Re: [RFC PATCH 3/8] kmod - teach call_usermodehelper() to use a namespace

[Oleg Nesterov <oleg@redhat.com>]

On 02/10, Ian Kent wrote:
>
> On Mon, 2015-02-09 at 17:03 +0100, Oleg Nesterov wrote:
> >
> > I understand. but I still can't understand why we can't implement something
> > like
> > 	enter_ns(struct nsproxy *p)
> > 	{
> > 		new_nsproxy = create_new_namespaces(...);
> >
> > 		p->mnt_ns->ns->ops->install(new_nsproxy, ...);
> > 		p->pid_ns_for_children->ns->ops->install(new_nsproxy, ...);
> > 		...
> >
> > 		switch_task_namespaces(new_nsproxy);
> > 	}
> >
> > Why we should abuse fs/proc ?
>
> That sounds like a much better approach.
> Your saying just take a reference to the nsproxy from the located
> process and use it instead, right?

Yes,

> Working out if there's a difference with what you from the open is
> challenging (I already tried), I'll have another go at it.

I thinks there should not be any difference, but please re-check ;)

> > And. Whatever we do, ops->install() or setns_inode() can't solve the problem with
> > pid_ns. You need the additional clone() to "activate" it. pidns_install() does not
> > actually change task_active_pid_ns().
>
> Right, but all this is done in preparation for the following do_execve()
> call. Isn't that enough or am I missing something?

Yes, but do_execve() doesn't (and shouldn't) change task_active_pid_ns(). Note
the ->pid_ns_for_children's name. It is only used by copy_process()->alloc_pid().

task_active_pid_ns() uses task_pid() and we obviously can't change it.

I am wondering if we can do something like

	kernel_thread_in_ns(struct nsproxy *ns, ...)
	{
		struct nsproxy *saved_ns = current->nsproxy;
		pid_t pid;

		task_lock(current);
		current->nsproxy = ns;
		task_unlock(current);

		pid = kernel_thread(...);

		task_lock(current);
		current->nsproxy = saved_ns;
		task_unlock(current);

		return pid;
	}

used by __call_usermodehelper/wait_for_helper, instead of "enter_ns" from
sub_info->init()...

Oleg.