Merge branch '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/next-queue

Tony Nguyen says:

====================
100GbE Intel Wired LAN Driver Updates 2021-10-19

This series contains updates to ice driver only.

Brett implements support for ndo_set_vf_rate allowing for min_tx_rate
and max_tx_rate to be set for a VF.

Jesse updates DIM moderation to improve latency and resolves problems
with reported rate limit and extra software generated interrupts.

Wojciech moves a check for trusted VFs to the correct function,
disables lb_en for switchdev offloads, and refactors ethtool ops due
to differences in support for PF and port representor support.

Cai Huoqing utilizes the helper function devm_add_action_or_reset().

Gustavo A. R. Silva replaces uses of allocation to devm_kcalloc() as
applicable.

Dan Carpenter propagates an error instead of returning success.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/intel/ice/ice_devlink.c

@@ -452,10 +452,8 @@ struct ice_pf *ice_allocate_pf(struct device *dev)
 		return NULL;
 
 	/* Add an action to teardown the devlink when unwinding the driver */
-	if (devm_add_action(dev, ice_devlink_free, devlink)) {
-		devlink_free(devlink);
+	if (devm_add_action_or_reset(dev, ice_devlink_free, devlink))
 		return NULL;
-	}
 
 	return devlink_priv(devlink);
 }

drivers/net/ethernet/intel/ice/ice_ethtool.c

@@ -192,7 +192,6 @@ __ice_get_drvinfo(struct net_device *netdev, struct ethtool_drvinfo *drvinfo,
 
 	strscpy(drvinfo->bus_info, pci_name(pf->pdev),
 		sizeof(drvinfo->bus_info));
-	drvinfo->n_priv_flags = ICE_PRIV_FLAG_ARRAY_SIZE;
 }
 
 static void
@@ -201,18 +200,8 @@ ice_get_drvinfo(struct net_device *netdev, struct ethtool_drvinfo *drvinfo)
 	struct ice_netdev_priv *np = netdev_priv(netdev);
 
 	__ice_get_drvinfo(netdev, drvinfo, np->vsi);
-}
-
-static void
-ice_repr_get_drvinfo(struct net_device *netdev,
-		     struct ethtool_drvinfo *drvinfo)
-{
-	struct ice_repr *repr = ice_netdev_to_repr(netdev);
-
-	if (ice_check_vf_ready_for_cfg(repr->vf))
-		return;
 
-	__ice_get_drvinfo(netdev, drvinfo, repr->src_vsi);
+	drvinfo->n_priv_flags = ICE_PRIV_FLAG_ARRAY_SIZE;
 }
 
 static int ice_get_regs_len(struct net_device __always_unused *netdev)
@@ -886,10 +875,10 @@ ice_self_test(struct net_device *netdev, struct ethtool_test *eth_test,
 	netdev_info(netdev, "testing finished\n");
 }
 
-static void ice_get_strings(struct net_device *netdev, u32 stringset, u8 *data)
+static void
+__ice_get_strings(struct net_device *netdev, u32 stringset, u8 *data,
+		  struct ice_vsi *vsi)
 {
-	struct ice_netdev_priv *np = netdev_priv(netdev);
-	struct ice_vsi *vsi = ice_get_netdev_priv_vsi(np);
 	unsigned int i;
 	u8 *p = data;
 
@@ -940,6 +929,13 @@ static void ice_get_strings(struct net_device *netdev, u32 stringset, u8 *data)
 	}
 }
 
+static void ice_get_strings(struct net_device *netdev, u32 stringset, u8 *data)
+{
+	struct ice_netdev_priv *np = netdev_priv(netdev);
+
+	__ice_get_strings(netdev, stringset, data, np->vsi);
+}
+
 static int
 ice_set_phys_id(struct net_device *netdev, enum ethtool_phys_id_state state)
 {
@@ -1331,9 +1327,6 @@ static int ice_get_sset_count(struct net_device *netdev, int sset)
 		 * order of strings will suffer from race conditions and are
 		 * not safe.
 		 */
-		if (ice_is_port_repr_netdev(netdev))
-			return ICE_VSI_STATS_LEN;
-
 		return ICE_ALL_STATS_LEN(netdev);
 	case ETH_SS_TEST:
 		return ICE_TEST_LEN;
@@ -1345,11 +1338,10 @@ static int ice_get_sset_count(struct net_device *netdev, int sset)
 }
 
 static void
-ice_get_ethtool_stats(struct net_device *netdev,
-		      struct ethtool_stats __always_unused *stats, u64 *data)
+__ice_get_ethtool_stats(struct net_device *netdev,
+			struct ethtool_stats __always_unused *stats, u64 *data,
+			struct ice_vsi *vsi)
 {
-	struct ice_netdev_priv *np = netdev_priv(netdev);
-	struct ice_vsi *vsi = ice_get_netdev_priv_vsi(np);
 	struct ice_pf *pf = vsi->back;
 	struct ice_tx_ring *tx_ring;
 	struct ice_rx_ring *rx_ring;
@@ -1416,6 +1408,15 @@ ice_get_ethtool_stats(struct net_device *netdev,
 	}
 }
 
+static void
+ice_get_ethtool_stats(struct net_device *netdev,
+		      struct ethtool_stats __always_unused *stats, u64 *data)
+{
+	struct ice_netdev_priv *np = netdev_priv(netdev);
+
+	__ice_get_ethtool_stats(netdev, stats, data, np->vsi);
+}
+
 #define ICE_PHY_TYPE_LOW_MASK_MIN_1G	(ICE_PHY_TYPE_LOW_100BASE_TX | \
 					 ICE_PHY_TYPE_LOW_100M_SGMII)
 
@@ -3640,6 +3641,9 @@ ice_set_rc_coalesce(struct ethtool_coalesce *ec,
 
 	switch (rc->type) {
 	case ICE_RX_CONTAINER:
+	{
+		struct ice_q_vector *q_vector = rc->rx_ring->q_vector;
+
 		if (ec->rx_coalesce_usecs_high > ICE_MAX_INTRL ||
 		    (ec->rx_coalesce_usecs_high &&
 		     ec->rx_coalesce_usecs_high < pf->hw.intrl_gran)) {
@@ -3648,22 +3652,20 @@ ice_set_rc_coalesce(struct ethtool_coalesce *ec,
 				    ICE_MAX_INTRL);
 			return -EINVAL;
 		}
-		if (ec->rx_coalesce_usecs_high != rc->rx_ring->q_vector->intrl &&
+		if (ec->rx_coalesce_usecs_high != q_vector->intrl &&
 		    (ec->use_adaptive_rx_coalesce || ec->use_adaptive_tx_coalesce)) {
 			netdev_info(vsi->netdev, "Invalid value, %s-usecs-high cannot be changed if adaptive-tx or adaptive-rx is enabled\n",
 				    c_type_str);
 			return -EINVAL;
 		}
-		if (ec->rx_coalesce_usecs_high != rc->rx_ring->q_vector->intrl) {
-			rc->rx_ring->q_vector->intrl = ec->rx_coalesce_usecs_high;
-			ice_write_intrl(rc->rx_ring->q_vector,
-					ec->rx_coalesce_usecs_high);
-		}
+		if (ec->rx_coalesce_usecs_high != q_vector->intrl)
+			q_vector->intrl = ec->rx_coalesce_usecs_high;
 
 		use_adaptive_coalesce = ec->use_adaptive_rx_coalesce;
 		coalesce_usecs = ec->rx_coalesce_usecs;
 
 		break;
+	}
 	case ICE_TX_CONTAINER:
 		use_adaptive_coalesce = ec->use_adaptive_tx_coalesce;
 		coalesce_usecs = ec->tx_coalesce_usecs;
@@ -3808,6 +3810,8 @@ __ice_set_coalesce(struct net_device *netdev, struct ethtool_coalesce *ec,
 
 			if (ice_set_q_coalesce(vsi, ec, v_idx))
 				return -EINVAL;
+
+			ice_set_q_vector_intrl(vsi->q_vectors[v_idx]);
 		}
 		goto set_complete;
 	}
@@ -3815,6 +3819,8 @@ __ice_set_coalesce(struct net_device *netdev, struct ethtool_coalesce *ec,
 	if (ice_set_q_coalesce(vsi, ec, q_num))
 		return -EINVAL;
 
+	ice_set_q_vector_intrl(vsi->q_vectors[q_num]);
+
 set_complete:
 	return 0;
 }
@@ -3834,6 +3840,54 @@ ice_set_per_q_coalesce(struct net_device *netdev, u32 q_num,
 	return __ice_set_coalesce(netdev, ec, q_num);
 }
 
