Merge master.kernel.org:/home/davem/BK/net-2.5

into home.transmeta.com:/home/torvalds/v2.5/linux

drivers/net/tg3.h

@@ -1528,7 +1528,7 @@ struct tg3_internal_buffer_desc {
 	u32				__cookie3;
 };
 
-#define TG3_HW_STATUS_SIZE		0x80
+#define TG3_HW_STATUS_SIZE		0x50
 struct tg3_hw_status {
 	u32				status;
 #define SD_STATUS_UPDATED		0x00000001

include/linux/netlink.h

@@ -7,6 +7,7 @@
 #define NETLINK_FIREWALL	3	/* Firewalling hook				*/
 #define NETLINK_TCPDIAG		4	/* TCP socket monitoring			*/
 #define NETLINK_NFLOG		5	/* netfilter/iptables ULOG */
+#define NETLINK_XFRM		6	/* ipsec */
 #define NETLINK_ARPD		8
 #define NETLINK_ROUTE6		11	/* af_inet6 route comm channel */
 #define NETLINK_IP6_FW		13
@@ -86,6 +87,8 @@ struct nlmsgerr
 
 #ifdef __KERNEL__
 
+#include <linux/capability.h>
+
 struct netlink_skb_parms
 {
 	struct ucred		creds;		/* Skb credentials	*/

include/linux/xfrm.h

+#ifndef _LINUX_XFRM_H
+#define _LINUX_XFRM_H
+
+#include <linux/types.h>
+
+/* All of the structures in this file may not change size as they are
+ * passed into the kernel from userspace via netlink sockets.
+ */
+
+/* Structure to encapsulate addresses. I do not want to use
+ * "standard" structure. My apologies.
+ */
+typedef union
+{
+	struct {
+		__u32	addr;
+		__u32	mask;	/* Use unused bits to cache mask. */
+	} a4;
+#define xfrm4_addr a4.addr
+#define xfrm4_mask a4.mask
+	__u32		a6[4];
+} xfrm_address_t;
+
+/* Ident of a specific xfrm_state. It is used on input to lookup
+ * the state by (spi,daddr,ah/esp) or to store information about
+ * spi, protocol and tunnel address on output.
+ */
+struct xfrm_id
+{
+	xfrm_address_t	daddr;
+	__u32		spi;
+	__u8		proto;
+};
+
+/* Selector, used as selector both on policy rules (SPD) and SAs. */
+
+struct xfrm_selector
+{
+	xfrm_address_t	daddr;
+	xfrm_address_t	saddr;
+	__u16	dport;
+	__u16	dport_mask;
+	__u16	sport;
+	__u16	sport_mask;
+	__u8	prefixlen_d;
+	__u8	prefixlen_s;
+	__u8	proto;
+	int	ifindex;
+	uid_t	user;
+	void	*owner;
+};
+
+#define XFRM_INF (~(u64)0)
+
+struct xfrm_lifetime_cfg
+{
+	__u64	soft_byte_limit;
+	__u64	hard_byte_limit;
+	__u64	soft_packet_limit;
+	__u64	hard_packet_limit;
+	__u64	soft_add_expires_seconds;
+	__u64	hard_add_expires_seconds;
+	__u64	soft_use_expires_seconds;
+	__u64	hard_use_expires_seconds;
+};
+
+struct xfrm_lifetime_cur
+{
+	__u64	bytes;
+	__u64	packets;
+	__u64	add_time;
+	__u64	use_time;
+};
+
+struct xfrm_replay_state
+{
+	__u32	oseq;
+	__u32	seq;
+	__u32	bitmap;
+};
+
+struct xfrm_algo {
+	char	alg_name[64];
+	int	alg_key_len;    /* in bits */
+	char	alg_key[0];
+};
+
+struct xfrm_stats {
+	__u32	replay_window;
+	__u32	replay;
+	__u32	integrity_failed;
+};
+
+/* Netlink configuration messages.  */
+#define XFRM_MSG_BASE		0x10
+
+#define XFRM_MSG_NEWSA		(RTM_BASE + 0)
+#define XFRM_MSG_DELSA		(RTM_BASE + 1)
+#define XFRM_MSG_GETSA		(RTM_BASE + 2)
+
+#define XFRM_MSG_NEWPOLICY	(RTM_BASE + 3)
+#define XFRM_MSG_DELPOLICY	(RTM_BASE + 4)
+#define XFRM_MSG_GETPOLICY	(RTM_BASE + 5)
+
+#define XFRM_MSG_ALLOCSPI	(RTM_BASE + 6)
+#define XFRM_MSG_ACQUIRE	(RTM_BASE + 7)
+
+#define XFRM_MSG_MAX		(XFRM_MSG_ACQUIRE+1)
+
+struct xfrm_user_tmpl {
+	struct xfrm_id		id;
+	xfrm_address_t		saddr;
+	__u16			reqid;
+	__u8			mode;
+	__u8			share;
+	__u32			aalgos;
+	__u32			ealgos;
+	__u32			calgos;
+};
+
+/* Netlink message attributes.  */
+enum xfrm_attr_type_t {
+	XFRMA_UNSPEC,
+	XFRMA_ALG_AUTH,		/* struct xfrm_algo */
+	XFRMA_ALG_CRYPT,	/* struct xfrm_algo */
+	XFRMA_ALG_COMP,		/* struct xfrm_algo */
+	XFRMA_TMPL,		/* 1 or more struct xfrm_user_tmpl */
+
+#define XFRMA_MAX XFRMA_TMPL
+};
+
+struct xfrm_usersa_info {
+	struct xfrm_selector		sel;
+	struct xfrm_id			id;
+	struct xfrm_lifetime_cfg	lft;
+	struct xfrm_lifetime_cur	curlft;
+	struct xfrm_stats		stats;
+	__u16				family;
+	__u16				reqid;
+	__u8				sa_type;
+	__u8				mode; /* 0=transport,1=tunnel */
+	__u8				replay_window;
+};
+
+struct xfrm_usersa_id {
+	xfrm_address_t			saddr;
+	__u32				spi;
+	__u8				proto;
+};
+
+struct xfrm_userpolicy_info {
+	struct xfrm_selector		sel;
+	struct xfrm_id			id;
+	struct xfrm_lifetime_cfg	lft;
+	struct xfrm_lifetime_cur	curlft;
+	__u32				index;
+	__u16				family;
+	__u8				dir;
+	__u8				action;
+};
+
+struct xfrm_userpolicy_id {
+	struct xfrm_selector		sel;
+	__u32				index;
+	__u8				dir;
+};
+
+#endif /* _LINUX_XFRM_H */

include/net/route.h

@@ -140,7 +140,9 @@ static inline char rt_tos2priority(u8 tos)
 	return ip_tos2prio[IPTOS_TOS(tos)>>1];
 }
 
