Merge branch 'ethtool-mm-api-consolidation'

Vladimir Oltean says:

====================
ethtool mm API consolidation

This series consolidates the behavior of the 2 drivers that implement
the ethtool MAC Merge layer by making NXP ENETC commit its preemptible
traffic classes to hardware only when MM TX is active (same as Ocelot).

Then, after resolving an issue with the ENETC driver, it restricts user
space from entering 2 states which don't make sense:

- pmac-enabled off tx-enabled on  verify-enabled *
- pmac-enabled *   tx-enabled off verify-enabled on

Then, it introduces a selftest (ethtool_mm.sh) which puts everything
together and tests all valid configurations known to me.

This is simultaneously the v2 of "[PATCH net-next 0/2] ethtool mm API
improvements":
https://lore.kernel.org/netdev/20230415173454.3970647-1-vladimir.oltean@nxp.com/
which had caused some problems to openlldp. Those were solved in the
meantime, see:
https://github.com/intel/openlldp/commit/11171b474f6f3cbccac5d608b7f26b32ff72c651

and of "[RFC PATCH net-next] selftests: forwarding: add a test for MAC
Merge layer":
https://lore.kernel.org/netdev/20230210221243.228932-1-vladimir.oltean@nxp.com/
====================

Link: https://lore.kernel.org/r/20230418111459.811553-1-vladimir.oltean@nxp.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

drivers/net/ethernet/freescale/enetc/enetc.c

@@ -25,23 +25,12 @@ void enetc_port_mac_wr(struct enetc_si *si, u32 reg, u32 val)
 }
 EXPORT_SYMBOL_GPL(enetc_port_mac_wr);
 
-void enetc_set_ptcfpr(struct enetc_hw *hw, unsigned long preemptible_tcs)
+static void enetc_change_preemptible_tcs(struct enetc_ndev_priv *priv,
+					 u8 preemptible_tcs)
 {
-	u32 val;
-	int tc;
-
-	for (tc = 0; tc < 8; tc++) {
-		val = enetc_port_rd(hw, ENETC_PTCFPR(tc));
-
-		if (preemptible_tcs & BIT(tc))
-			val |= ENETC_PTCFPR_FPE;
-		else
-			val &= ~ENETC_PTCFPR_FPE;
-
-		enetc_port_wr(hw, ENETC_PTCFPR(tc), val);
-	}
+	priv->preemptible_tcs = preemptible_tcs;
+	enetc_mm_commit_preemptible_tcs(priv);
 }
-EXPORT_SYMBOL_GPL(enetc_set_ptcfpr);
 
 static int enetc_num_stack_tx_queues(struct enetc_ndev_priv *priv)
 {
@@ -2659,7 +2648,7 @@ static void enetc_reset_tc_mqprio(struct net_device *ndev)
 
 	enetc_debug_tx_ring_prios(priv);
 
-	enetc_set_ptcfpr(hw, 0);
+	enetc_change_preemptible_tcs(priv, 0);
 }
 
 int enetc_setup_tc_mqprio(struct net_device *ndev, void *type_data)
@@ -2714,7 +2703,7 @@ int enetc_setup_tc_mqprio(struct net_device *ndev, void *type_data)
 
 	enetc_debug_tx_ring_prios(priv);
 
-	enetc_set_ptcfpr(hw, mqprio->preemptible_tcs);
+	enetc_change_preemptible_tcs(priv, mqprio->preemptible_tcs);
 
 	return 0;
 

drivers/net/ethernet/freescale/enetc/enetc.h

@@ -355,6 +355,9 @@ struct enetc_ndev_priv {
 	u16 rx_bd_count, tx_bd_count;
 
 	u16 msg_enable;
+
+	u8 preemptible_tcs;
+
 	enum enetc_active_offloads active_offloads;
 
 	u32 speed; /* store speed for compare update pspeed */
@@ -433,6 +436,7 @@ int enetc_xdp_xmit(struct net_device *ndev, int num_frames,
 /* ethtool */
 void enetc_set_ethtool_ops(struct net_device *ndev);
 void enetc_mm_link_state_update(struct enetc_ndev_priv *priv, bool link);
+void enetc_mm_commit_preemptible_tcs(struct enetc_ndev_priv *priv);
 
