[PATCH v3 02/26] net/bnxt: code reorg to properly allocate resources for PF/VF

[Ajit Khaparde <ajit.khaparde@broadcom.com>]

1) Move the function reset to bnxt_dev_init.
On the same lines, setup, enable and request interrupt to init path.
Memory allocation is also being done in the init path.
2) After a function reset, configure the VFs.  Distribute resources
evenly between all functions (PF and VF) for now. In the future, this
should be controllable.
3) The bnxt_vf_info and bnxt_pf_info had lot of duplication. Move the
common items to struct bnxt. And only unique items specific to PF remain
in the struct bnxt_pf_info.
4) Program the firmware to allow certain commands sent by a VF.
Disallowing these will prevent clean VF driver cleanup.
5) Since PF/VF need to allocate resources from a pool in the hardware,
use func_qcaps and func_qcfg to appropriately query the capabilities
and available resources.
6) If a PF is being initialized and no VFs are allocated, explicitly
call func_cfg to allocate the resources.
7) Once resources are requested from the firmware, update local copy
of resource count in struct bnxt only after sending the func_qcfg to
make sure the allocation request in the firmware went through.

The changes in this patch will be used by the subsequent patches
to allow proper initialization of PF/VF instance.

Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>

--
v1->v2: Regroup related patches per review comment
  Also split bigger patch into smaller patches if possible.

v2->v3:
  - rebasing to next-net tree
  - Use net/bnxt instead of just bnxt in patch subject
---
 drivers/net/bnxt/bnxt.h        |  80 ++++--
 drivers/net/bnxt/bnxt_ethdev.c | 155 ++++++++---
 drivers/net/bnxt/bnxt_filter.c |  30 +-
 drivers/net/bnxt/bnxt_hwrm.c   | 614 +++++++++++++++++++++++++++++++++++++----
 drivers/net/bnxt/bnxt_hwrm.h   |  17 +-
 drivers/net/bnxt/bnxt_vnic.c   |  45 +--
 drivers/net/bnxt/bnxt_vnic.h   |  11 +-
 7 files changed, 764 insertions(+), 188 deletions(-)

diff --git a/drivers/net/bnxt/bnxt.h b/drivers/net/bnxt/bnxt.h
index 4418c7f..0c59c66 100644
--- a/drivers/net/bnxt/bnxt.h
+++ b/drivers/net/bnxt/bnxt.h
@@ -35,6 +35,7 @@
 #define _BNXT_H_
 
 #include <inttypes.h>
+#include <stdbool.h>
 #include <sys/queue.h>
 
 #include <rte_ethdev.h>
@@ -54,17 +55,19 @@ enum bnxt_hw_context {
 	HW_CONTEXT_IS_LB    = 3,
 };
 
-struct bnxt_vf_info {
-	uint16_t		fw_fid;
-	uint8_t			mac_addr[ETHER_ADDR_LEN];
-	uint16_t		max_rsscos_ctx;
-	uint16_t		max_cp_rings;
-	uint16_t		max_tx_rings;
-	uint16_t		max_rx_rings;
-	uint16_t		max_l2_ctx;
-	uint16_t		max_vnics;
-	uint16_t		vlan;
-	struct bnxt_pf_info	*pf;
+struct bnxt_vlan_table_entry {
+	uint16_t		tpid;
+	uint16_t		vid;
+} __attribute__((packed));
+
+struct bnxt_child_vf_info {
+	void			*req_buf;
+	struct bnxt_vlan_table_entry	*vlan_table;
+	STAILQ_HEAD(, bnxt_filter_info)	filter;
+	uint32_t		func_cfg_flags;
+	uint32_t		l2_rx_mask;
+	uint16_t		fid;
+	bool			random_mac;
 };
 
 struct bnxt_pf_info {
@@ -73,22 +76,20 @@ struct bnxt_pf_info {
 #define BNXT_FIRST_VF_FID	128
 #define BNXT_PF_RINGS_USED(bp)	bnxt_get_num_queues(bp)
 #define BNXT_PF_RINGS_AVAIL(bp)	(bp->pf.max_cp_rings - BNXT_PF_RINGS_USED(bp))
-	uint32_t		fw_fid;
 	uint8_t			port_id;
-	uint8_t			mac_addr[ETHER_ADDR_LEN];
-	uint16_t		max_rsscos_ctx;
-	uint16_t		max_cp_rings;
-	uint16_t		max_tx_rings;
-	uint16_t		max_rx_rings;
-	uint16_t		max_l2_ctx;
-	uint16_t		max_vnics;
 	uint16_t		first_vf_id;
 	uint16_t		active_vfs;
 	uint16_t		max_vfs;
+	uint32_t		func_cfg_flags;
 	void			*vf_req_buf;
 	phys_addr_t		vf_req_buf_dma_addr;
 	uint32_t		vf_req_fwd[8];
-	struct bnxt_vf_info	*vf;
+	uint16_t		total_vnics;
+	struct bnxt_child_vf_info	*vf_info;
+#define BNXT_EVB_MODE_NONE	0
+#define BNXT_EVB_MODE_VEB	1
+#define BNXT_EVB_MODE_VEPA	2
+	uint8_t			evb_mode;
 };
 
 /* Max wait time is 10 * 100ms = 1s */
@@ -174,12 +175,49 @@ struct bnxt {
 	struct bnxt_link_info	link_info;
 	struct bnxt_cos_queue_info	cos_queue[BNXT_COS_QUEUE_COUNT];
 
+	uint16_t		fw_fid;
+	uint8_t			dflt_mac_addr[ETHER_ADDR_LEN];
+	uint16_t		max_rsscos_ctx;
+	uint16_t		max_cp_rings;
+	uint16_t		max_tx_rings;
+	uint16_t		max_rx_rings;
+	uint16_t		max_l2_ctx;
+	uint16_t		max_vnics;
+	uint16_t		max_stat_ctx;
+	uint16_t		vlan;
 	struct bnxt_pf_info		pf;
-	struct bnxt_vf_info		vf;
 	uint8_t			port_partition_type;
 	uint8_t			dev_stopped;
+	uint32_t		fw_ver;
+};
+
+/*
+ * Response sent back to the caller after callback
+ */
+enum rte_pmd_bnxt_mb_event_rsp {
+	RTE_PMD_BNXT_MB_EVENT_NOOP_ACK,  /**< skip mbox request and ACK */
+	RTE_PMD_BNXT_MB_EVENT_NOOP_NACK, /**< skip mbox request and NACK */
+	RTE_PMD_BNXT_MB_EVENT_PROCEED,  /**< proceed with mbox request  */
+	RTE_PMD_BNXT_MB_EVENT_MAX       /**< max value of this enum */
+};
+
+/* mailbox message types */
+#define BNXT_VF_RESET			0x01 /* VF requests reset */
+#define BNXT_VF_SET_MAC_ADDR	0x02 /* VF requests PF to set MAC addr */
+#define BNXT_VF_SET_VLAN		0x03 /* VF requests PF to set VLAN */
+#define BNXT_VF_SET_MTU			0x04 /* VF requests PF to set MTU */
+#define BNXT_VF_SET_MRU			0x05 /* VF requests PF to set MRU */
+
+/*
+ * Data sent to the caller when the callback is executed.
+ */
+struct rte_pmd_bnxt_mb_event_param {
+	uint16_t vf_id;	/* Virtual Function number */
+	int	retval;	/* return value */
+	void	*msg;	/* pointer to message */
 };
 
 int bnxt_link_update_op(struct rte_eth_dev *eth_dev, int wait_to_complete);
+int bnxt_rcv_msg_from_vf(struct bnxt *bp, uint16_t vf_id, void *msg);
 
 #endif
diff --git a/drivers/net/bnxt/bnxt_ethdev.c b/drivers/net/bnxt/bnxt_ethdev.c
index c230394..6e7e4f6 100644
--- a/drivers/net/bnxt/bnxt_ethdev.c
+++ b/drivers/net/bnxt/bnxt_ethdev.c
@@ -345,18 +345,12 @@ static void bnxt_dev_info_get_op(struct rte_eth_dev *eth_dev,
 	dev_info->max_hash_mac_addrs = 0;
 
 	/* PF/VF specifics */
-	if (BNXT_PF(bp)) {
-		dev_info->max_rx_queues = bp->pf.max_rx_rings;
-		dev_info->max_tx_queues = bp->pf.max_tx_rings;
-		dev_info->max_vfs = bp->pf.active_vfs;
-		dev_info->reta_size = bp->pf.max_rsscos_ctx;
-		max_vnics = bp->pf.max_vnics;
-	} else {
-		dev_info->max_rx_queues = bp->vf.max_rx_rings;
-		dev_info->max_tx_queues = bp->vf.max_tx_rings;
-		dev_info->reta_size = bp->vf.max_rsscos_ctx;
-		max_vnics = bp->vf.max_vnics;
-	}
+	if (BNXT_PF(bp))
+		dev_info->max_vfs = bp->pdev->max_vfs;
+	dev_info->max_rx_queues = bp->max_rx_rings;
+	dev_info->max_tx_queues = bp->max_tx_rings;
+	dev_info->reta_size = bp->max_rsscos_ctx;
+	max_vnics = bp->max_vnics;
 