-static inline int ip_route_connect(struct rtable **rp, u32 dst, u32 src, u32 tos, int oif, u8 protocol, u16 sport, u16 dport)
+static inline int ip_route_connect(struct rtable **rp, u32 dst,
+				   u32 src, u32 tos, int oif, u8 protocol,
+				   u16 sport, u16 dport, struct sock *sk)
 {
 	struct flowi fl = { .oif = oif,
 			    .nl_u = { .ip4_u = { .daddr = dst,
@@ -161,10 +163,11 @@ static inline int ip_route_connect(struct rtable **rp, u32 dst, u32 src, u32 tos
 		ip_rt_put(*rp);
 		*rp = NULL;
 	}
-	return ip_route_output_key(rp, &fl);
+	return ip_route_output_flow(rp, &fl, sk, 0);
 }
 
-static inline int ip_route_newports(struct rtable **rp, u16 sport, u16 dport)
+static inline int ip_route_newports(struct rtable **rp, u16 sport, u16 dport,
+				    struct sock *sk)
 {
 	if (sport != (*rp)->fl.uli_u.ports.sport ||
 	    dport != (*rp)->fl.uli_u.ports.dport) {
@@ -175,7 +178,7 @@ static inline int ip_route_newports(struct rtable **rp, u16 sport, u16 dport)
 		fl.uli_u.ports.dport = dport;
 		ip_rt_put(*rp);
 		*rp = NULL;
-		return ip_route_output_key(rp, &fl);
+		return ip_route_output_flow(rp, &fl, sk, 0);
 	}
 	return 0;
 }

include/net/xfrm.h

-#include <linux/types.h>
+#include <linux/xfrm.h>
 #include <linux/spinlock.h>
 #include <linux/list.h>
 #include <linux/skbuff.h>
@@ -8,6 +8,8 @@
 #include <net/dst.h>
 #include <net/route.h>
 
+extern struct semaphore xfrm_cfg_sem;
+
 /* Organization of SPD aka "XFRM rules"
    ------------------------------------
 
@@ -69,84 +71,6 @@
       metrics. Plus, it will be made via sk->dst_cache. Solved.
  */
 
-/* Structure to encapsulate addresses. I do not want to use
- * "standard" structure. My apologies. */
-
-typedef union
-{
-	struct {
-		u32	addr;
-		u32	mask;	/* Use unused bits to cache mask. */
-	} a4;
-#define xfrm4_addr a4.addr
-#define xfrm4_mask a4.mask
-	u32		a6[4];
-} xfrm_address_t;
-
-/* Ident of a specific xfrm_state. It is used on input to lookup
- * the state by (spi,daddr,ah/esp) or to store information about
- * spi, protocol and tunnel address on output. */
-
-struct xfrm_id
-{
-	xfrm_address_t	daddr;
-	__u32		spi;
-	__u8		proto;
-};
-
-/* Selector, used as selector both on policy rules (SPD) and SAs. */
-
-struct xfrm_selector
-{
-	xfrm_address_t	daddr;
-	xfrm_address_t	saddr;
-	__u16	dport;
-	__u16	dport_mask;
-	__u16	sport;
-	__u16	sport_mask;
-	__u8	prefixlen_d;
-	__u8	prefixlen_s;
-	__u8	proto;
-	int	ifindex;
-	uid_t	user;
-	void	*owner;
-};
-
-#define XFRM_INF (~(u64)0)
-
-struct xfrm_lifetime_cfg
-{
-	u64	soft_byte_limit;
-	u64	hard_byte_limit;
-	u64	soft_packet_limit;
-	u64	hard_packet_limit;
-	u64	soft_add_expires_seconds;
-	u64	hard_add_expires_seconds;
-	u64	soft_use_expires_seconds;
-	u64	hard_use_expires_seconds;
-};
-
-struct xfrm_lifetime_cur
-{
-	u64	bytes;
-	u64	packets;
-	u64	add_time;
-	u64	use_time;
-};
-
-struct xfrm_replay_state
-{
-	u32	oseq;
-	u32	seq;
-	u32	bitmap;
-};
-
-struct xfrm_algo {
-	char	alg_name[CRYPTO_MAX_ALG_NAME];
-	int	alg_key_len;    /* in bits */
-	char	alg_key[0];
-};
-
 /* Full description of state of transformer. */
 struct xfrm_state
 {
@@ -188,11 +112,7 @@ struct xfrm_state
 	struct xfrm_replay_state replay;
 
 	/* Statistics */
-	struct {
-		u32		replay_window;
-		u32		replay;
-		u32		integrity_failed;
-	} stats;
+	struct xfrm_stats	stats;
 
 	struct xfrm_lifetime_cur curlft;
 	struct timer_list	timer;
@@ -320,7 +240,7 @@ extern int xfrm_register_km(struct xfrm_mgr *km);
 extern int xfrm_unregister_km(struct xfrm_mgr *km);
 
 
-extern struct xfrm_policy *xfrm_policy_list[XFRM_POLICY_MAX];
+extern struct xfrm_policy *xfrm_policy_list[XFRM_POLICY_MAX*2];
 
 static inline void xfrm_pol_hold(struct xfrm_policy *policy)
 {
@@ -429,16 +349,16 @@ static inline int xfrm_sk_clone_policy(struct sock *sk)
 	return 0;
 }
 
