[mirror_ubuntu-bionic-kernel.git] / kernel / pid_namespace.c

/*
 * Pid namespaces
 *
 * Authors:
 *    (C) 2007 Pavel Emelyanov <xemul@openvz.org>, OpenVZ, SWsoft Inc.
 *    (C) 2007 Sukadev Bhattiprolu <sukadev@us.ibm.com>, IBM
 *     Many thanks to Oleg Nesterov for comments and help
 *
 */

#include <linux/pid.h>
#include <linux/pid_namespace.h>
#include <linux/user_namespace.h>
#include <linux/syscalls.h>
#include <linux/cred.h>
#include <linux/err.h>
#include <linux/acct.h>
#include <linux/slab.h>
#include <linux/proc_ns.h>
#include <linux/reboot.h>
#include <linux/export.h>
#include <linux/sched/task.h>
#include <linux/sched/signal.h>
#include <linux/idr.h>

struct pid_cache {
	int nr_ids;
	char name[16];
	struct kmem_cache *cachep;
	struct list_head list;
};

static LIST_HEAD(pid_caches_lh);
static DEFINE_MUTEX(pid_caches_mutex);
static struct kmem_cache *pid_ns_cachep;

/*
 * creates the kmem cache to allocate pids from.
 * @nr_ids: the number of numerical ids this pid will have to carry
 */

static struct kmem_cache *create_pid_cachep(int nr_ids)
{
	struct pid_cache *pcache;
	struct kmem_cache *cachep;

	mutex_lock(&pid_caches_mutex);
	list_for_each_entry(pcache, &pid_caches_lh, list)
		if (pcache->nr_ids == nr_ids)
			goto out;

	pcache = kmalloc(sizeof(struct pid_cache), GFP_KERNEL);
	if (pcache == NULL)
		goto err_alloc;

	snprintf(pcache->name, sizeof(pcache->name), "pid_%d", nr_ids);
	cachep = kmem_cache_create(pcache->name,
			sizeof(struct pid) + (nr_ids - 1) * sizeof(struct upid),
			0, SLAB_HWCACHE_ALIGN, NULL);
	if (cachep == NULL)
		goto err_cachep;

	pcache->nr_ids = nr_ids;
	pcache->cachep = cachep;
	list_add(&pcache->list, &pid_caches_lh);
out:
	mutex_unlock(&pid_caches_mutex);
	return pcache->cachep;

err_cachep:
	kfree(pcache);
err_alloc:
	mutex_unlock(&pid_caches_mutex);
	return NULL;
}

static void proc_cleanup_work(struct work_struct *work)
{
	struct pid_namespace *ns = container_of(work, struct pid_namespace, proc_work);
	pid_ns_release_proc(ns);
}

/* MAX_PID_NS_LEVEL is needed for limiting size of 'struct pid' */
#define MAX_PID_NS_LEVEL 32

static struct ucounts *inc_pid_namespaces(struct user_namespace *ns)
{
	return inc_ucount(ns, current_euid(), UCOUNT_PID_NAMESPACES);
}

static void dec_pid_namespaces(struct ucounts *ucounts)
{
	dec_ucount(ucounts, UCOUNT_PID_NAMESPACES);
}

static struct pid_namespace *create_pid_namespace(struct user_namespace *user_ns,
	struct pid_namespace *parent_pid_ns)
{
	struct pid_namespace *ns;
	unsigned int level = parent_pid_ns->level + 1;
	struct ucounts *ucounts;
	int err;

	err = -EINVAL;
	if (!in_userns(parent_pid_ns->user_ns, user_ns))
		goto out;

	err = -ENOSPC;
	if (level > MAX_PID_NS_LEVEL)
		goto out;
	ucounts = inc_pid_namespaces(user_ns);
	if (!ucounts)
		goto out;

	err = -ENOMEM;
	ns = kmem_cache_zalloc(pid_ns_cachep, GFP_KERNEL);
	if (ns == NULL)
		goto out_dec;

	idr_init(&ns->idr);

	ns->pid_cachep = create_pid_cachep(level + 1);
	if (ns->pid_cachep == NULL)
		goto out_free_idr;

	err = ns_alloc_inum(&ns->ns);
	if (err)
		goto out_free_idr;
	ns->ns.ops = &pidns_operations;

	kref_init(&ns->kref);
	ns->level = level;
	ns->parent = get_pid_ns(parent_pid_ns);
	ns->user_ns = get_user_ns(user_ns);
	ns->ucounts = ucounts;
	ns->pid_allocated = PIDNS_ADDING;
	INIT_WORK(&ns->proc_work, proc_cleanup_work);

	return ns;

