This series backports the netfilter/defrag-related changes made recently
upstream to our compat code, which should bring conntrack.c up-to-date just
prior to the NAT changes. Patch 5 introduced some breakage which is fixed in
patches 6 and 7; I have left these separate to mirror the upstream commits.

Tested using kmod tests on Ubuntu 3.13.0-24, 3.16.0-70, 3.19.0-58, and
4.2.0-35, and RHEL 3.10.0-327, plus compilation against vanilla
kernel targets on Travis:
https://travis-ci.org/joestringer/openvswitch/builds/127290050

v2:
- 3 new backport patches, bringing the series up to date with upstream prior
  to the NAT series.
- Added a final patch to document the defrag backport strategy in v4 and v6.
- Fixed upstream commit references to use the correct IDs.
v1:
- Initial Post

Joe Stringer (9):
  compat: Add a struct net parameter to l4_pkt_to_tuple.
  compat: ipv4: Pass struct net into ip_defrag.
  compat: ipv6: Pass struct net into nf_ct_frag6_gather.
  compat: nf_defrag_ipv6: avoid/free clone operations.
  compat: nf_defrag_ipv6: avoid nf_iterate recursion.
  compat: nf_defrag_ipv6: fix NULL deref panic.
  datapath: Orphan skbs before IPv6 defrag
  datapath: Fix template leak in error cases.
  compat: Document nf_defrag_ipv[46] backport.

 acinclude.m4                                       |   5 +
 datapath/conntrack.c                               |  36 ++---
 datapath/linux/Modules.mk                          |   1 +
 datapath/linux/compat/include/net/ip.h             |  15 +-
 .../include/net/netfilter/ipv6/nf_defrag_ipv6.h    |  21 ++-
 .../compat/include/net/netfilter/nf_conntrack.h    |  17 ++
 datapath/linux/compat/ip_fragment.c                |   3 +-
 datapath/linux/compat/nf_conntrack_reasm.c         | 172 +++++++++------------
 8 files changed, 133 insertions(+), 137 deletions(-)
 create mode 100644 datapath/linux/compat/include/net/netfilter/nf_conntrack.h

-- 
2.1.4

_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev

Reply via email to