[deliverable/linux.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/fdtable.h>
#include <linux/init.h>
#include <linux/module.h>
#include <linux/fs.h>
#include <linux/security.h>
#include <linux/eventpoll.h>
#include <linux/rcupdate.h>
#include <linux/mount.h>
#include <linux/capability.h>
#include <linux/cdev.h>
#include <linux/fsnotify.h>
#include <linux/sysctl.h>
#include <linux/lglock.h>
#include <linux/percpu_counter.h>
#include <linux/percpu.h>
#include <linux/hardirq.h>
#include <linux/task_work.h>
#include <linux/ima.h>

#include <linux/atomic.h>

#include "internal.h"

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

/* SLAB cache for file structures */
static struct kmem_cache *filp_cachep __read_mostly;

static struct percpu_counter nr_files __cacheline_aligned_in_smp;

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

	put_cred(f->f_cred);
	kmem_cache_free(filp_cachep, f);
}

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

/*
 * Return the total number of open files in the system
 */
static long get_nr_files(void)
{
	return percpu_counter_read_positive(&nr_files);
}

/*
 * Return the maximum number of open files in the system
 */
unsigned long get_max_files(void)
{
	return files_stat.max_files;
}
EXPORT_SYMBOL_GPL(get_max_files);

/*
 * Handle nr_files sysctl
 */
#if defined(CONFIG_SYSCTL) && defined(CONFIG_PROC_FS)
int proc_nr_files(ctl_table *table, int write,
                     void __user *buffer, size_t *lenp, loff_t *ppos)
{
	files_stat.nr_files = get_nr_files();
	return proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
}
#else
int proc_nr_files(ctl_table *table, int write,
                     void __user *buffer, size_t *lenp, loff_t *ppos)
{
	return -ENOSYS;
}
#endif

/* Find an unused file structure and return a pointer to it.
 * Returns an error pointer if some error happend e.g. we over file
 * structures limit, run out of memory or operation is not permitted.
 *
 * Be very careful using this.  You are responsible for
 * getting write access to any mount that you might assign
 * to this filp, if it is opened for write.  If this is not
 * done, you will imbalance int the mount's writer count
 * and a warning at __fput() time.
 */
struct file *get_empty_filp(void)
{
	const struct cred *cred = current_cred();
	static long old_max;
	struct file *f;
	int error;

	/*
	 * Privileged users can go above max_files
	 */
	if (get_nr_files() >= files_stat.max_files && !capable(CAP_SYS_ADMIN)) {
		/*
		 * percpu_counters are inaccurate.  Do an expensive check before
		 * we go and fail.
		 */
		if (percpu_counter_sum_positive(&nr_files) >= files_stat.max_files)
			goto over;
	}

	f = kmem_cache_zalloc(filp_cachep, GFP_KERNEL);
	if (unlikely(!f))
		return ERR_PTR(-ENOMEM);

	percpu_counter_inc(&nr_files);
	f->f_cred = get_cred(cred);
	error = security_file_alloc(f);
	if (unlikely(error)) {
		file_free(f);
		return ERR_PTR(error);
	}

	atomic_long_set(&f->f_count, 1);
	rwlock_init(&f->f_owner.lock);
	spin_lock_init(&f->f_lock);
	eventpoll_init_file(f);
	/* f->f_version: 0 */
	return f;

over:
	/* Ran out of filps - report that */
	if (get_nr_files() > old_max) {
		pr_info("VFS: file-max limit %lu reached\n", get_max_files());
		old_max = get_nr_files();
	}
	return ERR_PTR(-ENFILE);
}

/**
 * alloc_file - allocate and initialize a 'struct file'
 * @mnt: the vfsmount on which the file will reside
 * @dentry: the dentry representing the new file
 * @mode: the mode with which the new file will be opened
 * @fop: the 'struct file_operations' for the new file
 *
 * Use this instead of get_empty_filp() to get a new
 * 'struct file'.  Do so because of the same initialization
 * pitfalls reasons listed for init_file().  This is a
 * preferred interface to using init_file().
 *
 * If all the callers of init_file() are eliminated, its
 * code should be moved into this function.
 */
struct file *alloc_file(struct path *path, fmode_t mode,
		const struct file_operations *fop)
{
	struct file *file;

	file = get_empty_filp();
	if (IS_ERR(file))
		return file;

	file->f_path = *path;
	file->f_inode = path->dentry->d_inode;
	file->f_mapping = path->dentry->d_inode->i_mapping;
	file->f_mode = mode;
	file->f_op = fop;