 /* control buffer descriptor ring (CBDR) */
 int enetc_setup_cbdr(struct device *dev, struct enetc_hw *hw, int bd_count,
@@ -486,7 +490,6 @@ static inline void enetc_cbd_free_data_mem(struct enetc_si *si, int size,
 
 void enetc_reset_ptcmsdur(struct enetc_hw *hw);
 void enetc_set_ptcmsdur(struct enetc_hw *hw, u32 *queue_max_sdu);
-void enetc_set_ptcfpr(struct enetc_hw *hw, unsigned long preemptible_tcs);
 
 #ifdef CONFIG_FSL_ENETC_QOS
 int enetc_qos_query_caps(struct net_device *ndev, void *type_data);

drivers/net/ethernet/freescale/enetc/enetc_ethtool.c

@@ -32,6 +32,12 @@ static const u32 enetc_port_regs[] = {
 	ENETC_PM0_CMD_CFG, ENETC_PM0_MAXFRM, ENETC_PM0_IF_MODE
 };
 
+static const u32 enetc_port_mm_regs[] = {
+	ENETC_MMCSR, ENETC_PFPMR, ENETC_PTCFPR(0), ENETC_PTCFPR(1),
+	ENETC_PTCFPR(2), ENETC_PTCFPR(3), ENETC_PTCFPR(4), ENETC_PTCFPR(5),
+	ENETC_PTCFPR(6), ENETC_PTCFPR(7),
+};
+
 static int enetc_get_reglen(struct net_device *ndev)
 {
 	struct enetc_ndev_priv *priv = netdev_priv(ndev);
@@ -45,6 +51,9 @@ static int enetc_get_reglen(struct net_device *ndev)
 	if (hw->port)
 		len += ARRAY_SIZE(enetc_port_regs);
 
+	if (hw->port && !!(priv->si->hw_features & ENETC_SI_F_QBU))
+		len += ARRAY_SIZE(enetc_port_mm_regs);
+
 	len *= sizeof(u32) * 2; /* store 2 entries per reg: addr and value */
 
 	return len;
@@ -90,6 +99,14 @@ static void enetc_get_regs(struct net_device *ndev, struct ethtool_regs *regs,
 		*buf++ = addr;
 		*buf++ = enetc_rd(hw, addr);
 	}
+
+	if (priv->si->hw_features & ENETC_SI_F_QBU) {
+		for (i = 0; i < ARRAY_SIZE(enetc_port_mm_regs); i++) {
+			addr = ENETC_PORT_BASE + enetc_port_mm_regs[i];
+			*buf++ = addr;
+			*buf++ = enetc_rd(hw, addr);
+		}
+	}
 }
 
 static const struct {
@@ -976,7 +993,9 @@ static int enetc_get_mm(struct net_device *ndev, struct ethtool_mm_state *state)
 	lafs = ENETC_MMCSR_GET_LAFS(val);
 	state->rx_min_frag_size = ethtool_mm_frag_size_add_to_min(lafs);
 	state->tx_enabled = !!(val & ENETC_MMCSR_LPE); /* mirror of MMCSR_ME */
-	state->tx_active = !!(val & ENETC_MMCSR_LPA);
+	state->tx_active = state->tx_enabled &&
+			   (state->verify_status == ETHTOOL_MM_VERIFY_STATUS_SUCCEEDED ||
+			    state->verify_status == ETHTOOL_MM_VERIFY_STATUS_DISABLED);
 	state->verify_enabled = !(val & ENETC_MMCSR_VDIS);
 	state->verify_time = ENETC_MMCSR_GET_VT(val);
 	/* A verifyTime of 128 ms would exceed the 7 bit width
@@ -989,6 +1008,64 @@ static int enetc_get_mm(struct net_device *ndev, struct ethtool_mm_state *state)
 	return 0;
 }
 