+static void
+ice_repr_get_drvinfo(struct net_device *netdev,
+		     struct ethtool_drvinfo *drvinfo)
+{
+	struct ice_repr *repr = ice_netdev_to_repr(netdev);
+
+	if (ice_check_vf_ready_for_cfg(repr->vf))
+		return;
+
+	__ice_get_drvinfo(netdev, drvinfo, repr->src_vsi);
+}
+
+static void
+ice_repr_get_strings(struct net_device *netdev, u32 stringset, u8 *data)
+{
+	struct ice_repr *repr = ice_netdev_to_repr(netdev);
+
+	/* for port representors only ETH_SS_STATS is supported */
+	if (ice_check_vf_ready_for_cfg(repr->vf) ||
+	    stringset != ETH_SS_STATS)
+		return;
+
+	__ice_get_strings(netdev, stringset, data, repr->src_vsi);
+}
+
+static void
+ice_repr_get_ethtool_stats(struct net_device *netdev,
+			   struct ethtool_stats __always_unused *stats,
+			   u64 *data)
+{
+	struct ice_repr *repr = ice_netdev_to_repr(netdev);
+
+	if (ice_check_vf_ready_for_cfg(repr->vf))
+		return;
+
+	__ice_get_ethtool_stats(netdev, stats, data, repr->src_vsi);
+}
+
+static int ice_repr_get_sset_count(struct net_device *netdev, int sset)
+{
+	switch (sset) {
+	case ETH_SS_STATS:
+		return ICE_VSI_STATS_LEN;
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
 #define ICE_I2C_EEPROM_DEV_ADDR		0xA0
 #define ICE_I2C_EEPROM_DEV_ADDR2	0xA2
 #define ICE_MODULE_TYPE_SFP		0x03
@@ -4088,9 +4142,9 @@ void ice_set_ethtool_safe_mode_ops(struct net_device *netdev)
 static const struct ethtool_ops ice_ethtool_repr_ops = {
 	.get_drvinfo		= ice_repr_get_drvinfo,
 	.get_link		= ethtool_op_get_link,
-	.get_strings		= ice_get_strings,
-	.get_ethtool_stats      = ice_get_ethtool_stats,
-	.get_sset_count		= ice_get_sset_count,
+	.get_strings		= ice_repr_get_strings,
+	.get_ethtool_stats      = ice_repr_get_ethtool_stats,
+	.get_sset_count		= ice_repr_get_sset_count,
 };
 
 /**

drivers/net/ethernet/intel/ice/ice_ethtool_fdir.c

@@ -706,7 +706,7 @@ ice_create_init_fdir_rule(struct ice_pf *pf, enum ice_fltr_ptype flow)
 	if (!seg)
 		return -ENOMEM;
 
-	tun_seg = devm_kzalloc(dev, sizeof(*seg) * ICE_FD_HW_SEG_MAX,
+	tun_seg = devm_kcalloc(dev, sizeof(*seg), ICE_FD_HW_SEG_MAX,
 			       GFP_KERNEL);
 	if (!tun_seg) {
 		devm_kfree(dev, seg);
@@ -1068,7 +1068,7 @@ ice_cfg_fdir_xtrct_seq(struct ice_pf *pf, struct ethtool_rx_flow_spec *fsp,
 	if (!seg)
 		return -ENOMEM;
 
-	tun_seg = devm_kzalloc(dev, sizeof(*seg) * ICE_FD_HW_SEG_MAX,
+	tun_seg = devm_kcalloc(dev, sizeof(*seg), ICE_FD_HW_SEG_MAX,
 			       GFP_KERNEL);
 	if (!tun_seg) {
 		devm_kfree(dev, seg);

drivers/net/ethernet/intel/ice/ice_fltr.c

@@ -453,133 +453,6 @@ static u32 ice_fltr_build_action(u16 vsi_id)
 		ICE_SINGLE_ACT_VSI_FORWARDING | ICE_SINGLE_ACT_VALID_BIT;
 }
 
-/**
- * ice_fltr_find_adv_entry - find advanced rule
- * @rules: list of rules
- * @rule_id: id of wanted rule
- */
-static struct ice_adv_fltr_mgmt_list_entry *
-ice_fltr_find_adv_entry(struct list_head *rules, u16 rule_id)
-{
-	struct ice_adv_fltr_mgmt_list_entry *entry;
-
-	list_for_each_entry(entry, rules, list_entry) {
-		if (entry->rule_info.fltr_rule_id == rule_id)
-			return entry;
-	}
-
-	return NULL;
-}
-
-/**
- * ice_fltr_update_adv_rule_flags - update flags on advanced rule
- * @vsi: pointer to VSI
- * @recipe_id: id of recipe
- * @entry: advanced rule entry
- * @new_flags: flags to update
- */
-static enum ice_status
-ice_fltr_update_adv_rule_flags(struct ice_vsi *vsi, u16 recipe_id,
-			       struct ice_adv_fltr_mgmt_list_entry *entry,
-			       u32 new_flags)
-{
-	struct ice_adv_rule_info *info = &entry->rule_info;
-	struct ice_sw_act_ctrl *act = &info->sw_act;
-	u32 action;
-
-	if (act->fltr_act != ICE_FWD_TO_VSI)
-		return ICE_ERR_NOT_SUPPORTED;
-
-	action = ice_fltr_build_action(act->fwd_id.hw_vsi_id);
-
-	return ice_fltr_update_rule_flags(&vsi->back->hw, info->fltr_rule_id,
-					  recipe_id, action, info->sw_act.flag,
-					  act->src, new_flags);
-}
-
-/**
- * ice_fltr_find_regular_entry - find regular rule
- * @rules: list of rules
- * @rule_id: id of wanted rule
- */
-static struct ice_fltr_mgmt_list_entry *
-ice_fltr_find_regular_entry(struct list_head *rules, u16 rule_id)
-{
-	struct ice_fltr_mgmt_list_entry *entry;
-
-	list_for_each_entry(entry, rules, list_entry) {
-		if (entry->fltr_info.fltr_rule_id == rule_id)
-			return entry;
-	}
-
-	return NULL;
-}
-
-/**
- * ice_fltr_update_regular_rule - update flags on regular rule
- * @vsi: pointer to VSI
- * @recipe_id: id of recipe
- * @entry: regular rule entry
- * @new_flags: flags to update
- */
-static enum ice_status
-ice_fltr_update_regular_rule(struct ice_vsi *vsi, u16 recipe_id,
-			     struct ice_fltr_mgmt_list_entry *entry,
-			     u32 new_flags)
-{
-	struct ice_fltr_info *info = &entry->fltr_info;
-	u32 action;
-
-	if (info->fltr_act != ICE_FWD_TO_VSI)
-		return ICE_ERR_NOT_SUPPORTED;
-
-	action = ice_fltr_build_action(info->fwd_id.hw_vsi_id);
-
-	return ice_fltr_update_rule_flags(&vsi->back->hw, info->fltr_rule_id,
-					  recipe_id, action, info->flag,
-					  info->src, new_flags);
-}
-
-/**
- * ice_fltr_update_flags - update flags on rule
- * @vsi: pointer to VSI
- * @rule_id: id of rule
- * @recipe_id: id of recipe
- * @new_flags: flags to update
- *
- * Function updates flags on regular and advance rule.
- *
- * Flags should be a combination of ICE_SINGLE_ACT_LB_ENABLE and
- * ICE_SINGLE_ACT_LAN_ENABLE.
- */
-enum ice_status
-ice_fltr_update_flags(struct ice_vsi *vsi, u16 rule_id, u16 recipe_id,
-		      u32 new_flags)
-{
-	struct ice_adv_fltr_mgmt_list_entry *adv_entry;
-	struct ice_fltr_mgmt_list_entry *regular_entry;
-	struct ice_hw *hw = &vsi->back->hw;
-	struct ice_sw_recipe *recp_list;
-	struct list_head *fltr_rules;
-
-	recp_list = &hw->switch_info->recp_list[recipe_id];
-	if (!recp_list)
-		return ICE_ERR_DOES_NOT_EXIST;
-
-	fltr_rules = &recp_list->filt_rules;
-	regular_entry = ice_fltr_find_regular_entry(fltr_rules, rule_id);
-	if (regular_entry)
-		return ice_fltr_update_regular_rule(vsi, recipe_id,
-						    regular_entry, new_flags);
-
-	adv_entry = ice_fltr_find_adv_entry(fltr_rules, rule_id);
-	if (adv_entry)
-		return ice_fltr_update_adv_rule_flags(vsi, recipe_id,
-						      adv_entry, new_flags);
-
-	return ICE_ERR_DOES_NOT_EXIST;
-}
-
 /**
  * ice_fltr_update_flags_dflt_rule - update flags on default rule
  * @vsi: pointer to VSI

