From: ebiederm@xmission.com (Eric W. Biederman)
To: "Serge E. Hallyn" <serue@us.ibm.com>
Cc: linux-kernel@vger.kernel.org,
"Alan Cox <alan@lxorguk.ukuu.org.uk> Dave Hansen"
<haveblue@us.ibm.com>, Arjan van de Ven <arjan@infradead.org>,
Suleiman Souhlal <ssouhlal@FreeBSD.org>,
Hubertus Franke <frankeh@watson.ibm.com>,
Cedric Le Goater <clg@fr.ibm.com>
Subject: Re: RFC: Multiple instances of kernel namespaces.
Date: Sat, 21 Jan 2006 03:04:16 -0700 [thread overview]
Message-ID: <m13bjhoq1r.fsf@ebiederm.dsl.xmission.com> (raw)
In-Reply-To: <20060120201353.GA13265@sergelap.austin.ibm.com> (Serge E. Hallyn's message of "Fri, 20 Jan 2006 14:13:53 -0600")
"Serge E. Hallyn" <serue@us.ibm.com> writes:
> Quoting Eric W. Biederman (ebiederm@xmission.com):
>>
>> At this point I have to confess I have been working on something
>> similar, to IBM's pid virtualization work. But I have what is at
>> least for me a unifying concept, that makes things easier to think
>> about.
>>
>> The idea is to think about things in terms of namespaces. Currently
>> in the kernel we have the fs/mount namespace already implemented.
>>
>> Partly this helps on what the interface for creating a new namespace
>> instance should be. 'clone(CLONE_NEW<NAMESPACE_TYPE>)', and how
>> it should be managed from the kernel data structures.
>>
>> Partly thinking of things as namespaces helps me scope the problem.
>>
>> Does this sound like a sane approach?
>
> And a bonus of this is that for security and vserver-type applications,
> the CLONE_NEWPID and CLONE_NEWFS will often happen at the same time.
>
> How do you (or do you?) address naming namespaces? This would be
> necessary for transitioning into an existing namespace, performing
> actions on existing namespaces (i.e. checkpoint, migrate to another
> machine, enter the namespace and kill pid 521), and would just be
> useful for accounting purposes, i.e. how else do you have a
> "ps --all-namespaces" specify a process' namespace?
So I address naming indirectly. The last thing I want to have
is to add yet another namespace to the kernel for naming namespaces.
We have enough namespaces already.
In any sane context for a pid-namespace we need a pid that
we can call waitpid on, so we don't break the process tree.
Which means at least the init process has 2 pids, one
that it's parent sees, and another (1) that it and it's
children see.
So I name pidspaces like we do sessions of process groups
and sessions by the pid of the leader.
So in the simple case I have names like:
1178/1632
> Doubt we want to add an argument to clone(), so do we just add a new
> proc, sysfs, or syscall for setting a pid-namespace name?
That shouldn't be necessary.
> Do we need a new syscall for transitioning into an existing namespace?
That is a good question. The FS namespaces that we already have
has much the same problem. A completely different solution to
this problem seems to have been implemented but I don't grasp it
yet.
Inherently transitioning to an existing namespace is something
that is straight forward to implement, so it is worth thinking
about.
If I want a guest that can keep secrets from the host sysadmin I don't
want transitioning into a guest namespace to come too easily.
Currently I can always just create an extra child of pid 1
that I will be my slave. The problem is that this is an extra
process laying around.
Eric
next prev parent reply other threads:[~2006-01-21 10:05 UTC|newest]
Thread overview: 136+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-01-17 14:32 RFC [patch 00/34] PID Virtualization Overview Serge Hallyn
2006-01-17 14:32 ` RFC [patch 01/34] PID Virtualization Change pid accesses: drivers Serge Hallyn
2006-01-17 14:33 ` RFC [patch 02/34] PID Virtualization Change pid accesses: most archs Serge Hallyn
2006-01-17 14:33 ` RFC [patch 03/34] PID Virtualization Change pid accesses: filesystems Serge Hallyn
2006-01-17 14:33 ` RFC [patch 04/34] PID Virtualization Change pid accesses: include/ Serge Hallyn
2006-01-17 14:33 ` RFC [patch 05/34] PID Virtualization Change pid accesses: ipc Serge Hallyn
2006-01-17 14:33 ` RFC [patch 06/34] PID Virtualization Change pid accesses: kernel/ Serge Hallyn
2006-01-17 14:33 ` RFC [patch 07/34] PID Virtualization Change pid accesses: lib/ Serge Hallyn
2006-01-17 14:33 ` RFC [patch 08/34] PID Virtualization Change pid accesses: mm/ Serge Hallyn
2006-01-17 14:33 ` RFC [patch 09/34] PID Virtualization Change pid accesses: net/ Serge Hallyn
2006-01-17 14:33 ` RFC [patch 10/34] PID Virtualization Change pid accesses: security/ Serge Hallyn
2006-01-17 14:33 ` RFC [patch 11/34] PID Virtualization Change pid accesses: sound/ Serge Hallyn
2006-01-17 14:33 ` RFC [patch 12/34] PID Virtualization Change pid accesses: ia64 and mips Serge Hallyn
2006-01-17 14:33 ` RFC [patch 13/34] PID Virtualization Define new task_pid api Serge Hallyn
2006-01-17 15:32 ` Arjan van de Ven
2006-01-17 15:56 ` Serge E. Hallyn
2006-01-17 16:02 ` Arjan van de Ven
2006-01-17 16:03 ` Alan Cox
2006-01-17 17:16 ` Kyle Moffett
2006-01-17 17:25 ` Dave Hansen
2006-01-18 4:54 ` Greg KH
2006-01-18 4:55 ` Greg KH
2006-01-18 16:23 ` Dave Hansen
2006-01-20 17:00 ` Eric W. Biederman
2006-01-20 20:18 ` Hubertus Franke
2006-01-21 10:25 ` Eric W. Biederman
2006-01-23 18:38 ` Hubertus Franke
2006-01-23 18:48 ` Eric W. Biederman
2006-01-21 14:42 ` Eric W. Biederman
2006-01-22 6:43 ` Kyle Moffett
2006-01-22 15:48 ` Eric W. Biederman
2006-01-22 15:55 ` Arjan van de Ven
2006-01-22 16:24 ` Eric W. Biederman
2006-01-26 20:01 ` Herbert Poetzl
2006-01-27 9:04 ` Eric W. Biederman
2006-01-27 12:27 ` Kyle Moffett
2006-01-27 13:15 ` Eric W. Biederman
2006-01-23 18:50 ` Hubertus Franke
2006-01-23 19:28 ` Eric W. Biederman
2006-01-23 21:11 ` Alan Cox
2006-01-23 21:30 ` Eric W. Biederman
2006-01-23 22:15 ` Hubertus Franke
2006-01-24 6:56 ` Arjan van de Ven
2006-01-24 19:34 ` Eric W. Biederman
2006-01-24 21:09 ` Hubertus Franke
2006-01-24 0:22 ` Alan Cox
2006-01-24 19:26 ` Eric W. Biederman
2006-01-24 21:11 ` Alan Cox
2006-01-24 21:15 ` Arjan van de Ven
2006-01-25 9:58 ` Eric W. Biederman
2006-01-25 15:10 ` Trond Myklebust
2006-01-25 18:01 ` Eric W. Biederman
2006-01-25 19:30 ` Trond Myklebust
2006-01-25 21:59 ` Eric W. Biederman
2006-01-25 9:13 ` Eric W. Biederman
2006-01-25 9:51 ` Eric W. Biederman
2006-01-26 20:23 ` Herbert Poetzl
2006-01-27 8:28 ` Eric W. Biederman
[not found] ` <m1k6cqlmfe.fsf_-_@ebiederm.dsl.xmission.com>
2006-01-23 21:57 ` RFC: [PATCH] pids as weak references Dave Hansen
2006-01-31 21:02 ` RFC [patch 13/34] PID Virtualization Define new task_pid api Linus Torvalds
2006-02-01 0:01 ` Hubertus Franke
2006-02-01 4:18 ` Eric W. Biederman
2006-02-01 4:39 ` Linus Torvalds
2006-02-01 7:14 ` Eric W. Biederman
2006-02-01 16:41 ` Dave Hansen
2006-02-02 5:14 ` Herbert Poetzl
2006-02-01 16:29 ` Greg
2006-02-01 16:44 ` Eric W. Biederman
2006-02-02 13:50 ` Greg
2006-02-02 14:09 ` Eric W. Biederman
2006-02-02 14:48 ` Kirill Korotaev
2006-02-02 15:13 ` Eric W. Biederman
2006-02-02 15:26 ` Kirill Korotaev
2006-02-02 15:51 ` Eric W. Biederman
2006-02-02 16:05 ` Kirill Korotaev
2006-02-02 16:27 ` Eric W. Biederman
2006-02-02 21:32 ` Cedric Le Goater
2006-02-02 21:43 ` Hubertus Franke
2006-02-02 21:46 ` Eric W. Biederman
2006-02-03 10:07 ` Kirill Korotaev
