[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/slab.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 void eventfd_ctx_do_read(struct eventfd_ctx *ctx, __u64 *cnt)
{
	*cnt = (ctx->flags & EFD_SEMAPHORE) ? 1 : ctx->count;
	ctx->count -= *cnt;
}

/**
 * eventfd_ctx_remove_wait_queue - Read the current counter and removes wait queue.
 * @ctx: [in] Pointer to eventfd context.
 * @wait: [in] Wait queue to be removed.
 * @cnt: [out] Pointer to the 64bit conter value.
 *
 * Returns zero if successful, or the following error codes:
 *
 * -EAGAIN      : The operation would have blocked.
 *
 * This is used to atomically remove a wait queue entry from the eventfd wait
 * queue head, and read/reset the counter value.
 */
int eventfd_ctx_remove_wait_queue(struct eventfd_ctx *ctx, wait_queue_t *wait,
				  __u64 *cnt)
{
	unsigned long flags;

	spin_lock_irqsave(&ctx->wqh.lock, flags);
	eventfd_ctx_do_read(ctx, cnt);
	__remove_wait_queue(&ctx->wqh, wait);
	if (*cnt != 0 && waitqueue_active(&ctx->wqh))
		wake_up_locked_poll(&ctx->wqh, POLLOUT);
	spin_unlock_irqrestore(&ctx->wqh.lock, flags);

	return *cnt != 0 ? 0 : -EAGAIN;
}
EXPORT_SYMBOL_GPL(eventfd_ctx_remove_wait_queue);

/**
 * eventfd_ctx_read - Reads the eventfd counter or wait if it is zero.
 * @ctx: [in] Pointer to eventfd context.
 * @no_wait: [in] Different from zero if the operation should not block.
 * @cnt: [out] Pointer to the 64bit conter value.
 *
 * Returns zero if successful, or the following error codes:
 *
 * -EAGAIN      : The operation would have blocked but @no_wait was nonzero.
 * -ERESTARTSYS : A signal interrupted the wait operation.
 *
 * If @no_wait is zero, the function might sleep until the eventfd internal
 * counter becomes greater than zero.
 */
ssize_t eventfd_ctx_read(struct eventfd_ctx *ctx, int no_wait, __u64 *cnt)
{
	ssize_t res;
	DECLARE_WAITQUEUE(wait, current);

	spin_lock_irq(&ctx->wqh.lock);
	*cnt = 0;
	res = -EAGAIN;
	if (ctx->count > 0)
		res = 0;
	else if (!no_wait) {
		__add_wait_queue(&ctx->wqh, &wait);
		for (;;) {
			set_current_state(TASK_INTERRUPTIBLE);
			if (ctx->count > 0) {
				res = 0;
				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)) {
		eventfd_ctx_do_read(ctx, cnt);
		if (waitqueue_active(&ctx->wqh))
			wake_up_locked_poll(&ctx->wqh, POLLOUT);
	}
	spin_unlock_irq(&ctx->wqh.lock);

	return res;
}
EXPORT_SYMBOL_GPL(eventfd_ctx_read);

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 cnt;

	if (count < sizeof(cnt))
		return -EINVAL;
	res = eventfd_ctx_read(ctx, file->f_flags & O_NONBLOCK, &cnt);
	if (res < 0)
		return res;

