[thirdparty/linux.git] / fs / eventfd.c

// SPDX-License-Identifier: GPL-2.0-only
/*
 *  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/signal.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>
#include <linux/idr.h>
#include <linux/uio.h>

static DEFINE_IDA(eventfd_ida);

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". If EFD_SEMAPHORE flag was not
	 * specified, 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;
	int id;
};

/**
 * eventfd_signal_mask - Increment the event counter
 * @ctx: [in] Pointer to the eventfd context.
 * @mask: [in] poll mask
 *
 * 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 returning a EPOLLERR
 * to poll(2).
 */
void eventfd_signal_mask(struct eventfd_ctx *ctx, __poll_t mask)
{
	unsigned long flags;

	/*
	 * Deadlock or stack overflow issues can happen if we recurse here
	 * through waitqueue wakeup handlers. If the caller users potentially
	 * nested waitqueues with custom wakeup handlers, then it should
	 * check eventfd_signal_allowed() before calling this function. If
	 * it returns false, the eventfd_signal() call should be deferred to a
	 * safe context.
	 */
	if (WARN_ON_ONCE(current->in_eventfd))
		return;

	spin_lock_irqsave(&ctx->wqh.lock, flags);
	current->in_eventfd = 1;
	if (ctx->count < ULLONG_MAX)
		ctx->count++;
	if (waitqueue_active(&ctx->wqh))
		wake_up_locked_poll(&ctx->wqh, EPOLLIN | mask);
	current->in_eventfd = 0;
	spin_unlock_irqrestore(&ctx->wqh.lock, flags);
}
EXPORT_SYMBOL_GPL(eventfd_signal_mask);

static void eventfd_free_ctx(struct eventfd_ctx *ctx)
{
	if (ctx->id >= 0)
		ida_free(&eventfd_ida, ctx->id);
	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_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_fdget() or eventfd_ctx_fileget().
 */
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, EPOLLHUP);
	eventfd_ctx_put(ctx);
	return 0;
}

static __poll_t eventfd_poll(struct file *file, poll_table *wait)
{
	struct eventfd_ctx *ctx = file->private_data;
	__poll_t events = 0;
	u64 count;

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

	/*
	 * All writes to ctx->count occur within ctx->wqh.lock.  This read
	 * can be done outside ctx->wqh.lock because we know that poll_wait
	 * takes that lock (through add_wait_queue) if our caller will sleep.
	 *
	 * The read _can_ therefore seep into add_wait_queue's critical
	 * section, but cannot move above it!  add_wait_queue's spin_lock acts
	 * as an acquire barrier and ensures that the read be ordered properly
	 * against the writes.  The following CAN happen and is safe:
	 *
	 *     poll                               write
	 *     -----------------                  ------------
	 *     lock ctx->wqh.lock (in poll_wait)
	 *     count = ctx->count
	 *     __add_wait_queue
	 *     unlock ctx->wqh.lock
	 *                                        lock ctx->qwh.lock
	 *                                        ctx->count += n
	 *                                        if (waitqueue_active)
	 *                                          wake_up_locked_poll
	 *                                        unlock ctx->qwh.lock
	 *     eventfd_poll returns 0
	 *
	 * but the following, which would miss a wakeup, cannot happen:
	 *
	 *     poll                               write
	 *     -----------------                  ------------
	 *     count = ctx->count (INVALID!)
	 *                                        lock ctx->qwh.lock
	 *                                        ctx->count += n
	 *                                        **waitqueue_active is false**
	 *                                        **no wake_up_locked_poll!**
	 *                                        unlock ctx->qwh.lock
	 *     lock ctx->wqh.lock (in poll_wait)
	 *     __add_wait_queue
	 *     unlock ctx->wqh.lock
	 *     eventfd_poll returns 0
	 */
	count = READ_ONCE(ctx->count);

	if (count > 0)
		events |= EPOLLIN;
	if (count == ULLONG_MAX)
		events |= EPOLLERR;
	if (ULLONG_MAX - 1 > count)
		events |= EPOLLOUT;

	return events;
}

void eventfd_ctx_do_read(struct eventfd_ctx *ctx, __u64 *cnt)
{
	lockdep_assert_held(&ctx->wqh.lock);

