Calico Enterprise KubeVirt docs: say live-migrating VMs are IPv4-only (bridge binding) #3161

The Calico Enterprise KubeVirt pages don't say that VMs using live migration are IPv4-only. One page implies the opposite.

Why VMs are IPv4-only

• KubeVirt gives a VM IPv6 only with the masquerade binding (DHCPv6 plus per-family NAT). With the bridge binding, it deliberately drops the pod's IPv6 address, because the address would never be reachable from the guest (pkg/network/setup/netpod/discover.go).
• networking/kubevirt/kubevirt-networking.mdx says masquerade is not supported for live migration, so bridge binding is required.
• So a VM that can live-migrate with Calico gets IPv4 only.
• We confirmed this on a cluster. On a dual-stack IP pool, the virt-launcher pod gets an IPv6 address, but the VMI status and the guest don't.

Gaps

1. networking/kubevirt/kubevirt-networking.mdx and index.mdx never mention IPv6 either way. They should say that KubeVirt VMs on Calico (bridge binding) get IPv4 only.
2. networking/kubevirt/live-migration-bgp.mdx: the example BGPFilter has exportV6 / importV6 rules, which suggests IPv6 live migration works for KubeVirt. Either drop the v6 rules from the KubeVirt example, or say they only apply to non-KubeVirt workloads.
3. networking/kubevirt/install-forklift-kubernetes.mdx / install-forklift-openshift.mdx and networking/l2-bridge/vm-identity.mdx don't say that only a source VM's IPv4 address is kept as its Calico identity. A guest's IPv6 configuration isn't.

reference/l2-bridge-support.mdx already says IPv6 isn't supported on L2. A similar line on the KubeVirt pages would close this.