 	/* Fast path specifics */
 	dev_info->min_rx_bufsize = 1;
@@ -488,41 +482,18 @@ static int bnxt_dev_start_op(struct rte_eth_dev *eth_dev)
 	int rc;
 
 	bp->dev_stopped = 0;
-	rc = bnxt_hwrm_func_reset(bp);
-	if (rc) {
-		RTE_LOG(ERR, PMD, "hwrm chip reset failure rc: %x\n", rc);
-		rc = -1;
-		goto error;
-	}
-
-	rc = bnxt_setup_int(bp);
-	if (rc)
-		goto error;
-
-	rc = bnxt_alloc_mem(bp);
-	if (rc)
-		goto error;
-
-	rc = bnxt_request_int(bp);
-	if (rc)
-		goto error;
 
 	rc = bnxt_init_nic(bp);
 	if (rc)
 		goto error;
 
-	bnxt_enable_int(bp);
-
 	bnxt_link_update_op(eth_dev, 0);
 	return 0;
 
 error:
 	bnxt_shutdown_nic(bp);
-	bnxt_disable_int(bp);
-	bnxt_free_int(bp);
 	bnxt_free_tx_mbufs(bp);
 	bnxt_free_rx_mbufs(bp);
-	bnxt_free_mem(bp);
 	return rc;
 }
 
@@ -554,8 +525,6 @@ static void bnxt_dev_stop_op(struct rte_eth_dev *eth_dev)
 		eth_dev->data->dev_link.link_status = 0;
 	}
 	bnxt_set_hwrm_link_config(bp, false);
-	bnxt_disable_int(bp);
-	bnxt_free_int(bp);
 	bnxt_shutdown_nic(bp);
 	bp->dev_stopped = 1;
 }
@@ -1082,6 +1051,12 @@ static int bnxt_init_board(struct rte_eth_dev *eth_dev)
 
 static int bnxt_dev_uninit(struct rte_eth_dev *eth_dev);
 
+#define ALLOW_FUNC(x)	\
+	{ \
+		typeof(x) arg = (x); \
+		bp->pf.vf_req_fwd[((arg) >> 5)] &= \
+		~rte_cpu_to_le_32(1 << ((arg) & 0x1f)); \
+	}
 static int
 bnxt_dev_init(struct rte_eth_dev *eth_dev)
 {
@@ -1097,6 +1072,7 @@ bnxt_dev_init(struct rte_eth_dev *eth_dev)
 	eth_dev->data->dev_flags |= RTE_ETH_DEV_DETACHABLE;
 
 	bp = eth_dev->data->dev_private;
+	bp->dev_stopped = 1;
 
 	if (bnxt_vf_pciid(pci_dev->id.device_id))
 		bp->flags |= BNXT_FLAG_VF;
@@ -1130,6 +1106,11 @@ bnxt_dev_init(struct rte_eth_dev *eth_dev)
 		RTE_LOG(ERR, PMD, "hwrm query capability failure rc: %x\n", rc);
 		goto error_free;
 	}
+	if (bp->max_tx_rings == 0) {
+		RTE_LOG(ERR, PMD, "No TX rings available!\n");
+		rc = -EBUSY;
+		goto error_free;
+	}
 	eth_dev->data->mac_addrs = rte_zmalloc("bnxt_mac_addr_tbl",
 					ETHER_ADDR_LEN * MAX_NUM_MAC_ADDR, 0);
 	if (eth_dev->data->mac_addrs == NULL) {
@@ -1140,10 +1121,7 @@ bnxt_dev_init(struct rte_eth_dev *eth_dev)
 		goto error_free;
 	}
 	/* Copy the permanent MAC from the qcap response address now. */
-	if (BNXT_PF(bp))
-		memcpy(bp->mac_addr, bp->pf.mac_addr, sizeof(bp->mac_addr));
-	else
-		memcpy(bp->mac_addr, bp->vf.mac_addr, sizeof(bp->mac_addr));
+	memcpy(bp->mac_addr, bp->dflt_mac_addr, sizeof(bp->mac_addr));
 	memcpy(&eth_dev->data->mac_addrs[0], bp->mac_addr, ETHER_ADDR_LEN);
 	bp->grp_info = rte_zmalloc("bnxt_grp_info",
 				sizeof(*bp->grp_info) * bp->max_ring_grps, 0);
@@ -1155,8 +1133,29 @@ bnxt_dev_init(struct rte_eth_dev *eth_dev)
 		goto error_free;
 	}
 
-	rc = bnxt_hwrm_func_driver_register(bp, 0,
-					    bp->pf.vf_req_fwd);
+	/* Forward all requests if firmware is new enough */
+	if (((bp->fw_ver >= ((20 << 24) | (6 << 16) | (100 << 8))) &&
+	    (bp->fw_ver < ((20 << 24) | (7 << 16)))) ||
+	    ((bp->fw_ver >= ((20 << 24) | (8 << 16))))) {
+		memset(bp->pf.vf_req_fwd, 0xff, sizeof(bp->pf.vf_req_fwd));
+	} else {
+		RTE_LOG(WARNING, PMD,
+			"Firmware too old for VF mailbox functionality\n");
+		memset(bp->pf.vf_req_fwd, 0, sizeof(bp->pf.vf_req_fwd));
+	}
+
+	/*
+	 * The following are used for driver cleanup.  If we disallow these,
+	 * VF drivers can't clean up cleanly.
+	 */
+	ALLOW_FUNC(HWRM_FUNC_DRV_UNRGTR);
+	ALLOW_FUNC(HWRM_VNIC_FREE);
+	ALLOW_FUNC(HWRM_RING_FREE);
+	ALLOW_FUNC(HWRM_RING_GRP_FREE);
+	ALLOW_FUNC(HWRM_VNIC_RSS_COS_LB_CTX_FREE);
+	ALLOW_FUNC(HWRM_CFA_L2_FILTER_FREE);
+	ALLOW_FUNC(HWRM_STAT_CTX_FREE);
+	rc = bnxt_hwrm_func_driver_register(bp);
 	if (rc) {
 		RTE_LOG(ERR, PMD,
 			"Failed to register driver");
@@ -1169,10 +1168,55 @@ bnxt_dev_init(struct rte_eth_dev *eth_dev)
 		pci_dev->mem_resource[0].phys_addr,
 		pci_dev->mem_resource[0].addr);
 
-	bp->dev_stopped = 0;
+	rc = bnxt_hwrm_func_reset(bp);
+	if (rc) {
+		RTE_LOG(ERR, PMD, "hwrm chip reset failure rc: %x\n", rc);
+		rc = -1;
+		goto error_free;
+	}
+
+	if (BNXT_PF(bp)) {
+		//if (bp->pf.active_vfs) {
+			// TODO: Deallocate VF resources?
+		//}
+		if (bp->pdev->max_vfs) {
+			rc = bnxt_hwrm_allocate_vfs(bp, bp->pdev->max_vfs);
+			if (rc) {
+				RTE_LOG(ERR, PMD, "Failed to allocate VFs\n");
+				goto error_free;
+			}
+		} else {
+			rc = bnxt_hwrm_allocate_pf_only(bp);
+			if (rc) {
+				RTE_LOG(ERR, PMD,
+					"Failed to allocate PF resources\n");
+				goto error_free;
+			}
+		}
+	}
+
+	rc = bnxt_setup_int(bp);
+	if (rc)
+		goto error_free;
+
+	rc = bnxt_alloc_mem(bp);
+	if (rc)
+		goto error_free_int;
+
+	rc = bnxt_request_int(bp);
+	if (rc)
+		goto error_free_int;
+
+	bnxt_enable_int(bp);
 
 	return 0;
 
+error_free_int:
+	bnxt_disable_int(bp);
+	bnxt_free_def_cp_ring(bp);
+	bnxt_hwrm_func_buf_unrgtr(bp);
+	bnxt_free_int(bp);
+	bnxt_free_mem(bp);
 error_free:
 	bnxt_dev_uninit(eth_dev);
 error:
@@ -1184,6 +1228,9 @@ bnxt_dev_uninit(struct rte_eth_dev *eth_dev) {
 	struct bnxt *bp = eth_dev->data->dev_private;
 	int rc;
 
+	bnxt_disable_int(bp);
+	bnxt_free_int(bp);
+	bnxt_free_mem(bp);
 	if (eth_dev->data->mac_addrs != NULL) {
 		rte_free(eth_dev->data->mac_addrs);
 		eth_dev->data->mac_addrs = NULL;
@@ -1196,6 +1243,8 @@ bnxt_dev_uninit(struct rte_eth_dev *eth_dev) {
 	bnxt_free_hwrm_resources(bp);
 	if (bp->dev_stopped == 0)
 		bnxt_dev_close_op(eth_dev);
+	if (bp->pf.vf_info)
+		rte_free(bp->pf.vf_info);
 	eth_dev->dev_ops = NULL;
 	eth_dev->rx_pkt_burst = NULL;
 	eth_dev->tx_pkt_burst = NULL;
@@ -1203,6 +1252,24 @@ bnxt_dev_uninit(struct rte_eth_dev *eth_dev) {
 	return rc;
 }
 
