[mirror_ubuntu-zesty-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/err.h>
#include <linux/acct.h>
#include <linux/slab.h>
#include <linux/proc_ns.h>
#include <linux/reboot.h>
#include <linux/export.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 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,
};

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 PE	14	#include <linux/syscalls.h>
74bd59bb PE	15	#include <linux/err.h>
0b6b030f	16	#include <linux/acct.h>
5a0e3ad6	17	#include <linux/slab.h>
0bb80f24	18	#include <linux/proc_ns.h>
cf3f8921	19	#include <linux/reboot.h>
523a6a94	20	#include <linux/export.h>
74bd59bb	21
74bd59bb PE	22	struct pid_cache {
	23	int nr_ids;
	24	char name[16];
	25	struct kmem_cache *cachep;
	26	struct list_head list;
	27	};
	28
	29	static LIST_HEAD(pid_caches_lh);
	30	static DEFINE_MUTEX(pid_caches_mutex);
	31	static struct kmem_cache *pid_ns_cachep;
	32
	33	/*
	34	* creates the kmem cache to allocate pids from.
	35	* @nr_ids: the number of numerical ids this pid will have to carry
	36	*/
	37
	38	static struct kmem_cache *create_pid_cachep(int nr_ids)
	39	{
	40	struct pid_cache *pcache;
	41	struct kmem_cache *cachep;
	42
	43	mutex_lock(&pid_caches_mutex);
	44	list_for_each_entry(pcache, &pid_caches_lh, list)
	45	if (pcache->nr_ids == nr_ids)
	46	goto out;
	47
	48	pcache = kmalloc(sizeof(struct pid_cache), GFP_KERNEL);
	49	if (pcache == NULL)
	50	goto err_alloc;
	51
	52	snprintf(pcache->name, sizeof(pcache->name), "pid_%d", nr_ids);
	53	cachep = kmem_cache_create(pcache->name,
	54	sizeof(struct pid) + (nr_ids - 1) * sizeof(struct upid),
	55	0, SLAB_HWCACHE_ALIGN, NULL);
	56	if (cachep == NULL)
	57	goto err_cachep;
	58
	59	pcache->nr_ids = nr_ids;
	60	pcache->cachep = cachep;
	61	list_add(&pcache->list, &pid_caches_lh);
	62	out:
	63	mutex_unlock(&pid_caches_mutex);
	64	return pcache->cachep;
	65
	66	err_cachep:
	67	kfree(pcache);
	68	err_alloc:
	69	mutex_unlock(&pid_caches_mutex);
	70	return NULL;
	71	}
	72
0a01f2cc EB	73	static void proc_cleanup_work(struct work_struct *work)
	74	{
	75	struct pid_namespace *ns = container_of(work, struct pid_namespace, proc_work);
	76	pid_ns_release_proc(ns);
	77	}
	78
f2302505 AV	79	/* MAX_PID_NS_LEVEL is needed for limiting size of 'struct pid' */
	80	#define MAX_PID_NS_LEVEL 32
	81
f333c700 EB	82	static struct ucounts inc_pid_namespaces(struct user_namespace ns)
	83	{
	84	return inc_ucount(ns, current_euid(), UCOUNT_PID_NAMESPACES);
	85	}
	86
	87	static void dec_pid_namespaces(struct ucounts *ucounts)
	88	{
	89	dec_ucount(ucounts, UCOUNT_PID_NAMESPACES);
	90	}
	91
49f4d8b9 EB	92	static struct pid_namespace create_pid_namespace(struct user_namespace user_ns,
49f4d8b9 EB	93	struct pid_namespace *parent_pid_ns)
74bd59bb PE	94	{
74bd59bb PE	95	struct pid_namespace *ns;
ed469a63	96	unsigned int level = parent_pid_ns->level + 1;
f333c700	97	struct ucounts *ucounts;
f2302505 AV	98	int i;
	99	int err;
	100
df75e774	101	err = -ENOSPC;
f333c700 EB	102	if (level > MAX_PID_NS_LEVEL)
	103	goto out;
	104	ucounts = inc_pid_namespaces(user_ns);
	105	if (!ucounts)
f2302505	106	goto out;
74bd59bb	107
f2302505	108	err = -ENOMEM;
84406c15	109	ns = kmem_cache_zalloc(pid_ns_cachep, GFP_KERNEL);
74bd59bb	110	if (ns == NULL)
f333c700	111	goto out_dec;
74bd59bb PE	112
