[linux-2.6-block.git] / fs / signalfd.c

// SPDX-License-Identifier: GPL-2.0
/*
 *  fs/signalfd.c
 *
 *  Copyright (C) 2003  Linus Torvalds
 *
 *  Mon Mar 5, 2007: Davide Libenzi <davidel@xmailserver.org>
 *      Changed ->read() to return a siginfo strcture instead of signal number.
 *      Fixed locking in ->poll().
 *      Added sighand-detach notification.
 *      Added fd re-use in sys_signalfd() syscall.
 *      Now using anonymous inode source.
 *      Thanks to Oleg Nesterov for useful code review and suggestions.
 *      More comments and suggestions from Arnd Bergmann.
 *  Sat May 19, 2007: Davi E. M. Arnaut <davi@haxent.com.br>
 *      Retrieve multiple signals with one read() call
 *  Sun Jul 15, 2007: Davide Libenzi <davidel@xmailserver.org>
 *      Attach to the sighand only during read() and poll().
 */

#include <linux/file.h>
#include <linux/poll.h>
#include <linux/init.h>
#include <linux/fs.h>
#include <linux/sched.h>
#include <linux/slab.h>
#include <linux/kernel.h>
#include <linux/signal.h>
#include <linux/list.h>
#include <linux/anon_inodes.h>
#include <linux/signalfd.h>
#include <linux/syscalls.h>
#include <linux/proc_fs.h>
#include <linux/compat.h>

void signalfd_cleanup(struct sighand_struct *sighand)
{
	wait_queue_head_t *wqh = &sighand->signalfd_wqh;
	/*
	 * The lockless check can race with remove_wait_queue() in progress,
	 * but in this case its caller should run under rcu_read_lock() and
	 * sighand_cachep is SLAB_TYPESAFE_BY_RCU, we can safely return.
	 */
	if (likely(!waitqueue_active(wqh)))
		return;

	/* wait_queue_entry_t->func(POLLFREE) should do remove_wait_queue() */
	wake_up_poll(wqh, EPOLLHUP | POLLFREE);
}

struct signalfd_ctx {
	sigset_t sigmask;
};

static int signalfd_release(struct inode *inode, struct file *file)
{
	kfree(file->private_data);
	return 0;
}

static __poll_t signalfd_poll(struct file *file, poll_table *wait)
{
	struct signalfd_ctx *ctx = file->private_data;
	__poll_t events = 0;

	poll_wait(file, &current->sighand->signalfd_wqh, wait);

	spin_lock_irq(&current->sighand->siglock);
	if (next_signal(&current->pending, &ctx->sigmask) ||
	    next_signal(&current->signal->shared_pending,
			&ctx->sigmask))
		events |= EPOLLIN;
	spin_unlock_irq(&current->sighand->siglock);

	return events;
}

/*
 * Copied from copy_siginfo_to_user() in kernel/signal.c
 */
static int signalfd_copyinfo(struct signalfd_siginfo __user *uinfo,
			     kernel_siginfo_t const *kinfo)
{
	struct signalfd_siginfo new;

	BUILD_BUG_ON(sizeof(struct signalfd_siginfo) != 128);

	/*
	 * Unused members should be zero ...
	 */
	memset(&new, 0, sizeof(new));

