[PATCH net-next 0/5] netdevsim: get link via ethtool, set via sysfs

[PATCH net-next 0/5] netdevsim: get link via ethtool, set via sysfs

[Íñigo Huguet]

Allow to read the carrier status of a netdevsim device via ethtool, like
most of Ethernet devices allow.

Also, allow to set its carrier status via sysfs. Currently it's only
possible by linking a peer device and setting its admin state to
UP/DOWN. By editing the carrier status via sysfs tests that need to
check carrier related stuff may be simpler.

The third patch add the proper selftest for both things, and the fourth
and fifth fix a few small bugs in the selftest itself.

Filip Pokryvka (1):
  netdevsim: Add ethtool get_link support

Íñigo Huguet (4):
  netdevsim: allow changing carrier state via sysfs
  selftest: netdevsim: test carrier change via sysfs
  selftest: netdevsim: cleanup the devices and module on failure
  selftest: netdevsim: suppress socat warnings

 drivers/net/netdevsim/ethtool.c               |  1 +
 drivers/net/netdevsim/netdev.c                | 23 +++++
 .../selftests/drivers/net/netdevsim/peer.sh   | 84 ++++++++++++-------
 3 files changed, 80 insertions(+), 28 deletions(-)

-- 
2.53.0

[PATCH net-next 1/5] netdevsim: Add ethtool get_link support

[Íñigo Huguet]

From: Filip Pokryvka <fpokryvk@redhat.com>

Add ethtool get_link callback to netdevsim. They are devices that have
the concept of a carrier, so it makes sense that it can be queried via
ethtool too, like ethernet devices.

Signed-off-by: Filip Pokryvka <fpokryvk@redhat.com>
---
 drivers/net/netdevsim/ethtool.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/net/netdevsim/ethtool.c b/drivers/net/netdevsim/ethtool.c
index 36a201533aae..4d47eb989c24 100644
--- a/drivers/net/netdevsim/ethtool.c
+++ b/drivers/net/netdevsim/ethtool.c
@@ -209,6 +209,7 @@ static const struct ethtool_ops nsim_ethtool_ops = {
 	.supported_coalesce_params	= ETHTOOL_COALESCE_ALL_PARAMS,
 	.supported_ring_params		= ETHTOOL_RING_USE_TCP_DATA_SPLIT |
 					  ETHTOOL_RING_USE_HDS_THRS,
+	.get_link			= ethtool_op_get_link,
 	.get_pause_stats	        = nsim_get_pause_stats,
 	.get_pauseparam		        = nsim_get_pauseparam,
 	.set_pauseparam		        = nsim_set_pauseparam,
-- 
2.53.0

[PATCH net-next 2/5] netdevsim: allow changing carrier state via sysfs

[Íñigo Huguet]

Currently it is only possible to change it by linking a peer netdevsim
and setting the peer's state to down.

Allow to change the carrier state when having a single netdevsim device
without any peer, by writing to it's carrier attribute in sysfs:
/sys/bus/netdevsim/devices/netdevsimX/net/eniYynpZ/carrier. This might
be useful for carrier related tests where a peer device is not needed
for anything else.

If it's linked with a peer, change the carrier of both, while they stay
administratively up. The alternative would be to return EOPNOTSUPP, but
that doesn't seem very intuitive. Also, there is no reason why it should
not be supported, it simulates the cable being cut or disconnected, for
example.

Signed-off-by: Íñigo Huguet <ihuguet@riseup.net>
---
 drivers/net/netdevsim/netdev.c | 23 +++++++++++++++++++++++
 1 file changed, 23 insertions(+)

diff --git a/drivers/net/netdevsim/netdev.c b/drivers/net/netdevsim/netdev.c
index a750768912b5..4f77c9171b06 100644
--- a/drivers/net/netdevsim/netdev.c
+++ b/drivers/net/netdevsim/netdev.c
@@ -616,6 +616,28 @@ static void nsim_shaper_cap(struct net_shaper_binding *binding,
 	*flags = ULONG_MAX;
 }
 
