Merge http://bk.skbuff.net:20608/linux-2.6-xfrm6tunnel

into nuts.davemloft.net:/disk1/BK/net-2.6

include/net/xfrm.h

@@ -497,10 +497,6 @@ xfrm_selector_match(struct xfrm_selector *sel, struct flowi *fl,
 	return 0;
 }
 
-/* placeholder until xfrm6_tunnel.c is written */
-static inline int xfrm6_tunnel_check_size(struct sk_buff *skb)
-{ return 0; }
-
 /* A struct encoding bundle of transformations to apply to some set of flow.
  *
  * dst->child points to the next element of bundle.
@@ -783,6 +779,12 @@ struct xfrm_tunnel {
 	void (*err_handler)(struct sk_buff *skb, void *info);
 };
 
+struct xfrm6_tunnel {
+	int (*handler)(struct sk_buff **pskb, unsigned int *nhoffp);
+	void (*err_handler)(struct sk_buff *skb, struct inet6_skb_parm *opt,
+			    int type, int code, int offset, __u32 info);
+};
+
 extern void xfrm_init(void);
 extern void xfrm4_init(void);
 extern void xfrm4_fini(void);
@@ -793,6 +795,8 @@ extern void xfrm4_state_init(void);
 extern void xfrm4_state_fini(void);
 extern void xfrm6_state_init(void);
 extern void xfrm6_state_fini(void);
+extern void xfrm6_tunnel_init(void);
+extern void xfrm6_tunnel_fini(void);
 
 extern int xfrm_state_walk(u8 proto, int (*func)(struct xfrm_state *, int, void*), void *);
 extern struct xfrm_state *xfrm_state_alloc(void);
@@ -812,12 +816,18 @@ extern void xfrm_state_flush(u8 proto);
 extern int xfrm_replay_check(struct xfrm_state *x, u32 seq);
 extern void xfrm_replay_advance(struct xfrm_state *x, u32 seq);
 extern int xfrm_check_selectors(struct xfrm_state **x, int n, struct flowi *fl);
-extern int xfrm_check_output(struct xfrm_state *x, struct sk_buff *skb, unsigned short family);
+extern int xfrm_state_check(struct xfrm_state *x, struct sk_buff *skb);
 extern int xfrm4_rcv(struct sk_buff *skb);
 extern int xfrm4_tunnel_register(struct xfrm_tunnel *handler);
 extern int xfrm4_tunnel_deregister(struct xfrm_tunnel *handler);
 extern int xfrm4_tunnel_check_size(struct sk_buff *skb);
 extern int xfrm6_rcv(struct sk_buff **pskb, unsigned int *nhoffp);
+extern int xfrm6_tunnel_register(struct xfrm6_tunnel *handler);
+extern int xfrm6_tunnel_deregister(struct xfrm6_tunnel *handler);
+extern int xfrm6_tunnel_check_size(struct sk_buff *skb);
+extern u32 xfrm6_tunnel_alloc_spi(xfrm_address_t *saddr);
+extern void xfrm6_tunnel_free_spi(xfrm_address_t *saddr);
+extern u32 xfrm6_tunnel_spi_lookup(xfrm_address_t *saddr);
 
 #ifdef CONFIG_XFRM
 extern int xfrm4_rcv_encap(struct sk_buff *skb, __u16 encap_type);

net/ipv4/ah4.c

@@ -74,12 +74,15 @@ static int ah_output(struct sk_buff **pskb)
 	}
 
 	spin_lock_bh(&x->lock);
-	err = xfrm_check_output(x, *pskb, AF_INET);
+	err = xfrm_state_check(x, *pskb);
 	if (err)
 		goto error;
 
 	iph = (*pskb)->nh.iph;
 	if (x->props.mode) {
+		err = xfrm4_tunnel_check_size(*pskb);
+		if (err)
+			goto error;
 		top_iph = (struct iphdr*)skb_push(*pskb, x->props.header_len);
 		top_iph->ihl = 5;
 		top_iph->version = 4;

net/ipv4/esp4.c

@@ -49,19 +49,24 @@ int esp_output(struct sk_buff **pskb)
 	}
 
 	spin_lock_bh(&x->lock);
-	err = xfrm_check_output(x, *pskb, AF_INET);
+	err = xfrm_state_check(x, *pskb);
 	if (err)
 		goto error;
-	err = -ENOMEM;
 
-	/* Strip IP header in transport mode. Save it. */
-	if (!x->props.mode) {
+	if (x->props.mode) {
+		err = xfrm4_tunnel_check_size(*pskb);
+		if (err)
+			goto error;
+	} else {
+		/* Strip IP header in transport mode. Save it. */
 		iph = (*pskb)->nh.iph;
 		memcpy(&tmp_iph, iph, iph->ihl*4);
 		__skb_pull(*pskb, iph->ihl*4);
 	}
 	/* Now skb is pure payload to encrypt */
 
+	err = -ENOMEM;
+
 	/* Round to block size */
 	clen = (*pskb)->len;
 

net/ipv4/ipcomp.c

@@ -164,12 +164,16 @@ static int ipcomp_output(struct sk_buff **pskb)
 	}
 
 	spin_lock_bh(&x->lock);
-	err = xfrm_check_output(x, *pskb, AF_INET);
+	err = xfrm_state_check(x, *pskb);
 	if (err)
 		goto error;
 
-	/* Don't bother compressing */
-	if (!x->props.mode) {
+	if (x->props.mode) {
+		err = xfrm4_tunnel_check_size(*pskb);
+		if (err)
+			goto error;
+	} else {
+		/* Don't bother compressing */
 		iph = (*pskb)->nh.iph;
 		hdr_len = iph->ihl * 4;
 	}

net/ipv6/Makefile

@@ -10,7 +10,8 @@ ipv6-objs :=	af_inet6.o anycast.o ip6_output.o ip6_input.o addrconf.o sit.o \
 		exthdrs.o sysctl_net_ipv6.o datagram.o proc.o \
 		ip6_flowlabel.o ipv6_syms.o
 