+static int enetc_mm_wait_tx_active(struct enetc_hw *hw, int verify_time)
+{
+	int timeout = verify_time * USEC_PER_MSEC * ENETC_MM_VERIFY_RETRIES;
+	u32 val;
+
+	/* This will time out after the standard value of 3 verification
+	 * attempts. To not sleep forever, it relies on a non-zero verify_time,
+	 * guarantee which is provided by the ethtool nlattr policy.
+	 */
+	return read_poll_timeout(enetc_port_rd, val,
+				 ENETC_MMCSR_GET_VSTS(val) == 3,
+				 ENETC_MM_VERIFY_SLEEP_US, timeout,
+				 true, hw, ENETC_MMCSR);
+}
+
+static void enetc_set_ptcfpr(struct enetc_hw *hw, u8 preemptible_tcs)
+{
+	u32 val;
+	int tc;
+
+	for (tc = 0; tc < 8; tc++) {
+		val = enetc_port_rd(hw, ENETC_PTCFPR(tc));
+
+		if (preemptible_tcs & BIT(tc))
+			val |= ENETC_PTCFPR_FPE;
+		else
+			val &= ~ENETC_PTCFPR_FPE;
+
+		enetc_port_wr(hw, ENETC_PTCFPR(tc), val);
+	}
+}
+
+/* ENETC does not have an IRQ to notify changes to the MAC Merge TX status
+ * (active/inactive), but the preemptible traffic classes should only be
+ * committed to hardware once TX is active. Resort to polling.
+ */
+void enetc_mm_commit_preemptible_tcs(struct enetc_ndev_priv *priv)
+{
+	struct enetc_hw *hw = &priv->si->hw;
+	u8 preemptible_tcs = 0;
+	u32 val;
+	int err;
+
+	val = enetc_port_rd(hw, ENETC_MMCSR);
+	if (!(val & ENETC_MMCSR_ME))
+		goto out;
+
+	if (!(val & ENETC_MMCSR_VDIS)) {
+		err = enetc_mm_wait_tx_active(hw, ENETC_MMCSR_GET_VT(val));
+		if (err)
+			goto out;
+	}
+
+	preemptible_tcs = priv->preemptible_tcs;
+out:
+	enetc_set_ptcfpr(hw, preemptible_tcs);
+}
+
 /* FIXME: Workaround for the link partner's verification failing if ENETC
  * priorly received too much express traffic. The documentation doesn't
  * suggest this is needed.
@@ -1041,10 +1118,13 @@ static int enetc_set_mm(struct net_device *ndev, struct ethtool_mm_cfg *cfg,
 	else
 		priv->active_offloads &= ~ENETC_F_QBU;
 
-	/* If link is up, enable MAC Merge right away */
-	if (!!(priv->active_offloads & ENETC_F_QBU) &&
-	    !(val & ENETC_MMCSR_LINK_FAIL))
-		val |= ENETC_MMCSR_ME;
+	/* If link is up, enable/disable MAC Merge right away */
+	if (!(val & ENETC_MMCSR_LINK_FAIL)) {
+		if (!!(priv->active_offloads & ENETC_F_QBU))
+			val |= ENETC_MMCSR_ME;
+		else
+			val &= ~ENETC_MMCSR_ME;
+	}
 
 	val &= ~ENETC_MMCSR_VT_MASK;
 	val |= ENETC_MMCSR_VT(cfg->verify_time);
@@ -1056,6 +1136,8 @@ static int enetc_set_mm(struct net_device *ndev, struct ethtool_mm_cfg *cfg,
 
 	enetc_restart_emac_rx(priv->si);
 
+	enetc_mm_commit_preemptible_tcs(priv);
+
 	mutex_unlock(&priv->mm_lock);
 
 	return 0;
@@ -1089,6 +1171,8 @@ void enetc_mm_link_state_update(struct enetc_ndev_priv *priv, bool link)
 
 	enetc_port_wr(hw, ENETC_MMCSR, val);
 
