[linux.git] / kernel / pid_namespace.c

/*
 * Pid namespaces
 *
 * Authors:
 *    (C) 2007 Pavel Emelyanov <[email protected]>, OpenVZ, SWsoft Inc.
 *    (C) 2007 Sukadev Bhattiprolu <[email protected]>, 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>

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 i;
	int err;

	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;

	ns->pidmap[0].page = kzalloc(PAGE_SIZE, GFP_KERNEL);
	if (!ns->pidmap[0].page)
		goto out_free;

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

	err = ns_alloc_inum(&ns->ns);
	if (err)
		goto out_free_map;
	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->nr_hashed = PIDNS_HASH_ADDING;
	INIT_WORK(&ns->proc_work, proc_cleanup_work);

	set_bit(0, ns->pidmap[0].page);
	atomic_set(&ns->pidmap[0].nr_free, BITS_PER_PAGE - 1);

	for (i = 1; i < PIDMAP_ENTRIES; i++)
		atomic_set(&ns->pidmap[i].nr_free, BITS_PER_PAGE);

	return ns;

out_free_map:
	kfree(ns->pidmap[0].page);
out_free:
	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)
{
	int i;

	ns_free_inum(&ns->ns);
	for (i = 0; i < PIDMAP_ENTRIES; i++)
		kfree(ns->pidmap[i].page);
	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;

	/* 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.
	 *
	 */
	read_lock(&tasklist_lock);
	nr = next_pidmap(pid_ns, 1);
	while (nr > 0) {
		rcu_read_lock();

		task = pid_task(find_vpid(nr), PIDTYPE_PID);
		if (task && !__fatal_signal_pending(task))
			send_sig_info(SIGKILL, SEND_SIG_FORCED, task);

		rcu_read_unlock();

		nr = next_pidmap(pid_ns, nr);
	}
	read_unlock(&tasklist_lock);

	/*
	 * 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(nr_hashed == 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->nr_hashed == 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;

	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.
	 */

	tmp.data = &pid_ns->last_pid;
	return proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
}

