1. Patchew
  2. MPTCP
  3. [PATCH mptcp-next v2 0/9] selftests: mptcp: skip the v6 subtests when CONFIG_MPTCP_IPV6=n
  4. View patch

[PATCH mptcp-next v2 5/9] selftests: mptcp: mptcp_sockopt.sh skips v6 paths when MPTCP IPv6 is missing

Gang Yan posted 9 patches 1 month, 3 weeks ago
[PATCH mptcp-next v2 5/9] selftests: mptcp: mptcp_sockopt.sh skips v6 paths when MPTCP IPv6 is missing
Posted by Gang Yan 1 month, 3 weeks ago 
From: Gang Yan <yangang@kylinos.cn>

Confine all of ipv6-related commands to mptcp_lib_is_v6_enabled:
  - call mptcp_lib_check_ipv6() once after the existing init-time
    mptcp_lib_check_*() calls to populate the cached flag;
  - guard the ip6tables loop in add_mark_rules() and every v6
    "ip addr add" / mptcp_lib_pm_nl_add_endpoint() in init();
  - in do_mptcp_sockopt_tests(), do_tcpinq_tests() and the top-level
    run_tests dispatcher, emit a SKIP TAP line for each v6 subtest
    instead of running it.

v4 transfers, v4 mark checks and the AF_INET sockopt / TCP_INQ tests
are untouched. In a full-featured kernel the new guards always take
their v6 branch and the test set runs as before.

Assisted-by: GLM:5.1 Z.ai
Signed-off-by: Gang Yan <yangang@kylinos.cn>
---
 .../selftests/net/mptcp/mptcp_sockopt.sh      | 94 +++++++++++++++----
 1 file changed, 75 insertions(+), 19 deletions(-)

diff --git a/tools/testing/selftests/net/mptcp/mptcp_sockopt.sh b/tools/testing/selftests/net/mptcp/mptcp_sockopt.sh
index e850a87429b6..ae7776b17c33 100755
--- a/tools/testing/selftests/net/mptcp/mptcp_sockopt.sh
+++ b/tools/testing/selftests/net/mptcp/mptcp_sockopt.sh
@@ -50,7 +50,7 @@ add_mark_rules()
 	local m=$2
 
 	local t
-	for t in ${iptables} ${ip6tables}; do
+	for t in ${iptables}; do
 		# just to debug: check we have multiple subflows connection requests
 		ip netns exec $ns $t -A OUTPUT -p tcp --syn -m mark --mark $m -j ACCEPT
 
@@ -60,6 +60,21 @@ add_mark_rules()
 		ip netns exec $ns $t -A OUTPUT -p tcp -m mark --mark $m -j ACCEPT
 		ip netns exec $ns $t -A OUTPUT -p tcp -m mark --mark 0 -j DROP
 	done
+	# ip6tables rules require MPTCP IPv6 support in the kernel, otherwise
+	# ip6tables will reject them (or, in some environments, the binary itself
+	# may behave oddly without IPv6 available).
+	if mptcp_lib_is_v6_enabled; then
+		for t in ${ip6tables}; do
+			# just to debug: check we have multiple subflows connection requests
+			ip netns exec $ns $t -A OUTPUT -p tcp --syn -m mark --mark $m -j ACCEPT
+
+			# RST packets might be handled by a internal dummy socket
+			ip netns exec $ns $t -A OUTPUT -p tcp --tcp-flags RST RST -m mark --mark 0 -j ACCEPT
+
+			ip netns exec $ns $t -A OUTPUT -p tcp -m mark --mark $m -j ACCEPT
+			ip netns exec $ns $t -A OUTPUT -p tcp -m mark --mark 0 -j DROP
+		done
+	fi
 }
 
 init()
@@ -70,21 +85,29 @@ init()
 	for i in $(seq 1 4); do
 		ip link add ns1eth$i netns "$ns1" type veth peer name ns2eth$i netns "$ns2"
 		ip -net "$ns1" addr add 10.0.$i.1/24 dev ns1eth$i
-		ip -net "$ns1" addr add dead:beef:$i::1/64 dev ns1eth$i nodad
+		if mptcp_lib_is_v6_enabled; then
+			ip -net "$ns1" addr add dead:beef:$i::1/64 dev ns1eth$i nodad
+		fi
 		ip -net "$ns1" link set ns1eth$i up
 
