);
-       for(mfc = mfc_unres_queue; mfc; mfc = mfc->next) 
-               if (pos-- == 0)
+       for (mfc = mfc_unres_queue; mfc; mfc = mfc->next)
+               if (net_eq(mfc_net(mfc), net) &&
+                   pos-- == 0)
                        return mfc;
        spin_unlock_bh(&mfc_unres_lock);
 
@@ -1747,9 +1818,11 @@ static struct mfc_cache *ipmr_mfc_seq_idx(struct ipmr_mfc_iter *it, loff_t pos)
 static void *ipmr_mfc_seq_start(struct seq_file *seq, loff_t *pos)
 {
        struct ipmr_mfc_iter *it = seq->private;
+       struct net *net = seq_file_net(seq);
+
        it->cache = NULL;
        it->ct = 0;
-       return *pos ? ipmr_mfc_seq_idx(seq->private, *pos - 1) 
+       return *pos ? ipmr_mfc_seq_idx(net, seq->private, *pos - 1)
                : SEQ_START_TOKEN;
 }
 
@@ -1757,22 +1830,23 @@ static void *ipmr_mfc_seq_next(struct seq_file *seq, void *v, loff_t *pos)
 {
        struct mfc_cache *mfc = v;
        struct ipmr_mfc_iter *it = seq->private;
+       struct net *net = seq_file_net(seq);
 
        ++*pos;
 
        if (v == SEQ_START_TOKEN)
-               return ipmr_mfc_seq_idx(seq->private, 0);
+               return ipmr_mfc_seq_idx(net, seq->private, 0);
 
        if (mfc->next)
                return mfc->next;
-       
-       if (it->cache == &mfc_unres_queue) 
+
+       if (it->cache == &mfc_unres_queue)
                goto end_of_list;
 
-       BUG_ON(it->cache != mfc_cache_array);
+       BUG_ON(it->cache != net->ipv4.mfc_cache_array);
 
        while (++it->ct < MFC_LINES) {
-               mfc = mfc_cache_array[it->ct];
+               mfc = net->ipv4.mfc_cache_array[it->ct];
                if (mfc)
                        return mfc;
        }
@@ -1781,10 +1855,12 @@ static void *ipmr_mfc_seq_next(struct seq_file *seq, void *v, loff_t *pos)
        read_unlock(&mrt_lock);
        it->cache = &mfc_unres_queue;
        it->ct = 0;
-               
+
        spin_lock_bh(&mfc_unres_lock);
        mfc = mfc_unres_queue;
-       if (mfc) 
+       while (mfc && !net_eq(mfc_net(mfc), net))
+               mfc = mfc->next;
+       if (mfc)
                return mfc;
 
  end_of_list:
@@ -1797,48 +1873,56 @@ static void *ipmr_mfc_seq_next(struct seq_file *seq, void *v, loff_t *pos)
 static void ipmr_mfc_seq_stop(struct seq_file *seq, void *v)
 {
        struct ipmr_mfc_iter *it = seq->private;
+       struct net *net = seq_file_net(seq);
 
        if (it->cache == &mfc_unres_queue)
                spin_unlock_bh(&mfc_unres_lock);
-       else if (it->cache == mfc_cache_array)
+       else if (it->cache == net->ipv4.mfc_cache_array)
                read_unlock(&mrt_lock);
 }
 
 static int ipmr_mfc_seq_show(struct seq_file *seq, void *v)
 {
        int n;
+       struct net *net = seq_file_net(seq);
 
