[thirdparty/kernel/stable.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_fs.h>
#include <linux/reboot.h>
#include <linux/export.h>

#define BITS_PER_PAGE		(PAGE_SIZE*8)

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;
}

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

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;
	int i;
	int err;

	if (level > MAX_PID_NS_LEVEL) {
		err = -EINVAL;
		goto out;
	}

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

	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;

	kref_init(&ns->kref);
	ns->level = level;
	ns->parent = get_pid_ns(parent_pid_ns);
	ns->user_ns = get_user_ns(user_ns);

	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);

	err = pid_ns_prepare_proc(ns);
	if (err)
		goto out_put_parent_pid_ns;

	return ns;

out_put_parent_pid_ns:
	put_pid_ns(parent_pid_ns);
	put_user_ns(user_ns);
out_free_map:
	kfree(ns->pidmap[0].page);
out_free:
	kmem_cache_free(pid_ns_cachep, ns);
out:
	return ERR_PTR(err);
}

static void destroy_pid_namespace(struct pid_namespace *ns)
{
	int i;

	for (i = 0; i < PIDMAP_ENTRIES; i++)
		kfree(ns->pidmap[i].page);
	put_user_ns(ns->user_ns);
	kmem_cache_free(pid_ns_cachep, ns);
}

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 (flags & (CLONE_THREAD|CLONE_PARENT))
		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;

	/* Ignore SIGCHLD causing any terminated children to autoreap */
	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);

	/* Firstly reap the EXIT_ZOMBIE children we may have. */
	do {
		clear_thread_flag(TIF_SIGPENDING);
		rc = sys_wait4(-1, NULL, __WALL, NULL);
	} while (rc != -ECHILD);

	/*
	 * sys_wait4() above can't reap the TASK_DEAD children.
	 * Make sure they all go away, see __unhash_process().
	 */
	for (;;) {
		bool need_wait = false;

		read_lock(&tasklist_lock);
		if (!list_empty(&current->children)) {
			__set_current_state(TASK_UNINTERRUPTIBLE);
			need_wait = true;
		}
		read_unlock(&tasklist_lock);

		if (!need_wait)
			break;
		schedule();
	}

	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 __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>
4308eebb	18	#include <linux/proc_fs.h>
cf3f8921	19	#include <linux/reboot.h>
523a6a94	20	#include <linux/export.h>
74bd59bb PE	21
	22	#define BITS_PER_PAGE (PAGE_SIZE*8)
	23
	24	struct pid_cache {
	25	int nr_ids;
	26	char name[16];
	27	struct kmem_cache *cachep;
	28	struct list_head list;
	29	};
	30
	31	static LIST_HEAD(pid_caches_lh);
	32	static DEFINE_MUTEX(pid_caches_mutex);
	33	static struct kmem_cache *pid_ns_cachep;
	34
	35	/*
	36	* creates the kmem cache to allocate pids from.
	37	* @nr_ids: the number of numerical ids this pid will have to carry
	38	*/
	39
	40	static struct kmem_cache *create_pid_cachep(int nr_ids)
	41	{
	42	struct pid_cache *pcache;
	43	struct kmem_cache *cachep;
	44
	45	mutex_lock(&pid_caches_mutex);
	46	list_for_each_entry(pcache, &pid_caches_lh, list)
	47	if (pcache->nr_ids == nr_ids)
	48	goto out;
	49
	50	pcache = kmalloc(sizeof(struct pid_cache), GFP_KERNEL);
	51	if (pcache == NULL)
	52	goto err_alloc;
	53
	54	snprintf(pcache->name, sizeof(pcache->name), "pid_%d", nr_ids);
	55	cachep = kmem_cache_create(pcache->name,
	56	sizeof(struct pid) + (nr_ids - 1) * sizeof(struct upid),
	57	0, SLAB_HWCACHE_ALIGN, NULL);
	58	if (cachep == NULL)
	59	goto err_cachep;
	60
	61	pcache->nr_ids = nr_ids;
	62	pcache->cachep = cachep;
	63	list_add(&pcache->list, &pid_caches_lh);
	64	out:
	65	mutex_unlock(&pid_caches_mutex);
	66	return pcache->cachep;
	67
	68	err_cachep:
	69	kfree(pcache);
	70	err_alloc:
	71	mutex_unlock(&pid_caches_mutex);
	72	return NULL;
	73	}
	74
f2302505 AV	75	/* MAX_PID_NS_LEVEL is needed for limiting size of 'struct pid' */
	76	#define MAX_PID_NS_LEVEL 32
	77
49f4d8b9 EB	78	static struct pid_namespace create_pid_namespace(struct user_namespace user_ns,
49f4d8b9 EB	79	struct pid_namespace *parent_pid_ns)
74bd59bb PE	80	{
74bd59bb PE	81	struct pid_namespace *ns;
ed469a63	82	unsigned int level = parent_pid_ns->level + 1;
f2302505 AV	83	int i;
	84	int err;
	85
	86	if (level > MAX_PID_NS_LEVEL) {
	87	err = -EINVAL;
	88	goto out;
	89	}
74bd59bb	90
f2302505	91	err = -ENOMEM;
84406c15	92	ns = kmem_cache_zalloc(pid_ns_cachep, GFP_KERNEL);
74bd59bb PE	93	if (ns == NULL)
	94	goto out;
	95
	96	ns->pidmap[0].page = kzalloc(PAGE_SIZE, GFP_KERNEL);
	97	if (!ns->pidmap[0].page)
	98	goto out_free;
	99
	100	ns->pid_cachep = create_pid_cachep(level + 1);
	101	if (ns->pid_cachep == NULL)
	102	goto out_free_map;
	103
	104	kref_init(&ns->kref);
74bd59bb	105	ns->level = level;
ed469a63	106	ns->parent = get_pid_ns(parent_pid_ns);
49f4d8b9	107	ns->user_ns = get_user_ns(user_ns);
74bd59bb PE	108
	109	set_bit(0, ns->pidmap[0].page);
	110	atomic_set(&ns->pidmap[0].nr_free, BITS_PER_PAGE - 1);
	111
84406c15	112	for (i = 1; i < PIDMAP_ENTRIES; i++)
74bd59bb	113	atomic_set(&ns->pidmap[i].nr_free, BITS_PER_PAGE);
74bd59bb	114
4308eebb EB	115	err = pid_ns_prepare_proc(ns);
	116	if (err)
	117	goto out_put_parent_pid_ns;
	118
74bd59bb PE	119	return ns;
74bd59bb PE	120
4308eebb EB	121	out_put_parent_pid_ns:
4308eebb EB	122	put_pid_ns(parent_pid_ns);
49f4d8b9	123	put_user_ns(user_ns);
74bd59bb PE	124	out_free_map:
	125	kfree(ns->pidmap[0].page);
	126	out_free:
	127	kmem_cache_free(pid_ns_cachep, ns);
	128	out:
4308eebb	129	return ERR_PTR(err);
74bd59bb PE	130	}
	131
	132	static void destroy_pid_namespace(struct pid_namespace *ns)
	133	{
	134	int i;
	135
	136	for (i = 0; i < PIDMAP_ENTRIES; i++)
	137	kfree(ns->pidmap[i].page);
49f4d8b9	138	put_user_ns(ns->user_ns);
74bd59bb PE	139	kmem_cache_free(pid_ns_cachep, ns);
	140	}
	141
49f4d8b9 EB	142	struct pid_namespace *copy_pid_ns(unsigned long flags,
49f4d8b9 EB	143	struct user_namespace user_ns, struct pid_namespace old_ns)
74bd59bb	144	{
74bd59bb	145	if (!(flags & CLONE_NEWPID))
dca4a979	146	return get_pid_ns(old_ns);
e5a47386	147	if (flags & (CLONE_THREAD\|CLONE_PARENT))
dca4a979	148	return ERR_PTR(-EINVAL);
49f4d8b9	149	return create_pid_namespace(user_ns, old_ns);
74bd59bb PE	150	}
74bd59bb PE	151
bbc2e3ef	152	static void free_pid_ns(struct kref *kref)
74bd59bb	153	{
bbc2e3ef	154	struct pid_namespace *ns;
74bd59bb PE	155
74bd59bb PE	156	ns = container_of(kref, struct pid_namespace, kref);
74bd59bb	157	destroy_pid_namespace(ns);
bbc2e3ef	158	}
74bd59bb	159
bbc2e3ef CG	160	void put_pid_ns(struct pid_namespace *ns)
	161	{
	162	struct pid_namespace *parent;
	163
	164	while (ns != &init_pid_ns) {
	165	parent = ns->parent;
	166	if (!kref_put(&ns->kref, free_pid_ns))
	167	break;
	168	ns = parent;
	169	}
74bd59bb	170	}
bbc2e3ef	171	EXPORT_SYMBOL_GPL(put_pid_ns);
