Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/gi…

…t/kaber/nf-next-2.6

include/linux/list_nulls.h

@@ -56,6 +56,18 @@ static inline int hlist_nulls_empty(const struct hlist_nulls_head *h)
 	return is_a_nulls(h->first);
 }
 
+static inline void hlist_nulls_add_head(struct hlist_nulls_node *n,
+					struct hlist_nulls_head *h)
+{
+	struct hlist_nulls_node *first = h->first;
+
+	n->next = first;
+	n->pprev = &h->first;
+	h->first = n;
+	if (!is_a_nulls(first))
+		first->pprev = &n->next;
+}
+
 static inline void __hlist_nulls_del(struct hlist_nulls_node *n)
 {
 	struct hlist_nulls_node *next = n->next;
@@ -65,6 +77,12 @@ static inline void __hlist_nulls_del(struct hlist_nulls_node *n)
 		next->pprev = pprev;
 }
 
+static inline void hlist_nulls_del(struct hlist_nulls_node *n)
+{
+	__hlist_nulls_del(n);
+	n->pprev = LIST_POISON2;
+}
+
 /**
  * hlist_nulls_for_each_entry	- iterate over list of given type
  * @tpos:	the type * to use as a loop cursor.

include/net/netfilter/nf_conntrack.h

@@ -201,6 +201,8 @@ extern struct nf_conntrack_tuple_hash *
 __nf_conntrack_find(struct net *net, const struct nf_conntrack_tuple *tuple);
 
 extern void nf_conntrack_hash_insert(struct nf_conn *ct);
+extern void nf_ct_delete_from_lists(struct nf_conn *ct);
+extern void nf_ct_insert_dying_list(struct nf_conn *ct);
 
 extern void nf_conntrack_flush_report(struct net *net, u32 pid, int report);
 

include/net/netfilter/nf_conntrack_ecache.h

@@ -6,61 +6,54 @@
 #define _NF_CONNTRACK_ECACHE_H
 #include <net/netfilter/nf_conntrack.h>
 
-#include <linux/interrupt.h>
 #include <net/net_namespace.h>
 #include <net/netfilter/nf_conntrack_expect.h>
+#include <linux/netfilter/nf_conntrack_common.h>
+#include <linux/netfilter/nf_conntrack_tuple_common.h>
+#include <net/netfilter/nf_conntrack_extend.h>
 
-/* Connection tracking event bits */
+/* Connection tracking event types */
 enum ip_conntrack_events
 {
-	/* New conntrack */
-	IPCT_NEW_BIT = 0,
-	IPCT_NEW = (1 << IPCT_NEW_BIT),
-
-	/* Expected connection */
-	IPCT_RELATED_BIT = 1,
-	IPCT_RELATED = (1 << IPCT_RELATED_BIT),
-
-	/* Destroyed conntrack */
-	IPCT_DESTROY_BIT = 2,
-	IPCT_DESTROY = (1 << IPCT_DESTROY_BIT),
-
-	/* Status has changed */
-	IPCT_STATUS_BIT = 3,
-	IPCT_STATUS = (1 << IPCT_STATUS_BIT),
+	IPCT_NEW		= 0,	/* new conntrack */
+	IPCT_RELATED		= 1,	/* related conntrack */
+	IPCT_DESTROY		= 2,	/* destroyed conntrack */
+	IPCT_STATUS		= 3,	/* status has changed */
+	IPCT_PROTOINFO		= 4,	/* protocol information has changed */
+	IPCT_HELPER		= 5,	/* new helper has been set */
+	IPCT_MARK		= 6,	/* new mark has been set */
+	IPCT_NATSEQADJ		= 7,	/* NAT is doing sequence adjustment */
+	IPCT_SECMARK		= 8,	/* new security mark has been set */
+};
 
-	/* Update of protocol info */
-	IPCT_PROTOINFO_BIT = 4,
-	IPCT_PROTOINFO = (1 << IPCT_PROTOINFO_BIT),
+enum ip_conntrack_expect_events {
+	IPEXP_NEW		= 0,	/* new expectation */
+};
 
-	/* New helper for conntrack */
-	IPCT_HELPER_BIT = 5,
-	IPCT_HELPER = (1 << IPCT_HELPER_BIT),
+struct nf_conntrack_ecache {
+	unsigned long cache;		/* bitops want long */
+	unsigned long missed;		/* missed events */
+	u32 pid;			/* netlink pid of destroyer */
+};
 
-	/* Mark is set */
-	IPCT_MARK_BIT = 6,
-	IPCT_MARK = (1 << IPCT_MARK_BIT),
+static inline struct nf_conntrack_ecache *
+nf_ct_ecache_find(const struct nf_conn *ct)
+{
+	return nf_ct_ext_find(ct, NF_CT_EXT_ECACHE);
+}
 
