On Fri, Jun 30, 2017 at 12:23:28PM -0700, Yongseok Koh wrote: > This is to introduce more efficient Rx/Tx burst functions using SIMD > instructions. Currently it is only supported by 64bit x86 having SSE4.1. > > From functional perspective, Rx burst function is equivalent to the > existing mlx5_rx_burst() except for scatter support, which will be added > soon. Tx burst function supports multi-segment packets and offload flags > unless it is disabled by txq_flags. However, disabling those features will > bring a little higher performance. > > v2: > * Streamline redundant conditional clauses in txq_complete(). > * Remove the mempool pointer in txq->mp2mr structure. > * Fix indentation and spacing. > > Yongseok Koh (5): > net/mlx5: change indexing for Tx SW ring > net/mlx5: free buffers in bulk on Tx completion > net/mlx5: use buffer address for LKEY search > net/mlx5: select Rx/Tx callbacks when starting device > net/mlx5: add vectorized Rx/Tx burst for SSE4.1 > > drivers/net/mlx5/Makefile | 10 + > drivers/net/mlx5/mlx5_defs.h | 18 + > drivers/net/mlx5/mlx5_ethdev.c | 28 +- > drivers/net/mlx5/mlx5_mr.c | 17 +- > drivers/net/mlx5/mlx5_rxq.c | 57 +- > drivers/net/mlx5/mlx5_rxtx.c | 425 +++------- > drivers/net/mlx5/mlx5_rxtx.h | 291 ++++++- > drivers/net/mlx5/mlx5_rxtx_vec_sse.c | 1451 > ++++++++++++++++++++++++++++++++++ > drivers/net/mlx5/mlx5_trigger.c | 3 + > drivers/net/mlx5/mlx5_txq.c | 23 +- > 10 files changed, 1980 insertions(+), 343 deletions(-) > create mode 100644 drivers/net/mlx5/mlx5_rxtx_vec_sse.c > FYI: I've upstreamed a patchset to bump the minimum specs for DPDK on x86 to include SSE4.2, so that would lead to some simplification of your code, I expect. See this set: http://dpdk.org/dev/patchwork/patch/25444/
Regards, /Bruce
