X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=lib%2Flibrte_vhost%2Fvirtio_net.c;h=4ddf265679f55bc7ded0f43c5c120aa357bc8fa8;hb=ef861692c3980b906d433dcd17318c218eaf2988;hp=2b7ffcf922a9757ab60fd565476e707274ec35d0;hpb=b1cce26af1dca2389e140d07fb32902abce9fd86;p=dpdk.git diff --git a/lib/librte_vhost/virtio_net.c b/lib/librte_vhost/virtio_net.c index 2b7ffcf922..4ddf265679 100644 --- a/lib/librte_vhost/virtio_net.c +++ b/lib/librte_vhost/virtio_net.c @@ -37,45 +37,6 @@ is_valid_virt_queue_idx(uint32_t idx, int is_tx, uint32_t nr_vring) return (is_tx ^ (idx & 1)) == 0 && idx < nr_vring; } -static __rte_always_inline void * -alloc_copy_ind_table(struct virtio_net *dev, struct vhost_virtqueue *vq, - uint64_t desc_addr, uint64_t desc_len) -{ - void *idesc; - uint64_t src, dst; - uint64_t len, remain = desc_len; - - idesc = rte_malloc(__func__, desc_len, 0); - if (unlikely(!idesc)) - return 0; - - dst = (uint64_t)(uintptr_t)idesc; - - while (remain) { - len = remain; - src = vhost_iova_to_vva(dev, vq, desc_addr, &len, - VHOST_ACCESS_RO); - if (unlikely(!src || !len)) { - rte_free(idesc); - return 0; - } - - rte_memcpy((void *)(uintptr_t)dst, (void *)(uintptr_t)src, len); - - remain -= len; - dst += len; - desc_addr += len; - } - - return idesc; -} - -static __rte_always_inline void -free_ind_table(void *idesc) -{ - rte_free(idesc); -} - static __rte_always_inline void do_flush_shadow_used_ring_split(struct virtio_net *dev, struct vhost_virtqueue *vq, @@ -122,7 +83,7 @@ flush_shadow_used_ring_split(struct virtio_net *dev, struct vhost_virtqueue *vq) static __rte_always_inline void update_shadow_used_ring_split(struct vhost_virtqueue *vq, - uint16_t desc_idx, uint16_t len) + uint16_t desc_idx, uint32_t len) { uint16_t i = vq->shadow_used_idx++; @@ -136,6 +97,8 @@ flush_shadow_used_ring_packed(struct virtio_net *dev, { int i; uint16_t used_idx = vq->last_used_idx; + uint16_t head_idx = vq->last_used_idx; + uint16_t head_flags = 0; /* Split loop in two to save memory barriers */ for (i = 0; i < vq->shadow_used_idx; i++) { @@ -147,8 +110,6 @@ flush_shadow_used_ring_packed(struct virtio_net *dev, used_idx -= vq->size; } - rte_smp_wmb(); - for (i = 0; i < vq->shadow_used_idx; i++) { uint16_t flags; @@ -165,28 +126,36 @@ flush_shadow_used_ring_packed(struct virtio_net *dev, flags &= ~VRING_DESC_F_AVAIL; } - vq->desc_packed[vq->last_used_idx].flags = flags; + if (i > 0) { + vq->desc_packed[vq->last_used_idx].flags = flags; - vhost_log_cache_used_vring(dev, vq, + vhost_log_cache_used_vring(dev, vq, vq->last_used_idx * sizeof(struct vring_packed_desc), sizeof(struct vring_packed_desc)); - - vq->last_used_idx += vq->shadow_used_packed[i].count; - if (vq->last_used_idx >= vq->size) { - vq->used_wrap_counter ^= 1; - vq->last_used_idx -= vq->size; + } else { + head_idx = vq->last_used_idx; + head_flags = flags; } + + vq_inc_last_used_packed(vq, vq->shadow_used_packed[i].count); } - rte_smp_wmb(); + __atomic_store_n(&vq->desc_packed[head_idx].flags, head_flags, + __ATOMIC_RELEASE); + + vhost_log_cache_used_vring(dev, vq, + head_idx * + sizeof(struct vring_packed_desc), + sizeof(struct vring_packed_desc)); + vq->shadow_used_idx = 0; vhost_log_cache_sync(dev, vq); } static __rte_always_inline void update_shadow_used_ring_packed(struct vhost_virtqueue *vq, - uint16_t desc_idx, uint16_t len, uint16_t count) + uint16_t desc_idx, uint32_t len, uint16_t count) { uint16_t i = vq->shadow_used_idx++; @@ -204,7 +173,8 @@ do_data_copy_enqueue(struct virtio_net *dev, struct vhost_virtqueue *vq) for (i = 0; i < count; i++) { rte_memcpy(elem[i].dst, elem[i].src, elem[i].len); - vhost_log_cache_write(dev, vq, elem[i].log_addr, elem[i].len); + vhost_log_cache_write_iova(dev, vq, elem[i].log_addr, + elem[i].len); PRINT_PACKET(dev, (uintptr_t)elem[i].dst, elem[i].len, 0); } @@ -244,15 +214,15 @@ virtio_enqueue_offload(struct rte_mbuf *m_buf, struct virtio_net_hdr *net_hdr) switch (csum_l4) { case PKT_TX_TCP_CKSUM: - net_hdr->csum_offset = (offsetof(struct tcp_hdr, + net_hdr->csum_offset = (offsetof(struct rte_tcp_hdr, cksum)); break; case