Merge davem@nuts.ninka.net:/home/davem/src/BK/net-2.5

into kernel.bkbits.net:/home/davem/net-2.5

MAINTAINERS

@@ -316,6 +316,15 @@ M:	marcel@holtmann.org
 W:	http://www.holtmann.org/linux/bluetooth/
 S:	Maintained
 
+BONDING DRIVER
+P:   Chad Tindel
+M:   ctindel@users.sourceforge.net
+P:   Jay Vosburgh
+M:   fubar@us.ibm.com
+L:   bonding-devel@lists.sourceforge.net
+W:   http://sourceforge.net/projects/bonding/
+S:   Supported
+
 BTTV VIDEO4LINUX DRIVER
 P:	Gerd Knorr
 M:	kraxel@bytesex.org

include/linux/skbuff.h

@@ -1102,7 +1102,7 @@ static inline int skb_cow(struct sk_buff *skb, unsigned int headroom)
  
 static inline struct sk_buff *skb_padto(struct sk_buff *skb, unsigned int len)
 {
-	unsigned int size = skb->len + skb->data_len;
+	unsigned int size = skb->len;
 	if (likely(size >= len))
 		return skb;
 	return skb_pad(skb, len-size);

include/net/dn_route.h

@@ -122,7 +122,7 @@ static inline void dn_nsp_send(struct sk_buff *skb)
 	if ((dst = sk->dst_cache) && !dst->obsolete) {
 try_again:
 		skb->dst = dst_clone(dst);
-		dst->output(skb);
+		dst_output(skb);
 		return;
 	}
 

net/decnet/dn_nsp_out.c

@@ -593,7 +593,7 @@ static __inline__ void dn_nsp_do_disc(struct sock *sk, unsigned char msgflg,
 	 * associations.
 	 */
 	skb->dst = dst_clone(dst);
-	skb->dst->output(skb);
+	dst_output(skb);
 }
 
 

net/decnet/dn_route.c

@@ -389,7 +389,7 @@ static int dn_route_rx_packet(struct sk_buff *skb)
 	int err;
 
 	if ((err = dn_route_input(skb)) == 0)
-		return skb->dst->input(skb);
+		return dst_input(skb);
 
 	if (decnet_debug_level & 4) {
 		char *devname = skb->dev ? skb->dev->name : "???";

net/ipv4/igmp.c

@@ -184,14 +184,6 @@ static void igmp_mod_timer(struct ip_mc_list *im, int max_delay)
 
 #define IGMP_SIZE (sizeof(struct igmphdr)+sizeof(struct iphdr)+4)
 
-/* Don't just hand NF_HOOK skb->dst->output, in case netfilter hook
-   changes route */
-static inline int
-output_maybe_reroute(struct sk_buff *skb)
-{
-	return skb->dst->output(skb);
-}
-
 static int igmp_send_report(struct net_device *dev, u32 group, int type)
 {
 	struct sk_buff *skb;
@@ -255,7 +247,7 @@ static int igmp_send_report(struct net_device *dev, u32 group, int type)
 	ih->csum=ip_compute_csum((void *)ih, sizeof(struct igmphdr));
 
 	return NF_HOOK(PF_INET, NF_IP_LOCAL_OUT, skb, NULL, rt->u.dst.dev,
-		       output_maybe_reroute);
+		       dst_output);
 }
 
 

net/ipv4/ip_fragment.c

@@ -19,6 +19,7 @@
  *		Bill Hawes	:	Frag accounting and evictor fixes.
  *		John McDonald	:	0 length frag bug.
  *		Alexey Kuznetsov:	SMP races, threading, cleanup.
+ *		Patrick McHardy :	LRU queue of frag heads for evictor.
  */
 
 #include <linux/config.h>
@@ -26,6 +27,7 @@
 #include <linux/mm.h>
 #include <linux/jiffies.h>
 #include <linux/skbuff.h>
+#include <linux/list.h>
 #include <linux/ip.h>
 #include <linux/icmp.h>
 #include <linux/netdevice.h>
@@ -67,6 +69,7 @@ struct ipfrag_skb_cb
 /* Describe an entry in the "incomplete datagrams" queue. */
 struct ipq {
 	struct ipq	*next;		/* linked list pointers			*/
+	struct list_head lru_list;	/* lru list member 			*/
 	u32		saddr;
 	u32		daddr;
 	u16		id;
@@ -94,6 +97,7 @@ struct ipq {
 /* Per-bucket lock is easy to add now. */
 static struct ipq *ipq_hash[IPQ_HASHSZ];
 static rwlock_t ipfrag_lock = RW_LOCK_UNLOCKED;
+static LIST_HEAD(ipq_lru_list);
 int ip_frag_nqueues = 0;
 
 static __inline__ void __ipq_unlink(struct ipq *qp)
@@ -101,6 +105,7 @@ static __inline__ void __ipq_unlink(struct ipq *qp)
 	if(qp->next)
 		qp->next->pprev = qp->pprev;
 	*qp->pprev = qp->next;
+	list_del(&qp->lru_list);
 	ip_frag_nqueues--;
 }
 