	/*
	 * If you change siginfo_t structure, please be sure
	 * this code is fixed accordingly.
	 */
	new.ssi_signo = kinfo->si_signo;
	new.ssi_errno = kinfo->si_errno;
	new.ssi_code  = kinfo->si_code;
	switch (siginfo_layout(kinfo->si_signo, kinfo->si_code)) {
	case SIL_KILL:
		new.ssi_pid = kinfo->si_pid;
		new.ssi_uid = kinfo->si_uid;
		break;
	case SIL_TIMER:
		new.ssi_tid = kinfo->si_tid;
		new.ssi_overrun = kinfo->si_overrun;
		new.ssi_ptr = (long) kinfo->si_ptr;
		new.ssi_int = kinfo->si_int;
		break;
	case SIL_POLL:
		new.ssi_band = kinfo->si_band;
		new.ssi_fd   = kinfo->si_fd;
		break;
	case SIL_FAULT_BNDERR:
	case SIL_FAULT_PKUERR:
	case SIL_FAULT_PERF_EVENT:
		/*
		 * Fall through to the SIL_FAULT case.  SIL_FAULT_BNDERR,
		 * SIL_FAULT_PKUERR, and SIL_FAULT_PERF_EVENT are only
		 * generated by faults that deliver them synchronously to
		 * userspace.  In case someone injects one of these signals
		 * and signalfd catches it treat it as SIL_FAULT.
		 */
	case SIL_FAULT:
		new.ssi_addr = (long) kinfo->si_addr;
		break;
	case SIL_FAULT_TRAPNO:
		new.ssi_addr = (long) kinfo->si_addr;
		new.ssi_trapno = kinfo->si_trapno;
		break;
	case SIL_FAULT_MCEERR:
		new.ssi_addr = (long) kinfo->si_addr;
		new.ssi_addr_lsb = (short) kinfo->si_addr_lsb;
		break;
	case SIL_CHLD:
		new.ssi_pid    = kinfo->si_pid;
		new.ssi_uid    = kinfo->si_uid;
		new.ssi_status = kinfo->si_status;
		new.ssi_utime  = kinfo->si_utime;
		new.ssi_stime  = kinfo->si_stime;
		break;
	case SIL_RT:
		/*
		 * This case catches also the signals queued by sigqueue().
		 */
		new.ssi_pid = kinfo->si_pid;
		new.ssi_uid = kinfo->si_uid;
		new.ssi_ptr = (long) kinfo->si_ptr;
		new.ssi_int = kinfo->si_int;
		break;
	case SIL_SYS:
		new.ssi_call_addr = (long) kinfo->si_call_addr;
		new.ssi_syscall   = kinfo->si_syscall;
		new.ssi_arch      = kinfo->si_arch;
		break;
	}

	if (copy_to_user(uinfo, &new, sizeof(struct signalfd_siginfo)))
		return -EFAULT;

	return sizeof(*uinfo);
}

static ssize_t signalfd_dequeue(struct signalfd_ctx *ctx, kernel_siginfo_t *info,
				int nonblock)
{
	ssize_t ret;
	DECLARE_WAITQUEUE(wait, current);

	spin_lock_irq(&current->sighand->siglock);
	ret = dequeue_signal(current, &ctx->sigmask, info);
	switch (ret) {
	case 0:
		if (!nonblock)
			break;
		ret = -EAGAIN;
		fallthrough;
	default:
		spin_unlock_irq(&current->sighand->siglock);
		return ret;
	}

	add_wait_queue(&current->sighand->signalfd_wqh, &wait);
	for (;;) {
		set_current_state(TASK_INTERRUPTIBLE);
		ret = dequeue_signal(current, &ctx->sigmask, info);
		if (ret != 0)
			break;
		if (signal_pending(current)) {
			ret = -ERESTARTSYS;
			break;
		}
		spin_unlock_irq(&current->sighand->siglock);
		schedule();
		spin_lock_irq(&current->sighand->siglock);
	}
	spin_unlock_irq(&current->sighand->siglock);

	remove_wait_queue(&current->sighand->signalfd_wqh, &wait);
	__set_current_state(TASK_RUNNING);

	return ret;
}

/*
 * Returns a multiple of the size of a "struct signalfd_siginfo", or a negative
 * error code. The "count" parameter must be at least the size of a
 * "struct signalfd_siginfo".
 */
static ssize_t signalfd_read(struct file *file, char __user *buf, size_t count,
			     loff_t *ppos)
{
	struct signalfd_ctx *ctx = file->private_data;
	struct signalfd_siginfo __user *siginfo;
	int nonblock = file->f_flags & O_NONBLOCK;
	ssize_t ret, total = 0;
	kernel_siginfo_t info;

	count /= sizeof(struct signalfd_siginfo);
	if (!count)
		return -EINVAL;

	siginfo = (struct signalfd_siginfo __user *) buf;
	do {
		ret = signalfd_dequeue(ctx, &info, nonblock);
		if (unlikely(ret <= 0))
			break;
		ret = signalfd_copyinfo(siginfo, &info);
		if (ret < 0)
			break;
		siginfo++;
		total += ret;
		nonblock = 1;
	} while (--count);

	return total ? total: ret;
}

#ifdef CONFIG_PROC_FS
static void signalfd_show_fdinfo(struct seq_file *m, struct file *f)
{
	struct signalfd_ctx *ctx = f->private_data;
	sigset_t sigmask;

	sigmask = ctx->sigmask;
	signotset(&sigmask);
	render_sigset_t(m, "sigmask:\t", &sigmask);
}
#endif

static const struct file_operations signalfd_fops = {
#ifdef CONFIG_PROC_FS
	.show_fdinfo	= signalfd_show_fdinfo,
#endif
	.release	= signalfd_release,
	.poll		= signalfd_poll,
	.read		= signalfd_read,
	.llseek		= noop_llseek,
};

static int do_signalfd4(int ufd, sigset_t *mask, int flags)
{
	struct signalfd_ctx *ctx;

