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

/*
 *  fs/eventfd.c
 *
 *  Copyright (C) 2007  Davide Libenzi <davidel@xmailserver.org>
 *
 */

#include <linux/file.h>
#include <linux/poll.h>
#include <linux/init.h>
#include <linux/fs.h>
#include <linux/sched.h>
#include <linux/kernel.h>
#include <linux/list.h>
#include <linux/spinlock.h>
#include <linux/anon_inodes.h>
#include <linux/syscalls.h>
#include <linux/module.h>
#include <linux/kref.h>
#include <linux/eventfd.h>

struct eventfd_ctx {
	struct kref kref;
	wait_queue_head_t wqh;
	/*
	 * Every time that a write(2) is performed on an eventfd, the
	 * value of the __u64 being written is added to "count" and a
	 * wakeup is performed on "wqh". A read(2) will return the "count"
	 * value to userspace, and will reset "count" to zero. The kernel
	 * side eventfd_signal() also, adds to the "count" counter and
	 * issue a wakeup.
	 */
	__u64 count;
	unsigned int flags;
};

/**
 * eventfd_signal - Adds @n to the eventfd counter.
 * @ctx: [in] Pointer to the eventfd context.
 * @n: [in] Value of the counter to be added to the eventfd internal counter.
 *          The value cannot be negative.
 *
 * This function is supposed to be called by the kernel in paths that do not
 * allow sleeping. In this function we allow the counter to reach the ULLONG_MAX
 * value, and we signal this as overflow condition by returining a POLLERR
 * to poll(2).
 *
 * Returns @n in case of success, a non-negative number lower than @n in case
 * of overflow, or the following error codes:
 *
 * -EINVAL    : The value of @n is negative.
 */
int eventfd_signal(struct eventfd_ctx *ctx, int n)
{
	unsigned long flags;

	if (n < 0)
		return -EINVAL;
	spin_lock_irqsave(&ctx->wqh.lock, flags);
	if (ULLONG_MAX - ctx->count < n)
		n = (int) (ULLONG_MAX - ctx->count);
	ctx->count += n;
	if (waitqueue_active(&ctx->wqh))
		wake_up_locked_poll(&ctx->wqh, POLLIN);
	spin_unlock_irqrestore(&ctx->wqh.lock, flags);

	return n;
}
EXPORT_SYMBOL_GPL(eventfd_signal);

static void eventfd_free_ctx(struct eventfd_ctx *ctx)
{
	kfree(ctx);
}

static void eventfd_free(struct kref *kref)
{
	struct eventfd_ctx *ctx = container_of(kref, struct eventfd_ctx, kref);

	eventfd_free_ctx(ctx);
}

/**
 * eventfd_ctx_get - Acquires a reference to the internal eventfd context.
 * @ctx: [in] Pointer to the eventfd context.
 *
 * Returns: In case of success, returns a pointer to the eventfd context.
 */
struct eventfd_ctx *eventfd_ctx_get(struct eventfd_ctx *ctx)
{
	kref_get(&ctx->kref);
	return ctx;
}
EXPORT_SYMBOL_GPL(eventfd_ctx_get);

/**
 * eventfd_ctx_put - Releases a reference to the internal eventfd context.
 * @ctx: [in] Pointer to eventfd context.
 *
 * The eventfd context reference must have been previously acquired either
 * with eventfd_ctx_get() or eventfd_ctx_fdget()).
 */
void eventfd_ctx_put(struct eventfd_ctx *ctx)
{
	kref_put(&ctx->kref, eventfd_free);
}
EXPORT_SYMBOL_GPL(eventfd_ctx_put);

static int eventfd_release(struct inode *inode, struct file *file)
{
	struct eventfd_ctx *ctx = file->private_data;

	wake_up_poll(&ctx->wqh, POLLHUP);
	eventfd_ctx_put(ctx);
	return 0;
}

static unsigned int eventfd_poll(struct file *file, poll_table *wait)
{
	struct eventfd_ctx *ctx = file->private_data;
	unsigned int events = 0;
	unsigned long flags;

