[deliverable/linux.git] / fs / anon_inodes.c

/*
 *  fs/anon_inodes.c
 *
 *  Copyright (C) 2007  Davide Libenzi <davidel@xmailserver.org>
 *
 *  Thanks to Arnd Bergmann for code review and suggestions.
 *  More changes for Thomas Gleixner suggestions.
 *
 */

#include <linux/cred.h>
#include <linux/file.h>
#include <linux/poll.h>
#include <linux/sched.h>
#include <linux/init.h>
#include <linux/fs.h>
#include <linux/mount.h>
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/magic.h>
#include <linux/anon_inodes.h>

#include <asm/uaccess.h>

static struct vfsmount *anon_inode_mnt __read_mostly;
static struct inode *anon_inode_inode;
static const struct file_operations anon_inode_fops;

/*
 * anon_inodefs_dname() is called from d_path().
 */
static char *anon_inodefs_dname(struct dentry *dentry, char *buffer, int buflen)
{
	return dynamic_dname(dentry, buffer, buflen, "anon_inode:%s",
				dentry->d_name.name);
}

static const struct dentry_operations anon_inodefs_dentry_operations = {
	.d_dname	= anon_inodefs_dname,
};

/*
 * nop .set_page_dirty method so that people can use .page_mkwrite on
 * anon inodes.
 */
static int anon_set_page_dirty(struct page *page)
{
	return 0;
};

static const struct address_space_operations anon_aops = {
	.set_page_dirty = anon_set_page_dirty,
};

/*
 * A single inode exists for all anon_inode files. Contrary to pipes,
 * anon_inode inodes have no associated per-instance data, so we need
 * only allocate one of them.
 */
static struct inode *anon_inode_mkinode(struct super_block *s)
{
	struct inode *inode = new_inode_pseudo(s);

	if (!inode)
		return ERR_PTR(-ENOMEM);

	inode->i_ino = get_next_ino();
	inode->i_fop = &anon_inode_fops;

	inode->i_mapping->a_ops = &anon_aops;

	/*
	 * Mark the inode dirty from the very beginning,
	 * that way it will never be moved to the dirty
	 * list because mark_inode_dirty() will think
	 * that it already _is_ on the dirty list.
	 */
	inode->i_state = I_DIRTY;
	inode->i_mode = S_IRUSR | S_IWUSR;
	inode->i_uid = current_fsuid();
	inode->i_gid = current_fsgid();
	inode->i_flags |= S_PRIVATE;
	inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
	return inode;
}

static struct dentry *anon_inodefs_mount(struct file_system_type *fs_type,
				int flags, const char *dev_name, void *data)
{
	struct dentry *root;
	root = mount_pseudo(fs_type, "anon_inode:", NULL,
			&anon_inodefs_dentry_operations, ANON_INODE_FS_MAGIC);
	if (!IS_ERR(root)) {
		struct super_block *s = root->d_sb;
		anon_inode_inode = anon_inode_mkinode(s);
		if (IS_ERR(anon_inode_inode)) {
			dput(root);
			deactivate_locked_super(s);
			root = ERR_CAST(anon_inode_inode);
		}
	}
	return root;
}

static struct file_system_type anon_inode_fs_type = {
	.name		= "anon_inodefs",
	.mount		= anon_inodefs_mount,
	.kill_sb	= kill_anon_super,
};

/**
 * anon_inode_getfile_private - creates a new file instance by hooking it up to an
 *                      anonymous inode, and a dentry that describe the "class"
 *                      of the file
 *
 * @name:    [in]    name of the "class" of the new file
 * @fops:    [in]    file operations for the new file
 * @priv:    [in]    private data for the new file (will be file's private_data)
 * @flags:   [in]    flags
 *
 *
 * Similar to anon_inode_getfile, but each file holds a single inode.
 *
 */
struct file *anon_inode_getfile_private(const char *name,
					const struct file_operations *fops,
					void *priv, int flags)
{
	struct qstr this;
	struct path path;
	struct file *file;
	struct inode *inode;

	if (fops->owner && !try_module_get(fops->owner))
		return ERR_PTR(-ENOENT);

	inode = anon_inode_mkinode(anon_inode_mnt->mnt_sb);
	if (IS_ERR(inode)) {
		file = ERR_PTR(-ENOMEM);
		goto err_module;
	}