@@ -202,39 +207,30 @@ static void ipq_kill(struct ipq *ipq)
  */
 static void ip_evictor(void)
 {
-	int i, progress;
+	struct ipq *qp;
+	struct list_head *tmp;
 
-	do {
+	for(;;) {
 		if (atomic_read(&ip_frag_mem) <= sysctl_ipfrag_low_thresh)
 			return;
-		progress = 0;
-		/* FIXME: Make LRU queue of frag heads. -DaveM */
-		for (i = 0; i < IPQ_HASHSZ; i++) {
-			struct ipq *qp;
-			if (ipq_hash[i] == NULL)
-				continue;
-
-			read_lock(&ipfrag_lock);
-			if ((qp = ipq_hash[i]) != NULL) {
-				/* find the oldest queue for this hash bucket */
-				while (qp->next)
-					qp = qp->next;
-				atomic_inc(&qp->refcnt);
-				read_unlock(&ipfrag_lock);
-
-				spin_lock(&qp->lock);
-				if (!(qp->last_in&COMPLETE))
-					ipq_kill(qp);
-				spin_unlock(&qp->lock);
-
-				ipq_put(qp);
-				IP_INC_STATS_BH(IpReasmFails);
-				progress = 1;
-				continue;
-			}
+		read_lock(&ipfrag_lock);
+		if (list_empty(&ipq_lru_list)) {
 			read_unlock(&ipfrag_lock);
+			return;
 		}
-	} while (progress);
+		tmp = ipq_lru_list.next;
+		qp = list_entry(tmp, struct ipq, lru_list);
+		atomic_inc(&qp->refcnt);
+		read_unlock(&ipfrag_lock);
+
+		spin_lock(&qp->lock);
+		if (!(qp->last_in&COMPLETE))
+			ipq_kill(qp);
+		spin_unlock(&qp->lock);
+
+		ipq_put(qp);
+		IP_INC_STATS_BH(IpReasmFails);
+	}
 }
 
 /*
@@ -302,6 +298,8 @@ static struct ipq *ip_frag_intern(unsigned int hash, struct ipq *qp_in)
 		qp->next->pprev = &qp->next;
 	ipq_hash[hash] = qp;
 	qp->pprev = &ipq_hash[hash];
+	INIT_LIST_HEAD(&qp->lru_list);
+	list_add_tail(&qp->lru_list, &ipq_lru_list);
 	ip_frag_nqueues++;
 	write_unlock(&ipfrag_lock);
 	return qp;
@@ -496,6 +494,10 @@ static void ip_frag_queue(struct ipq *qp, struct sk_buff *skb)
 	if (offset == 0)
 		qp->last_in |= FIRST_IN;
 
+	write_lock(&ipfrag_lock);
+	list_move_tail(&qp->lru_list, &ipq_lru_list);
+	write_unlock(&ipfrag_lock);
+
 	return;
 
 err:

net/ipv4/ip_input.c

@@ -344,7 +344,7 @@ static inline int ip_rcv_finish(struct sk_buff *skb)
 		}
 	}
 
-	return skb->dst->input(skb);
+	return dst_input(skb);
 
 inhdr_error:
 	IP_INC_STATS_BH(IpInHdrErrors);

net/ipv4/ipmr.c

@@ -1112,9 +1112,9 @@ static inline int ipmr_forward_finish(struct sk_buff *skb)
 	struct dst_entry *dst = skb->dst;
 
 	if (skb->len <= dst_pmtu(dst))
-		return dst->output(skb);
+		return dst_output(skb);
 	else
-		return ip_fragment(skb, dst->output);
+		return ip_fragment(skb, dst_output);
 }
 