	poll_wait(file, &ctx->wqh, wait);

	spin_lock_irqsave(&ctx->wqh.lock, flags);
	if (ctx->count > 0)
		events |= POLLIN;
	if (ctx->count == ULLONG_MAX)
		events |= POLLERR;
	if (ULLONG_MAX - 1 > ctx->count)
		events |= POLLOUT;
	spin_unlock_irqrestore(&ctx->wqh.lock, flags);

	return events;
}

static ssize_t eventfd_read(struct file *file, char __user *buf, size_t count,
			    loff_t *ppos)
{
	struct eventfd_ctx *ctx = file->private_data;
	ssize_t res;
	__u64 ucnt = 0;
	DECLARE_WAITQUEUE(wait, current);

	if (count < sizeof(ucnt))
		return -EINVAL;
	spin_lock_irq(&ctx->wqh.lock);
	res = -EAGAIN;
	if (ctx->count > 0)
		res = sizeof(ucnt);
	else if (!(file->f_flags & O_NONBLOCK)) {
		__add_wait_queue(&ctx->wqh, &wait);
		for (res = 0;;) {
			set_current_state(TASK_INTERRUPTIBLE);
			if (ctx->count > 0) {
				res = sizeof(ucnt);
				break;
			}
			if (signal_pending(current)) {
				res = -ERESTARTSYS;
				break;
			}
			spin_unlock_irq(&ctx->wqh.lock);
			schedule();
			spin_lock_irq(&ctx->wqh.lock);
		}
		__remove_wait_queue(&ctx->wqh, &wait);
		__set_current_state(TASK_RUNNING);
	}
	if (likely(res > 0)) {
		ucnt = (ctx->flags & EFD_SEMAPHORE) ? 1 : ctx->count;
		ctx->count -= ucnt;
		if (waitqueue_active(&ctx->wqh))
			wake_up_locked_poll(&ctx->wqh, POLLOUT);
	}
	spin_unlock_irq(&ctx->wqh.lock);
	if (res > 0 && put_user(ucnt, (__u64 __user *) buf))
		return -EFAULT;

	return res;
}

static ssize_t eventfd_write(struct file *file, const char __user *buf, size_t count,
			     loff_t *ppos)
{
	struct eventfd_ctx *ctx = file->private_data;
	ssize_t res;
	__u64 ucnt;
	DECLARE_WAITQUEUE(wait, current);

	if (count < sizeof(ucnt))
		return -EINVAL;
	if (copy_from_user(&ucnt, buf, sizeof(ucnt)))
		return -EFAULT;
	if (ucnt == ULLONG_MAX)
		return -EINVAL;
	spin_lock_irq(&ctx->wqh.lock);
	res = -EAGAIN;
	if (ULLONG_MAX - ctx->count > ucnt)
		res = sizeof(ucnt);
	else if (!(file->f_flags & O_NONBLOCK)) {
		__add_wait_queue(&ctx->wqh, &wait);
		for (res = 0;;) {
			set_current_state(TASK_INTERRUPTIBLE);
			if (ULLONG_MAX - ctx->count > ucnt) {
				res = sizeof(ucnt);
				break;
			}
			if (signal_pending(current)) {
				res = -ERESTARTSYS;
				break;
			}
			spin_unlock_irq(&ctx->wqh.lock);
			schedule();
			spin_lock_irq(&ctx->wqh.lock);
		}
		__remove_wait_queue(&ctx->wqh, &wait);
		__set_current_state(TASK_RUNNING);
	}
	if (likely(res > 0)) {
		ctx->count += ucnt;
		if (waitqueue_active(&ctx->wqh))
			wake_up_locked_poll(&ctx->wqh, POLLIN);
	}
	spin_unlock_irq(&ctx->wqh.lock);

	return res;
}

static const struct file_operations eventfd_fops = {
	.release	= eventfd_release,
	.poll		= eventfd_poll,
	.read		= eventfd_read,
	.write		= eventfd_write,
};

