[linux-block.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/filelock.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/kmemleak.h>

#include <linux/atomic.h>

#include "internal.h"

/* sysctl tunables... */
static 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;

/* Container for backing file with optional real path */
struct backing_file {
	struct file file;
	struct path real_path;
};

static inline struct backing_file *backing_file(struct file *f)
{
	return container_of(f, struct backing_file, file);
}

struct path *backing_file_real_path(struct file *f)
{
	return &backing_file(f)->real_path;
}
EXPORT_SYMBOL_GPL(backing_file_real_path);

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

	put_cred(f->f_cred);
	if (unlikely(f->f_mode & FMODE_BACKING))
		kfree(backing_file(f));
	else
		kmem_cache_free(filp_cachep, f);
}

static inline void file_free(struct file *f)
{
	security_file_free(f);
	if (unlikely(f->f_mode & FMODE_BACKING))
		path_put(backing_file_real_path(f));
	if (likely(!(f->f_mode & FMODE_NOACCOUNT)))
		percpu_counter_dec(&nr_files);
	call_rcu(&f->f_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);

#if defined(CONFIG_SYSCTL) && defined(CONFIG_PROC_FS)

/*
 * Handle nr_files sysctl
 */
static int proc_nr_files(struct ctl_table *table, int write, void *buffer,
			 size_t *lenp, loff_t *ppos)
{
	files_stat.nr_files = get_nr_files();
	return proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
}

static struct ctl_table fs_stat_sysctls[] = {
	{
		.procname	= "file-nr",
		.data		= &files_stat,
		.maxlen		= sizeof(files_stat),
		.mode		= 0444,
		.proc_handler	= proc_nr_files,
	},
	{
		.procname	= "file-max",
		.data		= &files_stat.max_files,
		.maxlen		= sizeof(files_stat.max_files),
		.mode		= 0644,
		.proc_handler	= proc_doulongvec_minmax,
		.extra1		= SYSCTL_LONG_ZERO,
		.extra2		= SYSCTL_LONG_MAX,
	},
	{
		.procname	= "nr_open",
		.data		= &sysctl_nr_open,
		.maxlen		= sizeof(unsigned int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_minmax,
		.extra1		= &sysctl_nr_open_min,
		.extra2		= &sysctl_nr_open_max,
	},
	{ }
};

static int __init init_fs_stat_sysctls(void)
{
	register_sysctl_init("fs", fs_stat_sysctls);
	if (IS_ENABLED(CONFIG_BINFMT_MISC)) {
		struct ctl_table_header *hdr;
		hdr = register_sysctl_mount_point("fs/binfmt_misc");
		kmemleak_not_leak(hdr);
	}
	return 0;
}
fs_initcall(init_fs_stat_sysctls);
#endif

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

	f->f_cred = get_cred(cred);
	error = security_file_alloc(f);
	if (unlikely(error)) {
		put_cred(f->f_cred);
		return 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);
	f->f_flags = flags;
	f->f_mode = OPEN_FMODE(flags);
	/* f->f_version: 0 */

	return 0;
}

/* 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;
	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);

	error = init_file(f, flags, cred);
	if (unlikely(error)) {
		kmem_cache_free(filp_cachep, f);
		return ERR_PTR(error);
	}

	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.
 *
 * This is only for kernel internal use, and the allocate file must not be
 * installed into file tables or such.
 */
struct file *alloc_empty_file_noaccount(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);

	error = init_file(f, flags, cred);
	if (unlikely(error)) {
		kmem_cache_free(filp_cachep, f);
		return ERR_PTR(error);
	}

	f->f_mode |= FMODE_NOACCOUNT;

	return f;
}

/*
 * Variant of alloc_empty_file() that allocates a backing_file container
 * and doesn't check and modify nr_files.
 *
 * This is only for kernel internal use, and the allocate file must not be
 * installed into file tables or such.
 */
struct file *alloc_empty_backing_file(int flags, const struct cred *cred)
{
	struct backing_file *ff;
	int error;

	ff = kzalloc(sizeof(struct backing_file), GFP_KERNEL);
	if (unlikely(!ff))
		return ERR_PTR(-ENOMEM);

	error = init_file(&ff->file, flags, cred);
	if (unlikely(error)) {
		kfree(ff);
		return ERR_PTR(error);
	}

	ff->file.f_mode |= FMODE_BACKING | FMODE_NOACCOUNT;
	return &ff->file;
}

/**
 * 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);
	file->f_sb_err = file_sample_sb_err(file);
	if (fop->llseek)
		file->f_mode |= FMODE_LSEEK;
	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_iocb_flags = iocb_flags(file);
	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);
	put_file_access(file);
	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_llist)
		__fput(f);
}

static void ____fput(struct callback_head *work)
{
	__fput(container_of(work, struct file, f_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);
}
EXPORT_SYMBOL_GPL(flush_delayed_fput);

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_rcuhead, ____fput);
			if (!task_work_add(task, &file->f_rcuhead, TWA_RESUME))
				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_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);
EXPORT_SYMBOL(__fput_sync);

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	15	#include <linux/fs.h>
5970e15d	16	#include <linux/filelock.h>
1da177e4	17	#include <linux/security.h>
5b825c3a	18	#include <linux/cred.h>
1da177e4	19	#include <linux/eventpoll.h>
ab2af1f5	20	#include <linux/rcupdate.h>
1da177e4	21	#include <linux/mount.h>
16f7e0fe	22	#include <linux/capability.h>
1da177e4	23	#include <linux/cdev.h>
0eeca283	24	#include <linux/fsnotify.h>
529bf6be DS	25	#include <linux/sysctl.h>
529bf6be DS	26	#include <linux/percpu_counter.h>
6416ccb7	27	#include <linux/percpu.h>
4a9d4b02	28	#include <linux/task_work.h>
0552f879	29	#include <linux/ima.h>
4248b0da	30	#include <linux/swap.h>
a3580ac9	31	#include <linux/kmemleak.h>
529bf6be	32
60063497	33	#include <linux/atomic.h>
1da177e4	34
e81e3f4d EP	35	#include "internal.h"
e81e3f4d EP	36
1da177e4	37	/* sysctl tunables... */
204d5a24	38	static struct files_stat_struct files_stat = {
1da177e4 LT	39	.max_files = NR_FILE
	40	};
	41
b6b3fdea ED	42	/* SLAB cache for file structures */
	43	static struct kmem_cache *filp_cachep __read_mostly;
	44
529bf6be	45	static struct percpu_counter nr_files __cacheline_aligned_in_smp;
1da177e4	46
62d53c4a AG	47	/* Container for backing file with optional real path */
	48	struct backing_file {
	49	struct file file;
	50	struct path real_path;
	51	};
	52
	53	static inline struct backing_file backing_file(struct file f)
	54	{
	55	return container_of(f, struct backing_file, file);
	56	}
	57
	58	struct path backing_file_real_path(struct file f)
	59	{
	60	return &backing_file(f)->real_path;
	61	}
	62	EXPORT_SYMBOL_GPL(backing_file_real_path);
	63
5c33b183	64	static void file_free_rcu(struct rcu_head *head)
1da177e4	65	{
e87f2c26	66	struct file *f = container_of(head, struct file, f_rcuhead);
d76b0d9b DH	67
d76b0d9b DH	68	put_cred(f->f_cred);
