tcp: annotate tp->copied_seq lockless reads

There are few places where we fetch tp->copied_seq while
this field can change from IRQ or other cpu.

We need to add READ_ONCE() annotations, and also make
sure write sides use corresponding WRITE_ONCE() to avoid
store-tearing.

Note that tcp_inq_hint() was already using READ_ONCE(tp->copied_seq)

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

6 files changed, 18 insertions, 17 deletions

diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 883ee863db43..c322ad071e17 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -477,7 +477,7 @@ static void tcp_tx_timestamp(struct sock *sk, u16 tsflags)
 static inline bool tcp_stream_is_readable(const struct tcp_sock *tp,
 					  int target, struct sock *sk)
 {
-	return (READ_ONCE(tp->rcv_nxt) - tp->copied_seq >= target) ||
+	return (READ_ONCE(tp->rcv_nxt) - READ_ONCE(tp->copied_seq) >= target) ||
 		(sk->sk_prot->stream_memory_read ?
 		sk->sk_prot->stream_memory_read(sk) : false);
 }
@@ -546,7 +546,7 @@ __poll_t tcp_poll(struct file *file, struct socket *sock, poll_table *wait)
 	    (state != TCP_SYN_RECV || rcu_access_pointer(tp->fastopen_rsk))) {
 		int target = sock_rcvlowat(sk, 0, INT_MAX);
 
-		if (tp->urg_seq == tp->copied_seq &&
+		if (tp->urg_seq == READ_ONCE(tp->copied_seq) &&
 		    !sock_flag(sk, SOCK_URGINLINE) &&
 		    tp->urg_data)
 			target++;
@@ -607,7 +607,7 @@ int tcp_ioctl(struct sock *sk, int cmd, unsigned long arg)
 		unlock_sock_fast(sk, slow);
 		break;
 	case SIOCATMARK:
-		answ = tp->urg_data && tp->urg_seq == tp->copied_seq;
+		answ = tp->urg_data && tp->urg_seq == READ_ONCE(tp->copied_seq);
 		break;
 	case SIOCOUTQ:
 		if (sk->sk_state == TCP_LISTEN)
@@ -1668,9 +1668,9 @@ int tcp_read_sock(struct sock *sk, read_descriptor_t *desc,
 		sk_eat_skb(sk, skb);
 		if (!desc->count)
 			break;
-		tp->copied_seq = seq;
+		WRITE_ONCE(tp->copied_seq, seq);
 	}
-	tp->copied_seq = seq;
+	WRITE_ONCE(tp->copied_seq, seq);
 
 	tcp_rcv_space_adjust(sk);
 
@@ -1819,7 +1819,7 @@ static int tcp_zerocopy_receive(struct sock *sk,
 out:
 	up_read(&current->mm->mmap_sem);
 	if (length) {
-		tp->copied_seq = seq;
+		WRITE_ONCE(tp->copied_seq, seq);
 		tcp_rcv_space_adjust(sk);
 
 		/* Clean up data we have read: This will do ACK frames. */
@@ -2117,7 +2117,7 @@ found_ok_skb:
 			if (urg_offset < used) {
 				if (!urg_offset) {
 					if (!sock_flag(sk, SOCK_URGINLINE)) {
-						++*seq;
+						WRITE_ONCE(*seq, *seq + 1);
 						urg_hole++;
 						offset++;
 						used--;
@@ -2139,7 +2139,7 @@ found_ok_skb:
 			}
 		}
 
-		*seq += used;
+		WRITE_ONCE(*seq, *seq + used);
 		copied += used;
 		len -= used;
 
@@ -2166,7 +2166,7 @@ skip_copy:
 
