[net-next-2.6.git] / fs / file_table.c

/*
 *  linux/fs/file_table.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 *  Copyright (C) 1997 David S. Miller (davem@caip.rutgers.edu)
 */

#include <linux/string.h>
#include <linux/slab.h>
#include <linux/file.h>
#include <linux/init.h>
#include <linux/module.h>
#include <linux/smp_lock.h>
#include <linux/fs.h>
#include <linux/security.h>
#include <linux/eventpoll.h>
#include <linux/rcupdate.h>
#include <linux/mount.h>
#include <linux/cdev.h>
#include <linux/fsnotify.h>

/* sysctl tunables... */
struct files_stat_struct files_stat = {
	.max_files = NR_FILE
};

EXPORT_SYMBOL(files_stat); /* Needed by unix.o */

/* public. Not pretty! */
 __cacheline_aligned_in_smp DEFINE_SPINLOCK(files_lock);

static DEFINE_SPINLOCK(filp_count_lock);

/* slab constructors and destructors are called from arbitrary
 * context and must be fully threaded - use a local spinlock
 * to protect files_stat.nr_files
 */
void filp_ctor(void *objp, struct kmem_cache *cachep, unsigned long cflags)
{
	if ((cflags & (SLAB_CTOR_VERIFY|SLAB_CTOR_CONSTRUCTOR)) ==
	    SLAB_CTOR_CONSTRUCTOR) {
		unsigned long flags;
		spin_lock_irqsave(&filp_count_lock, flags);
		files_stat.nr_files++;
		spin_unlock_irqrestore(&filp_count_lock, flags);
	}
}

void filp_dtor(void *objp, struct kmem_cache *cachep, unsigned long dflags)
{
	unsigned long flags;
	spin_lock_irqsave(&filp_count_lock, flags);
	files_stat.nr_files--;
	spin_unlock_irqrestore(&filp_count_lock, flags);
}

static inline void file_free_rcu(struct rcu_head *head)
{
	struct file *f =  container_of(head, struct file, f_u.fu_rcuhead);
	kmem_cache_free(filp_cachep, f);
}

static inline void file_free(struct file *f)
{
	call_rcu(&f->f_u.fu_rcuhead, file_free_rcu);
}

/* Find an unused file structure and return a pointer to it.
 * Returns NULL, if there are no more free file structures or
 * we run out of memory.
 */
struct file *get_empty_filp(void)
{
	static int old_max;
	struct file * f;

	/*
	 * Privileged users can go above max_files
	 */
	if (files_stat.nr_files >= files_stat.max_files &&
				!capable(CAP_SYS_ADMIN))
		goto over;

	f = kmem_cache_alloc(filp_cachep, GFP_KERNEL);
	if (f == NULL)
		goto fail;

	memset(f, 0, sizeof(*f));
	if (security_file_alloc(f))
		goto fail_sec;

	eventpoll_init_file(f);
	atomic_set(&f->f_count, 1);
	f->f_uid = current->fsuid;
	f->f_gid = current->fsgid;
	rwlock_init(&f->f_owner.lock);
	/* f->f_version: 0 */
	INIT_LIST_HEAD(&f->f_u.fu_list);
	return f;

over:
	/* Ran out of filps - report that */
	if (files_stat.nr_files > old_max) {
		printk(KERN_INFO "VFS: file-max limit %d reached\n",
					files_stat.max_files);
		old_max = files_stat.nr_files;
	}
	goto fail;

fail_sec:
	file_free(f);
fail:
	return NULL;
}

EXPORT_SYMBOL(get_empty_filp);

void fastcall fput(struct file *file)
{
	if (atomic_dec_and_test(&file->f_count))
		__fput(file);
}

EXPORT_SYMBOL(fput);

/* __fput is called from task context when aio completion releases the last
 * last use of a struct file *.  Do not use otherwise.
 */
void fastcall __fput(struct file *file)
{
	struct dentry *dentry = file->f_dentry;
	struct vfsmount *mnt = file->f_vfsmnt;
	struct inode *inode = dentry->d_inode;

	might_sleep();

	fsnotify_close(file);
	/*
	 * The function eventpoll_release() should be the first called
	 * in the file cleanup chain.
	 */
	eventpoll_release(file);
	locks_remove_flock(file);

	if (file->f_op && file->f_op->release)
		file->f_op->release(inode, file);
	security_file_free(file);
	if (unlikely(inode->i_cdev != NULL))
		cdev_put(inode->i_cdev);
	fops_put(file->f_op);
	if (file->f_mode & FMODE_WRITE)
		put_write_access(inode);
	file_kill(file);
	file->f_dentry = NULL;
	file->f_vfsmnt = NULL;
	file_free(file);
	dput(dentry);
	mntput(mnt);
}

struct file fastcall *fget(unsigned int fd)
{
	struct file *file;
	struct files_struct *files = current->files;

	rcu_read_lock();
	file = fcheck_files(files, fd);
	if (file) {
		if (!atomic_inc_not_zero(&file->f_count)) {
			/* File object ref couldn't be taken */
			rcu_read_unlock();
			return NULL;
		}
	}
	rcu_read_unlock();

	return file;
}

EXPORT_SYMBOL(fget);

