DPDK patches and discussions
 help / color / mirror / Atom feed
From: Somnath Kotur <somnath.kotur@broadcom.com>
To: dev@dpdk.org
Cc: ferruh.yigit@intel.com
Subject: [dpdk-dev] [PATCH 09/50] net/bnxt: add support for Exact Match
Date: Fri, 12 Jun 2020 18:58:53 +0530	[thread overview]
Message-ID: <20200612132934.16488-10-somnath.kotur@broadcom.com> (raw)
In-Reply-To: <20200612132934.16488-1-somnath.kotur@broadcom.com>

From: Pete Spreadborough <peter.spreadborough@broadcom.com>

- Add Exact Match support
- Create EM table pool of memory indices
- Insert exact match internal entry API
- Sends EM internal insert and delete request to firmware

Signed-off-by: Pete Spreadborough <peter.spreadborough@broadcom.com>
Reviewed-by: Randy Schacher <stuart.schacher@broadcom.com>
Reviewed-by: Ajit Kumar Khaparde <ajit.khaparde@broadcom.com>
Signed-off-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com>
---
 drivers/net/bnxt/hsi_struct_def_dpdk.h        | 3095 +++++++++++++++++++++----
 drivers/net/bnxt/tf_core/hwrm_tf.h            |    9 +
 drivers/net/bnxt/tf_core/lookup3.h            |    1 -
 drivers/net/bnxt/tf_core/stack.c              |    8 +
 drivers/net/bnxt/tf_core/stack.h              |   10 +
 drivers/net/bnxt/tf_core/tf_core.c            |  144 +-
 drivers/net/bnxt/tf_core/tf_core.h            |  383 +--
 drivers/net/bnxt/tf_core/tf_em.c              |   98 +-
 drivers/net/bnxt/tf_core/tf_em.h              |   31 +
 drivers/net/bnxt/tf_core/tf_ext_flow_handle.h |   16 +
 drivers/net/bnxt/tf_core/tf_msg.c             |   86 +-
 drivers/net/bnxt/tf_core/tf_msg.h             |   13 +
 drivers/net/bnxt/tf_core/tf_session.h         |   18 +
 drivers/net/bnxt/tf_core/tf_tbl.c             |  123 +-
 drivers/net/bnxt/tf_core/tf_tbl.h             |   57 +-
 drivers/net/bnxt/tf_core/tfp.h                |  123 +-
 16 files changed, 3511 insertions(+), 704 deletions(-)

diff --git a/drivers/net/bnxt/hsi_struct_def_dpdk.h b/drivers/net/bnxt/hsi_struct_def_dpdk.h
index 7e30c9f..e51f42f 100644
--- a/drivers/net/bnxt/hsi_struct_def_dpdk.h
+++ b/drivers/net/bnxt/hsi_struct_def_dpdk.h
@@ -611,6 +611,10 @@ struct cmd_nums {
 	#define HWRM_FUNC_VF_BW_QCFG                      UINT32_C(0x196)
 	/* Queries pf ids belong to specified host(s) */
 	#define HWRM_FUNC_HOST_PF_IDS_QUERY               UINT32_C(0x197)
+	/* Queries extended stats per function */
+	#define HWRM_FUNC_QSTATS_EXT                      UINT32_C(0x198)
+	/* Queries exteded statitics context */
+	#define HWRM_STAT_EXT_CTX_QUERY                   UINT32_C(0x199)
 	/* Experimental */
 	#define HWRM_SELFTEST_QLIST                       UINT32_C(0x200)
 	/* Experimental */
@@ -647,41 +651,49 @@ struct cmd_nums {
 	/* Experimental */
 	#define HWRM_TF_SESSION_ATTACH                    UINT32_C(0x2c7)
 	/* Experimental */
-	#define HWRM_TF_SESSION_CLOSE                     UINT32_C(0x2c8)
+	#define HWRM_TF_SESSION_REGISTER                  UINT32_C(0x2c8)
 	/* Experimental */
-	#define HWRM_TF_SESSION_QCFG                      UINT32_C(0x2c9)
+	#define HWRM_TF_SESSION_UNREGISTER                UINT32_C(0x2c9)
 	/* Experimental */
-	#define HWRM_TF_SESSION_RESC_QCAPS                UINT32_C(0x2ca)
+	#define HWRM_TF_SESSION_CLOSE                     UINT32_C(0x2ca)
 	/* Experimental */
-	#define HWRM_TF_SESSION_RESC_ALLOC                UINT32_C(0x2cb)
+	#define HWRM_TF_SESSION_QCFG                      UINT32_C(0x2cb)
 	/* Experimental */
-	#define HWRM_TF_SESSION_RESC_FREE                 UINT32_C(0x2cc)
+	#define HWRM_TF_SESSION_RESC_QCAPS                UINT32_C(0x2cc)
 	/* Experimental */
-	#define HWRM_TF_SESSION_RESC_FLUSH                UINT32_C(0x2cd)
+	#define HWRM_TF_SESSION_RESC_ALLOC                UINT32_C(0x2cd)
 	/* Experimental */
-	#define HWRM_TF_TBL_TYPE_GET                      UINT32_C(0x2d0)
+	#define HWRM_TF_SESSION_RESC_FREE                 UINT32_C(0x2ce)
 	/* Experimental */
-	#define HWRM_TF_TBL_TYPE_SET                      UINT32_C(0x2d1)
+	#define HWRM_TF_SESSION_RESC_FLUSH                UINT32_C(0x2cf)
 	/* Experimental */
-	#define HWRM_TF_CTXT_MEM_RGTR                     UINT32_C(0x2da)
+	#define HWRM_TF_TBL_TYPE_GET                      UINT32_C(0x2da)
 	/* Experimental */
-	#define HWRM_TF_CTXT_MEM_UNRGTR                   UINT32_C(0x2db)
+	#define HWRM_TF_TBL_TYPE_SET                      UINT32_C(0x2db)
 	/* Experimental */
-	#define HWRM_TF_EXT_EM_QCAPS                      UINT32_C(0x2dc)
+	#define HWRM_TF_CTXT_MEM_RGTR                     UINT32_C(0x2e4)
 	/* Experimental */
-	#define HWRM_TF_EXT_EM_OP                         UINT32_C(0x2dd)
+	#define HWRM_TF_CTXT_MEM_UNRGTR                   UINT32_C(0x2e5)
 	/* Experimental */
-	#define HWRM_TF_EXT_EM_CFG                        UINT32_C(0x2de)
+	#define HWRM_TF_EXT_EM_QCAPS                      UINT32_C(0x2e6)
 	/* Experimental */
-	#define HWRM_TF_EXT_EM_QCFG                       UINT32_C(0x2df)
+	#define HWRM_TF_EXT_EM_OP                         UINT32_C(0x2e7)
 	/* Experimental */
-	#define HWRM_TF_TCAM_SET                          UINT32_C(0x2ee)
+	#define HWRM_TF_EXT_EM_CFG                        UINT32_C(0x2e8)
 	/* Experimental */
-	#define HWRM_TF_TCAM_GET                          UINT32_C(0x2ef)
+	#define HWRM_TF_EXT_EM_QCFG                       UINT32_C(0x2e9)
 	/* Experimental */
-	#define HWRM_TF_TCAM_MOVE                         UINT32_C(0x2f0)
+	#define HWRM_TF_EM_INSERT                         UINT32_C(0x2ea)
 	/* Experimental */
-	#define HWRM_TF_TCAM_FREE                         UINT32_C(0x2f1)
+	#define HWRM_TF_EM_DELETE                         UINT32_C(0x2eb)
+	/* Experimental */
+	#define HWRM_TF_TCAM_SET                          UINT32_C(0x2f8)
+	/* Experimental */
+	#define HWRM_TF_TCAM_GET                          UINT32_C(0x2f9)
+	/* Experimental */
+	#define HWRM_TF_TCAM_MOVE                         UINT32_C(0x2fa)
+	/* Experimental */
+	#define HWRM_TF_TCAM_FREE                         UINT32_C(0x2fb)
 	/* Experimental */
 	#define HWRM_SV                                   UINT32_C(0x400)
 	/* Experimental */
@@ -715,6 +727,13 @@ struct cmd_nums {
 	#define HWRM_DBG_CRASHDUMP_ERASE                  UINT32_C(0xff1e)
 	/* Send driver debug information to firmware */
 	#define HWRM_DBG_DRV_TRACE                        UINT32_C(0xff1f)
+	/* Query debug capabilities of firmware */
+	#define HWRM_DBG_QCAPS                            UINT32_C(0xff20)
+	/* Retrieve debug settings of firmware */
+	#define HWRM_DBG_QCFG                             UINT32_C(0xff21)
+	/* Set destination parameters for crashdump medium */
+	#define HWRM_DBG_CRASHDUMP_MEDIUM_CFG             UINT32_C(0xff22)
+	#define HWRM_NVM_REQ_ARBITRATION                  UINT32_C(0xffed)
 	/* Experimental */
 	#define HWRM_NVM_FACTORY_DEFAULTS                 UINT32_C(0xffee)
 	#define HWRM_NVM_VALIDATE_OPTION                  UINT32_C(0xffef)
@@ -914,8 +933,8 @@ struct hwrm_err_output {
 #define HWRM_VERSION_MINOR 10
 #define HWRM_VERSION_UPDATE 1
 /* non-zero means beta version */
-#define HWRM_VERSION_RSVD 30
-#define HWRM_VERSION_STR "1.10.1.30"
+#define HWRM_VERSION_RSVD 45
+#define HWRM_VERSION_STR "1.10.1.45"
 
 /****************
  * hwrm_ver_get *
@@ -2293,6 +2312,35 @@ struct cmpl_base {
 	 */
 	#define CMPL_BASE_TYPE_TX_L2             UINT32_C(0x0)
 	/*
+	 * NO-OP completion:
+	 * Completion of NO-OP. Length = 16B
+	 */
+	#define CMPL_BASE_TYPE_NO_OP             UINT32_C(0x1)
+	/*
+	 * TX L2 coalesced completion:
+	 * Completion of coalesced TX packet. Length = 16B
+	 */
+	#define CMPL_BASE_TYPE_TX_L2_COAL        UINT32_C(0x2)
+	/*
+	 * TX L2 PTP completion:
+	 * Completion of PTP TX packet. Length = 32B
+	 */
+	#define CMPL_BASE_TYPE_TX_L2_PTP         UINT32_C(0x3)
+	/*
+	 * RX L2 TPA Start V2 Completion:
+	 * Completion of and L2 RX packet. Length = 32B
+	 * This is the new version of the RX_TPA_START completion used
+	 * in SR2 and later chips.
+	 */
+	#define CMPL_BASE_TYPE_RX_TPA_START_V2   UINT32_C(0xd)
+	/*
+	 * RX L2 V2 completion:
+	 * Completion of and L2 RX packet. Length = 32B
+	 * This is the new version of the RX_L2 completion used in SR2
+	 * and later chips.
+	 */
+	#define CMPL_BASE_TYPE_RX_L2_V2          UINT32_C(0xf)
+	/*
 	 * RX L2 completion:
 	 * Completion of and L2 RX packet. Length = 32B
 	 */
@@ -2322,6 +2370,24 @@ struct cmpl_base {
 	 */
 	#define CMPL_BASE_TYPE_STAT_EJECT        UINT32_C(0x1a)
 	/*
+	 * VEE Flush Completion:
+	 * This completion is inserted manually by
+	 * the Primate and processed by the VEE hardware to ensure that
+	 * all completions on a VEE function have been processed by the
+	 * VEE hardware before FLR process is completed.
+	 */
+	#define CMPL_BASE_TYPE_VEE_FLUSH         UINT32_C(0x1c)
+	/*
+	 * Mid Path Short Completion :
+	 * Completion of a Mid Path Command. Length = 16B
+	 */
+	#define CMPL_BASE_TYPE_MID_PATH_SHORT    UINT32_C(0x1e)
+	/*
+	 * Mid Path Long Completion :
+	 * Completion of a Mid Path Command. Length = 32B
+	 */
+	#define CMPL_BASE_TYPE_MID_PATH_LONG     UINT32_C(0x1f)
+	/*
 	 * HWRM Command Completion:
 	 * Completion of an HWRM command.
 	 */
@@ -2398,7 +2464,9 @@ struct tx_cmpl {
 	uint16_t	unused_0;
 	/*
 	 * This is a copy of the opaque field from the first TX BD of this
-	 * transmitted packet.
+	 * transmitted packet. Note that, if the packet was described by a short
+	 * CSO or short CSO inline BD, then the 16-bit opaque field from the
+	 * short CSO BD will appear in the bottom 16 bits of this field.
 	 */
 	uint32_t	opaque;
 	uint16_t	errors_v;
@@ -2407,58 +2475,352 @@ struct tx_cmpl {
 	 * for each pass through the completion queue. The even passes
 	 * will write 1. The odd passes will write 0.
 	 */
-	#define TX_CMPL_V                              UINT32_C(0x1)
-	#define TX_CMPL_ERRORS_MASK                    UINT32_C(0xfffe)
-	#define TX_CMPL_ERRORS_SFT                     1
+	#define TX_CMPL_V                                  UINT32_C(0x1)
+	#define TX_CMPL_ERRORS_MASK                        UINT32_C(0xfffe)
+	#define TX_CMPL_ERRORS_SFT                         1
 	/*
 	 * This error indicates that there was some sort of problem
 	 * with the BDs for the packet.
 	 */
-	#define TX_CMPL_ERRORS_BUFFER_ERROR_MASK        UINT32_C(0xe)
-	#define TX_CMPL_ERRORS_BUFFER_ERROR_SFT         1
+	#define TX_CMPL_ERRORS_BUFFER_ERROR_MASK            UINT32_C(0xe)
+	#define TX_CMPL_ERRORS_BUFFER_ERROR_SFT             1
 	/* No error */
-	#define TX_CMPL_ERRORS_BUFFER_ERROR_NO_ERROR      (UINT32_C(0x0) << 1)
+	#define TX_CMPL_ERRORS_BUFFER_ERROR_NO_ERROR \
+		(UINT32_C(0x0) << 1)
 	/*
 	 * Bad Format:
 	 * BDs were not formatted correctly.
 	 */
-	#define TX_CMPL_ERRORS_BUFFER_ERROR_BAD_FMT       (UINT32_C(0x2) << 1)
+	#define TX_CMPL_ERRORS_BUFFER_ERROR_BAD_FMT \
+		(UINT32_C(0x2) << 1)
 	#define TX_CMPL_ERRORS_BUFFER_ERROR_LAST \
 		TX_CMPL_ERRORS_BUFFER_ERROR_BAD_FMT
 	/*
 	 * When this bit is '1', it indicates that the length of
 	 * the packet was zero. No packet was transmitted.
 	 */
-	#define TX_CMPL_ERRORS_ZERO_LENGTH_PKT          UINT32_C(0x10)
+	#define TX_CMPL_ERRORS_ZERO_LENGTH_PKT              UINT32_C(0x10)
 	/*
 	 * When this bit is '1', it indicates that the packet
 	 * was longer than the programmed limit in TDI. No
 	 * packet was transmitted.
 	 */
-	#define TX_CMPL_ERRORS_EXCESSIVE_BD_LENGTH      UINT32_C(0x20)
+	#define TX_CMPL_ERRORS_EXCESSIVE_BD_LENGTH          UINT32_C(0x20)
 	/*
 	 * When this bit is '1', it indicates that one or more of the
 	 * BDs associated with this packet generated a PCI error.
 	 * This probably means the address was not valid.
 	 */
-	#define TX_CMPL_ERRORS_DMA_ERROR                UINT32_C(0x40)
+	#define TX_CMPL_ERRORS_DMA_ERROR                    UINT32_C(0x40)
 	/*
 	 * When this bit is '1', it indicates that the packet was longer
 	 * than indicated by the hint. No packet was transmitted.
 	 */
-	#define TX_CMPL_ERRORS_HINT_TOO_SHORT           UINT32_C(0x80)
+	#define TX_CMPL_ERRORS_HINT_TOO_SHORT               UINT32_C(0x80)
 	/*
 	 * When this bit is '1', it indicates that the packet was
 	 * dropped due to Poison TLP error on one or more of the
 	 * TLPs in the PXP completion.
 	 */
-	#define TX_CMPL_ERRORS_POISON_TLP_ERROR         UINT32_C(0x100)
+	#define TX_CMPL_ERRORS_POISON_TLP_ERROR             UINT32_C(0x100)
+	/*
+	 * When this bit is '1', it indicates that the packet was dropped
+	 * due to a transient internal error in TDC. The packet or LSO can
+	 * be retried and may transmit successfully on a subsequent attempt.
+	 */
+	#define TX_CMPL_ERRORS_INTERNAL_ERROR               UINT32_C(0x200)
+	/*
+	 * When this bit is '1', it was not possible to collect a a timestamp
+	 * for a PTP completion, in which case the timestamp_hi and
+	 * timestamp_lo fields are invalid. When this bit is '0' for a PTP
+	 * completion, the timestamp_hi and timestamp_lo fields are valid.
+	 * RJRN will copy the value of this bit into the field of the same
+	 * name in all TX completions, regardless of whether such completions
+	 * are PTP completions or other TX completions.
+	 */
+	#define TX_CMPL_ERRORS_TIMESTAMP_INVALID_ERROR      UINT32_C(0x400)
 	/* unused2 is 16 b */
 	uint16_t	unused_1;
 	/* unused3 is 32 b */
 	uint32_t	unused_2;
 } __rte_packed;
 
+/* tx_cmpl_coal (size:128b/16B) */
+struct tx_cmpl_coal {
+	uint16_t	flags_type;
+	/*
+	 * This field indicates the exact type of the completion.
+	 * By convention, the LSB identifies the length of the
+	 * record in 16B units. Even values indicate 16B
+	 * records. Odd values indicate 32B
+	 * records.
+	 */
+	#define TX_CMPL_COAL_TYPE_MASK       UINT32_C(0x3f)
+	#define TX_CMPL_COAL_TYPE_SFT        0
+	/*
+	 * TX L2 coalesced completion:
+	 * Completion of TX packet. Length = 16B
+	 */
+	#define TX_CMPL_COAL_TYPE_TX_L2_COAL   UINT32_C(0x2)
+	#define TX_CMPL_COAL_TYPE_LAST        TX_CMPL_COAL_TYPE_TX_L2_COAL
+	#define TX_CMPL_COAL_FLAGS_MASK      UINT32_C(0xffc0)
+	#define TX_CMPL_COAL_FLAGS_SFT       6
+	/*
+	 * When this bit is '1', it indicates a packet that has an
+	 * error of some type. Type of error is indicated in
+	 * error_flags.
+	 */
+	#define TX_CMPL_COAL_FLAGS_ERROR      UINT32_C(0x40)
+	/*
+	 * When this bit is '1', it indicates that the packet completed
+	 * was transmitted using the push acceleration data provided
+	 * by the driver. When this bit is '0', it indicates that the
+	 * packet had not push acceleration data written or was executed
+	 * as a normal packet even though push data was provided.
+	 */
+	#define TX_CMPL_COAL_FLAGS_PUSH       UINT32_C(0x80)
+	/* unused1 is 16 b */
+	uint16_t	unused_0;
+	/*
+	 * This is a copy of the opaque field from the first TX BD of the packet
+	 * which corresponds with the reported sq_cons_idx. Note that, with
+	 * coalesced completions, completions are generated for only some of the
+	 * packets. The driver will see the opaque field for only those packets.
+	 * Note that, if the packet was described by a short CSO or short CSO
+	 * inline BD, then the 16-bit opaque field from the short CSO BD will
+	 * appear in the bottom 16 bits of this field. For TX rings with
+	 * completion coalescing enabled (which would use the coalesced
+	 * completion record), it is suggested that the driver populate the
+	 * opaque field to indicate the specific TX ring with which the
+	 * completion is associated, then utilize the opaque and sq_cons_idx
+	 * fields in the coalesced completion record to determine the specific
+	 * packets that are to be completed on that ring.
+	 */
+	uint32_t	opaque;
+	uint16_t	errors_v;
+	/*
+	 * This value is written by the NIC such that it will be different
+	 * for each pass through the completion queue. The even passes
+	 * will write 1. The odd passes will write 0.
+	 */
+	#define TX_CMPL_COAL_V                                  UINT32_C(0x1)
+	#define TX_CMPL_COAL_ERRORS_MASK \
+		UINT32_C(0xfffe)
+	#define TX_CMPL_COAL_ERRORS_SFT                         1
+	/*
+	 * This error indicates that there was some sort of problem
+	 * with the BDs for the packet.
+	 */
+	#define TX_CMPL_COAL_ERRORS_BUFFER_ERROR_MASK            UINT32_C(0xe)
+	#define TX_CMPL_COAL_ERRORS_BUFFER_ERROR_SFT             1
+	/* No error */
+	#define TX_CMPL_COAL_ERRORS_BUFFER_ERROR_NO_ERROR \
+		(UINT32_C(0x0) << 1)
+	/*
+	 * Bad Format:
+	 * BDs were not formatted correctly.
+	 */
+	#define TX_CMPL_COAL_ERRORS_BUFFER_ERROR_BAD_FMT \
+		(UINT32_C(0x2) << 1)
+	#define TX_CMPL_COAL_ERRORS_BUFFER_ERROR_LAST \
+		TX_CMPL_COAL_ERRORS_BUFFER_ERROR_BAD_FMT
+	/*
+	 * When this bit is '1', it indicates that the length of
+	 * the packet was zero. No packet was transmitted.
+	 */
+	#define TX_CMPL_COAL_ERRORS_ZERO_LENGTH_PKT              UINT32_C(0x10)
+	/*
+	 * When this bit is '1', it indicates that the packet
+	 * was longer than the programmed limit in TDI. No
+	 * packet was transmitted.
+	 */
+	#define TX_CMPL_COAL_ERRORS_EXCESSIVE_BD_LENGTH          UINT32_C(0x20)
+	/*
+	 * When this bit is '1', it indicates that one or more of the
+	 * BDs associated with this packet generated a PCI error.
+	 * This probably means the address was not valid.
+	 */
+	#define TX_CMPL_COAL_ERRORS_DMA_ERROR                    UINT32_C(0x40)
+	/*
+	 * When this bit is '1', it indicates that the packet was longer
+	 * than indicated by the hint. No packet was transmitted.
+	 */
+	#define TX_CMPL_COAL_ERRORS_HINT_TOO_SHORT               UINT32_C(0x80)
+	/*
+	 * When this bit is '1', it indicates that the packet was
+	 * dropped due to Poison TLP error on one or more of the
+	 * TLPs in the PXP completion.
+	 */
+	#define TX_CMPL_COAL_ERRORS_POISON_TLP_ERROR \
+		UINT32_C(0x100)
+	/*
+	 * When this bit is '1', it indicates that the packet was dropped
+	 * due to a transient internal error in TDC. The packet or LSO can
+	 * be retried and may transmit successfully on a subsequent attempt.
+	 */
+	#define TX_CMPL_COAL_ERRORS_INTERNAL_ERROR \
+		UINT32_C(0x200)
+	/*
+	 * When this bit is '1', it was not possible to collect a a timestamp
+	 * for a PTP completion, in which case the timestamp_hi and
+	 * timestamp_lo fields are invalid. When this bit is '0' for a PTP
+	 * completion, the timestamp_hi and timestamp_lo fields are valid.
+	 * RJRN will copy the value of this bit into the field of the same
+	 * name in all TX completions, regardless of whether such
+	 * completions are PTP completions or other TX completions.
+	 */
+	#define TX_CMPL_COAL_ERRORS_TIMESTAMP_INVALID_ERROR \
+		UINT32_C(0x400)
+	/* unused2 is 16 b */
+	uint16_t	unused_1;
+	uint32_t	sq_cons_idx;
+	/*
+	 * This value is SQ index for the start of the packet following the
+	 * last completed packet.
+	 */
+	#define TX_CMPL_COAL_SQ_CONS_IDX_MASK UINT32_C(0xffffff)
+	#define TX_CMPL_COAL_SQ_CONS_IDX_SFT 0
+} __rte_packed;
+
+/* tx_cmpl_ptp (size:128b/16B) */
+struct tx_cmpl_ptp {
+	uint16_t	flags_type;
+	/*
+	 * This field indicates the exact type of the completion.
+	 * By convention, the LSB identifies the length of the
+	 * record in 16B units. Even values indicate 16B
+	 * records. Odd values indicate 32B
+	 * records.
+	 */
+	#define TX_CMPL_PTP_TYPE_MASK       UINT32_C(0x3f)
+	#define TX_CMPL_PTP_TYPE_SFT        0
+	/*
+	 * TX L2 PTP completion:
+	 * Completion of TX packet. Length = 32B
+	 */
+	#define TX_CMPL_PTP_TYPE_TX_L2_PTP    UINT32_C(0x2)
+	#define TX_CMPL_PTP_TYPE_LAST        TX_CMPL_PTP_TYPE_TX_L2_PTP
+	#define TX_CMPL_PTP_FLAGS_MASK      UINT32_C(0xffc0)
+	#define TX_CMPL_PTP_FLAGS_SFT       6
+	/*
+	 * When this bit is '1', it indicates a packet that has an
+	 * error of some type. Type of error is indicated in
+	 * error_flags.
+	 */
+	#define TX_CMPL_PTP_FLAGS_ERROR      UINT32_C(0x40)
+	/*
+	 * When this bit is '1', it indicates that the packet completed
+	 * was transmitted using the push acceleration data provided
+	 * by the driver. When this bit is '0', it indicates that the
+	 * packet had not push acceleration data written or was executed
+	 * as a normal packet even though push data was provided.
+	 */
+	#define TX_CMPL_PTP_FLAGS_PUSH       UINT32_C(0x80)
+	/* unused1 is 16 b */
+	uint16_t	unused_0;
+	/*
+	 * This is a copy of the opaque field from the first TX BD of this
+	 * transmitted packet. Note that, if the packet was described by a short
+	 * CSO or short CSO inline BD, then the 16-bit opaque field from the
+	 * short CSO BD will appear in the bottom 16 bits of this field.
+	 */
+	uint32_t	opaque;
+	uint16_t	errors_v;
+	/*
+	 * This value is written by the NIC such that it will be different
+	 * for each pass through the completion queue. The even passes
+	 * will write 1. The odd passes will write 0.
+	 */
+	#define TX_CMPL_PTP_V                                  UINT32_C(0x1)
+	#define TX_CMPL_PTP_ERRORS_MASK                        UINT32_C(0xfffe)
+	#define TX_CMPL_PTP_ERRORS_SFT                         1
+	/*
+	 * This error indicates that there was some sort of problem
+	 * with the BDs for the packet.
+	 */
+	#define TX_CMPL_PTP_ERRORS_BUFFER_ERROR_MASK            UINT32_C(0xe)
+	#define TX_CMPL_PTP_ERRORS_BUFFER_ERROR_SFT             1
+	/* No error */
+	#define TX_CMPL_PTP_ERRORS_BUFFER_ERROR_NO_ERROR \
+		(UINT32_C(0x0) << 1)
+	/*
+	 * Bad Format:
+	 * BDs were not formatted correctly.
+	 */
+	#define TX_CMPL_PTP_ERRORS_BUFFER_ERROR_BAD_FMT \
+		(UINT32_C(0x2) << 1)
+	#define TX_CMPL_PTP_ERRORS_BUFFER_ERROR_LAST \
+		TX_CMPL_PTP_ERRORS_BUFFER_ERROR_BAD_FMT
+	/*
+	 * When this bit is '1', it indicates that the length of
+	 * the packet was zero. No packet was transmitted.
+	 */
+	#define TX_CMPL_PTP_ERRORS_ZERO_LENGTH_PKT              UINT32_C(0x10)
+	/*
+	 * When this bit is '1', it indicates that the packet
+	 * was longer than the programmed limit in TDI. No
+	 * packet was transmitted.
+	 */
+	#define TX_CMPL_PTP_ERRORS_EXCESSIVE_BD_LENGTH          UINT32_C(0x20)
+	/*
+	 * When this bit is '1', it indicates that one or more of the
+	 * BDs associated with this packet generated a PCI error.
+	 * This probably means the address was not valid.
+	 */
+	#define TX_CMPL_PTP_ERRORS_DMA_ERROR                    UINT32_C(0x40)
+	/*
+	 * When this bit is '1', it indicates that the packet was longer
+	 * than indicated by the hint. No packet was transmitted.
+	 */
+	#define TX_CMPL_PTP_ERRORS_HINT_TOO_SHORT               UINT32_C(0x80)
+	/*
+	 * When this bit is '1', it indicates that the packet was
+	 * dropped due to Poison TLP error on one or more of the
+	 * TLPs in the PXP completion.
+	 */
+	#define TX_CMPL_PTP_ERRORS_POISON_TLP_ERROR             UINT32_C(0x100)
+	/*
+	 * When this bit is '1', it indicates that the packet was dropped due
+	 * to a transient internal error in TDC. The packet or LSO can be
+	 * retried and may transmit successfully on a subsequent attempt.
+	 */
+	#define TX_CMPL_PTP_ERRORS_INTERNAL_ERROR               UINT32_C(0x200)
+	/*
+	 * When this bit is '1', it was not possible to collect a a timestamp
+	 * for a PTP completion, in which case the timestamp_hi and
+	 * timestamp_lo fields are invalid. When this bit is '0' for a PTP
+	 * completion, the timestamp_hi and timestamp_lo fields are valid.
+	 * RJRN will copy the value of this bit into the field of the same
+	 * name in all TX completions, regardless of whether such
+	 * completions are PTP completions or other TX completions.
+	 */
+	#define TX_CMPL_PTP_ERRORS_TIMESTAMP_INVALID_ERROR      UINT32_C(0x400)
+	/* unused2 is 16 b */
+	uint16_t	unused_1;
+	/*
+	 * This is timestamp value (lower 32bits) read from PM for the PTP
+	 * timestamp enabled packet.
+	 */
+	uint32_t	timestamp_lo;
+} __rte_packed;
+
+/* tx_cmpl_ptp_hi (size:128b/16B) */
+struct tx_cmpl_ptp_hi {
+	/*
+	 * This is timestamp value (lower 32bits) read from PM for the PTP
+	 * timestamp enabled packet.
+	 */
+	uint16_t	timestamp_hi[3];
+	uint16_t	reserved16;
+	uint64_t	v2;
+	/*
+	 * This value is written by the NIC such that it will be different for
+	 * each pass through the completion queue.The even passes will write 1.
+	 * The odd passes will write 0
+	 */
+	#define TX_CMPL_PTP_HI_V2     UINT32_C(0x1)
+} __rte_packed;
+
 /* rx_pkt_cmpl (size:128b/16B) */
 struct rx_pkt_cmpl {
 	uint16_t	flags_type;
@@ -3003,12 +3365,8 @@ struct rx_pkt_cmpl_hi {
 	#define RX_PKT_CMPL_REORDER_SFT 0
 } __rte_packed;
 