-ipv6-$(CONFIG_XFRM) += xfrm6_policy.o xfrm6_state.o xfrm6_input.o
+ipv6-$(CONFIG_XFRM) += xfrm6_policy.o xfrm6_state.o xfrm6_input.o \
+	xfrm6_tunnel.o
 ipv6-objs += $(ipv6-y)
 
 obj-$(CONFIG_INET6_AH) += ah6.o

net/ipv6/ah6.c

@@ -163,11 +163,15 @@ int ah6_output(struct sk_buff **pskb)
 	}
 
 	spin_lock_bh(&x->lock);
-	err = xfrm_check_output(x, *pskb, AF_INET6);
+	err = xfrm_state_check(x, *pskb);
 	if (err)
 		goto error;
 
 	if (x->props.mode) {
+		err = xfrm6_tunnel_check_size(*pskb);
+		if (err)
+			goto error;
+
 		iph = (*pskb)->nh.ipv6h;
 		(*pskb)->nh.ipv6h = (struct ipv6hdr*)skb_push(*pskb, x->props.header_len);
 		(*pskb)->nh.ipv6h->version = 6;

net/ipv6/esp6.c

@@ -65,14 +65,16 @@ int esp6_output(struct sk_buff **pskb)
 	}
 
 	spin_lock_bh(&x->lock);
-	err = xfrm_check_output(x, *pskb, AF_INET6);
+	err = xfrm_state_check(x, *pskb);
 	if (err)
 		goto error;
-	err = -ENOMEM;
-
-	/* Strip IP header in transport mode. Save it. */
 
-	if (!x->props.mode) {
+	if (x->props.mode) {
+		err = xfrm6_tunnel_check_size(*pskb);
+		if (err)
+			goto error;
+	} else {
+		/* Strip IP header in transport mode. Save it. */
 		hdr_len = ip6_find_1stfragopt(*pskb, &prevhdr);
 		nexthdr = *prevhdr;
 		*prevhdr = IPPROTO_ESP;
@@ -86,6 +88,7 @@ int esp6_output(struct sk_buff **pskb)
 	}
 
 	/* Now skb is pure payload to encrypt */
+	err = -ENOMEM;
 
 	/* Round to block size */
 	clen = (*pskb)->len;

net/ipv6/ip6_tunnel.c

@@ -123,7 +123,7 @@ static inline void ip6_tnl_dst_store(struct ip6_tnl *t, struct dst_entry *dst)
  *   else %NULL
  **/
 
-struct ip6_tnl *
+static struct ip6_tnl *
 ip6ip6_tnl_lookup(struct in6_addr *remote, struct in6_addr *local)
 {
 	unsigned h0 = HASH(remote);
@@ -387,8 +387,9 @@ parse_tlv_tnl_enc_lim(struct sk_buff *skb, __u8 * raw)
  *   to the specifications in RFC 2473.
  **/
 
-void ip6ip6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
-		   int type, int code, int offset, __u32 info)
+static void 
+ip6ip6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
+	   int type, int code, int offset, __u32 info)
 {
 	struct ipv6hdr *ipv6h = (struct ipv6hdr *) skb->data;
 	struct ip6_tnl *t;
@@ -496,7 +497,8 @@ void ip6ip6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
  * Return: 0
  **/
 
-int ip6ip6_rcv(struct sk_buff **pskb, unsigned int *nhoffp)
+static int 
+ip6ip6_rcv(struct sk_buff **pskb, unsigned int *nhoffp)
 {
 	struct sk_buff *skb = *pskb;
 	struct ipv6hdr *ipv6h;
@@ -510,6 +512,11 @@ int ip6ip6_rcv(struct sk_buff **pskb, unsigned int *nhoffp)
 	read_lock(&ip6ip6_lock);
 
 	if ((t = ip6ip6_tnl_lookup(&ipv6h->saddr, &ipv6h->daddr)) != NULL) {
+		if (!xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb)) {
+			kfree_skb(skb);
+			return 0;
+		}
+
 		if (!(t->parms.flags & IP6_TNL_F_CAP_RCV)) {
 			t->stat.rx_dropped++;
 			read_unlock(&ip6ip6_lock);
@@ -533,8 +540,7 @@ int ip6ip6_rcv(struct sk_buff **pskb, unsigned int *nhoffp)
 	read_unlock(&ip6ip6_lock);
 	icmpv6_send(skb, ICMPV6_DEST_UNREACH, ICMPV6_ADDR_UNREACH, 0, skb->dev);
 discard:
-	kfree_skb(skb);
-	return 0;
+	return 1;
 }
 
 static inline struct ipv6_txoptions *create_tel(__u8 encap_limit)
@@ -598,7 +604,8 @@ ip6ip6_tnl_addr_conflict(struct ip6_tnl *t, struct ipv6hdr *hdr)
  *   0
  **/
 
-int ip6ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev)
+static int 
+ip6ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev)
 {
 	struct ip6_tnl *t = (struct ip6_tnl *) dev->priv;
 	struct net_device_stats *stats = &t->stat;
@@ -1079,7 +1086,8 @@ ip6ip6_tnl_dev_init(struct net_device *dev)
  * Return: 0
  **/
 
-int ip6ip6_fb_tnl_dev_init(struct net_device *dev)
+static int 
+ip6ip6_fb_tnl_dev_init(struct net_device *dev)
 {
 	struct ip6_tnl *t = dev->priv;
 	ip6ip6_tnl_dev_init_gen(dev);
@@ -1088,10 +1096,9 @@ int ip6ip6_fb_tnl_dev_init(struct net_device *dev)
 	return 0;
 }
 