-extern void __xfrm_sk_free_policy(struct xfrm_policy *);
+extern void __xfrm_sk_free_policy(struct xfrm_policy *, int dir);
 
 static inline void xfrm_sk_free_policy(struct sock *sk)
 {
 	if (unlikely(sk->policy[0] != NULL)) {
-		__xfrm_sk_free_policy(sk->policy[0]);
+		__xfrm_sk_free_policy(sk->policy[0], 0);
 		sk->policy[0] = NULL;
 	}
 	if (unlikely(sk->policy[1] != NULL)) {
-		__xfrm_sk_free_policy(sk->policy[1]);
+		__xfrm_sk_free_policy(sk->policy[1], 1);
 		sk->policy[1] = NULL;
 	}
 }
@@ -477,7 +397,7 @@ extern int xfrm_sk_policy_insert(struct sock *sk, int dir, struct xfrm_policy *p
 extern struct xfrm_policy *xfrm_sk_policy_lookup(struct sock *sk, int dir, struct flowi *fl);
 extern int xfrm_flush_bundles(struct xfrm_state *x);
 
-extern wait_queue_head_t *km_waitq;
+extern wait_queue_head_t km_waitq;
 extern void km_warn_expired(struct xfrm_state *x);
 extern void km_expired(struct xfrm_state *x);
 extern int km_query(struct xfrm_state *x, struct xfrm_tmpl *, struct xfrm_policy *pol);

net/ipv4/arp.c

@@ -1142,65 +1142,115 @@ struct arp_iter_state {
 	int is_pneigh, bucket;
 };
 
-static __inline__ struct neighbour *neigh_get_bucket(struct seq_file *seq,
-						     loff_t *pos)
+static struct neighbour *neigh_get_first(struct seq_file *seq)
 {
-	struct neighbour *n = NULL;
 	struct arp_iter_state* state = seq->private;
-	loff_t l = *pos;
-	int i;
-
-	for (; state->bucket <= NEIGH_HASHMASK; ++state->bucket)
-		for (i = 0, n = arp_tbl.hash_buckets[state->bucket]; n;
-		     ++i, n = n->next)
-			/* Do not confuse users "arp -a" with magic entries */
-			if ((n->nud_state & ~NUD_NOARP) && !l--) {
-				*pos = i;
-				goto out;
-			}
-out:
+	struct neighbour *n = NULL;
+
+	state->is_pneigh = 0;
+
+	for (state->bucket = 0;
+	     state->bucket <= NEIGH_HASHMASK;
+	     ++state->bucket) {
+		n = arp_tbl.hash_buckets[state->bucket];
+		while (n && !(n->nud_state & ~NUD_NOARP))
+			n = n->next;
+		if (n)
+			break;
+	}
+
 	return n;
 }
 
-static __inline__ struct pneigh_entry *pneigh_get_bucket(struct seq_file *seq,
-							 loff_t *pos)
+static struct neighbour *neigh_get_next(struct seq_file *seq,
+					struct neighbour *n)
 {
-	struct pneigh_entry *n = NULL;
 	struct arp_iter_state* state = seq->private;
-	loff_t l = *pos;
-	int i;
-
-	for (; state->bucket <= PNEIGH_HASHMASK; ++state->bucket)
-		for (i = 0, n = arp_tbl.phash_buckets[state->bucket]; n;
-		     ++i, n = n->next)
-			if (!l--) {
-				*pos = i;
-				goto out;
-			}
+
+	do {
+		n = n->next;
+		/* Don't confuse "arp -a" w/ magic entries */
+try_again:
+	} while (n && !(n->nud_state & ~NUD_NOARP));
+
+	if (n)
+		goto out;
+	if (++state->bucket > NEIGH_HASHMASK)
+		goto out;
+	n = arp_tbl.hash_buckets[state->bucket];
+	goto try_again;
 out:
 	return n;
 }
 
-static __inline__ void *arp_get_bucket(struct seq_file *seq, loff_t *pos)
+static struct neighbour *neigh_get_idx(struct seq_file *seq, loff_t *pos)
 {
-	void *rc = neigh_get_bucket(seq, pos);
+	struct neighbour *n = neigh_get_first(seq);
 
-	if (!rc) {
-		struct arp_iter_state* state = seq->private;
+	if (n)
+		while (*pos && (n = neigh_get_next(seq, n)))
+			--*pos;
+	return *pos ? NULL : n;
+}
+
+static struct pneigh_entry *pneigh_get_first(struct seq_file *seq)
+{
+	struct arp_iter_state* state = seq->private;
+	struct pneigh_entry *pn;
+
+	state->is_pneigh = 1;
+
+	for (state->bucket = 0;
+	     state->bucket <= PNEIGH_HASHMASK;
+	     ++state->bucket) {
+		pn = arp_tbl.phash_buckets[state->bucket];
+		if (pn)
+			break;
+	}
+	return pn;
+}
 
+static struct pneigh_entry *pneigh_get_next(struct seq_file *seq,
+					    struct pneigh_entry *pn)
+{
+	struct arp_iter_state* state = seq->private;
+
+	pn = pn->next;
+	while (!pn) {
+		if (++state->bucket > PNEIGH_HASHMASK)
+			break;
+		pn = arp_tbl.phash_buckets[state->bucket];
+	}
+	return pn;
+}
+
+static struct pneigh_entry *pneigh_get_idx(struct seq_file *seq, loff_t pos)
+{
+	struct pneigh_entry *pn = pneigh_get_first(seq);
+
+	if (pn)
+		while (pos && (pn = pneigh_get_next(seq, pn)))
+			--pos;
+	return pos ? NULL : pn;
+}
+
+static void *arp_get_idx(struct seq_file *seq, loff_t pos)
+{
+	void *rc;
+
+	read_lock_bh(&arp_tbl.lock);
+	rc = neigh_get_idx(seq, &pos);
+
+	if (!rc) {
 		read_unlock_bh(&arp_tbl.lock);
-		state->is_pneigh = 1;
-		state->bucket	 = 0;
-		*pos		 = 0;
-		rc = pneigh_get_bucket(seq, pos);
+		rc = pneigh_get_idx(seq, pos);
 	}
 	return rc;
 }
 
 static void *arp_seq_start(struct seq_file *seq, loff_t *pos)
 {
-	read_lock_bh(&arp_tbl.lock);
-	return *pos ? arp_get_bucket(seq, pos) : (void *)1;
+	return *pos ? arp_get_idx(seq, *pos - 1) : (void *)1;
 }
 
