net/txgbe: add Rx and Tx init

[dpdk.git] / drivers / net / virtio / virtio_rxtx_packed_avx.c

diff --git a/drivers/net/virtio/virtio_rxtx_packed_avx.c b/drivers/net/virtio/virtio_rxtx_packed_avx.c
index d130d68..9bc6271 100644 (file)
--- a/drivers/net/virtio/virtio_rxtx_packed_avx.c
+++ b/drivers/net/virtio/virtio_rxtx_packed_avx.c
@@ -207,19 +207,26 @@ virtqueue_enqueue_single_packed_vec(struct virtnet_tx *txvq,
        struct virtqueue *vq = txvq->vq;
        struct virtio_hw *hw = vq->hw;
        uint16_t hdr_size = hw->vtnet_hdr_size;
-       uint16_t slots, can_push;
+       uint16_t slots, can_push = 0, use_indirect = 0;
        int16_t need;
 
+       /* optimize ring usage */
+       if ((vtpci_with_feature(hw, VIRTIO_F_ANY_LAYOUT) ||
+             vtpci_with_feature(hw, VIRTIO_F_VERSION_1)) &&
+           rte_mbuf_refcnt_read(txm) == 1 &&
+           RTE_MBUF_DIRECT(txm) &&
+           txm->nb_segs == 1 &&
+           rte_pktmbuf_headroom(txm) >= hdr_size)
+               can_push = 1;
+       else if (vtpci_with_feature(hw, VIRTIO_RING_F_INDIRECT_DESC) &&
+                txm->nb_segs < VIRTIO_MAX_TX_INDIRECT)
+               use_indirect = 1;
        /* How many main ring entries are needed to this Tx?
+        * indirect   => 1
         * any_layout => number of segments
         * default    => number of segments + 1
         */
-       can_push = rte_mbuf_refcnt_read(txm) == 1 &&
-                  RTE_MBUF_DIRECT(txm) &&
-                  txm->nb_segs == 1 &&
-                  rte_pktmbuf_headroom(txm) >= hdr_size;
-
-       slots = txm->nb_segs + !can_push;
+       slots = use_indirect ? 1 : (txm->nb_segs + !can_push);
        need = slots - vq->vq_free_cnt;
 
        /* Positive value indicates it need free vring descriptors */
@@ -234,7 +241,8 @@ virtqueue_enqueue_single_packed_vec(struct virtnet_tx *txvq,
        }
 
        /* Enqueue Packet buffers */
-       virtqueue_enqueue_xmit_packed(txvq, txm, slots, can_push, 1);
+       virtqueue_enqueue_xmit_packed(txvq, txm, slots, use_indirect,
+                               can_push, 1);
 
        txvq->stats.bytes += txm->pkt_len;
        return 0;
@@ -328,9 +336,10 @@ virtio_vec_rx_offload(struct rte_mbuf *m, struct virtio_net_hdr *hdr)
                         */
                        uint16_t csum = 0, off;
 
-                       rte_raw_cksum_mbuf(m, hdr->csum_start,
+                       if (rte_raw_cksum_mbuf(m, hdr->csum_start,
                                rte_pktmbuf_pkt_len(m) - hdr->csum_start,
-                               &csum);
+                               &csum) < 0)
+                               return -1;
                        if (likely(csum != 0xffff))
                                csum = ~csum;
                        off = hdr->csum_offset + hdr->csum_start;
@@ -365,7 +374,12 @@ virtqueue_dequeue_batch_packed_vec(struct virtnet_rx *rxvq,
                return -1;
 
        /* only care avail/used bits */
+#if defined(RTE_ARCH_I686)
+       __m512i v_mask = _mm512_set4_epi64(PACKED_FLAGS_MASK, 0x0,
+                                          PACKED_FLAGS_MASK, 0x0);
+#else
        __m512i v_mask = _mm512_maskz_set1_epi64(0xaa, PACKED_FLAGS_MASK);
+#endif
        desc_addr = &vq->vq_packed.ring.desc[id];
 
        __m512i v_desc = _mm512_loadu_si512(desc_addr);
@@ -373,7 +387,12 @@ virtqueue_dequeue_batch_packed_vec(struct virtnet_rx *rxvq,
 
        __m512i v_used_flag = _mm512_setzero_si512();
        if (vq->vq_packed.used_wrap_counter)
+#if defined(RTE_ARCH_I686)
+               v_used_flag = _mm512_set4_epi64(PACKED_FLAGS_MASK, 0x0,
+                                               PACKED_FLAGS_MASK, 0x0);
+#else
                v_used_flag = _mm512_maskz_set1_epi64(0xaa, PACKED_FLAGS_MASK);
+#endif
 
        /* Check all descs are used */
        desc_stats = _mm512_cmpneq_epu64_mask(v_flag, v_used_flag);