netfilter: nf_conntrack: use atomic64 for accounting counters

We can use atomic64_t infrastructure to avoid taking a spinlock in fast
path, and remove inaccuracies while reading values in
ctnetlink_dump_counters() and connbytes_mt() on 32bit arches.

Suggested by Pablo.

Signed-off-by: Eric Dumazet <[e-mail]>
Signed-off-by: Pablo Neira Ayuso <[e-mail]>

include/net/netfilter/nf_conntrack_acct.h

···
 #include <net/netfilter/nf_conntrack_extend.h>
 
 struct nf_conn_counter {
-	u_int64_t packets;
-	u_int64_t bytes;
+	atomic64_t packets;
+	atomic64_t bytes;
 };
 
 static inline

net/netfilter/nf_conntrack_acct.c

···
 		return 0;
 
 	return seq_printf(s, "packets=%llu bytes=%llu ",
-			  (unsigned long long)acct[dir].packets,
-			  (unsigned long long)acct[dir].bytes);
+			  (unsigned long long)atomic64_read(&acct[dir].packets),
+			  (unsigned long long)atomic64_read(&acct[dir].bytes));
 };
 EXPORT_SYMBOL_GPL(seq_print_acct);
 

net/netfilter/nf_conntrack_core.c

···
 
 		acct = nf_conn_acct_find(ct);
 		if (acct) {
-			spin_lock_bh(&ct->lock);
-			acct[CTINFO2DIR(ctinfo)].packets++;
-			acct[CTINFO2DIR(ctinfo)].bytes += skb->len;
-			spin_unlock_bh(&ct->lock);
+			atomic64_inc(&acct[CTINFO2DIR(ctinfo)].packets);
+			atomic64_add(skb->len, &acct[CTINFO2DIR(ctinfo)].bytes);
 		}
 	}
 }
···
 
 		acct = nf_conn_acct_find(ct);
 		if (acct) {
-			spin_lock_bh(&ct->lock);
-			acct[CTINFO2DIR(ctinfo)].packets++;
-			acct[CTINFO2DIR(ctinfo)].bytes +=
-				skb->len - skb_network_offset(skb);
-			spin_unlock_bh(&ct->lock);
+			atomic64_inc(&acct[CTINFO2DIR(ctinfo)].packets);
+			atomic64_add(skb->len - skb_network_offset(skb),
+				     &acct[CTINFO2DIR(ctinfo)].bytes);
 		}
 	}
 

net/netfilter/nf_conntrack_netlink.c

···
 		goto nla_put_failure;
 
 	NLA_PUT_BE64(skb, CTA_COUNTERS_PACKETS,
-		     cpu_to_be64(acct[dir].packets));
+		     cpu_to_be64(atomic64_read(&acct[dir].packets)));
 	NLA_PUT_BE64(skb, CTA_COUNTERS_BYTES,
-		     cpu_to_be64(acct[dir].bytes));
+		     cpu_to_be64(atomic64_read(&acct[dir].bytes)));
 
 	nla_nest_end(skb, nest_count);
 
···
 				struct nf_conn_counter *acct;
 
 				acct = nf_conn_acct_find(ct);
-				if (acct)
-					memset(acct, 0, sizeof(struct nf_conn_counter[IP_CT_DIR_MAX]));
+				if (acct) {
+					atomic64_set(&acct[IP_CT_DIR_ORIGINAL].bytes, 0);
+					atomic64_set(&acct[IP_CT_DIR_ORIGINAL].packets, 0);
+					atomic64_set(&acct[IP_CT_DIR_REPLY].bytes, 0);
+					atomic64_set(&acct[IP_CT_DIR_REPLY].packets, 0);
+					}
 			}
 		}
 		if (cb->args[1]) {

net/netfilter/xt_connbytes.c

···
 	case XT_CONNBYTES_PKTS:
 		switch (sinfo->direction) {
 		case XT_CONNBYTES_DIR_ORIGINAL:
-			what = counters[IP_CT_DIR_ORIGINAL].packets;
+			what = atomic64_read(&counters[IP_CT_DIR_ORIGINAL].packets);
 			break;
 		case XT_CONNBYTES_DIR_REPLY:
-			what = counters[IP_CT_DIR_REPLY].packets;
+			what = atomic64_read(&counters[IP_CT_DIR_REPLY].packets);
 			break;
 		case XT_CONNBYTES_DIR_BOTH:
-			what = counters[IP_CT_DIR_ORIGINAL].packets;
-			what += counters[IP_CT_DIR_REPLY].packets;
+			what = atomic64_read(&counters[IP_CT_DIR_ORIGINAL].packets);
+			what += atomic64_read(&counters[IP_CT_DIR_REPLY].packets);
 			break;
 		}
 		break;
 	case XT_CONNBYTES_BYTES:
 		switch (sinfo->direction) {
 		case XT_CONNBYTES_DIR_ORIGINAL:
-			what = counters[IP_CT_DIR_ORIGINAL].bytes;
+			what = atomic64_read(&counters[IP_CT_DIR_ORIGINAL].bytes);
 			break;
 		case XT_CONNBYTES_DIR_REPLY:
-			what = counters[IP_CT_DIR_REPLY].bytes;
+			what = atomic64_read(&counters[IP_CT_DIR_REPLY].bytes);
 			break;
 		case XT_CONNBYTES_DIR_BOTH:
-			what = counters[IP_CT_DIR_ORIGINAL].bytes;
-			what += counters[IP_CT_DIR_REPLY].bytes;
+			what = atomic64_read(&counters[IP_CT_DIR_ORIGINAL].bytes);
+			what += atomic64_read(&counters[IP_CT_DIR_REPLY].bytes);
 			break;
 		}
 		break;
 	case XT_CONNBYTES_AVGPKT:
 		switch (sinfo->direction) {
 		case XT_CONNBYTES_DIR_ORIGINAL:
-			bytes = counters[IP_CT_DIR_ORIGINAL].bytes;
-			pkts  = counters[IP_CT_DIR_ORIGINAL].packets;
+			bytes = atomic64_read(&counters[IP_CT_DIR_ORIGINAL].bytes);
+			pkts  = atomic64_read(&counters[IP_CT_DIR_ORIGINAL].packets);
 			break;
 		case XT_CONNBYTES_DIR_REPLY:
-			bytes = counters[IP_CT_DIR_REPLY].bytes;
-			pkts  = counters[IP_CT_DIR_REPLY].packets;
+			bytes = atomic64_read(&counters[IP_CT_DIR_REPLY].bytes);
+			pkts  = atomic64_read(&counters[IP_CT_DIR_REPLY].packets);
 			break;
 		case XT_CONNBYTES_DIR_BOTH:
-			bytes = counters[IP_CT_DIR_ORIGINAL].bytes +
-				counters[IP_CT_DIR_REPLY].bytes;
-			pkts  = counters[IP_CT_DIR_ORIGINAL].packets +
-				counters[IP_CT_DIR_REPLY].packets;
+			bytes = atomic64_read(&counters[IP_CT_DIR_ORIGINAL].bytes) +
+				atomic64_read(&counters[IP_CT_DIR_REPLY].bytes);
+			pkts  = atomic64_read(&counters[IP_CT_DIR_ORIGINAL].packets) +
+				atomic64_read(&counters[IP_CT_DIR_REPLY].packets);
 			break;
 		}
 		if (pkts != 0)