+static int nsim_change_carrier(struct net_device *dev, bool new_carrier)
+{
+	struct netdevsim *ns = netdev_priv(dev);
+	struct netdevsim *peer;
+
+	netdev_assert_locked(dev);
+
+	peer = rtnl_dereference(ns->peer);
+
+	if (new_carrier) {
+		netif_carrier_on(dev);
+		if (peer)
+			netif_carrier_on(peer->netdev);
+	} else {
+		netif_carrier_off(dev);
+		if (peer)
+			netif_carrier_off(peer->netdev);
+	}
+
+	return 0;
+}
+
 static const struct net_shaper_ops nsim_shaper_ops = {
 	.set			= nsim_shaper_set,
 	.delete			= nsim_shaper_del,
@@ -645,6 +667,7 @@ static const struct net_device_ops nsim_netdev_ops = {
 	.ndo_stop		= nsim_stop,
 	.ndo_vlan_rx_add_vid	= nsim_vlan_rx_add_vid,
 	.ndo_vlan_rx_kill_vid	= nsim_vlan_rx_kill_vid,
+	.ndo_change_carrier     = nsim_change_carrier,
 	.net_shaper_ops		= &nsim_shaper_ops,
 };
 
-- 
2.53.0

[PATCH net-next 3/5] selftest: netdevsim: test carrier change via sysfs

[Íñigo Huguet]

The previous commits allowed to change the carrier state via the sysfs
carrier attribute and to get the carrier state via ethtool. Add
selftest for it.

Signed-off-by: Íñigo Huguet <ihuguet@riseup.net>
---
 .../selftests/drivers/net/netdevsim/peer.sh   | 59 +++++++++++++++----
 1 file changed, 48 insertions(+), 11 deletions(-)

diff --git a/tools/testing/selftests/drivers/net/netdevsim/peer.sh b/tools/testing/selftests/drivers/net/netdevsim/peer.sh
index f4721f7636dd..5e7b6c1d0bfb 100755
--- a/tools/testing/selftests/drivers/net/netdevsim/peer.sh
+++ b/tools/testing/selftests/drivers/net/netdevsim/peer.sh
@@ -52,21 +52,17 @@ cleanup_ns()
 	ip netns del nssv
 }
 
-is_carrier_up()
-{
-	local netns="$1"
-	local nsim_dev="$2"
-
-	test "$(ip netns exec "$netns"	\
-		cat /sys/class/net/"$nsim_dev"/carrier 2>/dev/null)" -eq 1
-}
-
 assert_carrier_up()
 {
 	local netns="$1"
 	local nsim_dev="$2"
 
-	if ! is_carrier_up "$netns" "$nsim_dev"; then
+	local sysfs_carrier="$(ip netns exec "$netns" \
+		cat /sys/class/net/"$nsim_dev"/carrier 2>/dev/null)"
+	local ethtool_carrier="$(ip netns exec "$netns" \
+		ethtool -j "$nsim_dev" | jq '.[0]["link-detected"]' 2>/dev/null)"
+
+	if [ "$sysfs_carrier" -ne 1 -o "$ethtool_carrier" != "true" ]; then
 		echo "$nsim_dev's carrier should be UP, but it isn't"
 		cleanup_ns
 		exit 1
@@ -78,13 +74,36 @@ assert_carrier_down()
 	local netns="$1"
 	local nsim_dev="$2"
 
-	if is_carrier_up "$netns" "$nsim_dev"; then
+	local sysfs_carrier="$(ip netns exec "$netns" \
+		cat /sys/class/net/"$nsim_dev"/carrier 2>/dev/null)"
+	local ethtool_carrier="$(ip netns exec "$netns" \
+		ethtool -j "$nsim_dev" | jq '.[0]["link-detected"]' 2>/dev/null)"
+
+	if [ "$sysfs_carrier" -ne 0 -o "$ethtool_carrier" != "false" ]; then
 		echo "$nsim_dev's carrier should be DOWN, but it isn't"
 		cleanup_ns
 		exit 1
 	fi
 }
 