+	enetc_mm_commit_preemptible_tcs(priv);
+
 	mutex_unlock(&priv->mm_lock);
 }
 EXPORT_SYMBOL_GPL(enetc_mm_link_state_update);

drivers/net/ethernet/freescale/enetc/enetc_hw.h

@@ -3,6 +3,9 @@
 
 #include <linux/bitops.h>
 
+#define ENETC_MM_VERIFY_SLEEP_US	USEC_PER_MSEC
+#define ENETC_MM_VERIFY_RETRIES		3
+
 /* ENETC device IDs */
 #define ENETC_DEV_ID_PF		0xe100
 #define ENETC_DEV_ID_VF		0xef00

net/ethtool/mm.c

@@ -214,6 +214,16 @@ static int ethnl_set_mm(struct ethnl_req_info *req_info, struct genl_info *info)
 		return -ERANGE;
 	}
 
+	if (cfg.verify_enabled && !cfg.tx_enabled) {
+		NL_SET_ERR_MSG(extack, "Verification requires TX enabled");
+		return -EINVAL;
+	}
+
+	if (cfg.tx_enabled && !cfg.pmac_enabled) {
+		NL_SET_ERR_MSG(extack, "TX enabled requires pMAC enabled");
+		return -EINVAL;
+	}
+
 	ret = dev->ethtool_ops->set_mm(dev, &cfg, extack);
 	return ret < 0 ? ret : 1;
 }

tools/testing/selftests/drivers/net/mlxsw/qos_headroom.sh

@@ -18,7 +18,6 @@ lib_dir=$(dirname $0)/../../../net/forwarding
 NUM_NETIFS=0
 source $lib_dir/lib.sh
 source $lib_dir/devlink_lib.sh
-source qos_lib.sh
 
 swp=$NETIF_NO_CABLE
 
@@ -371,7 +370,7 @@ test_tc_int_buf()
 	tc qdisc delete dev $swp root
 }
 
-bail_on_lldpad
+bail_on_lldpad "configure DCB" "configure Qdiscs"
 
 trap cleanup EXIT
 setup_wait

tools/testing/selftests/drivers/net/mlxsw/qos_lib.sh

@@ -54,31 +54,3 @@ measure_rate()
 	echo $ir $er
 	return $ret
 }
-
-bail_on_lldpad()
-{
-	if systemctl is-active --quiet lldpad; then
-
-		cat >/dev/stderr <<-EOF
-		WARNING: lldpad is running
-
-			lldpad will likely configure DCB, and this test will
-			configure Qdiscs. mlxsw does not support both at the
-			same time, one of them is arbitrarily going to overwrite
-			the other. That will cause spurious failures (or,
-			unlikely, passes) of this test.
-		EOF
-
-		if [[ -z $ALLOW_LLDPAD ]]; then
-			cat >/dev/stderr <<-EOF
-
-				If you want to run the test anyway, please set
-				an environment variable ALLOW_LLDPAD to a
-				non-empty string.
-			EOF
-			exit 1
-		else
-			return
-		fi
-	fi
-}

tools/testing/selftests/drivers/net/mlxsw/qos_pfc.sh

@@ -79,7 +79,6 @@ lib_dir=$(dirname $0)/../../../net/forwarding
 NUM_NETIFS=6
 source $lib_dir/lib.sh
 source $lib_dir/devlink_lib.sh
-source qos_lib.sh
 
 _1KB=1000
 _100KB=$((100 * _1KB))
@@ -393,7 +392,7 @@ test_qos_pfc()
 	log_test "PFC"
 }
 
-bail_on_lldpad
+bail_on_lldpad "configure DCB" "configure Qdiscs"
 
 trap cleanup EXIT
 setup_prepare

tools/testing/selftests/drivers/net/mlxsw/sch_ets.sh

@@ -5,7 +5,6 @@
 lib_dir=$(dirname $0)/../../../net/forwarding
 source $lib_dir/sch_ets_core.sh
 source $lib_dir/devlink_lib.sh
-source qos_lib.sh
 