 /*

net/ipv4/route.c

@@ -2409,149 +2409,149 @@ static int ipv4_sysctl_rtcache_flush_strategy(ctl_table *table, int *name,
 
 ctl_table ipv4_route_table[] = {
         {
-		.ctl_name =	NET_IPV4_ROUTE_FLUSH,
-		.procname =	"flush",
-		.data =	&flush_delay,
-		.maxlen =		sizeof(int),
-		.mode =	0644,
-		.proc_handler =&ipv4_sysctl_rtcache_flush,
-		.strategy =	&ipv4_sysctl_rtcache_flush_strategy,
+		.ctl_name 	= NET_IPV4_ROUTE_FLUSH,
+		.procname	= "flush",
+		.data		= &flush_delay,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &ipv4_sysctl_rtcache_flush,
+		.strategy	= &ipv4_sysctl_rtcache_flush_strategy,
 	},
 	{
-		.ctl_name =	NET_IPV4_ROUTE_MIN_DELAY,
-		.procname =	"min_delay",
-		.data =	&ip_rt_min_delay,
-		.maxlen =		sizeof(int),
-		.mode =	0644,
-		.proc_handler =&proc_dointvec_jiffies,
-		.strategy =	&sysctl_jiffies,
+		.ctl_name	= NET_IPV4_ROUTE_MIN_DELAY,
+		.procname	= "min_delay",
+		.data		= &ip_rt_min_delay,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec_jiffies,
+		.strategy	= &sysctl_jiffies,
 	},
 	{
-		.ctl_name =	NET_IPV4_ROUTE_MAX_DELAY,
-		.procname =	"max_delay",
-		.data =	&ip_rt_max_delay,
-		.maxlen =		sizeof(int),
-		.mode =	0644,
-		.proc_handler =&proc_dointvec_jiffies,
-		.strategy =	&sysctl_jiffies,
+		.ctl_name	= NET_IPV4_ROUTE_MAX_DELAY,
+		.procname	= "max_delay",
+		.data		= &ip_rt_max_delay,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec_jiffies,
+		.strategy	= &sysctl_jiffies,
 	},
 	{
-		.ctl_name =	NET_IPV4_ROUTE_GC_THRESH,
-		.procname =	"gc_thresh",
-		.data =	&ipv4_dst_ops.gc_thresh,
-		.maxlen =		sizeof(int),
-		.mode =	0644,
-		.proc_handler =&proc_dointvec,
+		.ctl_name	= NET_IPV4_ROUTE_GC_THRESH,
+		.procname	= "gc_thresh",
+		.data		= &ipv4_dst_ops.gc_thresh,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec,
 	},
 	{
-		.ctl_name =	NET_IPV4_ROUTE_MAX_SIZE,
-		.procname =	"max_size",
-		.data =	&ip_rt_max_size,
-		.maxlen =		sizeof(int),
-		.mode =	0644,
-		.proc_handler =&proc_dointvec,
+		.ctl_name	= NET_IPV4_ROUTE_MAX_SIZE,
+		.procname	= "max_size",
+		.data		= &ip_rt_max_size,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec,
 	},
 	{
-		.ctl_name =	NET_IPV4_ROUTE_GC_MIN_INTERVAL,
-		.procname =	"gc_min_interval",
-		.data =	&ip_rt_gc_min_interval,
-		.maxlen =		sizeof(int),
-		.mode =	0644,
-		.proc_handler =&proc_dointvec_jiffies,
-		.strategy =	&sysctl_jiffies,
+		.ctl_name	= NET_IPV4_ROUTE_GC_MIN_INTERVAL,
+		.procname	= "gc_min_interval",
+		.data		= &ip_rt_gc_min_interval,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec_jiffies,
+		.strategy	= &sysctl_jiffies,
 	},
 	{
-		.ctl_name =	NET_IPV4_ROUTE_GC_TIMEOUT,
-		.procname =	"gc_timeout",
-		.data =	&ip_rt_gc_timeout,
-		.maxlen =		sizeof(int),
-		.mode =	0644,
-		.proc_handler =&proc_dointvec_jiffies,
-		.strategy =	&sysctl_jiffies,
+		.ctl_name	= NET_IPV4_ROUTE_GC_TIMEOUT,
+		.procname	= "gc_timeout",
+		.data		= &ip_rt_gc_timeout,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec_jiffies,
+		.strategy	= &sysctl_jiffies,
 	},
 	{
-		.ctl_name =	NET_IPV4_ROUTE_GC_INTERVAL,
-		.procname =	"gc_interval",
-		.data =	&ip_rt_gc_interval,
-		.maxlen =		sizeof(int),
-		.mode =	0644,
-		.proc_handler =&proc_dointvec_jiffies,
-		.strategy =	&sysctl_jiffies,
+		.ctl_name	= NET_IPV4_ROUTE_GC_INTERVAL,
+		.procname	= "gc_interval",
+		.data		= &ip_rt_gc_interval,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec_jiffies,
