19 files changed, 195 insertions, 148 deletions

diff --git a/net/ipv4/fib_semantics.c b/net/ipv4/fib_semantics.c
index c886b28ba9f5..e278cb9d0075 100644
--- a/net/ipv4/fib_semantics.c
+++ b/net/ipv4/fib_semantics.c
@@ -593,10 +593,13 @@ static void fib_hash_move(struct hlist_head *new_info_hash,
 			  struct hlist_head *new_laddrhash,
 			  unsigned int new_size)
 {
+	struct hlist_head *old_info_hash, *old_laddrhash;
 	unsigned int old_size = fib_hash_size;
-	unsigned int i;
+	unsigned int i, bytes;
 
 	write_lock(&fib_info_lock);
+	old_info_hash = fib_info_hash;
+	old_laddrhash = fib_info_laddrhash;
 	fib_hash_size = new_size;
 
 	for (i = 0; i < old_size; i++) {
@@ -636,6 +639,10 @@ static void fib_hash_move(struct hlist_head *new_info_hash,
 	fib_info_laddrhash = new_laddrhash;
 
 	write_unlock(&fib_info_lock);
+
+	bytes = old_size * sizeof(struct hlist_head *);
+	fib_hash_free(old_info_hash, bytes);
+	fib_hash_free(old_laddrhash, bytes);
 }
 
 struct fib_info *
diff --git a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c
index a701405fab0b..45efd5f4741b 100644
--- a/net/ipv4/fib_trie.c
+++ b/net/ipv4/fib_trie.c
@@ -1333,9 +1333,9 @@ err:;
 }
 
 static inline int check_leaf(struct trie *t, struct leaf *l,  t_key key, int *plen, const struct flowi *flp,
-			     struct fib_result *res, int *err)
+			     struct fib_result *res)
 {
-	int i;
+	int err, i;
 	t_key mask;
 	struct leaf_info *li;
 	struct hlist_head *hhead = &l->list;
@@ -1348,18 +1348,18 @@ static inline int check_leaf(struct trie *t, struct leaf *l,  t_key key, int *pl
 		if (l->key != (key & mask))
 			continue;
 
-		if (((*err) = fib_semantic_match(&li->falh, flp, res, l->key, mask, i)) == 0) {
+		if ((err = fib_semantic_match(&li->falh, flp, res, l->key, mask, i)) <= 0) {
 			*plen = i;
 #ifdef CONFIG_IP_FIB_TRIE_STATS
 			t->stats.semantic_match_passed++;
 #endif
-			return 1;
+			return err;
 		}
 #ifdef CONFIG_IP_FIB_TRIE_STATS
 		t->stats.semantic_match_miss++;
 #endif
 	}
-	return 0;
+	return 1;
 }
 
 static int
@@ -1386,7 +1386,7 @@ fn_trie_lookup(struct fib_table *tb, const struct flowi *flp, struct fib_result
 
 	/* Just a leaf? */
 	if (IS_LEAF(n)) {
-		if (check_leaf(t, (struct leaf *)n, key, &plen, flp, res, &ret))
+		if ((ret = check_leaf(t, (struct leaf *)n, key, &plen, flp, res)) <= 0)
 			goto found;
 		goto failed;
 	}
@@ -1508,7 +1508,7 @@ fn_trie_lookup(struct fib_table *tb, const struct flowi *flp, struct fib_result
 		       continue;
 		}
 		if (IS_LEAF(n)) {
-			if (check_leaf(t, (struct leaf *)n, key, &plen, flp, res, &ret))
+			if ((ret = check_leaf(t, (struct leaf *)n, key, &plen, flp, res)) <= 0)
 				goto found;
 	       }
 backtrace:
diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c
index 279f57abfecb..badfc5849973 100644
--- a/net/ipv4/icmp.c
+++ b/net/ipv4/icmp.c
@@ -349,12 +349,12 @@ static void icmp_push_reply(struct icmp_bxm *icmp_param,
 {
 	struct sk_buff *skb;
 
