Merge master.kernel.org:/pub/scm/linux/kernel/git/acme/net-2.6.15

include/linux/skbuff.h

@@ -171,7 +171,6 @@ enum {
  *	struct sk_buff - socket buffer
  *	@next: Next buffer in list
  *	@prev: Previous buffer in list
- *	@list: List we are on
  *	@sk: Socket we are owned by
  *	@tstamp: Time we arrived
  *	@dev: Device we arrived on/are leaving by
@@ -190,6 +189,7 @@ enum {
  *	@cloned: Head may be cloned (check refcnt to be sure)
  *	@nohdr: Payload reference only, must not modify header
  *	@pkt_type: Packet class
+ *	@fclone: skbuff clone status
  *	@ip_summed: Driver fed us an IP checksum
  *	@priority: Packet queueing priority
  *	@users: User count - see {datagram,tcp}.c
@@ -202,6 +202,7 @@ enum {
  *	@destructor: Destruct function
  *	@nfmark: Can be used for communication between hooks
  *	@nfct: Associated connection, if any
+ *	@ipvs_property: skbuff is owned by ipvs
  *	@nfctinfo: Relationship of this skb to the connection
  *	@nf_bridge: Saved data about a bridged frame - see br_netfilter.c
  *	@tc_index: Traffic control index

include/net/dst.h

@@ -94,7 +94,6 @@ struct dst_ops
 	struct dst_entry *	(*negative_advice)(struct dst_entry *);
 	void			(*link_failure)(struct sk_buff *);
 	void			(*update_pmtu)(struct dst_entry *dst, u32 mtu);
-	int			(*get_mss)(struct dst_entry *dst, u32 mtu);
 	int			entry_size;
 
 	atomic_t		entries;

net/core/neighbour.c

@@ -1625,12 +1625,9 @@ static int neightbl_fill_info(struct neigh_table *tbl, struct sk_buff *skb,
 
 		memset(&ndst, 0, sizeof(ndst));
 
-		for (cpu = 0; cpu < NR_CPUS; cpu++) {
+		for_each_cpu(cpu) {
 			struct neigh_statistics	*st;
 
-			if (!cpu_possible(cpu))
-				continue;
-
 			st = per_cpu_ptr(tbl->stats, cpu);
 			ndst.ndts_allocs		+= st->allocs;
 			ndst.ndts_destroys		+= st->destroys;

net/core/skbuff.c

@@ -122,6 +122,8 @@ void skb_under_panic(struct sk_buff *skb, int sz, void *here)
  *	__alloc_skb	-	allocate a network buffer
  *	@size: size to allocate
  *	@gfp_mask: allocation mask
+ *	@fclone: allocate from fclone cache instead of head cache
+ *		and allocate a cloned (child) skb
  *
  *	Allocate a new &sk_buff. The returned buffer has no headroom and a
  *	tail room of size bytes. The object has a reference count of one.

net/decnet/af_decnet.c

@@ -719,22 +719,9 @@ static int dn_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
 	if (saddr->sdn_flags & ~SDF_WILD)
 		return -EINVAL;
 
-#if 1
 	if (!capable(CAP_NET_BIND_SERVICE) && (saddr->sdn_objnum ||
 	    (saddr->sdn_flags & SDF_WILD)))
 		return -EACCES;
-#else
-	/*
-	 * Maybe put the default actions in the default security ops for
-	 * dn_prot_sock ? Would be nice if the capable call would go there
-	 * too.
-	 */
-	if (security_dn_prot_sock(saddr) &&
-	    !capable(CAP_NET_BIND_SERVICE) || 
-	    saddr->sdn_objnum || (saddr->sdn_flags & SDF_WILD))
-		return -EACCES;
-#endif
-
 