-static struct inet6_protocol ip6ip6_protocol = {
+static struct xfrm6_tunnel ip6ip6_handler = {
 	.handler = ip6ip6_rcv,
 	.err_handler = ip6ip6_err,
-	.flags = INET6_PROTO_FINAL
 };
 
 /**
@@ -1104,9 +1111,9 @@ static int __init ip6_tunnel_init(void)
 {
 	int  err;
 
-	if ((err = inet6_add_protocol(&ip6ip6_protocol, IPPROTO_IPV6)) < 0) {
-		printk(KERN_ERR "Failed to register IPv6 protocol\n");
-		return err;
+	if (xfrm6_tunnel_register(&ip6ip6_handler) < 0) {
+		printk(KERN_ERR "ip6ip6 init: can't register tunnel\n");
+		return -EAGAIN;
 	}
 	ip6ip6_fb_tnl_dev = alloc_netdev(sizeof(struct ip6_tnl), "ip6tnl0",
 					 ip6ip6_tnl_dev_setup);
@@ -1123,7 +1130,7 @@ static int __init ip6_tunnel_init(void)
 	}
 	return 0;
 fail:
-	inet6_del_protocol(&ip6ip6_protocol, IPPROTO_IPV6);
+	xfrm6_tunnel_deregister(&ip6ip6_handler);
 	return err;
 }
 
@@ -1133,8 +1140,10 @@ static int __init ip6_tunnel_init(void)
 
 static void __exit ip6_tunnel_cleanup(void)
 {
+	if (xfrm6_tunnel_deregister(&ip6ip6_handler) < 0)
+		printk(KERN_INFO "ip6ip6 close: can't deregister tunnel\n");
+
 	unregister_netdev(ip6ip6_fb_tnl_dev);
-	inet6_del_protocol(&ip6ip6_protocol, IPPROTO_IPV6);
 }
 
 module_init(ip6_tunnel_init);

net/ipv6/ipcomp6.c

@@ -140,11 +140,15 @@ static int ipcomp6_output(struct sk_buff **pskb)
 
 	spin_lock_bh(&x->lock);
 
-	err = xfrm_check_output(x, *pskb, AF_INET6);
+	err = xfrm_state_check(x, *pskb);
 	if (err)
 		goto error;
 
 	if (x->props.mode) {
+		err = xfrm6_tunnel_check_size(*pskb);
+		if (err)
+			goto error;
+
 		hdr_len = sizeof(struct ipv6hdr);
 		nexthdr = IPPROTO_IPV6;
 		iph = (*pskb)->nh.ipv6h;
@@ -258,6 +262,66 @@ static void ipcomp6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
 	xfrm_state_put(x);
 }
 
+static struct xfrm_state *ipcomp6_tunnel_create(struct xfrm_state *x)
+{
+	struct xfrm_state *t = NULL;
+
+	t = xfrm_state_alloc();
+	if (!t)
+		goto out;
+
+	t->id.proto = IPPROTO_IPV6;
+	t->id.spi = xfrm6_tunnel_alloc_spi((xfrm_address_t *)&x->props.saddr);
+	memcpy(t->id.daddr.a6, x->id.daddr.a6, sizeof(struct in6_addr));
+	memcpy(&t->sel, &x->sel, sizeof(t->sel));
+	t->props.family = AF_INET6;
+	t->props.mode = 1;
+	memcpy(t->props.saddr.a6, x->props.saddr.a6, sizeof(struct in6_addr));
+
+	t->type = xfrm_get_type(IPPROTO_IPV6, t->props.family);
+	if (t->type == NULL)
+		goto error;
+
+	if (t->type->init_state(t, NULL))
+		goto error;
+
+	t->km.state = XFRM_STATE_VALID;
+	atomic_set(&t->tunnel_users, 1);
+
+out:
+	return t;
+
+error:
+	xfrm_state_put(t);
+	goto out;
+}
+
+static int ipcomp6_tunnel_attach(struct xfrm_state *x)
+{
+	int err = 0;
+	struct xfrm_state *t = NULL;
+	u32 spi;
+
+	spi = xfrm6_tunnel_spi_lookup((xfrm_address_t *)&x->props.saddr);
+	if (spi)
+		t = xfrm_state_lookup((xfrm_address_t *)&x->id.daddr,
+					      spi, IPPROTO_IPV6, AF_INET6);
+	if (!t) {
+		t = ipcomp6_tunnel_create(x);
+		if (!t) {
+			err = -EINVAL;
+			goto out;
+		}
+		xfrm_state_insert(t);
+		xfrm_state_hold(t);
+	}
+	x->tunnel = t;
+	atomic_inc(&t->tunnel_users);
+
+out:
+	return err;
+}
+
 static void ipcomp6_free_data(struct ipcomp_data *ipcd)
 {
 	if (ipcd->tfm)
@@ -271,8 +335,11 @@ static void ipcomp6_destroy(struct xfrm_state *x)
 	struct ipcomp_data *ipcd = x->data;
 	if (!ipcd)
 		return;
+	xfrm_state_delete_tunnel(x);
 	ipcomp6_free_data(ipcd);
 	kfree(ipcd);
+
+	xfrm6_tunnel_free_spi((xfrm_address_t *)&x->props.saddr);
 }
 
 static int ipcomp6_init_state(struct xfrm_state *x, void *args)
@@ -303,6 +370,12 @@ static int ipcomp6_init_state(struct xfrm_state *x, void *args)
 	if (!ipcd->tfm)
 		goto error;
 
