On Wed, May 22, 2013 at 1:20 PM, Pravin Shelar <pshe...@nicira.com> wrote: > On Tue, May 21, 2013 at 4:39 PM, Jesse Gross <je...@nicira.com> wrote: >> On Mon, May 20, 2013 at 2:33 PM, Pravin B Shelar <pshe...@nicira.com> wrote: >>> Following patch restructures ovs tunneling and gre vport >>> implementation to make ovs tunneling more in sync with >>> upstream kernel tunneling. Doing this tunneling code is >>> simplified as most of protocol processing on send and >>> recv is pushed to kernel tunneling. For external ovs >>> module the code is moved to kernel compatibility code. >>> >>> Signed-off-by: Pravin B Shelar <pshe...@nicira.com> >> >> A couple more comments (just on the non-compat portions): >> >>> diff --git a/datapath/vport-gre.c b/datapath/vport-gre.c >>> index add17d9..f551976 100644 >>> --- a/datapath/vport-gre.c >>> +++ b/datapath/vport-gre.c >>> -static __be32 be64_get_high32(__be64 x) >>> +static __be16 ovs_tnl_flags_to_gre_flags(u16 tun_flags) >> [...] >>> + __be16 flags = 0; >>> + if (tun_flags & OVS_TNL_F_CSUM) >>> + flags |= TUNNEL_CSUM; >>> + if (tun_flags & OVS_TNL_F_KEY) >>> + flags |= TUNNEL_KEY; >>> + return flags; >>> } >> >> Is there any way that we can unify the tunnel flags values to avoid >> needing to do this conversion? I know the existing flags are based on >> the GRE values, so maybe we should push the conversion down there. >> > OVS can directly use TUNNEL_* flags, we just need to define > OVS_TNL_F_DONT_FRAGMENT equivalent there.
I'm not sure that it makes sense to have OVS use the TUNNEL_ flags directly since they are just GRE protocol values. I was wondering if we could make the protocol handlers use the OVS values. For example, with VXLAN it seems undesirable to define yet another set of flags and odd to use the GRE ones. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