drivers/net/ethernet/intel/ice/ice_fltr.h

@@ -36,10 +36,6 @@ enum ice_status
 ice_fltr_remove_eth(struct ice_vsi *vsi, u16 ethertype, u16 flag,
 		    enum ice_sw_fwd_act_type action);
 void ice_fltr_remove_all(struct ice_vsi *vsi);
-
-enum ice_status
-ice_fltr_update_flags(struct ice_vsi *vsi, u16 rule_id, u16 recipe_id,
-		      u32 new_flags);
 enum ice_status
 ice_fltr_update_flags_dflt_rule(struct ice_vsi *vsi, u16 rule_id, u8 direction,
 				u32 new_flags);

drivers/net/ethernet/intel/ice/ice_hw_autogen.h

@@ -182,6 +182,7 @@
 #define GLINT_DYN_CTL_INTERVAL_S		5
 #define GLINT_DYN_CTL_INTERVAL_M		ICE_M(0xFFF, 5)
 #define GLINT_DYN_CTL_SW_ITR_INDX_ENA_M		BIT(24)
+#define GLINT_DYN_CTL_SW_ITR_INDX_S		25
 #define GLINT_DYN_CTL_SW_ITR_INDX_M		ICE_M(0x3, 25)
 #define GLINT_DYN_CTL_WB_ON_ITR_M		BIT(30)
 #define GLINT_DYN_CTL_INTENA_MSK_M		BIT(31)

drivers/net/ethernet/intel/ice/ice_lib.c

@@ -1941,6 +1941,31 @@ void ice_write_itr(struct ice_ring_container *rc, u16 itr)
 	__ice_write_itr(q_vector, rc, itr);
 }
 
+/**
+ * ice_set_q_vector_intrl - set up interrupt rate limiting
+ * @q_vector: the vector to be configured
+ *
+ * Interrupt rate limiting is local to the vector, not per-queue so we must
+ * detect if either ring container has dynamic moderation enabled to decide
+ * what to set the interrupt rate limit to via INTRL settings. In the case that
+ * dynamic moderation is disabled on both, write the value with the cached
+ * setting to make sure INTRL register matches the user visible value.
+ */
+void ice_set_q_vector_intrl(struct ice_q_vector *q_vector)
+{
+	if (ITR_IS_DYNAMIC(&q_vector->tx) || ITR_IS_DYNAMIC(&q_vector->rx)) {
+		/* in the case of dynamic enabled, cap each vector to no more
+		 * than (4 us) 250,000 ints/sec, which allows low latency
+		 * but still less than 500,000 interrupts per second, which
+		 * reduces CPU a bit in the case of the lowest latency
+		 * setting. The 4 here is a value in microseconds.
+		 */
+		ice_write_intrl(q_vector, 4);
+	} else {
+		ice_write_intrl(q_vector, q_vector->intrl);
+	}
+}
+
 /**
  * ice_vsi_cfg_msix - MSIX mode Interrupt Config in the HW
  * @vsi: the VSI being configured
@@ -3096,7 +3121,7 @@ ice_vsi_rebuild_set_coalesce(struct ice_vsi *vsi,
 		}
 
 		vsi->q_vectors[i]->intrl = coalesce[i].intrl;
-		ice_write_intrl(vsi->q_vectors[i], coalesce[i].intrl);
+		ice_set_q_vector_intrl(vsi->q_vectors[i]);
 	}
 
 	/* the number of queue vectors increased so write whatever is in
@@ -3114,7 +3139,7 @@ ice_vsi_rebuild_set_coalesce(struct ice_vsi *vsi,
 		ice_write_itr(rc, rc->itr_setting);
 
 		vsi->q_vectors[i]->intrl = coalesce[0].intrl;
-		ice_write_intrl(vsi->q_vectors[i], coalesce[0].intrl);
+		ice_set_q_vector_intrl(vsi->q_vectors[i]);
 	}
 }
 
@@ -3600,6 +3625,180 @@ int ice_clear_dflt_vsi(struct ice_sw *sw)
 	return 0;
 }
 
+/**
+ * ice_get_link_speed_mbps - get link speed in Mbps
+ * @vsi: the VSI whose link speed is being queried
+ *
+ * Return current VSI link speed and 0 if the speed is unknown.
+ */
+int ice_get_link_speed_mbps(struct ice_vsi *vsi)
+{
+	switch (vsi->port_info->phy.link_info.link_speed) {
+	case ICE_AQ_LINK_SPEED_100GB:
+		return SPEED_100000;
+	case ICE_AQ_LINK_SPEED_50GB:
+		return SPEED_50000;
+	case ICE_AQ_LINK_SPEED_40GB:
+		return SPEED_40000;
+	case ICE_AQ_LINK_SPEED_25GB:
+		return SPEED_25000;
+	case ICE_AQ_LINK_SPEED_20GB:
+		return SPEED_20000;
+	case ICE_AQ_LINK_SPEED_10GB:
+		return SPEED_10000;
+	case ICE_AQ_LINK_SPEED_5GB:
+		return SPEED_5000;
+	case ICE_AQ_LINK_SPEED_2500MB:
+		return SPEED_2500;
+	case ICE_AQ_LINK_SPEED_1000MB:
+		return SPEED_1000;
+	case ICE_AQ_LINK_SPEED_100MB:
+		return SPEED_100;
+	case ICE_AQ_LINK_SPEED_10MB:
+		return SPEED_10;
+	case ICE_AQ_LINK_SPEED_UNKNOWN:
+	default:
+		return 0;
+	}
+}
+
+/**
+ * ice_get_link_speed_kbps - get link speed in Kbps
+ * @vsi: the VSI whose link speed is being queried
+ *
+ * Return current VSI link speed and 0 if the speed is unknown.
+ */
+static int ice_get_link_speed_kbps(struct ice_vsi *vsi)
+{
+	int speed_mbps;
+
+	speed_mbps = ice_get_link_speed_mbps(vsi);
+
+	return speed_mbps * 1000;
+}
+
+/**
+ * ice_set_min_bw_limit - setup minimum BW limit for Tx based on min_tx_rate
+ * @vsi: VSI to be configured
+ * @min_tx_rate: min Tx rate in Kbps to be configured as BW limit
+ *
+ * If the min_tx_rate is specified as 0 that means to clear the minimum BW limit
+ * profile, otherwise a non-zero value will force a minimum BW limit for the VSI
+ * on TC 0.
+ */
+int ice_set_min_bw_limit(struct ice_vsi *vsi, u64 min_tx_rate)
+{
+	struct ice_pf *pf = vsi->back;
+	enum ice_status status;
+	struct device *dev;
+	int speed;
+
+	dev = ice_pf_to_dev(pf);
+	if (!vsi->port_info) {
+		dev_dbg(dev, "VSI %d, type %u specified doesn't have valid port_info\n",
+			vsi->idx, vsi->type);
+		return -EINVAL;
+	}
+
+	speed = ice_get_link_speed_kbps(vsi);
+	if (min_tx_rate > (u64)speed) {
+		dev_err(dev, "invalid min Tx rate %llu Kbps specified for %s %d is greater than current link speed %u Kbps\n",
+			min_tx_rate, ice_vsi_type_str(vsi->type), vsi->idx,
+			speed);
+		return -EINVAL;
+	}
+
+	/* Configure min BW for VSI limit */
+	if (min_tx_rate) {
+		status = ice_cfg_vsi_bw_lmt_per_tc(vsi->port_info, vsi->idx, 0,
+						   ICE_MIN_BW, min_tx_rate);
+		if (status) {
+			dev_err(dev, "failed to set min Tx rate(%llu Kbps) for %s %d\n",
+				min_tx_rate, ice_vsi_type_str(vsi->type),
+				vsi->idx);
+			return -EIO;
+		}
+
+		dev_dbg(dev, "set min Tx rate(%llu Kbps) for %s\n",
+			min_tx_rate, ice_vsi_type_str(vsi->type));
+	} else {
+		status = ice_cfg_vsi_bw_dflt_lmt_per_tc(vsi->port_info,
+							vsi->idx, 0,
+							ICE_MIN_BW);
+		if (status) {
+			dev_err(dev, "failed to clear min Tx rate configuration for %s %d\n",
+				ice_vsi_type_str(vsi->type), vsi->idx);
+			return -EIO;
+		}
+
+		dev_dbg(dev, "cleared min Tx rate configuration for %s %d\n",
+			ice_vsi_type_str(vsi->type), vsi->idx);
+	}
+
+	return 0;
+}
+
+/**
+ * ice_set_max_bw_limit - setup maximum BW limit for Tx based on max_tx_rate
+ * @vsi: VSI to be configured
+ * @max_tx_rate: max Tx rate in Kbps to be configured as BW limit
+ *
+ * If the max_tx_rate is specified as 0 that means to clear the maximum BW limit
+ * profile, otherwise a non-zero value will force a maximum BW limit for the VSI
+ * on TC 0.
+ */
+int ice_set_max_bw_limit(struct ice_vsi *vsi, u64 max_tx_rate)
+{
+	struct ice_pf *pf = vsi->back;
+	enum ice_status status;
+	struct device *dev;
+	int speed;
+
+	dev = ice_pf_to_dev(pf);
+	if (!vsi->port_info) {
+		dev_dbg(dev, "VSI %d, type %u specified doesn't have valid port_info\n",
+			vsi->idx, vsi->type);
+		return -EINVAL;
+	}
+
+	speed = ice_get_link_speed_kbps(vsi);
+	if (max_tx_rate > (u64)speed) {
+		dev_err(dev, "invalid max Tx rate %llu Kbps specified for %s %d is greater than current link speed %u Kbps\n",
+			max_tx_rate, ice_vsi_type_str(vsi->type), vsi->idx,
+			speed);
+		return -EINVAL;
+	}
+
+	/* Configure max BW for VSI limit */
+	if (max_tx_rate) {
+		status = ice_cfg_vsi_bw_lmt_per_tc(vsi->port_info, vsi->idx, 0,
+						   ICE_MAX_BW, max_tx_rate);
+		if (status) {
+			dev_err(dev, "failed setting max Tx rate(%llu Kbps) for %s %d\n",
+				max_tx_rate, ice_vsi_type_str(vsi->type),
+				vsi->idx);
+			return -EIO;
+		}
+
+		dev_dbg(dev, "set max Tx rate(%llu Kbps) for %s %d\n",
+			max_tx_rate, ice_vsi_type_str(vsi->type), vsi->idx);
+	} else {
+		status = ice_cfg_vsi_bw_dflt_lmt_per_tc(vsi->port_info,
+							vsi->idx, 0,
+							ICE_MAX_BW);
+		if (status) {
+			dev_err(dev, "failed clearing max Tx rate configuration for %s %d\n",
+				ice_vsi_type_str(vsi->type), vsi->idx);
+			return -EIO;
+		}
+
+		dev_dbg(dev, "cleared max Tx rate configuration for %s %d\n",
+			ice_vsi_type_str(vsi->type), vsi->idx);
+	}
+
+	return 0;
+}
+
 /**
  * ice_set_link - turn on/off physical link
  * @vsi: VSI to modify physical link on