62d53c4a AG	69	if (unlikely(f->f_mode & FMODE_BACKING))
	70	kfree(backing_file(f));
	71	else
	72	kmem_cache_free(filp_cachep, f);
1da177e4 LT	73	}
1da177e4 LT	74
529bf6be	75	static inline void file_free(struct file *f)
1da177e4	76	{
e8cff84f	77	security_file_free(f);
62d53c4a AG	78	if (unlikely(f->f_mode & FMODE_BACKING))
	79	path_put(backing_file_real_path(f));
	80	if (likely(!(f->f_mode & FMODE_NOACCOUNT)))
d3b1084d	81	percpu_counter_dec(&nr_files);
e87f2c26	82	call_rcu(&f->f_rcuhead, file_free_rcu);
1da177e4 LT	83	}
1da177e4 LT	84
529bf6be DS	85	/*
	86	* Return the total number of open files in the system
	87	*/
518de9b3	88	static long get_nr_files(void)
1da177e4	89	{
529bf6be	90	return percpu_counter_read_positive(&nr_files);
1da177e4 LT	91	}
1da177e4 LT	92
529bf6be DS	93	/*
	94	* Return the maximum number of open files in the system
	95	*/
518de9b3	96	unsigned long get_max_files(void)
ab2af1f5	97	{
529bf6be	98	return files_stat.max_files;
ab2af1f5	99	}
529bf6be DS	100	EXPORT_SYMBOL_GPL(get_max_files);
529bf6be DS	101
204d5a24 LC	102	#if defined(CONFIG_SYSCTL) && defined(CONFIG_PROC_FS)
204d5a24 LC	103
529bf6be DS	104	/*
	105	* Handle nr_files sysctl
	106	*/
204d5a24 LC	107	static int proc_nr_files(struct ctl_table table, int write, void buffer,
204d5a24 LC	108	size_t lenp, loff_t ppos)
529bf6be DS	109	{
529bf6be DS	110	files_stat.nr_files = get_nr_files();
518de9b3	111	return proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
529bf6be	112	}
204d5a24 LC	113
	114	static struct ctl_table fs_stat_sysctls[] = {
	115	{
	116	.procname = "file-nr",
	117	.data = &files_stat,
	118	.maxlen = sizeof(files_stat),
	119	.mode = 0444,
	120	.proc_handler = proc_nr_files,
	121	},
	122	{
	123	.procname = "file-max",
	124	.data = &files_stat.max_files,
	125	.maxlen = sizeof(files_stat.max_files),
	126	.mode = 0644,
	127	.proc_handler = proc_doulongvec_minmax,
	128	.extra1 = SYSCTL_LONG_ZERO,
	129	.extra2 = SYSCTL_LONG_MAX,
	130	},
	131	{
	132	.procname = "nr_open",
	133	.data = &sysctl_nr_open,
	134	.maxlen = sizeof(unsigned int),
	135	.mode = 0644,
	136	.proc_handler = proc_dointvec_minmax,
	137	.extra1 = &sysctl_nr_open_min,
	138	.extra2 = &sysctl_nr_open_max,
	139	},
	140	{ }
	141	};
	142
	143	static int __init init_fs_stat_sysctls(void)
529bf6be	144	{
204d5a24	145	register_sysctl_init("fs", fs_stat_sysctls);
a3580ac9 LC	146	if (IS_ENABLED(CONFIG_BINFMT_MISC)) {
	147	struct ctl_table_header *hdr;
	148	hdr = register_sysctl_mount_point("fs/binfmt_misc");
	149	kmemleak_not_leak(hdr);
	150	}
204d5a24	151	return 0;
529bf6be	152	}
204d5a24	153	fs_initcall(init_fs_stat_sysctls);
529bf6be	154	#endif
ab2af1f5	155
8a05a8c3	156	static int init_file(struct file f, int flags, const struct cred cred)
d3b1084d	157	{
d3b1084d MS	158	int error;
d3b1084d MS	159
d3b1084d MS	160	f->f_cred = get_cred(cred);
	161	error = security_file_alloc(f);
	162	if (unlikely(error)) {
dff745c1	163	put_cred(f->f_cred);
8a05a8c3	164	return error;
