This patch implements such an egress function on the Vhost lib. When
data is received from a front-end, it will also send the data to its
mirrorfront-end.
Signed-off-by: Cheng Jiang <cheng1.ji...@intel.com>
Signed-off-by: Wenwu Ma <wenwux...@intel.com>
---
lib/vhost/rte_vhost_async.h | 12 +-
lib/vhost/version.map | 3 +
lib/vhost/virtio_net.c | 729 ++++++++++++++++++++++++++++++++++++
3 files changed, 742 insertions(+), 2 deletions(-)
diff --git a/lib/vhost/rte_vhost_async.h b/lib/vhost/rte_vhost_async.h
index b199af078c..b0772c2db7 100644
--- a/lib/vhost/rte_vhost_async.h
+++ b/lib/vhost/rte_vhost_async.h
@@ -85,11 +85,12 @@ struct rte_vhost_async_channel_ops {
};
/**
- * inflight async packet information
+ * in-flight async packet information
*/
struct async_inflight_info {
struct rte_mbuf *mbuf;
- uint16_t descs; /* num of descs inflight */
+ struct virtio_net_hdr nethdr;
+ uint16_t descs; /* num of descs in-flight */
uint16_t nr_buffers; /* num of buffers inflight for packed ring */
};
@@ -268,4 +269,11 @@ __rte_experimental
uint16_t rte_vhost_clear_queue_thread_unsafe(int vid, uint16_t queue_id,
struct rte_mbuf **pkts, uint16_t count);
+__rte_experimental
+uint16_t
+rte_vhost_async_try_egress_burst(int vid, uint16_t queue_id,
+ int mirr_vid, uint16_t mirr_queue_id,
+ struct rte_mempool *mbuf_pool, struct rte_mbuf **pkts, uint16_t count,
+ int *nr_inflight);
+
#endif /* _RTE_VHOST_ASYNC_H_ */
diff --git a/lib/vhost/version.map b/lib/vhost/version.map
index 4c35fa4555..2529943a91 100644
--- a/lib/vhost/version.map
+++ b/lib/vhost/version.map
@@ -87,4 +87,7 @@ EXPERIMENTAL {
rte_vhost_async_channel_register_thread_unsafe;
rte_vhost_async_channel_unregister_thread_unsafe;
rte_vhost_clear_queue_thread_unsafe;
+
+ # added in 21.11
+ rte_vhost_async_try_egress_burst;
};
diff --git a/lib/vhost/virtio_net.c b/lib/vhost/virtio_net.c
index c9f0bb22e5..92074377b2 100644
--- a/lib/vhost/virtio_net.c
+++ b/lib/vhost/virtio_net.c
@@ -3827,3 +3827,732 @@ rte_vhost_dequeue_burst(int vid, uint16_t queue_id,
return count;
}
+
+static __rte_always_inline uint16_t
+async_poll_egress_completed_split(struct virtio_net *dev,
+ struct vhost_virtqueue *vq, uint16_t queue_id,
+ struct virtio_net *mirr_dev, struct vhost_virtqueue *mirr_vq,
+ struct rte_mbuf **pkts, uint16_t count, bool legacy_ol_flags)
+{
+ uint16_t n_pkts_cpl = 0, n_pkts_put = 0;
+ uint16_t start_idx, pkt_idx, from;
+ struct async_inflight_info *pkts_info;
+ uint16_t mirror_n_pkts_cpl = 0, n_descs = 0;
+
+ pkt_idx = vq->async_pkts_idx & (vq->size - 1);
+ pkts_info = vq->async_pkts_info;
+ start_idx = virtio_dev_rx_async_get_info_idx(pkt_idx, vq->size,
+ vq->async_pkts_inflight_n);
+
+ if (count > vq->async_last_pkts_n) {
+ int ret;
+
