Merge branch 'master'

183 files changed, 6451 insertions, 3640 deletions

diff --git a/net/802/p8022.c b/net/802/p8022.c
index b24817c63ca..2530f35241c 100644
--- a/net/802/p8022.c
+++ b/net/802/p8022.c
@@ -56,7 +56,7 @@ struct datalink_proto *register_8022_client(unsigned char type,
 
 void unregister_8022_client(struct datalink_proto *proto)
 {
-	llc_sap_close(proto->sap);
+	llc_sap_put(proto->sap);
 	kfree(proto);
 }
 
diff --git a/net/802/psnap.c b/net/802/psnap.c
index ab80b1fab53..4d638944d93 100644
--- a/net/802/psnap.c
+++ b/net/802/psnap.c
@@ -106,7 +106,7 @@ module_init(snap_init);
 
 static void __exit snap_exit(void)
 {
-	llc_sap_close(snap_sap);
+	llc_sap_put(snap_sap);
 }
 
 module_exit(snap_exit);
diff --git a/net/802/tr.c b/net/802/tr.c
index 1bb7dc1b85c..afd8385c0c9 100644
--- a/net/802/tr.c
+++ b/net/802/tr.c
@@ -238,7 +238,7 @@ unsigned short tr_type_trans(struct sk_buff *skb, struct net_device *dev)
 		return trllc->ethertype;
 	}
 
-	return ntohs(ETH_P_802_2);
+	return ntohs(ETH_P_TR_802_2);
 }
 
 /*
@@ -340,9 +340,10 @@ static void tr_add_rif_info(struct trh_hdr *trh, struct net_device *dev)
 	unsigned int hash, rii_p = 0;
 	unsigned long flags;
 	struct rif_cache *entry;
-
+	unsigned char saddr0;
 
 	spin_lock_irqsave(&rif_lock, flags);
+	saddr0 = trh->saddr[0];
 	
 	/*
 	 *	Firstly see if the entry exists
@@ -395,7 +396,6 @@ printk("adding rif_entry: addr:%02X:%02X:%02X:%02X:%02X:%02X rcf:%04X\n",
 			entry->rcf = trh->rcf & htons((unsigned short)~TR_RCF_BROADCAST_MASK);
 			memcpy(&(entry->rseg[0]),&(trh->rseg[0]),8*sizeof(unsigned short));
 			entry->local_ring = 0;
-			trh->saddr[0]|=TR_RII; /* put the routing indicator back for tcpdump */
 		}
 		else
 		{
@@ -422,6 +422,7 @@ printk("updating rif_entry: addr:%02X:%02X:%02X:%02X:%02X:%02X rcf:%04X\n",
 		    }                                         
            	entry->last_used=jiffies;               
 	}
+	trh->saddr[0]=saddr0; /* put the routing indicator back for tcpdump */
 	spin_unlock_irqrestore(&rif_lock, flags);
 }
 
diff --git a/net/appletalk/ddp.c b/net/appletalk/ddp.c
index 1d31b3a3f1e..7982656b9c8 100644
--- a/net/appletalk/ddp.c
+++ b/net/appletalk/ddp.c
@@ -100,8 +100,7 @@ static struct sock *atalk_search_socket(struct sockaddr_at *to,
 			continue;
 
 	    	if (to->sat_addr.s_net == ATADDR_ANYNET &&
-		    to->sat_addr.s_node == ATADDR_BCAST &&
-		    at->src_net == atif->address.s_net)
+		    to->sat_addr.s_node == ATADDR_BCAST)
 			goto found;
 
 	    	if (to->sat_addr.s_net == at->src_net &&
@@ -1443,8 +1442,10 @@ static int atalk_rcv(struct sk_buff *skb, struct net_device *dev,
 	else
 		atif = atalk_find_interface(ddp->deh_dnet, ddp->deh_dnode);
 
-	/* Not ours, so we route the packet via the correct AppleTalk iface */
 	if (!atif) {
+		/* Not ours, so we route the packet via the correct
+		 * AppleTalk iface
+		 */
 		atalk_route_packet(skb, dev, ddp, &ddphv, origlen);
 		goto out;
 	}
@@ -1592,9 +1593,6 @@ static int atalk_sendmsg(struct kiocb *iocb, struct socket *sock, struct msghdr
 
 	if (usat->sat_addr.s_net || usat->sat_addr.s_node == ATADDR_ANYNODE) {
 		rt = atrtr_find(&usat->sat_addr);
-		if (!rt)
-			return -ENETUNREACH;
-
 		dev = rt->dev;
 	} else {
 		struct atalk_addr at_hint;
@@ -1603,11 +1601,12 @@ static int atalk_sendmsg(struct kiocb *iocb, struct socket *sock, struct msghdr
 		at_hint.s_net  = at->src_net;
 
 		rt = atrtr_find(&at_hint);
-		if (!rt)
-			return -ENETUNREACH;
-
 		dev = rt->dev;
 	}
+	if (!rt)
+		return -ENETUNREACH;
+
+	dev = rt->dev;
 
 	SOCK_DEBUG(sk, "SK %p: Size needed %d, device %s\n",
 			sk, size, dev->name);
@@ -1677,6 +1676,20 @@ static int atalk_sendmsg(struct kiocb *iocb, struct socket *sock, struct msghdr
 		SOCK_DEBUG(sk, "SK %p: Loop back.\n", sk);
 		/* loop back */
 		skb_orphan(skb);
+		if (ddp->deh_dnode == ATADDR_BCAST) {
+			struct atalk_addr at_lo;
+
+			at_lo.s_node = 0;
+			at_lo.s_net  = 0;
+
+			rt = atrtr_find(&at_lo);
+			if (!rt) {
+				kfree_skb(skb);
+				return -ENETUNREACH;
+			}
+			dev = rt->dev;
+			skb->dev = dev;
+		}
 		ddp_dl->request(ddp_dl, skb, dev->dev_addr);
 	} else {
 		SOCK_DEBUG(sk, "SK %p: send out.\n", sk);
diff --git a/net/atm/addr.c b/net/atm/addr.c
index 1c8867f7f54..3060fd0ba4b 100644
--- a/net/atm/addr.c
+++ b/net/atm/addr.c
@@ -44,29 +44,43 @@ static void notify_sigd(struct atm_dev *dev)
 	sigd_enq(NULL, as_itf_notify, NULL, &pvc, NULL);
 }
 
-void atm_reset_addr(struct atm_dev *dev)
+void atm_reset_addr(struct atm_dev *dev, enum atm_addr_type_t atype)
 {
 	unsigned long flags;
 	struct atm_dev_addr *this, *p;
+	struct list_head *head;
 
 	spin_lock_irqsave(&dev->lock, flags);
-	list_for_each_entry_safe(this, p, &dev->local, entry)
-	    kfree(this);
+	if (atype == ATM_ADDR_LECS)
+		head = &dev->lecs;
+	else
+		head = &dev->local;
+	list_for_each_entry_safe(this, p, head, entry) {
+		list_del(&this->entry);
+		kfree(this);
+	}
 	spin_unlock_irqrestore(&dev->lock, flags);
-	notify_sigd(dev);
+	if (head == &dev->local)
+		notify_sigd(dev);
 }
 
-int atm_add_addr(struct atm_dev *dev, struct sockaddr_atmsvc *addr)
+int atm_add_addr(struct atm_dev *dev, struct sockaddr_atmsvc *addr,
+		 enum atm_addr_type_t atype)
 {
 	unsigned long flags;
 	struct atm_dev_addr *this;
+	struct list_head *head;
 	int error;
 
 	error = check_addr(addr);
 	if (error)
 		return error;
 	spin_lock_irqsave(&dev->lock, flags);
-	list_for_each_entry(this, &dev->local, entry) {
+	if (atype == ATM_ADDR_LECS)
+		head = &dev->lecs;
+	else
+		head = &dev->local;
+	list_for_each_entry(this, head, entry) {
 		if (identical(&this->addr, addr)) {
 			spin_unlock_irqrestore(&dev->lock, flags);
 			return -EEXIST;
@@ -78,28 +92,36 @@ int atm_add_addr(struct atm_dev *dev, struct sockaddr_atmsvc *addr)
 		return -ENOMEM;
 	}
 	this->addr = *addr;
-	list_add(&this->entry, &dev->local);
+	list_add(&this->entry, head);
 	spin_unlock_irqrestore(&dev->lock, flags);
-	notify_sigd(dev);
+	if (head == &dev->local)
+		notify_sigd(dev);
 	return 0;
 }
 
-int atm_del_addr(struct atm_dev *dev, struct sockaddr_atmsvc *addr)
+int atm_del_addr(struct atm_dev *dev, struct sockaddr_atmsvc *addr,
+		 enum atm_addr_type_t atype)
 {
 	unsigned long flags;
 	struct atm_dev_addr *this;
+	struct list_head *head;
 	int error;
 
 	error = check_addr(addr);
 	if (error)
 		return error;
 	spin_lock_irqsave(&dev->lock, flags);
-	list_for_each_entry(this, &dev->local, entry) {
+	if (atype == ATM_ADDR_LECS)
+		head = &dev->lecs;
+	else
+		head = &dev->local;
+	list_for_each_entry(this, head, entry) {
 		if (identical(&this->addr, addr)) {
 			list_del(&this->entry);
 			spin_unlock_irqrestore(&dev->lock, flags);
 			kfree(this);
-			notify_sigd(dev);
+			if (head == &dev->local)
+				notify_sigd(dev);
 			return 0;
 		}
 	}
@@ -108,22 +130,27 @@ int atm_del_addr(struct atm_dev *dev, struct sockaddr_atmsvc *addr)
 }
 
 int atm_get_addr(struct atm_dev *dev, struct sockaddr_atmsvc __user * buf,
-		 size_t size)
+		 size_t size, enum atm_addr_type_t atype)
 {
 	unsigned long flags;
 	struct atm_dev_addr *this;
+	struct list_head *head;
 	int total = 0, error;
 	struct sockaddr_atmsvc *tmp_buf, *tmp_bufp;
 
 	spin_lock_irqsave(&dev->lock, flags);
-	list_for_each_entry(this, &dev->local, entry)
+	if (atype == ATM_ADDR_LECS)
+		head = &dev->lecs;
+	else
+		head = &dev->local;
+	list_for_each_entry(this, head, entry)
 	    total += sizeof(struct sockaddr_atmsvc);
 	tmp_buf = tmp_bufp = kmalloc(total, GFP_ATOMIC);
 	if (!tmp_buf) {
 		spin_unlock_irqrestore(&dev->lock, flags);
 		return -ENOMEM;
 	}
-	list_for_each_entry(this, &dev->local, entry)
+	list_for_each_entry(this, head, entry)
 	    memcpy(tmp_bufp++, &this->addr, sizeof(struct sockaddr_atmsvc));
 	spin_unlock_irqrestore(&dev->lock, flags);
 	error = total > size ? -E2BIG : total;
diff --git a/net/atm/addr.h b/net/atm/addr.h
index 3099d21feea..f39433ad45d 100644
--- a/net/atm/addr.h
+++ b/net/atm/addr.h
@@ -9,10 +9,12 @@
 #include <linux/atm.h>
 #include <linux/atmdev.h>
 
-
-void atm_reset_addr(struct atm_dev *dev);
-int atm_add_addr(struct atm_dev *dev,struct sockaddr_atmsvc *addr);
-int atm_del_addr(struct atm_dev *dev,struct sockaddr_atmsvc *addr);
-int atm_get_addr(struct atm_dev *dev,struct sockaddr_atmsvc __user *buf,size_t size);
+void atm_reset_addr(struct atm_dev *dev, enum atm_addr_type_t type);
+int atm_add_addr(struct atm_dev *dev, struct sockaddr_atmsvc *addr,
+		 enum atm_addr_type_t type);
+int atm_del_addr(struct atm_dev *dev, struct sockaddr_atmsvc *addr,
+		 enum atm_addr_type_t type);
+int atm_get_addr(struct atm_dev *dev, struct sockaddr_atmsvc __user *buf,
+		 size_t size, enum atm_addr_type_t type);
 
 #endif
diff --git a/net/atm/atm_misc.c b/net/atm/atm_misc.c
index b2113c3454a..223c7ad5bd0 100644
--- a/net/atm/atm_misc.c
+++ b/net/atm/atm_misc.c
@@ -25,7 +25,7 @@ int atm_charge(struct atm_vcc *vcc,int truesize)
 
 
 struct sk_buff *atm_alloc_charge(struct atm_vcc *vcc,int pdu_size,
-    int gfp_flags)
+    gfp_t gfp_flags)
 {
 	struct sock *sk = sk_atm(vcc);
 	int guess = atm_guess_pdu2truesize(pdu_size);
diff --git a/net/atm/br2684.c b/net/atm/br2684.c
index 289956c4dd3..72f3f7b8de8 100644
--- a/net/atm/br2684.c
+++ b/net/atm/br2684.c
@@ -220,7 +220,7 @@ static int br2684_start_xmit(struct sk_buff *skb, struct net_device *dev)
 		/* netif_stop_queue(dev); */
 		dev_kfree_skb(skb);
 		read_unlock(&devs_lock);
-		return -EUNATCH;
+		return 0;
 	}
 	if (!br2684_xmit_vcc(skb, brdev, brvcc)) {
 		/*
diff --git a/net/atm/clip.c b/net/atm/clip.c
index 28dab55a438..4f54c9a5e84 100644
--- a/net/atm/clip.c
+++ b/net/atm/clip.c
@@ -310,7 +310,7 @@ static int clip_constructor(struct neighbour *neigh)
 	if (neigh->type != RTN_UNICAST) return -EINVAL;
 
 	rcu_read_lock();
-	in_dev = rcu_dereference(__in_dev_get(dev));
+	in_dev = __in_dev_get_rcu(dev);
 	if (!in_dev) {
 		rcu_read_unlock();
 		return -EINVAL;
diff --git a/net/atm/common.c b/net/atm/common.c
index e93e838069e..63feea49fb1 100644
--- a/net/atm/common.c
+++ b/net/atm/common.c
@@ -46,7 +46,7 @@ static void __vcc_insert_socket(struct sock *sk)
 	struct atm_vcc *vcc = atm_sk(sk);
 	struct hlist_head *head = &vcc_hash[vcc->vci &
 					(VCC_HTABLE_SIZE - 1)];
-	sk->sk_hashent = vcc->vci & (VCC_HTABLE_SIZE - 1);
+	sk->sk_hash = vcc->vci & (VCC_HTABLE_SIZE - 1);
 	sk_add_node(sk, head);
 }
 
@@ -178,8 +178,6 @@ static void vcc_destroy_socket(struct sock *sk)
 		if (vcc->push)
 			vcc->push(vcc, NULL); /* atmarpd has no push */
 
-		vcc_remove_socket(sk);	/* no more receive */
-
 		while ((skb = skb_dequeue(&sk->sk_receive_queue)) != NULL) {
 			atm_return(vcc,skb->truesize);
 			kfree_skb(skb);
@@ -188,6 +186,8 @@ static void vcc_destroy_socket(struct sock *sk)
 		module_put(vcc->dev->ops->owner);
 		atm_dev_put(vcc->dev);
 	}
+
+	vcc_remove_socket(sk);
 }
 
 
diff --git a/net/atm/ioctl.c b/net/atm/ioctl.c
index d89056ec44d..a150198b05a 100644
--- a/net/atm/ioctl.c
+++ b/net/atm/ioctl.c
@@ -105,17 +105,35 @@ int vcc_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
 			if (!error)
 				sock->state = SS_CONNECTED;
 			goto done;
-		default:
+		case ATM_SETBACKEND:
+		case ATM_NEWBACKENDIF:
+			{
+				atm_backend_t backend;
+				error = get_user(backend, (atm_backend_t __user *) argp);
+				if (error)
+					goto done;
+				switch (backend) {
+					case ATM_BACKEND_PPP:
+						request_module("pppoatm");
+						break;
+					case ATM_BACKEND_BR2684:
+						request_module("br2684");
+						break;
+				}
+			}
+			break;
+		case ATMMPC_CTRL:
+		case ATMMPC_DATA:
+			request_module("mpoa");
+			break;
+		case ATMARPD_CTRL:
+			request_module("clip");
+			break;
+		case ATMLEC_CTRL:
+			request_module("lec");
 			break;
 	}
 
-	if (cmd == ATMMPC_CTRL || cmd == ATMMPC_DATA)
-		request_module("mpoa");
-	if (cmd == ATMARPD_CTRL)
-		request_module("clip");
-	if (cmd == ATMLEC_CTRL)
-		request_module("lec");
-
 	error = -ENOIOCTLCMD;
 
 	down(&ioctl_mutex);
diff --git a/net/atm/lec.c b/net/atm/lec.c
index a0752487026..ad840b9afba 100644
--- a/net/atm/lec.c
+++ b/net/atm/lec.c
@@ -686,9 +686,19 @@ static unsigned char lec_ctrl_magic[] = {
         0x01,
         0x01 };
 
+#define LEC_DATA_DIRECT_8023  2
+#define LEC_DATA_DIRECT_8025  3
+
+static int lec_is_data_direct(struct atm_vcc *vcc)
+{ 
+	return ((vcc->sap.blli[0].l3.tr9577.snap[4] == LEC_DATA_DIRECT_8023) ||
+		(vcc->sap.blli[0].l3.tr9577.snap[4] == LEC_DATA_DIRECT_8025));
+} 
+
 static void 
 lec_push(struct atm_vcc *vcc, struct sk_buff *skb)
 {
+	unsigned long flags;
         struct net_device *dev = (struct net_device *)vcc->proto_data;
         struct lec_priv *priv = (struct lec_priv *)dev->priv; 
 
@@ -728,7 +738,8 @@ lec_push(struct atm_vcc *vcc, struct sk_buff *skb)
                 skb_queue_tail(&sk->sk_receive_queue, skb);
                 sk->sk_data_ready(sk, skb->len);
         } else { /* Data frame, queue to protocol handlers */
-                unsigned char *dst;
+		struct lec_arp_table *entry;
+                unsigned char *src, *dst;
 
                 atm_return(vcc,skb->truesize);
                 if (*(uint16_t *)skb->data == htons(priv->lecid) ||
@@ -741,10 +752,30 @@ lec_push(struct atm_vcc *vcc, struct sk_buff *skb)
                         return;
                 }
 #ifdef CONFIG_TR
-                if (priv->is_trdev) dst = ((struct lecdatahdr_8025 *)skb->data)->h_dest;
+                if (priv->is_trdev)
+			dst = ((struct lecdatahdr_8025 *) skb->data)->h_dest;
                 else
 #endif
-                dst = ((struct lecdatahdr_8023 *)skb->data)->h_dest;
+		dst = ((struct lecdatahdr_8023 *) skb->data)->h_dest;
+
+		/* If this is a Data Direct VCC, and the VCC does not match
+		 * the LE_ARP cache entry, delete the LE_ARP cache entry.
+		 */
+		spin_lock_irqsave(&priv->lec_arp_lock, flags);
+		if (lec_is_data_direct(vcc)) {
+#ifdef CONFIG_TR
+			if (priv->is_trdev)
+				src = ((struct lecdatahdr_8025 *) skb->data)->h_source;
+			else
+#endif
+			src = ((struct lecdatahdr_8023 *) skb->data)->h_source;
+			entry = lec_arp_find(priv, src);
+			if (entry && entry->vcc != vcc) {
+				lec_arp_remove(priv, entry);
+				kfree(entry);
+			}
+		}
+		spin_unlock_irqrestore(&priv->lec_arp_lock, flags);
 
                 if (!(dst[0]&0x01) &&   /* Never filter Multi/Broadcast */
                     !priv->is_proxy &&  /* Proxy wants all the packets */
@@ -1990,6 +2021,12 @@ lec_arp_resolve(struct lec_priv *priv, unsigned char *mac_to_find,
                         found = entry->vcc;
 			goto out;
                 }
+		/* If the LE_ARP cache entry is still pending, reset count to 0
+		 * so another LE_ARP request can be made for this frame.
+		 */
+		if (entry->status == ESI_ARP_PENDING) {
+			entry->no_tries = 0;
+		}
                 /* Data direct VC not yet set up, check to see if the unknown
                    frame count is greater than the limit. If the limit has
                    not been reached, allow the caller to send packet to
diff --git a/net/atm/resources.c b/net/atm/resources.c
index a57a9268bd2..415d2615d47 100644
--- a/net/atm/resources.c
+++ b/net/atm/resources.c
@@ -40,6 +40,7 @@ static struct atm_dev *__alloc_atm_dev(const char *type)
 	dev->link_rate = ATM_OC3_PCR;
 	spin_lock_init(&dev->lock);
 	INIT_LIST_HEAD(&dev->local);
+	INIT_LIST_HEAD(&dev->lecs);
 
 	return dev;
 }
@@ -320,10 +321,12 @@ int atm_dev_ioctl(unsigned int cmd, void __user *arg)
 				error = -EPERM;
 				goto done;
 			}
-			atm_reset_addr(dev);
+			atm_reset_addr(dev, ATM_ADDR_LOCAL);
 			break;
 		case ATM_ADDADDR:
 		case ATM_DELADDR:
+		case ATM_ADDLECSADDR:
+		case ATM_DELLECSADDR:
 			if (!capable(CAP_NET_ADMIN)) {
 				error = -EPERM;
 				goto done;
@@ -335,14 +338,21 @@ int atm_dev_ioctl(unsigned int cmd, void __user *arg)
 					error = -EFAULT;
 					goto done;
 				}
-				if (cmd == ATM_ADDADDR)
-					error = atm_add_addr(dev, &addr);
+				if (cmd == ATM_ADDADDR || cmd == ATM_ADDLECSADDR)
+					error = atm_add_addr(dev, &addr,
+							     (cmd == ATM_ADDADDR ?
+							      ATM_ADDR_LOCAL : ATM_ADDR_LECS));
 				else
-					error = atm_del_addr(dev, &addr);
+					error = atm_del_addr(dev, &addr,
+							     (cmd == ATM_DELADDR ?
+							      ATM_ADDR_LOCAL : ATM_ADDR_LECS));
 				goto done;
 			}
 		case ATM_GETADDR:
-			error = atm_get_addr(dev, buf, len);
+		case ATM_GETLECSADDR:
+			error = atm_get_addr(dev, buf, len,
+					     (cmd == ATM_GETADDR ?
+					      ATM_ADDR_LOCAL : ATM_ADDR_LECS));
 			if (error < 0)
 				goto done;
 			size = error;
diff --git a/net/atm/signaling.c b/net/atm/signaling.c
index f7c449ac180..e7211a7f382 100644
--- a/net/atm/signaling.c
+++ b/net/atm/signaling.c
@@ -217,8 +217,9 @@ void sigd_enq(struct atm_vcc *vcc,enum atmsvc_msg_type type,
 static void purge_vcc(struct atm_vcc *vcc)
 {
 	if (sk_atm(vcc)->sk_family == PF_ATMSVC &&
-	    !test_bit(ATM_VF_META,&vcc->flags)) {
-		set_bit(ATM_VF_RELEASED,&vcc->flags);
+	    !test_bit(ATM_VF_META, &vcc->flags)) {
+		set_bit(ATM_VF_RELEASED, &vcc->flags);
+		clear_bit(ATM_VF_REGIS, &vcc->flags);
 		vcc_release_async(vcc, -EUNATCH);
 	}
 }
@@ -243,8 +244,7 @@ static void sigd_close(struct atm_vcc *vcc)
 		sk_for_each(s, node, head) {
 			struct atm_vcc *vcc = atm_sk(s);
 
-			if (vcc->dev)
-				purge_vcc(vcc);
+			purge_vcc(vcc);
 		}
 	}
 	read_unlock(&vcc_sklist_lock);
diff --git a/net/atm/svc.c b/net/atm/svc.c
index 08e46052a3e..d7b266136bf 100644
--- a/net/atm/svc.c
+++ b/net/atm/svc.c
@@ -302,6 +302,7 @@ static int svc_listen(struct socket *sock,int backlog)
 		error = -EINVAL;
 		goto out;
 	}
+	vcc_insert_socket(sk);
 	set_bit(ATM_VF_WAITING, &vcc->flags);
 	prepare_to_wait(sk->sk_sleep, &wait, TASK_UNINTERRUPTIBLE);
 	sigd_enq(vcc,as_listen,NULL,NULL,&vcc->local);
diff --git a/net/ax25/ax25_in.c b/net/ax25/ax25_in.c
index 810c9c76c2e..73cfc3411c4 100644
--- a/net/ax25/ax25_in.c
+++ b/net/ax25/ax25_in.c
@@ -123,7 +123,7 @@ int ax25_rx_iframe(ax25_cb *ax25, struct sk_buff *skb)
 		}
 
 		skb_pull(skb, 1);	/* Remove PID */
-		skb->h.raw    = skb->data;
+		skb->mac.raw  = skb->nh.raw;
 		skb->nh.raw   = skb->data;
 		skb->dev      = ax25->ax25_dev->dev;
 		skb->pkt_type = PACKET_HOST;
diff --git a/net/bluetooth/l2cap.c b/net/bluetooth/l2cap.c
index d3d6bc54721..59b2dd36baa 100644
--- a/net/bluetooth/l2cap.c
+++ b/net/bluetooth/l2cap.c
@@ -372,7 +372,7 @@ static struct proto l2cap_proto = {
 	.obj_size	= sizeof(struct l2cap_pinfo)
 };
 
-static struct sock *l2cap_sock_alloc(struct socket *sock, int proto, unsigned int __nocast prio)
+static struct sock *l2cap_sock_alloc(struct socket *sock, int proto, gfp_t prio)
 {
 	struct sock *sk;
 
diff --git a/net/bluetooth/rfcomm/core.c b/net/bluetooth/rfcomm/core.c
index 173f46e8cda..35adce6482b 100644
--- a/net/bluetooth/rfcomm/core.c
+++ b/net/bluetooth/rfcomm/core.c
@@ -229,7 +229,7 @@ static void rfcomm_dlc_clear_state(struct rfcomm_dlc *d)
 	d->rx_credits = RFCOMM_DEFAULT_CREDITS;
 }
 
-struct rfcomm_dlc *rfcomm_dlc_alloc(unsigned int __nocast prio)
+struct rfcomm_dlc *rfcomm_dlc_alloc(gfp_t prio)
 {
 	struct rfcomm_dlc *d = kmalloc(sizeof(*d), prio);
 	if (!d)
diff --git a/net/bluetooth/rfcomm/sock.c b/net/bluetooth/rfcomm/sock.c
index f49e7e938bf..a2b30f0aedb 100644
--- a/net/bluetooth/rfcomm/sock.c
+++ b/net/bluetooth/rfcomm/sock.c
@@ -284,7 +284,7 @@ static struct proto rfcomm_proto = {
 	.obj_size	= sizeof(struct rfcomm_pinfo)
 };
 
-static struct sock *rfcomm_sock_alloc(struct socket *sock, int proto, unsigned int __nocast prio)
+static struct sock *rfcomm_sock_alloc(struct socket *sock, int proto, gfp_t prio)
 {
 	struct rfcomm_dlc *d;
 	struct sock *sk;
diff --git a/net/bluetooth/rfcomm/tty.c b/net/bluetooth/rfcomm/tty.c
index 1bca860a610..158a9c46d86 100644
--- a/net/bluetooth/rfcomm/tty.c
+++ b/net/bluetooth/rfcomm/tty.c
@@ -286,7 +286,7 @@ static inline void rfcomm_set_owner_w(struct sk_buff *skb, struct rfcomm_dev *de
 	skb->destructor = rfcomm_wfree;
 }
 
-static struct sk_buff *rfcomm_wmalloc(struct rfcomm_dev *dev, unsigned long size, unsigned int __nocast priority)
+static struct sk_buff *rfcomm_wmalloc(struct rfcomm_dev *dev, unsigned long size, gfp_t priority)
 {
 	if (atomic_read(&dev->wmem_alloc) < rfcomm_room(dev->dlc)) {
 		struct sk_buff *skb = alloc_skb(size, priority);
diff --git a/net/bluetooth/sco.c b/net/bluetooth/sco.c
index ce7ab7dfa0b..997e42df115 100644
--- a/net/bluetooth/sco.c
+++ b/net/bluetooth/sco.c
@@ -418,7 +418,7 @@ static struct proto sco_proto = {
 	.obj_size	= sizeof(struct sco_pinfo)
 };
 
-static struct sock *sco_sock_alloc(struct socket *sock, int proto, unsigned int __nocast prio)
+static struct sock *sco_sock_alloc(struct socket *sock, int proto, gfp_t prio)
 {
 	struct sock *sk;
 
diff --git a/net/bridge/br_forward.c b/net/bridge/br_forward.c
index 069253f830c..2d24fb400e0 100644
--- a/net/bridge/br_forward.c
+++ b/net/bridge/br_forward.c
@@ -31,7 +31,8 @@ static inline int should_deliver(const struct net_bridge_port *p,
 
 int br_dev_queue_push_xmit(struct sk_buff *skb)
 {
-	if (skb->len > skb->dev->mtu) 
+	/* drop mtu oversized packets except tso */
+	if (skb->len > skb->dev->mtu && !skb_shinfo(skb)->tso_size)
 		kfree_skb(skb);
 	else {
 #ifdef CONFIG_BRIDGE_NETFILTER
diff --git a/net/bridge/br_if.c b/net/bridge/br_if.c
index 91bb895375f..defcf6a8607 100644
--- a/net/bridge/br_if.c
+++ b/net/bridge/br_if.c
@@ -79,7 +79,6 @@ static void destroy_nbp(struct net_bridge_port *p)
 {
 	struct net_device *dev = p->dev;
 
-	dev->br_port = NULL;
 	p->br = NULL;
 	p->dev = NULL;
 	dev_put(dev);
@@ -100,6 +99,7 @@ static void del_nbp(struct net_bridge_port *p)
 	struct net_bridge *br = p->br;
 	struct net_device *dev = p->dev;
 
+	dev->br_port = NULL;
 	dev_set_promiscuity(dev, -1);
 
 	spin_lock_bh(&br->lock);
diff --git a/net/bridge/netfilter/ebtables.c b/net/bridge/netfilter/ebtables.c
index c4540144f0f..f8ffbf6e233 100644
--- a/net/bridge/netfilter/ebtables.c
+++ b/net/bridge/netfilter/ebtables.c
@@ -26,6 +26,7 @@
 #include <linux/spinlock.h>
 #include <asm/uaccess.h>
 #include <linux/smp.h>
+#include <linux/cpumask.h>
 #include <net/sock.h>
 /* needed for logical [in,out]-dev filtering */
 #include "../br_private.h"
@@ -823,10 +824,11 @@ static int translate_table(struct ebt_replace *repl,
 		/* this will get free'd in do_replace()/ebt_register_table()
 		   if an error occurs */
 		newinfo->chainstack = (struct ebt_chainstack **)
-		   vmalloc(num_possible_cpus() * sizeof(struct ebt_chainstack));
+		   vmalloc((highest_possible_processor_id()+1) 
+				   		* sizeof(struct ebt_chainstack));
 		if (!newinfo->chainstack)
 			return -ENOMEM;
-		for (i = 0; i < num_possible_cpus(); i++) {
+		for_each_cpu(i) {
 			newinfo->chainstack[i] =
 			   vmalloc(udc_cnt * sizeof(struct ebt_chainstack));
 			if (!newinfo->chainstack[i]) {
@@ -895,9 +897,12 @@ static void get_counters(struct ebt_counter *oldcounters,
 
 	/* counters of cpu 0 */
 	memcpy(counters, oldcounters,
-	   sizeof(struct ebt_counter) * nentries);
+	       sizeof(struct ebt_counter) * nentries);
+
 	/* add other counters to those of cpu 0 */
-	for (cpu = 1; cpu < num_possible_cpus(); cpu++) {
+	for_each_cpu(cpu) {
+		if (cpu == 0)
+			continue;
 		counter_base = COUNTER_BASE(oldcounters, nentries, cpu);
 		for (i = 0; i < nentries; i++) {
 			counters[i].pcnt += counter_base[i].pcnt;
@@ -929,7 +934,8 @@ static int do_replace(void __user *user, unsigned int len)
 		BUGPRINT("Entries_size never zero\n");
 		return -EINVAL;
 	}
-	countersize = COUNTER_OFFSET(tmp.nentries) * num_possible_cpus();
+	countersize = COUNTER_OFFSET(tmp.nentries) * 
+					(highest_possible_processor_id()+1);
 	newinfo = (struct ebt_table_info *)
 	   vmalloc(sizeof(struct ebt_table_info) + countersize);
 	if (!newinfo)
@@ -1022,7 +1028,7 @@ static int do_replace(void __user *user, unsigned int len)
 
 	vfree(table->entries);
 	if (table->chainstack) {
-		for (i = 0; i < num_possible_cpus(); i++)
+		for_each_cpu(i)
 			vfree(table->chainstack[i]);
 		vfree(table->chainstack);
 	}
@@ -1040,7 +1046,7 @@ free_counterstmp:
 	vfree(counterstmp);
 	/* can be initialized in translate_table() */
 	if (newinfo->chainstack) {
-		for (i = 0; i < num_possible_cpus(); i++)
+		for_each_cpu(i)
 			vfree(newinfo->chainstack[i]);
 		vfree(newinfo->chainstack);
 	}
@@ -1132,7 +1138,8 @@ int ebt_register_table(struct ebt_table *table)
 		return -EINVAL;
 	}
 
-	countersize = COUNTER_OFFSET(table->table->nentries) * num_possible_cpus();
+	countersize = COUNTER_OFFSET(table->table->nentries) *
+					(highest_possible_processor_id()+1);
 	newinfo = (struct ebt_table_info *)
 	   vmalloc(sizeof(struct ebt_table_info) + countersize);
 	ret = -ENOMEM;
@@ -1186,7 +1193,7 @@ free_unlock:
 	up(&ebt_mutex);
 free_chainstack:
 	if (newinfo->chainstack) {
-		for (i = 0; i < num_possible_cpus(); i++)
+		for_each_cpu(i)
 			vfree(newinfo->chainstack[i]);
 		vfree(newinfo->chainstack);
 	}
@@ -1209,7 +1216,7 @@ void ebt_unregister_table(struct ebt_table *table)
 	up(&ebt_mutex);
 	vfree(table->private->entries);
 	if (table->private->chainstack) {
-		for (i = 0; i < num_possible_cpus(); i++)
+		for_each_cpu(i)
 			vfree(table->private->chainstack[i]);
 		vfree(table->private->chainstack);
 	}
diff --git a/net/core/datagram.c b/net/core/datagram.c
index da9bf71421a..81987df536e 100644
--- a/net/core/datagram.c
+++ b/net/core/datagram.c
@@ -211,74 +211,45 @@ void skb_free_datagram(struct sock *sk, struct sk_buff *skb)
 int skb_copy_datagram_iovec(const struct sk_buff *skb, int offset,
 			    struct iovec *to, int len)
 {
-	int start = skb_headlen(skb);
-	int i, copy = start - offset;
-
-	/* Copy header. */
-	if (copy > 0) {
-		if (copy > len)
-			copy = len;
-		if (memcpy_toiovec(to, skb->data + offset, copy))
-			goto fault;
-		if ((len -= copy) == 0)
-			return 0;
-		offset += copy;
-	}
+	int i, err, fraglen, end = 0;
+	struct sk_buff *next = skb_shinfo(skb)->frag_list;
+next_skb:
+	fraglen = skb_headlen(skb);
+	i = -1;
 
-	/* Copy paged appendix. Hmm... why does this look so complicated? */
-	for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
-		int end;
-
-		BUG_TRAP(start <= offset + len);
+	while (1) {
+		int start = end;
 
-		end = start + skb_shinfo(skb)->frags[i].size;
-		if ((copy = end - offset) > 0) {
-			int err;
-			u8  *vaddr;
-			skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
-			struct page *page = frag->page;
+		if ((end += fraglen) > offset) {
+			int copy = end - offset, o = offset - start;
 
 			if (copy > len)
 				copy = len;
-			vaddr = kmap(page);
-			err = memcpy_toiovec(to, vaddr + frag->page_offset +
-					     offset - start, copy);
-			kunmap(page);
+			if (i == -1)
+				err = memcpy_toiovec(to, skb->data + o, copy);
+			else {
+				skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
+				struct page *page = frag->page;
+				void *p = kmap(page) + frag->page_offset + o;
+				err = memcpy_toiovec(to, p, copy);
+				kunmap(page);
+			}
 			if (err)
 				goto fault;
 			if (!(len -= copy))
 				return 0;
 			offset += copy;
 		}
-		start = end;
+		if (++i >= skb_shinfo(skb)->nr_frags)
+			break;
+		fraglen = skb_shinfo(skb)->frags[i].size;
 	}
-
-	if (skb_shinfo(skb)->frag_list) {
-		struct sk_buff *list = skb_shinfo(skb)->frag_list;
-
-		for (; list; list = list->next) {
-			int end;
-
-			BUG_TRAP(start <= offset + len);
-
-			end = start + list->len;
-			if ((copy = end - offset) > 0) {
-				if (copy > len)
-					copy = len;
-				if (skb_copy_datagram_iovec(list,
-							    offset - start,
-							    to, copy))
-					goto fault;
-				if ((len -= copy) == 0)
-					return 0;
-				offset += copy;
-			}
-			start = end;
-		}
+	if (next) {
+		skb = next;
+		BUG_ON(skb_shinfo(skb)->frag_list);
+		next = skb->next;
+		goto next_skb;
 	}
-	if (!len)
-		return 0;
-
 fault:
 	return -EFAULT;
 }
diff --git a/net/core/dev.c b/net/core/dev.c
index c01511e3d0c..a44eeef24ed 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -574,6 +574,8 @@ struct net_device *dev_getbyhwaddr(unsigned short type, char *ha)
 	return dev;
 }
 
+EXPORT_SYMBOL(dev_getbyhwaddr);
+
 struct net_device *dev_getfirstbyhwtype(unsigned short type)
 {
 	struct net_device *dev;
@@ -1130,7 +1132,7 @@ static inline int illegal_highdma(struct net_device *dev, struct sk_buff *skb)
 #endif
 
 /* Keep head the same: replace data */
-int __skb_linearize(struct sk_buff *skb, unsigned int __nocast gfp_mask)
+int __skb_linearize(struct sk_buff *skb, gfp_t gfp_mask)
 {
 	unsigned int size;
 	u8 *data;
@@ -1257,6 +1259,8 @@ int dev_queue_xmit(struct sk_buff *skb)
 	      	if (skb_checksum_help(skb, 0))
 	      		goto out_kfree_skb;
 
+	spin_lock_prefetch(&dev->queue_lock);
+
 	/* Disable soft irqs for various locks below. Also 
 	 * stops preemption for RCU. 
 	 */
diff --git a/net/core/neighbour.c b/net/core/neighbour.c
index 39fc55edf69..e68700f950a 100644
--- a/net/core/neighbour.c
+++ b/net/core/neighbour.c
@@ -61,7 +61,9 @@ static int pneigh_ifdown(struct neigh_table *tbl, struct net_device *dev);
 void neigh_changeaddr(struct neigh_table *tbl, struct net_device *dev);
 
 static struct neigh_table *neigh_tables;
+#ifdef CONFIG_PROC_FS
 static struct file_operations neigh_stat_seq_fops;
+#endif
 
 /*
    Neighbour hash table buckets are protected with rwlock tbl->lock.
@@ -173,39 +175,10 @@ static void pneigh_queue_purge(struct sk_buff_head *list)
 	}
 }
 
-void neigh_changeaddr(struct neigh_table *tbl, struct net_device *dev)
-{
-	int i;
-
-	write_lock_bh(&tbl->lock);
-
-	for (i=0; i <= tbl->hash_mask; i++) {
-		struct neighbour *n, **np;
-
-		np = &tbl->hash_buckets[i];
-		while ((n = *np) != NULL) {
-			if (dev && n->dev != dev) {
-				np = &n->next;
-				continue;
-			}
-			*np = n->next;
-			write_lock_bh(&n->lock);
-			n->dead = 1;
-			neigh_del_timer(n);
-			write_unlock_bh(&n->lock);
-			neigh_release(n);
-		}
-	}
-
-        write_unlock_bh(&tbl->lock);
-}
-
-int neigh_ifdown(struct neigh_table *tbl, struct net_device *dev)
+static void neigh_flush_dev(struct neigh_table *tbl, struct net_device *dev)
 {
 	int i;
 
-	write_lock_bh(&tbl->lock);
-
 	for (i = 0; i <= tbl->hash_mask; i++) {
 		struct neighbour *n, **np = &tbl->hash_buckets[i];
 
@@ -241,7 +214,19 @@ int neigh_ifdown(struct neigh_table *tbl, struct net_device *dev)
 			neigh_release(n);
 		}
 	}
+}
+
+void neigh_changeaddr(struct neigh_table *tbl, struct net_device *dev)
+{
+	write_lock_bh(&tbl->lock);
+	neigh_flush_dev(tbl, dev);
+	write_unlock_bh(&tbl->lock);
+}
 
+int neigh_ifdown(struct neigh_table *tbl, struct net_device *dev)
+{
+	write_lock_bh(&tbl->lock);
+	neigh_flush_dev(tbl, dev);
 	pneigh_ifdown(tbl, dev);
 	write_unlock_bh(&tbl->lock);
 
@@ -725,6 +710,14 @@ static __inline__ int neigh_max_probes(struct neighbour *n)
 		p->ucast_probes + p->app_probes + p->mcast_probes);
 }
 
+static inline void neigh_add_timer(struct neighbour *n, unsigned long when)
+{
+	if (unlikely(mod_timer(&n->timer, when))) {
+		printk("NEIGH: BUG, double timer add, state is %x\n",
+		       n->nud_state);
+		dump_stack();
+	}
+}
 
 /* Called when a timer expires for a neighbour entry. */
 
@@ -806,11 +799,10 @@ static void neigh_timer_handler(unsigned long arg)
 	}
 
 	if (neigh->nud_state & NUD_IN_TIMER) {
-		neigh_hold(neigh);
 		if (time_before(next, jiffies + HZ/2))
 			next = jiffies + HZ/2;
-		neigh->timer.expires = next;
-		add_timer(&neigh->timer);
+		if (!mod_timer(&neigh->timer, next))
+			neigh_hold(neigh);
 	}
 	if (neigh->nud_state & (NUD_INCOMPLETE | NUD_PROBE)) {
 		struct sk_buff *skb = skb_peek(&neigh->arp_queue);
@@ -852,8 +844,7 @@ int __neigh_event_send(struct neighbour *neigh, struct sk_buff *skb)
 			atomic_set(&neigh->probes, neigh->parms->ucast_probes);
 			neigh->nud_state     = NUD_INCOMPLETE;
 			neigh_hold(neigh);
-			neigh->timer.expires = now + 1;
-			add_timer(&neigh->timer);
+			neigh_add_timer(neigh, now + 1);
 		} else {
 			neigh->nud_state = NUD_FAILED;
 			write_unlock_bh(&neigh->lock);
@@ -866,8 +857,8 @@ int __neigh_event_send(struct neighbour *neigh, struct sk_buff *skb)
 		NEIGH_PRINTK2("neigh %p is delayed.\n", neigh);
 		neigh_hold(neigh);
 		neigh->nud_state = NUD_DELAY;
-		neigh->timer.expires = jiffies + neigh->parms->delay_probe_time;
-		add_timer(&neigh->timer);
+		neigh_add_timer(neigh,
+				jiffies + neigh->parms->delay_probe_time);
 	}
 
 	if (neigh->nud_state == NUD_INCOMPLETE) {
@@ -1013,10 +1004,10 @@ int neigh_update(struct neighbour *neigh, const u8 *lladdr, u8 new,
 		neigh_del_timer(neigh);
 		if (new & NUD_IN_TIMER) {
 			neigh_hold(neigh);
-			neigh->timer.expires = jiffies + 
+			neigh_add_timer(neigh, (jiffies + 
 						((new & NUD_REACHABLE) ? 
-						 neigh->parms->reachable_time : 0);
-			add_timer(&neigh->timer);
+						 neigh->parms->reachable_time :
+						 0)));
 		}
 		neigh->nud_state = new;
 	}
@@ -1634,12 +1625,9 @@ static int neightbl_fill_info(struct neigh_table *tbl, struct sk_buff *skb,
 
 		memset(&ndst, 0, sizeof(ndst));
 
-		for (cpu = 0; cpu < NR_CPUS; cpu++) {
+		for_each_cpu(cpu) {
 			struct neigh_statistics	*st;
 
-			if (!cpu_possible(cpu))
-				continue;
-
 			st = per_cpu_ptr(tbl->stats, cpu);
 			ndst.ndts_allocs		+= st->allocs;
 			ndst.ndts_destroys		+= st->destroys;
diff --git a/net/core/netpoll.c b/net/core/netpoll.c
index 5265dfd6992..802fe11efad 100644
--- a/net/core/netpoll.c
+++ b/net/core/netpoll.c
@@ -703,7 +703,7 @@ int netpoll_setup(struct netpoll *np)
 
 	if (!np->local_ip) {
 		rcu_read_lock();
-		in_dev = __in_dev_get(ndev);
+		in_dev = __in_dev_get_rcu(ndev);
 
 		if (!in_dev || !in_dev->ifa_list) {
 			rcu_read_unlock();
diff --git a/net/core/pktgen.c b/net/core/pktgen.c
index ef430b1e8e4..7fc3e9e28c3 100644
--- a/net/core/pktgen.c
+++ b/net/core/pktgen.c
@@ -75,7 +75,7 @@
  * By design there should only be *one* "controlling" process. In practice 
  * multiple write accesses gives unpredictable result. Understood by "write" 
  * to /proc gives result code thats should be read be the "writer".
- * For pratical use this should be no problem.
+ * For practical use this should be no problem.
  *
  * Note when adding devices to a specific CPU there good idea to also assign 
  * /proc/irq/XX/smp_affinity so TX-interrupts gets bound to the same CPU. 
@@ -96,7 +96,7 @@
  * New xmit() return, do_div and misc clean up by Stephen Hemminger 
  * <shemminger@osdl.org> 040923
  *
- * Rany Dunlap fixed u64 printk compiler waring 
+ * Randy Dunlap fixed u64 printk compiler waring 
  *
  * Remove FCS from BW calculation.  Lennert Buytenhek <buytenh@wantstofly.org>
  * New time handling. Lennert Buytenhek <buytenh@wantstofly.org> 041213
@@ -137,6 +137,7 @@
 #include <linux/ipv6.h>
 #include <linux/udp.h>
 #include <linux/proc_fs.h>
+#include <linux/seq_file.h>
 #include <linux/wait.h>
 #include <net/checksum.h>
 #include <net/ipv6.h>
@@ -151,7 +152,7 @@
 #include <asm/timex.h>
 
 
-#define VERSION  "pktgen v2.62: Packet Generator for packet performance testing.\n"
+#define VERSION  "pktgen v2.63: Packet Generator for packet performance testing.\n"
 
 /* #define PG_DEBUG(a) a */
 #define PG_DEBUG(a) 
@@ -177,8 +178,8 @@
 #define T_REMDEV      (1<<3)  /* Remove all devs */
 
 /* Locks */
-#define   thread_lock()        spin_lock(&_thread_lock)
-#define   thread_unlock()      spin_unlock(&_thread_lock)
+#define   thread_lock()        down(&pktgen_sem)
+#define   thread_unlock()      up(&pktgen_sem)
 
 /* If lock -- can be removed after some work */
 #define   if_lock(t)           spin_lock(&(t->if_lock));
@@ -187,6 +188,8 @@
 /* Used to help with determining the pkts on receive */
 #define PKTGEN_MAGIC 0xbe9be955
 #define PG_PROC_DIR "pktgen"
+#define PGCTRL	    "pgctrl"
+static struct proc_dir_entry *pg_proc_dir = NULL;
 
 #define MAX_CFLOWS  65536
 
@@ -202,11 +205,8 @@ struct pktgen_dev {
 	 * Try to keep frequent/infrequent used vars. separated.
 	 */
 
-        char ifname[32];
-        struct proc_dir_entry *proc_ent;
+        char ifname[IFNAMSIZ];
         char result[512];
-        /* proc file names */
-        char fname[80];
 
         struct pktgen_thread* pg_thread; /* the owner */
         struct pktgen_dev *next; /* Used for chaining in the thread's run-queue */
@@ -244,7 +244,7 @@ struct pktgen_dev {
         __u32 seq_num;
         
         int clone_skb; /* Use multiple SKBs during packet gen.  If this number
-                          * is greater than 1, then that many coppies of the same
+                          * is greater than 1, then that many copies of the same
                           * packet will be sent before a new packet is allocated.
                           * For instance, if you want to send 1024 identical packets
                           * before creating a new packet, set clone_skb to 1024.
@@ -330,8 +330,6 @@ struct pktgen_thread {
         struct pktgen_dev *if_list;           /* All device here */
         struct pktgen_thread* next;
         char name[32];
-        char fname[128]; /* name of proc file */
-        struct proc_dir_entry *proc_ent;
         char result[512];
         u32 max_before_softirq; /* We'll call do_softirq to prevent starvation. */
         
@@ -396,7 +394,7 @@ static inline s64 divremdi3(s64 x, s64 y, int type)
 
 /* End of hacks to deal with 64-bit math on x86 */
 
-/** Convert to miliseconds */
+/** Convert to milliseconds */
 static inline __u64 tv_to_ms(const struct timeval* tv) 
 {
         __u64 ms = tv->tv_usec / 1000;
@@ -425,7 +423,7 @@ static inline __u64 pg_div64(__u64 n, __u64 base)
 {
         __u64 tmp = n;
 /*
- * How do we know if the architectrure we are running on
+ * How do we know if the architecture we are running on
  * supports division with 64 bit base?
  * 
  */
@@ -473,16 +471,6 @@ static inline __u64 tv_diff(const struct timeval* a, const struct timeval* b)
 
 static char version[] __initdata = VERSION;
 
-static ssize_t proc_pgctrl_read(struct file* file, char __user * buf, size_t count, loff_t *ppos);
-static ssize_t proc_pgctrl_write(struct file* file, const char __user * buf, size_t count, loff_t *ppos);
-static int proc_if_read(char *buf , char **start, off_t offset, int len, int *eof, void *data);
-
-static int proc_thread_read(char *buf , char **start, off_t offset, int len, int *eof, void *data);
-static int proc_if_write(struct file *file, const char __user *user_buffer, unsigned long count, void *data);
-static int proc_thread_write(struct file *file, const char __user *user_buffer, unsigned long count, void *data);
-static int create_proc_dir(void);
-static int remove_proc_dir(void);
-
 static int pktgen_remove_device(struct pktgen_thread* t, struct pktgen_dev *i);
 static int pktgen_add_device(struct pktgen_thread* t, const char* ifname);
 static struct pktgen_thread* pktgen_find_thread(const char* name);
@@ -503,83 +491,41 @@ static int pg_delay_d = 0;
 static int pg_clone_skb_d = 0;
 static int debug = 0;
 
-static DEFINE_SPINLOCK(_thread_lock);
+static DECLARE_MUTEX(pktgen_sem);
 static struct pktgen_thread *pktgen_threads = NULL;
 
-static char module_fname[128];
-static struct proc_dir_entry *module_proc_ent = NULL;
-
 static struct notifier_block pktgen_notifier_block = {
 	.notifier_call = pktgen_device_event,
 };
 
-static struct file_operations pktgen_fops = {
-        .read     = proc_pgctrl_read,
-        .write    = proc_pgctrl_write,
-	/*  .ioctl    = pktgen_ioctl, later maybe */
-};
-
 /*
  * /proc handling functions 
  *
  */
 
-static struct proc_dir_entry *pg_proc_dir = NULL;
-static int proc_pgctrl_read_eof=0;
-
-static ssize_t proc_pgctrl_read(struct file* file, char __user * buf,
-                                 size_t count, loff_t *ppos)
+static int pgctrl_show(struct seq_file *seq, void *v)
 { 
-	char data[200];
-	int len = 0;
-
-	if(proc_pgctrl_read_eof) {
-		proc_pgctrl_read_eof=0;
-		len = 0;
-		goto out;
-	}
-
-	sprintf(data, "%s", VERSION); 
-
-	len = strlen(data);
-
-	if(len > count) {
-		len =-EFAULT;
-		goto out;
-	}  	
-
-	if (copy_to_user(buf, data, len)) {
-		len =-EFAULT;
-		goto out;
-	}  
-
-	*ppos += len;
-	proc_pgctrl_read_eof=1; /* EOF next call */
-
- out:
-	return len;
+	seq_puts(seq, VERSION);
+	return 0;
 }
 
-static ssize_t proc_pgctrl_write(struct file* file,const char __user * buf,
-				 size_t count, loff_t *ppos)
+static ssize_t pgctrl_write(struct file* file,const char __user * buf,
+			    size_t count, loff_t *ppos)
 {
-	char *data = NULL;
 	int err = 0;
+	char data[128];
 
         if (!capable(CAP_NET_ADMIN)){
                 err = -EPERM;
 		goto out;
         }
 
-	data = (void*)vmalloc ((unsigned int)count);
+	if (count > sizeof(data))
+		count = sizeof(data);
 
-	if(!data) {
-		err = -ENOMEM;
-		goto out;
-	}
 	if (copy_from_user(data, buf, count)) {
-		err =-EFAULT;
-		goto out_free;
+		err = -EFAULT;
+		goto out;
 	}  
 	data[count-1] = 0; /* Make string */
 
@@ -594,31 +540,40 @@ static ssize_t proc_pgctrl_write(struct file* file,const char __user * buf,
 
 	err = count;
 
- out_free:
-	vfree (data);
  out:
         return err;
 }
 
-static int proc_if_read(char *buf , char **start, off_t offset,
-                           int len, int *eof, void *data)
+static int pgctrl_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, pgctrl_show, PDE(inode)->data);
+}
+
+static struct file_operations pktgen_fops = {
+	.owner	  = THIS_MODULE,
+	.open	  = pgctrl_open,
+        .read     = seq_read,
+	.llseek	  = seq_lseek,
+        .write    = pgctrl_write,
+	.release  = single_release,
+};
+
+static int pktgen_if_show(struct seq_file *seq, void *v)
 {
-	char *p;
 	int i;
-        struct pktgen_dev *pkt_dev = (struct pktgen_dev*)(data);
+        struct pktgen_dev *pkt_dev = seq->private;
         __u64 sa;
         __u64 stopped;
         __u64 now = getCurUs();
         
-	p = buf;
-	p += sprintf(p, "Params: count %llu  min_pkt_size: %u  max_pkt_size: %u\n",
-		     (unsigned long long) pkt_dev->count,
-		     pkt_dev->min_pkt_size, pkt_dev->max_pkt_size);
+	seq_printf(seq, "Params: count %llu  min_pkt_size: %u  max_pkt_size: %u\n",
+		   (unsigned long long) pkt_dev->count,
+		   pkt_dev->min_pkt_size, pkt_dev->max_pkt_size);
 
-	p += sprintf(p, "     frags: %d  delay: %u  clone_skb: %d  ifname: %s\n",
-                     pkt_dev->nfrags, 1000*pkt_dev->delay_us+pkt_dev->delay_ns, pkt_dev->clone_skb, pkt_dev->ifname);
+	seq_printf(seq, "     frags: %d  delay: %u  clone_skb: %d  ifname: %s\n",
+		   pkt_dev->nfrags, 1000*pkt_dev->delay_us+pkt_dev->delay_ns, pkt_dev->clone_skb, pkt_dev->ifname);
 
-	p += sprintf(p, "     flows: %u flowlen: %u\n", pkt_dev->cflows, pkt_dev->lflow);
+	seq_printf(seq, "     flows: %u flowlen: %u\n", pkt_dev->cflows, pkt_dev->lflow);
 
 
 	if(pkt_dev->flags & F_IPV6) {
@@ -626,19 +581,19 @@ static int proc_if_read(char *buf , char **start, off_t offset,
 		fmt_ip6(b1,  pkt_dev->in6_saddr.s6_addr);
 		fmt_ip6(b2,  pkt_dev->min_in6_saddr.s6_addr);
 		fmt_ip6(b3,  pkt_dev->max_in6_saddr.s6_addr);
-		p += sprintf(p, "     saddr: %s  min_saddr: %s  max_saddr: %s\n", b1, b2, b3);
+		seq_printf(seq, "     saddr: %s  min_saddr: %s  max_saddr: %s\n", b1, b2, b3);
 
 		fmt_ip6(b1,  pkt_dev->in6_daddr.s6_addr);
 		fmt_ip6(b2,  pkt_dev->min_in6_daddr.s6_addr);
 		fmt_ip6(b3,  pkt_dev->max_in6_daddr.s6_addr);
-		p += sprintf(p, "     daddr: %s  min_daddr: %s  max_daddr: %s\n", b1, b2, b3);
+		seq_printf(seq, "     daddr: %s  min_daddr: %s  max_daddr: %s\n", b1, b2, b3);
 
 	} 
 	else 
-		p += sprintf(p, "     dst_min: %s  dst_max: %s\n     src_min: %s  src_max: %s\n",
-                     pkt_dev->dst_min, pkt_dev->dst_max, pkt_dev->src_min, pkt_dev->src_max);
+		seq_printf(seq,"     dst_min: %s  dst_max: %s\n     src_min: %s  src_max: %s\n",
+			   pkt_dev->dst_min, pkt_dev->dst_max, pkt_dev->src_min, pkt_dev->src_max);
 
-        p += sprintf(p, "     src_mac: ");
+	seq_puts(seq, "     src_mac: ");
 
 	if ((pkt_dev->src_mac[0] == 0) && 
 	    (pkt_dev->src_mac[1] == 0) && 
@@ -648,89 +603,89 @@ static int proc_if_read(char *buf , char **start, off_t offset,
 	    (pkt_dev->src_mac[5] == 0)) 
 
 		for (i = 0; i < 6; i++) 
-			p += sprintf(p, "%02X%s", pkt_dev->odev->dev_addr[i], i == 5 ? "  " : ":");
+			seq_printf(seq,  "%02X%s", pkt_dev->odev->dev_addr[i], i == 5 ? "  " : ":");
 
 	else 
 		for (i = 0; i < 6; i++) 
-			p += sprintf(p, "%02X%s", pkt_dev->src_mac[i], i == 5 ? "  " : ":");
+			seq_printf(seq,  "%02X%s", pkt_dev->src_mac[i], i == 5 ? "  " : ":");
 
-        p += sprintf(p, "dst_mac: ");
+        seq_printf(seq,  "dst_mac: ");
 	for (i = 0; i < 6; i++) 
-		p += sprintf(p, "%02X%s", pkt_dev->dst_mac[i], i == 5 ? "\n" : ":");
+		seq_printf(seq,  "%02X%s", pkt_dev->dst_mac[i], i == 5 ? "\n" : ":");
 
-        p += sprintf(p, "     udp_src_min: %d  udp_src_max: %d  udp_dst_min: %d  udp_dst_max: %d\n",
-                     pkt_dev->udp_src_min, pkt_dev->udp_src_max, pkt_dev->udp_dst_min,
-                     pkt_dev->udp_dst_max);
+        seq_printf(seq,  "     udp_src_min: %d  udp_src_max: %d  udp_dst_min: %d  udp_dst_max: %d\n",
+		   pkt_dev->udp_src_min, pkt_dev->udp_src_max, pkt_dev->udp_dst_min,
+		   pkt_dev->udp_dst_max);
 
-        p += sprintf(p, "     src_mac_count: %d  dst_mac_count: %d \n     Flags: ",
-                     pkt_dev->src_mac_count, pkt_dev->dst_mac_count);
+        seq_printf(seq,  "     src_mac_count: %d  dst_mac_count: %d \n     Flags: ",
+		   pkt_dev->src_mac_count, pkt_dev->dst_mac_count);
 
 
         if (pkt_dev->flags &  F_IPV6) 
-                p += sprintf(p, "IPV6  ");
+                seq_printf(seq,  "IPV6  ");
 
         if (pkt_dev->flags &  F_IPSRC_RND) 
-                p += sprintf(p, "IPSRC_RND  ");
+                seq_printf(seq,  "IPSRC_RND  ");
 
         if (pkt_dev->flags & F_IPDST_RND) 
-                p += sprintf(p, "IPDST_RND  ");
+                seq_printf(seq,  "IPDST_RND  ");
         
         if (pkt_dev->flags & F_TXSIZE_RND) 
-                p += sprintf(p, "TXSIZE_RND  ");
+                seq_printf(seq,  "TXSIZE_RND  ");
         
         if (pkt_dev->flags & F_UDPSRC_RND) 
-                p += sprintf(p, "UDPSRC_RND  ");
+                seq_printf(seq,  "UDPSRC_RND  ");
         
         if (pkt_dev->flags & F_UDPDST_RND) 
-                p += sprintf(p, "UDPDST_RND  ");
+                seq_printf(seq,  "UDPDST_RND  ");
         
         if (pkt_dev->flags & F_MACSRC_RND) 
-                p += sprintf(p, "MACSRC_RND  ");
+                seq_printf(seq,  "MACSRC_RND  ");
         
         if (pkt_dev->flags & F_MACDST_RND) 
-                p += sprintf(p, "MACDST_RND  ");
+                seq_printf(seq,  "MACDST_RND  ");
 
         
-        p += sprintf(p, "\n");
+        seq_puts(seq,  "\n");
         
         sa = pkt_dev->started_at;
         stopped = pkt_dev->stopped_at;
         if (pkt_dev->running) 
                 stopped = now; /* not really stopped, more like last-running-at */
         
-        p += sprintf(p, "Current:\n     pkts-sofar: %llu  errors: %llu\n     started: %lluus  stopped: %lluus idle: %lluus\n",
-		     (unsigned long long) pkt_dev->sofar,
-		     (unsigned long long) pkt_dev->errors,
-		     (unsigned long long) sa,
-		     (unsigned long long) stopped, 
-		     (unsigned long long) pkt_dev->idle_acc);
+        seq_printf(seq,  "Current:\n     pkts-sofar: %llu  errors: %llu\n     started: %lluus  stopped: %lluus idle: %lluus\n",
+		   (unsigned long long) pkt_dev->sofar,
+		   (unsigned long long) pkt_dev->errors,
+		   (unsigned long long) sa,
+		   (unsigned long long) stopped,
+		   (unsigned long long) pkt_dev->idle_acc);
 
-        p += sprintf(p, "     seq_num: %d  cur_dst_mac_offset: %d  cur_src_mac_offset: %d\n",
-                     pkt_dev->seq_num, pkt_dev->cur_dst_mac_offset, pkt_dev->cur_src_mac_offset);
+        seq_printf(seq,  "     seq_num: %d  cur_dst_mac_offset: %d  cur_src_mac_offset: %d\n",
+		   pkt_dev->seq_num, pkt_dev->cur_dst_mac_offset,
+		   pkt_dev->cur_src_mac_offset);
 
 	if(pkt_dev->flags & F_IPV6) {
 		char b1[128], b2[128];
 		fmt_ip6(b1,  pkt_dev->cur_in6_daddr.s6_addr);
 		fmt_ip6(b2,  pkt_dev->cur_in6_saddr.s6_addr);
-		p += sprintf(p, "     cur_saddr: %s  cur_daddr: %s\n", b2, b1);
+		seq_printf(seq,  "     cur_saddr: %s  cur_daddr: %s\n", b2, b1);
 	} 
 	else 
-		p += sprintf(p, "     cur_saddr: 0x%x  cur_daddr: 0x%x\n",
-                     pkt_dev->cur_saddr, pkt_dev->cur_daddr);
+		seq_printf(seq,  "     cur_saddr: 0x%x  cur_daddr: 0x%x\n",
+			   pkt_dev->cur_saddr, pkt_dev->cur_daddr);
 
 
-	p += sprintf(p, "     cur_udp_dst: %d  cur_udp_src: %d\n",
-                     pkt_dev->cur_udp_dst, pkt_dev->cur_udp_src);
+	seq_printf(seq,  "     cur_udp_dst: %d  cur_udp_src: %d\n",
+		   pkt_dev->cur_udp_dst, pkt_dev->cur_udp_src);
 
-	p += sprintf(p, "     flows: %u\n", pkt_dev->nflows);
+	seq_printf(seq,  "     flows: %u\n", pkt_dev->nflows);
 
 	if (pkt_dev->result[0])
-		p += sprintf(p, "Result: %s\n", pkt_dev->result);
+		seq_printf(seq,  "Result: %s\n", pkt_dev->result);
 	else
-		p += sprintf(p, "Result: Idle\n");
-	*eof = 1;
+		seq_printf(seq,  "Result: Idle\n");
 
-	return p - buf;
+	return 0;
 }
 
 
@@ -802,13 +757,14 @@ done_str:
 	return i;
 }
 
-static int proc_if_write(struct file *file, const char __user *user_buffer,
-                            unsigned long count, void *data)
+static ssize_t pktgen_if_write(struct file *file, const char __user *user_buffer,
+			       size_t count, loff_t *offset)
 {
+	struct seq_file *seq = (struct seq_file *) file->private_data;
+        struct pktgen_dev *pkt_dev = seq->private;
 	int i = 0, max, len;
 	char name[16], valstr[32];
 	unsigned long value = 0;
-        struct pktgen_dev *pkt_dev = (struct pktgen_dev*)(data);
         char* pg_result = NULL;
         int tmp = 0;
 	char buf[128];
@@ -849,7 +805,8 @@ static int proc_if_write(struct file *file, const char __user *user_buffer,
                 if (copy_from_user(tb, user_buffer, count))
 			return -EFAULT;
                 tb[count] = 0;
-		printk("pktgen: %s,%lu  buffer -:%s:-\n", name, count, tb);
+		printk("pktgen: %s,%lu  buffer -:%s:-\n", name,
+		       (unsigned long) count, tb);
         }
 
 	if (!strcmp(name, "min_pkt_size")) {
@@ -1335,92 +1292,98 @@ static int proc_if_write(struct file *file, const char __user *user_buffer,
 	return -EINVAL;
 }
 
-static int proc_thread_read(char *buf , char **start, off_t offset,
-                               int len, int *eof, void *data)
+static int pktgen_if_open(struct inode *inode, struct file *file)
 {
-	char *p;
-        struct pktgen_thread *t = (struct pktgen_thread*)(data);
-        struct pktgen_dev *pkt_dev = NULL;
+	return single_open(file, pktgen_if_show, PDE(inode)->data);
+}
 
+static struct file_operations pktgen_if_fops = {
+	.owner	  = THIS_MODULE,
+	.open	  = pktgen_if_open,
+        .read     = seq_read,
+	.llseek	  = seq_lseek,
+        .write    = pktgen_if_write,
+	.release  = single_release,
+};
 
-        if (!t) {
-                printk("pktgen: ERROR: could not find thread in proc_thread_read\n");
-                return -EINVAL;
-        }
+static int pktgen_thread_show(struct seq_file *seq, void *v)
+{
+        struct pktgen_thread *t = seq->private;
+        struct pktgen_dev *pkt_dev = NULL;
 
-	p = buf;
-	p += sprintf(p, "Name: %s  max_before_softirq: %d\n",
+	BUG_ON(!t);
+
+	seq_printf(seq, "Name: %s  max_before_softirq: %d\n",
                      t->name, t->max_before_softirq);
 
-        p += sprintf(p, "Running: ");
+        seq_printf(seq, "Running: ");
         
         if_lock(t);
         for(pkt_dev = t->if_list;pkt_dev; pkt_dev = pkt_dev->next) 
 		if(pkt_dev->running)
-			p += sprintf(p, "%s ", pkt_dev->ifname);
+			seq_printf(seq, "%s ", pkt_dev->ifname);
         
-        p += sprintf(p, "\nStopped: ");
+        seq_printf(seq, "\nStopped: ");
 
         for(pkt_dev = t->if_list;pkt_dev; pkt_dev = pkt_dev->next) 
 		if(!pkt_dev->running)
-			p += sprintf(p, "%s ", pkt_dev->ifname);
+			seq_printf(seq, "%s ", pkt_dev->ifname);
 
 	if (t->result[0])
-		p += sprintf(p, "\nResult: %s\n", t->result);
+		seq_printf(seq, "\nResult: %s\n", t->result);
 	else
-		p += sprintf(p, "\nResult: NA\n");
-
-	*eof = 1;
+		seq_printf(seq, "\nResult: NA\n");
 
         if_unlock(t);
 
-	return p - buf;
+	return 0;
 }
 
-static int proc_thread_write(struct file *file, const char __user *user_buffer,
-                                unsigned long count, void *data)
+static ssize_t pktgen_thread_write(struct file *file,
+				   const char __user *user_buffer,
+				   size_t count, loff_t *offset)
 {
+	struct seq_file *seq = (struct seq_file *) file->private_data;
+        struct pktgen_thread *t = seq->private;
 	int i = 0, max, len, ret;
 	char name[40];
-        struct pktgen_thread *t;
         char *pg_result;
         unsigned long value = 0;
-        
+
 	if (count < 1) {
 		//	sprintf(pg_result, "Wrong command format");
 		return -EINVAL;
 	}
-  
+
 	max = count - i;
         len = count_trail_chars(&user_buffer[i], max);
-        if (len < 0) 
-		return len; 
-     
+        if (len < 0)
+		return len;
+
 	i += len;
-  
+
 	/* Read variable name */
 
 	len = strn_len(&user_buffer[i], sizeof(name) - 1);
-        if (len < 0)  
-		return len; 
+        if (len < 0)
+		return len;
 	
 	memset(name, 0, sizeof(name));
 	if (copy_from_user(name, &user_buffer[i], len))
 		return -EFAULT;
 	i += len;
-  
+
 	max = count -i;
 	len = count_trail_chars(&user_buffer[i], max);
-        if (len < 0)  
-		return len; 
-	
+        if (len < 0)
+		return len;
+
 	i += len;
 
-	if (debug) 
-		printk("pktgen: t=%s, count=%lu\n", name, count);
-        
+	if (debug)
+		printk("pktgen: t=%s, count=%lu\n", name,
+		       (unsigned long) count);
 
-        t = (struct pktgen_thread*)(data);
 	if(!t) {
 		printk("pktgen: ERROR: No thread\n");
 		ret = -EINVAL;
@@ -1474,32 +1437,19 @@ static int proc_thread_write(struct file *file, const char __user *user_buffer,
 	return ret;
 }
 
-static int create_proc_dir(void)
+static int pktgen_thread_open(struct inode *inode, struct file *file)
 {
-        int     len;
-        /*  does proc_dir already exists */
-        len = strlen(PG_PROC_DIR);
-
-        for (pg_proc_dir = proc_net->subdir; pg_proc_dir; pg_proc_dir=pg_proc_dir->next) {
-                if ((pg_proc_dir->namelen == len) &&
-		    (! memcmp(pg_proc_dir->name, PG_PROC_DIR, len))) 
-                        break;
-        }
-        
-        if (!pg_proc_dir) 
-                pg_proc_dir = create_proc_entry(PG_PROC_DIR, S_IFDIR, proc_net);
-        
-        if (!pg_proc_dir) 
-                return -ENODEV;
-        
-        return 0;
+	return single_open(file, pktgen_thread_show, PDE(inode)->data);
 }
 
-static int remove_proc_dir(void)
-{
-        remove_proc_entry(PG_PROC_DIR, proc_net);
-        return 0;
-}
+static struct file_operations pktgen_thread_fops = {
+	.owner	  = THIS_MODULE,
+	.open	  = pktgen_thread_open,
+        .read     = seq_read,
+	.llseek	  = seq_lseek,
+        .write    = pktgen_thread_write,
+	.release  = single_release,
+};
 
 /* Think find or remove for NN */
 static struct pktgen_dev *__pktgen_NN_threads(const char* ifname, int remove) 
@@ -1678,13 +1628,12 @@ static void pktgen_setup_inject(struct pktgen_dev *pkt_dev)
 			struct in_device *in_dev; 
 
 			rcu_read_lock();
-			in_dev = __in_dev_get(pkt_dev->odev);
+			in_dev = __in_dev_get_rcu(pkt_dev->odev);
 			if (in_dev) {
 				if (in_dev->ifa_list) {
 					pkt_dev->saddr_min = in_dev->ifa_list->ifa_address;
 					pkt_dev->saddr_max = pkt_dev->saddr_min;
 				}
-				__in_dev_put(in_dev);	
 			}
 			rcu_read_unlock();
 		}
@@ -1714,7 +1663,7 @@ static void spin(struct pktgen_dev *pkt_dev, __u64 spin_until_us)
 	start = now = getCurUs();
 	printk(KERN_INFO "sleeping for %d\n", (int)(spin_until_us - now));
 	while (now < spin_until_us) {
-		/* TODO: optimise sleeping behavior */
+		/* TODO: optimize sleeping behavior */
 		if (spin_until_us - now > jiffies_to_usecs(1)+1)
 			schedule_timeout_interruptible(1);
 		else if (spin_until_us - now > 100) {
@@ -2373,7 +2322,7 @@ static void pktgen_stop_all_threads_ifs(void)
 		pktgen_stop(t);
 		t = t->next;
 	}
-       thread_unlock();
+	thread_unlock();
 }
 
 static int thread_is_running(struct pktgen_thread *t )
@@ -2564,10 +2513,9 @@ static void pktgen_rem_thread(struct pktgen_thread *t)
 
 	struct pktgen_thread *tmp = pktgen_threads;
 
-        if (strlen(t->fname))
-                remove_proc_entry(t->fname, NULL);
+	remove_proc_entry(t->name, pg_proc_dir);
 
-       thread_lock();
+	thread_lock();
 
 	if (tmp == t)
 		pktgen_threads = tmp->next;
@@ -2837,7 +2785,7 @@ static struct pktgen_dev *pktgen_find_dev(struct pktgen_thread *t, const char* i
         if_lock(t);
 
         for(pkt_dev=t->if_list; pkt_dev; pkt_dev = pkt_dev->next ) {
-                if (strcmp(pkt_dev->ifname, ifname) == 0) {
+                if (strncmp(pkt_dev->ifname, ifname, IFNAMSIZ) == 0) {
                         break;
                 }
         }
@@ -2876,74 +2824,70 @@ static int add_dev_to_thread(struct pktgen_thread *t, struct pktgen_dev *pkt_dev
 static int pktgen_add_device(struct pktgen_thread *t, const char* ifname) 
 {
         struct pktgen_dev *pkt_dev;
+	struct proc_dir_entry *pe;
 	
 	/* We don't allow a device to be on several threads */
 
-	if( (pkt_dev = __pktgen_NN_threads(ifname, FIND)) == NULL) {
-						   
-		pkt_dev = kmalloc(sizeof(struct pktgen_dev), GFP_KERNEL);
-                if (!pkt_dev) 
-                        return -ENOMEM;
+	pkt_dev = __pktgen_NN_threads(ifname, FIND);
+	if (pkt_dev) {
+                printk("pktgen: ERROR: interface already used.\n");
+                return -EBUSY;
+        }
 
-                memset(pkt_dev, 0, sizeof(struct pktgen_dev));
+	pkt_dev = kzalloc(sizeof(struct pktgen_dev), GFP_KERNEL);
+	if (!pkt_dev)
+		return -ENOMEM;
 
-		pkt_dev->flows = vmalloc(MAX_CFLOWS*sizeof(struct flow_state));
-		if (pkt_dev->flows == NULL) {
-			kfree(pkt_dev);
-			return -ENOMEM;
-		}
-		memset(pkt_dev->flows, 0, MAX_CFLOWS*sizeof(struct flow_state));
-
-		pkt_dev->min_pkt_size = ETH_ZLEN;
-                pkt_dev->max_pkt_size = ETH_ZLEN;
-                pkt_dev->nfrags = 0;
-                pkt_dev->clone_skb = pg_clone_skb_d;
-                pkt_dev->delay_us = pg_delay_d / 1000;
-                pkt_dev->delay_ns = pg_delay_d % 1000;
-                pkt_dev->count = pg_count_d;
-                pkt_dev->sofar = 0;
-                pkt_dev->udp_src_min = 9; /* sink port */
-                pkt_dev->udp_src_max = 9;
-                pkt_dev->udp_dst_min = 9;
-                pkt_dev->udp_dst_max = 9;
-
-                strncpy(pkt_dev->ifname, ifname, 31);
-                sprintf(pkt_dev->fname, "net/%s/%s", PG_PROC_DIR, ifname);
-
-                if (! pktgen_setup_dev(pkt_dev)) {
-                        printk("pktgen: ERROR: pktgen_setup_dev failed.\n");
-			if (pkt_dev->flows)
-				vfree(pkt_dev->flows);
-                        kfree(pkt_dev);
-                        return -ENODEV;
-                }
+	pkt_dev->flows = vmalloc(MAX_CFLOWS*sizeof(struct flow_state));
+	if (pkt_dev->flows == NULL) {
+		kfree(pkt_dev);
+		return -ENOMEM;
+	}
+	memset(pkt_dev->flows, 0, MAX_CFLOWS*sizeof(struct flow_state));
 
-                pkt_dev->proc_ent = create_proc_entry(pkt_dev->fname, 0600, NULL);
-                if (!pkt_dev->proc_ent) {
-                        printk("pktgen: cannot create %s procfs entry.\n", pkt_dev->fname);
-			if (pkt_dev->flows)
-				vfree(pkt_dev->flows);
-                        kfree(pkt_dev);
-                        return -EINVAL;
-                }
-                pkt_dev->proc_ent->read_proc = proc_if_read;
-                pkt_dev->proc_ent->write_proc = proc_if_write;
-                pkt_dev->proc_ent->data = (void*)(pkt_dev);
-		pkt_dev->proc_ent->owner = THIS_MODULE;
+	pkt_dev->min_pkt_size = ETH_ZLEN;
+	pkt_dev->max_pkt_size = ETH_ZLEN;
+	pkt_dev->nfrags = 0;
+	pkt_dev->clone_skb = pg_clone_skb_d;
+	pkt_dev->delay_us = pg_delay_d / 1000;
+	pkt_dev->delay_ns = pg_delay_d % 1000;
+	pkt_dev->count = pg_count_d;
+	pkt_dev->sofar = 0;
+	pkt_dev->udp_src_min = 9; /* sink port */
+	pkt_dev->udp_src_max = 9;
+	pkt_dev->udp_dst_min = 9;
+	pkt_dev->udp_dst_max = 9;
+
+	strncpy(pkt_dev->ifname, ifname, IFNAMSIZ);
+
+	if (! pktgen_setup_dev(pkt_dev)) {
+		printk("pktgen: ERROR: pktgen_setup_dev failed.\n");
+		if (pkt_dev->flows)
+			vfree(pkt_dev->flows);
+		kfree(pkt_dev);
+		return -ENODEV;
+	}
+
+	pe = create_proc_entry(ifname, 0600, pg_proc_dir);
+	if (!pe) {
+		printk("pktgen: cannot create %s/%s procfs entry.\n",
+		       PG_PROC_DIR, ifname);
+		if (pkt_dev->flows)
+			vfree(pkt_dev->flows);
+		kfree(pkt_dev);
+		return -EINVAL;
+	}
+	pe->proc_fops = &pktgen_if_fops;
+	pe->data = pkt_dev;
 
-                return add_dev_to_thread(t, pkt_dev);
-        }
-        else {
-                printk("pktgen: ERROR: interface already used.\n");
-                return -EBUSY;
-        }
+	return add_dev_to_thread(t, pkt_dev);
 }
 
 static struct pktgen_thread *pktgen_find_thread(const char* name) 
 {
         struct pktgen_thread *t = NULL;
 
-       thread_lock();
+	thread_lock();
 
         t = pktgen_threads;
         while (t) {
@@ -2959,6 +2903,7 @@ static struct pktgen_thread *pktgen_find_thread(const char* name)
 static int pktgen_create_thread(const char* name, int cpu) 
 {
         struct pktgen_thread *t = NULL;
+	struct proc_dir_entry *pe;
 
         if (strlen(name) > 31) {
                 printk("pktgen: ERROR:  Thread name cannot be more than 31 characters.\n");
@@ -2970,28 +2915,26 @@ static int pktgen_create_thread(const char* name, int cpu)
                 return -EINVAL;
         }
 
-        t = (struct pktgen_thread*)(kmalloc(sizeof(struct pktgen_thread), GFP_KERNEL));
+        t = kzalloc(sizeof(struct pktgen_thread), GFP_KERNEL);
         if (!t) {
                 printk("pktgen: ERROR: out of memory, can't create new thread.\n");
                 return -ENOMEM;
         }
 
-        memset(t, 0, sizeof(struct pktgen_thread));
         strcpy(t->name, name);
         spin_lock_init(&t->if_lock);
 	t->cpu = cpu;
         
-        sprintf(t->fname, "net/%s/%s", PG_PROC_DIR, t->name);
-        t->proc_ent = create_proc_entry(t->fname, 0600, NULL);
-        if (!t->proc_ent) {
-                printk("pktgen: cannot create %s procfs entry.\n", t->fname);
+        pe = create_proc_entry(t->name, 0600, pg_proc_dir);
+        if (!pe) {
+                printk("pktgen: cannot create %s/%s procfs entry.\n",
+		       PG_PROC_DIR, t->name);
                 kfree(t);
                 return -EINVAL;
         }
-        t->proc_ent->read_proc = proc_thread_read;
-        t->proc_ent->write_proc = proc_thread_write;
-        t->proc_ent->data = (void*)(t);
-        t->proc_ent->owner = THIS_MODULE;
+
+	pe->proc_fops = &pktgen_thread_fops;
+	pe->data = t;
 
         t->next = pktgen_threads;
         pktgen_threads = t;
@@ -3046,8 +2989,7 @@ static int pktgen_remove_device(struct pktgen_thread *t, struct pktgen_dev *pkt_
 
         /* Clean up proc file system */
 
-        if (strlen(pkt_dev->fname)) 
-                remove_proc_entry(pkt_dev->fname, NULL);
+	remove_proc_entry(pkt_dev->ifname, pg_proc_dir);
 
 	if (pkt_dev->flows)
 		vfree(pkt_dev->flows);
@@ -3058,31 +3000,31 @@ static int pktgen_remove_device(struct pktgen_thread *t, struct pktgen_dev *pkt_
 static int __init pg_init(void) 
 {
 	int cpu;
-	printk(version);
+	struct proc_dir_entry *pe;
 
-        module_fname[0] = 0;
+	printk(version);
 
-	create_proc_dir();
+	pg_proc_dir = proc_mkdir(PG_PROC_DIR, proc_net);
+	if (!pg_proc_dir)
+		return -ENODEV;
+	pg_proc_dir->owner = THIS_MODULE;
 
-        sprintf(module_fname, "net/%s/pgctrl", PG_PROC_DIR);
-        module_proc_ent = create_proc_entry(module_fname, 0600, NULL);
-        if (!module_proc_ent) {
-                printk("pktgen: ERROR: cannot create %s procfs entry.\n", module_fname);
+	pe = create_proc_entry(PGCTRL, 0600, pg_proc_dir);
+        if (pe == NULL) {
+		printk("pktgen: ERROR: cannot create %s procfs entry.\n", PGCTRL);
+		proc_net_remove(PG_PROC_DIR);
                 return -EINVAL;
         }
 
-        module_proc_ent->proc_fops =  &pktgen_fops;
-        module_proc_ent->data = NULL;
+        pe->proc_fops = &pktgen_fops;
+        pe->data      = NULL;
 
 	/* Register us to receive netdevice events */
 	register_netdevice_notifier(&pktgen_notifier_block);
         
-	for (cpu = 0; cpu < NR_CPUS ; cpu++) {
+	for_each_online_cpu(cpu) {
 		char buf[30];
 
-		if (!cpu_online(cpu))
-			continue;
-
                 sprintf(buf, "kpktgend_%i", cpu);
                 pktgen_create_thread(buf, cpu);
         }
@@ -3107,10 +3049,8 @@ static void __exit pg_cleanup(void)
 	unregister_netdevice_notifier(&pktgen_notifier_block);
 
         /* Clean up proc file system */
-
-        remove_proc_entry(module_fname, NULL);
-        
-	remove_proc_dir();
+	remove_proc_entry(PGCTRL, pg_proc_dir);
+	proc_net_remove(PG_PROC_DIR);
 }
 
 
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index f80a2878561..ef9d46b91eb 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -71,8 +71,6 @@
 static kmem_cache_t *skbuff_head_cache __read_mostly;
 static kmem_cache_t *skbuff_fclone_cache __read_mostly;
 
-struct timeval __read_mostly skb_tv_base;
-
 /*
  *	Keep out-of-line to prevent kernel bloat.
  *	__builtin_return_address is not used because it is not always
@@ -124,6 +122,8 @@ void skb_under_panic(struct sk_buff *skb, int sz, void *here)
  *	__alloc_skb	-	allocate a network buffer
  *	@size: size to allocate
  *	@gfp_mask: allocation mask
+ *	@fclone: allocate from fclone cache instead of head cache
+ *		and allocate a cloned (child) skb
  *
  *	Allocate a new &sk_buff. The returned buffer has no headroom and a
  *	tail room of size bytes. The object has a reference count of one.
@@ -132,7 +132,7 @@ void skb_under_panic(struct sk_buff *skb, int sz, void *here)
  *	Buffers may only be allocated from interrupts using a @gfp_mask of
  *	%GFP_ATOMIC.
  */
-struct sk_buff *__alloc_skb(unsigned int size, unsigned int __nocast gfp_mask,
+struct sk_buff *__alloc_skb(unsigned int size, gfp_t gfp_mask,
 			    int fclone)
 {
 	struct sk_buff *skb;
@@ -200,7 +200,7 @@ nodata:
  */
 struct sk_buff *alloc_skb_from_cache(kmem_cache_t *cp,
 				     unsigned int size,
-				     unsigned int __nocast gfp_mask)
+				     gfp_t gfp_mask)
 {
 	struct sk_buff *skb;
 	u8 *data;
@@ -363,7 +363,7 @@ void __kfree_skb(struct sk_buff *skb)
  *	%GFP_ATOMIC.
  */
 
-struct sk_buff *skb_clone(struct sk_buff *skb, unsigned int __nocast gfp_mask)
+struct sk_buff *skb_clone(struct sk_buff *skb, gfp_t gfp_mask)
 {
 	struct sk_buff *n;
 
@@ -412,6 +412,9 @@ struct sk_buff *skb_clone(struct sk_buff *skb, unsigned int __nocast gfp_mask)
 	C(nfct);
 	nf_conntrack_get(skb->nfct);
 	C(nfctinfo);
+#if defined(CONFIG_IP_VS) || defined(CONFIG_IP_VS_MODULE)
+	C(ipvs_property);
+#endif
 #ifdef CONFIG_BRIDGE_NETFILTER
 	C(nf_bridge);
 	nf_bridge_get(skb->nf_bridge);
@@ -469,6 +472,9 @@ static void copy_skb_header(struct sk_buff *new, const struct sk_buff *old)
 	new->nfct	= old->nfct;
 	nf_conntrack_get(old->nfct);
 	new->nfctinfo	= old->nfctinfo;
+#if defined(CONFIG_IP_VS) || defined(CONFIG_IP_VS_MODULE)
+	new->ipvs_property = old->ipvs_property;
+#endif
 #ifdef CONFIG_BRIDGE_NETFILTER
 	new->nf_bridge	= old->nf_bridge;
 	nf_bridge_get(old->nf_bridge);
@@ -502,7 +508,7 @@ static void copy_skb_header(struct sk_buff *new, const struct sk_buff *old)
  *	header is going to be modified. Use pskb_copy() instead.
  */
 
-struct sk_buff *skb_copy(const struct sk_buff *skb, unsigned int __nocast gfp_mask)
+struct sk_buff *skb_copy(const struct sk_buff *skb, gfp_t gfp_mask)
 {
 	int headerlen = skb->data - skb->head;
 	/*
@@ -541,7 +547,7 @@ struct sk_buff *skb_copy(const struct sk_buff *skb, unsigned int __nocast gfp_ma
  *	The returned buffer has a reference count of 1.
  */
 
-struct sk_buff *pskb_copy(struct sk_buff *skb, unsigned int __nocast gfp_mask)
+struct sk_buff *pskb_copy(struct sk_buff *skb, gfp_t gfp_mask)
 {
 	/*
 	 *	Allocate the copy buffer
@@ -600,7 +606,7 @@ out:
  */
 
 int pskb_expand_head(struct sk_buff *skb, int nhead, int ntail,
-		     unsigned int __nocast gfp_mask)
+		     gfp_t gfp_mask)
 {
 	int i;
 	u8 *data;
@@ -691,7 +697,7 @@ struct sk_buff *skb_realloc_headroom(struct sk_buff *skb, unsigned int headroom)
  */
 struct sk_buff *skb_copy_expand(const struct sk_buff *skb,
 				int newheadroom, int newtailroom,
-				unsigned int __nocast gfp_mask)
+				gfp_t gfp_mask)
 {
 	/*
 	 *	Allocate the copy buffer
@@ -1708,8 +1714,6 @@ void __init skb_init(void)
 						NULL, NULL);
 	if (!skbuff_fclone_cache)
 		panic("cannot create skbuff cache");
-
-	do_gettimeofday(&skb_tv_base);
 }
 
 EXPORT_SYMBOL(___pskb_trim);
@@ -1743,4 +1747,3 @@ EXPORT_SYMBOL(skb_prepare_seq_read);
 EXPORT_SYMBOL(skb_seq_read);
 EXPORT_SYMBOL(skb_abort_seq_read);
 EXPORT_SYMBOL(skb_find_text);
-EXPORT_SYMBOL(skb_tv_base);
diff --git a/net/core/sock.c b/net/core/sock.c
index ac63b56e23b..9602ceb3bac 100644
--- a/net/core/sock.c
+++ b/net/core/sock.c
@@ -637,7 +637,7 @@ lenout:
  *	@prot: struct proto associated with this new sock instance
  *	@zero_it: if we should zero the newly allocated sock
  */
-struct sock *sk_alloc(int family, unsigned int __nocast priority,
+struct sock *sk_alloc(int family, gfp_t priority,
 		      struct proto *prot, int zero_it)
 {
 	struct sock *sk = NULL;
@@ -660,16 +660,20 @@ struct sock *sk_alloc(int family, unsigned int __nocast priority,
 			sock_lock_init(sk);
 		}
 		
-		if (security_sk_alloc(sk, family, priority)) {
-			if (slab != NULL)
-				kmem_cache_free(slab, sk);
-			else
-				kfree(sk);
-			sk = NULL;
-		} else
-			__module_get(prot->owner);
+		if (security_sk_alloc(sk, family, priority))
+			goto out_free;
+
+		if (!try_module_get(prot->owner))
+			goto out_free;
 	}
 	return sk;
+
+out_free:
+	if (slab != NULL)
+		kmem_cache_free(slab, sk);
+	else
+		kfree(sk);
+	return NULL;
 }
 
 void sk_free(struct sock *sk)
@@ -700,7 +704,7 @@ void sk_free(struct sock *sk)
 	module_put(owner);
 }
 
-struct sock *sk_clone(const struct sock *sk, const unsigned int __nocast priority)
+struct sock *sk_clone(const struct sock *sk, const gfp_t priority)
 {
 	struct sock *newsk = sk_alloc(sk->sk_family, priority, sk->sk_prot, 0);
 
@@ -841,7 +845,7 @@ unsigned long sock_i_ino(struct sock *sk)
  * Allocate a skb from the socket's send buffer.
  */
 struct sk_buff *sock_wmalloc(struct sock *sk, unsigned long size, int force,
-			     unsigned int __nocast priority)
+			     gfp_t priority)
 {
 	if (force || atomic_read(&sk->sk_wmem_alloc) < sk->sk_sndbuf) {
 		struct sk_buff * skb = alloc_skb(size, priority);
@@ -857,7 +861,7 @@ struct sk_buff *sock_wmalloc(struct sock *sk, unsigned long size, int force,
  * Allocate a skb from the socket's receive buffer.
  */ 
 struct sk_buff *sock_rmalloc(struct sock *sk, unsigned long size, int force,
-			     unsigned int __nocast priority)
+			     gfp_t priority)
 {
 	if (force || atomic_read(&sk->sk_rmem_alloc) < sk->sk_rcvbuf) {
 		struct sk_buff *skb = alloc_skb(size, priority);
@@ -872,7 +876,7 @@ struct sk_buff *sock_rmalloc(struct sock *sk, unsigned long size, int force,
 /* 
  * Allocate a memory block from the socket's option memory buffer.
  */ 
-void *sock_kmalloc(struct sock *sk, int size, unsigned int __nocast priority)
+void *sock_kmalloc(struct sock *sk, int size, gfp_t priority)
 {
 	if ((unsigned)size <= sysctl_optmem_max &&
 	    atomic_read(&sk->sk_omem_alloc) + size < sysctl_optmem_max) {
@@ -936,7 +940,7 @@ static struct sk_buff *sock_alloc_send_pskb(struct sock *sk,
 					    int noblock, int *errcode)
 {
 	struct sk_buff *skb;
-	unsigned int gfp_mask;
+	gfp_t gfp_mask;
 	long timeo;
 	int err;
 
diff --git a/net/core/wireless.c b/net/core/wireless.c
index d17f1583ea3..271ddb35b0b 100644
--- a/net/core/wireless.c
+++ b/net/core/wireless.c
@@ -455,10 +455,15 @@ static inline struct iw_statistics *get_wireless_stats(struct net_device *dev)
 
 	/* Old location, field to be removed in next WE */
 	if(dev->get_wireless_stats) {
-		printk(KERN_DEBUG "%s (WE) : Driver using old /proc/net/wireless support, please fix driver !\n",
-		       dev->name);
+		static int printed_message;
+
+		if (!printed_message++)
+			printk(KERN_DEBUG "%s (WE) : Driver using old /proc/net/wireless support, please fix driver !\n",
+				dev->name);
+
 		return dev->get_wireless_stats(dev);
 	}
+
 	/* Not found */
 	return (struct iw_statistics *) NULL;
 }
diff --git a/net/dccp/ackvec.c b/net/dccp/ackvec.c
index 6530283eafc..c9a62cca22f 100644
--- a/net/dccp/ackvec.c
+++ b/net/dccp/ackvec.c
@@ -91,7 +91,7 @@ int dccp_insert_option_ackvec(struct sock *sk, struct sk_buff *skb)
 }
 
 struct dccp_ackvec *dccp_ackvec_alloc(const unsigned int len,
-				      const unsigned int __nocast priority)
+				      const gfp_t priority)
 {
 	struct dccp_ackvec *av = kmalloc(sizeof(*av) + len, priority);
 
diff --git a/net/dccp/ackvec.h b/net/dccp/ackvec.h
index 8ca51c9191f..d0fd6c60c57 100644
--- a/net/dccp/ackvec.h
+++ b/net/dccp/ackvec.h
@@ -74,7 +74,7 @@ struct sk_buff;
 
 #ifdef CONFIG_IP_DCCP_ACKVEC
 extern struct dccp_ackvec *dccp_ackvec_alloc(unsigned int len,
-					  const unsigned int __nocast priority);
+					  const gfp_t priority);
 extern void dccp_ackvec_free(struct dccp_ackvec *av);
 
 extern int dccp_ackvec_add(struct dccp_ackvec *av, const struct sock *sk,
@@ -93,7 +93,7 @@ static inline int dccp_ackvec_pending(const struct dccp_ackvec *av)
 }
 #else /* CONFIG_IP_DCCP_ACKVEC */
 static inline struct dccp_ackvec *dccp_ackvec_alloc(unsigned int len,
-					   const unsigned int __nocast priority)
+					   const gfp_t priority)
 {
 	return NULL;
 }
diff --git a/net/dccp/ccid.h b/net/dccp/ccid.h
index 21e55142dcd..c37eeeaf5c6 100644
--- a/net/dccp/ccid.h
+++ b/net/dccp/ccid.h
@@ -110,14 +110,14 @@ static inline int ccid_hc_tx_init(struct ccid *ccid, struct sock *sk)
 
 static inline void ccid_hc_rx_exit(struct ccid *ccid, struct sock *sk)
 {
-	if (ccid->ccid_hc_rx_exit != NULL &&
+	if (ccid != NULL && ccid->ccid_hc_rx_exit != NULL &&
 	    dccp_sk(sk)->dccps_hc_rx_ccid_private != NULL)
 		ccid->ccid_hc_rx_exit(sk);
 }
 
 static inline void ccid_hc_tx_exit(struct ccid *ccid, struct sock *sk)
 {
-	if (ccid->ccid_hc_tx_exit != NULL &&
+	if (ccid != NULL && ccid->ccid_hc_tx_exit != NULL &&
 	    dccp_sk(sk)->dccps_hc_tx_ccid_private != NULL)
 		ccid->ccid_hc_tx_exit(sk);
 }
diff --git a/net/dccp/ccids/lib/loss_interval.h b/net/dccp/ccids/lib/loss_interval.h
index 13ad47ba142..417d9d82df3 100644
--- a/net/dccp/ccids/lib/loss_interval.h
+++ b/net/dccp/ccids/lib/loss_interval.h
@@ -36,7 +36,7 @@ struct dccp_li_hist_entry {
 
 static inline struct dccp_li_hist_entry *
 		dccp_li_hist_entry_new(struct dccp_li_hist *hist,
-				       const unsigned int __nocast prio)
+				       const gfp_t prio)
 {
 	return kmem_cache_alloc(hist->dccplih_slab, prio);
 }
diff --git a/net/dccp/ccids/lib/packet_history.h b/net/dccp/ccids/lib/packet_history.h
index b375ebdb7dc..122e96737ff 100644
--- a/net/dccp/ccids/lib/packet_history.h
+++ b/net/dccp/ccids/lib/packet_history.h
@@ -86,7 +86,7 @@ extern struct dccp_rx_hist_entry *
 
 static inline struct dccp_tx_hist_entry *
 		dccp_tx_hist_entry_new(struct dccp_tx_hist *hist,
-				       const unsigned int __nocast prio)
+				       const gfp_t prio)
 {
 	struct dccp_tx_hist_entry *entry = kmem_cache_alloc(hist->dccptxh_slab,
 							    prio);
@@ -137,7 +137,7 @@ static inline struct dccp_rx_hist_entry *
 				     	    const struct sock *sk, 
 				     	    const u32 ndp, 
 					    const struct sk_buff *skb,
-					    const unsigned int __nocast prio)
+					    const gfp_t prio)
 {
 	struct dccp_rx_hist_entry *entry = kmem_cache_alloc(hist->dccprxh_slab,
 							    prio);
diff --git a/net/dccp/input.c b/net/dccp/input.c
index 1b6b2cb1237..3454d594190 100644
--- a/net/dccp/input.c
+++ b/net/dccp/input.c
@@ -375,6 +375,9 @@ static int dccp_rcv_respond_partopen_state_process(struct sock *sk,
 	case DCCP_PKT_RESET:
 		inet_csk_clear_xmit_timer(sk, ICSK_TIME_DACK);
 		break;
+	case DCCP_PKT_DATA:
+		if (sk->sk_state == DCCP_RESPOND)
+			break;
 	case DCCP_PKT_DATAACK:
 	case DCCP_PKT_ACK:
 		/*
@@ -393,7 +396,8 @@ static int dccp_rcv_respond_partopen_state_process(struct sock *sk,
 		dccp_sk(sk)->dccps_osr = DCCP_SKB_CB(skb)->dccpd_seq;
 		dccp_set_state(sk, DCCP_OPEN);
 
-		if (dh->dccph_type == DCCP_PKT_DATAACK) {
+		if (dh->dccph_type == DCCP_PKT_DATAACK ||
+		    dh->dccph_type == DCCP_PKT_DATA) {
 			dccp_rcv_established(sk, skb, dh, len);
 			queued = 1; /* packet was queued
 				       (by dccp_rcv_established) */
diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c
index 40fe6afacde..6298cf58ff9 100644
--- a/net/dccp/ipv4.c
+++ b/net/dccp/ipv4.c
@@ -62,27 +62,27 @@ static int __dccp_v4_check_established(struct sock *sk, const __u16 lport,
 	const int dif = sk->sk_bound_dev_if;
 	INET_ADDR_COOKIE(acookie, saddr, daddr)
 	const __u32 ports = INET_COMBINED_PORTS(inet->dport, lport);
-	const int hash = inet_ehashfn(daddr, lport, saddr, inet->dport,
-				      dccp_hashinfo.ehash_size);
-	struct inet_ehash_bucket *head = &dccp_hashinfo.ehash[hash];
+	unsigned int hash = inet_ehashfn(daddr, lport, saddr, inet->dport);
+	struct inet_ehash_bucket *head = inet_ehash_bucket(&dccp_hashinfo, hash);
 	const struct sock *sk2;
 	const struct hlist_node *node;
 	struct inet_timewait_sock *tw;
 
+	prefetch(head->chain.first);
 	write_lock(&head->lock);
 
 	/* Check TIME-WAIT sockets first. */
 	sk_for_each(sk2, node, &(head + dccp_hashinfo.ehash_size)->chain) {
 		tw = inet_twsk(sk2);
 
-		if (INET_TW_MATCH(sk2, acookie, saddr, daddr, ports, dif))
+		if (INET_TW_MATCH(sk2, hash, acookie, saddr, daddr, ports, dif))
 			goto not_unique;
 	}
 	tw = NULL;
 
 	/* And established part... */
 	sk_for_each(sk2, node, &head->chain) {
-		if (INET_MATCH(sk2, acookie, saddr, daddr, ports, dif))
+		if (INET_MATCH(sk2, hash, acookie, saddr, daddr, ports, dif))
 			goto not_unique;
 	}
 
@@ -90,7 +90,7 @@ static int __dccp_v4_check_established(struct sock *sk, const __u16 lport,
 	 * in hash table socket with a funny identity. */
 	inet->num = lport;
 	inet->sport = htons(lport);
-	sk->sk_hashent = hash;
+	sk->sk_hash = hash;
 	BUG_TRAP(sk_unhashed(sk));
 	__sk_add_node(sk, &head->chain);
 	sock_prot_inc_use(sk->sk_prot);
@@ -463,6 +463,7 @@ static int dccp_v4_send_response(struct sock *sk, struct request_sock *req,
 	if (skb != NULL) {
 		const struct inet_request_sock *ireq = inet_rsk(req);
 
+		memset(&(IPCB(skb)->opt), 0, sizeof(IPCB(skb)->opt));
 		err = ip_build_and_send_pkt(skb, sk, ireq->loc_addr,
 					    ireq->rmt_addr,
 					    ireq->opt);
@@ -647,6 +648,7 @@ int dccp_v4_send_reset(struct sock *sk, enum dccp_reset_codes code)
 	if (skb != NULL) {
 		const struct inet_sock *inet = inet_sk(sk);
 
+		memset(&(IPCB(skb)->opt), 0, sizeof(IPCB(skb)->opt));
 		err = ip_build_and_send_pkt(skb, sk,
 					    inet->saddr, inet->daddr, NULL);
 		if (err == NET_XMIT_CN)
diff --git a/net/dccp/output.c b/net/dccp/output.c
index 4786bdcddcc..d59f86f7cea 100644
--- a/net/dccp/output.c
+++ b/net/dccp/output.c
@@ -62,10 +62,8 @@ int dccp_transmit_skb(struct sock *sk, struct sk_buff *skb)
 		
 		skb->h.raw = skb_push(skb, dccp_header_size);
 		dh = dccp_hdr(skb);
-		/*
-		 * Data packets are not cloned as they are never retransmitted
-		 */
-		if (skb_cloned(skb))
+
+		if (!skb->sk)
 			skb_set_owner_w(skb, sk);
 
 		/* Build DCCP header and checksum it. */
@@ -102,6 +100,7 @@ int dccp_transmit_skb(struct sock *sk, struct sk_buff *skb)
 
 		DCCP_INC_STATS(DCCP_MIB_OUTSEGS);
 
+		memset(&(IPCB(skb)->opt), 0, sizeof(IPCB(skb)->opt));
 		err = ip_queue_xmit(skb, 0);
 		if (err <= 0)
 			return err;
@@ -243,7 +242,8 @@ int dccp_write_xmit(struct sock *sk, struct sk_buff *skb, long *timeo)
 
 		err = dccp_transmit_skb(sk, skb);
 		ccid_hc_tx_packet_sent(dp->dccps_hc_tx_ccid, sk, 0, len);
-	}
+	} else
+		kfree_skb(skb);
 
 	return err;
 }
@@ -495,7 +495,7 @@ void dccp_send_close(struct sock *sk, const int active)
 {
 	struct dccp_sock *dp = dccp_sk(sk);
 	struct sk_buff *skb;
-	const unsigned int prio = active ? GFP_KERNEL : GFP_ATOMIC;
+	const gfp_t prio = active ? GFP_KERNEL : GFP_ATOMIC;
 
 	skb = alloc_skb(sk->sk_prot->max_header, prio);
 	if (skb == NULL)
diff --git a/net/dccp/proto.c b/net/dccp/proto.c
index a1cfd0e9e3b..a021c3422f6 100644
--- a/net/dccp/proto.c
+++ b/net/dccp/proto.c
@@ -402,8 +402,6 @@ int dccp_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
 	 *     This bug was _quickly_ found & fixed by just looking at an OSTRA
 	 *     generated callgraph 8) -acme
 	 */
-	if (rc != 0)
-		goto out_discard;
 out_release:
 	release_sock(sk);
 	return rc ? : len;
diff --git a/net/decnet/af_decnet.c b/net/decnet/af_decnet.c
index 348f36b529f..3f25cadccdd 100644
--- a/net/decnet/af_decnet.c
+++ b/net/decnet/af_decnet.c
@@ -452,7 +452,7 @@ static struct proto dn_proto = {
 	.obj_size = sizeof(struct dn_sock),
 };
 
-static struct sock *dn_alloc_sock(struct socket *sock, int gfp)
+static struct sock *dn_alloc_sock(struct socket *sock, gfp_t gfp)
 {
 	struct dn_scp *scp;
 	struct sock *sk = sk_alloc(PF_DECnet, gfp, &dn_proto, 1);
@@ -719,22 +719,9 @@ static int dn_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
 	if (saddr->sdn_flags & ~SDF_WILD)
 		return -EINVAL;
 
-#if 1
 	if (!capable(CAP_NET_BIND_SERVICE) && (saddr->sdn_objnum ||
 	    (saddr->sdn_flags & SDF_WILD)))
 		return -EACCES;
-#else
-	/*
-	 * Maybe put the default actions in the default security ops for
-	 * dn_prot_sock ? Would be nice if the capable call would go there
-	 * too.
-	 */
-	if (security_dn_prot_sock(saddr) &&
-	    !capable(CAP_NET_BIND_SERVICE) || 
-	    saddr->sdn_objnum || (saddr->sdn_flags & SDF_WILD))
-		return -EACCES;
-#endif
-
 
 	if (!(saddr->sdn_flags & SDF_WILD)) {
 		if (dn_ntohs(saddr->sdn_nodeaddrl)) {
@@ -804,7 +791,7 @@ static int dn_auto_bind(struct socket *sock)
 	return rv;
 }
 
-static int dn_confirm_accept(struct sock *sk, long *timeo, int allocation)
+static int dn_confirm_accept(struct sock *sk, long *timeo, gfp_t allocation)
 {
 	struct dn_scp *scp = DN_SK(sk);
 	DEFINE_WAIT(wait);
diff --git a/net/decnet/dn_nsp_out.c b/net/decnet/dn_nsp_out.c
index 53633d35286..c96c767b1f7 100644
--- a/net/decnet/dn_nsp_out.c
+++ b/net/decnet/dn_nsp_out.c
@@ -117,7 +117,7 @@ try_again:
  * The eventual aim is for each socket to have a cached header size
  * for its outgoing packets, and to set hdr from this when sk != NULL.
  */
-struct sk_buff *dn_alloc_skb(struct sock *sk, int size, int pri)
+struct sk_buff *dn_alloc_skb(struct sock *sk, int size, gfp_t pri)
 {
 	struct sk_buff *skb;
 	int hdr = 64;
@@ -210,7 +210,8 @@ static void dn_nsp_rtt(struct sock *sk, long rtt)
  *
  * Returns: The number of times the packet has been sent previously
  */
-static inline unsigned dn_nsp_clone_and_send(struct sk_buff *skb, int gfp)
+static inline unsigned dn_nsp_clone_and_send(struct sk_buff *skb,
+					     gfp_t gfp)
 {
 	struct dn_skb_cb *cb = DN_SKB_CB(skb);
 	struct sk_buff *skb2;
@@ -350,7 +351,8 @@ static unsigned short *dn_nsp_mk_data_header(struct sock *sk, struct sk_buff *sk
 	return ptr;
 }
 
-void dn_nsp_queue_xmit(struct sock *sk, struct sk_buff *skb, int gfp, int oth)
+void dn_nsp_queue_xmit(struct sock *sk, struct sk_buff *skb,
+			gfp_t gfp, int oth)
 {
 	struct dn_scp *scp = DN_SK(sk);
 	struct dn_skb_cb *cb = DN_SKB_CB(skb);
@@ -517,7 +519,7 @@ static int dn_nsp_retrans_conn_conf(struct sock *sk)
 	return 0;
 }
 
-void dn_send_conn_conf(struct sock *sk, int gfp)
+void dn_send_conn_conf(struct sock *sk, gfp_t gfp)
 {
 	struct dn_scp *scp = DN_SK(sk);
 	struct sk_buff *skb = NULL;
@@ -549,7 +551,8 @@ void dn_send_conn_conf(struct sock *sk, int gfp)
 
 
 static __inline__ void dn_nsp_do_disc(struct sock *sk, unsigned char msgflg, 
-			unsigned short reason, int gfp, struct dst_entry *dst,
+			unsigned short reason, gfp_t gfp,
+			struct dst_entry *dst,
 			int ddl, unsigned char *dd, __u16 rem, __u16 loc)
 {
 	struct sk_buff *skb = NULL;
@@ -591,7 +594,7 @@ static __inline__ void dn_nsp_do_disc(struct sock *sk, unsigned char msgflg,
 
 
 void dn_nsp_send_disc(struct sock *sk, unsigned char msgflg, 
-			unsigned short reason, int gfp)
+			unsigned short reason, gfp_t gfp)
 {
 	struct dn_scp *scp = DN_SK(sk);
 	int ddl = 0;
@@ -612,7 +615,7 @@ void dn_nsp_return_disc(struct sk_buff *skb, unsigned char msgflg,
 {
 	struct dn_skb_cb *cb = DN_SKB_CB(skb);
 	int ddl = 0;
-	int gfp = GFP_ATOMIC;
+	gfp_t gfp = GFP_ATOMIC;
 
 	dn_nsp_do_disc(NULL, msgflg, reason, gfp, skb->dst, ddl, 
 			NULL, cb->src_port, cb->dst_port);
@@ -624,7 +627,7 @@ void dn_nsp_send_link(struct sock *sk, unsigned char lsflags, char fcval)
 	struct dn_scp *scp = DN_SK(sk);
 	struct sk_buff *skb;
 	unsigned char *ptr;
-	int gfp = GFP_ATOMIC;
+	gfp_t gfp = GFP_ATOMIC;
 
 	if ((skb = dn_alloc_skb(sk, DN_MAX_NSP_DATA_HEADER + 2, gfp)) == NULL)
 		return;
@@ -659,7 +662,7 @@ void dn_nsp_send_conninit(struct sock *sk, unsigned char msgflg)
 	unsigned char menuver;
 	struct dn_skb_cb *cb;
 	unsigned char type = 1;
-	int allocation = (msgflg == NSP_CI) ? sk->sk_allocation : GFP_ATOMIC;
+	gfp_t allocation = (msgflg == NSP_CI) ? sk->sk_allocation : GFP_ATOMIC;
 	struct sk_buff *skb = dn_alloc_skb(sk, 200, allocation);
 
 	if (!skb)
diff --git a/net/econet/af_econet.c b/net/econet/af_econet.c
index 4a62093eb34..34fdac51df9 100644
--- a/net/econet/af_econet.c
+++ b/net/econet/af_econet.c
@@ -406,7 +406,7 @@ static int econet_sendmsg(struct kiocb *iocb, struct socket *sock,
 		unsigned long network = 0;
 
 		rcu_read_lock();
-		idev = __in_dev_get(dev);
+		idev = __in_dev_get_rcu(dev);
 		if (idev) {
 			if (idev->ifa_list)
 				network = ntohl(idev->ifa_list->ifa_address) & 
diff --git a/net/ethernet/eth.c b/net/ethernet/eth.c
index 87a052a9a84..68a5ca86644 100644
--- a/net/ethernet/eth.c
+++ b/net/ethernet/eth.c
@@ -146,6 +146,19 @@ int eth_rebuild_header(struct sk_buff *skb)
 	return 0;
 }
 
+static inline unsigned int compare_eth_addr(const unsigned char *__a, const unsigned char *__b)
+{
+	const unsigned short *dest = (unsigned short *) __a;
+	const unsigned short *devaddr = (unsigned short *) __b;
+	unsigned int res;
+
+	BUILD_BUG_ON(ETH_ALEN != 6);
+	res = ((dest[0] ^ devaddr[0]) |
+	       (dest[1] ^ devaddr[1]) |
+	       (dest[2] ^ devaddr[2])) != 0;
+
+	return res;
+}
 
 /*
  *	Determine the packet's protocol ID. The rule here is that we 
@@ -158,16 +171,15 @@ __be16 eth_type_trans(struct sk_buff *skb, struct net_device *dev)
 	struct ethhdr *eth;
 	unsigned char *rawp;
 	
-	skb->mac.raw=skb->data;
+	skb->mac.raw = skb->data;
 	skb_pull(skb,ETH_HLEN);
 	eth = eth_hdr(skb);
 	
-	if(*eth->h_dest&1)
-	{
-		if(memcmp(eth->h_dest,dev->broadcast, ETH_ALEN)==0)
-			skb->pkt_type=PACKET_BROADCAST;
+	if (*eth->h_dest&1) {
+		if (!compare_eth_addr(eth->h_dest, dev->broadcast))
+			skb->pkt_type = PACKET_BROADCAST;
 		else
-			skb->pkt_type=PACKET_MULTICAST;
+			skb->pkt_type = PACKET_MULTICAST;
 	}
 	
 	/*
@@ -178,10 +190,9 @@ __be16 eth_type_trans(struct sk_buff *skb, struct net_device *dev)
 	 *	seems to set IFF_PROMISC.
 	 */
 	 
-	else if(1 /*dev->flags&IFF_PROMISC*/)
-	{
-		if(memcmp(eth->h_dest,dev->dev_addr, ETH_ALEN))
-			skb->pkt_type=PACKET_OTHERHOST;
+	else if(1 /*dev->flags&IFF_PROMISC*/) {
+		if (unlikely(compare_eth_addr(eth->h_dest, dev->dev_addr)))
+			skb->pkt_type = PACKET_OTHERHOST;
 	}
 	
 	if (ntohs(eth->h_proto) >= 1536)
diff --git a/net/ieee80211/Makefile b/net/ieee80211/Makefile
index a6ccac5baea..f988417121d 100644
--- a/net/ieee80211/Makefile
+++ b/net/ieee80211/Makefile
@@ -7,5 +7,6 @@ ieee80211-objs := \
 	ieee80211_module.o \
 	ieee80211_tx.o \
 	ieee80211_rx.o \
-	ieee80211_wx.o
+	ieee80211_wx.o \
+	ieee80211_geo.o
 
diff --git a/net/ieee80211/ieee80211_crypt.c b/net/ieee80211/ieee80211_crypt.c
index 61a9d92e455..f3b6aa3be63 100644
--- a/net/ieee80211/ieee80211_crypt.c
+++ b/net/ieee80211/ieee80211_crypt.c
@@ -41,6 +41,12 @@ void ieee80211_crypt_deinit_entries(struct ieee80211_device *ieee, int force)
 {
 	struct list_head *ptr, *n;
 	struct ieee80211_crypt_data *entry;
+	unsigned long flags;
+
+	spin_lock_irqsave(&ieee->lock, flags);
+
+	if (list_empty(&ieee->crypt_deinit_list))
+		goto unlock;
 
 	for (ptr = ieee->crypt_deinit_list.next, n = ptr->next;
 	     ptr != &ieee->crypt_deinit_list; ptr = n, n = ptr->next) {
@@ -57,6 +63,18 @@ void ieee80211_crypt_deinit_entries(struct ieee80211_device *ieee, int force)
 		}
 		kfree(entry);
 	}
+      unlock:
+	spin_unlock_irqrestore(&ieee->lock, flags);
+}
+
+/* After this, crypt_deinit_list won't accept new members */
+void ieee80211_crypt_quiescing(struct ieee80211_device *ieee)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&ieee->lock, flags);
+	ieee->crypt_quiesced = 1;
+	spin_unlock_irqrestore(&ieee->lock, flags);
 }
 
 void ieee80211_crypt_deinit_handler(unsigned long data)
@@ -64,16 +82,16 @@ void ieee80211_crypt_deinit_handler(unsigned long data)
 	struct ieee80211_device *ieee = (struct ieee80211_device *)data;
 	unsigned long flags;
 
-	spin_lock_irqsave(&ieee->lock, flags);
 	ieee80211_crypt_deinit_entries(ieee, 0);
-	if (!list_empty(&ieee->crypt_deinit_list)) {
+
+	spin_lock_irqsave(&ieee->lock, flags);
+	if (!list_empty(&ieee->crypt_deinit_list) && !ieee->crypt_quiesced) {
 		printk(KERN_DEBUG "%s: entries remaining in delayed crypt "
 		       "deletion list\n", ieee->dev->name);
 		ieee->crypt_deinit_timer.expires = jiffies + HZ;
 		add_timer(&ieee->crypt_deinit_timer);
 	}
 	spin_unlock_irqrestore(&ieee->lock, flags);
-
 }
 
 void ieee80211_crypt_delayed_deinit(struct ieee80211_device *ieee,
@@ -93,10 +111,12 @@ void ieee80211_crypt_delayed_deinit(struct ieee80211_device *ieee,
 	 * locking. */
 
 	spin_lock_irqsave(&ieee->lock, flags);
-	list_add(&tmp->list, &ieee->crypt_deinit_list);
-	if (!timer_pending(&ieee->crypt_deinit_timer)) {
-		ieee->crypt_deinit_timer.expires = jiffies + HZ;
-		add_timer(&ieee->crypt_deinit_timer);
+	if (!ieee->crypt_quiesced) {
+		list_add(&tmp->list, &ieee->crypt_deinit_list);
+		if (!timer_pending(&ieee->crypt_deinit_timer)) {
+			ieee->crypt_deinit_timer.expires = jiffies + HZ;
+			add_timer(&ieee->crypt_deinit_timer);
+		}
 	}
 	spin_unlock_irqrestore(&ieee->lock, flags);
 }
@@ -191,18 +211,18 @@ static void ieee80211_crypt_null_deinit(void *priv)
 }
 
 static struct ieee80211_crypto_ops ieee80211_crypt_null = {
-	.name			= "NULL",
-	.init			= ieee80211_crypt_null_init,
-	.deinit			= ieee80211_crypt_null_deinit,
-	.encrypt_mpdu		= NULL,
-	.decrypt_mpdu		= NULL,
-	.encrypt_msdu		= NULL,
-	.decrypt_msdu		= NULL,
-	.set_key		= NULL,
-	.get_key		= NULL,
-	.extra_prefix_len	= 0,
-	.extra_postfix_len	= 0,
-	.owner			= THIS_MODULE,
+	.name = "NULL",
+	.init = ieee80211_crypt_null_init,
+	.deinit = ieee80211_crypt_null_deinit,
+	.encrypt_mpdu = NULL,
+	.decrypt_mpdu = NULL,
+	.encrypt_msdu = NULL,
+	.decrypt_msdu = NULL,
+	.set_key = NULL,
+	.get_key = NULL,
+	.extra_mpdu_prefix_len = 0,
+	.extra_mpdu_postfix_len = 0,
+	.owner = THIS_MODULE,
 };
 
 static int __init ieee80211_crypto_init(void)
@@ -249,6 +269,7 @@ static void __exit ieee80211_crypto_deinit(void)
 EXPORT_SYMBOL(ieee80211_crypt_deinit_entries);
 EXPORT_SYMBOL(ieee80211_crypt_deinit_handler);
 EXPORT_SYMBOL(ieee80211_crypt_delayed_deinit);
+EXPORT_SYMBOL(ieee80211_crypt_quiescing);
 
 EXPORT_SYMBOL(ieee80211_register_crypto_ops);
 EXPORT_SYMBOL(ieee80211_unregister_crypto_ops);
diff --git a/net/ieee80211/ieee80211_crypt_ccmp.c b/net/ieee80211/ieee80211_crypt_ccmp.c
index 8fc13f45971..05a853c1301 100644
--- a/net/ieee80211/ieee80211_crypt_ccmp.c
+++ b/net/ieee80211/ieee80211_crypt_ccmp.c
@@ -119,7 +119,7 @@ static inline void xor_block(u8 * b, u8 * a, size_t len)
 }
 
 static void ccmp_init_blocks(struct crypto_tfm *tfm,
-			     struct ieee80211_hdr *hdr,
+			     struct ieee80211_hdr_4addr *hdr,
 			     u8 * pn, size_t dlen, u8 * b0, u8 * auth, u8 * s0)
 {
 	u8 *pos, qc = 0;
@@ -191,26 +191,18 @@ static void ccmp_init_blocks(struct crypto_tfm *tfm,
 	ieee80211_ccmp_aes_encrypt(tfm, b0, s0);
 }
 
-static int ieee80211_ccmp_encrypt(struct sk_buff *skb, int hdr_len, void *priv)
+static int ieee80211_ccmp_hdr(struct sk_buff *skb, int hdr_len, void *priv)
 {
 	struct ieee80211_ccmp_data *key = priv;
-	int data_len, i, blocks, last, len;
-	u8 *pos, *mic;
-	struct ieee80211_hdr *hdr;
-	u8 *b0 = key->tx_b0;
-	u8 *b = key->tx_b;
-	u8 *e = key->tx_e;
-	u8 *s0 = key->tx_s0;
+	int i;
+	u8 *pos;
 
-	if (skb_headroom(skb) < CCMP_HDR_LEN ||
-	    skb_tailroom(skb) < CCMP_MIC_LEN || skb->len < hdr_len)
+	if (skb_headroom(skb) < CCMP_HDR_LEN || skb->len < hdr_len)
 		return -1;
 
-	data_len = skb->len - hdr_len;
 	pos = skb_push(skb, CCMP_HDR_LEN);
 	memmove(pos, pos + CCMP_HDR_LEN, hdr_len);
 	pos += hdr_len;
-	mic = skb_put(skb, CCMP_MIC_LEN);
 
 	i = CCMP_PN_LEN - 1;
 	while (i >= 0) {
@@ -229,7 +221,31 @@ static int ieee80211_ccmp_encrypt(struct sk_buff *skb, int hdr_len, void *priv)
 	*pos++ = key->tx_pn[1];
 	*pos++ = key->tx_pn[0];
 
-	hdr = (struct ieee80211_hdr *)skb->data;
+	return CCMP_HDR_LEN;
+}
+
+static int ieee80211_ccmp_encrypt(struct sk_buff *skb, int hdr_len, void *priv)
+{
+	struct ieee80211_ccmp_data *key = priv;
+	int data_len, i, blocks, last, len;
+	u8 *pos, *mic;
+	struct ieee80211_hdr_4addr *hdr;
+	u8 *b0 = key->tx_b0;
+	u8 *b = key->tx_b;
+	u8 *e = key->tx_e;
+	u8 *s0 = key->tx_s0;
+
+	if (skb_tailroom(skb) < CCMP_MIC_LEN || skb->len < hdr_len)
+		return -1;
+
+	data_len = skb->len - hdr_len;
+	len = ieee80211_ccmp_hdr(skb, hdr_len, priv);
+	if (len < 0)
+		return -1;
+
+	pos = skb->data + hdr_len + CCMP_HDR_LEN;
+	mic = skb_put(skb, CCMP_MIC_LEN);
+	hdr = (struct ieee80211_hdr_4addr *)skb->data;
 	ccmp_init_blocks(key->tfm, hdr, key->tx_pn, data_len, b0, b, s0);
 
 	blocks = (data_len + AES_BLOCK_LEN - 1) / AES_BLOCK_LEN;
@@ -258,7 +274,7 @@ static int ieee80211_ccmp_decrypt(struct sk_buff *skb, int hdr_len, void *priv)
 {
 	struct ieee80211_ccmp_data *key = priv;
 	u8 keyidx, *pos;
-	struct ieee80211_hdr *hdr;
+	struct ieee80211_hdr_4addr *hdr;
 	u8 *b0 = key->rx_b0;
 	u8 *b = key->rx_b;
 	u8 *a = key->rx_a;
@@ -272,7 +288,7 @@ static int ieee80211_ccmp_decrypt(struct sk_buff *skb, int hdr_len, void *priv)
 		return -1;
 	}
 
-	hdr = (struct ieee80211_hdr *)skb->data;
+	hdr = (struct ieee80211_hdr_4addr *)skb->data;
 	pos = skb->data + hdr_len;
 	keyidx = pos[3];
 	if (!(keyidx & (1 << 5))) {
@@ -426,19 +442,20 @@ static char *ieee80211_ccmp_print_stats(char *p, void *priv)
 }
 
 static struct ieee80211_crypto_ops ieee80211_crypt_ccmp = {
-	.name			= "CCMP",
-	.init			= ieee80211_ccmp_init,
-	.deinit			= ieee80211_ccmp_deinit,
-	.encrypt_mpdu		= ieee80211_ccmp_encrypt,
-	.decrypt_mpdu		= ieee80211_ccmp_decrypt,
-	.encrypt_msdu		= NULL,
-	.decrypt_msdu		= NULL,
-	.set_key		= ieee80211_ccmp_set_key,
-	.get_key		= ieee80211_ccmp_get_key,
-	.print_stats		= ieee80211_ccmp_print_stats,
-	.extra_prefix_len	= CCMP_HDR_LEN,
-	.extra_postfix_len	= CCMP_MIC_LEN,
-	.owner			= THIS_MODULE,
+	.name = "CCMP",
+	.init = ieee80211_ccmp_init,
+	.deinit = ieee80211_ccmp_deinit,
+	.build_iv = ieee80211_ccmp_hdr,
+	.encrypt_mpdu = ieee80211_ccmp_encrypt,
+	.decrypt_mpdu = ieee80211_ccmp_decrypt,
+	.encrypt_msdu = NULL,
+	.decrypt_msdu = NULL,
+	.set_key = ieee80211_ccmp_set_key,
+	.get_key = ieee80211_ccmp_get_key,
+	.print_stats = ieee80211_ccmp_print_stats,
+	.extra_mpdu_prefix_len = CCMP_HDR_LEN,
+	.extra_mpdu_postfix_len = CCMP_MIC_LEN,
+	.owner = THIS_MODULE,
 };
 
 static int __init ieee80211_crypto_ccmp_init(void)
diff --git a/net/ieee80211/ieee80211_crypt_tkip.c b/net/ieee80211/ieee80211_crypt_tkip.c
index d4f9164be1a..2e34f29b795 100644
--- a/net/ieee80211/ieee80211_crypt_tkip.c
+++ b/net/ieee80211/ieee80211_crypt_tkip.c
@@ -59,8 +59,24 @@ struct ieee80211_tkip_data {
 
 	/* scratch buffers for virt_to_page() (crypto API) */
 	u8 rx_hdr[16], tx_hdr[16];
+
+	unsigned long flags;
 };
 
+static unsigned long ieee80211_tkip_set_flags(unsigned long flags, void *priv)
+{
+	struct ieee80211_tkip_data *_priv = priv;
+	unsigned long old_flags = _priv->flags;
+	_priv->flags = flags;
+	return old_flags;
+}
+
+static unsigned long ieee80211_tkip_get_flags(void *priv)
+{
+	struct ieee80211_tkip_data *_priv = priv;
+	return _priv->flags;
+}
+
 static void *ieee80211_tkip_init(int key_idx)
 {
 	struct ieee80211_tkip_data *priv;
@@ -69,6 +85,7 @@ static void *ieee80211_tkip_init(int key_idx)
 	if (priv == NULL)
 		goto fail;
 	memset(priv, 0, sizeof(*priv));
+
 	priv->key_idx = key_idx;
 
 	priv->tfm_arc4 = crypto_alloc_tfm("arc4", 0);
@@ -255,25 +272,27 @@ static void tkip_mixing_phase2(u8 * WEPSeed, const u8 * TK, const u16 * TTAK,
 #endif
 }
 
-static int ieee80211_tkip_encrypt(struct sk_buff *skb, int hdr_len, void *priv)
+static u8 *ieee80211_tkip_hdr(struct sk_buff *skb, int hdr_len, void *priv)
 {
 	struct ieee80211_tkip_data *tkey = priv;
 	int len;
-	u8 rc4key[16], *pos, *icv;
-	struct ieee80211_hdr *hdr;
+	u8 *rc4key, *pos, *icv;
+	struct ieee80211_hdr_4addr *hdr;
 	u32 crc;
-	struct scatterlist sg;
 
-	if (skb_headroom(skb) < 8 || skb_tailroom(skb) < 4 ||
-	    skb->len < hdr_len)
-		return -1;
+	hdr = (struct ieee80211_hdr_4addr *)skb->data;
+
+	if (skb_headroom(skb) < 8 || skb->len < hdr_len)
+		return NULL;
 
-	hdr = (struct ieee80211_hdr *)skb->data;
 	if (!tkey->tx_phase1_done) {
 		tkip_mixing_phase1(tkey->tx_ttak, tkey->key, hdr->addr2,
 				   tkey->tx_iv32);
 		tkey->tx_phase1_done = 1;
 	}
+	rc4key = kmalloc(16, GFP_ATOMIC);
+	if (!rc4key)
+		return NULL;
 	tkip_mixing_phase2(rc4key, tkey->key, tkey->tx_ttak, tkey->tx_iv16);
 
 	len = skb->len - hdr_len;
@@ -282,9 +301,9 @@ static int ieee80211_tkip_encrypt(struct sk_buff *skb, int hdr_len, void *priv)
 	pos += hdr_len;
 	icv = skb_put(skb, 4);
 
-	*pos++ = rc4key[0];
-	*pos++ = rc4key[1];
-	*pos++ = rc4key[2];
+	*pos++ = *rc4key;
+	*pos++ = *(rc4key + 1);
+	*pos++ = *(rc4key + 2);
 	*pos++ = (tkey->key_idx << 6) | (1 << 5) /* Ext IV included */ ;
 	*pos++ = tkey->tx_iv32 & 0xff;
 	*pos++ = (tkey->tx_iv32 >> 8) & 0xff;
@@ -297,6 +316,38 @@ static int ieee80211_tkip_encrypt(struct sk_buff *skb, int hdr_len, void *priv)
 	icv[2] = crc >> 16;
 	icv[3] = crc >> 24;
 
+	return rc4key;
+}
+
+static int ieee80211_tkip_encrypt(struct sk_buff *skb, int hdr_len, void *priv)
+{
+	struct ieee80211_tkip_data *tkey = priv;
+	int len;
+	const u8 *rc4key;
+	u8 *pos;
+	struct scatterlist sg;
+
+	if (tkey->flags & IEEE80211_CRYPTO_TKIP_COUNTERMEASURES) {
+		if (net_ratelimit()) {
+			struct ieee80211_hdr_4addr *hdr =
+			    (struct ieee80211_hdr_4addr *)skb->data;
+			printk(KERN_DEBUG "TKIP countermeasures: dropped "
+			       "TX packet to " MAC_FMT "\n",
+			       MAC_ARG(hdr->addr1));
+		}
+		return -1;
+	}
+
+	if (skb_tailroom(skb) < 4 || skb->len < hdr_len)
+		return -1;
+
+	len = skb->len - hdr_len;
+	pos = skb->data + hdr_len;
+
+	rc4key = ieee80211_tkip_hdr(skb, hdr_len, priv);
+	if (!rc4key)
+		return -1;
+
 	crypto_cipher_setkey(tkey->tfm_arc4, rc4key, 16);
 	sg.page = virt_to_page(pos);
 	sg.offset = offset_in_page(pos);
@@ -319,16 +370,26 @@ static int ieee80211_tkip_decrypt(struct sk_buff *skb, int hdr_len, void *priv)
 	u8 keyidx, *pos;
 	u32 iv32;
 	u16 iv16;
-	struct ieee80211_hdr *hdr;
+	struct ieee80211_hdr_4addr *hdr;
 	u8 icv[4];
 	u32 crc;
 	struct scatterlist sg;
 	int plen;
 
+	hdr = (struct ieee80211_hdr_4addr *)skb->data;
+
+	if (tkey->flags & IEEE80211_CRYPTO_TKIP_COUNTERMEASURES) {
+		if (net_ratelimit()) {
+			printk(KERN_DEBUG "TKIP countermeasures: dropped "
+			       "received packet from " MAC_FMT "\n",
+			       MAC_ARG(hdr->addr2));
+		}
+		return -1;
+	}
+
 	if (skb->len < hdr_len + 8 + 4)
 		return -1;
 
-	hdr = (struct ieee80211_hdr *)skb->data;
 	pos = skb->data + hdr_len;
 	keyidx = pos[3];
 	if (!(keyidx & (1 << 5))) {
@@ -441,9 +502,9 @@ static int michael_mic(struct ieee80211_tkip_data *tkey, u8 * key, u8 * hdr,
 
 static void michael_mic_hdr(struct sk_buff *skb, u8 * hdr)
 {
-	struct ieee80211_hdr *hdr11;
+	struct ieee80211_hdr_4addr *hdr11;
 
-	hdr11 = (struct ieee80211_hdr *)skb->data;
+	hdr11 = (struct ieee80211_hdr_4addr *)skb->data;
 	switch (le16_to_cpu(hdr11->frame_ctl) &
 		(IEEE80211_FCTL_FROMDS | IEEE80211_FCTL_TODS)) {
 	case IEEE80211_FCTL_TODS:
@@ -490,9 +551,9 @@ static int ieee80211_michael_mic_add(struct sk_buff *skb, int hdr_len,
 	return 0;
 }
 
-#if WIRELESS_EXT >= 18
 static void ieee80211_michael_mic_failure(struct net_device *dev,
-					  struct ieee80211_hdr *hdr, int keyidx)
+					  struct ieee80211_hdr_4addr *hdr,
+					  int keyidx)
 {
 	union iwreq_data wrqu;
 	struct iw_michaelmicfailure ev;
@@ -510,28 +571,6 @@ static void ieee80211_michael_mic_failure(struct net_device *dev,
 	wrqu.data.length = sizeof(ev);
 	wireless_send_event(dev, IWEVMICHAELMICFAILURE, &wrqu, (char *)&ev);
 }
-#elif WIRELESS_EXT >= 15
-static void ieee80211_michael_mic_failure(struct net_device *dev,
-					  struct ieee80211_hdr *hdr, int keyidx)
-{
-	union iwreq_data wrqu;
-	char buf[128];
-
-	/* TODO: needed parameters: count, keyid, key type, TSC */
-	sprintf(buf, "MLME-MICHAELMICFAILURE.indication(keyid=%d %scast addr="
-		MAC_FMT ")", keyidx, hdr->addr1[0] & 0x01 ? "broad" : "uni",
-		MAC_ARG(hdr->addr2));
-	memset(&wrqu, 0, sizeof(wrqu));
-	wrqu.data.length = strlen(buf);
-	wireless_send_event(dev, IWEVCUSTOM, &wrqu, buf);
-}
-#else				/* WIRELESS_EXT >= 15 */
-static inline void ieee80211_michael_mic_failure(struct net_device *dev,
-						 struct ieee80211_hdr *hdr,
-						 int keyidx)
-{
-}
-#endif				/* WIRELESS_EXT >= 15 */
 
 static int ieee80211_michael_mic_verify(struct sk_buff *skb, int keyidx,
 					int hdr_len, void *priv)
@@ -547,8 +586,8 @@ static int ieee80211_michael_mic_verify(struct sk_buff *skb, int keyidx,
 			skb->data + hdr_len, skb->len - 8 - hdr_len, mic))
 		return -1;
 	if (memcmp(mic, skb->data + skb->len - 8, 8) != 0) {
-		struct ieee80211_hdr *hdr;
-		hdr = (struct ieee80211_hdr *)skb->data;
+		struct ieee80211_hdr_4addr *hdr;
+		hdr = (struct ieee80211_hdr_4addr *)skb->data;
 		printk(KERN_DEBUG "%s: Michael MIC verification failed for "
 		       "MSDU from " MAC_FMT " keyidx=%d\n",
 		       skb->dev ? skb->dev->name : "N/A", MAC_ARG(hdr->addr2),
@@ -654,19 +693,22 @@ static char *ieee80211_tkip_print_stats(char *p, void *priv)
 }
 
 static struct ieee80211_crypto_ops ieee80211_crypt_tkip = {
-	.name			= "TKIP",
-	.init			= ieee80211_tkip_init,
-	.deinit			= ieee80211_tkip_deinit,
-	.encrypt_mpdu		= ieee80211_tkip_encrypt,
-	.decrypt_mpdu		= ieee80211_tkip_decrypt,
-	.encrypt_msdu		= ieee80211_michael_mic_add,
-	.decrypt_msdu		= ieee80211_michael_mic_verify,
-	.set_key		= ieee80211_tkip_set_key,
-	.get_key		= ieee80211_tkip_get_key,
-	.print_stats		= ieee80211_tkip_print_stats,
-	.extra_prefix_len	= 4 + 4,	/* IV + ExtIV */
-	.extra_postfix_len	= 8 + 4,	/* MIC + ICV */
-	.owner			= THIS_MODULE,
+	.name = "TKIP",
+	.init = ieee80211_tkip_init,
+	.deinit = ieee80211_tkip_deinit,
+	.encrypt_mpdu = ieee80211_tkip_encrypt,
+	.decrypt_mpdu = ieee80211_tkip_decrypt,
+	.encrypt_msdu = ieee80211_michael_mic_add,
+	.decrypt_msdu = ieee80211_michael_mic_verify,
+	.set_key = ieee80211_tkip_set_key,
+	.get_key = ieee80211_tkip_get_key,
+	.print_stats = ieee80211_tkip_print_stats,
+	.extra_mpdu_prefix_len = 4 + 4,	/* IV + ExtIV */
+	.extra_mpdu_postfix_len = 4,	/* ICV */
+	.extra_msdu_postfix_len = 8,	/* MIC */
+	.get_flags = ieee80211_tkip_get_flags,
+	.set_flags = ieee80211_tkip_set_flags,
+	.owner = THIS_MODULE,
 };
 
 static int __init ieee80211_crypto_tkip_init(void)
diff --git a/net/ieee80211/ieee80211_crypt_wep.c b/net/ieee80211/ieee80211_crypt_wep.c
index b4d2514a090..7c08ed2f262 100644
--- a/net/ieee80211/ieee80211_crypt_wep.c
+++ b/net/ieee80211/ieee80211_crypt_wep.c
@@ -229,19 +229,19 @@ static char *prism2_wep_print_stats(char *p, void *priv)
 }
 
 static struct ieee80211_crypto_ops ieee80211_crypt_wep = {
-	.name			= "WEP",
-	.init			= prism2_wep_init,
-	.deinit			= prism2_wep_deinit,
-	.encrypt_mpdu		= prism2_wep_encrypt,
-	.decrypt_mpdu		= prism2_wep_decrypt,
-	.encrypt_msdu		= NULL,
-	.decrypt_msdu		= NULL,
-	.set_key		= prism2_wep_set_key,
-	.get_key		= prism2_wep_get_key,
-	.print_stats		= prism2_wep_print_stats,
-	.extra_prefix_len	= 4,	/* IV */
-	.extra_postfix_len	= 4,	/* ICV */
-	.owner			= THIS_MODULE,
+	.name = "WEP",
+	.init = prism2_wep_init,
+	.deinit = prism2_wep_deinit,
+	.encrypt_mpdu = prism2_wep_encrypt,
+	.decrypt_mpdu = prism2_wep_decrypt,
+	.encrypt_msdu = NULL,
+	.decrypt_msdu = NULL,
+	.set_key = prism2_wep_set_key,
+	.get_key = prism2_wep_get_key,
+	.print_stats = prism2_wep_print_stats,
+	.extra_mpdu_prefix_len = 4,	/* IV */
+	.extra_mpdu_postfix_len = 4,	/* ICV */
+	.owner = THIS_MODULE,
 };
 
 static int __init ieee80211_crypto_wep_init(void)
diff --git a/net/ieee80211/ieee80211_geo.c b/net/ieee80211/ieee80211_geo.c
new file mode 100644
index 00000000000..c4b54ef8f6d
--- /dev/null
+++ b/net/ieee80211/ieee80211_geo.c
@@ -0,0 +1,141 @@
+/******************************************************************************
+
+  Copyright(c) 2005 Intel Corporation. All rights reserved.
+
+  This program is free software; you can redistribute it and/or modify it
+  under the terms of version 2 of the GNU General Public License as
+  published by the Free Software Foundation.
+
+  This program is distributed in the hope that it will be useful, but WITHOUT
+  ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+  FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+  more details.
+
+  You should have received a copy of the GNU General Public License along with
+  this program; if not, write to the Free Software Foundation, Inc., 59
+  Temple Place - Suite 330, Boston, MA  02111-1307, USA.
+
+  The full GNU General Public License is included in this distribution in the
+  file called LICENSE.
+
+  Contact Information:
+  James P. Ketrenos <ipw2100-admin@linux.intel.com>
+  Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497
+
+******************************************************************************/
+#include <linux/compiler.h>
+#include <linux/config.h>
+#include <linux/errno.h>
+#include <linux/if_arp.h>
+#include <linux/in6.h>
+#include <linux/in.h>
+#include <linux/ip.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/netdevice.h>
+#include <linux/proc_fs.h>
+#include <linux/skbuff.h>
+#include <linux/slab.h>
+#include <linux/tcp.h>
+#include <linux/types.h>
+#include <linux/version.h>
+#include <linux/wireless.h>
+#include <linux/etherdevice.h>
+#include <asm/uaccess.h>
+
+#include <net/ieee80211.h>
+
+int ieee80211_is_valid_channel(struct ieee80211_device *ieee, u8 channel)
+{
+	int i;
+
+	/* Driver needs to initialize the geography map before using
+	 * these helper functions */
+	BUG_ON(ieee->geo.bg_channels == 0 && ieee->geo.a_channels == 0);
+
+	if (ieee->freq_band & IEEE80211_24GHZ_BAND)
+		for (i = 0; i < ieee->geo.bg_channels; i++)
+			/* NOTE: If G mode is currently supported but
+			 * this is a B only channel, we don't see it
+			 * as valid. */
+			if ((ieee->geo.bg[i].channel == channel) &&
+			    (!(ieee->mode & IEEE_G) ||
+			     !(ieee->geo.bg[i].flags & IEEE80211_CH_B_ONLY)))
+				return IEEE80211_24GHZ_BAND;
+
+	if (ieee->freq_band & IEEE80211_52GHZ_BAND)
+		for (i = 0; i < ieee->geo.a_channels; i++)
+			if (ieee->geo.a[i].channel == channel)
+				return IEEE80211_52GHZ_BAND;
+
+	return 0;
+}
+
+int ieee80211_channel_to_index(struct ieee80211_device *ieee, u8 channel)
+{
+	int i;
+
+	/* Driver needs to initialize the geography map before using
+	 * these helper functions */
+	BUG_ON(ieee->geo.bg_channels == 0 && ieee->geo.a_channels == 0);
+
+	if (ieee->freq_band & IEEE80211_24GHZ_BAND)
+		for (i = 0; i < ieee->geo.bg_channels; i++)
+			if (ieee->geo.bg[i].channel == channel)
+				return i;
+
+	if (ieee->freq_band & IEEE80211_52GHZ_BAND)
+		for (i = 0; i < ieee->geo.a_channels; i++)
+			if (ieee->geo.a[i].channel == channel)
+				return i;
+
+	return -1;
+}
+
+u8 ieee80211_freq_to_channel(struct ieee80211_device * ieee, u32 freq)
+{
+	int i;
+
+	/* Driver needs to initialize the geography map before using
+	 * these helper functions */
+	BUG_ON(ieee->geo.bg_channels == 0 && ieee->geo.a_channels == 0);
+
+	freq /= 100000;
+
+	if (ieee->freq_band & IEEE80211_24GHZ_BAND)
+		for (i = 0; i < ieee->geo.bg_channels; i++)
+			if (ieee->geo.bg[i].freq == freq)
+				return ieee->geo.bg[i].channel;
+
+	if (ieee->freq_band & IEEE80211_52GHZ_BAND)
+		for (i = 0; i < ieee->geo.a_channels; i++)
+			if (ieee->geo.a[i].freq == freq)
+				return ieee->geo.a[i].channel;
+
+	return 0;
+}
+
+int ieee80211_set_geo(struct ieee80211_device *ieee,
+		      const struct ieee80211_geo *geo)
+{
+	memcpy(ieee->geo.name, geo->name, 3);
+	ieee->geo.name[3] = '\0';
+	ieee->geo.bg_channels = geo->bg_channels;
+	ieee->geo.a_channels = geo->a_channels;
+	memcpy(ieee->geo.bg, geo->bg, geo->bg_channels *
+	       sizeof(struct ieee80211_channel));
+	memcpy(ieee->geo.a, geo->a, ieee->geo.a_channels *
+	       sizeof(struct ieee80211_channel));
+	return 0;
+}
+
+const struct ieee80211_geo *ieee80211_get_geo(struct ieee80211_device *ieee)
+{
+	return &ieee->geo;
+}
+
+EXPORT_SYMBOL(ieee80211_is_valid_channel);
+EXPORT_SYMBOL(ieee80211_freq_to_channel);
+EXPORT_SYMBOL(ieee80211_channel_to_index);
+EXPORT_SYMBOL(ieee80211_set_geo);
+EXPORT_SYMBOL(ieee80211_get_geo);
diff --git a/net/ieee80211/ieee80211_module.c b/net/ieee80211/ieee80211_module.c
index 03a47343ddc..f66d792cd20 100644
--- a/net/ieee80211/ieee80211_module.c
+++ b/net/ieee80211/ieee80211_module.c
@@ -1,6 +1,6 @@
 /*******************************************************************************
 
-  Copyright(c) 2004 Intel Corporation. All rights reserved.
+  Copyright(c) 2004-2005 Intel Corporation. All rights reserved.
 
   Portions of this file are based on the WEP enablement code provided by the
   Host AP project hostap-drivers v0.1.3
@@ -53,12 +53,15 @@
 
 #include <net/ieee80211.h>
 
-MODULE_DESCRIPTION("802.11 data/management/control stack");
-MODULE_AUTHOR
-    ("Copyright (C) 2004 Intel Corporation <jketreno@linux.intel.com>");
-MODULE_LICENSE("GPL");
+#define DRV_DESCRIPTION "802.11 data/management/control stack"
+#define DRV_NAME        "ieee80211"
+#define DRV_VERSION	IEEE80211_VERSION
+#define DRV_COPYRIGHT   "Copyright (C) 2004-2005 Intel Corporation <jketreno@linux.intel.com>"
 
-#define DRV_NAME "ieee80211"
+MODULE_VERSION(DRV_VERSION);
+MODULE_DESCRIPTION(DRV_DESCRIPTION);
+MODULE_AUTHOR(DRV_COPYRIGHT);
+MODULE_LICENSE("GPL");
 
 static inline int ieee80211_networks_allocate(struct ieee80211_device *ieee)
 {
@@ -126,26 +129,34 @@ struct net_device *alloc_ieee80211(int sizeof_priv)
 
 	/* Default fragmentation threshold is maximum payload size */
 	ieee->fts = DEFAULT_FTS;
+	ieee->rts = DEFAULT_FTS;
 	ieee->scan_age = DEFAULT_MAX_SCAN_AGE;
 	ieee->open_wep = 1;
 
 	/* Default to enabling full open WEP with host based encrypt/decrypt */
 	ieee->host_encrypt = 1;
 	ieee->host_decrypt = 1;
+	ieee->host_mc_decrypt = 1;
+
+	/* Host fragementation in Open mode. Default is enabled.
+	 * Note: host fragmentation is always enabled if host encryption
+	 * is enabled. For cards can do hardware encryption, they must do
+	 * hardware fragmentation as well. So we don't need a variable
+	 * like host_enc_frag. */
+	ieee->host_open_frag = 1;
 	ieee->ieee802_1x = 1;	/* Default to supporting 802.1x */
 
 	INIT_LIST_HEAD(&ieee->crypt_deinit_list);
 	init_timer(&ieee->crypt_deinit_timer);
 	ieee->crypt_deinit_timer.data = (unsigned long)ieee;
 	ieee->crypt_deinit_timer.function = ieee80211_crypt_deinit_handler;
+	ieee->crypt_quiesced = 0;
 
 	spin_lock_init(&ieee->lock);
 
 	ieee->wpa_enabled = 0;
-	ieee->tkip_countermeasures = 0;
 	ieee->drop_unencrypted = 0;
 	ieee->privacy_invoked = 0;
-	ieee->ieee802_1x = 1;
 
 	return dev;
 
@@ -161,6 +172,7 @@ void free_ieee80211(struct net_device *dev)
 
 	int i;
 
+	ieee80211_crypt_quiescing(ieee);
 	del_timer_sync(&ieee->crypt_deinit_timer);
 	ieee80211_crypt_deinit_entries(ieee, 1);
 
@@ -195,42 +207,30 @@ static int show_debug_level(char *page, char **start, off_t offset,
 static int store_debug_level(struct file *file, const char __user * buffer,
 			     unsigned long count, void *data)
 {
-	char buf[] = "0x00000000";
-	char *p = (char *)buf;
+	char buf[] = "0x00000000\n";
+	unsigned long len = min((unsigned long)sizeof(buf) - 1, count);
 	unsigned long val;
 
-	if (count > sizeof(buf) - 1)
-		count = sizeof(buf) - 1;
-
-	if (copy_from_user(buf, buffer, count))
+	if (copy_from_user(buf, buffer, len))
 		return count;
-	buf[count] = 0;
-	/*
-	 * what a FPOS...  What, sscanf(buf, "%i", &val) would be too
-	 * scary?
-	 */
-	if (p[1] == 'x' || p[1] == 'X' || p[0] == 'x' || p[0] == 'X') {
-		p++;
-		if (p[0] == 'x' || p[0] == 'X')
-			p++;
-		val = simple_strtoul(p, &p, 16);
-	} else
-		val = simple_strtoul(p, &p, 10);
-	if (p == buf)
+	buf[len] = 0;
+	if (sscanf(buf, "%li", &val) != 1)
 		printk(KERN_INFO DRV_NAME
 		       ": %s is not in hex or decimal form.\n", buf);
 	else
 		ieee80211_debug_level = val;
 
-	return strlen(buf);
+	return strnlen(buf, len);
 }
+#endif				/* CONFIG_IEEE80211_DEBUG */
 
 static int __init ieee80211_init(void)
 {
+#ifdef CONFIG_IEEE80211_DEBUG
 	struct proc_dir_entry *e;
 
 	ieee80211_debug_level = debug;
-	ieee80211_proc = create_proc_entry(DRV_NAME, S_IFDIR, proc_net);
+	ieee80211_proc = proc_mkdir(DRV_NAME, proc_net);
 	if (ieee80211_proc == NULL) {
 		IEEE80211_ERROR("Unable to create " DRV_NAME
 				" proc directory\n");
@@ -246,26 +246,33 @@ static int __init ieee80211_init(void)
 	e->read_proc = show_debug_level;
 	e->write_proc = store_debug_level;
 	e->data = NULL;
+#endif				/* CONFIG_IEEE80211_DEBUG */
+
+	printk(KERN_INFO DRV_NAME ": " DRV_DESCRIPTION ", " DRV_VERSION "\n");
+	printk(KERN_INFO DRV_NAME ": " DRV_COPYRIGHT "\n");
 
 	return 0;
 }
 
 static void __exit ieee80211_exit(void)
 {
+#ifdef CONFIG_IEEE80211_DEBUG
 	if (ieee80211_proc) {
 		remove_proc_entry("debug_level", ieee80211_proc);
 		remove_proc_entry(DRV_NAME, proc_net);
 		ieee80211_proc = NULL;
 	}
+#endif				/* CONFIG_IEEE80211_DEBUG */
 }
 
+#ifdef CONFIG_IEEE80211_DEBUG
 #include <linux/moduleparam.h>
 module_param(debug, int, 0444);
 MODULE_PARM_DESC(debug, "debug output mask");
+#endif				/* CONFIG_IEEE80211_DEBUG */
 
 module_exit(ieee80211_exit);
 module_init(ieee80211_init);
-#endif
 
 const char *escape_essid(const char *essid, u8 essid_len)
 {
diff --git a/net/ieee80211/ieee80211_rx.c b/net/ieee80211/ieee80211_rx.c
index f7dcd854139..ce694cf5c16 100644
--- a/net/ieee80211/ieee80211_rx.c
+++ b/net/ieee80211/ieee80211_rx.c
@@ -5,7 +5,7 @@
  * Copyright (c) 2001-2002, SSH Communications Security Corp and Jouni Malinen
  * <jkmaline@cc.hut.fi>
  * Copyright (c) 2002-2003, Jouni Malinen <jkmaline@cc.hut.fi>
- * Copyright (c) 2004, Intel Corporation
+ * Copyright (c) 2004-2005, Intel Corporation
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License version 2 as
@@ -87,7 +87,7 @@ static struct ieee80211_frag_entry *ieee80211_frag_cache_find(struct
 
 /* Called only as a tasklet (software IRQ) */
 static struct sk_buff *ieee80211_frag_cache_get(struct ieee80211_device *ieee,
-						struct ieee80211_hdr *hdr)
+						struct ieee80211_hdr_4addr *hdr)
 {
 	struct sk_buff *skb = NULL;
 	u16 sc;
@@ -101,7 +101,7 @@ static struct sk_buff *ieee80211_frag_cache_get(struct ieee80211_device *ieee,
 	if (frag == 0) {
 		/* Reserve enough space to fit maximum frame length */
 		skb = dev_alloc_skb(ieee->dev->mtu +
-				    sizeof(struct ieee80211_hdr) +
+				    sizeof(struct ieee80211_hdr_4addr) +
 				    8 /* LLC */  +
 				    2 /* alignment */  +
 				    8 /* WEP */  + ETH_ALEN /* WDS */ );
@@ -138,7 +138,7 @@ static struct sk_buff *ieee80211_frag_cache_get(struct ieee80211_device *ieee,
 
 /* Called only as a tasklet (software IRQ) */
 static int ieee80211_frag_cache_invalidate(struct ieee80211_device *ieee,
-					   struct ieee80211_hdr *hdr)
+					   struct ieee80211_hdr_4addr *hdr)
 {
 	u16 sc;
 	unsigned int seq;
@@ -176,7 +176,7 @@ ieee80211_rx_frame_mgmt(struct ieee80211_device *ieee, struct sk_buff *skb,
 		       ieee->dev->name);
 		return 0;
 /*
-  hostap_update_sta_ps(ieee, (struct hostap_ieee80211_hdr *)
+  hostap_update_sta_ps(ieee, (struct hostap_ieee80211_hdr_4addr *)
   skb->data);*/
 	}
 
@@ -232,13 +232,13 @@ static int ieee80211_is_eapol_frame(struct ieee80211_device *ieee,
 {
 	struct net_device *dev = ieee->dev;
 	u16 fc, ethertype;
-	struct ieee80211_hdr *hdr;
+	struct ieee80211_hdr_3addr *hdr;
 	u8 *pos;
 
 	if (skb->len < 24)
 		return 0;
 
-	hdr = (struct ieee80211_hdr *)skb->data;
+	hdr = (struct ieee80211_hdr_3addr *)skb->data;
 	fc = le16_to_cpu(hdr->frame_ctl);
 
 	/* check that the frame is unicast frame to us */
@@ -271,26 +271,15 @@ static inline int
 ieee80211_rx_frame_decrypt(struct ieee80211_device *ieee, struct sk_buff *skb,
 			   struct ieee80211_crypt_data *crypt)
 {
-	struct ieee80211_hdr *hdr;
+	struct ieee80211_hdr_3addr *hdr;
 	int res, hdrlen;
 
 	if (crypt == NULL || crypt->ops->decrypt_mpdu == NULL)
 		return 0;
 
-	hdr = (struct ieee80211_hdr *)skb->data;
+	hdr = (struct ieee80211_hdr_3addr *)skb->data;
 	hdrlen = ieee80211_get_hdrlen(le16_to_cpu(hdr->frame_ctl));
 
-#ifdef CONFIG_IEEE80211_CRYPT_TKIP
-	if (ieee->tkip_countermeasures && strcmp(crypt->ops->name, "TKIP") == 0) {
-		if (net_ratelimit()) {
-			printk(KERN_DEBUG "%s: TKIP countermeasures: dropped "
-			       "received packet from " MAC_FMT "\n",
-			       ieee->dev->name, MAC_ARG(hdr->addr2));
-		}
-		return -1;
-	}
-#endif
-
 	atomic_inc(&crypt->refcnt);
 	res = crypt->ops->decrypt_mpdu(skb, hdrlen, crypt->priv);
 	atomic_dec(&crypt->refcnt);
@@ -314,13 +303,13 @@ ieee80211_rx_frame_decrypt_msdu(struct ieee80211_device *ieee,
 				struct sk_buff *skb, int keyidx,
 				struct ieee80211_crypt_data *crypt)
 {
-	struct ieee80211_hdr *hdr;
+	struct ieee80211_hdr_3addr *hdr;
 	int res, hdrlen;
 
 	if (crypt == NULL || crypt->ops->decrypt_msdu == NULL)
 		return 0;
 
-	hdr = (struct ieee80211_hdr *)skb->data;
+	hdr = (struct ieee80211_hdr_3addr *)skb->data;
 	hdrlen = ieee80211_get_hdrlen(le16_to_cpu(hdr->frame_ctl));
 
 	atomic_inc(&crypt->refcnt);
@@ -343,7 +332,7 @@ int ieee80211_rx(struct ieee80211_device *ieee, struct sk_buff *skb,
 		 struct ieee80211_rx_stats *rx_stats)
 {
 	struct net_device *dev = ieee->dev;
-	struct ieee80211_hdr *hdr;
+	struct ieee80211_hdr_4addr *hdr;
 	size_t hdrlen;
 	u16 fc, type, stype, sc;
 	struct net_device_stats *stats;
@@ -363,7 +352,7 @@ int ieee80211_rx(struct ieee80211_device *ieee, struct sk_buff *skb,
 	struct ieee80211_crypt_data *crypt = NULL;
 	int keyidx = 0;
 
-	hdr = (struct ieee80211_hdr *)skb->data;
+	hdr = (struct ieee80211_hdr_4addr *)skb->data;
 	stats = &ieee->stats;
 
 	if (skb->len < 10) {
@@ -378,35 +367,51 @@ int ieee80211_rx(struct ieee80211_device *ieee, struct sk_buff *skb,
 	frag = WLAN_GET_SEQ_FRAG(sc);
 	hdrlen = ieee80211_get_hdrlen(fc);
 
-#ifdef NOT_YET
-#if WIRELESS_EXT > 15
 	/* Put this code here so that we avoid duplicating it in all
 	 * Rx paths. - Jean II */
 #ifdef IW_WIRELESS_SPY		/* defined in iw_handler.h */
 	/* If spy monitoring on */
-	if (iface->spy_data.spy_number > 0) {
+	if (ieee->spy_data.spy_number > 0) {
 		struct iw_quality wstats;
-		wstats.level = rx_stats->signal;
-		wstats.noise = rx_stats->noise;
-		wstats.updated = 6;	/* No qual value */
+
+		wstats.updated = 0;
+		if (rx_stats->mask & IEEE80211_STATMASK_RSSI) {
+			wstats.level = rx_stats->rssi;
+			wstats.updated |= IW_QUAL_LEVEL_UPDATED;
+		} else
+			wstats.updated |= IW_QUAL_LEVEL_INVALID;
+
+		if (rx_stats->mask & IEEE80211_STATMASK_NOISE) {
+			wstats.noise = rx_stats->noise;
+			wstats.updated |= IW_QUAL_NOISE_UPDATED;
+		} else
+			wstats.updated |= IW_QUAL_NOISE_INVALID;
+
+		if (rx_stats->mask & IEEE80211_STATMASK_SIGNAL) {
+			wstats.qual = rx_stats->signal;
+			wstats.updated |= IW_QUAL_QUAL_UPDATED;
+		} else
+			wstats.updated |= IW_QUAL_QUAL_INVALID;
+
 		/* Update spy records */
-		wireless_spy_update(dev, hdr->addr2, &wstats);
+		wireless_spy_update(ieee->dev, hdr->addr2, &wstats);
 	}
 #endif				/* IW_WIRELESS_SPY */
-#endif				/* WIRELESS_EXT > 15 */
+
+#ifdef NOT_YET
 	hostap_update_rx_stats(local->ap, hdr, rx_stats);
 #endif
 
-#if WIRELESS_EXT > 15
 	if (ieee->iw_mode == IW_MODE_MONITOR) {
 		ieee80211_monitor_rx(ieee, skb, rx_stats);
 		stats->rx_packets++;
 		stats->rx_bytes += skb->len;
 		return 1;
 	}
-#endif
 
-	if (ieee->host_decrypt) {
+	if ((is_multicast_ether_addr(hdr->addr1) ||
+	     is_broadcast_ether_addr(hdr->addr2)) ? ieee->host_mc_decrypt :
+	    ieee->host_decrypt) {
 		int idx = 0;
 		if (skb->len >= hdrlen + 3)
 			idx = skb->data[hdrlen + 3] >> 6;
@@ -531,6 +536,9 @@ int ieee80211_rx(struct ieee80211_device *ieee, struct sk_buff *skb,
 
 	/* Nullfunc frames may have PS-bit set, so they must be passed to
 	 * hostap_handle_sta_rx() before being dropped here. */
+
+	stype &= ~IEEE80211_STYPE_QOS_DATA;
+
 	if (stype != IEEE80211_STYPE_DATA &&
 	    stype != IEEE80211_STYPE_DATA_CFACK &&
 	    stype != IEEE80211_STYPE_DATA_CFPOLL &&
@@ -549,7 +557,7 @@ int ieee80211_rx(struct ieee80211_device *ieee, struct sk_buff *skb,
 	    (keyidx = ieee80211_rx_frame_decrypt(ieee, skb, crypt)) < 0)
 		goto rx_dropped;
 
-	hdr = (struct ieee80211_hdr *)skb->data;
+	hdr = (struct ieee80211_hdr_4addr *)skb->data;
 
 	/* skb: hdr + (possibly fragmented) plaintext payload */
 	// PR: FIXME: hostap has additional conditions in the "if" below:
@@ -603,7 +611,7 @@ int ieee80211_rx(struct ieee80211_device *ieee, struct sk_buff *skb,
 		/* this was the last fragment and the frame will be
 		 * delivered, so remove skb from fragment cache */
 		skb = frag_skb;
-		hdr = (struct ieee80211_hdr *)skb->data;
+		hdr = (struct ieee80211_hdr_4addr *)skb->data;
 		ieee80211_frag_cache_invalidate(ieee, hdr);
 	}
 
@@ -613,7 +621,7 @@ int ieee80211_rx(struct ieee80211_device *ieee, struct sk_buff *skb,
 	    ieee80211_rx_frame_decrypt_msdu(ieee, skb, keyidx, crypt))
 		goto rx_dropped;
 
-	hdr = (struct ieee80211_hdr *)skb->data;
+	hdr = (struct ieee80211_hdr_4addr *)skb->data;
 	if (crypt && !(fc & IEEE80211_FCTL_PROTECTED) && !ieee->open_wep) {
 		if (		/*ieee->ieee802_1x && */
 			   ieee80211_is_eapol_frame(ieee, skb)) {
@@ -755,69 +763,179 @@ int ieee80211_rx(struct ieee80211_device *ieee, struct sk_buff *skb,
 
 #define MGMT_FRAME_FIXED_PART_LENGTH		0x24
 
-static inline int ieee80211_is_ofdm_rate(u8 rate)
+static u8 qos_oui[QOS_OUI_LEN] = { 0x00, 0x50, 0xF2 };
+
+/*
+* Make ther structure we read from the beacon packet has
+* the right values
+*/
+static int ieee80211_verify_qos_info(struct ieee80211_qos_information_element
+				     *info_element, int sub_type)
 {
-	switch (rate & ~IEEE80211_BASIC_RATE_MASK) {
-	case IEEE80211_OFDM_RATE_6MB:
-	case IEEE80211_OFDM_RATE_9MB:
-	case IEEE80211_OFDM_RATE_12MB:
-	case IEEE80211_OFDM_RATE_18MB:
-	case IEEE80211_OFDM_RATE_24MB:
-	case IEEE80211_OFDM_RATE_36MB:
-	case IEEE80211_OFDM_RATE_48MB:
-	case IEEE80211_OFDM_RATE_54MB:
-		return 1;
-	}
+
+	if (info_element->qui_subtype != sub_type)
+		return -1;
+	if (memcmp(info_element->qui, qos_oui, QOS_OUI_LEN))
+		return -1;
+	if (info_element->qui_type != QOS_OUI_TYPE)
+		return -1;
+	if (info_element->version != QOS_VERSION_1)
+		return -1;
+
 	return 0;
 }
 
-static inline int ieee80211_network_init(struct ieee80211_device *ieee,
-					 struct ieee80211_probe_response
-					 *beacon,
-					 struct ieee80211_network *network,
-					 struct ieee80211_rx_stats *stats)
+/*
+ * Parse a QoS parameter element
+ */
+static int ieee80211_read_qos_param_element(struct ieee80211_qos_parameter_info
+					    *element_param, struct ieee80211_info_element
+					    *info_element)
 {
-#ifdef CONFIG_IEEE80211_DEBUG
-	char rates_str[64];
-	char *p;
-#endif
-	struct ieee80211_info_element *info_element;
-	u16 left;
-	u8 i;
+	int ret = 0;
+	u16 size = sizeof(struct ieee80211_qos_parameter_info) - 2;
 
-	/* Pull out fixed field data */
-	memcpy(network->bssid, beacon->header.addr3, ETH_ALEN);
-	network->capability = beacon->capability;
-	network->last_scanned = jiffies;
-	network->time_stamp[0] = beacon->time_stamp[0];
-	network->time_stamp[1] = beacon->time_stamp[1];
-	network->beacon_interval = beacon->beacon_interval;
-	/* Where to pull this? beacon->listen_interval; */
-	network->listen_interval = 0x0A;
-	network->rates_len = network->rates_ex_len = 0;
-	network->last_associate = 0;
-	network->ssid_len = 0;
-	network->flags = 0;
-	network->atim_window = 0;
+	if ((info_element == NULL) || (element_param == NULL))
+		return -1;
 
-	if (stats->freq == IEEE80211_52GHZ_BAND) {
-		/* for A band (No DS info) */
-		network->channel = stats->received_channel;
+	if (info_element->id == QOS_ELEMENT_ID && info_element->len == size) {
+		memcpy(element_param->info_element.qui, info_element->data,
+		       info_element->len);
+		element_param->info_element.elementID = info_element->id;
+		element_param->info_element.length = info_element->len;
 	} else
-		network->flags |= NETWORK_HAS_CCK;
+		ret = -1;
+	if (ret == 0)
+		ret = ieee80211_verify_qos_info(&element_param->info_element,
+						QOS_OUI_PARAM_SUB_TYPE);
+	return ret;
+}
 
-	network->wpa_ie_len = 0;
-	network->rsn_ie_len = 0;
+/*
+ * Parse a QoS information element
+ */
+static int ieee80211_read_qos_info_element(struct
+					   ieee80211_qos_information_element
+					   *element_info, struct ieee80211_info_element
+					   *info_element)
+{
+	int ret = 0;
+	u16 size = sizeof(struct ieee80211_qos_information_element) - 2;
+
+	if (element_info == NULL)
+		return -1;
+	if (info_element == NULL)
+		return -1;
+
+	if ((info_element->id == QOS_ELEMENT_ID) && (info_element->len == size)) {
+		memcpy(element_info->qui, info_element->data,
+		       info_element->len);
+		element_info->elementID = info_element->id;
+		element_info->length = info_element->len;
+	} else
+		ret = -1;
+
+	if (ret == 0)
+		ret = ieee80211_verify_qos_info(element_info,
+						QOS_OUI_INFO_SUB_TYPE);
+	return ret;
+}
+
+/*
+ * Write QoS parameters from the ac parameters.
+ */
+static int ieee80211_qos_convert_ac_to_parameters(struct
+						  ieee80211_qos_parameter_info
+						  *param_elm, struct
+						  ieee80211_qos_parameters
+						  *qos_param)
+{
+	int rc = 0;
+	int i;
+	struct ieee80211_qos_ac_parameter *ac_params;
+	u32 txop;
+	u8 cw_min;
+	u8 cw_max;
+
+	for (i = 0; i < QOS_QUEUE_NUM; i++) {
+		ac_params = &(param_elm->ac_params_record[i]);
+
+		qos_param->aifs[i] = (ac_params->aci_aifsn) & 0x0F;
+		qos_param->aifs[i] -= (qos_param->aifs[i] < 2) ? 0 : 2;
+
+		cw_min = ac_params->ecw_min_max & 0x0F;
+		qos_param->cw_min[i] = (u16) ((1 << cw_min) - 1);
+
+		cw_max = (ac_params->ecw_min_max & 0xF0) >> 4;
+		qos_param->cw_max[i] = (u16) ((1 << cw_max) - 1);
+
+		qos_param->flag[i] =
+		    (ac_params->aci_aifsn & 0x10) ? 0x01 : 0x00;
+
+		txop = le16_to_cpu(ac_params->tx_op_limit) * 32;
+		qos_param->tx_op_limit[i] = (u16) txop;
+	}
+	return rc;
+}
+
+/*
+ * we have a generic data element which it may contain QoS information or
+ * parameters element. check the information element length to decide
+ * which type to read
+ */
+static int ieee80211_parse_qos_info_param_IE(struct ieee80211_info_element
+					     *info_element,
+					     struct ieee80211_network *network)
+{
+	int rc = 0;
+	struct ieee80211_qos_parameters *qos_param = NULL;
+	struct ieee80211_qos_information_element qos_info_element;
+
+	rc = ieee80211_read_qos_info_element(&qos_info_element, info_element);
+
+	if (rc == 0) {
+		network->qos_data.param_count = qos_info_element.ac_info & 0x0F;
+		network->flags |= NETWORK_HAS_QOS_INFORMATION;
+	} else {
+		struct ieee80211_qos_parameter_info param_element;
+
+		rc = ieee80211_read_qos_param_element(&param_element,
+						      info_element);
+		if (rc == 0) {
+			qos_param = &(network->qos_data.parameters);
+			ieee80211_qos_convert_ac_to_parameters(&param_element,
+							       qos_param);
+			network->flags |= NETWORK_HAS_QOS_PARAMETERS;
+			network->qos_data.param_count =
+			    param_element.info_element.ac_info & 0x0F;
+		}
+	}
+
+	if (rc == 0) {
+		IEEE80211_DEBUG_QOS("QoS is supported\n");
+		network->qos_data.supported = 1;
+	}
+	return rc;
+}
+
+static int ieee80211_parse_info_param(struct ieee80211_info_element
+				      *info_element, u16 length,
+				      struct ieee80211_network *network)
+{
+	u8 i;
+#ifdef CONFIG_IEEE80211_DEBUG
+	char rates_str[64];
+	char *p;
+#endif
 
-	info_element = &beacon->info_element;
-	left = stats->len - ((void *)info_element - (void *)beacon);
-	while (left >= sizeof(struct ieee80211_info_element_hdr)) {
-		if (sizeof(struct ieee80211_info_element_hdr) +
-		    info_element->len > left) {
-			IEEE80211_DEBUG_SCAN
-			    ("SCAN: parse failed: info_element->len + 2 > left : info_element->len+2=%Zd left=%d.\n",
-			     info_element->len +
-			     sizeof(struct ieee80211_info_element), left);
+	while (length >= sizeof(*info_element)) {
+		if (sizeof(*info_element) + info_element->len > length) {
+			IEEE80211_DEBUG_MGMT("Info elem: parse failed: "
+					     "info_element->len + 2 > left : "
+					     "info_element->len+2=%zd left=%d, id=%d.\n",
+					     info_element->len +
+					     sizeof(*info_element),
+					     length, info_element->id);
 			return 1;
 		}
 
@@ -837,7 +955,7 @@ static inline int ieee80211_network_init(struct ieee80211_device *ieee,
 				memset(network->ssid + network->ssid_len, 0,
 				       IW_ESSID_MAX_SIZE - network->ssid_len);
 
-			IEEE80211_DEBUG_SCAN("MFIE_TYPE_SSID: '%s' len=%d.\n",
+			IEEE80211_DEBUG_MGMT("MFIE_TYPE_SSID: '%s' len=%d.\n",
 					     network->ssid, network->ssid_len);
 			break;
 
@@ -845,15 +963,14 @@ static inline int ieee80211_network_init(struct ieee80211_device *ieee,
 #ifdef CONFIG_IEEE80211_DEBUG
 			p = rates_str;
 #endif
-			network->rates_len =
-			    min(info_element->len, MAX_RATES_LENGTH);
+			network->rates_len = min(info_element->len,
+						 MAX_RATES_LENGTH);
 			for (i = 0; i < network->rates_len; i++) {
 				network->rates[i] = info_element->data[i];
 #ifdef CONFIG_IEEE80211_DEBUG
-				p += snprintf(p,
-					      sizeof(rates_str) - (p -
-								   rates_str),
-					      "%02X ", network->rates[i]);
+				p += snprintf(p, sizeof(rates_str) -
+					      (p - rates_str), "%02X ",
+					      network->rates[i]);
 #endif
 				if (ieee80211_is_ofdm_rate
 				    (info_element->data[i])) {
@@ -865,7 +982,7 @@ static inline int ieee80211_network_init(struct ieee80211_device *ieee,
 				}
 			}
 
-			IEEE80211_DEBUG_SCAN("MFIE_TYPE_RATES: '%s' (%d)\n",
+			IEEE80211_DEBUG_MGMT("MFIE_TYPE_RATES: '%s' (%d)\n",
 					     rates_str, network->rates_len);
 			break;
 
@@ -873,15 +990,14 @@ static inline int ieee80211_network_init(struct ieee80211_device *ieee,
 #ifdef CONFIG_IEEE80211_DEBUG
 			p = rates_str;
 #endif
-			network->rates_ex_len =
-			    min(info_element->len, MAX_RATES_EX_LENGTH);
+			network->rates_ex_len = min(info_element->len,
+						    MAX_RATES_EX_LENGTH);
 			for (i = 0; i < network->rates_ex_len; i++) {
 				network->rates_ex[i] = info_element->data[i];
 #ifdef CONFIG_IEEE80211_DEBUG
-				p += snprintf(p,
-					      sizeof(rates_str) - (p -
-								   rates_str),
-					      "%02X ", network->rates[i]);
+				p += snprintf(p, sizeof(rates_str) -
+					      (p - rates_str), "%02X ",
+					      network->rates[i]);
 #endif
 				if (ieee80211_is_ofdm_rate
 				    (info_element->data[i])) {
@@ -893,40 +1009,51 @@ static inline int ieee80211_network_init(struct ieee80211_device *ieee,
 				}
 			}
 
-			IEEE80211_DEBUG_SCAN("MFIE_TYPE_RATES_EX: '%s' (%d)\n",
+			IEEE80211_DEBUG_MGMT("MFIE_TYPE_RATES_EX: '%s' (%d)\n",
 					     rates_str, network->rates_ex_len);
 			break;
 
 		case MFIE_TYPE_DS_SET:
-			IEEE80211_DEBUG_SCAN("MFIE_TYPE_DS_SET: %d\n",
+			IEEE80211_DEBUG_MGMT("MFIE_TYPE_DS_SET: %d\n",
 					     info_element->data[0]);
-			if (stats->freq == IEEE80211_24GHZ_BAND)
-				network->channel = info_element->data[0];
+			network->channel = info_element->data[0];
 			break;
 
 		case MFIE_TYPE_FH_SET:
-			IEEE80211_DEBUG_SCAN("MFIE_TYPE_FH_SET: ignored\n");
+			IEEE80211_DEBUG_MGMT("MFIE_TYPE_FH_SET: ignored\n");
 			break;
 
 		case MFIE_TYPE_CF_SET:
-			IEEE80211_DEBUG_SCAN("MFIE_TYPE_CF_SET: ignored\n");
+			IEEE80211_DEBUG_MGMT("MFIE_TYPE_CF_SET: ignored\n");
 			break;
 
 		case MFIE_TYPE_TIM:
-			IEEE80211_DEBUG_SCAN("MFIE_TYPE_TIM: ignored\n");
+			IEEE80211_DEBUG_MGMT("MFIE_TYPE_TIM: ignored\n");
+			break;
+
+		case MFIE_TYPE_ERP_INFO:
+			network->erp_value = info_element->data[0];
+			IEEE80211_DEBUG_MGMT("MFIE_TYPE_ERP_SET: %d\n",
+					     network->erp_value);
 			break;
 
 		case MFIE_TYPE_IBSS_SET:
-			IEEE80211_DEBUG_SCAN("MFIE_TYPE_IBSS_SET: ignored\n");
+			network->atim_window = info_element->data[0];
+			IEEE80211_DEBUG_MGMT("MFIE_TYPE_IBSS_SET: %d\n",
+					     network->atim_window);
 			break;
 
 		case MFIE_TYPE_CHALLENGE:
-			IEEE80211_DEBUG_SCAN("MFIE_TYPE_CHALLENGE: ignored\n");
+			IEEE80211_DEBUG_MGMT("MFIE_TYPE_CHALLENGE: ignored\n");
 			break;
 
 		case MFIE_TYPE_GENERIC:
-			IEEE80211_DEBUG_SCAN("MFIE_TYPE_GENERIC: %d bytes\n",
+			IEEE80211_DEBUG_MGMT("MFIE_TYPE_GENERIC: %d bytes\n",
 					     info_element->len);
+			if (!ieee80211_parse_qos_info_param_IE(info_element,
+							       network))
+				break;
+
 			if (info_element->len >= 4 &&
 			    info_element->data[0] == 0x00 &&
 			    info_element->data[1] == 0x50 &&
@@ -940,7 +1067,7 @@ static inline int ieee80211_network_init(struct ieee80211_device *ieee,
 			break;
 
 		case MFIE_TYPE_RSN:
-			IEEE80211_DEBUG_SCAN("MFIE_TYPE_RSN: %d bytes\n",
+			IEEE80211_DEBUG_MGMT("MFIE_TYPE_RSN: %d bytes\n",
 					     info_element->len);
 			network->rsn_ie_len = min(info_element->len + 2,
 						  MAX_WPA_IE_LEN);
@@ -948,18 +1075,127 @@ static inline int ieee80211_network_init(struct ieee80211_device *ieee,
 			       network->rsn_ie_len);
 			break;
 
+		case MFIE_TYPE_QOS_PARAMETER:
+			printk(KERN_ERR
+			       "QoS Error need to parse QOS_PARAMETER IE\n");
+			break;
+
 		default:
-			IEEE80211_DEBUG_SCAN("unsupported IE %d\n",
+			IEEE80211_DEBUG_MGMT("unsupported IE %d\n",
 					     info_element->id);
 			break;
 		}
 
-		left -= sizeof(struct ieee80211_info_element_hdr) +
-		    info_element->len;
-		info_element = (struct ieee80211_info_element *)
-		    &info_element->data[info_element->len];
+		length -= sizeof(*info_element) + info_element->len;
+		info_element =
+		    (struct ieee80211_info_element *)&info_element->
+		    data[info_element->len];
+	}
+
+	return 0;
+}
+
+static int ieee80211_handle_assoc_resp(struct ieee80211_device *ieee, struct ieee80211_assoc_response
+				       *frame, struct ieee80211_rx_stats *stats)
+{
+	struct ieee80211_network network_resp;
+	struct ieee80211_network *network = &network_resp;
+	struct net_device *dev = ieee->dev;
+
+	network->flags = 0;
+	network->qos_data.active = 0;
+	network->qos_data.supported = 0;
+	network->qos_data.param_count = 0;
+	network->qos_data.old_param_count = 0;
+
+	//network->atim_window = le16_to_cpu(frame->aid) & (0x3FFF);
+	network->atim_window = le16_to_cpu(frame->aid);
+	network->listen_interval = le16_to_cpu(frame->status);
+	memcpy(network->bssid, frame->header.addr3, ETH_ALEN);
+	network->capability = le16_to_cpu(frame->capability);
+	network->last_scanned = jiffies;
+	network->rates_len = network->rates_ex_len = 0;
+	network->last_associate = 0;
+	network->ssid_len = 0;
+	network->erp_value =
+	    (network->capability & WLAN_CAPABILITY_IBSS) ? 0x3 : 0x0;
+
+	if (stats->freq == IEEE80211_52GHZ_BAND) {
+		/* for A band (No DS info) */
+		network->channel = stats->received_channel;
+	} else
+		network->flags |= NETWORK_HAS_CCK;
+
+	network->wpa_ie_len = 0;
+	network->rsn_ie_len = 0;
+
+	if (ieee80211_parse_info_param
+	    (frame->info_element, stats->len - sizeof(*frame), network))
+		return 1;
+
+	network->mode = 0;
+	if (stats->freq == IEEE80211_52GHZ_BAND)
+		network->mode = IEEE_A;
+	else {
+		if (network->flags & NETWORK_HAS_OFDM)
+			network->mode |= IEEE_G;
+		if (network->flags & NETWORK_HAS_CCK)
+			network->mode |= IEEE_B;
 	}
 
+	if (ieee80211_is_empty_essid(network->ssid, network->ssid_len))
+		network->flags |= NETWORK_EMPTY_ESSID;
+
+	memcpy(&network->stats, stats, sizeof(network->stats));
+
+	if (ieee->handle_assoc_response != NULL)
+		ieee->handle_assoc_response(dev, frame, network);
+
+	return 0;
+}
+
+/***************************************************/
+
+static inline int ieee80211_network_init(struct ieee80211_device *ieee, struct ieee80211_probe_response
+					 *beacon,
+					 struct ieee80211_network *network,
+					 struct ieee80211_rx_stats *stats)
+{
+	network->qos_data.active = 0;
+	network->qos_data.supported = 0;
+	network->qos_data.param_count = 0;
+	network->qos_data.old_param_count = 0;
+
+	/* Pull out fixed field data */
+	memcpy(network->bssid, beacon->header.addr3, ETH_ALEN);
+	network->capability = le16_to_cpu(beacon->capability);
+	network->last_scanned = jiffies;
+	network->time_stamp[0] = le32_to_cpu(beacon->time_stamp[0]);
+	network->time_stamp[1] = le32_to_cpu(beacon->time_stamp[1]);
+	network->beacon_interval = le16_to_cpu(beacon->beacon_interval);
+	/* Where to pull this? beacon->listen_interval; */
+	network->listen_interval = 0x0A;
+	network->rates_len = network->rates_ex_len = 0;
+	network->last_associate = 0;
+	network->ssid_len = 0;
+	network->flags = 0;
+	network->atim_window = 0;
+	network->erp_value = (network->capability & WLAN_CAPABILITY_IBSS) ?
+	    0x3 : 0x0;
+
+	if (stats->freq == IEEE80211_52GHZ_BAND) {
+		/* for A band (No DS info) */
+		network->channel = stats->received_channel;
+	} else
+		network->flags |= NETWORK_HAS_CCK;
+
+	network->wpa_ie_len = 0;
+	network->rsn_ie_len = 0;
+
+	if (ieee80211_parse_info_param
+	    (beacon->info_element, stats->len - sizeof(*beacon), network))
+		return 1;
+
 	network->mode = 0;
 	if (stats->freq == IEEE80211_52GHZ_BAND)
 		network->mode = IEEE_A;
@@ -1002,6 +1238,9 @@ static inline int is_same_network(struct ieee80211_network *src,
 static inline void update_network(struct ieee80211_network *dst,
 				  struct ieee80211_network *src)
 {
+	int qos_active;
+	u8 old_param;
+
 	memcpy(&dst->stats, &src->stats, sizeof(struct ieee80211_rx_stats));
 	dst->capability = src->capability;
 	memcpy(dst->rates, src->rates, src->rates_len);
@@ -1017,6 +1256,7 @@ static inline void update_network(struct ieee80211_network *dst,
 	dst->beacon_interval = src->beacon_interval;
 	dst->listen_interval = src->listen_interval;
 	dst->atim_window = src->atim_window;
+	dst->erp_value = src->erp_value;
 
 	memcpy(dst->wpa_ie, src->wpa_ie, src->wpa_ie_len);
 	dst->wpa_ie_len = src->wpa_ie_len;
@@ -1024,22 +1264,48 @@ static inline void update_network(struct ieee80211_network *dst,
 	dst->rsn_ie_len = src->rsn_ie_len;
 
 	dst->last_scanned = jiffies;
+	qos_active = src->qos_data.active;
+	old_param = dst->qos_data.old_param_count;
+	if (dst->flags & NETWORK_HAS_QOS_MASK)
+		memcpy(&dst->qos_data, &src->qos_data,
+		       sizeof(struct ieee80211_qos_data));
+	else {
+		dst->qos_data.supported = src->qos_data.supported;
+		dst->qos_data.param_count = src->qos_data.param_count;
+	}
+
+	if (dst->qos_data.supported == 1) {
+		if (dst->ssid_len)
+			IEEE80211_DEBUG_QOS
+			    ("QoS the network %s is QoS supported\n",
+			     dst->ssid);
+		else
+			IEEE80211_DEBUG_QOS
+			    ("QoS the network is QoS supported\n");
+	}
+	dst->qos_data.active = qos_active;
+	dst->qos_data.old_param_count = old_param;
+
 	/* dst->last_associate is not overwritten */
 }
 
+static inline int is_beacon(int fc)
+{
+	return (WLAN_FC_GET_STYPE(le16_to_cpu(fc)) == IEEE80211_STYPE_BEACON);
+}
+
 static inline void ieee80211_process_probe_response(struct ieee80211_device
-						    *ieee,
-						    struct
+						    *ieee, struct
 						    ieee80211_probe_response
-						    *beacon,
-						    struct ieee80211_rx_stats
+						    *beacon, struct ieee80211_rx_stats
 						    *stats)
 {
+	struct net_device *dev = ieee->dev;
 	struct ieee80211_network network;
 	struct ieee80211_network *target;
 	struct ieee80211_network *oldest = NULL;
 #ifdef CONFIG_IEEE80211_DEBUG
-	struct ieee80211_info_element *info_element = &beacon->info_element;
+	struct ieee80211_info_element *info_element = beacon->info_element;
 #endif
 	unsigned long flags;
 
@@ -1070,10 +1336,10 @@ static inline void ieee80211_process_probe_response(struct ieee80211_device
 				     escape_essid(info_element->data,
 						  info_element->len),
 				     MAC_ARG(beacon->header.addr3),
-				     WLAN_FC_GET_STYPE(beacon->header.
-						       frame_ctl) ==
-				     IEEE80211_STYPE_PROBE_RESP ?
-				     "PROBE RESPONSE" : "BEACON");
+				     is_beacon(le16_to_cpu
+					       (beacon->header.
+						frame_ctl)) ?
+				     "BEACON" : "PROBE RESPONSE");
 		return;
 	}
 
@@ -1122,10 +1388,10 @@ static inline void ieee80211_process_probe_response(struct ieee80211_device
 				     escape_essid(network.ssid,
 						  network.ssid_len),
 				     MAC_ARG(network.bssid),
-				     WLAN_FC_GET_STYPE(beacon->header.
-						       frame_ctl) ==
-				     IEEE80211_STYPE_PROBE_RESP ?
-				     "PROBE RESPONSE" : "BEACON");
+				     is_beacon(le16_to_cpu
+					       (beacon->header.
+						frame_ctl)) ?
+				     "BEACON" : "PROBE RESPONSE");
 #endif
 		memcpy(target, &network, sizeof(*target));
 		list_add_tail(&target->list, &ieee->network_list);
@@ -1134,34 +1400,60 @@ static inline void ieee80211_process_probe_response(struct ieee80211_device
 				     escape_essid(target->ssid,
 						  target->ssid_len),
 				     MAC_ARG(target->bssid),
-				     WLAN_FC_GET_STYPE(beacon->header.
-						       frame_ctl) ==
-				     IEEE80211_STYPE_PROBE_RESP ?
-				     "PROBE RESPONSE" : "BEACON");
+				     is_beacon(le16_to_cpu
+					       (beacon->header.
+						frame_ctl)) ?
+				     "BEACON" : "PROBE RESPONSE");
 		update_network(target, &network);
 	}
 
 	spin_unlock_irqrestore(&ieee->lock, flags);
+
+	if (is_beacon(le16_to_cpu(beacon->header.frame_ctl))) {
+		if (ieee->handle_beacon != NULL)
+			ieee->handle_beacon(dev, beacon, &network);
+	} else {
+		if (ieee->handle_probe_response != NULL)
+			ieee->handle_probe_response(dev, beacon, &network);
+	}
 }
 
 void ieee80211_rx_mgt(struct ieee80211_device *ieee,
-		      struct ieee80211_hdr *header,
+		      struct ieee80211_hdr_4addr *header,
 		      struct ieee80211_rx_stats *stats)
 {
-	switch (WLAN_FC_GET_STYPE(header->frame_ctl)) {
+	switch (WLAN_FC_GET_STYPE(le16_to_cpu(header->frame_ctl))) {
 	case IEEE80211_STYPE_ASSOC_RESP:
 		IEEE80211_DEBUG_MGMT("received ASSOCIATION RESPONSE (%d)\n",
-				     WLAN_FC_GET_STYPE(header->frame_ctl));
+				     WLAN_FC_GET_STYPE(le16_to_cpu
+						       (header->frame_ctl)));
+		ieee80211_handle_assoc_resp(ieee,
+					    (struct ieee80211_assoc_response *)
+					    header, stats);
 		break;
 
 	case IEEE80211_STYPE_REASSOC_RESP:
 		IEEE80211_DEBUG_MGMT("received REASSOCIATION RESPONSE (%d)\n",
-				     WLAN_FC_GET_STYPE(header->frame_ctl));
+				     WLAN_FC_GET_STYPE(le16_to_cpu
+						       (header->frame_ctl)));
+		break;
+
+	case IEEE80211_STYPE_PROBE_REQ:
+		IEEE80211_DEBUG_MGMT("recieved auth (%d)\n",
+				     WLAN_FC_GET_STYPE(le16_to_cpu
+						       (header->frame_ctl)));
+
+		if (ieee->handle_probe_request != NULL)
+			ieee->handle_probe_request(ieee->dev,
+						   (struct
+						    ieee80211_probe_request *)
+						   header, stats);
 		break;
 
 	case IEEE80211_STYPE_PROBE_RESP:
 		IEEE80211_DEBUG_MGMT("received PROBE RESPONSE (%d)\n",
-				     WLAN_FC_GET_STYPE(header->frame_ctl));
+				     WLAN_FC_GET_STYPE(le16_to_cpu
+						       (header->frame_ctl)));
 		IEEE80211_DEBUG_SCAN("Probe response\n");
 		ieee80211_process_probe_response(ieee,
 						 (struct
@@ -1171,20 +1463,46 @@ void ieee80211_rx_mgt(struct ieee80211_device *ieee,
 
 	case IEEE80211_STYPE_BEACON:
 		IEEE80211_DEBUG_MGMT("received BEACON (%d)\n",
-				     WLAN_FC_GET_STYPE(header->frame_ctl));
+				     WLAN_FC_GET_STYPE(le16_to_cpu
+						       (header->frame_ctl)));
 		IEEE80211_DEBUG_SCAN("Beacon\n");
 		ieee80211_process_probe_response(ieee,
 						 (struct
 						  ieee80211_probe_response *)
 						 header, stats);
 		break;
+	case IEEE80211_STYPE_AUTH:
 
+		IEEE80211_DEBUG_MGMT("recieved auth (%d)\n",
+				     WLAN_FC_GET_STYPE(le16_to_cpu
+						       (header->frame_ctl)));
+
+		if (ieee->handle_auth != NULL)
+			ieee->handle_auth(ieee->dev,
+					  (struct ieee80211_auth *)header);
+		break;
+
+	case IEEE80211_STYPE_DISASSOC:
+		if (ieee->handle_disassoc != NULL)
+			ieee->handle_disassoc(ieee->dev,
+					      (struct ieee80211_disassoc *)
+					      header);
+		break;
+
+	case IEEE80211_STYPE_DEAUTH:
+		printk("DEAUTH from AP\n");
+		if (ieee->handle_deauth != NULL)
+			ieee->handle_deauth(ieee->dev, (struct ieee80211_auth *)
+					    header);
+		break;
 	default:
 		IEEE80211_DEBUG_MGMT("received UNKNOWN (%d)\n",
-				     WLAN_FC_GET_STYPE(header->frame_ctl));
+				     WLAN_FC_GET_STYPE(le16_to_cpu
+						       (header->frame_ctl)));
 		IEEE80211_WARNING("%s: Unknown management packet: %d\n",
 				  ieee->dev->name,
-				  WLAN_FC_GET_STYPE(header->frame_ctl));
+				  WLAN_FC_GET_STYPE(le16_to_cpu
+						    (header->frame_ctl)));
 		break;
 	}
 }
diff --git a/net/ieee80211/ieee80211_tx.c b/net/ieee80211/ieee80211_tx.c
index c9aaff3fea1..95ccbadbf55 100644
--- a/net/ieee80211/ieee80211_tx.c
+++ b/net/ieee80211/ieee80211_tx.c
@@ -1,6 +1,6 @@
 /******************************************************************************
 
-  Copyright(c) 2003 - 2004 Intel Corporation. All rights reserved.
+  Copyright(c) 2003 - 2005 Intel Corporation. All rights reserved.
 