d3b1084d MS	165	}
	166
	167	atomic_long_set(&f->f_count, 1);
	168	rwlock_init(&f->f_owner.lock);
	169	spin_lock_init(&f->f_lock);
	170	mutex_init(&f->f_pos_lock);
d3b1084d MS	171	f->f_flags = flags;
	172	f->f_mode = OPEN_FMODE(flags);
	173	/* f->f_version: 0 */
	174
8a05a8c3	175	return 0;
d3b1084d MS	176	}
d3b1084d MS	177
1da177e4	178	/* Find an unused file structure and return a pointer to it.
1afc99be AV	179	* Returns an error pointer if some error happend e.g. we over file
1afc99be AV	180	* structures limit, run out of memory or operation is not permitted.
430e285e DH	181	*
	182	* Be very careful using this. You are responsible for
	183	* getting write access to any mount that you might assign
	184	* to this filp, if it is opened for write. If this is not
	185	* done, you will imbalance int the mount's writer count
	186	* and a warning at __fput() time.
1da177e4	187	*/
ea73ea72	188	struct file alloc_empty_file(int flags, const struct cred cred)
1da177e4	189	{
518de9b3	190	static long old_max;
1afc99be	191	struct file *f;
8a05a8c3	192	int error;
1da177e4 LT	193
	194	/*
	195	* Privileged users can go above max_files
	196	*/
529bf6be DS	197	if (get_nr_files() >= files_stat.max_files && !capable(CAP_SYS_ADMIN)) {
	198	/*
	199	* percpu_counters are inaccurate. Do an expensive check before
	200	* we go and fail.
	201	*/
52d9f3b4	202	if (percpu_counter_sum_positive(&nr_files) >= files_stat.max_files)
529bf6be DS	203	goto over;
529bf6be DS	204	}
af4d2ecb	205
8a05a8c3 AG	206	f = kmem_cache_zalloc(filp_cachep, GFP_KERNEL);
	207	if (unlikely(!f))
	208	return ERR_PTR(-ENOMEM);
	209
	210	error = init_file(f, flags, cred);
dff745c1 AG	211	if (unlikely(error)) {
dff745c1 AG	212	kmem_cache_free(filp_cachep, f);
8a05a8c3	213	return ERR_PTR(error);
dff745c1	214	}
8a05a8c3 AG	215
8a05a8c3 AG	216	percpu_counter_inc(&nr_files);
1da177e4	217
af4d2ecb KK	218	return f;
	219
	220	over:
1da177e4	221	/* Ran out of filps - report that */
529bf6be	222	if (get_nr_files() > old_max) {
518de9b3	223	pr_info("VFS: file-max limit %lu reached\n", get_max_files());
529bf6be	224	old_max = get_nr_files();
1da177e4	225	}
1afc99be	226	return ERR_PTR(-ENFILE);
1da177e4 LT	227	}
1da177e4 LT	228
d3b1084d MS	229	/*
	230	* Variant of alloc_empty_file() that doesn't check and modify nr_files.
	231	*
8a05a8c3 AG	232	* This is only for kernel internal use, and the allocate file must not be
8a05a8c3 AG	233	* installed into file tables or such.
d3b1084d MS	234	*/
	235	struct file alloc_empty_file_noaccount(int flags, const struct cred cred)
	236	{
8a05a8c3 AG	237	struct file *f;
	238	int error;
	239
	240	f = kmem_cache_zalloc(filp_cachep, GFP_KERNEL);
	241	if (unlikely(!f))
	242	return ERR_PTR(-ENOMEM);
	243
	244	error = init_file(f, flags, cred);
dff745c1 AG	245	if (unlikely(error)) {
dff745c1 AG	246	kmem_cache_free(filp_cachep, f);
8a05a8c3	247	return ERR_PTR(error);
dff745c1	248	}
d3b1084d	249
8a05a8c3	250	f->f_mode \|= FMODE_NOACCOUNT;
d3b1084d MS	251
	252	return f;
	253	}
	254
62d53c4a AG	255	/*
	256	* Variant of alloc_empty_file() that allocates a backing_file container
	257	* and doesn't check and modify nr_files.
	258	*
	259	* This is only for kernel internal use, and the allocate file must not be
	260	* installed into file tables or such.
	261	*/
	262	struct file alloc_empty_backing_file(int flags, const struct cred cred)
	263	{
	264	struct backing_file *ff;
	265	int error;
	266
	267	ff = kzalloc(sizeof(struct backing_file), GFP_KERNEL);
	268	if (unlikely(!ff))
	269	return ERR_PTR(-ENOMEM);
	270
	271	error = init_file(&ff->file, flags, cred);
dff745c1 AG	272	if (unlikely(error)) {
dff745c1 AG	273	kfree(ff);
62d53c4a	274	return ERR_PTR(error);
dff745c1	275	}
62d53c4a AG	276
	277	ff->file.f_mode \|= FMODE_BACKING \| FMODE_NOACCOUNT;
	278	return &ff->file;
	279	}
	280
ce8d2cdf DH	281	/**
ce8d2cdf DH	282	* alloc_file - allocate and initialize a 'struct file'
a457606a EB	283	*
a457606a EB	284	* @path: the (dentry, vfsmount) pair for the new file
c9c554f2	285	* @flags: O_... flags with which the new file will be opened
ce8d2cdf	286	* @fop: the 'struct file_operations' for the new file
ce8d2cdf	287	*/
ee1904ba	288	static struct file alloc_file(const struct path path, int flags,
2c48b9c4	289	const struct file_operations *fop)
ce8d2cdf DH	290	{
ce8d2cdf DH	291	struct file *file;
ce8d2cdf	292
ea73ea72	293	file = alloc_empty_file(flags, current_cred());
1afc99be	294	if (IS_ERR(file))
39b65252	295	return file;
ce8d2cdf	296
2c48b9c4	297	file->f_path = *path;
dd37978c	298	file->f_inode = path->dentry->d_inode;
2c48b9c4	299	file->f_mapping = path->dentry->d_inode->i_mapping;
5660e13d	300	file->f_wb_err = filemap_sample_wb_err(file->f_mapping);
735e4ae5	301	file->f_sb_err = file_sample_sb_err(file);
868941b1	302	if (fop->llseek)
e7478158	303	file->f_mode \|= FMODE_LSEEK;
c9c554f2	304	if ((file->f_mode & FMODE_READ) &&
84363182	305	likely(fop->read \|\| fop->read_iter))
c9c554f2 AV	306	file->f_mode \|= FMODE_CAN_READ;
c9c554f2 AV	307	if ((file->f_mode & FMODE_WRITE) &&
84363182	308	likely(fop->write \|\| fop->write_iter))
c9c554f2	309	file->f_mode \|= FMODE_CAN_WRITE;
164f4064	310	file->f_iocb_flags = iocb_flags(file);
f5d11409	311	file->f_mode \|= FMODE_OPENED;
ce8d2cdf	312	file->f_op = fop;
c9c554f2	313	if ((file->f_mode & (FMODE_READ \| FMODE_WRITE)) == FMODE_READ)
890275b5	314	i_readcount_inc(path->dentry->d_inode);
3d1e4631	315	return file;
ce8d2cdf	316	}
ce8d2cdf	317
d93aa9d8 AV	318	struct file alloc_file_pseudo(struct inode inode, struct vfsmount *mnt,
	319	const char *name, int flags,
	320	const struct file_operations *fops)
	321	{
	322	static const struct dentry_operations anon_ops = {
	323	.d_dname = simple_dname
	324	};
	325	struct qstr this = QSTR_INIT(name, strlen(name));
	326	struct path path;
	327	struct file *file;
	328
	329	path.dentry = d_alloc_pseudo(mnt->mnt_sb, &this);
	330	if (!path.dentry)
	331	return ERR_PTR(-ENOMEM);
	332	if (!mnt->mnt_sb->s_d_op)