 found_fin_ok:
 		/* Process the FIN. */
-		++*seq;
+		WRITE_ONCE(*seq, *seq + 1);
 		if (!(flags & MSG_PEEK))
 			sk_eat_skb(sk, skb);
 		break;
@@ -2588,7 +2588,7 @@ int tcp_disconnect(struct sock *sk, int flags)
 		__kfree_skb(sk->sk_rx_skb_cache);
 		sk->sk_rx_skb_cache = NULL;
 	}
-	tp->copied_seq = tp->rcv_nxt;
+	WRITE_ONCE(tp->copied_seq, tp->rcv_nxt);
 	tp->urg_data = 0;
 	tcp_write_queue_purge(sk);
 	tcp_fastopen_active_disable_ofo_check(sk);
diff --git a/net/ipv4/tcp_diag.c b/net/ipv4/tcp_diag.c
index cd219161f106..66273c8a55c2 100644
--- a/net/ipv4/tcp_diag.c
+++ b/net/ipv4/tcp_diag.c
@@ -26,7 +26,8 @@ static void tcp_diag_get_info(struct sock *sk, struct inet_diag_msg *r,
 	} else if (sk->sk_type == SOCK_STREAM) {
 		const struct tcp_sock *tp = tcp_sk(sk);
 
-		r->idiag_rqueue = max_t(int, READ_ONCE(tp->rcv_nxt) - tp->copied_seq, 0);
+		r->idiag_rqueue = max_t(int, READ_ONCE(tp->rcv_nxt) -
+					     READ_ONCE(tp->copied_seq), 0);
 		r->idiag_wqueue = tp->write_seq - tp->snd_una;
 	}
 	if (info)
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 5b7c8768ed5f..a30aae3a6a18 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -5961,7 +5961,7 @@ static int tcp_rcv_synsent_state_process(struct sock *sk, struct sk_buff *skb,
 		/* Remember, tcp_poll() does not lock socket!
 		 * Change state from SYN-SENT only after copied_seq
 		 * is initialized. */
-		tp->copied_seq = tp->rcv_nxt;
+		WRITE_ONCE(tp->copied_seq, tp->rcv_nxt);
 
 		smc_check_reset_syn(tp);
 
@@ -6036,7 +6036,7 @@ discard:
 		}
 
 		WRITE_ONCE(tp->rcv_nxt, TCP_SKB_CB(skb)->seq + 1);
-		tp->copied_seq = tp->rcv_nxt;
+		WRITE_ONCE(tp->copied_seq, tp->rcv_nxt);
 		tp->rcv_wup = TCP_SKB_CB(skb)->seq + 1;
 
 		/* RFC1323: The window in SYN & SYN/ACK segments is
@@ -6216,7 +6216,7 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb)
 			tcp_try_undo_spurious_syn(sk);
 			tp->retrans_stamp = 0;
 			tcp_init_transfer(sk, BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB);
-			tp->copied_seq = tp->rcv_nxt;
+			WRITE_ONCE(tp->copied_seq, tp->rcv_nxt);
 		}
 		smp_mb();
 		tcp_set_state(sk, TCP_ESTABLISHED);
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 5089dd6bee0f..39560f482e0b 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -2456,7 +2456,7 @@ static void get_tcp4_sock(struct sock *sk, struct seq_file *f, int i)
 		 * we might find a transient negative value.
 		 */
 		rx_queue = max_t(int, READ_ONCE(tp->rcv_nxt) -
-				      tp->copied_seq, 0);
+				      READ_ONCE(tp->copied_seq), 0);
 
 	seq_printf(f, "%4d: %08X:%04X %08X:%04X %02X %08X:%08X %02X:%08lX "
 			"%08X %5u %8d %lu %d %pK %lu %lu %u %u %d",
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index adc6ce486a38..c4731d26ab4a 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -478,7 +478,7 @@ struct sock *tcp_create_openreq_child(const struct sock *sk,
 
 	seq = treq->rcv_isn + 1;
 	newtp->rcv_wup = seq;
-	newtp->copied_seq = seq;
+	WRITE_ONCE(newtp->copied_seq, seq);
 	WRITE_ONCE(newtp->rcv_nxt, seq);
 	newtp->segs_in = 1;
 
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 84ae4d1449ea..7dda12720169 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -3433,7 +3433,7 @@ static void tcp_connect_init(struct sock *sk)
 	else
 		tp->rcv_tstamp = tcp_jiffies32;
 	tp->rcv_wup = tp->rcv_nxt;
-	tp->copied_seq = tp->rcv_nxt;
+	WRITE_ONCE(tp->copied_seq, tp->rcv_nxt);
 
 	inet_csk(sk)->icsk_rto = tcp_timeout_init(sk);
 	inet_csk(sk)->icsk_retransmits = 0;