-	ip_append_data(icmp_socket->sk, icmp_glue_bits, icmp_param,
-		       icmp_param->data_len+icmp_param->head_len,
-		       icmp_param->head_len,
-		       ipc, rt, MSG_DONTWAIT);
-
-	if ((skb = skb_peek(&icmp_socket->sk->sk_write_queue)) != NULL) {
+	if (ip_append_data(icmp_socket->sk, icmp_glue_bits, icmp_param,
+		           icmp_param->data_len+icmp_param->head_len,
+		           icmp_param->head_len,
+		           ipc, rt, MSG_DONTWAIT) < 0)
+		ip_flush_pending_frames(icmp_socket->sk);
+	else if ((skb = skb_peek(&icmp_socket->sk->sk_write_queue)) != NULL) {
 		struct icmphdr *icmph = skb->h.icmph;
 		unsigned int csum = 0;
 		struct sk_buff *skb1;
@@ -936,8 +936,7 @@ int icmp_rcv(struct sk_buff *skb)
 	case CHECKSUM_HW:
 		if (!(u16)csum_fold(skb->csum))
 			break;
-		NETDEBUG(if (net_ratelimit())
-				printk(KERN_DEBUG "icmp v4 hw csum failure\n"));
+		LIMIT_NETDEBUG(printk(KERN_DEBUG "icmp v4 hw csum failure\n"));
 	case CHECKSUM_NONE:
 		if ((u16)csum_fold(skb_checksum(skb, 0, skb->len, 0)))
 			goto error;
diff --git a/net/ipv4/inetpeer.c b/net/ipv4/inetpeer.c
index 95473953c406..ab18a853d7ce 100644
--- a/net/ipv4/inetpeer.c
+++ b/net/ipv4/inetpeer.c
@@ -450,10 +450,13 @@ static void peer_check_expire(unsigned long dummy)
 	/* Trigger the timer after inet_peer_gc_mintime .. inet_peer_gc_maxtime
 	 * interval depending on the total number of entries (more entries,
 	 * less interval). */
-	peer_periodic_timer.expires = jiffies
-		+ inet_peer_gc_maxtime
-		- (inet_peer_gc_maxtime - inet_peer_gc_mintime) / HZ *
-			peer_total / inet_peer_threshold * HZ;
+	if (peer_total >= inet_peer_threshold)
+		peer_periodic_timer.expires = jiffies + inet_peer_gc_mintime;
+	else
+		peer_periodic_timer.expires = jiffies
+			+ inet_peer_gc_maxtime
+			- (inet_peer_gc_maxtime - inet_peer_gc_mintime) / HZ *
+				peer_total / inet_peer_threshold * HZ;
 	add_timer(&peer_periodic_timer);
 }
 
diff --git a/net/ipv4/ip_fragment.c b/net/ipv4/ip_fragment.c
index 7f68e27eb4ea..eb377ae15305 100644
--- a/net/ipv4/ip_fragment.c
+++ b/net/ipv4/ip_fragment.c
@@ -377,7 +377,7 @@ static struct ipq *ip_frag_create(unsigned hash, struct iphdr *iph, u32 user)
 	return ip_frag_intern(hash, qp);
 
 out_nomem:
-	NETDEBUG(if (net_ratelimit()) printk(KERN_ERR "ip_frag_create: no memory left !\n"));
+	LIMIT_NETDEBUG(printk(KERN_ERR "ip_frag_create: no memory left !\n"));
 	return NULL;
 }
 
@@ -625,10 +625,8 @@ static struct sk_buff *ip_frag_reasm(struct ipq *qp, struct net_device *dev)
 	return head;
 
 out_nomem:
- 	NETDEBUG(if (net_ratelimit())
-	         printk(KERN_ERR 
-			"IP: queue_glue: no memory for gluing queue %p\n",
-			qp));
+ 	LIMIT_NETDEBUG(printk(KERN_ERR "IP: queue_glue: no memory for gluing "
+			      "queue %p\n", qp));
 	goto out_fail;
 out_oversize:
 	if (net_ratelimit())
diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c
index 884835522224..f0d5740d7e22 100644
--- a/net/ipv4/ip_gre.c
+++ b/net/ipv4/ip_gre.c
@@ -290,7 +290,6 @@ static struct ip_tunnel * ipgre_tunnel_locate(struct ip_tunnel_parm *parms, int
 
 	dev_hold(dev);
 	ipgre_tunnel_link(nt);
-	/* Do not decrement MOD_USE_COUNT here. */
 	return nt;
 
 failed:
@@ -1277,12 +1276,28 @@ err1:
 	goto out;
 }
 
-static void ipgre_fini(void)
+static void __exit ipgre_destroy_tunnels(void)
+{
+	int prio;
+
+	for (prio = 0; prio < 4; prio++) {
+		int h;
+		for (h = 0; h < HASH_SIZE; h++) {
+			struct ip_tunnel *t;
+			while ((t = tunnels[prio][h]) != NULL)
+				unregister_netdevice(t->dev);
+		}
+	}
+}
+
+static void __exit ipgre_fini(void)
 {
 	if (inet_del_protocol(&ipgre_protocol, IPPROTO_GRE) < 0)
 		printk(KERN_INFO "ipgre close: can't remove protocol\n");
 
-	unregister_netdev(ipgre_fb_tunnel_dev);
+	rtnl_lock();
+	ipgre_destroy_tunnels();
+	rtnl_unlock();
 }
 
 module_init(ipgre_init);
diff --git a/net/ipv4/ip_sockglue.c b/net/ipv4/ip_sockglue.c
index fc7c481d0d79..ff4bd067b397 100644
--- a/net/ipv4/ip_sockglue.c
+++ b/net/ipv4/ip_sockglue.c
@@ -848,6 +848,9 @@ mc_msf_out:
  
 		case IP_IPSEC_POLICY:
 		case IP_XFRM_POLICY:
+			err = -EPERM;
+			if (!capable(CAP_NET_ADMIN))
+				break;
 			err = xfrm_user_policy(sk, optname, optval, optlen);
 			break;
 
diff --git a/net/ipv4/ipcomp.c b/net/ipv4/ipcomp.c
index 2065944fd9e5..7ded6e60f43a 100644
--- a/net/ipv4/ipcomp.c
+++ b/net/ipv4/ipcomp.c
@@ -358,7 +358,7 @@ static struct crypto_tfm **ipcomp_alloc_tfms(const char *alg_name)
 	int cpu;
 