	/* Check the SFD_* constants for consistency.  */
	BUILD_BUG_ON(SFD_CLOEXEC != O_CLOEXEC);
	BUILD_BUG_ON(SFD_NONBLOCK != O_NONBLOCK);

	if (flags & ~(SFD_CLOEXEC | SFD_NONBLOCK))
		return -EINVAL;

	sigdelsetmask(mask, sigmask(SIGKILL) | sigmask(SIGSTOP));
	signotset(mask);

	if (ufd == -1) {
		ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
		if (!ctx)
			return -ENOMEM;

		ctx->sigmask = *mask;

		/*
		 * When we call this, the initialization must be complete, since
		 * anon_inode_getfd() will install the fd.
		 */
		ufd = anon_inode_getfd("[signalfd]", &signalfd_fops, ctx,
				       O_RDWR | (flags & (O_CLOEXEC | O_NONBLOCK)));
		if (ufd < 0)
			kfree(ctx);
	} else {
		struct fd f = fdget(ufd);
		if (!f.file)
			return -EBADF;
		ctx = f.file->private_data;
		if (f.file->f_op != &signalfd_fops) {
			fdput(f);
			return -EINVAL;
		}
		spin_lock_irq(&current->sighand->siglock);
		ctx->sigmask = *mask;
		spin_unlock_irq(&current->sighand->siglock);

		wake_up(&current->sighand->signalfd_wqh);
		fdput(f);
	}

	return ufd;
}

SYSCALL_DEFINE4(signalfd4, int, ufd, sigset_t __user *, user_mask,
		size_t, sizemask, int, flags)
{
	sigset_t mask;

	if (sizemask != sizeof(sigset_t))
		return -EINVAL;
	if (copy_from_user(&mask, user_mask, sizeof(mask)))
		return -EFAULT;
	return do_signalfd4(ufd, &mask, flags);
}

SYSCALL_DEFINE3(signalfd, int, ufd, sigset_t __user *, user_mask,
		size_t, sizemask)
{
	sigset_t mask;

	if (sizemask != sizeof(sigset_t))
		return -EINVAL;
	if (copy_from_user(&mask, user_mask, sizeof(mask)))
		return -EFAULT;
	return do_signalfd4(ufd, &mask, 0);
}

#ifdef CONFIG_COMPAT
static long do_compat_signalfd4(int ufd,
			const compat_sigset_t __user *user_mask,
			compat_size_t sigsetsize, int flags)
{
	sigset_t mask;

	if (sigsetsize != sizeof(compat_sigset_t))
		return -EINVAL;
	if (get_compat_sigset(&mask, user_mask))
		return -EFAULT;
	return do_signalfd4(ufd, &mask, flags);
}

COMPAT_SYSCALL_DEFINE4(signalfd4, int, ufd,
		     const compat_sigset_t __user *, user_mask,
		     compat_size_t, sigsetsize,
		     int, flags)
{
	return do_compat_signalfd4(ufd, user_mask, sigsetsize, flags);
}

COMPAT_SYSCALL_DEFINE3(signalfd, int, ufd,
		     const compat_sigset_t __user *, user_mask,
		     compat_size_t, sigsetsize)
{
	return do_compat_signalfd4(ufd, user_mask, sigsetsize, 0);
}
#endif
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
fba2afaa DL	2	/*
	3	* fs/signalfd.c
	4	*
	5	* Copyright (C) 2003 Linus Torvalds
	6	*
	7	* Mon Mar 5, 2007: Davide Libenzi <davidel@xmailserver.org>
	8	* Changed ->read() to return a siginfo strcture instead of signal number.
	9	* Fixed locking in ->poll().
	10	* Added sighand-detach notification.
	11	* Added fd re-use in sys_signalfd() syscall.
	12	* Now using anonymous inode source.
	13	* Thanks to Oleg Nesterov for useful code review and suggestions.
	14	* More comments and suggestions from Arnd Bergmann.
b8fceee1	15	* Sat May 19, 2007: Davi E. M. Arnaut <davi@haxent.com.br>
b3762bfc	16	* Retrieve multiple signals with one read() call
b8fceee1 DL	17	* Sun Jul 15, 2007: Davide Libenzi <davidel@xmailserver.org>
b8fceee1 DL	18	* Attach to the sighand only during read() and poll().
fba2afaa DL	19	*/
	20
	21	#include <linux/file.h>
	22	#include <linux/poll.h>
	23	#include <linux/init.h>
	24	#include <linux/fs.h>
	25	#include <linux/sched.h>
5a0e3ad6	26	#include <linux/slab.h>
fba2afaa DL	27	#include <linux/kernel.h>
	28	#include <linux/signal.h>
	29	#include <linux/list.h>
	30	#include <linux/anon_inodes.h>
	31	#include <linux/signalfd.h>
7ec37dfd	32	#include <linux/syscalls.h>
138d22b5	33	#include <linux/proc_fs.h>
7d197ed4	34	#include <linux/compat.h>
fba2afaa	35
d80e731e ON	36	void signalfd_cleanup(struct sighand_struct *sighand)
	37	{
	38	wait_queue_head_t *wqh = &sighand->signalfd_wqh;
971316f0 ON	39	/*
	40	* The lockless check can race with remove_wait_queue() in progress,
	41	* but in this case its caller should run under rcu_read_lock() and
5f0d5a3a	42	* sighand_cachep is SLAB_TYPESAFE_BY_RCU, we can safely return.
971316f0	43	*/
d80e731e ON	44	if (likely(!waitqueue_active(wqh)))
	45	return;
	46
ac6424b9	47	/* wait_queue_entry_t->func(POLLFREE) should do remove_wait_queue() */
a9a08845	48	wake_up_poll(wqh, EPOLLHUP \| POLLFREE);
d80e731e ON	49	}
d80e731e ON	50
fba2afaa	51	struct signalfd_ctx {
fba2afaa	52	sigset_t sigmask;
fba2afaa DL	53	};
fba2afaa DL	54
fba2afaa DL	55	static int signalfd_release(struct inode inode, struct file file)
fba2afaa DL	56	{
b8fceee1	57	kfree(file->private_data);
fba2afaa DL	58	return 0;
	59	}
	60
076ccb76	61	static __poll_t signalfd_poll(struct file file, poll_table wait)
fba2afaa DL	62	{
fba2afaa DL	63	struct signalfd_ctx *ctx = file->private_data;
076ccb76	64	__poll_t events = 0;
fba2afaa	65
b8fceee1	66	poll_wait(file, &current->sighand->signalfd_wqh, wait);
fba2afaa	67
b8fceee1 DL	68	spin_lock_irq(&current->sighand->siglock);
	69	if (next_signal(&current->pending, &ctx->sigmask) \|\|
	70	next_signal(&current->signal->shared_pending,
	71	&ctx->sigmask))
a9a08845	72	events \|= EPOLLIN;
b8fceee1	73	spin_unlock_irq(&current->sighand->siglock);
fba2afaa DL	74
	75	return events;
	76	}
	77
	78	/*
	79	* Copied from copy_siginfo_to_user() in kernel/signal.c
	80	*/
	81	static int signalfd_copyinfo(struct signalfd_siginfo __user *uinfo,