-/*
- * This TPA completion structure is used on devices where the
- * `hwrm_vnic_qcaps.max_aggs_supported` value is 0.
- */
-/* rx_tpa_start_cmpl (size:128b/16B) */
-struct rx_tpa_start_cmpl {
+/* rx_pkt_v2_cmpl (size:128b/16B) */
+struct rx_pkt_v2_cmpl {
 	uint16_t	flags_type;
 	/*
 	 * This field indicates the exact type of the completion.
@@ -3017,84 +3375,143 @@ struct rx_tpa_start_cmpl {
 	 * records. Odd values indicate 32B
 	 * records.
 	 */
-	#define RX_TPA_START_CMPL_TYPE_MASK                UINT32_C(0x3f)
-	#define RX_TPA_START_CMPL_TYPE_SFT                 0
+	#define RX_PKT_V2_CMPL_TYPE_MASK                      UINT32_C(0x3f)
+	#define RX_PKT_V2_CMPL_TYPE_SFT                       0
 	/*
-	 * RX L2 TPA Start Completion:
-	 * Completion at the beginning of a TPA operation.
-	 * Length = 32B
+	 * RX L2 V2 completion:
+	 * Completion of and L2 RX packet. Length = 32B
+	 * This is the new version of the RX_L2 completion used in SR2
+	 * and later chips.
 	 */
-	#define RX_TPA_START_CMPL_TYPE_RX_TPA_START          UINT32_C(0x13)
-	#define RX_TPA_START_CMPL_TYPE_LAST \
-		RX_TPA_START_CMPL_TYPE_RX_TPA_START
-	#define RX_TPA_START_CMPL_FLAGS_MASK               UINT32_C(0xffc0)
-	#define RX_TPA_START_CMPL_FLAGS_SFT                6
-	/* This bit will always be '0' for TPA start completions. */
-	#define RX_TPA_START_CMPL_FLAGS_ERROR               UINT32_C(0x40)
+	#define RX_PKT_V2_CMPL_TYPE_RX_L2_V2                    UINT32_C(0xf)
+	#define RX_PKT_V2_CMPL_TYPE_LAST \
+		RX_PKT_V2_CMPL_TYPE_RX_L2_V2
+	#define RX_PKT_V2_CMPL_FLAGS_MASK                     UINT32_C(0xffc0)
+	#define RX_PKT_V2_CMPL_FLAGS_SFT                      6
+	/*
+	 * When this bit is '1', it indicates a packet that has an
+	 * error of some type. Type of error is indicated in
+	 * error_flags.
+	 */
+	#define RX_PKT_V2_CMPL_FLAGS_ERROR                     UINT32_C(0x40)
 	/* This field indicates how the packet was placed in the buffer. */
-	#define RX_TPA_START_CMPL_FLAGS_PLACEMENT_MASK      UINT32_C(0x380)
-	#define RX_TPA_START_CMPL_FLAGS_PLACEMENT_SFT       7
+	#define RX_PKT_V2_CMPL_FLAGS_PLACEMENT_MASK            UINT32_C(0x380)
+	#define RX_PKT_V2_CMPL_FLAGS_PLACEMENT_SFT             7
+	/*
+	 * Normal:
+	 * Packet was placed using normal algorithm.
+	 */
+	#define RX_PKT_V2_CMPL_FLAGS_PLACEMENT_NORMAL \
+		(UINT32_C(0x0) << 7)
 	/*
 	 * Jumbo:
-	 * TPA Packet was placed using jumbo algorithm. This means
-	 * that the first buffer will be filled with data before
-	 * moving to aggregation buffers. Each aggregation buffer
-	 * will be filled before moving to the next aggregation
-	 * buffer.
+	 * Packet was placed using jumbo algorithm.
 	 */
-	#define RX_TPA_START_CMPL_FLAGS_PLACEMENT_JUMBO \
+	#define RX_PKT_V2_CMPL_FLAGS_PLACEMENT_JUMBO \
 		(UINT32_C(0x1) << 7)
 	/*
 	 * Header/Data Separation:
 	 * Packet was placed using Header/Data separation algorithm.
 	 * The separation location is indicated by the itype field.
 	 */
-	#define RX_TPA_START_CMPL_FLAGS_PLACEMENT_HDS \
+	#define RX_PKT_V2_CMPL_FLAGS_PLACEMENT_HDS \
 		(UINT32_C(0x2) << 7)
 	/*
-	 * GRO/Jumbo:
-	 * Packet will be placed using GRO/Jumbo where the first
-	 * packet is filled with data. Subsequent packets will be
-	 * placed such that any one packet does not span two
-	 * aggregation buffers unless it starts at the beginning of
-	 * an aggregation buffer.
+	 * Truncation:
+	 * Packet was placed using truncation algorithm. The
+	 * placed (truncated) length is indicated in the payload_offset
+	 * field. The original length is indicated in the len field.
 	 */
-	#define RX_TPA_START_CMPL_FLAGS_PLACEMENT_GRO_JUMBO \
-		(UINT32_C(0x5) << 7)
+	#define RX_PKT_V2_CMPL_FLAGS_PLACEMENT_TRUNCATION \
+		(UINT32_C(0x3) << 7)
+	#define RX_PKT_V2_CMPL_FLAGS_PLACEMENT_LAST \
+		RX_PKT_V2_CMPL_FLAGS_PLACEMENT_TRUNCATION
+	/* This bit is '1' if the RSS field in this completion is valid. */
+	#define RX_PKT_V2_CMPL_FLAGS_RSS_VALID                 UINT32_C(0x400)
 	/*
-	 * GRO/Header-Data Separation:
-	 * Packet will be placed using GRO/HDS where the header
-	 * is in the first packet.
-	 * Payload of each packet will be
-	 * placed such that any one packet does not span two
-	 * aggregation buffers unless it starts at the beginning of
-	 * an aggregation buffer.
+	 * This bit is '1' if metadata has been added to the end of the
+	 * packet in host memory. Metadata starts at the first 32B boundary
+	 * after the end of the packet for regular and jumbo placement.
+	 * It starts at the first 32B boundary after the end of the header
+	 * for HDS placement. The length of the metadata is indicated in the
+	 * metadata itself.
 	 */
-	#define RX_TPA_START_CMPL_FLAGS_PLACEMENT_GRO_HDS \
-		(UINT32_C(0x6) << 7)
-	#define RX_TPA_START_CMPL_FLAGS_PLACEMENT_LAST \
-		RX_TPA_START_CMPL_FLAGS_PLACEMENT_GRO_HDS
-	/* This bit is '1' if the RSS field in this completion is valid. */
-	#define RX_TPA_START_CMPL_FLAGS_RSS_VALID           UINT32_C(0x400)
-	/* unused is 1 b */
-	#define RX_TPA_START_CMPL_FLAGS_UNUSED              UINT32_C(0x800)
+	#define RX_PKT_V2_CMPL_FLAGS_PKT_METADATA_PRESENT      UINT32_C(0x800)
 	/*
 	 * This value indicates what the inner packet determined for the
 	 * packet was.
 	 */
-	#define RX_TPA_START_CMPL_FLAGS_ITYPE_MASK          UINT32_C(0xf000)
-	#define RX_TPA_START_CMPL_FLAGS_ITYPE_SFT           12
+	#define RX_PKT_V2_CMPL_FLAGS_ITYPE_MASK                UINT32_C(0xf000)
+	#define RX_PKT_V2_CMPL_FLAGS_ITYPE_SFT                 12
+	/*
+	 * Not Known:
+	 * Indicates that the packet type was not known.
+	 */
+	#define RX_PKT_V2_CMPL_FLAGS_ITYPE_NOT_KNOWN \
+		(UINT32_C(0x0) << 12)
+	/*
+	 * IP Packet:
+	 * Indicates that the packet was an IP packet, but further
+	 * classification was not possible.
+	 */
+	#define RX_PKT_V2_CMPL_FLAGS_ITYPE_IP \
+		(UINT32_C(0x1) << 12)
 	/*
 	 * TCP Packet:
 	 * Indicates that the packet was IP and TCP.
+	 * This indicates that the payload_offset field is valid.
 	 */
-	#define RX_TPA_START_CMPL_FLAGS_ITYPE_TCP \
+	#define RX_PKT_V2_CMPL_FLAGS_ITYPE_TCP \
 		(UINT32_C(0x2) << 12)
-	#define RX_TPA_START_CMPL_FLAGS_ITYPE_LAST \
-		RX_TPA_START_CMPL_FLAGS_ITYPE_TCP
 	/*
-	 * This value indicates the amount of packet data written to the
-	 * buffer the opaque field in this completion corresponds to.
+	 * UDP Packet:
+	 * Indicates that the packet was IP and UDP.
+	 * This indicates that the payload_offset field is valid.
+	 */
+	#define RX_PKT_V2_CMPL_FLAGS_ITYPE_UDP \
+		(UINT32_C(0x3) << 12)
+	/*
+	 * FCoE Packet:
+	 * Indicates that the packet was recognized as a FCoE.
+	 * This also indicates that the payload_offset field is valid.
+	 */
+	#define RX_PKT_V2_CMPL_FLAGS_ITYPE_FCOE \
+		(UINT32_C(0x4) << 12)
+	/*
+	 * RoCE Packet:
+	 * Indicates that the packet was recognized as a RoCE.
+	 * This also indicates that the payload_offset field is valid.
+	 */
+	#define RX_PKT_V2_CMPL_FLAGS_ITYPE_ROCE \
+		(UINT32_C(0x5) << 12)
+	/*
+	 * ICMP Packet:
+	 * Indicates that the packet was recognized as ICMP.
+	 * This indicates that the payload_offset field is valid.
+	 */
+	#define RX_PKT_V2_CMPL_FLAGS_ITYPE_ICMP \
+		(UINT32_C(0x7) << 12)
+	/*
+	 * PtP packet wo/timestamp:
+	 * Indicates that the packet was recognized as a PtP
+	 * packet.
+	 */
+	#define RX_PKT_V2_CMPL_FLAGS_ITYPE_PTP_WO_TIMESTAMP \
+		(UINT32_C(0x8) << 12)
+	/*
+	 * PtP packet w/timestamp:
+	 * Indicates that the packet was recognized as a PtP
+	 * packet and that a timestamp was taken for the packet.
+	 */
+	#define RX_PKT_V2_CMPL_FLAGS_ITYPE_PTP_W_TIMESTAMP \
+		(UINT32_C(0x9) << 12)
+	#define RX_PKT_V2_CMPL_FLAGS_ITYPE_LAST \
+		RX_PKT_V2_CMPL_FLAGS_ITYPE_PTP_W_TIMESTAMP
+	/*
+	 * This is the length of the data for the packet stored in the
+	 * buffer(s) identified by the opaque value. This includes
+	 * the packet BD and any associated buffer BDs. This does not include
+	 * the length of any data places in aggregation BDs.
 	 */
 	uint16_t	len;
 	/*
@@ -3102,19 +3519,597 @@ struct rx_tpa_start_cmpl {
 	 * corresponds to.
 	 */
 	uint32_t	opaque;
+	uint8_t	agg_bufs_v1;
 	/*
 	 * This value is written by the NIC such that it will be different
 	 * for each pass through the completion queue. The even passes
 	 * will write 1. The odd passes will write 0.
 	 */
-	uint8_t	v1;
+	#define RX_PKT_V2_CMPL_V1           UINT32_C(0x1)
 	/*
-	 * This value is written by the NIC such that it will be different
-	 * for each pass through the completion queue. The even passes
-	 * will write 1. The odd passes will write 0.
+	 * This value is the number of aggregation buffers that follow this
+	 * entry in the completion ring that are a part of this packet.
+	 * If the value is zero, then the packet is completely contained
+	 * in the buffer space provided for the packet in the RX ring.
 	 */
-	#define RX_TPA_START_CMPL_V1 UINT32_C(0x1)
-	#define RX_TPA_START_CMPL_LAST RX_TPA_START_CMPL_V1
+	#define RX_PKT_V2_CMPL_AGG_BUFS_MASK UINT32_C(0x3e)
+	#define RX_PKT_V2_CMPL_AGG_BUFS_SFT 1
+	/* unused1 is 2 b */
+	#define RX_PKT_V2_CMPL_UNUSED1_MASK UINT32_C(0xc0)
+	#define RX_PKT_V2_CMPL_UNUSED1_SFT  6
+	/*
+	 * This is the RSS hash type for the packet. The value is packed
+	 * {tuple_extrac_op[1:0],rss_profile_id[4:0],tuple_extrac_op[2]}.
+	 *
+	 * The value of tuple_extrac_op provides the information about
+	 * what fields the hash was computed on.
+	 * * 0: The RSS hash was computed over source IP address,
+	 * destination IP address, source port, and destination port of inner
+	 * IP and TCP or UDP headers. Note: For non-tunneled packets,
+	 * the packet headers are considered inner packet headers for the RSS
+	 * hash computation purpose.
+	 * * 1: The RSS hash was computed over source IP address and destination
+	 * IP address of inner IP header. Note: For non-tunneled packets,
+	 * the packet headers are considered inner packet headers for the RSS
+	 * hash computation purpose.
+	 * * 2: The RSS hash was computed over source IP address,
+	 * destination IP address, source port, and destination port of
+	 * IP and TCP or UDP headers of outer tunnel headers.
+	 * Note: For non-tunneled packets, this value is not applicable.
+	 * * 3: The RSS hash was computed over source IP address and
+	 * destination IP address of IP header of outer tunnel headers.
+	 * Note: For non-tunneled packets, this value is not applicable.
+	 *
+	 * Note that 4-tuples values listed above are applicable
+	 * for layer 4 protocols supported and enabled for RSS in the hardware,
+	 * HWRM firmware, and drivers. For example, if RSS hash is supported and
+	 * enabled for TCP traffic only, then the values of tuple_extract_op
+	 * corresponding to 4-tuples are only valid for TCP traffic.
+	 */
+	uint8_t	rss_hash_type;
+	uint16_t	metadata1_payload_offset;
+	/*
+	 * This is data from the CFA as indicated by the meta_format field.
+	 * If truncation placement is not used, this value indicates the offset
+	 * in bytes from the beginning of the packet where the inner payload
+	 * starts. This value is valid for TCP, UDP, FCoE, and RoCE packets. If
+	 * truncation placement is used, this value represents the placed
+	 * (truncated) length of the packet.
+	 */
+	#define RX_PKT_V2_CMPL_PAYLOAD_OFFSET_MASK    UINT32_C(0x1ff)
+	#define RX_PKT_V2_CMPL_PAYLOAD_OFFSET_SFT     0
+	/* This is data from the CFA as indicated by the meta_format field. */
+	#define RX_PKT_V2_CMPL_METADATA1_MASK         UINT32_C(0xf000)
+	#define RX_PKT_V2_CMPL_METADATA1_SFT          12
+	/* When meta_format != 0, this value is the VLAN TPID_SEL. */
+	#define RX_PKT_V2_CMPL_METADATA1_TPID_SEL_MASK UINT32_C(0x7000)
+	#define RX_PKT_V2_CMPL_METADATA1_TPID_SEL_SFT  12
+	/* When meta_format != 0, this value is the VLAN TPID_SEL. */
+	#define RX_PKT_V2_CMPL_METADATA1_VALID         UINT32_C(0x8000)
+	/*
+	 * This value is the RSS hash value calculated for the packet
+	 * based on the mode bits and key value in the VNIC. When vee_cmpl_mode
+	 * is set in VNIC context, this is the lower 32b of the host address
+	 * from the first BD used to place the packet.
+	 */
+	uint32_t	rss_hash;
+} __rte_packed;
+
+/* Last 16 bytes of RX Packet V2 Completion Record */
+/* rx_pkt_v2_cmpl_hi (size:128b/16B) */
+struct rx_pkt_v2_cmpl_hi {
+	uint32_t	flags2;
+	/*
+	 * When this bit is '0', the cs_ok field has the following definition:-
+	 * ip_cs_ok[2:0] = The number of header groups with a valid IP checksum
+	 * in the delivered packet, counted from the outer-most header group to
+	 * the inner-most header group, stopping at the first error. -
+	 * l4_cs_ok[5:3] = The number of header groups with a valid L4 checksum
+	 * in the delivered packet, counted from the outer-most header group to
+	 * the inner-most header group, stopping at the first error. When this
+	 * bit is '1', the cs_ok field has the following definition: -
+	 * hdr_cnt[2:0] = The number of header groups that were parsed by the
+	 * chip and passed in the delivered packet. - ip_cs_all_ok[3] =This bit
+	 * will be '1' if all the parsed header groups with an IP checksum are
+	 * valid. - l4_cs_all_ok[4] = This bit will be '1' if all the parsed
+	 * header groups with an L4 checksum are valid.
+	 */
+	#define RX_PKT_V2_CMPL_HI_FLAGS2_CS_ALL_OK_MODE \
+		UINT32_C(0x8)
+	/* This value indicates what format the metadata field is. */
+	#define RX_PKT_V2_CMPL_HI_FLAGS2_META_FORMAT_MASK \
+		UINT32_C(0xf0)
+	#define RX_PKT_V2_CMPL_HI_FLAGS2_META_FORMAT_SFT            4
+	/* There is no metadata information. Values are zero. */
+	#define RX_PKT_V2_CMPL_HI_FLAGS2_META_FORMAT_NONE \
+		(UINT32_C(0x0) << 4)
+	/*
+	 * The {metadata1, metadata0} fields contain the vtag
+	 * information: - vtag[19:0] = {valid, tpid_sel[2:0], pri[2:0],
+	 * de, vid[11:0]} The metadata2 field contains the table scope
+	 * and action record pointer. - metadata2[25:0] contains the
+	 * action record pointer. - metadata2[31:26] contains the table
+	 * scope.
+	 */
+	#define RX_PKT_V2_CMPL_HI_FLAGS2_META_FORMAT_ACT_REC_PTR \
+		(UINT32_C(0x1) << 4)
+	/*
+	 * The {metadata1, metadata0} fields contain the vtag
+	 * information:
+	 * - vtag[19:0] = {valid, tpid_sel[2:0], pri[2:0], de, vid[11:0]}
+	 * The metadata2 field contains the Tunnel ID
+	 * value, justified to LSB. i
+	 * - VXLAN = VNI[23:0] -> VXLAN Network ID
+	 * - Geneve (NGE) = VNI[23:0] a-> Virtual Network Identifier
+	 * - NVGRE = TNI[23:0] -> Tenant Network ID
+	 * - GRE = KEY[31:0] -> key field with bit mask. zero if K=0
+	 * - IPv4 = 0 (not populated)
+	 * - IPv6 = Flow Label[19:0]
+	 * - PPPoE = sessionID[15:0]
+	 * - MPLs = Outer label[19:0]
+	 * - UPAR = Selected[31:0] with bit mask
+	 */
+	#define RX_PKT_V2_CMPL_HI_FLAGS2_META_FORMAT_TUNNEL_ID \
+		(UINT32_C(0x2) << 4)
+	/*
+	 * The {metadata1, metadata0} fields contain the vtag
+	 * information:
+	 * - vtag[19:0] = {valid, tpid_sel[2:0], pri[2:0],de, vid[11:0]}
+	 * The metadata2 field contains the 32b metadata from the prepended
+	 * header (chdr_data).
+	 */
+	#define RX_PKT_V2_CMPL_HI_FLAGS2_META_FORMAT_CHDR_DATA \
+		(UINT32_C(0x3) << 4)
+	/*
+	 * The {metadata1, metadata0} fields contain the vtag
+	 * information:
+	 * - vtag[19:0] = {valid, tpid_sel[2:0], pri[2:0], de, vid[11:0]}
+	 * The metadata2 field contains the outer_l3_offset,
+	 * inner_l2_offset, inner_l3_offset, and inner_l4_size.
+	 * - metadata2[8:0] contains the outer_l3_offset.
+	 * - metadata2[17:9] contains the inner_l2_offset.
+	 * - metadata2[26:18] contains the inner_l3_offset.
+	 * - metadata2[31:27] contains the inner_l4_size.
+	 */
+	#define RX_PKT_V2_CMPL_HI_FLAGS2_META_FORMAT_HDR_OFFSET \
+		(UINT32_C(0x4) << 4)
+	#define RX_PKT_V2_CMPL_HI_FLAGS2_META_FORMAT_LAST \
+		RX_PKT_V2_CMPL_HI_FLAGS2_META_FORMAT_HDR_OFFSET
+	/*
+	 * This field indicates the IP type for the inner-most IP header.
+	 * A value of '0' indicates IPv4. A value of '1' indicates IPv6.
+	 * This value is only valid if itype indicates a packet
+	 * with an IP header.
+	 */
+	#define RX_PKT_V2_CMPL_HI_FLAGS2_IP_TYPE \
+		UINT32_C(0x100)
+	/*
+	 * This indicates that the complete 1's complement checksum was
+	 * calculated for the packet.
+	 */
+	#define RX_PKT_V2_CMPL_HI_FLAGS2_COMPLETE_CHECKSUM_CALC \
+		UINT32_C(0x200)
+	/*
+	 * This field indicates the status of IP and L4 CS calculations done
+	 * by the chip. The format of this field is indicated by the
+	 * cs_all_ok_mode bit.
+	 */
+	#define RX_PKT_V2_CMPL_HI_FLAGS2_CS_OK_MASK \
+		UINT32_C(0xfc00)
+	#define RX_PKT_V2_CMPL_HI_FLAGS2_CS_OK_SFT                  10
+	/*
+	 * This value is the complete 1's complement checksum calculated from
+	 * the start of the outer L3 header to the end of the packet (not
+	 * including the ethernet crc). It is valid when the
+	 * 'complete_checksum_calc' flag is set.
+	 */
+	#define RX_PKT_V2_CMPL_HI_FLAGS2_COMPLETE_CHECKSUM_MASK \
+		UINT32_C(0xffff0000)
+	#define RX_PKT_V2_CMPL_HI_FLAGS2_COMPLETE_CHECKSUM_SFT      16
+	/*
+	 * This is data from the CFA block as indicated by the meta_format
+	 * field.
+	 * - meta_format 0 - none - metadata2 = 0 - not valid/not stripped
+	 * - meta_format 1 - act_rec_ptr - metadata2 = {table_scope[5:0],
+	 *   act_rec_ptr[25:0]}
+	 * - meta_format 2 - tunnel_id - metadata2 = tunnel_id[31:0]
+	 * - meta_format 3 - chdr_data - metadata2 = updated_chdr_data[31:0]
+	 * - meta_format 4 - hdr_offsets - metadata2 = hdr_offsets[31:0]
+	 * When vee_cmpl_mode is set in VNIC context, this is the upper 32b
+	 * of the host address from the first BD used to place the packet.
+	 */
+	uint32_t	metadata2;
+	uint16_t	errors_v2;
+	/*
+	 * This value is written by the NIC such that it will be different
+	 * for each pass through the completion queue. The even passes
+	 * will write 1. The odd passes will write 0.
+	 */
+	#define RX_PKT_V2_CMPL_HI_V2 \
+		UINT32_C(0x1)
+	#define RX_PKT_V2_CMPL_HI_ERRORS_MASK \
+		UINT32_C(0xfffe)
+	#define RX_PKT_V2_CMPL_HI_ERRORS_SFT                               1
+	/*
+	 * This error indicates that there was some sort of problem with
+	 * the BDs for the packet that was found after part of the
+	 * packet was already placed. The packet should be treated as
+	 * invalid.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_BUFFER_ERROR_MASK \
+		UINT32_C(0xe)
+	#define RX_PKT_V2_CMPL_HI_ERRORS_BUFFER_ERROR_SFT                   1
+	/* No buffer error */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_BUFFER_ERROR_NO_BUFFER \
+		(UINT32_C(0x0) << 1)
+	/*
+	 * Did Not Fit: Packet did not fit into packet buffer provided.
+	 * For regular placement, this means the packet did not fit in
+	 * the buffer provided. For HDS and jumbo placement, this means
+	 * that the packet could not be placed into 8 physical buffers
+	 * (if fixed-size buffers are used), or that the packet could
+	 * not be placed in the number of physical buffers configured
+	 * for the VNIC (if variable-size buffers are used)
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_BUFFER_ERROR_DID_NOT_FIT \
+		(UINT32_C(0x1) << 1)
+	/*
+	 * Not On Chip: All BDs needed for the packet were not on-chip
+	 * when the packet arrived. For regular placement, this error is
+	 * not valid. For HDS and jumbo placement, this means that not
+	 * enough agg BDs were posted to place the packet.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_BUFFER_ERROR_NOT_ON_CHIP \
+		(UINT32_C(0x2) << 1)
+	/*
+	 * Bad Format:
+	 * BDs were not formatted correctly.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_BUFFER_ERROR_BAD_FORMAT \
+		(UINT32_C(0x3) << 1)
+	/*
+	 * Flush:
+	 * There was a bad_format error on the previous operation
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_BUFFER_ERROR_FLUSH \
+		(UINT32_C(0x5) << 1)
+	#define RX_PKT_V2_CMPL_HI_ERRORS_BUFFER_ERROR_LAST \
+		RX_PKT_V2_CMPL_HI_ERRORS_BUFFER_ERROR_FLUSH
+	/*
+	 * This indicates that there was an error in the outer tunnel
+	 * portion of the packet when this field is non-zero.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_OT_PKT_ERROR_MASK \
+		UINT32_C(0x70)
+	#define RX_PKT_V2_CMPL_HI_ERRORS_OT_PKT_ERROR_SFT                   4
+	/*
+	 * No additional error occurred on the outer tunnel portion
+	 * of the packet or the packet does not have a outer tunnel.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_OT_PKT_ERROR_NO_ERROR \
+		(UINT32_C(0x0) << 4)
+	/*
+	 * Indicates that IP header version does not match expectation
+	 * from L2 Ethertype for IPv4 and IPv6 in the outer tunnel header.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_OT_PKT_ERROR_OT_L3_BAD_VERSION \
+		(UINT32_C(0x1) << 4)
+	/*
+	 * Indicates that header length is out of range in the outer
+	 * tunnel header. Valid for IPv4.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_OT_PKT_ERROR_OT_L3_BAD_HDR_LEN \
+		(UINT32_C(0x2) << 4)
+	/*
+	 * Indicates that physical packet is shorter than that claimed
+	 * by the outer tunnel l3 header length. Valid for IPv4, or
+	 * IPv6 outer tunnel packets.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_OT_PKT_ERROR_OT_IP_TOTAL_ERROR \
+		(UINT32_C(0x3) << 4)
+	/*
+	 * Indicates that the physical packet is shorter than that
+	 * claimed by the outer tunnel UDP header length for a outer
+	 * tunnel UDP packet that is not fragmented.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_OT_PKT_ERROR_OT_UDP_TOTAL_ERROR \
+		(UINT32_C(0x4) << 4)
+	/*
+	 * Indicates that the IPv4 TTL or IPv6 hop limit check have
+	 * failed (e.g. TTL = 0) in the outer tunnel header. Valid for
+	 * IPv4, and IPv6.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_OT_PKT_ERROR_OT_L3_BAD_TTL \
+		(UINT32_C(0x5) << 4)
+	/*
+	 * Indicates that the IP checksum failed its check in the outer
+	 * tunnel header.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_OT_PKT_ERROR_OT_IP_CS_ERROR \
+		(UINT32_C(0x6) << 4)
+	/*
+	 * Indicates that the L4 checksum failed its check in the outer
+	 * tunnel header.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_OT_PKT_ERROR_OT_L4_CS_ERROR \
+		(UINT32_C(0x7) << 4)
+	#define RX_PKT_V2_CMPL_HI_ERRORS_OT_PKT_ERROR_LAST \
+		RX_PKT_V2_CMPL_HI_ERRORS_OT_PKT_ERROR_OT_L4_CS_ERROR
+	/*
+	 * This indicates that there was a CRC error on either an FCoE
+	 * or RoCE packet. The itype indicates the packet type.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_CRC_ERROR \
+		UINT32_C(0x100)
+	/*
+	 * This indicates that there was an error in the tunnel portion
+	 * of the packet when this field is non-zero.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_T_PKT_ERROR_MASK \
+		UINT32_C(0xe00)
+	#define RX_PKT_V2_CMPL_HI_ERRORS_T_PKT_ERROR_SFT                    9
+	/*
+	 * No additional error occurred on the tunnel portion
+	 * of the packet or the packet does not have a tunnel.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_T_PKT_ERROR_NO_ERROR \
+		(UINT32_C(0x0) << 9)
+	/*
+	 * Indicates that IP header version does not match expectation
+	 * from L2 Ethertype for IPv4 and IPv6 in the tunnel header.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_T_PKT_ERROR_T_L3_BAD_VERSION \
+		(UINT32_C(0x1) << 9)
+	/*
+	 * Indicates that header length is out of range in the tunnel
+	 * header. Valid for IPv4.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_T_PKT_ERROR_T_L3_BAD_HDR_LEN \
+		(UINT32_C(0x2) << 9)
+	/*
+	 * Indicates that physical packet is shorter than that claimed
+	 * by the tunnel l3 header length. Valid for IPv4, or IPv6 tunnel
+	 * packet packets.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_T_PKT_ERROR_T_IP_TOTAL_ERROR \
+		(UINT32_C(0x3) << 9)
+	/*
+	 * Indicates that the physical packet is shorter than that claimed
+	 * by the tunnel UDP header length for a tunnel UDP packet that is
+	 * not fragmented.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_T_PKT_ERROR_T_UDP_TOTAL_ERROR \
+		(UINT32_C(0x4) << 9)
+	/*
+	 * Indicates that the IPv4 TTL or IPv6 hop limit check have failed
+	 * (e.g. TTL = 0) in the tunnel header. Valid for IPv4, and IPv6.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_T_PKT_ERROR_T_L3_BAD_TTL \
+		(UINT32_C(0x5) << 9)
+	/*
+	 * Indicates that the IP checksum failed its check in the tunnel
+	 * header.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_T_PKT_ERROR_T_IP_CS_ERROR \
+		(UINT32_C(0x6) << 9)
+	/*
+	 * Indicates that the L4 checksum failed its check in the tunnel
+	 * header.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_T_PKT_ERROR_T_L4_CS_ERROR \
+		(UINT32_C(0x7) << 9)
+	#define RX_PKT_V2_CMPL_HI_ERRORS_T_PKT_ERROR_LAST \
+		RX_PKT_V2_CMPL_HI_ERRORS_T_PKT_ERROR_T_L4_CS_ERROR
+	/*
+	 * This indicates that there was an error in the inner
+	 * portion of the packet when this
+	 * field is non-zero.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_PKT_ERROR_MASK \
+		UINT32_C(0xf000)
+	#define RX_PKT_V2_CMPL_HI_ERRORS_PKT_ERROR_SFT                      12
+	/*
+	 * No additional error occurred on the tunnel portion
+	 * or the packet of the packet does not have a tunnel.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_PKT_ERROR_NO_ERROR \
+		(UINT32_C(0x0) << 12)
+	/*
+	 * Indicates that IP header version does not match
+	 * expectation from L2 Ethertype for IPv4 and IPv6 or that
+	 * option other than VFT was parsed on
+	 * FCoE packet.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_PKT_ERROR_L3_BAD_VERSION \
+		(UINT32_C(0x1) << 12)
+	/*
+	 * indicates that header length is out of range. Valid for
+	 * IPv4 and RoCE
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_PKT_ERROR_L3_BAD_HDR_LEN \
+		(UINT32_C(0x2) << 12)
+	/*
+	 * indicates that the IPv4 TTL or IPv6 hop limit check
+	 * have failed (e.g. TTL = 0). Valid for IPv4, and IPv6
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_PKT_ERROR_L3_BAD_TTL \
+		(UINT32_C(0x3) << 12)
+	/*
+	 * Indicates that physical packet is shorter than that
+	 * claimed by the l3 header length. Valid for IPv4,
+	 * IPv6 packet or RoCE packets.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_PKT_ERROR_IP_TOTAL_ERROR \
+		(UINT32_C(0x4) << 12)
+	/*
+	 * Indicates that the physical packet is shorter than that
+	 * claimed by the UDP header length for a UDP packet that is
+	 * not fragmented.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_PKT_ERROR_UDP_TOTAL_ERROR \
+		(UINT32_C(0x5) << 12)
+	/*
+	 * Indicates that TCP header length > IP payload. Valid for
+	 * TCP packets only.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_PKT_ERROR_L4_BAD_HDR_LEN \
+		(UINT32_C(0x6) << 12)
+	/* Indicates that TCP header length < 5. Valid for TCP. */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_PKT_ERROR_L4_BAD_HDR_LEN_TOO_SMALL \
+		(UINT32_C(0x7) << 12)
+	/*
+	 * Indicates that TCP option headers result in a TCP header
+	 * size that does not match data offset in TCP header. Valid
+	 * for TCP.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_PKT_ERROR_L4_BAD_OPT_LEN \
+		(UINT32_C(0x8) << 12)
+	/*
+	 * Indicates that the IP checksum failed its check in the
+	 * inner header.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_PKT_ERROR_IP_CS_ERROR \
+		(UINT32_C(0x9) << 12)
+	/*
+	 * Indicates that the L4 checksum failed its check in the
+	 * inner header.
+	 */
+	#define RX_PKT_V2_CMPL_HI_ERRORS_PKT_ERROR_L4_CS_ERROR \
+		(UINT32_C(0xa) << 12)
+	#define RX_PKT_V2_CMPL_HI_ERRORS_PKT_ERROR_LAST \
+		RX_PKT_V2_CMPL_HI_ERRORS_PKT_ERROR_L4_CS_ERROR
+	/*
+	 * This is data from the CFA block as indicated by the meta_format
+	 * field.
+	 */
+	uint16_t	metadata0;
+	/* When meta_format=1, this value is the VLAN VID. */
+	#define RX_PKT_V2_CMPL_HI_METADATA0_VID_MASK UINT32_C(0xfff)
+	#define RX_PKT_V2_CMPL_HI_METADATA0_VID_SFT 0
+	/* When meta_format=1, this value is the VLAN DE. */
+	#define RX_PKT_V2_CMPL_HI_METADATA0_DE      UINT32_C(0x1000)
+	/* When meta_format=1, this value is the VLAN PRI. */
+	#define RX_PKT_V2_CMPL_HI_METADATA0_PRI_MASK UINT32_C(0xe000)
+	#define RX_PKT_V2_CMPL_HI_METADATA0_PRI_SFT 13
+	/*
+	 * The timestamp field contains the 32b timestamp for the packet from
+	 * the MAC.
+	 */
+	uint32_t	timestamp;
+} __rte_packed;
+
+/*
+ * This TPA completion structure is used on devices where the
+ * `hwrm_vnic_qcaps.max_aggs_supported` value is 0.
+ */
+/* rx_tpa_start_cmpl (size:128b/16B) */
+struct rx_tpa_start_cmpl {
+	uint16_t	flags_type;
+	/*
+	 * This field indicates the exact type of the completion.
+	 * By convention, the LSB identifies the length of the
+	 * record in 16B units. Even values indicate 16B
+	 * records. Odd values indicate 32B
+	 * records.
+	 */
+	#define RX_TPA_START_CMPL_TYPE_MASK                UINT32_C(0x3f)
+	#define RX_TPA_START_CMPL_TYPE_SFT                 0
+	/*
+	 * RX L2 TPA Start Completion:
+	 * Completion at the beginning of a TPA operation.
+	 * Length = 32B
+	 */
+	#define RX_TPA_START_CMPL_TYPE_RX_TPA_START          UINT32_C(0x13)
+	#define RX_TPA_START_CMPL_TYPE_LAST \
+		RX_TPA_START_CMPL_TYPE_RX_TPA_START
+	#define RX_TPA_START_CMPL_FLAGS_MASK               UINT32_C(0xffc0)
+	#define RX_TPA_START_CMPL_FLAGS_SFT                6
+	/* This bit will always be '0' for TPA start completions. */
+	#define RX_TPA_START_CMPL_FLAGS_ERROR               UINT32_C(0x40)
+	/* This field indicates how the packet was placed in the buffer. */
+	#define RX_TPA_START_CMPL_FLAGS_PLACEMENT_MASK      UINT32_C(0x380)
+	#define RX_TPA_START_CMPL_FLAGS_PLACEMENT_SFT       7
+	/*
+	 * Jumbo:
+	 * TPA Packet was placed using jumbo algorithm. This means
+	 * that the first buffer will be filled with data before
+	 * moving to aggregation buffers. Each aggregation buffer
+	 * will be filled before moving to the next aggregation
+	 * buffer.
+	 */
+	#define RX_TPA_START_CMPL_FLAGS_PLACEMENT_JUMBO \
+		(UINT32_C(0x1) << 7)
+	/*
+	 * Header/Data Separation:
+	 * Packet was placed using Header/Data separation algorithm.
+	 * The separation location is indicated by the itype field.
+	 */
+	#define RX_TPA_START_CMPL_FLAGS_PLACEMENT_HDS \
+		(UINT32_C(0x2) << 7)
+	/*
+	 * GRO/Jumbo:
+	 * Packet will be placed using GRO/Jumbo where the first
+	 * packet is filled with data. Subsequent packets will be
+	 * placed such that any one packet does not span two
+	 * aggregation buffers unless it starts at the beginning of
+	 * an aggregation buffer.
+	 */
+	#define RX_TPA_START_CMPL_FLAGS_PLACEMENT_GRO_JUMBO \
+		(UINT32_C(0x5) << 7)
+	/*
+	 * GRO/Header-Data Separation:
+	 * Packet will be placed using GRO/HDS where the header
+	 * is in the first packet.
+	 * Payload of each packet will be
+	 * placed such that any one packet does not span two
+	 * aggregation buffers unless it starts at the beginning of
+	 * an aggregation buffer.
+	 */
+	#define RX_TPA_START_CMPL_FLAGS_PLACEMENT_GRO_HDS \
+		(UINT32_C(0x6) << 7)
+	#define RX_TPA_START_CMPL_FLAGS_PLACEMENT_LAST \
+		RX_TPA_START_CMPL_FLAGS_PLACEMENT_GRO_HDS
+	/* This bit is '1' if the RSS field in this completion is valid. */
+	#define RX_TPA_START_CMPL_FLAGS_RSS_VALID           UINT32_C(0x400)
+	/* unused is 1 b */
+	#define RX_TPA_START_CMPL_FLAGS_UNUSED              UINT32_C(0x800)
+	/*
+	 * This value indicates what the inner packet determined for the
+	 * packet was.
+	 */
+	#define RX_TPA_START_CMPL_FLAGS_ITYPE_MASK          UINT32_C(0xf000)
+	#define RX_TPA_START_CMPL_FLAGS_ITYPE_SFT           12
+	/*
+	 * TCP Packet:
+	 * Indicates that the packet was IP and TCP.
+	 */
+	#define RX_TPA_START_CMPL_FLAGS_ITYPE_TCP \
+		(UINT32_C(0x2) << 12)
+	#define RX_TPA_START_CMPL_FLAGS_ITYPE_LAST \
+		RX_TPA_START_CMPL_FLAGS_ITYPE_TCP
+	/*
+	 * This value indicates the amount of packet data written to the
+	 * buffer the opaque field in this completion corresponds to.
+	 */
+	uint16_t	len;
+	/*
+	 * This is a copy of the opaque field from the RX BD this completion
+	 * corresponds to.
+	 */
+	uint32_t	opaque;
+	/*
+	 * This value is written by the NIC such that it will be different
+	 * for each pass through the completion queue. The even passes
+	 * will write 1. The odd passes will write 0.
+	 */
+	uint8_t	v1;
+	/*
+	 * This value is written by the NIC such that it will be different
+	 * for each pass through the completion queue. The even passes
+	 * will write 1. The odd passes will write 0.
+	 */
+	#define RX_TPA_START_CMPL_V1 UINT32_C(0x1)
+	#define RX_TPA_START_CMPL_LAST RX_TPA_START_CMPL_V1
 	/*
 	 * This is the RSS hash type for the packet. The value is packed
 	 * {tuple_extrac_op[1:0],rss_profile_id[4:0],tuple_extrac_op[2]}.
@@ -3288,6 +4283,430 @@ struct rx_tpa_start_cmpl_hi {
 /*
  * This TPA completion structure is used on devices where the
  * `hwrm_vnic_qcaps.max_aggs_supported` value is 0.
+ * RX L2 TPA Start V2 Completion Record (32 bytes split to 2 16-byte
+ * struct)
+ */
+/* rx_tpa_start_v2_cmpl (size:128b/16B) */
+struct rx_tpa_start_v2_cmpl {
+	uint16_t	flags_type;
+	/*
+	 * This field indicates the exact type of the completion.
+	 * By convention, the LSB identifies the length of the
+	 * record in 16B units. Even values indicate 16B
+	 * records. Odd values indicate 32B
+	 * records.
+	 */
+	#define RX_TPA_START_V2_CMPL_TYPE_MASK \
+		UINT32_C(0x3f)
+	#define RX_TPA_START_V2_CMPL_TYPE_SFT                       0
+	/*
+	 * RX L2 TPA Start V2 Completion:
+	 * Completion at the beginning of a TPA operation.
+	 * Length = 32B
+	 * This is the new version of the RX_TPA_START completion used
+	 * in SR2 and later chips.
+	 */
+	#define RX_TPA_START_V2_CMPL_TYPE_RX_TPA_START_V2 \
+		UINT32_C(0xd)
+	#define RX_TPA_START_V2_CMPL_TYPE_LAST \
+		RX_TPA_START_V2_CMPL_TYPE_RX_TPA_START_V2
+	#define RX_TPA_START_V2_CMPL_FLAGS_MASK \
+		UINT32_C(0xffc0)
+	#define RX_TPA_START_V2_CMPL_FLAGS_SFT                      6
+	/*
+	 * When this bit is '1', it indicates a packet that has an error
+	 * of some type. Type of error is indicated in error_flags.
+	 */
+	#define RX_TPA_START_V2_CMPL_FLAGS_ERROR \
+		UINT32_C(0x40)
+	/* This field indicates how the packet was placed in the buffer. */
+	#define RX_TPA_START_V2_CMPL_FLAGS_PLACEMENT_MASK \
+		UINT32_C(0x380)
+	#define RX_TPA_START_V2_CMPL_FLAGS_PLACEMENT_SFT             7
+	/*
+	 * Jumbo:
+	 * TPA Packet was placed using jumbo algorithm. This means
+	 * that the first buffer will be filled with data before
+	 * moving to aggregation buffers. Each aggregation buffer
+	 * will be filled before moving to the next aggregation
+	 * buffer.
+	 */
+	#define RX_TPA_START_V2_CMPL_FLAGS_PLACEMENT_JUMBO \
+		(UINT32_C(0x1) << 7)
+	/*
+	 * Header/Data Separation:
+	 * Packet was placed using Header/Data separation algorithm.
+	 * The separation location is indicated by the itype field.
+	 */
+	#define RX_TPA_START_V2_CMPL_FLAGS_PLACEMENT_HDS \
+		(UINT32_C(0x2) << 7)
+	/*
+	 * IOC/Jumbo:
+	 * Packet will be placed using In-Order Completion/Jumbo where
+	 * the first packet of the aggregation is placed using Jumbo
+	 * Placement. Subsequent packets will be placed such that each
+	 * packet starts at the beginning of an aggregation buffer.
+	 */
+	#define RX_TPA_START_V2_CMPL_FLAGS_PLACEMENT_IOC_JUMBO \
+		(UINT32_C(0x4) << 7)
+	/*
+	 * GRO/Jumbo:
+	 * Packet will be placed using GRO/Jumbo where the first
+	 * packet is filled with data. Subsequent packets will be
+	 * placed such that any one packet does not span two
+	 * aggregation buffers unless it starts at the beginning of
+	 * an aggregation buffer.
+	 */
+	#define RX_TPA_START_V2_CMPL_FLAGS_PLACEMENT_GRO_JUMBO \
+		(UINT32_C(0x5) << 7)
+	/*
+	 * GRO/Header-Data Separation:
+	 * Packet will be placed using GRO/HDS where the header
+	 * is in the first packet.
+	 * Payload of each packet will be
+	 * placed such that any one packet does not span two
+	 * aggregation buffers unless it starts at the beginning of
+	 * an aggregation buffer.
+	 */
+	#define RX_TPA_START_V2_CMPL_FLAGS_PLACEMENT_GRO_HDS \
+		(UINT32_C(0x6) << 7)
+	/*
+	 * IOC/Header-Data Separation:
+	 * Packet will be placed using In-Order Completion/HDS where
+	 * the header is in the first packet buffer. Payload of each
+	 * packet will be placed such that each packet starts at the
+	 * beginning of an aggregation buffer.
+	 */
+	#define RX_TPA_START_V2_CMPL_FLAGS_PLACEMENT_IOC_HDS \
+		(UINT32_C(0x7) << 7)
+	#define RX_TPA_START_V2_CMPL_FLAGS_PLACEMENT_LAST \
+		RX_TPA_START_V2_CMPL_FLAGS_PLACEMENT_IOC_HDS
+	/* This bit is '1' if the RSS field in this completion is valid. */
+	#define RX_TPA_START_V2_CMPL_FLAGS_RSS_VALID \
+		UINT32_C(0x400)
+	/*
+	 * This bit is '1' if metadata has been added to the end of the
+	 * packet in host memory. Metadata starts at the first 32B boundary
+	 * after the end of the packet for regular and jumbo placement. It
+	 * starts at the first 32B boundary after the end of the header for
+	 * HDS placement. The length of the metadata is indicated in the
+	 * metadata itself.
+	 */
+	#define RX_TPA_START_V2_CMPL_FLAGS_PKT_METADATA_PRESENT \
+		UINT32_C(0x800)
+	/*
+	 * This value indicates what the inner packet determined for the
+	 * packet was.
+	 */
+	#define RX_TPA_START_V2_CMPL_FLAGS_ITYPE_MASK \
+		UINT32_C(0xf000)
+	#define RX_TPA_START_V2_CMPL_FLAGS_ITYPE_SFT                 12
+	/*
+	 * TCP Packet:
+	 * Indicates that the packet was IP and TCP.
+	 */
+	#define RX_TPA_START_V2_CMPL_FLAGS_ITYPE_TCP \
+		(UINT32_C(0x2) << 12)
+	#define RX_TPA_START_V2_CMPL_FLAGS_ITYPE_LAST \
+		RX_TPA_START_V2_CMPL_FLAGS_ITYPE_TCP
+	/*
+	 * This value indicates the amount of packet data written to the
+	 * buffer the opaque field in this completion corresponds to.
+	 */
+	uint16_t	len;
+	/*
+	 * This is a copy of the opaque field from the RX BD this completion
+	 * corresponds to. If the VNIC is configured to not use an Rx BD for
+	 * the TPA Start completion, then this is a copy of the opaque field
+	 * from the first BD used to place the TPA Start packet.
+	 */
+	uint32_t	opaque;
+	/*
+	 * This value is written by the NIC such that it will be different
+	 * for each pass through the completion queue. The even passes
+	 * will write 1. The odd passes will write 0.
+	 */
+	uint8_t	v1;
+	/*
+	 * This value is written by the NIC such that it will be different
+	 * for each pass through the completion queue. The even passes
+	 * will write 1. The odd passes will write 0.
+	 */
+	#define RX_TPA_START_V2_CMPL_V1 UINT32_C(0x1)
+	#define RX_TPA_START_V2_CMPL_LAST RX_TPA_START_V2_CMPL_V1
+	/*
+	 * This is the RSS hash type for the packet. The value is packed
+	 * {tuple_extrac_op[1:0],rss_profile_id[4:0],tuple_extrac_op[2]}.
+	 *
+	 * The value of tuple_extrac_op provides the information about
+	 * what fields the hash was computed on.
+	 * * 0: The RSS hash was computed over source IP address,
+	 * destination IP address, source port, and destination port of inner
+	 * IP and TCP or UDP headers. Note: For non-tunneled packets,
+	 * the packet headers are considered inner packet headers for the RSS
+	 * hash computation purpose.
+	 * * 1: The RSS hash was computed over source IP address and destination
+	 * IP address of inner IP header. Note: For non-tunneled packets,
+	 * the packet headers are considered inner packet headers for the RSS
+	 * hash computation purpose.
+	 * * 2: The RSS hash was computed over source IP address,
+	 * destination IP address, source port, and destination port of
+	 * IP and TCP or UDP headers of outer tunnel headers.
+	 * Note: For non-tunneled packets, this value is not applicable.
+	 * * 3: The RSS hash was computed over source IP address and
+	 * destination IP address of IP header of outer tunnel headers.
+	 * Note: For non-tunneled packets, this value is not applicable.
+	 *
+	 * Note that 4-tuples values listed above are applicable
+	 * for layer 4 protocols supported and enabled for RSS in the hardware,
+	 * HWRM firmware, and drivers. For example, if RSS hash is supported and
+	 * enabled for TCP traffic only, then the values of tuple_extract_op
+	 * corresponding to 4-tuples are only valid for TCP traffic.
+	 */
+	uint8_t	rss_hash_type;
+	/*
+	 * This is the aggregation ID that the completion is associated
+	 * with. Use this number to correlate the TPA start completion
+	 * with the TPA end completion.
+	 */
+	uint16_t	agg_id;
+	/*
+	 * This is the aggregation ID that the completion is associated
+	 * with. Use this number to correlate the TPA start completion
+	 * with the TPA end completion.
+	 */
+	#define RX_TPA_START_V2_CMPL_AGG_ID_MASK            UINT32_C(0xfff)
+	#define RX_TPA_START_V2_CMPL_AGG_ID_SFT             0
+	#define RX_TPA_START_V2_CMPL_METADATA1_MASK         UINT32_C(0xf000)
+	#define RX_TPA_START_V2_CMPL_METADATA1_SFT          12
+	/* When meta_format != 0, this value is the VLAN TPID_SEL. */
+	#define RX_TPA_START_V2_CMPL_METADATA1_TPID_SEL_MASK UINT32_C(0x7000)
+	#define RX_TPA_START_V2_CMPL_METADATA1_TPID_SEL_SFT  12
+	/* When meta_format != 0, this value is the VLAN valid. */
+	#define RX_TPA_START_V2_CMPL_METADATA1_VALID         UINT32_C(0x8000)
+	/*
+	 * This value is the RSS hash value calculated for the packet
+	 * based on the mode bits and key value in the VNIC.
+	 * When vee_cmpl_mode is set in VNIC context, this is the lower
+	 * 32b of the host address from the first BD used to place the packet.
+	 */
+	uint32_t	rss_hash;
+} __rte_packed;
+
+/*
+ * Last 16 bytes of RX L2 TPA Start V2 Completion Record
+ *
+ * This TPA completion structure is used on devices where the
+ * `hwrm_vnic_qcaps.max_aggs_supported` value is 0.
+ */
+/* rx_tpa_start_v2_cmpl_hi (size:128b/16B) */
+struct rx_tpa_start_v2_cmpl_hi {
+	uint32_t	flags2;
+	/* This indicates that the aggregation was done using GRO rules. */
+	#define RX_TPA_START_V2_CMPL_FLAGS2_AGG_GRO \
+		UINT32_C(0x4)
+	/*
+	 * When this bit is '0', the cs_ok field has the following definition:-
+	 * ip_cs_ok[2:0] = The number of header groups with a valid IP checksum
+	 * in the delivered packet, counted from the outer-most header group to
+	 * the inner-most header group, stopping at the first error. -
+	 * l4_cs_ok[5:3] = The number of header groups with a valid L4 checksum
+	 * in the delivered packet, counted from the outer-most header group to
+	 * the inner-most header group, stopping at the first error. When this
+	 * bit is '1', the cs_ok field has the following definition: -
+	 * hdr_cnt[2:0] = The number of header groups that were parsed by the
+	 * chip and passed in the delivered packet. - ip_cs_all_ok[3] =This bit
+	 * will be '1' if all the parsed header groups with an IP checksum are
+	 * valid. - l4_cs_all_ok[4] = This bit will be '1' if all the parsed
+	 * header groups with an L4 checksum are valid.
+	 */
+	#define RX_TPA_START_V2_CMPL_FLAGS2_CS_ALL_OK_MODE \
+		UINT32_C(0x8)
+	/* This value indicates what format the metadata field is. */
+	#define RX_TPA_START_V2_CMPL_FLAGS2_META_FORMAT_MASK \
+		UINT32_C(0xf0)
+	#define RX_TPA_START_V2_CMPL_FLAGS2_META_FORMAT_SFT            4
+	/* There is no metadata information. Values are zero. */
+	#define RX_TPA_START_V2_CMPL_FLAGS2_META_FORMAT_NONE \
+		(UINT32_C(0x0) << 4)
+	/*
+	 * The {metadata1, metadata0} fields contain the vtag
+	 * information: - vtag[19:0] = {valid, tpid_sel[2:0], pri[2:0],
+	 * de, vid[11:0]} The metadata2 field contains the table scope
+	 * and action record pointer. - metadata2[25:0] contains the
+	 * action record pointer. - metadata2[31:26] contains the table
+	 * scope.
+	 */
+	#define RX_TPA_START_V2_CMPL_FLAGS2_META_FORMAT_ACT_REC_PTR \
+		(UINT32_C(0x1) << 4)
+	/*
+	 * The {metadata1, metadata0} fields contain the vtag
+	 * information:
+	 * - vtag[19:0] = {valid, tpid_sel[2:0], pri[2:0], de, vid[11:0]}
+	 * The metadata2 field contains the Tunnel ID
+	 * value, justified to LSB. i
+	 * - VXLAN = VNI[23:0] -> VXLAN Network ID
+	 * - Geneve (NGE) = VNI[23:0] a-> Virtual Network Identifier
+	 * - NVGRE = TNI[23:0] -> Tenant Network ID
+	 * - GRE = KEY[31:0] -> key field with bit mask. zero if K=0
+	 * - IPv4 = 0 (not populated)
+	 * - IPv6 = Flow Label[19:0]
+	 * - PPPoE = sessionID[15:0]
+	 * - MPLs = Outer label[19:0]
+	 * - UPAR = Selected[31:0] with bit mask
+	 */
+	#define RX_TPA_START_V2_CMPL_FLAGS2_META_FORMAT_TUNNEL_ID \
+		(UINT32_C(0x2) << 4)
+	/*
+	 * The {metadata1, metadata0} fields contain the vtag
+	 * information:
+	 * - vtag[19:0] = {valid, tpid_sel[2:0], pri[2:0],de, vid[11:0]}
+	 * The metadata2 field contains the 32b metadata from the prepended
+	 * header (chdr_data).
+	 */
+	#define RX_TPA_START_V2_CMPL_FLAGS2_META_FORMAT_CHDR_DATA \
+		(UINT32_C(0x3) << 4)
+	/*
+	 * The {metadata1, metadata0} fields contain the vtag
+	 * information:
+	 * - vtag[19:0] = {valid, tpid_sel[2:0], pri[2:0], de, vid[11:0]}
+	 * The metadata2 field contains the outer_l3_offset,
+	 * inner_l2_offset, inner_l3_offset, and inner_l4_size.
+	 * - metadata2[8:0] contains the outer_l3_offset.
+	 * - metadata2[17:9] contains the inner_l2_offset.
+	 * - metadata2[26:18] contains the inner_l3_offset.
+	 * - metadata2[31:27] contains the inner_l4_size.
+	 */
+	#define RX_TPA_START_V2_CMPL_FLAGS2_META_FORMAT_HDR_OFFSET \
+		(UINT32_C(0x4) << 4)
+	#define RX_TPA_START_V2_CMPL_FLAGS2_META_FORMAT_LAST \
+		RX_TPA_START_V2_CMPL_FLAGS2_META_FORMAT_HDR_OFFSET
+	/*
+	 * This field indicates the IP type for the inner-most IP header.
+	 * A value of '0' indicates IPv4. A value of '1' indicates IPv6.
+	 * This value is only valid if itype indicates a packet
+	 * with an IP header.
+	 */
+	#define RX_TPA_START_V2_CMPL_FLAGS2_IP_TYPE \
+		UINT32_C(0x100)
+	/*
+	 * This indicates that the complete 1's complement checksum was
+	 * calculated for the packet in the affregation.
+	 */
+	#define RX_TPA_START_V2_CMPL_FLAGS2_COMPLETE_CHECKSUM_CALC \
+		UINT32_C(0x200)
+	/*
+	 * This field indicates the status of IP and L4 CS calculations done
+	 * by the chip. The format of this field is indicated by the
+	 * cs_all_ok_mode bit.
+	 * CS status for TPA packets is always valid. This means that "all_ok"
+	 * status will always be set. The ok count status will be set
+	 * appropriately for the packet header, such that all existing CS
+	 * values are ok.
+	 */
+	#define RX_TPA_START_V2_CMPL_FLAGS2_CS_OK_MASK \
+		UINT32_C(0xfc00)
+	#define RX_TPA_START_V2_CMPL_FLAGS2_CS_OK_SFT                  10
+	/*
+	 * This value is the complete 1's complement checksum calculated from
+	 * the start of the outer L3 header to the end of the packet (not
+	 * including the ethernet crc). It is valid when the
+	 * 'complete_checksum_calc' flag is set. For TPA Start completions,
+	 * the complete checksum is calculated for the first packet in the
+	 * aggregation only.
+	 */
+	#define RX_TPA_START_V2_CMPL_FLAGS2_COMPLETE_CHECKSUM_MASK \
+		UINT32_C(0xffff0000)
+	#define RX_TPA_START_V2_CMPL_FLAGS2_COMPLETE_CHECKSUM_SFT      16
+	/*
+	 * This is data from the CFA block as indicated by the meta_format
+	 * field.
+	 * - meta_format 0 - none - metadata2 = 0 - not valid/not stripped
+	 * - meta_format 1 - act_rec_ptr - metadata2 = {table_scope[5:0],
+	 *   act_rec_ptr[25:0]}
+	 * - meta_format 2 - tunnel_id - metadata2 = tunnel_id[31:0]
+	 * - meta_format 3 - chdr_data - metadata2 = updated_chdr_data[31:0]
+	 * - meta_format 4 - hdr_offsets - metadata2 = hdr_offsets[31:0]
+	 * When vee_cmpl_mode is set in VNIC context, this is the upper 32b
+	 * of the host address from the first BD used to place the packet.
+	 */
+	uint32_t	metadata2;
+	uint16_t	errors_v2;
+	/*
+	 * This value is written by the NIC such that it will be different
+	 * for each pass through the completion queue. The even passes
+	 * will write 1. The odd passes will write 0.
+	 */
+	#define RX_TPA_START_V2_CMPL_V2 \
+		UINT32_C(0x1)
+	#define RX_TPA_START_V2_CMPL_ERRORS_MASK \
+		UINT32_C(0xfffe)
+	#define RX_TPA_START_V2_CMPL_ERRORS_SFT                     1
+	/*
+	 * This error indicates that there was some sort of problem with
+	 * the BDs for the packetThe packet should be treated as
+	 * invalid.
+	 */
+	#define RX_TPA_START_V2_CMPL_ERRORS_BUFFER_ERROR_MASK \
+		UINT32_C(0xe)
+	#define RX_TPA_START_V2_CMPL_ERRORS_BUFFER_ERROR_SFT         1
+	/* No buffer error */
+	#define RX_TPA_START_V2_CMPL_ERRORS_BUFFER_ERROR_NO_BUFFER \
+		(UINT32_C(0x0) << 1)
+	/*
+	 * Did Not Fit:
+	 * Packet did not fit into packet buffer provided. This means
+	 * that the TPA Start packet was too big to be placed into the
+	 * per-packet maximum number of physical buffers configured for
+	 * the VNIC, or that it was too big to be placed into the
+	 * per-aggregation maximum number of physical buffers configured
+	 * for the VNIC. This error only occurs when the VNIC is
+	 * configured for variable size receive buffers.
+	 */
+	#define RX_TPA_START_V2_CMPL_ERRORS_BUFFER_ERROR_DID_NOT_FIT \
+		(UINT32_C(0x1) << 1)
+	/*
+	 * Bad Format:
+	 * BDs were not formatted correctly.
+	 */
+	#define RX_TPA_START_V2_CMPL_ERRORS_BUFFER_ERROR_BAD_FORMAT \
+		(UINT32_C(0x3) << 1)
+	/*
+	 * Flush:
+	 * There was a bad_format error on the previous operation
+	 */
+	#define RX_TPA_START_V2_CMPL_ERRORS_BUFFER_ERROR_FLUSH \
+		(UINT32_C(0x5) << 1)
+	#define RX_TPA_START_V2_CMPL_ERRORS_BUFFER_ERROR_LAST \
+		RX_TPA_START_V2_CMPL_ERRORS_BUFFER_ERROR_FLUSH
+	/*
+	 * This is data from the CFA block as indicated by the meta_format
+	 * field.
+	 */
+	uint16_t	metadata0;
+	/* When meta_format != 0, this value is the VLAN VID. */
+	#define RX_TPA_START_V2_CMPL_METADATA0_VID_MASK UINT32_C(0xfff)
+	#define RX_TPA_START_V2_CMPL_METADATA0_VID_SFT 0
+	/* When meta_format != 0, this value is the VLAN DE. */
+	#define RX_TPA_START_V2_CMPL_METADATA0_DE      UINT32_C(0x1000)
+	/* When meta_format != 0, this value is the VLAN PRI. */
+	#define RX_TPA_START_V2_CMPL_METADATA0_PRI_MASK UINT32_C(0xe000)
+	#define RX_TPA_START_V2_CMPL_METADATA0_PRI_SFT 13
+	/*
+	 * This field contains the outer_l3_offset, inner_l2_offset,
+	 * inner_l3_offset, and inner_l4_size.
+	 *
+	 * hdr_offsets[8:0] contains the outer_l3_offset.
+	 * hdr_offsets[17:9] contains the inner_l2_offset.
+	 * hdr_offsets[26:18] contains the inner_l3_offset.
+	 * hdr_offsets[31:27] contains the inner_l4_size.
+	 */
+	uint32_t	hdr_offsets;
+} __rte_packed;
+
+/*
+ * This TPA completion structure is used on devices where the
+ * `hwrm_vnic_qcaps.max_aggs_supported` value is 0.
  */
 /* rx_tpa_end_cmpl (size:128b/16B) */
 struct rx_tpa_end_cmpl {
@@ -3299,27 +4718,27 @@ struct rx_tpa_end_cmpl {
 	 * records. Odd values indicate 32B
 	 * records.
 	 */
-	#define RX_TPA_END_CMPL_TYPE_MASK                UINT32_C(0x3f)
-	#define RX_TPA_END_CMPL_TYPE_SFT                 0
+	#define RX_TPA_END_CMPL_TYPE_MASK                      UINT32_C(0x3f)
+	#define RX_TPA_END_CMPL_TYPE_SFT                       0
 	/*
 	 * RX L2 TPA End Completion:
 	 * Completion at the end of a TPA operation.
 	 * Length = 32B
 	 */
-	#define RX_TPA_END_CMPL_TYPE_RX_TPA_END            UINT32_C(0x15)
+	#define RX_TPA_END_CMPL_TYPE_RX_TPA_END                  UINT32_C(0x15)
 	#define RX_TPA_END_CMPL_TYPE_LAST \
 		RX_TPA_END_CMPL_TYPE_RX_TPA_END
-	#define RX_TPA_END_CMPL_FLAGS_MASK               UINT32_C(0xffc0)
-	#define RX_TPA_END_CMPL_FLAGS_SFT                6
+	#define RX_TPA_END_CMPL_FLAGS_MASK                     UINT32_C(0xffc0)
+	#define RX_TPA_END_CMPL_FLAGS_SFT                      6
 	/*
 	 * When this bit is '1', it indicates a packet that has an
 	 * error of some type. Type of error is indicated in
 	 * error_flags.
 	 */
-	#define RX_TPA_END_CMPL_FLAGS_ERROR               UINT32_C(0x40)
+	#define RX_TPA_END_CMPL_FLAGS_ERROR                     UINT32_C(0x40)
 	/* This field indicates how the packet was placed in the buffer. */
-	#define RX_TPA_END_CMPL_FLAGS_PLACEMENT_MASK      UINT32_C(0x380)
-	#define RX_TPA_END_CMPL_FLAGS_PLACEMENT_SFT       7
+	#define RX_TPA_END_CMPL_FLAGS_PLACEMENT_MASK            UINT32_C(0x380)
+	#define RX_TPA_END_CMPL_FLAGS_PLACEMENT_SFT             7
 	/*
 	 * Jumbo:
 	 * TPA Packet was placed using jumbo algorithm. This means
@@ -3338,6 +4757,15 @@ struct rx_tpa_end_cmpl {
 	#define RX_TPA_END_CMPL_FLAGS_PLACEMENT_HDS \
 		(UINT32_C(0x2) << 7)
 	/*
+	 * IOC/Jumbo:
+	 * Packet will be placed using In-Order Completion/Jumbo where
+	 * the first packet of the aggregation is placed using Jumbo
+	 * Placement. Subsequent packets will be placed such that each
+	 * packet starts at the beginning of an aggregation buffer.
+	 */
+	#define RX_TPA_END_CMPL_FLAGS_PLACEMENT_IOC_JUMBO \
+		(UINT32_C(0x4) << 7)
+	/*
 	 * GRO/Jumbo:
 	 * Packet will be placed using GRO/Jumbo where the first
 	 * packet is filled with data. Subsequent packets will be
@@ -3358,11 +4786,28 @@ struct rx_tpa_end_cmpl {
 	 */
 	#define RX_TPA_END_CMPL_FLAGS_PLACEMENT_GRO_HDS \
 		(UINT32_C(0x6) << 7)
+	/*
+	 * IOC/Header-Data Separation:
+	 * Packet will be placed using In-Order Completion/HDS where
+	 * the header is in the first packet buffer. Payload of each
+	 * packet will be placed such that each packet starts at the
+	 * beginning of an aggregation buffer.
+	 */
+	#define RX_TPA_END_CMPL_FLAGS_PLACEMENT_IOC_HDS \
+		(UINT32_C(0x7) << 7)
 	#define RX_TPA_END_CMPL_FLAGS_PLACEMENT_LAST \
-		RX_TPA_END_CMPL_FLAGS_PLACEMENT_GRO_HDS
-	/* unused is 2 b */
-	#define RX_TPA_END_CMPL_FLAGS_UNUSED_MASK         UINT32_C(0xc00)
-	#define RX_TPA_END_CMPL_FLAGS_UNUSED_SFT          10
+		RX_TPA_END_CMPL_FLAGS_PLACEMENT_IOC_HDS
+	/* unused is 1 b */
+	#define RX_TPA_END_CMPL_FLAGS_UNUSED                    UINT32_C(0x400)
+	/*
+	 * This bit is '1' if metadata has been added to the end of the
+	 * packet in host memory. Metadata starts at the first 32B boundary
+	 * after the end of the packet for regular and jumbo placement.
+	 * It starts at the first 32B boundary after the end of the header
+	 * for HDS placement. The length of the metadata is indicated in the
+	 * metadata itself.
+	 */
+	#define RX_TPA_END_CMPL_FLAGS_PKT_METADATA_PRESENT      UINT32_C(0x800)
 	/*
 	 * This value indicates what the inner packet determined for the
 	 * packet was.
@@ -3372,8 +4817,9 @@ struct rx_tpa_end_cmpl {
 	 *     field is valid and contains the TCP checksum.
 	 *     This also indicates that the payload_offset field is valid.
 	 */
-	#define RX_TPA_END_CMPL_FLAGS_ITYPE_MASK          UINT32_C(0xf000)
-	#define RX_TPA_END_CMPL_FLAGS_ITYPE_SFT           12
+	#define RX_TPA_END_CMPL_FLAGS_ITYPE_MASK \
+		UINT32_C(0xf000)
+	#define RX_TPA_END_CMPL_FLAGS_ITYPE_SFT                 12
 	/*
 	 * This value is zero for TPA End completions.
 	 * There is no data in the buffer that corresponds to the opaque
@@ -4243,6 +5689,52 @@ struct rx_abuf_cmpl {
 	uint32_t	unused_2;
 } __rte_packed;
 