+	if (x->props.mode) {
+		err = ipcomp6_tunnel_attach(x);
+		if (err)
+			goto error;
+	}
+
 	calg_desc = xfrm_calg_get_byname(x->calg->alg_name);
 	BUG_ON(!calg_desc);
 	ipcd->threshold = calg_desc->uinfo.comp.threshold;

net/ipv6/xfrm6_policy.c

@@ -277,10 +277,12 @@ void __init xfrm6_init(void)
 {
 	xfrm6_policy_init();
 	xfrm6_state_init();
+	xfrm6_tunnel_init();
 }
 
 void __exit xfrm6_fini(void)
 {
+	xfrm6_tunnel_fini();
 	//xfrm6_input_fini();
 	xfrm6_policy_fini();
 	xfrm6_state_fini();

net/ipv6/xfrm6_tunnel.c

+/*
+ * Copyright (C)2003,2004 USAGI/WIDE Project
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ * 
+ * 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
+ *
+ * Authors	Mitsuru KANDA  <mk@linux-ipv6.org>
+ * 		YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
+ *
+ * Based on net/ipv4/xfrm4_tunnel.c
+ *
+ */
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/xfrm.h>
+#include <linux/list.h>
+#include <net/ip.h>
+#include <net/xfrm.h>
+#include <net/icmp.h>
+#include <net/ipv6.h>
+#include <linux/ipv6.h>
+#include <linux/icmpv6.h>
+
+#ifdef CONFIG_IPV6_XFRM6_TUNNEL_DEBUG
+# define X6TDEBUG	3
+#else
+# define X6TDEBUG	1
+#endif
+
+#define X6TPRINTK(fmt, args...)		printk(fmt, ## args)
+#define X6TNOPRINTK(fmt, args...)	do { ; } while(0)
+
+#if X6TDEBUG >= 1
+# define X6TPRINTK1	X6TPRINTK
+#else
+# define X6TPRINTK1	X6TNOPRINTK
+#endif
+
+#if X6TDEBUG >= 3
+# define X6TPRINTK3	X6TPRINTK
+#else
+# define X6TPRINTK3	X6TNOPRINTK
+#endif
+
+/*
+ * xfrm_tunnel_spi things are for allocating unique id ("spi") 
+ * per xfrm_address_t.
+ */
+struct xfrm6_tunnel_spi {
+	struct hlist_node list_byaddr;
+	struct hlist_node list_byspi;
+	xfrm_address_t addr;
+	u32 spi;
+	atomic_t refcnt;
+#ifdef XFRM6_TUNNEL_SPI_MAGIC
+	u32 magic;
+#endif
+};
+
+#ifdef CONFIG_IPV6_XFRM6_TUNNEL_DEBUG
+# define XFRM6_TUNNEL_SPI_MAGIC 0xdeadbeef
+#endif
+
+static rwlock_t xfrm6_tunnel_spi_lock = RW_LOCK_UNLOCKED;
+
+static u32 xfrm6_tunnel_spi;
+
+#define XFRM6_TUNNEL_SPI_MIN	1
+#define XFRM6_TUNNEL_SPI_MAX	0xffffffff
+
+static kmem_cache_t *xfrm6_tunnel_spi_kmem;
+
+#define XFRM6_TUNNEL_SPI_BYADDR_HSIZE 256
+#define XFRM6_TUNNEL_SPI_BYSPI_HSIZE 256
+
+static struct hlist_head xfrm6_tunnel_spi_byaddr[XFRM6_TUNNEL_SPI_BYADDR_HSIZE];
+static struct hlist_head xfrm6_tunnel_spi_byspi[XFRM6_TUNNEL_SPI_BYSPI_HSIZE];
+
+#ifdef XFRM6_TUNNEL_SPI_MAGIC
+static int x6spi_check_magic(const struct xfrm6_tunnel_spi *x6spi,
+			     const char *name)
+{
+	if (unlikely(x6spi->magic != XFRM6_TUNNEL_SPI_MAGIC)) {
+		X6TPRINTK3(KERN_DEBUG "%s(): x6spi object "
+				      "at %p has corrupted magic %08x "
+				      "(should be %08x)\n",
+			   name, x6spi, x6spi->magic, XFRM6_TUNNEL_SPI_MAGIC);
+		return -1;
+	}
+	return 0;
+}
+#else
+static int inline x6spi_check_magic(const struct xfrm6_tunnel_spi *x6spi,
+				    const char *name)
+{
+	return 0;
+}
+#endif
+
+#define X6SPI_CHECK_MAGIC(x6spi) x6spi_check_magic((x6spi), __FUNCTION__)
+
+
+static unsigned inline xfrm6_tunnel_spi_hash_byaddr(xfrm_address_t *addr)
+{
+	unsigned h;
+
+	X6TPRINTK3(KERN_DEBUG "%s(addr=%p)\n", __FUNCTION__, addr);
+
+	h = addr->a6[0] ^ addr->a6[1] ^ addr->a6[2] ^ addr->a6[3];
+	h ^= h >> 16;
+	h ^= h >> 8;
+	h &= XFRM6_TUNNEL_SPI_BYADDR_HSIZE - 1;
+
+	X6TPRINTK3(KERN_DEBUG "%s() = %u\n", __FUNCTION__, h);
+
+	return h;
+}
+
+static unsigned inline xfrm6_tunnel_spi_hash_byspi(u32 spi)
+{
+	return spi % XFRM6_TUNNEL_SPI_BYSPI_HSIZE;
+}
+
+
+static int xfrm6_tunnel_spi_init(void)
+{
+	int i;
+
+	X6TPRINTK3(KERN_DEBUG "%s()\n", __FUNCTION__);
+
+	xfrm6_tunnel_spi = 0;
+	xfrm6_tunnel_spi_kmem = kmem_cache_create("xfrm6_tunnel_spi",
+						  sizeof(struct xfrm6_tunnel_spi),
+						  0, SLAB_HWCACHE_ALIGN,
+						  NULL, NULL);
+	if (!xfrm6_tunnel_spi_kmem) {
+		X6TPRINTK1(KERN_ERR
+			   "%s(): failed to allocate xfrm6_tunnel_spi_kmem\n",
+		           __FUNCTION__);
+		return -ENOMEM;
+	}
+