+ ret = vq->async_ops.check_completed_copies(dev->vid, queue_id,
+ 0, count - vq->async_last_pkts_n);
+ if (unlikely(ret < 0)) {
+ VHOST_LOG_DATA(ERR, "(%d) async channel poll error\n",
dev->vid);
+ ret = 0;
+ }
+ n_pkts_cpl = ret;
+ }
+
+ n_pkts_cpl += vq->async_last_pkts_n;
+ mirror_n_pkts_cpl += mirr_vq->async_last_pkts_n;
+ if (unlikely(n_pkts_cpl == 0)) {
+ mirr_vq->async_last_pkts_n = mirror_n_pkts_cpl;
+ return 0;
+ }
+
+ n_pkts_put = RTE_MIN(count, n_pkts_cpl);
+
+ for (pkt_idx = 0; pkt_idx < n_pkts_put; pkt_idx++) {
+ from = (start_idx + pkt_idx) & (vq->size - 1);
+ pkts[pkt_idx] = pkts_info[from].mbuf;
+ n_descs += pkts_info[from].descs;
+
+ if (virtio_net_with_host_offload(dev))
+ vhost_dequeue_offload(&pkts_info[from].nethdr,
+ pkts[pkt_idx], legacy_ol_flags);
+ }
+
+ /* write back completed descs to used ring and update used idx */
+ write_back_completed_descs_split(vq, n_pkts_put);
+ __atomic_add_fetch(&vq->used->idx, n_pkts_put, __ATOMIC_RELEASE);
+ vhost_vring_call_split(dev, vq);
+
+ vq->async_last_pkts_n = n_pkts_cpl - n_pkts_put;
+ vq->async_pkts_inflight_n -= n_pkts_put;
+
+ if (likely(mirr_vq->enabled && mirr_vq->access_ok)) {
+ write_back_completed_descs_split(mirr_vq, n_descs);
+
+ __atomic_add_fetch(&mirr_vq->used->idx, n_descs,
+ __ATOMIC_RELEASE);
+ vhost_vring_call_split(mirr_dev, mirr_vq);
+ } else {
+ mirr_vq->last_async_desc_idx_split += n_descs;
+ }
+
+ return n_pkts_put;
+}
+
+static __rte_always_inline void
+egress_async_fill_desc(struct rte_vhost_async_desc *desc,
+ struct rte_vhost_iov_iter *src, struct rte_vhost_iov_iter *dst,
+ struct rte_vhost_iov_iter *mirror_dst)
+{
+ desc->src = src;
+ desc->dst = dst;
+ desc->mirror_dst = mirror_dst;
+}
+
+static __rte_always_inline int
+egress_async_desc_to_mbuf(struct virtio_net *dev,
+ struct buf_vector *buf_vec, uint16_t nr_vec,
+ struct rte_mbuf *m, struct rte_mempool *mbuf_pool,
+ struct iovec *src_iovec, struct iovec *dst_iovec,
+ struct rte_vhost_iov_iter *src_it,
+ struct rte_vhost_iov_iter *dst_it,
+ struct virtio_net *mirr_dev, struct vhost_virtqueue *mirr_vq
__rte_unused,
+ struct buf_vector *mirr_buf_vec, uint16_t mirr_nr_vec,
uint16_t mirr_num_buffers,
+ struct iovec *mirr_dst_iovec, struct rte_vhost_iov_iter
*mirr_dst_it,
+ struct virtio_net_hdr *nethdr,
+ int nr_iovec)
+{
+ uint64_t buf_addr, buf_iova;
+ uint64_t mapped_len;
+ uint32_t tlen = 0;
+ uint32_t buf_avail, buf_offset, buf_len;
+ uint32_t mbuf_avail, mbuf_offset;
+ uint32_t cpy_len;
+ /* A counter to avoid desc dead loop chain */
+ uint16_t vec_idx = 0;
+ int tvec_idx = 0;
+ struct rte_mbuf *cur = m, *prev = m;
+ struct virtio_net_hdr tmp_hdr;
+ struct virtio_net_hdr *hdr = NULL;
+
+ int error = 0;
+ uint64_t mirror_mapped_len;
+ uint64_t mirror_buf_addr, mirror_buf_iova;
