The code virtio_dev_rx_async_submit has mismatch between locking the
access_lock for write, and then releases it with read_unlock.  This could cause
all sorts of deadlock depending on the implementation of rwlock.

static __rte_always_inline uint32_t
virtio_dev_rx_async_submit(struct virtio_net *dev, struct vhost_virtqueue *vq,
        struct rte_mbuf **pkts, uint32_t count, int16_t dma_id, uint16_t
vchan_id)
{
        uint32_t nb_tx = 0;

        VHOST_DATA_LOG(dev->ifname, DEBUG, "%s", __func__);

        if (unlikely(!dma_copy_track[dma_id].vchans ||
                               
!dma_copy_track[dma_id].vchans[vchan_id].pkts_cmpl_flag_addr)) {
                VHOST_DATA_LOG(dev->ifname, ERR,
                        "%s: invalid channel %d:%u.",
                         __func__, dma_id, vchan_id);
                return 0;
        }

        rte_rwlock_write_lock(&vq->access_lock);

        if (unlikely(!vq->enabled || !vq->async))
                goto out_access_unlock;

        vhost_user_iotlb_rd_lock(vq);

        if (unlikely(!vq->access_ok)) {
                vhost_user_iotlb_rd_unlock(vq);
                rte_rwlock_read_unlock(&vq->access_lock);  <<<<<<<< HERE

                virtio_dev_vring_translate(dev, vq);
                goto out_no_unlock;
        }