/**
 * eventfd_fget - Acquire a reference of an eventfd file descriptor.
 * @fd: [in] Eventfd file descriptor.
 *
 * Returns a pointer to the eventfd file structure in case of success, or the
 * following error pointer:
 *
 * -EBADF    : Invalid @fd file descriptor.
 * -EINVAL   : The @fd file descriptor is not an eventfd file.
 */
struct file *eventfd_fget(int fd)
{
	struct file *file;

	file = fget(fd);
	if (!file)
		return ERR_PTR(-EBADF);
	if (file->f_op != &eventfd_fops) {
		fput(file);
		return ERR_PTR(-EINVAL);
	}

	return file;
}
EXPORT_SYMBOL_GPL(eventfd_fget);

/**
 * eventfd_ctx_fdget - Acquires a reference to the internal eventfd context.
 * @fd: [in] Eventfd file descriptor.
 *
 * Returns a pointer to the internal eventfd context, otherwise the error
 * pointers returned by the following functions:
 *
 * eventfd_fget
 */
struct eventfd_ctx *eventfd_ctx_fdget(int fd)
{
	struct file *file;
	struct eventfd_ctx *ctx;

	file = eventfd_fget(fd);
	if (IS_ERR(file))
		return (struct eventfd_ctx *) file;
	ctx = eventfd_ctx_get(file->private_data);
	fput(file);

	return ctx;
}
EXPORT_SYMBOL_GPL(eventfd_ctx_fdget);

/**
 * eventfd_ctx_fileget - Acquires a reference to the internal eventfd context.
 * @file: [in] Eventfd file pointer.
 *
 * Returns a pointer to the internal eventfd context, otherwise the error
 * pointer:
 *
 * -EINVAL   : The @fd file descriptor is not an eventfd file.
 */
struct eventfd_ctx *eventfd_ctx_fileget(struct file *file)
{
	if (file->f_op != &eventfd_fops)
		return ERR_PTR(-EINVAL);

	return eventfd_ctx_get(file->private_data);
}
EXPORT_SYMBOL_GPL(eventfd_ctx_fileget);

/**
 * eventfd_file_create - Creates an eventfd file pointer.
 * @count: Initial eventfd counter value.
 * @flags: Flags for the eventfd file.
 *
 * This function creates an eventfd file pointer, w/out installing it into
 * the fd table. This is useful when the eventfd file is used during the
 * initialization of data structures that require extra setup after the eventfd
 * creation. So the eventfd creation is split into the file pointer creation
 * phase, and the file descriptor installation phase.
 * In this way races with userspace closing the newly installed file descriptor
 * can be avoided.
 * Returns an eventfd file pointer, or a proper error pointer.
 */
struct file *eventfd_file_create(unsigned int count, int flags)
{
	struct file *file;
	struct eventfd_ctx *ctx;

	/* Check the EFD_* constants for consistency.  */
	BUILD_BUG_ON(EFD_CLOEXEC != O_CLOEXEC);
	BUILD_BUG_ON(EFD_NONBLOCK != O_NONBLOCK);

	if (flags & ~EFD_FLAGS_SET)
		return ERR_PTR(-EINVAL);

	ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
	if (!ctx)
		return ERR_PTR(-ENOMEM);

	kref_init(&ctx->kref);
	init_waitqueue_head(&ctx->wqh);
	ctx->count = count;
	ctx->flags = flags;

	file = anon_inode_getfile("[eventfd]", &eventfd_fops, ctx,
				  flags & EFD_SHARED_FCNTL_FLAGS);
	if (IS_ERR(file))
		eventfd_free_ctx(ctx);

	return file;
}

SYSCALL_DEFINE2(eventfd2, unsigned int, count, int, flags)
{
	int fd, error;
	struct file *file;

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

	file = eventfd_file_create(count, 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;
}