+	for (i = 0; i < XFRM6_TUNNEL_SPI_BYADDR_HSIZE; i++)
+		INIT_HLIST_HEAD(&xfrm6_tunnel_spi_byaddr[i]);
+	for (i = 0; i < XFRM6_TUNNEL_SPI_BYSPI_HSIZE; i++)
+		INIT_HLIST_HEAD(&xfrm6_tunnel_spi_byspi[i]);
+	return 0;
+}
+
+static void xfrm6_tunnel_spi_fini(void)
+{
+	int i;
+
+	X6TPRINTK3(KERN_DEBUG "%s()\n", __FUNCTION__);
+
+	for (i = 0; i < XFRM6_TUNNEL_SPI_BYADDR_HSIZE; i++) {
+		if (!hlist_empty(&xfrm6_tunnel_spi_byaddr[i]))
+			goto err;
+	}
+	for (i = 0; i < XFRM6_TUNNEL_SPI_BYSPI_HSIZE; i++) {
+		if (!hlist_empty(&xfrm6_tunnel_spi_byspi[i]))
+			goto err;
+	}
+	kmem_cache_destroy(xfrm6_tunnel_spi_kmem);
+	xfrm6_tunnel_spi_kmem = NULL;
+	return;
+err:
+	X6TPRINTK1(KERN_ERR "%s(): table is not empty\n", __FUNCTION__);
+	return;
+}
+
+static struct xfrm6_tunnel_spi *__xfrm6_tunnel_spi_lookup(xfrm_address_t *saddr)
+{
+	struct xfrm6_tunnel_spi *x6spi;
+	struct hlist_node *pos;
+
+	X6TPRINTK3(KERN_DEBUG "%s(saddr=%p)\n", __FUNCTION__, saddr);
+
+	hlist_for_each_entry(x6spi, pos,
+			     &xfrm6_tunnel_spi_byaddr[xfrm6_tunnel_spi_hash_byaddr(saddr)],
+			     list_byaddr) {
+		if (memcmp(&x6spi->addr, saddr, sizeof(x6spi->addr)) == 0) {
+			X6SPI_CHECK_MAGIC(x6spi);
+			X6TPRINTK3(KERN_DEBUG "%s() = %p(%u)\n", __FUNCTION__, x6spi, x6spi->spi);
+			return x6spi;
+		}
+	}
+
+	X6TPRINTK3(KERN_DEBUG "%s() = NULL(0)\n", __FUNCTION__);
+	return NULL;
+}
+
+u32 xfrm6_tunnel_spi_lookup(xfrm_address_t *saddr)
+{
+	struct xfrm6_tunnel_spi *x6spi;
+	u32 spi;
+
+	X6TPRINTK3(KERN_DEBUG "%s(saddr=%p)\n", __FUNCTION__, saddr);
+
+	read_lock_bh(&xfrm6_tunnel_spi_lock);
+	x6spi = __xfrm6_tunnel_spi_lookup(saddr);
+	spi = x6spi ? x6spi->spi : 0;
+	read_unlock_bh(&xfrm6_tunnel_spi_lock);
+	return spi;
+}
+
+static u32 __xfrm6_tunnel_alloc_spi(xfrm_address_t *saddr)
+{
+	u32 spi;
+	struct xfrm6_tunnel_spi *x6spi;
+	struct hlist_node *pos;
+	unsigned index;
+
+	X6TPRINTK3(KERN_DEBUG "%s(saddr=%p)\n", __FUNCTION__, saddr);
+
+	if (xfrm6_tunnel_spi < XFRM6_TUNNEL_SPI_MIN ||
+	    xfrm6_tunnel_spi >= XFRM6_TUNNEL_SPI_MAX)
+		xfrm6_tunnel_spi = XFRM6_TUNNEL_SPI_MIN;
+	else
+		xfrm6_tunnel_spi++;
+
+	for (spi = xfrm6_tunnel_spi; spi <= XFRM6_TUNNEL_SPI_MAX; spi++) {
+		index = xfrm6_tunnel_spi_hash_byspi(spi);
+		hlist_for_each_entry(x6spi, pos, 
+				     &xfrm6_tunnel_spi_byspi[index], 
+				     list_byspi) {
+			if (x6spi->spi == spi)
+				goto try_next_1;
+		}
+		xfrm6_tunnel_spi = spi;
+		goto alloc_spi;
+try_next_1:;
+	}
+	for (spi = XFRM6_TUNNEL_SPI_MIN; spi < xfrm6_tunnel_spi; spi++) {
+		index = xfrm6_tunnel_spi_hash_byspi(spi);
+		hlist_for_each_entry(x6spi, pos, 
+				     &xfrm6_tunnel_spi_byspi[index], 
+				     list_byspi) {
+			if (x6spi->spi == spi)
+				goto try_next_2;
+		}
+		xfrm6_tunnel_spi = spi;
+		goto alloc_spi;
+try_next_2:;
+	}
+	spi = 0;
+	goto out;
+alloc_spi:
+	X6TPRINTK3(KERN_DEBUG "%s(): allocate new spi for "
+			      "%04x:%04x:%04x:%04x:%04x:%04x:%04x:%04x\n", 
+			      __FUNCTION__, 
+			      NIP6(*(struct in6_addr *)saddr));
+	x6spi = kmem_cache_alloc(xfrm6_tunnel_spi_kmem, SLAB_ATOMIC);
+	if (!x6spi) {
+		X6TPRINTK1(KERN_ERR "%s(): kmem_cache_alloc() failed\n", 
+			   __FUNCTION__);
+		goto out;
+	}
+#ifdef XFRM6_TUNNEL_SPI_MAGIC
+	x6spi->magic = XFRM6_TUNNEL_SPI_MAGIC;
+#endif
+	memcpy(&x6spi->addr, saddr, sizeof(x6spi->addr));
+	x6spi->spi = spi;
+	atomic_set(&x6spi->refcnt, 1);
+
+	hlist_add_head(&x6spi->list_byspi, &xfrm6_tunnel_spi_byspi[index]);
+
+	index = xfrm6_tunnel_spi_hash_byaddr(saddr);
+	hlist_add_head(&x6spi->list_byaddr, &xfrm6_tunnel_spi_byaddr[index]);
+	X6SPI_CHECK_MAGIC(x6spi);
+out:
+	X6TPRINTK3(KERN_DEBUG "%s() = %u\n", __FUNCTION__, spi);
+	return spi;
+}
+
+u32 xfrm6_tunnel_alloc_spi(xfrm_address_t *saddr)
+{
+	struct xfrm6_tunnel_spi *x6spi;
+	u32 spi;
+
+	X6TPRINTK3(KERN_DEBUG "%s(saddr=%p)\n", __FUNCTION__, saddr);
+
+	write_lock_bh(&xfrm6_tunnel_spi_lock);
+	x6spi = __xfrm6_tunnel_spi_lookup(saddr);
+	if (x6spi) {