   This program is free software; you can redistribute it and/or modify it
   under the terms of version 2 of the GNU General Public License as
@@ -128,7 +128,7 @@ payload of each frame is reduced to 492 bytes.
 static u8 P802_1H_OUI[P80211_OUI_LEN] = { 0x00, 0x00, 0xf8 };
 static u8 RFC1042_OUI[P80211_OUI_LEN] = { 0x00, 0x00, 0x00 };
 
-static inline int ieee80211_put_snap(u8 * data, u16 h_proto)
+static inline int ieee80211_copy_snap(u8 * data, u16 h_proto)
 {
 	struct ieee80211_snap_hdr *snap;
 	u8 *oui;
@@ -157,31 +157,14 @@ static inline int ieee80211_encrypt_fragment(struct ieee80211_device *ieee,
 	struct ieee80211_crypt_data *crypt = ieee->crypt[ieee->tx_keyidx];
 	int res;
 
-#ifdef CONFIG_IEEE80211_CRYPT_TKIP
-	struct ieee80211_hdr *header;
-
-	if (ieee->tkip_countermeasures &&
-	    crypt && crypt->ops && strcmp(crypt->ops->name, "TKIP") == 0) {
-		header = (struct ieee80211_hdr *)frag->data;
-		if (net_ratelimit()) {
-			printk(KERN_DEBUG "%s: TKIP countermeasures: dropped "
-			       "TX packet to " MAC_FMT "\n",
-			       ieee->dev->name, MAC_ARG(header->addr1));
-		}
+	if (crypt == NULL)
 		return -1;
-	}
-#endif
+
 	/* To encrypt, frame format is:
 	 * IV (4 bytes), clear payload (including SNAP), ICV (4 bytes) */
-
-	// PR: FIXME: Copied from hostap. Check fragmentation/MSDU/MPDU encryption.
-	/* Host-based IEEE 802.11 fragmentation for TX is not yet supported, so
-	 * call both MSDU and MPDU encryption functions from here. */
 	atomic_inc(&crypt->refcnt);
 	res = 0;
-	if (crypt->ops->encrypt_msdu)
-		res = crypt->ops->encrypt_msdu(frag, hdr_len, crypt->priv);
-	if (res == 0 && crypt->ops->encrypt_mpdu)
+	if (crypt->ops && crypt->ops->encrypt_mpdu)
 		res = crypt->ops->encrypt_mpdu(frag, hdr_len, crypt->priv);
 
 	atomic_dec(&crypt->refcnt);
@@ -207,7 +190,7 @@ void ieee80211_txb_free(struct ieee80211_txb *txb)
 }
 
 static struct ieee80211_txb *ieee80211_alloc_txb(int nr_frags, int txb_size,
-						 int gfp_mask)
+						 int headroom, gfp_t gfp_mask)
 {
 	struct ieee80211_txb *txb;
 	int i;
@@ -221,11 +204,13 @@ static struct ieee80211_txb *ieee80211_alloc_txb(int nr_frags, int txb_size,
 	txb->frag_size = txb_size;
 
 	for (i = 0; i < nr_frags; i++) {
-		txb->fragments[i] = dev_alloc_skb(txb_size);
+		txb->fragments[i] = __dev_alloc_skb(txb_size + headroom,
+						    gfp_mask);
 		if (unlikely(!txb->fragments[i])) {
 			i--;
 			break;
 		}
+		skb_reserve(txb->fragments[i], headroom);
 	}
 	if (unlikely(i != nr_frags)) {
 		while (i >= 0)
@@ -236,25 +221,31 @@ static struct ieee80211_txb *ieee80211_alloc_txb(int nr_frags, int txb_size,
 	return txb;
 }
 
-/* SKBs are added to the ieee->tx_queue. */
+/* Incoming skb is converted to a txb which consists of
+ * a block of 802.11 fragment packets (stored as skbs) */
 int ieee80211_xmit(struct sk_buff *skb, struct net_device *dev)
 {
 	struct ieee80211_device *ieee = netdev_priv(dev);
 	struct ieee80211_txb *txb = NULL;
-	struct ieee80211_hdr *frag_hdr;
-	int i, bytes_per_frag, nr_frags, bytes_last_frag, frag_size;
+	struct ieee80211_hdr_3addr *frag_hdr;
+	int i, bytes_per_frag, nr_frags, bytes_last_frag, frag_size,
+	    rts_required;
 	unsigned long flags;
 	struct net_device_stats *stats = &ieee->stats;
-	int ether_type, encrypt;
+	int ether_type, encrypt, host_encrypt, host_encrypt_msdu, host_build_iv;
 	int bytes, fc, hdr_len;
 	struct sk_buff *skb_frag;
-	struct ieee80211_hdr header = {	/* Ensure zero initialized */
+	struct ieee80211_hdr_3addr header = {	/* Ensure zero initialized */
 		.duration_id = 0,
 		.seq_ctl = 0
 	};
 	u8 dest[ETH_ALEN], src[ETH_ALEN];
-
 	struct ieee80211_crypt_data *crypt;
+	int priority = skb->priority;
+	int snapped = 0;
+
+	if (ieee->is_queue_full && (*ieee->is_queue_full) (dev, priority))
+		return NETDEV_TX_BUSY;
 
 	spin_lock_irqsave(&ieee->lock, flags);
 
@@ -276,7 +267,11 @@ int ieee80211_xmit(struct sk_buff *skb, struct net_device *dev)
 	crypt = ieee->crypt[ieee->tx_keyidx];
 
 	encrypt = !(ether_type == ETH_P_PAE && ieee->ieee802_1x) &&
-	    ieee->host_encrypt && crypt && crypt->ops;
+	    ieee->sec.encrypt;
+
+	host_encrypt = ieee->host_encrypt && encrypt && crypt;
+	host_encrypt_msdu = ieee->host_encrypt_msdu && encrypt && crypt;
+	host_build_iv = ieee->host_build_iv && encrypt && crypt;
 
 	if (!encrypt && ieee->ieee802_1x &&
 	    ieee->drop_unencrypted && ether_type != ETH_P_PAE) {
@@ -285,8 +280,8 @@ int ieee80211_xmit(struct sk_buff *skb, struct net_device *dev)
 	}
 
 	/* Save source and destination addresses */
-	memcpy(&dest, skb->data, ETH_ALEN);
-	memcpy(&src, skb->data + ETH_ALEN, ETH_ALEN);
+	memcpy(dest, skb->data, ETH_ALEN);
+	memcpy(src, skb->data + ETH_ALEN, ETH_ALEN);
 
 	/* Advance the SKB to the start of the payload */
 	skb_pull(skb, sizeof(struct ethhdr));
@@ -294,7 +289,7 @@ int ieee80211_xmit(struct sk_buff *skb, struct net_device *dev)
 	/* Determine total amount of storage required for TXB packets */
 	bytes = skb->len + SNAP_SIZE + sizeof(u16);
 
-	if (encrypt)
+	if (host_encrypt)
 		fc = IEEE80211_FTYPE_DATA | IEEE80211_STYPE_DATA |
 		    IEEE80211_FCTL_PROTECTED;
 	else
@@ -302,70 +297,144 @@ int ieee80211_xmit(struct sk_buff *skb, struct net_device *dev)
 
 	if (ieee->iw_mode == IW_MODE_INFRA) {
 		fc |= IEEE80211_FCTL_TODS;
-		/* To DS: Addr1 = BSSID, Addr2 = SA,
-		   Addr3 = DA */
-		memcpy(&header.addr1, ieee->bssid, ETH_ALEN);
-		memcpy(&header.addr2, &src, ETH_ALEN);
-		memcpy(&header.addr3, &dest, ETH_ALEN);
+		/* To DS: Addr1 = BSSID, Addr2 = SA, Addr3 = DA */
+		memcpy(header.addr1, ieee->bssid, ETH_ALEN);
+		memcpy(header.addr2, src, ETH_ALEN);
+		memcpy(header.addr3, dest, ETH_ALEN);
 	} else if (ieee->iw_mode == IW_MODE_ADHOC) {
-		/* not From/To DS: Addr1 = DA, Addr2 = SA,
-		   Addr3 = BSSID */
-		memcpy(&header.addr1, dest, ETH_ALEN);
-		memcpy(&header.addr2, src, ETH_ALEN);
-		memcpy(&header.addr3, ieee->bssid, ETH_ALEN);
+		/* not From/To DS: Addr1 = DA, Addr2 = SA, Addr3 = BSSID */
+		memcpy(header.addr1, dest, ETH_ALEN);
+		memcpy(header.addr2, src, ETH_ALEN);
+		memcpy(header.addr3, ieee->bssid, ETH_ALEN);
 	}
 	header.frame_ctl = cpu_to_le16(fc);
 	hdr_len = IEEE80211_3ADDR_LEN;
 
-	/* Determine fragmentation size based on destination (multicast
-	 * and broadcast are not fragmented) */
-	if (is_multicast_ether_addr(dest) || is_broadcast_ether_addr(dest))
-		frag_size = MAX_FRAG_THRESHOLD;
-	else
-		frag_size = ieee->fts;
+	/* Encrypt msdu first on the whole data packet. */
+	if ((host_encrypt || host_encrypt_msdu) &&
+	    crypt && crypt->ops && crypt->ops->encrypt_msdu) {
+		int res = 0;
+		int len = bytes + hdr_len + crypt->ops->extra_msdu_prefix_len +
+		    crypt->ops->extra_msdu_postfix_len;
+		struct sk_buff *skb_new = dev_alloc_skb(len);
+
+		if (unlikely(!skb_new))
+			goto failed;
+
+		skb_reserve(skb_new, crypt->ops->extra_msdu_prefix_len);
+		memcpy(skb_put(skb_new, hdr_len), &header, hdr_len);
+		snapped = 1;
+		ieee80211_copy_snap(skb_put(skb_new, SNAP_SIZE + sizeof(u16)),
+				    ether_type);
+		memcpy(skb_put(skb_new, skb->len), skb->data, skb->len);
+		res = crypt->ops->encrypt_msdu(skb_new, hdr_len, crypt->priv);
+		if (res < 0) {
+			IEEE80211_ERROR("msdu encryption failed\n");
+			dev_kfree_skb_any(skb_new);
+			goto failed;
+		}
+		dev_kfree_skb_any(skb);
+		skb = skb_new;
+		bytes += crypt->ops->extra_msdu_prefix_len +
+		    crypt->ops->extra_msdu_postfix_len;
+		skb_pull(skb, hdr_len);
+	}
 
-	/* Determine amount of payload per fragment.  Regardless of if
-	 * this stack is providing the full 802.11 header, one will
-	 * eventually be affixed to this fragment -- so we must account for
-	 * it when determining the amount of payload space. */
-	bytes_per_frag = frag_size - IEEE80211_3ADDR_LEN;
-	if (ieee->config &
-	    (CFG_IEEE80211_COMPUTE_FCS | CFG_IEEE80211_RESERVE_FCS))
-		bytes_per_frag -= IEEE80211_FCS_LEN;
-
-	/* Each fragment may need to have room for encryptiong pre/postfix */
-	if (encrypt)
-		bytes_per_frag -= crypt->ops->extra_prefix_len +
-		    crypt->ops->extra_postfix_len;
-
-	/* Number of fragments is the total bytes_per_frag /
-	 * payload_per_fragment */
-	nr_frags = bytes / bytes_per_frag;
-	bytes_last_frag = bytes % bytes_per_frag;
-	if (bytes_last_frag)
+	if (host_encrypt || ieee->host_open_frag) {
+		/* Determine fragmentation size based on destination (multicast
+		 * and broadcast are not fragmented) */
+		if (is_multicast_ether_addr(dest) ||
+		    is_broadcast_ether_addr(dest))
+			frag_size = MAX_FRAG_THRESHOLD;
+		else
+			frag_size = ieee->fts;
+
+		/* Determine amount of payload per fragment.  Regardless of if
+		 * this stack is providing the full 802.11 header, one will
+		 * eventually be affixed to this fragment -- so we must account
+		 * for it when determining the amount of payload space. */
+		bytes_per_frag = frag_size - IEEE80211_3ADDR_LEN;
+		if (ieee->config &
+		    (CFG_IEEE80211_COMPUTE_FCS | CFG_IEEE80211_RESERVE_FCS))
+			bytes_per_frag -= IEEE80211_FCS_LEN;
+
+		/* Each fragment may need to have room for encryptiong
+		 * pre/postfix */
+		if (host_encrypt)
+			bytes_per_frag -= crypt->ops->extra_mpdu_prefix_len +
+			    crypt->ops->extra_mpdu_postfix_len;
+
+		/* Number of fragments is the total
+		 * bytes_per_frag / payload_per_fragment */
+		nr_frags = bytes / bytes_per_frag;
+		bytes_last_frag = bytes % bytes_per_frag;
+		if (bytes_last_frag)
+			nr_frags++;
+		else
+			bytes_last_frag = bytes_per_frag;
+	} else {
+		nr_frags = 1;
+		bytes_per_frag = bytes_last_frag = bytes;
+		frag_size = bytes + IEEE80211_3ADDR_LEN;
+	}
+
+	rts_required = (frag_size > ieee->rts
+			&& ieee->config & CFG_IEEE80211_RTS);
+	if (rts_required)
 		nr_frags++;
-	else
-		bytes_last_frag = bytes_per_frag;
 
 	/* When we allocate the TXB we allocate enough space for the reserve
 	 * and full fragment bytes (bytes_per_frag doesn't include prefix,
 	 * postfix, header, FCS, etc.) */
-	txb = ieee80211_alloc_txb(nr_frags, frag_size, GFP_ATOMIC);
+	txb = ieee80211_alloc_txb(nr_frags, frag_size,
+				  ieee->tx_headroom, GFP_ATOMIC);
 	if (unlikely(!txb)) {
 		printk(KERN_WARNING "%s: Could not allocate TXB\n",
 		       ieee->dev->name);
 		goto failed;
 	}
 	txb->encrypted = encrypt;
-	txb->payload_size = bytes;
+	if (host_encrypt)
+		txb->payload_size = frag_size * (nr_frags - 1) +
+		    bytes_last_frag;
+	else
+		txb->payload_size = bytes;
+
+	if (rts_required) {
+		skb_frag = txb->fragments[0];
+		frag_hdr =
+		    (struct ieee80211_hdr_3addr *)skb_put(skb_frag, hdr_len);
+
+		/*
+		 * Set header frame_ctl to the RTS.
+		 */
+		header.frame_ctl =
+		    cpu_to_le16(IEEE80211_FTYPE_CTL | IEEE80211_STYPE_RTS);
+		memcpy(frag_hdr, &header, hdr_len);
 
-	for (i = 0; i < nr_frags; i++) {
+		/*
+		 * Restore header frame_ctl to the original data setting.
+		 */
+		header.frame_ctl = cpu_to_le16(fc);
+
+		if (ieee->config &
+		    (CFG_IEEE80211_COMPUTE_FCS | CFG_IEEE80211_RESERVE_FCS))
+			skb_put(skb_frag, 4);
+
+		txb->rts_included = 1;
+		i = 1;
+	} else
+		i = 0;
+
+	for (; i < nr_frags; i++) {
 		skb_frag = txb->fragments[i];
 
-		if (encrypt)
-			skb_reserve(skb_frag, crypt->ops->extra_prefix_len);
+		if (host_encrypt || host_build_iv)
+			skb_reserve(skb_frag,
+				    crypt->ops->extra_mpdu_prefix_len);
 
-		frag_hdr = (struct ieee80211_hdr *)skb_put(skb_frag, hdr_len);
+		frag_hdr =
+		    (struct ieee80211_hdr_3addr *)skb_put(skb_frag, hdr_len);
 		memcpy(frag_hdr, &header, hdr_len);
 
 		/* If this is not the last fragment, then add the MOREFRAGS
@@ -379,11 +448,10 @@ int ieee80211_xmit(struct sk_buff *skb, struct net_device *dev)
 			bytes = bytes_last_frag;
 		}
 
-		/* Put a SNAP header on the first fragment */
-		if (i == 0) {
-			ieee80211_put_snap(skb_put
-					   (skb_frag, SNAP_SIZE + sizeof(u16)),
-					   ether_type);
+		if (i == 0 && !snapped) {
+			ieee80211_copy_snap(skb_put
+					    (skb_frag, SNAP_SIZE + sizeof(u16)),
+					    ether_type);
 			bytes -= SNAP_SIZE + sizeof(u16);
 		}
 
@@ -394,8 +462,19 @@ int ieee80211_xmit(struct sk_buff *skb, struct net_device *dev)
 
 		/* Encryption routine will move the header forward in order
 		 * to insert the IV between the header and the payload */
-		if (encrypt)
+		if (host_encrypt)
 			ieee80211_encrypt_fragment(ieee, skb_frag, hdr_len);
+		else if (host_build_iv) {
+			struct ieee80211_crypt_data *crypt;
+
+			crypt = ieee->crypt[ieee->tx_keyidx];
+			atomic_inc(&crypt->refcnt);
+			if (crypt->ops->build_iv)
+				crypt->ops->build_iv(skb_frag, hdr_len,
+						     crypt->priv);
+			atomic_dec(&crypt->refcnt);
+		}
+
 		if (ieee->config &
 		    (CFG_IEEE80211_COMPUTE_FCS | CFG_IEEE80211_RESERVE_FCS))
 			skb_put(skb_frag, 4);
@@ -407,11 +486,20 @@ int ieee80211_xmit(struct sk_buff *skb, struct net_device *dev)
 	dev_kfree_skb_any(skb);
 
 	if (txb) {
-		if ((*ieee->hard_start_xmit) (txb, dev) == 0) {
+		int ret = (*ieee->hard_start_xmit) (txb, dev, priority);
+		if (ret == 0) {
 			stats->tx_packets++;
 			stats->tx_bytes += txb->payload_size;
 			return 0;
 		}
+
+		if (ret == NETDEV_TX_BUSY) {
+			printk(KERN_ERR "%s: NETDEV_TX_BUSY returned; "
+			       "driver should report queue full via "
+			       "ieee_device->is_queue_full.\n",
+			       ieee->dev->name);
+		}
+
 		ieee80211_txb_free(txb);
 	}
 
@@ -422,7 +510,72 @@ int ieee80211_xmit(struct sk_buff *skb, struct net_device *dev)
 	netif_stop_queue(dev);
 	stats->tx_errors++;
 	return 1;
+}
+
+/* Incoming 802.11 strucure is converted to a TXB
+ * a block of 802.11 fragment packets (stored as skbs) */
+int ieee80211_tx_frame(struct ieee80211_device *ieee,
+		       struct ieee80211_hdr *frame, int len)
+{
+	struct ieee80211_txb *txb = NULL;
+	unsigned long flags;
+	struct net_device_stats *stats = &ieee->stats;
+	struct sk_buff *skb_frag;
+	int priority = -1;
+
+	spin_lock_irqsave(&ieee->lock, flags);
 
+	/* If there is no driver handler to take the TXB, dont' bother
+	 * creating it... */
+	if (!ieee->hard_start_xmit) {
+		printk(KERN_WARNING "%s: No xmit handler.\n", ieee->dev->name);
+		goto success;
+	}
+
+	if (unlikely(len < 24)) {
+		printk(KERN_WARNING "%s: skb too small (%d).\n",
+		       ieee->dev->name, len);
+		goto success;
+	}
+
+	/* When we allocate the TXB we allocate enough space for the reserve
+	 * and full fragment bytes (bytes_per_frag doesn't include prefix,
+	 * postfix, header, FCS, etc.) */
+	txb = ieee80211_alloc_txb(1, len, ieee->tx_headroom, GFP_ATOMIC);
+	if (unlikely(!txb)) {
+		printk(KERN_WARNING "%s: Could not allocate TXB\n",
+		       ieee->dev->name);
+		goto failed;
+	}
+	txb->encrypted = 0;
+	txb->payload_size = len;
+
+	skb_frag = txb->fragments[0];
+
+	memcpy(skb_put(skb_frag, len), frame, len);
+
+	if (ieee->config &
+	    (CFG_IEEE80211_COMPUTE_FCS | CFG_IEEE80211_RESERVE_FCS))
+		skb_put(skb_frag, 4);
+
+      success:
+	spin_unlock_irqrestore(&ieee->lock, flags);
+
+	if (txb) {
+		if ((*ieee->hard_start_xmit) (txb, ieee->dev, priority) == 0) {
+			stats->tx_packets++;
+			stats->tx_bytes += txb->payload_size;
+			return 0;
+		}
+		ieee80211_txb_free(txb);
+	}
+	return 0;
+
+      failed:
+	spin_unlock_irqrestore(&ieee->lock, flags);
+	stats->tx_errors++;
+	return 1;
 }
 
+EXPORT_SYMBOL(ieee80211_tx_frame);
 EXPORT_SYMBOL(ieee80211_txb_free);
diff --git a/net/ieee80211/ieee80211_wx.c b/net/ieee80211/ieee80211_wx.c
index 94882f39b07..1ce7af9bec3 100644
--- a/net/ieee80211/ieee80211_wx.c
+++ b/net/ieee80211/ieee80211_wx.c
@@ -1,6 +1,6 @@
 /******************************************************************************
 
-  Copyright(c) 2004 Intel Corporation. All rights reserved.
+  Copyright(c) 2004-2005 Intel Corporation. All rights reserved.
 
   Portions of this file are based on the WEP enablement code provided by the
   Host AP project hostap-drivers v0.1.3
@@ -32,6 +32,7 @@
 
 #include <linux/kmod.h>
 #include <linux/module.h>
+#include <linux/jiffies.h>
 
 #include <net/ieee80211.h>
 #include <linux/wireless.h>
@@ -140,18 +141,41 @@ static inline char *ipw2100_translate_scan(struct ieee80211_device *ieee,
 		start = iwe_stream_add_point(start, stop, &iwe, custom);
 
 	/* Add quality statistics */
-	/* TODO: Fix these values... */
 	iwe.cmd = IWEVQUAL;
-	iwe.u.qual.qual = network->stats.signal;
-	iwe.u.qual.level = network->stats.rssi;
-	iwe.u.qual.noise = network->stats.noise;
-	iwe.u.qual.updated = network->stats.mask & IEEE80211_STATMASK_WEMASK;
-	if (!(network->stats.mask & IEEE80211_STATMASK_RSSI))
-		iwe.u.qual.updated |= IW_QUAL_LEVEL_INVALID;
-	if (!(network->stats.mask & IEEE80211_STATMASK_NOISE))
+	iwe.u.qual.updated = IW_QUAL_QUAL_UPDATED | IW_QUAL_LEVEL_UPDATED |
+	    IW_QUAL_NOISE_UPDATED;
+
+	if (!(network->stats.mask & IEEE80211_STATMASK_RSSI)) {
+		iwe.u.qual.updated |= IW_QUAL_QUAL_INVALID |
+		    IW_QUAL_LEVEL_INVALID;
+		iwe.u.qual.qual = 0;
+		iwe.u.qual.level = 0;
+	} else {
+		iwe.u.qual.level = network->stats.rssi;
+		if (ieee->perfect_rssi == ieee->worst_rssi)
+			iwe.u.qual.qual = 100;
+		else
+			iwe.u.qual.qual =
+			    (100 *
+			     (ieee->perfect_rssi - ieee->worst_rssi) *
+			     (ieee->perfect_rssi - ieee->worst_rssi) -
+			     (ieee->perfect_rssi - network->stats.rssi) *
+			     (15 * (ieee->perfect_rssi - ieee->worst_rssi) +
+			      62 * (ieee->perfect_rssi - network->stats.rssi))) /
+			    ((ieee->perfect_rssi - ieee->worst_rssi) *
+			     (ieee->perfect_rssi - ieee->worst_rssi));
+		if (iwe.u.qual.qual > 100)
+			iwe.u.qual.qual = 100;
+		else if (iwe.u.qual.qual < 1)
+			iwe.u.qual.qual = 0;
+	}
+
+	if (!(network->stats.mask & IEEE80211_STATMASK_NOISE)) {
 		iwe.u.qual.updated |= IW_QUAL_NOISE_INVALID;
-	if (!(network->stats.mask & IEEE80211_STATMASK_SIGNAL))
-		iwe.u.qual.updated |= IW_QUAL_QUAL_INVALID;
+		iwe.u.qual.noise = 0;
+	} else {
+		iwe.u.qual.noise = network->stats.noise;
+	}
 
 	start = iwe_stream_add_event(start, stop, &iwe, IW_EV_QUAL_LEN);
 
@@ -162,7 +186,7 @@ static inline char *ipw2100_translate_scan(struct ieee80211_device *ieee,
 	if (iwe.u.data.length)
 		start = iwe_stream_add_point(start, stop, &iwe, custom);
 
-	if (ieee->wpa_enabled && network->wpa_ie_len) {
+	if (network->wpa_ie_len) {
 		char buf[MAX_WPA_IE_LEN * 2 + 30];
 
 		u8 *p = buf;
@@ -177,7 +201,7 @@ static inline char *ipw2100_translate_scan(struct ieee80211_device *ieee,
 		start = iwe_stream_add_point(start, stop, &iwe, buf);
 	}
 
-	if (ieee->wpa_enabled && network->rsn_ie_len) {
+	if (network->rsn_ie_len) {
 		char buf[MAX_WPA_IE_LEN * 2 + 30];
 
 		u8 *p = buf;
@@ -197,8 +221,8 @@ static inline char *ipw2100_translate_scan(struct ieee80211_device *ieee,
 	iwe.cmd = IWEVCUSTOM;
 	p = custom;
 	p += snprintf(p, MAX_CUSTOM_LEN - (p - custom),
-		      " Last beacon: %lums ago",
-		      (jiffies - network->last_scanned) / (HZ / 100));
+		      " Last beacon: %dms ago",
+		      jiffies_to_msecs(jiffies - network->last_scanned));
 	iwe.u.data.length = p - custom;
 	if (iwe.u.data.length)
 		start = iwe_stream_add_point(start, stop, &iwe, custom);
@@ -228,13 +252,13 @@ int ieee80211_wx_get_scan(struct ieee80211_device *ieee,
 			ev = ipw2100_translate_scan(ieee, ev, stop, network);
 		else
 			IEEE80211_DEBUG_SCAN("Not showing network '%s ("
-					     MAC_FMT ")' due to age (%lums).\n",
+					     MAC_FMT ")' due to age (%dms).\n",
 					     escape_essid(network->ssid,
 							  network->ssid_len),
 					     MAC_ARG(network->bssid),
-					     (jiffies -
-					      network->last_scanned) / (HZ /
-									100));
+					     jiffies_to_msecs(jiffies -
+							      network->
+							      last_scanned));
 	}
 
 	spin_unlock_irqrestore(&ieee->lock, flags);
@@ -258,6 +282,7 @@ int ieee80211_wx_set_encode(struct ieee80211_device *ieee,
 	};
 	int i, key, key_provided, len;
 	struct ieee80211_crypt_data **crypt;
+	int host_crypto = ieee->host_encrypt || ieee->host_decrypt;
 
 	IEEE80211_DEBUG_WX("SET_ENCODE\n");
 
@@ -298,15 +323,17 @@ int ieee80211_wx_set_encode(struct ieee80211_device *ieee,
 
 		if (i == WEP_KEYS) {
 			sec.enabled = 0;
+			sec.encrypt = 0;
 			sec.level = SEC_LEVEL_0;
-			sec.flags |= SEC_ENABLED | SEC_LEVEL;
+			sec.flags |= SEC_ENABLED | SEC_LEVEL | SEC_ENCRYPT;
 		}
 
 		goto done;
 	}
 
 	sec.enabled = 1;
-	sec.flags |= SEC_ENABLED;
+	sec.encrypt = 1;
+	sec.flags |= SEC_ENABLED | SEC_ENCRYPT;
 
 	if (*crypt != NULL && (*crypt)->ops != NULL &&
 	    strcmp((*crypt)->ops->name, "WEP") != 0) {
@@ -315,7 +342,7 @@ int ieee80211_wx_set_encode(struct ieee80211_device *ieee,
 		ieee80211_crypt_delayed_deinit(ieee, crypt);
 	}
 
-	if (*crypt == NULL) {
+	if (*crypt == NULL && host_crypto) {
 		struct ieee80211_crypt_data *new_crypt;
 
 		/* take WEP into use */
@@ -355,49 +382,56 @@ int ieee80211_wx_set_encode(struct ieee80211_device *ieee,
 				   key, escape_essid(sec.keys[key], len),
 				   erq->length, len);
 		sec.key_sizes[key] = len;
-		(*crypt)->ops->set_key(sec.keys[key], len, NULL,
-				       (*crypt)->priv);
+		if (*crypt)
+			(*crypt)->ops->set_key(sec.keys[key], len, NULL,
+					       (*crypt)->priv);
 		sec.flags |= (1 << key);
 		/* This ensures a key will be activated if no key is
 		 * explicitely set */
 		if (key == sec.active_key)
 			sec.flags |= SEC_ACTIVE_KEY;
+
 	} else {
-		len = (*crypt)->ops->get_key(sec.keys[key], WEP_KEY_LEN,
-					     NULL, (*crypt)->priv);
-		if (len == 0) {
-			/* Set a default key of all 0 */
-			IEEE80211_DEBUG_WX("Setting key %d to all zero.\n",
-					   key);
-			memset(sec.keys[key], 0, 13);
-			(*crypt)->ops->set_key(sec.keys[key], 13, NULL,
-					       (*crypt)->priv);
-			sec.key_sizes[key] = 13;
-			sec.flags |= (1 << key);
+		if (host_crypto) {
+			len = (*crypt)->ops->get_key(sec.keys[key], WEP_KEY_LEN,
+						     NULL, (*crypt)->priv);
+			if (len == 0) {
+				/* Set a default key of all 0 */
+				IEEE80211_DEBUG_WX("Setting key %d to all "
+						   "zero.\n", key);
+				memset(sec.keys[key], 0, 13);
+				(*crypt)->ops->set_key(sec.keys[key], 13, NULL,
+						       (*crypt)->priv);
+				sec.key_sizes[key] = 13;
+				sec.flags |= (1 << key);
+			}
 		}
-
 		/* No key data - just set the default TX key index */
 		if (key_provided) {
-			IEEE80211_DEBUG_WX
-			    ("Setting key %d to default Tx key.\n", key);
+			IEEE80211_DEBUG_WX("Setting key %d to default Tx "
+					   "key.\n", key);
 			ieee->tx_keyidx = key;
 			sec.active_key = key;
 			sec.flags |= SEC_ACTIVE_KEY;
 		}
 	}
-
-      done:
-	ieee->open_wep = !(erq->flags & IW_ENCODE_RESTRICTED);
-	sec.auth_mode = ieee->open_wep ? WLAN_AUTH_OPEN : WLAN_AUTH_SHARED_KEY;
-	sec.flags |= SEC_AUTH_MODE;
-	IEEE80211_DEBUG_WX("Auth: %s\n", sec.auth_mode == WLAN_AUTH_OPEN ?
-			   "OPEN" : "SHARED KEY");
+	if (erq->flags & (IW_ENCODE_OPEN | IW_ENCODE_RESTRICTED)) {
+		ieee->open_wep = !(erq->flags & IW_ENCODE_RESTRICTED);
+		sec.auth_mode = ieee->open_wep ? WLAN_AUTH_OPEN :
+		    WLAN_AUTH_SHARED_KEY;
+		sec.flags |= SEC_AUTH_MODE;
+		IEEE80211_DEBUG_WX("Auth: %s\n",
+				   sec.auth_mode == WLAN_AUTH_OPEN ?
+				   "OPEN" : "SHARED KEY");
+	}
 
 	/* For now we just support WEP, so only set that security level...
 	 * TODO: When WPA is added this is one place that needs to change */
 	sec.flags |= SEC_LEVEL;
 	sec.level = SEC_LEVEL_1;	/* 40 and 104 bit WEP */
+	sec.encode_alg[key] = SEC_ALG_WEP;
 
+      done:
 	if (ieee->set_security)
 		ieee->set_security(dev, &sec);
 
@@ -422,6 +456,7 @@ int ieee80211_wx_get_encode(struct ieee80211_device *ieee,
 	struct iw_point *erq = &(wrqu->encoding);
 	int len, key;
 	struct ieee80211_crypt_data *crypt;
+	struct ieee80211_security *sec = &ieee->sec;
 
 	IEEE80211_DEBUG_WX("GET_ENCODE\n");
 
@@ -436,23 +471,16 @@ int ieee80211_wx_get_encode(struct ieee80211_device *ieee,
 	crypt = ieee->crypt[key];
 	erq->flags = key + 1;
 
-	if (crypt == NULL || crypt->ops == NULL) {
+	if (!sec->enabled) {
 		erq->length = 0;
 		erq->flags |= IW_ENCODE_DISABLED;
 		return 0;
 	}
 
-	if (strcmp(crypt->ops->name, "WEP") != 0) {
-		/* only WEP is supported with wireless extensions, so just
-		 * report that encryption is used */
-		erq->length = 0;
-		erq->flags |= IW_ENCODE_ENABLED;
-		return 0;
-	}
+	len = sec->key_sizes[key];
+	memcpy(keybuf, sec->keys[key], len);
 
-	len = crypt->ops->get_key(keybuf, WEP_KEY_LEN, NULL, crypt->priv);
 	erq->length = (len >= 0 ? len : 0);
-
 	erq->flags |= IW_ENCODE_ENABLED;
 
 	if (ieee->open_wep)
@@ -463,6 +491,240 @@ int ieee80211_wx_get_encode(struct ieee80211_device *ieee,
 	return 0;
 }
 
+int ieee80211_wx_set_encodeext(struct ieee80211_device *ieee,
+			       struct iw_request_info *info,
+			       union iwreq_data *wrqu, char *extra)
+{
+	struct net_device *dev = ieee->dev;
+	struct iw_point *encoding = &wrqu->encoding;
+	struct iw_encode_ext *ext = (struct iw_encode_ext *)extra;
+	int i, idx, ret = 0;
+	int group_key = 0;
+	const char *alg, *module;
+	struct ieee80211_crypto_ops *ops;
+	struct ieee80211_crypt_data **crypt;
+
+	struct ieee80211_security sec = {
+		.flags = 0,
+	};
+
+	idx = encoding->flags & IW_ENCODE_INDEX;
+	if (idx) {
+		if (idx < 1 || idx > WEP_KEYS)
+			return -EINVAL;
+		idx--;
+	} else
+		idx = ieee->tx_keyidx;
+
+	if (ext->ext_flags & IW_ENCODE_EXT_GROUP_KEY) {
+		crypt = &ieee->crypt[idx];
+		group_key = 1;
+	} else {
+		if (idx != 0)
+			return -EINVAL;
+		if (ieee->iw_mode == IW_MODE_INFRA)
+			crypt = &ieee->crypt[idx];
+		else
+			return -EINVAL;
+	}
+
+	sec.flags |= SEC_ENABLED | SEC_ENCRYPT;
+	if ((encoding->flags & IW_ENCODE_DISABLED) ||
+	    ext->alg == IW_ENCODE_ALG_NONE) {
+		if (*crypt)
+			ieee80211_crypt_delayed_deinit(ieee, crypt);
+
+		for (i = 0; i < WEP_KEYS; i++)
+			if (ieee->crypt[i] != NULL)
+				break;
+
+		if (i == WEP_KEYS) {
+			sec.enabled = 0;
+			sec.encrypt = 0;
+			sec.level = SEC_LEVEL_0;
+			sec.flags |= SEC_LEVEL;
+		}
+		goto done;
+	}
+
+	sec.enabled = 1;
+	sec.encrypt = 1;
+
+	if (group_key ? !ieee->host_mc_decrypt :
+	    !(ieee->host_encrypt || ieee->host_decrypt ||
+	      ieee->host_encrypt_msdu))
+		goto skip_host_crypt;
+
+	switch (ext->alg) {
+	case IW_ENCODE_ALG_WEP:
+		alg = "WEP";
+		module = "ieee80211_crypt_wep";
+		break;
+	case IW_ENCODE_ALG_TKIP:
+		alg = "TKIP";
+		module = "ieee80211_crypt_tkip";
+		break;
+	case IW_ENCODE_ALG_CCMP:
+		alg = "CCMP";
+		module = "ieee80211_crypt_ccmp";
+		break;
+	default:
+		IEEE80211_DEBUG_WX("%s: unknown crypto alg %d\n",
+				   dev->name, ext->alg);
+		ret = -EINVAL;
+		goto done;
+	}
+
+	ops = ieee80211_get_crypto_ops(alg);
+	if (ops == NULL) {
+		request_module(module);
+		ops = ieee80211_get_crypto_ops(alg);
+	}
+	if (ops == NULL) {
+		IEEE80211_DEBUG_WX("%s: unknown crypto alg %d\n",
+				   dev->name, ext->alg);
+		ret = -EINVAL;
+		goto done;
+	}
+
+	if (*crypt == NULL || (*crypt)->ops != ops) {
+		struct ieee80211_crypt_data *new_crypt;
+
+		ieee80211_crypt_delayed_deinit(ieee, crypt);
+
+		new_crypt = (struct ieee80211_crypt_data *)
+		    kmalloc(sizeof(*new_crypt), GFP_KERNEL);
+		if (new_crypt == NULL) {
+			ret = -ENOMEM;
+			goto done;
+		}
+		memset(new_crypt, 0, sizeof(struct ieee80211_crypt_data));
+		new_crypt->ops = ops;
+		if (new_crypt->ops && try_module_get(new_crypt->ops->owner))
+			new_crypt->priv = new_crypt->ops->init(idx);
+		if (new_crypt->priv == NULL) {
+			kfree(new_crypt);
+			ret = -EINVAL;
+			goto done;
+		}
+		*crypt = new_crypt;
+	}
+
+	if (ext->key_len > 0 && (*crypt)->ops->set_key &&
+	    (*crypt)->ops->set_key(ext->key, ext->key_len, ext->rx_seq,
+				   (*crypt)->priv) < 0) {
+		IEEE80211_DEBUG_WX("%s: key setting failed\n", dev->name);
+		ret = -EINVAL;
+		goto done;
+	}
+
+      skip_host_crypt:
+	if (ext->ext_flags & IW_ENCODE_EXT_SET_TX_KEY) {
+		ieee->tx_keyidx = idx;
+		sec.active_key = idx;
+		sec.flags |= SEC_ACTIVE_KEY;
+	}
+
+	if (ext->alg != IW_ENCODE_ALG_NONE) {
+		memcpy(sec.keys[idx], ext->key, ext->key_len);
+		sec.key_sizes[idx] = ext->key_len;
+		sec.flags |= (1 << idx);
+		if (ext->alg == IW_ENCODE_ALG_WEP) {
+			sec.encode_alg[idx] = SEC_ALG_WEP;
+			sec.flags |= SEC_LEVEL;
+			sec.level = SEC_LEVEL_1;
+		} else if (ext->alg == IW_ENCODE_ALG_TKIP) {
+			sec.encode_alg[idx] = SEC_ALG_TKIP;
+			sec.flags |= SEC_LEVEL;
+			sec.level = SEC_LEVEL_2;
+		} else if (ext->alg == IW_ENCODE_ALG_CCMP) {
+			sec.encode_alg[idx] = SEC_ALG_CCMP;
+			sec.flags |= SEC_LEVEL;
+			sec.level = SEC_LEVEL_3;
+		}
+		/* Don't set sec level for group keys. */
+		if (group_key)
+			sec.flags &= ~SEC_LEVEL;
+	}
+      done:
+	if (ieee->set_security)
+		ieee->set_security(ieee->dev, &sec);
+
+	/*
+	 * Do not reset port if card is in Managed mode since resetting will
+	 * generate new IEEE 802.11 authentication which may end up in looping
+	 * with IEEE 802.1X. If your hardware requires a reset after WEP
+	 * configuration (for example... Prism2), implement the reset_port in
+	 * the callbacks structures used to initialize the 802.11 stack.
+	 */
+	if (ieee->reset_on_keychange &&
+	    ieee->iw_mode != IW_MODE_INFRA &&
+	    ieee->reset_port && ieee->reset_port(dev)) {
+		IEEE80211_DEBUG_WX("%s: reset_port failed\n", dev->name);
+		return -EINVAL;
+	}
+
+	return ret;
+}
+
+int ieee80211_wx_get_encodeext(struct ieee80211_device *ieee,
+			       struct iw_request_info *info,
+			       union iwreq_data *wrqu, char *extra)
+{
+	struct iw_point *encoding = &wrqu->encoding;
+	struct iw_encode_ext *ext = (struct iw_encode_ext *)extra;
+	struct ieee80211_security *sec = &ieee->sec;
+	int idx, max_key_len;
+
+	max_key_len = encoding->length - sizeof(*ext);
+	if (max_key_len < 0)
+		return -EINVAL;
+
+	idx = encoding->flags & IW_ENCODE_INDEX;
+	if (idx) {
+		if (idx < 1 || idx > WEP_KEYS)
+			return -EINVAL;
+		idx--;
+	} else
+		idx = ieee->tx_keyidx;
+
+	if (!ext->ext_flags & IW_ENCODE_EXT_GROUP_KEY)
+		if (idx != 0 || ieee->iw_mode != IW_MODE_INFRA)
+			return -EINVAL;
+
+	encoding->flags = idx + 1;
+	memset(ext, 0, sizeof(*ext));
+
+	if (!sec->enabled) {
+		ext->alg = IW_ENCODE_ALG_NONE;
+		ext->key_len = 0;
+		encoding->flags |= IW_ENCODE_DISABLED;
+	} else {
+		if (sec->encode_alg[idx] == SEC_ALG_WEP)
+			ext->alg = IW_ENCODE_ALG_WEP;
+		else if (sec->encode_alg[idx] == SEC_ALG_TKIP)
+			ext->alg = IW_ENCODE_ALG_TKIP;
+		else if (sec->encode_alg[idx] == SEC_ALG_CCMP)
+			ext->alg = IW_ENCODE_ALG_CCMP;
+		else
+			return -EINVAL;
+
+		ext->key_len = sec->key_sizes[idx];
+		memcpy(ext->key, sec->keys[idx], ext->key_len);
+		encoding->flags |= IW_ENCODE_ENABLED;
+		if (ext->key_len &&
+		    (ext->alg == IW_ENCODE_ALG_TKIP ||
+		     ext->alg == IW_ENCODE_ALG_CCMP))
+			ext->ext_flags |= IW_ENCODE_EXT_TX_SEQ_VALID;
+
+	}
+
+	return 0;
+}
+
+EXPORT_SYMBOL(ieee80211_wx_set_encodeext);
+EXPORT_SYMBOL(ieee80211_wx_get_encodeext);
+
 EXPORT_SYMBOL(ieee80211_wx_get_scan);
 EXPORT_SYMBOL(ieee80211_wx_set_encode);
 EXPORT_SYMBOL(ieee80211_wx_get_encode);
diff --git a/net/ipv4/arp.c b/net/ipv4/arp.c
index 8bf312bdea1..b425748f02d 100644
--- a/net/ipv4/arp.c
+++ b/net/ipv4/arp.c
@@ -241,7 +241,7 @@ static int arp_constructor(struct neighbour *neigh)
 	neigh->type = inet_addr_type(addr);
 
 	rcu_read_lock();
-	in_dev = rcu_dereference(__in_dev_get(dev));
+	in_dev = __in_dev_get_rcu(dev);
 	if (in_dev == NULL) {
 		rcu_read_unlock();
 		return -EINVAL;
@@ -697,12 +697,6 @@ void arp_send(int type, int ptype, u32 dest_ip,
 	arp_xmit(skb);
 }
 
-static void parp_redo(struct sk_buff *skb)
-{
-	nf_reset(skb);
-	arp_rcv(skb, skb->dev, NULL, skb->dev);
-}
-
 /*
  *	Process an arp request.
  */
@@ -922,6 +916,11 @@ out:
 	return 0;
 }
 
+static void parp_redo(struct sk_buff *skb)
+{
+	arp_process(skb);
+}
+
 
 /*
  *	Receive an arp request from the device layer.
@@ -990,8 +989,8 @@ static int arp_req_set(struct arpreq *r, struct net_device * dev)
 			ipv4_devconf.proxy_arp = 1;
 			return 0;
 		}
-		if (__in_dev_get(dev)) {
-			__in_dev_get(dev)->cnf.proxy_arp = 1;
+		if (__in_dev_get_rtnl(dev)) {
+			__in_dev_get_rtnl(dev)->cnf.proxy_arp = 1;
 			return 0;
 		}
 		return -ENXIO;
@@ -1096,8 +1095,8 @@ static int arp_req_delete(struct arpreq *r, struct net_device * dev)
 				ipv4_devconf.proxy_arp = 0;
 				return 0;
 			}
-			if (__in_dev_get(dev)) {
-				__in_dev_get(dev)->cnf.proxy_arp = 0;
+			if (__in_dev_get_rtnl(dev)) {
+				__in_dev_get_rtnl(dev)->cnf.proxy_arp = 0;
 				return 0;
 			}
 			return -ENXIO;
diff --git a/net/ipv4/devinet.c b/net/ipv4/devinet.c
index ba2895ae815..4ec4b2ca6ab 100644
--- a/net/ipv4/devinet.c
+++ b/net/ipv4/devinet.c
@@ -351,7 +351,7 @@ static int inet_insert_ifa(struct in_ifaddr *ifa)
 
 static int inet_set_ifa(struct net_device *dev, struct in_ifaddr *ifa)
 {
-	struct in_device *in_dev = __in_dev_get(dev);
+	struct in_device *in_dev = __in_dev_get_rtnl(dev);
 
 	ASSERT_RTNL();
 
@@ -449,7 +449,7 @@ static int inet_rtm_newaddr(struct sk_buff *skb, struct nlmsghdr *nlh, void *arg
 		goto out;
 
 	rc = -ENOBUFS;
-	if ((in_dev = __in_dev_get(dev)) == NULL) {
+	if ((in_dev = __in_dev_get_rtnl(dev)) == NULL) {
 		in_dev = inetdev_init(dev);
 		if (!in_dev)
 			goto out;
@@ -584,7 +584,7 @@ int devinet_ioctl(unsigned int cmd, void __user *arg)
 	if (colon)
 		*colon = ':';
 
-	if ((in_dev = __in_dev_get(dev)) != NULL) {
+	if ((in_dev = __in_dev_get_rtnl(dev)) != NULL) {
 		if (tryaddrmatch) {
 			/* Matthias Andree */
 			/* compare label and address (4.4BSD style) */
@@ -715,6 +715,7 @@ int devinet_ioctl(unsigned int cmd, void __user *arg)
 			break;
 		ret = 0;
 		if (ifa->ifa_mask != sin->sin_addr.s_addr) {
+			u32 old_mask = ifa->ifa_mask;
 			inet_del_ifa(in_dev, ifap, 0);
 			ifa->ifa_mask = sin->sin_addr.s_addr;
 			ifa->ifa_prefixlen = inet_mask_len(ifa->ifa_mask);
@@ -728,7 +729,7 @@ int devinet_ioctl(unsigned int cmd, void __user *arg)
 			if ((dev->flags & IFF_BROADCAST) &&
 			    (ifa->ifa_prefixlen < 31) &&
 			    (ifa->ifa_broadcast ==
-			     (ifa->ifa_local|~ifa->ifa_mask))) {
+			     (ifa->ifa_local|~old_mask))) {
 				ifa->ifa_broadcast = (ifa->ifa_local |
 						      ~sin->sin_addr.s_addr);
 			}
@@ -748,7 +749,7 @@ rarok:
 
 static int inet_gifconf(struct net_device *dev, char __user *buf, int len)
 {
-	struct in_device *in_dev = __in_dev_get(dev);
+	struct in_device *in_dev = __in_dev_get_rtnl(dev);
 	struct in_ifaddr *ifa;
 	struct ifreq ifr;
 	int done = 0;
@@ -791,7 +792,7 @@ u32 inet_select_addr(const struct net_device *dev, u32 dst, int scope)
 	struct in_device *in_dev;
 
 	rcu_read_lock();
-	in_dev = __in_dev_get(dev);
+	in_dev = __in_dev_get_rcu(dev);
 	if (!in_dev)
 		goto no_in_dev;
 
@@ -818,7 +819,7 @@ no_in_dev:
 	read_lock(&dev_base_lock);
 	rcu_read_lock();
 	for (dev = dev_base; dev; dev = dev->next) {
-		if ((in_dev = __in_dev_get(dev)) == NULL)
+		if ((in_dev = __in_dev_get_rcu(dev)) == NULL)
 			continue;
 
 		for_primary_ifa(in_dev) {
@@ -887,7 +888,7 @@ u32 inet_confirm_addr(const struct net_device *dev, u32 dst, u32 local, int scop
 
 	if (dev) {
 		rcu_read_lock();
-		if ((in_dev = __in_dev_get(dev)))
+		if ((in_dev = __in_dev_get_rcu(dev)))
 			addr = confirm_addr_indev(in_dev, dst, local, scope);
 		rcu_read_unlock();
 
@@ -897,7 +898,7 @@ u32 inet_confirm_addr(const struct net_device *dev, u32 dst, u32 local, int scop
 	read_lock(&dev_base_lock);
 	rcu_read_lock();
 	for (dev = dev_base; dev; dev = dev->next) {
-		if ((in_dev = __in_dev_get(dev))) {
+		if ((in_dev = __in_dev_get_rcu(dev))) {
 			addr = confirm_addr_indev(in_dev, dst, local, scope);
 			if (addr)
 				break;
@@ -957,7 +958,7 @@ static int inetdev_event(struct notifier_block *this, unsigned long event,
 			 void *ptr)
 {
 	struct net_device *dev = ptr;
-	struct in_device *in_dev = __in_dev_get(dev);
+	struct in_device *in_dev = __in_dev_get_rtnl(dev);
 
 	ASSERT_RTNL();
 
@@ -1078,7 +1079,7 @@ static int inet_dump_ifaddr(struct sk_buff *skb, struct netlink_callback *cb)
 		if (idx > s_idx)
 			s_ip_idx = 0;
 		rcu_read_lock();
-		if ((in_dev = __in_dev_get(dev)) == NULL) {
+		if ((in_dev = __in_dev_get_rcu(dev)) == NULL) {
 			rcu_read_unlock();
 			continue;
 		}
@@ -1149,7 +1150,7 @@ void inet_forward_change(void)
 	for (dev = dev_base; dev; dev = dev->next) {
 		struct in_device *in_dev;
 		rcu_read_lock();
-		in_dev = __in_dev_get(dev);
+		in_dev = __in_dev_get_rcu(dev);
 		if (in_dev)
 			in_dev->cnf.forwarding = on;
 		rcu_read_unlock();
diff --git a/net/ipv4/esp4.c b/net/ipv4/esp4.c
index 1b5a09d1b90..1b18ce66e7b 100644
--- a/net/ipv4/esp4.c
+++ b/net/ipv4/esp4.c
@@ -5,6 +5,7 @@
 #include <net/esp.h>
 #include <asm/scatterlist.h>
 #include <linux/crypto.h>
+#include <linux/kernel.h>
 #include <linux/pfkeyv2.h>
 #include <linux/random.h>
 #include <net/icmp.h>
@@ -42,10 +43,10 @@ static int esp_output(struct xfrm_state *x, struct sk_buff *skb)
 	esp = x->data;
 	alen = esp->auth.icv_trunc_len;
 	tfm = esp->conf.tfm;
-	blksize = (crypto_tfm_alg_blocksize(tfm) + 3) & ~3;
-	clen = (clen + 2 + blksize-1)&~(blksize-1);
+	blksize = ALIGN(crypto_tfm_alg_blocksize(tfm), 4);
+	clen = ALIGN(clen + 2, blksize);
 	if (esp->conf.padlen)
-		clen = (clen + esp->conf.padlen-1)&~(esp->conf.padlen-1);
+		clen = ALIGN(clen, esp->conf.padlen);
 
 	if ((nfrags = skb_cow_data(skb, clen-skb->len+alen, &trailer)) < 0)
 		goto error;
@@ -143,7 +144,7 @@ static int esp_input(struct xfrm_state *x, struct xfrm_decap_state *decap, struc
 	struct ip_esp_hdr *esph;
 	struct esp_data *esp = x->data;
 	struct sk_buff *trailer;
-	int blksize = crypto_tfm_alg_blocksize(esp->conf.tfm);
+	int blksize = ALIGN(crypto_tfm_alg_blocksize(esp->conf.tfm), 4);
 	int alen = esp->auth.icv_trunc_len;
 	int elen = skb->len - sizeof(struct ip_esp_hdr) - esp->conf.ivlen - alen;
 	int nfrags;
@@ -304,16 +305,16 @@ static int esp_post_input(struct xfrm_state *x, struct xfrm_decap_state *decap,
 static u32 esp4_get_max_size(struct xfrm_state *x, int mtu)
 {
 	struct esp_data *esp = x->data;
-	u32 blksize = crypto_tfm_alg_blocksize(esp->conf.tfm);
+	u32 blksize = ALIGN(crypto_tfm_alg_blocksize(esp->conf.tfm), 4);
 
 	if (x->props.mode) {
-		mtu = (mtu + 2 + blksize-1)&~(blksize-1);
+		mtu = ALIGN(mtu + 2, blksize);
 	} else {
 		/* The worst case. */
-		mtu += 2 + blksize;
+		mtu = ALIGN(mtu + 2, 4) + blksize - 4;
 	}
 	if (esp->conf.padlen)
-		mtu = (mtu + esp->conf.padlen-1)&~(esp->conf.padlen-1);
+		mtu = ALIGN(mtu, esp->conf.padlen);
 
 	return mtu + x->props.header_len + esp->auth.icv_trunc_len;
 }
diff --git a/net/ipv4/fib_frontend.c b/net/ipv4/fib_frontend.c
index 4e1379f7126..e61bc7177eb 100644
--- a/net/ipv4/fib_frontend.c
+++ b/net/ipv4/fib_frontend.c
@@ -173,7 +173,7 @@ int fib_validate_source(u32 src, u32 dst, u8 tos, int oif,
 
 	no_addr = rpf = 0;
 	rcu_read_lock();
-	in_dev = __in_dev_get(dev);
+	in_dev = __in_dev_get_rcu(dev);
 	if (in_dev) {
 		no_addr = in_dev->ifa_list == NULL;
 		rpf = IN_DEV_RPFILTER(in_dev);
@@ -607,7 +607,7 @@ static int fib_inetaddr_event(struct notifier_block *this, unsigned long event,
 static int fib_netdev_event(struct notifier_block *this, unsigned long event, void *ptr)
 {
 	struct net_device *dev = ptr;
-	struct in_device *in_dev = __in_dev_get(dev);
+	struct in_device *in_dev = __in_dev_get_rtnl(dev);
 
 	if (event == NETDEV_UNREGISTER) {
 		fib_disable_ip(dev, 2);
diff --git a/net/ipv4/fib_semantics.c b/net/ipv4/fib_semantics.c
index d41219e8037..186f20c4a45 100644
--- a/net/ipv4/fib_semantics.c
+++ b/net/ipv4/fib_semantics.c
@@ -1087,7 +1087,7 @@ fib_convert_rtentry(int cmd, struct nlmsghdr *nl, struct rtmsg *rtm,
 		rta->rta_oif = &dev->ifindex;
 		if (colon) {
 			struct in_ifaddr *ifa;
-			struct in_device *in_dev = __in_dev_get(dev);
+			struct in_device *in_dev = __in_dev_get_rtnl(dev);
 			if (!in_dev)
 				return -ENODEV;
 			*colon = ':';
@@ -1268,7 +1268,7 @@ int fib_sync_up(struct net_device *dev)
 			}
 			if (nh->nh_dev == NULL || !(nh->nh_dev->flags&IFF_UP))
 				continue;
-			if (nh->nh_dev != dev || __in_dev_get(dev) == NULL)
+			if (nh->nh_dev != dev || !__in_dev_get_rtnl(dev))
 				continue;
 			alive++;
 			spin_lock_bh(&fib_multipath_lock);
diff --git a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c
index 50c0519cd70..66247f38b37 100644
--- a/net/ipv4/fib_trie.c
+++ b/net/ipv4/fib_trie.c
@@ -286,6 +286,8 @@ static inline void check_tnode(const struct tnode *tn)
 
 static int halve_threshold = 25;
 static int inflate_threshold = 50;
+static int halve_threshold_root = 15;
+static int inflate_threshold_root = 25; 
 
 
 static void __alias_free_mem(struct rcu_head *head)
@@ -449,6 +451,8 @@ static struct node *resize(struct trie *t, struct tnode *tn)
 	int i;
 	int err = 0;
 	struct tnode *old_tn;
+	int inflate_threshold_use;
+	int halve_threshold_use;
 
  	if (!tn)
 		return NULL;
@@ -541,10 +545,17 @@ static struct node *resize(struct trie *t, struct tnode *tn)
 
 	check_tnode(tn);
 
+	/* Keep root node larger  */
+
+	if(!tn->parent)
+		inflate_threshold_use = inflate_threshold_root;
+	else 
+		inflate_threshold_use = inflate_threshold;
+
 	err = 0;
 	while ((tn->full_children > 0 &&
 	       50 * (tn->full_children + tnode_child_length(tn) - tn->empty_children) >=
-				inflate_threshold * tnode_child_length(tn))) {
+				inflate_threshold_use * tnode_child_length(tn))) {
 
 		old_tn = tn;
 		tn = inflate(t, tn);
@@ -564,10 +575,18 @@ static struct node *resize(struct trie *t, struct tnode *tn)
 	 * node is above threshold.
 	 */
 
+
+	/* Keep root node larger  */
+
+	if(!tn->parent)
+		halve_threshold_use = halve_threshold_root;
+	else 
+		halve_threshold_use = halve_threshold;
+
 	err = 0;
 	while (tn->bits > 1 &&
 	       100 * (tnode_child_length(tn) - tn->empty_children) <
-	       halve_threshold * tnode_child_length(tn)) {
+	       halve_threshold_use * tnode_child_length(tn)) {
 
 		old_tn = tn;
 		tn = halve(t, tn);
@@ -2385,7 +2404,7 @@ static int fib_route_seq_show(struct seq_file *seq, void *v)
 		prefix = htonl(l->key);
 
 		list_for_each_entry_rcu(fa, &li->falh, fa_list) {
-			const struct fib_info *fi = rcu_dereference(fa->fa_info);
+			const struct fib_info *fi = fa->fa_info;
 			unsigned flags = fib_flag_trans(fa->fa_type, mask, fi);
 
 			if (fa->fa_type == RTN_BROADCAST
diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c
index 24eb56ae1b5..175e093ec56 100644
--- a/net/ipv4/icmp.c
+++ b/net/ipv4/icmp.c
@@ -188,7 +188,7 @@ struct icmp_err icmp_err_convert[] = {
 
 /* Control parameters for ECHO replies. */
 int sysctl_icmp_echo_ignore_all;
-int sysctl_icmp_echo_ignore_broadcasts;
+int sysctl_icmp_echo_ignore_broadcasts = 1;
 
 /* Control parameter - ignore bogus broadcast responses? */
 int sysctl_icmp_ignore_bogus_error_responses;
@@ -1108,12 +1108,9 @@ void __init icmp_init(struct net_proto_family *ops)
 	struct inet_sock *inet;
 	int i;
 
-	for (i = 0; i < NR_CPUS; i++) {
+	for_each_cpu(i) {
 		int err;
 
-		if (!cpu_possible(i))
-			continue;
-
 		err = sock_create_kern(PF_INET, SOCK_RAW, IPPROTO_ICMP,
 				       &per_cpu(__icmp_socket, i));
 
diff --git a/net/ipv4/igmp.c b/net/ipv4/igmp.c
index 70c44e4c3ce..8b6d3939e1e 100644
--- a/net/ipv4/igmp.c
+++ b/net/ipv4/igmp.c
@@ -1323,7 +1323,7 @@ static struct in_device * ip_mc_find_dev(struct ip_mreqn *imr)
 	}
 	if (dev) {
 		imr->imr_ifindex = dev->ifindex;
-		idev = __in_dev_get(dev);
+		idev = __in_dev_get_rtnl(dev);
 	}
 	return idev;
 }
diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index fe3c6d3d0c9..94468a76c5b 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c
@@ -494,7 +494,7 @@ void inet_csk_reqsk_queue_prune(struct sock *parent,
 EXPORT_SYMBOL_GPL(inet_csk_reqsk_queue_prune);
 
 struct sock *inet_csk_clone(struct sock *sk, const struct request_sock *req,
-			    const unsigned int __nocast priority)
+			    const gfp_t priority)
 {
 	struct sock *newsk = sk_clone(sk, priority);
 
diff --git a/net/ipv4/inet_timewait_sock.c b/net/ipv4/inet_timewait_sock.c
index 4d1502a4985..a010e9a6881 100644
--- a/net/ipv4/inet_timewait_sock.c
+++ b/net/ipv4/inet_timewait_sock.c
@@ -20,7 +20,7 @@ void __inet_twsk_kill(struct inet_timewait_sock *tw, struct inet_hashinfo *hashi
 	struct inet_bind_hashbucket *bhead;
 	struct inet_bind_bucket *tb;
 	/* Unlink from established hashes. */
-	struct inet_ehash_bucket *ehead = &hashinfo->ehash[tw->tw_hashent];
+	struct inet_ehash_bucket *ehead = inet_ehash_bucket(hashinfo, tw->tw_hash);
 
 	write_lock(&ehead->lock);
 	if (hlist_unhashed(&tw->tw_node)) {
@@ -60,7 +60,7 @@ void __inet_twsk_hashdance(struct inet_timewait_sock *tw, struct sock *sk,
 {
 	const struct inet_sock *inet = inet_sk(sk);
 	const struct inet_connection_sock *icsk = inet_csk(sk);
-	struct inet_ehash_bucket *ehead = &hashinfo->ehash[sk->sk_hashent];
+	struct inet_ehash_bucket *ehead = inet_ehash_bucket(hashinfo, sk->sk_hash);
 	struct inet_bind_hashbucket *bhead;
 	/* Step 1: Put TW into bind hash. Original socket stays there too.
 	   Note, that any socket with inet->num != 0 MUST be bound in
@@ -106,11 +106,12 @@ struct inet_timewait_sock *inet_twsk_alloc(const struct sock *sk, const int stat
 		tw->tw_dport	    = inet->dport;
 		tw->tw_family	    = sk->sk_family;
 		tw->tw_reuse	    = sk->sk_reuse;
-		tw->tw_hashent	    = sk->sk_hashent;
+		tw->tw_hash	    = sk->sk_hash;
 		tw->tw_ipv6only	    = 0;
 		tw->tw_prot	    = sk->sk_prot_creator;
 		atomic_set(&tw->tw_refcnt, 1);
 		inet_twsk_dead_node_init(tw);
+		__module_get(tw->tw_prot->owner);
 	}
 
 	return tw;
diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c
index f0d5740d7e2..896ce3f8f53 100644
--- a/net/ipv4/ip_gre.c
+++ b/net/ipv4/ip_gre.c
@@ -1104,10 +1104,10 @@ static int ipgre_open(struct net_device *dev)
 			return -EADDRNOTAVAIL;
 		dev = rt->u.dst.dev;
 		ip_rt_put(rt);
-		if (__in_dev_get(dev) == NULL)
+		if (__in_dev_get_rtnl(dev) == NULL)
 			return -EADDRNOTAVAIL;
 		t->mlink = dev->ifindex;
-		ip_mc_inc_group(__in_dev_get(dev), t->parms.iph.daddr);
+		ip_mc_inc_group(__in_dev_get_rtnl(dev), t->parms.iph.daddr);
 	}
 	return 0;
 }
diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c
index 3f1a263e124..87e350069ab 100644
--- a/net/ipv4/ip_output.c
+++ b/net/ipv4/ip_output.c
@@ -391,6 +391,9 @@ static void ip_copy_metadata(struct sk_buff *to, struct sk_buff *from)
 	to->nfct = from->nfct;
 	nf_conntrack_get(to->nfct);
 	to->nfctinfo = from->nfctinfo;
+#if defined(CONFIG_IP_VS) || defined(CONFIG_IP_VS_MODULE)
+	to->ipvs_property = from->ipvs_property;
+#endif
 #ifdef CONFIG_BRIDGE_NETFILTER
 	nf_bridge_put(to->nf_bridge);
 	to->nf_bridge = from->nf_bridge;
@@ -1020,10 +1023,7 @@ ssize_t	ip_append_page(struct sock *sk, struct page *page,
 			int alloclen;
 
 			skb_prev = skb;
-			if (skb_prev)
-				fraggap = skb_prev->len - maxfraglen;
-			else
-				fraggap = 0;
+			fraggap = skb_prev->len - maxfraglen;
 
 			alloclen = fragheaderlen + hh_len + fraggap + 15;
 			skb = sock_wmalloc(sk, alloclen, 1, sk->sk_allocation);
diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c
index 9dbf5909f3a..302b7eb507c 100644
--- a/net/ipv4/ipmr.c
+++ b/net/ipv4/ipmr.c
@@ -149,7 +149,7 @@ struct net_device *ipmr_new_tunnel(struct vifctl *v)
 		if (err == 0 && (dev = __dev_get_by_name(p.name)) != NULL) {
 			dev->flags |= IFF_MULTICAST;
 
-			in_dev = __in_dev_get(dev);
+			in_dev = __in_dev_get_rtnl(dev);
 			if (in_dev == NULL && (in_dev = inetdev_init(dev)) == NULL)
 				goto failure;
 			in_dev->cnf.rp_filter = 0;
@@ -278,7 +278,7 @@ static int vif_delete(int vifi)
 
 	dev_set_allmulti(dev, -1);
 
-	if ((in_dev = __in_dev_get(dev)) != NULL) {
+	if ((in_dev = __in_dev_get_rtnl(dev)) != NULL) {
 		in_dev->cnf.mc_forwarding--;
 		ip_rt_multicast_event(in_dev);
 	}
@@ -421,7 +421,7 @@ static int vif_add(struct vifctl *vifc, int mrtsock)
 		return -EINVAL;
 	}
 
-	if ((in_dev = __in_dev_get(dev)) == NULL)
+	if ((in_dev = __in_dev_get_rtnl(dev)) == NULL)
 		return -EADDRNOTAVAIL;
 	in_dev->cnf.mc_forwarding++;
 	dev_set_allmulti(dev, +1);
diff --git a/net/ipv4/ipvs/ip_vs_app.c b/net/ipv4/ipvs/ip_vs_app.c
index 6e092dadb38..fc6f95aaa96 100644
--- a/net/ipv4/ipvs/ip_vs_app.c
+++ b/net/ipv4/ipvs/ip_vs_app.c
@@ -604,7 +604,7 @@ static struct file_operations ip_vs_app_fops = {
 /*
  *	Replace a segment of data with a new segment
  */
-int ip_vs_skb_replace(struct sk_buff *skb, int pri,
+int ip_vs_skb_replace(struct sk_buff *skb, gfp_t pri,
 		      char *o_buf, int o_len, char *n_buf, int n_len)
 {
 	struct iphdr *iph;
diff --git a/net/ipv4/netfilter/Kconfig b/net/ipv4/netfilter/Kconfig
index 3cf9b451675..7d917e4ce1d 100644
--- a/net/ipv4/netfilter/Kconfig
+++ b/net/ipv4/netfilter/Kconfig
@@ -139,9 +139,10 @@ config IP_NF_AMANDA
 
 config IP_NF_PPTP
 	tristate  'PPTP protocol support'
+	depends on IP_NF_CONNTRACK
 	help
 	  This module adds support for PPTP (Point to Point Tunnelling
-	  Protocol, RFC2637) conncection tracking and NAT. 
+	  Protocol, RFC2637) connection tracking and NAT. 
 	
 	  If you are running PPTP sessions over a stateful firewall or NAT
 	  box, you may want to enable this feature.  
@@ -498,9 +499,14 @@ config IP_NF_TARGET_LOG
 	  To compile it as a module, choose M here.  If unsure, say N.
 
 config IP_NF_TARGET_ULOG
-	tristate "ULOG target support"
+	tristate "ULOG target support (OBSOLETE)"
 	depends on IP_NF_IPTABLES
 	---help---
+
+	  This option enables the old IPv4-only "ipt_ULOG" implementation
+	  which has been obsoleted by the new "nfnetlink_log" code (see
+	  CONFIG_NETFILTER_NETLINK_LOG).
+
 	  This option adds a `ULOG' target, which allows you to create rules in
 	  any iptables table. The packet is passed to a userspace logging
 	  daemon using netlink multicast sockets; unlike the LOG target
@@ -537,6 +543,17 @@ config IP_NF_TARGET_TCPMSS
 
 	  To compile it as a module, choose M here.  If unsure, say N.
 
+config IP_NF_TARGET_NFQUEUE
+	tristate "NFQUEUE Target Support"
+	depends on IP_NF_IPTABLES
+	help
+	  This Target replaced the old obsolete QUEUE target.
+
+	  As opposed to QUEUE, it supports 65535 different queues,
+	  not just one.
+
+	  To compile it as a module, choose M here.  If unsure, say N.
+
 # NAT + specific targets
 config IP_NF_NAT
 	tristate "Full NAT"
diff --git a/net/ipv4/netfilter/Makefile b/net/ipv4/netfilter/Makefile
index 3d45d3c0283..dab4b58dd31 100644
--- a/net/ipv4/netfilter/Makefile
+++ b/net/ipv4/netfilter/Makefile
@@ -4,7 +4,8 @@
 
 # objects for the standalone - connection tracking / NAT
 ip_conntrack-objs	:= ip_conntrack_standalone.o ip_conntrack_core.o ip_conntrack_proto_generic.o ip_conntrack_proto_tcp.o ip_conntrack_proto_udp.o ip_conntrack_proto_icmp.o
-iptable_nat-objs	:= ip_nat_standalone.o ip_nat_rule.o ip_nat_core.o ip_nat_helper.o ip_nat_proto_unknown.o ip_nat_proto_tcp.o ip_nat_proto_udp.o ip_nat_proto_icmp.o
+ip_nat-objs	:= ip_nat_core.o ip_nat_helper.o ip_nat_proto_unknown.o ip_nat_proto_tcp.o ip_nat_proto_udp.o ip_nat_proto_icmp.o
+iptable_nat-objs	:= ip_nat_rule.o ip_nat_standalone.o
 
 ip_conntrack_pptp-objs	:= ip_conntrack_helper_pptp.o ip_conntrack_proto_gre.o
 ip_nat_pptp-objs	:= ip_nat_helper_pptp.o ip_nat_proto_gre.o
@@ -40,7 +41,7 @@ obj-$(CONFIG_IP_NF_IPTABLES) += ip_tables.o
 # the three instances of ip_tables
 obj-$(CONFIG_IP_NF_FILTER) += iptable_filter.o
 obj-$(CONFIG_IP_NF_MANGLE) += iptable_mangle.o
-obj-$(CONFIG_IP_NF_NAT) += iptable_nat.o
+obj-$(CONFIG_IP_NF_NAT) += iptable_nat.o ip_nat.o
 obj-$(CONFIG_IP_NF_RAW) += iptable_raw.o
 
 # matches
@@ -92,6 +93,7 @@ obj-$(CONFIG_IP_NF_TARGET_TCPMSS) += ipt_TCPMSS.o
 obj-$(CONFIG_IP_NF_TARGET_NOTRACK) += ipt_NOTRACK.o
 obj-$(CONFIG_IP_NF_TARGET_CLUSTERIP) += ipt_CLUSTERIP.o
 obj-$(CONFIG_IP_NF_TARGET_TTL) += ipt_TTL.o
+obj-$(CONFIG_IP_NF_TARGET_NFQUEUE) += ipt_NFQUEUE.o
 
 # generic ARP tables
 obj-$(CONFIG_IP_NF_ARPTABLES) += arp_tables.o
@@ -101,4 +103,3 @@ obj-$(CONFIG_IP_NF_ARP_MANGLE) += arpt_mangle.o
 obj-$(CONFIG_IP_NF_ARPFILTER) += arptable_filter.o
 
 obj-$(CONFIG_IP_NF_QUEUE) += ip_queue.o
-obj-$(CONFIG_NETFILTER_NETLINK_QUEUE) += ipt_NFQUEUE.o
diff --git a/net/ipv4/netfilter/arp_tables.c b/net/ipv4/netfilter/arp_tables.c
index fa163425668..a7969286e6e 100644
--- a/net/ipv4/netfilter/arp_tables.c
+++ b/net/ipv4/netfilter/arp_tables.c
@@ -716,8 +716,10 @@ static int translate_table(const char *name,
 	}
 
 	/* And one copy for every other CPU */
-	for (i = 1; i < num_possible_cpus(); i++) {
-		memcpy(newinfo->entries + SMP_ALIGN(newinfo->size)*i,
+	for_each_cpu(i) {
+		if (i == 0)
+			continue;
+		memcpy(newinfo->entries + SMP_ALIGN(newinfo->size) * i,
 		       newinfo->entries,
 		       SMP_ALIGN(newinfo->size));
 	}
@@ -767,7 +769,7 @@ static void get_counters(const struct arpt_table_info *t,
 	unsigned int cpu;
 	unsigned int i;
 
-	for (cpu = 0; cpu < num_possible_cpus(); cpu++) {
+	for_each_cpu(cpu) {
 		i = 0;
 		ARPT_ENTRY_ITERATE(t->entries + TABLE_OFFSET(t, cpu),
 				   t->size,
@@ -885,7 +887,8 @@ static int do_replace(void __user *user, unsigned int len)
 		return -ENOMEM;
 
 	newinfo = vmalloc(sizeof(struct arpt_table_info)
-			  + SMP_ALIGN(tmp.size) * num_possible_cpus());
+			  + SMP_ALIGN(tmp.size) *
+			  		(highest_possible_processor_id()+1));
 	if (!newinfo)
 		return -ENOMEM;
 
@@ -1158,7 +1161,8 @@ int arpt_register_table(struct arpt_table *table,
 		= { 0, 0, 0, { 0 }, { 0 }, { } };
 
 	newinfo = vmalloc(sizeof(struct arpt_table_info)
-			  + SMP_ALIGN(repl->size) * num_possible_cpus());
+			  + SMP_ALIGN(repl->size) *
+			  		(highest_possible_processor_id()+1));
 	if (!newinfo) {
 		ret = -ENOMEM;
 		return ret;
diff --git a/net/ipv4/netfilter/ip_conntrack_amanda.c b/net/ipv4/netfilter/ip_conntrack_amanda.c
index dc20881004b..fa3f914117e 100644
--- a/net/ipv4/netfilter/ip_conntrack_amanda.c
+++ b/net/ipv4/netfilter/ip_conntrack_amanda.c
@@ -65,7 +65,7 @@ static int help(struct sk_buff **pskb,
 
 	/* increase the UDP timeout of the master connection as replies from
 	 * Amanda clients to the server can be quite delayed */
-	ip_ct_refresh_acct(ct, ctinfo, NULL, master_timeout * HZ);
+	ip_ct_refresh(ct, *pskb, master_timeout * HZ);
 
 	/* No data? */
 	dataoff = (*pskb)->nh.iph->ihl*4 + sizeof(struct udphdr);
diff --git a/net/ipv4/netfilter/ip_conntrack_core.c b/net/ipv4/netfilter/ip_conntrack_core.c
index c1f82e0c81c..422ab68ee7f 100644
--- a/net/ipv4/netfilter/ip_conntrack_core.c
+++ b/net/ipv4/netfilter/ip_conntrack_core.c
@@ -50,7 +50,7 @@
 #include <linux/netfilter_ipv4/ip_conntrack_core.h>
 #include <linux/netfilter_ipv4/listhelp.h>
 
-#define IP_CONNTRACK_VERSION	"2.3"
+#define IP_CONNTRACK_VERSION	"2.4"
 
 #if 0
 #define DEBUGP printk
@@ -148,16 +148,20 @@ DEFINE_PER_CPU(struct ip_conntrack_stat, ip_conntrack_stat);
 static int ip_conntrack_hash_rnd_initted;
 static unsigned int ip_conntrack_hash_rnd;
 
-static u_int32_t
-hash_conntrack(const struct ip_conntrack_tuple *tuple)
+static u_int32_t __hash_conntrack(const struct ip_conntrack_tuple *tuple,
+			    unsigned int size, unsigned int rnd)
 {
-#if 0
-	dump_tuple(tuple);
-#endif
 	return (jhash_3words(tuple->src.ip,
 	                     (tuple->dst.ip ^ tuple->dst.protonum),
 	                     (tuple->src.u.all | (tuple->dst.u.all << 16)),
-	                     ip_conntrack_hash_rnd) % ip_conntrack_htable_size);
+	                     rnd) % size);
+}
+
+static u_int32_t
+hash_conntrack(const struct ip_conntrack_tuple *tuple)
+{
+	return __hash_conntrack(tuple, ip_conntrack_htable_size,
+				ip_conntrack_hash_rnd);
 }
 
 int
@@ -1112,45 +1116,49 @@ void ip_conntrack_helper_unregister(struct ip_conntrack_helper *me)
 	synchronize_net();
 }
 
-static inline void ct_add_counters(struct ip_conntrack *ct,
-				   enum ip_conntrack_info ctinfo,
-				   const struct sk_buff *skb)
-{
-#ifdef CONFIG_IP_NF_CT_ACCT
-	if (skb) {
-		ct->counters[CTINFO2DIR(ctinfo)].packets++;
-		ct->counters[CTINFO2DIR(ctinfo)].bytes += 
-					ntohs(skb->nh.iph->tot_len);
-	}
-#endif
-}
-
-/* Refresh conntrack for this many jiffies and do accounting (if skb != NULL) */
-void ip_ct_refresh_acct(struct ip_conntrack *ct, 
+/* Refresh conntrack for this many jiffies and do accounting if do_acct is 1 */
+void __ip_ct_refresh_acct(struct ip_conntrack *ct, 
 		        enum ip_conntrack_info ctinfo,
 			const struct sk_buff *skb,
-			unsigned long extra_jiffies)
+			unsigned long extra_jiffies,
+			int do_acct)
 {
+	int event = 0;
+
 	IP_NF_ASSERT(ct->timeout.data == (unsigned long)ct);
+	IP_NF_ASSERT(skb);
+
+	write_lock_bh(&ip_conntrack_lock);
 
 	/* If not in hash table, timer will not be active yet */
 	if (!is_confirmed(ct)) {
 		ct->timeout.expires = extra_jiffies;
-		ct_add_counters(ct, ctinfo, skb);
+		event = IPCT_REFRESH;
 	} else {
-		write_lock_bh(&ip_conntrack_lock);
 		/* Need del_timer for race avoidance (may already be dying). */
 		if (del_timer(&ct->timeout)) {
 			ct->timeout.expires = jiffies + extra_jiffies;
 			add_timer(&ct->timeout);
-			/* FIXME: We loose some REFRESH events if this function
-			 * is called without an skb.  I'll fix this later -HW */
-			if (skb)
-				ip_conntrack_event_cache(IPCT_REFRESH, skb);
+			event = IPCT_REFRESH;
 		}
-		ct_add_counters(ct, ctinfo, skb);
-		write_unlock_bh(&ip_conntrack_lock);
 	}
+
+#ifdef CONFIG_IP_NF_CT_ACCT
+	if (do_acct) {
+		ct->counters[CTINFO2DIR(ctinfo)].packets++;
+		ct->counters[CTINFO2DIR(ctinfo)].bytes += 
+						ntohs(skb->nh.iph->tot_len);
+		if ((ct->counters[CTINFO2DIR(ctinfo)].packets & 0x80000000)
+		    || (ct->counters[CTINFO2DIR(ctinfo)].bytes & 0x80000000))
+			event |= IPCT_COUNTER_FILLING;
+	}
+#endif
+
+	write_unlock_bh(&ip_conntrack_lock);
+
+	/* must be unlocked when calling event cache */
+	if (event)
+		ip_conntrack_event_cache(event, skb);
 }
 
 #if defined(CONFIG_IP_NF_CONNTRACK_NETLINK) || \
@@ -1337,14 +1345,13 @@ static int kill_all(struct ip_conntrack *i, void *data)
 	return 1;
 }
 
-static void free_conntrack_hash(void)
+static void free_conntrack_hash(struct list_head *hash, int vmalloced,int size)
 {
-	if (ip_conntrack_vmalloc)
-		vfree(ip_conntrack_hash);
+	if (vmalloced)
+		vfree(hash);
 	else
-		free_pages((unsigned long)ip_conntrack_hash, 
-			   get_order(sizeof(struct list_head)
-				     * ip_conntrack_htable_size));
+		free_pages((unsigned long)hash, 
+			   get_order(sizeof(struct list_head) * size));
 }
 
 void ip_conntrack_flush()
@@ -1374,12 +1381,83 @@ void ip_conntrack_cleanup(void)
 	ip_conntrack_flush();
 	kmem_cache_destroy(ip_conntrack_cachep);
 	kmem_cache_destroy(ip_conntrack_expect_cachep);
-	free_conntrack_hash();
+	free_conntrack_hash(ip_conntrack_hash, ip_conntrack_vmalloc,
+			    ip_conntrack_htable_size);
 	nf_unregister_sockopt(&so_getorigdst);
 }
 
-static int hashsize;
-module_param(hashsize, int, 0400);
+static struct list_head *alloc_hashtable(int size, int *vmalloced)
+{
+	struct list_head *hash;
+	unsigned int i;
+
+	*vmalloced = 0; 
+	hash = (void*)__get_free_pages(GFP_KERNEL, 
+				       get_order(sizeof(struct list_head)
+						 * size));
+	if (!hash) { 
+		*vmalloced = 1;
+		printk(KERN_WARNING"ip_conntrack: falling back to vmalloc.\n");
+		hash = vmalloc(sizeof(struct list_head) * size);
+	}
+
+	if (hash)
+		for (i = 0; i < size; i++)
+			INIT_LIST_HEAD(&hash[i]);
+
+	return hash;
+}
+
+int set_hashsize(const char *val, struct kernel_param *kp)
+{
+	int i, bucket, hashsize, vmalloced;
+	int old_vmalloced, old_size;
+	int rnd;
+	struct list_head *hash, *old_hash;
+	struct ip_conntrack_tuple_hash *h;
+
+	/* On boot, we can set this without any fancy locking. */
+	if (!ip_conntrack_htable_size)
+		return param_set_int(val, kp);
+
+	hashsize = simple_strtol(val, NULL, 0);
+	if (!hashsize)
+		return -EINVAL;
+
+	hash = alloc_hashtable(hashsize, &vmalloced);
+	if (!hash)
+		return -ENOMEM;
+
+	/* We have to rehash for the new table anyway, so we also can 
+	 * use a new random seed */
+	get_random_bytes(&rnd, 4);
+
+	write_lock_bh(&ip_conntrack_lock);
+	for (i = 0; i < ip_conntrack_htable_size; i++) {
+		while (!list_empty(&ip_conntrack_hash[i])) {
+			h = list_entry(ip_conntrack_hash[i].next,
+				       struct ip_conntrack_tuple_hash, list);
+			list_del(&h->list);
+			bucket = __hash_conntrack(&h->tuple, hashsize, rnd);
+			list_add_tail(&h->list, &hash[bucket]);
+		}
+	}
+	old_size = ip_conntrack_htable_size;
+	old_vmalloced = ip_conntrack_vmalloc;
+	old_hash = ip_conntrack_hash;
+
+	ip_conntrack_htable_size = hashsize;
+	ip_conntrack_vmalloc = vmalloced;
+	ip_conntrack_hash = hash;
+	ip_conntrack_hash_rnd = rnd;
+	write_unlock_bh(&ip_conntrack_lock);
+
+	free_conntrack_hash(old_hash, old_vmalloced, old_size);
+	return 0;
+}
+
+module_param_call(hashsize, set_hashsize, param_get_uint,
+		  &ip_conntrack_htable_size, 0600);
 
 int __init ip_conntrack_init(void)
 {
@@ -1388,9 +1466,7 @@ int __init ip_conntrack_init(void)
 
