Re: [MPTCP][PATCH mptcp-next 3/3] mptcp: send out MP_FAIL when data checksum fail

[Mat Martineau <mathew.j.martineau@linux.intel.com>]

On Thu, 6 May 2021, Geliang Tang wrote:

> When a bad checksum is detected, send out the MP_FAIL suboption.
>
> Signed-off-by: Geliang Tang <geliangtang@gmail.com>
> ---
> net/mptcp/subflow.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/net/mptcp/subflow.c b/net/mptcp/subflow.c
> index 659b8842ae3b..efd84ff61015 100644
> --- a/net/mptcp/subflow.c
> +++ b/net/mptcp/subflow.c
> @@ -906,6 +906,8 @@ static enum mapping_status validate_data_csum(struct sock *ssk, struct sk_buff *
> 	csum = csum_partial(&header, sizeof(header), subflow->map_data_csum);
> 	if (unlikely(csum_fold(csum))) {
> 		MPTCP_INC_STATS(sock_net(ssk), MPTCP_MIB_DSSCSUMERR);
> +		subflow->send_mp_fail = 1;
> +		subflow->fail_seq = subflow->map_seq;

map_seq is the sequence number of the last good data *received* by this 
peer, but fail_seq will be used to discard data *sent* by this peer. So I 
think fail_seq needs to be based on the sequence number for outgoing data 
on this subflow? I admit the RFC has me confused here. If MP_FAIL is 
implemented on the multipath-tcp.org kernel it might be helpful to see how 
it is used there.


> 		return subflow->mp_join ? MAPPING_INVALID : MAPPING_DUMMY;
> 	}
>
> -- 
> 2.31.1
>
>
>

--
Mat Martineau
Intel