	*cnt = ((ctx->flags & EFD_SEMAPHORE) && ctx->count) ? 1 : ctx->count;
	ctx->count -= *cnt;
}
EXPORT_SYMBOL_GPL(eventfd_ctx_do_read);

/**
 * 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_entry_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, EPOLLOUT);
	spin_unlock_irqrestore(&ctx->wqh.lock, flags);

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

static ssize_t eventfd_read(struct kiocb *iocb, struct iov_iter *to)
{
	struct file *file = iocb->ki_filp;
	struct eventfd_ctx *ctx = file->private_data;
	__u64 ucnt = 0;

	if (iov_iter_count(to) < sizeof(ucnt))
		return -EINVAL;
	spin_lock_irq(&ctx->wqh.lock);
	if (!ctx->count) {
		if ((file->f_flags & O_NONBLOCK) ||
		    (iocb->ki_flags & IOCB_NOWAIT)) {
			spin_unlock_irq(&ctx->wqh.lock);
			return -EAGAIN;
		}

		if (wait_event_interruptible_locked_irq(ctx->wqh, ctx->count)) {
			spin_unlock_irq(&ctx->wqh.lock);
			return -ERESTARTSYS;
		}
	}
	eventfd_ctx_do_read(ctx, &ucnt);
	current->in_eventfd = 1;
	if (waitqueue_active(&ctx->wqh))
		wake_up_locked_poll(&ctx->wqh, EPOLLOUT);
	current->in_eventfd = 0;
	spin_unlock_irq(&ctx->wqh.lock);
	if (unlikely(copy_to_iter(&ucnt, sizeof(ucnt), to) != sizeof(ucnt)))
		return -EFAULT;

	return sizeof(ucnt);
}

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;

	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)) {
		res = wait_event_interruptible_locked_irq(ctx->wqh,
				ULLONG_MAX - ctx->count > ucnt);
		if (!res)
			res = sizeof(ucnt);
	}
	if (likely(res > 0)) {
		ctx->count += ucnt;
		current->in_eventfd = 1;
		if (waitqueue_active(&ctx->wqh))
			wake_up_locked_poll(&ctx->wqh, EPOLLIN);
		current->in_eventfd = 0;
	}
	spin_unlock_irq(&ctx->wqh.lock);

	return res;
}

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

	spin_lock_irq(&ctx->wqh.lock);
	seq_printf(m, "eventfd-count: %16llx\n",
		   (unsigned long long)ctx->count);
	spin_unlock_irq(&ctx->wqh.lock);
	seq_printf(m, "eventfd-id: %d\n", ctx->id);
	seq_printf(m, "eventfd-semaphore: %d\n",
		   !!(ctx->flags & EFD_SEMAPHORE));
}
#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_iter	= 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 eventfd_ctx *ctx;
	struct fd f = fdget(fd);
	if (!f.file)
		return ERR_PTR(-EBADF);
	ctx = eventfd_ctx_fileget(f.file);
	fdput(f);
	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)
{
	struct eventfd_ctx *ctx;

	if (file->f_op != &eventfd_fops)
		return ERR_PTR(-EINVAL);

	ctx = file->private_data;
	kref_get(&ctx->kref);
	return ctx;
}
EXPORT_SYMBOL_GPL(eventfd_ctx_fileget);

static int do_eventfd(unsigned int count, int flags)
{
	struct eventfd_ctx *ctx;
	struct file *file;
	int fd;

	/* 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 -EINVAL;

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

	kref_init(&ctx->kref);
	init_waitqueue_head(&ctx->wqh);
	ctx->count = count;
	ctx->flags = flags;
	ctx->id = ida_alloc(&eventfd_ida, GFP_KERNEL);

	flags &= EFD_SHARED_FCNTL_FLAGS;
	flags |= O_RDWR;
	fd = get_unused_fd_flags(flags);
	if (fd < 0)
		goto err;

	file = anon_inode_getfile("[eventfd]", &eventfd_fops, ctx, flags);
	if (IS_ERR(file)) {
		put_unused_fd(fd);
		fd = PTR_ERR(file);
		goto err;
	}

	file->f_mode |= FMODE_NOWAIT;
	fd_install(fd, file);
	return fd;
err:
	eventfd_free_ctx(ctx);
	return fd;
}