 	/* Idea from tcp.c: use 1/16384 of memory.  On i386: 32MB
 	 * machine has 256 buckets.  >= 1GB machines have 8192 buckets. */
- 	if (hashsize) {
- 		ip_conntrack_htable_size = hashsize;
- 	} else {
+ 	if (!ip_conntrack_htable_size) {
 		ip_conntrack_htable_size
 			= (((num_physpages << PAGE_SHIFT) / 16384)
 			   / sizeof(struct list_head));
@@ -1412,20 +1488,8 @@ int __init ip_conntrack_init(void)
 		return ret;
 	}
 
-	/* AK: the hash table is twice as big than needed because it
-	   uses list_head.  it would be much nicer to caches to use a
-	   single pointer list head here. */
-	ip_conntrack_vmalloc = 0; 
-	ip_conntrack_hash 
-		=(void*)__get_free_pages(GFP_KERNEL, 
-					 get_order(sizeof(struct list_head)
-						   *ip_conntrack_htable_size));
-	if (!ip_conntrack_hash) { 
-		ip_conntrack_vmalloc = 1;
-		printk(KERN_WARNING "ip_conntrack: falling back to vmalloc.\n");
-		ip_conntrack_hash = vmalloc(sizeof(struct list_head)
-					    * ip_conntrack_htable_size);
-	}
+	ip_conntrack_hash = alloc_hashtable(ip_conntrack_htable_size,
+					    &ip_conntrack_vmalloc);
 	if (!ip_conntrack_hash) {
 		printk(KERN_ERR "Unable to create ip_conntrack_hash\n");
 		goto err_unreg_sockopt;
@@ -1457,9 +1521,6 @@ int __init ip_conntrack_init(void)
 	ip_ct_protos[IPPROTO_ICMP] = &ip_conntrack_protocol_icmp;
 	write_unlock_bh(&ip_conntrack_lock);
 
-	for (i = 0; i < ip_conntrack_htable_size; i++)
-		INIT_LIST_HEAD(&ip_conntrack_hash[i]);
-
 	/* For use by ipt_REJECT */
 	ip_ct_attach = ip_conntrack_attach;
 
@@ -1474,7 +1535,8 @@ int __init ip_conntrack_init(void)
 err_free_conntrack_slab:
 	kmem_cache_destroy(ip_conntrack_cachep);
 err_free_hash:
-	free_conntrack_hash();
+	free_conntrack_hash(ip_conntrack_hash, ip_conntrack_vmalloc,
+			    ip_conntrack_htable_size);
 err_unreg_sockopt:
 	nf_unregister_sockopt(&so_getorigdst);
 
diff --git a/net/ipv4/netfilter/ip_conntrack_helper_pptp.c b/net/ipv4/netfilter/ip_conntrack_helper_pptp.c
index 79db5b70d5f..926a6684643 100644
--- a/net/ipv4/netfilter/ip_conntrack_helper_pptp.c
+++ b/net/ipv4/netfilter/ip_conntrack_helper_pptp.c
@@ -172,7 +172,6 @@ static int destroy_sibling_or_exp(const struct ip_conntrack_tuple *t)
 		DEBUGP("setting timeout of conntrack %p to 0\n", sibling);
 		sibling->proto.gre.timeout = 0;
 		sibling->proto.gre.stream_timeout = 0;
-		/* refresh_acct will not modify counters if skb == NULL */
 		if (del_timer(&sibling->timeout))
 			sibling->timeout.function((unsigned long)sibling);
 		ip_conntrack_put(sibling);
@@ -223,8 +222,8 @@ static void pptp_destroy_siblings(struct ip_conntrack *ct)
 static inline int
 exp_gre(struct ip_conntrack *master,
 	u_int32_t seq,
-	u_int16_t callid,
-	u_int16_t peer_callid)
+	__be16 callid,
+	__be16 peer_callid)
 {
 	struct ip_conntrack_tuple inv_tuple;
 	struct ip_conntrack_tuple exp_tuples[] = {
@@ -263,7 +262,7 @@ exp_gre(struct ip_conntrack *master,
 	exp_orig->mask.src.ip = 0xffffffff;
 	exp_orig->mask.src.u.all = 0;
 	exp_orig->mask.dst.u.all = 0;
-	exp_orig->mask.dst.u.gre.key = 0xffff;
+	exp_orig->mask.dst.u.gre.key = htons(0xffff);
 	exp_orig->mask.dst.ip = 0xffffffff;
 	exp_orig->mask.dst.protonum = 0xff;
 		
@@ -340,7 +339,8 @@ pptp_inbound_pkt(struct sk_buff **pskb,
 	unsigned int reqlen;
 	union pptp_ctrl_union _pptpReq, *pptpReq;
 	struct ip_ct_pptp_master *info = &ct->help.ct_pptp_info;
-	u_int16_t msg, *cid, *pcid;
+	u_int16_t msg;
+	__be16 *cid, *pcid;
 	u_int32_t seq;	
 
 	ctlh = skb_header_pointer(*pskb, nexthdr_off, sizeof(_ctlh), &_ctlh);
@@ -485,7 +485,7 @@ pptp_inbound_pkt(struct sk_buff **pskb,
 
 		if (info->pns_call_id != ntohs(*pcid)) {
 			DEBUGP("%s for unknown CallID %u\n", 
-				pptp_msg_name[msg], ntohs(*cid));
+				pptp_msg_name[msg], ntohs(*pcid));
 			break;
 		}
 
@@ -551,7 +551,8 @@ pptp_outbound_pkt(struct sk_buff **pskb,
 	unsigned int reqlen;
 	union pptp_ctrl_union _pptpReq, *pptpReq;
 	struct ip_ct_pptp_master *info = &ct->help.ct_pptp_info;
-	u_int16_t msg, *cid, *pcid;
+	u_int16_t msg;
+	__be16 *cid, *pcid;
 
 	ctlh = skb_header_pointer(*pskb, nexthdr_off, sizeof(_ctlh), &_ctlh);
 	if (!ctlh)
@@ -755,7 +756,7 @@ static struct ip_conntrack_helper pptp = {
 			  } 
 		 },
 	.mask = { .src = { .ip = 0, 
-			   .u = { .tcp = { .port = 0xffff } } 
+			   .u = { .tcp = { .port = __constant_htons(0xffff) } } 
 			 }, 
 		  .dst = { .ip = 0, 
 			   .u = { .all = 0 },
diff --git a/net/ipv4/netfilter/ip_conntrack_netbios_ns.c b/net/ipv4/netfilter/ip_conntrack_netbios_ns.c
index 71ef19d126d..186646eb249 100644
--- a/net/ipv4/netfilter/ip_conntrack_netbios_ns.c
+++ b/net/ipv4/netfilter/ip_conntrack_netbios_ns.c
@@ -58,7 +58,7 @@ static int help(struct sk_buff **pskb,
 		goto out;
 
 	rcu_read_lock();
-	in_dev = __in_dev_get(rt->u.dst.dev);
+	in_dev = __in_dev_get_rcu(rt->u.dst.dev);
 	if (in_dev != NULL) {
 		for_primary_ifa(in_dev) {
 			if (ifa->ifa_broadcast == iph->daddr) {
@@ -91,7 +91,7 @@ static int help(struct sk_buff **pskb,
 	ip_conntrack_expect_related(exp);
 	ip_conntrack_expect_put(exp);
 
-	ip_ct_refresh_acct(ct, ctinfo, NULL, timeout * HZ);
+	ip_ct_refresh(ct, *pskb, timeout * HZ);
 out:
 	return NF_ACCEPT;
 }
diff --git a/net/ipv4/netfilter/ip_conntrack_netlink.c b/net/ipv4/netfilter/ip_conntrack_netlink.c
index b08a432efcf..166e6069f12 100644
--- a/net/ipv4/netfilter/ip_conntrack_netlink.c
+++ b/net/ipv4/netfilter/ip_conntrack_netlink.c
@@ -177,11 +177,11 @@ ctnetlink_dump_counters(struct sk_buff *skb, const struct ip_conntrack *ct,
 	struct nfattr *nest_count = NFA_NEST(skb, type);
 	u_int64_t tmp;
 
-	tmp = cpu_to_be64(ct->counters[dir].packets);
-	NFA_PUT(skb, CTA_COUNTERS_PACKETS, sizeof(u_int64_t), &tmp);
+	tmp = htonl(ct->counters[dir].packets);
+	NFA_PUT(skb, CTA_COUNTERS32_PACKETS, sizeof(u_int32_t), &tmp);
 
-	tmp = cpu_to_be64(ct->counters[dir].bytes);
-	NFA_PUT(skb, CTA_COUNTERS_BYTES, sizeof(u_int64_t), &tmp);
+	tmp = htonl(ct->counters[dir].bytes);
+	NFA_PUT(skb, CTA_COUNTERS32_BYTES, sizeof(u_int32_t), &tmp);
 
 	NFA_NEST_END(skb, nest_count);
 
@@ -833,7 +833,8 @@ out:
 static inline int
 ctnetlink_change_status(struct ip_conntrack *ct, struct nfattr *cda[])
 {
-	unsigned long d, status = *(u_int32_t *)NFA_DATA(cda[CTA_STATUS-1]);
+	unsigned long d;
+	unsigned status = ntohl(*(u_int32_t *)NFA_DATA(cda[CTA_STATUS-1]));
 	d = ct->status ^ status;
 
 	if (d & (IPS_EXPECTED|IPS_CONFIRMED|IPS_DYING))
@@ -948,6 +949,31 @@ ctnetlink_change_timeout(struct ip_conntrack *ct, struct nfattr *cda[])
 	return 0;
 }
 
+static inline int
+ctnetlink_change_protoinfo(struct ip_conntrack *ct, struct nfattr *cda[])
+{
+	struct nfattr *tb[CTA_PROTOINFO_MAX], *attr = cda[CTA_PROTOINFO-1];
+	struct ip_conntrack_protocol *proto;
+	u_int16_t npt = ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple.dst.protonum;
+	int err = 0;
+
+	if (nfattr_parse_nested(tb, CTA_PROTOINFO_MAX, attr) < 0)
+		goto nfattr_failure;
+
+	proto = ip_conntrack_proto_find_get(npt);
+	if (!proto)
+		return -EINVAL;
+
+	if (proto->from_nfattr)
+		err = proto->from_nfattr(tb, ct);
+	ip_conntrack_proto_put(proto); 
+
+	return err;
+
+nfattr_failure:
+	return -ENOMEM;
+}
+
 static int
 ctnetlink_change_conntrack(struct ip_conntrack *ct, struct nfattr *cda[])
 {
@@ -973,6 +999,12 @@ ctnetlink_change_conntrack(struct ip_conntrack *ct, struct nfattr *cda[])
 			return err;
 	}
 
+	if (cda[CTA_PROTOINFO-1]) {
+		err = ctnetlink_change_protoinfo(ct, cda);
+		if (err < 0)
+			return err;
+	}
+
 	DEBUGP("all done\n");
 	return 0;
 }
@@ -1002,6 +1034,12 @@ ctnetlink_create_conntrack(struct nfattr *cda[],
 	if (err < 0)
 		goto err;
 
+	if (cda[CTA_PROTOINFO-1]) {
+		err = ctnetlink_change_protoinfo(ct, cda);
+		if (err < 0)
+			return err;
+	}
+
 	ct->helper = ip_conntrack_helper_find_get(rtuple);
 
 	add_timer(&ct->timeout);
diff --git a/net/ipv4/netfilter/ip_conntrack_proto_gre.c b/net/ipv4/netfilter/ip_conntrack_proto_gre.c
index de3cb9db6f8..744abb9d377 100644
--- a/net/ipv4/netfilter/ip_conntrack_proto_gre.c
+++ b/net/ipv4/netfilter/ip_conntrack_proto_gre.c
@@ -247,6 +247,7 @@ static int gre_packet(struct ip_conntrack *ct,
 				   ct->proto.gre.stream_timeout);
 		/* Also, more likely to be important, and not a probe. */
 		set_bit(IPS_ASSURED_BIT, &ct->status);
+		ip_conntrack_event_cache(IPCT_STATUS, skb);
 	} else
 		ip_ct_refresh_acct(ct, conntrackinfo, skb,
 				   ct->proto.gre.timeout);
diff --git a/net/ipv4/netfilter/ip_conntrack_proto_icmp.c b/net/ipv4/netfilter/ip_conntrack_proto_icmp.c
index 838d1d69b36..98f0015dd25 100644
--- a/net/ipv4/netfilter/ip_conntrack_proto_icmp.c
+++ b/net/ipv4/netfilter/ip_conntrack_proto_icmp.c
@@ -296,8 +296,7 @@ static int icmp_nfattr_to_tuple(struct nfattr *tb[],
 				struct ip_conntrack_tuple *tuple)
 {
 	if (!tb[CTA_PROTO_ICMP_TYPE-1]
-	    || !tb[CTA_PROTO_ICMP_CODE-1]
-	    || !tb[CTA_PROTO_ICMP_ID-1])
+	    || !tb[CTA_PROTO_ICMP_CODE-1])
 		return -1;
 
 	tuple->dst.u.icmp.type = 
diff --git a/net/ipv4/netfilter/ip_conntrack_proto_sctp.c b/net/ipv4/netfilter/ip_conntrack_proto_sctp.c
index a875f35e576..59a4a0111dd 100644
--- a/net/ipv4/netfilter/ip_conntrack_proto_sctp.c
+++ b/net/ipv4/netfilter/ip_conntrack_proto_sctp.c
@@ -416,6 +416,7 @@ static int sctp_packet(struct ip_conntrack *conntrack,
 		&& newconntrack == SCTP_CONNTRACK_ESTABLISHED) {
 		DEBUGP("Setting assured bit\n");
 		set_bit(IPS_ASSURED_BIT, &conntrack->status);
+		ip_conntrack_event_cache(IPCT_STATUS, skb);
 	}
 
 	return NF_ACCEPT;
diff --git a/net/ipv4/netfilter/ip_conntrack_proto_tcp.c b/net/ipv4/netfilter/ip_conntrack_proto_tcp.c
index 1985abc59d2..d6701cafbcc 100644
--- a/net/ipv4/netfilter/ip_conntrack_proto_tcp.c
+++ b/net/ipv4/netfilter/ip_conntrack_proto_tcp.c
@@ -341,17 +341,43 @@ static int tcp_print_conntrack(struct seq_file *s,
 static int tcp_to_nfattr(struct sk_buff *skb, struct nfattr *nfa,
 			 const struct ip_conntrack *ct)
 {
+	struct nfattr *nest_parms = NFA_NEST(skb, CTA_PROTOINFO_TCP);
+	
 	read_lock_bh(&tcp_lock);
 	NFA_PUT(skb, CTA_PROTOINFO_TCP_STATE, sizeof(u_int8_t),
 		&ct->proto.tcp.state);
 	read_unlock_bh(&tcp_lock);
 
+	NFA_NEST_END(skb, nest_parms);
+
 	return 0;
 
 nfattr_failure:
 	read_unlock_bh(&tcp_lock);
 	return -1;
 }
+
+static int nfattr_to_tcp(struct nfattr *cda[], struct ip_conntrack *ct)
+{
+	struct nfattr *attr = cda[CTA_PROTOINFO_TCP-1];
+	struct nfattr *tb[CTA_PROTOINFO_TCP_MAX];
+
+        if (nfattr_parse_nested(tb, CTA_PROTOINFO_TCP_MAX, attr) < 0)
+                goto nfattr_failure;
+
+	if (!tb[CTA_PROTOINFO_TCP_STATE-1])
+		return -EINVAL;
+
+	write_lock_bh(&tcp_lock);
+	ct->proto.tcp.state = 
+		*(u_int8_t *)NFA_DATA(tb[CTA_PROTOINFO_TCP_STATE-1]);
+	write_unlock_bh(&tcp_lock);
+
+	return 0;
+
+nfattr_failure:
+	return -1;
+}
 #endif
 
 static unsigned int get_conntrack_index(const struct tcphdr *tcph)
@@ -1014,7 +1040,8 @@ static int tcp_packet(struct ip_conntrack *conntrack,
 		/* Set ASSURED if we see see valid ack in ESTABLISHED 
 		   after SYN_RECV or a valid answer for a picked up 
 		   connection. */
-			set_bit(IPS_ASSURED_BIT, &conntrack->status);
+		set_bit(IPS_ASSURED_BIT, &conntrack->status);
+		ip_conntrack_event_cache(IPCT_STATUS, skb);
 	}
 	ip_ct_refresh_acct(conntrack, ctinfo, skb, timeout);
 
@@ -1122,6 +1149,7 @@ struct ip_conntrack_protocol ip_conntrack_protocol_tcp =
 #if defined(CONFIG_IP_NF_CONNTRACK_NETLINK) || \
     defined(CONFIG_IP_NF_CONNTRACK_NETLINK_MODULE)
 	.to_nfattr		= tcp_to_nfattr,
+	.from_nfattr		= nfattr_to_tcp,
 	.tuple_to_nfattr	= ip_ct_port_tuple_to_nfattr,
 	.nfattr_to_tuple	= ip_ct_port_nfattr_to_tuple,
 #endif
diff --git a/net/ipv4/netfilter/ip_conntrack_standalone.c b/net/ipv4/netfilter/ip_conntrack_standalone.c
index d3c7808010e..dd476b191f4 100644
--- a/net/ipv4/netfilter/ip_conntrack_standalone.c
+++ b/net/ipv4/netfilter/ip_conntrack_standalone.c
@@ -989,7 +989,7 @@ EXPORT_SYMBOL(need_ip_conntrack);
 EXPORT_SYMBOL(ip_conntrack_helper_register);
 EXPORT_SYMBOL(ip_conntrack_helper_unregister);
 EXPORT_SYMBOL(ip_ct_iterate_cleanup);
-EXPORT_SYMBOL(ip_ct_refresh_acct);
+EXPORT_SYMBOL(__ip_ct_refresh_acct);
 
 EXPORT_SYMBOL(ip_conntrack_expect_alloc);
 EXPORT_SYMBOL(ip_conntrack_expect_put);
diff --git a/net/ipv4/netfilter/ip_nat_core.c b/net/ipv4/netfilter/ip_nat_core.c
index c3ea891d38e..c5e3abd2467 100644
--- a/net/ipv4/netfilter/ip_nat_core.c
+++ b/net/ipv4/netfilter/ip_nat_core.c
@@ -74,12 +74,14 @@ ip_nat_proto_find_get(u_int8_t protonum)
 
 	return p;
 }
+EXPORT_SYMBOL_GPL(ip_nat_proto_find_get);
 
 void
 ip_nat_proto_put(struct ip_nat_protocol *p)
 {
 	module_put(p->me);
 }
+EXPORT_SYMBOL_GPL(ip_nat_proto_put);
 
 /* We keep an extra hash for each conntrack, for fast searching. */
 static inline unsigned int
@@ -111,6 +113,7 @@ ip_nat_cheat_check(u_int32_t oldvalinv, u_int32_t newval, u_int16_t oldcheck)
 	return csum_fold(csum_partial((char *)diffs, sizeof(diffs),
 				      oldcheck^0xFFFF));
 }
+EXPORT_SYMBOL(ip_nat_cheat_check);
 
 /* Is this tuple already taken? (not by us) */
 int
@@ -127,6 +130,7 @@ ip_nat_used_tuple(const struct ip_conntrack_tuple *tuple,
 	invert_tuplepr(&reply, tuple);
 	return ip_conntrack_tuple_taken(&reply, ignored_conntrack);
 }
+EXPORT_SYMBOL(ip_nat_used_tuple);
 
 /* If we source map this tuple so reply looks like reply_tuple, will
  * that meet the constraints of range. */
@@ -347,6 +351,7 @@ ip_nat_setup_info(struct ip_conntrack *conntrack,
 
 	return NF_ACCEPT;
 }
+EXPORT_SYMBOL(ip_nat_setup_info);
 
 /* Returns true if succeeded. */
 static int
@@ -387,10 +392,10 @@ manip_pkt(u_int16_t proto,
 }
 
 /* Do packet manipulations according to ip_nat_setup_info. */
-unsigned int nat_packet(struct ip_conntrack *ct,
-			enum ip_conntrack_info ctinfo,
-			unsigned int hooknum,
-			struct sk_buff **pskb)
+unsigned int ip_nat_packet(struct ip_conntrack *ct,
+			   enum ip_conntrack_info ctinfo,
+			   unsigned int hooknum,
+			   struct sk_buff **pskb)
 {
 	enum ip_conntrack_dir dir = CTINFO2DIR(ctinfo);
 	unsigned long statusbit;
@@ -417,12 +422,13 @@ unsigned int nat_packet(struct ip_conntrack *ct,
 	}
 	return NF_ACCEPT;
 }
+EXPORT_SYMBOL_GPL(ip_nat_packet);
 
 /* Dir is direction ICMP is coming from (opposite to packet it contains) */
-int icmp_reply_translation(struct sk_buff **pskb,
-			   struct ip_conntrack *ct,
-			   enum ip_nat_manip_type manip,
-			   enum ip_conntrack_dir dir)
+int ip_nat_icmp_reply_translation(struct sk_buff **pskb,
+				  struct ip_conntrack *ct,
+				  enum ip_nat_manip_type manip,
+				  enum ip_conntrack_dir dir)
 {
 	struct {
 		struct icmphdr icmp;
@@ -509,6 +515,7 @@ int icmp_reply_translation(struct sk_buff **pskb,
 
 	return 1;
 }
+EXPORT_SYMBOL_GPL(ip_nat_icmp_reply_translation);
 
 /* Protocol registration. */
 int ip_nat_protocol_register(struct ip_nat_protocol *proto)
@@ -525,6 +532,7 @@ int ip_nat_protocol_register(struct ip_nat_protocol *proto)
 	write_unlock_bh(&ip_nat_lock);
 	return ret;
 }
+EXPORT_SYMBOL(ip_nat_protocol_register);
 
 /* Noone stores the protocol anywhere; simply delete it. */
 void ip_nat_protocol_unregister(struct ip_nat_protocol *proto)
@@ -536,6 +544,7 @@ void ip_nat_protocol_unregister(struct ip_nat_protocol *proto)
 	/* Someone could be still looking at the proto in a bh. */
 	synchronize_net();
 }
+EXPORT_SYMBOL(ip_nat_protocol_unregister);
 
 #if defined(CONFIG_IP_NF_CONNTRACK_NETLINK) || \
     defined(CONFIG_IP_NF_CONNTRACK_NETLINK_MODULE)
@@ -582,7 +591,7 @@ EXPORT_SYMBOL_GPL(ip_nat_port_nfattr_to_range);
 EXPORT_SYMBOL_GPL(ip_nat_port_range_to_nfattr);
 #endif
 
-int __init ip_nat_init(void)
+static int __init ip_nat_init(void)
 {
 	size_t i;
 
@@ -624,10 +633,14 @@ static int clean_nat(struct ip_conntrack *i, void *data)
 	return 0;
 }
 
-/* Not __exit: called from ip_nat_standalone.c:init_or_cleanup() --RR */
-void ip_nat_cleanup(void)
+static void __exit ip_nat_cleanup(void)
 {
 	ip_ct_iterate_cleanup(&clean_nat, NULL);
 	ip_conntrack_destroyed = NULL;
 	vfree(bysource);
 }
+
+MODULE_LICENSE("GPL");
+
+module_init(ip_nat_init);
+module_exit(ip_nat_cleanup);
diff --git a/net/ipv4/netfilter/ip_nat_helper.c b/net/ipv4/netfilter/ip_nat_helper.c
index d2dd5d31355..5d506e0564d 100644
--- a/net/ipv4/netfilter/ip_nat_helper.c
+++ b/net/ipv4/netfilter/ip_nat_helper.c
@@ -199,6 +199,7 @@ ip_nat_mangle_tcp_packet(struct sk_buff **pskb,
 	}
 	return 1;
 }
+EXPORT_SYMBOL(ip_nat_mangle_tcp_packet);
 			
 /* Generic function for mangling variable-length address changes inside
  * NATed UDP connections (like the CONNECT DATA XXXXX MESG XXXXX INDEX XXXXX
@@ -256,6 +257,7 @@ ip_nat_mangle_udp_packet(struct sk_buff **pskb,
 
 	return 1;
 }
+EXPORT_SYMBOL(ip_nat_mangle_udp_packet);
 
 /* Adjust one found SACK option including checksum correction */
 static void
@@ -399,6 +401,7 @@ ip_nat_seq_adjust(struct sk_buff **pskb,
 
 	return 1;
 }
+EXPORT_SYMBOL(ip_nat_seq_adjust);
 
 /* Setup NAT on this expected conntrack so it follows master. */
 /* If we fail to get a free NAT slot, we'll get dropped on confirm */
@@ -425,3 +428,4 @@ void ip_nat_follow_master(struct ip_conntrack *ct,
 	/* hook doesn't matter, but it has to do destination manip */
 	ip_nat_setup_info(ct, &range, NF_IP_PRE_ROUTING);
 }
+EXPORT_SYMBOL(ip_nat_follow_master);
diff --git a/net/ipv4/netfilter/ip_nat_standalone.c b/net/ipv4/netfilter/ip_nat_standalone.c
index 0ff368b131f..30cd4e18c12 100644
--- a/net/ipv4/netfilter/ip_nat_standalone.c
+++ b/net/ipv4/netfilter/ip_nat_standalone.c
@@ -108,8 +108,8 @@ ip_nat_fn(unsigned int hooknum,
 	case IP_CT_RELATED:
 	case IP_CT_RELATED+IP_CT_IS_REPLY:
 		if ((*pskb)->nh.iph->protocol == IPPROTO_ICMP) {
-			if (!icmp_reply_translation(pskb, ct, maniptype,
-						    CTINFO2DIR(ctinfo)))
+			if (!ip_nat_icmp_reply_translation(pskb, ct, maniptype,
+							   CTINFO2DIR(ctinfo)))
 				return NF_DROP;
 			else
 				return NF_ACCEPT;
@@ -152,7 +152,7 @@ ip_nat_fn(unsigned int hooknum,
 	}
 
 	IP_NF_ASSERT(info);
-	return nat_packet(ct, ctinfo, hooknum, pskb);
+	return ip_nat_packet(ct, ctinfo, hooknum, pskb);
 }
 
 static unsigned int
@@ -325,15 +325,10 @@ static int init_or_cleanup(int init)
 		printk("ip_nat_init: can't setup rules.\n");
 		goto cleanup_nothing;
 	}
-	ret = ip_nat_init();
-	if (ret < 0) {
-		printk("ip_nat_init: can't setup rules.\n");
-		goto cleanup_rule_init;
-	}
 	ret = nf_register_hook(&ip_nat_in_ops);
 	if (ret < 0) {
 		printk("ip_nat_init: can't register in hook.\n");
-		goto cleanup_nat;
+		goto cleanup_rule_init;
 	}
 	ret = nf_register_hook(&ip_nat_out_ops);
 	if (ret < 0) {
@@ -374,8 +369,6 @@ static int init_or_cleanup(int init)
 	nf_unregister_hook(&ip_nat_out_ops);
  cleanup_inops:
 	nf_unregister_hook(&ip_nat_in_ops);
- cleanup_nat:
-	ip_nat_cleanup();
  cleanup_rule_init:
 	ip_nat_rule_cleanup();
  cleanup_nothing:
@@ -395,14 +388,4 @@ static void __exit fini(void)
 module_init(init);
 module_exit(fini);
 
-EXPORT_SYMBOL(ip_nat_setup_info);
-EXPORT_SYMBOL(ip_nat_protocol_register);
-EXPORT_SYMBOL(ip_nat_protocol_unregister);
-EXPORT_SYMBOL_GPL(ip_nat_proto_find_get);
-EXPORT_SYMBOL_GPL(ip_nat_proto_put);
-EXPORT_SYMBOL(ip_nat_cheat_check);
-EXPORT_SYMBOL(ip_nat_mangle_tcp_packet);
-EXPORT_SYMBOL(ip_nat_mangle_udp_packet);
-EXPORT_SYMBOL(ip_nat_used_tuple);
-EXPORT_SYMBOL(ip_nat_follow_master);
 MODULE_LICENSE("GPL");
diff --git a/net/ipv4/netfilter/ip_queue.c b/net/ipv4/netfilter/ip_queue.c
index d54f14d926f..36339eb39e1 100644
--- a/net/ipv4/netfilter/ip_queue.c
+++ b/net/ipv4/netfilter/ip_queue.c
@@ -240,8 +240,8 @@ ipq_build_packet_message(struct ipq_queue_entry *entry, int *errp)
 
 	pmsg->packet_id       = (unsigned long )entry;
 	pmsg->data_len        = data_len;
-	pmsg->timestamp_sec   = skb_tv_base.tv_sec + entry->skb->tstamp.off_sec;
-	pmsg->timestamp_usec  = skb_tv_base.tv_usec + entry->skb->tstamp.off_usec;
+	pmsg->timestamp_sec   = entry->skb->tstamp.off_sec;
+	pmsg->timestamp_usec  = entry->skb->tstamp.off_usec;
 	pmsg->mark            = entry->skb->nfmark;
 	pmsg->hook            = entry->info->hook;
 	pmsg->hw_protocol     = entry->skb->protocol;
diff --git a/net/ipv4/netfilter/ip_tables.c b/net/ipv4/netfilter/ip_tables.c
index eef99a1b5de..75c27e92f6a 100644
--- a/net/ipv4/netfilter/ip_tables.c
+++ b/net/ipv4/netfilter/ip_tables.c
@@ -27,6 +27,7 @@
 #include <asm/semaphore.h>
 #include <linux/proc_fs.h>
 #include <linux/err.h>
+#include <linux/cpumask.h>
 
 #include <linux/netfilter_ipv4/ip_tables.h>
 
@@ -921,8 +922,10 @@ translate_table(const char *name,
 	}
 
 	/* And one copy for every other CPU */
-	for (i = 1; i < num_possible_cpus(); i++) {
-		memcpy(newinfo->entries + SMP_ALIGN(newinfo->size)*i,
+	for_each_cpu(i) {
+		if (i == 0)
+			continue;
+		memcpy(newinfo->entries + SMP_ALIGN(newinfo->size) * i,
 		       newinfo->entries,
 		       SMP_ALIGN(newinfo->size));
 	}
@@ -943,7 +946,7 @@ replace_table(struct ipt_table *table,
 		struct ipt_entry *table_base;
 		unsigned int i;
 
-		for (i = 0; i < num_possible_cpus(); i++) {
+		for_each_cpu(i) {
 			table_base =
 				(void *)newinfo->entries
 				+ TABLE_OFFSET(newinfo, i);
@@ -990,7 +993,7 @@ get_counters(const struct ipt_table_info *t,
 	unsigned int cpu;
 	unsigned int i;
 
-	for (cpu = 0; cpu < num_possible_cpus(); cpu++) {
+	for_each_cpu(cpu) {
 		i = 0;
 		IPT_ENTRY_ITERATE(t->entries + TABLE_OFFSET(t, cpu),
 				  t->size,
@@ -1128,7 +1131,8 @@ do_replace(void __user *user, unsigned int len)
 		return -ENOMEM;
 
 	newinfo = vmalloc(sizeof(struct ipt_table_info)
-			  + SMP_ALIGN(tmp.size) * num_possible_cpus());
+			  + SMP_ALIGN(tmp.size) * 
+			  	(highest_possible_processor_id()+1));
 	if (!newinfo)
 		return -ENOMEM;
 
@@ -1458,7 +1462,8 @@ int ipt_register_table(struct ipt_table *table, const struct ipt_replace *repl)
 		= { 0, 0, 0, { 0 }, { 0 }, { } };
 
 	newinfo = vmalloc(sizeof(struct ipt_table_info)
-			  + SMP_ALIGN(repl->size) * num_possible_cpus());
+			  + SMP_ALIGN(repl->size) * 
+			  		(highest_possible_processor_id()+1));
 	if (!newinfo)
 		return -ENOMEM;
 
diff --git a/net/ipv4/netfilter/ipt_REDIRECT.c b/net/ipv4/netfilter/ipt_REDIRECT.c
index 715cb613405..5245bfd33d5 100644
--- a/net/ipv4/netfilter/ipt_REDIRECT.c
+++ b/net/ipv4/netfilter/ipt_REDIRECT.c
@@ -93,7 +93,7 @@ redirect_target(struct sk_buff **pskb,
 		newdst = 0;
 		
 		rcu_read_lock();
-		indev = __in_dev_get((*pskb)->dev);
+		indev = __in_dev_get_rcu((*pskb)->dev);
 		if (indev && (ifa = indev->ifa_list))
 			newdst = ifa->ifa_local;
 		rcu_read_unlock();
diff --git a/net/ipv4/netfilter/ipt_ULOG.c b/net/ipv4/netfilter/ipt_ULOG.c
index e2c14f3cb2f..2883ccd8a91 100644
--- a/net/ipv4/netfilter/ipt_ULOG.c
+++ b/net/ipv4/netfilter/ipt_ULOG.c
@@ -225,8 +225,8 @@ static void ipt_ulog_packet(unsigned int hooknum,
 
 	/* copy hook, prefix, timestamp, payload, etc. */
 	pm->data_len = copy_len;
-	pm->timestamp_sec = skb_tv_base.tv_sec + skb->tstamp.off_sec;
-	pm->timestamp_usec = skb_tv_base.tv_usec + skb->tstamp.off_usec;
+	pm->timestamp_sec = skb->tstamp.off_sec;
+	pm->timestamp_usec = skb->tstamp.off_usec;
 	pm->mark = skb->nfmark;
 	pm->hook = hooknum;
 	if (prefix != NULL)
diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c
index f7943ba1f43..a65e508fbd4 100644
--- a/net/ipv4/proc.c
+++ b/net/ipv4/proc.c
@@ -90,9 +90,7 @@ fold_field(void *mib[], int offt)
 	unsigned long res = 0;
 	int i;
 
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!cpu_possible(i))
-			continue;
+	for_each_cpu(i) {
 		res += *(((unsigned long *) per_cpu_ptr(mib[0], i)) + offt);
 		res += *(((unsigned long *) per_cpu_ptr(mib[1], i)) + offt);
 	}
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 8549f26e249..381dd6a6aeb 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -2128,7 +2128,7 @@ int ip_route_input(struct sk_buff *skb, u32 daddr, u32 saddr,
 		struct in_device *in_dev;
 
 		rcu_read_lock();
-		if ((in_dev = __in_dev_get(dev)) != NULL) {
+		if ((in_dev = __in_dev_get_rcu(dev)) != NULL) {
 			int our = ip_check_mc(in_dev, daddr, saddr,
 				skb->nh.iph->protocol);
 			if (our
@@ -2443,7 +2443,9 @@ static int ip_route_output_slow(struct rtable **rp, const struct flowi *oldflp)
 		err = -ENODEV;
 		if (dev_out == NULL)
 			goto out;
-		if (__in_dev_get(dev_out) == NULL) {
+
+		/* RACE: Check return value of inet_select_addr instead. */
+		if (__in_dev_get_rtnl(dev_out) == NULL) {
 			dev_put(dev_out);
 			goto out;	/* Wrong error code */
 		}
diff --git a/net/ipv4/tcp_bic.c b/net/ipv4/tcp_bic.c
index b940346de4e..6d80e063c18 100644
--- a/net/ipv4/tcp_bic.c
+++ b/net/ipv4/tcp_bic.c
@@ -136,7 +136,7 @@ static inline void bictcp_update(struct bictcp *ca, u32 cwnd)
 		else if (cwnd < ca->last_max_cwnd + max_increment*(BICTCP_B-1))
 			/* slow start */
 			ca->cnt = (cwnd * (BICTCP_B-1))
-				/ cwnd-ca->last_max_cwnd;
+				/ (cwnd - ca->last_max_cwnd);
 		else
 			/* linear increase */
 			ca->cnt = cwnd / max_increment;
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index a7537c7bbd0..3e98b57578d 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -355,8 +355,6 @@ static void tcp_clamp_window(struct sock *sk, struct tcp_sock *tp)
 			app_win -= icsk->icsk_ack.rcv_mss;
 		app_win = max(app_win, 2U*tp->advmss);
 
-		if (!ofo_win)
-			tp->window_clamp = min(tp->window_clamp, app_win);
 		tp->rcv_ssthresh = min(tp->window_clamp, 2U*tp->advmss);
 	}
 }
@@ -2241,6 +2239,7 @@ static int tcp_ack_update_window(struct sock *sk, struct tcp_sock *tp,
 			/* Note, it is the only place, where
 			 * fast path is recovered for sending TCP.
 			 */
+			tp->pred_flags = 0;
 			tcp_fast_path_check(sk, tp);
 
 			if (nwin > tp->max_window) {
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 13dfb391cdf..c85819d8474 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -130,19 +130,20 @@ static int __tcp_v4_check_established(struct sock *sk, __u16 lport,
 	int dif = sk->sk_bound_dev_if;
 	INET_ADDR_COOKIE(acookie, saddr, daddr)
 	const __u32 ports = INET_COMBINED_PORTS(inet->dport, lport);
-	const int hash = inet_ehashfn(daddr, lport, saddr, inet->dport, tcp_hashinfo.ehash_size);
-	struct inet_ehash_bucket *head = &tcp_hashinfo.ehash[hash];
+	unsigned int hash = inet_ehashfn(daddr, lport, saddr, inet->dport);
+	struct inet_ehash_bucket *head = inet_ehash_bucket(&tcp_hashinfo, hash);
 	struct sock *sk2;
 	const struct hlist_node *node;
 	struct inet_timewait_sock *tw;
 
+	prefetch(head->chain.first);
 	write_lock(&head->lock);
 
 	/* Check TIME-WAIT sockets first. */
 	sk_for_each(sk2, node, &(head + tcp_hashinfo.ehash_size)->chain) {
 		tw = inet_twsk(sk2);
 
-		if (INET_TW_MATCH(sk2, acookie, saddr, daddr, ports, dif)) {
+		if (INET_TW_MATCH(sk2, hash, acookie, saddr, daddr, ports, dif)) {
 			const struct tcp_timewait_sock *tcptw = tcp_twsk(sk2);
 			struct tcp_sock *tp = tcp_sk(sk);
 
@@ -179,7 +180,7 @@ static int __tcp_v4_check_established(struct sock *sk, __u16 lport,
 
 	/* And established part... */
 	sk_for_each(sk2, node, &head->chain) {
-		if (INET_MATCH(sk2, acookie, saddr, daddr, ports, dif))
+		if (INET_MATCH(sk2, hash, acookie, saddr, daddr, ports, dif))
 			goto not_unique;
 	}
 
@@ -188,7 +189,7 @@ unique:
 	 * in hash table socket with a funny identity. */
 	inet->num = lport;
 	inet->sport = htons(lport);
-	sk->sk_hashent = hash;
+	sk->sk_hash = hash;
 	BUG_TRAP(sk_unhashed(sk));
 	__sk_add_node(sk, &head->chain);
 	sock_prot_inc_use(sk->sk_prot);
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 5dd6dd7d091..b907456a79f 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -190,7 +190,7 @@ void tcp_select_initial_window(int __space, __u32 mss,
 	}
 
 	/* Set initial window to value enough for senders,
-	 * following RFC1414. Senders, not following this RFC,
+	 * following RFC2414. Senders, not following this RFC,
 	 * will be satisfied with 2.
 	 */
 	if (mss > (1<<*rcv_wscale)) {
@@ -435,8 +435,7 @@ int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len, unsigned int mss
 	int nsize, old_factor;
 	u16 flags;
 
-	BUG_ON(len >= skb->len);
-
+	BUG_ON(len > skb->len);
 	nsize = skb_headlen(skb) - len;
 	if (nsize < 0)
 		nsize = 0;
@@ -509,7 +508,16 @@ int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len, unsigned int mss
 			tp->lost_out -= diff;
 			tp->left_out -= diff;
 		}
+
 		if (diff > 0) {
+			/* Adjust Reno SACK estimate. */
+			if (!tp->rx_opt.sack_ok) {
+				tp->sacked_out -= diff;
+				if ((int)tp->sacked_out < 0)
+					tp->sacked_out = 0;
+				tcp_sync_left_out(tp);
+			}
+
 			tp->fackets_out -= diff;
 			if ((int)tp->fackets_out < 0)
 				tp->fackets_out = 0;
@@ -1601,7 +1609,7 @@ void tcp_send_fin(struct sock *sk)
  * was unread data in the receive queue.  This behavior is recommended
  * by draft-ietf-tcpimpl-prob-03.txt section 3.10.  -DaveM
  */
-void tcp_send_active_reset(struct sock *sk, unsigned int __nocast priority)
+void tcp_send_active_reset(struct sock *sk, gfp_t priority)
 {
 	struct tcp_sock *tp = tcp_sk(sk);
 	struct sk_buff *skb;
diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
index 2fea3f4402a..a970b4727ce 100644
--- a/net/ipv6/addrconf.c
+++ b/net/ipv6/addrconf.c
@@ -1806,7 +1806,7 @@ static void sit_add_v4_addrs(struct inet6_dev *idev)
 	}
 
         for (dev = dev_base; dev != NULL; dev = dev->next) {
-		struct in_device * in_dev = __in_dev_get(dev);
+		struct in_device * in_dev = __in_dev_get_rtnl(dev);
 		if (in_dev && (dev->flags & IFF_UP)) {
 			struct in_ifaddr * ifa;
 
@@ -3520,6 +3520,8 @@ int __init addrconf_init(void)
 	if (err)
 		return err;
 
+	ip6_null_entry.rt6i_idev = in6_dev_get(&loopback_dev);
+
 	register_netdevice_notifier(&ipv6_dev_notf);
 
 #ifdef CONFIG_IPV6_PRIVACY
diff --git a/net/ipv6/esp6.c b/net/ipv6/esp6.c
index 9b27460f0cc..40d9a1935ab 100644
--- a/net/ipv6/esp6.c
+++ b/net/ipv6/esp6.c
@@ -31,6 +31,7 @@
 #include <net/esp.h>
 #include <asm/scatterlist.h>
 #include <linux/crypto.h>
+#include <linux/kernel.h>
 #include <linux/pfkeyv2.h>
 #include <linux/random.h>
 #include <net/icmp.h>
@@ -66,10 +67,10 @@ static int esp6_output(struct xfrm_state *x, struct sk_buff *skb)
 
 	alen = esp->auth.icv_trunc_len;
 	tfm = esp->conf.tfm;
-	blksize = (crypto_tfm_alg_blocksize(tfm) + 3) & ~3;
-	clen = (clen + 2 + blksize-1)&~(blksize-1);
+	blksize = ALIGN(crypto_tfm_alg_blocksize(tfm), 4);
+	clen = ALIGN(clen + 2, blksize);
 	if (esp->conf.padlen)
-		clen = (clen + esp->conf.padlen-1)&~(esp->conf.padlen-1);
+		clen = ALIGN(clen, esp->conf.padlen);
 
 	if ((nfrags = skb_cow_data(skb, clen-skb->len+alen, &trailer)) < 0) {
 		goto error;
@@ -133,7 +134,7 @@ static int esp6_input(struct xfrm_state *x, struct xfrm_decap_state *decap, stru
 	struct ipv6_esp_hdr *esph;
 	struct esp_data *esp = x->data;
 	struct sk_buff *trailer;
-	int blksize = crypto_tfm_alg_blocksize(esp->conf.tfm);
+	int blksize = ALIGN(crypto_tfm_alg_blocksize(esp->conf.tfm), 4);
 	int alen = esp->auth.icv_trunc_len;
 	int elen = skb->len - sizeof(struct ipv6_esp_hdr) - esp->conf.ivlen - alen;
 
@@ -235,16 +236,17 @@ out_nofree:
 static u32 esp6_get_max_size(struct xfrm_state *x, int mtu)
 {
 	struct esp_data *esp = x->data;
-	u32 blksize = crypto_tfm_alg_blocksize(esp->conf.tfm);
+	u32 blksize = ALIGN(crypto_tfm_alg_blocksize(esp->conf.tfm), 4);
 
 	if (x->props.mode) {
-		mtu = (mtu + 2 + blksize-1)&~(blksize-1);
+		mtu = ALIGN(mtu + 2, blksize);
 	} else {
 		/* The worst case. */
-		mtu += 2 + blksize;
+		u32 padsize = ((blksize - 1) & 7) + 1;
+		mtu = ALIGN(mtu + 2, padsize) + blksize - padsize;
 	}
 	if (esp->conf.padlen)
-		mtu = (mtu + esp->conf.padlen-1)&~(esp->conf.padlen-1);
+		mtu = ALIGN(mtu, esp->conf.padlen);
 
 	return mtu + x->props.header_len + esp->auth.icv_full_len;
 }
diff --git a/net/ipv6/icmp.c b/net/ipv6/icmp.c
index b7185fb3377..23e540365a1 100644
--- a/net/ipv6/icmp.c
+++ b/net/ipv6/icmp.c
@@ -700,10 +700,7 @@ int __init icmpv6_init(struct net_proto_family *ops)
 	struct sock *sk;
 	int err, i, j;
 
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!cpu_possible(i))
-			continue;
-
+	for_each_cpu(i) {
 		err = sock_create_kern(PF_INET6, SOCK_RAW, IPPROTO_ICMPV6,
 				       &per_cpu(__icmpv6_socket, i));
 		if (err < 0) {
@@ -749,9 +746,7 @@ void icmpv6_cleanup(void)
 {
 	int i;
 
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!cpu_possible(i))
-			continue;
+	for_each_cpu(i) {
 		sock_release(per_cpu(__icmpv6_socket, i));
 	}
 	inet6_del_protocol(&icmpv6_protocol, IPPROTO_ICMPV6);
diff --git a/net/ipv6/ip6_flowlabel.c b/net/ipv6/ip6_flowlabel.c
index f841bde30c1..bbbe80cdaf7 100644
--- a/net/ipv6/ip6_flowlabel.c
+++ b/net/ipv6/ip6_flowlabel.c
@@ -483,7 +483,7 @@ int ipv6_flowlabel_opt(struct sock *sk, char __user *optval, int optlen)
 						goto done;
 					}
 					fl1 = sfl->fl;
-					atomic_inc(&fl->users);
+					atomic_inc(&fl1->users);
 					break;
 				}
 			}
diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
index 2f589f24c09..563b442ffab 100644
--- a/net/ipv6/ip6_output.c
+++ b/net/ipv6/ip6_output.c
@@ -666,7 +666,7 @@ slow_path:
 		 */
 		fh->nexthdr = nexthdr;
 		fh->reserved = 0;
-		if (frag_id) {
+		if (!frag_id) {
 			ipv6_select_ident(skb, fh);
 			frag_id = fh->identification;
 		} else
diff --git a/net/ipv6/mcast.c b/net/ipv6/mcast.c
index 519899fb11d..39a96c76810 100644
--- a/net/ipv6/mcast.c
+++ b/net/ipv6/mcast.c
@@ -1393,7 +1393,7 @@ static void mld_sendpack(struct sk_buff *skb)
 
 static int grec_size(struct ifmcaddr6 *pmc, int type, int gdel, int sdel)
 {
-	return sizeof(struct mld2_grec) + 4*mld_scount(pmc,type,gdel,sdel);
+	return sizeof(struct mld2_grec) + 16 * mld_scount(pmc,type,gdel,sdel);
 }
 
 static struct sk_buff *add_grhead(struct sk_buff *skb, struct ifmcaddr6 *pmc,
diff --git a/net/ipv6/ndisc.c b/net/ipv6/ndisc.c
index 555a31347ed..305d9ee6d7d 100644
--- a/net/ipv6/ndisc.c
+++ b/net/ipv6/ndisc.c
@@ -1450,7 +1450,7 @@ void ndisc_send_redirect(struct sk_buff *skb, struct neighbour *neigh,
 
 static void pndisc_redo(struct sk_buff *skb)
 {
-	ndisc_rcv(skb);
+	ndisc_recv_ns(skb);
 	kfree_skb(skb);
 }
 
diff --git a/net/ipv6/netfilter/Kconfig b/net/ipv6/netfilter/Kconfig
index 216fbe1ac65..bb7ccfe33f2 100644
--- a/net/ipv6/netfilter/Kconfig
+++ b/net/ipv6/netfilter/Kconfig
@@ -209,6 +209,17 @@ config IP6_NF_TARGET_REJECT
 
 	  To compile it as a module, choose M here.  If unsure, say N.
 
+config IP6_NF_TARGET_NFQUEUE
+	tristate "NFQUEUE Target Support"
+	depends on IP_NF_IPTABLES
+	help
+	  This Target replaced the old obsolete QUEUE target.
+
+	  As opposed to QUEUE, it supports 65535 different queues,
+	  not just one.
+
+	  To compile it as a module, choose M here.  If unsure, say N.
+
 #  if [ "$CONFIG_IP6_NF_FILTER" != "n" ]; then
 #    dep_tristate '    REJECT target support' CONFIG_IP6_NF_TARGET_REJECT $CONFIG_IP6_NF_FILTER
 #    if [ "$CONFIG_EXPERIMENTAL" = "y" ]; then
diff --git a/net/ipv6/netfilter/Makefile b/net/ipv6/netfilter/Makefile
index bd9a16a5cbb..2b2c370e8b1 100644
--- a/net/ipv6/netfilter/Makefile
+++ b/net/ipv6/netfilter/Makefile
@@ -21,9 +21,9 @@ obj-$(CONFIG_IP6_NF_FILTER) += ip6table_filter.o
 obj-$(CONFIG_IP6_NF_MANGLE) += ip6table_mangle.o
 obj-$(CONFIG_IP6_NF_TARGET_MARK) += ip6t_MARK.o
 obj-$(CONFIG_IP6_NF_TARGET_HL) += ip6t_HL.o
+obj-$(CONFIG_IP6_NF_TARGET_NFQUEUE) += ip6t_NFQUEUE.o
 obj-$(CONFIG_IP6_NF_QUEUE) += ip6_queue.o
 obj-$(CONFIG_IP6_NF_TARGET_LOG) += ip6t_LOG.o
 obj-$(CONFIG_IP6_NF_RAW) += ip6table_raw.o
 obj-$(CONFIG_IP6_NF_MATCH_HL) += ip6t_hl.o
 obj-$(CONFIG_IP6_NF_TARGET_REJECT) += ip6t_REJECT.o
-obj-$(CONFIG_NETFILTER_NETLINK_QUEUE) += ip6t_NFQUEUE.o
diff --git a/net/ipv6/netfilter/ip6_queue.c b/net/ipv6/netfilter/ip6_queue.c
index aa11cf366ef..5027bbe6415 100644
--- a/net/ipv6/netfilter/ip6_queue.c
+++ b/net/ipv6/netfilter/ip6_queue.c
@@ -238,8 +238,8 @@ ipq_build_packet_message(struct ipq_queue_entry *entry, int *errp)
 
 	pmsg->packet_id       = (unsigned long )entry;
 	pmsg->data_len        = data_len;
-	pmsg->timestamp_sec   = skb_tv_base.tv_sec + entry->skb->tstamp.off_sec;
-	pmsg->timestamp_usec  = skb_tv_base.tv_usec + entry->skb->tstamp.off_usec;
+	pmsg->timestamp_sec   = entry->skb->tstamp.off_sec;
+	pmsg->timestamp_usec  = entry->skb->tstamp.off_usec;
 	pmsg->mark            = entry->skb->nfmark;
 	pmsg->hook            = entry->info->hook;
 	pmsg->hw_protocol     = entry->skb->protocol;
diff --git a/net/ipv6/netfilter/ip6_tables.c b/net/ipv6/netfilter/ip6_tables.c
index 2da514b16d9..21deec25a12 100644
--- a/net/ipv6/netfilter/ip6_tables.c
+++ b/net/ipv6/netfilter/ip6_tables.c
@@ -28,6 +28,7 @@
 #include <asm/uaccess.h>
 #include <asm/semaphore.h>
 #include <linux/proc_fs.h>
+#include <linux/cpumask.h>
 
 #include <linux/netfilter_ipv6/ip6_tables.h>
 
@@ -950,8 +951,10 @@ translate_table(const char *name,
 	}
 
 	/* And one copy for every other CPU */
-	for (i = 1; i < num_possible_cpus(); i++) {
-		memcpy(newinfo->entries + SMP_ALIGN(newinfo->size)*i,
+	for_each_cpu(i) {
+		if (i == 0)
+			continue;
+		memcpy(newinfo->entries + SMP_ALIGN(newinfo->size) * i,
 		       newinfo->entries,
 		       SMP_ALIGN(newinfo->size));
 	}
@@ -972,7 +975,7 @@ replace_table(struct ip6t_table *table,
 		struct ip6t_entry *table_base;
 		unsigned int i;
 
-		for (i = 0; i < num_possible_cpus(); i++) {
+		for_each_cpu(i) {
 			table_base =
 				(void *)newinfo->entries
 				+ TABLE_OFFSET(newinfo, i);
@@ -1019,7 +1022,7 @@ get_counters(const struct ip6t_table_info *t,
 	unsigned int cpu;
 	unsigned int i;
 
-	for (cpu = 0; cpu < num_possible_cpus(); cpu++) {
+	for_each_cpu(cpu) {
 		i = 0;
 		IP6T_ENTRY_ITERATE(t->entries + TABLE_OFFSET(t, cpu),
 				  t->size,
@@ -1153,7 +1156,8 @@ do_replace(void __user *user, unsigned int len)
 		return -ENOMEM;
 
 	newinfo = vmalloc(sizeof(struct ip6t_table_info)
-			  + SMP_ALIGN(tmp.size) * num_possible_cpus());
+			  + SMP_ALIGN(tmp.size) *
+			  		(highest_possible_processor_id()+1));
 	if (!newinfo)
 		return -ENOMEM;
 
@@ -1467,7 +1471,8 @@ int ip6t_register_table(struct ip6t_table *table,
 		= { 0, 0, 0, { 0 }, { 0 }, { } };
 
 	newinfo = vmalloc(sizeof(struct ip6t_table_info)
-			  + SMP_ALIGN(repl->size) * num_possible_cpus());
+			  + SMP_ALIGN(repl->size) *
+			  		(highest_possible_processor_id()+1));
 	if (!newinfo)
 		return -ENOMEM;
 
diff --git a/net/ipv6/proc.c b/net/ipv6/proc.c
index 334a5967831..50a13e75d70 100644
--- a/net/ipv6/proc.c
+++ b/net/ipv6/proc.c
@@ -140,9 +140,7 @@ fold_field(void *mib[], int offt)
         unsigned long res = 0;
         int i;
  
-        for (i = 0; i < NR_CPUS; i++) {
-                if (!cpu_possible(i))
-                        continue;
+        for_each_cpu(i) {
                 res += *(((unsigned long *)per_cpu_ptr(mib[0], i)) + offt);
                 res += *(((unsigned long *)per_cpu_ptr(mib[1], i)) + offt);
         }
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 80643e6b346..d693cb988b7 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -209,9 +209,11 @@ static __inline__ void __tcp_v6_hash(struct sock *sk)
 		lock = &tcp_hashinfo.lhash_lock;
 		inet_listen_wlock(&tcp_hashinfo);
 	} else {
-		sk->sk_hashent = inet6_sk_ehashfn(sk, tcp_hashinfo.ehash_size);
-		list = &tcp_hashinfo.ehash[sk->sk_hashent].chain;
-		lock = &tcp_hashinfo.ehash[sk->sk_hashent].lock;
+		unsigned int hash;
+		sk->sk_hash = hash = inet6_sk_ehashfn(sk);
+		hash &= (tcp_hashinfo.ehash_size - 1);
+		list = &tcp_hashinfo.ehash[hash].chain;
+		lock = &tcp_hashinfo.ehash[hash].lock;
 		write_lock(lock);
 	}
 
@@ -322,13 +324,13 @@ static int __tcp_v6_check_established(struct sock *sk, const __u16 lport,
 	const struct in6_addr *saddr = &np->daddr;
 	const int dif = sk->sk_bound_dev_if;
 	const u32 ports = INET_COMBINED_PORTS(inet->dport, lport);
-	const int hash = inet6_ehashfn(daddr, inet->num, saddr, inet->dport,
-				       tcp_hashinfo.ehash_size);
-	struct inet_ehash_bucket *head = &tcp_hashinfo.ehash[hash];
+	unsigned int hash = inet6_ehashfn(daddr, inet->num, saddr, inet->dport);
+	struct inet_ehash_bucket *head = inet_ehash_bucket(&tcp_hashinfo, hash);
 	struct sock *sk2;
 	const struct hlist_node *node;
 	struct inet_timewait_sock *tw;
 
+	prefetch(head->chain.first);
 	write_lock(&head->lock);
 
 	/* Check TIME-WAIT sockets first. */
@@ -365,14 +367,14 @@ static int __tcp_v6_check_established(struct sock *sk, const __u16 lport,
 
 	/* And established part... */
 	sk_for_each(sk2, node, &head->chain) {
-		if (INET6_MATCH(sk2, saddr, daddr, ports, dif))
+		if (INET6_MATCH(sk2, hash, saddr, daddr, ports, dif))
 			goto not_unique;
 	}
 
 unique:
 	BUG_TRAP(sk_unhashed(sk));
 	__sk_add_node(sk, &head->chain);
-	sk->sk_hashent = hash;
+	sk->sk_hash = hash;
 	sock_prot_inc_use(sk->sk_prot);
 	write_unlock(&head->lock);
 
diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c
index 6001948600f..bf9519341fd 100644
--- a/net/ipv6/udp.c
+++ b/net/ipv6/udp.c
@@ -99,7 +99,7 @@ static int udp_v6_get_port(struct sock *sk, unsigned short snum)
 		next:;
 		}
 		result = best;
-		for(;; result += UDP_HTABLE_SIZE) {
+		for(i = 0; i < (1 << 16) / UDP_HTABLE_SIZE; i++, result += UDP_HTABLE_SIZE) {
 			if (result > sysctl_local_port_range[1])
 				result = sysctl_local_port_range[0]
 					+ ((result - sysctl_local_port_range[0]) &
@@ -107,6 +107,8 @@ static int udp_v6_get_port(struct sock *sk, unsigned short snum)
 			if (!udp_lport_inuse(result))
 				break;
 		}
+		if (i >= (1 << 16) / UDP_HTABLE_SIZE)
+			goto fail;
 gotit:
 		udp_port_rover = snum = result;
 	} else {
@@ -852,10 +854,16 @@ do_append_data:
 	else if (!corkreq)
 		err = udp_v6_push_pending_frames(sk, up);
 
-	if (dst && connected)
-		ip6_dst_store(sk, dst,
-			      ipv6_addr_equal(&fl->fl6_dst, &np->daddr) ?
-			      &np->daddr : NULL);
+	if (dst) {
+		if (connected) {
+			ip6_dst_store(sk, dst,
+				      ipv6_addr_equal(&fl->fl6_dst, &np->daddr) ?
+				      &np->daddr : NULL);
+		} else {
+			dst_release(dst);
+		}
+	}
+
 	if (err > 0)
 		err = np->recverr ? net_xmit_errno(err) : 0;
 	release_sock(sk);
diff --git a/net/irda/irlan/irlan_eth.c b/net/irda/irlan/irlan_eth.c
index 071cd2cefd8..953e255d2bc 100644
--- a/net/irda/irlan/irlan_eth.c
+++ b/net/irda/irlan/irlan_eth.c
@@ -310,7 +310,7 @@ void irlan_eth_send_gratuitous_arp(struct net_device *dev)
 #ifdef CONFIG_INET
 	IRDA_DEBUG(4, "IrLAN: Sending gratuitous ARP\n");
 	rcu_read_lock();
-	in_dev = __in_dev_get(dev);
+	in_dev = __in_dev_get_rcu(dev);
 	if (in_dev == NULL)
 		goto out;
 	if (in_dev->ifa_list)
diff --git a/net/irda/irttp.c b/net/irda/irttp.c
index 6602d901f8b..8aff254cb41 100644
--- a/net/irda/irttp.c
+++ b/net/irda/irttp.c
@@ -38,7 +38,7 @@
 #include <net/irda/parameters.h>
 #include <net/irda/irttp.h>
 
-static struct irttp_cb *irttp = NULL;
+static struct irttp_cb *irttp;
 
 static void __irttp_close_tsap(struct tsap_cb *self);
 
@@ -86,12 +86,9 @@ static pi_param_info_t param_info = { pi_major_call_table, 1, 0x0f, 4 };
  */
 int __init irttp_init(void)
 {
-	/* Initialize the irttp structure. */
-	if (irttp == NULL) {
-		irttp = kmalloc(sizeof(struct irttp_cb), GFP_KERNEL);
-		if (irttp == NULL)
-			return -ENOMEM;
-	}
+	irttp = kmalloc(sizeof(struct irttp_cb), GFP_KERNEL);
+	if (irttp == NULL)
+		return -ENOMEM;
 	memset(irttp, 0, sizeof(struct irttp_cb));
 
 	irttp->magic = TTP_MAGIC;
@@ -100,6 +97,7 @@ int __init irttp_init(void)
 	if (!irttp->tsaps) {
 		IRDA_ERROR("%s: can't allocate IrTTP hashbin!\n",
 			   __FUNCTION__);
+		kfree(irttp);
 		return -ENOMEM;
 	}
 
@@ -115,7 +113,6 @@ int __init irttp_init(void)
 void __exit irttp_cleanup(void) 
 {
 	/* Check for main structure */
-	IRDA_ASSERT(irttp != NULL, return;);
 	IRDA_ASSERT(irttp->magic == TTP_MAGIC, return;);
 
 	/*
@@ -382,7 +379,6 @@ struct tsap_cb *irttp_open_tsap(__u8 stsap_sel, int credit, notify_t *notify)
 	struct lsap_cb *lsap;
 	notify_t ttp_notify;
 
-	IRDA_ASSERT(irttp != NULL, return NULL;);
 	IRDA_ASSERT(irttp->magic == TTP_MAGIC, return NULL;);
 
 	/* The IrLMP spec (IrLMP 1.1 p10) says that we have the right to
@@ -1880,8 +1876,6 @@ static int irttp_seq_open(struct inode *inode, struct file *file)
 	struct seq_file *seq;
 	int rc = -ENOMEM;
 	struct irttp_iter_state *s;
-       
-	IRDA_ASSERT(irttp != NULL, return -EINVAL;);
 
 	s = kmalloc(sizeof(*s), GFP_KERNEL);
 	if (!s)
diff --git a/net/key/af_key.c b/net/key/af_key.c
index 4879743b945..39031684b65 100644
--- a/net/key/af_key.c
+++ b/net/key/af_key.c
@@ -185,7 +185,7 @@ static int pfkey_release(struct socket *sock)
 }
 
 static int pfkey_broadcast_one(struct sk_buff *skb, struct sk_buff **skb2,
-			       int allocation, struct sock *sk)
+			       gfp_t allocation, struct sock *sk)
 {
 	int err = -ENOBUFS;
 
@@ -217,7 +217,7 @@ static int pfkey_broadcast_one(struct sk_buff *skb, struct sk_buff **skb2,
 #define BROADCAST_ONE		1
 #define BROADCAST_REGISTERED	2
 #define BROADCAST_PROMISC_ONLY	4
-static int pfkey_broadcast(struct sk_buff *skb, int allocation,
+static int pfkey_broadcast(struct sk_buff *skb, gfp_t allocation,
 			   int broadcast_flags, struct sock *one_sk)
 {
 	struct sock *sk;
@@ -1416,7 +1416,8 @@ static int pfkey_get(struct sock *sk, struct sk_buff *skb, struct sadb_msg *hdr,
 	return 0;
 }
 
-static struct sk_buff *compose_sadb_supported(struct sadb_msg *orig, int allocation)
+static struct sk_buff *compose_sadb_supported(struct sadb_msg *orig,
+					      gfp_t allocation)
 {
 	struct sk_buff *skb;
 	struct sadb_msg *hdr;
@@ -2153,6 +2154,7 @@ out:
 
 static int pfkey_spdget(struct sock *sk, struct sk_buff *skb, struct sadb_msg *hdr, void **ext_hdrs)
 {
+	unsigned int dir;
 	int err;
 	struct sadb_x_policy *pol;
 	struct xfrm_policy *xp;
@@ -2161,7 +2163,11 @@ static int pfkey_spdget(struct sock *sk, struct sk_buff *skb, struct sadb_msg *h
 	if ((pol = ext_hdrs[SADB_X_EXT_POLICY-1]) == NULL)
 		return -EINVAL;
 
-	xp = xfrm_policy_byid(0, pol->sadb_x_policy_id,
+	dir = xfrm_policy_id2dir(pol->sadb_x_policy_id);
+	if (dir >= XFRM_POLICY_MAX)
+		return -EINVAL;
+
+	xp = xfrm_policy_byid(dir, pol->sadb_x_policy_id,
 			      hdr->sadb_msg_type == SADB_X_SPDDELETE2);
 	if (xp == NULL)
 		return -ENOENT;
@@ -2173,9 +2179,9 @@ static int pfkey_spdget(struct sock *sk, struct sk_buff *skb, struct sadb_msg *h
 	if (hdr->sadb_msg_type == SADB_X_SPDDELETE2) {
 		c.data.byid = 1;
 		c.event = XFRM_MSG_DELPOLICY;
-		km_policy_notify(xp, pol->sadb_x_policy_dir-1, &c);
+		km_policy_notify(xp, dir, &c);
 	} else {
-		err = key_pol_get_resp(sk, xp, hdr, pol->sadb_x_policy_dir-1);
+		err = key_pol_get_resp(sk, xp, hdr, dir);
 	}
 
 	xfrm_pol_put(xp);
diff --git a/net/llc/Makefile b/net/llc/Makefile
index 5ebd4ed2bd4..4e260cff3c5 100644
--- a/net/llc/Makefile
+++ b/net/llc/Makefile
@@ -22,3 +22,4 @@ llc2-y := llc_if.o llc_c_ev.o llc_c_ac.o llc_conn.o llc_c_st.o llc_pdu.o \
 	  llc_sap.o llc_s_ac.o llc_s_ev.o llc_s_st.o af_llc.o llc_station.o
 
 llc2-$(CONFIG_PROC_FS) += llc_proc.o
+llc2-$(CONFIG_SYSCTL)  += sysctl_net_llc.o
diff --git a/net/llc/af_llc.c b/net/llc/af_llc.c
index 66f55e514b5..59d02cbbeb9 100644
--- a/net/llc/af_llc.c
+++ b/net/llc/af_llc.c
@@ -21,6 +21,7 @@
  * See the GNU General Public License for more details.
  */
 #include <linux/config.h>
+#include <linux/compiler.h>
 #include <linux/kernel.h>
 #include <linux/module.h>
 #include <linux/rtnetlink.h>
@@ -37,10 +38,9 @@ static u16 llc_ui_sap_link_no_max[256];
 static struct sockaddr_llc llc_ui_addrnull;
 static struct proto_ops llc_ui_ops;
 
-static int llc_ui_wait_for_conn(struct sock *sk, int timeout);
-static int llc_ui_wait_for_disc(struct sock *sk, int timeout);
-static int llc_ui_wait_for_data(struct sock *sk, int timeout);
-static int llc_ui_wait_for_busy_core(struct sock *sk, int timeout);
+static int llc_ui_wait_for_conn(struct sock *sk, long timeout);
+static int llc_ui_wait_for_disc(struct sock *sk, long timeout);
+static int llc_ui_wait_for_busy_core(struct sock *sk, long timeout);
 
 #if 0
 #define dprintk(args...) printk(KERN_DEBUG args)
@@ -116,12 +116,12 @@ static int llc_ui_send_data(struct sock* sk, struct sk_buff *skb, int noblock)
 	struct llc_sock* llc = llc_sk(sk);
 	int rc = 0;
 
-	if (llc_data_accept_state(llc->state) || llc->p_flag) {
-		int timeout = sock_sndtimeo(sk, noblock);
+	if (unlikely(llc_data_accept_state(llc->state) || llc->p_flag)) {
+		long timeout = sock_sndtimeo(sk, noblock);
 
 		rc = llc_ui_wait_for_busy_core(sk, timeout);
 	}
-	if (!rc)
+	if (unlikely(!rc))
 		rc = llc_build_and_send_pkt(sk, skb);
 	return rc;
 }
@@ -155,7 +155,7 @@ static int llc_ui_create(struct socket *sock, int protocol)
 	struct sock *sk;
 	int rc = -ESOCKTNOSUPPORT;
 
-	if (sock->type == SOCK_DGRAM || sock->type == SOCK_STREAM) {
+	if (likely(sock->type == SOCK_DGRAM || sock->type == SOCK_STREAM)) {
 		rc = -ENOMEM;
 		sk = llc_sk_alloc(PF_LLC, GFP_KERNEL, &llc_proto);
 		if (sk) {
@@ -177,7 +177,7 @@ static int llc_ui_release(struct socket *sock)
 	struct sock *sk = sock->sk;
 	struct llc_sock *llc;
 
-	if (!sk)
+	if (unlikely(sk == NULL))
 		goto out;
 	sock_hold(sk);
 	lock_sock(sk);
@@ -189,10 +189,6 @@ static int llc_ui_release(struct socket *sock)
 	if (!sock_flag(sk, SOCK_ZAPPED))
 		llc_sap_remove_socket(llc->sap, sk);
 	release_sock(sk);
-	if (llc->sap && hlist_empty(&llc->sap->sk_list.list)) {
-		llc_release_sockets(llc->sap);
-		llc_sap_close(llc->sap);
-	}
 	if (llc->dev)
 		dev_put(llc->dev);
 	sock_put(sk);
@@ -221,6 +217,7 @@ static int llc_ui_autoport(void)
 				llc_ui_sap_last_autoport = i + 2;
 				goto out;
 			}
+			llc_sap_put(sap);
 		}
 		llc_ui_sap_last_autoport = LLC_SAP_DYN_START;
 		tries++;
@@ -231,20 +228,13 @@ out:
 }
 
 /**
- *	llc_ui_autobind - Bind a socket to a specific address.
- *	@sk: Socket to bind an address to.
- *	@addr: Address the user wants the socket bound to.
+ *	llc_ui_autobind - automatically bind a socket to a sap
+ *	@sock: socket to bind
+ *	@addr: address to connect to
+ *
+ * 	Used by llc_ui_connect and llc_ui_sendmsg when the user hasn't
+ * 	specifically used llc_ui_bind to bind to an specific address/sap
  *
- *	Bind a socket to a specific address. For llc a user is able to bind to
- *	a specific sap only or mac + sap. If the user only specifies a sap and
- *	a null dmac (all zeros) the user is attempting to bind to an entire
- *	sap. This will stop anyone else on the local system from using that
- *	sap.  If someone else has a mac + sap open the bind to null + sap will
- *	fail.
- *	If the user desires to bind to a specific mac + sap, it is possible to
- *	have multiple sap connections via multiple macs.
- *	Bind and autobind for that matter must enforce the correct sap usage
- *	otherwise all hell will break loose.
  *	Returns: 0 upon success, negative otherwise.
  */
 static int llc_ui_autobind(struct socket *sock, struct sockaddr_llc *addr)
@@ -285,11 +275,7 @@ out:
  *	@addrlen: Length of the uaddr structure.
  *
  *	Bind a socket to a specific address. For llc a user is able to bind to
- *	a specific sap only or mac + sap. If the user only specifies a sap and
- *	a null dmac (all zeros) the user is attempting to bind to an entire
- *	sap. This will stop anyone else on the local system from using that
- *	sap. If someone else has a mac + sap open the bind to null + sap will
- *	fail.
+ *	a specific sap only or mac + sap.
  *	If the user desires to bind to a specific mac + sap, it is possible to
  *	have multiple sap connections via multiple macs.
  *	Bind and autobind for that matter must enforce the correct sap usage
@@ -305,10 +291,16 @@ static int llc_ui_bind(struct socket *sock, struct sockaddr *uaddr, int addrlen)
 	int rc = -EINVAL;
 
 	dprintk("%s: binding %02X\n", __FUNCTION__, addr->sllc_sap);
-	if (!sock_flag(sk, SOCK_ZAPPED) || addrlen != sizeof(*addr))
+	if (unlikely(!sock_flag(sk, SOCK_ZAPPED) || addrlen != sizeof(*addr)))
 		goto out;
 	rc = -EAFNOSUPPORT;
-	if (addr->sllc_family != AF_LLC)
+	if (unlikely(addr->sllc_family != AF_LLC))
+		goto out;
+	rc = -ENODEV;
+	rtnl_lock();
+	llc->dev = dev_getbyhwaddr(addr->sllc_arphrd, addr->sllc_mac);
+	rtnl_unlock();
+	if (!llc->dev)
 		goto out;
 	if (!addr->sllc_sap) {
 		rc = -EUSERS;
@@ -322,6 +314,7 @@ static int llc_ui_bind(struct socket *sock, struct sockaddr *uaddr, int addrlen)
 		rc = -EBUSY; /* some other network layer is using the sap */
 		if (!sap)
 			goto out;
+		llc_sap_hold(sap);
 	} else {
 		struct llc_addr laddr, daddr;
 		struct sock *ask;
@@ -338,7 +331,7 @@ static int llc_ui_bind(struct socket *sock, struct sockaddr *uaddr, int addrlen)
 		ask = llc_lookup_established(sap, &daddr, &laddr);
 		if (ask) {
 			sock_put(ask);
-			goto out;
+			goto out_put;
 		}
 	}
 	llc->laddr.lsap = addr->sllc_sap;
@@ -348,6 +341,8 @@ static int llc_ui_bind(struct socket *sock, struct sockaddr *uaddr, int addrlen)
 	llc_sap_add_socket(sap, sk);
 	sock_reset_flag(sk, SOCK_ZAPPED);
 	rc = 0;
+out_put:
+	llc_sap_put(sap);
 out:
 	return rc;
 }
@@ -369,7 +364,7 @@ static int llc_ui_shutdown(struct socket *sock, int how)
 	int rc = -ENOTCONN;
 
 	lock_sock(sk);
-	if (sk->sk_state != TCP_ESTABLISHED)
+	if (unlikely(sk->sk_state != TCP_ESTABLISHED))
 		goto out;
 	rc = -EINVAL;
 	if (how != 2)
@@ -404,14 +399,18 @@ static int llc_ui_connect(struct socket *sock, struct sockaddr *uaddr,
 	struct sock *sk = sock->sk;
 	struct llc_sock *llc = llc_sk(sk);
 	struct sockaddr_llc *addr = (struct sockaddr_llc *)uaddr;
-	struct net_device *dev;
 	int rc = -EINVAL;
 
 	lock_sock(sk);
-	if (addrlen != sizeof(*addr))
+	if (unlikely(addrlen != sizeof(*addr)))
 		goto out;
 	rc = -EAFNOSUPPORT;
-	if (addr->sllc_family != AF_LLC)
+	if (unlikely(addr->sllc_family != AF_LLC))
+		goto out;
+	if (unlikely(sk->sk_type != SOCK_STREAM))
+		goto out;
+	rc = -EALREADY;
+	if (unlikely(sock->state == SS_CONNECTING))
 		goto out;
 	/* bind connection to sap if user hasn't done it. */
 	if (sock_flag(sk, SOCK_ZAPPED)) {
@@ -419,19 +418,13 @@ static int llc_ui_connect(struct socket *sock, struct sockaddr *uaddr,
 		rc = llc_ui_autobind(sock, addr);
 		if (rc)
 			goto out;
-		llc->daddr.lsap = addr->sllc_sap;
-		memcpy(llc->daddr.mac, addr->sllc_mac, IFHWADDRLEN);
 	}
-	dev = llc->dev;
-	if (sk->sk_type != SOCK_STREAM)
-		goto out;
-	rc = -EALREADY;
-	if (sock->state == SS_CONNECTING)
-		goto out;
+	llc->daddr.lsap = addr->sllc_sap;
+	memcpy(llc->daddr.mac, addr->sllc_mac, IFHWADDRLEN);
 	sock->state = SS_CONNECTING;
 	sk->sk_state   = TCP_SYN_SENT;
 	llc->link   = llc_ui_next_link_no(llc->sap->laddr.lsap);
-	rc = llc_establish_connection(sk, dev->dev_addr,
+	rc = llc_establish_connection(sk, llc->dev->dev_addr,
 				      addr->sllc_mac, addr->sllc_sap);
 	if (rc) {
 		dprintk("%s: llc_ui_send_conn failed :-(\n", __FUNCTION__);
@@ -439,12 +432,30 @@ static int llc_ui_connect(struct socket *sock, struct sockaddr *uaddr,
 		sk->sk_state = TCP_CLOSE;
 		goto out;
 	}
-	rc = llc_ui_wait_for_conn(sk, sk->sk_rcvtimeo);
-	if (rc)
-		dprintk("%s: llc_ui_wait_for_conn failed=%d\n", __FUNCTION__, rc);
+
+	if (sk->sk_state == TCP_SYN_SENT) {
+		const long timeo = sock_sndtimeo(sk, flags & O_NONBLOCK);
+
+		if (!timeo || !llc_ui_wait_for_conn(sk, timeo))
+			goto out;
+
+		rc = sock_intr_errno(timeo);
+		if (signal_pending(current))
+			goto out;
+	}
+
+	if (sk->sk_state == TCP_CLOSE)
+		goto sock_error;
+
+	sock->state = SS_CONNECTED;
+	rc = 0;
 out:
 	release_sock(sk);
 	return rc;
+sock_error:
+	rc = sock_error(sk) ? : -ECONNABORTED;
+	sock->state = SS_UNCONNECTED;
+	goto out;
 }
 
 /**
@@ -461,10 +472,10 @@ static int llc_ui_listen(struct socket *sock, int backlog)
 	int rc = -EINVAL;
 
 	lock_sock(sk);
-	if (sock->state != SS_UNCONNECTED)
+	if (unlikely(sock->state != SS_UNCONNECTED))
 		goto out;
 	rc = -EOPNOTSUPP;
-	if (sk->sk_type != SOCK_STREAM)
+	if (unlikely(sk->sk_type != SOCK_STREAM))
 		goto out;
 	rc = -EAGAIN;
 	if (sock_flag(sk, SOCK_ZAPPED))
@@ -483,20 +494,14 @@ out:
 	return rc;
 }
 
-static int llc_ui_wait_for_disc(struct sock *sk, int timeout)
+static int llc_ui_wait_for_disc(struct sock *sk, long timeout)
 {
-	DECLARE_WAITQUEUE(wait, current);
-	int rc;
+	DEFINE_WAIT(wait);
+	int rc = 0;
 
-	add_wait_queue_exclusive(sk->sk_sleep, &wait);
-	for (;;) {
-		__set_current_state(TASK_INTERRUPTIBLE);
-		rc = 0;
-		if (sk->sk_state != TCP_CLOSE) {
-			release_sock(sk);
-			timeout = schedule_timeout(timeout);
-			lock_sock(sk);
-		} else
+	while (1) {
+		prepare_to_wait(sk->sk_sleep, &wait, TASK_INTERRUPTIBLE);
+		if (sk_wait_event(sk, &timeout, sk->sk_state == TCP_CLOSE))
 			break;
 		rc = -ERESTARTSYS;
 		if (signal_pending(current))
@@ -504,65 +509,40 @@ static int llc_ui_wait_for_disc(struct sock *sk, int timeout)
 		rc = -EAGAIN;
 		if (!timeout)
 			break;
+		rc = 0;
 	}
-	__set_current_state(TASK_RUNNING);
-	remove_wait_queue(sk->sk_sleep, &wait);
+	finish_wait(sk->sk_sleep, &wait);
 	return rc;
 }
 
-static int llc_ui_wait_for_conn(struct sock *sk, int timeout)
+static int llc_ui_wait_for_conn(struct sock *sk, long timeout)
 {
-	DECLARE_WAITQUEUE(wait, current);
-	int rc;
+	DEFINE_WAIT(wait);
 
-	add_wait_queue_exclusive(sk->sk_sleep, &wait);
-	for (;;) {
-		__set_current_state(TASK_INTERRUPTIBLE);
-		rc = -EAGAIN;
-		if (sk->sk_state == TCP_CLOSE)
-			break;
-		rc = 0;
-		if (sk->sk_state != TCP_ESTABLISHED) {
-			release_sock(sk);
-			timeout = schedule_timeout(timeout);
-			lock_sock(sk);
-		} else
+	while (1) {
+		prepare_to_wait(sk->sk_sleep, &wait, TASK_INTERRUPTIBLE);
+		if (sk_wait_event(sk, &timeout, sk->sk_state != TCP_SYN_SENT))
 			break;
-		rc = -ERESTARTSYS;
-		if (signal_pending(current))
-			break;
-		rc = -EAGAIN;
-		if (!timeout)
+		if (signal_pending(current) || !timeout)
 			break;
 	}
-	__set_current_state(TASK_RUNNING);
-	remove_wait_queue(sk->sk_sleep, &wait);
-	return rc;
+	finish_wait(sk->sk_sleep, &wait);
+	return timeout;
 }
 
-static int llc_ui_wait_for_data(struct sock *sk, int timeout)
+static int llc_ui_wait_for_busy_core(struct sock *sk, long timeout)
 {
-	DECLARE_WAITQUEUE(wait, current);
-	int rc = 0;
+	DEFINE_WAIT(wait);
+	struct llc_sock *llc = llc_sk(sk);
+	int rc;
 
-	add_wait_queue_exclusive(sk->sk_sleep, &wait);
-	for (;;) {
-		__set_current_state(TASK_INTERRUPTIBLE);
-		if (sk->sk_shutdown & RCV_SHUTDOWN)
-			break;
-		/*
-		 * Well, if we have backlog, try to process it now.
-		 */
-                if (sk->sk_backlog.tail) {
-			release_sock(sk);
-			lock_sock(sk);
-		}
+	while (1) {
+		prepare_to_wait(sk->sk_sleep, &wait, TASK_INTERRUPTIBLE);
 		rc = 0;
-		if (skb_queue_empty(&sk->sk_receive_queue)) {
-			release_sock(sk);
-			timeout = schedule_timeout(timeout);
-			lock_sock(sk);
-		} else
+		if (sk_wait_event(sk, &timeout,
+				  (sk->sk_shutdown & RCV_SHUTDOWN) ||
+				  (!llc_data_accept_state(llc->state) &&
+				   !llc->p_flag)))
 			break;
 		rc = -ERESTARTSYS;
 		if (signal_pending(current))
@@ -571,40 +551,35 @@ static int llc_ui_wait_for_data(struct sock *sk, int timeout)
 		if (!timeout)
 			break;
 	}
-	__set_current_state(TASK_RUNNING);
-	remove_wait_queue(sk->sk_sleep, &wait);
+	finish_wait(sk->sk_sleep, &wait);
 	return rc;
 }
 
-static int llc_ui_wait_for_busy_core(struct sock *sk, int timeout)
+static int llc_wait_data(struct sock *sk, long timeo)
 {
-	DECLARE_WAITQUEUE(wait, current);
-	struct llc_sock *llc = llc_sk(sk);
 	int rc;
 
-	add_wait_queue_exclusive(sk->sk_sleep, &wait);
-	for (;;) {
-		dprintk("%s: looping...\n", __FUNCTION__);
-		__set_current_state(TASK_INTERRUPTIBLE);
-		rc = -ENOTCONN;
-		if (sk->sk_shutdown & RCV_SHUTDOWN)
+	while (1) {
+		/*
+		 * POSIX 1003.1g mandates this order.
+		 */
+		if (sk->sk_err) {
+			rc = sock_error(sk);
 			break;
+		}
 		rc = 0;
-		if (llc_data_accept_state(llc->state) || llc->p_flag) {
-			release_sock(sk);
-			timeout = schedule_timeout(timeout);
-			lock_sock(sk);
-		} else
+		if (sk->sk_shutdown & RCV_SHUTDOWN)
 			break;
-		rc = -ERESTARTSYS;
+		rc = -EAGAIN;
+		if (!timeo)
+			break;
+		rc = sock_intr_errno(timeo);
 		if (signal_pending(current))
 			break;
-		rc = -EAGAIN;
-		if (!timeout)
+		rc = 0;
+		if (sk_wait_data(sk, &timeo))
 			break;
 	}
-	__set_current_state(TASK_RUNNING);
-	remove_wait_queue(sk->sk_sleep, &wait);
 	return rc;
 }
 
@@ -627,15 +602,18 @@ static int llc_ui_accept(struct socket *sock, struct socket *newsock, int flags)
 	dprintk("%s: accepting on %02X\n", __FUNCTION__,
 	        llc_sk(sk)->laddr.lsap);
 	lock_sock(sk);
-	if (sk->sk_type != SOCK_STREAM)
+	if (unlikely(sk->sk_type != SOCK_STREAM))
 		goto out;
 	rc = -EINVAL;
-	if (sock->state != SS_UNCONNECTED || sk->sk_state != TCP_LISTEN)
+	if (unlikely(sock->state != SS_UNCONNECTED ||
+		     sk->sk_state != TCP_LISTEN))
 		goto out;
 	/* wait for a connection to arrive. */
-	rc = llc_ui_wait_for_data(sk, sk->sk_rcvtimeo);
-	if (rc)
-		goto out;
+	if (skb_queue_empty(&sk->sk_receive_queue)) {
+		rc = llc_wait_data(sk, sk->sk_rcvtimeo);
+		if (rc)
+			goto out;
+	}
 	dprintk("%s: got a new connection on %02X\n", __FUNCTION__,
 	        llc_sk(sk)->laddr.lsap);
 	skb = skb_dequeue(&sk->sk_receive_queue);
@@ -657,7 +635,6 @@ static int llc_ui_accept(struct socket *sock, struct socket *newsock, int flags)
 	/* put original socket back into a clean listen state. */
 	sk->sk_state = TCP_LISTEN;
 	sk->sk_ack_backlog--;
-	skb->sk = NULL;
 	dprintk("%s: ok success on %02X, client on %02X\n", __FUNCTION__,
 		llc_sk(sk)->addr.sllc_sap, newllc->daddr.lsap);
 frees:
@@ -671,56 +648,167 @@ out:
  *	llc_ui_recvmsg - copy received data to the socket user.
  *	@sock: Socket to copy data from.
  *	@msg: Various user space related information.
- *	@size: Size of user buffer.
+ *	@len: Size of user buffer.
  *	@flags: User specified flags.
  *
  *	Copy received data to the socket user.
  *	Returns non-negative upon success, negative otherwise.
  */
 static int llc_ui_recvmsg(struct kiocb *iocb, struct socket *sock,
-			  struct msghdr *msg, size_t size, int flags)
+			  struct msghdr *msg, size_t len, int flags)
 {
-	struct sock *sk = sock->sk;
 	struct sockaddr_llc *uaddr = (struct sockaddr_llc *)msg->msg_name;
-	struct sk_buff *skb;
+	const int nonblock = flags & MSG_DONTWAIT;
+	struct sk_buff *skb = NULL;
+	struct sock *sk = sock->sk;
+	struct llc_sock *llc = llc_sk(sk);
 	size_t copied = 0;
-	int rc = -ENOMEM, timeout;
-	int noblock = flags & MSG_DONTWAIT;
+	u32 peek_seq = 0;
+	u32 *seq;
+	unsigned long used;
+	int target;	/* Read at least this many bytes */
+	long timeo;
 
-	dprintk("%s: receiving in %02X from %02X\n", __FUNCTION__,
-		llc_sk(sk)->laddr.lsap, llc_sk(sk)->daddr.lsap);
 	lock_sock(sk);
-	timeout = sock_rcvtimeo(sk, noblock);
-	rc = llc_ui_wait_for_data(sk, timeout);
-	if (rc) {
-		dprintk("%s: llc_ui_wait_for_data failed recv "
-			"in %02X from %02X\n", __FUNCTION__,
-			llc_sk(sk)->laddr.lsap, llc_sk(sk)->daddr.lsap);
+	copied = -ENOTCONN;
+	if (sk->sk_state == TCP_LISTEN)
 		goto out;
-	}
-	skb = skb_dequeue(&sk->sk_receive_queue);
-	if (!skb) /* shutdown */
-		goto out;
-	copied = skb->len;
-	if (copied > size)
-		copied = size;
-	rc = skb_copy_datagram_iovec(skb, 0, msg->msg_iov, copied);
-	if (rc)
-		goto dgram_free;
-	if (skb->len > copied) {
-		skb_pull(skb, copied);
-		skb_queue_head(&sk->sk_receive_queue, skb);
-	}
-	if (uaddr)
-		memcpy(uaddr, llc_ui_skb_cb(skb), sizeof(*uaddr));
-	msg->msg_namelen = sizeof(*uaddr);
-	if (!skb->next) {
-dgram_free:
-		kfree_skb(skb);
-	}
+
+	timeo = sock_rcvtimeo(sk, nonblock);
+
+	seq = &llc->copied_seq;
+	if (flags & MSG_PEEK) {
+		peek_seq = llc->copied_seq;
+		seq = &peek_seq;
+ 	}
+
+	target = sock_rcvlowat(sk, flags & MSG_WAITALL, len);
+	copied = 0;
+
+	do {
+		u32 offset;
+
+		/*
+		 * We need to check signals first, to get correct SIGURG
+		 * handling. FIXME: Need to check this doesn't impact 1003.1g
+		 * and move it down to the bottom of the loop
+		 */
+		if (signal_pending(current)) {
+			if (copied)
+				break;
+			copied = timeo ? sock_intr_errno(timeo) : -EAGAIN;
+			break;
+		}
+
+		/* Next get a buffer. */
+
+		skb = skb_peek(&sk->sk_receive_queue);
+		if (skb) {
+			offset = *seq;
+			goto found_ok_skb;
+		}
+		/* Well, if we have backlog, try to process it now yet. */
+
+		if (copied >= target && !sk->sk_backlog.tail)
+			break;
+
+		if (copied) {
+			if (sk->sk_err ||
+			    sk->sk_state == TCP_CLOSE ||
+			    (sk->sk_shutdown & RCV_SHUTDOWN) ||
+			    !timeo ||
+			    (flags & MSG_PEEK))
+				break;
+		} else {
+			if (sock_flag(sk, SOCK_DONE))
+				break;
+
+			if (sk->sk_err) {
+				copied = sock_error(sk);
+				break;
+			}
+			if (sk->sk_shutdown & RCV_SHUTDOWN)
+				break;
+
+			if (sk->sk_state == TCP_CLOSE) {
+				if (!sock_flag(sk, SOCK_DONE)) {
+					/*
+					 * This occurs when user tries to read
+					 * from never connected socket.
+					 */
+					copied = -ENOTCONN;
+					break;
+				}
+				break;
+			}
+			if (!timeo) {
+				copied = -EAGAIN;
+				break;
+			}
+		}
+
+		if (copied >= target) { /* Do not sleep, just process backlog. */
+			release_sock(sk);
+			lock_sock(sk);
+		} else
+			sk_wait_data(sk, &timeo);
+
+		if ((flags & MSG_PEEK) && peek_seq != llc->copied_seq) {
+			if (net_ratelimit())
+				printk(KERN_DEBUG "LLC(%s:%d): Application "
+						  "bug, race in MSG_PEEK.\n",
+				       current->comm, current->pid);
+			peek_seq = llc->copied_seq;
+		}
+		continue;
+	found_ok_skb:
+		/* Ok so how much can we use? */
+		used = skb->len - offset;
+		if (len < used)
+			used = len;
+
+		if (!(flags & MSG_TRUNC)) {
+			int rc = skb_copy_datagram_iovec(skb, offset,
+							 msg->msg_iov, used);
+			if (rc) {
+				/* Exception. Bailout! */
+				if (!copied)
+					copied = -EFAULT;
+				break;
+			}
+		}
+
+		*seq += used;
+		copied += used;
+		len -= used;
+
+		if (used + offset < skb->len)
+			continue;
+
+		if (!(flags & MSG_PEEK)) {
+			sk_eat_skb(sk, skb);
+			*seq = 0;
+		}
+	} while (len > 0);
+
+	/* 
+	 * According to UNIX98, msg_name/msg_namelen are ignored
+	 * on connected socket. -ANK
+	 * But... af_llc still doesn't have separate sets of methods for
+	 * SOCK_DGRAM and SOCK_STREAM :-( So we have to do this test, will
+	 * eventually fix this tho :-) -acme
+	 */
+	if (sk->sk_type == SOCK_DGRAM)
+		goto copy_uaddr;
 out:
 	release_sock(sk);
-	return rc ? : copied;
+	return copied;
+copy_uaddr:
+	if (uaddr != NULL && skb != NULL) {
+		memcpy(uaddr, llc_ui_skb_cb(skb), sizeof(*uaddr));
+		msg->msg_namelen = sizeof(*uaddr);
+	}
+	goto out;
 }
 