	return put_user(cnt, (__u64 __user *) buf) ? -EFAULT : sizeof(cnt);
}

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,
				  O_RDWR | (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>
5a0e3ad6	14	#include <linux/slab.h>
e1ad7468 DL	15	#include <linux/list.h>
	16	#include <linux/spinlock.h>
	17	#include <linux/anon_inodes.h>
7747cdb2	18	#include <linux/syscalls.h>
5718607b	19	#include <linux/module.h>
13389010 DL	20	#include <linux/kref.h>
13389010 DL	21	#include <linux/eventfd.h>
e1ad7468 DL	22
e1ad7468 DL	23	struct eventfd_ctx {
13389010	24	struct kref kref;
e1ad7468 DL	25	wait_queue_head_t wqh;
	26	/*
	27	* Every time that a write(2) is performed on an eventfd, the
	28	* value of the __u64 being written is added to "count" and a
	29	* wakeup is performed on "wqh". A read(2) will return the "count"
	30	* value to userspace, and will reset "count" to zero. The kernel
13389010	31	* side eventfd_signal() also, adds to the "count" counter and
e1ad7468 DL	32	* issue a wakeup.
	33	*/
	34	__u64 count;
bcd0b235	35	unsigned int flags;
e1ad7468 DL	36	};
e1ad7468 DL	37
13389010 DL	38	/**
	39	* eventfd_signal - Adds @n to the eventfd counter.
	40	* @ctx: [in] Pointer to the eventfd context.
	41	* @n: [in] Value of the counter to be added to the eventfd internal counter.
	42	* The value cannot be negative.
	43	*
	44	* This function is supposed to be called by the kernel in paths that do not
	45	* allow sleeping. In this function we allow the counter to reach the ULLONG_MAX
	46	* value, and we signal this as overflow condition by returining a POLLERR
	47	* to poll(2).
	48	*
	49	* Returns @n in case of success, a non-negative number lower than @n in case
	50	* of overflow, or the following error codes:
	51	*
	52	* -EINVAL : The value of @n is negative.
e1ad7468	53	*/
13389010	54	int eventfd_signal(struct eventfd_ctx *ctx, int n)
e1ad7468	55	{
e1ad7468 DL	56	unsigned long flags;
	57
	58	if (n < 0)
	59	return -EINVAL;
d48eb233	60	spin_lock_irqsave(&ctx->wqh.lock, flags);
e1ad7468 DL	61	if (ULLONG_MAX - ctx->count < n)
	62	n = (int) (ULLONG_MAX - ctx->count);
	63	ctx->count += n;
	64	if (waitqueue_active(&ctx->wqh))
39510888	65	wake_up_locked_poll(&ctx->wqh, POLLIN);
d48eb233	66	spin_unlock_irqrestore(&ctx->wqh.lock, flags);
e1ad7468 DL	67
	68	return n;
	69	}
5718607b	70	EXPORT_SYMBOL_GPL(eventfd_signal);
e1ad7468	71
562787a5 DL	72	static void eventfd_free_ctx(struct eventfd_ctx *ctx)
	73	{
	74	kfree(ctx);
	75	}
	76
13389010 DL	77	static void eventfd_free(struct kref *kref)
	78	{
	79	struct eventfd_ctx *ctx = container_of(kref, struct eventfd_ctx, kref);
	80
562787a5	81	eventfd_free_ctx(ctx);
13389010 DL	82	}
	83
	84	/**
	85	* eventfd_ctx_get - Acquires a reference to the internal eventfd context.
	86	* @ctx: [in] Pointer to the eventfd context.
	87	*
	88	* Returns: In case of success, returns a pointer to the eventfd context.
	89	*/
	90	struct eventfd_ctx eventfd_ctx_get(struct eventfd_ctx ctx)
	91	{
	92	kref_get(&ctx->kref);
	93	return ctx;
	94	}
	95	EXPORT_SYMBOL_GPL(eventfd_ctx_get);
	96
	97	/**
	98	* eventfd_ctx_put - Releases a reference to the internal eventfd context.
	99	* @ctx: [in] Pointer to eventfd context.
	100	*
	101	* The eventfd context reference must have been previously acquired either
	102	* with eventfd_ctx_get() or eventfd_ctx_fdget()).
	103	*/
	104	void eventfd_ctx_put(struct eventfd_ctx *ctx)
	105	{
	106	kref_put(&ctx->kref, eventfd_free);
	107	}