out_free_idr:
	idr_destroy(&ns->idr);
	kmem_cache_free(pid_ns_cachep, ns);
out_dec:
	dec_pid_namespaces(ucounts);
out:
	return ERR_PTR(err);
}

static void delayed_free_pidns(struct rcu_head *p)
{
	struct pid_namespace *ns = container_of(p, struct pid_namespace, rcu);

	dec_pid_namespaces(ns->ucounts);
	put_user_ns(ns->user_ns);

	kmem_cache_free(pid_ns_cachep, ns);
}

static void destroy_pid_namespace(struct pid_namespace *ns)
{
	ns_free_inum(&ns->ns);

	idr_destroy(&ns->idr);
	call_rcu(&ns->rcu, delayed_free_pidns);
}

struct pid_namespace *copy_pid_ns(unsigned long flags,
	struct user_namespace *user_ns, struct pid_namespace *old_ns)
{
	if (!(flags & CLONE_NEWPID))
		return get_pid_ns(old_ns);
	if (task_active_pid_ns(current) != old_ns)
		return ERR_PTR(-EINVAL);
	return create_pid_namespace(user_ns, old_ns);
}

static void free_pid_ns(struct kref *kref)
{
	struct pid_namespace *ns;

	ns = container_of(kref, struct pid_namespace, kref);
	destroy_pid_namespace(ns);
}

void put_pid_ns(struct pid_namespace *ns)
{
	struct pid_namespace *parent;

	while (ns != &init_pid_ns) {
		parent = ns->parent;
		if (!kref_put(&ns->kref, free_pid_ns))
			break;
		ns = parent;
	}
}
EXPORT_SYMBOL_GPL(put_pid_ns);

void zap_pid_ns_processes(struct pid_namespace *pid_ns)
{
	int nr;
	int rc;
	struct task_struct *task, *me = current;
	int init_pids = thread_group_leader(me) ? 1 : 2;
	struct pid *pid;

	/* Don't allow any more processes into the pid namespace */
	disable_pid_allocation(pid_ns);

	/*
	 * Ignore SIGCHLD causing any terminated children to autoreap.
	 * This speeds up the namespace shutdown, plus see the comment
	 * below.
	 */
	spin_lock_irq(&me->sighand->siglock);
	me->sighand->action[SIGCHLD - 1].sa.sa_handler = SIG_IGN;
	spin_unlock_irq(&me->sighand->siglock);

	/*
	 * The last thread in the cgroup-init thread group is terminating.
	 * Find remaining pid_ts in the namespace, signal and wait for them
	 * to exit.
	 *
	 * Note:  This signals each threads in the namespace - even those that
	 * 	  belong to the same thread group, To avoid this, we would have
	 * 	  to walk the entire tasklist looking a processes in this
	 * 	  namespace, but that could be unnecessarily expensive if the
	 * 	  pid namespace has just a few processes. Or we need to
	 * 	  maintain a tasklist for each pid namespace.
	 *
	 */
	rcu_read_lock();
	read_lock(&tasklist_lock);
	nr = 2;
	idr_for_each_entry_continue(&pid_ns->idr, pid, nr) {
		task = pid_task(pid, PIDTYPE_PID);
		if (task && !__fatal_signal_pending(task))
			send_sig_info(SIGKILL, SEND_SIG_FORCED, task);
	}
	read_unlock(&tasklist_lock);
	rcu_read_unlock();

	/*
	 * Reap the EXIT_ZOMBIE children we had before we ignored SIGCHLD.
	 * sys_wait4() will also block until our children traced from the
	 * parent namespace are detached and become EXIT_DEAD.
	 */
	do {
		clear_thread_flag(TIF_SIGPENDING);
		rc = sys_wait4(-1, NULL, __WALL, NULL);
	} while (rc != -ECHILD);

	/*
	 * sys_wait4() above can't reap the EXIT_DEAD children but we do not
	 * really care, we could reparent them to the global init. We could
	 * exit and reap ->child_reaper even if it is not the last thread in
	 * this pid_ns, free_pid(pid_allocated == 0) calls proc_cleanup_work(),
	 * pid_ns can not go away until proc_kill_sb() drops the reference.
	 *
	 * But this ns can also have other tasks injected by setns()+fork().
	 * Again, ignoring the user visible semantics we do not really need
	 * to wait until they are all reaped, but they can be reparented to
	 * us and thus we need to ensure that pid->child_reaper stays valid
	 * until they all go away. See free_pid()->wake_up_process().
	 *
	 * We rely on ignored SIGCHLD, an injected zombie must be autoreaped
	 * if reparented.
	 */
	for (;;) {
		set_current_state(TASK_INTERRUPTIBLE);
		if (pid_ns->pid_allocated == init_pids)
			break;
		schedule();
	}
	__set_current_state(TASK_RUNNING);

