[deliverable/linux.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/export.h>
#include <linux/kref.h>
#include <linux/eventfd.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.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;
}

#ifdef CONFIG_PROC_FS
static int eventfd_show_fdinfo(struct seq_file *m, struct file *f)
{
	struct eventfd_ctx *ctx = f->private_data;
	int ret;

	spin_lock_irq(&ctx->wqh.lock);
	ret = seq_printf(m, "eventfd-count: %16llx\n",
			 (unsigned long long)ctx->count);
	spin_unlock_irq(&ctx->wqh.lock);

	return ret;
}
#endif

static const struct file_operations eventfd_fops = {
#ifdef CONFIG_PROC_FS
	.show_fdinfo	= eventfd_show_fdinfo,
#endif
	.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 <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>
630d9c47	19	#include <linux/export.h>
13389010 DL	20	#include <linux/kref.h>
13389010 DL	21	#include <linux/eventfd.h>
cbac5542 CG	22	#include <linux/proc_fs.h>
cbac5542 CG	23	#include <linux/seq_file.h>
e1ad7468 DL	24
e1ad7468 DL	25	struct eventfd_ctx {
13389010	26	struct kref kref;
e1ad7468 DL	27	wait_queue_head_t wqh;
	28	/*
	29	* Every time that a write(2) is performed on an eventfd, the
	30	* value of the __u64 being written is added to "count" and a
	31	* wakeup is performed on "wqh". A read(2) will return the "count"
	32	* value to userspace, and will reset "count" to zero. The kernel
13389010	33	* side eventfd_signal() also, adds to the "count" counter and
e1ad7468 DL	34	* issue a wakeup.
	35	*/
	36	__u64 count;
bcd0b235	37	unsigned int flags;
e1ad7468 DL	38	};
e1ad7468 DL	39
13389010 DL	40	/**
	41	* eventfd_signal - Adds @n to the eventfd counter.
	42	* @ctx: [in] Pointer to the eventfd context.
	43	* @n: [in] Value of the counter to be added to the eventfd internal counter.
	44	* The value cannot be negative.
	45	*
	46	* This function is supposed to be called by the kernel in paths that do not
	47	* allow sleeping. In this function we allow the counter to reach the ULLONG_MAX
	48	* value, and we signal this as overflow condition by returining a POLLERR
	49	* to poll(2).
	50	*
ee62c6b2 SZ	51	* Returns the amount by which the counter was incrememnted. This will be less
ee62c6b2 SZ	52	* than @n if the counter has overflowed.
e1ad7468	53	*/
ee62c6b2	54	__u64 eventfd_signal(struct eventfd_ctx *ctx, __u64 n)
e1ad7468	55	{
e1ad7468 DL	56	unsigned long flags;
e1ad7468 DL	57
d48eb233	58	spin_lock_irqsave(&ctx->wqh.lock, flags);
e1ad7468	59	if (ULLONG_MAX - ctx->count < n)
ee62c6b2	60	n = ULLONG_MAX - ctx->count;
e1ad7468 DL	61	ctx->count += n;
e1ad7468 DL	62	if (waitqueue_active(&ctx->wqh))
39510888	63	wake_up_locked_poll(&ctx->wqh, POLLIN);
d48eb233	64	spin_unlock_irqrestore(&ctx->wqh.lock, flags);
e1ad7468 DL	65
	66	return n;
	67	}
5718607b	68	EXPORT_SYMBOL_GPL(eventfd_signal);
e1ad7468	69
562787a5 DL	70	static void eventfd_free_ctx(struct eventfd_ctx *ctx)
	71	{
	72	kfree(ctx);
	73	}
	74
13389010 DL	75	static void eventfd_free(struct kref *kref)
	76	{
	77	struct eventfd_ctx *ctx = container_of(kref, struct eventfd_ctx, kref);
	78
562787a5	79	eventfd_free_ctx(ctx);
13389010 DL	80	}
	81
	82	/**
	83	* eventfd_ctx_get - Acquires a reference to the internal eventfd context.
	84	* @ctx: [in] Pointer to the eventfd context.
	85	*
	86	* Returns: In case of success, returns a pointer to the eventfd context.
	87	*/
	88	struct eventfd_ctx eventfd_ctx_get(struct eventfd_ctx ctx)
	89	{
	90	kref_get(&ctx->kref);
	91	return ctx;
	92	}
	93	EXPORT_SYMBOL_GPL(eventfd_ctx_get);
	94
	95	/**
	96	* eventfd_ctx_put - Releases a reference to the internal eventfd context.
	97	* @ctx: [in] Pointer to eventfd context.
	98	*
	99	* The eventfd context reference must have been previously acquired either
36182185	100	* with eventfd_ctx_get() or eventfd_ctx_fdget().
13389010 DL	101	*/
	102	void eventfd_ctx_put(struct eventfd_ctx *ctx)
	103	{
	104	kref_put(&ctx->kref, eventfd_free);
	105	}
	106	EXPORT_SYMBOL_GPL(eventfd_ctx_put);
	107
e1ad7468 DL	108	static int eventfd_release(struct inode inode, struct file file)
e1ad7468 DL	109	{
13389010 DL	110	struct eventfd_ctx *ctx = file->private_data;
	111
	112	wake_up_poll(&ctx->wqh, POLLHUP);
	113	eventfd_ctx_put(ctx);
e1ad7468 DL	114	return 0;
	115	}
	116
	117	static unsigned int eventfd_poll(struct file file, poll_table wait)
	118	{
	119	struct eventfd_ctx *ctx = file->private_data;
	120	unsigned int events = 0;
	121	unsigned long flags;
	122
	123	poll_wait(file, &ctx->wqh, wait);
	124
d48eb233	125	spin_lock_irqsave(&ctx->wqh.lock, flags);
e1ad7468 DL	126	if (ctx->count > 0)
	127	events \|= POLLIN;
	128	if (ctx->count == ULLONG_MAX)
	129	events \|= POLLERR;
	130	if (ULLONG_MAX - 1 > ctx->count)
	131	events \|= POLLOUT;
d48eb233	132	spin_unlock_irqrestore(&ctx->wqh.lock, flags);
e1ad7468 DL	133
	134	return events;
	135	}
	136
cb289d62 DL	137	static void eventfd_ctx_do_read(struct eventfd_ctx ctx, __u64 cnt)
	138	{
	139	*cnt = (ctx->flags & EFD_SEMAPHORE) ? 1 : ctx->count;
	140	ctx->count -= *cnt;
	141	}
	142
	143	/**
	144	* eventfd_ctx_remove_wait_queue - Read the current counter and removes wait queue.
	145	* @ctx: [in] Pointer to eventfd context.
	146	* @wait: [in] Wait queue to be removed.
36182185	147	* @cnt: [out] Pointer to the 64-bit counter value.
cb289d62	148	*
36182185	149	* Returns %0 if successful, or the following error codes:
cb289d62 DL	150	*
	151	* -EAGAIN : The operation would have blocked.
	152	*
	153	* This is used to atomically remove a wait queue entry from the eventfd wait
	154	* queue head, and read/reset the counter value.
	155	*/
	156	int eventfd_ctx_remove_wait_queue(struct eventfd_ctx ctx, wait_queue_t wait,
	157	__u64 *cnt)
	158	{
	159	unsigned long flags;
	160
	161	spin_lock_irqsave(&ctx->wqh.lock, flags);
	162	eventfd_ctx_do_read(ctx, cnt);
	163	__remove_wait_queue(&ctx->wqh, wait);
	164	if (*cnt != 0 && waitqueue_active(&ctx->wqh))
	165	wake_up_locked_poll(&ctx->wqh, POLLOUT);
	166	spin_unlock_irqrestore(&ctx->wqh.lock, flags);
	167
	168	return *cnt != 0 ? 0 : -EAGAIN;
	169	}
	170	EXPORT_SYMBOL_GPL(eventfd_ctx_remove_wait_queue);
	171
	172	/**
	173	* eventfd_ctx_read - Reads the eventfd counter or wait if it is zero.