+int bnxt_rcv_msg_from_vf(struct bnxt *bp, uint16_t vf_id, void *msg)
+{
+	struct rte_pmd_bnxt_mb_event_param cb_param;
+
+	cb_param.retval = RTE_PMD_BNXT_MB_EVENT_PROCEED;
+	cb_param.vf_id = vf_id;
+	cb_param.msg = msg;
+
+	_rte_eth_dev_callback_process(bp->eth_dev, RTE_ETH_EVENT_VF_MBOX,
+			&cb_param);
+
+	/* Default to approve */
+	if (cb_param.retval == RTE_PMD_BNXT_MB_EVENT_PROCEED)
+		cb_param.retval = RTE_PMD_BNXT_MB_EVENT_NOOP_ACK;
+
+	return cb_param.retval == RTE_PMD_BNXT_MB_EVENT_NOOP_ACK ? true : false;
+}
+
 static int bnxt_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
 	struct rte_pci_device *pci_dev)
 {
diff --git a/drivers/net/bnxt/bnxt_filter.c b/drivers/net/bnxt/bnxt_filter.c
index df1042c..137c7b7 100644
--- a/drivers/net/bnxt/bnxt_filter.c
+++ b/drivers/net/bnxt/bnxt_filter.c
@@ -73,15 +73,7 @@ void bnxt_init_filters(struct bnxt *bp)
 	struct bnxt_filter_info *filter;
 	int i, max_filters;
 
-	if (BNXT_PF(bp)) {
-		struct bnxt_pf_info *pf = &bp->pf;
-
-		max_filters = pf->max_l2_ctx;
-	} else {
-		struct bnxt_vf_info *vf = &bp->vf;
-
-		max_filters = vf->max_l2_ctx;
-	}
+	max_filters = bp->max_l2_ctx;
 	STAILQ_INIT(&bp->free_filter_list);
 	for (i = 0; i < max_filters; i++) {
 		filter = &bp->filter_info[i];
@@ -122,15 +114,7 @@ void bnxt_free_filter_mem(struct bnxt *bp)
 		return;
 
 	/* Ensure that all filters are freed */
-	if (BNXT_PF(bp)) {
-		struct bnxt_pf_info *pf = &bp->pf;
-
-		max_filters = pf->max_l2_ctx;
-	} else {
-		struct bnxt_vf_info *vf = &bp->vf;
-
-		max_filters = vf->max_l2_ctx;
-	}
+	max_filters = bp->max_l2_ctx;
 	for (i = 0; i < max_filters; i++) {
 		filter = &bp->filter_info[i];
 		if (filter->fw_l2_filter_id != ((uint64_t)-1)) {
@@ -155,15 +139,7 @@ int bnxt_alloc_filter_mem(struct bnxt *bp)
 	struct bnxt_filter_info *filter_mem;
 	uint16_t max_filters;
 
-	if (BNXT_PF(bp)) {
-		struct bnxt_pf_info *pf = &bp->pf;
-
-		max_filters = pf->max_l2_ctx;
-	} else {
-		struct bnxt_vf_info *vf = &bp->vf;
-
-		max_filters = vf->max_l2_ctx;
-	}
+	max_filters = bp->max_l2_ctx;
 	/* Allocate memory for VNIC pool and filter pool */
 	filter_mem = rte_zmalloc("bnxt_filter_info",
 				 max_filters * sizeof(struct bnxt_filter_info),
diff --git a/drivers/net/bnxt/bnxt_hwrm.c b/drivers/net/bnxt/bnxt_hwrm.c
index 9100b7b..7aece5d 100644
--- a/drivers/net/bnxt/bnxt_hwrm.c
+++ b/drivers/net/bnxt/bnxt_hwrm.c
@@ -31,6 +31,8 @@
  *   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  */
 
+#include <unistd.h>
+
 #include <rte_byteorder.h>
 #include <rte_common.h>
 #include <rte_cycles.h>
@@ -54,6 +56,38 @@
 
 #define HWRM_CMD_TIMEOUT		2000
 
+struct bnxt_plcmodes_cfg {
+	uint32_t	flags;
+	uint16_t	jumbo_thresh;
+	uint16_t	hds_offset;
+	uint16_t	hds_threshold;
+};
+
+static int page_getenum(size_t size)
+{
+	if (size <= 1 << 4)
+		return 4;
+	if (size <= 1 << 12)
+		return 12;
+	if (size <= 1 << 13)
+		return 13;
+	if (size <= 1 << 16)
+		return 16;
+	if (size <= 1 << 21)
+		return 21;
+	if (size <= 1 << 22)
+		return 22;
+	if (size <= 1 << 30)
+		return 30;
+	RTE_LOG(ERR, PMD, "Page size %zu out of range\n", size);
+	return sizeof(void *) * 8 - 1;
+}
+
+static int page_roundup(size_t size)
+{
+	return 1 << page_getenum(size);
+}
+
 /*
  * HWRM Functions (sent to HWRM)
  * These are named bnxt_hwrm_*() and return -1 if bnxt_hwrm_send_message()
@@ -275,6 +309,8 @@ int bnxt_hwrm_func_qcaps(struct bnxt *bp)
 	int rc = 0;
 	struct hwrm_func_qcaps_input req = {.req_type = 0 };
 	struct hwrm_func_qcaps_output *resp = bp->hwrm_cmd_resp_addr;
+	uint16_t new_max_vfs;
+	int i;
 
 	HWRM_PREP(req, FUNC_QCAPS, -1, resp);
 
@@ -286,31 +322,52 @@ int bnxt_hwrm_func_qcaps(struct bnxt *bp)
 
 	bp->max_ring_grps = rte_le_to_cpu_32(resp->max_hw_ring_grps);
 	if (BNXT_PF(bp)) {
-		struct bnxt_pf_info *pf = &bp->pf;
-
-		pf->fw_fid = rte_le_to_cpu_32(resp->fid);
-		pf->port_id = resp->port_id;
-		memcpy(pf->mac_addr, resp->mac_address, ETHER_ADDR_LEN);
-		pf->max_rsscos_ctx = rte_le_to_cpu_16(resp->max_rsscos_ctx);
-		pf->max_cp_rings = rte_le_to_cpu_16(resp->max_cmpl_rings);
-		pf->max_tx_rings = rte_le_to_cpu_16(resp->max_tx_rings);
-		pf->max_rx_rings = rte_le_to_cpu_16(resp->max_rx_rings);
-		pf->max_l2_ctx = rte_le_to_cpu_16(resp->max_l2_ctxs);
-		pf->max_vnics = rte_le_to_cpu_16(resp->max_vnics);
-		pf->first_vf_id = rte_le_to_cpu_16(resp->first_vf_id);
-		pf->max_vfs = rte_le_to_cpu_16(resp->max_vfs);
+		bp->pf.port_id = resp->port_id;
+		bp->pf.first_vf_id = rte_le_to_cpu_16(resp->first_vf_id);
+		new_max_vfs = bp->pdev->max_vfs;
+		if (new_max_vfs != bp->pf.max_vfs) {
+			if (bp->pf.vf_info)
+				rte_free(bp->pf.vf_info);
+			bp->pf.vf_info = rte_malloc("bnxt_vf_info",
+			    sizeof(bp->pf.vf_info[0]) * new_max_vfs, 0);
+			bp->pf.max_vfs = new_max_vfs;
+			for (i = 0; i < new_max_vfs; i++) {
+				bp->pf.vf_info[i].fid = bp->pf.first_vf_id + i;
+				bp->pf.vf_info[i].vlan_table =
+					rte_zmalloc("VF VLAN table",
+						    getpagesize(),
+						    getpagesize());
+				if (bp->pf.vf_info[i].vlan_table == NULL)
+					RTE_LOG(ERR, PMD,
+					"Fail to alloc VLAN table for VF %d\n",
+					i);
+				else
+					rte_mem_lock_page(
+						bp->pf.vf_info[i].vlan_table);
+				STAILQ_INIT(&bp->pf.vf_info[i].filter);
+			}
+		}
+	}
+
+	bp->fw_fid = rte_le_to_cpu_32(resp->fid);
+	memcpy(bp->dflt_mac_addr, &resp->mac_address, ETHER_ADDR_LEN);
+	bp->max_rsscos_ctx = rte_le_to_cpu_16(resp->max_rsscos_ctx);
+	bp->max_cp_rings = rte_le_to_cpu_16(resp->max_cmpl_rings);
+	bp->max_tx_rings = rte_le_to_cpu_16(resp->max_tx_rings);
+	bp->max_rx_rings = rte_le_to_cpu_16(resp->max_rx_rings);
+	bp->max_l2_ctx = rte_le_to_cpu_16(resp->max_l2_ctxs);
+	/* TODO: For now, do not support VMDq/RFS on VFs. */
+	if (BNXT_PF(bp)) {
+		if (bp->pf.max_vfs)
+			bp->max_vnics = 1;
+		else
+			bp->max_vnics = rte_le_to_cpu_16(resp->max_vnics);
 	} else {
-		struct bnxt_vf_info *vf = &bp->vf;
-
-		vf->fw_fid = rte_le_to_cpu_32(resp->fid);
-		memcpy(vf->mac_addr, &resp->mac_address, ETHER_ADDR_LEN);
-		vf->max_rsscos_ctx = rte_le_to_cpu_16(resp->max_rsscos_ctx);
-		vf->max_cp_rings = rte_le_to_cpu_16(resp->max_cmpl_rings);
-		vf->max_tx_rings = rte_le_to_cpu_16(resp->max_tx_rings);
-		vf->max_rx_rings = rte_le_to_cpu_16(resp->max_rx_rings);
-		vf->max_l2_ctx = rte_le_to_cpu_16(resp->max_l2_ctxs);
-		vf->max_vnics = rte_le_to_cpu_16(resp->max_vnics);
+		bp->max_vnics = 1;
 	}
+	bp->max_stat_ctx = rte_le_to_cpu_16(resp->max_stat_ctx);
+	if (BNXT_PF(bp))
+		bp->pf.total_vnics = rte_le_to_cpu_16(resp->max_vnics);
 