+		atomic_inc(&x6spi->refcnt);
+		spi = x6spi->spi;
+	} else
+		spi = __xfrm6_tunnel_alloc_spi(saddr);
+	write_unlock_bh(&xfrm6_tunnel_spi_lock);
+
+	X6TPRINTK3(KERN_DEBUG "%s() = %u\n", __FUNCTION__, spi);
+
+	return spi;
+}
+
+void xfrm6_tunnel_free_spi(xfrm_address_t *saddr)
+{
+	struct xfrm6_tunnel_spi *x6spi;
+	struct hlist_node *pos, *n;
+
+	X6TPRINTK3(KERN_DEBUG "%s(saddr=%p)\n", __FUNCTION__, saddr);
+
+	write_lock_bh(&xfrm6_tunnel_spi_lock);
+
+	hlist_for_each_entry_safe(x6spi, pos, n, 
+				  &xfrm6_tunnel_spi_byaddr[xfrm6_tunnel_spi_hash_byaddr(saddr)],
+				  list_byaddr)
+	{
+		if (memcmp(&x6spi->addr, saddr, sizeof(x6spi->addr)) == 0) {
+			X6TPRINTK3(KERN_DEBUG "%s(): x6spi object "
+					      "for %04x:%04x:%04x:%04x:%04x:%04x:%04x:%04x "
+					      "found at %p\n",
+				   __FUNCTION__, 
+				   NIP6(*(struct in6_addr *)saddr),
+				   x6spi);
+			X6SPI_CHECK_MAGIC(x6spi);
+			if (atomic_dec_and_test(&x6spi->refcnt)) {
+				hlist_del(&x6spi->list_byaddr);
+				hlist_del(&x6spi->list_byspi);
+				kmem_cache_free(xfrm6_tunnel_spi_kmem, x6spi);
+				break;
+			}
+		}
+	}
+	write_unlock_bh(&xfrm6_tunnel_spi_lock);
+}
+
+
+int xfrm6_tunnel_check_size(struct sk_buff *skb)
+{
+	int mtu, ret = 0;
+	struct dst_entry *dst = skb->dst;
+
+	mtu = dst_pmtu(dst) - sizeof(struct ipv6hdr);
+	if (mtu < IPV6_MIN_MTU)
+		mtu = IPV6_MIN_MTU;
+
+	if (skb->len > mtu) {
+		icmpv6_send(skb, ICMPV6_PKT_TOOBIG, 0, mtu, skb->dev);
+		ret = -EMSGSIZE;
+	}
+
+	return ret;
+}
+
+static int xfrm6_tunnel_output(struct sk_buff **pskb)
+{
+	struct sk_buff *skb = *pskb;
+	struct dst_entry *dst = skb->dst;
+	struct xfrm_state *x = dst->xfrm;
+	struct ipv6hdr *iph, *top_iph;
+	int err;
+
+	if ((err = xfrm6_tunnel_check_size(skb)) != 0)
+		goto error_nolock;
+
+	iph = skb->nh.ipv6h;
+
+	top_iph = (struct ipv6hdr *)skb_push(skb, x->props.header_len);
+	top_iph->version = 6;
+	top_iph->priority = iph->priority;
+	top_iph->flow_lbl[0] = iph->flow_lbl[0];
+	top_iph->flow_lbl[1] = iph->flow_lbl[1];
+	top_iph->flow_lbl[2] = iph->flow_lbl[2];
+	top_iph->nexthdr = IPPROTO_IPV6; 
+	top_iph->payload_len = htons(skb->len - sizeof(struct ipv6hdr));
+	top_iph->hop_limit = iph->hop_limit;
+	memcpy(&top_iph->saddr, (struct in6_addr *)&x->props.saddr, sizeof(struct in6_addr));
+	memcpy(&top_iph->daddr, (struct in6_addr *)&x->id.daddr, sizeof(struct in6_addr));
+	skb->nh.raw = skb->data;
+	skb->h.raw = skb->nh.raw + sizeof(struct ipv6hdr);
+
+	x->curlft.bytes += skb->len;
+	x->curlft.packets++;
+
+	spin_unlock_bh(&x->lock);
+
+	if ((skb->dst = dst_pop(dst)) == NULL) { 
+		kfree_skb(skb);
+		err = -EHOSTUNREACH;
+		goto error_nolock;
+	}
+
+	return NET_XMIT_BYPASS;
+
+error_nolock:
+	kfree_skb(skb);
+	return err;
+}
+
+static int xfrm6_tunnel_input(struct xfrm_state *x, struct xfrm_decap_state *decap, struct sk_buff *skb)
+{
+	if (!pskb_may_pull(skb, sizeof(struct ipv6hdr))) 
+		return -EINVAL;
+
+	skb->mac.raw = skb->nh.raw;
+	skb->nh.raw = skb->data;
+	dst_release(skb->dst);
+	skb->dst = NULL;
+	skb->protocol = htons(ETH_P_IPV6);
+	skb->pkt_type = PACKET_HOST;
+	netif_rx(skb);
+
+	return 0;
+}
+
+static struct xfrm6_tunnel *xfrm6_tunnel_handler;
+static DECLARE_MUTEX(xfrm6_tunnel_sem);
+
+int xfrm6_tunnel_register(struct xfrm6_tunnel *handler)
+{
+	int ret;
+
+	down(&xfrm6_tunnel_sem);
+	ret = 0;
+	if (xfrm6_tunnel_handler != NULL)
+		ret = -EINVAL;
+	if (!ret)
+		xfrm6_tunnel_handler = handler;
+	up(&xfrm6_tunnel_sem);
+
+	return ret;
+}
+
+int xfrm6_tunnel_deregister(struct xfrm6_tunnel *handler)
+{
+	int ret;
+
+	down(&xfrm6_tunnel_sem);
+	ret = 0;
+	if (xfrm6_tunnel_handler != handler)
+		ret = -EINVAL;
+	if (!ret)
+		xfrm6_tunnel_handler = NULL;
+	up(&xfrm6_tunnel_sem);
+
+	synchronize_net();
+
+	return ret;
+}
+
+static int xfrm6_tunnel_rcv(struct sk_buff **pskb, unsigned int *nhoffp)
+{
+	struct sk_buff *skb = *pskb;
+	struct xfrm6_tunnel *handler = xfrm6_tunnel_handler;
+	struct xfrm_state *x = NULL;