	if (pid_ns->reboot)
		current->signal->group_exit_code = pid_ns->reboot;

	acct_exit_ns(pid_ns);
	return;
}

#ifdef CONFIG_CHECKPOINT_RESTORE
static int pid_ns_ctl_handler(struct ctl_table *table, int write,
		void __user *buffer, size_t *lenp, loff_t *ppos)
{
	struct pid_namespace *pid_ns = task_active_pid_ns(current);
	struct ctl_table tmp = *table;
	int ret, next;

	if (write && !ns_capable(pid_ns->user_ns, CAP_SYS_ADMIN))
		return -EPERM;

	/*
	 * Writing directly to ns' last_pid field is OK, since this field
	 * is volatile in a living namespace anyway and a code writing to
	 * it should synchronize its usage with external means.
	 */

	next = idr_get_cursor(&pid_ns->idr) - 1;

	tmp.data = &next;
	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
	if (!ret && write)
		idr_set_cursor(&pid_ns->idr, next + 1);

	return ret;
}

extern int pid_max;
static int zero = 0;
static struct ctl_table pid_ns_ctl_table[] = {
	{
		.procname = "ns_last_pid",
		.maxlen = sizeof(int),
		.mode = 0666, /* permissions are checked in the handler */
		.proc_handler = pid_ns_ctl_handler,
		.extra1 = &zero,
		.extra2 = &pid_max,
	},
	{ }
};
static struct ctl_path kern_path[] = { { .procname = "kernel", }, { } };
#endif	/* CONFIG_CHECKPOINT_RESTORE */

int reboot_pid_ns(struct pid_namespace *pid_ns, int cmd)
{
	if (pid_ns == &init_pid_ns)
		return 0;

	switch (cmd) {
	case LINUX_REBOOT_CMD_RESTART2:
	case LINUX_REBOOT_CMD_RESTART:
		pid_ns->reboot = SIGHUP;
		break;

	case LINUX_REBOOT_CMD_POWER_OFF:
	case LINUX_REBOOT_CMD_HALT:
		pid_ns->reboot = SIGINT;
		break;
	default:
		return -EINVAL;
	}

	read_lock(&tasklist_lock);
	force_sig(SIGKILL, pid_ns->child_reaper);
	read_unlock(&tasklist_lock);

	do_exit(0);

	/* Not reached */
	return 0;
}

static inline struct pid_namespace *to_pid_ns(struct ns_common *ns)
{
	return container_of(ns, struct pid_namespace, ns);
}

static struct ns_common *pidns_get(struct task_struct *task)
{
	struct pid_namespace *ns;

	rcu_read_lock();
	ns = task_active_pid_ns(task);
	if (ns)
		get_pid_ns(ns);
	rcu_read_unlock();

	return ns ? &ns->ns : NULL;
}

static struct ns_common *pidns_for_children_get(struct task_struct *task)
{
	struct pid_namespace *ns = NULL;

	task_lock(task);
	if (task->nsproxy) {
		ns = task->nsproxy->pid_ns_for_children;
		get_pid_ns(ns);
	}
	task_unlock(task);

	if (ns) {
		read_lock(&tasklist_lock);
		if (!ns->child_reaper) {
			put_pid_ns(ns);
			ns = NULL;
		}
		read_unlock(&tasklist_lock);
	}

	return ns ? &ns->ns : NULL;
}

static void pidns_put(struct ns_common *ns)
{
	put_pid_ns(to_pid_ns(ns));
}

static int pidns_install(struct nsproxy *nsproxy, struct ns_common *ns)
{
	struct pid_namespace *active = task_active_pid_ns(current);
	struct pid_namespace *ancestor, *new = to_pid_ns(ns);

	if (!ns_capable(new->user_ns, CAP_SYS_ADMIN) ||
	    !ns_capable(current_user_ns(), CAP_SYS_ADMIN))
		return -EPERM;

	/*
	 * Only allow entering the current active pid namespace
	 * or a child of the current active pid namespace.
	 *
	 * This is required for fork to return a usable pid value and
	 * this maintains the property that processes and their
	 * children can not escape their current pid namespace.
	 */
	if (new->level < active->level)
		return -EINVAL;

	ancestor = new;
	while (ancestor->level > active->level)
		ancestor = ancestor->parent;
	if (ancestor != active)
		return -EINVAL;

	put_pid_ns(nsproxy->pid_ns_for_children);
	nsproxy->pid_ns_for_children = get_pid_ns(new);
	return 0;
}

static struct ns_common *pidns_get_parent(struct ns_common *ns)
{
	struct pid_namespace *active = task_active_pid_ns(current);
	struct pid_namespace *pid_ns, *p;