SYSCALL_DEFINE2(eventfd2, unsigned int, count, int, flags)
{
	return do_eventfd(count, flags);
}

SYSCALL_DEFINE1(eventfd, unsigned int, count)
{
	return do_eventfd(count, 0);
}
Commit	Line	Data
457c8996	1	// SPDX-License-Identifier: GPL-2.0-only
e1ad7468 DL	2	/*
	3	* fs/eventfd.c
	4	*
	5	* Copyright (C) 2007 Davide Libenzi <davidel@xmailserver.org>
	6	*
	7	*/
	8
	9	#include <linux/file.h>
	10	#include <linux/poll.h>
	11	#include <linux/init.h>
	12	#include <linux/fs.h>
174cd4b1	13	#include <linux/sched/signal.h>
e1ad7468	14	#include <linux/kernel.h>
5a0e3ad6	15	#include <linux/slab.h>
e1ad7468 DL	16	#include <linux/list.h>
	17	#include <linux/spinlock.h>
	18	#include <linux/anon_inodes.h>
7747cdb2	19	#include <linux/syscalls.h>
630d9c47	20	#include <linux/export.h>
13389010 DL	21	#include <linux/kref.h>
13389010 DL	22	#include <linux/eventfd.h>
cbac5542 CG	23	#include <linux/proc_fs.h>
cbac5542 CG	24	#include <linux/seq_file.h>
b556db17	25	#include <linux/idr.h>
12aceb89	26	#include <linux/uio.h>
b556db17	27
ce528c4c	28	static DEFINE_IDA(eventfd_ida);
e1ad7468 DL	29
e1ad7468 DL	30	struct eventfd_ctx {
13389010	31	struct kref kref;
e1ad7468 DL	32	wait_queue_head_t wqh;
	33	/*
	34	* Every time that a write(2) is performed on an eventfd, the
	35	* value of the __u64 being written is added to "count" and a
33d8b5d7 WY	36	* wakeup is performed on "wqh". If EFD_SEMAPHORE flag was not
	37	* specified, a read(2) will return the "count" value to userspace,
	38	* and will reset "count" to zero. The kernel side eventfd_signal()
	39	* also, adds to the "count" counter and issue a wakeup.
e1ad7468 DL	40	*/
e1ad7468 DL	41	__u64 count;
bcd0b235	42	unsigned int flags;
b556db17	43	int id;
e1ad7468 DL	44	};
e1ad7468 DL	45
b7638ad0 CB	46	/**
	47	* eventfd_signal_mask - Increment the event counter
	48	* @ctx: [in] Pointer to the eventfd context.
	49	* @mask: [in] poll mask
	50	*
	51	* This function is supposed to be called by the kernel in paths that do not
	52	* allow sleeping. In this function we allow the counter to reach the ULLONG_MAX
	53	* value, and we signal this as overflow condition by returning a EPOLLERR
	54	* to poll(2).
	55	*/
	56	void eventfd_signal_mask(struct eventfd_ctx *ctx, __poll_t mask)
e1ad7468	57	{
e1ad7468 DL	58	unsigned long flags;
e1ad7468 DL	59
b5e683d5 JA	60	/*
	61	* Deadlock or stack overflow issues can happen if we recurse here
	62	* through waitqueue wakeup handlers. If the caller users potentially
	63	* nested waitqueues with custom wakeup handlers, then it should
b542e383 TG	64	* check eventfd_signal_allowed() before calling this function. If
b542e383 TG	65	* it returns false, the eventfd_signal() call should be deferred to a
b5e683d5 JA	66	* safe context.
b5e683d5 JA	67	*/
9f0deaa1	68	if (WARN_ON_ONCE(current->in_eventfd))
b7638ad0	69	return;
b5e683d5	70
d48eb233	71	spin_lock_irqsave(&ctx->wqh.lock, flags);