 		ip -net "$ns2" addr add 10.0.$i.2/24 dev ns2eth$i
-		ip -net "$ns2" addr add dead:beef:$i::2/64 dev ns2eth$i nodad
+		if mptcp_lib_is_v6_enabled; then
+			ip -net "$ns2" addr add dead:beef:$i::2/64 dev ns2eth$i nodad
+		fi
 		ip -net "$ns2" link set ns2eth$i up
 
 		# let $ns2 reach any $ns1 address from any interface
 		ip -net "$ns2" route add default via 10.0.$i.1 dev ns2eth$i metric 10$i
 
 		mptcp_lib_pm_nl_add_endpoint "${ns1}" "10.0.${i}.1" flags signal
-		mptcp_lib_pm_nl_add_endpoint "${ns1}" "dead:beef:${i}::1" flags signal
+		if mptcp_lib_is_v6_enabled; then
+			mptcp_lib_pm_nl_add_endpoint "${ns1}" "dead:beef:${i}::1" flags signal
+		fi
 
 		mptcp_lib_pm_nl_add_endpoint "${ns2}" "10.0.${i}.2" flags signal
-		mptcp_lib_pm_nl_add_endpoint "${ns2}" "dead:beef:${i}::2" flags signal
+		if mptcp_lib_is_v6_enabled; then
+			mptcp_lib_pm_nl_add_endpoint "${ns2}" "dead:beef:${i}::2" flags signal
+		fi
 	done
 
 	mptcp_lib_pm_nl_set_limits "${ns1}" 8 8
@@ -106,6 +129,7 @@ cleanup()
 mptcp_lib_check_mptcp
 mptcp_lib_check_kallsyms
 mptcp_lib_check_tools ip "${iptables}" "${ip6tables}"
+mptcp_lib_check_ipv6
 
 check_mark()
 {
@@ -164,6 +188,20 @@ do_transfer()
 		ip=ipv4
 	fi
 
+	# Skip v6 subtests when the running kernel has no MPTCP IPv6 support.
+	# Reuses the same ${ip}/${ip:2} naming as the normal path below so the
+	# console counter (MPTCP_LIB_TEST_COUNTER, bumped by print_title) and
+	# the TAP index (bumped by mptcp_lib_result_skip) stay in lock-step.
+	if [ "${ip}" = "ipv6" ] && ! mptcp_lib_is_v6_enabled; then
+		print_title "Transfer ${ip:2}"
+		mptcp_lib_pr_skip "MPTCP IPv6 not available"
+		mptcp_lib_result_skip "transfer ${ip}"
+		print_title "Mark ${ip:2}"
+		mptcp_lib_pr_skip "MPTCP IPv6 not available"
+		mptcp_lib_result_skip "mark ${ip}"
+		return 0
+	fi
+
 	cmsg="TIMESTAMPNS"
 	if mptcp_lib_kallsyms_has "mptcp_ioctl$"; then
 		cmsg+=",TCPINQ"
@@ -274,18 +312,23 @@ do_mptcp_sockopt_tests()
 	mptcp_lib_pr_ok
 	mptcp_lib_result_pass "sockopt v4"
 
-	ip netns exec "$ns_sbox" ./mptcp_sockopt -6
-	lret=$?
-
 	print_title "SOL_MPTCP sockopt v6"
-	if [ $lret -ne 0 ]; then
-		mptcp_lib_pr_fail
-		mptcp_lib_result_fail "sockopt v6"
-		ret=$lret
-		return
+	if ! mptcp_lib_is_v6_enabled; then
+		mptcp_lib_pr_skip "MPTCP IPv6 not available"
+		mptcp_lib_result_skip "sockopt v6"
+	else
+		ip netns exec "$ns_sbox" ./mptcp_sockopt -6
+		lret=$?
+
+		if [ $lret -ne 0 ]; then
+			mptcp_lib_pr_fail
+			mptcp_lib_result_fail "sockopt v6"
+			ret=$lret
+			return
+		fi
+		mptcp_lib_pr_ok
+		mptcp_lib_result_pass "sockopt v6"
 	fi
-	mptcp_lib_pr_ok
-	mptcp_lib_result_pass "sockopt v6"
 }
 
