[NETFILTER]: conntrack: fix race condition in early_drop

On SMP environments the maximum number of conntracks can be overpassed
under heavy stress situations due to an existing race condition.

        CPU A                   CPU B
     atomic_read()               ...
     early_drop()                ...
        ...                  atomic_read()
   allocate conntrack      allocate conntrack
     atomic_inc()             atomic_inc()

This patch moves the counter incrementation before the early drop stage.

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

1 files changed, 6 insertions, 3 deletions

diff --git a/net/ipv4/netfilter/ip_conntrack_core.c b/net/ipv4/netfilter/ip_conntrack_core.c
index 2568d480e9a9..422a662194cc 100644
--- a/net/ipv4/netfilter/ip_conntrack_core.c
+++ b/net/ipv4/netfilter/ip_conntrack_core.c
@@ -622,11 +622,15 @@ struct ip_conntrack *ip_conntrack_alloc(struct ip_conntrack_tuple *orig,
                 ip_conntrack_hash_rnd_initted = 1;
         }
 
+        /* We don't want any race condition at early drop stage */
+        atomic_inc(&ip_conntrack_count);
+
         if (ip_conntrack_max
-            && atomic_read(&ip_conntrack_count) >= ip_conntrack_max) {
+            && atomic_read(&ip_conntrack_count) > ip_conntrack_max) {
                 unsigned int hash = hash_conntrack(orig);
                 /* Try dropping from this hash chain. */
                 if (!early_drop(&ip_conntrack_hash[hash])) {
+                        atomic_dec(&ip_conntrack_count);
                         if (net_ratelimit())
                                 printk(KERN_WARNING
                                        "ip_conntrack: table full, dropping"
@@ -638,6 +642,7 @@ struct ip_conntrack *ip_conntrack_alloc(struct ip_conntrack_tuple *orig,
         conntrack = kmem_cache_alloc(ip_conntrack_cachep, GFP_ATOMIC);
         if (!conntrack) {
                 DEBUGP("Can't allocate conntrack.\n");
+                atomic_dec(&ip_conntrack_count);
                 return ERR_PTR(-ENOMEM);
         }
 
@@ -651,8 +656,6 @@ struct ip_conntrack *ip_conntrack_alloc(struct ip_conntrack_tuple *orig,
         conntrack->timeout.data = (unsigned long)conntrack;
         conntrack->timeout.function = death_by_timeout;
 
-        atomic_inc(&ip_conntrack_count);
-
         return conntrack;
 }