X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fiavf%2Fiavf_rxtx_vec_avx2.c;h=96c05d93199192da0574ca92c9c67b0fea31754f;hb=689a56742a560cc1baff7255a57efcdfdb789f87;hp=2587083d87f3519c511f9de3d60b033d94c4ae1f;hpb=040b44551f77b1173f2c1c830856e1e1f771da26;p=dpdk.git diff --git a/drivers/net/iavf/iavf_rxtx_vec_avx2.c b/drivers/net/iavf/iavf_rxtx_vec_avx2.c index 2587083d87..96c05d9319 100644 --- a/drivers/net/iavf/iavf_rxtx_vec_avx2.c +++ b/drivers/net/iavf/iavf_rxtx_vec_avx2.c @@ -4,132 +4,16 @@ #include "iavf_rxtx_vec_common.h" -#include +#include #ifndef __INTEL_COMPILER #pragma GCC diagnostic ignored "-Wcast-qual" #endif -static inline void +static __rte_always_inline void iavf_rxq_rearm(struct iavf_rx_queue *rxq) { - int i; - uint16_t rx_id; - volatile union iavf_rx_desc *rxdp; - struct rte_mbuf **rxp = &rxq->sw_ring[rxq->rxrearm_start]; - - rxdp = rxq->rx_ring + rxq->rxrearm_start; - - /* Pull 'n' more MBUFs into the software ring */ - if (rte_mempool_get_bulk(rxq->mp, - (void *)rxp, - IAVF_RXQ_REARM_THRESH) < 0) { - if (rxq->rxrearm_nb + IAVF_RXQ_REARM_THRESH >= - rxq->nb_rx_desc) { - __m128i dma_addr0; - - dma_addr0 = _mm_setzero_si128(); - for (i = 0; i < IAVF_VPMD_DESCS_PER_LOOP; i++) { - rxp[i] = &rxq->fake_mbuf; - _mm_store_si128((__m128i *)&rxdp[i].read, - dma_addr0); - } - } - rte_eth_devices[rxq->port_id].data->rx_mbuf_alloc_failed += - IAVF_RXQ_REARM_THRESH; - return; - } - -#ifndef RTE_LIBRTE_IAVF_16BYTE_RX_DESC - struct rte_mbuf *mb0, *mb1; - __m128i dma_addr0, dma_addr1; - __m128i hdr_room = _mm_set_epi64x(RTE_PKTMBUF_HEADROOM, - RTE_PKTMBUF_HEADROOM); - /* Initialize the mbufs in vector, process 2 mbufs in one loop */ - for (i = 0; i < IAVF_RXQ_REARM_THRESH; i += 2, rxp += 2) { - __m128i vaddr0, vaddr1; - - mb0 = rxp[0]; - mb1 = rxp[1]; - - /* load buf_addr(lo 64bit) and buf_physaddr(hi 64bit) */ - RTE_BUILD_BUG_ON(offsetof(struct rte_mbuf, buf_physaddr) != - offsetof(struct rte_mbuf, buf_addr) + 8); - vaddr0 = _mm_loadu_si128((__m128i *)&mb0->buf_addr); - vaddr1 = _mm_loadu_si128((__m128i *)&mb1->buf_addr); - - /* convert pa to dma_addr hdr/data */ - dma_addr0 = _mm_unpackhi_epi64(vaddr0, vaddr0); - dma_addr1 = _mm_unpackhi_epi64(vaddr1, vaddr1); - - /* add headroom to pa values */ - dma_addr0 = _mm_add_epi64(dma_addr0, hdr_room); - dma_addr1 = _mm_add_epi64(dma_addr1, hdr_room); - - /* flush desc with pa dma_addr */ - _mm_store_si128((__m128i *)&rxdp++->read, dma_addr0); - _mm_store_si128((__m128i *)&rxdp++->read, dma_addr1); - } -#else - struct rte_mbuf *mb0, *mb1, *mb2, *mb3; - __m256i dma_addr0_1, dma_addr2_3; - __m256i hdr_room = _mm256_set1_epi64x(RTE_PKTMBUF_HEADROOM); - /* Initialize the mbufs in vector, process 4 mbufs in one loop */ - for (i = 0; i < IAVF_RXQ_REARM_THRESH; - i += 4, rxp += 4, rxdp += 4) { - __m128i vaddr0, vaddr1, vaddr2, vaddr3; - __m256i vaddr0_1, vaddr2_3; - - mb0 = rxp[0]; - mb1 = rxp[1]; - mb2 = rxp[2]; - mb3 = rxp[3]; - - /* load buf_addr(lo 64bit) and buf_physaddr(hi 64bit) */ - RTE_BUILD_BUG_ON(offsetof(struct rte_mbuf, buf_physaddr) != - offsetof(struct rte_mbuf, buf_addr) + 8); - vaddr0 = _mm_loadu_si128((__m128i *)&mb0->buf_addr); - vaddr1 = _mm_loadu_si128((__m128i *)&mb1->buf_addr); - vaddr2 = _mm_loadu_si128((__m128i *)&mb2->buf_addr); - vaddr3 = _mm_loadu_si128((__m128i *)&mb3->buf_addr); - - /** - * merge 0 & 1, by casting 0 to 256-bit and inserting 1 - * into the high lanes. Similarly for 2 & 3 - */ - vaddr0_1 = - _mm256_inserti128_si256(_mm256_castsi128_si256(vaddr0), - vaddr1, 1); - vaddr2_3 = - _mm256_inserti128_si256(_mm256_castsi128_si256(vaddr2), - vaddr3, 1); - - /* convert pa to dma_addr hdr/data */ - dma_addr0_1 = _mm256_unpackhi_epi64(vaddr0_1, vaddr0_1); - dma_addr2_3 = _mm256_unpackhi_epi64(vaddr2_3, vaddr2_3); - - /* add headroom to pa values */ - dma_addr0_1 = _mm256_add_epi64(dma_addr0_1, hdr_room); - dma_addr2_3 = _mm256_add_epi64(dma_addr2_3, hdr_room); - - /* flush desc with pa dma_addr */ - _mm256_store_si256((__m256i *)&rxdp->read, dma_addr0_1); - _mm256_store_si256((__m256i *)&(rxdp + 2)->read, dma_addr2_3); - } - -#endif - - rxq->rxrearm_start += IAVF_RXQ_REARM_THRESH; - if (rxq->rxrearm_start >= rxq->nb_rx_desc) - rxq->rxrearm_start = 0; - - rxq->rxrearm_nb -= IAVF_RXQ_REARM_THRESH; - - rx_id = (uint16_t)((rxq->rxrearm_start == 0) ? - (rxq->nb_rx_desc - 1) : (rxq->rxrearm_start - 1)); - - /* Update the tail pointer on the NIC */ - IAVF_PCI_REG_WRITE(rxq->qrx_tail, rx_id); + return iavf_rxq_rearm_common(rxq, false); } #define PKTLEN_SHIFT 10 @@ -264,24 +148,24 @@ _iavf_recv_raw_pkts_vec_avx2(struct iavf_rx_queue *rxq, */ 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_OUTER_IP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1, - (PKT_RX_IP_CKSUM_GOOD | PKT_RX_EIP_CKSUM_BAD | + (PKT_RX_IP_CKSUM_GOOD | PKT_RX_OUTER_IP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD) >> 1, - (PKT_RX_EIP_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1, - (PKT_RX_IP_CKSUM_GOOD | PKT_RX_EIP_CKSUM_BAD) >> 1, + (PKT_RX_OUTER_IP_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1, + (PKT_RX_IP_CKSUM_GOOD | PKT_RX_OUTER_IP_CKSUM_BAD) >> 1, (PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1, (PKT_RX_IP_CKSUM_GOOD | PKT_RX_L4_CKSUM_BAD) >> 1, PKT_RX_IP_CKSUM_BAD >> 1, (PKT_RX_IP_CKSUM_GOOD | PKT_RX_L4_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_OUTER_IP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1, - (PKT_RX_IP_CKSUM_GOOD | PKT_RX_EIP_CKSUM_BAD | + (PKT_RX_IP_CKSUM_GOOD | PKT_RX_OUTER_IP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD) >> 1, - (PKT_RX_EIP_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1, - (PKT_RX_IP_CKSUM_GOOD | PKT_RX_EIP_CKSUM_BAD) >> 1, + (PKT_RX_OUTER_IP_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1, + (PKT_RX_IP_CKSUM_GOOD | PKT_RX_OUTER_IP_CKSUM_BAD) >> 1, (PKT_RX_L4_CKSUM_BAD | PKT_RX_IP_CKSUM_BAD) >> 1, (PKT_RX_IP_CKSUM_GOOD | PKT_RX_L4_CKSUM_BAD) >> 1, PKT_RX_IP_CKSUM_BAD >> 1, @@ -290,7 +174,7 @@ _iavf_recv_raw_pkts_vec_avx2(struct iavf_rx_queue *rxq, 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); + PKT_RX_OUTER_IP_CKSUM_BAD); RTE_SET_USED(avx_aligned); /* for 32B descriptors we don't use this */ @@ -614,6 +498,719 @@ _iavf_recv_raw_pkts_vec_avx2(struct iavf_rx_queue *rxq, return received; } +static inline __m256i +flex_rxd_to_fdir_flags_vec_avx2(const __m256i fdir_id0_7) +{ +#define FDID_MIS_MAGIC 0xFFFFFFFF + RTE_BUILD_BUG_ON(PKT_RX_FDIR != (1 << 2)); + RTE_BUILD_BUG_ON(PKT_RX_FDIR_ID != (1 << 13)); + const __m256i pkt_fdir_bit = _mm256_set1_epi32(PKT_RX_FDIR | + PKT_RX_FDIR_ID); + /* desc->flow_id field == 0xFFFFFFFF means fdir mismatch */ + const __m256i fdir_mis_mask = _mm256_set1_epi32(FDID_MIS_MAGIC); + __m256i fdir_mask = _mm256_cmpeq_epi32(fdir_id0_7, + fdir_mis_mask); + /* this XOR op results to bit-reverse the fdir_mask */ + fdir_mask = _mm256_xor_si256(fdir_mask, fdir_mis_mask); + const __m256i fdir_flags = _mm256_and_si256(fdir_mask, pkt_fdir_bit); + + return fdir_flags; +} + +static inline uint16_t +_iavf_recv_raw_pkts_vec_avx2_flex_rxd(struct iavf_rx_queue *rxq, + struct rte_mbuf **rx_pkts, + uint16_t nb_pkts, uint8_t *split_packet) +{ +#define IAVF_DESCS_PER_LOOP_AVX 8 + + 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 __m256i crc_adjust = + _mm256_set_epi16 + (/* first 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 */ + /* second 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 (2 descriptors)*/ + const __m256i shuf_msk = + _mm256_set_epi8 + (/* first 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 */ + /* second 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_OUTER_IP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD | + PKT_RX_IP_CKSUM_BAD) >> 1, + (PKT_RX_OUTER_IP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD | + PKT_RX_IP_CKSUM_GOOD) >> 1, + (PKT_RX_OUTER_IP_CKSUM_BAD | PKT_RX_L4_CKSUM_GOOD | + PKT_RX_IP_CKSUM_BAD) >> 1, + (PKT_RX_OUTER_IP_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_OUTER_IP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD | + PKT_RX_IP_CKSUM_BAD) >> 1, + (PKT_RX_OUTER_IP_CKSUM_BAD | PKT_RX_L4_CKSUM_BAD | + PKT_RX_IP_CKSUM_GOOD) >> 1, + (PKT_RX_OUTER_IP_CKSUM_BAD | PKT_RX_L4_CKSUM_GOOD | + PKT_RX_IP_CKSUM_BAD) >> 1, + (PKT_RX_OUTER_IP_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_OUTER_IP_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_flags_shuf = _mm256_set_epi8(0, 0, 0, 0, + 0, 0, 0, 0, + 0, 0, 0, 0, + PKT_RX_RSS_HASH, 0, + 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, 0, + PKT_RX_RSS_HASH, 0); + + const __m256i vlan_flags_shuf = _mm256_set_epi8(0, 0, 0, 0, + 0, 0, 0, 0, + 0, 0, 0, 