        if (v == SEQ_START_TOKEN) {
-               seq_puts(seq, 
+               seq_puts(seq,
                 "Group    Origin   Iif     Pkts    Bytes    Wrong Oifs\n");
        } else {
                const struct mfc_cache *mfc = v;
                const struct ipmr_mfc_iter *it = seq->private;
-               
-               seq_printf(seq, "%08lX %08lX %-3d %8ld %8ld %8ld",
+
+               seq_printf(seq, "%08lX %08lX %-3hd",
                           (unsigned long) mfc->mfc_mcastgrp,
                           (unsigned long) mfc->mfc_origin,
-                          mfc->mfc_parent,
-                          mfc->mfc_un.res.pkt,
-                          mfc->mfc_un.res.bytes,
-                          mfc->mfc_un.res.wrong_if);
+                          mfc->mfc_parent);
 
                if (it->cache != &mfc_unres_queue) {
-                       for(n = mfc->mfc_un.res.minvif; 
-                           n < mfc->mfc_un.res.maxvif; n++ ) {
-                               if(VIF_EXISTS(n) 
-                                  && mfc->mfc_un.res.ttls[n] < 255)
-                               seq_printf(seq, 
-                                          " %2d:%-3d", 
+                       seq_printf(seq, " %8lu %8lu %8lu",
+                                  mfc->mfc_un.res.pkt,
+                                  mfc->mfc_un.res.bytes,
+                                  mfc->mfc_un.res.wrong_if);
+                       for (n = mfc->mfc_un.res.minvif;
+                            n < mfc->mfc_un.res.maxvif; n++ ) {
+                               if (VIF_EXISTS(net, n) &&
+                                   mfc->mfc_un.res.ttls[n] < 255)
+                                       seq_printf(seq,
+                                          " %2d:%-3d",
                                           n, mfc->mfc_un.res.ttls[n]);
                        }
+               } else {
+                       /* unresolved mfc_caches don't contain
+                        * pkt, bytes and wrong_if values
+                        */
+                       seq_printf(seq, " %8lu %8lu %8lu", 0ul, 0ul, 0ul);
                }
                seq_putc(seq, '\n');
        }
        return 0;
 }
 
-static struct seq_operations ipmr_mfc_seq_ops = {
+static const struct seq_operations ipmr_mfc_seq_ops = {
        .start = ipmr_mfc_seq_start,
        .next  = ipmr_mfc_seq_next,
        .stop  = ipmr_mfc_seq_stop,
@@ -1847,39 +1931,23 @@ static struct seq_operations ipmr_mfc_seq_ops = {
 
 static int ipmr_mfc_open(struct inode *inode, struct file *file)
 {
-       struct seq_file *seq;
-       int rc = -ENOMEM;
-       struct ipmr_mfc_iter *s = kmalloc(sizeof(*s), GFP_KERNEL);
-       
-       if (!s)
-               goto out;
-
-       rc = seq_open(file, &ipmr_mfc_seq_ops);
-       if (rc)
-               goto out_kfree;
-
-       seq = file->private_data;
-       seq->private = s;
-out:
-       return rc;
-out_kfree:
-       kfree(s);
-       goto out;
-
+       return seq_open_net(inode, file, &ipmr_mfc_seq_ops,
+                           sizeof(struct ipmr_mfc_iter));
 }
 
-static struct file_operations ipmr_mfc_fops = {
+static const struct file_operations ipmr_mfc_fops = {
        .owner   = THIS_MODULE,
        .open    = ipmr_mfc_open,
        .read    = seq_read,
        .llseek  = seq_lseek,
-       .release = seq_release_private,
+       .release = seq_release_net,
 };
-#endif 
+#endif
 
 #ifdef CONFIG_IP_PIMSM_V2
 static struct net_protocol pim_protocol = {
        .handler        =       pim_rcv,
+       .netns_ok       =       1,
 };
 #endif
 