	174	* @ctx: [in] Pointer to eventfd context.
	175	* @no_wait: [in] Different from zero if the operation should not block.
36182185	176	* @cnt: [out] Pointer to the 64-bit counter value.
cb289d62	177	*
36182185	178	* Returns %0 if successful, or the following error codes:
cb289d62	179	*
36182185	180	* -EAGAIN : The operation would have blocked but @no_wait was non-zero.
cb289d62 DL	181	* -ERESTARTSYS : A signal interrupted the wait operation.
	182	*
	183	* If @no_wait is zero, the function might sleep until the eventfd internal
	184	* counter becomes greater than zero.
	185	*/
	186	ssize_t eventfd_ctx_read(struct eventfd_ctx ctx, int no_wait, __u64 cnt)
e1ad7468	187	{
e1ad7468	188	ssize_t res;
e1ad7468 DL	189	DECLARE_WAITQUEUE(wait, current);
e1ad7468 DL	190
d48eb233	191	spin_lock_irq(&ctx->wqh.lock);
cb289d62	192	*cnt = 0;
e1ad7468	193	res = -EAGAIN;
bcd0b235	194	if (ctx->count > 0)
cb289d62 DL	195	res = 0;
cb289d62 DL	196	else if (!no_wait) {
e1ad7468	197	__add_wait_queue(&ctx->wqh, &wait);
cb289d62	198	for (;;) {
e1ad7468 DL	199	set_current_state(TASK_INTERRUPTIBLE);
e1ad7468 DL	200	if (ctx->count > 0) {
cb289d62	201	res = 0;
e1ad7468 DL	202	break;
	203	}
	204	if (signal_pending(current)) {
	205	res = -ERESTARTSYS;
	206	break;
	207	}
d48eb233	208	spin_unlock_irq(&ctx->wqh.lock);
e1ad7468	209	schedule();
d48eb233	210	spin_lock_irq(&ctx->wqh.lock);
e1ad7468 DL	211	}
	212	__remove_wait_queue(&ctx->wqh, &wait);
	213	__set_current_state(TASK_RUNNING);
	214	}
cb289d62 DL	215	if (likely(res == 0)) {
cb289d62 DL	216	eventfd_ctx_do_read(ctx, cnt);
e1ad7468	217	if (waitqueue_active(&ctx->wqh))
39510888	218	wake_up_locked_poll(&ctx->wqh, POLLOUT);
e1ad7468	219	}
d48eb233	220	spin_unlock_irq(&ctx->wqh.lock);
e1ad7468 DL	221
	222	return res;
	223	}
cb289d62 DL	224	EXPORT_SYMBOL_GPL(eventfd_ctx_read);
	225
	226	static ssize_t eventfd_read(struct file file, char __user buf, size_t count,
	227	loff_t *ppos)
	228	{
	229	struct eventfd_ctx *ctx = file->private_data;
	230	ssize_t res;
	231	__u64 cnt;
	232
	233	if (count < sizeof(cnt))
	234	return -EINVAL;
	235	res = eventfd_ctx_read(ctx, file->f_flags & O_NONBLOCK, &cnt);
	236	if (res < 0)
	237	return res;
	238
	239	return put_user(cnt, (__u64 __user *) buf) ? -EFAULT : sizeof(cnt);
	240	}
e1ad7468 DL	241
	242	static ssize_t eventfd_write(struct file file, const char __user buf, size_t count,
	243	loff_t *ppos)
	244	{
	245	struct eventfd_ctx *ctx = file->private_data;
	246	ssize_t res;
	247	__u64 ucnt;
	248	DECLARE_WAITQUEUE(wait, current);
	249
	250	if (count < sizeof(ucnt))
	251	return -EINVAL;
	252	if (copy_from_user(&ucnt, buf, sizeof(ucnt)))
	253	return -EFAULT;
	254	if (ucnt == ULLONG_MAX)
	255	return -EINVAL;
d48eb233	256	spin_lock_irq(&ctx->wqh.lock);
e1ad7468 DL	257	res = -EAGAIN;
	258	if (ULLONG_MAX - ctx->count > ucnt)
	259	res = sizeof(ucnt);
	260	else if (!(file->f_flags & O_NONBLOCK)) {
	261	__add_wait_queue(&ctx->wqh, &wait);
	262	for (res = 0;;) {