+ uint32_t mirror_buf_avail, mirror_buf_offset, mirror_buf_len;
+ uint32_t mirror_vec_idx = 0;
+
+ mirror_buf_addr = mirr_buf_vec[mirror_vec_idx].buf_addr;
+ mirror_buf_iova = mirr_buf_vec[mirror_vec_idx].buf_iova;
+ mirror_buf_len = mirr_buf_vec[mirror_vec_idx].buf_len;
+
+ if (unlikely(mirror_buf_len < mirr_dev->vhost_hlen && mirr_nr_vec <=
1)) {
+ error = -1;
+ goto out;
+ }
+
+ struct virtio_net_hdr_mrg_rxbuf *mirror_hdr = NULL;
+ mirror_hdr = (struct virtio_net_hdr_mrg_rxbuf
*)(uintptr_t)mirror_buf_addr;
+ if (mirror_hdr && rxvq_is_mergeable(mirr_dev))
+ ASSIGN_UNLESS_EQUAL(mirror_hdr->num_buffers, mirr_num_buffers);
+
+ if (unlikely(mirror_buf_len < mirr_dev->vhost_hlen)) {
+ mirror_buf_offset = mirr_dev->vhost_hlen - mirror_buf_len;
+ mirror_vec_idx++;
+ mirror_buf_addr = mirr_buf_vec[mirror_vec_idx].buf_addr;
+ mirror_buf_iova = mirr_buf_vec[mirror_vec_idx].buf_iova;
+ mirror_buf_len = mirr_buf_vec[mirror_vec_idx].buf_len;
+ mirror_buf_avail = mirror_buf_len - mirror_buf_offset;
+ } else {
+ mirror_buf_offset = mirr_dev->vhost_hlen;
+ mirror_buf_avail = mirror_buf_len - mirr_dev->vhost_hlen;
+ }
+
+ buf_addr = buf_vec[vec_idx].buf_addr;
+ buf_len = buf_vec[vec_idx].buf_len;
+ buf_iova = buf_vec[vec_idx].buf_iova;
+
+ if (unlikely(buf_len < dev->vhost_hlen && nr_vec <= 1))
+ return -1;
+
+ if (virtio_net_with_host_offload(dev)) {
+ if (unlikely(buf_len < sizeof(struct virtio_net_hdr))) {
+ /*
+ * No luck, the virtio-net header doesn't fit
+ * in a contiguous virtual area.
+ */
+ copy_vnet_hdr_from_desc(&tmp_hdr, buf_vec);
+ hdr = &tmp_hdr;
+ } else {
+ hdr = (struct virtio_net_hdr *)((uintptr_t)buf_addr);
+ }
+ }
+
+ /*
+ * A virtio driver normally uses at least 2 desc buffers
+ * for Tx: the first for storing the header, and others
+ * for storing the data.
+ */
+ if (unlikely(buf_len < dev->vhost_hlen)) {
+ 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))
+ return -1;
+ 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;
+ buf_avail = buf_len;
+ } else {
+ buf_offset = dev->vhost_hlen;
+ buf_avail = buf_vec[vec_idx].buf_len - dev->vhost_hlen;
+ }
+
+ PRINT_PACKET(dev, (uintptr_t)(buf_addr + buf_offset),
(uint32_t)buf_avail, 0);
+
+ mbuf_offset = 0;
+ mbuf_avail = m->buf_len - RTE_PKTMBUF_HEADROOM;
+ int flag = 0;
+ static uint64_t total = 0;
+ while (1) {
+ cpy_len = RTE_MIN(buf_avail, mbuf_avail);
+ cpy_len = RTE_MIN(cpy_len, mirror_buf_avail);
+
+ while (cpy_len) {
+ void *hpa = (void *)(uintptr_t)gpa_to_first_hpa(dev,
+ buf_iova + buf_offset, cpy_len,
+ &mapped_len);
+ if (unlikely(!hpa)) {
+ VHOST_LOG_DATA(ERR, "(%d) %s: failed to get
hpa.\n",
+ dev->vid, __func__);
+ return -1;
+ }
+
+ void *mirror_hpa = (void
*)(uintptr_t)gpa_to_first_hpa(mirr_dev,
+ mirror_buf_iova + mirror_buf_offset,
cpy_len,