0, + PKT_RX_VLAN | PKT_RX_VLAN_STRIPPED, + PKT_RX_VLAN | PKT_RX_VLAN_STRIPPED, + 0, 0, + /* end up 128-bits */ + 0, 0, 0, 0, + 0, 0, 0, 0, + 0, 0, 0, 0, + PKT_RX_VLAN | PKT_RX_VLAN_STRIPPED, + PKT_RX_VLAN | PKT_RX_VLAN_STRIPPED, + 0, 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 + + __m256i raw_desc0_1, raw_desc2_3, raw_desc4_5, raw_desc6_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_desc6_7 = + _mm256_inserti128_si256 + (_mm256_castsi128_si256(raw_desc6), + raw_desc7, 1); + raw_desc4_5 = + _mm256_inserti128_si256 + (_mm256_castsi128_si256(raw_desc4), + raw_desc5, 1); + raw_desc2_3 = + _mm256_inserti128_si256 + (_mm256_castsi128_si256(raw_desc2), + raw_desc3, 1); + raw_desc0_1 = + _mm256_inserti128_si256 + (_mm256_castsi128_si256(raw_desc0), + raw_desc1, 1); + + 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. + */ + __m256i mb6_7 = _mm256_shuffle_epi8(raw_desc6_7, shuf_msk); + __m256i mb4_5 = _mm256_shuffle_epi8(raw_desc4_5, shuf_msk); + + mb6_7 = _mm256_add_epi16(mb6_7, crc_adjust); + mb4_5 = _mm256_add_epi16(mb4_5, crc_adjust); + /** + * to get packet types, ptype is located in bit16-25 + * of each 128bits + */ + const __m256i ptype_mask = + _mm256_set1_epi16(IAVF_RX_FLEX_DESC_PTYPE_M); + const __m256i ptypes6_7 = + _mm256_and_si256(raw_desc6_7, ptype_mask); + const __m256i ptypes4_5 = + _mm256_and_si256(raw_desc4_5, ptype_mask); + 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); + + mb6_7 = _mm256_insert_epi32(mb6_7, type_table[ptype7], 4); + mb6_7 = _mm256_insert_epi32(mb6_7, type_table[ptype6], 0); + mb4_5 = _mm256_insert_epi32(mb4_5, type_table[ptype5], 4); + mb4_5 = _mm256_insert_epi32(mb4_5, type_table[ptype4], 0); + /* merge the status bits into one register */ + const __m256i status4_7 = _mm256_unpackhi_epi32(raw_desc6_7, + raw_desc4_5); + + /** + * convert descriptors 0-3 into mbufs, re-arrange fields. + * Then write into the mbuf. + */ + __m256i mb2_3 = _mm256_shuffle_epi8(raw_desc2_3, shuf_msk); + __m256i mb0_1 = _mm256_shuffle_epi8(raw_desc0_1, shuf_msk); + + mb2_3 = _mm256_add_epi16(mb2_3, crc_adjust); + mb0_1 = _mm256_add_epi16(mb0_1, crc_adjust); + /** + * to get packet types, ptype is located in bit16-25 + * of each 128bits + */ + const __m256i ptypes2_3 = + _mm256_and_si256(raw_desc2_3, ptype_mask); + const __m256i ptypes0_1 = + _mm256_and_si256(raw_desc0_1, ptype_mask); + 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); + + mb2_3 = _mm256_insert_epi32(mb2_3, type_table[ptype3], 4); + mb2_3 = _mm256_insert_epi32(mb2_3, type_table[ptype2], 0); + mb0_1 = _mm256_insert_epi32(mb0_1, type_table[ptype1], 4); + mb0_1 = _mm256_insert_epi32(mb0_1, type_table[ptype0], 0); + /* merge the status bits into one register */ + const __m256i status0_3 = _mm256_unpackhi_epi32(raw_desc2_3, + raw_desc0_1); + + /** + * take the two sets of status bits and merge to one + * After merge, the packets status