[linux-2.6-block.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);
	int err = 0;
	struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt };
	struct super_block *sb = mnt_path.dentry->d_sb;

	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 = 0;

	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);
	else
		seq_dentry(m, mnt->mnt_root, " \t\n\\");
	if (err)
		goto out;
	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);
	else
		mangle(m, r->mnt_devname ? r->mnt_devname : "none");
	if (err)
		goto out;
	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 = 0;

	/* device */
	if (sb->s_op->show_devname) {
		seq_puts(m, "device ");
		err = sb->s_op->show_devname(m, mnt_path.dentry);
	} 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, ' ');
		if (!err)
			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 AV	97	struct mount *r = real_mount(mnt);
	98	int err = 0;
	99	struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt };
d861c630	100	struct super_block *sb = mnt_path.dentry->d_sb;
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 };
0226f492 AV	134	int err = 0;
	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));
	138	if (sb->s_op->show_path)
a6322de6	139	err = sb->s_op->show_path(m, mnt->mnt_root);
0226f492 AV	140	else
	141	seq_dentry(m, mnt->mnt_root, " \t\n\\");
	142	if (err)
	143	goto out;
	144	seq_putc(m, ' ');
	145
	146	/* mountpoints outside of chroot jail will give SEQ_SKIP on this */
9ad4dc4f	147	err = seq_path_root(m, &mnt_path, &p->root, " \t\n\\");
0226f492 AV	148	if (err)
	149	goto out;
	150
	151	seq_puts(m, mnt->mnt_flags & MNT_READONLY ? " ro" : " rw");
	152	show_mnt_opts(m, mnt);
	153
	154	/* Tagged fields ("foo:X" or "bar") */
	155	if (IS_MNT_SHARED(r))
	156	seq_printf(m, " shared:%i", r->mnt_group_id);
	157	if (IS_MNT_SLAVE(r)) {
	158	int master = r->mnt_master->mnt_group_id;
	159	int dom = get_dominating_id(r, &p->root);
	160	seq_printf(m, " master:%i", master);
	161	if (dom && dom != master)
	162	seq_printf(m, " propagate_from:%i", dom);
	163	}
	164	if (IS_MNT_UNBINDABLE(r))
	165	seq_puts(m, " unbindable");
	166
	167	/* Filesystem specific data */
	168	seq_puts(m, " - ");
	169	show_type(m, sb);
	170	seq_putc(m, ' ');
	171	if (sb->s_op->show_devname)
d861c630	172	err = sb->s_op->show_devname(m, mnt->mnt_root);
0226f492 AV	173	else
	174	mangle(m, r->mnt_devname ? r->mnt_devname : "none");
	175	if (err)
	176	goto out;
	177	seq_puts(m, sb->s_flags & MS_RDONLY ? " ro" : " rw");
	178	err = show_sb_opts(m, sb);
	179	if (err)
	180	goto out;
	181	if (sb->s_op->show_options)
34c80b1d	182	err = sb->s_op->show_options(m, mnt->mnt_root);
0226f492 AV	183	seq_putc(m, '\n');
	184	out:
	185	return err;
	186	}
	187
	188	static int show_vfsstat(struct seq_file m, struct vfsmount mnt)
	189	{
ede1bf0d	190	struct proc_mounts *p = m->private;
0226f492 AV	191	struct mount *r = real_mount(mnt);
0226f492 AV	192	struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt };
64132379	193	struct super_block *sb = mnt_path.dentry->d_sb;
0226f492 AV	194	int err = 0;
	195
	196	/* device */
64132379	197	if (sb->s_op->show_devname) {
0226f492	198	seq_puts(m, "device ");
d861c630	199	err = sb->s_op->show_devname(m, mnt_path.dentry);
0226f492 AV	200	} else {
	201	if (r->mnt_devname) {
	202	seq_puts(m, "device ");
	203	mangle(m, r->mnt_devname);
	204	} else
	205	seq_puts(m, "no device");
	206	}
	207
	208	/* mount point */
	209	seq_puts(m, " mounted on ");
