Merge branch 'bnxt_en-updates'

Michael Chan says:

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

This series includes updating the firmware interface, adding
methods to get and set VEPA/VEB bridge modes, some minor DCBX and ETS
refinements, and 3 patches from Sathya Perla to implement initial
VF representors for SRIOV switching.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/broadcom/Kconfig

@@ -193,6 +193,7 @@ config SYSTEMPORT
 config BNXT
 	tristate "Broadcom NetXtreme-C/E support"
 	depends on PCI
+	depends on MAY_USE_DEVLINK
 	select FW_LOADER
 	select LIBCRC32C
 	---help---

drivers/net/ethernet/broadcom/bnxt/Makefile

 obj-$(CONFIG_BNXT) += bnxt_en.o
 
-bnxt_en-y := bnxt.o bnxt_sriov.o bnxt_ethtool.o bnxt_dcb.o bnxt_ulp.o bnxt_xdp.o
+bnxt_en-y := bnxt.o bnxt_sriov.o bnxt_ethtool.o bnxt_dcb.o bnxt_ulp.o bnxt_xdp.o bnxt_vfr.o

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

@@ -33,6 +33,7 @@
 #include <linux/mii.h>
 #include <linux/if.h>
 #include <linux/if_vlan.h>
+#include <linux/if_bridge.h>
 #include <linux/rtc.h>
 #include <linux/bpf.h>
 #include <net/ip.h>
@@ -56,6 +57,7 @@
 #include "bnxt_ethtool.h"
 #include "bnxt_dcb.h"
 #include "bnxt_xdp.h"
+#include "bnxt_vfr.h"
 
 #define BNXT_TX_TIMEOUT		(5 * HZ)
 
@@ -243,6 +245,16 @@ const u16 bnxt_lhint_arr[] = {
 	TX_BD_FLAGS_LHINT_2048_AND_LARGER,
 };
 
+static u16 bnxt_xmit_get_cfa_action(struct sk_buff *skb)
+{
+	struct metadata_dst *md_dst = skb_metadata_dst(skb);
+
+	if (!md_dst || md_dst->type != METADATA_HW_PORT_MUX)
+		return 0;
+
+	return md_dst->u.port_info.port_id;
+}
+
 static netdev_tx_t bnxt_start_xmit(struct sk_buff *skb, struct net_device *dev)
 {
 	struct bnxt *bp = netdev_priv(dev);
@@ -287,7 +299,7 @@ static netdev_tx_t bnxt_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	tx_buf->nr_frags = last_frag;
 
 	vlan_tag_flags = 0;
-	cfa_action = 0;
+	cfa_action = bnxt_xmit_get_cfa_action(skb);
 	if (skb_vlan_tag_present(skb)) {
 		vlan_tag_flags = TX_BD_CFA_META_KEY_VLAN |
 				 skb_vlan_tag_get(skb);
@@ -322,7 +334,8 @@ static netdev_tx_t bnxt_start_xmit(struct sk_buff *skb, struct net_device *dev)
 			tx_push1->tx_bd_hsize_lflags = 0;
 
 		tx_push1->tx_bd_cfa_meta = cpu_to_le32(vlan_tag_flags);
-		tx_push1->tx_bd_cfa_action = cpu_to_le32(cfa_action);
+		tx_push1->tx_bd_cfa_action =
+			cpu_to_le32(cfa_action << TX_BD_CFA_ACTION_SHIFT);
 
 		end = pdata + length;
 		end = PTR_ALIGN(end, 8) - 1;
@@ -427,7 +440,8 @@ static netdev_tx_t bnxt_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	txbd->tx_bd_len_flags_type = cpu_to_le32(flags);
 
 	txbd1->tx_bd_cfa_meta = cpu_to_le32(vlan_tag_flags);
-	txbd1->tx_bd_cfa_action = cpu_to_le32(cfa_action);
+	txbd1->tx_bd_cfa_action =
+			cpu_to_le32(cfa_action << TX_BD_CFA_ACTION_SHIFT);
 	for (i = 0; i < last_frag; i++) {
 		skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
 
@@ -1032,7 +1046,10 @@ static void bnxt_tpa_start(struct bnxt *bp, struct bnxt_rx_ring_info *rxr,
 		bnxt_sched_reset(bp, rxr);
 		return;
 	}
-
+	/* Store cfa_code in tpa_info to use in tpa_end
+	 * completion processing.
+	 */
+	tpa_info->cfa_code = TPA_START_CFA_CODE(tpa_start1);
 	prod_rx_buf->data = tpa_info->data;
 	prod_rx_buf->data_ptr = tpa_info->data_ptr;
 
@@ -1267,6 +1284,17 @@ static inline struct sk_buff *bnxt_gro_skb(struct bnxt *bp,
 	return skb;
 }
 
+/* Given the cfa_code of a received packet determine which
+ * netdev (vf-rep or PF) the packet is destined to.
+ */
+static struct net_device *bnxt_get_pkt_dev(struct bnxt *bp, u16 cfa_code)
+{
+	struct net_device *dev = bnxt_get_vf_rep(bp, cfa_code);
+
+	/* if vf-rep dev is NULL, the must belongs to the PF */
+	return dev ? dev : bp->dev;
+}
+
 static inline struct sk_buff *bnxt_tpa_end(struct bnxt *bp,
 					   struct bnxt_napi *bnapi,
 					   u32 *raw_cons,
@@ -1360,7 +1388,9 @@ static inline struct sk_buff *bnxt_tpa_end(struct bnxt *bp,
 			return NULL;
 		}
 	}
-	skb->protocol = eth_type_trans(skb, bp->dev);
+
+	skb->protocol =
+		eth_type_trans(skb, bnxt_get_pkt_dev(bp, tpa_info->cfa_code));
 
 	if (tpa_info->hash_type != PKT_HASH_TYPE_NONE)
 		skb_set_hash(skb, tpa_info->rss_hash, tpa_info->hash_type);
@@ -1387,6 +1417,18 @@ static inline struct sk_buff *bnxt_tpa_end(struct bnxt *bp,
 	return skb;
 }
 
+static void bnxt_deliver_skb(struct bnxt *bp, struct bnxt_napi *bnapi,
+			     struct sk_buff *skb)
+{
+	if (skb->dev != bp->dev) {
+		/* this packet belongs to a vf-rep */
+		bnxt_vf_rep_rx(bp, skb);
+		return;
+	}
+	skb_record_rx_queue(skb, bnapi->index);
+	napi_gro_receive(&bnapi->napi, skb);
+}
+
 /* returns the following:
  * 1       - 1 packet successfully received
  * 0       - successful TPA_START, packet not completed yet
@@ -1403,7 +1445,7 @@ static int bnxt_rx_pkt(struct bnxt *bp, struct bnxt_napi *bnapi, u32 *raw_cons,
 	struct rx_cmp *rxcmp;
 	struct rx_cmp_ext *rxcmp1;
 	u32 tmp_raw_cons = *raw_cons;
-	u16 cons, prod, cp_cons = RING_CMP(tmp_raw_cons);
+	u16 cfa_code, cons, prod, cp_cons = RING_CMP(tmp_raw_cons);
 	struct bnxt_sw_rx_bd *rx_buf;
 	unsigned int len;
 	u8 *data_ptr, agg_bufs, cmp_type;
@@ -1445,8 +1487,7 @@ static int bnxt_rx_pkt(struct bnxt *bp, struct bnxt_napi *bnapi, u32 *raw_cons,
 
 		rc = -ENOMEM;
 		if (likely(skb)) {
-			skb_record_rx_queue(skb, bnapi->index);
-			napi_gro_receive(&bnapi->napi, skb);
+			bnxt_deliver_skb(bp, bnapi, skb);
 			rc = 1;
 		}
 		*event |= BNXT_RX_EVENT;
@@ -1535,7 +1576,8 @@ static int bnxt_rx_pkt(struct bnxt *bp, struct bnxt_napi *bnapi, u32 *raw_cons,
 		skb_set_hash(skb, le32_to_cpu(rxcmp->rx_cmp_rss_hash), type);
 	}
 
-	skb->protocol = eth_type_trans(skb, dev);
+	cfa_code = RX_CMP_CFA_CODE(rxcmp1);
+	skb->protocol = eth_type_trans(skb, bnxt_get_pkt_dev(bp, cfa_code));
 
 	if ((rxcmp1->rx_cmp_flags2 &
 	     cpu_to_le32(RX_CMP_FLAGS2_META_FORMAT_VLAN)) &&
@@ -1560,8 +1602,7 @@ static int bnxt_rx_pkt(struct bnxt *bp, struct bnxt_napi *bnapi, u32 *raw_cons,
 		}
 	}
 
-	skb_record_rx_queue(skb, bnapi->index);
-	napi_gro_receive(&bnapi->napi, skb);
+	bnxt_deliver_skb(bp, bnapi, skb);
 	rc = 1;
 
 next_rx:
@@ -4577,6 +4618,7 @@ static int bnxt_hwrm_func_qcfg(struct bnxt *bp)
 {
 	struct hwrm_func_qcfg_input req = {0};
 	struct hwrm_func_qcfg_output *resp = bp->hwrm_cmd_resp_addr;
+	u16 flags;
 	int rc;
 
 	bnxt_hwrm_cmd_hdr_init(bp, &req, HWRM_FUNC_QCFG, -1, -1);
@@ -4593,15 +4635,15 @@ static int bnxt_hwrm_func_qcfg(struct bnxt *bp)
 		vf->vlan = le16_to_cpu(resp->vlan) & VLAN_VID_MASK;
 	}
 #endif
-	if (BNXT_PF(bp)) {
-		u16 flags = le16_to_cpu(resp->flags);
-
-		if (flags & (FUNC_QCFG_RESP_FLAGS_FW_DCBX_AGENT_ENABLED |
-			     FUNC_QCFG_RESP_FLAGS_FW_LLDP_AGENT_ENABLED))
-			bp->flags |= BNXT_FLAG_FW_LLDP_AGENT;
-		if (flags & FUNC_QCFG_RESP_FLAGS_MULTI_HOST)
-			bp->flags |= BNXT_FLAG_MULTI_HOST;
+	flags = le16_to_cpu(resp->flags);
+	if (flags & (FUNC_QCFG_RESP_FLAGS_FW_DCBX_AGENT_ENABLED |
+		     FUNC_QCFG_RESP_FLAGS_FW_LLDP_AGENT_ENABLED)) {
+		bp->flags |= BNXT_FLAG_FW_LLDP_AGENT;
+		if (flags & FUNC_QCFG_RESP_FLAGS_FW_DCBX_AGENT_ENABLED)
+			bp->flags |= BNXT_FLAG_FW_DCBX_AGENT;
 	}
+	if (BNXT_PF(bp) && (flags & FUNC_QCFG_RESP_FLAGS_MULTI_HOST))
+		bp->flags |= BNXT_FLAG_MULTI_HOST;
 
