[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;
}

SYSCALL_DEFINE3(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;
}

SYSCALL_DEFINE2(dup2, unsigned int, oldfd, unsigned int, newfd)
{
	if (unlikely(newfd == oldfd)) { /* corner case */
		struct files_struct *files = current->files;
		int retval = oldfd;

		rcu_read_lock();
		if (!fcheck_files(files, oldfd))
			retval = -EBADF;
		rcu_read_unlock();
		return retval;
	}
	return sys_dup3(oldfd, newfd, 0);
}

SYSCALL_DEFINE1(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 | 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;

	/*
	 * ->fasync() is responsible for setting the FASYNC bit.
	 */
	if (((arg ^ filp->f_flags) & FASYNC) && filp->f_op &&
			filp->f_op->fasync) {
		error = filp->f_op->fasync(fd, filp, (arg & FASYNC) != 0);
		if (error < 0)
			goto out;
		if (error > 0)
			error = 0;
	}
	spin_lock(&filp->f_lock);
	filp->f_flags = (arg & SETFL_MASK) | (filp->f_flags & ~SETFL_MASK);
	spin_unlock(&filp->f_lock);

 out:
	return error;
}

static void f_modown(struct file *filp, struct pid *pid, enum pid_type type,
                     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;

		if (pid) {
			const struct cred *cred = current_cred();
			filp->f_owner.uid = cred->uid;
			filp->f_owner.euid = cred->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, 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, 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 int f_setown_ex(struct file *filp, unsigned long arg)
{
	struct f_owner_ex * __user owner_p = (void * __user)arg;
	struct f_owner_ex owner;
	struct pid *pid;
	int type;
	int ret;

	ret = copy_from_user(&owner, owner_p, sizeof(owner));
	if (ret)
		return ret;

	switch (owner.type) {
	case F_OWNER_TID:
		type = PIDTYPE_MAX;
		break;

	case F_OWNER_PID:
		type = PIDTYPE_PID;
		break;

	case F_OWNER_PGRP:
		type = PIDTYPE_PGID;
		break;

	default:
		return -EINVAL;
	}

	rcu_read_lock();
	pid = find_vpid(owner.pid);
	if (owner.pid && !pid)
		ret = -ESRCH;
	else
		ret = __f_setown(filp, pid, type, 1);
	rcu_read_unlock();

	return ret;
}

static int f_getown_ex(struct file *filp, unsigned long arg)
{
	struct f_owner_ex * __user owner_p = (void * __user)arg;
	struct f_owner_ex owner;
	int ret = 0;

	read_lock(&filp->f_owner.lock);
	owner.pid = pid_vnr(filp->f_owner.pid);
	switch (filp->f_owner.pid_type) {
	case PIDTYPE_MAX:
		owner.type = F_OWNER_TID;
		break;

	case PIDTYPE_PID:
		owner.type = F_OWNER_PID;
		break;

	case PIDTYPE_PGID:
		owner.type = F_OWNER_PGRP;
		break;

	default:
		WARN_ON(1);
		ret = -EINVAL;
		break;
	}
	read_unlock(&filp->f_owner.lock);

	if (!ret)
		ret = copy_to_user(owner_p, &owner, sizeof(owner));
	return ret;
}

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_GETOWN_EX:
		err = f_getown_ex(filp, arg);
		break;
	case F_SETOWN_EX:
		err = f_setown_ex(filp, arg);
		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;
}

SYSCALL_DEFINE3(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
SYSCALL_DEFINE3(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)
{
	const struct cred *cred;
	int ret;

	rcu_read_lock();
	cred = __task_cred(p);
	ret = ((fown->euid == 0 ||
		fown->euid == cred->suid || fown->euid == cred->uid ||
		fown->uid  == cred->suid || fown->uid  == cred->uid) &&
	       !security_file_send_sigiotask(p, fown, sig));
	rcu_read_unlock();
	return ret;
}

static void send_sigio_to_task(struct task_struct *p,
			       struct fown_struct *fown,
			       int fd, int reason, int group)
{
	/*
	 * F_SETSIG can change ->signum lockless in parallel, make
	 * sure we read it once and use the same value throughout.
	 */
	int signum = ACCESS_ONCE(fown->signum);

	if (!sigio_perm(p, fown, signum))
		return;

	switch (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 = 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 (!do_send_sig_info(signum, &si, p, group))
				break;
		/* fall-through: fall back on the old plain SIGIO signal */
		case 0:
			do_send_sig_info(SIGIO, SEND_SIG_PRIV, p, group);
	}
}

void send_sigio(struct fown_struct *fown, int fd, int band)
{
	struct task_struct *p;
	enum pid_type type;
	struct pid *pid;
	int group = 1;
	
	read_lock(&fown->lock);

	type = fown->pid_type;
	if (type == PIDTYPE_MAX) {
		group = 0;
		type = PIDTYPE_PID;
	}

	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, group);
	} 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, int group)
{
	if (sigio_perm(p, fown, SIGURG))
		do_send_sig_info(SIGURG, SEND_SIG_PRIV, p, group);
}

int send_sigurg(struct fown_struct *fown)
{
	struct task_struct *p;
	enum pid_type type;
	struct pid *pid;
	int group = 1;
	int ret = 0;
	
	read_lock(&fown->lock);

	type = fown->pid_type;
	if (type == PIDTYPE_MAX) {
		group = 0;
		type = PIDTYPE_PID;
	}

	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, group);
	} 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 almost all character device drivers
 * 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;
	}