drivers/net/ethernet/intel/ice/ice_lib.h

@@ -103,6 +103,7 @@ int ice_status_to_errno(enum ice_status err);
 
 void ice_write_intrl(struct ice_q_vector *q_vector, u8 intrl);
 void ice_write_itr(struct ice_ring_container *rc, u16 itr);
+void ice_set_q_vector_intrl(struct ice_q_vector *q_vector);
 
 enum ice_status
 ice_vsi_cfg_mac_fltr(struct ice_vsi *vsi, const u8 *macaddr, bool set);
@@ -116,7 +117,9 @@ bool ice_is_vsi_dflt_vsi(struct ice_sw *sw, struct ice_vsi *vsi);
 int ice_set_dflt_vsi(struct ice_sw *sw, struct ice_vsi *vsi);
 
 int ice_clear_dflt_vsi(struct ice_sw *sw);
-
+int ice_set_min_bw_limit(struct ice_vsi *vsi, u64 min_tx_rate);
+int ice_set_max_bw_limit(struct ice_vsi *vsi, u64 max_tx_rate);
+int ice_get_link_speed_mbps(struct ice_vsi *vsi);
 int
 ice_vsi_update_security(struct ice_vsi *vsi, void (*fill)(struct ice_vsi_ctx *));
 

drivers/net/ethernet/intel/ice/ice_main.c

@@ -5502,77 +5502,59 @@ int ice_vsi_cfg(struct ice_vsi *vsi)
 }
 
 /* THEORY OF MODERATION:
- * The below code creates custom DIM profiles for use by this driver, because
- * the ice driver hardware works differently than the hardware that DIMLIB was
+ * The ice driver hardware works differently than the hardware that DIMLIB was
  * originally made for. ice hardware doesn't have packet count limits that
  * can trigger an interrupt, but it *does* have interrupt rate limit support,
- * and this code adds that capability to be used by the driver when it's using
- * DIMLIB. The DIMLIB code was always designed to be a suggestion to the driver
- * for how to "respond" to traffic and interrupts, so this driver uses a
- * slightly different set of moderation parameters to get best performance.
+ * which is hard-coded to a limit of 250,000 ints/second.
+ * If not using dynamic moderation, the INTRL value can be modified
+ * by ethtool rx-usecs-high.
  */
 struct ice_dim {
 	/* the throttle rate for interrupts, basically worst case delay before
 	 * an initial interrupt fires, value is stored in microseconds.
 	 */
 	u16 itr;
-	/* the rate limit for interrupts, which can cap a delay from a small
-	 * ITR at a certain amount of interrupts per second. f.e. a 2us ITR
-	 * could yield as much as 500,000 interrupts per second, but with a
-	 * 10us rate limit, it limits to 100,000 interrupts per second. Value
-	 * is stored in microseconds.
-	 */
-	u16 intrl;
 };
 
 /* Make a different profile for Rx that doesn't allow quite so aggressive
- * moderation at the high end (it maxes out at 128us or about 8k interrupts a
- * second. The INTRL/rate parameters here are only useful to cap small ITR
- * values, which is why for larger ITR's - like 128, which can only generate
- * 8k interrupts per second, there is no point to rate limit and the values
- * are set to zero. The rate limit values do affect latency, and so must
- * be reasonably small so to not impact latency sensitive tests.
+ * moderation at the high end (it maxes out at 126us or about 8k interrupts a
+ * second.
  */
 static const struct ice_dim rx_profile[] = {
-	{2, 10},
-	{8, 16},
-	{32, 0},
-	{96, 0},
-	{128, 0}
+	{2},    /* 500,000 ints/s, capped at 250K by INTRL */
+	{8},    /* 125,000 ints/s */
+	{16},   /*  62,500 ints/s */
+	{62},   /*  16,129 ints/s */
+	{126}   /*   7,936 ints/s */
 };
 
 /* The transmit profile, which has the same sorts of values
  * as the previous struct
  */
 static const struct ice_dim tx_profile[] = {
-	{2, 10},
-	{8, 16},
-	{64, 0},
-	{128, 0},
-	{256, 0}
+	{2},    /* 500,000 ints/s, capped at 250K by INTRL */
+	{8},    /* 125,000 ints/s */
+	{40},   /*  16,125 ints/s */
+	{128},  /*   7,812 ints/s */
+	{256}   /*   3,906 ints/s */
 };
 
 static void ice_tx_dim_work(struct work_struct *work)
 {
 	struct ice_ring_container *rc;
-	struct ice_q_vector *q_vector;
 	struct dim *dim;
-	u16 itr, intrl;
+	u16 itr;
 
 	dim = container_of(work, struct dim, work);
-	rc = container_of(dim, struct ice_ring_container, dim);
-	q_vector = container_of(rc, struct ice_q_vector, tx);
+	rc = (struct ice_ring_container *)dim->priv;
 
-	if (dim->profile_ix >= ARRAY_SIZE(tx_profile))
-		dim->profile_ix = ARRAY_SIZE(tx_profile) - 1;
+	WARN_ON(dim->profile_ix >= ARRAY_SIZE(tx_profile));
 
 	/* look up the values in our local table */
 	itr = tx_profile[dim->profile_ix].itr;
-	intrl = tx_profile[dim->profile_ix].intrl;
 