	263	set_current_state(TASK_INTERRUPTIBLE);
	264	if (ULLONG_MAX - ctx->count > ucnt) {
	265	res = sizeof(ucnt);
	266	break;
	267	}
	268	if (signal_pending(current)) {
	269	res = -ERESTARTSYS;
	270	break;
	271	}
d48eb233	272	spin_unlock_irq(&ctx->wqh.lock);
e1ad7468	273	schedule();
d48eb233	274	spin_lock_irq(&ctx->wqh.lock);
e1ad7468 DL	275	}
	276	__remove_wait_queue(&ctx->wqh, &wait);
	277	__set_current_state(TASK_RUNNING);
	278	}
bcd0b235	279	if (likely(res > 0)) {
e1ad7468 DL	280	ctx->count += ucnt;
e1ad7468 DL	281	if (waitqueue_active(&ctx->wqh))
39510888	282	wake_up_locked_poll(&ctx->wqh, POLLIN);
e1ad7468	283	}
d48eb233	284	spin_unlock_irq(&ctx->wqh.lock);
e1ad7468 DL	285
	286	return res;
	287	}
	288
cbac5542 CG	289	#ifdef CONFIG_PROC_FS
	290	static int eventfd_show_fdinfo(struct seq_file m, struct file f)
	291	{
	292	struct eventfd_ctx *ctx = f->private_data;
	293	int ret;
	294
	295	spin_lock_irq(&ctx->wqh.lock);
	296	ret = seq_printf(m, "eventfd-count: %16llx\n",
	297	(unsigned long long)ctx->count);
	298	spin_unlock_irq(&ctx->wqh.lock);
	299
	300	return ret;
	301	}
	302	#endif
	303
e1ad7468	304	static const struct file_operations eventfd_fops = {
cbac5542 CG	305	#ifdef CONFIG_PROC_FS
	306	.show_fdinfo = eventfd_show_fdinfo,
	307	#endif
e1ad7468 DL	308	.release = eventfd_release,
	309	.poll = eventfd_poll,
	310	.read = eventfd_read,
	311	.write = eventfd_write,
6038f373	312	.llseek = noop_llseek,
e1ad7468 DL	313	};
e1ad7468 DL	314
13389010 DL	315	/**
	316	* eventfd_fget - Acquire a reference of an eventfd file descriptor.
	317	* @fd: [in] Eventfd file descriptor.
	318	*
	319	* Returns a pointer to the eventfd file structure in case of success, or the
	320	* following error pointer:
	321	*
	322	* -EBADF : Invalid @fd file descriptor.
	323	* -EINVAL : The @fd file descriptor is not an eventfd file.
	324	*/
e1ad7468 DL	325	struct file *eventfd_fget(int fd)
	326	{
	327	struct file *file;
	328
	329	file = fget(fd);
	330	if (!file)
	331	return ERR_PTR(-EBADF);
	332	if (file->f_op != &eventfd_fops) {
	333	fput(file);
	334	return ERR_PTR(-EINVAL);
	335	}
	336
	337	return file;
	338	}
5718607b	339	EXPORT_SYMBOL_GPL(eventfd_fget);
e1ad7468	340
13389010 DL	341	/**
	342	* eventfd_ctx_fdget - Acquires a reference to the internal eventfd context.
	343	* @fd: [in] Eventfd file descriptor.
	344	*
	345	* Returns a pointer to the internal eventfd context, otherwise the error
	346	* pointers returned by the following functions:
	347	*
	348	* eventfd_fget
	349	*/
	350	struct eventfd_ctx *eventfd_ctx_fdget(int fd)
	351	{
	352	struct file *file;
	353	struct eventfd_ctx *ctx;
	354
	355	file = eventfd_fget(fd);
	356	if (IS_ERR(file))
	357	return (struct eventfd_ctx *) file;
	358	ctx = eventfd_ctx_get(file->private_data);
	359	fput(file);
	360
	361	return ctx;
	362	}
	363	EXPORT_SYMBOL_GPL(eventfd_ctx_fdget);
	364
	365	/**
	366	* eventfd_ctx_fileget - Acquires a reference to the internal eventfd context.
