On Tue, Oct 30, 2012 at 11:47:45AM -0400, Vlad Yasevich wrote:

On 10/29/2012 11:49 AM, Michael S. Tsirkin wrote:

quoted

It seems that to avoid deadlocks it is enough to poll vq before
 we are going to use the last buffer.  This should be faster than
c70aa540c7a9f67add11ad3161096fb95233aa2e.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
 drivers/vhost/net.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c
index 8e9de79..3967f82 100644
--- a/drivers/vhost/net.c
+++ b/drivers/vhost/net.c

@@ -197,8 +197,16 @@ static void vhost_zerocopy_callback(struct ubuf_info *ubuf, int status)
 {
 	struct vhost_ubuf_ref *ubufs = ubuf->ctx;
 	struct vhost_virtqueue *vq = ubufs->vq;
-
-	vhost_poll_queue(&vq->poll);
+	int cnt = atomic_read(&ubufs->kref.refcount);
+
+	/*
+	 * Trigger polling thread if guest stopped submitting new buffers:
+	 * in this case, the refcount after decrement will eventually reach 1
+	 * so here it is 2.
+	 * We also trigger polling periodically after each 16 packets.
+	 */
+	if (cnt <= 2 || !(cnt % 16))

Why 16?  Does it make sense to make it configurable?

-vlad

Probably not but I'll add a comment explaining why.

quoted

+		vhost_poll_queue(&vq->poll);
 	/* set len to mark this desc buffers done DMA */
 	vq->heads[ubuf->desc].len = status ?
 		VHOST_DMA_FAILED_LEN : VHOST_DMA_DONE_LEN;