+		.strategy	= &sysctl_jiffies,
 	},
 	{
-		.ctl_name =	NET_IPV4_ROUTE_REDIRECT_LOAD,
-		.procname =	"redirect_load",
-		.data =	&ip_rt_redirect_load,
-		.maxlen =		sizeof(int),
-		.mode =	0644,
-		.proc_handler =&proc_dointvec,
+		.ctl_name	= NET_IPV4_ROUTE_REDIRECT_LOAD,
+		.procname	= "redirect_load",
+		.data		= &ip_rt_redirect_load,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec,
 	},
 	{
-		.ctl_name =	NET_IPV4_ROUTE_REDIRECT_NUMBER,
-		.procname =	"redirect_number",
-		.data =	&ip_rt_redirect_number,
-		.maxlen =		sizeof(int),
-		.mode =	0644,
-		.proc_handler =&proc_dointvec,
+		.ctl_name	= NET_IPV4_ROUTE_REDIRECT_NUMBER,
+		.procname	= "redirect_number",
+		.data		= &ip_rt_redirect_number,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec,
 	},
 	{
-		.ctl_name =	NET_IPV4_ROUTE_REDIRECT_SILENCE,
-		.procname =	"redirect_silence",
-		.data =	&ip_rt_redirect_silence,
-		.maxlen =		sizeof(int),
-		.mode =	0644,
-		.proc_handler =&proc_dointvec,
+		.ctl_name	= NET_IPV4_ROUTE_REDIRECT_SILENCE,
+		.procname	= "redirect_silence",
+		.data		= &ip_rt_redirect_silence,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec,
 	},
 	{
-		.ctl_name =	NET_IPV4_ROUTE_ERROR_COST,
-		.procname =	"error_cost",
-		.data =	&ip_rt_error_cost,
-		.maxlen =		sizeof(int),
-		.mode =	0644,
-		.proc_handler =&proc_dointvec,
+		.ctl_name	= NET_IPV4_ROUTE_ERROR_COST,
+		.procname	= "error_cost",
+		.data		= &ip_rt_error_cost,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec,
 	},
 	{
-		.ctl_name =	NET_IPV4_ROUTE_ERROR_BURST,
-		.procname =	"error_burst",
-		.data =	&ip_rt_error_burst,
-		.maxlen =		sizeof(int),
-		.mode =	0644,
-		.proc_handler =&proc_dointvec,
+		.ctl_name	= NET_IPV4_ROUTE_ERROR_BURST,
+		.procname	= "error_burst",
+		.data		= &ip_rt_error_burst,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec,
 	},
 	{
-		.ctl_name =	NET_IPV4_ROUTE_GC_ELASTICITY,
-		.procname =	"gc_elasticity",
-		.data =	&ip_rt_gc_elasticity,
-		.maxlen =		sizeof(int),
-		.mode =	0644,
-		.proc_handler =&proc_dointvec,
+		.ctl_name	= NET_IPV4_ROUTE_GC_ELASTICITY,
+		.procname	= "gc_elasticity",
+		.data		= &ip_rt_gc_elasticity,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec,
 	},
 	{
-		.ctl_name =	NET_IPV4_ROUTE_MTU_EXPIRES,
-		.procname =	"mtu_expires",
-		.data =	&ip_rt_mtu_expires,
-		.maxlen =		sizeof(int),
-		.mode =	0644,
-		.proc_handler =&proc_dointvec_jiffies,
-		.strategy =	&sysctl_jiffies,
+		.ctl_name	= NET_IPV4_ROUTE_MTU_EXPIRES,
+		.procname	= "mtu_expires",
+		.data		= &ip_rt_mtu_expires,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec_jiffies,
+		.strategy	= &sysctl_jiffies,
 	},
 	{
-		.ctl_name =	NET_IPV4_ROUTE_MIN_PMTU,
-		.procname =	"min_pmtu",
-		.data =	&ip_rt_min_pmtu,
-		.maxlen =		sizeof(int),
-		.mode =	0644,
-		.proc_handler =&proc_dointvec,
+		.ctl_name	= NET_IPV4_ROUTE_MIN_PMTU,
+		.procname	= "min_pmtu",
+		.data		= &ip_rt_min_pmtu,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec,
 	},
 	{
-		.ctl_name =	NET_IPV4_ROUTE_MIN_ADVMSS,
-		.procname =	"min_adv_mss",
-		.data =	&ip_rt_min_advmss,
-		.maxlen =		sizeof(int),
-		.mode =	0644,
-		.proc_handler =&proc_dointvec,
+		.ctl_name	= NET_IPV4_ROUTE_MIN_ADVMSS,
+		.procname	= "min_adv_mss",
+		.data		= &ip_rt_min_advmss,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec,
 	},
-	 { 0 }
+	{ .ctl_name = 0 }
 };
 #endif
 