9f0deaa1	72	current->in_eventfd = 1;
b7638ad0 CB	73	if (ctx->count < ULLONG_MAX)
b7638ad0 CB	74	ctx->count++;
e1ad7468	75	if (waitqueue_active(&ctx->wqh))
03e02acd	76	wake_up_locked_poll(&ctx->wqh, EPOLLIN \| mask);
9f0deaa1	77	current->in_eventfd = 0;
d48eb233	78	spin_unlock_irqrestore(&ctx->wqh.lock, flags);
03e02acd	79	}
b7638ad0	80	EXPORT_SYMBOL_GPL(eventfd_signal_mask);
e1ad7468	81
562787a5 DL	82	static void eventfd_free_ctx(struct eventfd_ctx *ctx)
562787a5 DL	83	{
b556db17	84	if (ctx->id >= 0)
43422523	85	ida_free(&eventfd_ida, ctx->id);
562787a5 DL	86	kfree(ctx);
	87	}
	88
13389010 DL	89	static void eventfd_free(struct kref *kref)
	90	{
	91	struct eventfd_ctx *ctx = container_of(kref, struct eventfd_ctx, kref);
	92
562787a5	93	eventfd_free_ctx(ctx);
13389010 DL	94	}
13389010 DL	95
13389010 DL	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
105f2b70	101	* with eventfd_ctx_fdget() or eventfd_ctx_fileget().
13389010 DL	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;
13389010 DL	112
a9a08845	113	wake_up_poll(&ctx->wqh, EPOLLHUP);
13389010	114	eventfd_ctx_put(ctx);
e1ad7468 DL	115	return 0;
	116	}
	117
a11e1d43	118	static __poll_t eventfd_poll(struct file file, poll_table wait)
e1ad7468 DL	119	{
e1ad7468 DL	120	struct eventfd_ctx *ctx = file->private_data;
076ccb76	121	__poll_t events = 0;
e22553e2	122	u64 count;
e1ad7468	123
a11e1d43 LT	124	poll_wait(file, &ctx->wqh, wait);
a11e1d43 LT	125
a484c3dd PB	126	/*
	127	* All writes to ctx->count occur within ctx->wqh.lock. This read
	128	* can be done outside ctx->wqh.lock because we know that poll_wait
	129	* takes that lock (through add_wait_queue) if our caller will sleep.
	130	*
	131	* The read _can_ therefore seep into add_wait_queue's critical
	132	* section, but cannot move above it! add_wait_queue's spin_lock acts
	133	* as an acquire barrier and ensures that the read be ordered properly
	134	* against the writes. The following CAN happen and is safe:
	135	*
	136	* poll write
	137	* ----------------- ------------
	138	* lock ctx->wqh.lock (in poll_wait)
	139	* count = ctx->count
	140	* __add_wait_queue
	141	* unlock ctx->wqh.lock
	142	* lock ctx->qwh.lock
	143	* ctx->count += n
	144	* if (waitqueue_active)
	145	* wake_up_locked_poll
	146	* unlock ctx->qwh.lock
	147	* eventfd_poll returns 0
	148	*
	149	* but the following, which would miss a wakeup, cannot happen:
	150	*
	151	* poll write
	152	* ----------------- ------------
	153	* count = ctx->count (INVALID!)
	154	* lock ctx->qwh.lock
	155	* ctx->count += n
	156	* waitqueue_active is false
	157	* no wake_up_locked_poll!
	158	* unlock ctx->qwh.lock
	159	* lock ctx->wqh.lock (in poll_wait)
	160	* __add_wait_queue
	161	* unlock ctx->wqh.lock
	162	* eventfd_poll returns 0
	163	*/
	164	count = READ_ONCE(ctx->count);
e1ad7468	165
e22553e2	166	if (count > 0)
a11e1d43	167	events \|= EPOLLIN;