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 <[email protected]>, OpenVZ, SWsoft Inc.
	6	* (C) 2007 Sukadev Bhattiprolu <[email protected]>, 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>
74bd59bb	24
74bd59bb PE	25	struct pid_cache {
	26	int nr_ids;
	27	char name[16];
	28	struct kmem_cache *cachep;
	29	struct list_head list;
	30	};
	31
	32	static LIST_HEAD(pid_caches_lh);
	33	static DEFINE_MUTEX(pid_caches_mutex);
	34	static struct kmem_cache *pid_ns_cachep;
	35
	36	/*
	37	* creates the kmem cache to allocate pids from.
	38	* @nr_ids: the number of numerical ids this pid will have to carry
	39	*/
	40
	41	static struct kmem_cache *create_pid_cachep(int nr_ids)
	42	{
	43	struct pid_cache *pcache;
	44	struct kmem_cache *cachep;
	45
	46	mutex_lock(&pid_caches_mutex);
	47	list_for_each_entry(pcache, &pid_caches_lh, list)
	48	if (pcache->nr_ids == nr_ids)
	49	goto out;
	50
	51	pcache = kmalloc(sizeof(struct pid_cache), GFP_KERNEL);
	52	if (pcache == NULL)
	53	goto err_alloc;
	54
	55	snprintf(pcache->name, sizeof(pcache->name), "pid_%d", nr_ids);
	56	cachep = kmem_cache_create(pcache->name,
	57	sizeof(struct pid) + (nr_ids - 1) * sizeof(struct upid),
	58	0, SLAB_HWCACHE_ALIGN, NULL);
	59	if (cachep == NULL)
	60	goto err_cachep;
	61
	62	pcache->nr_ids = nr_ids;
	63	pcache->cachep = cachep;
	64	list_add(&pcache->list, &pid_caches_lh);
	65	out:
	66	mutex_unlock(&pid_caches_mutex);
	67	return pcache->cachep;
	68
	69	err_cachep:
	70	kfree(pcache);
	71	err_alloc:
	72	mutex_unlock(&pid_caches_mutex);
	73	return NULL;
	74	}
	75
0a01f2cc EB	76	static void proc_cleanup_work(struct work_struct *work)
	77	{
	78	struct pid_namespace *ns = container_of(work, struct pid_namespace, proc_work);
	79	pid_ns_release_proc(ns);
	80	}
	81
f2302505 AV	82	/* MAX_PID_NS_LEVEL is needed for limiting size of 'struct pid' */
	83	#define MAX_PID_NS_LEVEL 32
	84
f333c700 EB	85	static struct ucounts inc_pid_namespaces(struct user_namespace ns)
	86	{
	87	return inc_ucount(ns, current_euid(), UCOUNT_PID_NAMESPACES);
	88	}
	89
	90	static void dec_pid_namespaces(struct ucounts *ucounts)
	91	{
	92	dec_ucount(ucounts, UCOUNT_PID_NAMESPACES);
	93	}
	94
49f4d8b9 EB	95	static struct pid_namespace create_pid_namespace(struct user_namespace user_ns,
49f4d8b9 EB	96	struct pid_namespace *parent_pid_ns)
74bd59bb PE	97	{
74bd59bb PE	98	struct pid_namespace *ns;
ed469a63	99	unsigned int level = parent_pid_ns->level + 1;
f333c700	100	struct ucounts *ucounts;
f2302505 AV	101	int i;
	102	int err;
	103
df75e774	104	err = -ENOSPC;
f333c700 EB	105	if (level > MAX_PID_NS_LEVEL)
	106	goto out;
	107	ucounts = inc_pid_namespaces(user_ns);
	108	if (!ucounts)
f2302505	109	goto out;
74bd59bb	110
f2302505	111	err = -ENOMEM;
84406c15	112	ns = kmem_cache_zalloc(pid_ns_cachep, GFP_KERNEL);
74bd59bb	113	if (ns == NULL)
f333c700	114	goto out_dec;
74bd59bb PE	115
	116	ns->pidmap[0].page = kzalloc(PAGE_SIZE, GFP_KERNEL);
	117	if (!ns->pidmap[0].page)
	118	goto out_free;
	119
	120	ns->pid_cachep = create_pid_cachep(level + 1);
	121	if (ns->pid_cachep == NULL)
	122	goto out_free_map;
	123
6344c433	124	err = ns_alloc_inum(&ns->ns);
98f842e6 EB	125	if (err)
98f842e6 EB	126	goto out_free_map;
33c42940	127	ns->ns.ops = &pidns_operations;
98f842e6	128
74bd59bb	129	kref_init(&ns->kref);
74bd59bb	130	ns->level = level;
ed469a63	131	ns->parent = get_pid_ns(parent_pid_ns);
49f4d8b9	132	ns->user_ns = get_user_ns(user_ns);
f333c700	133	ns->ucounts = ucounts;
c876ad76	134	ns->nr_hashed = PIDNS_HASH_ADDING;
0a01f2cc	135	INIT_WORK(&ns->proc_work, proc_cleanup_work);
74bd59bb PE	136
	137	set_bit(0, ns->pidmap[0].page);
	138	atomic_set(&ns->pidmap[0].nr_free, BITS_PER_PAGE - 1);
	139
84406c15	140	for (i = 1; i < PIDMAP_ENTRIES; i++)
