[thirdparty/kernel/stable.git] / fs / file_table.c

// SPDX-License-Identifier: GPL-2.0-only
/*
 *  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/cred.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/percpu_counter.h>
#include <linux/percpu.h>
#include <linux/task_work.h>
#include <linux/ima.h>
#include <linux/swap.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)
{
	security_file_free(f);
	if (!(f->f_mode & FMODE_NOACCOUNT))
		percpu_counter_dec(&nr_files);
	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(struct 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(struct ctl_table *table, int write,
                     void __user *buffer, size_t *lenp, loff_t *ppos)
{
	return -ENOSYS;
}
#endif

static struct file *__alloc_file(int flags, const struct cred *cred)
{
	struct file *f;
	int error;

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

	f->f_cred = get_cred(cred);
	error = security_file_alloc(f);
	if (unlikely(error)) {
		file_free_rcu(&f->f_u.fu_rcuhead);
		return ERR_PTR(error);
	}

	atomic_long_set(&f->f_count, 1);
	rwlock_init(&f->f_owner.lock);
	spin_lock_init(&f->f_lock);
	mutex_init(&f->f_pos_lock);
	eventpoll_init_file(f);
	f->f_flags = flags;
	f->f_mode = OPEN_FMODE(flags);
	/* f->f_version: 0 */

	return f;
}

/* 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 *alloc_empty_file(int flags, const struct cred *cred)
{
	static long old_max;
	struct file *f;

	/*
	 * 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 = __alloc_file(flags, cred);
	if (!IS_ERR(f))
		percpu_counter_inc(&nr_files);

	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);
}

/*
 * Variant of alloc_empty_file() that doesn't check and modify nr_files.
 *
 * Should not be used unless there's a very good reason to do so.
 */
struct file *alloc_empty_file_noaccount(int flags, const struct cred *cred)
{
	struct file *f = __alloc_file(flags, cred);

	if (!IS_ERR(f))
		f->f_mode |= FMODE_NOACCOUNT;

	return f;
}

/**
 * alloc_file - allocate and initialize a 'struct file'
 *
 * @path: the (dentry, vfsmount) pair for the new file
 * @flags: O_... flags with which the new file will be opened
 * @fop: the 'struct file_operations' for the new file
 */
static struct file *alloc_file(const struct path *path, int flags,
		const struct file_operations *fop)
{
	struct file *file;

	file = alloc_empty_file(flags, current_cred());
	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_wb_err = filemap_sample_wb_err(file->f_mapping);
	if ((file->f_mode & FMODE_READ) &&
	     likely(fop->read || fop->read_iter))
		file->f_mode |= FMODE_CAN_READ;
	if ((file->f_mode & FMODE_WRITE) &&
	     likely(fop->write || fop->write_iter))
		file->f_mode |= FMODE_CAN_WRITE;
	file->f_mode |= FMODE_OPENED;
	file->f_op = fop;
	if ((file->f_mode & (FMODE_READ | FMODE_WRITE)) == FMODE_READ)
		i_readcount_inc(path->dentry->d_inode);
	return file;
}

struct file *alloc_file_pseudo(struct inode *inode, struct vfsmount *mnt,
				const char *name, int flags,
				const struct file_operations *fops)
{
	static const struct dentry_operations anon_ops = {
		.d_dname = simple_dname
	};
	struct qstr this = QSTR_INIT(name, strlen(name));
	struct path path;
	struct file *file;

	path.dentry = d_alloc_pseudo(mnt->mnt_sb, &this);
	if (!path.dentry)
		return ERR_PTR(-ENOMEM);
	if (!mnt->mnt_sb->s_d_op)
		d_set_d_op(path.dentry, &anon_ops);
	path.mnt = mntget(mnt);
	d_instantiate(path.dentry, inode);
	file = alloc_file(&path, flags, fops);
	if (IS_ERR(file)) {
		ihold(inode);
		path_put(&path);
	}
	return file;
}
EXPORT_SYMBOL(alloc_file_pseudo);

struct file *alloc_file_clone(struct file *base, int flags,
				const struct file_operations *fops)
{
	struct file *f = alloc_file(&base->f_path, flags, fops);
	if (!IS_ERR(f)) {
		path_get(&f->f_path);
		f->f_mapping = base->f_mapping;
	}
	return f;
}

/* 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;
	fmode_t mode = file->f_mode;

	if (unlikely(!(file->f_mode & FMODE_OPENED)))
		goto out;

	might_sleep();

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

	ima_file_free(file);
	if (unlikely(file->f_flags & FASYNC)) {
		if (file->f_op->fasync)
			file->f_op->fasync(-1, file, 0);
	}
	if (file->f_op->release)
		file->f_op->release(inode, file);
	if (unlikely(S_ISCHR(inode->i_mode) && inode->i_cdev != NULL &&
		     !(mode & FMODE_PATH))) {
		cdev_put(inode->i_cdev);
	}
	fops_put(file->f_op);
	put_pid(file->f_owner.pid);
	if ((mode & (FMODE_READ | FMODE_WRITE)) == FMODE_READ)
		i_readcount_dec(inode);
	if (mode & FMODE_WRITER) {
		put_write_access(inode);
		__mnt_drop_write(mnt);
	}
	dput(dentry);
	if (unlikely(mode & FMODE_NEED_UNMOUNT))
		dissolve_on_fput(mnt);
	mntput(mnt);
out:
	file_free(file);
}

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 file *f, *t;

	llist_for_each_entry_safe(f, t, node, f_u.fu_llist)
		__fput(f);
}

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_many(struct file *file, unsigned int refs)
{
	if (atomic_long_sub_and_test(refs, &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);
	}
}

void fput(struct file *file)
{
	fput_many(file, 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 __init files_init(void)
{
	filp_cachep = kmem_cache_create("filp", sizeof(struct file), 0,
			SLAB_HWCACHE_ALIGN | SLAB_PANIC | SLAB_ACCOUNT, NULL);
	percpu_counter_init(&nr_files, 0, GFP_KERNEL);
}

/*
 * One file with associated inode and dcache is very roughly 1K. Per default
 * do not use more than 10% of our memory for files.
 */
void __init files_maxfiles_init(void)
{
	unsigned long n;
	unsigned long nr_pages = totalram_pages();
	unsigned long memreserve = (nr_pages - nr_free_pages()) * 3/2;

	memreserve = min(memreserve, nr_pages - 1);
	n = ((nr_pages - memreserve) * (PAGE_SIZE / 1024)) / 10;

	files_stat.max_files = max_t(unsigned long, n, NR_FILE);
}
Commit	Line	Data
457c8996	1	// SPDX-License-Identifier: GPL-2.0-only
1da177e4 LT	2	/*
	3	* linux/fs/file_table.c
	4	*
	5	* Copyright (C) 1991, 1992 Linus Torvalds
	6	* Copyright (C) 1997 David S. Miller (davem@caip.rutgers.edu)
	7	*/
	8
	9	#include <linux/string.h>
	10	#include <linux/slab.h>
	11	#include <linux/file.h>
9f3acc31	12	#include <linux/fdtable.h>
1da177e4 LT	13	#include <linux/init.h>
1da177e4 LT	14	#include <linux/module.h>
1da177e4 LT	15	#include <linux/fs.h>
1da177e4 LT	16	#include <linux/security.h>
5b825c3a	17	#include <linux/cred.h>
1da177e4	18	#include <linux/eventpoll.h>
ab2af1f5	19	#include <linux/rcupdate.h>
1da177e4	20	#include <linux/mount.h>
16f7e0fe	21	#include <linux/capability.h>
1da177e4	22	#include <linux/cdev.h>
0eeca283	23	#include <linux/fsnotify.h>
529bf6be DS	24	#include <linux/sysctl.h>
529bf6be DS	25	#include <linux/percpu_counter.h>
6416ccb7	26	#include <linux/percpu.h>
4a9d4b02	27	#include <linux/task_work.h>
0552f879	28	#include <linux/ima.h>
4248b0da	29	#include <linux/swap.h>
529bf6be	30
60063497	31	#include <linux/atomic.h>
1da177e4	32
e81e3f4d EP	33	#include "internal.h"
e81e3f4d EP	34
1da177e4 LT	35	/* sysctl tunables... */
	36	struct files_stat_struct files_stat = {
	37	.max_files = NR_FILE
	38	};
	39
b6b3fdea ED	40	/* SLAB cache for file structures */
	41	static struct kmem_cache *filp_cachep __read_mostly;
	42
529bf6be	43	static struct percpu_counter nr_files __cacheline_aligned_in_smp;
1da177e4	44
5c33b183	45	static void file_free_rcu(struct rcu_head *head)
1da177e4	46	{
d76b0d9b DH	47	struct file *f = container_of(head, struct file, f_u.fu_rcuhead);
	48
	49	put_cred(f->f_cred);
529bf6be	50	kmem_cache_free(filp_cachep, f);
1da177e4 LT	51	}
1da177e4 LT	52
529bf6be	53	static inline void file_free(struct file *f)
1da177e4	54	{
e8cff84f	55	security_file_free(f);
d3b1084d MS	56	if (!(f->f_mode & FMODE_NOACCOUNT))
d3b1084d MS	57	percpu_counter_dec(&nr_files);
529bf6be	58	call_rcu(&f->f_u.fu_rcuhead, file_free_rcu);
1da177e4 LT	59	}
1da177e4 LT	60
529bf6be DS	61	/*
	62	* Return the total number of open files in the system
	63	*/
518de9b3	64	static long get_nr_files(void)
1da177e4	65	{
529bf6be	66	return percpu_counter_read_positive(&nr_files);
1da177e4 LT	67	}
1da177e4 LT	68
529bf6be DS	69	/*
	70	* Return the maximum number of open files in the system
	71	*/
518de9b3	72	unsigned long get_max_files(void)
ab2af1f5	73	{
529bf6be	74	return files_stat.max_files;
ab2af1f5	75	}
529bf6be DS	76	EXPORT_SYMBOL_GPL(get_max_files);
	77
	78	/*
	79	* Handle nr_files sysctl
	80	*/
	81	#if defined(CONFIG_SYSCTL) && defined(CONFIG_PROC_FS)
1f7e0616	82	int proc_nr_files(struct ctl_table *table, int write,
529bf6be DS	83	void __user buffer, size_t lenp, loff_t *ppos)
	84	{
	85	files_stat.nr_files = get_nr_files();
518de9b3	86	return proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
529bf6be DS	87	}
529bf6be DS	88	#else
1f7e0616	89	int proc_nr_files(struct ctl_table *table, int write,
529bf6be DS	90	void __user buffer, size_t lenp, loff_t *ppos)
	91	{
	92	return -ENOSYS;
	93	}
	94	#endif
ab2af1f5	95
d3b1084d MS	96	static struct file __alloc_file(int flags, const struct cred cred)
	97	{
	98	struct file *f;
	99	int error;
	100
	101	f = kmem_cache_zalloc(filp_cachep, GFP_KERNEL);
	102	if (unlikely(!f))
	103	return ERR_PTR(-ENOMEM);
	104
	105	f->f_cred = get_cred(cred);
	106	error = security_file_alloc(f);
	107	if (unlikely(error)) {