+/* VEE FLUSH Completion Record (16 bytes) */
+/* vee_flush (size:128b/16B) */
+struct vee_flush {
+	uint32_t	downstream_path_type;
+	/*
+	 * This field indicates the exact type of the completion.
+	 * By convention, the LSB identifies the length of the
+	 * record in 16B units. Even values indicate 16B
+	 * records. Odd values indicate 32B
+	 * records.
+	 */
+	#define VEE_FLUSH_TYPE_MASK           UINT32_C(0x3f)
+	#define VEE_FLUSH_TYPE_SFT            0
+	/*
+	 * VEE Flush Completion:
+	 * This completion is inserted manually by the Primate and processed
+	 * by the VEE hardware to ensure that all completions on a VEE
+	 * function have been processed by the VEE hardware before FLR
+	 * process is completed.
+	 */
+	#define VEE_FLUSH_TYPE_VEE_FLUSH        UINT32_C(0x1c)
+	#define VEE_FLUSH_TYPE_LAST            VEE_FLUSH_TYPE_VEE_FLUSH
+	/* downstream_path is 1 b */
+	#define VEE_FLUSH_DOWNSTREAM_PATH     UINT32_C(0x40)
+	/* This completion is associated with VEE Transmit */
+	#define VEE_FLUSH_DOWNSTREAM_PATH_TX    (UINT32_C(0x0) << 6)
+	/* This completion is associated with VEE Receive */
+	#define VEE_FLUSH_DOWNSTREAM_PATH_RX    (UINT32_C(0x1) << 6)
+	#define VEE_FLUSH_DOWNSTREAM_PATH_LAST VEE_FLUSH_DOWNSTREAM_PATH_RX
+	/*
+	 * This is an opaque value that is passed through the completion
+	 * to the VEE handler SW and is used to indicate what VEE VQ or
+	 * function has completed FLR processing.
+	 */
+	uint32_t	opaque;
+	uint32_t	v;
+	/*
+	 * This value is written by the NIC such that it will be different
+	 * for each pass through the completion queue. The even passes will
+	 * write 1. The odd passes will write 0.
+	 */
+	#define VEE_FLUSH_V     UINT32_C(0x1)
+	/* unused3 is 32 b */
+	uint32_t	unused_3;
+} __rte_packed;
+
 /* eject_cmpl (size:128b/16B) */
 struct eject_cmpl {
 	uint16_t	type;
@@ -6562,7 +8054,7 @@ struct hwrm_async_event_cmpl_deferred_response {
 	/*
 	 * The PF's mailbox is clear to issue another command.
 	 * A command with this seq_id is still in progress
-	 * and will return a regular HWRM completion when done.
+	 * and will return a regualr HWRM completion when done.
 	 * 'event_data1' field, if non-zero, contains the estimated
 	 * execution time for the command.
 	 */
@@ -7476,6 +8968,8 @@ struct hwrm_func_qcaps_input {
 	 * Function ID of the function that is being queried.
 	 * 0xFF... (All Fs) if the query is for the requesting
 	 * function.
+	 * 0xFFFE (REQUESTING_PARENT_FID) This is a special FID
+	 * to be used by a trusted VF to query its parent PF.
 	 */
 	uint16_t	fid;
 	uint8_t	unused_0[6];
@@ -7730,6 +9224,12 @@ struct hwrm_func_qcaps_output {
 	#define HWRM_FUNC_QCAPS_OUTPUT_FLAGS_PFC_WD_STATS_SUPPORTED \
 		UINT32_C(0x40000000)
 	/*
+	 * When this bit is '1', it indicates that core firmware supports
+	 * DBG_QCAPS command
+	 */
+	#define HWRM_FUNC_QCAPS_OUTPUT_FLAGS_DBG_QCAPS_CMD_SUPPORTED \
+		UINT32_C(0x80000000)
+	/*
 	 * This value is current MAC address configured for this
 	 * function. A value of 00-00-00-00-00-00 indicates no
 	 * MAC address is currently configured.
@@ -7854,6 +9354,19 @@ struct hwrm_func_qcaps_output {
 	 */
 	#define HWRM_FUNC_QCAPS_OUTPUT_FLAGS_EXT_ECN_STATS_SUPPORTED \
 		UINT32_C(0x2)
+	/*
+	 * If 1, the device can report extended hw statistics (including
+	 * additional tpa statistics).
+	 */
+	#define HWRM_FUNC_QCAPS_OUTPUT_FLAGS_EXT_EXT_HW_STATS_SUPPORTED \
+		UINT32_C(0x4)
+	/*
+	 * If set to 1, then the core firmware has support to enable/
+	 * disable hot reset support for interface dynamically through
+	 * HWRM_FUNC_CFG.
+	 */
+	#define HWRM_FUNC_QCAPS_OUTPUT_FLAGS_EXT_HOT_RESET_IF_SUPPORT \
+		UINT32_C(0x8)
 	uint8_t	unused_1[3];
 	/*
 	 * This field is used in Output records to indicate that the output
@@ -7904,6 +9417,8 @@ struct hwrm_func_qcfg_input {
 	 * Function ID of the function that is being queried.
 	 * 0xFF... (All Fs) if the query is for the requesting
 	 * function.
+	 * 0xFFFE (REQUESTING_PARENT_FID) This is a special FID
+	 * to be used by a trusted VF to query its parent PF.
 	 */
 	uint16_t	fid;
 	uint8_t	unused_0[6];
@@ -8014,6 +9529,15 @@ struct hwrm_func_qcfg_output {
 	#define HWRM_FUNC_QCFG_OUTPUT_FLAGS_PREBOOT_LEGACY_L2_RINGS \
 		UINT32_C(0x100)
 	/*
+	 * If set to 1, then the firmware and all currently registered driver
+	 * instances support hot reset. The hot reset support will be updated
+	 * dynamically based on the driver interface advertisement.
+	 * If set to 0, then the adapter is not currently able to initiate
+	 * hot reset.
+	 */
+	#define HWRM_FUNC_QCFG_OUTPUT_FLAGS_HOT_RESET_ALLOWED \
+		UINT32_C(0x200)
+	/*
 	 * This value is current MAC address configured for this
 	 * function. A value of 00-00-00-00-00-00 indicates no
 	 * MAC address is currently configured.
@@ -8565,6 +10089,17 @@ struct hwrm_func_cfg_input {
 	 */
 	#define HWRM_FUNC_CFG_INPUT_FLAGS_PREBOOT_LEGACY_L2_RINGS \
 		UINT32_C(0x2000000)
+	/*
+	 * If this bit is set to 0, then the interface does not support hot
+	 * reset capability which it advertised with the hot_reset_support
+	 * flag in HWRM_FUNC_DRV_RGTR. If any of the function has set this
+	 * flag to 0, adapter cannot do the hot reset. In this state, if the
+	 * firmware receives a hot reset request, firmware must fail the
+	 * request. If this bit is set to 1, then interface is renabling the
+	 * hot reset capability.
+	 */
+	#define HWRM_FUNC_CFG_INPUT_FLAGS_HOT_RESET_IF_EN_DIS \
+		UINT32_C(0x4000000)
 	uint32_t	enables;
 	/*
 	 * This bit must be '1' for the mtu field to be
@@ -8705,6 +10240,12 @@ struct hwrm_func_cfg_input {
 	#define HWRM_FUNC_CFG_INPUT_ENABLES_ADMIN_LINK_STATE \
 		UINT32_C(0x400000)
 	/*
+	 * This bit must be '1' for the hot_reset_if_en_dis field to be
+	 * configured.
+	 */
+	#define HWRM_FUNC_CFG_INPUT_ENABLES_HOT_RESET_IF_SUPPORT \
+		UINT32_C(0x800000)
+	/*
 	 * The maximum transmission unit of the function.
 	 * The HWRM should make sure that the mtu of
 	 * the function does not exceed the mtu of the physical
@@ -9036,15 +10577,21 @@ struct hwrm_func_qstats_input {
 	/* This flags indicates the type of statistics request. */
 	uint8_t	flags;
 	/* This value is not used to avoid backward compatibility issues. */
-	#define HWRM_FUNC_QSTATS_INPUT_FLAGS_UNUSED    UINT32_C(0x0)
+	#define HWRM_FUNC_QSTATS_INPUT_FLAGS_UNUSED       UINT32_C(0x0)
 	/*
 	 * flags should be set to 1 when request is for only RoCE statistics.
 	 * This will be honored only if the caller_fid is a privileged PF.
 	 * In all other cases FID and caller_fid should be the same.
 	 */
-	#define HWRM_FUNC_QSTATS_INPUT_FLAGS_ROCE_ONLY UINT32_C(0x1)
+	#define HWRM_FUNC_QSTATS_INPUT_FLAGS_ROCE_ONLY    UINT32_C(0x1)
+	/*
+	 * flags should be set to 2 when request is for the counter mask,
+	 * representing the width of each of the stats counters, rather
+	 * than counters themselves.
+	 */
+	#define HWRM_FUNC_QSTATS_INPUT_FLAGS_COUNTER_MASK UINT32_C(0x2)
 	#define HWRM_FUNC_QSTATS_INPUT_FLAGS_LAST \
-		HWRM_FUNC_QSTATS_INPUT_FLAGS_ROCE_ONLY
+		HWRM_FUNC_QSTATS_INPUT_FLAGS_COUNTER_MASK
 	uint8_t	unused_0[5];
 } __rte_packed;
 