-	ice_trace(tx_dim_work, q_vector, dim);
+	ice_trace(tx_dim_work, container_of(rc, struct ice_q_vector, tx), dim);
 	ice_write_itr(rc, itr);
-	ice_write_intrl(q_vector, intrl);
 
 	dim->state = DIM_START_MEASURE;
 }
@@ -5580,28 +5562,65 @@ static void ice_tx_dim_work(struct work_struct *work)
 static void ice_rx_dim_work(struct work_struct *work)
 {
 	struct ice_ring_container *rc;
-	struct ice_q_vector *q_vector;
 	struct dim *dim;
-	u16 itr, intrl;
+	u16 itr;
 
 	dim = container_of(work, struct dim, work);
-	rc = container_of(dim, struct ice_ring_container, dim);
-	q_vector = container_of(rc, struct ice_q_vector, rx);
+	rc = (struct ice_ring_container *)dim->priv;
 
-	if (dim->profile_ix >= ARRAY_SIZE(rx_profile))
-		dim->profile_ix = ARRAY_SIZE(rx_profile) - 1;
+	WARN_ON(dim->profile_ix >= ARRAY_SIZE(rx_profile));
 
 	/* look up the values in our local table */
 	itr = rx_profile[dim->profile_ix].itr;
-	intrl = rx_profile[dim->profile_ix].intrl;
 
-	ice_trace(rx_dim_work, q_vector, dim);
+	ice_trace(rx_dim_work, container_of(rc, struct ice_q_vector, rx), dim);
 	ice_write_itr(rc, itr);
-	ice_write_intrl(q_vector, intrl);
 
 	dim->state = DIM_START_MEASURE;
 }
 
+#define ICE_DIM_DEFAULT_PROFILE_IX 1
+
+/**
+ * ice_init_moderation - set up interrupt moderation
+ * @q_vector: the vector containing rings to be configured
+ *
+ * Set up interrupt moderation registers, with the intent to do the right thing
+ * when called from reset or from probe, and whether or not dynamic moderation
+ * is enabled or not. Take special care to write all the registers in both
+ * dynamic moderation mode or not in order to make sure hardware is in a known
+ * state.
+ */
+static void ice_init_moderation(struct ice_q_vector *q_vector)
+{
+	struct ice_ring_container *rc;
+	bool tx_dynamic, rx_dynamic;
+
+	rc = &q_vector->tx;
+	INIT_WORK(&rc->dim.work, ice_tx_dim_work);
+	rc->dim.mode = DIM_CQ_PERIOD_MODE_START_FROM_EQE;
+	rc->dim.profile_ix = ICE_DIM_DEFAULT_PROFILE_IX;
+	rc->dim.priv = rc;
+	tx_dynamic = ITR_IS_DYNAMIC(rc);
+
+	/* set the initial TX ITR to match the above */
+	ice_write_itr(rc, tx_dynamic ?
+		      tx_profile[rc->dim.profile_ix].itr : rc->itr_setting);
+
+	rc = &q_vector->rx;
+	INIT_WORK(&rc->dim.work, ice_rx_dim_work);
+	rc->dim.mode = DIM_CQ_PERIOD_MODE_START_FROM_EQE;
+	rc->dim.profile_ix = ICE_DIM_DEFAULT_PROFILE_IX;
+	rc->dim.priv = rc;
+	rx_dynamic = ITR_IS_DYNAMIC(rc);
+
+	/* set the initial RX ITR to match the above */
+	ice_write_itr(rc, rx_dynamic ? rx_profile[rc->dim.profile_ix].itr :
+				       rc->itr_setting);
+
+	ice_set_q_vector_intrl(q_vector);
+}
+
 /**
  * ice_napi_enable_all - Enable NAPI for all q_vectors in the VSI
  * @vsi: the VSI being configured
@@ -5616,11 +5635,7 @@ static void ice_napi_enable_all(struct ice_vsi *vsi)
 	ice_for_each_q_vector(vsi, q_idx) {
 		struct ice_q_vector *q_vector = vsi->q_vectors[q_idx];
 
-		INIT_WORK(&q_vector->tx.dim.work, ice_tx_dim_work);
-		q_vector->tx.dim.mode = DIM_CQ_PERIOD_MODE_START_FROM_EQE;
-
-		INIT_WORK(&q_vector->rx.dim.work, ice_rx_dim_work);
-		q_vector->rx.dim.mode = DIM_CQ_PERIOD_MODE_START_FROM_EQE;
+		ice_init_moderation(q_vector);
 
 		if (q_vector->rx.rx_ring || q_vector->tx.tx_ring)
 			napi_enable(&q_vector->napi);
@@ -7390,6 +7405,7 @@ static const struct net_device_ops ice_netdev_ops = {
 	.ndo_set_vf_vlan = ice_set_vf_port_vlan,
 	.ndo_set_vf_link_state = ice_set_vf_link_state,
 	.ndo_get_vf_stats = ice_get_vf_stats,
+	.ndo_set_vf_rate = ice_set_vf_bw,
 	.ndo_vlan_rx_add_vid = ice_vlan_rx_add_vid,
 	.ndo_vlan_rx_kill_vid = ice_vlan_rx_kill_vid,
 	.ndo_setup_tc = ice_setup_tc,

drivers/net/ethernet/intel/ice/ice_sched.c

@@ -3770,6 +3770,136 @@ ice_cfg_q_bw_dflt_lmt(struct ice_port_info *pi, u16 vsi_handle, u8 tc,
 				      ICE_SCHED_DFLT_BW);
 }
 
+/**
+ * ice_sched_get_node_by_id_type - get node from ID type
+ * @pi: port information structure
+ * @id: identifier
+ * @agg_type: type of aggregator
+ * @tc: traffic class
+ *
+ * This function returns node identified by ID of type aggregator, and
+ * based on traffic class (TC). This function needs to be called with
+ * the scheduler lock held.
+ */
+static struct ice_sched_node *
+ice_sched_get_node_by_id_type(struct ice_port_info *pi, u32 id,
+			      enum ice_agg_type agg_type, u8 tc)
+{
+	struct ice_sched_node *node = NULL;
+
+	switch (agg_type) {
+	case ICE_AGG_TYPE_VSI: {
+		struct ice_vsi_ctx *vsi_ctx;
+		u16 vsi_handle = (u16)id;
+
+		if (!ice_is_vsi_valid(pi->hw, vsi_handle))
+			break;
+		/* Get sched_vsi_info */
+		vsi_ctx = ice_get_vsi_ctx(pi->hw, vsi_handle);
+		if (!vsi_ctx)
+			break;
+		node = vsi_ctx->sched.vsi_node[tc];
+		break;
+	}
+
+	case ICE_AGG_TYPE_AGG: {
+		struct ice_sched_node *tc_node;
+
+		tc_node = ice_sched_get_tc_node(pi, tc);
+		if (tc_node)
+			node = ice_sched_get_agg_node(pi, tc_node, id);
+		break;
+	}
+
+	default:
+		break;
+	}
+
+	return node;
+}
+
+/**
+ * ice_sched_set_node_bw_lmt_per_tc - set node BW limit per TC
+ * @pi: port information structure
+ * @id: ID (software VSI handle or AGG ID)
+ * @agg_type: aggregator type (VSI or AGG type node)
+ * @tc: traffic class
+ * @rl_type: min or max
+ * @bw: bandwidth in Kbps
+ *
+ * This function sets BW limit of VSI or Aggregator scheduling node
+ * based on TC information from passed in argument BW.
+ */
+static enum ice_status
+ice_sched_set_node_bw_lmt_per_tc(struct ice_port_info *pi, u32 id,
+				 enum ice_agg_type agg_type, u8 tc,
+				 enum ice_rl_type rl_type, u32 bw)
+{
+	enum ice_status status = ICE_ERR_PARAM;
+	struct ice_sched_node *node;
+
+	if (!pi)
+		return status;
+
+	if (rl_type == ICE_UNKNOWN_BW)
+		return status;
+
+	mutex_lock(&pi->sched_lock);
+	node = ice_sched_get_node_by_id_type(pi, id, agg_type, tc);
+	if (!node) {
+		ice_debug(pi->hw, ICE_DBG_SCHED, "Wrong id, agg type, or tc\n");
+		goto exit_set_node_bw_lmt_per_tc;
+	}
+	if (bw == ICE_SCHED_DFLT_BW)
+		status = ice_sched_set_node_bw_dflt_lmt(pi, node, rl_type);
+	else
+		status = ice_sched_set_node_bw_lmt(pi, node, rl_type, bw);
+
+exit_set_node_bw_lmt_per_tc:
+	mutex_unlock(&pi->sched_lock);
+	return status;
+}
+
+/**
+ * ice_cfg_vsi_bw_lmt_per_tc - configure VSI BW limit per TC
+ * @pi: port information structure
+ * @vsi_handle: software VSI handle
+ * @tc: traffic class
+ * @rl_type: min or max
+ * @bw: bandwidth in Kbps
+ *
+ * This function configures BW limit of VSI scheduling node based on TC
+ * information.
+ */
+enum ice_status
+ice_cfg_vsi_bw_lmt_per_tc(struct ice_port_info *pi, u16 vsi_handle, u8 tc,
+			  enum ice_rl_type rl_type, u32 bw)
+{
+	return ice_sched_set_node_bw_lmt_per_tc(pi, vsi_handle,
+						ICE_AGG_TYPE_VSI,
+						tc, rl_type, bw);
+}
+
+/**
+ * ice_cfg_vsi_bw_dflt_lmt_per_tc - configure default VSI BW limit per TC
+ * @pi: port information structure
+ * @vsi_handle: software VSI handle
+ * @tc: traffic class
+ * @rl_type: min or max
+ *
+ * This function configures default BW limit of VSI scheduling node based on TC
+ * information.
+ */
+enum ice_status
+ice_cfg_vsi_bw_dflt_lmt_per_tc(struct ice_port_info *pi, u16 vsi_handle, u8 tc,
+			       enum ice_rl_type rl_type)
+{
+	return ice_sched_set_node_bw_lmt_per_tc(pi, vsi_handle,
+						ICE_AGG_TYPE_VSI,
+						tc, rl_type,
+						ICE_SCHED_DFLT_BW);
+}
+
 /**
  * ice_cfg_rl_burst_size - Set burst size value
  * @hw: pointer to the HW struct