ae7795bc	82	kernel_siginfo_t const *kinfo)
fba2afaa	83	{
5611f55e	84	struct signalfd_siginfo new;
fba2afaa DL	85
	86	BUILD_BUG_ON(sizeof(struct signalfd_siginfo) != 128);
	87
	88	/*
14e4a0f2	89	* Unused members should be zero ...
fba2afaa	90	*/
5611f55e	91	memset(&new, 0, sizeof(new));
fba2afaa DL	92
	93	/*
	94	* If you change siginfo_t structure, please be sure
	95	* this code is fixed accordingly.
	96	*/
5611f55e EB	97	new.ssi_signo = kinfo->si_signo;
	98	new.ssi_errno = kinfo->si_errno;
	99	new.ssi_code = kinfo->si_code;
cc731525 EB	100	switch (siginfo_layout(kinfo->si_signo, kinfo->si_code)) {
cc731525 EB	101	case SIL_KILL:
5611f55e EB	102	new.ssi_pid = kinfo->si_pid;
5611f55e EB	103	new.ssi_uid = kinfo->si_uid;
fba2afaa	104	break;
cc731525	105	case SIL_TIMER:
5611f55e EB	106	new.ssi_tid = kinfo->si_tid;
	107	new.ssi_overrun = kinfo->si_overrun;
	108	new.ssi_ptr = (long) kinfo->si_ptr;
	109	new.ssi_int = kinfo->si_int;
fba2afaa	110	break;
cc731525	111	case SIL_POLL:
5611f55e EB	112	new.ssi_band = kinfo->si_band;
5611f55e EB	113	new.ssi_fd = kinfo->si_fd;
fba2afaa	114	break;
31931c93 EB	115	case SIL_FAULT_BNDERR:
31931c93 EB	116	case SIL_FAULT_PKUERR:
f4ac7302	117	case SIL_FAULT_PERF_EVENT:
31931c93	118	/*
922e3013	119	* Fall through to the SIL_FAULT case. SIL_FAULT_BNDERR,
f4ac7302	120	* SIL_FAULT_PKUERR, and SIL_FAULT_PERF_EVENT are only
922e3013 EB	121	* generated by faults that deliver them synchronously to
	122	* userspace. In case someone injects one of these signals
	123	* and signalfd catches it treat it as SIL_FAULT.
31931c93	124	*/
cc731525	125	case SIL_FAULT:
5611f55e	126	new.ssi_addr = (long) kinfo->si_addr;
9abcabe3 EB	127	break;
	128	case SIL_FAULT_TRAPNO:
	129	new.ssi_addr = (long) kinfo->si_addr;
5611f55e	130	new.ssi_trapno = kinfo->si_trapno;
31931c93 EB	131	break;
	132	case SIL_FAULT_MCEERR:
	133	new.ssi_addr = (long) kinfo->si_addr;
31931c93	134	new.ssi_addr_lsb = (short) kinfo->si_addr_lsb;
fba2afaa	135	break;
cc731525	136	case SIL_CHLD:
5611f55e EB	137	new.ssi_pid = kinfo->si_pid;
	138	new.ssi_uid = kinfo->si_uid;
	139	new.ssi_status = kinfo->si_status;
	140	new.ssi_utime = kinfo->si_utime;
	141	new.ssi_stime = kinfo->si_stime;
fba2afaa	142	break;
cc731525	143	case SIL_RT:
0859ab59 DL	144	/*
	145	* This case catches also the signals queued by sigqueue().
	146	*/
5611f55e EB	147	new.ssi_pid = kinfo->si_pid;
	148	new.ssi_uid = kinfo->si_uid;
	149	new.ssi_ptr = (long) kinfo->si_ptr;
	150	new.ssi_int = kinfo->si_int;
fba2afaa	151	break;
76b7f670 EB	152	case SIL_SYS:
	153	new.ssi_call_addr = (long) kinfo->si_call_addr;
	154	new.ssi_syscall = kinfo->si_syscall;
	155	new.ssi_arch = kinfo->si_arch;
	156	break;
fba2afaa DL	157	}
fba2afaa DL	158
5611f55e EB	159	if (copy_to_user(uinfo, &new, sizeof(struct signalfd_siginfo)))
	160	return -EFAULT;
	161
	162	return sizeof(*uinfo);
fba2afaa DL	163	}
fba2afaa DL	164
ae7795bc	165	static ssize_t signalfd_dequeue(struct signalfd_ctx ctx, kernel_siginfo_t info,
b3762bfc DA	166	int nonblock)
	167	{
	168	ssize_t ret;
b3762bfc DA	169	DECLARE_WAITQUEUE(wait, current);
b3762bfc DA	170
b8fceee1 DL	171	spin_lock_irq(&current->sighand->siglock);
