[linux.git] / fs / eventfd.c

/*
 *  fs/eventfd.c
 *
 *  Copyright (C) 2007  Davide Libenzi <[email protected]>
 *
 */

#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/export.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 the amount by which the counter was incrememnted.  This will be less
 * than @n if the counter has overflowed.
 */
__u64 eventfd_signal(struct eventfd_ctx *ctx, __u64 n)
{
	unsigned long flags;

	spin_lock_irqsave(&ctx->wqh.lock, flags);
	if (ULLONG_MAX - ctx->count < n)
		n = 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 64-bit counter value.
 *
 * Returns %0 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 64-bit counter value.
 *
 * Returns %0 if successful, or the following error codes:
 *
 * -EAGAIN      : The operation would have blocked but @no_wait was non-zero.
 * -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,
	.llseek		= noop_llseek,
};

/**
 * 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 <[email protected]>
	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>
630d9c47	19	#include <linux/export.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	*
ee62c6b2 SZ	49	* Returns the amount by which the counter was incrememnted. This will be less
ee62c6b2 SZ	50	* than @n if the counter has overflowed.
e1ad7468	51	*/
ee62c6b2	52	__u64 eventfd_signal(struct eventfd_ctx *ctx, __u64 n)
e1ad7468	53	{
e1ad7468 DL	54	unsigned long flags;
e1ad7468 DL	55
d48eb233	56	spin_lock_irqsave(&ctx->wqh.lock, flags);
e1ad7468	57	if (ULLONG_MAX - ctx->count < n)
ee62c6b2	58	n = ULLONG_MAX - ctx->count;
e1ad7468 DL	59	ctx->count += n;
e1ad7468 DL	60	if (waitqueue_active(&ctx->wqh))
39510888	61	wake_up_locked_poll(&ctx->wqh, POLLIN);
d48eb233	62	spin_unlock_irqrestore(&ctx->wqh.lock, flags);
e1ad7468 DL	63
	64	return n;
	65	}
5718607b	66	EXPORT_SYMBOL_GPL(eventfd_signal);
e1ad7468	67
562787a5 DL	68	static void eventfd_free_ctx(struct eventfd_ctx *ctx)
	69	{
	70	kfree(ctx);
	71	}
	72
13389010 DL	73	static void eventfd_free(struct kref *kref)
	74	{
	75	struct eventfd_ctx *ctx = container_of(kref, struct eventfd_ctx, kref);
	76
562787a5	77	eventfd_free_ctx(ctx);
13389010 DL	78	}
	79
	80	/**
	81	* eventfd_ctx_get - Acquires a reference to the internal eventfd context.
	82	* @ctx: [in] Pointer to the eventfd context.
	83	*
	84	* Returns: In case of success, returns a pointer to the eventfd context.
	85	*/
	86	struct eventfd_ctx eventfd_ctx_get(struct eventfd_ctx ctx)
	87	{
	88	kref_get(&ctx->kref);
	89	return ctx;
	90	}
	91	EXPORT_SYMBOL_GPL(eventfd_ctx_get);
	92
	93	/**
	94	* eventfd_ctx_put - Releases a reference to the internal eventfd context.
	95	* @ctx: [in] Pointer to eventfd context.
	96	*
	97	* The eventfd context reference must have been previously acquired either
36182185	98	* with eventfd_ctx_get() or eventfd_ctx_fdget().
13389010 DL	99	*/
	100	void eventfd_ctx_put(struct eventfd_ctx *ctx)
	101	{
	102	kref_put(&ctx->kref, eventfd_free);
	103	}
	104	EXPORT_SYMBOL_GPL(eventfd_ctx_put);
	105
e1ad7468 DL	106	static int eventfd_release(struct inode inode, struct file file)
e1ad7468 DL	107	{
13389010 DL	108	struct eventfd_ctx *ctx = file->private_data;
	109
	110	wake_up_poll(&ctx->wqh, POLLHUP);
	111	eventfd_ctx_put(ctx);
e1ad7468 DL	112	return 0;
	113	}
	114
	115	static unsigned int eventfd_poll(struct file file, poll_table wait)
	116	{
	117	struct eventfd_ctx *ctx = file->private_data;
	118	unsigned int events = 0;
	119	unsigned long flags;
	120
	121	poll_wait(file, &ctx->wqh, wait);
	122
d48eb233	123	spin_lock_irqsave(&ctx->wqh.lock, flags);
e1ad7468 DL	124	if (ctx->count > 0)
	125	events \|= POLLIN;
	126	if (ctx->count == ULLONG_MAX)
	127	events \|= POLLERR;
	128	if (ULLONG_MAX - 1 > ctx->count)
	129	events \|= POLLOUT;
d48eb233	130	spin_unlock_irqrestore(&ctx->wqh.lock, flags);
e1ad7468 DL	131
	132	return events;
	133	}
	134
cb289d62 DL	135	static void eventfd_ctx_do_read(struct eventfd_ctx ctx, __u64 cnt)
	136	{
	137	*cnt = (ctx->flags & EFD_SEMAPHORE) ? 1 : ctx->count;
	138	ctx->count -= *cnt;
	139	}
	140
	141	/**
	142	* eventfd_ctx_remove_wait_queue - Read the current counter and removes wait queue.
