[ovs-dev] [PATCHv2 00/11] Backport upstream Linux patches and support 4.12 kernel

[Greg Rose]

On 07/24/2017 10:19 AM, Joe Stringer wrote:
> On 24 July 2017 at 08:31, Greg Rose <gvrose8192 at gmail.com> wrote:
> > On 07/21/2017 04:46 PM, Joe Stringer wrote:
> >>
> >> This series takes Greg's recent proposed series, rolls in a few missed
> >> upstream commits, and orders them the same as the upstream net-next tree.
> >>
> >> Greg, would you mind taking a look at patches #1, #2, #4 and #7? Note that
> >> for
> >> some of the patches it wasn't possible to introduce a function in the
> >> compat
> >> that hides the differences between kernels; for example, when there are
> >> different fields in a structure, we can't hide differences in this way.
> >>
> >> I've added my sign-off to all of the other patches, which I have reviewed.
> >
> >
> > The patch series looks good to me.  I note that you went ahead and
> > backported
> > patches for the current release candidate kernel.  I had held off on them
> > (as noted
> > in my original patch series cover letter) because the kernel wasn't released
> > yet and I
> > was unsure of policy on that.  As it stands I don't think any of the patches
> > for
> > the unreleased kernel will actually be used yet unless someone is building
> > against a
> > 4.13-rc* kernel.  I suppose that can happen but even then they would have to
> > update
> > acinclude.m4 to allow compile against a 4.13 kernel.
>
> I don't think we have a very specific policy in mind for this so far.
> I've discussed with a few people recently that we should probably try
> to ensure that when we do an OVS release branch, we sync up with
> net-next, so that during the pre-release testing we can test the
> latest version of all of the code.

Right.  OK, that's what I was unsure about.

Thanks for the clarification.

- Greg


> I think that this is the most
> useful aspect to this topic. We can, of course, consider syncing up
> more frequently than this, in which case people can start testing it
> out on master as well.
>
> > In any case, going forward I'll go ahead and backport patches even though
> > the kernel
> > they are targeted for has not been released yet.
>
> That sounds reasonable to me. They need to be done at some point, so
> the main aspect is just how much you want to batch this kind of work.
>
> Thanks for the reviews!
>