tcp: Add stats for socket migration.

This commit adds two stats for the socket migration feature to evaluate the
effectiveness: LINUX_MIB_TCPMIGRATEREQ(SUCCESS|FAILURE).

If the migration fails because of the own_req race in receiving ACK and
sending SYN+ACK paths, we do not increment the failure stat. Then another
CPU is responsible for the req.

Link: https://lore.kernel.org/bpf/CAK6E8=[e-mail]/
Suggested-by: Yuchung Cheng <[e-mail]>
Signed-off-by: Kuniyuki Iwashima <[e-mail]>
Acked-by: Yuchung Cheng <[e-mail]>
Signed-off-by: David S. Miller <[e-mail]>

include/uapi/linux/snmp.h

···
 	LINUX_MIB_TCPDUPLICATEDATAREHASH,	/* TCPDuplicateDataRehash */
 	LINUX_MIB_TCPDSACKRECVSEGS,		/* TCPDSACKRecvSegs */
 	LINUX_MIB_TCPDSACKIGNOREDDUBIOUS,	/* TCPDSACKIgnoredDubious */
+	LINUX_MIB_TCPMIGRATEREQSUCCESS,		/* TCPMigrateReqSuccess */
+	LINUX_MIB_TCPMIGRATEREQFAILURE,		/* TCPMigrateReqFailure */
 	__LINUX_MIB_MAX
 };
 

net/core/sock_reuseport.c

···
  * selecting the socket index from the array of available sockets.
  */
 
+#include <net/ip.h>
 #include <net/sock_reuseport.h>
 #include <linux/bpf.h>
 #include <linux/idr.h>
···
 
 	socks = READ_ONCE(reuse->num_socks);
 	if (unlikely(!socks))
-		goto out;
+		goto failure;
 
 	/* paired with smp_wmb() in __reuseport_add_sock() */
 	smp_rmb();
···
 	if (!prog || prog->expected_attach_type != BPF_SK_REUSEPORT_SELECT_OR_MIGRATE) {
 		if (sock_net(sk)->ipv4.sysctl_tcp_migrate_req)
 			goto select_by_hash;
-		goto out;
+		goto failure;
 	}
 
 	if (!skb) {
 		skb = alloc_skb(0, GFP_ATOMIC);
 		if (!skb)
-			goto out;
+			goto failure;
 		allocated = true;
 	}
 
···
 	if (!nsk)
 		nsk = reuseport_select_sock_by_hash(reuse, hash, socks);
 
-	if (IS_ERR_OR_NULL(nsk) || unlikely(!refcount_inc_not_zero(&nsk->sk_refcnt)))
+	if (IS_ERR_OR_NULL(nsk) || unlikely(!refcount_inc_not_zero(&nsk->sk_refcnt))) {
 		nsk = NULL;
+		goto failure;
+	}
 
 out:
 	rcu_read_unlock();
 	return nsk;
+
+failure:
+	__NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMIGRATEREQFAILURE);
+	goto out;
 }
 EXPORT_SYMBOL(reuseport_migrate_sock);
 

net/ipv4/inet_connection_sock.c

···
 
 	nreq = kmem_cache_alloc(req->rsk_ops->slab, GFP_ATOMIC | __GFP_NOWARN);
 	if (!nreq) {
+		__NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMIGRATEREQFAILURE);
+
 		/* paired with refcount_inc_not_zero() in reuseport_migrate_sock() */
 		sock_put(sk);
 		return NULL;
···
 		if (!inet_ehash_insert(req_to_sk(nreq), req_to_sk(oreq), NULL)) {
 			/* delete timer */
 			inet_csk_reqsk_queue_drop(sk_listener, nreq);
-			goto drop;
+			goto no_ownership;
 		}
 
+		__NET_INC_STATS(net, LINUX_MIB_TCPMIGRATEREQSUCCESS);
 		reqsk_migrate_reset(oreq);
 		reqsk_queue_removed(&inet_csk(oreq->rsk_listener)->icsk_accept_queue, oreq);
 		reqsk_put(oreq);
···
 		return;
 	}
 
-drop:
 	/* Even if we can clone the req, we may need not retransmit any more
 	 * SYN+ACKs (nreq->num_timeout > max_syn_ack_retries, etc), or another
 	 * CPU may win the "own_req" race so that inet_ehash_insert() fails.
 	 */
 	if (nreq) {
+		__NET_INC_STATS(net, LINUX_MIB_TCPMIGRATEREQFAILURE);
+no_ownership:
 		reqsk_migrate_reset(nreq);
 		reqsk_queue_removed(queue, nreq);
 		__reqsk_free(nreq);
 	}
 
+drop:
 	inet_csk_reqsk_queue_drop_and_put(oreq->rsk_listener, oreq);
 }
 
···
 
 			refcount_set(&nreq->rsk_refcnt, 1);
 			if (inet_csk_reqsk_queue_add(sk, nreq, child)) {
+				__NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMIGRATEREQSUCCESS);
 				reqsk_migrate_reset(req);
 				reqsk_put(req);
 				return child;
 			}
 
+			__NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMIGRATEREQFAILURE);
 			reqsk_migrate_reset(nreq);
 			__reqsk_free(nreq);
 		} else if (inet_csk_reqsk_queue_add(sk, req, child)) {
···
 				refcount_set(&nreq->rsk_refcnt, 1);
 
 				if (inet_csk_reqsk_queue_add(nsk, nreq, child)) {
+					__NET_INC_STATS(sock_net(nsk),
+							LINUX_MIB_TCPMIGRATEREQSUCCESS);
 					reqsk_migrate_reset(req);
 				} else {
+					__NET_INC_STATS(sock_net(nsk),
+							LINUX_MIB_TCPMIGRATEREQFAILURE);
 					reqsk_migrate_reset(nreq);
 					__reqsk_free(nreq);
 				}

net/ipv4/proc.c

···
 	SNMP_MIB_ITEM("TcpDuplicateDataRehash", LINUX_MIB_TCPDUPLICATEDATAREHASH),
 	SNMP_MIB_ITEM("TCPDSACKRecvSegs", LINUX_MIB_TCPDSACKRECVSEGS),
 	SNMP_MIB_ITEM("TCPDSACKIgnoredDubious", LINUX_MIB_TCPDSACKIGNOREDDUBIOUS),
+	SNMP_MIB_ITEM("TCPMigrateReqSuccess", LINUX_MIB_TCPMIGRATEREQSUCCESS),
+	SNMP_MIB_ITEM("TCPMigrateReqFailure", LINUX_MIB_TCPMIGRATEREQFAILURE),
 	SNMP_MIB_SENTINEL
 };
 

net/ipv4/tcp_minisocks.c

···
 	return inet_csk_complete_hashdance(sk, child, req, own_req);
 
 listen_overflow:
+	if (sk != req->rsk_listener)
+		__NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMIGRATEREQFAILURE);
+
 	if (!sock_net(sk)->ipv4.sysctl_tcp_abort_on_overflow) {
 		inet_rsk(req)->acked = 1;
 		return NULL;