	143	* @ctx: [in] Pointer to eventfd context.
	144	* @wait: [in] Wait queue to be removed.
36182185	145	* @cnt: [out] Pointer to the 64-bit counter value.
cb289d62	146	*
36182185	147	* Returns %0 if successful, or the following error codes:
cb289d62 DL	148	*
	149	* -EAGAIN : The operation would have blocked.
	150	*
	151	* This is used to atomically remove a wait queue entry from the eventfd wait
	152	* queue head, and read/reset the counter value.
	153	*/
	154	int eventfd_ctx_remove_wait_queue(struct eventfd_ctx ctx, wait_queue_t wait,
	155	__u64 *cnt)
	156	{
	157	unsigned long flags;
	158
	159	spin_lock_irqsave(&ctx->wqh.lock, flags);
	160	eventfd_ctx_do_read(ctx, cnt);
	161	__remove_wait_queue(&ctx->wqh, wait);
	162	if (*cnt != 0 && waitqueue_active(&ctx->wqh))
	163	wake_up_locked_poll(&ctx->wqh, POLLOUT);
	164	spin_unlock_irqrestore(&ctx->wqh.lock, flags);
	165
	166	return *cnt != 0 ? 0 : -EAGAIN;
	167	}
	168	EXPORT_SYMBOL_GPL(eventfd_ctx_remove_wait_queue);
	169
	170	/**
	171	* eventfd_ctx_read - Reads the eventfd counter or wait if it is zero.
	172	* @ctx: [in] Pointer to eventfd context.
	173	* @no_wait: [in] Different from zero if the operation should not block.
36182185	174	* @cnt: [out] Pointer to the 64-bit counter value.
cb289d62	175	*
36182185	176	* Returns %0 if successful, or the following error codes:
cb289d62	177	*
36182185	178	* -EAGAIN : The operation would have blocked but @no_wait was non-zero.
cb289d62 DL	179	* -ERESTARTSYS : A signal interrupted the wait operation.
	180	*
	181	* If @no_wait is zero, the function might sleep until the eventfd internal
	182	* counter becomes greater than zero.
	183	*/
	184	ssize_t eventfd_ctx_read(struct eventfd_ctx ctx, int no_wait, __u64 cnt)
e1ad7468	185	{
e1ad7468	186	ssize_t res;
e1ad7468 DL	187	DECLARE_WAITQUEUE(wait, current);
e1ad7468 DL	188
d48eb233	189	spin_lock_irq(&ctx->wqh.lock);
cb289d62	190	*cnt = 0;
e1ad7468	191	res = -EAGAIN;
bcd0b235	192	if (ctx->count > 0)
cb289d62 DL	193	res = 0;
cb289d62 DL	194	else if (!no_wait) {
e1ad7468	195	__add_wait_queue(&ctx->wqh, &wait);
cb289d62	196	for (;;) {
e1ad7468 DL	197	set_current_state(TASK_INTERRUPTIBLE);
e1ad7468 DL	198	if (ctx->count > 0) {
cb289d62	199	res = 0;
e1ad7468 DL	200	break;
	201	}
	202	if (signal_pending(current)) {
	203	res = -ERESTARTSYS;
	204	break;
	205	}
d48eb233	206	spin_unlock_irq(&ctx->wqh.lock);
e1ad7468	207	schedule();
d48eb233	208	spin_lock_irq(&ctx->wqh.lock);
e1ad7468 DL	209	}
	210	__remove_wait_queue(&ctx->wqh, &wait);
	211	__set_current_state(TASK_RUNNING);
	212	}
cb289d62 DL	213	if (likely(res == 0)) {
cb289d62 DL	214	eventfd_ctx_do_read(ctx, cnt);
e1ad7468	215	if (waitqueue_active(&ctx->wqh))
39510888	216	wake_up_locked_poll(&ctx->wqh, POLLOUT);
e1ad7468	217	}
d48eb233	218	spin_unlock_irq(&ctx->wqh.lock);
e1ad7468 DL	219
	220	return res;
	221	}
cb289d62 DL	222	EXPORT_SYMBOL_GPL(eventfd_ctx_read);
	223
