[v2,2/3] net/iavf: enable AVX512 for flexible RX
diff mbox series

Message ID 1600306778-46470-3-git-send-email-wenzhuo.lu@intel.com
State Superseded, archived
Delegated to: Qi Zhang
Headers show
Series
  • enable AVX512 for iavf
Related show

Checks

Context Check Description
ci/checkpatch success coding style OK

Commit Message

Wenzhuo Lu Sept. 17, 2020, 1:39 a.m. UTC
To enhance the per-core performance, this patch adds some AVX512
instructions to the data path to handle the flexible RX descriptors.

Signed-off-by: Wenzhuo Lu <wenzhuo.lu@intel.com>
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
Signed-off-by: Leyi Rong <leyi.rong@intel.com>
---
 drivers/net/iavf/iavf_rxtx.c            |  10 +
 drivers/net/iavf/iavf_rxtx.h            |   6 +
 drivers/net/iavf/iavf_rxtx_vec_avx512.c | 699 ++++++++++++++++++++++++++++++++
 3 files changed, 715 insertions(+)

Patch
diff mbox series

diff --git a/drivers/net/iavf/iavf_rxtx.c b/drivers/net/iavf/iavf_rxtx.c
index c36e809..0818107 100644
--- a/drivers/net/iavf/iavf_rxtx.c
+++ b/drivers/net/iavf/iavf_rxtx.c
@@ -2132,6 +2132,11 @@ 
 				dev->rx_pkt_burst = use_avx2 ?
 					iavf_recv_scattered_pkts_vec_avx2_flex_rxd :
 					iavf_recv_scattered_pkts_vec_flex_rxd;
+#ifdef CC_AVX512_SUPPORT
+				if (use_avx512)
+					dev->rx_pkt_burst =
+						iavf_recv_scattered_pkts_vec_avx512_flex_rxd;
+#endif
 			} else {
 				dev->rx_pkt_burst = use_avx2 ?
 					iavf_recv_scattered_pkts_vec_avx2 :
@@ -2151,6 +2156,11 @@ 
 				dev->rx_pkt_burst = use_avx2 ?
 					iavf_recv_pkts_vec_avx2_flex_rxd :
 					iavf_recv_pkts_vec_flex_rxd;
+#ifdef CC_AVX512_SUPPORT
+				if (use_avx512)
+					dev->rx_pkt_burst =
+						iavf_recv_pkts_vec_avx512_flex_rxd;
+#endif
 			} else {
 				dev->rx_pkt_burst = use_avx2 ?
 					iavf_recv_pkts_vec_avx2 :
diff --git a/drivers/net/iavf/iavf_rxtx.h b/drivers/net/iavf/iavf_rxtx.h
index cb12888..9653e0c 100644
--- a/drivers/net/iavf/iavf_rxtx.h
+++ b/drivers/net/iavf/iavf_rxtx.h
@@ -439,9 +439,15 @@  uint16_t iavf_xmit_pkts_vec_avx2(void *tx_queue, struct rte_mbuf **tx_pkts,
 int iavf_txq_vec_setup(struct iavf_tx_queue *txq);
 uint16_t iavf_recv_pkts_vec_avx512(void *rx_queue, struct rte_mbuf **rx_pkts,
 				   uint16_t nb_pkts);
+uint16_t iavf_recv_pkts_vec_avx512_flex_rxd(void *rx_queue,
+					    struct rte_mbuf **rx_pkts,
+					    uint16_t nb_pkts);
 uint16_t iavf_recv_scattered_pkts_vec_avx512(void *rx_queue,
 					     struct rte_mbuf **rx_pkts,
 					     uint16_t nb_pkts);
+uint16_t iavf_recv_scattered_pkts_vec_avx512_flex_rxd(void *rx_queue,
+						      struct rte_mbuf **rx_pkts,
+						      uint16_t nb_pkts);
 
 const uint32_t *iavf_get_default_ptype_table(void);
 
diff --git a/drivers/net/iavf/iavf_rxtx_vec_avx512.c b/drivers/net/iavf/iavf_rxtx_vec_avx512.c
index b528ed3..8c33661 100644
--- a/drivers/net/iavf/iavf_rxtx_vec_avx512.c
+++ b/drivers/net/iavf/iavf_rxtx_vec_avx512.c
@@ -643,6 +643,624 @@ 
 	return fdir_flags;
 }
 
+static inline uint16_t
+_iavf_recv_raw_pkts_vec_avx512_flex_rxd(struct iavf_rx_queue *rxq,
+					struct rte_mbuf **rx_pkts,
+					uint16_t nb_pkts, uint8_t *split_packet)
+{
+	const uint32_t *type_table = rxq->vsi->adapter->ptype_tbl;
+
+	const __m256i mbuf_init = _mm256_set_epi64x(0, 0,
+			0, rxq->mbuf_initializer);
+	struct rte_mbuf **sw_ring = &rxq->sw_ring[rxq->rx_tail];
+	volatile union iavf_rx_flex_desc *rxdp =
+		(union iavf_rx_flex_desc *)rxq->rx_ring + rxq->rx_tail;
+
+	rte_prefetch0(rxdp);
+
+	/* nb_pkts has to be floor-aligned to IAVF_DESCS_PER_LOOP_AVX */
+	nb_pkts = RTE_ALIGN_FLOOR(nb_pkts, IAVF_DESCS_PER_LOOP_AVX);
