Message-ID: <4ADF56D4.8030405@librato.com>
Date: Wed, 21 Oct 2009 14:45:40 -0400
From: Oren Laadan <orenl@librato.com>
Organization: Librato
User-Agent: Thunderbird 2.0.0.23 (X11/20090817)
MIME-Version: 1.0
To: Daniel Lezcano <daniel.lezcano@free.fr>
CC: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>, randy.dunlap@oracle.com,
       arnd@arndb.de, linux-api@vger.kernel.org,
       Containers <containers@lists.linux-foundation.org>,
       Nathan Lynch <nathanl@austin.ibm.com>, linux-kernel@vger.kernel.org,
       Louis.Rilling@kerlabs.com, "Eric W. Biederman" <ebiederm@xmission.com>,
       kosaki.motohiro@jp.fujitsu.com, hpa@zytor.com, mingo@elte.hu,
       torvalds@linux-foundation.org, Alexey Dobriyan <adobriyan@gmail.com>,
       roland@redhat.com, Pavel Emelyanov <xemul@openvz.org>
Subject: Re: [RFC][v8][PATCH 0/10] Implement clone3() system call
References: <20091013044925.GA28181@us.ibm.com> <4AD8C7E4.9000903@free.fr>	<20091016194451.GA28706@us.ibm.com> <4ADCCD68.9030003@free.fr> <4ADCDE7F.4090501@librato.com> <4ADF2E75.1020801@free.fr>
In-Reply-To: <4ADF2E75.1020801@free.fr>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 5350
Lines: 153


Daniel Lezcano wrote:
> Oren Laadan wrote:
>>
>> Daniel Lezcano wrote:
> [ ... ]
> 
>>> I forgot to mention a constraint with the specified pid : P2 has to
>>> be child of P1.
>>> In other word, you can not specify a pid to clonat which is not your
>>> descendant (including yourself).
>>> With this constraint I think there is no security issues.
>>
>> Sounds dangerous. What if your descendant executed a setuid program ?
> 
> That does not happen because you inherit the context of the caller.
> 
>>> Concerning of forking on behalf of another process, we can consider
>>> it is up to the caller / programmer to know what it does. If a
>>> process in 
>>
>> Before the user can program with this syscall, _you_ need to define
>> the semantics of this syscall. 
> Yes, you are right. Here it is the proposition of the semantics.
> 
> Function prototype is:
> 
> pid_t cloneat(pid_t pid, pid_t hint, struct clone_args *args);
> 
> Structure types are:
> 
> typedef int clone_flag_t;
> 
> struct clone_args {
>     clone_flag_t *flags;
>     int flags_size;
>     u32 reserved1;
>     u32 reserved2;
>     u64 child_stack_base;
>     u64 child_stack_size;
>     u64 parent_tid_ptr;
>     u64 child_tid_ptr;
>     u64 reserved3;
> };
> 
> With the helper macros:
> 
> void CLONE_SET(int flag, clone_flag_t *flags);
> void CLONE_CLR(int flag, clone_flag_t *flags);
> bool CLONE_ISSET(int flag, clone_flag_t *flags);
> void CLONE_ZERO(flag_t *clone_flags);
> 
> And:
> 
> #define CLONEXT_VM      0x20  /* CLONE_VM>>3 */ #define CLONEXT_FS     
> 0x21
> #define CLONEXT_FILES   0x22
> ...
> 

The main motivation for your new syscall is to make it possible to
inject a process into a namespace. IOW, what you are proposing is
a new incarnation of sys_hijack().

This is _orthogonal_ to the current discussion, which is about an
extension for clone to allow (a) choosing target pid(s), (b) more
flags, and (c) future extensions.

(Your suggested syscall may, too, allow the request a specific set
of pids for the child process, and reuse the current code for that).

I suggest that you start a new thread about your RFC. This will
reduce distractions on the current thread, and bring more focus to
your proposal. I surely will post some comments there :)

[...]

> The cloneat syscall can be used for the following use cases:
> 
> * checkpoint / restart:
> 
> The restart can be done with a clone(.., CLONE_NEWPID|...);
> Then the new pid (aka pid 1) retrieves the proctree from the statefile
> and creates the different tasks with the process hierarchy with the
> cloneat syscall.

s/cloneat/$CLONE3/
(hint: this is how it's done now)

> 
> The proctree creation can be done from outside of the pid namespace or
> from inside.

Ew .. why would you do that ?

> Concerning nested pid namespaces, IMHO I would not try to checkpoint /
> restart them. The checkpoint of a nested pid namespace should be
> forbidden except for the leaf of a pid namespaces tree. That should

Others (me included) *will* try and may get upset if forbidden...
Seriously, there is no technical reason to restrict this.

 >> Can you define more precisely what you mean by "enter" the container ?
>>
>> If you simply want create a new process in the container, you can
>> achieve the same thing with a daemon, or a smart init process (in
>> there), or even ptrace tricks.
> 
> Yes, you can launch a daemon inside the container, that works for a
> system container because the container is killed by killing the first
> process of the container or by a shutdown inside the container (not
> fully implemented in the kernel).
> But this is unreliable for application containers, I won't enter in the
> details but the container exits when the application exits, with a
> daemon inside the container, this is no longer the case because you can
> not detect the application death as the daemon is always there.
> 
> With cloneat you restrict the life cycle of the command you launched,
> that is the container exits as soon as all the processes exited the
> container, including the spawned command itself.

Then start a daemon _in addition_ to the application, or write a
daemon that will launch the application and monitor it... And also
there is ptrace -

But, please let's take this off to a new thread about adding how to
add a process into a namespace from the outside. FYI, I do think
such an interface may be useful and nicer than the two alternatives
I suggested above.

>> Also, there is a reason why sys_hijack() was hijacked away ... And
>> I honestly think that a syscall to force another process to clone
>> would be shot down by the kernel guys.
> Maybe, maybe not. CLONE_PARENT exists and looks similar to cloneat.

Actually, I misread previously; I mean not forcing another process
to clone, but instead forcing another process to become a parent (and
I shall ignore the ethical issues :)

I still suspect it won't be welcome. Several people would have liked
to see CLONE_PARENT go away, too, if that was possible without breaking
userspace applications. Yet another reason to take it to a discussion
of its own.

Oren.


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/