+set_carrier_up()
+{
+	local netns="$1"
+	local nsim_dev="$2"
+
+	ip netns exec "$netns" \
+		sh -c "echo 1 > /sys/class/net/$nsim_dev/carrier" 2>/dev/null
+}
+
+set_carrier_down()
+{
+	local netns="$1"
+	local nsim_dev="$2"
+
+	ip netns exec "$netns" \
+		sh -c "echo 0 > /sys/class/net/$nsim_dev/carrier" 2>/dev/null
+}
+
 ###
 ### Code start
 ###
@@ -161,6 +180,14 @@ ip netns exec nssv ip link set dev "$NSIM_DEV_1_NAME" up
 assert_carrier_down nssv "$NSIM_DEV_1_NAME"
 assert_carrier_down nscl "$NSIM_DEV_2_NAME"
 
+set_carrier_up nssv "$NSIM_DEV_1_NAME"
+
+assert_carrier_up nssv "$NSIM_DEV_1_NAME"
+assert_carrier_down nscl "$NSIM_DEV_2_NAME"
+
+set_carrier_down nssv "$NSIM_DEV_1_NAME"
+assert_carrier_down nssv "$NSIM_DEV_1_NAME"
+
 echo "$NSIM_DEV_1_FD:$NSIM_DEV_1_IFIDX $NSIM_DEV_2_FD:$NSIM_DEV_2_IFIDX" > $NSIM_DEV_SYS_LINK
 
 assert_carrier_up nssv "$NSIM_DEV_1_NAME"
@@ -172,6 +199,16 @@ ip netns exec nssv ip link set dev "$NSIM_DEV_1_NAME" up
 assert_carrier_up nssv "$NSIM_DEV_1_NAME"
 assert_carrier_up nscl "$NSIM_DEV_2_NAME"
 
+set_carrier_down nssv "$NSIM_DEV_1_NAME"
+
+assert_carrier_down nssv "$NSIM_DEV_1_NAME"
+assert_carrier_down nscl "$NSIM_DEV_2_NAME"
+
+set_carrier_up nscl "$NSIM_DEV_2_NAME"
+
+assert_carrier_up nssv "$NSIM_DEV_1_NAME"
+assert_carrier_up nscl "$NSIM_DEV_2_NAME"
+
 # send/recv packets
 
 tmp_file=$(mktemp)
-- 
2.53.0

[PATCH net-next 4/5] selftest: netdevsim: cleanup the devices and module on failure

[Íñigo Huguet]

In case of failures only the netns were cleaned, leaving behind the
netdevsim devices created for the test and the module loaded. On
failure, delete the devices and unload the module.

Signed-off-by: Íñigo Huguet <ihuguet@riseup.net>
---
 .../selftests/drivers/net/netdevsim/peer.sh   | 21 ++++++-------------
 1 file changed, 6 insertions(+), 15 deletions(-)

diff --git a/tools/testing/selftests/drivers/net/netdevsim/peer.sh b/tools/testing/selftests/drivers/net/netdevsim/peer.sh
index 5e7b6c1d0bfb..ec96015bc9eb 100755
--- a/tools/testing/selftests/drivers/net/netdevsim/peer.sh
+++ b/tools/testing/selftests/drivers/net/netdevsim/peer.sh
@@ -46,10 +46,13 @@ setup_ns()
 	set +e
 }
 
-cleanup_ns()
+cleanup()
 {
 	ip netns del nscl
 	ip netns del nssv
+	echo $NSIM_DEV_2_ID > $NSIM_DEV_SYS_DEL
+	echo $NSIM_DEV_1_ID > $NSIM_DEV_SYS_DEL
+	modprobe -r netdevsim
 }
 
 assert_carrier_up()
@@ -64,7 +67,6 @@ assert_carrier_up()
 
 	if [ "$sysfs_carrier" -ne 1 -o "$ethtool_carrier" != "true" ]; then
 		echo "$nsim_dev's carrier should be UP, but it isn't"
-		cleanup_ns
 		exit 1
 	fi
 }
@@ -81,7 +83,6 @@ assert_carrier_down()
 