 /**
@@ -740,7 +828,6 @@ static int llc_ui_sendmsg(struct kiocb *iocb, struct socket *sock,
 	struct sockaddr_llc *addr = (struct sockaddr_llc *)msg->msg_name;
 	int flags = msg->msg_flags;
 	int noblock = flags & MSG_DONTWAIT;
-	struct net_device *dev;
 	struct sk_buff *skb;
 	size_t size = 0;
 	int rc = -EINVAL, copied = 0, hdrlen;
@@ -763,19 +850,17 @@ static int llc_ui_sendmsg(struct kiocb *iocb, struct socket *sock,
 		if (rc)
 			goto release;
 	}
-	dev = llc->dev;
-	hdrlen = dev->hard_header_len + llc_ui_header_len(sk, addr);
+	hdrlen = llc->dev->hard_header_len + llc_ui_header_len(sk, addr);
 	size = hdrlen + len;
-	if (size > dev->mtu)
-		size = dev->mtu;
+	if (size > llc->dev->mtu)
+		size = llc->dev->mtu;
 	copied = size - hdrlen;
 	release_sock(sk);
 	skb = sock_alloc_send_skb(sk, size, noblock, &rc);
 	lock_sock(sk);
 	if (!skb)
 		goto release;
-	skb->sk	      = sk;
-	skb->dev      = dev;
+	skb->dev      = llc->dev;
 	skb->protocol = llc_proto_type(addr->sllc_arphrd);
 	skb_reserve(skb, hdrlen); 
 	rc = memcpy_fromiovec(skb_put(skb, copied), msg->msg_iov, copied);
@@ -800,15 +885,13 @@ static int llc_ui_sendmsg(struct kiocb *iocb, struct socket *sock,
 	if (!(sk->sk_type == SOCK_STREAM && !addr->sllc_ua))
 		goto out;
 	rc = llc_ui_send_data(sk, skb, noblock);
-	if (rc)
-		dprintk("%s: llc_ui_send_data failed: %d\n", __FUNCTION__, rc);
 out:
-	if (rc)
+	if (rc) {
 		kfree_skb(skb);
 release:
-	if (rc)
 		dprintk("%s: failed sending from %02X to %02X: %d\n",
 			__FUNCTION__, llc->laddr.lsap, llc->daddr.lsap, rc);
+	}
 	release_sock(sk);
 	return rc ? : copied;
 }
@@ -895,7 +978,7 @@ static int llc_ui_setsockopt(struct socket *sock, int level, int optname,
 	int rc = -EINVAL, opt;
 
 	lock_sock(sk);
-	if (level != SOL_LLC || optlen != sizeof(int))
+	if (unlikely(level != SOL_LLC || optlen != sizeof(int)))
 		goto out;
 	rc = get_user(opt, (int __user *)optval);
 	if (rc)
@@ -915,22 +998,22 @@ static int llc_ui_setsockopt(struct socket *sock, int level, int optname,
 	case LLC_OPT_ACK_TMR_EXP:
 		if (opt > LLC_OPT_MAX_ACK_TMR_EXP)
 			goto out;
-		llc->ack_timer.expire = opt;
+		llc->ack_timer.expire = opt * HZ;
 		break;
 	case LLC_OPT_P_TMR_EXP:
 		if (opt > LLC_OPT_MAX_P_TMR_EXP)
 			goto out;
-		llc->pf_cycle_timer.expire = opt;
+		llc->pf_cycle_timer.expire = opt * HZ;
 		break;
 	case LLC_OPT_REJ_TMR_EXP:
 		if (opt > LLC_OPT_MAX_REJ_TMR_EXP)
 			goto out;
-		llc->rej_sent_timer.expire = opt;
+		llc->rej_sent_timer.expire = opt * HZ;
 		break;
 	case LLC_OPT_BUSY_TMR_EXP:
 		if (opt > LLC_OPT_MAX_BUSY_TMR_EXP)
 			goto out;
-		llc->busy_state_timer.expire = opt;
+		llc->busy_state_timer.expire = opt * HZ;
 		break;
 	case LLC_OPT_TX_WIN:
 		if (opt > LLC_OPT_MAX_WIN)
@@ -970,7 +1053,7 @@ static int llc_ui_getsockopt(struct socket *sock, int level, int optname,
 	int val = 0, len = 0, rc = -EINVAL;
 
 	lock_sock(sk);
-	if (level != SOL_LLC)
+	if (unlikely(level != SOL_LLC))
 		goto out;
 	rc = get_user(len, optlen);
 	if (rc)
@@ -980,17 +1063,17 @@ static int llc_ui_getsockopt(struct socket *sock, int level, int optname,
 		goto out;
 	switch (optname) {
 	case LLC_OPT_RETRY:
-		val = llc->n2;				break;
+		val = llc->n2;					break;
 	case LLC_OPT_SIZE:
-		val = llc->n1;				break;
+		val = llc->n1;					break;
 	case LLC_OPT_ACK_TMR_EXP:
-		val = llc->ack_timer.expire;		break;
+		val = llc->ack_timer.expire / HZ;		break;
 	case LLC_OPT_P_TMR_EXP:
-		val = llc->pf_cycle_timer.expire;	break;
+		val = llc->pf_cycle_timer.expire / HZ;		break;
 	case LLC_OPT_REJ_TMR_EXP:
-		val = llc->rej_sent_timer.expire;	break;
+		val = llc->rej_sent_timer.expire / HZ;		break;
 	case LLC_OPT_BUSY_TMR_EXP:
-		val = llc->busy_state_timer.expire;	break;
+		val = llc->busy_state_timer.expire / HZ;	break;
 	case LLC_OPT_TX_WIN:
 		val = llc->k;				break;
 	case LLC_OPT_RX_WIN:
@@ -1034,8 +1117,12 @@ static struct proto_ops llc_ui_ops = {
 	.sendpage    = sock_no_sendpage,
 };
 
-extern void llc_sap_handler(struct llc_sap *sap, struct sk_buff *skb);
-extern void llc_conn_handler(struct llc_sap *sap, struct sk_buff *skb);
+static char llc_proc_err_msg[] __initdata =
+        KERN_CRIT "LLC: Unable to register the proc_fs entries\n";
+static char llc_sysctl_err_msg[] __initdata =
+        KERN_CRIT "LLC: Unable to register the sysctl entries\n";
+static char llc_sock_err_msg[] __initdata =
+        KERN_CRIT "LLC: Unable to register the network family\n";
 
 static int __init llc2_init(void)
 {
@@ -1048,13 +1135,28 @@ static int __init llc2_init(void)
 	llc_station_init();
 	llc_ui_sap_last_autoport = LLC_SAP_DYN_START;
 	rc = llc_proc_init();
-	if (rc != 0)
+	if (rc != 0) {
+		printk(llc_proc_err_msg);
 		goto out_unregister_llc_proto;
-	sock_register(&llc_ui_family_ops);
+	}
+	rc = llc_sysctl_init();
+	if (rc) {
+		printk(llc_sysctl_err_msg);
+		goto out_proc;
+	}
+	rc = sock_register(&llc_ui_family_ops);
+	if (rc) {
+		printk(llc_sock_err_msg);
+		goto out_sysctl;
+	}
 	llc_add_pack(LLC_DEST_SAP, llc_sap_handler);
 	llc_add_pack(LLC_DEST_CONN, llc_conn_handler);
 out:
 	return rc;
+out_sysctl:
+	llc_sysctl_exit();
+out_proc:
+	llc_proc_exit();
 out_unregister_llc_proto:
 	proto_unregister(&llc_proto);
 	goto out;
@@ -1067,6 +1169,7 @@ static void __exit llc2_exit(void)
 	llc_remove_pack(LLC_DEST_CONN);
 	sock_unregister(PF_LLC);
 	llc_proc_exit();
+	llc_sysctl_exit();
 	proto_unregister(&llc_proto);
 }
 
diff --git a/net/llc/llc_c_ac.c b/net/llc/llc_c_ac.c
index b218be4c10e..b0bcfb1f12d 100644
--- a/net/llc/llc_c_ac.c
+++ b/net/llc/llc_c_ac.c
@@ -60,23 +60,10 @@ int llc_conn_ac_clear_remote_busy(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ac_conn_ind(struct sock *sk, struct sk_buff *skb)
 {
-	int rc = -ENOTCONN;
-	u8 dsap;
-	struct llc_sap *sap;
-
-	llc_pdu_decode_dsap(skb, &dsap);
-	sap = llc_sap_find(dsap);
-	if (sap) {
-		struct llc_conn_state_ev *ev = llc_conn_ev(skb);
-		struct llc_sock *llc = llc_sk(sk);
+	struct llc_conn_state_ev *ev = llc_conn_ev(skb);
 
-		llc_pdu_decode_sa(skb, llc->daddr.mac);
-		llc_pdu_decode_da(skb, llc->laddr.mac);
-		llc->dev = skb->dev;
-		ev->ind_prim = LLC_CONN_PRIM;
-		rc = 0;
-	}
-	return rc;
+	ev->ind_prim = LLC_CONN_PRIM;
+	return 0;
 }
 
 int llc_conn_ac_conn_confirm(struct sock *sk, struct sk_buff *skb)
@@ -120,10 +107,8 @@ int llc_conn_ac_disc_ind(struct sock *sk, struct sk_buff *skb)
 			reason = LLC_DISC_REASON_RX_DISC_CMD_PDU;
 	} else if (ev->type == LLC_CONN_EV_TYPE_ACK_TMR)
 		reason = LLC_DISC_REASON_ACK_TMR_EXP;
-	else {
-		reason = 0;
+	else
 		rc = -EINVAL;
-	}
 	if (!rc) {
 		ev->reason   = reason;
 		ev->ind_prim = LLC_DISC_PRIM;
@@ -160,9 +145,6 @@ int llc_conn_ac_rst_ind(struct sock *sk, struct sk_buff *skb)
 			   LLC_U_PDU_CMD(pdu) == LLC_2_PDU_CMD_SABME) {
 			reason = LLC_RESET_REASON_REMOTE;
 			rc = 0;
-		} else {
-			reason = 0;
-			rc  = 1;
 		}
 		break;
 	case LLC_CONN_EV_TYPE_ACK_TMR:
@@ -172,8 +154,7 @@ int llc_conn_ac_rst_ind(struct sock *sk, struct sk_buff *skb)
 		if (llc->retry_count > llc->n2) {
 			reason = LLC_RESET_REASON_LOCAL;
 			rc = 0;
-		} else
-			rc = 1;
+		}
 		break;
 	}
 	if (!rc) {
@@ -217,18 +198,17 @@ int llc_conn_ac_stop_rej_tmr_if_data_flag_eq_2(struct sock *sk,
 int llc_conn_ac_send_disc_cmd_p_set_x(struct sock *sk, struct sk_buff *skb)
 {
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	if (nskb) {
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
 
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_U, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_CMD);
 		llc_pdu_init_as_disc_cmd(nskb, 1);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 		llc_conn_ac_set_p_flag_1(sk, skb);
@@ -243,20 +223,19 @@ free:
 int llc_conn_ac_send_dm_rsp_f_set_p(struct sock *sk, struct sk_buff *skb)
 {
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	if (nskb) {
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
 		u8 f_bit;
 
-		nskb->dev = llc->dev;
 		llc_pdu_decode_pf_bit(skb, &f_bit);
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_U, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_RSP);
 		llc_pdu_init_as_dm_rsp(nskb, f_bit);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 	}
@@ -270,19 +249,17 @@ free:
 int llc_conn_ac_send_dm_rsp_f_set_1(struct sock *sk, struct sk_buff *skb)
 {
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	if (nskb) {
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
-		u8 f_bit = 1;
 
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_U, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_RSP);
-		llc_pdu_init_as_dm_rsp(nskb, f_bit);
+		llc_pdu_init_as_dm_rsp(nskb, 1);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 	}
@@ -306,17 +283,16 @@ int llc_conn_ac_send_frmr_rsp_f_set_x(struct sock *sk, struct sk_buff *skb)
 		llc_pdu_decode_pf_bit(skb, &f_bit);
 	else
 		f_bit = 0;
-	nskb = llc_alloc_frame();
+	nskb = llc_alloc_frame(sk, llc->dev);
 	if (nskb) {
 		struct llc_sap *sap = llc->sap;
 
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_U, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_RSP);
 		llc_pdu_init_as_frmr_rsp(nskb, pdu, f_bit, llc->vS,
 					 llc->vR, INCORRECT);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 	}
@@ -330,21 +306,19 @@ free:
 int llc_conn_ac_resend_frmr_rsp_f_set_0(struct sock *sk, struct sk_buff *skb)
 {
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	if (nskb) {
-		u8 f_bit = 0;
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
 		struct llc_pdu_sn *pdu = (struct llc_pdu_sn *)&llc->rx_pdu_hdr;
 
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_U, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_RSP);
-		llc_pdu_init_as_frmr_rsp(nskb, pdu, f_bit, llc->vS,
+		llc_pdu_init_as_frmr_rsp(nskb, pdu, 0, llc->vS,
 					 llc->vR, INCORRECT);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 	}
@@ -360,21 +334,20 @@ int llc_conn_ac_resend_frmr_rsp_f_set_p(struct sock *sk, struct sk_buff *skb)
 	u8 f_bit;
 	int rc = -ENOBUFS;
 	struct sk_buff *nskb;
+	struct llc_sock *llc = llc_sk(sk);
 
 	llc_pdu_decode_pf_bit(skb, &f_bit);
-	nskb = llc_alloc_frame();
+	nskb = llc_alloc_frame(sk, llc->dev);
 	if (nskb) {
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
 		struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_U, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_RSP);
 		llc_pdu_init_as_frmr_rsp(nskb, pdu, f_bit, llc->vS,
 					 llc->vR, INCORRECT);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 	}
@@ -395,7 +368,7 @@ int llc_conn_ac_send_i_cmd_p_set_1(struct sock *sk, struct sk_buff *skb)
 			    llc->daddr.lsap, LLC_PDU_CMD);
 	llc_pdu_init_as_i_cmd(skb, 1, llc->vS, llc->vR);
 	rc = llc_mac_hdr_init(skb, llc->dev->dev_addr, llc->daddr.mac);
-	if (!rc) {
+	if (likely(!rc)) {
 		llc_conn_send_pdu(sk, skb);
 		llc_conn_ac_inc_vs_by_1(sk, skb);
 	}
@@ -412,7 +385,7 @@ static int llc_conn_ac_send_i_cmd_p_set_0(struct sock *sk, struct sk_buff *skb)
 			    llc->daddr.lsap, LLC_PDU_CMD);
 	llc_pdu_init_as_i_cmd(skb, 0, llc->vS, llc->vR);
 	rc = llc_mac_hdr_init(skb, llc->dev->dev_addr, llc->daddr.mac);
-	if (!rc) {
+	if (likely(!rc)) {
 		llc_conn_send_pdu(sk, skb);
 		llc_conn_ac_inc_vs_by_1(sk, skb);
 	}
@@ -429,7 +402,7 @@ int llc_conn_ac_send_i_xxx_x_set_0(struct sock *sk, struct sk_buff *skb)
 			    llc->daddr.lsap, LLC_PDU_CMD);
 	llc_pdu_init_as_i_cmd(skb, 0, llc->vS, llc->vR);
 	rc = llc_mac_hdr_init(skb, llc->dev->dev_addr, llc->daddr.mac);
-	if (!rc) {
+	if (likely(!rc)) {
 		llc_conn_send_pdu(sk, skb);
 		llc_conn_ac_inc_vs_by_1(sk, skb);
 	}
@@ -451,18 +424,17 @@ int llc_conn_ac_resend_i_xxx_x_set_0_or_send_rr(struct sock *sk,
 	u8 nr;
 	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	if (nskb) {
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
 
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_U, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_RSP);
 		llc_pdu_init_as_rr_rsp(nskb, 0, llc->vR);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (!rc)
+		if (likely(!rc))
 			llc_conn_send_pdu(sk, nskb);
 		else
 			kfree_skb(skb);
@@ -487,18 +459,17 @@ int llc_conn_ac_resend_i_rsp_f_set_1(struct sock *sk, struct sk_buff *skb)
 int llc_conn_ac_send_rej_cmd_p_set_1(struct sock *sk, struct sk_buff *skb)
 {
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	if (nskb) {
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
 
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_S, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_CMD);
 		llc_pdu_init_as_rej_cmd(nskb, 1, llc->vR);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 	}
@@ -512,19 +483,17 @@ free:
 int llc_conn_ac_send_rej_rsp_f_set_1(struct sock *sk, struct sk_buff *skb)
 {
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	if (nskb) {
-		u8 f_bit = 1;
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
 
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_S, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_RSP);
-		llc_pdu_init_as_rej_rsp(nskb, f_bit, llc->vR);
+		llc_pdu_init_as_rej_rsp(nskb, 1, llc->vR);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 	}
@@ -538,19 +507,17 @@ free:
 int llc_conn_ac_send_rej_xxx_x_set_0(struct sock *sk, struct sk_buff *skb)
 {
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	if (nskb) {
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
-		u8 f_bit = 0;
 
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_S, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_RSP);
-		llc_pdu_init_as_rej_rsp(nskb, f_bit, llc->vR);
+		llc_pdu_init_as_rej_rsp(nskb, 0, llc->vR);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 	}
@@ -564,18 +531,17 @@ free:
 int llc_conn_ac_send_rnr_cmd_p_set_1(struct sock *sk, struct sk_buff *skb)
 {
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	if (nskb) {
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
 
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_S, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_CMD);
 		llc_pdu_init_as_rnr_cmd(nskb, 1, llc->vR);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 	}
@@ -589,19 +555,17 @@ free:
 int llc_conn_ac_send_rnr_rsp_f_set_1(struct sock *sk, struct sk_buff *skb)
 {
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	if (nskb) {
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
-		u8 f_bit = 1;
 
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_S, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_RSP);
-		llc_pdu_init_as_rnr_rsp(nskb, f_bit, llc->vR);
+		llc_pdu_init_as_rnr_rsp(nskb, 1, llc->vR);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 	}
@@ -615,19 +579,17 @@ free:
 int llc_conn_ac_send_rnr_xxx_x_set_0(struct sock *sk, struct sk_buff *skb)
 {
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	if (nskb) {
-		u8 f_bit = 0;
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
 
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_S, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_RSP);
-		llc_pdu_init_as_rnr_rsp(nskb, f_bit, llc->vR);
+		llc_pdu_init_as_rnr_rsp(nskb, 0, llc->vR);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 	}
@@ -645,7 +607,7 @@ int llc_conn_ac_set_remote_busy(struct sock *sk, struct sk_buff *skb)
 	if (!llc->remote_busy_flag) {
 		llc->remote_busy_flag = 1;
 		mod_timer(&llc->busy_state_timer.timer,
-			 jiffies + llc->busy_state_timer.expire * HZ);
+			 jiffies + llc->busy_state_timer.expire);
 	}
 	return 0;
 }
@@ -653,18 +615,17 @@ int llc_conn_ac_set_remote_busy(struct sock *sk, struct sk_buff *skb)
 int llc_conn_ac_opt_send_rnr_xxx_x_set_0(struct sock *sk, struct sk_buff *skb)
 {
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	if (nskb) {
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
 
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_S, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_RSP);
 		llc_pdu_init_as_rnr_rsp(nskb, 0, llc->vR);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 	}
@@ -678,18 +639,17 @@ free:
 int llc_conn_ac_send_rr_cmd_p_set_1(struct sock *sk, struct sk_buff *skb)
 {
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	if (nskb) {
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
 
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_S, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_CMD);
 		llc_pdu_init_as_rr_cmd(nskb, 1, llc->vR);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 	}
@@ -703,19 +663,18 @@ free:
 int llc_conn_ac_send_rr_rsp_f_set_1(struct sock *sk, struct sk_buff *skb)
 {
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	if (nskb) {
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
 		u8 f_bit = 1;
 
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_S, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_RSP);
 		llc_pdu_init_as_rr_rsp(nskb, f_bit, llc->vR);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 	}
@@ -729,19 +688,17 @@ free:
 int llc_conn_ac_send_ack_rsp_f_set_1(struct sock *sk, struct sk_buff *skb)
 {
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	if (nskb) {
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
-		u8 f_bit = 1;
 
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_S, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_RSP);
-		llc_pdu_init_as_rr_rsp(nskb, f_bit, llc->vR);
+		llc_pdu_init_as_rr_rsp(nskb, 1, llc->vR);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 	}
@@ -755,18 +712,17 @@ free:
 int llc_conn_ac_send_rr_xxx_x_set_0(struct sock *sk, struct sk_buff *skb)
 {
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	if (nskb) {
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
 
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_S, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_RSP);
 		llc_pdu_init_as_rr_rsp(nskb, 0, llc->vR);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 	}
@@ -780,18 +736,17 @@ free:
 int llc_conn_ac_send_ack_xxx_x_set_0(struct sock *sk, struct sk_buff *skb)
 {
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	if (nskb) {
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
 
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_S, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_RSP);
 		llc_pdu_init_as_rr_rsp(nskb, 0, llc->vR);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 	}
@@ -815,8 +770,8 @@ void llc_conn_set_p_flag(struct sock *sk, u8 value)
 int llc_conn_ac_send_sabme_cmd_p_set_x(struct sock *sk, struct sk_buff *skb)
 {
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
 	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	if (nskb) {
 		struct llc_sap *sap = llc->sap;
@@ -824,12 +779,11 @@ int llc_conn_ac_send_sabme_cmd_p_set_x(struct sock *sk, struct sk_buff *skb)
 
 		if (llc->dev->flags & IFF_LOOPBACK)
 			dmac = llc->dev->dev_addr;
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_U, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_CMD);
 		llc_pdu_init_as_sabme_cmd(nskb, 1);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, dmac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 		llc_conn_set_p_flag(sk, 1);
@@ -845,11 +799,11 @@ int llc_conn_ac_send_ua_rsp_f_set_p(struct sock *sk, struct sk_buff *skb)
 {
 	u8 f_bit;
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	llc_pdu_decode_pf_bit(skb, &f_bit);
 	if (nskb) {
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
 
 		nskb->dev = llc->dev;
@@ -857,7 +811,7 @@ int llc_conn_ac_send_ua_rsp_f_set_p(struct sock *sk, struct sk_buff *skb)
 				    llc->daddr.lsap, LLC_PDU_RSP);
 		llc_pdu_init_as_ua_rsp(nskb, f_bit);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 	}
@@ -886,7 +840,7 @@ int llc_conn_ac_start_p_timer(struct sock *sk, struct sk_buff *skb)
 
 	llc_conn_set_p_flag(sk, 1);
 	mod_timer(&llc->pf_cycle_timer.timer,
-		  jiffies + llc->pf_cycle_timer.expire * HZ);
+		  jiffies + llc->pf_cycle_timer.expire);
 	return 0;
 }
 
@@ -957,7 +911,7 @@ static int llc_conn_ac_send_i_rsp_f_set_ackpf(struct sock *sk,
 			    llc->daddr.lsap, LLC_PDU_RSP);
 	llc_pdu_init_as_i_cmd(skb, llc->ack_pf, llc->vS, llc->vR);
 	rc = llc_mac_hdr_init(skb, llc->dev->dev_addr, llc->daddr.mac);
-	if (!rc) {
+	if (likely(!rc)) {
 		llc_conn_send_pdu(sk, skb);
 		llc_conn_ac_inc_vs_by_1(sk, skb);
 	}
@@ -1001,18 +955,17 @@ static int llc_conn_ac_send_rr_rsp_f_set_ackpf(struct sock *sk,
 					       struct sk_buff *skb)
 {
 	int rc = -ENOBUFS;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct llc_sock *llc = llc_sk(sk);
+	struct sk_buff *nskb = llc_alloc_frame(sk, llc->dev);
 
 	if (nskb) {
-		struct llc_sock *llc = llc_sk(sk);
 		struct llc_sap *sap = llc->sap;
 
-		nskb->dev = llc->dev;
 		llc_pdu_header_init(nskb, LLC_PDU_TYPE_S, sap->laddr.lsap,
 				    llc->daddr.lsap, LLC_PDU_RSP);
 		llc_pdu_init_as_rr_rsp(nskb, llc->ack_pf, llc->vR);
 		rc = llc_mac_hdr_init(nskb, llc->dev->dev_addr, llc->daddr.mac);
-		if (rc)
+		if (unlikely(rc))
 			goto free;
 		llc_conn_send_pdu(sk, nskb);
 	}
@@ -1165,7 +1118,7 @@ int llc_conn_ac_start_ack_timer(struct sock *sk, struct sk_buff *skb)
 {
 	struct llc_sock *llc = llc_sk(sk);
 
-	mod_timer(&llc->ack_timer.timer, jiffies + llc->ack_timer.expire * HZ);
+	mod_timer(&llc->ack_timer.timer, jiffies + llc->ack_timer.expire);
 	return 0;
 }
 
@@ -1174,7 +1127,7 @@ int llc_conn_ac_start_rej_timer(struct sock *sk, struct sk_buff *skb)
 	struct llc_sock *llc = llc_sk(sk);
 
 	mod_timer(&llc->rej_sent_timer.timer,
-		  jiffies + llc->rej_sent_timer.expire * HZ);
+		  jiffies + llc->rej_sent_timer.expire);
 	return 0;
 }
 
@@ -1185,7 +1138,7 @@ int llc_conn_ac_start_ack_tmr_if_not_running(struct sock *sk,
 
 	if (!timer_pending(&llc->ack_timer.timer))
 		mod_timer(&llc->ack_timer.timer,
-			  jiffies + llc->ack_timer.expire * HZ);
+			  jiffies + llc->ack_timer.expire);
 	return 0;
 }
 
@@ -1233,7 +1186,7 @@ int llc_conn_ac_upd_nr_received(struct sock *sk, struct sk_buff *skb)
 		}
 		if (unacked)
 			mod_timer(&llc->ack_timer.timer,
-				  jiffies + llc->ack_timer.expire * HZ);
+				  jiffies + llc->ack_timer.expire);
 	} else if (llc->failed_data_req) {
 		u8 f_bit;
 
@@ -1354,13 +1307,13 @@ int llc_conn_ac_set_vs_nr(struct sock *sk, struct sk_buff *skb)
 	return 0;
 }
 
-int llc_conn_ac_inc_vs_by_1(struct sock *sk, struct sk_buff *skb)
+static int llc_conn_ac_inc_vs_by_1(struct sock *sk, struct sk_buff *skb)
 {
 	llc_sk(sk)->vS = (llc_sk(sk)->vS + 1) % 128;
 	return 0;
 }
 
-void llc_conn_pf_cycle_tmr_cb(unsigned long timeout_data)
+static void llc_conn_tmr_common_cb(unsigned long timeout_data, u8 type)
 {
 	struct sock *sk = (struct sock *)timeout_data;
 	struct sk_buff *skb = alloc_skb(0, GFP_ATOMIC);
@@ -1369,59 +1322,31 @@ void llc_conn_pf_cycle_tmr_cb(unsigned long timeout_data)
 	if (skb) {
 		struct llc_conn_state_ev *ev = llc_conn_ev(skb);
 
-		skb->sk  = sk;
-		ev->type = LLC_CONN_EV_TYPE_P_TMR;
+		skb_set_owner_r(skb, sk);
+		ev->type = type;
 		llc_process_tmr_ev(sk, skb);
 	}
 	bh_unlock_sock(sk);
 }
 
-void llc_conn_busy_tmr_cb(unsigned long timeout_data)
+void llc_conn_pf_cycle_tmr_cb(unsigned long timeout_data)
 {
-	struct sock *sk = (struct sock *)timeout_data;
-	struct sk_buff *skb = alloc_skb(0, GFP_ATOMIC);
-
-	bh_lock_sock(sk);
-	if (skb) {
-		struct llc_conn_state_ev *ev = llc_conn_ev(skb);
+	llc_conn_tmr_common_cb(timeout_data, LLC_CONN_EV_TYPE_P_TMR);
+}
 
-		skb->sk  = sk;
-		ev->type = LLC_CONN_EV_TYPE_BUSY_TMR;
-		llc_process_tmr_ev(sk, skb);
-	}
-	bh_unlock_sock(sk);
+void llc_conn_busy_tmr_cb(unsigned long timeout_data)
+{
+	llc_conn_tmr_common_cb(timeout_data, LLC_CONN_EV_TYPE_BUSY_TMR);
 }
 
 void llc_conn_ack_tmr_cb(unsigned long timeout_data)
 {
-	struct sock* sk = (struct sock *)timeout_data;
-	struct sk_buff *skb = alloc_skb(0, GFP_ATOMIC);
-
-	bh_lock_sock(sk);
-	if (skb) {
-		struct llc_conn_state_ev *ev = llc_conn_ev(skb);
-
-		skb->sk  = sk;
-		ev->type = LLC_CONN_EV_TYPE_ACK_TMR;
-		llc_process_tmr_ev(sk, skb);
-	}
-	bh_unlock_sock(sk);
+	llc_conn_tmr_common_cb(timeout_data, LLC_CONN_EV_TYPE_ACK_TMR);
 }
 
 void llc_conn_rej_tmr_cb(unsigned long timeout_data)
 {
-	struct sock *sk = (struct sock *)timeout_data;
-	struct sk_buff *skb = alloc_skb(0, GFP_ATOMIC);
-
-	bh_lock_sock(sk);
-	if (skb) {
-		struct llc_conn_state_ev *ev = llc_conn_ev(skb);
-
-		skb->sk  = sk;
-		ev->type = LLC_CONN_EV_TYPE_REJ_TMR;
-		llc_process_tmr_ev(sk, skb);
-	}
-	bh_unlock_sock(sk);
+	llc_conn_tmr_common_cb(timeout_data, LLC_CONN_EV_TYPE_REJ_TMR);
 }
 
 int llc_conn_ac_rst_vs(struct sock *sk, struct sk_buff *skb)
diff --git a/net/llc/llc_c_ev.c b/net/llc/llc_c_ev.c
index d5bdb53a348..c5deda24661 100644
--- a/net/llc/llc_c_ev.c
+++ b/net/llc/llc_c_ev.c
@@ -37,6 +37,7 @@
 #include <net/llc_conn.h>
 #include <net/llc_sap.h>
 #include <net/sock.h>
+#include <net/llc_c_ac.h>
 #include <net/llc_c_ev.h>
 #include <net/llc_pdu.h>
 
@@ -46,8 +47,6 @@
 #define dprintk(args...)
 #endif
 
-extern u16 llc_circular_between(u8 a, u8 b, u8 c);
-
 /**
  *	llc_util_ns_inside_rx_window - check if sequence number is in rx window
  *	@ns: sequence number of received pdu.
@@ -99,7 +98,7 @@ out:
 
 int llc_conn_ev_conn_req(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_conn_state_ev *ev = llc_conn_ev(skb);
+	const struct llc_conn_state_ev *ev = llc_conn_ev(skb);
 
 	return ev->prim == LLC_CONN_PRIM &&
 	       ev->prim_type == LLC_PRIM_TYPE_REQ ? 0 : 1;
@@ -107,7 +106,7 @@ int llc_conn_ev_conn_req(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_data_req(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_conn_state_ev *ev = llc_conn_ev(skb);
+	const struct llc_conn_state_ev *ev = llc_conn_ev(skb);
 
 	return ev->prim == LLC_DATA_PRIM &&
 	       ev->prim_type == LLC_PRIM_TYPE_REQ ? 0 : 1;
@@ -115,7 +114,7 @@ int llc_conn_ev_data_req(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_disc_req(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_conn_state_ev *ev = llc_conn_ev(skb);
+	const struct llc_conn_state_ev *ev = llc_conn_ev(skb);
 
 	return ev->prim == LLC_DISC_PRIM &&
 	       ev->prim_type == LLC_PRIM_TYPE_REQ ? 0 : 1;
@@ -123,7 +122,7 @@ int llc_conn_ev_disc_req(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rst_req(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_conn_state_ev *ev = llc_conn_ev(skb);
+	const struct llc_conn_state_ev *ev = llc_conn_ev(skb);
 
 	return ev->prim == LLC_RESET_PRIM &&
 	       ev->prim_type == LLC_PRIM_TYPE_REQ ? 0 : 1;
@@ -131,7 +130,7 @@ int llc_conn_ev_rst_req(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_local_busy_detected(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_conn_state_ev *ev = llc_conn_ev(skb);
+	const struct llc_conn_state_ev *ev = llc_conn_ev(skb);
 
 	return ev->type == LLC_CONN_EV_TYPE_SIMPLE &&
 	       ev->prim_type == LLC_CONN_EV_LOCAL_BUSY_DETECTED ? 0 : 1;
@@ -139,7 +138,7 @@ int llc_conn_ev_local_busy_detected(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_local_busy_cleared(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_conn_state_ev *ev = llc_conn_ev(skb);
+	const struct llc_conn_state_ev *ev = llc_conn_ev(skb);
 
 	return ev->type == LLC_CONN_EV_TYPE_SIMPLE &&
 	       ev->prim_type == LLC_CONN_EV_LOCAL_BUSY_CLEARED ? 0 : 1;
@@ -152,7 +151,7 @@ int llc_conn_ev_rx_bad_pdu(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_disc_cmd_pbit_set_x(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_un *pdu = llc_pdu_un_hdr(skb);
+	const struct llc_pdu_un *pdu = llc_pdu_un_hdr(skb);
 
 	return LLC_PDU_IS_CMD(pdu) && LLC_PDU_TYPE_IS_U(pdu) &&
 	       LLC_U_PDU_CMD(pdu) == LLC_2_PDU_CMD_DISC ? 0 : 1;
@@ -160,7 +159,7 @@ int llc_conn_ev_rx_disc_cmd_pbit_set_x(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_dm_rsp_fbit_set_x(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_un *pdu = llc_pdu_un_hdr(skb);
+	const struct llc_pdu_un *pdu = llc_pdu_un_hdr(skb);
 
 	return LLC_PDU_IS_RSP(pdu) && LLC_PDU_TYPE_IS_U(pdu) &&
 	       LLC_U_PDU_RSP(pdu) == LLC_2_PDU_RSP_DM ? 0 : 1;
@@ -168,7 +167,7 @@ int llc_conn_ev_rx_dm_rsp_fbit_set_x(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_frmr_rsp_fbit_set_x(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_un *pdu = llc_pdu_un_hdr(skb);
+	const struct llc_pdu_un *pdu = llc_pdu_un_hdr(skb);
 
 	return LLC_PDU_IS_RSP(pdu) && LLC_PDU_TYPE_IS_U(pdu) &&
 	       LLC_U_PDU_RSP(pdu) == LLC_2_PDU_RSP_FRMR ? 0 : 1;
@@ -176,7 +175,7 @@ int llc_conn_ev_rx_frmr_rsp_fbit_set_x(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_i_cmd_pbit_set_0(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 
 	return llc_conn_space(sk, skb) &&
 	       LLC_PDU_IS_CMD(pdu) && LLC_PDU_TYPE_IS_I(pdu) &&
@@ -186,7 +185,7 @@ int llc_conn_ev_rx_i_cmd_pbit_set_0(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_i_cmd_pbit_set_1(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 
 	return llc_conn_space(sk, skb) &&
 	       LLC_PDU_IS_CMD(pdu) && LLC_PDU_TYPE_IS_I(pdu) &&
@@ -197,9 +196,9 @@ int llc_conn_ev_rx_i_cmd_pbit_set_1(struct sock *sk, struct sk_buff *skb)
 int llc_conn_ev_rx_i_cmd_pbit_set_0_unexpd_ns(struct sock *sk,
 					      struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
-	u8 vr = llc_sk(sk)->vR;
-	u8 ns = LLC_I_GET_NS(pdu);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const u8 vr = llc_sk(sk)->vR;
+	const u8 ns = LLC_I_GET_NS(pdu);
 
 	return LLC_PDU_IS_CMD(pdu) && LLC_PDU_TYPE_IS_I(pdu) &&
 	       LLC_I_PF_IS_0(pdu) && ns != vr &&
@@ -209,9 +208,9 @@ int llc_conn_ev_rx_i_cmd_pbit_set_0_unexpd_ns(struct sock *sk,
 int llc_conn_ev_rx_i_cmd_pbit_set_1_unexpd_ns(struct sock *sk,
 					      struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
-	u8 vr = llc_sk(sk)->vR;
-	u8 ns = LLC_I_GET_NS(pdu);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const u8 vr = llc_sk(sk)->vR;
+	const u8 ns = LLC_I_GET_NS(pdu);
 
 	return LLC_PDU_IS_CMD(pdu) && LLC_PDU_TYPE_IS_I(pdu) &&
 	       LLC_I_PF_IS_1(pdu) && ns != vr &&
@@ -221,10 +220,11 @@ int llc_conn_ev_rx_i_cmd_pbit_set_1_unexpd_ns(struct sock *sk,
 int llc_conn_ev_rx_i_cmd_pbit_set_x_inval_ns(struct sock *sk,
 					     struct sk_buff *skb)
 {
-	struct llc_pdu_sn * pdu = llc_pdu_sn_hdr(skb);
-	u8 vr = llc_sk(sk)->vR;
-	u8 ns = LLC_I_GET_NS(pdu);
-	u16 rc = LLC_PDU_IS_CMD(pdu) && LLC_PDU_TYPE_IS_I(pdu) && ns != vr &&
+	const struct llc_pdu_sn * pdu = llc_pdu_sn_hdr(skb);
+	const u8 vr = llc_sk(sk)->vR;
+	const u8 ns = LLC_I_GET_NS(pdu);
+	const u16 rc = LLC_PDU_IS_CMD(pdu) && LLC_PDU_TYPE_IS_I(pdu) &&
+		ns != vr &&
 		 llc_util_ns_inside_rx_window(ns, vr, llc_sk(sk)->rw) ? 0 : 1;
 	if (!rc)
 		dprintk("%s: matched, state=%d, ns=%d, vr=%d\n",
@@ -234,7 +234,7 @@ int llc_conn_ev_rx_i_cmd_pbit_set_x_inval_ns(struct sock *sk,
 
 int llc_conn_ev_rx_i_rsp_fbit_set_0(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 
 	return llc_conn_space(sk, skb) &&
 	       LLC_PDU_IS_RSP(pdu) && LLC_PDU_TYPE_IS_I(pdu) &&
@@ -244,7 +244,7 @@ int llc_conn_ev_rx_i_rsp_fbit_set_0(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_i_rsp_fbit_set_1(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 
 	return LLC_PDU_IS_RSP(pdu) && LLC_PDU_TYPE_IS_I(pdu) &&
 	       LLC_I_PF_IS_1(pdu) &&
@@ -253,7 +253,7 @@ int llc_conn_ev_rx_i_rsp_fbit_set_1(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_i_rsp_fbit_set_x(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 
 	return llc_conn_space(sk, skb) &&
 	       LLC_PDU_IS_RSP(pdu) && LLC_PDU_TYPE_IS_I(pdu) &&
@@ -263,9 +263,9 @@ int llc_conn_ev_rx_i_rsp_fbit_set_x(struct sock *sk, struct sk_buff *skb)
 int llc_conn_ev_rx_i_rsp_fbit_set_0_unexpd_ns(struct sock *sk,
 					      struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
-	u8 vr = llc_sk(sk)->vR;
-	u8 ns = LLC_I_GET_NS(pdu);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const u8 vr = llc_sk(sk)->vR;
+	const u8 ns = LLC_I_GET_NS(pdu);
 
 	return LLC_PDU_IS_RSP(pdu) && LLC_PDU_TYPE_IS_I(pdu) &&
 	       LLC_I_PF_IS_0(pdu) && ns != vr &&
@@ -275,9 +275,9 @@ int llc_conn_ev_rx_i_rsp_fbit_set_0_unexpd_ns(struct sock *sk,
 int llc_conn_ev_rx_i_rsp_fbit_set_1_unexpd_ns(struct sock *sk,
 					      struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
-	u8 vr = llc_sk(sk)->vR;
-	u8 ns = LLC_I_GET_NS(pdu);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const u8 vr = llc_sk(sk)->vR;
+	const u8 ns = LLC_I_GET_NS(pdu);
 
 	return LLC_PDU_IS_RSP(pdu) && LLC_PDU_TYPE_IS_I(pdu) &&
 	       LLC_I_PF_IS_1(pdu) && ns != vr &&
@@ -287,9 +287,9 @@ int llc_conn_ev_rx_i_rsp_fbit_set_1_unexpd_ns(struct sock *sk,
 int llc_conn_ev_rx_i_rsp_fbit_set_x_unexpd_ns(struct sock *sk,
 					      struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
-	u8 vr = llc_sk(sk)->vR;
-	u8 ns = LLC_I_GET_NS(pdu);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const u8 vr = llc_sk(sk)->vR;
+	const u8 ns = LLC_I_GET_NS(pdu);
 
 	return LLC_PDU_IS_RSP(pdu) && LLC_PDU_TYPE_IS_I(pdu) && ns != vr &&
 	       !llc_util_ns_inside_rx_window(ns, vr, llc_sk(sk)->rw) ? 0 : 1;
@@ -298,10 +298,11 @@ int llc_conn_ev_rx_i_rsp_fbit_set_x_unexpd_ns(struct sock *sk,
 int llc_conn_ev_rx_i_rsp_fbit_set_x_inval_ns(struct sock *sk,
 					     struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
-	u8 vr = llc_sk(sk)->vR;
-	u8 ns = LLC_I_GET_NS(pdu);
-	u16 rc = LLC_PDU_IS_RSP(pdu) && LLC_PDU_TYPE_IS_I(pdu) && ns != vr &&
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const u8 vr = llc_sk(sk)->vR;
+	const u8 ns = LLC_I_GET_NS(pdu);
+	const u16 rc = LLC_PDU_IS_RSP(pdu) && LLC_PDU_TYPE_IS_I(pdu) &&
+		ns != vr &&
 		 llc_util_ns_inside_rx_window(ns, vr, llc_sk(sk)->rw) ? 0 : 1;
 	if (!rc)
 		dprintk("%s: matched, state=%d, ns=%d, vr=%d\n",
@@ -311,7 +312,7 @@ int llc_conn_ev_rx_i_rsp_fbit_set_x_inval_ns(struct sock *sk,
 
 int llc_conn_ev_rx_rej_cmd_pbit_set_0(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 
 	return LLC_PDU_IS_CMD(pdu) && LLC_PDU_TYPE_IS_S(pdu) &&
 	       LLC_S_PF_IS_0(pdu) &&
@@ -320,7 +321,7 @@ int llc_conn_ev_rx_rej_cmd_pbit_set_0(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_rej_cmd_pbit_set_1(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 
 	return LLC_PDU_IS_CMD(pdu) && LLC_PDU_TYPE_IS_S(pdu) &&
 	       LLC_S_PF_IS_1(pdu) &&
@@ -329,7 +330,7 @@ int llc_conn_ev_rx_rej_cmd_pbit_set_1(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_rej_rsp_fbit_set_0(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 
 	return LLC_PDU_IS_RSP(pdu) && LLC_PDU_TYPE_IS_S(pdu) &&
 	       LLC_S_PF_IS_0(pdu) &&
@@ -338,7 +339,7 @@ int llc_conn_ev_rx_rej_rsp_fbit_set_0(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_rej_rsp_fbit_set_1(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 
 	return LLC_PDU_IS_RSP(pdu) && LLC_PDU_TYPE_IS_S(pdu) &&
 	       LLC_S_PF_IS_1(pdu) &&
@@ -347,7 +348,7 @@ int llc_conn_ev_rx_rej_rsp_fbit_set_1(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_rej_rsp_fbit_set_x(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_un *pdu = llc_pdu_un_hdr(skb);
+	const struct llc_pdu_un *pdu = llc_pdu_un_hdr(skb);
 
 	return LLC_PDU_IS_RSP(pdu) && LLC_PDU_TYPE_IS_S(pdu) &&
 	       LLC_S_PDU_RSP(pdu) == LLC_2_PDU_RSP_REJ ? 0 : 1;
@@ -355,7 +356,7 @@ int llc_conn_ev_rx_rej_rsp_fbit_set_x(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_rnr_cmd_pbit_set_0(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 
 	return LLC_PDU_IS_CMD(pdu) && LLC_PDU_TYPE_IS_S(pdu) &&
 	       LLC_S_PF_IS_0(pdu) &&
@@ -364,7 +365,7 @@ int llc_conn_ev_rx_rnr_cmd_pbit_set_0(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_rnr_cmd_pbit_set_1(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 
 	return LLC_PDU_IS_CMD(pdu) && LLC_PDU_TYPE_IS_S(pdu) &&
 	       LLC_S_PF_IS_1(pdu) &&
@@ -373,7 +374,7 @@ int llc_conn_ev_rx_rnr_cmd_pbit_set_1(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_rnr_rsp_fbit_set_0(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 
 	return LLC_PDU_IS_RSP(pdu) && LLC_PDU_TYPE_IS_S(pdu) &&
 	       LLC_S_PF_IS_0(pdu) &&
@@ -382,7 +383,7 @@ int llc_conn_ev_rx_rnr_rsp_fbit_set_0(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_rnr_rsp_fbit_set_1(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 
 	return LLC_PDU_IS_RSP(pdu) && LLC_PDU_TYPE_IS_S(pdu) &&
 	       LLC_S_PF_IS_1(pdu) &&
@@ -391,7 +392,7 @@ int llc_conn_ev_rx_rnr_rsp_fbit_set_1(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_rr_cmd_pbit_set_0(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 
 	return LLC_PDU_IS_CMD(pdu) && LLC_PDU_TYPE_IS_S(pdu) &&
 	       LLC_S_PF_IS_0(pdu) &&
@@ -400,7 +401,7 @@ int llc_conn_ev_rx_rr_cmd_pbit_set_0(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_rr_cmd_pbit_set_1(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 
 	return LLC_PDU_IS_CMD(pdu) && LLC_PDU_TYPE_IS_S(pdu) &&
 	       LLC_S_PF_IS_1(pdu) &&
@@ -409,7 +410,7 @@ int llc_conn_ev_rx_rr_cmd_pbit_set_1(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_rr_rsp_fbit_set_0(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 
 	return llc_conn_space(sk, skb) &&
 	       LLC_PDU_IS_RSP(pdu) && LLC_PDU_TYPE_IS_S(pdu) &&
@@ -419,7 +420,7 @@ int llc_conn_ev_rx_rr_rsp_fbit_set_0(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_rr_rsp_fbit_set_1(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 
 	return llc_conn_space(sk, skb) &&
 	       LLC_PDU_IS_RSP(pdu) && LLC_PDU_TYPE_IS_S(pdu) &&
@@ -429,7 +430,7 @@ int llc_conn_ev_rx_rr_rsp_fbit_set_1(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_rx_sabme_cmd_pbit_set_x(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_pdu_un *pdu = llc_pdu_un_hdr(skb);
+	const struct llc_pdu_un *pdu = llc_pdu_un_hdr(skb);
 
 	return LLC_PDU_IS_CMD(pdu) && LLC_PDU_TYPE_IS_U(pdu) &&
 	       LLC_U_PDU_CMD(pdu) == LLC_2_PDU_CMD_SABME ? 0 : 1;
@@ -446,7 +447,7 @@ int llc_conn_ev_rx_ua_rsp_fbit_set_x(struct sock *sk, struct sk_buff *skb)
 int llc_conn_ev_rx_xxx_cmd_pbit_set_1(struct sock *sk, struct sk_buff *skb)
 {
 	u16 rc = 1;
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
 
 	if (LLC_PDU_IS_CMD(pdu)) {
 		if (LLC_PDU_TYPE_IS_I(pdu) || LLC_PDU_TYPE_IS_S(pdu)) {
@@ -461,7 +462,7 @@ int llc_conn_ev_rx_xxx_cmd_pbit_set_1(struct sock *sk, struct sk_buff *skb)
 int llc_conn_ev_rx_xxx_cmd_pbit_set_x(struct sock *sk, struct sk_buff *skb)
 {
 	u16 rc = 1;
-	struct llc_pdu_un *pdu = llc_pdu_un_hdr(skb);
+	const struct llc_pdu_un *pdu = llc_pdu_un_hdr(skb);
 
 	if (LLC_PDU_IS_CMD(pdu)) {
 		if (LLC_PDU_TYPE_IS_I(pdu) || LLC_PDU_TYPE_IS_S(pdu))
@@ -477,32 +478,10 @@ int llc_conn_ev_rx_xxx_cmd_pbit_set_x(struct sock *sk, struct sk_buff *skb)
 	return rc;
 }
 
-int llc_conn_ev_rx_xxx_rsp_fbit_set_1(struct sock *sk, struct sk_buff *skb)
-{
-	u16 rc = 1;
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
-
-	if (LLC_PDU_IS_RSP(pdu)) {
-		if (LLC_PDU_TYPE_IS_I(pdu) || LLC_PDU_TYPE_IS_S(pdu)) {
-			if (LLC_I_PF_IS_1(pdu))
-				rc = 0;
-		} else if (LLC_PDU_TYPE_IS_U(pdu))
-			switch (LLC_U_PDU_RSP(pdu)) {
-			case LLC_2_PDU_RSP_UA:
-			case LLC_2_PDU_RSP_DM:
-			case LLC_2_PDU_RSP_FRMR:
-				if (LLC_U_PF_IS_1(pdu))
-					rc = 0;
-				break;
-			}
-	}
-	return rc;
-}
-
 int llc_conn_ev_rx_xxx_rsp_fbit_set_x(struct sock *sk, struct sk_buff *skb)
 {
 	u16 rc = 1;
-	struct llc_pdu_un *pdu = llc_pdu_un_hdr(skb);
+	const struct llc_pdu_un *pdu = llc_pdu_un_hdr(skb);
 
 	if (LLC_PDU_IS_RSP(pdu)) {
 		if (LLC_PDU_TYPE_IS_I(pdu) || LLC_PDU_TYPE_IS_S(pdu))
@@ -524,9 +503,9 @@ int llc_conn_ev_rx_zzz_cmd_pbit_set_x_inval_nr(struct sock *sk,
 					       struct sk_buff *skb)
 {
 	u16 rc = 1;
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
-	u8 vs = llc_sk(sk)->vS;
-	u8 nr = LLC_I_GET_NR(pdu);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const u8 vs = llc_sk(sk)->vS;
+	const u8 nr = LLC_I_GET_NR(pdu);
 
 	if (LLC_PDU_IS_CMD(pdu) &&
 	    (LLC_PDU_TYPE_IS_I(pdu) || LLC_PDU_TYPE_IS_S(pdu)) &&
@@ -542,9 +521,9 @@ int llc_conn_ev_rx_zzz_rsp_fbit_set_x_inval_nr(struct sock *sk,
 					       struct sk_buff *skb)
 {
 	u16 rc = 1;
-	struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
-	u8 vs = llc_sk(sk)->vS;
-	u8 nr = LLC_I_GET_NR(pdu);
+	const struct llc_pdu_sn *pdu = llc_pdu_sn_hdr(skb);
+	const u8 vs = llc_sk(sk)->vS;
+	const u8 nr = LLC_I_GET_NR(pdu);
 
 	if (LLC_PDU_IS_RSP(pdu) &&
 	    (LLC_PDU_TYPE_IS_I(pdu) || LLC_PDU_TYPE_IS_S(pdu)) &&
@@ -563,28 +542,28 @@ int llc_conn_ev_rx_any_frame(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_p_tmr_exp(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_conn_state_ev *ev = llc_conn_ev(skb);
+	const struct llc_conn_state_ev *ev = llc_conn_ev(skb);
 
 	return ev->type != LLC_CONN_EV_TYPE_P_TMR;
 }
 
 int llc_conn_ev_ack_tmr_exp(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_conn_state_ev *ev = llc_conn_ev(skb);
+	const struct llc_conn_state_ev *ev = llc_conn_ev(skb);
 
 	return ev->type != LLC_CONN_EV_TYPE_ACK_TMR;
 }
 
 int llc_conn_ev_rej_tmr_exp(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_conn_state_ev *ev = llc_conn_ev(skb);
+	const struct llc_conn_state_ev *ev = llc_conn_ev(skb);
 
 	return ev->type != LLC_CONN_EV_TYPE_REJ_TMR;
 }
 
 int llc_conn_ev_busy_tmr_exp(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_conn_state_ev *ev = llc_conn_ev(skb);
+	const struct llc_conn_state_ev *ev = llc_conn_ev(skb);
 
 	return ev->type != LLC_CONN_EV_TYPE_BUSY_TMR;
 }
@@ -596,7 +575,7 @@ int llc_conn_ev_init_p_f_cycle(struct sock *sk, struct sk_buff *skb)
 
 int llc_conn_ev_tx_buffer_full(struct sock *sk, struct sk_buff *skb)
 {
-	struct llc_conn_state_ev *ev = llc_conn_ev(skb);
+	const struct llc_conn_state_ev *ev = llc_conn_ev(skb);
 
 	return ev->type == LLC_CONN_EV_TYPE_SIMPLE &&
 	       ev->prim_type == LLC_CONN_EV_TX_BUFF_FULL ? 0 : 1;
diff --git a/net/llc/llc_conn.c b/net/llc/llc_conn.c
index 4c644bc70ea..c761c15da42 100644
--- a/net/llc/llc_conn.c
+++ b/net/llc/llc_conn.c
@@ -40,6 +40,11 @@ static struct llc_conn_state_trans *llc_qualify_conn_ev(struct sock *sk,
 /* Offset table on connection states transition diagram */
 static int llc_offset_table[NBR_CONN_STATES][NBR_CONN_EV];
 
+int sysctl_llc2_ack_timeout = LLC2_ACK_TIME * HZ;
+int sysctl_llc2_p_timeout = LLC2_P_TIME * HZ;
+int sysctl_llc2_rej_timeout = LLC2_REJ_TIME * HZ;
+int sysctl_llc2_busy_timeout = LLC2_BUSY_TIME * HZ;
+
 /**
  *	llc_conn_state_process - sends event to connection state machine
  *	@sk: connection
@@ -53,7 +58,7 @@ static int llc_offset_table[NBR_CONN_STATES][NBR_CONN_EV];
 int llc_conn_state_process(struct sock *sk, struct sk_buff *skb)
 {
 	int rc;
-	struct llc_sock *llc = llc_sk(sk);
+	struct llc_sock *llc = llc_sk(skb->sk);
 	struct llc_conn_state_ev *ev = llc_conn_ev(skb);
 
 	/*
@@ -63,13 +68,16 @@ int llc_conn_state_process(struct sock *sk, struct sk_buff *skb)
 	 */
 	skb_get(skb);
 	ev->ind_prim = ev->cfm_prim = 0;
-	rc = llc_conn_service(sk, skb); /* sending event to state machine */
-	if (rc) {
+	/*
+	 * Send event to state machine
+	 */
+	rc = llc_conn_service(skb->sk, skb);
+	if (unlikely(rc != 0)) {
 		printk(KERN_ERR "%s: llc_conn_service failed\n", __FUNCTION__);
 		goto out_kfree_skb;
 	}
 
-	if (!ev->ind_prim && !ev->cfm_prim) {
+	if (unlikely(!ev->ind_prim && !ev->cfm_prim)) {
 		/* indicate or confirm not required */
 		/* XXX this is not very pretty, perhaps we should store
 		 * XXX indicate/confirm-needed state in the llc_conn_state_ev
@@ -80,13 +88,13 @@ int llc_conn_state_process(struct sock *sk, struct sk_buff *skb)
 		goto out_skb_put;
 	}
 
-	if (ev->ind_prim && ev->cfm_prim) /* Paranoia */
+	if (unlikely(ev->ind_prim && ev->cfm_prim)) /* Paranoia */
 		skb_get(skb);
 
 	switch (ev->ind_prim) {
 	case LLC_DATA_PRIM:
-		llc_save_primitive(skb, LLC_DATA_PRIM);
-		if (sock_queue_rcv_skb(sk, skb)) {
+		llc_save_primitive(sk, skb, LLC_DATA_PRIM);
+		if (unlikely(sock_queue_rcv_skb(sk, skb))) {
 			/*
 			 * shouldn't happen
 			 */
@@ -95,13 +103,14 @@ int llc_conn_state_process(struct sock *sk, struct sk_buff *skb)
 			kfree_skb(skb);
 		}
 		break;
-	case LLC_CONN_PRIM: {
-		struct sock *parent = skb->sk;
-
-		skb->sk = sk;
-		skb_queue_tail(&parent->sk_receive_queue, skb);
-		sk->sk_state_change(parent);
-	}
+	case LLC_CONN_PRIM:
+		/*
+		 * Can't be sock_queue_rcv_skb, because we have to leave the
+		 * skb->sk pointing to the newly created struct sock in
+		 * llc_conn_handler. -acme
+		 */
+		skb_queue_tail(&sk->sk_receive_queue, skb);
+		sk->sk_state_change(sk);
 		break;
 	case LLC_DISC_PRIM:
 		sock_hold(sk);
@@ -111,8 +120,8 @@ int llc_conn_state_process(struct sock *sk, struct sk_buff *skb)
 			sk->sk_socket->state  = SS_UNCONNECTED;
 			sk->sk_state          = TCP_CLOSE;
 			if (!sock_flag(sk, SOCK_DEAD)) {
-				sk->sk_state_change(sk);
 				sock_set_flag(sk, SOCK_DEAD);
+				sk->sk_state_change(sk);
 			}
 		}
 		kfree_skb(skb);
@@ -465,7 +474,7 @@ static int llc_exec_conn_trans_actions(struct sock *sk,
 }
 
 /**
- *	llc_lookup_established - Finds connection for the remote/local sap/mac
+ *	__llc_lookup_established - Finds connection for the remote/local sap/mac
  *	@sap: SAP
  *	@daddr: address of remote LLC (MAC + SAP)
  *	@laddr: address of local LLC (MAC + SAP)
@@ -473,14 +482,16 @@ static int llc_exec_conn_trans_actions(struct sock *sk,
  *	Search connection list of the SAP and finds connection using the remote
  *	mac, remote sap, local mac, and local sap. Returns pointer for
  *	connection found, %NULL otherwise.
+ *	Caller has to make sure local_bh is disabled.
  */
-struct sock *llc_lookup_established(struct llc_sap *sap, struct llc_addr *daddr,
-				    struct llc_addr *laddr)
+static struct sock *__llc_lookup_established(struct llc_sap *sap,
+					     struct llc_addr *daddr,
+					     struct llc_addr *laddr)
 {
 	struct sock *rc;
 	struct hlist_node *node;
 
-	read_lock_bh(&sap->sk_list.lock);
+	read_lock(&sap->sk_list.lock);
 	sk_for_each(rc, node, &sap->sk_list.list) {
 		struct llc_sock *llc = llc_sk(rc);
 
@@ -494,10 +505,22 @@ struct sock *llc_lookup_established(struct llc_sap *sap, struct llc_addr *daddr,
 	}
 	rc = NULL;
 found:
-	read_unlock_bh(&sap->sk_list.lock);
+	read_unlock(&sap->sk_list.lock);
 	return rc;
 }
 
+struct sock *llc_lookup_established(struct llc_sap *sap,
+				    struct llc_addr *daddr,
+				    struct llc_addr *laddr)
+{
+	struct sock *sk;
+
+	local_bh_disable();
+	sk = __llc_lookup_established(sap, daddr, laddr);
+	local_bh_enable();
+	return sk;
+}
+
 /**
  *	llc_lookup_listener - Finds listener for local MAC + SAP
  *	@sap: SAP
@@ -506,6 +529,7 @@ found:
  *	Search connection list of the SAP and finds connection listening on
  *	local mac, and local sap. Returns pointer for parent socket found,
  *	%NULL otherwise.
+ *	Caller has to make sure local_bh is disabled.
  */
 static struct sock *llc_lookup_listener(struct llc_sap *sap,
 					struct llc_addr *laddr)
@@ -513,7 +537,7 @@ static struct sock *llc_lookup_listener(struct llc_sap *sap,
 	struct sock *rc;
 	struct hlist_node *node;
 
-	read_lock_bh(&sap->sk_list.lock);
+	read_lock(&sap->sk_list.lock);
 	sk_for_each(rc, node, &sap->sk_list.list) {
 		struct llc_sock *llc = llc_sk(rc);
 
@@ -527,10 +551,19 @@ static struct sock *llc_lookup_listener(struct llc_sap *sap,
 	}
 	rc = NULL;
 found:
-	read_unlock_bh(&sap->sk_list.lock);
+	read_unlock(&sap->sk_list.lock);
 	return rc;
 }
 
+static struct sock *__llc_lookup(struct llc_sap *sap,
+				 struct llc_addr *daddr,
+				 struct llc_addr *laddr)
+{
+	struct sock *sk = __llc_lookup_established(sap, daddr, laddr);
+
+	return sk ? : llc_lookup_listener(sap, laddr);
+}
+
 /**
  *	llc_data_accept_state - designates if in this state data can be sent.
  *	@state: state of connection.
@@ -544,14 +577,14 @@ u8 llc_data_accept_state(u8 state)
 }
 
 /**
- *	find_next_offset - finds offset for next category of transitions
+ *	llc_find_next_offset - finds offset for next category of transitions
  *	@state: state table.
  *	@offset: start offset.
  *
  *	Finds offset of next category of transitions in transition table.
  *	Returns the start index of next category.
  */
-static u16 find_next_offset(struct llc_conn_state *state, u16 offset)
+static u16 __init llc_find_next_offset(struct llc_conn_state *state, u16 offset)
 {
 	u16 cnt = 0;
 	struct llc_conn_state_trans **next_trans;
@@ -578,8 +611,8 @@ void __init llc_build_offset_table(void)
 		next_offset = 0;
 		for (ev_type = 0; ev_type < NBR_CONN_EV; ev_type++) {
 			llc_offset_table[state][ev_type] = next_offset;
-			next_offset += find_next_offset(curr_state,
-							next_offset) + 1;
+			next_offset += llc_find_next_offset(curr_state,
+							    next_offset) + 1;
 		}
 	}
 }
@@ -623,6 +656,7 @@ static int llc_find_offset(int state, int ev_type)
  */
 void llc_sap_add_socket(struct llc_sap *sap, struct sock *sk)
 {
+	llc_sap_hold(sap);
 	write_lock_bh(&sap->sk_list.lock);
 	llc_sk(sk)->sap = sap;
 	sk_add_node(sk, &sap->sk_list.list);
@@ -642,6 +676,7 @@ void llc_sap_remove_socket(struct llc_sap *sap, struct sock *sk)
 	write_lock_bh(&sap->sk_list.lock);
 	sk_del_node_init(sk);
 	write_unlock_bh(&sap->sk_list.lock);
+	llc_sap_put(sap);
 }
 
 /**
@@ -654,15 +689,34 @@ void llc_sap_remove_socket(struct llc_sap *sap, struct sock *sk)
 static int llc_conn_rcv(struct sock* sk, struct sk_buff *skb)
 {
 	struct llc_conn_state_ev *ev = llc_conn_ev(skb);
-	struct llc_sock *llc = llc_sk(sk);
 
-	if (!llc->dev)
-		llc->dev = skb->dev;
 	ev->type   = LLC_CONN_EV_TYPE_PDU;
 	ev->reason = 0;
 	return llc_conn_state_process(sk, skb);
 }
 
+static struct sock *llc_create_incoming_sock(struct sock *sk,
+					     struct net_device *dev,
+					     struct llc_addr *saddr,
+					     struct llc_addr *daddr)
+{
+	struct sock *newsk = llc_sk_alloc(sk->sk_family, GFP_ATOMIC,
+					  sk->sk_prot);
+	struct llc_sock *newllc, *llc = llc_sk(sk);
+
+	if (!newsk)
+		goto out;
+	newllc = llc_sk(newsk);
+	memcpy(&newllc->laddr, daddr, sizeof(newllc->laddr));
+	memcpy(&newllc->daddr, saddr, sizeof(newllc->daddr));
+	newllc->dev = dev;
+	dev_hold(dev);
+	llc_sap_add_socket(llc->sap, newsk);
+	llc_sap_hold(llc->sap);
+out:
+	return newsk;
+}
+
 void llc_conn_handler(struct llc_sap *sap, struct sk_buff *skb)
 {
 	struct llc_addr saddr, daddr;
@@ -673,35 +727,35 @@ void llc_conn_handler(struct llc_sap *sap, struct sk_buff *skb)
 	llc_pdu_decode_da(skb, daddr.mac);
 	llc_pdu_decode_dsap(skb, &daddr.lsap);
 
-	sk = llc_lookup_established(sap, &saddr, &daddr);
-	if (!sk) {
+	sk = __llc_lookup(sap, &saddr, &daddr);
+	if (!sk)
+		goto drop;
+
+	bh_lock_sock(sk);
+	/*
+	 * This has to be done here and not at the upper layer ->accept
+	 * method because of the way the PROCOM state machine works:
+	 * it needs to set several state variables (see, for instance,
+	 * llc_adm_actions_2 in net/llc/llc_c_st.c) and send a packet to
+	 * the originator of the new connection, and this state has to be
+	 * in the newly created struct sock private area. -acme
+	 */
+	if (unlikely(sk->sk_state == TCP_LISTEN)) {
+		struct sock *newsk = llc_create_incoming_sock(sk, skb->dev,
+							      &saddr, &daddr);
+		if (!newsk)
+			goto drop_unlock;
+		skb_set_owner_r(skb, newsk);
+	} else {
 		/*
-		 * Didn't find an active connection; verify if there
-		 * is a listening socket for this llc addr
+		 * Can't be skb_set_owner_r, this will be done at the
+		 * llc_conn_state_process function, later on, when we will use
+		 * skb_queue_rcv_skb to send it to upper layers, this is
+		 * another trick required to cope with how the PROCOM state
+		 * machine works. -acme
 		 */
-		struct llc_sock *llc;
-		struct sock *parent = llc_lookup_listener(sap, &daddr);
-
-		if (!parent) {
-			dprintk("llc_lookup_listener failed!\n");
-			goto drop;
-		}
-
-		sk = llc_sk_alloc(parent->sk_family, GFP_ATOMIC, parent->sk_prot);
-		if (!sk) {
-			sock_put(parent);
-			goto drop;
-		}
-		llc = llc_sk(sk);
-		memcpy(&llc->laddr, &daddr, sizeof(llc->laddr));
-		memcpy(&llc->daddr, &saddr, sizeof(llc->daddr));
-		llc_sap_add_socket(sap, sk);
-		sock_hold(sk);
-		sock_put(parent);
-		skb->sk = parent;
-	} else
 		skb->sk = sk;
-	bh_lock_sock(sk);
+	}
 	if (!sock_owned_by_user(sk))
 		llc_conn_rcv(sk, skb);
 	else {
@@ -709,11 +763,16 @@ void llc_conn_handler(struct llc_sap *sap, struct sk_buff *skb)
 		llc_set_backlog_type(skb, LLC_PACKET);
 		sk_add_backlog(sk, skb);
 	}
+out:
 	bh_unlock_sock(sk);
 	sock_put(sk);
 	return;
 drop:
 	kfree_skb(skb);
+	return;
+drop_unlock:
+	kfree_skb(skb);
+	goto out;
 }
 
 #undef LLC_REFCNT_DEBUG
@@ -722,32 +781,6 @@ static atomic_t llc_sock_nr;
 #endif
 
 /**
- *	llc_release_sockets - releases all sockets in a sap
- *	@sap: sap to release its sockets
- *
- *	Releases all connections of a sap. Returns 0 if all actions complete
- *	successfully, nonzero otherwise
- */
-int llc_release_sockets(struct llc_sap *sap)
-{
-	int rc = 0;
-	struct sock *sk;
-	struct hlist_node *node;
-
-	write_lock_bh(&sap->sk_list.lock);
-
-	sk_for_each(sk, node, &sap->sk_list.list) {
-		llc_sk(sk)->state = LLC_CONN_STATE_TEMP;
-
-		if (llc_send_disc(sk))
-			rc = 1;
-	}
-
-	write_unlock_bh(&sap->sk_list.lock);
-	return rc;
-}
-
-/**
  *	llc_backlog_rcv - Processes rx frames and expired timers.
  *	@sk: LLC sock (p8022 connection)
  *	@skb: queued rx frame or event
@@ -762,14 +795,14 @@ static int llc_backlog_rcv(struct sock *sk, struct sk_buff *skb)
 	int rc = 0;
 	struct llc_sock *llc = llc_sk(sk);
 
-	if (llc_backlog_type(skb) == LLC_PACKET) {
-		if (llc->state > 1) /* not closed */
+	if (likely(llc_backlog_type(skb) == LLC_PACKET)) {
+		if (likely(llc->state > 1)) /* not closed */
 			rc = llc_conn_rcv(sk, skb);
 		else
 			goto out_kfree_skb;
 	} else if (llc_backlog_type(skb) == LLC_EVENT) {
 		/* timer expiration event */
-		if (llc->state > 1)  /* not closed */
+		if (likely(llc->state > 1))  /* not closed */
 			rc = llc_conn_state_process(sk, skb);
 		else
 			goto out_kfree_skb;
@@ -799,22 +832,22 @@ static void llc_sk_init(struct sock* sk)
 	llc->dec_step = llc->connect_step = 1;
 
 	init_timer(&llc->ack_timer.timer);
-	llc->ack_timer.expire	      = LLC_ACK_TIME;
+	llc->ack_timer.expire	      = sysctl_llc2_ack_timeout;
 	llc->ack_timer.timer.data     = (unsigned long)sk;
 	llc->ack_timer.timer.function = llc_conn_ack_tmr_cb;
 
 	init_timer(&llc->pf_cycle_timer.timer);
-	llc->pf_cycle_timer.expire	   = LLC_P_TIME;
+	llc->pf_cycle_timer.expire	   = sysctl_llc2_p_timeout;
 	llc->pf_cycle_timer.timer.data     = (unsigned long)sk;
 	llc->pf_cycle_timer.timer.function = llc_conn_pf_cycle_tmr_cb;
 
 	init_timer(&llc->rej_sent_timer.timer);
-	llc->rej_sent_timer.expire	   = LLC_REJ_TIME;
+	llc->rej_sent_timer.expire	   = sysctl_llc2_rej_timeout;
 	llc->rej_sent_timer.timer.data     = (unsigned long)sk;
 	llc->rej_sent_timer.timer.function = llc_conn_rej_tmr_cb;
 
 	init_timer(&llc->busy_state_timer.timer);
-	llc->busy_state_timer.expire	     = LLC_BUSY_TIME;
+	llc->busy_state_timer.expire	     = sysctl_llc2_busy_timeout;
 	llc->busy_state_timer.timer.data     = (unsigned long)sk;
 	llc->busy_state_timer.timer.function = llc_conn_busy_tmr_cb;
 
@@ -834,7 +867,7 @@ static void llc_sk_init(struct sock* sk)
  *	Allocates a LLC sock and initializes it. Returns the new LLC sock
  *	or %NULL if there's no memory available for one
  */
-struct sock *llc_sk_alloc(int family, int priority, struct proto *prot)
+struct sock *llc_sk_alloc(int family, gfp_t priority, struct proto *prot)
 {
 	struct sock *sk = sk_alloc(family, priority, prot, 1);
 
diff --git a/net/llc/llc_core.c b/net/llc/llc_core.c
index 9727455bf0e..ab0fcd32fd8 100644
--- a/net/llc/llc_core.c
+++ b/net/llc/llc_core.c
@@ -40,6 +40,7 @@ static struct llc_sap *llc_sap_alloc(void)
 		sap->state = LLC_SAP_STATE_ACTIVE;
 		memcpy(sap->laddr.mac, llc_station_mac_sa, ETH_ALEN);
 		rwlock_init(&sap->sk_list.lock);
+		atomic_set(&sap->refcnt, 1);
 	}
 	return sap;
 }
@@ -52,9 +53,7 @@ static struct llc_sap *llc_sap_alloc(void)
  */
 static void llc_add_sap(struct llc_sap *sap)
 {
-	write_lock_bh(&llc_sap_list_lock);
 	list_add_tail(&sap->node, &llc_sap_list);
-	write_unlock_bh(&llc_sap_list_lock);
 }
 
 /**
@@ -70,11 +69,25 @@ static void llc_del_sap(struct llc_sap *sap)
 	write_unlock_bh(&llc_sap_list_lock);
 }
 
+static struct llc_sap *__llc_sap_find(unsigned char sap_value)
+{
+	struct llc_sap* sap;
+
+	list_for_each_entry(sap, &llc_sap_list, node)
+		if (sap->laddr.lsap == sap_value)
+			goto out;
+	sap = NULL;
+out:
+	return sap;
+}
+
 /**
  *	llc_sap_find - searchs a SAP in station
  *	@sap_value: sap to be found
  *
  *	Searchs for a sap in the sap list of the LLC's station upon the sap ID.
+ *	If the sap is found it will be refcounted and the user will have to do
+ *	a llc_sap_put after use.
  *	Returns the sap or %NULL if not found.
  */
 struct llc_sap *llc_sap_find(unsigned char sap_value)
@@ -82,11 +95,9 @@ struct llc_sap *llc_sap_find(unsigned char sap_value)
 	struct llc_sap* sap;
 
 	read_lock_bh(&llc_sap_list_lock);
-	list_for_each_entry(sap, &llc_sap_list, node)
-		if (sap->laddr.lsap == sap_value)
-			goto out;
-	sap = NULL;
-out:
+	sap = __llc_sap_find(sap_value);
+	if (sap)
+		llc_sap_hold(sap);
 	read_unlock_bh(&llc_sap_list_lock);
 	return sap;
 }
@@ -106,19 +117,20 @@ struct llc_sap *llc_sap_open(unsigned char lsap,
 					 struct packet_type *pt,
 					 struct net_device *orig_dev))
 {
-	struct llc_sap *sap = llc_sap_find(lsap);
+	struct llc_sap *sap = NULL;
 
-	if (sap) { /* SAP already exists */
-		sap = NULL;
+	write_lock_bh(&llc_sap_list_lock);
+	if (__llc_sap_find(lsap)) /* SAP already exists */
 		goto out;
-	}
 	sap = llc_sap_alloc();
 	if (!sap)
 		goto out;
 	sap->laddr.lsap = lsap;
 	sap->rcv_func	= func;
+	llc_sap_hold(sap);
 	llc_add_sap(sap);
 out:
+	write_unlock_bh(&llc_sap_list_lock);
 	return sap;
 }
 
diff --git a/net/llc/llc_if.c b/net/llc/llc_if.c
index 0f84f66018e..ba90f7f0801 100644
--- a/net/llc/llc_if.c
+++ b/net/llc/llc_if.c
@@ -47,14 +47,11 @@ int llc_build_and_send_pkt(struct sock *sk, struct sk_buff *skb)
 	int rc = -ECONNABORTED;
 	struct llc_sock *llc = llc_sk(sk);
 
-	if (llc->state == LLC_CONN_STATE_ADM)
+	if (unlikely(llc->state == LLC_CONN_STATE_ADM))
 		goto out;
 	rc = -EBUSY;
-	if (llc_data_accept_state(llc->state)) { /* data_conn_refuse */
-		llc->failed_data_req = 1;
-		goto out;
-	}
-	if (llc->p_flag) {
+	if (unlikely(llc_data_accept_state(llc->state) || /* data_conn_refuse */
+		     llc->p_flag)) {
 		llc->failed_data_req = 1;
 		goto out;
 	}
@@ -110,6 +107,7 @@ int llc_establish_connection(struct sock *sk, u8 *lmac, u8 *dmac, u8 dsap)
 		ev->type      = LLC_CONN_EV_TYPE_PRIM;
 		ev->prim      = LLC_CONN_PRIM;
 		ev->prim_type = LLC_PRIM_TYPE_REQ;
+		skb_set_owner_w(skb, sk);
 		rc = llc_conn_state_process(sk, skb);
 	}
 out_put:
@@ -144,6 +142,7 @@ int llc_send_disc(struct sock *sk)
 	skb = alloc_skb(0, GFP_ATOMIC);
 	if (!skb)
 		goto out;
+	skb_set_owner_w(skb, sk);
 	sk->sk_state  = TCP_CLOSING;
 	ev	      = llc_conn_ev(skb);
 	ev->type      = LLC_CONN_EV_TYPE_PRIM;
diff --git a/net/llc/llc_input.c b/net/llc/llc_input.c
index 13b46240b7a..8f3addf0724 100644
--- a/net/llc/llc_input.c
+++ b/net/llc/llc_input.c
@@ -99,15 +99,19 @@ out:
 static inline int llc_fixup_skb(struct sk_buff *skb)
 {
 	u8 llc_len = 2;
-	struct llc_pdu_sn *pdu;
+	struct llc_pdu_un *pdu;
 
-	if (!pskb_may_pull(skb, sizeof(*pdu)))
+	if (unlikely(!pskb_may_pull(skb, sizeof(*pdu))))
 		return 0;
 
-	pdu = (struct llc_pdu_sn *)skb->data;
+	pdu = (struct llc_pdu_un *)skb->data;
 	if ((pdu->ctrl_1 & LLC_PDU_TYPE_MASK) == LLC_PDU_TYPE_U)
 		llc_len = 1;
 	llc_len += 2;
+
+	if (unlikely(!pskb_may_pull(skb, llc_len)))
+		return 0;
+
 	skb->h.raw += llc_len;
 	skb_pull(skb, llc_len);
 	if (skb->protocol == htons(ETH_P_802_2)) {
@@ -166,17 +170,22 @@ int llc_rcv(struct sk_buff *skb, struct net_device *dev,
 	 */
 	if (sap->rcv_func) {
 		sap->rcv_func(skb, dev, pt, orig_dev);
-		goto out;
+		goto out_put;
 	}
 	dest = llc_pdu_type(skb);
 	if (unlikely(!dest || !llc_type_handlers[dest - 1]))
-		goto drop;
+		goto drop_put;
 	llc_type_handlers[dest - 1](sap, skb);
+out_put:
+	llc_sap_put(sap);
 out:
 	return 0;
 drop:
 	kfree_skb(skb);
 	goto out;
+drop_put:
+	kfree_skb(skb);
+	goto out_put;
 handle_station:
 	if (!llc_station_handler)
 		goto drop;
diff --git a/net/llc/llc_output.c b/net/llc/llc_output.c
index ab5784cf163..b4d55b6abb6 100644
--- a/net/llc/llc_output.c
+++ b/net/llc/llc_output.c
@@ -98,7 +98,7 @@ int llc_build_and_send_ui_pkt(struct llc_sap *sap, struct sk_buff *skb,
 			    dsap, LLC_PDU_CMD);
 	llc_pdu_init_as_ui_cmd(skb);
 	rc = llc_mac_hdr_init(skb, skb->dev->dev_addr, dmac);
-	if (!rc)
+	if (likely(!rc))
 		rc = dev_queue_xmit(skb);
 	return rc;
 }
diff --git a/net/llc/llc_proc.c b/net/llc/llc_proc.c
index 36e8db3fa1a..bd531cb235a 100644
--- a/net/llc/llc_proc.c
+++ b/net/llc/llc_proc.c
@@ -134,7 +134,7 @@ static int llc_seq_socket_show(struct seq_file *seq, void *v)
 	llc_ui_format_mac(seq, llc->daddr.mac);
 	seq_printf(seq, "@%02X %8d %8d %2d %3d %4d\n", llc->daddr.lsap,
 		   atomic_read(&sk->sk_wmem_alloc),
-		   atomic_read(&sk->sk_rmem_alloc),
+		   atomic_read(&sk->sk_rmem_alloc) - llc->copied_seq,
 		   sk->sk_state,
 		   sk->sk_socket ? SOCK_INODE(sk->sk_socket)->i_uid : -1,
 		   llc->link);
diff --git a/net/llc/llc_s_ac.c b/net/llc/llc_s_ac.c
index ed8ba7de612..bb3580fb8cf 100644
--- a/net/llc/llc_s_ac.c
+++ b/net/llc/llc_s_ac.c
@@ -58,7 +58,7 @@ int llc_sap_action_send_ui(struct llc_sap *sap, struct sk_buff *skb)
 			    ev->daddr.lsap, LLC_PDU_CMD);
 	llc_pdu_init_as_ui_cmd(skb);
 	rc = llc_mac_hdr_init(skb, ev->saddr.mac, ev->daddr.mac);
-	if (!rc)
+	if (likely(!rc))
 		rc = dev_queue_xmit(skb);
 	return rc;
 }
@@ -81,7 +81,7 @@ int llc_sap_action_send_xid_c(struct llc_sap *sap, struct sk_buff *skb)
 			    ev->daddr.lsap, LLC_PDU_CMD);
 	llc_pdu_init_as_xid_cmd(skb, LLC_XID_NULL_CLASS_2, 0);
 	rc = llc_mac_hdr_init(skb, ev->saddr.mac, ev->daddr.mac);
-	if (!rc)
+	if (likely(!rc))
 		rc = dev_queue_xmit(skb);
 	return rc;
 }
@@ -103,15 +103,14 @@ int llc_sap_action_send_xid_r(struct llc_sap *sap, struct sk_buff *skb)
 	llc_pdu_decode_sa(skb, mac_da);
 	llc_pdu_decode_da(skb, mac_sa);
 	llc_pdu_decode_ssap(skb, &dsap);
-	nskb = llc_alloc_frame();
+	nskb = llc_alloc_frame(NULL, skb->dev);
 	if (!nskb)
 		goto out;
-	nskb->dev = skb->dev;
 	llc_pdu_header_init(nskb, LLC_PDU_TYPE_U, sap->laddr.lsap, dsap,
 			    LLC_PDU_RSP);
 	llc_pdu_init_as_xid_rsp(nskb, LLC_XID_NULL_CLASS_2, 0);
 	rc = llc_mac_hdr_init(nskb, mac_sa, mac_da);
-	if (!rc)
+	if (likely(!rc))
 		rc = dev_queue_xmit(nskb);
 out:
 	return rc;
@@ -135,7 +134,7 @@ int llc_sap_action_send_test_c(struct llc_sap *sap, struct sk_buff *skb)
 			    ev->daddr.lsap, LLC_PDU_CMD);
 	llc_pdu_init_as_test_cmd(skb);
 	rc = llc_mac_hdr_init(skb, ev->saddr.mac, ev->daddr.mac);
-	if (!rc)
+	if (likely(!rc))
 		rc = dev_queue_xmit(skb);
 	return rc;
 }
@@ -149,15 +148,14 @@ int llc_sap_action_send_test_r(struct llc_sap *sap, struct sk_buff *skb)
 	llc_pdu_decode_sa(skb, mac_da);
 	llc_pdu_decode_da(skb, mac_sa);
 	llc_pdu_decode_ssap(skb, &dsap);
-	nskb = llc_alloc_frame();
+	nskb = llc_alloc_frame(NULL, skb->dev);
 	if (!nskb)
 		goto out;
-	nskb->dev = skb->dev;
 	llc_pdu_header_init(nskb, LLC_PDU_TYPE_U, sap->laddr.lsap, dsap,
 			    LLC_PDU_RSP);
 	llc_pdu_init_as_test_rsp(nskb, skb);
 	rc = llc_mac_hdr_init(nskb, mac_sa, mac_da);
-	if (!rc)
+	if (likely(!rc))
 		rc = dev_queue_xmit(nskb);
 out:
 	return rc;
diff --git a/net/llc/llc_sap.c b/net/llc/llc_sap.c
index 34228ef1498..4029ceee9b9 100644
--- a/net/llc/llc_sap.c
+++ b/net/llc/llc_sap.c
@@ -26,11 +26,12 @@
 
 /**
  *	llc_alloc_frame - allocates sk_buff for frame
+ *	@dev: network device this skb will be sent over
  *
  *	Allocates an sk_buff for frame and initializes sk_buff fields.
  *	Returns allocated skb or %NULL when out of memory.
  */
-struct sk_buff *llc_alloc_frame(void)
+struct sk_buff *llc_alloc_frame(struct sock *sk, struct net_device *dev)
 {
 	struct sk_buff *skb = alloc_skb(128, GFP_ATOMIC);
 
@@ -38,18 +39,23 @@ struct sk_buff *llc_alloc_frame(void)
 		skb_reserve(skb, 50);
 		skb->nh.raw   = skb->h.raw = skb->data;
 		skb->protocol = htons(ETH_P_802_2);
-		skb->dev      = dev_base->next;
+		skb->dev      = dev;
 		skb->mac.raw  = skb->head;
+		if (sk != NULL)
+			skb_set_owner_w(skb, sk);
 	}
 	return skb;
 }
 
-void llc_save_primitive(struct sk_buff* skb, u8 prim)
+void llc_save_primitive(struct sock *sk, struct sk_buff* skb, u8 prim)
 {
-	struct sockaddr_llc *addr = llc_ui_skb_cb(skb);
+	struct sockaddr_llc *addr;
 
+	if (skb->sk->sk_type == SOCK_STREAM) /* See UNIX98 */
+		return;
        /* save primitive for use by the user. */
-	addr->sllc_family = skb->sk->sk_family;
+	addr		  = llc_ui_skb_cb(skb);
+	addr->sllc_family = sk->sk_family;
 	addr->sllc_arphrd = skb->dev->type;
 	addr->sllc_test   = prim == LLC_TEST_PRIM;
 	addr->sllc_xid    = prim == LLC_XID_PRIM;
@@ -189,7 +195,7 @@ static void llc_sap_state_process(struct llc_sap *sap, struct sk_buff *skb)
 		if (skb->sk->sk_state == TCP_LISTEN)
 			kfree_skb(skb);
 		else {
-			llc_save_primitive(skb, ev->prim);
+			llc_save_primitive(skb->sk, skb, ev->prim);
 
 			/* queue skb to the user. */
 			if (sock_queue_rcv_skb(skb->sk, skb))
@@ -308,7 +314,7 @@ void llc_sap_handler(struct llc_sap *sap, struct sk_buff *skb)
 
 	sk = llc_lookup_dgram(sap, &laddr);
 	if (sk) {
-		skb->sk = sk;
+		skb_set_owner_r(skb, sk);
 		llc_sap_rcv(sap, skb);
 		sock_put(sk);
 	} else
diff --git a/net/llc/llc_station.c b/net/llc/llc_station.c
index 8fe48a24bad..f37dbf8ef12 100644
--- a/net/llc/llc_station.c
+++ b/net/llc/llc_station.c
@@ -50,6 +50,10 @@ struct llc_station {
 	struct sk_buff_head	    mac_pdu_q;
 };
 
+#define LLC_STATION_ACK_TIME (3 * HZ)
+
+int sysctl_llc_station_ack_timeout = LLC_STATION_ACK_TIME;
+
 /* Types of events (possible values in 'ev->type') */
 #define LLC_STATION_EV_TYPE_SIMPLE	1
 #define LLC_STATION_EV_TYPE_CONDITION	2
@@ -218,7 +222,8 @@ static void llc_station_send_pdu(struct sk_buff *skb)
 
 static int llc_station_ac_start_ack_timer(struct sk_buff *skb)
 {
-	mod_timer(&llc_main_station.ack_timer, jiffies + LLC_ACK_TIME * HZ);
+	mod_timer(&llc_main_station.ack_timer,
+		  jiffies + sysctl_llc_station_ack_timeout);
 	return 0;
 }
 