	224	static ssize_t eventfd_read(struct file file, char __user buf, size_t count,
	225	loff_t *ppos)
	226	{
	227	struct eventfd_ctx *ctx = file->private_data;
	228	ssize_t res;
	229	__u64 cnt;
	230
	231	if (count < sizeof(cnt))
	232	return -EINVAL;
	233	res = eventfd_ctx_read(ctx, file->f_flags & O_NONBLOCK, &cnt);
	234	if (res < 0)
	235	return res;
	236
	237	return put_user(cnt, (__u64 __user *) buf) ? -EFAULT : sizeof(cnt);
	238	}
e1ad7468 DL	239
	240	static ssize_t eventfd_write(struct file file, const char __user buf, size_t count,
	241	loff_t *ppos)
	242	{
	243	struct eventfd_ctx *ctx = file->private_data;
	244	ssize_t res;
	245	__u64 ucnt;
	246	DECLARE_WAITQUEUE(wait, current);
	247
	248	if (count < sizeof(ucnt))
	249	return -EINVAL;
	250	if (copy_from_user(&ucnt, buf, sizeof(ucnt)))
	251	return -EFAULT;
	252	if (ucnt == ULLONG_MAX)
	253	return -EINVAL;
d48eb233	254	spin_lock_irq(&ctx->wqh.lock);
e1ad7468 DL	255	res = -EAGAIN;
	256	if (ULLONG_MAX - ctx->count > ucnt)
	257	res = sizeof(ucnt);
	258	else if (!(file->f_flags & O_NONBLOCK)) {
	259	__add_wait_queue(&ctx->wqh, &wait);
	260	for (res = 0;;) {
	261	set_current_state(TASK_INTERRUPTIBLE);
	262	if (ULLONG_MAX - ctx->count > ucnt) {
	263	res = sizeof(ucnt);
	264	break;
	265	}
	266	if (signal_pending(current)) {
	267	res = -ERESTARTSYS;
	268	break;
	269	}
d48eb233	270	spin_unlock_irq(&ctx->wqh.lock);
e1ad7468	271	schedule();
d48eb233	272	spin_lock_irq(&ctx->wqh.lock);
e1ad7468 DL	273	}
	274	__remove_wait_queue(&ctx->wqh, &wait);
	275	__set_current_state(TASK_RUNNING);
	276	}
bcd0b235	277	if (likely(res > 0)) {
e1ad7468 DL	278	ctx->count += ucnt;
e1ad7468 DL	279	if (waitqueue_active(&ctx->wqh))
39510888	280	wake_up_locked_poll(&ctx->wqh, POLLIN);
e1ad7468	281	}
d48eb233	282	spin_unlock_irq(&ctx->wqh.lock);
e1ad7468 DL	283
	284	return res;
	285	}
	286
	287	static const struct file_operations eventfd_fops = {
	288	.release = eventfd_release,
	289	.poll = eventfd_poll,
	290	.read = eventfd_read,
	291	.write = eventfd_write,
6038f373	292	.llseek = noop_llseek,
e1ad7468 DL	293	};
e1ad7468 DL	294
13389010 DL	295	/**
	296	* eventfd_fget - Acquire a reference of an eventfd file descriptor.
	297	* @fd: [in] Eventfd file descriptor.
	298	*
	299	* Returns a pointer to the eventfd file structure in case of success, or the
	300	* following error pointer:
	301	*
	302	* -EBADF : Invalid @fd file descriptor.
	303	* -EINVAL : The @fd file descriptor is not an eventfd file.
	304	*/
e1ad7468 DL	305	struct file *eventfd_fget(int fd)
	306	{
	307	struct file *file;
	308
	309	file = fget(fd);
	310	if (!file)
	311	return ERR_PTR(-EBADF);
	312	if (file->f_op != &eventfd_fops) {
	313	fput(file);
	314	return ERR_PTR(-EINVAL);
	315	}
	316
	317	return file;
	318	}
5718607b	319	EXPORT_SYMBOL_GPL(eventfd_fget);
e1ad7468	320
13389010 DL	321	/**
	322	* eventfd_ctx_fdget - Acquires a reference to the internal eventfd context.
	323	* @fd: [in] Eventfd file descriptor.