	108	file_free_rcu(&f->f_u.fu_rcuhead);
	109	return ERR_PTR(error);
	110	}
	111
	112	atomic_long_set(&f->f_count, 1);
	113	rwlock_init(&f->f_owner.lock);
	114	spin_lock_init(&f->f_lock);
	115	mutex_init(&f->f_pos_lock);
	116	eventpoll_init_file(f);
	117	f->f_flags = flags;
	118	f->f_mode = OPEN_FMODE(flags);
	119	/* f->f_version: 0 */
	120
	121	return f;
	122	}
	123
1da177e4	124	/* Find an unused file structure and return a pointer to it.
1afc99be AV	125	* Returns an error pointer if some error happend e.g. we over file
1afc99be AV	126	* structures limit, run out of memory or operation is not permitted.
430e285e DH	127	*
	128	* Be very careful using this. You are responsible for
	129	* getting write access to any mount that you might assign
	130	* to this filp, if it is opened for write. If this is not
	131	* done, you will imbalance int the mount's writer count
	132	* and a warning at __fput() time.
1da177e4	133	*/
ea73ea72	134	struct file alloc_empty_file(int flags, const struct cred cred)
1da177e4	135	{
518de9b3	136	static long old_max;
1afc99be	137	struct file *f;
1da177e4 LT	138
	139	/*
	140	* Privileged users can go above max_files
	141	*/
529bf6be DS	142	if (get_nr_files() >= files_stat.max_files && !capable(CAP_SYS_ADMIN)) {
	143	/*
	144	* percpu_counters are inaccurate. Do an expensive check before
	145	* we go and fail.
	146	*/
52d9f3b4	147	if (percpu_counter_sum_positive(&nr_files) >= files_stat.max_files)
529bf6be DS	148	goto over;
529bf6be DS	149	}
af4d2ecb	150
d3b1084d MS	151	f = __alloc_file(flags, cred);
	152	if (!IS_ERR(f))
	153	percpu_counter_inc(&nr_files);
1da177e4	154
af4d2ecb KK	155	return f;
	156
	157	over:
1da177e4	158	/* Ran out of filps - report that */
529bf6be	159	if (get_nr_files() > old_max) {
518de9b3	160	pr_info("VFS: file-max limit %lu reached\n", get_max_files());
529bf6be	161	old_max = get_nr_files();
1da177e4	162	}
1afc99be	163	return ERR_PTR(-ENFILE);
1da177e4 LT	164	}
1da177e4 LT	165
d3b1084d MS	166	/*
	167	* Variant of alloc_empty_file() that doesn't check and modify nr_files.
	168	*
	169	* Should not be used unless there's a very good reason to do so.
	170	*/
	171	struct file alloc_empty_file_noaccount(int flags, const struct cred cred)
	172	{
	173	struct file *f = __alloc_file(flags, cred);
	174
	175	if (!IS_ERR(f))
	176	f->f_mode \|= FMODE_NOACCOUNT;
	177
	178	return f;
	179	}
	180
ce8d2cdf DH	181	/**
ce8d2cdf DH	182	* alloc_file - allocate and initialize a 'struct file'
a457606a EB	183	*
a457606a EB	184	* @path: the (dentry, vfsmount) pair for the new file
c9c554f2	185	* @flags: O_... flags with which the new file will be opened
ce8d2cdf	186	* @fop: the 'struct file_operations' for the new file
ce8d2cdf	187	*/
ee1904ba	188	static struct file alloc_file(const struct path path, int flags,
2c48b9c4	189	const struct file_operations *fop)
ce8d2cdf DH	190	{
ce8d2cdf DH	191	struct file *file;
ce8d2cdf	192
ea73ea72	193	file = alloc_empty_file(flags, current_cred());
1afc99be	194	if (IS_ERR(file))
39b65252	195	return file;
ce8d2cdf	196
2c48b9c4	197	file->f_path = *path;
dd37978c	198	file->f_inode = path->dentry->d_inode;
2c48b9c4	199	file->f_mapping = path->dentry->d_inode->i_mapping;
5660e13d	200	file->f_wb_err = filemap_sample_wb_err(file->f_mapping);
c9c554f2	201	if ((file->f_mode & FMODE_READ) &&
84363182	202	likely(fop->read \|\| fop->read_iter))
c9c554f2 AV	203	file->f_mode \|= FMODE_CAN_READ;
c9c554f2 AV	204	if ((file->f_mode & FMODE_WRITE) &&
84363182	205	likely(fop->write \|\| fop->write_iter))