 	return rc;
 }
@@ -332,8 +389,7 @@ int bnxt_hwrm_func_reset(struct bnxt *bp)
 	return rc;
 }
 
-int bnxt_hwrm_func_driver_register(struct bnxt *bp, uint32_t flags,
-				   uint32_t *vf_req_fwd)
+int bnxt_hwrm_func_driver_register(struct bnxt *bp)
 {
 	int rc;
 	struct hwrm_func_drv_rgtr_input req = {.req_type = 0 };
@@ -343,16 +399,22 @@ int bnxt_hwrm_func_driver_register(struct bnxt *bp, uint32_t flags,
 		return 0;
 
 	HWRM_PREP(req, FUNC_DRV_RGTR, -1, resp);
-	req.flags = flags;
-	req.enables = HWRM_FUNC_DRV_RGTR_INPUT_ENABLES_VER |
-			HWRM_FUNC_DRV_RGTR_INPUT_ENABLES_ASYNC_EVENT_FWD;
+	req.enables = rte_cpu_to_le_32(HWRM_FUNC_DRV_RGTR_INPUT_ENABLES_VER |
+			HWRM_FUNC_DRV_RGTR_INPUT_ENABLES_ASYNC_EVENT_FWD);
 	req.ver_maj = RTE_VER_YEAR;
 	req.ver_min = RTE_VER_MONTH;
 	req.ver_upd = RTE_VER_MINOR;
 
-	memcpy(req.vf_req_fwd, vf_req_fwd, sizeof(req.vf_req_fwd));
+	if (BNXT_PF(bp)) {
+		req.enables |= rte_cpu_to_le_32(
+			HWRM_FUNC_DRV_RGTR_INPUT_ENABLES_VF_INPUT_FWD);
+		memcpy(req.vf_req_fwd, bp->pf.vf_req_fwd,
+		       RTE_MIN(sizeof(req.vf_req_fwd),
+			       sizeof(bp->pf.vf_req_fwd)));
+	}
 
 	req.async_event_fwd[0] |= rte_cpu_to_le_32(0x1);   /* TODO: Use MACRO */
+	memset(req.async_event_fwd, 0xff, sizeof(req.async_event_fwd));
 
 	rc = bnxt_hwrm_send_message(bp, &req, sizeof(req));
 
@@ -391,6 +453,8 @@ int bnxt_hwrm_ver_get(struct bnxt *bp)
 		resp->hwrm_intf_maj, resp->hwrm_intf_min,
 		resp->hwrm_intf_upd,
 		resp->hwrm_fw_maj, resp->hwrm_fw_min, resp->hwrm_fw_bld);
+	bp->fw_ver = (resp->hwrm_fw_maj << 24) | (resp->hwrm_fw_min << 16) |
+			(resp->hwrm_fw_bld << 8) | resp->hwrm_fw_rsvd;
 	RTE_LOG(INFO, PMD, "Driver HWRM version: %d.%d.%d\n",
 		HWRM_VERSION_MAJOR, HWRM_VERSION_MINOR, HWRM_VERSION_UPDATE);
 
@@ -825,10 +889,12 @@ int bnxt_hwrm_vnic_alloc(struct bnxt *bp, struct bnxt_vnic_info *vnic)
 		}
 		vnic->fw_grp_ids[j] = bp->grp_info[i].fw_grp_id;
 	}
-
-	vnic->fw_rss_cos_lb_ctx = (uint16_t)HWRM_NA_SIGNATURE;
-	vnic->ctx_is_rss_cos_lb = HW_CONTEXT_NONE;
-
+	vnic->dflt_ring_grp = bp->grp_info[vnic->start_grp_id].fw_grp_id;
+	vnic->rss_rule = (uint16_t)HWRM_NA_SIGNATURE;
+	vnic->cos_rule = (uint16_t)HWRM_NA_SIGNATURE;
+	vnic->lb_rule = (uint16_t)HWRM_NA_SIGNATURE;
+	vnic->mru = bp->eth_dev->data->mtu + ETHER_HDR_LEN +
+				ETHER_CRC_LEN + VLAN_TAG_SIZE;
 	HWRM_PREP(req, VNIC_ALLOC, -1, resp);
 
 	rc = bnxt_hwrm_send_message(bp, &req, sizeof(req));
@@ -844,27 +910,45 @@ int bnxt_hwrm_vnic_cfg(struct bnxt *bp, struct bnxt_vnic_info *vnic)
 	int rc = 0;
 	struct hwrm_vnic_cfg_input req = {.req_type = 0 };
 	struct hwrm_vnic_cfg_output *resp = bp->hwrm_cmd_resp_addr;
+	uint32_t ctx_enable_flag = HWRM_VNIC_CFG_INPUT_ENABLES_RSS_RULE;
 
 	HWRM_PREP(req, VNIC_CFG, -1, resp);
 
 	/* Only RSS support for now TBD: COS & LB */
 	req.enables =
 	    rte_cpu_to_le_32(HWRM_VNIC_CFG_INPUT_ENABLES_DFLT_RING_GRP |
-			     HWRM_VNIC_CFG_INPUT_ENABLES_RSS_RULE |
 			     HWRM_VNIC_CFG_INPUT_ENABLES_MRU);
+	if (vnic->lb_rule != 0xffff)
+		ctx_enable_flag = HWRM_VNIC_CFG_INPUT_ENABLES_LB_RULE;
+	if (vnic->cos_rule != 0xffff)
+		ctx_enable_flag = HWRM_VNIC_CFG_INPUT_ENABLES_COS_RULE;
+	if (vnic->rss_rule != 0xffff)
+		ctx_enable_flag = HWRM_VNIC_CFG_INPUT_ENABLES_RSS_RULE;
+	req.enables |= rte_cpu_to_le_32(ctx_enable_flag);
 	req.vnic_id = rte_cpu_to_le_16(vnic->fw_vnic_id);
-	req.dflt_ring_grp =
-		rte_cpu_to_le_16(bp->grp_info[vnic->start_grp_id].fw_grp_id);
-	req.rss_rule = rte_cpu_to_le_16(vnic->fw_rss_cos_lb_ctx);
-	req.cos_rule = rte_cpu_to_le_16(0xffff);
-	req.lb_rule = rte_cpu_to_le_16(0xffff);
-	req.mru = rte_cpu_to_le_16(bp->eth_dev->data->mtu + ETHER_HDR_LEN +
-				   ETHER_CRC_LEN + VLAN_TAG_SIZE);
+	req.dflt_ring_grp = rte_cpu_to_le_16(vnic->dflt_ring_grp);
+	req.rss_rule = rte_cpu_to_le_16(vnic->rss_rule);
+	req.cos_rule = rte_cpu_to_le_16(vnic->cos_rule);
+	req.lb_rule = rte_cpu_to_le_16(vnic->lb_rule);
+	req.mru = rte_cpu_to_le_16(vnic->mru);
 	if (vnic->func_default)
-		req.flags = 1;
+		req.flags |=
+		    rte_cpu_to_le_32(HWRM_VNIC_CFG_INPUT_FLAGS_DEFAULT);
 	if (vnic->vlan_strip)
 		req.flags |=
 		    rte_cpu_to_le_32(HWRM_VNIC_CFG_INPUT_FLAGS_VLAN_STRIP_MODE);
+	if (vnic->bd_stall)
+		req.flags |=
+		    rte_cpu_to_le_32(HWRM_VNIC_CFG_INPUT_FLAGS_BD_STALL_MODE);
+	if (vnic->roce_dual)
+		req.flags |= rte_cpu_to_le_32(
+			HWRM_VNIC_QCFG_OUTPUT_FLAGS_ROCE_DUAL_VNIC_MODE);
+	if (vnic->roce_only)
+		req.flags |= rte_cpu_to_le_32(
+			HWRM_VNIC_QCFG_OUTPUT_FLAGS_ROCE_ONLY_VNIC_MODE);
+	if (vnic->rss_dflt_cr)
+		req.flags |= rte_cpu_to_le_32(
+			HWRM_VNIC_QCFG_OUTPUT_FLAGS_RSS_DFLT_CR_MODE);
 
 	rc = bnxt_hwrm_send_message(bp, &req, sizeof(req));
 