@@ -249,14 +254,14 @@ static int llc_station_ac_inc_xid_r_cnt_by_1(struct sk_buff *skb)
 static int llc_station_ac_send_null_dsap_xid_c(struct sk_buff *skb)
 {
 	int rc = 1;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct sk_buff *nskb = llc_alloc_frame(NULL, skb->dev);
 
 	if (!nskb)
 		goto out;
 	llc_pdu_header_init(nskb, LLC_PDU_TYPE_U, 0, 0, LLC_PDU_CMD);
 	llc_pdu_init_as_xid_cmd(nskb, LLC_XID_NULL_CLASS_2, 127);
 	rc = llc_mac_hdr_init(nskb, llc_station_mac_sa, llc_station_mac_sa);
-	if (rc)
+	if (unlikely(rc))
 		goto free;
 	llc_station_send_pdu(nskb);
 out:
@@ -270,18 +275,17 @@ static int llc_station_ac_send_xid_r(struct sk_buff *skb)
 {
 	u8 mac_da[ETH_ALEN], dsap;
 	int rc = 1;
-	struct sk_buff* nskb = llc_alloc_frame();
+	struct sk_buff* nskb = llc_alloc_frame(NULL, skb->dev);
 
 	if (!nskb)
 		goto out;
 	rc = 0;
-	nskb->dev = skb->dev;
 	llc_pdu_decode_sa(skb, mac_da);
 	llc_pdu_decode_ssap(skb, &dsap);
 	llc_pdu_header_init(nskb, LLC_PDU_TYPE_U, 0, dsap, LLC_PDU_RSP);
 	llc_pdu_init_as_xid_rsp(nskb, LLC_XID_NULL_CLASS_2, 127);
 	rc = llc_mac_hdr_init(nskb, llc_station_mac_sa, mac_da);
-	if (rc)
+	if (unlikely(rc))
 		goto free;
 	llc_station_send_pdu(nskb);
 out:
@@ -295,18 +299,17 @@ static int llc_station_ac_send_test_r(struct sk_buff *skb)
 {
 	u8 mac_da[ETH_ALEN], dsap;
 	int rc = 1;
-	struct sk_buff *nskb = llc_alloc_frame();
+	struct sk_buff *nskb = llc_alloc_frame(NULL, skb->dev);
 
 	if (!nskb)
 		goto out;
 	rc = 0;
-	nskb->dev = skb->dev;
 	llc_pdu_decode_sa(skb, mac_da);
 	llc_pdu_decode_ssap(skb, &dsap);
 	llc_pdu_header_init(nskb, LLC_PDU_TYPE_U, 0, dsap, LLC_PDU_RSP);
        	llc_pdu_init_as_test_rsp(nskb, skb);
 	rc = llc_mac_hdr_init(nskb, llc_station_mac_sa, mac_da);
-	if (rc)
+	if (unlikely(rc))
 		goto free;
 	llc_station_send_pdu(nskb);
 out:
@@ -689,7 +692,8 @@ int __init llc_station_init(void)
 	init_timer(&llc_main_station.ack_timer);
 	llc_main_station.ack_timer.data     = (unsigned long)&llc_main_station;
 	llc_main_station.ack_timer.function = llc_station_ack_tmr_cb;
-
+	llc_main_station.ack_timer.expires  = jiffies +
+						sysctl_llc_station_ack_timeout;
 	skb = alloc_skb(0, GFP_ATOMIC);
 	if (!skb)
 		goto out;
@@ -697,7 +701,6 @@ int __init llc_station_init(void)
 	llc_set_station_handler(llc_station_rcv);
 	ev = llc_station_ev(skb);
 	memset(ev, 0, sizeof(*ev));
-	llc_main_station.ack_timer.expires = jiffies + 3 * HZ;
 	llc_main_station.maximum_retry	= 1;
 	llc_main_station.state		= LLC_STATION_STATE_DOWN;
 	ev->type	= LLC_STATION_EV_TYPE_SIMPLE;
diff --git a/net/llc/sysctl_net_llc.c b/net/llc/sysctl_net_llc.c
new file mode 100644
index 00000000000..d1eaddb1363
--- /dev/null
+++ b/net/llc/sysctl_net_llc.c
@@ -0,0 +1,131 @@
+/*
+ * sysctl_net_llc.c: sysctl interface to LLC net subsystem.
+ * 
+ * Arnaldo Carvalho de Melo <acme@conectiva.com.br>
+ */
+
+#include <linux/config.h>
+#include <linux/mm.h>
+#include <linux/init.h>
+#include <linux/sysctl.h>
+#include <net/llc.h>
+
+#ifndef CONFIG_SYSCTL
+#error This file should not be compiled without CONFIG_SYSCTL defined
+#endif
+
+static struct ctl_table llc2_timeout_table[] = {
+	{
+		.ctl_name	= NET_LLC2_ACK_TIMEOUT,
+		.procname	= "ack",
+		.data		= &sysctl_llc2_ack_timeout,
+		.maxlen		= sizeof(long),
+		.mode		= 0644,
+		.proc_handler   = &proc_dointvec_jiffies,
+		.strategy       = &sysctl_jiffies,
+	},
+	{
+		.ctl_name	= NET_LLC2_BUSY_TIMEOUT,
+		.procname	= "busy",
+		.data		= &sysctl_llc2_busy_timeout,
+		.maxlen		= sizeof(long),
+		.mode		= 0644,
+		.proc_handler   = &proc_dointvec_jiffies,
+		.strategy       = &sysctl_jiffies,
+	},
+	{
+		.ctl_name	= NET_LLC2_P_TIMEOUT,
+		.procname	= "p",
+		.data		= &sysctl_llc2_p_timeout,
+		.maxlen		= sizeof(long),
+		.mode		= 0644,
+		.proc_handler   = &proc_dointvec_jiffies,
+		.strategy       = &sysctl_jiffies,
+	},
+	{
+		.ctl_name	= NET_LLC2_REJ_TIMEOUT,
+		.procname	= "rej",
+		.data		= &sysctl_llc2_rej_timeout,
+		.maxlen		= sizeof(long),
+		.mode		= 0644,
+		.proc_handler   = &proc_dointvec_jiffies,
+		.strategy       = &sysctl_jiffies,
+	},
+	{ 0 },
+};
+
+static struct ctl_table llc_station_table[] = {
+	{
+		.ctl_name	= NET_LLC_STATION_ACK_TIMEOUT,
+		.procname	= "ack_timeout",
+		.data		= &sysctl_llc_station_ack_timeout,
+		.maxlen		= sizeof(long),
+		.mode		= 0644,
+		.proc_handler   = &proc_dointvec_jiffies,
+		.strategy       = &sysctl_jiffies,
+	},
+	{ 0 },
+};
+
+static struct ctl_table llc2_dir_timeout_table[] = {
+	{
+		.ctl_name	= NET_LLC2,
+		.procname	= "timeout",
+		.mode		= 0555,
+		.child		= llc2_timeout_table,
+       	},
+	{ 0 },
+};
+
+static struct ctl_table llc_table[] = {
+	{
+		.ctl_name	= NET_LLC2,
+		.procname	= "llc2",
+		.mode		= 0555,
+		.child		= llc2_dir_timeout_table,
+       	},
+	{
+		.ctl_name       = NET_LLC_STATION,
+		.procname       = "station",
+		.mode           = 0555,
+		.child          = llc_station_table,
+	},
+	{ 0 },
+};
+
+static struct ctl_table llc_dir_table[] = {
+	{
+		.ctl_name	= NET_LLC,
+		.procname	= "llc",
+		.mode		= 0555,
+		.child		= llc_table,
+       	},
+	{ 0 },
+};
+
+static struct ctl_table llc_root_table[] = {
+	{
+		.ctl_name	= CTL_NET,
+		.procname	= "net",
+		.mode		= 0555,
+		.child		= llc_dir_table,
+	},
+	{ 0 },
+};
+
+static struct ctl_table_header *llc_table_header;
+
+int __init llc_sysctl_init(void)
+{
+	llc_table_header = register_sysctl_table(llc_root_table, 1);
+
+	return llc_table_header ? 0 : -ENOMEM;
+}
+
+void llc_sysctl_exit(void)
+{
+	if (llc_table_header) {
+		unregister_sysctl_table(llc_table_header);
+		llc_table_header = NULL;
+	}
+}
diff --git a/net/netfilter/nfnetlink.c b/net/netfilter/nfnetlink.c
index 49a3900e3d3..4bc27a6334c 100644
--- a/net/netfilter/nfnetlink.c
+++ b/net/netfilter/nfnetlink.c
@@ -133,7 +133,7 @@ int nfattr_parse(struct nfattr *tb[], int maxattr, struct nfattr *nfa, int len)
 	memset(tb, 0, sizeof(struct nfattr *) * maxattr);
 
 	while (NFA_OK(nfa, len)) {
-		unsigned flavor = nfa->nfa_type;
+		unsigned flavor = NFA_TYPE(nfa);
 		if (flavor && flavor <= maxattr)
 			tb[flavor-1] = nfa;
 		nfa = NFA_NEXT(nfa, len);
@@ -177,7 +177,7 @@ nfnetlink_check_attributes(struct nfnetlink_subsystem *subsys,
 		int attrlen = nlh->nlmsg_len - NLMSG_ALIGN(min_len);
 
 		while (NFA_OK(attr, attrlen)) {
-			unsigned flavor = attr->nfa_type;
+			unsigned flavor = NFA_TYPE(attr);
 			if (flavor) {
 				if (flavor > attr_count)
 					return -EINVAL;
@@ -195,7 +195,7 @@ nfnetlink_check_attributes(struct nfnetlink_subsystem *subsys,
 
 int nfnetlink_send(struct sk_buff *skb, u32 pid, unsigned group, int echo)
 {
-	int allocation = in_interrupt() ? GFP_ATOMIC : GFP_KERNEL;
+	gfp_t allocation = in_interrupt() ? GFP_ATOMIC : GFP_KERNEL;
 	int err = 0;
 
 	NETLINK_CB(skb).dst_group = group;
diff --git a/net/netfilter/nfnetlink_log.c b/net/netfilter/nfnetlink_log.c
index ff5601ceedc..efcd10f996b 100644
--- a/net/netfilter/nfnetlink_log.c
+++ b/net/netfilter/nfnetlink_log.c
@@ -494,8 +494,8 @@ __build_packet_message(struct nfulnl_instance *inst,
 	if (skb->tstamp.off_sec) {
 		struct nfulnl_msg_packet_timestamp ts;
 
-		ts.sec = cpu_to_be64(skb_tv_base.tv_sec + skb->tstamp.off_sec);
-		ts.usec = cpu_to_be64(skb_tv_base.tv_usec + skb->tstamp.off_usec);
+		ts.sec = cpu_to_be64(skb->tstamp.off_sec);
+		ts.usec = cpu_to_be64(skb->tstamp.off_usec);
 
 		NFA_PUT(inst->skb, NFULA_TIMESTAMP, sizeof(ts), &ts);
 	}
diff --git a/net/netfilter/nfnetlink_queue.c b/net/netfilter/nfnetlink_queue.c
index f81fe8c52e9..eaa44c49567 100644
--- a/net/netfilter/nfnetlink_queue.c
+++ b/net/netfilter/nfnetlink_queue.c
@@ -492,8 +492,8 @@ nfqnl_build_packet_message(struct nfqnl_instance *queue,
 	if (entry->skb->tstamp.off_sec) {
 		struct nfqnl_msg_packet_timestamp ts;
 
-		ts.sec = cpu_to_be64(skb_tv_base.tv_sec + entry->skb->tstamp.off_sec);
-		ts.usec = cpu_to_be64(skb_tv_base.tv_usec + entry->skb->tstamp.off_usec);
+		ts.sec = cpu_to_be64(entry->skb->tstamp.off_sec);
+		ts.usec = cpu_to_be64(entry->skb->tstamp.off_usec);
 
 		NFA_PUT(skb, NFQA_TIMESTAMP, sizeof(ts), &ts);
 	}
diff --git a/net/netlink/af_netlink.c b/net/netlink/af_netlink.c
index a64e1d5ce3c..5ca283537bc 100644
--- a/net/netlink/af_netlink.c
+++ b/net/netlink/af_netlink.c
@@ -740,11 +740,8 @@ int netlink_attachskb(struct sock *sk, struct sk_buff *skb, int nonblock, long t
 
 int netlink_sendskb(struct sock *sk, struct sk_buff *skb, int protocol)
 {
-	struct netlink_sock *nlk;
 	int len = skb->len;
 
-	nlk = nlk_sk(sk);
-
 	skb_queue_tail(&sk->sk_receive_queue, skb);
 	sk->sk_data_ready(sk, len);
 	sock_put(sk);
@@ -758,7 +755,7 @@ void netlink_detachskb(struct sock *sk, struct sk_buff *skb)
 }
 
 static inline struct sk_buff *netlink_trim(struct sk_buff *skb,
-					   unsigned int __nocast allocation)
+					   gfp_t allocation)
 {
 	int delta;
 
@@ -827,7 +824,7 @@ struct netlink_broadcast_data {
 	int failure;
 	int congested;
 	int delivered;
-	unsigned int allocation;
+	gfp_t allocation;
 	struct sk_buff *skb, *skb2;
 };
 
@@ -880,7 +877,7 @@ out:
 }
 
 int netlink_broadcast(struct sock *ssk, struct sk_buff *skb, u32 pid,
-		      u32 group, unsigned int __nocast allocation)
+		      u32 group, gfp_t allocation)
 {
 	struct netlink_broadcast_data info;
 	struct hlist_node *node;
diff --git a/net/netrom/nr_dev.c b/net/netrom/nr_dev.c
index 4e66eef9a03..509afddae56 100644
--- a/net/netrom/nr_dev.c
+++ b/net/netrom/nr_dev.c
@@ -58,7 +58,7 @@ int nr_rx_ip(struct sk_buff *skb, struct net_device *dev)
 
 	/* Spoof incoming device */
 	skb->dev      = dev;
-	skb->h.raw    = skb->data;
+	skb->mac.raw  = skb->nh.raw;
 	skb->nh.raw   = skb->data;
 	skb->pkt_type = PACKET_HOST;
 
diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c
index ee865d88183..499ae3df4a4 100644
--- a/net/packet/af_packet.c
+++ b/net/packet/af_packet.c
@@ -654,8 +654,8 @@ static int tpacket_rcv(struct sk_buff *skb, struct net_device *dev, struct packe
 		__net_timestamp(skb);
 		sock_enable_timestamp(sk);
 	}
-	h->tp_sec = skb_tv_base.tv_sec + skb->tstamp.off_sec;
-	h->tp_usec = skb_tv_base.tv_usec + skb->tstamp.off_usec;
+	h->tp_sec = skb->tstamp.off_sec;
+	h->tp_usec = skb->tstamp.off_usec;
 
 	sll = (struct sockaddr_ll*)((u8*)h + TPACKET_ALIGN(sizeof(*h)));
 	sll->sll_halen = 0;
@@ -761,12 +761,6 @@ static int packet_sendmsg(struct kiocb *iocb, struct socket *sock,
 	if (dev->hard_header) {
 		int res;
 		err = -EINVAL;
-		if (saddr) {
-			if (saddr->sll_halen != dev->addr_len)
-				goto out_free;
-			if (saddr->sll_hatype != dev->type)
-				goto out_free;
-		}
 		res = dev->hard_header(skb, dev, ntohs(proto), addr, NULL, len);
 		if (sock->type != SOCK_DGRAM) {
 			skb->tail = skb->data;
diff --git a/net/rose/af_rose.c b/net/rose/af_rose.c
index 5acb1680524..829fdbc4400 100644
--- a/net/rose/af_rose.c
+++ b/net/rose/af_rose.c
@@ -1472,22 +1472,25 @@ static const char banner[] = KERN_INFO "F6FBB/G4KLX ROSE for Linux. Version 0.62
 static int __init rose_proto_init(void)
 {
 	int i;
-	int rc = proto_register(&rose_proto, 0);
+	int rc;
 
+	if (rose_ndevs > 0x7FFFFFFF/sizeof(struct net_device *)) {
+		printk(KERN_ERR "ROSE: rose_proto_init - rose_ndevs parameter to large\n");
+		rc = -EINVAL;
+		goto out;
+	}
+
+	rc = proto_register(&rose_proto, 0);
 	if (rc != 0)
 		goto out;
 
 	rose_callsign = null_ax25_address;
 
-	if (rose_ndevs > 0x7FFFFFFF/sizeof(struct net_device *)) {
-		printk(KERN_ERR "ROSE: rose_proto_init - rose_ndevs parameter to large\n");
-		return -1;
-	}
-
 	dev_rose = kmalloc(rose_ndevs * sizeof(struct net_device *), GFP_KERNEL);
 	if (dev_rose == NULL) {
 		printk(KERN_ERR "ROSE: rose_proto_init - unable to allocate device structure\n");
-		return -1;
+		rc = -ENOMEM;
+		goto out_proto_unregister;
 	}
 
 	memset(dev_rose, 0x00, rose_ndevs * sizeof(struct net_device*));
@@ -1500,10 +1503,12 @@ static int __init rose_proto_init(void)
 				   name, rose_setup);
 		if (!dev) {
 			printk(KERN_ERR "ROSE: rose_proto_init - unable to allocate memory\n");
+			rc = -ENOMEM;
 			goto fail;
 		}
-		if (register_netdev(dev)) {
-			printk(KERN_ERR "ROSE: netdevice regeistration failed\n");
+		rc = register_netdev(dev);
+		if (rc) {
+			printk(KERN_ERR "ROSE: netdevice registration failed\n");
 			free_netdev(dev);
 			goto fail;
 		}
@@ -1536,8 +1541,9 @@ fail:
 		free_netdev(dev_rose[i]);
 	}
 	kfree(dev_rose);
+out_proto_unregister:
 	proto_unregister(&rose_proto);
-	return -ENOMEM;
+	goto out;
 }
 module_init(rose_proto_init);
 
diff --git a/net/rose/rose_route.c b/net/rose/rose_route.c
index e556d92c0bc..b18fe504301 100644
--- a/net/rose/rose_route.c
+++ b/net/rose/rose_route.c
@@ -727,7 +727,7 @@ int rose_rt_ioctl(unsigned int cmd, void __user *arg)
 		}
 		if (rose_route.mask > 10) /* Mask can't be more than 10 digits */
 			return -EINVAL;
-		if (rose_route.ndigis > 8) /* No more than 8 digipeats */
+		if (rose_route.ndigis > AX25_MAX_DIGIS)
 			return -EINVAL;
 		err = rose_add_node(&rose_route, dev);
 		dev_put(dev);
diff --git a/net/rxrpc/call.c b/net/rxrpc/call.c
index 5cfd4cadee4..c4aeb7d4026 100644
--- a/net/rxrpc/call.c
+++ b/net/rxrpc/call.c
@@ -1923,7 +1923,7 @@ int rxrpc_call_write_data(struct rxrpc_call *call,
 			  size_t sioc,
 			  struct kvec *siov,
 			  u8 rxhdr_flags,
-			  int alloc_flags,
+			  gfp_t alloc_flags,
 			  int dup_data,
 			  size_t *size_sent)
 {
diff --git a/net/rxrpc/connection.c b/net/rxrpc/connection.c
index 61463c74f8c..2ba14a75dbb 100644
--- a/net/rxrpc/connection.c
+++ b/net/rxrpc/connection.c
@@ -522,7 +522,7 @@ int rxrpc_conn_newmsg(struct rxrpc_connection *conn,
 		      uint8_t type,
 		      int dcount,
 		      struct kvec diov[],
-		      int alloc_flags,
+		      gfp_t alloc_flags,
 		      struct rxrpc_message **_msg)
 {
 	struct rxrpc_message *msg;
diff --git a/net/sched/Kconfig b/net/sched/Kconfig
index 45d3bc0812c..81510da3179 100644
--- a/net/sched/Kconfig
+++ b/net/sched/Kconfig
@@ -72,9 +72,11 @@ config NET_SCH_CLK_GETTIMEOFDAY
 	  Choose this if you need a high resolution clock source but can't use
 	  the CPU's cycle counter.
 
+# don't allow on SMP x86 because they can have unsynchronized TSCs.
+# gettimeofday is a good alternative
 config NET_SCH_CLK_CPU
 	bool "CPU cycle counter"
-	depends on X86_TSC || X86_64 || ALPHA || SPARC64 || PPC64 || IA64
+	depends on ((X86_TSC || X86_64) && !SMP) || ALPHA || SPARC64 || PPC64 || IA64
 	help
 	  Say Y here if you want to use the CPU's cycle counter as clock source.
 	  This is a cheap and high resolution clock source, but on some
diff --git a/net/sched/em_meta.c b/net/sched/em_meta.c
index 00eae5f9a01..cf68a59fdc5 100644
--- a/net/sched/em_meta.c
+++ b/net/sched/em_meta.c
@@ -393,10 +393,10 @@ META_COLLECTOR(int_sk_route_caps)
 	dst->value = skb->sk->sk_route_caps;
 }
 
-META_COLLECTOR(int_sk_hashent)
+META_COLLECTOR(int_sk_hash)
 {
 	SKIP_NONLOCAL(skb);
-	dst->value = skb->sk->sk_hashent;
+	dst->value = skb->sk->sk_hash;
 }
 
 META_COLLECTOR(int_sk_lingertime)
@@ -515,7 +515,7 @@ static struct meta_ops __meta_ops[TCF_META_TYPE_MAX+1][TCF_META_ID_MAX+1] = {
 		[META_ID(SK_FORWARD_ALLOCS)]	= META_FUNC(int_sk_fwd_alloc),
 		[META_ID(SK_ALLOCS)]		= META_FUNC(int_sk_alloc),
 		[META_ID(SK_ROUTE_CAPS)]	= META_FUNC(int_sk_route_caps),
-		[META_ID(SK_HASHENT)]		= META_FUNC(int_sk_hashent),
+		[META_ID(SK_HASH)]		= META_FUNC(int_sk_hash),
 		[META_ID(SK_LINGERTIME)]	= META_FUNC(int_sk_lingertime),
 		[META_ID(SK_ACK_BACKLOG)]	= META_FUNC(int_sk_ack_bl),
 		[META_ID(SK_MAX_ACK_BACKLOG)]	= META_FUNC(int_sk_max_ack_bl),
diff --git a/net/sctp/associola.c b/net/sctp/associola.c
index 5b24ae0650d..12b0f582a66 100644
--- a/net/sctp/associola.c
+++ b/net/sctp/associola.c
@@ -71,7 +71,7 @@ static struct sctp_association *sctp_association_init(struct sctp_association *a
 					  const struct sctp_endpoint *ep,
 					  const struct sock *sk,
 					  sctp_scope_t scope,
-					  unsigned int __nocast gfp)
+					  gfp_t gfp)
 {
 	struct sctp_sock *sp;
 	int i;
@@ -273,7 +273,7 @@ fail_init:
 struct sctp_association *sctp_association_new(const struct sctp_endpoint *ep,
 					 const struct sock *sk,
 					 sctp_scope_t scope,
-					 unsigned int __nocast gfp)
+					 gfp_t gfp)
 {
 	struct sctp_association *asoc;
 
@@ -479,7 +479,7 @@ void sctp_assoc_rm_peer(struct sctp_association *asoc,
 /* Add a transport address to an association.  */
 struct sctp_transport *sctp_assoc_add_peer(struct sctp_association *asoc,
 					   const union sctp_addr *addr,
-					   const unsigned int __nocast gfp,
+					   const gfp_t gfp,
 					   const int peer_state)
 {
 	struct sctp_transport *peer;
@@ -1231,7 +1231,7 @@ void sctp_assoc_rwnd_decrease(struct sctp_association *asoc, unsigned len)
  * local endpoint and the remote peer.
  */
 int sctp_assoc_set_bind_addr_from_ep(struct sctp_association *asoc,
-				     unsigned int __nocast gfp)
+				     gfp_t gfp)
 {
 	sctp_scope_t scope;
 	int flags;
@@ -1254,7 +1254,7 @@ int sctp_assoc_set_bind_addr_from_ep(struct sctp_association *asoc,
 /* Build the association's bind address list from the cookie.  */
 int sctp_assoc_set_bind_addr_from_cookie(struct sctp_association *asoc,
 					 struct sctp_cookie *cookie,
-					 unsigned int __nocast gfp)
+					 gfp_t gfp)
 {
 	int var_size2 = ntohs(cookie->peer_init->chunk_hdr.length);
 	int var_size3 = cookie->raw_addr_list_len;
diff --git a/net/sctp/bind_addr.c b/net/sctp/bind_addr.c
index f71549710f2..2b962627f63 100644
--- a/net/sctp/bind_addr.c
+++ b/net/sctp/bind_addr.c
@@ -53,7 +53,7 @@
 
 /* Forward declarations for internal helpers. */
 static int sctp_copy_one_addr(struct sctp_bind_addr *, union sctp_addr *,
-			      sctp_scope_t scope, unsigned int __nocast gfp,
+			      sctp_scope_t scope, gfp_t gfp,
 			      int flags);
 static void sctp_bind_addr_clean(struct sctp_bind_addr *);
 
@@ -64,7 +64,7 @@ static void sctp_bind_addr_clean(struct sctp_bind_addr *);
  */
 int sctp_bind_addr_copy(struct sctp_bind_addr *dest, 
 			const struct sctp_bind_addr *src,
-			sctp_scope_t scope, unsigned int __nocast gfp,
+			sctp_scope_t scope, gfp_t gfp,
 			int flags)
 {
 	struct sctp_sockaddr_entry *addr;
@@ -146,7 +146,7 @@ void sctp_bind_addr_free(struct sctp_bind_addr *bp)
 
 /* Add an address to the bind address list in the SCTP_bind_addr structure. */
 int sctp_add_bind_addr(struct sctp_bind_addr *bp, union sctp_addr *new,
-		       unsigned int __nocast gfp)
+		       gfp_t gfp)
 {
 	struct sctp_sockaddr_entry *addr;
 
@@ -200,7 +200,7 @@ int sctp_del_bind_addr(struct sctp_bind_addr *bp, union sctp_addr *del_addr)
  */
 union sctp_params sctp_bind_addrs_to_raw(const struct sctp_bind_addr *bp,
 					 int *addrs_len,
-					 unsigned int __nocast gfp)
+					 gfp_t gfp)
 {
 	union sctp_params addrparms;
 	union sctp_params retval;
@@ -252,7 +252,7 @@ end_raw:
  * address parameters).
  */
 int sctp_raw_to_bind_addrs(struct sctp_bind_addr *bp, __u8 *raw_addr_list,
-			   int addrs_len, __u16 port, unsigned int __nocast gfp)
+			   int addrs_len, __u16 port, gfp_t gfp)
 {
 	union sctp_addr_param *rawaddr;
 	struct sctp_paramhdr *param;
@@ -350,7 +350,7 @@ union sctp_addr *sctp_find_unmatch_addr(struct sctp_bind_addr	*bp,
 /* Copy out addresses from the global local address list. */
 static int sctp_copy_one_addr(struct sctp_bind_addr *dest, 
 			      union sctp_addr *addr,
-			      sctp_scope_t scope, unsigned int __nocast gfp,
+			      sctp_scope_t scope, gfp_t gfp,
 			      int flags)
 {
 	int error = 0;
diff --git a/net/sctp/chunk.c b/net/sctp/chunk.c
index 61da2937e64..83ef411772f 100644
--- a/net/sctp/chunk.c
+++ b/net/sctp/chunk.c
@@ -62,7 +62,7 @@ static void sctp_datamsg_init(struct sctp_datamsg *msg)
 }
 
 /* Allocate and initialize datamsg. */
-SCTP_STATIC struct sctp_datamsg *sctp_datamsg_new(unsigned int __nocast gfp)
+SCTP_STATIC struct sctp_datamsg *sctp_datamsg_new(gfp_t gfp)
 {
 	struct sctp_datamsg *msg;
 	msg = kmalloc(sizeof(struct sctp_datamsg), gfp);
diff --git a/net/sctp/endpointola.c b/net/sctp/endpointola.c
index e22ccd65596..96984f7a2d6 100644
--- a/net/sctp/endpointola.c
+++ b/net/sctp/endpointola.c
@@ -68,7 +68,7 @@ static void sctp_endpoint_bh_rcv(struct sctp_endpoint *ep);
  */
 static struct sctp_endpoint *sctp_endpoint_init(struct sctp_endpoint *ep,
 						struct sock *sk,
-						unsigned int __nocast gfp)
+						gfp_t gfp)
 {
 	struct sctp_sock *sp = sctp_sk(sk);
 	memset(ep, 0, sizeof(struct sctp_endpoint));
@@ -138,8 +138,7 @@ static struct sctp_endpoint *sctp_endpoint_init(struct sctp_endpoint *ep,
 /* Create a sctp_endpoint with all that boring stuff initialized.
  * Returns NULL if there isn't enough memory.
  */
-struct sctp_endpoint *sctp_endpoint_new(struct sock *sk,
-					unsigned int __nocast gfp)
+struct sctp_endpoint *sctp_endpoint_new(struct sock *sk, gfp_t gfp)
 {
 	struct sctp_endpoint *ep;
 
diff --git a/net/sctp/proc.c b/net/sctp/proc.c
index b74f7772b57..6e4dc28874d 100644
--- a/net/sctp/proc.c
+++ b/net/sctp/proc.c
@@ -69,9 +69,7 @@ fold_field(void *mib[], int nr)
 	unsigned long res = 0;
 	int i;
 
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!cpu_possible(i))
-			continue;
+	for_each_cpu(i) {
 		res +=
 		    *((unsigned long *) (((void *) per_cpu_ptr(mib[0], i)) +
 					 sizeof (unsigned long) * nr));
diff --git a/net/sctp/protocol.c b/net/sctp/protocol.c
index e7025be7769..26de4d3e1bd 100644
--- a/net/sctp/protocol.c
+++ b/net/sctp/protocol.c
@@ -147,7 +147,7 @@ static void sctp_v4_copy_addrlist(struct list_head *addrlist,
 	struct sctp_sockaddr_entry *addr;
 
 	rcu_read_lock();
-	if ((in_dev = __in_dev_get(dev)) == NULL) {
+	if ((in_dev = __in_dev_get_rcu(dev)) == NULL) {
 		rcu_read_unlock();
 		return;
 	}
@@ -219,7 +219,7 @@ static void sctp_free_local_addr_list(void)
 
 /* Copy the local addresses which are valid for 'scope' into 'bp'.  */
 int sctp_copy_local_addr_list(struct sctp_bind_addr *bp, sctp_scope_t scope,
-			      unsigned int __nocast gfp, int copy_flags)
+			      gfp_t gfp, int copy_flags)
 {
 	struct sctp_sockaddr_entry *addr;
 	int error = 0;
diff --git a/net/sctp/sm_make_chunk.c b/net/sctp/sm_make_chunk.c
index 3868a8d70cc..10e82ec2ebd 100644
--- a/net/sctp/sm_make_chunk.c
+++ b/net/sctp/sm_make_chunk.c
@@ -78,7 +78,7 @@ static sctp_cookie_param_t *sctp_pack_cookie(const struct sctp_endpoint *ep,
 static int sctp_process_param(struct sctp_association *asoc,
 			      union sctp_params param,
 			      const union sctp_addr *peer_addr,
-			      unsigned int __nocast gfp);
+			      gfp_t gfp);
 
 /* What was the inbound interface for this chunk? */
 int sctp_chunk_iif(const struct sctp_chunk *chunk)
@@ -174,7 +174,7 @@ void  sctp_init_cause(struct sctp_chunk *chunk, __u16 cause_code,
  */
 struct sctp_chunk *sctp_make_init(const struct sctp_association *asoc,
 			     const struct sctp_bind_addr *bp,
-			     unsigned int __nocast gfp, int vparam_len)
+			     gfp_t gfp, int vparam_len)
 {
 	sctp_inithdr_t init;
 	union sctp_params addrs;
@@ -261,7 +261,7 @@ nodata:
 
 struct sctp_chunk *sctp_make_init_ack(const struct sctp_association *asoc,
 				 const struct sctp_chunk *chunk,
-				 unsigned int __nocast gfp, int unkparam_len)
+				 gfp_t gfp, int unkparam_len)
 {
 	sctp_inithdr_t initack;
 	struct sctp_chunk *retval;
@@ -1234,7 +1234,7 @@ void sctp_chunk_assign_tsn(struct sctp_chunk *chunk)
 /* Create a CLOSED association to use with an incoming packet.  */
 struct sctp_association *sctp_make_temp_asoc(const struct sctp_endpoint *ep,
 					struct sctp_chunk *chunk,
-					unsigned int __nocast gfp)
+					gfp_t gfp)
 {
 	struct sctp_association *asoc;
 	struct sk_buff *skb;
@@ -1349,7 +1349,7 @@ nodata:
 struct sctp_association *sctp_unpack_cookie(
 	const struct sctp_endpoint *ep,
 	const struct sctp_association *asoc,
-	struct sctp_chunk *chunk, unsigned int __nocast gfp,
+	struct sctp_chunk *chunk, gfp_t gfp,
 	int *error, struct sctp_chunk **errp)
 {
 	struct sctp_association *retval = NULL;
@@ -1814,7 +1814,7 @@ int sctp_verify_init(const struct sctp_association *asoc,
  */
 int sctp_process_init(struct sctp_association *asoc, sctp_cid_t cid,
 		      const union sctp_addr *peer_addr,
-		      sctp_init_chunk_t *peer_init, unsigned int __nocast gfp)
+		      sctp_init_chunk_t *peer_init, gfp_t gfp)
 {
 	union sctp_params param;
 	struct sctp_transport *transport;
@@ -1985,7 +1985,7 @@ nomem:
 static int sctp_process_param(struct sctp_association *asoc,
 			      union sctp_params param,
 			      const union sctp_addr *peer_addr,
-			      unsigned int __nocast gfp)
+			      gfp_t gfp)
 {
 	union sctp_addr addr;
 	int i;
diff --git a/net/sctp/sm_sideeffect.c b/net/sctp/sm_sideeffect.c
index 39c970b5b19..f84173ea8ec 100644
--- a/net/sctp/sm_sideeffect.c
+++ b/net/sctp/sm_sideeffect.c
@@ -63,7 +63,7 @@ static int sctp_cmd_interpreter(sctp_event_t event_type,
 				void *event_arg,
 			 	sctp_disposition_t status,
 				sctp_cmd_seq_t *commands,
-				unsigned int __nocast gfp);
+				gfp_t gfp);
 static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
 			     sctp_state_t state,
 			     struct sctp_endpoint *ep,
@@ -71,7 +71,7 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
 			     void *event_arg,
 			     sctp_disposition_t status,
 			     sctp_cmd_seq_t *commands,
-			     unsigned int __nocast gfp);
+			     gfp_t gfp);
 
 /********************************************************************
  * Helper functions
@@ -498,7 +498,7 @@ static int sctp_cmd_process_init(sctp_cmd_seq_t *commands,
 				 struct sctp_association *asoc,
 				 struct sctp_chunk *chunk,
 				 sctp_init_chunk_t *peer_init,
-				 unsigned int __nocast gfp)
+				 gfp_t gfp)
 {
 	int error;
 
@@ -853,7 +853,7 @@ int sctp_do_sm(sctp_event_t event_type, sctp_subtype_t subtype,
 	       struct sctp_endpoint *ep,
 	       struct sctp_association *asoc,
 	       void *event_arg,
-	       unsigned int __nocast gfp)
+	       gfp_t gfp)
 {
 	sctp_cmd_seq_t commands;
 	const sctp_sm_table_entry_t *state_fn;
@@ -898,7 +898,7 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
 			     void *event_arg,
 			     sctp_disposition_t status,
 			     sctp_cmd_seq_t *commands,
-			     unsigned int __nocast gfp)
+			     gfp_t gfp)
 {
 	int error;
 
@@ -986,7 +986,7 @@ static int sctp_cmd_interpreter(sctp_event_t event_type,
 				void *event_arg,
 			 	sctp_disposition_t status,
 				sctp_cmd_seq_t *commands,
-				unsigned int __nocast gfp)
+				gfp_t gfp)
 {
 	int error = 0;
 	int force;
diff --git a/net/sctp/sm_statefuns.c b/net/sctp/sm_statefuns.c
index 86073df418f..505c7de10c5 100644
--- a/net/sctp/sm_statefuns.c
+++ b/net/sctp/sm_statefuns.c
@@ -2414,6 +2414,17 @@ sctp_disposition_t sctp_sf_do_9_2_shutdown(const struct sctp_endpoint *ep,
 	skb_pull(chunk->skb, sizeof(sctp_shutdownhdr_t));
 	chunk->subh.shutdown_hdr = sdh;
 
+	/* API 5.3.1.5 SCTP_SHUTDOWN_EVENT
+	 * When a peer sends a SHUTDOWN, SCTP delivers this notification to
+	 * inform the application that it should cease sending data.
+	 */
+	ev = sctp_ulpevent_make_shutdown_event(asoc, 0, GFP_ATOMIC);
+	if (!ev) {
+		disposition = SCTP_DISPOSITION_NOMEM;
+		goto out;	
+	}
+	sctp_add_cmd_sf(commands, SCTP_CMD_EVENT_ULP, SCTP_ULPEVENT(ev));
+
 	/* Upon the reception of the SHUTDOWN, the peer endpoint shall
 	 *  - enter the SHUTDOWN-RECEIVED state,
 	 *  - stop accepting new data from its SCTP user
@@ -2439,17 +2450,6 @@ sctp_disposition_t sctp_sf_do_9_2_shutdown(const struct sctp_endpoint *ep,
 	sctp_add_cmd_sf(commands, SCTP_CMD_PROCESS_CTSN,
 			SCTP_U32(chunk->subh.shutdown_hdr->cum_tsn_ack));
 
-	/* API 5.3.1.5 SCTP_SHUTDOWN_EVENT
-	 * When a peer sends a SHUTDOWN, SCTP delivers this notification to
-	 * inform the application that it should cease sending data.
-	 */
-	ev = sctp_ulpevent_make_shutdown_event(asoc, 0, GFP_ATOMIC);
-	if (!ev) {
-		disposition = SCTP_DISPOSITION_NOMEM;
-		goto out;	
-	}
-	sctp_add_cmd_sf(commands, SCTP_CMD_EVENT_ULP, SCTP_ULPEVENT(ev));
-
 out:
 	return disposition;
 }
diff --git a/net/sctp/socket.c b/net/sctp/socket.c
index 91ec8c93691..02e068d3450 100644
--- a/net/sctp/socket.c
+++ b/net/sctp/socket.c
@@ -3159,8 +3159,9 @@ static int sctp_getsockopt_initmsg(struct sock *sk, int len, char __user *optval
 	return 0;
 }
 
-static int sctp_getsockopt_peer_addrs_num(struct sock *sk, int len,
-					  char __user *optval, int __user *optlen)
+static int sctp_getsockopt_peer_addrs_num_old(struct sock *sk, int len,
+					      char __user *optval,
+					      int __user *optlen)
 {
 	sctp_assoc_t id;
 	struct sctp_association *asoc;
@@ -3185,23 +3186,28 @@ static int sctp_getsockopt_peer_addrs_num(struct sock *sk, int len,
 	return cnt;
 }
 
-static int sctp_getsockopt_peer_addrs(struct sock *sk, int len,
-				      char __user *optval, int __user *optlen)
+/* 
+ * Old API for getting list of peer addresses. Does not work for 32-bit
+ * programs running on a 64-bit kernel
+ */
+static int sctp_getsockopt_peer_addrs_old(struct sock *sk, int len,
+					  char __user *optval,
+					  int __user *optlen)
 {
 	struct sctp_association *asoc;
 	struct list_head *pos;
 	int cnt = 0;
-	struct sctp_getaddrs getaddrs;
+	struct sctp_getaddrs_old getaddrs;
 	struct sctp_transport *from;
 	void __user *to;
 	union sctp_addr temp;
 	struct sctp_sock *sp = sctp_sk(sk);
 	int addrlen;
 
-	if (len != sizeof(struct sctp_getaddrs))
+	if (len != sizeof(struct sctp_getaddrs_old))
 		return -EINVAL;
 
-	if (copy_from_user(&getaddrs, optval, sizeof(struct sctp_getaddrs)))
+	if (copy_from_user(&getaddrs, optval, sizeof(struct sctp_getaddrs_old)))
 		return -EFAULT;
 
 	if (getaddrs.addr_num <= 0) return -EINVAL;
@@ -3225,15 +3231,69 @@ static int sctp_getsockopt_peer_addrs(struct sock *sk, int len,
 		if (cnt >= getaddrs.addr_num) break;
 	}
 	getaddrs.addr_num = cnt;
-	if (copy_to_user(optval, &getaddrs, sizeof(struct sctp_getaddrs)))
+	if (copy_to_user(optval, &getaddrs, sizeof(struct sctp_getaddrs_old)))
+		return -EFAULT;
+
+	return 0;
+}
+
+static int sctp_getsockopt_peer_addrs(struct sock *sk, int len,
+				      char __user *optval, int __user *optlen)
+{
+	struct sctp_association *asoc;
+	struct list_head *pos;
+	int cnt = 0;
+	struct sctp_getaddrs getaddrs;
+	struct sctp_transport *from;
+	void __user *to;
+	union sctp_addr temp;
+	struct sctp_sock *sp = sctp_sk(sk);
+	int addrlen;
+	size_t space_left;
+	int bytes_copied;
+
+	if (len < sizeof(struct sctp_getaddrs))
+		return -EINVAL;
+
+	if (copy_from_user(&getaddrs, optval, sizeof(struct sctp_getaddrs)))
+		return -EFAULT;
+
+	/* For UDP-style sockets, id specifies the association to query.  */
+	asoc = sctp_id2assoc(sk, getaddrs.assoc_id);
+	if (!asoc)
+		return -EINVAL;
+
+	to = optval + offsetof(struct sctp_getaddrs,addrs);
+	space_left = len - sizeof(struct sctp_getaddrs) - 
+			offsetof(struct sctp_getaddrs,addrs);
+
+	list_for_each(pos, &asoc->peer.transport_addr_list) {
+		from = list_entry(pos, struct sctp_transport, transports);
+		memcpy(&temp, &from->ipaddr, sizeof(temp));
+		sctp_get_pf_specific(sk->sk_family)->addr_v4map(sp, &temp);
+		addrlen = sctp_get_af_specific(sk->sk_family)->sockaddr_len;
+		if(space_left < addrlen)
+			return -ENOMEM;
+		temp.v4.sin_port = htons(temp.v4.sin_port);
+		if (copy_to_user(to, &temp, addrlen))
+			return -EFAULT;
+		to += addrlen;
+		cnt++;
+		space_left -= addrlen;
+	}
+
+	if (put_user(cnt, &((struct sctp_getaddrs __user *)optval)->addr_num))
+		return -EFAULT;
+	bytes_copied = ((char __user *)to) - optval;
+	if (put_user(bytes_copied, optlen))
 		return -EFAULT;
 
 	return 0;
 }
 
-static int sctp_getsockopt_local_addrs_num(struct sock *sk, int len,
-						char __user *optval,
-						int __user *optlen)
+static int sctp_getsockopt_local_addrs_num_old(struct sock *sk, int len,
+					       char __user *optval,
+					       int __user *optlen)
 {
 	sctp_assoc_t id;
 	struct sctp_bind_addr *bp;
@@ -3306,8 +3366,8 @@ done:
 /* Helper function that copies local addresses to user and returns the number
  * of addresses copied.
  */
-static int sctp_copy_laddrs_to_user(struct sock *sk, __u16 port, int max_addrs,
-				    void __user *to)
+static int sctp_copy_laddrs_to_user_old(struct sock *sk, __u16 port, int max_addrs,
+					void __user *to)
 {
 	struct list_head *pos;
 	struct sctp_sockaddr_entry *addr;
@@ -3341,14 +3401,54 @@ static int sctp_copy_laddrs_to_user(struct sock *sk, __u16 port, int max_addrs,
 	return cnt;
 }
 
-static int sctp_getsockopt_local_addrs(struct sock *sk, int len,
-				       char __user *optval, int __user *optlen)
+static int sctp_copy_laddrs_to_user(struct sock *sk, __u16 port,
+				    void * __user *to, size_t space_left)
+{
+	struct list_head *pos;
+	struct sctp_sockaddr_entry *addr;
+	unsigned long flags;
+	union sctp_addr temp;
+	int cnt = 0;
+	int addrlen;
+
+	sctp_spin_lock_irqsave(&sctp_local_addr_lock, flags);
+	list_for_each(pos, &sctp_local_addr_list) {
+		addr = list_entry(pos, struct sctp_sockaddr_entry, list);
+		if ((PF_INET == sk->sk_family) && 
+		    (AF_INET6 == addr->a.sa.sa_family))
+			continue;
+		memcpy(&temp, &addr->a, sizeof(temp));
+		sctp_get_pf_specific(sk->sk_family)->addr_v4map(sctp_sk(sk),
+								&temp);
+		addrlen = sctp_get_af_specific(temp.sa.sa_family)->sockaddr_len;
+		if(space_left<addrlen)
+			return -ENOMEM;
+		temp.v4.sin_port = htons(port);
+		if (copy_to_user(*to, &temp, addrlen)) {
+			sctp_spin_unlock_irqrestore(&sctp_local_addr_lock,
+						    flags);
+			return -EFAULT;
+		}
+		*to += addrlen;
+		cnt ++;
+		space_left -= addrlen;
+	}
+	sctp_spin_unlock_irqrestore(&sctp_local_addr_lock, flags);
+
+	return cnt;
+}
+
+/* Old API for getting list of local addresses. Does not work for 32-bit
+ * programs running on a 64-bit kernel
+ */
+static int sctp_getsockopt_local_addrs_old(struct sock *sk, int len,
+					   char __user *optval, int __user *optlen)
 {
 	struct sctp_bind_addr *bp;
 	struct sctp_association *asoc;
 	struct list_head *pos;
 	int cnt = 0;
-	struct sctp_getaddrs getaddrs;
+	struct sctp_getaddrs_old getaddrs;
 	struct sctp_sockaddr_entry *addr;
 	void __user *to;
 	union sctp_addr temp;
@@ -3357,10 +3457,10 @@ static int sctp_getsockopt_local_addrs(struct sock *sk, int len,
 	rwlock_t *addr_lock;
 	int err = 0;
 
-	if (len != sizeof(struct sctp_getaddrs))
+	if (len != sizeof(struct sctp_getaddrs_old))
 		return -EINVAL;
 
-	if (copy_from_user(&getaddrs, optval, sizeof(struct sctp_getaddrs)))
+	if (copy_from_user(&getaddrs, optval, sizeof(struct sctp_getaddrs_old)))
 		return -EFAULT;
 
 	if (getaddrs.addr_num <= 0) return -EINVAL;
@@ -3392,8 +3492,9 @@ static int sctp_getsockopt_local_addrs(struct sock *sk, int len,
 		addr = list_entry(bp->address_list.next,
 				  struct sctp_sockaddr_entry, list);
 		if (sctp_is_any(&addr->a)) {
-			cnt = sctp_copy_laddrs_to_user(sk, bp->port,
-						       getaddrs.addr_num, to);
+			cnt = sctp_copy_laddrs_to_user_old(sk, bp->port,
+							   getaddrs.addr_num,
+							   to);
 			if (cnt < 0) {
 				err = cnt;
 				goto unlock;
@@ -3419,7 +3520,7 @@ static int sctp_getsockopt_local_addrs(struct sock *sk, int len,
 
 copy_getaddrs:
 	getaddrs.addr_num = cnt;
-	if (copy_to_user(optval, &getaddrs, sizeof(struct sctp_getaddrs)))
+	if (copy_to_user(optval, &getaddrs, sizeof(struct sctp_getaddrs_old)))
 		err = -EFAULT;
 
 unlock:
@@ -3427,6 +3528,99 @@ unlock:
 	return err;
 }
 
+static int sctp_getsockopt_local_addrs(struct sock *sk, int len,
+				       char __user *optval, int __user *optlen)
+{
+	struct sctp_bind_addr *bp;
+	struct sctp_association *asoc;
+	struct list_head *pos;
+	int cnt = 0;
+	struct sctp_getaddrs getaddrs;
+	struct sctp_sockaddr_entry *addr;
+	void __user *to;
+	union sctp_addr temp;
+	struct sctp_sock *sp = sctp_sk(sk);
+	int addrlen;
+	rwlock_t *addr_lock;
+	int err = 0;
+	size_t space_left;
+	int bytes_copied;
+
+	if (len <= sizeof(struct sctp_getaddrs))
+		return -EINVAL;
+
+	if (copy_from_user(&getaddrs, optval, sizeof(struct sctp_getaddrs)))
+		return -EFAULT;
+
+	/*
+	 *  For UDP-style sockets, id specifies the association to query.
+	 *  If the id field is set to the value '0' then the locally bound
+	 *  addresses are returned without regard to any particular
+	 *  association.
+	 */
+	if (0 == getaddrs.assoc_id) {
+		bp = &sctp_sk(sk)->ep->base.bind_addr;
+		addr_lock = &sctp_sk(sk)->ep->base.addr_lock;
+	} else {
+		asoc = sctp_id2assoc(sk, getaddrs.assoc_id);
+		if (!asoc)
+			return -EINVAL;
+		bp = &asoc->base.bind_addr;
+		addr_lock = &asoc->base.addr_lock;
+	}
+
+	to = optval + offsetof(struct sctp_getaddrs,addrs);
+	space_left = len - sizeof(struct sctp_getaddrs) -
+			 offsetof(struct sctp_getaddrs,addrs);
+
+	sctp_read_lock(addr_lock);
+
+	/* If the endpoint is bound to 0.0.0.0 or ::0, get the valid
+	 * addresses from the global local address list.
+	 */
+	if (sctp_list_single_entry(&bp->address_list)) {
+		addr = list_entry(bp->address_list.next,
+				  struct sctp_sockaddr_entry, list);
+		if (sctp_is_any(&addr->a)) {
+			cnt = sctp_copy_laddrs_to_user(sk, bp->port,
+						       &to, space_left);
+			if (cnt < 0) {
+				err = cnt;
+				goto unlock;
+			}
+			goto copy_getaddrs;		
+		}
+	}
+
+	list_for_each(pos, &bp->address_list) {
+		addr = list_entry(pos, struct sctp_sockaddr_entry, list);
+		memcpy(&temp, &addr->a, sizeof(temp));
+		sctp_get_pf_specific(sk->sk_family)->addr_v4map(sp, &temp);
+		addrlen = sctp_get_af_specific(temp.sa.sa_family)->sockaddr_len;
+		if(space_left < addrlen)
+			return -ENOMEM; /*fixme: right error?*/
+		temp.v4.sin_port = htons(temp.v4.sin_port);
+		if (copy_to_user(to, &temp, addrlen)) {
+			err = -EFAULT;
+			goto unlock;
+		}
+		to += addrlen;
+		cnt ++;
+		space_left -= addrlen;
+	}
+
+copy_getaddrs:
+	if (put_user(cnt, &((struct sctp_getaddrs __user *)optval)->addr_num))
+		return -EFAULT;
+	bytes_copied = ((char __user *)to) - optval;
+	if (put_user(bytes_copied, optlen))
+		return -EFAULT;
+
+unlock:
+	sctp_read_unlock(addr_lock);
+	return err;
+}
+
 /* 7.1.10 Set Primary Address (SCTP_PRIMARY_ADDR)
  *
  * Requests that the local SCTP stack use the enclosed peer address as
@@ -3807,12 +4001,20 @@ SCTP_STATIC int sctp_getsockopt(struct sock *sk, int level, int optname,
 	case SCTP_INITMSG:
 		retval = sctp_getsockopt_initmsg(sk, len, optval, optlen);
 		break;
-	case SCTP_GET_PEER_ADDRS_NUM:
-		retval = sctp_getsockopt_peer_addrs_num(sk, len, optval,
+	case SCTP_GET_PEER_ADDRS_NUM_OLD:
+		retval = sctp_getsockopt_peer_addrs_num_old(sk, len, optval,
+							    optlen);
+		break;
+	case SCTP_GET_LOCAL_ADDRS_NUM_OLD:
+		retval = sctp_getsockopt_local_addrs_num_old(sk, len, optval,
+							     optlen);
+		break;
+	case SCTP_GET_PEER_ADDRS_OLD:
+		retval = sctp_getsockopt_peer_addrs_old(sk, len, optval,
 							optlen);
 		break;
-	case SCTP_GET_LOCAL_ADDRS_NUM:
-		retval = sctp_getsockopt_local_addrs_num(sk, len, optval,
+	case SCTP_GET_LOCAL_ADDRS_OLD:
+		retval = sctp_getsockopt_local_addrs_old(sk, len, optval,
 							 optlen);
 		break;
 	case SCTP_GET_PEER_ADDRS:
diff --git a/net/sctp/ssnmap.c b/net/sctp/ssnmap.c
index 25037daf3fa..cbe2513d282 100644
--- a/net/sctp/ssnmap.c
+++ b/net/sctp/ssnmap.c
@@ -58,7 +58,7 @@ static inline size_t sctp_ssnmap_size(__u16 in, __u16 out)
  * Allocate room to store at least 'len' contiguous TSNs.
  */
 struct sctp_ssnmap *sctp_ssnmap_new(__u16 in, __u16 out,
-				    unsigned int __nocast gfp)
+				    gfp_t gfp)
 {
 	struct sctp_ssnmap *retval;
 	int size;
diff --git a/net/sctp/transport.c b/net/sctp/transport.c
index d2f04ebe508..6bc27200e6c 100644
--- a/net/sctp/transport.c
+++ b/net/sctp/transport.c
@@ -57,7 +57,7 @@
 /* Initialize a new transport from provided memory.  */
 static struct sctp_transport *sctp_transport_init(struct sctp_transport *peer,
 						  const union sctp_addr *addr,
-						  unsigned int __nocast gfp)
+						  gfp_t gfp)
 {
 	/* Copy in the address.  */
 	peer->ipaddr = *addr;
@@ -122,7 +122,7 @@ static struct sctp_transport *sctp_transport_init(struct sctp_transport *peer,
 
 /* Allocate and initialize a new transport.  */
 struct sctp_transport *sctp_transport_new(const union sctp_addr *addr,
-					  unsigned int __nocast gfp)
+					  gfp_t gfp)
 {
         struct sctp_transport *transport;
 
diff --git a/net/sctp/ulpevent.c b/net/sctp/ulpevent.c
index 0abd5101107..057e7fac3af 100644
--- a/net/sctp/ulpevent.c
+++ b/net/sctp/ulpevent.c
@@ -74,7 +74,7 @@ SCTP_STATIC void sctp_ulpevent_init(struct sctp_ulpevent *event, int msg_flags)
 
 /* Create a new sctp_ulpevent.  */
 SCTP_STATIC struct sctp_ulpevent *sctp_ulpevent_new(int size, int msg_flags,
-						    unsigned int __nocast gfp)
+						    gfp_t gfp)
 {
 	struct sctp_ulpevent *event;
 	struct sk_buff *skb;
@@ -136,7 +136,7 @@ static inline void sctp_ulpevent_release_owner(struct sctp_ulpevent *event)
 struct sctp_ulpevent  *sctp_ulpevent_make_assoc_change(
 	const struct sctp_association *asoc,
 	__u16 flags, __u16 state, __u16 error, __u16 outbound,
-	__u16 inbound, unsigned int __nocast gfp)
+	__u16 inbound, gfp_t gfp)
 {
 	struct sctp_ulpevent *event;
 	struct sctp_assoc_change *sac;
@@ -237,7 +237,7 @@ fail:
 struct sctp_ulpevent *sctp_ulpevent_make_peer_addr_change(
 	const struct sctp_association *asoc,
 	const struct sockaddr_storage *aaddr,
-	int flags, int state, int error, unsigned int __nocast gfp)
+	int flags, int state, int error, gfp_t gfp)
 {
 	struct sctp_ulpevent *event;
 	struct sctp_paddr_change  *spc;
@@ -350,7 +350,7 @@ fail:
  */
 struct sctp_ulpevent *sctp_ulpevent_make_remote_error(
 	const struct sctp_association *asoc, struct sctp_chunk *chunk,
-	__u16 flags, unsigned int __nocast gfp)
+	__u16 flags, gfp_t gfp)
 {
 	struct sctp_ulpevent *event;
 	struct sctp_remote_error *sre;
@@ -448,7 +448,7 @@ fail:
  */
 struct sctp_ulpevent *sctp_ulpevent_make_send_failed(
 	const struct sctp_association *asoc, struct sctp_chunk *chunk,
-	__u16 flags, __u32 error, unsigned int __nocast gfp)
+	__u16 flags, __u32 error, gfp_t gfp)
 {
 	struct sctp_ulpevent *event;
 	struct sctp_send_failed *ssf;
@@ -557,7 +557,7 @@ fail:
  */
 struct sctp_ulpevent *sctp_ulpevent_make_shutdown_event(
 	const struct sctp_association *asoc,
-	__u16 flags, unsigned int __nocast gfp)
+	__u16 flags, gfp_t gfp)
 {
 	struct sctp_ulpevent *event;
 	struct sctp_shutdown_event *sse;
@@ -620,7 +620,7 @@ fail:
  * 5.3.1.6 SCTP_ADAPTION_INDICATION
  */
 struct sctp_ulpevent *sctp_ulpevent_make_adaption_indication(
-	const struct sctp_association *asoc, unsigned int __nocast gfp)
+	const struct sctp_association *asoc, gfp_t gfp)
 {
 	struct sctp_ulpevent *event;
 	struct sctp_adaption_event *sai;
@@ -657,7 +657,7 @@ fail:
  */
 struct sctp_ulpevent *sctp_ulpevent_make_rcvmsg(struct sctp_association *asoc,
 						struct sctp_chunk *chunk,
-						unsigned int __nocast gfp)
+						gfp_t gfp)
 {
 	struct sctp_ulpevent *event = NULL;
 	struct sk_buff *skb;
@@ -719,7 +719,7 @@ fail:
  */
 struct sctp_ulpevent *sctp_ulpevent_make_pdapi(
 	const struct sctp_association *asoc, __u32 indication,
-	unsigned int __nocast gfp)
+	gfp_t gfp)
 {
 	struct sctp_ulpevent *event;
 	struct sctp_pdapi_event *pd;
diff --git a/net/sctp/ulpqueue.c b/net/sctp/ulpqueue.c
index ec2c857eae7..2080b2d28c9 100644
--- a/net/sctp/ulpqueue.c
+++ b/net/sctp/ulpqueue.c
@@ -100,7 +100,7 @@ void sctp_ulpq_free(struct sctp_ulpq *ulpq)
 
 /* Process an incoming DATA chunk.  */
 int sctp_ulpq_tail_data(struct sctp_ulpq *ulpq, struct sctp_chunk *chunk,
-			unsigned int __nocast gfp)
+			gfp_t gfp)
 {
 	struct sk_buff_head temp;
 	sctp_data_chunk_t *hdr;
@@ -792,7 +792,7 @@ static __u16 sctp_ulpq_renege_frags(struct sctp_ulpq *ulpq, __u16 needed)
 /* Partial deliver the first message as there is pressure on rwnd. */
 void sctp_ulpq_partial_delivery(struct sctp_ulpq *ulpq,
 				struct sctp_chunk *chunk,
-				unsigned int __nocast gfp)
+				gfp_t gfp)
 {
 	struct sctp_ulpevent *event;
 	struct sctp_association *asoc;
@@ -816,7 +816,7 @@ void sctp_ulpq_partial_delivery(struct sctp_ulpq *ulpq,
 
 /* Renege some packets to make room for an incoming chunk.  */
 void sctp_ulpq_renege(struct sctp_ulpq *ulpq, struct sctp_chunk *chunk,
-		      unsigned int __nocast gfp)
+		      gfp_t gfp)
 {
 	struct sctp_association *asoc;
 	__u16 needed, freed;
@@ -855,7 +855,7 @@ void sctp_ulpq_renege(struct sctp_ulpq *ulpq, struct sctp_chunk *chunk,
 /* Notify the application if an association is aborted and in
  * partial delivery mode.  Send up any pending received messages.
  */
-void sctp_ulpq_abort_pd(struct sctp_ulpq *ulpq, unsigned int __nocast gfp)
+void sctp_ulpq_abort_pd(struct sctp_ulpq *ulpq, gfp_t gfp)
 {
 	struct sctp_ulpevent *ev = NULL;
 	struct sock *sk;
diff --git a/net/socket.c b/net/socket.c
index f9264472377..3145103cdf5 100644
--- a/net/socket.c
+++ b/net/socket.c
@@ -1145,8 +1145,11 @@ static int __sock_create(int family, int type, int protocol, struct socket **res
 	if (!try_module_get(net_families[family]->owner))
 		goto out_release;
 
-	if ((err = net_families[family]->create(sock, protocol)) < 0)
+	if ((err = net_families[family]->create(sock, protocol)) < 0) {
+		sock->ops = NULL;
 		goto out_module_put;
+	}
+
 	/*
 	 * Now to bump the refcnt of the [loadable] module that owns this
 	 * socket at sock_release time we decrement its refcnt.
@@ -1360,16 +1363,16 @@ asmlinkage long sys_accept(int fd, struct sockaddr __user *upeer_sockaddr, int _
 	newsock->type = sock->type;
 	newsock->ops = sock->ops;
 
-	err = security_socket_accept(sock, newsock);
-	if (err)
-		goto out_release;
-
 	/*
 	 * We don't need try_module_get here, as the listening socket (sock)
 	 * has the protocol module (sock->ops->owner) held.
 	 */
 	__module_get(newsock->ops->owner);
 
+	err = security_socket_accept(sock, newsock);
+	if (err)
+		goto out_release;
+
 	err = sock->ops->accept(sock, newsock, sock->file->f_flags);
 	if (err < 0)
 		goto out_release;
@@ -1700,7 +1703,9 @@ asmlinkage long sys_sendmsg(int fd, struct msghdr __user *msg, unsigned flags)
 	struct socket *sock;
 	char address[MAX_SOCK_ADDR];
 	struct iovec iovstack[UIO_FASTIOV], *iov = iovstack;
-	unsigned char ctl[sizeof(struct cmsghdr) + 20];	/* 20 is size of ipv6_pktinfo */
+	unsigned char ctl[sizeof(struct cmsghdr) + 20]
+			__attribute__ ((aligned (sizeof(__kernel_size_t))));
+			/* 20 is size of ipv6_pktinfo */
 	unsigned char *ctl_buf = ctl;
 	struct msghdr msg_sys;
 	int err, ctl_len, iov_size, total_len;
diff --git a/net/sunrpc/Makefile b/net/sunrpc/Makefile
index 46a2ce00a29..cdcab9ca4c6 100644
--- a/net/sunrpc/Makefile
+++ b/net/sunrpc/Makefile
@@ -6,7 +6,7 @@
 obj-$(CONFIG_SUNRPC) += sunrpc.o
 obj-$(CONFIG_SUNRPC_GSS) += auth_gss/
 
-sunrpc-y := clnt.o xprt.o sched.o \
+sunrpc-y := clnt.o xprt.o socklib.o xprtsock.o sched.o \
 	    auth.o auth_null.o auth_unix.o \
 	    svc.o svcsock.o svcauth.o svcauth_unix.o \
 	    pmap_clnt.o timer.o xdr.o \
diff --git a/net/sunrpc/auth.c b/net/sunrpc/auth.c
index 505e2d4b3d6..a415d99c394 100644
--- a/net/sunrpc/auth.c
+++ b/net/sunrpc/auth.c
@@ -11,7 +11,6 @@
 #include <linux/module.h>
 #include <linux/slab.h>
 #include <linux/errno.h>
-#include <linux/socket.h>
 #include <linux/sunrpc/clnt.h>
 #include <linux/spinlock.h>
 
diff --git a/net/sunrpc/auth_gss/Makefile b/net/sunrpc/auth_gss/Makefile
index fe1b874084b..f3431a7e33d 100644
--- a/net/sunrpc/auth_gss/Makefile
+++ b/net/sunrpc/auth_gss/Makefile
@@ -10,7 +10,7 @@ auth_rpcgss-objs := auth_gss.o gss_generic_token.o \
 obj-$(CONFIG_RPCSEC_GSS_KRB5) += rpcsec_gss_krb5.o
 
 rpcsec_gss_krb5-objs := gss_krb5_mech.o gss_krb5_seal.o gss_krb5_unseal.o \
-	gss_krb5_seqnum.o
+	gss_krb5_seqnum.o gss_krb5_wrap.o
 
 obj-$(CONFIG_RPCSEC_GSS_SPKM3) += rpcsec_gss_spkm3.o
 
diff --git a/net/sunrpc/auth_gss/auth_gss.c b/net/sunrpc/auth_gss/auth_gss.c
index 2f7b867161d..f44f46f1d8e 100644
--- a/net/sunrpc/auth_gss/auth_gss.c
+++ b/net/sunrpc/auth_gss/auth_gss.c
@@ -42,9 +42,8 @@
 #include <linux/init.h>
 #include <linux/types.h>
 #include <linux/slab.h>
-#include <linux/socket.h>
-#include <linux/in.h>
 #include <linux/sched.h>
+#include <linux/pagemap.h>
 #include <linux/sunrpc/clnt.h>
 #include <linux/sunrpc/auth.h>
 #include <linux/sunrpc/auth_gss.h>
@@ -846,10 +845,8 @@ gss_marshal(struct rpc_task *task, u32 *p)
 
 	/* We compute the checksum for the verifier over the xdr-encoded bytes
 	 * starting with the xid and ending at the end of the credential: */
-	iov.iov_base = req->rq_snd_buf.head[0].iov_base;
-	if (task->tk_client->cl_xprt->stream)
-		/* See clnt.c:call_header() */
-		iov.iov_base += 4;
+	iov.iov_base = xprt_skip_transport_header(task->tk_xprt,
+					req->rq_snd_buf.head[0].iov_base);
 	iov.iov_len = (u8 *)p - (u8 *)iov.iov_base;
 	xdr_buf_from_iov(&iov, &verf_buf);
 
@@ -857,9 +854,7 @@ gss_marshal(struct rpc_task *task, u32 *p)
 	*p++ = htonl(RPC_AUTH_GSS);
 
 	mic.data = (u8 *)(p + 1);
-	maj_stat = gss_get_mic(ctx->gc_gss_ctx,
-			       GSS_C_QOP_DEFAULT, 
-			       &verf_buf, &mic);
+	maj_stat = gss_get_mic(ctx->gc_gss_ctx, &verf_buf, &mic);
 	if (maj_stat == GSS_S_CONTEXT_EXPIRED) {
 		cred->cr_flags &= ~RPCAUTH_CRED_UPTODATE;
 	} else if (maj_stat != 0) {
@@ -890,10 +885,8 @@ static u32 *
 gss_validate(struct rpc_task *task, u32 *p)
 {
 	struct rpc_cred *cred = task->tk_msg.rpc_cred;
-	struct gss_cred	*gss_cred = container_of(cred, struct gss_cred,
-						gc_base);
 	struct gss_cl_ctx *ctx = gss_cred_get_ctx(cred);
-	u32		seq, qop_state;
+	u32		seq;
 	struct kvec	iov;
 	struct xdr_buf	verf_buf;
 	struct xdr_netobj mic;
@@ -914,23 +907,14 @@ gss_validate(struct rpc_task *task, u32 *p)
 	mic.data = (u8 *)p;
 	mic.len = len;
 
-	maj_stat = gss_verify_mic(ctx->gc_gss_ctx, &verf_buf, &mic, &qop_state);
+	maj_stat = gss_verify_mic(ctx->gc_gss_ctx, &verf_buf, &mic);
 	if (maj_stat == GSS_S_CONTEXT_EXPIRED)
 		cred->cr_flags &= ~RPCAUTH_CRED_UPTODATE;
 	if (maj_stat)
 		goto out_bad;
-       switch (gss_cred->gc_service) {
-       case RPC_GSS_SVC_NONE:
-	       /* verifier data, flavor, length: */
-	       task->tk_auth->au_rslack = XDR_QUADLEN(len) + 2;
-	       break;
-       case RPC_GSS_SVC_INTEGRITY:
-	       /* verifier data, flavor, length, length, sequence number: */
-	       task->tk_auth->au_rslack = XDR_QUADLEN(len) + 4;
-	       break;
-       case RPC_GSS_SVC_PRIVACY:
-	       goto out_bad;
-       }
+	/* We leave it to unwrap to calculate au_rslack. For now we just
+	 * calculate the length of the verifier: */
+	task->tk_auth->au_verfsize = XDR_QUADLEN(len) + 2;
 	gss_put_ctx(ctx);
 	dprintk("RPC: %4u GSS gss_validate: gss_verify_mic succeeded.\n",
 			task->tk_pid);
@@ -975,8 +959,7 @@ gss_wrap_req_integ(struct rpc_cred *cred, struct gss_cl_ctx *ctx,
 	p = iov->iov_base + iov->iov_len;
 	mic.data = (u8 *)(p + 1);
 
-	maj_stat = gss_get_mic(ctx->gc_gss_ctx,
-			GSS_C_QOP_DEFAULT, &integ_buf, &mic);
+	maj_stat = gss_get_mic(ctx->gc_gss_ctx, &integ_buf, &mic);
 	status = -EIO; /* XXX? */
 	if (maj_stat == GSS_S_CONTEXT_EXPIRED)
 		cred->cr_flags &= ~RPCAUTH_CRED_UPTODATE;
@@ -990,6 +973,113 @@ gss_wrap_req_integ(struct rpc_cred *cred, struct gss_cl_ctx *ctx,
 	return 0;
 }
 
+static void
+priv_release_snd_buf(struct rpc_rqst *rqstp)
+{
+	int i;
+
+	for (i=0; i < rqstp->rq_enc_pages_num; i++)
+		__free_page(rqstp->rq_enc_pages[i]);
+	kfree(rqstp->rq_enc_pages);
+}
+
+static int
+alloc_enc_pages(struct rpc_rqst *rqstp)
+{
+	struct xdr_buf *snd_buf = &rqstp->rq_snd_buf;
+	int first, last, i;
+
+	if (snd_buf->page_len == 0) {
+		rqstp->rq_enc_pages_num = 0;
+		return 0;
+	}
+
+	first = snd_buf->page_base >> PAGE_CACHE_SHIFT;
+	last = (snd_buf->page_base + snd_buf->page_len - 1) >> PAGE_CACHE_SHIFT;
+	rqstp->rq_enc_pages_num = last - first + 1 + 1;
+	rqstp->rq_enc_pages
+		= kmalloc(rqstp->rq_enc_pages_num * sizeof(struct page *),
+				GFP_NOFS);
+	if (!rqstp->rq_enc_pages)
+		goto out;
+	for (i=0; i < rqstp->rq_enc_pages_num; i++) {
+		rqstp->rq_enc_pages[i] = alloc_page(GFP_NOFS);
+		if (rqstp->rq_enc_pages[i] == NULL)
+			goto out_free;
+	}
+	rqstp->rq_release_snd_buf = priv_release_snd_buf;
+	return 0;
+out_free:
+	for (i--; i >= 0; i--) {
+		__free_page(rqstp->rq_enc_pages[i]);
+	}
+out:
+	return -EAGAIN;
+}
+
+static inline int
+gss_wrap_req_priv(struct rpc_cred *cred, struct gss_cl_ctx *ctx,
+		kxdrproc_t encode, struct rpc_rqst *rqstp, u32 *p, void *obj)
+{
+	struct xdr_buf	*snd_buf = &rqstp->rq_snd_buf;
+	u32		offset;
+	u32             maj_stat;
+	int		status;
+	u32		*opaque_len;
+	struct page	**inpages;
+	int		first;
+	int		pad;
+	struct kvec	*iov;
+	char		*tmp;
+
+	opaque_len = p++;
+	offset = (u8 *)p - (u8 *)snd_buf->head[0].iov_base;
+	*p++ = htonl(rqstp->rq_seqno);
+
+	status = encode(rqstp, p, obj);
+	if (status)
+		return status;
+
+	status = alloc_enc_pages(rqstp);
+	if (status)
+		return status;
+	first = snd_buf->page_base >> PAGE_CACHE_SHIFT;
+	inpages = snd_buf->pages + first;
+	snd_buf->pages = rqstp->rq_enc_pages;
+	snd_buf->page_base -= first << PAGE_CACHE_SHIFT;
+	/* Give the tail its own page, in case we need extra space in the
+	 * head when wrapping: */
+	if (snd_buf->page_len || snd_buf->tail[0].iov_len) {
+		tmp = page_address(rqstp->rq_enc_pages[rqstp->rq_enc_pages_num - 1]);
+		memcpy(tmp, snd_buf->tail[0].iov_base, snd_buf->tail[0].iov_len);
+		snd_buf->tail[0].iov_base = tmp;
+	}
+	maj_stat = gss_wrap(ctx->gc_gss_ctx, offset, snd_buf, inpages);
+	/* RPC_SLACK_SPACE should prevent this ever happening: */
+	BUG_ON(snd_buf->len > snd_buf->buflen);
+        status = -EIO;
+	/* We're assuming that when GSS_S_CONTEXT_EXPIRED, the encryption was
+	 * done anyway, so it's safe to put the request on the wire: */
+	if (maj_stat == GSS_S_CONTEXT_EXPIRED)
+		cred->cr_flags &= ~RPCAUTH_CRED_UPTODATE;
+	else if (maj_stat)
+		return status;
+
+	*opaque_len = htonl(snd_buf->len - offset);
+	/* guess whether we're in the head or the tail: */
+	if (snd_buf->page_len || snd_buf->tail[0].iov_len)
+		iov = snd_buf->tail;
+	else
+		iov = snd_buf->head;
+	p = iov->iov_base + iov->iov_len;
+	pad = 3 - ((snd_buf->len - offset - 1) & 3);
+	memset(p, 0, pad);
+	iov->iov_len += pad;
+	snd_buf->len += pad;
+
+	return 0;
+}
+
 static int
 gss_wrap_req(struct rpc_task *task,
 	     kxdrproc_t encode, void *rqstp, u32 *p, void *obj)
@@ -1017,6 +1107,8 @@ gss_wrap_req(struct rpc_task *task,
 								rqstp, p, obj);
 			break;
        		case RPC_GSS_SVC_PRIVACY:
+			status = gss_wrap_req_priv(cred, ctx, encode,
+					rqstp, p, obj);
 			break;
 	}
 out:
@@ -1054,8 +1146,7 @@ gss_unwrap_resp_integ(struct rpc_cred *cred, struct gss_cl_ctx *ctx,
 	if (xdr_buf_read_netobj(rcv_buf, &mic, mic_offset))
 		return status;
 
-	maj_stat = gss_verify_mic(ctx->gc_gss_ctx, &integ_buf,
-			&mic, NULL);
+	maj_stat = gss_verify_mic(ctx->gc_gss_ctx, &integ_buf, &mic);
 	if (maj_stat == GSS_S_CONTEXT_EXPIRED)
 		cred->cr_flags &= ~RPCAUTH_CRED_UPTODATE;
 	if (maj_stat != GSS_S_COMPLETE)
@@ -1063,6 +1154,35 @@ gss_unwrap_resp_integ(struct rpc_cred *cred, struct gss_cl_ctx *ctx,
 	return 0;
 }
 
+static inline int
+gss_unwrap_resp_priv(struct rpc_cred *cred, struct gss_cl_ctx *ctx,
+		struct rpc_rqst *rqstp, u32 **p)
+{
+	struct xdr_buf  *rcv_buf = &rqstp->rq_rcv_buf;
+	u32 offset;
+	u32 opaque_len;
+	u32 maj_stat;
+	int status = -EIO;
+
+	opaque_len = ntohl(*(*p)++);
+	offset = (u8 *)(*p) - (u8 *)rcv_buf->head[0].iov_base;
+	if (offset + opaque_len > rcv_buf->len)
+		return status;
+	/* remove padding: */
+	rcv_buf->len = offset + opaque_len;
+
+	maj_stat = gss_unwrap(ctx->gc_gss_ctx, offset, rcv_buf);
+	if (maj_stat == GSS_S_CONTEXT_EXPIRED)
+		cred->cr_flags &= ~RPCAUTH_CRED_UPTODATE;
+	if (maj_stat != GSS_S_COMPLETE)
+		return status;
+	if (ntohl(*(*p)++) != rqstp->rq_seqno)
+		return status;
+
+	return 0;
+}
+
+
 static int
 gss_unwrap_resp(struct rpc_task *task,
 		kxdrproc_t decode, void *rqstp, u32 *p, void *obj)
@@ -1071,6 +1191,9 @@ gss_unwrap_resp(struct rpc_task *task,
 	struct gss_cred *gss_cred = container_of(cred, struct gss_cred,
 			gc_base);
 	struct gss_cl_ctx *ctx = gss_cred_get_ctx(cred);
+	u32		*savedp = p;
+	struct kvec	*head = ((struct rpc_rqst *)rqstp)->rq_rcv_buf.head;
+	int		savedlen = head->iov_len;
 	int             status = -EIO;
 
 	if (ctx->gc_proc != RPC_GSS_PROC_DATA)
@@ -1084,8 +1207,14 @@ gss_unwrap_resp(struct rpc_task *task,
 				goto out;
 			break;
        		case RPC_GSS_SVC_PRIVACY:
+			status = gss_unwrap_resp_priv(cred, ctx, rqstp, &p);
+			if (status)
+				goto out;
 			break;
 	}
+	/* take into account extra slack for integrity and privacy cases: */
+	task->tk_auth->au_rslack = task->tk_auth->au_verfsize + (p - savedp)
+						+ (savedlen - head->iov_len);
 out_decode:
 	status = decode(rqstp, p, obj);
 out:
diff --git a/net/sunrpc/auth_gss/gss_krb5_crypto.c b/net/sunrpc/auth_gss/gss_krb5_crypto.c
index ee6ae74cd1b..3f3d5437f02 100644
--- a/net/sunrpc/auth_gss/gss_krb5_crypto.c
+++ b/net/sunrpc/auth_gss/gss_krb5_crypto.c
@@ -139,17 +139,91 @@ buf_to_sg(struct scatterlist *sg, char *ptr, int len) {
 	sg->length = len;
 }
 
+static int
+process_xdr_buf(struct xdr_buf *buf, int offset, int len,
+		int (*actor)(struct scatterlist *, void *), void *data)
+{
+	int i, page_len, thislen, page_offset, ret = 0;
+	struct scatterlist	sg[1];
+
+	if (offset >= buf->head[0].iov_len) {
+		offset -= buf->head[0].iov_len;
+	} else {
+		thislen = buf->head[0].iov_len - offset;
+		if (thislen > len)
+			thislen = len;
+		buf_to_sg(sg, buf->head[0].iov_base + offset, thislen);
+		ret = actor(sg, data);
+		if (ret)
+			goto out;
+		offset = 0;
+		len -= thislen;
+	}
+	if (len == 0)
+		goto out;
+
+	if (offset >= buf->page_len) {
+		offset -= buf->page_len;
+	} else {
+		page_len = buf->page_len - offset;
+		if (page_len > len)
+			page_len = len;
+		len -= page_len;
+		page_offset = (offset + buf->page_base) & (PAGE_CACHE_SIZE - 1);
+		i = (offset + buf->page_base) >> PAGE_CACHE_SHIFT;
+		thislen = PAGE_CACHE_SIZE - page_offset;
+		do {
+			if (thislen > page_len)
+				thislen = page_len;
+			sg->page = buf->pages[i];
+			sg->offset = page_offset;
+			sg->length = thislen;
+			ret = actor(sg, data);
+			if (ret)
+				goto out;
+			page_len -= thislen;
+			i++;
+			page_offset = 0;
+			thislen = PAGE_CACHE_SIZE;
+		} while (page_len != 0);
+		offset = 0;
+	}
+	if (len == 0)
+		goto out;
+
+	if (offset < buf->tail[0].iov_len) {
+		thislen = buf->tail[0].iov_len - offset;
+		if (thislen > len)
+			thislen = len;
+		buf_to_sg(sg, buf->tail[0].iov_base + offset, thislen);
+		ret = actor(sg, data);
+		len -= thislen;
+	}
+	if (len != 0)
+		ret = -EINVAL;
+out:
+	return ret;
+}
+
+static int
+checksummer(struct scatterlist *sg, void *data)
+{
+	struct crypto_tfm *tfm = (struct crypto_tfm *)data;
+
+	crypto_digest_update(tfm, sg, 1);
+
+	return 0;
+}
+
 /* checksum the plaintext data and hdrlen bytes of the token header */
 s32
 make_checksum(s32 cksumtype, char *header, int hdrlen, struct xdr_buf *body,
-		   struct xdr_netobj *cksum)
+		   int body_offset, struct xdr_netobj *cksum)
 {
 	char                            *cksumname;
 	struct crypto_tfm               *tfm = NULL; /* XXX add to ctx? */
 	struct scatterlist              sg[1];
 	u32                             code = GSS_S_FAILURE;
-	int				len, thislen, offset;
-	int				i;
 
 	switch (cksumtype) {
 		case CKSUMTYPE_RSA_MD5:
@@ -169,33 +243,8 @@ make_checksum(s32 cksumtype, char *header, int hdrlen, struct xdr_buf *body,
 	crypto_digest_init(tfm);
 	buf_to_sg(sg, header, hdrlen);
 	crypto_digest_update(tfm, sg, 1);
-	if (body->head[0].iov_len) {
-		buf_to_sg(sg, body->head[0].iov_base, body->head[0].iov_len);
-		crypto_digest_update(tfm, sg, 1);
-	}
-
-	len = body->page_len;
-	if (len != 0) {
-		offset = body->page_base & (PAGE_CACHE_SIZE - 1);
-		i = body->page_base >> PAGE_CACHE_SHIFT;
-		thislen = PAGE_CACHE_SIZE - offset;
-		do {
-			if (thislen > len)
-				thislen = len;
-			sg->page = body->pages[i];
-			sg->offset = offset;
-			sg->length = thislen;
-			crypto_digest_update(tfm, sg, 1);
-			len -= thislen;
-			i++;
-			offset = 0;
-			thislen = PAGE_CACHE_SIZE;
-		} while(len != 0);
-	}
-	if (body->tail[0].iov_len) {
-		buf_to_sg(sg, body->tail[0].iov_base, body->tail[0].iov_len);
-		crypto_digest_update(tfm, sg, 1);
-	}
+	process_xdr_buf(body, body_offset, body->len - body_offset,
+			checksummer, tfm);
 	crypto_digest_final(tfm, cksum->data);
 	code = 0;
 out:
@@ -204,3 +253,154 @@ out:
 }
 
 EXPORT_SYMBOL(make_checksum);
+
+struct encryptor_desc {
+	u8 iv[8]; /* XXX hard-coded blocksize */
+	struct crypto_tfm *tfm;
+	int pos;
+	struct xdr_buf *outbuf;
+	struct page **pages;
+	struct scatterlist infrags[4];
+	struct scatterlist outfrags[4];
+	int fragno;
+	int fraglen;
+};
+
+static int
+encryptor(struct scatterlist *sg, void *data)
+{
+	struct encryptor_desc *desc = data;
+	struct xdr_buf *outbuf = desc->outbuf;
+	struct page *in_page;
+	int thislen = desc->fraglen + sg->length;
+	int fraglen, ret;
+	int page_pos;
+
+	/* Worst case is 4 fragments: head, end of page 1, start
+	 * of page 2, tail.  Anything more is a bug. */
+	BUG_ON(desc->fragno > 3);
+	desc->infrags[desc->fragno] = *sg;
+	desc->outfrags[desc->fragno] = *sg;
+
+	page_pos = desc->pos - outbuf->head[0].iov_len;
+	if (page_pos >= 0 && page_pos < outbuf->page_len) {
+		/* pages are not in place: */
+		int i = (page_pos + outbuf->page_base) >> PAGE_CACHE_SHIFT;
+		in_page = desc->pages[i];
+	} else {
+		in_page = sg->page;
+	}
+	desc->infrags[desc->fragno].page = in_page;
+	desc->fragno++;
+	desc->fraglen += sg->length;
+	desc->pos += sg->length;
+
+	fraglen = thislen & 7; /* XXX hardcoded blocksize */
+	thislen -= fraglen;
+
+	if (thislen == 0)
+		return 0;
+
+	ret = crypto_cipher_encrypt_iv(desc->tfm, desc->outfrags, desc->infrags,
+					thislen, desc->iv);
+	if (ret)
+		return ret;
+	if (fraglen) {
+		desc->outfrags[0].page = sg->page;
+		desc->outfrags[0].offset = sg->offset + sg->length - fraglen;
+		desc->outfrags[0].length = fraglen;
+		desc->infrags[0] = desc->outfrags[0];
+		desc->infrags[0].page = in_page;
+		desc->fragno = 1;
+		desc->fraglen = fraglen;
+	} else {
+		desc->fragno = 0;
+		desc->fraglen = 0;
+	}
+	return 0;
+}
+
+int
+gss_encrypt_xdr_buf(struct crypto_tfm *tfm, struct xdr_buf *buf, int offset,
+		struct page **pages)
+{
+	int ret;
+	struct encryptor_desc desc;
+
+	BUG_ON((buf->len - offset) % crypto_tfm_alg_blocksize(tfm) != 0);
+
+	memset(desc.iv, 0, sizeof(desc.iv));
+	desc.tfm = tfm;
+	desc.pos = offset;
+	desc.outbuf = buf;
+	desc.pages = pages;
+	desc.fragno = 0;
+	desc.fraglen = 0;
+
+	ret = process_xdr_buf(buf, offset, buf->len - offset, encryptor, &desc);
+	return ret;
+}
+
+EXPORT_SYMBOL(gss_encrypt_xdr_buf);
+
+struct decryptor_desc {
+	u8 iv[8]; /* XXX hard-coded blocksize */
+	struct crypto_tfm *tfm;
+	struct scatterlist frags[4];
+	int fragno;
+	int fraglen;
+};
+
+static int
+decryptor(struct scatterlist *sg, void *data)
+{
+	struct decryptor_desc *desc = data;
+	int thislen = desc->fraglen + sg->length;
+	int fraglen, ret;
+
+	/* Worst case is 4 fragments: head, end of page 1, start
+	 * of page 2, tail.  Anything more is a bug. */
+	BUG_ON(desc->fragno > 3);
+	desc->frags[desc->fragno] = *sg;
+	desc->fragno++;
+	desc->fraglen += sg->length;
+
+	fraglen = thislen & 7; /* XXX hardcoded blocksize */
+	thislen -= fraglen;
+
+	if (thislen == 0)
+		return 0;
+
+	ret = crypto_cipher_decrypt_iv(desc->tfm, desc->frags, desc->frags,
+					thislen, desc->iv);
+	if (ret)
+		return ret;
+	if (fraglen) {
+		desc->frags[0].page = sg->page;
+		desc->frags[0].offset = sg->offset + sg->length - fraglen;
+		desc->frags[0].length = fraglen;
+		desc->fragno = 1;
+		desc->fraglen = fraglen;
+	} else {
+		desc->fragno = 0;
+		desc->fraglen = 0;
+	}
+	return 0;
+}
+
+int
+gss_decrypt_xdr_buf(struct crypto_tfm *tfm, struct xdr_buf *buf, int offset)
+{
+	struct decryptor_desc desc;
+
+	/* XXXJBF: */
+	BUG_ON((buf->len - offset) % crypto_tfm_alg_blocksize(tfm) != 0);
+
+	memset(desc.iv, 0, sizeof(desc.iv));
+	desc.tfm = tfm;
+	desc.fragno = 0;
+	desc.fraglen = 0;
+	return process_xdr_buf(buf, offset, buf->len - offset, decryptor, &desc);
+}
+
+EXPORT_SYMBOL(gss_decrypt_xdr_buf);
diff --git a/net/sunrpc/auth_gss/gss_krb5_mech.c b/net/sunrpc/auth_gss/gss_krb5_mech.c
index 606a8a82caf..5f1f806a0b1 100644
--- a/net/sunrpc/auth_gss/gss_krb5_mech.c
+++ b/net/sunrpc/auth_gss/gss_krb5_mech.c
@@ -39,7 +39,6 @@
 #include <linux/types.h>
 #include <linux/slab.h>
 #include <linux/sunrpc/auth.h>
-#include <linux/in.h>
 #include <linux/sunrpc/gss_krb5.h>
 #include <linux/sunrpc/xdr.h>
 #include <linux/crypto.h>
@@ -191,43 +190,12 @@ gss_delete_sec_context_kerberos(void *internal_ctx) {
 	kfree(kctx);
 }
 
-static u32
-gss_verify_mic_kerberos(struct gss_ctx		*ctx,
-			struct xdr_buf		*message,
-			struct xdr_netobj	*mic_token,
-			u32			*qstate) {
-	u32 maj_stat = 0;
-	int qop_state;
-	struct krb5_ctx *kctx = ctx->internal_ctx_id;
-
-	maj_stat = krb5_read_token(kctx, mic_token, message, &qop_state,
-				   KG_TOK_MIC_MSG);
-	if (!maj_stat && qop_state)
-	    *qstate = qop_state;
-
-	dprintk("RPC:      gss_verify_mic_kerberos returning %d\n", maj_stat);
-	return maj_stat;
-}
-
-static u32
-gss_get_mic_kerberos(struct gss_ctx	*ctx,
-		     u32		qop,
-		     struct xdr_buf 	*message,
-		     struct xdr_netobj	*mic_token) {
-	u32 err = 0;
-	struct krb5_ctx *kctx = ctx->internal_ctx_id;
-
-	err = krb5_make_token(kctx, qop, message, mic_token, KG_TOK_MIC_MSG);
-
-	dprintk("RPC:      gss_get_mic_kerberos returning %d\n",err);
-
-	return err;
-}
-
 static struct gss_api_ops gss_kerberos_ops = {
 	.gss_import_sec_context	= gss_import_sec_context_kerberos,
 	.gss_get_mic		= gss_get_mic_kerberos,
 	.gss_verify_mic		= gss_verify_mic_kerberos,
+	.gss_wrap		= gss_wrap_kerberos,
+	.gss_unwrap		= gss_unwrap_kerberos,
 	.gss_delete_sec_context	= gss_delete_sec_context_kerberos,
 };
 
@@ -242,6 +210,11 @@ static struct pf_desc gss_kerberos_pfs[] = {
 		.service = RPC_GSS_SVC_INTEGRITY,
 		.name = "krb5i",
 	},
+	[2] = {
+		.pseudoflavor = RPC_AUTH_GSS_KRB5P,
+		.service = RPC_GSS_SVC_PRIVACY,
+		.name = "krb5p",
+	},
 };
 
 static struct gss_api_mech gss_kerberos_mech = {
diff --git a/net/sunrpc/auth_gss/gss_krb5_seal.c b/net/sunrpc/auth_gss/gss_krb5_seal.c
index afeeb8715a7..13f8ae97945 100644
--- a/net/sunrpc/auth_gss/gss_krb5_seal.c
+++ b/net/sunrpc/auth_gss/gss_krb5_seal.c
@@ -70,22 +70,13 @@
 # define RPCDBG_FACILITY        RPCDBG_AUTH
 #endif
 
-static inline int
-gss_krb5_padding(int blocksize, int length) {
-	/* Most of the code is block-size independent but in practice we
-	 * use only 8: */
-	BUG_ON(blocksize != 8);
-	return 8 - (length & 7);
-}
-
 u32
-krb5_make_token(struct krb5_ctx *ctx, int qop_req,
-		   struct xdr_buf *text, struct xdr_netobj *token,
-		   int toktype)
+gss_get_mic_kerberos(struct gss_ctx *gss_ctx, struct xdr_buf *text,
+		struct xdr_netobj *token)
 {
+	struct krb5_ctx		*ctx = gss_ctx->internal_ctx_id;
 	s32			checksum_type;
 	struct xdr_netobj	md5cksum = {.len = 0, .data = NULL};
-	int			blocksize = 0, tmsglen;
 	unsigned char		*ptr, *krb5_hdr, *msg_start;
 	s32			now;
 
@@ -93,9 +84,6 @@ krb5_make_token(struct krb5_ctx *ctx, int qop_req,
 
 	now = get_seconds();
 
-	if (qop_req != 0)
-		goto out_err;
-
 	switch (ctx->signalg) {
 		case SGN_ALG_DES_MAC_MD5:
 			checksum_type = CKSUMTYPE_RSA_MD5;
@@ -111,21 +99,13 @@ krb5_make_token(struct krb5_ctx *ctx, int qop_req,
 		goto out_err;
 	}
 
-	if (toktype == KG_TOK_WRAP_MSG) {
-		blocksize = crypto_tfm_alg_blocksize(ctx->enc);
-		tmsglen = blocksize + text->len
-			+ gss_krb5_padding(blocksize, blocksize + text->len);
-	} else {
-		tmsglen = 0;
-	}
-
-	token->len = g_token_size(&ctx->mech_used, 22 + tmsglen);
+	token->len = g_token_size(&ctx->mech_used, 22);
 
 	ptr = token->data;
-	g_make_token_header(&ctx->mech_used, 22 + tmsglen, &ptr);
+	g_make_token_header(&ctx->mech_used, 22, &ptr);
 
-	*ptr++ = (unsigned char) ((toktype>>8)&0xff);
-	*ptr++ = (unsigned char) (toktype&0xff);
+	*ptr++ = (unsigned char) ((KG_TOK_MIC_MSG>>8)&0xff);
+	*ptr++ = (unsigned char) (KG_TOK_MIC_MSG&0xff);
 
 	/* ptr now at byte 2 of header described in rfc 1964, section 1.2.1: */
 	krb5_hdr = ptr - 2;
@@ -133,17 +113,9 @@ krb5_make_token(struct krb5_ctx *ctx, int qop_req,
 
 	*(u16 *)(krb5_hdr + 2) = htons(ctx->signalg);
 	memset(krb5_hdr + 4, 0xff, 4);
-	if (toktype == KG_TOK_WRAP_MSG)
-		*(u16 *)(krb5_hdr + 4) = htons(ctx->sealalg);
 
-	if (toktype == KG_TOK_WRAP_MSG) {
-		/* XXX removing support for now */
-		goto out_err;
-	} else { /* Sign only.  */
-		if (make_checksum(checksum_type, krb5_hdr, 8, text,
-				       &md5cksum))
+	if (make_checksum(checksum_type, krb5_hdr, 8, text, 0, &md5cksum))
 			goto out_err;
-	}
 
 	switch (ctx->signalg) {
 	case SGN_ALG_DES_MAC_MD5:
diff --git a/net/sunrpc/auth_gss/gss_krb5_unseal.c b/net/sunrpc/auth_gss/gss_krb5_unseal.c
index 8767fc53183..2030475d98e 100644
--- a/net/sunrpc/auth_gss/gss_krb5_unseal.c
+++ b/net/sunrpc/auth_gss/gss_krb5_unseal.c
@@ -68,21 +68,14 @@
 #endif
 
 
-/* message_buffer is an input if toktype is MIC and an output if it is WRAP:
- * If toktype is MIC: read_token is a mic token, and message_buffer is the
- *   data that the mic was supposedly taken over.
- * If toktype is WRAP: read_token is a wrap token, and message_buffer is used
- *   to return the decrypted data.
- */
+/* read_token is a mic token, and message_buffer is the data that the mic was
+ * supposedly taken over. */
 
-/* XXX will need to change prototype and/or just split into a separate function
- * when we add privacy (because read_token will be in pages too). */
 u32
-krb5_read_token(struct krb5_ctx *ctx,
-		struct xdr_netobj *read_token,
-		struct xdr_buf *message_buffer,
-		int *qop_state, int toktype)
+gss_verify_mic_kerberos(struct gss_ctx *gss_ctx,
+		struct xdr_buf *message_buffer, struct xdr_netobj *read_token)
 {
+	struct krb5_ctx		*ctx = gss_ctx->internal_ctx_id;
 	int			signalg;
 	int			sealalg;
 	s32			checksum_type;
@@ -100,16 +93,12 @@ krb5_read_token(struct krb5_ctx *ctx,
 					read_token->len))
 		goto out;
 
-	if ((*ptr++ != ((toktype>>8)&0xff)) || (*ptr++ != (toktype&0xff)))
+	if ((*ptr++ != ((KG_TOK_MIC_MSG>>8)&0xff)) ||
+	    (*ptr++ != ( KG_TOK_MIC_MSG    &0xff))   )
 		goto out;
 
 	/* XXX sanity-check bodysize?? */
 
-	if (toktype == KG_TOK_WRAP_MSG) {
-		/* XXX gone */
-		goto out;
-	}
-
 	/* get the sign and seal algorithms */
 
 	signalg = ptr[0] + (ptr[1] << 8);
@@ -120,14 +109,7 @@ krb5_read_token(struct krb5_ctx *ctx,
 	if ((ptr[4] != 0xff) || (ptr[5] != 0xff))
 		goto out;
 
