Kernel development for Yakkety

Seth Forshee seth.forshee at canonical.com
Wed Apr 27 18:12:22 UTC 2016


On Tue, Apr 26, 2016 at 02:19:20PM -0600, Tim Gardner wrote:
> On 04/26/2016 02:11 PM, Seth Forshee wrote:
> >On Tue, Apr 26, 2016 at 01:25:28PM -0600, Tim Gardner wrote:
> >>I've been working on getting some semblance of a Yakkety kernel working. You
> >>can follow along by watching
> >>'git://git.launchpad.net/~ubuntu-kernel/ubuntu/+source/linux/+git/unstable
> >>master'. Beware, some days this branch gets massively rebased as often as
> >>several times in an hour (especially on Mondays when the rc's are released).
> >>
> >>There are currently PIE related toolchain issues in Yakkety, so I've
> >>continued to build using the Xenial toolchain.
> >>
> >>The initial rebase from Xenial had some patch fallout:
> >>
> >>Seth - Could you have a look at these patches. I assume they are still
> >>necessary.
> >>
> >>UBUNTU: SAUCE: ext4: Add support for unprivileged mounts from user
> >>namespaces
> >>UBUNTU: SAUCE: ext4: Add module parameter to enable user namespace mounts
> >
> >Yes, though I do have updates for the former related to changes which
> >came in after 4.4.
> >
> >>You should also have a look at this backport to ensure that I did it
> >>correctly:
> >>
> >>UBUNTU: SAUCE: cgroup: Use a new super block when mounting in a cgroup
> >>namespace
> >
> >That looks okay ... I'm surprised that there were even any conflicts.
> >
> >I'm already rebasing that series onto the current -rc kernel pretty
> >frequently, so it seems kind of silly for both of us to be resolving
> >conflicts. If you want I could just start giving you a branch as needed
> >for you to cherry pick updated patches, or something like that.
> >
> >Seth
> >
> 
> That works for me.

git://git.launchpad.net/~sforshee/+git/linux-unstable nsmount

I've smoke tested that in a yakkity vm and it worked fine there.

What I ended up doing was dropping all the namespace mount commits
carried forward from xenial and applying patches from my upstream branch
on the end, plus the patches from xenial for the module parameters which
are not part of my upstream branch. I also had to fix up a couple of
minor conflicts which resulted from the reordering. 

Let me know how that works out, and if there's anything I should do
differently in the future to make things easier.




More information about the kernel-team mailing list