+
+	/* See if we need to rearm the RX queue - gives the prefetch a bit
+	 * of time to act
+	 */
+	if (rxq->rxrearm_nb > IAVF_RXQ_REARM_THRESH)
+		iavf_rxq_rearm(rxq);
+
+	/* Before we start moving massive data around, check to see if
+	 * there is actually a packet available
+	 */
+	if (!(rxdp->wb.status_error0 &
+			rte_cpu_to_le_32(1 << IAVF_RX_FLEX_DESC_STATUS0_DD_S)))
+		return 0;
+
+	/* constants used in processing loop */
+	const __m512i crc_adjust =
+		_mm512_set_epi16
+			(/* 1st descriptor */
+			 0, 0, 0,       /* ignore non-length fields */
+			 -rxq->crc_len, /* sub crc on data_len */
+			 0,             /* ignore high-16bits of pkt_len */
+			 -rxq->crc_len, /* sub crc on pkt_len */
+			 0, 0,          /* ignore pkt_type field */
+			 /* 2nd descriptor */
+			 0, 0, 0,       /* ignore non-length fields */
+			 -rxq->crc_len, /* sub crc on data_len */
+			 0,             /* ignore high-16bits of pkt_len */
+			 -rxq->crc_len, /* sub crc on pkt_len */
+			 0, 0,          /* ignore pkt_type field */
+			 /* 3rd descriptor */
+			 0, 0, 0,       /* ignore non-length fields */
+			 -rxq->crc_len, /* sub crc on data_len */
+			 0,             /* ignore high-16bits of pkt_len */
+			 -rxq->crc_len, /* sub crc on pkt_len */
+			 0, 0,          /* ignore pkt_type field */
+			 /* 4th descriptor */
+			 0, 0, 0,       /* ignore non-length fields */
+			 -rxq->crc_len, /* sub crc on data_len */
+			 0,             /* ignore high-16bits of pkt_len */
+			 -rxq->crc_len, /* sub crc on pkt_len */
+			 0, 0           /* ignore pkt_type field */
+			);
+
+	/* 8 packets DD mask, LSB in each 32-bit value */
+	const __m256i dd_check = _mm256_set1_epi32(1);
+
+	/* 8 packets EOP mask, second-LSB in each 32-bit value */
+	const __m256i eop_check = _mm256_slli_epi32(dd_check,
+			IAVF_RX_FLEX_DESC_STATUS0_EOF_S);
+
+	/* mask to shuffle from desc. to mbuf (4 descriptors)*/
+	const __m512i shuf_msk =
+		_mm512_set_epi8
+			(/* 1st descriptor */
+			 0xFF, 0xFF,
+			 0xFF, 0xFF,    /* rss hash parsed separately */
+			 11, 10,	/* octet 10~11, 16 bits vlan_macip */
+			 5, 4,		/* octet 4~5, 16 bits data_len */
+			 0xFF, 0xFF,	/* skip hi 16 bits pkt_len, zero out */
+			 5, 4,		/* octet 4~5, 16 bits pkt_len */
+			 0xFF, 0xFF,	/* pkt_type set as unknown */
+			 0xFF, 0xFF,	/*pkt_type set as unknown */
+			 /* 2nd descriptor */
+			 0xFF, 0xFF,
+			 0xFF, 0xFF,    /* rss hash parsed separately */
+			 11, 10,	/* octet 10~11, 16 bits vlan_macip */
+			 5, 4,		/* octet 4~5, 16 bits data_len */
+			 0xFF, 0xFF,	/* skip hi 16 bits pkt_len, zero out */
+			 5, 4,		/* octet 4~5, 16 bits pkt_len */
+			 0xFF, 0xFF,	/* pkt_type set as unknown */
+			 0xFF, 0xFF,	/*pkt_type set as unknown */
+			 /* 3rd descriptor */
+			 0xFF, 0xFF,
+			 0xFF, 0xFF,    /* rss hash parsed separately */
+			 11, 10,	/* octet 10~11, 16 bits vlan_macip */
+			 5, 4,		/* octet 4~5, 16 bits data_len */
+			 0xFF, 0xFF,	/* skip hi 16 bits pkt_len, zero out */
+			 5, 4,		/* octet 4~5, 16 bits pkt_len */
+			 0xFF, 0xFF,	/* pkt_type set as unknown */
+			 0xFF, 0xFF,	/*pkt_type set as unknown */
+			 /* 4th descriptor */
+			 0xFF, 0xFF,
+			 0xFF, 0xFF,    /* rss hash parsed separately */
+			 11, 10,	/* octet 10~11, 16 bits vlan_macip */
+			 5, 4,		/* octet 4~5, 16 bits data_len */
+			 0xFF, 0xFF,	/* skip hi 16 bits pkt_len, zero out */
+			 5, 4,		/* octet 4~5, 16 bits pkt_len */
+			 0xFF, 0xFF,	/* pkt_type set as unknown */
+			 0xFF, 0xFF	/*pkt_type set as unknown */
+			);
+	/**
+	 * compile-time check the above crc and shuffle layout is correct.
