On Thu, Feb 19, 2015 at 12:02 PM, Jesse Gross <je...@nicira.com> wrote: > On Thu, Feb 19, 2015 at 11:29 AM, Pravin Shelar <pshe...@nicira.com> wrote: >> On Wed, Feb 18, 2015 at 7:02 PM, Jesse Gross <je...@nicira.com> wrote: >>> This series backports much of the UDP tunnel infrastructure, ultimately >>> in support of syncing Geneve with upstream and providing outer UDP >>> checksum support. >>> >>> During the backporting process, I realized that there are several >>> other lurking bugs that were fixed through a more complete UDP >>> tunnel layer. The ones that I am aware of are: >>> * VXLAN TSO fails in some cases because the OVS handle_offload() >>> function is used with the upstream GSO handler. >>> * Hardware VXLAN offload was disabled on kernels that is otherwise >>> supported on since version checks were incremented by the GBP changes. >>> * LISP attempts to use hardware offload with NICs that only support >>> VXLAN >>> >>> Jesse Gross (14): >>> datapath: Remove compat vxlan_src_port(). >>> datapath: Account for "udp: Add udp_sock_create for UDP tunnels to >>> open listener socket" >>> datapath: Account for "vxlan: Call udp_sock_create" >>> datapath: Use udp_sock_create() for LISP. >>> datapath: Account for "udp: Generic functions to set checksum" >>> datapath: Enable OVS GSO to be used up to 3.18 if necessary. >>> datapath: Account for "udp-tunnel: Add a few more UDP tunnel APIs" >>> datapath: Consistently set skb->inner_protocol for tunnels. >>> datapath: Account for "vxlan: Refactor vxlan driver to make use of the >>> common UDP tunnel functions." >>> datapath: Use additional common UDP functions for LISP. >>> datapath: Account for "vlan: introduce *vlan_hwaccel_push_inside >>> helpers" >>> datapath: Account for "udp: Do not require sock in >>> udp_tunnel_xmit_skb" >>> datapath: Backport upstream Geneve implementation. >>> datapath: Account for "openvswitch: Add support for checksums on UDP >>> tunnels." >>> >> >> I have not finished review yet, but wanted to let you know that this >> series does not compile on 3.12, 3.17, so that you can fix it in >> parallel. > > Thanks, I'll take a look at that.
I fixed the compilation problems on these kernels. The only affected patch was #6 (enabling GSO on up to 3.18) so I resent that one only in case you are still in the process of looking at the others. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