SYSCALL_DEFINE1(eventfd, unsigned int, count)
{
	return sys_eventfd2(count, 0);
}
Commit	Line	Data
e1ad7468 DL	1	/*
	2	* fs/eventfd.c
	3	*
	4	* Copyright (C) 2007 Davide Libenzi <davidel@xmailserver.org>
	5	*
	6	*/
	7
	8	#include <linux/file.h>
	9	#include <linux/poll.h>
	10	#include <linux/init.h>
	11	#include <linux/fs.h>
	12	#include <linux/sched.h>
	13	#include <linux/kernel.h>
	14	#include <linux/list.h>
	15	#include <linux/spinlock.h>
	16	#include <linux/anon_inodes.h>
7747cdb2	17	#include <linux/syscalls.h>
5718607b	18	#include <linux/module.h>
13389010 DL	19	#include <linux/kref.h>
13389010 DL	20	#include <linux/eventfd.h>
e1ad7468 DL	21
e1ad7468 DL	22	struct eventfd_ctx {
13389010	23	struct kref kref;
e1ad7468 DL	24	wait_queue_head_t wqh;
	25	/*
	26	* Every time that a write(2) is performed on an eventfd, the
	27	* value of the __u64 being written is added to "count" and a
	28	* wakeup is performed on "wqh". A read(2) will return the "count"
	29	* value to userspace, and will reset "count" to zero. The kernel
13389010	30	* side eventfd_signal() also, adds to the "count" counter and
e1ad7468 DL	31	* issue a wakeup.
	32	*/
	33	__u64 count;
bcd0b235	34	unsigned int flags;
e1ad7468 DL	35	};
e1ad7468 DL	36
13389010 DL	37	/**
	38	* eventfd_signal - Adds @n to the eventfd counter.
	39	* @ctx: [in] Pointer to the eventfd context.
	40	* @n: [in] Value of the counter to be added to the eventfd internal counter.
	41	* The value cannot be negative.
	42	*
	43	* This function is supposed to be called by the kernel in paths that do not
	44	* allow sleeping. In this function we allow the counter to reach the ULLONG_MAX
	45	* value, and we signal this as overflow condition by returining a POLLERR
	46	* to poll(2).
	47	*
	48	* Returns @n in case of success, a non-negative number lower than @n in case
	49	* of overflow, or the following error codes:
	50	*
	51	* -EINVAL : The value of @n is negative.
e1ad7468	52	*/
13389010	53	int eventfd_signal(struct eventfd_ctx *ctx, int n)
e1ad7468	54	{
e1ad7468 DL	55	unsigned long flags;
	56
	57	if (n < 0)
	58	return -EINVAL;
d48eb233	59	spin_lock_irqsave(&ctx->wqh.lock, flags);
e1ad7468 DL	60	if (ULLONG_MAX - ctx->count < n)
	61	n = (int) (ULLONG_MAX - ctx->count);
	62	ctx->count += n;
	63	if (waitqueue_active(&ctx->wqh))
39510888	64	wake_up_locked_poll(&ctx->wqh, POLLIN);
d48eb233	65	spin_unlock_irqrestore(&ctx->wqh.lock, flags);
e1ad7468 DL	66
	67	return n;
	68	}
5718607b	69	EXPORT_SYMBOL_GPL(eventfd_signal);
e1ad7468	70
562787a5 DL	71	static void eventfd_free_ctx(struct eventfd_ctx *ctx)
	72	{
	73	kfree(ctx);
	74	}
	75
13389010 DL	76	static void eventfd_free(struct kref *kref)
	77	{
	78	struct eventfd_ctx *ctx = container_of(kref, struct eventfd_ctx, kref);
	79
562787a5	80	eventfd_free_ctx(ctx);
13389010 DL	81	}
	82
	83	/**
	84	* eventfd_ctx_get - Acquires a reference to the internal eventfd context.
	85	* @ctx: [in] Pointer to the eventfd context.
	86	*