	/*
	 * We need to take f_lock first since it's not an IRQ-safe
	 * lock.
	 */
	spin_lock(&filp->f_lock);
	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:
	if (on)
		filp->f_flags |= FASYNC;
	else
		filp->f_flags &= ~FASYNC;
	write_unlock_irq(&fasync_lock);
	spin_unlock(&filp->f_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
a26eab24	52	SYSCALL_DEFINE3(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	114
a26eab24	115	SYSCALL_DEFINE2(dup2, unsigned int, oldfd, unsigned int, newfd)
336dd1f7	116	{
6c5d0512 AV	117	if (unlikely(newfd == oldfd)) { /* corner case */
6c5d0512 AV	118	struct files_struct *files = current->files;
2b79bc4f JM	119	int retval = oldfd;
2b79bc4f JM	120
6c5d0512 AV	121	rcu_read_lock();
6c5d0512 AV	122	if (!fcheck_files(files, oldfd))
2b79bc4f	123	retval = -EBADF;
6c5d0512	124	rcu_read_unlock();
2b79bc4f	125	return retval;
6c5d0512	126	}
336dd1f7 UD	127	return sys_dup3(oldfd, newfd, 0);
336dd1f7 UD	128	}
1da177e4	129
a26eab24	130	SYSCALL_DEFINE1(dup, unsigned int, fildes)
1da177e4 LT	131	{
1da177e4 LT	132	int ret = -EBADF;
1027abe8 AV	133	struct file *file = fget(fildes);
	134
	135	if (file) {
	136	ret = get_unused_fd();
	137	if (ret >= 0)
	138	fd_install(ret, file);
	139	else
	140	fput(file);
	141	}
1da177e4 LT	142	return ret;
	143	}
	144
76398425	145	#define SETFL_MASK (O_APPEND \| O_NONBLOCK \| O_NDELAY \| O_DIRECT \| O_NOATIME)
1da177e4 LT	146
	147	static int setfl(int fd, struct file * filp, unsigned long arg)
	148	{
0f7fc9e4	149	struct inode * inode = filp->f_path.dentry->d_inode;
1da177e4 LT	150	int error = 0;
1da177e4 LT	151
7d95c8f2	152	/*
	153	* O_APPEND cannot be cleared if the file is marked as append-only
	154	* and the file is open for write.
	155	*/
	156	if (((arg ^ filp->f_flags) & O_APPEND) && IS_APPEND(inode))
1da177e4 LT	157	return -EPERM;
	158
	159	/* O_NOATIME can only be set by the owner or superuser */
	160	if ((arg & O_NOATIME) && !(filp->f_flags & O_NOATIME))
3bd858ab	161	if (!is_owner_or_cap(inode))
1da177e4 LT	162	return -EPERM;
	163
	164	/* required for strict SunOS emulation */
	165	if (O_NONBLOCK != O_NDELAY)
	166	if (arg & O_NDELAY)
	167	arg \|= O_NONBLOCK;
	168
	169	if (arg & O_DIRECT) {
	170	if (!filp->f_mapping \|\| !filp->f_mapping->a_ops \|\|
	171	!filp->f_mapping->a_ops->direct_IO)
	172	return -EINVAL;
	173	}
	174
	175	if (filp->f_op && filp->f_op->check_flags)
	176	error = filp->f_op->check_flags(arg);
	177	if (error)
	178	return error;
	179
218d11a8	180	/*
76398425	181	* ->fasync() is responsible for setting the FASYNC bit.
218d11a8	182	*/
76398425 JC	183	if (((arg ^ filp->f_flags) & FASYNC) && filp->f_op &&
	184	filp->f_op->fasync) {
	185	error = filp->f_op->fasync(fd, filp, (arg & FASYNC) != 0);
	186	if (error < 0)
	187	goto out;
60aa4924 JC	188	if (error > 0)
60aa4924 JC	189	error = 0;
1da177e4	190	}
db1dd4d3	191	spin_lock(&filp->f_lock);
1da177e4	192	filp->f_flags = (arg & SETFL_MASK) \| (filp->f_flags & ~SETFL_MASK);
db1dd4d3	193	spin_unlock(&filp->f_lock);
76398425	194
1da177e4	195	out:
1da177e4 LT	196	return error;
	197	}
	198
609d7fa9	199	static void f_modown(struct file filp, struct pid pid, enum pid_type type,
2f38d70f	200	int force)
1da177e4 LT	201	{
	202	write_lock_irq(&filp->f_owner.lock);
	203	if (force \|\| !filp->f_owner.pid) {
609d7fa9 EB	204	put_pid(filp->f_owner.pid);
	205	filp->f_owner.pid = get_pid(pid);
	206	filp->f_owner.pid_type = type;
2f38d70f ON	207
