[linux.git] / fs / proc_namespace.c

/*
 * fs/proc_namespace.c - handling of /proc/<pid>/{mounts,mountinfo,mountstats}
 *
 * In fact, that's a piece of procfs; it's *almost* isolated from
 * the rest of fs/proc, but has rather close relationships with
 * fs/namespace.c, thus here instead of fs/proc
 *
 */
#include <linux/mnt_namespace.h>
#include <linux/nsproxy.h>
#include <linux/security.h>
#include <linux/fs_struct.h>
#include "proc/internal.h" /* only for get_proc_task() in ->open() */

#include "pnode.h"
#include "internal.h"

static unsigned mounts_poll(struct file *file, poll_table *wait)
{
	struct seq_file *m = file->private_data;
	struct proc_mounts *p = m->private;
	struct mnt_namespace *ns = p->ns;
	unsigned res = POLLIN | POLLRDNORM;
	int event;

	poll_wait(file, &p->ns->poll, wait);

	event = ACCESS_ONCE(ns->event);
	if (m->poll_event != event) {
		m->poll_event = event;
		res |= POLLERR | POLLPRI;
	}

	return res;
}

struct proc_fs_info {
	int flag;
	const char *str;
};

static int show_sb_opts(struct seq_file *m, struct super_block *sb)
{
	static const struct proc_fs_info fs_info[] = {
		{ MS_SYNCHRONOUS, ",sync" },
		{ MS_DIRSYNC, ",dirsync" },
		{ MS_MANDLOCK, ",mand" },
		{ MS_LAZYTIME, ",lazytime" },
		{ 0, NULL }
	};
	const struct proc_fs_info *fs_infop;

	for (fs_infop = fs_info; fs_infop->flag; fs_infop++) {
		if (sb->s_flags & fs_infop->flag)
			seq_puts(m, fs_infop->str);
	}

	return security_sb_show_options(m, sb);
}

static void show_mnt_opts(struct seq_file *m, struct vfsmount *mnt)
{
	static const struct proc_fs_info mnt_info[] = {
		{ MNT_NOSUID, ",nosuid" },
		{ MNT_NODEV, ",nodev" },
		{ MNT_NOEXEC, ",noexec" },
		{ MNT_NOATIME, ",noatime" },
		{ MNT_NODIRATIME, ",nodiratime" },
		{ MNT_RELATIME, ",relatime" },
		{ 0, NULL }
	};
	const struct proc_fs_info *fs_infop;

	for (fs_infop = mnt_info; fs_infop->flag; fs_infop++) {
		if (mnt->mnt_flags & fs_infop->flag)
			seq_puts(m, fs_infop->str);
	}
}

static inline void mangle(struct seq_file *m, const char *s)
{
	seq_escape(m, s, " \t\n\\");
}

static void show_type(struct seq_file *m, struct super_block *sb)
{
	mangle(m, sb->s_type->name);
	if (sb->s_subtype && sb->s_subtype[0]) {
		seq_putc(m, '.');
		mangle(m, sb->s_subtype);
	}
}

static int show_vfsmnt(struct seq_file *m, struct vfsmount *mnt)
{
	struct proc_mounts *p = m->private;
	struct mount *r = real_mount(mnt);
	struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt };
	struct super_block *sb = mnt_path.dentry->d_sb;
	int err;

	if (sb->s_op->show_devname) {
		err = sb->s_op->show_devname(m, mnt_path.dentry);
		if (err)
			goto out;
	} else {
		mangle(m, r->mnt_devname ? r->mnt_devname : "none");
	}
	seq_putc(m, ' ');
	/* mountpoints outside of chroot jail will give SEQ_SKIP on this */
	err = seq_path_root(m, &mnt_path, &p->root, " \t\n\\");
	if (err)
		goto out;
	seq_putc(m, ' ');
	show_type(m, sb);
	seq_puts(m, __mnt_is_readonly(mnt) ? " ro" : " rw");
	err = show_sb_opts(m, sb);
	if (err)
		goto out;
	show_mnt_opts(m, mnt);
	if (sb->s_op->show_options)
		err = sb->s_op->show_options(m, mnt_path.dentry);
	seq_puts(m, " 0 0\n");
out:
	return err;
}

