[deliverable/linux.git] / fs / proc / namespaces.c

#include <linux/proc_fs.h>
#include <linux/nsproxy.h>
#include <linux/sched.h>
#include <linux/ptrace.h>
#include <linux/fs_struct.h>
#include <linux/mount.h>
#include <linux/path.h>
#include <linux/namei.h>
#include <linux/file.h>
#include <linux/utsname.h>
#include <net/net_namespace.h>
#include <linux/ipc_namespace.h>
#include <linux/pid_namespace.h>
#include <linux/user_namespace.h>
#include "internal.h"


static const struct proc_ns_operations *ns_entries[] = {
#ifdef CONFIG_NET_NS
	&netns_operations,
#endif
#ifdef CONFIG_UTS_NS
	&utsns_operations,
#endif
#ifdef CONFIG_IPC_NS
	&ipcns_operations,
#endif
#ifdef CONFIG_PID_NS
	&pidns_operations,
#endif
#ifdef CONFIG_USER_NS
	&userns_operations,
#endif
	&mntns_operations,
};

static const struct file_operations ns_file_operations = {
	.llseek		= no_llseek,
};

static const struct inode_operations ns_inode_operations = {
	.setattr	= proc_setattr,
};

static int ns_delete_dentry(const struct dentry *dentry)
{
	/* Don't cache namespace inodes when not in use */
	return 1;
}

static char *ns_dname(struct dentry *dentry, char *buffer, int buflen)
{
	struct inode *inode = dentry->d_inode;
	const struct proc_ns_operations *ns_ops = PROC_I(inode)->ns.ns_ops;

	return dynamic_dname(dentry, buffer, buflen, "%s:[%lu]",
		ns_ops->name, inode->i_ino);
}

const struct dentry_operations ns_dentry_operations =
{
	.d_delete	= ns_delete_dentry,
	.d_dname	= ns_dname,
};

static struct dentry *proc_ns_get_dentry(struct super_block *sb,
	struct task_struct *task, const struct proc_ns_operations *ns_ops)
{
	struct dentry *dentry, *result;
	struct inode *inode;
	struct proc_inode *ei;
	struct qstr qname = { .name = "", };
	void *ns;

	ns = ns_ops->get(task);
	if (!ns)
		return ERR_PTR(-ENOENT);

	dentry = d_alloc_pseudo(sb, &qname);
	if (!dentry) {
		ns_ops->put(ns);
		return ERR_PTR(-ENOMEM);
	}

	inode = iget_locked(sb, ns_ops->inum(ns));
	if (!inode) {
		dput(dentry);
		ns_ops->put(ns);
		return ERR_PTR(-ENOMEM);
	}

	ei = PROC_I(inode);
	if (inode->i_state & I_NEW) {
		inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
		inode->i_op = &ns_inode_operations;
		inode->i_mode = S_IFREG | S_IRUGO;
		inode->i_fop = &ns_file_operations;
		ei->ns.ns_ops = ns_ops;
		ei->ns.ns = ns;
		unlock_new_inode(inode);
	} else {
		ns_ops->put(ns);
	}

	d_set_d_op(dentry, &ns_dentry_operations);
	result = d_instantiate_unique(dentry, inode);
	if (result) {
		dput(dentry);
		dentry = result;
	}

	return dentry;
}

static void *proc_ns_follow_link(struct dentry *dentry, struct nameidata *nd)
{
	struct inode *inode = dentry->d_inode;
	struct super_block *sb = inode->i_sb;
	struct proc_inode *ei = PROC_I(inode);
	struct task_struct *task;
	struct path ns_path;
	void *error = ERR_PTR(-EACCES);

	task = get_proc_task(inode);
	if (!task)
		goto out;

	if (!ptrace_may_access(task, PTRACE_MODE_READ))
		goto out_put_task;

	ns_path.dentry = proc_ns_get_dentry(sb, task, ei->ns.ns_ops);
	if (IS_ERR(ns_path.dentry)) {
		error = ERR_CAST(ns_path.dentry);
		goto out_put_task;
	}

	ns_path.mnt = mntget(nd->path.mnt);
	nd_jump_link(nd, &ns_path);
	error = NULL;

out_put_task:
	put_task_struct(task);
out:
	return error;
}

static int proc_ns_readlink(struct dentry *dentry, char __user *buffer, int buflen)
{
	struct inode *inode = dentry->d_inode;
	struct proc_inode *ei = PROC_I(inode);
	const struct proc_ns_operations *ns_ops = ei->ns.ns_ops;
	struct task_struct *task;
	void *ns;
	char name[50];
	int len = -EACCES;

	task = get_proc_task(inode);
	if (!task)
		goto out;

	if (!ptrace_may_access(task, PTRACE_MODE_READ))
		goto out_put_task;

	len = -ENOENT;
	ns = ns_ops->get(task);
	if (!ns)
		goto out_put_task;

	snprintf(name, sizeof(name), "%s:[%u]", ns_ops->name, ns_ops->inum(ns));
	len = strlen(name);

	if (len > buflen)
		len = buflen;
	if (copy_to_user(buffer, name, len))
		len = -EFAULT;

	ns_ops->put(ns);
out_put_task:
	put_task_struct(task);
out:
	return len;
}

static const struct inode_operations proc_ns_link_inode_operations = {
	.readlink	= proc_ns_readlink,
	.follow_link	= proc_ns_follow_link,
	.setattr	= proc_setattr,
};

static int proc_ns_instantiate(struct inode *dir,
	struct dentry *dentry, struct task_struct *task, const void *ptr)
{
	const struct proc_ns_operations *ns_ops = ptr;
	struct inode *inode;
	struct proc_inode *ei;

	inode = proc_pid_make_inode(dir->i_sb, task);
	if (!inode)
		goto out;

	ei = PROC_I(inode);
	inode->i_mode = S_IFLNK|S_IRWXUGO;
	inode->i_op = &proc_ns_link_inode_operations;
	ei->ns.ns_ops = ns_ops;

	d_set_d_op(dentry, &pid_dentry_operations);
	d_add(dentry, inode);
	/* Close the race of the process dying before we return the dentry */
	if (pid_revalidate(dentry, 0))
		return 0;
out:
	return -ENOENT;
}

static int proc_ns_dir_readdir(struct file *file, struct dir_context *ctx)
{
	struct task_struct *task = get_proc_task(file_inode(file));
	const struct proc_ns_operations **entry, **last;

	if (!task)
		return -ENOENT;

	if (!dir_emit_dots(file, ctx))
		goto out;
	if (ctx->pos >= 2 + ARRAY_SIZE(ns_entries))
		goto out;
	entry = ns_entries + (ctx->pos - 2);
	last = &ns_entries[ARRAY_SIZE(ns_entries) - 1];
	while (entry <= last) {
		const struct proc_ns_operations *ops = *entry;
		if (!proc_fill_cache(file, ctx, ops->name, strlen(ops->name),
				     proc_ns_instantiate, task, ops))
			break;
		ctx->pos++;
		entry++;
	}
out:
	put_task_struct(task);
	return 0;
}

const struct file_operations proc_ns_dir_operations = {
	.read		= generic_read_dir,
	.iterate	= proc_ns_dir_readdir,
};

static struct dentry *proc_ns_dir_lookup(struct inode *dir,
				struct dentry *dentry, unsigned int flags)
{
	int error;
	struct task_struct *task = get_proc_task(dir);
	const struct proc_ns_operations **entry, **last;
	unsigned int len = dentry->d_name.len;

	error = -ENOENT;

	if (!task)
		goto out_no_task;

	last = &ns_entries[ARRAY_SIZE(ns_entries)];
	for (entry = ns_entries; entry < last; entry++) {
		if (strlen((*entry)->name) != len)
			continue;
		if (!memcmp(dentry->d_name.name, (*entry)->name, len))
			break;
	}
	if (entry == last)
		goto out;

	error = proc_ns_instantiate(dir, dentry, task, *entry);
out:
	put_task_struct(task);
out_no_task:
	return ERR_PTR(error);
}

const struct inode_operations proc_ns_dir_inode_operations = {
	.lookup		= proc_ns_dir_lookup,
	.getattr	= pid_getattr,
	.setattr	= proc_setattr,
};

struct file *proc_ns_fget(int fd)
{
	struct file *file;