+ &mirror_mapped_len);
+ if (unlikely(!mirror_hpa)) {
+ VHOST_LOG_DATA(ERR, "(%d) %s: failed to get
mirror hpa.\n",
+ mirr_dev->vid, __func__);
+ error = -1;
+ goto out;
+ }
+
+ if (unlikely(tvec_idx >= nr_iovec)) {
+ VHOST_LOG_DATA(ERR, "iovec is not enough for
offloading\n");
+ return -1;
+ }
+
+ if (unlikely(mirror_mapped_len != mapped_len)) {
+ VHOST_LOG_DATA(ERR, "original mapped len is not
equal to mirror len\n");
+ return -1;
+ }
+
+ async_fill_vec(src_iovec + tvec_idx, hpa,
(size_t)mapped_len);
+ async_fill_vec(dst_iovec + tvec_idx,
+ (void *)(uintptr_t)rte_pktmbuf_iova_offset(cur,
mbuf_offset),
+ (size_t)mapped_len);
+ async_fill_vec(mirr_dst_iovec + tvec_idx,
+ mirror_hpa, (size_t)mirror_mapped_len);
+
+ if ((((uintptr_t)rte_pktmbuf_iova_offset(cur,
mbuf_offset) & 0xFFF) ^ ((uint64_t)mirror_hpa & 0xFFF)) != 0 && flag == 0)
+ {
+ total++;
+ VHOST_LOG_DATA(ERR, "%lu.........mbuf=%p
mirror_hpa=%p.\n", total, (void *)(uintptr_t)rte_pktmbuf_iova_offset(cur,
mbuf_offset), mirror_hpa);
+ }
+
+ tvec_idx++;
+ 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;
+ mirror_buf_avail -= (uint32_t)mapped_len;
+ mirror_buf_offset += (uint32_t)mapped_len;
+ }
+
+ /* This buf reaches to its end, get the next one */
+ if (buf_avail == 0) {
+ if (++vec_idx >= nr_vec)
+ 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;
+ buf_avail = buf_len;
+
+ PRINT_PACKET(dev, (uintptr_t)buf_addr,
(uint32_t)buf_avail, 0);
+ }
+
+ if (mirror_buf_avail == 0) {
+ mirror_vec_idx++;
+ if (unlikely(mirror_vec_idx >= mirr_nr_vec)) {
+ error = -1;
+ goto out;
+ }
+
+ mirror_buf_addr = mirr_buf_vec[mirror_vec_idx].buf_addr;
+ mirror_buf_iova = mirr_buf_vec[mirror_vec_idx].buf_iova;
+ mirror_buf_len = mirr_buf_vec[mirror_vec_idx].buf_len;
+
+ mirror_buf_offset = 0;
+ mirror_buf_avail = mirror_buf_len;
+ }
+
+ /*
+ * This mbuf reaches to its end, get a new one
+ * to hold more data.
+ */
+ if (mbuf_avail == 0) {
+ cur = rte_pktmbuf_alloc(mbuf_pool);
+ if (unlikely(cur == NULL)) {
+ VHOST_LOG_DATA(ERR, "Failed to allocate memory
for mbuf.\n");
+ return -1;
+ }
+
+ prev->next = cur;
+ prev->data_len = mbuf_offset;
+ m->nb_segs += 1;
+ m->pkt_len += mbuf_offset;
+ prev = cur;
+
+ mbuf_offset = 0;
+ mbuf_avail = cur->buf_len - RTE_PKTMBUF_HEADROOM;
+ }
+ flag++;
+ }
+
+ prev->data_len = mbuf_offset;
+ m->pkt_len += mbuf_offset;
+
+ if (tlen) {
+ async_fill_iter(src_it, tlen, src_iovec, tvec_idx);
+ async_fill_iter(dst_it, tlen, dst_iovec, tvec_idx);
+ async_fill_iter(mirr_dst_it, tlen, mirr_dst_iovec, tvec_idx);
+ if (hdr)
+ *nethdr = *hdr;
+ }
+out:
+ return error;
+}
+
+static __rte_always_inline uint16_t
+virtio_mirror_dev_tx_async_split(struct virtio_net *dev,
+ struct vhost_virtqueue *vq, uint16_t queue_id,
+ struct virtio_net *mirr_dev,