	108	EXPORT_SYMBOL_GPL(eventfd_ctx_put);
	109
e1ad7468 DL	110	static int eventfd_release(struct inode inode, struct file file)
e1ad7468 DL	111	{
13389010 DL	112	struct eventfd_ctx *ctx = file->private_data;
	113
	114	wake_up_poll(&ctx->wqh, POLLHUP);
	115	eventfd_ctx_put(ctx);
e1ad7468 DL	116	return 0;
	117	}
	118
	119	static unsigned int eventfd_poll(struct file file, poll_table wait)
	120	{
	121	struct eventfd_ctx *ctx = file->private_data;
	122	unsigned int events = 0;
	123	unsigned long flags;
	124
	125	poll_wait(file, &ctx->wqh, wait);
	126
d48eb233	127	spin_lock_irqsave(&ctx->wqh.lock, flags);
e1ad7468 DL	128	if (ctx->count > 0)
	129	events \|= POLLIN;
	130	if (ctx->count == ULLONG_MAX)
	131	events \|= POLLERR;
	132	if (ULLONG_MAX - 1 > ctx->count)
	133	events \|= POLLOUT;
d48eb233	134	spin_unlock_irqrestore(&ctx->wqh.lock, flags);
e1ad7468 DL	135
	136	return events;
	137	}
	138
cb289d62 DL	139	static void eventfd_ctx_do_read(struct eventfd_ctx ctx, __u64 cnt)
	140	{
	141	*cnt = (ctx->flags & EFD_SEMAPHORE) ? 1 : ctx->count;
	142	ctx->count -= *cnt;
	143	}
	144
	145	/**
	146	* eventfd_ctx_remove_wait_queue - Read the current counter and removes wait queue.
	147	* @ctx: [in] Pointer to eventfd context.
	148	* @wait: [in] Wait queue to be removed.
	149	* @cnt: [out] Pointer to the 64bit conter value.
	150	*
	151	* Returns zero if successful, or the following error codes:
	152	*
	153	* -EAGAIN : The operation would have blocked.
	154	*
	155	* This is used to atomically remove a wait queue entry from the eventfd wait
	156	* queue head, and read/reset the counter value.
	157	*/
	158	int eventfd_ctx_remove_wait_queue(struct eventfd_ctx ctx, wait_queue_t wait,
	159	__u64 *cnt)
	160	{
	161	unsigned long flags;
	162
	163	spin_lock_irqsave(&ctx->wqh.lock, flags);
	164	eventfd_ctx_do_read(ctx, cnt);
	165	__remove_wait_queue(&ctx->wqh, wait);
	166	if (*cnt != 0 && waitqueue_active(&ctx->wqh))
	167	wake_up_locked_poll(&ctx->wqh, POLLOUT);
	168	spin_unlock_irqrestore(&ctx->wqh.lock, flags);
	169
	170	return *cnt != 0 ? 0 : -EAGAIN;
	171	}
	172	EXPORT_SYMBOL_GPL(eventfd_ctx_remove_wait_queue);
	173
	174	/**
	175	* eventfd_ctx_read - Reads the eventfd counter or wait if it is zero.
	176	* @ctx: [in] Pointer to eventfd context.
	177	* @no_wait: [in] Different from zero if the operation should not block.
	178	* @cnt: [out] Pointer to the 64bit conter value.
	179	*
	180	* Returns zero if successful, or the following error codes:
	181	*
	182	* -EAGAIN : The operation would have blocked but @no_wait was nonzero.
	183	* -ERESTARTSYS : A signal interrupted the wait operation.
	184	*
	185	* If @no_wait is zero, the function might sleep until the eventfd internal
	186	* counter becomes greater than zero.
	187	*/
	188	ssize_t eventfd_ctx_read(struct eventfd_ctx ctx, int no_wait, __u64 cnt)
e1ad7468	189	{
e1ad7468	190	ssize_t res;
e1ad7468 DL	191	DECLARE_WAITQUEUE(wait, current);
e1ad7468 DL	192
d48eb233	193	spin_lock_irq(&ctx->wqh.lock);
cb289d62	194	*cnt = 0;
e1ad7468	195	res = -EAGAIN;
bcd0b235	196	if (ctx->count > 0)
cb289d62 DL	197	res = 0;
cb289d62 DL	198	else if (!no_wait) {
e1ad7468	199	__add_wait_queue(&ctx->wqh, &wait);
cb289d62	200	for (;;) {
e1ad7468 DL	201	set_current_state(TASK_INTERRUPTIBLE);
e1ad7468 DL	202	if (ctx->count > 0) {
cb289d62	203	res = 0;
e1ad7468 DL	204	break;
	205	}
	206	if (signal_pending(current)) {
	207	res = -ERESTARTSYS;
	208	break;
	209	}
d48eb233	210	spin_unlock_irq(&ctx->wqh.lock);