@@ -1887,21 +1955,102 @@ static struct net_protocol pim_protocol = {
 /*
  *     Setup for IP multicast routing
  */
-void __init ip_mr_init(void)
+static int __net_init ipmr_net_init(struct net *net)
 {
+       int err = 0;
+
+       net->ipv4.vif_table = kcalloc(MAXVIFS, sizeof(struct vif_device),
+                                     GFP_KERNEL);
+       if (!net->ipv4.vif_table) {
+               err = -ENOMEM;
+               goto fail;
+       }
+
+       /* Forwarding cache */
+       net->ipv4.mfc_cache_array = kcalloc(MFC_LINES,
+                                           sizeof(struct mfc_cache *),
+                                           GFP_KERNEL);
+       if (!net->ipv4.mfc_cache_array) {
+               err = -ENOMEM;
+               goto fail_mfc_cache;
+       }
+
+#ifdef CONFIG_IP_PIMSM
+       net->ipv4.mroute_reg_vif_num = -1;
+#endif
+
+#ifdef CONFIG_PROC_FS
+       err = -ENOMEM;
+       if (!proc_net_fops_create(net, "ip_mr_vif", 0, &ipmr_vif_fops))
+               goto proc_vif_fail;
+       if (!proc_net_fops_create(net, "ip_mr_cache", 0, &ipmr_mfc_fops))
+               goto proc_cache_fail;
+#endif
+       return 0;
+
+#ifdef CONFIG_PROC_FS
+proc_cache_fail:
+       proc_net_remove(net, "ip_mr_vif");
+proc_vif_fail:
+       kfree(net->ipv4.mfc_cache_array);
+#endif
+fail_mfc_cache:
+       kfree(net->ipv4.vif_table);
+fail:
+       return err;
+}
+
+static void __net_exit ipmr_net_exit(struct net *net)
+{
+#ifdef CONFIG_PROC_FS
+       proc_net_remove(net, "ip_mr_cache");
+       proc_net_remove(net, "ip_mr_vif");
+#endif
+       kfree(net->ipv4.mfc_cache_array);
+       kfree(net->ipv4.vif_table);
+}
+
+static struct pernet_operations ipmr_net_ops = {
+       .init = ipmr_net_init,
+       .exit = ipmr_net_exit,
+};
+
+int __init ip_mr_init(void)
+{
+       int err;
+
        mrt_cachep = kmem_cache_create("ip_mrt_cache",
                                       sizeof(struct mfc_cache),
-                                      0, SLAB_HWCACHE_ALIGN,
-                                      NULL, NULL);
+                                      0, SLAB_HWCACHE_ALIGN|SLAB_PANIC,
+                                      NULL);
        if (!mrt_cachep)
-               panic("cannot allocate ip_mrt_cache");
-
-       init_timer(&ipmr_expire_timer);
-       ipmr_expire_timer.function=ipmr_expire_process;
-       register_netdevice_notifier(&ip_mr_notifier);
-#ifdef CONFIG_PROC_FS  
-       proc_net_fops_create("ip_mr_vif", 0, &ipmr_vif_fops);
-       proc_net_fops_create("ip_mr_cache", 0, &ipmr_mfc_fops);
-#endif 
+               return -ENOMEM;
+
+       err = register_pernet_subsys(&ipmr_net_ops);
+       if (err)
+               goto reg_pernet_fail;
+
+       setup_timer(&ipmr_expire_timer, ipmr_expire_process, 0);
+       err = register_netdevice_notifier(&ip_mr_notifier);
+       if (err)
+               goto reg_notif_fail;
+#ifdef CONFIG_IP_PIMSM_V2
+       if (inet_add_protocol(&pim_protocol, IPPROTO_PIM) < 0) {
+               printk(KERN_ERR "ip_mr_init: can't add PIM protocol\n");
+               err = -EAGAIN;
+               goto add_proto_fail;
+       }
+#endif
+       return 0;
+
+#ifdef CONFIG_IP_PIMSM_V2
+add_proto_fail:
+       unregister_netdevice_notifier(&ip_mr_notifier);
+#endif
+reg_notif_fail:
+       del_timer(&ipmr_expire_timer);
+       unregister_pernet_subsys(&ipmr_net_ops);
+reg_pernet_fail:
+       kmem_cache_destroy(mrt_cachep);
+       return err;
 }