	208	if (pid) {
	209	const struct cred *cred = current_cred();
	210	filp->f_owner.uid = cred->uid;
	211	filp->f_owner.euid = cred->euid;
	212	}
1da177e4 LT	213	}
	214	write_unlock_irq(&filp->f_owner.lock);
	215	}
	216
609d7fa9 EB	217	int __f_setown(struct file filp, struct pid pid, enum pid_type type,
609d7fa9 EB	218	int force)
1da177e4 LT	219	{
1da177e4 LT	220	int err;
2f38d70f	221
1da177e4 LT	222	err = security_file_set_fowner(filp);
	223	if (err)
	224	return err;
	225
2f38d70f	226	f_modown(filp, pid, type, force);
1da177e4 LT	227	return 0;
1da177e4 LT	228	}
609d7fa9	229	EXPORT_SYMBOL(__f_setown);
1da177e4	230
609d7fa9 EB	231	int f_setown(struct file *filp, unsigned long arg, int force)
	232	{
	233	enum pid_type type;
	234	struct pid *pid;
	235	int who = arg;
	236	int result;
	237	type = PIDTYPE_PID;
	238	if (who < 0) {
	239	type = PIDTYPE_PGID;
	240	who = -who;
	241	}
	242	rcu_read_lock();
b488893a	243	pid = find_vpid(who);
609d7fa9 EB	244	result = __f_setown(filp, pid, type, force);
	245	rcu_read_unlock();
	246	return result;
	247	}
1da177e4 LT	248	EXPORT_SYMBOL(f_setown);
	249
	250	void f_delown(struct file *filp)
	251	{
2f38d70f	252	f_modown(filp, NULL, PIDTYPE_PID, 1);
609d7fa9 EB	253	}
	254
	255	pid_t f_getown(struct file *filp)
	256	{
	257	pid_t pid;
43fa1adb	258	read_lock(&filp->f_owner.lock);
6c5f3e7b	259	pid = pid_vnr(filp->f_owner.pid);
609d7fa9 EB	260	if (filp->f_owner.pid_type == PIDTYPE_PGID)
609d7fa9 EB	261	pid = -pid;
43fa1adb	262	read_unlock(&filp->f_owner.lock);
609d7fa9	263	return pid;
1da177e4 LT	264	}
1da177e4 LT	265
ba0a6c9f PZ	266	static int f_setown_ex(struct file *filp, unsigned long arg)
	267	{
	268	struct f_owner_ex * __user owner_p = (void * __user)arg;
	269	struct f_owner_ex owner;
	270	struct pid *pid;
	271	int type;
	272	int ret;
	273
	274	ret = copy_from_user(&owner, owner_p, sizeof(owner));
	275	if (ret)
	276	return ret;
	277
	278	switch (owner.type) {
	279	case F_OWNER_TID:
	280	type = PIDTYPE_MAX;
	281	break;
	282
	283	case F_OWNER_PID:
	284	type = PIDTYPE_PID;
	285	break;
	286
978b4053	287	case F_OWNER_PGRP:
ba0a6c9f PZ	288	type = PIDTYPE_PGID;
	289	break;
	290
	291	default:
	292	return -EINVAL;
	293	}
	294
	295	rcu_read_lock();
	296	pid = find_vpid(owner.pid);
	297	if (owner.pid && !pid)
	298	ret = -ESRCH;
	299	else
	300	ret = __f_setown(filp, pid, type, 1);
	301	rcu_read_unlock();
	302
	303	return ret;
	304	}
	305
	306	static int f_getown_ex(struct file *filp, unsigned long arg)
	307	{
	308	struct f_owner_ex * __user owner_p = (void * __user)arg;
	309	struct f_owner_ex owner;
	310	int ret = 0;
	311
	312	read_lock(&filp->f_owner.lock);
	313	owner.pid = pid_vnr(filp->f_owner.pid);
	314	switch (filp->f_owner.pid_type) {
	315	case PIDTYPE_MAX:
	316	owner.type = F_OWNER_TID;
	317	break;
	318
	319	case PIDTYPE_PID:
	320	owner.type = F_OWNER_PID;
	321	break;
	322
	323	case PIDTYPE_PGID:
978b4053	324	owner.type = F_OWNER_PGRP;
ba0a6c9f PZ	325	break;
	326
	327	default:
	328	WARN_ON(1);
	329	ret = -EINVAL;
	330	break;
	331	}
	332	read_unlock(&filp->f_owner.lock);
	333
	334	if (!ret)
	335	ret = copy_to_user(owner_p, &owner, sizeof(owner));
	336	return ret;
	337	}
	338
1da177e4 LT	339	static long do_fcntl(int fd, unsigned int cmd, unsigned long arg,
	340	struct file *filp)
	341	{
	342	long err = -EINVAL;
	343
	344	switch (cmd) {
	345	case F_DUPFD:
22d2b35b	346	case F_DUPFD_CLOEXEC:
4e1e018e AV	347	if (arg >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur)