/*
 * Lightweight file lookup - no refcnt increment if fd table isn't shared. 
 * You can use this only if it is guranteed that the current task already 
 * holds a refcnt to that file. That check has to be done at fget() only
 * and a flag is returned to be passed to the corresponding fput_light().
 * There must not be a cloning between an fget_light/fput_light pair.
 */
struct file fastcall *fget_light(unsigned int fd, int *fput_needed)
{
	struct file *file;
	struct files_struct *files = current->files;

	*fput_needed = 0;
	if (likely((atomic_read(&files->count) == 1))) {
		file = fcheck_files(files, fd);
	} else {
		rcu_read_lock();
		file = fcheck_files(files, fd);
		if (file) {
			if (atomic_inc_not_zero(&file->f_count))
				*fput_needed = 1;
			else
				/* Didn't get the reference, someone's freed */
				file = NULL;
		}
		rcu_read_unlock();
	}

	return file;
}


void put_filp(struct file *file)
{
	if (atomic_dec_and_test(&file->f_count)) {
		security_file_free(file);
		file_kill(file);
		file_free(file);
	}
}

void file_move(struct file *file, struct list_head *list)
{
	if (!list)
		return;
	file_list_lock();
	list_move(&file->f_u.fu_list, list);
	file_list_unlock();
}

void file_kill(struct file *file)
{
	if (!list_empty(&file->f_u.fu_list)) {
		file_list_lock();
		list_del_init(&file->f_u.fu_list);
		file_list_unlock();
	}
}

int fs_may_remount_ro(struct super_block *sb)
{
	struct list_head *p;

	/* Check that no files are currently opened for writing. */
	file_list_lock();
	list_for_each(p, &sb->s_files) {
		struct file *file = list_entry(p, struct file, f_u.fu_list);
		struct inode *inode = file->f_dentry->d_inode;

		/* File with pending delete? */
		if (inode->i_nlink == 0)
			goto too_bad;

		/* Writeable file? */
		if (S_ISREG(inode->i_mode) && (file->f_mode & FMODE_WRITE))
			goto too_bad;
	}
	file_list_unlock();
	return 1; /* Tis' cool bro. */
too_bad:
	file_list_unlock();
	return 0;
}

void __init files_init(unsigned long mempages)
{ 
	int n; 
	/* One file with associated inode and dcache is very roughly 1K. 
	 * Per default don't use more than 10% of our memory for files. 
	 */ 

