X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fmlx5%2Fmlx5_rxtx_vec_neon.h;h=af924b70924b6fee85f89a27f97f2128cca56da4;hb=4ffab7b9e199d6253e0b7fdcf4bef137930d0914;hp=ecafbf8004e2b0e0705f966bb6ff51dca8edbef5;hpb=0c55591588e017f1ef2c1ac845a1fa23ea1e19fb;p=dpdk.git diff --git a/drivers/net/mlx5/mlx5_rxtx_vec_neon.h b/drivers/net/mlx5/mlx5_rxtx_vec_neon.h index ecafbf8004..af924b7092 100644 --- a/drivers/net/mlx5/mlx5_rxtx_vec_neon.h +++ b/drivers/net/mlx5/mlx5_rxtx_vec_neon.h @@ -145,6 +145,7 @@ rxq_cq_decompress_v(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cq, -1UL << ((mcqe_n - pos) * sizeof(uint16_t) * 8) : 0); #endif + for (i = 0; i < MLX5_VPMD_DESCS_PER_LOOP; ++i) if (likely(pos + i < mcqe_n)) rte_prefetch0((void *)(cq + pos + i)); @@ -227,6 +228,8 @@ rxq_cq_decompress_v(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cq, pos += MLX5_VPMD_DESCS_PER_LOOP; /* Move to next CQE and invalidate consumed CQEs. */ if (!(pos & 0x7) && pos < mcqe_n) { + if (pos + 8 < mcqe_n) + rte_prefetch0((void *)(cq + pos + 8)); mcq = (void *)&(cq + pos)->pkt_info; for (i = 0; i < 8; ++i) cq[inv++].op_own = MLX5_CQE_INVALIDATE; @@ -378,13 +381,15 @@ rxq_cq_to_ptype_oflags_v(struct mlx5_rxq_data *rxq, * @param[out] err * Pointer to a flag. Set non-zero value if pkts array has at least one error * packet to handle. + * @param[out] no_cq + * Pointer to a boolean. Set true if no new CQE seen. * * @return * Number of packets received including errors (<= pkts_n). */ static inline uint16_t rxq_burst_v(struct mlx5_rxq_data *rxq, struct rte_mbuf **pkts, uint16_t pkts_n, - uint64_t *err) + uint64_t *err, bool *no_cq) { const uint16_t q_n = 1 << rxq->cqe_n; const uint16_t q_mask = q_n - 1; @@ -485,8 +490,10 @@ rxq_burst_v(struct mlx5_rxq_data *rxq, struct rte_mbuf **pkts, uint16_t pkts_n, /* Not to cross queue end. */ pkts_n = RTE_MIN(pkts_n, q_n - elts_idx); pkts_n = RTE_MIN(pkts_n, q_n - cq_idx); - if (!pkts_n) + if (!pkts_n) { + *no_cq = !rcvd_pkt; return rcvd_pkt; + } /* At this point, there shouldn't be any remained packets. */ MLX5_ASSERT(rxq->decompressed == 0); /* @@ -690,37 +697,59 @@ rxq_burst_v(struct mlx5_rxq_data *rxq, struct rte_mbuf **pkts, uint16_t pkts_n, rxq_cq_to_ptype_oflags_v(rxq, ptype_info, flow_tag, opcode, &elts[pos]); if (rxq->hw_timestamp) { - elts[pos]->timestamp = - rte_be_to_cpu_64( - container_of(p0, struct mlx5_cqe, - pkt_info)->timestamp); - elts[pos + 1]->timestamp = - rte_be_to_cpu_64( - container_of(p1, struct mlx5_cqe, - pkt_info)->timestamp); - elts[pos + 2]->timestamp = - rte_be_to_cpu_64( - container_of(p2, struct mlx5_cqe, - pkt_info)->timestamp); - elts[pos + 3]->timestamp = - rte_be_to_cpu_64( - container_of(p3, struct