Re: [RFCv2 net-next 0/7] OVS conntrack support

From: Tom Herbert
Date: Mon Mar 02 2015 - 19:59:34 EST


On Mon, Mar 2, 2015 at 1:54 PM, Joe Stringer <joestringer@xxxxxxxxxx> wrote:
> The goal of this series is to allow OVS to send packets through the Linux
> kernel connection tracker, and subsequently match on fields populated by
> conntrack.
>
> Sending this out as another RFC change as this is the first time IP fragment
> support is included. Only IPv4 is added right now, as we'd like to get some
> feedback on that approach before we implement IPv6 frag support.
>
> Helper support is also yet to be addressed, for tracking a particular flow a la
> iptables CT targets. I think this is just a matter of having userspace specify
> the helper to use (eg via 8-bit field in conntrack action), and setting up the
> conntrack template accordingly when OVS first installs the flow containing a
> conntrack action.
>
> There are some additional related items that I intend to work on, which I do
> not see as prerequisite for this series:
> - OVS Connlabel support.
> - Allow OVS to register logging facilities for conntrack.
> - Conntrack per-zone configuration.
>
> The branch below has been updated with the corresponding userspace pieces:
> https://github.com/justinpettit/ovs/tree/conntrack
>
>
> RFCv2:
> - Support IPv4 fragments
> - Warn when ct->net is different from skb net in skb_has_valid_nfct().
> - Set OVS_CS_F_TRACKED when a flow cannot be identified ("invalid")
> - Continue processing packets when conntrack marks the flow invalid.
> - Use PF_INET6 family when sending IPv6 packets to conntrack.
> - Verify conn_* matches when deserializing metadata from netlink.
> - Only allow conntrack action on IPv4/IPv6 packets.
> - Remove explicit dependencies on conn_zone, conn_mark.
> - General tidyups
>
> RFCv1:
> - Rebase to net-next.
> - Add conn_zone field to the flow key.
> - Add explicit dependencies on conn_zone, conn_mark.
> - Refactor conntrack changes into net/openvswitch/ovs_conntrack.*.
> - Don't allow set_field() actions to change conn_state, conn_zone.
> - Add OVS_CS_F_* flags to indicate connection state.
> - Add "invalid" connection state.
>
>
> Andy Zhou (3):
> net: refactor ip_fragment()
> net: Refactor ip_defrag() APIs
> openvswitch: Support fragmented IPv4 packets for conntrack
>
> Joe Stringer (2):
> openvswitch: Serialize acts with original netlink len
> openvswitch: Move MASKED* macros to datapath.h
>
> Justin Pettit (2):
> openvswitch: Add conntrack action
> openvswitch: Allow matching on conntrack mark
>
> drivers/net/macvlan.c | 2 +-
> include/net/ip.h | 13 +-
> include/uapi/linux/openvswitch.h | 42 +++-
> net/ipv4/ip_fragment.c | 46 ++--
> net/ipv4/ip_input.c | 5 +-
> net/ipv4/ip_output.c | 113 +++++----

This is a lot of change to core IP. It probably should be done in its
own patch set.

> net/ipv4/netfilter/nf_defrag_ipv4.c | 2 +-
> net/netfilter/ipvs/ip_vs_core.c | 2 +-
> net/openvswitch/Kconfig | 11 +
> net/openvswitch/Makefile | 1 +
> net/openvswitch/actions.c | 140 +++++++++---
> net/openvswitch/conntrack.c | 427 +++++++++++++++++++++++++++++++++++
> net/openvswitch/conntrack.h | 91 ++++++++
> net/openvswitch/datapath.c | 60 +++--
> net/openvswitch/datapath.h | 10 +
> net/openvswitch/flow.c | 4 +
> net/openvswitch/flow.h | 4 +
> net/openvswitch/flow_netlink.c | 95 ++++++--
> net/openvswitch/flow_netlink.h | 4 +-
> net/openvswitch/vport.c | 1 +
> net/packet/af_packet.c | 2 +-
> 21 files changed, 938 insertions(+), 137 deletions(-)
> create mode 100644 net/openvswitch/conntrack.c
> create mode 100644 net/openvswitch/conntrack.h
>
> --
> 1.7.10.4
>
> --
> To unsubscribe from this list: send the line "unsubscribe netdev" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/