+	struct ipv6hdr *iph = skb->nh.ipv6h;
+	int err = 0;
+	u32 spi;
+
+	/* device-like_ip6ip6_handler() */
+	if (handler) {
+		err = handler->handler(pskb, nhoffp);
+		if (!err)
+			goto out;
+	}
+
+	spi = xfrm6_tunnel_spi_lookup((xfrm_address_t *)&iph->saddr);
+	x = xfrm_state_lookup((xfrm_address_t *)&iph->daddr, 
+			spi,
+			IPPROTO_IPV6, AF_INET6);
+
+	if (!x)
+		goto drop;
+
+	spin_lock(&x->lock);
+
+	if (unlikely(x->km.state != XFRM_STATE_VALID))
+		goto drop_unlock;
+
+	err = xfrm6_tunnel_input(x, NULL, skb);
+	if (err)
+		goto drop_unlock;
+
+	x->curlft.bytes += skb->len;
+	x->curlft.packets++; 
+	spin_unlock(&x->lock); 
+	xfrm_state_put(x); 
+
+out:
+	return 0;
+
+drop_unlock:
+	spin_unlock(&x->lock);
+	xfrm_state_put(x);
+drop:
+	kfree_skb(skb);
+	return -1;
+}
+
+static void xfrm6_tunnel_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
+			     int type, int code, int offset, __u32 info)
+{
+	struct xfrm6_tunnel *handler = xfrm6_tunnel_handler;
+
+	/* call here first for device-like ip6ip6 err handling */
+	if (handler) {
+		handler->err_handler(skb, opt, type, code, offset, info);
+		return;
+	}
+
+	/* xfrm6_tunnel native err handling */
+	switch (type) {
+	case ICMPV6_DEST_UNREACH: 
+		switch (code) {
+		case ICMPV6_NOROUTE: 
+		case ICMPV6_ADM_PROHIBITED:
+		case ICMPV6_NOT_NEIGHBOUR:
+		case ICMPV6_ADDR_UNREACH:
+		case ICMPV6_PORT_UNREACH:
+		default:
+			X6TPRINTK3(KERN_DEBUG
+				   "xfrm6_tunnel: Destination Unreach.\n");
+			break;
+		}
+		break;
+	case ICMPV6_PKT_TOOBIG:
+			X6TPRINTK3(KERN_DEBUG 
+				   "xfrm6_tunnel: Packet Too Big.\n");
+		break;
+	case ICMPV6_TIME_EXCEED:
+		switch (code) {
+		case ICMPV6_EXC_HOPLIMIT:
+			X6TPRINTK3(KERN_DEBUG
+				   "xfrm6_tunnel: Too small Hoplimit.\n");
+			break;
+		case ICMPV6_EXC_FRAGTIME:
+		default: 
+			break;
+		}
+		break;
+	case ICMPV6_PARAMPROB:
+		switch (code) {
+		case ICMPV6_HDR_FIELD: break;
+		case ICMPV6_UNK_NEXTHDR: break;
+		case ICMPV6_UNK_OPTION: break;
+		}
+		break;
+	default:
+		break;
+	}
+	return;
+}
+
+static int xfrm6_tunnel_init_state(struct xfrm_state *x, void *args)
+{
+	if (!x->props.mode)
+		return -EINVAL;
+
+	x->props.header_len = sizeof(struct ipv6hdr);
+
+	return 0;
+}
+
+static void xfrm6_tunnel_destroy(struct xfrm_state *x)
+{
+	xfrm6_tunnel_free_spi((xfrm_address_t *)&x->props.saddr);
+}
+
+static struct xfrm_type xfrm6_tunnel_type = {
+	.description	= "IP6IP6",
+	.owner          = THIS_MODULE,
+	.proto		= IPPROTO_IPV6,
+	.init_state	= xfrm6_tunnel_init_state,
+	.destructor	= xfrm6_tunnel_destroy,
+	.input		= xfrm6_tunnel_input,
+	.output		= xfrm6_tunnel_output,
+};
+
+static struct inet6_protocol xfrm6_tunnel_protocol = {
+	.handler	= xfrm6_tunnel_rcv,
+	.err_handler	= xfrm6_tunnel_err, 
+	.flags          = INET6_PROTO_NOPOLICY|INET6_PROTO_FINAL,
+};
+
+void __init xfrm6_tunnel_init(void)
+{
+	X6TPRINTK3(KERN_DEBUG "%s()\n", __FUNCTION__);
+
+	if (xfrm_register_type(&xfrm6_tunnel_type, AF_INET6) < 0) {
+		X6TPRINTK1(KERN_ERR
+			   "xfrm6_tunnel init: can't add xfrm type\n");
+		return;
+	}
+	if (inet6_add_protocol(&xfrm6_tunnel_protocol, IPPROTO_IPV6) < 0) {
+		X6TPRINTK1(KERN_ERR
+			   "xfrm6_tunnel init(): can't add protocol\n");
+		xfrm_unregister_type(&xfrm6_tunnel_type, AF_INET6);
+		return;
+	}
+	if (xfrm6_tunnel_spi_init() < 0) {
+		X6TPRINTK1(KERN_ERR
+			   "xfrm6_tunnel init: failed to initialize spi\n");
+		inet6_del_protocol(&xfrm6_tunnel_protocol, IPPROTO_IPV6);
+		xfrm_unregister_type(&xfrm6_tunnel_type, AF_INET6);
+		return;
+	}
+}
+
+void __exit xfrm6_tunnel_fini(void)
+{
+	X6TPRINTK3(KERN_DEBUG "%s()\n", __FUNCTION__);
+
+	xfrm6_tunnel_spi_fini();
+	if (inet6_del_protocol(&xfrm6_tunnel_protocol, IPPROTO_IPV6) < 0)
+		X6TPRINTK1(KERN_ERR 
+			   "xfrm6_tunnel close: can't remove protocol\n");
+	if (xfrm_unregister_type(&xfrm6_tunnel_type, AF_INET6) < 0)
+		X6TPRINTK1(KERN_ERR
+			   "xfrm6_tunnel close: can't remove xfrm type\n");
+}