74bd59bb	141	atomic_set(&ns->pidmap[i].nr_free, BITS_PER_PAGE);
74bd59bb PE	142
	143	return ns;
	144
	145	out_free_map:
	146	kfree(ns->pidmap[0].page);
	147	out_free:
	148	kmem_cache_free(pid_ns_cachep, ns);
f333c700 EB	149	out_dec:
f333c700 EB	150	dec_pid_namespaces(ucounts);
74bd59bb	151	out:
4308eebb	152	return ERR_PTR(err);
74bd59bb PE	153	}
74bd59bb PE	154
1adfcb03 AV	155	static void delayed_free_pidns(struct rcu_head *p)
1adfcb03 AV	156	{
add7c65c AV	157	struct pid_namespace *ns = container_of(p, struct pid_namespace, rcu);
	158
	159	dec_pid_namespaces(ns->ucounts);
	160	put_user_ns(ns->user_ns);
	161
	162	kmem_cache_free(pid_ns_cachep, ns);
1adfcb03 AV	163	}
1adfcb03 AV	164
74bd59bb PE	165	static void destroy_pid_namespace(struct pid_namespace *ns)
	166	{
	167	int i;
	168
6344c433	169	ns_free_inum(&ns->ns);
74bd59bb PE	170	for (i = 0; i < PIDMAP_ENTRIES; i++)
74bd59bb PE	171	kfree(ns->pidmap[i].page);
1adfcb03	172	call_rcu(&ns->rcu, delayed_free_pidns);
74bd59bb PE	173	}
74bd59bb PE	174
49f4d8b9 EB	175	struct pid_namespace *copy_pid_ns(unsigned long flags,
49f4d8b9 EB	176	struct user_namespace user_ns, struct pid_namespace old_ns)
74bd59bb	177	{
74bd59bb	178	if (!(flags & CLONE_NEWPID))
dca4a979	179	return get_pid_ns(old_ns);
225778d6 EB	180	if (task_active_pid_ns(current) != old_ns)
225778d6 EB	181	return ERR_PTR(-EINVAL);
49f4d8b9	182	return create_pid_namespace(user_ns, old_ns);
74bd59bb PE	183	}
74bd59bb PE	184
bbc2e3ef	185	static void free_pid_ns(struct kref *kref)
74bd59bb	186	{
bbc2e3ef	187	struct pid_namespace *ns;
74bd59bb PE	188
74bd59bb PE	189	ns = container_of(kref, struct pid_namespace, kref);
74bd59bb	190	destroy_pid_namespace(ns);
bbc2e3ef	191	}
74bd59bb	192
bbc2e3ef CG	193	void put_pid_ns(struct pid_namespace *ns)
	194	{
	195	struct pid_namespace *parent;
	196
	197	while (ns != &init_pid_ns) {
	198	parent = ns->parent;
	199	if (!kref_put(&ns->kref, free_pid_ns))
	200	break;
	201	ns = parent;
	202	}
74bd59bb	203	}
bbc2e3ef	204	EXPORT_SYMBOL_GPL(put_pid_ns);
74bd59bb PE	205
	206	void zap_pid_ns_processes(struct pid_namespace *pid_ns)
	207	{
	208	int nr;
	209	int rc;
00c10bc1	210	struct task_struct task, me = current;
751c644b	211	int init_pids = thread_group_leader(me) ? 1 : 2;
00c10bc1	212
c876ad76 EB	213	/* Don't allow any more processes into the pid namespace */
	214	disable_pid_allocation(pid_ns);
	215
a53b8315 ON	216	/*
	217	* Ignore SIGCHLD causing any terminated children to autoreap.
	218	* This speeds up the namespace shutdown, plus see the comment
	219	* below.
	220	*/
00c10bc1 EB	221	spin_lock_irq(&me->sighand->siglock);
	222	me->sighand->action[SIGCHLD - 1].sa.sa_handler = SIG_IGN;
	223	spin_unlock_irq(&me->sighand->siglock);
74bd59bb PE	224
	225	/*
	226	* The last thread in the cgroup-init thread group is terminating.
	227	* Find remaining pid_ts in the namespace, signal and wait for them
	228	* to exit.
	229	*
	230	* Note: This signals each threads in the namespace - even those that
	231	* belong to the same thread group, To avoid this, we would have
	232	* to walk the entire tasklist looking a processes in this
	233	* namespace, but that could be unnecessarily expensive if the
	234	* pid namespace has just a few processes. Or we need to
	235	* maintain a tasklist for each pid namespace.
	236	*
	237	*/
	238	read_lock(&tasklist_lock);
	239	nr = next_pidmap(pid_ns, 1);
	240	while (nr > 0) {
e4da026f SB	241	rcu_read_lock();
e4da026f SB	242
e4da026f	243	task = pid_task(find_vpid(nr), PIDTYPE_PID);
a02d6fd6 ON	244	if (task && !__fatal_signal_pending(task))
a02d6fd6 ON	245	send_sig_info(SIGKILL, SEND_SIG_FORCED, task);
e4da026f SB	246
	247	rcu_read_unlock();
	248
74bd59bb PE	249	nr = next_pidmap(pid_ns, nr);
	250	}
	251	read_unlock(&tasklist_lock);
	252
a53b8315 ON	253	/*