+	 * NOTE: the first field (lowest address) is given last in set_epi
+	 * calls above.
+	 */
+	RTE_BUILD_BUG_ON(offsetof(struct rte_mbuf, pkt_len) !=
+			offsetof(struct rte_mbuf, rx_descriptor_fields1) + 4);
+	RTE_BUILD_BUG_ON(offsetof(struct rte_mbuf, data_len) !=
+			offsetof(struct rte_mbuf, rx_descriptor_fields1) + 8);
+	RTE_BUILD_BUG_ON(offsetof(struct rte_mbuf, vlan_tci) !=
+			offsetof(struct rte_mbuf, rx_descriptor_fields1) + 10);
+	RTE_BUILD_BUG_ON(offsetof(struct rte_mbuf, hash) !=
+			offsetof(struct rte_mbuf, rx_descriptor_fields1) + 12);
+
+	/* Status/Error flag masks */
+	/**
+	 * mask everything except Checksum Reports, RSS indication
+	 * and VLAN indication.
+	 * bit6:4 for IP/L4 checksum errors.
+	 * bit12 is for RSS indication.
+	 * bit13 is for VLAN indication.
+	 */
+	const __m256i flags_mask =
+		 _mm256_set1_epi32((7 << 4) | (1 << 12) | (1 << 13));
+	/**
+	 * data to be shuffled by the result of the flags mask shifted by 4
+	 * bits.  This gives use the l3_l4 flags.
+	 */
+	const __m256i l3_l4_flags_shuf = _mm256_set_epi8(0, 0, 0, 0, 0, 0, 0, 0,
+			/* shift right 1 bit to make sure it not exceed 255 */
+			(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD |
+			 PKT_RX_IP_CKSUM_BAD) >> 1,
+			(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD |
+			 PKT_RX_IP_CKSUM_GOOD) >> 1,
+			(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_GOOD |
+			 PKT_RX_IP_CKSUM_BAD) >> 1,
+			(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_GOOD |
+			 PKT_RX_IP_CKSUM_GOOD) >> 1,
+			(PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1,
+			(PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_GOOD) >> 1,
+			(PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD) >> 1,
+			(PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_GOOD) >> 1,
+			/* second 128-bits */
+			0, 0, 0, 0, 0, 0, 0, 0,
+			(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD |
+			 PKT_RX_IP_CKSUM_BAD) >> 1,
+			(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD |
+			 PKT_RX_IP_CKSUM_GOOD) >> 1,
+			(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_GOOD |
+			 PKT_RX_IP_CKSUM_BAD) >> 1,
+			(PKT_RX_EIP_CKSUM_BAD | PKT_RX_L4_CKSUM_GOOD |
+			 PKT_RX_IP_CKSUM_GOOD) >> 1,
+			(PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1,
+			(PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_GOOD) >> 1,
+			(PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD) >> 1,
+			(PKT_RX_L4_CKSUM_GOOD | PKT_RX_IP_CKSUM_GOOD) >> 1);
+	const __m256i cksum_mask =
+		 _mm256_set1_epi32(PKT_RX_IP_CKSUM_GOOD | PKT_RX_IP_CKSUM_BAD |
+				   PKT_RX_L4_CKSUM_GOOD | PKT_RX_L4_CKSUM_BAD |
+				   PKT_RX_EIP_CKSUM_BAD);
+	/**
+	 * data to be shuffled by result of flag mask, shifted down 12.
+	 * If RSS(bit12)/VLAN(bit13) are set,
+	 * shuffle moves appropriate flags in place.
