Merge branch 'bnxt_en-next'

Michael Chan says:

====================
bnxt_en: Updates for net-next.

This series includes firmware interface update, some optimizations,
some new PCI IDs, new MTU checks, ethtool reset method, interrupt coalescing
code cleanup, and TC flower offload for vxlan encap/decap from Sathya
Perla.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/broadcom/bnxt/bnxt.c

@@ -108,9 +108,11 @@ enum board_idx {
 	BCM57452,
 	BCM57454,
 	BCM58802,
+	BCM58804,
 	BCM58808,
 	NETXTREME_E_VF,
 	NETXTREME_C_VF,
+	NETXTREME_S_VF,
 };
 
 /* indexed by enum above */
@@ -146,9 +148,11 @@ static const struct {
 	[BCM57452] = { "Broadcom BCM57452 NetXtreme-E 10Gb/25Gb/40Gb/50Gb Ethernet" },
 	[BCM57454] = { "Broadcom BCM57454 NetXtreme-E 10Gb/25Gb/40Gb/50Gb/100Gb Ethernet" },
 	[BCM58802] = { "Broadcom BCM58802 NetXtreme-S 10Gb/25Gb/40Gb/50Gb Ethernet" },
+	[BCM58804] = { "Broadcom BCM58804 NetXtreme-S 10Gb/25Gb/40Gb/50Gb/100Gb Ethernet" },
 	[BCM58808] = { "Broadcom BCM58808 NetXtreme-S 10Gb/25Gb/40Gb/50Gb/100Gb Ethernet" },
 	[NETXTREME_E_VF] = { "Broadcom NetXtreme-E Ethernet Virtual Function" },
 	[NETXTREME_C_VF] = { "Broadcom NetXtreme-C Ethernet Virtual Function" },
+	[NETXTREME_S_VF] = { "Broadcom NetXtreme-S Ethernet Virtual Function" },
 };
 
 static const struct pci_device_id bnxt_pci_tbl[] = {
@@ -186,6 +190,7 @@ static const struct pci_device_id bnxt_pci_tbl[] = {
 	{ PCI_VDEVICE(BROADCOM, 0x16f0), .driver_data = BCM58808 },
 	{ PCI_VDEVICE(BROADCOM, 0x16f1), .driver_data = BCM57452 },
 	{ PCI_VDEVICE(BROADCOM, 0xd802), .driver_data = BCM58802 },
+	{ PCI_VDEVICE(BROADCOM, 0xd804), .driver_data = BCM58804 },
 #ifdef CONFIG_BNXT_SRIOV
 	{ PCI_VDEVICE(BROADCOM, 0x1606), .driver_data = NETXTREME_E_VF },
 	{ PCI_VDEVICE(BROADCOM, 0x1609), .driver_data = NETXTREME_E_VF },
@@ -195,6 +200,7 @@ static const struct pci_device_id bnxt_pci_tbl[] = {
 	{ PCI_VDEVICE(BROADCOM, 0x16dc), .driver_data = NETXTREME_E_VF },
 	{ PCI_VDEVICE(BROADCOM, 0x16e1), .driver_data = NETXTREME_C_VF },
 	{ PCI_VDEVICE(BROADCOM, 0x16e5), .driver_data = NETXTREME_C_VF },
+	{ PCI_VDEVICE(BROADCOM, 0xd800), .driver_data = NETXTREME_S_VF },
 #endif
 	{ 0 }
 };
@@ -219,7 +225,8 @@ static struct workqueue_struct *bnxt_pf_wq;
 
 static bool bnxt_vf_pciid(enum board_idx idx)
 {
-	return (idx == NETXTREME_C_VF || idx == NETXTREME_E_VF);
+	return (idx == NETXTREME_C_VF || idx == NETXTREME_E_VF ||
+		idx == NETXTREME_S_VF);
 }
 
 #define DB_CP_REARM_FLAGS	(DB_KEY_CP | DB_IDX_VALID)
@@ -2828,7 +2835,8 @@ int bnxt_set_rx_skb_mode(struct bnxt *bp, bool page_mode)
 	if (page_mode) {
 		if (bp->dev->mtu > BNXT_MAX_PAGE_MODE_MTU)
 			return -EOPNOTSUPP;
-		bp->dev->max_mtu = BNXT_MAX_PAGE_MODE_MTU;
+		bp->dev->max_mtu =
+			min_t(u16, bp->max_mtu, BNXT_MAX_PAGE_MODE_MTU);
 		bp->flags &= ~BNXT_FLAG_AGG_RINGS;
 		bp->flags |= BNXT_FLAG_NO_AGG_RINGS | BNXT_FLAG_RX_PAGE_MODE;
 		bp->dev->hw_features &= ~NETIF_F_LRO;
@@ -2836,7 +2844,7 @@ int bnxt_set_rx_skb_mode(struct bnxt *bp, bool page_mode)
 		bp->rx_dir = DMA_BIDIRECTIONAL;
 		bp->rx_skb_func = bnxt_rx_page_skb;
 	} else {
-		bp->dev->max_mtu = BNXT_MAX_MTU;
+		bp->dev->max_mtu = bp->max_mtu;
 		bp->flags &= ~BNXT_FLAG_RX_PAGE_MODE;
 		bp->rx_dir = DMA_FROM_DEVICE;
 		bp->rx_skb_func = bnxt_rx_skb;
@@ -4529,19 +4537,42 @@ static int bnxt_hwrm_check_tx_rings(struct bnxt *bp, int tx_rings)
 	return 0;
 }
 
-static void bnxt_hwrm_set_coal_params(struct bnxt *bp, u32 max_bufs,
-	u32 buf_tmrs, u16 flags,
+static void bnxt_hwrm_set_coal_params(struct bnxt_coal *hw_coal,
 	struct hwrm_ring_cmpl_ring_cfg_aggint_params_input *req)
 {
+	u16 val, tmr, max, flags;
+
+	max = hw_coal->bufs_per_record * 128;
+	if (hw_coal->budget)
+		max = hw_coal->bufs_per_record * hw_coal->budget;
+
+	val = clamp_t(u16, hw_coal->coal_bufs, 1, max);
+	req->num_cmpl_aggr_int = cpu_to_le16(val);
+	req->num_cmpl_dma_aggr = cpu_to_le16(val);
+
+	val = clamp_t(u16, hw_coal->coal_bufs_irq, 1, max);
+	req->num_cmpl_dma_aggr_during_int = cpu_to_le16(val);
+
+	tmr = BNXT_USEC_TO_COAL_TIMER(hw_coal->coal_ticks);
+	tmr = max_t(u16, tmr, 1);
+	req->int_lat_tmr_max = cpu_to_le16(tmr);
+
+	/* min timer set to 1/2 of interrupt timer */
+	val = tmr / 2;
+	req->int_lat_tmr_min = cpu_to_le16(val);
+
+	/* buf timer set to 1/4 of interrupt timer */
+	val = max_t(u16, tmr / 4, 1);
+	req->cmpl_aggr_dma_tmr = cpu_to_le16(val);
+
+	tmr = BNXT_USEC_TO_COAL_TIMER(hw_coal->coal_ticks_irq);
+	tmr = max_t(u16, tmr, 1);
+	req->cmpl_aggr_dma_tmr_during_int = cpu_to_le16(tmr);
+
+	flags = RING_CMPL_RING_CFG_AGGINT_PARAMS_REQ_FLAGS_TIMER_RESET;
+	if (hw_coal->idle_thresh && hw_coal->coal_ticks < hw_coal->idle_thresh)
+		flags |= RING_CMPL_RING_CFG_AGGINT_PARAMS_REQ_FLAGS_RING_IDLE;
 	req->flags = cpu_to_le16(flags);
-	req->num_cmpl_dma_aggr = cpu_to_le16((u16)max_bufs);
-	req->num_cmpl_dma_aggr_during_int = cpu_to_le16(max_bufs >> 16);
-	req->cmpl_aggr_dma_tmr = cpu_to_le16((u16)buf_tmrs);
-	req->cmpl_aggr_dma_tmr_during_int = cpu_to_le16(buf_tmrs >> 16);
-	/* Minimum time between 2 interrupts set to buf_tmr x 2 */
-	req->int_lat_tmr_min = cpu_to_le16((u16)buf_tmrs * 2);
-	req->int_lat_tmr_max = cpu_to_le16((u16)buf_tmrs * 4);
-	req->num_cmpl_aggr_int = cpu_to_le16((u16)max_bufs * 4);
 }
 
 int bnxt_hwrm_set_coal(struct bnxt *bp)
@@ -4549,51 +4580,14 @@ int bnxt_hwrm_set_coal(struct bnxt *bp)
 	int i, rc = 0;
 	struct hwrm_ring_cmpl_ring_cfg_aggint_params_input req_rx = {0},
 							   req_tx = {0}, *req;
-	u16 max_buf, max_buf_irq;
-	u16 buf_tmr, buf_tmr_irq;
-	u32 flags;
 
 	bnxt_hwrm_cmd_hdr_init(bp, &req_rx,
 			       HWRM_RING_CMPL_RING_CFG_AGGINT_PARAMS, -1, -1);
 	bnxt_hwrm_cmd_hdr_init(bp, &req_tx,
 			       HWRM_RING_CMPL_RING_CFG_AGGINT_PARAMS, -1, -1);
 
-	/* Each rx completion (2 records) should be DMAed immediately.
-	 * DMA 1/4 of the completion buffers at a time.
-	 */
-	max_buf = min_t(u16, bp->rx_coal_bufs / 4, 2);
-	/* max_buf must not be zero */
-	max_buf = clamp_t(u16, max_buf, 1, 63);
-	max_buf_irq = clamp_t(u16, bp->rx_coal_bufs_irq, 1, 63);
-	buf_tmr = BNXT_USEC_TO_COAL_TIMER(bp->rx_coal_ticks);
-	/* buf timer set to 1/4 of interrupt timer */
-	buf_tmr = max_t(u16, buf_tmr / 4, 1);
-	buf_tmr_irq = BNXT_USEC_TO_COAL_TIMER(bp->rx_coal_ticks_irq);
-	buf_tmr_irq = max_t(u16, buf_tmr_irq, 1);
-
-	flags = RING_CMPL_RING_CFG_AGGINT_PARAMS_REQ_FLAGS_TIMER_RESET;
-
-	/* RING_IDLE generates more IRQs for lower latency.  Enable it only
-	 * if coal_ticks is less than 25 us.
-	 */
-	if (bp->rx_coal_ticks < 25)
-		flags |= RING_CMPL_RING_CFG_AGGINT_PARAMS_REQ_FLAGS_RING_IDLE;
-
-	bnxt_hwrm_set_coal_params(bp, max_buf_irq << 16 | max_buf,
-				  buf_tmr_irq << 16 | buf_tmr, flags, &req_rx);
-
-	/* max_buf must not be zero */
-	max_buf = clamp_t(u16, bp->tx_coal_bufs, 1, 63);
-	max_buf_irq = clamp_t(u16, bp->tx_coal_bufs_irq, 1, 63);
-	buf_tmr = BNXT_USEC_TO_COAL_TIMER(bp->tx_coal_ticks);
-	/* buf timer set to 1/4 of interrupt timer */
-	buf_tmr = max_t(u16, buf_tmr / 4, 1);
-	buf_tmr_irq = BNXT_USEC_TO_COAL_TIMER(bp->tx_coal_ticks_irq);
-	buf_tmr_irq = max_t(u16, buf_tmr_irq, 1);
-
-	flags = RING_CMPL_RING_CFG_AGGINT_PARAMS_REQ_FLAGS_TIMER_RESET;
-	bnxt_hwrm_set_coal_params(bp, max_buf_irq << 16 | max_buf,
-				  buf_tmr_irq << 16 | buf_tmr, flags, &req_tx);
+	bnxt_hwrm_set_coal_params(&bp->rx_coal, &req_rx);
+	bnxt_hwrm_set_coal_params(&bp->tx_coal, &req_tx);
 
 	mutex_lock(&bp->hwrm_cmd_lock);
 	for (i = 0; i < bp->cp_nr_rings; i++) {
@@ -4725,6 +4719,10 @@ static int bnxt_hwrm_func_qcfg(struct bnxt *bp)
 	else
 		bp->br_mode = BRIDGE_MODE_UNDEF;
 
+	bp->max_mtu = le16_to_cpu(resp->max_mtu_configured);
+	if (!bp->max_mtu)
+		bp->max_mtu = BNXT_MAX_MTU;
+
 func_qcfg_exit:
 	mutex_unlock(&bp->hwrm_cmd_lock);
 	return rc;
@@ -4885,9 +4883,9 @@ static int bnxt_hwrm_ver_get(struct bnxt *bp)
 			    resp->hwrm_intf_upd);
 		netdev_warn(bp->dev, "Please update firmware with HWRM interface 1.0.0 or newer.\n");
 	}
-	snprintf(bp->fw_ver_str, BC_HWRM_STR_LEN, "%d.%d.%d/%d.%d.%d",
+	snprintf(bp->fw_ver_str, BC_HWRM_STR_LEN, "%d.%d.%d.%d",
 		 resp->hwrm_fw_maj, resp->hwrm_fw_min, resp->hwrm_fw_bld,
-		 resp->hwrm_intf_maj, resp->hwrm_intf_min, resp->hwrm_intf_upd);
+		 resp->hwrm_fw_rsvd);
 
 	bp->hwrm_cmd_timeout = le16_to_cpu(resp->def_req_timeout);
 	if (!bp->hwrm_cmd_timeout)
@@ -6981,6 +6979,11 @@ static void bnxt_timer(unsigned long data)
 		set_bit(BNXT_PERIODIC_STATS_SP_EVENT, &bp->sp_event);
 		bnxt_queue_sp_work(bp);
 	}
+
+	if (bnxt_tc_flower_enabled(bp)) {
+		set_bit(BNXT_FLOW_STATS_SP_EVENT, &bp->sp_event);
+		bnxt_queue_sp_work(bp);
+	}
 bnxt_restart_timer:
 	mod_timer(&bp->timer, jiffies + bp->current_interval);
 }
@@ -7071,6 +7074,10 @@ static void bnxt_sp_task(struct work_struct *work)
 		bnxt_get_port_module_status(bp);
 		mutex_unlock(&bp->link_lock);
 	}
+
+	if (test_and_clear_bit(BNXT_FLOW_STATS_SP_EVENT, &bp->sp_event))
+		bnxt_tc_flow_stats_work(bp);
+
 	/* These functions below will clear BNXT_STATE_IN_SP_TASK.  They
 	 * must be the last functions to be called before exiting.
 	 */
@@ -7134,6 +7141,32 @@ static void bnxt_cleanup_pci(struct bnxt *bp)
 	pci_disable_device(bp->pdev);
 }
 
+static void bnxt_init_dflt_coal(struct bnxt *bp)
+{
+	struct bnxt_coal *coal;
+
+	/* Tick values in micro seconds.
+	 * 1 coal_buf x bufs_per_record = 1 completion record.
+	 */
+	coal = &bp->rx_coal;
+	coal->coal_ticks = 14;
+	coal->coal_bufs = 30;
+	coal->coal_ticks_irq = 1;
+	coal->coal_bufs_irq = 2;
+	coal->idle_thresh = 25;
+	coal->bufs_per_record = 2;
+	coal->budget = 64;		/* NAPI budget */
+
+	coal = &bp->tx_coal;
+	coal->coal_ticks = 28;
+	coal->coal_bufs = 30;
+	coal->coal_ticks_irq = 2;
+	coal->coal_bufs_irq = 2;
+	coal->bufs_per_record = 1;
+
+	bp->stats_coal_ticks = BNXT_DEF_STATS_COAL_TICKS;
+}
+
 static int bnxt_init_board(struct pci_dev *pdev, struct net_device *dev)
 {
 	int rc;
@@ -7202,18 +7235,7 @@ static int bnxt_init_board(struct pci_dev *pdev, struct net_device *dev)
 	bp->rx_ring_size = BNXT_DEFAULT_RX_RING_SIZE;
 	bp->tx_ring_size = BNXT_DEFAULT_TX_RING_SIZE;
 