PKT_TX_UDP_CKSUM: - net_hdr->csum_offset = (offsetof(struct udp_hdr, + net_hdr->csum_offset = (offsetof(struct rte_udp_hdr, dgram_cksum)); break; case PKT_TX_SCTP_CKSUM: - net_hdr->csum_offset = (offsetof(struct sctp_hdr, + net_hdr->csum_offset = (offsetof(struct rte_sctp_hdr, cksum)); break; } @@ -264,9 +234,9 @@ virtio_enqueue_offload(struct rte_mbuf *m_buf, struct virtio_net_hdr *net_hdr) /* IP cksum verification cannot be bypassed, then calculate here */ if (m_buf->ol_flags & PKT_TX_IP_CKSUM) { - struct ipv4_hdr *ipv4_hdr; + struct rte_ipv4_hdr *ipv4_hdr; - ipv4_hdr = rte_pktmbuf_mtod_offset(m_buf, struct ipv4_hdr *, + ipv4_hdr = rte_pktmbuf_mtod_offset(m_buf, struct rte_ipv4_hdr *, m_buf->l2_len); ipv4_hdr->hdr_checksum = rte_ipv4_cksum(ipv4_hdr); } @@ -312,6 +282,8 @@ map_one_desc(struct virtio_net *dev, struct vhost_virtqueue *vq, if (unlikely(!desc_addr)) return -1; + rte_prefetch0((void *)(uintptr_t)desc_addr); + buf_vec[vec_id].buf_iova = desc_iova; buf_vec[vec_id].buf_addr = desc_addr; buf_vec[vec_id].buf_len = desc_chunck_len; @@ -329,19 +301,28 @@ static __rte_always_inline int fill_vec_buf_split(struct virtio_net *dev, struct vhost_virtqueue *vq, uint32_t avail_idx, uint16_t *vec_idx, struct buf_vector *buf_vec, uint16_t *desc_chain_head, - uint16_t *desc_chain_len, uint8_t perm) + uint32_t *desc_chain_len, uint8_t perm) { uint16_t idx = vq->avail->ring[avail_idx & (vq->size - 1)]; uint16_t vec_id = *vec_idx; uint32_t len = 0; uint64_t dlen; + uint32_t nr_descs = vq->size; + uint32_t cnt = 0; struct vring_desc *descs = vq->desc; struct vring_desc *idesc = NULL; + if (unlikely(idx >= vq->size)) + return -1; + *desc_chain_head = idx; if (vq->desc[idx].flags & VRING_DESC_F_INDIRECT) { dlen = vq->desc[idx].len; + nr_descs = dlen / sizeof(struct vring_desc); + if (unlikely(nr_descs > vq->size)) + return -1; + descs = (struct vring_desc *)(uintptr_t) vhost_iova_to_vva(dev, vq, vq->desc[idx].addr, &dlen, @@ -354,7 +335,7 @@ fill_vec_buf_split(struct virtio_net *dev, struct vhost_virtqueue *vq, * The indirect desc table is not contiguous * in process VA space, we have to copy it. */ - idesc = alloc_copy_ind_table(dev, vq, + idesc = vhost_alloc_copy_ind_table(dev, vq, vq->desc[idx].addr, vq->desc[idx].len); if (unlikely(!idesc)) return -1; @@ -366,7 +347,7 @@ fill_vec_buf_split(struct virtio_net *dev, struct vhost_virtqueue *vq, } while (1) { - if (unlikely(idx >= vq->size)) { + if (unlikely(idx >= nr_descs || cnt++ >= nr_descs)) { free_ind_table(idesc); return -1; } @@ -409,19 +390,26 @@ reserve_avail_buf_split(struct virtio_net *dev, struct vhost_virtqueue *vq, uint16_t max_tries, tries = 0; uint16_t head_idx = 0; - uint16_t len = 0; + uint32_t len = 0; *num_buffers = 0; cur_idx = vq->last_avail_idx; if (rxvq_is_mergeable(dev)) - max_tries = vq->size; + max_tries = vq->size - 1; else max_tries = 1; while (size > 0) { if (unlikely(cur_idx == avail_head)) return -1; + /* + * if we tried all available ring items, and still + * can't get enough buf, it means something abnormal + * happened. + */ + if (unlikely(++tries > max_tries)) + return -1; if (unlikely(fill_vec_buf_split(dev, vq, cur_idx, &vec_idx, buf_vec, @@ -433,16 +421,7 @@ reserve_avail_buf_split(struct virtio_net *dev, struct vhost_virtqueue *vq, size -= len; cur_idx++; - tries++; *num_buffers += 1; - - /* - * if we tried all available ring items, and still - * can't get enough buf, it means something abnormal - * happened. - */ - if (unlikely(tries > max_tries)) - return -1; } *nr_vec = vec_idx; @@ -454,7 +433,7 @@ static __rte_always_inline int fill_vec_buf_packed_indirect(struct virtio_net *dev, struct vhost_virtqueue *vq, struct vring_packed_desc *desc, uint16_t *vec_idx, - struct buf_vector *buf_vec, uint16_t *len, uint8_t perm) + struct buf_vector *buf_vec, uint32_t *len, uint8_t perm) { uint16_t i; uint32_t nr_descs; @@ -473,7 +452,8 @@ fill_vec_buf_packed_indirect(struct virtio_net *dev, * The indirect desc table is not contiguous * in process VA space, we have to copy it. */ - idescs = alloc_copy_ind_table(dev, vq, desc->addr, desc->len); + idescs = vhost_alloc_copy_ind_table(dev, + vq, desc->addr, desc->len); if (unlikely(!idescs)) return -1; @@ -510,7 +490,7 @@ static __rte_always_inline int fill_vec_buf_packed(struct virtio_net *dev, struct vhost_virtqueue *vq, uint16_t avail_idx, uint16_t *desc_count, struct buf_vector *buf_vec, uint16_t *vec_idx, - uint16_t *buf_id, uint16_t *len, uint8_t perm) + uint16_t *buf_id, uint32_t *len, uint8_t perm) { bool wrap_counter = vq->avail_wrap_counter; struct vring_packed_desc *descs = vq->desc_packed; @@ -519,15 +499,24 @@ fill_vec_buf_packed(struct virtio_net *dev, struct vhost_virtqueue *vq, if (avail_idx < vq->last_avail_idx) wrap_counter ^= 1; + /* + * Perform a load-acquire barrier in desc_is_avail to + * enforce the ordering between desc flags and desc + * content. + */ if (unlikely(!desc_is_avail(&descs[avail_idx], wrap_counter))) return -1; *desc_count = 0; + *len = 0; while (1) { if (unlikely(vec_id >= BUF_VECTOR_MAX)) return -1; + if (unlikely(*desc_count >= vq->size)) + return -1; + *desc_count += 1; *buf_id = descs[avail_idx].id; @@ -575,23 +564,31 @@ reserve_avail_buf_packed(struct virtio_net *dev, struct vhost_virtqueue *vq, uint16_t max_tries, tries = 0; uint16_t buf_id = 0; - uint16_t len = 0; + uint32_t len = 0; uint16_t desc_count; *num_buffers = 0; avail_idx = vq->last_avail_idx; if (rxvq_is_mergeable(dev)) - max_tries = vq->size; + max_tries = vq->size - 1; else max_tries = 1; while (size > 0) { + /* + * if we tried all available ring items, and still + * can't get enough buf, it means something abnormal + * happened. + */ + if (unlikely(++tries > max_tries)) + return -1; + if (unlikely(fill_vec_buf_packed(dev, vq, avail_idx, &desc_count, buf_vec, &vec_idx, &buf_id, &len, - VHOST_ACCESS_RO) < 0)) + VHOST_ACCESS_RW) < 0)) return -1; len = RTE_MIN(len, size); @@ -603,16 +600,7 @@ reserve_avail_buf_packed(struct virtio_net *dev, struct vhost_virtqueue *vq, avail_idx -= vq->size; *nr_descs += desc_count; - tries++; *num_buffers += 1; - - /* - * if we tried all available ring items, and still - * can't get enough buf, it means something abnormal - * happened. - */ - if (unlikely(tries > max_tries)) - return -1; } *nr_vec = vec_idx; @@ -620,6 +608,36 @@ reserve_avail_buf_packed(struct virtio_net *dev, struct vhost_virtqueue *vq, return 0; } +static __rte_noinline void +copy_vnet_hdr_to_desc(struct virtio_net *dev, struct vhost_virtqueue *vq, + struct buf_vector *buf_vec, + struct virtio_net_hdr_mrg_rxbuf *hdr) +{ + uint64_t len; + uint64_t remain = dev->vhost_hlen; + uint64_t src = (uint64_t)(uintptr_t)hdr, dst; + uint64_t iova = buf_vec->buf_iova; + + while (remain) { + len = RTE_MIN(remain, + buf_vec->buf_len); + dst = buf_vec->buf_addr; + rte_memcpy((void *)(uintptr_t)dst, + (void *)(uintptr_t)src, + len); + + PRINT_PACKET(dev, (uintptr_t)dst, + (uint32_t)len, 0); + vhost_log_cache_write_iova(dev, vq, + iova, len); + + remain -= len; + iova += len; + src += len; + buf_vec++; + } +} + static __rte_always_inline int copy_mbuf_to_desc(struct virtio_net *dev, struct vhost_virtqueue *vq, struct rte_mbuf *m, struct buf_vector *buf_vec, @@ -645,9 +663,6 @@ copy_mbuf_to_desc(struct virtio_net *dev, struct vhost_virtqueue *vq, buf_iova = buf_vec[vec_idx].buf_iova; buf_len = buf_vec[vec_idx].buf_len; - if (nr_vec > 1) - rte_prefetch0((void *)(uintptr_t)buf_vec[1].buf_addr); - if (unlikely(buf_len < dev->vhost_hlen && nr_vec <= 1)) { error = -1; goto out; @@ -690,10 +705,6 @@ copy_mbuf_to_desc(struct virtio_net *dev, struct vhost_virtqueue *vq, buf_iova = buf_vec[vec_idx].buf_iova; buf_len = buf_vec[vec_idx].buf_len; - /* Prefetch next buffer address. */ - if (vec_idx + 1 < nr_vec) - rte_prefetch0((void *)(uintptr_t) - buf_vec[vec_idx + 1].buf_addr); buf_offset = 0; buf_avail = buf_len; } @@ -713,33 +724,11 @@ copy_mbuf_to_desc(struct virtio_net *dev, struct vhost_virtqueue *vq, num_buffers); if (unlikely(hdr == &tmp_hdr)) { - uint64_t len; - uint64_t remain = dev->vhost_hlen; - uint64_t src = (uint64_t)(uintptr_t)hdr, dst; - uint64_t iova = buf_vec[0].buf_iova; - uint16_t hdr_vec_idx = 0; - - while (remain) { - len = remain; - dst = buf_vec[hdr_vec_idx].buf_addr; - rte_memcpy((void *)(uintptr_t)dst, - (void *)(uintptr_t)src, - len); - - PRINT_PACKET(dev, (uintptr_t)dst, - (uint32_t)len, 0); - vhost_log_cache_write(dev, vq, - iova, len); - - remain -= len; - iova += len; - src += len; - hdr_vec_idx++; - } + copy_vnet_hdr_to_desc(dev, vq, buf_vec, hdr); } else { PRINT_PACKET(dev, (uintptr_t)hdr_addr, dev->vhost_hlen, 0); - vhost_log_cache_write(dev, vq, + vhost_log_cache_write_iova(dev, vq, buf_vec[0].buf_iova, dev->vhost_hlen); } @@ -747,15 +736,16 @@ copy_mbuf_to_desc(struct virtio_net *dev, struct vhost_virtqueue *vq, hdr_addr = 0; } - cpy_len = RTE_MIN(buf_len, mbuf_avail); + cpy_len = RTE_MIN(buf_avail, mbuf_avail); if (likely(cpy_len > MAX_BATCH_LEN || vq->batch_copy_nb_elems >= vq->size)) { rte_memcpy((void *)((uintptr_t)(buf_addr + buf_offset)), rte_pktmbuf_mtod_offset(m, void *, mbuf_offset), cpy_len); - vhost_log_cache_write(dev, vq, buf_iova + buf_offset, - cpy_len); + vhost_log_cache_write_iova(dev, vq, + buf_iova + buf_offset, + cpy_len); PRINT_PACKET(dev, (uintptr_t)(buf_addr + buf_offset), cpy_len, 0); } else { @@ -780,7 +770,61 @@ out: return error; } -static __rte_always_inline uint32_t +static __rte_always_inline int +vhost_enqueue_single_packed(struct virtio_net *dev, + struct vhost_virtqueue *vq, + struct rte_mbuf *pkt, + struct buf_vector *buf_vec, + uint16_t *nr_descs) +{ + uint16_t nr_vec = 0; + uint16_t avail_idx = vq->last_avail_idx; + uint16_t max_tries, tries = 0; + uint16_t buf_id = 0; + uint32_t len = 0; + uint16_t desc_count; + uint32_t size = pkt->pkt_len + dev->vhost_hlen; + uint16_t num_buffers = 0; + + if (rxvq_is_mergeable(dev)) + max_tries = vq->size - 1; + else + max_tries = 1; + + while (size > 0) { + /* + * if we tried all available ring items, and still + * can't get enough buf, it means something abnormal + * happened. + */ + if (unlikely(++tries > max_tries)) + return -1; + + if (unlikely(fill_vec_buf_packed(dev, vq, + avail_idx, &desc_count, + buf_vec, &nr_vec, + &buf_id, &len, + VHOST_ACCESS_RW) < 0)) + return -1; + + len = RTE_MIN(len, size); + size -= len; + + num_buffers += 1; + + *nr_descs += desc_count; + avail_idx += desc_count; + if (avail_idx >= vq->size) + avail_idx -= vq->size; + } + + if (copy_mbuf_to_desc(dev, vq, pkt, buf_vec, nr_vec, num_buffers) < 0) + return -1; + + return 0; +} + +static __rte_noinline uint32_t virtio_dev_rx_split(struct virtio_net *dev, struct vhost_virtqueue *vq, struct rte_mbuf **pkts, uint32_t count) { @@ -789,9 +833,16 @@ virtio_dev_rx_split(struct virtio_net *dev, struct vhost_virtqueue *vq, struct buf_vector buf_vec[BUF_VECTOR_MAX]; uint16_t avail_head; - rte_prefetch0(&vq->avail->ring[vq->last_avail_idx & (vq->size - 1)]); avail_head = *((volatile uint16_t *)&vq->avail->idx); + /* + * The ordering between avail index and + * desc reads needs to be enforced. + */ + rte_smp_rmb(); + + rte_prefetch0(&vq->avail->ring[vq->last_avail_idx & (vq->size - 1)]); + for (pkt_idx = 0; pkt_idx < count; pkt_idx++) { uint32_t pkt_len = pkts[pkt_idx]->pkt_len + dev->vhost_hlen; uint16_t nr_vec = 0; @@ -806,8 +857,6 @@ virtio_dev_rx_split(struct virtio_net *dev, struct vhost_virtqueue *vq, break; } - rte_prefetch0((void *)(uintptr_t)buf_vec[0].buf_addr); - VHOST_LOG_DEBUG(VHOST_DATA, "(%d) current index %d | end index %d\n", dev->vid, vq->last_avail_idx, vq->last_avail_idx + num_buffers); @@ -832,7 +881,103 @@ virtio_dev_rx_split(struct virtio_net *dev, struct vhost_virtqueue *vq, return pkt_idx; } -static __rte_always_inline uint32_t +static __rte_unused int +virtio_dev_rx_batch_packed(struct virtio_net *dev, + struct vhost_virtqueue *vq, + struct rte_mbuf **pkts) +{ + bool wrap_counter = vq->avail_wrap_counter; + struct vring_packed_desc *descs = vq->desc_packed; + 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; + uint64_t lens[PACKED_BATCH_SIZE]; + uint16_t i; + + if (unlikely(avail_idx & PACKED_BATCH_MASK)) + return -1; + + if (unlikely((avail_idx + PACKED_BATCH_SIZE) > vq->size)) + return -1; + + vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) { + if (unlikely(pkts[i]->next != NULL)) + return -1; + if (unlikely(!desc_is_avail(&descs[avail_idx + i], + wrap_counter))) + return -1; + } + + rte_smp_rmb(); + + vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) + lens[i] = descs[avail_idx + i].len; + + vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) { + if (unlikely(pkts[i]->pkt_len > (lens[i] - buf_offset))) + return -1; + } + + vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) + desc_addrs[i] = vhost_iova_to_vva(dev, vq, + descs[avail_idx + i].addr, + &lens[i], + VHOST_ACCESS_RW); + + vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) { + if (unlikely(lens[i] != descs[avail_idx + i].len)) + return -1; + } + + vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) { + 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; + } + + vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) + virtio_enqueue_offload(pkts[i], &hdrs[i]->hdr); + + vq_inc_last_avail_packed(vq, PACKED_BATCH_SIZE); + + vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) { + rte_memcpy((void *)(uintptr_t)(desc_addrs[i] + buf_offset), + rte_pktmbuf_mtod_offset(pkts[i], void *, 0), + pkts[i]->pkt_len); + } + + return 0; +} + +static __rte_unused int16_t +virtio_dev_rx_single_packed(struct virtio_net *dev, + struct vhost_virtqueue *vq, + struct rte_mbuf *pkt) +{ + 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_DEBUG(VHOST_DATA, + "(%d) failed to get enough desc from vring\n", + dev->vid); + return -1; + } + + VHOST_LOG_DEBUG(VHOST_DATA, "(%d) current index %d | end index %d\n", + dev->vid, vq->last_avail_idx, + vq->last_avail_idx + nr_descs); + + vq_inc_last_avail_packed(vq, nr_descs); + + return 0; +} + +static __rte_noinline uint32_t virtio_dev_rx_packed(struct virtio_net *dev, struct vhost_virtqueue *vq, struct rte_mbuf **pkts, uint32_t count) { @@ -855,8 +1000,6 @@ virtio_dev_rx_packed(struct virtio_net *dev, struct vhost_virtqueue *vq, break; } - rte_prefetch0((void *)(uintptr_t)buf_vec[0].buf_addr); - VHOST_LOG_DEBUG(VHOST_DATA, "(%d) current index %d | end index %d\n", dev->vid, vq->last_avail_idx, vq->last_avail_idx + num_buffers); @@ -868,11 +1011,7 @@ virtio_dev_rx_packed(struct virtio_net *dev, struct vhost_virtqueue *vq, break; } - vq->last_avail_idx += nr_descs; - if (vq->last_avail_idx >= vq->size) { - vq->last_avail_idx -= vq->size; - vq->avail_wrap_counter ^= 1; - } + vq_inc_last_avail_packed(vq, nr_descs); } do_data_copy_enqueue(dev, vq); @@ -890,6 +1029,7 @@ virtio_dev_rx(struct virtio_net *dev, uint16_t queue_id, struct rte_mbuf **pkts, uint32_t count) { struct vhost_virtqueue *vq; + uint32_t nb_tx = 0; VHOST_LOG_DEBUG(VHOST_DATA, "(%d) %s\n", dev->vid, __func__); if (unlikely(!is_valid_virt_queue_idx(queue_id, 0, dev->nr_vring))) { @@ -917,9 +1057,9 @@ virtio_dev_rx(struct virtio_net *dev, uint16_t queue_id, goto out; if (vq_is_packed(dev)) - count = virtio_dev_rx_packed(dev, vq, pkts, count); + nb_tx = virtio_dev_rx_packed(dev, vq, pkts, count); else - count = virtio_dev_rx_split(dev, vq, pkts, count); + nb_tx = virtio_dev_rx_split(dev, vq, pkts, count); out: if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM)) @@ -928,7 +1068,7 @@ out: out_access_unlock: rte_spinlock_unlock(&vq->access_lock); - return count; + return nb_tx; } uint16_t @@ -967,38 +1107,39 @@ virtio_net_with_host_offload(struct virtio_net *dev) static void parse_ethernet(struct rte_mbuf *m, uint16_t *l4_proto, void **l4_hdr) { - struct ipv4_hdr *ipv4_hdr; - struct ipv6_hdr *ipv6_hdr; + struct rte_ipv4_hdr *ipv4_hdr; + struct rte_ipv6_hdr *ipv6_hdr; void *l3_hdr = NULL; - struct ether_hdr *eth_hdr; + struct rte_ether_hdr *eth_hdr; uint16_t ethertype; - eth_hdr = rte_pktmbuf_mtod(m, struct ether_hdr *); + eth_hdr = rte_pktmbuf_mtod(m, struct rte_ether_hdr *); - m->l2_len = sizeof(struct ether_hdr); + m->l2_len = sizeof(struct rte_ether_hdr); ethertype = rte_be_to_cpu_16(eth_hdr->ether_type); - if (ethertype == ETHER_TYPE_VLAN) { - struct vlan_hdr *vlan_hdr = (struct vlan_hdr *)(eth_hdr + 1); + if (ethertype == RTE_ETHER_TYPE_VLAN) { + struct