b8fceee1 DL	172	ret = dequeue_signal(current, &ctx->sigmask, info);
b3762bfc DA	173	switch (ret) {
	174	case 0:
	175	if (!nonblock)
	176	break;
	177	ret = -EAGAIN;
df561f66	178	fallthrough;
b3762bfc	179	default:
b8fceee1	180	spin_unlock_irq(&current->sighand->siglock);
b3762bfc DA	181	return ret;
	182	}
	183
b8fceee1	184	add_wait_queue(&current->sighand->signalfd_wqh, &wait);
b3762bfc DA	185	for (;;) {
b3762bfc DA	186	set_current_state(TASK_INTERRUPTIBLE);
b8fceee1	187	ret = dequeue_signal(current, &ctx->sigmask, info);
b3762bfc DA	188	if (ret != 0)
	189	break;
	190	if (signal_pending(current)) {
	191	ret = -ERESTARTSYS;
	192	break;
	193	}
b8fceee1	194	spin_unlock_irq(&current->sighand->siglock);
b3762bfc	195	schedule();
b8fceee1	196	spin_lock_irq(&current->sighand->siglock);
b3762bfc	197	}
b8fceee1	198	spin_unlock_irq(&current->sighand->siglock);
b3762bfc	199
b8fceee1	200	remove_wait_queue(&current->sighand->signalfd_wqh, &wait);
b3762bfc DA	201	__set_current_state(TASK_RUNNING);
	202
	203	return ret;
	204	}
	205
fba2afaa	206	/*
b8fceee1 DL	207	* Returns a multiple of the size of a "struct signalfd_siginfo", or a negative
	208	* error code. The "count" parameter must be at least the size of a
	209	* "struct signalfd_siginfo".
fba2afaa DL	210	*/
	211	static ssize_t signalfd_read(struct file file, char __user buf, size_t count,
	212	loff_t *ppos)
	213	{
	214	struct signalfd_ctx *ctx = file->private_data;
b3762bfc DA	215	struct signalfd_siginfo __user *siginfo;
	216	int nonblock = file->f_flags & O_NONBLOCK;
	217	ssize_t ret, total = 0;
ae7795bc	218	kernel_siginfo_t info;
fba2afaa	219
b3762bfc DA	220	count /= sizeof(struct signalfd_siginfo);
b3762bfc DA	221	if (!count)
fba2afaa	222	return -EINVAL;
fba2afaa	223
b3762bfc	224	siginfo = (struct signalfd_siginfo __user *) buf;
b3762bfc DA	225	do {
	226	ret = signalfd_dequeue(ctx, &info, nonblock);
	227	if (unlikely(ret <= 0))
	228	break;
	229	ret = signalfd_copyinfo(siginfo, &info);
	230	if (ret < 0)
	231	break;
	232	siginfo++;
	233	total += ret;
	234	nonblock = 1;
	235	} while (--count);
	236
b8fceee1	237	return total ? total: ret;
fba2afaa DL	238	}
fba2afaa DL	239
138d22b5	240	#ifdef CONFIG_PROC_FS
a3816ab0	241	static void signalfd_show_fdinfo(struct seq_file m, struct file f)
138d22b5 CG	242	{
	243	struct signalfd_ctx *ctx = f->private_data;
	244	sigset_t sigmask;
	245
	246	sigmask = ctx->sigmask;
	247	signotset(&sigmask);
	248	render_sigset_t(m, "sigmask:\t", &sigmask);
138d22b5 CG	249	}
	250	#endif
	251
fba2afaa	252	static const struct file_operations signalfd_fops = {
138d22b5 CG	253	#ifdef CONFIG_PROC_FS
	254	.show_fdinfo = signalfd_show_fdinfo,
	255	#endif
fba2afaa DL	256	.release = signalfd_release,
	257	.poll = signalfd_poll,
	258	.read = signalfd_read,
6038f373	259	.llseek = noop_llseek,
fba2afaa DL	260	};
fba2afaa DL	261
5ed0127f	262	static int do_signalfd4(int ufd, sigset_t *mask, int flags)
fba2afaa	263	{
fba2afaa	264	struct signalfd_ctx *ctx;
fba2afaa	265
e38b36f3 UD	266	/* Check the SFD_* constants for consistency. */
	267	BUILD_BUG_ON(SFD_CLOEXEC != O_CLOEXEC);
	268	BUILD_BUG_ON(SFD_NONBLOCK != O_NONBLOCK);
	269
5fb5e049	270	if (flags & ~(SFD_CLOEXEC \| SFD_NONBLOCK))