-	/* tick values in micro seconds */
-	bp->rx_coal_ticks = 12;
-	bp->rx_coal_bufs = 30;
-	bp->rx_coal_ticks_irq = 1;
-	bp->rx_coal_bufs_irq = 2;
-
-	bp->tx_coal_ticks = 25;
-	bp->tx_coal_bufs = 30;
-	bp->tx_coal_ticks_irq = 2;
-	bp->tx_coal_bufs_irq = 2;
-
-	bp->stats_coal_ticks = BNXT_DEF_STATS_COAL_TICKS;
+	bnxt_init_dflt_coal(bp);
 
 	setup_timer(&bp->timer, bnxt_timer, (unsigned long)bp);
 	bp->current_interval = BNXT_TIMER_INTERVAL;
@@ -7242,13 +7264,13 @@ static int bnxt_change_mac_addr(struct net_device *dev, void *p)
 	if (!is_valid_ether_addr(addr->sa_data))
 		return -EADDRNOTAVAIL;
 
+	if (ether_addr_equal(addr->sa_data, dev->dev_addr))
+		return 0;
+
 	rc = bnxt_approve_mac(bp, addr->sa_data);
 	if (rc)
 		return rc;
 
-	if (ether_addr_equal(addr->sa_data, dev->dev_addr))
-		return 0;
-
 	memcpy(dev->dev_addr, addr->sa_data, dev->addr_len);
 	if (netif_running(dev)) {
 		bnxt_close_nic(bp, false, false);
@@ -7325,7 +7347,7 @@ static int bnxt_setup_tc_block_cb(enum tc_setup_type type, void *type_data,
 {
 	struct bnxt *bp = cb_priv;
 
-	if (BNXT_VF(bp))
+	if (!bnxt_tc_flower_enabled(bp))
 		return -EOPNOTSUPP;
 
 	switch (type) {
@@ -8088,10 +8110,6 @@ static int bnxt_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 	dev->features |= dev->hw_features | NETIF_F_HIGHDMA;
 	dev->priv_flags |= IFF_UNICAST_FLT;
 
-	/* MTU range: 60 - 9500 */
-	dev->min_mtu = ETH_ZLEN;
-	dev->max_mtu = BNXT_MAX_MTU;
-
 #ifdef CONFIG_BNXT_SRIOV
 	init_waitqueue_head(&bp->sriov_cfg_wait);
 	mutex_init(&bp->sriov_lock);
@@ -8139,6 +8157,10 @@ static int bnxt_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 	bnxt_ethtool_init(bp);
 	bnxt_dcb_init(bp);
 
+	/* MTU range: 60 - FW defined max */
+	dev->min_mtu = ETH_ZLEN;
+	dev->max_mtu = bp->max_mtu;
+
 	rc = bnxt_probe_phy(bp);
 	if (rc)
 		goto init_err_pci_clean;

drivers/net/ethernet/broadcom/bnxt/bnxt.h

@@ -944,6 +944,22 @@ struct bnxt_test_info {
 #define BNXT_CAG_REG_LEGACY_INT_STATUS	0x4014
 #define BNXT_CAG_REG_BASE		0x300000
 
+struct bnxt_coal {
+	u16			coal_ticks;
+	u16			coal_ticks_irq;
+	u16			coal_bufs;
+	u16			coal_bufs_irq;
+			/* RING_IDLE enabled when coal ticks < idle_thresh  */
+	u16			idle_thresh;
+	u8			bufs_per_record;
+	u8			budget;
+};
+
+struct bnxt_tc_flow_stats {
+	u64		packets;
+	u64		bytes;
+};
+
 struct bnxt_tc_info {
 	bool				enabled;
 
@@ -954,12 +970,29 @@ struct bnxt_tc_info {
 	/* hash table to store L2 keys of TC flows */
 	struct rhashtable		l2_table;
 	struct rhashtable_params	l2_ht_params;
+	/* hash table to store L2 keys for TC tunnel decap */
+	struct rhashtable		decap_l2_table;
+	struct rhashtable_params	decap_l2_ht_params;
+	/* hash table to store tunnel decap entries */
+	struct rhashtable		decap_table;
+	struct rhashtable_params	decap_ht_params;
+	/* hash table to store tunnel encap entries */
+	struct rhashtable		encap_table;
+	struct rhashtable_params	encap_ht_params;
 
 	/* lock to atomically add/del an l2 node when a flow is
 	 * added or deleted.
 	 */
 	struct mutex			lock;
 
+	/* Fields used for batching stats query */
+	struct rhashtable_iter		iter;
+#define BNXT_FLOW_STATS_BATCH_MAX	10
+	struct bnxt_tc_stats_batch {
+		void			  *flow_node;
+		struct bnxt_tc_flow_stats hw_stats;
+	} stats_batch[BNXT_FLOW_STATS_BATCH_MAX];
+
 	/* Stat counter mask (width) */
 	u64				bytes_mask;
 	u64				packets_mask;
@@ -1013,6 +1046,7 @@ struct bnxt {
 #define CHIP_NUM_5745X		0xd730
 
 #define CHIP_NUM_58802		0xd802
+#define CHIP_NUM_58804		0xd804
 #define CHIP_NUM_58808		0xd808
 
 #define BNXT_CHIP_NUM_5730X(chip_num)		\
@@ -1048,6 +1082,7 @@ struct bnxt {
 
 #define BNXT_CHIP_NUM_588XX(chip_num)		\
 	((chip_num) == CHIP_NUM_58802 ||	\
+	 (chip_num) == CHIP_NUM_58804 ||        \
 	 (chip_num) == CHIP_NUM_58808)
 
 	struct net_device	*dev;
@@ -1170,6 +1205,7 @@ struct bnxt {
 	int			nr_vnics;
 	u32			rss_hash_cfg;
 
+	u16			max_mtu;
 	u8			max_tc;
 	u8			max_lltc;	/* lossless TCs */
 	struct bnxt_queue_info	q_info[BNXT_MAX_QUEUE];
@@ -1232,14 +1268,8 @@ struct bnxt {
 	u8			port_count;
 	u16			br_mode;
 
-	u16			rx_coal_ticks;
-	u16			rx_coal_ticks_irq;
-	u16			rx_coal_bufs;
-	u16			rx_coal_bufs_irq;
-	u16			tx_coal_ticks;
-	u16			tx_coal_ticks_irq;
-	u16			tx_coal_bufs;
-	u16			tx_coal_bufs_irq;
+	struct bnxt_coal	rx_coal;
+	struct bnxt_coal	tx_coal;
 
 #define BNXT_USEC_TO_COAL_TIMER(x)	((x) * 25 / 2)
 
@@ -1265,6 +1295,7 @@ struct bnxt {
 #define BNXT_GENEVE_ADD_PORT_SP_EVENT	12
 #define BNXT_GENEVE_DEL_PORT_SP_EVENT	13
 #define BNXT_LINK_SPEED_CHNG_SP_EVENT	14
+#define BNXT_FLOW_STATS_SP_EVENT	15
 
 	struct bnxt_pf_info	pf;
 #ifdef CONFIG_BNXT_SRIOV
@@ -1315,7 +1346,7 @@ struct bnxt {
 	enum devlink_eswitch_mode eswitch_mode;
 	struct bnxt_vf_rep	**vf_reps; /* array of vf-rep ptrs */
 	u16			*cfa_code_map; /* cfa_code -> vf_idx map */
-	struct bnxt_tc_info	tc_info;
+	struct bnxt_tc_info	*tc_info;
 };
 
 #define BNXT_RX_STATS_OFFSET(counter)			\

drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.c

@@ -29,7 +29,7 @@ int bnxt_dl_register(struct bnxt *bp)
 	if (!pci_find_ext_capability(bp->pdev, PCI_EXT_CAP_ID_SRIOV))
 		return 0;
 
-	if (bp->hwrm_spec_code < 0x10800) {
+	if (bp->hwrm_spec_code < 0x10803) {
 		netdev_warn(bp->dev, "Firmware does not support SR-IOV E-Switch SWITCHDEV mode.\n");
 		return -ENOTSUPP;
 	}

drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c

@@ -26,8 +26,6 @@
 #define FLASH_PACKAGE_TIMEOUT	((HWRM_CMD_TIMEOUT) * 200)
 #define INSTALL_PACKAGE_TIMEOUT	((HWRM_CMD_TIMEOUT) * 200)
 
-static char *bnxt_get_pkgver(struct net_device *dev, char *buf, size_t buflen);
-
 static u32 bnxt_get_msglevel(struct net_device *dev)
 {
 	struct bnxt *bp = netdev_priv(dev);
@@ -46,19 +44,24 @@ static int bnxt_get_coalesce(struct net_device *dev,
 			     struct ethtool_coalesce *coal)
 {
 	struct bnxt *bp = netdev_priv(dev);
+	struct bnxt_coal *hw_coal;
+	u16 mult;
 
 	memset(coal, 0, sizeof(*coal));
 
-	coal->rx_coalesce_usecs = bp->rx_coal_ticks;
-	/* 2 completion records per rx packet */
-	coal->rx_max_coalesced_frames = bp->rx_coal_bufs / 2;
-	coal->rx_coalesce_usecs_irq = bp->rx_coal_ticks_irq;
-	coal->rx_max_coalesced_frames_irq = bp->rx_coal_bufs_irq / 2;
+	hw_coal = &bp->rx_coal;
+	mult = hw_coal->bufs_per_record;
+	coal->rx_coalesce_usecs = hw_coal->coal_ticks;
+	coal->rx_max_coalesced_frames = hw_coal->coal_bufs / mult;
+	coal->rx_coalesce_usecs_irq = hw_coal->coal_ticks_irq;
+	coal->rx_max_coalesced_frames_irq = hw_coal->coal_bufs_irq / mult;
 
-	coal->tx_coalesce_usecs = bp->tx_coal_ticks;
-	coal->tx_max_coalesced_frames = bp->tx_coal_bufs;
-	coal->tx_coalesce_usecs_irq = bp->tx_coal_ticks_irq;
-	coal->tx_max_coalesced_frames_irq = bp->tx_coal_bufs_irq;
+	hw_coal = &bp->tx_coal;
+	mult = hw_coal->bufs_per_record;
+	coal->tx_coalesce_usecs = hw_coal->coal_ticks;
+	coal->tx_max_coalesced_frames = hw_coal->coal_bufs / mult;
+	coal->tx_coalesce_usecs_irq = hw_coal->coal_ticks_irq;
+	coal->tx_max_coalesced_frames_irq = hw_coal->coal_bufs_irq / mult;
 
 	coal->stats_block_coalesce_usecs = bp->stats_coal_ticks;
 
@@ -70,18 +73,23 @@ static int bnxt_set_coalesce(struct net_device *dev,
 {
 	struct bnxt *bp = netdev_priv(dev);
 	bool update_stats = false;
+	struct bnxt_coal *hw_coal;
 	int rc = 0;
-
-	bp->rx_coal_ticks = coal->rx_coalesce_usecs;
-	/* 2 completion records per rx packet */
-	bp->rx_coal_bufs = coal->rx_max_coalesced_frames * 2;
-	bp->rx_coal_ticks_irq = coal->rx_coalesce_usecs_irq;
-	bp->rx_coal_bufs_irq = coal->rx_max_coalesced_frames_irq * 2;
-
-	bp->tx_coal_ticks = coal->tx_coalesce_usecs;
-	bp->tx_coal_bufs = coal->tx_max_coalesced_frames;
-	bp->tx_coal_ticks_irq = coal->tx_coalesce_usecs_irq;
-	bp->tx_coal_bufs_irq = coal->tx_max_coalesced_frames_irq;
+	u16 mult;
+
+	hw_coal = &bp->rx_coal;
+	mult = hw_coal->bufs_per_record;
+	hw_coal->coal_ticks = coal->rx_coalesce_usecs;
+	hw_coal->coal_bufs = coal->rx_max_coalesced_frames * mult;
+	hw_coal->coal_ticks_irq = coal->rx_coalesce_usecs_irq;
+	hw_coal->coal_bufs_irq = coal->rx_max_coalesced_frames_irq * mult;
+
+	hw_coal = &bp->rx_coal;
+	mult = hw_coal->bufs_per_record;
+	hw_coal->coal_ticks = coal->tx_coalesce_usecs;
+	hw_coal->coal_bufs = coal->tx_max_coalesced_frames * mult;
+	hw_coal->coal_ticks_irq = coal->tx_coalesce_usecs_irq;
+	hw_coal->coal_bufs_irq = coal->tx_max_coalesced_frames_irq * mult;
 
 	if (bp->stats_coal_ticks != coal->stats_block_coalesce_usecs) {
 		u32 stats_ticks = coal->stats_block_coalesce_usecs;
@@ -822,20 +830,10 @@ static void bnxt_get_drvinfo(struct net_device *dev,
 			     struct ethtool_drvinfo *info)
 {
 	struct bnxt *bp = netdev_priv(dev);
-	char *pkglog;
-	char *pkgver = NULL;
 
-	pkglog = kmalloc(BNX_PKG_LOG_MAX_LENGTH, GFP_KERNEL);
-	if (pkglog)
-		pkgver = bnxt_get_pkgver(dev, pkglog, BNX_PKG_LOG_MAX_LENGTH);
 	strlcpy(info->driver, DRV_MODULE_NAME, sizeof(info->driver));
 	strlcpy(info->version, DRV_MODULE_VERSION, sizeof(info->version));
-	if (pkgver && *pkgver != 0 && isdigit(*pkgver))
-		snprintf(info->fw_version, sizeof(info->fw_version) - 1,
-			 "%s pkg %s", bp->fw_ver_str, pkgver);
-	else
-		strlcpy(info->fw_version, bp->fw_ver_str,
-			sizeof(info->fw_version));
+	strlcpy(info->fw_version, bp->fw_ver_str, sizeof(info->fw_version));
 	strlcpy(info->bus_info, pci_name(bp->pdev), sizeof(info->bus_info));
 	info->n_stats = bnxt_get_num_stats(bp);
 	info->testinfo_len = bp->num_tests;
@@ -843,7 +841,6 @@ static void bnxt_get_drvinfo(struct net_device *dev,
 	info->eedump_len = 0;
 	/* TODO CHIMP FW: reg dump details */
 	info->regdump_len = 0;
-	kfree(pkglog);
 }
 
 static void bnxt_get_wol(struct net_device *dev, struct ethtool_wolinfo *wol)
@@ -1350,7 +1347,6 @@ static int bnxt_firmware_reset(struct net_device *dev,
 
 	bnxt_hwrm_cmd_hdr_init(bp, &req, HWRM_FW_RESET, -1, -1);
 
-	/* TODO: Support ASAP ChiMP self-reset (e.g. upon PF driver unload) */
 	/* TODO: Address self-reset of APE/KONG/BONO/TANG or ungraceful reset */
 	/*       (e.g. when firmware isn't already running) */
 	switch (dir_type) {
@@ -1376,6 +1372,10 @@ static int bnxt_firmware_reset(struct net_device *dev,
 	case BNX_DIR_TYPE_BONO_PATCH:
 		req.embedded_proc_type = FW_RESET_REQ_EMBEDDED_PROC_TYPE_ROCE;
 		break;
+	case BNXT_FW_RESET_CHIP:
+		req.embedded_proc_type = FW_RESET_REQ_EMBEDDED_PROC_TYPE_CHIP;
+		req.selfrst_status = FW_RESET_REQ_SELFRST_STATUS_SELFRSTASAP;
+		break;
 	default:
 		return -EINVAL;
 	}
@@ -1773,6 +1773,9 @@ static int bnxt_get_nvram_item(struct net_device *dev, u32 index, u32 offset,
 	dma_addr_t dma_handle;
 	struct hwrm_nvm_read_input req = {0};
 
+	if (!length)
+		return -EINVAL;
+
 	buf = dma_alloc_coherent(&bp->pdev->dev, length, &dma_handle,
 				 GFP_KERNEL);
 	if (!buf) {
@@ -2495,13 +2498,59 @@ static void bnxt_self_test(struct net_device *dev, struct ethtool_test *etest,
 	}
 }
 