@@ -873,6 +957,45 @@ int bnxt_hwrm_vnic_cfg(struct bnxt *bp, struct bnxt_vnic_info *vnic)
 	return rc;
 }
 
+int bnxt_hwrm_vnic_qcfg(struct bnxt *bp, struct bnxt_vnic_info *vnic,
+		int16_t fw_vf_id)
+{
+	int rc = 0;
+	struct hwrm_vnic_qcfg_input req = {.req_type = 0 };
+	struct hwrm_vnic_qcfg_output *resp = bp->hwrm_cmd_resp_addr;
+
+	HWRM_PREP(req, VNIC_QCFG, -1, resp);
+
+	req.enables =
+		rte_cpu_to_le_32(HWRM_VNIC_QCFG_INPUT_ENABLES_VF_ID_VALID);
+	req.vnic_id = rte_cpu_to_le_16(vnic->fw_vnic_id);
+	req.vf_id = rte_cpu_to_le_16(fw_vf_id);
+
+	rc = bnxt_hwrm_send_message(bp, &req, sizeof(req));
+
+	HWRM_CHECK_RESULT;
+
+	vnic->dflt_ring_grp = rte_le_to_cpu_16(resp->dflt_ring_grp);
+	vnic->rss_rule = rte_le_to_cpu_16(resp->rss_rule);
+	vnic->cos_rule = rte_le_to_cpu_16(resp->cos_rule);
+	vnic->lb_rule = rte_le_to_cpu_16(resp->lb_rule);
+	vnic->mru = rte_le_to_cpu_16(resp->mru);
+	vnic->func_default = rte_le_to_cpu_32(
+			resp->flags) & HWRM_VNIC_QCFG_OUTPUT_FLAGS_DEFAULT;
+	vnic->vlan_strip = rte_le_to_cpu_32(resp->flags) &
+			HWRM_VNIC_QCFG_OUTPUT_FLAGS_VLAN_STRIP_MODE;
+	vnic->bd_stall = rte_le_to_cpu_32(resp->flags) &
+			HWRM_VNIC_QCFG_OUTPUT_FLAGS_BD_STALL_MODE;
+	vnic->roce_dual = rte_le_to_cpu_32(resp->flags) &
+			HWRM_VNIC_QCFG_OUTPUT_FLAGS_ROCE_DUAL_VNIC_MODE;
+	vnic->roce_only = rte_le_to_cpu_32(resp->flags) &
+			HWRM_VNIC_QCFG_OUTPUT_FLAGS_ROCE_ONLY_VNIC_MODE;
+	vnic->rss_dflt_cr = rte_le_to_cpu_32(resp->flags) &
+			HWRM_VNIC_QCFG_OUTPUT_FLAGS_RSS_DFLT_CR_MODE;
+
+	return rc;
+}
+
 int bnxt_hwrm_vnic_ctx_alloc(struct bnxt *bp, struct bnxt_vnic_info *vnic)
 {
 	int rc = 0;
@@ -886,7 +1009,7 @@ int bnxt_hwrm_vnic_ctx_alloc(struct bnxt *bp, struct bnxt_vnic_info *vnic)
 
 	HWRM_CHECK_RESULT;
 
-	vnic->fw_rss_cos_lb_ctx = rte_le_to_cpu_16(resp->rss_cos_lb_ctx_id);
+	vnic->rss_rule = rte_le_to_cpu_16(resp->rss_cos_lb_ctx_id);
 
 	return rc;
 }
@@ -900,13 +1023,13 @@ int bnxt_hwrm_vnic_ctx_free(struct bnxt *bp, struct bnxt_vnic_info *vnic)
 
 	HWRM_PREP(req, VNIC_RSS_COS_LB_CTX_FREE, -1, resp);
 
-	req.rss_cos_lb_ctx_id = rte_cpu_to_le_16(vnic->fw_rss_cos_lb_ctx);
+	req.rss_cos_lb_ctx_id = rte_cpu_to_le_16(vnic->rss_rule);
 
 	rc = bnxt_hwrm_send_message(bp, &req, sizeof(req));
 
 	HWRM_CHECK_RESULT;
 
-	vnic->fw_rss_cos_lb_ctx = INVALID_HW_RING_ID;
+	vnic->rss_rule = INVALID_HW_RING_ID;
 
 	return rc;
 }
@@ -947,7 +1070,7 @@ int bnxt_hwrm_vnic_rss_cfg(struct bnxt *bp,
 	    rte_cpu_to_le_64(vnic->rss_table_dma_addr);
 	req.hash_key_tbl_addr =
 	    rte_cpu_to_le_64(vnic->rss_hash_key_dma_addr);
-	req.rss_ctx_idx = rte_cpu_to_le_16(vnic->fw_rss_cos_lb_ctx);
+	req.rss_ctx_idx = rte_cpu_to_le_16(vnic->rss_rule);
 
 	rc = bnxt_hwrm_send_message(bp, &req, sizeof(req));
 
@@ -956,6 +1079,28 @@ int bnxt_hwrm_vnic_rss_cfg(struct bnxt *bp,
 	return rc;
 }
 
+int bnxt_hwrm_func_vf_mac(struct bnxt *bp, uint16_t vf, const uint8_t *mac_addr)
+{
+	struct hwrm_func_cfg_input req = {0};
+	struct hwrm_func_cfg_output *resp = bp->hwrm_cmd_resp_addr;
+	int rc;
+
+	req.flags = rte_cpu_to_le_32(bp->pf.vf_info[vf].func_cfg_flags);
+	req.enables = rte_cpu_to_le_32(
+			HWRM_FUNC_CFG_INPUT_ENABLES_DFLT_MAC_ADDR);
+	memcpy(req.dflt_mac_addr, mac_addr, sizeof(req.dflt_mac_addr));
+	req.fid = rte_cpu_to_le_16(bp->pf.vf_info[vf].fid);
+
+	HWRM_PREP(req, FUNC_CFG, -1, resp);
+
+	rc = bnxt_hwrm_send_message(bp, &req, sizeof(req));
+	HWRM_CHECK_RESULT;
+
+	bp->pf.vf_info[vf].random_mac = false;
+
+	return rc;
+}
+
 /*
  * HWRM utility functions
  */
@@ -1217,7 +1362,8 @@ void bnxt_free_all_hwrm_resources(struct bnxt *bp)
 		return;
 
 	vnic = &bp->vnic_info[0];
-	bnxt_hwrm_cfa_l2_clear_rx_mask(bp, vnic);
+	if (BNXT_PF(bp))
+		bnxt_hwrm_cfa_l2_clear_rx_mask(bp, vnic);
 
 	/* VNIC resources */
 	for (i = 0; i < bp->nr_vnics; i++) {
@@ -1512,12 +1658,8 @@ int bnxt_hwrm_func_qcfg(struct bnxt *bp)
 
 	HWRM_CHECK_RESULT;
 
-	if (BNXT_VF(bp)) {
-		struct bnxt_vf_info *vf = &bp->vf;
-
-		/* Hard Coded.. 0xfff VLAN ID mask */
-		vf->vlan = rte_le_to_cpu_16(resp->vlan) & 0xfff;
-	}
+	/* Hard Coded.. 0xfff VLAN ID mask */
+	bp->vlan = rte_le_to_cpu_16(resp->vlan) & 0xfff;
 
 	switch (resp->port_partition_type) {
 	case HWRM_FUNC_QCFG_OUTPUT_PORT_PARTITION_TYPE_NPAR1_0:
@@ -1532,3 +1674,369 @@ int bnxt_hwrm_func_qcfg(struct bnxt *bp)
 
