[PATCH net v3 1/2] net: ipv4: fix regression in local-broadcast routes

[PATCH net v3 1/2] net: ipv4: fix regression in local-broadcast routes

[Oscar Maes]

Commit 9e30ecf23b1b ("net: ipv4: fix incorrect MTU in broadcast routes")
introduced a regression where local-broadcast packets would have their
gateway set in __mkroute_output, which was caused by fi = NULL being
removed.

Fix this by resetting the fib_info for local-broadcast packets. This
preserves the intended changes for directed-broadcast packets.

Cc: stable@vger.kernel.org
Fixes: 9e30ecf23b1b ("net: ipv4: fix incorrect MTU in broadcast routes")
Reported-by: Brett A C Sheffield <bacs@librecast.net>
Closes: https://lore.kernel.org/regressions/20250822165231.4353-4-bacs@librecast.net
Signed-off-by: Oscar Maes <oscmaes92@gmail.com>
---
Link to discussion:
https://lore.kernel.org/netdev/20250822165231.4353-4-bacs@librecast.net/

Thanks to Brett Sheffield for finding the regression and writing
the initial fix!

 net/ipv4/route.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index f639a2ae881a..baa43e5966b1 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -2575,12 +2575,16 @@ static struct rtable *__mkroute_output(const struct fib_result *res,
 		    !netif_is_l3_master(dev_out))
 			return ERR_PTR(-EINVAL);
 
-	if (ipv4_is_lbcast(fl4->daddr))
+	if (ipv4_is_lbcast(fl4->daddr)) {
 		type = RTN_BROADCAST;
-	else if (ipv4_is_multicast(fl4->daddr))
+
+		/* reset fi to prevent gateway resolution */
+		fi = NULL;
+	} else if (ipv4_is_multicast(fl4->daddr)) {
 		type = RTN_MULTICAST;
-	else if (ipv4_is_zeronet(fl4->daddr))
+	} else if (ipv4_is_zeronet(fl4->daddr)) {
 		return ERR_PTR(-EINVAL);
+	}
 
 	if (dev_out->flags & IFF_LOOPBACK)
 		flags |= RTCF_LOCAL;
-- 
2.39.5

[PATCH net v3 2/2] selftests: net: add test for destination in broadcast packets

[Oscar Maes]

Add test to check the broadcast ethernet destination field is set
correctly.

This test sends a broadcast ping, captures it using tcpdump and
ensures that all bits of the 6 octet ethernet destination address
are correctly set by examining the output capture file.

Signed-off-by: Oscar Maes <oscmaes92@gmail.com>
---
Link to discussion:
https://lore.kernel.org/netdev/20250822165231.4353-4-bacs@librecast.net/

Thanks to Brett Sheffield for writing the initial version of this
selftest!

 tools/testing/selftests/net/Makefile          |  1 +
 .../selftests/net/broadcast_ether_dst.sh      | 82 +++++++++++++++++++
 2 files changed, 83 insertions(+)
 create mode 100755 tools/testing/selftests/net/broadcast_ether_dst.sh

diff --git a/tools/testing/selftests/net/Makefile b/tools/testing/selftests/net/Makefile
index b31a71f2b372..56ad10ea6628 100644
--- a/tools/testing/selftests/net/Makefile
+++ b/tools/testing/selftests/net/Makefile
@@ -115,6 +115,7 @@ TEST_PROGS += skf_net_off.sh
 TEST_GEN_FILES += skf_net_off
 TEST_GEN_FILES += tfo
 TEST_PROGS += tfo_passive.sh
+TEST_PROGS += broadcast_ether_dst.sh
 TEST_PROGS += broadcast_pmtu.sh
 TEST_PROGS += ipv6_force_forwarding.sh
 
