[Qemu-devel] [PATCH V3] vhost_net: start/stop guest notifiers properly

[Jason Wang]

commit a9f98bb5ebe6fb1869321dcc58e72041ae626ad8 vhost: multiqueue
support changed the order of stopping the device. Previously
vhost_dev_stop would disable backend and only afterwards, unset guest
notifiers. We now unset guest notifiers while vhost is still
active. This can lose interrupts causing guest networking to fail. In
particular, this has been observed during migration.

To adapt this, several other changes are needed:
- remove the hdev->started assertion in vhost.c since we may want to
start the guest notifiers before vhost starts and stop the guest
notifiers after vhost is stopped.
- introduce the vhost_net_set_vq_index() and call it before setting
guest notifiers. This is used to guarantee vhost_net has the correct
virtqueue index when setting guest notifiers.

Reported-by: "Zhangjie (HZ)" <address@hidden>
Cc: William Dauchy <address@hidden>
Signed-off-by: Michael S. Tsirkin <address@hidden>
Signed-off-by: Jason Wang <address@hidden>

--
Changes from Michael's patch:
- Remove the assertion
Changes from V1:
- Rebase to latest
Changes from V2:
- Introduce vhost_net_set_vq_index() to unbreak multiqueue

Zhang Jie, please test this patch to see if it fixes the issue.
---
 hw/net/vhost_net.c | 31 +++++++++++++++++++------------
 hw/virtio/vhost.c  |  2 --
 2 files changed, 19 insertions(+), 14 deletions(-)

diff --git a/hw/net/vhost_net.c b/hw/net/vhost_net.c
index f87c798..fe0203d 100644
--- a/hw/net/vhost_net.c
+++ b/hw/net/vhost_net.c
@@ -188,9 +188,13 @@ bool vhost_net_query(VHostNetState *net, VirtIODevice *dev)
     return vhost_dev_query(&net->dev, dev);
 }
 
+static void vhost_net_set_vq_index(struct vhost_net *net, int vq_index)
+{
+    net->dev.vq_index = vq_index;
+}
+
 static int vhost_net_start_one(struct vhost_net *net,
-                               VirtIODevice *dev,
-                               int vq_index)
+                       VirtIODevice *dev)
 {
     struct vhost_vring_file file = { };
     int r;
@@ -201,7 +205,6 @@ static int vhost_net_start_one(struct vhost_net *net,
 
     net->dev.nvqs = 2;
     net->dev.vqs = net->vqs;
-    net->dev.vq_index = vq_index;
 
     r = vhost_dev_enable_notifiers(&net->dev, dev);
     if (r < 0) {
@@ -309,11 +312,7 @@ int vhost_net_start(VirtIODevice *dev, NetClientState *ncs,
     }
 
     for (i = 0; i < total_queues; i++) {
-        r = vhost_net_start_one(get_vhost_net(ncs[i].peer), dev, i * 2);
-
-        if (r < 0) {
-            goto err;
-        }
+        vhost_net_set_vq_index(get_vhost_net(ncs[i].peer), i * 2);
     }
 
     r = k->set_guest_notifiers(qbus->parent, total_queues * 2, true);
@@ -322,6 +321,14 @@ int vhost_net_start(VirtIODevice *dev, NetClientState *ncs,
         goto err;
     }
 
+    for (i = 0; i < total_queues; i++) {
+        r = vhost_net_start_one(get_vhost_net(ncs[i].peer), dev);
+
+        if (r < 0) {
+            goto err;
+        }
+    }
+
     return 0;
 
 err:
@@ -339,16 +346,16 @@ void vhost_net_stop(VirtIODevice *dev, NetClientState 
*ncs,
     VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(vbus);
     int i, r;
 
+    for (i = 0; i < total_queues; i++) {
+        vhost_net_stop_one(get_vhost_net(ncs[i].peer), dev);
+    }
+
     r = k->set_guest_notifiers(qbus->parent, total_queues * 2, false);
     if (r < 0) {
         fprintf(stderr, "vhost guest notifier cleanup failed: %d\n", r);
         fflush(stderr);
     }
     assert(r >= 0);
-
-    for (i = 0; i < total_queues; i++) {
-        vhost_net_stop_one(get_vhost_net(ncs[i].peer), dev);
-    }
 }
 
 void vhost_net_cleanup(struct vhost_net *net)
diff --git a/hw/virtio/vhost.c b/hw/virtio/vhost.c
index e55fe1c..5d7c40a 100644
--- a/hw/virtio/vhost.c
+++ b/hw/virtio/vhost.c
@@ -976,7 +976,6 @@ void vhost_dev_disable_notifiers(struct vhost_dev *hdev, 
VirtIODevice *vdev)
 bool vhost_virtqueue_pending(struct vhost_dev *hdev, int n)
 {
     struct vhost_virtqueue *vq = hdev->vqs + n - hdev->vq_index;
-    assert(hdev->started);
     assert(n >= hdev->vq_index && n < hdev->vq_index + hdev->nvqs);
     return event_notifier_test_and_clear(&vq->masked_notifier);
 }
@@ -988,7 +987,6 @@ void vhost_virtqueue_mask(struct vhost_dev *hdev, 
VirtIODevice *vdev, int n,
     struct VirtQueue *vvq = virtio_get_queue(vdev, n);
     int r, index = n - hdev->vq_index;
 
-    assert(hdev->started);
     assert(n >= hdev->vq_index && n < hdev->vq_index + hdev->nvqs);
 
     struct vhost_vring_file file = {
-- 
1.8.3.1