This bug may need to be evaluated as an exception for RHEL 9.6, refer to https://issues.redhat.com/browse/RHEL-69455?focusedId=26627330&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-26627330

What were you trying to do that didn't work?

After migration failure, the passt process will be killed, the interface is down on the vm

What is the impact of this issue to you?

When migration fails, the vm network will broken

Please provide the package NVR for which the bug is seen:

# rpm -q libvirt  passt qemu-kvm
libvirt-10.10.0-7.el9.x86_64
passt-0^20250217.ga1e48a0-1.el9.x86_64
qemu-kvm-9.1.0-15.el9.x86_64

How reproducible is this bug?:

100%

Steps to reproduce

1.  Prepare 2 hosts, and prepare enough hugepages for the vm, prepare the migration env;
2.  On source host, start a vm with vhostuser passt interface, run wget on the vm;

   [guest] #wget --limit-rate=1M http://dell-per740-36.lab.eng.pek2.redhat.com/libvirt-CI-resources/RHEL-7.5-x86_64-non-release-ovmf.qcow2
   

3.  During the wget running, migrate the vm, the migration will fail:

   # virsh migrate rhel --live --verbose qemu+tcp://dell-xxx-xx.lab.eng.pek2.redhat.com/system --p2p 
   Migration: [98.22 %]error: operation failed: job 'migration out' failed: Save of field virtio-net/virtio failed
   

4. After then, check the vm status on the source host, the passt process is killed, and the interface is down;

   [host] # pidof passt
   (none)
   [guest] # ip addr show enp1s0
   2: enp1s0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc fq_codel state DOWN group default qlen 1000
       link/ether 52:54:00:5a:2d:da brd ff:ff:ff:ff:ff:ff
   

   Expected results

   After migration fail, the vm should work well on the source host
   Or by not failing the migration but simply not transferring TCP connections.

   Actual results

   After migration fail with tcp traffic on vm, the vm network is broken
   but for scenarios like cancel migration, the network continue works