diff --git a/tools/testing/selftests/net/broadcast_ether_dst.sh b/tools/testing/selftests/net/broadcast_ether_dst.sh
new file mode 100755
index 000000000000..865b5c7c8c8a
--- /dev/null
+++ b/tools/testing/selftests/net/broadcast_ether_dst.sh
@@ -0,0 +1,82 @@
+#!/bin/bash
+# SPDX-License-Identifier: GPL-2.0
+#
+# Author: Brett A C Sheffield <bacs@librecast.net>
+# Author: Oscar Maes <oscmaes92@gmail.com>
+#
+# Ensure destination ethernet field is correctly set for
+# broadcast packets
+
+source lib.sh
+
+CLIENT_IP4="192.168.0.1"
+GW_IP4="192.168.0.2"
+
+setup() {
+	setup_ns CLIENT_NS SERVER_NS
+
+	ip -net "${SERVER_NS}" link add link1 type veth \
+		peer name link0 netns "${CLIENT_NS}"
+
+	ip -net "${CLIENT_NS}" link set link0 up
+	ip -net "${CLIENT_NS}" addr add "${CLIENT_IP4}"/24 dev link0
+
+	ip -net "${SERVER_NS}" link set link1 up
+
+	ip -net "${CLIENT_NS}" route add default via "${GW_IP4}"
+	ip netns exec "${CLIENT_NS}" arp -s "${GW_IP4}" 00:11:22:33:44:55
+}
+
+cleanup() {
+	rm -f "${CAPFILE}"
+	ip -net "${SERVER_NS}" link del link1
+	cleanup_ns "${CLIENT_NS}" "${SERVER_NS}"
+}
+
+test_broadcast_ether_dst() {
+	local rc=0
+	CAPFILE=$(mktemp -u cap.XXXXXXXXXX)
+
+	echo "Testing ethernet broadcast destination"
+
+	# start tcpdump listening for icmp
+	# tcpdump will exit after receiving a single packet
+	# timeout will kill tcpdump if it is still running after 2s
+	timeout 2s ip netns exec "${CLIENT_NS}" \
+		tcpdump -i link0 -c 1 -w "${CAPFILE}" icmp &> /dev/null &
+	pid=$!
+	sleep 0.1 # let tcpdump wake up
+
+	# send broadcast ping
+	ip netns exec "${CLIENT_NS}" \
+		ping -W0.01 -c1 -b 255.255.255.255 &> /dev/null
+
+	# wait for tcpdump for exit after receiving packet
+	wait "${pid}"
+
+	# compare ethernet destination field to ff:ff:ff:ff:ff:ff
+	ether_dst=$(tcpdump -r "${CAPFILE}" -tnne 2>/dev/null | \
+			awk '{sub(/,/,"",$3); print $3}')
+	if [[ "${ether_dst}" == "ff:ff:ff:ff:ff:ff" ]]; then
+		echo "[ OK ]"
+		rc="${ksft_pass}"
+	else
+		echo "[FAIL] expected dst ether addr to be ff:ff:ff:ff:ff:ff," \
+			"got ${ether_dst}"
+		rc="${ksft_fail}"
+	fi
+
+	return "${rc}"
+}
+
+if [ ! -x "$(command -v tcpdump)" ]; then
+	echo "SKIP: Could not run test without tcpdump tool"
+	exit "${ksft_skip}"
+fi
+
+trap cleanup EXIT
+
+setup
+test_broadcast_ether_dst
+
+exit $?
-- 
2.39.5

Re: [PATCH net v3 2/2] selftests: net: add test for destination in broadcast packets

[Brett A C Sheffield]

On 2025-08-27 08:23, Oscar Maes wrote:
> Add test to check the broadcast ethernet destination field is set
> correctly.
> 
> This test sends a broadcast ping, captures it using tcpdump and
> ensures that all bits of the 6 octet ethernet destination address
> are correctly set by examining the output capture file.
> 
> Signed-off-by: Oscar Maes <oscmaes92@gmail.com>
> ---
> Link to discussion:
> https://lore.kernel.org/netdev/20250822165231.4353-4-bacs@librecast.net/
> 
> Thanks to Brett Sheffield for writing the initial version of this
> selftest!

Thanks for leaving my author name in the file.  Perhaps you might consider
adding:

Co-Authored-By: Brett A C Sheffield <bacs@librecast.net>

to your commit message. I spend quite a bit of my Saturday bisecting and
diagnosing,  and writing the patch and test.