	/*
	 * Link the inode to a directory entry by creating a unique name
	 * using the inode sequence number.
	 */
	file = ERR_PTR(-ENOMEM);
	this.name = name;
	this.len = strlen(name);
	this.hash = 0;
	path.dentry = d_alloc_pseudo(anon_inode_mnt->mnt_sb, &this);
	if (!path.dentry)
		goto err_module;

	path.mnt = mntget(anon_inode_mnt);

	d_instantiate(path.dentry, inode);

	file = alloc_file(&path, OPEN_FMODE(flags), fops);
	if (IS_ERR(file))
		goto err_dput;

	file->f_mapping = inode->i_mapping;
	file->f_flags = flags & (O_ACCMODE | O_NONBLOCK);
	file->private_data = priv;

	return file;

err_dput:
	path_put(&path);
err_module:
	module_put(fops->owner);
	return file;
}
EXPORT_SYMBOL_GPL(anon_inode_getfile_private);

/**
 * anon_inode_getfile - creates a new file instance by hooking it up to an
 *                      anonymous inode, and a dentry that describe the "class"
 *                      of the file
 *
 * @name:    [in]    name of the "class" of the new file
 * @fops:    [in]    file operations for the new file
 * @priv:    [in]    private data for the new file (will be file's private_data)
 * @flags:   [in]    flags
 *
 * Creates a new file by hooking it on a single inode. This is useful for files
 * that do not need to have a full-fledged inode in order to operate correctly.
 * All the files created with anon_inode_getfile() will share a single inode,
 * hence saving memory and avoiding code duplication for the file/inode/dentry
 * setup.  Returns the newly created file* or an error pointer.
 */
struct file *anon_inode_getfile(const char *name,
				const struct file_operations *fops,
				void *priv, int flags)
{
	struct qstr this;
	struct path path;
	struct file *file;

	if (IS_ERR(anon_inode_inode))
		return ERR_PTR(-ENODEV);

	if (fops->owner && !try_module_get(fops->owner))
		return ERR_PTR(-ENOENT);

	/*
	 * Link the inode to a directory entry by creating a unique name
	 * using the inode sequence number.
	 */
	file = ERR_PTR(-ENOMEM);
	this.name = name;
	this.len = strlen(name);
	this.hash = 0;
	path.dentry = d_alloc_pseudo(anon_inode_mnt->mnt_sb, &this);
	if (!path.dentry)
		goto err_module;

	path.mnt = mntget(anon_inode_mnt);
	/*
	 * We know the anon_inode inode count is always greater than zero,
	 * so ihold() is safe.
	 */
	ihold(anon_inode_inode);

	d_instantiate(path.dentry, anon_inode_inode);

	file = alloc_file(&path, OPEN_FMODE(flags), fops);
	if (IS_ERR(file))
		goto err_dput;
	file->f_mapping = anon_inode_inode->i_mapping;

	file->f_flags = flags & (O_ACCMODE | O_NONBLOCK);
	file->private_data = priv;

	return file;

err_dput:
	path_put(&path);
err_module:
	module_put(fops->owner);
	return file;
}
EXPORT_SYMBOL_GPL(anon_inode_getfile);

/**
 * anon_inode_getfd - creates a new file instance by hooking it up to an
 *                    anonymous inode, and a dentry that describe the "class"
 *                    of the file
 *
 * @name:    [in]    name of the "class" of the new file
 * @fops:    [in]    file operations for the new file
 * @priv:    [in]    private data for the new file (will be file's private_data)
 * @flags:   [in]    flags
 *
 * Creates a new file by hooking it on a single inode. This is useful for files
 * that do not need to have a full-fledged inode in order to operate correctly.
 * All the files created with anon_inode_getfd() will share a single inode,
 * hence saving memory and avoiding code duplication for the file/inode/dentry
 * setup.  Returns new descriptor or an error code.
 */
int anon_inode_getfd(const char *name, const struct file_operations *fops,
		     void *priv, int flags)
{
	int error, fd;
	struct file *file;

	error = get_unused_fd_flags(flags);
	if (error < 0)
		return error;
	fd = error;

	file = anon_inode_getfile(name, fops, priv, flags);
	if (IS_ERR(file)) {
		error = PTR_ERR(file);
		goto err_put_unused_fd;
	}
	fd_install(fd, file);

	return fd;

err_put_unused_fd:
	put_unused_fd(fd);
	return error;
}
EXPORT_SYMBOL_GPL(anon_inode_getfd);

