On 7/18/25 1:17 PM, Stefano Garzarella wrote:
> On Fri, Jul 18, 2025 at 10:52:38AM +0200, Paolo Abeni wrote:
>> @@ -4245,6 +4274,14 @@ static const Property virtio_net_properties[] = {
>> rss_data.specified_hash_types,
>> VIRTIO_NET_HASH_REPORT_UDPv6_EX - 1,
>> ON_OFF_AUTO_AUTO),
>> + DEFINE_PROP_FEATURE("host_tunnel", VirtIONet, host_features_ex,
>> + VIRTIO_NET_F_HOST_UDP_TUNNEL_GSO, true),
>> + DEFINE_PROP_FEATURE("host_tunnel_csum", VirtIONet, host_features_ex,
>> + VIRTIO_NET_F_HOST_UDP_TUNNEL_GSO_CSUM, true),
>> + DEFINE_PROP_FEATURE("guest_tunnel", VirtIONet, host_features_ex,
>> + VIRTIO_NET_F_GUEST_UDP_TUNNEL_GSO, true),
>> + DEFINE_PROP_FEATURE("guest_tunnel_csum", VirtIONet, host_features_ex,
>> + VIRTIO_NET_F_GUEST_UDP_TUNNEL_GSO_CSUM, true),
>
> IIUC, to avoid issue with live-migration, we should disable all of them
> in `hw_compat_10_0` in hw/core/machine.c (e.g. like `host_uso`,
> `guest_uso*` in hw_compat_8_1).
I guess the relevant compat entry should be 10.1, right? AFAICS such
entry should added with a separate/specific before 10.1 will be tagged.
I'll add the relevant the entries after that the hw_compat_10_1 will be
added.
Thanks,
Paolo
