[net-next-2.6.git] / fs / fcntl.c

/*
 *  linux/fs/fcntl.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 */

#include <linux/syscalls.h>
#include <linux/init.h>
#include <linux/mm.h>
#include <linux/fs.h>
#include <linux/file.h>
#include <linux/fdtable.h>
#include <linux/capability.h>
#include <linux/dnotify.h>
#include <linux/slab.h>
#include <linux/module.h>
#include <linux/security.h>
#include <linux/ptrace.h>
#include <linux/signal.h>
#include <linux/rcupdate.h>
#include <linux/pid_namespace.h>

#include <asm/poll.h>
#include <asm/siginfo.h>
#include <asm/uaccess.h>

void set_close_on_exec(unsigned int fd, int flag)
{
	struct files_struct *files = current->files;
	struct fdtable *fdt;
	spin_lock(&files->file_lock);
	fdt = files_fdtable(files);
	if (flag)
		FD_SET(fd, fdt->close_on_exec);
	else
		FD_CLR(fd, fdt->close_on_exec);
	spin_unlock(&files->file_lock);
}

static int get_close_on_exec(unsigned int fd)
{
	struct files_struct *files = current->files;
	struct fdtable *fdt;
	int res;
	rcu_read_lock();
	fdt = files_fdtable(files);
	res = FD_ISSET(fd, fdt->close_on_exec);
	rcu_read_unlock();
	return res;
}

asmlinkage long sys_dup3(unsigned int oldfd, unsigned int newfd, int flags)
{
	int err = -EBADF;
	struct file * file, *tofree;
	struct files_struct * files = current->files;
	struct fdtable *fdt;

	if ((flags & ~O_CLOEXEC) != 0)
		return -EINVAL;

	if (unlikely(oldfd == newfd))
		return -EINVAL;

	spin_lock(&files->file_lock);
	err = expand_files(files, newfd);
	file = fcheck(oldfd);
	if (unlikely(!file))
		goto Ebadf;
	if (unlikely(err < 0)) {
		if (err == -EMFILE)
			goto Ebadf;
		goto out_unlock;
	}
	/*
	 * We need to detect attempts to do dup2() over allocated but still
	 * not finished descriptor.  NB: OpenBSD avoids that at the price of
	 * extra work in their equivalent of fget() - they insert struct
	 * file immediately after grabbing descriptor, mark it larval if
	 * more work (e.g. actual opening) is needed and make sure that
	 * fget() treats larval files as absent.  Potentially interesting,
	 * but while extra work in fget() is trivial, locking implications
	 * and amount of surgery on open()-related paths in VFS are not.
	 * FreeBSD fails with -EBADF in the same situation, NetBSD "solution"
	 * deadlocks in rather amusing ways, AFAICS.  All of that is out of
	 * scope of POSIX or SUS, since neither considers shared descriptor
	 * tables and this condition does not arise without those.
	 */
	err = -EBUSY;
	fdt = files_fdtable(files);
	tofree = fdt->fd[newfd];
	if (!tofree && FD_ISSET(newfd, fdt->open_fds))
		goto out_unlock;
	get_file(file);
	rcu_assign_pointer(fdt->fd[newfd], file);
	FD_SET(newfd, fdt->open_fds);
	if (flags & O_CLOEXEC)
		FD_SET(newfd, fdt->close_on_exec);
	else
		FD_CLR(newfd, fdt->close_on_exec);
	spin_unlock(&files->file_lock);

	if (tofree)
		filp_close(tofree, files);

	return newfd;

Ebadf:
	err = -EBADF;
out_unlock:
	spin_unlock(&files->file_lock);
	return err;
}

asmlinkage long sys_dup2(unsigned int oldfd, unsigned int newfd)
{
	if (unlikely(newfd == oldfd)) { /* corner case */
		struct files_struct *files = current->files;
		rcu_read_lock();
		if (!fcheck_files(files, oldfd))
			oldfd = -EBADF;
		rcu_read_unlock();
		return oldfd;
	}
	return sys_dup3(oldfd, newfd, 0);
}

asmlinkage long sys_dup(unsigned int fildes)
{
	int ret = -EBADF;
	struct file *file = fget(fildes);

	if (file) {
		ret = get_unused_fd();
		if (ret >= 0)
			fd_install(ret, file);
		else
			fput(file);
	}
	return ret;
}

#define SETFL_MASK (O_APPEND | O_NONBLOCK | O_NDELAY | FASYNC | O_DIRECT | O_NOATIME)

static int setfl(int fd, struct file * filp, unsigned long arg)
{
	struct inode * inode = filp->f_path.dentry->d_inode;
	int error = 0;

	/*
	 * O_APPEND cannot be cleared if the file is marked as append-only
	 * and the file is open for write.
	 */
	if (((arg ^ filp->f_flags) & O_APPEND) && IS_APPEND(inode))
		return -EPERM;

	/* O_NOATIME can only be set by the owner or superuser */
	if ((arg & O_NOATIME) && !(filp->f_flags & O_NOATIME))
		if (!is_owner_or_cap(inode))
			return -EPERM;

	/* required for strict SunOS emulation */
	if (O_NONBLOCK != O_NDELAY)
	       if (arg & O_NDELAY)
		   arg |= O_NONBLOCK;

	if (arg & O_DIRECT) {
		if (!filp->f_mapping || !filp->f_mapping->a_ops ||
			!filp->f_mapping->a_ops->direct_IO)
				return -EINVAL;
	}

	if (filp->f_op && filp->f_op->check_flags)
		error = filp->f_op->check_flags(arg);
	if (error)
		return error;

	if ((arg ^ filp->f_flags) & FASYNC) {
		if (filp->f_op && filp->f_op->fasync) {
			error = filp->f_op->fasync(fd, filp, (arg & FASYNC) != 0);
			if (error < 0)
				goto out;
		}
	}

	filp->f_flags = (arg & SETFL_MASK) | (filp->f_flags & ~SETFL_MASK);
 out:
	return error;
}

static void f_modown(struct file *filp, struct pid *pid, enum pid_type type,
                     uid_t uid, uid_t euid, int force)
{
	write_lock_irq(&filp->f_owner.lock);
	if (force || !filp->f_owner.pid) {
		put_pid(filp->f_owner.pid);
		filp->f_owner.pid = get_pid(pid);
		filp->f_owner.pid_type = type;
		filp->f_owner.uid = uid;
		filp->f_owner.euid = euid;
	}
	write_unlock_irq(&filp->f_owner.lock);
}

int __f_setown(struct file *filp, struct pid *pid, enum pid_type type,
		int force)
{
	int err;
	
	err = security_file_set_fowner(filp);
	if (err)
		return err;

	f_modown(filp, pid, type, current->uid, current->euid, force);
	return 0;
}
EXPORT_SYMBOL(__f_setown);

int f_setown(struct file *filp, unsigned long arg, int force)
{
	enum pid_type type;
	struct pid *pid;
	int who = arg;
	int result;
	type = PIDTYPE_PID;
	if (who < 0) {
		type = PIDTYPE_PGID;
		who = -who;
	}
	rcu_read_lock();
	pid = find_vpid(who);
	result = __f_setown(filp, pid, type, force);
	rcu_read_unlock();
	return result;
}
EXPORT_SYMBOL(f_setown);

void f_delown(struct file *filp)
{
	f_modown(filp, NULL, PIDTYPE_PID, 0, 0, 1);
}

pid_t f_getown(struct file *filp)
{
	pid_t pid;
	read_lock(&filp->f_owner.lock);
	pid = pid_vnr(filp->f_owner.pid);
	if (filp->f_owner.pid_type == PIDTYPE_PGID)
		pid = -pid;
	read_unlock(&filp->f_owner.lock);
	return pid;
}

static long do_fcntl(int fd, unsigned int cmd, unsigned long arg,
		struct file *filp)
{
	long err = -EINVAL;

	switch (cmd) {
	case F_DUPFD:
	case F_DUPFD_CLOEXEC:
		if (arg >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur)
			break;
		err = alloc_fd(arg, cmd == F_DUPFD_CLOEXEC ? O_CLOEXEC : 0);
		if (err >= 0) {
			get_file(filp);
			fd_install(err, filp);
		}
		break;
	case F_GETFD:
		err = get_close_on_exec(fd) ? FD_CLOEXEC : 0;
		break;
	case F_SETFD:
		err = 0;
		set_close_on_exec(fd, arg & FD_CLOEXEC);
		break;
	case F_GETFL:
		err = filp->f_flags;
		break;
	case F_SETFL:
		err = setfl(fd, filp, arg);
		break;
	case F_GETLK:
		err = fcntl_getlk(filp, (struct flock __user *) arg);
		break;
	case F_SETLK:
	case F_SETLKW:
		err = fcntl_setlk(fd, filp, cmd, (struct flock __user *) arg);
		break;
	case F_GETOWN:
		/*
		 * XXX If f_owner is a process group, the
		 * negative return value will get converted
		 * into an error.  Oops.  If we keep the
		 * current syscall conventions, the only way
		 * to fix this will be in libc.
		 */
		err = f_getown(filp);
		force_successful_syscall_return();
		break;
	case F_SETOWN:
		err = f_setown(filp, arg, 1);
		break;
	case F_GETSIG:
		err = filp->f_owner.signum;
		break;
	case F_SETSIG:
		/* arg == 0 restores default behaviour. */
		if (!valid_signal(arg)) {
			break;
		}
		err = 0;
		filp->f_owner.signum = arg;
		break;
	case F_GETLEASE:
		err = fcntl_getlease(filp);
		break;
	case F_SETLEASE:
		err = fcntl_setlease(fd, filp, arg);
		break;
	case F_NOTIFY:
		err = fcntl_dirnotify(fd, filp, arg);
		break;
	default:
		break;
	}
	return err;
}

asmlinkage long sys_fcntl(unsigned int fd, unsigned int cmd, unsigned long arg)
{	
	struct file *filp;
	long err = -EBADF;

	filp = fget(fd);
	if (!filp)
		goto out;

	err = security_file_fcntl(filp, cmd, arg);
	if (err) {
		fput(filp);
		return err;
	}

	err = do_fcntl(fd, cmd, arg, filp);

 	fput(filp);
out:
	return err;
}

#if BITS_PER_LONG == 32
asmlinkage long sys_fcntl64(unsigned int fd, unsigned int cmd, unsigned long arg)
{	
	struct file * filp;
	long err;

	err = -EBADF;
	filp = fget(fd);
	if (!filp)
		goto out;

	err = security_file_fcntl(filp, cmd, arg);
	if (err) {
		fput(filp);
		return err;
	}
	err = -EBADF;
	
	switch (cmd) {
		case F_GETLK64:
			err = fcntl_getlk64(filp, (struct flock64 __user *) arg);
			break;
		case F_SETLK64:
		case F_SETLKW64:
			err = fcntl_setlk64(fd, filp, cmd,
					(struct flock64 __user *) arg);
			break;
		default:
			err = do_fcntl(fd, cmd, arg, filp);
			break;
	}
	fput(filp);
out:
	return err;
}
#endif

/* Table to convert sigio signal codes into poll band bitmaps */

static const long band_table[NSIGPOLL] = {
	POLLIN | POLLRDNORM,			/* POLL_IN */
	POLLOUT | POLLWRNORM | POLLWRBAND,	/* POLL_OUT */
	POLLIN | POLLRDNORM | POLLMSG,		/* POLL_MSG */
	POLLERR,				/* POLL_ERR */
	POLLPRI | POLLRDBAND,			/* POLL_PRI */
	POLLHUP | POLLERR			/* POLL_HUP */
};

static inline int sigio_perm(struct task_struct *p,
                             struct fown_struct *fown, int sig)
{
	return (((fown->euid == 0) ||
		 (fown->euid == p->suid) || (fown->euid == p->uid) ||
		 (fown->uid == p->suid) || (fown->uid == p->uid)) &&
		!security_file_send_sigiotask(p, fown, sig));
}

static void send_sigio_to_task(struct task_struct *p,
			       struct fown_struct *fown, 
			       int fd,
			       int reason)
{
	if (!sigio_perm(p, fown, fown->signum))
		return;

	switch (fown->signum) {
		siginfo_t si;
		default:
			/* Queue a rt signal with the appropriate fd as its
			   value.  We use SI_SIGIO as the source, not 
			   SI_KERNEL, since kernel signals always get 
			   delivered even if we can't queue.  Failure to
			   queue in this case _should_ be reported; we fall
			   back to SIGIO in that case. --sct */
			si.si_signo = fown->signum;
			si.si_errno = 0;
		        si.si_code  = reason;
			/* Make sure we are called with one of the POLL_*
			   reasons, otherwise we could leak kernel stack into
			   userspace.  */
			BUG_ON((reason & __SI_MASK) != __SI_POLL);
			if (reason - POLL_IN >= NSIGPOLL)
				si.si_band  = ~0L;
			else
				si.si_band = band_table[reason - POLL_IN];
			si.si_fd    = fd;
			if (!group_send_sig_info(fown->signum, &si, p))
				break;
		/* fall-through: fall back on the old plain SIGIO signal */
		case 0:
			group_send_sig_info(SIGIO, SEND_SIG_PRIV, p);
	}
}

void send_sigio(struct fown_struct *fown, int fd, int band)
{
	struct task_struct *p;
	enum pid_type type;
	struct pid *pid;
	
	read_lock(&fown->lock);
	type = fown->pid_type;
	pid = fown->pid;
	if (!pid)
		goto out_unlock_fown;
	
	read_lock(&tasklist_lock);
	do_each_pid_task(pid, type, p) {
		send_sigio_to_task(p, fown, fd, band);
	} while_each_pid_task(pid, type, p);
	read_unlock(&tasklist_lock);
 out_unlock_fown:
	read_unlock(&fown->lock);
}

static void send_sigurg_to_task(struct task_struct *p,
                                struct fown_struct *fown)
{
	if (sigio_perm(p, fown, SIGURG))
		group_send_sig_info(SIGURG, SEND_SIG_PRIV, p);
}

int send_sigurg(struct fown_struct *fown)
{
	struct task_struct *p;
	enum pid_type type;
	struct pid *pid;
	int ret = 0;
	
	read_lock(&fown->lock);
	type = fown->pid_type;
	pid = fown->pid;
	if (!pid)
		goto out_unlock_fown;

	ret = 1;
	
	read_lock(&tasklist_lock);
	do_each_pid_task(pid, type, p) {
		send_sigurg_to_task(p, fown);
	} while_each_pid_task(pid, type, p);
	read_unlock(&tasklist_lock);
 out_unlock_fown:
	read_unlock(&fown->lock);
	return ret;
}

static DEFINE_RWLOCK(fasync_lock);
static struct kmem_cache *fasync_cache __read_mostly;

/*
 * fasync_helper() is used by some character device drivers (mainly mice)
 * to set up the fasync queue. It returns negative on error, 0 if it did
 * no changes and positive if it added/deleted the entry.
 */
int fasync_helper(int fd, struct file * filp, int on, struct fasync_struct **fapp)
{
	struct fasync_struct *fa, **fp;
	struct fasync_struct *new = NULL;
	int result = 0;

	if (on) {
		new = kmem_cache_alloc(fasync_cache, GFP_KERNEL);
		if (!new)
			return -ENOMEM;
	}
	write_lock_irq(&fasync_lock);
	for (fp = fapp; (fa = *fp) != NULL; fp = &fa->fa_next) {
		if (fa->fa_file == filp) {
			if(on) {
				fa->fa_fd = fd;
				kmem_cache_free(fasync_cache, new);
			} else {
				*fp = fa->fa_next;
				kmem_cache_free(fasync_cache, fa);
				result = 1;
			}
			goto out;
		}
	}

	if (on) {
		new->magic = FASYNC_MAGIC;
		new->fa_file = filp;
		new->fa_fd = fd;
		new->fa_next = *fapp;
		*fapp = new;
		result = 1;
	}
out:
	write_unlock_irq(&fasync_lock);
	return result;
}

EXPORT_SYMBOL(fasync_helper);

void __kill_fasync(struct fasync_struct *fa, int sig, int band)
{
	while (fa) {
		struct fown_struct * fown;
		if (fa->magic != FASYNC_MAGIC) {
			printk(KERN_ERR "kill_fasync: bad magic number in "
			       "fasync_struct!\n");
			return;
		}
		fown = &fa->fa_file->f_owner;
		/* Don't send SIGURG to processes which have not set a
		   queued signum: SIGURG has its own default signalling
		   mechanism. */
		if (!(sig == SIGURG && fown->signum == 0))
			send_sigio(fown, fa->fa_fd, band);
		fa = fa->fa_next;
	}
}

EXPORT_SYMBOL(__kill_fasync);

void kill_fasync(struct fasync_struct **fp, int sig, int band)
{
	/* First a quick test without locking: usually
	 * the list is empty.
	 */
	if (*fp) {
		read_lock(&fasync_lock);
		/* reread *fp after obtaining the lock */
		__kill_fasync(*fp, sig, band);
		read_unlock(&fasync_lock);
	}
}
EXPORT_SYMBOL(kill_fasync);

static int __init fasync_init(void)
{
	fasync_cache = kmem_cache_create("fasync_cache",
		sizeof(struct fasync_struct), 0, SLAB_PANIC, NULL);
	return 0;
}

module_init(fasync_init)
Commit	Line	Data
1da177e4 LT	1	/*
	2	* linux/fs/fcntl.c
	3	*
	4	* Copyright (C) 1991, 1992 Linus Torvalds
	5	*/
	6
	7	#include <linux/syscalls.h>
	8	#include <linux/init.h>
	9	#include <linux/mm.h>
	10	#include <linux/fs.h>
	11	#include <linux/file.h>
9f3acc31	12	#include <linux/fdtable.h>
16f7e0fe	13	#include <linux/capability.h>
1da177e4	14	#include <linux/dnotify.h>
1da177e4 LT	15	#include <linux/slab.h>
	16	#include <linux/module.h>
	17	#include <linux/security.h>
	18	#include <linux/ptrace.h>
7ed20e1a	19	#include <linux/signal.h>
ab2af1f5	20	#include <linux/rcupdate.h>
b488893a	21	#include <linux/pid_namespace.h>
1da177e4 LT	22
	23	#include <asm/poll.h>
	24	#include <asm/siginfo.h>
	25	#include <asm/uaccess.h>
	26
fc9b52cd	27	void set_close_on_exec(unsigned int fd, int flag)
1da177e4 LT	28	{
1da177e4 LT	29	struct files_struct *files = current->files;
badf1662	30	struct fdtable *fdt;
1da177e4	31	spin_lock(&files->file_lock);
badf1662	32	fdt = files_fdtable(files);
1da177e4	33	if (flag)
badf1662	34	FD_SET(fd, fdt->close_on_exec);
1da177e4	35	else
badf1662	36	FD_CLR(fd, fdt->close_on_exec);
1da177e4 LT	37	spin_unlock(&files->file_lock);
	38	}
	39
858119e1	40	static int get_close_on_exec(unsigned int fd)
1da177e4 LT	41	{
1da177e4 LT	42	struct files_struct *files = current->files;
badf1662	43	struct fdtable *fdt;
1da177e4	44	int res;
b835996f	45	rcu_read_lock();
badf1662 DS	46	fdt = files_fdtable(files);
badf1662 DS	47	res = FD_ISSET(fd, fdt->close_on_exec);
b835996f	48	rcu_read_unlock();
1da177e4 LT	49	return res;
	50	}
	51
336dd1f7	52	asmlinkage long sys_dup3(unsigned int oldfd, unsigned int newfd, int flags)
1da177e4 LT	53	{
	54	int err = -EBADF;
	55	struct file * file, *tofree;
	56	struct files_struct * files = current->files;
badf1662	57	struct fdtable *fdt;
1da177e4	58
336dd1f7 UD	59	if ((flags & ~O_CLOEXEC) != 0)
	60	return -EINVAL;
	61
6c5d0512 AV	62	if (unlikely(oldfd == newfd))
	63	return -EINVAL;
	64
1da177e4	65	spin_lock(&files->file_lock);
1da177e4	66	err = expand_files(files, newfd);
1b7e190b AV	67	file = fcheck(oldfd);
	68	if (unlikely(!file))
	69	goto Ebadf;
4e1e018e AV	70	if (unlikely(err < 0)) {
4e1e018e AV	71	if (err == -EMFILE)
1b7e190b AV	72	goto Ebadf;
1b7e190b AV	73	goto out_unlock;
4e1e018e	74	}
1b7e190b AV	75	/*
	76	* We need to detect attempts to do dup2() over allocated but still
	77	* not finished descriptor. NB: OpenBSD avoids that at the price of
	78	* extra work in their equivalent of fget() - they insert struct
	79	* file immediately after grabbing descriptor, mark it larval if
	80	* more work (e.g. actual opening) is needed and make sure that
	81	* fget() treats larval files as absent. Potentially interesting,
	82	* but while extra work in fget() is trivial, locking implications
	83	* and amount of surgery on open()-related paths in VFS are not.
	84	* FreeBSD fails with -EBADF in the same situation, NetBSD "solution"
	85	* deadlocks in rather amusing ways, AFAICS. All of that is out of
	86	* scope of POSIX or SUS, since neither considers shared descriptor
	87	* tables and this condition does not arise without those.
	88	*/
1da177e4	89	err = -EBUSY;
badf1662 DS	90	fdt = files_fdtable(files);
	91	tofree = fdt->fd[newfd];
	92	if (!tofree && FD_ISSET(newfd, fdt->open_fds))
1b7e190b AV	93	goto out_unlock;
1b7e190b AV	94	get_file(file);
ab2af1f5	95	rcu_assign_pointer(fdt->fd[newfd], file);
badf1662	96	FD_SET(newfd, fdt->open_fds);
336dd1f7 UD	97	if (flags & O_CLOEXEC)
	98	FD_SET(newfd, fdt->close_on_exec);
	99	else
	100	FD_CLR(newfd, fdt->close_on_exec);
1da177e4 LT	101	spin_unlock(&files->file_lock);
	102
	103	if (tofree)
	104	filp_close(tofree, files);
1da177e4	105
1b7e190b AV	106	return newfd;
	107
	108	Ebadf:
	109	err = -EBADF;
	110	out_unlock:
1da177e4	111	spin_unlock(&files->file_lock);
1b7e190b	112	return err;
1da177e4	113	}
336dd1f7 UD	114
	115	asmlinkage long sys_dup2(unsigned int oldfd, unsigned int newfd)
	116	{
6c5d0512 AV	117	if (unlikely(newfd == oldfd)) { /* corner case */
	118	struct files_struct *files = current->files;
	119	rcu_read_lock();
	120	if (!fcheck_files(files, oldfd))
	121	oldfd = -EBADF;
	122	rcu_read_unlock();
	123	return oldfd;
	124	}
336dd1f7 UD	125	return sys_dup3(oldfd, newfd, 0);
336dd1f7 UD	126	}
1da177e4 LT	127
	128	asmlinkage long sys_dup(unsigned int fildes)
	129	{
	130	int ret = -EBADF;
1027abe8 AV	131	struct file *file = fget(fildes);
	132
	133	if (file) {
	134	ret = get_unused_fd();
	135	if (ret >= 0)
	136	fd_install(ret, file);
	137	else
	138	fput(file);
	139	}
1da177e4 LT	140	return ret;
	141	}
	142
	143	#define SETFL_MASK (O_APPEND \| O_NONBLOCK \| O_NDELAY \| FASYNC \| O_DIRECT \| O_NOATIME)
	144
	145	static int setfl(int fd, struct file * filp, unsigned long arg)
	146	{
0f7fc9e4	147	struct inode * inode = filp->f_path.dentry->d_inode;
1da177e4 LT	148	int error = 0;
1da177e4 LT	149
7d95c8f2	150	/*
	151	* O_APPEND cannot be cleared if the file is marked as append-only
	152	* and the file is open for write.
	153	*/
	154	if (((arg ^ filp->f_flags) & O_APPEND) && IS_APPEND(inode))
1da177e4 LT	155	return -EPERM;
	156
	157	/* O_NOATIME can only be set by the owner or superuser */
	158	if ((arg & O_NOATIME) && !(filp->f_flags & O_NOATIME))
3bd858ab	159	if (!is_owner_or_cap(inode))
1da177e4 LT	160	return -EPERM;
	161
	162	/* required for strict SunOS emulation */
	163	if (O_NONBLOCK != O_NDELAY)
	164	if (arg & O_NDELAY)
	165	arg \|= O_NONBLOCK;
	166
	167	if (arg & O_DIRECT) {
	168	if (!filp->f_mapping \|\| !filp->f_mapping->a_ops \|\|
	169	!filp->f_mapping->a_ops->direct_IO)
	170	return -EINVAL;
	171	}
	172
	173	if (filp->f_op && filp->f_op->check_flags)
	174	error = filp->f_op->check_flags(arg);
	175	if (error)
	176	return error;
	177
1da177e4 LT	178	if ((arg ^ filp->f_flags) & FASYNC) {
	179	if (filp->f_op && filp->f_op->fasync) {
	180	error = filp->f_op->fasync(fd, filp, (arg & FASYNC) != 0);
	181	if (error < 0)
	182	goto out;
	183	}
	184	}
	185
	186	filp->f_flags = (arg & SETFL_MASK) \| (filp->f_flags & ~SETFL_MASK);
	187	out:
1da177e4 LT	188	return error;
	189	}
	190
609d7fa9	191	static void f_modown(struct file filp, struct pid pid, enum pid_type type,
1da177e4 LT	192	uid_t uid, uid_t euid, int force)
	193	{
	194	write_lock_irq(&filp->f_owner.lock);
	195	if (force \|\| !filp->f_owner.pid) {
609d7fa9 EB	196	put_pid(filp->f_owner.pid);
	197	filp->f_owner.pid = get_pid(pid);
	198	filp->f_owner.pid_type = type;
1da177e4 LT	199	filp->f_owner.uid = uid;
	200	filp->f_owner.euid = euid;
	201	}
	202	write_unlock_irq(&filp->f_owner.lock);
	203	}
	204
609d7fa9 EB	205	int __f_setown(struct file filp, struct pid pid, enum pid_type type,
609d7fa9 EB	206	int force)
1da177e4 LT	207	{
	208	int err;
	209
	210	err = security_file_set_fowner(filp);
	211	if (err)
	212	return err;
	213
609d7fa9	214	f_modown(filp, pid, type, current->uid, current->euid, force);
1da177e4 LT	215	return 0;
1da177e4 LT	216	}
609d7fa9	217	EXPORT_SYMBOL(__f_setown);
1da177e4	218
609d7fa9 EB	219	int f_setown(struct file *filp, unsigned long arg, int force)
	220	{
	221	enum pid_type type;
	222	struct pid *pid;
	223	int who = arg;
	224	int result;
	225	type = PIDTYPE_PID;
	226	if (who < 0) {
	227	type = PIDTYPE_PGID;
	228	who = -who;
	229	}
	230	rcu_read_lock();
b488893a	231	pid = find_vpid(who);
609d7fa9 EB	232	result = __f_setown(filp, pid, type, force);
	233	rcu_read_unlock();
	234	return result;
	235	}
1da177e4 LT	236	EXPORT_SYMBOL(f_setown);
	237
	238	void f_delown(struct file *filp)
	239	{
609d7fa9 EB	240	f_modown(filp, NULL, PIDTYPE_PID, 0, 0, 1);
	241	}
	242
	243	pid_t f_getown(struct file *filp)
	244	{
	245	pid_t pid;
43fa1adb	246	read_lock(&filp->f_owner.lock);
6c5f3e7b	247	pid = pid_vnr(filp->f_owner.pid);
609d7fa9 EB	248	if (filp->f_owner.pid_type == PIDTYPE_PGID)
609d7fa9 EB	249	pid = -pid;
43fa1adb	250	read_unlock(&filp->f_owner.lock);
609d7fa9	251	return pid;
1da177e4 LT	252	}
	253
	254	static long do_fcntl(int fd, unsigned int cmd, unsigned long arg,
	255	struct file *filp)
	256	{
	257	long err = -EINVAL;
	258
	259	switch (cmd) {
	260	case F_DUPFD:
22d2b35b	261	case F_DUPFD_CLOEXEC:
4e1e018e AV	262	if (arg >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur)
4e1e018e AV	263	break;
1027abe8 AV	264	err = alloc_fd(arg, cmd == F_DUPFD_CLOEXEC ? O_CLOEXEC : 0);
	265	if (err >= 0) {
	266	get_file(filp);
	267	fd_install(err, filp);
	268	}
1da177e4 LT	269	break;
	270	case F_GETFD:
	271	err = get_close_on_exec(fd) ? FD_CLOEXEC : 0;
	272	break;
	273	case F_SETFD:
	274	err = 0;
	275	set_close_on_exec(fd, arg & FD_CLOEXEC);
	276	break;
	277	case F_GETFL:
	278	err = filp->f_flags;
	279	break;
	280	case F_SETFL:
	281	err = setfl(fd, filp, arg);
	282	break;
	283	case F_GETLK:
	284	err = fcntl_getlk(filp, (struct flock __user *) arg);
	285	break;
	286	case F_SETLK:
	287	case F_SETLKW:
c293621b	288	err = fcntl_setlk(fd, filp, cmd, (struct flock __user *) arg);
1da177e4 LT	289	break;
	290	case F_GETOWN:
	291	/*
	292	* XXX If f_owner is a process group, the
	293	* negative return value will get converted
	294	* into an error. Oops. If we keep the
	295	* current syscall conventions, the only way
	296	* to fix this will be in libc.
	297	*/
609d7fa9	298	err = f_getown(filp);
1da177e4 LT	299	force_successful_syscall_return();
	300	break;
	301	case F_SETOWN:
	302	err = f_setown(filp, arg, 1);
	303	break;
	304	case F_GETSIG:
	305	err = filp->f_owner.signum;
	306	break;
	307	case F_SETSIG:
	308	/* arg == 0 restores default behaviour. */
7ed20e1a	309	if (!valid_signal(arg)) {
1da177e4 LT	310	break;
	311	}
	312	err = 0;
	313	filp->f_owner.signum = arg;
	314	break;
	315	case F_GETLEASE:
	316	err = fcntl_getlease(filp);
	317	break;
	318	case F_SETLEASE:
	319	err = fcntl_setlease(fd, filp, arg);
	320	break;
	321	case F_NOTIFY:
	322	err = fcntl_dirnotify(fd, filp, arg);
	323	break;
	324	default:
	325	break;
	326	}
	327	return err;
	328	}
	329
	330	asmlinkage long sys_fcntl(unsigned int fd, unsigned int cmd, unsigned long arg)
	331	{
	332	struct file *filp;
	333	long err = -EBADF;
	334
	335	filp = fget(fd);
	336	if (!filp)
	337	goto out;
	338
	339	err = security_file_fcntl(filp, cmd, arg);
	340	if (err) {
	341	fput(filp);
	342	return err;
	343	}
	344
	345	err = do_fcntl(fd, cmd, arg, filp);
	346
	347	fput(filp);
	348	out:
	349	return err;
	350	}
	351
	352	#if BITS_PER_LONG == 32
	353	asmlinkage long sys_fcntl64(unsigned int fd, unsigned int cmd, unsigned long arg)
	354	{
	355	struct file * filp;
	356	long err;
	357
	358	err = -EBADF;
	359	filp = fget(fd);
	360	if (!filp)
	361	goto out;
	362
	363	err = security_file_fcntl(filp, cmd, arg);
	364	if (err) {
	365	fput(filp);
	366	return err;
	367	}
	368	err = -EBADF;
	369
	370	switch (cmd) {
	371	case F_GETLK64:
	372	err = fcntl_getlk64(filp, (struct flock64 __user *) arg);
	373	break;
374	case F_SETLK64:
375	case F_SETLKW64:
c293621b PS	376	err = fcntl_setlk64(fd, filp, cmd,
c293621b PS	377	(struct flock64 __user *) arg);
1da177e4 LT	378	break;
	379	default:
	380	err = do_fcntl(fd, cmd, arg, filp);
	381	break;
	382	}
	383	fput(filp);
	384	out:
	385	return err;
	386	}
	387	#endif
	388
	389	/* Table to convert sigio signal codes into poll band bitmaps */
	390
fa3536cc	391	static const long band_table[NSIGPOLL] = {
1da177e4 LT	392	POLLIN \| POLLRDNORM, /* POLL_IN */
	393	POLLOUT \| POLLWRNORM \| POLLWRBAND, /* POLL_OUT */
	394	POLLIN \| POLLRDNORM \| POLLMSG, /* POLL_MSG */
	395	POLLERR, /* POLL_ERR */
	396	POLLPRI \| POLLRDBAND, /* POLL_PRI */
	397	POLLHUP \| POLLERR /* POLL_HUP */
	398	};
	399
	400	static inline int sigio_perm(struct task_struct *p,
	401	struct fown_struct *fown, int sig)
	402	{
	403	return (((fown->euid == 0) \|\|
	404	(fown->euid == p->suid) \|\| (fown->euid == p->uid) \|\|
	405	(fown->uid == p->suid) \|\| (fown->uid == p->uid)) &&
	406	!security_file_send_sigiotask(p, fown, sig));
	407	}
	408
	409	static void send_sigio_to_task(struct task_struct *p,
	410	struct fown_struct *fown,
	411	int fd,
	412	int reason)
	413	{
	414	if (!sigio_perm(p, fown, fown->signum))
	415	return;
	416
	417	switch (fown->signum) {
	418	siginfo_t si;
	419	default:
	420	/* Queue a rt signal with the appropriate fd as its
	421	value. We use SI_SIGIO as the source, not
	422	SI_KERNEL, since kernel signals always get
	423	delivered even if we can't queue. Failure to
	424	queue in this case _should_ be reported; we fall
	425	back to SIGIO in that case. --sct */
	426	si.si_signo = fown->signum;
	427	si.si_errno = 0;
	428	si.si_code = reason;
	429	/* Make sure we are called with one of the POLL_*
	430	reasons, otherwise we could leak kernel stack into
	431	userspace. */
f6298aab	432	BUG_ON((reason & __SI_MASK) != __SI_POLL);
1da177e4 LT	433	if (reason - POLL_IN >= NSIGPOLL)
	434	si.si_band = ~0L;
	435	else
	436	si.si_band = band_table[reason - POLL_IN];
	437	si.si_fd = fd;
850d6fbe	438	if (!group_send_sig_info(fown->signum, &si, p))
1da177e4 LT	439	break;
	440	/* fall-through: fall back on the old plain SIGIO signal */
	441	case 0:
850d6fbe	442	group_send_sig_info(SIGIO, SEND_SIG_PRIV, p);
1da177e4 LT	443	}
	444	}
	445
	446	void send_sigio(struct fown_struct *fown, int fd, int band)
	447	{
	448	struct task_struct *p;
609d7fa9 EB	449	enum pid_type type;
609d7fa9 EB	450	struct pid *pid;
1da177e4 LT	451
1da177e4 LT	452	read_lock(&fown->lock);
609d7fa9	453	type = fown->pid_type;
1da177e4 LT	454	pid = fown->pid;
	455	if (!pid)
	456	goto out_unlock_fown;
	457
	458	read_lock(&tasklist_lock);
609d7fa9 EB	459	do_each_pid_task(pid, type, p) {
	460	send_sigio_to_task(p, fown, fd, band);
	461	} while_each_pid_task(pid, type, p);
1da177e4 LT	462	read_unlock(&tasklist_lock);
	463	out_unlock_fown:
	464	read_unlock(&fown->lock);
	465	}
	466
	467	static void send_sigurg_to_task(struct task_struct *p,
	468	struct fown_struct *fown)
	469	{
	470	if (sigio_perm(p, fown, SIGURG))
850d6fbe	471	group_send_sig_info(SIGURG, SEND_SIG_PRIV, p);
1da177e4 LT	472	}
	473
	474	int send_sigurg(struct fown_struct *fown)
	475	{
	476	struct task_struct *p;
609d7fa9 EB	477	enum pid_type type;
	478	struct pid *pid;
	479	int ret = 0;
1da177e4 LT	480
1da177e4 LT	481	read_lock(&fown->lock);
609d7fa9	482	type = fown->pid_type;
1da177e4 LT	483	pid = fown->pid;
	484	if (!pid)
	485	goto out_unlock_fown;
	486
	487	ret = 1;
	488
	489	read_lock(&tasklist_lock);
609d7fa9 EB	490	do_each_pid_task(pid, type, p) {
	491	send_sigurg_to_task(p, fown);
	492	} while_each_pid_task(pid, type, p);
1da177e4 LT	493	read_unlock(&tasklist_lock);
	494	out_unlock_fown:
	495	read_unlock(&fown->lock);
	496	return ret;
	497	}
	498
	499	static DEFINE_RWLOCK(fasync_lock);
e18b890b	500	static struct kmem_cache *fasync_cache __read_mostly;
1da177e4 LT	501
	502	/*
	503	* fasync_helper() is used by some character device drivers (mainly mice)
	504	* to set up the fasync queue. It returns negative on error, 0 if it did
	505	* no changes and positive if it added/deleted the entry.
	506	*/
	507	int fasync_helper(int fd, struct file * filp, int on, struct fasync_struct **fapp)
	508	{
	509	struct fasync_struct fa, *fp;
	510	struct fasync_struct *new = NULL;
	511	int result = 0;
	512
	513	if (on) {
e94b1766	514	new = kmem_cache_alloc(fasync_cache, GFP_KERNEL);
1da177e4 LT	515	if (!new)
	516	return -ENOMEM;
	517	}
	518	write_lock_irq(&fasync_lock);
	519	for (fp = fapp; (fa = *fp) != NULL; fp = &fa->fa_next) {
	520	if (fa->fa_file == filp) {
	521	if(on) {
	522	fa->fa_fd = fd;
	523	kmem_cache_free(fasync_cache, new);
	524	} else {
	525	*fp = fa->fa_next;
	526	kmem_cache_free(fasync_cache, fa);
	527	result = 1;
	528	}
	529	goto out;
	530	}
	531	}
	532
	533	if (on) {
	534	new->magic = FASYNC_MAGIC;
	535	new->fa_file = filp;
	536	new->fa_fd = fd;
	537	new->fa_next = *fapp;
	538	*fapp = new;
	539	result = 1;
	540	}
	541	out:
	542	write_unlock_irq(&fasync_lock);
	543	return result;
	544	}
	545
	546	EXPORT_SYMBOL(fasync_helper);
	547
	548	void __kill_fasync(struct fasync_struct *fa, int sig, int band)
	549	{
	550	while (fa) {
	551	struct fown_struct * fown;
	552	if (fa->magic != FASYNC_MAGIC) {
	553	printk(KERN_ERR "kill_fasync: bad magic number in "
	554	"fasync_struct!\n");
	555	return;
	556	}
	557	fown = &fa->fa_file->f_owner;
	558	/* Don't send SIGURG to processes which have not set a
	559	queued signum: SIGURG has its own default signalling
	560	mechanism. */
	561	if (!(sig == SIGURG && fown->signum == 0))
	562	send_sigio(fown, fa->fa_fd, band);
	563	fa = fa->fa_next;
	564	}
	565	}
	566
	567	EXPORT_SYMBOL(__kill_fasync);
	568
	569	void kill_fasync(struct fasync_struct **fp, int sig, int band)
	570	{
	571	/* First a quick test without locking: usually
	572	* the list is empty.
	573	*/
	574	if (*fp) {
	575	read_lock(&fasync_lock);
	576	/* reread fp after obtaining the lock /
	577	__kill_fasync(*fp, sig, band);
	578	read_unlock(&fasync_lock);
579	}
580	}
581	EXPORT_SYMBOL(kill_fasync);
582
583	static int __init fasync_init(void)
584	{
585	fasync_cache = kmem_cache_create("fasync_cache",
20c2df83	586	sizeof(struct fasync_struct), 0, SLAB_PANIC, NULL);
1da177e4 LT	587	return 0;
	588	}
	589
	590	module_init(fasync_init)