-	/* NAT sequence adjustment */
-	IPCT_NATSEQADJ_BIT = 7,
-	IPCT_NATSEQADJ = (1 << IPCT_NATSEQADJ_BIT),
+static inline struct nf_conntrack_ecache *
+nf_ct_ecache_ext_add(struct nf_conn *ct, gfp_t gfp)
+{
+	struct net *net = nf_ct_net(ct);
 
-	/* Secmark is set */
-	IPCT_SECMARK_BIT = 8,
-	IPCT_SECMARK = (1 << IPCT_SECMARK_BIT),
-};
+	if (!net->ct.sysctl_events)
+		return NULL;
 
-enum ip_conntrack_expect_events {
-	IPEXP_NEW_BIT = 0,
-	IPEXP_NEW = (1 << IPEXP_NEW_BIT),
+	return nf_ct_ext_add(ct, NF_CT_EXT_ECACHE, gfp);
 };
 
 #ifdef CONFIG_NF_CONNTRACK_EVENTS
-struct nf_conntrack_ecache {
-	struct nf_conn *ct;
-	unsigned int events;
-};
-
 /* This structure is passed to event handler */
 struct nf_ct_event {
 	struct nf_conn *ct;
@@ -76,53 +69,88 @@ extern struct nf_ct_event_notifier *nf_conntrack_event_cb;
 extern int nf_conntrack_register_notifier(struct nf_ct_event_notifier *nb);
 extern void nf_conntrack_unregister_notifier(struct nf_ct_event_notifier *nb);
 
-extern void nf_ct_deliver_cached_events(const struct nf_conn *ct);
-extern void __nf_ct_event_cache_init(struct nf_conn *ct);
-extern void nf_ct_event_cache_flush(struct net *net);
+extern void nf_ct_deliver_cached_events(struct nf_conn *ct);
 
 static inline void
 nf_conntrack_event_cache(enum ip_conntrack_events event, struct nf_conn *ct)
 {
-	struct net *net = nf_ct_net(ct);
-	struct nf_conntrack_ecache *ecache;
-
-	local_bh_disable();
-	ecache = per_cpu_ptr(net->ct.ecache, raw_smp_processor_id());
-	if (ct != ecache->ct)
-		__nf_ct_event_cache_init(ct);
-	ecache->events |= event;
-	local_bh_enable();
+	struct nf_conntrack_ecache *e;
+
+	if (nf_conntrack_event_cb == NULL)
+		return;
+
+	e = nf_ct_ecache_find(ct);
+	if (e == NULL)
+		return;
+
+	set_bit(event, &e->cache);
 }
 
-static inline void
-nf_conntrack_event_report(enum ip_conntrack_events event,
-			  struct nf_conn *ct,
-			  u32 pid,
-			  int report)
+static inline int
+nf_conntrack_eventmask_report(unsigned int eventmask,
+			      struct nf_conn *ct,
+			      u32 pid,
+			      int report)
 {
+	int ret = 0;
+	struct net *net = nf_ct_net(ct);
 	struct nf_ct_event_notifier *notify;
+	struct nf_conntrack_ecache *e;
 
 	rcu_read_lock();
 	notify = rcu_dereference(nf_conntrack_event_cb);
 	if (notify == NULL)
 		goto out_unlock;
 
+	if (!net->ct.sysctl_events)
+		goto out_unlock;
+
+	e = nf_ct_ecache_find(ct);
+	if (e == NULL)
+		goto out_unlock;
+
 	if (nf_ct_is_confirmed(ct) && !nf_ct_is_dying(ct)) {
 		struct nf_ct_event item = {
 			.ct 	= ct,
-			.pid	= pid,
+			.pid	= e->pid ? e->pid : pid,
 			.report = report
 		};
-		notify->fcn(event, &item);
+		/* This is a resent of a destroy event? If so, skip missed */
+		unsigned long missed = e->pid ? 0 : e->missed;
+
+		ret = notify->fcn(eventmask | missed, &item);
+		if (unlikely(ret < 0 || missed)) {
+			spin_lock_bh(&ct->lock);
+			if (ret < 0) {
+				/* This is a destroy event that has been
+				 * triggered by a process, we store the PID
+				 * to include it in the retransmission. */
+				if (eventmask & (1 << IPCT_DESTROY) &&
+				    e->pid == 0 && pid != 0)
+					e->pid = pid;
+				else
+					e->missed |= eventmask;
+			} else
+				e->missed &= ~missed;
+			spin_unlock_bh(&ct->lock);
+		}
 	}
 out_unlock:
 	rcu_read_unlock();
+	return ret;
 }
 
-static inline void
+static inline int
+nf_conntrack_event_report(enum ip_conntrack_events event, struct nf_conn *ct,
+			  u32 pid, int report)
+{
+	return nf_conntrack_eventmask_report(1 << event, ct, pid, report);
+}
+
+static inline int
 nf_conntrack_event(enum ip_conntrack_events event, struct nf_conn *ct)
 {
-	nf_conntrack_event_report(event, ct, 0, 0);
+	return nf_conntrack_eventmask_report(1 << event, ct, 0, 0);
 }
 
 struct nf_exp_event {
@@ -145,20 +173,24 @@ nf_ct_expect_event_report(enum ip_conntrack_expect_events event,
 			  u32 pid,
 			  int report)
 {
+	struct net *net = nf_ct_exp_net(exp);
 	struct nf_exp_event_notifier *notify;
 
 	rcu_read_lock();
 	notify = rcu_dereference(nf_expect_event_cb);
 	if (notify == NULL)
 		goto out_unlock;
 
+	if (!net->ct.sysctl_events)
+		goto out_unlock;
+
 	{
 		struct nf_exp_event item = {
 			.exp	= exp,
 			.pid	= pid,
 			.report = report
 		};
-		notify->fcn(event, &item);
+		notify->fcn(1 << event, &item);
 	}
 out_unlock:
 	rcu_read_unlock();
@@ -178,20 +210,23 @@ extern void nf_conntrack_ecache_fini(struct net *net);
 
 static inline void nf_conntrack_event_cache(enum ip_conntrack_events event,
 					    struct nf_conn *ct) {}
-static inline void nf_conntrack_event(enum ip_conntrack_events event,
-				      struct nf_conn *ct) {}
-static inline void nf_conntrack_event_report(enum ip_conntrack_events event,
-					     struct nf_conn *ct,
-					     u32 pid,
-					     int report) {}
+static inline int nf_conntrack_eventmask_report(unsigned int eventmask,
+						struct nf_conn *ct,
+						u32 pid,
+						int report) { return 0; }
+static inline int nf_conntrack_event(enum ip_conntrack_events event,
+				     struct nf_conn *ct) { return 0; }
+static inline int nf_conntrack_event_report(enum ip_conntrack_events event,
+					    struct nf_conn *ct,
+					    u32 pid,
+					    int report) { return 0; }
 static inline void nf_ct_deliver_cached_events(const struct nf_conn *ct) {}
 static inline void nf_ct_expect_event(enum ip_conntrack_expect_events event,
 				      struct nf_conntrack_expect *exp) {}
 static inline void nf_ct_expect_event_report(enum ip_conntrack_expect_events e,
 					     struct nf_conntrack_expect *exp,
  					     u32 pid,
  					     int report) {}
-static inline void nf_ct_event_cache_flush(struct net *net) {}
 
 static inline int nf_conntrack_ecache_init(struct net *net)
 {

include/net/netfilter/nf_conntrack_extend.h

@@ -8,12 +8,14 @@ enum nf_ct_ext_id
 	NF_CT_EXT_HELPER,
 	NF_CT_EXT_NAT,
 	NF_CT_EXT_ACCT,
+	NF_CT_EXT_ECACHE,
 	NF_CT_EXT_NUM,
 };
 
 #define NF_CT_EXT_HELPER_TYPE struct nf_conn_help
 #define NF_CT_EXT_NAT_TYPE struct nf_conn_nat
 #define NF_CT_EXT_ACCT_TYPE struct nf_conn_counter
+#define NF_CT_EXT_ECACHE_TYPE struct nf_conntrack_ecache
 
 /* Extensions: optional stuff which isn't permanently in struct. */
 struct nf_ct_ext {

include/net/netfilter/nf_conntrack_helper.h

@@ -50,6 +50,8 @@ extern struct nf_conn_help *nf_ct_helper_ext_add(struct nf_conn *ct, gfp_t gfp);
 
 extern int __nf_ct_try_assign_helper(struct nf_conn *ct, gfp_t flags);
 
+extern void nf_ct_helper_destroy(struct nf_conn *ct);
+
 static inline struct nf_conn_help *nfct_help(const struct nf_conn *ct)
 {
 	return nf_ct_ext_find(ct, NF_CT_EXT_HELPER);

include/net/netns/conntrack.h

@@ -14,16 +14,17 @@ struct netns_ct {
 	struct hlist_nulls_head	*hash;
 	struct hlist_head	*expect_hash;
 	struct hlist_nulls_head	unconfirmed;
+	struct hlist_nulls_head	dying;
 	struct ip_conntrack_stat *stat;
-#ifdef CONFIG_NF_CONNTRACK_EVENTS
-	struct nf_conntrack_ecache *ecache;
-#endif
+	int			sysctl_events;
+	unsigned int		sysctl_events_retry_timeout;
 	int			sysctl_acct;
 	int			sysctl_checksum;
 	unsigned int		sysctl_log_invalid; /* Log invalid packets */
 #ifdef CONFIG_SYSCTL
 	struct ctl_table_header	*sysctl_header;
 	struct ctl_table_header	*acct_sysctl_header;
+	struct ctl_table_header	*event_sysctl_header;
 #endif
 	int			hash_vmalloc;
 	int			expect_vmalloc;