	113	ns->pidmap[0].page = kzalloc(PAGE_SIZE, GFP_KERNEL);
	114	if (!ns->pidmap[0].page)
	115	goto out_free;
	116
	117	ns->pid_cachep = create_pid_cachep(level + 1);
	118	if (ns->pid_cachep == NULL)
	119	goto out_free_map;
	120
6344c433	121	err = ns_alloc_inum(&ns->ns);
98f842e6 EB	122	if (err)
98f842e6 EB	123	goto out_free_map;
33c42940	124	ns->ns.ops = &pidns_operations;
98f842e6	125
74bd59bb	126	kref_init(&ns->kref);
74bd59bb	127	ns->level = level;
ed469a63	128	ns->parent = get_pid_ns(parent_pid_ns);
49f4d8b9	129	ns->user_ns = get_user_ns(user_ns);
f333c700	130	ns->ucounts = ucounts;
c876ad76	131	ns->nr_hashed = PIDNS_HASH_ADDING;
0a01f2cc	132	INIT_WORK(&ns->proc_work, proc_cleanup_work);
74bd59bb PE	133
	134	set_bit(0, ns->pidmap[0].page);
	135	atomic_set(&ns->pidmap[0].nr_free, BITS_PER_PAGE - 1);
	136
84406c15	137	for (i = 1; i < PIDMAP_ENTRIES; i++)
74bd59bb	138	atomic_set(&ns->pidmap[i].nr_free, BITS_PER_PAGE);
74bd59bb PE	139
	140	return ns;
	141
	142	out_free_map:
	143	kfree(ns->pidmap[0].page);
	144	out_free:
	145	kmem_cache_free(pid_ns_cachep, ns);
f333c700 EB	146	out_dec:
f333c700 EB	147	dec_pid_namespaces(ucounts);
74bd59bb	148	out:
4308eebb	149	return ERR_PTR(err);
74bd59bb PE	150	}
74bd59bb PE	151
1adfcb03 AV	152	static void delayed_free_pidns(struct rcu_head *p)
1adfcb03 AV	153	{
add7c65c AV	154	struct pid_namespace *ns = container_of(p, struct pid_namespace, rcu);
	155
	156	dec_pid_namespaces(ns->ucounts);
	157	put_user_ns(ns->user_ns);
	158
	159	kmem_cache_free(pid_ns_cachep, ns);
1adfcb03 AV	160	}
1adfcb03 AV	161
74bd59bb PE	162	static void destroy_pid_namespace(struct pid_namespace *ns)
	163	{
	164	int i;
	165
6344c433	166	ns_free_inum(&ns->ns);
74bd59bb PE	167	for (i = 0; i < PIDMAP_ENTRIES; i++)
74bd59bb PE	168	kfree(ns->pidmap[i].page);
1adfcb03	169	call_rcu(&ns->rcu, delayed_free_pidns);
74bd59bb PE	170	}
74bd59bb PE	171
49f4d8b9 EB	172	struct pid_namespace *copy_pid_ns(unsigned long flags,
49f4d8b9 EB	173	struct user_namespace user_ns, struct pid_namespace old_ns)
74bd59bb	174	{
74bd59bb	175	if (!(flags & CLONE_NEWPID))
dca4a979	176	return get_pid_ns(old_ns);
225778d6 EB	177	if (task_active_pid_ns(current) != old_ns)
225778d6 EB	178	return ERR_PTR(-EINVAL);
49f4d8b9	179	return create_pid_namespace(user_ns, old_ns);
74bd59bb PE	180	}
74bd59bb PE	181
bbc2e3ef	182	static void free_pid_ns(struct kref *kref)
74bd59bb	183	{
bbc2e3ef	184	struct pid_namespace *ns;
74bd59bb PE	185
74bd59bb PE	186	ns = container_of(kref, struct pid_namespace, kref);
74bd59bb	187	destroy_pid_namespace(ns);
bbc2e3ef	188	}
74bd59bb	189
bbc2e3ef CG	190	void put_pid_ns(struct pid_namespace *ns)
	191	{
	192	struct pid_namespace *parent;
	193
	194	while (ns != &init_pid_ns) {
	195	parent = ns->parent;
	196	if (!kref_put(&ns->kref, free_pid_ns))
	197	break;
	198	ns = parent;
	199	}
74bd59bb	200	}
bbc2e3ef	201	EXPORT_SYMBOL_GPL(put_pid_ns);
74bd59bb PE	202
	203	void zap_pid_ns_processes(struct pid_namespace *pid_ns)
	204	{
	205	int nr;
	206	int rc;
00c10bc1	207	struct task_struct task, me = current;
751c644b	208	int init_pids = thread_group_leader(me) ? 1 : 2;
00c10bc1	209
c876ad76 EB	210	/* Don't allow any more processes into the pid namespace */
	211	disable_pid_allocation(pid_ns);
	212
a53b8315 ON	213	/*
	214	* Ignore SIGCHLD causing any terminated children to autoreap.
