This reverts the patch: when add_addr_timeout is set to 1, chk_add_nr()
and chk_add_tx_nr() are more tolerant with retransmissions, see commit
6ef84b1517e0 ("selftests: mptcp: more robust signal race test").

But the main reason is that this revert breaks the selftests running on
old kernels.

It is then important to keep this. Plus it still acts as a maximum, just
in case everything is very slow, we still have retransmissions in time,
especially with the new exponential backoff.

Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
---
 tools/testing/selftests/net/mptcp/mptcp_join.sh | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/tools/testing/selftests/net/mptcp/mptcp_join.sh b/tools/testing/selftests/net/mptcp/mptcp_join.sh
index a97b568104bc284f050b2f0e09fe3fdd3341c5cb..2f046167a0b6cc6fb5531a033d8d95c9ea399cf9 100755
--- a/tools/testing/selftests/net/mptcp/mptcp_join.sh
+++ b/tools/testing/selftests/net/mptcp/mptcp_join.sh
@@ -358,6 +358,8 @@ reset_with_add_addr_timeout()
 		tables="${ip6tables}"
 	fi
 
+	ip netns exec $ns1 sysctl -q net.mptcp.add_addr_timeout=1
+
 	if ! ip netns exec $ns2 $tables -A OUTPUT -p tcp \
 			-m tcp --tcp-option 30 \
 			-m bpf --bytecode \
@@ -2303,6 +2305,7 @@ signal_address_tests()
 		pm_nl_add_endpoint $ns2 10.0.4.2 flags signal
 
 		# the peer could possibly miss some addr notification, allow retransmission
+		ip netns exec $ns1 sysctl -q net.mptcp.add_addr_timeout=1
 		speed=slow \
 			run_tests $ns1 $ns2 10.0.1.1
 

-- 
2.51.0

Hi Matt,

On Tue, 2025-09-02 at 21:01 +0200, Matthieu Baerts (NGI0) wrote:
> This reverts the patch: when add_addr_timeout is set to 1,
> chk_add_nr()
> and chk_add_tx_nr() are more tolerant with retransmissions, see
> commit
> 6ef84b1517e0 ("selftests: mptcp: more robust signal race test").
> 
> But the main reason is that this revert breaks the selftests running
> on
> old kernels.
> 
> It is then important to keep this. Plus it still acts as a maximum,
> just
> in case everything is very slow, we still have retransmissions in
> time,
> especially with the new exponential backoff.
> 
> Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>

I agree.

Reviewed-by: Geliang Tang <geliang@kernel.org>

Thanks,
-Geliang

> ---
>  tools/testing/selftests/net/mptcp/mptcp_join.sh | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/tools/testing/selftests/net/mptcp/mptcp_join.sh
> b/tools/testing/selftests/net/mptcp/mptcp_join.sh
> index
> a97b568104bc284f050b2f0e09fe3fdd3341c5cb..2f046167a0b6cc6fb5531a033d8
> d95c9ea399cf9 100755
> --- a/tools/testing/selftests/net/mptcp/mptcp_join.sh
> +++ b/tools/testing/selftests/net/mptcp/mptcp_join.sh
> @@ -358,6 +358,8 @@ reset_with_add_addr_timeout()
>  		tables="${ip6tables}"
>  	fi
>  
> +	ip netns exec $ns1 sysctl -q net.mptcp.add_addr_timeout=1
> +
>  	if ! ip netns exec $ns2 $tables -A OUTPUT -p tcp \
>  			-m tcp --tcp-option 30 \
>  			-m bpf --bytecode \
> @@ -2303,6 +2305,7 @@ signal_address_tests()
>  		pm_nl_add_endpoint $ns2 10.0.4.2 flags signal
>  
>  		# the peer could possibly miss some addr
> notification, allow retransmission
> +		ip netns exec $ns1 sysctl -q
> net.mptcp.add_addr_timeout=1
>  		speed=slow \
>  			run_tests $ns1 $ns2 10.0.1.1
>