e1ad7468	211	schedule();
d48eb233	212	spin_lock_irq(&ctx->wqh.lock);
e1ad7468 DL	213	}
	214	__remove_wait_queue(&ctx->wqh, &wait);
	215	__set_current_state(TASK_RUNNING);
	216	}
cb289d62 DL	217	if (likely(res == 0)) {
cb289d62 DL	218	eventfd_ctx_do_read(ctx, cnt);
e1ad7468	219	if (waitqueue_active(&ctx->wqh))
39510888	220	wake_up_locked_poll(&ctx->wqh, POLLOUT);
e1ad7468	221	}
d48eb233	222	spin_unlock_irq(&ctx->wqh.lock);
e1ad7468 DL	223
	224	return res;
	225	}
cb289d62 DL	226	EXPORT_SYMBOL_GPL(eventfd_ctx_read);
	227
	228	static ssize_t eventfd_read(struct file file, char __user buf, size_t count,
	229	loff_t *ppos)
	230	{
	231	struct eventfd_ctx *ctx = file->private_data;
	232	ssize_t res;
	233	__u64 cnt;
	234
	235	if (count < sizeof(cnt))
	236	return -EINVAL;
	237	res = eventfd_ctx_read(ctx, file->f_flags & O_NONBLOCK, &cnt);
	238	if (res < 0)
	239	return res;
	240
	241	return put_user(cnt, (__u64 __user *) buf) ? -EFAULT : sizeof(cnt);
	242	}
e1ad7468 DL	243
	244	static ssize_t eventfd_write(struct file file, const char __user buf, size_t count,
	245	loff_t *ppos)
	246	{
	247	struct eventfd_ctx *ctx = file->private_data;
	248	ssize_t res;
	249	__u64 ucnt;
	250	DECLARE_WAITQUEUE(wait, current);
	251
	252	if (count < sizeof(ucnt))
	253	return -EINVAL;
	254	if (copy_from_user(&ucnt, buf, sizeof(ucnt)))
	255	return -EFAULT;
	256	if (ucnt == ULLONG_MAX)
	257	return -EINVAL;
d48eb233	258	spin_lock_irq(&ctx->wqh.lock);
e1ad7468 DL	259	res = -EAGAIN;
	260	if (ULLONG_MAX - ctx->count > ucnt)
	261	res = sizeof(ucnt);
	262	else if (!(file->f_flags & O_NONBLOCK)) {
	263	__add_wait_queue(&ctx->wqh, &wait);
	264	for (res = 0;;) {
	265	set_current_state(TASK_INTERRUPTIBLE);
	266	if (ULLONG_MAX - ctx->count > ucnt) {
	267	res = sizeof(ucnt);
	268	break;
	269	}
	270	if (signal_pending(current)) {
	271	res = -ERESTARTSYS;
	272	break;
	273	}
d48eb233	274	spin_unlock_irq(&ctx->wqh.lock);
e1ad7468	275	schedule();
d48eb233	276	spin_lock_irq(&ctx->wqh.lock);
e1ad7468 DL	277	}
	278	__remove_wait_queue(&ctx->wqh, &wait);
	279	__set_current_state(TASK_RUNNING);
	280	}
bcd0b235	281	if (likely(res > 0)) {
e1ad7468 DL	282	ctx->count += ucnt;
e1ad7468 DL	283	if (waitqueue_active(&ctx->wqh))
39510888	284	wake_up_locked_poll(&ctx->wqh, POLLIN);
e1ad7468	285	}
d48eb233	286	spin_unlock_irq(&ctx->wqh.lock);
e1ad7468 DL	287
	288	return res;
	289	}
	290
	291	static const struct file_operations eventfd_fops = {
	292	.release = eventfd_release,
	293	.poll = eventfd_poll,
	294	.read = eventfd_read,
	295	.write = eventfd_write,
	296	};
	297
13389010 DL	298	/**
	299	* eventfd_fget - Acquire a reference of an eventfd file descriptor.
	300	* @fd: [in] Eventfd file descriptor.
	301	*
	302	* Returns a pointer to the eventfd file structure in case of success, or the
	303	* following error pointer:
	304	*
	305	* -EBADF : Invalid @fd file descriptor.
	306	* -EINVAL : The @fd file descriptor is not an eventfd file.
	307	*/
e1ad7468 DL	308	struct file *eventfd_fget(int fd)
	309	{
	310	struct file *file;
	311
	312	file = fget(fd);
	313	if (!file)
	314	return ERR_PTR(-EBADF);
	315	if (file->f_op != &eventfd_fops) {
	316	fput(file);
	317	return ERR_PTR(-EINVAL);
	318	}
	319
	320	return file;
	321	}
5718607b	322	EXPORT_SYMBOL_GPL(eventfd_fget);
e1ad7468	323
13389010 DL	324	/**
	325	* eventfd_ctx_fdget - Acquires a reference to the internal eventfd context.
	326	* @fd: [in] Eventfd file descriptor.
	327	*