+static int bnxt_reset(struct net_device *dev, u32 *flags)
+{
+	struct bnxt *bp = netdev_priv(dev);
+	int rc = 0;
+
+	if (!BNXT_PF(bp)) {
+		netdev_err(dev, "Reset is not supported from a VF\n");
+		return -EOPNOTSUPP;
+	}
+
+	if (pci_vfs_assigned(bp->pdev)) {
+		netdev_err(dev,
+			   "Reset not allowed when VFs are assigned to VMs\n");
+		return -EBUSY;
+	}
+
+	if (*flags == ETH_RESET_ALL) {
+		/* This feature is not supported in older firmware versions */
+		if (bp->hwrm_spec_code < 0x10803)
+			return -EOPNOTSUPP;
+
+		rc = bnxt_firmware_reset(dev, BNXT_FW_RESET_CHIP);
+		if (!rc)
+			netdev_info(dev, "Reset request successful. Reload driver to complete reset\n");
+	} else {
+		rc = -EINVAL;
+	}
+
+	return rc;
+}
+
 void bnxt_ethtool_init(struct bnxt *bp)
 {
 	struct hwrm_selftest_qlist_output *resp = bp->hwrm_cmd_resp_addr;
 	struct hwrm_selftest_qlist_input req = {0};
 	struct bnxt_test_info *test_info;
+	struct net_device *dev = bp->dev;
+	char *pkglog;
 	int i, rc;
 
+	pkglog = kzalloc(BNX_PKG_LOG_MAX_LENGTH, GFP_KERNEL);
+	if (pkglog) {
+		char *pkgver;
+		int len;
+
+		pkgver = bnxt_get_pkgver(dev, pkglog, BNX_PKG_LOG_MAX_LENGTH);
+		if (pkgver && *pkgver != 0 && isdigit(*pkgver)) {
+			len = strlen(bp->fw_ver_str);
+			snprintf(bp->fw_ver_str + len, FW_VER_STR_LEN - len - 1,
+				 "/pkg %s", pkgver);
+		}
+		kfree(pkglog);
+	}
 	if (bp->hwrm_spec_code < 0x10704 || !BNXT_SINGLE_PF(bp))
 		return;
 
@@ -2592,4 +2641,5 @@ const struct ethtool_ops bnxt_ethtool_ops = {
 	.nway_reset		= bnxt_nway_reset,
 	.set_phys_id		= bnxt_set_phys_id,
 	.self_test		= bnxt_self_test,
+	.reset			= bnxt_reset,
 };

drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.h

@@ -34,6 +34,8 @@ struct bnxt_led_cfg {
 #define BNXT_LED_DFLT_ENABLES(x)			\
 	cpu_to_le32(BNXT_LED_DFLT_ENA << (BNXT_LED_DFLT_ENA_SHIFT * (x)))
 
+#define BNXT_FW_RESET_CHIP	0xffff
+
 extern const struct ethtool_ops bnxt_ethtool_ops;
 
 u32 _bnxt_fw_to_ethtool_adv_spds(u16, u8);

drivers/net/ethernet/broadcom/bnxt/bnxt_hsi.h

@@ -11,21 +11,21 @@
 #ifndef BNXT_HSI_H
 #define BNXT_HSI_H
 
-/* HSI and HWRM Specification 1.8.1 */
+/* HSI and HWRM Specification 1.8.3 */
 #define HWRM_VERSION_MAJOR	1
 #define HWRM_VERSION_MINOR	8
-#define HWRM_VERSION_UPDATE	1
+#define HWRM_VERSION_UPDATE	3
 
-#define HWRM_VERSION_RSVD	4 /* non-zero means beta version */
+#define HWRM_VERSION_RSVD	1 /* non-zero means beta version */
 
-#define HWRM_VERSION_STR	"1.8.1.4"
+#define HWRM_VERSION_STR	"1.8.3.1"
 /*
  * Following is the signature for HWRM message field that indicates not
  * applicable (All F's). Need to cast it the size of the field if needed.
  */
 #define HWRM_NA_SIGNATURE	((__le32)(-1))
 #define HWRM_MAX_REQ_LEN    (128)  /* hwrm_func_buf_rgtr */
-#define HWRM_MAX_RESP_LEN    (248)  /* hwrm_selftest_qlist */
+#define HWRM_MAX_RESP_LEN    (280)  /* hwrm_selftest_qlist */
 #define HW_HASH_INDEX_SIZE      0x80    /* 7 bit indirection table index. */
 #define HW_HASH_KEY_SIZE	40
 #define HWRM_RESP_VALID_KEY      1 /* valid key for HWRM response */
@@ -111,6 +111,7 @@ struct hwrm_async_event_cmpl {
 	#define ASYNC_EVENT_CMPL_EVENT_ID_VF_MAC_ADDR_CHANGE      0x31UL
 	#define ASYNC_EVENT_CMPL_EVENT_ID_PF_VF_COMM_STATUS_CHANGE 0x32UL
 	#define ASYNC_EVENT_CMPL_EVENT_ID_VF_CFG_CHANGE	   0x33UL
+	#define ASYNC_EVENT_CMPL_EVENT_ID_LLFC_PFC_CHANGE	   0x34UL
 	#define ASYNC_EVENT_CMPL_EVENT_ID_HWRM_ERROR		   0xffUL
 	__le32 event_data2;
 	u8 opaque_v;
@@ -835,8 +836,7 @@ struct hwrm_func_qcfg_output {
 	u8 port_pf_cnt;
 	#define FUNC_QCFG_RESP_PORT_PF_CNT_UNAVAIL		   0x0UL
 	__le16 dflt_vnic_id;
-	u8 unused_0;
-	u8 unused_1;
+	__le16 max_mtu_configured;
 	__le32 min_bw;
 	#define FUNC_QCFG_RESP_MIN_BW_BW_VALUE_MASK		    0xfffffffUL
 	#define FUNC_QCFG_RESP_MIN_BW_BW_VALUE_SFT		    0
@@ -873,12 +873,12 @@ struct hwrm_func_qcfg_output {
 	#define FUNC_QCFG_RESP_EVB_MODE_NO_EVB			   0x0UL
 	#define FUNC_QCFG_RESP_EVB_MODE_VEB			   0x1UL
 	#define FUNC_QCFG_RESP_EVB_MODE_VEPA			   0x2UL
-	u8 unused_2;
+	u8 unused_0;
 	__le16 alloc_vfs;
 	__le32 alloc_mcast_filters;
 	__le32 alloc_hw_ring_grps;
 	__le16 alloc_sp_tx_rings;
-	u8 unused_3;
+	u8 unused_1;
 	u8 valid;
 };
 
@@ -3407,6 +3407,7 @@ struct hwrm_vnic_cfg_input {
 	#define VNIC_CFG_REQ_FLAGS_ROCE_DUAL_VNIC_MODE		    0x8UL
 	#define VNIC_CFG_REQ_FLAGS_ROCE_ONLY_VNIC_MODE		    0x10UL
 	#define VNIC_CFG_REQ_FLAGS_RSS_DFLT_CR_MODE		    0x20UL
+	#define VNIC_CFG_REQ_FLAGS_ROCE_MIRRORING_CAPABLE_VNIC_MODE 0x40UL
 	__le32 enables;
 	#define VNIC_CFG_REQ_ENABLES_DFLT_RING_GRP		    0x1UL
 	#define VNIC_CFG_REQ_ENABLES_RSS_RULE			    0x2UL
@@ -3463,6 +3464,7 @@ struct hwrm_vnic_qcaps_output {
 	#define VNIC_QCAPS_RESP_FLAGS_ROCE_DUAL_VNIC_CAP	    0x8UL
 	#define VNIC_QCAPS_RESP_FLAGS_ROCE_ONLY_VNIC_CAP	    0x10UL
 	#define VNIC_QCAPS_RESP_FLAGS_RSS_DFLT_CR_CAP		    0x20UL
+	#define VNIC_QCAPS_RESP_FLAGS_ROCE_MIRROING_CAPABLE_VNIC_CAP 0x40UL
 	__le32 unused_2;
 	u8 unused_3;
 	u8 unused_4;
@@ -3994,6 +3996,7 @@ struct hwrm_cfa_l2_filter_alloc_input {
 	#define CFA_L2_FILTER_ALLOC_REQ_TUNNEL_TYPE_MPLS	   0x6UL
 	#define CFA_L2_FILTER_ALLOC_REQ_TUNNEL_TYPE_STT	   0x7UL
 	#define CFA_L2_FILTER_ALLOC_REQ_TUNNEL_TYPE_IPGRE	   0x8UL
+	#define CFA_L2_FILTER_ALLOC_REQ_TUNNEL_TYPE_VXLAN_V4      0x9UL
 	#define CFA_L2_FILTER_ALLOC_REQ_TUNNEL_TYPE_ANYTUNNEL     0xffUL
 	u8 unused_7;
 	__le16 dst_id;
@@ -4122,6 +4125,14 @@ struct hwrm_cfa_l2_set_rx_mask_output {
 	u8 valid;
 };
 
+/* Command specific Error Codes (8 bytes) */
+struct hwrm_cfa_l2_set_rx_mask_cmd_err {
+	u8 code;
+	#define CFA_L2_SET_RX_MASK_CMD_ERR_CODE_UNKNOWN	   0x0UL
+	#define CFA_L2_SET_RX_MASK_CMD_ERR_CODE_NTUPLE_FILTER_CONFLICT_ERR 0x1UL
+	u8 unused_0[7];
+};
+
 /* hwrm_cfa_tunnel_filter_alloc */
 /* Input (88 bytes) */
 struct hwrm_cfa_tunnel_filter_alloc_input {
@@ -4161,6 +4172,7 @@ struct hwrm_cfa_tunnel_filter_alloc_input {
 	#define CFA_TUNNEL_FILTER_ALLOC_REQ_TUNNEL_TYPE_MPLS      0x6UL
 	#define CFA_TUNNEL_FILTER_ALLOC_REQ_TUNNEL_TYPE_STT       0x7UL
 	#define CFA_TUNNEL_FILTER_ALLOC_REQ_TUNNEL_TYPE_IPGRE     0x8UL
+	#define CFA_TUNNEL_FILTER_ALLOC_REQ_TUNNEL_TYPE_VXLAN_V4  0x9UL
 	#define CFA_TUNNEL_FILTER_ALLOC_REQ_TUNNEL_TYPE_ANYTUNNEL 0xffUL
 	u8 unused_0;
 	__le32 vni;
@@ -4323,6 +4335,7 @@ struct hwrm_cfa_ntuple_filter_alloc_input {
 	#define CFA_NTUPLE_FILTER_ALLOC_REQ_TUNNEL_TYPE_MPLS      0x6UL
 	#define CFA_NTUPLE_FILTER_ALLOC_REQ_TUNNEL_TYPE_STT       0x7UL
 	#define CFA_NTUPLE_FILTER_ALLOC_REQ_TUNNEL_TYPE_IPGRE     0x8UL
+	#define CFA_NTUPLE_FILTER_ALLOC_REQ_TUNNEL_TYPE_VXLAN_V4  0x9UL
 	#define CFA_NTUPLE_FILTER_ALLOC_REQ_TUNNEL_TYPE_ANYTUNNEL 0xffUL
 	u8 pri_hint;
 	#define CFA_NTUPLE_FILTER_ALLOC_REQ_PRI_HINT_NO_PREFER    0x0UL
@@ -4355,6 +4368,14 @@ struct hwrm_cfa_ntuple_filter_alloc_output {
 	u8 valid;
 };
 
+/* Command specific Error Codes (8 bytes) */
+struct hwrm_cfa_ntuple_filter_alloc_cmd_err {
+	u8 code;
+	#define CFA_NTUPLE_FILTER_ALLOC_CMD_ERR_CODE_UNKNOWN      0x0UL
+	#define CFA_NTUPLE_FILTER_ALLOC_CMD_ERR_CODE_RX_MASK_VLAN_CONFLICT_ERR 0x1UL
+	u8 unused_0[7];
+};
+
 /* hwrm_cfa_ntuple_filter_free */
 /* Input (24 bytes) */
 struct hwrm_cfa_ntuple_filter_free_input {
@@ -4413,6 +4434,116 @@ struct hwrm_cfa_ntuple_filter_cfg_output {
 	u8 valid;
 };
 
+/* hwrm_cfa_decap_filter_alloc */
+/* Input (104 bytes) */
+struct hwrm_cfa_decap_filter_alloc_input {
+	__le16 req_type;
+	__le16 cmpl_ring;
+	__le16 seq_id;
+	__le16 target_id;
+	__le64 resp_addr;
+	__le32 flags;
+	#define CFA_DECAP_FILTER_ALLOC_REQ_FLAGS_OVS_TUNNEL	    0x1UL
+	__le32 enables;
+	#define CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_TUNNEL_TYPE     0x1UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_TUNNEL_ID       0x2UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_SRC_MACADDR     0x4UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_DST_MACADDR     0x8UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_OVLAN_VID       0x10UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_IVLAN_VID       0x20UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_T_OVLAN_VID     0x40UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_T_IVLAN_VID     0x80UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_ETHERTYPE       0x100UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_SRC_IPADDR      0x200UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_DST_IPADDR      0x400UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_IPADDR_TYPE     0x800UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_IP_PROTOCOL     0x1000UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_SRC_PORT	    0x2000UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_DST_PORT	    0x4000UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_DST_ID	    0x8000UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_MIRROR_VNIC_ID  0x10000UL
+	__be32 tunnel_id;
+	u8 tunnel_type;
+	#define CFA_DECAP_FILTER_ALLOC_REQ_TUNNEL_TYPE_NONTUNNEL  0x0UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_TUNNEL_TYPE_VXLAN      0x1UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_TUNNEL_TYPE_NVGRE      0x2UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_TUNNEL_TYPE_L2GRE      0x3UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_TUNNEL_TYPE_IPIP       0x4UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_TUNNEL_TYPE_GENEVE     0x5UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_TUNNEL_TYPE_MPLS       0x6UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_TUNNEL_TYPE_STT	   0x7UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_TUNNEL_TYPE_IPGRE      0x8UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_TUNNEL_TYPE_VXLAN_V4   0x9UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_TUNNEL_TYPE_ANYTUNNEL  0xffUL
+	u8 unused_0;
+	__le16 unused_1;
+	u8 src_macaddr[6];
+	u8 unused_2;
+	u8 unused_3;
+	u8 dst_macaddr[6];
+	__be16 ovlan_vid;
+	__be16 ivlan_vid;
+	__be16 t_ovlan_vid;
+	__be16 t_ivlan_vid;
+	__be16 ethertype;
+	u8 ip_addr_type;
+	#define CFA_DECAP_FILTER_ALLOC_REQ_IP_ADDR_TYPE_UNKNOWN   0x0UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_IP_ADDR_TYPE_IPV4      0x4UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_IP_ADDR_TYPE_IPV6      0x6UL
+	u8 ip_protocol;
+	#define CFA_DECAP_FILTER_ALLOC_REQ_IP_PROTOCOL_UNKNOWN    0x0UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_IP_PROTOCOL_TCP	   0x6UL
+	#define CFA_DECAP_FILTER_ALLOC_REQ_IP_PROTOCOL_UDP	   0x11UL
+	u8 unused_4;
+	u8 unused_5;
+	u8 unused_6[3];
+	u8 unused_7;
+	__be32 src_ipaddr[4];
+	__be32 dst_ipaddr[4];
+	__be16 src_port;
+	__be16 dst_port;
+	__le16 dst_id;
+	__le16 l2_ctxt_ref_id;
+};
+
+/* Output (16 bytes) */
+struct hwrm_cfa_decap_filter_alloc_output {
+	__le16 error_code;
+	__le16 req_type;
+	__le16 seq_id;
+	__le16 resp_len;
+	__le32 decap_filter_id;
+	u8 unused_0;
+	u8 unused_1;
+	u8 unused_2;
+	u8 valid;
+};
+
+/* hwrm_cfa_decap_filter_free */
+/* Input (24 bytes) */
+struct hwrm_cfa_decap_filter_free_input {
+	__le16 req_type;
+	__le16 cmpl_ring;
+	__le16 seq_id;
+	__le16 target_id;
+	__le64 resp_addr;
+	__le32 decap_filter_id;
+	__le32 unused_0;
+};
+
+/* Output (16 bytes) */
+struct hwrm_cfa_decap_filter_free_output {
+	__le16 error_code;
+	__le16 req_type;
+	__le16 seq_id;
+	__le16 resp_len;
+	__le32 unused_0;
+	u8 unused_1;
+	u8 unused_2;
+	u8 unused_3;
+	u8 valid;
+};
+
 /* hwrm_cfa_flow_alloc */
 /* Input (128 bytes) */
 struct hwrm_cfa_flow_alloc_input {
@@ -4634,6 +4765,7 @@ struct hwrm_tunnel_dst_port_query_input {
 	u8 tunnel_type;
 	#define TUNNEL_DST_PORT_QUERY_REQ_TUNNEL_TYPE_VXLAN       0x1UL
 	#define TUNNEL_DST_PORT_QUERY_REQ_TUNNEL_TYPE_GENEVE      0x5UL
+	#define TUNNEL_DST_PORT_QUERY_REQ_TUNNEL_TYPE_VXLAN_V4    0x9UL
 	u8 unused_0[7];
 };
 