+ struct vhost_virtqueue *mirr_vq, uint16_t mirr_queue_id
__rte_unused,
+ struct rte_mempool *mbuf_pool, struct rte_mbuf **pkts,
+ uint16_t count, bool legacy_ol_flags)
+{
+ static bool allocerr_warned;
+ bool dropped = false;
+ uint16_t free_entries;
+ uint16_t pkt_idx, slot_idx = 0;
+ uint16_t nr_done_pkts = 0;
+ uint16_t nr_async_burst = 0;
+ uint16_t pkt_err = 0;
+ uint16_t iovec_idx = 0, it_idx = 0;
+ struct rte_vhost_iov_iter *it_pool = vq->it_pool;
+ struct iovec *vec_pool = vq->vec_pool;
+ struct iovec *src_iovec = vec_pool;
+ struct iovec *dst_iovec = vec_pool + (VHOST_MAX_ASYNC_VEC >> 1);
+ struct rte_vhost_async_desc tdes[MAX_PKT_BURST];
+ struct async_inflight_info *pkts_info = vq->async_pkts_info;
+ struct rte_mbuf *pkts_prealloc[MAX_PKT_BURST];
+
+
+ struct buf_vector mirror_buf_vec[BUF_VECTOR_MAX];
+ uint16_t mirror_num_buffers;
+ uint16_t mirror_avail_head;
+ mirror_avail_head = __atomic_load_n(&mirr_vq->avail->idx,
__ATOMIC_ACQUIRE);
+ rte_prefetch0(&mirr_vq->avail->ring[mirr_vq->last_avail_idx &
(mirr_vq->size - 1)]);
+
+ struct rte_vhost_iov_iter *mirror_it_pool = mirr_vq->it_pool;
+ struct iovec *mirror_dst_iovec = mirr_vq->vec_pool;
+ uint16_t mirror_iovec_idx = 0, mirror_it_idx = 0;
+
+ /**
+ * The ordering between avail index and
+ * desc reads needs to be enforced.
+ */
+ free_entries = __atomic_load_n(&vq->avail->idx, __ATOMIC_ACQUIRE) -
vq->last_avail_idx;
+ if (free_entries == 0)
+ goto out;
+
+ rte_prefetch0(&vq->avail->ring[vq->last_avail_idx & (vq->size - 1)]);
+
+ count = RTE_MIN(count, MAX_PKT_BURST);
+ count = RTE_MIN(count, free_entries);
+ VHOST_LOG_DATA(DEBUG, "(%d) about to dequeue %u buffers\n", dev->vid,
count);
+
+ if (rte_pktmbuf_alloc_bulk(mbuf_pool, pkts_prealloc, count))
+ goto out;
+
+ for (pkt_idx = 0; pkt_idx < count; pkt_idx++) {
+ uint16_t head_idx = 0;
+ uint16_t nr_vec = 0;
+ uint16_t to;
+ uint32_t buf_len;
+ int err;
+ struct buf_vector buf_vec[BUF_VECTOR_MAX];
+ struct rte_mbuf *pkt = pkts_prealloc[pkt_idx];
+ uint16_t mirror_nr_vec = 0;
+
+ if (unlikely(fill_vec_buf_split(dev, vq, vq->last_avail_idx,
+ &nr_vec, buf_vec,
+ &head_idx, &buf_len,
+ VHOST_ACCESS_RO) < 0)) {
+ dropped = true;
+ break;
+ }
+
+ if (unlikely(reserve_avail_buf_split(mirr_dev, mirr_vq,
+ buf_len, mirror_buf_vec,
&mirror_num_buffers,
+ mirror_avail_head,
&mirror_nr_vec) < 0)) {
+ VHOST_LOG_DATA(DEBUG,
+ "(%d) failed to get enough desc from mirror
vring\n",
+ mirr_dev->vid);
+ mirr_vq->shadow_used_idx -= mirror_num_buffers;
+ dropped = true;
+ break;
+ }
+
+ err = virtio_dev_pktmbuf_prep(dev, pkt, buf_len);
+ if (unlikely(err)) {
+ /**
+ * mbuf allocation fails for jumbo packets when external
+ * buffer allocation is not allowed and linear buffer
+ * is required. Drop this packet.
+ */
+ if (!allocerr_warned) {
+ VHOST_LOG_DATA(ERR,
+ "Failed mbuf alloc of size %d from %s
on %s.\n",
+ buf_len, mbuf_pool->name, dev->ifname);
+ allocerr_warned = true;
+ }
+ dropped = true;
+ break;
+ }
+
+ slot_idx = (vq->async_pkts_idx + pkt_idx) & (vq->size - 1);
+ err = egress_async_desc_to_mbuf(dev, buf_vec, nr_vec, pkt,
+ mbuf_pool, &src_iovec[iovec_idx],
+ &dst_iovec[iovec_idx], &it_pool[it_idx],
+ &it_pool[it_idx + 1],
+ mirr_dev, mirr_vq,
+ mirror_buf_vec, mirror_nr_vec,
mirror_num_buffers,
+ &mirror_dst_iovec[mirror_iovec_idx],
+ &mirror_it_pool[mirror_it_idx],
+ &pkts_info[slot_idx].nethdr,
+ (VHOST_MAX_ASYNC_VEC >> 1) - iovec_idx);
+ if (unlikely(err)) {
+ if (!allocerr_warned) {
+ VHOST_LOG_DATA(ERR,
+ "Failed to offload copies to async
channel %s.\n",
+ dev->ifname);
+ allocerr_warned = true;
+ }
+ dropped = true;
+ break;
+ }
+
+ egress_async_fill_desc(&tdes[nr_async_burst],
+ &it_pool[it_idx], &it_pool[it_idx + 1],
&mirror_it_pool[mirror_it_idx]);
+ pkts_info[slot_idx].mbuf = pkt;
+ nr_async_burst++;
+
+ iovec_idx += it_pool[it_idx].nr_segs;
+ it_idx += 2;
+
+ mirror_iovec_idx += mirror_it_pool[mirror_it_idx].nr_segs;
+ mirror_it_idx += 1;
+ mirr_vq->last_avail_idx += mirror_num_buffers;
+ pkts_info[slot_idx].descs = mirror_num_buffers;
+
+ /* store used descs */
+ to = vq->async_desc_idx_split & (vq->size - 1);
+ vq->async_descs_split[to].id = head_idx;
+ vq->async_descs_split[to].len = 0;
+ vq->async_desc_idx_split++;
+
+ vq->last_avail_idx++;
+
+ if (unlikely(nr_async_burst >= VHOST_ASYNC_BATCH_THRESHOLD)) {
+ uint16_t nr_pkts;
+ int32_t ret;
+
+ ret = vq->async_ops.transfer_data(dev->vid, queue_id,
+ tdes, 0, nr_async_burst, true);
+ if (unlikely(ret < 0)) {
+ VHOST_LOG_DATA(ERR, "(%d) async channel submit
error\n", dev->vid);
+ ret = 0;
+ }
+ nr_pkts = ret;
+
+ vq->async_pkts_inflight_n += nr_pkts;
+ it_idx = 0;
+ iovec_idx = 0;
+
+ mirror_iovec_idx = 0;
+ mirror_it_idx = 0;
+
+ if (unlikely(nr_pkts < nr_async_burst)) {
+ pkt_err = nr_async_burst - nr_pkts;
+ nr_async_burst = 0;
+ pkt_idx++;
+ break;
+ }
+ nr_async_burst = 0;
+ }
+ }
+
+ if (unlikely(dropped))
+ rte_pktmbuf_free_bulk(&pkts_prealloc[pkt_idx], count - pkt_idx);
+
+ if (nr_async_burst) {
+ uint16_t nr_pkts;
+ int32_t ret;
+
+ ret = vq->async_ops.transfer_data(dev->vid, queue_id, tdes, 0,
nr_async_burst, true);
+ if (unlikely(ret < 0)) {
+ VHOST_LOG_DATA(ERR, "(%d) async channel submit
error\n", dev->vid);
+ ret = 0;
+ }
+ nr_pkts = ret;
+
+ vq->async_pkts_inflight_n += nr_pkts;
+
+ if (unlikely(nr_pkts < nr_async_burst))
+ pkt_err = nr_async_burst - nr_pkts;
+ }
+
+ if (unlikely(pkt_err)) {
+ uint16_t nr_err_dma = pkt_err;
+ uint16_t num_descs = 0;
+
+ pkt_idx -= nr_err_dma;
+
+ /**
+ * recover async channel copy related structures and free
pktmbufs
+ * for error pkts.