 	return rc;
 }
+
+static void copy_func_cfg_to_qcaps(struct hwrm_func_cfg_input *fcfg,
+				   struct hwrm_func_qcaps_output *qcaps)
+{
+	qcaps->max_rsscos_ctx = fcfg->num_rsscos_ctxs;
+	memcpy(qcaps->mac_address, fcfg->dflt_mac_addr,
+	       sizeof(qcaps->mac_address));
+	qcaps->max_l2_ctxs = fcfg->num_l2_ctxs;
+	qcaps->max_rx_rings = fcfg->num_rx_rings;
+	qcaps->max_tx_rings = fcfg->num_tx_rings;
+	qcaps->max_cmpl_rings = fcfg->num_cmpl_rings;
+	qcaps->max_stat_ctx = fcfg->num_stat_ctxs;
+	qcaps->max_vfs = 0;
+	qcaps->first_vf_id = 0;
+	qcaps->max_vnics = fcfg->num_vnics;
+	qcaps->max_decap_records = 0;
+	qcaps->max_encap_records = 0;
+	qcaps->max_tx_wm_flows = 0;
+	qcaps->max_tx_em_flows = 0;
+	qcaps->max_rx_wm_flows = 0;
+	qcaps->max_rx_em_flows = 0;
+	qcaps->max_flow_id = 0;
+	qcaps->max_mcast_filters = fcfg->num_mcast_filters;
+	qcaps->max_sp_tx_rings = 0;
+	qcaps->max_hw_ring_grps = fcfg->num_hw_ring_grps;
+}
+
+static int bnxt_hwrm_pf_func_cfg(struct bnxt *bp, int tx_rings)
+{
+	struct hwrm_func_cfg_input req = {0};
+	struct hwrm_func_cfg_output *resp = bp->hwrm_cmd_resp_addr;
+	int rc;
+
+	req.enables = rte_cpu_to_le_32(HWRM_FUNC_CFG_INPUT_ENABLES_MTU |
+			HWRM_FUNC_CFG_INPUT_ENABLES_MRU |
+			HWRM_FUNC_CFG_INPUT_ENABLES_NUM_RSSCOS_CTXS |
+			HWRM_FUNC_CFG_INPUT_ENABLES_NUM_STAT_CTXS |
+			HWRM_FUNC_CFG_INPUT_ENABLES_NUM_CMPL_RINGS |
+			HWRM_FUNC_CFG_INPUT_ENABLES_NUM_TX_RINGS |
+			HWRM_FUNC_CFG_INPUT_ENABLES_NUM_RX_RINGS |
+			HWRM_FUNC_CFG_INPUT_ENABLES_NUM_L2_CTXS |
+			HWRM_FUNC_CFG_INPUT_ENABLES_NUM_VNICS |
+			HWRM_FUNC_CFG_INPUT_ENABLES_NUM_HW_RING_GRPS);
+	req.flags = rte_cpu_to_le_32(bp->pf.func_cfg_flags);
+	req.mtu = rte_cpu_to_le_16(bp->eth_dev->data->mtu + ETHER_HDR_LEN +
+				   ETHER_CRC_LEN + VLAN_TAG_SIZE);
+	req.mru = rte_cpu_to_le_16(bp->eth_dev->data->mtu + ETHER_HDR_LEN +
+				   ETHER_CRC_LEN + VLAN_TAG_SIZE);
+	req.num_rsscos_ctxs = rte_cpu_to_le_16(bp->max_rsscos_ctx);
+	req.num_stat_ctxs = rte_cpu_to_le_16(bp->max_stat_ctx);
+	req.num_cmpl_rings = rte_cpu_to_le_16(bp->max_cp_rings);
+	req.num_tx_rings = rte_cpu_to_le_16(tx_rings);
+	req.num_rx_rings = rte_cpu_to_le_16(bp->max_rx_rings);
+	req.num_l2_ctxs = rte_cpu_to_le_16(bp->max_l2_ctx);
+	req.num_vnics = rte_cpu_to_le_16(bp->max_vnics);
+	req.num_hw_ring_grps = rte_cpu_to_le_16(bp->max_ring_grps);
+	req.fid = rte_cpu_to_le_16(0xffff);
+
+	HWRM_PREP(req, FUNC_CFG, -1, resp);
+
+	rc = bnxt_hwrm_send_message(bp, &req, sizeof(req));
+	HWRM_CHECK_RESULT;
+
+	return rc;
+}
+
+static void populate_vf_func_cfg_req(struct bnxt *bp,
+				     struct hwrm_func_cfg_input *req,
+				     int num_vfs)
+{
+	req->enables = rte_cpu_to_le_32(HWRM_FUNC_CFG_INPUT_ENABLES_MTU |
+			HWRM_FUNC_CFG_INPUT_ENABLES_MRU |
+			HWRM_FUNC_CFG_INPUT_ENABLES_NUM_RSSCOS_CTXS |
+			HWRM_FUNC_CFG_INPUT_ENABLES_NUM_STAT_CTXS |
+			HWRM_FUNC_CFG_INPUT_ENABLES_NUM_CMPL_RINGS |
+			HWRM_FUNC_CFG_INPUT_ENABLES_NUM_TX_RINGS |
+			HWRM_FUNC_CFG_INPUT_ENABLES_NUM_RX_RINGS |
+			HWRM_FUNC_CFG_INPUT_ENABLES_NUM_L2_CTXS |
+			HWRM_FUNC_CFG_INPUT_ENABLES_NUM_VNICS |
+			HWRM_FUNC_CFG_INPUT_ENABLES_NUM_HW_RING_GRPS);
+
+	req->mtu = rte_cpu_to_le_16(bp->eth_dev->data->mtu + ETHER_HDR_LEN +
+				    ETHER_CRC_LEN + VLAN_TAG_SIZE);
+	req->mru = rte_cpu_to_le_16(bp->eth_dev->data->mtu + ETHER_HDR_LEN +
+				    ETHER_CRC_LEN + VLAN_TAG_SIZE);
+	req->num_rsscos_ctxs = rte_cpu_to_le_16(bp->max_rsscos_ctx /
+						(num_vfs + 1));
+	req->num_stat_ctxs = rte_cpu_to_le_16(bp->max_stat_ctx / (num_vfs + 1));
+	req->num_cmpl_rings = rte_cpu_to_le_16(bp->max_cp_rings /
+					       (num_vfs + 1));
+	req->num_tx_rings = rte_cpu_to_le_16(bp->max_tx_rings / (num_vfs + 1));
+	req->num_rx_rings = rte_cpu_to_le_16(bp->max_rx_rings / (num_vfs + 1));
+	req->num_l2_ctxs = rte_cpu_to_le_16(bp->max_l2_ctx / (num_vfs + 1));
+	/* TODO: For now, do not support VMDq/RFS on VFs. */
+	req->num_vnics = rte_cpu_to_le_16(1);
+	req->num_hw_ring_grps = rte_cpu_to_le_16(bp->max_ring_grps /
+						 (num_vfs + 1));
+}
+
+static void reserve_resources_from_vf(struct bnxt *bp,
+				      struct hwrm_func_cfg_input *cfg_req,
+				      int vf)
+{
+	struct hwrm_func_qcaps_input req = {0};
+	struct hwrm_func_qcaps_output *resp = bp->hwrm_cmd_resp_addr;
+	int rc;
+
+	/* Get the actual allocated values now */
+	HWRM_PREP(req, FUNC_QCAPS, -1, resp);
+	req.fid = rte_cpu_to_le_16(bp->pf.vf_info[vf].fid);
+	rc = bnxt_hwrm_send_message(bp, &req, sizeof(req));
+
+	if (rc) {
+		RTE_LOG(ERR, PMD, "hwrm_func_qcaps failed rc:%d\n", rc);
+		copy_func_cfg_to_qcaps(cfg_req, resp);
+	} else if (resp->error_code) {
+		rc = rte_le_to_cpu_16(resp->error_code);
+		RTE_LOG(ERR, PMD, "hwrm_func_qcaps error %d\n", rc);
+		copy_func_cfg_to_qcaps(cfg_req, resp);
+	}
+
+	bp->max_rsscos_ctx -= rte_le_to_cpu_16(resp->max_rsscos_ctx);
+	bp->max_stat_ctx -= rte_le_to_cpu_16(resp->max_stat_ctx);
+	bp->max_cp_rings -= rte_le_to_cpu_16(resp->max_cmpl_rings);
+	bp->max_tx_rings -= rte_le_to_cpu_16(resp->max_tx_rings);
+	bp->max_rx_rings -= rte_le_to_cpu_16(resp->max_rx_rings);
+	bp->max_l2_ctx -= rte_le_to_cpu_16(resp->max_l2_ctxs);
+	/*
+	 * TODO: While not supporting VMDq with VFs, max_vnics is always
+	 * forced to 1 in this case
+	 */
+	//bp->max_vnics -= rte_le_to_cpu_16(esp->max_vnics);
+	bp->max_ring_grps -= rte_le_to_cpu_16(resp->max_hw_ring_grps);
+}
+
+static int update_pf_resource_max(struct bnxt *bp)
+{
+	struct hwrm_func_qcfg_input req = {0};
+	struct hwrm_func_qcfg_output *resp = bp->hwrm_cmd_resp_addr;
+	int rc;
+
+	/* And copy the allocated numbers into the pf struct */
+	HWRM_PREP(req, FUNC_QCFG, -1, resp);
+	req.fid = rte_cpu_to_le_16(0xffff);
+	rc = bnxt_hwrm_send_message(bp, &req, sizeof(req));
+	HWRM_CHECK_RESULT;
+
+	/* Only TX ring value reflects actual allocation? TODO */
+	bp->max_tx_rings = rte_le_to_cpu_16(resp->alloc_tx_rings);
+	bp->pf.evb_mode = resp->evb_mode;
+
+	return rc;
+}
+
+int bnxt_hwrm_allocate_pf_only(struct bnxt *bp)
+{
+	int rc;
+
+	if (!BNXT_PF(bp)) {
+		RTE_LOG(ERR, PMD, "Attempt to allcoate VFs on a VF!\n");