@@ -4662,9 +4794,10 @@ struct hwrm_tunnel_dst_port_alloc_input {
 	u8 tunnel_type;
 	#define TUNNEL_DST_PORT_ALLOC_REQ_TUNNEL_TYPE_VXLAN       0x1UL
 	#define TUNNEL_DST_PORT_ALLOC_REQ_TUNNEL_TYPE_GENEVE      0x5UL
+	#define TUNNEL_DST_PORT_ALLOC_REQ_TUNNEL_TYPE_VXLAN_V4    0x9UL
 	u8 unused_0;
 	__be16 tunnel_dst_port_val;
-	__le32 unused_1;
+	__be32 unused_1;
 };
 
 /* Output (16 bytes) */
@@ -4693,6 +4826,7 @@ struct hwrm_tunnel_dst_port_free_input {
 	u8 tunnel_type;
 	#define TUNNEL_DST_PORT_FREE_REQ_TUNNEL_TYPE_VXLAN	   0x1UL
 	#define TUNNEL_DST_PORT_FREE_REQ_TUNNEL_TYPE_GENEVE       0x5UL
+	#define TUNNEL_DST_PORT_FREE_REQ_TUNNEL_TYPE_VXLAN_V4     0x9UL
 	u8 unused_0;
 	__le16 tunnel_dst_port_id;
 	__le32 unused_1;
@@ -4848,6 +4982,8 @@ struct hwrm_fw_reset_input {
 	#define FW_RESET_REQ_EMBEDDED_PROC_TYPE_NETCTRL	   0x2UL
 	#define FW_RESET_REQ_EMBEDDED_PROC_TYPE_ROCE		   0x3UL
 	#define FW_RESET_REQ_EMBEDDED_PROC_TYPE_HOST		   0x4UL
+	#define FW_RESET_REQ_EMBEDDED_PROC_TYPE_AP		   0x5UL
+	#define FW_RESET_REQ_EMBEDDED_PROC_TYPE_CHIP		   0x6UL
 	u8 selfrst_status;
 	#define FW_RESET_REQ_SELFRST_STATUS_SELFRSTNONE	   0x0UL
 	#define FW_RESET_REQ_SELFRST_STATUS_SELFRSTASAP	   0x1UL
@@ -4888,6 +5024,8 @@ struct hwrm_fw_qstatus_input {
 	#define FW_QSTATUS_REQ_EMBEDDED_PROC_TYPE_NETCTRL	   0x2UL
 	#define FW_QSTATUS_REQ_EMBEDDED_PROC_TYPE_ROCE		   0x3UL
 	#define FW_QSTATUS_REQ_EMBEDDED_PROC_TYPE_HOST		   0x4UL
+	#define FW_QSTATUS_REQ_EMBEDDED_PROC_TYPE_AP		   0x5UL
+	#define FW_QSTATUS_REQ_EMBEDDED_PROC_TYPE_CHIP		   0x6UL
 	u8 unused_0[7];
 };
 
@@ -5324,6 +5462,32 @@ struct hwrm_wol_reason_qcfg_output {
 	u8 valid;
 };
 
+/* hwrm_dbg_read_direct */
+/* Input (32 bytes) */
+struct hwrm_dbg_read_direct_input {
+	__le16 req_type;
+	__le16 cmpl_ring;
+	__le16 seq_id;
+	__le16 target_id;
+	__le64 resp_addr;
+	__le64 host_dest_addr;
+	__le32 read_addr;
+	__le32 read_len32;
+};
+
+/* Output (16 bytes) */
+struct hwrm_dbg_read_direct_output {
+	__le16 error_code;
+	__le16 req_type;
+	__le16 seq_id;
+	__le16 resp_len;
+	__le32 unused_0;
+	u8 unused_1;
+	u8 unused_2;
+	u8 unused_3;
+	u8 valid;
+};
+
 /* hwrm_nvm_read */
 /* Input (40 bytes) */
 struct hwrm_nvm_read_input {
@@ -5676,6 +5840,105 @@ struct hwrm_nvm_install_update_cmd_err {
 	u8 unused_0[7];
 };
 
+/* hwrm_nvm_get_variable */
+/* Input (40 bytes) */
+struct hwrm_nvm_get_variable_input {
+	__le16 req_type;
+	__le16 cmpl_ring;
+	__le16 seq_id;
+	__le16 target_id;
+	__le64 resp_addr;
+	__le64 dest_data_addr;
+	__le16 data_len;
+	__le16 option_num;
+	#define NVM_GET_VARIABLE_REQ_OPTION_NUM_RSVD_0		   0x0UL
+	#define NVM_GET_VARIABLE_REQ_OPTION_NUM_RSVD_FFFF	   0xffffUL
+	__le16 dimensions;
+	__le16 index_0;
+	__le16 index_1;
+	__le16 index_2;
+	__le16 index_3;
+	u8 flags;
+	#define NVM_GET_VARIABLE_REQ_FLAGS_FACTORY_DFLT	    0x1UL
+	u8 unused_0;
+};
+
+/* Output (16 bytes) */
+struct hwrm_nvm_get_variable_output {
+	__le16 error_code;
+	__le16 req_type;
+	__le16 seq_id;
+	__le16 resp_len;
+	__le16 data_len;
+	__le16 option_num;
+	#define NVM_GET_VARIABLE_RESP_OPTION_NUM_RSVD_0	   0x0UL
+	#define NVM_GET_VARIABLE_RESP_OPTION_NUM_RSVD_FFFF	   0xffffUL
+	u8 unused_0;
+	u8 unused_1;
+	u8 unused_2;
+	u8 valid;
+};
+
+/* Command specific Error Codes (8 bytes) */
+struct hwrm_nvm_get_variable_cmd_err {
+	u8 code;
+	#define NVM_GET_VARIABLE_CMD_ERR_CODE_UNKNOWN		   0x0UL
+	#define NVM_GET_VARIABLE_CMD_ERR_CODE_VAR_NOT_EXIST       0x1UL
+	#define NVM_GET_VARIABLE_CMD_ERR_CODE_CORRUPT_VAR	   0x2UL
+	#define NVM_GET_VARIABLE_CMD_ERR_CODE_LEN_TOO_SHORT       0x3UL
+	u8 unused_0[7];
+};
+
+/* hwrm_nvm_set_variable */
+/* Input (40 bytes) */
+struct hwrm_nvm_set_variable_input {
+	__le16 req_type;
+	__le16 cmpl_ring;
+	__le16 seq_id;
+	__le16 target_id;
+	__le64 resp_addr;
+	__le64 src_data_addr;
+	__le16 data_len;
+	__le16 option_num;
+	#define NVM_SET_VARIABLE_REQ_OPTION_NUM_RSVD_0		   0x0UL
+	#define NVM_SET_VARIABLE_REQ_OPTION_NUM_RSVD_FFFF	   0xffffUL
+	__le16 dimensions;
+	__le16 index_0;
+	__le16 index_1;
+	__le16 index_2;
+	__le16 index_3;
+	u8 flags;
+	#define NVM_SET_VARIABLE_REQ_FLAGS_FORCE_FLUSH		    0x1UL
+	#define NVM_SET_VARIABLE_REQ_FLAGS_ENCRYPT_MODE_MASK       0xeUL
+	#define NVM_SET_VARIABLE_REQ_FLAGS_ENCRYPT_MODE_SFT	    1
+	#define NVM_SET_VARIABLE_REQ_FLAGS_ENCRYPT_MODE_NONE      (0x0UL << 1)
+	#define NVM_SET_VARIABLE_REQ_FLAGS_ENCRYPT_MODE_HMAC_SHA1 (0x1UL << 1)
+	#define NVM_SET_VARIABLE_REQ_FLAGS_ENCRYPT_MODE_LAST    NVM_SET_VARIABLE_REQ_FLAGS_ENCRYPT_MODE_HMAC_SHA1
+	u8 unused_0;
+};
+
+/* Output (16 bytes) */
+struct hwrm_nvm_set_variable_output {
+	__le16 error_code;
+	__le16 req_type;
+	__le16 seq_id;
+	__le16 resp_len;
+	__le32 unused_0;
+	u8 unused_1;
+	u8 unused_2;
+	u8 unused_3;
+	u8 valid;
+};
+
+/* Command specific Error Codes (8 bytes) */
+struct hwrm_nvm_set_variable_cmd_err {
+	u8 code;
+	#define NVM_SET_VARIABLE_CMD_ERR_CODE_UNKNOWN		   0x0UL
+	#define NVM_SET_VARIABLE_CMD_ERR_CODE_VAR_NOT_EXIST       0x1UL
+	#define NVM_SET_VARIABLE_CMD_ERR_CODE_CORRUPT_VAR	   0x2UL
+	u8 unused_0[7];
+};
+
 /* hwrm_selftest_qlist */
 /* Input (16 bytes) */
 struct hwrm_selftest_qlist_input {
@@ -5686,7 +5949,7 @@ struct hwrm_selftest_qlist_input {
 	__le64 resp_addr;
 };
 
-/* Output (248 bytes) */
+/* Output (280 bytes) */
 struct hwrm_selftest_qlist_output {
 	__le16 error_code;
 	__le16 req_type;
@@ -5698,15 +5961,15 @@ struct hwrm_selftest_qlist_output {
 	#define SELFTEST_QLIST_RESP_AVAILABLE_TESTS_LINK_TEST      0x2UL
 	#define SELFTEST_QLIST_RESP_AVAILABLE_TESTS_REGISTER_TEST  0x4UL
 	#define SELFTEST_QLIST_RESP_AVAILABLE_TESTS_MEMORY_TEST    0x8UL
-	#define SELFTEST_QLIST_RESP_AVAILABLE_TESTS_PCIE_EYE_TEST  0x10UL
-	#define SELFTEST_QLIST_RESP_AVAILABLE_TESTS_ETHERNET_EYE_TEST 0x20UL
+	#define SELFTEST_QLIST_RESP_AVAILABLE_TESTS_PCIE_SERDES_TEST 0x10UL
+	#define SELFTEST_QLIST_RESP_AVAILABLE_TESTS_ETHERNET_SERDES_TEST 0x20UL
 	u8 offline_tests;
 	#define SELFTEST_QLIST_RESP_OFFLINE_TESTS_NVM_TEST	    0x1UL
 	#define SELFTEST_QLIST_RESP_OFFLINE_TESTS_LINK_TEST	    0x2UL
 	#define SELFTEST_QLIST_RESP_OFFLINE_TESTS_REGISTER_TEST    0x4UL
 	#define SELFTEST_QLIST_RESP_OFFLINE_TESTS_MEMORY_TEST      0x8UL
-	#define SELFTEST_QLIST_RESP_OFFLINE_TESTS_PCIE_EYE_TEST    0x10UL
-	#define SELFTEST_QLIST_RESP_OFFLINE_TESTS_ETHERNET_EYE_TEST 0x20UL
+	#define SELFTEST_QLIST_RESP_OFFLINE_TESTS_PCIE_SERDES_TEST 0x10UL
+	#define SELFTEST_QLIST_RESP_OFFLINE_TESTS_ETHERNET_SERDES_TEST 0x20UL
 	u8 unused_0;
 	__le16 test_timeout;
 	u8 unused_1;
@@ -5719,6 +5982,11 @@ struct hwrm_selftest_qlist_output {
 	char test5_name[32];
 	char test6_name[32];
 	char test7_name[32];
+	__le32 unused_3;
+	u8 unused_4;
+	u8 unused_5;
+	u8 unused_6;
+	u8 valid;
 };
 
 /* hwrm_selftest_exec */
@@ -5734,8 +6002,8 @@ struct hwrm_selftest_exec_input {
 	#define SELFTEST_EXEC_REQ_FLAGS_LINK_TEST		    0x2UL
 	#define SELFTEST_EXEC_REQ_FLAGS_REGISTER_TEST		    0x4UL
 	#define SELFTEST_EXEC_REQ_FLAGS_MEMORY_TEST		    0x8UL
-	#define SELFTEST_EXEC_REQ_FLAGS_PCIE_EYE_TEST		    0x10UL
-	#define SELFTEST_EXEC_REQ_FLAGS_ETHERNET_EYE_TEST	    0x20UL
+	#define SELFTEST_EXEC_REQ_FLAGS_PCIE_SERDES_TEST	    0x10UL
+	#define SELFTEST_EXEC_REQ_FLAGS_ETHERNET_SERDES_TEST       0x20UL
 	u8 unused_0[7];
 };
 
@@ -5750,16 +6018,21 @@ struct hwrm_selftest_exec_output {
 	#define SELFTEST_EXEC_RESP_REQUESTED_TESTS_LINK_TEST       0x2UL
 	#define SELFTEST_EXEC_RESP_REQUESTED_TESTS_REGISTER_TEST   0x4UL
 	#define SELFTEST_EXEC_RESP_REQUESTED_TESTS_MEMORY_TEST     0x8UL
-	#define SELFTEST_EXEC_RESP_REQUESTED_TESTS_PCIE_EYE_TEST   0x10UL
-	#define SELFTEST_EXEC_RESP_REQUESTED_TESTS_ETHERNET_EYE_TEST 0x20UL
+	#define SELFTEST_EXEC_RESP_REQUESTED_TESTS_PCIE_SERDES_TEST 0x10UL
+	#define SELFTEST_EXEC_RESP_REQUESTED_TESTS_ETHERNET_SERDES_TEST 0x20UL
 	u8 test_success;
 	#define SELFTEST_EXEC_RESP_TEST_SUCCESS_NVM_TEST	    0x1UL
 	#define SELFTEST_EXEC_RESP_TEST_SUCCESS_LINK_TEST	    0x2UL
 	#define SELFTEST_EXEC_RESP_TEST_SUCCESS_REGISTER_TEST      0x4UL
 	#define SELFTEST_EXEC_RESP_TEST_SUCCESS_MEMORY_TEST	    0x8UL
-	#define SELFTEST_EXEC_RESP_TEST_SUCCESS_PCIE_EYE_TEST      0x10UL
-	#define SELFTEST_EXEC_RESP_TEST_SUCCESS_ETHERNET_EYE_TEST  0x20UL
-	__le16 unused_0[3];
+	#define SELFTEST_EXEC_RESP_TEST_SUCCESS_PCIE_SERDES_TEST   0x10UL
+	#define SELFTEST_EXEC_RESP_TEST_SUCCESS_ETHERNET_SERDES_TEST 0x20UL
+	u8 unused_0;
+	u8 unused_1;
+	u8 unused_2;
+	u8 unused_3;
+	u8 unused_4;
+	u8 valid;
 };
 
 /* hwrm_selftest_irq */
@@ -5772,12 +6045,50 @@ struct hwrm_selftest_irq_input {
 	__le64 resp_addr;
 };
 
-/* Output (8 bytes) */
+/* Output (16 bytes) */
 struct hwrm_selftest_irq_output {
 	__le16 error_code;
 	__le16 req_type;
 	__le16 seq_id;
 	__le16 resp_len;
+	__le32 unused_0;
+	u8 unused_1;
+	u8 unused_2;
+	u8 unused_3;
+	u8 valid;
+};
+
+/* hwrm_selftest_retrieve_serdes_data */
+/* Input (32 bytes) */
+struct hwrm_selftest_retrieve_serdes_data_input {
+	__le16 req_type;
+	__le16 cmpl_ring;
+	__le16 seq_id;
+	__le16 target_id;
+	__le64 resp_addr;
+	__le64 resp_data_addr;
+	__le32 resp_data_offset;
+	__le16 data_len;
+	u8 flags;
+	#define SELFTEST_RETRIEVE_SERDES_DATA_REQ_FLAGS_UNUSED_TEST_MASK 0xfUL
+	#define SELFTEST_RETRIEVE_SERDES_DATA_REQ_FLAGS_UNUSED_TEST_SFT 0
+	#define SELFTEST_RETRIEVE_SERDES_DATA_REQ_FLAGS_PCIE_SERDES_TEST 0x10UL
+	#define SELFTEST_RETRIEVE_SERDES_DATA_REQ_FLAGS_ETHERNET_SERDES_TEST 0x20UL
+	u8 unused_0;
+};
+
+/* Output (16 bytes) */
+struct hwrm_selftest_retrieve_serdes_data_output {
+	__le16 error_code;
+	__le16 req_type;
+	__le16 seq_id;
+	__le16 resp_len;
+	__le16 total_data_len;
+	__le16 copied_data_len;
+	u8 unused_0;
+	u8 unused_1;
+	u8 unused_2;
+	u8 valid;
 };
 