 	/* This can be any valid CPU ID so we don't need locking. */
-	cpu = smp_processor_id();
+	cpu = raw_smp_processor_id();
 
 	list_for_each_entry(pos, &ipcomp_tfms_list, list) {
 		struct crypto_tfm *tfm;
diff --git a/net/ipv4/ipip.c b/net/ipv4/ipip.c
index c3947cd566b7..c05c1df0bb04 100644
--- a/net/ipv4/ipip.c
+++ b/net/ipv4/ipip.c
@@ -255,7 +255,6 @@ static struct ip_tunnel * ipip_tunnel_locate(struct ip_tunnel_parm *parms, int c
 
 	dev_hold(dev);
 	ipip_tunnel_link(nt);
-	/* Do not decrement MOD_USE_COUNT here. */
 	return nt;
 
 failed:
@@ -920,12 +919,29 @@ static int __init ipip_init(void)
 	goto out;
 }
 
+static void __exit ipip_destroy_tunnels(void)
+{
+	int prio;
+
+	for (prio = 1; prio < 4; prio++) {
+		int h;
+		for (h = 0; h < HASH_SIZE; h++) {
+			struct ip_tunnel *t;
+			while ((t = tunnels[prio][h]) != NULL)
+				unregister_netdevice(t->dev);
+		}
+	}
+}
+
 static void __exit ipip_fini(void)
 {
 	if (ipip_unregister() < 0)
 		printk(KERN_INFO "ipip close: can't deregister tunnel\n");
 
-	unregister_netdev(ipip_fb_tunnel_dev);
+	rtnl_lock();
+	ipip_destroy_tunnels();
+	unregister_netdevice(ipip_fb_tunnel_dev);
+	rtnl_unlock();
 }
 
 module_init(ipip_init);
diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c
index 7833d920bdba..dc806b578427 100644
--- a/net/ipv4/ipmr.c
+++ b/net/ipv4/ipmr.c
@@ -362,7 +362,7 @@ out:
 
 /* Fill oifs list. It is called under write locked mrt_lock. */
 
-static void ipmr_update_threshoulds(struct mfc_cache *cache, unsigned char *ttls)
+static void ipmr_update_thresholds(struct mfc_cache *cache, unsigned char *ttls)
 {
 	int vifi;
 
@@ -727,7 +727,7 @@ static int ipmr_mfc_add(struct mfcctl *mfc, int mrtsock)
 	if (c != NULL) {
 		write_lock_bh(&mrt_lock);
 		c->mfc_parent = mfc->mfcc_parent;
-		ipmr_update_threshoulds(c, mfc->mfcc_ttls);
+		ipmr_update_thresholds(c, mfc->mfcc_ttls);
 		if (!mrtsock)
 			c->mfc_flags |= MFC_STATIC;
 		write_unlock_bh(&mrt_lock);
@@ -744,7 +744,7 @@ static int ipmr_mfc_add(struct mfcctl *mfc, int mrtsock)
 	c->mfc_origin=mfc->mfcc_origin.s_addr;
 	c->mfc_mcastgrp=mfc->mfcc_mcastgrp.s_addr;
 	c->mfc_parent=mfc->mfcc_parent;
-	ipmr_update_threshoulds(c, mfc->mfcc_ttls);
+	ipmr_update_thresholds(c, mfc->mfcc_ttls);
 	if (!mrtsock)
 		c->mfc_flags |= MFC_STATIC;
 
diff --git a/net/ipv4/netfilter/ip_conntrack_core.c b/net/ipv4/netfilter/ip_conntrack_core.c
index 86f04e41dd8e..a7f0c821a9b2 100644
--- a/net/ipv4/netfilter/ip_conntrack_core.c
+++ b/net/ipv4/netfilter/ip_conntrack_core.c
@@ -513,6 +513,11 @@ init_conntrack(const struct ip_conntrack_tuple *tuple,
 #ifdef CONFIG_IP_NF_CONNTRACK_MARK
 		conntrack->mark = exp->master->mark;
 #endif
+#if defined(CONFIG_IP_NF_TARGET_MASQUERADE) || \
+    defined(CONFIG_IP_NF_TARGET_MASQUERADE_MODULE)
+		/* this is ugly, but there is no other place where to put it */
+		conntrack->nat.masq_index = exp->master->nat.masq_index;
+#endif
 		nf_conntrack_get(&conntrack->master->ct_general);
 		CONNTRACK_STAT_INC(expect_new);
 	} else {
diff --git a/net/ipv4/netfilter/ip_nat_standalone.c b/net/ipv4/netfilter/ip_nat_standalone.c
index bc59d0d6e89e..91d5ea1dbbc9 100644
--- a/net/ipv4/netfilter/ip_nat_standalone.c
+++ b/net/ipv4/netfilter/ip_nat_standalone.c
@@ -102,6 +102,10 @@ ip_nat_fn(unsigned int hooknum,
 		return NF_ACCEPT;
 	}
 