+		return -1;
+	}
+
+	rc = bnxt_hwrm_func_qcaps(bp);
+	if (rc)
+		return rc;
+
+	bp->pf.func_cfg_flags &=
+		~(HWRM_FUNC_CFG_INPUT_FLAGS_STD_TX_RING_MODE_ENABLE |
+		  HWRM_FUNC_CFG_INPUT_FLAGS_STD_TX_RING_MODE_DISABLE);
+	bp->pf.func_cfg_flags |=
+		HWRM_FUNC_CFG_INPUT_FLAGS_STD_TX_RING_MODE_DISABLE;
+	rc = bnxt_hwrm_pf_func_cfg(bp, bp->max_tx_rings);
+	return rc;
+}
+
+int bnxt_hwrm_allocate_vfs(struct bnxt *bp, int num_vfs)
+{
+	struct hwrm_func_cfg_input req = {0};
+	struct hwrm_func_cfg_output *resp = bp->hwrm_cmd_resp_addr;
+	int i;
+	size_t sz;
+	int rc = 0;
+	size_t req_buf_sz;
+
+	if (!BNXT_PF(bp)) {
+		RTE_LOG(ERR, PMD, "Attempt to allcoate VFs on a VF!\n");
+		return -1;
+	}
+
+	rc = bnxt_hwrm_func_qcaps(bp);
+
+	if (rc)
+		return rc;
+
+	bp->pf.active_vfs = num_vfs;
+
+	/*
+	 * First, configure the PF to only use one TX ring.  This ensures that
+	 * there are enough rings for all VFs.
+	 *
+	 * If we don't do this, when we call func_alloc() later, we will lock
+	 * extra rings to the PF that won't be available during func_cfg() of
+	 * the VFs.
+	 *
+	 * This has been fixed with firmware versions above 20.6.54
+	 */
+	bp->pf.func_cfg_flags &=
+		~(HWRM_FUNC_CFG_INPUT_FLAGS_STD_TX_RING_MODE_ENABLE |
+		  HWRM_FUNC_CFG_INPUT_FLAGS_STD_TX_RING_MODE_DISABLE);
+	bp->pf.func_cfg_flags |=
+		HWRM_FUNC_CFG_INPUT_FLAGS_STD_TX_RING_MODE_ENABLE;
+	rc = bnxt_hwrm_pf_func_cfg(bp, 1);
+	if (rc)
+		return rc;
+
+	/*
+	 * Now, create and register a buffer to hold forwarded VF requests
+	 */
+	req_buf_sz = num_vfs * HWRM_MAX_REQ_LEN;
+	bp->pf.vf_req_buf = rte_malloc("bnxt_vf_fwd", req_buf_sz,
+		page_roundup(num_vfs * HWRM_MAX_REQ_LEN));
+	if (bp->pf.vf_req_buf == NULL) {
+		rc = -ENOMEM;
+		goto error_free;
+	}
+	for (sz = 0; sz < req_buf_sz; sz += getpagesize())
+		rte_mem_lock_page(((char *)bp->pf.vf_req_buf) + sz);
+	for (i = 0; i < num_vfs; i++)
+		bp->pf.vf_info[i].req_buf = ((char *)bp->pf.vf_req_buf) +
+					(i * HWRM_MAX_REQ_LEN);
+
+	rc = bnxt_hwrm_func_buf_rgtr(bp);
+	if (rc)
+		goto error_free;
+
+	populate_vf_func_cfg_req(bp, &req, num_vfs);
+
+	bp->pf.active_vfs = 0;
+	for (i = 0; i < num_vfs; i++) {
+		HWRM_PREP(req, FUNC_CFG, -1, resp);
+		req.flags = rte_cpu_to_le_32(bp->pf.vf_info[i].func_cfg_flags);
+		req.fid = rte_cpu_to_le_16(bp->pf.vf_info[i].fid);
+		rc = bnxt_hwrm_send_message(bp, &req, sizeof(req));
+
+		/* Clear enable flag for next pass */
+		req.enables &= ~rte_cpu_to_le_32(
+				HWRM_FUNC_CFG_INPUT_ENABLES_DFLT_MAC_ADDR);
+
+		if (rc || resp->error_code) {
+			RTE_LOG(ERR, PMD,
+				"Failed to initizlie VF %d\n", i);
+			RTE_LOG(ERR, PMD,
+				"Not all VFs available. (%d, %d)\n",
+				rc, resp->error_code);
+			break;
+		}
+
+		reserve_resources_from_vf(bp, &req, i);
+		bp->pf.active_vfs++;
+	}
+
+	/*
+	 * Now configure the PF to use "the rest" of the resources
+	 * We're using STD_TX_RING_MODE here though which will limit the TX
+	 * rings.  This will allow QoS to function properly.  Not setting this
+	 * will cause PF rings to break bandwidth settings.
+	 */
+	rc = bnxt_hwrm_pf_func_cfg(bp, bp->max_tx_rings);
+	if (rc)
+		goto error_free;
+
+	rc = update_pf_resource_max(bp);
+	if (rc)
+		goto error_free;
+
+	return rc;
+
+error_free:
+	bnxt_hwrm_func_buf_unrgtr(bp);
+	return rc;
+}
+
+
+int bnxt_hwrm_func_buf_rgtr(struct bnxt *bp)
+{
+	int rc = 0;
+	struct hwrm_func_buf_rgtr_input req = {.req_type = 0 };
+	struct hwrm_func_buf_rgtr_output *resp = bp->hwrm_cmd_resp_addr;
+
+	HWRM_PREP(req, FUNC_BUF_RGTR, -1, resp);
+
+	req.req_buf_num_pages = rte_cpu_to_le_16(1);
+	req.req_buf_page_size = rte_cpu_to_le_16(
+			 page_getenum(bp->pf.active_vfs * HWRM_MAX_REQ_LEN));
+	req.req_buf_len = rte_cpu_to_le_16(HWRM_MAX_REQ_LEN);
+	req.req_buf_page_addr[0] =
+		rte_cpu_to_le_64(rte_mem_virt2phy(bp->pf.vf_req_buf));
+	if (req.req_buf_page_addr[0] == 0) {
+		RTE_LOG(ERR, PMD,
+			"unable to map buffer address to physical memory\n");
+		return -ENOMEM;
+	}
+
+	rc = bnxt_hwrm_send_message(bp, &req, sizeof(req));
+
+	HWRM_CHECK_RESULT;
+
+	return rc;
+}
+
+int bnxt_hwrm_func_buf_unrgtr(struct bnxt *bp)
+{
+	int rc = 0;
+	struct hwrm_func_buf_unrgtr_input req = {.req_type = 0 };
+	struct hwrm_func_buf_unrgtr_output *resp = bp->hwrm_cmd_resp_addr;
+
+	HWRM_PREP(req, FUNC_BUF_UNRGTR, -1, resp);
+
+	rc = bnxt_hwrm_send_message(bp, &req, sizeof(req));
+
+	HWRM_CHECK_RESULT;
+
+	return rc;
+}
+
+int bnxt_hwrm_func_cfg_def_cp(struct bnxt *bp)
+{
+	struct hwrm_func_cfg_output *resp = bp->hwrm_cmd_resp_addr;
+	struct hwrm_func_cfg_input req = {0};
+	int rc;
+
+	HWRM_PREP(req, FUNC_CFG, -1, resp);
+	req.fid = rte_cpu_to_le_16(0xffff);
+	req.flags = rte_cpu_to_le_32(bp->pf.func_cfg_flags);
+	req.enables = rte_cpu_to_le_32(
+			HWRM_FUNC_CFG_INPUT_ENABLES_ASYNC_EVENT_CR);
+	req.async_event_cr = rte_cpu_to_le_16(
+			bp->def_cp_ring->cp_ring_struct->fw_ring_id);
+	rc = bnxt_hwrm_send_message(bp, &req, sizeof(req));
+	HWRM_CHECK_RESULT;
+
+	return rc;
+}
+
+int bnxt_hwrm_reject_fwd_resp(struct bnxt *bp, uint16_t target_id,
+			      void *encaped, size_t ec_size)
+{
+	int rc = 0;
+	struct hwrm_reject_fwd_resp_input req = {.req_type = 0};
+	struct hwrm_reject_fwd_resp_output *resp = bp->hwrm_cmd_resp_addr;
+
+	if (ec_size > sizeof(req.encap_request))
+		return -1;
+
+	HWRM_PREP(req, REJECT_FWD_RESP, -1, resp);
+
+	req.encap_resp_target_id = rte_cpu_to_le_16(target_id);
+	memcpy(req.encap_request, encaped, ec_size);
+
+	rc = bnxt_hwrm_send_message(bp, &req, sizeof(req));
+
+	HWRM_CHECK_RESULT;
+
+	return rc;
+}
diff --git a/drivers/net/bnxt/bnxt_hwrm.h b/drivers/net/bnxt/bnxt_hwrm.h
index 6519ef2..0046ba4 100644
--- a/drivers/net/bnxt/bnxt_hwrm.h
+++ b/drivers/net/bnxt/bnxt_hwrm.h
@@ -51,14 +51,17 @@ int bnxt_hwrm_clear_filter(struct bnxt *bp,
 int bnxt_hwrm_set_filter(struct bnxt *bp,
 			 struct bnxt_vnic_info *vnic,
 			 struct bnxt_filter_info *filter);
-
 int bnxt_hwrm_exec_fwd_resp(struct bnxt *bp, void *fwd_cmd);
+int bnxt_hwrm_reject_fwd_resp(struct bnxt *bp, uint16_t target_id,
+			      void *encaped, size_t ec_size);
 
