Commit 5f923853 authored by Eric Dumazet's avatar Eric Dumazet Committed by Jakub Kicinski
Browse files

tcp: fix __tcp_close() to only send RST when required 



If the receive queue contains payload that was already
received, __tcp_close() can send an unexpected RST.

Refine the code to take tp->copied_seq into account,
as we already do in tcp recvmsg().

Fixes: 1da177e4 ("Linux-2.6.12-rc2")
Signed-off-by: default avatarEric Dumazet <edumazet@google.com>
Reviewed-by: default avatarNeal Cardwell <ncardwell@google.com>
Reviewed-by: default avatarKuniyuki Iwashima <kuniyu@google.com>
Reviewed-by: default avatarJason Xing <kerneljasonxing@gmail.com>
Link: https://patch.msgid.link/20250903084720.1168904-2-edumazet@google.com


Signed-off-by: default avatarJakub Kicinski <kuba@kernel.org>
parent 69777753

net/ipv4/tcp.c

+5 −4
Original line number Diff line number Diff line
@@ -3099,8 +3099,8 @@ bool tcp_check_oom(const struct sock *sk, int shift) 


void __tcp_close(struct sock *sk, long timeout)

{

	bool data_was_unread = false;

	struct sk_buff *skb;

	int data_was_unread = 0;

	int state;



	WRITE_ONCE(sk->sk_shutdown, SHUTDOWN_MASK);
@@ -3119,11 +3119,12 @@ void __tcp_close(struct sock *sk, long timeout) 
	 *  reader process may not have drained the data yet!

	 */

	while ((skb = __skb_dequeue(&sk->sk_receive_queue)) != NULL) {

		u32 len = TCP_SKB_CB(skb)->end_seq - TCP_SKB_CB(skb)->seq;

		u32 end_seq = TCP_SKB_CB(skb)->end_seq;



		if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN)

			len--;

		data_was_unread += len;

			end_seq--;

		if (after(end_seq, tcp_sk(sk)->copied_seq))

			data_was_unread = true;

		__kfree_skb(skb);

	}