On Thu, Oct 8, 2020 at 5:36 PM Xie He <xie.he.0...@gmail.com> wrote: > > On Thu, Oct 8, 2020 at 1:32 PM Willem de Bruijn > <willemdebruijn.ker...@gmail.com> wrote: > > > > On Thu, Oct 8, 2020 at 4:11 PM Xie He <xie.he.0...@gmail.com> wrote: > > > > > > OK. I understand that t->tun_hlen is the GRE header length. What is > > > t->encap_hlen? > > > > I've looked at that closely either. > > > > Appears to be to account for additional FOU/GUE encap: > > > > " > > commit 56328486539ddd07cbaafec7a542a2c8a3043623 > > Author: Tom Herbert <therb...@google.com> > > Date: Wed Sep 17 12:25:58 2014 -0700 > > net: Changes to ip_tunnel to support foo-over-udp encapsulation > > > > This patch changes IP tunnel to support (secondary) encapsulation, > > Foo-over-UDP. Changes include: > > > > 1) Adding tun_hlen as the tunnel header length, encap_hlen as the > > encapsulation header length, and hlen becomes the grand total > > of these. > > 2) Added common netlink define to support FOU encapsulation. > > 3) Routines to perform FOU encapsulation. > > > > Signed-off-by: Tom Herbert <therb...@google.com> > > Signed-off-by: David S. Miller <da...@davemloft.net> > > " > > I see the ipgre_xmit function would pull the header our header_ops > creates, and then call __gre_xmit. __gre_xmit will call > gre_build_header to complete the GRE header. gre_build_header expects > to find the base GRE header after pushing tunnel->tun_hlen. However, > if tunnel->encap_hlen is not 0, it couldn't find the base GRE header > there. Is there a problem? > > Where exactly should we put the tunnel->encap_hlen header? Before the > GRE header or after?
The L4 tunnel infra uses the two callbacks encap_hlen (e.g., fou_encap_hlen) and build_header (fou_build_header) in struct ip_tunnel_encap_ops to first allocate more space and later call back into the specific implementation to fill that data. build_header is called from __gre6_xmit -> ip6_tnl_xmit (or its ipv4 equivalent). This happens after gre has pushed its header, so the headers will come before that.