 /* Hardware Resource Manager Specification */
@@ -5938,10 +6249,16 @@ struct cmd_nums {
 	#define HWRM_CFA_DECAP_FILTER_ALLOC			   (0x108UL)
 	#define HWRM_CFA_DECAP_FILTER_FREE			   (0x109UL)
 	#define HWRM_CFA_VLAN_ANTISPOOF_QCFG			   (0x10aUL)
+	#define HWRM_CFA_REDIRECT_TUNNEL_TYPE_ALLOC		   (0x10bUL)
+	#define HWRM_CFA_REDIRECT_TUNNEL_TYPE_FREE		   (0x10cUL)
+	#define HWRM_CFA_PAIR_ALLOC				   (0x10dUL)
+	#define HWRM_CFA_PAIR_FREE				   (0x10eUL)
+	#define HWRM_CFA_PAIR_INFO				   (0x10fUL)
+	#define HWRM_FW_IPC_MSG				   (0x110UL)
 	#define HWRM_SELFTEST_QLIST				   (0x200UL)
 	#define HWRM_SELFTEST_EXEC				   (0x201UL)
 	#define HWRM_SELFTEST_IRQ				   (0x202UL)
-	#define HWRM_SELFTEST_RETREIVE_EYE_DATA		   (0x203UL)
+	#define HWRM_SELFTEST_RETRIEVE_SERDES_DATA		   (0x203UL)
 	#define HWRM_DBG_READ_DIRECT				   (0xff10UL)
 	#define HWRM_DBG_READ_INDIRECT				   (0xff11UL)
 	#define HWRM_DBG_WRITE_DIRECT				   (0xff12UL)
@@ -5949,6 +6266,9 @@ struct cmd_nums {
 	#define HWRM_DBG_DUMP					   (0xff14UL)
 	#define HWRM_DBG_ERASE_NVM				   (0xff15UL)
 	#define HWRM_DBG_CFG					   (0xff16UL)
+	#define HWRM_DBG_COREDUMP_LIST				   (0xff17UL)
+	#define HWRM_DBG_COREDUMP_INITIATE			   (0xff18UL)
+	#define HWRM_DBG_COREDUMP_RETRIEVE			   (0xff19UL)
 	#define HWRM_NVM_FACTORY_DEFAULTS			   (0xffeeUL)
 	#define HWRM_NVM_VALIDATE_OPTION			   (0xffefUL)
 	#define HWRM_NVM_FLUSH					   (0xfff0UL)
@@ -6123,6 +6443,58 @@ struct rx_port_stats {
 	__le64 rx_stat_err;
 };
 
+/* VXLAN IPv4 encapsulation structure (16 bytes) */
+struct hwrm_vxlan_ipv4_hdr {
+	u8 ver_hlen;
+	#define VXLAN_IPV4_HDR_VER_HLEN_HEADER_LENGTH_MASK	    0xfUL
+	#define VXLAN_IPV4_HDR_VER_HLEN_HEADER_LENGTH_SFT	    0
+	#define VXLAN_IPV4_HDR_VER_HLEN_VERSION_MASK		    0xf0UL
+	#define VXLAN_IPV4_HDR_VER_HLEN_VERSION_SFT		    4
+	u8 tos;
+	__be16 ip_id;
+	__be16 flags_frag_offset;
+	u8 ttl;
+	u8 protocol;
+	__be32 src_ip_addr;
+	__be32 dest_ip_addr;
+};
+
+/* VXLAN IPv6 encapsulation structure (32 bytes) */
+struct hwrm_vxlan_ipv6_hdr {
+	__be32 ver_tc_flow_label;
+	#define VXLAN_IPV6_HDR_VER_TC_FLOW_LABEL_VER_SFT	   0x1cUL
+	#define VXLAN_IPV6_HDR_VER_TC_FLOW_LABEL_VER_MASK	   0xf0000000UL
+	#define VXLAN_IPV6_HDR_VER_TC_FLOW_LABEL_TC_SFT	   0x14UL
+	#define VXLAN_IPV6_HDR_VER_TC_FLOW_LABEL_TC_MASK	   0xff00000UL
+	#define VXLAN_IPV6_HDR_VER_TC_FLOW_LABEL_FLOW_LABEL_SFT   0x0UL
+	#define VXLAN_IPV6_HDR_VER_TC_FLOW_LABEL_FLOW_LABEL_MASK  0xfffffUL
+	__be16 payload_len;
+	u8 next_hdr;
+	u8 ttl;
+	__be32 src_ip_addr[4];
+	__be32 dest_ip_addr[4];
+};
+
+/* VXLAN encapsulation structure (72 bytes) */
+struct hwrm_cfa_encap_data_vxlan {
+	u8 src_mac_addr[6];
+	__le16 unused_0;
+	u8 dst_mac_addr[6];
+	u8 num_vlan_tags;
+	u8 unused_1;
+	__be16 ovlan_tpid;
+	__be16 ovlan_tci;
+	__be16 ivlan_tpid;
+	__be16 ivlan_tci;
+	__le32 l3[10];
+	#define CFA_ENCAP_DATA_VXLAN_L3_VER_MASK		   0xfUL
+	#define CFA_ENCAP_DATA_VXLAN_L3_VER_IPV4		   0x4UL
+	#define CFA_ENCAP_DATA_VXLAN_L3_VER_IPV6		   0x6UL
+	__be16 src_port;
+	__be16 dst_port;
+	__be32 vni;
+};
+
 /* Periodic Statistics Context DMA to host (160 bytes) */
 struct ctx_hw_stats {
 	__le64 rx_ucast_pkts;

drivers/net/ethernet/broadcom/bnxt/bnxt_tc.c

@@ -16,6 +16,7 @@
 #include <net/tc_act/tc_skbedit.h>
 #include <net/tc_act/tc_mirred.h>
 #include <net/tc_act/tc_vlan.h>
+#include <net/tc_act/tc_tunnel_key.h>
 
 #include "bnxt_hsi.h"
 #include "bnxt.h"
@@ -89,6 +90,23 @@ static void bnxt_tc_parse_vlan(struct bnxt *bp,
 	}
 }
 
+static int bnxt_tc_parse_tunnel_set(struct bnxt *bp,
+				    struct bnxt_tc_actions *actions,
+				    const struct tc_action *tc_act)
+{
+	struct ip_tunnel_info *tun_info = tcf_tunnel_info(tc_act);
+	struct ip_tunnel_key *tun_key = &tun_info->key;
+
+	if (ip_tunnel_info_af(tun_info) != AF_INET) {
+		netdev_info(bp->dev, "only IPv4 tunnel-encap is supported");
+		return -EOPNOTSUPP;
+	}
+
+	actions->tun_encap_key = *tun_key;
+	actions->flags |= BNXT_TC_ACTION_FLAG_TUNNEL_ENCAP;
+	return 0;
+}
+
 static int bnxt_tc_parse_actions(struct bnxt *bp,
 				 struct bnxt_tc_actions *actions,
 				 struct tcf_exts *tc_exts)
@@ -123,9 +141,35 @@ static int bnxt_tc_parse_actions(struct bnxt *bp,
 			bnxt_tc_parse_vlan(bp, actions, tc_act);
 			continue;
 		}
+
+		/* Tunnel encap */
+		if (is_tcf_tunnel_set(tc_act)) {
+			rc = bnxt_tc_parse_tunnel_set(bp, actions, tc_act);
+			if (rc)
+				return rc;
+			continue;
+		}
+
+		/* Tunnel decap */
+		if (is_tcf_tunnel_release(tc_act)) {
+			actions->flags |= BNXT_TC_ACTION_FLAG_TUNNEL_DECAP;
+			continue;
+		}
 	}
 
-	return 0;
+	if (rc)
+		return rc;
+
+	/* Tunnel encap/decap action must be accompanied by a redirect action */
+	if ((actions->flags & BNXT_TC_ACTION_FLAG_TUNNEL_ENCAP ||
+	     actions->flags & BNXT_TC_ACTION_FLAG_TUNNEL_DECAP) &&
+	    !(actions->flags & BNXT_TC_ACTION_FLAG_FWD)) {
+		netdev_info(bp->dev,
+			    "error: no redir action along with encap/decap");
+		return -EINVAL;
+	}
+
+	return rc;
 }
 
 #define GET_KEY(flow_cmd, key_type)					\
@@ -252,6 +296,54 @@ static int bnxt_tc_parse_flow(struct bnxt *bp,
 		flow->l4_mask.icmp.code = mask->code;
 	}
 
+	if (dissector_uses_key(dissector, FLOW_DISSECTOR_KEY_ENC_CONTROL)) {
+		struct flow_dissector_key_control *key =
+			GET_KEY(tc_flow_cmd, FLOW_DISSECTOR_KEY_ENC_CONTROL);
+
+		addr_type = key->addr_type;
+	}
+
+	if (dissector_uses_key(dissector, FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS)) {
+		struct flow_dissector_key_ipv4_addrs *key =
+			GET_KEY(tc_flow_cmd, FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS);
+		struct flow_dissector_key_ipv4_addrs *mask =
+				GET_MASK(tc_flow_cmd,
+					 FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS);
+
+		flow->flags |= BNXT_TC_FLOW_FLAGS_TUNL_IPV4_ADDRS;
+		flow->tun_key.u.ipv4.dst = key->dst;
+		flow->tun_mask.u.ipv4.dst = mask->dst;
+		flow->tun_key.u.ipv4.src = key->src;
+		flow->tun_mask.u.ipv4.src = mask->src;
+	} else if (dissector_uses_key(dissector,
+				      FLOW_DISSECTOR_KEY_ENC_IPV6_ADDRS)) {
+		return -EOPNOTSUPP;
+	}
+
+	if (dissector_uses_key(dissector, FLOW_DISSECTOR_KEY_ENC_KEYID)) {
+		struct flow_dissector_key_keyid *key =
+			GET_KEY(tc_flow_cmd, FLOW_DISSECTOR_KEY_ENC_KEYID);
+		struct flow_dissector_key_keyid *mask =
+			GET_MASK(tc_flow_cmd, FLOW_DISSECTOR_KEY_ENC_KEYID);
+
+		flow->flags |= BNXT_TC_FLOW_FLAGS_TUNL_ID;
+		flow->tun_key.tun_id = key32_to_tunnel_id(key->keyid);
+		flow->tun_mask.tun_id = key32_to_tunnel_id(mask->keyid);
+	}
+
+	if (dissector_uses_key(dissector, FLOW_DISSECTOR_KEY_ENC_PORTS)) {
+		struct flow_dissector_key_ports *key =
+			GET_KEY(tc_flow_cmd, FLOW_DISSECTOR_KEY_ENC_PORTS);
+		struct flow_dissector_key_ports *mask =
+			GET_MASK(tc_flow_cmd, FLOW_DISSECTOR_KEY_ENC_PORTS);
+
+		flow->flags |= BNXT_TC_FLOW_FLAGS_TUNL_PORTS;
+		flow->tun_key.tp_dst = key->dst;
+		flow->tun_mask.tp_dst = mask->dst;
+		flow->tun_key.tp_src = key->src;
+		flow->tun_mask.tp_src = mask->src;
+	}
+
 	return bnxt_tc_parse_actions(bp, &flow->actions, tc_flow_cmd->exts);
 }
 
@@ -293,7 +385,8 @@ static bool is_wildcard(void *mask, int len)
 }
 
 static int bnxt_hwrm_cfa_flow_alloc(struct bnxt *bp, struct bnxt_tc_flow *flow,
-				    __le16 ref_flow_handle, __le16 *flow_handle)
+				    __le16 ref_flow_handle,
+				    __le32 tunnel_handle, __le16 *flow_handle)
 {
 	struct hwrm_cfa_flow_alloc_output *resp = bp->hwrm_cmd_resp_addr;
 	struct bnxt_tc_actions *actions = &flow->actions;
@@ -307,6 +400,14 @@ static int bnxt_hwrm_cfa_flow_alloc(struct bnxt *bp, struct bnxt_tc_flow *flow,
 
 	req.src_fid = cpu_to_le16(flow->src_fid);
 	req.ref_flow_handle = ref_flow_handle;
+
+	if (actions->flags & BNXT_TC_ACTION_FLAG_TUNNEL_DECAP ||
+	    actions->flags & BNXT_TC_ACTION_FLAG_TUNNEL_ENCAP) {
+		req.tunnel_handle = tunnel_handle;
+		flow_flags |= CFA_FLOW_ALLOC_REQ_FLAGS_TUNNEL;
+		action_flags |= CFA_FLOW_ALLOC_REQ_ACTION_FLAGS_TUNNEL;
+	}
+
 	req.ethertype = flow->l2_key.ether_type;
 	req.ip_proto = flow->l4_key.ip_proto;
 
@@ -403,78 +504,153 @@ static int bnxt_hwrm_cfa_flow_alloc(struct bnxt *bp, struct bnxt_tc_flow *flow,
 	return rc;
 }
 
-/* Add val to accum while handling a possible wraparound
- * of val. Eventhough val is of type u64, its actual width
- * is denoted by mask and will wrap-around beyond that width.
- */
-static void accumulate_val(u64 *accum, u64 val, u64 mask)
+static int hwrm_cfa_decap_filter_alloc(struct bnxt *bp,
+				       struct bnxt_tc_flow *flow,
+				       struct bnxt_tc_l2_key *l2_info,
+				       __le32 ref_decap_handle,
+				       __le32 *decap_filter_handle)
 {
-#define low_bits(x, mask)		((x) & (mask))
-#define high_bits(x, mask)		((x) & ~(mask))
-	bool wrapped = val < low_bits(*accum, mask);
+	struct hwrm_cfa_decap_filter_alloc_output *resp =
+						bp->hwrm_cmd_resp_addr;
+	struct hwrm_cfa_decap_filter_alloc_input req = { 0 };
+	struct ip_tunnel_key *tun_key = &flow->tun_key;
+	u32 enables = 0;
+	int rc;
 
-	*accum = high_bits(*accum, mask) + val;
-	if (wrapped)
-		*accum += (mask + 1);
-}
+	bnxt_hwrm_cmd_hdr_init(bp, &req, HWRM_CFA_DECAP_FILTER_ALLOC, -1, -1);
 