	/* See if the parent is in the current namespace */
	pid_ns = p = to_pid_ns(ns)->parent;
	for (;;) {
		if (!p)
			return ERR_PTR(-EPERM);
		if (p == active)
			break;
		p = p->parent;
	}

	return &get_pid_ns(pid_ns)->ns;
}

static struct user_namespace *pidns_owner(struct ns_common *ns)
{
	return to_pid_ns(ns)->user_ns;
}

const struct proc_ns_operations pidns_operations = {
	.name		= "pid",
	.type		= CLONE_NEWPID,
	.get		= pidns_get,
	.put		= pidns_put,
	.install	= pidns_install,
	.owner		= pidns_owner,
	.get_parent	= pidns_get_parent,
};

const struct proc_ns_operations pidns_for_children_operations = {
	.name		= "pid_for_children",
	.real_ns_name	= "pid",
	.type		= CLONE_NEWPID,
	.get		= pidns_for_children_get,
	.put		= pidns_put,
	.install	= pidns_install,
	.owner		= pidns_owner,
	.get_parent	= pidns_get_parent,
};

static __init int pid_namespaces_init(void)
{
	pid_ns_cachep = KMEM_CACHE(pid_namespace, SLAB_PANIC);

#ifdef CONFIG_CHECKPOINT_RESTORE
	register_sysctl_paths(kern_path, pid_ns_ctl_table);
#endif
	return 0;
}

__initcall(pid_namespaces_init);
Commit	Line	Data
74bd59bb PE	1	/*
	2	* Pid namespaces
	3	*
	4	* Authors:
	5	* (C) 2007 Pavel Emelyanov <xemul@openvz.org>, OpenVZ, SWsoft Inc.
	6	* (C) 2007 Sukadev Bhattiprolu <sukadev@us.ibm.com>, IBM
	7	* Many thanks to Oleg Nesterov for comments and help
	8	*
	9	*/
	10
	11	#include <linux/pid.h>
	12	#include <linux/pid_namespace.h>
49f4d8b9	13	#include <linux/user_namespace.h>
74bd59bb	14	#include <linux/syscalls.h>
5b825c3a	15	#include <linux/cred.h>
74bd59bb	16	#include <linux/err.h>
0b6b030f	17	#include <linux/acct.h>
5a0e3ad6	18	#include <linux/slab.h>
0bb80f24	19	#include <linux/proc_ns.h>
cf3f8921	20	#include <linux/reboot.h>
523a6a94	21	#include <linux/export.h>
29930025	22	#include <linux/sched/task.h>
f361bf4a	23	#include <linux/sched/signal.h>
95846ecf	24	#include <linux/idr.h>
74bd59bb	25
74bd59bb PE	26	struct pid_cache {
	27	int nr_ids;
	28	char name[16];
	29	struct kmem_cache *cachep;
	30	struct list_head list;
	31	};
	32
	33	static LIST_HEAD(pid_caches_lh);
	34	static DEFINE_MUTEX(pid_caches_mutex);
	35	static struct kmem_cache *pid_ns_cachep;
	36
	37	/*
	38	* creates the kmem cache to allocate pids from.
	39	* @nr_ids: the number of numerical ids this pid will have to carry
	40	*/
	41
	42	static struct kmem_cache *create_pid_cachep(int nr_ids)
	43	{
	44	struct pid_cache *pcache;
	45	struct kmem_cache *cachep;
	46
	47	mutex_lock(&pid_caches_mutex);
	48	list_for_each_entry(pcache, &pid_caches_lh, list)
	49	if (pcache->nr_ids == nr_ids)
	50	goto out;
	51
	52	pcache = kmalloc(sizeof(struct pid_cache), GFP_KERNEL);
	53	if (pcache == NULL)
	54	goto err_alloc;
	55
	56	snprintf(pcache->name, sizeof(pcache->name), "pid_%d", nr_ids);
	57	cachep = kmem_cache_create(pcache->name,
	58	sizeof(struct pid) + (nr_ids - 1) * sizeof(struct upid),
	59	0, SLAB_HWCACHE_ALIGN, NULL);
	60	if (cachep == NULL)
	61	goto err_cachep;
	62
	63	pcache->nr_ids = nr_ids;
	64	pcache->cachep = cachep;
	65	list_add(&pcache->list, &pid_caches_lh);
	66	out:
	67	mutex_unlock(&pid_caches_mutex);
	68	return pcache->cachep;
	69
	70	err_cachep:
	71	kfree(pcache);
	72	err_alloc:
	73	mutex_unlock(&pid_caches_mutex);
	74	return NULL;
	75	}
	76
0a01f2cc EB	77	static void proc_cleanup_work(struct work_struct *work)
	78	{
	79	struct pid_namespace *ns = container_of(work, struct pid_namespace, proc_work);
	80	pid_ns_release_proc(ns);
	81	}
	82
f2302505 AV	83	/* MAX_PID_NS_LEVEL is needed for limiting size of 'struct pid' */
	84	#define MAX_PID_NS_LEVEL 32
	85
f333c700 EB	86	static struct ucounts inc_pid_namespaces(struct user_namespace ns)
	87	{
	88	return inc_ucount(ns, current_euid(), UCOUNT_PID_NAMESPACES);
	89	}
	90
	91	static void dec_pid_namespaces(struct ucounts *ucounts)
	92	{
	93	dec_ucount(ucounts, UCOUNT_PID_NAMESPACES);
	94	}
	95
49f4d8b9 EB	96	static struct pid_namespace create_pid_namespace(struct user_namespace user_ns,
49f4d8b9 EB	97	struct pid_namespace *parent_pid_ns)
74bd59bb PE	98	{
74bd59bb PE	99	struct pid_namespace *ns;
ed469a63	100	unsigned int level = parent_pid_ns->level + 1;
f333c700	101	struct ucounts *ucounts;
f2302505 AV	102	int err;
f2302505 AV	103
a2b42626 EB	104	err = -EINVAL;
	105	if (!in_userns(parent_pid_ns->user_ns, user_ns))
	106	goto out;
	107
df75e774	108	err = -ENOSPC;
f333c700 EB	109	if (level > MAX_PID_NS_LEVEL)
	110	goto out;
	111	ucounts = inc_pid_namespaces(user_ns);
	112	if (!ucounts)
f2302505	113	goto out;
74bd59bb	114
f2302505	115	err = -ENOMEM;
84406c15	116	ns = kmem_cache_zalloc(pid_ns_cachep, GFP_KERNEL);
74bd59bb	117	if (ns == NULL)
f333c700	118	goto out_dec;
74bd59bb	119
95846ecf	120	idr_init(&ns->idr);
74bd59bb PE	121
	122	ns->pid_cachep = create_pid_cachep(level + 1);
	123	if (ns->pid_cachep == NULL)
95846ecf	124	goto out_free_idr;
74bd59bb	125
6344c433	126	err = ns_alloc_inum(&ns->ns);
98f842e6	127	if (err)
95846ecf	128	goto out_free_idr;
33c42940	129	ns->ns.ops = &pidns_operations;