 static void *arp_seq_next(struct seq_file *seq, void *v, loff_t *pos)
@@ -1209,38 +1259,19 @@ static void *arp_seq_next(struct seq_file *seq, void *v, loff_t *pos)
 	struct arp_iter_state* state;
 
 	if (v == (void *)1) {
-		rc = arp_get_bucket(seq, pos);
+		rc = arp_get_idx(seq, 0);
 		goto out;
 	}
 
 	state = seq->private;
 	if (!state->is_pneigh) {
-		struct neighbour *n = v;
-
-		rc = n = n->next;
-		if (n)
-			goto out;
-		*pos = 0;
-		++state->bucket;
-		rc = neigh_get_bucket(seq, pos);
+		rc = neigh_get_next(seq, v);
 		if (rc)
 			goto out;
 		read_unlock_bh(&arp_tbl.lock);
-		state->is_pneigh = 1;
-		state->bucket	 = 0;
-		*pos		 = 0;
-		rc = pneigh_get_bucket(seq, pos);
-	} else {
-		struct pneigh_entry *pn = v;
-
-		pn = pn->next;
-		if (!pn) {
-			++state->bucket;
-			*pos = 0;
-			pn   = pneigh_get_bucket(seq, pos);
-		}
-		rc = pn;
-	}
+		rc = pneigh_get_first(seq);
+	} else
+		rc = pneigh_get_next(seq, v);
 out:
 	++*pos;
 	return rc;
@@ -1291,7 +1322,6 @@ static __inline__ void arp_format_neigh_entry(struct seq_file *seq,
 static __inline__ void arp_format_pneigh_entry(struct seq_file *seq,
 					       struct pneigh_entry *n)
 {
-
 	struct net_device *dev = n->dev;
 	int hatype = dev ? dev->type : 0;
 	char tbuf[16];

net/ipv4/ip_output.c

@@ -815,14 +815,16 @@ int ip_append_data(struct sock *sk,
 				alloclen = maxfraglen;
 			else
 				alloclen = datalen + fragheaderlen;
-			if (!(flags & MSG_DONTWAIT) || transhdrlen) {
+			if (transhdrlen) {
 				skb = sock_alloc_send_skb(sk, 
 						alloclen + hh_len + 15,
 						(flags & MSG_DONTWAIT), &err);
 			} else {
-				skb = sock_wmalloc(sk, 
-						alloclen + hh_len + 15, 1,
-						sk->allocation);
+				skb = NULL;
+				if (atomic_read(&sk->wmem_alloc) <= 2*sk->sndbuf)
+					skb = sock_wmalloc(sk, 
+							   alloclen + hh_len + 15, 1,
+							   sk->allocation);
 				if (unlikely(skb == NULL))
 					err = -ENOBUFS;
 			}

net/ipv4/proc.c

@@ -164,13 +164,11 @@ int snmp_get_info(char *buffer, char **start, off_t offset, int length)
 /* 
  *	Output /proc/net/netstat
  */
- 
-int netstat_get_info(char *buffer, char **start, off_t offset, int length)
+static int netstat_seq_show(struct seq_file *seq, void *v)
 {
-	int len, i;
+	int i;
 
-	len = sprintf(buffer,
-		      "TcpExt: SyncookiesSent SyncookiesRecv SyncookiesFailed"
+	seq_puts(seq, "TcpExt: SyncookiesSent SyncookiesRecv SyncookiesFailed"
 		      " EmbryonicRsts PruneCalled RcvPruned OfoPruned"
 		      " OutOfWindowIcmps LockDroppedIcmps ArpFilter"
 		      " TW TWRecycled TWKilled"
@@ -196,31 +194,37 @@ int netstat_get_info(char *buffer, char **start, off_t offset, int length)
 		      " TCPAbortOnMemory TCPAbortOnTimeout TCPAbortOnLinger"
 		      " TCPAbortFailed TCPMemoryPressures\n"
 		      "TcpExt:");
-	for (i=0; i<offsetof(struct linux_mib, __pad)/sizeof(unsigned long); i++)
-		len += sprintf(buffer+len, " %lu", fold_field((unsigned long*)net_statistics, sizeof(struct linux_mib), i));
-
-	len += sprintf (buffer + len, "\n");
+	for (i = 0;
+	     i < offsetof(struct linux_mib, __pad) / sizeof(unsigned long);
+	     i++)
+		seq_printf(seq, " %lu",
+			   fold_field((unsigned long *)net_statistics,
+				      sizeof(struct linux_mib), i));
+	seq_putc(seq, '\n');
+	return 0;
+}
 
-	if (offset >= len)
-	{
-		*start = buffer;
-		return 0;
-	}
-	*start = buffer + offset;
-	len -= offset;
-	if (len > length)
-		len = length;
-	if (len < 0)
-		len = 0; 
-	return len;
+static int netstat_seq_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, netstat_seq_show, NULL);
 }
 
+static struct file_operations netstat_seq_fops = {
+	.open	 = netstat_seq_open,
+	.read	 = seq_read,
+	.llseek	 = seq_lseek,
+	.release = single_release,
+};
+
 int __init ip_misc_proc_init(void)
 {
 	int rc = 0;
+	struct proc_dir_entry *p = create_proc_entry("netstat", S_IRUGO, proc_net);
 
-	if (!proc_net_create("netstat", 0, netstat_get_info))
+	if (!p)
 		goto out_netstat;
+	p->proc_fops = &netstat_seq_fops;
+
 	if (!proc_net_create("snmp", 0, snmp_get_info))
 		goto out_snmp;
 	if (!proc_net_create("sockstat", 0, afinet_get_info))
@@ -230,7 +234,7 @@ int __init ip_misc_proc_init(void)
 out_sockstat:
 	proc_net_remove("snmp");
 out_snmp:
-	proc_net_remove("netstat");
+	remove_proc_entry("netstat", proc_net);
 out_netstat:
 	rc = -ENOMEM;
 	goto out;

net/ipv4/raw.c

@@ -431,7 +431,7 @@ static int raw_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
 						.saddr = saddr,
 						.tos = tos } },
 				    .proto = inet->hdrincl ? IPPROTO_RAW : sk->protocol };
-		err = ip_route_output_flow(&rt, &fl, sk, msg->msg_flags&MSG_DONTWAIT);
+		err = ip_route_output_flow(&rt, &fl, sk, !(msg->msg_flags&MSG_DONTWAIT));
 	}
 	if (err)
 		goto done;

net/ipv4/tcp_ipv4.c

@@ -780,7 +780,7 @@ int tcp_v4_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)
 	tmp = ip_route_connect(&rt, nexthop, inet->saddr,
 			       RT_CONN_FLAGS(sk), sk->bound_dev_if,
 			       IPPROTO_TCP,
-			       inet->sport, usin->sin_port);
+			       inet->sport, usin->sin_port, sk);
 	if (tmp < 0)
 		return tmp;
 
