[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/capability.h>
#include <linux/dnotify.h>
#include <linux/smp_lock.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 <asm/poll.h>
#include <asm/siginfo.h>
#include <asm/uaccess.h>

void fastcall 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;
}

/*
 * locate_fd finds a free file descriptor in the open_fds fdset,
 * expanding the fd arrays if necessary.  Must be called with the
 * file_lock held for write.
 */

static int locate_fd(struct files_struct *files, 
			    struct file *file, unsigned int orig_start)
{
	unsigned int newfd;
	unsigned int start;
	int error;
	struct fdtable *fdt;

	error = -EINVAL;
	if (orig_start >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur)
		goto out;

repeat:
	fdt = files_fdtable(files);
	/*
	 * Someone might have closed fd's in the range
	 * orig_start..fdt->next_fd
	 */
	start = orig_start;
	if (start < files->next_fd)
		start = files->next_fd;

	newfd = start;
	if (start < fdt->max_fdset) {
		newfd = find_next_zero_bit(fdt->open_fds->fds_bits,
			fdt->max_fdset, start);
	}
	
	error = -EMFILE;
	if (newfd >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur)
		goto out;

	error = expand_files(files, newfd);
	if (error < 0)
		goto out;

	/*
	 * If we needed to expand the fs array we
	 * might have blocked - try again.
	 */
	if (error)
		goto repeat;

	/*
	 * We reacquired files_lock, so we are safe as long as
	 * we reacquire the fdtable pointer and use it while holding
	 * the lock, no one can free it during that time.
	 */
	if (start <= files->next_fd)
		files->next_fd = newfd + 1;

	error = newfd;
	
out:
	return error;
}

static int dupfd(struct file *file, unsigned int start)
{
	struct files_struct * files = current->files;
	struct fdtable *fdt;
	int fd;

	spin_lock(&files->file_lock);
	fd = locate_fd(files, file, start);
	if (fd >= 0) {
		/* locate_fd() may have expanded fdtable, load the ptr */
		fdt = files_fdtable(files);
		FD_SET(fd, fdt->open_fds);
		FD_CLR(fd, fdt->close_on_exec);
		spin_unlock(&files->file_lock);
		fd_install(fd, file);
	} else {
		spin_unlock(&files->file_lock);
		fput(file);
	}

	return fd;
}

asmlinkage long sys_dup2(unsigned int oldfd, unsigned int newfd)
{
	int err = -EBADF;
	struct file * file, *tofree;
	struct files_struct * files = current->files;
	struct fdtable *fdt;

	spin_lock(&files->file_lock);
	if (!(file = fcheck(oldfd)))
		goto out_unlock;
	err = newfd;
	if (newfd == oldfd)
		goto out_unlock;
	err = -EBADF;
	if (newfd >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur)
		goto out_unlock;
	get_file(file);			/* We are now finished with oldfd */

	err = expand_files(files, newfd);
	if (err < 0)
		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);
	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_dup(unsigned int fildes)
{
	int ret = -EBADF;
	struct file * file = fget(fildes);

	if (file)
		ret = dupfd(file, 0);
	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_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 (current->fsuid != inode->i_uid && !capable(CAP_FOWNER))
			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;

	lock_kernel();
	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:
	unlock_kernel();
	return error;
}

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

int f_setown(struct file *filp, unsigned long arg, int force)
{
	int err;
	
	err = security_file_set_fowner(filp);
	if (err)
		return err;

	f_modown(filp, arg, current->uid, current->euid, force);
	return 0;
}

EXPORT_SYMBOL(f_setown);

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

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

	switch (cmd) {
	case F_DUPFD:
		get_file(filp);
		err = dupfd(filp, arg);
		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 = filp->f_owner.pid;
		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.  */
			if ((reason & __SI_MASK) != __SI_POLL)
				BUG();
			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;
	int pid;
	
	read_lock(&fown->lock);
	pid = fown->pid;
	if (!pid)
		goto out_unlock_fown;
	
	read_lock(&tasklist_lock);
	if (pid > 0) {
		p = find_task_by_pid(pid);
		if (p) {
			send_sigio_to_task(p, fown, fd, band);
		}
	} else {
		do_each_task_pid(-pid, PIDTYPE_PGID, p) {
			send_sigio_to_task(p, fown, fd, band);
		} while_each_task_pid(-pid, PIDTYPE_PGID, 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;
	int pid, ret = 0;
	
	read_lock(&fown->lock);
	pid = fown->pid;
	if (!pid)
		goto out_unlock_fown;

	ret = 1;
	
	read_lock(&tasklist_lock);
	if (pid > 0) {
		p = find_task_by_pid(pid);
		if (p) {
			send_sigurg_to_task(p, fown);
		}
	} else {
		do_each_task_pid(-pid, PIDTYPE_PGID, p) {
			send_sigurg_to_task(p, fown);
		} while_each_task_pid(-pid, PIDTYPE_PGID, p);
	}
	read_unlock(&tasklist_lock);
 out_unlock_fown:
	read_unlock(&fown->lock);
	return ret;
}

static DEFINE_RWLOCK(fasync_lock);
static kmem_cache_t *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, SLAB_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, 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>
16f7e0fe	12	#include <linux/capability.h>
1da177e4 LT	13	#include <linux/dnotify.h>
	14	#include <linux/smp_lock.h>
	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>
1da177e4 LT	21
	22	#include <asm/poll.h>
	23	#include <asm/siginfo.h>
	24	#include <asm/uaccess.h>
	25
	26	void fastcall set_close_on_exec(unsigned int fd, int flag)
	27	{
	28	struct files_struct *files = current->files;
badf1662	29	struct fdtable *fdt;
1da177e4	30	spin_lock(&files->file_lock);
badf1662	31	fdt = files_fdtable(files);
1da177e4	32	if (flag)
badf1662	33	FD_SET(fd, fdt->close_on_exec);
1da177e4	34	else
badf1662	35	FD_CLR(fd, fdt->close_on_exec);
1da177e4 LT	36	spin_unlock(&files->file_lock);
	37	}
	38
858119e1	39	static int get_close_on_exec(unsigned int fd)
1da177e4 LT	40	{
1da177e4 LT	41	struct files_struct *files = current->files;
badf1662	42	struct fdtable *fdt;
1da177e4	43	int res;
b835996f	44	rcu_read_lock();
badf1662 DS	45	fdt = files_fdtable(files);
badf1662 DS	46	res = FD_ISSET(fd, fdt->close_on_exec);
b835996f	47	rcu_read_unlock();
1da177e4 LT	48	return res;
	49	}
	50
	51	/*
	52	* locate_fd finds a free file descriptor in the open_fds fdset,
	53	* expanding the fd arrays if necessary. Must be called with the
	54	* file_lock held for write.
	55	*/
	56
	57	static int locate_fd(struct files_struct *files,
	58	struct file *file, unsigned int orig_start)
	59	{
	60	unsigned int newfd;
	61	unsigned int start;
	62	int error;
badf1662	63	struct fdtable *fdt;
1da177e4 LT	64
	65	error = -EINVAL;
	66	if (orig_start >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur)
	67	goto out;
	68
	69	repeat:
ab2af1f5	70	fdt = files_fdtable(files);
1da177e4 LT	71	/*
1da177e4 LT	72	* Someone might have closed fd's in the range
badf1662	73	* orig_start..fdt->next_fd
1da177e4 LT	74	*/
1da177e4 LT	75	start = orig_start;
0c9e63fd ED	76	if (start < files->next_fd)
0c9e63fd ED	77	start = files->next_fd;
1da177e4 LT	78
1da177e4 LT	79	newfd = start;
badf1662 DS	80	if (start < fdt->max_fdset) {
	81	newfd = find_next_zero_bit(fdt->open_fds->fds_bits,
	82	fdt->max_fdset, start);
1da177e4 LT	83	}
	84
	85	error = -EMFILE;
	86	if (newfd >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur)
	87	goto out;
	88
	89	error = expand_files(files, newfd);
	90	if (error < 0)
	91	goto out;
	92
	93	/*
	94	* If we needed to expand the fs array we
	95	* might have blocked - try again.
	96	*/
	97	if (error)
	98	goto repeat;
	99
ab2af1f5 DS	100	/*
	101	* We reacquired files_lock, so we are safe as long as
	102	* we reacquire the fdtable pointer and use it while holding
	103	* the lock, no one can free it during that time.
	104	*/
0c9e63fd ED	105	if (start <= files->next_fd)
0c9e63fd ED	106	files->next_fd = newfd + 1;
ab2af1f5	107
1da177e4 LT	108	error = newfd;
	109
	110	out:
	111	return error;
	112	}
	113
	114	static int dupfd(struct file *file, unsigned int start)
	115	{
	116	struct files_struct * files = current->files;
badf1662	117	struct fdtable *fdt;
1da177e4 LT	118	int fd;
	119
	120	spin_lock(&files->file_lock);
	121	fd = locate_fd(files, file, start);
	122	if (fd >= 0) {
badf1662 DS	123	/* locate_fd() may have expanded fdtable, load the ptr */
	124	fdt = files_fdtable(files);
	125	FD_SET(fd, fdt->open_fds);
	126	FD_CLR(fd, fdt->close_on_exec);
1da177e4 LT	127	spin_unlock(&files->file_lock);
	128	fd_install(fd, file);
	129	} else {
	130	spin_unlock(&files->file_lock);
	131	fput(file);
	132	}
	133
	134	return fd;
	135	}
	136
	137	asmlinkage long sys_dup2(unsigned int oldfd, unsigned int newfd)
	138	{
	139	int err = -EBADF;
	140	struct file * file, *tofree;
	141	struct files_struct * files = current->files;
badf1662	142	struct fdtable *fdt;
1da177e4 LT	143
	144	spin_lock(&files->file_lock);
	145	if (!(file = fcheck(oldfd)))
	146	goto out_unlock;
	147	err = newfd;
	148	if (newfd == oldfd)
	149	goto out_unlock;
	150	err = -EBADF;
	151	if (newfd >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur)
	152	goto out_unlock;
	153	get_file(file); /* We are now finished with oldfd */
	154
	155	err = expand_files(files, newfd);
	156	if (err < 0)
	157	goto out_fput;
	158
	159	/* To avoid races with open() and dup(), we will mark the fd as
	160	* in-use in the open-file bitmap throughout the entire dup2()
	161	* process. This is quite safe: do_close() uses the fd array
	162	* entry, not the bitmap, to decide what work needs to be
	163	* done. --sct */
	164	/* Doesn't work. open() might be there first. --AV */
	165
	166	/* Yes. It's a race. In user space. Nothing sane to do */
	167	err = -EBUSY;
badf1662 DS	168	fdt = files_fdtable(files);
	169	tofree = fdt->fd[newfd];
	170	if (!tofree && FD_ISSET(newfd, fdt->open_fds))
1da177e4 LT	171	goto out_fput;
1da177e4 LT	172
ab2af1f5	173	rcu_assign_pointer(fdt->fd[newfd], file);
badf1662 DS	174	FD_SET(newfd, fdt->open_fds);
badf1662 DS	175	FD_CLR(newfd, fdt->close_on_exec);
1da177e4 LT	176	spin_unlock(&files->file_lock);
	177
	178	if (tofree)
	179	filp_close(tofree, files);
	180	err = newfd;
	181	out:
	182	return err;
	183	out_unlock:
	184	spin_unlock(&files->file_lock);
	185	goto out;
	186
	187	out_fput:
	188	spin_unlock(&files->file_lock);
	189	fput(file);
	190	goto out;
	191	}
	192
	193	asmlinkage long sys_dup(unsigned int fildes)
	194	{
	195	int ret = -EBADF;
	196	struct file * file = fget(fildes);
	197
	198	if (file)
	199	ret = dupfd(file, 0);
	200	return ret;
	201	}
	202
	203	#define SETFL_MASK (O_APPEND \| O_NONBLOCK \| O_NDELAY \| FASYNC \| O_DIRECT \| O_NOATIME)
	204
	205	static int setfl(int fd, struct file * filp, unsigned long arg)
	206	{
	207	struct inode * inode = filp->f_dentry->d_inode;
	208	int error = 0;
	209
7d95c8f2	210	/*
	211	* O_APPEND cannot be cleared if the file is marked as append-only
	212	* and the file is open for write.
	213	*/
	214	if (((arg ^ filp->f_flags) & O_APPEND) && IS_APPEND(inode))
1da177e4 LT	215	return -EPERM;
	216
	217	/* O_NOATIME can only be set by the owner or superuser */
	218	if ((arg & O_NOATIME) && !(filp->f_flags & O_NOATIME))
	219	if (current->fsuid != inode->i_uid && !capable(CAP_FOWNER))
	220	return -EPERM;
	221
	222	/* required for strict SunOS emulation */
	223	if (O_NONBLOCK != O_NDELAY)
	224	if (arg & O_NDELAY)
	225	arg \|= O_NONBLOCK;
	226
	227	if (arg & O_DIRECT) {
	228	if (!filp->f_mapping \|\| !filp->f_mapping->a_ops \|\|
	229	!filp->f_mapping->a_ops->direct_IO)
	230	return -EINVAL;
	231	}
	232
	233	if (filp->f_op && filp->f_op->check_flags)
	234	error = filp->f_op->check_flags(arg);
	235	if (error)
	236	return error;
	237
	238	lock_kernel();
	239	if ((arg ^ filp->f_flags) & FASYNC) {
	240	if (filp->f_op && filp->f_op->fasync) {
	241	error = filp->f_op->fasync(fd, filp, (arg & FASYNC) != 0);
	242	if (error < 0)
	243	goto out;
	244	}
	245	}
	246
	247	filp->f_flags = (arg & SETFL_MASK) \| (filp->f_flags & ~SETFL_MASK);
	248	out:
	249	unlock_kernel();
	250	return error;
	251	}
	252
	253	static void f_modown(struct file *filp, unsigned long pid,
	254	uid_t uid, uid_t euid, int force)
	255	{
	256	write_lock_irq(&filp->f_owner.lock);
	257	if (force \|\| !filp->f_owner.pid) {
	258	filp->f_owner.pid = pid;
	259	filp->f_owner.uid = uid;
	260	filp->f_owner.euid = euid;
	261	}
	262	write_unlock_irq(&filp->f_owner.lock);
	263	}
	264
	265	int f_setown(struct file *filp, unsigned long arg, int force)
	266	{
	267	int err;
	268
	269	err = security_file_set_fowner(filp);
	270	if (err)
	271	return err;
	272
	273	f_modown(filp, arg, current->uid, current->euid, force);
	274	return 0;
	275	}
	276
	277	EXPORT_SYMBOL(f_setown);
	278
279	void f_delown(struct file *filp)
280	{
281	f_modown(filp, 0, 0, 0, 1);
282	}
283
284	static long do_fcntl(int fd, unsigned int cmd, unsigned long arg,
285	struct file *filp)
286	{
287	long err = -EINVAL;
288
289	switch (cmd) {
290	case F_DUPFD:
291	get_file(filp);
292	err = dupfd(filp, arg);
293	break;
294	case F_GETFD:
295	err = get_close_on_exec(fd) ? FD_CLOEXEC : 0;
296	break;
297	case F_SETFD:
298	err = 0;
299	set_close_on_exec(fd, arg & FD_CLOEXEC);
300	break;
301	case F_GETFL:
302	err = filp->f_flags;
303	break;
304	case F_SETFL:
305	err = setfl(fd, filp, arg);
306	break;
307	case F_GETLK:
308	err = fcntl_getlk(filp, (struct flock __user *) arg);
309	break;
310	case F_SETLK:
311	case F_SETLKW:
c293621b	312	err = fcntl_setlk(fd, filp, cmd, (struct flock __user *) arg);
1da177e4 LT	313	break;
	314	case F_GETOWN:
	315	/*
	316	* XXX If f_owner is a process group, the
	317	* negative return value will get converted
	318	* into an error. Oops. If we keep the
	319	* current syscall conventions, the only way
	320	* to fix this will be in libc.
	321	*/
	322	err = filp->f_owner.pid;
	323	force_successful_syscall_return();
	324	break;
	325	case F_SETOWN:
	326	err = f_setown(filp, arg, 1);
	327	break;
	328	case F_GETSIG:
	329	err = filp->f_owner.signum;
	330	break;
	331	case F_SETSIG:
	332	/* arg == 0 restores default behaviour. */
7ed20e1a	333	if (!valid_signal(arg)) {
1da177e4 LT	334	break;
	335	}
	336	err = 0;
	337	filp->f_owner.signum = arg;
	338	break;
	339	case F_GETLEASE:
	340	err = fcntl_getlease(filp);
	341	break;
	342	case F_SETLEASE:
	343	err = fcntl_setlease(fd, filp, arg);
	344	break;
	345	case F_NOTIFY:
	346	err = fcntl_dirnotify(fd, filp, arg);
	347	break;
	348	default:
	349	break;
	350	}
	351	return err;
	352	}
	353
	354	asmlinkage long sys_fcntl(unsigned int fd, unsigned int cmd, unsigned long arg)
	355	{
	356	struct file *filp;
	357	long err = -EBADF;
	358
	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
	369	err = do_fcntl(fd, cmd, arg, filp);
	370
	371	fput(filp);
	372	out:
	373	return err;
	374	}
	375
	376	#if BITS_PER_LONG == 32
	377	asmlinkage long sys_fcntl64(unsigned int fd, unsigned int cmd, unsigned long arg)
	378	{
	379	struct file * filp;
	380	long err;
	381
	382	err = -EBADF;
	383	filp = fget(fd);
	384	if (!filp)
	385	goto out;
	386
	387	err = security_file_fcntl(filp, cmd, arg);
	388	if (err) {
	389	fput(filp);
	390	return err;
	391	}
	392	err = -EBADF;
	393
	394	switch (cmd) {
	395	case F_GETLK64:
	396	err = fcntl_getlk64(filp, (struct flock64 __user *) arg);
	397	break;
398	case F_SETLK64:
399	case F_SETLKW64:
c293621b PS	400	err = fcntl_setlk64(fd, filp, cmd,
c293621b PS	401	(struct flock64 __user *) arg);
1da177e4 LT	402	break;
	403	default:
	404	err = do_fcntl(fd, cmd, arg, filp);
	405	break;
	406	}
	407	fput(filp);
	408	out:
	409	return err;
	410	}
	411	#endif
	412
	413	/* Table to convert sigio signal codes into poll band bitmaps */
	414
fa3536cc	415	static const long band_table[NSIGPOLL] = {
1da177e4 LT	416	POLLIN \| POLLRDNORM, /* POLL_IN */
	417	POLLOUT \| POLLWRNORM \| POLLWRBAND, /* POLL_OUT */
	418	POLLIN \| POLLRDNORM \| POLLMSG, /* POLL_MSG */
	419	POLLERR, /* POLL_ERR */
	420	POLLPRI \| POLLRDBAND, /* POLL_PRI */
	421	POLLHUP \| POLLERR /* POLL_HUP */
	422	};
	423
	424	static inline int sigio_perm(struct task_struct *p,
	425	struct fown_struct *fown, int sig)
	426	{
	427	return (((fown->euid == 0) \|\|
	428	(fown->euid == p->suid) \|\| (fown->euid == p->uid) \|\|
	429	(fown->uid == p->suid) \|\| (fown->uid == p->uid)) &&
	430	!security_file_send_sigiotask(p, fown, sig));
	431	}
	432
	433	static void send_sigio_to_task(struct task_struct *p,
	434	struct fown_struct *fown,
	435	int fd,
	436	int reason)
	437	{
	438	if (!sigio_perm(p, fown, fown->signum))
	439	return;
	440
	441	switch (fown->signum) {
	442	siginfo_t si;
	443	default:
	444	/* Queue a rt signal with the appropriate fd as its
	445	value. We use SI_SIGIO as the source, not
	446	SI_KERNEL, since kernel signals always get
	447	delivered even if we can't queue. Failure to
	448	queue in this case _should_ be reported; we fall
	449	back to SIGIO in that case. --sct */
	450	si.si_signo = fown->signum;
	451	si.si_errno = 0;
	452	si.si_code = reason;
	453	/* Make sure we are called with one of the POLL_*
	454	reasons, otherwise we could leak kernel stack into
	455	userspace. */
	456	if ((reason & __SI_MASK) != __SI_POLL)
	457	BUG();
	458	if (reason - POLL_IN >= NSIGPOLL)
	459	si.si_band = ~0L;
	460	else
	461	si.si_band = band_table[reason - POLL_IN];
	462	si.si_fd = fd;
850d6fbe	463	if (!group_send_sig_info(fown->signum, &si, p))
1da177e4 LT	464	break;
	465	/* fall-through: fall back on the old plain SIGIO signal */
	466	case 0:
850d6fbe	467	group_send_sig_info(SIGIO, SEND_SIG_PRIV, p);
1da177e4 LT	468	}
	469	}
	470
	471	void send_sigio(struct fown_struct *fown, int fd, int band)
	472	{
	473	struct task_struct *p;
	474	int pid;
	475
	476	read_lock(&fown->lock);
	477	pid = fown->pid;
	478	if (!pid)
	479	goto out_unlock_fown;
	480
	481	read_lock(&tasklist_lock);
	482	if (pid > 0) {
	483	p = find_task_by_pid(pid);
	484	if (p) {
	485	send_sigio_to_task(p, fown, fd, band);
	486	}
	487	} else {
	488	do_each_task_pid(-pid, PIDTYPE_PGID, p) {
	489	send_sigio_to_task(p, fown, fd, band);
	490	} while_each_task_pid(-pid, PIDTYPE_PGID, p);
	491	}
	492	read_unlock(&tasklist_lock);
	493	out_unlock_fown:
	494	read_unlock(&fown->lock);
	495	}
	496
	497	static void send_sigurg_to_task(struct task_struct *p,
	498	struct fown_struct *fown)
	499	{
	500	if (sigio_perm(p, fown, SIGURG))
850d6fbe	501	group_send_sig_info(SIGURG, SEND_SIG_PRIV, p);
1da177e4 LT	502	}
	503
	504	int send_sigurg(struct fown_struct *fown)
	505	{
	506	struct task_struct *p;
	507	int pid, ret = 0;
	508
	509	read_lock(&fown->lock);
	510	pid = fown->pid;
	511	if (!pid)
	512	goto out_unlock_fown;
	513
	514	ret = 1;
	515
	516	read_lock(&tasklist_lock);
	517	if (pid > 0) {
	518	p = find_task_by_pid(pid);
	519	if (p) {
	520	send_sigurg_to_task(p, fown);
	521	}
	522	} else {
	523	do_each_task_pid(-pid, PIDTYPE_PGID, p) {
	524	send_sigurg_to_task(p, fown);
	525	} while_each_task_pid(-pid, PIDTYPE_PGID, p);
	526	}
	527	read_unlock(&tasklist_lock);
	528	out_unlock_fown:
	529	read_unlock(&fown->lock);
	530	return ret;
	531	}
	532
	533	static DEFINE_RWLOCK(fasync_lock);
fa3536cc	534	static kmem_cache_t *fasync_cache __read_mostly;
1da177e4 LT	535
	536	/*
	537	* fasync_helper() is used by some character device drivers (mainly mice)
	538	* to set up the fasync queue. It returns negative on error, 0 if it did
	539	* no changes and positive if it added/deleted the entry.
	540	*/
	541	int fasync_helper(int fd, struct file * filp, int on, struct fasync_struct **fapp)
	542	{
	543	struct fasync_struct fa, *fp;
	544	struct fasync_struct *new = NULL;
	545	int result = 0;
	546
	547	if (on) {
	548	new = kmem_cache_alloc(fasync_cache, SLAB_KERNEL);
	549	if (!new)
	550	return -ENOMEM;
	551	}
	552	write_lock_irq(&fasync_lock);
	553	for (fp = fapp; (fa = *fp) != NULL; fp = &fa->fa_next) {
	554	if (fa->fa_file == filp) {
	555	if(on) {
	556	fa->fa_fd = fd;
	557	kmem_cache_free(fasync_cache, new);
	558	} else {
	559	*fp = fa->fa_next;
	560	kmem_cache_free(fasync_cache, fa);
	561	result = 1;
	562	}
	563	goto out;
	564	}
	565	}
	566
	567	if (on) {
	568	new->magic = FASYNC_MAGIC;
	569	new->fa_file = filp;
	570	new->fa_fd = fd;
	571	new->fa_next = *fapp;
	572	*fapp = new;
	573	result = 1;
	574	}
	575	out:
	576	write_unlock_irq(&fasync_lock);
	577	return result;
	578	}
	579
	580	EXPORT_SYMBOL(fasync_helper);
	581
	582	void __kill_fasync(struct fasync_struct *fa, int sig, int band)
	583	{
	584	while (fa) {
	585	struct fown_struct * fown;
	586	if (fa->magic != FASYNC_MAGIC) {
	587	printk(KERN_ERR "kill_fasync: bad magic number in "
	588	"fasync_struct!\n");
	589	return;
	590	}
	591	fown = &fa->fa_file->f_owner;
	592	/* Don't send SIGURG to processes which have not set a
	593	queued signum: SIGURG has its own default signalling
	594	mechanism. */
	595	if (!(sig == SIGURG && fown->signum == 0))
	596	send_sigio(fown, fa->fa_fd, band);
	597	fa = fa->fa_next;
	598	}
599	}
600
601	EXPORT_SYMBOL(__kill_fasync);
602
603	void kill_fasync(struct fasync_struct **fp, int sig, int band)
604	{
605	/* First a quick test without locking: usually
606	* the list is empty.
607	*/
608	if (*fp) {
609	read_lock(&fasync_lock);
610	/* reread fp after obtaining the lock /
611	__kill_fasync(*fp, sig, band);
612	read_unlock(&fasync_lock);
613	}
614	}
615	EXPORT_SYMBOL(kill_fasync);
616
617	static int __init fasync_init(void)
618	{
619	fasync_cache = kmem_cache_create("fasync_cache",
620	sizeof(struct fasync_struct), 0, SLAB_PANIC, NULL, NULL);
621	return 0;
622	}
623
624	module_init(fasync_init)