static int show_mountinfo(struct seq_file *m, struct vfsmount *mnt)
{
	struct proc_mounts *p = m->private;
	struct mount *r = real_mount(mnt);
	struct super_block *sb = mnt->mnt_sb;
	struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt };
	int err;

	seq_printf(m, "%i %i %u:%u ", r->mnt_id, r->mnt_parent->mnt_id,
		   MAJOR(sb->s_dev), MINOR(sb->s_dev));
	if (sb->s_op->show_path) {
		err = sb->s_op->show_path(m, mnt->mnt_root);
		if (err)
			goto out;
	} else {
		seq_dentry(m, mnt->mnt_root, " \t\n\\");
	}
	seq_putc(m, ' ');

	/* mountpoints outside of chroot jail will give SEQ_SKIP on this */
	err = seq_path_root(m, &mnt_path, &p->root, " \t\n\\");
	if (err)
		goto out;

	seq_puts(m, mnt->mnt_flags & MNT_READONLY ? " ro" : " rw");
	show_mnt_opts(m, mnt);

	/* Tagged fields ("foo:X" or "bar") */
	if (IS_MNT_SHARED(r))
		seq_printf(m, " shared:%i", r->mnt_group_id);
	if (IS_MNT_SLAVE(r)) {
		int master = r->mnt_master->mnt_group_id;
		int dom = get_dominating_id(r, &p->root);
		seq_printf(m, " master:%i", master);
		if (dom && dom != master)
			seq_printf(m, " propagate_from:%i", dom);
	}
	if (IS_MNT_UNBINDABLE(r))
		seq_puts(m, " unbindable");

	/* Filesystem specific data */
	seq_puts(m, " - ");
	show_type(m, sb);
	seq_putc(m, ' ');
	if (sb->s_op->show_devname) {
		err = sb->s_op->show_devname(m, mnt->mnt_root);
		if (err)
			goto out;
	} else {
		mangle(m, r->mnt_devname ? r->mnt_devname : "none");
	}
	seq_puts(m, sb->s_flags & MS_RDONLY ? " ro" : " rw");
	err = show_sb_opts(m, sb);
	if (err)
		goto out;
	if (sb->s_op->show_options)
		err = sb->s_op->show_options(m, mnt->mnt_root);
	seq_putc(m, '\n');
out:
	return err;
}

static int show_vfsstat(struct seq_file *m, struct vfsmount *mnt)
{
	struct proc_mounts *p = m->private;
	struct mount *r = real_mount(mnt);
	struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt };
	struct super_block *sb = mnt_path.dentry->d_sb;
	int err;

	/* device */
	if (sb->s_op->show_devname) {
		seq_puts(m, "device ");
		err = sb->s_op->show_devname(m, mnt_path.dentry);
		if (err)
			goto out;
	} else {
		if (r->mnt_devname) {
			seq_puts(m, "device ");
			mangle(m, r->mnt_devname);
		} else
			seq_puts(m, "no device");
	}

	/* mount point */
	seq_puts(m, " mounted on ");
	/* mountpoints outside of chroot jail will give SEQ_SKIP on this */
	err = seq_path_root(m, &mnt_path, &p->root, " \t\n\\");
	if (err)
		goto out;
	seq_putc(m, ' ');

	/* file system type */
	seq_puts(m, "with fstype ");
	show_type(m, sb);