e22553e2	168	if (count == ULLONG_MAX)
a9a08845	169	events \|= EPOLLERR;
e22553e2	170	if (ULLONG_MAX - 1 > count)
a11e1d43	171	events \|= EPOLLOUT;
e1ad7468 DL	172
	173	return events;
	174	}
	175
28f13267	176	void eventfd_ctx_do_read(struct eventfd_ctx ctx, __u64 cnt)
cb289d62	177	{
28f13267 DW	178	lockdep_assert_held(&ctx->wqh.lock);
28f13267 DW	179
758b4920	180	*cnt = ((ctx->flags & EFD_SEMAPHORE) && ctx->count) ? 1 : ctx->count;
cb289d62 DL	181	ctx->count -= *cnt;
cb289d62 DL	182	}
28f13267	183	EXPORT_SYMBOL_GPL(eventfd_ctx_do_read);
cb289d62 DL	184
	185	/**
	186	* eventfd_ctx_remove_wait_queue - Read the current counter and removes wait queue.
	187	* @ctx: [in] Pointer to eventfd context.
	188	* @wait: [in] Wait queue to be removed.
36182185	189	* @cnt: [out] Pointer to the 64-bit counter value.
cb289d62	190	*
36182185	191	* Returns %0 if successful, or the following error codes:
cb289d62 DL	192	*
	193	* -EAGAIN : The operation would have blocked.
	194	*
	195	* This is used to atomically remove a wait queue entry from the eventfd wait
	196	* queue head, and read/reset the counter value.
	197	*/
ac6424b9	198	int eventfd_ctx_remove_wait_queue(struct eventfd_ctx ctx, wait_queue_entry_t wait,
cb289d62 DL	199	__u64 *cnt)
	200	{
	201	unsigned long flags;
	202
	203	spin_lock_irqsave(&ctx->wqh.lock, flags);
	204	eventfd_ctx_do_read(ctx, cnt);
	205	__remove_wait_queue(&ctx->wqh, wait);
	206	if (*cnt != 0 && waitqueue_active(&ctx->wqh))
a9a08845	207	wake_up_locked_poll(&ctx->wqh, EPOLLOUT);
cb289d62 DL	208	spin_unlock_irqrestore(&ctx->wqh.lock, flags);
	209
	210	return *cnt != 0 ? 0 : -EAGAIN;
	211	}
	212	EXPORT_SYMBOL_GPL(eventfd_ctx_remove_wait_queue);
	213
12aceb89	214	static ssize_t eventfd_read(struct kiocb iocb, struct iov_iter to)
e1ad7468	215	{
12aceb89	216	struct file *file = iocb->ki_filp;
b6364572	217	struct eventfd_ctx *ctx = file->private_data;
b6364572	218	__u64 ucnt = 0;
e1ad7468	219
12aceb89	220	if (iov_iter_count(to) < sizeof(ucnt))
b6364572	221	return -EINVAL;
d48eb233	222	spin_lock_irq(&ctx->wqh.lock);
12aceb89 JA	223	if (!ctx->count) {
	224	if ((file->f_flags & O_NONBLOCK) \|\|
	225	(iocb->ki_flags & IOCB_NOWAIT)) {
	226	spin_unlock_irq(&ctx->wqh.lock);
	227	return -EAGAIN;
	228	}
113348a4 WY	229
113348a4 WY	230	if (wait_event_interruptible_locked_irq(ctx->wqh, ctx->count)) {
d48eb233	231	spin_unlock_irq(&ctx->wqh.lock);
113348a4	232	return -ERESTARTSYS;
e1ad7468	233	}
e1ad7468	234	}
12aceb89	235	eventfd_ctx_do_read(ctx, &ucnt);
9f0deaa1	236	current->in_eventfd = 1;
12aceb89 JA	237	if (waitqueue_active(&ctx->wqh))
12aceb89 JA	238	wake_up_locked_poll(&ctx->wqh, EPOLLOUT);
9f0deaa1	239	current->in_eventfd = 0;
d48eb233	240	spin_unlock_irq(&ctx->wqh.lock);
12aceb89	241	if (unlikely(copy_to_iter(&ucnt, sizeof(ucnt), to) != sizeof(ucnt)))
b6364572	242	return -EFAULT;
cb289d62	243
12aceb89	244	return sizeof(ucnt);