drivers/net/ethernet/intel/ice/ice_sched.h

@@ -103,6 +103,12 @@ ice_cfg_q_bw_lmt(struct ice_port_info *pi, u16 vsi_handle, u8 tc,
 enum ice_status
 ice_cfg_q_bw_dflt_lmt(struct ice_port_info *pi, u16 vsi_handle, u8 tc,
 		      u16 q_handle, enum ice_rl_type rl_type);
+enum ice_status
+ice_cfg_vsi_bw_lmt_per_tc(struct ice_port_info *pi, u16 vsi_handle, u8 tc,
+			  enum ice_rl_type rl_type, u32 bw);
+enum ice_status
+ice_cfg_vsi_bw_dflt_lmt_per_tc(struct ice_port_info *pi, u16 vsi_handle, u8 tc,
+			       enum ice_rl_type rl_type);
 enum ice_status ice_cfg_rl_burst_size(struct ice_hw *hw, u32 bytes);
 void ice_sched_replay_agg_vsi_preinit(struct ice_hw *hw);
 void ice_sched_replay_agg(struct ice_hw *hw);

drivers/net/ethernet/intel/ice/ice_switch.c

@@ -4783,7 +4783,14 @@ ice_add_adv_rule(struct ice_hw *hw, struct ice_adv_lkup_elem *lkups,
 	s_rule = kzalloc(rule_buf_sz, GFP_KERNEL);
 	if (!s_rule)
 		return ICE_ERR_NO_MEMORY;
-	act |= ICE_SINGLE_ACT_LB_ENABLE | ICE_SINGLE_ACT_LAN_ENABLE;
+	if (!rinfo->flags_info.act_valid) {
+		act |= ICE_SINGLE_ACT_LAN_ENABLE;
+		act |= ICE_SINGLE_ACT_LB_ENABLE;
+	} else {
+		act |= rinfo->flags_info.act & (ICE_SINGLE_ACT_LAN_ENABLE |
+						ICE_SINGLE_ACT_LB_ENABLE);
+	}
+
 	switch (rinfo->sw_act.fltr_act) {
 	case ICE_FWD_TO_VSI:
 		act |= (rinfo->sw_act.fwd_id.hw_vsi_id <<

drivers/net/ethernet/intel/ice/ice_switch.h

@@ -160,11 +160,22 @@ struct ice_rule_query_data {
 	u16 vsi_handle;
 };
 
+/* This structure allows to pass info about lb_en and lan_en
+ * flags to ice_add_adv_rule. Values in act would be used
+ * only if act_valid was set to true, otherwise default
+ * values would be used.
+ */
+struct ice_adv_rule_flags_info {
+	u32 act;
+	u8 act_valid;		/* indicate if flags in act are valid */
+};
+
 struct ice_adv_rule_info {
 	struct ice_sw_act_ctrl sw_act;
 	u32 priority;
 	u8 rx; /* true means LOOKUP_RX otherwise LOOKUP_TX */
 	u16 fltr_rule_id;
+	struct ice_adv_rule_flags_info flags_info;
 };
 
 /* A collection of one or more four word recipe */

drivers/net/ethernet/intel/ice/ice_tc_lib.c

@@ -274,6 +274,8 @@ ice_eswitch_add_tc_fltr(struct ice_vsi *vsi, struct ice_tc_flower_fltr *fltr)
 		rule_info.sw_act.flag |= ICE_FLTR_TX;
 		rule_info.sw_act.src = vsi->idx;
 		rule_info.rx = false;
+		rule_info.flags_info.act = ICE_SINGLE_ACT_LAN_ENABLE;
+		rule_info.flags_info.act_valid = true;
 	}
 
 	/* specify the cookie as filter_rule_id */
@@ -296,12 +298,6 @@ ice_eswitch_add_tc_fltr(struct ice_vsi *vsi, struct ice_tc_flower_fltr *fltr)
 	fltr->rid = rule_added.rid;
 	fltr->rule_id = rule_added.rule_id;
 
-	if (fltr->direction == ICE_ESWITCH_FLTR_EGRESS) {
-		if (ice_fltr_update_flags(vsi, fltr->rule_id, fltr->rid,
-					  ICE_SINGLE_ACT_LAN_ENABLE))
-			ice_rem_adv_rule_by_id(hw, &rule_added);
-	}
-
 exit:
 	kfree(list);
 	return ret;

drivers/net/ethernet/intel/ice/ice_txrx.c

@@ -343,7 +343,7 @@ int ice_setup_tx_ring(struct ice_tx_ring *tx_ring)
 	/* warn if we are about to overwrite the pointer */
 	WARN_ON(tx_ring->tx_buf);
 	tx_ring->tx_buf =
-		devm_kzalloc(dev, sizeof(*tx_ring->tx_buf) * tx_ring->count,
+		devm_kcalloc(dev, sizeof(*tx_ring->tx_buf), tx_ring->count,
 			     GFP_KERNEL);
 	if (!tx_ring->tx_buf)
 		return -ENOMEM;
@@ -475,7 +475,7 @@ int ice_setup_rx_ring(struct ice_rx_ring *rx_ring)
 	/* warn if we are about to overwrite the pointer */
 	WARN_ON(rx_ring->rx_buf);
 	rx_ring->rx_buf =
-		devm_kzalloc(dev, sizeof(*rx_ring->rx_buf) * rx_ring->count,
+		devm_kcalloc(dev, sizeof(*rx_ring->rx_buf), rx_ring->count,
 			     GFP_KERNEL);
 	if (!rx_ring->rx_buf)
 		return -ENOMEM;
@@ -1259,6 +1259,41 @@ int ice_clean_rx_irq(struct ice_rx_ring *rx_ring, int budget)
 	return failure ? budget : (int)total_rx_pkts;
 }
 