-	if (((toktype != KG_TOK_WRAP_MSG) && (sealalg != 0xffff)) ||
-	    ((toktype == KG_TOK_WRAP_MSG) && (sealalg == 0xffff)))
-		goto out;
-
-	/* in the current spec, there is only one valid seal algorithm per
-	   key type, so a simple comparison is ok */
-
-	if ((toktype == KG_TOK_WRAP_MSG) && !(sealalg == ctx->sealalg))
+	if (sealalg != 0xffff)
 		goto out;
 
 	/* there are several mappings of seal algorithms to sign algorithms,
@@ -154,7 +136,7 @@ krb5_read_token(struct krb5_ctx *ctx,
 	switch (signalg) {
 	case SGN_ALG_DES_MAC_MD5:
 		ret = make_checksum(checksum_type, ptr - 2, 8,
-					 message_buffer, &md5cksum);
+					 message_buffer, 0, &md5cksum);
 		if (ret)
 			goto out;
 
@@ -175,9 +157,6 @@ krb5_read_token(struct krb5_ctx *ctx,
 
 	/* it got through unscathed.  Make sure the context is unexpired */
 
-	if (qop_state)
-		*qop_state = GSS_C_QOP_DEFAULT;
-
 	now = get_seconds();
 
 	ret = GSS_S_CONTEXT_EXPIRED;
diff --git a/net/sunrpc/auth_gss/gss_krb5_wrap.c b/net/sunrpc/auth_gss/gss_krb5_wrap.c
new file mode 100644
index 00000000000..af777cf9f25
--- /dev/null
+++ b/net/sunrpc/auth_gss/gss_krb5_wrap.c
@@ -0,0 +1,363 @@
+#include <linux/types.h>
+#include <linux/slab.h>
+#include <linux/jiffies.h>
+#include <linux/sunrpc/gss_krb5.h>
+#include <linux/random.h>
+#include <linux/pagemap.h>
+#include <asm/scatterlist.h>
+#include <linux/crypto.h>
+
+#ifdef RPC_DEBUG
+# define RPCDBG_FACILITY	RPCDBG_AUTH
+#endif
+
+static inline int
+gss_krb5_padding(int blocksize, int length)
+{
+	/* Most of the code is block-size independent but currently we
+	 * use only 8: */
+	BUG_ON(blocksize != 8);
+	return 8 - (length & 7);
+}
+
+static inline void
+gss_krb5_add_padding(struct xdr_buf *buf, int offset, int blocksize)
+{
+	int padding = gss_krb5_padding(blocksize, buf->len - offset);
+	char *p;
+	struct kvec *iov;
+
+	if (buf->page_len || buf->tail[0].iov_len)
+		iov = &buf->tail[0];
+	else
+		iov = &buf->head[0];
+	p = iov->iov_base + iov->iov_len;
+	iov->iov_len += padding;
+	buf->len += padding;
+	memset(p, padding, padding);
+}
+
+static inline int
+gss_krb5_remove_padding(struct xdr_buf *buf, int blocksize)
+{
+	u8 *ptr;
+	u8 pad;
+	int len = buf->len;
+
+	if (len <= buf->head[0].iov_len) {
+		pad = *(u8 *)(buf->head[0].iov_base + len - 1);
+		if (pad > buf->head[0].iov_len)
+			return -EINVAL;
+		buf->head[0].iov_len -= pad;
+		goto out;
+	} else
+		len -= buf->head[0].iov_len;
+	if (len <= buf->page_len) {
+		int last = (buf->page_base + len - 1)
+					>>PAGE_CACHE_SHIFT;
+		int offset = (buf->page_base + len - 1)
+					& (PAGE_CACHE_SIZE - 1);
+		ptr = kmap_atomic(buf->pages[last], KM_SKB_SUNRPC_DATA);
+		pad = *(ptr + offset);
+		kunmap_atomic(ptr, KM_SKB_SUNRPC_DATA);
+		goto out;
+	} else
+		len -= buf->page_len;
+	BUG_ON(len > buf->tail[0].iov_len);
+	pad = *(u8 *)(buf->tail[0].iov_base + len - 1);
+out:
+	/* XXX: NOTE: we do not adjust the page lengths--they represent
+	 * a range of data in the real filesystem page cache, and we need
+	 * to know that range so the xdr code can properly place read data.
+	 * However adjusting the head length, as we do above, is harmless.
+	 * In the case of a request that fits into a single page, the server
+	 * also uses length and head length together to determine the original
+	 * start of the request to copy the request for deferal; so it's
+	 * easier on the server if we adjust head and tail length in tandem.
+	 * It's not really a problem that we don't fool with the page and
+	 * tail lengths, though--at worst badly formed xdr might lead the
+	 * server to attempt to parse the padding.
+	 * XXX: Document all these weird requirements for gss mechanism
+	 * wrap/unwrap functions. */
+	if (pad > blocksize)
+		return -EINVAL;
+	if (buf->len > pad)
+		buf->len -= pad;
+	else
+		return -EINVAL;
+	return 0;
+}
+
+static inline void
+make_confounder(char *p, int blocksize)
+{
+	static u64 i = 0;
+	u64 *q = (u64 *)p;
+
+	/* rfc1964 claims this should be "random".  But all that's really
+	 * necessary is that it be unique.  And not even that is necessary in
+	 * our case since our "gssapi" implementation exists only to support
+	 * rpcsec_gss, so we know that the only buffers we will ever encrypt
+	 * already begin with a unique sequence number.  Just to hedge my bets
+	 * I'll make a half-hearted attempt at something unique, but ensuring
+	 * uniqueness would mean worrying about atomicity and rollover, and I
+	 * don't care enough. */
+
+	BUG_ON(blocksize != 8);
+	*q = i++;
+}
+
+/* Assumptions: the head and tail of inbuf are ours to play with.
+ * The pages, however, may be real pages in the page cache and we replace
+ * them with scratch pages from **pages before writing to them. */
+/* XXX: obviously the above should be documentation of wrap interface,
+ * and shouldn't be in this kerberos-specific file. */
+
+/* XXX factor out common code with seal/unseal. */
+
+u32
+gss_wrap_kerberos(struct gss_ctx *ctx, int offset,
+		struct xdr_buf *buf, struct page **pages)
+{
+	struct krb5_ctx		*kctx = ctx->internal_ctx_id;
+	s32			checksum_type;
+	struct xdr_netobj	md5cksum = {.len = 0, .data = NULL};
+	int			blocksize = 0, plainlen;
+	unsigned char		*ptr, *krb5_hdr, *msg_start;
+	s32			now;
+	int			headlen;
+	struct page		**tmp_pages;
+
+	dprintk("RPC:     gss_wrap_kerberos\n");
+
+	now = get_seconds();
+
+	switch (kctx->signalg) {
+		case SGN_ALG_DES_MAC_MD5:
+			checksum_type = CKSUMTYPE_RSA_MD5;
+			break;
+		default:
+			dprintk("RPC:      gss_krb5_seal: kctx->signalg %d not"
+				" supported\n", kctx->signalg);
+			goto out_err;
+	}
+	if (kctx->sealalg != SEAL_ALG_NONE && kctx->sealalg != SEAL_ALG_DES) {
+		dprintk("RPC:      gss_krb5_seal: kctx->sealalg %d not supported\n",
+			kctx->sealalg);
+		goto out_err;
+	}
+
+	blocksize = crypto_tfm_alg_blocksize(kctx->enc);
+	gss_krb5_add_padding(buf, offset, blocksize);
+	BUG_ON((buf->len - offset) % blocksize);
+	plainlen = blocksize + buf->len - offset;
+
+	headlen = g_token_size(&kctx->mech_used, 22 + plainlen) -
+						(buf->len - offset);
+
+	ptr = buf->head[0].iov_base + offset;
+	/* shift data to make room for header. */
+	/* XXX Would be cleverer to encrypt while copying. */
+	/* XXX bounds checking, slack, etc. */
+	memmove(ptr + headlen, ptr, buf->head[0].iov_len - offset);
+	buf->head[0].iov_len += headlen;
+	buf->len += headlen;
+	BUG_ON((buf->len - offset - headlen) % blocksize);
+
+	g_make_token_header(&kctx->mech_used, 22 + plainlen, &ptr);
+
+
+	*ptr++ = (unsigned char) ((KG_TOK_WRAP_MSG>>8)&0xff);
+	*ptr++ = (unsigned char) (KG_TOK_WRAP_MSG&0xff);
+
+	/* ptr now at byte 2 of header described in rfc 1964, section 1.2.1: */
+	krb5_hdr = ptr - 2;
+	msg_start = krb5_hdr + 24;
+	/* XXXJBF: */ BUG_ON(buf->head[0].iov_base + offset + headlen != msg_start + blocksize);
+
+	*(u16 *)(krb5_hdr + 2) = htons(kctx->signalg);
+	memset(krb5_hdr + 4, 0xff, 4);
+	*(u16 *)(krb5_hdr + 4) = htons(kctx->sealalg);
+
+	make_confounder(msg_start, blocksize);
+
+	/* XXXJBF: UGH!: */
+	tmp_pages = buf->pages;
+	buf->pages = pages;
+	if (make_checksum(checksum_type, krb5_hdr, 8, buf,
+				offset + headlen - blocksize, &md5cksum))
+		goto out_err;
+	buf->pages = tmp_pages;
+
+	switch (kctx->signalg) {
+	case SGN_ALG_DES_MAC_MD5:
+		if (krb5_encrypt(kctx->seq, NULL, md5cksum.data,
+				  md5cksum.data, md5cksum.len))
+			goto out_err;
+		memcpy(krb5_hdr + 16,
+		       md5cksum.data + md5cksum.len - KRB5_CKSUM_LENGTH,
+		       KRB5_CKSUM_LENGTH);
+
+		dprintk("RPC:      make_seal_token: cksum data: \n");
+		print_hexl((u32 *) (krb5_hdr + 16), KRB5_CKSUM_LENGTH, 0);
+		break;
+	default:
+		BUG();
+	}
+
+	kfree(md5cksum.data);
+
+	/* XXX would probably be more efficient to compute checksum
+	 * and encrypt at the same time: */
+	if ((krb5_make_seq_num(kctx->seq, kctx->initiate ? 0 : 0xff,
+			       kctx->seq_send, krb5_hdr + 16, krb5_hdr + 8)))
+		goto out_err;
+
+	if (gss_encrypt_xdr_buf(kctx->enc, buf, offset + headlen - blocksize,
+									pages))
+		goto out_err;
+
+	kctx->seq_send++;
+
+	return ((kctx->endtime < now) ? GSS_S_CONTEXT_EXPIRED : GSS_S_COMPLETE);
+out_err:
+	if (md5cksum.data) kfree(md5cksum.data);
+	return GSS_S_FAILURE;
+}
+
+u32
+gss_unwrap_kerberos(struct gss_ctx *ctx, int offset, struct xdr_buf *buf)
+{
+	struct krb5_ctx		*kctx = ctx->internal_ctx_id;
+	int			signalg;
+	int			sealalg;
+	s32			checksum_type;
+	struct xdr_netobj	md5cksum = {.len = 0, .data = NULL};
+	s32			now;
+	int			direction;
+	s32			seqnum;
+	unsigned char		*ptr;
+	int			bodysize;
+	u32			ret = GSS_S_DEFECTIVE_TOKEN;
+	void			*data_start, *orig_start;
+	int			data_len;
+	int			blocksize;
+
+	dprintk("RPC:      gss_unwrap_kerberos\n");
+
+	ptr = (u8 *)buf->head[0].iov_base + offset;
+	if (g_verify_token_header(&kctx->mech_used, &bodysize, &ptr,
+					buf->len - offset))
+		goto out;
+
+	if ((*ptr++ != ((KG_TOK_WRAP_MSG>>8)&0xff)) ||
+	    (*ptr++ !=  (KG_TOK_WRAP_MSG    &0xff))   )
+		goto out;
+
+	/* XXX sanity-check bodysize?? */
+
+	/* get the sign and seal algorithms */
+
+	signalg = ptr[0] + (ptr[1] << 8);
+	sealalg = ptr[2] + (ptr[3] << 8);
+
+	/* Sanity checks */
+
+	if ((ptr[4] != 0xff) || (ptr[5] != 0xff))
+		goto out;
+
+	if (sealalg == 0xffff)
+		goto out;
+
+	/* in the current spec, there is only one valid seal algorithm per
+	   key type, so a simple comparison is ok */
+
+	if (sealalg != kctx->sealalg)
+		goto out;
+
+	/* there are several mappings of seal algorithms to sign algorithms,
+	   but few enough that we can try them all. */
+
+	if ((kctx->sealalg == SEAL_ALG_NONE && signalg > 1) ||
+	    (kctx->sealalg == SEAL_ALG_1 && signalg != SGN_ALG_3) ||
+	    (kctx->sealalg == SEAL_ALG_DES3KD &&
+	     signalg != SGN_ALG_HMAC_SHA1_DES3_KD))
+		goto out;
+
+	if (gss_decrypt_xdr_buf(kctx->enc, buf,
+			ptr + 22 - (unsigned char *)buf->head[0].iov_base))
+		goto out;
+
+	/* compute the checksum of the message */
+
+	/* initialize the the cksum */
+	switch (signalg) {
+	case SGN_ALG_DES_MAC_MD5:
+		checksum_type = CKSUMTYPE_RSA_MD5;
+		break;
+	default:
+		ret = GSS_S_DEFECTIVE_TOKEN;
+		goto out;
+	}
+
+	switch (signalg) {
+	case SGN_ALG_DES_MAC_MD5:
+		ret = make_checksum(checksum_type, ptr - 2, 8, buf,
+			 ptr + 22 - (unsigned char *)buf->head[0].iov_base, &md5cksum);
+		if (ret)
+			goto out;
+
+		ret = krb5_encrypt(kctx->seq, NULL, md5cksum.data,
+				   md5cksum.data, md5cksum.len);
+		if (ret)
+			goto out;
+
+		if (memcmp(md5cksum.data + 8, ptr + 14, 8)) {
+			ret = GSS_S_BAD_SIG;
+			goto out;
+		}
+		break;
+	default:
+		ret = GSS_S_DEFECTIVE_TOKEN;
+		goto out;
+	}
+
+	/* it got through unscathed.  Make sure the context is unexpired */
+
+	now = get_seconds();
+
+	ret = GSS_S_CONTEXT_EXPIRED;
+	if (now > kctx->endtime)
+		goto out;
+
+	/* do sequencing checks */
+
+	ret = GSS_S_BAD_SIG;
+	if ((ret = krb5_get_seq_num(kctx->seq, ptr + 14, ptr + 6, &direction,
+				    &seqnum)))
+		goto out;
+
+	if ((kctx->initiate && direction != 0xff) ||
+	    (!kctx->initiate && direction != 0))
+		goto out;
+
+	/* Copy the data back to the right position.  XXX: Would probably be
+	 * better to copy and encrypt at the same time. */
+
+	blocksize = crypto_tfm_alg_blocksize(kctx->enc);
+	data_start = ptr + 22 + blocksize;
+	orig_start = buf->head[0].iov_base + offset;
+	data_len = (buf->head[0].iov_base + buf->head[0].iov_len) - data_start;
+	memmove(orig_start, data_start, data_len);
+	buf->head[0].iov_len -= (data_start - orig_start);
+	buf->len -= (data_start - orig_start);
+
+	ret = GSS_S_DEFECTIVE_TOKEN;
+	if (gss_krb5_remove_padding(buf, blocksize))
+		goto out;
+
+	ret = GSS_S_COMPLETE;
+out:
+	if (md5cksum.data) kfree(md5cksum.data);
+	return ret;
+}
diff --git a/net/sunrpc/auth_gss/gss_mech_switch.c b/net/sunrpc/auth_gss/gss_mech_switch.c
index 9dfb68377d6..b048bf672da 100644
--- a/net/sunrpc/auth_gss/gss_mech_switch.c
+++ b/net/sunrpc/auth_gss/gss_mech_switch.c
@@ -35,7 +35,6 @@
 
 #include <linux/types.h>
 #include <linux/slab.h>
-#include <linux/socket.h>
 #include <linux/module.h>
 #include <linux/sunrpc/msg_prot.h>
 #include <linux/sunrpc/gss_asn1.h>
@@ -251,13 +250,11 @@ gss_import_sec_context(const void *input_token, size_t bufsize,
 
 u32
 gss_get_mic(struct gss_ctx	*context_handle,
-	    u32			qop,
 	    struct xdr_buf	*message,
 	    struct xdr_netobj	*mic_token)
 {
 	 return context_handle->mech_type->gm_ops
 		->gss_get_mic(context_handle,
-			      qop,
 			      message,
 			      mic_token);
 }
@@ -267,16 +264,34 @@ gss_get_mic(struct gss_ctx	*context_handle,
 u32
 gss_verify_mic(struct gss_ctx		*context_handle,
 	       struct xdr_buf		*message,
-	       struct xdr_netobj	*mic_token,
-	       u32			*qstate)
+	       struct xdr_netobj	*mic_token)
 {
 	return context_handle->mech_type->gm_ops
 		->gss_verify_mic(context_handle,
 				 message,
-				 mic_token,
-				 qstate);
+				 mic_token);
 }
 
+u32
+gss_wrap(struct gss_ctx	*ctx_id,
+	 int		offset,
+	 struct xdr_buf	*buf,
+	 struct page	**inpages)
+{
+	return ctx_id->mech_type->gm_ops
+		->gss_wrap(ctx_id, offset, buf, inpages);
+}
+
+u32
+gss_unwrap(struct gss_ctx	*ctx_id,
+	   int			offset,
+	   struct xdr_buf	*buf)
+{
+	return ctx_id->mech_type->gm_ops
+		->gss_unwrap(ctx_id, offset, buf);
+}
+
+
 /* gss_delete_sec_context: free all resources associated with context_handle.
  * Note this differs from the RFC 2744-specified prototype in that we don't
  * bother returning an output token, since it would never be used anyway. */
diff --git a/net/sunrpc/auth_gss/gss_spkm3_mech.c b/net/sunrpc/auth_gss/gss_spkm3_mech.c
index 6c97d61baa9..39b3edc1469 100644
--- a/net/sunrpc/auth_gss/gss_spkm3_mech.c
+++ b/net/sunrpc/auth_gss/gss_spkm3_mech.c
@@ -224,18 +224,13 @@ gss_delete_sec_context_spkm3(void *internal_ctx) {
 static u32
 gss_verify_mic_spkm3(struct gss_ctx		*ctx,
 			struct xdr_buf		*signbuf,
-			struct xdr_netobj	*checksum,
-			u32		*qstate) {
+			struct xdr_netobj	*checksum)
+{
 	u32 maj_stat = 0;
-	int qop_state = 0;
 	struct spkm3_ctx *sctx = ctx->internal_ctx_id;
 
 	dprintk("RPC: gss_verify_mic_spkm3 calling spkm3_read_token\n");
-	maj_stat = spkm3_read_token(sctx, checksum, signbuf, &qop_state,
-				   SPKM_MIC_TOK);
-
-	if (!maj_stat && qop_state)
-	    *qstate = qop_state;
+	maj_stat = spkm3_read_token(sctx, checksum, signbuf, SPKM_MIC_TOK);
 
 	dprintk("RPC: gss_verify_mic_spkm3 returning %d\n", maj_stat);
 	return maj_stat;
@@ -243,15 +238,15 @@ gss_verify_mic_spkm3(struct gss_ctx		*ctx,
 
 static u32
 gss_get_mic_spkm3(struct gss_ctx	*ctx,
-		     u32		qop,
 		     struct xdr_buf	*message_buffer,
-		     struct xdr_netobj	*message_token) {
+		     struct xdr_netobj	*message_token)
+{
 	u32 err = 0;
 	struct spkm3_ctx *sctx = ctx->internal_ctx_id;
 
 	dprintk("RPC: gss_get_mic_spkm3\n");
 
-	err = spkm3_make_token(sctx, qop, message_buffer,
+	err = spkm3_make_token(sctx, message_buffer,
 			      message_token, SPKM_MIC_TOK);
 	return err;
 }
@@ -264,8 +259,8 @@ static struct gss_api_ops gss_spkm3_ops = {
 };
 
 static struct pf_desc gss_spkm3_pfs[] = {
-	{RPC_AUTH_GSS_SPKM, 0, RPC_GSS_SVC_NONE, "spkm3"},
-	{RPC_AUTH_GSS_SPKMI, 0, RPC_GSS_SVC_INTEGRITY, "spkm3i"},
+	{RPC_AUTH_GSS_SPKM, RPC_GSS_SVC_NONE, "spkm3"},
+	{RPC_AUTH_GSS_SPKMI, RPC_GSS_SVC_INTEGRITY, "spkm3i"},
 };
 
 static struct gss_api_mech gss_spkm3_mech = {
diff --git a/net/sunrpc/auth_gss/gss_spkm3_seal.c b/net/sunrpc/auth_gss/gss_spkm3_seal.c
index 25339868d46..148201e929d 100644
--- a/net/sunrpc/auth_gss/gss_spkm3_seal.c
+++ b/net/sunrpc/auth_gss/gss_spkm3_seal.c
@@ -51,7 +51,7 @@
  */
 
 u32
-spkm3_make_token(struct spkm3_ctx *ctx, int qop_req,
+spkm3_make_token(struct spkm3_ctx *ctx,
 		   struct xdr_buf * text, struct xdr_netobj * token,
 		   int toktype)
 {
@@ -68,8 +68,6 @@ spkm3_make_token(struct spkm3_ctx *ctx, int qop_req,
 	dprintk("RPC: spkm3_make_token\n");
 
 	now = jiffies;
-	if (qop_req != 0)
-		goto out_err;
 
 	if (ctx->ctx_id.len != 16) {
 		dprintk("RPC: spkm3_make_token BAD ctx_id.len %d\n",
diff --git a/net/sunrpc/auth_gss/gss_spkm3_unseal.c b/net/sunrpc/auth_gss/gss_spkm3_unseal.c
index 65ce81bf0bc..c3c0d958610 100644
--- a/net/sunrpc/auth_gss/gss_spkm3_unseal.c
+++ b/net/sunrpc/auth_gss/gss_spkm3_unseal.c
@@ -52,7 +52,7 @@ u32
 spkm3_read_token(struct spkm3_ctx *ctx,
 		struct xdr_netobj *read_token,    /* checksum */
 		struct xdr_buf *message_buffer, /* signbuf */
-		int *qop_state, int toktype)
+		int toktype)
 {
 	s32			code;
 	struct xdr_netobj	wire_cksum = {.len =0, .data = NULL};
diff --git a/net/sunrpc/auth_gss/svcauth_gss.c b/net/sunrpc/auth_gss/svcauth_gss.c
index e3308195374..e4ada15ed85 100644
--- a/net/sunrpc/auth_gss/svcauth_gss.c
+++ b/net/sunrpc/auth_gss/svcauth_gss.c
@@ -566,8 +566,7 @@ gss_verify_header(struct svc_rqst *rqstp, struct rsc *rsci,
 
 	if (rqstp->rq_deferred) /* skip verification of revisited request */
 		return SVC_OK;
-	if (gss_verify_mic(ctx_id, &rpchdr, &checksum, NULL)
-							!= GSS_S_COMPLETE) {
+	if (gss_verify_mic(ctx_id, &rpchdr, &checksum) != GSS_S_COMPLETE) {
 		*authp = rpcsec_gsserr_credproblem;
 		return SVC_DENIED;
 	}
@@ -604,7 +603,7 @@ gss_write_verf(struct svc_rqst *rqstp, struct gss_ctx *ctx_id, u32 seq)
 	xdr_buf_from_iov(&iov, &verf_data);
 	p = rqstp->rq_res.head->iov_base + rqstp->rq_res.head->iov_len;
 	mic.data = (u8 *)(p + 1);
-	maj_stat = gss_get_mic(ctx_id, 0, &verf_data, &mic);
+	maj_stat = gss_get_mic(ctx_id, &verf_data, &mic);
 	if (maj_stat != GSS_S_COMPLETE)
 		return -1;
 	*p++ = htonl(mic.len);
@@ -710,7 +709,7 @@ unwrap_integ_data(struct xdr_buf *buf, u32 seq, struct gss_ctx *ctx)
 		goto out;
 	if (read_bytes_from_xdr_buf(buf, integ_len + 4, mic.data, mic.len))
 		goto out;
-	maj_stat = gss_verify_mic(ctx, &integ_buf, &mic, NULL);
+	maj_stat = gss_verify_mic(ctx, &integ_buf, &mic);
 	if (maj_stat != GSS_S_COMPLETE)
 		goto out;
 	if (ntohl(svc_getu32(&buf->head[0])) != seq)
@@ -1012,7 +1011,7 @@ svcauth_gss_release(struct svc_rqst *rqstp)
 			resv = &resbuf->tail[0];
 		}
 		mic.data = (u8 *)resv->iov_base + resv->iov_len + 4;
-		if (gss_get_mic(gsd->rsci->mechctx, 0, &integ_buf, &mic))
+		if (gss_get_mic(gsd->rsci->mechctx, &integ_buf, &mic))
 			goto out_err;
 		svc_putu32(resv, htonl(mic.len));
 		memset(mic.data + mic.len, 0,
diff --git a/net/sunrpc/auth_null.c b/net/sunrpc/auth_null.c
index 9b72d3abf82..f56767aaa92 100644
--- a/net/sunrpc/auth_null.c
+++ b/net/sunrpc/auth_null.c
@@ -7,9 +7,7 @@
  */
 
 #include <linux/types.h>
-#include <linux/socket.h>
 #include <linux/module.h>
-#include <linux/in.h>
 #include <linux/utsname.h>
 #include <linux/sunrpc/clnt.h>
 #include <linux/sched.h>
diff --git a/net/sunrpc/auth_unix.c b/net/sunrpc/auth_unix.c
index 4ff297a9b15..890fb5ea0dc 100644
--- a/net/sunrpc/auth_unix.c
+++ b/net/sunrpc/auth_unix.c
@@ -9,8 +9,6 @@
 #include <linux/types.h>
 #include <linux/sched.h>
 #include <linux/module.h>
-#include <linux/socket.h>
-#include <linux/in.h>
 #include <linux/sunrpc/clnt.h>
 #include <linux/sunrpc/auth.h>
 
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index f17e6153b68..702ede309b0 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -1,5 +1,5 @@
 /*
- *  linux/net/sunrpc/rpcclnt.c
+ *  linux/net/sunrpc/clnt.c
  *
  *  This file contains the high-level RPC interface.
  *  It is modeled as a finite state machine to support both synchronous
@@ -27,7 +27,6 @@
 #include <linux/types.h>
 #include <linux/mm.h>
 #include <linux/slab.h>
-#include <linux/in.h>
 #include <linux/utsname.h>
 
 #include <linux/sunrpc/clnt.h>
@@ -53,6 +52,7 @@ static void	call_allocate(struct rpc_task *task);
 static void	call_encode(struct rpc_task *task);
 static void	call_decode(struct rpc_task *task);
 static void	call_bind(struct rpc_task *task);
+static void	call_bind_status(struct rpc_task *task);
 static void	call_transmit(struct rpc_task *task);
 static void	call_status(struct rpc_task *task);
 static void	call_refresh(struct rpc_task *task);
@@ -517,15 +517,8 @@ void
 rpc_setbufsize(struct rpc_clnt *clnt, unsigned int sndsize, unsigned int rcvsize)
 {
 	struct rpc_xprt *xprt = clnt->cl_xprt;
-
-	xprt->sndsize = 0;
-	if (sndsize)
-		xprt->sndsize = sndsize + RPC_SLACK_SPACE;
-	xprt->rcvsize = 0;
-	if (rcvsize)
-		xprt->rcvsize = rcvsize + RPC_SLACK_SPACE;
-	if (xprt_connected(xprt))
-		xprt_sock_setbufsize(xprt);
+	if (xprt->ops->set_buffer_size)
+		xprt->ops->set_buffer_size(xprt, sndsize, rcvsize);
 }
 
 /*
@@ -685,13 +678,11 @@ call_allocate(struct rpc_task *task)
 static void
 call_encode(struct rpc_task *task)
 {
-	struct rpc_clnt	*clnt = task->tk_client;
 	struct rpc_rqst	*req = task->tk_rqstp;
 	struct xdr_buf *sndbuf = &req->rq_snd_buf;
 	struct xdr_buf *rcvbuf = &req->rq_rcv_buf;
 	unsigned int	bufsiz;
 	kxdrproc_t	encode;
-	int		status;
 	u32		*p;
 
 	dprintk("RPC: %4d call_encode (status %d)\n", 
@@ -719,11 +710,15 @@ call_encode(struct rpc_task *task)
 		rpc_exit(task, -EIO);
 		return;
 	}
-	if (encode && (status = rpcauth_wrap_req(task, encode, req, p,
-						 task->tk_msg.rpc_argp)) < 0) {
-		printk(KERN_WARNING "%s: can't encode arguments: %d\n",
-				clnt->cl_protname, -status);
-		rpc_exit(task, status);
+	if (encode == NULL)
+		return;
+
+	task->tk_status = rpcauth_wrap_req(task, encode, req, p,
+			task->tk_msg.rpc_argp);
+	if (task->tk_status == -ENOMEM) {
+		/* XXX: Is this sane? */
+		rpc_delay(task, 3*HZ);
+		task->tk_status = -EAGAIN;
 	}
 }
 
@@ -734,43 +729,95 @@ static void
 call_bind(struct rpc_task *task)
 {
 	struct rpc_clnt	*clnt = task->tk_client;
-	struct rpc_xprt *xprt = clnt->cl_xprt;
-
-	dprintk("RPC: %4d call_bind xprt %p %s connected\n", task->tk_pid,
-			xprt, (xprt_connected(xprt) ? "is" : "is not"));
 
-	task->tk_action = (xprt_connected(xprt)) ? call_transmit : call_connect;
+	dprintk("RPC: %4d call_bind (status %d)\n",
+				task->tk_pid, task->tk_status);
 
+	task->tk_action = call_connect;
 	if (!clnt->cl_port) {
-		task->tk_action = call_connect;
-		task->tk_timeout = RPC_CONNECT_TIMEOUT;
+		task->tk_action = call_bind_status;
+		task->tk_timeout = task->tk_xprt->bind_timeout;
 		rpc_getport(task, clnt);
 	}
 }
 
 /*
- * 4a.	Connect to the RPC server (TCP case)
+ * 4a.	Sort out bind result
+ */
+static void
+call_bind_status(struct rpc_task *task)
+{
+	int status = -EACCES;
+
+	if (task->tk_status >= 0) {
+		dprintk("RPC: %4d call_bind_status (status %d)\n",
+					task->tk_pid, task->tk_status);
+		task->tk_status = 0;
+		task->tk_action = call_connect;
+		return;
+	}
+
+	switch (task->tk_status) {
+	case -EACCES:
+		dprintk("RPC: %4d remote rpcbind: RPC program/version unavailable\n",
+				task->tk_pid);
+		rpc_delay(task, 3*HZ);
+		goto retry_bind;
+	case -ETIMEDOUT:
+		dprintk("RPC: %4d rpcbind request timed out\n",
+				task->tk_pid);
+		if (RPC_IS_SOFT(task)) {
+			status = -EIO;
+			break;
+		}
+		goto retry_bind;
+	case -EPFNOSUPPORT:
+		dprintk("RPC: %4d remote rpcbind service unavailable\n",
+				task->tk_pid);
+		break;
+	case -EPROTONOSUPPORT:
+		dprintk("RPC: %4d remote rpcbind version 2 unavailable\n",
+				task->tk_pid);
+		break;
+	default:
+		dprintk("RPC: %4d unrecognized rpcbind error (%d)\n",
+				task->tk_pid, -task->tk_status);
+		status = -EIO;
+		break;
+	}
+
+	rpc_exit(task, status);
+	return;
+
+retry_bind:
+	task->tk_status = 0;
+	task->tk_action = call_bind;
+	return;
+}
+
+/*
+ * 4b.	Connect to the RPC server
  */
 static void
 call_connect(struct rpc_task *task)
 {
-	struct rpc_clnt *clnt = task->tk_client;
+	struct rpc_xprt *xprt = task->tk_xprt;
 
-	dprintk("RPC: %4d call_connect status %d\n",
-				task->tk_pid, task->tk_status);
+	dprintk("RPC: %4d call_connect xprt %p %s connected\n",
+			task->tk_pid, xprt,
+			(xprt_connected(xprt) ? "is" : "is not"));
 
-	if (xprt_connected(clnt->cl_xprt)) {
-		task->tk_action = call_transmit;
-		return;
+	task->tk_action = call_transmit;
+	if (!xprt_connected(xprt)) {
+		task->tk_action = call_connect_status;
+		if (task->tk_status < 0)
+			return;
+		xprt_connect(task);
 	}
-	task->tk_action = call_connect_status;
-	if (task->tk_status < 0)
-		return;
-	xprt_connect(task);
 }
 
 /*
- * 4b. Sort out connect result
+ * 4c.	Sort out connect result
  */
 static void
 call_connect_status(struct rpc_task *task)
@@ -778,6 +825,9 @@ call_connect_status(struct rpc_task *task)
 	struct rpc_clnt *clnt = task->tk_client;
 	int status = task->tk_status;
 
+	dprintk("RPC: %5u call_connect_status (status %d)\n", 
+				task->tk_pid, task->tk_status);
+
 	task->tk_status = 0;
 	if (status >= 0) {
 		clnt->cl_stats->netreconn++;
@@ -785,17 +835,19 @@ call_connect_status(struct rpc_task *task)
 		return;
 	}
 
-	/* Something failed: we may have to rebind */
+	/* Something failed: remote service port may have changed */
 	if (clnt->cl_autobind)
 		clnt->cl_port = 0;
+
 	switch (status) {
 	case -ENOTCONN:
 	case -ETIMEDOUT:
 	case -EAGAIN:
-		task->tk_action = (clnt->cl_port == 0) ? call_bind : call_connect;
+		task->tk_action = call_bind;
 		break;
 	default:
 		rpc_exit(task, -EIO);
+		break;
 	}
 }
 
@@ -815,10 +867,12 @@ call_transmit(struct rpc_task *task)
 	if (task->tk_status != 0)
 		return;
 	/* Encode here so that rpcsec_gss can use correct sequence number. */
-	if (!task->tk_rqstp->rq_bytes_sent)
+	if (task->tk_rqstp->rq_bytes_sent == 0) {
 		call_encode(task);
-	if (task->tk_status < 0)
-		return;
+		/* Did the encode result in an error condition? */
+		if (task->tk_status != 0)
+			goto out_nosend;
+	}
 	xprt_transmit(task);
 	if (task->tk_status < 0)
 		return;
@@ -826,6 +880,10 @@ call_transmit(struct rpc_task *task)
 		task->tk_action = NULL;
 		rpc_wake_up_task(task);
 	}
+	return;
+out_nosend:
+	/* release socket write lock before attempting to handle error */
+	xprt_abort_transmit(task);
 }
 
 /*
@@ -1020,13 +1078,12 @@ static u32 *
 call_header(struct rpc_task *task)
 {
 	struct rpc_clnt *clnt = task->tk_client;
-	struct rpc_xprt *xprt = clnt->cl_xprt;
 	struct rpc_rqst	*req = task->tk_rqstp;
 	u32		*p = req->rq_svec[0].iov_base;
 
 	/* FIXME: check buffer size? */
-	if (xprt->stream)
-		*p++ = 0;		/* fill in later */
+
+	p = xprt_skip_transport_header(task->tk_xprt, p);
 	*p++ = req->rq_xid;		/* XID */
 	*p++ = htonl(RPC_CALL);		/* CALL */
 	*p++ = htonl(RPC_VERSION);	/* RPC version */
diff --git a/net/sunrpc/pmap_clnt.c b/net/sunrpc/pmap_clnt.c
index 4e81f276692..a398575f94b 100644
--- a/net/sunrpc/pmap_clnt.c
+++ b/net/sunrpc/pmap_clnt.c
@@ -26,7 +26,7 @@
 #define PMAP_GETPORT		3
 
 static struct rpc_procinfo	pmap_procedures[];
-static struct rpc_clnt *	pmap_create(char *, struct sockaddr_in *, int);
+static struct rpc_clnt *	pmap_create(char *, struct sockaddr_in *, int, int);
 static void			pmap_getport_done(struct rpc_task *);
 static struct rpc_program	pmap_program;
 static DEFINE_SPINLOCK(pmap_lock);
@@ -65,7 +65,7 @@ rpc_getport(struct rpc_task *task, struct rpc_clnt *clnt)
 	map->pm_binding = 1;
 	spin_unlock(&pmap_lock);
 
-	pmap_clnt = pmap_create(clnt->cl_server, sap, map->pm_prot);
+	pmap_clnt = pmap_create(clnt->cl_server, sap, map->pm_prot, 0);
 	if (IS_ERR(pmap_clnt)) {
 		task->tk_status = PTR_ERR(pmap_clnt);
 		goto bailout;
@@ -112,7 +112,7 @@ rpc_getport_external(struct sockaddr_in *sin, __u32 prog, __u32 vers, int prot)
 			NIPQUAD(sin->sin_addr.s_addr), prog, vers, prot);
 
 	sprintf(hostname, "%u.%u.%u.%u", NIPQUAD(sin->sin_addr.s_addr));
-	pmap_clnt = pmap_create(hostname, sin, prot);
+	pmap_clnt = pmap_create(hostname, sin, prot, 0);
 	if (IS_ERR(pmap_clnt))
 		return PTR_ERR(pmap_clnt);
 
@@ -171,7 +171,7 @@ rpc_register(u32 prog, u32 vers, int prot, unsigned short port, int *okay)
 
 	sin.sin_family = AF_INET;
 	sin.sin_addr.s_addr = htonl(INADDR_LOOPBACK);
-	pmap_clnt = pmap_create("localhost", &sin, IPPROTO_UDP);
+	pmap_clnt = pmap_create("localhost", &sin, IPPROTO_UDP, 1);
 	if (IS_ERR(pmap_clnt)) {
 		error = PTR_ERR(pmap_clnt);
 		dprintk("RPC: couldn't create pmap client. Error = %d\n", error);
@@ -198,7 +198,7 @@ rpc_register(u32 prog, u32 vers, int prot, unsigned short port, int *okay)
 }
 
 static struct rpc_clnt *
-pmap_create(char *hostname, struct sockaddr_in *srvaddr, int proto)
+pmap_create(char *hostname, struct sockaddr_in *srvaddr, int proto, int privileged)
 {
 	struct rpc_xprt	*xprt;
 	struct rpc_clnt	*clnt;
@@ -208,6 +208,8 @@ pmap_create(char *hostname, struct sockaddr_in *srvaddr, int proto)
 	if (IS_ERR(xprt))
 		return (struct rpc_clnt *)xprt;
 	xprt->addr.sin_port = htons(RPC_PMAP_PORT);
+	if (!privileged)
+		xprt->resvport = 0;
 
 	/* printk("pmap: create clnt\n"); */
 	clnt = rpc_new_client(xprt, hostname,
diff --git a/net/sunrpc/rpc_pipe.c b/net/sunrpc/rpc_pipe.c
index ded6c63f11e..4f188d0a5d1 100644
--- a/net/sunrpc/rpc_pipe.c
+++ b/net/sunrpc/rpc_pipe.c
@@ -76,25 +76,35 @@ int
 rpc_queue_upcall(struct inode *inode, struct rpc_pipe_msg *msg)
 {
 	struct rpc_inode *rpci = RPC_I(inode);
-	int res = 0;
+	int res = -EPIPE;
 
 	down(&inode->i_sem);
+	if (rpci->ops == NULL)
+		goto out;
 	if (rpci->nreaders) {
 		list_add_tail(&msg->list, &rpci->pipe);
 		rpci->pipelen += msg->len;
+		res = 0;
 	} else if (rpci->flags & RPC_PIPE_WAIT_FOR_OPEN) {
 		if (list_empty(&rpci->pipe))
 			schedule_delayed_work(&rpci->queue_timeout,
 					RPC_UPCALL_TIMEOUT);
 		list_add_tail(&msg->list, &rpci->pipe);
 		rpci->pipelen += msg->len;
-	} else
-		res = -EPIPE;
+		res = 0;
+	}
+out:
 	up(&inode->i_sem);
 	wake_up(&rpci->waitq);
 	return res;
 }
 
+static inline void
+rpc_inode_setowner(struct inode *inode, void *private)
+{
+	RPC_I(inode)->private = private;
+}
+
 static void
 rpc_close_pipes(struct inode *inode)
 {
@@ -111,15 +121,10 @@ rpc_close_pipes(struct inode *inode)
 			rpci->ops->release_pipe(inode);
 		rpci->ops = NULL;
 	}
+	rpc_inode_setowner(inode, NULL);
 	up(&inode->i_sem);
 }
 
-static inline void
-rpc_inode_setowner(struct inode *inode, void *private)
-{
-	RPC_I(inode)->private = private;
-}
-
 static struct inode *
 rpc_alloc_inode(struct super_block *sb)
 {
@@ -501,7 +506,6 @@ repeat:
 			dentry = dvec[--n];
 			if (dentry->d_inode) {
 				rpc_close_pipes(dentry->d_inode);
-				rpc_inode_setowner(dentry->d_inode, NULL);
 				simple_unlink(dir, dentry);
 			}
 			dput(dentry);
@@ -576,10 +580,8 @@ __rpc_rmdir(struct inode *dir, struct dentry *dentry)
 	int error;
 
 	shrink_dcache_parent(dentry);
-	if (dentry->d_inode) {
+	if (dentry->d_inode)
 		rpc_close_pipes(dentry->d_inode);
-		rpc_inode_setowner(dentry->d_inode, NULL);
-	}
 	if ((error = simple_rmdir(dir, dentry)) != 0)
 		return error;
 	if (!error) {
@@ -732,7 +734,6 @@ rpc_unlink(char *path)
 	d_drop(dentry);
 	if (dentry->d_inode) {
 		rpc_close_pipes(dentry->d_inode);
-		rpc_inode_setowner(dentry->d_inode, NULL);
 		error = simple_unlink(dir, dentry);
 	}
 	dput(dentry);
diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c
index f3104035e35..54e60a65750 100644
--- a/net/sunrpc/sched.c
+++ b/net/sunrpc/sched.c
@@ -719,7 +719,7 @@ static void rpc_async_schedule(void *arg)
 void *
 rpc_malloc(struct rpc_task *task, size_t size)
 {
-	int	gfp;
+	gfp_t	gfp;
 
 	if (task->tk_flags & RPC_TASK_SWAPPER)
 		gfp = GFP_ATOMIC;
diff --git a/net/sunrpc/socklib.c b/net/sunrpc/socklib.c
new file mode 100644
index 00000000000..8f97e90f36c
--- /dev/null
+++ b/net/sunrpc/socklib.c
@@ -0,0 +1,175 @@
+/*
+ * linux/net/sunrpc/socklib.c
+ *
+ * Common socket helper routines for RPC client and server
+ *
+ * Copyright (C) 1995, 1996 Olaf Kirch <okir@monad.swb.de>
+ */
+
+#include <linux/types.h>
+#include <linux/pagemap.h>
+#include <linux/udp.h>
+#include <linux/sunrpc/xdr.h>
+
+
+/**
+ * skb_read_bits - copy some data bits from skb to internal buffer
+ * @desc: sk_buff copy helper
+ * @to: copy destination
+ * @len: number of bytes to copy
+ *
+ * Possibly called several times to iterate over an sk_buff and copy
+ * data out of it.
+ */
+static size_t skb_read_bits(skb_reader_t *desc, void *to, size_t len)
+{
+	if (len > desc->count)
+		len = desc->count;
+	if (skb_copy_bits(desc->skb, desc->offset, to, len))
+		return 0;
+	desc->count -= len;
+	desc->offset += len;
+	return len;
+}
+
+/**
+ * skb_read_and_csum_bits - copy and checksum from skb to buffer
+ * @desc: sk_buff copy helper
+ * @to: copy destination
+ * @len: number of bytes to copy
+ *
+ * Same as skb_read_bits, but calculate a checksum at the same time.
+ */
+static size_t skb_read_and_csum_bits(skb_reader_t *desc, void *to, size_t len)
+{
+	unsigned int	csum2, pos;
+
+	if (len > desc->count)
+		len = desc->count;
+	pos = desc->offset;
+	csum2 = skb_copy_and_csum_bits(desc->skb, pos, to, len, 0);
+	desc->csum = csum_block_add(desc->csum, csum2, pos);
+	desc->count -= len;
+	desc->offset += len;
+	return len;
+}
+
+/**
+ * xdr_partial_copy_from_skb - copy data out of an skb
+ * @xdr: target XDR buffer
+ * @base: starting offset
+ * @desc: sk_buff copy helper
+ * @copy_actor: virtual method for copying data
+ *
+ */
+ssize_t xdr_partial_copy_from_skb(struct xdr_buf *xdr, unsigned int base, skb_reader_t *desc, skb_read_actor_t copy_actor)
+{
+	struct page	**ppage = xdr->pages;
+	unsigned int	len, pglen = xdr->page_len;
+	ssize_t		copied = 0;
+	int		ret;
+
+	len = xdr->head[0].iov_len;
+	if (base < len) {
+		len -= base;
+		ret = copy_actor(desc, (char *)xdr->head[0].iov_base + base, len);
+		copied += ret;
+		if (ret != len || !desc->count)
+			goto out;
+		base = 0;
+	} else
+		base -= len;
+
+	if (unlikely(pglen == 0))
+		goto copy_tail;
+	if (unlikely(base >= pglen)) {
+		base -= pglen;
+		goto copy_tail;
+	}
+	if (base || xdr->page_base) {
+		pglen -= base;
+		base += xdr->page_base;
+		ppage += base >> PAGE_CACHE_SHIFT;
+		base &= ~PAGE_CACHE_MASK;
+	}
+	do {
+		char *kaddr;
+
+		/* ACL likes to be lazy in allocating pages - ACLs
+		 * are small by default but can get huge. */
+		if (unlikely(*ppage == NULL)) {
+			*ppage = alloc_page(GFP_ATOMIC);
+			if (unlikely(*ppage == NULL)) {
+				if (copied == 0)
+					copied = -ENOMEM;
+				goto out;
+			}
+		}
+
+		len = PAGE_CACHE_SIZE;
+		kaddr = kmap_atomic(*ppage, KM_SKB_SUNRPC_DATA);
+		if (base) {
+			len -= base;
+			if (pglen < len)
+				len = pglen;
+			ret = copy_actor(desc, kaddr + base, len);
+			base = 0;
+		} else {
+			if (pglen < len)
+				len = pglen;
+			ret = copy_actor(desc, kaddr, len);
+		}
+		flush_dcache_page(*ppage);
+		kunmap_atomic(kaddr, KM_SKB_SUNRPC_DATA);
+		copied += ret;
+		if (ret != len || !desc->count)
+			goto out;
+		ppage++;
+	} while ((pglen -= len) != 0);
+copy_tail:
+	len = xdr->tail[0].iov_len;
+	if (base < len)
+		copied += copy_actor(desc, (char *)xdr->tail[0].iov_base + base, len - base);
+out:
+	return copied;
+}
+
+/**
+ * csum_partial_copy_to_xdr - checksum and copy data
+ * @xdr: target XDR buffer
+ * @skb: source skb
+ *
+ * We have set things up such that we perform the checksum of the UDP
+ * packet in parallel with the copies into the RPC client iovec.  -DaveM
+ */
+int csum_partial_copy_to_xdr(struct xdr_buf *xdr, struct sk_buff *skb)
+{
+	skb_reader_t	desc;
+
+	desc.skb = skb;
+	desc.offset = sizeof(struct udphdr);
+	desc.count = skb->len - desc.offset;
+
+	if (skb->ip_summed == CHECKSUM_UNNECESSARY)
+		goto no_checksum;
+
+	desc.csum = csum_partial(skb->data, desc.offset, skb->csum);
+	if (xdr_partial_copy_from_skb(xdr, 0, &desc, skb_read_and_csum_bits) < 0)
+		return -1;
+	if (desc.offset != skb->len) {
+		unsigned int csum2;
+		csum2 = skb_checksum(skb, desc.offset, skb->len - desc.offset, 0);
+		desc.csum = csum_block_add(desc.csum, csum2, desc.offset);
+	}
+	if (desc.count)
+		return -1;
+	if ((unsigned short)csum_fold(desc.csum))
+		return -1;
+	return 0;
+no_checksum:
+	if (xdr_partial_copy_from_skb(xdr, 0, &desc, skb_read_bits) < 0)
+		return -1;
+	if (desc.count)
+		return -1;
+	return 0;
+}
diff --git a/net/sunrpc/sunrpc_syms.c b/net/sunrpc/sunrpc_syms.c
index ed48ff022d3..2387e7b823f 100644
--- a/net/sunrpc/sunrpc_syms.c
+++ b/net/sunrpc/sunrpc_syms.c
@@ -10,7 +10,6 @@
 #include <linux/module.h>
 
 #include <linux/types.h>
-#include <linux/socket.h>
 #include <linux/sched.h>
 #include <linux/uio.h>
 #include <linux/unistd.h>
diff --git a/net/sunrpc/svcsock.c b/net/sunrpc/svcsock.c
index 30ec3efc48a..f16e7cdd615 100644
--- a/net/sunrpc/svcsock.c
+++ b/net/sunrpc/svcsock.c
@@ -548,9 +548,6 @@ svc_write_space(struct sock *sk)
 /*
  * Receive a datagram from a UDP socket.
  */
-extern int
-csum_partial_copy_to_xdr(struct xdr_buf *xdr, struct sk_buff *skb);
-
 static int
 svc_udp_recvfrom(struct svc_rqst *rqstp)
 {
@@ -587,7 +584,7 @@ svc_udp_recvfrom(struct svc_rqst *rqstp)
 		struct timeval tv;
 
 		tv.tv_sec = xtime.tv_sec;
-		tv.tv_usec = xtime.tv_nsec * 1000;
+		tv.tv_usec = xtime.tv_nsec / NSEC_PER_USEC;
 		skb_set_timestamp(skb, &tv);
 		/* Don't enable netstamp, sunrpc doesn't 
 		   need that much accuracy */
diff --git a/net/sunrpc/sysctl.c b/net/sunrpc/sysctl.c
index 1b9616a12e2..d0c9f460e41 100644
--- a/net/sunrpc/sysctl.c
+++ b/net/sunrpc/sysctl.c
@@ -119,8 +119,18 @@ done:
 	return 0;
 }
 
+unsigned int xprt_udp_slot_table_entries = RPC_DEF_SLOT_TABLE;
+unsigned int xprt_tcp_slot_table_entries = RPC_DEF_SLOT_TABLE;
+unsigned int xprt_min_resvport = RPC_DEF_MIN_RESVPORT;
+EXPORT_SYMBOL(xprt_min_resvport);
+unsigned int xprt_max_resvport = RPC_DEF_MAX_RESVPORT;
+EXPORT_SYMBOL(xprt_max_resvport);
+
+
 static unsigned int min_slot_table_size = RPC_MIN_SLOT_TABLE;
 static unsigned int max_slot_table_size = RPC_MAX_SLOT_TABLE;
+static unsigned int xprt_min_resvport_limit = RPC_MIN_RESVPORT;
+static unsigned int xprt_max_resvport_limit = RPC_MAX_RESVPORT;
 
 static ctl_table debug_table[] = {
 	{
@@ -177,6 +187,28 @@ static ctl_table debug_table[] = {
 		.extra1		= &min_slot_table_size,
 		.extra2		= &max_slot_table_size
 	},
+	{
+		.ctl_name	= CTL_MIN_RESVPORT,
+		.procname	= "min_resvport",
+		.data		= &xprt_min_resvport,
+		.maxlen		= sizeof(unsigned int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec_minmax,
+		.strategy	= &sysctl_intvec,
+		.extra1		= &xprt_min_resvport_limit,
+		.extra2		= &xprt_max_resvport_limit
+	},
+	{
+		.ctl_name	= CTL_MAX_RESVPORT,
+		.procname	= "max_resvport",
+		.data		= &xprt_max_resvport,
+		.maxlen		= sizeof(unsigned int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec_minmax,
+		.strategy	= &sysctl_intvec,
+		.extra1		= &xprt_min_resvport_limit,
+		.extra2		= &xprt_max_resvport_limit
+	},
 	{ .ctl_name = 0 }
 };
 
diff --git a/net/sunrpc/xdr.c b/net/sunrpc/xdr.c
index fde16f40a58..32df43372ee 100644
--- a/net/sunrpc/xdr.c
+++ b/net/sunrpc/xdr.c
@@ -6,15 +6,12 @@
  * Copyright (C) 1995, 1996 Olaf Kirch <okir@monad.swb.de>
  */
 
+#include <linux/module.h>
 #include <linux/types.h>
-#include <linux/socket.h>
 #include <linux/string.h>
 #include <linux/kernel.h>
 #include <linux/pagemap.h>
 #include <linux/errno.h>
-#include <linux/in.h>
-#include <linux/net.h>
-#include <net/sock.h>
 #include <linux/sunrpc/xdr.h>
 #include <linux/sunrpc/msg_prot.h>
 
@@ -176,178 +173,6 @@ xdr_inline_pages(struct xdr_buf *xdr, unsigned int offset,
 	xdr->buflen += len;
 }
 
-ssize_t
-xdr_partial_copy_from_skb(struct xdr_buf *xdr, unsigned int base,
-			  skb_reader_t *desc,
-			  skb_read_actor_t copy_actor)
-{
-	struct page	**ppage = xdr->pages;
-	unsigned int	len, pglen = xdr->page_len;
-	ssize_t		copied = 0;
-	int		ret;
-
-	len = xdr->head[0].iov_len;
-	if (base < len) {
-		len -= base;
-		ret = copy_actor(desc, (char *)xdr->head[0].iov_base + base, len);
-		copied += ret;
-		if (ret != len || !desc->count)
-			goto out;
-		base = 0;
-	} else
-		base -= len;
-
-	if (pglen == 0)
-		goto copy_tail;
-	if (base >= pglen) {
-		base -= pglen;
-		goto copy_tail;
-	}
-	if (base || xdr->page_base) {
-		pglen -= base;
-		base  += xdr->page_base;
-		ppage += base >> PAGE_CACHE_SHIFT;
-		base &= ~PAGE_CACHE_MASK;
-	}
-	do {
-		char *kaddr;
-
-		/* ACL likes to be lazy in allocating pages - ACLs
-		 * are small by default but can get huge. */
-		if (unlikely(*ppage == NULL)) {
-			*ppage = alloc_page(GFP_ATOMIC);
-			if (unlikely(*ppage == NULL)) {
-				if (copied == 0)
-					copied = -ENOMEM;
-				goto out;
-			}
-		}
-
-		len = PAGE_CACHE_SIZE;
-		kaddr = kmap_atomic(*ppage, KM_SKB_SUNRPC_DATA);
-		if (base) {
-			len -= base;
-			if (pglen < len)
-				len = pglen;
-			ret = copy_actor(desc, kaddr + base, len);
-			base = 0;
-		} else {
-			if (pglen < len)
-				len = pglen;
-			ret = copy_actor(desc, kaddr, len);
-		}
-		flush_dcache_page(*ppage);
-		kunmap_atomic(kaddr, KM_SKB_SUNRPC_DATA);
-		copied += ret;
-		if (ret != len || !desc->count)
-			goto out;
-		ppage++;
-	} while ((pglen -= len) != 0);
-copy_tail:
-	len = xdr->tail[0].iov_len;
-	if (base < len)
-		copied += copy_actor(desc, (char *)xdr->tail[0].iov_base + base, len - base);
-out:
-	return copied;
-}
-
-
-int
-xdr_sendpages(struct socket *sock, struct sockaddr *addr, int addrlen,
-		struct xdr_buf *xdr, unsigned int base, int msgflags)
-{
-	struct page **ppage = xdr->pages;
-	unsigned int len, pglen = xdr->page_len;
-	int err, ret = 0;
-	ssize_t (*sendpage)(struct socket *, struct page *, int, size_t, int);
-
-	len = xdr->head[0].iov_len;
-	if (base < len || (addr != NULL && base == 0)) {
-		struct kvec iov = {
-			.iov_base = xdr->head[0].iov_base + base,
-			.iov_len  = len - base,
-		};
-		struct msghdr msg = {
-			.msg_name    = addr,
-			.msg_namelen = addrlen,
-			.msg_flags   = msgflags,
-		};
-		if (xdr->len > len)
-			msg.msg_flags |= MSG_MORE;
-
-		if (iov.iov_len != 0)
-			err = kernel_sendmsg(sock, &msg, &iov, 1, iov.iov_len);
-		else
-			err = kernel_sendmsg(sock, &msg, NULL, 0, 0);
-		if (ret == 0)
-			ret = err;
-		else if (err > 0)
-			ret += err;
-		if (err != iov.iov_len)
-			goto out;
-		base = 0;
-	} else
-		base -= len;
-
-	if (pglen == 0)
-		goto copy_tail;
-	if (base >= pglen) {
-		base -= pglen;
-		goto copy_tail;
-	}
-	if (base || xdr->page_base) {
-		pglen -= base;
-		base  += xdr->page_base;
-		ppage += base >> PAGE_CACHE_SHIFT;
-		base &= ~PAGE_CACHE_MASK;
-	}
-
-	sendpage = sock->ops->sendpage ? : sock_no_sendpage;
-	do {
-		int flags = msgflags;
-
-		len = PAGE_CACHE_SIZE;
-		if (base)
-			len -= base;
-		if (pglen < len)
-			len = pglen;
-
-		if (pglen != len || xdr->tail[0].iov_len != 0)
-			flags |= MSG_MORE;
-
-		/* Hmm... We might be dealing with highmem pages */
-		if (PageHighMem(*ppage))
-			sendpage = sock_no_sendpage;
-		err = sendpage(sock, *ppage, base, len, flags);
-		if (ret == 0)
-			ret = err;
-		else if (err > 0)
-			ret += err;
-		if (err != len)
-			goto out;
-		base = 0;
-		ppage++;
-	} while ((pglen -= len) != 0);
-copy_tail:
-	len = xdr->tail[0].iov_len;
-	if (base < len) {
-		struct kvec iov = {
-			.iov_base = xdr->tail[0].iov_base + base,
-			.iov_len  = len - base,
-		};
-		struct msghdr msg = {
-			.msg_flags   = msgflags,
-		};
-		err = kernel_sendmsg(sock, &msg, &iov, 1, iov.iov_len);
-		if (ret == 0)
-			ret = err;
-		else if (err > 0)
-			ret += err;
-	}
-out:
-	return ret;
-}
-
 
 /*
  * Helper routines for doing 'memmove' like operations on a struct xdr_buf
diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index 3c654e06b08..6dda3860351 100644
--- a/net/sunrpc/xprt.c
+++ b/net/sunrpc/xprt.c
@@ -10,12 +10,12 @@
  *	one is available. Otherwise, it sleeps on the backlog queue
  *	(xprt_reserve).
  *  -	Next, the caller puts together the RPC message, stuffs it into
- *	the request struct, and calls xprt_call().
- *  -	xprt_call transmits the message and installs the caller on the
- *	socket's wait list. At the same time, it installs a timer that
+ *	the request struct, and calls xprt_transmit().
+ *  -	xprt_transmit sends the message and installs the caller on the
+ *	transport's wait list. At the same time, it installs a timer that
  *	is run after the packet's timeout has expired.
  *  -	When a packet arrives, the data_ready handler walks the list of
- *	pending requests for that socket. If a matching XID is found, the
+ *	pending requests for that transport. If a matching XID is found, the
  *	caller is woken up, and the timer removed.
  *  -	When no reply arrives within the timeout interval, the timer is
  *	fired by the kernel and runs xprt_timer(). It either adjusts the
@@ -33,36 +33,17 @@
  *
  *  Copyright (C) 1995-1997, Olaf Kirch <okir@monad.swb.de>
  *
- *  TCP callback races fixes (C) 1998 Red Hat Software <alan@redhat.com>
- *  TCP send fixes (C) 1998 Red Hat Software <alan@redhat.com>
- *  TCP NFS related read + write fixes
- *   (C) 1999 Dave Airlie, University of Limerick, Ireland <airlied@linux.ie>
- *
- *  Rewrite of larges part of the code in order to stabilize TCP stuff.
- *  Fix behaviour when socket buffer is full.
- *   (C) 1999 Trond Myklebust <trond.myklebust@fys.uio.no>
+ *  Transport switch API copyright (C) 2005, Chuck Lever <cel@netapp.com>
  */
 
+#include <linux/module.h>
+
 #include <linux/types.h>
-#include <linux/slab.h>
-#include <linux/capability.h>
-#include <linux/sched.h>
-#include <linux/errno.h>
-#include <linux/socket.h>
-#include <linux/in.h>
-#include <linux/net.h>
-#include <linux/mm.h>
-#include <linux/udp.h>
-#include <linux/tcp.h>
-#include <linux/sunrpc/clnt.h>
-#include <linux/file.h>
+#include <linux/interrupt.h>
 #include <linux/workqueue.h>
 #include <linux/random.h>
 
-#include <net/sock.h>
-#include <net/checksum.h>
-#include <net/udp.h>
-#include <net/tcp.h>
+#include <linux/sunrpc/clnt.h>
 
 /*
  * Local variables
@@ -73,81 +54,90 @@
 # define RPCDBG_FACILITY	RPCDBG_XPRT
 #endif
 
-#define XPRT_MAX_BACKOFF	(8)
-#define XPRT_IDLE_TIMEOUT	(5*60*HZ)
-#define XPRT_MAX_RESVPORT	(800)
-
 /*
  * Local functions
  */
 static void	xprt_request_init(struct rpc_task *, struct rpc_xprt *);
 static inline void	do_xprt_reserve(struct rpc_task *);
-static void	xprt_disconnect(struct rpc_xprt *);
 static void	xprt_connect_status(struct rpc_task *task);
-static struct rpc_xprt * xprt_setup(int proto, struct sockaddr_in *ap,
-						struct rpc_timeout *to);
-static struct socket *xprt_create_socket(struct rpc_xprt *, int, int);
-static void	xprt_bind_socket(struct rpc_xprt *, struct socket *);
 static int      __xprt_get_cong(struct rpc_xprt *, struct rpc_task *);
 
-static int	xprt_clear_backlog(struct rpc_xprt *xprt);
-
-#ifdef RPC_DEBUG_DATA
 /*
- * Print the buffer contents (first 128 bytes only--just enough for
- * diropres return).
+ * The transport code maintains an estimate on the maximum number of out-
+ * standing RPC requests, using a smoothed version of the congestion
+ * avoidance implemented in 44BSD. This is basically the Van Jacobson
+ * congestion algorithm: If a retransmit occurs, the congestion window is
+ * halved; otherwise, it is incremented by 1/cwnd when
+ *
+ *	-	a reply is received and
+ *	-	a full number of requests are outstanding and
+ *	-	the congestion window hasn't been updated recently.
  */
-static void
-xprt_pktdump(char *msg, u32 *packet, unsigned int count)
-{
-	u8	*buf = (u8 *) packet;
-	int	j;
-
-	dprintk("RPC:      %s\n", msg);
-	for (j = 0; j < count && j < 128; j += 4) {
-		if (!(j & 31)) {
-			if (j)
-				dprintk("\n");
-			dprintk("0x%04x ", j);
-		}
-		dprintk("%02x%02x%02x%02x ",
-			buf[j], buf[j+1], buf[j+2], buf[j+3]);
-	}
-	dprintk("\n");
-}
-#else
-static inline void
-xprt_pktdump(char *msg, u32 *packet, unsigned int count)
-{
-	/* NOP */
-}
-#endif
+#define RPC_CWNDSHIFT		(8U)
+#define RPC_CWNDSCALE		(1U << RPC_CWNDSHIFT)
+#define RPC_INITCWND		RPC_CWNDSCALE
+#define RPC_MAXCWND(xprt)	((xprt)->max_reqs << RPC_CWNDSHIFT)
 
-/*
- * Look up RPC transport given an INET socket
+#define RPCXPRT_CONGESTED(xprt) ((xprt)->cong >= (xprt)->cwnd)
+
+/**
+ * xprt_reserve_xprt - serialize write access to transports
+ * @task: task that is requesting access to the transport
+ *
+ * This prevents mixing the payload of separate requests, and prevents
+ * transport connects from colliding with writes.  No congestion control
+ * is provided.
  */
-static inline struct rpc_xprt *
-xprt_from_sock(struct sock *sk)
+int xprt_reserve_xprt(struct rpc_task *task)
 {
-	return (struct rpc_xprt *) sk->sk_user_data;
+	struct rpc_xprt	*xprt = task->tk_xprt;
+	struct rpc_rqst *req = task->tk_rqstp;
+
+	if (test_and_set_bit(XPRT_LOCKED, &xprt->state)) {
+		if (task == xprt->snd_task)
+			return 1;
+		if (task == NULL)
+			return 0;
+		goto out_sleep;
+	}
+	xprt->snd_task = task;
+	if (req) {
+		req->rq_bytes_sent = 0;
+		req->rq_ntrans++;
+	}
+	return 1;
+
+out_sleep:
+	dprintk("RPC: %4d failed to lock transport %p\n",
+			task->tk_pid, xprt);
+	task->tk_timeout = 0;
+	task->tk_status = -EAGAIN;
+	if (req && req->rq_ntrans)
+		rpc_sleep_on(&xprt->resend, task, NULL, NULL);
+	else
+		rpc_sleep_on(&xprt->sending, task, NULL, NULL);
+	return 0;
 }
 
 /*
- * Serialize write access to sockets, in order to prevent different
- * requests from interfering with each other.
- * Also prevents TCP socket connects from colliding with writes.
+ * xprt_reserve_xprt_cong - serialize write access to transports
+ * @task: task that is requesting access to the transport
+ *
+ * Same as xprt_reserve_xprt, but Van Jacobson congestion control is
+ * integrated into the decision of whether a request is allowed to be
+ * woken up and given access to the transport.
  */
-static int
-__xprt_lock_write(struct rpc_xprt *xprt, struct rpc_task *task)
+int xprt_reserve_xprt_cong(struct rpc_task *task)
 {
+	struct rpc_xprt	*xprt = task->tk_xprt;
 	struct rpc_rqst *req = task->tk_rqstp;
 
-	if (test_and_set_bit(XPRT_LOCKED, &xprt->sockstate)) {
+	if (test_and_set_bit(XPRT_LOCKED, &xprt->state)) {
 		if (task == xprt->snd_task)
 			return 1;
 		goto out_sleep;
 	}
-	if (xprt->nocong || __xprt_get_cong(xprt, task)) {
+	if (__xprt_get_cong(xprt, task)) {
 		xprt->snd_task = task;
 		if (req) {
 			req->rq_bytes_sent = 0;
@@ -156,10 +146,10 @@ __xprt_lock_write(struct rpc_xprt *xprt, struct rpc_task *task)
 		return 1;
 	}
 	smp_mb__before_clear_bit();
-	clear_bit(XPRT_LOCKED, &xprt->sockstate);
+	clear_bit(XPRT_LOCKED, &xprt->state);
 	smp_mb__after_clear_bit();
 out_sleep:
-	dprintk("RPC: %4d failed to lock socket %p\n", task->tk_pid, xprt);
+	dprintk("RPC: %4d failed to lock transport %p\n", task->tk_pid, xprt);
 	task->tk_timeout = 0;
 	task->tk_status = -EAGAIN;
 	if (req && req->rq_ntrans)
@@ -169,26 +159,52 @@ out_sleep:
 	return 0;
 }
 
-static inline int
-xprt_lock_write(struct rpc_xprt *xprt, struct rpc_task *task)
+static inline int xprt_lock_write(struct rpc_xprt *xprt, struct rpc_task *task)
 {
 	int retval;
 
-	spin_lock_bh(&xprt->sock_lock);
-	retval = __xprt_lock_write(xprt, task);
-	spin_unlock_bh(&xprt->sock_lock);
+	spin_lock_bh(&xprt->transport_lock);
+	retval = xprt->ops->reserve_xprt(task);
+	spin_unlock_bh(&xprt->transport_lock);
 	return retval;
 }
 
+static void __xprt_lock_write_next(struct rpc_xprt *xprt)
+{
+	struct rpc_task *task;
+	struct rpc_rqst *req;
 
-static void
-__xprt_lock_write_next(struct rpc_xprt *xprt)
+	if (test_and_set_bit(XPRT_LOCKED, &xprt->state))
+		return;
+
+	task = rpc_wake_up_next(&xprt->resend);
+	if (!task) {
+		task = rpc_wake_up_next(&xprt->sending);
+		if (!task)
+			goto out_unlock;
+	}
+
+	req = task->tk_rqstp;
+	xprt->snd_task = task;
+	if (req) {
+		req->rq_bytes_sent = 0;
+		req->rq_ntrans++;
+	}
+	return;
+
+out_unlock:
+	smp_mb__before_clear_bit();
+	clear_bit(XPRT_LOCKED, &xprt->state);
+	smp_mb__after_clear_bit();
+}
+
+static void __xprt_lock_write_next_cong(struct rpc_xprt *xprt)
 {
 	struct rpc_task *task;
 
-	if (test_and_set_bit(XPRT_LOCKED, &xprt->sockstate))
+	if (test_and_set_bit(XPRT_LOCKED, &xprt->state))
 		return;
-	if (!xprt->nocong && RPCXPRT_CONGESTED(xprt))
+	if (RPCXPRT_CONGESTED(xprt))
 		goto out_unlock;
 	task = rpc_wake_up_next(&xprt->resend);
 	if (!task) {
@@ -196,7 +212,7 @@ __xprt_lock_write_next(struct rpc_xprt *xprt)
 		if (!task)
 			goto out_unlock;
 	}
-	if (xprt->nocong || __xprt_get_cong(xprt, task)) {
+	if (__xprt_get_cong(xprt, task)) {
 		struct rpc_rqst *req = task->tk_rqstp;
 		xprt->snd_task = task;
 		if (req) {
@@ -207,87 +223,52 @@ __xprt_lock_write_next(struct rpc_xprt *xprt)
 	}
 out_unlock:
 	smp_mb__before_clear_bit();
-	clear_bit(XPRT_LOCKED, &xprt->sockstate);
+	clear_bit(XPRT_LOCKED, &xprt->state);
 	smp_mb__after_clear_bit();
 }
 
-/*
- * Releases the socket for use by other requests.
+/**
+ * xprt_release_xprt - allow other requests to use a transport
+ * @xprt: transport with other tasks potentially waiting
+ * @task: task that is releasing access to the transport
+ *
+ * Note that "task" can be NULL.  No congestion control is provided.
  */
-static void
-__xprt_release_write(struct rpc_xprt *xprt, struct rpc_task *task)
+void xprt_release_xprt(struct rpc_xprt *xprt, struct rpc_task *task)
 {
 	if (xprt->snd_task == task) {
 		xprt->snd_task = NULL;
 		smp_mb__before_clear_bit();
-		clear_bit(XPRT_LOCKED, &xprt->sockstate);
+		clear_bit(XPRT_LOCKED, &xprt->state);
 		smp_mb__after_clear_bit();
 		__xprt_lock_write_next(xprt);
 	}
 }
 
-static inline void
-xprt_release_write(struct rpc_xprt *xprt, struct rpc_task *task)
-{
-	spin_lock_bh(&xprt->sock_lock);
-	__xprt_release_write(xprt, task);
-	spin_unlock_bh(&xprt->sock_lock);
-}
-
-/*
- * Write data to socket.
+/**
+ * xprt_release_xprt_cong - allow other requests to use a transport
+ * @xprt: transport with other tasks potentially waiting
+ * @task: task that is releasing access to the transport
+ *
+ * Note that "task" can be NULL.  Another task is awoken to use the
+ * transport if the transport's congestion window allows it.
  */
-static inline int
-xprt_sendmsg(struct rpc_xprt *xprt, struct rpc_rqst *req)
+void xprt_release_xprt_cong(struct rpc_xprt *xprt, struct rpc_task *task)
 {
-	struct socket	*sock = xprt->sock;
-	struct xdr_buf	*xdr = &req->rq_snd_buf;
-	struct sockaddr *addr = NULL;
-	int addrlen = 0;
-	unsigned int	skip;
-	int		result;
-
-	if (!sock)
-		return -ENOTCONN;
-
-	xprt_pktdump("packet data:",
-				req->rq_svec->iov_base,
-				req->rq_svec->iov_len);
-
-	/* For UDP, we need to provide an address */
-	if (!xprt->stream) {
-		addr = (struct sockaddr *) &xprt->addr;
-		addrlen = sizeof(xprt->addr);
+	if (xprt->snd_task == task) {
+		xprt->snd_task = NULL;
+		smp_mb__before_clear_bit();
+		clear_bit(XPRT_LOCKED, &xprt->state);
+		smp_mb__after_clear_bit();
+		__xprt_lock_write_next_cong(xprt);
 	}
-	/* Dont repeat bytes */
-	skip = req->rq_bytes_sent;
-
-	clear_bit(SOCK_ASYNC_NOSPACE, &sock->flags);
-	result = xdr_sendpages(sock, addr, addrlen, xdr, skip, MSG_DONTWAIT);
-
-	dprintk("RPC:      xprt_sendmsg(%d) = %d\n", xdr->len - skip, result);
-
-	if (result >= 0)
-		return result;
+}
 
-	switch (result) {
-	case -ECONNREFUSED:
-		/* When the server has died, an ICMP port unreachable message
-		 * prompts ECONNREFUSED.
-		 */
-	case -EAGAIN:
-		break;
-	case -ECONNRESET:
-	case -ENOTCONN:
-	case -EPIPE:
-		/* connection broken */
-		if (xprt->stream)
-			result = -ENOTCONN;
-		break;
-	default:
-		printk(KERN_NOTICE "RPC: sendmsg returned error %d\n", -result);
-	}
-	return result;
+static inline void xprt_release_write(struct rpc_xprt *xprt, struct rpc_task *task)
+{
+	spin_lock_bh(&xprt->transport_lock);
+	xprt->ops->release_xprt(xprt, task);
+	spin_unlock_bh(&xprt->transport_lock);
 }
 
 /*
@@ -321,26 +302,40 @@ __xprt_put_cong(struct rpc_xprt *xprt, struct rpc_rqst *req)
 		return;
 	req->rq_cong = 0;
 	xprt->cong -= RPC_CWNDSCALE;
-	__xprt_lock_write_next(xprt);
+	__xprt_lock_write_next_cong(xprt);
 }
 
-/*
- * Adjust RPC congestion window
+/**
+ * xprt_release_rqst_cong - housekeeping when request is complete
+ * @task: RPC request that recently completed
+ *
+ * Useful for transports that require congestion control.
+ */
+void xprt_release_rqst_cong(struct rpc_task *task)
+{
+	__xprt_put_cong(task->tk_xprt, task->tk_rqstp);
+}
+
+/**
+ * xprt_adjust_cwnd - adjust transport congestion window
+ * @task: recently completed RPC request used to adjust window
+ * @result: result code of completed RPC request
+ *
  * We use a time-smoothed congestion estimator to avoid heavy oscillation.
  */
-static void
-xprt_adjust_cwnd(struct rpc_xprt *xprt, int result)
+void xprt_adjust_cwnd(struct rpc_task *task, int result)
 {
-	unsigned long	cwnd;
+	struct rpc_rqst *req = task->tk_rqstp;
+	struct rpc_xprt *xprt = task->tk_xprt;
+	unsigned long cwnd = xprt->cwnd;
 
-	cwnd = xprt->cwnd;
 	if (result >= 0 && cwnd <= xprt->cong) {
 		/* The (cwnd >> 1) term makes sure
 		 * the result gets rounded properly. */
 		cwnd += (RPC_CWNDSCALE * RPC_CWNDSCALE + (cwnd >> 1)) / cwnd;
 		if (cwnd > RPC_MAXCWND(xprt))
 			cwnd = RPC_MAXCWND(xprt);
-		__xprt_lock_write_next(xprt);
+		__xprt_lock_write_next_cong(xprt);
 	} else if (result == -ETIMEDOUT) {
 		cwnd >>= 1;
 		if (cwnd < RPC_CWNDSCALE)
@@ -349,11 +344,89 @@ xprt_adjust_cwnd(struct rpc_xprt *xprt, int result)
 	dprintk("RPC:      cong %ld, cwnd was %ld, now %ld\n",
 			xprt->cong, xprt->cwnd, cwnd);
 	xprt->cwnd = cwnd;
+	__xprt_put_cong(xprt, req);
+}
+
+/**
+ * xprt_wake_pending_tasks - wake all tasks on a transport's pending queue
+ * @xprt: transport with waiting tasks
+ * @status: result code to plant in each task before waking it
+ *
+ */
+void xprt_wake_pending_tasks(struct rpc_xprt *xprt, int status)
+{
+	if (status < 0)
+		rpc_wake_up_status(&xprt->pending, status);
+	else
+		rpc_wake_up(&xprt->pending);
+}
+
+/**
+ * xprt_wait_for_buffer_space - wait for transport output buffer to clear
+ * @task: task to be put to sleep
+ *
+ */
+void xprt_wait_for_buffer_space(struct rpc_task *task)
+{
+	struct rpc_rqst *req = task->tk_rqstp;
+	struct rpc_xprt *xprt = req->rq_xprt;
+
+	task->tk_timeout = req->rq_timeout;
+	rpc_sleep_on(&xprt->pending, task, NULL, NULL);
+}
+
+/**
+ * xprt_write_space - wake the task waiting for transport output buffer space
+ * @xprt: transport with waiting tasks
+ *
+ * Can be called in a soft IRQ context, so xprt_write_space never sleeps.
+ */
+void xprt_write_space(struct rpc_xprt *xprt)
+{
+	if (unlikely(xprt->shutdown))
+		return;
+
+	spin_lock_bh(&xprt->transport_lock);
+	if (xprt->snd_task) {
+		dprintk("RPC:      write space: waking waiting task on xprt %p\n",
+				xprt);
+		rpc_wake_up_task(xprt->snd_task);
+	}
+	spin_unlock_bh(&xprt->transport_lock);
+}
+
+/**
+ * xprt_set_retrans_timeout_def - set a request's retransmit timeout
+ * @task: task whose timeout is to be set
+ *
+ * Set a request's retransmit timeout based on the transport's
+ * default timeout parameters.  Used by transports that don't adjust
+ * the retransmit timeout based on round-trip time estimation.
+ */
+void xprt_set_retrans_timeout_def(struct rpc_task *task)
+{
+	task->tk_timeout = task->tk_rqstp->rq_timeout;
 }
 
 /*
- * Reset the major timeout value
+ * xprt_set_retrans_timeout_rtt - set a request's retransmit timeout
+ * @task: task whose timeout is to be set
+ * 
+ * Set a request's retransmit timeout using the RTT estimator.
  */
+void xprt_set_retrans_timeout_rtt(struct rpc_task *task)
+{
+	int timer = task->tk_msg.rpc_proc->p_timer;
+	struct rpc_rtt *rtt = task->tk_client->cl_rtt;
+	struct rpc_rqst *req = task->tk_rqstp;
+	unsigned long max_timeout = req->rq_xprt->timeout.to_maxval;
+
+	task->tk_timeout = rpc_calc_rto(rtt, timer);
+	task->tk_timeout <<= rpc_ntimeo(rtt, timer) + req->rq_retries;
+	if (task->tk_timeout > max_timeout || task->tk_timeout == 0)
+		task->tk_timeout = max_timeout;
+}
+
 static void xprt_reset_majortimeo(struct rpc_rqst *req)
 {
 	struct rpc_timeout *to = &req->rq_xprt->timeout;
@@ -368,8 +441,10 @@ static void xprt_reset_majortimeo(struct rpc_rqst *req)
 	req->rq_majortimeo += jiffies;
 }
 
-/*
- * Adjust timeout values etc for next retransmit
+/**
+ * xprt_adjust_timeout - adjust timeout values for next retransmit
+ * @req: RPC request containing parameters to use for the adjustment
+ *
  */
 int xprt_adjust_timeout(struct rpc_rqst *req)
 {
@@ -391,9 +466,9 @@ int xprt_adjust_timeout(struct rpc_rqst *req)
 		req->rq_retries = 0;
 		xprt_reset_majortimeo(req);
 		/* Reset the RTT counters == "slow start" */
-		spin_lock_bh(&xprt->sock_lock);
+		spin_lock_bh(&xprt->transport_lock);
 		rpc_init_rtt(req->rq_task->tk_client->cl_rtt, to->to_initval);
-		spin_unlock_bh(&xprt->sock_lock);
+		spin_unlock_bh(&xprt->transport_lock);
 		pprintk("RPC: %lu timeout\n", jiffies);
 		status = -ETIMEDOUT;
 	}
@@ -405,133 +480,52 @@ int xprt_adjust_timeout(struct rpc_rqst *req)
 	return status;
 }
 
-/*
- * Close down a transport socket
- */
-static void
-xprt_close(struct rpc_xprt *xprt)
-{
-	struct socket	*sock = xprt->sock;
-	struct sock	*sk = xprt->inet;
-
-	if (!sk)
-		return;
-
-	write_lock_bh(&sk->sk_callback_lock);
-	xprt->inet = NULL;
-	xprt->sock = NULL;
-
-	sk->sk_user_data    = NULL;
-	sk->sk_data_ready   = xprt->old_data_ready;
-	sk->sk_state_change = xprt->old_state_change;
-	sk->sk_write_space  = xprt->old_write_space;
-	write_unlock_bh(&sk->sk_callback_lock);
-
-	sk->sk_no_check	 = 0;
-
-	sock_release(sock);
-}
-
-static void
-xprt_socket_autoclose(void *args)
+static void xprt_autoclose(void *args)
 {
 	struct rpc_xprt *xprt = (struct rpc_xprt *)args;
 
 	xprt_disconnect(xprt);
-	xprt_close(xprt);
+	xprt->ops->close(xprt);
 	xprt_release_write(xprt, NULL);
 }
 
-/*
- * Mark a transport as disconnected
+/**
+ * xprt_disconnect - mark a transport as disconnected
+ * @xprt: transport to flag for disconnect
+ *
  */
-static void
-xprt_disconnect(struct rpc_xprt *xprt)
+void xprt_disconnect(struct rpc_xprt *xprt)
 {
 	dprintk("RPC:      disconnected transport %p\n", xprt);
-	spin_lock_bh(&xprt->sock_lock);
+	spin_lock_bh(&xprt->transport_lock);
 	xprt_clear_connected(xprt);
-	rpc_wake_up_status(&xprt->pending, -ENOTCONN);
-	spin_unlock_bh(&xprt->sock_lock);
+	xprt_wake_pending_tasks(xprt, -ENOTCONN);
+	spin_unlock_bh(&xprt->transport_lock);
 }
 
-/*
- * Used to allow disconnection when we've been idle
- */
 static void
 xprt_init_autodisconnect(unsigned long data)
 {
 	struct rpc_xprt *xprt = (struct rpc_xprt *)data;
 
-	spin_lock(&xprt->sock_lock);
+	spin_lock(&xprt->transport_lock);
 	if (!list_empty(&xprt->recv) || xprt->shutdown)
 		goto out_abort;
-	if (test_and_set_bit(XPRT_LOCKED, &xprt->sockstate))
+	if (test_and_set_bit(XPRT_LOCKED, &xprt->state))
 		goto out_abort;
-	spin_unlock(&xprt->sock_lock);
-	/* Let keventd close the socket */
-	if (test_bit(XPRT_CONNECTING, &xprt->sockstate) != 0)
+	spin_unlock(&xprt->transport_lock);
+	if (xprt_connecting(xprt))
 		xprt_release_write(xprt, NULL);
 	else
 		schedule_work(&xprt->task_cleanup);
 	return;
 out_abort:
-	spin_unlock(&xprt->sock_lock);
-}
-
-static void xprt_socket_connect(void *args)
-{
-	struct rpc_xprt *xprt = (struct rpc_xprt *)args;
-	struct socket *sock = xprt->sock;
-	int status = -EIO;
-
-	if (xprt->shutdown || xprt->addr.sin_port == 0)
-		goto out;
-
-	/*
-	 * Start by resetting any existing state
-	 */
-	xprt_close(xprt);
-	sock = xprt_create_socket(xprt, xprt->prot, xprt->resvport);
-	if (sock == NULL) {
-		/* couldn't create socket or bind to reserved port;
-		 * this is likely a permanent error, so cause an abort */
-		goto out;
-	}
-	xprt_bind_socket(xprt, sock);
-	xprt_sock_setbufsize(xprt);
-
-	status = 0;
-	if (!xprt->stream)
-		goto out;
-
-	/*
-	 * Tell the socket layer to start connecting...
-	 */
-	status = sock->ops->connect(sock, (struct sockaddr *) &xprt->addr,
-			sizeof(xprt->addr), O_NONBLOCK);
-	dprintk("RPC: %p  connect status %d connected %d sock state %d\n",
-			xprt, -status, xprt_connected(xprt), sock->sk->sk_state);
-	if (status < 0) {
-		switch (status) {
-			case -EINPROGRESS:
-			case -EALREADY:
-				goto out_clear;
-		}
-	}
-out:
-	if (status < 0)
-		rpc_wake_up_status(&xprt->pending, status);
-	else
-		rpc_wake_up(&xprt->pending);
-out_clear:
-	smp_mb__before_clear_bit();
-	clear_bit(XPRT_CONNECTING, &xprt->sockstate);
-	smp_mb__after_clear_bit();
+	spin_unlock(&xprt->transport_lock);
 }
 
-/*
- * Attempt to connect a TCP socket.
+/**
+ * xprt_connect - schedule a transport connect operation
+ * @task: RPC task that is requesting the connect
  *
  */
 void xprt_connect(struct rpc_task *task)
@@ -552,37 +546,19 @@ void xprt_connect(struct rpc_task *task)
 	if (!xprt_lock_write(xprt, task))
 		return;
 	if (xprt_connected(xprt))
-		goto out_write;
+		xprt_release_write(xprt, task);
+	else {
+		if (task->tk_rqstp)
+			task->tk_rqstp->rq_bytes_sent = 0;
 
-	if (task->tk_rqstp)
-		task->tk_rqstp->rq_bytes_sent = 0;
-
-	task->tk_timeout = RPC_CONNECT_TIMEOUT;
-	rpc_sleep_on(&xprt->pending, task, xprt_connect_status, NULL);
-	if (!test_and_set_bit(XPRT_CONNECTING, &xprt->sockstate)) {
-		/* Note: if we are here due to a dropped connection
-		 * 	 we delay reconnecting by RPC_REESTABLISH_TIMEOUT/HZ
-		 * 	 seconds
-		 */
-		if (xprt->sock != NULL)
-			schedule_delayed_work(&xprt->sock_connect,
-					RPC_REESTABLISH_TIMEOUT);
-		else {
-			schedule_work(&xprt->sock_connect);
-			if (!RPC_IS_ASYNC(task))
-				flush_scheduled_work();
-		}
+		task->tk_timeout = xprt->connect_timeout;
+		rpc_sleep_on(&xprt->pending, task, xprt_connect_status, NULL);
+		xprt->ops->connect(task);
 	}
 	return;
- out_write:
-	xprt_release_write(xprt, task);
 }
 
-/*
- * We arrive here when awoken from waiting on connection establishment.
- */
-static void
-xprt_connect_status(struct rpc_task *task)
+static void xprt_connect_status(struct rpc_task *task)
 {
 	struct rpc_xprt	*xprt = task->tk_xprt;
 
@@ -592,31 +568,42 @@ xprt_connect_status(struct rpc_task *task)
 		return;
 	}
 
-	/* if soft mounted, just cause this RPC to fail */
-	if (RPC_IS_SOFT(task))
-		task->tk_status = -EIO;
-
 	switch (task->tk_status) {
 	case -ECONNREFUSED:
 	case -ECONNRESET:
+		dprintk("RPC: %4d xprt_connect_status: server %s refused connection\n",
+				task->tk_pid, task->tk_client->cl_server);
+		break;
 	case -ENOTCONN:
-		return;
+		dprintk("RPC: %4d xprt_connect_status: connection broken\n",
+				task->tk_pid);
+		break;
 	case -ETIMEDOUT:
-		dprintk("RPC: %4d xprt_connect_status: timed out\n",
+		dprintk("RPC: %4d xprt_connect_status: connect attempt timed out\n",
 				task->tk_pid);
 		break;
 	default:
-		printk(KERN_ERR "RPC: error %d connecting to server %s\n",
-				-task->tk_status, task->tk_client->cl_server);
+		dprintk("RPC: %4d xprt_connect_status: error %d connecting to server %s\n",
+				task->tk_pid, -task->tk_status, task->tk_client->cl_server);
+		xprt_release_write(xprt, task);
+		task->tk_status = -EIO;
+		return;
+	}
+
+	/* if soft mounted, just cause this RPC to fail */
+	if (RPC_IS_SOFT(task)) {
+		xprt_release_write(xprt, task);
+		task->tk_status = -EIO;
 	}
-	xprt_release_write(xprt, task);
 }
 
-/*
- * Look up the RPC request corresponding to a reply, and then lock it.
+/**
+ * xprt_lookup_rqst - find an RPC request corresponding to an XID
+ * @xprt: transport on which the original request was transmitted
+ * @xid: RPC XID of incoming reply
+ *
  */
-static inline struct rpc_rqst *
-xprt_lookup_rqst(struct rpc_xprt *xprt, u32 xid)
+struct rpc_rqst *xprt_lookup_rqst(struct rpc_xprt *xprt, u32 xid)
 {
 	struct list_head *pos;
 	struct rpc_rqst	*req = NULL;
@@ -631,556 +618,68 @@ xprt_lookup_rqst(struct rpc_xprt *xprt, u32 xid)
 	return req;
 }
 