 ALL_TESTS="
 	ping_ipv4
@@ -78,5 +77,5 @@ collect_stats()
 	done
 }
 
-bail_on_lldpad
+bail_on_lldpad "configure DCB" "configure Qdiscs"
 ets_run

tools/testing/selftests/drivers/net/mlxsw/sch_red_core.sh

@@ -74,7 +74,6 @@ lib_dir=$(dirname $0)/../../../net/forwarding
 source $lib_dir/lib.sh
 source $lib_dir/devlink_lib.sh
 source mlxsw_lib.sh
-source qos_lib.sh
 
 ipaddr()
 {

tools/testing/selftests/drivers/net/mlxsw/sch_red_ets.sh

@@ -166,7 +166,7 @@ ecn_mirror_test()
 	uninstall_qdisc
 }
 
-bail_on_lldpad
+bail_on_lldpad "configure DCB" "configure Qdiscs"
 
 trap cleanup EXIT
 setup_prepare

tools/testing/selftests/drivers/net/mlxsw/sch_red_root.sh

@@ -73,7 +73,7 @@ red_mirror_test()
 	uninstall_qdisc
 }
 
-bail_on_lldpad
+bail_on_lldpad "configure DCB" "configure Qdiscs"
 
 trap cleanup EXIT
 setup_prepare

tools/testing/selftests/drivers/net/mlxsw/sch_tbf_ets.sh

 #!/bin/bash
 # SPDX-License-Identifier: GPL-2.0
 
-source qos_lib.sh
-bail_on_lldpad
+sch_tbf_pre_hook()
+{
+	bail_on_lldpad "configure DCB" "configure Qdiscs"
+}
 
 lib_dir=$(dirname $0)/../../../net/forwarding
 TCFLAGS=skip_sw

tools/testing/selftests/drivers/net/mlxsw/sch_tbf_prio.sh

 #!/bin/bash
 # SPDX-License-Identifier: GPL-2.0
 
-source qos_lib.sh
-bail_on_lldpad
+sch_tbf_pre_hook()
+{
+	bail_on_lldpad "configure DCB" "configure Qdiscs"
+}
 
 lib_dir=$(dirname $0)/../../../net/forwarding
 TCFLAGS=skip_sw

tools/testing/selftests/drivers/net/mlxsw/sch_tbf_root.sh

 #!/bin/bash
 # SPDX-License-Identifier: GPL-2.0
 
-source qos_lib.sh
-bail_on_lldpad
+sch_tbf_pre_hook()
+{
+	bail_on_lldpad "configure DCB" "configure Qdiscs"
+}
 
 lib_dir=$(dirname $0)/../../../net/forwarding
 TCFLAGS=skip_sw

tools/testing/selftests/net/forwarding/Makefile

@@ -15,6 +15,7 @@ TEST_PROGS = bridge_igmp.sh \
 	custom_multipath_hash.sh \
 	dual_vxlan_bridge.sh \
 	ethtool_extended_state.sh \
+	ethtool_mm.sh \
 	ethtool.sh \
 	gre_custom_multipath_hash.sh \
 	gre_inner_v4_multipath.sh \

