packet: rollover statistics

Rollover indicates exceptional conditions. Export a counter to inform
socket owners of this state.

If no socket with sufficient room is found, rollover fails. Also count
these events.

Finally, also count when flows are rolled over early thanks to huge
flow detection, to validate its correctness.

Tested:
  Read counters in bench_rollover on all other tests in the patchset

Signed-off-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

include/uapi/linux/if_packet.h

@@ -54,6 +54,7 @@ struct sockaddr_ll {
 #define PACKET_FANOUT			18
 #define PACKET_TX_HAS_OFF		19
 #define PACKET_QDISC_BYPASS		20
+#define PACKET_ROLLOVER_STATS		21
 
 #define PACKET_FANOUT_HASH		0
 #define PACKET_FANOUT_LB		1
@@ -75,6 +76,12 @@ struct tpacket_stats_v3 {
 	unsigned int	tp_freeze_q_cnt;
 };
 
+struct tpacket_rollover_stats {
+	__aligned_u64	tp_all;
+	__aligned_u64	tp_huge;
+	__aligned_u64	tp_failed;
+};
+
 union tpacket_stats_u {
 	struct tpacket_stats stats1;
 	struct tpacket_stats_v3 stats3;

net/packet/af_packet.c

@@ -1395,7 +1395,7 @@ static unsigned int fanout_demux_rollover(struct packet_fanout *f,
 					  unsigned int num)
 {
 	struct packet_sock *po, *po_next;
-	unsigned int i, j, room;
+	unsigned int i, j, room = ROOM_NONE;
 
 	po = pkt_sk(f->arr[idx]);
 
@@ -1413,6 +1413,9 @@ static unsigned int fanout_demux_rollover(struct packet_fanout *f,
 		    packet_rcv_has_room(po_next, skb) == ROOM_NORMAL) {
 			if (i != j)
 				po->rollover->sock = i;
+			atomic_long_inc(&po->rollover->num);
+			if (room == ROOM_LOW)
+				atomic_long_inc(&po->rollover->num_huge);
 			return i;
 		}
 
@@ -1420,6 +1423,7 @@ static unsigned int fanout_demux_rollover(struct packet_fanout *f,
 			i = 0;
 	} while (i != j);
 
+	atomic_long_inc(&po->rollover->num_failed);
 	return idx;
 }
 
@@ -1554,6 +1558,9 @@ static int fanout_add(struct sock *sk, u16 id, u16 type_flags)
 		po->rollover = kzalloc(sizeof(*po->rollover), GFP_KERNEL);
 		if (!po->rollover)
 			return -ENOMEM;
+		atomic_long_set(&po->rollover->num, 0);
+		atomic_long_set(&po->rollover->num_huge, 0);
+		atomic_long_set(&po->rollover->num_failed, 0);
 	}
 
 	mutex_lock(&fanout_mutex);
@@ -3584,6 +3591,7 @@ static int packet_getsockopt(struct socket *sock, int level, int optname,
 	struct packet_sock *po = pkt_sk(sk);
 	void *data = &val;
 	union tpacket_stats_u st;
+	struct tpacket_rollover_stats rstats;
 
 	if (level != SOL_PACKET)
 		return -ENOPROTOOPT;
@@ -3659,6 +3667,15 @@ static int packet_getsockopt(struct socket *sock, int level, int optname,
 			((u32)po->fanout->flags << 24)) :
 		       0);
 		break;
+	case PACKET_ROLLOVER_STATS:
+		if (!po->rollover)
+			return -EINVAL;
+		rstats.tp_all = atomic_long_read(&po->rollover->num);
+		rstats.tp_huge = atomic_long_read(&po->rollover->num_huge);
+		rstats.tp_failed = atomic_long_read(&po->rollover->num_failed);
+		data = &rstats;
+		lv = sizeof(rstats);
+		break;
 	case PACKET_TX_HAS_OFF:
 		val = po->tp_tx_has_off;
 		break;

net/packet/internal.h

@@ -89,6 +89,9 @@ struct packet_fanout {
 
 struct packet_rollover {
 	int			sock;
+	atomic_long_t		num;
+	atomic_long_t		num_huge;
+	atomic_long_t		num_failed;
 #define ROLLOVER_HLEN	(L1_CACHE_BYTES / sizeof(u32))
 	u32			history[ROLLOVER_HLEN] ____cacheline_aligned;
 } ____cacheline_aligned_in_smp;