X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=lib%2Flibrte_vhost%2Fvirtio_net.c;h=583bf379c65ae7ef936eb4ba5bc22f63dc573ef0;hb=e863fe3a13da89787fdf3b5c590101a3c0f10af6;hp=236498f71eeb9d1e0de3e9192418ea5600c229d5;hpb=cd6760da1076d43c3ef9624c4efca0921b5cb58d;p=dpdk.git diff --git a/lib/librte_vhost/virtio_net.c b/lib/librte_vhost/virtio_net.c index 236498f71e..583bf379c6 100644 --- a/lib/librte_vhost/virtio_net.c +++ b/lib/librte_vhost/virtio_net.c @@ -117,31 +117,6 @@ flush_shadow_used_ring_split(struct virtio_net *dev, struct vhost_virtqueue *vq) sizeof(vq->used->idx)); } -static __rte_always_inline void -async_flush_shadow_used_ring_split(struct virtio_net *dev, - struct vhost_virtqueue *vq) -{ - uint16_t used_idx = vq->last_used_idx & (vq->size - 1); - - if (used_idx + vq->shadow_used_idx <= vq->size) { - do_flush_shadow_used_ring_split(dev, vq, used_idx, 0, - vq->shadow_used_idx); - } else { - uint16_t size; - - /* update used ring interval [used_idx, vq->size] */ - size = vq->size - used_idx; - do_flush_shadow_used_ring_split(dev, vq, used_idx, 0, size); - - /* update the left half used ring interval [0, left_size] */ - do_flush_shadow_used_ring_split(dev, vq, 0, size, - vq->shadow_used_idx - size); - } - - vq->last_used_idx += vq->shadow_used_idx; - vq->shadow_used_idx = 0; -} - static __rte_always_inline void update_shadow_used_ring_split(struct vhost_virtqueue *vq, uint16_t desc_idx, uint32_t len) @@ -171,7 +146,8 @@ vhost_flush_enqueue_shadow_packed(struct virtio_net *dev, used_idx -= vq->size; } - rte_smp_wmb(); + /* The ordering for storing desc flags needs to be enforced. */ + rte_atomic_thread_fence(__ATOMIC_RELEASE); for (i = 0; i < vq->shadow_used_idx; i++) { uint16_t flags; @@ -222,8 +198,9 @@ vhost_flush_dequeue_shadow_packed(struct virtio_net *dev, struct vring_used_elem_packed *used_elem = &vq->shadow_used_packed[0]; vq->desc_packed[vq->shadow_last_used_idx].id = used_elem->id; - rte_smp_wmb(); - vq->desc_packed[vq->shadow_last_used_idx].flags = used_elem->flags; + /* desc flags is the synchronization point for virtio packed vring */ + __atomic_store_n(&vq->desc_packed[vq->shadow_last_used_idx].flags, + used_elem->flags, __ATOMIC_RELEASE); vhost_log_cache_used_vring(dev, vq, vq->shadow_last_used_idx * sizeof(struct vring_packed_desc), @@ -253,7 +230,7 @@ vhost_flush_enqueue_batch_packed(struct virtio_net *dev, vq->desc_packed[vq->last_used_idx + i].len = lens[i]; } - rte_smp_wmb(); + rte_atomic_thread_fence(__ATOMIC_RELEASE); vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) vq->desc_packed[vq->last_used_idx + i].flags = flags; @@ -312,7 +289,7 @@ vhost_shadow_dequeue_batch_packed(struct virtio_net *dev, vq->desc_packed[vq->last_used_idx + i].len = 0; } - rte_smp_wmb(); + rte_atomic_thread_fence(__ATOMIC_RELEASE); vhost_for_each_try_unroll(i, begin, PACKED_BATCH_SIZE) vq->desc_packed[vq->last_used_idx + i].flags = flags; @@ -980,6 +957,7 @@ async_mbuf_to_desc(struct virtio_net *dev, struct vhost_virtqueue *vq, struct batch_copy_elem *batch_copy = vq->batch_copy_elems; struct virtio_net_hdr_mrg_rxbuf tmp_hdr, *hdr = NULL; int error = 0; + uint64_t mapped_len; uint32_t tlen = 0; int tvec_idx = 0; @@ -1072,24 +1050,31 @@ async_mbuf_to_desc(struct virtio_net *dev, struct vhost_virtqueue *vq, cpy_len = RTE_MIN(buf_avail, mbuf_avail); - if (unlikely(cpy_len >= cpy_threshold)) { - hpa = (void *)(uintptr_t)gpa_to_hpa(dev, - buf_iova + buf_offset, cpy_len); + while (unlikely(cpy_len && cpy_len >= cpy_threshold)) { + hpa = (void *)(uintptr_t)gpa_to_first_hpa(dev, + buf_iova + buf_offset, + cpy_len, &mapped_len); - if (unlikely(!hpa)) { - error = -1; - goto out; - } + if (unlikely(!hpa || mapped_len < cpy_threshold)) + break; async_fill_vec(src_iovec + tvec_idx, (void *)(uintptr_t)rte_pktmbuf_iova_offset(m, - mbuf_offset), cpy_len); + mbuf_offset), (size_t)mapped_len); - async_fill_vec(dst_iovec + tvec_idx, hpa, cpy_len); + async_fill_vec(dst_iovec + tvec_idx, + hpa, (size_t)mapped_len); - tlen += cpy_len; + tlen += (uint32_t)mapped_len; + cpy_len -= (uint32_t)mapped_len; + mbuf_avail -= (uint32_t)mapped_len; + mbuf_offset += (uint32_t)mapped_len; + buf_avail -= (uint32_t)mapped_len; + buf_offset += (uint32_t)mapped_len; tvec_idx++; - } else { + } + + if (likely(cpy_len)) { if (unlikely(vq->batch_copy_nb_elems >= vq->size)) { rte_memcpy( (void *)((uintptr_t)(buf_addr + buf_offset)), @@ -1110,17 +1095,22 @@ async_mbuf_to_desc(struct virtio_net *dev, struct vhost_virtqueue *vq, cpy_len; vq->batch_copy_nb_elems++; } + + mbuf_avail -= cpy_len; + mbuf_offset += cpy_len; + buf_avail -= cpy_len; + buf_offset += cpy_len; } - mbuf_avail -= cpy_len; - mbuf_offset += cpy_len; - buf_avail -= cpy_len; - buf_offset += cpy_len; } out: - async_fill_iter(src_it, tlen, src_iovec, tvec_idx); - async_fill_iter(dst_it, tlen, dst_iovec, tvec_idx); + if (tlen) { + async_fill_iter(src_it, tlen, src_iovec, tvec_idx); + async_fill_iter(dst_it, tlen, dst_iovec, tvec_idx); + } else { + src_it->count = 0; + } return error; } @@ -1138,7 +1128,7 @@ vhost_enqueue_single_packed(struct virtio_net *dev, uint16_t buf_id = 0; uint32_t len = 0; uint16_t desc_count; - uint32_t size = pkt->pkt_len + dev->vhost_hlen; + uint32_t size = pkt->pkt_len + sizeof(struct virtio_net_hdr_mrg_rxbuf); uint16_t num_buffers = 0; uint32_t buffer_len[vq->size]; uint16_t buffer_buf_id[vq->size]; @@ -1253,7 +1243,7 @@ virtio_dev_rx_batch_packed(struct virtio_net *dev, uint16_t avail_idx = vq->last_avail_idx; uint64_t desc_addrs[PACKED_BATCH_SIZE]; struct virtio_net_hdr_mrg_rxbuf *hdrs[PACKED_BATCH_SIZE]; - uint32_t buf_offset = dev->vhost_hlen; + uint32_t buf_offset = sizeof(struct virtio_net_hdr_mrg_rxbuf); uint64_t lens[PACKED_BATCH_SIZE]; uint16_t ids[PACKED_BATCH_SIZE]; uint16_t i; @@ -1272,8 +1262,6 @@ virtio_dev_rx_batch_packed(struct virtio_net *dev, return -1; } - rte_smp_rmb(); - vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) lens[i] = descs[avail_idx + i].len; @@ -1299,7 +1287,8 @@ virtio_dev_rx_batch_packed(struct virtio_net *dev, rte_prefetch0((void *)(uintptr_t)desc_addrs[i]); hdrs[i] = (struct virtio_net_hdr_mrg_rxbuf *) (uintptr_t)desc_addrs[i]; - lens[i] = pkts[i]->pkt_len + dev->vhost_hlen; + lens[i] = pkts[i]->pkt_len + + sizeof(struct virtio_net_hdr_mrg_rxbuf); } vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) @@ -1333,7 +1322,6 @@ virtio_dev_rx_single_packed(struct virtio_net *dev, struct buf_vector buf_vec[BUF_VECTOR_MAX]; uint16_t nr_descs = 0; - rte_smp_rmb(); if (unlikely(vhost_enqueue_single_packed(dev, vq, pkt, buf_vec, &nr_descs) < 0)) { VHOST_LOG_DATA(DEBUG, @@ -1353,8 +1341,8 @@ virtio_dev_rx_single_packed(struct virtio_net *dev, static __rte_noinline uint32_t virtio_dev_rx_packed(struct virtio_net *dev, - struct vhost_virtqueue *vq, - struct rte_mbuf **pkts, + struct vhost_virtqueue *__rte_restrict vq, + struct rte_mbuf **__rte_restrict pkts, uint32_t count) { uint32_t pkt_idx = 0; @@ -1439,7 +1427,7 @@ out_access_unlock: uint16_t rte_vhost_enqueue_burst(int vid, uint16_t queue_id, - struct rte_mbuf **pkts, uint16_t count) + struct rte_mbuf **__rte_restrict pkts, uint16_t count) { struct virtio_net *dev = get_device(vid); @@ -1464,46 +1452,16 @@ virtio_dev_rx_async_get_info_idx(uint16_t pkts_idx, (vq_size - n_inflight + pkts_idx) & (vq_size - 1); } -static __rte_always_inline void -virtio_dev_rx_async_submit_split_err(struct virtio_net *dev, - struct vhost_virtqueue *vq, uint16_t queue_id, - uint16_t last_idx, uint16_t shadow_idx) -{ - uint16_t start_idx, pkts_idx, vq_size; - uint64_t *async_pending_info; - - pkts_idx = vq->async_pkts_idx; - async_pending_info = vq->async_pending_info; - vq_size = vq->size; - start_idx = virtio_dev_rx_async_get_info_idx(pkts_idx, - vq_size, vq->async_pkts_inflight_n); - - while (likely((start_idx & (vq_size - 1)) != pkts_idx)) { - uint64_t n_seg = - async_pending_info[(start_idx) & (vq_size - 1)] >> - ASYNC_PENDING_INFO_N_SFT; - - while (n_seg) - n_seg -= vq->async_ops.check_completed_copies(dev->vid, - queue_id, 0, 1); - } - - vq->async_pkts_inflight_n = 0; - vq->batch_copy_nb_elems = 0; - - vq->shadow_used_idx = shadow_idx; - vq->last_avail_idx = last_idx; -} - static __rte_noinline uint32_t virtio_dev_rx_async_submit_split(struct virtio_net *dev, struct vhost_virtqueue *vq, uint16_t queue_id, - struct rte_mbuf **pkts, uint32_t count) + struct rte_mbuf **pkts, uint32_t count, + struct rte_mbuf **comp_pkts, uint32_t *comp_count) { uint32_t pkt_idx = 0, pkt_burst_idx = 0; uint16_t num_buffers; struct buf_vector buf_vec[BUF_VECTOR_MAX]; - uint16_t avail_head, last_idx, shadow_idx; + uint16_t avail_head; struct rte_vhost_iov_iter *it_pool = vq->it_pool; struct iovec *vec_pool = vq->vec_pool; @@ -1512,18 +1470,20 @@ virtio_dev_rx_async_submit_split(struct virtio_net *dev, struct iovec *dst_iovec = vec_pool + (VHOST_MAX_ASYNC_VEC >> 1); struct rte_vhost_iov_iter *src_it = it_pool; struct rte_vhost_iov_iter *dst_it = it_pool + 1; - uint16_t n_free_slot, slot_idx; - int n_pkts = 0; - - avail_head = __atomic_load_n(&vq->avail->idx, __ATOMIC_ACQUIRE); - last_idx = vq->last_avail_idx; - shadow_idx = vq->shadow_used_idx; + uint16_t slot_idx = 0; + uint16_t segs_await = 0; + struct async_inflight_info *pkts_info = vq->async_pkts_info; + uint32_t n_pkts = 0, pkt_err = 0; + uint32_t num_async_pkts = 0, num_done_pkts = 0; + struct { + uint16_t pkt_idx; + uint16_t last_avail_idx; + } async_pkts_log[MAX_PKT_BURST]; /* - * The ordering between avail index and - * desc reads needs to be enforced. + * The ordering between avail index and desc reads need to be enforced. */ - rte_smp_rmb(); + avail_head = __atomic_load_n(&vq->avail->idx, __ATOMIC_ACQUIRE); rte_prefetch0(&vq->avail->ring[vq->last_avail_idx & (vq->size - 1)]); @@ -1552,76 +1512,123 @@ virtio_dev_rx_async_submit_split(struct virtio_net *dev, break; } - slot_idx = (vq->async_pkts_idx + pkt_idx) & (vq->size - 1); + slot_idx = (vq->async_pkts_idx + num_async_pkts) & + (vq->size - 1); if (src_it->count) { - async_fill_desc(&tdes[pkt_burst_idx], src_it, dst_it); - pkt_burst_idx++; - vq->async_pending_info[slot_idx] = - num_buffers | (src_it->nr_segs << 16); + uint16_t from, to; + + async_fill_desc(&tdes[pkt_burst_idx++], src_it, dst_it); + pkts_info[slot_idx].descs = num_buffers; + pkts_info[slot_idx].mbuf = pkts[pkt_idx]; + async_pkts_log[num_async_pkts].pkt_idx = pkt_idx; + async_pkts_log[num_async_pkts++].last_avail_idx = + vq->last_avail_idx; src_iovec += src_it->nr_segs; dst_iovec += dst_it->nr_segs; src_it += 2; dst_it += 2; - } else { - vq->async_pending_info[slot_idx] = num_buffers; - vq->async_pkts_inflight_n++; - } + segs_await += src_it->nr_segs; + + /** + * recover shadow used ring and keep DMA-occupied + * descriptors. + */ + from = vq->shadow_used_idx - num_buffers; + to = vq->async_desc_idx & (vq->size - 1); + if (num_buffers + to <= vq->size) { + rte_memcpy(&vq->async_descs_split[to], + &vq->shadow_used_split[from], + num_buffers * + sizeof(struct vring_used_elem)); + } else { + int size = vq->size - to; + + rte_memcpy(&vq->async_descs_split[to], + &vq->shadow_used_split[from], + size * + sizeof(struct vring_used_elem)); + rte_memcpy(vq->async_descs_split, + &vq->shadow_used_split[from + + size], (num_buffers - size) * + sizeof(struct vring_used_elem)); + } + vq->async_desc_idx += num_buffers; + vq->shadow_used_idx -= num_buffers; + } else + comp_pkts[num_done_pkts++] = pkts[pkt_idx]; vq->last_avail_idx += num_buffers; - if (pkt_burst_idx >= VHOST_ASYNC_BATCH_THRESHOLD || - (pkt_idx == count - 1 && pkt_burst_idx)) { + /* + * conditions to trigger async device transfer: + * - buffered packet number reaches transfer threshold + * - unused async iov number is less than max vhost vector + */ + if (unlikely(pkt_burst_idx >= VHOST_ASYNC_BATCH_THRESHOLD || + ((VHOST_MAX_ASYNC_VEC >> 1) - segs_await < + BUF_VECTOR_MAX))) { n_pkts = vq->async_ops.transfer_data(dev->vid, queue_id, tdes, 0, pkt_burst_idx); src_iovec = vec_pool; dst_iovec = vec_pool + (VHOST_MAX_ASYNC_VEC >> 1); src_it = it_pool; dst_it = it_pool + 1; + segs_await = 0; + vq->async_pkts_inflight_n += n_pkts; - if (unlikely(n_pkts < (int)pkt_burst_idx)) { - vq->async_pkts_inflight_n += - n_pkts > 0 ? n_pkts : 0; - virtio_dev_rx_async_submit_split_err(dev, - vq, queue_id, last_idx, shadow_idx); - return 0; + if (unlikely(n_pkts < pkt_burst_idx)) { + /* + * log error packets number here and do actual + * error processing when applications poll + * completion + */ + pkt_err = pkt_burst_idx - n_pkts; + pkt_burst_idx = 0; + break; } pkt_burst_idx = 0; - vq->async_pkts_inflight_n += n_pkts; } } if (pkt_burst_idx) { n_pkts = vq->async_ops.transfer_data(dev->vid, queue_id, tdes, 0, pkt_burst_idx); - if (unlikely(n_pkts < (int)pkt_burst_idx)) { - vq->async_pkts_inflight_n += n_pkts > 0 ? n_pkts : 0; - virtio_dev_rx_async_submit_split_err(dev, vq, queue_id, - last_idx, shadow_idx); - return 0; - } - vq->async_pkts_inflight_n += n_pkts; + + if (unlikely(n_pkts < pkt_burst_idx)) + pkt_err = pkt_burst_idx - n_pkts; } do_data_copy_enqueue(dev, vq); - n_free_slot = vq->size - vq->async_pkts_idx; - if (n_free_slot > pkt_idx) { - rte_memcpy(&vq->async_pkts_pending[vq->async_pkts_idx], - pkts, pkt_idx * sizeof(uintptr_t)); - vq->async_pkts_idx += pkt_idx; - } else { - rte_memcpy(&vq->async_pkts_pending[vq->async_pkts_idx], - pkts, n_free_slot * sizeof(uintptr_t)); - rte_memcpy(&vq->async_pkts_pending[0], - &pkts[n_free_slot], - (pkt_idx - n_free_slot) * sizeof(uintptr_t)); - vq->async_pkts_idx = pkt_idx - n_free_slot; + if (unlikely(pkt_err)) { + uint16_t num_descs = 0; + + num_async_pkts -= pkt_err; + /* calculate the sum of descriptors of DMA-error packets. */ + while (pkt_err-- > 0) { + num_descs += pkts_info[slot_idx & (vq->size - 1)].descs; + slot_idx--; + } + vq->async_desc_idx -= num_descs; + /* recover shadow used ring and available ring */ + vq->shadow_used_idx -= (vq->last_avail_idx - + async_pkts_log[num_async_pkts].last_avail_idx - + num_descs); + vq->last_avail_idx = + async_pkts_log[num_async_pkts].last_avail_idx; + pkt_idx = async_pkts_log[num_async_pkts].pkt_idx; + num_done_pkts = pkt_idx - num_async_pkts; } - if (likely(vq->shadow_used_idx)) - async_flush_shadow_used_ring_split(dev, vq); + vq->async_pkts_idx += num_async_pkts; + *comp_count = num_done_pkts; + + if (likely(vq->shadow_used_idx)) { + flush_shadow_used_ring_split(dev, vq); + vhost_vring_call_split(dev, vq); + } return pkt_idx; } @@ -1631,9 +1638,13 @@ uint16_t rte_vhost_poll_enqueue_completed(int vid, uint16_t queue_id, { struct virtio_net *dev = get_device(vid); struct vhost_virtqueue *vq; - uint16_t n_pkts_cpl, n_pkts_put = 0, n_descs = 0; + uint16_t n_pkts_cpl = 0, n_pkts_put = 0, n_descs = 0; uint16_t start_idx, pkts_idx, vq_size; - uint64_t *async_pending_info; + struct async_inflight_info *pkts_info; + uint16_t from, i; + + if (!dev) + return 0; VHOST_LOG_DATA(DEBUG, "(%d) %s\n", dev->vid, __func__); if (unlikely(!is_valid_virt_queue_idx(queue_id, 0, dev->nr_vring))) { @@ -1644,63 +1655,80 @@ uint16_t rte_vhost_poll_enqueue_completed(int vid, uint16_t queue_id, vq = dev->virtqueue[queue_id]; + if (unlikely(!vq->async_registered)) { + VHOST_LOG_DATA(ERR, "(%d) %s: async not registered for queue id %d.\n", + dev->vid, __func__, queue_id); + return 0; + } + rte_spinlock_lock(&vq->access_lock); - pkts_idx = vq->async_pkts_idx; - async_pending_info = vq->async_pending_info; + pkts_idx = vq->async_pkts_idx & (vq->size - 1); + pkts_info = vq->async_pkts_info; vq_size = vq->size; start_idx = virtio_dev_rx_async_get_info_idx(pkts_idx, vq_size, vq->async_pkts_inflight_n); - n_pkts_cpl = - vq->async_ops.check_completed_copies(vid, queue_id, 0, count); - - rte_smp_wmb(); - - while (likely(((start_idx + n_pkts_put) & (vq_size - 1)) != pkts_idx)) { - uint64_t info = async_pending_info[ - (start_idx + n_pkts_put) & (vq_size - 1)]; - uint64_t n_segs; - n_pkts_put++; - n_descs += info & ASYNC_PENDING_INFO_N_MSK; - n_segs = info >> ASYNC_PENDING_INFO_N_SFT; - - if (n_segs) { - if (!n_pkts_cpl || n_pkts_cpl < n_segs) { - n_pkts_put--; - n_descs -= info & ASYNC_PENDING_INFO_N_MSK; - if (n_pkts_cpl) { - async_pending_info[ - (start_idx + n_pkts_put) & - (vq_size - 1)] = - ((n_segs - n_pkts_cpl) << - ASYNC_PENDING_INFO_N_SFT) | - (info & ASYNC_PENDING_INFO_N_MSK); - n_pkts_cpl = 0; - } - break; + if (count > vq->async_last_pkts_n) + n_pkts_cpl = vq->async_ops.check_completed_copies(vid, + queue_id, 0, count - vq->async_last_pkts_n); + n_pkts_cpl += vq->async_last_pkts_n; + + n_pkts_put = RTE_MIN(count, n_pkts_cpl); + if (unlikely(n_pkts_put == 0)) { + vq->async_last_pkts_n = n_pkts_cpl; + goto done; + } + + for (i = 0; i < n_pkts_put; i++) { + from = (start_idx + i) & (vq_size - 1); + n_descs += pkts_info[from].descs; + pkts[i] = pkts_info[from].mbuf; + } + vq->async_last_pkts_n = n_pkts_cpl - n_pkts_put; + vq->async_pkts_inflight_n -= n_pkts_put; + + if (likely(vq->enabled && vq->access_ok)) { + uint16_t nr_left = n_descs; + uint16_t nr_copy; + uint16_t to; + + /* write back completed descriptors to used ring */ + do { + from = vq->last_async_desc_idx & (vq->size - 1); + nr_copy = nr_left + from <= vq->size ? nr_left : + vq->size - from; + to = vq->last_used_idx & (vq->size - 1); + + if (to + nr_copy <= vq->size) { + rte_memcpy(&vq->used->ring[to], + &vq->async_descs_split[from], + nr_copy * + sizeof(struct vring_used_elem)); + } else { + uint16_t size = vq->size - to; + + rte_memcpy(&vq->used->ring[to], + &vq->async_descs_split[from], + size * + sizeof(struct vring_used_elem)); + rte_memcpy(vq->used->ring, + &vq->async_descs_split[from + + size], (nr_copy - size) * + sizeof(struct vring_used_elem)); } - n_pkts_cpl -= n_segs; - } - } - if (n_pkts_put) { - vq->async_pkts_inflight_n -= n_pkts_put; - __atomic_add_fetch(&vq->used->idx, n_descs, __ATOMIC_RELEASE); + vq->last_async_desc_idx += nr_copy; + vq->last_used_idx += nr_copy; + nr_left -= nr_copy; + } while (nr_left > 0); + __atomic_add_fetch(&vq->used->idx, n_descs, __ATOMIC_RELEASE); vhost_vring_call_split(dev, vq); - } - - if (start_idx + n_pkts_put <= vq_size) { - rte_memcpy(pkts, &vq->async_pkts_pending[start_idx], - n_pkts_put * sizeof(uintptr_t)); - } else { - rte_memcpy(pkts, &vq->async_pkts_pending[start_idx], - (vq_size - start_idx) * sizeof(uintptr_t)); - rte_memcpy(&pkts[vq_size - start_idx], vq->async_pkts_pending, - (n_pkts_put - vq_size + start_idx) * sizeof(uintptr_t)); - } + } else + vq->last_async_desc_idx += n_descs; +done: rte_spinlock_unlock(&vq->access_lock); return n_pkts_put; @@ -1708,11 +1736,11 @@ uint16_t rte_vhost_poll_enqueue_completed(int vid, uint16_t queue_id, static __rte_always_inline uint32_t virtio_dev_rx_async_submit(struct virtio_net *dev, uint16_t queue_id, - struct rte_mbuf **pkts, uint32_t count) + struct rte_mbuf **pkts, uint32_t count, + struct rte_mbuf **comp_pkts, uint32_t *comp_count) { struct vhost_virtqueue *vq; uint32_t nb_tx = 0; - bool drawback = false; VHOST_LOG_DATA(DEBUG, "(%d) %s\n", dev->vid, __func__); if (unlikely(!is_valid_virt_queue_idx(queue_id, 0, dev->nr_vring))) { @@ -1725,13 +1753,8 @@ virtio_dev_rx_async_submit(struct virtio_net *dev, uint16_t queue_id, rte_spinlock_lock(&vq->access_lock); - if (unlikely(vq->enabled == 0)) - goto out_access_unlock; - - if (unlikely(!vq->async_registered)) { - drawback = true; + if (unlikely(vq->enabled == 0 || !vq->async_registered)) goto out_access_unlock; - } if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM)) vhost_user_iotlb_rd_lock(vq); @@ -1749,7 +1772,8 @@ virtio_dev_rx_async_submit(struct virtio_net *dev, uint16_t queue_id, nb_tx = 0; else nb_tx = virtio_dev_rx_async_submit_split(dev, - vq, queue_id, pkts, count); + vq, queue_id, pkts, count, comp_pkts, + comp_count); out: if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM)) @@ -1758,18 +1782,17 @@ out: out_access_unlock: rte_spinlock_unlock(&vq->access_lock); - if (drawback) - return rte_vhost_enqueue_burst(dev->vid, queue_id, pkts, count); - return nb_tx; } uint16_t rte_vhost_submit_enqueue_burst(int vid, uint16_t queue_id, - struct rte_mbuf **pkts, uint16_t count) + struct rte_mbuf **pkts, uint16_t count, + struct rte_mbuf **comp_pkts, uint32_t *comp_count) { struct virtio_net *dev = get_device(vid); + *comp_count = 0; if (!dev) return 0; @@ -1780,7 +1803,8 @@ rte_vhost_submit_enqueue_burst(int vid, uint16_t queue_id, return 0; } - return virtio_dev_rx_async_submit(dev, queue_id, pkts, count); + return virtio_dev_rx_async_submit(dev, queue_id, pkts, count, comp_pkts, + comp_count); } static inline bool @@ -1825,7 +1849,7 @@ parse_ethernet(struct rte_mbuf *m, uint16_t *l4_proto, void **l4_hdr) case RTE_ETHER_TYPE_IPV4: ipv4_hdr = l3_hdr; *l4_proto = ipv4_hdr->next_proto_id; - m->l3_len = (ipv4_hdr->version_ihl & 0x0f) * 4; + m->l3_len = rte_ipv4_hdr_len(ipv4_hdr); *l4_hdr = (char *)l3_hdr + m->l3_len; m->ol_flags |= PKT_TX_IPV4; break; @@ -1925,7 +1949,7 @@ copy_desc_to_mbuf(struct virtio_net *dev, struct vhost_virtqueue *vq, struct rte_mbuf *m, struct rte_mempool *mbuf_pool) { uint32_t buf_avail, buf_offset; - uint64_t buf_addr, buf_iova, buf_len; + uint64_t buf_addr, buf_len; uint32_t mbuf_avail, mbuf_offset; uint32_t cpy_len; struct rte_mbuf *cur = m, *prev = m; @@ -1937,7 +1961,6 @@ copy_desc_to_mbuf(struct virtio_net *dev, struct vhost_virtqueue *vq, int error = 0; buf_addr = buf_vec[vec_idx].buf_addr; - buf_iova = buf_vec[vec_idx].buf_iova; buf_len = buf_vec[vec_idx].buf_len; if (unlikely(buf_len < dev->vhost_hlen && nr_vec <= 1)) { @@ -1967,14 +1990,12 @@ copy_desc_to_mbuf(struct virtio_net *dev, struct vhost_virtqueue *vq, buf_offset = dev->vhost_hlen - buf_len; vec_idx++; buf_addr = buf_vec[vec_idx].buf_addr; - buf_iova = buf_vec[vec_idx].buf_iova; buf_len = buf_vec[vec_idx].buf_len; buf_avail = buf_len - buf_offset; } else if (buf_len == dev->vhost_hlen) { if (unlikely(++vec_idx >= nr_vec)) goto out; buf_addr = buf_vec[vec_idx].buf_addr; - buf_iova = buf_vec[vec_idx].buf_iova; buf_len = buf_vec[vec_idx].buf_len; buf_offset = 0; @@ -1991,48 +2012,23 @@ copy_desc_to_mbuf(struct virtio_net *dev, struct vhost_virtqueue *vq, mbuf_offset = 0; mbuf_avail = m->buf_len - RTE_PKTMBUF_HEADROOM; while (1) { - uint64_t hpa; - cpy_len = RTE_MIN(buf_avail, mbuf_avail); - /* - * A desc buf might across two host physical pages that are - * not continuous. In such case (gpa_to_hpa returns 0), data - * will be copied even though zero copy is enabled. - */ - if (unlikely(dev->dequeue_zero_copy && (hpa = gpa_to_hpa(dev, - buf_iova + buf_offset, cpy_len)))) { - cur->data_len = cpy_len; - cur->data_off = 0; - cur->buf_addr = - (void *)(uintptr_t)(buf_addr + buf_offset); - cur->buf_iova = hpa; - - /* - * In zero copy mode, one mbuf can only reference data - * for one or partial of one desc buff. - */ - mbuf_avail = cpy_len; - } else { - if (likely(cpy_len > MAX_BATCH_LEN || - vq->batch_copy_nb_elems >= vq->size || - (hdr && cur == m))) { - rte_memcpy(rte_pktmbuf_mtod_offset(cur, void *, - mbuf_offset), - (void *)((uintptr_t)(buf_addr + - buf_offset)), - cpy_len); - } else { - batch_copy[vq->batch_copy_nb_elems].dst = - rte_pktmbuf_mtod_offset(cur, void *, - mbuf_offset); - batch_copy[vq->batch_copy_nb_elems].src = + if (likely(cpy_len > MAX_BATCH_LEN || + vq->batch_copy_nb_elems >= vq->size || + (hdr && cur == m))) { + rte_memcpy(rte_pktmbuf_mtod_offset(cur, void *, + mbuf_offset), (void *)((uintptr_t)(buf_addr + - buf_offset)); - batch_copy[vq->batch_copy_nb_elems].len = - cpy_len; - vq->batch_copy_nb_elems++; - } + buf_offset)), cpy_len); + } else { + batch_copy[vq->batch_copy_nb_elems].dst = + rte_pktmbuf_mtod_offset(cur, void *, + mbuf_offset); + batch_copy[vq->batch_copy_nb_elems].src = + (void *)((uintptr_t)(buf_addr + buf_offset)); + batch_copy[vq->batch_copy_nb_elems].len = cpy_len; + vq->batch_copy_nb_elems++; } mbuf_avail -= cpy_len; @@ -2046,7 +2042,6 @@ copy_desc_to_mbuf(struct virtio_net *dev, struct vhost_virtqueue *vq, break; buf_addr = buf_vec[vec_idx].buf_addr; - buf_iova = buf_vec[vec_idx].buf_iova; buf_len = buf_vec[vec_idx].buf_len; buf_offset = 0; @@ -2068,8 +2063,6 @@ copy_desc_to_mbuf(struct virtio_net *dev, struct vhost_virtqueue *vq, error = -1; goto out; } - if (unlikely(dev->dequeue_zero_copy)) - rte_mbuf_refcnt_update(cur, 1); prev->next = cur; prev->data_len = mbuf_offset; @@ -2093,37 +2086,6 @@ out: return error; } -static __rte_always_inline struct zcopy_mbuf * -get_zmbuf(struct vhost_virtqueue *vq) -{ - uint16_t i; - uint16_t last; - int tries = 0; - - /* search [last_zmbuf_idx, zmbuf_size) */ - i = vq->last_zmbuf_idx; - last = vq->zmbuf_size; - -again: - for (; i < last; i++) { - if (vq->zmbufs[i].in_use == 0) { - vq->last_zmbuf_idx = i + 1; - vq->zmbufs[i].in_use = 1; - return &vq->zmbufs[i]; - } - } - - tries++; - if (tries == 1) { - /* search [0, last_zmbuf_idx) */ - i = 0; - last = vq->last_zmbuf_idx; - goto again; - } - - return NULL; -} - static void virtio_dev_extbuf_free(void *addr __rte_unused, void *opaque) { @@ -2139,16 +2101,8 @@ virtio_dev_extbuf_alloc(struct rte_mbuf *pkt, uint32_t size) rte_iova_t iova; void *buf; - /* Try to use pkt buffer to store shinfo to reduce the amount of memory - * required, otherwise store shinfo in the new buffer. - */ - if (rte_pktmbuf_tailroom(pkt) >= sizeof(*shinfo)) - shinfo = rte_pktmbuf_mtod(pkt, - struct rte_mbuf_ext_shared_info *); - else { - total_len += sizeof(*shinfo) + sizeof(uintptr_t); - total_len = RTE_ALIGN_CEIL(total_len, sizeof(uintptr_t)); - } + total_len += sizeof(*shinfo) + sizeof(uintptr_t); + total_len = RTE_ALIGN_CEIL(total_len, sizeof(uintptr_t)); if (unlikely(total_len > UINT16_MAX)) return -ENOSPC; @@ -2159,18 +2113,12 @@ virtio_dev_extbuf_alloc(struct rte_mbuf *pkt, uint32_t size) return -ENOMEM; /* Initialize shinfo */ - if (shinfo) { - shinfo->free_cb = virtio_dev_extbuf_free; - shinfo->fcb_opaque = buf; - rte_mbuf_ext_refcnt_set(shinfo, 1); - } else { - shinfo = rte_pktmbuf_ext_shinfo_init_helper(buf, &buf_len, - virtio_dev_extbuf_free, buf); - if (unlikely(shinfo == NULL)) { - rte_free(buf); - VHOST_LOG_DATA(ERR, "Failed to init shinfo\n"); - return -1; - } + shinfo = rte_pktmbuf_ext_shinfo_init_helper(buf, &buf_len, + virtio_dev_extbuf_free, buf); + if (unlikely(shinfo == NULL)) { + rte_free(buf); + VHOST_LOG_DATA(ERR, "Failed to init shinfo\n"); + return -1; } iova = rte_malloc_virt2iova(buf); @@ -2223,30 +2171,6 @@ virtio_dev_tx_split(struct virtio_net *dev, struct vhost_virtqueue *vq, uint16_t dropped = 0; static bool allocerr_warned; - if (unlikely(dev->dequeue_zero_copy)) { - struct zcopy_mbuf *zmbuf, *next; - - for (zmbuf = TAILQ_FIRST(&vq->zmbuf_list); - zmbuf != NULL; zmbuf = next) { - next = TAILQ_NEXT(zmbuf, next); - - if (mbuf_is_consumed(zmbuf->mbuf)) { - update_shadow_used_ring_split(vq, - zmbuf->desc_idx, 0); - TAILQ_REMOVE(&vq->zmbuf_list, zmbuf, next); - restore_mbuf(zmbuf->mbuf); - rte_pktmbuf_free(zmbuf->mbuf); - put_zmbuf(zmbuf); - vq->nr_zmbuf -= 1; - } - } - - if (likely(vq->shadow_used_idx)) { - flush_shadow_used_ring_split(dev, vq); - vhost_vring_call_split(dev, vq); - } - } - /* * The ordering between avail index and * desc reads needs to be enforced. @@ -2279,8 +2203,7 @@ virtio_dev_tx_split(struct virtio_net *dev, struct vhost_virtqueue *vq, VHOST_ACCESS_RO) < 0)) break; - if (likely(dev->dequeue_zero_copy == 0)) - update_shadow_used_ring_split(vq, head_idx, 0); + update_shadow_used_ring_split(vq, head_idx, 0); pkts[i] = virtio_dev_pktmbuf_alloc(dev, mbuf_pool, buf_len); if (unlikely(pkts[i] == NULL)) { @@ -2314,42 +2237,16 @@ virtio_dev_tx_split(struct virtio_net *dev, struct vhost_virtqueue *vq, i++; break; } - - if (unlikely(dev->dequeue_zero_copy)) { - struct zcopy_mbuf *zmbuf; - - zmbuf = get_zmbuf(vq); - if (!zmbuf) { - rte_pktmbuf_free(pkts[i]); - dropped += 1; - i++; - break; - } - zmbuf->mbuf = pkts[i]; - zmbuf->desc_idx = head_idx; - - /* - * Pin lock the mbuf; we will check later to see - * whether the mbuf is freed (when we are the last - * user) or not. If that's the case, we then could - * update the used ring safely. - */ - rte_mbuf_refcnt_update(pkts[i], 1); - - vq->nr_zmbuf += 1; - TAILQ_INSERT_TAIL(&vq->zmbuf_list, zmbuf, next); - } } + vq->last_avail_idx += i; - if (likely(dev->dequeue_zero_copy == 0)) { - do_data_copy_dequeue(vq); - if (unlikely(i < count)) - vq->shadow_used_idx = i; - if (likely(vq->shadow_used_idx)) { - flush_shadow_used_ring_split(dev, vq); - vhost_vring_call_split(dev, vq); - } + do_data_copy_dequeue(vq); + if (unlikely(i < count)) + vq->shadow_used_idx = i; + if (likely(vq->shadow_used_idx)) { + flush_shadow_used_ring_split(dev, vq); + vhost_vring_call_split(dev, vq); } return (i - dropped); @@ -2366,10 +2263,9 @@ vhost_reserve_avail_batch_packed(struct virtio_net *dev, { bool wrap = vq->avail_wrap_counter; struct vring_packed_desc *descs = vq->desc_packed; - struct virtio_net_hdr *hdr; uint64_t lens[PACKED_BATCH_SIZE]; uint64_t buf_lens[PACKED_BATCH_SIZE]; - uint32_t buf_offset = dev->vhost_hlen; + uint32_t buf_offset = sizeof(struct virtio_net_hdr_mrg_rxbuf); uint16_t flags, i; if (unlikely(avail_idx & PACKED_BATCH_MASK)) @@ -2385,7 +2281,7 @@ vhost_reserve_avail_batch_packed(struct virtio_net *dev, return -1; } - rte_smp_rmb(); + rte_atomic_thread_fence(__ATOMIC_ACQUIRE); vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) lens[i] = descs[avail_idx + i].len; @@ -2423,13 +2319,6 @@ vhost_reserve_avail_batch_packed(struct virtio_net *dev, ids[i] = descs[avail_idx + i].id; } - if (virtio_net_with_host_offload(dev)) { - vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) { - hdr = (struct virtio_net_hdr *)(desc_addrs[i]); - vhost_dequeue_offload(hdr, pkts[i]); - } - } - return 0; free_buf: @@ -2446,7 +2335,8 @@ virtio_dev_tx_batch_packed(struct virtio_net *dev, struct rte_mbuf **pkts) { uint16_t avail_idx = vq->last_avail_idx; - uint32_t buf_offset = dev->vhost_hlen; + uint32_t buf_offset = sizeof(struct virtio_net_hdr_mrg_rxbuf); + struct virtio_net_hdr *hdr; uintptr_t desc_addrs[PACKED_BATCH_SIZE]; uint16_t ids[PACKED_BATCH_SIZE]; uint16_t i; @@ -2463,6 +2353,13 @@ virtio_dev_tx_batch_packed(struct virtio_net *dev, (void *)(uintptr_t)(desc_addrs[i] + buf_offset), pkts[i]->pkt_len); + if (virtio_net_with_host_offload(dev)) { + vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) { + hdr = (struct virtio_net_hdr *)(desc_addrs[i]); + vhost_dequeue_offload(hdr, pkts[i]); + } + } + if (virtio_net_is_inorder(dev)) vhost_shadow_dequeue_batch_packed_inorder(vq, ids[PACKED_BATCH_SIZE - 1]); @@ -2549,167 +2446,11 @@ virtio_dev_tx_single_packed(struct virtio_net *dev, return ret; } -static __rte_always_inline int -virtio_dev_tx_batch_packed_zmbuf(struct virtio_net *dev, - struct vhost_virtqueue *vq, - struct rte_mempool *mbuf_pool, - struct rte_mbuf **pkts) -{ - struct zcopy_mbuf *zmbufs[PACKED_BATCH_SIZE]; - uintptr_t desc_addrs[PACKED_BATCH_SIZE]; - uint16_t ids[PACKED_BATCH_SIZE]; - uint16_t i; - - uint16_t avail_idx = vq->last_avail_idx; - - if (vhost_reserve_avail_batch_packed(dev, vq, mbuf_pool, pkts, - avail_idx, desc_addrs, ids)) - return -1; - - vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) - zmbufs[i] = get_zmbuf(vq); - - vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) { - if (!zmbufs[i]) - goto free_pkt; - } - - vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) { - zmbufs[i]->mbuf = pkts[i]; - zmbufs[i]->desc_idx = ids[i]; - zmbufs[i]->desc_count = 1; - } - - vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) - rte_mbuf_refcnt_update(pkts[i], 1); - - vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) - TAILQ_INSERT_TAIL(&vq->zmbuf_list, zmbufs[i], next); - - vq->nr_zmbuf += PACKED_BATCH_SIZE; - vq_inc_last_avail_packed(vq, PACKED_BATCH_SIZE); - - return 0; - -free_pkt: - vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) - rte_pktmbuf_free(pkts[i]); - - return -1; -} - -static __rte_always_inline int -virtio_dev_tx_single_packed_zmbuf(struct virtio_net *dev, - struct vhost_virtqueue *vq, - struct rte_mempool *mbuf_pool, - struct rte_mbuf **pkts) -{ - uint16_t buf_id, desc_count; - struct zcopy_mbuf *zmbuf; - - if (vhost_dequeue_single_packed(dev, vq, mbuf_pool, pkts, &buf_id, - &desc_count)) - return -1; - - zmbuf = get_zmbuf(vq); - if (!zmbuf) { - rte_pktmbuf_free(*pkts); - return -1; - } - zmbuf->mbuf = *pkts; - zmbuf->desc_idx = buf_id; - zmbuf->desc_count = desc_count; - - rte_mbuf_refcnt_update(*pkts, 1); - - vq->nr_zmbuf += 1; - TAILQ_INSERT_TAIL(&vq->zmbuf_list, zmbuf, next); - - vq_inc_last_avail_packed(vq, desc_count); - return 0; -} - -static __rte_always_inline void -free_zmbuf(struct vhost_virtqueue *vq) -{ - struct zcopy_mbuf *next = NULL; - struct zcopy_mbuf *zmbuf; - - for (zmbuf = TAILQ_FIRST(&vq->zmbuf_list); - zmbuf != NULL; zmbuf = next) { - next = TAILQ_NEXT(zmbuf, next); - - uint16_t last_used_idx = vq->last_used_idx; - - if (mbuf_is_consumed(zmbuf->mbuf)) { - uint16_t flags; - flags = vq->desc_packed[last_used_idx].flags; - if (vq->used_wrap_counter) { - flags |= VRING_DESC_F_USED; - flags |= VRING_DESC_F_AVAIL; - } else { - flags &= ~VRING_DESC_F_USED; - flags &= ~VRING_DESC_F_AVAIL; - } - - vq->desc_packed[last_used_idx].id = zmbuf->desc_idx; - vq->desc_packed[last_used_idx].len = 0; - - rte_smp_wmb(); - vq->desc_packed[last_used_idx].flags = flags; - - vq_inc_last_used_packed(vq, zmbuf->desc_count); - - TAILQ_REMOVE(&vq->zmbuf_list, zmbuf, next); - restore_mbuf(zmbuf->mbuf); - rte_pktmbuf_free(zmbuf->mbuf); - put_zmbuf(zmbuf); - vq->nr_zmbuf -= 1; - } - } -} - -static __rte_noinline uint16_t -virtio_dev_tx_packed_zmbuf(struct virtio_net *dev, - struct vhost_virtqueue *vq, - struct rte_mempool *mbuf_pool, - struct rte_mbuf **pkts, - uint32_t count) -{ - uint32_t pkt_idx = 0; - uint32_t remained = count; - - free_zmbuf(vq); - - do { - if (remained >= PACKED_BATCH_SIZE) { - if (!virtio_dev_tx_batch_packed_zmbuf(dev, vq, - mbuf_pool, &pkts[pkt_idx])) { - pkt_idx += PACKED_BATCH_SIZE; - remained -= PACKED_BATCH_SIZE; - continue; - } - } - - if (virtio_dev_tx_single_packed_zmbuf(dev, vq, mbuf_pool, - &pkts[pkt_idx])) - break; - pkt_idx++; - remained--; - - } while (remained); - - if (pkt_idx) - vhost_vring_call_packed(dev, vq); - - return pkt_idx; -} - static __rte_noinline uint16_t virtio_dev_tx_packed(struct virtio_net *dev, - struct vhost_virtqueue *vq, + struct vhost_virtqueue *__rte_restrict vq, struct rte_mempool *mbuf_pool, - struct rte_mbuf **pkts, + struct rte_mbuf **__rte_restrict pkts, uint32_t count) { uint32_t pkt_idx = 0; @@ -2820,14 +2561,9 @@ rte_vhost_dequeue_burst(int vid, uint16_t queue_id, count -= 1; } - if (vq_is_packed(dev)) { - if (unlikely(dev->dequeue_zero_copy)) - count = virtio_dev_tx_packed_zmbuf(dev, vq, mbuf_pool, - pkts, count); - else - count = virtio_dev_tx_packed(dev, vq, mbuf_pool, pkts, - count); - } else + if (vq_is_packed(dev)) + count = virtio_dev_tx_packed(dev, vq, mbuf_pool, pkts, count); + else count = virtio_dev_tx_split(dev, vq, mbuf_pool, pkts, count); out: