[deliverable/linux.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/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 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);

	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);
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);
	kmem_cache_free(pid_ns_cachep, ns);
}

struct pid_namespace *copy_pid_ns(unsigned long flags, 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(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 ctl_table tmp = *table;

	if (write && !capable(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 = &current->nsproxy->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>
	13	#include <linux/syscalls.h>
	14	#include <linux/err.h>
0b6b030f	15	#include <linux/acct.h>
5a0e3ad6	16	#include <linux/slab.h>
4308eebb	17	#include <linux/proc_fs.h>
cf3f8921	18	#include <linux/reboot.h>
523a6a94	19	#include <linux/export.h>
74bd59bb PE	20
	21	#define BITS_PER_PAGE (PAGE_SIZE*8)
	22
	23	struct pid_cache {
	24	int nr_ids;
	25	char name[16];
	26	struct kmem_cache *cachep;
	27	struct list_head list;
	28	};
	29
	30	static LIST_HEAD(pid_caches_lh);
	31	static DEFINE_MUTEX(pid_caches_mutex);
	32	static struct kmem_cache *pid_ns_cachep;
	33
	34	/*
	35	* creates the kmem cache to allocate pids from.
	36	* @nr_ids: the number of numerical ids this pid will have to carry
	37	*/
	38
	39	static struct kmem_cache *create_pid_cachep(int nr_ids)
	40	{
	41	struct pid_cache *pcache;
	42	struct kmem_cache *cachep;
	43
	44	mutex_lock(&pid_caches_mutex);
	45	list_for_each_entry(pcache, &pid_caches_lh, list)
	46	if (pcache->nr_ids == nr_ids)
	47	goto out;
	48
	49	pcache = kmalloc(sizeof(struct pid_cache), GFP_KERNEL);
	50	if (pcache == NULL)
	51	goto err_alloc;
	52
	53	snprintf(pcache->name, sizeof(pcache->name), "pid_%d", nr_ids);
	54	cachep = kmem_cache_create(pcache->name,
	55	sizeof(struct pid) + (nr_ids - 1) * sizeof(struct upid),
	56	0, SLAB_HWCACHE_ALIGN, NULL);
	57	if (cachep == NULL)
	58	goto err_cachep;
	59
	60	pcache->nr_ids = nr_ids;
	61	pcache->cachep = cachep;
	62	list_add(&pcache->list, &pid_caches_lh);
	63	out:
	64	mutex_unlock(&pid_caches_mutex);
	65	return pcache->cachep;
	66
	67	err_cachep:
	68	kfree(pcache);
	69	err_alloc:
	70	mutex_unlock(&pid_caches_mutex);
	71	return NULL;
	72	}
	73
f2302505 AV	74	/* MAX_PID_NS_LEVEL is needed for limiting size of 'struct pid' */
	75	#define MAX_PID_NS_LEVEL 32
	76
ed469a63	77	static struct pid_namespace create_pid_namespace(struct pid_namespace parent_pid_ns)
74bd59bb PE	78	{
74bd59bb PE	79	struct pid_namespace *ns;
ed469a63	80	unsigned int level = parent_pid_ns->level + 1;
f2302505 AV	81	int i;
	82	int err;
	83
	84	if (level > MAX_PID_NS_LEVEL) {
	85	err = -EINVAL;
	86	goto out;
	87	}
74bd59bb	88
f2302505	89	err = -ENOMEM;
84406c15	90	ns = kmem_cache_zalloc(pid_ns_cachep, GFP_KERNEL);
74bd59bb PE	91	if (ns == NULL)
	92	goto out;
	93
	94	ns->pidmap[0].page = kzalloc(PAGE_SIZE, GFP_KERNEL);
	95	if (!ns->pidmap[0].page)
	96	goto out_free;
	97
	98	ns->pid_cachep = create_pid_cachep(level + 1);
	99	if (ns->pid_cachep == NULL)
	100	goto out_free_map;
	101
	102	kref_init(&ns->kref);
74bd59bb	103	ns->level = level;
ed469a63	104	ns->parent = get_pid_ns(parent_pid_ns);
74bd59bb PE	105
	106	set_bit(0, ns->pidmap[0].page);
	107	atomic_set(&ns->pidmap[0].nr_free, BITS_PER_PAGE - 1);
	108