@@ -837,7 +837,7 @@ int tcp_v4_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)
 	if (err)
 		goto failure;
 
-	err = ip_route_newports(&rt, inet->sport, inet->dport);
+	err = ip_route_newports(&rt, inet->sport, inet->dport, sk);
 	if (err)
 		goto failure;
 
@@ -1896,7 +1896,7 @@ static int tcp_v4_reselect_saddr(struct sock *sk)
 			       RT_TOS(inet->tos) | sk->localroute,
 			       sk->bound_dev_if,
 			       IPPROTO_TCP,
-			       inet->sport, inet->dport);
+			       inet->sport, inet->dport, sk);
 	if (err)
 		return err;
 

net/ipv4/udp.c

@@ -601,7 +601,7 @@ int udp_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
 				    .uli_u = { .ports =
 					       { .sport = inet->sport,
 						 .dport = dport } } };
-		err = ip_route_output_flow(&rt, &fl, sk, msg->msg_flags&MSG_DONTWAIT);
+		err = ip_route_output_flow(&rt, &fl, sk, !(msg->msg_flags&MSG_DONTWAIT));
 		if (err)
 			goto out;
 
@@ -884,7 +884,7 @@ int udp_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)
 	err = ip_route_connect(&rt, usin->sin_addr.s_addr, saddr,
 			       RT_CONN_FLAGS(sk), oif,
 			       IPPROTO_UDP,
-			       inet->sport, usin->sin_port);
+			       inet->sport, usin->sin_port, sk);
 	if (err)
 		return err;
 	if ((rt->rt_flags&RTCF_BROADCAST) && !sk->broadcast) {

net/ipv4/xfrm_policy.c

 #include <net/xfrm.h>
 #include <net/ip.h>
 
+DECLARE_MUTEX(xfrm_cfg_sem);
+
 static u32      xfrm_policy_genid;
 static rwlock_t xfrm_policy_lock = RW_LOCK_UNLOCKED;
 
-struct xfrm_policy *xfrm_policy_list[XFRM_POLICY_MAX];
+struct xfrm_policy *xfrm_policy_list[XFRM_POLICY_MAX*2];
 
 extern struct dst_ops xfrm4_dst_ops;
 
@@ -263,10 +265,13 @@ static u32 xfrm_gen_index(int dir)
 {
 	u32 idx;
 	struct xfrm_policy *p;
-	static u32 pol_id;
+	static u32 idx_generator;
 
 	for (;;) {
-		idx = (++pol_id ? : ++pol_id);
+		idx = (idx_generator | dir);
+		idx_generator += 8;
+		if (idx == 0)
+			idx = 8;
 		for (p = xfrm_policy_list[dir]; p; p = p->next) {
 			if (p->index == idx)
 				break;
@@ -300,6 +305,7 @@ int xfrm_policy_insert(int dir, struct xfrm_policy *policy, int excl)
 	write_unlock_bh(&xfrm_policy_lock);
 
 	if (pol) {
+		atomic_dec(&pol->refcnt);
 		xfrm_policy_kill(pol);
 		xfrm_pol_put(pol);
 	}
@@ -328,7 +334,7 @@ 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[dir]; (pol=*p)!=NULL; p = &pol->next) {
+	for (p = &xfrm_policy_list[id & 7]; (pol=*p)!=NULL; p = &pol->next) {
 		if (pol->index == id) {
 			if (delete)
 				*p = pol->next;
@@ -375,7 +381,7 @@ int xfrm_policy_walk(int (*func)(struct xfrm_policy *, int, int, void*),
 	int error = 0;
 
 	read_lock(&xfrm_policy_lock);
-	for (dir = 0; dir < XFRM_POLICY_MAX; dir++) {
+	for (dir = 0; dir < 2*XFRM_POLICY_MAX; dir++) {
 		for (xp = xfrm_policy_list[dir]; xp; xp = xp->next)
 			count++;
 	}
@@ -385,9 +391,9 @@ int xfrm_policy_walk(int (*func)(struct xfrm_policy *, int, int, void*),
 		goto out;
 	}
 
-	for (dir = 0; dir < XFRM_POLICY_MAX; dir++) {
+	for (dir = 0; dir < 2*XFRM_POLICY_MAX; dir++) {
 		for (xp = xfrm_policy_list[dir]; xp; xp = xp->next) {
-			error = func(xp, dir, --count, data);
+			error = func(xp, dir%XFRM_POLICY_MAX, --count, data);
 			if (error)
 				goto out;
 		}
@@ -423,18 +429,37 @@ struct xfrm_policy *xfrm_sk_policy_lookup(struct sock *sk, int dir, struct flowi
 	struct xfrm_policy *pol;
 
 	read_lock(&xfrm_policy_lock);
-	for (pol = sk->policy[dir]; pol; pol = pol->next) {
-		struct xfrm_selector *sel = &pol->selector;
-
-		if (xfrm4_selector_match(sel, fl)) {
+	if ((pol = sk->policy[dir]) != NULL) {
+		if (xfrm4_selector_match(&pol->selector, fl))
 			atomic_inc(&pol->refcnt);
-			break;
-		}
+		else
+			pol = NULL;
 	}
 	read_unlock(&xfrm_policy_lock);
 	return pol;
 }
 