	215	* This speeds up the namespace shutdown, plus see the comment
	216	* below.
	217	*/
00c10bc1 EB	218	spin_lock_irq(&me->sighand->siglock);
	219	me->sighand->action[SIGCHLD - 1].sa.sa_handler = SIG_IGN;
	220	spin_unlock_irq(&me->sighand->siglock);
74bd59bb PE	221
	222	/*
	223	* The last thread in the cgroup-init thread group is terminating.
	224	* Find remaining pid_ts in the namespace, signal and wait for them
	225	* to exit.
	226	*
	227	* Note: This signals each threads in the namespace - even those that
	228	* belong to the same thread group, To avoid this, we would have
	229	* to walk the entire tasklist looking a processes in this
	230	* namespace, but that could be unnecessarily expensive if the
	231	* pid namespace has just a few processes. Or we need to
	232	* maintain a tasklist for each pid namespace.
	233	*
	234	*/
	235	read_lock(&tasklist_lock);
	236	nr = next_pidmap(pid_ns, 1);
	237	while (nr > 0) {
e4da026f SB	238	rcu_read_lock();
e4da026f SB	239
e4da026f	240	task = pid_task(find_vpid(nr), PIDTYPE_PID);
a02d6fd6 ON	241	if (task && !__fatal_signal_pending(task))
a02d6fd6 ON	242	send_sig_info(SIGKILL, SEND_SIG_FORCED, task);
e4da026f SB	243
	244	rcu_read_unlock();
	245
74bd59bb PE	246	nr = next_pidmap(pid_ns, nr);
	247	}
	248	read_unlock(&tasklist_lock);
	249
a53b8315 ON	250	/*
	251	* Reap the EXIT_ZOMBIE children we had before we ignored SIGCHLD.
	252	* sys_wait4() will also block until our children traced from the
	253	* parent namespace are detached and become EXIT_DEAD.
	254	*/
74bd59bb PE	255	do {
	256	clear_thread_flag(TIF_SIGPENDING);
	257	rc = sys_wait4(-1, NULL, __WALL, NULL);
	258	} while (rc != -ECHILD);
	259
6347e900	260	/*
a53b8315 ON	261	* sys_wait4() above can't reap the EXIT_DEAD children but we do not
	262	* really care, we could reparent them to the global init. We could
	263	* exit and reap ->child_reaper even if it is not the last thread in
	264	* this pid_ns, free_pid(nr_hashed == 0) calls proc_cleanup_work(),
	265	* pid_ns can not go away until proc_kill_sb() drops the reference.
	266	*
	267	* But this ns can also have other tasks injected by setns()+fork().
	268	* Again, ignoring the user visible semantics we do not really need
	269	* to wait until they are all reaped, but they can be reparented to
	270	* us and thus we need to ensure that pid->child_reaper stays valid
	271	* until they all go away. See free_pid()->wake_up_process().
	272	*
	273	* We rely on ignored SIGCHLD, an injected zombie must be autoreaped
	274	* if reparented.