98f842e6	130
74bd59bb	131	kref_init(&ns->kref);
74bd59bb	132	ns->level = level;
ed469a63	133	ns->parent = get_pid_ns(parent_pid_ns);
49f4d8b9	134	ns->user_ns = get_user_ns(user_ns);
f333c700	135	ns->ucounts = ucounts;
e8cfbc24	136	ns->pid_allocated = PIDNS_ADDING;
0a01f2cc	137	INIT_WORK(&ns->proc_work, proc_cleanup_work);
74bd59bb	138
74bd59bb PE	139	return ns;
74bd59bb PE	140
95846ecf GS	141	out_free_idr:
95846ecf GS	142	idr_destroy(&ns->idr);
74bd59bb	143	kmem_cache_free(pid_ns_cachep, ns);
f333c700 EB	144	out_dec:
f333c700 EB	145	dec_pid_namespaces(ucounts);
74bd59bb	146	out:
4308eebb	147	return ERR_PTR(err);
74bd59bb PE	148	}
74bd59bb PE	149
1adfcb03 AV	150	static void delayed_free_pidns(struct rcu_head *p)
1adfcb03 AV	151	{
add7c65c AV	152	struct pid_namespace *ns = container_of(p, struct pid_namespace, rcu);
	153
	154	dec_pid_namespaces(ns->ucounts);
	155	put_user_ns(ns->user_ns);
	156
	157	kmem_cache_free(pid_ns_cachep, ns);
1adfcb03 AV	158	}
1adfcb03 AV	159
74bd59bb PE	160	static void destroy_pid_namespace(struct pid_namespace *ns)
74bd59bb PE	161	{
6344c433	162	ns_free_inum(&ns->ns);
95846ecf GS	163
95846ecf GS	164	idr_destroy(&ns->idr);
1adfcb03	165	call_rcu(&ns->rcu, delayed_free_pidns);
74bd59bb PE	166	}
74bd59bb PE	167
49f4d8b9 EB	168	struct pid_namespace *copy_pid_ns(unsigned long flags,
49f4d8b9 EB	169	struct user_namespace user_ns, struct pid_namespace old_ns)
74bd59bb	170	{
74bd59bb	171	if (!(flags & CLONE_NEWPID))
dca4a979	172	return get_pid_ns(old_ns);
225778d6 EB	173	if (task_active_pid_ns(current) != old_ns)
225778d6 EB	174	return ERR_PTR(-EINVAL);
49f4d8b9	175	return create_pid_namespace(user_ns, old_ns);
74bd59bb PE	176	}
74bd59bb PE	177
bbc2e3ef	178	static void free_pid_ns(struct kref *kref)
74bd59bb	179	{
bbc2e3ef	180	struct pid_namespace *ns;
74bd59bb PE	181
74bd59bb PE	182	ns = container_of(kref, struct pid_namespace, kref);
74bd59bb	183	destroy_pid_namespace(ns);
bbc2e3ef	184	}
74bd59bb	185
bbc2e3ef CG	186	void put_pid_ns(struct pid_namespace *ns)
	187	{
	188	struct pid_namespace *parent;
	189
	190	while (ns != &init_pid_ns) {
	191	parent = ns->parent;
	192	if (!kref_put(&ns->kref, free_pid_ns))
	193	break;
	194	ns = parent;
	195	}
74bd59bb	196	}
bbc2e3ef	197	EXPORT_SYMBOL_GPL(put_pid_ns);
74bd59bb PE	198
	199	void zap_pid_ns_processes(struct pid_namespace *pid_ns)
	200	{
	201	int nr;
	202	int rc;
00c10bc1	203	struct task_struct task, me = current;
751c644b	204	int init_pids = thread_group_leader(me) ? 1 : 2;
95846ecf	205	struct pid *pid;
00c10bc1	206
c876ad76 EB	207	/* Don't allow any more processes into the pid namespace */
	208	disable_pid_allocation(pid_ns);
	209
a53b8315 ON	210	/*
	211	* Ignore SIGCHLD causing any terminated children to autoreap.
	212	* This speeds up the namespace shutdown, plus see the comment
	213	* below.
	214	*/
00c10bc1 EB	215	spin_lock_irq(&me->sighand->siglock);
	216	me->sighand->action[SIGCHLD - 1].sa.sa_handler = SIG_IGN;
	217	spin_unlock_irq(&me->sighand->siglock);
74bd59bb PE	218
	219	/*
	220	* The last thread in the cgroup-init thread group is terminating.
	221	* Find remaining pid_ts in the namespace, signal and wait for them
	222	* to exit.
	223	*
	224	* Note: This signals each threads in the namespace - even those that
	225	* belong to the same thread group, To avoid this, we would have
	226	* to walk the entire tasklist looking a processes in this
	227	* namespace, but that could be unnecessarily expensive if the
	228	* pid namespace has just a few processes. Or we need to
	229	* maintain a tasklist for each pid namespace.
	230	*
	231	*/
95846ecf	232	rcu_read_lock();
74bd59bb	233	read_lock(&tasklist_lock);
95846ecf GS	234	nr = 2;
	235	idr_for_each_entry_continue(&pid_ns->idr, pid, nr) {
	236	task = pid_task(pid, PIDTYPE_PID);
a02d6fd6 ON	237	if (task && !__fatal_signal_pending(task))
a02d6fd6 ON	238	send_sig_info(SIGKILL, SEND_SIG_FORCED, task);