c9c554f2	206	file->f_mode \|= FMODE_CAN_WRITE;
f5d11409	207	file->f_mode \|= FMODE_OPENED;
ce8d2cdf	208	file->f_op = fop;
c9c554f2	209	if ((file->f_mode & (FMODE_READ \| FMODE_WRITE)) == FMODE_READ)
890275b5	210	i_readcount_inc(path->dentry->d_inode);
3d1e4631	211	return file;
ce8d2cdf	212	}
ce8d2cdf	213
d93aa9d8 AV	214	struct file alloc_file_pseudo(struct inode inode, struct vfsmount *mnt,
	215	const char *name, int flags,
	216	const struct file_operations *fops)
	217	{
	218	static const struct dentry_operations anon_ops = {
	219	.d_dname = simple_dname
	220	};
	221	struct qstr this = QSTR_INIT(name, strlen(name));
	222	struct path path;
	223	struct file *file;
	224
	225	path.dentry = d_alloc_pseudo(mnt->mnt_sb, &this);
	226	if (!path.dentry)
	227	return ERR_PTR(-ENOMEM);
	228	if (!mnt->mnt_sb->s_d_op)
	229	d_set_d_op(path.dentry, &anon_ops);
	230	path.mnt = mntget(mnt);
	231	d_instantiate(path.dentry, inode);
	232	file = alloc_file(&path, flags, fops);
	233	if (IS_ERR(file)) {
	234	ihold(inode);
	235	path_put(&path);
	236	}
	237	return file;
	238	}
	239	EXPORT_SYMBOL(alloc_file_pseudo);
	240
183266f2 AV	241	struct file alloc_file_clone(struct file base, int flags,
	242	const struct file_operations *fops)
	243	{
	244	struct file *f = alloc_file(&base->f_path, flags, fops);
	245	if (!IS_ERR(f)) {
	246	path_get(&f->f_path);
	247	f->f_mapping = base->f_mapping;
	248	}
	249	return f;
	250	}
	251
d7065da0	252	/* the real guts of fput() - releasing the last reference to file
1da177e4	253	*/
d7065da0	254	static void __fput(struct file *file)
1da177e4	255	{
0f7fc9e4 JJS	256	struct dentry *dentry = file->f_path.dentry;
0f7fc9e4 JJS	257	struct vfsmount *mnt = file->f_path.mnt;
c77cecee	258	struct inode *inode = file->f_inode;
a07b2000	259	fmode_t mode = file->f_mode;
1da177e4	260
4d27f326 AV	261	if (unlikely(!(file->f_mode & FMODE_OPENED)))
	262	goto out;
	263
1da177e4	264	might_sleep();
0eeca283 RL	265
0eeca283 RL	266	fsnotify_close(file);
1da177e4 LT	267	/*
	268	* The function eventpoll_release() should be the first called
	269	* in the file cleanup chain.
	270	*/
	271	eventpoll_release(file);
78ed8a13	272	locks_remove_file(file);
1da177e4	273
bb02b186	274	ima_file_free(file);
233e70f4	275	if (unlikely(file->f_flags & FASYNC)) {
72c2d531	276	if (file->f_op->fasync)
233e70f4 AV	277	file->f_op->fasync(-1, file, 0);
233e70f4 AV	278	}
72c2d531	279	if (file->f_op->release)
1da177e4	280	file->f_op->release(inode, file);
60ed8cf7	281	if (unlikely(S_ISCHR(inode->i_mode) && inode->i_cdev != NULL &&
a07b2000	282	!(mode & FMODE_PATH))) {
1da177e4	283	cdev_put(inode->i_cdev);
60ed8cf7	284	}
1da177e4	285	fops_put(file->f_op);
609d7fa9	286	put_pid(file->f_owner.pid);
a07b2000	287	if ((mode & (FMODE_READ \| FMODE_WRITE)) == FMODE_READ)
890275b5	288	i_readcount_dec(inode);
a07b2000	289	if (mode & FMODE_WRITER) {
83f936c7 AV	290	put_write_access(inode);
	291	__mnt_drop_write(mnt);
	292	}
1da177e4	293	dput(dentry);
a07b2000 AV	294	if (unlikely(mode & FMODE_NEED_UNMOUNT))
a07b2000 AV	295	dissolve_on_fput(mnt);
1da177e4	296	mntput(mnt);
4d27f326 AV	297	out:
4d27f326 AV	298	file_free(file);
1da177e4 LT	299	}
1da177e4 LT	300
4f5e65a1	301	static LLIST_HEAD(delayed_fput_list);
4a9d4b02 AV	302	static void delayed_fput(struct work_struct *unused)
4a9d4b02 AV	303	{
4f5e65a1	304	struct llist_node *node = llist_del_all(&delayed_fput_list);
b9ea557e	305	struct file f, t;
4f5e65a1	306
b9ea557e BP	307	llist_for_each_entry_safe(f, t, node, f_u.fu_llist)