>  tools/testing/selftests/net/Makefile          |  1 +
>  .../selftests/net/broadcast_ether_dst.sh      | 82 +++++++++++++++++++
>  2 files changed, 83 insertions(+)
>  create mode 100755 tools/testing/selftests/net/broadcast_ether_dst.sh
> 
> diff --git a/tools/testing/selftests/net/Makefile b/tools/testing/selftests/net/Makefile
> index b31a71f2b372..56ad10ea6628 100644
> --- a/tools/testing/selftests/net/Makefile
> +++ b/tools/testing/selftests/net/Makefile
> @@ -115,6 +115,7 @@ TEST_PROGS += skf_net_off.sh
>  TEST_GEN_FILES += skf_net_off
>  TEST_GEN_FILES += tfo
>  TEST_PROGS += tfo_passive.sh
> +TEST_PROGS += broadcast_ether_dst.sh
>  TEST_PROGS += broadcast_pmtu.sh
>  TEST_PROGS += ipv6_force_forwarding.sh
>  
> diff --git a/tools/testing/selftests/net/broadcast_ether_dst.sh b/tools/testing/selftests/net/broadcast_ether_dst.sh
> new file mode 100755
> index 000000000000..865b5c7c8c8a
> --- /dev/null
> +++ b/tools/testing/selftests/net/broadcast_ether_dst.sh
> @@ -0,0 +1,82 @@
> +#!/bin/bash
> +# SPDX-License-Identifier: GPL-2.0
> +#
> +# Author: Brett A C Sheffield <bacs@librecast.net>
> +# Author: Oscar Maes <oscmaes92@gmail.com>
> +#
> +# Ensure destination ethernet field is correctly set for
> +# broadcast packets
> +
> +source lib.sh
> +
> +CLIENT_IP4="192.168.0.1"
> +GW_IP4="192.168.0.2"
> +
> +setup() {
> +	setup_ns CLIENT_NS SERVER_NS
> +
> +	ip -net "${SERVER_NS}" link add link1 type veth \
> +		peer name link0 netns "${CLIENT_NS}"
> +
> +	ip -net "${CLIENT_NS}" link set link0 up
> +	ip -net "${CLIENT_NS}" addr add "${CLIENT_IP4}"/24 dev link0
> +
> +	ip -net "${SERVER_NS}" link set link1 up
> +
> +	ip -net "${CLIENT_NS}" route add default via "${GW_IP4}"
> +	ip netns exec "${CLIENT_NS}" arp -s "${GW_IP4}" 00:11:22:33:44:55
> +}
> +
> +cleanup() {
> +	rm -f "${CAPFILE}"
> +	ip -net "${SERVER_NS}" link del link1
> +	cleanup_ns "${CLIENT_NS}" "${SERVER_NS}"
> +}
> +
> +test_broadcast_ether_dst() {
> +	local rc=0
> +	CAPFILE=$(mktemp -u cap.XXXXXXXXXX)
> +
> +	echo "Testing ethernet broadcast destination"
> +
> +	# start tcpdump listening for icmp
> +	# tcpdump will exit after receiving a single packet
> +	# timeout will kill tcpdump if it is still running after 2s
> +	timeout 2s ip netns exec "${CLIENT_NS}" \
> +		tcpdump -i link0 -c 1 -w "${CAPFILE}" icmp &> /dev/null &
> +	pid=$!
> +	sleep 0.1 # let tcpdump wake up
> +
> +	# send broadcast ping
> +	ip netns exec "${CLIENT_NS}" \
> +		ping -W0.01 -c1 -b 255.255.255.255 &> /dev/null
> +
> +	# wait for tcpdump for exit after receiving packet
> +	wait "${pid}"
> +
> +	# compare ethernet destination field to ff:ff:ff:ff:ff:ff
> +	ether_dst=$(tcpdump -r "${CAPFILE}" -tnne 2>/dev/null | \
> +			awk '{sub(/,/,"",$3); print $3}')
> +	if [[ "${ether_dst}" == "ff:ff:ff:ff:ff:ff" ]]; then
> +		echo "[ OK ]"
> +		rc="${ksft_pass}"
> +	else
> +		echo "[FAIL] expected dst ether addr to be ff:ff:ff:ff:ff:ff," \
> +			"got ${ether_dst}"
> +		rc="${ksft_fail}"
> +	fi
> +
> +	return "${rc}"
> +}
> +
> +if [ ! -x "$(command -v tcpdump)" ]; then
> +	echo "SKIP: Could not run test without tcpdump tool"
> +	exit "${ksft_skip}"
> +fi
> +
> +trap cleanup EXIT
> +
> +setup
> +test_broadcast_ether_dst
> +
> +exit $?
> -- 
> 2.39.5
> 
> 