 	if (!(saddr->sdn_flags & SDF_WILD)) {
 		if (dn_ntohs(saddr->sdn_nodeaddrl)) {

net/ipv4/devinet.c

@@ -715,6 +715,7 @@ int devinet_ioctl(unsigned int cmd, void __user *arg)
 			break;
 		ret = 0;
 		if (ifa->ifa_mask != sin->sin_addr.s_addr) {
+			u32 old_mask = ifa->ifa_mask;
 			inet_del_ifa(in_dev, ifap, 0);
 			ifa->ifa_mask = sin->sin_addr.s_addr;
 			ifa->ifa_prefixlen = inet_mask_len(ifa->ifa_mask);
@@ -728,7 +729,7 @@ int devinet_ioctl(unsigned int cmd, void __user *arg)
 			if ((dev->flags & IFF_BROADCAST) &&
 			    (ifa->ifa_prefixlen < 31) &&
 			    (ifa->ifa_broadcast ==
-			     (ifa->ifa_local|~ifa->ifa_mask))) {
+			     (ifa->ifa_local|~old_mask))) {
 				ifa->ifa_broadcast = (ifa->ifa_local |
 						      ~sin->sin_addr.s_addr);
 			}

net/ipv4/fib_trie.c

@@ -2404,7 +2404,7 @@ static int fib_route_seq_show(struct seq_file *seq, void *v)
 		prefix = htonl(l->key);
 
 		list_for_each_entry_rcu(fa, &li->falh, fa_list) {
-			const struct fib_info *fi = rcu_dereference(fa->fa_info);
+			const struct fib_info *fi = fa->fa_info;
 			unsigned flags = fib_flag_trans(fa->fa_type, mask, fi);
 
 			if (fa->fa_type == RTN_BROADCAST

net/ipv4/icmp.c

@@ -1108,12 +1108,9 @@ void __init icmp_init(struct net_proto_family *ops)
 	struct inet_sock *inet;
 	int i;
 
-	for (i = 0; i < NR_CPUS; i++) {
+	for_each_cpu(i) {
 		int err;
 
-		if (!cpu_possible(i))
-			continue;
-
 		err = sock_create_kern(PF_INET, SOCK_RAW, IPPROTO_ICMP,
 				       &per_cpu(__icmp_socket, i));
 

net/ipv4/ip_output.c

@@ -1023,10 +1023,7 @@ ssize_t	ip_append_page(struct sock *sk, struct page *page,
 			int alloclen;
 
 			skb_prev = skb;
-			if (skb_prev)
-				fraggap = skb_prev->len - maxfraglen;
-			else
-				fraggap = 0;
+			fraggap = skb_prev->len - maxfraglen;
 
 			alloclen = fragheaderlen + hh_len + fraggap + 15;
 			skb = sock_wmalloc(sk, alloclen, 1, sk->sk_allocation);

net/ipv4/netfilter/ip_conntrack_core.c

@@ -50,7 +50,7 @@
 #include <linux/netfilter_ipv4/ip_conntrack_core.h>
 #include <linux/netfilter_ipv4/listhelp.h>
 
-#define IP_CONNTRACK_VERSION	"2.3"
+#define IP_CONNTRACK_VERSION	"2.4"
 
 #if 0
 #define DEBUGP printk
@@ -148,16 +148,20 @@ DEFINE_PER_CPU(struct ip_conntrack_stat, ip_conntrack_stat);
 static int ip_conntrack_hash_rnd_initted;
 static unsigned int ip_conntrack_hash_rnd;
 
-static u_int32_t
-hash_conntrack(const struct ip_conntrack_tuple *tuple)
+static u_int32_t __hash_conntrack(const struct ip_conntrack_tuple *tuple,
+			    unsigned int size, unsigned int rnd)
 {
-#if 0
-	dump_tuple(tuple);
-#endif
 	return (jhash_3words(tuple->src.ip,
 	                     (tuple->dst.ip ^ tuple->dst.protonum),
 	                     (tuple->src.u.all | (tuple->dst.u.all << 16)),
-	                     ip_conntrack_hash_rnd) % ip_conntrack_htable_size);
+	                     rnd) % size);
+}
+
+static u_int32_t
+hash_conntrack(const struct ip_conntrack_tuple *tuple)
+{
+	return __hash_conntrack(tuple, ip_conntrack_htable_size,
+				ip_conntrack_hash_rnd);
 }
 
 int
@@ -1341,14 +1345,13 @@ static int kill_all(struct ip_conntrack *i, void *data)
 	return 1;
 }
 