+ */
+ vq->async_desc_idx_split -= nr_err_dma;
+ while (nr_err_dma-- > 0) {
+ rte_pktmbuf_free(pkts_info[slot_idx & (vq->size -
1)].mbuf);
+ num_descs += pkts_info[slot_idx & (vq->size - 1)].descs;
+ slot_idx--;
+ }
+
+ /* recover available ring */
+ vq->last_avail_idx -= pkt_err;
+
+ mirr_vq->shadow_used_idx -= num_descs;
+ mirr_vq->last_avail_idx -= num_descs;
+ }
+
+ vq->async_pkts_idx += pkt_idx;
+
+ if (likely(mirr_vq->shadow_used_idx)) {
+ uint16_t to = mirr_vq->async_desc_idx_split & (mirr_vq->size -
1);
+
+ store_dma_desc_info_split(mirr_vq->shadow_used_split,
+ mirr_vq->async_descs_split, mirr_vq->size, 0, to,
mirr_vq->shadow_used_idx);
+
+ mirr_vq->async_desc_idx_split += mirr_vq->shadow_used_idx;
+ mirr_vq->async_pkts_idx += pkt_idx;
+ mirr_vq->async_pkts_inflight_n += pkt_idx;
+ mirr_vq->shadow_used_idx = 0;
+ }
+
+out:
+ if (vq->async_pkts_inflight_n > 0) {
+ nr_done_pkts = async_poll_egress_completed_split(dev, vq,
+ queue_id,
+ mirr_dev, mirr_vq,
+ pkts, count, legacy_ol_flags);
+ }
+
+ return nr_done_pkts;
+}
+
+__rte_noinline
+static uint16_t
+virtio_mirror_dev_tx_async_split_legacy(struct virtio_net *dev,
+ struct vhost_virtqueue *vq, uint16_t queue_id,
+ struct virtio_net *mirr_dev,
+ struct vhost_virtqueue *mirr_vq, uint16_t mirr_queue_id,
+ struct rte_mempool *mbuf_pool, struct rte_mbuf **pkts,
+ uint16_t count)
+{
+ return virtio_mirror_dev_tx_async_split(dev, vq, queue_id,
+ mirr_dev, mirr_vq, mirr_queue_id,
+ mbuf_pool,
+ pkts, count, true);
+}
+
+__rte_noinline
+static uint16_t
+virtio_mirror_dev_tx_async_split_compliant(struct virtio_net *dev,
+ struct vhost_virtqueue *vq, uint16_t queue_id,
+ struct virtio_net *mirr_dev,
+ struct vhost_virtqueue *mirr_vq, uint16_t mirr_queue_id,
+ struct rte_mempool *mbuf_pool, struct rte_mbuf **pkts,
+ uint16_t count)
+{
+ return virtio_mirror_dev_tx_async_split(dev, vq, queue_id,
+ mirr_dev, mirr_vq, mirr_queue_id,
+ mbuf_pool,
+ pkts, count, false);
+}
+
+uint16_t
+rte_vhost_async_try_egress_burst(int vid, uint16_t queue_id,
+ int mirr_vid, uint16_t mirr_queue_id,
+ struct rte_mempool *mbuf_pool, struct rte_mbuf **pkts, uint16_t count,
+ int *nr_inflight)
+{
+ struct virtio_net *dev;
+ struct virtio_net *mirror_dev;
+ struct rte_mbuf *rarp_mbuf = NULL;
+ struct vhost_virtqueue *vq;
+ struct vhost_virtqueue *mirror_vq;
+ int16_t success = 1;
+
+ *nr_inflight = -1;
+
+ dev = get_device(vid);
+ mirror_dev = get_device(mirr_vid);
+ if (!dev || !mirror_dev)
+ return 0;
+
+ if (unlikely(!(dev->flags & VIRTIO_DEV_BUILTIN_VIRTIO_NET))
+ || unlikely(!(mirror_dev->flags &
VIRTIO_DEV_BUILTIN_VIRTIO_NET))) {
+ VHOST_LOG_DATA(ERR,
+ "(%d) %s: built-in vhost net backend is disabled.\n",
+ dev->vid, __func__);
+ return 0;
+ }
+
+ if (unlikely(!is_valid_virt_queue_idx(queue_id, 1, dev->nr_vring))
+ || unlikely(!is_valid_virt_queue_idx(mirr_queue_id, 0,
mirror_dev->nr_vring))) {
+ VHOST_LOG_DATA(ERR,
+ "(%d) %s: invalid virtqueue idx %d.\n",
+ dev->vid, __func__, queue_id);
+ return 0;
+ }
+
+ vq = dev->virtqueue[queue_id];
+ mirror_vq = mirror_dev->virtqueue[mirr_queue_id];
+
+ if (unlikely(rte_spinlock_trylock(&vq->access_lock) == 0)
+ || unlikely(rte_spinlock_trylock(&mirror_vq->access_lock) == 0))
+ return 0;
+
+ if (unlikely(vq->enabled == 0) || unlikely(mirror_vq->enabled == 0)) {
+ count = 0;
+ goto out_access_unlock;
+ }
+
+ if (unlikely(!vq->async_registered) ||
unlikely(!mirror_vq->async_registered)) {
+ VHOST_LOG_DATA(ERR, "(%d) %s: async not registered for queue id
%d.\n",
+ dev->vid, __func__, queue_id);
+ 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) || unlikely(mirror_vq->access_ok == 0))
+ if (unlikely(vring_translate(dev, vq) < 0)
+ || unlikely(vring_translate(mirror_dev, mirror_vq) <
0)) {
+ count = 0;
+ goto out_access_unlock;
+ }
+
+ /*
+ * Construct a RARP broadcast packet, and inject it to the "pkts"
+ * array, to looks like that guest actually send such packet.
