From: Willem de Bruijn <willemdebruijn.ker...@gmail.com> Date: Wed, 5 Jul 2017 13:21:02 -0400
> On Wed, Jul 5, 2017 at 12:06 PM, Willem de Bruijn > <willemdebruijn.ker...@gmail.com> wrote: >>> diff --git a/include/linux/virtio_net.h b/include/linux/virtio_net.h >>> index 5209b5e..32fb046 100644 >>> --- a/include/linux/virtio_net.h >>> +++ b/include/linux/virtio_net.h >>> @@ -18,9 +18,6 @@ static inline int virtio_net_hdr_to_skb(struct sk_buff >>> *skb, >>> case VIRTIO_NET_HDR_GSO_TCPV6: >>> gso_type = SKB_GSO_TCPV6; >>> break; >>> - case VIRTIO_NET_HDR_GSO_UDP: >>> - gso_type = SKB_GSO_UDP; >>> - break; >> >> Virtio devices negotiate feature support before using this, but >> tuntap and pf_packet may be passing these packets unconditionally. >> Perhaps we should fragment those on the spot with skb_segment. > > Tun has ioctl TUNSETIFF to probe for features and it can be argued > that packet sockets should query device features with ethtool before > relying on them. So perhaps we don't need to fix this up, after all. Yes, this is the same thought process I went through, and the same conclusion I arrived at. :-)