-static void free_conntrack_hash(void)
+static void free_conntrack_hash(struct list_head *hash, int vmalloced,int size)
 {
-	if (ip_conntrack_vmalloc)
-		vfree(ip_conntrack_hash);
+	if (vmalloced)
+		vfree(hash);
 	else
-		free_pages((unsigned long)ip_conntrack_hash, 
-			   get_order(sizeof(struct list_head)
-				     * ip_conntrack_htable_size));
+		free_pages((unsigned long)hash, 
+			   get_order(sizeof(struct list_head) * size));
 }
 
 void ip_conntrack_flush()
@@ -1378,12 +1381,83 @@ void ip_conntrack_cleanup(void)
 	ip_conntrack_flush();
 	kmem_cache_destroy(ip_conntrack_cachep);
 	kmem_cache_destroy(ip_conntrack_expect_cachep);
-	free_conntrack_hash();
+	free_conntrack_hash(ip_conntrack_hash, ip_conntrack_vmalloc,
+			    ip_conntrack_htable_size);
 	nf_unregister_sockopt(&so_getorigdst);
 }
 
-static int hashsize;
-module_param(hashsize, int, 0400);
+static struct list_head *alloc_hashtable(int size, int *vmalloced)
+{
+	struct list_head *hash;
+	unsigned int i;
+
+	*vmalloced = 0; 
+	hash = (void*)__get_free_pages(GFP_KERNEL, 
+				       get_order(sizeof(struct list_head)
+						 * size));
+	if (!hash) { 
+		*vmalloced = 1;
+		printk(KERN_WARNING"ip_conntrack: falling back to vmalloc.\n");
+		hash = vmalloc(sizeof(struct list_head) * size);
+	}
+
+	if (hash)
+		for (i = 0; i < size; i++)
+			INIT_LIST_HEAD(&hash[i]);
+
+	return hash;
+}
+
+int set_hashsize(const char *val, struct kernel_param *kp)
+{
+	int i, bucket, hashsize, vmalloced;
+	int old_vmalloced, old_size;
+	int rnd;
+	struct list_head *hash, *old_hash;
+	struct ip_conntrack_tuple_hash *h;
+
+	/* On boot, we can set this without any fancy locking. */
+	if (!ip_conntrack_htable_size)
+		return param_set_int(val, kp);
+
+	hashsize = simple_strtol(val, NULL, 0);
+	if (!hashsize)
+		return -EINVAL;
+
+	hash = alloc_hashtable(hashsize, &vmalloced);
+	if (!hash)
+		return -ENOMEM;
+
+	/* We have to rehash for the new table anyway, so we also can 
+	 * use a new random seed */
+	get_random_bytes(&rnd, 4);
+
+	write_lock_bh(&ip_conntrack_lock);
+	for (i = 0; i < ip_conntrack_htable_size; i++) {
+		while (!list_empty(&ip_conntrack_hash[i])) {
+			h = list_entry(ip_conntrack_hash[i].next,
+				       struct ip_conntrack_tuple_hash, list);
+			list_del(&h->list);
+			bucket = __hash_conntrack(&h->tuple, hashsize, rnd);
+			list_add_tail(&h->list, &hash[bucket]);
+		}
+	}
+	old_size = ip_conntrack_htable_size;
+	old_vmalloced = ip_conntrack_vmalloc;
+	old_hash = ip_conntrack_hash;
+
+	ip_conntrack_htable_size = hashsize;
+	ip_conntrack_vmalloc = vmalloced;
+	ip_conntrack_hash = hash;
+	ip_conntrack_hash_rnd = rnd;
+	write_unlock_bh(&ip_conntrack_lock);
+
+	free_conntrack_hash(old_hash, old_vmalloced, old_size);
+	return 0;
+}
+
+module_param_call(hashsize, set_hashsize, param_get_uint,
+		  &ip_conntrack_htable_size, 0600);
 
 int __init ip_conntrack_init(void)
 {
@@ -1392,9 +1466,7 @@ int __init ip_conntrack_init(void)
 
 	/* Idea from tcp.c: use 1/16384 of memory.  On i386: 32MB
 	 * machine has 256 buckets.  >= 1GB machines have 8192 buckets. */
- 	if (hashsize) {
- 		ip_conntrack_htable_size = hashsize;
- 	} else {
+ 	if (!ip_conntrack_htable_size) {
 		ip_conntrack_htable_size
 			= (((num_physpages << PAGE_SHIFT) / 16384)
 			   / sizeof(struct list_head));
@@ -1416,20 +1488,8 @@ int __init ip_conntrack_init(void)
 		return ret;
 	}
 