	254	* Reap the EXIT_ZOMBIE children we had before we ignored SIGCHLD.
	255	* sys_wait4() will also block until our children traced from the
	256	* parent namespace are detached and become EXIT_DEAD.
	257	*/
74bd59bb PE	258	do {
	259	clear_thread_flag(TIF_SIGPENDING);
	260	rc = sys_wait4(-1, NULL, __WALL, NULL);
	261	} while (rc != -ECHILD);
	262
6347e900	263	/*
a53b8315 ON	264	* sys_wait4() above can't reap the EXIT_DEAD children but we do not
	265	* really care, we could reparent them to the global init. We could
	266	* exit and reap ->child_reaper even if it is not the last thread in
	267	* this pid_ns, free_pid(nr_hashed == 0) calls proc_cleanup_work(),
	268	* pid_ns can not go away until proc_kill_sb() drops the reference.
	269	*
	270	* But this ns can also have other tasks injected by setns()+fork().
	271	* Again, ignoring the user visible semantics we do not really need
	272	* to wait until they are all reaped, but they can be reparented to
	273	* us and thus we need to ensure that pid->child_reaper stays valid
	274	* until they all go away. See free_pid()->wake_up_process().
	275	*
	276	* We rely on ignored SIGCHLD, an injected zombie must be autoreaped
	277	* if reparented.
6347e900 EB	278	*/
6347e900 EB	279	for (;;) {
b9a985db	280	set_current_state(TASK_INTERRUPTIBLE);
751c644b	281	if (pid_ns->nr_hashed == init_pids)
6347e900 EB	282	break;
	283	schedule();
	284	}
af4b8a83	285	__set_current_state(TASK_RUNNING);
6347e900	286
cf3f8921 DL	287	if (pid_ns->reboot)
	288	current->signal->group_exit_code = pid_ns->reboot;
	289
0b6b030f	290	acct_exit_ns(pid_ns);
74bd59bb PE	291	return;
	292	}
	293
98ed57ee	294	#ifdef CONFIG_CHECKPOINT_RESTORE
b8f566b0 PE	295	static int pid_ns_ctl_handler(struct ctl_table *table, int write,
	296	void __user buffer, size_t lenp, loff_t *ppos)
	297	{
49f4d8b9	298	struct pid_namespace *pid_ns = task_active_pid_ns(current);
b8f566b0 PE	299	struct ctl_table tmp = *table;
b8f566b0 PE	300
49f4d8b9	301	if (write && !ns_capable(pid_ns->user_ns, CAP_SYS_ADMIN))
b8f566b0 PE	302	return -EPERM;
	303
	304	/*
	305	* Writing directly to ns' last_pid field is OK, since this field
	306	* is volatile in a living namespace anyway and a code writing to
	307	* it should synchronize its usage with external means.
	308	*/
	309
49f4d8b9	310	tmp.data = &pid_ns->last_pid;
579035dc	311	return proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
b8f566b0 PE	312	}
b8f566b0 PE	313
579035dc AV	314	extern int pid_max;
579035dc AV	315	static int zero = 0;
b8f566b0 PE	316	static struct ctl_table pid_ns_ctl_table[] = {
	317	{
	318	.procname = "ns_last_pid",
	319	.maxlen = sizeof(int),
	320	.mode = 0666, /* permissions are checked in the handler */
	321	.proc_handler = pid_ns_ctl_handler,
579035dc AV	322	.extra1 = &zero,
579035dc AV	323	.extra2 = &pid_max,
b8f566b0 PE	324	},
	325	{ }
	326	};
b8f566b0	327	static struct ctl_path kern_path[] = { { .procname = "kernel", }, { } };
98ed57ee	328	#endif /* CONFIG_CHECKPOINT_RESTORE */
b8f566b0	329
cf3f8921 DL	330	int reboot_pid_ns(struct pid_namespace *pid_ns, int cmd)
	331	{
	332	if (pid_ns == &init_pid_ns)
	333	return 0;
	334
	335	switch (cmd) {
	336	case LINUX_REBOOT_CMD_RESTART2:
	337	case LINUX_REBOOT_CMD_RESTART:
	338	pid_ns->reboot = SIGHUP;
	339	break;
	340
	341	case LINUX_REBOOT_CMD_POWER_OFF:
	342	case LINUX_REBOOT_CMD_HALT:
	343	pid_ns->reboot = SIGINT;
	344	break;
	345	default:
	346	return -EINVAL;
	347	}
	348
	349	read_lock(&tasklist_lock);
	350	force_sig(SIGKILL, pid_ns->child_reaper);
	351	read_unlock(&tasklist_lock);
	352
	353	do_exit(0);
	354
	355	/* Not reached */
	356	return 0;
	357	}
	358
3c041184 AV	359	static inline struct pid_namespace to_pid_ns(struct ns_common ns)
	360	{
	361	return container_of(ns, struct pid_namespace, ns);