	/* optional statistics */
	if (sb->s_op->show_stats) {
		seq_putc(m, ' ');
		err = sb->s_op->show_stats(m, mnt_path.dentry);
	}

	seq_putc(m, '\n');
out:
	return err;
}

static int mounts_open_common(struct inode *inode, struct file *file,
			      int (*show)(struct seq_file *, struct vfsmount *))
{
	struct task_struct *task = get_proc_task(inode);
	struct nsproxy *nsp;
	struct mnt_namespace *ns = NULL;
	struct path root;
	struct proc_mounts *p;
	struct seq_file *m;
	int ret = -EINVAL;

	if (!task)
		goto err;

	task_lock(task);
	nsp = task->nsproxy;
	if (!nsp || !nsp->mnt_ns) {
		task_unlock(task);
		put_task_struct(task);
		goto err;
	}
	ns = nsp->mnt_ns;
	get_mnt_ns(ns);
	if (!task->fs) {
		task_unlock(task);
		put_task_struct(task);
		ret = -ENOENT;
		goto err_put_ns;
	}
	get_fs_root(task->fs, &root);
	task_unlock(task);
	put_task_struct(task);

	ret = seq_open_private(file, &mounts_op, sizeof(struct proc_mounts));
	if (ret)
		goto err_put_path;

	m = file->private_data;
	m->poll_event = ns->event;

	p = m->private;
	p->ns = ns;
	p->root = root;
	p->show = show;
	p->cached_event = ~0ULL;

	return 0;

 err_put_path:
	path_put(&root);
 err_put_ns:
	put_mnt_ns(ns);
 err:
	return ret;
}

static int mounts_release(struct inode *inode, struct file *file)
{
	struct seq_file *m = file->private_data;
	struct proc_mounts *p = m->private;
	path_put(&p->root);
	put_mnt_ns(p->ns);
	return seq_release_private(inode, file);
}

static int mounts_open(struct inode *inode, struct file *file)
{
	return mounts_open_common(inode, file, show_vfsmnt);
}

static int mountinfo_open(struct inode *inode, struct file *file)
{
	return mounts_open_common(inode, file, show_mountinfo);
}

static int mountstats_open(struct inode *inode, struct file *file)
{
	return mounts_open_common(inode, file, show_vfsstat);
}

const struct file_operations proc_mounts_operations = {
	.open		= mounts_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= mounts_release,
	.poll		= mounts_poll,
};

const struct file_operations proc_mountinfo_operations = {
	.open		= mountinfo_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= mounts_release,
	.poll		= mounts_poll,
};