-/* The HW counters' width is much less than 64bits.
- * Handle possible wrap-around while updating the stat counters
- */
-static void bnxt_flow_stats_fix_wraparound(struct bnxt_tc_info *tc_info,
-					   struct bnxt_tc_flow_stats *stats,
-					   struct bnxt_tc_flow_stats *hw_stats)
-{
-	accumulate_val(&stats->bytes, hw_stats->bytes, tc_info->bytes_mask);
-	accumulate_val(&stats->packets, hw_stats->packets,
-		       tc_info->packets_mask);
+	req.flags = cpu_to_le32(CFA_DECAP_FILTER_ALLOC_REQ_FLAGS_OVS_TUNNEL);
+	enables |= CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_TUNNEL_TYPE |
+		   CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_IP_PROTOCOL;
+	req.tunnel_type = CFA_DECAP_FILTER_ALLOC_REQ_TUNNEL_TYPE_VXLAN;
+	req.ip_protocol = CFA_DECAP_FILTER_ALLOC_REQ_IP_PROTOCOL_UDP;
+
+	if (flow->flags & BNXT_TC_FLOW_FLAGS_TUNL_ID) {
+		enables |= CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_TUNNEL_ID;
+		/* tunnel_id is wrongly defined in hsi defn. as __le32 */
+		req.tunnel_id = tunnel_id_to_key32(tun_key->tun_id);
+	}
+
+	if (flow->flags & BNXT_TC_FLOW_FLAGS_TUNL_ETH_ADDRS) {
+		enables |= CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_DST_MACADDR |
+			   CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_SRC_MACADDR;
+		ether_addr_copy(req.dst_macaddr, l2_info->dmac);
+		ether_addr_copy(req.src_macaddr, l2_info->smac);
+	}
+	if (l2_info->num_vlans) {
+		enables |= CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_T_IVLAN_VID;
+		req.t_ivlan_vid = l2_info->inner_vlan_tci;
+	}
+
+	enables |= CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_ETHERTYPE;
+	req.ethertype = htons(ETH_P_IP);
+
+	if (flow->flags & BNXT_TC_FLOW_FLAGS_TUNL_IPV4_ADDRS) {
+		enables |= CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_SRC_IPADDR |
+			   CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_DST_IPADDR |
+			   CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_IPADDR_TYPE;
+		req.ip_addr_type = CFA_DECAP_FILTER_ALLOC_REQ_IP_ADDR_TYPE_IPV4;
+		req.dst_ipaddr[0] = tun_key->u.ipv4.dst;
+		req.src_ipaddr[0] = tun_key->u.ipv4.src;
+	}
+
+	if (flow->flags & BNXT_TC_FLOW_FLAGS_TUNL_PORTS) {
+		enables |= CFA_DECAP_FILTER_ALLOC_REQ_ENABLES_DST_PORT;
+		req.dst_port = tun_key->tp_dst;
+	}
+
+	/* Eventhough the decap_handle returned by hwrm_cfa_decap_filter_alloc
+	 * is defined as __le32, l2_ctxt_ref_id is defined in HSI as __le16.
+	 */
+	req.l2_ctxt_ref_id = (__force __le16)ref_decap_handle;
+	req.enables = cpu_to_le32(enables);
+
+	mutex_lock(&bp->hwrm_cmd_lock);
+	rc = _hwrm_send_message(bp, &req, sizeof(req), HWRM_CMD_TIMEOUT);
+	if (!rc)
+		*decap_filter_handle = resp->decap_filter_id;
+	else
+		netdev_info(bp->dev, "%s: Error rc=%d", __func__, rc);
+	mutex_unlock(&bp->hwrm_cmd_lock);
+
+	return rc;
 }
 
-/* Fix possible wraparound of the stats queried from HW, calculate
- * the delta from prev_stats, and also update the prev_stats.
- * The HW flow stats are fetched under the hwrm_cmd_lock mutex.
- * This routine is best called while under the mutex so that the
- * stats processing happens atomically.
- */
-static void bnxt_flow_stats_calc(struct bnxt_tc_info *tc_info,
-				 struct bnxt_tc_flow *flow,
-				 struct bnxt_tc_flow_stats *stats)
+static int hwrm_cfa_decap_filter_free(struct bnxt *bp,
+				      __le32 decap_filter_handle)
 {
-	struct bnxt_tc_flow_stats *acc_stats, *prev_stats;
+	struct hwrm_cfa_decap_filter_free_input req = { 0 };
+	int rc;
 
-	acc_stats = &flow->stats;
-	bnxt_flow_stats_fix_wraparound(tc_info, acc_stats, stats);
+	bnxt_hwrm_cmd_hdr_init(bp, &req, HWRM_CFA_DECAP_FILTER_FREE, -1, -1);
+	req.decap_filter_id = decap_filter_handle;
 
-	prev_stats = &flow->prev_stats;
-	stats->bytes = acc_stats->bytes - prev_stats->bytes;
-	stats->packets = acc_stats->packets - prev_stats->packets;
-	*prev_stats = *acc_stats;
+	rc = hwrm_send_message(bp, &req, sizeof(req), HWRM_CMD_TIMEOUT);
+	if (rc)
+		netdev_info(bp->dev, "%s: Error rc=%d", __func__, rc);
+	return rc;
 }
 
-static int bnxt_hwrm_cfa_flow_stats_get(struct bnxt *bp,
-					__le16 flow_handle,
-					struct bnxt_tc_flow *flow,
-					struct bnxt_tc_flow_stats *stats)
+static int hwrm_cfa_encap_record_alloc(struct bnxt *bp,
+				       struct ip_tunnel_key *encap_key,
+				       struct bnxt_tc_l2_key *l2_info,
+				       __le32 *encap_record_handle)
 {
-	struct hwrm_cfa_flow_stats_output *resp = bp->hwrm_cmd_resp_addr;
-	struct hwrm_cfa_flow_stats_input req = { 0 };
+	struct hwrm_cfa_encap_record_alloc_output *resp =
+						bp->hwrm_cmd_resp_addr;
+	struct hwrm_cfa_encap_record_alloc_input req = { 0 };
+	struct hwrm_cfa_encap_data_vxlan *encap =
+			(struct hwrm_cfa_encap_data_vxlan *)&req.encap_data;
+	struct hwrm_vxlan_ipv4_hdr *encap_ipv4 =
+				(struct hwrm_vxlan_ipv4_hdr *)encap->l3;
 	int rc;
 
-	bnxt_hwrm_cmd_hdr_init(bp, &req, HWRM_CFA_FLOW_STATS, -1, -1);
-	req.num_flows = cpu_to_le16(1);
-	req.flow_handle_0 = flow_handle;
+	bnxt_hwrm_cmd_hdr_init(bp, &req, HWRM_CFA_ENCAP_RECORD_ALLOC, -1, -1);
 
-	mutex_lock(&bp->hwrm_cmd_lock);
-	rc = _hwrm_send_message(bp, &req, sizeof(req), HWRM_CMD_TIMEOUT);
-	if (!rc) {
-		stats->packets = le64_to_cpu(resp->packet_0);
-		stats->bytes = le64_to_cpu(resp->byte_0);
-		bnxt_flow_stats_calc(&bp->tc_info, flow, stats);
-	} else {
-		netdev_info(bp->dev, "error rc=%d", rc);
+	req.encap_type = CFA_ENCAP_RECORD_ALLOC_REQ_ENCAP_TYPE_VXLAN;
+
+	ether_addr_copy(encap->dst_mac_addr, l2_info->dmac);
+	ether_addr_copy(encap->src_mac_addr, l2_info->smac);
+	if (l2_info->num_vlans) {
+		encap->num_vlan_tags = l2_info->num_vlans;
+		encap->ovlan_tci = l2_info->inner_vlan_tci;
+		encap->ovlan_tpid = l2_info->inner_vlan_tpid;
 	}
 
+	encap_ipv4->ver_hlen = 4 << VXLAN_IPV4_HDR_VER_HLEN_VERSION_SFT;
+	encap_ipv4->ver_hlen |= 5 << VXLAN_IPV4_HDR_VER_HLEN_HEADER_LENGTH_SFT;
+	encap_ipv4->ttl = encap_key->ttl;
+
+	encap_ipv4->dest_ip_addr = encap_key->u.ipv4.dst;
+	encap_ipv4->src_ip_addr = encap_key->u.ipv4.src;
+	encap_ipv4->protocol = IPPROTO_UDP;
+
+	encap->dst_port = encap_key->tp_dst;
+	encap->vni = tunnel_id_to_key32(encap_key->tun_id);
+
+	mutex_lock(&bp->hwrm_cmd_lock);
+	rc = _hwrm_send_message(bp, &req, sizeof(req), HWRM_CMD_TIMEOUT);
+	if (!rc)
+		*encap_record_handle = resp->encap_record_id;
+	else
+		netdev_info(bp->dev, "%s: Error rc=%d", __func__, rc);
 	mutex_unlock(&bp->hwrm_cmd_lock);
+
+	return rc;
+}
+
+static int hwrm_cfa_encap_record_free(struct bnxt *bp,
+				      __le32 encap_record_handle)
+{
+	struct hwrm_cfa_encap_record_free_input req = { 0 };
+	int rc;
+
+	bnxt_hwrm_cmd_hdr_init(bp, &req, HWRM_CFA_ENCAP_RECORD_FREE, -1, -1);
+	req.encap_record_id = encap_record_handle;
+
+	rc = hwrm_send_message(bp, &req, sizeof(req), HWRM_CMD_TIMEOUT);
+	if (rc)
+		netdev_info(bp->dev, "%s: Error rc=%d", __func__, rc);
 	return rc;
 }
 
@@ -482,7 +658,7 @@ static int bnxt_tc_put_l2_node(struct bnxt *bp,
 			       struct bnxt_tc_flow_node *flow_node)
 {
 	struct bnxt_tc_l2_node *l2_node = flow_node->l2_node;
-	struct bnxt_tc_info *tc_info = &bp->tc_info;
+	struct bnxt_tc_info *tc_info = bp->tc_info;
 	int rc;
 
 	/* remove flow_node from the L2 shared flow list */
@@ -519,7 +695,7 @@ bnxt_tc_get_l2_node(struct bnxt *bp, struct rhashtable *l2_table,
 		rc = rhashtable_insert_fast(l2_table, &l2_node->node,
 					    ht_params);
 		if (rc) {
-			kfree(l2_node);
+			kfree_rcu(l2_node, rcu);
 			netdev_err(bp->dev,
 				   "Error: %s: rhashtable_insert_fast: %d",
 				   __func__, rc);
@@ -538,7 +714,7 @@ bnxt_tc_get_ref_flow_handle(struct bnxt *bp, struct bnxt_tc_flow *flow,
 			    struct bnxt_tc_flow_node *flow_node,
 			    __le16 *ref_flow_handle)
 {
-	struct bnxt_tc_info *tc_info = &bp->tc_info;
+	struct bnxt_tc_info *tc_info = bp->tc_info;
 	struct bnxt_tc_flow_node *ref_flow_node;
 	struct bnxt_tc_l2_node *l2_node;
 
@@ -588,10 +764,380 @@ static bool bnxt_tc_can_offload(struct bnxt *bp, struct bnxt_tc_flow *flow)
 	return true;
 }
 
