From: Serge Hallyn <serge@hallyn.com>
To: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>,
Aristeu Rozanski <aris@ruivo.org>,
Neil Horman <nhorman@tuxdriver.com>,
"Serge E. Hallyn" <serue@us.ibm.com>,
containers@lists.linux-foundation.org,
linux-kernel@vger.kernel.org, Michal Hocko <mhocko@suse.cz>,
Thomas Graf <tgraf@suug.ch>, Paul Mackerras <paulus@samba.org>,
"Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>,
Arnaldo Carvalho de Melo <acme@ghostprotocols.net>,
Johannes Weiner <hannes@cmpxchg.org>, Tejun Heo <tj@kernel.org>,
cgroups@vger.kernel.org, Paul Turner <pjt@google.com>,
Ingo Molnar <mingo@redhat.com>
Subject: Re: Controlling devices and device namespaces
Date: Sun, 16 Sep 2012 11:15:38 -0500 [thread overview]
Message-ID: <5055FB2A.1020103@hallyn.com> (raw)
In-Reply-To: <87k3vuqc5l.fsf@xmission.com>
On 09/16/2012 09:23 AM, Eric W. Biederman wrote:
> Serge Hallyn <serge@hallyn.com> writes:
>
>> On 09/16/2012 07:17 AM, Eric W. Biederman wrote:
>>> ebiederm@xmission.com (Eric W. Biederman) writes:
>>>
>>>> Alan Cox <alan@lxorguk.ukuu.org.uk> writes:
>>>>
>>>>>> One piece of the puzzle is that we should be able to allow unprivileged
>>>>>> device node creation and access for any device on any filesystem
>>>>>> for which it unprivileged access is safe.
>>>>>
>>>>> Which devices are "safe" is policy for all interesting and useful cases,
>>>>> as are file permissions, security tags, chroot considerations and the
>>>>> like.
>>>>>
>>>>> It's a complete non starter.
>>>
>>> Come to think of it mknod is completely unnecessary.
>>>
>>> Without mknod. Without being able to mount filesystems containing
>>> device nodes.
>>
>> Hm? That sounds like it will really upset init/udev/upgrades in the
>> container.
>
> udev does not create device nodes. For an older udev the worst
> I can see it doing is having mknod failing with EEXIST because
> the device node already exists.
>
> We should be able to make it look to init like a ramdisk mounted the
> filesystems.
>
> Why should upgrades care? Package installation shouldn't be calling
> mknod.
>
> At least with a recent modern distro I can't imagine this to be an
> issue. I expect we could have a kernel build option that removed the
> mknod system call and a modern distro wouldn't notice.
>
>> Are you saying all filesystems containing device nodes will need to be
>> mounted in advance by the process setting up the container?
>
> As a general rule.
>
> I think in practice there is wiggle room for special cases
> like mounting a fresh devpts. devpts at least in always create a new
> instance on mount mode seems safe, as it can not give you access to
> any existing devices.
>
> You can also do a lot of what would normally be done with mknod
> with bind mounts to the original devices location.
>
>>> The mount namespace is sufficient to prevent all of the
>>> cases that the device control group prevents (open and mknod on device
>>> nodes).
>>>
>>> So I honestly think the device control group is superflous, and it is
>>> probably wise to deprecate it and move to a model where it does not
>>> exist.
>>>
>>> Eric
>>>
>>
>> That's what I said a few emails ago :) The device cgroup was meant as
>> a short-term workaround for lack of user (and device) namespaces.
>
> I am saying something stronger. The device cgroup doesn't seem to have
> a practical function now.
"Now" is wrong. The user namespace is not complete and not yet usable
for a full system container. We still need the device control group.
I'd like us to have a sprint (either a day at UDS in person, or a few
days with a virtual sprint) with the focus of getting a full system
container working the way you envision it, as cleanly as possible. I
can take two or three consecutave days sometime in the next 2-3 weeks,
we can sit on irc and share a few instances on which to experiment?
> That for the general case we don't need any
> kernel support. That all of this should be a matter of some user space
> glue code, and just the tiniest bit of sorting out how hotplug events are
> sent.
>
> The only thing I can think we would need a device namespace for is
> for migration.
>
> For migration with direct access to real hardware devices we must treat
> it as hardware hotunplug. There is nothing else we can do.
>
> If there is any other case where we need to preserve device numbers
> etc we have the example of devpts.
>
> So at this point I really don't think we need a device namespace or a
> device control group. (Just emulate devtmpfs, sysfs and uevents).