84406c15	109	for (i = 1; i < PIDMAP_ENTRIES; i++)
74bd59bb	110	atomic_set(&ns->pidmap[i].nr_free, BITS_PER_PAGE);
74bd59bb	111
4308eebb EB	112	err = pid_ns_prepare_proc(ns);
	113	if (err)
	114	goto out_put_parent_pid_ns;
	115
74bd59bb PE	116	return ns;
74bd59bb PE	117
4308eebb EB	118	out_put_parent_pid_ns:
4308eebb EB	119	put_pid_ns(parent_pid_ns);
74bd59bb PE	120	out_free_map:
	121	kfree(ns->pidmap[0].page);
	122	out_free:
	123	kmem_cache_free(pid_ns_cachep, ns);
	124	out:
4308eebb	125	return ERR_PTR(err);
74bd59bb PE	126	}
	127
	128	static void destroy_pid_namespace(struct pid_namespace *ns)
	129	{
	130	int i;
	131
	132	for (i = 0; i < PIDMAP_ENTRIES; i++)
	133	kfree(ns->pidmap[i].page);
	134	kmem_cache_free(pid_ns_cachep, ns);
	135	}
	136
	137	struct pid_namespace copy_pid_ns(unsigned long flags, struct pid_namespace old_ns)
	138	{
74bd59bb	139	if (!(flags & CLONE_NEWPID))
dca4a979	140	return get_pid_ns(old_ns);
e5a47386	141	if (flags & (CLONE_THREAD\|CLONE_PARENT))
dca4a979 AD	142	return ERR_PTR(-EINVAL);
dca4a979 AD	143	return create_pid_namespace(old_ns);
74bd59bb PE	144	}
74bd59bb PE	145
bbc2e3ef	146	static void free_pid_ns(struct kref *kref)
74bd59bb	147	{
bbc2e3ef	148	struct pid_namespace *ns;
74bd59bb PE	149
74bd59bb PE	150	ns = container_of(kref, struct pid_namespace, kref);
74bd59bb	151	destroy_pid_namespace(ns);
bbc2e3ef	152	}
74bd59bb	153
bbc2e3ef CG	154	void put_pid_ns(struct pid_namespace *ns)
	155	{
	156	struct pid_namespace *parent;
	157
	158	while (ns != &init_pid_ns) {
	159	parent = ns->parent;
	160	if (!kref_put(&ns->kref, free_pid_ns))
	161	break;
	162	ns = parent;
	163	}
74bd59bb	164	}
bbc2e3ef	165	EXPORT_SYMBOL_GPL(put_pid_ns);
74bd59bb PE	166
	167	void zap_pid_ns_processes(struct pid_namespace *pid_ns)
	168	{
	169	int nr;
	170	int rc;
00c10bc1 EB	171	struct task_struct task, me = current;
	172
	173	/* Ignore SIGCHLD causing any terminated children to autoreap */
	174	spin_lock_irq(&me->sighand->siglock);
	175	me->sighand->action[SIGCHLD - 1].sa.sa_handler = SIG_IGN;
	176	spin_unlock_irq(&me->sighand->siglock);
74bd59bb PE	177
	178	/*
	179	* The last thread in the cgroup-init thread group is terminating.
	180	* Find remaining pid_ts in the namespace, signal and wait for them
	181	* to exit.
	182	*
	183	* Note: This signals each threads in the namespace - even those that
	184	* belong to the same thread group, To avoid this, we would have
	185	* to walk the entire tasklist looking a processes in this
	186	* namespace, but that could be unnecessarily expensive if the
	187	* pid namespace has just a few processes. Or we need to
	188	* maintain a tasklist for each pid namespace.
	189	*
	190	*/
	191	read_lock(&tasklist_lock);
	192	nr = next_pidmap(pid_ns, 1);
	193	while (nr > 0) {
e4da026f SB	194	rcu_read_lock();
e4da026f SB	195
e4da026f	196	task = pid_task(find_vpid(nr), PIDTYPE_PID);
a02d6fd6 ON	197	if (task && !__fatal_signal_pending(task))
a02d6fd6 ON	198	send_sig_info(SIGKILL, SEND_SIG_FORCED, task);
e4da026f SB	199
	200	rcu_read_unlock();
	201
74bd59bb PE	202	nr = next_pidmap(pid_ns, nr);
	203	}
	204	read_unlock(&tasklist_lock);
	205
6347e900	206	/* Firstly reap the EXIT_ZOMBIE children we may have. */
74bd59bb PE	207	do {
	208	clear_thread_flag(TIF_SIGPENDING);
	209	rc = sys_wait4(-1, NULL, __WALL, NULL);
	210	} while (rc != -ECHILD);
	211
6347e900 EB	212	/*
	213	* sys_wait4() above can't reap the TASK_DEAD children.