@@ -9130,6 +10677,132 @@ struct hwrm_func_qstats_output {
 	uint8_t	valid;
 } __rte_packed;
 
+/************************
+ * hwrm_func_qstats_ext *
+ ************************/
+
+
+/* hwrm_func_qstats_ext_input (size:192b/24B) */
+struct hwrm_func_qstats_ext_input {
+	/* The HWRM command request type. */
+	uint16_t	req_type;
+	/*
+	 * The completion ring to send the completion event on. This should
+	 * be the NQ ID returned from the `nq_alloc` HWRM command.
+	 */
+	uint16_t	cmpl_ring;
+	/*
+	 * The sequence ID is used by the driver for tracking multiple
+	 * commands. This ID is treated as opaque data by the firmware and
+	 * the value is returned in the `hwrm_resp_hdr` upon completion.
+	 */
+	uint16_t	seq_id;
+	/*
+	 * The target ID of the command:
+	 * * 0x0-0xFFF8 - The function ID
+	 * * 0xFFF8-0xFFFC, 0xFFFE - Reserved for internal processors
+	 * * 0xFFFD - Reserved for user-space HWRM interface
+	 * * 0xFFFF - HWRM
+	 */
+	uint16_t	target_id;
+	/*
+	 * A physical address pointer pointing to a host buffer that the
+	 * command's response data will be written. This can be either a host
+	 * physical address (HPA) or a guest physical address (GPA) and must
+	 * point to a physically contiguous block of memory.
+	 */
+	uint64_t	resp_addr;
+	/*
+	 * Function ID of the function that is being queried.
+	 * 0xFF... (All Fs) if the query is for the requesting
+	 * function.
+	 * A privileged PF can query for other function's statistics.
+	 */
+	uint16_t	fid;
+	/* This flags indicates the type of statistics request. */
+	uint8_t	flags;
+	/* This value is not used to avoid backward compatibility issues. */
+	#define HWRM_FUNC_QSTATS_EXT_INPUT_FLAGS_UNUSED       UINT32_C(0x0)
+	/*
+	 * flags should be set to 1 when request is for only RoCE statistics.
+	 * This will be honored only if the caller_fid is a privileged PF.
+	 * In all other cases FID and caller_fid should be the same.
+	 */
+	#define HWRM_FUNC_QSTATS_EXT_INPUT_FLAGS_ROCE_ONLY    UINT32_C(0x1)
+	/*
+	 * flags should be set to 2 when request is for the counter mask
+	 * representing the width of each of the stats counters, rather
+	 * than counters themselves.
+	 */
+	#define HWRM_FUNC_QSTATS_EXT_INPUT_FLAGS_COUNTER_MASK UINT32_C(0x2)
+	#define HWRM_FUNC_QSTATS_EXT_INPUT_FLAGS_LAST \
+		HWRM_FUNC_QSTATS_EXT_INPUT_FLAGS_COUNTER_MASK
+	uint8_t	unused_0[5];
+} __rte_packed;
+
+/* hwrm_func_qstats_ext_output (size:1472b/184B) */
+struct hwrm_func_qstats_ext_output {
+	/* The specific error status for the command. */
+	uint16_t	error_code;
+	/* The HWRM command request type. */
+	uint16_t	req_type;
+	/* The sequence ID from the original command. */
+	uint16_t	seq_id;
+	/* The length of the response data in number of bytes. */
+	uint16_t	resp_len;
+	/* Number of received unicast packets */
+	uint64_t	rx_ucast_pkts;
+	/* Number of received multicast packets */
+	uint64_t	rx_mcast_pkts;
+	/* Number of received broadcast packets */
+	uint64_t	rx_bcast_pkts;
+	/* Number of discarded packets on received path */
+	uint64_t	rx_discard_pkts;
+	/* Number of packets on receive path with error */
+	uint64_t	rx_error_pkts;
+	/* Number of received bytes for unicast traffic */
+	uint64_t	rx_ucast_bytes;
+	/* Number of received bytes for multicast traffic */
+	uint64_t	rx_mcast_bytes;
+	/* Number of received bytes for broadcast traffic */
+	uint64_t	rx_bcast_bytes;
+	/* Number of transmitted unicast packets */
+	uint64_t	tx_ucast_pkts;
+	/* Number of transmitted multicast packets */
+	uint64_t	tx_mcast_pkts;
+	/* Number of transmitted broadcast packets */
+	uint64_t	tx_bcast_pkts;
+	/* Number of packets on transmit path with error */
+	uint64_t	tx_error_pkts;
+	/* Number of discarded packets on transmit path */
+	uint64_t	tx_discard_pkts;
+	/* Number of transmitted bytes for unicast traffic */
+	uint64_t	tx_ucast_bytes;
+	/* Number of transmitted bytes for multicast traffic */
+	uint64_t	tx_mcast_bytes;
+	/* Number of transmitted bytes for broadcast traffic */
+	uint64_t	tx_bcast_bytes;
+	/* Number of TPA eligible packets */
+	uint64_t	rx_tpa_eligible_pkt;
+	/* Number of TPA eligible bytes */
+	uint64_t	rx_tpa_eligible_bytes;
+	/* Number of TPA packets */
+	uint64_t	rx_tpa_pkt;
+	/* Number of TPA bytes */
+	uint64_t	rx_tpa_bytes;
+	/* Number of TPA errors */
+	uint64_t	rx_tpa_errors;
+	uint8_t	unused_0[7];
+	/*
+	 * This field is used in Output records to indicate that the output
+	 * is completely written to RAM.  This field should be read as '1'
+	 * to indicate that the output has been completely written.
+	 * When writing a command completion or response to an internal processor,
+	 * the order of writes has to be such that this field is written last.
+	 */
+	uint8_t	valid;
+} __rte_packed;
+
 /***********************
  * hwrm_func_clr_stats *
  ***********************/
@@ -10116,7 +11789,7 @@ struct hwrm_func_backing_store_qcaps_output {
 	 *
 	 * TQM slowpath rings should be sized as follows:
 	 *
-	 * num_entries = num_vnics + num_l2_tx_rings + num_roce_qps + tqm_min_size
+	 * num_entries = num_vnics + num_l2_tx_rings + 2 * num_roce_qps + tqm_min_size
 	 *
 	 * Where:
 	 *   num_vnics is the number of VNICs allocated in the VNIC backing store
@@ -11039,7 +12712,7 @@ struct hwrm_func_backing_store_cfg_input {
 	 *
 	 * TQM slowpath rings should be sized as follows:
 	 *
-	 * num_entries = num_vnics + num_l2_tx_rings + num_roce_qps + tqm_min_size
+	 * num_entries = num_vnics + num_l2_tx_rings + 2 * num_roce_qps + tqm_min_size
 	 *
 	 * Where:
 	 *   num_vnics is the number of VNICs allocated in the VNIC backing store
@@ -14403,7 +16076,7 @@ struct hwrm_port_phy_qcfg_output {
 	/* Module is not inserted. */
 	#define HWRM_PORT_PHY_QCFG_OUTPUT_MODULE_STATUS_NOTINSERTED \
 		UINT32_C(0x4)
-	/* Module is powered down because of over current fault. */
+	/* Module is powered down becuase of over current fault. */
 	#define HWRM_PORT_PHY_QCFG_OUTPUT_MODULE_STATUS_CURRENTFAULT \
 		UINT32_C(0x5)
 	/* Module status is not applicable. */
@@ -16149,7 +17822,18 @@ struct hwrm_port_qstats_input {
 	uint64_t	resp_addr;
 	/* Port ID of port that is being queried. */
 	uint16_t	port_id;
-	uint8_t	unused_0[6];
+	uint8_t	flags;
+	/* This value is not used to avoid backward compatibility issues. */
+	#define HWRM_PORT_QSTATS_INPUT_FLAGS_UNUSED       UINT32_C(0x0)
+	/*
+	 * This bit is set to 1 when request is for a counter mask,
+	 * representing the width of each of the stats counters, rather
+	 * than counters themselves.
+	 */
+	#define HWRM_PORT_QSTATS_INPUT_FLAGS_COUNTER_MASK UINT32_C(0x1)
+	#define HWRM_PORT_QSTATS_INPUT_FLAGS_LAST \
+		HWRM_PORT_QSTATS_INPUT_FLAGS_COUNTER_MASK
+	uint8_t	unused_0[5];
 	/*
 	 * This is the host address where
 	 * Tx port statistics will be stored
@@ -16382,7 +18066,7 @@ struct rx_port_stats_ext {
  * Port Rx Statistics extended PFC WatchDog Format.
  * StormDetect and StormRevert event determination is based
  * on an integration period and a percentage threshold.
- * StormDetect event - when percentage of XOFF frames received
+ * StormDetect event - when percentage of XOFF frames receieved
  * within an integration period exceeds the configured threshold.
  * StormRevert event - when percentage of XON frames received
  * within an integration period exceeds the configured threshold.
@@ -16843,7 +18527,18 @@ struct hwrm_port_qstats_ext_input {
 	 * statistics block in bytes
 	 */
 	uint16_t	rx_stat_size;
-	uint8_t	unused_0[2];
+	uint8_t	flags;
+	/* This value is not used to avoid backward compatibility issues. */
+	#define HWRM_PORT_QSTATS_EXT_INPUT_FLAGS_UNUSED       UINT32_C(0x0)
+	/*
+	 * This bit is set to 1 when request is for the counter mask,
+	 * representing width of each of the stats counters, rather than
+	 * counters themselves.
+	 */
+	#define HWRM_PORT_QSTATS_EXT_INPUT_FLAGS_COUNTER_MASK UINT32_C(0x1)
+	#define HWRM_PORT_QSTATS_EXT_INPUT_FLAGS_LAST \
+		HWRM_PORT_QSTATS_EXT_INPUT_FLAGS_COUNTER_MASK
+	uint8_t	unused_0;
 	/*
 	 * This is the host address where
 	 * Tx port statistics will be stored
@@ -19283,7 +20978,7 @@ struct hwrm_port_phy_mdio_bus_acquire_input {
 	 * Timeout in milli seconds, MDIO BUS will be released automatically
 	 * after this time, if another mdio acquire command is not received
 	 * within the timeout window from the same client.
-	 * A 0xFFFF will hold the bus until this bus is released.
+	 * A 0xFFFF will hold the bus untill this bus is released.
 	 */
 	uint16_t	mdio_bus_timeout;
 	uint8_t	unused_0[2];
@@ -25312,95 +27007,104 @@ struct hwrm_ring_free_input {
 	/* Ring Type. */
 	uint8_t	ring_type;
 	/* L2 Completion Ring (CR) */
-	#define HWRM_RING_FREE_INPUT_RING_TYPE_L2_CMPL   UINT32_C(0x0)
+	#define HWRM_RING_FREE_INPUT_RING_TYPE_L2_CMPL   UINT32_C(0x0)
+	/* TX Ring (TR) */
+	#define HWRM_RING_FREE_INPUT_RING_TYPE_TX        UINT32_C(0x1)
+	/* RX Ring (RR) */
+	#define HWRM_RING_FREE_INPUT_RING_TYPE_RX        UINT32_C(0x2)
+	/* RoCE Notification Completion Ring (ROCE_CR) */
+	#define HWRM_RING_FREE_INPUT_RING_TYPE_ROCE_CMPL UINT32_C(0x3)
+	/* RX Aggregation Ring */
+	#define HWRM_RING_FREE_INPUT_RING_TYPE_RX_AGG    UINT32_C(0x4)
+	/* Notification Queue */
+	#define HWRM_RING_FREE_INPUT_RING_TYPE_NQ        UINT32_C(0x5)
+	#define HWRM_RING_FREE_INPUT_RING_TYPE_LAST \
+		HWRM_RING_FREE_INPUT_RING_TYPE_NQ
+	uint8_t	unused_0;
+	/* Physical number of ring allocated. */
+	uint16_t	ring_id;
+	uint8_t	unused_1[4];
+} __rte_packed;
+
+/* hwrm_ring_free_output (size:128b/16B) */
+struct hwrm_ring_free_output {
+	/* The specific error status for the command. */
+	uint16_t	error_code;
+	/* The HWRM command request type. */
+	uint16_t	req_type;
+	/* The sequence ID from the original command. */
+	uint16_t	seq_id;
+	/* The length of the response data in number of bytes. */
+	uint16_t	resp_len;
+	uint8_t	unused_0[7];
+	/*
+	 * This field is used in Output records to indicate that the output
+	 * is completely written to RAM.  This field should be read as '1'
+	 * to indicate that the output has been completely written.
+	 * When writing a command completion or response to an internal processor,
+	 * the order of writes has to be such that this field is written last.
+	 */
+	uint8_t	valid;
+} __rte_packed;
+
+/*******************
+ * hwrm_ring_reset *
+ *******************/
+
+
+/* hwrm_ring_reset_input (size:192b/24B) */
+struct hwrm_ring_reset_input {
+	/* The HWRM command request type. */
+	uint16_t	req_type;
+	/*
+	 * The completion ring to send the completion event on. This should
+	 * be the NQ ID returned from the `nq_alloc` HWRM command.
+	 */
+	uint16_t	cmpl_ring;
+	/*
+	 * The sequence ID is used by the driver for tracking multiple
+	 * commands. This ID is treated as opaque data by the firmware and
+	 * the value is returned in the `hwrm_resp_hdr` upon completion.
+	 */
+	uint16_t	seq_id;
+	/*
+	 * The target ID of the command:
+	 * * 0x0-0xFFF8 - The function ID
+	 * * 0xFFF8-0xFFFC, 0xFFFE - Reserved for internal processors
+	 * * 0xFFFD - Reserved for user-space HWRM interface
+	 * * 0xFFFF - HWRM
+	 */
+	uint16_t	target_id;
+	/*
+	 * A physical address pointer pointing to a host buffer that the
+	 * command's response data will be written. This can be either a host
+	 * physical address (HPA) or a guest physical address (GPA) and must
+	 * point to a physically contiguous block of memory.
+	 */
+	uint64_t	resp_addr;
+	/* Ring Type. */
+	uint8_t	ring_type;
+	/* L2 Completion Ring (CR) */
+	#define HWRM_RING_RESET_INPUT_RING_TYPE_L2_CMPL     UINT32_C(0x0)
 	/* TX Ring (TR) */
-	#define HWRM_RING_FREE_INPUT_RING_TYPE_TX        UINT32_C(0x1)
+	#define HWRM_RING_RESET_INPUT_RING_TYPE_TX          UINT32_C(0x1)
 	/* RX Ring (RR) */
-	#define HWRM_RING_FREE_INPUT_RING_TYPE_RX        UINT32_C(0x2)
+	#define HWRM_RING_RESET_INPUT_RING_TYPE_RX          UINT32_C(0x2)
 	/* RoCE Notification Completion Ring (ROCE_CR) */
-	#define HWRM_RING_FREE_INPUT_RING_TYPE_ROCE_CMPL UINT32_C(0x3)
-	/* RX Aggregation Ring */
-	#define HWRM_RING_FREE_INPUT_RING_TYPE_RX_AGG    UINT32_C(0x4)
-	/* Notification Queue */
-	#define HWRM_RING_FREE_INPUT_RING_TYPE_NQ        UINT32_C(0x5)
-	#define HWRM_RING_FREE_INPUT_RING_TYPE_LAST \
-		HWRM_RING_FREE_INPUT_RING_TYPE_NQ
-	uint8_t	unused_0;
-	/* Physical number of ring allocated. */
-	uint16_t	ring_id;
-	uint8_t	unused_1[4];
-} __rte_packed;
-
-/* hwrm_ring_free_output (size:128b/16B) */
-struct hwrm_ring_free_output {
-	/* The specific error status for the command. */
-	uint16_t	error_code;
-	/* The HWRM command request type. */
-	uint16_t	req_type;
-	/* The sequence ID from the original command. */
-	uint16_t	seq_id;
-	/* The length of the response data in number of bytes. */
-	uint16_t	resp_len;
-	uint8_t	unused_0[7];
+	#define HWRM_RING_RESET_INPUT_RING_TYPE_ROCE_CMPL   UINT32_C(0x3)
 	/*
-	 * This field is used in Output records to indicate that the output
-	 * is completely written to RAM.  This field should be read as '1'
-	 * to indicate that the output has been completely written.
-	 * When writing a command completion or response to an internal processor,
-	 * the order of writes has to be such that this field is written last.
+	 * Rx Ring Group.  This is to reset rx and aggregation in an atomic
+	 * operation. Completion ring associated with this ring group is
+	 * not reset.
 	 */
-	uint8_t	valid;
-} __rte_packed;
-
-/*******************
- * hwrm_ring_reset *
- *******************/
-
-
-/* hwrm_ring_reset_input (size:192b/24B) */
-struct hwrm_ring_reset_input {
-	/* The HWRM command request type. */
-	uint16_t	req_type;
-	/*
-	 * The completion ring to send the completion event on. This should
-	 * be the NQ ID returned from the `nq_alloc` HWRM command.
-	 */
-	uint16_t	cmpl_ring;
-	/*
-	 * The sequence ID is used by the driver for tracking multiple
-	 * commands. This ID is treated as opaque data by the firmware and
-	 * the value is returned in the `hwrm_resp_hdr` upon completion.
-	 */
-	uint16_t	seq_id;
-	/*
-	 * The target ID of the command:
-	 * * 0x0-0xFFF8 - The function ID
-	 * * 0xFFF8-0xFFFC, 0xFFFE - Reserved for internal processors
-	 * * 0xFFFD - Reserved for user-space HWRM interface
-	 * * 0xFFFF - HWRM
-	 */
-	uint16_t	target_id;
-	/*
-	 * A physical address pointer pointing to a host buffer that the
-	 * command's response data will be written. This can be either a host
-	 * physical address (HPA) or a guest physical address (GPA) and must
-	 * point to a physically contiguous block of memory.
-	 */
-	uint64_t	resp_addr;
-	/* Ring Type. */
-	uint8_t	ring_type;
-	/* L2 Completion Ring (CR) */
-	#define HWRM_RING_RESET_INPUT_RING_TYPE_L2_CMPL   UINT32_C(0x0)
-	/* TX Ring (TR) */
-	#define HWRM_RING_RESET_INPUT_RING_TYPE_TX        UINT32_C(0x1)
-	/* RX Ring (RR) */
-	#define HWRM_RING_RESET_INPUT_RING_TYPE_RX        UINT32_C(0x2)
-	/* RoCE Notification Completion Ring (ROCE_CR) */
-	#define HWRM_RING_RESET_INPUT_RING_TYPE_ROCE_CMPL UINT32_C(0x3)
+	#define HWRM_RING_RESET_INPUT_RING_TYPE_RX_RING_GRP UINT32_C(0x6)
 	#define HWRM_RING_RESET_INPUT_RING_TYPE_LAST \
-		HWRM_RING_RESET_INPUT_RING_TYPE_ROCE_CMPL
+		HWRM_RING_RESET_INPUT_RING_TYPE_RX_RING_GRP
 	uint8_t	unused_0;
-	/* Physical number of the ring. */
+	/*
+	 * Physical number of the ring. When ring type is rx_ring_grp, ring id
+	 * actually refers to ring group id.
+	 */
 	uint16_t	ring_id;
 	uint8_t	unused_1[4];
 } __rte_packed;
@@ -25615,7 +27319,18 @@ struct hwrm_ring_cmpl_ring_qaggint_params_input {
 	uint64_t	resp_addr;
 	/* Physical number of completion ring. */
 	uint16_t	ring_id;
-	uint8_t	unused_0[6];
+	uint16_t	flags;
+	#define HWRM_RING_CMPL_RING_QAGGINT_PARAMS_INPUT_FLAGS_UNUSED_0_MASK \
+		UINT32_C(0x3)
+	#define HWRM_RING_CMPL_RING_QAGGINT_PARAMS_INPUT_FLAGS_UNUSED_0_SFT 0
+	/*
+	 * Set this flag to 1 when querying parameters on a notification
+	 * queue. Set this flag to 0 when querying parameters on a
+	 * completion queue or completion ring.
+	 */
+	#define HWRM_RING_CMPL_RING_QAGGINT_PARAMS_INPUT_FLAGS_IS_NQ \
+		UINT32_C(0x4)
+	uint8_t	unused_0[4];
 } __rte_packed;
 
 /* hwrm_ring_cmpl_ring_qaggint_params_output (size:256b/32B) */
@@ -25652,19 +27367,19 @@ struct hwrm_ring_cmpl_ring_qaggint_params_output {
 	 */
 	uint16_t	num_cmpl_dma_aggr_during_int;
 	/*
-	 * Timer in unit of 80-nsec used to aggregate completions before
+	 * Timer used to aggregate completions before
 	 * DMA during the normal mode (not in interrupt mode).
 	 */
 	uint16_t	cmpl_aggr_dma_tmr;
 	/*
-	 * Timer in unit of 80-nsec used to aggregate completions before
-	 * DMA during the interrupt mode.
+	 * Timer used to aggregate completions before
+	 * DMA when in interrupt mode.
 	 */
 	uint16_t	cmpl_aggr_dma_tmr_during_int;
-	/* Minimum time (in unit of 80-nsec) between two interrupts. */
+	/* Minimum time between two interrupts. */
 	uint16_t	int_lat_tmr_min;
 	/*
-	 * Maximum wait time (in unit of 80-nsec) spent aggregating
+	 * Maximum wait time spent aggregating
 	 * completions before signaling the interrupt after the
 	 * interrupt is enabled.
 	 */
@@ -25738,7 +27453,7 @@ struct hwrm_ring_cmpl_ring_cfg_aggint_params_input {
 	/*
 	 * Set this flag to 1 when configuring parameters on a
 	 * notification queue. Set this flag to 0 when configuring
-	 * parameters on a completion queue.
+	 * parameters on a completion queue or completion ring.
 	 */
 	#define HWRM_RING_CMPL_RING_CFG_AGGINT_PARAMS_INPUT_FLAGS_IS_NQ \
 		UINT32_C(0x4)
@@ -25753,20 +27468,20 @@ struct hwrm_ring_cmpl_ring_cfg_aggint_params_input {
 	 */
 	uint16_t	num_cmpl_dma_aggr_during_int;
 	/*
-	 * Timer in unit of 80-nsec used to aggregate completions before
+	 * Timer used to aggregate completions before
 	 * DMA during the normal mode (not in interrupt mode).
 	 */
 	uint16_t	cmpl_aggr_dma_tmr;
 	/*
-	 * Timer in unit of 80-nsec used to aggregate completions before
-	 * DMA during the interrupt mode.
+	 * Timer used to aggregate completions before
+	 * DMA while in interrupt mode.
 	 */
 	uint16_t	cmpl_aggr_dma_tmr_during_int;
-	/* Minimum time (in unit of 80-nsec) between two interrupts. */
+	/* Minimum time between two interrupts. */
 	uint16_t	int_lat_tmr_min;
 	/*
-	 * Maximum wait time (in unit of 80-nsec) spent aggregating
-	 * cmpls before signaling the interrupt after the
+	 * Maximum wait time spent aggregating
+	 * completions before signaling the interrupt after the
 	 * interrupt is enabled.
 	 */
 	uint16_t	int_lat_tmr_max;
@@ -33339,78 +35054,246 @@ struct hwrm_tf_version_get_input {
 	 * point to a physically contiguous block of memory.
 	 */
 	uint64_t	resp_addr;
-} __rte_packed;
-
-/* hwrm_tf_version_get_output (size:128b/16B) */
-struct hwrm_tf_version_get_output {
-	/* The specific error status for the command. */
-	uint16_t	error_code;
-	/* The HWRM command request type. */
-	uint16_t	req_type;
-	/* The sequence ID from the original command. */
-	uint16_t	seq_id;
-	/* The length of the response data in number of bytes. */
-	uint16_t	resp_len;
-	/* Version Major number. */
-	uint8_t	major;
-	/* Version Minor number. */
-	uint8_t	minor;
-	/* Version Update number. */
-	uint8_t	update;
-	/* unused. */
-	uint8_t	unused0[4];
+} __rte_packed;
+
+/* hwrm_tf_version_get_output (size:128b/16B) */
+struct hwrm_tf_version_get_output {
+	/* The specific error status for the command. */
+	uint16_t	error_code;
+	/* The HWRM command request type. */
+	uint16_t	req_type;
+	/* The sequence ID from the original command. */
+	uint16_t	seq_id;
+	/* The length of the response data in number of bytes. */
+	uint16_t	resp_len;
+	/* Version Major number. */
+	uint8_t	major;
+	/* Version Minor number. */
+	uint8_t	minor;
+	/* Version Update number. */
+	uint8_t	update;
+	/* unused. */
+	uint8_t	unused0[4];
+	/*
+	 * This field is used in Output records to indicate that the output
+	 * is completely written to RAM. This field should be read as '1'
+	 * to indicate that the output has been completely written.
+	 * When writing a command completion or response to an internal
+	 * processor, the order of writes has to be such that this field is
+	 * written last.
+	 */
+	uint8_t	valid;
+} __rte_packed;
+
+/************************
+ * hwrm_tf_session_open *
+ ************************/
+
+
+/* hwrm_tf_session_open_input (size:640b/80B) */
+struct hwrm_tf_session_open_input {
+	/* The HWRM command request type. */
+	uint16_t	req_type;
+	/*
+	 * The completion ring to send the completion event on. This should
+	 * be the NQ ID returned from the `nq_alloc` HWRM command.
+	 */
+	uint16_t	cmpl_ring;
+	/*
+	 * The sequence ID is used by the driver for tracking multiple
+	 * commands. This ID is treated as opaque data by the firmware and
+	 * the value is returned in the `hwrm_resp_hdr` upon completion.
+	 */
+	uint16_t	seq_id;
+	/*
+	 * The target ID of the command:
+	 * * 0x0-0xFFF8 - The function ID
+	 * * 0xFFF8-0xFFFC, 0xFFFE - Reserved for internal processors
+	 * * 0xFFFD - Reserved for user-space HWRM interface
+	 * * 0xFFFF - HWRM
+	 */
+	uint16_t	target_id;
+	/*
+	 * A physical address pointer pointing to a host buffer that the
+	 * command's response data will be written. This can be either a host
+	 * physical address (HPA) or a guest physical address (GPA) and must
+	 * point to a physically contiguous block of memory.
+	 */
+	uint64_t	resp_addr;
+	/* Name of the session. */
+	uint8_t	session_name[64];
+} __rte_packed;
+
+/* hwrm_tf_session_open_output (size:192b/24B) */
+struct hwrm_tf_session_open_output {
+	/* The specific error status for the command. */
+	uint16_t	error_code;
+	/* The HWRM command request type. */
+	uint16_t	req_type;
+	/* The sequence ID from the original command. */
+	uint16_t	seq_id;
+	/* The length of the response data in number of bytes. */
+	uint16_t	resp_len;
+	/*
+	 * Unique session identifier for the session created by the
+	 * firmware.
+	 */
+	uint32_t	fw_session_id;
+	/*
+	 * Unique session client identifier for the first client on
+	 * the newly created session.
+	 */
+	uint32_t	fw_session_client_id;
+	/* unused. */
+	uint32_t	unused0;
+	/* unused. */
+	uint8_t	unused1[3];
+	/*
+	 * This field is used in Output records to indicate that the output
+	 * is completely written to RAM. This field should be read as '1'
+	 * to indicate that the output has been completely written.
+	 * When writing a command completion or response to an internal
+	 * processor, the order of writes has to be such that this field is
+	 * written last.
+	 */
+	uint8_t	valid;
+} __rte_packed;
+
+/**************************
+ * hwrm_tf_session_attach *
+ **************************/
+
+
+/* hwrm_tf_session_attach_input (size:704b/88B) */
+struct hwrm_tf_session_attach_input {
+	/* The HWRM command request type. */
+	uint16_t	req_type;
+	/*
+	 * The completion ring to send the completion event on. This should
+	 * be the NQ ID returned from the `nq_alloc` HWRM command.
+	 */
+	uint16_t	cmpl_ring;
+	/*
+	 * The sequence ID is used by the driver for tracking multiple
+	 * commands. This ID is treated as opaque data by the firmware and
+	 * the value is returned in the `hwrm_resp_hdr` upon completion.
+	 */
+	uint16_t	seq_id;
+	/*
+	 * The target ID of the command:
+	 * * 0x0-0xFFF8 - The function ID
+	 * * 0xFFF8-0xFFFC, 0xFFFE - Reserved for internal processors
+	 * * 0xFFFD - Reserved for user-space HWRM interface
+	 * * 0xFFFF - HWRM
+	 */
+	uint16_t	target_id;
+	/*
+	 * A physical address pointer pointing to a host buffer that the
+	 * command's response data will be written. This can be either a host
+	 * physical address (HPA) or a guest physical address (GPA) and must
+	 * point to a physically contiguous block of memory.
+	 */
+	uint64_t	resp_addr;
+	/*
+	 * Unique session identifier for the session that the attach
+	 * request want to attach to. This value originates from the
+	 * shared session memory that the attach request opened by
+	 * way of the 'attach name' that was passed in to the core
+	 * attach API.
+	 * The fw_session_id of the attach session includes PCIe bus
+	 * info to distinguish the PF and session info to identify
+	 * the associated TruFlow session.
+	 */
+	uint32_t	attach_fw_session_id;
+	/* unused. */
+	uint32_t	unused0;
+	/* Name of the session it self. */
+	uint8_t	session_name[64];
+} __rte_packed;
+
+/* hwrm_tf_session_attach_output (size:128b/16B) */
+struct hwrm_tf_session_attach_output {
+	/* The specific error status for the command. */
+	uint16_t	error_code;
+	/* The HWRM command request type. */
+	uint16_t	req_type;
+	/* The sequence ID from the original command. */
+	uint16_t	seq_id;
+	/* The length of the response data in number of bytes. */
+	uint16_t	resp_len;
+	/*
+	 * Unique session identifier for the session created by the
+	 * firmware. It includes PCIe bus info to distinguish the PF
+	 * and session info to identify the associated TruFlow
+	 * session. This fw_session_id is unique to the attach
+	 * request.
+	 */
+	uint32_t	fw_session_id;
+	/* unused. */
+	uint8_t	unused0[3];
+	/*
+	 * This field is used in Output records to indicate that the output
+	 * is completely written to RAM. This field should be read as '1'
+	 * to indicate that the output has been completely written.
+	 * When writing a command completion or response to an internal
+	 * processor, the order of writes has to be such that this field is
+	 * written last.
+	 */
+	uint8_t	valid;
+} __rte_packed;
+
+/****************************
+ * hwrm_tf_session_register *
+ ****************************/
+
+
+/* hwrm_tf_session_register_input (size:704b/88B) */
+struct hwrm_tf_session_register_input {
+	/* The HWRM command request type. */
+	uint16_t	req_type;
+	/*
+	 * The completion ring to send the completion event on. This should
+	 * be the NQ ID returned from the `nq_alloc` HWRM command.
+	 */
+	uint16_t	cmpl_ring;
+	/*
+	 * The sequence ID is used by the driver for tracking multiple
+	 * commands. This ID is treated as opaque data by the firmware and
+	 * the value is returned in the `hwrm_resp_hdr` upon completion.
+	 */
+	uint16_t	seq_id;
+	/*
+	 * The target ID of the command:
+	 * * 0x0-0xFFF8 - The function ID
+	 * * 0xFFF8-0xFFFC, 0xFFFE - Reserved for internal processors
+	 * * 0xFFFD - Reserved for user-space HWRM interface
+	 * * 0xFFFF - HWRM
+	 */
+	uint16_t	target_id;
+	/*
+	 * A physical address pointer pointing to a host buffer that the
+	 * command's response data will be written. This can be either a host
+	 * physical address (HPA) or a guest physical address (GPA) and must
+	 * point to a physically contiguous block of memory.
+	 */
+	uint64_t	resp_addr;
 	/*
-	 * This field is used in Output records to indicate that the output
-	 * is completely written to RAM. This field should be read as '1'
-	 * to indicate that the output has been completely written.
-	 * When writing a command completion or response to an internal
-	 * processor, the order of writes has to be such that this field is
-	 * written last.
-	 */
-	uint8_t	valid;
-} __rte_packed;
-
-/************************
- * hwrm_tf_session_open *
- ************************/
-
-
-/* hwrm_tf_session_open_input (size:640b/80B) */
-struct hwrm_tf_session_open_input {
-	/* The HWRM command request type. */
-	uint16_t	req_type;
-	/*
-	 * The completion ring to send the completion event on. This should
-	 * be the NQ ID returned from the `nq_alloc` HWRM command.
-	 */
-	uint16_t	cmpl_ring;
-	/*
-	 * The sequence ID is used by the driver for tracking multiple
-	 * commands. This ID is treated as opaque data by the firmware and
-	 * the value is returned in the `hwrm_resp_hdr` upon completion.
-	 */
-	uint16_t	seq_id;
-	/*
-	 * The target ID of the command:
-	 * * 0x0-0xFFF8 - The function ID
-	 * * 0xFFF8-0xFFFC, 0xFFFE - Reserved for internal processors
-	 * * 0xFFFD - Reserved for user-space HWRM interface
-	 * * 0xFFFF - HWRM
-	 */
-	uint16_t	target_id;
-	/*
-	 * A physical address pointer pointing to a host buffer that the
-	 * command's response data will be written. This can be either a host
-	 * physical address (HPA) or a guest physical address (GPA) and must
-	 * point to a physically contiguous block of memory.
+	 * Unique session identifier for the session that the
+	 * register request want to create a new client on. This
+	 * value originates from the first open request.
+	 * The fw_session_id of the attach session includes PCIe bus
+	 * info to distinguish the PF and session info to identify
+	 * the associated TruFlow session.
 	 */
-	uint64_t	resp_addr;
-	/* Name of the session. */
-	uint8_t	session_name[64];
+	uint32_t	fw_session_id;
+	/* unused. */
+	uint32_t	unused0;
+	/* Name of the session client. */
+	uint8_t	session_client_name[64];
 } __rte_packed;
 