	328	* Returns a pointer to the internal eventfd context, otherwise the error
	329	* pointers returned by the following functions:
	330	*
	331	* eventfd_fget
	332	*/
	333	struct eventfd_ctx *eventfd_ctx_fdget(int fd)
	334	{
	335	struct file *file;
	336	struct eventfd_ctx *ctx;
	337
	338	file = eventfd_fget(fd);
	339	if (IS_ERR(file))
	340	return (struct eventfd_ctx *) file;
	341	ctx = eventfd_ctx_get(file->private_data);
	342	fput(file);
	343
	344	return ctx;
	345	}
	346	EXPORT_SYMBOL_GPL(eventfd_ctx_fdget);
	347
	348	/**
	349	* eventfd_ctx_fileget - Acquires a reference to the internal eventfd context.
	350	* @file: [in] Eventfd file pointer.
	351	*
	352	* Returns a pointer to the internal eventfd context, otherwise the error
	353	* pointer:
	354	*
	355	* -EINVAL : The @fd file descriptor is not an eventfd file.
	356	*/
	357	struct eventfd_ctx eventfd_ctx_fileget(struct file file)
	358	{
	359	if (file->f_op != &eventfd_fops)
	360	return ERR_PTR(-EINVAL);
	361
	362	return eventfd_ctx_get(file->private_data);
	363	}
	364	EXPORT_SYMBOL_GPL(eventfd_ctx_fileget);
	365
562787a5 DL	366	/**
	367	* eventfd_file_create - Creates an eventfd file pointer.
	368	* @count: Initial eventfd counter value.
	369	* @flags: Flags for the eventfd file.
	370	*
	371	* This function creates an eventfd file pointer, w/out installing it into
	372	* the fd table. This is useful when the eventfd file is used during the
	373	* initialization of data structures that require extra setup after the eventfd
	374	* creation. So the eventfd creation is split into the file pointer creation
	375	* phase, and the file descriptor installation phase.
	376	* In this way races with userspace closing the newly installed file descriptor
	377	* can be avoided.
	378	* Returns an eventfd file pointer, or a proper error pointer.
	379	*/
	380	struct file *eventfd_file_create(unsigned int count, int flags)
e1ad7468	381	{
562787a5	382	struct file *file;
e1ad7468	383	struct eventfd_ctx *ctx;
e1ad7468	384
e38b36f3 UD	385	/* Check the EFD_* constants for consistency. */
	386	BUILD_BUG_ON(EFD_CLOEXEC != O_CLOEXEC);
	387	BUILD_BUG_ON(EFD_NONBLOCK != O_NONBLOCK);
	388
bcd0b235	389	if (flags & ~EFD_FLAGS_SET)
562787a5	390	return ERR_PTR(-EINVAL);
b087498e	391
e1ad7468 DL	392	ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
e1ad7468 DL	393	if (!ctx)
562787a5	394	return ERR_PTR(-ENOMEM);
e1ad7468	395
13389010	396	kref_init(&ctx->kref);
e1ad7468	397	init_waitqueue_head(&ctx->wqh);
e1ad7468	398	ctx->count = count;
bcd0b235	399	ctx->flags = flags;
e1ad7468	400
562787a5	401	file = anon_inode_getfile("[eventfd]", &eventfd_fops, ctx,
628ff7c1	402	O_RDWR \| (flags & EFD_SHARED_FCNTL_FLAGS));
562787a5 DL	403	if (IS_ERR(file))
	404	eventfd_free_ctx(ctx);
	405
	406	return file;
	407	}
	408
	409	SYSCALL_DEFINE2(eventfd2, unsigned int, count, int, flags)
	410	{
	411	int fd, error;
	412	struct file *file;
	413
	414	error = get_unused_fd_flags(flags & EFD_SHARED_FCNTL_FLAGS);
	415	if (error < 0)
	416	return error;
	417	fd = error;
	418
	419	file = eventfd_file_create(count, flags);
	420	if (IS_ERR(file)) {
	421	error = PTR_ERR(file);
	422	goto err_put_unused_fd;
	423	}
	424	fd_install(fd, file);
	425
2030a42c	426	return fd;
562787a5 DL	427
	428	err_put_unused_fd:
	429	put_unused_fd(fd);
	430
	431	return error;
e1ad7468 DL	432	}
e1ad7468 DL	433
d4e82042	434	SYSCALL_DEFINE1(eventfd, unsigned int, count)
b087498e UD	435	{
	436	return sys_eventfd2(count, 0);
	437	}
bcd0b235	438