+void xfrm_sk_policy_link(struct xfrm_policy *pol, int dir)
+{
+	pol->next = xfrm_policy_list[XFRM_POLICY_MAX+dir];
+	xfrm_policy_list[XFRM_POLICY_MAX+dir] = pol;
+	atomic_inc(&pol->refcnt);
+}
+
+void xfrm_sk_policy_unlink(struct xfrm_policy *pol, int dir)
+{
+	struct xfrm_policy **polp;
+
+	for (polp = &xfrm_policy_list[XFRM_POLICY_MAX+dir];
+	     *polp != NULL; polp = &(*polp)->next) {
+		if (*polp == pol) {
+			*polp = pol->next;
+			atomic_dec(&pol->refcnt);
+			return;
+		}
+	}
+}
+
 int xfrm_sk_policy_insert(struct sock *sk, int dir, struct xfrm_policy *pol)
 {
 	struct xfrm_policy *old_pol;
@@ -442,6 +467,13 @@ int xfrm_sk_policy_insert(struct sock *sk, int dir, struct xfrm_policy *pol)
 	write_lock_bh(&xfrm_policy_lock);
 	old_pol = sk->policy[dir];
 	sk->policy[dir] = pol;
+	if (pol) {
+		pol->curlft.add_time = (unsigned long)xtime.tv_sec;
+		pol->index = xfrm_gen_index(XFRM_POLICY_MAX+dir);
+		xfrm_sk_policy_link(pol, dir);
+	}
+	if (old_pol)
+		xfrm_sk_policy_unlink(old_pol, dir);
 	write_unlock_bh(&xfrm_policy_lock);
 
 	if (old_pol) {
@@ -451,7 +483,7 @@ int xfrm_sk_policy_insert(struct sock *sk, int dir, struct xfrm_policy *pol)
 	return 0;
 }
 
-static struct xfrm_policy *clone_policy(struct xfrm_policy *old)
+static struct xfrm_policy *clone_policy(struct xfrm_policy *old, int dir)
 {
 	struct xfrm_policy *newp = xfrm_policy_alloc(GFP_ATOMIC);
 
@@ -462,8 +494,12 @@ static struct xfrm_policy *clone_policy(struct xfrm_policy *old)
 		newp->action = old->action;
 		newp->flags = old->flags;
 		newp->xfrm_nr = old->xfrm_nr;
+		newp->index = old->index;
 		memcpy(newp->xfrm_vec, old->xfrm_vec,
 		       newp->xfrm_nr*sizeof(struct xfrm_tmpl));
+		write_lock_bh(&xfrm_policy_lock);
+		xfrm_sk_policy_link(newp, dir);
+		write_unlock_bh(&xfrm_policy_lock);
 	}
 	return newp;
 }
@@ -475,15 +511,19 @@ int __xfrm_sk_clone_policy(struct sock *sk)
 	p1 = sk->policy[1];
 	sk->policy[0] = NULL;
 	sk->policy[1] = NULL;
-	if (p0 && (sk->policy[0] = clone_policy(p0)) == NULL)
+	if (p0 && (sk->policy[0] = clone_policy(p0, 0)) == NULL)
 		return -ENOMEM;
-	if (p1 && (sk->policy[1] = clone_policy(p1)) == NULL)
+	if (p1 && (sk->policy[1] = clone_policy(p1, 1)) == NULL)
 		return -ENOMEM;
 	return 0;
 }
 
-void __xfrm_sk_free_policy(struct xfrm_policy *pol)
+void __xfrm_sk_free_policy(struct xfrm_policy *pol, int dir)
 {
+	write_lock_bh(&xfrm_policy_lock);
+	xfrm_sk_policy_unlink(pol, dir);
+	write_unlock_bh(&xfrm_policy_lock);
+
 	xfrm_policy_kill(pol);
 	xfrm_pol_put(pol);
 }
@@ -734,12 +774,12 @@ int xfrm_lookup(struct dst_entry **dst_p, struct flowi *fl,
 					goto error;
 
 				__set_task_state(tsk, TASK_INTERRUPTIBLE);
-				add_wait_queue(km_waitq, &wait);
+				add_wait_queue(&km_waitq, &wait);
 				err = xfrm_tmpl_resolve(policy, fl, xfrm);
 				if (err == -EAGAIN)
 					schedule();
 				__set_task_state(tsk, TASK_RUNNING);
-				remove_wait_queue(km_waitq, &wait);
+				remove_wait_queue(&km_waitq, &wait);
 
 				if (err == -EAGAIN && signal_pending(current)) {
 					err = -ERESTART;
@@ -888,7 +928,7 @@ int __xfrm_policy_check(struct sock *sk, int dir, struct sk_buff *skb)
 
 	pol = NULL;
 	if (sk && sk->policy[dir])
-		pol =xfrm_sk_policy_lookup(sk, dir, &fl);
+		pol = xfrm_sk_policy_lookup(sk, dir, &fl);
 
 	if (!pol)
 		pol = flow_lookup(dir, &fl);
@@ -985,7 +1025,7 @@ static int xfrm4_garbage_collect(void)
 	struct dst_entry *dst, **dstp, *gc_list = NULL;
 
 	read_lock_bh(&xfrm_policy_lock);
-	for (i=0; i<XFRM_POLICY_MAX; i++) {
+	for (i=0; i<2*XFRM_POLICY_MAX; i++) {
 		for (pol = xfrm_policy_list[i]; pol; pol = pol->next) {
 			write_lock(&pol->lock);
 			dstp = &pol->bundles;
@@ -1028,7 +1068,7 @@ int xfrm_flush_bundles(struct xfrm_state *x)
 	struct dst_entry *dst, **dstp, *gc_list = NULL;
 
 	read_lock_bh(&xfrm_policy_lock);
-	for (i=0; i<XFRM_POLICY_MAX; i++) {
+	for (i=0; i<2*XFRM_POLICY_MAX; i++) {
 		for (pol = xfrm_policy_list[i]; pol; pol = pol->next) {
 			write_lock(&pol->lock);
 			dstp = &pol->bundles;

net/ipv4/xfrm_state.c

@@ -22,7 +22,7 @@ static spinlock_t xfrm_state_lock = SPIN_LOCK_UNLOCKED;
 static struct list_head xfrm_state_bydst[XFRM_DST_HSIZE];
 static struct list_head xfrm_state_byspi[XFRM_DST_HSIZE];
 