-/* hwrm_tf_session_open_output (size:128b/16B) */
-struct hwrm_tf_session_open_output {
+/* hwrm_tf_session_register_output (size:128b/16B) */
+struct hwrm_tf_session_register_output {
 	/* The specific error status for the command. */
 	uint16_t	error_code;
 	/* The HWRM command request type. */
@@ -33420,12 +35303,11 @@ struct hwrm_tf_session_open_output {
 	/* The length of the response data in number of bytes. */
 	uint16_t	resp_len;
 	/*
-	 * Unique session identifier for the session created by the
-	 * firmware. It includes PCIe bus info to distinguish the PF
-	 * and session info to identify the associated TruFlow
-	 * session.
+	 * Unique session client identifier for the session created
+	 * by the firmware. It includes the session the client it
+	 * attached to and session client info.
 	 */
-	uint32_t	fw_session_id;
+	uint32_t	fw_session_client_id;
 	/* unused. */
 	uint8_t	unused0[3];
 	/*
@@ -33439,13 +35321,13 @@ struct hwrm_tf_session_open_output {
 	uint8_t	valid;
 } __rte_packed;
 
-/**************************
- * hwrm_tf_session_attach *
- **************************/
+/******************************
+ * hwrm_tf_session_unregister *
+ ******************************/
 
 
-/* hwrm_tf_session_attach_input (size:704b/88B) */
-struct hwrm_tf_session_attach_input {
+/* hwrm_tf_session_unregister_input (size:192b/24B) */
+struct hwrm_tf_session_unregister_input {
 	/* The HWRM command request type. */
 	uint16_t	req_type;
 	/*
@@ -33475,24 +35357,19 @@ struct hwrm_tf_session_attach_input {
 	 */
 	uint64_t	resp_addr;
 	/*
-	 * Unique session identifier for the session that the attach
-	 * request want to attach to. This value originates from the
-	 * shared session memory that the attach request opened by
-	 * way of the 'attach name' that was passed in to the core
-	 * attach API.
-	 * The fw_session_id of the attach session includes PCIe bus
-	 * info to distinguish the PF and session info to identify
-	 * the associated TruFlow session.
+	 * Unique session identifier for the session that the
+	 * unregister request want to close a session client on.
 	 */
-	uint32_t	attach_fw_session_id;
-	/* unused. */
-	uint32_t	unused0;
-	/* Name of the session it self. */
-	uint8_t	session_name[64];
+	uint32_t	fw_session_id;
+	/*
+	 * Unique session client identifier for the session that the
+	 * unregister request want to close.
+	 */
+	uint32_t	fw_session_client_id;
 } __rte_packed;
 
-/* hwrm_tf_session_attach_output (size:128b/16B) */
-struct hwrm_tf_session_attach_output {
+/* hwrm_tf_session_unregister_output (size:128b/16B) */
+struct hwrm_tf_session_unregister_output {
 	/* The specific error status for the command. */
 	uint16_t	error_code;
 	/* The HWRM command request type. */
@@ -33501,16 +35378,8 @@ struct hwrm_tf_session_attach_output {
 	uint16_t	seq_id;
 	/* The length of the response data in number of bytes. */
 	uint16_t	resp_len;
-	/*
-	 * Unique session identifier for the session created by the
-	 * firmware. It includes PCIe bus info to distinguish the PF
-	 * and session info to identify the associated TruFlow
-	 * session. This fw_session_id is unique to the attach
-	 * request.
-	 */
-	uint32_t	fw_session_id;
 	/* unused. */
-	uint8_t	unused0[3];
+	uint8_t	unused0[7];
 	/*
 	 * This field is used in Output records to indicate that the output
 	 * is completely written to RAM. This field should be read as '1'
@@ -33746,15 +35615,17 @@ struct hwrm_tf_session_resc_qcaps_input {
 	#define HWRM_TF_SESSION_RESC_QCAPS_INPUT_FLAGS_DIR_LAST \
 		HWRM_TF_SESSION_RESC_QCAPS_INPUT_FLAGS_DIR_TX
 	/*
-	 * Defines the size, in bytes, of the provided qcaps_addr
+	 * Defines the size of the provided qcaps_addr array
 	 * buffer. The size should be set to the Resource Manager
-	 * provided max qcaps value that is device specific. This is
-	 * the max size possible.
+	 * provided max number of qcaps entries which is device
+	 * specific. Resource Manager gets the max size from HCAPI
+	 * RM.
 	 */
-	uint16_t	size;
+	uint16_t	qcaps_size;
 	/*
-	 * This is the DMA address for the qcaps output data
-	 * array. Array is of tf_rm_cap type and is device specific.
+	 * This is the DMA address for the qcaps output data array
+	 * buffer. Array is of tf_rm_resc_req_entry type and is
+	 * device specific.
 	 */
 	uint64_t	qcaps_addr;
 } __rte_packed;
@@ -33772,29 +35643,28 @@ struct hwrm_tf_session_resc_qcaps_output {
 	/* Control flags. */
 	uint32_t	flags;
 	/* Session reservation strategy. */
-	#define HWRM_TF_SESSION_RESC_QCAPS_OUTPUT_FLAGS_SESS_RES_STRATEGY_MASK \
+	#define HWRM_TF_SESSION_RESC_QCAPS_OUTPUT_FLAGS_SESS_RESV_STRATEGY_MASK \
 		UINT32_C(0x3)
-	#define HWRM_TF_SESSION_RESC_QCAPS_OUTPUT_FLAGS_SESS_RES_STRATEGY_SFT \
+	#define HWRM_TF_SESSION_RESC_QCAPS_OUTPUT_FLAGS_SESS_RESV_STRATEGY_SFT \
 		0
 	/* Static partitioning. */
-	#define HWRM_TF_SESSION_RESC_QCAPS_OUTPUT_FLAGS_SESS_RES_STRATEGY_STATIC \
+	#define HWRM_TF_SESSION_RESC_QCAPS_OUTPUT_FLAGS_SESS_RESV_STRATEGY_STATIC \
 		UINT32_C(0x0)
 	/* Strategy 1. */
-	#define HWRM_TF_SESSION_RESC_QCAPS_OUTPUT_FLAGS_SESS_RES_STRATEGY_1 \
+	#define HWRM_TF_SESSION_RESC_QCAPS_OUTPUT_FLAGS_SESS_RESV_STRATEGY_1 \
 		UINT32_C(0x1)
 	/* Strategy 2. */
-	#define HWRM_TF_SESSION_RESC_QCAPS_OUTPUT_FLAGS_SESS_RES_STRATEGY_2 \
+	#define HWRM_TF_SESSION_RESC_QCAPS_OUTPUT_FLAGS_SESS_RESV_STRATEGY_2 \
 		UINT32_C(0x2)
 	/* Strategy 3. */
-	#define HWRM_TF_SESSION_RESC_QCAPS_OUTPUT_FLAGS_SESS_RES_STRATEGY_3 \
+	#define HWRM_TF_SESSION_RESC_QCAPS_OUTPUT_FLAGS_SESS_RESV_STRATEGY_3 \
 		UINT32_C(0x3)
-	#define HWRM_TF_SESSION_RESC_QCAPS_OUTPUT_FLAGS_SESS_RES_STRATEGY_LAST \
-		HWRM_TF_SESSION_RESC_QCAPS_OUTPUT_FLAGS_SESS_RES_STRATEGY_3
+	#define HWRM_TF_SESSION_RESC_QCAPS_OUTPUT_FLAGS_SESS_RESV_STRATEGY_LAST \
+		HWRM_TF_SESSION_RESC_QCAPS_OUTPUT_FLAGS_SESS_RESV_STRATEGY_3
 	/*
-	 * Size of the returned tf_rm_cap data array. The value
-	 * cannot exceed the size defined by the input msg. The data
-	 * array is returned using the qcaps_addr specified DMA
-	 * address also provided by the input msg.
+	 * Size of the returned qcaps_addr data array buffer. The
+	 * value cannot exceed the size defined by the input msg,
+	 * qcaps_size.
 	 */
 	uint16_t	size;
 	/* unused. */
@@ -33817,7 +35687,7 @@ struct hwrm_tf_session_resc_qcaps_output {
  ******************************/
 
 
-/* hwrm_tf_session_resc_alloc_input (size:256b/32B) */
+/* hwrm_tf_session_resc_alloc_input (size:320b/40B) */
 struct hwrm_tf_session_resc_alloc_input {
 	/* The HWRM command request type. */
 	uint16_t	req_type;
@@ -33860,16 +35730,25 @@ struct hwrm_tf_session_resc_alloc_input {
 	#define HWRM_TF_SESSION_RESC_ALLOC_INPUT_FLAGS_DIR_LAST \
 		HWRM_TF_SESSION_RESC_ALLOC_INPUT_FLAGS_DIR_TX
 	/*
-	 * Defines the size, in bytes, of the provided num_addr
-	 * buffer.
+	 * Defines the array size of the provided req_addr and
+	 * resv_addr array buffers. Should be set to the number of
+	 * request entries.
 	 */
-	uint16_t	size;
+	uint16_t	req_size;
+	/*
+	 * This is the DMA address for the request input data array
+	 * buffer. Array is of tf_rm_resc_req_entry type. Size of the
+	 * array buffer is provided by the 'req_size' field in this
+	 * message.
+	 */
+	uint64_t	req_addr;
 	/*
-	 * This is the DMA address for the num input data array
-	 * buffer. Array is of tf_rm_num type. Size of the buffer is
-	 * provided by the 'size' field in this message.
+	 * This is the DMA address for the resc output data array
+	 * buffer. Array is of tf_rm_resc_entry type. Size of the array
+	 * buffer is provided by the 'req_size' field in this
+	 * message.
 	 */
-	uint64_t	num_addr;
+	uint64_t	resc_addr;
 } __rte_packed;
 
 /* hwrm_tf_session_resc_alloc_output (size:128b/16B) */
@@ -33882,8 +35761,15 @@ struct hwrm_tf_session_resc_alloc_output {
 	uint16_t	seq_id;
 	/* The length of the response data in number of bytes. */
 	uint16_t	resp_len;
+	/*
+	 * Size of the returned tf_rm_resc_entry data array. The value
+	 * cannot exceed the req_size defined by the input msg. The data
+	 * array is returned using the resv_addr specified DMA
+	 * address also provided by the input msg.
+	 */
+	uint16_t	size;
 	/* unused. */
-	uint8_t	unused0[7];
+	uint8_t	unused0[5];
 	/*
 	 * This field is used in Output records to indicate that the output
 	 * is completely written to RAM. This field should be read as '1'
@@ -33946,11 +35832,12 @@ struct hwrm_tf_session_resc_free_input {
 	 * Defines the size, in bytes, of the provided free_addr
 	 * buffer.
 	 */
-	uint16_t	size;
+	uint16_t	free_size;
 	/*
 	 * This is the DMA address for the free input data array
-	 * buffer.  Array of tf_rm_res type. Size of the buffer is
-	 * provided by the 'size field of this message.
+	 * buffer.  Array is of tf_rm_resc_entry type. Size of the
+	 * buffer is provided by the 'free_size' field of this
+	 * message.
 	 */
 	uint64_t	free_addr;
 } __rte_packed;
@@ -34029,11 +35916,12 @@ struct hwrm_tf_session_resc_flush_input {
 	 * Defines the size, in bytes, of the provided flush_addr
 	 * buffer.
 	 */
-	uint16_t	size;
+	uint16_t	flush_size;
 	/*
 	 * This is the DMA address for the flush input data array
-	 * buffer.  Array of tf_rm_res type. Size of the buffer is
-	 * provided by the 'size' field in this message.
+	 * buffer.  Array of tf_rm_resc_entry type. Size of the
+	 * buffer is provided by the 'flush_size' field in this
+	 * message.
 	 */
 	uint64_t	flush_addr;
 } __rte_packed;
@@ -34062,12 +35950,9 @@ struct hwrm_tf_session_resc_flush_output {
 } __rte_packed;
 
 /* TruFlow RM capability of a resource. */
-/* tf_rm_cap (size:64b/8B) */
-struct tf_rm_cap {
-	/*
-	 * Type of the resource, defined globally in the
-	 * hwrm_tf_resc_type enum.
-	 */
+/* tf_rm_resc_req_entry (size:64b/8B) */
+struct tf_rm_resc_req_entry {
+	/* Type of the resource, defined globally in HCAPI RM. */
 	uint32_t	type;
 	/* Minimum value. */
 	uint16_t	min;
@@ -34075,25 +35960,10 @@ struct tf_rm_cap {
 	uint16_t	max;
 } __rte_packed;
 
-/* TruFlow RM number of a resource. */
-/* tf_rm_num (size:64b/8B) */
-struct tf_rm_num {
-	/*
-	 * Type of the resource, defined globally in the
-	 * hwrm_tf_resc_type enum.
-	 */
-	uint32_t	type;
-	/* Number of resources. */
-	uint32_t	num;
-} __rte_packed;
-
 /* TruFlow RM reservation information. */
-/* tf_rm_res (size:64b/8B) */
-struct tf_rm_res {
-	/*
-	 * Type of the resource, defined globally in the
-	 * hwrm_tf_resc_type enum.
-	 */
+/* tf_rm_resc_entry (size:64b/8B) */
+struct tf_rm_resc_entry {
+	/* Type of the resource, defined globally in HCAPI RM. */
 	uint32_t	type;
 	/* Start offset. */
 	uint16_t	start;
@@ -34925,6 +36795,162 @@ struct hwrm_tf_ext_em_qcfg_output {
 	uint8_t	valid;
 } __rte_packed;
 
+/*********************
+ * hwrm_tf_em_insert *
+ *********************/
+
+
+/* hwrm_tf_em_insert_input (size:832b/104B) */
+struct hwrm_tf_em_insert_input {
+	/* The HWRM command request type. */
+	uint16_t	req_type;
+	/*
+	 * The completion ring to send the completion event on. This should
+	 * be the NQ ID returned from the `nq_alloc` HWRM command.
+	 */
+	uint16_t	cmpl_ring;
+	/*
+	 * The sequence ID is used by the driver for tracking multiple
+	 * commands. This ID is treated as opaque data by the firmware and
+	 * the value is returned in the `hwrm_resp_hdr` upon completion.
+	 */
+	uint16_t	seq_id;
+	/*
+	 * The target ID of the command:
+	 * * 0x0-0xFFF8 - The function ID
+	 * * 0xFFF8-0xFFFC, 0xFFFE - Reserved for internal processors
+	 * * 0xFFFD - Reserved for user-space HWRM interface
+	 * * 0xFFFF - HWRM
+	 */
+	uint16_t	target_id;
+	/*
+	 * A physical address pointer pointing to a host buffer that the
+	 * command's response data will be written. This can be either a host
+	 * physical address (HPA) or a guest physical address (GPA) and must
+	 * point to a physically contiguous block of memory.
+	 */
+	uint64_t	resp_addr;
+	/* Firmware Session Id. */
+	uint32_t	fw_session_id;
+	/* Control Flags. */
+	uint16_t	flags;
+	/* Indicates the flow direction. */
+	#define HWRM_TF_EM_INSERT_INPUT_FLAGS_DIR     UINT32_C(0x1)
+	/* If this bit set to 0, then it indicates rx flow. */
+	#define HWRM_TF_EM_INSERT_INPUT_FLAGS_DIR_RX    UINT32_C(0x0)
+	/* If this bit is set to 1, then it indicates that tx flow. */
+	#define HWRM_TF_EM_INSERT_INPUT_FLAGS_DIR_TX    UINT32_C(0x1)
+	#define HWRM_TF_EM_INSERT_INPUT_FLAGS_DIR_LAST \
+		HWRM_TF_EM_INSERT_INPUT_FLAGS_DIR_TX
+	/* Reported match strength. */
+	uint16_t	strength;
+	/* Index to action. */
+	uint32_t	action_ptr;
+	/* Index of EM record. */
+	uint32_t	em_record_idx;
+	/* EM Key value. */
+	uint64_t	em_key[8];
+	/* Number of bits in em_key. */
+	uint16_t	em_key_bitlen;
+	/* unused. */
+	uint16_t	unused0[3];
+} __rte_packed;
+
+/* hwrm_tf_em_insert_output (size:128b/16B) */
+struct hwrm_tf_em_insert_output {
+	/* The specific error status for the command. */
+	uint16_t	error_code;
+	/* The HWRM command request type. */
+	uint16_t	req_type;
+	/* The sequence ID from the original command. */
+	uint16_t	seq_id;
+	/* The length of the response data in number of bytes. */
+	uint16_t	resp_len;
+	/* EM record pointer index. */
+	uint16_t	rptr_index;
+	/* EM record offset 0~3. */
+	uint8_t	rptr_entry;
+	/* Number of word entries consumed by the key. */
+	uint8_t	num_of_entries;
+	/* unused. */
+	uint32_t	unused0;
+} __rte_packed;
+
+/*********************
+ * hwrm_tf_em_delete *
+ *********************/
+
+
+/* hwrm_tf_em_delete_input (size:832b/104B) */
+struct hwrm_tf_em_delete_input {
+	/* The HWRM command request type. */
+	uint16_t	req_type;
+	/*
+	 * The completion ring to send the completion event on. This should
+	 * be the NQ ID returned from the `nq_alloc` HWRM command.
+	 */
+	uint16_t	cmpl_ring;
+	/*
+	 * The sequence ID is used by the driver for tracking multiple
+	 * commands. This ID is treated as opaque data by the firmware and
+	 * the value is returned in the `hwrm_resp_hdr` upon completion.
+	 */
+	uint16_t	seq_id;
+	/*
+	 * The target ID of the command:
+	 * * 0x0-0xFFF8 - The function ID
+	 * * 0xFFF8-0xFFFC, 0xFFFE - Reserved for internal processors
+	 * * 0xFFFD - Reserved for user-space HWRM interface
+	 * * 0xFFFF - HWRM
+	 */
+	uint16_t	target_id;
+	/*
+	 * A physical address pointer pointing to a host buffer that the
+	 * command's response data will be written. This can be either a host
+	 * physical address (HPA) or a guest physical address (GPA) and must
+	 * point to a physically contiguous block of memory.
+	 */
+	uint64_t	resp_addr;
+	/* Session Id. */
+	uint32_t	fw_session_id;
+	/* Control flags. */
+	uint16_t	flags;
+	/* Indicates the flow direction. */
+	#define HWRM_TF_EM_DELETE_INPUT_FLAGS_DIR     UINT32_C(0x1)
+	/* If this bit set to 0, then it indicates rx flow. */
+	#define HWRM_TF_EM_DELETE_INPUT_FLAGS_DIR_RX    UINT32_C(0x0)
+	/* If this bit is set to 1, then it indicates that tx flow. */
+	#define HWRM_TF_EM_DELETE_INPUT_FLAGS_DIR_TX    UINT32_C(0x1)
+	#define HWRM_TF_EM_DELETE_INPUT_FLAGS_DIR_LAST \
+		HWRM_TF_EM_DELETE_INPUT_FLAGS_DIR_TX
+	/* Unused0 */
+	uint16_t	unused0;
+	/* EM internal flow hanndle. */
+	uint64_t	flow_handle;
+	/* EM Key value */
+	uint64_t	em_key[8];
+	/* Number of bits in em_key. */
+	uint16_t	em_key_bitlen;
+	/* unused. */
+	uint16_t	unused1[3];
+} __rte_packed;
+
+/* hwrm_tf_em_delete_output (size:128b/16B) */
+struct hwrm_tf_em_delete_output {
+	/* The specific error status for the command. */
+	uint16_t	error_code;
+	/* The HWRM command request type. */
+	uint16_t	req_type;
+	/* The sequence ID from the original command. */
+	uint16_t	seq_id;
+	/* The length of the response data in number of bytes. */
+	uint16_t	resp_len;
+	/* Original stack allocation index. */
+	uint16_t	em_index;
+	/* unused. */
+	uint16_t	unused0[3];
+} __rte_packed;
+
 /********************
  * hwrm_tf_tcam_set *
  ********************/
@@ -35582,10 +37608,10 @@ struct ctx_hw_stats {
 	uint64_t	rx_mcast_pkts;
 	/* Number of received broadcast packets */
 	uint64_t	rx_bcast_pkts;
-	/* Number of discarded packets on received path */
+	/* Number of discarded packets on receive path */
 	uint64_t	rx_discard_pkts;
-	/* Number of dropped packets on received path */
-	uint64_t	rx_drop_pkts;
+	/* Number of packets on receive path with error */
+	uint64_t	rx_error_pkts;
 	/* Number of received bytes for unicast traffic */
 	uint64_t	rx_ucast_bytes;
 	/* Number of received bytes for multicast traffic */
@@ -35598,10 +37624,10 @@ struct ctx_hw_stats {
 	uint64_t	tx_mcast_pkts;
 	/* Number of transmitted broadcast packets */
 	uint64_t	tx_bcast_pkts;
+	/* Number of packets on transmit path with error */
+	uint64_t	tx_error_pkts;
 	/* Number of discarded packets on transmit path */
 	uint64_t	tx_discard_pkts;
-	/* Number of dropped packets on transmit path */
-	uint64_t	tx_drop_pkts;
 	/* Number of transmitted bytes for unicast traffic */
 	uint64_t	tx_ucast_bytes;
 	/* Number of transmitted bytes for multicast traffic */
@@ -35618,7 +37644,11 @@ struct ctx_hw_stats {
 	uint64_t	tpa_aborts;
 } __rte_packed;
 
-/* Periodic statistics context DMA to host. */
+/*
+ * Extended periodic statistics context DMA to host. On cards that
+ * support TPA v2, additional TPA related stats exist and can be retrieved
+ * by DMA of ctx_hw_stats_ext, rather than legacy ctx_hw_stats structure.
+ */
 /* ctx_hw_stats_ext (size:1344b/168B) */
 struct ctx_hw_stats_ext {
 	/* Number of received unicast packets */
@@ -35627,10 +37657,10 @@ struct ctx_hw_stats_ext {
 	uint64_t	rx_mcast_pkts;
 	/* Number of received broadcast packets */
 	uint64_t	rx_bcast_pkts;
-	/* Number of discarded packets on received path */
+	/* Number of discarded packets on receive path */
 	uint64_t	rx_discard_pkts;
-	/* Number of dropped packets on received path */
-	uint64_t	rx_drop_pkts;
+	/* Number of packets on receive path with error */
+	uint64_t	rx_error_pkts;
 	/* Number of received bytes for unicast traffic */
 	uint64_t	rx_ucast_bytes;
 	/* Number of received bytes for multicast traffic */
@@ -35643,10 +37673,10 @@ struct ctx_hw_stats_ext {
 	uint64_t	tx_mcast_pkts;
 	/* Number of transmitted broadcast packets */
 	uint64_t	tx_bcast_pkts;
+	/* Number of packets on transmit path with error */
+	uint64_t	tx_error_pkts;
 	/* Number of discarded packets on transmit path */
 	uint64_t	tx_discard_pkts;
-	/* Number of dropped packets on transmit path */
-	uint64_t	tx_drop_pkts;
 	/* Number of transmitted bytes for unicast traffic */
 	uint64_t	tx_ucast_bytes;
 	/* Number of transmitted bytes for multicast traffic */
@@ -35912,7 +37942,14 @@ struct hwrm_stat_ctx_query_input {
 	uint64_t	resp_addr;
 	/* ID of the statistics context that is being queried. */
 	uint32_t	stat_ctx_id;
-	uint8_t	unused_0[4];
+	uint8_t	flags;
+	/*
+	 * This bit is set to 1 when request is for a counter mask,
+	 * representing the width of each of the stats counters, rather
+	 * than counters themselves.
+	 */
+	#define HWRM_STAT_CTX_QUERY_INPUT_FLAGS_COUNTER_MASK     UINT32_C(0x1)
+	uint8_t	unused_0[3];
 } __rte_packed;
 
 /* hwrm_stat_ctx_query_output (size:1408b/176B) */
@@ -35949,7 +37986,7 @@ struct hwrm_stat_ctx_query_output {
 	uint64_t	rx_bcast_pkts;
 	/* Number of received packets with error */
 	uint64_t	rx_err_pkts;
-	/* Number of dropped packets on received path */
+	/* Number of dropped packets on receive path */
 	uint64_t	rx_drop_pkts;
 	/* Number of received bytes for unicast traffic */
 	uint64_t	rx_ucast_bytes;
@@ -35977,6 +38014,117 @@ struct hwrm_stat_ctx_query_output {
 } __rte_packed;
 
 /***************************
+ * hwrm_stat_ext_ctx_query *
+ ***************************/
+
+
+/* hwrm_stat_ext_ctx_query_input (size:192b/24B) */
+struct hwrm_stat_ext_ctx_query_input {
+	/* The HWRM command request type. */
+	uint16_t	req_type;
+	/*
+	 * The completion ring to send the completion event on. This should
+	 * be the NQ ID returned from the `nq_alloc` HWRM command.
+	 */
+	uint16_t	cmpl_ring;
+	/*
+	 * The sequence ID is used by the driver for tracking multiple
+	 * commands. This ID is treated as opaque data by the firmware and
+	 * the value is returned in the `hwrm_resp_hdr` upon completion.
+	 */
+	uint16_t	seq_id;
+	/*
+	 * The target ID of the command:
+	 * * 0x0-0xFFF8 - The function ID
+	 * * 0xFFF8-0xFFFC, 0xFFFE - Reserved for internal processors
+	 * * 0xFFFD - Reserved for user-space HWRM interface
+	 * * 0xFFFF - HWRM
+	 */
+	uint16_t	target_id;
+	/*
+	 * A physical address pointer pointing to a host buffer that the
+	 * command's response data will be written. This can be either a host
+	 * physical address (HPA) or a guest physical address (GPA) and must
+	 * point to a physically contiguous block of memory.
+	 */
+	uint64_t	resp_addr;
+	/* ID of the extended statistics context that is being queried. */
+	uint32_t	stat_ctx_id;
+	uint8_t	flags;
+	/*
+	 * This bit is set to 1 when request is for a counter mask,
+	 * representing the width of each of the stats counters, rather
+	 * than counters themselves.
+	 */
+	#define HWRM_STAT_EXT_CTX_QUERY_INPUT_FLAGS_COUNTER_MASK \
+		UINT32_C(0x1)
+	uint8_t	unused_0[3];
+} __rte_packed;
+
+/* hwrm_stat_ext_ctx_query_output (size:1472b/184B) */
+struct hwrm_stat_ext_ctx_query_output {
+	/* The specific error status for the command. */
+	uint16_t	error_code;
+	/* The HWRM command request type. */
+	uint16_t	req_type;
+	/* The sequence ID from the original command. */
+	uint16_t	seq_id;
+	/* The length of the response data in number of bytes. */
+	uint16_t	resp_len;
+	/* Number of received unicast packets */
+	uint64_t	rx_ucast_pkts;
+	/* Number of received multicast packets */
+	uint64_t	rx_mcast_pkts;
+	/* Number of received broadcast packets */
+	uint64_t	rx_bcast_pkts;
+	/* Number of discarded packets on receive path */
+	uint64_t	rx_discard_pkts;
+	/* Number of packets on receive path with error */
+	uint64_t	rx_error_pkts;
+	/* Number of received bytes for unicast traffic */
+	uint64_t	rx_ucast_bytes;
+	/* Number of received bytes for multicast traffic */
+	uint64_t	rx_mcast_bytes;
+	/* Number of received bytes for broadcast traffic */
+	uint64_t	rx_bcast_bytes;
+	/* Number of transmitted unicast packets */
+	uint64_t	tx_ucast_pkts;
+	/* Number of transmitted multicast packets */
+	uint64_t	tx_mcast_pkts;
+	/* Number of transmitted broadcast packets */
+	uint64_t	tx_bcast_pkts;
+	/* Number of packets on transmit path with error */
+	uint64_t	tx_error_pkts;
+	/* Number of discarded packets on transmit path */
+	uint64_t	tx_discard_pkts;
+	/* Number of transmitted bytes for unicast traffic */
+	uint64_t	tx_ucast_bytes;
+	/* Number of transmitted bytes for multicast traffic */
+	uint64_t	tx_mcast_bytes;
+	/* Number of transmitted bytes for broadcast traffic */
+	uint64_t	tx_bcast_bytes;
+	/* Number of TPA eligible packets */
+	uint64_t	rx_tpa_eligible_pkt;
+	/* Number of TPA eligible bytes */
+	uint64_t	rx_tpa_eligible_bytes;
+	/* Number of TPA packets */
+	uint64_t	rx_tpa_pkt;
+	/* Number of TPA bytes */
+	uint64_t	rx_tpa_bytes;
+	/* Number of TPA errors */
+	uint64_t	rx_tpa_errors;
+	uint8_t	unused_0[7];
+	/*
+	 * This field is used in Output records to indicate that the output
+	 * is completely written to RAM.  This field should be read as '1'
+	 * to indicate that the output has been completely written.
+	 * When writing a command completion or response to an internal processor,
+	 * the order of writes has to be such that this field is written last.
+	 */
+	uint8_t	valid;
+} __rte_packed;
+
+/***************************
  * hwrm_stat_ctx_eng_query *
  ***************************/
 