9d4d6574 DL	210	/* mountpoints outside of chroot jail will give SEQ_SKIP on this */
	211	err = seq_path_root(m, &mnt_path, &p->root, " \t\n\\");
	212	if (err)
	213	goto out;
0226f492 AV	214	seq_putc(m, ' ');
	215
	216	/* file system type */
	217	seq_puts(m, "with fstype ");
64132379	218	show_type(m, sb);
0226f492 AV	219
0226f492 AV	220	/* optional statistics */
64132379	221	if (sb->s_op->show_stats) {
0226f492 AV	222	seq_putc(m, ' ');
0226f492 AV	223	if (!err)
64132379	224	err = sb->s_op->show_stats(m, mnt_path.dentry);
0226f492 AV	225	}
	226
	227	seq_putc(m, '\n');
9d4d6574	228	out:
0226f492 AV	229	return err;
	230	}
	231
	232	static int mounts_open_common(struct inode inode, struct file file,
	233	int (show)(struct seq_file , struct vfsmount *))
	234	{
	235	struct task_struct *task = get_proc_task(inode);
	236	struct nsproxy *nsp;
	237	struct mnt_namespace *ns = NULL;
	238	struct path root;
	239	struct proc_mounts *p;
ede1bf0d	240	struct seq_file *m;
0226f492 AV	241	int ret = -EINVAL;
	242
	243	if (!task)
	244	goto err;
	245
728dba3a EB	246	task_lock(task);
728dba3a EB	247	nsp = task->nsproxy;
3d93116c	248	if (!nsp \|\| !nsp->mnt_ns) {
728dba3a	249	task_unlock(task);
0226f492 AV	250	put_task_struct(task);
	251	goto err;
	252	}
	253	ns = nsp->mnt_ns;
0226f492	254	get_mnt_ns(ns);
0226f492 AV	255	if (!task->fs) {
	256	task_unlock(task);
	257	put_task_struct(task);
	258	ret = -ENOENT;
	259	goto err_put_ns;
	260	}
	261	get_fs_root(task->fs, &root);
	262	task_unlock(task);
	263	put_task_struct(task);
	264
ede1bf0d YD	265	ret = seq_open_private(file, &mounts_op, sizeof(struct proc_mounts));
ede1bf0d YD	266	if (ret)
0226f492 AV	267	goto err_put_path;
0226f492 AV	268
ede1bf0d YD	269	m = file->private_data;
ede1bf0d YD	270	m->poll_event = ns->event;
0226f492	271
ede1bf0d	272	p = m->private;
0226f492 AV	273	p->ns = ns;
0226f492 AV	274	p->root = root;
0226f492	275	p->show = show;
c7999c36	276	p->cached_event = ~0ULL;
0226f492 AV	277
	278	return 0;
	279
0226f492 AV	280	err_put_path:
	281	path_put(&root);
	282	err_put_ns:
	283	put_mnt_ns(ns);
	284	err:
	285	return ret;
	286	}
	287
	288	static int mounts_release(struct inode inode, struct file file)
	289	{
ede1bf0d YD	290	struct seq_file *m = file->private_data;
ede1bf0d YD	291	struct proc_mounts *p = m->private;
0226f492 AV	292	path_put(&p->root);
0226f492 AV	293	put_mnt_ns(p->ns);
ede1bf0d	294	return seq_release_private(inode, file);
0226f492 AV	295	}
	296
	297	static int mounts_open(struct inode inode, struct file file)
	298	{
	299	return mounts_open_common(inode, file, show_vfsmnt);
	300	}
	301
	302	static int mountinfo_open(struct inode inode, struct file file)
	303	{
	304	return mounts_open_common(inode, file, show_mountinfo);
	305	}
	306
	307	static int mountstats_open(struct inode inode, struct file file)
	308	{
	309	return mounts_open_common(inode, file, show_vfsstat);
	310	}
	311
	312	const struct file_operations proc_mounts_operations = {
	313	.open = mounts_open,
	314	.read = seq_read,
	315	.llseek = seq_lseek,
	316	.release = mounts_release,
	317	.poll = mounts_poll,
	318	};
	319
	320	const struct file_operations proc_mountinfo_operations = {
	321	.open = mountinfo_open,
	322	.read = seq_read,
	323	.llseek = seq_lseek,
	324	.release = mounts_release,
	325	.poll = mounts_poll,
	326	};
	327
	328	const struct file_operations proc_mountstats_operations = {
	329	.open = mountstats_open,
	330	.read = seq_read,
	331	.llseek = seq_lseek,
	332	.release = mounts_release,
	333	};