6347e900 EB	275	*/
6347e900 EB	276	for (;;) {
6f48766e	277	set_current_state(TASK_INTERRUPTIBLE);
751c644b	278	if (pid_ns->nr_hashed == init_pids)
6347e900 EB	279	break;
	280	schedule();
	281	}
af4b8a83	282	__set_current_state(TASK_RUNNING);
6347e900	283
cf3f8921 DL	284	if (pid_ns->reboot)
	285	current->signal->group_exit_code = pid_ns->reboot;
	286
0b6b030f	287	acct_exit_ns(pid_ns);
74bd59bb PE	288	return;
	289	}
	290
98ed57ee	291	#ifdef CONFIG_CHECKPOINT_RESTORE
b8f566b0 PE	292	static int pid_ns_ctl_handler(struct ctl_table *table, int write,
	293	void __user buffer, size_t lenp, loff_t *ppos)
	294	{
49f4d8b9	295	struct pid_namespace *pid_ns = task_active_pid_ns(current);
b8f566b0 PE	296	struct ctl_table tmp = *table;
b8f566b0 PE	297
49f4d8b9	298	if (write && !ns_capable(pid_ns->user_ns, CAP_SYS_ADMIN))
b8f566b0 PE	299	return -EPERM;
	300
	301	/*
	302	* Writing directly to ns' last_pid field is OK, since this field
	303	* is volatile in a living namespace anyway and a code writing to
	304	* it should synchronize its usage with external means.
	305	*/
	306
49f4d8b9	307	tmp.data = &pid_ns->last_pid;
579035dc	308	return proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
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
64964528	374	static void pidns_put(struct ns_common *ns)
57e8391d	375	{
3c041184	376	put_pid_ns(to_pid_ns(ns));
57e8391d EB	377	}
57e8391d EB	378
64964528	379	static int pidns_install(struct nsproxy nsproxy, struct ns_common ns)
57e8391d EB	380	{
57e8391d EB	381	struct pid_namespace *active = task_active_pid_ns(current);
3c041184	382	struct pid_namespace ancestor, new = to_pid_ns(ns);
57e8391d	383
5e4a0847	384	if (!ns_capable(new->user_ns, CAP_SYS_ADMIN) \|\|
c7b96acf	385	!ns_capable(current_user_ns(), CAP_SYS_ADMIN))
57e8391d EB	386	return -EPERM;
	387
	388	/*
	389	* Only allow entering the current active pid namespace
	390	* or a child of the current active pid namespace.
	391	*
	392	* This is required for fork to return a usable pid value and
	393	* this maintains the property that processes and their
	394	* children can not escape their current pid namespace.
	395	*/
	396	if (new->level < active->level)
	397	return -EINVAL;
	398
	399	ancestor = new;
	400	while (ancestor->level > active->level)
	401	ancestor = ancestor->parent;
	402	if (ancestor != active)
	403	return -EINVAL;
	404
c2b1df2e AL	405	put_pid_ns(nsproxy->pid_ns_for_children);
c2b1df2e AL	406	nsproxy->pid_ns_for_children = get_pid_ns(new);
57e8391d EB	407	return 0;
	408	}
	409
a7306ed8 AV	410	static struct ns_common pidns_get_parent(struct ns_common ns)
	411	{
	412	struct pid_namespace *active = task_active_pid_ns(current);
	413	struct pid_namespace pid_ns, p;
	414
	415	/* See if the parent is in the current namespace */
	416	pid_ns = p = to_pid_ns(ns)->parent;
	417	for (;;) {
	418	if (!p)
	419	return ERR_PTR(-EPERM);
	420	if (p == active)
	421	break;
	422	p = p->parent;
	423	}
	424
	425	return &get_pid_ns(pid_ns)->ns;
	426	}
	427
bcac25a5 AV	428	static struct user_namespace pidns_owner(struct ns_common ns)
	429	{
	430	return to_pid_ns(ns)->user_ns;
	431	}
	432
57e8391d EB	433	const struct proc_ns_operations pidns_operations = {
	434	.name = "pid",
	435	.type = CLONE_NEWPID,
	436	.get = pidns_get,
	437	.put = pidns_put,
	438	.install = pidns_install,
bcac25a5	439	.owner = pidns_owner,
a7306ed8	440	.get_parent = pidns_get_parent,
57e8391d EB	441	};
57e8391d EB	442
74bd59bb PE	443	static __init int pid_namespaces_init(void)
	444	{
	445	pid_ns_cachep = KMEM_CACHE(pid_namespace, SLAB_PANIC);
98ed57ee CG	446
98ed57ee CG	447	#ifdef CONFIG_CHECKPOINT_RESTORE
b8f566b0	448	register_sysctl_paths(kern_path, pid_ns_ctl_table);
98ed57ee	449	#endif
74bd59bb PE	450	return 0;
	451	}
	452
	453	__initcall(pid_namespaces_init);