rte_vlan_hdr *vlan_hdr = + (struct rte_vlan_hdr *)(eth_hdr + 1); - m->l2_len += sizeof(struct vlan_hdr); + m->l2_len += sizeof(struct rte_vlan_hdr); ethertype = rte_be_to_cpu_16(vlan_hdr->eth_proto); } l3_hdr = (char *)eth_hdr + m->l2_len; switch (ethertype) { - case ETHER_TYPE_IPv4: + 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; *l4_hdr = (char *)l3_hdr + m->l3_len; m->ol_flags |= PKT_TX_IPV4; break; - case ETHER_TYPE_IPv6: + case RTE_ETHER_TYPE_IPV6: ipv6_hdr = l3_hdr; *l4_proto = ipv6_hdr->proto; - m->l3_len = sizeof(struct ipv6_hdr); + m->l3_len = sizeof(struct rte_ipv6_hdr); *l4_hdr = (char *)l3_hdr + m->l3_len; m->ol_flags |= PKT_TX_IPV6; break; @@ -1015,7 +1156,7 @@ vhost_dequeue_offload(struct virtio_net_hdr *hdr, struct rte_mbuf *m) { uint16_t l4_proto = 0; void *l4_hdr = NULL; - struct tcp_hdr *tcp_hdr = NULL; + struct rte_tcp_hdr *tcp_hdr = NULL; if (hdr->flags == 0 && hdr->gso_type == VIRTIO_NET_HDR_GSO_NONE) return; @@ -1024,15 +1165,15 @@ vhost_dequeue_offload(struct virtio_net_hdr *hdr, struct rte_mbuf *m) if (hdr->flags == VIRTIO_NET_HDR_F_NEEDS_CSUM) { if (hdr->csum_start == (m->l2_len + m->l3_len)) { switch (hdr->csum_offset) { - case (offsetof(struct tcp_hdr, cksum)): + case (offsetof(struct rte_tcp_hdr, cksum)): if (l4_proto == IPPROTO_TCP) m->ol_flags |= PKT_TX_TCP_CKSUM; break; - case (offsetof(struct udp_hdr, dgram_cksum)): + case (offsetof(struct rte_udp_hdr, dgram_cksum)): if (l4_proto == IPPROTO_UDP) m->ol_flags |= PKT_TX_UDP_CKSUM; break; - case (offsetof(struct sctp_hdr, cksum)): + case (offsetof(struct rte_sctp_hdr, cksum)): if (l4_proto == IPPROTO_SCTP) m->ol_flags |= PKT_TX_SCTP_CKSUM; break; @@ -1054,7 +1195,7 @@ vhost_dequeue_offload(struct virtio_net_hdr *hdr, struct rte_mbuf *m) case VIRTIO_NET_HDR_GSO_UDP: m->ol_flags |= PKT_TX_UDP_SEG; m->tso_segsz = hdr->gso_size; - m->l4_len = sizeof(struct udp_hdr); + m->l4_len = sizeof(struct rte_udp_hdr); break; default: RTE_LOG(WARNING, VHOST_DATA, @@ -1064,10 +1205,25 @@ vhost_dequeue_offload(struct virtio_net_hdr *hdr, struct rte_mbuf *m) } } -static __rte_always_inline void -put_zmbuf(struct zcopy_mbuf *zmbuf) +static __rte_noinline void +copy_vnet_hdr_from_desc(struct virtio_net_hdr *hdr, + struct buf_vector *buf_vec) { - zmbuf->in_use = 0; + uint64_t len; + uint64_t remain = sizeof(struct virtio_net_hdr); + uint64_t src; + uint64_t dst = (uint64_t)(uintptr_t)hdr; + + while (remain) { + len = RTE_MIN(remain, buf_vec->buf_len); + src = buf_vec->buf_addr; + rte_memcpy((void *)(uintptr_t)dst, + (void *)(uintptr_t)src, len); + + remain -= len; + dst += len; + buf_vec++; + } } static __rte_always_inline int @@ -1096,36 +1252,16 @@ copy_desc_to_mbuf(struct virtio_net *dev, struct vhost_virtqueue *vq, goto out; } - if (likely(nr_vec > 1)) - rte_prefetch0((void *)(uintptr_t)buf_vec[1].buf_addr); - if (virtio_net_with_host_offload(dev)) { if (unlikely(buf_len < sizeof(struct virtio_net_hdr))) { - uint64_t len; - uint64_t remain = sizeof(struct virtio_net_hdr); - uint64_t src; - uint64_t dst = (uint64_t)(uintptr_t)&tmp_hdr; - uint16_t hdr_vec_idx = 0; - /* * No luck, the virtio-net header doesn't fit * in a contiguous virtual area. */ - while (remain) { - len = remain; - src = buf_vec[hdr_vec_idx].buf_addr; - rte_memcpy((void *)(uintptr_t)dst, - (void *)(uintptr_t)src, len); - - remain -= len; - dst += len; - hdr_vec_idx++; - } - + copy_vnet_hdr_from_desc(&tmp_hdr, buf_vec); hdr = &tmp_hdr; } else { hdr = (struct virtio_net_hdr *)((uintptr_t)buf_addr); - rte_prefetch0(hdr); } } @@ -1155,9 +1291,6 @@ copy_desc_to_mbuf(struct virtio_net *dev, struct vhost_virtqueue *vq, buf_avail = buf_vec[vec_idx].buf_len - dev->vhost_hlen; } - rte_prefetch0((void *)(uintptr_t) - (buf_addr + buf_offset)); - PRINT_PACKET(dev, (uintptr_t)(buf_addr + buf_offset), (uint32_t)buf_avail, 0); @@ -1223,14 +1356,6 @@ copy_desc_to_mbuf(struct virtio_net *dev, struct vhost_virtqueue *vq, buf_iova = buf_vec[vec_idx].buf_iova; buf_len = buf_vec[vec_idx].buf_len; - /* - * Prefecth desc n + 1 buffer while - * desc n buffer is processed. - */ - if (vec_idx + 1 < nr_vec) - rte_prefetch0((void *)(uintptr_t) - buf_vec[vec_idx + 1].buf_addr); - buf_offset = 0; buf_avail = buf_len; @@ -1306,35 +1431,94 @@ again: return NULL; } -static __rte_always_inline bool -mbuf_is_consumed(struct rte_mbuf *m) +static void +virtio_dev_extbuf_free(void *addr __rte_unused, void *opaque) +{ + rte_free(opaque); +} + +static int +virtio_dev_extbuf_alloc(struct rte_mbuf *pkt, uint32_t size) { - while (m) { - if (rte_mbuf_refcnt_read(m) > 1) - return false; - m = m->next; + struct rte_mbuf_ext_shared_info *shinfo = NULL; + uint32_t total_len = RTE_PKTMBUF_HEADROOM + size; + uint16_t buf_len; + 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)); + } + + if (unlikely(total_len > UINT16_MAX)) + return -ENOSPC; + + buf_len = total_len; + buf = rte_malloc(NULL, buf_len, RTE_CACHE_LINE_SIZE); + if (unlikely(buf == NULL)) + 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); + RTE_LOG(ERR, VHOST_DATA, "Failed to init shinfo\n"); + return -1; + } } - return true; + iova = rte_malloc_virt2iova(buf); + rte_pktmbuf_attach_extbuf(pkt, buf, iova, buf_len, shinfo); + rte_pktmbuf_reset_headroom(pkt); + + return 0; } -static __rte_always_inline void -restore_mbuf(struct rte_mbuf *m) +/* + * Allocate a host supported pktmbuf. + */ +static __rte_always_inline struct rte_mbuf * +virtio_dev_pktmbuf_alloc(struct virtio_net *dev, struct rte_mempool *mp, + uint32_t data_len) { - uint32_t mbuf_size, priv_size; + struct rte_mbuf *pkt = rte_pktmbuf_alloc(mp); - while (m) { - priv_size = rte_pktmbuf_priv_size(m->pool); - mbuf_size = sizeof(struct rte_mbuf) + priv_size; - /* start of buffer is after mbuf structure and priv data */ + if (unlikely(pkt == NULL)) + return NULL; - m->buf_addr = (char *)m + mbuf_size; - m->buf_iova = rte_mempool_virt2iova(m) + mbuf_size; - m = m->next; - } + if (rte_pktmbuf_tailroom(pkt) >= data_len) + return pkt; + + /* attach an external buffer if supported */ + if (dev->extbuf && !virtio_dev_extbuf_alloc(pkt, data_len)) + return pkt; + + /* check if chained buffers are allowed */ + if (!dev->linearbuf) + return pkt; + + /* Data doesn't fit into the buffer and the host supports + * only linear buffers + */ + rte_pktmbuf_free(pkt); + + return NULL; } -static __rte_always_inline uint16_t +static __rte_noinline uint16_t virtio_dev_tx_split(struct virtio_net *dev, struct vhost_virtqueue *vq, struct rte_mempool *mbuf_pool, struct rte_mbuf **pkts, uint16_t count) { @@ -1343,7 +1527,6 @@ virtio_dev_tx_split(struct virtio_net *dev, struct vhost_virtqueue *vq, if (unlikely(dev->dequeue_zero_copy)) { struct zcopy_mbuf *zmbuf, *next; - int nr_updated = 0; for (zmbuf = TAILQ_FIRST(&vq->zmbuf_list); zmbuf != NULL; zmbuf = next) { @@ -1352,8 +1535,6 @@ virtio_dev_tx_split(struct virtio_net *dev, struct vhost_virtqueue *vq, if (mbuf_is_consumed(zmbuf->mbuf)) { update_shadow_used_ring_split(vq, zmbuf->desc_idx, 0); - nr_updated += 1; - TAILQ_REMOVE(&vq->zmbuf_list, zmbuf, next); restore_mbuf(zmbuf->mbuf); rte_pktmbuf_free(zmbuf->mbuf); @@ -1362,17 +1543,25 @@ virtio_dev_tx_split(struct virtio_net *dev, struct vhost_virtqueue *vq, } } - flush_shadow_used_ring_split(dev, vq); - vhost_vring_call_split(dev, vq); + if (likely(vq->shadow_used_idx)) { + flush_shadow_used_ring_split(dev, vq); + vhost_vring_call_split(dev, vq); + } } - rte_prefetch0(&vq->avail->ring[vq->last_avail_idx & (vq->size - 1)]); - free_entries = *((volatile uint16_t *)&vq->avail->idx) - vq->last_avail_idx; if (free_entries == 0) return 0; + /* + * The ordering between avail index and + * desc reads needs to be enforced. + */ + rte_smp_rmb(); + + rte_prefetch0(&vq->avail->ring[vq->last_avail_idx & (vq->size - 1)]); + VHOST_LOG_DEBUG(VHOST_DATA, "(%d) %s\n", dev->vid, __func__); count = RTE_MIN(count, MAX_PKT_BURST); @@ -1382,28 +1571,24 @@ virtio_dev_tx_split(struct virtio_net *dev, struct vhost_virtqueue *vq, for (i = 0; i < count; i++) { struct buf_vector buf_vec[BUF_VECTOR_MAX]; - uint16_t head_idx, dummy_len; + uint16_t head_idx; + uint32_t buf_len; uint16_t nr_vec = 0; int err; if (unlikely(fill_vec_buf_split(dev, vq, vq->last_avail_idx + i, &nr_vec, buf_vec, - &head_idx, &dummy_len, + &head_idx, &buf_len, VHOST_ACCESS_RO) < 0)) break; if (likely(dev->dequeue_zero_copy == 0)) update_shadow_used_ring_split(vq, head_idx, 0); - rte_prefetch0((void *)(uintptr_t)buf_vec[0].buf_addr); - - pkts[i] = rte_pktmbuf_alloc(mbuf_pool); - if (unlikely(pkts[i] == NULL)) { - RTE_LOG(ERR, VHOST_DATA, - "Failed to allocate memory for mbuf.