 run_tests()
@@ -308,6 +351,13 @@ run_tests()
 do_tcpinq_test()
 {
 	print_title "TCP_INQ cmsg/ioctl $*"
+
+	if [ "${1}" = "-6" ] && ! mptcp_lib_is_v6_enabled; then
+		mptcp_lib_pr_skip "MPTCP IPv6 not available"
+		mptcp_lib_result_skip "TCP_INQ: $*"
+		return 0
+	fi
+
 	ip netns exec "$ns_sbox" ./mptcp_inq "$@"
 	local lret=$?
 	if [ $lret -ne 0 ];then
@@ -339,10 +389,16 @@ do_tcpinq_tests()
 		if [ $lret -ne 0 ] ; then
 			return $lret
 		fi
-		do_tcpinq_test -6 $args
-		lret=$?
-		if [ $lret -ne 0 ] ; then
-			return $lret
+		if mptcp_lib_is_v6_enabled; then
+			do_tcpinq_test -6 $args
+			lret=$?
+			if [ $lret -ne 0 ] ; then
+				return $lret
+			fi
+		else
+			print_title "TCP_INQ cmsg/ioctl -6 $args"
+			mptcp_lib_pr_skip "MPTCP IPv6 not available"
+			mptcp_lib_result_skip "TCP_INQ: -6 $args"
 		fi
 	done
 
-- 
2.43.0
Re: [PATCH mptcp-next v2 5/9] selftests: mptcp: mptcp_sockopt.sh skips v6 paths when MPTCP IPv6 is missing
Posted by Matthieu Baerts 1 month, 3 weeks ago 
On 17/06/2026 16:46, Gang Yan wrote:
> From: Gang Yan <yangang@kylinos.cn>
> 
> Confine all of ipv6-related commands to mptcp_lib_is_v6_enabled:
>   - call mptcp_lib_check_ipv6() once after the existing init-time
>     mptcp_lib_check_*() calls to populate the cached flag;
>   - guard the ip6tables loop in add_mark_rules() and every v6
>     "ip addr add" / mptcp_lib_pm_nl_add_endpoint() in init();
>   - in do_mptcp_sockopt_tests(), do_tcpinq_tests() and the top-level
>     run_tests dispatcher, emit a SKIP TAP line for each v6 subtest
>     instead of running it.
> 
> v4 transfers, v4 mark checks and the AF_INET sockopt / TCP_INQ tests
> are untouched. In a full-featured kernel the new guards always take
> their v6 branch and the test set runs as before.
> 
> Assisted-by: GLM:5.1 Z.ai
> Signed-off-by: Gang Yan <yangang@kylinos.cn>
> ---
>  .../selftests/net/mptcp/mptcp_sockopt.sh      | 94 +++++++++++++++----
>  1 file changed, 75 insertions(+), 19 deletions(-)
> 
> diff --git a/tools/testing/selftests/net/mptcp/mptcp_sockopt.sh b/tools/testing/selftests/net/mptcp/mptcp_sockopt.sh
> index e850a87429b6..ae7776b17c33 100755
> --- a/tools/testing/selftests/net/mptcp/mptcp_sockopt.sh
> +++ b/tools/testing/selftests/net/mptcp/mptcp_sockopt.sh
> @@ -50,7 +50,7 @@ add_mark_rules()
>  	local m=$2
>  
>  	local t
> -	for t in ${iptables} ${ip6tables}; do
> +	for t in ${iptables}; do

Instead, you could keep most of the code here, and have something like:

  if [ ${t} = ${ip6tables} ] && ! mptcp_lib_is_v6_enabled; then
      continue
  fi

>  		# just to debug: check we have multiple subflows connection requests
>  		ip netns exec $ns $t -A OUTPUT -p tcp --syn -m mark --mark $m -j ACCEPT
>  
> @@ -60,6 +60,21 @@ add_mark_rules()
>  		ip netns exec $ns $t -A OUTPUT -p tcp -m mark --mark $m -j ACCEPT
>  		ip netns exec $ns $t -A OUTPUT -p tcp -m mark --mark 0 -j DROP
>  	done
> +	# ip6tables rules require MPTCP IPv6 support in the kernel, otherwise
> +	# ip6tables will reject them (or, in some environments, the binary itself
> +	# may behave oddly without IPv6 available).

I'm not sure to understand here: what's the link between ip6tables and
MPTCP IPv6 support? Or only check if the binary is there. Still it looks
like this modification is not needed.

> +	if mptcp_lib_is_v6_enabled; then
> +		for t in ${ip6tables}; do
> +			# just to debug: check we have multiple subflows connection requests
> +			ip netns exec $ns $t -A OUTPUT -p tcp --syn -m mark --mark $m -j ACCEPT
> +
> +			# RST packets might be handled by a internal dummy socket
> +			ip netns exec $ns $t -A OUTPUT -p tcp --tcp-flags RST RST -m mark --mark 0 -j ACCEPT
> +
> +			ip netns exec $ns $t -A OUTPUT -p tcp -m mark --mark $m -j ACCEPT
> +			ip netns exec $ns $t -A OUTPUT -p tcp -m mark --mark 0 -j DROP
> +		done
> +	fi
>  }
>  
>  init()
> @@ -70,21 +85,29 @@ init()
>  	for i in $(seq 1 4); do
>  		ip link add ns1eth$i netns "$ns1" type veth peer name ns2eth$i netns "$ns2"
>  		ip -net "$ns1" addr add 10.0.$i.1/24 dev ns1eth$i
> -		ip -net "$ns1" addr add dead:beef:$i::1/64 dev ns1eth$i nodad
> +		if mptcp_lib_is_v6_enabled; then
> +			ip -net "$ns1" addr add dead:beef:$i::1/64 dev ns1eth$i nodad
> +		fi
>  		ip -net "$ns1" link set ns1eth$i up
>  
>  		ip -net "$ns2" addr add 10.0.$i.2/24 dev ns2eth$i
> -		ip -net "$ns2" addr add dead:beef:$i::2/64 dev ns2eth$i nodad
> +		if mptcp_lib_is_v6_enabled; then
> +			ip -net "$ns2" addr add dead:beef:$i::2/64 dev ns2eth$i nodad
> +		fi
>  		ip -net "$ns2" link set ns2eth$i up
>  
>  		# let $ns2 reach any $ns1 address from any interface
>  		ip -net "$ns2" route add default via 10.0.$i.1 dev ns2eth$i metric 10$i
>  
>  		mptcp_lib_pm_nl_add_endpoint "${ns1}" "10.0.${i}.1" flags signal
> -		mptcp_lib_pm_nl_add_endpoint "${ns1}" "dead:beef:${i}::1" flags signal
> +		if mptcp_lib_is_v6_enabled; then
> +			mptcp_lib_pm_nl_add_endpoint "${ns1}" "dead:beef:${i}::1" flags signal
> +		fi
>  
>  		mptcp_lib_pm_nl_add_endpoint "${ns2}" "10.0.${i}.2" flags signal
> -		mptcp_lib_pm_nl_add_endpoint "${ns2}" "dead:beef:${i}::2" flags signal
> +		if mptcp_lib_is_v6_enabled; then
> +			mptcp_lib_pm_nl_add_endpoint "${ns2}" "dead:beef:${i}::2" flags signal
> +		fi

Maybe better to regroup all commands under on "if v6".