net/ipv4/xfrm_policy.c

+#include <linux/config.h>
 #include <net/xfrm.h>
 #include <net/ip.h>
 
@@ -198,6 +199,49 @@ struct xfrm_type *xfrm_get_type(u8 proto)
 	return type;
 }
 
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+static struct xfrm_type *xfrm6_type_map[256];
+static rwlock_t xfrm6_type_lock = RW_LOCK_UNLOCKED;
+
+int xfrm6_register_type(struct xfrm_type *type)
+{
+	int err = 0;
+
+	write_lock(&xfrm6_type_lock);
+	if (xfrm6_type_map[type->proto] == NULL)
+		xfrm6_type_map[type->proto] = type;
+	else
+		err = -EEXIST;
+	write_unlock(&xfrm6_type_lock);
+	return err;
+}
+
+int xfrm6_unregister_type(struct xfrm_type *type)
+{
+	int err = 0;
+
+	write_lock(&xfrm6_type_lock);
+	if (xfrm6_type_map[type->proto] != type)
+		err = -ENOENT;
+	else
+		xfrm6_type_map[type->proto] = NULL;
+	write_unlock(&xfrm6_type_lock);
+	return err;
+}
+
+struct xfrm_type *xfrm6_get_type(u8 proto)
+{
+	struct xfrm_type *type;
+
+	read_lock(&xfrm6_type_lock);
+	type = xfrm6_type_map[proto];
+	if (type && !try_module_get(type->owner))
+		type = NULL;
+	read_unlock(&xfrm6_type_lock);
+	return type;
+}
+#endif /* CONFIG_IPV6 || CONFIG_IPV6_MODULE */
+
 void xfrm_put_type(struct xfrm_type *type)
 {
 	module_put(type->owner);

net/ipv4/xfrm_user.c

 /* xfrm_user.c: User interface to configure xfrm engine.
  *
  * Copyright (C) 2002 David S. Miller (davem@redhat.com)
+ *
+ * Changes
+ *
+ *	Mitsuru KANDA @USAGI       : IPv6 Support 
+ * 	Kazunori MIYAZAWA @USAGI   :
+ * 	Kunihiro Ishiguro          :
+ * 	
  */
 
 #include <linux/module.h>
@@ -17,6 +24,9 @@
 #include <linux/ipsec.h>
 #include <linux/init.h>
 #include <linux/security.h>
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+#include <linux/in6.h>
+#endif
 #include <net/sock.h>
 #include <net/xfrm.h>
 
@@ -63,10 +73,14 @@ static int verify_newsa_info(struct xfrm_usersa_info *p,
 	case AF_INET:
 		break;
 
-	case AF_INET6: /* XXX */
+	case AF_INET6:
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+		break;
+#else
 		err = -EAFNOSUPPORT;
+		goto out;
+#endif
 
-		/* fallthru */
 	default:
 		goto out;
 	};
@@ -171,7 +185,19 @@ static struct xfrm_state *xfrm_state_construct(struct xfrm_usersa_info *p,
 		goto error;
 
 	err = -ENOENT;
-	x->type = xfrm_get_type(x->id.proto);
+	switch (x->props.family) {
+	case AF_INET:
+		x->type = xfrm_get_type(x->id.proto);
+		break;
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+	case AF_INET6:
+		x->type = xfrm6_get_type(x->id.proto);
+		break;
+#endif
+	default:
+		x->type = NULL;
+		break;
+	}
 	if (x->type == NULL)
 		goto error;
 
@@ -206,8 +232,21 @@ static int xfrm_add_sa(struct sk_buff *skb, struct nlmsghdr *nlh, void **xfrma)
 	if (!x)
 		return err;
 