74bd59bb PE	239	}
74bd59bb PE	240	read_unlock(&tasklist_lock);
95846ecf	241	rcu_read_unlock();
74bd59bb	242
a53b8315 ON	243	/*
	244	* Reap the EXIT_ZOMBIE children we had before we ignored SIGCHLD.
	245	* sys_wait4() will also block until our children traced from the
	246	* parent namespace are detached and become EXIT_DEAD.
	247	*/
74bd59bb PE	248	do {
	249	clear_thread_flag(TIF_SIGPENDING);
	250	rc = sys_wait4(-1, NULL, __WALL, NULL);
	251	} while (rc != -ECHILD);
	252
6347e900	253	/*
a53b8315 ON	254	* sys_wait4() above can't reap the EXIT_DEAD children but we do not
	255	* really care, we could reparent them to the global init. We could
	256	* exit and reap ->child_reaper even if it is not the last thread in
e8cfbc24	257	* this pid_ns, free_pid(pid_allocated == 0) calls proc_cleanup_work(),
a53b8315 ON	258	* pid_ns can not go away until proc_kill_sb() drops the reference.
	259	*
	260	* But this ns can also have other tasks injected by setns()+fork().
	261	* Again, ignoring the user visible semantics we do not really need
	262	* to wait until they are all reaped, but they can be reparented to
	263	* us and thus we need to ensure that pid->child_reaper stays valid
	264	* until they all go away. See free_pid()->wake_up_process().
	265	*
	266	* We rely on ignored SIGCHLD, an injected zombie must be autoreaped
	267	* if reparented.
6347e900 EB	268	*/
6347e900 EB	269	for (;;) {
b9a985db	270	set_current_state(TASK_INTERRUPTIBLE);
e8cfbc24	271	if (pid_ns->pid_allocated == init_pids)
6347e900 EB	272	break;
	273	schedule();
	274	}
af4b8a83	275	__set_current_state(TASK_RUNNING);
6347e900	276
cf3f8921 DL	277	if (pid_ns->reboot)
	278	current->signal->group_exit_code = pid_ns->reboot;
	279
0b6b030f	280	acct_exit_ns(pid_ns);
74bd59bb PE	281	return;
	282	}
	283
98ed57ee	284	#ifdef CONFIG_CHECKPOINT_RESTORE
b8f566b0 PE	285	static int pid_ns_ctl_handler(struct ctl_table *table, int write,
	286	void __user buffer, size_t lenp, loff_t *ppos)
	287	{
49f4d8b9	288	struct pid_namespace *pid_ns = task_active_pid_ns(current);
b8f566b0	289	struct ctl_table tmp = *table;
95846ecf	290	int ret, next;
b8f566b0	291
49f4d8b9	292	if (write && !ns_capable(pid_ns->user_ns, CAP_SYS_ADMIN))
b8f566b0 PE	293	return -EPERM;
	294
	295	/*
	296	* Writing directly to ns' last_pid field is OK, since this field
	297	* is volatile in a living namespace anyway and a code writing to
	298	* it should synchronize its usage with external means.
	299	*/
	300
95846ecf GS	301	next = idr_get_cursor(&pid_ns->idr) - 1;
	302
	303	tmp.data = &next;
	304	ret = proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
	305	if (!ret && write)
	306	idr_set_cursor(&pid_ns->idr, next + 1);
	307
	308	return ret;
b8f566b0 PE	309	}
b8f566b0 PE	310
579035dc AV	311	extern int pid_max;
579035dc AV	312	static int zero = 0;
b8f566b0 PE	313	static struct ctl_table pid_ns_ctl_table[] = {
	314	{
	315	.procname = "ns_last_pid",
	316	.maxlen = sizeof(int),
	317	.mode = 0666, /* permissions are checked in the handler */
	318	.proc_handler = pid_ns_ctl_handler,
579035dc AV	319	.extra1 = &zero,
579035dc AV	320	.extra2 = &pid_max,
b8f566b0 PE	321	},
	322	{ }
	323	};
b8f566b0	324	static struct ctl_path kern_path[] = { { .procname = "kernel", }, { } };
98ed57ee	325	#endif /* CONFIG_CHECKPOINT_RESTORE */
b8f566b0	326
cf3f8921 DL	327	int reboot_pid_ns(struct pid_namespace *pid_ns, int cmd)
	328	{
	329	if (pid_ns == &init_pid_ns)
	330	return 0;
	331
	332	switch (cmd) {
	333	case LINUX_REBOOT_CMD_RESTART2:
	334	case LINUX_REBOOT_CMD_RESTART:
	335	pid_ns->reboot = SIGHUP;
	336	break;
	337
	338	case LINUX_REBOOT_CMD_POWER_OFF:
	339	case LINUX_REBOOT_CMD_HALT:
	340	pid_ns->reboot = SIGINT;
	341	break;
	342	default:
	343	return -EINVAL;
	344	}
	345
	346	read_lock(&tasklist_lock);
	347	force_sig(SIGKILL, pid_ns->child_reaper);
	348	read_unlock(&tasklist_lock);
	349
	350	do_exit(0);
	351
	352	/* Not reached */
	353	return 0;
	354	}
	355
3c041184 AV	356	static inline struct pid_namespace to_pid_ns(struct ns_common ns)
	357	{
	358	return container_of(ns, struct pid_namespace, ns);
	359	}
	360
64964528	361	static struct ns_common pidns_get(struct task_struct task)
57e8391d EB	362	{
	363	struct pid_namespace *ns;
	364
	365	rcu_read_lock();
d2308225 ON	366	ns = task_active_pid_ns(task);
	367	if (ns)
	368	get_pid_ns(ns);
57e8391d EB	369	rcu_read_unlock();
57e8391d EB	370
3c041184	371	return ns ? &ns->ns : NULL;
57e8391d EB	372	}
57e8391d EB	373
eaa0d190 KT	374	static struct ns_common pidns_for_children_get(struct task_struct task)
	375	{
	376	struct pid_namespace *ns = NULL;
	377
	378	task_lock(task);
	379	if (task->nsproxy) {
	380	ns = task->nsproxy->pid_ns_for_children;
	381	get_pid_ns(ns);
	382	}
	383	task_unlock(task);
	384
	385	if (ns) {
	386	read_lock(&tasklist_lock);
	387	if (!ns->child_reaper) {
	388	put_pid_ns(ns);
	389	ns = NULL;
	390	}
	391	read_unlock(&tasklist_lock);
	392	}
	393
	394	return ns ? &ns->ns : NULL;
	395	}
	396
64964528	397	static void pidns_put(struct ns_common *ns)
57e8391d	398	{
3c041184	399	put_pid_ns(to_pid_ns(ns));
57e8391d EB	400	}
57e8391d EB	401
64964528	402	static int pidns_install(struct nsproxy nsproxy, struct ns_common ns)
57e8391d EB	403	{
57e8391d EB	404	struct pid_namespace *active = task_active_pid_ns(current);
3c041184	405	struct pid_namespace ancestor, new = to_pid_ns(ns);
57e8391d	406
5e4a0847	407	if (!ns_capable(new->user_ns, CAP_SYS_ADMIN) \|\|
c7b96acf	408	!ns_capable(current_user_ns(), CAP_SYS_ADMIN))
57e8391d EB	409	return -EPERM;
	410
	411	/*
	412	* Only allow entering the current active pid namespace
	413	* or a child of the current active pid namespace.
	414	*
	415	* This is required for fork to return a usable pid value and
	416	* this maintains the property that processes and their
	417	* children can not escape their current pid namespace.
	418	*/
	419	if (new->level < active->level)
	420	return -EINVAL;
	421
	422	ancestor = new;
	423	while (ancestor->level > active->level)
	424	ancestor = ancestor->parent;
	425	if (ancestor != active)
	426	return -EINVAL;
	427
c2b1df2e AL	428	put_pid_ns(nsproxy->pid_ns_for_children);
c2b1df2e AL	429	nsproxy->pid_ns_for_children = get_pid_ns(new);
57e8391d EB	430	return 0;
	431	}
	432
a7306ed8 AV	433	static struct ns_common pidns_get_parent(struct ns_common ns)
	434	{
	435	struct pid_namespace *active = task_active_pid_ns(current);
	436	struct pid_namespace pid_ns, p;
	437
	438	/* See if the parent is in the current namespace */
	439	pid_ns = p = to_pid_ns(ns)->parent;
	440	for (;;) {
	441	if (!p)
	442	return ERR_PTR(-EPERM);
	443	if (p == active)
	444	break;
	445	p = p->parent;
	446	}
	447
	448	return &get_pid_ns(pid_ns)->ns;
	449	}
	450
bcac25a5 AV	451	static struct user_namespace pidns_owner(struct ns_common ns)
	452	{
	453	return to_pid_ns(ns)->user_ns;
	454	}
	455
57e8391d EB	456	const struct proc_ns_operations pidns_operations = {
	457	.name = "pid",
	458	.type = CLONE_NEWPID,
	459	.get = pidns_get,
	460	.put = pidns_put,
	461	.install = pidns_install,
bcac25a5	462	.owner = pidns_owner,
a7306ed8	463	.get_parent = pidns_get_parent,
57e8391d EB	464	};
57e8391d EB	465
eaa0d190 KT	466	const struct proc_ns_operations pidns_for_children_operations = {
	467	.name = "pid_for_children",
	468	.real_ns_name = "pid",
	469	.type = CLONE_NEWPID,
	470	.get = pidns_for_children_get,
	471	.put = pidns_put,
	472	.install = pidns_install,
	473	.owner = pidns_owner,
	474	.get_parent = pidns_get_parent,
	475	};
	476
74bd59bb PE	477	static __init int pid_namespaces_init(void)
	478	{
	479	pid_ns_cachep = KMEM_CACHE(pid_namespace, SLAB_PANIC);
98ed57ee CG	480
98ed57ee CG	481	#ifdef CONFIG_CHECKPOINT_RESTORE
b8f566b0	482	register_sysctl_paths(kern_path, pid_ns_ctl_table);
98ed57ee	483	#endif
74bd59bb PE	484	return 0;
	485	}
	486
	487	__initcall(pid_namespaces_init);