>
> Eric
>
next prev parent reply other threads:[~2012-09-16 16:16 UTC|newest]
Thread overview: 70+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-09-13 20:58 [RFC] cgroup TODOs Tejun Heo
2012-09-14 9:04 ` Mike Galbraith
2012-09-14 17:17 ` Tejun Heo
2012-09-14 9:10 ` Daniel P. Berrange
2012-09-14 13:58 ` Vivek Goyal
2012-09-14 19:29 ` Tejun Heo
2012-09-14 21:51 ` Kay Sievers
[not found] ` <5052E7DF.7040000@parallels.com>
2012-09-14 9:12 ` Li Zefan
2012-09-14 11:22 ` Peter Zijlstra
2012-09-14 17:59 ` Tejun Heo
2012-09-14 18:23 ` Peter Zijlstra
2012-09-14 18:33 ` Tejun Heo
2012-09-14 17:43 ` Tejun Heo
2012-09-17 8:50 ` Glauber Costa
2012-09-17 17:21 ` Tejun Heo
2012-09-14 11:15 ` Peter Zijlstra
2012-09-14 12:54 ` Daniel P. Berrange
2012-09-14 17:53 ` Tejun Heo
2012-09-14 14:25 ` Vivek Goyal
2012-09-14 14:53 ` Peter Zijlstra
2012-09-14 15:14 ` Vivek Goyal
2012-09-14 21:57 ` Tejun Heo
2012-09-17 15:27 ` Vivek Goyal
2012-09-18 18:08 ` Vivek Goyal
2012-09-14 21:39 ` Tejun Heo
2012-09-17 15:05 ` Vivek Goyal
2012-09-17 16:40 ` Tejun Heo
2012-09-14 15:03 ` Michal Hocko
2012-09-19 14:02 ` Michal Hocko
2012-09-19 14:03 ` [PATCH 2.6.32] memcg: warn on deeper hierarchies with use_hierarchy==0 Michal Hocko
2012-09-19 19:38 ` David Rientjes
2012-09-20 13:24 ` Michal Hocko
2012-09-20 22:33 ` David Rientjes
2012-09-21 7:16 ` Michal Hocko
2012-09-19 14:03 ` [PATCH 3.0] " Michal Hocko
2012-09-19 14:05 ` [PATCH 3.2+] " Michal Hocko
2012-09-14 18:07 ` [RFC] cgroup TODOs Vivek Goyal
2012-09-14 18:53 ` Tejun Heo
2012-09-14 19:28 ` Vivek Goyal
2012-09-14 19:44 ` Tejun Heo
2012-09-14 19:49 ` Tejun Heo
2012-09-14 20:39 ` Tejun Heo
2012-09-17 8:40 ` Glauber Costa
2012-09-17 17:30 ` Tejun Heo
2012-09-17 14:37 ` Vivek Goyal
2012-09-14 18:36 ` Aristeu Rozanski
2012-09-14 18:54 ` Tejun Heo
2012-09-15 2:20 ` Serge E. Hallyn
2012-09-15 9:27 ` Controlling devices and device namespaces Eric W. Biederman
2012-09-15 22:05 ` Serge E. Hallyn
2012-09-16 0:24 ` Eric W. Biederman
2012-09-16 3:31 ` Serge E. Hallyn
2012-09-16 11:21 ` Alan Cox
2012-09-16 11:56 ` Eric W. Biederman
2012-09-16 12:17 ` Eric W. Biederman
2012-09-16 13:32 ` Serge Hallyn
2012-09-16 14:23 ` Eric W. Biederman
2012-09-16 16:13 ` Alan Cox
2012-09-16 17:49 ` Eric W. Biederman
2012-09-16 16:15 ` Serge Hallyn [this message]
2012-09-16 16:53 ` Eric W. Biederman
2012-09-16 8:19 ` [RFC] cgroup TODOs James Bottomley
2012-09-16 14:41 ` Eric W. Biederman
2012-09-17 13:21 ` Aristeu Rozanski
2012-09-14 22:03 ` Dhaval Giani
2012-09-14 22:06 ` Tejun Heo
2012-09-20 1:33 ` Andy Lutomirski
2012-09-20 18:26 ` Tejun Heo
2012-09-20 18:39 ` Andy Lutomirski
2012-09-21 21:40 ` Tejun Heo
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=5055FB2A.1020103@hallyn.com \
--to=serge@hallyn.com \
--cc=acme@ghostprotocols.net \
--cc=alan@lxorguk.ukuu.org.uk \
--cc=aneesh.kumar@linux.vnet.ibm.com \
--cc=aris@ruivo.org \
--cc=cgroups@vger.kernel.org \
--cc=containers@lists.linux-foundation.org \
--cc=ebiederm@xmission.com \
--cc=hannes@cmpxchg.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mhocko@suse.cz \
--cc=mingo@redhat.com \
--cc=nhorman@tuxdriver.com \
--cc=paulus@samba.org \
--cc=pjt@google.com \
--cc=serue@us.ibm.com \
--cc=tgraf@suug.ch \
--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).