net/mlx5: prefetch CQEs for a faster decompression

[dpdk.git] / drivers / net / mlx5 / mlx5_rxtx_vec_sse.h

diff --git a/drivers/net/mlx5/mlx5_rxtx_vec_sse.h b/drivers/net/mlx5/mlx5_rxtx_vec_sse.h
index a4086df..4b711f0 100644 (file)
--- a/drivers/net/mlx5/mlx5_rxtx_vec_sse.h
+++ b/drivers/net/mlx5/mlx5_rxtx_vec_sse.h
@@ -132,8 +132,10 @@ rxq_cq_decompress_v(struct mlx5_rxq_data *rxq, volatile struct mlx5_cqe *cq,
                __m128i byte_cnt, invalid_mask;
 #endif
 
-               if (!(pos & 0x7) && pos + 8 < mcqe_n)
-                       rte_prefetch0((void *)(cq + pos + 8));
+               for (i = 0; i < MLX5_VPMD_DESCS_PER_LOOP; ++i)
+                       if (likely(pos + i < mcqe_n))
+                               rte_prefetch0((void *)(cq + pos + i));
+
                /* A.1 load mCQEs into a 128bit register. */
                mcqe1 = _mm_loadu_si128((__m128i *)&mcq[pos % 8]);
                mcqe2 = _mm_loadu_si128((__m128i *)&mcq[pos % 8 + 2]);