	214	* Make sure they all go away, see __unhash_process().
	215	*/
	216	for (;;) {
	217	bool need_wait = false;
	218
	219	read_lock(&tasklist_lock);
	220	if (!list_empty(&current->children)) {
	221	__set_current_state(TASK_UNINTERRUPTIBLE);
	222	need_wait = true;
	223	}
	224	read_unlock(&tasklist_lock);
	225
	226	if (!need_wait)
	227	break;
	228	schedule();
	229	}
	230
cf3f8921 DL	231	if (pid_ns->reboot)
	232	current->signal->group_exit_code = pid_ns->reboot;
	233
0b6b030f	234	acct_exit_ns(pid_ns);
74bd59bb PE	235	return;
	236	}
	237
98ed57ee	238	#ifdef CONFIG_CHECKPOINT_RESTORE
b8f566b0 PE	239	static int pid_ns_ctl_handler(struct ctl_table *table, int write,
	240	void __user buffer, size_t lenp, loff_t *ppos)
	241	{
	242	struct ctl_table tmp = *table;
	243
	244	if (write && !capable(CAP_SYS_ADMIN))
	245	return -EPERM;
	246
	247	/*
	248	* Writing directly to ns' last_pid field is OK, since this field
	249	* is volatile in a living namespace anyway and a code writing to
	250	* it should synchronize its usage with external means.
	251	*/
	252
	253	tmp.data = &current->nsproxy->pid_ns->last_pid;
579035dc	254	return proc_dointvec_minmax(&tmp, write, buffer, lenp, ppos);
b8f566b0 PE	255	}
b8f566b0 PE	256
579035dc AV	257	extern int pid_max;
579035dc AV	258	static int zero = 0;
b8f566b0 PE	259	static struct ctl_table pid_ns_ctl_table[] = {
	260	{
	261	.procname = "ns_last_pid",
	262	.maxlen = sizeof(int),
	263	.mode = 0666, /* permissions are checked in the handler */
	264	.proc_handler = pid_ns_ctl_handler,
579035dc AV	265	.extra1 = &zero,
579035dc AV	266	.extra2 = &pid_max,
b8f566b0 PE	267	},
	268	{ }
	269	};
b8f566b0	270	static struct ctl_path kern_path[] = { { .procname = "kernel", }, { } };
98ed57ee	271	#endif /* CONFIG_CHECKPOINT_RESTORE */
b8f566b0	272
cf3f8921 DL	273	int reboot_pid_ns(struct pid_namespace *pid_ns, int cmd)
	274	{
	275	if (pid_ns == &init_pid_ns)
	276	return 0;
	277
	278	switch (cmd) {
	279	case LINUX_REBOOT_CMD_RESTART2:
	280	case LINUX_REBOOT_CMD_RESTART:
	281	pid_ns->reboot = SIGHUP;
	282	break;
	283
	284	case LINUX_REBOOT_CMD_POWER_OFF:
	285	case LINUX_REBOOT_CMD_HALT:
	286	pid_ns->reboot = SIGINT;
	287	break;
	288	default:
	289	return -EINVAL;
	290	}
	291
	292	read_lock(&tasklist_lock);
	293	force_sig(SIGKILL, pid_ns->child_reaper);
	294	read_unlock(&tasklist_lock);
	295
	296	do_exit(0);
	297
	298	/* Not reached */
	299	return 0;
	300	}
	301
74bd59bb PE	302	static __init int pid_namespaces_init(void)
	303	{
	304	pid_ns_cachep = KMEM_CACHE(pid_namespace, SLAB_PANIC);
98ed57ee CG	305
98ed57ee CG	306	#ifdef CONFIG_CHECKPOINT_RESTORE
b8f566b0	307	register_sysctl_paths(kern_path, pid_ns_ctl_table);
98ed57ee	308	#endif
74bd59bb PE	309	return 0;
	310	}
	311
	312	__initcall(pid_namespaces_init);