	87	* Returns: In case of success, returns a pointer to the eventfd context.
	88	*/
	89	struct eventfd_ctx eventfd_ctx_get(struct eventfd_ctx ctx)
	90	{
	91	kref_get(&ctx->kref);
	92	return ctx;
	93	}
	94	EXPORT_SYMBOL_GPL(eventfd_ctx_get);
	95
	96	/**
	97	* eventfd_ctx_put - Releases a reference to the internal eventfd context.
	98	* @ctx: [in] Pointer to eventfd context.
	99	*
	100	* The eventfd context reference must have been previously acquired either
	101	* with eventfd_ctx_get() or eventfd_ctx_fdget()).
	102	*/
	103	void eventfd_ctx_put(struct eventfd_ctx *ctx)
	104	{
	105	kref_put(&ctx->kref, eventfd_free);
	106	}
	107	EXPORT_SYMBOL_GPL(eventfd_ctx_put);
	108
e1ad7468 DL	109	static int eventfd_release(struct inode inode, struct file file)
e1ad7468 DL	110	{
13389010 DL	111	struct eventfd_ctx *ctx = file->private_data;
	112
	113	wake_up_poll(&ctx->wqh, POLLHUP);
	114	eventfd_ctx_put(ctx);
e1ad7468 DL	115	return 0;
	116	}
	117
	118	static unsigned int eventfd_poll(struct file file, poll_table wait)
	119	{
	120	struct eventfd_ctx *ctx = file->private_data;
	121	unsigned int events = 0;
	122	unsigned long flags;
	123
	124	poll_wait(file, &ctx->wqh, wait);
	125
d48eb233	126	spin_lock_irqsave(&ctx->wqh.lock, flags);
e1ad7468 DL	127	if (ctx->count > 0)
	128	events \|= POLLIN;
	129	if (ctx->count == ULLONG_MAX)
	130	events \|= POLLERR;
	131	if (ULLONG_MAX - 1 > ctx->count)
	132	events \|= POLLOUT;
d48eb233	133	spin_unlock_irqrestore(&ctx->wqh.lock, flags);
e1ad7468 DL	134
	135	return events;
	136	}
	137
	138	static ssize_t eventfd_read(struct file file, char __user buf, size_t count,
	139	loff_t *ppos)
	140	{
	141	struct eventfd_ctx *ctx = file->private_data;
	142	ssize_t res;
bcd0b235	143	__u64 ucnt = 0;
e1ad7468 DL	144	DECLARE_WAITQUEUE(wait, current);
	145
	146	if (count < sizeof(ucnt))
	147	return -EINVAL;
d48eb233	148	spin_lock_irq(&ctx->wqh.lock);
e1ad7468	149	res = -EAGAIN;
bcd0b235	150	if (ctx->count > 0)
e1ad7468 DL	151	res = sizeof(ucnt);
	152	else if (!(file->f_flags & O_NONBLOCK)) {
	153	__add_wait_queue(&ctx->wqh, &wait);
	154	for (res = 0;;) {
	155	set_current_state(TASK_INTERRUPTIBLE);
	156	if (ctx->count > 0) {
e1ad7468 DL	157	res = sizeof(ucnt);
	158	break;
	159	}
	160	if (signal_pending(current)) {
	161	res = -ERESTARTSYS;
	162	break;
	163	}
d48eb233	164	spin_unlock_irq(&ctx->wqh.lock);
e1ad7468	165	schedule();
d48eb233	166	spin_lock_irq(&ctx->wqh.lock);
e1ad7468 DL	167	}
	168	__remove_wait_queue(&ctx->wqh, &wait);
	169	__set_current_state(TASK_RUNNING);
	170	}
bcd0b235 DL	171	if (likely(res > 0)) {
	172	ucnt = (ctx->flags & EFD_SEMAPHORE) ? 1 : ctx->count;
	173	ctx->count -= ucnt;
e1ad7468	174	if (waitqueue_active(&ctx->wqh))
39510888	175	wake_up_locked_poll(&ctx->wqh, POLLOUT);
e1ad7468	176	}
d48eb233	177	spin_unlock_irq(&ctx->wqh.lock);
e1ad7468 DL	178	if (res > 0 && put_user(ucnt, (__u64 __user *) buf))
	179	return -EFAULT;
	180
	181	return res;
	182	}
	183