+	/* Don't try to NAT if this packet is not conntracked */
+	if (ct == &ip_conntrack_untracked)
+		return NF_ACCEPT;
+
 	switch (ctinfo) {
 	case IP_CT_RELATED:
 	case IP_CT_RELATED+IP_CT_IS_REPLY:
diff --git a/net/ipv4/netfilter/ip_queue.c b/net/ipv4/netfilter/ip_queue.c
index eda1fba431a4..c6baa8174389 100644
--- a/net/ipv4/netfilter/ip_queue.c
+++ b/net/ipv4/netfilter/ip_queue.c
@@ -214,6 +214,12 @@ ipq_build_packet_message(struct ipq_queue_entry *entry, int *errp)
 		break;
 	
 	case IPQ_COPY_PACKET:
+		if (entry->skb->ip_summed == CHECKSUM_HW &&
+		    (*errp = skb_checksum_help(entry->skb,
+		                               entry->info->outdev == NULL))) {
+			read_unlock_bh(&queue_lock);
+			return NULL;
+		}
 		if (copy_range == 0 || copy_range > entry->skb->len)
 			data_len = entry->skb->len;
 		else
@@ -385,6 +391,7 @@ ipq_mangle_ipv4(ipq_verdict_msg_t *v, struct ipq_queue_entry *e)
 	if (!skb_ip_make_writable(&e->skb, v->data_len))
 		return -ENOMEM;
 	memcpy(e->skb->data, v->payload, v->data_len);
+	e->skb->ip_summed = CHECKSUM_NONE;
 	e->skb->nfcache |= NFC_ALTERED;
 
 	/*
diff --git a/net/ipv4/netfilter/ipt_ECN.c b/net/ipv4/netfilter/ipt_ECN.c
index ada9911118e9..94a0ce1c1c9d 100644
--- a/net/ipv4/netfilter/ipt_ECN.c
+++ b/net/ipv4/netfilter/ipt_ECN.c
@@ -61,16 +61,20 @@ set_ect_tcp(struct sk_buff **pskb, const struct ipt_ECN_info *einfo, int inward)
 	if (!tcph)
 		return 0;
 
-	if (!(einfo->operation & IPT_ECN_OP_SET_ECE
-	      || tcph->ece == einfo->proto.tcp.ece)
-	    && (!(einfo->operation & IPT_ECN_OP_SET_CWR
-		  || tcph->cwr == einfo->proto.tcp.cwr)))
+	if ((!(einfo->operation & IPT_ECN_OP_SET_ECE) ||
+	     tcph->ece == einfo->proto.tcp.ece) &&
+	    ((!(einfo->operation & IPT_ECN_OP_SET_CWR) ||
+	     tcph->cwr == einfo->proto.tcp.cwr)))
 		return 1;
 
 	if (!skb_ip_make_writable(pskb, (*pskb)->nh.iph->ihl*4+sizeof(*tcph)))
 		return 0;
 	tcph = (void *)(*pskb)->nh.iph + (*pskb)->nh.iph->ihl*4;
 
+	if ((*pskb)->ip_summed == CHECKSUM_HW &&
+	    skb_checksum_help(*pskb, inward))
+		return 0;
+
 	diffs[0] = ((u_int16_t *)tcph)[6];
 	if (einfo->operation & IPT_ECN_OP_SET_ECE)
 		tcph->ece = einfo->proto.tcp.ece;
@@ -79,13 +83,10 @@ set_ect_tcp(struct sk_buff **pskb, const struct ipt_ECN_info *einfo, int inward)
 	diffs[1] = ((u_int16_t *)tcph)[6];
 	diffs[0] = diffs[0] ^ 0xFFFF;
 
-	if ((*pskb)->ip_summed != CHECKSUM_HW)
+	if ((*pskb)->ip_summed != CHECKSUM_UNNECESSARY)
 		tcph->check = csum_fold(csum_partial((char *)diffs,
 						     sizeof(diffs),
 						     tcph->check^0xFFFF));
-	else
-		if (skb_checksum_help(*pskb, inward))
-			return 0;
 	(*pskb)->nfcache |= NFC_ALTERED;
 	return 1;
 }
diff --git a/net/ipv4/netfilter/ipt_TCPMSS.c b/net/ipv4/netfilter/ipt_TCPMSS.c
index 1049050b2bfb..7b84a254440e 100644
--- a/net/ipv4/netfilter/ipt_TCPMSS.c
+++ b/net/ipv4/netfilter/ipt_TCPMSS.c
@@ -61,6 +61,10 @@ ipt_tcpmss_target(struct sk_buff **pskb,
 	if (!skb_ip_make_writable(pskb, (*pskb)->len))
 		return NF_DROP;
 
