Commit a2daaa49 authored by Jason Wang's avatar Jason Wang Committed by Khalid Elmously

vhost_net: tx batching

This patch tries to utilize tuntap rx batching by peeking the tx
virtqueue during transmission, if there's more available buffers in
the virtqueue, set MSG_MORE flag for a hint for backend (e.g tuntap)
to batch the packets.
Reviewed-by: default avatarStefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: default avatarJason Wang <jasowang@redhat.com>
Acked-by: default avatarMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>

CVE-2019-3900

(backported from commit 0ed005ce)
[tyhicks: Minor context adjustment due to missing commit 03088137
 ("vhost_net: basic polling support")]
Signed-off-by: default avatarTyler Hicks <tyhicks@canonical.com>
Acked-by: default avatarStefan Bader <stefan.bader@canonical.com>
Acked-by: default avatarConnor Kuehl <connor.kuehl@canonical.com>
Signed-off-by: default avatarKhalid Elmously <khalid.elmously@canonical.com>
parent 1eed6074
...@@ -287,6 +287,15 @@ static void vhost_zerocopy_callback(struct ubuf_info *ubuf, bool success) ...@@ -287,6 +287,15 @@ static void vhost_zerocopy_callback(struct ubuf_info *ubuf, bool success)
rcu_read_unlock_bh(); rcu_read_unlock_bh();
} }
static bool vhost_exceeds_maxpend(struct vhost_net *net)
{
struct vhost_net_virtqueue *nvq = &net->vqs[VHOST_NET_VQ_TX];
struct vhost_virtqueue *vq = &nvq->vq;
return (nvq->upend_idx + vq->num - VHOST_MAX_PEND) % UIO_MAXIOV
== nvq->done_idx;
}
/* Expects to be always run from workqueue - which acts as /* Expects to be always run from workqueue - which acts as
* read-size critical section for our kind of RCU. */ * read-size critical section for our kind of RCU. */
static void handle_tx(struct vhost_net *net) static void handle_tx(struct vhost_net *net)
...@@ -327,8 +336,7 @@ static void handle_tx(struct vhost_net *net) ...@@ -327,8 +336,7 @@ static void handle_tx(struct vhost_net *net)
/* If more outstanding DMAs, queue the work. /* If more outstanding DMAs, queue the work.
* Handle upend_idx wrap around * Handle upend_idx wrap around
*/ */
if (unlikely((nvq->upend_idx + vq->num - VHOST_MAX_PEND) if (unlikely(vhost_exceeds_maxpend(net)))
% UIO_MAXIOV == nvq->done_idx))
break; break;
head = vhost_get_vq_desc(vq, vq->iov, head = vhost_get_vq_desc(vq, vq->iov,
...@@ -388,6 +396,16 @@ static void handle_tx(struct vhost_net *net) ...@@ -388,6 +396,16 @@ static void handle_tx(struct vhost_net *net)
msg.msg_control = NULL; msg.msg_control = NULL;
ubufs = NULL; ubufs = NULL;
} }
total_len += len;
if (total_len < VHOST_NET_WEIGHT &&
!vhost_vq_avail_empty(&net->dev, vq) &&
likely(!vhost_exceeds_maxpend(net))) {
msg.msg_flags |= MSG_MORE;
} else {
msg.msg_flags &= ~MSG_MORE;
}
/* TODO: Check specific error and bomb out unless ENOBUFS? */ /* TODO: Check specific error and bomb out unless ENOBUFS? */
err = sock->ops->sendmsg(sock, &msg, len); err = sock->ops->sendmsg(sock, &msg, len);
if (unlikely(err < 0)) { if (unlikely(err < 0)) {
...@@ -406,7 +424,6 @@ static void handle_tx(struct vhost_net *net) ...@@ -406,7 +424,6 @@ static void handle_tx(struct vhost_net *net)
vhost_add_used_and_signal(&net->dev, vq, head, 0); vhost_add_used_and_signal(&net->dev, vq, head, 0);
else else
vhost_zerocopy_signal_used(net, vq); vhost_zerocopy_signal_used(net, vq);
total_len += len;
vhost_net_tx_packet(net); vhost_net_tx_packet(net);
if (unlikely(total_len >= VHOST_NET_WEIGHT)) { if (unlikely(total_len >= VHOST_NET_WEIGHT)) {
vhost_poll_queue(&vq->poll); vhost_poll_queue(&vq->poll);
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment