Description of problem:
The nic inside guest changed to down when change mtu many times with ping traffic

Version-Release number of selected component (if applicable):
openvswitch-2.7.2-8.git20170719.el7fdp.x86_64.rpm 

How reproducible:


Steps to Reproduce:
In the testing, there were two machines, one server was "cisco-c220m3-01.rhts.eng.pek2.redhat.com", the other was "dell-per730-02.rhts.eng.pek2.redhat.com", the two machines connect directly.
1. 
configured ovs on cisco machine as following,
/usr/bin/ovs-vsctl --no-wait set Open_vSwitch . other_config:dpdk-init=true
 /usr/bin/ovs-vsctl --no-wait set Open_vSwitch . other_config:dpdk-lcore-mask=0x2
/usr/bin/ovs-vsctl --no-wait set Open_vSwitch . other_config:dpdk-socket-mem=1024,0
/usr/bin/ovs-vsctl --timeout 10 set Open_vSwitch . other_config:pmd-cpu-mask=1
ovs-vsctl --timeout 10 set Open_vSwitch . other_config:pmd-cpu-mask=30
systemctl restart openvswitch
/usr/bin/ovs-vsctl --timeout 10 add-br br0 -- set bridge br0 datapath_type=netdev
/usr/bin/ovs-vsctl --timeout 10 add-port br0 dpdk0 -- set Interface dpdk0 type=dpdk options:dpdk-devargs=0000:0b:00.0 
/usr/bin/ovs-vsctl --timeout 10 add-port br0 dpdkvhostuser0 -- set Interface dpdkvhostuser0 type=dpdkvhostuser 
sudo /usr/bin/ovs-ofctl -O OpenFlow13 --timeout 10 del-flows br0 
 /usr/bin/ovs-ofctl -O OpenFlow13 --timeout 10 add-flow br0 idle_timeout=0,in_port=1,action=output:2
/usr/bin/ovs-ofctl -O OpenFlow13 --timeout 10 add-flow br0 idle_timeout=0,in_port=2,action=output:1
chmod 777 /var/run/openvswitch/dpdkvhostuser0

There is guest inside cisco machine, configured ip of guest eth1 as following:
ip addr add 192.168.10.1/24 dev eth1

2.
configured ip on dell02 machine:
ip addr add 192.168.10.2/24 dev p3p1
and ping the guest inside cisco machine as following:
ping -n -i 0.001 192.168.10.1

3.
changed mtu many times on cisco machine, such as following commands.
ovs-vsctl set int dpdkvhostuser0 mtu_request=1900
ovs-vsctl set int dpdkvhostuser0 mtu_request=2000
ovs-vsctl set int dpdkvhostuser0 mtu_request=2200
ovs-vsctl set int dpdkvhostuser0 mtu_request=2300
ovs-vsctl set int dpdkvhostuser0 mtu_request=9000
ovs-vsctl set int dpdkvhostuser0 mtu_request=2000
ovs-vsctl set int dpdkvhostuser0 mtu_request=1500


Actual results:
After changed mtu many times with ping traffic, the eth1 inside cisco machine' guest changed to down as following, and dell02 ping "192.168.10.1" failed. 
eth1: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast state DOWN qlen 1000
    link/ether 56:48:4f:53:54:01 brd ff:ff:ff:ff:ff:ff
    inet 192.168.10.1/24 scope global eth1
       valid_lft forever preferred_lft forever
    inet6 fe80::5448:4fff:fe53:5401/64 scope link 
       valid_lft forever preferred_lft forever

There was no segfault in /var/log/messages. 
And the ovs service seems worked well as following.
[root@cisco-c220m3-01 openvswitch]# ovs-vsctl show
8ab35f85-52e0-498d-a1fd-3dd3408ba1ce
    Bridge "br0"
        Port "dpdkvhostuser0"
            Interface "dpdkvhostuser0"
                type: dpdkvhostuser
        Port "dpdk0"
            Interface "dpdk0"
                type: dpdk
                options: {dpdk-devargs="0000:0b:00.0"}
        Port "br0"
            Interface "br0"
                type: internal
    ovs_version: "2.7.2"

Run "ip link set eth1 up" failed, it need to restart guest,and then the eth1 change to up.

I tested it on ixgbe nic and i40e nic, they also had this issue.

Expected results:
The eth1 should be keep up when changed mtu many times with ping traffic.

Additional info: