From: Andy Lutomirski <luto@amacapital.net>
To: Aditya Kali <adityakali@google.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>,
"Serge E. Hallyn" <serge@hallyn.com>,
Linux API <linux-api@vger.kernel.org>,
Linux Containers <containers@lists.linux-foundation.org>,
Serge Hallyn <serge.hallyn@ubuntu.com>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
Tejun Heo <tj@kernel.org>,
cgroups@vger.kernel.org, Ingo Molnar <mingo@redhat.com>
Subject: Re: [PATCHv1 7/8] cgroup: cgroup namespace setns support
Date: Tue, 21 Oct 2014 12:02:37 -0700 [thread overview]
Message-ID: <CALCETrWXDMRsexfvmh2CiMW4WX0ZLJ4pJvzHU55PEBk=NmnyZg@mail.gmail.com> (raw)
In-Reply-To: <CAGr1F2Ee2MCKOwALR2YV7ppDmyHxO6+EsHqSc1+WcwKFPPQB0w@mail.gmail.com>
On Tue, Oct 21, 2014 at 11:49 AM, Aditya Kali <adityakali@google.com> wrote:
> On Mon, Oct 20, 2014 at 10:49 PM, Andy Lutomirski <luto@amacapital.net> wrote:
>> On Mon, Oct 20, 2014 at 10:42 PM, Eric W. Biederman
>> <ebiederm@xmission.com> wrote:
>>>
>>> I do wonder if we think of this as chcgrouproot if there is a simpler
>>> implementation.
>>
>> Could be. I'll defer to Aditya for that one.
>>
>
> More than chcgrouproot, its probably closer to pivot_cgroup_root. In
> addition to restricting the process to a cgroup-root, new processes
> entering the container should also be implicitly contained within the
> cgroup-root of that container.
Why? Concretely, why should this be in the kernel namespace code
instead of in userspace?
> Implementing pivot_cgroup_root would
> probably involve overloading mount-namespace to now understand cgroup
> filesystem too. I did attempt combining cgroupns-root with mntns
> earlier (not via a new syscall though), but came to the conclusion
> that its just simpler to have a separate cgroup namespace and get
> clear semantics. One of the issues was that implicitly changing cgroup
> on setns to mntns seemed like a huge undesirable side-effect.
>
> About pinning: I really feel that it should be OK to pin processes
> within cgroupns-root. I think thats one of the most important feature
> of cgroup-namespace since its most common usecase is to containerize
> un-trusted processes - processes that, for their entire lifetime, need
> to remain inside their container.
So don't let them out. None of the other namespaces have this kind of
constraint:
- If you're in a mntns, you can still use fds from outside.
- If you're in a netns, you can still use sockets from outside the namespace.
- If you're in an ipcns, you can still use ipc handles from outside.
etc.
> And with explicit permission from
> cgroup subsystem (something like cgroup.may_unshare as you had
> suggested previously), we can make sure that unprivileged processes
> cannot pin themselves. Also, maintaining this invariant (your current
> cgroup is always under your cgroupns-root) keeps the code and the
> semantics simple.
I actually think it makes the semantics more complex. The less policy
you stick in the kernel, the easier it is to understand the impact of
that policy.
>
> If we ditch the pinning requirement and allow the containarized
> process to move outside of its cgroupns-root, we will have to address
> atleast the following:
> * what does its /proc/self/cgroup (and /proc/<pid>/cgroup in general)
> look like? We might need to just not show anything in
> /proc/<pid>/cgroup in such case (for default hierarchy).
The process should see the cgroup path relative to its cgroup ns.
Whether this requires a new /proc mount or happens automatically is an
open question. (I *hate* procfs for reasons like this.)
> * how should future setns() and unshare() by such process behave?
Open question.
> * 'mount -t cgroup cgroup <mnt>' by such a process will yield unexpected result
You could disallow that and instead require 'mount -t cgroup -o
cgrouproot=. cgroup mnt' where '.' will be resolved at mount time
relative to the caller's cgroupns.
> * container will not remain migratable
Why not?
> * added code complexity to handle above scenarios
>
> I understand that having process pinned to a cgroup hierarchy might
> seem inconvenient. But even today (without cgroup namespaces), moving
> a task from one cgroup to another can fail for reasons outside of
> control of the task attempting the move (even if its privileged). So
> the userspace should already handle this scenario. I feel its not
> worth to add complexity in the kernel for this.
--Andy
next prev parent reply other threads:[~2014-10-21 19:03 UTC|newest]
Thread overview: 157+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <adityakali-cgroupns>
2014-07-17 19:52 ` [PATCH 0/5] RFC: CGroup Namespaces Aditya Kali
2014-07-17 19:52 ` [PATCH 1/5] kernfs: Add API to get generate relative kernfs path Aditya Kali
2014-07-24 15:10 ` Serge Hallyn
2014-07-17 19:52 ` [PATCH 2/5] sched: new clone flag CLONE_NEWCGROUP for cgroup namespace Aditya Kali
2014-07-24 17:01 ` Serge Hallyn
2014-07-31 19:48 ` Aditya Kali
2014-08-04 23:12 ` Serge Hallyn
2014-07-17 19:52 ` [PATCH 3/5] cgroup: add function to get task's cgroup on default hierarchy Aditya Kali
2014-07-24 16:59 ` Serge Hallyn
2014-07-17 19:52 ` [PATCH 4/5] cgroup: export cgroup_get() and cgroup_put() Aditya Kali
2014-07-24 17:03 ` Serge Hallyn
2014-07-17 19:52 ` [PATCH 5/5] cgroup: introduce cgroup namespaces Aditya Kali
2014-07-17 19:57 ` Andy Lutomirski
2014-07-17 20:55 ` Aditya Kali
2014-07-18 16:51 ` Andy Lutomirski
2014-07-18 18:51 ` Aditya Kali
2014-07-18 18:57 ` Andy Lutomirski
2014-07-21 22:11 ` Aditya Kali
2014-07-21 22:16 ` Andy Lutomirski
2014-07-23 19:52 ` Aditya Kali
2014-07-18 16:00 ` [PATCH 0/5] RFC: CGroup Namespaces Serge Hallyn
2014-07-24 16:10 ` Serge Hallyn
2014-07-24 16:36 ` Serge Hallyn
2014-07-25 19:29 ` Aditya Kali
2014-07-25 20:27 ` Andy Lutomirski
2014-07-29 4:51 ` Serge E. Hallyn
2014-07-29 15:08 ` Andy Lutomirski
2014-07-29 16:06 ` Serge E. Hallyn
2014-10-13 21:23 ` [PATCHv1 0/8] " Aditya Kali
2014-10-13 21:23 ` [PATCHv1 1/8] kernfs: Add API to generate relative kernfs path Aditya Kali
2014-10-16 16:07 ` Serge E. Hallyn
2014-10-13 21:23 ` [PATCHv1 2/8] sched: new clone flag CLONE_NEWCGROUP for cgroup namespace Aditya Kali
2014-10-16 16:08 ` Serge E. Hallyn
2014-10-13 21:23 ` [PATCHv1 3/8] cgroup: add function to get task's cgroup on default hierarchy Aditya Kali
2014-10-16 16:13 ` Serge E. Hallyn
2014-10-13 21:23 ` [PATCHv1 4/8] cgroup: export cgroup_get() and cgroup_put() Aditya Kali
2014-10-16 16:14 ` Serge E. Hallyn
2014-10-13 21:23 ` [PATCHv1 5/8] cgroup: introduce cgroup namespaces Aditya Kali
2014-10-16 16:37 ` Serge E. Hallyn
2014-10-24 1:03 ` Aditya Kali
2014-10-25 3:16 ` Serge E. Hallyn
2014-10-13 21:23 ` [PATCHv1 6/8] cgroup: restrict cgroup operations within task's cgroupns Aditya Kali
2014-10-17 9:28 ` Serge E. Hallyn
2014-10-22 19:06 ` Aditya Kali
2014-10-19 4:57 ` Eric W. Biederman
2014-10-13 21:23 ` [PATCHv1 7/8] cgroup: cgroup namespace setns support Aditya Kali
2014-10-16 21:12 ` Serge E. Hallyn
2014-10-16 21:17 ` Andy Lutomirski
2014-10-16 21:22 ` Aditya Kali
2014-10-16 21:47 ` Serge E. Hallyn
2014-10-19 5:23 ` Eric W. Biederman
2014-10-19 18:26 ` Andy Lutomirski
2014-10-20 4:55 ` Eric W.Biederman
2014-10-21 0:20 ` Andy Lutomirski
2014-10-21 4:49 ` Eric W. Biederman
2014-10-21 5:03 ` Andy Lutomirski
2014-10-21 5:42 ` Eric W. Biederman
2014-10-21 5:49 ` Andy Lutomirski
2014-10-21 18:49 ` Aditya Kali
2014-10-21 19:02 ` Andy Lutomirski [this message]
2014-10-21 22:33 ` Aditya Kali
2014-10-21 22:42 ` Andy Lutomirski
2014-10-22 0:46 ` Aditya Kali
2014-10-22 0:58 ` Andy Lutomirski
2014-10-22 18:37 ` Aditya Kali
2014-10-22 18:50 ` Andy Lutomirski
2014-10-22 19:42 ` Tejun Heo
2014-10-17 9:52 ` Serge E. Hallyn
2014-10-13 21:23 ` [PATCHv1 8/8] cgroup: mount cgroupns-root when inside non-init cgroupns Aditya Kali
2014-10-17 12:19 ` Serge E. Hallyn
2014-10-14 22:42 ` [PATCHv1 0/8] CGroup Namespaces Andy Lutomirski
2014-10-14 23:33 ` Aditya Kali
2014-10-19 4:54 ` Eric W. Biederman
2015-07-22 18:10 ` Vincent Batts
2014-10-31 19:18 ` [PATCHv2 0/7] " Aditya Kali
2014-10-31 19:18 ` [PATCHv2 1/7] kernfs: Add API to generate relative kernfs path Aditya Kali
2014-10-31 19:18 ` [PATCHv2 2/7] sched: new clone flag CLONE_NEWCGROUP for cgroup namespace Aditya Kali
2014-10-31 19:18 ` [PATCHv2 3/7] cgroup: add function to get task's cgroup on default hierarchy Aditya Kali
2014-10-31 19:18 ` [PATCHv2 4/7] cgroup: export cgroup_get() and cgroup_put() Aditya Kali
2014-10-31 19:18 ` [PATCHv2 5/7] cgroup: introduce cgroup namespaces Aditya Kali
2014-11-01 0:02 ` Andy Lutomirski
2014-11-01 0:58 ` Eric W. Biederman
2014-11-03 23:42 ` Aditya Kali
2014-11-03 23:40 ` Aditya Kali
2014-11-04 1:56 ` Aditya Kali
2014-10-31 19:19 ` [PATCHv2 6/7] cgroup: cgroup namespace setns support Aditya Kali
2014-10-31 19:19 ` [PATCHv2 7/7] cgroup: mount cgroupns-root when inside non-init cgroupns Aditya Kali
2014-11-01 0:07 ` Andy Lutomirski
2014-11-01 2:59 ` Eric W. Biederman
2014-11-01 3:29 ` Andy Lutomirski
2014-11-03 23:12 ` Aditya Kali
2014-11-03 23:15 ` Andy Lutomirski
2014-11-03 23:23 ` Aditya Kali
2014-11-03 23:48 ` Andy Lutomirski
2014-11-04 0:12 ` Aditya Kali
2014-11-04 0:17 ` Andy Lutomirski
2014-11-04 0:49 ` Aditya Kali
2014-11-04 13:57 ` Tejun Heo
2014-11-06 17:28 ` Aditya Kali
2014-11-01 1:09 ` Eric W. Biederman
2014-11-03 22:46 ` Aditya Kali
[not found] ` <CAGr1F2Hd_PS_AscBGMXdZC9qkHGRUp-MeQvJksDOQkRBB3RGoA@mail.gmail.com>
2014-11-03 22:56 ` Andy Lutomirski
2014-11-04 13:46 ` Tejun Heo
2014-11-04 15:00 ` Andy Lutomirski