+	if ((*pskb)->ip_summed == CHECKSUM_HW &&
+	    skb_checksum_help(*pskb, out == NULL))
+		return NF_DROP;
+
 	iph = (*pskb)->nh.iph;
 	tcplen = (*pskb)->len - iph->ihl*4;
 
@@ -186,9 +190,6 @@ ipt_tcpmss_target(struct sk_buff **pskb,
 	       newmss);
 
  retmodified:
-	/* We never hw checksum SYN packets.  */
-	BUG_ON((*pskb)->ip_summed == CHECKSUM_HW);
-
 	(*pskb)->nfcache |= NFC_UNKNOWN | NFC_ALTERED;
 	return IPT_CONTINUE;
 }
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index ddb6ce4ecff2..69b1fcf70077 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -584,7 +584,7 @@ static inline void skb_entail(struct sock *sk, struct tcp_sock *tp,
 	sk_charge_skb(sk, skb);
 	if (!sk->sk_send_head)
 		sk->sk_send_head = skb;
-	else if (tp->nonagle&TCP_NAGLE_PUSH)
+	if (tp->nonagle & TCP_NAGLE_PUSH)
 		tp->nonagle &= ~TCP_NAGLE_PUSH; 
 }
 
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 62f62bb05c2a..67c670886c1f 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -242,9 +242,14 @@ static int tcp_v4_get_port(struct sock *sk, unsigned short snum)
 		tcp_port_rover = rover;
 		spin_unlock(&tcp_portalloc_lock);
 
-		/* Exhausted local port range during search? */
+		/* Exhausted local port range during search?  It is not
+		 * possible for us to be holding one of the bind hash
+		 * locks if this test triggers, because if 'remaining'
+		 * drops to zero, we broke out of the do/while loop at
+		 * the top level, not from the 'break;' statement.
+		 */
 		ret = 1;
-		if (remaining <= 0)
+		if (unlikely(remaining <= 0))
 			goto fail;
 
 		/* OK, here is the one we will use.  HEAD is
@@ -1494,12 +1499,11 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
 			 * to destinations, already remembered
 			 * to the moment of synflood.
 			 */
-			NETDEBUG(if (net_ratelimit()) \
-					printk(KERN_DEBUG "TCP: drop open "
-							  "request from %u.%u."
-							  "%u.%u/%u\n", \
-					       NIPQUAD(saddr),
-					       ntohs(skb->h.th->source)));
+			LIMIT_NETDEBUG(printk(KERN_DEBUG "TCP: drop open "
+					      "request from %u.%u."
+					      "%u.%u/%u\n",
+					      NIPQUAD(saddr),
+					      ntohs(skb->h.th->source)));
 			dst_release(dst);
 			goto drop_and_free;
 		}
@@ -1627,8 +1631,7 @@ static int tcp_v4_checksum_init(struct sk_buff *skb)
 				  skb->nh.iph->daddr, skb->csum))
 			return 0;
 
-		NETDEBUG(if (net_ratelimit())
-				printk(KERN_DEBUG "hw tcp v4 csum failed\n"));
+		LIMIT_NETDEBUG(printk(KERN_DEBUG "hw tcp v4 csum failed\n"));
 		skb->ip_summed = CHECKSUM_NONE;
 	}
 	if (skb->len <= 76) {
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index e3f8ea1bfa9c..dd30dd137b74 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -403,11 +403,9 @@ static void tcp_queue_skb(struct sock *sk, struct sk_buff *skb)
 		sk->sk_send_head = skb;
 }
 
-static void tcp_set_skb_tso_segs(struct sock *sk, struct sk_buff *skb)
+static void tcp_set_skb_tso_segs(struct sock *sk, struct sk_buff *skb, unsigned int mss_now)
 {
-	struct tcp_sock *tp = tcp_sk(sk);
-
-	if (skb->len <= tp->mss_cache ||
+	if (skb->len <= mss_now ||
 	    !(sk->sk_route_caps & NETIF_F_TSO)) {
 		/* Avoid the costly divide in the normal
 		 * non-TSO case.
@@ -417,10 +415,10 @@ static void tcp_set_skb_tso_segs(struct sock *sk, struct sk_buff *skb)
 	} else {
 		unsigned int factor;
 
-		factor = skb->len + (tp->mss_cache - 1);
-		factor /= tp->mss_cache;
+		factor = skb->len + (mss_now - 1);
+		factor /= mss_now;
 		skb_shinfo(skb)->tso_segs = factor;
-		skb_shinfo(skb)->tso_size = tp->mss_cache;
+		skb_shinfo(skb)->tso_size = mss_now;
 	}
 }
 
@@ -429,7 +427,7 @@ static void tcp_set_skb_tso_segs(struct sock *sk, struct sk_buff *skb)
  * packet to the list.  This won't be called frequently, I hope. 
  * Remember, these are still headerless SKBs at this point.
  */
-static int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len)
+static int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len, unsigned int mss_now)
 {
 	struct tcp_sock *tp = tcp_sk(sk);
 	struct sk_buff *buff;
@@ -492,8 +490,8 @@ static int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len)
 	}
 