 	switch (resp->port_partition_type) {
 	case FUNC_QCFG_RESP_PORT_PARTITION_TYPE_NPAR1_0:
@@ -4610,6 +4652,13 @@ static int bnxt_hwrm_func_qcfg(struct bnxt *bp)
 		bp->port_partition_type = resp->port_partition_type;
 		break;
 	}
+	if (bp->hwrm_spec_code < 0x10707 ||
+	    resp->evb_mode == FUNC_QCFG_RESP_EVB_MODE_VEB)
+		bp->br_mode = BRIDGE_MODE_VEB;
+	else if (resp->evb_mode == FUNC_QCFG_RESP_EVB_MODE_VEPA)
+		bp->br_mode = BRIDGE_MODE_VEPA;
+	else
+		bp->br_mode = BRIDGE_MODE_UNDEF;
 
 func_qcfg_exit:
 	mutex_unlock(&bp->hwrm_cmd_lock);
@@ -4911,6 +4960,26 @@ static void bnxt_hwrm_resource_free(struct bnxt *bp, bool close_path,
 	}
 }
 
+static int bnxt_hwrm_set_br_mode(struct bnxt *bp, u16 br_mode)
+{
+	struct hwrm_func_cfg_input req = {0};
+	int rc;
+
+	bnxt_hwrm_cmd_hdr_init(bp, &req, HWRM_FUNC_CFG, -1, -1);
+	req.fid = cpu_to_le16(0xffff);
+	req.enables = cpu_to_le32(FUNC_CFG_REQ_ENABLES_EVB_MODE);
+	if (br_mode == BRIDGE_MODE_VEB)
+		req.evb_mode = FUNC_CFG_REQ_EVB_MODE_VEB;
+	else if (br_mode == BRIDGE_MODE_VEPA)
+		req.evb_mode = FUNC_CFG_REQ_EVB_MODE_VEPA;
+	else
+		return -EINVAL;
+	rc = hwrm_send_message(bp, &req, sizeof(req), HWRM_CMD_TIMEOUT);
+	if (rc)
+		rc = -EIO;
+	return rc;
+}
+
 static int bnxt_setup_vnic(struct bnxt *bp, u16 vnic_id)
 {
 	struct bnxt_vnic_info *vnic = &bp->vnic_info[vnic_id];
@@ -5646,7 +5715,7 @@ static int bnxt_hwrm_phy_qcaps(struct bnxt *bp)
 	if (rc)
 		goto hwrm_phy_qcaps_exit;
 
-	if (resp->eee_supported & PORT_PHY_QCAPS_RESP_EEE_SUPPORTED) {
+	if (resp->flags & PORT_PHY_QCAPS_RESP_FLAGS_EEE_SUPPORTED) {
 		struct ethtool_eee *eee = &bp->eee;
 		u16 fw_speeds = le16_to_cpu(resp->supported_speeds_eee_mode);
 
@@ -5686,13 +5755,15 @@ static int bnxt_update_link(struct bnxt *bp, bool chng_link_state)
 
 	memcpy(&link_info->phy_qcfg_resp, resp, sizeof(*resp));
 	link_info->phy_link_status = resp->link;
-	link_info->duplex =  resp->duplex;
+	link_info->duplex = resp->duplex_cfg;
+	if (bp->hwrm_spec_code >= 0x10800)
+		link_info->duplex = resp->duplex_state;
 	link_info->pause = resp->pause;
 	link_info->auto_mode = resp->auto_mode;
 	link_info->auto_pause_setting = resp->auto_pause;
 	link_info->lp_pause = resp->link_partner_adv_pause;
 	link_info->force_pause_setting = resp->force_pause;
-	link_info->duplex_setting = resp->duplex;
+	link_info->duplex_setting = resp->duplex_cfg;
 	if (link_info->phy_link_status == BNXT_LINK_LINK)
 		link_info->link_speed = le16_to_cpu(resp->link_speed);
 	else
@@ -6214,6 +6285,9 @@ static int __bnxt_open_nic(struct bnxt *bp, bool irq_re_init, bool link_re_init)
 	/* Poll link status and check for SFP+ module status */
 	bnxt_get_port_module_status(bp);
 
+	/* VF-reps may need to be re-opened after the PF is re-opened */
+	if (BNXT_PF(bp))
+		bnxt_vf_reps_open(bp);
 	return 0;
 
 open_err:
@@ -6302,6 +6376,10 @@ int bnxt_close_nic(struct bnxt *bp, bool irq_re_init, bool link_re_init)
 		if (rc)
 			netdev_warn(bp->dev, "timeout waiting for SRIOV config operation to complete!\n");
 	}
+
+	/* Close the VF-reps before closing PF */
+	if (BNXT_PF(bp))
+		bnxt_vf_reps_close(bp);
 #endif
 	/* Change device state to avoid TX queue wake up's */
 	bnxt_tx_disable(bp);
@@ -6813,7 +6891,8 @@ static void bnxt_timer(unsigned long data)
 	if (atomic_read(&bp->intr_sem) != 0)
 		goto bnxt_restart_timer;
 
-	if (bp->link_info.link_up && (bp->flags & BNXT_FLAG_PORT_STATS)) {
+	if (bp->link_info.link_up && (bp->flags & BNXT_FLAG_PORT_STATS) &&
+	    bp->stats_coal_ticks) {
 		set_bit(BNXT_PERIODIC_STATS_SP_EVENT, &bp->sp_event);
 		schedule_work(&bp->sp_task);
 	}
@@ -7422,6 +7501,106 @@ static void bnxt_udp_tunnel_del(struct net_device *dev,
 	schedule_work(&bp->sp_task);
 }
 
+static int bnxt_bridge_getlink(struct sk_buff *skb, u32 pid, u32 seq,
+			       struct net_device *dev, u32 filter_mask,
+			       int nlflags)
+{
+	struct bnxt *bp = netdev_priv(dev);
+
+	return ndo_dflt_bridge_getlink(skb, pid, seq, dev, bp->br_mode, 0, 0,
+				       nlflags, filter_mask, NULL);
+}
+
+static int bnxt_bridge_setlink(struct net_device *dev, struct nlmsghdr *nlh,
+			       u16 flags)
+{
+	struct bnxt *bp = netdev_priv(dev);
+	struct nlattr *attr, *br_spec;
+	int rem, rc = 0;
+
+	if (bp->hwrm_spec_code < 0x10708 || !BNXT_SINGLE_PF(bp))
+		return -EOPNOTSUPP;
+
+	br_spec = nlmsg_find_attr(nlh, sizeof(struct ifinfomsg), IFLA_AF_SPEC);
+	if (!br_spec)
+		return -EINVAL;
+
+	nla_for_each_nested(attr, br_spec, rem) {
+		u16 mode;
+
+		if (nla_type(attr) != IFLA_BRIDGE_MODE)
+			continue;
+
+		if (nla_len(attr) < sizeof(mode))
+			return -EINVAL;
+
+		mode = nla_get_u16(attr);
+		if (mode == bp->br_mode)
+			break;
+
+		rc = bnxt_hwrm_set_br_mode(bp, mode);
+		if (!rc)
+			bp->br_mode = mode;
+		break;
+	}
+	return rc;
+}
+
+static int bnxt_get_phys_port_name(struct net_device *dev, char *buf,
+				   size_t len)
+{
+	struct bnxt *bp = netdev_priv(dev);
+	int rc;
+
+	/* The PF and it's VF-reps only support the switchdev framework */
+	if (!BNXT_PF(bp))
+		return -EOPNOTSUPP;
+
+	/* The switch-id that the pf belongs to is exported by
+	 * the switchdev ndo. This name is just to distinguish from the
+	 * vf-rep ports.
+	 */
+	rc = snprintf(buf, len, "pf%d", bp->pf.port_id);
+
+	if (rc >= len)
+		return -EOPNOTSUPP;
+	return 0;
+}
+
+int bnxt_port_attr_get(struct bnxt *bp, struct switchdev_attr *attr)
+{
+	if (bp->eswitch_mode != DEVLINK_ESWITCH_MODE_SWITCHDEV)
+		return -EOPNOTSUPP;
+
+	/* The PF and it's VF-reps only support the switchdev framework */
+	if (!BNXT_PF(bp))
+		return -EOPNOTSUPP;
+
+	switch (attr->id) {
+	case SWITCHDEV_ATTR_ID_PORT_PARENT_ID:
+		/* In SRIOV each PF-pool (PF + child VFs) serves as a
+		 * switching domain, the PF's perm mac-addr can be used
+		 * as the unique parent-id
+		 */
+		attr->u.ppid.id_len = ETH_ALEN;
+		ether_addr_copy(attr->u.ppid.id, bp->pf.mac_addr);
+		break;
+	default:
+		return -EOPNOTSUPP;
+	}
+	return 0;
+}
+
+static int bnxt_swdev_port_attr_get(struct net_device *dev,
+				    struct switchdev_attr *attr)
+{
+	return bnxt_port_attr_get(netdev_priv(dev), attr);
+}
+
+static const struct switchdev_ops bnxt_switchdev_ops = {
+	.switchdev_port_attr_get	= bnxt_swdev_port_attr_get
+};
+
 static const struct net_device_ops bnxt_netdev_ops = {
 	.ndo_open		= bnxt_open,
 	.ndo_start_xmit		= bnxt_start_xmit,
@@ -7453,6 +7632,9 @@ static const struct net_device_ops bnxt_netdev_ops = {
 	.ndo_udp_tunnel_add	= bnxt_udp_tunnel_add,
 	.ndo_udp_tunnel_del	= bnxt_udp_tunnel_del,
 	.ndo_xdp		= bnxt_xdp,
+	.ndo_bridge_getlink	= bnxt_bridge_getlink,
+	.ndo_bridge_setlink	= bnxt_bridge_setlink,
+	.ndo_get_phys_port_name = bnxt_get_phys_port_name
 };
 
 static void bnxt_remove_one(struct pci_dev *pdev)
@@ -7460,8 +7642,10 @@ static void bnxt_remove_one(struct pci_dev *pdev)
 	struct net_device *dev = pci_get_drvdata(pdev);
 	struct bnxt *bp = netdev_priv(dev);
 