	n = (mempages * (PAGE_SIZE / 1024)) / 10;
	files_stat.max_files = n; 
	if (files_stat.max_files < NR_FILE)
		files_stat.max_files = NR_FILE;
	files_defer_init();
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* linux/fs/file_table.c
	3	*
	4	* Copyright (C) 1991, 1992 Linus Torvalds
	5	* Copyright (C) 1997 David S. Miller (davem@caip.rutgers.edu)
	6	*/
	7
	8	#include <linux/string.h>
	9	#include <linux/slab.h>
	10	#include <linux/file.h>
	11	#include <linux/init.h>
	12	#include <linux/module.h>
	13	#include <linux/smp_lock.h>
	14	#include <linux/fs.h>
	15	#include <linux/security.h>
	16	#include <linux/eventpoll.h>
ab2af1f5	17	#include <linux/rcupdate.h>
1da177e4 LT	18	#include <linux/mount.h>
1da177e4 LT	19	#include <linux/cdev.h>
0eeca283	20	#include <linux/fsnotify.h>
1da177e4 LT	21
	22	/* sysctl tunables... */
	23	struct files_stat_struct files_stat = {
	24	.max_files = NR_FILE
	25	};
	26
	27	EXPORT_SYMBOL(files_stat); /* Needed by unix.o */
	28
	29	/* public. Not pretty! */
	30	__cacheline_aligned_in_smp DEFINE_SPINLOCK(files_lock);
	31
	32	static DEFINE_SPINLOCK(filp_count_lock);
	33
	34	/* slab constructors and destructors are called from arbitrary
	35	* context and must be fully threaded - use a local spinlock
	36	* to protect files_stat.nr_files
	37	*/
2109a2d1	38	void filp_ctor(void objp, struct kmem_cache cachep, unsigned long cflags)
1da177e4 LT	39	{
	40	if ((cflags & (SLAB_CTOR_VERIFY\|SLAB_CTOR_CONSTRUCTOR)) ==
	41	SLAB_CTOR_CONSTRUCTOR) {
	42	unsigned long flags;
	43	spin_lock_irqsave(&filp_count_lock, flags);
	44	files_stat.nr_files++;
	45	spin_unlock_irqrestore(&filp_count_lock, flags);
	46	}
	47	}
	48
2109a2d1	49	void filp_dtor(void objp, struct kmem_cache cachep, unsigned long dflags)
1da177e4 LT	50	{
	51	unsigned long flags;
	52	spin_lock_irqsave(&filp_count_lock, flags);
	53	files_stat.nr_files--;
	54	spin_unlock_irqrestore(&filp_count_lock, flags);
	55	}
	56
ab2af1f5	57	static inline void file_free_rcu(struct rcu_head *head)
1da177e4	58	{
2f512016	59	struct file *f = container_of(head, struct file, f_u.fu_rcuhead);
1da177e4 LT	60	kmem_cache_free(filp_cachep, f);
	61	}
	62
ab2af1f5 DS	63	static inline void file_free(struct file *f)
ab2af1f5 DS	64	{
2f512016	65	call_rcu(&f->f_u.fu_rcuhead, file_free_rcu);
ab2af1f5 DS	66	}
ab2af1f5 DS	67
1da177e4 LT	68	/* Find an unused file structure and return a pointer to it.
	69	* Returns NULL, if there are no more free file structures or
	70	* we run out of memory.
	71	*/
	72	struct file *get_empty_filp(void)
	73	{
af4d2ecb	74	static int old_max;
1da177e4 LT	75	struct file * f;
	76
	77	/*
	78	* Privileged users can go above max_files
	79	*/
af4d2ecb KK	80	if (files_stat.nr_files >= files_stat.max_files &&
	81	!capable(CAP_SYS_ADMIN))
	82	goto over;
	83
	84	f = kmem_cache_alloc(filp_cachep, GFP_KERNEL);
	85	if (f == NULL)
	86	goto fail;
	87
	88	memset(f, 0, sizeof(*f));
	89	if (security_file_alloc(f))
	90	goto fail_sec;
1da177e4	91
af4d2ecb KK	92	eventpoll_init_file(f);
	93	atomic_set(&f->f_count, 1);
	94	f->f_uid = current->fsuid;
	95	f->f_gid = current->fsgid;
	96	rwlock_init(&f->f_owner.lock);
	97	/* f->f_version: 0 */
2f512016	98	INIT_LIST_HEAD(&f->f_u.fu_list);
af4d2ecb KK	99	return f;
	100
	101	over:
1da177e4	102	/* Ran out of filps - report that */
af4d2ecb	103	if (files_stat.nr_files > old_max) {
1da177e4 LT	104	printk(KERN_INFO "VFS: file-max limit %d reached\n",
1da177e4 LT	105	files_stat.max_files);
af4d2ecb	106	old_max = files_stat.nr_files;
1da177e4	107	}
af4d2ecb KK	108	goto fail;
	109
	110	fail_sec:
	111	file_free(f);
1da177e4 LT	112	fail:
	113	return NULL;
	114	}
	115
	116	EXPORT_SYMBOL(get_empty_filp);
	117
	118	void fastcall fput(struct file *file)
	119	{
095975da	120	if (atomic_dec_and_test(&file->f_count))
1da177e4 LT	121	__fput(file);
	122	}
	123
	124	EXPORT_SYMBOL(fput);
	125
	126	/* __fput is called from task context when aio completion releases the last
	127	* last use of a struct file *. Do not use otherwise.
	128	*/
	129	void fastcall __fput(struct file *file)
	130	{
	131	struct dentry *dentry = file->f_dentry;
	132	struct vfsmount *mnt = file->f_vfsmnt;
	133	struct inode *inode = dentry->d_inode;
	134
	135	might_sleep();
0eeca283 RL	136
0eeca283 RL	137	fsnotify_close(file);
1da177e4 LT	138	/*
	139	* The function eventpoll_release() should be the first called
	140	* in the file cleanup chain.
	141	*/
	142	eventpoll_release(file);
	143	locks_remove_flock(file);
	144
	145	if (file->f_op && file->f_op->release)
	146	file->f_op->release(inode, file);
	147	security_file_free(file);
	148	if (unlikely(inode->i_cdev != NULL))
	149	cdev_put(inode->i_cdev);
	150	fops_put(file->f_op);
	151	if (file->f_mode & FMODE_WRITE)
	152	put_write_access(inode);
	153	file_kill(file);
	154	file->f_dentry = NULL;
	155	file->f_vfsmnt = NULL;
	156	file_free(file);
	157	dput(dentry);
	158	mntput(mnt);
	159	}
	160
	161	struct file fastcall *fget(unsigned int fd)
	162	{
	163	struct file *file;
	164	struct files_struct *files = current->files;
	165
ab2af1f5	166	rcu_read_lock();
1da177e4	167	file = fcheck_files(files, fd);
ab2af1f5	168	if (file) {
095975da	169	if (!atomic_inc_not_zero(&file->f_count)) {
ab2af1f5 DS	170	/* File object ref couldn't be taken */
	171	rcu_read_unlock();
	172	return NULL;
	173	}
	174	}
	175	rcu_read_unlock();
	176
1da177e4 LT	177	return file;
	178	}
	179
	180	EXPORT_SYMBOL(fget);
	181
	182	/*
	183	* Lightweight file lookup - no refcnt increment if fd table isn't shared.
	184	* You can use this only if it is guranteed that the current task already
	185	* holds a refcnt to that file. That check has to be done at fget() only
	186	* and a flag is returned to be passed to the corresponding fput_light().
	187	* There must not be a cloning between an fget_light/fput_light pair.
	188	*/
	189	struct file fastcall fget_light(unsigned int fd, int fput_needed)
	190	{
	191	struct file *file;
	192	struct files_struct *files = current->files;
	193
	194	*fput_needed = 0;
	195	if (likely((atomic_read(&files->count) == 1))) {
	196	file = fcheck_files(files, fd);
	197	} else {
ab2af1f5	198	rcu_read_lock();
1da177e4 LT	199	file = fcheck_files(files, fd);
1da177e4 LT	200	if (file) {
095975da	201	if (atomic_inc_not_zero(&file->f_count))
ab2af1f5 DS	202	*fput_needed = 1;
	203	else
	204	/* Didn't get the reference, someone's freed */
	205	file = NULL;
1da177e4	206	}
ab2af1f5	207	rcu_read_unlock();
1da177e4	208	}
ab2af1f5	209
1da177e4 LT	210	return file;
	211	}
	212
	213
	214	void put_filp(struct file *file)
	215	{
095975da	216	if (atomic_dec_and_test(&file->f_count)) {
1da177e4 LT	217	security_file_free(file);
	218	file_kill(file);
	219	file_free(file);
	220	}
	221	}
	222
	223	void file_move(struct file file, struct list_head list)
	224	{
	225	if (!list)
	226	return;
	227	file_list_lock();
2f512016	228	list_move(&file->f_u.fu_list, list);
1da177e4 LT	229	file_list_unlock();
	230	}
	231
	232	void file_kill(struct file *file)
	233	{
2f512016	234	if (!list_empty(&file->f_u.fu_list)) {
1da177e4	235	file_list_lock();
2f512016	236	list_del_init(&file->f_u.fu_list);
1da177e4 LT	237	file_list_unlock();
	238	}
	239	}
	240
	241	int fs_may_remount_ro(struct super_block *sb)
	242	{
	243	struct list_head *p;
	244
	245	/* Check that no files are currently opened for writing. */
	246	file_list_lock();
	247	list_for_each(p, &sb->s_files) {
2f512016	248	struct file *file = list_entry(p, struct file, f_u.fu_list);
1da177e4 LT	249	struct inode *inode = file->f_dentry->d_inode;
	250
	251	/* File with pending delete? */
	252	if (inode->i_nlink == 0)
	253	goto too_bad;
	254
	255	/* Writeable file? */
	256	if (S_ISREG(inode->i_mode) && (file->f_mode & FMODE_WRITE))
	257	goto too_bad;
	258	}
	259	file_list_unlock();
	260	return 1; /* Tis' cool bro. */
	261	too_bad:
	262	file_list_unlock();
	263	return 0;
	264	}
	265
	266	void __init files_init(unsigned long mempages)
	267	{
	268	int n;
	269	/* One file with associated inode and dcache is very roughly 1K.
	270	* Per default don't use more than 10% of our memory for files.
	271	*/
	272
	273	n = (mempages * (PAGE_SIZE / 1024)) / 10;
	274	files_stat.max_files = n;
	275	if (files_stat.max_files < NR_FILE)
	276	files_stat.max_files = NR_FILE;
ab2af1f5	277	files_defer_init();
1da177e4	278	}