 	/* Fix up tso_factor for both original and new SKB.  */
-	tcp_set_skb_tso_segs(sk, skb);
-	tcp_set_skb_tso_segs(sk, buff);
+	tcp_set_skb_tso_segs(sk, skb, mss_now);
+	tcp_set_skb_tso_segs(sk, buff, mss_now);
 
 	if (TCP_SKB_CB(skb)->sacked & TCPCB_LOST) {
 		tp->lost_out += tcp_skb_pcount(skb);
@@ -569,7 +567,7 @@ int tcp_trim_head(struct sock *sk, struct sk_buff *skb, u32 len)
 	 * factor and mss.
 	 */
 	if (tcp_skb_pcount(skb) > 1)
-		tcp_set_skb_tso_segs(sk, skb);
+		tcp_set_skb_tso_segs(sk, skb, tcp_current_mss(sk, 1));
 
 	return 0;
 }
@@ -734,12 +732,14 @@ static inline unsigned int tcp_cwnd_test(struct tcp_sock *tp, struct sk_buff *sk
 /* This must be invoked the first time we consider transmitting
  * SKB onto the wire.
  */
-static inline int tcp_init_tso_segs(struct sock *sk, struct sk_buff *skb)
+static inline int tcp_init_tso_segs(struct sock *sk, struct sk_buff *skb, unsigned int mss_now)
 {
 	int tso_segs = tcp_skb_pcount(skb);
 
-	if (!tso_segs) {
-		tcp_set_skb_tso_segs(sk, skb);
+	if (!tso_segs ||
+	    (tso_segs > 1 &&
+	     skb_shinfo(skb)->tso_size != mss_now)) {
+		tcp_set_skb_tso_segs(sk, skb, mss_now);
 		tso_segs = tcp_skb_pcount(skb);
 	}
 	return tso_segs;
@@ -817,7 +817,7 @@ static unsigned int tcp_snd_test(struct sock *sk, struct sk_buff *skb,
 	struct tcp_sock *tp = tcp_sk(sk);
 	unsigned int cwnd_quota;
 
-	tcp_init_tso_segs(sk, skb);
+	tcp_init_tso_segs(sk, skb, cur_mss);
 
 	if (!tcp_nagle_test(tp, skb, cur_mss, nonagle))
 		return 0;
@@ -854,14 +854,15 @@ int tcp_may_send_now(struct sock *sk, struct tcp_sock *tp)
  * know that all the data is in scatter-gather pages, and that the
  * packet has never been sent out before (and thus is not cloned).
  */
-static int tso_fragment(struct sock *sk, struct sk_buff *skb, unsigned int len)
+static int tso_fragment(struct sock *sk, struct sk_buff *skb, unsigned int len, unsigned int mss_now)
 {
 	struct sk_buff *buff;
 	int nlen = skb->len - len;
 	u16 flags;
 
 	/* All of a TSO frame must be composed of paged data.  */
-	BUG_ON(skb->len != skb->data_len);
+	if (skb->len != skb->data_len)
+		return tcp_fragment(sk, skb, len, mss_now);
 
 	buff = sk_stream_alloc_pskb(sk, 0, 0, GFP_ATOMIC);
 	if (unlikely(buff == NULL))
@@ -887,8 +888,8 @@ static int tso_fragment(struct sock *sk, struct sk_buff *skb, unsigned int len)
 	skb_split(skb, buff, len);
 
 	/* Fix up tso_factor for both original and new SKB.  */
-	tcp_set_skb_tso_segs(sk, skb);
-	tcp_set_skb_tso_segs(sk, buff);
+	tcp_set_skb_tso_segs(sk, skb, mss_now);
+	tcp_set_skb_tso_segs(sk, buff, mss_now);
 
 	/* Link BUFF into the send queue. */
 	skb_header_release(buff);
@@ -924,10 +925,6 @@ static int tcp_tso_should_defer(struct sock *sk, struct tcp_sock *tp, struct sk_
 
 	limit = min(send_win, cong_win);
 
-	/* If sk_send_head can be sent fully now, just do it.  */
-	if (skb->len <= limit)
-		return 0;
-
 	if (sysctl_tcp_tso_win_divisor) {
 		u32 chunk = min(tp->snd_wnd, tp->snd_cwnd * tp->mss_cache);
 
@@ -972,19 +969,20 @@ static int tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle)
 	if (unlikely(sk->sk_state == TCP_CLOSE))
 		return 0;
 
-	skb = sk->sk_send_head;
-	if (unlikely(!skb))
-		return 0;
-
-	tso_segs = tcp_init_tso_segs(sk, skb);
-	cwnd_quota = tcp_cwnd_test(tp, skb);
-	if (unlikely(!cwnd_quota))
-		goto out;
-
 	sent_pkts = 0;
-	while (likely(tcp_snd_wnd_test(tp, skb, mss_now))) {
+	while ((skb = sk->sk_send_head)) {
+		unsigned int limit;
+
+		tso_segs = tcp_init_tso_segs(sk, skb, mss_now);
 		BUG_ON(!tso_segs);
 
