Re: CGroup Namespaces (v8)

From: "Serge E. Hallyn" <serge.hallyn@ubuntu.com>
To: Dan Williams <dan.j.williams@intel.com>
Cc: "Serge E. Hallyn" <serge.hallyn@ubuntu.com>,
	Stephen Rothwell <sfr@canb.auug.org.au>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	linux-api@vger.kernel.org, containers@lists.linux-foundation.org,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	"Eric W. Biederman" <ebiederm@xmission.com>,
	lxc-devel@lists.linuxcontainers.org,
	Johannes Weiner <hannes@cmpxchg.org>, Tejun Heo <tj@kernel.org>,
	cgroups@vger.kernel.org,
	Andrew Morton <akpm@linux-foundation.org>
Subject: Re: CGroup Namespaces (v8)
Date: Fri, 1 Jan 2016 12:06:33 -0600	[thread overview]
Message-ID: <20160101180633.GA30941@mail.hallyn.com> (raw)
In-Reply-To: <CAPcyv4iig_sujK=p0TbNjTmtv39N7a0HSqpVe8mLnfX9Es7zNA@mail.gmail.com>

On Fri, Jan 01, 2016 at 01:42:57AM -0800, Dan Williams wrote:
> On Fri, Jan 1, 2016 at 12:59 AM, Serge E. Hallyn
> <serge.hallyn@ubuntu.com> wrote:
> > On Fri, Jan 01, 2016 at 12:19:44AM -0800, Dan Williams wrote:
> >> On Tue, Dec 22, 2015 at 8:23 PM,  <serge.hallyn@ubuntu.com> wrote:
> >> > Hi,
> >> >
> >> > following is a revised set of the CGroup Namespace patchset which Aditya
> >> > Kali has previously sent.  The code can also be found in the cgroupns.v8
> >> > branch of
> >> >
> >> > https://git.kernel.org/cgit/linux/kernel/git/sergeh/linux-security.git/
> >>
> >> Happy New Year!
> >>
> >> I'm hitting both a build and boot regression in -next with these
> >> patches.  The bisect log reports:
> >>
> >> git bisect skip f176ae3a5df610e0373c39538cc7744c7788b3df
> >> # bad: [54b39d263704b66cc8006ce8f7dd4743fdabae70] cgroup: cgroup
> >> namespace setns support
> >> git bisect bad 54b39d263704b66cc8006ce8f7dd4743fdabae70
> >> # good: [bf9b3590879afbaa54148d817ed68a0b51180375] sched: new clone
> >> flag CLONE_NEWCGROUP for c
> >> group namespace
> >> git bisect good bf9b3590879afbaa54148d817ed68a0b51180375
> >> # only skipped commits left to test
> >> # possible first bad commit:
> >> [54b39d263704b66cc8006ce8f7dd4743fdabae70] cgroup: cgroup namespa
> >> ce setns support
> >> # possible first bad commit:
> >> [f176ae3a5df610e0373c39538cc7744c7788b3df] cgroup: introduce cgro
> >> up namespaces
> >>
> >> Where compiling "f176ae3a5df6"  fails with:
> >>
> >> kernel/cgroup.c: In function ‘cgroupns_put’:
> >> kernel/cgroup.c:5956:2: error: implicit declaration of function
> >> ‘to_cg_ns’ [-Werror=implicit-f
> >> unction-declaration]
> >>  put_cgroup_ns(to_cg_ns(ns));
> >
> > Hi,
> >
> > is it possible you have a bad tree?
> 
> No, I don't believe so...
> 
> >  Looking at
> > https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git/tree/kernel/cgroup.c?id=refs/tags/next-20151231
> > to_cg_ns should definately be defined there.
> >
> 
> Right, but it's not defined until commit 54b39d263704 "cgroup: cgroup
> namespace setns support", however it is first used in the preceeding
> commit f176ae3a5df6 "cgroup: introduce cgroup namespaces".  The bad
> patch ordering breaks bisection.
> 
> Commit 54b39d263704 "cgroup: cgroup namespace setns support" not
> booting is a separate issue.

Oh - been there since my first version of the set (v4).  Odd, I
thought that the automated korg testing caught those.

What is the simplest way to fix this?  Do I send new versions of
patches v3 and v4?  Does Tejun or Stephen just do it inline in the
git tree?  Do we leave it be?