+static void __ice_update_sample(struct ice_q_vector *q_vector,
+				struct ice_ring_container *rc,
+				struct dim_sample *sample,
+				bool is_tx)
+{
+	u64 packets = 0, bytes = 0;
+
+	if (is_tx) {
+		struct ice_tx_ring *tx_ring;
+
+		ice_for_each_tx_ring(tx_ring, *rc) {
+			packets += tx_ring->stats.pkts;
+			bytes += tx_ring->stats.bytes;
+		}
+	} else {
+		struct ice_rx_ring *rx_ring;
+
+		ice_for_each_rx_ring(rx_ring, *rc) {
+			packets += rx_ring->stats.pkts;
+			bytes += rx_ring->stats.bytes;
+		}
+	}
+
+	dim_update_sample(q_vector->total_events, packets, bytes, sample);
+	sample->comp_ctr = 0;
+
+	/* if dim settings get stale, like when not updated for 1
+	 * second or longer, force it to start again. This addresses the
+	 * frequent case of an idle queue being switched to by the
+	 * scheduler. The 1,000 here means 1,000 milliseconds.
+	 */
+	if (ktime_ms_delta(sample->time, rc->dim.start_sample.time) >= 1000)
+		rc->dim.state = DIM_START_MEASURE;
+}
+
 /**
  * ice_net_dim - Update net DIM algorithm
  * @q_vector: the vector associated with the interrupt
@@ -1274,34 +1309,16 @@ static void ice_net_dim(struct ice_q_vector *q_vector)
 	struct ice_ring_container *rx = &q_vector->rx;
 
 	if (ITR_IS_DYNAMIC(tx)) {
-		struct dim_sample dim_sample = {};
-		u64 packets = 0, bytes = 0;
-		struct ice_tx_ring *ring;
-
-		ice_for_each_tx_ring(ring, q_vector->tx) {
-			packets += ring->stats.pkts;
-			bytes += ring->stats.bytes;
-		}
-
-		dim_update_sample(q_vector->total_events, packets, bytes,
-				  &dim_sample);
+		struct dim_sample dim_sample;
 
+		__ice_update_sample(q_vector, tx, &dim_sample, true);
 		net_dim(&tx->dim, dim_sample);
 	}
 
 	if (ITR_IS_DYNAMIC(rx)) {
-		struct dim_sample dim_sample = {};
-		u64 packets = 0, bytes = 0;
-		struct ice_rx_ring *ring;
-
-		ice_for_each_rx_ring(ring, q_vector->rx) {
-			packets += ring->stats.pkts;
-			bytes += ring->stats.bytes;
-		}
-
-		dim_update_sample(q_vector->total_events, packets, bytes,
-				  &dim_sample);
+		struct dim_sample dim_sample;
 
+		__ice_update_sample(q_vector, rx, &dim_sample, false);
 		net_dim(&rx->dim, dim_sample);
 	}
 }
@@ -1328,15 +1345,14 @@ static u32 ice_buildreg_itr(u16 itr_idx, u16 itr)
 }
 
 /**
- * ice_update_ena_itr - Update ITR moderation and re-enable MSI-X interrupt
+ * ice_enable_interrupt - re-enable MSI-X interrupt
  * @q_vector: the vector associated with the interrupt to enable
  *
- * Update the net_dim() algorithm and re-enable the interrupt associated with
- * this vector.
- *
- * If the VSI is down, the interrupt will not be re-enabled.
+ * If the VSI is down, the interrupt will not be re-enabled. Also,
+ * when enabling the interrupt always reset the wb_on_itr to false
+ * and trigger a software interrupt to clean out internal state.
  */
-static void ice_update_ena_itr(struct ice_q_vector *q_vector)
+static void ice_enable_interrupt(struct ice_q_vector *q_vector)
 {
 	struct ice_vsi *vsi = q_vector->vsi;
 	bool wb_en = q_vector->wb_on_itr;
@@ -1345,25 +1361,25 @@ static void ice_update_ena_itr(struct ice_q_vector *q_vector)
 	if (test_bit(ICE_DOWN, vsi->state))
 		return;
 
-	/* When exiting WB_ON_ITR, let ITR resume its normal
-	 * interrupts-enabled path.
+	/* trigger an ITR delayed software interrupt when exiting busy poll, to
+	 * make sure to catch any pending cleanups that might have been missed
+	 * due to interrupt state transition. If busy poll or poll isn't
+	 * enabled, then don't update ITR, and just enable the interrupt.
 	 */
-	if (wb_en)
+	if (!wb_en) {
+		itr_val = ice_buildreg_itr(ICE_ITR_NONE, 0);
+	} else {
 		q_vector->wb_on_itr = false;
 
-	/* This will do nothing if dynamic updates are not enabled. */
-	ice_net_dim(q_vector);
-
-	/* net_dim() updates ITR out-of-band using a work item */
-	itr_val = ice_buildreg_itr(ICE_ITR_NONE, 0);
-	/* trigger an immediate software interrupt when exiting
-	 * busy poll, to make sure to catch any pending cleanups
-	 * that might have been missed due to interrupt state
-	 * transition.
-	 */
-	if (wb_en) {
+		/* do two things here with a single write. Set up the third ITR
+		 * index to be used for software interrupt moderation, and then
+		 * trigger a software interrupt with a rate limit of 20K on
+		 * software interrupts, this will help avoid high interrupt
+		 * loads due to frequently polling and exiting polling.
+		 */
+		itr_val = ice_buildreg_itr(ICE_IDX_ITR2, ICE_ITR_20K);
 		itr_val |= GLINT_DYN_CTL_SWINT_TRIG_M |
-			   GLINT_DYN_CTL_SW_ITR_INDX_M |
+			   ICE_IDX_ITR2 << GLINT_DYN_CTL_SW_ITR_INDX_S |
 			   GLINT_DYN_CTL_SW_ITR_INDX_ENA_M;
 	}
 	wr32(&vsi->back->hw, GLINT_DYN_CTL(q_vector->reg_idx), itr_val);
@@ -1482,10 +1498,12 @@ int ice_napi_poll(struct napi_struct *napi, int budget)
 	/* Exit the polling mode, but don't re-enable interrupts if stack might
 	 * poll us due to busy-polling
 	 */
-	if (likely(napi_complete_done(napi, work_done)))
-		ice_update_ena_itr(q_vector);
-	else
+	if (likely(napi_complete_done(napi, work_done))) {
+		ice_net_dim(q_vector);
+		ice_enable_interrupt(q_vector);
+	} else {
 		ice_set_wb_on_itr(q_vector);
+	}
 
 	return min_t(int, work_done, budget - 1);
 }

drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c

@@ -5,6 +5,7 @@
 #include "ice_base.h"
 #include "ice_lib.h"
 #include "ice_fltr.h"
+#include "ice_dcb_lib.h"
 #include "ice_flow.h"
 #include "ice_eswitch.h"
 #include "ice_virtchnl_allowlist.h"
@@ -884,6 +885,40 @@ static int ice_calc_vf_first_vector_idx(struct ice_pf *pf, struct ice_vf *vf)
 	return pf->sriov_base_vector + vf->vf_id * pf->num_msix_per_vf;
 }
 
+/**
+ * ice_vf_rebuild_host_tx_rate_cfg - re-apply the Tx rate limiting configuration
+ * @vf: VF to re-apply the configuration for
+ *
+ * Called after a VF VSI has been re-added/rebuild during reset. The PF driver
+ * needs to re-apply the host configured Tx rate limiting configuration.
+ */
+static int ice_vf_rebuild_host_tx_rate_cfg(struct ice_vf *vf)
+{
+	struct device *dev = ice_pf_to_dev(vf->pf);
+	struct ice_vsi *vsi = ice_get_vf_vsi(vf);
+	int err;
+
+	if (vf->min_tx_rate) {
+		err = ice_set_min_bw_limit(vsi, (u64)vf->min_tx_rate * 1000);
+		if (err) {
+			dev_err(dev, "failed to set min Tx rate to %d Mbps for VF %u, error %d\n",
+				vf->min_tx_rate, vf->vf_id, err);
+			return err;
+		}
+	}
+
+	if (vf->max_tx_rate) {
+		err = ice_set_max_bw_limit(vsi, (u64)vf->max_tx_rate * 1000);
+		if (err) {
+			dev_err(dev, "failed to set max Tx rate to %d Mbps for VF %u, error %d\n",
+				vf->max_tx_rate, vf->vf_id, err);
+			return err;
+		}
+	}
+
+	return 0;
+}
+
 /**
  * ice_vf_rebuild_host_vlan_cfg - add VLAN 0 filter or rebuild the Port VLAN
  * @vf: VF to add MAC filters for
@@ -1420,6 +1455,11 @@ static void ice_vf_rebuild_host_cfg(struct ice_vf *vf)
 	if (ice_vf_rebuild_host_vlan_cfg(vf))
 		dev_err(dev, "failed to rebuild VLAN configuration for VF %u\n",
 			vf->vf_id);
+
+	if (ice_vf_rebuild_host_tx_rate_cfg(vf))
+		dev_err(dev, "failed to rebuild Tx rate limiting configuration for VF %u\n",
+			vf->vf_id);
+
 	/* rebuild aggregator node config for main VF VSI */
 	ice_vf_rebuild_aggregator_node_cfg(vsi);
 }
@@ -1975,7 +2015,8 @@ static int ice_ena_vfs(struct ice_pf *pf, u16 num_vfs)
 
 	clear_bit(ICE_VF_DIS, pf->state);
 