	333	d_set_d_op(path.dentry, &anon_ops);
	334	path.mnt = mntget(mnt);
	335	d_instantiate(path.dentry, inode);
b6509f6a	336	file = alloc_file(&path, flags, fops);
d93aa9d8 AV	337	if (IS_ERR(file)) {
	338	ihold(inode);
	339	path_put(&path);
	340	}
	341	return file;
	342	}
	343	EXPORT_SYMBOL(alloc_file_pseudo);
	344
183266f2 AV	345	struct file alloc_file_clone(struct file base, int flags,
	346	const struct file_operations *fops)
	347	{
	348	struct file *f = alloc_file(&base->f_path, flags, fops);
	349	if (!IS_ERR(f)) {
	350	path_get(&f->f_path);
	351	f->f_mapping = base->f_mapping;
	352	}
	353	return f;
	354	}
	355
d7065da0	356	/* the real guts of fput() - releasing the last reference to file
1da177e4	357	*/
d7065da0	358	static void __fput(struct file *file)
1da177e4	359	{
0f7fc9e4 JJS	360	struct dentry *dentry = file->f_path.dentry;
0f7fc9e4 JJS	361	struct vfsmount *mnt = file->f_path.mnt;
c77cecee	362	struct inode *inode = file->f_inode;
a07b2000	363	fmode_t mode = file->f_mode;
1da177e4	364
4d27f326 AV	365	if (unlikely(!(file->f_mode & FMODE_OPENED)))
	366	goto out;
	367
1da177e4	368	might_sleep();
0eeca283 RL	369
0eeca283 RL	370	fsnotify_close(file);
1da177e4 LT	371	/*
	372	* The function eventpoll_release() should be the first called
	373	* in the file cleanup chain.
	374	*/
	375	eventpoll_release(file);
78ed8a13	376	locks_remove_file(file);
1da177e4	377
bb02b186	378	ima_file_free(file);
233e70f4	379	if (unlikely(file->f_flags & FASYNC)) {
72c2d531	380	if (file->f_op->fasync)
233e70f4 AV	381	file->f_op->fasync(-1, file, 0);
233e70f4 AV	382	}
72c2d531	383	if (file->f_op->release)
1da177e4	384	file->f_op->release(inode, file);
60ed8cf7	385	if (unlikely(S_ISCHR(inode->i_mode) && inode->i_cdev != NULL &&
a07b2000	386	!(mode & FMODE_PATH))) {
1da177e4	387	cdev_put(inode->i_cdev);
60ed8cf7	388	}
1da177e4	389	fops_put(file->f_op);
609d7fa9	390	put_pid(file->f_owner.pid);
d6da19c9	391	put_file_access(file);
1da177e4	392	dput(dentry);
a07b2000 AV	393	if (unlikely(mode & FMODE_NEED_UNMOUNT))
a07b2000 AV	394	dissolve_on_fput(mnt);
1da177e4	395	mntput(mnt);
4d27f326 AV	396	out:
4d27f326 AV	397	file_free(file);
1da177e4 LT	398	}
1da177e4 LT	399
4f5e65a1	400	static LLIST_HEAD(delayed_fput_list);
4a9d4b02 AV	401	static void delayed_fput(struct work_struct *unused)
4a9d4b02 AV	402	{
4f5e65a1	403	struct llist_node *node = llist_del_all(&delayed_fput_list);
b9ea557e	404	struct file f, t;
4f5e65a1	405
e87f2c26	406	llist_for_each_entry_safe(f, t, node, f_llist)
b9ea557e	407	__fput(f);
4a9d4b02 AV	408	}
	409
	410	static void ____fput(struct callback_head *work)
	411	{
e87f2c26	412	__fput(container_of(work, struct file, f_rcuhead));
4a9d4b02 AV	413	}
	414
	415	/*
	416	* If kernel thread really needs to have the final fput() it has done
	417	* to complete, call this. The only user right now is the boot - we
	418	* do need to make sure our writes to binaries on initramfs has
	419	* not left us with opened struct file waiting for __fput() - execve()