tools/testing/selftests/net/forwarding/ethtool_mm.sh

+#!/bin/bash
+# SPDX-License-Identifier: GPL-2.0
+
+ALL_TESTS="
+	manual_with_verification_h1_to_h2
+	manual_with_verification_h2_to_h1
+	manual_without_verification_h1_to_h2
+	manual_without_verification_h2_to_h1
+	manual_failed_verification_h1_to_h2
+	manual_failed_verification_h2_to_h1
+	lldp
+"
+
+NUM_NETIFS=2
+REQUIRE_MZ=no
+PREEMPTIBLE_PRIO=0
+source lib.sh
+
+traffic_test()
+{
+	local if=$1; shift
+	local src=$1; shift
+	local num_pkts=10000
+	local before=
+	local after=
+	local delta=
+
+	before=$(ethtool_std_stats_get $if "eth-mac" "FramesTransmittedOK" $src)
+
+	$MZ $if -q -c $num_pkts -p 64 -b bcast -t ip -R $PREEMPTIBLE_PRIO
+
+	after=$(ethtool_std_stats_get $if "eth-mac" "FramesTransmittedOK" $src)
+
+	delta=$((after - before))
+
+	# Allow an extra 1% tolerance for random packets sent by the stack
+	[ $delta -ge $num_pkts ] && [ $delta -le $((num_pkts + 100)) ]
+}
+
+manual_with_verification()
+{
+	local tx=$1; shift
+	local rx=$1; shift
+
+	RET=0
+
+	# It isn't completely clear from IEEE 802.3-2018 Figure 99-5: Transmit
+	# Processing state diagram whether the "send_r" variable (send response
+	# to verification frame) should be taken into consideration while the
+	# MAC Merge TX direction is disabled. That being said, at least the
+	# NXP ENETC does not, and requires tx-enabled on in order to respond to
+	# the link partner's verification frames.
+	ethtool --set-mm $rx tx-enabled on
+	ethtool --set-mm $tx verify-enabled on tx-enabled on
+
+	# Wait for verification to finish
+	sleep 1
+
+	ethtool --json --show-mm $tx | jq -r '.[]."verify-status"' | \
+		grep -q 'SUCCEEDED'
+	check_err "$?" "Verification did not succeed"
+
+	ethtool --json --show-mm $tx | jq -r '.[]."tx-active"' | grep -q 'true'
+	check_err "$?" "pMAC TX is not active"
+
+	traffic_test $tx "pmac"
+	check_err "$?" "Traffic did not get sent through $tx's pMAC"
+
+	ethtool --set-mm $tx verify-enabled off tx-enabled off
+	ethtool --set-mm $rx tx-enabled off
+
+	log_test "Manual configuration with verification: $tx to $rx"
+}
+
+manual_with_verification_h1_to_h2()
+{
+	manual_with_verification $h1 $h2
+}
+
+manual_with_verification_h2_to_h1()
+{
+	manual_with_verification $h2 $h1
+}
+
+manual_without_verification()
+{
+	local tx=$1; shift
+	local rx=$1; shift
+
+	RET=0
+
+	ethtool --set-mm $tx verify-enabled off tx-enabled on
+
+	ethtool --json --show-mm $tx | jq -r '.[]."verify-status"' | \
+		grep -q 'DISABLED'
+	check_err "$?" "Verification is not disabled"
+
+	ethtool --json --show-mm $tx | jq -r '.[]."tx-active"' | grep -q 'true'
+	check_err "$?" "pMAC TX is not active"
+
+	traffic_test $tx "pmac"
+	check_err "$?" "Traffic did not get sent through $tx's pMAC"
+
+	ethtool --set-mm $tx verify-enabled off tx-enabled off
+
+	log_test "Manual configuration without verification: $tx to $rx"
+}
+
+manual_without_verification_h1_to_h2()
+{
+	manual_without_verification $h1 $h2
+}
+
+manual_without_verification_h2_to_h1()
+{
+	manual_without_verification $h2 $h1
+}
+
+manual_failed_verification()
+{
+	local tx=$1; shift
+	local rx=$1; shift
+
+	RET=0
+
+	ethtool --set-mm $rx pmac-enabled off
+	ethtool --set-mm $tx verify-enabled on tx-enabled on
+
+	# Wait for verification to time out
+	sleep 1
+
+	ethtool --json --show-mm $tx | jq -r '.[]."verify-status"' | \