static int __init anon_inode_init(void)
{
	int error;

	error = register_filesystem(&anon_inode_fs_type);
	if (error)
		goto err_exit;
	anon_inode_mnt = kern_mount(&anon_inode_fs_type);
	if (IS_ERR(anon_inode_mnt)) {
		error = PTR_ERR(anon_inode_mnt);
		goto err_unregister_filesystem;
	}
	return 0;

err_unregister_filesystem:
	unregister_filesystem(&anon_inode_fs_type);
err_exit:
	panic(KERN_ERR "anon_inode_init() failed (%d)\n", error);
}

fs_initcall(anon_inode_init);
Commit	Line	Data
5dc8bf81 DL	1	/*
	2	* fs/anon_inodes.c
	3	*
	4	* Copyright (C) 2007 Davide Libenzi <davidel@xmailserver.org>
	5	*
	6	* Thanks to Arnd Bergmann for code review and suggestions.
	7	* More changes for Thomas Gleixner suggestions.
	8	*
	9	*/
	10
a99bbaf5	11	#include <linux/cred.h>
5dc8bf81 DL	12	#include <linux/file.h>
5dc8bf81 DL	13	#include <linux/poll.h>
a99bbaf5	14	#include <linux/sched.h>
5dc8bf81 DL	15	#include <linux/init.h>
	16	#include <linux/fs.h>
	17	#include <linux/mount.h>
	18	#include <linux/module.h>
	19	#include <linux/kernel.h>
	20	#include <linux/magic.h>
	21	#include <linux/anon_inodes.h>
	22
	23	#include <asm/uaccess.h>
	24
	25	static struct vfsmount *anon_inode_mnt __read_mostly;
	26	static struct inode *anon_inode_inode;
	27	static const struct file_operations anon_inode_fops;
	28
b9aff027 NP	29	/*
	30	* anon_inodefs_dname() is called from d_path().
	31	*/
	32	static char anon_inodefs_dname(struct dentry dentry, char *buffer, int buflen)
	33	{
	34	return dynamic_dname(dentry, buffer, buflen, "anon_inode:%s",
	35	dentry->d_name.name);
	36	}
	37
c74a1cbb AV	38	static const struct dentry_operations anon_inodefs_dentry_operations = {
	39	.d_dname = anon_inodefs_dname,
	40	};
	41
d3a9262e PZ	42	/*
	43	* nop .set_page_dirty method so that people can use .page_mkwrite on
	44	* anon inodes.
	45	*/
	46	static int anon_set_page_dirty(struct page *page)
	47	{
	48	return 0;
	49	};
	50
	51	static const struct address_space_operations anon_aops = {
	52	.set_page_dirty = anon_set_page_dirty,
	53	};
	54
ca7068c4 AV	55	/*
	56	* A single inode exists for all anon_inode files. Contrary to pipes,
	57	* anon_inode inodes have no associated per-instance data, so we need
	58	* only allocate one of them.
	59	*/
	60	static struct inode anon_inode_mkinode(struct super_block s)
	61	{
	62	struct inode *inode = new_inode_pseudo(s);
	63
	64	if (!inode)
	65	return ERR_PTR(-ENOMEM);
	66
	67	inode->i_ino = get_next_ino();
	68	inode->i_fop = &anon_inode_fops;
	69
	70	inode->i_mapping->a_ops = &anon_aops;
	71
	72	/*
	73	* Mark the inode dirty from the very beginning,
	74	* that way it will never be moved to the dirty
	75	* list because mark_inode_dirty() will think
	76	* that it already _is_ on the dirty list.
	77	*/
	78	inode->i_state = I_DIRTY;
	79	inode->i_mode = S_IRUSR \| S_IWUSR;
	80	inode->i_uid = current_fsuid();
	81	inode->i_gid = current_fsgid();
	82	inode->i_flags \|= S_PRIVATE;
	83	inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
	84	return inode;
	85	}
	86
	87	static struct dentry anon_inodefs_mount(struct file_system_type fs_type,
	88	int flags, const char dev_name, void data)
	89	{
	90	struct dentry *root;
	91	root = mount_pseudo(fs_type, "anon_inode:", NULL,
	92	&anon_inodefs_dentry_operations, ANON_INODE_FS_MAGIC);
	93	if (!IS_ERR(root)) {
	94	struct super_block *s = root->d_sb;
	95	anon_inode_inode = anon_inode_mkinode(s);
	96	if (IS_ERR(anon_inode_inode)) {
	97	dput(root);