-	x1 = xfrm_state_lookup(x->props.saddr.xfrm4_addr,
-			       x->id.spi, x->id.proto);
+	switch (x->props.family) {
+	case AF_INET:
+		x1 = xfrm_state_lookup(x->props.saddr.xfrm4_addr,
+				       x->id.spi, x->id.proto);
+		break;
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+	case AF_INET6:
+		x1 = xfrm6_state_lookup((struct in6_addr*)x->props.saddr.a6,
+					x->id.spi, x->id.proto);
+		break;
+#endif
+	default:
+		x1 = NULL;
+		break;
+	}
 	if (x1) {
 		xfrm_state_put(x);
 		xfrm_state_put(x1);
@@ -224,7 +263,19 @@ static int xfrm_del_sa(struct sk_buff *skb, struct nlmsghdr *nlh, void **xfrma)
 	struct xfrm_state *x;
 	struct xfrm_usersa_id *p = NLMSG_DATA(nlh);
 
-	x = xfrm_state_lookup(p->saddr.xfrm4_addr, p->spi, p->proto);
+	switch (p->family) {
+	case AF_INET:
+		x = xfrm_state_lookup(p->saddr.xfrm4_addr, p->spi, p->proto);
+		break;
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+	case AF_INET6:
+		x = xfrm6_state_lookup((struct in6_addr*)p->saddr.a6, p->spi, p->proto);
+		break;
+#endif
+	default:
+		x = NULL;
+		break;
+	}
 	if (x == NULL)
 		return -ESRCH;
 
@@ -342,7 +393,19 @@ static int xfrm_get_sa(struct sk_buff *skb, struct nlmsghdr *nlh, void **xfrma)
 	struct sk_buff *resp_skb;
 	int err;
 
-	x = xfrm_state_lookup(p->saddr.xfrm4_addr, p->spi, p->proto);
+	switch (p->family) {
+	case AF_INET:
+		x = xfrm_state_lookup(p->saddr.xfrm4_addr, p->spi, p->proto);
+		break;
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+	case AF_INET6:
+		x = xfrm6_state_lookup((struct in6_addr*)p->saddr.a6, p->spi, p->proto);
+		break;
+#endif
+	default:
+		x = NULL;
+		break;
+	}
 	err = -ESRCH;
 	if (x == NULL)
 		goto out_noput;
@@ -393,9 +456,23 @@ static int xfrm_alloc_userspi(struct sk_buff *skb, struct nlmsghdr *nlh, void **
 	err = verify_userspi_info(p);
 	if (err)
 		goto out_noput;
-	x = xfrm_find_acq(p->info.mode, p->info.reqid, p->info.id.proto,
-			  p->info.sel.daddr.xfrm4_addr,
-			  p->info.sel.saddr.xfrm4_addr, 1);
+	switch (p->info.family) {
+	case AF_INET:
+		x = xfrm_find_acq(p->info.mode, p->info.reqid, p->info.id.proto,
+				  p->info.sel.daddr.xfrm4_addr,
+				  p->info.sel.saddr.xfrm4_addr, 1);
+		break;
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+	case AF_INET6:
+		x = xfrm6_find_acq(p->info.mode, p->info.reqid, p->info.id.proto,
+				   (struct in6_addr*)p->info.sel.daddr.a6,
+				   (struct in6_addr*)p->info.sel.saddr.a6, 1);
+		break;
+#endif
+	default:
+		x = NULL;
+		break;
+	}
 	err = -ENOENT;
 	if (x == NULL)
 		goto out_noput;

net/ipv6/Makefile

@@ -11,5 +11,3 @@ ipv6-objs :=	af_inet6.o ip6_output.o ip6_input.o addrconf.o sit.o \
 		ip6_flowlabel.o ipv6_syms.o
 
 obj-$(CONFIG_NETFILTER)	+= netfilter/
-
-obj-y += xfrm_policy.o

net/ipv6/exthdrs.c

@@ -288,7 +288,7 @@ static int ipv6_routing_header(struct sk_buff **skb_ptr, int nhoff)
 	dst_release(xchg(&skb->dst, NULL));
 	ip6_route_input(skb);
 	if (skb->dst->error) {
-		skb->dst->input(skb);
+		dst_input(skb);
 		return -1;
 	}
 	if (skb->dst->dev->flags&IFF_LOOPBACK) {
@@ -302,7 +302,7 @@ static int ipv6_routing_header(struct sk_buff **skb_ptr, int nhoff)
 		goto looped_back;
 	}
 
-	skb->dst->input(skb);
+	dst_input(skb);
 	return -1;
 }
 

net/ipv6/ip6_input.c

@@ -47,7 +47,7 @@ static inline int ip6_rcv_finish( struct sk_buff *skb)
 	if (skb->dst == NULL)
 		ip6_route_input(skb);
 