-- 

Re: [PATCH net v3 2/2] selftests: net: add test for destination in broadcast packets

[Paolo Abeni]

On 8/27/25 4:26 PM, Brett A C Sheffield wrote:
> On 2025-08-27 08:23, Oscar Maes wrote:
>> Add test to check the broadcast ethernet destination field is set
>> correctly.
>>
>> This test sends a broadcast ping, captures it using tcpdump and
>> ensures that all bits of the 6 octet ethernet destination address
>> are correctly set by examining the output capture file.
>>
>> Signed-off-by: Oscar Maes <oscmaes92@gmail.com>
>> ---
>> Link to discussion:
>> https://lore.kernel.org/netdev/20250822165231.4353-4-bacs@librecast.net/
>>
>> Thanks to Brett Sheffield for writing the initial version of this
>> selftest!
> 
> Thanks for leaving my author name in the file.  Perhaps you might consider
> adding:
> 
> Co-Authored-By: Brett A C Sheffield <bacs@librecast.net>
> 
> to your commit message. I spend quite a bit of my Saturday bisecting and
> diagnosing,  and writing the patch and test.

I don't want to delay the fix, since I received other reports for the
same problem, but I think proper recognition should be agreed by all the
involved parties.

I'm going to apply patch 1/2 standalone, to allow repost for this one.

>>  tools/testing/selftests/net/Makefile          |  1 +
>>  .../selftests/net/broadcast_ether_dst.sh      | 82 +++++++++++++++++++
>>  2 files changed, 83 insertions(+)
>>  create mode 100755 tools/testing/selftests/net/broadcast_ether_dst.sh
>>
>> diff --git a/tools/testing/selftests/net/Makefile b/tools/testing/selftests/net/Makefile
>> index b31a71f2b372..56ad10ea6628 100644
>> --- a/tools/testing/selftests/net/Makefile
>> +++ b/tools/testing/selftests/net/Makefile
>> @@ -115,6 +115,7 @@ TEST_PROGS += skf_net_off.sh
>>  TEST_GEN_FILES += skf_net_off
>>  TEST_GEN_FILES += tfo
>>  TEST_PROGS += tfo_passive.sh
>> +TEST_PROGS += broadcast_ether_dst.sh
>>  TEST_PROGS += broadcast_pmtu.sh
>>  TEST_PROGS += ipv6_force_forwarding.sh
>>  
>> diff --git a/tools/testing/selftests/net/broadcast_ether_dst.sh b/tools/testing/selftests/net/broadcast_ether_dst.sh
>> new file mode 100755
>> index 000000000000..865b5c7c8c8a
>> --- /dev/null
>> +++ b/tools/testing/selftests/net/broadcast_ether_dst.sh
>> @@ -0,0 +1,82 @@
>> +#!/bin/bash
>> +# SPDX-License-Identifier: GPL-2.0
>> +#
>> +# Author: Brett A C Sheffield <bacs@librecast.net>
>> +# Author: Oscar Maes <oscmaes92@gmail.com>
>> +#
>> +# Ensure destination ethernet field is correctly set for
>> +# broadcast packets
>> +
>> +source lib.sh
>> +
>> +CLIENT_IP4="192.168.0.1"
>> +GW_IP4="192.168.0.2"
>> +
>> +setup() {
>> +	setup_ns CLIENT_NS SERVER_NS
>> +
>> +	ip -net "${SERVER_NS}" link add link1 type veth \
>> +		peer name link0 netns "${CLIENT_NS}"
>> +
>> +	ip -net "${CLIENT_NS}" link set link0 up
>> +	ip -net "${CLIENT_NS}" addr add "${CLIENT_IP4}"/24 dev link0
>> +
>> +	ip -net "${SERVER_NS}" link set link1 up
>> +
>> +	ip -net "${CLIENT_NS}" route add default via "${GW_IP4}"
>> +	ip netns exec "${CLIENT_NS}" arp -s "${GW_IP4}" 00:11:22:33:44:55
>> +}
>> +
>> +cleanup() {
>> +	rm -f "${CAPFILE}"
>> +	ip -net "${SERVER_NS}" link del link1
>> +	cleanup_ns "${CLIENT_NS}" "${SERVER_NS}"
>> +}
>> +
>> +test_broadcast_ether_dst() {
>> +	local rc=0
>> +	CAPFILE=$(mktemp -u cap.XXXXXXXXXX)
>> +
>> +	echo "Testing ethernet broadcast destination"
>> +
>> +	# start tcpdump listening for icmp
>> +	# tcpdump will exit after receiving a single packet
>> +	# timeout will kill tcpdump if it is still running after 2s
>> +	timeout 2s ip netns exec "${CLIENT_NS}" \
>> +		tcpdump -i link0 -c 1 -w "${CAPFILE}" icmp &> /dev/null &
>> +	pid=$!
>> +	sleep 0.1 # let tcpdump wake up