+	 */
+	const __m256i rss_vlan_flags_shuf = _mm256_set_epi8(0, 0, 0, 0,
+			0, 0, 0, 0,
+			0, 0, 0, 0,
+			PKT_RX_RSS_HASH | PKT_RX_VLAN | PKT_RX_VLAN_STRIPPED,
+			PKT_RX_VLAN | PKT_RX_VLAN_STRIPPED,
+			PKT_RX_RSS_HASH, 0,
+			/* end up 128-bits */
+			0, 0, 0, 0,
+			0, 0, 0, 0,
+			0, 0, 0, 0,
+			PKT_RX_RSS_HASH | PKT_RX_VLAN | PKT_RX_VLAN_STRIPPED,
+			PKT_RX_VLAN | PKT_RX_VLAN_STRIPPED,
+			PKT_RX_RSS_HASH, 0);
+
+	uint16_t i, received;
+
+	for (i = 0, received = 0; i < nb_pkts;
+	     i += IAVF_DESCS_PER_LOOP_AVX,
+	     rxdp += IAVF_DESCS_PER_LOOP_AVX) {
+		/* step 1, copy over 8 mbuf pointers to rx_pkts array */
+		_mm256_storeu_si256((void *)&rx_pkts[i],
+				    _mm256_loadu_si256((void *)&sw_ring[i]));
+#ifdef RTE_ARCH_X86_64
+		_mm256_storeu_si256
+			((void *)&rx_pkts[i + 4],
+			 _mm256_loadu_si256((void *)&sw_ring[i + 4]));
+#endif
+
+		__m512i raw_desc0_3, raw_desc4_7;
+
+		const __m128i raw_desc7 =
+			_mm_load_si128((void *)(rxdp + 7));
+		rte_compiler_barrier();
+		const __m128i raw_desc6 =
+			_mm_load_si128((void *)(rxdp + 6));
+		rte_compiler_barrier();
+		const __m128i raw_desc5 =
+			_mm_load_si128((void *)(rxdp + 5));
+		rte_compiler_barrier();
+		const __m128i raw_desc4 =
+			_mm_load_si128((void *)(rxdp + 4));
+		rte_compiler_barrier();
+		const __m128i raw_desc3 =
+			_mm_load_si128((void *)(rxdp + 3));
+		rte_compiler_barrier();
+		const __m128i raw_desc2 =
+			_mm_load_si128((void *)(rxdp + 2));
+		rte_compiler_barrier();
+		const __m128i raw_desc1 =
+			_mm_load_si128((void *)(rxdp + 1));
+		rte_compiler_barrier();
+		const __m128i raw_desc0 =
+			_mm_load_si128((void *)(rxdp + 0));
+
+		raw_desc4_7 = _mm512_broadcast_i32x4(raw_desc4);
+		raw_desc4_7 = _mm512_inserti32x4(raw_desc4_7, raw_desc5, 1);
+		raw_desc4_7 = _mm512_inserti32x4(raw_desc4_7, raw_desc6, 2);
+		raw_desc4_7 = _mm512_inserti32x4(raw_desc4_7, raw_desc7, 3);
+		raw_desc0_3 = _mm512_broadcast_i32x4(raw_desc0);
+		raw_desc0_3 = _mm512_inserti32x4(raw_desc0_3, raw_desc1, 1);
+		raw_desc0_3 = _mm512_inserti32x4(raw_desc0_3, raw_desc2, 2);
+		raw_desc0_3 = _mm512_inserti32x4(raw_desc0_3, raw_desc3, 3);
+
+		if (split_packet) {
+			int j;
+
+			for (j = 0; j < IAVF_DESCS_PER_LOOP_AVX; j++)
+				rte_mbuf_prefetch_part2(rx_pkts[i + j]);
+		}
+
+		/**
+		 * convert descriptors 4-7 into mbufs, re-arrange fields.
+		 * Then write into the mbuf.
+		 */
+		__m512i mb4_7 = _mm512_shuffle_epi8(raw_desc4_7, shuf_msk);
+
+		mb4_7 = _mm512_add_epi16(mb4_7, crc_adjust);
+		/**
+		 * to get packet types, ptype is located in bit16-25
+		 * of each 128bits
+		 */
+		const __m512i ptype_mask =
+			_mm512_set1_epi16(IAVF_RX_FLEX_DESC_PTYPE_M);
+		const __m512i ptypes4_7 =
+			_mm512_and_si512(raw_desc4_7, ptype_mask);
+		const __m256i ptypes6_7 = _mm512_extracti64x4_epi64(ptypes4_7, 1);
+		const __m256i ptypes4_5 = _mm512_extracti64x4_epi64(ptypes4_7, 0);
+		const uint16_t ptype7 = _mm256_extract_epi16(ptypes6_7, 9);
+		const uint16_t ptype6 = _mm256_extract_epi16(ptypes6_7, 1);
+		const uint16_t ptype5 = _mm256_extract_epi16(ptypes4_5, 9);
+		const uint16_t ptype4 = _mm256_extract_epi16(ptypes4_5, 1);
+
+		const __m512i ptype4_7 = _mm512_set_epi32
+			(0, 0, 0, type_table[ptype7],
+			 0, 0, 0, type_table[ptype6],
+			 0, 0, 0, type_table[ptype5],
+			 0, 0, 0, type_table[ptype4]);
+		mb4_7 = _mm512_mask_blend_epi32(0x1111, mb4_7, ptype4_7);
+
+		/**
+		 * convert descriptors 0-3 into mbufs, re-arrange fields.
+		 * Then write into the mbuf.