-	if (BNXT_PF(bp))
+	if (BNXT_PF(bp)) {
 		bnxt_sriov_disable(bp);
+		bnxt_dl_unregister(bp);
+	}
 
 	pci_disable_pcie_error_reporting(pdev);
 	unregister_netdev(dev);
@@ -7710,6 +7894,7 @@ static int bnxt_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 	dev->netdev_ops = &bnxt_netdev_ops;
 	dev->watchdog_timeo = BNXT_TX_TIMEOUT;
 	dev->ethtool_ops = &bnxt_ethtool_ops;
+	dev->switchdev_ops = &bnxt_switchdev_ops;
 	pci_set_drvdata(pdev, dev);
 
 	rc = bnxt_alloc_hwrm_resources(bp);
@@ -7764,6 +7949,7 @@ static int bnxt_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 
 #ifdef CONFIG_BNXT_SRIOV
 	init_waitqueue_head(&bp->sriov_cfg_wait);
+	mutex_init(&bp->sriov_lock);
 #endif
 	bp->gro_func = bnxt_gro_func_5730x;
 	if (BNXT_CHIP_P4_PLUS(bp))
@@ -7855,6 +8041,9 @@ static int bnxt_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 	if (rc)
 		goto init_err_clr_int;
 
+	if (BNXT_PF(bp))
+		bnxt_dl_register(bp);
+
 	netdev_info(dev, "%s found at mem %lx, node addr %pM\n",
 		    board_info[ent->driver_data].name,
 		    (long)pci_resource_start(pdev, 0), dev->dev_addr);

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

@@ -12,13 +12,16 @@
 #define BNXT_H
 
 #define DRV_MODULE_NAME		"bnxt_en"
-#define DRV_MODULE_VERSION	"1.7.0"
+#define DRV_MODULE_VERSION	"1.8.0"
 
 #define DRV_VER_MAJ	1
-#define DRV_VER_MIN	7
+#define DRV_VER_MIN	8
 #define DRV_VER_UPD	0
 
 #include <linux/interrupt.h>
+#include <net/devlink.h>
+#include <net/dst_metadata.h>
+#include <net/switchdev.h>
 
 struct tx_bd {
 	__le32 tx_bd_len_flags_type;
@@ -242,6 +245,10 @@ struct rx_cmp_ext {
 	    ((le32_to_cpu((rxcmp1)->rx_cmp_flags2) &			\
 	     RX_CMP_FLAGS2_T_L4_CS_CALC) >> 3)
 
+#define RX_CMP_CFA_CODE(rxcmpl1)					\
+	((le32_to_cpu((rxcmpl1)->rx_cmp_cfa_code_errors_v2) &		\
+	  RX_CMPL_CFA_CODE_MASK) >> RX_CMPL_CFA_CODE_SFT)
+
 struct rx_agg_cmp {
 	__le32 rx_agg_cmp_len_flags_type;
 	#define RX_AGG_CMP_TYPE					(0x3f << 0)
@@ -311,6 +318,10 @@ struct rx_tpa_start_cmp_ext {
 	__le32 rx_tpa_start_cmp_hdr_info;
 };
 
+#define TPA_START_CFA_CODE(rx_tpa_start)				\
+	((le32_to_cpu((rx_tpa_start)->rx_tpa_start_cmp_cfa_code_v2) &	\
+	 RX_TPA_START_CMP_CFA_CODE) >> RX_TPA_START_CMPL_CFA_CODE_SHIFT)
+
 struct rx_tpa_end_cmp {
 	__le32 rx_tpa_end_cmp_len_flags_type;
 	#define RX_TPA_END_CMP_TYPE				(0x3f << 0)
@@ -618,6 +629,8 @@ struct bnxt_tpa_info {
 
 #define BNXT_TPA_OUTER_L3_OFF(hdr_info)	\
 	((hdr_info) & 0x1ff)
+
+	u16			cfa_code; /* cfa_code in TPA start compl */
 };
 
 struct bnxt_rx_ring_info {
@@ -825,8 +838,8 @@ struct bnxt_link_info {
 	u8			loop_back;
 	u8			link_up;
 	u8			duplex;
-#define BNXT_LINK_DUPLEX_HALF	PORT_PHY_QCFG_RESP_DUPLEX_HALF
-#define BNXT_LINK_DUPLEX_FULL	PORT_PHY_QCFG_RESP_DUPLEX_FULL
+#define BNXT_LINK_DUPLEX_HALF	PORT_PHY_QCFG_RESP_DUPLEX_STATE_HALF
+#define BNXT_LINK_DUPLEX_FULL	PORT_PHY_QCFG_RESP_DUPLEX_STATE_FULL
 	u8			pause;
 #define BNXT_LINK_PAUSE_TX	PORT_PHY_QCFG_RESP_PAUSE_TX
 #define BNXT_LINK_PAUSE_RX	PORT_PHY_QCFG_RESP_PAUSE_RX
@@ -928,6 +941,24 @@ struct bnxt_test_info {
 #define BNXT_CAG_REG_LEGACY_INT_STATUS	0x4014
 #define BNXT_CAG_REG_BASE		0x300000
 
+struct bnxt_vf_rep_stats {
+	u64			packets;
+	u64			bytes;
+	u64			dropped;
+};
+
+struct bnxt_vf_rep {
+	struct bnxt			*bp;
+	struct net_device		*dev;
+	struct metadata_dst		*dst;
+	u16				vf_idx;
+	u16				tx_cfa_action;
+	u16				rx_cfa_code;
+
+	struct bnxt_vf_rep_stats	rx_stats;
+	struct bnxt_vf_rep_stats	tx_stats;
+};
+
 struct bnxt {
 	void __iomem		*bar0;
 	void __iomem		*bar1;
@@ -1027,6 +1058,7 @@ struct bnxt {
 	#define BNXT_FLAG_MULTI_HOST	0x100000
 	#define BNXT_FLAG_SHORT_CMD	0x200000
 	#define BNXT_FLAG_DOUBLE_DB	0x400000
+	#define BNXT_FLAG_FW_DCBX_AGENT	0x800000
 	#define BNXT_FLAG_CHIP_NITRO_A0	0x1000000
 
 	#define BNXT_FLAG_ALL_CONFIG_FEATS (BNXT_FLAG_TPA |		\
@@ -1164,6 +1196,7 @@ struct bnxt {
 	u8			nge_port_cnt;
 	__le16			nge_fw_dst_port_id;
 	u8			port_partition_type;
+	u16			br_mode;
 
 	u16			rx_coal_ticks;
 	u16			rx_coal_ticks_irq;
@@ -1206,6 +1239,12 @@ struct bnxt {
 	wait_queue_head_t	sriov_cfg_wait;
 	bool			sriov_cfg;
 #define BNXT_SRIOV_CFG_WAIT_TMO	msecs_to_jiffies(10000)
+
+	/* lock to protect VF-rep creation/cleanup via
+	 * multiple paths such as ->sriov_configure() and
+	 * devlink ->eswitch_mode_set()
+	 */
+	struct mutex		sriov_lock;
 #endif
 
 #define BNXT_NTP_FLTR_MAX_FLTR	4096
@@ -1232,6 +1271,12 @@ struct bnxt {
 	struct bnxt_led_info	leds[BNXT_MAX_LED];
 
 	struct bpf_prog		*xdp_prog;
+
+	/* devlink interface and vf-rep structs */
+	struct devlink		*dl;
+	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 */
 };
 
 #define BNXT_RX_STATS_OFFSET(counter)			\
@@ -1306,4 +1351,5 @@ int bnxt_reserve_rings(struct bnxt *bp, int tx, int rx, bool sh, int tcs,
 int bnxt_setup_mq_tc(struct net_device *dev, u8 tc);
 int bnxt_get_max_rings(struct bnxt *, int *, int *, bool);
 void bnxt_restore_pf_fw_resources(struct bnxt *bp);
+int bnxt_port_attr_get(struct bnxt *bp, struct switchdev_attr *attr);
 #endif

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

@@ -93,6 +93,12 @@ static int bnxt_hwrm_queue_cos2bw_cfg(struct bnxt *bp, struct ieee_ets *ets,
 			cos2bw.tsa =
 				QUEUE_COS2BW_QCFG_RESP_QUEUE_ID0_TSA_ASSIGN_ETS;
 			cos2bw.bw_weight = ets->tc_tx_bw[i];
+			/* older firmware requires min_bw to be set to the
+			 * same weight value in percent.
+			 */
+			cos2bw.min_bw =
+				cpu_to_le32((ets->tc_tx_bw[i] * 100) |
+					    BW_VALUE_UNIT_PERCENT1_100);
 		}
 		memcpy(data, &cos2bw.queue_id, sizeof(cos2bw) - 4);
 		if (i == 0) {
@@ -549,13 +555,18 @@ static u8 bnxt_dcbnl_setdcbx(struct net_device *dev, u8 mode)
 {
 	struct bnxt *bp = netdev_priv(dev);
 
-	/* only support IEEE */
-	if ((mode & DCB_CAP_DCBX_VER_CEE) || !(mode & DCB_CAP_DCBX_VER_IEEE))
+	/* All firmware DCBX settings are set in NVRAM */
+	if (bp->dcbx_cap & DCB_CAP_DCBX_LLD_MANAGED)
 		return 1;
 
 	if (mode & DCB_CAP_DCBX_HOST) {
 		if (BNXT_VF(bp) || (bp->flags & BNXT_FLAG_FW_LLDP_AGENT))
 			return 1;
+
+		/* only support IEEE */
+		if ((mode & DCB_CAP_DCBX_VER_CEE) ||
+		    !(mode & DCB_CAP_DCBX_VER_IEEE))
+			return 1;
 	}
 
 	if (mode == bp->dcbx_cap)
@@ -584,7 +595,7 @@ void bnxt_dcb_init(struct bnxt *bp)
 	bp->dcbx_cap = DCB_CAP_DCBX_VER_IEEE;
 	if (BNXT_PF(bp) && !(bp->flags & BNXT_FLAG_FW_LLDP_AGENT))
 		bp->dcbx_cap |= DCB_CAP_DCBX_HOST;
-	else
+	else if (bp->flags & BNXT_FLAG_FW_DCBX_AGENT)
 		bp->dcbx_cap |= DCB_CAP_DCBX_LLD_MANAGED;
 	bp->dev->dcbnl_ops = &dcbnl_ops;
 }