Here you could use slowwait checking for packet socket creation, to be
more robust WRT very slow env.

/P

Re: [PATCH net v3 1/2] net: ipv4: fix regression in local-broadcast routes

[David Ahern]

On 8/27/25 12:23 AM, Oscar Maes wrote:
> Commit 9e30ecf23b1b ("net: ipv4: fix incorrect MTU in broadcast routes")
> introduced a regression where local-broadcast packets would have their
> gateway set in __mkroute_output, which was caused by fi = NULL being
> removed.
> 
> Fix this by resetting the fib_info for local-broadcast packets. This
> preserves the intended changes for directed-broadcast packets.
> 
> Cc: stable@vger.kernel.org
> Fixes: 9e30ecf23b1b ("net: ipv4: fix incorrect MTU in broadcast routes")
> Reported-by: Brett A C Sheffield <bacs@librecast.net>
> Closes: https://lore.kernel.org/regressions/20250822165231.4353-4-bacs@librecast.net
> Signed-off-by: Oscar Maes <oscmaes92@gmail.com>
> ---
> Link to discussion:
> https://lore.kernel.org/netdev/20250822165231.4353-4-bacs@librecast.net/
> 
> Thanks to Brett Sheffield for finding the regression and writing
> the initial fix!
> 
>  net/ipv4/route.c | 10 +++++++---
>  1 file changed, 7 insertions(+), 3 deletions(-)
> 

Reviewed-by: David Ahern <dsahern@kernel.org>

Re: [PATCH net v3 1/2] net: ipv4: fix regression in local-broadcast routes

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net.git (main)
by Paolo Abeni <pabeni@redhat.com>:

On Wed, 27 Aug 2025 08:23:21 +0200 you wrote:
> Commit 9e30ecf23b1b ("net: ipv4: fix incorrect MTU in broadcast routes")
> introduced a regression where local-broadcast packets would have their
> gateway set in __mkroute_output, which was caused by fi = NULL being
> removed.
> 
> Fix this by resetting the fib_info for local-broadcast packets. This
> preserves the intended changes for directed-broadcast packets.
> 
> [...]

Here is the summary with links:
  - [net,v3,1/2] net: ipv4: fix regression in local-broadcast routes
    https://git.kernel.org/netdev/net/c/5189446ba995
  - [net,v3,2/2] selftests: net: add test for destination in broadcast packets
    (no matching commit)

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html