[ovs-dev] [PATCHv2 00/20] Conntrack datapath backport.

Joe Stringer joe at ovn.org
Fri Dec 4 01:45:41 UTC 2015


On 3 December 2015 at 16:54, Pravin Shelar <pshelar at nicira.com> wrote:

> On Wed, Dec 2, 2015 at 11:53 PM, Joe Stringer <joe at ovn.org> wrote:
> > From: Joe Stringer <joestringer at nicira.com>
> >
> > This series backports the support for connection tracking introduced in
> > Linux 4.3, allowing the kernel module in the tree to support this feature
> > on kernels 3.10 and newer.
> >
> > This series applies on top of the lwtunnel backport series v7 from
> Pravin.
> > For ease of review, I have made a combined branch available here:
> >
> > https://github.com/joestringer/openvswitch/tree/submit/ct_backport_v2
> >
> > v2:
> > - Fixed nf_conntrack_ftp module refcount issue
> > - Fix template free on backported kernels
> > - Fixed build on RHEL7.2 kernels
> > - Fixed various compiler warnings
> > - Dropped testsuite patches that were applied
> > - Rebase
> >
> > Joe Stringer (20):
> >   compat: Backport conntrack zones headers.
> >   compat: Backport nf_ct_tmpl_alloc().
> >   compat: Backport nf_connlabels_{get,put}().
> >   compat: Backport 'dst' functions.
> >   compat: Backport prandom_u32_max().
> >   compat: Backport dev_recursion_level().
> >   compat: Backport ip_skb_dst_mtu().
> >   compat: Wrap IPv4 fragmentation.
> >   compat: Backport IPv4 reassembly.
> >   compat: Backport IPv6 fragmentation.
> >   compat: Backport IPv6 reassembly.
> >   datapath: Move MASKED* macros to datapath.h
> >   datapath: Serialize acts with original netlink len
> >   datapath: Add conntrack action
> >   datapath: Allow matching on conntrack mark
> >   datapath: Allow matching on conntrack label
> >   datapath: Allow attaching helpers to ct action
> >   datapath: Backport conntrack fixes.
> >   datapath: Scrub skb between namespaces
> >   FAQ: Document kernel feature support.
> >
>
> Thanks for the patch.
>
> This patch set is also really big so it is not possible to review to
> in detail in given time. I do not see any problem with the basic
> structure of the ip-frag backport so I am fine with it going in.
>
> Acked-by: Pravin B Shelar <pshelar at nicira.com>
>

Thanks for the review. I figure that it is unlikely to get further review -
the diff from v1 is small, and v1 didn't receive any further review over
the past week, so I did another run through of the kernel testsuite on a
variety of kernel targets and with them passing, pushed the series to
master. It's been a while in the works :-)



More information about the dev mailing list