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

@@ -26,6 +26,7 @@ struct bnxt_cos2bw_cfg {
 	u8			queue_id;
 	__le32			min_bw;
 	__le32			max_bw;
+#define BW_VALUE_UNIT_PERCENT1_100		(0x1UL << 29)
 	u8			tsa;
 	u8			pri_lvl;
 	u8			bw_weight;

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

@@ -86,9 +86,11 @@ static int bnxt_set_coalesce(struct net_device *dev,
 	if (bp->stats_coal_ticks != coal->stats_block_coalesce_usecs) {
 		u32 stats_ticks = coal->stats_block_coalesce_usecs;
 
-		stats_ticks = clamp_t(u32, stats_ticks,
-				      BNXT_MIN_STATS_COAL_TICKS,
-				      BNXT_MAX_STATS_COAL_TICKS);
+		/* Allow 0, which means disable. */
+		if (stats_ticks)
+			stats_ticks = clamp_t(u32, stats_ticks,
+					      BNXT_MIN_STATS_COAL_TICKS,
+					      BNXT_MAX_STATS_COAL_TICKS);
 		stats_ticks = rounddown(stats_ticks, BNXT_MIN_STATS_COAL_TICKS);
 		bp->stats_coal_ticks = stats_ticks;
 		update_stats = true;
@@ -198,19 +200,23 @@ static const struct {
 
 #define BNXT_NUM_PORT_STATS ARRAY_SIZE(bnxt_port_stats_arr)
 
+static int bnxt_get_num_stats(struct bnxt *bp)
+{
+	int num_stats = BNXT_NUM_STATS * bp->cp_nr_rings;
+
+	if (bp->flags & BNXT_FLAG_PORT_STATS)
+		num_stats += BNXT_NUM_PORT_STATS;
+
+	return num_stats;
+}
+
 static int bnxt_get_sset_count(struct net_device *dev, int sset)
 {
 	struct bnxt *bp = netdev_priv(dev);
 
 	switch (sset) {
-	case ETH_SS_STATS: {
-		int num_stats = BNXT_NUM_STATS * bp->cp_nr_rings;
-
-		if (bp->flags & BNXT_FLAG_PORT_STATS)
-			num_stats += BNXT_NUM_PORT_STATS;
-
-		return num_stats;
-	}
+	case ETH_SS_STATS:
+		return bnxt_get_num_stats(bp);
 	case ETH_SS_TEST:
 		if (!bp->num_tests)
 			return -EOPNOTSUPP;
@@ -225,11 +231,8 @@ static void bnxt_get_ethtool_stats(struct net_device *dev,
 {
 	u32 i, j = 0;
 	struct bnxt *bp = netdev_priv(dev);
-	u32 buf_size = sizeof(struct ctx_hw_stats) * bp->cp_nr_rings;
 	u32 stat_fields = sizeof(struct ctx_hw_stats) / 8;
 
-	memset(buf, 0, buf_size);
-
 	if (!bp->bnapi)
 		return;
 
@@ -835,7 +838,7 @@ static void bnxt_get_drvinfo(struct net_device *dev,
 		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_NUM_STATS * bp->cp_nr_rings;
+	info->n_stats = bnxt_get_num_stats(bp);
 	info->testinfo_len = bp->num_tests;
 	/* TODO CHIMP_FW: eeprom dump details */
 	info->eedump_len = 0;

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

@@ -11,14 +11,14 @@
 #ifndef BNXT_HSI_H
 #define BNXT_HSI_H
 
-/* HSI and HWRM Specification 1.7.6 */
+/* HSI and HWRM Specification 1.8.0 */
 #define HWRM_VERSION_MAJOR	1
-#define HWRM_VERSION_MINOR	7
-#define HWRM_VERSION_UPDATE	6
+#define HWRM_VERSION_MINOR	8
+#define HWRM_VERSION_UPDATE	0
 
-#define HWRM_VERSION_RSVD	2 /* non-zero means beta version */
+#define HWRM_VERSION_RSVD	0 /* non-zero means beta version */
 
-#define HWRM_VERSION_STR	"1.7.6.2"
+#define HWRM_VERSION_STR	"1.8.0.0"
 /*
  * 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.
@@ -813,7 +813,7 @@ struct hwrm_func_qcfg_output {
 	#define FUNC_QCFG_RESP_FLAGS_FW_DCBX_AGENT_ENABLED	    0x4UL
 	#define FUNC_QCFG_RESP_FLAGS_STD_TX_RING_MODE_ENABLED      0x8UL
 	#define FUNC_QCFG_RESP_FLAGS_FW_LLDP_AGENT_ENABLED	    0x10UL
-	#define FUNC_QCFG_RESP_FLAGS_MULTI_HOST			    0x20UL
+	#define FUNC_QCFG_RESP_FLAGS_MULTI_HOST		    0x20UL
 	u8 mac_address[6];
 	__le16 pci_id;
 	__le16 alloc_rsscos_ctx;
@@ -835,9 +835,8 @@ struct hwrm_func_qcfg_output {
 	u8 port_pf_cnt;
 	#define FUNC_QCFG_RESP_PORT_PF_CNT_UNAVAIL		   0x0UL
 	__le16 dflt_vnic_id;
-	u8 host_cnt;
-	#define FUNC_QCFG_RESP_HOST_CNT_UNAVAIL		   0x0UL
 	u8 unused_0;
+	u8 unused_1;
 	__le32 min_bw;
 	#define FUNC_QCFG_RESP_MIN_BW_BW_VALUE_MASK		    0xfffffffUL
 	#define FUNC_QCFG_RESP_MIN_BW_BW_VALUE_SFT		    0
@@ -874,12 +873,56 @@ 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_1;
+	u8 unused_2;
 	__le16 alloc_vfs;
 	__le32 alloc_mcast_filters;
 	__le32 alloc_hw_ring_grps;
 	__le16 alloc_sp_tx_rings;
+	u8 unused_3;
+	u8 valid;
+};
+
+/* hwrm_func_vlan_cfg */
+/* Input (48 bytes) */
+struct hwrm_func_vlan_cfg_input {
+	__le16 req_type;
+	__le16 cmpl_ring;
+	__le16 seq_id;
+	__le16 target_id;
+	__le64 resp_addr;
+	__le16 fid;
+	u8 unused_0;
+	u8 unused_1;
+	__le32 enables;
+	#define FUNC_VLAN_CFG_REQ_ENABLES_STAG_VID		    0x1UL
+	#define FUNC_VLAN_CFG_REQ_ENABLES_CTAG_VID		    0x2UL
+	#define FUNC_VLAN_CFG_REQ_ENABLES_STAG_PCP		    0x4UL
+	#define FUNC_VLAN_CFG_REQ_ENABLES_CTAG_PCP		    0x8UL
+	#define FUNC_VLAN_CFG_REQ_ENABLES_STAG_TPID		    0x10UL
+	#define FUNC_VLAN_CFG_REQ_ENABLES_CTAG_TPID		    0x20UL
+	__le16 stag_vid;
+	u8 stag_pcp;
 	u8 unused_2;
+	__be16 stag_tpid;
+	__le16 ctag_vid;
+	u8 ctag_pcp;
+	u8 unused_3;
+	__be16 ctag_tpid;
+	__le32 rsvd1;
+	__le32 rsvd2;
+	__le32 unused_4;
+};
+
+/* Output (16 bytes) */
+struct hwrm_func_vlan_cfg_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;
 };
 
@@ -902,6 +945,7 @@ struct hwrm_func_cfg_input {
 	#define FUNC_CFG_REQ_FLAGS_STD_TX_RING_MODE_ENABLE	    0x200UL
 	#define FUNC_CFG_REQ_FLAGS_STD_TX_RING_MODE_DISABLE	    0x400UL
 	#define FUNC_CFG_REQ_FLAGS_VIRT_MAC_PERSIST		    0x800UL
+	#define FUNC_CFG_REQ_FLAGS_NO_AUTOCLEAR_STATISTIC	    0x1000UL
 	__le32 enables;
 	#define FUNC_CFG_REQ_ENABLES_MTU			    0x1UL
 	#define FUNC_CFG_REQ_ENABLES_MRU			    0x2UL
@@ -1456,9 +1500,9 @@ struct hwrm_port_phy_qcfg_output {
 	#define PORT_PHY_QCFG_RESP_LINK_SPEED_50GB		   0x1f4UL
 	#define PORT_PHY_QCFG_RESP_LINK_SPEED_100GB		   0x3e8UL
 	#define PORT_PHY_QCFG_RESP_LINK_SPEED_10MB		   0xffffUL
-	u8 duplex;
-	#define PORT_PHY_QCFG_RESP_DUPLEX_HALF			   0x0UL
-	#define PORT_PHY_QCFG_RESP_DUPLEX_FULL			   0x1UL
+	u8 duplex_cfg;
+	#define PORT_PHY_QCFG_RESP_DUPLEX_CFG_HALF		   0x0UL
+	#define PORT_PHY_QCFG_RESP_DUPLEX_CFG_FULL		   0x1UL
 	u8 pause;
 	#define PORT_PHY_QCFG_RESP_PAUSE_TX			    0x1UL
 	#define PORT_PHY_QCFG_RESP_PAUSE_RX			    0x2UL
@@ -1573,6 +1617,9 @@ struct hwrm_port_phy_qcfg_output {
 	#define PORT_PHY_QCFG_RESP_PHY_TYPE_40G_BASELR4	   0x16UL
 	#define PORT_PHY_QCFG_RESP_PHY_TYPE_40G_BASEER4	   0x17UL
 	#define PORT_PHY_QCFG_RESP_PHY_TYPE_40G_ACTIVE_CABLE      0x18UL
+	#define PORT_PHY_QCFG_RESP_PHY_TYPE_1G_BASET		   0x19UL
+	#define PORT_PHY_QCFG_RESP_PHY_TYPE_1G_BASESX		   0x1aUL
+	#define PORT_PHY_QCFG_RESP_PHY_TYPE_1G_BASECX		   0x1bUL
 	u8 media_type;
 	#define PORT_PHY_QCFG_RESP_MEDIA_TYPE_UNKNOWN		   0x0UL
 	#define PORT_PHY_QCFG_RESP_MEDIA_TYPE_TP		   0x1UL
@@ -1651,14 +1698,16 @@ struct hwrm_port_phy_qcfg_output {
 	#define PORT_PHY_QCFG_RESP_FEC_CFG_FEC_CLAUSE74_ENABLED    0x10UL
 	#define PORT_PHY_QCFG_RESP_FEC_CFG_FEC_CLAUSE91_SUPPORTED  0x20UL
 	#define PORT_PHY_QCFG_RESP_FEC_CFG_FEC_CLAUSE91_ENABLED    0x40UL
+	u8 duplex_state;
+	#define PORT_PHY_QCFG_RESP_DUPLEX_STATE_HALF		   0x0UL
+	#define PORT_PHY_QCFG_RESP_DUPLEX_STATE_FULL		   0x1UL
 	u8 unused_1;
-	u8 unused_2;
 	char phy_vendor_name[16];
 	char phy_vendor_partnumber[16];
-	__le32 unused_3;
+	__le32 unused_2;
+	u8 unused_3;
 	u8 unused_4;
 	u8 unused_5;
-	u8 unused_6;
 	u8 valid;
 };
 