mlx5_cqe, - pkt_info)->timestamp); + if (rxq->rt_timestamp) { + struct mlx5_dev_ctx_shared *sh = rxq->sh; + uint64_t ts; + + ts = rte_be_to_cpu_64 + (container_of(p0, struct mlx5_cqe, + pkt_info)->timestamp); + elts[pos]->timestamp = + mlx5_txpp_convert_rx_ts(sh, ts); + ts = rte_be_to_cpu_64 + (container_of(p1, struct mlx5_cqe, + pkt_info)->timestamp); + elts[pos + 1]->timestamp = + mlx5_txpp_convert_rx_ts(sh, ts); + ts = rte_be_to_cpu_64 + (container_of(p2, struct mlx5_cqe, + pkt_info)->timestamp); + elts[pos + 2]->timestamp = + mlx5_txpp_convert_rx_ts(sh, ts); + ts = rte_be_to_cpu_64 + (container_of(p3, struct mlx5_cqe, + pkt_info)->timestamp); + elts[pos + 3]->timestamp = + mlx5_txpp_convert_rx_ts(sh, ts); + } else { + elts[pos]->timestamp = rte_be_to_cpu_64 + (container_of(p0, struct mlx5_cqe, + pkt_info)->timestamp); + elts[pos + 1]->timestamp = rte_be_to_cpu_64 + (container_of(p1, struct mlx5_cqe, + pkt_info)->timestamp); + elts[pos + 2]->timestamp = rte_be_to_cpu_64 + (container_of(p2, struct mlx5_cqe, + pkt_info)->timestamp); + elts[pos + 3]->timestamp = rte_be_to_cpu_64 + (container_of(p3, struct mlx5_cqe, + pkt_info)->timestamp); + } } - if (!!rxq->flow_meta_mask) { + if (rxq->dynf_meta) { /* This code is subject for futher optimization. */ int32_t offs = rxq->flow_meta_offset; *RTE_MBUF_DYNFIELD(pkts[pos], offs, uint32_t *) = container_of(p0, struct mlx5_cqe, pkt_info)->flow_table_metadata; - *RTE_MBUF_DYNFIELD(pkts[pos], offs, uint32_t *) = + *RTE_MBUF_DYNFIELD(pkts[pos + 1], offs, uint32_t *) = container_of(p1, struct mlx5_cqe, pkt_info)->flow_table_metadata; - *RTE_MBUF_DYNFIELD(pkts[pos], offs, uint32_t *) = + *RTE_MBUF_DYNFIELD(pkts[pos + 2], offs, uint32_t *) = container_of(p2, struct mlx5_cqe, pkt_info)->flow_table_metadata; - *RTE_MBUF_DYNFIELD(pkts[pos], offs, uint32_t *) = + *RTE_MBUF_DYNFIELD(pkts[pos + 3], offs, uint32_t *) = container_of(p3, struct mlx5_cqe, pkt_info)->flow_table_metadata; if (*RTE_MBUF_DYNFIELD(pkts[pos], offs, uint32_t *)) @@ -745,8 +774,10 @@ rxq_burst_v(struct mlx5_rxq_data *rxq, struct rte_mbuf **pkts, uint16_t pkts_n, break; } /* If no new CQE seen, return without updating cq_db. */ - if (unlikely(!nocmp_n && comp_idx == MLX5_VPMD_DESCS_PER_LOOP)) + if (unlikely(!nocmp_n && comp_idx == MLX5_VPMD_DESCS_PER_LOOP)) { + *no_cq = true; return rcvd_pkt; + } /* Update the consumer indexes for non-compressed CQEs. */ MLX5_ASSERT(nocmp_n <= pkts_n); rxq->cq_ci += nocmp_n; @@ -774,6 +805,7 @@ rxq_burst_v(struct mlx5_rxq_data *rxq, struct rte_mbuf **pkts, uint16_t pkts_n, } rte_cio_wmb(); *rxq->cq_db = rte_cpu_to_be_32(rxq->cq_ci); + *no_cq = !rcvd_pkt; return rcvd_pkt; }