+ *
+ * Check user_send_rarp() for more information.
+ *
+ * broadcast_rarp shares a cacheline in the virtio_net structure
+ * with some fields that are accessed during enqueue and
+ * __atomic_compare_exchange_n causes a write if performed compare
+ * and exchange. This could result in false sharing between enqueue
+ * and dequeue.
+ *
+ * Prevent unnecessary false sharing by reading broadcast_rarp first
+ * and only performing compare and exchange if the read indicates it
+ * is likely to be set.
+ */
+ if (unlikely(__atomic_load_n(&dev->broadcast_rarp, __ATOMIC_ACQUIRE) &&
+ __atomic_compare_exchange_n(&dev->broadcast_rarp,
+ &success, 0, 0, __ATOMIC_RELEASE, __ATOMIC_RELAXED))) {
+
+ rarp_mbuf = rte_net_make_rarp_packet(mbuf_pool, &dev->mac);
+ if (rarp_mbuf == NULL) {
+ VHOST_LOG_DATA(ERR, "Failed to make RARP packet.\n");
+ count = 0;
+ goto out;
+ }
+ count -= 1;
+ }
+
+#if 0
+ if (unlikely(__atomic_load_n(&mirror_dev->broadcast_rarp,
__ATOMIC_ACQUIRE) &&
+ __atomic_compare_exchange_n(&mirror_dev->broadcast_rarp,
+ &success, 0, 0, __ATOMIC_RELEASE, __ATOMIC_RELAXED))) {
+
+ rarp_mbuf = rte_net_make_rarp_packet(mbuf_pool,
&mirror_dev->mac);
+ if (rarp_mbuf == NULL) {
+ VHOST_LOG_DATA(ERR, "Failed to make RARP packet.\n");
+ count = 0;
+ goto out;
+ }
+ count -= 1;
+ }
+#endif
+ if (unlikely(vq_is_packed(dev)) || unlikely(vq_is_packed(mirror_dev))) {
+ VHOST_LOG_DATA(ERR,
+ "(%d) %s: async dequeue does not support packed
ring.\n",
+ dev->vid, __func__);
+ return 0;
+ }
+
+ if (dev->flags & VIRTIO_DEV_LEGACY_OL_FLAGS)
+ count = virtio_mirror_dev_tx_async_split_legacy(dev, vq,
queue_id,
+ mirror_dev, mirror_vq, mirr_queue_id,
+ mbuf_pool, pkts, count);
+ else
+ count = virtio_mirror_dev_tx_async_split_compliant(dev, vq,
queue_id,
+ mirror_dev, mirror_vq, mirr_queue_id,
+ mbuf_pool, pkts, count);
+
+out:
+ *nr_inflight = vq->async_pkts_inflight_n;
+
+ if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM))
+ vhost_user_iotlb_rd_unlock(vq);
+
+out_access_unlock:
+ rte_spinlock_unlock(&vq->access_lock);
+ rte_spinlock_unlock(&mirror_vq->access_lock);
+
+ if (unlikely(rarp_mbuf != NULL)) {
+ /*
+ * Inject it to the head of "pkts" array, so that switch's mac
+ * learning table will get updated first.
+ */
+ memmove(&pkts[1], pkts, count * sizeof(struct rte_mbuf *));
+ pkts[0] = rarp_mbuf;
+ count += 1;
+ }
+
+ return count;
+}
--
2.35.1