+		 */
+		__m512i mb0_3 = _mm512_shuffle_epi8(raw_desc0_3, shuf_msk);
+
+		mb0_3 = _mm512_add_epi16(mb0_3, crc_adjust);
+		/**
+		 * to get packet types, ptype is located in bit16-25
+		 * of each 128bits
+		 */
+		const __m512i ptypes0_3 =
+			_mm512_and_si512(raw_desc0_3, ptype_mask);
+		const __m256i ptypes2_3 = _mm512_extracti64x4_epi64(ptypes0_3, 1);
+		const __m256i ptypes0_1 = _mm512_extracti64x4_epi64(ptypes0_3, 0);
+		const uint16_t ptype3 = _mm256_extract_epi16(ptypes2_3, 9);
+		const uint16_t ptype2 = _mm256_extract_epi16(ptypes2_3, 1);
+		const uint16_t ptype1 = _mm256_extract_epi16(ptypes0_1, 9);
+		const uint16_t ptype0 = _mm256_extract_epi16(ptypes0_1, 1);
+
+		const __m512i ptype0_3 = _mm512_set_epi32
+			(0, 0, 0, type_table[ptype3],
+			 0, 0, 0, type_table[ptype2],
+			 0, 0, 0, type_table[ptype1],
+			 0, 0, 0, type_table[ptype0]);
+		mb0_3 = _mm512_mask_blend_epi32(0x1111, mb0_3, ptype0_3);
+
+		/**
+		 * use permute/extract to get status content
+		 * After the operations, the packets status flags are in the
+		 * order (hi->lo): [1, 3, 5, 7, 0, 2, 4, 6]
+		 */
+		/* merge the status bits into one register */
+		const __m512i status_permute_msk = _mm512_set_epi32
+			(0, 0, 0, 0,
+			 0, 0, 0, 0,
+			 22, 30, 6, 14,
+			 18, 26, 2, 10);
+		const __m512i raw_status0_7 = _mm512_permutex2var_epi32
+			(raw_desc4_7, status_permute_msk, raw_desc0_3);
+		__m256i status0_7 = _mm512_extracti64x4_epi64
+			(raw_status0_7, 0);
+
+		/* now do flag manipulation */
+
+		/* get only flag/error bits we want */
+		const __m256i flag_bits =
+			_mm256_and_si256(status0_7, flags_mask);
+		/**
+		 * l3_l4_error flags, shuffle, then shift to correct adjustment
+		 * of flags in flags_shuf, and finally mask out extra bits
+		 */
+		__m256i l3_l4_flags = _mm256_shuffle_epi8(l3_l4_flags_shuf,
+				_mm256_srli_epi32(flag_bits, 4));
+		l3_l4_flags = _mm256_slli_epi32(l3_l4_flags, 1);
+		l3_l4_flags = _mm256_and_si256(l3_l4_flags, cksum_mask);
+		/* set rss and vlan flags */
+		const __m256i rss_vlan_flag_bits =
+			_mm256_srli_epi32(flag_bits, 12);
+		const __m256i rss_vlan_flags =
+			_mm256_shuffle_epi8(rss_vlan_flags_shuf,
+					    rss_vlan_flag_bits);
+
+		/* merge flags */
+		__m256i mbuf_flags = _mm256_or_si256(l3_l4_flags,
+				rss_vlan_flags);
+
+		if (rxq->fdir_enabled) {
+			const __m512i fdir_permute_mask = _mm512_set_epi32
+				(0, 0, 0, 0,
+				 0, 0, 0, 0,
+				 15, 11, 7, 3,
+				 31, 27, 23, 19);
+			__m512i fdir_tmp = _mm512_permutex2var_epi32
+				(raw_desc4_7, fdir_permute_mask, raw_desc0_3);
+			const __m256i fdir_id0_7 = _mm512_extracti64x4_epi64
+				(fdir_tmp, 0);
+			const __m256i fdir_flags =
+				flex_rxd_to_fdir_flags_vec_avx512(fdir_id0_7);
+
+			/* merge with fdir_flags */
+			mbuf_flags = _mm256_or_si256(mbuf_flags, fdir_flags);
+
+			/* write to mbuf: have to use scalar store here */
+			rx_pkts[i + 0]->hash.fdir.hi =
+				_mm256_extract_epi32(fdir_id0_7, 3);
+
+			rx_pkts[i + 1]->hash.fdir.hi =
+				_mm256_extract_epi32(fdir_id0_7, 7);
+
+			rx_pkts[i + 2]->hash.fdir.hi =
+				_mm256_extract_epi32(fdir_id0_7, 2);
+
+			rx_pkts[i + 3]->hash.fdir.hi =
+				_mm256_extract_epi32(fdir_id0_7, 6);
+
+			rx_pkts[i + 4]->hash.fdir.hi =
+				_mm256_extract_epi32(fdir_id0_7, 1);
+
+			rx_pkts[i + 5]->hash.fdir.hi =
+				_mm256_extract_epi32(fdir_id0_7, 5);
+
+			rx_pkts[i + 6]->hash.fdir.hi =
+				_mm256_extract_epi32(fdir_id0_7, 0);
+
+			rx_pkts[i + 7]->hash.fdir.hi =
+				_mm256_extract_epi32(fdir_id0_7, 4);
+		} /* if() on fdir_enabled */
+
+		__m256i mb4_5 = _mm512_extracti64x4_epi64(mb4_7, 0);
+		__m256i mb6_7 = _mm512_extracti64x4_epi64(mb4_7, 1);
+		__m256i mb0_1 = _mm512_extracti64x4_epi64(mb0_3, 0);
+		__m256i mb2_3 = _mm512_extracti64x4_epi64(mb0_3, 1);
+
+#ifndef RTE_LIBRTE_IAVF_16BYTE_RX_DESC
+		/**
+		 * needs to load 2nd 16B of each desc for RSS hash parsing,
+		 * will cause performance drop to get into this context.