	/*
	 * These mounts don't really matter in practice
	 * for r/o bind mounts.  They aren't userspace-
	 * visible.  We do this for consistency, and so
	 * that we can do debugging checks at __fput()
	 */
	if ((mode & FMODE_WRITE) && !special_file(path->dentry->d_inode->i_mode)) {
		file_take_write(file);
		WARN_ON(mnt_clone_write(path->mnt));
	}
	if ((mode & (FMODE_READ | FMODE_WRITE)) == FMODE_READ)
		i_readcount_inc(path->dentry->d_inode);
	return file;
}
EXPORT_SYMBOL(alloc_file);

/**
 * drop_file_write_access - give up ability to write to a file
 * @file: the file to which we will stop writing
 *
 * This is a central place which will give up the ability
 * to write to @file, along with access to write through
 * its vfsmount.
 */
static void drop_file_write_access(struct file *file)
{
	struct vfsmount *mnt = file->f_path.mnt;
	struct dentry *dentry = file->f_path.dentry;
	struct inode *inode = dentry->d_inode;

	put_write_access(inode);

	if (special_file(inode->i_mode))
		return;
	if (file_check_writeable(file) != 0)
		return;
	__mnt_drop_write(mnt);
	file_release_write(file);
}

/* the real guts of fput() - releasing the last reference to file
 */
static void __fput(struct file *file)
{
	struct dentry *dentry = file->f_path.dentry;
	struct vfsmount *mnt = file->f_path.mnt;
	struct inode *inode = file->f_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 (unlikely(file->f_flags & FASYNC)) {
		if (file->f_op->fasync)
			file->f_op->fasync(-1, file, 0);
	}
	ima_file_free(file);
	if (file->f_op->release)
		file->f_op->release(inode, file);
	security_file_free(file);
	if (unlikely(S_ISCHR(inode->i_mode) && inode->i_cdev != NULL &&
		     !(file->f_mode & FMODE_PATH))) {
		cdev_put(inode->i_cdev);
	}
	fops_put(file->f_op);
	put_pid(file->f_owner.pid);
	if ((file->f_mode & (FMODE_READ | FMODE_WRITE)) == FMODE_READ)
		i_readcount_dec(inode);
	if (file->f_mode & FMODE_WRITE)
		drop_file_write_access(file);
	file->f_path.dentry = NULL;
	file->f_path.mnt = NULL;
	file->f_inode = NULL;
	file_free(file);
	dput(dentry);
	mntput(mnt);
}

static LLIST_HEAD(delayed_fput_list);
static void delayed_fput(struct work_struct *unused)
{
	struct llist_node *node = llist_del_all(&delayed_fput_list);
	struct llist_node *next;

	for (; node; node = next) {
		next = llist_next(node);
		__fput(llist_entry(node, struct file, f_u.fu_llist));
	}
}

static void ____fput(struct callback_head *work)
{
	__fput(container_of(work, struct file, f_u.fu_rcuhead));
}

/*
 * If kernel thread really needs to have the final fput() it has done
 * to complete, call this.  The only user right now is the boot - we
 * *do* need to make sure our writes to binaries on initramfs has
 * not left us with opened struct file waiting for __fput() - execve()
 * won't work without that.  Please, don't add more callers without
 * very good reasons; in particular, never call that with locks
 * held and never call that from a thread that might need to do
 * some work on any kind of umount.
 */
void flush_delayed_fput(void)
{
	delayed_fput(NULL);
}

static DECLARE_DELAYED_WORK(delayed_fput_work, delayed_fput);

void fput(struct file *file)
{
	if (atomic_long_dec_and_test(&file->f_count)) {
		struct task_struct *task = current;

		if (likely(!in_interrupt() && !(task->flags & PF_KTHREAD))) {
			init_task_work(&file->f_u.fu_rcuhead, ____fput);
			if (!task_work_add(task, &file->f_u.fu_rcuhead, true))
				return;
			/*
			 * After this task has run exit_task_work(),
			 * task_work_add() will fail.  Fall through to delayed
			 * fput to avoid leaking *file.
			 */
		}

		if (llist_add(&file->f_u.fu_llist, &delayed_fput_list))
			schedule_delayed_work(&delayed_fput_work, 1);
	}
}

/*
 * synchronous analog of fput(); for kernel threads that might be needed
 * in some umount() (and thus can't use flush_delayed_fput() without
 * risking deadlocks), need to wait for completion of __fput() and know
 * for this specific struct file it won't involve anything that would
 * need them.  Use only if you really need it - at the very least,
 * don't blindly convert fput() by kernel thread to that.
 */
void __fput_sync(struct file *file)
{
	if (atomic_long_dec_and_test(&file->f_count)) {
		struct task_struct *task = current;
		BUG_ON(!(task->flags & PF_KTHREAD));
		__fput(file);
	}
}

EXPORT_SYMBOL(fput);

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

void __init files_init(unsigned long mempages)
{ 
	unsigned long n;

	filp_cachep = kmem_cache_create("filp", sizeof(struct file), 0,
			SLAB_HWCACHE_ALIGN | SLAB_PANIC, NULL);

	/*
	 * 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 = max_t(unsigned long, n, NR_FILE);
	files_defer_init();
	percpu_counter_init(&nr_files, 0);
}
Commit	Line	Data
	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/fdtable.h>
	12	#include <linux/init.h>
	13	#include <linux/module.h>
	14	#include <linux/fs.h>
	15	#include <linux/security.h>
	16	#include <linux/eventpoll.h>
	17	#include <linux/rcupdate.h>
	18	#include <linux/mount.h>
	19	#include <linux/capability.h>
	20	#include <linux/cdev.h>
	21	#include <linux/fsnotify.h>
	22	#include <linux/sysctl.h>
	23	#include <linux/lglock.h>
	24	#include <linux/percpu_counter.h>
	25	#include <linux/percpu.h>
	26	#include <linux/hardirq.h>
	27	#include <linux/task_work.h>
	28	#include <linux/ima.h>
	29
	30	#include <linux/atomic.h>
	31
	32	#include "internal.h"
	33
	34	/* sysctl tunables... */
	35	struct files_stat_struct files_stat = {
	36	.max_files = NR_FILE
	37	};
	38
	39	/* SLAB cache for file structures */
	40	static struct kmem_cache *filp_cachep __read_mostly;
	41
	42	static struct percpu_counter nr_files __cacheline_aligned_in_smp;
	43
	44	static void file_free_rcu(struct rcu_head *head)
	45	{
	46	struct file *f = container_of(head, struct file, f_u.fu_rcuhead);
	47
	48	put_cred(f->f_cred);
	49	kmem_cache_free(filp_cachep, f);
	50	}
	51
	52	static inline void file_free(struct file *f)
	53	{
	54	percpu_counter_dec(&nr_files);
	55	file_check_state(f);
	56	call_rcu(&f->f_u.fu_rcuhead, file_free_rcu);
	57	}
	58
	59	/*
	60	* Return the total number of open files in the system
	61	*/
	62	static long get_nr_files(void)
	63	{
	64	return percpu_counter_read_positive(&nr_files);
	65	}
	66
	67	/*
	68	* Return the maximum number of open files in the system
	69	*/
	70	unsigned long get_max_files(void)
	71	{
	72	return files_stat.max_files;
	73	}
	74	EXPORT_SYMBOL_GPL(get_max_files);
	75
	76	/*
	77	* Handle nr_files sysctl
	78	*/
	79	#if defined(CONFIG_SYSCTL) && defined(CONFIG_PROC_FS)
	80	int proc_nr_files(ctl_table *table, int write,
	81	void __user buffer, size_t lenp, loff_t *ppos)
	82	{
	83	files_stat.nr_files = get_nr_files();
	84	return proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
	85	}
	86	#else
	87	int proc_nr_files(ctl_table *table, int write,
	88	void __user buffer, size_t lenp, loff_t *ppos)
	89	{
	90	return -ENOSYS;
	91	}
	92	#endif
	93
	94	/* Find an unused file structure and return a pointer to it.
	95	* Returns an error pointer if some error happend e.g. we over file
	96	* structures limit, run out of memory or operation is not permitted.
	97	*
	98	* Be very careful using this. You are responsible for
	99	* getting write access to any mount that you might assign
	100	* to this filp, if it is opened for write. If this is not
	101	* done, you will imbalance int the mount's writer count
	102	* and a warning at __fput() time.
	103	*/
	104	struct file *get_empty_filp(void)
	105	{
	106	const struct cred *cred = current_cred();
	107	static long old_max;
	108	struct file *f;
	109	int error;
	110
	111	/*
	112	* Privileged users can go above max_files
	113	*/
	114	if (get_nr_files() >= files_stat.max_files && !capable(CAP_SYS_ADMIN)) {
	115	/*
	116	* percpu_counters are inaccurate. Do an expensive check before
	117	* we go and fail.
	118	*/
	119	if (percpu_counter_sum_positive(&nr_files) >= files_stat.max_files)
	120	goto over;
	121	}
	122
	123	f = kmem_cache_zalloc(filp_cachep, GFP_KERNEL);
	124	if (unlikely(!f))
	125	return ERR_PTR(-ENOMEM);
	126
	127	percpu_counter_inc(&nr_files);
	128	f->f_cred = get_cred(cred);
	129	error = security_file_alloc(f);
	130	if (unlikely(error)) {
	131	file_free(f);
	132	return ERR_PTR(error);
	133	}
	134
	135	atomic_long_set(&f->f_count, 1);
	136	rwlock_init(&f->f_owner.lock);
	137	spin_lock_init(&f->f_lock);
	138	eventpoll_init_file(f);
	139	/* f->f_version: 0 */
	140	return f;
	141
	142	over:
	143	/* Ran out of filps - report that */
	144	if (get_nr_files() > old_max) {
	145	pr_info("VFS: file-max limit %lu reached\n", get_max_files());
	146	old_max = get_nr_files();
	147	}
	148	return ERR_PTR(-ENFILE);
	149	}
	150
	151	/**
	152	* alloc_file - allocate and initialize a 'struct file'
	153	* @mnt: the vfsmount on which the file will reside
	154	* @dentry: the dentry representing the new file
	155	* @mode: the mode with which the new file will be opened
	156	* @fop: the 'struct file_operations' for the new file
	157	*
	158	* Use this instead of get_empty_filp() to get a new
	159	* 'struct file'. Do so because of the same initialization
	160	* pitfalls reasons listed for init_file(). This is a
	161	* preferred interface to using init_file().
	162	*
	163	* If all the callers of init_file() are eliminated, its
	164	* code should be moved into this function.
	165	*/
	166	struct file alloc_file(struct path path, fmode_t mode,
	167	const struct file_operations *fop)
	168	{
	169	struct file *file;
	170
	171	file = get_empty_filp();
	172	if (IS_ERR(file))
	173	return file;
	174
	175	file->f_path = *path;
	176	file->f_inode = path->dentry->d_inode;
	177	file->f_mapping = path->dentry->d_inode->i_mapping;
	178	file->f_mode = mode;
	179	file->f_op = fop;
	180
	181	/*
	182	* These mounts don't really matter in practice
	183	* for r/o bind mounts. They aren't userspace-
	184	* visible. We do this for consistency, and so
	185	* that we can do debugging checks at __fput()
	186	*/
	187	if ((mode & FMODE_WRITE) && !special_file(path->dentry->d_inode->i_mode)) {
	188	file_take_write(file);
	189	WARN_ON(mnt_clone_write(path->mnt));
	190	}
	191	if ((mode & (FMODE_READ \| FMODE_WRITE)) == FMODE_READ)
	192	i_readcount_inc(path->dentry->d_inode);
	193	return file;
	194	}
	195	EXPORT_SYMBOL(alloc_file);
	196
	197	/**
	198	* drop_file_write_access - give up ability to write to a file
	199	* @file: the file to which we will stop writing
	200	*
	201	* This is a central place which will give up the ability
	202	* to write to @file, along with access to write through
	203	* its vfsmount.
	204	*/
	205	static void drop_file_write_access(struct file *file)
	206	{
	207	struct vfsmount *mnt = file->f_path.mnt;
	208	struct dentry *dentry = file->f_path.dentry;
	209	struct inode *inode = dentry->d_inode;
	210
	211	put_write_access(inode);
	212
	213	if (special_file(inode->i_mode))
	214	return;
	215	if (file_check_writeable(file) != 0)
	216	return;
	217	__mnt_drop_write(mnt);
	218	file_release_write(file);
	219	}
	220
	221	/* the real guts of fput() - releasing the last reference to file
	222	*/
	223	static void __fput(struct file *file)
	224	{
	225	struct dentry *dentry = file->f_path.dentry;
	226	struct vfsmount *mnt = file->f_path.mnt;
	227	struct inode *inode = file->f_inode;
	228
	229	might_sleep();
	230
	231	fsnotify_close(file);
	232	/*
	233	* The function eventpoll_release() should be the first called
	234	* in the file cleanup chain.
	235	*/
	236	eventpoll_release(file);
	237	locks_remove_flock(file);
	238
	239	if (unlikely(file->f_flags & FASYNC)) {
	240	if (file->f_op->fasync)
	241	file->f_op->fasync(-1, file, 0);
	242	}
	243	ima_file_free(file);
	244	if (file->f_op->release)
	245	file->f_op->release(inode, file);
	246	security_file_free(file);
	247	if (unlikely(S_ISCHR(inode->i_mode) && inode->i_cdev != NULL &&
	248	!(file->f_mode & FMODE_PATH))) {
	249	cdev_put(inode->i_cdev);
	250	}
	251	fops_put(file->f_op);
	252	put_pid(file->f_owner.pid);
	253	if ((file->f_mode & (FMODE_READ \| FMODE_WRITE)) == FMODE_READ)
	254	i_readcount_dec(inode);
	255	if (file->f_mode & FMODE_WRITE)
	256	drop_file_write_access(file);
	257	file->f_path.dentry = NULL;
	258	file->f_path.mnt = NULL;
	259	file->f_inode = NULL;
	260	file_free(file);
	261	dput(dentry);
	262	mntput(mnt);
	263	}
	264
	265	static LLIST_HEAD(delayed_fput_list);
	266	static void delayed_fput(struct work_struct *unused)
	267	{
	268	struct llist_node *node = llist_del_all(&delayed_fput_list);
	269	struct llist_node *next;
	270
	271	for (; node; node = next) {
	272	next = llist_next(node);
	273	__fput(llist_entry(node, struct file, f_u.fu_llist));
	274	}
	275	}
	276
	277	static void ____fput(struct callback_head *work)
	278	{
	279	__fput(container_of(work, struct file, f_u.fu_rcuhead));
	280	}
	281
	282	/*
	283	* If kernel thread really needs to have the final fput() it has done
	284	* to complete, call this. The only user right now is the boot - we
	285	* do need to make sure our writes to binaries on initramfs has
	286	* not left us with opened struct file waiting for __fput() - execve()
	287	* won't work without that. Please, don't add more callers without
	288	* very good reasons; in particular, never call that with locks
	289	* held and never call that from a thread that might need to do
	290	* some work on any kind of umount.
	291	*/
	292	void flush_delayed_fput(void)
	293	{
	294	delayed_fput(NULL);
	295	}
	296
	297	static DECLARE_DELAYED_WORK(delayed_fput_work, delayed_fput);
	298
	299	void fput(struct file *file)
	300	{
	301	if (atomic_long_dec_and_test(&file->f_count)) {
	302	struct task_struct *task = current;
	303
	304	if (likely(!in_interrupt() && !(task->flags & PF_KTHREAD))) {
	305	init_task_work(&file->f_u.fu_rcuhead, ____fput);
	306	if (!task_work_add(task, &file->f_u.fu_rcuhead, true))
	307	return;
	308	/*
	309	* After this task has run exit_task_work(),
	310	* task_work_add() will fail. Fall through to delayed
	311	* fput to avoid leaking *file.
	312	*/
	313	}
	314
	315	if (llist_add(&file->f_u.fu_llist, &delayed_fput_list))
	316	schedule_delayed_work(&delayed_fput_work, 1);
	317	}
	318	}
	319
	320	/*
	321	* synchronous analog of fput(); for kernel threads that might be needed
	322	* in some umount() (and thus can't use flush_delayed_fput() without
	323	* risking deadlocks), need to wait for completion of __fput() and know
	324	* for this specific struct file it won't involve anything that would
	325	* need them. Use only if you really need it - at the very least,
	326	* don't blindly convert fput() by kernel thread to that.
	327	*/
	328	void __fput_sync(struct file *file)
	329	{
	330	if (atomic_long_dec_and_test(&file->f_count)) {
	331	struct task_struct *task = current;
	332	BUG_ON(!(task->flags & PF_KTHREAD));
	333	__fput(file);
	334	}
	335	}
	336
	337	EXPORT_SYMBOL(fput);
	338
	339	void put_filp(struct file *file)
	340	{
	341	if (atomic_long_dec_and_test(&file->f_count)) {
	342	security_file_free(file);
	343	file_free(file);
	344	}
	345	}
	346
	347	void __init files_init(unsigned long mempages)
	348	{
	349	unsigned long n;
	350
	351	filp_cachep = kmem_cache_create("filp", sizeof(struct file), 0,
	352	SLAB_HWCACHE_ALIGN \| SLAB_PANIC, NULL);
	353
	354	/*
	355	* One file with associated inode and dcache is very roughly 1K.
	356	* Per default don't use more than 10% of our memory for files.
	357	*/
	358
	359	n = (mempages * (PAGE_SIZE / 1024)) / 10;
	360	files_stat.max_files = max_t(unsigned long, n, NR_FILE);
	361	files_defer_init();
	362	percpu_counter_init(&nr_files, 0);
	363	}