-wait_queue_head_t *km_waitq;
+DECLARE_WAIT_QUEUE_HEAD(km_waitq);
 
 #define ACQ_EXPIRES 30
 
@@ -163,7 +163,7 @@ static void __xfrm_state_delete(struct xfrm_state *x)
 
 	if (kill && x->type)
 		x->type->destructor(x);
-	wake_up(km_waitq);
+	wake_up(&km_waitq);
 }
 
 void xfrm_state_delete(struct xfrm_state *x)
@@ -195,7 +195,7 @@ void xfrm_state_flush(u8 proto)
 		}
 	}
 	spin_unlock_bh(&xfrm_state_lock);
-	wake_up(km_waitq);
+	wake_up(&km_waitq);
 }
 
 struct xfrm_state *
@@ -213,6 +213,7 @@ xfrm_state_find(u32 daddr, u32 saddr, struct flowi *fl, struct xfrm_tmpl *tmpl,
 	spin_lock_bh(&xfrm_state_lock);
 	list_for_each_entry(x, xfrm_state_bydst+h, bydst) {
 		if (daddr == x->id.daddr.xfrm4_addr &&
+		    x->props.reqid == tmpl->reqid &&
 		    (saddr == x->props.saddr.xfrm4_addr || !saddr || !x->props.saddr.xfrm4_addr) &&
 		    tmpl->mode == x->props.mode &&
 		    tmpl->id.proto == x->id.proto) {
@@ -278,6 +279,7 @@ xfrm_state_find(u32 daddr, u32 saddr, struct flowi *fl, struct xfrm_tmpl *tmpl,
 		if (x->props.saddr.xfrm4_addr == 0)
 			x->props.saddr.xfrm4_addr = saddr;
 		x->props.mode = tmpl->mode;
+		x->props.reqid = tmpl->reqid;
 
 		if (km_query(x, tmpl, pol) == 0) {
 			x->km.state = XFRM_STATE_ACQ;
@@ -323,7 +325,7 @@ void xfrm_state_insert(struct xfrm_state *x)
 		atomic_inc(&x->refcnt);
 
 	spin_unlock_bh(&xfrm_state_lock);
-	wake_up(km_waitq);
+	wake_up(&km_waitq);
 }
 
 int xfrm_state_check_expire(struct xfrm_state *x)
@@ -398,7 +400,7 @@ xfrm_find_acq(u8 mode, u16 reqid, u8 proto, u32 daddr, u32 saddr)
 		    mode == x->props.mode &&
 		    proto == x->id.proto &&
 		    saddr == x->props.saddr.xfrm4_addr &&
-		    (!reqid || reqid == x->props.reqid)) {
+		    reqid == x->props.reqid) {
 			    if (!x0)
 				    x0 = x;
 			    if (x->km.state != XFRM_STATE_ACQ)
@@ -427,7 +429,7 @@ xfrm_find_acq(u8 mode, u16 reqid, u8 proto, u32 daddr, u32 saddr)
 		mod_timer(&x0->timer, jiffies + ACQ_EXPIRES*HZ);
 		atomic_inc(&x0->refcnt);
 		list_add_tail(&x0->bydst, xfrm_state_bydst+h);
-		wake_up(km_waitq);
+		wake_up(&km_waitq);
 	}
 	spin_unlock_bh(&xfrm_state_lock);
 	return x0;
@@ -491,7 +493,7 @@ xfrm_alloc_spi(struct xfrm_state *x, u32 minspi, u32 maxspi)
 		list_add(&x->byspi, xfrm_state_byspi+h);
 		atomic_inc(&x->refcnt);
 		spin_unlock_bh(&xfrm_state_lock);
-		wake_up(km_waitq);
+		wake_up(&km_waitq);
 	}
 }
 

net/key/af_key.c

@@ -30,6 +30,7 @@
 #define _X2KEY(x) ((x) == XFRM_INF ? 0 : (x))
 #define _KEY2X(x) ((x) == 0 ? XFRM_INF : (x))
 
+
 /* List of all pfkey sockets. */
 static struct sock * pfkey_table;
 static DECLARE_WAIT_QUEUE_HEAD(pfkey_table_wait);
@@ -871,10 +872,6 @@ static struct xfrm_state * pfkey_msg2xfrm_state(struct sadb_msg *hdr,
 	    !!ext_hdrs[SADB_EXT_LIFETIME_SOFT-1])
 		return ERR_PTR(-EINVAL);
 
-	/* XXX Do we need this check ? */
-	if (((struct sadb_address *) ext_hdrs[SADB_EXT_ADDRESS_SRC-1])->sadb_address_prefixlen != 32 ||
-	    ((struct sadb_address *) ext_hdrs[SADB_EXT_ADDRESS_DST-1])->sadb_address_prefixlen != 32)
-		return ERR_PTR(-EINVAL);
 	proto = pfkey_satype2proto(hdr->sadb_msg_satype);
 	if (proto == 0)
 		return ERR_PTR(-EINVAL);
@@ -1379,6 +1376,34 @@ static int pfkey_promisc(struct sock *sk, struct sk_buff *skb, struct sadb_msg *
 	return 0;
 }
 
+static int check_reqid(struct xfrm_policy *xp, int dir, int count, void *ptr)
+{
+	int i;
+	u16 reqid = *(u16*)ptr;
+
+	for (i=0; i<xp->xfrm_nr; i++) {
+		if (xp->xfrm_vec[i].reqid == reqid)
+			return -EEXIST;
+	}
+	return 0;
+}
+
+static u16 gen_reqid(void)
+{
+	u16 start;
+	static u16 reqid = IPSEC_MANUAL_REQID_MAX;
+
+	start = reqid;
+	do {
+		++reqid;
+		if (reqid == 0)
+			reqid = IPSEC_MANUAL_REQID_MAX+1;
+		if (xfrm_policy_walk(check_reqid, (void*)&reqid) != -EEXIST)
+			return reqid;
+	} while (reqid != start);
+	return 0;
+}
+
 static int
 parse_ipsecrequest(struct xfrm_policy *xp, struct sadb_x_ipsecrequest *rq)
 {
@@ -1395,7 +1420,14 @@ parse_ipsecrequest(struct xfrm_policy *xp, struct sadb_x_ipsecrequest *rq)
 	t->mode = rq->sadb_x_ipsecrequest_mode-1;
 	if (rq->sadb_x_ipsecrequest_level == IPSEC_LEVEL_USE)
 		t->optional = 1;
-	t->reqid = rq->sadb_x_ipsecrequest_reqid;
+	else if (rq->sadb_x_ipsecrequest_level == IPSEC_LEVEL_UNIQUE) {
+		t->reqid = rq->sadb_x_ipsecrequest_reqid;
+		if (t->reqid > IPSEC_MANUAL_REQID_MAX)
+			t->reqid = 0;
+		if (!t->reqid && !(t->reqid = gen_reqid()))
+			return -ENOBUFS;
+	}
+
 	/* addresses present only in tunnel mode */
 	if (t->mode) {
 		addr = (void*)(rq+1);
@@ -1544,6 +1576,8 @@ static struct sk_buff * pfkey_xfrm_policy2msg(struct xfrm_policy *xp, int dir)
 		rq->sadb_x_ipsecrequest_proto = t->id.proto;
 		rq->sadb_x_ipsecrequest_mode = t->mode+1;
 		rq->sadb_x_ipsecrequest_level = IPSEC_LEVEL_REQUIRE;
+		if (t->reqid)
+			rq->sadb_x_ipsecrequest_level = IPSEC_LEVEL_UNIQUE;
 		if (t->optional)
 			rq->sadb_x_ipsecrequest_level = IPSEC_LEVEL_USE;
 		rq->sadb_x_ipsecrequest_reqid = t->reqid;
@@ -1590,7 +1624,6 @@ static int pfkey_spdadd(struct sock *sk, struct sk_buff *skb, struct sadb_msg *h
 	if (xp == NULL)
 		return -ENOBUFS;
 
-	xp->index = pol->sadb_x_policy_id;
 	xp->action = (pol->sadb_x_policy_type == IPSEC_POLICY_DISCARD ?
 		      XFRM_POLICY_BLOCK : XFRM_POLICY_ALLOW);
 
@@ -2013,6 +2046,18 @@ static int pfkey_send_notify(struct xfrm_state *x, int hard)
 	return 0;
 }
 
+static u32 get_acqseq(void)
+{
+	u32 res;
+	static u32 acqseq;
+	static spinlock_t acqseq_lock = SPIN_LOCK_UNLOCKED;
+
+	spin_lock_bh(&acqseq_lock);
+	res = (++acqseq ? : ++acqseq);
+	spin_unlock_bh(&acqseq_lock);
+	return res;
+}
+
 static int pfkey_send_acquire(struct xfrm_state *x, struct xfrm_tmpl *t, struct xfrm_policy *xp, int dir)
 {
 	struct sk_buff *skb;
@@ -2020,7 +2065,6 @@ static int pfkey_send_acquire(struct xfrm_state *x, struct xfrm_tmpl *t, struct
 	struct sadb_address *addr;
 	struct sadb_x_policy *pol;
 	int size;
-	static u32 acqseq;
 	
 	size = sizeof(struct sadb_msg) +
 		sizeof(struct sadb_address)*2 + 
@@ -2043,10 +2087,9 @@ static int pfkey_send_acquire(struct xfrm_state *x, struct xfrm_tmpl *t, struct
 	hdr->sadb_msg_len = size / sizeof(uint64_t);
 	hdr->sadb_msg_errno = 0;
 	hdr->sadb_msg_reserved = 0;
-	hdr->sadb_msg_seq = (++acqseq ? : ++acqseq);
-	x->km.seq = acqseq;
+	hdr->sadb_msg_seq = x->km.seq = get_acqseq();
 	hdr->sadb_msg_pid = 0;
-	
+
 	/* src address */
 	addr = (struct sadb_address*) skb_put(skb, 
 					      sizeof(struct sadb_address)+sizeof(struct sockaddr_in));
@@ -2076,7 +2119,7 @@ static int pfkey_send_acquire(struct xfrm_state *x, struct xfrm_tmpl *t, struct
 	((struct sockaddr_in*)(addr + 1))->sin_addr.s_addr = 
 		x->id.daddr.xfrm4_addr;
 	((struct sockaddr_in*)(addr + 1))->sin_port = 0;
-	
+
 	pol = (struct sadb_x_policy *)  skb_put(skb, sizeof(struct sadb_x_policy));
 	pol->sadb_x_policy_len = sizeof(struct sadb_x_policy)/sizeof(uint64_t);
 	pol->sadb_x_policy_exttype = SADB_X_EXT_POLICY;
@@ -2119,7 +2162,6 @@ static struct xfrm_policy *pfkey_compile_policy(int opt, u8 *data, int len, int
 		return NULL;
 	}
 
-	xp->index = pol->sadb_x_policy_id;
 	xp->action = (pol->sadb_x_policy_type == IPSEC_POLICY_DISCARD ?
 		      XFRM_POLICY_BLOCK : XFRM_POLICY_ALLOW);
 
@@ -2171,7 +2213,9 @@ static int pfkey_sendmsg(struct kiocb *kiocb,
 	if (!hdr)
 		goto out;
 
+	down(&xfrm_cfg_sem);
 	err = pfkey_process(sk, skb, hdr);
+	up(&xfrm_cfg_sem);
 
 out:
 	if (err && hdr && pfkey_error(hdr, err, sk) == 0)
@@ -2179,7 +2223,7 @@ static int pfkey_sendmsg(struct kiocb *kiocb,
 	if (skb)
 		kfree_skb(skb);
 
-	return err;
+	return err ? : len;
 }
 
 static int pfkey_recvmsg(struct kiocb *kiocb,

net/netsyms.c

@@ -283,6 +283,7 @@ extern int (*dlci_ioctl_hook)(unsigned int, void *);
 EXPORT_SYMBOL(dlci_ioctl_hook);
 #endif
 
+EXPORT_SYMBOL(xfrm_cfg_sem);
 EXPORT_SYMBOL(xfrm_policy_alloc);
 EXPORT_SYMBOL(__xfrm_policy_destroy);
 EXPORT_SYMBOL(xfrm_policy_lookup);