Re: pasta: set correct flags in reset message when outgoing connection setup fails

[Stefano Brivio <sbrivio@redhat.com>]

On Sat, 25 Jan 2025 15:16:47 -0500
Jon Maloy <jmaloy@redhat.com> wrote:

> Bug #93 states that a connect() will hang if the remote port of a
> connection setup doesn't exist. The remote host will respond with
> the (RST | SYN | ACK) flags set, but we only relay the RST flag back
> to the local user. This causes the situation described above.

Thanks for looking into this.

> We now add a check if the reset happens in the connection state
> TAP_SYN_RCVD, in which case we add the missing (SYN | ACK) flags.
> 
> This solves the problem.

I guess that, actually, this part of the problem (or the whole bug #93?
I'm not sure) was solved by commit db2c91ae86c7 ("tcp: Set ACK flag on
*all* RST segments, even for client in SYN-SENT state").

> Signed-off-by: Jon Maloy <jmaloy@redhat.com>
> ---
>  tcp.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/tcp.c b/tcp.c
> index c89f323..46df6bd 100644
> --- a/tcp.c
> +++ b/tcp.c
> @@ -1199,9 +1199,14 @@ static int tcp_send_flag(const struct ctx *c, struct tcp_tap_conn *conn,
>   */
>  void tcp_rst_do(const struct ctx *c, struct tcp_tap_conn *conn)
>  {
> +	int flags = RST;
> +
>  	if (conn->events == CLOSED)
>  		return;
>  
> +	if (conn->events == TAP_SYN_RCVD)
> +		flags |= SYN | ACK;
> +

...flags is not used, though. Could it be that you tested this after
rebasing on top of db2c91ae86c7?

>  	if (!tcp_send_flag(c, conn, RST))
>  		conn_event(c, conn, CLOSED);
>  }

-- 
Stefano