+		 */
+		if (rxq->vsi->adapter->eth_dev->data->dev_conf.rxmode.offloads &
+				DEV_RX_OFFLOAD_RSS_HASH) {
+			/* load bottom half of every 32B desc */
+			const __m128i raw_desc_bh7 =
+				_mm_load_si128
+					((void *)(&rxdp[7].wb.status_error1));
+			rte_compiler_barrier();
+			const __m128i raw_desc_bh6 =
+				_mm_load_si128
+					((void *)(&rxdp[6].wb.status_error1));
+			rte_compiler_barrier();
+			const __m128i raw_desc_bh5 =
+				_mm_load_si128
+					((void *)(&rxdp[5].wb.status_error1));
+			rte_compiler_barrier();
+			const __m128i raw_desc_bh4 =
+				_mm_load_si128
+					((void *)(&rxdp[4].wb.status_error1));
+			rte_compiler_barrier();
+			const __m128i raw_desc_bh3 =
+				_mm_load_si128
+					((void *)(&rxdp[3].wb.status_error1));
+			rte_compiler_barrier();
+			const __m128i raw_desc_bh2 =
+				_mm_load_si128
+					((void *)(&rxdp[2].wb.status_error1));
+			rte_compiler_barrier();
+			const __m128i raw_desc_bh1 =
+				_mm_load_si128
+					((void *)(&rxdp[1].wb.status_error1));
+			rte_compiler_barrier();
+			const __m128i raw_desc_bh0 =
+				_mm_load_si128
+					((void *)(&rxdp[0].wb.status_error1));
+
+			__m256i raw_desc_bh6_7 =
+				_mm256_inserti128_si256
+					(_mm256_castsi128_si256(raw_desc_bh6),
+					raw_desc_bh7, 1);
+			__m256i raw_desc_bh4_5 =
+				_mm256_inserti128_si256
+					(_mm256_castsi128_si256(raw_desc_bh4),
+					raw_desc_bh5, 1);
+			__m256i raw_desc_bh2_3 =
+				_mm256_inserti128_si256
+					(_mm256_castsi128_si256(raw_desc_bh2),
+					raw_desc_bh3, 1);
+			__m256i raw_desc_bh0_1 =
+				_mm256_inserti128_si256
+					(_mm256_castsi128_si256(raw_desc_bh0),
+					raw_desc_bh1, 1);
+
+			/**
+			 * to shift the 32b RSS hash value to the
+			 * highest 32b of each 128b before mask
+			 */
+			__m256i rss_hash6_7 =
+				_mm256_slli_epi64(raw_desc_bh6_7, 32);
+			__m256i rss_hash4_5 =
+				_mm256_slli_epi64(raw_desc_bh4_5, 32);
+			__m256i rss_hash2_3 =
+				_mm256_slli_epi64(raw_desc_bh2_3, 32);
+			__m256i rss_hash0_1 =
+				_mm256_slli_epi64(raw_desc_bh0_1, 32);
+
+			__m256i rss_hash_msk =
+				_mm256_set_epi32(0xFFFFFFFF, 0, 0, 0,
+						 0xFFFFFFFF, 0, 0, 0);
+
+			rss_hash6_7 = _mm256_and_si256
+					(rss_hash6_7, rss_hash_msk);
+			rss_hash4_5 = _mm256_and_si256
+					(rss_hash4_5, rss_hash_msk);
+			rss_hash2_3 = _mm256_and_si256
+					(rss_hash2_3, rss_hash_msk);
+			rss_hash0_1 = _mm256_and_si256
+					(rss_hash0_1, rss_hash_msk);
+
+			mb6_7 = _mm256_or_si256(mb6_7, rss_hash6_7);
+			mb4_5 = _mm256_or_si256(mb4_5, rss_hash4_5);
+			mb2_3 = _mm256_or_si256(mb2_3, rss_hash2_3);
+			mb0_1 = _mm256_or_si256(mb0_1, rss_hash0_1);
+		} /* if() on RSS hash parsing */
+#endif
+
+		/**
+		 * At this point, we have the 8 sets of flags in the low 16-bits
+		 * of each 32-bit value in vlan0.