-/*
- * Complete reply received.
- * The TCP code relies on us to remove the request from xprt->pending.
- */
-static void
-xprt_complete_rqst(struct rpc_xprt *xprt, struct rpc_rqst *req, int copied)
-{
-	struct rpc_task	*task = req->rq_task;
-	struct rpc_clnt *clnt = task->tk_client;
-
-	/* Adjust congestion window */
-	if (!xprt->nocong) {
-		unsigned timer = task->tk_msg.rpc_proc->p_timer;
-		xprt_adjust_cwnd(xprt, copied);
-		__xprt_put_cong(xprt, req);
-		if (timer) {
-			if (req->rq_ntrans == 1)
-				rpc_update_rtt(clnt->cl_rtt, timer,
-						(long)jiffies - req->rq_xtime);
-			rpc_set_timeo(clnt->cl_rtt, timer, req->rq_ntrans - 1);
-		}
-	}
-
-#ifdef RPC_PROFILE
-	/* Profile only reads for now */
-	if (copied > 1024) {
-		static unsigned long	nextstat;
-		static unsigned long	pkt_rtt, pkt_len, pkt_cnt;
-
-		pkt_cnt++;
-		pkt_len += req->rq_slen + copied;
-		pkt_rtt += jiffies - req->rq_xtime;
-		if (time_before(nextstat, jiffies)) {
-			printk("RPC: %lu %ld cwnd\n", jiffies, xprt->cwnd);
-			printk("RPC: %ld %ld %ld %ld stat\n",
-					jiffies, pkt_cnt, pkt_len, pkt_rtt);
-			pkt_rtt = pkt_len = pkt_cnt = 0;
-			nextstat = jiffies + 5 * HZ;
-		}
-	}
-#endif
-
-	dprintk("RPC: %4d has input (%d bytes)\n", task->tk_pid, copied);
-	list_del_init(&req->rq_list);
-	req->rq_received = req->rq_private_buf.len = copied;
-
-	/* ... and wake up the process. */
-	rpc_wake_up_task(task);
-	return;
-}
-
-static size_t
-skb_read_bits(skb_reader_t *desc, void *to, size_t len)
-{
-	if (len > desc->count)
-		len = desc->count;
-	if (skb_copy_bits(desc->skb, desc->offset, to, len))
-		return 0;
-	desc->count -= len;
-	desc->offset += len;
-	return len;
-}
-
-static size_t
-skb_read_and_csum_bits(skb_reader_t *desc, void *to, size_t len)
-{
-	unsigned int csum2, pos;
-
-	if (len > desc->count)
-		len = desc->count;
-	pos = desc->offset;
-	csum2 = skb_copy_and_csum_bits(desc->skb, pos, to, len, 0);
-	desc->csum = csum_block_add(desc->csum, csum2, pos);
-	desc->count -= len;
-	desc->offset += len;
-	return len;
-}
-
-/*
- * We have set things up such that we perform the checksum of the UDP
- * packet in parallel with the copies into the RPC client iovec.  -DaveM
- */
-int
-csum_partial_copy_to_xdr(struct xdr_buf *xdr, struct sk_buff *skb)
-{
-	skb_reader_t desc;
-
-	desc.skb = skb;
-	desc.offset = sizeof(struct udphdr);
-	desc.count = skb->len - desc.offset;
-
-	if (skb->ip_summed == CHECKSUM_UNNECESSARY)
-		goto no_checksum;
-
-	desc.csum = csum_partial(skb->data, desc.offset, skb->csum);
-	if (xdr_partial_copy_from_skb(xdr, 0, &desc, skb_read_and_csum_bits) < 0)
-		return -1;
-	if (desc.offset != skb->len) {
-		unsigned int csum2;
-		csum2 = skb_checksum(skb, desc.offset, skb->len - desc.offset, 0);
-		desc.csum = csum_block_add(desc.csum, csum2, desc.offset);
-	}
-	if (desc.count)
-		return -1;
-	if ((unsigned short)csum_fold(desc.csum))
-		return -1;
-	return 0;
-no_checksum:
-	if (xdr_partial_copy_from_skb(xdr, 0, &desc, skb_read_bits) < 0)
-		return -1;
-	if (desc.count)
-		return -1;
-	return 0;
-}
-
-/*
- * Input handler for RPC replies. Called from a bottom half and hence
- * atomic.
- */
-static void
-udp_data_ready(struct sock *sk, int len)
-{
-	struct rpc_task	*task;
-	struct rpc_xprt	*xprt;
-	struct rpc_rqst *rovr;
-	struct sk_buff	*skb;
-	int err, repsize, copied;
-	u32 _xid, *xp;
-
-	read_lock(&sk->sk_callback_lock);
-	dprintk("RPC:      udp_data_ready...\n");
-	if (!(xprt = xprt_from_sock(sk))) {
-		printk("RPC:      udp_data_ready request not found!\n");
-		goto out;
-	}
-
-	dprintk("RPC:      udp_data_ready client %p\n", xprt);
-
-	if ((skb = skb_recv_datagram(sk, 0, 1, &err)) == NULL)
-		goto out;
-
-	if (xprt->shutdown)
-		goto dropit;
-
-	repsize = skb->len - sizeof(struct udphdr);
-	if (repsize < 4) {
-		printk("RPC: impossible RPC reply size %d!\n", repsize);
-		goto dropit;
-	}
-
-	/* Copy the XID from the skb... */
-	xp = skb_header_pointer(skb, sizeof(struct udphdr),
-				sizeof(_xid), &_xid);
-	if (xp == NULL)
-		goto dropit;
-
-	/* Look up and lock the request corresponding to the given XID */
-	spin_lock(&xprt->sock_lock);
-	rovr = xprt_lookup_rqst(xprt, *xp);
-	if (!rovr)
-		goto out_unlock;
-	task = rovr->rq_task;
-
-	dprintk("RPC: %4d received reply\n", task->tk_pid);
-
-	if ((copied = rovr->rq_private_buf.buflen) > repsize)
-		copied = repsize;
-
-	/* Suck it into the iovec, verify checksum if not done by hw. */
-	if (csum_partial_copy_to_xdr(&rovr->rq_private_buf, skb))
-		goto out_unlock;
-
-	/* Something worked... */
-	dst_confirm(skb->dst);
-
-	xprt_complete_rqst(xprt, rovr, copied);
-
- out_unlock:
-	spin_unlock(&xprt->sock_lock);
- dropit:
-	skb_free_datagram(sk, skb);
- out:
-	read_unlock(&sk->sk_callback_lock);
-}
-
-/*
- * Copy from an skb into memory and shrink the skb.
- */
-static inline size_t
-tcp_copy_data(skb_reader_t *desc, void *p, size_t len)
-{
-	if (len > desc->count)
-		len = desc->count;
-	if (skb_copy_bits(desc->skb, desc->offset, p, len)) {
-		dprintk("RPC:      failed to copy %zu bytes from skb. %zu bytes remain\n",
-				len, desc->count);
-		return 0;
-	}
-	desc->offset += len;
-	desc->count -= len;
-	dprintk("RPC:      copied %zu bytes from skb. %zu bytes remain\n",
-			len, desc->count);
-	return len;
-}
-
-/*
- * TCP read fragment marker
- */
-static inline void
-tcp_read_fraghdr(struct rpc_xprt *xprt, skb_reader_t *desc)
-{
-	size_t len, used;
-	char *p;
-
-	p = ((char *) &xprt->tcp_recm) + xprt->tcp_offset;
-	len = sizeof(xprt->tcp_recm) - xprt->tcp_offset;
-	used = tcp_copy_data(desc, p, len);
-	xprt->tcp_offset += used;
-	if (used != len)
-		return;
-	xprt->tcp_reclen = ntohl(xprt->tcp_recm);
-	if (xprt->tcp_reclen & 0x80000000)
-		xprt->tcp_flags |= XPRT_LAST_FRAG;
-	else
-		xprt->tcp_flags &= ~XPRT_LAST_FRAG;
-	xprt->tcp_reclen &= 0x7fffffff;
-	xprt->tcp_flags &= ~XPRT_COPY_RECM;
-	xprt->tcp_offset = 0;
-	/* Sanity check of the record length */
-	if (xprt->tcp_reclen < 4) {
-		printk(KERN_ERR "RPC: Invalid TCP record fragment length\n");
-		xprt_disconnect(xprt);
-	}
-	dprintk("RPC:      reading TCP record fragment of length %d\n",
-			xprt->tcp_reclen);
-}
-
-static void
-tcp_check_recm(struct rpc_xprt *xprt)
-{
-	dprintk("RPC:      xprt = %p, tcp_copied = %lu, tcp_offset = %u, tcp_reclen = %u, tcp_flags = %lx\n",
-			xprt, xprt->tcp_copied, xprt->tcp_offset, xprt->tcp_reclen, xprt->tcp_flags);
-	if (xprt->tcp_offset == xprt->tcp_reclen) {
-		xprt->tcp_flags |= XPRT_COPY_RECM;
-		xprt->tcp_offset = 0;
-		if (xprt->tcp_flags & XPRT_LAST_FRAG) {
-			xprt->tcp_flags &= ~XPRT_COPY_DATA;
-			xprt->tcp_flags |= XPRT_COPY_XID;
-			xprt->tcp_copied = 0;
-		}
-	}
-}
-
-/*
- * TCP read xid
- */
-static inline void
-tcp_read_xid(struct rpc_xprt *xprt, skb_reader_t *desc)
-{
-	size_t len, used;
-	char *p;
-
-	len = sizeof(xprt->tcp_xid) - xprt->tcp_offset;
-	dprintk("RPC:      reading XID (%Zu bytes)\n", len);
-	p = ((char *) &xprt->tcp_xid) + xprt->tcp_offset;
-	used = tcp_copy_data(desc, p, len);
-	xprt->tcp_offset += used;
-	if (used != len)
-		return;
-	xprt->tcp_flags &= ~XPRT_COPY_XID;
-	xprt->tcp_flags |= XPRT_COPY_DATA;
-	xprt->tcp_copied = 4;
-	dprintk("RPC:      reading reply for XID %08x\n",
-						ntohl(xprt->tcp_xid));
-	tcp_check_recm(xprt);
-}
-
-/*
- * TCP read and complete request
- */
-static inline void
-tcp_read_request(struct rpc_xprt *xprt, skb_reader_t *desc)
-{
-	struct rpc_rqst *req;
-	struct xdr_buf *rcvbuf;
-	size_t len;
-	ssize_t r;
-
-	/* Find and lock the request corresponding to this xid */
-	spin_lock(&xprt->sock_lock);
-	req = xprt_lookup_rqst(xprt, xprt->tcp_xid);
-	if (!req) {
-		xprt->tcp_flags &= ~XPRT_COPY_DATA;
-		dprintk("RPC:      XID %08x request not found!\n",
-				ntohl(xprt->tcp_xid));
-		spin_unlock(&xprt->sock_lock);
-		return;
-	}
-
-	rcvbuf = &req->rq_private_buf;
-	len = desc->count;
-	if (len > xprt->tcp_reclen - xprt->tcp_offset) {
-		skb_reader_t my_desc;
-
-		len = xprt->tcp_reclen - xprt->tcp_offset;
-		memcpy(&my_desc, desc, sizeof(my_desc));
-		my_desc.count = len;
-		r = xdr_partial_copy_from_skb(rcvbuf, xprt->tcp_copied,
-					  &my_desc, tcp_copy_data);
-		desc->count -= r;
-		desc->offset += r;
-	} else
-		r = xdr_partial_copy_from_skb(rcvbuf, xprt->tcp_copied,
-					  desc, tcp_copy_data);
-
-	if (r > 0) {
-		xprt->tcp_copied += r;
-		xprt->tcp_offset += r;
-	}
-	if (r != len) {
-		/* Error when copying to the receive buffer,
-		 * usually because we weren't able to allocate
-		 * additional buffer pages. All we can do now
-		 * is turn off XPRT_COPY_DATA, so the request
-		 * will not receive any additional updates,
-		 * and time out.
-		 * Any remaining data from this record will
-		 * be discarded.
-		 */
-		xprt->tcp_flags &= ~XPRT_COPY_DATA;
-		dprintk("RPC:      XID %08x truncated request\n",
-				ntohl(xprt->tcp_xid));
-		dprintk("RPC:      xprt = %p, tcp_copied = %lu, tcp_offset = %u, tcp_reclen = %u\n",
-				xprt, xprt->tcp_copied, xprt->tcp_offset, xprt->tcp_reclen);
-		goto out;
-	}
-
-	dprintk("RPC:      XID %08x read %Zd bytes\n",
-			ntohl(xprt->tcp_xid), r);
-	dprintk("RPC:      xprt = %p, tcp_copied = %lu, tcp_offset = %u, tcp_reclen = %u\n",
-			xprt, xprt->tcp_copied, xprt->tcp_offset, xprt->tcp_reclen);
-
-	if (xprt->tcp_copied == req->rq_private_buf.buflen)
-		xprt->tcp_flags &= ~XPRT_COPY_DATA;
-	else if (xprt->tcp_offset == xprt->tcp_reclen) {
-		if (xprt->tcp_flags & XPRT_LAST_FRAG)
-			xprt->tcp_flags &= ~XPRT_COPY_DATA;
-	}
-
-out:
-	if (!(xprt->tcp_flags & XPRT_COPY_DATA)) {
-		dprintk("RPC: %4d received reply complete\n",
-				req->rq_task->tk_pid);
-		xprt_complete_rqst(xprt, req, xprt->tcp_copied);
-	}
-	spin_unlock(&xprt->sock_lock);
-	tcp_check_recm(xprt);
-}
-
-/*
- * TCP discard extra bytes from a short read
- */
-static inline void
-tcp_read_discard(struct rpc_xprt *xprt, skb_reader_t *desc)
-{
-	size_t len;
-
-	len = xprt->tcp_reclen - xprt->tcp_offset;
-	if (len > desc->count)
-		len = desc->count;
-	desc->count -= len;
-	desc->offset += len;
-	xprt->tcp_offset += len;
-	dprintk("RPC:      discarded %Zu bytes\n", len);
-	tcp_check_recm(xprt);
-}
-
-/*
- * TCP record receive routine
- * We first have to grab the record marker, then the XID, then the data.
+/**
+ * xprt_update_rtt - update an RPC client's RTT state after receiving a reply
+ * @task: RPC request that recently completed
+ *
  */
-static int
-tcp_data_recv(read_descriptor_t *rd_desc, struct sk_buff *skb,
-		unsigned int offset, size_t len)
-{
-	struct rpc_xprt *xprt = rd_desc->arg.data;
-	skb_reader_t desc = {
-		.skb	= skb,
-		.offset	= offset,
-		.count	= len,
-		.csum	= 0
-       	};
-
-	dprintk("RPC:      tcp_data_recv\n");
-	do {
-		/* Read in a new fragment marker if necessary */
-		/* Can we ever really expect to get completely empty fragments? */
-		if (xprt->tcp_flags & XPRT_COPY_RECM) {
-			tcp_read_fraghdr(xprt, &desc);
-			continue;
-		}
-		/* Read in the xid if necessary */
-		if (xprt->tcp_flags & XPRT_COPY_XID) {
-			tcp_read_xid(xprt, &desc);
-			continue;
-		}
-		/* Read in the request data */
-		if (xprt->tcp_flags & XPRT_COPY_DATA) {
-			tcp_read_request(xprt, &desc);
-			continue;
-		}
-		/* Skip over any trailing bytes on short reads */
-		tcp_read_discard(xprt, &desc);
-	} while (desc.count);
-	dprintk("RPC:      tcp_data_recv done\n");
-	return len - desc.count;
-}
-
-static void tcp_data_ready(struct sock *sk, int bytes)
+void xprt_update_rtt(struct rpc_task *task)
 {
-	struct rpc_xprt *xprt;
-	read_descriptor_t rd_desc;
-
-	read_lock(&sk->sk_callback_lock);
-	dprintk("RPC:      tcp_data_ready...\n");
-	if (!(xprt = xprt_from_sock(sk))) {
-		printk("RPC:      tcp_data_ready socket info not found!\n");
-		goto out;
-	}
-	if (xprt->shutdown)
-		goto out;
-
-	/* We use rd_desc to pass struct xprt to tcp_data_recv */
-	rd_desc.arg.data = xprt;
-	rd_desc.count = 65536;
-	tcp_read_sock(sk, &rd_desc, tcp_data_recv);
-out:
-	read_unlock(&sk->sk_callback_lock);
-}
-
-static void
-tcp_state_change(struct sock *sk)
-{
-	struct rpc_xprt	*xprt;
+	struct rpc_rqst *req = task->tk_rqstp;
+	struct rpc_rtt *rtt = task->tk_client->cl_rtt;
+	unsigned timer = task->tk_msg.rpc_proc->p_timer;
 
-	read_lock(&sk->sk_callback_lock);
-	if (!(xprt = xprt_from_sock(sk)))
-		goto out;
-	dprintk("RPC:      tcp_state_change client %p...\n", xprt);
-	dprintk("RPC:      state %x conn %d dead %d zapped %d\n",
-				sk->sk_state, xprt_connected(xprt),
-				sock_flag(sk, SOCK_DEAD),
-				sock_flag(sk, SOCK_ZAPPED));
-
-	switch (sk->sk_state) {
-	case TCP_ESTABLISHED:
-		spin_lock_bh(&xprt->sock_lock);
-		if (!xprt_test_and_set_connected(xprt)) {
-			/* Reset TCP record info */
-			xprt->tcp_offset = 0;
-			xprt->tcp_reclen = 0;
-			xprt->tcp_copied = 0;
-			xprt->tcp_flags = XPRT_COPY_RECM | XPRT_COPY_XID;
-			rpc_wake_up(&xprt->pending);
-		}
-		spin_unlock_bh(&xprt->sock_lock);
-		break;
-	case TCP_SYN_SENT:
-	case TCP_SYN_RECV:
-		break;
-	default:
-		xprt_disconnect(xprt);
-		break;
+	if (timer) {
+		if (req->rq_ntrans == 1)
+			rpc_update_rtt(rtt, timer,
+					(long)jiffies - req->rq_xtime);
+		rpc_set_timeo(rtt, timer, req->rq_ntrans - 1);
 	}
- out:
-	read_unlock(&sk->sk_callback_lock);
 }
 
-/*
- * Called when more output buffer space is available for this socket.
- * We try not to wake our writers until they can make "significant"
- * progress, otherwise we'll waste resources thrashing sock_sendmsg
- * with a bunch of small requests.
+/**
+ * xprt_complete_rqst - called when reply processing is complete
+ * @task: RPC request that recently completed
+ * @copied: actual number of bytes received from the transport
+ *
+ * Caller holds transport lock.
  */
-static void
-xprt_write_space(struct sock *sk)
+void xprt_complete_rqst(struct rpc_task *task, int copied)
 {
-	struct rpc_xprt	*xprt;
-	struct socket	*sock;
-
-	read_lock(&sk->sk_callback_lock);
-	if (!(xprt = xprt_from_sock(sk)) || !(sock = sk->sk_socket))
-		goto out;
-	if (xprt->shutdown)
-		goto out;
-
-	/* Wait until we have enough socket memory */
-	if (xprt->stream) {
-		/* from net/core/stream.c:sk_stream_write_space */
-		if (sk_stream_wspace(sk) < sk_stream_min_wspace(sk))
-			goto out;
-	} else {
-		/* from net/core/sock.c:sock_def_write_space */
-		if (!sock_writeable(sk))
-			goto out;
-	}
+	struct rpc_rqst *req = task->tk_rqstp;
 
-	if (!test_and_clear_bit(SOCK_NOSPACE, &sock->flags))
-		goto out;
+	dprintk("RPC: %5u xid %08x complete (%d bytes received)\n",
+			task->tk_pid, ntohl(req->rq_xid), copied);
 
-	spin_lock_bh(&xprt->sock_lock);
-	if (xprt->snd_task)
-		rpc_wake_up_task(xprt->snd_task);
-	spin_unlock_bh(&xprt->sock_lock);
-out:
-	read_unlock(&sk->sk_callback_lock);
+	list_del_init(&req->rq_list);
+	req->rq_received = req->rq_private_buf.len = copied;
+	rpc_wake_up_task(task);
 }
 
-/*
- * RPC receive timeout handler.
- */
-static void
-xprt_timer(struct rpc_task *task)
+static void xprt_timer(struct rpc_task *task)
 {
-	struct rpc_rqst	*req = task->tk_rqstp;
+	struct rpc_rqst *req = task->tk_rqstp;
 	struct rpc_xprt *xprt = req->rq_xprt;
 
-	spin_lock(&xprt->sock_lock);
-	if (req->rq_received)
-		goto out;
-
-	xprt_adjust_cwnd(req->rq_xprt, -ETIMEDOUT);
-	__xprt_put_cong(xprt, req);
+	dprintk("RPC: %4d xprt_timer\n", task->tk_pid);
 
-	dprintk("RPC: %4d xprt_timer (%s request)\n",
-		task->tk_pid, req ? "pending" : "backlogged");
-
-	task->tk_status  = -ETIMEDOUT;
-out:
+	spin_lock(&xprt->transport_lock);
+	if (!req->rq_received) {
+		if (xprt->ops->timer)
+			xprt->ops->timer(task);
+		task->tk_status = -ETIMEDOUT;
+	}
 	task->tk_timeout = 0;
 	rpc_wake_up_task(task);
-	spin_unlock(&xprt->sock_lock);
+	spin_unlock(&xprt->transport_lock);
 }
 
-/*
- * Place the actual RPC call.
- * We have to copy the iovec because sendmsg fiddles with its contents.
+/**
+ * xprt_prepare_transmit - reserve the transport before sending a request
+ * @task: RPC task about to send a request
+ *
  */
-int
-xprt_prepare_transmit(struct rpc_task *task)
+int xprt_prepare_transmit(struct rpc_task *task)
 {
 	struct rpc_rqst	*req = task->tk_rqstp;
 	struct rpc_xprt	*xprt = req->rq_xprt;
@@ -1191,12 +690,12 @@ xprt_prepare_transmit(struct rpc_task *task)
 	if (xprt->shutdown)
 		return -EIO;
 
-	spin_lock_bh(&xprt->sock_lock);
+	spin_lock_bh(&xprt->transport_lock);
 	if (req->rq_received && !req->rq_bytes_sent) {
 		err = req->rq_received;
 		goto out_unlock;
 	}
-	if (!__xprt_lock_write(xprt, task)) {
+	if (!xprt->ops->reserve_xprt(task)) {
 		err = -EAGAIN;
 		goto out_unlock;
 	}
@@ -1206,39 +705,42 @@ xprt_prepare_transmit(struct rpc_task *task)
 		goto out_unlock;
 	}
 out_unlock:
-	spin_unlock_bh(&xprt->sock_lock);
+	spin_unlock_bh(&xprt->transport_lock);
 	return err;
 }
 
 void
-xprt_transmit(struct rpc_task *task)
+xprt_abort_transmit(struct rpc_task *task)
+{
+	struct rpc_xprt	*xprt = task->tk_xprt;
+
+	xprt_release_write(xprt, task);
+}
+
+/**
+ * xprt_transmit - send an RPC request on a transport
+ * @task: controlling RPC task
+ *
+ * We have to copy the iovec because sendmsg fiddles with its contents.
+ */
+void xprt_transmit(struct rpc_task *task)
 {
-	struct rpc_clnt *clnt = task->tk_client;
 	struct rpc_rqst	*req = task->tk_rqstp;
 	struct rpc_xprt	*xprt = req->rq_xprt;
-	int status, retry = 0;
-
+	int status;
 
 	dprintk("RPC: %4d xprt_transmit(%u)\n", task->tk_pid, req->rq_slen);
 
-	/* set up everything as needed. */
-	/* Write the record marker */
-	if (xprt->stream) {
-		u32	*marker = req->rq_svec[0].iov_base;
-
-		*marker = htonl(0x80000000|(req->rq_slen-sizeof(*marker)));
-	}
-
 	smp_rmb();
 	if (!req->rq_received) {
 		if (list_empty(&req->rq_list)) {
-			spin_lock_bh(&xprt->sock_lock);
+			spin_lock_bh(&xprt->transport_lock);
 			/* Update the softirq receive buffer */
 			memcpy(&req->rq_private_buf, &req->rq_rcv_buf,
 					sizeof(req->rq_private_buf));
 			/* Add request to the receive list */
 			list_add_tail(&req->rq_list, &xprt->recv);
-			spin_unlock_bh(&xprt->sock_lock);
+			spin_unlock_bh(&xprt->transport_lock);
 			xprt_reset_majortimeo(req);
 			/* Turn off autodisconnect */
 			del_singleshot_timer_sync(&xprt->timer);
@@ -1246,40 +748,19 @@ xprt_transmit(struct rpc_task *task)
 	} else if (!req->rq_bytes_sent)
 		return;
 
-	/* Continue transmitting the packet/record. We must be careful
-	 * to cope with writespace callbacks arriving _after_ we have
-	 * called xprt_sendmsg().
-	 */
-	while (1) {
-		req->rq_xtime = jiffies;
-		status = xprt_sendmsg(xprt, req);
-
-		if (status < 0)
-			break;
-
-		if (xprt->stream) {
-			req->rq_bytes_sent += status;
-
-			/* If we've sent the entire packet, immediately
-			 * reset the count of bytes sent. */
-			if (req->rq_bytes_sent >= req->rq_slen) {
-				req->rq_bytes_sent = 0;
-				goto out_receive;
-			}
-		} else {
-			if (status >= req->rq_slen)
-				goto out_receive;
-			status = -EAGAIN;
-			break;
-		}
-
-		dprintk("RPC: %4d xmit incomplete (%d left of %d)\n",
-				task->tk_pid, req->rq_slen - req->rq_bytes_sent,
-				req->rq_slen);
-
-		status = -EAGAIN;
-		if (retry++ > 50)
-			break;
+	status = xprt->ops->send_request(task);
+	if (status == 0) {
+		dprintk("RPC: %4d xmit complete\n", task->tk_pid);
+		spin_lock_bh(&xprt->transport_lock);
+		xprt->ops->set_retrans_timeout(task);
+		/* Don't race with disconnect */
+		if (!xprt_connected(xprt))
+			task->tk_status = -ENOTCONN;
+		else if (!req->rq_received)
+			rpc_sleep_on(&xprt->pending, task, NULL, xprt_timer);
+		xprt->ops->release_xprt(xprt, task);
+		spin_unlock_bh(&xprt->transport_lock);
+		return;
 	}
 
 	/* Note: at this point, task->tk_sleeping has not yet been set,
@@ -1289,60 +770,19 @@ xprt_transmit(struct rpc_task *task)
 	task->tk_status = status;
 
 	switch (status) {
-	case -EAGAIN:
-		if (test_bit(SOCK_ASYNC_NOSPACE, &xprt->sock->flags)) {
-			/* Protect against races with xprt_write_space */
-			spin_lock_bh(&xprt->sock_lock);
-			/* Don't race with disconnect */
-			if (!xprt_connected(xprt))
-				task->tk_status = -ENOTCONN;
-			else if (test_bit(SOCK_NOSPACE, &xprt->sock->flags)) {
-				task->tk_timeout = req->rq_timeout;
-				rpc_sleep_on(&xprt->pending, task, NULL, NULL);
-			}
-			spin_unlock_bh(&xprt->sock_lock);
-			return;
-		}
-		/* Keep holding the socket if it is blocked */
-		rpc_delay(task, HZ>>4);
-		return;
 	case -ECONNREFUSED:
-		task->tk_timeout = RPC_REESTABLISH_TIMEOUT;
 		rpc_sleep_on(&xprt->sending, task, NULL, NULL);
+	case -EAGAIN:
 	case -ENOTCONN:
 		return;
 	default:
-		if (xprt->stream)
-			xprt_disconnect(xprt);
+		break;
 	}
 	xprt_release_write(xprt, task);
 	return;
- out_receive:
-	dprintk("RPC: %4d xmit complete\n", task->tk_pid);
-	/* Set the task's receive timeout value */
-	spin_lock_bh(&xprt->sock_lock);
-	if (!xprt->nocong) {
-		int timer = task->tk_msg.rpc_proc->p_timer;
-		task->tk_timeout = rpc_calc_rto(clnt->cl_rtt, timer);
-		task->tk_timeout <<= rpc_ntimeo(clnt->cl_rtt, timer) + req->rq_retries;
-		if (task->tk_timeout > xprt->timeout.to_maxval || task->tk_timeout == 0)
-			task->tk_timeout = xprt->timeout.to_maxval;
-	} else
-		task->tk_timeout = req->rq_timeout;
-	/* Don't race with disconnect */
-	if (!xprt_connected(xprt))
-		task->tk_status = -ENOTCONN;
-	else if (!req->rq_received)
-		rpc_sleep_on(&xprt->pending, task, NULL, xprt_timer);
-	__xprt_release_write(xprt, task);
-	spin_unlock_bh(&xprt->sock_lock);
 }
 
-/*
- * Reserve an RPC call slot.
- */
-static inline void
-do_xprt_reserve(struct rpc_task *task)
+static inline void do_xprt_reserve(struct rpc_task *task)
 {
 	struct rpc_xprt	*xprt = task->tk_xprt;
 
@@ -1362,22 +802,25 @@ do_xprt_reserve(struct rpc_task *task)
 	rpc_sleep_on(&xprt->backlog, task, NULL, NULL);
 }
 
-void
-xprt_reserve(struct rpc_task *task)
+/**
+ * xprt_reserve - allocate an RPC request slot
+ * @task: RPC task requesting a slot allocation
+ *
+ * If no more slots are available, place the task on the transport's
+ * backlog queue.
+ */
+void xprt_reserve(struct rpc_task *task)
 {
 	struct rpc_xprt	*xprt = task->tk_xprt;
 
 	task->tk_status = -EIO;
 	if (!xprt->shutdown) {
-		spin_lock(&xprt->xprt_lock);
+		spin_lock(&xprt->reserve_lock);
 		do_xprt_reserve(task);
-		spin_unlock(&xprt->xprt_lock);
+		spin_unlock(&xprt->reserve_lock);
 	}
 }
 
-/*
- * Allocate a 'unique' XID
- */
 static inline u32 xprt_alloc_xid(struct rpc_xprt *xprt)
 {
 	return xprt->xid++;
@@ -1388,11 +831,7 @@ static inline void xprt_init_xid(struct rpc_xprt *xprt)
 	get_random_bytes(&xprt->xid, sizeof(xprt->xid));
 }
 
-/*
- * Initialize RPC request
- */
-static void
-xprt_request_init(struct rpc_task *task, struct rpc_xprt *xprt)
+static void xprt_request_init(struct rpc_task *task, struct rpc_xprt *xprt)
 {
 	struct rpc_rqst	*req = task->tk_rqstp;
 
@@ -1400,128 +839,104 @@ xprt_request_init(struct rpc_task *task, struct rpc_xprt *xprt)
 	req->rq_task	= task;
 	req->rq_xprt    = xprt;
 	req->rq_xid     = xprt_alloc_xid(xprt);
+	req->rq_release_snd_buf = NULL;
 	dprintk("RPC: %4d reserved req %p xid %08x\n", task->tk_pid,
 			req, ntohl(req->rq_xid));
 }
 
-/*
- * Release an RPC call slot
+/**
+ * xprt_release - release an RPC request slot
+ * @task: task which is finished with the slot
+ *
  */
-void
-xprt_release(struct rpc_task *task)
+void xprt_release(struct rpc_task *task)
 {
 	struct rpc_xprt	*xprt = task->tk_xprt;
 	struct rpc_rqst	*req;
 
 	if (!(req = task->tk_rqstp))
 		return;
-	spin_lock_bh(&xprt->sock_lock);
-	__xprt_release_write(xprt, task);
-	__xprt_put_cong(xprt, req);
+	spin_lock_bh(&xprt->transport_lock);
+	xprt->ops->release_xprt(xprt, task);
+	if (xprt->ops->release_request)
+		xprt->ops->release_request(task);
 	if (!list_empty(&req->rq_list))
 		list_del(&req->rq_list);
 	xprt->last_used = jiffies;
 	if (list_empty(&xprt->recv) && !xprt->shutdown)
-		mod_timer(&xprt->timer, xprt->last_used + XPRT_IDLE_TIMEOUT);
-	spin_unlock_bh(&xprt->sock_lock);
+		mod_timer(&xprt->timer,
+				xprt->last_used + xprt->idle_timeout);
+	spin_unlock_bh(&xprt->transport_lock);
 	task->tk_rqstp = NULL;
+	if (req->rq_release_snd_buf)
+		req->rq_release_snd_buf(req);
 	memset(req, 0, sizeof(*req));	/* mark unused */
 
 	dprintk("RPC: %4d release request %p\n", task->tk_pid, req);
 
-	spin_lock(&xprt->xprt_lock);
+	spin_lock(&xprt->reserve_lock);
 	list_add(&req->rq_list, &xprt->free);
-	xprt_clear_backlog(xprt);
-	spin_unlock(&xprt->xprt_lock);
-}
-
-/*
- * Set default timeout parameters
- */
-static void
-xprt_default_timeout(struct rpc_timeout *to, int proto)
-{
-	if (proto == IPPROTO_UDP)
-		xprt_set_timeout(to, 5,  5 * HZ);
-	else
-		xprt_set_timeout(to, 5, 60 * HZ);
+	rpc_wake_up_next(&xprt->backlog);
+	spin_unlock(&xprt->reserve_lock);
 }
 
-/*
- * Set constant timeout
+/**
+ * xprt_set_timeout - set constant RPC timeout
+ * @to: RPC timeout parameters to set up
+ * @retr: number of retries
+ * @incr: amount of increase after each retry
+ *
  */
-void
-xprt_set_timeout(struct rpc_timeout *to, unsigned int retr, unsigned long incr)
+void xprt_set_timeout(struct rpc_timeout *to, unsigned int retr, unsigned long incr)
 {
 	to->to_initval   = 
 	to->to_increment = incr;
-	to->to_maxval    = incr * retr;
+	to->to_maxval    = to->to_initval + (incr * retr);
 	to->to_retries   = retr;
 	to->to_exponential = 0;
 }
 
-unsigned int xprt_udp_slot_table_entries = RPC_DEF_SLOT_TABLE;
-unsigned int xprt_tcp_slot_table_entries = RPC_DEF_SLOT_TABLE;
-
-/*
- * Initialize an RPC client
- */
-static struct rpc_xprt *
-xprt_setup(int proto, struct sockaddr_in *ap, struct rpc_timeout *to)
+static struct rpc_xprt *xprt_setup(int proto, struct sockaddr_in *ap, struct rpc_timeout *to)
 {
+	int result;
 	struct rpc_xprt	*xprt;
-	unsigned int entries;
-	size_t slot_table_size;
 	struct rpc_rqst	*req;
 
-	dprintk("RPC:      setting up %s transport...\n",
-				proto == IPPROTO_UDP? "UDP" : "TCP");
-
-	entries = (proto == IPPROTO_TCP)?
-		xprt_tcp_slot_table_entries : xprt_udp_slot_table_entries;
-
 	if ((xprt = kmalloc(sizeof(struct rpc_xprt), GFP_KERNEL)) == NULL)
 		return ERR_PTR(-ENOMEM);
 	memset(xprt, 0, sizeof(*xprt)); /* Nnnngh! */
-	xprt->max_reqs = entries;
-	slot_table_size = entries * sizeof(xprt->slot[0]);
-	xprt->slot = kmalloc(slot_table_size, GFP_KERNEL);
-	if (xprt->slot == NULL) {
-		kfree(xprt);
-		return ERR_PTR(-ENOMEM);
-	}
-	memset(xprt->slot, 0, slot_table_size);
 
 	xprt->addr = *ap;
-	xprt->prot = proto;
-	xprt->stream = (proto == IPPROTO_TCP)? 1 : 0;
-	if (xprt->stream) {
-		xprt->cwnd = RPC_MAXCWND(xprt);
-		xprt->nocong = 1;
-		xprt->max_payload = (1U << 31) - 1;
-	} else {
-		xprt->cwnd = RPC_INITCWND;
-		xprt->max_payload = (1U << 16) - (MAX_HEADER << 3);
+
+	switch (proto) {
+	case IPPROTO_UDP:
+		result = xs_setup_udp(xprt, to);
+		break;
+	case IPPROTO_TCP:
+		result = xs_setup_tcp(xprt, to);
+		break;
+	default:
+		printk(KERN_ERR "RPC: unrecognized transport protocol: %d\n",
+				proto);
+		result = -EIO;
+		break;
+	}
+	if (result) {
+		kfree(xprt);
+		return ERR_PTR(result);
 	}
-	spin_lock_init(&xprt->sock_lock);
-	spin_lock_init(&xprt->xprt_lock);
-	init_waitqueue_head(&xprt->cong_wait);
+
+	spin_lock_init(&xprt->transport_lock);
+	spin_lock_init(&xprt->reserve_lock);
 
 	INIT_LIST_HEAD(&xprt->free);
 	INIT_LIST_HEAD(&xprt->recv);
-	INIT_WORK(&xprt->sock_connect, xprt_socket_connect, xprt);
-	INIT_WORK(&xprt->task_cleanup, xprt_socket_autoclose, xprt);
+	INIT_WORK(&xprt->task_cleanup, xprt_autoclose, xprt);
 	init_timer(&xprt->timer);
 	xprt->timer.function = xprt_init_autodisconnect;
 	xprt->timer.data = (unsigned long) xprt;
 	xprt->last_used = jiffies;
-	xprt->port = XPRT_MAX_RESVPORT;
-
-	/* Set timeout parameters */
-	if (to) {
-		xprt->timeout = *to;
-	} else
-		xprt_default_timeout(&xprt->timeout, xprt->prot);
+	xprt->cwnd = RPC_INITCWND;
 
 	rpc_init_wait_queue(&xprt->pending, "xprt_pending");
 	rpc_init_wait_queue(&xprt->sending, "xprt_sending");
@@ -1529,139 +944,25 @@ xprt_setup(int proto, struct sockaddr_in *ap, struct rpc_timeout *to)
 	rpc_init_priority_wait_queue(&xprt->backlog, "xprt_backlog");
 
 	/* initialize free list */
-	for (req = &xprt->slot[entries-1]; req >= &xprt->slot[0]; req--)
+	for (req = &xprt->slot[xprt->max_reqs-1]; req >= &xprt->slot[0]; req--)
 		list_add(&req->rq_list, &xprt->free);
 
 	xprt_init_xid(xprt);
 
-	/* Check whether we want to use a reserved port */
-	xprt->resvport = capable(CAP_NET_BIND_SERVICE) ? 1 : 0;
-
 	dprintk("RPC:      created transport %p with %u slots\n", xprt,
 			xprt->max_reqs);
 	
 	return xprt;
 }
 
-/*
- * Bind to a reserved port
- */
-static inline int xprt_bindresvport(struct rpc_xprt *xprt, struct socket *sock)
-{
-	struct sockaddr_in myaddr = {
-		.sin_family = AF_INET,
-	};
-	int		err, port;
-
-	/* Were we already bound to a given port? Try to reuse it */
-	port = xprt->port;
-	do {
-		myaddr.sin_port = htons(port);
-		err = sock->ops->bind(sock, (struct sockaddr *) &myaddr,
-						sizeof(myaddr));
-		if (err == 0) {
-			xprt->port = port;
-			return 0;
-		}
-		if (--port == 0)
-			port = XPRT_MAX_RESVPORT;
-	} while (err == -EADDRINUSE && port != xprt->port);
-
-	printk("RPC: Can't bind to reserved port (%d).\n", -err);
-	return err;
-}
-
-static void
-xprt_bind_socket(struct rpc_xprt *xprt, struct socket *sock)
-{
-	struct sock	*sk = sock->sk;
-
-	if (xprt->inet)
-		return;
-
-	write_lock_bh(&sk->sk_callback_lock);
-	sk->sk_user_data = xprt;
-	xprt->old_data_ready = sk->sk_data_ready;
-	xprt->old_state_change = sk->sk_state_change;
-	xprt->old_write_space = sk->sk_write_space;
-	if (xprt->prot == IPPROTO_UDP) {
-		sk->sk_data_ready = udp_data_ready;
-		sk->sk_no_check = UDP_CSUM_NORCV;
-		xprt_set_connected(xprt);
-	} else {
-		tcp_sk(sk)->nonagle = 1;	/* disable Nagle's algorithm */
-		sk->sk_data_ready = tcp_data_ready;
-		sk->sk_state_change = tcp_state_change;
-		xprt_clear_connected(xprt);
-	}
-	sk->sk_write_space = xprt_write_space;
-
-	/* Reset to new socket */
-	xprt->sock = sock;
-	xprt->inet = sk;
-	write_unlock_bh(&sk->sk_callback_lock);
-
-	return;
-}
-
-/*
- * Set socket buffer length
- */
-void
-xprt_sock_setbufsize(struct rpc_xprt *xprt)
-{
-	struct sock *sk = xprt->inet;
-
-	if (xprt->stream)
-		return;
-	if (xprt->rcvsize) {
-		sk->sk_userlocks |= SOCK_RCVBUF_LOCK;
-		sk->sk_rcvbuf = xprt->rcvsize * xprt->max_reqs *  2;
-	}
-	if (xprt->sndsize) {
-		sk->sk_userlocks |= SOCK_SNDBUF_LOCK;
-		sk->sk_sndbuf = xprt->sndsize * xprt->max_reqs * 2;
-		sk->sk_write_space(sk);
-	}
-}
-
-/*
- * Datastream sockets are created here, but xprt_connect will create
- * and connect stream sockets.
- */
-static struct socket * xprt_create_socket(struct rpc_xprt *xprt, int proto, int resvport)
-{
-	struct socket	*sock;
-	int		type, err;
-
-	dprintk("RPC:      xprt_create_socket(%s %d)\n",
-			   (proto == IPPROTO_UDP)? "udp" : "tcp", proto);
-
-	type = (proto == IPPROTO_UDP)? SOCK_DGRAM : SOCK_STREAM;
-
-	if ((err = sock_create_kern(PF_INET, type, proto, &sock)) < 0) {
-		printk("RPC: can't create socket (%d).\n", -err);
-		return NULL;
-	}
-
-	/* If the caller has the capability, bind to a reserved port */
-	if (resvport && xprt_bindresvport(xprt, sock) < 0) {
-		printk("RPC: can't bind to reserved port.\n");
-		goto failed;
-	}
-
-	return sock;
-
-failed:
-	sock_release(sock);
-	return NULL;
-}
-
-/*
- * Create an RPC client transport given the protocol and peer address.
+/**
+ * xprt_create_proto - create an RPC client transport
+ * @proto: requested transport protocol
+ * @sap: remote peer's address
+ * @to: timeout parameters for new transport
+ *
  */
-struct rpc_xprt *
-xprt_create_proto(int proto, struct sockaddr_in *sap, struct rpc_timeout *to)
+struct rpc_xprt *xprt_create_proto(int proto, struct sockaddr_in *sap, struct rpc_timeout *to)
 {
 	struct rpc_xprt	*xprt;
 
@@ -1673,46 +974,26 @@ xprt_create_proto(int proto, struct sockaddr_in *sap, struct rpc_timeout *to)
 	return xprt;
 }
 
-/*
- * Prepare for transport shutdown.
- */
-static void
-xprt_shutdown(struct rpc_xprt *xprt)
+static void xprt_shutdown(struct rpc_xprt *xprt)
 {
 	xprt->shutdown = 1;
 	rpc_wake_up(&xprt->sending);
 	rpc_wake_up(&xprt->resend);
-	rpc_wake_up(&xprt->pending);
+	xprt_wake_pending_tasks(xprt, -EIO);
 	rpc_wake_up(&xprt->backlog);
-	wake_up(&xprt->cong_wait);
 	del_timer_sync(&xprt->timer);
-
-	/* synchronously wait for connect worker to finish */
-	cancel_delayed_work(&xprt->sock_connect);
-	flush_scheduled_work();
 }
 
-/*
- * Clear the xprt backlog queue
- */
-static int
-xprt_clear_backlog(struct rpc_xprt *xprt) {
-	rpc_wake_up_next(&xprt->backlog);
-	wake_up(&xprt->cong_wait);
-	return 1;
-}
-
-/*
- * Destroy an RPC transport, killing off all requests.
+/**
+ * xprt_destroy - destroy an RPC transport, killing off all requests.
+ * @xprt: transport to destroy
+ *
  */
-int
-xprt_destroy(struct rpc_xprt *xprt)
+int xprt_destroy(struct rpc_xprt *xprt)
 {
 	dprintk("RPC:      destroying transport %p\n", xprt);
 	xprt_shutdown(xprt);
-	xprt_disconnect(xprt);
-	xprt_close(xprt);
-	kfree(xprt->slot);
+	xprt->ops->destroy(xprt);
 	kfree(xprt);
 
 	return 0;
diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c
new file mode 100644
index 00000000000..2e1529217e6
--- /dev/null
+++ b/net/sunrpc/xprtsock.c
@@ -0,0 +1,1252 @@
+/*
+ * linux/net/sunrpc/xprtsock.c
+ *
+ * Client-side transport implementation for sockets.
+ *
+ * TCP callback races fixes (C) 1998 Red Hat Software <alan@redhat.com>
+ * TCP send fixes (C) 1998 Red Hat Software <alan@redhat.com>
+ * TCP NFS related read + write fixes
+ *  (C) 1999 Dave Airlie, University of Limerick, Ireland <airlied@linux.ie>
+ *
+ * Rewrite of larges part of the code in order to stabilize TCP stuff.
+ * Fix behaviour when socket buffer is full.
+ *  (C) 1999 Trond Myklebust <trond.myklebust@fys.uio.no>
+ *
+ * IP socket transport implementation, (C) 2005 Chuck Lever <cel@netapp.com>
+ */
+
+#include <linux/types.h>
+#include <linux/slab.h>
+#include <linux/capability.h>
+#include <linux/sched.h>
+#include <linux/pagemap.h>
+#include <linux/errno.h>
+#include <linux/socket.h>
+#include <linux/in.h>
+#include <linux/net.h>
+#include <linux/mm.h>
+#include <linux/udp.h>
+#include <linux/tcp.h>
+#include <linux/sunrpc/clnt.h>
+#include <linux/file.h>
+
+#include <net/sock.h>
+#include <net/checksum.h>
+#include <net/udp.h>
+#include <net/tcp.h>
+
+/*
+ * How many times to try sending a request on a socket before waiting
+ * for the socket buffer to clear.
+ */
+#define XS_SENDMSG_RETRY	(10U)
+
+/*
+ * Time out for an RPC UDP socket connect.  UDP socket connects are
+ * synchronous, but we set a timeout anyway in case of resource
+ * exhaustion on the local host.
+ */
+#define XS_UDP_CONN_TO		(5U * HZ)
+
+/*
+ * Wait duration for an RPC TCP connection to be established.  Solaris
+ * NFS over TCP uses 60 seconds, for example, which is in line with how
+ * long a server takes to reboot.
+ */
+#define XS_TCP_CONN_TO		(60U * HZ)
+
+/*
+ * Wait duration for a reply from the RPC portmapper.
+ */
+#define XS_BIND_TO		(60U * HZ)
+
+/*
+ * Delay if a UDP socket connect error occurs.  This is most likely some
+ * kind of resource problem on the local host.
+ */
+#define XS_UDP_REEST_TO		(2U * HZ)
+
+/*
+ * The reestablish timeout allows clients to delay for a bit before attempting
+ * to reconnect to a server that just dropped our connection.
+ *
+ * We implement an exponential backoff when trying to reestablish a TCP
+ * transport connection with the server.  Some servers like to drop a TCP
+ * connection when they are overworked, so we start with a short timeout and
+ * increase over time if the server is down or not responding.
+ */
+#define XS_TCP_INIT_REEST_TO	(3U * HZ)
+#define XS_TCP_MAX_REEST_TO	(5U * 60 * HZ)
+
+/*
+ * TCP idle timeout; client drops the transport socket if it is idle
+ * for this long.  Note that we also timeout UDP sockets to prevent
+ * holding port numbers when there is no RPC traffic.
+ */
+#define XS_IDLE_DISC_TO		(5U * 60 * HZ)
+
+#ifdef RPC_DEBUG
+# undef  RPC_DEBUG_DATA
+# define RPCDBG_FACILITY	RPCDBG_TRANS
+#endif
+
+#ifdef RPC_DEBUG_DATA
+static void xs_pktdump(char *msg, u32 *packet, unsigned int count)
+{
+	u8 *buf = (u8 *) packet;
+	int j;
+
+	dprintk("RPC:      %s\n", msg);
+	for (j = 0; j < count && j < 128; j += 4) {
+		if (!(j & 31)) {
+			if (j)
+				dprintk("\n");
+			dprintk("0x%04x ", j);
+		}
+		dprintk("%02x%02x%02x%02x ",
+			buf[j], buf[j+1], buf[j+2], buf[j+3]);
+	}
+	dprintk("\n");
+}
+#else
+static inline void xs_pktdump(char *msg, u32 *packet, unsigned int count)
+{
+	/* NOP */
+}
+#endif
+
+#define XS_SENDMSG_FLAGS	(MSG_DONTWAIT | MSG_NOSIGNAL)
+
+static inline int xs_send_head(struct socket *sock, struct sockaddr *addr, int addrlen, struct xdr_buf *xdr, unsigned int base, unsigned int len)
+{
+	struct kvec iov = {
+		.iov_base	= xdr->head[0].iov_base + base,
+		.iov_len	= len - base,
+	};
+	struct msghdr msg = {
+		.msg_name	= addr,
+		.msg_namelen	= addrlen,
+		.msg_flags	= XS_SENDMSG_FLAGS,
+	};
+
+	if (xdr->len > len)
+		msg.msg_flags |= MSG_MORE;
+
+	if (likely(iov.iov_len))
+		return kernel_sendmsg(sock, &msg, &iov, 1, iov.iov_len);
+	return kernel_sendmsg(sock, &msg, NULL, 0, 0);
+}
+
+static int xs_send_tail(struct socket *sock, struct xdr_buf *xdr, unsigned int base, unsigned int len)
+{
+	struct kvec iov = {
+		.iov_base	= xdr->tail[0].iov_base + base,
+		.iov_len	= len - base,
+	};
+	struct msghdr msg = {
+		.msg_flags	= XS_SENDMSG_FLAGS,
+	};
+
+	return kernel_sendmsg(sock, &msg, &iov, 1, iov.iov_len);
+}
+
+/**
+ * xs_sendpages - write pages directly to a socket
+ * @sock: socket to send on
+ * @addr: UDP only -- address of destination
+ * @addrlen: UDP only -- length of destination address
+ * @xdr: buffer containing this request
+ * @base: starting position in the buffer
+ *
+ */
+static inline int xs_sendpages(struct socket *sock, struct sockaddr *addr, int addrlen, struct xdr_buf *xdr, unsigned int base)
+{
+	struct page **ppage = xdr->pages;
+	unsigned int len, pglen = xdr->page_len;
+	int err, ret = 0;
+	ssize_t (*sendpage)(struct socket *, struct page *, int, size_t, int);
+
+	if (unlikely(!sock))
+		return -ENOTCONN;
+
+	clear_bit(SOCK_ASYNC_NOSPACE, &sock->flags);
+
+	len = xdr->head[0].iov_len;
+	if (base < len || (addr != NULL && base == 0)) {
+		err = xs_send_head(sock, addr, addrlen, xdr, base, len);
+		if (ret == 0)
+			ret = err;
+		else if (err > 0)
+			ret += err;
+		if (err != (len - base))
+			goto out;
+		base = 0;
+	} else
+		base -= len;
+
+	if (unlikely(pglen == 0))
+		goto copy_tail;
+	if (unlikely(base >= pglen)) {
+		base -= pglen;
+		goto copy_tail;
+	}
+	if (base || xdr->page_base) {
+		pglen -= base;
+		base += xdr->page_base;
+		ppage += base >> PAGE_CACHE_SHIFT;
+		base &= ~PAGE_CACHE_MASK;
+	}
+
+	sendpage = sock->ops->sendpage ? : sock_no_sendpage;
+	do {
+		int flags = XS_SENDMSG_FLAGS;
+
+		len = PAGE_CACHE_SIZE;
+		if (base)
+			len -= base;
+		if (pglen < len)
+			len = pglen;
+
+		if (pglen != len || xdr->tail[0].iov_len != 0)
+			flags |= MSG_MORE;
+
+		/* Hmm... We might be dealing with highmem pages */
+		if (PageHighMem(*ppage))
+			sendpage = sock_no_sendpage;
+		err = sendpage(sock, *ppage, base, len, flags);
+		if (ret == 0)
+			ret = err;
+		else if (err > 0)
+			ret += err;
+		if (err != len)
+			goto out;
+		base = 0;
+		ppage++;
+	} while ((pglen -= len) != 0);
+copy_tail:
+	len = xdr->tail[0].iov_len;
+	if (base < len) {
+		err = xs_send_tail(sock, xdr, base, len);
+		if (ret == 0)
+			ret = err;
+		else if (err > 0)
+			ret += err;
+	}
+out:
+	return ret;
+}
+
+/**
+ * xs_nospace - place task on wait queue if transmit was incomplete
+ * @task: task to put to sleep
+ *
+ */
+static void xs_nospace(struct rpc_task *task)
+{
+	struct rpc_rqst *req = task->tk_rqstp;
+	struct rpc_xprt *xprt = req->rq_xprt;
+
+	dprintk("RPC: %4d xmit incomplete (%u left of %u)\n",
+			task->tk_pid, req->rq_slen - req->rq_bytes_sent,
+			req->rq_slen);
+
+	if (test_bit(SOCK_ASYNC_NOSPACE, &xprt->sock->flags)) {
+		/* Protect against races with write_space */
+		spin_lock_bh(&xprt->transport_lock);
+
+		/* Don't race with disconnect */
+		if (!xprt_connected(xprt))
+			task->tk_status = -ENOTCONN;
+		else if (test_bit(SOCK_NOSPACE, &xprt->sock->flags))
+			xprt_wait_for_buffer_space(task);
+
+		spin_unlock_bh(&xprt->transport_lock);
+	} else
+		/* Keep holding the socket if it is blocked */
+		rpc_delay(task, HZ>>4);
+}
+
+/**
+ * xs_udp_send_request - write an RPC request to a UDP socket
+ * @task: address of RPC task that manages the state of an RPC request
+ *
+ * Return values:
+ *        0:	The request has been sent
+ *   EAGAIN:	The socket was blocked, please call again later to
+ *		complete the request
+ * ENOTCONN:	Caller needs to invoke connect logic then call again
+ *    other:	Some other error occured, the request was not sent
+ */
+static int xs_udp_send_request(struct rpc_task *task)
+{
+	struct rpc_rqst *req = task->tk_rqstp;
+	struct rpc_xprt *xprt = req->rq_xprt;
+	struct xdr_buf *xdr = &req->rq_snd_buf;
+	int status;
+
+	xs_pktdump("packet data:",
+				req->rq_svec->iov_base,
+				req->rq_svec->iov_len);
+
+	req->rq_xtime = jiffies;
+	status = xs_sendpages(xprt->sock, (struct sockaddr *) &xprt->addr,
+				sizeof(xprt->addr), xdr, req->rq_bytes_sent);
+
+	dprintk("RPC:      xs_udp_send_request(%u) = %d\n",
+			xdr->len - req->rq_bytes_sent, status);
+
+	if (likely(status >= (int) req->rq_slen))
+		return 0;
+
+	/* Still some bytes left; set up for a retry later. */
+	if (status > 0)
+		status = -EAGAIN;
+
+	switch (status) {
+	case -ENETUNREACH:
+	case -EPIPE:
+	case -ECONNREFUSED:
+		/* When the server has died, an ICMP port unreachable message
+		 * prompts ECONNREFUSED. */
+		break;
+	case -EAGAIN:
+		xs_nospace(task);
+		break;
+	default:
+		dprintk("RPC:      sendmsg returned unrecognized error %d\n",
+			-status);
+		break;
+	}
+
+	return status;
+}
+
+static inline void xs_encode_tcp_record_marker(struct xdr_buf *buf)
+{
+	u32 reclen = buf->len - sizeof(rpc_fraghdr);
+	rpc_fraghdr *base = buf->head[0].iov_base;
+	*base = htonl(RPC_LAST_STREAM_FRAGMENT | reclen);
+}
+
+/**
+ * xs_tcp_send_request - write an RPC request to a TCP socket
+ * @task: address of RPC task that manages the state of an RPC request
+ *
+ * Return values:
+ *        0:	The request has been sent
+ *   EAGAIN:	The socket was blocked, please call again later to
+ *		complete the request
+ * ENOTCONN:	Caller needs to invoke connect logic then call again
+ *    other:	Some other error occured, the request was not sent
+ *
+ * XXX: In the case of soft timeouts, should we eventually give up
+ *	if sendmsg is not able to make progress?
+ */
+static int xs_tcp_send_request(struct rpc_task *task)
+{
+	struct rpc_rqst *req = task->tk_rqstp;
+	struct rpc_xprt *xprt = req->rq_xprt;
+	struct xdr_buf *xdr = &req->rq_snd_buf;
+	int status, retry = 0;
+
+	xs_encode_tcp_record_marker(&req->rq_snd_buf);
+
+	xs_pktdump("packet data:",
+				req->rq_svec->iov_base,
+				req->rq_svec->iov_len);
+
+	/* Continue transmitting the packet/record. We must be careful
+	 * to cope with writespace callbacks arriving _after_ we have
+	 * called sendmsg(). */
+	while (1) {
+		req->rq_xtime = jiffies;
+		status = xs_sendpages(xprt->sock, NULL, 0, xdr,
+						req->rq_bytes_sent);
+
+		dprintk("RPC:      xs_tcp_send_request(%u) = %d\n",
+				xdr->len - req->rq_bytes_sent, status);
+
+		if (unlikely(status < 0))
+			break;
+
+		/* If we've sent the entire packet, immediately
+		 * reset the count of bytes sent. */
+		req->rq_bytes_sent += status;
+		if (likely(req->rq_bytes_sent >= req->rq_slen)) {
+			req->rq_bytes_sent = 0;
+			return 0;
+		}
+
+		status = -EAGAIN;
+		if (retry++ > XS_SENDMSG_RETRY)
+			break;
+	}
+
+	switch (status) {
+	case -EAGAIN:
+		xs_nospace(task);
+		break;
+	case -ECONNREFUSED:
+	case -ECONNRESET:
+	case -ENOTCONN:
+	case -EPIPE:
+		status = -ENOTCONN;
+		break;
+	default:
+		dprintk("RPC:      sendmsg returned unrecognized error %d\n",
+			-status);
+		xprt_disconnect(xprt);
+		break;
+	}
+
+	return status;
+}
+
+/**
+ * xs_close - close a socket
+ * @xprt: transport
+ *
+ * This is used when all requests are complete; ie, no DRC state remains
+ * on the server we want to save.
+ */
+static void xs_close(struct rpc_xprt *xprt)
+{
+	struct socket *sock = xprt->sock;
+	struct sock *sk = xprt->inet;
+
+	if (!sk)
+		return;
+
+	dprintk("RPC:      xs_close xprt %p\n", xprt);
+
+	write_lock_bh(&sk->sk_callback_lock);
+	xprt->inet = NULL;
+	xprt->sock = NULL;
+
+	sk->sk_user_data = NULL;
+	sk->sk_data_ready = xprt->old_data_ready;
+	sk->sk_state_change = xprt->old_state_change;
+	sk->sk_write_space = xprt->old_write_space;
+	write_unlock_bh(&sk->sk_callback_lock);
+
+	sk->sk_no_check = 0;
+
+	sock_release(sock);
+}
+
+/**
+ * xs_destroy - prepare to shutdown a transport
+ * @xprt: doomed transport
+ *
+ */
+static void xs_destroy(struct rpc_xprt *xprt)
+{
+	dprintk("RPC:      xs_destroy xprt %p\n", xprt);
+
+	cancel_delayed_work(&xprt->connect_worker);
+	flush_scheduled_work();
+
+	xprt_disconnect(xprt);
+	xs_close(xprt);
+	kfree(xprt->slot);
+}
+
+static inline struct rpc_xprt *xprt_from_sock(struct sock *sk)
+{
+	return (struct rpc_xprt *) sk->sk_user_data;
+}
+
+/**
+ * xs_udp_data_ready - "data ready" callback for UDP sockets
+ * @sk: socket with data to read
+ * @len: how much data to read
+ *
+ */
+static void xs_udp_data_ready(struct sock *sk, int len)
+{
+	struct rpc_task *task;
+	struct rpc_xprt *xprt;
+	struct rpc_rqst *rovr;
+	struct sk_buff *skb;
+	int err, repsize, copied;
+	u32 _xid, *xp;
+
+	read_lock(&sk->sk_callback_lock);
+	dprintk("RPC:      xs_udp_data_ready...\n");
+	if (!(xprt = xprt_from_sock(sk)))
+		goto out;
+
+	if ((skb = skb_recv_datagram(sk, 0, 1, &err)) == NULL)
+		goto out;
+
+	if (xprt->shutdown)
+		goto dropit;
+
+	repsize = skb->len - sizeof(struct udphdr);
+	if (repsize < 4) {
+		dprintk("RPC:      impossible RPC reply size %d!\n", repsize);
+		goto dropit;
+	}
+
+	/* Copy the XID from the skb... */
+	xp = skb_header_pointer(skb, sizeof(struct udphdr),
+				sizeof(_xid), &_xid);
+	if (xp == NULL)
+		goto dropit;
+
+	/* Look up and lock the request corresponding to the given XID */
+	spin_lock(&xprt->transport_lock);
+	rovr = xprt_lookup_rqst(xprt, *xp);
+	if (!rovr)
+		goto out_unlock;
+	task = rovr->rq_task;
+
+	if ((copied = rovr->rq_private_buf.buflen) > repsize)
+		copied = repsize;
+
+	/* Suck it into the iovec, verify checksum if not done by hw. */
+	if (csum_partial_copy_to_xdr(&rovr->rq_private_buf, skb))
+		goto out_unlock;
+
+	/* Something worked... */
+	dst_confirm(skb->dst);
+
+	xprt_adjust_cwnd(task, copied);
+	xprt_update_rtt(task);
+	xprt_complete_rqst(task, copied);
+
+ out_unlock:
+	spin_unlock(&xprt->transport_lock);
+ dropit:
+	skb_free_datagram(sk, skb);
+ out:
+	read_unlock(&sk->sk_callback_lock);
+}
+
+static inline size_t xs_tcp_copy_data(skb_reader_t *desc, void *p, size_t len)
+{
+	if (len > desc->count)
+		len = desc->count;
+	if (skb_copy_bits(desc->skb, desc->offset, p, len)) {
+		dprintk("RPC:      failed to copy %zu bytes from skb. %zu bytes remain\n",
+				len, desc->count);
+		return 0;
+	}
+	desc->offset += len;
+	desc->count -= len;
+	dprintk("RPC:      copied %zu bytes from skb. %zu bytes remain\n",
+			len, desc->count);
+	return len;
+}
+
+static inline void xs_tcp_read_fraghdr(struct rpc_xprt *xprt, skb_reader_t *desc)
+{
+	size_t len, used;
+	char *p;
+
+	p = ((char *) &xprt->tcp_recm) + xprt->tcp_offset;
+	len = sizeof(xprt->tcp_recm) - xprt->tcp_offset;
+	used = xs_tcp_copy_data(desc, p, len);
+	xprt->tcp_offset += used;
+	if (used != len)
+		return;
+
+	xprt->tcp_reclen = ntohl(xprt->tcp_recm);
+	if (xprt->tcp_reclen & RPC_LAST_STREAM_FRAGMENT)
+		xprt->tcp_flags |= XPRT_LAST_FRAG;
+	else
+		xprt->tcp_flags &= ~XPRT_LAST_FRAG;
+	xprt->tcp_reclen &= RPC_FRAGMENT_SIZE_MASK;
+
+	xprt->tcp_flags &= ~XPRT_COPY_RECM;
+	xprt->tcp_offset = 0;
+
+	/* Sanity check of the record length */
+	if (unlikely(xprt->tcp_reclen < 4)) {
+		dprintk("RPC:      invalid TCP record fragment length\n");
+		xprt_disconnect(xprt);
+		return;
+	}
+	dprintk("RPC:      reading TCP record fragment of length %d\n",
+			xprt->tcp_reclen);
+}
+
+static void xs_tcp_check_recm(struct rpc_xprt *xprt)
+{
+	dprintk("RPC:      xprt = %p, tcp_copied = %lu, tcp_offset = %u, tcp_reclen = %u, tcp_flags = %lx\n",
+			xprt, xprt->tcp_copied, xprt->tcp_offset, xprt->tcp_reclen, xprt->tcp_flags);
+	if (xprt->tcp_offset == xprt->tcp_reclen) {
+		xprt->tcp_flags |= XPRT_COPY_RECM;
+		xprt->tcp_offset = 0;
+		if (xprt->tcp_flags & XPRT_LAST_FRAG) {
+			xprt->tcp_flags &= ~XPRT_COPY_DATA;
+			xprt->tcp_flags |= XPRT_COPY_XID;
+			xprt->tcp_copied = 0;
+		}
+	}
+}
+
+static inline void xs_tcp_read_xid(struct rpc_xprt *xprt, skb_reader_t *desc)
+{
+	size_t len, used;
+	char *p;
+
+	len = sizeof(xprt->tcp_xid) - xprt->tcp_offset;
+	dprintk("RPC:      reading XID (%Zu bytes)\n", len);
+	p = ((char *) &xprt->tcp_xid) + xprt->tcp_offset;
+	used = xs_tcp_copy_data(desc, p, len);
+	xprt->tcp_offset += used;
+	if (used != len)
+		return;
+	xprt->tcp_flags &= ~XPRT_COPY_XID;
+	xprt->tcp_flags |= XPRT_COPY_DATA;
+	xprt->tcp_copied = 4;
+	dprintk("RPC:      reading reply for XID %08x\n",
+						ntohl(xprt->tcp_xid));
+	xs_tcp_check_recm(xprt);
+}
+
+static inline void xs_tcp_read_request(struct rpc_xprt *xprt, skb_reader_t *desc)
+{
+	struct rpc_rqst *req;
+	struct xdr_buf *rcvbuf;
+	size_t len;
+	ssize_t r;
+
+	/* Find and lock the request corresponding to this xid */
+	spin_lock(&xprt->transport_lock);
+	req = xprt_lookup_rqst(xprt, xprt->tcp_xid);
+	if (!req) {
+		xprt->tcp_flags &= ~XPRT_COPY_DATA;
+		dprintk("RPC:      XID %08x request not found!\n",
+				ntohl(xprt->tcp_xid));
+		spin_unlock(&xprt->transport_lock);
+		return;
+	}
+
+	rcvbuf = &req->rq_private_buf;
+	len = desc->count;
+	if (len > xprt->tcp_reclen - xprt->tcp_offset) {
+		skb_reader_t my_desc;
+
+		len = xprt->tcp_reclen - xprt->tcp_offset;
+		memcpy(&my_desc, desc, sizeof(my_desc));
+		my_desc.count = len;
+		r = xdr_partial_copy_from_skb(rcvbuf, xprt->tcp_copied,
+					  &my_desc, xs_tcp_copy_data);
+		desc->count -= r;
+		desc->offset += r;
+	} else
+		r = xdr_partial_copy_from_skb(rcvbuf, xprt->tcp_copied,
+					  desc, xs_tcp_copy_data);
+
+	if (r > 0) {
+		xprt->tcp_copied += r;
+		xprt->tcp_offset += r;
+	}
+	if (r != len) {
+		/* Error when copying to the receive buffer,
+		 * usually because we weren't able to allocate
+		 * additional buffer pages. All we can do now
+		 * is turn off XPRT_COPY_DATA, so the request
+		 * will not receive any additional updates,
+		 * and time out.
+		 * Any remaining data from this record will
+		 * be discarded.
+		 */
+		xprt->tcp_flags &= ~XPRT_COPY_DATA;
+		dprintk("RPC:      XID %08x truncated request\n",
+				ntohl(xprt->tcp_xid));
+		dprintk("RPC:      xprt = %p, tcp_copied = %lu, tcp_offset = %u, tcp_reclen = %u\n",
+				xprt, xprt->tcp_copied, xprt->tcp_offset, xprt->tcp_reclen);
+		goto out;
+	}
+
+	dprintk("RPC:      XID %08x read %Zd bytes\n",
+			ntohl(xprt->tcp_xid), r);
+	dprintk("RPC:      xprt = %p, tcp_copied = %lu, tcp_offset = %u, tcp_reclen = %u\n",
+			xprt, xprt->tcp_copied, xprt->tcp_offset, xprt->tcp_reclen);
+
+	if (xprt->tcp_copied == req->rq_private_buf.buflen)
+		xprt->tcp_flags &= ~XPRT_COPY_DATA;
+	else if (xprt->tcp_offset == xprt->tcp_reclen) {
+		if (xprt->tcp_flags & XPRT_LAST_FRAG)
+			xprt->tcp_flags &= ~XPRT_COPY_DATA;
+	}
+
+out:
+	if (!(xprt->tcp_flags & XPRT_COPY_DATA))
+		xprt_complete_rqst(req->rq_task, xprt->tcp_copied);
+	spin_unlock(&xprt->transport_lock);
+	xs_tcp_check_recm(xprt);
+}
+
+static inline void xs_tcp_read_discard(struct rpc_xprt *xprt, skb_reader_t *desc)
+{
+	size_t len;
+
+	len = xprt->tcp_reclen - xprt->tcp_offset;
+	if (len > desc->count)
+		len = desc->count;
+	desc->count -= len;
+	desc->offset += len;
+	xprt->tcp_offset += len;
+	dprintk("RPC:      discarded %Zu bytes\n", len);
+	xs_tcp_check_recm(xprt);
+}
+
+static int xs_tcp_data_recv(read_descriptor_t *rd_desc, struct sk_buff *skb, unsigned int offset, size_t len)
+{
+	struct rpc_xprt *xprt = rd_desc->arg.data;
+	skb_reader_t desc = {
+		.skb	= skb,
+		.offset	= offset,
+		.count	= len,
+		.csum	= 0
+	};
+
+	dprintk("RPC:      xs_tcp_data_recv started\n");
+	do {
+		/* Read in a new fragment marker if necessary */
+		/* Can we ever really expect to get completely empty fragments? */
+		if (xprt->tcp_flags & XPRT_COPY_RECM) {
+			xs_tcp_read_fraghdr(xprt, &desc);
+			continue;
+		}
+		/* Read in the xid if necessary */
+		if (xprt->tcp_flags & XPRT_COPY_XID) {
+			xs_tcp_read_xid(xprt, &desc);
+			continue;
+		}
+		/* Read in the request data */
+		if (xprt->tcp_flags & XPRT_COPY_DATA) {
+			xs_tcp_read_request(xprt, &desc);
+			continue;
+		}
+		/* Skip over any trailing bytes on short reads */
+		xs_tcp_read_discard(xprt, &desc);
+	} while (desc.count);
+	dprintk("RPC:      xs_tcp_data_recv done\n");
+	return len - desc.count;
+}
+
+/**
+ * xs_tcp_data_ready - "data ready" callback for TCP sockets
+ * @sk: socket with data to read
+ * @bytes: how much data to read
+ *
+ */
+static void xs_tcp_data_ready(struct sock *sk, int bytes)
+{
+	struct rpc_xprt *xprt;
+	read_descriptor_t rd_desc;
+
+	read_lock(&sk->sk_callback_lock);
+	dprintk("RPC:      xs_tcp_data_ready...\n");
+	if (!(xprt = xprt_from_sock(sk)))
+		goto out;
+	if (xprt->shutdown)
+		goto out;
+
+	/* We use rd_desc to pass struct xprt to xs_tcp_data_recv */
+	rd_desc.arg.data = xprt;
+	rd_desc.count = 65536;
+	tcp_read_sock(sk, &rd_desc, xs_tcp_data_recv);
+out:
+	read_unlock(&sk->sk_callback_lock);
+}
+
+/**
+ * xs_tcp_state_change - callback to handle TCP socket state changes
+ * @sk: socket whose state has changed
+ *
+ */
+static void xs_tcp_state_change(struct sock *sk)
+{
+	struct rpc_xprt *xprt;
+
+	read_lock(&sk->sk_callback_lock);
+	if (!(xprt = xprt_from_sock(sk)))
+		goto out;
+	dprintk("RPC:      xs_tcp_state_change client %p...\n", xprt);
+	dprintk("RPC:      state %x conn %d dead %d zapped %d\n",
+				sk->sk_state, xprt_connected(xprt),
+				sock_flag(sk, SOCK_DEAD),
+				sock_flag(sk, SOCK_ZAPPED));
+
+	switch (sk->sk_state) {
+	case TCP_ESTABLISHED:
+		spin_lock_bh(&xprt->transport_lock);
+		if (!xprt_test_and_set_connected(xprt)) {
+			/* Reset TCP record info */
+			xprt->tcp_offset = 0;
+			xprt->tcp_reclen = 0;
+			xprt->tcp_copied = 0;
+			xprt->tcp_flags = XPRT_COPY_RECM | XPRT_COPY_XID;
+			xprt->reestablish_timeout = XS_TCP_INIT_REEST_TO;
+			xprt_wake_pending_tasks(xprt, 0);
+		}
+		spin_unlock_bh(&xprt->transport_lock);
+		break;
+	case TCP_SYN_SENT:
+	case TCP_SYN_RECV:
+		break;
+	default:
+		xprt_disconnect(xprt);
+		break;
+	}
+ out:
+	read_unlock(&sk->sk_callback_lock);
+}
+
+/**
+ * xs_udp_write_space - callback invoked when socket buffer space
+ *                             becomes available
+ * @sk: socket whose state has changed
+ *
+ * Called when more output buffer space is available for this socket.
+ * We try not to wake our writers until they can make "significant"
+ * progress, otherwise we'll waste resources thrashing kernel_sendmsg
+ * with a bunch of small requests.
+ */
+static void xs_udp_write_space(struct sock *sk)
+{
+	read_lock(&sk->sk_callback_lock);
+
+	/* from net/core/sock.c:sock_def_write_space */
+	if (sock_writeable(sk)) {
+		struct socket *sock;
+		struct rpc_xprt *xprt;
+
+		if (unlikely(!(sock = sk->sk_socket)))
+			goto out;
+		if (unlikely(!(xprt = xprt_from_sock(sk))))
+			goto out;
+		if (unlikely(!test_and_clear_bit(SOCK_NOSPACE, &sock->flags)))
+			goto out;
+
+		xprt_write_space(xprt);
+	}
+
+ out:
+	read_unlock(&sk->sk_callback_lock);
+}
+
+/**
+ * xs_tcp_write_space - callback invoked when socket buffer space
+ *                             becomes available
+ * @sk: socket whose state has changed
+ *
+ * Called when more output buffer space is available for this socket.
+ * We try not to wake our writers until they can make "significant"
+ * progress, otherwise we'll waste resources thrashing kernel_sendmsg
+ * with a bunch of small requests.
+ */
+static void xs_tcp_write_space(struct sock *sk)
+{
+	read_lock(&sk->sk_callback_lock);
+
+	/* from net/core/stream.c:sk_stream_write_space */
+	if (sk_stream_wspace(sk) >= sk_stream_min_wspace(sk)) {
+		struct socket *sock;
+		struct rpc_xprt *xprt;
+
+		if (unlikely(!(sock = sk->sk_socket)))
+			goto out;
+		if (unlikely(!(xprt = xprt_from_sock(sk))))
+			goto out;
+		if (unlikely(!test_and_clear_bit(SOCK_NOSPACE, &sock->flags)))
+			goto out;
+
+		xprt_write_space(xprt);
+	}
+
+ out:
+	read_unlock(&sk->sk_callback_lock);
+}
+
+static void xs_udp_do_set_buffer_size(struct rpc_xprt *xprt)
+{
+	struct sock *sk = xprt->inet;
+
+	if (xprt->rcvsize) {
+		sk->sk_userlocks |= SOCK_RCVBUF_LOCK;
+		sk->sk_rcvbuf = xprt->rcvsize * xprt->max_reqs *  2;
+	}
+	if (xprt->sndsize) {
+		sk->sk_userlocks |= SOCK_SNDBUF_LOCK;
+		sk->sk_sndbuf = xprt->sndsize * xprt->max_reqs * 2;
+		sk->sk_write_space(sk);
+	}
+}
+
+/**
+ * xs_udp_set_buffer_size - set send and receive limits
+ * @xprt: generic transport
+ * @sndsize: requested size of send buffer, in bytes
+ * @rcvsize: requested size of receive buffer, in bytes
+ *
+ * Set socket send and receive buffer size limits.
+ */
+static void xs_udp_set_buffer_size(struct rpc_xprt *xprt, size_t sndsize, size_t rcvsize)
+{
+	xprt->sndsize = 0;
+	if (sndsize)
+		xprt->sndsize = sndsize + 1024;
+	xprt->rcvsize = 0;
+	if (rcvsize)
+		xprt->rcvsize = rcvsize + 1024;
+
+	xs_udp_do_set_buffer_size(xprt);
+}
+
+/**
+ * xs_udp_timer - called when a retransmit timeout occurs on a UDP transport
+ * @task: task that timed out
+ *
+ * Adjust the congestion window after a retransmit timeout has occurred.
+ */
+static void xs_udp_timer(struct rpc_task *task)
+{
+	xprt_adjust_cwnd(task, -ETIMEDOUT);
+}
+
+static int xs_bindresvport(struct rpc_xprt *xprt, struct socket *sock)
+{
+	struct sockaddr_in myaddr = {
+		.sin_family = AF_INET,
+	};
+	int err;
+	unsigned short port = xprt->port;
+
+	do {
+		myaddr.sin_port = htons(port);
+		err = sock->ops->bind(sock, (struct sockaddr *) &myaddr,
+						sizeof(myaddr));
+		if (err == 0) {
+			xprt->port = port;
+			dprintk("RPC:      xs_bindresvport bound to port %u\n",
+					port);
+			return 0;
+		}
+		if (port <= xprt_min_resvport)
+			port = xprt_max_resvport;
+		else
+			port--;
+	} while (err == -EADDRINUSE && port != xprt->port);
+
+	dprintk("RPC:      can't bind to reserved port (%d).\n", -err);
+	return err;
+}
+
+/**
+ * xs_udp_connect_worker - set up a UDP socket
+ * @args: RPC transport to connect
+ *
+ * Invoked by a work queue tasklet.
+ */
+static void xs_udp_connect_worker(void *args)
+{
+	struct rpc_xprt *xprt = (struct rpc_xprt *) args;
+	struct socket *sock = xprt->sock;
+	int err, status = -EIO;
+
+	if (xprt->shutdown || xprt->addr.sin_port == 0)
+		goto out;
+
+	dprintk("RPC:      xs_udp_connect_worker for xprt %p\n", xprt);
+
+	/* Start by resetting any existing state */
+	xs_close(xprt);
+
+	if ((err = sock_create_kern(PF_INET, SOCK_DGRAM, IPPROTO_UDP, &sock)) < 0) {
+		dprintk("RPC:      can't create UDP transport socket (%d).\n", -err);
+		goto out;
+	}
+
+	if (xprt->resvport && xs_bindresvport(xprt, sock) < 0) {
+		sock_release(sock);
+		goto out;
+	}
+
+	if (!xprt->inet) {
+		struct sock *sk = sock->sk;
+
+		write_lock_bh(&sk->sk_callback_lock);
+
+		sk->sk_user_data = xprt;
+		xprt->old_data_ready = sk->sk_data_ready;
+		xprt->old_state_change = sk->sk_state_change;
+		xprt->old_write_space = sk->sk_write_space;
+		sk->sk_data_ready = xs_udp_data_ready;
+		sk->sk_write_space = xs_udp_write_space;
+		sk->sk_no_check = UDP_CSUM_NORCV;
+
+		xprt_set_connected(xprt);
+
+		/* Reset to new socket */
+		xprt->sock = sock;
+		xprt->inet = sk;
+
+		write_unlock_bh(&sk->sk_callback_lock);
+	}
+	xs_udp_do_set_buffer_size(xprt);
+	status = 0;
+out:
+	xprt_wake_pending_tasks(xprt, status);
+	xprt_clear_connecting(xprt);
+}
+
+/*
+ * We need to preserve the port number so the reply cache on the server can
+ * find our cached RPC replies when we get around to reconnecting.
+ */
+static void xs_tcp_reuse_connection(struct rpc_xprt *xprt)
+{
+	int result;
+	struct socket *sock = xprt->sock;
+	struct sockaddr any;
+
+	dprintk("RPC:      disconnecting xprt %p to reuse port\n", xprt);
+
+	/*
+	 * Disconnect the transport socket by doing a connect operation
+	 * with AF_UNSPEC.  This should return immediately...
+	 */
+	memset(&any, 0, sizeof(any));
+	any.sa_family = AF_UNSPEC;
+	result = sock->ops->connect(sock, &any, sizeof(any), 0);
+	if (result)
+		dprintk("RPC:      AF_UNSPEC connect return code %d\n",
+				result);
+}
+
+/**
+ * xs_tcp_connect_worker - connect a TCP socket to a remote endpoint
+ * @args: RPC transport to connect
+ *
+ * Invoked by a work queue tasklet.
+ */
+static void xs_tcp_connect_worker(void *args)
+{
+	struct rpc_xprt *xprt = (struct rpc_xprt *)args;
+	struct socket *sock = xprt->sock;
+	int err, status = -EIO;
+
+	if (xprt->shutdown || xprt->addr.sin_port == 0)
+		goto out;
+
+	dprintk("RPC:      xs_tcp_connect_worker for xprt %p\n", xprt);
+
+	if (!xprt->sock) {
+		/* start from scratch */
+		if ((err = sock_create_kern(PF_INET, SOCK_STREAM, IPPROTO_TCP, &sock)) < 0) {
+			dprintk("RPC:      can't create TCP transport socket (%d).\n", -err);
+			goto out;
+		}
+
+		if (xprt->resvport && xs_bindresvport(xprt, sock) < 0) {
+			sock_release(sock);
+			goto out;
+		}
+	} else
+		/* "close" the socket, preserving the local port */
+		xs_tcp_reuse_connection(xprt);
+
+	if (!xprt->inet) {
+		struct sock *sk = sock->sk;
+
+		write_lock_bh(&sk->sk_callback_lock);
+
+		sk->sk_user_data = xprt;
+		xprt->old_data_ready = sk->sk_data_ready;
+		xprt->old_state_change = sk->sk_state_change;
+		xprt->old_write_space = sk->sk_write_space;
+		sk->sk_data_ready = xs_tcp_data_ready;
+		sk->sk_state_change = xs_tcp_state_change;
+		sk->sk_write_space = xs_tcp_write_space;
+
+		/* socket options */
+		sk->sk_userlocks |= SOCK_BINDPORT_LOCK;
+		sock_reset_flag(sk, SOCK_LINGER);
+		tcp_sk(sk)->linger2 = 0;
+		tcp_sk(sk)->nonagle |= TCP_NAGLE_OFF;
+
+		xprt_clear_connected(xprt);
+
+		/* Reset to new socket */
+		xprt->sock = sock;
+		xprt->inet = sk;
+
+		write_unlock_bh(&sk->sk_callback_lock);
+	}
+
+	/* Tell the socket layer to start connecting... */
+	status = sock->ops->connect(sock, (struct sockaddr *) &xprt->addr,
+			sizeof(xprt->addr), O_NONBLOCK);
+	dprintk("RPC: %p  connect status %d connected %d sock state %d\n",
+			xprt, -status, xprt_connected(xprt), sock->sk->sk_state);
+	if (status < 0) {
+		switch (status) {
+			case -EINPROGRESS:
+			case -EALREADY:
+				goto out_clear;
+			case -ECONNREFUSED:
+			case -ECONNRESET:
+				/* retry with existing socket, after a delay */
+				break;
+			default:
+				/* get rid of existing socket, and retry */
+				xs_close(xprt);
+				break;
+		}
+	}
+out:
+	xprt_wake_pending_tasks(xprt, status);
+out_clear:
+	xprt_clear_connecting(xprt);
+}
+
+/**
+ * xs_connect - connect a socket to a remote endpoint
+ * @task: address of RPC task that manages state of connect request
+ *
+ * TCP: If the remote end dropped the connection, delay reconnecting.
+ *
+ * UDP socket connects are synchronous, but we use a work queue anyway
+ * to guarantee that even unprivileged user processes can set up a
+ * socket on a privileged port.
+ *
+ * If a UDP socket connect fails, the delay behavior here prevents
+ * retry floods (hard mounts).
+ */
+static void xs_connect(struct rpc_task *task)
+{
+	struct rpc_xprt *xprt = task->tk_xprt;
+
+	if (xprt_test_and_set_connecting(xprt))
+		return;
+
+	if (xprt->sock != NULL) {
+		dprintk("RPC:      xs_connect delayed xprt %p for %lu seconds\n",
+				xprt, xprt->reestablish_timeout / HZ);
+		schedule_delayed_work(&xprt->connect_worker,
+					xprt->reestablish_timeout);
+		xprt->reestablish_timeout <<= 1;
+		if (xprt->reestablish_timeout > XS_TCP_MAX_REEST_TO)
+			xprt->reestablish_timeout = XS_TCP_MAX_REEST_TO;
+	} else {
+		dprintk("RPC:      xs_connect scheduled xprt %p\n", xprt);
+		schedule_work(&xprt->connect_worker);
+
+		/* flush_scheduled_work can sleep... */
+		if (!RPC_IS_ASYNC(task))
+			flush_scheduled_work();
+	}
+}
+
+static struct rpc_xprt_ops xs_udp_ops = {
+	.set_buffer_size	= xs_udp_set_buffer_size,
+	.reserve_xprt		= xprt_reserve_xprt_cong,
+	.release_xprt		= xprt_release_xprt_cong,
+	.connect		= xs_connect,
+	.send_request		= xs_udp_send_request,
+	.set_retrans_timeout	= xprt_set_retrans_timeout_rtt,
+	.timer			= xs_udp_timer,
+	.release_request	= xprt_release_rqst_cong,
+	.close			= xs_close,
+	.destroy		= xs_destroy,
+};
+
+static struct rpc_xprt_ops xs_tcp_ops = {
+	.reserve_xprt		= xprt_reserve_xprt,
+	.release_xprt		= xprt_release_xprt,
+	.connect		= xs_connect,
+	.send_request		= xs_tcp_send_request,
+	.set_retrans_timeout	= xprt_set_retrans_timeout_def,
+	.close			= xs_close,
+	.destroy		= xs_destroy,
+};
+
+/**
+ * xs_setup_udp - Set up transport to use a UDP socket
+ * @xprt: transport to set up
+ * @to:   timeout parameters
+ *
+ */
+int xs_setup_udp(struct rpc_xprt *xprt, struct rpc_timeout *to)
+{
+	size_t slot_table_size;
+
+	dprintk("RPC:      setting up udp-ipv4 transport...\n");
+
+	xprt->max_reqs = xprt_udp_slot_table_entries;
+	slot_table_size = xprt->max_reqs * sizeof(xprt->slot[0]);
+	xprt->slot = kmalloc(slot_table_size, GFP_KERNEL);
+	if (xprt->slot == NULL)
+		return -ENOMEM;
+	memset(xprt->slot, 0, slot_table_size);
+
+	xprt->prot = IPPROTO_UDP;
+	xprt->port = xprt_max_resvport;
+	xprt->tsh_size = 0;
+	xprt->resvport = capable(CAP_NET_BIND_SERVICE) ? 1 : 0;
+	/* XXX: header size can vary due to auth type, IPv6, etc. */
+	xprt->max_payload = (1U << 16) - (MAX_HEADER << 3);
+
+	INIT_WORK(&xprt->connect_worker, xs_udp_connect_worker, xprt);
+	xprt->bind_timeout = XS_BIND_TO;
+	xprt->connect_timeout = XS_UDP_CONN_TO;
+	xprt->reestablish_timeout = XS_UDP_REEST_TO;
+	xprt->idle_timeout = XS_IDLE_DISC_TO;
+
+	xprt->ops = &xs_udp_ops;
+
+	if (to)
+		xprt->timeout = *to;
+	else
+		xprt_set_timeout(&xprt->timeout, 5, 5 * HZ);
+
+	return 0;
+}
+
+/**
+ * xs_setup_tcp - Set up transport to use a TCP socket
+ * @xprt: transport to set up
+ * @to: timeout parameters
+ *
+ */
+int xs_setup_tcp(struct rpc_xprt *xprt, struct rpc_timeout *to)
+{
+	size_t slot_table_size;
+
+	dprintk("RPC:      setting up tcp-ipv4 transport...\n");
+
+	xprt->max_reqs = xprt_tcp_slot_table_entries;
+	slot_table_size = xprt->max_reqs * sizeof(xprt->slot[0]);
+	xprt->slot = kmalloc(slot_table_size, GFP_KERNEL);
+	if (xprt->slot == NULL)
+		return -ENOMEM;
+	memset(xprt->slot, 0, slot_table_size);
+
+	xprt->prot = IPPROTO_TCP;
+	xprt->port = xprt_max_resvport;
+	xprt->tsh_size = sizeof(rpc_fraghdr) / sizeof(u32);
+	xprt->resvport = capable(CAP_NET_BIND_SERVICE) ? 1 : 0;
+	xprt->max_payload = RPC_MAX_FRAGMENT_SIZE;
+
+	INIT_WORK(&xprt->connect_worker, xs_tcp_connect_worker, xprt);
+	xprt->bind_timeout = XS_BIND_TO;
+	xprt->connect_timeout = XS_TCP_CONN_TO;
+	xprt->reestablish_timeout = XS_TCP_INIT_REEST_TO;
+	xprt->idle_timeout = XS_IDLE_DISC_TO;
+
+	xprt->ops = &xs_tcp_ops;
+
+	if (to)
+		xprt->timeout = *to;
+	else
+		xprt_set_timeout(&xprt->timeout, 2, 60 * HZ);
+
+	return 0;
+}
diff --git a/net/sysctl_net.c b/net/sysctl_net.c
index c5241fcbb96..55538f6b60f 100644
--- a/net/sysctl_net.c
+++ b/net/sysctl_net.c
@@ -16,6 +16,8 @@
 #include <linux/mm.h>
 #include <linux/sysctl.h>
 
+#include <net/sock.h>
+
 #ifdef CONFIG_INET
 #include <net/ip.h>
 #endif
diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c
index fda737d77ed..0db9e57013f 100644
--- a/net/xfrm/xfrm_policy.c
+++ b/net/xfrm/xfrm_policy.c
@@ -163,7 +163,7 @@ static void xfrm_policy_timer(unsigned long data)
 	if (xp->dead)
 		goto out;
 
-	dir = xp->index & 7;
+	dir = xfrm_policy_id2dir(xp->index);
 
 	if (xp->lft.hard_add_expires_seconds) {
 		long tmo = xp->lft.hard_add_expires_seconds +
@@ -225,7 +225,7 @@ expired:
  * SPD calls.
  */
 
-struct xfrm_policy *xfrm_policy_alloc(int gfp)
+struct xfrm_policy *xfrm_policy_alloc(gfp_t gfp)
 {
 	struct xfrm_policy *policy;
 
@@ -417,7 +417,7 @@ struct xfrm_policy *xfrm_policy_byid(int dir, u32 id, int delete)
 	struct xfrm_policy *pol, **p;
 
 	write_lock_bh(&xfrm_policy_lock);
-	for (p = &xfrm_policy_list[id & 7]; (pol=*p)!=NULL; p = &pol->next) {
+	for (p = &xfrm_policy_list[dir]; (pol=*p)!=NULL; p = &pol->next) {
 		if (pol->index == id) {
 			xfrm_pol_hold(pol);
 			if (delete)
@@ -1192,46 +1192,6 @@ int xfrm_bundle_ok(struct xfrm_dst *first, struct flowi *fl, int family)
 
 EXPORT_SYMBOL(xfrm_bundle_ok);
 
-/* Well... that's _TASK_. We need to scan through transformation
- * list and figure out what mss tcp should generate in order to
- * final datagram fit to mtu. Mama mia... :-)
- *
- * Apparently, some easy way exists, but we used to choose the most
- * bizarre ones. :-) So, raising Kalashnikov... tra-ta-ta.
- *
- * Consider this function as something like dark humour. :-)
- */
-static int xfrm_get_mss(struct dst_entry *dst, u32 mtu)
-{
-	int res = mtu - dst->header_len;
-
-	for (;;) {
-		struct dst_entry *d = dst;
-		int m = res;
-
-		do {
-			struct xfrm_state *x = d->xfrm;
-			if (x) {
-				spin_lock_bh(&x->lock);
-				if (x->km.state == XFRM_STATE_VALID &&
-				    x->type && x->type->get_max_size)
-					m = x->type->get_max_size(d->xfrm, m);
-				else
-					m += x->props.header_len;
-				spin_unlock_bh(&x->lock);
-			}
-		} while ((d = d->child) != NULL);
-
-		if (m <= mtu)
-			break;
-		res -= (m - mtu);
-		if (res < 88)
-			return mtu;
-	}
-
-	return res + dst->header_len;
-}
-
 int xfrm_policy_register_afinfo(struct xfrm_policy_afinfo *afinfo)
 {
 	int err = 0;
@@ -1252,8 +1212,6 @@ int xfrm_policy_register_afinfo(struct xfrm_policy_afinfo *afinfo)
 			dst_ops->negative_advice = xfrm_negative_advice;
 		if (likely(dst_ops->link_failure == NULL))
 			dst_ops->link_failure = xfrm_link_failure;
-		if (likely(dst_ops->get_mss == NULL))
-			dst_ops->get_mss = xfrm_get_mss;
 		if (likely(afinfo->garbage_collect == NULL))
 			afinfo->garbage_collect = __xfrm_garbage_collect;
 		xfrm_policy_afinfo[afinfo->family] = afinfo;
@@ -1281,7 +1239,6 @@ int xfrm_policy_unregister_afinfo(struct xfrm_policy_afinfo *afinfo)
 			dst_ops->check = NULL;
 			dst_ops->negative_advice = NULL;
 			dst_ops->link_failure = NULL;
-			dst_ops->get_mss = NULL;
 			afinfo->garbage_collect = NULL;
 		}
 	}
diff --git a/net/xfrm/xfrm_state.c b/net/xfrm/xfrm_state.c
index 9d206c282cf..8b9a4747417 100644
--- a/net/xfrm/xfrm_state.c
+++ b/net/xfrm/xfrm_state.c
@@ -1026,6 +1026,12 @@ void xfrm_state_delete_tunnel(struct xfrm_state *x)
 }
 EXPORT_SYMBOL(xfrm_state_delete_tunnel);
 
+/*
+ * This function is NOT optimal.  For example, with ESP it will give an
+ * MTU that's usually two bytes short of being optimal.  However, it will
+ * usually give an answer that's a multiple of 4 provided the input is
+ * also a multiple of 4.
+ */
 int xfrm_state_mtu(struct xfrm_state *x, int mtu)
 {
 	int res = mtu;