-	/* AK: the hash table is twice as big than needed because it
-	   uses list_head.  it would be much nicer to caches to use a
-	   single pointer list head here. */
-	ip_conntrack_vmalloc = 0; 
-	ip_conntrack_hash 
-		=(void*)__get_free_pages(GFP_KERNEL, 
-					 get_order(sizeof(struct list_head)
-						   *ip_conntrack_htable_size));
-	if (!ip_conntrack_hash) { 
-		ip_conntrack_vmalloc = 1;
-		printk(KERN_WARNING "ip_conntrack: falling back to vmalloc.\n");
-		ip_conntrack_hash = vmalloc(sizeof(struct list_head)
-					    * ip_conntrack_htable_size);
-	}
+	ip_conntrack_hash = alloc_hashtable(ip_conntrack_htable_size,
+					    &ip_conntrack_vmalloc);
 	if (!ip_conntrack_hash) {
 		printk(KERN_ERR "Unable to create ip_conntrack_hash\n");
 		goto err_unreg_sockopt;
@@ -1461,9 +1521,6 @@ int __init ip_conntrack_init(void)
 	ip_ct_protos[IPPROTO_ICMP] = &ip_conntrack_protocol_icmp;
 	write_unlock_bh(&ip_conntrack_lock);
 
-	for (i = 0; i < ip_conntrack_htable_size; i++)
-		INIT_LIST_HEAD(&ip_conntrack_hash[i]);
-
 	/* For use by ipt_REJECT */
 	ip_ct_attach = ip_conntrack_attach;
 
@@ -1478,7 +1535,8 @@ int __init ip_conntrack_init(void)
 err_free_conntrack_slab:
 	kmem_cache_destroy(ip_conntrack_cachep);
 err_free_hash:
-	free_conntrack_hash();
+	free_conntrack_hash(ip_conntrack_hash, ip_conntrack_vmalloc,
+			    ip_conntrack_htable_size);
 err_unreg_sockopt:
 	nf_unregister_sockopt(&so_getorigdst);
 

net/ipv4/proc.c

@@ -90,9 +90,7 @@ fold_field(void *mib[], int offt)
 	unsigned long res = 0;
 	int i;
 
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!cpu_possible(i))
-			continue;
+	for_each_cpu(i) {
 		res += *(((unsigned long *) per_cpu_ptr(mib[0], i)) + offt);
 		res += *(((unsigned long *) per_cpu_ptr(mib[1], i)) + offt);
 	}

net/ipv6/icmp.c

@@ -700,10 +700,7 @@ int __init icmpv6_init(struct net_proto_family *ops)
 	struct sock *sk;
 	int err, i, j;
 
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!cpu_possible(i))
-			continue;
-
+	for_each_cpu(i) {
 		err = sock_create_kern(PF_INET6, SOCK_RAW, IPPROTO_ICMPV6,
 				       &per_cpu(__icmpv6_socket, i));
 		if (err < 0) {
@@ -749,9 +746,7 @@ void icmpv6_cleanup(void)
 {
 	int i;
 
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!cpu_possible(i))
-			continue;
+	for_each_cpu(i) {
 		sock_release(per_cpu(__icmpv6_socket, i));
 	}
 	inet6_del_protocol(&icmpv6_protocol, IPPROTO_ICMPV6);

net/ipv6/proc.c

@@ -140,9 +140,7 @@ fold_field(void *mib[], int offt)
         unsigned long res = 0;
         int i;
  
-        for (i = 0; i < NR_CPUS; i++) {
-                if (!cpu_possible(i))
-                        continue;
+        for_each_cpu(i) {
                 res += *(((unsigned long *)per_cpu_ptr(mib[0], i)) + offt);
                 res += *(((unsigned long *)per_cpu_ptr(mib[1], i)) + offt);
         }

net/netlink/af_netlink.c

@@ -740,11 +740,8 @@ int netlink_attachskb(struct sock *sk, struct sk_buff *skb, int nonblock, long t
 
 int netlink_sendskb(struct sock *sk, struct sk_buff *skb, int protocol)
 {
-	struct netlink_sock *nlk;
 	int len = skb->len;
 