+/* Returns the final refcount of the node on success
+ * or a -ve error code on failure
+ */
+static int bnxt_tc_put_tunnel_node(struct bnxt *bp,
+				   struct rhashtable *tunnel_table,
+				   struct rhashtable_params *ht_params,
+				   struct bnxt_tc_tunnel_node *tunnel_node)
+{
+	int rc;
+
+	if (--tunnel_node->refcount == 0) {
+		rc =  rhashtable_remove_fast(tunnel_table, &tunnel_node->node,
+					     *ht_params);
+		if (rc) {
+			netdev_err(bp->dev, "rhashtable_remove_fast rc=%d", rc);
+			rc = -1;
+		}
+		kfree_rcu(tunnel_node, rcu);
+		return rc;
+	} else {
+		return tunnel_node->refcount;
+	}
+}
+
+/* Get (or add) either encap or decap tunnel node from/to the supplied
+ * hash table.
+ */
+static struct bnxt_tc_tunnel_node *
+bnxt_tc_get_tunnel_node(struct bnxt *bp, struct rhashtable *tunnel_table,
+			struct rhashtable_params *ht_params,
+			struct ip_tunnel_key *tun_key)
+{
+	struct bnxt_tc_tunnel_node *tunnel_node;
+	int rc;
+
+	tunnel_node = rhashtable_lookup_fast(tunnel_table, tun_key, *ht_params);
+	if (!tunnel_node) {
+		tunnel_node = kzalloc(sizeof(*tunnel_node), GFP_KERNEL);
+		if (!tunnel_node) {
+			rc = -ENOMEM;
+			goto err;
+		}
+
+		tunnel_node->key = *tun_key;
+		tunnel_node->tunnel_handle = INVALID_TUNNEL_HANDLE;
+		rc = rhashtable_insert_fast(tunnel_table, &tunnel_node->node,
+					    *ht_params);
+		if (rc) {
+			kfree_rcu(tunnel_node, rcu);
+			goto err;
+		}
+	}
+	tunnel_node->refcount++;
+	return tunnel_node;
+err:
+	netdev_info(bp->dev, "error rc=%d", rc);
+	return NULL;
+}
+
+static int bnxt_tc_get_ref_decap_handle(struct bnxt *bp,
+					struct bnxt_tc_flow *flow,
+					struct bnxt_tc_l2_key *l2_key,
+					struct bnxt_tc_flow_node *flow_node,
+					__le32 *ref_decap_handle)
+{
+	struct bnxt_tc_info *tc_info = bp->tc_info;
+	struct bnxt_tc_flow_node *ref_flow_node;
+	struct bnxt_tc_l2_node *decap_l2_node;
+
+	decap_l2_node = bnxt_tc_get_l2_node(bp, &tc_info->decap_l2_table,
+					    tc_info->decap_l2_ht_params,
+					    l2_key);
+	if (!decap_l2_node)
+		return -1;
+
+	/* If any other flow is using this decap_l2_node, use it's decap_handle
+	 * as the ref_decap_handle
+	 */
+	if (decap_l2_node->refcount > 0) {
+		ref_flow_node =
+			list_first_entry(&decap_l2_node->common_l2_flows,
+					 struct bnxt_tc_flow_node,
+					 decap_l2_list_node);
+		*ref_decap_handle = ref_flow_node->decap_node->tunnel_handle;
+	} else {
+		*ref_decap_handle = INVALID_TUNNEL_HANDLE;
+	}
+
+	/* Insert the l2_node into the flow_node so that subsequent flows
+	 * with a matching decap l2 key can use the decap_filter_handle of
+	 * this flow as their ref_decap_handle
+	 */
+	flow_node->decap_l2_node = decap_l2_node;
+	list_add(&flow_node->decap_l2_list_node,
+		 &decap_l2_node->common_l2_flows);
+	decap_l2_node->refcount++;
+	return 0;
+}
+
+static void bnxt_tc_put_decap_l2_node(struct bnxt *bp,
+				      struct bnxt_tc_flow_node *flow_node)
+{
+	struct bnxt_tc_l2_node *decap_l2_node = flow_node->decap_l2_node;
+	struct bnxt_tc_info *tc_info = bp->tc_info;
+	int rc;
+
+	/* remove flow_node from the decap L2 sharing flow list */
+	list_del(&flow_node->decap_l2_list_node);
+	if (--decap_l2_node->refcount == 0) {
+		rc =  rhashtable_remove_fast(&tc_info->decap_l2_table,
+					     &decap_l2_node->node,
+					     tc_info->decap_l2_ht_params);
+		if (rc)
+			netdev_err(bp->dev, "rhashtable_remove_fast rc=%d", rc);
+		kfree_rcu(decap_l2_node, rcu);
+	}
+}
+
+static void bnxt_tc_put_decap_handle(struct bnxt *bp,
+				     struct bnxt_tc_flow_node *flow_node)
+{
+	__le32 decap_handle = flow_node->decap_node->tunnel_handle;
+	struct bnxt_tc_info *tc_info = bp->tc_info;
+	int rc;
+
+	if (flow_node->decap_l2_node)
+		bnxt_tc_put_decap_l2_node(bp, flow_node);
+
+	rc = bnxt_tc_put_tunnel_node(bp, &tc_info->decap_table,
+				     &tc_info->decap_ht_params,
+				     flow_node->decap_node);
+	if (!rc && decap_handle != INVALID_TUNNEL_HANDLE)
+		hwrm_cfa_decap_filter_free(bp, decap_handle);
+}
+
+static int bnxt_tc_resolve_tunnel_hdrs(struct bnxt *bp,
+				       struct ip_tunnel_key *tun_key,
+				       struct bnxt_tc_l2_key *l2_info,
+				       struct net_device *real_dst_dev)
+{
+	struct flowi4 flow = { {0} };
+	struct net_device *dst_dev;
+	struct neighbour *nbr;
+	struct rtable *rt;
+	int rc;
+
+	flow.flowi4_proto = IPPROTO_UDP;
+	flow.fl4_dport = tun_key->tp_dst;
+	flow.daddr = tun_key->u.ipv4.dst;
+
+	rt = ip_route_output_key(dev_net(real_dst_dev), &flow);
+	if (IS_ERR(rt)) {
+		netdev_info(bp->dev, "no route to %pI4b", &flow.daddr);
+		return -EOPNOTSUPP;
+	}
+
+	/* The route must either point to the real_dst_dev or a dst_dev that
+	 * uses the real_dst_dev.
+	 */
+	dst_dev = rt->dst.dev;
+	if (is_vlan_dev(dst_dev)) {
+		struct vlan_dev_priv *vlan = vlan_dev_priv(dst_dev);
+
+		if (vlan->real_dev != real_dst_dev) {
+			netdev_info(bp->dev,
+				    "dst_dev(%s) doesn't use PF-if(%s)",
+				    netdev_name(dst_dev),
+				    netdev_name(real_dst_dev));
+			rc = -EOPNOTSUPP;
+			goto put_rt;
+		}
+		l2_info->inner_vlan_tci = htons(vlan->vlan_id);
+		l2_info->inner_vlan_tpid = vlan->vlan_proto;
+		l2_info->num_vlans = 1;
+	} else if (dst_dev != real_dst_dev) {
+		netdev_info(bp->dev,
+			    "dst_dev(%s) for %pI4b is not PF-if(%s)",
+			    netdev_name(dst_dev), &flow.daddr,
+			    netdev_name(real_dst_dev));
+		rc = -EOPNOTSUPP;
+		goto put_rt;
+	}
+
+	nbr = dst_neigh_lookup(&rt->dst, &flow.daddr);
+	if (!nbr) {
+		netdev_info(bp->dev, "can't lookup neighbor for %pI4b",
+			    &flow.daddr);
+		rc = -EOPNOTSUPP;
+		goto put_rt;
+	}
+
+	tun_key->u.ipv4.src = flow.saddr;
+	tun_key->ttl = ip4_dst_hoplimit(&rt->dst);
+	neigh_ha_snapshot(l2_info->dmac, nbr, dst_dev);
+	ether_addr_copy(l2_info->smac, dst_dev->dev_addr);
+	neigh_release(nbr);
+	ip_rt_put(rt);
+
+	return 0;
+put_rt:
+	ip_rt_put(rt);
+	return rc;
+}
+
+static int bnxt_tc_get_decap_handle(struct bnxt *bp, struct bnxt_tc_flow *flow,
+				    struct bnxt_tc_flow_node *flow_node,
+				    __le32 *decap_filter_handle)
+{
+	struct ip_tunnel_key *decap_key = &flow->tun_key;
+	struct bnxt_tc_info *tc_info = bp->tc_info;
+	struct bnxt_tc_l2_key l2_info = { {0} };
+	struct bnxt_tc_tunnel_node *decap_node;
+	struct ip_tunnel_key tun_key = { 0 };
+	struct bnxt_tc_l2_key *decap_l2_info;
+	__le32 ref_decap_handle;
+	int rc;
+
+	/* Check if there's another flow using the same tunnel decap.
+	 * If not, add this tunnel to the table and resolve the other
+	 * tunnel header fileds
+	 */
+	decap_node = bnxt_tc_get_tunnel_node(bp, &tc_info->decap_table,
+					     &tc_info->decap_ht_params,
+					     decap_key);
+	if (!decap_node)
+		return -ENOMEM;
+
+	flow_node->decap_node = decap_node;
+
+	if (decap_node->tunnel_handle != INVALID_TUNNEL_HANDLE)
+		goto done;
+
+	/* Resolve the L2 fields for tunnel decap
+	 * Resolve the route for remote vtep (saddr) of the decap key
+	 * Find it's next-hop mac addrs
+	 */
+	tun_key.u.ipv4.dst = flow->tun_key.u.ipv4.src;
+	tun_key.tp_dst = flow->tun_key.tp_dst;
+	rc = bnxt_tc_resolve_tunnel_hdrs(bp, &tun_key, &l2_info, bp->dev);
+	if (rc)
+		goto put_decap;
+
+	decap_key->ttl = tun_key.ttl;
+	decap_l2_info = &decap_node->l2_info;
+	ether_addr_copy(decap_l2_info->dmac, l2_info.smac);
+	ether_addr_copy(decap_l2_info->smac, l2_info.dmac);
+	if (l2_info.num_vlans) {
+		decap_l2_info->num_vlans = l2_info.num_vlans;
+		decap_l2_info->inner_vlan_tpid = l2_info.inner_vlan_tpid;
+		decap_l2_info->inner_vlan_tci = l2_info.inner_vlan_tci;
+	}
+	flow->flags |= BNXT_TC_FLOW_FLAGS_TUNL_ETH_ADDRS;
+
+	/* For getting a decap_filter_handle we first need to check if
+	 * there are any other decap flows that share the same tunnel L2
+	 * key and if so, pass that flow's decap_filter_handle as the
+	 * ref_decap_handle for this flow.
+	 */
+	rc = bnxt_tc_get_ref_decap_handle(bp, flow, decap_l2_info, flow_node,
+					  &ref_decap_handle);
+	if (rc)
+		goto put_decap;
+
+	/* Issue the hwrm cmd to allocate a decap filter handle */
+	rc = hwrm_cfa_decap_filter_alloc(bp, flow, decap_l2_info,
+					 ref_decap_handle,
+					 &decap_node->tunnel_handle);
+	if (rc)
+		goto put_decap_l2;
+
+done:
+	*decap_filter_handle = decap_node->tunnel_handle;
+	return 0;
+
+put_decap_l2:
+	bnxt_tc_put_decap_l2_node(bp, flow_node);
+put_decap:
+	bnxt_tc_put_tunnel_node(bp, &tc_info->decap_table,
+				&tc_info->decap_ht_params,
+				flow_node->decap_node);
+	return rc;
+}
+
+static void bnxt_tc_put_encap_handle(struct bnxt *bp,
+				     struct bnxt_tc_tunnel_node *encap_node)
+{
+	__le32 encap_handle = encap_node->tunnel_handle;
+	struct bnxt_tc_info *tc_info = bp->tc_info;
+	int rc;
+
+	rc = bnxt_tc_put_tunnel_node(bp, &tc_info->encap_table,
+				     &tc_info->encap_ht_params, encap_node);
+	if (!rc && encap_handle != INVALID_TUNNEL_HANDLE)
+		hwrm_cfa_encap_record_free(bp, encap_handle);
+}
+
+/* Lookup the tunnel encap table and check if there's an encap_handle
+ * alloc'd already.
+ * If not, query L2 info via a route lookup and issue an encap_record_alloc
+ * cmd to FW.
+ */
+static int bnxt_tc_get_encap_handle(struct bnxt *bp, struct bnxt_tc_flow *flow,
+				    struct bnxt_tc_flow_node *flow_node,
+				    __le32 *encap_handle)
+{
+	struct ip_tunnel_key *encap_key = &flow->actions.tun_encap_key;
+	struct bnxt_tc_info *tc_info = bp->tc_info;
+	struct bnxt_tc_tunnel_node *encap_node;
+	int rc;
+
+	/* Check if there's another flow using the same tunnel encap.
+	 * If not, add this tunnel to the table and resolve the other
+	 * tunnel header fileds
+	 */
+	encap_node = bnxt_tc_get_tunnel_node(bp, &tc_info->encap_table,
+					     &tc_info->encap_ht_params,
+					     encap_key);
+	if (!encap_node)
+		return -ENOMEM;
+
+	flow_node->encap_node = encap_node;
+
+	if (encap_node->tunnel_handle != INVALID_TUNNEL_HANDLE)
+		goto done;
+
+	rc = bnxt_tc_resolve_tunnel_hdrs(bp, encap_key, &encap_node->l2_info,
+					 flow->actions.dst_dev);
+	if (rc)
+		goto put_encap;
+
+	/* Allocate a new tunnel encap record */
+	rc = hwrm_cfa_encap_record_alloc(bp, encap_key, &encap_node->l2_info,
+					 &encap_node->tunnel_handle);
+	if (rc)
+		goto put_encap;
+
+done:
+	*encap_handle = encap_node->tunnel_handle;
+	return 0;
+
+put_encap:
+	bnxt_tc_put_tunnel_node(bp, &tc_info->encap_table,
+				&tc_info->encap_ht_params, encap_node);
+	return rc;
+}
+
+static void bnxt_tc_put_tunnel_handle(struct bnxt *bp,
+				      struct bnxt_tc_flow *flow,
+				      struct bnxt_tc_flow_node *flow_node)
+{
+	if (flow->actions.flags & BNXT_TC_ACTION_FLAG_TUNNEL_DECAP)
+		bnxt_tc_put_decap_handle(bp, flow_node);
+	else if (flow->actions.flags & BNXT_TC_ACTION_FLAG_TUNNEL_ENCAP)
+		bnxt_tc_put_encap_handle(bp, flow_node->encap_node);
+}
+
+static int bnxt_tc_get_tunnel_handle(struct bnxt *bp,
+				     struct bnxt_tc_flow *flow,
+				     struct bnxt_tc_flow_node *flow_node,
+				     __le32 *tunnel_handle)
+{
+	if (flow->actions.flags & BNXT_TC_ACTION_FLAG_TUNNEL_DECAP)
+		return bnxt_tc_get_decap_handle(bp, flow, flow_node,
+						tunnel_handle);
+	else if (flow->actions.flags & BNXT_TC_ACTION_FLAG_TUNNEL_ENCAP)
+		return bnxt_tc_get_encap_handle(bp, flow, flow_node,
+						tunnel_handle);
+	else
+		return 0;
+}
 static int __bnxt_tc_del_flow(struct bnxt *bp,
 			      struct bnxt_tc_flow_node *flow_node)
 {
-	struct bnxt_tc_info *tc_info = &bp->tc_info;
+	struct bnxt_tc_info *tc_info = bp->tc_info;
 	int rc;
 
 	/* send HWRM cmd to free the flow-id */
@@ -599,6 +1145,9 @@ static int __bnxt_tc_del_flow(struct bnxt *bp,
 
 	mutex_lock(&tc_info->lock);
 
+	/* release references to any tunnel encap/decap nodes */
+	bnxt_tc_put_tunnel_handle(bp, &flow_node->flow, flow_node);
+
 	/* release reference to l2 node */
 	bnxt_tc_put_l2_node(bp, flow_node);
 
@@ -631,8 +1180,9 @@ static int bnxt_tc_add_flow(struct bnxt *bp, u16 src_fid,
 			    struct tc_cls_flower_offload *tc_flow_cmd)
 {
 	struct bnxt_tc_flow_node *new_node, *old_node;
-	struct bnxt_tc_info *tc_info = &bp->tc_info;
+	struct bnxt_tc_info *tc_info = bp->tc_info;
 	struct bnxt_tc_flow *flow;
+	__le32 tunnel_handle = 0;
 	__le16 ref_flow_handle;
 	int rc;
 
@@ -670,12 +1220,19 @@ static int bnxt_tc_add_flow(struct bnxt *bp, u16 src_fid,
 	if (rc)
 		goto unlock;
 
+	/* If the flow involves tunnel encap/decap, get tunnel_handle */
+	rc = bnxt_tc_get_tunnel_handle(bp, flow, new_node, &tunnel_handle);
+	if (rc)
+		goto put_l2;
+
 	/* send HWRM cmd to alloc the flow */
 	rc = bnxt_hwrm_cfa_flow_alloc(bp, flow, ref_flow_handle,
-				      &new_node->flow_handle);
+				      tunnel_handle, &new_node->flow_handle);
 	if (rc)
-		goto put_l2;
+		goto put_tunnel;
 
+	flow->lastused = jiffies;
+	spin_lock_init(&flow->stats_lock);
 	/* add new flow to flow-table */
 	rc = rhashtable_insert_fast(&tc_info->flow_table, &new_node->node,
 				    tc_info->flow_ht_params);
@@ -687,12 +1244,14 @@ static int bnxt_tc_add_flow(struct bnxt *bp, u16 src_fid,
 
 hwrm_flow_free:
 	bnxt_hwrm_cfa_flow_free(bp, new_node->flow_handle);
+put_tunnel:
+	bnxt_tc_put_tunnel_handle(bp, flow, new_node);
 put_l2:
 	bnxt_tc_put_l2_node(bp, new_node);
 unlock:
 	mutex_unlock(&tc_info->lock);
 free_node:
-	kfree(new_node);
+	kfree_rcu(new_node, rcu);
 done:
 	netdev_err(bp->dev, "Error: %s: cookie=0x%lx error=%d",
 		   __func__, tc_flow_cmd->cookie, rc);
@@ -702,7 +1261,7 @@ done:
 static int bnxt_tc_del_flow(struct bnxt *bp,
 			    struct tc_cls_flower_offload *tc_flow_cmd)
 {
-	struct bnxt_tc_info *tc_info = &bp->tc_info;
+	struct bnxt_tc_info *tc_info = bp->tc_info;
 	struct bnxt_tc_flow_node *flow_node;
 
 	flow_node = rhashtable_lookup_fast(&tc_info->flow_table,
@@ -720,10 +1279,11 @@ static int bnxt_tc_del_flow(struct bnxt *bp,
 static int bnxt_tc_get_flow_stats(struct bnxt *bp,
 				  struct tc_cls_flower_offload *tc_flow_cmd)
 {
-	struct bnxt_tc_info *tc_info = &bp->tc_info;
+	struct bnxt_tc_flow_stats stats, *curr_stats, *prev_stats;
+	struct bnxt_tc_info *tc_info = bp->tc_info;
 	struct bnxt_tc_flow_node *flow_node;
-	struct bnxt_tc_flow_stats stats;
-	int rc;
+	struct bnxt_tc_flow *flow;
+	unsigned long lastused;
 
 	flow_node = rhashtable_lookup_fast(&tc_info->flow_table,
 					   &tc_flow_cmd->cookie,
@@ -734,15 +1294,183 @@ static int bnxt_tc_get_flow_stats(struct bnxt *bp,
 		return -1;
 	}
 
-	rc = bnxt_hwrm_cfa_flow_stats_get(bp, flow_node->flow_handle,
-					  &flow_node->flow, &stats);
+	flow = &flow_node->flow;
+	curr_stats = &flow->stats;
+	prev_stats = &flow->prev_stats;
+
+	spin_lock(&flow->stats_lock);
+	stats.packets = curr_stats->packets - prev_stats->packets;
+	stats.bytes = curr_stats->bytes - prev_stats->bytes;
+	*prev_stats = *curr_stats;
+	lastused = flow->lastused;
+	spin_unlock(&flow->stats_lock);
+
+	tcf_exts_stats_update(tc_flow_cmd->exts, stats.bytes, stats.packets,
+			      lastused);
+	return 0;
+}
+
+static int
+bnxt_hwrm_cfa_flow_stats_get(struct bnxt *bp, int num_flows,
+			     struct bnxt_tc_stats_batch stats_batch[])
+{
+	struct hwrm_cfa_flow_stats_output *resp = bp->hwrm_cmd_resp_addr;
+	struct hwrm_cfa_flow_stats_input req = { 0 };
+	__le16 *req_flow_handles = &req.flow_handle_0;
+	int rc, i;
+
+	bnxt_hwrm_cmd_hdr_init(bp, &req, HWRM_CFA_FLOW_STATS, -1, -1);
+	req.num_flows = cpu_to_le16(num_flows);
+	for (i = 0; i < num_flows; i++) {
+		struct bnxt_tc_flow_node *flow_node = stats_batch[i].flow_node;
+
+		req_flow_handles[i] = flow_node->flow_handle;
+	}
+
+	mutex_lock(&bp->hwrm_cmd_lock);
+	rc = _hwrm_send_message(bp, &req, sizeof(req), HWRM_CMD_TIMEOUT);
+	if (!rc) {
+		__le64 *resp_packets = &resp->packet_0;
+		__le64 *resp_bytes = &resp->byte_0;
+
+		for (i = 0; i < num_flows; i++) {
+			stats_batch[i].hw_stats.packets =
+						le64_to_cpu(resp_packets[i]);
+			stats_batch[i].hw_stats.bytes =
+						le64_to_cpu(resp_bytes[i]);
+		}
+	} else {
+		netdev_info(bp->dev, "error rc=%d", rc);
+	}
+
+	mutex_unlock(&bp->hwrm_cmd_lock);
+	return rc;
+}
+
+/* Add val to accum while handling a possible wraparound
+ * of val. Eventhough val is of type u64, its actual width
+ * is denoted by mask and will wrap-around beyond that width.
+ */
+static void accumulate_val(u64 *accum, u64 val, u64 mask)
+{
+#define low_bits(x, mask)		((x) & (mask))
+#define high_bits(x, mask)		((x) & ~(mask))
+	bool wrapped = val < low_bits(*accum, mask);
+
+	*accum = high_bits(*accum, mask) + val;
+	if (wrapped)
+		*accum += (mask + 1);
+}
+
+/* The HW counters' width is much less than 64bits.
+ * Handle possible wrap-around while updating the stat counters
+ */
+static void bnxt_flow_stats_accum(struct bnxt_tc_info *tc_info,
+				  struct bnxt_tc_flow_stats *acc_stats,
+				  struct bnxt_tc_flow_stats *hw_stats)
+{
+	accumulate_val(&acc_stats->bytes, hw_stats->bytes, tc_info->bytes_mask);
+	accumulate_val(&acc_stats->packets, hw_stats->packets,
+		       tc_info->packets_mask);
+}
+
+static int
+bnxt_tc_flow_stats_batch_update(struct bnxt *bp, int num_flows,
+				struct bnxt_tc_stats_batch stats_batch[])
+{
+	struct bnxt_tc_info *tc_info = bp->tc_info;
+	int rc, i;
+
+	rc = bnxt_hwrm_cfa_flow_stats_get(bp, num_flows, stats_batch);
 	if (rc)
 		return rc;
 