	184	static ssize_t eventfd_write(struct file file, const char __user buf, size_t count,
	185	loff_t *ppos)
	186	{
	187	struct eventfd_ctx *ctx = file->private_data;
	188	ssize_t res;
	189	__u64 ucnt;
	190	DECLARE_WAITQUEUE(wait, current);
	191
	192	if (count < sizeof(ucnt))
	193	return -EINVAL;
	194	if (copy_from_user(&ucnt, buf, sizeof(ucnt)))
	195	return -EFAULT;
	196	if (ucnt == ULLONG_MAX)
	197	return -EINVAL;
d48eb233	198	spin_lock_irq(&ctx->wqh.lock);
e1ad7468 DL	199	res = -EAGAIN;
	200	if (ULLONG_MAX - ctx->count > ucnt)
	201	res = sizeof(ucnt);
	202	else if (!(file->f_flags & O_NONBLOCK)) {
	203	__add_wait_queue(&ctx->wqh, &wait);
	204	for (res = 0;;) {
	205	set_current_state(TASK_INTERRUPTIBLE);
	206	if (ULLONG_MAX - ctx->count > ucnt) {
	207	res = sizeof(ucnt);
	208	break;
	209	}
	210	if (signal_pending(current)) {
	211	res = -ERESTARTSYS;
	212	break;
	213	}
d48eb233	214	spin_unlock_irq(&ctx->wqh.lock);
e1ad7468	215	schedule();
d48eb233	216	spin_lock_irq(&ctx->wqh.lock);
e1ad7468 DL	217	}
	218	__remove_wait_queue(&ctx->wqh, &wait);
	219	__set_current_state(TASK_RUNNING);
	220	}
bcd0b235	221	if (likely(res > 0)) {
e1ad7468 DL	222	ctx->count += ucnt;
e1ad7468 DL	223	if (waitqueue_active(&ctx->wqh))
39510888	224	wake_up_locked_poll(&ctx->wqh, POLLIN);
e1ad7468	225	}
d48eb233	226	spin_unlock_irq(&ctx->wqh.lock);
e1ad7468 DL	227
	228	return res;
	229	}
	230
	231	static const struct file_operations eventfd_fops = {
	232	.release = eventfd_release,
	233	.poll = eventfd_poll,
	234	.read = eventfd_read,
	235	.write = eventfd_write,
	236	};
	237
13389010 DL	238	/**
	239	* eventfd_fget - Acquire a reference of an eventfd file descriptor.
	240	* @fd: [in] Eventfd file descriptor.
	241	*
	242	* Returns a pointer to the eventfd file structure in case of success, or the
	243	* following error pointer:
	244	*
	245	* -EBADF : Invalid @fd file descriptor.
	246	* -EINVAL : The @fd file descriptor is not an eventfd file.
	247	*/
e1ad7468 DL	248	struct file *eventfd_fget(int fd)
	249	{
	250	struct file *file;
	251
	252	file = fget(fd);
	253	if (!file)
	254	return ERR_PTR(-EBADF);
	255	if (file->f_op != &eventfd_fops) {
	256	fput(file);
	257	return ERR_PTR(-EINVAL);
	258	}
	259
	260	return file;
	261	}
5718607b	262	EXPORT_SYMBOL_GPL(eventfd_fget);
e1ad7468	263
13389010 DL	264	/**
	265	* eventfd_ctx_fdget - Acquires a reference to the internal eventfd context.
	266	* @fd: [in] Eventfd file descriptor.
	267	*
	268	* Returns a pointer to the internal eventfd context, otherwise the error
	269	* pointers returned by the following functions:
	270	*
	271	* eventfd_fget
	272	*/
	273	struct eventfd_ctx *eventfd_ctx_fdget(int fd)
	274	{
	275	struct file *file;
	276	struct eventfd_ctx *ctx;
	277
	278	file = eventfd_fget(fd);
	279	if (IS_ERR(file))
	280	return (struct eventfd_ctx *) file;
	281	ctx = eventfd_ctx_get(file->private_data);
	282	fput(file);
	283
	284	return ctx;
	285	}
	286	EXPORT_SYMBOL_GPL(eventfd_ctx_fdget);