net/xfrm/Makefile

@@ -2,7 +2,7 @@
 # Makefile for the XFRM subsystem.
 #
 
-obj-$(CONFIG_XFRM) := xfrm_policy.o xfrm_state.o xfrm_input.o xfrm_algo.o xfrm_output.o \
+obj-$(CONFIG_XFRM) := xfrm_policy.o xfrm_state.o xfrm_input.o xfrm_algo.o \
 	xfrm_export.o
 obj-$(CONFIG_XFRM_USER) += xfrm_user.o
 

net/xfrm/xfrm_export.c

@@ -18,6 +18,7 @@ EXPORT_SYMBOL(xfrm_state_add);
 EXPORT_SYMBOL(xfrm_state_update);
 EXPORT_SYMBOL(xfrm_state_check_expire);
 EXPORT_SYMBOL(xfrm_state_check_space);
+EXPORT_SYMBOL(xfrm_state_check);
 EXPORT_SYMBOL(xfrm_state_lookup);
 EXPORT_SYMBOL(xfrm_state_register_afinfo);
 EXPORT_SYMBOL(xfrm_state_unregister_afinfo);
@@ -27,7 +28,6 @@ EXPORT_SYMBOL(xfrm_state_delete_tunnel);
 EXPORT_SYMBOL(xfrm_replay_check);
 EXPORT_SYMBOL(xfrm_replay_advance);
 EXPORT_SYMBOL(xfrm_check_selectors);
-EXPORT_SYMBOL(xfrm_check_output);
 EXPORT_SYMBOL(__secpath_destroy);
 EXPORT_SYMBOL(secpath_dup);
 EXPORT_SYMBOL(xfrm_get_acqseq);

net/xfrm/xfrm_output.c

-/* 
- * generic xfrm output routines
- *
- * Copyright (c) 2003 James Morris <jmorris@intercode.com.au>
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License as published by the Free
- * Software Foundation; either version 2 of the License, or (at your option) 
- * any later version.
- */
-#include <linux/config.h>
-#include <linux/kernel.h>
-#include <linux/skbuff.h>
-#include <net/xfrm.h>
-
-int xfrm_check_output(struct xfrm_state *x,
-                      struct sk_buff *skb, unsigned short family)
-{
-	int err;
-	
-	err = xfrm_state_check_expire(x);
-	if (err)
-		goto out;
-		
-	if (x->props.mode) {
-		switch (family) {
-		case AF_INET:
-			err = xfrm4_tunnel_check_size(skb);
-			break;
-			
-		case AF_INET6:
-			err = xfrm6_tunnel_check_size(skb);
-			break;
-			
-		default:
-			err = -EINVAL;
-		}
-		
-		if (err)
-			goto out;
-	}
-
-	err = xfrm_state_check_space(x, skb);
-out:
-	return err;
-}

net/xfrm/xfrm_state.c

@@ -531,6 +531,16 @@ int xfrm_state_check_space(struct xfrm_state *x, struct sk_buff *skb)
 	return 0;
 }
 
+int xfrm_state_check(struct xfrm_state *x, struct sk_buff *skb)
+{
+	int err = xfrm_state_check_expire(x);
+	if (err < 0)
+		goto err;
+	err = xfrm_state_check_space(x, skb);
+err:
+	return err;
+}
+
 struct xfrm_state *
 xfrm_state_lookup(xfrm_address_t *daddr, u32 spi, u8 proto,
 		  unsigned short family)