74bd59bb PE	172
	173	void zap_pid_ns_processes(struct pid_namespace *pid_ns)
	174	{
	175	int nr;
	176	int rc;
00c10bc1 EB	177	struct task_struct task, me = current;
	178
	179	/* Ignore SIGCHLD causing any terminated children to autoreap */
	180	spin_lock_irq(&me->sighand->siglock);
	181	me->sighand->action[SIGCHLD - 1].sa.sa_handler = SIG_IGN;
	182	spin_unlock_irq(&me->sighand->siglock);
74bd59bb PE	183
	184	/*
	185	* The last thread in the cgroup-init thread group is terminating.
	186	* Find remaining pid_ts in the namespace, signal and wait for them
	187	* to exit.
	188	*
	189	* Note: This signals each threads in the namespace - even those that
	190	* belong to the same thread group, To avoid this, we would have
	191	* to walk the entire tasklist looking a processes in this
	192	* namespace, but that could be unnecessarily expensive if the
	193	* pid namespace has just a few processes. Or we need to
	194	* maintain a tasklist for each pid namespace.
	195	*
	196	*/
	197	read_lock(&tasklist_lock);
	198	nr = next_pidmap(pid_ns, 1);
	199	while (nr > 0) {
e4da026f SB	200	rcu_read_lock();
e4da026f SB	201
e4da026f	202	task = pid_task(find_vpid(nr), PIDTYPE_PID);
a02d6fd6 ON	203	if (task && !__fatal_signal_pending(task))
a02d6fd6 ON	204	send_sig_info(SIGKILL, SEND_SIG_FORCED, task);
e4da026f SB	205
	206	rcu_read_unlock();
	207
74bd59bb PE	208	nr = next_pidmap(pid_ns, nr);
	209	}
	210	read_unlock(&tasklist_lock);
	211
6347e900	212	/* Firstly reap the EXIT_ZOMBIE children we may have. */
74bd59bb PE	213	do {
	214	clear_thread_flag(TIF_SIGPENDING);
	215	rc = sys_wait4(-1, NULL, __WALL, NULL);
	216	} while (rc != -ECHILD);
	217
6347e900 EB	218	/*
	219	* sys_wait4() above can't reap the TASK_DEAD children.
	220	* Make sure they all go away, see __unhash_process().
	221	*/
	222	for (;;) {
	223	bool need_wait = false;
	224
	225	read_lock(&tasklist_lock);
	226	if (!list_empty(&current->children)) {
	227	__set_current_state(TASK_UNINTERRUPTIBLE);
	228	need_wait = true;
	229	}
	230	read_unlock(&tasklist_lock);
	231
	232	if (!need_wait)
	233	break;
	234	schedule();
	235	}
	236
cf3f8921 DL	237	if (pid_ns->reboot)
	238	current->signal->group_exit_code = pid_ns->reboot;
	239
0b6b030f	240	acct_exit_ns(pid_ns);
74bd59bb PE	241	return;
	242	}
	243
98ed57ee	244	#ifdef CONFIG_CHECKPOINT_RESTORE
b8f566b0 PE	245	static int pid_ns_ctl_handler(struct ctl_table *table, int write,
	246	void __user buffer, size_t lenp, loff_t *ppos)
	247	{
49f4d8b9	248	struct pid_namespace *pid_ns = task_active_pid_ns(current);
b8f566b0 PE	249	struct ctl_table tmp = *table;
b8f566b0 PE	250
49f4d8b9	251	if (write && !ns_capable(pid_ns->user_ns, CAP_SYS_ADMIN))
b8f566b0 PE	252	return -EPERM;
	253
	254	/*
	255	* Writing directly to ns' last_pid field is OK, since this field
	256	* is volatile in a living namespace anyway and a code writing to
	257	* it should synchronize its usage with external means.
	258	*/
	259
49f4d8b9	260	tmp.data = &pid_ns->last_pid;
579035dc	261	return proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
b8f566b0 PE	262	}
b8f566b0 PE	263
579035dc AV	264	extern int pid_max;
579035dc AV	265	static int zero = 0;
b8f566b0 PE	266	static struct ctl_table pid_ns_ctl_table[] = {
	267	{
	268	.procname = "ns_last_pid",
	269	.maxlen = sizeof(int),
	270	.mode = 0666, /* permissions are checked in the handler */
	271	.proc_handler = pid_ns_ctl_handler,
579035dc AV	272	.extra1 = &zero,
579035dc AV	273	.extra2 = &pid_max,
b8f566b0 PE	274	},
	275	{ }
	276	};
b8f566b0	277	static struct ctl_path kern_path[] = { { .procname = "kernel", }, { } };
98ed57ee	278	#endif /* CONFIG_CHECKPOINT_RESTORE */
b8f566b0	279
cf3f8921 DL	280	int reboot_pid_ns(struct pid_namespace *pid_ns, int cmd)
	281	{
	282	if (pid_ns == &init_pid_ns)
	283	return 0;
	284
	285	switch (cmd) {
	286	case LINUX_REBOOT_CMD_RESTART2:
	287	case LINUX_REBOOT_CMD_RESTART:
	288	pid_ns->reboot = SIGHUP;
	289	break;
	290
	291	case LINUX_REBOOT_CMD_POWER_OFF:
	292	case LINUX_REBOOT_CMD_HALT:
	293	pid_ns->reboot = SIGINT;
	294	break;
	295	default:
	296	return -EINVAL;
	297	}
	298
	299	read_lock(&tasklist_lock);
	300	force_sig(SIGKILL, pid_ns->child_reaper);
	301	read_unlock(&tasklist_lock);
	302
	303	do_exit(0);
	304
	305	/* Not reached */
	306	return 0;
	307	}
	308
74bd59bb PE	309	static __init int pid_namespaces_init(void)
	310	{
	311	pid_ns_cachep = KMEM_CACHE(pid_namespace, SLAB_PANIC);
98ed57ee CG	312
98ed57ee CG	313	#ifdef CONFIG_CHECKPOINT_RESTORE
b8f566b0	314	register_sysctl_paths(kern_path, pid_ns_ctl_table);
98ed57ee	315	#endif
74bd59bb PE	316	return 0;
	317	}
	318
	319	__initcall(pid_namespaces_init);