netfilter: conntrack: move event caching to conntrack extension infrastructure

This patch reworks the per-cpu event caching to use the conntrack
extension infrastructure.

The main drawback is that we consume more memory per conntrack
if event delivery is enabled. This patch is required by the
reliable event delivery that follows to this patch.

BTW, this patch allows you to enable/disable event delivery via
/proc/sys/net/netfilter/nf_conntrack_events in runtime, although
you can still disable event caching as compilation option.

Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Patrick McHardy <kaber@trash.net>

3 files changed, 73 insertions, 64 deletions

diff --git a/include/net/netfilter/nf_conntrack_ecache.h b/include/net/netfilter/nf_conntrack_ecache.h
index 1afb907e015a..e7ae297ba383 100644
--- a/include/net/netfilter/nf_conntrack_ecache.h
+++ b/include/net/netfilter/nf_conntrack_ecache.h
@@ -6,61 +6,52 @@
 #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 */
+};
 
-        /* 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;
@@ -76,30 +67,30 @@ 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)
+nf_conntrack_eventmask_report(unsigned int eventmask,
+                              struct nf_conn *ct,
+                              u32 pid,
+                              int report)
 {
+        struct net *net = nf_ct_net(ct);
         struct nf_ct_event_notifier *notify;
 
         rcu_read_lock();
@@ -107,22 +98,32 @@ nf_conntrack_event_report(enum ip_conntrack_events event,
         if (notify == NULL)
                 goto out_unlock;
 
+        if (!net->ct.sysctl_events)
+                goto out_unlock;
+
         if (nf_ct_is_confirmed(ct) && !nf_ct_is_dying(ct)) {
                 struct nf_ct_event item = {
                         .ct     = ct,
                         .pid    = pid,
                         .report = report
                 };
-                notify->fcn(event, &item);
+                notify->fcn(eventmask, &item);
         }
 out_unlock:
         rcu_read_unlock();
 }
 
 static inline void
+nf_conntrack_event_report(enum ip_conntrack_events event, struct nf_conn *ct,
+                          u32 pid, int report)
+{
+        nf_conntrack_eventmask_report(1 << event, ct, pid, report);
+}
+
+static inline void
 nf_conntrack_event(enum ip_conntrack_events event, struct nf_conn *ct)
 {
-        nf_conntrack_event_report(event, ct, 0, 0);
+        nf_conntrack_eventmask_report(1 << event, ct, 0, 0);
 }
 
 struct nf_exp_event {
@@ -145,6 +146,7 @@ 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();
@@ -152,13 +154,16 @@ nf_ct_expect_event_report(enum ip_conntrack_expect_events event,
         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();
@@ -178,6 +183,10 @@ 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_eventmask_report(unsigned int eventmask,
+                                                 struct nf_conn *ct,
+                                                 u32 pid,
+                                                 int report) {}
 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,
@@ -191,7 +200,6 @@ 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)
 {
diff --git a/include/net/netfilter/nf_conntrack_extend.h b/include/net/netfilter/nf_conntrack_extend.h
index da8ee52613a5..7f8fc5d123c5 100644
--- a/include/net/netfilter/nf_conntrack_extend.h
+++ b/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 {
diff --git a/include/net/netns/conntrack.h b/include/net/netns/conntrack.h
index 9dc58402bc09..505a51cd8c63 100644
--- a/include/net/netns/conntrack.h
+++ b/include/net/netns/conntrack.h
@@ -15,15 +15,14 @@ struct netns_ct {
         struct hlist_head       *expect_hash;
         struct hlist_nulls_head unconfirmed;
         struct ip_conntrack_stat *stat;
-#ifdef CONFIG_NF_CONNTRACK_EVENTS
-        struct nf_conntrack_ecache *ecache;
-#endif
+        int                     sysctl_events;
         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;