@@ -1744,6 +1793,51 @@ struct hwrm_port_mac_cfg_output {
 	u8 valid;
 };
 
+/* hwrm_port_mac_ptp_qcfg */
+/* Input (24 bytes) */
+struct hwrm_port_mac_ptp_qcfg_input {
+	__le16 req_type;
+	__le16 cmpl_ring;
+	__le16 seq_id;
+	__le16 target_id;
+	__le64 resp_addr;
+	__le16 port_id;
+	__le16 unused_0[3];
+};
+
+/* Output (80 bytes) */
+struct hwrm_port_mac_ptp_qcfg_output {
+	__le16 error_code;
+	__le16 req_type;
+	__le16 seq_id;
+	__le16 resp_len;
+	u8 flags;
+	#define PORT_MAC_PTP_QCFG_RESP_FLAGS_DIRECT_ACCESS	    0x1UL
+	#define PORT_MAC_PTP_QCFG_RESP_FLAGS_HWRM_ACCESS	    0x2UL
+	u8 unused_0;
+	__le16 unused_1;
+	__le32 rx_ts_reg_off_lower;
+	__le32 rx_ts_reg_off_upper;
+	__le32 rx_ts_reg_off_seq_id;
+	__le32 rx_ts_reg_off_src_id_0;
+	__le32 rx_ts_reg_off_src_id_1;
+	__le32 rx_ts_reg_off_src_id_2;
+	__le32 rx_ts_reg_off_domain_id;
+	__le32 rx_ts_reg_off_fifo;
+	__le32 rx_ts_reg_off_fifo_adv;
+	__le32 rx_ts_reg_off_granularity;
+	__le32 tx_ts_reg_off_lower;
+	__le32 tx_ts_reg_off_upper;
+	__le32 tx_ts_reg_off_seq_id;
+	__le32 tx_ts_reg_off_fifo;
+	__le32 tx_ts_reg_off_granularity;
+	__le32 unused_2;
+	u8 unused_3;
+	u8 unused_4;
+	u8 unused_5;
+	u8 valid;
+};
+
 /* hwrm_port_qstats */
 /* Input (40 bytes) */
 struct hwrm_port_qstats_input {
@@ -1874,10 +1968,10 @@ struct hwrm_port_phy_qcaps_output {
 	__le16 req_type;
 	__le16 seq_id;
 	__le16 resp_len;
-	u8 eee_supported;
-	#define PORT_PHY_QCAPS_RESP_EEE_SUPPORTED		    0x1UL
-	#define PORT_PHY_QCAPS_RESP_RSVD1_MASK			    0xfeUL
-	#define PORT_PHY_QCAPS_RESP_RSVD1_SFT			    1
+	u8 flags;
+	#define PORT_PHY_QCAPS_RESP_FLAGS_EEE_SUPPORTED	    0x1UL
+	#define PORT_PHY_QCAPS_RESP_FLAGS_RSVD1_MASK		    0xfeUL
+	#define PORT_PHY_QCAPS_RESP_FLAGS_RSVD1_SFT		    1
 	u8 unused_0;
 	__le16 supported_speeds_force_mode;
 	#define PORT_PHY_QCAPS_RESP_SUPPORTED_SPEEDS_FORCE_MODE_100MBHD 0x1UL
@@ -3152,6 +3246,95 @@ struct hwrm_queue_cos2bw_cfg_output {
 	u8 valid;
 };
 
+/* hwrm_queue_dscp_qcaps */
+/* Input (24 bytes) */
+struct hwrm_queue_dscp_qcaps_input {
+	__le16 req_type;
+	__le16 cmpl_ring;
+	__le16 seq_id;
+	__le16 target_id;
+	__le64 resp_addr;
+	u8 port_id;
+	u8 unused_0[7];
+};
+
+/* Output (16 bytes) */
+struct hwrm_queue_dscp_qcaps_output {
+	__le16 error_code;
+	__le16 req_type;
+	__le16 seq_id;
+	__le16 resp_len;
+	u8 num_dscp_bits;
+	u8 unused_0;
+	__le16 max_entries;
+	u8 unused_1;
+	u8 unused_2;
+	u8 unused_3;
+	u8 valid;
+};
+
+/* hwrm_queue_dscp2pri_qcfg */
+/* Input (32 bytes) */
+struct hwrm_queue_dscp2pri_qcfg_input {
+	__le16 req_type;
+	__le16 cmpl_ring;
+	__le16 seq_id;
+	__le16 target_id;
+	__le64 resp_addr;
+	__le64 dest_data_addr;
+	u8 port_id;
+	u8 unused_0;
+	__le16 dest_data_buffer_size;
+	__le32 unused_1;
+};
+
+/* Output (16 bytes) */
+struct hwrm_queue_dscp2pri_qcfg_output {
+	__le16 error_code;
+	__le16 req_type;
+	__le16 seq_id;
+	__le16 resp_len;
+	__le16 entry_cnt;
+	u8 default_pri;
+	u8 unused_0;
+	u8 unused_1;
+	u8 unused_2;
+	u8 unused_3;
+	u8 valid;
+};
+
+/* hwrm_queue_dscp2pri_cfg */
+/* Input (40 bytes) */
+struct hwrm_queue_dscp2pri_cfg_input {
+	__le16 req_type;
+	__le16 cmpl_ring;
+	__le16 seq_id;
+	__le16 target_id;
+	__le64 resp_addr;
+	__le64 src_data_addr;
+	__le32 flags;
+	#define QUEUE_DSCP2PRI_CFG_REQ_FLAGS_USE_HW_DEFAULT_PRI    0x1UL
+	__le32 enables;
+	#define QUEUE_DSCP2PRI_CFG_REQ_ENABLES_DEFAULT_PRI	    0x1UL
+	u8 port_id;
+	u8 default_pri;
+	__le16 entry_cnt;
+	__le32 unused_0;
+};
+
+/* Output (16 bytes) */
+struct hwrm_queue_dscp2pri_cfg_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_vnic_alloc */
 /* Input (24 bytes) */
 struct hwrm_vnic_alloc_input {
@@ -4038,7 +4221,7 @@ struct hwrm_cfa_encap_record_alloc_input {
 	#define CFA_ENCAP_RECORD_ALLOC_REQ_ENCAP_TYPE_IPGRE       0x8UL
 	u8 unused_0;
 	__le16 unused_1;
-	__le32 encap_data[16];
+	__le32 encap_data[20];
 };
 
 /* Output (16 bytes) */
@@ -4120,8 +4303,8 @@ struct hwrm_cfa_ntuple_filter_alloc_input {
 	#define CFA_NTUPLE_FILTER_ALLOC_REQ_IP_ADDR_TYPE_IPV6     0x6UL
 	u8 ip_protocol;
 	#define CFA_NTUPLE_FILTER_ALLOC_REQ_IP_PROTOCOL_UNKNOWN   0x0UL
-	#define CFA_NTUPLE_FILTER_ALLOC_REQ_IP_PROTOCOL_UDP       0x6UL
-	#define CFA_NTUPLE_FILTER_ALLOC_REQ_IP_PROTOCOL_TCP       0x11UL
+	#define CFA_NTUPLE_FILTER_ALLOC_REQ_IP_PROTOCOL_TCP       0x6UL
+	#define CFA_NTUPLE_FILTER_ALLOC_REQ_IP_PROTOCOL_UDP       0x11UL
 	__le16 dst_id;
 	__le16 mirror_vnic_id;
 	u8 tunnel_type;
@@ -4224,6 +4407,58 @@ struct hwrm_cfa_ntuple_filter_cfg_output {
 	u8 valid;
 };
 
+/* hwrm_cfa_vfr_alloc */
+/* Input (32 bytes) */
+struct hwrm_cfa_vfr_alloc_input {
+	__le16 req_type;
+	__le16 cmpl_ring;
+	__le16 seq_id;
+	__le16 target_id;
+	__le64 resp_addr;
+	__le16 vf_id;
+	__le16 reserved;
+	__le32 unused_0;
+	char vfr_name[32];
+};
+
+/* Output (16 bytes) */
+struct hwrm_cfa_vfr_alloc_output {
+	__le16 error_code;
+	__le16 req_type;
+	__le16 seq_id;
+	__le16 resp_len;
+	__le16 rx_cfa_code;
+	__le16 tx_cfa_action;
+	u8 unused_0;
+	u8 unused_1;
+	u8 unused_2;
+	u8 valid;
+};
+
+/* hwrm_cfa_vfr_free */
+/* Input (24 bytes) */
+struct hwrm_cfa_vfr_free_input {
+	__le16 req_type;
+	__le16 cmpl_ring;
+	__le16 seq_id;
+	__le16 target_id;
+	__le64 resp_addr;
+	char vfr_name[32];
+};
+
+/* Output (16 bytes) */
+struct hwrm_cfa_vfr_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_tunnel_dst_port_query */
 /* Input (24 bytes) */
 struct hwrm_tunnel_dst_port_query_input {
@@ -4448,12 +4683,13 @@ struct hwrm_fw_reset_input {
 	#define FW_RESET_REQ_EMBEDDED_PROC_TYPE_MGMT		   0x1UL
 	#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_RSVD		   0x4UL
+	#define FW_RESET_REQ_EMBEDDED_PROC_TYPE_HOST		   0x4UL
 	u8 selfrst_status;
 	#define FW_RESET_REQ_SELFRST_STATUS_SELFRSTNONE	   0x0UL
 	#define FW_RESET_REQ_SELFRST_STATUS_SELFRSTASAP	   0x1UL
 	#define FW_RESET_REQ_SELFRST_STATUS_SELFRSTPCIERST	   0x2UL
-	__le16 unused_0[3];
+	u8 host_idx;
+	u8 unused_0[5];
 };
 