 	if [ "$sysfs_carrier" -ne 0 -o "$ethtool_carrier" != "false" ]; then
 		echo "$nsim_dev's carrier should be DOWN, but it isn't"
-		cleanup_ns
 		exit 1
 	fi
 }
@@ -120,6 +121,8 @@ udevadm settle
 
 setup_ns
 
+trap cleanup EXIT
+
 NSIM_DEV_1_FD=$((256 + RANDOM % 256))
 exec {NSIM_DEV_1_FD}</var/run/netns/nssv
 NSIM_DEV_1_IFIDX=$(ip netns exec nssv cat /sys/class/net/$NSIM_DEV_1_NAME/ifindex)
@@ -131,28 +134,24 @@ NSIM_DEV_2_IFIDX=$(ip netns exec nscl cat /sys/class/net/$NSIM_DEV_2_NAME/ifinde
 echo "$NSIM_DEV_1_FD:$NSIM_DEV_1_IFIDX $NSIM_DEV_2_FD:2000" > $NSIM_DEV_SYS_LINK 2>/dev/null
 if [ $? -eq 0 ]; then
 	echo "linking with non-existent netdevsim should fail"
-	cleanup_ns
 	exit 1
 fi
 
 echo "$NSIM_DEV_1_FD:$NSIM_DEV_1_IFIDX 2000:$NSIM_DEV_2_IFIDX" > $NSIM_DEV_SYS_LINK 2>/dev/null
 if [ $? -eq 0 ]; then
 	echo "linking with non-existent netnsid should fail"
-	cleanup_ns
 	exit 1
 fi
 
 echo "$NSIM_DEV_1_FD:$NSIM_DEV_1_IFIDX $NSIM_DEV_1_FD:$NSIM_DEV_1_IFIDX" > $NSIM_DEV_SYS_LINK 2>/dev/null
 if [ $? -eq 0 ]; then
 	echo "linking with self should fail"
-	cleanup_ns
 	exit 1
 fi
 
 echo "$NSIM_DEV_1_FD:$NSIM_DEV_1_IFIDX $NSIM_DEV_2_FD:$NSIM_DEV_2_IFIDX" > $NSIM_DEV_SYS_LINK
 if [ $? -ne 0 ]; then
 	echo "linking netdevsim1 with netdevsim2 should succeed"
-	cleanup_ns
 	exit 1
 fi
 
@@ -161,7 +160,6 @@ fi
 echo "$NSIM_DEV_1_FD:$NSIM_DEV_1_IFIDX $NSIM_DEV_2_FD:a" > $NSIM_DEV_SYS_LINK 2>/dev/null
 if [ $? -eq 0 ]; then
 	echo "invalid arg should fail"
-	cleanup_ns
 	exit 1
 fi
 
@@ -228,13 +226,6 @@ fi
 
 echo "$NSIM_DEV_1_FD:$NSIM_DEV_1_IFIDX" > $NSIM_DEV_SYS_UNLINK
 
-echo $NSIM_DEV_2_ID > $NSIM_DEV_SYS_DEL
-
 kill $pid
-echo $NSIM_DEV_1_ID > $NSIM_DEV_SYS_DEL
-
-cleanup_ns
-
-modprobe -r netdevsim
 
 exit $res
-- 
2.53.0

[PATCH net-next 5/5] selftest: netdevsim: suppress socat warnings

[Íñigo Huguet]

When executing the peer.sh selftest, socat emits these 2 warnings:
  socat[116839] W address is opened in read-write mode but only supports read-only
  socat[116833] W exiting on signal 15

The first one is because socat tries to open in read-write mode by
default, but STDIN can only be opened in read mode. Fix it by adding the
-u option (unidirectional left-to-right).

The second is because 'kill' is used to terminate the first socat process.
Suppress this warning by redirecting its stderr to /dev/null

Signed-off-by: Íñigo Huguet <ihuguet@riseup.net>
---
 tools/testing/selftests/drivers/net/netdevsim/peer.sh | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/tools/testing/selftests/drivers/net/netdevsim/peer.sh b/tools/testing/selftests/drivers/net/netdevsim/peer.sh
index ec96015bc9eb..183a130c6e4b 100755
--- a/tools/testing/selftests/drivers/net/netdevsim/peer.sh
+++ b/tools/testing/selftests/drivers/net/netdevsim/peer.sh
@@ -210,13 +210,13 @@ assert_carrier_up nscl "$NSIM_DEV_2_NAME"
 # send/recv packets
 
 tmp_file=$(mktemp)
-ip netns exec nssv socat TCP-LISTEN:1234,fork $tmp_file &
+ip netns exec nssv socat TCP-LISTEN:1234,fork $tmp_file 2>/dev/null &
 pid=$!
 res=0
 
 wait_local_port_listen nssv 1234 tcp
 
-echo "HI" | ip netns exec nscl socat STDIN TCP:192.168.1.1:1234
+echo "HI" | ip netns exec nscl socat -u STDIN TCP:192.168.1.1:1234
 
 count=$(cat $tmp_file | wc -c)
 if [[ $count -ne 3 ]]; then
-- 
2.53.0

Re: [PATCH net-next 0/5] netdevsim: get link via ethtool, set via sysfs

[Jakub Kicinski]

On Sun, 10 May 2026 18:43:32 +0200 Íñigo Huguet wrote:
> Allow to read the carrier status of a netdevsim device via ethtool, like
> most of Ethernet devices allow.
> 
> Also, allow to set its carrier status via sysfs. Currently it's only
> possible by linking a peer device and setting its admin state to
> UP/DOWN. By editing the carrier status via sysfs tests that need to
> check carrier related stuff may be simpler.
> 
> The third patch add the proper selftest for both things, and the fourth
> and fifth fix a few small bugs in the selftest itself.

Could you please explain your motivation? Feels quite trivial,
so it's not immediately obvious why we need tests.

Re: [PATCH net-next 0/5] netdevsim: get link via ethtool, set via sysfs

[Íñigo Huguet]

Hi Jakub,

On Sun, May 10, 2026 at 6:52 PM Jakub Kicinski <kuba@kernel.org> wrote:
> Could you please explain your motivation? Feels quite trivial,
> so it's not immediately obvious why we need tests.
>

We use netdevsim for some userspace CI testing here at Red Hat. Specifically,
There are some tests that require reacting to carrier changes on a device
matched by permanent MAC address from userspace. As veths don't have a
permanent MAC address, we need netdevsim for this test case.

As you say, it's pretty trivial, but IMHO there's nothing against it, as it
will help to resemble better a typical Ethernet device.

Note: we know that there's no guarantee regarding netdevsim's uAPI and we're
ready to adapt for any change, but the truth is that there's no reasonable
replacement of netdevsim for these kinds of cases where veth is not enough.



-- 
Íñigo Huguet

Re: [PATCH net-next 0/5] netdevsim: get link via ethtool, set via sysfs

[Jakub Kicinski]

On Mon, 11 May 2026 11:50:32 +0200 Íñigo Huguet wrote:
> On Sun, May 10, 2026 at 6:52 PM Jakub Kicinski <kuba@kernel.org> wrote:
> > Could you please explain your motivation? Feels quite trivial,
> > so it's not immediately obvious why we need tests.
> >  
> 
> We use netdevsim for some userspace CI testing here at Red Hat. Specifically,
> There are some tests that require reacting to carrier changes on a device
> matched by permanent MAC address from userspace. As veths don't have a
> permanent MAC address, we need netdevsim for this test case.
> 
> As you say, it's pretty trivial, but IMHO there's nothing against it, as it
> will help to resemble better a typical Ethernet device.
> 
> Note: we know that there's no guarantee regarding netdevsim's uAPI and we're
> ready to adapt for any change, but the truth is that there's no reasonable
> replacement of netdevsim for these kinds of cases where veth is not enough.

All code require maintenance. I'm trying to push back on netdevsim use
across the board. It was essentially a workaround for us not having any
real HW CI upstream. But that's no longer the case. Well, we still don't
have any crypto-capable HW :/ I think you'd be better off implementing
whatever you need in a small OOT module instead adding fake tests
upstream.