-	return skb->dst->input(skb);
+	return dst_input(skb);
 }
 
 int ipv6_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt)
@@ -235,7 +235,7 @@ int ip6_mc_input(struct sk_buff *skb)
 				skb2 = skb;
 			}
 
-			dst->output(skb2);
+			dst_output(skb2);
 		}
 	}
 #endif

net/ipv6/ip6_output.c

@@ -174,7 +174,7 @@ static inline int ip6_maybe_reroute(struct sk_buff *skb)
 		}
 	}
 #endif /* CONFIG_NETFILTER */
-	return skb->dst->output(skb);
+	return dst_output(skb);
 }
 
 /*
@@ -722,7 +722,7 @@ int ip6_call_ra_chain(struct sk_buff *skb, int sel)
 
 static inline int ip6_forward_finish(struct sk_buff *skb)
 {
-	return skb->dst->output(skb);
+	return dst_output(skb);
 }
 
 int ip6_forward(struct sk_buff *skb)

net/ipv6/ipv6_syms.c

@@ -4,7 +4,6 @@
 #include <net/ipv6.h>
 #include <net/addrconf.h>
 #include <net/ip6_route.h>
-#include <net/xfrm.h>
 
 EXPORT_SYMBOL(ipv6_addr_type);
 EXPORT_SYMBOL(icmpv6_send);
@@ -26,6 +25,3 @@ EXPORT_SYMBOL(inet6_getname);
 EXPORT_SYMBOL(inet6_ioctl);
 EXPORT_SYMBOL(ipv6_get_saddr);
 EXPORT_SYMBOL(ipv6_chk_addr);
-EXPORT_SYMBOL(xfrm6_register_type);
-EXPORT_SYMBOL(xfrm6_unregister_type);
-EXPORT_SYMBOL(xfrm6_get_type);

net/ipv6/reassembly.c

@@ -22,6 +22,7 @@
  *
  *      Horst von Brand Add missing #include <linux/string.h>
  *	Alexey Kuznetsov	SMP races, threading, cleanup.
+ *	Patrick McHardy		LRU queue of frag heads for evictor.
  */
 #include <linux/config.h>
 #include <linux/errno.h>
@@ -31,6 +32,7 @@
 #include <linux/sockios.h>
 #include <linux/jiffies.h>
 #include <linux/net.h>
+#include <linux/list.h>
 #include <linux/netdevice.h>
 #include <linux/in6.h>
 #include <linux/ipv6.h>
@@ -67,6 +69,7 @@ struct ip6frag_skb_cb
 struct frag_queue
 {
 	struct frag_queue	*next;
+	struct list_head lru_list;		/* lru list member	*/
 
 	__u32			id;		/* fragment id		*/
 	struct in6_addr		saddr;
@@ -95,6 +98,7 @@ struct frag_queue
 
 static struct frag_queue *ip6_frag_hash[IP6Q_HASHSZ];
 static rwlock_t ip6_frag_lock = RW_LOCK_UNLOCKED;
+static LIST_HEAD(ip6_frag_lru_list);
 int ip6_frag_nqueues = 0;
 
 static __inline__ void __fq_unlink(struct frag_queue *fq)
@@ -102,6 +106,7 @@ static __inline__ void __fq_unlink(struct frag_queue *fq)
 	if(fq->next)
 		fq->next->pprev = fq->pprev;
 	*fq->pprev = fq->next;
+	list_del(&fq->lru_list);
 	ip6_frag_nqueues--;
 }
 
@@ -193,38 +198,30 @@ static __inline__ void fq_kill(struct frag_queue *fq)
 
 static void ip6_evictor(void)
 {
-	int i, progress;
+	struct frag_queue *fq;
+	struct list_head *tmp;
 