4e1e018e AV	348	break;
1027abe8 AV	349	err = alloc_fd(arg, cmd == F_DUPFD_CLOEXEC ? O_CLOEXEC : 0);
	350	if (err >= 0) {
	351	get_file(filp);
	352	fd_install(err, filp);
	353	}
1da177e4 LT	354	break;
	355	case F_GETFD:
	356	err = get_close_on_exec(fd) ? FD_CLOEXEC : 0;
	357	break;
	358	case F_SETFD:
	359	err = 0;
	360	set_close_on_exec(fd, arg & FD_CLOEXEC);
	361	break;
	362	case F_GETFL:
	363	err = filp->f_flags;
	364	break;
	365	case F_SETFL:
	366	err = setfl(fd, filp, arg);
	367	break;
	368	case F_GETLK:
	369	err = fcntl_getlk(filp, (struct flock __user *) arg);
	370	break;
	371	case F_SETLK:
	372	case F_SETLKW:
c293621b	373	err = fcntl_setlk(fd, filp, cmd, (struct flock __user *) arg);
1da177e4 LT	374	break;
	375	case F_GETOWN:
	376	/*
	377	* XXX If f_owner is a process group, the
	378	* negative return value will get converted
	379	* into an error. Oops. If we keep the
	380	* current syscall conventions, the only way
	381	* to fix this will be in libc.
	382	*/
609d7fa9	383	err = f_getown(filp);
1da177e4 LT	384	force_successful_syscall_return();
	385	break;
	386	case F_SETOWN:
	387	err = f_setown(filp, arg, 1);
	388	break;
ba0a6c9f PZ	389	case F_GETOWN_EX:
	390	err = f_getown_ex(filp, arg);
	391	break;
	392	case F_SETOWN_EX:
	393	err = f_setown_ex(filp, arg);
	394	break;
1da177e4 LT	395	case F_GETSIG:
	396	err = filp->f_owner.signum;
	397	break;
	398	case F_SETSIG:
	399	/* arg == 0 restores default behaviour. */
7ed20e1a	400	if (!valid_signal(arg)) {
1da177e4 LT	401	break;
	402	}
	403	err = 0;
	404	filp->f_owner.signum = arg;
	405	break;
	406	case F_GETLEASE:
	407	err = fcntl_getlease(filp);
	408	break;
	409	case F_SETLEASE:
	410	err = fcntl_setlease(fd, filp, arg);
	411	break;
	412	case F_NOTIFY:
	413	err = fcntl_dirnotify(fd, filp, arg);
	414	break;
	415	default:
	416	break;
	417	}
	418	return err;
	419	}
	420
a26eab24	421	SYSCALL_DEFINE3(fcntl, unsigned int, fd, unsigned int, cmd, unsigned long, arg)
1da177e4 LT	422	{
	423	struct file *filp;
	424	long err = -EBADF;
	425
	426	filp = fget(fd);
	427	if (!filp)
	428	goto out;
	429
	430	err = security_file_fcntl(filp, cmd, arg);
	431	if (err) {
	432	fput(filp);
	433	return err;
	434	}
	435
	436	err = do_fcntl(fd, cmd, arg, filp);
	437
	438	fput(filp);
	439	out:
	440	return err;
	441	}
	442
	443	#if BITS_PER_LONG == 32
a26eab24 HC	444	SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd,
a26eab24 HC	445	unsigned long, arg)
1da177e4 LT	446	{
	447	struct file * filp;
	448	long err;
	449
	450	err = -EBADF;
	451	filp = fget(fd);
	452	if (!filp)
	453	goto out;
	454
	455	err = security_file_fcntl(filp, cmd, arg);
	456	if (err) {
	457	fput(filp);
	458	return err;
	459	}
	460	err = -EBADF;
	461
	462	switch (cmd) {
	463	case F_GETLK64:
	464	err = fcntl_getlk64(filp, (struct flock64 __user *) arg);
	465	break;
	466	case F_SETLK64:
	467	case F_SETLKW64:
c293621b PS	468	err = fcntl_setlk64(fd, filp, cmd,
c293621b PS	469	(struct flock64 __user *) arg);
1da177e4 LT	470	break;
	471	default:
	472	err = do_fcntl(fd, cmd, arg, filp);
	473	break;
	474	}
	475	fput(filp);
	476	out:
	477	return err;
	478	}
	479	#endif
	480
	481	/* Table to convert sigio signal codes into poll band bitmaps */
	482
fa3536cc	483	static const long band_table[NSIGPOLL] = {
1da177e4 LT	484	POLLIN \| POLLRDNORM, /* POLL_IN */
	485	POLLOUT \| POLLWRNORM \| POLLWRBAND, /* POLL_OUT */
	486	POLLIN \| POLLRDNORM \| POLLMSG, /* POLL_MSG */
	487	POLLERR, /* POLL_ERR */
	488	POLLPRI \| POLLRDBAND, /* POLL_PRI */
	489	POLLHUP \| POLLERR /* POLL_HUP */
	490	};
	491
	492	static inline int sigio_perm(struct task_struct *p,
	493	struct fown_struct *fown, int sig)
	494	{
c69e8d9c DH	495	const struct cred *cred;
	496	int ret;
	497
	498	rcu_read_lock();
	499	cred = __task_cred(p);
	500	ret = ((fown->euid == 0 \|\|
	501	fown->euid == cred->suid \|\| fown->euid == cred->uid \|\|
	502	fown->uid == cred->suid \|\| fown->uid == cred->uid) &&
	503	!security_file_send_sigiotask(p, fown, sig));
	504	rcu_read_unlock();
	505	return ret;
1da177e4 LT	506	}
	507
	508	static void send_sigio_to_task(struct task_struct *p,
8eeee4e2	509	struct fown_struct *fown,
ba0a6c9f	510	int fd, int reason, int group)
1da177e4	511	{
8eeee4e2 ON	512	/*
	513	* F_SETSIG can change ->signum lockless in parallel, make
	514	* sure we read it once and use the same value throughout.
	515	*/
	516	int signum = ACCESS_ONCE(fown->signum);
	517
	518	if (!sigio_perm(p, fown, signum))
1da177e4 LT	519	return;
1da177e4 LT	520
8eeee4e2	521	switch (signum) {
1da177e4 LT	522	siginfo_t si;
	523	default:
	524	/* Queue a rt signal with the appropriate fd as its
	525	value. We use SI_SIGIO as the source, not
	526	SI_KERNEL, since kernel signals always get
	527	delivered even if we can't queue. Failure to
	528	queue in this case _should_ be reported; we fall
	529	back to SIGIO in that case. --sct */
8eeee4e2	530	si.si_signo = signum;
1da177e4 LT	531	si.si_errno = 0;
	532	si.si_code = reason;
	533	/* Make sure we are called with one of the POLL_*
	534	reasons, otherwise we could leak kernel stack into
	535	userspace. */
f6298aab	536	BUG_ON((reason & __SI_MASK) != __SI_POLL);
1da177e4 LT	537	if (reason - POLL_IN >= NSIGPOLL)
	538	si.si_band = ~0L;
	539	else
	540	si.si_band = band_table[reason - POLL_IN];
	541	si.si_fd = fd;
ba0a6c9f	542	if (!do_send_sig_info(signum, &si, p, group))
1da177e4 LT	543	break;
	544	/* fall-through: fall back on the old plain SIGIO signal */
	545	case 0:
ba0a6c9f	546	do_send_sig_info(SIGIO, SEND_SIG_PRIV, p, group);
1da177e4 LT	547	}
	548	}
	549
	550	void send_sigio(struct fown_struct *fown, int fd, int band)
	551	{
	552	struct task_struct *p;
609d7fa9 EB	553	enum pid_type type;
609d7fa9 EB	554	struct pid *pid;
ba0a6c9f	555	int group = 1;
1da177e4 LT	556
1da177e4 LT	557	read_lock(&fown->lock);
ba0a6c9f	558
609d7fa9	559	type = fown->pid_type;
ba0a6c9f PZ	560	if (type == PIDTYPE_MAX) {
	561	group = 0;
	562	type = PIDTYPE_PID;
	563	}
	564
1da177e4 LT	565	pid = fown->pid;
	566	if (!pid)
	567	goto out_unlock_fown;
	568
	569	read_lock(&tasklist_lock);
609d7fa9	570	do_each_pid_task(pid, type, p) {
ba0a6c9f	571	send_sigio_to_task(p, fown, fd, band, group);
609d7fa9	572	} while_each_pid_task(pid, type, p);
1da177e4 LT	573	read_unlock(&tasklist_lock);
	574	out_unlock_fown:
	575	read_unlock(&fown->lock);
	576	}
	577
	578	static void send_sigurg_to_task(struct task_struct *p,
ba0a6c9f	579	struct fown_struct *fown, int group)
1da177e4 LT	580	{
1da177e4 LT	581	if (sigio_perm(p, fown, SIGURG))
ba0a6c9f	582	do_send_sig_info(SIGURG, SEND_SIG_PRIV, p, group);
1da177e4 LT	583	}
	584
	585	int send_sigurg(struct fown_struct *fown)
	586	{
	587	struct task_struct *p;
609d7fa9 EB	588	enum pid_type type;
609d7fa9 EB	589	struct pid *pid;
ba0a6c9f	590	int group = 1;
609d7fa9	591	int ret = 0;
1da177e4 LT	592
1da177e4 LT	593	read_lock(&fown->lock);
ba0a6c9f	594
609d7fa9	595	type = fown->pid_type;
ba0a6c9f PZ	596	if (type == PIDTYPE_MAX) {
	597	group = 0;
	598	type = PIDTYPE_PID;
	599	}
	600
1da177e4 LT	601	pid = fown->pid;
	602	if (!pid)
	603	goto out_unlock_fown;
	604
	605	ret = 1;
	606
	607	read_lock(&tasklist_lock);
609d7fa9	608	do_each_pid_task(pid, type, p) {
ba0a6c9f	609	send_sigurg_to_task(p, fown, group);
609d7fa9	610	} while_each_pid_task(pid, type, p);
1da177e4 LT	611	read_unlock(&tasklist_lock);
	612	out_unlock_fown:
	613	read_unlock(&fown->lock);
	614	return ret;
	615	}
	616
	617	static DEFINE_RWLOCK(fasync_lock);
e18b890b	618	static struct kmem_cache *fasync_cache __read_mostly;
1da177e4 LT	619
1da177e4 LT	620	/*
76398425	621	* fasync_helper() is used by almost all character device drivers
1da177e4 LT	622	* to set up the fasync queue. It returns negative on error, 0 if it did
	623	* no changes and positive if it added/deleted the entry.
	624	*/
	625	int fasync_helper(int fd, struct file * filp, int on, struct fasync_struct **fapp)
	626	{
	627	struct fasync_struct fa, *fp;
	628	struct fasync_struct *new = NULL;
	629	int result = 0;
	630
	631	if (on) {
e94b1766	632	new = kmem_cache_alloc(fasync_cache, GFP_KERNEL);
1da177e4 LT	633	if (!new)
	634	return -ENOMEM;
	635	}
4a6a4499 JC	636
	637	/*
	638	* We need to take f_lock first since it's not an IRQ-safe
	639	* lock.
	640	*/
	641	spin_lock(&filp->f_lock);
1da177e4 LT	642	write_lock_irq(&fasync_lock);
	643	for (fp = fapp; (fa = *fp) != NULL; fp = &fa->fa_next) {
	644	if (fa->fa_file == filp) {
	645	if(on) {
	646	fa->fa_fd = fd;
	647	kmem_cache_free(fasync_cache, new);
	648	} else {
	649	*fp = fa->fa_next;
	650	kmem_cache_free(fasync_cache, fa);
	651	result = 1;
	652	}
	653	goto out;
	654	}
	655	}
	656
	657	if (on) {
	658	new->magic = FASYNC_MAGIC;
	659	new->fa_file = filp;
	660	new->fa_fd = fd;
	661	new->fa_next = *fapp;
	662	*fapp = new;
	663	result = 1;
	664	}
	665	out:
76398425 JC	666	if (on)
	667	filp->f_flags \|= FASYNC;
	668	else
	669	filp->f_flags &= ~FASYNC;
1da177e4	670	write_unlock_irq(&fasync_lock);
4a6a4499	671	spin_unlock(&filp->f_lock);
1da177e4 LT	672	return result;
	673	}
	674
	675	EXPORT_SYMBOL(fasync_helper);
	676
	677	void __kill_fasync(struct fasync_struct *fa, int sig, int band)
	678	{
	679	while (fa) {
	680	struct fown_struct * fown;
	681	if (fa->magic != FASYNC_MAGIC) {
	682	printk(KERN_ERR "kill_fasync: bad magic number in "
	683	"fasync_struct!\n");
	684	return;
	685	}
	686	fown = &fa->fa_file->f_owner;
	687	/* Don't send SIGURG to processes which have not set a
	688	queued signum: SIGURG has its own default signalling
	689	mechanism. */
	690	if (!(sig == SIGURG && fown->signum == 0))
	691	send_sigio(fown, fa->fa_fd, band);
	692	fa = fa->fa_next;
	693	}
	694	}
	695
	696	EXPORT_SYMBOL(__kill_fasync);
	697
	698	void kill_fasync(struct fasync_struct **fp, int sig, int band)
	699	{
	700	/* First a quick test without locking: usually
	701	* the list is empty.
	702	*/
	703	if (*fp) {
	704	read_lock(&fasync_lock);
	705	/* reread fp after obtaining the lock /
	706	__kill_fasync(*fp, sig, band);
	707	read_unlock(&fasync_lock);
	708	}
	709	}
	710	EXPORT_SYMBOL(kill_fasync);
	711
	712	static int __init fasync_init(void)
	713	{
	714	fasync_cache = kmem_cache_create("fasync_cache",
20c2df83	715	sizeof(struct fasync_struct), 0, SLAB_PANIC, NULL);
1da177e4 LT	716	return 0;
	717	}
	718
	719	module_init(fasync_init)