-	tcf_exts_stats_update(tc_flow_cmd->exts, stats.bytes, stats.packets, 0);
+	for (i = 0; i < num_flows; i++) {
+		struct bnxt_tc_flow_node *flow_node = stats_batch[i].flow_node;
+		struct bnxt_tc_flow *flow = &flow_node->flow;
+
+		spin_lock(&flow->stats_lock);
+		bnxt_flow_stats_accum(tc_info, &flow->stats,
+				      &stats_batch[i].hw_stats);
+		if (flow->stats.packets != flow->prev_stats.packets)
+			flow->lastused = jiffies;
+		spin_unlock(&flow->stats_lock);
+	}
+
 	return 0;
 }
 
+static int
+bnxt_tc_flow_stats_batch_prep(struct bnxt *bp,
+			      struct bnxt_tc_stats_batch stats_batch[],
+			      int *num_flows)
+{
+	struct bnxt_tc_info *tc_info = bp->tc_info;
+	struct rhashtable_iter *iter = &tc_info->iter;
+	void *flow_node;
+	int rc, i;
+
+	rc = rhashtable_walk_start(iter);
+	if (rc && rc != -EAGAIN) {
+		i = 0;
+		goto done;
+	}
+
+	rc = 0;
+	for (i = 0; i < BNXT_FLOW_STATS_BATCH_MAX; i++) {
+		flow_node = rhashtable_walk_next(iter);
+		if (IS_ERR(flow_node)) {
+			i = 0;
+			if (PTR_ERR(flow_node) == -EAGAIN) {
+				continue;
+			} else {
+				rc = PTR_ERR(flow_node);
+				goto done;
+			}
+		}
+
+		/* No more flows */
+		if (!flow_node)
+			goto done;
+
+		stats_batch[i].flow_node = flow_node;
+	}
+done:
+	rhashtable_walk_stop(iter);
+	*num_flows = i;
+	return rc;
+}
+
+void bnxt_tc_flow_stats_work(struct bnxt *bp)
+{
+	struct bnxt_tc_info *tc_info = bp->tc_info;
+	int num_flows, rc;
+
+	num_flows = atomic_read(&tc_info->flow_table.nelems);
+	if (!num_flows)
+		return;
+
+	rhashtable_walk_enter(&tc_info->flow_table, &tc_info->iter);
+
+	for (;;) {
+		rc = bnxt_tc_flow_stats_batch_prep(bp, tc_info->stats_batch,
+						   &num_flows);
+		if (rc) {
+			if (rc == -EAGAIN)
+				continue;
+			break;
+		}
+
+		if (!num_flows)
+			break;
+
+		bnxt_tc_flow_stats_batch_update(bp, num_flows,
+						tc_info->stats_batch);
+	}
+
+	rhashtable_walk_exit(&tc_info->iter);
+}
+
 int bnxt_tc_setup_flower(struct bnxt *bp, u16 src_fid,
 			 struct tc_cls_flower_offload *cls_flower)
 {
@@ -781,19 +1509,37 @@ static const struct rhashtable_params bnxt_tc_l2_ht_params = {
 	.automatic_shrinking = true
 };
 
+static const struct rhashtable_params bnxt_tc_decap_l2_ht_params = {
+	.head_offset = offsetof(struct bnxt_tc_l2_node, node),
+	.key_offset = offsetof(struct bnxt_tc_l2_node, key),
+	.key_len = BNXT_TC_L2_KEY_LEN,
+	.automatic_shrinking = true
+};
+
+static const struct rhashtable_params bnxt_tc_tunnel_ht_params = {
+	.head_offset = offsetof(struct bnxt_tc_tunnel_node, node),
+	.key_offset = offsetof(struct bnxt_tc_tunnel_node, key),
+	.key_len = sizeof(struct ip_tunnel_key),
+	.automatic_shrinking = true
+};
+
 /* convert counter width in bits to a mask */
 #define mask(width)		((u64)~0 >> (64 - (width)))
 
 int bnxt_init_tc(struct bnxt *bp)
 {
-	struct bnxt_tc_info *tc_info = &bp->tc_info;
+	struct bnxt_tc_info *tc_info;
 	int rc;
 
-	if (bp->hwrm_spec_code < 0x10800) {
+	if (bp->hwrm_spec_code < 0x10803) {
 		netdev_warn(bp->dev,
 			    "Firmware does not support TC flower offload.\n");
 		return -ENOTSUPP;
 	}
+
+	tc_info = kzalloc(sizeof(*tc_info), GFP_KERNEL);
+	if (!tc_info)
+		return -ENOMEM;
 	mutex_init(&tc_info->lock);
 
 	/* Counter widths are programmed by FW */
@@ -803,30 +1549,62 @@ int bnxt_init_tc(struct bnxt *bp)
 	tc_info->flow_ht_params = bnxt_tc_flow_ht_params;
 	rc = rhashtable_init(&tc_info->flow_table, &tc_info->flow_ht_params);
 	if (rc)
-		return rc;
+		goto free_tc_info;
 
 	tc_info->l2_ht_params = bnxt_tc_l2_ht_params;
 	rc = rhashtable_init(&tc_info->l2_table, &tc_info->l2_ht_params);
 	if (rc)
 		goto destroy_flow_table;
 
+	tc_info->decap_l2_ht_params = bnxt_tc_decap_l2_ht_params;
+	rc = rhashtable_init(&tc_info->decap_l2_table,
+			     &tc_info->decap_l2_ht_params);
+	if (rc)
+		goto destroy_l2_table;
+
+	tc_info->decap_ht_params = bnxt_tc_tunnel_ht_params;
+	rc = rhashtable_init(&tc_info->decap_table,
+			     &tc_info->decap_ht_params);
+	if (rc)
+		goto destroy_decap_l2_table;
+
+	tc_info->encap_ht_params = bnxt_tc_tunnel_ht_params;
+	rc = rhashtable_init(&tc_info->encap_table,
+			     &tc_info->encap_ht_params);
+	if (rc)
+		goto destroy_decap_table;
+
 	tc_info->enabled = true;
 	bp->dev->hw_features |= NETIF_F_HW_TC;
 	bp->dev->features |= NETIF_F_HW_TC;
+	bp->tc_info = tc_info;
 	return 0;
 
+destroy_decap_table:
+	rhashtable_destroy(&tc_info->decap_table);
+destroy_decap_l2_table:
+	rhashtable_destroy(&tc_info->decap_l2_table);
+destroy_l2_table:
+	rhashtable_destroy(&tc_info->l2_table);
 destroy_flow_table:
 	rhashtable_destroy(&tc_info->flow_table);
+free_tc_info:
+	kfree(tc_info);
 	return rc;
 }
 
 void bnxt_shutdown_tc(struct bnxt *bp)
 {
-	struct bnxt_tc_info *tc_info = &bp->tc_info;
+	struct bnxt_tc_info *tc_info = bp->tc_info;
 
-	if (!tc_info->enabled)
+	if (!bnxt_tc_flower_enabled(bp))
 		return;
 
 	rhashtable_destroy(&tc_info->flow_table);
 	rhashtable_destroy(&tc_info->l2_table);
+	rhashtable_destroy(&tc_info->decap_l2_table);
+	rhashtable_destroy(&tc_info->decap_table);
+	rhashtable_destroy(&tc_info->encap_table);
+	kfree(tc_info);
+	bp->tc_info = NULL;
 }

drivers/net/ethernet/broadcom/bnxt/bnxt_tc.h

@@ -12,6 +12,8 @@
 
 #ifdef CONFIG_BNXT_FLOWER_OFFLOAD
 
+#include <net/ip_tunnels.h>
+
 /* Structs used for storing the filter/actions of the TC cmd.
  */
 struct bnxt_tc_l2_key {
@@ -50,6 +52,13 @@ struct bnxt_tc_l4_key {
 	};
 };
 
+struct bnxt_tc_tunnel_key {
+	struct bnxt_tc_l2_key	l2;
+	struct bnxt_tc_l3_key	l3;
+	struct bnxt_tc_l4_key	l4;
+	__be32			id;
+};
+
 struct bnxt_tc_actions {
 	u32				flags;
 #define BNXT_TC_ACTION_FLAG_FWD			BIT(0)
@@ -57,16 +66,16 @@ struct bnxt_tc_actions {
 #define BNXT_TC_ACTION_FLAG_PUSH_VLAN		BIT(3)
 #define BNXT_TC_ACTION_FLAG_POP_VLAN		BIT(4)
 #define BNXT_TC_ACTION_FLAG_DROP		BIT(5)
+#define BNXT_TC_ACTION_FLAG_TUNNEL_ENCAP	BIT(6)
+#define BNXT_TC_ACTION_FLAG_TUNNEL_DECAP	BIT(7)
 
 	u16				dst_fid;
 	struct net_device		*dst_dev;
 	__be16				push_vlan_tpid;
 	__be16				push_vlan_tci;
-};
 
-struct bnxt_tc_flow_stats {
-	u64		packets;
-	u64		bytes;
+	/* tunnel encap */
+	struct ip_tunnel_key		tun_encap_key;
 };
 
 struct bnxt_tc_flow {
@@ -76,6 +85,16 @@ struct bnxt_tc_flow {
 #define BNXT_TC_FLOW_FLAGS_IPV6_ADDRS		BIT(3)
 #define BNXT_TC_FLOW_FLAGS_PORTS		BIT(4)
 #define BNXT_TC_FLOW_FLAGS_ICMP			BIT(5)
+#define BNXT_TC_FLOW_FLAGS_TUNL_ETH_ADDRS	BIT(6)
+#define BNXT_TC_FLOW_FLAGS_TUNL_IPV4_ADDRS	BIT(7)
+#define BNXT_TC_FLOW_FLAGS_TUNL_IPV6_ADDRS	BIT(8)
+#define BNXT_TC_FLOW_FLAGS_TUNL_PORTS		BIT(9)
+#define BNXT_TC_FLOW_FLAGS_TUNL_ID		BIT(10)
+#define BNXT_TC_FLOW_FLAGS_TUNNEL	(BNXT_TC_FLOW_FLAGS_TUNL_ETH_ADDRS | \
+					 BNXT_TC_FLOW_FLAGS_TUNL_IPV4_ADDRS | \
+					 BNXT_TC_FLOW_FLAGS_TUNL_IPV6_ADDRS |\
+					 BNXT_TC_FLOW_FLAGS_TUNL_PORTS |\
+					 BNXT_TC_FLOW_FLAGS_TUNL_ID)
 
 	/* flow applicable to pkts ingressing on this fid */
 	u16				src_fid;
@@ -85,6 +104,8 @@ struct bnxt_tc_flow {
 	struct bnxt_tc_l3_key		l3_mask;
 	struct bnxt_tc_l4_key		l4_key;
 	struct bnxt_tc_l4_key		l4_mask;
+	struct ip_tunnel_key		tun_key;
+	struct ip_tunnel_key		tun_mask;
 
 	struct bnxt_tc_actions		actions;
 
@@ -93,13 +114,39 @@ struct bnxt_tc_flow {
 	/* previous snap-shot of stats */
 	struct bnxt_tc_flow_stats	prev_stats;
 	unsigned long			lastused; /* jiffies */
+	/* for calculating delta from prev_stats and
+	 * updating prev_stats atomically.
+	 */
+	spinlock_t			stats_lock;
+};
+
+/* Tunnel encap/decap hash table
+ * This table is used to maintain a list of flows that use
+ * the same tunnel encap/decap params (ip_daddrs, vni, udp_dport)
+ * and the FW returned handle.
+ * A separate table is maintained for encap and decap
+ */
+struct bnxt_tc_tunnel_node {
+	struct ip_tunnel_key		key;
+	struct rhash_head		node;
+
+	/* tunnel l2 info */
+	struct bnxt_tc_l2_key		l2_info;
+
+#define	INVALID_TUNNEL_HANDLE		cpu_to_le32(0xffffffff)
+	/* tunnel handle returned by FW */
+	__le32				tunnel_handle;
+
+	u32				refcount;
+	struct rcu_head			rcu;
 };
 
 /* L2 hash table
- * This data-struct is used for L2-flow table.
- * The L2 part of a flow is stored in a hash table.
+ * The same data-struct is used for L2-flow table and L2-tunnel table.
+ * The L2 part of a flow or tunnel is stored in a hash table.
  * A flow that shares the same L2 key/mask with an
- * already existing flow must refer to it's flow handle.
+ * already existing flow/tunnel must refer to it's flow handle or
+ * decap_filter_id respectively.
  */
 struct bnxt_tc_l2_node {
 	/* hash key: first 16b of key */
@@ -110,7 +157,7 @@ struct bnxt_tc_l2_node {
 	/* a linked list of flows that share the same l2 key */
 	struct list_head	common_l2_flows;
 
-	/* number of flows sharing the l2 key */
+	/* number of flows/tunnels sharing the l2 key */
 	u16			refcount;
 
 	struct rcu_head		rcu;
@@ -130,6 +177,16 @@ struct bnxt_tc_flow_node {
 	/* for the shared_flows list maintained in l2_node */
 	struct list_head		l2_list_node;
 
+	/* tunnel encap related */
+	struct bnxt_tc_tunnel_node	*encap_node;
+
+	/* tunnel decap related */
+	struct bnxt_tc_tunnel_node	*decap_node;
+	/* L2 node in tunnel-l2 hashtable that shares flow's tunnel l2 key */
+	struct bnxt_tc_l2_node		*decap_l2_node;
+	/* for the shared_flows list maintained in tunnel decap l2_node */
+	struct list_head		decap_l2_list_node;
+
 	struct rcu_head			rcu;
 };
 
@@ -137,6 +194,12 @@ int bnxt_tc_setup_flower(struct bnxt *bp, u16 src_fid,
 			 struct tc_cls_flower_offload *cls_flower);
 int bnxt_init_tc(struct bnxt *bp);
 void bnxt_shutdown_tc(struct bnxt *bp);
+void bnxt_tc_flow_stats_work(struct bnxt *bp);
+
+static inline bool bnxt_tc_flower_enabled(struct bnxt *bp)
+{
+	return bp->tc_info && bp->tc_info->enabled;
+}
 
 #else /* CONFIG_BNXT_FLOWER_OFFLOAD */
 
@@ -154,5 +217,14 @@ static inline int bnxt_init_tc(struct bnxt *bp)
 static inline void bnxt_shutdown_tc(struct bnxt *bp)
 {
 }
+
+static inline void bnxt_tc_flow_stats_work(struct bnxt *bp)
+{
+}
+
+static inline bool bnxt_tc_flower_enabled(struct bnxt *bp)
+{
+	return false;
+}
 #endif /* CONFIG_BNXT_FLOWER_OFFLOAD */
 #endif /* BNXT_TC_H */

drivers/net/ethernet/broadcom/bnxt/bnxt_vfr.c

@@ -124,6 +124,9 @@ static int bnxt_vf_rep_setup_tc_block_cb(enum tc_setup_type type,
 	struct bnxt *bp = vf_rep->bp;
 	int vf_fid = bp->pf.vf[vf_rep->vf_idx].fw_fid;
 
+	if (!bnxt_tc_flower_enabled(vf_rep->bp))
+		return -EOPNOTSUPP;
+
 	switch (type) {
 	case TC_SETUP_CLSFLOWER:
 		return bnxt_tc_setup_flower(bp, vf_fid, type_data);