On Wed, Apr 02, 2025 at 09:16:22AM +0200, Stefano Brivio wrote: [snip]I agree that supporting IFF_TUN would be a good idea in general. I think trying to do it in the middle of this will be unnecessary difficulty though. There are a bunch of places throughout the code that assume we have Ethernet headers, and changing that will require a pretty wide ranging rework. -- David Gibson (he or they) | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you, not the other way | around. http://www.ozlabs.org/~dgibsonstatic int tap_ns_tun(void *arg) { + /* TODO we need to check if vhost support VIRTIO_NET_F_MRG_RXBUF and VHOST_NET_F_VIRTIO_NET_HDR actually */ + static const uint64_t features = + (1ULL << VIRTIO_F_VERSION_1) | (1ULL << VIRTIO_NET_F_MRG_RXBUF) | (1ULL << VHOST_NET_F_VIRTIO_NET_HDR); struct ifreq ifr = { .ifr_flags = IFF_TAP | IFF_NO_PI };I kind of wonder, by the way, if IFF_TUN simplifies things here. It's something we should already add, as an option, see also: https://bugs.passt.top/show_bug.cgi?id=49, but if it makes your life easier for any reason you might consider adding it right away.