[linux.git] / fs / signalfd.c

/*
 *  fs/signalfd.c
 *
 *  Copyright (C) 2003  Linus Torvalds
 *
 *  Mon Mar 5, 2007: Davide Libenzi <[email protected]>
 *      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 <[email protected]>
 *      Retrieve multiple signals with one read() call
 *  Sun Jul 15, 2007: Davide Libenzi <[email protected]>
 *      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>

struct signalfd_ctx {
	sigset_t sigmask;
};

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

static unsigned int signalfd_poll(struct file *file, poll_table *wait)
{
	struct signalfd_ctx *ctx = file->private_data;
	unsigned int 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 |= POLLIN;
	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,
			     siginfo_t const *kinfo)
{
	long err;

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

	/*
	 * Unused members should be zero ...
	 */
	err = __clear_user(uinfo, sizeof(*uinfo));

	/*
	 * If you change siginfo_t structure, please be sure
	 * this code is fixed accordingly.
	 */
	err |= __put_user(kinfo->si_signo, &uinfo->ssi_signo);
	err |= __put_user(kinfo->si_errno, &uinfo->ssi_errno);
	err |= __put_user((short) kinfo->si_code, &uinfo->ssi_code);
	switch (kinfo->si_code & __SI_MASK) {
	case __SI_KILL:
		err |= __put_user(kinfo->si_pid, &uinfo->ssi_pid);
		err |= __put_user(kinfo->si_uid, &uinfo->ssi_uid);
		break;
	case __SI_TIMER:
		 err |= __put_user(kinfo->si_tid, &uinfo->ssi_tid);
		 err |= __put_user(kinfo->si_overrun, &uinfo->ssi_overrun);
		 err |= __put_user((long) kinfo->si_ptr, &uinfo->ssi_ptr);
		 err |= __put_user(kinfo->si_int, &uinfo->ssi_int);
		break;
	case __SI_POLL:
		err |= __put_user(kinfo->si_band, &uinfo->ssi_band);
		err |= __put_user(kinfo->si_fd, &uinfo->ssi_fd);
		break;
	case __SI_FAULT:
		err |= __put_user((long) kinfo->si_addr, &uinfo->ssi_addr);
#ifdef __ARCH_SI_TRAPNO
		err |= __put_user(kinfo->si_trapno, &uinfo->ssi_trapno);
#endif
#ifdef BUS_MCEERR_AO
		/* 
		 * Other callers might not initialize the si_lsb field,
		 * so check explicitly for the right codes here.
		 */
		if (kinfo->si_code == BUS_MCEERR_AR ||
		    kinfo->si_code == BUS_MCEERR_AO)
			err |= __put_user((short) kinfo->si_addr_lsb,
					  &uinfo->ssi_addr_lsb);
#endif
		break;
	case __SI_CHLD:
		err |= __put_user(kinfo->si_pid, &uinfo->ssi_pid);
		err |= __put_user(kinfo->si_uid, &uinfo->ssi_uid);
		err |= __put_user(kinfo->si_status, &uinfo->ssi_status);
		err |= __put_user(kinfo->si_utime, &uinfo->ssi_utime);
		err |= __put_user(kinfo->si_stime, &uinfo->ssi_stime);
		break;
	case __SI_RT: /* This is not generated by the kernel as of now. */
	case __SI_MESGQ: /* But this is */
		err |= __put_user(kinfo->si_pid, &uinfo->ssi_pid);
		err |= __put_user(kinfo->si_uid, &uinfo->ssi_uid);
		err |= __put_user((long) kinfo->si_ptr, &uinfo->ssi_ptr);
		err |= __put_user(kinfo->si_int, &uinfo->ssi_int);
		break;
	default:
		/*
		 * This case catches also the signals queued by sigqueue().
		 */
		err |= __put_user(kinfo->si_pid, &uinfo->ssi_pid);
		err |= __put_user(kinfo->si_uid, &uinfo->ssi_uid);
		err |= __put_user((long) kinfo->si_ptr, &uinfo->ssi_ptr);
		err |= __put_user(kinfo->si_int, &uinfo->ssi_int);
		break;
	}

	return err ? -EFAULT: sizeof(*uinfo);
}

static ssize_t signalfd_dequeue(struct signalfd_ctx *ctx, 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;
	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;
	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;
}

static const struct file_operations signalfd_fops = {
	.release	= signalfd_release,
	.poll		= signalfd_poll,
	.read		= signalfd_read,
	.llseek		= noop_llseek,
};