-	nlk = nlk_sk(sk);
-
 	skb_queue_tail(&sk->sk_receive_queue, skb);
 	sk->sk_data_ready(sk, len);
 	sock_put(sk);

net/rose/rose_route.c

@@ -727,7 +727,7 @@ int rose_rt_ioctl(unsigned int cmd, void __user *arg)
 		}
 		if (rose_route.mask > 10) /* Mask can't be more than 10 digits */
 			return -EINVAL;
-		if (rose_route.ndigis > 8) /* No more than 8 digipeats */
+		if (rose_route.ndigis > AX25_MAX_DIGIS)
 			return -EINVAL;
 		err = rose_add_node(&rose_route, dev);
 		dev_put(dev);

net/sctp/proc.c

@@ -69,9 +69,7 @@ fold_field(void *mib[], int nr)
 	unsigned long res = 0;
 	int i;
 
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!cpu_possible(i))
-			continue;
+	for_each_cpu(i) {
 		res +=
 		    *((unsigned long *) (((void *) per_cpu_ptr(mib[0], i)) +
 					 sizeof (unsigned long) * nr));

net/xfrm/xfrm_policy.c

@@ -1192,46 +1192,6 @@ int xfrm_bundle_ok(struct xfrm_dst *first, struct flowi *fl, int family)
 
 EXPORT_SYMBOL(xfrm_bundle_ok);
 
-/* Well... that's _TASK_. We need to scan through transformation
- * list and figure out what mss tcp should generate in order to
- * final datagram fit to mtu. Mama mia... :-)
- *
- * Apparently, some easy way exists, but we used to choose the most
- * bizarre ones. :-) So, raising Kalashnikov... tra-ta-ta.
- *
- * Consider this function as something like dark humour. :-)
- */
-static int xfrm_get_mss(struct dst_entry *dst, u32 mtu)
-{
-	int res = mtu - dst->header_len;
-
-	for (;;) {
-		struct dst_entry *d = dst;
-		int m = res;
-
-		do {
-			struct xfrm_state *x = d->xfrm;
-			if (x) {
-				spin_lock_bh(&x->lock);
-				if (x->km.state == XFRM_STATE_VALID &&
-				    x->type && x->type->get_max_size)
-					m = x->type->get_max_size(d->xfrm, m);
-				else
-					m += x->props.header_len;
-				spin_unlock_bh(&x->lock);
-			}
-		} while ((d = d->child) != NULL);
-
-		if (m <= mtu)
-			break;
-		res -= (m - mtu);
-		if (res < 88)
-			return mtu;
-	}
-
-	return res + dst->header_len;
-}
-
 int xfrm_policy_register_afinfo(struct xfrm_policy_afinfo *afinfo)
 {
 	int err = 0;
@@ -1252,8 +1212,6 @@ int xfrm_policy_register_afinfo(struct xfrm_policy_afinfo *afinfo)
 			dst_ops->negative_advice = xfrm_negative_advice;
 		if (likely(dst_ops->link_failure == NULL))
 			dst_ops->link_failure = xfrm_link_failure;
-		if (likely(dst_ops->get_mss == NULL))
-			dst_ops->get_mss = xfrm_get_mss;
 		if (likely(afinfo->garbage_collect == NULL))
 			afinfo->garbage_collect = __xfrm_garbage_collect;
 		xfrm_policy_afinfo[afinfo->family] = afinfo;
@@ -1281,7 +1239,6 @@ int xfrm_policy_unregister_afinfo(struct xfrm_policy_afinfo *afinfo)
 			dst_ops->check = NULL;
 			dst_ops->negative_advice = NULL;
 			dst_ops->link_failure = NULL;
-			dst_ops->get_mss = NULL;
 			afinfo->garbage_collect = NULL;
 		}
 	}

net/xfrm/xfrm_state.c

@@ -1026,6 +1026,12 @@ void xfrm_state_delete_tunnel(struct xfrm_state *x)
 }
 EXPORT_SYMBOL(xfrm_state_delete_tunnel);
 
+/*
+ * This function is NOT optimal.  For example, with ESP it will give an
+ * MTU that's usually two bytes short of being optimal.  However, it will
+ * usually give an answer that's a multiple of 4 provided the input is
+ * also a multiple of 4.
+ */
 int xfrm_state_mtu(struct xfrm_state *x, int mtu)
 {
 	int res = mtu;