Hi, For all NICs (except virtio-net) emulated by qemu,such as e1000, rtl8139, pcnet and ne2k_pci, Qemu can still receive packets when VM is not running.
If this happened in *migration's* last PAUSE VM stage, The new dirty RAM related to the packets will be missed, And this will lead serious network fault in VM. We have discussed this question long time ago, you can find it in http://qemu.11.n7.nabble.com/PATCH-e1000-rtl8139-forbid-dealing-with-packets-when-VM-is-paused-td261168.html The problem is still remain, we can reproduct the bug as the follow steps: (1) Start a VM which configured at least one NIC (2) In VM, open several Terminal and do *Ping IP -i 0.1* (3) Migrate the VM repeatly between two Hosts. And the *PING* command in VM will very likely fail with message: 'Destination HOST Unreachable', the NIC in VM will stay unavailable unless you run 'service network restart'. This patch set is to solve this problem, and according to the suggestion of Peter Maydell and Stefan Hajnoczi, we implement this in net layer of qemu. PS. In the old discussion, Michael S. Tsirkin advanced that other devices may have the same problem, he suggested to stop all io threads with the VM when VM is not running. I have test serial, virtio-console, etc. And did not find any obvious fault in the migration scene. Also the scheme of stopping all io threads is complex and maybe a lot work to do. And for now i think it is ok to fix the obvious net bug first. Thanks, zhanghailiang zhanghailiang (3): net: forbid dealing with packets when VM is not running net: Flush queues when runstate changes back to running virtio-net: Remove checking vm state in virtio_net_can_receive hw/net/virtio-net.c | 4 ---- include/net/net.h | 2 ++ net/net.c | 28 ++++++++++++++++++++++++++++ 3 files changed, 30 insertions(+), 4 deletions(-) -- 1.7.12.4