	file = fget(fd);
	if (!file)
		return ERR_PTR(-EBADF);

	if (file->f_op != &ns_file_operations)
		goto out_invalid;

	return file;

out_invalid:
	fput(file);
	return ERR_PTR(-EINVAL);
}

struct proc_ns *get_proc_ns(struct inode *inode)
{
	return &PROC_I(inode)->ns;
}

bool proc_ns_inode(struct inode *inode)
{
	return inode->i_fop == &ns_file_operations;
}
Commit	Line	Data
6b4e306a EB	1	#include <linux/proc_fs.h>
	2	#include <linux/nsproxy.h>
	3	#include <linux/sched.h>
	4	#include <linux/ptrace.h>
	5	#include <linux/fs_struct.h>
	6	#include <linux/mount.h>
	7	#include <linux/path.h>
	8	#include <linux/namei.h>
	9	#include <linux/file.h>
	10	#include <linux/utsname.h>
	11	#include <net/net_namespace.h>
6b4e306a EB	12	#include <linux/ipc_namespace.h>
6b4e306a EB	13	#include <linux/pid_namespace.h>
cde1975b	14	#include <linux/user_namespace.h>
6b4e306a EB	15	#include "internal.h"
	16
	17
	18	static const struct proc_ns_operations *ns_entries[] = {
13b6f576 EB	19	#ifdef CONFIG_NET_NS
	20	&netns_operations,
	21	#endif
34482e89 EB	22	#ifdef CONFIG_UTS_NS
	23	&utsns_operations,
	24	#endif
a00eaf11 EB	25	#ifdef CONFIG_IPC_NS
	26	&ipcns_operations,
	27	#endif
57e8391d EB	28	#ifdef CONFIG_PID_NS
57e8391d EB	29	&pidns_operations,
cde1975b EB	30	#endif
	31	#ifdef CONFIG_USER_NS
	32	&userns_operations,
57e8391d	33	#endif
8823c079	34	&mntns_operations,
6b4e306a EB	35	};
	36
	37	static const struct file_operations ns_file_operations = {
	38	.llseek = no_llseek,
	39	};
	40
bf056bfa EB	41	static const struct inode_operations ns_inode_operations = {
	42	.setattr = proc_setattr,
	43	};
	44
	45	static int ns_delete_dentry(const struct dentry *dentry)
	46	{
	47	/* Don't cache namespace inodes when not in use */
	48	return 1;
	49	}
	50
	51	static char ns_dname(struct dentry dentry, char *buffer, int buflen)
	52	{
	53	struct inode *inode = dentry->d_inode;
0bb80f24	54	const struct proc_ns_operations *ns_ops = PROC_I(inode)->ns.ns_ops;
bf056bfa EB	55
	56	return dynamic_dname(dentry, buffer, buflen, "%s:[%lu]",
	57	ns_ops->name, inode->i_ino);
	58	}
	59
	60	const struct dentry_operations ns_dentry_operations =
	61	{
	62	.d_delete = ns_delete_dentry,
	63	.d_dname = ns_dname,
	64	};
	65
	66	static struct dentry proc_ns_get_dentry(struct super_block sb,
	67	struct task_struct task, const struct proc_ns_operations ns_ops)
	68	{
	69	struct dentry dentry, result;
	70	struct inode *inode;
	71	struct proc_inode *ei;
	72	struct qstr qname = { .name = "", };
	73	void *ns;
	74
	75	ns = ns_ops->get(task);
	76	if (!ns)
	77	return ERR_PTR(-ENOENT);
	78
	79	dentry = d_alloc_pseudo(sb, &qname);
	80	if (!dentry) {
	81	ns_ops->put(ns);
	82	return ERR_PTR(-ENOMEM);
	83	}
	84
98f842e6	85	inode = iget_locked(sb, ns_ops->inum(ns));
bf056bfa EB	86	if (!inode) {
	87	dput(dentry);
	88	ns_ops->put(ns);
	89	return ERR_PTR(-ENOMEM);
	90	}
	91
	92	ei = PROC_I(inode);
98f842e6 EB	93	if (inode->i_state & I_NEW) {
	94	inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
	95	inode->i_op = &ns_inode_operations;
	96	inode->i_mode = S_IFREG \| S_IRUGO;
	97	inode->i_fop = &ns_file_operations;
0bb80f24 DH	98	ei->ns.ns_ops = ns_ops;
0bb80f24 DH	99	ei->ns.ns = ns;
98f842e6 EB	100	unlock_new_inode(inode);
	101	} else {
	102	ns_ops->put(ns);
	103	}
bf056bfa EB	104
	105	d_set_d_op(dentry, &ns_dentry_operations);
	106	result = d_instantiate_unique(dentry, inode);
	107	if (result) {
	108	dput(dentry);
	109	dentry = result;
	110	}
	111
	112	return dentry;
	113	}
	114
	115	static void proc_ns_follow_link(struct dentry dentry, struct nameidata *nd)
	116	{
	117	struct inode *inode = dentry->d_inode;
	118	struct super_block *sb = inode->i_sb;
	119	struct proc_inode *ei = PROC_I(inode);
	120	struct task_struct *task;
db04dc67	121	struct path ns_path;
bf056bfa EB	122	void *error = ERR_PTR(-EACCES);
	123
	124	task = get_proc_task(inode);
	125	if (!task)
	126	goto out;
	127
	128	if (!ptrace_may_access(task, PTRACE_MODE_READ))
	129	goto out_put_task;
	130
0bb80f24	131	ns_path.dentry = proc_ns_get_dentry(sb, task, ei->ns.ns_ops);
db04dc67 EB	132	if (IS_ERR(ns_path.dentry)) {
db04dc67 EB	133	error = ERR_CAST(ns_path.dentry);
bf056bfa EB	134	goto out_put_task;
	135	}
	136
db04dc67 EB	137	ns_path.mnt = mntget(nd->path.mnt);
db04dc67 EB	138	nd_jump_link(nd, &ns_path);
bf056bfa EB	139	error = NULL;
	140
	141	out_put_task:
	142	put_task_struct(task);
	143	out:
	144	return error;
	145	}
	146
	147	static int proc_ns_readlink(struct dentry dentry, char __user buffer, int buflen)
	148	{
	149	struct inode *inode = dentry->d_inode;
	150	struct proc_inode *ei = PROC_I(inode);
0bb80f24	151	const struct proc_ns_operations *ns_ops = ei->ns.ns_ops;
bf056bfa EB	152	struct task_struct *task;
	153	void *ns;
	154	char name[50];
	155	int len = -EACCES;
	156
	157	task = get_proc_task(inode);
	158	if (!task)
	159	goto out;
	160
	161	if (!ptrace_may_access(task, PTRACE_MODE_READ))
	162	goto out_put_task;
	163
	164	len = -ENOENT;
	165	ns = ns_ops->get(task);
	166	if (!ns)
	167	goto out_put_task;
	168
98f842e6	169	snprintf(name, sizeof(name), "%s:[%u]", ns_ops->name, ns_ops->inum(ns));
bf056bfa EB	170	len = strlen(name);
	171
	172	if (len > buflen)
	173	len = buflen;
98f842e6	174	if (copy_to_user(buffer, name, len))
bf056bfa EB	175	len = -EFAULT;
	176
	177	ns_ops->put(ns);
	178	out_put_task:
	179	put_task_struct(task);
	180	out:
	181	return len;
	182	}
	183
	184	static const struct inode_operations proc_ns_link_inode_operations = {
	185	.readlink = proc_ns_readlink,
	186	.follow_link = proc_ns_follow_link,
	187	.setattr = proc_setattr,
	188	};
	189
c52a47ac	190	static int proc_ns_instantiate(struct inode *dir,
6b4e306a EB	191	struct dentry dentry, struct task_struct task, const void *ptr)
	192	{
	193	const struct proc_ns_operations *ns_ops = ptr;
	194	struct inode *inode;
	195	struct proc_inode *ei;
6b4e306a EB	196
	197	inode = proc_pid_make_inode(dir->i_sb, task);
	198	if (!inode)
	199	goto out;
	200
	201	ei = PROC_I(inode);
bf056bfa EB	202	inode->i_mode = S_IFLNK\|S_IRWXUGO;
bf056bfa EB	203	inode->i_op = &proc_ns_link_inode_operations;
0bb80f24	204	ei->ns.ns_ops = ns_ops;