+		 * We want to extract these, and merge them with the mbuf init
+		 * data so we can do a single write to the mbuf to set the flags
+		 * and all the other initialization fields. Extracting the
+		 * appropriate flags means that we have to do a shift and blend
+		 * for each mbuf before we do the write. However, we can also
+		 * add in the previously computed rx_descriptor fields to
+		 * make a single 256-bit write per mbuf
+		 */
+		/* check the structure matches expectations */
+		RTE_BUILD_BUG_ON(offsetof(struct rte_mbuf, ol_flags) !=
+				 offsetof(struct rte_mbuf, rearm_data) + 8);
+		RTE_BUILD_BUG_ON(offsetof(struct rte_mbuf, rearm_data) !=
+				 RTE_ALIGN(offsetof(struct rte_mbuf,
+						    rearm_data),
+					   16));
+		/* build up data and do writes */
+		__m256i rearm0, rearm1, rearm2, rearm3, rearm4, rearm5,
+			rearm6, rearm7;
+		rearm6 = _mm256_blend_epi32(mbuf_init,
+					    _mm256_slli_si256(mbuf_flags, 8),
+					    0x04);
+		rearm4 = _mm256_blend_epi32(mbuf_init,
+					    _mm256_slli_si256(mbuf_flags, 4),
+					    0x04);
+		rearm2 = _mm256_blend_epi32(mbuf_init, mbuf_flags, 0x04);
+		rearm0 = _mm256_blend_epi32(mbuf_init,
+					    _mm256_srli_si256(mbuf_flags, 4),
+					    0x04);
+		/* permute to add in the rx_descriptor e.g. rss fields */
+		rearm6 = _mm256_permute2f128_si256(rearm6, mb6_7, 0x20);
+		rearm4 = _mm256_permute2f128_si256(rearm4, mb4_5, 0x20);
+		rearm2 = _mm256_permute2f128_si256(rearm2, mb2_3, 0x20);
+		rearm0 = _mm256_permute2f128_si256(rearm0, mb0_1, 0x20);
+		/* write to mbuf */
+		_mm256_storeu_si256((__m256i *)&rx_pkts[i + 6]->rearm_data,
+				    rearm6);
+		_mm256_storeu_si256((__m256i *)&rx_pkts[i + 4]->rearm_data,
+				    rearm4);
+		_mm256_storeu_si256((__m256i *)&rx_pkts[i + 2]->rearm_data,
+				    rearm2);
+		_mm256_storeu_si256((__m256i *)&rx_pkts[i + 0]->rearm_data,
+				    rearm0);
+
+		/* repeat for the odd mbufs */
+		const __m256i odd_flags =
+			_mm256_castsi128_si256
+				(_mm256_extracti128_si256(mbuf_flags, 1));
+		rearm7 = _mm256_blend_epi32(mbuf_init,
+					    _mm256_slli_si256(odd_flags, 8),
+					    0x04);
+		rearm5 = _mm256_blend_epi32(mbuf_init,
+					    _mm256_slli_si256(odd_flags, 4),
+					    0x04);
+		rearm3 = _mm256_blend_epi32(mbuf_init, odd_flags, 0x04);
+		rearm1 = _mm256_blend_epi32(mbuf_init,
+					    _mm256_srli_si256(odd_flags, 4),
+					    0x04);
+		/* since odd mbufs are already in hi 128-bits use blend */
+		rearm7 = _mm256_blend_epi32(rearm7, mb6_7, 0xF0);
+		rearm5 = _mm256_blend_epi32(rearm5, mb4_5, 0xF0);
+		rearm3 = _mm256_blend_epi32(rearm3, mb2_3, 0xF0);
+		rearm1 = _mm256_blend_epi32(rearm1, mb0_1, 0xF0);
+		/* again write to mbufs */
+		_mm256_storeu_si256((__m256i *)&rx_pkts[i + 7]->rearm_data,
+				    rearm7);
+		_mm256_storeu_si256((__m256i *)&rx_pkts[i + 5]->rearm_data,
+				    rearm5);
+		_mm256_storeu_si256((__m256i *)&rx_pkts[i + 3]->rearm_data,
+				    rearm3);
+		_mm256_storeu_si256((__m256i *)&rx_pkts[i + 1]->rearm_data,
+				    rearm1);
+
+		/* extract and record EOP bit */
+		if (split_packet) {
+			const __m128i eop_mask =
+				_mm_set1_epi16(1 <<
+					       IAVF_RX_FLEX_DESC_STATUS0_EOF_S);
+			const __m256i eop_bits256 = _mm256_and_si256(status0_7,
+								     eop_check);
+			/* pack status bits into a single 128-bit register */
+			const __m128i eop_bits =
+				_mm_packus_epi32
+					(_mm256_castsi256_si128(eop_bits256),
+					 _mm256_extractf128_si256(eop_bits256,
+								  1));
+			/**
+			 * flip bits, and mask out the EOP bit, which is now
+			 * a split-packet bit i.e. !EOP, rather than EOP one.
+			 */
+			__m128i split_bits = _mm_andnot_si128(eop_bits,
+					eop_mask);
+			/**
+			 * eop bits are out of order, so we need to shuffle them
+			 * back into order again. In doing so, only use low 8
+			 * bits, which acts like another pack instruction
+			 * The original order is (hi->lo): 1,3,5,7,0,2,4,6
+			 * [Since we use epi8, the 16-bit positions are
+			 * multiplied by 2 in the eop_shuffle value.]