	287
	288	/**
	289	* eventfd_ctx_fileget - Acquires a reference to the internal eventfd context.
	290	* @file: [in] Eventfd file pointer.
	291	*
	292	* Returns a pointer to the internal eventfd context, otherwise the error
	293	* pointer:
	294	*
	295	* -EINVAL : The @fd file descriptor is not an eventfd file.
	296	*/
	297	struct eventfd_ctx eventfd_ctx_fileget(struct file file)
	298	{
	299	if (file->f_op != &eventfd_fops)
	300	return ERR_PTR(-EINVAL);
	301
	302	return eventfd_ctx_get(file->private_data);
	303	}
	304	EXPORT_SYMBOL_GPL(eventfd_ctx_fileget);
	305
562787a5 DL	306	/**
	307	* eventfd_file_create - Creates an eventfd file pointer.
	308	* @count: Initial eventfd counter value.
	309	* @flags: Flags for the eventfd file.
	310	*
	311	* This function creates an eventfd file pointer, w/out installing it into
	312	* the fd table. This is useful when the eventfd file is used during the
	313	* initialization of data structures that require extra setup after the eventfd
	314	* creation. So the eventfd creation is split into the file pointer creation
	315	* phase, and the file descriptor installation phase.
	316	* In this way races with userspace closing the newly installed file descriptor
	317	* can be avoided.
	318	* Returns an eventfd file pointer, or a proper error pointer.
	319	*/
	320	struct file *eventfd_file_create(unsigned int count, int flags)
e1ad7468	321	{
562787a5	322	struct file *file;
e1ad7468	323	struct eventfd_ctx *ctx;
e1ad7468	324
e38b36f3 UD	325	/* Check the EFD_* constants for consistency. */
	326	BUILD_BUG_ON(EFD_CLOEXEC != O_CLOEXEC);
	327	BUILD_BUG_ON(EFD_NONBLOCK != O_NONBLOCK);
	328
bcd0b235	329	if (flags & ~EFD_FLAGS_SET)
562787a5	330	return ERR_PTR(-EINVAL);
b087498e	331
e1ad7468 DL	332	ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
e1ad7468 DL	333	if (!ctx)
562787a5	334	return ERR_PTR(-ENOMEM);
e1ad7468	335
13389010	336	kref_init(&ctx->kref);
e1ad7468	337	init_waitqueue_head(&ctx->wqh);
e1ad7468	338	ctx->count = count;
bcd0b235	339	ctx->flags = flags;
e1ad7468	340
562787a5 DL	341	file = anon_inode_getfile("[eventfd]", &eventfd_fops, ctx,
	342	flags & EFD_SHARED_FCNTL_FLAGS);
	343	if (IS_ERR(file))
	344	eventfd_free_ctx(ctx);
	345
	346	return file;
	347	}
	348
	349	SYSCALL_DEFINE2(eventfd2, unsigned int, count, int, flags)
	350	{
	351	int fd, error;
	352	struct file *file;
	353
	354	error = get_unused_fd_flags(flags & EFD_SHARED_FCNTL_FLAGS);
	355	if (error < 0)
	356	return error;
	357	fd = error;
	358
	359	file = eventfd_file_create(count, flags);
	360	if (IS_ERR(file)) {
	361	error = PTR_ERR(file);
	362	goto err_put_unused_fd;
	363	}
	364	fd_install(fd, file);
	365
2030a42c	366	return fd;
562787a5 DL	367
	368	err_put_unused_fd:
	369	put_unused_fd(fd);
	370
	371	return error;
e1ad7468 DL	372	}
e1ad7468 DL	373
d4e82042	374	SYSCALL_DEFINE1(eventfd, unsigned int, count)
b087498e UD	375	{
	376	return sys_eventfd2(count, 0);
	377	}
bcd0b235	378