[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);
	if (!(file = fcheck(oldfd)))
		goto out_unlock;
	get_file(file);			/* We are now finished with oldfd */

	err = expand_files(files, newfd);
	if (unlikely(err < 0)) {
		if (err == -EMFILE)
			err = -EBADF;
		goto out_fput;
	}

	/* To avoid races with open() and dup(), we will mark the fd as
	 * in-use in the open-file bitmap throughout the entire dup2()
	 * process.  This is quite safe: do_close() uses the fd array
	 * entry, not the bitmap, to decide what work needs to be
	 * done.  --sct */
	/* Doesn't work. open() might be there first. --AV */

	/* Yes. It's a race. In user space. Nothing sane to do */
	err = -EBUSY;
	fdt = files_fdtable(files);
	tofree = fdt->fd[newfd];
	if (!tofree && FD_ISSET(newfd, fdt->open_fds))
		goto out_fput;

	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);
	err = newfd;
out:
	return err;
out_unlock:
	spin_unlock(&files->file_lock);
	goto out;

out_fput:
	spin_unlock(&files->file_lock);
	fput(file);
	goto out;
}

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