flags are in the + * order (hi->lo): [1, 3, 5, 7, 0, 2, 4, 6] + */ + __m256i status0_7 = _mm256_unpacklo_epi64(status4_7, + status0_3); + + /* 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_flags = + _mm256_shuffle_epi8(rss_flags_shuf, + rss_vlan_flag_bits); + + __m256i vlan_flags = _mm256_setzero_si256(); + + if (rxq->rx_flags == IAVF_RX_FLAGS_VLAN_TAG_LOC_L2TAG1) + vlan_flags = + _mm256_shuffle_epi8(vlan_flags_shuf, + rss_vlan_flag_bits); + + const __m256i rss_vlan_flags = + _mm256_or_si256(rss_flags, vlan_flags); + + /* merge flags */ + __m256i mbuf_flags = _mm256_or_si256(l3_l4_flags, + rss_vlan_flags); + + if (rxq->fdir_enabled) { + const __m256i fdir_id4_7 = + _mm256_unpackhi_epi32(raw_desc6_7, raw_desc4_5); + + const __m256i fdir_id0_3 = + _mm256_unpackhi_epi32(raw_desc2_3, raw_desc0_1); + + const __m256i fdir_id0_7 = + _mm256_unpackhi_epi64(fdir_id4_7, fdir_id0_3); + + const __m256i fdir_flags = + flex_rxd_to_fdir_flags_vec_avx2(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 */ + +#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 || + rxq->rx_flags & IAVF_RX_FLAGS_VLAN_TAG_LOC_L2TAG2_2) { + /* 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); + + if (rxq->vsi->adapter->eth_dev->data->dev_conf.rxmode.offloads & + DEV_RX_OFFLOAD_RSS_HASH) { + /** + * 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); + + const __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 (rxq->rx_flags & IAVF_RX_FLAGS_VLAN_TAG_LOC_L2TAG2_2) { + /* merge the status/error-1 bits into one register */ + const __m256i status1_4_7 = + _mm256_unpacklo_epi32(raw_desc_bh6_7, + raw_desc_bh4_5); + const __m256i status1_0_3 = + _mm256_unpacklo_epi32(raw_desc_bh2_3, + raw_desc_bh0_1); + + const __m256i status1_0_7 = + _mm256_unpacklo_epi64(status1_4_7, + status1_0_3); + + const __m256i l2tag2p_flag_mask = + _mm256_set1_epi32 + (1 << IAVF_RX_FLEX_DESC_STATUS1_L2TAG2P_S); + + __m256i l2tag2p_flag_bits = + _mm256_and_si256 + (status1_0_7, l2tag2p_flag_mask); + + l2tag2p_flag_bits = + _mm256_srli_epi32(l2tag2p_flag_bits, + IAVF_RX_FLEX_DESC_STATUS1_L2TAG2P_S); + + const __m256i l2tag2_flags_shuf = + _mm256_set_epi8(0, 0, 0, 0, + 0, 0, 0, 0, + 0, 0, 0, 0, + 0, 0, 0, 0, + /* end up 128-bits */ + 0, 0, 0, 0, + 0, 0, 0, 0, + 0, 0, 0, 0, + 0, 0, + PKT_RX_VLAN | + PKT_RX_VLAN_STRIPPED, + 0); + + vlan_flags = + _mm256_shuffle_epi8(l2tag2_flags_shuf, + l2tag2p_flag_bits); + + /* merge with vlan_flags */ + mbuf_flags = _mm256_or_si256 + (mbuf_flags, vlan_flags); + + /* L2TAG2_2 */ + __m256i vlan_tci6_7 = + _mm256_slli_si256(raw_desc_bh6_7, 4); + __m256i vlan_tci4_5 = + _mm256_slli_si256(raw_desc_bh4_5, 4); + __m256i vlan_tci2_3 = + _mm256_slli_si256(raw_desc_bh2_3, 4); + __m256i vlan_tci0_1 = + _mm256_slli_si256(raw_desc_bh0_1, 4); + + const __m256i vlan_tci_msk = + _mm256_set_epi32(0, 0xFFFF0000, 0, 0, + 0, 0xFFFF0000, 0, 0); + + vlan_tci6_7 = _mm256_and_si256 + (vlan_tci6_7, vlan_tci_msk); + vlan_tci4_5 = _mm256_and_si256 + (vlan_tci4_5, vlan_tci_msk); + vlan_tci2_3 = _mm256_and_si256 + (vlan_tci2_3, vlan_tci_msk); + vlan_tci0_1 = _mm256_and_si256 + (vlan_tci0_1, vlan_tci_msk); + + mb6_7 = _mm256_or_si256(mb6_7, vlan_tci6_7); + mb4_5 = _mm256_or_si256(mb4_5, vlan_tci4_5); + mb2_3 = _mm256_or_si256(mb2_3, vlan_tci2_3); + mb0_1 = _mm256_or_si256(mb0_1, vlan_tci0_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 avx2 aligned */ + rxq->rx_tail--; + received--; + } + rxq->rxrearm_nb += received; + return received; +} + /** * Notice: * - nb_pkts < IAVF_DESCS_PER_LOOP, just return no packet @@ -625,6 +1222,18 @@ iavf_recv_pkts_vec_avx2(void *rx_queue, struct rte_mbuf **rx_pkts, return _iavf_recv_raw_pkts_vec_avx2(rx_queue, rx_pkts, nb_pkts, NULL); } +/** + * Notice: + * - nb_pkts < IAVF_DESCS_PER_LOOP, just return no packet + */ +uint16_t +iavf_recv_pkts_vec_avx2_flex_rxd(void *rx_queue, struct rte_mbuf **rx_pkts, + uint16_t nb_pkts) +{ + return _iavf_recv_raw_pkts_vec_avx2_flex_rxd(rx_queue, rx_pkts, + nb_pkts, NULL); +} + /** * vPMD receive routine that reassembles single burst of 32 scattered packets * Notice: @@ -690,6 +1299,75 @@ iavf_recv_scattered_pkts_vec_avx2(void *rx_queue, struct rte_mbuf **rx_pkts, 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_avx2_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_avx2_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_avx2_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_avx2_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_avx2_flex_rxd(rx_queue, + rx_pkts + retval, nb_pkts); +} + static inline void iavf_vtx1(volatile struct iavf_tx_desc *txdp, struct rte_mbuf *pkt, uint64_t flags) @@ -700,7 +1378,7 @@ iavf_vtx1(volatile struct iavf_tx_desc *txdp, ((uint64_t)pkt->data_len << IAVF_TXD_QW1_TX_BUF_SZ_SHIFT)); __m128i descriptor = _mm_set_epi64x(high_qw, - pkt->buf_physaddr + pkt->data_off); + pkt->buf_iova + pkt->data_off); _mm_store_si128((__m128i *)txdp, descriptor); } @@ -739,15 +1417,15 @@ iavf_vtx(volatile struct iavf_tx_desc *txdp, __m256i desc2_3 = _mm256_set_epi64x (hi_qw3, - pkt[3]->buf_physaddr + pkt[3]->data_off, + pkt[3]->buf_iova + pkt[3]->data_off, hi_qw2, - pkt[2]->buf_physaddr + pkt[2]->data_off); + pkt[2]->buf_iova + pkt[2]->data_off); __m256i desc0_1 = _mm256_set_epi64x (hi_qw1, - pkt[1]->buf_physaddr + pkt[1]->data_off, + pkt[1]->buf_iova + pkt[1]->data_off, hi_qw0, - pkt[0]->buf_physaddr + pkt[0]->data_off); + pkt[0]->buf_iova + pkt[0]->data_off); _mm256_store_si256((void *)(txdp + 2), desc2_3); _mm256_store_si256((void *)txdp, desc0_1); } @@ -822,7 +1500,7 @@ iavf_xmit_fixed_burst_vec_avx2(void *tx_queue, struct rte_mbuf **tx_pkts, txq->tx_tail = tx_id; - IAVF_PCI_REG_WRITE(txq->qtx_tail, txq->tx_tail); + IAVF_PCI_REG_WC_WRITE(txq->qtx_tail, txq->tx_tail); return nb_pkts; }