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

	if (sizemask != sizeof(sigset_t) ||
	    copy_from_user(&sigmask, user_mask, sizeof(sigmask)))
		return -EINVAL;
	sigdelsetmask(&sigmask, sigmask(SIGKILL) | sigmask(SIGSTOP));
	signotset(&sigmask);

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

		ctx->sigmask = sigmask;

		/*
		 * 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 file *file = fget(ufd);
		if (!file)
			return -EBADF;
		ctx = file->private_data;
		if (file->f_op != &signalfd_fops) {
			fput(file);
			return -EINVAL;
		}
		spin_lock_irq(&current->sighand->siglock);
		ctx->sigmask = sigmask;
		spin_unlock_irq(&current->sighand->siglock);

		wake_up(&current->sighand->signalfd_wqh);
		fput(file);
	}

	return ufd;
}

SYSCALL_DEFINE3(signalfd, int, ufd, sigset_t __user *, user_mask,
		size_t, sizemask)
{
	return sys_signalfd4(ufd, user_mask, sizemask, 0);
}
Commit	Line	Data
fba2afaa DL	1	/*
	2	* fs/signalfd.c
	3	*
	4	* Copyright (C) 2003 Linus Torvalds
	5	*
	6	* Mon Mar 5, 2007: Davide Libenzi <[email protected]>
	7	* Changed ->read() to return a siginfo strcture instead of signal number.
	8	* Fixed locking in ->poll().
	9	* Added sighand-detach notification.
	10	* Added fd re-use in sys_signalfd() syscall.
	11	* Now using anonymous inode source.
	12	* Thanks to Oleg Nesterov for useful code review and suggestions.
	13	* More comments and suggestions from Arnd Bergmann.
b8fceee1	14	* Sat May 19, 2007: Davi E. M. Arnaut <[email protected]>
b3762bfc	15	* Retrieve multiple signals with one read() call
b8fceee1 DL	16	* Sun Jul 15, 2007: Davide Libenzi <[email protected]>
b8fceee1 DL	17	* Attach to the sighand only during read() and poll().
fba2afaa DL	18	*/
	19
	20	#include <linux/file.h>
	21	#include <linux/poll.h>
	22	#include <linux/init.h>
	23	#include <linux/fs.h>
	24	#include <linux/sched.h>
5a0e3ad6	25	#include <linux/slab.h>
fba2afaa DL	26	#include <linux/kernel.h>
	27	#include <linux/signal.h>
	28	#include <linux/list.h>
	29	#include <linux/anon_inodes.h>
	30	#include <linux/signalfd.h>
7ec37dfd	31	#include <linux/syscalls.h>
fba2afaa DL	32
fba2afaa DL	33	struct signalfd_ctx {
fba2afaa	34	sigset_t sigmask;
fba2afaa DL	35	};
fba2afaa DL	36
fba2afaa DL	37	static int signalfd_release(struct inode inode, struct file file)
fba2afaa DL	38	{
b8fceee1	39	kfree(file->private_data);
fba2afaa DL	40	return 0;
	41	}
	42
	43	static unsigned int signalfd_poll(struct file file, poll_table wait)
	44	{
	45	struct signalfd_ctx *ctx = file->private_data;
	46	unsigned int events = 0;
fba2afaa	47
b8fceee1	48	poll_wait(file, &current->sighand->signalfd_wqh, wait);
fba2afaa	49
b8fceee1 DL	50	spin_lock_irq(&current->sighand->siglock);