 /* Output (16 bytes) */
@@ -4487,7 +4723,7 @@ struct hwrm_fw_qstatus_input {
 	#define FW_QSTATUS_REQ_EMBEDDED_PROC_TYPE_MGMT		   0x1UL
 	#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_RSVD		   0x4UL
+	#define FW_QSTATUS_REQ_EMBEDDED_PROC_TYPE_HOST		   0x4UL
 	u8 unused_0[7];
 };
 
@@ -4572,6 +4808,16 @@ struct hwrm_fw_set_structured_data_output {
 	u8 valid;
 };
 
+/* Command specific Error Codes (8 bytes) */
+struct hwrm_fw_set_structured_data_cmd_err {
+	u8 code;
+	#define FW_SET_STRUCTURED_DATA_CMD_ERR_CODE_UNKNOWN       0x0UL
+	#define FW_SET_STRUCTURED_DATA_CMD_ERR_CODE_BAD_HDR_CNT   0x1UL
+	#define FW_SET_STRUCTURED_DATA_CMD_ERR_CODE_BAD_FMT       0x2UL
+	#define FW_SET_STRUCTURED_DATA_CMD_ERR_CODE_BAD_ID	   0x3UL
+	u8 unused_0[7];
+};
+
 /* hwrm_fw_get_structured_data */
 /* Input (32 bytes) */
 struct hwrm_fw_get_structured_data_input {
@@ -4611,6 +4857,14 @@ struct hwrm_fw_get_structured_data_output {
 	u8 valid;
 };
 
+/* Command specific Error Codes (8 bytes) */
+struct hwrm_fw_get_structured_data_cmd_err {
+	u8 code;
+	#define FW_GET_STRUCTURED_DATA_CMD_ERR_CODE_UNKNOWN       0x0UL
+	#define FW_GET_STRUCTURED_DATA_CMD_ERR_CODE_BAD_ID	   0x3UL
+	u8 unused_0[7];
+};
+
 /* hwrm_exec_fwd_resp */
 /* Input (128 bytes) */
 struct hwrm_exec_fwd_resp_input {
@@ -5411,7 +5665,7 @@ struct cmd_nums {
 	#define HWRM_PORT_LPBK_CLR_STATS			   (0x26UL)
 	#define HWRM_PORT_PHY_QCFG				   (0x27UL)
 	#define HWRM_PORT_MAC_QCFG				   (0x28UL)
-	#define RESERVED7					   (0x29UL)
+	#define HWRM_PORT_MAC_PTP_QCFG				   (0x29UL)
 	#define HWRM_PORT_PHY_QCAPS				   (0x2aUL)
 	#define HWRM_PORT_PHY_I2C_WRITE			   (0x2bUL)
 	#define HWRM_PORT_PHY_I2C_READ				   (0x2cUL)
@@ -5421,14 +5675,17 @@ struct cmd_nums {
 	#define HWRM_QUEUE_QPORTCFG				   (0x30UL)
 	#define HWRM_QUEUE_QCFG				   (0x31UL)
 	#define HWRM_QUEUE_CFG					   (0x32UL)
-	#define RESERVED2					   (0x33UL)
-	#define RESERVED3					   (0x34UL)
+	#define HWRM_FUNC_VLAN_CFG				   (0x33UL)
+	#define HWRM_FUNC_VLAN_QCFG				   (0x34UL)
 	#define HWRM_QUEUE_PFCENABLE_QCFG			   (0x35UL)
 	#define HWRM_QUEUE_PFCENABLE_CFG			   (0x36UL)
 	#define HWRM_QUEUE_PRI2COS_QCFG			   (0x37UL)
 	#define HWRM_QUEUE_PRI2COS_CFG				   (0x38UL)
 	#define HWRM_QUEUE_COS2BW_QCFG				   (0x39UL)
 	#define HWRM_QUEUE_COS2BW_CFG				   (0x3aUL)
+	#define HWRM_QUEUE_DSCP_QCAPS				   (0x3bUL)
+	#define HWRM_QUEUE_DSCP2PRI_QCFG			   (0x3cUL)
+	#define HWRM_QUEUE_DSCP2PRI_CFG			   (0x3dUL)
 	#define HWRM_VNIC_ALLOC				   (0x40UL)
 	#define HWRM_VNIC_FREE					   (0x41UL)
 	#define HWRM_VNIC_CFG					   (0x42UL)
@@ -5455,7 +5712,7 @@ struct cmd_nums {
 	#define HWRM_CFA_L2_FILTER_FREE			   (0x91UL)
 	#define HWRM_CFA_L2_FILTER_CFG				   (0x92UL)
 	#define HWRM_CFA_L2_SET_RX_MASK			   (0x93UL)
-	#define RESERVED4					   (0x94UL)
+	#define HWRM_CFA_VLAN_ANTISPOOF_CFG			   (0x94UL)
 	#define HWRM_CFA_TUNNEL_FILTER_ALLOC			   (0x95UL)
 	#define HWRM_CFA_TUNNEL_FILTER_FREE			   (0x96UL)
 	#define HWRM_CFA_ENCAP_RECORD_ALLOC			   (0x97UL)
@@ -5494,6 +5751,8 @@ struct cmd_nums {
 	#define HWRM_CFA_METER_PROFILE_CFG			   (0xf7UL)
 	#define HWRM_CFA_METER_INSTANCE_ALLOC			   (0xf8UL)
 	#define HWRM_CFA_METER_INSTANCE_FREE			   (0xf9UL)
+	#define HWRM_CFA_VFR_ALLOC				   (0xfdUL)
+	#define HWRM_CFA_VFR_FREE				   (0xfeUL)
 	#define HWRM_CFA_VF_PAIR_ALLOC				   (0x100UL)
 	#define HWRM_CFA_VF_PAIR_FREE				   (0x101UL)
 	#define HWRM_CFA_VF_PAIR_INFO				   (0x102UL)
@@ -5502,6 +5761,9 @@ struct cmd_nums {
 	#define HWRM_CFA_FLOW_FLUSH				   (0x105UL)
 	#define HWRM_CFA_FLOW_STATS				   (0x106UL)
 	#define HWRM_CFA_FLOW_INFO				   (0x107UL)
+	#define HWRM_CFA_DECAP_FILTER_ALLOC			   (0x108UL)
+	#define HWRM_CFA_DECAP_FILTER_FREE			   (0x109UL)
+	#define HWRM_CFA_VLAN_ANTISPOOF_QCFG			   (0x10aUL)
 	#define HWRM_SELFTEST_QLIST				   (0x200UL)
 	#define HWRM_SELFTEST_EXEC				   (0x201UL)
 	#define HWRM_SELFTEST_IRQ				   (0x202UL)
@@ -5510,6 +5772,8 @@ struct cmd_nums {
 	#define HWRM_DBG_WRITE_DIRECT				   (0xff12UL)
 	#define HWRM_DBG_WRITE_INDIRECT			   (0xff13UL)
 	#define HWRM_DBG_DUMP					   (0xff14UL)
+	#define HWRM_DBG_ERASE_NVM				   (0xff15UL)
+	#define HWRM_DBG_CFG					   (0xff16UL)
 	#define HWRM_NVM_FACTORY_DEFAULTS			   (0xffeeUL)
 	#define HWRM_NVM_VALIDATE_OPTION			   (0xffefUL)
 	#define HWRM_NVM_FLUSH					   (0xfff0UL)

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

@@ -18,6 +18,7 @@
 #include "bnxt.h"
 #include "bnxt_ulp.h"
 #include "bnxt_sriov.h"
+#include "bnxt_vfr.h"
 #include "bnxt_ethtool.h"
 
 #ifdef CONFIG_BNXT_SRIOV
@@ -587,6 +588,10 @@ void bnxt_sriov_disable(struct bnxt *bp)
 	if (!num_vfs)
 		return;
 
+	/* synchronize VF and VF-rep create and destroy */
+	mutex_lock(&bp->sriov_lock);
+	bnxt_vf_reps_destroy(bp);
+
 	if (pci_vfs_assigned(bp->pdev)) {
 		bnxt_hwrm_fwd_async_event_cmpl(
 			bp, NULL, ASYNC_EVENT_CMPL_EVENT_ID_PF_DRVR_UNLOAD);
@@ -597,6 +602,7 @@ void bnxt_sriov_disable(struct bnxt *bp)
 		/* Free the HW resources reserved for various VF's */
 		bnxt_hwrm_func_vf_resource_free(bp, num_vfs);
 	}
+	mutex_unlock(&bp->sriov_lock);
 
 	bnxt_free_vf_resources(bp);
 
@@ -794,8 +800,10 @@ static int bnxt_vf_set_link(struct bnxt *bp, struct bnxt_vf_info *vf)
 					PORT_PHY_QCFG_RESP_LINK_LINK;
 				phy_qcfg_resp.link_speed = cpu_to_le16(
 					PORT_PHY_QCFG_RESP_LINK_SPEED_10GB);
-				phy_qcfg_resp.duplex =
-					PORT_PHY_QCFG_RESP_DUPLEX_FULL;
+				phy_qcfg_resp.duplex_cfg =
+					PORT_PHY_QCFG_RESP_DUPLEX_CFG_FULL;
+				phy_qcfg_resp.duplex_state =
+					PORT_PHY_QCFG_RESP_DUPLEX_STATE_FULL;
 				phy_qcfg_resp.pause =
 					(PORT_PHY_QCFG_RESP_PAUSE_TX |
 					 PORT_PHY_QCFG_RESP_PAUSE_RX);
@@ -804,7 +812,8 @@ static int bnxt_vf_set_link(struct bnxt *bp, struct bnxt_vf_info *vf)
 			/* force link down */
 			phy_qcfg_resp.link = PORT_PHY_QCFG_RESP_LINK_NO_LINK;
 			phy_qcfg_resp.link_speed = 0;
-			phy_qcfg_resp.duplex = PORT_PHY_QCFG_RESP_DUPLEX_HALF;
+			phy_qcfg_resp.duplex_state =
+				PORT_PHY_QCFG_RESP_DUPLEX_STATE_HALF;
 			phy_qcfg_resp.pause = 0;
 		}
 		rc = bnxt_hwrm_fwd_resp(bp, vf, &phy_qcfg_resp,