	362	}
	363
64964528	364	static struct ns_common pidns_get(struct task_struct task)
57e8391d EB	365	{
	366	struct pid_namespace *ns;
	367
	368	rcu_read_lock();
d2308225 ON	369	ns = task_active_pid_ns(task);
	370	if (ns)
	371	get_pid_ns(ns);
57e8391d EB	372	rcu_read_unlock();
57e8391d EB	373
3c041184	374	return ns ? &ns->ns : NULL;
57e8391d EB	375	}
57e8391d EB	376
eaa0d190 KT	377	static struct ns_common pidns_for_children_get(struct task_struct task)
	378	{
	379	struct pid_namespace *ns = NULL;
	380
	381	task_lock(task);
	382	if (task->nsproxy) {
	383	ns = task->nsproxy->pid_ns_for_children;
	384	get_pid_ns(ns);
	385	}
	386	task_unlock(task);
	387
	388	if (ns) {
	389	read_lock(&tasklist_lock);
	390	if (!ns->child_reaper) {
	391	put_pid_ns(ns);
	392	ns = NULL;
	393	}
	394	read_unlock(&tasklist_lock);
	395	}
	396
	397	return ns ? &ns->ns : NULL;
	398	}
	399
64964528	400	static void pidns_put(struct ns_common *ns)
57e8391d	401	{
3c041184	402	put_pid_ns(to_pid_ns(ns));
57e8391d EB	403	}
57e8391d EB	404
64964528	405	static int pidns_install(struct nsproxy nsproxy, struct ns_common ns)
57e8391d EB	406	{
57e8391d EB	407	struct pid_namespace *active = task_active_pid_ns(current);
3c041184	408	struct pid_namespace ancestor, new = to_pid_ns(ns);
57e8391d	409
5e4a0847	410	if (!ns_capable(new->user_ns, CAP_SYS_ADMIN) \|\|
c7b96acf	411	!ns_capable(current_user_ns(), CAP_SYS_ADMIN))
57e8391d EB	412	return -EPERM;
	413
	414	/*
	415	* Only allow entering the current active pid namespace
	416	* or a child of the current active pid namespace.
	417	*
	418	* This is required for fork to return a usable pid value and
	419	* this maintains the property that processes and their
	420	* children can not escape their current pid namespace.
	421	*/
	422	if (new->level < active->level)
	423	return -EINVAL;
	424
	425	ancestor = new;
	426	while (ancestor->level > active->level)
	427	ancestor = ancestor->parent;
	428	if (ancestor != active)
	429	return -EINVAL;
	430
c2b1df2e AL	431	put_pid_ns(nsproxy->pid_ns_for_children);
c2b1df2e AL	432	nsproxy->pid_ns_for_children = get_pid_ns(new);
57e8391d EB	433	return 0;
	434	}
	435
a7306ed8 AV	436	static struct ns_common pidns_get_parent(struct ns_common ns)
	437	{
	438	struct pid_namespace *active = task_active_pid_ns(current);
	439	struct pid_namespace pid_ns, p;
	440
	441	/* See if the parent is in the current namespace */
	442	pid_ns = p = to_pid_ns(ns)->parent;
	443	for (;;) {
	444	if (!p)
	445	return ERR_PTR(-EPERM);
	446	if (p == active)
	447	break;
	448	p = p->parent;
	449	}
	450
	451	return &get_pid_ns(pid_ns)->ns;
	452	}
	453
bcac25a5 AV	454	static struct user_namespace pidns_owner(struct ns_common ns)
	455	{
	456	return to_pid_ns(ns)->user_ns;
	457	}
	458
57e8391d EB	459	const struct proc_ns_operations pidns_operations = {
	460	.name = "pid",
	461	.type = CLONE_NEWPID,
	462	.get = pidns_get,
	463	.put = pidns_put,
	464	.install = pidns_install,
bcac25a5	465	.owner = pidns_owner,
a7306ed8	466	.get_parent = pidns_get_parent,
57e8391d EB	467	};
57e8391d EB	468
eaa0d190 KT	469	const struct proc_ns_operations pidns_for_children_operations = {
	470	.name = "pid_for_children",
	471	.real_ns_name = "pid",
	472	.type = CLONE_NEWPID,
	473	.get = pidns_for_children_get,
	474	.put = pidns_put,
	475	.install = pidns_install,
	476	.owner = pidns_owner,
	477	.get_parent = pidns_get_parent,
	478	};
	479
74bd59bb PE	480	static __init int pid_namespaces_init(void)
	481	{
	482	pid_ns_cachep = KMEM_CACHE(pid_namespace, SLAB_PANIC);
98ed57ee CG	483
98ed57ee CG	484	#ifdef CONFIG_CHECKPOINT_RESTORE
b8f566b0	485	register_sysctl_paths(kern_path, pid_ns_ctl_table);
98ed57ee	486	#endif
74bd59bb PE	487	return 0;
	488	}
	489
	490	__initcall(pid_namespaces_init);