@@ -37565,6 +39713,13 @@ struct hwrm_nvm_install_update_input {
 	 */
 	#define HWRM_NVM_INSTALL_UPDATE_INPUT_FLAGS_ALLOWED_TO_DEFRAG \
 		UINT32_C(0x4)
+	/*
+	 * If set to 1, FW will verify the package in the "UPDATE" NVM item
+	 * without installing it. This flag is for FW internal use only.
+	 * Users should not set this flag. The request will otherwise fail.
+	 */
+	#define HWRM_NVM_INSTALL_UPDATE_INPUT_FLAGS_VERIFY_ONLY \
+		UINT32_C(0x8)
 	uint8_t	unused_0[2];
 } __rte_packed;
 
@@ -38115,6 +40270,72 @@ struct hwrm_nvm_validate_option_cmd_err {
 	uint8_t	unused_0[7];
 } __rte_packed;
 
+/****************
+ * hwrm_oem_cmd *
+ ****************/
+
+
+/* hwrm_oem_cmd_input (size:1024b/128B) */
+struct hwrm_oem_cmd_input {
+	/* The HWRM command request type. */
+	uint16_t	req_type;
+	/*
+	 * The completion ring to send the completion event on. This should
+	 * be the NQ ID returned from the `nq_alloc` HWRM command.
+	 */
+	uint16_t	cmpl_ring;
+	/*
+	 * The sequence ID is used by the driver for tracking multiple
+	 * commands. This ID is treated as opaque data by the firmware and
+	 * the value is returned in the `hwrm_resp_hdr` upon completion.
+	 */
+	uint16_t	seq_id;
+	/*
+	 * The target ID of the command:
+	 * * 0x0-0xFFF8 - The function ID
+	 * * 0xFFF8-0xFFFC, 0xFFFE - Reserved for internal processors
+	 * * 0xFFFD - Reserved for user-space HWRM interface
+	 * * 0xFFFF - HWRM
+	 */
+	uint16_t	target_id;
+	/*
+	 * A physical address pointer pointing to a host buffer that the
+	 * command's response data will be written. This can be either a host
+	 * physical address (HPA) or a guest physical address (GPA) and must
+	 * point to a physically contiguous block of memory.
+	 */
+	uint64_t	resp_addr;
+	uint32_t	IANA;
+	uint32_t	unused_0;
+	/* This field contains the vendor specific command data. */
+	uint32_t	oem_data[26];
+} __rte_packed;
+
+/* hwrm_oem_cmd_output (size:768b/96B) */
+struct hwrm_oem_cmd_output {
+	/* The specific error status for the command. */
+	uint16_t	error_code;
+	/* The HWRM command request type. */
+	uint16_t	req_type;
+	/* The sequence ID from the original command. */
+	uint16_t	seq_id;
+	/* The length of the response data in number of bytes. */
+	uint16_t	resp_len;
+	uint32_t	IANA;
+	uint32_t	unused_0;
+	/* This field contains the vendor specific response data. */
+	uint32_t	oem_data[18];
+	uint8_t	unused_1[7];
+	/*
+	 * This field is used in Output records to indicate that the output
+	 * is completely written to RAM.  This field should be read as '1'
+	 * to indicate that the output has been completely written.
+	 * When writing a command completion or response to an internal processor,
+	 * the order of writes has to be such that this field is written last.
+	 */
+	uint8_t	valid;
+} __rte_packed;
+
 /*****************
  * hwrm_fw_reset *
  ******************/
@@ -38338,6 +40559,55 @@ struct hwrm_port_ts_query_output {
 	uint8_t		valid;
 } __rte_packed;
 
+/*
+ * This structure is fixed at the beginning of the ChiMP SRAM (GRC
+ * offset: 0x31001F0). Host software is expected to read from this
+ * location for a defined signature. If it exists, the software can
+ * assume the presence of this structure and the validity of the
+ * FW_STATUS location in the next field.
+ */
+/* hcomm_status (size:64b/8B) */
+struct hcomm_status {
+	uint32_t	sig_ver;
+	/*
+	 * This field defines the version of the structure. The latest
+	 * version value is 1.
+	 */
+	#define HCOMM_STATUS_VER_MASK		UINT32_C(0xff)
+	#define HCOMM_STATUS_VER_SFT		0
+	#define HCOMM_STATUS_VER_LATEST		UINT32_C(0x1)
+	#define HCOMM_STATUS_VER_LAST		HCOMM_STATUS_VER_LATEST
+	/*
+	 * This field is to store the signature value to indicate the
+	 * presence of the structure.
+	 */
+	#define HCOMM_STATUS_SIGNATURE_MASK	UINT32_C(0xffffff00)
+	#define HCOMM_STATUS_SIGNATURE_SFT	8
+	#define HCOMM_STATUS_SIGNATURE_VAL	(UINT32_C(0x484353) << 8)
+	#define HCOMM_STATUS_SIGNATURE_LAST	HCOMM_STATUS_SIGNATURE_VAL
+	uint32_t	fw_status_loc;
+	#define HCOMM_STATUS_TRUE_ADDR_SPACE_MASK	UINT32_C(0x3)
+	#define HCOMM_STATUS_TRUE_ADDR_SPACE_SFT	0
+	/* PCIE configuration space */
+	#define HCOMM_STATUS_FW_STATUS_LOC_ADDR_SPACE_PCIE_CFG	UINT32_C(0x0)
+	/* GRC space */
+	#define HCOMM_STATUS_FW_STATUS_LOC_ADDR_SPACE_GRC	UINT32_C(0x1)
+	/* BAR0 space */
+	#define HCOMM_STATUS_FW_STATUS_LOC_ADDR_SPACE_BAR0	UINT32_C(0x2)
+	/* BAR1 space */
+	#define HCOMM_STATUS_FW_STATUS_LOC_ADDR_SPACE_BAR1	UINT32_C(0x3)
+	#define HCOMM_STATUS_FW_STATUS_LOC_ADDR_SPACE_LAST	\
+		HCOMM_STATUS_FW_STATUS_LOC_ADDR_SPACE_BAR1
+	/*
+	 * This offset where the fw_status register is located. The value
+	 * is generally 4-byte aligned.
+	 */
+	#define HCOMM_STATUS_TRUE_OFFSET_MASK		UINT32_C(0xfffffffc)
+	#define HCOMM_STATUS_TRUE_OFFSET_SFT		2
+} __rte_packed;
+/* This is the GRC offset where the hcomm_status struct resides. */
+#define HCOMM_STATUS_STRUCT_LOC		0x31001F0UL
+
 /**************************
  * hwrm_cfa_counter_qcaps *
  **************************/
@@ -38622,53 +40892,4 @@ struct hwrm_cfa_counter_qstats_output {
 	uint8_t	valid;
 } __rte_packed;
 
-/*
- * This structure is fixed at the beginning of the ChiMP SRAM (GRC
- * offset: 0x31001F0). Host software is expected to read from this
- * location for a defined signature. If it exists, the software can
- * assume the presence of this structure and the validity of the
- * FW_STATUS location in the next field.
- */
-/* hcomm_status (size:64b/8B) */
-struct hcomm_status {
-	uint32_t	sig_ver;
-	/*
-	 * This field defines the version of the structure. The latest
-	 * version value is 1.
-	 */
-	#define HCOMM_STATUS_VER_MASK		UINT32_C(0xff)
-	#define HCOMM_STATUS_VER_SFT		0
-	#define HCOMM_STATUS_VER_LATEST		UINT32_C(0x1)
-	#define HCOMM_STATUS_VER_LAST		HCOMM_STATUS_VER_LATEST
-	/*
-	 * This field is to store the signature value to indicate the
-	 * presence of the structure.
-	 */
-	#define HCOMM_STATUS_SIGNATURE_MASK	UINT32_C(0xffffff00)
-	#define HCOMM_STATUS_SIGNATURE_SFT	8
-	#define HCOMM_STATUS_SIGNATURE_VAL	(UINT32_C(0x484353) << 8)
-	#define HCOMM_STATUS_SIGNATURE_LAST	HCOMM_STATUS_SIGNATURE_VAL
-	uint32_t	fw_status_loc;
-	#define HCOMM_STATUS_TRUE_ADDR_SPACE_MASK	UINT32_C(0x3)
-	#define HCOMM_STATUS_TRUE_ADDR_SPACE_SFT	0
-	/* PCIE configuration space */
-	#define HCOMM_STATUS_FW_STATUS_LOC_ADDR_SPACE_PCIE_CFG	UINT32_C(0x0)
-	/* GRC space */
-	#define HCOMM_STATUS_FW_STATUS_LOC_ADDR_SPACE_GRC	UINT32_C(0x1)
-	/* BAR0 space */
-	#define HCOMM_STATUS_FW_STATUS_LOC_ADDR_SPACE_BAR0	UINT32_C(0x2)
-	/* BAR1 space */
-	#define HCOMM_STATUS_FW_STATUS_LOC_ADDR_SPACE_BAR1	UINT32_C(0x3)
-	#define HCOMM_STATUS_FW_STATUS_LOC_ADDR_SPACE_LAST	\
-		HCOMM_STATUS_FW_STATUS_LOC_ADDR_SPACE_BAR1
-	/*
-	 * This offset where the fw_status register is located. The value
-	 * is generally 4-byte aligned.
-	 */
-	#define HCOMM_STATUS_TRUE_OFFSET_MASK		UINT32_C(0xfffffffc)
-	#define HCOMM_STATUS_TRUE_OFFSET_SFT		2
-} __rte_packed;
-/* This is the GRC offset where the hcomm_status struct resides. */
-#define HCOMM_STATUS_STRUCT_LOC		0x31001F0UL
-
 #endif /* _HSI_STRUCT_DEF_DPDK_H_ */
diff --git a/drivers/net/bnxt/tf_core/hwrm_tf.h b/drivers/net/bnxt/tf_core/hwrm_tf.h
index 3419095..439950e 100644
--- a/drivers/net/bnxt/tf_core/hwrm_tf.h
+++ b/drivers/net/bnxt/tf_core/hwrm_tf.h
@@ -86,6 +86,7 @@ struct tf_tbl_type_get_output;
 struct tf_em_internal_insert_input;
 struct tf_em_internal_insert_output;
 struct tf_em_internal_delete_input;
+struct tf_em_internal_delete_output;
 /* Input params for session attach */
 typedef struct tf_session_attach_input {
 	/* Firmware session id returned when HWRM_TF_SESSION_OPEN is sent */
@@ -949,6 +950,8 @@ typedef struct tf_em_internal_insert_output {
 	uint16_t			 rptr_index;
 	/* EM record offset 0~3 */
 	uint8_t			  rptr_entry;
+	/* Number of word entries consumed by the key */
+	uint8_t			  num_of_entries;
 } tf_em_internal_insert_output_t, *ptf_em_internal_insert_output_t;
 
 /* Input params for EM INTERNAL rule delete */
@@ -969,4 +972,10 @@ typedef struct tf_em_internal_delete_input {
 	uint16_t			 em_key_bitlen;
 } tf_em_internal_delete_input_t, *ptf_em_internal_delete_input_t;
 
+/* Input params for EM INTERNAL rule delete */
+typedef struct tf_em_internal_delete_output {
+	/* Original stack allocation index */
+	uint16_t			 em_index;
+} tf_em_internal_delete_output_t, *ptf_em_internal_delete_output_t;
+
 #endif /* _HWRM_TF_H_ */
diff --git a/drivers/net/bnxt/tf_core/lookup3.h b/drivers/net/bnxt/tf_core/lookup3.h
index e5abcc2..b1fd2cd 100644
--- a/drivers/net/bnxt/tf_core/lookup3.h
+++ b/drivers/net/bnxt/tf_core/lookup3.h
@@ -152,7 +152,6 @@ static inline uint32_t hashword(const uint32_t *k,
 		final(a, b, c);
 		/* Falls through. */
 	case 0:	    /* case 0: nothing left to add */
-		/* FALLTHROUGH */
 		break;
 	}
 	/*------------------------------------------------- report the result */
diff --git a/drivers/net/bnxt/tf_core/stack.c b/drivers/net/bnxt/tf_core/stack.c
index 9cfbd24..9548063 100644
--- a/drivers/net/bnxt/tf_core/stack.c
+++ b/drivers/net/bnxt/tf_core/stack.c
@@ -27,6 +27,14 @@ stack_init(int num_entries, uint32_t *items, struct stack *st)
 	return 0;
 }
 
+/*
+ * Return the address of the items
+ */
+uint32_t *stack_items(struct stack *st)
+{
+	return st->items;
+}
+
 /* Return the size of the stack
  */
 int32_t
diff --git a/drivers/net/bnxt/tf_core/stack.h b/drivers/net/bnxt/tf_core/stack.h
index ebd0555..6732e03 100644
--- a/drivers/net/bnxt/tf_core/stack.h
+++ b/drivers/net/bnxt/tf_core/stack.h
@@ -36,6 +36,16 @@ int stack_init(int num_entries,
 	       uint32_t *items,
 	       struct stack *st);
 
+/** Return the address of the stack contents
+ *
+ *  [in] st
+ *    pointer to the stack
+ *
+ *  return
+ *    pointer to the stack contents
+ */
+uint32_t *stack_items(struct stack *st);
+
 /** Return the size of the stack
  *
  *  [in] st
diff --git a/drivers/net/bnxt/tf_core/tf_core.c b/drivers/net/bnxt/tf_core/tf_core.c
index cf9f36a..ba54df6 100644
--- a/drivers/net/bnxt/tf_core/tf_core.c
+++ b/drivers/net/bnxt/tf_core/tf_core.c
@@ -45,6 +45,100 @@ static void tf_seeds_init(struct tf_session *session)
 	}
 }
 
+/**
+ * Create EM Tbl pool of memory indexes.
+ *
+ * [in] session
+ *   Pointer to session
+ * [in] dir
+ *   direction
+ * [in] num_entries
+ *   number of entries to write
+ *
+ * Return:
+ *  0       - Success, entry allocated - no search support
+ *  -ENOMEM -EINVAL -EOPNOTSUPP
+ *          - Failure, entry not allocated, out of resources
+ */
+static int
+tf_create_em_pool(struct tf_session *session,
+		  enum tf_dir dir,
+		  uint32_t num_entries)
+{
+	struct tfp_calloc_parms parms;
+	uint32_t i, j;
+	int rc = 0;
+	struct stack *pool = &session->em_pool[dir];
+
+	parms.nitems = num_entries;
+	parms.size = sizeof(uint32_t);
+	parms.alignment = 0;
+
+	if (tfp_calloc(&parms) != 0) {
+		TFP_DRV_LOG(ERR, "EM pool allocation failure %s\n",
+			    strerror(-ENOMEM));
+		return -ENOMEM;
+	}
+
+	/* Create empty stack
+	 */
+	rc = stack_init(num_entries, parms.mem_va, pool);
+
+	if (rc != 0) {
+		TFP_DRV_LOG(ERR, "EM pool stack init failure %s\n",
+			    strerror(-rc));
+		goto cleanup;
+	}
+
+	/* Fill pool with indexes
+	 */
+	j = num_entries - 1;
+
+	for (i = 0; i < num_entries; i++) {
+		rc = stack_push(pool, j);
+		if (rc != 0) {
+			TFP_DRV_LOG(ERR, "EM pool stack push failure %s\n",
+				    strerror(-rc));
+			goto cleanup;
+		}
+		j--;
+	}
+
+	if (!stack_is_full(pool)) {
+		rc = -EINVAL;
+		TFP_DRV_LOG(ERR, "EM pool stack failure %s\n",
+			    strerror(-rc));
+		goto cleanup;
+	}
+
+	return 0;
+cleanup:
+	tfp_free((void *)parms.mem_va);
+	return rc;
+}
+
+/**
+ * Create EM Tbl pool of memory indexes.
+ *
+ * [in] session
+ *   Pointer to session
+ * [in] dir
+ *   direction
+ *
+ * Return:
+ */
+static void
+tf_free_em_pool(struct tf_session *session,
+		enum tf_dir dir)
+{
+	struct stack *pool = &session->em_pool[dir];
+	uint32_t *ptr;
+
+	ptr = stack_items(pool);
+
+	tfp_free(ptr);
+}
+
 int
 tf_open_session(struct tf                    *tfp,
 		struct tf_open_session_parms *parms)
@@ -54,6 +148,7 @@ tf_open_session(struct tf                    *tfp,
 	struct tfp_calloc_parms alloc_parms;
 	unsigned int domain, bus, slot, device;
 	uint8_t fw_session_id;
+	int dir;
 
 	if (tfp == NULL || parms == NULL)
 		return -EINVAL;
@@ -110,7 +205,7 @@ tf_open_session(struct tf                    *tfp,
 		goto cleanup;
 	}
 
-	tfp->session = (struct tf_session_info *)alloc_parms.mem_va;
+	tfp->session = alloc_parms.mem_va;
 
 	/* Allocate core data for the session */
 	alloc_parms.nitems = 1;
@@ -175,6 +270,16 @@ tf_open_session(struct tf                    *tfp,
 	/* Setup hash seeds */
 	tf_seeds_init(session);
 
+	/* Initialize EM pool */
+	for (dir = 0; dir < TF_DIR_MAX; dir++) {
+		rc = tf_create_em_pool(session, dir, TF_SESSION_EM_POOL_SIZE);
+		if (rc) {
+			TFP_DRV_LOG(ERR,
+				    "EM Pool initialization failed\n");
+			goto cleanup_close;
+		}
+	}
+
 	session->ref_count++;
 
 	/* Return session ID */
@@ -239,6 +344,7 @@ tf_close_session(struct tf *tfp)
 	int rc_close = 0;
 	struct tf_session *tfs;
 	union tf_session_id session_id;
+	int dir;
 
 	if (tfp == NULL || tfp->session == NULL)
 		return -EINVAL;
@@ -268,6 +374,10 @@ tf_close_session(struct tf *tfp)
 
 	/* Final cleanup as we're last user of the session */
 	if (tfs->ref_count == 0) {
+		/* Free EM pool */
+		for (dir = 0; dir < TF_DIR_MAX; dir++)
+			tf_free_em_pool(tfs, dir);
+
 		tfp_free(tfp->session->core_data);
 		tfp_free(tfp->session);
 		tfp->session = NULL;
@@ -301,16 +411,25 @@ int tf_insert_em_entry(struct tf *tfp,
 	if (tfp == NULL || parms == NULL)
 		return -EINVAL;
 
-	tbl_scope_cb =
-		tbl_scope_cb_find((struct tf_session *)tfp->session->core_data,
-				  parms->tbl_scope_id);
+	tbl_scope_cb = tbl_scope_cb_find(
+		(struct tf_session *)(tfp->session->core_data),
+		parms->tbl_scope_id);
 	if (tbl_scope_cb == NULL)
 		return -EINVAL;
 
 	/* Process the EM entry per Table Scope type */
-	return tf_insert_eem_entry((struct tf_session *)tfp->session->core_data,
-				   tbl_scope_cb,
-				   parms);
+	if (parms->mem == TF_MEM_EXTERNAL) {
+		/* External EEM */
+		return tf_insert_eem_entry
+			((struct tf_session *)(tfp->session->core_data),
+			tbl_scope_cb,
+			parms);
+	} else if (parms->mem == TF_MEM_INTERNAL) {
+		/* Internal EM */
+		return tf_insert_em_internal_entry(tfp,	parms);
+	}
+
+	return -EINVAL;
 }
 
 /** Delete EM hash entry API
@@ -327,13 +446,16 @@ int tf_delete_em_entry(struct tf *tfp,
 	if (tfp == NULL || parms == NULL)
 		return -EINVAL;
 
-	tbl_scope_cb =
-		tbl_scope_cb_find((struct tf_session *)tfp->session->core_data,
-				  parms->tbl_scope_id);
+	tbl_scope_cb = tbl_scope_cb_find(
+		(struct tf_session *)(tfp->session->core_data),
+		parms->tbl_scope_id);
 	if (tbl_scope_cb == NULL)
 		return -EINVAL;
 
-	return tf_delete_eem_entry(tfp, parms);
+	if (parms->mem == TF_MEM_EXTERNAL)
+		return tf_delete_eem_entry(tfp, parms);
+	else
+		return tf_delete_em_internal_entry(tfp, parms);
 }
 
 /** allocate identifier resource
diff --git a/drivers/net/bnxt/tf_core/tf_core.h b/drivers/net/bnxt/tf_core/tf_core.h
index 1eedd80..81ff760 100644
--- a/drivers/net/bnxt/tf_core/tf_core.h
+++ b/drivers/net/bnxt/tf_core/tf_core.h
@@ -44,44 +44,7 @@ enum tf_mem {
 };
 
 /**
- * The size of the external action record (Wh+/Brd2)
- *
- * Currently set to 512.
- *
- * AR (16B) + encap (256B) + stats_ptrs (8) + resvd (8)
- * + stats (16) = 304 aligned on a 16B boundary
- *
- * Theoretically, the size should be smaller. ~304B
- */
-#define TF_ACTION_RECORD_SZ 512
-
-/**
- * External pool size
- *
- * Defines a single pool of external action records of
- * fixed size.  Currently, this is an index.
- */
-#define TF_EXT_POOL_ENTRY_SZ_BYTES 1
-
-/**
- *  External pool entry count
- *
- *  Defines the number of entries in the external action pool
- */
-#define TF_EXT_POOL_ENTRY_CNT (1 * 1024)
-
-/**
- * Number of external pools
- */
-#define TF_EXT_POOL_CNT_MAX 1
-
-/**
- * External pool Id
- */
-#define TF_EXT_POOL_0      0 /**< matches TF_TBL_TYPE_EXT   */
-#define TF_EXT_POOL_1      1 /**< matches TF_TBL_TYPE_EXT_0 */
-
-/** EEM record AR helper
+ * EEM record AR helper
  *
  * Helper to handle the Action Record Pointer in the EEM Record Entry.
  *
@@ -109,7 +72,8 @@ enum tf_mem {
  */
 
 
-/** Session Version defines
+/**
+ * Session Version defines
  *
  * The version controls the format of the tf_session and
  * tf_session_info structure. This is to assure upgrade between
@@ -119,7 +83,8 @@ enum tf_mem {
 #define TF_SESSION_VER_MINOR  0   /**< Minor Version */
 #define TF_SESSION_VER_UPDATE 0   /**< Update Version */
 
-/** Session Name
+/**
+ * Session Name
  *
  * Name of the TruFlow control channel interface.  Expects
  * format to be RTE Name specific, i.e. rte_eth_dev_get_name_by_port()
@@ -128,7 +93,8 @@ enum tf_mem {
 
 #define TF_FW_SESSION_ID_INVALID  0xFF  /**< Invalid FW Session ID define */
 
-/** Session Identifier
+/**
+ * Session Identifier
  *
  * Unique session identifier which includes PCIe bus info to
  * distinguish the PF and session info to identify the associated
@@ -146,7 +112,8 @@ union tf_session_id {
 	} internal;
 };
 
-/** Session Version
+/**
+ * Session Version
  *
  * The version controls the format of the tf_session and
  * tf_session_info structure. This is to assure upgrade between
@@ -160,8 +127,8 @@ struct tf_session_version {
 	uint8_t update;
 };
 
-/** Session supported device types
- *
+/**
+ * Session supported device types
  */
 enum tf_device_type {
 	TF_DEVICE_TYPE_WH = 0, /**< Whitney+  */
@@ -171,6 +138,147 @@ enum tf_device_type {
 	TF_DEVICE_TYPE_MAX     /**< Maximum   */
 };
 
+/** Identifier resource types
+ */
+enum tf_identifier_type {
+	/** The L2 Context is returned from the L2 Ctxt TCAM lookup
+	 *  and can be used in WC TCAM or EM keys to virtualize further
+	 *  lookups.
+	 */
+	TF_IDENT_TYPE_L2_CTXT,
+	/** The WC profile func is returned from the L2 Ctxt TCAM lookup
+	 *  to enable virtualization of the profile TCAM.
+	 */
+	TF_IDENT_TYPE_PROF_FUNC,
+	/** The WC profile ID is included in the WC lookup key
+	 *  to enable virtualization of the WC TCAM hardware.
+	 */
+	TF_IDENT_TYPE_WC_PROF,
+	/** The EM profile ID is included in the EM lookup key
+	 *  to enable virtualization of the EM hardware. (not required for SR2
+	 *  as it has table scope)
+	 */
+	TF_IDENT_TYPE_EM_PROF,
+	/** The L2 func is included in the ILT result and from recycling to
+	 *  enable virtualization of further lookups.
+	 */
+	TF_IDENT_TYPE_L2_FUNC,
+	TF_IDENT_TYPE_MAX
+};
+
+/**
+ * Enumeration of TruFlow table types. A table type is used to identify a
+ * resource object.
+ *
+ * NOTE: The table type TF_TBL_TYPE_EXT is unique in that it is
+ * the only table type that is connected with a table scope.
+ */
+enum tf_tbl_type {
+	/* Internal */
+
+	/** Wh+/SR Action Record */
+	TF_TBL_TYPE_FULL_ACT_RECORD,
+	/** Wh+/SR/Th Multicast Groups */
+	TF_TBL_TYPE_MCAST_GROUPS,
+	/** Wh+/SR Action Encap 8 Bytes */
+	TF_TBL_TYPE_ACT_ENCAP_8B,
+	/** Wh+/SR Action Encap 16 Bytes */
+	TF_TBL_TYPE_ACT_ENCAP_16B,
+	/** Action Encap 32 Bytes */
+	TF_TBL_TYPE_ACT_ENCAP_32B,
+	/** Wh+/SR Action Encap 64 Bytes */
+	TF_TBL_TYPE_ACT_ENCAP_64B,
+	/** Action Source Properties SMAC */
+	TF_TBL_TYPE_ACT_SP_SMAC,
+	/** Wh+/SR Action Source Properties SMAC IPv4 */
+	TF_TBL_TYPE_ACT_SP_SMAC_IPV4,
+	/** Action Source Properties SMAC IPv6 */
+	TF_TBL_TYPE_ACT_SP_SMAC_IPV6,
+	/** Wh+/SR Action Statistics 64 Bits */
+	TF_TBL_TYPE_ACT_STATS_64,
+	/** Wh+/SR Action Modify L4 Src Port */
+	TF_TBL_TYPE_ACT_MODIFY_SPORT,
+	/** Wh+/SR Action Modify L4 Dest Port */
+	TF_TBL_TYPE_ACT_MODIFY_DPORT,
+	/** Wh+/SR Action Modify IPv4 Source */
+	TF_TBL_TYPE_ACT_MODIFY_IPV4_SRC,
+	/** Wh+/SR Action _Modify L4 Dest Port */
+	TF_TBL_TYPE_ACT_MODIFY_IPV4_DEST,
+	/** Action Modify IPv6 Source */
+	TF_TBL_TYPE_ACT_MODIFY_IPV6_SRC,
+	/** Action Modify IPv6 Destination */
+	TF_TBL_TYPE_ACT_MODIFY_IPV6_DEST,
+	/** Meter Profiles */
+	TF_TBL_TYPE_METER_PROF,
+	/** Meter Instance */
+	TF_TBL_TYPE_METER_INST,
+	/** Mirror Config */
+	TF_TBL_TYPE_MIRROR_CONFIG,
+	/** UPAR */
+	TF_TBL_TYPE_UPAR,
+	/** SR2 Epoch 0 table */
+	TF_TBL_TYPE_EPOCH0,
+	/** SR2 Epoch 1 table  */
+	TF_TBL_TYPE_EPOCH1,
+	/** SR2 Metadata  */
+	TF_TBL_TYPE_METADATA,
+	/** SR2 CT State  */
+	TF_TBL_TYPE_CT_STATE,
+	/** SR2 Range Profile  */
+	TF_TBL_TYPE_RANGE_PROF,
+	/** SR2 Range Entry  */
+	TF_TBL_TYPE_RANGE_ENTRY,
+	/** SR2 LAG Entry  */
+	TF_TBL_TYPE_LAG,
+	/** SR2 VNIC/SVIF Table */
+	TF_TBL_TYPE_VNIC_SVIF,
+	/** Th/SR2 EM Flexible Key builder */
+	TF_TBL_TYPE_EM_FKB,
+	/** Th/SR2 WC Flexible Key builder */
+	TF_TBL_TYPE_WC_FKB,
+
+	/* External */
+
+	/** External table type - initially 1 poolsize entries.
+	 * All External table types are associated with a table
+	 * scope. Internal types are not.
+	 */
+	TF_TBL_TYPE_EXT,
+	TF_TBL_TYPE_MAX
+};
+
+/**
+ * TCAM table type
+ */
+enum tf_tcam_tbl_type {
+	/** L2 Context TCAM */
+	TF_TCAM_TBL_TYPE_L2_CTXT_TCAM,
+	/** Profile TCAM */
+	TF_TCAM_TBL_TYPE_PROF_TCAM,
+	/** Wildcard TCAM */
+	TF_TCAM_TBL_TYPE_WC_TCAM,
+	/** Source Properties TCAM */
+	TF_TCAM_TBL_TYPE_SP_TCAM,
+	/** Connection Tracking Rule TCAM */
+	TF_TCAM_TBL_TYPE_CT_RULE_TCAM,
+	/** Virtual Edge Bridge TCAM */
+	TF_TCAM_TBL_TYPE_VEB_TCAM,
+	TF_TCAM_TBL_TYPE_MAX
+};
+
+/**
+ * EM Resources
+ * These defines are provisioned during
+ * tf_open_session()
+ */
+enum tf_em_tbl_type {
+	/** The number of internal EM records for the session */
+	TF_EM_TBL_TYPE_EM_RECORD,
+	/** The number of table scopes reequested */
+	TF_EM_TBL_TYPE_TBL_SCOPE,
+	TF_EM_TBL_TYPE_MAX
+};
+
 /** TruFlow Session Information
  *
  * Structure defining a TruFlow Session, also known as a Management
@@ -309,6 +417,30 @@ struct tf_open_session_parms {
 	 * Device type is passed, one of Wh+, Brd2, Brd3, Brd4
 	 */
 	enum tf_device_type device_type;
+	/** [in] Requested Identifier Resources
+	 *
+	 * The number of identifier resources requested for the session.
+	 * The index used is tf_identifier_type.
+	 */
+	uint16_t identifer_cnt[TF_IDENT_TYPE_MAX];
+	/** [in] Requested Index Table resource counts
+	 *
+	 * The number of index table resources requested for the session.
+	 * The index used is tf_tbl_type.
+	 */
+	uint16_t tbl_cnt[TF_TBL_TYPE_MAX];
+	/** [in] Requested TCAM Table resource counts
+	 *
+	 * The number of TCAM table resources requested for the session.
+	 * The index used is tf_tcam_tbl_type.
+	 */
+	uint16_t tcam_tbl_cnt[TF_TCAM_TBL_TYPE_MAX];
+	/** [in] Requested EM resource counts
+	 *
+	 * The number of internal EM table resources requested for the session
+	 * The index used is tf_em_tbl_type.
+	 */
+	uint16_t em_tbl_cnt[TF_EM_TBL_TYPE_MAX];
 };
 
 /**
@@ -417,31 +549,6 @@ int tf_close_session(struct tf *tfp);
  *
  * @ref tf_free_identifier
  */
-enum tf_identifier_type {
-	/** The L2 Context is returned from the L2 Ctxt TCAM lookup
-	 *  and can be used in WC TCAM or EM keys to virtualize further
-	 *  lookups.
-	 */
-	TF_IDENT_TYPE_L2_CTXT,
-	/** The WC profile func is returned from the L2 Ctxt TCAM lookup
-	 *  to enable virtualization of the profile TCAM.
-	 */
-	TF_IDENT_TYPE_PROF_FUNC,
-	/** The WC profile ID is included in the WC lookup key
-	 *  to enable virtualization of the WC TCAM hardware.
-	 */
-	TF_IDENT_TYPE_WC_PROF,
-	/** The EM profile ID is included in the EM lookup key
-	 *  to enable virtualization of the EM hardware. (not required for Brd4
-	 *  as it has table scope)
-	 */
-	TF_IDENT_TYPE_EM_PROF,
-	/** The L2 func is included in the ILT result and from recycling to
-	 *  enable virtualization of further lookups.
-	 */
-	TF_IDENT_TYPE_L2_FUNC
-};
-
 /** tf_alloc_identifier parameter definition
  */
 struct tf_alloc_identifier_parms {
@@ -631,19 +738,6 @@ int tf_alloc_tbl_scope(struct tf *tfp,
 int tf_free_tbl_scope(struct tf *tfp,
 		      struct tf_free_tbl_scope_parms *parms);
 
-/**
- * TCAM table type
- */
-enum tf_tcam_tbl_type {
-	TF_TCAM_TBL_TYPE_L2_CTXT_TCAM,
-	TF_TCAM_TBL_TYPE_PROF_TCAM,
-	TF_TCAM_TBL_TYPE_WC_TCAM,
-	TF_TCAM_TBL_TYPE_SP_TCAM,
-	TF_TCAM_TBL_TYPE_CT_RULE_TCAM,
-	TF_TCAM_TBL_TYPE_VEB_TCAM,
-	TF_TCAM_TBL_TYPE_MAX
-
-};
 
 /**
  * @page tcam TCAM Access
@@ -813,7 +907,8 @@ struct tf_get_tcam_entry_parms {
 	uint16_t result_sz_in_bits;
 };
 
-/** get TCAM entry
+/*
+ * get TCAM entry
  *
  * Program a TCAM table entry for a TruFlow session.
  *
@@ -824,7 +919,8 @@ struct tf_get_tcam_entry_parms {
 int tf_get_tcam_entry(struct tf *tfp,
 		      struct tf_get_tcam_entry_parms *parms);
 
-/** tf_free_tcam_entry parameter definition
+/*
+ * tf_free_tcam_entry parameter definition
  */
 struct tf_free_tcam_entry_parms {
 	/**
@@ -845,8 +941,7 @@ struct tf_free_tcam_entry_parms {
 	uint16_t ref_cnt;
 };
 
-/** free TCAM entry
- *
+/*
  * Free TCAM entry.
  *
  * Firmware checks to ensure the TCAM entries are owned by the TruFlow
@@ -873,84 +968,7 @@ int tf_free_tcam_entry(struct tf *tfp,
  */
 