+		cwnd_quota = tcp_cwnd_test(tp, skb);
+		if (!cwnd_quota)
+			break;
+
+		if (unlikely(!tcp_snd_wnd_test(tp, skb, mss_now)))
+			break;
+
 		if (tso_segs == 1) {
 			if (unlikely(!tcp_nagle_test(tp, skb, mss_now,
 						     (tcp_skb_is_last(sk, skb) ?
@@ -995,9 +993,10 @@ static int tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle)
 				break;
 		}
 
+		limit = mss_now;
 		if (tso_segs > 1) {
-			u32 limit = tcp_window_allows(tp, skb,
-						      mss_now, cwnd_quota);
+			limit = tcp_window_allows(tp, skb,
+						  mss_now, cwnd_quota);
 
 			if (skb->len < limit) {
 				unsigned int trim = skb->len % mss_now;
@@ -1005,15 +1004,12 @@ static int tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle)
 				if (trim)
 					limit = skb->len - trim;
 			}
-			if (skb->len > limit) {
-				if (tso_fragment(sk, skb, limit))
-					break;
-			}
-		} else if (unlikely(skb->len > mss_now)) {
-			if (unlikely(tcp_fragment(sk, skb,  mss_now)))
-				break;
 		}
 
+		if (skb->len > limit &&
+		    unlikely(tso_fragment(sk, skb, limit, mss_now)))
+			break;
+
 		TCP_SKB_CB(skb)->when = tcp_time_stamp;
 
 		if (unlikely(tcp_transmit_skb(sk, skb_clone(skb, GFP_ATOMIC))))
@@ -1026,27 +1022,12 @@ static int tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle)
 
 		tcp_minshall_update(tp, mss_now, skb);
 		sent_pkts++;
-
-		/* Do not optimize this to use tso_segs. If we chopped up
-		 * the packet above, tso_segs will no longer be valid.
-		 */
-		cwnd_quota -= tcp_skb_pcount(skb);
-
-		BUG_ON(cwnd_quota < 0);
-		if (!cwnd_quota)
-			break;
-
-		skb = sk->sk_send_head;
-		if (!skb)
-			break;
-		tso_segs = tcp_init_tso_segs(sk, skb);
 	}
 
 	if (likely(sent_pkts)) {
 		tcp_cwnd_validate(sk, tp);
 		return 0;
 	}
-out:
 	return !tp->packets_out && sk->sk_send_head;
 }
 
@@ -1076,15 +1057,18 @@ void tcp_push_one(struct sock *sk, unsigned int mss_now)
 
 	BUG_ON(!skb || skb->len < mss_now);
 
-	tso_segs = tcp_init_tso_segs(sk, skb);
+	tso_segs = tcp_init_tso_segs(sk, skb, mss_now);
 	cwnd_quota = tcp_snd_test(sk, skb, mss_now, TCP_NAGLE_PUSH);
 
 	if (likely(cwnd_quota)) {
+		unsigned int limit;
+
 		BUG_ON(!tso_segs);
 
+		limit = mss_now;
 		if (tso_segs > 1) {
-			u32 limit = tcp_window_allows(tp, skb,
-						      mss_now, cwnd_quota);
+			limit = tcp_window_allows(tp, skb,
+						  mss_now, cwnd_quota);
 
 			if (skb->len < limit) {
 				unsigned int trim = skb->len % mss_now;
@@ -1092,15 +1076,12 @@ void tcp_push_one(struct sock *sk, unsigned int mss_now)
 				if (trim)
 					limit = skb->len - trim;
 			}
-			if (skb->len > limit) {
-				if (unlikely(tso_fragment(sk, skb, limit)))
-					return;
-			}
-		} else if (unlikely(skb->len > mss_now)) {
-			if (unlikely(tcp_fragment(sk, skb, mss_now)))
-				return;
 		}
 
+		if (skb->len > limit &&
+		    unlikely(tso_fragment(sk, skb, limit, mss_now)))
+			return;
+
 		/* Send it out now. */
 		TCP_SKB_CB(skb)->when = tcp_time_stamp;
 
@@ -1386,15 +1367,21 @@ int tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb)
 
 	if (skb->len > cur_mss) {
 		int old_factor = tcp_skb_pcount(skb);
-		int new_factor;
+		int diff;
 
-		if (tcp_fragment(sk, skb, cur_mss))
+		if (tcp_fragment(sk, skb, cur_mss, cur_mss))
 			return -ENOMEM; /* We'll try again later. */
 
 		/* New SKB created, account for it. */
-		new_factor = tcp_skb_pcount(skb);
-		tp->packets_out -= old_factor - new_factor;
-		tp->packets_out += tcp_skb_pcount(skb->next);
+		diff = old_factor - tcp_skb_pcount(skb) -
+		       tcp_skb_pcount(skb->next);
+		tp->packets_out -= diff;
+
+		if (diff > 0) {
+			tp->fackets_out -= diff;
+			if ((int)tp->fackets_out < 0)
+				tp->fackets_out = 0;
+		}
 	}
 