+		grep -q 'SUCCEEDED'
+	check_fail "$?" "Verification succeeded when it shouldn't have"
+
+	ethtool --json --show-mm $tx | jq -r '.[]."tx-active"' | grep -q 'true'
+	check_fail "$?" "pMAC TX is active when it shouldn't have"
+
+	traffic_test $tx "emac"
+	check_err "$?" "Traffic did not get sent through $tx's eMAC"
+
+	ethtool --set-mm $tx verify-enabled off tx-enabled off
+	ethtool --set-mm $rx pmac-enabled on
+
+	log_test "Manual configuration with failed verification: $tx to $rx"
+}
+
+manual_failed_verification_h1_to_h2()
+{
+	manual_failed_verification $h1 $h2
+}
+
+manual_failed_verification_h2_to_h1()
+{
+	manual_failed_verification $h2 $h1
+}
+
+lldp_change_add_frag_size()
+{
+	local add_frag_size=$1
+
+	lldptool -T -i $h1 -V addEthCaps addFragSize=$add_frag_size >/dev/null
+	# Wait for TLVs to be received
+	sleep 2
+	lldptool -i $h2 -t -n -V addEthCaps | \
+		grep -q "Additional fragment size: $add_frag_size"
+}
+
+lldp()
+{
+	RET=0
+
+	systemctl start lldpad
+
+	# Configure the interfaces to receive and transmit LLDPDUs
+	lldptool -L -i $h1 adminStatus=rxtx >/dev/null
+	lldptool -L -i $h2 adminStatus=rxtx >/dev/null
+
+	# Enable the transmission of Additional Ethernet Capabilities TLV
+	lldptool -T -i $h1 -V addEthCaps enableTx=yes >/dev/null
+	lldptool -T -i $h2 -V addEthCaps enableTx=yes >/dev/null
+
+	# Wait for TLVs to be received
+	sleep 2
+
+	lldptool -i $h1 -t -n -V addEthCaps | \
+		grep -q "Preemption capability active"
+	check_err "$?" "$h1 pMAC TX is not active"
+
+	lldptool -i $h2 -t -n -V addEthCaps | \
+		grep -q "Preemption capability active"
+	check_err "$?" "$h2 pMAC TX is not active"
+
+	lldp_change_add_frag_size 3
+	check_err "$?" "addFragSize 3"
+
+	lldp_change_add_frag_size 2
+	check_err "$?" "addFragSize 2"
+
+	lldp_change_add_frag_size 1
+	check_err "$?" "addFragSize 1"
+
+	lldp_change_add_frag_size 0
+	check_err "$?" "addFragSize 0"
+
+	traffic_test $h1 "pmac"
+	check_err "$?" "Traffic did not get sent through $h1's pMAC"
+
+	traffic_test $h2 "pmac"
+	check_err "$?" "Traffic did not get sent through $h2's pMAC"
+
+	systemctl stop lldpad
+
+	log_test "LLDP"
+}
+
+h1_create()
+{
+	ip link set dev $h1 up
+
+	tc qdisc add dev $h1 root mqprio num_tc 4 map 0 1 2 3 \
+		queues 1@0 1@1 1@2 1@3 \
+		fp P E E E \
+		hw 1
+
+	ethtool --set-mm $h1 pmac-enabled on tx-enabled off verify-enabled off
+}
+
+h2_create()
+{
+	ip link set dev $h2 up
+
+	ethtool --set-mm $h2 pmac-enabled on tx-enabled off verify-enabled off
+
+	tc qdisc add dev $h2 root mqprio num_tc 4 map 0 1 2 3 \
+		queues 1@0 1@1 1@2 1@3 \
+		fp P E E E \
+		hw 1
+}
+
+h1_destroy()
+{
+	ethtool --set-mm $h1 pmac-enabled off tx-enabled off verify-enabled off
+
+	tc qdisc del dev $h1 root
+
+	ip link set dev $h1 down
+}
+
+h2_destroy()
+{
+	tc qdisc del dev $h2 root
+
+	ethtool --set-mm $h2 pmac-enabled off tx-enabled off verify-enabled off
+
+	ip link set dev $h2 down
+}
+
+setup_prepare()
+{
+	check_ethtool_mm_support
+	check_tc_fp_support
+	require_command lldptool
+	bail_on_lldpad "autoconfigure the MAC Merge layer" "configure it manually"
+
+	h1=${NETIFS[p1]}
+	h2=${NETIFS[p2]}
+
+	h1_create
+	h2_create
+}
+
+cleanup()
+{
+	pre_cleanup
+
+	h2_destroy
+	h1_destroy
+}
+
+trap cleanup EXIT
+
+setup_prepare
+setup_wait
+
+tests_run
+
+exit $EXIT_STATUS