+			 */
+			__m128i eop_shuffle =
+				_mm_set_epi8(/* zero hi 64b */
+					     0xFF, 0xFF, 0xFF, 0xFF,
+					     0xFF, 0xFF, 0xFF, 0xFF,
+					     /* move values to lo 64b */
+					     8, 0, 10, 2,
+					     12, 4, 14, 6);
+			split_bits = _mm_shuffle_epi8(split_bits, eop_shuffle);
+			*(uint64_t *)split_packet =
+				_mm_cvtsi128_si64(split_bits);
+			split_packet += IAVF_DESCS_PER_LOOP_AVX;
+		}
+
+		/* perform dd_check */
+		status0_7 = _mm256_and_si256(status0_7, dd_check);
+		status0_7 = _mm256_packs_epi32(status0_7,
+					       _mm256_setzero_si256());
+
+		uint64_t burst = __builtin_popcountll
+					(_mm_cvtsi128_si64
+						(_mm256_extracti128_si256
+							(status0_7, 1)));
+		burst += __builtin_popcountll
+				(_mm_cvtsi128_si64
+					(_mm256_castsi256_si128(status0_7)));
+		received += burst;
+		if (burst != IAVF_DESCS_PER_LOOP_AVX)
+			break;
+	}
+
+	/* update tail pointers */
+	rxq->rx_tail += received;
+	rxq->rx_tail &= (rxq->nb_rx_desc - 1);
+	if ((rxq->rx_tail & 1) == 1 && received > 1) { /* keep aligned */
+		rxq->rx_tail--;
+		received--;
+	}
+	rxq->rxrearm_nb += received;
+	return received;
+}
+
 /**
  * Notice:
  * - nb_pkts < IAVF_DESCS_PER_LOOP, just return no packet
@@ -655,6 +1273,18 @@ 
 }
 
 /**
+ * Notice:
+ * - nb_pkts < IAVF_DESCS_PER_LOOP, just return no packet
+ */
+uint16_t
+iavf_recv_pkts_vec_avx512_flex_rxd(void *rx_queue, struct rte_mbuf **rx_pkts,
+				   uint16_t nb_pkts)
+{
+	return _iavf_recv_raw_pkts_vec_avx512_flex_rxd(rx_queue, rx_pkts,
+						       nb_pkts, NULL);
+}
+
+/**
  * vPMD receive routine that reassembles single burst of 32 scattered packets
  * Notice:
  * - nb_pkts < IAVF_DESCS_PER_LOOP, just return no packet
@@ -718,3 +1348,72 @@ 
 	return retval + iavf_recv_scattered_burst_vec_avx512(rx_queue,
 				rx_pkts + retval, nb_pkts);
 }
+
+/**
+ * vPMD receive routine that reassembles single burst of
+ * 32 scattered packets for flex RxD
+ * Notice:
+ * - nb_pkts < IAVF_DESCS_PER_LOOP, just return no packet
+ */
+static uint16_t
+iavf_recv_scattered_burst_vec_avx512_flex_rxd(void *rx_queue,
+					      struct rte_mbuf **rx_pkts,
+					      uint16_t nb_pkts)
+{
+	struct iavf_rx_queue *rxq = rx_queue;
+	uint8_t split_flags[IAVF_VPMD_RX_MAX_BURST] = {0};
+
+	/* get some new buffers */
+	uint16_t nb_bufs = _iavf_recv_raw_pkts_vec_avx512_flex_rxd(rxq,
+					rx_pkts, nb_pkts, split_flags);
+	if (nb_bufs == 0)
+		return 0;
+
+	/* happy day case, full burst + no packets to be joined */
+	const uint64_t *split_fl64 = (uint64_t *)split_flags;
+
+	if (!rxq->pkt_first_seg &&
+	    split_fl64[0] == 0 && split_fl64[1] == 0 &&
+	    split_fl64[2] == 0 && split_fl64[3] == 0)
+		return nb_bufs;
+
+	/* reassemble any packets that need reassembly*/
+	unsigned int i = 0;
+
+	if (!rxq->pkt_first_seg) {
+		/* find the first split flag, and only reassemble then*/
+		while (i < nb_bufs && !split_flags[i])
+			i++;
+		if (i == nb_bufs)
+			return nb_bufs;
+		rxq->pkt_first_seg = rx_pkts[i];
+	}
+	return i + reassemble_packets(rxq, &rx_pkts[i], nb_bufs - i,
+					     &split_flags[i]);
+}
+
+/**
+ * vPMD receive routine that reassembles scattered packets for flex RxD.
+ * Main receive routine that can handle arbitrary burst sizes
+ * Notice:
+ * - nb_pkts < IAVF_DESCS_PER_LOOP, just return no packet
+ */
+uint16_t
+iavf_recv_scattered_pkts_vec_avx512_flex_rxd(void *rx_queue,
+					     struct rte_mbuf **rx_pkts,
+					     uint16_t nb_pkts)
+{
+	uint16_t retval = 0;
+
+	while (nb_pkts > IAVF_VPMD_RX_MAX_BURST) {
+		uint16_t burst =
+			iavf_recv_scattered_burst_vec_avx512_flex_rxd
+			(rx_queue, rx_pkts + retval, IAVF_VPMD_RX_MAX_BURST);
+		retval += burst;
+		nb_pkts -= burst;
+		if (burst < IAVF_VPMD_RX_MAX_BURST)
+			return retval;
+	}
+	return retval + iavf_recv_scattered_burst_vec_avx512_flex_rxd(rx_queue,
+				rx_pkts + retval, nb_pkts);
+}