 /**
- * Enumeration of TruFlow table types. A table type is used to identify a
- * resource object.
- *
- * NOTE: The table type TF_TBL_TYPE_EXT is unique in that it is
- * the only table type that is connected with a table scope.
- */
-enum tf_tbl_type {
-	/** Wh+/Brd2 Action Record */
-	TF_TBL_TYPE_FULL_ACT_RECORD,
-	/** Multicast Groups */
-	TF_TBL_TYPE_MCAST_GROUPS,
-	/** Action Encap 8 Bytes */
-	TF_TBL_TYPE_ACT_ENCAP_8B,
-	/** Action Encap 16 Bytes */
-	TF_TBL_TYPE_ACT_ENCAP_16B,
-	/** Action Encap 64 Bytes */
-	TF_TBL_TYPE_ACT_ENCAP_32B,
-	/** Action Encap 64 Bytes */
-	TF_TBL_TYPE_ACT_ENCAP_64B,
-	/** Action Source Properties SMAC */
-	TF_TBL_TYPE_ACT_SP_SMAC,
-	/** Action Source Properties SMAC IPv4 */
-	TF_TBL_TYPE_ACT_SP_SMAC_IPV4,
-	/** Action Source Properties SMAC IPv6 */
-	TF_TBL_TYPE_ACT_SP_SMAC_IPV6,
-	/** Action Statistics 64 Bits */
-	TF_TBL_TYPE_ACT_STATS_64,
-	/** Action Modify L4 Src Port */
-	TF_TBL_TYPE_ACT_MODIFY_SPORT,
-	/** Action Modify L4 Dest Port */
-	TF_TBL_TYPE_ACT_MODIFY_DPORT,
-	/** Action Modify IPv4 Source */
-	TF_TBL_TYPE_ACT_MODIFY_IPV4_SRC,
-	/** Action _Modify L4 Dest Port */
-	TF_TBL_TYPE_ACT_MODIFY_IPV4_DEST,
-	/** Action Modify IPv6 Source */
-	TF_TBL_TYPE_ACT_MODIFY_IPV6_SRC,
-	/** Action Modify IPv6 Destination */
-	TF_TBL_TYPE_ACT_MODIFY_IPV6_DEST,
-
-	/* HW */
-
-	/** Meter Profiles */
-	TF_TBL_TYPE_METER_PROF,
-	/** Meter Instance */
-	TF_TBL_TYPE_METER_INST,
-	/** Mirror Config */
-	TF_TBL_TYPE_MIRROR_CONFIG,
-	/** UPAR */
-	TF_TBL_TYPE_UPAR,
-	/** Brd4 Epoch 0 table */
-	TF_TBL_TYPE_EPOCH0,
-	/** Brd4 Epoch 1 table  */
-	TF_TBL_TYPE_EPOCH1,
-	/** Brd4 Metadata  */
-	TF_TBL_TYPE_METADATA,
-	/** Brd4 CT State  */
-	TF_TBL_TYPE_CT_STATE,
-	/** Brd4 Range Profile  */
-	TF_TBL_TYPE_RANGE_PROF,
-	/** Brd4 Range Entry  */
-	TF_TBL_TYPE_RANGE_ENTRY,
-	/** Brd4 LAG Entry  */
-	TF_TBL_TYPE_LAG,
-	/** Brd4 only VNIC/SVIF Table */
-	TF_TBL_TYPE_VNIC_SVIF,
-
-	/* External */
-
-	/** External table type - initially 1 poolsize entries.
-	 * All External table types are associated with a table
-	 * scope. Internal types are not.
-	 */
-	TF_TBL_TYPE_EXT,
-	TF_TBL_TYPE_MAX
-};
-
-/** tf_alloc_tbl_entry parameter definition
+ * tf_alloc_tbl_entry parameter definition
  */
 struct tf_alloc_tbl_entry_parms {
 	/**
@@ -993,7 +1011,8 @@ struct tf_alloc_tbl_entry_parms {
 	uint32_t idx;
 };
 
-/** allocate index table entries
+/**
+ * allocate index table entries
  *
  * Internal types:
  *
@@ -1023,7 +1042,8 @@ struct tf_alloc_tbl_entry_parms {
 int tf_alloc_tbl_entry(struct tf *tfp,
 		       struct tf_alloc_tbl_entry_parms *parms);
 
-/** tf_free_tbl_entry parameter definition
+/**
+ * tf_free_tbl_entry parameter definition
  */
 struct tf_free_tbl_entry_parms {
 	/**
@@ -1049,7 +1069,8 @@ struct tf_free_tbl_entry_parms {
 	uint16_t ref_cnt;
 };
 
-/** free index table entry
+/**
+ * free index table entry
  *
  * Used to free a previously allocated table entry.
  *
@@ -1075,7 +1096,8 @@ struct tf_free_tbl_entry_parms {
 int tf_free_tbl_entry(struct tf *tfp,
 		      struct tf_free_tbl_entry_parms *parms);
 
-/** tf_set_tbl_entry parameter definition
+/**
+ * tf_set_tbl_entry parameter definition
  */
 struct tf_set_tbl_entry_parms {
 	/**
@@ -1104,7 +1126,8 @@ struct tf_set_tbl_entry_parms {
 	uint32_t idx;
 };
 
-/** set index table entry
+/**
+ * set index table entry
  *
  * Used to insert an application programmed index table entry into a
  * previous allocated table location.  A shadow copy of the table
@@ -1115,7 +1138,8 @@ struct tf_set_tbl_entry_parms {
 int tf_set_tbl_entry(struct tf *tfp,
 		     struct tf_set_tbl_entry_parms *parms);
 
-/** tf_get_tbl_entry parameter definition
+/**
+ * tf_get_tbl_entry parameter definition
  */
 struct tf_get_tbl_entry_parms {
 	/**
@@ -1140,7 +1164,8 @@ struct tf_get_tbl_entry_parms {
 	uint32_t idx;
 };
 
-/** get index table entry
+/**
+ * get index table entry
  *
  * Used to retrieve a previous set index table entry.
  *
@@ -1163,7 +1188,8 @@ int tf_get_tbl_entry(struct tf *tfp,
  * @ref tf_search_em_entry
  *
  */
-/** tf_insert_em_entry parameter definition
+/**
+ * tf_insert_em_entry parameter definition
  */
 struct tf_insert_em_entry_parms {
 	/**
@@ -1240,6 +1266,10 @@ struct tf_delete_em_entry_parms {
 	 */
 	uint16_t *epochs;
 	/**
+	 * [out] The index of the entry
+	 */
+	uint16_t index;
+	/**
 	 * [in] structure containing flow delete handle information
 	 */
 	uint64_t flow_handle;
@@ -1291,7 +1321,8 @@ struct tf_search_em_entry_parms {
 	uint64_t flow_handle;
 };
 
-/** insert em hash entry in internal table memory
+/**
+ * insert em hash entry in internal table memory
  *
  * Internal:
  *
@@ -1328,7 +1359,8 @@ struct tf_search_em_entry_parms {
 int tf_insert_em_entry(struct tf *tfp,
 		       struct tf_insert_em_entry_parms *parms);
 
-/** delete em hash entry table memory
+/**
+ * delete em hash entry table memory
  *
  * Internal:
  *
@@ -1353,7 +1385,8 @@ int tf_insert_em_entry(struct tf *tfp,
 int tf_delete_em_entry(struct tf *tfp,
 		       struct tf_delete_em_entry_parms *parms);
 
-/** search em hash entry table memory
+/**
+ * search em hash entry table memory
  *
  * Internal:
 
diff --git a/drivers/net/bnxt/tf_core/tf_em.c b/drivers/net/bnxt/tf_core/tf_em.c
index bd8e2ba..fd1797e 100644
--- a/drivers/net/bnxt/tf_core/tf_em.c
+++ b/drivers/net/bnxt/tf_core/tf_em.c
@@ -287,7 +287,7 @@ static int tf_em_entry_exists(struct tf_tbl_scope_cb *tbl_scope_cb,
 }
 
 static void tf_em_create_key_entry(struct tf_eem_entry_hdr *result,
-				    uint8_t	       *in_key,
+				    uint8_t *in_key,
 				    struct tf_eem_64b_entry *key_entry)
 {
 	key_entry->hdr.word1 = result->word1;
@@ -308,7 +308,7 @@ static void tf_em_create_key_entry(struct tf_eem_entry_hdr *result,
  * EEXIST  - Key does exist in table at "index" in table "table".
  * TF_ERR     - Something went horribly wrong.
  */
-static int tf_em_select_inject_table(struct tf_tbl_scope_cb	*tbl_scope_cb,
+static int tf_em_select_inject_table(struct tf_tbl_scope_cb *tbl_scope_cb,
 					  enum tf_dir dir,
 					  struct tf_eem_64b_entry *entry,
 					  uint32_t key0_hash,
@@ -368,8 +368,8 @@ static int tf_em_select_inject_table(struct tf_tbl_scope_cb	*tbl_scope_cb,
  *   0
  *   TF_ERR_EM_DUP  - key is already in table
  */
-int tf_insert_eem_entry(struct tf_session	   *session,
-			struct tf_tbl_scope_cb	   *tbl_scope_cb,
+int tf_insert_eem_entry(struct tf_session *session,
+			struct tf_tbl_scope_cb *tbl_scope_cb,
 			struct tf_insert_em_entry_parms *parms)
 {
 	uint32_t	   mask;
@@ -457,6 +457,96 @@ int tf_insert_eem_entry(struct tf_session	   *session,
 	return -EINVAL;
 }
 
+/**
+ * Insert EM internal entry API
+ *
+ *  returns:
+ *     0 - Success
+ */
+int tf_insert_em_internal_entry(struct tf *tfp,
+				struct tf_insert_em_entry_parms *parms)
+{
+	int       rc;
+	uint32_t  gfid;
+	uint16_t  rptr_index = 0;
+	uint8_t   rptr_entry = 0;
+	uint8_t   num_of_entries = 0;
+	struct tf_session *session =
+		(struct tf_session *)(tfp->session->core_data);
+	struct stack *pool = &session->em_pool[parms->dir];
+	uint32_t index;
+
+	rc = stack_pop(pool, &index);
+
+	if (rc != 0) {
+		PMD_DRV_LOG
+		   (ERR,
+		   "dir:%d, EM entry index allocation failed\n",
+		   parms->dir);
+		return rc;
+	}
+
+	rptr_index = index * TF_SESSION_EM_ENTRY_SIZE;
+	rc = tf_msg_insert_em_internal_entry(tfp,
+					     parms,
+					     &rptr_index,
+					     &rptr_entry,
+					     &num_of_entries);
+	if (rc != 0)
+		return -1;
+
+	PMD_DRV_LOG
+		   (ERR,
+		   "Internal entry @ Index:%d rptr_index:0x%x rptr_entry:0x%x num_of_entries:%d\n",
+		   index * TF_SESSION_EM_ENTRY_SIZE,
+		   rptr_index,
+		   rptr_entry,
+		   num_of_entries);
+
+	TF_SET_GFID(gfid,
+		    ((rptr_index << TF_EM_INTERNAL_INDEX_SHIFT) |
+		     rptr_entry),
+		    0); /* N/A for internal table */
+
+	TF_SET_FLOW_ID(parms->flow_id,
+		       gfid,
+		       TF_GFID_TABLE_INTERNAL,
+		       parms->dir);
+
+	TF_SET_FIELDS_IN_FLOW_HANDLE(parms->flow_handle,
+				     num_of_entries,
+				     0,
+				     0,
+				     rptr_index,
+				     rptr_entry,
+				     0);
+	return 0;
+}
+
+/** Delete EM internal entry API
+ *
+ * returns:
+ * 0
+ * -EINVAL
+ */
+int tf_delete_em_internal_entry(struct tf *tfp,
+				struct tf_delete_em_entry_parms *parms)
+{
+	int rc;
+	struct tf_session *session =
+		(struct tf_session *)(tfp->session->core_data);
+	struct stack *pool = &session->em_pool[parms->dir];
+
+	rc = tf_msg_delete_em_entry(tfp, parms);
+
+	/* Return resource to pool */
+	if (rc == 0)
+		stack_push(pool, parms->index / TF_SESSION_EM_ENTRY_SIZE);
+
+	return rc;
+}
+
+
 /** delete EEM hash entry API
  *
  * returns:
diff --git a/drivers/net/bnxt/tf_core/tf_em.h b/drivers/net/bnxt/tf_core/tf_em.h
index 8a3584f..c1805df 100644
--- a/drivers/net/bnxt/tf_core/tf_em.h
+++ b/drivers/net/bnxt/tf_core/tf_em.h
@@ -12,6 +12,20 @@
 #define TF_HW_EM_KEY_MAX_SIZE 52
 #define TF_EM_KEY_RECORD_SIZE 64
 
+/*
+ * Used to build GFID:
+ *
+ *   15           2  0
+ *  +--------------+--+
+ *  |   Index      |E |
+ *  +--------------+--+
+ *
+ * E = Entry (bucket inndex)
+ */
+#define TF_EM_INTERNAL_INDEX_SHIFT 2
+#define TF_EM_INTERNAL_INDEX_MASK 0xFFFC
+#define TF_EM_INTERNAL_ENTRY_MASK  0x3
+
 /** EEM Entry header
  *
  */
@@ -53,6 +67,17 @@ struct tf_eem_64b_entry {
 	struct tf_eem_entry_hdr hdr;
 };
 
+/** EM Entry
+ *  Each EM entry is 512-bit (64-bytes) but ordered differently to
+ *  EEM.
+ */
+struct tf_em_64b_entry {
+	/** Header is 8 bytes long */
+	struct tf_eem_entry_hdr hdr;
+	/** Key is 448 bits - 56 bytes */
+	uint8_t key[TF_EM_KEY_RECORD_SIZE - sizeof(struct tf_eem_entry_hdr)];
+};
+
 /**
  * Allocates EEM Table scope
  *
@@ -106,9 +131,15 @@ int tf_insert_eem_entry(struct tf_session *session,
 			struct tf_tbl_scope_cb *tbl_scope_cb,
 			struct tf_insert_em_entry_parms *parms);
 
+int tf_insert_em_internal_entry(struct tf *tfp,
+				struct tf_insert_em_entry_parms *parms);
+
 int tf_delete_eem_entry(struct tf *tfp,
 			struct tf_delete_em_entry_parms *parms);
 
+int tf_delete_em_internal_entry(struct tf                       *tfp,
+				struct tf_delete_em_entry_parms *parms);
+
 void *tf_em_get_table_page(struct tf_tbl_scope_cb *tbl_scope_cb,
 			   enum tf_dir dir,
 			   uint32_t offset,
diff --git a/drivers/net/bnxt/tf_core/tf_ext_flow_handle.h b/drivers/net/bnxt/tf_core/tf_ext_flow_handle.h
index 417a99c..399f7d1 100644
--- a/drivers/net/bnxt/tf_core/tf_ext_flow_handle.h
+++ b/drivers/net/bnxt/tf_core/tf_ext_flow_handle.h
@@ -90,6 +90,22 @@ do {									\
 		     TF_HASH_TYPE_FLOW_HANDLE_SFT);			\
 } while (0)
 
+#define TF_GET_NUM_KEY_ENTRIES_FROM_FLOW_HANDLE(flow_handle,		\
+					  num_key_entries)		\
+do {									\
+	num_key_entries =						\
+		(((flow_handle) & TF_NUM_KEY_ENTRIES_FLOW_HANDLE_MASK) >> \
+		     TF_NUM_KEY_ENTRIES_FLOW_HANDLE_SFT);		\
+} while (0)
+
+#define TF_GET_ENTRY_NUM_FROM_FLOW_HANDLE(flow_handle,		\
+					  entry_num)		\
+do {									\
+	entry_num =						\
+		(((flow_handle) & TF_ENTRY_NUM_FLOW_HANDLE_MASK) >> \
+		     TF_ENTRY_NUM_FLOW_HANDLE_SFT);		\
+} while (0)
+
 /*
  * 32 bit Flow ID handlers
  */
diff --git a/drivers/net/bnxt/tf_core/tf_msg.c b/drivers/net/bnxt/tf_core/tf_msg.c
index beecafd..554a849 100644
--- a/drivers/net/bnxt/tf_core/tf_msg.c
+++ b/drivers/net/bnxt/tf_core/tf_msg.c
@@ -16,6 +16,7 @@
 #include "tf_msg.h"
 #include "hsi_struct_def_dpdk.h"
 #include "hwrm_tf.h"
+#include "tf_em.h"
 
 /**
  * Endian converts min and max values from the HW response to the query
@@ -1014,14 +1015,93 @@ int tf_msg_em_cfg(struct tf *tfp,
 }
 
 /**
+ * Sends EM internal insert request to Firmware
+ */
+int tf_msg_insert_em_internal_entry(struct tf *tfp,
+				struct tf_insert_em_entry_parms *em_parms,
+				uint16_t *rptr_index,
+				uint8_t *rptr_entry,
+				uint8_t *num_of_entries)
+{
+	int rc;
+	struct tfp_send_msg_parms parms = { 0 };
+	struct tf_em_internal_insert_input req = { 0 };
+	struct tf_em_internal_insert_output resp = { 0 };
+	struct tf_session *tfs = (struct tf_session *)(tfp->session->core_data);
+	struct tf_em_64b_entry *em_result =
+		(struct tf_em_64b_entry *)em_parms->em_record;
+
+	req.fw_session_id =
+		tfp_cpu_to_le_32(tfs->session_id.internal.fw_session_id);
+	memcpy(req.em_key, em_parms->key, ((em_parms->key_sz_in_bits + 7) / 8));
+	req.flags = tfp_cpu_to_le_16(em_parms->dir);
+	req.strength = (em_result->hdr.word1 & TF_LKUP_RECORD_STRENGTH_MASK) >>
+		TF_LKUP_RECORD_STRENGTH_SHIFT;
+	req.em_key_bitlen = em_parms->key_sz_in_bits;
+	req.action_ptr = em_result->hdr.pointer;
+	req.em_record_idx = *rptr_index;
+
+	MSG_PREP(parms,
+		 TF_KONG_MB,
+		 HWRM_TF,
+		 HWRM_TFT_EM_RULE_INSERT,
+		 req,
+		 resp);
+
+	rc = tfp_send_msg_tunneled(tfp, &parms);
+	if (rc)
+		return rc;
+
+	*rptr_entry = resp.rptr_entry;
+	*rptr_index = resp.rptr_index;
+	*num_of_entries = resp.num_of_entries;
+
+	return tfp_le_to_cpu_32(parms.tf_resp_code);
+}
+
+/**
+ * Sends EM delete insert request to Firmware
+ */
+int tf_msg_delete_em_entry(struct tf *tfp,
+			   struct tf_delete_em_entry_parms *em_parms)
+{
+	int rc;
+	struct tfp_send_msg_parms parms = { 0 };
+	struct tf_em_internal_delete_input req = { 0 };
+	struct tf_em_internal_delete_output resp = { 0 };
+	struct tf_session *tfs =
+		(struct tf_session *)(tfp->session->core_data);
+
+	req.tf_session_id =
+		tfp_cpu_to_le_32(tfs->session_id.internal.fw_session_id);
+	req.flags = tfp_cpu_to_le_16(em_parms->dir);
+	req.flow_handle = tfp_cpu_to_le_64(em_parms->flow_handle);
+
+	MSG_PREP(parms,
+		 TF_KONG_MB,
+		 HWRM_TF,
+		 HWRM_TFT_EM_RULE_DELETE,
+		 req,
+		resp);
+
+	rc = tfp_send_msg_tunneled(tfp, &parms);
+	if (rc)
+		return rc;
+
+	em_parms->index = tfp_le_to_cpu_16(resp.em_index);
+
+	return tfp_le_to_cpu_32(parms.tf_resp_code);
+}
+
+/**
  * Sends EM operation request to Firmware
  */
 int tf_msg_em_op(struct tf *tfp,
-		 int        dir,
-		 uint16_t   op)
+		 int dir,
+		 uint16_t op)
 {
 	int rc;
-	struct hwrm_tf_ext_em_op_input  req = {0};
+	struct hwrm_tf_ext_em_op_input req = {0};
 	struct hwrm_tf_ext_em_op_output resp = {0};
 	uint32_t flags;
 	struct tfp_send_msg_parms parms = { 0 };
diff --git a/drivers/net/bnxt/tf_core/tf_msg.h b/drivers/net/bnxt/tf_core/tf_msg.h
index 030d188..89f7370 100644
--- a/drivers/net/bnxt/tf_core/tf_msg.h
+++ b/drivers/net/bnxt/tf_core/tf_msg.h
@@ -122,6 +122,19 @@ int tf_msg_session_sram_resc_flush(struct tf *tfp,
 				   struct tf_rm_entry *sram_entry);
 
 /**
+ * Sends EM internal insert request to Firmware
+ */
+int tf_msg_insert_em_internal_entry(struct tf *tfp,
+				    struct tf_insert_em_entry_parms *params,
+				    uint16_t *rptr_index,
+				    uint8_t *rptr_entry,
+				    uint8_t *num_of_entries);
+/**
+ * Sends EM internal delete request to Firmware
+ */
+int tf_msg_delete_em_entry(struct tf *tfp,
+			   struct tf_delete_em_entry_parms *em_parms);
+/**
  * Sends EM mem register request to Firmware
  */
 int tf_msg_em_mem_rgtr(struct tf *tfp,
diff --git a/drivers/net/bnxt/tf_core/tf_session.h b/drivers/net/bnxt/tf_core/tf_session.h
index 50ef2d5..c9f4f8f 100644
--- a/drivers/net/bnxt/tf_core/tf_session.h
+++ b/drivers/net/bnxt/tf_core/tf_session.h
@@ -13,12 +13,25 @@
 #include "tf_core.h"
 #include "tf_rm.h"
 #include "tf_tbl.h"
+#include "stack.h"
 
 /** Session defines
  */
 #define TF_SESSIONS_MAX	          1          /** max # sessions */
 #define TF_SESSION_ID_INVALID     0xFFFFFFFF /** Invalid Session ID define */
 
+/**
+ * Number of EM entries. Static for now will be removed
+ * when parameter added at a later date. At this stage we
+ * are using fixed size entries so that each stack entry
+ * represents 4 RT (f/n)blocks. So we take the total block
+ * allocation for truflow and divide that by 4.
+ */
+#define TF_SESSION_TOTAL_FN_BLOCKS (1024 * 8) /* 8K blocks */
+#define TF_SESSION_EM_ENTRY_SIZE 4 /* 4 blocks per entry */
+#define TF_SESSION_EM_POOL_SIZE \
+	(TF_SESSION_TOTAL_FN_BLOCKS / TF_SESSION_EM_ENTRY_SIZE)
+
 /** Session
  *
  * Shared memory containing private TruFlow session information.
@@ -289,6 +302,11 @@ struct tf_session {
 
 	/** Table scope array */
 	struct tf_tbl_scope_cb tbl_scopes[TF_NUM_TBL_SCOPE];
+
+	/**
+	 * EM Pools
+	 */
+	struct stack em_pool[TF_DIR_MAX];
 };
 
 #endif /* _TF_SESSION_H_ */
diff --git a/drivers/net/bnxt/tf_core/tf_tbl.c b/drivers/net/bnxt/tf_core/tf_tbl.c
index d900c9c..b9c71d4 100644
--- a/drivers/net/bnxt/tf_core/tf_tbl.c
+++ b/drivers/net/bnxt/tf_core/tf_tbl.c
@@ -156,7 +156,7 @@ tf_em_alloc_pg_tbl(struct tf_em_page_tbl *tp,
 		if (tfp_calloc(&parms) != 0)
 			goto cleanup;
 
-		tp->pg_pa_tbl[i] = (uint64_t)(uintptr_t)parms.mem_pa;
+		tp->pg_pa_tbl[i] = (uintptr_t)parms.mem_pa;
 		tp->pg_va_tbl[i] = parms.mem_va;
 
 		memset(tp->pg_va_tbl[i], 0, pg_size);
@@ -727,13 +727,13 @@ tf_em_validate_num_entries(struct tf_tbl_scope_cb *tbl_scope_cb,
 	tbl_scope_cb->em_ctx_info[TF_DIR_RX].em_tables[KEY1_TABLE].entry_size =
 		parms->rx_max_key_sz_in_bits / 8;
 
-	tbl_scope_cb->em_ctx_info[TF_DIR_RX].em_tables[RECORD_TABLE].num_entries =
-		parms->rx_num_flows_in_k * TF_KILOBYTE;
-	tbl_scope_cb->em_ctx_info[TF_DIR_RX].em_tables[RECORD_TABLE].entry_size =
-		parms->rx_max_action_entry_sz_in_bits / 8;
+	tbl_scope_cb->em_ctx_info[TF_DIR_RX].em_tables[RECORD_TABLE].num_entries
+		= parms->rx_num_flows_in_k * TF_KILOBYTE;
+	tbl_scope_cb->em_ctx_info[TF_DIR_RX].em_tables[RECORD_TABLE].entry_size
+		= parms->rx_max_action_entry_sz_in_bits / 8;
 
-	tbl_scope_cb->em_ctx_info[TF_DIR_RX].em_tables[EFC_TABLE].num_entries =
-		0;
+	tbl_scope_cb->em_ctx_info[TF_DIR_RX].em_tables[EFC_TABLE].num_entries
+		= 0;
 
 	/* Tx */
 	tbl_scope_cb->em_ctx_info[TF_DIR_TX].em_tables[KEY0_TABLE].num_entries =
@@ -746,13 +746,13 @@ tf_em_validate_num_entries(struct tf_tbl_scope_cb *tbl_scope_cb,
 	tbl_scope_cb->em_ctx_info[TF_DIR_TX].em_tables[KEY1_TABLE].entry_size =
 		parms->tx_max_key_sz_in_bits / 8;
 
-	tbl_scope_cb->em_ctx_info[TF_DIR_TX].em_tables[RECORD_TABLE].num_entries =
-		parms->tx_num_flows_in_k * TF_KILOBYTE;
-	tbl_scope_cb->em_ctx_info[TF_DIR_TX].em_tables[RECORD_TABLE].entry_size =
-		parms->tx_max_action_entry_sz_in_bits / 8;
+	tbl_scope_cb->em_ctx_info[TF_DIR_TX].em_tables[RECORD_TABLE].num_entries
+		= parms->tx_num_flows_in_k * TF_KILOBYTE;
+	tbl_scope_cb->em_ctx_info[TF_DIR_TX].em_tables[RECORD_TABLE].entry_size
+		= parms->tx_max_action_entry_sz_in_bits / 8;
 
-	tbl_scope_cb->em_ctx_info[TF_DIR_TX].em_tables[EFC_TABLE].num_entries =
-		0;
+	tbl_scope_cb->em_ctx_info[TF_DIR_TX].em_tables[EFC_TABLE].num_entries
+		= 0;
 
 	return 0;
 }
@@ -792,7 +792,8 @@ tf_set_tbl_entry_internal(struct tf *tfp,
 	index = parms->idx;
 
 	if (parms->type != TF_TBL_TYPE_FULL_ACT_RECORD &&
-	    parms->type != TF_TBL_TYPE_ACT_SP_SMAC_IPV4) {
+	    parms->type != TF_TBL_TYPE_ACT_SP_SMAC_IPV4 &&
+	    parms->type != TF_TBL_TYPE_ACT_STATS_64) {
 		PMD_DRV_LOG(ERR,
 			    "dir:%d, Type not supported, type:%d\n",
 			    parms->dir,
@@ -1179,7 +1180,8 @@ tf_alloc_tbl_entry_pool_internal(struct tf *tfp,
 	    parms->type != TF_TBL_TYPE_ACT_SP_SMAC_IPV4 &&
 	    parms->type != TF_TBL_TYPE_ACT_ENCAP_8B &&
 	    parms->type != TF_TBL_TYPE_ACT_ENCAP_16B &&
-	    parms->type != TF_TBL_TYPE_ACT_ENCAP_64B) {
+	    parms->type != TF_TBL_TYPE_ACT_ENCAP_64B &&
+	    parms->type != TF_TBL_TYPE_ACT_STATS_64) {
 		PMD_DRV_LOG(ERR,
 			    "dir:%d, Type not supported, type:%d\n",
 			    parms->dir,
@@ -1330,7 +1332,8 @@ tf_free_tbl_entry_pool_internal(struct tf *tfp,
 	    parms->type != TF_TBL_TYPE_ACT_SP_SMAC_IPV4 &&
 	    parms->type != TF_TBL_TYPE_ACT_ENCAP_8B &&
 	    parms->type != TF_TBL_TYPE_ACT_ENCAP_16B &&
-	    parms->type != TF_TBL_TYPE_ACT_ENCAP_64B) {
+	    parms->type != TF_TBL_TYPE_ACT_ENCAP_64B &&
+	    parms->type != TF_TBL_TYPE_ACT_STATS_64) {
 		PMD_DRV_LOG(ERR,
 			    "dir:%d, Type not supported, type:%d\n",
 			    parms->dir,
@@ -1801,3 +1804,91 @@ tf_free_tbl_entry(struct tf *tfp,
 			    rc);
 	return rc;
 }
+
+
+static void
+tf_dump_link_page_table(struct tf_em_page_tbl *tp,
+			struct tf_em_page_tbl *tp_next)
+{
+	uint64_t *pg_va;
+	uint32_t i;
+	uint32_t j;
+	uint32_t k = 0;
+
+	printf("pg_count:%d pg_size:0x%x\n",
+	       tp->pg_count,
+	       tp->pg_size);
+	for (i = 0; i < tp->pg_count; i++) {
+		pg_va = tp->pg_va_tbl[i];
+		printf("\t%p\n", (void *)pg_va);
+		for (j = 0; j < MAX_PAGE_PTRS(tp->pg_size); j++) {
+			printf("\t\t%p\n", (void *)(uintptr_t)pg_va[j]);
+			if (((pg_va[j] & 0x7) ==
+			     tfp_cpu_to_le_64(PTU_PTE_LAST |
+					      PTU_PTE_VALID)))
+				return;
+
+			if (!(pg_va[j] & tfp_cpu_to_le_64(PTU_PTE_VALID))) {
+				printf("** Invalid entry **\n");
+				return;
+			}
+
+			if (++k >= tp_next->pg_count) {
+				printf("** Shouldn't get here **\n");
+				return;
+			}
+		}
+	}
+}
+
+void tf_dump_dma(struct tf *tfp, uint32_t tbl_scope_id);
+
+void tf_dump_dma(struct tf *tfp, uint32_t tbl_scope_id)
+{
+	struct tf_session      *session;
+	struct tf_tbl_scope_cb *tbl_scope_cb;
+	struct tf_em_page_tbl *tp;
+	struct tf_em_page_tbl *tp_next;
+	struct tf_em_table *tbl;
+	int i;
+	int j;
+	int dir;
+
+	printf("called %s\n", __func__);
+
+	/* find session struct */
+	session = (struct tf_session *)tfp->session->core_data;
+
+	/* find control block for table scope */
+	tbl_scope_cb = tbl_scope_cb_find(session,
+					 tbl_scope_id);
+	if (tbl_scope_cb == NULL)
+		PMD_DRV_LOG(ERR, "No table scope\n");
+
+	for (dir = 0; dir < TF_DIR_MAX; dir++) {
+		printf("Direction %s:\n", (dir == TF_DIR_RX ? "Rx" : "Tx"));
+
+		for (j = KEY0_TABLE; j < MAX_TABLE; j++) {
+			tbl = &tbl_scope_cb->em_ctx_info[dir].em_tables[j];
+			printf
+	("Table: j:%d type:%d num_entries:%d entry_size:0x%x num_lvl:%d ",
+			       j,
+			       tbl->type,
+			       tbl->num_entries,
+			       tbl->entry_size,
+			       tbl->num_lvl);
+			if (tbl->pg_tbl[0].pg_va_tbl &&
+			    tbl->pg_tbl[0].pg_pa_tbl)
+				printf("%p %p\n",
+			       tbl->pg_tbl[0].pg_va_tbl[0],
+			       (void *)(uintptr_t)tbl->pg_tbl[0].pg_pa_tbl[0]);
+			for (i = 0; i < tbl->num_lvl - 1; i++) {
+				printf("Level:%d\n", i);
+				tp = &tbl->pg_tbl[i];
+				tp_next = &tbl->pg_tbl[i + 1];
+				tf_dump_link_page_table(tp, tp_next);
+			}
+			printf("\n");
+		}
+	}
+}
diff --git a/drivers/net/bnxt/tf_core/tf_tbl.h b/drivers/net/bnxt/tf_core/tf_tbl.h
index bdc6288..6cda487 100644
--- a/drivers/net/bnxt/tf_core/tf_tbl.h
+++ b/drivers/net/bnxt/tf_core/tf_tbl.h
@@ -76,38 +76,51 @@ struct tf_tbl_scope_cb {
 	uint32_t                  *ext_act_pool_mem[TF_DIR_MAX];
 };
 
-/** Hardware Page sizes supported for EEM: 4K, 8K, 64K, 256K, 1M, 2M, 4M, 1G.
- * Round-down other page sizes to the lower hardware page size supported.
+/**
+ * Hardware Page sizes supported for EEM:
+ *   4K, 8K, 64K, 256K, 1M, 2M, 4M, 1G.
+ *
+ * Round-down other page sizes to the lower hardware page
+ * size supported.
  */
-#define BNXT_PAGE_SHIFT 22 /** 2M */
+#define TF_EM_PAGE_SIZE_4K 12
+#define TF_EM_PAGE_SIZE_8K 13
+#define TF_EM_PAGE_SIZE_64K 16
+#define TF_EM_PAGE_SIZE_256K 18
+#define TF_EM_PAGE_SIZE_1M 20
+#define TF_EM_PAGE_SIZE_2M 21
+#define TF_EM_PAGE_SIZE_4M 22
+#define TF_EM_PAGE_SIZE_1G 30
+
+/* Set page size */
+#define PAGE_SIZE TF_EM_PAGE_SIZE_2M
 