	98	deactivate_locked_super(s);
	99	root = ERR_CAST(anon_inode_inode);
	100	}
	101	}
	102	return root;
	103	}
	104
	105	static struct file_system_type anon_inode_fs_type = {
	106	.name = "anon_inodefs",
	107	.mount = anon_inodefs_mount,
	108	.kill_sb = kill_anon_super,
	109	};
	110
55708698 GZ	111	/**
	112	* anon_inode_getfile_private - creates a new file instance by hooking it up to an
	113	* anonymous inode, and a dentry that describe the "class"
	114	* of the file
	115	*
	116	* @name: [in] name of the "class" of the new file
	117	* @fops: [in] file operations for the new file
	118	* @priv: [in] private data for the new file (will be file's private_data)
	119	* @flags: [in] flags
	120	*
	121	*
	122	* Similar to anon_inode_getfile, but each file holds a single inode.
	123	*
	124	*/
	125	struct file anon_inode_getfile_private(const char name,
	126	const struct file_operations *fops,
	127	void *priv, int flags)
	128	{
	129	struct qstr this;
	130	struct path path;
	131	struct file *file;
	132	struct inode *inode;
	133
	134	if (fops->owner && !try_module_get(fops->owner))
	135	return ERR_PTR(-ENOENT);
	136
	137	inode = anon_inode_mkinode(anon_inode_mnt->mnt_sb);
	138	if (IS_ERR(inode)) {
	139	file = ERR_PTR(-ENOMEM);
	140	goto err_module;
	141	}
	142
	143	/*
	144	* Link the inode to a directory entry by creating a unique name
	145	* using the inode sequence number.
	146	*/
	147	file = ERR_PTR(-ENOMEM);
	148	this.name = name;
	149	this.len = strlen(name);
	150	this.hash = 0;
	151	path.dentry = d_alloc_pseudo(anon_inode_mnt->mnt_sb, &this);
	152	if (!path.dentry)
	153	goto err_module;
	154
	155	path.mnt = mntget(anon_inode_mnt);
	156
	157	d_instantiate(path.dentry, inode);
	158
	159	file = alloc_file(&path, OPEN_FMODE(flags), fops);
	160	if (IS_ERR(file))
	161	goto err_dput;
	162
	163	file->f_mapping = inode->i_mapping;
	164	file->f_flags = flags & (O_ACCMODE \| O_NONBLOCK);
	165	file->private_data = priv;
	166
	167	return file;
	168
	169	err_dput:
	170	path_put(&path);
	171	err_module:
	172	module_put(fops->owner);
	173	return file;
	174	}
175	EXPORT_SYMBOL_GPL(anon_inode_getfile_private);
176
5dc8bf81	177	/**
c0d8768a NK	178	* anon_inode_getfile - creates a new file instance by hooking it up to an
	179	* anonymous inode, and a dentry that describe the "class"
	180	* of the file
5dc8bf81	181	*
5dc8bf81	182	* @name: [in] name of the "class" of the new file
7d9dbca3 UD	183	* @fops: [in] file operations for the new file
	184	* @priv: [in] private data for the new file (will be file's private_data)
	185	* @flags: [in] flags
5dc8bf81 DL	186	*
	187	* Creates a new file by hooking it on a single inode. This is useful for files
	188	* that do not need to have a full-fledged inode in order to operate correctly.
562787a5	189	* All the files created with anon_inode_getfile() will share a single inode,
5dc8bf81	190	* hence saving memory and avoiding code duplication for the file/inode/dentry
562787a5	191	* setup. Returns the newly created file* or an error pointer.
5dc8bf81	192	*/
562787a5 DL	193	struct file anon_inode_getfile(const char name,
	194	const struct file_operations *fops,
	195	void *priv, int flags)
5dc8bf81 DL	196	{
5dc8bf81 DL	197	struct qstr this;
2c48b9c4	198	struct path path;
5dc8bf81	199	struct file *file;
5dc8bf81 DL	200
5dc8bf81 DL	201	if (IS_ERR(anon_inode_inode))
562787a5	202	return ERR_PTR(-ENODEV);
5dc8bf81	203
e3a2a0d4	204	if (fops->owner && !try_module_get(fops->owner))