	51	if (next_signal(&current->pending, &ctx->sigmask) \|\|
	52	next_signal(&current->signal->shared_pending,
	53	&ctx->sigmask))
fba2afaa	54	events \|= POLLIN;
b8fceee1	55	spin_unlock_irq(&current->sighand->siglock);
fba2afaa DL	56
	57	return events;
	58	}
	59
	60	/*
	61	* Copied from copy_siginfo_to_user() in kernel/signal.c
	62	*/
	63	static int signalfd_copyinfo(struct signalfd_siginfo __user *uinfo,
	64	siginfo_t const *kinfo)
	65	{
	66	long err;
	67
	68	BUILD_BUG_ON(sizeof(struct signalfd_siginfo) != 128);
	69
	70	/*
14e4a0f2	71	* Unused members should be zero ...
fba2afaa DL	72	*/
	73	err = __clear_user(uinfo, sizeof(*uinfo));
	74
	75	/*
	76	* If you change siginfo_t structure, please be sure
	77	* this code is fixed accordingly.
	78	*/
96358de6 DL	79	err \|= __put_user(kinfo->si_signo, &uinfo->ssi_signo);
	80	err \|= __put_user(kinfo->si_errno, &uinfo->ssi_errno);
	81	err \|= __put_user((short) kinfo->si_code, &uinfo->ssi_code);
fba2afaa DL	82	switch (kinfo->si_code & __SI_MASK) {
fba2afaa DL	83	case __SI_KILL:
96358de6 DL	84	err \|= __put_user(kinfo->si_pid, &uinfo->ssi_pid);
96358de6 DL	85	err \|= __put_user(kinfo->si_uid, &uinfo->ssi_uid);
fba2afaa DL	86	break;
fba2afaa DL	87	case __SI_TIMER:
96358de6 DL	88	err \|= __put_user(kinfo->si_tid, &uinfo->ssi_tid);
	89	err \|= __put_user(kinfo->si_overrun, &uinfo->ssi_overrun);
	90	err \|= __put_user((long) kinfo->si_ptr, &uinfo->ssi_ptr);
a2a20c41	91	err \|= __put_user(kinfo->si_int, &uinfo->ssi_int);
fba2afaa DL	92	break;
fba2afaa DL	93	case __SI_POLL:
96358de6 DL	94	err \|= __put_user(kinfo->si_band, &uinfo->ssi_band);
96358de6 DL	95	err \|= __put_user(kinfo->si_fd, &uinfo->ssi_fd);
fba2afaa DL	96	break;
fba2afaa DL	97	case __SI_FAULT:
96358de6	98	err \|= __put_user((long) kinfo->si_addr, &uinfo->ssi_addr);
fba2afaa	99	#ifdef __ARCH_SI_TRAPNO
96358de6	100	err \|= __put_user(kinfo->si_trapno, &uinfo->ssi_trapno);
b8aeec34 HS	101	#endif
	102	#ifdef BUS_MCEERR_AO
	103	/*
	104	* Other callers might not initialize the si_lsb field,
	105	* so check explicitly for the right codes here.
	106	*/
	107	if (kinfo->si_code == BUS_MCEERR_AR \|\|
	108	kinfo->si_code == BUS_MCEERR_AO)
	109	err \|= __put_user((short) kinfo->si_addr_lsb,
	110	&uinfo->ssi_addr_lsb);
fba2afaa DL	111	#endif
	112	break;
	113	case __SI_CHLD:
96358de6 DL	114	err \|= __put_user(kinfo->si_pid, &uinfo->ssi_pid);
	115	err \|= __put_user(kinfo->si_uid, &uinfo->ssi_uid);
	116	err \|= __put_user(kinfo->si_status, &uinfo->ssi_status);
	117	err \|= __put_user(kinfo->si_utime, &uinfo->ssi_utime);
	118	err \|= __put_user(kinfo->si_stime, &uinfo->ssi_stime);
fba2afaa DL	119	break;