const struct file_operations proc_mountstats_operations = {
	.open		= mountstats_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= mounts_release,
};
Commit	Line	Data
0226f492 AV	1	/*
	2	* fs/proc_namespace.c - handling of /proc/<pid>/{mounts,mountinfo,mountstats}
	3	*
	4	* In fact, that's a piece of procfs; it's almost isolated from
	5	* the rest of fs/proc, but has rather close relationships with
	6	* fs/namespace.c, thus here instead of fs/proc
	7	*
	8	*/
	9	#include <linux/mnt_namespace.h>
	10	#include <linux/nsproxy.h>
	11	#include <linux/security.h>
	12	#include <linux/fs_struct.h>
	13	#include "proc/internal.h" /* only for get_proc_task() in ->open() */
	14
	15	#include "pnode.h"
	16	#include "internal.h"
	17
	18	static unsigned mounts_poll(struct file file, poll_table wait)
	19	{
ede1bf0d YD	20	struct seq_file *m = file->private_data;
ede1bf0d YD	21	struct proc_mounts *p = m->private;
0226f492 AV	22	struct mnt_namespace *ns = p->ns;
0226f492 AV	23	unsigned res = POLLIN \| POLLRDNORM;
aab407fc	24	int event;
0226f492 AV	25
	26	poll_wait(file, &p->ns->poll, wait);
	27
aab407fc	28	event = ACCESS_ONCE(ns->event);
ede1bf0d YD	29	if (m->poll_event != event) {
ede1bf0d YD	30	m->poll_event = event;
0226f492 AV	31	res \|= POLLERR \| POLLPRI;
0226f492 AV	32	}
0226f492 AV	33
	34	return res;
	35	}
	36
	37	struct proc_fs_info {
	38	int flag;
	39	const char *str;
	40	};
	41
	42	static int show_sb_opts(struct seq_file m, struct super_block sb)
	43	{
	44	static const struct proc_fs_info fs_info[] = {
	45	{ MS_SYNCHRONOUS, ",sync" },
	46	{ MS_DIRSYNC, ",dirsync" },
	47	{ MS_MANDLOCK, ",mand" },
0ae45f63	48	{ MS_LAZYTIME, ",lazytime" },
0226f492 AV	49	{ 0, NULL }
	50	};
	51	const struct proc_fs_info *fs_infop;
	52
	53	for (fs_infop = fs_info; fs_infop->flag; fs_infop++) {
	54	if (sb->s_flags & fs_infop->flag)
	55	seq_puts(m, fs_infop->str);
	56	}
	57
	58	return security_sb_show_options(m, sb);
	59	}
	60
	61	static void show_mnt_opts(struct seq_file m, struct vfsmount mnt)
	62	{
	63	static const struct proc_fs_info mnt_info[] = {
	64	{ MNT_NOSUID, ",nosuid" },
	65	{ MNT_NODEV, ",nodev" },
	66	{ MNT_NOEXEC, ",noexec" },
	67	{ MNT_NOATIME, ",noatime" },
	68	{ MNT_NODIRATIME, ",nodiratime" },
	69	{ MNT_RELATIME, ",relatime" },
	70	{ 0, NULL }
	71	};
	72	const struct proc_fs_info *fs_infop;
	73
	74	for (fs_infop = mnt_info; fs_infop->flag; fs_infop++) {
	75	if (mnt->mnt_flags & fs_infop->flag)
	76	seq_puts(m, fs_infop->str);
	77	}
	78	}
	79
	80	static inline void mangle(struct seq_file m, const char s)
	81	{
	82	seq_escape(m, s, " \t\n\\");
	83	}
	84
	85	static void show_type(struct seq_file m, struct super_block sb)
	86	{
	87	mangle(m, sb->s_type->name);
	88	if (sb->s_subtype && sb->s_subtype[0]) {
	89	seq_putc(m, '.');
	90	mangle(m, sb->s_subtype);
	91	}
	92	}
	93
	94	static int show_vfsmnt(struct seq_file m, struct vfsmount mnt)
	95	{
ede1bf0d	96	struct proc_mounts *p = m->private;
0226f492	97	struct mount *r = real_mount(mnt);
0226f492	98	struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt };
d861c630	99	struct super_block *sb = mnt_path.dentry->d_sb;
5d9f3c7b	100	int err;
0226f492	101
d861c630 AV	102	if (sb->s_op->show_devname) {
d861c630 AV	103	err = sb->s_op->show_devname(m, mnt_path.dentry);
0226f492 AV	104	if (err)
	105	goto out;
	106	} else {
	107	mangle(m, r->mnt_devname ? r->mnt_devname : "none");
	108	}
	109	seq_putc(m, ' ');
9d4d6574 DL	110	/* mountpoints outside of chroot jail will give SEQ_SKIP on this */
	111	err = seq_path_root(m, &mnt_path, &p->root, " \t\n\\");
	112	if (err)
	113	goto out;
0226f492	114	seq_putc(m, ' ');
d861c630	115	show_type(m, sb);
0226f492	116	seq_puts(m, __mnt_is_readonly(mnt) ? " ro" : " rw");
d861c630	117	err = show_sb_opts(m, sb);
0226f492 AV	118	if (err)
	119	goto out;
	120	show_mnt_opts(m, mnt);
d861c630	121	if (sb->s_op->show_options)
34c80b1d	122	err = sb->s_op->show_options(m, mnt_path.dentry);
0226f492 AV	123	seq_puts(m, " 0 0\n");
	124	out:
	125	return err;
	126	}
	127
	128	static int show_mountinfo(struct seq_file m, struct vfsmount mnt)
	129	{
ede1bf0d	130	struct proc_mounts *p = m->private;
0226f492 AV	131	struct mount *r = real_mount(mnt);
	132	struct super_block *sb = mnt->mnt_sb;
	133	struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt };
6ce4bca0	134	int err;
0226f492 AV	135
	136	seq_printf(m, "%i %i %u:%u ", r->mnt_id, r->mnt_parent->mnt_id,
	137	MAJOR(sb->s_dev), MINOR(sb->s_dev));
6ce4bca0	138	if (sb->s_op->show_path) {
a6322de6	139	err = sb->s_op->show_path(m, mnt->mnt_root);
6ce4bca0 DL	140	if (err)
	141	goto out;
	142	} else {
0226f492	143	seq_dentry(m, mnt->mnt_root, " \t\n\\");
6ce4bca0	144	}
0226f492 AV	145	seq_putc(m, ' ');
	146
	147	/* mountpoints outside of chroot jail will give SEQ_SKIP on this */
9ad4dc4f	148	err = seq_path_root(m, &mnt_path, &p->root, " \t\n\\");
0226f492 AV	149	if (err)
	150	goto out;
	151
	152	seq_puts(m, mnt->mnt_flags & MNT_READONLY ? " ro" : " rw");
	153	show_mnt_opts(m, mnt);
	154
	155	/* Tagged fields ("foo:X" or "bar") */
	156	if (IS_MNT_SHARED(r))
	157	seq_printf(m, " shared:%i", r->mnt_group_id);
	158	if (IS_MNT_SLAVE(r)) {
	159	int master = r->mnt_master->mnt_group_id;
	160	int dom = get_dominating_id(r, &p->root);
	161	seq_printf(m, " master:%i", master);
	162	if (dom && dom != master)
	163	seq_printf(m, " propagate_from:%i", dom);
	164	}
	165	if (IS_MNT_UNBINDABLE(r))
	166	seq_puts(m, " unbindable");
	167
	168	/* Filesystem specific data */
	169	seq_puts(m, " - ");
	170	show_type(m, sb);
	171	seq_putc(m, ' ');
6ce4bca0	172	if (sb->s_op->show_devname) {
d861c630	173	err = sb->s_op->show_devname(m, mnt->mnt_root);
6ce4bca0 DL	174	if (err)
	175	goto out;
	176	} else {
0226f492	177	mangle(m, r->mnt_devname ? r->mnt_devname : "none");
6ce4bca0	178	}
0226f492 AV	179	seq_puts(m, sb->s_flags & MS_RDONLY ? " ro" : " rw");
	180	err = show_sb_opts(m, sb);
	181	if (err)
	182	goto out;
	183	if (sb->s_op->show_options)
34c80b1d	184	err = sb->s_op->show_options(m, mnt->mnt_root);
0226f492 AV	185	seq_putc(m, '\n');
	186	out:
	187	return err;
	188	}
	189
	190	static int show_vfsstat(struct seq_file m, struct vfsmount mnt)
	191	{
ede1bf0d	192	struct proc_mounts *p = m->private;
0226f492 AV	193	struct mount *r = real_mount(mnt);
0226f492 AV	194	struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt };
64132379	195	struct super_block *sb = mnt_path.dentry->d_sb;
b896fb35	196	int err;
0226f492 AV	197
0226f492 AV	198	/* device */
64132379	199	if (sb->s_op->show_devname) {
0226f492	200	seq_puts(m, "device ");
d861c630	201	err = sb->s_op->show_devname(m, mnt_path.dentry);
5f8d498d DL	202	if (err)
5f8d498d DL	203	goto out;
0226f492 AV	204	} else {
	205	if (r->mnt_devname) {
	206	seq_puts(m, "device ");
	207	mangle(m, r->mnt_devname);
	208	} else
	209	seq_puts(m, "no device");
	210	}
	211
	212	/* mount point */
	213	seq_puts(m, " mounted on ");