562787a5	205	return ERR_PTR(-ENOENT);
5dc8bf81 DL	206
	207	/*
	208	* Link the inode to a directory entry by creating a unique name
	209	* using the inode sequence number.
	210	*/
39b65252	211	file = ERR_PTR(-ENOMEM);
5dc8bf81 DL	212	this.name = name;
	213	this.len = strlen(name);
	214	this.hash = 0;
4b936885	215	path.dentry = d_alloc_pseudo(anon_inode_mnt->mnt_sb, &this);
2c48b9c4	216	if (!path.dentry)
562787a5	217	goto err_module;
96fdc72d	218
2c48b9c4	219	path.mnt = mntget(anon_inode_mnt);
96fdc72d DL	220	/*
96fdc72d DL	221	* We know the anon_inode inode count is always greater than zero,
7de9c6ee	222	* so ihold() is safe.
96fdc72d	223	*/
7de9c6ee	224	ihold(anon_inode_inode);
96fdc72d	225
2c48b9c4	226	d_instantiate(path.dentry, anon_inode_inode);
5dc8bf81	227
5300990c	228	file = alloc_file(&path, OPEN_FMODE(flags), fops);
39b65252	229	if (IS_ERR(file))
430e285e	230	goto err_dput;
96fdc72d	231	file->f_mapping = anon_inode_inode->i_mapping;
5dc8bf81	232
628ff7c1	233	file->f_flags = flags & (O_ACCMODE \| O_NONBLOCK);
5dc8bf81 DL	234	file->private_data = priv;
5dc8bf81 DL	235
562787a5 DL	236	return file;
	237
	238	err_dput:
2c48b9c4	239	path_put(&path);
562787a5 DL	240	err_module:
562787a5 DL	241	module_put(fops->owner);
39b65252	242	return file;
562787a5 DL	243	}
	244	EXPORT_SYMBOL_GPL(anon_inode_getfile);
	245
	246	/**
	247	* anon_inode_getfd - creates a new file instance by hooking it up to an
	248	* anonymous inode, and a dentry that describe the "class"
	249	* of the file
	250	*
	251	* @name: [in] name of the "class" of the new file
	252	* @fops: [in] file operations for the new file
	253	* @priv: [in] private data for the new file (will be file's private_data)
	254	* @flags: [in] flags
	255	*
	256	* Creates a new file by hooking it on a single inode. This is useful for files
	257	* that do not need to have a full-fledged inode in order to operate correctly.
	258	* All the files created with anon_inode_getfd() will share a single inode,
	259	* hence saving memory and avoiding code duplication for the file/inode/dentry
	260	* setup. Returns new descriptor or an error code.
	261	*/
	262	int anon_inode_getfd(const char name, const struct file_operations fops,
	263	void *priv, int flags)
	264	{
	265	int error, fd;
	266	struct file *file;
	267
	268	error = get_unused_fd_flags(flags);
	269	if (error < 0)
	270	return error;
	271	fd = error;
	272
	273	file = anon_inode_getfile(name, fops, priv, flags);
	274	if (IS_ERR(file)) {
	275	error = PTR_ERR(file);
	276	goto err_put_unused_fd;
	277	}
5dc8bf81 DL	278	fd_install(fd, file);
5dc8bf81 DL	279
2030a42c	280	return fd;
5dc8bf81 DL	281
	282	err_put_unused_fd:
	283	put_unused_fd(fd);
5dc8bf81 DL	284	return error;
5dc8bf81 DL	285	}
d6d28168	286	EXPORT_SYMBOL_GPL(anon_inode_getfd);
5dc8bf81	287
5dc8bf81 DL	288	static int __init anon_inode_init(void)
	289	{
	290	int error;
	291
	292	error = register_filesystem(&anon_inode_fs_type);
	293	if (error)
	294	goto err_exit;
	295	anon_inode_mnt = kern_mount(&anon_inode_fs_type);
	296	if (IS_ERR(anon_inode_mnt)) {
	297	error = PTR_ERR(anon_inode_mnt);
	298	goto err_unregister_filesystem;
	299	}
5dc8bf81 DL	300	return 0;
5dc8bf81 DL	301
5dc8bf81 DL	302	err_unregister_filesystem:
	303	unregister_filesystem(&anon_inode_fs_type);
	304	err_exit:
	305	panic(KERN_ERR "anon_inode_init() failed (%d)\n", error);
	306	}
	307
	308	fs_initcall(anon_inode_init);
	309