\n"); + pkts[i] = virtio_dev_pktmbuf_alloc(dev, mbuf_pool, buf_len); + if (unlikely(pkts[i] == NULL)) break; - } err = copy_desc_to_mbuf(dev, vq, buf_vec, nr_vec, pkts[i], mbuf_pool); @@ -1441,21 +1626,21 @@ virtio_dev_tx_split(struct virtio_net *dev, struct vhost_virtqueue *vq, do_data_copy_dequeue(vq); if (unlikely(i < count)) vq->shadow_used_idx = i; - flush_shadow_used_ring_split(dev, vq); - vhost_vring_call_split(dev, vq); + if (likely(vq->shadow_used_idx)) { + flush_shadow_used_ring_split(dev, vq); + vhost_vring_call_split(dev, vq); + } } return i; } -static __rte_always_inline uint16_t +static __rte_noinline uint16_t virtio_dev_tx_packed(struct virtio_net *dev, struct vhost_virtqueue *vq, struct rte_mempool *mbuf_pool, struct rte_mbuf **pkts, uint16_t count) { uint16_t i; - rte_prefetch0(&vq->desc_packed[vq->last_avail_idx]); - if (unlikely(dev->dequeue_zero_copy)) { struct zcopy_mbuf *zmbuf, *next; @@ -1477,8 +1662,10 @@ virtio_dev_tx_packed(struct virtio_net *dev, struct vhost_virtqueue *vq, } } - flush_shadow_used_ring_packed(dev, vq); - vhost_vring_call_packed(dev, vq); + if (likely(vq->shadow_used_idx)) { + flush_shadow_used_ring_packed(dev, vq); + vhost_vring_call_packed(dev, vq); + } } VHOST_LOG_DEBUG(VHOST_DATA, "(%d) %s\n", dev->vid, __func__); @@ -1489,29 +1676,25 @@ virtio_dev_tx_packed(struct virtio_net *dev, struct vhost_virtqueue *vq, for (i = 0; i < count; i++) { struct buf_vector buf_vec[BUF_VECTOR_MAX]; - uint16_t buf_id, dummy_len; + uint16_t buf_id; + uint32_t buf_len; uint16_t desc_count, nr_vec = 0; int err; if (unlikely(fill_vec_buf_packed(dev, vq, vq->last_avail_idx, &desc_count, buf_vec, &nr_vec, - &buf_id, &dummy_len, - VHOST_ACCESS_RW) < 0)) + &buf_id, &buf_len, + VHOST_ACCESS_RO) < 0)) break; if (likely(dev->dequeue_zero_copy == 0)) update_shadow_used_ring_packed(vq, buf_id, 0, desc_count); - rte_prefetch0((void *)(uintptr_t)buf_vec[0].buf_addr); - - pkts[i] = rte_pktmbuf_alloc(mbuf_pool); - if (unlikely(pkts[i] == NULL)) { - RTE_LOG(ERR, VHOST_DATA, - "Failed to allocate memory for mbuf.\n"); + pkts[i] = virtio_dev_pktmbuf_alloc(dev, mbuf_pool, buf_len); + if (unlikely(pkts[i] == NULL)) break; - } err = copy_desc_to_mbuf(dev, vq, buf_vec, nr_vec, pkts[i], mbuf_pool); @@ -1544,19 +1727,17 @@ virtio_dev_tx_packed(struct virtio_net *dev, struct vhost_virtqueue *vq, TAILQ_INSERT_TAIL(&vq->zmbuf_list, zmbuf, next); } - vq->last_avail_idx += desc_count; - if (vq->last_avail_idx >= vq->size) { - vq->last_avail_idx -= vq->size; - vq->avail_wrap_counter ^= 1; - } + vq_inc_last_avail_packed(vq, desc_count); } if (likely(dev->dequeue_zero_copy == 0)) { do_data_copy_dequeue(vq); if (unlikely(i < count)) vq->shadow_used_idx = i; - flush_shadow_used_ring_packed(dev, vq); - vhost_vring_call_packed(dev, vq); + if (likely(vq->shadow_used_idx)) { + flush_shadow_used_ring_packed(dev, vq); + vhost_vring_call_packed(dev, vq); + } } return i; @@ -1592,15 +1773,19 @@ rte_vhost_dequeue_burst(int vid, uint16_t queue_id, if (unlikely(rte_spinlock_trylock(&vq->access_lock) == 0)) return 0; - if (unlikely(vq->enabled == 0)) + if (unlikely(vq->enabled == 0)) { + count = 0; goto out_access_unlock; + } if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM)) vhost_user_iotlb_rd_lock(vq); if (unlikely(vq->access_ok == 0)) - if (unlikely(vring_translate(dev, vq) < 0)) + if (unlikely(vring_translate(dev, vq) < 0)) { + count = 0; goto out; + } /* * Construct a RARP broadcast packet, and inject it to the "pkts" @@ -1625,7 +1810,8 @@ rte_vhost_dequeue_burst(int vid, uint16_t queue_id, if (rarp_mbuf == NULL) { RTE_LOG(ERR, VHOST_DATA, "Failed to make RARP packet.\n"); - return 0; + count = 0; + goto out; } count -= 1; }