On 4.16, if a VM is created where total vcpus are more than the cpus the host has in a single numa node, it fails to start with this error:

default                                            0s          Warning   SyncFailed                          virtualmachineinstance/vm-big                                       server error. command SyncVMI failed: "LibvirtError(Code=67, Domain=10, Message='unsupported configuration: more than 255 vCPUs require extended interrupt mode enabled on the iommu device')"

OCP 4.16.6
CNV 4.16.1

reported across multiple internal clusters

1. Start VM w/ high vcpu count
2. Check event logs
3. VM fails to start

VM fails

VM can succeed w/ many vcpus

Using very simple VM definition, setting vcpus > 1numa node cpus (on a 128cpu host):


apiVersion: kubevirt.io/v1
kind: VirtualMachine
metadata:
  labels:
    app: vm-big
  name: vm-big
spec:
  running: true
  template:
    metadata:
      labels:
        kubevirt.io/domain: vm-big
    spec:
      domain:
        cpu:
          cores: 1
          sockets: 100
          threads: 1
        devices:
          disks:
          - disk:
              bus: virtio
            name: containerdisk
          - disk:
              bus: virtio
            name: cloudinitdisk
          interfaces:
          - masquerade: {}
            model: virtio
            name: default
          networkInterfaceMultiqueue: true
          rng: {}
        features:
          smm:
            enabled: true
        firmware:
          bootloader:
            efi: {}
        machine:
          type: pc-q35-rhel9.2.0
        memory:
          guest: 10Gi
      networks:
      - name: default
        pod: {}
      terminationGracePeriodSeconds: 180
      nodeSelector:
        kubernetes.io/hostname: worker00
      volumes:
      - containerDisk:
          image: quay.io/kubevirt/fedora-container-disk-images:35
          imagePullPolicy: IfNotPresent
        name: containerdisk
      - cloudInitNoCloud:
          userData: |-
            #cloud-config
            user: fedora
            password: perf
            chpasswd: { expire: False }
            runcmd:
             - sed -i -e "s/PasswordAuthentication.*/PasswordAuthentication yes/" /etc/ssh/sshd_config
             - systemctl restart sshd
        name: cloudinitdisk