-int bnxt_hwrm_func_driver_register(struct bnxt *bp, uint32_t flags,
-				   uint32_t *vf_req_fwd);
+int bnxt_hwrm_func_buf_rgtr(struct bnxt *bp);
+int bnxt_hwrm_func_buf_unrgtr(struct bnxt *bp);
+int bnxt_hwrm_func_driver_register(struct bnxt *bp);
 int bnxt_hwrm_func_qcaps(struct bnxt *bp);
 int bnxt_hwrm_func_reset(struct bnxt *bp);
 int bnxt_hwrm_func_driver_unregister(struct bnxt *bp, uint32_t flags);
+int bnxt_hwrm_func_cfg_def_cp(struct bnxt *bp);
 
 int bnxt_hwrm_queue_qportcfg(struct bnxt *bp);
 
@@ -81,6 +84,8 @@ int bnxt_hwrm_ver_get(struct bnxt *bp);
 
 int bnxt_hwrm_vnic_alloc(struct bnxt *bp, struct bnxt_vnic_info *vnic);
 int bnxt_hwrm_vnic_cfg(struct bnxt *bp, struct bnxt_vnic_info *vnic);
+int bnxt_hwrm_vnic_qcfg(struct bnxt *bp, struct bnxt_vnic_info *vnic,
+				int16_t fw_vf_id);
 int bnxt_hwrm_vnic_ctx_alloc(struct bnxt *bp, struct bnxt_vnic_info *vnic);
 int bnxt_hwrm_vnic_ctx_free(struct bnxt *bp, struct bnxt_vnic_info *vnic);
 int bnxt_hwrm_vnic_free(struct bnxt *bp, struct bnxt_vnic_info *vnic);
@@ -101,5 +106,11 @@ int bnxt_alloc_hwrm_resources(struct bnxt *bp);
 int bnxt_get_hwrm_link_config(struct bnxt *bp, struct rte_eth_link *link);
 int bnxt_set_hwrm_link_config(struct bnxt *bp, bool link_up);
 int bnxt_hwrm_func_qcfg(struct bnxt *bp);
+int bnxt_hwrm_allocate_pf_only(struct bnxt *bp);
+int bnxt_hwrm_allocate_vfs(struct bnxt *bp, int num_vfs);
+int bnxt_hwrm_func_vf_mac(struct bnxt *bp, uint16_t vf,
+			  const uint8_t *mac_addr);
+int bnxt_hwrm_func_qcfg_vf_default_mac(struct bnxt *bp, uint16_t vf,
+				       struct ether_addr *mac);
 
 #endif
diff --git a/drivers/net/bnxt/bnxt_vnic.c b/drivers/net/bnxt/bnxt_vnic.c
index 33fdde2..4e378a9 100644
--- a/drivers/net/bnxt/bnxt_vnic.c
+++ b/drivers/net/bnxt/bnxt_vnic.c
@@ -69,21 +69,14 @@ void bnxt_init_vnics(struct bnxt *bp)
 	uint16_t max_vnics;
 	int i, j;
 
-	if (BNXT_PF(bp)) {
-		struct bnxt_pf_info *pf = &bp->pf;
-
-		max_vnics = pf->max_vnics;
-	} else {
-		struct bnxt_vf_info *vf = &bp->vf;
-
-		max_vnics = vf->max_vnics;
-	}
+	max_vnics = bp->max_vnics;
 	STAILQ_INIT(&bp->free_vnic_list);
 	for (i = 0; i < max_vnics; i++) {
 		vnic = &bp->vnic_info[i];
 		vnic->fw_vnic_id = (uint16_t)HWRM_NA_SIGNATURE;
-		vnic->fw_rss_cos_lb_ctx = (uint16_t)HWRM_NA_SIGNATURE;
-		vnic->ctx_is_rss_cos_lb = HW_CONTEXT_NONE;
+		vnic->rss_rule = (uint16_t)HWRM_NA_SIGNATURE;
+		vnic->cos_rule = (uint16_t)HWRM_NA_SIGNATURE;
+		vnic->lb_rule = (uint16_t)HWRM_NA_SIGNATURE;
 
 		for (j = 0; j < MAX_QUEUES_PER_VNIC; j++)
 			vnic->fw_grp_ids[j] = (uint16_t)HWRM_NA_SIGNATURE;
@@ -181,15 +174,7 @@ int bnxt_alloc_vnic_attributes(struct bnxt *bp)
 	uint16_t max_vnics;
 	int i;
 
-	if (BNXT_PF(bp)) {
-		struct bnxt_pf_info *pf = &bp->pf;
-
-		max_vnics = pf->max_vnics;
-	} else {
-		struct bnxt_vf_info *vf = &bp->vf;
-
-		max_vnics = vf->max_vnics;
-	}
+	max_vnics = bp->max_vnics;
 	snprintf(mz_name, RTE_MEMZONE_NAMESIZE,
 		 "bnxt_%04x:%02x:%02x:%02x_vnicattr", pdev->addr.domain,
 		 pdev->addr.bus, pdev->addr.devid, pdev->addr.function);
@@ -232,15 +217,7 @@ void bnxt_free_vnic_mem(struct bnxt *bp)
 	if (bp->vnic_info == NULL)
 		return;
 
-	if (BNXT_PF(bp)) {
-		struct bnxt_pf_info *pf = &bp->pf;
-
-		max_vnics = pf->max_vnics;
-	} else {
-		struct bnxt_vf_info *vf = &bp->vf;
-
-		max_vnics = vf->max_vnics;
-	}
+	max_vnics = bp->max_vnics;
 	for (i = 0; i < max_vnics; i++) {
 		vnic = &bp->vnic_info[i];
 		if (vnic->fw_vnic_id != (uint16_t)HWRM_NA_SIGNATURE) {
@@ -258,15 +235,7 @@ int bnxt_alloc_vnic_mem(struct bnxt *bp)
 	struct bnxt_vnic_info *vnic_mem;
 	uint16_t max_vnics;
 
-	if (BNXT_PF(bp)) {
-		struct bnxt_pf_info *pf = &bp->pf;
-
-		max_vnics = pf->max_vnics;
-	} else {
-		struct bnxt_vf_info *vf = &bp->vf;
-
-		max_vnics = vf->max_vnics;
-	}
+	max_vnics = bp->max_vnics;
 	/* Allocate memory for VNIC pool and filter pool */
 	vnic_mem = rte_zmalloc("bnxt_vnic_info",
 			       max_vnics * sizeof(struct bnxt_vnic_info), 0);
diff --git a/drivers/net/bnxt/bnxt_vnic.h b/drivers/net/bnxt/bnxt_vnic.h
index 9671ba4..ca1c9cf 100644
--- a/drivers/net/bnxt/bnxt_vnic.h
+++ b/drivers/net/bnxt/bnxt_vnic.h
@@ -42,8 +42,7 @@ struct bnxt_vnic_info {
 	uint8_t		ff_pool_idx;
 
 	uint16_t	fw_vnic_id; /* returned by Chimp during alloc */
-	uint16_t	fw_rss_cos_lb_ctx;
-	uint16_t	ctx_is_rss_cos_lb;
+	uint16_t	rss_rule;
 #define MAX_NUM_TRAFFIC_CLASSES		8
 #define MAX_NUM_RSS_QUEUES_PER_VNIC	16
 #define MAX_QUEUES_PER_VNIC	(MAX_NUM_RSS_QUEUES_PER_VNIC + \
@@ -51,6 +50,8 @@ struct bnxt_vnic_info {
 	uint16_t	start_grp_id;
 	uint16_t	end_grp_id;
 	uint16_t	fw_grp_ids[MAX_QUEUES_PER_VNIC];
+	uint16_t	dflt_ring_grp;
+	uint16_t	mru;
 	uint16_t	hash_type;
 	phys_addr_t	rss_table_dma_addr;
 	uint16_t	*rss_table;
@@ -60,8 +61,14 @@ struct bnxt_vnic_info {
 #define BNXT_VNIC_INFO_PROMISC			(1 << 0)
 #define BNXT_VNIC_INFO_ALLMULTI			(1 << 1)
 
+	uint16_t	cos_rule;
+	uint16_t	lb_rule;
 	bool		vlan_strip;
 	bool		func_default;
+	bool		bd_stall;
+	bool		roce_dual;
+	bool		roce_only;
+	bool		rss_dflt_cr;
 
 	STAILQ_HEAD(, bnxt_filter_info)	filter;
 };
-- 
2.10.1 (Apple Git-78)