ACK: [0/3] cgroup namespace update

[Tim Gardner]

On 05/16/2016 09:08 AM, Serge Hallyn wrote:
> Quoting Tim Gardner (tim.gardner at canonical.com):
>> On 05/11/2016 09:02 AM, Tim Gardner wrote:
>>> On 05/10/2016 02:43 PM, Serge Hallyn wrote:
>>>> Hi.  During xenial cycle we applied a patch (and a followon fix for criu)
>>>> to fix a usability problem with cgroup namespaces and
>>>> /proc/self/mountinfo.
>>>> Since then, the approach to fix it upstream has changed.  Following are
>>>> reverts for the two patches, plus the new patch cherrypicked from the
>>>> official cgroup for-4.6 tree.
>>>>
>>>> Impact:
>>>> One or the other of these patchsets suffices for docker in lxd to work.
>>>> The patchset being reverted required a docker patch, the new one does
>>>> not.  Docker cannot work inside lxd with neither patchset.
>>>>
>>>>
>>>> thanks,
>>>> -serge
>>>>
>>>
>>> Given that there is no change in functionality, perhaps we can wait
>>> until this patch is upstream (in Linus' repo). Remember to add bug links
>>> next time.
>>>
>>> rtg
>>
>> 3/3 is a clean cherry-pick commit
>> 4f41fc59620fcedaa97cbdf3d7d2956d80fcd922 ('cgroup, kernfs: make
>> mountinfo show properly scoped path for cgroup namespaces')
>>
>> rtg
> 
> hi,
> 
> the following
> 
> https://lkml.org/lkml/2016/5/12/13
> 
> will also be needed.  As discussed above I was waiting until these hit
> Linus' tree to send new patches.
> 

commit 3cc9b23c8137e9bb4eaf6d97d825b7f0dc711d0c ('kernfs:
kernfs_sop_show_path: don't return 0 after seq_dentry call')

-- 
Tim Gardner tim.gardner at canonical.com