9d4d6574 DL	214	/* mountpoints outside of chroot jail will give SEQ_SKIP on this */
	215	err = seq_path_root(m, &mnt_path, &p->root, " \t\n\\");
	216	if (err)
	217	goto out;
0226f492 AV	218	seq_putc(m, ' ');
	219
	220	/* file system type */
	221	seq_puts(m, "with fstype ");
64132379	222	show_type(m, sb);
0226f492 AV	223
0226f492 AV	224	/* optional statistics */
64132379	225	if (sb->s_op->show_stats) {
0226f492	226	seq_putc(m, ' ');
b896fb35	227	err = sb->s_op->show_stats(m, mnt_path.dentry);
0226f492 AV	228	}
	229
	230	seq_putc(m, '\n');
9d4d6574	231	out:
0226f492 AV	232	return err;
	233	}
	234
	235	static int mounts_open_common(struct inode inode, struct file file,
	236	int (show)(struct seq_file , struct vfsmount *))
	237	{
	238	struct task_struct *task = get_proc_task(inode);
	239	struct nsproxy *nsp;
	240	struct mnt_namespace *ns = NULL;
	241	struct path root;
	242	struct proc_mounts *p;
ede1bf0d	243	struct seq_file *m;
0226f492 AV	244	int ret = -EINVAL;
	245
	246	if (!task)
	247	goto err;
	248
728dba3a EB	249	task_lock(task);
728dba3a EB	250	nsp = task->nsproxy;
3d93116c	251	if (!nsp \|\| !nsp->mnt_ns) {
728dba3a	252	task_unlock(task);
0226f492 AV	253	put_task_struct(task);
	254	goto err;
	255	}
	256	ns = nsp->mnt_ns;
0226f492	257	get_mnt_ns(ns);
0226f492 AV	258	if (!task->fs) {
	259	task_unlock(task);
	260	put_task_struct(task);
	261	ret = -ENOENT;
	262	goto err_put_ns;
	263	}
	264	get_fs_root(task->fs, &root);
	265	task_unlock(task);
	266	put_task_struct(task);
	267
ede1bf0d YD	268	ret = seq_open_private(file, &mounts_op, sizeof(struct proc_mounts));
ede1bf0d YD	269	if (ret)
0226f492 AV	270	goto err_put_path;
0226f492 AV	271
ede1bf0d YD	272	m = file->private_data;
ede1bf0d YD	273	m->poll_event = ns->event;
0226f492	274
ede1bf0d	275	p = m->private;
0226f492 AV	276	p->ns = ns;
0226f492 AV	277	p->root = root;
0226f492	278	p->show = show;
c7999c36	279	p->cached_event = ~0ULL;
0226f492 AV	280
	281	return 0;
	282
0226f492 AV	283	err_put_path:
	284	path_put(&root);
	285	err_put_ns:
	286	put_mnt_ns(ns);
	287	err:
	288	return ret;
	289	}
	290
	291	static int mounts_release(struct inode inode, struct file file)
	292	{
ede1bf0d YD	293	struct seq_file *m = file->private_data;
ede1bf0d YD	294	struct proc_mounts *p = m->private;
0226f492 AV	295	path_put(&p->root);
0226f492 AV	296	put_mnt_ns(p->ns);
ede1bf0d	297	return seq_release_private(inode, file);
0226f492 AV	298	}
	299
	300	static int mounts_open(struct inode inode, struct file file)
	301	{
	302	return mounts_open_common(inode, file, show_vfsmnt);
	303	}
	304
	305	static int mountinfo_open(struct inode inode, struct file file)
	306	{
	307	return mounts_open_common(inode, file, show_mountinfo);
	308	}
	309
	310	static int mountstats_open(struct inode inode, struct file file)
	311	{
	312	return mounts_open_common(inode, file, show_vfsstat);
	313	}
	314
	315	const struct file_operations proc_mounts_operations = {
	316	.open = mounts_open,
	317	.read = seq_read,
	318	.llseek = seq_lseek,
	319	.release = mounts_release,
	320	.poll = mounts_poll,
	321	};
	322
	323	const struct file_operations proc_mountinfo_operations = {
	324	.open = mountinfo_open,
	325	.read = seq_read,
	326	.llseek = seq_lseek,
	327	.release = mounts_release,
	328	.poll = mounts_poll,
	329	};
	330
	331	const struct file_operations proc_mountstats_operations = {
	332	.open = mountstats_open,
	333	.read = seq_read,
	334	.llseek = seq_lseek,
	335	.release = mounts_release,
	336	};