	367	* @file: [in] Eventfd file pointer.
	368	*
	369	* Returns a pointer to the internal eventfd context, otherwise the error
	370	* pointer:
	371	*
	372	* -EINVAL : The @fd file descriptor is not an eventfd file.
	373	*/
	374	struct eventfd_ctx eventfd_ctx_fileget(struct file file)
	375	{
	376	if (file->f_op != &eventfd_fops)
	377	return ERR_PTR(-EINVAL);
	378
	379	return eventfd_ctx_get(file->private_data);
	380	}
	381	EXPORT_SYMBOL_GPL(eventfd_ctx_fileget);
	382
562787a5 DL	383	/**
	384	* eventfd_file_create - Creates an eventfd file pointer.
	385	* @count: Initial eventfd counter value.
	386	* @flags: Flags for the eventfd file.
	387	*
	388	* This function creates an eventfd file pointer, w/out installing it into
	389	* the fd table. This is useful when the eventfd file is used during the
	390	* initialization of data structures that require extra setup after the eventfd
	391	* creation. So the eventfd creation is split into the file pointer creation
	392	* phase, and the file descriptor installation phase.
	393	* In this way races with userspace closing the newly installed file descriptor
	394	* can be avoided.
	395	* Returns an eventfd file pointer, or a proper error pointer.
	396	*/
	397	struct file *eventfd_file_create(unsigned int count, int flags)
e1ad7468	398	{
562787a5	399	struct file *file;
e1ad7468	400	struct eventfd_ctx *ctx;
e1ad7468	401
e38b36f3 UD	402	/* Check the EFD_* constants for consistency. */
	403	BUILD_BUG_ON(EFD_CLOEXEC != O_CLOEXEC);
	404	BUILD_BUG_ON(EFD_NONBLOCK != O_NONBLOCK);
	405
bcd0b235	406	if (flags & ~EFD_FLAGS_SET)
562787a5	407	return ERR_PTR(-EINVAL);
b087498e	408
e1ad7468 DL	409	ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
e1ad7468 DL	410	if (!ctx)
562787a5	411	return ERR_PTR(-ENOMEM);
e1ad7468	412
13389010	413	kref_init(&ctx->kref);
e1ad7468	414	init_waitqueue_head(&ctx->wqh);
e1ad7468	415	ctx->count = count;
bcd0b235	416	ctx->flags = flags;
e1ad7468	417
562787a5	418	file = anon_inode_getfile("[eventfd]", &eventfd_fops, ctx,
628ff7c1	419	O_RDWR \| (flags & EFD_SHARED_FCNTL_FLAGS));
562787a5 DL	420	if (IS_ERR(file))
	421	eventfd_free_ctx(ctx);
	422
	423	return file;
	424	}
	425
	426	SYSCALL_DEFINE2(eventfd2, unsigned int, count, int, flags)
	427	{
	428	int fd, error;
	429	struct file *file;
	430
	431	error = get_unused_fd_flags(flags & EFD_SHARED_FCNTL_FLAGS);
	432	if (error < 0)
	433	return error;
	434	fd = error;
	435
	436	file = eventfd_file_create(count, flags);
	437	if (IS_ERR(file)) {
	438	error = PTR_ERR(file);
	439	goto err_put_unused_fd;
	440	}
	441	fd_install(fd, file);
	442
2030a42c	443	return fd;
562787a5 DL	444
	445	err_put_unused_fd:
	446	put_unused_fd(fd);
	447
	448	return error;
e1ad7468 DL	449	}
e1ad7468 DL	450
d4e82042	451	SYSCALL_DEFINE1(eventfd, unsigned int, count)
b087498e UD	452	{
	453	return sys_eventfd2(count, 0);
	454	}
bcd0b235	455