tools/testing/selftests/net/forwarding/lib.sh

@@ -120,6 +120,15 @@ check_tc_action_hw_stats_support()
 	fi
 }
 
+check_tc_fp_support()
+{
+	tc qdisc add dev lo mqprio help 2>&1 | grep -q "fp "
+	if [[ $? -ne 0 ]]; then
+		echo "SKIP: iproute2 too old; tc is missing frame preemption support"
+		exit $ksft_skip
+	fi
+}
+
 check_ethtool_lanes_support()
 {
 	ethtool --help 2>&1| grep lanes &> /dev/null
@@ -129,6 +138,15 @@ check_ethtool_lanes_support()
 	fi
 }
 
+check_ethtool_mm_support()
+{
+	ethtool --help 2>&1| grep -- '--show-mm' &> /dev/null
+	if [[ $? -ne 0 ]]; then
+		echo "SKIP: ethtool too old; it is missing MAC Merge layer support"
+		exit $ksft_skip
+	fi
+}
+
 check_locked_port_support()
 {
 	if ! bridge -d link show | grep -q " locked"; then
@@ -787,6 +805,17 @@ ethtool_stats_get()
 	ethtool -S $dev | grep "^ *$stat:" | head -n 1 | cut -d: -f2
 }
 
+ethtool_std_stats_get()
+{
+	local dev=$1; shift
+	local grp=$1; shift
+	local name=$1; shift
+	local src=$1; shift
+
+	ethtool --json -S $dev --groups $grp -- --src $src | \
+		jq '.[]."'"$grp"'"."'$name'"'
+}
+
 qdisc_stats_get()
 {
 	local dev=$1; shift
@@ -1887,3 +1916,34 @@ mldv1_done_get()
 
 	payload_template_expand_checksum "$hbh$icmpv6" $checksum
 }
+
+bail_on_lldpad()
+{
+	local reason1="$1"; shift
+	local reason2="$1"; shift
+
+	if systemctl is-active --quiet lldpad; then
+
+		cat >/dev/stderr <<-EOF
+		WARNING: lldpad is running
+
+			lldpad will likely $reason1, and this test will
+			$reason2. Both are not supported at the same time,
+			one of them is arbitrarily going to overwrite the
+			other. That will cause spurious failures (or, unlikely,
+			passes) of this test.
+		EOF
+
+		if [[ -z $ALLOW_LLDPAD ]]; then
+			cat >/dev/stderr <<-EOF
+
+				If you want to run the test anyway, please set
+				an environment variable ALLOW_LLDPAD to a
+				non-empty string.
+			EOF
+			exit 1
+		else
+			return
+		fi
+	fi
+}

tools/testing/selftests/net/forwarding/sch_tbf_etsprio.sh

@@ -57,6 +57,10 @@ tbf_root_test()
 	tc qdisc del dev $swp2 root
 }
 
+if type -t sch_tbf_pre_hook >/dev/null; then
+	sch_tbf_pre_hook
+fi
+
 trap cleanup EXIT
 
 setup_prepare

tools/testing/selftests/net/forwarding/sch_tbf_root.sh

@@ -23,6 +23,10 @@ tbf_test()
 	tc qdisc del dev $swp2 root
 }
 
+if type -t sch_tbf_pre_hook >/dev/null; then
+	sch_tbf_pre_hook
+fi
+
 trap cleanup EXIT
 
 setup_prepare