 	/* Collapse two adjacent packets if worthwhile and we can. */
@@ -1991,7 +1978,7 @@ int tcp_write_wakeup(struct sock *sk)
 			    skb->len > mss) {
 				seg_size = min(seg_size, mss);
 				TCP_SKB_CB(skb)->flags |= TCPCB_FLAG_PSH;
-				if (tcp_fragment(sk, skb, seg_size))
+				if (tcp_fragment(sk, skb, seg_size, mss))
 					return -1;
 				/* SWS override triggered forced fragmentation.
 				 * Disable TSO, the connection is too sick. */
@@ -2000,7 +1987,7 @@ int tcp_write_wakeup(struct sock *sk)
 					sk->sk_route_caps &= ~NETIF_F_TSO;
 				}
 			} else if (!tcp_skb_pcount(skb))
-				tcp_set_skb_tso_segs(sk, skb);
+				tcp_set_skb_tso_segs(sk, skb, mss);
 
 			TCP_SKB_CB(skb)->flags |= TCPCB_FLAG_PSH;
 			TCP_SKB_CB(skb)->when = tcp_time_stamp;
diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c
index 7c24e64b443f..dc4d07357e3a 100644
--- a/net/ipv4/udp.c
+++ b/net/ipv4/udp.c
@@ -628,7 +628,7 @@ back_from_confirm:
 		/* ... which is an evident application bug. --ANK */
 		release_sock(sk);
 
-		NETDEBUG(if (net_ratelimit()) printk(KERN_DEBUG "udp cork app bug 2\n"));
+		LIMIT_NETDEBUG(printk(KERN_DEBUG "udp cork app bug 2\n"));
 		err = -EINVAL;
 		goto out;
 	}
@@ -693,7 +693,7 @@ static int udp_sendpage(struct sock *sk, struct page *page, int offset,
 	if (unlikely(!up->pending)) {
 		release_sock(sk);
 
-		NETDEBUG(if (net_ratelimit()) printk(KERN_DEBUG "udp cork app bug 3\n"));
+		LIMIT_NETDEBUG(printk(KERN_DEBUG "udp cork app bug 3\n"));
 		return -EINVAL;
 	}
 
@@ -1102,7 +1102,7 @@ static int udp_checksum_init(struct sk_buff *skb, struct udphdr *uh,
 		skb->ip_summed = CHECKSUM_UNNECESSARY;
 		if (!udp_check(uh, ulen, saddr, daddr, skb->csum))
 			return 0;
-		NETDEBUG(if (net_ratelimit()) printk(KERN_DEBUG "udp v4 hw csum failure.\n"));
+		LIMIT_NETDEBUG(printk(KERN_DEBUG "udp v4 hw csum failure.\n"));
 		skb->ip_summed = CHECKSUM_NONE;
 	}
 	if (skb->ip_summed != CHECKSUM_UNNECESSARY)
@@ -1181,14 +1181,13 @@ int udp_rcv(struct sk_buff *skb)
 	return(0);
 
 short_packet:
-	NETDEBUG(if (net_ratelimit())
-		printk(KERN_DEBUG "UDP: short packet: From %u.%u.%u.%u:%u %d/%d to %u.%u.%u.%u:%u\n",
-			NIPQUAD(saddr),
-			ntohs(uh->source),
-			ulen,
-			len,
-			NIPQUAD(daddr),
-			ntohs(uh->dest)));
+	LIMIT_NETDEBUG(printk(KERN_DEBUG "UDP: short packet: From %u.%u.%u.%u:%u %d/%d to %u.%u.%u.%u:%u\n",
+			      NIPQUAD(saddr),
+			      ntohs(uh->source),
+			      ulen,
+			      len,
+			      NIPQUAD(daddr),
+			      ntohs(uh->dest)));
 no_header:
 	UDP_INC_STATS_BH(UDP_MIB_INERRORS);
 	kfree_skb(skb);
@@ -1199,13 +1198,12 @@ csum_error:
 	 * RFC1122: OK.  Discards the bad packet silently (as far as 
 	 * the network is concerned, anyway) as per 4.1.3.4 (MUST). 
 	 */
-	NETDEBUG(if (net_ratelimit())
-		 printk(KERN_DEBUG "UDP: bad checksum. From %d.%d.%d.%d:%d to %d.%d.%d.%d:%d ulen %d\n",
-			NIPQUAD(saddr),
-			ntohs(uh->source),
-			NIPQUAD(daddr),
-			ntohs(uh->dest),
-			ulen));
+	LIMIT_NETDEBUG(printk(KERN_DEBUG "UDP: bad checksum. From %d.%d.%d.%d:%d to %d.%d.%d.%d:%d ulen %d\n",
+			      NIPQUAD(saddr),
+			      ntohs(uh->source),
+			      NIPQUAD(daddr),
+			      ntohs(uh->dest),
+			      ulen));
 drop:
 	UDP_INC_STATS_BH(UDP_MIB_INERRORS);
 	kfree_skb(skb);