Re: [PATCH mptcp-net v7 4/5] mptcp: avoid processing packet if a subflow reset

[Paolo Abeni <pabeni@redhat.com>]

Hi,

I'm sorry for the partial feedback in my previous reply.

On Fri, 2021-06-25 at 16:25 +0800, wujianguo106@163.com wrote:
> @@ -856,7 +856,8 @@ bool mptcp_synack_options(const struct request_sock *req, unsigned int *size,
>  static bool check_fully_established(struct mptcp_sock *msk, struct sock *ssk,
>  				    struct mptcp_subflow_context *subflow,
>  				    struct sk_buff *skb,
> -				    struct mptcp_options_received *mp_opt)
> +				    struct mptcp_options_received *mp_opt,
> +				    bool *subflow_is_rst)

This additional argument is not needed...

[...]
> @@ -1053,12 +1057,12 @@ void mptcp_incoming_options(struct sock *sk, struct sk_buff *skb)
>  			__mptcp_check_push(subflow->conn, sk);
>  		__mptcp_data_acked(subflow->conn);
>  		mptcp_data_unlock(subflow->conn);
> -		return;
> +		return 1;
>  	}
>  
>  	mptcp_get_options(sk, skb, &mp_opt);
> -	if (!check_fully_established(msk, sk, subflow, skb, &mp_opt))
> -		return;
> +	if (!check_fully_established(msk, sk, subflow, skb, &mp_opt, &subflow_is_rst))
> +		return subflow_is_rst ? 0 : 1;


... here you can simply:

		return sk->sk_state != TCP_CLOSE;

plus some comment above alike:

"""the subflow can be in close state only if check_fully_established()
just sent a reset. If so, tell the caller to ignore the current
packet"""

/P