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

+/* Broadcom NetXtreme-C/E network driver.
+ *
+ * Copyright (c) 2016-2017 Broadcom Limited
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation.
+ */
+#include <linux/pci.h>
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/rtnetlink.h>
+#include <linux/jhash.h>
+
+#include "bnxt_hsi.h"
+#include "bnxt.h"
+#include "bnxt_vfr.h"
+
+#define CFA_HANDLE_INVALID		0xffff
+#define VF_IDX_INVALID			0xffff
+
+static int hwrm_cfa_vfr_alloc(struct bnxt *bp, u16 vf_idx,
+			      u16 *tx_cfa_action, u16 *rx_cfa_code)
+{
+	struct hwrm_cfa_vfr_alloc_output *resp = bp->hwrm_cmd_resp_addr;
+	struct hwrm_cfa_vfr_alloc_input req = { 0 };
+	int rc;
+
+	bnxt_hwrm_cmd_hdr_init(bp, &req, HWRM_CFA_VFR_ALLOC, -1, -1);
+	req.vf_id = cpu_to_le16(vf_idx);
+	sprintf(req.vfr_name, "vfr%d", vf_idx);
+
+	mutex_lock(&bp->hwrm_cmd_lock);
+	rc = _hwrm_send_message(bp, &req, sizeof(req), HWRM_CMD_TIMEOUT);
+	if (!rc) {
+		*tx_cfa_action = le16_to_cpu(resp->tx_cfa_action);
+		*rx_cfa_code = le16_to_cpu(resp->rx_cfa_code);
+		netdev_dbg(bp->dev, "tx_cfa_action=0x%x, rx_cfa_code=0x%x",
+			   *tx_cfa_action, *rx_cfa_code);
+	} else {
+		netdev_info(bp->dev, "%s error rc=%d", __func__, rc);
+	}
+
+	mutex_unlock(&bp->hwrm_cmd_lock);
+	return rc;
+}
+
+static int hwrm_cfa_vfr_free(struct bnxt *bp, u16 vf_idx)
+{
+	struct hwrm_cfa_vfr_free_input req = { 0 };
+	int rc;
+
+	bnxt_hwrm_cmd_hdr_init(bp, &req, HWRM_CFA_VFR_FREE, -1, -1);
+	sprintf(req.vfr_name, "vfr%d", vf_idx);
+
+	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_vf_rep_open(struct net_device *dev)
+{
+	struct bnxt_vf_rep *vf_rep = netdev_priv(dev);
+	struct bnxt *bp = vf_rep->bp;
+
+	/* Enable link and TX only if the parent PF is open. */
+	if (netif_running(bp->dev)) {
+		netif_carrier_on(dev);
+		netif_tx_start_all_queues(dev);
+	}
+	return 0;
+}
+
+static int bnxt_vf_rep_close(struct net_device *dev)
+{
+	netif_carrier_off(dev);
+	netif_tx_disable(dev);
+
+	return 0;
+}
+
+static netdev_tx_t bnxt_vf_rep_xmit(struct sk_buff *skb,
+				    struct net_device *dev)
+{
+	struct bnxt_vf_rep *vf_rep = netdev_priv(dev);
+	int rc, len = skb->len;
+
+	skb_dst_drop(skb);
+	dst_hold((struct dst_entry *)vf_rep->dst);
+	skb_dst_set(skb, (struct dst_entry *)vf_rep->dst);
+	skb->dev = vf_rep->dst->u.port_info.lower_dev;
+
+	rc = dev_queue_xmit(skb);
+	if (!rc) {
+		vf_rep->tx_stats.packets++;
+		vf_rep->tx_stats.bytes += len;
+	}
+	return rc;
+}
+
+static void
+bnxt_vf_rep_get_stats64(struct net_device *dev,
+			struct rtnl_link_stats64 *stats)
+{
+	struct bnxt_vf_rep *vf_rep = netdev_priv(dev);
+
+	stats->rx_packets = vf_rep->rx_stats.packets;
+	stats->rx_bytes = vf_rep->rx_stats.bytes;
+	stats->tx_packets = vf_rep->tx_stats.packets;
+	stats->tx_bytes = vf_rep->tx_stats.bytes;
+}
+
+struct net_device *bnxt_get_vf_rep(struct bnxt *bp, u16 cfa_code)
+{
+	u16 vf_idx;
+
+	if (cfa_code && bp->cfa_code_map && BNXT_PF(bp)) {
+		vf_idx = bp->cfa_code_map[cfa_code];
+		if (vf_idx != VF_IDX_INVALID)
+			return bp->vf_reps[vf_idx]->dev;
+	}
+	return NULL;
+}
+
+void bnxt_vf_rep_rx(struct bnxt *bp, struct sk_buff *skb)
+{
+	struct bnxt_vf_rep *vf_rep = netdev_priv(skb->dev);
+	struct bnxt_vf_rep_stats *rx_stats;
+
+	rx_stats = &vf_rep->rx_stats;
+	vf_rep->rx_stats.bytes += skb->len;
+	vf_rep->rx_stats.packets++;
+
+	netif_receive_skb(skb);
+}
+
+static int bnxt_vf_rep_get_phys_port_name(struct net_device *dev, char *buf,
+					  size_t len)
+{
+	struct bnxt_vf_rep *vf_rep = netdev_priv(dev);
+	int rc;
+
+	rc = snprintf(buf, len, "vfr%d", vf_rep->vf_idx);
+	if (rc >= len)
+		return -EOPNOTSUPP;
+	return 0;
+}
+
+static void bnxt_vf_rep_get_drvinfo(struct net_device *dev,
+				    struct ethtool_drvinfo *info)
+{
+	strlcpy(info->driver, DRV_MODULE_NAME, sizeof(info->driver));
+	strlcpy(info->version, DRV_MODULE_VERSION, sizeof(info->version));
+}
+
+static int bnxt_vf_rep_port_attr_get(struct net_device *dev,
+				     struct switchdev_attr *attr)
+{
+	struct bnxt_vf_rep *vf_rep = netdev_priv(dev);
+
+	/* as only PORT_PARENT_ID is supported currently use common code
+	 * between PF and VF-rep for now.
+	 */
+	return bnxt_port_attr_get(vf_rep->bp, attr);
+}
+
+static const struct switchdev_ops bnxt_vf_rep_switchdev_ops = {
+	.switchdev_port_attr_get	= bnxt_vf_rep_port_attr_get
+};
+
+static const struct ethtool_ops bnxt_vf_rep_ethtool_ops = {
+	.get_drvinfo		= bnxt_vf_rep_get_drvinfo
+};
+
+static const struct net_device_ops bnxt_vf_rep_netdev_ops = {
+	.ndo_open		= bnxt_vf_rep_open,
+	.ndo_stop		= bnxt_vf_rep_close,
+	.ndo_start_xmit		= bnxt_vf_rep_xmit,
+	.ndo_get_stats64	= bnxt_vf_rep_get_stats64,
+	.ndo_get_phys_port_name = bnxt_vf_rep_get_phys_port_name
+};
+
+/* Called when the parent PF interface is closed:
+ * As the mode transition from SWITCHDEV to LEGACY
+ * happens under the rtnl_lock() this routine is safe
+ * under the rtnl_lock()
+ */
+void bnxt_vf_reps_close(struct bnxt *bp)
+{
+	struct bnxt_vf_rep *vf_rep;
+	u16 num_vfs, i;
+
+	if (bp->eswitch_mode != DEVLINK_ESWITCH_MODE_SWITCHDEV)
+		return;
+
+	num_vfs = pci_num_vf(bp->pdev);
+	for (i = 0; i < num_vfs; i++) {
+		vf_rep = bp->vf_reps[i];
+		if (netif_running(vf_rep->dev))
+			bnxt_vf_rep_close(vf_rep->dev);
+	}
+}
+
+/* Called when the parent PF interface is opened (re-opened):
+ * As the mode transition from SWITCHDEV to LEGACY
+ * happen under the rtnl_lock() this routine is safe
+ * under the rtnl_lock()
+ */
+void bnxt_vf_reps_open(struct bnxt *bp)
+{
+	int i;
+
+	if (bp->eswitch_mode != DEVLINK_ESWITCH_MODE_SWITCHDEV)
+		return;
+
+	for (i = 0; i < pci_num_vf(bp->pdev); i++)
+		bnxt_vf_rep_open(bp->vf_reps[i]->dev);
+}
+
+static void __bnxt_vf_reps_destroy(struct bnxt *bp)
+{
+	u16 num_vfs = pci_num_vf(bp->pdev);
+	struct bnxt_vf_rep *vf_rep;
+	int i;
+
+	for (i = 0; i < num_vfs; i++) {
+		vf_rep = bp->vf_reps[i];
+		if (vf_rep) {
+			dst_release((struct dst_entry *)vf_rep->dst);
+
+			if (vf_rep->tx_cfa_action != CFA_HANDLE_INVALID)
+				hwrm_cfa_vfr_free(bp, vf_rep->vf_idx);
+
+			if (vf_rep->dev) {
+				/* if register_netdev failed, then netdev_ops
+				 * would have been set to NULL
+				 */
+				if (vf_rep->dev->netdev_ops)
+					unregister_netdev(vf_rep->dev);
+				free_netdev(vf_rep->dev);
+			}
+		}
+	}
+
+	kfree(bp->vf_reps);
+	bp->vf_reps = NULL;
+}
+
+void bnxt_vf_reps_destroy(struct bnxt *bp)
+{
+	bool closed = false;
+
+	if (bp->eswitch_mode != DEVLINK_ESWITCH_MODE_SWITCHDEV)
+		return;
+
+	if (!bp->vf_reps)
+		return;
+
+	/* Ensure that parent PF's and VF-reps' RX/TX has been quiesced
+	 * before proceeding with VF-rep cleanup.
+	 */
+	rtnl_lock();
+	if (netif_running(bp->dev)) {
+		bnxt_close_nic(bp, false, false);
+		closed = true;
+	}
+	/* un-publish cfa_code_map so that RX path can't see it anymore */
+	kfree(bp->cfa_code_map);
+	bp->cfa_code_map = NULL;
+	bp->eswitch_mode = DEVLINK_ESWITCH_MODE_LEGACY;
+
+	if (closed)
+		bnxt_open_nic(bp, false, false);
+	rtnl_unlock();
+
+	/* Need to call vf_reps_destroy() outside of rntl_lock
+	 * as unregister_netdev takes rtnl_lock
+	 */
+	__bnxt_vf_reps_destroy(bp);
+}
+
+/* Use the OUI of the PF's perm addr and report the same mac addr
+ * for the same VF-rep each time
+ */
+static void bnxt_vf_rep_eth_addr_gen(u8 *src_mac, u16 vf_idx, u8 *mac)
+{
+	u32 addr;
+
+	ether_addr_copy(mac, src_mac);
+
+	addr = jhash(src_mac, ETH_ALEN, 0) + vf_idx;
+	mac[3] = (u8)(addr & 0xFF);
+	mac[4] = (u8)((addr >> 8) & 0xFF);
+	mac[5] = (u8)((addr >> 16) & 0xFF);
+}
+
+static void bnxt_vf_rep_netdev_init(struct bnxt *bp, struct bnxt_vf_rep *vf_rep,
+				    struct net_device *dev)
+{
+	struct net_device *pf_dev = bp->dev;
+
+	dev->netdev_ops = &bnxt_vf_rep_netdev_ops;
+	dev->ethtool_ops = &bnxt_vf_rep_ethtool_ops;
+	dev->switchdev_ops = &bnxt_vf_rep_switchdev_ops;
+	/* Just inherit all the featues of the parent PF as the VF-R
+	 * uses the RX/TX rings of the parent PF
+	 */
+	dev->hw_features = pf_dev->hw_features;
+	dev->gso_partial_features = pf_dev->gso_partial_features;
+	dev->vlan_features = pf_dev->vlan_features;
+	dev->hw_enc_features = pf_dev->hw_enc_features;
+	dev->features |= pf_dev->features;
+	bnxt_vf_rep_eth_addr_gen(bp->pf.mac_addr, vf_rep->vf_idx,
+				 dev->perm_addr);
+	ether_addr_copy(dev->dev_addr, dev->perm_addr);
+}
+
+static int bnxt_vf_reps_create(struct bnxt *bp)
+{
+	u16 *cfa_code_map = NULL, num_vfs = pci_num_vf(bp->pdev);
+	struct bnxt_vf_rep *vf_rep;
+	struct net_device *dev;
+	int rc, i;
+
+	bp->vf_reps = kcalloc(num_vfs, sizeof(vf_rep), GFP_KERNEL);
+	if (!bp->vf_reps)
+		return -ENOMEM;
+
+	/* storage for cfa_code to vf-idx mapping */
+	cfa_code_map = kmalloc(sizeof(*bp->cfa_code_map) * MAX_CFA_CODE,
+			       GFP_KERNEL);
+	if (!cfa_code_map) {
+		rc = -ENOMEM;
+		goto err;
+	}
+	for (i = 0; i < MAX_CFA_CODE; i++)
+		cfa_code_map[i] = VF_IDX_INVALID;
+
+	for (i = 0; i < num_vfs; i++) {
+		dev = alloc_etherdev(sizeof(*vf_rep));
+		if (!dev) {
+			rc = -ENOMEM;
+			goto err;
+		}
+
+		vf_rep = netdev_priv(dev);
+		bp->vf_reps[i] = vf_rep;
+		vf_rep->dev = dev;
+		vf_rep->bp = bp;
+		vf_rep->vf_idx = i;
+		vf_rep->tx_cfa_action = CFA_HANDLE_INVALID;
+
+		/* get cfa handles from FW */
+		rc = hwrm_cfa_vfr_alloc(bp, vf_rep->vf_idx,
+					&vf_rep->tx_cfa_action,
+					&vf_rep->rx_cfa_code);
+		if (rc) {
+			rc = -ENOLINK;
+			goto err;
+		}
+		cfa_code_map[vf_rep->rx_cfa_code] = vf_rep->vf_idx;
+
+		vf_rep->dst = metadata_dst_alloc(0, METADATA_HW_PORT_MUX,
+						 GFP_KERNEL);
+		if (!vf_rep->dst) {
+			rc = -ENOMEM;
+			goto err;
+		}
+		/* only cfa_action is needed to mux a packet while TXing */
+		vf_rep->dst->u.port_info.port_id = vf_rep->tx_cfa_action;
+		vf_rep->dst->u.port_info.lower_dev = bp->dev;
+
+		bnxt_vf_rep_netdev_init(bp, vf_rep, dev);
+		rc = register_netdev(dev);
+		if (rc) {
+			/* no need for unregister_netdev in cleanup */
+			dev->netdev_ops = NULL;
+			goto err;
+		}
+	}
+
+	/* publish cfa_code_map only after all VF-reps have been initialized */
+	bp->cfa_code_map = cfa_code_map;
+	bp->eswitch_mode = DEVLINK_ESWITCH_MODE_SWITCHDEV;
+	netif_keep_dst(bp->dev);
+	return 0;
+
+err:
+	netdev_info(bp->dev, "%s error=%d", __func__, rc);
+	kfree(cfa_code_map);
+	__bnxt_vf_reps_destroy(bp);
+	return rc;
+}
+
+/* Devlink related routines */
+static int bnxt_dl_eswitch_mode_get(struct devlink *devlink, u16 *mode)
+{
+	struct bnxt *bp = bnxt_get_bp_from_dl(devlink);
+
+	*mode = bp->eswitch_mode;
+	return 0;
+}
+
+static int bnxt_dl_eswitch_mode_set(struct devlink *devlink, u16 mode)
+{
+	struct bnxt *bp = bnxt_get_bp_from_dl(devlink);
+	int rc = 0;
+
+	mutex_lock(&bp->sriov_lock);
+	if (bp->eswitch_mode == mode) {
+		netdev_info(bp->dev, "already in %s eswitch mode",
+			    mode == DEVLINK_ESWITCH_MODE_LEGACY ?
+			    "legacy" : "switchdev");
+		rc = -EINVAL;
+		goto done;
+	}
+
+	switch (mode) {
+	case DEVLINK_ESWITCH_MODE_LEGACY:
+		bnxt_vf_reps_destroy(bp);
+		break;
+
+	case DEVLINK_ESWITCH_MODE_SWITCHDEV:
+		if (pci_num_vf(bp->pdev) == 0) {
+			netdev_info(bp->dev,
+				    "Enable VFs before setting swtichdev mode");
+			rc = -EPERM;
+			goto done;
+		}
+		rc = bnxt_vf_reps_create(bp);
+		break;
+
+	default:
+		rc = -EINVAL;
+		goto done;
+	}
+done:
+	mutex_unlock(&bp->sriov_lock);
+	return rc;
+}
+
+static const struct devlink_ops bnxt_dl_ops = {
+	.eswitch_mode_set = bnxt_dl_eswitch_mode_set,
+	.eswitch_mode_get = bnxt_dl_eswitch_mode_get
+};
+
+int bnxt_dl_register(struct bnxt *bp)
+{
+	struct devlink *dl;
+	int rc;
+
+	if (!pci_find_ext_capability(bp->pdev, PCI_EXT_CAP_ID_SRIOV))
+		return 0;
+
+	if (bp->hwrm_spec_code < 0x10800) {
+		netdev_warn(bp->dev, "Firmware does not support SR-IOV E-Switch SWITCHDEV mode.\n");
+		return -ENOTSUPP;
+	}
+
+	dl = devlink_alloc(&bnxt_dl_ops, sizeof(struct bnxt_dl));
+	if (!dl) {
+		netdev_warn(bp->dev, "devlink_alloc failed");
+		return -ENOMEM;
+	}
+
+	bnxt_link_bp_to_dl(dl, bp);
+	bp->eswitch_mode = DEVLINK_ESWITCH_MODE_LEGACY;
+	rc = devlink_register(dl, &bp->pdev->dev);
+	if (rc) {
+		bnxt_link_bp_to_dl(dl, NULL);
+		devlink_free(dl);
+		netdev_warn(bp->dev, "devlink_register failed. rc=%d", rc);
+		return rc;
+	}
+
+	return 0;
+}
+
+void bnxt_dl_unregister(struct bnxt *bp)
+{
+	struct devlink *dl = bp->dl;
+
+	if (!dl)
+		return;
+
+	devlink_unregister(dl);
+	devlink_free(dl);
+}