>  	done
>  
>  	mptcp_lib_pm_nl_set_limits "${ns1}" 8 8
> @@ -106,6 +129,7 @@ cleanup()
>  mptcp_lib_check_mptcp
>  mptcp_lib_check_kallsyms
>  mptcp_lib_check_tools ip "${iptables}" "${ip6tables}"
> +mptcp_lib_check_ipv6
>  
>  check_mark()
>  {
> @@ -164,6 +188,20 @@ do_transfer()
>  		ip=ipv4
>  	fi
>  
> +	# Skip v6 subtests when the running kernel has no MPTCP IPv6 support.
> +	# Reuses the same ${ip}/${ip:2} naming as the normal path below so the
> +	# console counter (MPTCP_LIB_TEST_COUNTER, bumped by print_title) and
> +	# the TAP index (bumped by mptcp_lib_result_skip) stay in lock-step.
> +	if [ "${ip}" = "ipv6" ] && ! mptcp_lib_is_v6_enabled; then
> +		print_title "Transfer ${ip:2}"
> +		mptcp_lib_pr_skip "MPTCP IPv6 not available"
> +		mptcp_lib_result_skip "transfer ${ip}"
> +		print_title "Mark ${ip:2}"
> +		mptcp_lib_pr_skip "MPTCP IPv6 not available"
> +		mptcp_lib_result_skip "mark ${ip}"
> +		return 0
> +	fi
> +
>  	cmsg="TIMESTAMPNS"
>  	if mptcp_lib_kallsyms_has "mptcp_ioctl$"; then
>  		cmsg+=",TCPINQ"
> @@ -274,18 +312,23 @@ do_mptcp_sockopt_tests()
>  	mptcp_lib_pr_ok
>  	mptcp_lib_result_pass "sockopt v4"
>  
> -	ip netns exec "$ns_sbox" ./mptcp_sockopt -6
> -	lret=$?
> -
>  	print_title "SOL_MPTCP sockopt v6"
> -	if [ $lret -ne 0 ]; then
> -		mptcp_lib_pr_fail
> -		mptcp_lib_result_fail "sockopt v6"
> -		ret=$lret
> -		return
> +	if ! mptcp_lib_is_v6_enabled; then
> +		mptcp_lib_pr_skip "MPTCP IPv6 not available"
> +		mptcp_lib_result_skip "sockopt v6"
> +	else
> +		ip netns exec "$ns_sbox" ./mptcp_sockopt -6
> +		lret=$?
> +
> +		if [ $lret -ne 0 ]; then
> +			mptcp_lib_pr_fail
> +			mptcp_lib_result_fail "sockopt v6"
> +			ret=$lret
> +			return
> +		fi
> +		mptcp_lib_pr_ok
> +		mptcp_lib_result_pass "sockopt v6"
>  	fi
> -	mptcp_lib_pr_ok
> -	mptcp_lib_result_pass "sockopt v6"

Maybe this chunk can be reduced?

  if ! mptcp_lib_is_v6_enabled; then
     skip
     return
  fi

and the rest is not modified, no?
>  }
>  
>  run_tests()
> @@ -308,6 +351,13 @@ run_tests()
>  do_tcpinq_test()
>  {
>  	print_title "TCP_INQ cmsg/ioctl $*"
> +
> +	if [ "${1}" = "-6" ] && ! mptcp_lib_is_v6_enabled; then
> +		mptcp_lib_pr_skip "MPTCP IPv6 not available"
> +		mptcp_lib_result_skip "TCP_INQ: $*"
> +		return 0
> +	fi
> +
>  	ip netns exec "$ns_sbox" ./mptcp_inq "$@"
>  	local lret=$?
>  	if [ $lret -ne 0 ];then
> @@ -339,10 +389,16 @@ do_tcpinq_tests()
>  		if [ $lret -ne 0 ] ; then
>  			return $lret
>  		fi
> -		do_tcpinq_test -6 $args
> -		lret=$?
> -		if [ $lret -ne 0 ] ; then
> -			return $lret
> +		if mptcp_lib_is_v6_enabled; then
> +			do_tcpinq_test -6 $args
> +			lret=$?
> +			if [ $lret -ne 0 ] ; then
> +				return $lret
> +			fi
> +		else
> +			print_title "TCP_INQ cmsg/ioctl -6 $args"
> +			mptcp_lib_pr_skip "MPTCP IPv6 not available"
> +			mptcp_lib_result_skip "TCP_INQ: -6 $args"
>  		fi
>  	done
>  

Cheers,
Matt
-- 
Sponsored by the NGI0 Core fund.

Patchew 2.1-devel-ea86937

© 2016 - 2026 Red Hat, Inc.