-	if (ice_eswitch_configure(pf))
+	ret = ice_eswitch_configure(pf);
+	if (ret)
 		goto err_unroll_sriov;
 
 	return 0;
@@ -4747,8 +4788,8 @@ ice_get_vf_cfg(struct net_device *netdev, int vf_id, struct ifla_vf_info *ivi)
 		ivi->linkstate = IFLA_VF_LINK_STATE_ENABLE;
 	else
 		ivi->linkstate = IFLA_VF_LINK_STATE_DISABLE;
-	ivi->max_tx_rate = vf->tx_rate;
-	ivi->min_tx_rate = 0;
+	ivi->max_tx_rate = vf->max_tx_rate;
+	ivi->min_tx_rate = vf->min_tx_rate;
 	return 0;
 }
 
@@ -4799,11 +4840,6 @@ int ice_set_vf_mac(struct net_device *netdev, int vf_id, u8 *mac)
 	struct ice_vf *vf;
 	int ret;
 
-	if (ice_is_eswitch_mode_switchdev(pf)) {
-		dev_info(ice_pf_to_dev(pf), "Trusted VF is forbidden in switchdev mode\n");
-		return -EOPNOTSUPP;
-	}
-
 	if (ice_validate_vf_id(pf, vf_id))
 		return -EINVAL;
 
@@ -4863,6 +4899,11 @@ int ice_set_vf_trust(struct net_device *netdev, int vf_id, bool trusted)
 	struct ice_vf *vf;
 	int ret;
 
+	if (ice_is_eswitch_mode_switchdev(pf)) {
+		dev_info(ice_pf_to_dev(pf), "Trusted VF is forbidden in switchdev mode\n");
+		return -EOPNOTSUPP;
+	}
+
 	if (ice_validate_vf_id(pf, vf_id))
 		return -EINVAL;
 
@@ -4926,6 +4967,122 @@ int ice_set_vf_link_state(struct net_device *netdev, int vf_id, int link_state)
 	return 0;
 }
 
+/**
+ * ice_calc_all_vfs_min_tx_rate - calculate cumulative min Tx rate on all VFs
+ * @pf: PF associated with VFs
+ */
+static int ice_calc_all_vfs_min_tx_rate(struct ice_pf *pf)
+{
+	int rate = 0, i;
+
+	ice_for_each_vf(pf, i)
+		rate += pf->vf[i].min_tx_rate;
+
+	return rate;
+}
+
+/**
+ * ice_min_tx_rate_oversubscribed - check if min Tx rate causes oversubscription
+ * @vf: VF trying to configure min_tx_rate
+ * @min_tx_rate: min Tx rate in Mbps
+ *
+ * Check if the min_tx_rate being passed in will cause oversubscription of total
+ * min_tx_rate based on the current link speed and all other VFs configured
+ * min_tx_rate
+ *
+ * Return true if the passed min_tx_rate would cause oversubscription, else
+ * return false
+ */
+static bool
+ice_min_tx_rate_oversubscribed(struct ice_vf *vf, int min_tx_rate)
+{
+	int link_speed_mbps = ice_get_link_speed_mbps(ice_get_vf_vsi(vf));
+	int all_vfs_min_tx_rate = ice_calc_all_vfs_min_tx_rate(vf->pf);
+
+	/* this VF's previous rate is being overwritten */
+	all_vfs_min_tx_rate -= vf->min_tx_rate;
+
+	if (all_vfs_min_tx_rate + min_tx_rate > link_speed_mbps) {
+		dev_err(ice_pf_to_dev(vf->pf), "min_tx_rate of %d Mbps on VF %u would cause oversubscription of %d Mbps based on the current link speed %d Mbps\n",
+			min_tx_rate, vf->vf_id,
+			all_vfs_min_tx_rate + min_tx_rate - link_speed_mbps,
+			link_speed_mbps);
+		return true;
+	}
+
+	return false;
+}
+
+/**
+ * ice_set_vf_bw - set min/max VF bandwidth
+ * @netdev: network interface device structure
+ * @vf_id: VF identifier
+ * @min_tx_rate: Minimum Tx rate in Mbps
+ * @max_tx_rate: Maximum Tx rate in Mbps
+ */
+int
+ice_set_vf_bw(struct net_device *netdev, int vf_id, int min_tx_rate,
+	      int max_tx_rate)
+{
+	struct ice_pf *pf = ice_netdev_to_pf(netdev);
+	struct ice_vsi *vsi;
+	struct device *dev;
+	struct ice_vf *vf;
+	int ret;
+
+	dev = ice_pf_to_dev(pf);
+	if (ice_validate_vf_id(pf, vf_id))
+		return -EINVAL;
+
+	vf = &pf->vf[vf_id];
+	ret = ice_check_vf_ready_for_cfg(vf);
+	if (ret)
+		return ret;
+
+	vsi = ice_get_vf_vsi(vf);
+
+	/* when max_tx_rate is zero that means no max Tx rate limiting, so only
+	 * check if max_tx_rate is non-zero
+	 */
+	if (max_tx_rate && min_tx_rate > max_tx_rate) {
+		dev_err(dev, "Cannot set min Tx rate %d Mbps greater than max Tx rate %d Mbps\n",
+			min_tx_rate, max_tx_rate);
+		return -EINVAL;
+	}
+
+	if (min_tx_rate && ice_is_dcb_active(pf)) {
+		dev_err(dev, "DCB on PF is currently enabled. VF min Tx rate limiting not allowed on this PF.\n");
+		return -EOPNOTSUPP;
+	}
+
+	if (ice_min_tx_rate_oversubscribed(vf, min_tx_rate))
+		return -EINVAL;
+
+	if (vf->min_tx_rate != (unsigned int)min_tx_rate) {
+		ret = ice_set_min_bw_limit(vsi, (u64)min_tx_rate * 1000);
+		if (ret) {
+			dev_err(dev, "Unable to set min-tx-rate for VF %d\n",
+				vf->vf_id);
+			return ret;
+		}
+
+		vf->min_tx_rate = min_tx_rate;
+	}
+
+	if (vf->max_tx_rate != (unsigned int)max_tx_rate) {
+		ret = ice_set_max_bw_limit(vsi, (u64)max_tx_rate * 1000);
+		if (ret) {
+			dev_err(dev, "Unable to set max-tx-rate for VF %d\n",
+				vf->vf_id);
+			return ret;
+		}
+
+		vf->max_tx_rate = max_tx_rate;
+	}
+
+	return 0;
+}
+
 /**
  * ice_get_vf_stats - populate some stats for the VF
  * @netdev: the netdev of the PF

drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h

@@ -125,7 +125,8 @@ struct ice_vf {
 	 * the main LAN VSI for the PF.
 	 */
 	u16 lan_vsi_num;		/* ID as used by firmware */
-	unsigned int tx_rate;		/* Tx bandwidth limit in Mbps */
+	unsigned int min_tx_rate;	/* Minimum Tx bandwidth limit in Mbps */
+	unsigned int max_tx_rate;	/* Maximum Tx bandwidth limit in Mbps */
 	DECLARE_BITMAP(vf_states, ICE_VF_STATES_NBITS);	/* VF runtime states */
 
 	u64 num_inval_msgs;		/* number of continuous invalid msgs */
@@ -172,6 +173,10 @@ int
 ice_set_vf_port_vlan(struct net_device *netdev, int vf_id, u16 vlan_id, u8 qos,
 		     __be16 vlan_proto);
 
+int
+ice_set_vf_bw(struct net_device *netdev, int vf_id, int min_tx_rate,
+	      int max_tx_rate);
+
 int ice_set_vf_trust(struct net_device *netdev, int vf_id, bool trusted);
 
 int ice_set_vf_link_state(struct net_device *netdev, int vf_id, int link_state);
@@ -303,6 +308,14 @@ ice_set_vf_link_state(struct net_device __always_unused *netdev,
 	return -EOPNOTSUPP;
 }
 
+static inline int
+ice_set_vf_bw(struct net_device __always_unused *netdev,
+	      int __always_unused vf_id, int __always_unused min_tx_rate,
+	      int __always_unused max_tx_rate)
+{
+	return -EOPNOTSUPP;
+}
+
 static inline int
 ice_calc_vf_reg_idx(struct ice_vf __always_unused *vf,
 		    struct ice_q_vector __always_unused *q_vector)