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

+/* Broadcom NetXtreme-C/E network driver.
+ *
+ * Copyright (c) 2016-2017 Broadcom Limited
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation.
+ */
+
+#ifndef BNXT_VFR_H
+#define BNXT_VFR_H
+
+#define	MAX_CFA_CODE			65536
+
+/* Struct to hold housekeeping info needed by devlink interface */
+struct bnxt_dl {
+	struct bnxt *bp;	/* back ptr to the controlling dev */
+};
+
+static inline struct bnxt *bnxt_get_bp_from_dl(struct devlink *dl)
+{
+	return ((struct bnxt_dl *)devlink_priv(dl))->bp;
+}
+
+static inline void bnxt_link_bp_to_dl(struct devlink *dl, struct bnxt *bp)
+{
+	struct bnxt_dl *bp_dl = devlink_priv(dl);
+
+	bp_dl->bp = bp;
+	if (bp)
+		bp->dl = dl;
+}
+
+int bnxt_dl_register(struct bnxt *bp);
+void bnxt_dl_unregister(struct bnxt *bp);
+void bnxt_vf_reps_destroy(struct bnxt *bp);
+void bnxt_vf_reps_close(struct bnxt *bp);
+void bnxt_vf_reps_open(struct bnxt *bp);
+void bnxt_vf_rep_rx(struct bnxt *bp, struct sk_buff *skb);
+struct net_device *bnxt_get_vf_rep(struct bnxt *bp, u16 cfa_code);
+
+#endif /* BNXT_VFR_H */