	420	* won't work without that. Please, don't add more callers without
	421	* very good reasons; in particular, never call that with locks
	422	* held and never call that from a thread that might need to do
	423	* some work on any kind of umount.
	424	*/
	425	void flush_delayed_fput(void)
	426	{
	427	delayed_fput(NULL);
	428	}
7239a40c	429	EXPORT_SYMBOL_GPL(flush_delayed_fput);
4a9d4b02	430
c7314d74	431	static DECLARE_DELAYED_WORK(delayed_fput_work, delayed_fput);
4a9d4b02	432
81132a39	433	void fput(struct file *file)
d7065da0	434	{
81132a39	435	if (atomic_long_dec_and_test(&file->f_count)) {
4a9d4b02	436	struct task_struct *task = current;
e7b2c406	437
e7b2c406	438	if (likely(!in_interrupt() && !(task->flags & PF_KTHREAD))) {
e87f2c26 AV	439	init_task_work(&file->f_rcuhead, ____fput);
e87f2c26 AV	440	if (!task_work_add(task, &file->f_rcuhead, TWA_RESUME))
e7b2c406	441	return;
64372501 AM	442	/*
64372501 AM	443	* After this task has run exit_task_work(),
be49b30a	444	* task_work_add() will fail. Fall through to delayed
64372501 AM	445	* fput to avoid leaking *file.
64372501 AM	446	*/
4a9d4b02	447	}
4f5e65a1	448
e87f2c26	449	if (llist_add(&file->f_llist, &delayed_fput_list))
c7314d74	450	schedule_delayed_work(&delayed_fput_work, 1);
4a9d4b02 AV	451	}
	452	}
	453
	454	/*
	455	* synchronous analog of fput(); for kernel threads that might be needed
	456	* in some umount() (and thus can't use flush_delayed_fput() without
	457	* risking deadlocks), need to wait for completion of __fput() and know
	458	* for this specific struct file it won't involve anything that would
	459	* need them. Use only if you really need it - at the very least,
	460	* don't blindly convert fput() by kernel thread to that.
	461	*/
	462	void __fput_sync(struct file *file)
	463	{
	464	if (atomic_long_dec_and_test(&file->f_count)) {
	465	struct task_struct *task = current;
4a9d4b02	466	BUG_ON(!(task->flags & PF_KTHREAD));
d7065da0	467	__fput(file);
4a9d4b02	468	}
d7065da0 AV	469	}
	470
	471	EXPORT_SYMBOL(fput);
f0043206	472	EXPORT_SYMBOL(__fput_sync);
d7065da0	473
4248b0da	474	void __init files_init(void)
b9ea557e	475	{
b6b3fdea	476	filp_cachep = kmem_cache_create("filp", sizeof(struct file), 0,
f3f7c093	477	SLAB_HWCACHE_ALIGN \| SLAB_PANIC \| SLAB_ACCOUNT, NULL);
4248b0da MG	478	percpu_counter_init(&nr_files, 0, GFP_KERNEL);
4248b0da MG	479	}
b6b3fdea	480
4248b0da MG	481	/*
	482	* One file with associated inode and dcache is very roughly 1K. Per default
	483	* do not use more than 10% of our memory for files.
	484	*/
	485	void __init files_maxfiles_init(void)
	486	{
	487	unsigned long n;
ca79b0c2	488	unsigned long nr_pages = totalram_pages();
3d6357de	489	unsigned long memreserve = (nr_pages - nr_free_pages()) * 3/2;
4248b0da	490
3d6357de AK	491	memreserve = min(memreserve, nr_pages - 1);
3d6357de AK	492	n = ((nr_pages - memreserve) * (PAGE_SIZE / 1024)) / 10;
1da177e4	493
518de9b3	494	files_stat.max_files = max_t(unsigned long, n, NR_FILE);
b9ea557e	495	}