2006-02-03 10:52 ` Kirill Korotaev
2006-02-03 11:09 ` Eric W. Biederman
2006-02-03 15:45 ` Dave Hansen
2006-02-03 16:35 ` Kirill Korotaev
2006-02-02 21:10 ` Cedric Le Goater
2006-02-02 21:24 ` Eric W. Biederman
2006-02-06 20:15 ` Pavel Machek
2006-02-06 20:34 ` Eric W. Biederman
2006-02-06 20:36 ` Kirill Korotaev
2006-02-06 20:40 ` Eric W. Biederman
2006-02-02 14:49 ` Kirill Korotaev
2006-01-17 14:33 ` RFC [patch 14/34] PID Virtualization const parameter for process group Serge Hallyn
2006-01-17 14:33 ` RFC [patch 15/34] PID Virtualization task virtual pid access functions Serge Hallyn
2006-01-17 14:33 ` RFC [patch 16/34] PID Virtualization return virtual pids where required Serge Hallyn
2006-01-17 14:33 ` RFC [patch 17/34] PID Virtualization return virtual process group ids Serge Hallyn
2006-01-17 14:33 ` RFC [patch 18/34] PID Virtualization code enhancements for virtual pids in /proc Serge Hallyn
2006-01-17 14:33 ` RFC [patch 19/34] PID Virtualization Define pid_to_vpid functions Serge Hallyn
2006-01-17 14:33 ` RFC [patch 20/34] PID Virtualization Use pid_to_vpid conversion functions Serge Hallyn
2006-01-17 14:33 ` RFC [patch 21/34] PID Virtualization file owner pid virtualization Serge Hallyn
2006-01-17 14:33 ` RFC [patch 22/34] PID Virtualization define vpid_to_pid functions Serge Hallyn
2006-01-17 14:33 ` RFC [patch 23/34] PID Virtualization Use " Serge Hallyn
2006-01-17 14:33 ` RFC [patch 24/34] PID Virtualization use vpgid_to_pgid function Serge Hallyn
2006-01-17 14:33 ` RFC [patch 25/34] PID Virtualization Context for pid_to_vpid conversition functions Serge Hallyn
2006-01-17 14:33 ` RFC [patch 26/34] PID Virtualization Documentation Serge Hallyn
2006-01-17 14:33 ` RFC [patch 27/34] PID Virtualization pidspace Serge Hallyn
2006-01-17 14:33 ` RFC [patch 28/34] PID Virtualization container object and functions Serge Hallyn
2006-01-17 14:33 ` RFC [patch 29/34] PID Virtualization container attach/detach calls Serge Hallyn
2006-01-17 14:33 ` RFC [patch 30/34] PID Virtualization /proc/container filesystem Serge Hallyn
2006-01-17 14:33 ` RFC [patch 31/34] PID Virtualization Implementation of low level virtualization functions Serge Hallyn
2006-01-17 14:33 ` RFC [patch 32/34] PID Virtualization Handle special case vpid return cases Serge Hallyn
2006-01-17 14:33 ` RFC [patch 33/34] PID Virtualization per container /proc filesystem Serge Hallyn
2006-01-17 14:33 ` RFC [patch 34/34] PID Virtualization pidspace parent : signal behavior Serge Hallyn
2006-01-17 16:19 ` RFC [patch 00/34] PID Virtualization Overview Suleiman Souhlal
2006-01-17 17:08 ` Dave Hansen
2006-01-17 18:09 ` Suleiman Souhlal
2006-01-17 18:12 ` Dave Hansen
2006-01-17 18:29 ` Alan Cox
2006-01-18 19:01 ` Dave Hansen
2006-01-18 19:28 ` Arjan van de Ven
2006-01-18 19:38 ` Dave Hansen
2006-01-18 19:50 ` Arjan van de Ven
2006-01-18 22:54 ` Alan Cox
2006-01-19 7:15 ` Arjan van de Ven
2006-01-20 5:11 ` Eric W. Biederman
2006-01-20 20:23 ` Serge E. Hallyn
2006-01-20 20:33 ` Hubertus Franke
2006-01-21 10:34 ` Eric W. Biederman
2006-01-20 19:53 ` RFC: Multiple instances of kernel namespaces Eric W. Biederman
2006-01-20 20:13 ` Serge E. Hallyn
2006-01-20 20:22 ` Hubertus Franke
[not found] ` <20060120203555.GC13265@sergelap.austin.ibm.com>
2006-01-20 21:47 ` Hubertus Franke
2006-01-21 10:04 ` Eric W. Biederman [this message]
2006-01-26 19:47 ` Herbert Poetzl
2006-01-26 20:13 ` Eric W. Biederman
2006-01-26 20:27 ` Herbert Poetzl
2006-01-21 10:31 ` RFC [patch 00/34] PID Virtualization Overview Pavel Machek
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=m13bjhoq1r.fsf@ebiederm.dsl.xmission.com \
--to=ebiederm@xmission.com \
--cc=arjan@infradead.org \
--cc=clg@fr.ibm.com \
--cc=frankeh@watson.ibm.com \
--cc=haveblue@us.ibm.com \
--cc=linux-kernel@vger.kernel.org \
--cc=serue@us.ibm.com \
--cc=ssouhlal@FreeBSD.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).