	324	*
	325	* Returns a pointer to the internal eventfd context, otherwise the error
	326	* pointers returned by the following functions:
	327	*
	328	* eventfd_fget
	329	*/
	330	struct eventfd_ctx *eventfd_ctx_fdget(int fd)
	331	{
	332	struct file *file;
	333	struct eventfd_ctx *ctx;
	334
	335	file = eventfd_fget(fd);
	336	if (IS_ERR(file))
	337	return (struct eventfd_ctx *) file;
	338	ctx = eventfd_ctx_get(file->private_data);
	339	fput(file);
	340
	341	return ctx;
	342	}
	343	EXPORT_SYMBOL_GPL(eventfd_ctx_fdget);
	344
	345	/**
	346	* eventfd_ctx_fileget - Acquires a reference to the internal eventfd context.
	347	* @file: [in] Eventfd file pointer.
	348	*
	349	* Returns a pointer to the internal eventfd context, otherwise the error
	350	* pointer:
	351	*
	352	* -EINVAL : The @fd file descriptor is not an eventfd file.
	353	*/
	354	struct eventfd_ctx eventfd_ctx_fileget(struct file file)
	355	{
	356	if (file->f_op != &eventfd_fops)
	357	return ERR_PTR(-EINVAL);
	358
	359	return eventfd_ctx_get(file->private_data);
	360	}
	361	EXPORT_SYMBOL_GPL(eventfd_ctx_fileget);
	362
562787a5 DL	363	/**
	364	* eventfd_file_create - Creates an eventfd file pointer.
	365	* @count: Initial eventfd counter value.
	366	* @flags: Flags for the eventfd file.
	367	*
	368	* This function creates an eventfd file pointer, w/out installing it into
	369	* the fd table. This is useful when the eventfd file is used during the
	370	* initialization of data structures that require extra setup after the eventfd
	371	* creation. So the eventfd creation is split into the file pointer creation
	372	* phase, and the file descriptor installation phase.
	373	* In this way races with userspace closing the newly installed file descriptor
	374	* can be avoided.
	375	* Returns an eventfd file pointer, or a proper error pointer.
	376	*/
	377	struct file *eventfd_file_create(unsigned int count, int flags)
e1ad7468	378	{
562787a5	379	struct file *file;
e1ad7468	380	struct eventfd_ctx *ctx;
e1ad7468	381
e38b36f3 UD	382	/* Check the EFD_* constants for consistency. */
	383	BUILD_BUG_ON(EFD_CLOEXEC != O_CLOEXEC);
	384	BUILD_BUG_ON(EFD_NONBLOCK != O_NONBLOCK);
	385
bcd0b235	386	if (flags & ~EFD_FLAGS_SET)
562787a5	387	return ERR_PTR(-EINVAL);
b087498e	388
e1ad7468 DL	389	ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
e1ad7468 DL	390	if (!ctx)
562787a5	391	return ERR_PTR(-ENOMEM);
e1ad7468	392
13389010	393	kref_init(&ctx->kref);
e1ad7468	394	init_waitqueue_head(&ctx->wqh);
e1ad7468	395	ctx->count = count;
bcd0b235	396	ctx->flags = flags;
e1ad7468	397
562787a5	398	file = anon_inode_getfile("[eventfd]", &eventfd_fops, ctx,
628ff7c1	399	O_RDWR \| (flags & EFD_SHARED_FCNTL_FLAGS));
562787a5 DL	400	if (IS_ERR(file))
	401	eventfd_free_ctx(ctx);
	402
	403	return file;
	404	}
	405
	406	SYSCALL_DEFINE2(eventfd2, unsigned int, count, int, flags)
	407	{
	408	int fd, error;
	409	struct file *file;
	410
	411	error = get_unused_fd_flags(flags & EFD_SHARED_FCNTL_FLAGS);
	412	if (error < 0)
	413	return error;
	414	fd = error;
	415
	416	file = eventfd_file_create(count, flags);
	417	if (IS_ERR(file)) {
	418	error = PTR_ERR(file);
	419	goto err_put_unused_fd;
	420	}
	421	fd_install(fd, file);
	422
2030a42c	423	return fd;
562787a5 DL	424
	425	err_put_unused_fd:
	426	put_unused_fd(fd);
	427
	428	return error;
e1ad7468 DL	429	}
e1ad7468 DL	430
d4e82042	431	SYSCALL_DEFINE1(eventfd, unsigned int, count)
b087498e UD	432	{
	433	return sys_eventfd2(count, 0);
	434	}
bcd0b235	435