-	do {
+	for(;;) {
 		if (atomic_read(&ip6_frag_mem) <= sysctl_ip6frag_low_thresh)
 			return;
-		progress = 0;
-		for (i = 0; i < IP6Q_HASHSZ; i++) {
-			struct frag_queue *fq;
-			if (ip6_frag_hash[i] == NULL)
-				continue;
-
-			read_lock(&ip6_frag_lock);
-			if ((fq = ip6_frag_hash[i]) != NULL) {
-				/* find the oldest queue for this hash bucket */
-				while (fq->next)
-					fq = fq->next;
-				atomic_inc(&fq->refcnt);
-				read_unlock(&ip6_frag_lock);
-
-				spin_lock(&fq->lock);
-				if (!(fq->last_in&COMPLETE))
-					fq_kill(fq);
-				spin_unlock(&fq->lock);
-
-				fq_put(fq);
-				IP6_INC_STATS_BH(Ip6ReasmFails);
-				progress = 1;
-				continue;
-			}
+		read_lock(&ip6_frag_lock);
+		if (list_empty(&ip6_frag_lru_list)) {
 			read_unlock(&ip6_frag_lock);
+			return;
 		}
-	} while (progress);
+		tmp = ip6_frag_lru_list.next;
+		fq = list_entry(tmp, struct frag_queue, lru_list);
+		atomic_inc(&fq->refcnt);
+		read_unlock(&ip6_frag_lock);
+
+		spin_lock(&fq->lock);
+		if (!(fq->last_in&COMPLETE))
+			fq_kill(fq);
+		spin_unlock(&fq->lock);
+
+		fq_put(fq);
+		IP6_INC_STATS_BH(Ip6ReasmFails);
+	}
 }
 
 static void ip6_frag_expire(unsigned long data)
@@ -294,6 +291,8 @@ static struct frag_queue *ip6_frag_intern(unsigned int hash,
 		fq->next->pprev = &fq->next;
 	ip6_frag_hash[hash] = fq;
 	fq->pprev = &ip6_frag_hash[hash];
+	INIT_LIST_HEAD(&fq->lru_list);
+	list_add_tail(&fq->lru_list, &ip6_frag_lru_list);
 	ip6_frag_nqueues++;
 	write_unlock(&ip6_frag_lock);
 	return fq;
@@ -501,6 +500,9 @@ static void ip6_frag_queue(struct frag_queue *fq, struct sk_buff *skb,
 		fq->nhoffset = nhoff;
 		fq->last_in |= FIRST_IN;
 	}
+	write_lock(&ip6_frag_lock);
+	list_move_tail(&fq->lru_list, &ip6_frag_lru_list);
+	write_unlock(&ip6_frag_lock);
 	return;
 
 err:

net/ipv6/xfrm_policy.c

-#include <net/xfrm.h>
-#include <net/ip.h>
-
-static struct xfrm_type *xfrm6_type_map[256];
-static rwlock_t xfrm6_type_lock = RW_LOCK_UNLOCKED;
-
-int xfrm6_register_type(struct xfrm_type *type)
-{
-	int err = 0;
-
-	write_lock(&xfrm6_type_lock);
-	if (xfrm6_type_map[type->proto] == NULL)
-		xfrm6_type_map[type->proto] = type;
-	else
-		err = -EEXIST;
-	write_unlock(&xfrm6_type_lock);
-	return err;
-}
-
-int xfrm6_unregister_type(struct xfrm_type *type)
-{
-	int err = 0;
-
-	write_lock(&xfrm6_type_lock);
-	if (xfrm6_type_map[type->proto] != type)
-		err = -ENOENT;
-	else
-		xfrm6_type_map[type->proto] = NULL;
-	write_unlock(&xfrm6_type_lock);
-	return err;
-}
-
-struct xfrm_type *xfrm6_get_type(u8 proto)
-{
-	struct xfrm_type *type;
-
-	read_lock(&xfrm6_type_lock);
-	type = xfrm6_type_map[proto];
-	if (type && !try_module_get(type->owner))
-		type = NULL;
-	read_unlock(&xfrm6_type_lock);
-	return type;
-}

net/key/af_key.c

@@ -10,6 +10,7 @@
  *		David S. Miller	<davem@redhat.com>
  *		Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
  *		Kunihiro Ishiguro <kunihiro@ipinfusion.com>
+ *		Kazunori MIYAZAWA / USAGI Project <miyazawa@linux-ipv6.org>
  */
 
 #include <linux/config.h>

net/netsyms.c

@@ -328,6 +328,9 @@ EXPORT_SYMBOL(xfrm_policy_list);
 EXPORT_SYMBOL(xfrm6_state_lookup);
 EXPORT_SYMBOL(xfrm6_find_acq);
 EXPORT_SYMBOL(xfrm6_alloc_spi);
+EXPORT_SYMBOL(xfrm6_register_type);
+EXPORT_SYMBOL(xfrm6_unregister_type);
+EXPORT_SYMBOL(xfrm6_get_type);
 #endif
 
 EXPORT_SYMBOL_GPL(xfrm_probe_algs);