b9ea557e BP	308	__fput(f);
4a9d4b02 AV	309	}
	310
	311	static void ____fput(struct callback_head *work)
	312	{
	313	__fput(container_of(work, struct file, f_u.fu_rcuhead));
	314	}
	315
	316	/*
	317	* If kernel thread really needs to have the final fput() it has done
	318	* to complete, call this. The only user right now is the boot - we
	319	* do need to make sure our writes to binaries on initramfs has
	320	* not left us with opened struct file waiting for __fput() - execve()
	321	* won't work without that. Please, don't add more callers without
	322	* very good reasons; in particular, never call that with locks
	323	* held and never call that from a thread that might need to do
	324	* some work on any kind of umount.
	325	*/
	326	void flush_delayed_fput(void)
	327	{
	328	delayed_fput(NULL);
	329	}
	330
c7314d74	331	static DECLARE_DELAYED_WORK(delayed_fput_work, delayed_fput);
4a9d4b02	332
091141a4	333	void fput_many(struct file *file, unsigned int refs)
d7065da0	334	{
091141a4	335	if (atomic_long_sub_and_test(refs, &file->f_count)) {
4a9d4b02	336	struct task_struct *task = current;
e7b2c406	337
e7b2c406 ON	338	if (likely(!in_interrupt() && !(task->flags & PF_KTHREAD))) {
	339	init_task_work(&file->f_u.fu_rcuhead, ____fput);
	340	if (!task_work_add(task, &file->f_u.fu_rcuhead, true))
	341	return;
64372501 AM	342	/*
64372501 AM	343	* After this task has run exit_task_work(),
be49b30a	344	* task_work_add() will fail. Fall through to delayed
64372501 AM	345	* fput to avoid leaking *file.
64372501 AM	346	*/
4a9d4b02	347	}
4f5e65a1 ON	348
4f5e65a1 ON	349	if (llist_add(&file->f_u.fu_llist, &delayed_fput_list))
c7314d74	350	schedule_delayed_work(&delayed_fput_work, 1);
4a9d4b02 AV	351	}
	352	}
	353
091141a4 JA	354	void fput(struct file *file)
	355	{
	356	fput_many(file, 1);
	357	}
	358
4a9d4b02 AV	359	/*
	360	* synchronous analog of fput(); for kernel threads that might be needed
	361	* in some umount() (and thus can't use flush_delayed_fput() without
	362	* risking deadlocks), need to wait for completion of __fput() and know
	363	* for this specific struct file it won't involve anything that would
	364	* need them. Use only if you really need it - at the very least,
	365	* don't blindly convert fput() by kernel thread to that.
	366	*/
	367	void __fput_sync(struct file *file)
	368	{
	369	if (atomic_long_dec_and_test(&file->f_count)) {
	370	struct task_struct *task = current;
4a9d4b02	371	BUG_ON(!(task->flags & PF_KTHREAD));
d7065da0	372	__fput(file);
4a9d4b02	373	}
d7065da0 AV	374	}
	375
	376	EXPORT_SYMBOL(fput);
	377
4248b0da	378	void __init files_init(void)
b9ea557e	379	{
b6b3fdea	380	filp_cachep = kmem_cache_create("filp", sizeof(struct file), 0,
f3f7c093	381	SLAB_HWCACHE_ALIGN \| SLAB_PANIC \| SLAB_ACCOUNT, NULL);
4248b0da MG	382	percpu_counter_init(&nr_files, 0, GFP_KERNEL);
4248b0da MG	383	}
b6b3fdea	384
4248b0da MG	385	/*
	386	* One file with associated inode and dcache is very roughly 1K. Per default
	387	* do not use more than 10% of our memory for files.
	388	*/
	389	void __init files_maxfiles_init(void)
	390	{
	391	unsigned long n;
ca79b0c2	392	unsigned long nr_pages = totalram_pages();
3d6357de	393	unsigned long memreserve = (nr_pages - nr_free_pages()) * 3/2;
4248b0da	394
3d6357de AK	395	memreserve = min(memreserve, nr_pages - 1);
3d6357de AK	396	n = ((nr_pages - memreserve) * (PAGE_SIZE / 1024)) / 10;
1da177e4	397
518de9b3	398	files_stat.max_files = max_t(unsigned long, n, NR_FILE);
b9ea557e	399	}