2014-11-04 15:50 ` Serge E. Hallyn
2014-11-12 17:48 ` Aditya Kali
2014-11-04 1:59 ` Aditya Kali
2014-11-04 13:10 ` [PATCHv2 0/7] CGroup Namespaces Vivek Goyal
2014-11-06 17:33 ` Aditya Kali
2014-11-26 22:58 ` Richard Weinberger
2014-12-02 19:14 ` Aditya Kali
2014-12-05 1:55 ` [PATCHv3 0/8] " Aditya Kali
2014-12-05 1:55 ` [PATCHv3 1/8] kernfs: Add API to generate relative kernfs path Aditya Kali
2014-12-05 1:55 ` [PATCHv3 2/8] sched: new clone flag CLONE_NEWCGROUP for cgroup namespace Aditya Kali
2014-12-05 1:55 ` [PATCHv3 3/8] cgroup: add function to get task's cgroup on default hierarchy Aditya Kali
2014-12-05 1:55 ` [PATCHv3 4/8] cgroup: export cgroup_get() and cgroup_put() Aditya Kali
2014-12-05 1:55 ` [PATCHv3 5/8] cgroup: introduce cgroup namespaces Aditya Kali
2014-12-12 8:54 ` Zefan Li
2014-12-05 1:55 ` [PATCHv3 6/8] cgroup: cgroup namespace setns support Aditya Kali
2014-12-05 1:55 ` [PATCHv3 7/8] cgroup: mount cgroupns-root when inside non-init cgroupns Aditya Kali
2014-12-12 8:55 ` Zefan Li
2014-12-05 1:55 ` [PATCHv3 8/8] cgroup: Add documentation for cgroup namespaces Aditya Kali
2014-12-12 8:54 ` Zefan Li
2015-01-05 22:54 ` Aditya Kali
2014-12-14 23:05 ` Richard Weinberger
2015-01-05 22:48 ` Aditya Kali
2015-01-05 22:52 ` Richard Weinberger
2015-01-05 23:53 ` Eric W. Biederman
2015-01-06 0:07 ` Richard Weinberger
2015-01-06 0:10 ` Aditya Kali
2015-01-06 0:17 ` Richard Weinberger
2015-01-06 23:20 ` Aditya Kali
2015-01-06 23:39 ` Richard Weinberger
2015-01-07 9:28 ` Richard Weinberger
2015-01-07 14:45 ` Eric W. Biederman
2015-01-07 19:30 ` Serge E. Hallyn
2015-01-07 22:14 ` Eric W. Biederman
2015-01-07 22:45 ` Tejun Heo
2015-01-07 23:02 ` Eric W. Biederman
2015-01-07 23:06 ` Tejun Heo
2015-01-07 23:09 ` Eric W. Biederman
2015-01-07 23:16 ` Tejun Heo
2015-01-07 23:27 ` Eric W. Biederman
2015-01-07 23:35 ` Tejun Heo
2015-02-11 3:46 ` Serge E. Hallyn
2015-02-11 4:09 ` Tejun Heo
2015-02-11 4:29 ` Serge E. Hallyn
2015-02-11 5:02 ` Eric W. Biederman
2015-02-11 5:17 ` Tejun Heo
2015-02-11 6:29 ` Eric W. Biederman
2015-02-11 14:36 ` Tejun Heo
2015-02-11 16:00 ` Serge E. Hallyn
2015-02-11 16:03 ` Tejun Heo
2015-02-11 16:18 ` Serge E. Hallyn
2015-02-11 5:10 ` Tejun Heo
2015-01-07 18:57 ` Aditya Kali
2014-12-05 3:20 ` [PATCHv3 0/8] CGroup Namespaces Aditya Kali
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CALCETrWXDMRsexfvmh2CiMW4WX0ZLJ4pJvzHU55PEBk=NmnyZg@mail.gmail.com' \
--to=luto@amacapital.net \
--cc=adityakali@google.com \
--cc=cgroups@vger.kernel.org \
--cc=containers@lists.linux-foundation.org \
--cc=ebiederm@xmission.com \
--cc=linux-api@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=serge.hallyn@ubuntu.com \
--cc=serge@hallyn.com \
--cc=tj@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).