6b4e306a	205
1b26c9b3	206	d_set_d_op(dentry, &pid_dentry_operations);
6b4e306a EB	207	d_add(dentry, inode);
6b4e306a EB	208	/* Close the race of the process dying before we return the dentry */
0b728e19	209	if (pid_revalidate(dentry, 0))
c52a47ac	210	return 0;
6b4e306a	211	out:
c52a47ac	212	return -ENOENT;
6b4e306a EB	213	}
6b4e306a EB	214
f0c3b509	215	static int proc_ns_dir_readdir(struct file file, struct dir_context ctx)
6b4e306a	216	{
f0c3b509	217	struct task_struct *task = get_proc_task(file_inode(file));
6b4e306a	218	const struct proc_ns_operations entry, last;
6b4e306a	219
6b4e306a	220	if (!task)
f0c3b509	221	return -ENOENT;
6b4e306a	222
f0c3b509 AV	223	if (!dir_emit_dots(file, ctx))
	224	goto out;
	225	if (ctx->pos >= 2 + ARRAY_SIZE(ns_entries))
	226	goto out;
	227	entry = ns_entries + (ctx->pos - 2);
	228	last = &ns_entries[ARRAY_SIZE(ns_entries) - 1];
	229	while (entry <= last) {
	230	const struct proc_ns_operations ops = entry;
	231	if (!proc_fill_cache(file, ctx, ops->name, strlen(ops->name),
	232	proc_ns_instantiate, task, ops))
	233	break;
	234	ctx->pos++;
	235	entry++;
	236	}
6b4e306a EB	237	out:
6b4e306a EB	238	put_task_struct(task);
f0c3b509	239	return 0;
6b4e306a EB	240	}
	241
	242	const struct file_operations proc_ns_dir_operations = {
	243	.read = generic_read_dir,
f0c3b509	244	.iterate = proc_ns_dir_readdir,
6b4e306a EB	245	};
	246
	247	static struct dentry proc_ns_dir_lookup(struct inode dir,
00cd8dd3	248	struct dentry *dentry, unsigned int flags)
6b4e306a	249	{
c52a47ac	250	int error;
6b4e306a EB	251	struct task_struct *task = get_proc_task(dir);
	252	const struct proc_ns_operations entry, last;
	253	unsigned int len = dentry->d_name.len;
	254
c52a47ac	255	error = -ENOENT;
6b4e306a EB	256
	257	if (!task)
	258	goto out_no_task;
	259
4c619aa0 AM	260	last = &ns_entries[ARRAY_SIZE(ns_entries)];
4c619aa0 AM	261	for (entry = ns_entries; entry < last; entry++) {
6b4e306a EB	262	if (strlen((*entry)->name) != len)
	263	continue;
	264	if (!memcmp(dentry->d_name.name, (*entry)->name, len))
	265	break;
	266	}
4c619aa0	267	if (entry == last)
6b4e306a EB	268	goto out;
	269
	270	error = proc_ns_instantiate(dir, dentry, task, *entry);
	271	out:
	272	put_task_struct(task);
	273	out_no_task:
c52a47ac	274	return ERR_PTR(error);
6b4e306a EB	275	}
	276
	277	const struct inode_operations proc_ns_dir_inode_operations = {
	278	.lookup = proc_ns_dir_lookup,
	279	.getattr = pid_getattr,
	280	.setattr = proc_setattr,
	281	};
	282
	283	struct file *proc_ns_fget(int fd)
	284	{
	285	struct file *file;
	286
	287	file = fget(fd);
	288	if (!file)
	289	return ERR_PTR(-EBADF);
	290
	291	if (file->f_op != &ns_file_operations)
	292	goto out_invalid;
	293
	294	return file;
	295
	296	out_invalid:
	297	fput(file);
	298	return ERR_PTR(-EINVAL);
	299	}
	300
0bb80f24 DH	301	struct proc_ns get_proc_ns(struct inode inode)
	302	{
	303	return &PROC_I(inode)->ns;
	304	}
	305
8823c079 EB	306	bool proc_ns_inode(struct inode *inode)
	307	{
	308	return inode->i_fop == &ns_file_operations;
	309	}