-#if (BNXT_PAGE_SHIFT < 12)				/** < 4K >> 4K */
-#define TF_EM_PAGE_SHIFT 12
+#if (PAGE_SIZE == TF_EM_PAGE_SIZE_4K)	/** 4K */
+#define TF_EM_PAGE_SHIFT TF_EM_PAGE_SIZE_4K
 #define TF_EM_PAGE_SIZE_ENUM HWRM_TF_CTXT_MEM_RGTR_INPUT_PAGE_SIZE_4K
-#elif (BNXT_PAGE_SHIFT <= 13)			/** 4K, 8K */
-#define TF_EM_PAGE_SHIFT 13
+#elif (PAGE_SIZE == TF_EM_PAGE_SIZE_8K)	/** 8K */
+#define TF_EM_PAGE_SHIFT TF_EM_PAGE_SIZE_8K
 #define TF_EM_PAGE_SIZE_ENUM HWRM_TF_CTXT_MEM_RGTR_INPUT_PAGE_SIZE_8K
-#elif (BNXT_PAGE_SHIFT < 16)				/** 16K, 32K >> 8K */
-#define TF_EM_PAGE_SHIFT 15
-#define TF_EM_PAGE_SIZE_ENUM HWRM_TF_CTXT_MEM_RGTR_INPUT_PAGE_SIZE_32K
-#elif (BNXT_PAGE_SHIFT <= 17)			/** 64K, 128K >> 64K */
-#define TF_EM_PAGE_SHIFT 16
+#elif (PAGE_SIZE == TF_EM_PAGE_SIZE_64K)	/** 64K */
+#define TF_EM_PAGE_SHIFT TF_EM_PAGE_SIZE_64K
 #define TF_EM_PAGE_SIZE_ENUM HWRM_TF_CTXT_MEM_RGTR_INPUT_PAGE_SIZE_64K
-#elif (BNXT_PAGE_SHIFT <= 19)			/** 256K, 512K >> 256K */
-#define TF_EM_PAGE_SHIFT 18
+#elif (PAGE_SIZE == TF_EM_PAGE_SIZE_256K)	/** 256K */
+#define TF_EM_PAGE_SHIFT TF_EM_PAGE_SIZE_256K
 #define TF_EM_PAGE_SIZE_ENUM HWRM_TF_CTXT_MEM_RGTR_INPUT_PAGE_SIZE_256K
-#elif (BNXT_PAGE_SHIFT <= 21)			/** 1M */
-#define TF_EM_PAGE_SHIFT 20
+#elif (PAGE_SIZE == TF_EM_PAGE_SIZE_1M)	/** 1M */
+#define TF_EM_PAGE_SHIFT TF_EM_PAGE_SIZE_1M
 #define TF_EM_PAGE_SIZE_ENUM HWRM_TF_CTXT_MEM_RGTR_INPUT_PAGE_SIZE_1M
-#elif (BNXT_PAGE_SHIFT <= 22)			/** 2M, 4M */
-#define TF_EM_PAGE_SHIFT 21
+#elif (PAGE_SIZE == TF_EM_PAGE_SIZE_2M)	/** 2M */
+#define TF_EM_PAGE_SHIFT TF_EM_PAGE_SIZE_2M
 #define TF_EM_PAGE_SIZE_ENUM HWRM_TF_CTXT_MEM_RGTR_INPUT_PAGE_SIZE_2M
-#elif (BNXT_PAGE_SHIFT <= 29)			/** 8M ... 512M >> 4M */
-#define TF_EM_PAGE_SHIFT 22
+#elif (PAGE_SIZE == TF_EM_PAGE_SIZE_4M)	/** 4M */
+#define TF_EM_PAGE_SHIFT TF_EM_PAGE_SIZE_4M
 #define TF_EM_PAGE_SIZE_ENUM HWRM_TF_CTXT_MEM_RGTR_INPUT_PAGE_SIZE_4M
-#else						/** >= 1G >> 1G */
-#define TF_EM_PAGE_SHIFT	30
+#elif (PAGE_SIZE == TF_EM_PAGE_SIZE_1G)	/** 1G */
+#define TF_EM_PAGE_SHIFT TF_EM_PAGE_SIZE_1G
 #define TF_EM_PAGE_SIZE_ENUM HWRM_TF_CTXT_MEM_RGTR_INPUT_PAGE_SIZE_1G
+#else
+#error "Invalid Page Size specified. Please use a TF_EM_PAGE_SIZE_n define"
 #endif
 
 #define TF_EM_PAGE_SIZE	(1 << TF_EM_PAGE_SHIFT)
diff --git a/drivers/net/bnxt/tf_core/tfp.h b/drivers/net/bnxt/tf_core/tfp.h
index 8d5e94e..fe49b63 100644
--- a/drivers/net/bnxt/tf_core/tfp.h
+++ b/drivers/net/bnxt/tf_core/tfp.h
@@ -3,14 +3,23 @@
  * All rights reserved.
  */
 
-/* This header file defines the Portability structures and APIs for
+/*
+ * This header file defines the Portability structures and APIs for
  * TruFlow.
  */
 
 #ifndef _TFP_H_
 #define _TFP_H_
 
+#include <rte_config.h>
 #include <rte_spinlock.h>
+#include <rte_log.h>
+#include <rte_byteorder.h>
+
+/**
+ * DPDK/Driver specific log level for the BNXT Eth driver.
+ */
+extern int bnxt_logtype_driver;
 
 /** Spinlock
  */
@@ -18,13 +27,21 @@ struct tfp_spinlock_parms {
 	rte_spinlock_t slock;
 };
 
+#define TFP_DRV_LOG_RAW(level, fmt, args...) \
+	rte_log(RTE_LOG_ ## level, bnxt_logtype_driver, "%s(): " fmt, \
+		__func__, ## args)
+
+#define TFP_DRV_LOG(level, fmt, args...) \
+	TFP_DRV_LOG_RAW(level, fmt, ## args)
+
 /**
  * @file
  *
  * TrueFlow Portability API Header File
  */
 
-/** send message parameter definition
+/**
+ * send message parameter definition
  */
 struct tfp_send_msg_parms {
 	/**
@@ -62,7 +79,8 @@ struct tfp_send_msg_parms {
 	uint32_t *resp_data;
 };
 
-/** calloc parameter definition
+/**
+ * calloc parameter definition
  */
 struct tfp_calloc_parms {
 	/**
@@ -96,43 +114,15 @@ struct tfp_calloc_parms {
  * @ref tfp_send_msg_tunneled
  *
  * @ref tfp_calloc
- * @ref tfp_free
  * @ref tfp_memcpy
+ * @ref tfp_free
  *
  * @ref tfp_spinlock_init
  * @ref tfp_spinlock_lock
  * @ref tfp_spinlock_unlock
  *
- * @ref tfp_cpu_to_le_16
- * @ref tfp_le_to_cpu_16
- * @ref tfp_cpu_to_le_32
- * @ref tfp_le_to_cpu_32
- * @ref tfp_cpu_to_le_64
- * @ref tfp_le_to_cpu_64
- * @ref tfp_cpu_to_be_16
- * @ref tfp_be_to_cpu_16
- * @ref tfp_cpu_to_be_32
- * @ref tfp_be_to_cpu_32
- * @ref tfp_cpu_to_be_64
- * @ref tfp_be_to_cpu_64
  */
 
-#define tfp_cpu_to_le_16(val) rte_cpu_to_le_16(val)
-#define tfp_le_to_cpu_16(val) rte_le_to_cpu_16(val)
-#define tfp_cpu_to_le_32(val) rte_cpu_to_le_32(val)
-#define tfp_le_to_cpu_32(val) rte_le_to_cpu_32(val)
-#define tfp_cpu_to_le_64(val) rte_cpu_to_le_64(val)
-#define tfp_le_to_cpu_64(val) rte_le_to_cpu_64(val)
-#define tfp_cpu_to_be_16(val) rte_cpu_to_be_16(val)
-#define tfp_be_to_cpu_16(val) rte_be_to_cpu_16(val)
-#define tfp_cpu_to_be_32(val) rte_cpu_to_be_32(val)
-#define tfp_be_to_cpu_32(val) rte_be_to_cpu_32(val)
-#define tfp_cpu_to_be_64(val) rte_cpu_to_be_64(val)
-#define tfp_be_to_cpu_64(val) rte_be_to_cpu_64(val)
-#define tfp_bswap_16(val) rte_bswap16(val)
-#define tfp_bswap_32(val) rte_bswap32(val)
-#define tfp_bswap_64(val) rte_bswap64(val)
-
 /**
  * Provides communication capability from the TrueFlow API layer to
  * the TrueFlow firmware. The portability layer internally provides
@@ -162,10 +152,25 @@ int tfp_send_msg_direct(struct tf *tfp,
  *   -1             - Global error like not supported
  *   -EINVAL        - Parameter Error
  */
-int tfp_send_msg_tunneled(struct tf                 *tfp,
+int tfp_send_msg_tunneled(struct tf *tfp,
 			  struct tfp_send_msg_parms *parms);
 
 /**
+ * Sends OEM command message to Chimp
+ *
+ * [in] session, pointer to session handle
+ * [in] max_flows, max number of flows requested
+ *
+ * Returns:
+ *   0              - Success
+ *   -1             - Global error like not supported
+ *   -EINVAL        - Parameter Error
+ */
+int
+tfp_msg_hwrm_oem_cmd(struct tf *tfp,
+		     uint32_t max_flows);
+
+/**
  * Allocates zero'ed memory from the heap.
  *
  * NOTE: Also performs virt2phy address conversion by default thus is
@@ -179,10 +184,58 @@ int tfp_send_msg_tunneled(struct tf                 *tfp,
  *   -EINVAL        - Parameter error
  */
 int tfp_calloc(struct tfp_calloc_parms *parms);
-
-void tfp_free(void *addr);
 void tfp_memcpy(void *dest, void *src, size_t n);
+void tfp_free(void *addr);
+
 void tfp_spinlock_init(struct tfp_spinlock_parms *slock);
 void tfp_spinlock_lock(struct tfp_spinlock_parms *slock);
 void tfp_spinlock_unlock(struct tfp_spinlock_parms *slock);
+
+/**
+ * Lookup of the FID in the platform specific structure.
+ *
+ * [in] session
+ *   Pointer to session handle
+ *
+ * [out] fw_fid
+ *   Pointer to the fw_fid
+ *
+ * Returns:
+ *   0       - Success
+ *   -EINVAL - Parameter error
+ */
+int tfp_get_fid(struct tf *tfp, uint16_t *fw_fid);
+
+
+/*
+ * @ref tfp_cpu_to_le_16
+ * @ref tfp_le_to_cpu_16
+ * @ref tfp_cpu_to_le_32
+ * @ref tfp_le_to_cpu_32
+ * @ref tfp_cpu_to_le_64
+ * @ref tfp_le_to_cpu_64
+ * @ref tfp_cpu_to_be_16
+ * @ref tfp_be_to_cpu_16
+ * @ref tfp_cpu_to_be_32
+ * @ref tfp_be_to_cpu_32
+ * @ref tfp_cpu_to_be_64
+ * @ref tfp_be_to_cpu_64
+ */
+
+#define tfp_cpu_to_le_16(val) rte_cpu_to_le_16(val)
+#define tfp_le_to_cpu_16(val) rte_le_to_cpu_16(val)
+#define tfp_cpu_to_le_32(val) rte_cpu_to_le_32(val)
+#define tfp_le_to_cpu_32(val) rte_le_to_cpu_32(val)
+#define tfp_cpu_to_le_64(val) rte_cpu_to_le_64(val)
+#define tfp_le_to_cpu_64(val) rte_le_to_cpu_64(val)
+#define tfp_cpu_to_be_16(val) rte_cpu_to_be_16(val)
+#define tfp_be_to_cpu_16(val) rte_be_to_cpu_16(val)
+#define tfp_cpu_to_be_32(val) rte_cpu_to_be_32(val)
+#define tfp_be_to_cpu_32(val) rte_be_to_cpu_32(val)
+#define tfp_cpu_to_be_64(val) rte_cpu_to_be_64(val)
+#define tfp_be_to_cpu_64(val) rte_be_to_cpu_64(val)
+#define tfp_bswap_16(val) rte_bswap16(val)
+#define tfp_bswap_32(val) rte_bswap32(val)
+#define tfp_bswap_64(val) rte_bswap64(val)
+
 #endif /* _TFP_H_ */
-- 
2.7.4


  parent reply	other threads:[~2020-06-12 13:37 UTC|newest]

Thread overview: 271+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-06-12 13:28 [dpdk-dev] [PATCH 00/50] add features for host-based flow management Somnath Kotur
2020-06-12 13:28 ` [dpdk-dev] [PATCH 01/50] net/bnxt: Basic infrastructure support for VF representors Somnath Kotur
2020-06-12 13:28 ` [dpdk-dev] [PATCH 02/50] net/bnxt: Infrastructure support for VF-reps data path Somnath Kotur
2020-06-12 13:28 ` [dpdk-dev] [PATCH 03/50] net/bnxt: add support to get FID, default vnic ID and svif of VF-Rep Endpoint Somnath Kotur
2020-06-12 13:28 ` [dpdk-dev] [PATCH 04/50] net/bnxt: initialize parent PF information Somnath Kotur
2020-06-12 13:28 ` [dpdk-dev] [PATCH 05/50] net/bnxt: modify ulp_port_db_dev_port_intf_update prototype Somnath Kotur
2020-06-12 13:28 ` [dpdk-dev] [PATCH 06/50] net/bnxt: get port & function related information Somnath Kotur
2020-06-12 13:28 ` [dpdk-dev] [PATCH 07/50] net/bnxt: add support for bnxt_hwrm_port_phy_qcaps Somnath Kotur
2020-06-12 13:28 ` [dpdk-dev] [PATCH 08/50] net/bnxt: modify port_db to store & retrieve more info Somnath Kotur
2020-06-12 13:28 ` Somnath Kotur [this message]
2020-06-12 13:28 ` [dpdk-dev] [PATCH 10/50] net/bnxt: modify EM insert and delete to use HWRM direct Somnath Kotur
2020-06-12 13:28 ` [dpdk-dev] [PATCH 11/50] net/bnxt: add multi device support Somnath Kotur
2020-06-12 13:28 ` [dpdk-dev] [PATCH 12/50] net/bnxt: support bulk table get and mirror Somnath Kotur
2020-06-12 13:28 ` [dpdk-dev] [PATCH 13/50] net/bnxt: update multi device design support Somnath Kotur
2020-06-12 13:28 ` [dpdk-dev] [PATCH 14/50] net/bnxt: support two-level priority for TCAMs Somnath Kotur
2020-06-12 13:28 ` [dpdk-dev] [PATCH 15/50] net/bnxt: add HCAPI interface support Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 16/50] net/bnxt: add core changes for EM and EEM lookups Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 17/50] net/bnxt: implement support for TCAM access Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 18/50] net/bnxt: multiple device implementation Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 19/50] net/bnxt: update identifier with remap support Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 20/50] net/bnxt: update RM with residual checker Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 21/50] net/bnxt: support two level priority for TCAMs Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 22/50] net/bnxt: support EM and TCAM lookup with table scope Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 23/50] net/bnxt: update table get to use new design Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 24/50] net/bnxt: update RM to support HCAPI only Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 25/50] net/bnxt: remove table scope from session Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 26/50] net/bnxt: add external action alloc and free Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 27/50] net/bnxt: align CFA resources with RM Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 28/50] net/bnxt: implement IF tables set and get Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 29/50] net/bnxt: add TF register and unregister Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 30/50] net/bnxt: add global config set and get APIs Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 31/50] net/bnxt: add support for EEM System memory Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 32/50] net/bnxt: integrate with the latest tf_core library Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 33/50] net/bnxt: add support for internal encap records Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 34/50] net/bnxt: add support for if table processing Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 35/50] net/bnxt: disable vector mode in tx direction when truflow is enabled Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 36/50] net/bnxt: add index opcode and index operand mapper table Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 37/50] net/bnxt: add support for global resource templates Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 38/50] net/bnxt: add support for internal exact match entries Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 39/50] net/bnxt: add support for conditional execution of mapper tables Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 40/50] net/bnxt: enable HWRM_PORT_MAC_QCFG for trusted vf Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 41/50] net/bnxt: enhancements for port db Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 42/50] net/bnxt: fix for VF to VFR conduit Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 43/50] net/bnxt: fix to parse representor along with other dev-args Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 44/50] net/bnxt: fill mapper parameters with default rules info Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 45/50] net/bnxt: add support for vf rep and stat templates Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 46/50] net/bnxt: create default flow rules for the VF-rep conduit Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 47/50] net/bnxt: add ingress & egress port default rules Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 48/50] net/bnxt: fill cfa_action in the tx buffer descriptor properly Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 49/50] net/bnxt: support for ULP Flow counter Manager Somnath Kotur
2020-06-12 13:29 ` [dpdk-dev] [PATCH 50/50] net/bnxt: Add support for flow query with action_type COUNT Somnath Kotur
2020-07-01  6:51 ` [dpdk-dev] [PATCH v2 00/51] add features for host-based flow management Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 01/51] net/bnxt: add basic infrastructure for VF representors Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 02/51] net/bnxt: add support for VF-reps data path Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 03/51] net/bnxt: get IDs for VF-Rep endpoint Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 04/51] net/bnxt: initialize parent PF information Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 05/51] net/bnxt: modify port db dev interface Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 06/51] net/bnxt: get port and function info Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 07/51] net/bnxt: add support for hwrm port phy qcaps Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 08/51] net/bnxt: modify port db to handle more info Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 09/51] net/bnxt: add support for exact match Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 10/51] net/bnxt: modify EM insert and delete to use HWRM direct Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 11/51] net/bnxt: add multi device support Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 12/51] net/bnxt: support bulk table get and mirror Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 13/51] net/bnxt: update multi device design support Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 14/51] net/bnxt: support two-level priority for TCAMs Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 15/51] net/bnxt: add HCAPI interface support Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 16/51] net/bnxt: add core changes for EM and EEM lookups Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 17/51] net/bnxt: implement support for TCAM access Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 18/51] net/bnxt: multiple device implementation Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 19/51] net/bnxt: update identifier with remap support Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 20/51] net/bnxt: update RM with residual checker Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 21/51] net/bnxt: support two level priority for TCAMs Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 22/51] net/bnxt: support EM and TCAM lookup with table scope Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 23/51] net/bnxt: update table get to use new design Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 24/51] net/bnxt: update RM to support HCAPI only Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 25/51] net/bnxt: remove table scope from session Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 26/51] net/bnxt: add external action alloc and free Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 27/51] net/bnxt: align CFA resources with RM Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 28/51] net/bnxt: implement IF tables set and get Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 29/51] net/bnxt: add TF register and unregister Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 30/51] net/bnxt: add global config set and get APIs Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 31/51] net/bnxt: add support for EEM System memory Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 32/51] net/bnxt: integrate with the latest tf core changes Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 33/51] net/bnxt: add support for internal encap records Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 34/51] net/bnxt: add support for if table processing Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 35/51] net/bnxt: disable Tx vector mode if truflow is enabled Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 36/51] net/bnxt: add index opcode and operand to mapper table Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 37/51] net/bnxt: add support for global resource templates Ajit Khaparde
2020-07-01  6:51   ` [dpdk-dev] [PATCH v2 38/51] net/bnxt: add support for internal exact match entries Ajit Khaparde
2020-07-01  6:52   ` [dpdk-dev] [PATCH v2 39/51] net/bnxt: add support for conditional execution of mapper tables Ajit Khaparde
2020-07-01  6:52   ` [dpdk-dev] [PATCH v2 40/51] net/bnxt: enable port MAC qcfg command for trusted VF Ajit Khaparde
2020-07-01  6:52   ` [dpdk-dev] [PATCH v2 41/51] net/bnxt: enhancements for port db Ajit Khaparde
2020-07-01  6:52   ` [dpdk-dev] [PATCH v2 42/51] net/bnxt: manage VF to VFR conduit Ajit Khaparde
2020-07-01  6:52   ` [dpdk-dev] [PATCH v2 43/51] net/bnxt: parse representor along with other dev-args Ajit Khaparde
2020-07-01  6:52   ` [dpdk-dev] [PATCH v2 44/51] net/bnxt: fill mapper parameters with default rules info Ajit Khaparde
2020-07-01  6:52   ` [dpdk-dev] [PATCH v2 45/51] net/bnxt: add VF-rep and stat templates Ajit Khaparde
2020-07-01  6:52   ` [dpdk-dev] [PATCH v2 46/51] net/bnxt: create default flow rules for the VF-rep conduit Ajit Khaparde
2020-07-01  6:52   ` [dpdk-dev] [PATCH v2 47/51] net/bnxt: add port default rules for ingress and egress Ajit Khaparde
2020-07-01  6:52   ` [dpdk-dev] [PATCH v2 48/51] net/bnxt: fill cfa action in the Tx descriptor Ajit Khaparde
2020-07-01  6:52   ` [dpdk-dev] [PATCH v2 49/51] net/bnxt: add ULP Flow counter Manager Ajit Khaparde
2020-07-01  6:52   ` [dpdk-dev] [PATCH v2 50/51] net/bnxt: add support for count action in flow query Ajit Khaparde
2020-07-01  6:52   ` [dpdk-dev] [PATCH v2 51/51] doc: update release notes Ajit Khaparde
2020-07-01 14:26   ` [dpdk-dev] [PATCH v2 00/51] add features for host-based flow management Ajit Khaparde
2020-07-01 21:31     ` Ferruh Yigit
2020-07-02  4:10       ` [dpdk-dev] [PATCH v3 " Ajit Khaparde
2020-07-02  4:10         ` [dpdk-dev] [PATCH v3 01/51] net/bnxt: add basic infrastructure for VF reps Ajit Khaparde
2020-07-02  4:10         ` [dpdk-dev] [PATCH v3 02/51] net/bnxt: add support for VF-reps data path Ajit Khaparde
2020-07-02  4:10         ` [dpdk-dev] [PATCH v3 03/51] net/bnxt: get IDs for VF-Rep endpoint Ajit Khaparde
2020-07-02  4:10         ` [dpdk-dev] [PATCH v3 04/51] net/bnxt: initialize parent PF information Ajit Khaparde
2020-07-02  4:10         ` [dpdk-dev] [PATCH v3 05/51] net/bnxt: modify port db dev interface Ajit Khaparde
2020-07-02  4:10         ` [dpdk-dev] [PATCH v3 06/51] net/bnxt: get port and function info Ajit Khaparde
2020-07-02  4:10         ` [dpdk-dev] [PATCH v3 07/51] net/bnxt: add support for hwrm port phy qcaps Ajit Khaparde
2020-07-02  4:10         ` [dpdk-dev] [PATCH v3 08/51] net/bnxt: modify port db to handle more info Ajit Khaparde
2020-07-02  4:10         ` [dpdk-dev] [PATCH v3 09/51] net/bnxt: add support for exact match Ajit Khaparde
2020-07-02  4:10         ` [dpdk-dev] [PATCH v3 10/51] net/bnxt: modify EM insert and delete to use HWRM direct Ajit Khaparde
2020-07-02  4:10         ` [dpdk-dev] [PATCH v3 11/51] net/bnxt: add multi device support Ajit Khaparde
2020-07-02  4:10         ` [dpdk-dev] [PATCH v3 12/51] net/bnxt: support bulk table get and mirror Ajit Khaparde
2020-07-02  4:10         ` [dpdk-dev] [PATCH v3 13/51] net/bnxt: update multi device design support Ajit Khaparde
2020-07-02  4:10         ` [dpdk-dev] [PATCH v3 14/51] net/bnxt: support two-level priority for TCAMs Ajit Khaparde
2020-07-02  4:10         ` [dpdk-dev] [PATCH v3 15/51] net/bnxt: add HCAPI interface support Ajit Khaparde
2020-07-02  4:10         ` [dpdk-dev] [PATCH v3 16/51] net/bnxt: add core changes for EM and EEM lookups Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 17/51] net/bnxt: implement support for TCAM access Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 18/51] net/bnxt: multiple device implementation Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 19/51] net/bnxt: update identifier with remap support Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 20/51] net/bnxt: update RM with residual checker Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 21/51] net/bnxt: support two level priority for TCAMs Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 22/51] net/bnxt: support EM and TCAM lookup with table scope Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 23/51] net/bnxt: update table get to use new design Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 24/51] net/bnxt: update RM to support HCAPI only Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 25/51] net/bnxt: remove table scope from session Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 26/51] net/bnxt: add external action alloc and free Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 27/51] net/bnxt: align CFA resources with RM Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 28/51] net/bnxt: implement IF tables set and get Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 29/51] net/bnxt: add TF register and unregister Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 30/51] net/bnxt: add global config set and get APIs Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 31/51] net/bnxt: add support for EEM System memory Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 32/51] net/bnxt: integrate with the latest tf core changes Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 33/51] net/bnxt: add support for internal encap records Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 34/51] net/bnxt: add support for if table processing Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 35/51] net/bnxt: disable Tx vector mode if truflow is enabled Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 36/51] net/bnxt: add index opcode and operand to mapper table Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 37/51] net/bnxt: add support for global resource templates Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 38/51] net/bnxt: add support for internal exact match entries Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 39/51] net/bnxt: add conditional execution of mapper tables Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 40/51] net/bnxt: enable port MAC qcfg for trusted VF Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 41/51] net/bnxt: enhancements for port db Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 42/51] net/bnxt: manage VF to VFR conduit Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 43/51] net/bnxt: parse reps along with other dev-args Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 44/51] net/bnxt: fill mapper parameters with default rules Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 45/51] net/bnxt: add VF-rep and stat templates Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 46/51] net/bnxt: create default flow rules for the VF-rep Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 47/51] net/bnxt: add port default rules for ingress and egress Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 48/51] net/bnxt: fill cfa action in the Tx descriptor Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 49/51] net/bnxt: add ULP Flow counter Manager Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 50/51] net/bnxt: add support for count action in flow query Ajit Khaparde
2020-07-02  4:11         ` [dpdk-dev] [PATCH v3 51/51] doc: update release notes Ajit Khaparde
2020-07-02 23:27       ` [dpdk-dev] [PATCH v4 00/51] add features for host-based flow management Ajit Khaparde
2020-07-02 23:27         ` [dpdk-dev] [PATCH v4 01/51] net/bnxt: add basic infrastructure for VF reps Ajit Khaparde
2020-07-02 23:27         ` [dpdk-dev] [PATCH v4 02/51] net/bnxt: add support for VF-reps data path Ajit Khaparde
2020-07-02 23:27         ` [dpdk-dev] [PATCH v4 03/51] net/bnxt: get IDs for VF-Rep endpoint Ajit Khaparde
2020-07-02 23:27         ` [dpdk-dev] [PATCH v4 04/51] net/bnxt: initialize parent PF information Ajit Khaparde
2020-07-02 23:27         ` [dpdk-dev] [PATCH v4 05/51] net/bnxt: modify port db dev interface Ajit Khaparde
2020-07-02 23:27         ` [dpdk-dev] [PATCH v4 06/51] net/bnxt: get port and function info Ajit Khaparde
2020-07-02 23:27         ` [dpdk-dev] [PATCH v4 07/51] net/bnxt: add support for hwrm port phy qcaps Ajit Khaparde
2020-07-02 23:27         ` [dpdk-dev] [PATCH v4 08/51] net/bnxt: modify port db to handle more info Ajit Khaparde
2020-07-02 23:27         ` [dpdk-dev] [PATCH v4 09/51] net/bnxt: add support for exact match Ajit Khaparde
2020-07-02 23:27         ` [dpdk-dev] [PATCH v4 10/51] net/bnxt: modify EM insert and delete to use HWRM direct Ajit Khaparde
2020-07-02 23:27         ` [dpdk-dev] [PATCH v4 11/51] net/bnxt: add multi device support Ajit Khaparde
2020-07-02 23:27         ` [dpdk-dev] [PATCH v4 12/51] net/bnxt: support bulk table get and mirror Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 13/51] net/bnxt: update multi device design support Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 14/51] net/bnxt: support two-level priority for TCAMs Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 15/51] net/bnxt: add HCAPI interface support Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 16/51] net/bnxt: add core changes for EM and EEM lookups Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 17/51] net/bnxt: implement support for TCAM access Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 18/51] net/bnxt: multiple device implementation Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 19/51] net/bnxt: update identifier with remap support Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 20/51] net/bnxt: update RM with residual checker Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 21/51] net/bnxt: support two level priority for TCAMs Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 22/51] net/bnxt: support EM and TCAM lookup with table scope Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 23/51] net/bnxt: update table get to use new design Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 24/51] net/bnxt: update RM to support HCAPI only Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 25/51] net/bnxt: remove table scope from session Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 26/51] net/bnxt: add external action alloc and free Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 27/51] net/bnxt: align CFA resources with RM Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 28/51] net/bnxt: implement IF tables set and get Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 29/51] net/bnxt: add TF register and unregister Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 30/51] net/bnxt: add global config set and get APIs Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 31/51] net/bnxt: add support for EEM System memory Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 32/51] net/bnxt: integrate with the latest tf core changes Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 33/51] net/bnxt: add support for internal encap records Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 34/51] net/bnxt: add support for if table processing Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 35/51] net/bnxt: disable Tx vector mode if truflow is enabled Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 36/51] net/bnxt: add index opcode and operand to mapper table Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 37/51] net/bnxt: add support for global resource templates Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 38/51] net/bnxt: add support for internal exact match entries Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 39/51] net/bnxt: add support for conditional execution of mapper tables Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 40/51] net/bnxt: enable port MAC qcfg command for trusted VF Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 41/51] net/bnxt: enhancements for port db Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 42/51] net/bnxt: manage VF to VFR conduit Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 43/51] net/bnxt: parse reps along with other dev-args Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 44/51] net/bnxt: fill mapper parameters with default rules Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 45/51] net/bnxt: add VF-rep and stat templates Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 46/51] net/bnxt: create default flow rules for the VF-rep Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 47/51] net/bnxt: add port default rules for ingress and egress Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 48/51] net/bnxt: fill cfa action in the Tx descriptor Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 49/51] net/bnxt: add ULP Flow counter Manager Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 50/51] net/bnxt: add support for count action in flow query Ajit Khaparde
2020-07-02 23:28         ` [dpdk-dev] [PATCH v4 51/51] doc: update release notes Ajit Khaparde
2020-07-03 21:01       ` [dpdk-dev] [PATCH v5 00/51] net/bnxt: add features for host-based flow management Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 01/51] net/bnxt: add basic infrastructure for VF reps Ajit Khaparde
2020-07-06 10:07           ` Ferruh Yigit
2020-07-06 14:04             ` Somnath Kotur
2020-07-06 14:14               ` Ajit Khaparde
2020-07-06 18:35                 ` Ferruh Yigit
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 02/51] net/bnxt: add support for VF-reps data path Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 03/51] net/bnxt: get IDs for VF-Rep endpoint Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 04/51] net/bnxt: initialize parent PF information Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 05/51] net/bnxt: modify port db dev interface Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 06/51] net/bnxt: get port and function info Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 07/51] net/bnxt: add support for hwrm port phy qcaps Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 08/51] net/bnxt: modify port db to handle more info Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 09/51] net/bnxt: add support for exact match Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 10/51] net/bnxt: use HWRM direct for EM insert and delete Ajit Khaparde
2020-07-06 18:47           ` Ferruh Yigit
2020-07-06 19:11           ` Ferruh Yigit
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 11/51] net/bnxt: add multi device support Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 12/51] net/bnxt: support bulk table get and mirror Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 13/51] net/bnxt: update multi device design support Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 14/51] net/bnxt: support two-level priority for TCAMs Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 15/51] net/bnxt: add HCAPI interface support Ajit Khaparde
2020-07-07  8:03           ` Ferruh Yigit
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 16/51] net/bnxt: add core changes for EM and EEM lookups Ajit Khaparde
2020-07-07  8:08           ` Ferruh Yigit
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 17/51] net/bnxt: implement support for TCAM access Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 18/51] net/bnxt: multiple device implementation Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 19/51] net/bnxt: update identifier with remap support Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 20/51] net/bnxt: update RM with residual checker Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 21/51] net/bnxt: support two level priority for TCAMs Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 22/51] net/bnxt: use table scope for EM and TCAM lookup Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 23/51] net/bnxt: update table get to use new design Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 24/51] net/bnxt: update RM to support HCAPI only Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 25/51] net/bnxt: remove table scope from session Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 26/51] net/bnxt: add external action alloc and free Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 27/51] net/bnxt: align CFA resources with RM Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 28/51] net/bnxt: implement IF tables set and get Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 29/51] net/bnxt: add TF register and unregister Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 30/51] net/bnxt: add global config set and get APIs Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 31/51] net/bnxt: add support for EEM System memory Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 32/51] net/bnxt: integrate with the latest tf core changes Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 33/51] net/bnxt: add support for internal encap records Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 34/51] net/bnxt: add support for if table processing Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 35/51] net/bnxt: disable Tx vector mode if truflow is set Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 36/51] net/bnxt: add index opcode and operand to mapper table Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 37/51] net/bnxt: add support for global resource templates Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 38/51] net/bnxt: add support for internal exact match Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 39/51] net/bnxt: add conditional execution of mapper tables Ajit Khaparde
2020-07-03 21:01         ` [dpdk-dev] [PATCH v5 40/51] net/bnxt: allow port MAC qcfg command for trusted VF Ajit Khaparde
2020-07-03 21:02         ` [dpdk-dev] [PATCH v5 41/51] net/bnxt: enhancements for port db Ajit Khaparde
2020-07-03 21:02         ` [dpdk-dev] [PATCH v5 42/51] net/bnxt: manage VF to VFR conduit Ajit Khaparde
2020-07-03 21:02         ` [dpdk-dev] [PATCH v5 43/51] net/bnxt: parse reps along with other dev-args Ajit Khaparde
2020-07-03 21:02         ` [dpdk-dev] [PATCH v5 44/51] net/bnxt: fill mapper parameters with default rules Ajit Khaparde
2020-07-03 21:02         ` [dpdk-dev] [PATCH v5 45/51] net/bnxt: add VF-rep and stat templates Ajit Khaparde
2020-07-03 21:02         ` [dpdk-dev] [PATCH v5 46/51] net/bnxt: create default flow rules for the VF-rep Ajit Khaparde
2020-07-03 21:02         ` [dpdk-dev] [PATCH v5 47/51] net/bnxt: add port default rules for ingress and egress Ajit Khaparde
2020-07-03 21:02         ` [dpdk-dev] [PATCH v5 48/51] net/bnxt: fill cfa action in the Tx descriptor Ajit Khaparde
2020-07-03 21:02         ` [dpdk-dev] [PATCH v5 49/51] net/bnxt: add ULP Flow counter Manager Ajit Khaparde
2020-07-03 21:02         ` [dpdk-dev] [PATCH v5 50/51] net/bnxt: add support for count action in flow query Ajit Khaparde
2020-07-03 21:02         ` [dpdk-dev] [PATCH v5 51/51] doc: update release notes Ajit Khaparde
2020-07-06  1:47         ` [dpdk-dev] [PATCH v5 00/51] net/bnxt: add features for host-based flow management Ajit Khaparde
2020-07-06 10:10         ` Ferruh Yigit

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20200612132934.16488-10-somnath.kotur@broadcom.com \
    --to=somnath.kotur@broadcom.com \
    --cc=dev@dpdk.org \
    --cc=ferruh.yigit@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).