cb289d62	245	}
e1ad7468 DL	246
	247	static ssize_t eventfd_write(struct file file, const char __user buf, size_t count,
	248	loff_t *ppos)
	249	{
	250	struct eventfd_ctx *ctx = file->private_data;
	251	ssize_t res;
	252	__u64 ucnt;
e1ad7468 DL	253
	254	if (count < sizeof(ucnt))
	255	return -EINVAL;
	256	if (copy_from_user(&ucnt, buf, sizeof(ucnt)))
	257	return -EFAULT;
	258	if (ucnt == ULLONG_MAX)
	259	return -EINVAL;
d48eb233	260	spin_lock_irq(&ctx->wqh.lock);
e1ad7468 DL	261	res = -EAGAIN;
	262	if (ULLONG_MAX - ctx->count > ucnt)
	263	res = sizeof(ucnt);
	264	else if (!(file->f_flags & O_NONBLOCK)) {
113348a4 WY	265	res = wait_event_interruptible_locked_irq(ctx->wqh,
	266	ULLONG_MAX - ctx->count > ucnt);
	267	if (!res)
	268	res = sizeof(ucnt);
e1ad7468	269	}
bcd0b235	270	if (likely(res > 0)) {
e1ad7468	271	ctx->count += ucnt;
9f0deaa1	272	current->in_eventfd = 1;
e1ad7468	273	if (waitqueue_active(&ctx->wqh))
a9a08845	274	wake_up_locked_poll(&ctx->wqh, EPOLLIN);
9f0deaa1	275	current->in_eventfd = 0;
e1ad7468	276	}
d48eb233	277	spin_unlock_irq(&ctx->wqh.lock);
e1ad7468 DL	278
	279	return res;
	280	}
	281
cbac5542	282	#ifdef CONFIG_PROC_FS
a3816ab0	283	static void eventfd_show_fdinfo(struct seq_file m, struct file f)
cbac5542 CG	284	{
cbac5542 CG	285	struct eventfd_ctx *ctx = f->private_data;
cbac5542 CG	286
cbac5542 CG	287	spin_lock_irq(&ctx->wqh.lock);
a3816ab0 JP	288	seq_printf(m, "eventfd-count: %16llx\n",
a3816ab0 JP	289	(unsigned long long)ctx->count);
cbac5542	290	spin_unlock_irq(&ctx->wqh.lock);
b556db17	291	seq_printf(m, "eventfd-id: %d\n", ctx->id);
33d8b5d7 WY	292	seq_printf(m, "eventfd-semaphore: %d\n",
33d8b5d7 WY	293	!!(ctx->flags & EFD_SEMAPHORE));
cbac5542 CG	294	}
	295	#endif
	296
e1ad7468	297	static const struct file_operations eventfd_fops = {
cbac5542 CG	298	#ifdef CONFIG_PROC_FS
	299	.show_fdinfo = eventfd_show_fdinfo,
	300	#endif
e1ad7468	301	.release = eventfd_release,
a11e1d43	302	.poll = eventfd_poll,
12aceb89	303	.read_iter = eventfd_read,
e1ad7468	304	.write = eventfd_write,
6038f373	305	.llseek = noop_llseek,
e1ad7468 DL	306	};
e1ad7468 DL	307
13389010 DL	308	/**
	309	* eventfd_fget - Acquire a reference of an eventfd file descriptor.
	310	* @fd: [in] Eventfd file descriptor.
	311	*
	312	* Returns a pointer to the eventfd file structure in case of success, or the
	313	* following error pointer:
	314	*
	315	* -EBADF : Invalid @fd file descriptor.
	316	* -EINVAL : The @fd file descriptor is not an eventfd file.
	317	*/
e1ad7468 DL	318	struct file *eventfd_fget(int fd)
	319	{
	320	struct file *file;
	321
	322	file = fget(fd);
	323	if (!file)
	324	return ERR_PTR(-EBADF);
	325	if (file->f_op != &eventfd_fops) {
	326	fput(file);
	327	return ERR_PTR(-EINVAL);
	328	}
	329
	330	return file;
	331	}
5718607b	332	EXPORT_SYMBOL_GPL(eventfd_fget);
e1ad7468	333
13389010 DL	334	/**
	335	* eventfd_ctx_fdget - Acquires a reference to the internal eventfd context.
	336	* @fd: [in] Eventfd file descriptor.
	337	*
	338	* Returns a pointer to the internal eventfd context, otherwise the error
	339	* pointers returned by the following functions:
	340	*
	341	* eventfd_fget
	342	*/
	343	struct eventfd_ctx *eventfd_ctx_fdget(int fd)
	344	{
13389010	345	struct eventfd_ctx *ctx;
36a74117 AV	346	struct fd f = fdget(fd);
	347	if (!f.file)
	348	return ERR_PTR(-EBADF);
	349	ctx = eventfd_ctx_fileget(f.file);
	350	fdput(f);
13389010 DL	351	return ctx;
	352	}
	353	EXPORT_SYMBOL_GPL(eventfd_ctx_fdget);
	354
	355	/**
	356	* eventfd_ctx_fileget - Acquires a reference to the internal eventfd context.
	357	* @file: [in] Eventfd file pointer.
	358	*
	359	* Returns a pointer to the internal eventfd context, otherwise the error
	360	* pointer:
	361	*
	362	* -EINVAL : The @fd file descriptor is not an eventfd file.
	363	*/
	364	struct eventfd_ctx eventfd_ctx_fileget(struct file file)
	365	{
105f2b70 EB	366	struct eventfd_ctx *ctx;
105f2b70 EB	367
13389010 DL	368	if (file->f_op != &eventfd_fops)
	369	return ERR_PTR(-EINVAL);
	370
105f2b70 EB	371	ctx = file->private_data;
	372	kref_get(&ctx->kref);
	373	return ctx;
13389010 DL	374	}
	375	EXPORT_SYMBOL_GPL(eventfd_ctx_fileget);
	376
2fc96f83	377	static int do_eventfd(unsigned int count, int flags)
e1ad7468	378	{
e1ad7468	379	struct eventfd_ctx *ctx;
12aceb89	380	struct file *file;
7d815165	381	int fd;
e1ad7468	382
e38b36f3 UD	383	/* Check the EFD_* constants for consistency. */
	384	BUILD_BUG_ON(EFD_CLOEXEC != O_CLOEXEC);
	385	BUILD_BUG_ON(EFD_NONBLOCK != O_NONBLOCK);
	386
bcd0b235	387	if (flags & ~EFD_FLAGS_SET)
7d815165	388	return -EINVAL;
b087498e	389
e1ad7468 DL	390	ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
e1ad7468 DL	391	if (!ctx)
7d815165	392	return -ENOMEM;
e1ad7468	393
13389010	394	kref_init(&ctx->kref);
e1ad7468	395	init_waitqueue_head(&ctx->wqh);
e1ad7468	396	ctx->count = count;
bcd0b235	397	ctx->flags = flags;
43422523	398	ctx->id = ida_alloc(&eventfd_ida, GFP_KERNEL);
e1ad7468	399
12aceb89 JA	400	flags &= EFD_SHARED_FCNTL_FLAGS;
	401	flags \|= O_RDWR;
	402	fd = get_unused_fd_flags(flags);
7d815165	403	if (fd < 0)
12aceb89 JA	404	goto err;
	405
	406	file = anon_inode_getfile("[eventfd]", &eventfd_fops, ctx, flags);
	407	if (IS_ERR(file)) {
	408	put_unused_fd(fd);
	409	fd = PTR_ERR(file);
	410	goto err;
	411	}
562787a5	412
12aceb89 JA	413	file->f_mode \|= FMODE_NOWAIT;
	414	fd_install(fd, file);
	415	return fd;
	416	err:
	417	eventfd_free_ctx(ctx);
2030a42c	418	return fd;
e1ad7468 DL	419	}
e1ad7468 DL	420
2fc96f83 DB	421	SYSCALL_DEFINE2(eventfd2, unsigned int, count, int, flags)
	422	{
	423	return do_eventfd(count, flags);
	424	}
	425
d4e82042	426	SYSCALL_DEFINE1(eventfd, unsigned int, count)
b087498e	427	{
2fc96f83	428	return do_eventfd(count, 0);
b087498e	429	}
bcd0b235	430