9deb27ba UD	271	return -EINVAL;
9deb27ba UD	272
5ed0127f AV	273	sigdelsetmask(mask, sigmask(SIGKILL) \| sigmask(SIGSTOP));
5ed0127f AV	274	signotset(mask);
fba2afaa DL	275
	276	if (ufd == -1) {
	277	ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
	278	if (!ctx)
	279	return -ENOMEM;
	280
5ed0127f	281	ctx->sigmask = *mask;
fba2afaa DL	282
	283	/*
	284	* When we call this, the initialization must be complete, since
	285	* anon_inode_getfd() will install the fd.
	286	*/
7d9dbca3	287	ufd = anon_inode_getfd("[signalfd]", &signalfd_fops, ctx,
628ff7c1	288	O_RDWR \| (flags & (O_CLOEXEC \| O_NONBLOCK)));
2030a42c AV	289	if (ufd < 0)
2030a42c AV	290	kfree(ctx);
fba2afaa	291	} else {
2903ff01 AV	292	struct fd f = fdget(ufd);
2903ff01 AV	293	if (!f.file)
fba2afaa	294	return -EBADF;
2903ff01 AV	295	ctx = f.file->private_data;
	296	if (f.file->f_op != &signalfd_fops) {
	297	fdput(f);
fba2afaa DL	298	return -EINVAL;
fba2afaa DL	299	}
b8fceee1	300	spin_lock_irq(&current->sighand->siglock);
5ed0127f	301	ctx->sigmask = *mask;
b8fceee1 DL	302	spin_unlock_irq(&current->sighand->siglock);
	303
	304	wake_up(&current->sighand->signalfd_wqh);
2903ff01	305	fdput(f);
fba2afaa DL	306	}
	307
	308	return ufd;
fba2afaa	309	}
9deb27ba	310
52fb6db0 DB	311	SYSCALL_DEFINE4(signalfd4, int, ufd, sigset_t __user *, user_mask,
	312	size_t, sizemask, int, flags)
	313	{
5ed0127f AV	314	sigset_t mask;
5ed0127f AV	315
a089e3fd	316	if (sizemask != sizeof(sigset_t))
5ed0127f	317	return -EINVAL;
a089e3fd HD	318	if (copy_from_user(&mask, user_mask, sizeof(mask)))
a089e3fd HD	319	return -EFAULT;
5ed0127f	320	return do_signalfd4(ufd, &mask, flags);
52fb6db0 DB	321	}
52fb6db0 DB	322
836f92ad HC	323	SYSCALL_DEFINE3(signalfd, int, ufd, sigset_t __user *, user_mask,
836f92ad HC	324	size_t, sizemask)
9deb27ba	325	{
5ed0127f AV	326	sigset_t mask;
5ed0127f AV	327
a089e3fd	328	if (sizemask != sizeof(sigset_t))
5ed0127f	329	return -EINVAL;
a089e3fd HD	330	if (copy_from_user(&mask, user_mask, sizeof(mask)))
a089e3fd HD	331	return -EFAULT;
5ed0127f	332	return do_signalfd4(ufd, &mask, 0);
9deb27ba	333	}
7d197ed4 AV	334
7d197ed4 AV	335	#ifdef CONFIG_COMPAT
570484bf	336	static long do_compat_signalfd4(int ufd,
5ed0127f	337	const compat_sigset_t __user *user_mask,
570484bf	338	compat_size_t sigsetsize, int flags)
7d197ed4	339	{
5ed0127f	340	sigset_t mask;
7d197ed4 AV	341
	342	if (sigsetsize != sizeof(compat_sigset_t))
	343	return -EINVAL;
5ed0127f	344	if (get_compat_sigset(&mask, user_mask))
7d197ed4	345	return -EFAULT;
5ed0127f	346	return do_signalfd4(ufd, &mask, flags);
7d197ed4 AV	347	}
7d197ed4 AV	348
570484bf	349	COMPAT_SYSCALL_DEFINE4(signalfd4, int, ufd,
5ed0127f	350	const compat_sigset_t __user *, user_mask,
570484bf DB	351	compat_size_t, sigsetsize,
	352	int, flags)
	353	{
5ed0127f	354	return do_compat_signalfd4(ufd, user_mask, sigsetsize, flags);
570484bf DB	355	}
570484bf DB	356
7d197ed4	357	COMPAT_SYSCALL_DEFINE3(signalfd, int, ufd,
5ed0127f	358	const compat_sigset_t __user *, user_mask,
7d197ed4 AV	359	compat_size_t, sigsetsize)
7d197ed4 AV	360	{
5ed0127f	361	return do_compat_signalfd4(ufd, user_mask, sigsetsize, 0);
7d197ed4 AV	362	}
7d197ed4 AV	363	#endif