]> git.proxmox.com Git - mirror_qemu.git/commit
vhost_net: start/stop guest notifiers properly
authorJason Wang <jasowang@redhat.com>
Tue, 19 Aug 2014 04:56:29 +0000 (12:56 +0800)
committerMichael S. Tsirkin <mst@redhat.com>
Wed, 3 Sep 2014 13:40:44 +0000 (16:40 +0300)
commitcd7d1d26b0a333bf2fca715e332690bbd738c097
tree885b98af8f49afefb04737618be0d72d10f917db
parentd7efb7e08e5edaac23b0dc824f72c3f353447c39
vhost_net: start/stop guest notifiers properly

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 fix 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 to guarantee vhost_net has the correct
virtqueue index when setting guest notifiers.

MST: fix up error handling.

Cc: qemu-stable@nongnu.org
Cc: Jason Wang <jasowang@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Andrey Korolyov <andrey@xdel.ru>
Reported-by: "Zhangjie (HZ)" <zhangjie14@huawei.com>
Tested-by: William Dauchy <william@gandi.net>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
hw/net/vhost_net.c
hw/virtio/vhost.c