	120	case __SI_RT: /* This is not generated by the kernel as of now. */
	121	case __SI_MESGQ: /* But this is */
96358de6 DL	122	err \|= __put_user(kinfo->si_pid, &uinfo->ssi_pid);
	123	err \|= __put_user(kinfo->si_uid, &uinfo->ssi_uid);
	124	err \|= __put_user((long) kinfo->si_ptr, &uinfo->ssi_ptr);
a2a20c41	125	err \|= __put_user(kinfo->si_int, &uinfo->ssi_int);
fba2afaa	126	break;
0859ab59 DL	127	default:
	128	/*
	129	* This case catches also the signals queued by sigqueue().
	130	*/
96358de6 DL	131	err \|= __put_user(kinfo->si_pid, &uinfo->ssi_pid);
96358de6 DL	132	err \|= __put_user(kinfo->si_uid, &uinfo->ssi_uid);
0859ab59 DL	133	err \|= __put_user((long) kinfo->si_ptr, &uinfo->ssi_ptr);
0859ab59 DL	134	err \|= __put_user(kinfo->si_int, &uinfo->ssi_int);
fba2afaa DL	135	break;
	136	}
	137
	138	return err ? -EFAULT: sizeof(*uinfo);
	139	}
	140
b3762bfc DA	141	static ssize_t signalfd_dequeue(struct signalfd_ctx ctx, siginfo_t info,
	142	int nonblock)
	143	{
	144	ssize_t ret;
b3762bfc DA	145	DECLARE_WAITQUEUE(wait, current);
b3762bfc DA	146
b8fceee1 DL	147	spin_lock_irq(&current->sighand->siglock);
b8fceee1 DL	148	ret = dequeue_signal(current, &ctx->sigmask, info);
b3762bfc DA	149	switch (ret) {
	150	case 0:
	151	if (!nonblock)
	152	break;
	153	ret = -EAGAIN;
	154	default:
b8fceee1	155	spin_unlock_irq(&current->sighand->siglock);
b3762bfc DA	156	return ret;
	157	}
	158
b8fceee1	159	add_wait_queue(&current->sighand->signalfd_wqh, &wait);
b3762bfc DA	160	for (;;) {
b3762bfc DA	161	set_current_state(TASK_INTERRUPTIBLE);
b8fceee1	162	ret = dequeue_signal(current, &ctx->sigmask, info);
b3762bfc DA	163	if (ret != 0)
	164	break;
	165	if (signal_pending(current)) {
	166	ret = -ERESTARTSYS;
	167	break;
	168	}
b8fceee1	169	spin_unlock_irq(&current->sighand->siglock);
b3762bfc	170	schedule();
b8fceee1	171	spin_lock_irq(&current->sighand->siglock);
b3762bfc	172	}
b8fceee1	173	spin_unlock_irq(&current->sighand->siglock);
b3762bfc	174
b8fceee1	175	remove_wait_queue(&current->sighand->signalfd_wqh, &wait);
b3762bfc DA	176	__set_current_state(TASK_RUNNING);
	177
	178	return ret;
	179	}
	180
fba2afaa	181	/*
b8fceee1 DL	182	* Returns a multiple of the size of a "struct signalfd_siginfo", or a negative
	183	* error code. The "count" parameter must be at least the size of a
	184	* "struct signalfd_siginfo".
fba2afaa DL	185	*/
	186	static ssize_t signalfd_read(struct file file, char __user buf, size_t count,
	187	loff_t *ppos)
	188	{
	189	struct signalfd_ctx *ctx = file->private_data;
b3762bfc DA	190	struct signalfd_siginfo __user *siginfo;
	191	int nonblock = file->f_flags & O_NONBLOCK;
	192	ssize_t ret, total = 0;
fba2afaa	193	siginfo_t info;
fba2afaa	194
b3762bfc DA	195	count /= sizeof(struct signalfd_siginfo);
b3762bfc DA	196	if (!count)
fba2afaa	197	return -EINVAL;
fba2afaa	198
b3762bfc	199	siginfo = (struct signalfd_siginfo __user *) buf;
b3762bfc DA	200	do {
	201	ret = signalfd_dequeue(ctx, &info, nonblock);
	202	if (unlikely(ret <= 0))
	203	break;
	204	ret = signalfd_copyinfo(siginfo, &info);
	205	if (ret < 0)
	206	break;
	207	siginfo++;
	208	total += ret;
	209	nonblock = 1;
	210	} while (--count);
	211
b8fceee1	212	return total ? total: ret;
fba2afaa DL	213	}
	214
	215	static const struct file_operations signalfd_fops = {
	216	.release = signalfd_release,
	217	.poll = signalfd_poll,
	218	.read = signalfd_read,
6038f373	219	.llseek = noop_llseek,
fba2afaa DL	220	};
fba2afaa DL	221
836f92ad HC	222	SYSCALL_DEFINE4(signalfd4, int, ufd, sigset_t __user *, user_mask,
836f92ad HC	223	size_t, sizemask, int, flags)
fba2afaa	224	{
fba2afaa DL	225	sigset_t sigmask;
fba2afaa DL	226	struct signalfd_ctx *ctx;
fba2afaa	227
e38b36f3 UD	228	/* Check the SFD_* constants for consistency. */
	229	BUILD_BUG_ON(SFD_CLOEXEC != O_CLOEXEC);
	230	BUILD_BUG_ON(SFD_NONBLOCK != O_NONBLOCK);
	231
5fb5e049	232	if (flags & ~(SFD_CLOEXEC \| SFD_NONBLOCK))
9deb27ba UD	233	return -EINVAL;
9deb27ba UD	234
fba2afaa DL	235	if (sizemask != sizeof(sigset_t) \|\|
fba2afaa DL	236	copy_from_user(&sigmask, user_mask, sizeof(sigmask)))
f50cadaa	237	return -EINVAL;
fba2afaa DL	238	sigdelsetmask(&sigmask, sigmask(SIGKILL) \| sigmask(SIGSTOP));
	239	signotset(&sigmask);
	240
	241	if (ufd == -1) {
	242	ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
	243	if (!ctx)
	244	return -ENOMEM;
	245
fba2afaa	246	ctx->sigmask = sigmask;
fba2afaa DL	247
	248	/*
	249	* When we call this, the initialization must be complete, since
	250	* anon_inode_getfd() will install the fd.
	251	*/
7d9dbca3	252	ufd = anon_inode_getfd("[signalfd]", &signalfd_fops, ctx,
628ff7c1	253	O_RDWR \| (flags & (O_CLOEXEC \| O_NONBLOCK)));
2030a42c AV	254	if (ufd < 0)
2030a42c AV	255	kfree(ctx);
fba2afaa	256	} else {
2030a42c	257	struct file *file = fget(ufd);
fba2afaa DL	258	if (!file)
	259	return -EBADF;
	260	ctx = file->private_data;
	261	if (file->f_op != &signalfd_fops) {
	262	fput(file);
	263	return -EINVAL;
	264	}
b8fceee1 DL	265	spin_lock_irq(&current->sighand->siglock);
	266	ctx->sigmask = sigmask;
	267	spin_unlock_irq(&current->sighand->siglock);
	268
	269	wake_up(&current->sighand->signalfd_wqh);
fba2afaa DL	270	fput(file);
	271	}
	272
	273	return ufd;
fba2afaa	274	}
9deb27ba	275
836f92ad HC	276	SYSCALL_DEFINE3(signalfd, int, ufd, sigset_t __user *, user_mask,
836f92ad HC	277	size_t, sizemask)
9deb27ba UD	278	{
	279	return sys_signalfd4(ufd, user_mask, sizemask, 0);
	280	}